91.9%), GSM8K (92% ->95.5%), AQuA (76.4% ->79.9%) and MATH (50.3% ->53.9%).","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new prompting method, named Progressive-Hint Prompting (PHP), that enables automatic multiple interactions between users and LLMs by using previously generated answers as hints to progressively guide toward the correct answers.'}",https://arxiv.org/pdf/2304.09797
-FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance,Lingjiao Chen,"There is a rapidly growing number of large language models (LLMs) that users can query for a fee. We review the cost associated with querying popular LLM APIs, e.g. GPT-4, ChatGPT, J1-Jumbo, and find that these models have heterogeneous pricing structures, with fees that can differ by two orders of magnitude. In particular, using LLMs on large collections of queries and text can be expensive. Motivated by this, we outline and discuss three types of strategies that users can exploit to reduce the inference cost associated with using LLMs: 1) prompt adaptation, 2) LLM approximation, and 3) LLM cascade. As an example, we propose FrugalGPT, a simple yet flexible instantiation of LLM cascade which learns which combinations of LLMs to use for different queries in order to reduce cost and improve accuracy. Our experiments show that FrugalGPT can match the performance of the best individual LLM (e.g. GPT-4) with up to 98% cost reduction or improve the accuracy over GPT-4 by 4% with the same cost. The ideas and findings presented here lay a foundation for using LLMs sustainably and efficiently.","{'model': 'tldr@v2.0.0', 'text': 'FrugalGPT is proposed, a simple yet flexible instantiation of LLM cascade which learns which combinations of LLMs to use for different queries in order to reduce cost and improve accuracy and lays a foundation for using LLMs sustainably and efficiently.'}",http://arxiv.org/pdf/2305.05176
-Reward Design with Language Models,Minae Kwon,"Reward design in reinforcement learning (RL) is challenging since specifying human notions of desired behavior may be difficult via reward functions or require many expert demonstrations. Can we instead cheaply design rewards using a natural language interface? This paper explores how to simplify reward design by prompting a large language model (LLM) such as GPT-3 as a proxy reward function, where the user provides a textual prompt containing a few examples (few-shot) or a description (zero-shot) of the desired behavior. Our approach leverages this proxy reward function in an RL framework. Specifically, users specify a prompt once at the beginning of training. During training, the LLM evaluates an RL agent's behavior against the desired behavior described by the prompt and outputs a corresponding reward signal. The RL agent then uses this reward to update its behavior. We evaluate whether our approach can train agents aligned with user objectives in the Ultimatum Game, matrix games, and the DealOrNoDeal negotiation task. In all three tasks, we show that RL agents trained with our framework are well-aligned with the user's objectives and outperform RL agents trained with reward functions learned via supervised learning","{'model': 'tldr@v2.0.0', 'text': 'This paper explores how to simplify reward design by prompting a large language model (LLM) such as GPT-3 as a proxy reward function, where the user provides a textual prompt containing a few examples or a description of the desired behavior.'}",http://arxiv.org/pdf/2303.00001
-Conversational Automated Program Repair,Chun Xia,"Automated Program Repair (APR) can help developers automatically generate patches for bugs. Due to the impressive performance obtained using Large Pre-Trained Language Models (LLMs) on many code related tasks, researchers have started to directly use LLMs for APR. However, prior approaches simply repeatedly sample the LLM given the same constructed input/prompt created from the original buggy code, which not only leads to generating the same incorrect patches repeatedly but also miss the critical information in testcases. To address these limitations, we propose conversational APR, a new paradigm for program repair that alternates between patch generation and validation in a conversational manner. In conversational APR, we iteratively build the input to the model by combining previously generated patches with validation feedback. As such, we leverage the long-term context window of LLMs to not only avoid generating previously incorrect patches but also incorporate validation feedback to help the model understand the semantic meaning of the program under test. We evaluate 10 different LLM including the newly developed ChatGPT model to demonstrate the improvement of conversational APR over the prior LLM for APR approach.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2301.13246
-Large Language Models as Optimizers,Chengrun Yang,"Optimization is ubiquitous. While derivative-based algorithms have been powerful tools for various problems, the absence of gradient imposes challenges on many real-world applications. In this work, we propose Optimization by PROmpting (OPRO), a simple and effective approach to leverage large language models (LLMs) as optimizers, where the optimization task is described in natural language. In each optimization step, the LLM generates new solutions from the prompt that contains previously generated solutions with their values, then the new solutions are evaluated and added to the prompt for the next optimization step. We first showcase OPRO on linear regression and traveling salesman problems, then move on to prompt optimization where the goal is to find instructions that maximize the task accuracy. With a variety of LLMs, we demonstrate that the best prompts optimized by OPRO outperform human-designed prompts by up to 8% on GSM8K, and by up to 50% on Big-Bench Hard tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Optimization by PROmpting (OPRO), a simple and effective approach to leverage large language models (LLMs) as optimizers, where the optimization task is described in natural language.'}",https://arxiv.org/pdf/2309.03409
-AnnoLLM: Making Large Language Models to Be Better Crowdsourced Annotators,Xingwei He,"Many natural language processing (NLP) tasks rely on labeled data to train machine learning models to achieve high performance. However, data annotation can be a time-consuming and expensive process, especially when the task involves a large amount of data or requires specialized domains. Recently, GPT-3.5 series models have demonstrated remarkable few-shot and zero-shot ability across various NLP tasks. In this paper, we first claim that large language models (LLMs), such as GPT-3.5, can serve as an excellent crowdsourced annotator by providing them with sufficient guidance and demonstrated examples. To make LLMs to be better annotators, we propose a two-step approach, 'explain-then-annotate'. To be more precise, we begin by creating prompts for every demonstrated example, which we subsequently utilize to prompt a LLM to provide an explanation for why the specific ground truth answer/label was chosen for that particular example. Following this, we construct the few-shot chain-of-thought prompt with the self-generated explanation and employ it to annotate the unlabeled data. We conduct experiments on three tasks, including user input and keyword relevance assessment, BoolQ and WiC. The annotation results from GPT-3.5 surpasses those from crowdsourced annotation for user input and keyword relevance assessment. Additionally, for the other two tasks, GPT-3.5 achieves results that are comparable to those obtained through crowdsourced annotation.","{'model': 'tldr@v2.0.0', 'text': 'This paper claims that large language models, such as GPT-3.5, can serve as an excellent crowdsourced annotator by providing them with sufficient guidance and demonstrated examples, and proposes a two-step approach, explain-then-annotate, to make LLMs to be better annotators.'}",http://arxiv.org/pdf/2303.16854
-Keep the Conversation Going: Fixing 162 out of 337 bugs for $0.42 each using ChatGPT,Chun Xia,"Automated Program Repair (APR) aims to automatically generate patches for buggy programs. Recent APR work has been focused on leveraging modern Large Language Models (LLMs) to directly generate patches for APR. Such LLM-based APR tools work by first constructing an input prompt built using the original buggy code and then queries the LLM to generate patches. While the LLM-based APR tools are able to achieve state-of-the-art results, it still follows the classic Generate and Validate repair paradigm of first generating lots of patches and then validating each one afterwards. This not only leads to many repeated patches that are incorrect but also miss the crucial information in test failures as well as in plausible patches. To address these limitations, we propose ChatRepair, the first fully automated conversation-driven APR approach that interleaves patch generation with instant feedback to perform APR in a conversational style. ChatRepair first feeds the LLM with relevant test failure information to start with, and then learns from both failures and successes of earlier patching attempts of the same bug for more powerful APR. For earlier patches that failed to pass all tests, we combine the incorrect patches with their corresponding relevant test failure information to construct a new prompt for the LLM to generate the next patch. In this way, we can avoid making the same mistakes. For earlier patches that passed all the tests, we further ask the LLM to generate alternative variations of the original plausible patches. In this way, we can further build on and learn from earlier successes to generate more plausible patches to increase the chance of having correct patches. While our approach is general, we implement ChatRepair using state-of-the-art dialogue-based LLM -- ChatGPT. By calculating the cost of accessing ChatGPT, we can fix 162 out of 337 bugs for \$0.42 each!","{'model': 'tldr@v2.0.0', 'text': 'ChatRepair is proposed, the first fully automated conversation-driven APR approach that interleaves patch generation with instant feedback to perform APR in a conversational style and is implemented using state-of-the-art dialogue-based LLM -- ChatGPT.'}",http://arxiv.org/pdf/2304.00385
-A Survey on Large Language Models for Recommendation,Likang Wu,"Large Language Models (LLMs) have emerged as powerful tools in the field of Natural Language Processing (NLP) and have recently gained significant attention in the domain of Recommendation Systems (RS). These models, trained on massive amounts of data using self-supervised learning, have demonstrated remarkable success in learning universal representations and have the potential to enhance various aspects of recommendation systems by some effective transfer techniques such as fine-tuning and prompt tuning, and so on. The crucial aspect of harnessing the power of language models in enhancing recommendation quality is the utilization of their high-quality representations of textual features and their extensive coverage of external knowledge to establish correlations between items and users. To provide a comprehensive understanding of the existing LLM-based recommendation systems, this survey presents a taxonomy that categorizes these models into two major paradigms, respectively Discriminative LLM for Recommendation (DLLM4Rec) and Generative LLM for Recommendation (GLLM4Rec), with the latter being systematically sorted out for the first time. Furthermore, we systematically review and analyze existing LLM-based recommendation systems within each paradigm, providing insights into their methodologies, techniques, and performance. Additionally, we identify key challenges and several valuable findings to provide researchers and practitioners with inspiration. We have also created a GitHub repository to index relevant papers on LLMs for recommendation, https://github.com/WLiK/LLM4Rec.","{'model': 'tldr@v2.0.0', 'text': 'This survey presents a taxonomy that categorizes existing LLM-based recommendation systems into two major paradigms, respectively Discriminative LLM for Recommendation (DLLM4Rec) and Generative LLL4Rec (GLLM 4Rec), with the latter being systematically sorted out for the first time.'}",https://arxiv.org/pdf/2305.19860
-Robots That Ask For Help: Uncertainty Alignment for Large Language Model Planners,Allen Z. Ren,"Large language models (LLMs) exhibit a wide range of promising capabilities -- from step-by-step planning to commonsense reasoning -- that may provide utility for robots, but remain prone to confidently hallucinated predictions. In this work, we present KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed. KnowNo builds on the theory of conformal prediction to provide statistical guarantees on task completion while minimizing human help in complex multi-step planning settings. Experiments across a variety of simulated and real robot setups that involve tasks with different modes of ambiguity (e.g., from spatial to numeric uncertainties, from human preferences to Winograd schemas) show that KnowNo performs favorably over modern baselines (which may involve ensembles or extensive prompt tuning) in terms of improving efficiency and autonomy, while providing formal assurances. KnowNo can be used with LLMs out of the box without model-finetuning, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models. Website: https://robot-help.github.io","{'model': 'tldr@v2.0.0', 'text': ""This work presents KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models.""}",https://arxiv.org/pdf/2307.01928
-Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models,Myra Cheng,"To recognize and mitigate harms from large language models (LLMs), we need to understand the prevalence and nuances of stereotypes in LLM outputs. Toward this end, we present Marked Personas, a prompt-based method to measure stereotypes in LLMs for intersectional demographic groups without any lexicon or data labeling.Grounded in the sociolinguistic concept of markedness (which characterizes explicitly linguistically marked categories versus unmarked defaults), our proposed method is twofold: 1) prompting an LLM to generate personas, i.e., natural language descriptions, of the target demographic group alongside personas of unmarked, default groups; 2) identifying the words that significantly distinguish personas of the target group from corresponding unmarked ones.We find that the portrayals generated by GPT-3.5 and GPT-4 contain higher rates of racial stereotypes than human-written portrayals using the same prompts. The words distinguishing personas of marked (non-white, non-male) groups reflect patterns of othering and exoticizing these demographics. An intersectional lens further reveals tropes that dominate portrayals of marginalized groups, such as tropicalism and the hypersexualization of minoritized women. These representational harms have concerning implications for downstream applications like story generation.",,http://arxiv.org/pdf/2305.18189
-Towards Measuring the Representation of Subjective Global Opinions in Language Models,Esin Durmus,"Large language models (LLMs) may not equitably represent diverse global perspectives on societal issues. In this paper, we develop a quantitative framework to evaluate whose opinions model-generated responses are more similar to. We first build a dataset, GlobalOpinionQA, comprised of questions and answers from cross-national surveys designed to capture diverse opinions on global issues across different countries. Next, we define a metric that quantifies the similarity between LLM-generated survey responses and human responses, conditioned on country. With our framework, we run three experiments on an LLM trained to be helpful, honest, and harmless with Constitutional AI. By default, LLM responses tend to be more similar to the opinions of certain populations, such as those from the USA, and some European and South American countries, highlighting the potential for biases. When we prompt the model to consider a particular country's perspective, responses shift to be more similar to the opinions of the prompted populations, but can reflect harmful cultural stereotypes. When we translate GlobalOpinionQA questions to a target language, the model's responses do not necessarily become the most similar to the opinions of speakers of those languages. We release our dataset for others to use and build on. Our data is at https://huggingface.co/datasets/Anthropic/llm_global_opinions. We also provide an interactive visualization at https://llmglobalvalues.anthropic.com.","{'model': 'tldr@v2.0.0', 'text': 'A quantitative framework to evaluate whose opinions model-generated responses are more similar to, and defines a metric that quantifies the similarity between LLM-generated survey responses and human responses, conditioned on country.'}",http://arxiv.org/pdf/2306.16388
-Supporting Qualitative Analysis with Large Language Models: Combining Codebook with GPT-3 for Deductive Coding,Ziang Xiao,"Qualitative analysis of textual contents unpacks rich and valuable information by assigning labels to the data. However, this process is often labor-intensive, particularly when working with large datasets. While recent AI-based tools demonstrate utility, researchers may not have readily available AI resources and expertise, let alone be challenged by the limited generalizability of those task-specific models. In this study, we explored the use of large language models (LLMs) in supporting deductive coding, a major category of qualitative analysis where researchers use pre-determined codebooks to label the data into a fixed set of codes. Instead of training task-specific models, a pre-trained LLM could be used directly for various tasks without fine-tuning through prompt learning. Using a curiosity-driven questions coding task as a case study, we found, by combining GPT-3 with expert-drafted codebooks, our proposed approach achieved fair to substantial agreements with expert-coded results. We lay out challenges and opportunities in using LLMs to support qualitative coding and beyond.","{'model': 'tldr@v2.0.0', 'text': 'This study explored the use of large language models (LLMs) in supporting deductive coding, a major category of qualitative analysis where researchers use pre-determined codebooks to label the data into a fixed set of codes, and found a pre-trained LLM could be used directly for various tasks without fine-tuning through prompt learning.'}",https://arxiv.org/pdf/2304.10548
-Assessment of chemistry knowledge in large language models that generate code,A. White,"In this work, we investigate the question: do code-generating large language models know chemistry? Our results indicate, mostly yes. To evaluate this, we introduce an expandable framework for evaluating chemistry knowledge in these models, through prompting models to solve chemistry problems posed as coding tasks. To do so, we produce a benchmark set of problems, and evaluate these models based on correctness of code by automated testing and evaluation by experts. We find that recent LLMs are able to write correct code across a variety of topics in chemistry and their accuracy can be increased by 30 percentage points via prompt engineering strategies, like putting copyright notices at the top of files. Our dataset and evaluation tools are open source which can be contributed to or built upon by future researchers, and will serve as a community resource for evaluating the performance of new models as they emerge. We also describe some good practices for employing LLMs in chemistry. The general success of these models demonstrates that their impact on chemistry teaching and research is poised to be enormous.","{'model': 'tldr@v2.0.0', 'text': 'It is found that recent LLMs are able to write correct code across a variety of topics in chemistry and their accuracy can be increased by 30 percentage points via prompt engineering strategies, like putting copyright notices at the top of files.'}",https://pubs.rsc.org/en/content/articlepdf/2023/dd/d2dd00087c
-Adaptive Machine Translation with Large Language Models,Yasmin Moslem,"Consistency is a key requirement of high-quality translation. It is especially important to adhere to pre-approved terminology and adapt to corrected translations in domain-specific projects. Machine translation (MT) has achieved significant progress in the area of domain adaptation. However, real-time adaptation remains challenging. Large-scale language models (LLMs) have recently shown interesting capabilities of in-context learning, where they learn to replicate certain input-output text generation patterns, without further fine-tuning. By feeding an LLM at inference time with a prompt that consists of a list of translation pairs, it can then simulate the domain and style characteristics. This work aims to investigate how we can utilize in-context learning to improve real-time adaptive MT. Our extensive experiments show promising results at translation time. For example, GPT-3.5 can adapt to a set of in-domain sentence pairs and/or terminology while translating a new sentence. We observe that the translation quality with few-shot in-context learning can surpass that of strong encoder-decoder MT systems, especially for high-resource languages. Moreover, we investigate whether we can combine MT from strong encoder-decoder models with fuzzy matches, which can further improve translation quality, especially for less supported languages. We conduct our experiments across five diverse language pairs, namely English-to-Arabic (EN-AR), English-to-Chinese (EN-ZH), English-to-French (EN-FR), English-to-Kinyarwanda (EN-RW), and English-to-Spanish (EN-ES).","{'model': 'tldr@v2.0.0', 'text': 'It is observed that the translation quality with few-shot in-context learning can surpass that of strong encoder-decoder MT systems, especially for high-resource languages.'}",http://arxiv.org/pdf/2301.13294
-Accuracy of Information and References Using ChatGPT-3 for Retrieval of Clinical Radiological Information.,M. Wagner,"Purpose: To assess the accuracy of answers provided by ChatGPT-3 when prompted with questions from the daily routine of radiologists and to evaluate the text response when ChatGPT-3 was prompted to provide references for a given answer. Methods: ChatGPT-3 (San Francisco, OpenAI) is an artificial intelligence chatbot based on a large language model (LLM) that has been designed to generate human-like text. A total of 88 questions were submitted to ChatGPT-3 using textual prompt. These 88 questions were equally dispersed across 8 subspecialty areas of radiology. The responses provided by ChatGPT-3 were assessed for correctness by cross-checking them with peer-reviewed, PubMed-listed references. In addition, the references provided by ChatGPT-3 were evaluated for authenticity. Results: A total of 59 of 88 responses (67%) to radiological questions were correct, while 29 responses (33%) had errors. Out of 343 references provided, only 124 references (36.2%) were available through internet search, while 219 references (63.8%) appeared to be generated by ChatGPT-3. When examining the 124 identified references, only 47 references (37.9%) were considered to provide enough background to correctly answer 24 questions (37.5%). Conclusion: In this pilot study, ChatGPT-3 provided correct responses to questions from the daily clinical routine of radiologists in only about two thirds, while the remainder of responses contained errors. The majority of provided references were not found and only a minority of the provided references contained the correct information to answer the question. Caution is advised when using ChatGPT-3 to retrieve radiological information.","{'model': 'tldr@v2.0.0', 'text': 'In this pilot study, ChatGPT-3 provided correct responses to questions from the daily clinical routine of radiologists in only about two thirds, while the remainder of responses contained errors.'}",
-In-Context Impersonation Reveals Large Language Models' Strengths and Biases,Leonard Salewski,"In everyday conversations, humans can take on different roles and adapt their vocabulary to their chosen roles. We explore whether LLMs can take on, that is impersonate, different roles when they generate text in-context. We ask LLMs to assume different personas before solving vision and language tasks. We do this by prefixing the prompt with a persona that is associated either with a social identity or domain expertise. In a multi-armed bandit task, we find that LLMs pretending to be children of different ages recover human-like developmental stages of exploration. In a language-based reasoning task, we find that LLMs impersonating domain experts perform better than LLMs impersonating non-domain experts. Finally, we test whether LLMs' impersonations are complementary to visual information when describing different categories. We find that impersonation can improve performance: an LLM prompted to be a bird expert describes birds better than one prompted to be a car expert. However, impersonation can also uncover LLMs' biases: an LLM prompted to be a man describes cars better than one prompted to be a woman. These findings demonstrate that LLMs are capable of taking on diverse roles and that this in-context impersonation can be used to uncover their hidden strengths and biases.",,http://arxiv.org/pdf/2305.14930
-kNN Prompting: Beyond-Context Learning with Calibration-Free Nearest Neighbor Inference,Benfeng Xu,"In-Context Learning (ICL), which formulates target tasks as prompt completion conditioned on in-context demonstrations, has become the prevailing utilization of LLMs. In this paper, we first disclose an actual predicament for this typical usage that it can not scale up with training data due to context length restriction. Besides, existing works have shown that ICL also suffers from various biases and requires delicate calibration treatment. To address both challenges, we advocate a simple and effective solution, $k$NN Prompting, which first queries LLM with training data for distributed representations, then predicts test instances by simply referring to nearest neighbors. We conduct comprehensive experiments to demonstrate its two-fold superiority: 1) Calibration-Free: $k$NN Prompting does not directly align LLM output distribution with task-specific label space, instead leverages such distribution to align test and training instances. It significantly outperforms state-of-the-art calibration-based methods under comparable few-shot scenario. 2) Beyond-Context: $k$NN Prompting can further scale up effectively with as many training data as are available, continually bringing substantial improvements. The scaling trend holds across 10 orders of magnitude ranging from 2 shots to 1024 shots as well as different LLMs scales ranging from 0.8B to 30B. It successfully bridges data scaling into model scaling, and brings new potentials for the gradient-free paradigm of LLM deployment. Code is publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper advocates a simple and effective solution, $k$NN Prompting, which first queries LLM with training data for distributed representations, then predicts test instances by simply referring to nearest neighbors, and significantly outperforms state-of-the-art calibration-based methods under comparable few-shot scenario.'}",http://arxiv.org/pdf/2303.13824
-"On Second Thought, Let’s Not Think Step by Step! Bias and Toxicity in Zero-Shot Reasoning",Omar Shaikh,"Generating a Chain of Thought (CoT) has been shown to consistently improve large language model (LLM) performance on a wide range of NLP tasks. However, prior work has mainly focused on logical reasoning tasks (e.g. arithmetic, commonsense QA); it remains unclear whether improvements hold for more diverse types of reasoning, especially in socially situated contexts. Concretely, we perform a controlled evaluation of zero-shot CoT across two socially sensitive domains: harmful questions and stereotype benchmarks. We find that zero-shot CoT reasoning in sensitive domains significantly increases a model’s likelihood to produce harmful or undesirable output, with trends holding across different prompt formats and model variants. Furthermore, we show that harmful CoTs increase with model size, but decrease with improved instruction following. Our work suggests that zero-shot CoT should be used with caution on socially important tasks, especially when marginalized groups or sensitive topics are involved.","{'model': 'tldr@v2.0.0', 'text': 'It is found that zero-shot CoT reasoning in sensitive domains significantly increases a model’s likelihood to produce harmful or undesirable output, with trends holding across different prompt formats and model variants.'}",http://arxiv.org/pdf/2212.08061
-AdaPlanner: Adaptive Planning from Feedback with Language Models,Haotian Sun,"Large language models (LLMs) have recently demonstrated the potential in acting as autonomous agents for sequential decision-making tasks. However, most existing methods either take actions greedily without planning or rely on static plans that are not adaptable to environmental feedback. Consequently, the sequential decision-making performance of LLM agents degenerates with problem complexity and plan horizons increase. We propose a closed-loop approach, AdaPlanner, which allows the LLM agent to refine its self-generated plan adaptively in response to environmental feedback. In AdaPlanner, the LLM agent adaptively refines its plan from feedback with both in-plan and out-of-plan refinement strategies. To mitigate hallucination, we develop a code-style LLM prompt structure that facilitates plan generation across a variety of tasks, environments, and agent capabilities. Furthermore, we propose a skill discovery mechanism that leverages successful plans as few-shot exemplars, enabling the agent to plan and refine with fewer task demonstrations. Our experiments in the ALFWorld and MiniWoB++ environments demonstrate that AdaPlanner outperforms state-of-the-art baselines by 3.73% and 4.11% while utilizing 2x and 600x fewer samples, respectively.","{'model': 'tldr@v2.0.0', 'text': 'A closed-loop approach, AdaPlanner, is proposed, which allows the LLM agent to refine its self-generated plan adaptively in response to environmental feedback, and develops a code-style LLM prompt structure that facilitates plan generation across a variety of tasks, environments, and agent capabilities.'}",http://arxiv.org/pdf/2305.16653
-Evaluation of ChatGPT for NLP-based Mental Health Applications,Bishal Lamichhane,"Large language models (LLM) have been successful in several natural language understanding tasks and could be relevant for natural language processing (NLP)-based mental health application research. In this work, we report the performance of LLM-based ChatGPT (with gpt-3.5-turbo backend) in three text-based mental health classification tasks: stress detection (2-class classification), depression detection (2-class classification), and suicidality detection (5-class classification). We obtained annotated social media posts for the three classification tasks from public datasets. Then ChatGPT API classified the social media posts with an input prompt for classification. We obtained F1 scores of 0.73, 0.86, and 0.37 for stress detection, depression detection, and suicidality detection, respectively. A baseline model that always predicted the dominant class resulted in F1 scores of 0.35, 0.60, and 0.19. The zero-shot classification accuracy obtained with ChatGPT indicates a potential use of language models for mental health classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'The zero-shot classification accuracy obtained with ChatGPT indicates a potential use of language models for mental health classification tasks, and large language models have been successful in several natural language understanding tasks.'}",http://arxiv.org/pdf/2303.15727
-Exploiting Asymmetry for Synthetic Training Data Generation: SynthIE and the Case of Information Extraction,Martin Josifoski,"Large language models (LLMs) have great potential for synthetic data generation. This work shows that useful data can be synthetically generated even for tasks that cannot be solved directly by LLMs: for problems with structured outputs, it is possible to prompt an LLM to perform the task in the reverse direction, by generating plausible input text for a target output structure. Leveraging this asymmetry in task difficulty makes it possible to produce large-scale, high-quality data for complex tasks. We demonstrate the effectiveness of this approach on closed information extraction, where collecting ground-truth data is challenging, and no satisfactory dataset exists to date. We synthetically generate a dataset of 1.8M data points, establish its superior quality compared to existing datasets in a human evaluation, and use it to finetune small models (220M and 770M parameters), termed SynthIE, that outperform the prior state of the art (with equal model size) by a substantial margin of 57 absolute points in micro-F1 and 79 points in macro-F1. Code, data, and models are available at https://github.com/epfl-dlab/SynthIE.","{'model': 'tldr@v2.0.0', 'text': 'This work synthetically generates a dataset of 1.8M data points, establishes its superior quality compared to existing datasets in a human evaluation, and uses it to finetune small models, termed SynthIE, that outperform the prior state of the art by a substantial margin.'}",http://arxiv.org/pdf/2303.04132
-Scalable 3D Captioning with Pretrained Models,Tiange Luo,"We introduce Cap3D, an automatic approach for generating descriptive text for 3D objects. This approach utilizes pretrained models from image captioning, image-text alignment, and LLM to consolidate captions from multiple views of a 3D asset, completely side-stepping the time-consuming and costly process of manual annotation. We apply Cap3D to the recently introduced large-scale 3D dataset, Objaverse, resulting in 660k 3D-text pairs. Our evaluation, conducted using 41k human annotations from the same dataset, demonstrates that Cap3D surpasses human-authored descriptions in terms of quality, cost, and speed. Through effective prompt engineering, Cap3D rivals human performance in generating geometric descriptions on 17k collected annotations from the ABO dataset. Finally, we finetune Text-to-3D models on Cap3D and human captions, and show Cap3D outperforms; and benchmark the SOTA including Point-E, Shape-E, and DreamFusion.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Cap3D, an automatic approach for generating descriptive text for 3D objects that surpasses human-authored descriptions in terms of quality, cost, and speed, and applies it to the recently introduced large-scale 3D dataset, Objaverse.'}",http://arxiv.org/pdf/2306.07279
-LLaVAR: Enhanced Visual Instruction Tuning for Text-Rich Image Understanding,Yanzhe Zhang,"Instruction tuning unlocks the superior capability of Large Language Models (LLM) to interact with humans. Furthermore, recent instruction-following datasets include images as visual inputs, collecting responses for image-based instructions. However, visual instruction-tuned models cannot comprehend textual details within images well. This work enhances the current visual instruction tuning pipeline with text-rich images (e.g., movie posters, book covers, etc.). Specifically, we first use publicly available OCR tools to collect results on 422K text-rich images from the LAION dataset. Moreover, we prompt text-only GPT-4 with recognized texts and image captions to generate 16K conversations, each containing question-answer pairs for text-rich images. By combining our collected data with previous multi-modal instruction-following data, our model, LLaVAR, substantially improves the LLaVA model's capability on text-based VQA datasets (up to 20% accuracy improvement) while achieving an accuracy of 91.42% on ScienceQA. The GPT-4-based instruction-following evaluation also demonstrates the improvement of our model on both natural images and text-rich images. Through qualitative analysis, LLaVAR shows promising interaction (e.g., reasoning, writing, and elaboration) skills with humans based on the latest real-world online content that combines text and images. We make our code/data/models publicly available at https://llavar.github.io/.","{'model': 'tldr@v2.0.0', 'text': ""This work enhances the current visual instruction tuning pipeline with text-rich images by substantially improves the LLaVA model's capability on text-based VQA datasets and shows promising interaction skills with humans based on the latest real-world online content that combines text and images.""}",http://arxiv.org/pdf/2306.17107
-Guiding Large Language Models via Directional Stimulus Prompting,Zekun Li,"We introduce Directional Stimulus Prompting, a novel framework for guiding black-box large language models (LLMs) toward specific desired outputs. Instead of directly adjusting LLMs, our method employs a small tunable policy model (e.g., T5) to generate an auxiliary directional stimulus prompt for each input instance. These directional stimulus prompts act as nuanced, instance-specific hints and clues to guide LLMs in generating desired outcomes, such as including specific keywords in the generated summary. Our approach sidesteps the challenges of direct LLM tuning by optimizing the policy model to explore directional stimulus prompts that align LLMs with desired behaviors. The policy model can be optimized through 1) supervised fine-tuning using labeled data and 2) reinforcement learning from offline or online rewards based on the LLM's output. We assess our method across summarization, dialogue response generation, and chain-of-thought reasoning tasks. Our experiments demonstrate that the framework consistently improves LLMs' (e.g., ChatGPT, Codex, InstructGPT) performance on these supervised tasks using minimal labeled data. Notably, using just 80 dialogues on the MultiWOZ dataset, our approach enhances ChatGPT's performance by an impressive 41.4%, matching or surpassing some fully supervised start-of-the-art models. Additionally, the instance-specific chain-of-thought prompt generated by our approach improves InstructGPT's reasoning accuracy compared to human-crafted or automatically generated prompts. The code and data are publicly available at \url{https://github.com/Leezekun/Directional-Stimulus-Prompting}.","{'model': 'tldr@v2.0.0', 'text': 'Directional Stimulus Prompting, a novel framework for guiding black-box large language models (LLMs) toward specific desired outputs, sidesteps the challenges of direct LLM tuning by optimizing the policy model to explore directional stimulus prompts that align LLMs with desired behaviors.'}",https://arxiv.org/pdf/2302.11520
-MotionGPT: Finetuned LLMs are General-Purpose Motion Generators,Yaqi Zhang,"Generating realistic human motion from given action descriptions has experienced significant advancements because of the emerging requirement of digital humans. While recent works have achieved impressive results in generating motion directly from textual action descriptions, they often support only a single modality of the control signal, which limits their application in the real digital human industry. This paper presents a Motion General-Purpose generaTor (MotionGPT) that can use multimodal control signals, e.g., text and single-frame poses, for generating consecutive human motions by treating multimodal signals as special input tokens in large language models (LLMs). Specifically, we first quantize multimodal control signals into discrete codes and then formulate them in a unified prompt instruction to ask the LLMs to generate the motion answer. Our MotionGPT demonstrates a unified human motion generation model with multimodal control signals by tuning a mere 0.4% of LLM parameters. To the best of our knowledge, MotionGPT is the first method to generate human motion by multimodal control signals, which we hope can shed light on this new direction. Codes shall be released upon acceptance.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a Motion General-Purpose generaTor (MotionGPT) that can use multimodal control signals, e.g., text and single-frame poses, for generating consecutive human motions by treating multimmodal signals as special input tokens in large language models (LLMs).'}",http://arxiv.org/pdf/2306.10900
-RoCo: Dialectic Multi-Robot Collaboration with Large Language Models,Zhao Mandi,"We propose a novel approach to multi-robot collaboration that harnesses the power of pre-trained large language models (LLMs) for both high-level communication and low-level path planning. Robots are equipped with LLMs to discuss and collectively reason task strategies. They then generate sub-task plans and task space waypoint paths, which are used by a multi-arm motion planner to accelerate trajectory planning. We also provide feedback from the environment, such as collision checking, and prompt the LLM agents to improve their plan and waypoints in-context. For evaluation, we introduce RoCoBench, a 6-task benchmark covering a wide range of multi-robot collaboration scenarios, accompanied by a text-only dataset for agent representation and reasoning. We experimentally demonstrate the effectiveness of our approach -- it achieves high success rates across all tasks in RoCoBench and adapts to variations in task semantics. Our dialog setup offers high interpretability and flexibility -- in real world experiments, we show RoCo easily incorporates human-in-the-loop, where a user can communicate and collaborate with a robot agent to complete tasks together. See project website https://project-roco.github.io for videos and code.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach to multi-robot collaboration that harnesses the power of pre-trained large language models (LLMs) for both high-level communication and low-level path planning, which shows RoCo easily incorporates human-in-the-loop, where a user can communicate and collaborate with a robot agent to complete tasks together.'}",https://arxiv.org/pdf/2307.04738
-UP5: Unbiased Foundation Model for Fairness-aware Recommendation,Wenyue Hua,"Recent advancements in foundation models such as large language models (LLM) have propelled them to the forefront of recommender systems (RS). Moreover, fairness in RS is critical since many users apply it for decision-making and demand fulfillment. However, at present, there is a lack of understanding regarding the level of fairness exhibited by recommendation foundation models and the appropriate methods for equitably treating different groups of users in foundation models. In this paper, we focus on user-side unfairness problem and show through a thorough examination that there is unfairness involved in LLMs that lead to unfair recommendation results. To eliminate bias from LLM for fairness-aware recommendation, we introduce a novel Unbiased P5 (UP5) foundation model based on Counterfactually-Fair-Prompting (CFP) techniques. CFP includes two sub-modules: a personalized prefix prompt that enhances fairness with respect to individual sensitive attributes, and a Prompt Mixture that integrates multiple counterfactually-fair prompts for a set of sensitive attributes. Experiments are conducted on two real-world datasets, MovieLens-1M and Insurance, and results are compared with both matching-based and sequential-based fairness-aware recommendation models. The results show that UP5 achieves better recommendation performance and meanwhile exhibits a high level of fairness.","{'model': 'tldr@v2.0.0', 'text': 'To eliminate bias from LLM for fairness-aware recommendation, a novel Unbiased P5 (UP5) foundation model based on Counterfactually-Fair-Prompting (CFP) techniques is introduced, which achieves better recommendation performance and meanwhile exhibits a high level of fairness.'}",http://arxiv.org/pdf/2305.12090
-Fill in the Blank: Context-aware Automated Text Input Generation for Mobile GUI Testing,Zhe Liu,"Automated GUI testing is widely used to help ensure the quality of mobile apps. However, many GUIs require appropriate text inputs to proceed to the next page, which remains a prominent obstacle for testing coverage. Considering the diversity and semantic requirement of valid inputs (e.g., flight departure, movie name), it is challenging to automate the text input generation. Inspired by the fact that the pre-trained Large Language Model (LLM) has made outstanding progress in text generation, we propose an approach named QTypist based on LLM for intelligently generating semantic input text according to the GUI context. To boost the performance of LLM in the mobile testing scenario, we develop a prompt-based data construction and tuning method which automatically extracts the prompts and answers for model tuning. We evaluate QTypist on 106 apps from Google Play, and the result shows that the passing rate of QTypist is 87%, which is 93% higher than the best baseline. We also integrate QTypist with the automated GUI testing tools and it can cover 42% more app activities, 52% more pages, and subsequently help reveal 122% more bugs compared with the raw tool.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an approach named QTypist based on LLM for intelligently generating semantic input text according to the GUI context which can cover 42% more app activities, 52% more pages, and subsequently help reveal 122% more bugs compared with the raw tool.'}",https://arxiv.org/pdf/2212.04732
-Explaining Patterns in Data with Language Models via Interpretable Autoprompting,Chandan Singh,"Large language models (LLMs) have displayed an impressive ability to harness natural language to perform complex tasks. In this work, we explore whether we can leverage this learned ability to find and explain patterns in data. Specifically, given a pre-trained LLM and data examples, we introduce interpretable autoprompting (iPrompt), an algorithm that generates a natural-language string explaining the data. iPrompt iteratively alternates between generating explanations with an LLM and reranking them based on their performance when used as a prompt. Experiments on a wide range of datasets, from synthetic mathematics to natural-language understanding, show that iPrompt can yield meaningful insights by accurately finding groundtruth dataset descriptions. Moreover, the prompts produced by iPrompt are simultaneously human-interpretable and highly effective for generalization: on real-world sentiment classification datasets, iPrompt produces prompts that match or even improve upon human-written prompts for GPT-3. Finally, experiments with an fMRI dataset show the potential for iPrompt to aid in scientific discovery.1","{'model': 'tldr@v2.0.0', 'text': 'This work introduces interpretable autoprompting (iPrompt), an algorithm that generates a natural-language string explaining the data that iteratively alternates between generating explanations with an LLM and reranking them based on their performance when used as a prompt.'}",http://arxiv.org/pdf/2210.01848
-InstructZero: Efficient Instruction Optimization for Black-Box Large Language Models,Lichang Chen,"Large language models~(LLMs) are instruction followers, but it can be challenging to find the best instruction for different situations, especially for black-box LLMs on which backpropagation is forbidden. Instead of directly optimizing the discrete instruction, we optimize a low-dimensional soft prompt applied to an open-source LLM to generate the instruction for the black-box LLM. On each iteration of the proposed method, which we call InstructZero, a soft prompt is converted into an instruction using the open-source LLM, which is then submitted to the black-box LLM for zero-shot evaluation, and the performance is sent to Bayesian optimization to produce new soft prompts improving the zero-shot performance. We evaluate InstructZero on different combinations of open-source LLMs and APIs including Vicuna and ChatGPT. Our results show that InstructZero outperforms SOTA auto-instruction methods across a variety of downstream tasks. Our code and data are publicly available at https://github.com/Lichang-Chen/InstructZero.","{'model': 'tldr@v2.0.0', 'text': 'This work optimized a low-dimensional soft prompt applied to an open-source LLM to generate the instruction for the black-box LLM, showing that InstructZero outperforms SOTA auto-instruction methods across a variety of downstream tasks.'}",https://arxiv.org/pdf/2306.03082
-Language Models Enable Simple Systems for Generating Structured Views of Heterogeneous Data Lakes,Simran Arora,"A long standing goal of the data management community is to develop general, automated systems that ingest semi-structured documents and output queryable tables without human effort or domain specific customization. Given the sheer variety of potential documents, state-of-the art systems make simplifying assumptions and use domain specific training. In this work, we ask whether we can maintain generality by using large language models (LLMs). LLMs, which are pretrained on broad data, can perform diverse downstream tasks simply conditioned on natural language task descriptions. We propose and evaluate EVAPORATE, a simple, prototype system powered by LLMs. We identify two fundamentally different strategies for implementing this system: prompt the LLM to directly extract values from documents or prompt the LLM to synthesize code that performs the extraction. Our evaluations show a cost-quality tradeoff between these two approaches. Code synthesis is cheap, but far less accurate than directly processing each document with the LLM. To improve quality while maintaining low cost, we propose an extended code synthesis implementation, EVAPORATE-CODE+, which achieves better quality than direct extraction. Our key insight is to generate many candidate functions and ensemble their extractions using weak supervision. EVAPORATE-CODE+ not only outperforms the state-of-the art systems, but does so using a sublinear pass over the documents with the LLM. This equates to a 110x reduction in the number of tokens the LLM needs to process, averaged across 16 real-world evaluation settings of 10k documents each.","{'model': 'tldr@v2.0.0', 'text': 'An extended code synthesis implementation, EVAPORATE-CODE+, is proposed, which achieves better quality than direct extraction and not only outperforms the state-of-the art systems, but does so using a sublinear pass over the documents with the LLM.'}",http://arxiv.org/pdf/2304.09433
-A New Era in Software Security: Towards Self-Healing Software via Large Language Models and Formal Verification,Yiannis Charalambous,"In this paper we present a novel solution that combines the capabilities of Large Language Models (LLMs) with Formal Verification strategies to verify and automatically repair software vulnerabilities. Initially, we employ Bounded Model Checking (BMC) to locate the software vulnerability and derive a counterexample. The counterexample provides evidence that the system behaves incorrectly or contains a vulnerability. The counterexample that has been detected, along with the source code, are provided to the LLM engine. Our approach involves establishing a specialized prompt language for conducting code debugging and generation to understand the vulnerability's root cause and repair the code. Finally, we use BMC to verify the corrected version of the code generated by the LLM. As a proof of concept, we create ESBMC-AI based on the Efficient SMT-based Context-Bounded Model Checker (ESBMC) and a pre-trained Transformer model, specifically gpt-3.5-turbo, to detect and fix errors in C programs. Our experimentation involved generating a dataset comprising 1000 C code samples, each consisting of 20 to 50 lines of code. Notably, our proposed method achieved an impressive success rate of up to 80% in repairing vulnerable code encompassing buffer overflow and pointer dereference failures. We assert that this automated approach can effectively incorporate into the software development lifecycle's continuous integration and deployment (CI/CD) process.","{'model': 'tldr@v2.0.0', 'text': ""A novel solution that combines the capabilities of Large Language Models (LLMs) with Formal Verification strategies to verify and automatically repair software vulnerabilities and it is asserted that this automated approach can effectively incorporate into the software development lifecycle's continuous integration and deployment (CI/CD) process.""}",http://arxiv.org/pdf/2305.14752
-RecurrentGPT: Interactive Generation of (Arbitrarily) Long Text,Wangchunshu Zhou,"The fixed-size context of Transformer makes GPT models incapable of generating arbitrarily long text. In this paper, we introduce RecurrentGPT, a language-based simulacrum of the recurrence mechanism in RNNs. RecurrentGPT is built upon a large language model (LLM) such as ChatGPT and uses natural language to simulate the Long Short-Term Memory mechanism in an LSTM. At each timestep, RecurrentGPT generates a paragraph of text and updates its language-based long-short term memory stored on the hard drive and the prompt, respectively. This recurrence mechanism enables RecurrentGPT to generate texts of arbitrary length without forgetting. Since human users can easily observe and edit the natural language memories, RecurrentGPT is interpretable and enables interactive generation of long text. RecurrentGPT is an initial step towards next-generation computer-assisted writing systems beyond local editing suggestions. In addition to producing AI-generated content (AIGC), we also demonstrate the possibility of using RecurrentGPT as an interactive fiction that directly interacts with consumers. We call this usage of generative models by ``AI As Contents'' (AIAC), which we believe is the next form of conventional AIGC. We further demonstrate the possibility of using RecurrentGPT to create personalized interactive fiction that directly interacts with readers instead of interacting with writers. More broadly, RecurrentGPT demonstrates the utility of borrowing ideas from popular model designs in cognitive science and deep learning for prompting LLMs. Our code is available at https://github.com/aiwaves-cn/RecurrentGPT and an online demo is available at https://www.aiwaves.org/recurrentgpt.","{'model': 'tldr@v2.0.0', 'text': 'RecurrentGPT is a language-based simulacrum of the recurrence mechanism in RNNs that uses natural language to simulate the Long Short-Term Memory mechanism in an LSTM and demonstrates the utility of borrowing ideas from popular model designs in cognitive science and deep learning for prompting LLMs.'}",http://arxiv.org/pdf/2305.13304
-PRD: Peer Rank and Discussion Improve Large Language Model based Evaluations,Ruosen Li,"Nowadays, the quality of responses generated by different modern large language models (LLMs) are hard to evaluate and compare automatically. Recent studies suggest and predominantly use LLMs as a reference-free metric for open-ended question answering. More specifically, they use the recognized""strongest""LLM as the evaluator, which conducts pairwise comparisons of candidate models' answers and provides a ranking score. However, this intuitive method has multiple problems, such as bringing in self-enhancement (favoring its own answers) and positional bias. We draw insights and lessons from the educational domain (Cho and MacArthur, 2011; Walsh, 2014) to improve LLM-based evaluations. Specifically, we propose the (1) peer rank (PR) algorithm that takes into account each peer LLM's pairwise preferences of all answer pairs, and outputs a final ranking of models; and (2) peer discussion (PD), where we prompt two LLMs to discuss and try to reach a mutual agreement on preferences of two answers. We conduct experiments on two benchmark datasets. We find that our approaches achieve higher accuracy and align better with human judgments, respectively. Interestingly, PR can induce a relatively accurate self-ranking of models under the anonymous setting, where each model's name is unrevealed. Our work provides space to explore evaluating models that are hard to compare for humans.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes the (1) peer rank (PR) algorithm that takes into account each peer LLM's pairwise preferences of all answer pairs, and outputs a final ranking of models; and (2) peer discussion (PD), where two LLMs are prompt to discuss and try to reach a mutual agreement on preferences of two answers.""}",https://arxiv.org/pdf/2307.02762
-GPT-3-driven pedagogical agents for training children's curious question-asking skills,Rania Abdelghani,,"{'model': 'tldr@v2.0.0', 'text': 'The efficiency of using large language model (LLM) to support children in generating more curious questions, using a natural language prompting approach that affords usability by teachers and other users not specialists of AI techniques is suggested.'}",https://arxiv.org/pdf/2211.14228
-Fixing Hardware Security Bugs with Large Language Models,Baleegh Ahmad,"Novel AI-based code-writing Large Language Models (LLMs) such as OpenAI's Codex have demonstrated capabilities in many coding-adjacent domains. In this work we consider how LLMs maybe leveraged to automatically repair security relevant bugs present in hardware designs. We focus on bug repair in code written in the Hardware Description Language Verilog. For this study we build a corpus of domain-representative hardware security bugs. We then design and implement a framework to quantitatively evaluate the performance of any LLM tasked with fixing the specified bugs. The framework supports design space exploration of prompts (i.e., prompt engineering) and identifying the best parameters for the LLM. We show that an ensemble of LLMs can repair all ten of our benchmarks. This ensemble outperforms the state-of-the-art Cirfix hardware bug repair tool on its own suite of bugs. These results show that LLMs can repair hardware security bugs and the framework is an important step towards the ultimate goal of an automated end-to-end bug repair framework.","{'model': 'tldr@v2.0.0', 'text': 'The results show that LLMs can repair hardware security bugs and the framework designed and implemented are an important step towards the ultimate goal of an automated end-to-end bug repair framework.'}",http://arxiv.org/pdf/2302.01215
-Red Teaming Language Model Detectors with Language Models,Zhouxing Shi,"The prevalence and strong capability of large language models (LLMs) present significant safety and ethical risks if exploited by malicious users. To prevent the potentially deceptive usage of LLMs, recent works have proposed algorithms to detect LLM-generated text and protect LLMs. In this paper, we investigate the robustness and reliability of these LLM detectors under adversarial attacks. We study two types of attack strategies: 1) replacing certain words in an LLM's output with their synonyms given the context; 2) automatically searching for an instructional prompt to alter the writing style of the generation. In both strategies, we leverage an auxiliary LLM to generate the word replacements or the instructional prompt. Different from previous works, we consider a challenging setting where the auxiliary LLM can also be protected by a detector. Experiments reveal that our attacks effectively compromise the performance of all detectors in the study with plausible generations, underscoring the urgent need to improve the robustness of LLM-generated text detection systems.","{'model': 'tldr@v2.0.0', 'text': 'Experiments reveal that attacks effectively compromise the performance of all detectors in the study with plausible generations, underscoring the urgent need to improve the robustness of LLM-generated text detection systems.'}",http://arxiv.org/pdf/2305.19713
-Open Sesame! Universal Black Box Jailbreaking of Large Language Models,Raz Lapid,"Large language models (LLMs), designed to provide helpful and safe responses, often rely on alignment techniques to align with user intent and social guidelines. Unfortunately, this alignment can be exploited by malicious actors seeking to manipulate an LLM's outputs for unintended purposes. In this paper we introduce a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible. The GA attack works by optimizing a universal adversarial prompt that -- when combined with a user's query -- disrupts the attacked model's alignment, resulting in unintended and potentially harmful outputs. Our novel approach systematically reveals a model's limitations and vulnerabilities by uncovering instances where its responses deviate from expected behavior. Through extensive experiments we demonstrate the efficacy of our technique, thus contributing to the ongoing discussion on responsible AI development by providing a diagnostic tool for evaluating and enhancing alignment of LLMs with human intent. To our knowledge this is the first automated universal black box jailbreak attack.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible, and is the first automated universal black box jailbreak attack.'}",https://arxiv.org/pdf/2309.01446
-What Language Reveals about Perception: Distilling Psychophysical Knowledge from Large Language Models,Raja Marjieh,"Understanding the extent to which the perceptual world can be recovered from language is a fundamental problem in cognitive science. We reformulate this problem as that of distilling psychophysical information from text and show how this can be done by combining large language models (LLMs) with a classic psychophysical method based on similarity judgments. Specifically, we use the prompt auto-completion functional- ity of GPT3, a state-of-the-art LLM, to elicit similarity scores between stimuli and then apply multidimensional scaling to uncover their underlying psychological space. We test our approach on six perceptual domains and show that the elicited judgments strongly correlate with human data and successfully recover well-known psychophysical structures such as the color wheel and pitch spiral. We also explore meaningful divergences between LLM and human representations. Our work showcases how combining state-of-the-art machine models with well-known cognitive paradigms can shed new light on fundamental questions in perception and language research.",,http://arxiv.org/pdf/2302.01308
-Large Language Models Can Be Used to Estimate the Latent Positions of Politicians,Patrick Y. Wu,"Existing approaches to estimating politicians' latent positions along specific dimensions often fail when relevant data is limited. We leverage the embedded knowledge in generative large language models (LLMs) to address this challenge and measure lawmakers' positions along specific political or policy dimensions. We prompt an instruction/dialogue-tuned LLM to pairwise compare lawmakers and then scale the resulting graph using the Bradley-Terry model. We estimate novel measures of U.S. senators' positions on liberal-conservative ideology, gun control, and abortion. Our liberal-conservative scale, used to validate LLM-driven scaling, strongly correlates with existing measures and offsets interpretive gaps, suggesting LLMs synthesize relevant data from internet and digitized media rather than memorizing existing measures. Our gun control and abortion measures -- the first of their kind -- differ from the liberal-conservative scale in face-valid ways and predict interest group ratings and legislator votes better than ideology alone. Our findings suggest LLMs hold promise for solving complex social science measurement problems.","{'model': 'tldr@v2.0.0', 'text': 'The liberal-conservative scale, used to validate LLM-driven scaling, strongly correlates with existing measures and offsets interpretive gaps, suggesting LLMs synthesize relevant data from internet and digitized media rather than memorizing existing measures.'}",
-Boosting Language Models Reasoning with Chain-of-Knowledge Prompting,J. Wang,"Recently, Chain-of-Thought (CoT) prompting has delivered success on complex reasoning tasks, which aims at designing a simple prompt like ``Let's think step by step'' or multiple in-context exemplars with well-designed rationales to elicit Large Language Models (LLMs) to generate intermediate reasoning steps. However, the generated rationales often come with mistakes, making unfactual and unfaithful reasoning chains. To mitigate this brittleness, we propose a novel Chain-of-Knowledge (CoK) prompting, where we aim at eliciting LLMs to generate explicit pieces of knowledge evidence in the form of structure triple. This is inspired by our human behaviors, i.e., we can draw a mind map or knowledge map as the reasoning evidence in the brain before answering a complex question. Benefiting from CoK, we additionally introduce a F^2-Verification method to estimate the reliability of the reasoning chains in terms of factuality and faithfulness. For the unreliable response, the wrong evidence can be indicated to prompt the LLM to rethink. Extensive experiments demonstrate that our method can further improve the performance of commonsense, factual, symbolic, and arithmetic reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Chain-of-Knowledge prompting, where it aims at eliciting LLMs to generate explicit pieces of knowledge evidence in the form of structure triple, and introduces a F^2-Verification method to estimate the reliability of the reasoning chains in terms of factuality and faithfulness.'}",https://arxiv.org/pdf/2306.06427
-Large language models will not replace healthcare professionals: curbing popular fears and hype,A. Thirunavukarasu,"Following the release of ChatGPT, large language models (LLMs) have entered the mainstream. ChatGPT and GPT-4 recently garnered particular attention for attaining expert-level performance in United States Medical Licensing Examinations. However, performance is not perfect, and has not been as impressive in more specialised tests, such as the Membership of the Royal College of General Practitioners Applied Knowledge Test. ChatGPT frequently ‘hallucinates’, providing false, unverified information in the same manner as which it delivers facts. While performance in clinical tasks is expected to improve dramatically with the release of GPT-4, remaining inaccuracy and lack of an uncertainty indicator preclude autonomous deployment of ChatGPT and LLM chatbots like it in clinical settings. LLM applications may nevertheless revolutionise cognitive work – tools such as ChatGPT excel in tasks where specialist knowledge is not required, or is provided by the user prompt: examples include correcting language and rephrasing information for different audiences or within other constraints (e.g. word limits), and it has already been proposed as a tool for administrative tasks, clinical work and patient education. While this does represent an impressive advance in natural language processing, and benefits may be manifold across fields including medicine, these limited use-cases do not live up to the hype surrounding LLMs and artificial intelligence (AI) more generally in 2023. This is due to a fundamental misunderstanding about the form of AI represented by LLMs. Do LLMs represent artificial generalised intelligence (AGI)? The answer is currently probably not, despite emergence of interactive conversational interfaces and few-shot or zero-shot properties – where models execute tasks that they have previously been exposed to only a few times before, or never before, respectively. This is demonstrated by observing how these models are trained, and the composition of their architecture. The backend LLM (GPT-3, from which GPT-3.5 was developed) underpinning older versions of ChatGPT was initially trained on a dataset of billions of words taken from books, Wikipedia and the wider internet. Through a process of machine learning, the GPT-3 accurately encoded the association between individual words in the training dataset. Through ‘reinforcement learning from human feedback’, GPT-3 was subsequently finetuned to provide appropriate responses to users’ queries – producing GPT-3.5. Through these processes, ChatGPT has developed an impressive ability to respond appropriately to diverse prompts, albeit equally lucidly with accurate and inaccurate statements. This lucidity, responsiveness and flexibility have led to sensational claims regarding attainment of AGI that could feasibly replace professionals in cognitive roles. The performance of GPT-4 – which powers newer versions of ChatGPT – dwarfs that of GPT-3.5 across tasks including logical reasoning and medical aptitude tests. Moreover, GPT-4 can be prompted to adopt different roles on demand, and will accept multimodal input, processing images as well as text. Prominent figures in industry and academia have advocated for a moratorium on development of more advanced AI systems in response to concerns regarding safety, ethics and fears of replacement. Despite these fears and hype, the barriers to implementation of LLMs replacing healthcare professionals in any capacity still look out of reach. Although GPT-4’s architecture and training are confidential, it likely relies on similar schemata to its predecessor as it exhibits similar (albeit fewer) hallucinations and reasoning errors, including in medicine. None of ChatGPT’s published autonomous training involved actual comprehension of language in context; the meaning (as we understand it) of words in the dataset was immaterial throughout. While this brute force linguistic processing may prove sufficient to develop a form of AGI, it appears that these LLMs will continue to be afflicted by mistakes and errors. Journal of the Royal Society of Medicine; 2023, Vol. 116(5) 181–182","{'model': 'tldr@v2.0.0', 'text': 'Following the release of ChatGPT, large language models (LLMs) have entered the mainstream and recently garnered particular attention for attaining expert-level performance in United States Medical Licensing Examinations, but performance has not been as impressive in more specialised tests.'}",https://journals.sagepub.com/doi/pdf/10.1177/01410768231173123
-Lion: Adversarial Distillation of Proprietary Large Language Models,Yuxin Jiang,"The practice of transferring knowledge from a sophisticated, proprietary large language model (LLM) to a compact, open-source LLM has garnered considerable attention. Previous works have focused on a unidirectional knowledge distillation way by aligning the responses of the student model with those of the teacher model to a set of instructions. Nevertheless, they overlooked the possibility of incorporating any reciprocal""feedback""--identifying challenging instructions where the student model's performance falls short--to boost the student model's proficiency iteratively. To this end, we propose a novel adversarial distillation framework for a more efficient knowledge transfer. Leveraging the versatile role adaptability of LLMs, we prompt the teacher model to identify""hard""instructions and generate new""hard""instructions for the student model, creating a three-stage adversarial loop of imitation, discrimination, and generation. By applying this adversarial framework, we successfully transfer knowledge from ChatGPT to a student model (named Lion), using a mere 70k training data. Our results show that Lion-13B not only achieves comparable open-ended generation capabilities to ChatGPT but surpasses conventional state-of-the-art (SOTA) instruction-tuned models like Vicuna-13B by 55.4% in challenging zero-shot reasoning benchmarks such as BIG-Bench Hard (BBH) and 16.7% on AGIEval. Code and model can be found at https://github.com/YJiangcm/Lion.","{'model': 'tldr@v2.0.0', 'text': 'Lion-13B not only achieves comparable open-ended generation capabilities to ChatGPT but surpasses conventional state-of-the-art instruction-tuned models like Vicuna- 13B by 55.4% in challenging zero-shot reasoning benchmarks such as BIG-Bench Hard (BBH) and 16.7% on AGIEval.'}",
-Surreal VR Pong: LLM approach to Game Design,Jasmine Roberts,"The increase in complexity from 2D to 3D game design makes it fascinating to study from a computational creativity perspective. Generating images given text descriptions using models like DALL-E has recently increased in popularity. However, these models are limited to generating 2-dimensional outputs. While outputs of these models can be used to stylize 3d objects with variable textures, they cannot produce mesh-level interactions. We introduce Codex VR Pong as a demonstration of controlled non-deterministic game mechanics leveraging generative models. We are proposing that prompt-based creation can become part of gameplay rather than just part of game development","{'model': 'tldr@v2.0.0', 'text': 'This work proposes that prompt-based creation can become part of gameplay rather than just part of game development, and introduces Codex VR Pong as a demonstration of controlled non-deterministic game mechanics leveraging generative models.'}",
-DePlot: One-shot visual language reasoning by plot-to-table translation,Fangyu Liu,"Visual language such as charts and plots is ubiquitous in the human world. Comprehending plots and charts requires strong reasoning skills. Prior state-of-the-art (SOTA) models require at least tens of thousands of training examples and their reasoning capabilities are still much limited, especially on complex human-written queries. This paper presents the first one-shot solution to visual language reasoning. We decompose the challenge of visual language reasoning into two steps: (1) plot-to-text translation, and (2) reasoning over the translated text. The key in this method is a modality conversion module, named as DePlot, which translates the image of a plot or chart to a linearized table. The output of DePlot can then be directly used to prompt a pretrained large language model (LLM), exploiting the few-shot reasoning capabilities of LLMs. To obtain DePlot, we standardize the plot-to-table task by establishing unified task formats and metrics, and train DePlot end-to-end on this task. DePlot can then be used off-the-shelf together with LLMs in a plug-and-play fashion. Compared with a SOTA model finetuned on more than>28k data points, DePlot+LLM with just one-shot prompting achieves a 24.0% improvement over finetuned SOTA on human-written queries from the task of chart QA.","{'model': 'tldr@v2.0.0', 'text': 'DePlot+LLM with just one-shot prompting achieves a 24.0% improvement over finetuned SOTA on human-written queries from the task of chart QA, and can be used off-the-shelf together with LLMs in a plug-and-play fashion.'}",http://arxiv.org/pdf/2212.10505
-Towards Parameter-Efficient Automation of Data Wrangling Tasks with Prefix-Tuning,David Vos,"Data wrangling tasks for data integration and cleaning arise in virtually every data-driven application scenario nowadays. Recent research indicated the astounding potential of Large Language Models (LLMs) for such tasks. However, the automation of data wrangling with LLMs poses additional challenges, as hand-tuning task- and data-specific prompts for LLMs requires high expertise and manual effort. On the other hand, finetuning a whole LLM is more amenable to automation, but incurs high storage costs, as a copy of the LLM has to be maintained. In this work, we explore the potential of a lightweight alternative to finetuning an LLM, which automatically learns a continuous prompt. This approach called prefix-tuning does not require updating the original LLM parameters, and can therefore re-use a single LLM instance across tasks. At the same time, it is amenable to automation, as continuous prompts can be automatically learned with standard techniques. We evaluate prefix-tuning on common data wrangling tasks for tabular data such as entity matching, error detection, and data imputation, with promising results. We find that in five out of ten cases, prefix-tuning is within 2.3% of the performance of finetuning, even though it leverages only 0.39% of the parameter updates required for finetuning the full model. These results highlight the potential of prefix-tuning as a parameter-efficient alternative to finetuning for data integration and data cleaning with LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the potential of a lightweight alternative to finetuning an LLM, which automatically learns a continuous prompt, called prefix-tuning, which does not require updating the original LLM parameters, and can therefore re-use a single LLM instance across tasks.'}",
-PromptNER: Prompting For Named Entity Recognition,D. Ashok,"In a surprising turn, Large Language Models (LLMs) together with a growing arsenal of prompt-based heuristics now offer powerful off-the-shelf approaches providing few-shot solutions to myriad classic NLP problems. However, despite promising early results, these LLM-based few-shot methods remain far from the state of the art in Named Entity Recognition (NER), where prevailing methods include learning representations via end-to-end structural understanding and fine-tuning on standard labeled corpora. In this paper, we introduce PromptNER, a new state-of-the-art algorithm for few-Shot and cross-domain NER. To adapt to any new NER task PromptNER requires a set of entity definitions in addition to the standard few-shot examples. Given a sentence, PromptNER prompts an LLM to produce a list of potential entities along with corresponding explanations justifying their compatibility with the provided entity type definitions. Remarkably, PromptNER achieves state-of-the-art performance on few-shot NER, achieving a 4% (absolute) improvement in F1 score on the ConLL dataset, a 9% (absolute) improvement on the GENIA dataset, and a 4% (absolute) improvement on the FewNERD dataset. PromptNER also moves the state of the art on Cross Domain NER, outperforming prior methods (including those not limited to the few-shot setting), setting a new mark on 3/5 CrossNER target domains, with an average F1 gain of 3%, despite using less than 2% of the available data.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces PromptNER, a new state-of-the-art algorithm for few-Shot and cross-domain NER, and prompts an LLM to produce a list of potential entities along with corresponding explanations justifying their compatibility with the provided entity type definitions.'}",http://arxiv.org/pdf/2305.15444
-Improving Few-Shot Prompts with Relevant Static Analysis Products,Toufique Ahmed,"Large Language Models (LLM) are a new class of computation engines,""programmed""via prompt engineering. We are still learning how to best""program""these LLMs to help developers. We start with the intuition that developers tend to consciously and unconsciously have a collection of semantics facts in mind when working on coding tasks. Mostly these are shallow, simple facts arising from a quick read. For a function, examples of facts might include parameter and local variable names, return expressions, simple pre- and post-conditions, and basic control and data flow, etc. One might assume that the powerful multi-layer architecture of transformer-style LLMs makes them inherently capable of doing this simple level of""code analysis""and extracting such information, implicitly, while processing code: but are they, really? If they aren't, could explicitly adding this information help? Our goal here is to investigate this question, using the code summarization task and evaluate whether automatically augmenting an LLM's prompt with semantic facts explicitly, actually helps. Prior work shows that LLM performance on code summarization benefits from few-shot samples drawn either from the same-project or from examples found via information retrieval methods (such as BM25). While summarization performance has steadily increased since the early days, there is still room for improvement: LLM performance on code summarization still lags its performance on natural-language tasks like translation and text summarization. We find that adding semantic facts actually does help! This approach improves performance in several different settings suggested by prior work, including for two different Large Language Models. In most cases, improvement nears or exceeds 2 BLEU; for the PHP language in the challenging CodeSearchNet dataset, this augmentation actually yields performance surpassing 30 BLEU.","{'model': 'tldr@v2.0.0', 'text': ""This work investigates whether automatically augmenting an LLM's prompt with semantic facts explicitly, actually helps, and finds that adding semantic facts actually does help!""}",https://arxiv.org/pdf/2304.06815
-Herding AI Cats: Lessons from Designing a Chatbot by Prompting GPT-3,J. Zamfirescu-Pereira,"Prompting Large Language Models (LLMs) is an exciting new approach to designing chatbots. But can it improve LLM’s user experience (UX) reliably enough to power chatbot products? Our attempt to design a robust chatbot by prompting GPT-3/4 alone suggests: not yet. Prompts made achieving “80%” UX goals easy, but not the remaining 20%. Fixing the few remaining interaction breakdowns resembled herding cats: We could not address one UX issue or test one design solution at a time; instead, we had to handle everything everywhere all at once. Moreover, because no prompt could make GPT reliably say “I don’t know” when it should, the user-GPT conversations had no guardrails after a breakdown occurred, often leading to UX downward spirals. These risks incentivized us to design highly prescriptive prompts and scripted bots, counter to the promises of LLM-powered chatbots. This paper describes this case study, unpacks prompting’s fickleness and its impact on UX design processes, and discusses implications for LLM-based design methods and tools.",,https://dl.acm.org/doi/pdf/10.1145/3563657.3596138
-MindAgent: Emergent Gaming Interaction,Ran Gong,"Large Language Models (LLMs) have the capacity of performing complex scheduling in a multi-agent system and can coordinate these agents into completing sophisticated tasks that require extensive collaboration. However, despite the introduction of numerous gaming frameworks, the community has insufficient benchmarks towards building general multi-agents collaboration infrastructure that encompass both LLM and human-NPCs collaborations. In this work, we propose a novel infrastructure - MindAgent - to evaluate planning and coordination emergent capabilities for gaming interaction. In particular, our infrastructure leverages existing gaming framework, to i) require understanding of the coordinator for a multi-agent system, ii) collaborate with human players via un-finetuned proper instructions, and iii) establish an in-context learning on few-shot prompt with feedback. Furthermore, we introduce CUISINEWORLD, a new gaming scenario and related benchmark that dispatch a multi-agent collaboration efficiency and supervise multiple agents playing the game simultaneously. We conduct comprehensive evaluations with new auto-metric CoS for calculating the collaboration efficiency. Finally, our infrastructure can be deployed into real-world gaming scenarios in a customized VR version of CUISINEWORLD and adapted in existing broader Minecraft gaming domain. We hope our findings on LLMs and the new infrastructure for general-purpose scheduling and coordination can help shed light on how such skills can be obtained by learning from large language corpora.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel infrastructure - MindAgent - to evaluate planning and coordination emergent capabilities for gaming interaction and introduces CUISINEWORLD, a new gaming scenario and related benchmark that dispatch a multi-agent collaboration efficiency and supervise multiple agents playing the game simultaneously.'}",https://arxiv.org/pdf/2309.09971
-Time Travel in LLMs: Tracing Data Contamination in Large Language Models,Shahriar Golchin,"Data contamination, i.e., the presence of test data from downstream tasks in the training data of large language models (LLMs), is a potential major issue in measuring LLMs' real effectiveness on other tasks. We propose a straightforward yet effective method for identifying data contamination within LLMs. At its core, our approach starts by identifying potential contamination at the instance level; using this information, our approach then assesses wider contamination at the partition level. To estimate contamination of individual instances, we employ""guided instruction:""a prompt consisting of the dataset name, partition type, and the random-length initial segment of a reference instance, asking the LLM to complete it. An instance is flagged as contaminated if the LLM's output either exactly or nearly matches the latter segment of the reference. To understand if an entire partition is contaminated, we propose two ideas. The first idea marks a dataset partition as contaminated if the average overlap score with the reference instances (as measured by ROUGE-L or BLEURT) is statistically significantly better with the completions from guided instruction compared to a""general instruction""that does not include the dataset and partition name. The second idea marks a dataset partition as contaminated if a classifier based on GPT-4 with few-shot in-context learning prompt marks multiple generated completions as exact/near-exact matches of the corresponding reference instances. Our best method achieves an accuracy between 92% and 100% in detecting if an LLM is contaminated with seven datasets, containing train and test/validation partitions, when contrasted with manual evaluation by human experts. Further, our findings indicate that GPT-4 is contaminated with AG News, WNLI, and XSum datasets.","{'model': 'tldr@v2.0.0', 'text': 'The best method achieves an accuracy between 92% and 100% in detecting if an LLM is contaminated with seven datasets, containing train and test/validation partitions, when contrasted with manual evaluation by human experts.'}",https://arxiv.org/pdf/2308.08493
-Exploring Large Language Model for Graph Data Understanding in Online Job Recommendations,Likang Wu,"Large Language Models (LLMs) have revolutionized natural language processing tasks, demonstrating their exceptional capabilities in various domains. However, their potential for behavior graph understanding in job recommendations remains largely unexplored. This paper focuses on unveiling the capability of large language models in understanding behavior graphs and leveraging this understanding to enhance recommendations in online recruitment, including the promotion of out-of-distribution (OOD) application. We present a novel framework that harnesses the rich contextual information and semantic representations provided by large language models to analyze behavior graphs and uncover underlying patterns and relationships. Specifically, we propose a meta-path prompt constructor that leverages LLM recommender to understand behavior graphs for the first time and design a corresponding path augmentation module to alleviate the prompt bias introduced by path-based sequence input. By leveraging this capability, our framework enables personalized and accurate job recommendations for individual users. We evaluate the effectiveness of our approach on a comprehensive dataset and demonstrate its ability to improve the relevance and quality of recommended quality. This research not only sheds light on the untapped potential of large language models but also provides valuable insights for developing advanced recommendation systems in the recruitment market. The findings contribute to the growing field of natural language processing and offer practical implications for enhancing job search experiences.","{'model': 'tldr@v2.0.0', 'text': 'A meta-path prompt constructor is proposed that leverages LLM recommender to understand behavior graphs for the first time and a corresponding path augmentation module is designed to alleviate the prompt bias introduced by path-based sequence input.'}",https://arxiv.org/pdf/2307.05722
-Fine-tuning Multimodal LLMs to Follow Zero-shot Demonstrative Instructions,Juncheng Li,"Recent advancements in Multimodal Large Language Models (MLLMs) have been utilizing Visual Prompt Generators (VPGs) to convert visual features into tokens that LLMs can recognize. This is achieved by training the VPGs on millions of image-caption pairs, where the VPG-generated tokens of images are fed into a frozen LLM to generate the corresponding captions. However, this image-captioning based training objective inherently biases the VPG to concentrate solely on the primary visual contents sufficient for caption generation, often neglecting other visual details. This shortcoming results in MLLMs' underperformance in comprehending demonstrative instructions consisting of multiple, interleaved, and multimodal instructions that demonstrate the required context to complete a task. To address this issue, we introduce a generic and lightweight Visual Prompt Generator Complete module (VPG-C), which can infer and complete the missing details essential for comprehending demonstrative instructions. Further, we propose a synthetic discriminative training strategy to fine-tune VPG-C, eliminating the need for supervised demonstrative instructions. As for evaluation, we build DEMON, a comprehensive benchmark for demonstrative instruction understanding. Synthetically trained with the proposed strategy, VPG-C achieves significantly stronger zero-shot performance across all tasks of DEMON. Further evaluation on the MME and OwlEval benchmarks also demonstrate the superiority of VPG-C. Our benchmark, code, and pre-trained models are available at https://github.com/DCDmllm/Cheetah.","{'model': 'tldr@v2.0.0', 'text': 'A generic and lightweight Visual Prompt Generator Complete module (VPG-C) is introduced, which can infer and complete the missing details essential for comprehending demonstrative instructions, and a synthetic discriminative training strategy is proposed to fine-tune V PG-C, eliminating the need for supervised demonstrative instruction understanding.'}",
-Ten Quick Tips for Harnessing the Power of ChatGPT/GPT-4 in Computational Biology,Tiago Lubiana,"The rise of advanced chatbots, such as ChatGPT, has sparked curiosity in the scientific community. ChatGPT is a general-purpose chatbot powered by large language models (LLMs) GPT-3.5 and GPT-4, with the potential to impact numerous fields, including computational biology. In this article, we offer ten tips based on our experience with ChatGPT to assist computational biologists in optimizing their workflows. We have collected relevant prompts and reviewed the nascent literature in the field, compiling tips we project to remain pertinent for future ChatGPT and LLM iterations, ranging from code refactoring to scientific writing to prompt engineering. We hope our work will help bioinformaticians to complement their workflows while staying aware of the various implications of using this technology. Additionally, to track new and creative applications for bioinformatics tools such as ChatGPT, we have established a GitHub repository at https://github.com/csbl-br/awesome-compbio-chatgpt. Our belief is that ethical adherence to ChatGPT and other LLMs will increase the efficiency of computational biologists, ultimately advancing the pace of scientific discovery in the life sciences.","{'model': 'tldr@v2.0.0', 'text': 'It is believed that ethical adherence to ChatGPT and other LLMs will increase the efficiency of computational biologists, ultimately advancing the pace of scientific discovery in the life sciences.'}",
-ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMs,Justin Chih-Yao Chen,"Large Language Models (LLMs) still struggle with complex reasoning tasks. Motivated by the society of minds (Minsky, 1988), we propose ReConcile, a multi-model multi-agent framework designed as a round table conference among diverse LLM agents to foster diverse thoughts and discussion for improved consensus. ReConcile enhances the reasoning capabilities of LLMs by holding multiple rounds of discussion, learning to convince other agents to improve their answers, and employing a confidence-weighted voting mechanism. In each round, ReConcile initiates discussion between agents via a 'discussion prompt' that consists of (a) grouped answers and explanations generated by each agent in the previous round, (b) their uncertainties, and (c) demonstrations of answer-rectifying human explanations, used for convincing other agents. This discussion prompt enables each agent to revise their responses in light of insights from other agents. Once a consensus is reached and the discussion ends, ReConcile determines the final answer by leveraging the confidence of each agent in a weighted voting scheme. We implement ReConcile with ChatGPT, Bard, and Claude2 as the three agents. Our experimental results on various benchmarks demonstrate that ReConcile significantly enhances the reasoning performance of the agents (both individually and as a team), surpassing prior single-agent and multi-agent baselines by 7.7% and also outperforming GPT-4 on some of these datasets. We also experiment with GPT-4 itself as one of the agents in ReConcile and demonstrate that its initial performance also improves by absolute 10.0% through discussion and feedback from other agents. Finally, we also analyze the accuracy after every round and observe that ReConcile achieves better and faster consensus between agents, compared to a multi-agent debate baseline. Our code is available at: https://github.com/dinobby/ReConcile","{'model': 'tldr@v2.0.0', 'text': 'ReConcile significantly enhances the reasoning performance of the agents (both individually and as a team), surpassing prior single-agent and multi-agent baselines by 7.7% and also outperforming GPT-4 on some of these datasets.'}",https://arxiv.org/pdf/2309.13007
-Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognition,Yuang Li,"The integration of Language Models (LMs) has proven to be an effective way to address domain shifts in speech recognition. However, these approaches usually require a significant amount of target domain text data for the training of LMs. Different from these methods, in this work, with only a domain-specific text prompt, we propose two zero-shot ASR domain adaptation methods using LLaMA, a 7-billion-parameter large language model (LLM). LLM is used in two ways: 1) second-pass rescoring: reranking N-best hypotheses of a given ASR system with LLaMA; 2) deep LLM-fusion: incorporating LLM into the decoder of an encoder-decoder based ASR system. Experiments show that, with only one domain prompt, both methods can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGISpeech datasets. Especially, the deep LLM-fusion has the advantage of better recall of entity and out-of-vocabulary words.","{'model': 'tldr@v2.0.0', 'text': 'Two zero-shot ASR domain adaptation methods using LLaMA, a 7-billion-parameter large language model (LLM), which can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGISpeech datasets.'}",http://arxiv.org/pdf/2306.16007
-Aligning Language Models to User Opinions,EunJeong Hwang,"An important aspect of developing LLMs that interact with humans is to align models' behavior to their users. It is possible to prompt an LLM into behaving as a certain persona, especially a user group or ideological persona the model captured during its pertaining stage. But, how to best align an LLM with a specific user and not a demographic or ideological group remains an open question. Mining public opinion surveys (by Pew Research), we find that the opinions of a user and their demographics and ideologies are not mutual predictors. We use this insight to align LLMs by modeling both user opinions as well as user demographics and ideology, achieving up to 7 points accuracy gains in predicting public opinions from survey questions across a broad set of topics. In addition to the typical approach of prompting LLMs with demographics and ideology, we discover that utilizing the most relevant past opinions from individual users enables the model to predict user opinions more accurately.","{'model': 'tldr@v2.0.0', 'text': 'It is discovered that utilizing the most relevant past opinions from individual users enables the model to predict user opinions more accurately, achieving up to 7 points accuracy gains in predicting public opinions from survey questions across a broad set of topics.'}",http://arxiv.org/pdf/2305.14929
-Prompting is not a substitute for probability measurements in large language models,Jennifer Hu,"Prompting is now a dominant method for evaluating the linguistic knowledge of large language models (LLMs). While other methods directly read out models' probability distributions over strings, prompting requires models to access this internal information by processing linguistic input, thereby implicitly testing a new type of emergent ability: metalinguistic judgment. In this study, we compare metalinguistic prompting and direct probability measurements as ways of measuring models' linguistic knowledge. Broadly, we find that LLMs' metalinguistic judgments are inferior to quantities directly derived from representations. Furthermore, consistency gets worse as the prompt query diverges from direct measurements of next-word probabilities. Our findings suggest that negative results relying on metalinguistic prompts cannot be taken as conclusive evidence that an LLM lacks a particular linguistic generalization. Our results also highlight the value that is lost with the move to closed APIs where access to probability distributions is limited.",,
-Automated Annotation with Generative AI Requires Validation,Nicholas Pangakis,"Generative large language models (LLMs) can be a powerful tool for augmenting text annotation procedures, but their performance varies across annotation tasks due to prompt quality, text data idiosyncrasies, and conceptual difficulty. Because these challenges will persist even as LLM technology improves, we argue that any automated annotation process using an LLM must validate the LLM's performance against labels generated by humans. To this end, we outline a workflow to harness the annotation potential of LLMs in a principled, efficient way. Using GPT-4, we validate this approach by replicating 27 annotation tasks across 11 datasets from recent social science articles in high-impact journals. We find that LLM performance for text annotation is promising but highly contingent on both the dataset and the type of annotation task, which reinforces the necessity to validate on a task-by-task basis. We make available easy-to-use software designed to implement our workflow and streamline the deployment of LLMs for automated annotation.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that any automated annotation process using an LLM must validate the LLM's performance against labels generated by humans, and an outline of a workflow to harness the annotation potential of LLMs in a principled, efficient way is outlined.""}",http://arxiv.org/pdf/2306.00176
-Statler: State-Maintaining Language Models for Embodied Reasoning,Takuma Yoneda,"Large language models (LLMs) provide a promising tool that enable robots to perform complex robot reasoning tasks. However, the limited context window of contemporary LLMs makes reasoning over long time horizons difficult. Embodied tasks such as those that one might expect a household robot to perform typically require that the planner consider information acquired a long time ago (e.g., properties of the many objects that the robot previously encountered in the environment). Attempts to capture the world state using an LLM's implicit internal representation is complicated by the paucity of task- and environment-relevant information available in a robot's action history, while methods that rely on the ability to convey information via the prompt to the LLM are subject to its limited context window. In this paper, we propose Statler, a framework that endows LLMs with an explicit representation of the world state as a form of ``memory'' that is maintained over time. Integral to Statler is its use of two instances of general LLMs -- a world-model reader and a world-model writer -- that interface with and maintain the world state. By providing access to this world state ``memory'', Statler improves the ability of existing LLMs to reason over longer time horizons without the constraint of context length. We evaluate the effectiveness of our approach on three simulated table-top manipulation domains and a real robot domain, and show that it improves the state-of-the-art in LLM-based robot reasoning. Project website: https://statler-lm.github.io/","{'model': 'tldr@v2.0.0', 'text': ""Statler is a framework that endows LLMs with an explicit representation of the world state as a form of ``memory'' that is maintained over time, and improves the ability of existing LLMs to reason over longer time horizons without the constraint of context length.""}",http://arxiv.org/pdf/2306.17840
-Federated Large Language Model: A Position Paper,Chaochao Chen,"Large scale language models (LLM) have received significant attention and found diverse applications across various domains, but their development encounters challenges in real-world scenarios. These challenges arise due to the scarcity of public domain data availability and the need to maintain privacy with respect to private domain data. To address these issues, federated learning (FL) has emerged as a promising technology that enables collaborative training of shared models while preserving decentralized data. We propose the concept of federated LLM, which comprises three key components, i.e., federated LLM pre-training, federated LLM fine-tuning, and federated LLM prompt engineering. For each component, we discuss its advantage over traditional LLM training methods and propose specific engineering strategies for implementation. Furthermore, we explore the novel challenges introduced by the integration of FL and LLM. We analyze existing solutions and identify potential obstacles faced by these solutions within the context of federated LLM.","{'model': 'tldr@v2.0.0', 'text': 'This work analyzes existing solutions and identifies potential obstacles faced by these solutions within the context of federated LLM, and explores the novel challenges introduced by the integration of FL and LLM.'}",https://arxiv.org/pdf/2307.08925
-StudentEval: A Benchmark of Student-Written Prompts for Large Language Models of Code,Hannah McLean Babe,"Code LLMs are being rapidly deployed and there is evidence that they can make professional programmers more productive. Current benchmarks for code generation measure whether models generate correct programs given an expert prompt. In this paper, we present a new benchmark containing multiple prompts per problem, written by a specific population of non-expert prompters: beginning programmers. StudentEval contains 1,749 prompts for 48 problems, written by 80 students who have only completed one semester of Python programming. Our students wrote these prompts while working interactively with a Code LLM, and we observed very mixed success rates. We use StudentEval to evaluate 5 Code LLMs and find that StudentEval is a better discriminator of model performance than existing benchmarks. We analyze the prompts and find significant variation in students' prompting techniques. We also find that nondeterministic LLM sampling could mislead students into thinking that their prompts are more (or less) effective than they actually are, which has implications for how to teach with Code LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new benchmark containing multiple prompts per problem, written by a specific population of non-expert prompters: beginning programmers, and finds that StudentEval is a better discriminator of model performance than existing benchmarks.'}",http://arxiv.org/pdf/2306.04556
-MindMap: Knowledge Graph Prompting Sparks Graph of Thoughts in Large Language Models,Yilin Wen,"LLMs usually exhibit limitations in their ability to incorporate new knowledge, the generation of hallucinations, and the transparency of their decision-making process. In this paper, we explore how to prompt LLMs with knowledge graphs (KG), working as a remedy to engage LLMs with up-to-date knowledge and elicit the reasoning pathways from LLMs. Specifically, we build a prompting pipeline that endows LLMs with the capability of comprehending KG inputs and inferring with a combined implicit knowledge and the retrieved external knowledge. In addition, we investigate eliciting the mind map on which LLMs perform the reasoning and generate the answers. It is identified that the produced mind map exhibits the reasoning pathways of LLMs grounded on the ontology of knowledge, hence bringing the prospects of probing and gauging LLM inference in production. The experiments on three question&answering datasets also show that MindMap prompting leads to a striking empirical gain. For instance, prompting a GPT-3.5 with MindMap yields an overwhelming performance over GPT-4 consistently. We also demonstrate that with structured facts retrieved from KG, MindMap can outperform a series of prompting-with-document-retrieval methods, benefiting from more accurate, concise, and comprehensive knowledge from KGs. To reproduce our results and extend the framework further, we make our codebase available at https://github.com/wyl.willing/MindMap.","{'model': 'tldr@v2.0.0', 'text': 'This paper builds a prompting pipeline that endows LLMs with the capability of comprehending KG inputs and inferring with a combined implicit knowledge and the retrieved external knowledge, and demonstrates that with structured facts retrieved from KG, MindMap can outperform a series of prompting-with-document-retrieval methods, benefiting from more accurate, concise, and comprehensive knowledge from KGs.'}",https://arxiv.org/pdf/2308.09729
-ProgPrompt: program generation for situated robot task planning using large language models,Ishika Singh,,"{'model': 'tldr@v2.0.0', 'text': 'This work presents a programmatic LLM prompt structure that enables plan generation functional across situated environments, robot capabilities, and tasks, and makes concrete recommendations about prompt structure and generation constraints through ablation experiments.'}",https://link.springer.com/content/pdf/10.1007/s10514-023-10135-3.pdf
-Interactive Data Synthesis for Systematic Vision Adaptation via LLMs-AIGCs Collaboration,Qifan Yu,"Recent text-to-image generation models have shown promising results in generating high-fidelity photo-realistic images. In parallel, the problem of data scarcity has brought a growing interest in employing AIGC technology for high-quality data expansion. However, this paradigm requires well-designed prompt engineering that cost-less data expansion and labeling remain under-explored. Inspired by LLM's powerful capability in task guidance, we propose a new paradigm of annotated data expansion named as ChatGenImage. The core idea behind it is to leverage the complementary strengths of diverse models to establish a highly effective and user-friendly pipeline for interactive data augmentation. In this work, we extensively study how LLMs communicate with AIGC model to achieve more controllable image generation and make the first attempt to collaborate them for automatic data augmentation for a variety of downstream tasks. Finally, we present fascinating results obtained from our ChatGenImage framework and demonstrate the powerful potential of our synthetic data for systematic vision adaptation. Our codes are available at https://github.com/Yuqifan1117/Labal-Anything-Pipeline.","{'model': 'tldr@v2.0.0', 'text': 'This work extensively study how LLMs communicate with AIGC model to achieve more controllable image generation and makes the first attempt to collaborate them for automatic data augmentation for a variety of downstream tasks.'}",http://arxiv.org/pdf/2305.12799
-Examining the Validity of ChatGPT in Identifying Relevant Nephrology Literature: Findings and Implications,S. Suppadungsuk,"Literature reviews are valuable for summarizing and evaluating the available evidence in various medical fields, including nephrology. However, identifying and exploring the potential sources requires focus and time devoted to literature searching for clinicians and researchers. ChatGPT is a novel artificial intelligence (AI) large language model (LLM) renowned for its exceptional ability to generate human-like responses across various tasks. However, whether ChatGPT can effectively assist medical professionals in identifying relevant literature is unclear. Therefore, this study aimed to assess the effectiveness of ChatGPT in identifying references to literature reviews in nephrology. We keyed the prompt “Please provide the references in Vancouver style and their links in recent literature on… name of the topic” into ChatGPT-3.5 (03/23 Version). We selected all the results provided by ChatGPT and assessed them for existence, relevance, and author/link correctness. We recorded each resource’s citations, authors, title, journal name, publication year, digital object identifier (DOI), and link. The relevance and correctness of each resource were verified by searching on Google Scholar. Of the total 610 references in the nephrology literature, only 378 (62%) of the references provided by ChatGPT existed, while 31% were fabricated, and 7% of citations were incomplete references. Notably, only 122 (20%) of references were authentic. Additionally, 256 (68%) of the links in the references were found to be incorrect, and the DOI was inaccurate in 206 (54%) of the references. Moreover, among those with a link provided, the link was correct in only 20% of cases, and 3% of the references were irrelevant. Notably, an analysis of specific topics in electrolyte, hemodialysis, and kidney stones found that >60% of the references were inaccurate or misleading, with less reliable authorship and links provided by ChatGPT. Based on our findings, the use of ChatGPT as a sole resource for identifying references to literature reviews in nephrology is not recommended. Future studies could explore ways to improve AI language models’ performance in identifying relevant nephrology literature.","{'model': 'tldr@v2.0.0', 'text': 'The use of ChatGPT as a sole resource for identifying references to literature reviews in nephrology is not recommended and future studies could explore ways to improve AI language models’ performance in identifying relevant nephrologic literature.'}",https://www.mdpi.com/2077-0383/12/17/5550/pdf?version=1692975950
-Multi-Modal Classifiers for Open-Vocabulary Object Detection,Prannay Kaul,"The goal of this paper is open-vocabulary object detection (OVOD) $\unicode{x2013}$ building a model that can detect objects beyond the set of categories seen at training, thus enabling the user to specify categories of interest at inference without the need for model retraining. We adopt a standard two-stage object detector architecture, and explore three ways for specifying novel categories: via language descriptions, via image exemplars, or via a combination of the two. We make three contributions: first, we prompt a large language model (LLM) to generate informative language descriptions for object classes, and construct powerful text-based classifiers; second, we employ a visual aggregator on image exemplars that can ingest any number of images as input, forming vision-based classifiers; and third, we provide a simple method to fuse information from language descriptions and image exemplars, yielding a multi-modal classifier. When evaluating on the challenging LVIS open-vocabulary benchmark we demonstrate that: (i) our text-based classifiers outperform all previous OVOD works; (ii) our vision-based classifiers perform as well as text-based classifiers in prior work; (iii) using multi-modal classifiers perform better than either modality alone; and finally, (iv) our text-based and multi-modal classifiers yield better performance than a fully-supervised detector.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2306.05493
-InstructEval: Systematic Evaluation of Instruction Selection Methods,Anirudh Ajith,"In-context learning (ICL) performs tasks by prompting a large language model (LLM) using an instruction and a small set of annotated examples called demonstrations. Recent work has shown that precise details of the inputs used in the ICL prompt significantly impact performance, which has incentivized instruction selection algorithms. The effect of instruction-choice however is severely underexplored, with existing analyses restricted to shallow subsets of models and tasks, limiting the generalizability of their insights. We develop InstructEval, an ICL evaluation suite to conduct a thorough assessment of these techniques. The suite includes 13 open-sourced LLMs of varying scales from four model families, and covers nine tasks across three categories. Using the suite, we evaluate the relative performance of seven popular instruction selection methods over five metrics relevant to ICL. Our experiments reveal that using curated manually-written instructions or simple instructions without any task-specific descriptions often elicits superior ICL performance overall than that of automatic instruction-induction methods, pointing to a lack of generalizability among the latter. We release our evaluation suite for benchmarking instruction selection approaches and enabling more generalizable methods in this space.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that using curated manually-written instructions or simple instructions without any task-specific descriptions often elicits superior ICL performance overall than that of automatic instruction-induction methods, pointing to a lack of generalizability among the latter.'}",https://arxiv.org/pdf/2307.00259
-ClusterLLM: Large Language Models as a Guide for Text Clustering,Yuwei Zhang,"We introduce ClusterLLM, a novel text clustering framework that leverages feedback from an instruction-tuned large language model, such as ChatGPT. Compared with traditional unsupervised methods that builds upon""small""embedders, ClusterLLM exhibits two intriguing advantages: (1) it enjoys the emergent capability of LLM even if its embeddings are inaccessible; and (2) it understands the user's preference on clustering through textual instruction and/or a few annotated data. First, we prompt ChatGPT for insights on clustering perspective by constructing hard triplet questions, where A, B and C are similar data points that belong to different clusters according to small embedder. We empirically show that this strategy is both effective for fine-tuning small embedder and cost-efficient to query ChatGPT. Second, we prompt ChatGPT for helps on clustering granularity by carefully designed pairwise questions, and tune the granularity from cluster hierarchies that is the most consistent with the ChatGPT answers. Extensive experiments on 14 datasets show that ClusterLLM consistently improves clustering quality, at an average cost of ~$0.6 per dataset.","{'model': 'tldr@v2.0.0', 'text': 'ClusterLLM, a novel text clustering framework that leverages feedback from an instruction-tuned large language model, such as ChatGPT, consistently improves clustering quality, at an average cost of ~$0.6 per dataset.'}",http://arxiv.org/pdf/2305.14871
-SayTap: Language to Quadrupedal Locomotion,Yujin Tang,"Large language models (LLMs) have demonstrated the potential to perform high-level planning. Yet, it remains a challenge for LLMs to comprehend low-level commands, such as joint angle targets or motor torques. This paper proposes an approach to use foot contact patterns as an interface that bridges human commands in natural language and a locomotion controller that outputs these low-level commands. This results in an interactive system for quadrupedal robots that allows the users to craft diverse locomotion behaviors flexibly. We contribute an LLM prompt design, a reward function, and a method to expose the controller to the feasible distribution of contact patterns. The results are a controller capable of achieving diverse locomotion patterns that can be transferred to real robot hardware. Compared with other design choices, the proposed approach enjoys more than 50% success rate in predicting the correct contact patterns and can solve 10 more tasks out of a total of 30 tasks. Our project site is: https://saytap.github.io.","{'model': 'tldr@v2.0.0', 'text': 'An approach to use foot contact patterns as an interface that bridges human commands in natural language and a locomotion controller that outputs these low-level commands results in an interactive system for quadrupedal robots that allows the users to craft diverse locomotion behaviors flexibly.'}",https://arxiv.org/pdf/2306.07580
-How to Unleash the Power of Large Language Models for Few-shot Relation Extraction?,Xin Xu,"Scaling language models have revolutionized widespread NLP tasks, yet little comprehensively explored few-shot relation extraction with large language models. In this paper, we investigate principal methodologies, in-context learning and data generation, for few-shot relation extraction via GPT-3.5 through exhaustive experiments. To enhance few-shot performance, we further propose task-related instructions and schema-constrained data generation. We observe that in-context learning can achieve performance on par with previous prompt learning approaches, and data generation with the large language model can boost previous solutions to obtain new state-of-the-art few-shot results on four widely-studied relation extraction datasets. We hope our work can inspire future research for the capabilities of large language models in few-shot relation extraction. Code is available in https://github.com/zjunlp/DeepKE/tree/main/example/llm.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that in-context learning can achieve performance on par with previous prompt learning approaches, and data generation with the large language model can boost previous solutions to obtain new state-of-the-art few-shot results on four widely-studied relation extraction datasets.'}",http://arxiv.org/pdf/2305.01555
-"Large Language Models and Simple, Stupid Bugs",Kevin Jesse,"With the advent of powerful neural language models, AI-based systems to assist developers in coding tasks are becoming widely available; Copilot is one such system. Copilot uses Codex, a large language model (LLM), to complete code conditioned on a preceding ""prompt"". Codex, however, is trained on public GitHub repositories, viz., on code that may include bugs and vulnerabilities. Previous studies [1], [2] show Codex reproduces vulnerabilities seen in training. In this study, we examine how prone Codex is to generate an interesting bug category, single statement bugs, commonly referred to as simple, stupid bugs or SStuBs in the MSR community. We find that Codex and similar LLMs do help avoid some SStuBs, but do produce known, verbatim SStuBs as much as 2x as likely than known, verbatim correct code. We explore the consequences of the Codex generated SStuBs and propose avoidance strategies that suggest the possibility of reducing the production of known, verbatim SStubs, and increase the possibility of producing known, verbatim fixes.","{'model': 'tldr@v2.0.0', 'text': 'This study examines how prone Codex is to generate an interesting bug category, single statement bugs, commonly referred to as simple, stupid bugs or SStuBs in the MSR community and proposes avoidance strategies that suggest the possibility of reducing the production of known, verbatim SStubs, and increase theossibility of producing known,Verbatim fixes.'}",https://arxiv.org/pdf/2303.11455
-Systematic Rectification of Language Models via Dead-end Analysis,Mengyao Cao,"With adversarial or otherwise normal prompts, existing large language models (LLM) can be pushed to generate toxic discourses. One way to reduce the risk of LLMs generating undesired discourses is to alter the training of the LLM. This can be very restrictive due to demanding computation requirements. Other methods rely on rule-based or prompt-based token elimination, which are limited as they dismiss future tokens and the overall meaning of the complete discourse. Here, we center detoxification on the probability that the finished discourse is ultimately considered toxic. That is, at each point, we advise against token selections proportional to how likely a finished text from this point will be toxic. To this end, we formally extend the dead-end theory from the recent reinforcement learning (RL) literature to also cover uncertain outcomes. Our approach, called rectification, utilizes a separate but significantly smaller model for detoxification, which can be applied to diverse LLMs as long as they share the same vocabulary. Importantly, our method does not require access to the internal representations of the LLM, but only the token probability distribution at each decoding step. This is crucial as many LLMs today are hosted in servers and only accessible through APIs. When applied to various LLMs, including GPT-3, our approach significantly improves the generated discourse compared to the base LLMs and other techniques in terms of both the overall language and detoxification performance.","{'model': 'tldr@v2.0.0', 'text': 'The dead-end theory from the recent reinforcement learning (RL) literature is formally extended to also cover uncertain outcomes and the approach, called rectification, utilizes a separate but significantly smaller model for detoxification, which can be applied to diverse LLMs as long as they share the same vocabulary.'}",http://arxiv.org/pdf/2302.14003
-Knowledge Refinement via Interaction Between Search Engines and Large Language Models,Jiazhan Feng,"Information retrieval (IR) plays a crucial role in locating relevant resources from vast amounts of data, and its applications have evolved from traditional knowledge bases to modern search engines (SEs). The emergence of large language models (LLMs) has further revolutionized the IR field by enabling users to interact with search systems in natural language. In this paper, we explore the advantages and disadvantages of LLMs and SEs, highlighting their respective strengths in understanding user-issued queries and retrieving up-to-date information. To leverage the benefits of both paradigms while circumventing their limitations, we propose InteR, a novel framework that facilitates knowledge refinement through interaction between SEs and LLMs. InteR allows SEs to expand knowledge in queries using LLM-generated knowledge collections and enables LLMs to enhance prompt formulation using SE-retrieved documents. This iterative refinement process augments the inputs of SEs and LLMs, leading to more accurate retrieval. Experiments on large-scale retrieval benchmarks involving web search and low-resource retrieval tasks demonstrate that InteR achieves overall superior zero-shot retrieval performance compared to state-of-the-art methods, even those using relevance judgment. Source code is available at https://github.com/Cyril-JZ/InteR","{'model': 'tldr@v2.0.0', 'text': 'InteR is proposed, a novel framework that facilitates knowledge refinement through interaction between SEs and LLMs, which allows SEs to expand knowledge in queries using LLM-generated knowledge collections and enables LLMs to enhance prompt formulation using SE-retrieved documents.'}",http://arxiv.org/pdf/2305.07402
-Introspective Tips: Large Language Model for In-Context Decision Making,Liting Chen,"The emergence of large language models (LLMs) has substantially influenced natural language processing, demonstrating exceptional results across various tasks. In this study, we employ ``Introspective Tips""to facilitate LLMs in self-optimizing their decision-making. By introspectively examining trajectories, LLM refines its policy by generating succinct and valuable tips. Our method enhances the agent's performance in both few-shot and zero-shot learning situations by considering three essential scenarios: learning from the agent's past experiences, integrating expert demonstrations, and generalizing across diverse games. Importantly, we accomplish these improvements without fine-tuning the LLM parameters; rather, we adjust the prompt to generalize insights from the three aforementioned situations. Our framework not only supports but also emphasizes the advantage of employing LLM in in-contxt decision-making. Experiments involving over 100 games in TextWorld illustrate the superior performance of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This study employs ``Introspective Tips"" to facilitate LLMs in self-optimizing their decision-making by introspectively examining trajectories, which enhances the agent\'s performance in both few-shot and zero-shot learning situations.'}",http://arxiv.org/pdf/2305.11598
-Hypothesis Search: Inductive Reasoning with Language Models,Ruocheng Wang,"Inductive reasoning is a core problem-solving capacity: humans can identify underlying principles from a few examples, which can then be robustly generalized to novel scenarios. Recent work has evaluated large language models (LLMs) on inductive reasoning tasks by directly prompting them yielding""in context learning.""This can work well for straightforward inductive tasks, but performs very poorly on more complex tasks such as the Abstraction and Reasoning Corpus (ARC). In this work, we propose to improve the inductive reasoning ability of LLMs by generating explicit hypotheses at multiple levels of abstraction: we prompt the LLM to propose multiple abstract hypotheses about the problem, in natural language, then implement the natural language hypotheses as concrete Python programs. These programs can be directly verified by running on the observed examples and generalized to novel inputs. Because of the prohibitive cost of generation with state-of-the-art LLMs, we consider a middle step to filter the set of hypotheses that will be implemented into programs: we either ask the LLM to summarize into a smaller set of hypotheses, or ask human annotators to select a subset of the hypotheses. We verify our pipeline's effectiveness on the ARC visual inductive reasoning benchmark, its variant 1D-ARC, and string transformation dataset SyGuS. On a random 40-problem subset of ARC, our automated pipeline using LLM summaries achieves 27.5% accuracy, significantly outperforming the direct prompting baseline (accuracy of 12.5%). With the minimal human input of selecting from LLM-generated candidates, the performance is boosted to 37.5%. (And we argue this is a lower bound on the performance of our approach without filtering.) Our ablation studies show that abstract hypothesis generation and concrete program representations are both beneficial for LLMs to perform inductive reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to improve the inductive reasoning ability of LLMs by generating explicit hypotheses at multiple levels of abstraction: it prompts the LLM to propose multiple abstract hypotheses about the problem, in natural language, then implements the natural language hypotheses as concrete Python programs.'}",https://arxiv.org/pdf/2309.05660
-Augmenting Greybox Fuzzing with Generative AI,Jie Hu,"Real-world programs expecting structured inputs often has a format-parsing stage gating the deeper program space. Neither a mutation-based approach nor a generative approach can provide a solution that is effective and scalable. Large language models (LLM) pre-trained with an enormous amount of natural language corpus have proved to be effective for understanding the implicit format syntax and generating format-conforming inputs. In this paper, propose ChatFuzz, a greybox fuzzer augmented by generative AI. More specifically, we pick a seed in the fuzzer's seed pool and prompt ChatGPT generative models to variations, which are more likely to be format-conforming and thus of high quality. We conduct extensive experiments to explore the best practice for harvesting the power of generative LLM models. The experiment results show that our approach improves the edge coverage by 12.77\% over the SOTA greybox fuzzer (AFL++) on 12 target programs from three well-tested benchmarks. As for vulnerability detection, \sys is able to perform similar to or better than AFL++ for programs with explicit syntax rules but not for programs with non-trivial syntax.","{'model': 'tldr@v2.0.0', 'text': ""ChatFuzz is proposed, a greybox fuzzer augmented by generative AI that picks a seed in the fuzzer's seed pool and prompts ChatGPT generative models to variations, which are more likely to be format-conforming and thus of high quality.""}",http://arxiv.org/pdf/2306.06782
-ThinkSum: Probabilistic reasoning over sets using large language models,Batu Mehmet Ozturkler,"Large language models (LLMs) have a substantial capacity for high-level analogical reasoning: reproducing patterns in linear text that occur in their training data (zero-shot evaluation) or in the provided context (few-shot in-context learning). However, recent studies show that even the more advanced LLMs fail in scenarios that require reasoning over multiple objects or facts and making sequences of logical deductions. We propose a two-stage probabilistic inference paradigm, ThinkSum, which reasons over sets of objects or facts in a structured manner. In the first stage (Think – retrieval of associations), a LLM is queried in parallel over a set of phrases extracted from the prompt or an auxiliary model call. In the second stage (Sum – probabilistic inference or reasoning), the results of these queries are aggregated to make the final prediction. We demonstrate the possibilities and advantages of ThinkSum on the BIG-bench suite of LLM evaluation tasks, achieving improvements over the state of the art using GPT-family models on thirteen difficult tasks, often with far smaller model variants. We also compare and contrast ThinkSum with other proposed modifications to direct prompting of LLMs, such as variants of chain-of-thought prompting. Our results suggest that because the probabilistic inference in ThinkSum is performed outside of calls to the LLM, ThinkSum is less sensitive to prompt design, yields more interpretable predictions, and can be flexibly combined with latent variable models to extract structured knowledge from LLMs. Overall, our proposed paradigm represents a promising approach for enhancing the reasoning capabilities of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Because the probabilistic inference in ThinkSum is performed outside of calls to the LLM, thinkSum is less sensitive to prompt design, yields more interpretable predictions, and can be flexibly combined with latent variable models to extract structured knowledge from LLMs.'}",http://arxiv.org/pdf/2210.01293
-Query Rewriting for Retrieval-Augmented Large Language Models,Xinbei Ma,"Large Language Models (LLMs) play powerful, black-box readers in the retrieve-then-read pipeline, making remarkable progress in knowledge-intensive tasks. This work introduces a new framework, Rewrite-Retrieve-Read instead of the previous retrieve-then-read for the retrieval-augmented LLMs from the perspective of the query rewriting. Unlike prior studies focusing on adapting either the retriever or the reader, our approach pays attention to the adaptation of the search query itself, for there is inevitably a gap between the input text and the needed knowledge in retrieval. We first prompt an LLM to generate the query, then use a web search engine to retrieve contexts. Furthermore, to better align the query to the frozen modules, we propose a trainable scheme for our pipeline. A small language model is adopted as a trainable rewriter to cater to the black-box LLM reader. The rewriter is trained using the feedback of the LLM reader by reinforcement learning. Evaluation is conducted on downstream tasks, open-domain QA and multiple-choice QA. Experiments results show consistent performance improvement, indicating that our framework is proven effective and scalable, and brings a new framework for retrieval-augmented LLM.","{'model': 'tldr@v2.0.0', 'text': 'A new framework, Rewrite-Retrieve-Read instead of the previous retrieve-then-read for the retrieval-augmented LLMs from the perspective of the query rewriting, which is proven effective and scalable, and brings a new framework for retrieval-AUgmentedLLM.'}",http://arxiv.org/pdf/2305.14283
-ChatRule: Mining Logical Rules with Large Language Models for Knowledge Graph Reasoning,Linhao Luo,"Logical rules are essential for uncovering the logical connections between relations, which could improve the reasoning performance and provide interpretable results on knowledge graphs (KGs). Although there have been many efforts to mine meaningful logical rules over KGs, existing methods suffer from the computationally intensive searches over the rule space and a lack of scalability for large-scale KGs. Besides, they often ignore the semantics of relations which is crucial for uncovering logical connections. Recently, large language models (LLMs) have shown impressive performance in the field of natural language processing and various applications, owing to their emergent ability and generalizability. In this paper, we propose a novel framework, ChatRule, unleashing the power of large language models for mining logical rules over knowledge graphs. Specifically, the framework is initiated with an LLM-based rule generator, leveraging both the semantic and structural information of KGs to prompt LLMs to generate logical rules. To refine the generated rules, a rule ranking module estimates the rule quality by incorporating facts from existing KGs. Last, a rule validator harnesses the reasoning ability of LLMs to validate the logical correctness of ranked rules through chain-of-thought reasoning. ChatRule is evaluated on four large-scale KGs, w.r.t. different rule quality metrics and downstream tasks, showing the effectiveness and scalability of our method.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, ChatRule, unleashing the power of large language models for mining logical rules over knowledge graphs, initiated with an LLM-based rule generator, leveraging both the semantic and structural information of KGs to prompt LLMs to generate logical rules.'}",https://arxiv.org/pdf/2309.01538
-Zero-shot Prompting for Code Complexity Prediction Using GitHub Copilot,Mohammed Latif Siddiq,"Code generation models are gaining popularity because they can produce correct code from a prompt, speeding up the software development process. GitHub Copilot is currently one of the most commonly used tools for code generation. This tool is based on GPT3, a Large Language Model (LLM), and can perform zero-shot prompting tasks i.e., tasks for which the model is not specifically trained. In this paper, we describe a preliminary study that investigates whether GitHub Copilot can predict the runtime complexity of a given program using zero- shot prompting. In our study, we found that GitHub Copilot can correctly predict the runtime complexity 45.44% times in the first suggestion and 56.38 % times considering all suggestions. We also compared Copilot to other machine learning, neural network, and transformer-based approaches for code complexity prediction. We observed that Copilot outperformed other approaches for predicting code with linear complexity $\mathbf{O}(n)$.","{'model': 'tldr@v2.0.0', 'text': 'This study found that GitHub Copilot can correctly predict the runtime complexity 45.44% times in the first suggestion and 56.38 % times considering all suggestions, and observed that Copilot outperformed other approaches for predicting code with linear complexity $\\mathbf{O}(n)$.'}",
-Taming AI Bots: Controllability of Neural States in Large Language Models,S. Soatto,"We tackle the question of whether an agent can, by suitable choice of prompts, control an AI bot to any state. To that end, we first introduce a formal definition of ``meaning'' that is amenable to analysis. Then, we characterize ``meaningful data'' on which large language models (LLMs) are ostensibly trained, and ``well-trained LLMs'' through conditions that are largely met by today's LLMs. While a well-trained LLM constructs an embedding space of meanings that is Euclidean, meanings themselves do not form a vector (linear) subspace, but rather a quotient space within. We then characterize the subset of meanings that can be reached by the state of the LLMs for some input prompt, and show that a well-trained bot can reach any meaning albeit with small probability. We then introduce a stronger notion of controllability as {\em almost certain reachability}, and show that, when restricted to the space of meanings, an AI bot is controllable. We do so after introducing a functional characterization of attentive AI bots, and finally derive necessary and sufficient conditions for controllability. The fact that AI bots are controllable means that an adversary could steer them towards any state. However, the sampling process can be designed to counteract adverse actions and avoid reaching undesirable regions of state space before their boundary is crossed.",,http://arxiv.org/pdf/2305.18449
-Spellburst: A Node-based Interface for Exploratory Creative Coding with Natural Language Prompts,Tyler Angert,"Creative coding tasks are often exploratory in nature. When producing digital artwork, artists usually begin with a high-level semantic construct such as a “stained glass filter” and programmatically implement it by varying code parameters such as shape, color, lines, and opacity to produce visually appealing results. Based on interviews with artists, it can be effortful to translate semantic constructs to program syntax, and current programming tools don’t lend well to rapid creative exploration. To address these challenges, we introduce Spellburst, a large language model (LLM) powered creative-coding environment. Spellburst provides (1) a node-based interface that allows artists to create generative art and explore variations through branching and merging operations, (2) expressive prompt-based interactions to engage in semantic programming, and (3) dynamic prompt-driven interfaces and direct code editing to seamlessly switch between semantic and syntactic exploration. Our evaluation with artists demonstrates Spellburst’s potential to enhance creative coding practices and inform the design of computational creativity tools that bridge semantic and syntactic spaces.","{'model': 'tldr@v2.0.0', 'text': 'Spellburst provides a node-based interface that allows artists to create generative art and explore variations through branching and merging operations, expressive prompt-based interactions to engage in semantic programming, and dynamic prompt-driven interfaces and direct code editing to seamlessly switch between semantic and syntactic exploration.'}",https://arxiv.org/pdf/2308.03921
-SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks,Alexander Robey,"Despite efforts to align large language models (LLMs) with human values, widely-used LLMs such as GPT, Llama, Claude, and PaLM are susceptible to jailbreaking attacks, wherein an adversary fools a targeted LLM into generating objectionable content. To address this vulnerability, we propose SmoothLLM, the first algorithm designed to mitigate jailbreaking attacks on LLMs. Based on our finding that adversarially-generated prompts are brittle to character-level changes, our defense first randomly perturbs multiple copies of a given input prompt, and then aggregates the corresponding predictions to detect adversarial inputs. SmoothLLM reduces the attack success rate on numerous popular LLMs to below one percentage point, avoids unnecessary conservatism, and admits provable guarantees on attack mitigation. Moreover, our defense uses exponentially fewer queries than existing attacks and is compatible with any LLM.","{'model': 'tldr@v2.0.0', 'text': 'Based on the finding that adversarially-generated prompts are brittle to character-level changes, the first algorithm designed to mitigate jailbreaking attacks on LLMs is proposed, which reduces the attack success rate on numerous popular LLMs to below one percentage point, avoids unnecessary conservatism, and admits provable guarantees on attack mitigation.'}",https://arxiv.org/pdf/2310.03684
-Fully Autonomous Programming with Large Language Models,Vadim Liventsev,"Current approaches to program synthesis with Large Language Models (LLMs) exhibit a ""near miss syndrome"": they tend to generate programs that semantically resemble the correct answer (as measured by text similarity metrics or human evaluation), but achieve a low or even zero accuracy as measured by unit tests due to small imperfections, such as the wrong input or output format. This calls for an approach known as Synthesize, Execute, Debug (SED), whereby a draft of the solution is generated first, followed by a program repair phase addressing the failed tests. To effectively apply this approach to instruction-driven LLMs, one needs to determine which prompts perform best as instructions for LLMs, as well as strike a balance between repairing unsuccessful programs and replacing them with newly generated ones. We explore these trade-offs empirically, comparing replace-focused, repair-focused, and hybrid debug strategies, as well as different template-based and model-based prompt-generation techniques. We use OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation. The resulting framework outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.","{'model': 'tldr@v2.0.0', 'text': 'This work uses OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation, and finds a framework that outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.'}",https://arxiv.org/pdf/2304.10423
-Large language models and (non-)linguistic recursion,M. Dąbkowski,"Recursion is one of the hallmarks of human language. While many design features of language have been shown to exist in animal communication systems, recursion has not. Previous research shows that GPT-4 is the first large language model (LLM) to exhibit metalinguistic abilities (Begu\v{s}, D\k{a}bkowski, and Rhodes 2023). Here, we propose several prompt designs aimed at eliciting and analyzing recursive behavior in LLMs, both linguistic and non-linguistic. We demonstrate that when explicitly prompted, GPT-4 can both produce and analyze recursive structures. Thus, we present one of the first studies investigating whether meta-linguistic awareness of recursion -- a uniquely human cognitive property -- can emerge in transformers with a high number of parameters such as GPT-4.","{'model': 'tldr@v2.0.0', 'text': 'One of the first studies investigating whether meta-linguistic awareness of recursion -- a uniquely human cognitive property -- can emerge in transformers with a high number of parameters such as GPT-4 is presented.'}",http://arxiv.org/pdf/2306.07195
-Revisit Input Perturbation Problems for LLMs: A Unified Robustness Evaluation Framework for Noisy Slot Filling Task,Guanting Dong,,"{'model': 'tldr@v2.0.0', 'text': 'A unified robustness evaluation framework based on the slot-filling task is proposed to systematically evaluate the dialogue understanding capability of LLMs in diverse input perturbation scenarios and some forward-looking suggestions are made to fuel the research in this direction.'}",
-Domain Knowledge Distillation from Large Language Model: An Empirical Study in the Autonomous Driving Domain,Yun Tang,"Engineering knowledge-based (or expert) systems require extensive manual effort and domain knowledge. As Large Language Models (LLMs) are trained using an enormous amount of cross-domain knowledge, it becomes possible to automate such engineering processes. This paper presents an empirical automation and semi-automation framework for domain knowledge distillation using prompt engineering and the LLM ChatGPT. We assess the framework empirically in the autonomous driving domain and present our key observations. In our implementation, we construct the domain knowledge ontology by""chatting""with ChatGPT. The key finding is that while fully automated domain ontology construction is possible, human supervision and early intervention typically improve efficiency and output quality as they lessen the effects of response randomness and the butterfly effect. We, therefore, also develop a web-based distillation assistant enabling supervision and flexible intervention at runtime. We hope our findings and tools could inspire future research toward revolutionizing the engineering of knowledge-based systems across application domains.","{'model': 'tldr@v2.0.0', 'text': 'The key finding is that while fully automated domain ontology construction is possible, human supervision and early intervention typically improve efficiency and output quality as they lessen the effects of response randomness and the butterfly effect.'}",https://arxiv.org/pdf/2307.11769
-Extensible Prompts for Language Models,Tao Ge,"We propose eXtensible Prompt (X-Prompt) for prompting a large language model (LLM) beyond natural language (NL). X-Prompt instructs an LLM with not only NL but also an extensible vocabulary of imaginary words that are introduced to help represent what NL words hardly describe, allowing a prompt to be more descriptive. Like NL prompts, X-Prompt is out-of-distribution (OOD) robust, for which we propose context-guided learning with prompt augmentation to learn its imaginary words for general usability, enabling them to use in different prompt contexts for fine-grain specifications. The promising results of X-Prompt demonstrate its potential of approaching advanced interaction between humans and LLMs to bridge their communication gap.","{'model': 'tldr@v2.0.0', 'text': 'X-Prompt instructs an LLM with not only NL but also an extensible vocabulary of imaginary words that are introduced to help represent what NL words hardly describe, allowing a prompt to be more descriptive.'}",https://arxiv.org/pdf/2212.00616
-Investigating causal understanding in LLMs,Marius Hobbhahn,"We investigate the quality of causal world models of LLMs in very simple settings. We test whether LLMs can identify cause and effect in natural language settings (taken from BigBench) such as “My car got dirty. I washed the car. Question: Which sentence is the cause of the other?” and in multiple other toy settings. We probe the LLM’s world model by changing the presentation of the prompt while keeping the meaning constant, e.g. by changing the order of the sentences or asking the opposite question. Additionally, we test if the model can be “tricked” into giving wrong answers when we present the shot in a different pattern than the prompt. We have three findings. Firstly, larger models yield better results. Secondly, k-shot outperforms one-shot and one-shot outperforms zero-shot in standard conditions. Thirdly, LLMs perform worse in conditions where form and content differ. We conclude that the form of the presentation matters for LLM predictions or, in other words, that LLMs don’t solely base their predictions on content. Finally, we detail some of the implications this research has on AI safety.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that the form of the presentation matters for LLM predictions or, in other words, that LLMs don’t solely base their predictions on content.'}",
-Reducing Retraining by Recycling Parameter-Efficient Prompts,Brian Lester,"Parameter-efficient methods are able to use a single frozen pre-trained large language model (LLM) to perform many tasks by learning task-specific soft prompts that modulate model behavior when concatenated to the input text. However, these learned prompts are tightly coupled to a given frozen model -- if the model is updated, corresponding new prompts need to be obtained. In this work, we propose and investigate several approaches to""Prompt Recycling'""where a prompt trained on a source model is transformed to work with the new target model. Our methods do not rely on supervised pairs of prompts, task-specific data, or training updates with the target model, which would be just as costly as re-tuning prompts with the target model from scratch. We show that recycling between models is possible (our best settings are able to successfully recycle $88.9\%$ of prompts, producing a prompt that out-performs baselines), but significant performance headroom remains, requiring improved recycling techniques.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes and investigates several approaches to Prompt Recycling, where a prompt trained on a source model is transformed to work with the new target model, and shows that recycling between models is possible.'}",http://arxiv.org/pdf/2208.05577
-Selecting Better Samples from Pre-trained LLMs: A Case Study on Question Generation,Xingdi Yuan,"Large Language Models (LLMs) have in recent years demonstrated impressive prowess in natural language generation. A common practice to improve generation diversity is to sample multiple outputs from the model. However, there lacks a simple and robust way of selecting the best output from these stochastic samples. As a case study framed in the context of question generation, we propose two prompt-based approaches to selecting high-quality questions from a set of LLM-generated candidates. Our method works under the constraints of 1) a black-box (non-modifiable) question generation model and 2) lack of access to human-annotated references -- both of which are realistic limitations for real-world deployment of LLMs. With automatic as well as human evaluations, we empirically demonstrate that our approach can effectively select questions of higher qualities than greedy generation.","{'model': 'tldr@v2.0.0', 'text': 'This case study framed in the context of question generation proposes two prompt-based approaches to selecting high-quality questions from a set of LLM-generated candidates and empirically demonstrates that the approach can effectively select questions of higher qualities than greedy generation.'}",
-Data Curation Alone Can Stabilize In-context Learning,Ting-Yun Chang,"In-context learning (ICL) enables large language models (LLMs) to perform new tasks by prompting them with a sequence of training examples. However, it is known that ICL is very sensitive to the choice of training examples: randomly sampling examples from a training set leads to high variance in performance. In this paper, we show that carefully curating a subset of training data greatly stabilizes ICL performance without any other changes to the ICL algorithm (e.g., prompt retrieval or calibration). We introduce two methods to choose training subsets—both score training examples individually, then select the highest-scoring ones. CondAcc scores a training example by its average dev-set ICL accuracy when combined with random training examples, while Datamodels learns linear regressors that estimate how the presence of each training example influences LLM outputs. Across five tasks and two LLMs, sampling from stable subsets selected by CondAcc and Datamodels improves average accuracy over sampling from the entire training set by 7.7% and 6.3%, respectively.Surprisingly, the stable subset examples are not especially diverse in content or low in perplexity, in contrast with other work suggesting that diversity and perplexity are important when prompting LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper shows that carefully curating a subset of training data greatly stabilizes ICL performance without any other changes to the ICL algorithm, and introduces two methods to choose training subsets—both score training examples individually, then select the highest-scoring ones.'}",https://aclanthology.org/2023.acl-long.452.pdf
-Validating Large Language Models with ReLM,Michael Kuchnik,"Although large language models (LLMs) have been touted for their ability to generate natural-sounding text, there are growing concerns around possible negative effects of LLMs such as data memorization, bias, and inappropriate language. Unfortunately, the complexity and generation capacities of LLMs make validating (and correcting) such concerns difficult. In this work, we introduce ReLM, a system for validating and querying LLMs using standard regular expressions. ReLM formalizes and enables a broad range of language model evaluations, reducing complex evaluation rules to simple regular expression queries. Our results exploring queries surrounding memorization, gender bias, toxicity, and language understanding show that ReLM achieves up to 15x higher system efficiency, 2.5x data efficiency, and increased statistical and prompt-tuning coverage compared to state-of-the-art ad-hoc queries. ReLM offers a competitive and general baseline for the increasingly important problem of LLM validation.","{'model': 'tldr@v2.0.0', 'text': 'Results exploring queries surrounding memorization, gender bias, toxicity, and language understanding show that ReLM achieves up to 15x higher system efficiency, 2.5x data efficiency, and increased statistical and prompt-tuning coverage compared to state-of-the-art ad-hoc queries.'}",http://arxiv.org/pdf/2211.15458
-Preserving In-Context Learning ability in Large Language Model Fine-tuning,Yihan Wang,"Pretrained large language models (LLMs) are strong in-context learners that are able to perform few-shot learning without changing model parameters. However, as we show, fine-tuning an LLM on any specific task generally destroys its in-context ability. We discover an important cause of this loss, format specialization, where the model overfits to the format of the fine-tuned task and is unable to output anything beyond this format. We further show that format specialization happens at the beginning of fine-tuning. To solve this problem, we propose Prompt Tuning with MOdel Tuning (ProMoT), a simple yet effective two-stage fine-tuning framework that preserves in-context abilities of the pretrained model. ProMoT first trains a soft prompt for the fine-tuning target task, and then fine-tunes the model itself with this soft prompt attached. ProMoT offloads task-specific formats into the soft prompt that can be removed when doing other in-context tasks. We fine-tune mT5 XXL with ProMoT on natural language inference (NLI) and English-French translation and evaluate the in-context abilities of the resulting models on 8 different NLP tasks. ProMoT achieves similar performance on the fine-tuned tasks compared with vanilla fine-tuning, but with much less reduction of in-context learning performances across the board. More importantly, ProMoT shows remarkable generalization ability on tasks that have different formats, e.g","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prompt Tuning with MOdel Tuning (ProMoT), a simple yet effective two-stage fine-tuning framework that preserves in-context abilities of the pretrained model and shows remarkable generalization ability on tasks that have different formats.'}",https://arxiv.org/pdf/2211.00635
-Improving Knowledge Extraction from LLMs for Robotic Task Learning through Agent Analysis,James R. Kirk,": Large language models (LLMs) offer significant promise as a knowledge source for robotic task learning. Prompt engineering has been shown to be effective for eliciting knowledge from an LLM but alone is insufficient for acquiring relevant, situationally grounded knowledge for an embodied robotic agent learning novel tasks. We describe a cognitive-agent approach that extends and complements prompt engineering, mitigating its limitations, and thus enabling a robot to acquire new task knowledge matched to its native language capabilities, embodiment, environment, and user preferences. The approach is to increase the response space of LLMs and deploy general strategies, embedded within the autonomous robot, to evaluate, repair, and select among candidate responses produced by the LLM. We describe the approach and experiments that show how a robot, by retrieving and evaluating a breadth of responses from the LLM, can achieve > 75% task completion in one-shot learning without user oversight. The approach achieves 100% task completion when human oversight (such as indication of preference) is provided, while greatly reducing how much human oversight is needed.","{'model': 'tldr@v2.0.0', 'text': 'A cognitive-agent approach is described that extends and complements prompt engineering, mitigating its limitations, and thus enabling a robot to acquire new task knowledge matched to its native language capabilities, embodiment, environment, and user preferences.'}",https://arxiv.org/pdf/2306.06770
-Large Language Models as Superpositions of Cultural Perspectives,Grgur Kovač,"Large Language Models (LLMs) are often misleadingly recognized as having a personality or a set of values. We argue that an LLM can be seen as a superposition of perspectives with different values and personality traits. LLMs exhibit context-dependent values and personality traits that change based on the induced perspective (as opposed to humans, who tend to have more coherent values and personality traits across contexts). We introduce the concept of perspective controllability, which refers to a model's affordance to adopt various perspectives with differing values and personality traits. In our experiments, we use questionnaires from psychology (PVQ, VSM, IPIP) to study how exhibited values and personality traits change based on different perspectives. Through qualitative experiments, we show that LLMs express different values when those are (implicitly or explicitly) implied in the prompt, and that LLMs express different values even when those are not obviously implied (demonstrating their context-dependent nature). We then conduct quantitative experiments to study the controllability of different models (GPT-4, GPT-3.5, OpenAssistant, StableVicuna, StableLM), the effectiveness of various methods for inducing perspectives, and the smoothness of the models' drivability. We conclude by examining the broader implications of our work and outline a variety of associated scientific questions. The project website is available at https://sites.google.com/view/llm-superpositions .",,https://arxiv.org/pdf/2307.07870
-Robot Task Planning Based on Large Language Model Representing Knowledge with Directed Graph Structures,Zhen Yue,"Traditional robot task planning methods face challenges when dealing with highly unstructured environments and complex tasks. We propose a task planning method that combines human expertise with an LLM and have designed an LLM prompt template, Think_Net_Prompt, with stronger expressive power to represent structured professional knowledge. We further propose a method to progressively decompose tasks and generate a task tree to reduce the planning volume for each task, and we have designed a strategy to decouple robot task planning. By dividing different planning entities and separating the task from the actual machine binding process, the task planning process becomes more flexible. Research results show that our method performs well in handling specified code formats, understanding the relationship between tasks and subtasks, and extracting parameters from text descriptions. However, there are also problems such as limited complexity of task logic handling, ambiguity in the quantity of parts and the precise location of assembly. Improving the precision of task description and cognitive structure can bring certain improvements. https://github.com/NOMIzy/Think_Net_Prompt","{'model': 'tldr@v2.0.0', 'text': 'A task planning method that combines human expertise with anLLM and an LLM prompt template with stronger expressive power to represent structured professional knowledge is proposed and a method to progressively decompose tasks and generate a task tree to reduce the planning volume for each task is proposed.'}",http://arxiv.org/pdf/2306.05171
-Using Large Language Models to Generate Engaging Captions for Data Visualizations,A. Liew,"Creating compelling captions for data visualizations has been a long- standing challenge. Visualization researchers are typically untrained in journalistic reporting and hence the captions that are placed be- low data visualizations tend to be not overly engaging and rather just stick to basic observations about the data. In this work we explore the opportunities offered by the newly emerging crop of large language models (LLM) which use sophisticated deep learning technology to produce human-like prose. We ask, can these power-ful software devices be purposed to produce engaging captions for generic data visualizations like a scatterplot. It turns out that the key challenge lies in designing the most effective prompt for the LLM, a task called prompt engineering . We report on first experiments using the popular LLM GPT-3 and deliver some promising results.","{'model': 'tldr@v2.0.0', 'text': 'The opportunities offered by the newly emerging crop of large language models (LLM) which use sophisticated deep learning technology to produce human-like prose are explored and can these power-ful software devices be purposed to produce engaging captions for generic data visualizations like a scatterplot.'}",http://arxiv.org/pdf/2212.14047
-SPeC: A Soft Prompt-Based Calibration on Performance Variability of Large Language Model in Clinical Notes Summarization,Yu-Neng Chuang,"Electronic health records (EHRs) store an extensive array of patient information, encompassing medical histories, diagnoses, treatments, and test outcomes. These records are crucial for enabling healthcare providers to make well-informed decisions regarding patient care. Summarizing clinical notes further assists healthcare professionals in pinpointing potential health risks and making better-informed decisions. This process contributes to reducing errors and enhancing patient outcomes by ensuring providers have access to the most pertinent and current patient data. Recent research has shown that incorporating prompts with large language models (LLMs) substantially boosts the efficacy of summarization tasks. However, we show that this approach also leads to increased output variance, resulting in notably divergent outputs even when prompts share similar meanings. To tackle this challenge, we introduce a model-agnostic Soft Prompt-Based Calibration (SPeC) pipeline that employs soft prompts to diminish variance while preserving the advantages of prompt-based summarization. Experimental findings on multiple clinical note tasks and LLMs indicate that our method not only bolsters performance but also effectively curbs variance for various LLMs, providing a more uniform and dependable solution for summarizing vital medical information.","{'model': 'tldr@v2.0.0', 'text': 'A model-agnostic Soft Prompt-Based Calibration (SPeC) pipeline is introduced that employs soft prompts to diminish variance while preserving the advantages of prompt-based summarization, providing a more uniform and dependable solution for summarizing vital medical information.'}",
-Using a Large Language Model to Control Speaking Style for Expressive TTS,A. Sigurgeirsson,"Appropriate prosody is critical for successful spoken communication. Contextual word embeddings are proven to be helpful in predicting prosody but do not allow for choosing between plausible prosodic renditions. Reference-based TTS models attempt to address this by conditioning speech generation on a reference speech sample. These models can generate expressive speech but this requires finding an appropriate reference. Sufficiently large generative language models have been used to solve various language-related tasks. We explore whether such models can be used to suggest appropriate prosody for expressive TTS. We train a TTS model on a non-expressive corpus and then prompt the language model to suggest changes to pitch, energy and duration. The prompt can be designed for any task and we prompt the model to make suggestions based on target speaking style and dialogue context. The proposed method is rated most appropriate in 49.9% of cases compared to 31.0% for a baseline model.","{'model': 'tldr@v2.0.0', 'text': 'This work trains a TTS model on a non-expressive corpus and then prompts the language model to suggest changes to pitch, energy and duration and prompts the model to make suggestions based on target speaking style and dialogue context.'}",https://arxiv.org/pdf/2305.10321
-GPT4Tools: Teaching Large Language Model to Use Tools via Self-instruction,Rui Yang,"This paper aims to efficiently enable Large Language Models (LLMs) to use multimodal tools. Advanced proprietary LLMs, such as ChatGPT and GPT-4, have shown great potential for tool usage through sophisticated prompt engineering. Nevertheless, these models typically rely on prohibitive computational costs and publicly inaccessible data. To address these challenges, we propose the GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools. It generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts. By using the Low-Rank Adaptation (LoRA) optimization, our approach facilitates the open-source LLMs to solve a range of visual problems, including visual comprehension and image generation. Moreover, we provide a benchmark to evaluate the ability of LLMs to use tools, which is performed in both zero-shot and fine-tuning ways. Extensive experiments demonstrate the effectiveness of our method on various language models, which not only significantly improves the accuracy of invoking seen tools, but also enables the zero-shot capacity for unseen tools. The code and demo are available at https://github.com/StevenGrove/GPT4Tools.","{'model': 'tldr@v2.0.0', 'text': 'The GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools, generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts using the Low-Rank Adaptation (LoRA) optimization.'}",http://arxiv.org/pdf/2305.18752
-Prompting Large Language Model for Machine Translation: A Case Study,Biao Zhang,"Research on prompting has shown excellent performance with little or even no supervised training across many tasks. However, prompting for machine translation is still under-explored in the literature. We fill this gap by offering a systematic study on prompting strategies for translation, examining various factors for prompt template and demonstration example selection. We further explore the use of monolingual data and the feasibility of cross-lingual, cross-domain, and sentence-to-document transfer learning in prompting. Extensive experiments with GLM-130B (Zeng et al., 2022) as the testbed show that 1) the number and the quality of prompt examples matter, where using suboptimal examples degenerates translation; 2) several features of prompt examples, such as semantic similarity, show significant Spearman correlation with their prompting performance; yet, none of the correlations are strong enough; 3) using pseudo parallel prompt examples constructed from monolingual data via zero-shot prompting could improve translation; and 4) improved performance is achievable by transferring knowledge from prompt examples selected in other settings. We finally provide an analysis on the model outputs and discuss several problems that prompting still suffers from.","{'model': 'tldr@v2.0.0', 'text': 'A systematic study on prompting strategies for translation, examining various factors for prompt template and demonstration example selection, and exploring the use of monolingual data and the feasibility of cross-lingual, cross-domain, and sentence-to-document transfer learning in prompting.'}",http://arxiv.org/pdf/2301.07069
-Simulating H.P. Lovecraft horror literature with the ChatGPT large language model,Eduardo C. Garrido-Merch'an,"In this paper, we present a novel approach to simulating H.P. Lovecraft's horror literature using the ChatGPT large language model, specifically the GPT-4 architecture. Our study aims to generate text that emulates Lovecraft's unique writing style and themes, while also examining the effectiveness of prompt engineering techniques in guiding the model's output. To achieve this, we curated a prompt containing several specialized literature references and employed advanced prompt engineering methods. We conducted an empirical evaluation of the generated text by administering a survey to a sample of undergraduate students. Utilizing statistical hypothesis testing, we assessed the students ability to distinguish between genuine Lovecraft works and those generated by our model. Our findings demonstrate that the participants were unable to reliably differentiate between the two, indicating the effectiveness of the GPT-4 model and our prompt engineering techniques in emulating Lovecraft's literary style. In addition to presenting the GPT model's capabilities, this paper provides a comprehensive description of its underlying architecture and offers a comparative analysis with related work that simulates other notable authors and philosophers, such as Dennett. By exploring the potential of large language models in the context of literary emulation, our study contributes to the body of research on the applications and limitations of these models in various creative domains.","{'model': 'tldr@v2.0.0', 'text': ""This study aims to generate text that emulates Lovecraft's unique writing style and themes, while also examining the effectiveness of prompt engineering techniques in guiding the model's output, and provides a comprehensive description of its underlying architecture.""}",http://arxiv.org/pdf/2305.03429
-S3: Social-network Simulation System with Large Language Model-Empowered Agents,Chen Gao,"Social network simulation plays a crucial role in addressing various challenges within social science. It offers extensive applications such as state prediction, phenomena explanation, and policy-making support, among others. In this work, we harness the formidable human-like capabilities exhibited by large language models (LLMs) in sensing, reasoning, and behaving, and utilize these qualities to construct the S$^3$ system (short for $\textbf{S}$ocial network $\textbf{S}$imulation $\textbf{S}$ystem). Adhering to the widely employed agent-based simulation paradigm, we employ prompt engineering and prompt tuning techniques to ensure that the agent's behavior closely emulates that of a genuine human within the social network. Specifically, we simulate three pivotal aspects: emotion, attitude, and interaction behaviors. By endowing the agent in the system with the ability to perceive the informational environment and emulate human actions, we observe the emergence of population-level phenomena, including the propagation of information, attitudes, and emotions. We conduct an evaluation encompassing two levels of simulation, employing real-world social network data. Encouragingly, the results demonstrate promising accuracy. This work represents an initial step in the realm of social network simulation empowered by LLM-based agents. We anticipate that our endeavors will serve as a source of inspiration for the development of simulation systems within, but not limited to, social science.","{'model': 'tldr@v2.0.0', 'text': 'This work harnesses the formidable human-like capabilities exhibited by large language models in sensing, reasoning, and behaving, and utilize these qualities to construct the S$^3 system, an initial step in the realm of social network simulation empowered by LLM-based agents.'}",https://arxiv.org/pdf/2307.14984
-Diminished Diversity-of-Thought in a Standard Large Language Model,Peter S. Park,"We test whether Large Language Models (LLMs) can be used to simulate human participants in social-science studies. To do this, we run replications of 14 studies from the Many Labs 2 replication project with OpenAI's text-davinci-003 model, colloquially known as GPT3.5. Based on our pre-registered analyses, we find that among the eight studies we could analyse, our GPT sample replicated 37.5% of the original results and 37.5% of the Many Labs 2 results. However, we were unable to analyse the remaining six studies due to an unexpected phenomenon we call the""correct answer""effect. Different runs of GPT3.5 answered nuanced questions probing political orientation, economic preference, judgement, and moral philosophy with zero or near-zero variation in responses: with the supposedly""correct answer.""In one exploratory follow-up study, we found that a""correct answer""was robust to changing the demographic details that precede the prompt. In another, we found that most but not all""correct answers""were robust to changing the order of answer choices. One of our most striking findings occurred in our replication of the Moral Foundations Theory survey results, where we found GPT3.5 identifying as a political conservative in 99.6% of the cases, and as a liberal in 99.3% of the cases in the reverse-order condition. However, both self-reported 'GPT conservatives' and 'GPT liberals' showed right-leaning moral foundations. Our results cast doubts on the validity of using LLMs as a general replacement for human participants in the social sciences. Our results also raise concerns that a hypothetical AI-led future may be subject to a diminished diversity-of-thought.","{'model': 'tldr@v2.0.0', 'text': ""Re-replicate 14 psychology studies from the Many Labs 2 replication project with OpenAI's text-davinci-003 model, colloquially known as GPT3.5, to hypothesise that this pattern was learned from a conservative bias in the model's largely Internet-based training data.""}",
-Hierarchical Prompting Assists Large Language Model on Web Navigation,Abishek Sridhar,"Large language models (LLMs) struggle on processing complicated observations in interactive decision making tasks. To alleviate this issue, we propose a simple hierarchical prompting approach. Diverging from previous prompting approaches that always put the full observation (e.g. a web page) to the prompt, we propose to first construct an action-aware observation which is more condensed and relevant with a dedicated SUMMARIZER prompt. The ACTOR prompt then predicts the next action based on the summarized observation. While our method has broad applicability, we particularly demonstrate its efficacy in the complex domain of web navigation where a full observation often contains redundant and irrelevant information. Our approach outperforms the previous state-of-the-art prompting mechanics by 6.2% on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple hierarchical prompting approach that outperforms the previous state-of-the-art prompting mechanics by 6.2% on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.'}",http://arxiv.org/pdf/2305.14257
-Co-training Improves Prompt-based Learning for Large Language Models,Hunter Lang,"We demonstrate that co-training (Blum&Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data. While prompting has emerged as a promising paradigm for few-shot and zero-shot learning, it is often brittle and requires much larger models compared to the standard supervised setup. We find that co-training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model. In the case where we only have partial access to a prompt model (e.g., output probabilities from GPT-3 (Brown et al., 2020)) we learn a calibration model over the prompt outputs. When we have full access to the prompt model's gradients but full finetuning remains prohibitively expensive (e.g., T0 (Sanh et al., 2021)), we learn a set of soft prompt continuous vectors to iteratively update the prompt model. We find that models trained in this manner can significantly improve performance on challenging datasets where there is currently a large gap between prompt-based learning and fully-supervised models.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that co-training (Blum&Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data and co- training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model.'}",
-Prompt text classifications with transformer models! An exemplary introduction to prompt-based learning with large language models,Christian W. F. Mayer,"Abstract This study investigates the potential of automated classification using prompt-based learning approaches with transformer models (large language models trained in an unsupervised manner) for a domain-specific classification task. Prompt-based learning with zero or few shots has the potential to (1) make use of artificial intelligence without sophisticated programming skills and (2) make use of artificial intelligence without fine-tuning models with large amounts of labeled training data. We apply this novel method to perform an experiment using so-called zero-shot classification as a baseline model and a few-shot approach for classification. For comparison, we also fine-tune a language model on the given classification task and conducted a second independent human rating to compare it with the given human ratings from the original study. The used dataset consists of 2,088 email responses to a domain-specific problem-solving task that were manually labeled for their professional communication style. With the novel prompt-based learning approach, we achieved a Cohen’s kappa of .40, while the fine-tuning approach yields a kappa of .59, and the new human rating achieved a kappa of .58 with the original human ratings. However, the classifications from the machine learning models have the advantage that each prediction is provided with a reliability estimate allowing us to identify responses that are difficult to score. We, therefore, argue that response ratings should be based on a reciprocal workflow of machine raters and human raters, where the machine rates easy-to-classify responses and the human raters focus and agree on the responses that are difficult to classify. Further, we believe that this new, more intuitive, prompt-based learning approach will enable more people to use artificial intelligence.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that response ratings should be based on a reciprocal workflow of machine raters and human rater, where the machine rates easy-to-classify responses and the human raters focus and agree on the responses that are difficult to classify.'}",
-Program Synthesis with Large Language Models,Jacob Austin,"This paper explores the limits of the current generation of large language models for program synthesis in general purpose programming languages. We evaluate a collection of such models (with between 244M and 137B parameters) on two new benchmarks, MBPP and MathQA-Python, in both the few-shot and fine-tuning regimes. Our benchmarks are designed to measure the ability of these models to synthesize short Python programs from natural language descriptions. The Mostly Basic Programming Problems (MBPP) dataset contains 974 programming tasks, designed to be solvable by entry-level programmers. The MathQA-Python dataset, a Python version of the MathQA benchmark, contains 23914 problems that evaluate the ability of the models to synthesize code from more complex text. On both datasets, we find that synthesis performance scales log-linearly with model size. Our largest models, even without finetuning on a code dataset, can synthesize solutions to 59.6 percent of the problems from MBPP using few-shot learning with a well-designed prompt. Fine-tuning on a held-out portion of the dataset improves performance by about 10 percentage points across most model sizes. On the MathQA-Python dataset, the largest fine-tuned model achieves 83.8 percent accuracy. Going further, we study the model's ability to engage in dialog about code, incorporating human feedback to improve its solutions. We find that natural language feedback from a human halves the error rate compared to the model's initial prediction. Additionally, we conduct an error analysis to shed light on where these models fall short and what types of programs are most difficult to generate. Finally, we explore the semantic grounding of these models by fine-tuning them to predict the results of program execution. We find that even our best models are generally unable to predict the output of a program given a specific input.","{'model': 'tldr@v2.0.0', 'text': 'The limits of the current generation of large language models for program synthesis in general purpose programming languages are explored, and the semantic grounding of these models is explored by fine-tuning them to predict the results of program execution.'}",
-On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on Codex,Terry Yue Zhuo,"Semantic parsing is a technique aimed at constructing a structured representation of the meaning of a natural-language question. Recent advances in language models trained on code have shown superior performance in generating these representations compared to language models trained solely on natural language text. The existing fine-tuned neural semantic parsers are vulnerable to adversarial attacks on natural-language inputs. While it has been established that the robustness of smaller semantic parsers can be enhanced through adversarial training, this approach is not feasible for large language models in real-world scenarios, as it requires both substantial computational resources and expensive human annotation on in-domain semantic parsing data. This paper presents the first empirical study on the adversarial robustness of a prompt-based semantic parser based on CODEX, a stateof-the-art (SOTA) language model trained on code. Our results demonstrate that the large language model of code is vulnerable to carefully crafted adversarial examples. To overcome this challenge, we propose methods for enhancing robustness without requiring substantial amounts of labelled data or intensive computational resources.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the first empirical study on the adversarial robustness of a prompt-based semantic parser based on CODEX, a stateof-the-art (SOTA) language model trained on code.'}",http://arxiv.org/pdf/2301.12868
-Investigating the Translation Performance of a Large Multilingual Language Model: the Case of BLOOM,Rachel Bawden,"The NLP community recently saw the release of a new large open-access multilingual language model, BLOOM (BigScience et al., 2022) covering 46 languages. We focus on BLOOM’s multilingual ability by evaluating its machine translation performance across several datasets (WMT, Flores-101 and DiaBLa) and language pairs (high- and low-resourced). Our results show that 0-shot performance suffers from overgeneration and generating in the wrong language, but this is greatly improved in the few-shot setting, with very good results for a number of language pairs. We study several aspects including prompt design, model sizes, cross-lingual transfer and the use of discursive context.","{'model': 'tldr@v2.0.0', 'text': 'The results show that 0-shot performance suffers from overgeneration and generating in the wrong language, but this is greatly improved in the few-shot setting, with very good results for a number of language pairs.'}",http://arxiv.org/pdf/2303.01911
-Cold-Start Data Selection for Few-shot Language Model Fine-tuning: A Prompt-Based Uncertainty Propagation Approach,Yue Yu,"Large Language Models have demonstrated remarkable few-shot performance, but the performance can be sensitive to the selection of few-shot instances. We propose PATRON, a new method that uses prompt-based uncertainty estimation for data selection for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON is available at \url{https://github.com/yueyu1030/Patron}.","{'model': 'tldr@v2.0.0', 'text': 'PATRON is a new method that uses prompt-based uncertainty estimation for data selection for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available.'}",http://arxiv.org/pdf/2209.06995
-Can large language models reason about medical questions?,Valentin Li'evin,"Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge. We set out to investigate whether GPT-3.5 (Codex and InstructGPT) can be applied to answer and reason about difficult real-world-based questions. We utilize two multiple-choice medical exam questions (USMLE and MedMCQA) and a medical reading comprehension dataset (PubMedQA). We investigate multiple prompting scenarios: Chain-of-Thought (CoT, think step-by-step), zero- and few-shot (prepending the question with question-answer exemplars) and retrieval augmentation (injecting Wikipedia passages into the prompt). For a subset of the USMLE questions, a medical expert reviewed and annotated the model's CoT. We found that InstructGPT can often read, reason and recall expert knowledge. Failure are primarily due to lack of knowledge and reasoning errors and trivial guessing heuristics are observed, e.g.\ too often predicting labels A and D on USMLE. Sampling and combining many completions overcome some of these limitations. Using 100 samples, Codex 5-shot CoT not only gives close to well-calibrated predictive probability but also achieves human-level performances on the three datasets. USMLE: 60.2%, MedMCQA: 62.7% and PubMedQA: 78.2%.","{'model': 'tldr@v2.0.0', 'text': 'It is found that InstructGPT can often read, reason and recall expert knowledge, and not only gives close to well-calibrated predictive probability but also achieves human-level performances on the three datasets.'}",http://arxiv.org/pdf/2207.08143
-TabLLM: Few-shot Classification of Tabular Data with Large Language Models,S. Hegselmann,"We study the application of large language models to zero-shot and few-shot classification of tabular data. We prompt the large language model with a serialization of the tabular data to a natural-language string, together with a short description of the classification problem. In the few-shot setting, we fine-tune the large language model using some labeled examples. We evaluate several serialization methods including templates, table-to-text models, and large language models. Despite its simplicity, we find that this technique outperforms prior deep-learning-based tabular classification methods on several benchmark datasets. In most cases, even zero-shot classification obtains non-trivial performance, illustrating the method's ability to exploit prior knowledge encoded in large language models. Unlike many deep learning methods for tabular datasets, this approach is also competitive with strong traditional baselines like gradient-boosted trees, especially in the very-few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'Despite its simplicity, this technique outperforms prior deep-learning-based tabular classification methods on several benchmark datasets and is also competitive with strong traditional baselines like gradient-boosted trees, especially in the very-few-shot setting.'}",http://arxiv.org/pdf/2210.10723
-Evaluating the Text-to-SQL Capabilities of Large Language Models,Nitarshan Rajkumar,"We perform an empirical evaluation of Text-to-SQL capabilities of the Codex language model. We find that, without any finetuning, Codex is a strong baseline on the Spider benchmark; we also analyze the failure modes of Codex in this setting. Furthermore, we demonstrate on the GeoQuery and Scholar benchmarks that a small number of in-domain examples provided in the prompt enables Codex to perform better than state-of-the-art models finetuned on such few-shot examples.","{'model': 'tldr@v2.0.0', 'text': 'An empirical evaluation of Text-to-SQL capabilities of the Codex language model is performed and it is demonstrated on the GeoQuery and Scholar benchmarks that a small number of in-domain examples provided in the prompt enables Codex to perform better than state-of-the-art models finetuned on such few-shot examples.'}",http://arxiv.org/pdf/2204.00498
-Prompting Is Programming: A Query Language for Large Language Models,Luca Beurer-Kellner,"Large language models have demonstrated outstanding performance on a wide range of tasks such as question answering and code generation. On a high level, given an input, a language model can be used to automatically complete the sequence in a statistically-likely way. Based on this, users prompt these models with language instructions or examples, to implement a variety of downstream tasks. Advanced prompting methods can even imply interaction between the language model, a user, and external tools such as calculators. However, to obtain state-of-the-art performance or adapt language models for specific tasks, complex task- and model-specific programs have to be implemented, which may still require ad-hoc interaction. Based on this, we present the novel idea of Language Model Programming (LMP). LMP generalizes language model prompting from pure text prompts to an intuitive combination of text prompting and scripting. Additionally, LMP allows constraints to be specified over the language model output. This enables easy adaption to many tasks while abstracting language model internals and providing high-level semantics. To enable LMP, we implement LMQL (short for Language Model Query Language), which leverages the constraints and control flow from an LMP prompt to generate an efficient inference procedure that minimizes the number of expensive calls to the underlying language model. We show that LMQL can capture a wide range of state-of-the-art prompting methods in an intuitive way, especially facilitating interactive flows that are challenging to implement with existing high-level APIs. Our evaluation shows that we retain or increase the accuracy on several downstream tasks, while also significantly reducing the required amount of computation or cost in the case of pay-to-use APIs (26-85% cost savings).","{'model': 'tldr@v2.0.0', 'text': 'LMQL is implemented, which leverages the constraints and control flow from an LMP prompt to generate an efficient inference procedure that minimizes the number of expensive calls to the underlying language model.'}",https://dl.acm.org/doi/pdf/10.1145/3591300
-Large Language Models Are Reasoning Teachers,Namgyu Ho,"Recent works have shown that chain-of-thought (CoT) prompting can elicit language models to solve complex reasoning tasks, step-by-step. However, prompt-based CoT methods are dependent on very large models such as GPT-3 175B which are prohibitive to deploy at scale. In this paper, we use these large models as reasoning teachers to enable complex reasoning in smaller models and reduce model size requirements by several orders of magnitude. We propose Fine-tune-CoT, a method that generates reasoning samples from very large teacher models to fine-tune smaller models. We evaluate our method on a wide range of public models and complex tasks. We find that Fine-tune-CoT enables substantial reasoning capability in small models, far outperforming prompt-based baselines and even the teacher model in many tasks. Additionally, we extend our method by leveraging the teacher model’s ability to generate multiple distinct rationales for each original sample. Enriching the fine-tuning data with such diverse reasoning results in a substantial performance boost across datasets, even for very small models. We conduct ablations and sample studies to understand the emergence of reasoning capabilities of student models. Our code implementation and data are available at https://github.com/itsnamgyu/reasoning-teacher.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses very large models as reasoning teachers to enable complex reasoning in smaller models and reduce model size requirements by several orders of magnitude, and proposes Fine-tune-CoT, a method that generates reasoning samples from very large teacher models to fine-tunes smaller models.'}",http://arxiv.org/pdf/2212.10071
-Black-Box Tuning for Language-Model-as-a-Service,Tianxiang Sun,"Extremely large pre-trained language models (PTMs) such as GPT-3 are usually released as a service. It allows users to design task-specific prompts to query the PTMs through some black-box APIs. In such a scenario, which we call Language-Model-as-a-Service (LMaaS), the gradients of PTMs are usually unavailable. Can we optimize the task prompts by only accessing the model inference APIs? This paper proposes the black-box tuning framework to optimize the continuous prompt prepended to the input text via derivative-free optimization. Instead of optimizing in the original high-dimensional prompt space, which is intractable for traditional derivative-free optimization, we perform optimization in a randomly generated subspace due to the low intrinsic dimensionality of large PTMs. The experimental results show that the black-box tuning with RoBERTa on a few labeled samples not only significantly outperforms manual prompt and GPT-3's in-context learning, but also surpasses the gradient-based counterparts, i.e., prompt tuning and full model tuning.","{'model': 'tldr@v2.0.0', 'text': ""The experimental results show that the black-box tuning with RoBERTa on a few labeled samples not only significantly outperforms manual prompt and GPT-3's in-context learning, but also surpasses the gradient-based counterparts, i.e., prompt tuning and full model tuning.""}",
-BBTv2: Towards a Gradient-Free Future with Large Language Models,Tianxiang Sun,"Most downstream adaptation methods tune all or part of the parameters of pre-trained models (PTMs) through gradient descent, where the tuning cost increases linearly with the growth of the model size.By contrast, gradient-free methods only require the forward computation of the PTM to tune the prompt, retaining the benefits of efficient tuning and deployment.Though, past work on gradient-free tuning often introduces gradient descent to seek a good initialization of prompt and lacks versatility across tasks and PTMs.In this paper, we present BBTv2, an improved version of Black-Box Tuning, to drive PTMs for few-shot learning.We prepend continuous prompts to every layer of the PTM and propose a divide-and-conquer gradient-free algorithm to optimize the prompts at different layers alternately.Extensive experiments across various tasks and PTMs show that BBTv2 can achieve comparable performance to full model tuning and state-of-the-art parameter-efficient methods (e.g., Adapter, LoRA, BitFit, etc.) under few-shot settings while maintaining much fewer tunable parameters.","{'model': 'tldr@v2.0.0', 'text': 'BBTv2 is presented, an improved version of Black-Box Tuning, to drive PTMs for few-shot learning and a divide-and-conquer gradient-free algorithm to optimize the prompts at different layers alternately is proposed.'}",https://aclanthology.org/2022.emnlp-main.259.pdf
-Class-Aware Visual Prompt Tuning for Vision-Language Pre-Trained Model,Yinghui Xing,"With the emergence of large pre-trained vison-language model like CLIP, transferrable representations can be adapted to a wide range of downstream tasks via prompt tuning. Prompt tuning tries to probe the beneficial information for downstream tasks from the general knowledge stored in both the image and text encoders of the pre-trained vision-language model. A recently proposed method named Con- text Optimization (CoOp) introduces a set of learnable vectors as text prompt from the language side, while tuning the text prompt alone can not affect the computed visual features of the image encoder, thus leading to sub-optimal. In this paper, we propose a dual modality prompt tuning paradigm through learning text prompts and visual prompts for both the text and image encoder simultaneously. In addi-tion, to make the visual prompt concentrate more on the tar- get visual concept, we propose Class-Aware Visual Prompt Tuning (CAVPT), which is generated dynamically by per- forming the cross attention between language descriptions of template prompts and visual class token embeddings. Our method provides a new paradigm for tuning the large pretrained vision-language model and extensive experimental re- sults on 8 datasets demonstrate the effectiveness of the proposed method. Our code is available in the supplementary materials.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a dual modality prompt tuning paradigm through learning text prompts and visual prompts for both the text and image encoder simultaneously, which provides a new paradigm for tuning the large pretrained vision-language model.'}",http://arxiv.org/pdf/2208.08340
-Analogy Generation by Prompting Large Language Models: A Case Study of InstructGPT,B. Bhavya,"We propose a novel application of prompting Pre-trained Language Models (PLMs) to generate analogies and study how to design effective prompts for two task settings: generating a source concept analogous to a given target concept (aka Analogous Concept Generation or ACG), and generating an explanation of the similarity between a given pair of target concept and source concept (aka Analogous Explanation Generation or AEG). We found that it is feasible to prompt InstructGPT to generate meaningful analogies and the best prompts tend to be precise imperative statements especially with a low temperature setting. We also systematically analyzed the sensitivity of the InstructGPT model to prompt design, temperature, and injected spelling errors, and found that the model is particularly sensitive to certain variations (e.g., questions vs. imperative statements). Further, we conducted human evaluation on 1.4k of the generated analogies and found that the quality of generations varies substantially by model size. The largest InstructGPT model can achieve human-level performance at generating meaningful analogies for a given target while there is still room for improvement on the AEG task.","{'model': 'tldr@v2.0.0', 'text': 'It is feasible to prompt InstructGPT to generate meaningful analogies and the best prompts tend to be precise imperative statements especially with a low temperature setting, and the quality of generations varies substantially by model size.'}",http://arxiv.org/pdf/2210.04186
-Using Large Language Models to Simulate Multiple Humans,Gati Aher,"We propose a method for using a large language model, such as GPT-3, to simulate responses of different humans in a given context. We test our method by attempting to repro- duce well-established economic, psycholinguistic, and social experiments. The method requires prompt templates for each experiment. Simulations are run by varying the (hypotheti-cal) subject details, such as name, and analyzing the text gen- erated by the language model. To validate our methodology, we use GPT-3 to simulate the Ultimatum Game , garden path sentences , risk aversion , and the Milgram Shock experiments. In order to address concerns of exposure to these studies in training data, we also evaluate simulations on novel variants of these studies. We show that it is possible to simulate re- sponses of different people and that their responses are consistent with prior human studies from the literature. Across all studies, the distributions generated by larger language models better align with prior experimental results, suggesting a trend that future language models may be used for even more faithful simulations of human responses. Our use of a lan- guage model for simulation is contrasted with anthropomor-phic views of a language model as having its own behavior.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that it is possible to simulate re- sponses of different people and that their responses are consistent with prior human studies from the literature, suggesting a trend that future language models may be used for even more faithful simulations of human responses.'}",https://arxiv.org/pdf/2208.10264
-Persistent Anti-Muslim Bias in Large Language Models,Abubakar Abid,"It has been observed that large-scale language models capture undesirable societal biases, e.g. relating to race and gender; yet religious bias has been relatively unexplored. We demonstrate that GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias. We probe GPT-3 in various ways, including prompt completion, analogical reasoning, and story generation, to understand this anti-Muslim bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups. For instance, Muslim is analogized to terrorist in 23% of test cases, while Jewish is mapped to its most common stereotype, money, in 5% of test cases. We quantify the positive distraction needed to overcome this bias with adversarial text prompts, and find that use of the most positive 6 adjectives reduces violent completions for Muslims from 66% to 20%, but which is still higher than for other religious groups.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups.'}",https://arxiv.org/pdf/2101.05783
-Large Language Models in the Workplace: A Case Study on Prompt Engineering for Job Type Classification,Benjamin Clavié,"This case study investigates the task of job classification in a real-world setting, where the goal is to determine whether an English-language job posting is appropriate for a graduate or entry-level position. We explore multiple approaches to text classification, including supervised approaches such as traditional models like Support Vector Machines (SVMs) and state-of-the-art deep learning methods such as DeBERTa. We compare them with Large Language Models (LLMs) used in both few-shot and zero-shot classification settings. To accomplish this task, we employ prompt engineering, a technique that involves designing prompts to guide the LLMs towards the desired output. Specifically, we evaluate the performance of two commercially available state-of-the-art GPT-3.5-based language models, text-davinci-003 and gpt-3.5-turbo. We also conduct a detailed analysis of the impact of different aspects of prompt engineering on the model's performance. Our results show that, with a well-designed prompt, a zero-shot gpt-3.5-turbo classifier outperforms all other models, achieving a 6% increase in Precision@95% Recall compared to the best supervised approach. Furthermore, we observe that the wording of the prompt is a critical factor in eliciting the appropriate""reasoning""in the model, and that seemingly minor aspects of the prompt significantly affect the model's performance.","{'model': 'tldr@v2.0.0', 'text': 'This case study investigates the task of job classification in a real-world setting, where the goal is to determine whether an English-language job posting is appropriate for a graduate or entry-level position, and evaluates the performance of two commercially available state-of-the-art GPT-3.5-turbo language models.'}",http://arxiv.org/pdf/2303.07142
-Boosted Prompt Ensembles for Large Language Models,Silviu Pitis,"Methods such as chain-of-thought prompting and self-consistency have pushed the frontier of language model reasoning performance with no additional training. To further improve performance, we propose a prompt ensembling method for large language models, which uses a small dataset to construct a set of few shot prompts that together comprise a ``boosted prompt ensemble''. The few shot examples for each prompt are chosen in a stepwise fashion to be ``hard'' examples on which the previous step's ensemble is uncertain. We show that this outperforms single-prompt output-space ensembles and bagged prompt-space ensembles on the GSM8k and AQuA datasets, among others. We propose both train-time and test-time versions of boosted prompting that use different levels of available annotation and conduct a detailed empirical study of our algorithm.","{'model': 'tldr@v2.0.0', 'text': ""A prompt ensembling method for large language models, which uses a small dataset to construct a set of few shot prompts that together comprise a ``boosted prompt ensemble'' that outperforms single-prompt output-space ensembles and bagged prompt-spaceEnsemble on the GSM8k and AQuA datasets, among others.""}",http://arxiv.org/pdf/2304.05970
-Soft-prompt Tuning for Large Language Models to Evaluate Bias,Jacob-Junqi Tian,"Prompting large language models has gained immense popularity in recent years due to the advantage of producing good results even without the need for labelled data. However, this requires prompt tuning to get optimal prompts that lead to better model performances. In this paper, we explore the use of soft-prompt tuning on sentiment classification task to quantify the biases of large language models (LLMs) such as Open Pre-trained Transformers (OPT) and Galactica language model. Since these models are trained on real-world data that could be prone to bias toward certain groups of populations, it is important to identify these underlying issues. Using soft-prompts to evaluate bias gives us the extra advantage of avoiding the human-bias injection that can be caused by manually designed prompts. We check the model biases on different sensitive attributes using the group fairness (bias) and find interesting bias patterns. Since LLMs have been used in the industry in various applications, it is crucial to identify the biases before deploying these models in practice. We open-source our pipeline and encourage industry researchers to adapt our work to their use cases.","{'model': 'tldr@v2.0.0', 'text': 'The use of soft-prompt tuning on sentiment classification task to quantify the biases of large language models such as Open Pre-trained Transformers (OPT) and Galactica language model and finds interesting bias patterns.'}",http://arxiv.org/pdf/2306.04735
-Promptify: Text-to-Image Generation through Interactive Prompt Exploration with Large Language Models,Stephen Brade,"Text-to-image generative models have demonstrated remarkable capabilities in generating high-quality images based on textual prompts. However, crafting prompts that accurately capture the user’s creative intent remains challenging. It often involves laborious trial-and-error procedures to ensure that the model interprets the prompts in alignment with the user’s intention. To address these challenges, we present Promptify, an interactive system that supports prompt exploration and refinement for text-to-image generative models. Promptify utilizes a suggestion engine powered by large language models to help users quickly explore and craft diverse prompts. Our interface allows users to organize the generated images flexibly, and based on their preferences, Promptify suggests potential changes to the original prompt. This feedback loop enables users to iteratively refine their prompts and enhance desired features while avoiding unwanted ones. Our user study shows that Promptify effectively facilitates the text-to-image workflow, allowing users to create visually appealing images on their first attempt while requiring significantly less cognitive load than a widely-used baseline tool.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Promptify, an interactive system that supports prompt exploration and refinement for text-to-image generative models, and utilizes a suggestion engine powered by large language models to help users quickly explore and craft diverse prompts.'}",https://arxiv.org/pdf/2304.09337
-You Only Prompt Once: On the Capabilities of Prompt Learning on Large Language Models to Tackle Toxic Content,Xinlei He,"The spread of toxic content online is an important problem that has adverse effects on user experience online and in our society at large. Motivated by the importance and impact of the problem, research focuses on developing solutions to detect toxic content, usually leveraging machine learning (ML) models trained on human-annotated datasets. While these efforts are important, these models usually do not generalize well and they can not cope with new trends (e.g., the emergence of new toxic terms). Currently, we are witnessing a shift in the approach to tackling societal issues online, particularly leveraging large language models (LLMs) like GPT-3 or T5 that are trained on vast corpora and have strong generalizability. In this work, we investigate how we can use LLMs and prompt learning to tackle the problem of toxic content, particularly focusing on three tasks; 1) Toxicity Classification, 2) Toxic Span Detection, and 3) Detoxification. We perform an extensive evaluation over five model architectures and eight datasets demonstrating that LLMs with prompt learning can achieve similar or even better performance compared to models trained on these specific tasks. We find that prompt learning achieves around 10\% improvement in the toxicity classification task compared to the baselines, while for the toxic span detection task we find better performance to the best baseline (0.643 vs. 0.640 in terms of $F_1$-score). Finally, for the detoxification task, we find that prompt learning can successfully reduce the average toxicity score (from 0.775 to 0.213) while preserving semantic meaning.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates how to use LLMs and prompt learning to tackle the problem of toxic content, particularly focusing on three tasks; 1) Toxicity Classification, 2) Toxic Span Detection, and 3) Detoxification.'}",https://arxiv.org/pdf/2308.05596
-Controlling the Extraction of Memorized Data from Large Language Models via Prompt-Tuning,Mustafa Safa Ozdayi,"Large Language Models (LLMs) are known to memorize significant portions of their training data. Parts of this memorized content have been shown to be extractable by simply querying the model, which poses a privacy risk. We present a novel approach which uses prompt-tuning to control the extraction rates of memorized content in LLMs. We present two prompt training strategies to increase and decrease extraction rates, which correspond to an attack and a defense, respectively. We demonstrate the effectiveness of our techniques by using models from the GPT-Neo family on a public benchmark. For the 1.3B parameter GPT-Neo model, our attack yields a 9.3 percentage point increase in extraction rate compared to our baseline. Our defense can be tuned to achieve different privacy-utility trade-offs by a user-specified hyperparameter. We achieve an extraction rate reduction of up to 97.7% relative to our baseline, with a perplexity increase of 16.9%.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach which uses prompt-tuning to control the extraction rates of memorized content in LLMs, and presents two prompt training strategies to increase and decrease extraction rates, which correspond to an attack and a defense, respectively.'}",http://arxiv.org/pdf/2305.11759
-Sensitivity and Robustness of Large Language Models to Prompt Template in Japanese Text Classification Tasks,Chengguang Gan,"Prompt engineering relevance research has seen a notable surge in recent years, primarily driven by advancements in pre-trained language models and large language models. However, a critical issue has been identified within this domain: the inadequate of sensitivity and robustness of these models towards Prompt Templates, particularly in lesser-studied languages such as Japanese. This paper explores this issue through a comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM). These models are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context. Our experimental results reveal startling discrepancies. A simple modification in the sentence structure of the Prompt Template led to a drastic drop in the accuracy of GPT-4 from 49.21 to 25.44. This observation underscores the fact that even the highly performance GPT-4 model encounters significant stability issues when dealing with diverse Japanese prompt templates, rendering the consistency of the model's output results questionable. In light of these findings, we conclude by proposing potential research trajectories to further enhance the development and performance of Large Language Models in their current stage.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM) is scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context.'}",
-Legal Prompting: Teaching a Language Model to Think Like a Lawyer,Fang Yu,"Large language models that are capable of zero or few-shot prompting approaches have given rise to the new research area of prompt engineering. Recent advances showed that for example Chain-of-Thought (CoT) prompts can improve arithmetic or common sense tasks significantly. We explore how such approaches fare with legal reasoning tasks and take the COLIEE entailment task based on the Japanese Bar exam for testing zero-shot/few-shot and fine-tuning approaches. Our findings show that while CoT prompting and fine-tuning with explanations approaches show improvements, the best results are produced by prompts that are derived from specific legal reasoning techniques such as IRAC (Issue, Rule, Application, Conclusion). Based on our experiments we improve the 2021 best result from 0.7037 accuracy to 0.8148 accuracy and beat the 2022 best system of 0.6789 accuracy with an accuracy of 0.7431.","{'model': 'tldr@v2.0.0', 'text': 'The findings show that while CoT prompting and fine-tuning with explanations approaches show improvements, the best results are produced by prompts that are derived from specific legal reasoning techniques such as IRAC (Issue, Rule, Application, Conclusion).'}",http://arxiv.org/pdf/2212.01326
-Sensitivity and Robustness of Large Language Models to Prompt in Japanese,Chengguang Gan,"Prompt Engineering has gained significant rel-evance in recent years, fueled by advance-ments in pre-trained and large language models. However, a critical issue has been iden-tified within this domain: the lack of sensitivity and robustness of these models towards Prompt Templates, particularly in lesser-studied languages such as Japanese. This paper explores this issue through a comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM), T5. These models are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context. Our experimental results reveal startling discrepancies. A simple modification in the sentence structure of the Prompt Template led to a drastic drop in the accuracy of GPT-4 from 49.21 to 25.44. This observation underscores the fact that even the highly performance GPT-4 model encoun-ters significant stability issues when dealing with diverse Japanese prompt templates, ren-dering the consistency of the model’s output results questionable. In light of these findings, we conclude by proposing potential research trajectories to further enhance the devel-opment and performance of Large Language Models in their current stage.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model (PLM), T5 are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context.'}",http://arxiv.org/pdf/2305.08714
-Bounding the Capabilities of Large Language Models in Open Text Generation with Prompt Constraints,Albert Lu,"The limits of open-ended generative models are unclear, yet increasingly important. What causes them to succeed and what causes them to fail? In this paper, we take a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models. We present a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic. These constraint types are categorized into a set of well-defined constraints that are analyzable by a single prompt. We then systematically create a diverse set of simple, natural, and useful prompts to robustly analyze each individual constraint. Using the GPT-3 text-davinci-002 model as a case study, we generate outputs from our collection of prompts and analyze the model’s generative failures. We also show the generalizability of our proposed method on other large models like BLOOM and OPT. Our results and our in-context mitigation strategies reveal open challenges for future research.","{'model': 'tldr@v2.0.0', 'text': 'This paper takes a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models, and presents a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic.'}",http://arxiv.org/pdf/2302.09185
-LINGUIST: Language Model Instruction Tuning to Generate Annotated Utterances for Intent Classification and Slot Tagging,Andrew Rosenbaum,"We present LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt. In a 10-shot novel intent setting for the SNIPS dataset, LINGUIST surpasses state-of-the-art approaches (Back-Translation and Example Extrapolation) by a wide margin, showing absolute improvement for the target intents of +1.9 points on IC Recall and +2.5 points on ST F1 Score. In the zero-shot cross-lingual setting of the mATIS++ dataset, LINGUIST out-performs a strong baseline of Machine Translation with Slot Alignment by +4.14 points absolute on ST F1 Score across 6 languages, while matching performance on IC. Finally, we verify our results on an internal large-scale multilingual dataset for conversational agent IC+ST and show significant improvements over a baseline which uses Back-Translation, Paraphrasing and Slot Catalog Resampling. To our knowledge, we are the first to demonstrate instruction fine-tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation.","{'model': 'tldr@v2.0.0', 'text': 'This work presents LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt, and is the first to demonstrate instruction fine- tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation'}",http://arxiv.org/pdf/2209.09900
-AugESC: Dialogue Augmentation with Large Language Models for Emotional Support Conversation,Chujie Zheng,"Crowdsourced dialogue corpora are usually limited in scale and topic coverage due to the expensive cost of data curation. This would hinder the generalization of downstream dialogue models to open-domain topics. In this work, we leverage large language models for dialogue augmentation in the task of emotional support conversation (ESC). By treating dialogue augmentation as a dialogue completion task, we prompt a fine-tuned language model to complete full dialogues from available dialogue posts of various topics, which are then postprocessed based on heuristics. Applying this approach, we construct AugESC, an augmented dataset for the ESC task, which largely extends the scale and topic coverage of the crowdsourced ESConv corpus. Through comprehensive human evaluation, we demonstrate that our approach is superior to strong baselines of dialogue augmentation and that AugESC has comparable dialogue quality to the crowdsourced corpus. We also conduct human interactive evaluation and prove that post-training on AugESC improves downstream dialogue models' generalization ability to open-domain topics. These results suggest the utility of AugESC and highlight the potential of large language models in improving data-scarce dialogue generation tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work leverages large language models for dialogue augmentation in the task of emotional support conversation (ESC) to prompt a fine-tuned language model to complete full dialogues from available dialogue posts of various topics, which are then postprocessed based on heuristics.'}",https://aclanthology.org/2023.findings-acl.99.pdf
-CoNAL: Anticipating Outliers with Large Language Models,Albert Xu,"In many task settings, text classification models are likely to encounter examples from novel classes on which they cannot predict correctly. Selective prediction, in which models abstain on low-confidence examples, provides a possible solution, but existing models are often overly confident on OOD examples. To remedy this overconfidence, we introduce Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them. First, we generate OOD examples by prompting a large language model twice: we prompt it to enumerate relevant novel labels, then generate examples from each novel class matching the task format. Second, we train our classifier with a novel contrastive objective that encourages lower confidence on generated OOD examples than training examples. When trained with CoNAL, classifiers improve in their ability to detect and abstain on OOD examples over prior methods by an average of 2.3% AUAC and 5.5% AUROC across 4 NLP datasets, with no cost to in-distribution accuracy.1","{'model': 'tldr@v2.0.0', 'text': ""Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them, which improves classifiers' ability to detect and abstain on OODExamples over prior methods.""}",http://arxiv.org/pdf/2211.15718
-Variational prompt tuning improves generalization of vision-language models,Mohammad Mahdi Derakhshani,"Prompt tuning provides an efficient mechanism to adapt large vision-language models to downstream tasks by treating part of the input language prompts as learnable parameters while freezing the rest of the model. Existing works for prompt tuning are however prone to damaging the generalization capabilities of the foundation models, because the learned prompts lack the capacity of covering certain concepts within the language model. To avoid such limitation, we propose a probabilistic modeling of the underlying distribution of prompts, allowing prompts within the support of an associated concept to be derived through stochastic sampling. This results in a more complete and richer transfer of the information captured by the language model, providing better generalization capabilities for downstream tasks. The resulting algorithm relies on a simple yet powerful variational framework that can be directly integrated with other developments. We show our approach is seamlessly integrated into both standard and conditional prompt learning frameworks, improving the performance on both cases considerably, especially with regards to preserving the generalization capability of the original model. Our method provides the current state-of-the-art for prompt learning, surpassing CoCoOp by 1.6% average Top-1 accuracy on the standard benchmark. Remarkably, it even surpasses the original CLIP model in terms of generalization to new classes. Implementation code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a probabilistic modeling of the underlying distribution of prompts, allowing prompts within the support of an associated concept to be derived through stochastic sampling, which results in a more complete and richer transfer of the information captured by the language model, providing better generalization capabilities for downstream tasks.'}",https://arxiv.org/pdf/2210.02390
-Prompt-and-Rerank: A Method for Zero-Shot and Few-Shot Arbitrary Textual Style Transfer with Small Language Models,Mirac Suzgun,"We propose a method for arbitrary textual style transfer (TST)—the task of transforming a text into any given style—utilizing general-purpose pre-trained language models. Our method, Prompt-and-Rerank, is based on a mathematical formulation of the TST task, decomposing it into three constituent components: textual similarity, target style strength, and fluency. Our method uses zero-shot or few-shot prompting to obtain a set of candidate generations in the target style, and then re-ranks them according to the three components. Our method enables small pre-trained language models to perform on par with state-of-the-art large-scale models while using two orders of magnitude less compute and memory. We also investigate the effect of model size and prompt design (e.g., prompt paraphrasing and delimiter-pair choice) on style transfer quality across seven diverse textual style transfer datasets, finding, among other things, that delimiter-pair choice has a large impact on performance, and that models have biases on the direction of style transfer.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for arbitrary textual style transfer (TST), based on a mathematical formulation of the TST task, that enables small pre-trained language models to perform on par with state-of-the-art large-scale models while using two orders of magnitude less compute and memory.'}",https://arxiv.org/pdf/2205.11503
-Visual-Language Navigation Pretraining via Prompt-based Environmental Self-exploration,Xiwen Liang,"Vision-language navigation (VLN) is a challenging task due to its large searching space in the environment. To address this problem, previous works have proposed some methods of fine-tuning a large model that pretrained on large-scale datasets. However, the conventional fine-tuning methods require extra human-labeled navigation data and lack self-exploration capabilities in environments, which hinders their generalization of unseen scenes. To improve the ability of fast cross-domain adaptation, we propose Prompt-based Environmental Self-exploration (ProbES), which can self-explore the environments by sampling trajectories and automatically generates structured instructions via a large-scale cross-modal pretrained model (CLIP). Our method fully utilizes the knowledge learned from CLIP to build an in-domain dataset by self-exploration without human labeling. Unlike the conventional approach of fine-tuning, we introduce prompt tuning to achieve fast adaptation for language embeddings, which substantially improves the learning efficiency by leveraging prior knowledge. By automatically synthesizing trajectory-instruction pairs in any environment without human supervision and instruction prompt tuning, our model can adapt to diverse vision-language navigation tasks, including VLN and REVERIE. Both qualitative and quantitative results show that our ProbES significantly improves the generalization ability of the navigation model.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces prompt tuning to achieve fast adaptation for language embeddings, which substantially improves the learning efficiency by leveraging prior knowledge and significantly improves the generalization ability of the navigation model.'}",http://arxiv.org/pdf/2203.04006
-PRCBERT: Prompt Learning for Requirement Classification using BERT-based Pretrained Language Models,Xianchang Luo,"Software requirement classification is a longstanding and important problem in requirement engineering. Previous studies have applied various machine learning techniques for this problem, including Support Vector Machine (SVM) and decision trees. With the recent popularity of NLP technique, the state-of-the-art approach NoRBERT utilizes the pre-trained language model BERT and achieves a satisfactory performance. However, the dataset PROMISE used by the existing approaches for this problem consists of only hundreds of requirements that are outdated according to today’s technology and market trends. Besides, the NLP technique applied in these approaches might be obsolete. In this paper, we propose an approach of prompt learning for requirement classification using BERT-based pretrained language models (PRCBERT), which applies flexible prompt templates to achieve accurate requirements classification. Experiments conducted on two existing small-size requirement datasets (PROMISE and NFR-Review) and our collected large-scale requirement dataset NFR-SO prove that PRCBERT exhibits moderately better classification performance than NoRBERT and MLM-BERT (BERT with the standard prompt template). On the de-labeled NFR-Review and NFR-SO datasets, Trans_PRCBERT (the version of PRCBERT which is fine-tuned on PROMISE) is able to have a satisfactory zero-shot performance with 53.27% and 72.96% F1-score when enabling a self-learning strategy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an approach of prompt learning for requirement classification using BERT-based pretrained language models (PRCBERT), which applies flexible prompt templates to achieve accurate requirements classification.'}",https://dl.acm.org/doi/pdf/10.1145/3551349.3560417
-Large Language Models Can Be Easily Distracted by Irrelevant Context,Freda Shi,"Large language models have achieved impressive performance on various natural language processing tasks. However, so far they have been evaluated primarily on benchmarks where all information in the input context is relevant for solving the task. In this work, we investigate the distractibility of large language models, i.e., how the model problem-solving accuracy can be influenced by irrelevant context. In particular, we introduce Grade-School Math with Irrelevant Context (GSM-IC), an arithmetic reasoning dataset with irrelevant information in the problem description. We use this benchmark to measure the distractibility of cutting-edge prompting techniques for large language models, and find that the model performance is dramatically decreased when irrelevant information is included. We also identify several approaches for mitigating this deficiency, such as decoding with self-consistency and adding to the prompt an instruction that tells the language model to ignore the irrelevant information.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the distractibility of large language models, i.e., how the model problem-solving accuracy can be influenced by irrelevant context, and introduces Grade-School Math with Irrelevant Context (GSM-IC), an arithmetic reasoning dataset with irrelevant information in the problem description.'}",http://arxiv.org/pdf/2302.00093
-ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs,Yujia Qin,"Despite the advancements of open-source large language models (LLMs), e.g., LLaMA, they remain significantly limited in tool-use capabilities, i.e., using external tools (APIs) to fulfill human instructions. The reason is that current instruction tuning largely focuses on basic language tasks but ignores the tool-use domain. This is in contrast to the excellent tool-use capabilities of state-of-the-art (SOTA) closed-source LLMs, e.g., ChatGPT. To bridge this gap, we introduce ToolLLM, a general tool-use framework encompassing data construction, model training, and evaluation. We first present ToolBench, an instruction-tuning dataset for tool use, which is constructed automatically using ChatGPT. Specifically, the construction can be divided into three stages: (i) API collection: we collect 16,464 real-world RESTful APIs spanning 49 categories from RapidAPI Hub; (ii) instruction generation: we prompt ChatGPT to generate diverse instructions involving these APIs, covering both single-tool and multi-tool scenarios; (iii) solution path annotation: we use ChatGPT to search for a valid solution path (chain of API calls) for each instruction. To enhance the reasoning capabilities of LLMs, we develop a novel depth-first search-based decision tree algorithm. It enables LLMs to evaluate multiple reasoning traces and expand the search space. Moreover, to evaluate the tool-use capabilities of LLMs, we develop an automatic evaluator: ToolEval. Based on ToolBench, we fine-tune LLaMA to obtain an LLM ToolLLaMA, and equip it with a neural API retriever to recommend appropriate APIs for each instruction. Experiments show that ToolLLaMA demonstrates a remarkable ability to execute complex instructions and generalize to unseen APIs, and exhibits comparable performance to ChatGPT. Our ToolLLaMA also demonstrates strong zero-shot generalization ability in an out-of-distribution tool-use dataset: APIBench.","{'model': 'tldr@v2.0.0', 'text': 'ToolLLM is introduced, a general tool-use framework encompassing data construction, model training, and evaluation, and a novel depth-first search-based decision tree algorithm that enables LLMs to evaluate multiple reasoning traces and expand the search space.'}",https://arxiv.org/pdf/2307.16789
-Fundamental Limitations of Alignment in Large Language Models,Yotam Wolf,"An important aspect in developing language models that interact with humans is aligning their behavior to be useful and unharmful for their human users. This is usually achieved by tuning the model in a way that enhances desired behaviors and inhibits undesired ones, a process referred to as alignment. In this paper, we propose a theoretical approach called Behavior Expectation Bounds (BEB) which allows us to formally investigate several inherent characteristics and limitations of alignment in large language models. Importantly, we prove that within the limits of this framework, for any behavior that has a finite probability of being exhibited by the model, there exist prompts that can trigger the model into outputting this behavior, with probability that increases with the length of the prompt. This implies that any alignment process that attenuates an undesired behavior but does not remove it altogether, is not safe against adversarial prompting attacks. Furthermore, our framework hints at the mechanism by which leading alignment approaches such as reinforcement learning from human feedback make the LLM prone to being prompted into the undesired behaviors. This theoretical result is being experimentally demonstrated in large scale by the so called contemporary""chatGPT jailbreaks"", where adversarial users trick the LLM into breaking its alignment guardrails by triggering it into acting as a malicious persona. Our results expose fundamental limitations in alignment of LLMs and bring to the forefront the need to devise reliable mechanisms for ensuring AI safety.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that within the limits of this framework, for any behavior that has a finite probability of being exhibited by the model, there exist prompts that can trigger the model into outputting this behavior, with probability that increases with the length of the prompt.'}",https://arxiv.org/pdf/2304.11082
-Synthetic Prompting: Generating Chain-of-Thought Demonstrations for Large Language Models,Zhihong Shao,"Large language models can perform various reasoning tasks by using chain-of-thought prompting, which guides them to find answers through step-by-step demonstrations. However, the quality of the prompts depends on the demonstrations given to the models, and creating many of them by hand is costly. We introduce Synthetic prompting, a method that leverages a few handcrafted examples to prompt the model to generate more examples by itself, and selects effective demonstrations to elicit better reasoning. Our method alternates between a backward and forward process to generate new examples. The backward process generates a question that match a sampled reasoning chain, so that the question is solvable and clear. The forward process produces a more detailed reasoning chain for the question, improving the quality of the example. We evaluate our method on numerical, symbolic, and algorithmic reasoning tasks, and show that it outperforms existing prompting techniques.","{'model': 'tldr@v2.0.0', 'text': 'Synthetic prompting is introduced, a method that leverages a few handcrafted examples to prompt the model to generate more examples by itself, and selects effective demonstrations to elicit better reasoning.'}",http://arxiv.org/pdf/2302.00618
-Universal Captioner: Inducing Content-Style Separation in Vision-and-Language Model Training,M. Cornia,"While captioning models have obtained compelling results in describing natural images, there is a growing effort to increase their capability of dealing with real-world concepts. In this paper, we address the task of generating fluent descriptions by training on a non-uniform combination of data sources, containing both human- and automatically-collected captions. To this end, we propose a model which induces a separation between content and descriptive style through the incorporation of stylistic parameters and keywords extracted from large-scale multi-modal models as pivotal data. In terms of visual features, our model avoids the need of object detectors and employs grid-like features together with a single objective of prompt language modeling. Experimentally, we consistently outperform existing methods in terms of caption quality and capability of describing out-of-domain concepts. Finally, our model obtains a new state of the art on both COCO and nocaps.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a model which induces a separation between content and descriptive style through the incorporation of stylistic parameters and keywords extracted from large-scale multi-modal models as pivotal data, and obtains a new state of the art on both COCO and nocaps.'}",
-Prompting Large Language Models With the Socratic Method,Edward Y. Chang,"This paper presents a systematic approach to using the Socratic method in developing prompt templates that effectively interact with large language models, including GPT-3. Various methods are examined, and those that yield precise answers and justifications while fostering creativity and imagination to enhance creative writing are identified. Techniques such as definition, elenchus, dialectic, maieutics, generalization, and counterfactual reasoning are discussed for their application in engineering prompt templates and their connections to inductive, deductive, and abductive reasoning. Through examples, the effectiveness of these dialogue and reasoning methods is demonstrated. An interesting observation is made that when the task's goal and user intent are conveyed to GPT-3 via ChatGPT before the start of a dialogue, the large language model seems to connect to the external context expressed in the intent and perform more effectively.","{'model': 'tldr@v2.0.0', 'text': ""An interesting observation is made that when the task's goal and user intent are conveyed to GPT-3 via ChatGPT before the start of a dialogue, the large language model seems to connect to the external context expressed in the intent and perform more effectively.""}",https://arxiv.org/pdf/2303.08769
-Prompting Large Language Models with Answer Heuristics for Knowledge-Based Visual Question Answering,Zhenwei Shao,"Knowledge-based visual question answering (VQA) requires external knowledge beyond the image to answer the question. Early studies retrieve required knowledge from explicit knowledge bases (KBs), which often introduces irrelevant information to the question, hence restricting the performance of their models. Recent works have sought to use a large language model (i.e., GPT-3 [3]) as an implicit knowledge engine to acquire the necessary knowledge for answering. Despite the encouraging results achieved by these methods, we argue that they have not fully activated the capacity of GPT-3 as the provided input information is insufficient. In this paper, we present Prophet-a conceptually simple framework designed to $prompt$ GPT-3 with answer heuristics for knowledge-based VQA. Specifically, we first train a vanilla VQA model on a specific knowledge-based VQA dataset without external knowledge. After that, we extract two types of complementary answer heuristics from the model: answer candidates and answer-aware examples. Finally, the two types of answer heuristics are encoded into the prompts to enable GPT-3 to better comprehend the task thus enhancing its capacity. Prophet significantly outperforms all existing state-of-the-art methods on two challenging knowledge-based VQA datasets, OK-VQA and A-OKVQA, delivering 61.1% and 55.7% accuracies on their testing sets, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Prophet-a conceptually simple framework designed to enable GPT-3 to better comprehend the task thus enhancing its capacity, which significantly outperforms all existing state-of-the-art methods on two challenging knowledge-based VQA datasets, OK-VQA and A-OKVZA.'}",https://arxiv.org/pdf/2303.01903
-Large Language Models are Effective Text Rankers with Pairwise Ranking Prompting,Zhen Qin,"Ranking documents using Large Language Models (LLMs) by directly feeding the query and candidate documents into the prompt is an interesting and practical problem. However, there has been limited success so far, as researchers have found it difficult to outperform fine-tuned baseline rankers on benchmark datasets. We analyze pointwise and listwise ranking prompts used by existing methods and argue that off-the-shelf LLMs do not fully understand these ranking formulations, possibly due to the nature of how LLMs are trained. In this paper, we propose to significantly reduce the burden on LLMs by using a new technique called Pairwise Ranking Prompting (PRP). Our results are the first in the literature to achieve state-of-the-art ranking performance on standard benchmarks using moderate-sized open-sourced LLMs. On TREC-DL2020, PRP based on the Flan-UL2 model with 20B parameters outperforms the previous best approach in the literature, which is based on the blackbox commercial GPT-4 that has 50x (estimated) model size, by over 5% at NDCG@1. On TREC-DL2019, PRP is only inferior to the GPT-4 solution on the NDCG@5 and NDCG@10 metrics, while outperforming other existing solutions, such as InstructGPT which has 175B parameters, by over 10% for nearly all ranking metrics. Furthermore, we propose several variants of PRP to improve efficiency and show that it is possible to achieve competitive results even with linear complexity. We also discuss other benefits of PRP, such as supporting both generation and scoring LLM APIs, as well as being insensitive to input ordering.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to significantly reduce the burden on LLMs by using a new technique called Pairwise Ranking Prompting (PRP), and is the first in the literature to achieve state-of-the-art ranking performance on standard benchmarks using moderate-sized open-sourced LLMs.'}",http://arxiv.org/pdf/2306.17563
-"ChatGPT is fun, but it is not funny! Humor is still challenging Large Language Models",Sophie F. Jentzsch,"Humor is a central aspect of human communication that has not been solved for artificial agents so far. Large language models (LLMs) are increasingly able to capture implicit and contextual information. Especially, OpenAI’s ChatGPT recently gained immense public attention. The GPT3-based model almost seems to communicate on a human level and can even tell jokes. Humor is an essential component of human communication. But is ChatGPT really funny?We put ChatGPT’s sense of humor to the test. In a series of exploratory experiments around jokes, i.e., generation, explanation, and detection, we seek to understand ChatGPT’s capability to grasp and reproduce human humor. Since the model itself is not accessible, we applied prompt-based experiments. Our empirical evidence indicates that jokes are not hard-coded but mostly also not newly generated by the model. Over 90% of 1008 generated jokes were the same 25 Jokes. The system accurately explains valid jokes but also comes up with fictional explanations for invalid jokes. Joke-typical characteristics can mislead ChatGPT in the classification of jokes. ChatGPT has not solved computational humor yet but it can be a big leap toward “funny” machines.","{'model': 'tldr@v2.0.0', 'text': 'The empirical evidence indicates that jokes are not hard-coded but mostly also not newly generated by the model, which means ChatGPT can be a big leap toward “funny” machines.'}",http://arxiv.org/pdf/2306.04563
-Exploring the MIT Mathematics and EECS Curriculum Using Large Language Models,Sarah J. Zhang,"We curate a comprehensive dataset of 4,550 questions and solutions from problem sets, midterm exams, and final exams across all MIT Mathematics and Electrical Engineering and Computer Science (EECS) courses required for obtaining a degree. We evaluate the ability of large language models to fulfill the graduation requirements for any MIT major in Mathematics and EECS. Our results demonstrate that GPT-3.5 successfully solves a third of the entire MIT curriculum, while GPT-4, with prompt engineering, achieves a perfect solve rate on a test set excluding questions based on images. We fine-tune an open-source large language model on this dataset. We employ GPT-4 to automatically grade model responses, providing a detailed performance breakdown by course, question, and answer type. By embedding questions in a low-dimensional space, we explore the relationships between questions, topics, and classes and discover which questions and classes are required for solving other questions and classes through few-shot learning. Our analysis offers valuable insights into course prerequisites and curriculum design, highlighting language models' potential for learning and improving Mathematics and EECS education.","{'model': 'tldr@v2.0.0', 'text': ""A comprehensive dataset of 4,550 questions and solutions from problem sets, midterm exams, and final exams across all MIT Mathematics and Electrical Engineering and Computer Science courses required for obtaining a degree is curated, highlighting language models' potential for learning and improving Mathematics and EECS education.""}",http://arxiv.org/pdf/2306.08997
-Retroformer: Retrospective Large Language Agents with Policy Gradient Optimization,Weiran Yao,"Recent months have seen the emergence of a powerful new trend in which large language models (LLMs) are augmented to become autonomous language agents capable of performing objective oriented multi-step tasks on their own, rather than merely responding to queries from human users. Most existing language agents, however, are not optimized using environment-specific rewards. Although some agents enable iterative refinement through verbal feedback, they do not reason and plan in ways that are compatible with gradient-based learning from rewards. This paper introduces a principled framework for reinforcing large language agents by learning a retrospective model, which automatically tunes the language agent prompts from environment feedback through policy gradient. Specifically, our proposed agent architecture learns from rewards across multiple environments and tasks, for fine-tuning a pre-trained language model which refines the language agent prompt by summarizing the root cause of prior failed attempts and proposing action plans. Experimental results on various tasks demonstrate that the language agents improve over time and that our approach considerably outperforms baselines that do not properly leverage gradients from the environment. This demonstrates that using policy gradient optimization to improve language agents, for which we believe our work is one of the first, seems promising and can be applied to optimize other models in the agent architecture to enhance agent performances over time.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the language agents improve over time and that the approach considerably outperforms baselines that do not properly leverage gradients from the environment, and demonstrates that using policy gradient optimization to improve language agents seems promising and can be applied to optimize other models in the agent architecture to enhance agent performances over time.'}",https://arxiv.org/pdf/2308.02151
-"ChatGPT, Large Language Models, and Generative AI as Future Augments of Surgical Cancer Care",A. Kothari,,"{'model': 'tldr@v2.0.0', 'text': 'ChatGPT’s popularity is a result of its advanced technology, versatility, and accessibility, which make it a valuable resource for a wide range of users and applications, and where it excelled by showcasing the capabilities of AI in understanding and responding to natural language.'}",https://link.springer.com/content/pdf/10.1245/s10434-023-13442-2.pdf
-Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programs,Alexander K. Lew,"Even after fine-tuning and reinforcement learning, large language models (LLMs) can be difficult, if not impossible, to control reliably with prompts alone. We propose a new inference-time approach to enforcing syntactic and semantic constraints on the outputs of LLMs, called sequential Monte Carlo (SMC) steering. The key idea is to specify language generation tasks as posterior inference problems in a class of discrete probabilistic sequence models, and replace standard decoding with sequential Monte Carlo inference. For a computational cost similar to that of beam search, SMC can steer LLMs to solve diverse tasks, including infilling, generation under syntactic constraints, and prompt intersection. To facilitate experimentation with SMC steering, we present a probabilistic programming library, LLaMPPL (https://github.com/probcomp/LLaMPPL), for concisely specifying new generation tasks as language model probabilistic programs, and automating steering of LLaMA-family Transformers.","{'model': 'tldr@v2.0.0', 'text': 'A Probabilistic programming library, LLaMPPL, is presented, for concisely specifying new generation tasks as language model probabilistic programs, and automating steering of LLaMA-family Transformers, to facilitate experimentation with SMC steering.'}",http://arxiv.org/pdf/2306.03081
-FinEval: A Chinese Financial Domain Knowledge Evaluation Benchmark for Large Language Models,Liwen Zhang,"Large language models (LLMs) have demonstrated exceptional performance in various natural language processing tasks, yet their efficacy in more challenging and domain-specific tasks remains largely unexplored. This paper presents FinEval, a benchmark specifically designed for the financial domain knowledge in the LLMs. FinEval is a collection of high-quality multiple-choice questions covering Finance, Economy, Accounting, and Certificate. It includes 4,661 questions spanning 34 different academic subjects. To ensure a comprehensive model performance evaluation, FinEval employs a range of prompt types, including zero-shot and few-shot prompts, as well as answer-only and chain-of-thought prompts. Evaluating state-of-the-art Chinese and English LLMs on FinEval, the results show that only GPT-4 achieved an accuracy close to 70% in different prompt settings, indicating significant growth potential for LLMs in the financial domain knowledge. Our work offers a more comprehensive financial knowledge evaluation benchmark, utilizing data of mock exams and covering a wide range of evaluated LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Evaluating state-of-the-art Chinese and English LLMs on FinEval, the results show that only GPT-4 achieved an accuracy close to 70% in different prompt settings, indicating significant growth potential for LLMs in the financial domain knowledge.'}",https://arxiv.org/pdf/2308.09975
-Leveraging Large Language Models for Mental Health Prediction via Online Text Data,Xuhai Xu,"The recent technology boost of large language models (LLMs) has empowered a variety of applications. However, there is very little research on understanding and improving LLMs’ capability for the mental health domain. In this work, we present the first comprehensive evaluation of multiple LLMs, including Alpaca, Alpaca-LoRA, and GPT-3.5, on various mental health prediction tasks via online text data. We conduct a wide range of experiments, covering zero-shot prompting, few-shot prompting, and instruction finetuning. The results indicate the promising yet limited performance of LLMs with zero-shot and few-shot prompt designs for mental health tasks. More importantly, our experiments show that instruction finetuning can significantly boost the performance of LLMs for all tasks simultaneously. Our best-finetuned model, Mental-Alpaca, outperforms GPT-3.5 (25 times bigger) by 16.7% on balanced accuracy and performs on par with the state-of-the-art task-specific model. We summarize our findings into a set of action guidelines for future researchers, engineers, and practitioners on how to empower LLMs with better mental health domain knowledge and become an expert in mental health prediction tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the first comprehensive evaluation of multiple LLMs, including AlPaca, Alpaca-LoRA, and GPT-3.5, on various mental health prediction tasks via online text data and shows that instruction finetuning can significantly boost the performance of LLMs for all tasks simultaneously.'}",https://arxiv.org/pdf/2307.14385
-Analyzing Chain-of-Thought Prompting in Large Language Models via Gradient-based Feature Attributions,Skyler Wu,"Chain-of-thought (CoT) prompting has been shown to empirically improve the accuracy of large language models (LLMs) on various question answering tasks. While understanding why CoT prompting is effective is crucial to ensuring that this phenomenon is a consequence of desired model behavior, little work has addressed this; nonetheless, such an understanding is a critical prerequisite for responsible model deployment. We address this question by leveraging gradient-based feature attribution methods which produce saliency scores that capture the influence of input tokens on model output. Specifically, we probe several open-source LLMs to investigate whether CoT prompting affects the relative importances they assign to particular input tokens. Our results indicate that while CoT prompting does not increase the magnitude of saliency scores attributed to semantically relevant tokens in the prompt compared to standard few-shot prompting, it increases the robustness of saliency scores to question perturbations and variations in model output.","{'model': 'tldr@v2.0.0', 'text': 'This work probes several open-source LLMs to investigate whether CoT prompting affects the relative importances they assign to particular input tokens, and results indicate that while coT prompting does not increase the magnitude of saliency scores attributed to semantically relevant tokens in the prompt compared to standard few-shot prompting, it increases the robustness ofsaliency scores to question perturbations and variations in model output.'}",https://arxiv.org/pdf/2307.13339
-"Utilizing Large Language Models to Simplify Radiology Reports: a comparative analysis of ChatGPT-3.5, ChatGPT-4.0, Google Bard, and Microsoft Bing",Bing,"This paper investigates the application of Large Language Models (LLMs), specifically OpenAI's ChatGPT-3.5, ChatGPT-4.0, Google Bard, and Microsoft Bing, in simplifying radiology reports, thus potentially enhancing patient understanding. We examined 254 anonymized radiology reports from diverse examination types and used three different prompts to guide the LLMs' simplification processes. The resulting simplified reports were evaluated using four established readability indices. All LLMs significantly simplified the reports, but performance varied based on the prompt used and the specific model. The ChatGPT models performed best when additional context was provided (i.e., specifying user as a patient or requesting simplification at the 7th grade level). Our findings suggest that LLMs can effectively simplify radiology reports, although improvements are needed to ensure accurate clinical representation and optimal readability. These models have the potential to improve patient health literacy, patient-provider communication, and ultimately, health outcomes.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that LLMs can effectively simplify radiology reports, although improvements are needed to ensure accurate clinical representation and optimal readability.'}",https://www.medrxiv.org/content/medrxiv/early/2023/06/07/2023.06.04.23290786.full.pdf
-Can Large Language Models Empower Molecular Property Prediction?,Chen Qian,"Molecular property prediction has gained significant attention due to its transformative potential in multiple scientific disciplines. Conventionally, a molecule graph can be represented either as a graph-structured data or a SMILES text. Recently, the rapid development of Large Language Models (LLMs) has revolutionized the field of NLP. Although it is natural to utilize LLMs to assist in understanding molecules represented by SMILES, the exploration of how LLMs will impact molecular property prediction is still in its early stage. In this work, we advance towards this objective through two perspectives: zero/few-shot molecular classification, and using the new explanations generated by LLMs as representations of molecules. To be specific, we first prompt LLMs to do in-context molecular classification and evaluate their performance. After that, we employ LLMs to generate semantically enriched explanations for the original SMILES and then leverage that to fine-tune a small-scale LM model for multiple downstream tasks. The experimental results highlight the superiority of text explanations as molecular representations across multiple benchmark datasets, and confirm the immense potential of LLMs in molecular property prediction tasks. Codes are available at \url{https://github.com/ChnQ/LLM4Mol}.","{'model': 'tldr@v2.0.0', 'text': 'This work employs LLMs to generate semantically enriched explanations for the original SMILES and then uses that to fine-tune a small-scale LM model for multiple downstream tasks and highlights the superiority of text explanations as molecular representations across multiple benchmark datasets.'}",https://arxiv.org/pdf/2307.07443
-Understanding the Effectiveness of Very Large Language Models on Dialog Evaluation,Jessica Huynh,"Language models have steadily increased in size over the past few years. They achieve a high level of performance on various natural language processing (NLP) tasks such as question answering and summarization. Large language models (LLMs) have been used for generation and can now output human-like text. Due to this, there are other downstream tasks in the realm of dialog that can now harness the LLMs' language understanding capabilities. Dialog evaluation is one task that this paper will explore. It concentrates on prompting with LLMs: BLOOM, OPT, GPT-3, Flan-T5, InstructDial and TNLGv2. The paper shows that the choice of datasets used for training a model contributes to how well it performs on a task as well as on how the prompt should be structured. Specifically, the more diverse and relevant the group of datasets that a model is trained on, the better dialog evaluation performs. This paper also investigates how the number of examples in the prompt and the type of example selection used affect the model's performance.","{'model': 'tldr@v2.0.0', 'text': ""The paper shows that the choice of datasets used for training a model contributes to how well it performs on a task as well as on how the prompt should be structured, and investigates how the number of examples in the prompt and the type of example selection used affect the model's performance.""}",http://arxiv.org/pdf/2301.12004
-Generating Data for Symbolic Language with Large Language Models,Jiacheng Ye,"While large language models (LLMs) bring not only performance but also complexity, recent work has started to turn LLMs into data generators rather than task inferencers, where another affordable task model is trained for efficient deployment and inference. However, such an approach has primarily been applied to natural language tasks and has not yet been explored for symbolic language tasks with complex structured outputs (e.g., semantic parsing and code generation). In this paper, we propose SymGen which utilizes LLMs for generating various annotation-expensive symbolic language data. SymGen consists of an informative prompt to steer generation and an agreement-based verifier to improve data correctness. We conduct extensive experiments on six symbolic language tasks across various settings. Compared with the LLMs, we demonstrate the 1\%-sized task model can achieve comparable or better performance, largely cutting inference and deployment costs. We also show that generated data with only a few human demonstrations can be as effective as over 10 times the amount of human-annotated data when training the task model, saving a considerable amount of annotation effort. SymGen sheds new light on data generation for complex tasks, and we release the code at \href{https://github.com/HKUNLP/SymGen}{https://github.com/HKUNLP/SymGen}.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes SymGen which utilizes LLMs for generating various annotation-expensive symbolic language data and demonstrates the 1\\%-sized task model can achieve comparable or better performance, largely cutting inference and deployment costs.'}",http://arxiv.org/pdf/2305.13917
-Identifying and Extracting Rare Disease Phenotypes with Large Language Models,Cathy Shyr,"Rare diseases (RDs) are collectively common and affect 300 million people worldwide. Accurate phenotyping is critical for informing diagnosis and treatment, but RD phenotypes are often embedded in unstructured text and time-consuming to extract manually. While natural language processing (NLP) models can perform named entity recognition (NER) to automate extraction, a major bottleneck is the development of a large, annotated corpus for model training. Recently, prompt learning emerged as an NLP paradigm that can lead to more generalizable results without any (zero-shot) or few labeled samples (few-shot). Despite growing interest in ChatGPT, a revolutionary large language model capable of following complex human prompts and generating high-quality responses, none have studied its NER performance for RDs in the zero- and few-shot settings. To this end, we engineered novel prompts aimed at extracting RD phenotypes and, to the best of our knowledge, are the first the establish a benchmark for evaluating ChatGPT's performance in these settings. We compared its performance to the traditional fine-tuning approach and conducted an in-depth error analysis. Overall, fine-tuning BioClinicalBERT resulted in higher performance (F1 of 0.689) than ChatGPT (F1 of 0.472 and 0.591 in the zero- and few-shot settings, respectively). Despite this, ChatGPT achieved similar or higher accuracy for certain entities (i.e., rare diseases and signs) in the one-shot setting (F1 of 0.776 and 0.725). This suggests that with appropriate prompt engineering, ChatGPT has the potential to match or outperform fine-tuned language models for certain entity types with just one labeled sample. While the proliferation of large language models may provide opportunities for supporting RD diagnosis and treatment, researchers and clinicians should critically evaluate model outputs and be well-informed of their limitations.","{'model': 'tldr@v2.0.0', 'text': ""This work engineered novel prompts aimed at extracting RD phenotypes and is the first the establish a benchmark for evaluating ChatGPT's performance in these settings, and compared its performance to the traditional fine-tuning approach and conducted an in-depth error analysis.""}",http://arxiv.org/pdf/2306.12656
-Prompt Tuning or Fine-Tuning - Investigating Relational Knowledge in Pre-Trained Language Models,L. Fichtel,"Extracting relational knowledge from large pre-trained language models by a cloze-style sentence serving as a query has shown promising results. In particular, language models can be queried similar to knowledge graphs. The performance of the relational fact extraction task depends significantly on the query sentence, also known under the term prompt . Tuning these prompts has shown to increase the precision on standard language models by a maximum of around 12% points. However, usually large amounts of data in the form of existing knowledge graph facts and large text corpora are needed to train the required additional model. In this work, we propose using a completely different approach: Instead of spending resources on training an additional model, we simply perform an adaptive fine-tuning of the pre-trained language model on the standard fill-mask task using a small training dataset of existing facts from a knowledge graph. We investigate the differences between complex prompting techniques and adaptive fine-tuning in an extensive evaluation. Remarkably, adaptive fine-tuning outperforms all baselines, even by using significantly fewer training facts. Additionally, we analyze the transfer learning capabilities of this adapted language model by training on a restricted set of relations to show that even fewer training relations are needed to achieve high knowledge extraction quality.","{'model': 'tldr@v2.0.0', 'text': 'This work performs an adaptive tuning of the pre-trained language model on the standard task using a small training dataset of existing facts from a knowledge graph to show that even fewer training relations are needed to achieve high knowledge extraction quality.'}",
-On Transferability of Prompt Tuning for Natural Language Understanding,Yusheng Su,"Prompt tuning (PT) is a promising parameter-efficient method to utilize extremely large pre-trained language models (PLMs), which could achieve comparable performance to full-parameter fine-tuning by only tuning a few soft prompts. However, compared to fine-tuning, PT empirically requires much more training steps. To explore whether we can improve the efficiency of PT by reusing trained soft prompts and sharing learned knowledge, we empirically investigate the transferability of soft prompts across different tasks and models. In cross-task transfer , we find that trained soft prompts can well transfer to similar tasks and initialize PT for them to accelerate training and improve performance. Moreover, to explore what factors influence prompts’ transferability across tasks, we investigate how to measure the prompt similarity and find that the overlapping rate of activated neurons highly correlates to the transferability. In cross-model transfer , we explore how to project the prompts of a PLM to another PLM and successfully train a kind of projector which can achieve non-trivial transfer performance on similar tasks. However, initializing PT with the projected prompts does not work well, which may be caused by optimization preferences and PLMs’ high redundancy. Our findings show that improving PT with knowledge transfer is possible and promising, while prompts’ cross-task transferability is generally better than the cross-model transferability.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that improving PT with knowledge transfer is possible and promising, while prompts’ cross-task transferability is generally better than the cross-model transferability.'}",
-Knowledge-Augmented Language Model Prompting for Zero-Shot Knowledge Graph Question Answering,Jinheon Baek,"Large Language Models (LLMs) are capable of performing zero-shot closed-book question answering tasks, based on their internal knowledge stored in parameters during pre-training. However, such internalized knowledge might be insufficient and incorrect, which could lead LLMs to generate factually wrong answers. Furthermore, fine-tuning LLMs to update their knowledge is expensive. To this end, we propose to augment the knowledge directly in the input of LLMs. Specifically, we first retrieve the relevant facts to the input question from the knowledge graph based on semantic similarities between the question and its associated facts. After that, we prepend the retrieved facts to the input question in the form of the prompt, which is then forwarded to LLMs to generate the answer. Our framework, Knowledge-Augmented language model PromptING (KAPING), requires no model training, thus completely zero-shot. We validate the performance of our KAPING framework on the knowledge graph question answering task, that aims to answer the user’s question based on facts over a knowledge graph, on which ours outperforms relevant zero-shot baselines by up to 48% in average, across multiple LLMs of various sizes.","{'model': 'tldr@v2.0.0', 'text': ""The framework, Knowledge-Augmented language model PromptING (KAPING), requires no model training, thus completely zero-shot, and aims to answer the user’s question based on facts over a knowledge graph, on which the authors' outperforms relevant zero- shot baselines by up to 48% in average, across multiple LLMs of various sizes.""}",http://arxiv.org/pdf/2306.04136
-PURR: Efficiently Editing Language Model Hallucinations by Denoising Language Model Corruptions,Anthony Chen,"The remarkable capabilities of large language models have been accompanied by a persistent drawback: the generation of false and unsubstantiated claims commonly known as""hallucinations"". To combat this issue, recent research has introduced approaches that involve editing and attributing the outputs of language models, particularly through prompt-based editing. However, the inference cost and speed of using large language models for editing currently bottleneck prompt-based methods. These bottlenecks motivate the training of compact editors, which is challenging due to the scarcity of training data for this purpose. To overcome these challenges, we exploit the power of large language models to introduce corruptions (i.e., noise) into text and subsequently fine-tune compact editors to denoise the corruptions by incorporating relevant evidence. Our methodology is entirely unsupervised and provides us with faux hallucinations for training in any domain. Our Petite Unsupervised Research and Revision model, PURR, not only improves attribution over existing editing methods based on fine-tuning and prompting, but also achieves faster execution times by orders of magnitude.","{'model': 'tldr@v2.0.0', 'text': 'The Petite Unsupervised Research and Revision model, PURR, not only improves attribution over existing editing methods based on fine-tuning and prompting, but also achieves faster execution times by orders of magnitude.'}",http://arxiv.org/pdf/2305.14908
-Revealing the Unwritten: Visual Investigation of Beam Search Trees to Address Language Model Prompting Challenges,Thilo Spinner,"The growing popularity of generative language models has amplified interest in interactive methods to guide model outputs. Prompt refinement is considered one of the most effective means to influence output among these methods. We identify several challenges associated with prompting large language models, categorized into data- and model-specific, linguistic, and socio-linguistic challenges. A comprehensive examination of model outputs, including runner-up candidates and their corresponding probabilities, is needed to address these issues. The beam search tree, the prevalent algorithm to sample model outputs, can inherently supply this information. Consequently, we introduce an interactive visual method for investigating the beam search tree, facilitating analysis of the decisions made by the model during generation. We quantitatively show the value of exposing the beam search tree and present five detailed analysis scenarios addressing the identified challenges. Our methodology validates existing results and offers additional insights.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces an interactive visual method for investigating the beam search tree, facilitating analysis of the decisions made by the model during generation, and quantitatively shows the value of exposing thebeam search tree.'}",
-Training language models to follow instructions with human feedback,Long Ouyang,"Making language models bigger does not inherently make them better at following a user's intent. For example, large language models can generate outputs that are untruthful, toxic, or simply not helpful to the user. In other words, these models are not aligned with their users. In this paper, we show an avenue for aligning language models with user intent on a wide range of tasks by fine-tuning with human feedback. Starting with a set of labeler-written prompts and prompts submitted through the OpenAI API, we collect a dataset of labeler demonstrations of the desired model behavior, which we use to fine-tune GPT-3 using supervised learning. We then collect a dataset of rankings of model outputs, which we use to further fine-tune this supervised model using reinforcement learning from human feedback. We call the resulting models InstructGPT. In human evaluations on our prompt distribution, outputs from the 1.3B parameter InstructGPT model are preferred to outputs from the 175B GPT-3, despite having 100x fewer parameters. Moreover, InstructGPT models show improvements in truthfulness and reductions in toxic output generation while having minimal performance regressions on public NLP datasets. Even though InstructGPT still makes simple mistakes, our results show that fine-tuning with human feedback is a promising direction for aligning language models with human intent.","{'model': 'tldr@v2.0.0', 'text': 'The results show that fine-tuning with human feedback is a promising direction for aligning language models with human intent and showing improvements in truthfulness and reductions in toxic output generation while having minimal performance regressions on public NLP datasets.'}",http://arxiv.org/pdf/2203.02155
-"Translating radiology reports into plain language using ChatGPT and GPT-4 with prompt learning: results, limitations, and potential",Qing Lyu,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.'}",https://vciba.springeropen.com/counter/pdf/10.1186/s42492-023-00136-5
-"Translating Radiology Reports into Plain Language using ChatGPT and GPT-4 with Prompt Learning: Promising Results, Limitations, and Potential",Qing Lyu,"The large language model called ChatGPT has drawn extensively attention because of its human-like expression and reasoning abilities. In this study, we investigate the feasibility of using ChatGPT in experiments on using ChatGPT to translate radiology reports into plain language for patients and healthcare providers so that they are educated for improved healthcare. Radiology reports from 62 low-dose chest CT lung cancer screening scans and 76 brain MRI metastases screening scans were collected in the first half of February for this study. According to the evaluation by radiologists, ChatGPT can successfully translate radiology reports into plain language with an average score of 4.27 in the five-point system with 0.08 places of information missing and 0.07 places of misinformation. In terms of the suggestions provided by ChatGPT, they are general relevant such as keeping following-up with doctors and closely monitoring any symptoms, and for about 37% of 138 cases in total ChatGPT offers specific suggestions based on findings in the report. ChatGPT also presents some randomness in its responses with occasionally over-simplified or neglected information, which can be mitigated using a more detailed prompt. Furthermore, ChatGPT results are compared with a newly released large model GPT-4, showing that GPT-4 can significantly improve the quality of translated reports. Our results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.","{'model': 'tldr@v2.0.0', 'text': 'The results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.'}",http://arxiv.org/pdf/2303.09038
-A Systematic Survey of Prompt Engineering on Vision-Language Foundation Models,Jindong Gu,"—Prompt engineering is a technique that involves augmenting a large pre-trained model with task-specific hints, known as prompts, to adapt the model to new tasks. Prompts can be created manually as natural language instructions or generated automatically as either natural language instructions or vector representations. Prompt engineering enables the ability to perform predictions based solely on prompts without updating model parameters, and the easier application of large pre-trained models in real-world tasks. In past years, Prompt engineering has been well-studied in natural language processing. Recently, it has also been intensively studied in vision-language modeling. However, there is currently a lack of a systematic overview of prompt engineering on pre-trained vision-language models. This paper aims to provide a comprehensive survey of cutting-edge research in prompt engineering on three types of vision-language models: multimodal-to-text generation models ( e.g., Flamingo), image-text matching models ( e.g., CLIP), and text-to-image generation models ( e.g., Stable Diffusion). For each type of model, a brief model summary, prompting methods, prompting-based applications, and the corresponding responsibility and integrity issues are summarized and discussed. Furthermore, the commonalities and differences between prompting on vision-language models, language models, and vision models are also discussed. The challenges, future directions, and research opportunities are summarized to foster future research on this topic.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to provide a comprehensive survey of cutting-edge research in prompt engineering on three types of vision-language models: multimodal-to-text generation models, image-text matching models, and text- to-image generation models.'}",https://arxiv.org/pdf/2307.12980
-Towards using Few-Shot Prompt Learning for Automating Model Completion,Meriem Ben Chaaben,We propose a simple yet a novel approach to improve completion in domain modeling activities. Our approach exploits the power of large language models by using few-shot prompt learning without the need to train or fine-tune those models with large datasets that are scarce in this field. We implemented our approach and tested it on the completion of static and dynamic domain diagrams. Our initial evaluation shows that such an approach is effective and can be integrated in different ways during the modeling activities.,"{'model': 'tldr@v2.0.0', 'text': 'The initial evaluation shows that the approach exploits the power of large language models by using few-shot prompt learning without the need to train or fine-tune those models with large datasets that are scarce in this field.'}",https://arxiv.org/pdf/2212.03404
-POUF: Prompt-oriented unsupervised fine-tuning for large pre-trained models,Korawat Tanwisuth,"Through prompting, large-scale pre-trained models have become more expressive and powerful, gaining significant attention in recent years. Though these big models have zero-shot capabilities, in general, labeled data are still required to adapt them to downstream tasks. To overcome this critical limitation, we propose an unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data. We demonstrate how to apply our method to both language-augmented vision and masked-language models by aligning the discrete distributions extracted from the prompts and target data. To verify our approach's applicability, we conduct extensive experiments on image classification, sentiment analysis, and natural language inference tasks. Across 13 image-related tasks and 15 language-related ones, the proposed approach achieves consistent improvements over the baselines.","{'model': 'tldr@v2.0.0', 'text': 'An unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data by aligning the discrete distributions extracted from the prompts and target data is proposed.'}",http://arxiv.org/pdf/2305.00350
-Model ensemble instead of prompt fusion: a sample-specific knowledge transfer method for few-shot prompt tuning,Xiangyu Peng,"Prompt tuning approaches, which learn task-specific soft prompts for a downstream task conditioning on frozen pre-trained models, have attracted growing interest due to its parameter efficiency. With large language models and sufficient training data, prompt tuning performs comparably to full-model tuning. However, with limited training samples in few-shot settings, prompt tuning fails to match the performance of full-model fine-tuning. In this work, we focus on improving the few-shot performance of prompt tuning by transferring knowledge from soft prompts of source tasks. Recognizing the good generalization capabilities of ensemble methods in low-data regime, we first experiment and show that a simple ensemble of model predictions based on different source prompts, outperforms existing multi-prompt knowledge transfer approaches such as source prompt fusion in the few-shot setting. Motivated by this observation, we further investigate model ensembles and propose Sample-specific Ensemble of Source Models (SESoM). SESoM learns to adjust the contribution of each source model for each target sample separately when ensembling source model outputs. Through this way, SESoM inherits the superior generalization of model ensemble approaches and simultaneously captures the sample-specific competence of each source prompt. We conduct experiments across a diverse set of eight NLP tasks using models of different scales (T5-{base, large, XL}) and find that SESoM consistently outperforms the existing models of the same as well as larger parametric scale by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates model ensembles and proposes Sample-specific Ensemble of Source Models (SESoM), a simple ensemble of model predictions based on different source prompts that consistently outperforms the existing models of the same as well as larger parametric scale by a large margin.'}",http://arxiv.org/pdf/2210.12587
-Attentional Mixtures of Soft Prompt Tuning for Parameter-efficient Multi-task Knowledge Sharing,Akari Asai,"This work introduces ATTEMPT (ATTEntional Mixture of Prompt Tuning), a new modular, multi-task, and parameterefficient language model (LM) tuning approach that combines knowledge transferred across different tasks via a mixture of soft prompts while keeping original LM unchanged. ATTEMPT interpolates a set of prompts trained on large-scale source tasks and a newly initialized target task prompt using instance-wise attention computed by a light-weight sub-network trained on multiple target tasks. ATTEMPT is parameter-efficient (e.g., updates 1,600 times fewer parameters than fine-tuning) and enables multi-task learning and flexible extensions; importantly, it is also more interpretable because it demonstrates which source tasks affect the final model decision on target tasks. Experimental results across 17 diverse datasets show that ATTEMPT improves prompt tuning by up to a 22% absolute performance gain and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use over ten times more parameters.1","{'model': 'tldr@v2.0.0', 'text': 'Experimental results across 17 diverse datasets show that ATTEMPT improves prompt tuning by up to a 22% absolute performance gain and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use over ten times more parameters.'}",https://arxiv.org/pdf/2205.11961
-Prompting Large Pre-trained Vision-Language Models For Compositional Concept Learning,Guangyue Xu,"This work explores the zero-shot compositional learning ability of large pre-trained vision-language models(VLMs) within the prompt-based learning framework and propose a model (\textit{PromptCompVL}) to solve the compositonal zero-shot learning (CZSL) problem. \textit{PromptCompVL} makes two design choices: first, it uses a soft-prompting instead of hard-prompting to inject learnable parameters to reprogram VLMs for compositional learning. Second, to address the compositional challenge, it uses the soft-embedding layer to learn primitive concepts in different combinations. By combining both soft-embedding and soft-prompting, \textit{PromptCompVL} achieves state-of-the-art performance on the MIT-States dataset. Furthermore, our proposed model achieves consistent improvement compared to other CLIP-based methods which shows the effectiveness of the proposed prompting strategies for CZSL.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the zero-shot compositional learning ability of large pre-trained vision-language models within the prompt-based learning framework and proposes a model that achieves consistent improvement compared to other CLIP-based methods which shows the effectiveness of the proposed prompting strategies for CZSL.'}",https://arxiv.org/pdf/2211.05077
-Quantifying Memorization Across Neural Language Models,Nicholas Carlini,"Large language models (LMs) have been shown to memorize parts of their training data, and when prompted appropriately, they will emit the memorized training data verbatim. This is undesirable because memorization violates privacy (exposing user data), degrades utility (repeated easy-to-memorize text is often low quality), and hurts fairness (some texts are memorized over others). We describe three log-linear relationships that quantify the degree to which LMs emit memorized training data. Memorization significantly grows as we increase (1) the capacity of a model, (2) the number of times an example has been duplicated, and (3) the number of tokens of context used to prompt the model. Surprisingly, we find the situation becomes more complicated when generalizing these results across model families. On the whole, we find that memorization in LMs is more prevalent than previously believed and will likely get worse as models continues to scale, at least without active mitigations.","{'model': 'tldr@v2.0.0', 'text': 'On the whole, this work finds that memorization in LMs is more prevalent than previously believed and will likely get worse as models continues to scale, at least without active mitigations.'}",
-ProQA: Structural Prompt-based Pre-training for Unified Question Answering,Wanjun Zhong,"Question Answering (QA) is a longstanding challenge in natural language processing. Existing QA works mostly focus on specific question types, knowledge domains, or reasoning skills. The specialty in QA research hinders systems from modeling commonalities between tasks and generalization for wider applications. To address this issue, we present ProQA, a unified QA paradigm that solves various tasks through a single model. ProQA takes a unified structural prompt as the bridge and improves the QA-centric ability by structural prompt-based pre-training. Through a structurally designed prompt-based input schema, ProQA concurrently models the knowledge generalization for all QA tasks while keeping the knowledge customization for every specific QA task. Furthermore, ProQA is pre-trained with structural prompt-formatted large-scale synthesized corpus, which empowers the model with the commonly-required QA ability. Experimental results on 11 QA benchmarks demonstrate that ProQA consistently boosts performance on both full data fine-tuning, few-shot learning, and zero-shot testing scenarios. Furthermore, ProQA exhibits strong ability in both continual learning and transfer learning by taking the advantages of the structural prompt.","{'model': 'tldr@v2.0.0', 'text': 'ProQA is a unified QA paradigm that solves various tasks through a single model that takes a unified structural prompt as the bridge and improves the QA-centric ability by structural prompt-based pre-training.'}",http://arxiv.org/pdf/2205.04040
-Novelty Controlled Paraphrase Generation with Retrieval Augmented Conditional Prompt Tuning,Jishnu Ray Chowdhury,"Paraphrase generation is a fundamental and long-standing task in natural language processing. In this paper, we concentrate on two contributions to the task: (1) we propose Retrieval Augmented Prompt Tuning (RAPT) as a parameter-efficient method to adapt large pre-trained language models for paraphrase generation; (2) we propose Novelty Conditioned RAPT (NC-RAPT) as a simple model-agnostic method of using specialized prompt tokens for controlled paraphrase generation with varying levels of lexical novelty. By conducting extensive experiments on four datasets, we demonstrate the effectiveness of the proposed approaches for retaining the semantic content of the original text while inducing lexical novelty in the generation.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of the proposed approaches for retaining the semantic content of the original text while inducing lexical novelty in the generation of paraphrase generation is demonstrated.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21297/21046
-Discovering Latent Knowledge in Language Models Without Supervision,Collin Burns,"Existing techniques for training language models can be misaligned with the truth: if we train models with imitation learning, they may reproduce errors that humans make; if we train them to generate text that humans rate highly, they may output errors that human evaluators can't detect. We propose circumventing this issue by directly finding latent knowledge inside the internal activations of a language model in a purely unsupervised way. Specifically, we introduce a method for accurately answering yes-no questions given only unlabeled model activations. It works by finding a direction in activation space that satisfies logical consistency properties, such as that a statement and its negation have opposite truth values. We show that despite using no supervision and no model outputs, our method can recover diverse knowledge represented in large language models: across 6 models and 10 question-answering datasets, it outperforms zero-shot accuracy by 4\% on average. We also find that it cuts prompt sensitivity in half and continues to maintain high accuracy even when models are prompted to generate incorrect answers. Our results provide an initial step toward discovering what language models know, distinct from what they say, even when we don't have access to explicit ground truth labels.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a method for accurately answering yes-no questions given only unlabeled model activations, and shows that despite using no supervision and no model outputs, the method can recover diverse knowledge represented in large language models.'}",https://arxiv.org/pdf/2212.03827
-DisCup: Discriminator Cooperative Unlikelihood Prompt-tuning for Controllable Text Generation,Hanqing Zhang,"Prompt learning with immensely large Casual Language Models (CLMs) has been shown promising for attribute-controllable text generation (CTG). However, vanilla prompt tuning tends to imitate training corpus characteristics beyond the control attributes, resulting in a poor generalization ability. Moreover, it is less able to capture the relationship between different attributes, further limiting the control performance. In this paper, we propose a new CTG approach, namely DisCup, which incorporates the attribute knowledge of discriminator to optimize the control-prompts, steering a frozen CLM to produce attribute-specific texts. Specifically, the frozen CLM model, capable of producing multitudinous texts, is first used to generate the next-token candidates based on the context, so as to ensure the diversity of tokens to be predicted. Then, we leverage an attribute-discriminator to select desired/undesired tokens from those candidates, providing the inter-attribute knowledge. Finally, we bridge the above two traits by an unlikelihood objective for prompt-tuning. Extensive experimental results show that DisCup can achieve a new state-of-the-art control performance while maintaining an efficient and high-quality text generation, only relying on around 10 virtual tokens.","{'model': 'tldr@v2.0.0', 'text': 'A new CTG approach, namely DisCup, is proposed, which incorporates the attribute knowledge of discriminator to optimize the control-prompts, steering a frozen CLM to produce attribute-specific texts.'}",http://arxiv.org/pdf/2210.09551
-Deep Continuous Prompt for Contrastive Learning of Sentence Embeddings,Yuxin Jiang,"The performance of sentence representation 001 has been remarkably improved by the frame- 002 work of contrastive learning. However, recent 003 works still require full fine-tuning, which is 004 quite inefficient for large-scaled pre-trained lan- 005 guage models. To this end, we present a novel 006 method which freezes the whole language 007 model and only optimizes the prefix deep con- 008 tinuous prompts. It not only tunes around 0.1% 009 parameters of the original language model, but 010 avoids the cumbersome computation of search- 011 ing handcrafted prompts. Experimental results 012 show that our proposed DCPCSE outperforms 013 the state-of-the-art method SimCSE by a large 014 margin. We raise the performance of unsuper- 015 vised BERT base and supervised RoBERTa large 016 by 2.24 and 1.00 points, respectively. Our code 017 will be released at Github. 018","{'model': 'tldr@v2.0.0', 'text': 'A novel 006 method which freezes the whole language 007 model and only optimizes the prefix deep con- 008 tinuous prompts and avoids the cumbersome computation of search- 011 ing handcrafted prompts is presented.'}",http://arxiv.org/pdf/2203.06875
-Language Models with Image Descriptors are Strong Few-Shot Video-Language Learners,Zhenhailong Wang,"The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples, such as domain-specific captioning, question answering, and future event prediction. Existing few-shot video-language learners focus exclusively on the encoder, resulting in the absence of a video-to-text decoder to handle generative tasks. Video captioners have been pretrained on large-scale video-language datasets, but they rely heavily on finetuning and lack the ability to generate text for unseen tasks in a few-shot setting. We propose VidIL, a few-shot Video-language Learner via Image and Language models, which demonstrates strong performance on few-shot video-to-text tasks without the necessity of pretraining or finetuning on any video datasets. We use the image-language models to translate the video content into frame captions, object, attribute, and event phrases, and compose them into a temporal structure template. We then instruct a language model, with a prompt containing a few in-context examples, to generate a target output from the composed content. The flexibility of prompting allows the model to capture any form of text input, such as automatic speech recognition (ASR) transcripts. Our experiments demonstrate the power of language models in understanding videos on a wide variety of video-language tasks, including video captioning, video question answering, video caption retrieval, and video future event prediction. Especially, on video future event prediction, our few-shot model significantly outperforms state-of-the-art supervised models trained on large-scale video datasets. Code and resources are publicly available for research purposes at https://github.com/MikeWangWZHL/VidIL .","{'model': 'tldr@v2.0.0', 'text': 'The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples, such as domain-specific captioning, question answering, and future event prediction, and to demonstrate the power of language models in understanding videos on a wide variety of video- language tasks.'}",http://arxiv.org/pdf/2205.10747
-Improving the Sample Efficiency of Prompt Tuning with Domain Adaptation,Xu Guo,"Prompt tuning, or the conditioning of a frozen pretrained language model (PLM) with soft prompts learned from data, has demonstrated impressive performance on a wide range of NLP tasks. However, prompt tuning requires a large training dataset to be effective and is outperformed by finetuning the entire PLM in data-scarce regimes. Previous work (Gu et al., 2022, Vu et al., 2022) proposed to transfer soft prompts pretrained on the source domain to the target domain. In this paper, we explore domain adaptation for prompt tuning, a problem setting where unlabeled data from the target domain are available during pretraining. We propose bOosting Prompt TunIng with doMain Adaptation (OPTIMA), which regularizes the decision boundary to be smooth around regions where source and target data distributions are similar. Extensive experiments demonstrate that OPTIMA significantly enhances the transferability and sample-efficiency of prompt tuning compared to strong baselines. Moreover, in few-shot settings, OPTIMA exceeds full-model tuning by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes bOosting Prompt TunIng with doMain Adaptation (OPTIMA), which regularizes the decision boundary to be smooth around regions where source and target data distributions are similar, and demonstrates that OPTIMA significantly enhances the transferability and sample-efficiency of prompt tuning compared to strong baselines.'}",http://arxiv.org/pdf/2210.02952
-Prompt-Augmented Linear Probing: Scaling Beyond The Limit of Few-shot In-Context Learners,Hyunsoo Cho,"Through in-context learning (ICL), large-scale language models are effective few-shot learners without additional model fine-tuning. However, the ICL performance does not scale well with the number of available training sample as it is limited by the inherent input length constraint of the underlying language model. Meanwhile, many studies have revealed that language models are also powerful feature extractors, allowing them to be utilized in a black-box manner and enabling the linear probing paradigm, where lightweight discriminators are trained on top of the pre-extracted input representations. This paper proposes prompt-augmented linear probing (PALP), a hybrid of linear probing and ICL, which leverages the best of both worlds. PALP inherits the scalability of linear probing and the capability of enforcing language models to derive more meaningful representations via tailoring input into a more conceivable form. Throughout in-depth investigations on various datasets, we verified that PALP significantly closes the gap between ICL in the data-hungry scenario and fine-tuning in the data-abundant scenario with little training overhead, potentially making PALP a strong alternative in a black-box scenario.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2212.10873
-Are Prompt-based Models Clueless?,Pride Kavumba,"Finetuning large pre-trained language models with a task-specific head has advanced the state-of-the-art on many natural language understanding benchmarks. However, models with a task-specific head require a lot of training data, making them susceptible to learning and exploiting dataset-specific superficial cues that do not generalize to other datasets.Prompting has reduced the data requirement by reusing the language model head and formatting the task input to match the pre-training objective. Therefore, it is expected that few-shot prompt-based models do not exploit superficial cues.This paper presents an empirical examination of whether few-shot prompt-based models also exploit superficial cues.Analyzing few-shot prompt-based models on MNLI, SNLI, HANS, and COPA has revealed that prompt-based models also exploit superficial cues. While the models perform well on instances with superficial cues, they often underperform or only marginally outperform random accuracy on instances without superficial cues.","{'model': 'tldr@v2.0.0', 'text': 'Analyzing few-shot prompt- based models on MNLI, SNLI, HANS, and COPA has revealed that prompt-based models also exploit superficial cues, and while the models perform well on instances with superficial cue, they often underperform or only marginally outperform random accuracy on instances without superficial cues.'}",https://arxiv.org/pdf/2205.09295
-On Measuring Social Biases in Prompt-Based Multi-Task Learning,Afra Feyza Akyurek,"Large language models trained on a mixture of NLP tasks that are converted into a text-to-text format using prompts, can generalize into novel forms of language and handle novel tasks. A large body of work within prompt engineering attempts to understand the effects of input forms and prompts in achieving superior performance. We consider an alternative measure and inquire whether the way in which an input is encoded affects social biases promoted in outputs. In this paper, we study T0, a large-scale multi-task text-to-text language model trained using prompt-based learning. We consider two different forms of semantically equivalent inputs: question-answer format and premise-hypothesis format. We use an existing bias benchmark for the former BBQ and create the first bias benchmark in natural language inference BBNLI with hand-written hypotheses while also converting each benchmark into the other form. The results on two benchmarks suggest that given two different formulations of essentially the same input, T0 conspicuously acts more biased in question answering form, which is seen during training, compared to premise-hypothesis form which is unlike its training examples. Code and data are released under https://github.com/feyzaakyurek/bbnli.","{'model': 'tldr@v2.0.0', 'text': 'The results on two benchmarks suggest that given two different formulations of essentially the same input, T0 conspicuously acts more biased in question answering form, which is seen during training, compared to premise-hypothesis form which is unlike its training examples.'}",https://arxiv.org/pdf/2205.11605
-Reduce Communication Costs and Preserve Privacy: Prompt Tuning Method in Federated Learning,Haodong Zhao,"Federated learning (FL) has enabled global model training on decentralized data in a privacy-preserving way by aggregating model updates. However, for many natural language processing (NLP) tasks that utilize pre-trained language models (PLMs) with large numbers of parameters, there are con- siderable communication costs associated with FL. Recently, prompt tuning, which tunes some soft prompts without mod- ifying PLMs, has achieved excellent performance as a new learning paradigm. Therefore we want to combine the two methods and explore the effect of prompt tuning under FL. In this paper, we propose ”FedPrompt” as the first work study prompt tuning in a model split learning way using FL, and prove that split learning greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution. This improves the efficiency of FL method while also protect-ing the data privacy in prompt tuning. In addition, like PLMs, prompts are uploaded and downloaded between public platforms and personal users, so we try to figure out whether there is still a backdoor threat using only soft prompt in FL scenarios. We further conduct backdoor attacks by data poisoning on FedPrompt. Our experiments show that normal backdoor attack can not achieve a high attack success rate, proving the robustness of FedPrompt. We hope this work can promote the application of prompt in FL and raise the awareness of the possible security threats.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that split learning greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution, which improves theency of FL method while also protect-ing the data privacy in prompt tuning.'}",http://arxiv.org/pdf/2208.12268
-Doubly Right Object Recognition: A Why Prompt for Visual Rationales,Chengzhi Mao,"Many visual recognition models are evaluated only on their classification accuracy, a metric for which they obtain strong performance. In this paper, we investigate whether computer vision models can also provide correct rationales for their predictions. We propose a “doubly right” object recognition benchmark, where the metric requires the model to simultaneously produce both the right labels as well as the right rationales. We find that state-of-the-art visual models, such as CLIP, often provide incorrect rationales for their categorical predictions. However, by transferring the rationales from language models into visual representations through a tailored dataset, we show that we can learn a “why prompt,” which adapts large visual representations to produce correct rationales. Visualizations and empirical experiments show that our prompts significantly improve performance on doubly right object recognition, in addition to zero-shot transfer to unseen tasks and datasets.","{'model': 'tldr@v2.0.0', 'text': 'It is found that state-of-the-art visual models, such as CLIP, often provide incorrect rationales for their categorical predictions, but by transferring the rationales from language models into visual representations through a tailored dataset, it is shown that a “why prompt” can be learned, which adapts large visual representations to produce correct rationales.'}",https://arxiv.org/pdf/2212.06202
-XPrompt: Exploring the Extreme of Prompt Tuning,Fang Ma,"Prompt tuning learns soft prompts to condition the frozen Pre-trained Language Models (PLMs) for performing downstream tasks in a parameter-efficient manner. While prompt tuning has gradually reached the performance level of fine-tuning as the model scale increases, there is still a large performance gap between prompt tuning and fine-tuning for models of moderate and small scales (typically less than 11B parameters). In this paper, we empirically show that the trained prompt tokens can have a negative impact on a downstream task and thus degrade its performance. To bridge the gap, we propose a novel Prompt tuning model with an eXtremely small scale (XPrompt) under the regime of lottery tickets hypothesis. Specifically, XPrompt eliminates the negative prompt tokens at different granularity levels through a hierarchical structured pruning, yielding a more parameter-efficient prompt yet with a competitive performance. Comprehensive experiments are carried out on the SuperGLUE tasks, and the results indicate that XPrompt is able to close the performance gap at smaller model scales.","{'model': 'tldr@v2.0.0', 'text': 'This paper empirically shows that the trained prompt tokens can have a negative impact on a downstream task and thus degrade its performance, and proposes a novel Prompt tuning model with an eXtremely small scale (XPrompt) under the regime of lottery tickets hypothesis that eliminates the negative prompt tokens at different granularity levels through a hierarchical structured pruning.'}",http://arxiv.org/pdf/2210.04457
-Automatic Prompt Augmentation and Selection with Chain-of-Thought from Labeled Data,Kashun Shum,"Chain-of-thought prompting (CoT) advances the reasoning abilities of large language models (LLMs) and achieves superior performance in arithmetic, commonsense, and symbolic reasoning tasks. However, most CoT studies rely on carefully designed human-annotated rational chains to prompt the language model, which poses challenges for real-world applications where labeled training data is available without human-annotated rational chains. This creates barriers to applications of CoT prompting to these general tasks. This paper proposes a new strategy, Automate-CoT (Automatic Prompt Augmentation and Selection with Chain-of-Thought), that can bypass human engineering of CoTs by automatically augmenting rational chains from a small labeled dataset, and then pruning low-quality chains to construct a candidate pool of machine-generated rationale chains based on the labels. Finally, it selects the optimal combination of several rationale chains from the pool for CoT prompting by employing a variance-reduced policy gradient strategy to estimate the significance of each example in a black-box language model. Automate-CoT enables a quick adaptation of the CoT technique to different tasks. Experimental results demonstrate the effectiveness of our method, where state-of-the-art results are achieved on arithmetic reasoning (+2.7\%), commonsense reasoning (+3.4\%), symbolic reasoning (+3.2\%), and non-reasoning tasks (+2.5\%). Our code will be available at https://github.com/shizhediao/automate-cot.","{'model': 'tldr@v2.0.0', 'text': 'A new strategy, Automate-CoT (Automatic Prompt Augmentation and Selection with Chain-of-Thought), that can bypass human engineering of CoTs by automatically augmenting rational chains from a small labeled dataset, and then pruning low-quality chains to construct a candidate pool of machine-generated rationale chains based on the labels.'}",http://arxiv.org/pdf/2302.12822
-Multitask Prompt Tuning Enables Parameter-Efficient Transfer Learning,Zhen Wang,"Prompt tuning, in which a base pretrained model is adapted to each task via conditioning on learned prompt vectors, has emerged as a promising approach for efficiently adapting large language models to multiple downstream tasks. However, existing methods typically learn soft prompt vectors from scratch, and it has not been clear how to exploit the rich cross-task knowledge with prompt vectors in a multitask learning setting. We propose multitask prompt tuning (MPT), which first learns a single transferable prompt by distilling knowledge from multiple task-specific source prompts. We then learn multiplicative low rank updates to this shared prompt to efficiently adapt it to each downstream target task. Extensive experiments on 23 NLP datasets demonstrate that our proposed approach outperforms the state-of-the-art methods, including the full finetuning baseline in some cases, despite only tuning 0.035% as many task-specific parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes multitask prompt tuning (MPT), which first learns a single transferable prompt by distilling knowledge from multiple task-specific source prompts, then learns multiplicative low rank updates to this shared prompt to efficiently adapt it to each downstream target task.'}",http://arxiv.org/pdf/2303.02861
-Declaration-based Prompt Tuning for Visual Question Answering,Yuhang Liu,"In recent years, the pre-training-then-fine-tuning paradigm has yielded immense success on a wide spectrum of cross-modal tasks, such as visual question answering (VQA), in which a visual-language (VL) model is first optimized via self-supervised task objectives, e.g., masked language modeling (MLM) and image-text matching (ITM), and then fine-tuned to adapt to downstream task (e.g., VQA) via a brand-new objective function, e.g., answer prediction. However, the inconsistency of the objective forms not only severely limits the generalization of pre-trained VL models to downstream tasks, but also requires a large amount of labeled data for fine-tuning. To alleviate the problem, we propose an innovative VL fine-tuning paradigm (named Declaration-based Prompt Tuning, abbreviated as DPT), which fine-tunes the model for downstream VQA using the pre-training objectives, boosting the effective adaptation of pre-trained models to the downstream task. Specifically, DPT reformulates the VQA task via (1) textual adaptation, which converts the given questions into declarative sentence form for prompt-tuning, and (2) task adaptation, which optimizes the objective function of VQA problem in the manner of pre-training phase. Experimental results on GQA dataset show that DPT outperforms the fine-tuned counterpart by a large margin regarding accuracy in both fully-supervised (2.68%) and zero-shot/fewshot (over 31%) settings. All the data and codes will be available to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'An innovative VL fine-tuning paradigm (named Declaration-based Prompt Tuning, abbreviated as DPT), which fine-tunes the model for downstream VQA using the pre-training objectives, boosting the effective adaptation of pre-trained models to the downstream task.'}",http://arxiv.org/pdf/2205.02456
-Prompt Generation Networks for Efficient Adaptation of Frozen Vision Transformers,J. Loedeman,"Large-scale pretrained models, especially those trained from vision-language data have demonstrated the tremendous value that can be gained from both larger training datasets and models. Thus, in order to benefit from these developments, there is renewed interest in transfer learning and adapting models from large-scale general pretraining to particular downstream tasks. However, the continuously increasing size of the models means that even the classic approach of finetuning is becoming infeasible for all but big institutions. Prompt leaning has emerged as a flexible way to adapt models by solely learning additional inputs to a model that is kept frozen, but so far performances remained inferior to finetuning. To address this, we propose the Prompt Generation Network (PGN) that generates input-dependent prompts by sampling from a learned library of tokens. We show the PGN is effective in adapting pretrained models to various new datasets. It surpasses previous prompt-learning methods by a large margin and even fullfinetuning on 5 out of 12 datasets while requiring 100x less parameters. PGN can even be used for training and inferring on multiple datasets simultaneously and learns to allocate tokens between domains. Given these findings, we conclude that PGN is a viable and scalable approach for downstream adaptation of frozen models. Code is available at https://github.com/jochemloedeman/PGN.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt Generation Network (PGN) is proposed that generates input-dependent prompts by sampling from a learned library of tokens and is shown to be effective in adapting pretrained models to various new datasets.'}",http://arxiv.org/pdf/2210.06466
-SPT: Semi-Parametric Prompt Tuning for Multitask Prompted Learning,M Saiful Bari,"Pre-trained large language models can efficiently interpolate human-written prompts in a natural way. Multitask prompted learning can help generalization through a diverse set of tasks at once, thus enhancing the potential for more effective downstream fine-tuning. To perform efficient multitask-inference in the same batch, parameter-efficient fine-tuning methods such as prompt tuning have been proposed. However, the existing prompt tuning methods may lack generalization. We propose SPT, a semi-parametric prompt tuning method for multitask prompted learning. The novel component of SPT is a memory bank from where memory prompts are retrieved based on discrete prompts. Extensive experiments, such as (i) fine-tuning a full language model with SPT on 31 different tasks from 8 different domains and evaluating zero-shot generalization on 9 heldout datasets under 5 NLP task categories and (ii) pretraining SPT on the GLUE datasets and evaluating fine-tuning on the SuperGLUE datasets, demonstrate effectiveness of SPT.","{'model': 'tldr@v2.0.0', 'text': 'SPT, a semi-parametric prompt tuning method for multitask prompted learning, which is a memory bank from where memory prompts are retrieved based on discrete prompts, is proposed.'}",http://arxiv.org/pdf/2212.10929
-GPS: Genetic Prompt Search for Efficient Few-Shot Learning,Hanwei Xu,"Prompt-based techniques have demostrated great potential for improving the few-shot generalization of pretrained language models. However, their performance heavily relies on the manual design of prompts and thus requiring a lot of human efforts. In this paper, we introduce Genetic Prompt Search (GPS) to improve few-shot learning with prompts, which utilizes a genetic algorithm to automatically search for the best prompt.GPS is gradient-free and requires no update of model parameters but only a small validation set. Experiments on diverse datasets proved the effectiveness of GPS, which outperforms manual prompts by a large margin of 2.6 points. Our method is also better than other parameter-efficient tuning methods such as prompt tuning.","{'model': 'tldr@v2.0.0', 'text': 'GPS is gradient-free and requires no update of model parameters but only a small validation set, which outperforms manual prompts by a large margin and is better than other parameter-efficient tuning methods such as prompt tuning.'}",https://arxiv.org/pdf/2210.17041
-CUP: Curriculum Learning based Prompt Tuning for Implicit Event Argument Extraction,Jiaju Lin,"Implicit event argument extraction (EAE) aims to identify arguments that could scatter over the document. Most previous work focuses on learning the direct relations between arguments and the given trigger, while the implicit relations with long-range dependency are not well studied. Moreover, recent neural network based approaches rely on a large amount of labeled data for training, which is unavailable due to the high labelling cost. In this paper, we propose a Curriculum learning based Prompt tuning (CUP) approach, which resolves implicit EAE by four learning stages. The stages are defined according to the relations with the trigger node in a semantic graph, which well captures the long-range dependency between arguments and the trigger. In addition, we integrate a prompt-based encoder-decoder model to elicit related knowledge from pre-trained language models (PLMs) in each stage, where the prompt templates are adapted with the learning progress to enhance the reasoning for arguments. Experimental results on two well-known benchmark datasets show the great advantages of our proposed approach. In particular, we outperform the state-of-the-art models in both fully-supervised and low-data scenarios.","{'model': 'tldr@v2.0.0', 'text': 'A Curriculum learning based Prompt tuning (CUP) approach, which resolves implicit EAE by four learning stages, which outperform the state-of-the-art models in both fully-supervised and low-data scenarios.'}",https://arxiv.org/pdf/2205.00498
-Zero-Label Prompt Selection,Chonghua Liao,"Natural language prompts have been shown to facilitate cross-task generalization for large language models. However, with no or limited labeled examples, the cross-task performance is highly sensitive to the choice of prompts, while selecting a high-performing prompt is challenging given the scarcity of labels. To address the issue, we propose a Zero-Label Prompt Selection (ZPS) method that selects prompts without any labeled data or gradient update. Specifically, given the candidate human-written prompts for a task, ZPS labels a set of unlabeled data with a prompt ensemble and uses the pseudo-labels for prompt selection. Experiments show that ZPS improves over prior methods by a sizeable margin in zero-label performance. We also extend ZPS to a few-shot setting and show its advantages over strong baselines such as prompt tuning and model tuning.","{'model': 'tldr@v2.0.0', 'text': 'A Zero-Label Prompt Selection (ZPS) method that selects prompts without any labeled data or gradient update is proposed that improves over prior methods by a sizeable margin in zero-label performance.'}",http://arxiv.org/pdf/2211.04668
-Clip-Tuning: Towards Derivative-free Prompt Learning with a Mixture of Rewards,Yekun Chai,"Derivative-free prompt learning has emerged as a lightweight alternative to prompt tuning, which only requires model inference to optimize the prompts. However, existing work did not take full advantage of the over-parameterized characteristics of large pre-trained language models (PLMs). In this paper, we propose Clip-Tuning, a simple yet effective method that adopts diverse frozen""thinned""networks of PLMs to obtain a mixture of rewards and thus advance the derivative-free prompt learning. The thinned networks consist of all the hidden units that survive a stationary dropout strategy, whose inference predictions reflect an ensemble of partial views over prompted training samples. Our method outperforms previous gradient-free prompt learning methods and achieves parity with gradient-based counterparts on seven language understanding benchmarks under few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Clip-Tuning is proposed, a simple yet effective method that adopts diverse frozen""thinned""networks of PLMs to obtain a mixture of rewards and thus advance the derivative-free prompt learning.'}",http://arxiv.org/pdf/2210.12050
-DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting,Yongming Rao,"Recent progress has shown that large-scale pre-training using contrastive image-text pairs can be a promising alternative for high-quality visual representation learning from natural language supervision. Benefiting from a broader source of supervision, this new paradigm exhibits impressive transferability to downstream classification tasks and datasets. However, the problem of transferring the knowledge learned from image-text pairs to more complex dense prediction tasks has barely been visited. In this work, we present a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP. Specifically, we convert the original image-text matching problem in CLIP to a pixel-text matching problem and use the pixel-text score maps to guide the learning of dense prediction models. By further using the contextual information from the image to prompt the language model, we are able to facilitate our model to better exploit the pretrained knowledge. Our method is model-agnostic, which can be applied to arbitrary dense prediction systems and various pre-trained visual backbones including both CLIP models and ImageNet pre-trained models. Extensive experiments demonstrate the superior performance of our methods on semantic segmentation, object detection, and instance segmentation tasks. Code is available at https://github.com/raoyongming/DenseCLIP.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP and converts the original image-text matching problem in CLIP to a pixel- text matching problem and uses the pixel-text score maps to guide the learning of dense prediction models.'}",https://arxiv.org/pdf/2112.01518
-UoM&MMU at TSAR-2022 Shared Task: Prompt Learning for Lexical Simplification,Laura Vásquez-Rodríguez,"We present PromptLS, a method for fine-tuning large pre-trained Language Models (LM) to perform the task of Lexical Simplification. We use a predefined template to attain appropriate replacements for a term, and fine-tune a LM using this template on language specific datasets. We filter candidate lists in post-processing to improve accuracy. We demonstrate that our model can work in a) a zero shot setting (where we only require a pre-trained LM), b) a fine-tuned setting (where language-specific data is required), and c) a multilingual setting (where the model is pre-trained across multiple languages and fine-tuned in an specific language). Experimental results show that, although the zero-shot setting is competitive, its performance is still far from the fine-tuned setting. Also, the multilingual is unsurprisingly worse than the fine-tuned model. Among all TSAR-2022 Shared Task participants, our team was ranked second in Spanish and third in English.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2022.tsar-1.23.pdf
-Bidirectional Language Models Are Also Few-shot Learners,Ajay Patel,"Large language models such as GPT-3 (Brown et al., 2020) can perform arbitrary tasks without undergoing fine-tuning after being prompted with only a few labeled examples. An arbitrary task can be reformulated as a natural language prompt, and a language model can be asked to generate the completion, indirectly performing the task in a paradigm known as prompt-based learning. To date, emergent prompt-based learning capabilities have mainly been demonstrated for unidirectional language models. However, bidirectional language models pre-trained on denoising objectives such as masked language modeling produce stronger learned representations for transfer learning. This motivates the possibility of prompting bidirectional models, but their pre-training objectives have made them largely incompatible with the existing prompting paradigm. We present SAP (Sequential Autoregressive Prompting), a technique that enables the prompting of bidirectional models. Utilizing the machine translation task as a case study, we prompt the bidirectional mT5 model (Xue et al., 2021) with SAP and demonstrate its few-shot and zero-shot translations outperform the few-shot translations of unidirectional models like GPT-3 and XGLM (Lin et al., 2021), despite mT5's approximately 50% fewer parameters. We further show SAP is effective on question answering and summarization. For the first time, our results demonstrate prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models.","{'model': 'tldr@v2.0.0', 'text': 'For the first time, prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models, and is shown to be effective on question answering and summarization.'}",http://arxiv.org/pdf/2209.14500
-SpeechPrompt v2: Prompt Tuning for Speech Classification Tasks,Kai-Wei Chang,"Prompt tuning is a technology that tunes a small set of parameters to steer a pre-trained language model (LM) to directly generate the output for downstream tasks. Recently, prompt tuning has demonstrated its storage and computation efficiency in both natural language processing (NLP) and speech processing fields. These advantages have also revealed prompt tuning as a candidate approach to serving pre-trained LM for multiple tasks in a unified manner. For speech processing, SpeechPrompt shows its high parameter efficiency and competitive performance on a few speech classification tasks. However, whether SpeechPrompt is capable of serving a large number of tasks is unanswered. In this work, we propose SpeechPrompt v2, a prompt tuning framework capable of performing a wide variety of speech classification tasks, covering multiple languages and prosody-related tasks. The experiment result shows that SpeechPrompt v2 achieves performance on par with prior works with less than 0.15M trainable parameters in a unified framework.","{'model': 'tldr@v2.0.0', 'text': 'SpeechPrompt v2 is proposed, a prompt tuning framework capable of performing a wide variety of speech classification tasks, covering multiple languages and prosody-related tasks, and achieves performance on par with prior works with less than 0.15M trainable parameters in a unified framework.'}",http://arxiv.org/pdf/2303.00733
-Language Models in the Loop: Incorporating Prompting into Weak Supervision,Ryan Smith,"We propose a new strategy for applying large pre-trained language models to novel tasks when labeled training data is limited. Rather than apply the model in a typical zero-shot or few-shot fashion, we treat the model as the basis for labeling functions in a weak supervision framework. To create a classifier, we first prompt the model to answer multiple distinct queries about an example and define how the possible responses should be mapped to votes for labels and abstentions. We then denoise these noisy label sources using the Snorkel system and train an end classifier with the resulting training data. Our experimental evaluation shows that prompting large language models within a weak supervision framework can provide significant gains in accuracy. On the WRENCH weak supervision benchmark, this approach can significantly improve over zero-shot performance, an average 19.5% reduction in errors. We also find that this approach produces classifiers with comparable or superior accuracy to those trained from hand-engineered rules.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new strategy for applying large pre-trained language models to novel tasks when labeled training data is limited and treats the model as the basis for labeling functions in a weak supervision framework, and produces classifiers with comparable or superior accuracy to those trained from hand-engineered rules.'}",http://arxiv.org/pdf/2205.02318
-Prompting as Probing: Using Language Models for Knowledge Base Construction,Dimitrios Alivanistos,"Language Models (LMs) have proven to be useful in various downstream applications, such as summarisation, translation, question answering and text classification. LMs are becoming increasingly important tools in Artificial Intelligence, because of the vast quantity of information they can store. In this work, we present ProP (Prompting as Probing), which utilizes GPT-3, a large Language Model originally proposed by OpenAI in 2020, to perform the task of Knowledge Base Construction (KBC). ProP implements a multi-step approach that combines a variety of prompting techniques to achieve this. Our results show that manual prompt curation is essential, that the LM must be encouraged to give answer sets of variable lengths, in particular including empty answer sets, that true/false questions are a useful device to increase precision on suggestions generated by the LM, that the size of the LM is a crucial factor, and that a dictionary of entity aliases improves the LM score. Our evaluation study indicates that these proposed techniques can substantially enhance the quality of the final predictions: ProP won track 2 of the LM-KBC competition, outperforming the baseline by 36.4 percentage points. Our implementation is available on https://github.com/HEmile/iswc-challenge.","{'model': 'tldr@v2.0.0', 'text': 'ProP (Prompting as Probing), which utilizes GPT-3, a large Language Model originally proposed by OpenAI in 2020, to perform the task of Knowledge Base Construction (KBC), implements a multi-step approach that combines a variety of prompting techniques to achieve this.'}",http://arxiv.org/pdf/2208.11057
-Understanding Zero-Shot Adversarial Robustness for Large-Scale Models,Chengzhi Mao,"Pretrained large-scale vision-language models like CLIP have exhibited strong generalization over unseen tasks. Yet imperceptible adversarial perturbations can significantly reduce CLIP's performance on new tasks. In this work, we identify and explore the problem of \emph{adapting large-scale models for zero-shot adversarial robustness}. We first identify two key factors during model adaption -- training losses and adaptation methods -- that affect the model's zero-shot adversarial robustness. We then propose a text-guided contrastive adversarial training loss, which aligns the text embeddings and the adversarial visual features with contrastive learning on a small set of training data. We apply this training loss to two adaption methods, model finetuning and visual prompt tuning. We find that visual prompt tuning is more effective in the absence of texts, while finetuning wins in the existence of text guidance. Overall, our approach significantly improves the zero-shot adversarial robustness over CLIP, seeing an average improvement of over 31 points over ImageNet and 15 zero-shot datasets. We hope this work can shed light on understanding the zero-shot adversarial robustness of large-scale models.","{'model': 'tldr@v2.0.0', 'text': ""This work identifies two key factors during model adaption -- training losses and adaptation methods -- that affect the model's zero-shot adversarial robustness, and proposes a text-guided contrastive adversarial training loss, which aligns the text embeddings and the adversarial visual features with contrastive learning on a small set of training data.""}",http://arxiv.org/pdf/2212.07016
-What does CLIP know about a red circle? Visual prompt engineering for VLMs,Aleksandar Shtedritski,"Large-scale Vision-Language Models, such as CLIP, learn powerful image-text representations that have found numerous applications, from zero-shot classification to text-to-image generation. Despite that, their capabilities for solving novel discriminative tasks via prompting fall behind those of large language models, such as GPT-3. Here we explore the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text. In particular, we discover an emergent ability of CLIP, where, by simply drawing a red circle around an object, we can direct the model's attention to that region, while also maintaining global information. We show the power of this simple approach by achieving state-of-the-art in zero-shot referring expressions comprehension and strong performance in keypoint localization tasks. Finally, we draw attention to some potential ethical concerns of large language-vision models.","{'model': 'tldr@v2.0.0', 'text': ""This work explores the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text, and discovers an emergent ability of CLIP, where, by simply drawing a red circle around an object, it can direct the model's attention to that region, while also maintaining global information.""}",https://arxiv.org/pdf/2304.06712
-Prompting Visual-Language Models for Efficient Video Understanding,Chen Ju,,"{'model': 'tldr@v2.0.0', 'text': 'A simple but strong baseline is presented to efficiently adapt the pre-trained I-VL model, and exploit its powerful ability for resource-hungry video understanding tasks, with minimal training, to optimise a few random vectors that convert video-related tasks into the same format as thePre-training objectives.'}",https://arxiv.org/pdf/2112.04478
-HealthPrompt: A Zero-shot Learning Paradigm for Clinical Natural Language Processing,S. Sivarajkumar,"Developing clinical natural language systems based on machine learning and deep learning is dependent on the availability of large-scale annotated clinical text datasets, most of which are time-consuming to create and not publicly available. The lack of such annotated datasets is the biggest bottleneck for the development of clinical NLP systems. Zero-Shot Learning (ZSL) refers to the use of deep learning models to classify instances from new classes of which no training data have been seen before. Prompt-based learning is an emerging ZSL technique in NLP where we define task-based templates for different tasks. In this study, we developed a novel prompt-based clinical NLP framework called HealthPrompt and applied the paradigm of prompt-based learning on clinical texts. In this technique, rather than fine-tuning a Pre-trained Language Model (PLM), the task definitions are tuned by defining a prompt template. We performed an in-depth analysis of HealthPrompt on six different PLMs in a no-training-data setting. Our experiments show that HealthPrompt could effectively capture the context of clinical texts and perform well for clinical NLP tasks without any training data.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt- based clinical NLP framework called HealthPrompt is developed and applied the paradigm of prompt-based learning on clinical texts and shows that it could effectively capture the context ofclinical texts and perform well forclinical NLP tasks without any training data.'}",
-VL-ADAPTER: Parameter-Efficient Transfer Learning for Vision-and-Language Tasks,Yi-Lin Sung,"Recently, fine-tuning language models pre-trained on large text corpora have provided huge improvements on vision-and-language (V&L) tasks as well as on pure language tasks. However, fine-tuning the entire parameter set of pre-trained models becomes impractical since the model size is growing rapidly. Hence, in this paper, we introduce adapter-based parameter-efficient transfer learning techniques to V&L models such as VL-BART and VL-T5. We evaluate our methods in a unified multi-task setup on both image-text and video-text benchmarks. For the image-text tasks, we use four diverse V&L datasets: VQAv2, GQA, NLVR2, and MSCOCO image captioning. For video-text tasks, we use TVQA, How2QA, TVC, and YC2C. With careful training and thorough experiments, we benchmark three popular adapter-based methods (Adapter, Hyperformer, Compacter) against the standard full fine-tuning and the recently proposed prompt-tuning approach. We also enhance the efficiency and performance of adapters by sharing their weights to attain knowledge across tasks. Our results demonstrate that training the adapter with the weight-sharing technique (4.18% of total parameters for image-text tasks and 3.39% for video-text tasks) can match the performance of fine-tuning the entire model. Lastly, we present a comprehensive analysis including the combination of adapter and task-specific prompts and the impact of V&L pre-training on adapters. 11The code for our CVPR 2022 paper is available at: https://github.com/ylsung/VL_adapter.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that training the adapter with the weight-sharing technique can match the performance of fine-tuning the entire model, and enhance the efficiency and performance of adapters by sharing their weights to attain knowledge across tasks.'}",https://arxiv.org/pdf/2112.06825
-An automatically discovered chain-of-thought prompt generalizes to novel models and datasets,Konstantin Hebenstreit,"Emergent chain-of-thought (CoT) reasoning capabilities promise to improve performance and explainability of large language models (LLMs). However, uncertainties remain about how reasoning strategies formulated for previous model generations generalize to new model generations and different datasets. In this small-scale study, we compare different reasoning strategies induced by zero-shot prompting across six recently released LLMs (davinci-002, davinci-003, GPT-3.5-turbo, GPT-4, Flan-T5-xxl and Cohere command-xlarge) on a mixture of six question-answering datasets, including datasets from scientific and medical domains. Our findings demonstrate that while some variations in effectiveness occur, gains from CoT reasoning strategies remain robust across different models and datasets. GPT-4 has the most benefit from current state-of-the-art reasoning strategies and exhibits the best performance by applying a prompt previously discovered through automated discovery.","{'model': 'tldr@v2.0.0', 'text': 'GPT-4 has the most benefit from current state-of-the-art reasoning strategies and exhibits the best performance by applying a prompt previously discovered through automated discovery.'}",https://arxiv.org/pdf/2305.02897
-PBNR: Prompt-based News Recommender System,Xinyi Li,"Online news platforms often use personalized news recommendation methods to help users discover articles that align with their interests. These methods typically predict a matching score between a user and a candidate article to reflect the user's preference for the article. Some previous works have used language model techniques, such as the attention mechanism, to capture users' interests based on their past behaviors, and to understand the content of articles. However, these existing model architectures require adjustments if additional information is taken into account. Pre-trained large language models, which can better capture word relationships and comprehend contexts, have seen a significant development in recent years, and these pre-trained models have the advantages of transfer learning and reducing the training time for downstream tasks. Meanwhile, prompt learning is a newly developed technique that leverages pre-trained language models by building task-specific guidance for output generations. To leverage textual information in news articles, this paper introduces the pre-trained large language model and prompt-learning to the community of news recommendation. The proposed model""prompt-based news recommendation""(PBNR) treats the personalized news recommendation as a text-to-text language task and designs personalized prompts to adapt to the pre-trained language model -- text-to-text transfer transformer (T5). Experimental studies using the Microsoft News dataset show that PBNR is capable of making accurate recommendations by taking into account various lengths of past behaviors of different users. PBNR can also easily adapt to new information without changing the model architecture and the training objective. Additionally, PBNR can make recommendations based on users' specific requirements, allowing human-computer interaction in the news recommendation field.","{'model': 'tldr@v2.0.0', 'text': 'The proposed model ""prompt-based news recommendation""(PBNR) treats the personalized news recommendation as a text-to-text language task and designs personalized prompts to adapt to the pre-trained language model -- text- to-text transfer transformer (T5).'}",http://arxiv.org/pdf/2304.07862
-How Does In-Context Learning Help Prompt Tuning?,Simeng Sun,"Fine-tuning large language models is becoming ever more impractical due to their rapidly-growing scale. This motivates the use of parameter-efficient adaptation methods such as prompt tuning (PT), which adds a small number of tunable embeddings to an otherwise frozen model, and in-context learning (ICL), in which demonstrations of the task are provided to the model in natural language without any additional training. Recently, Singhal et al. (2022) propose ``instruction prompt tuning'' (IPT), which combines PT with ICL by concatenating a natural language demonstration with learned prompt embeddings. While all of these methods have proven effective on different tasks, how they interact with each other remains unexplored. In this paper, we empirically study when and how in-context examples improve prompt tuning by measuring the effectiveness of ICL, PT, and IPT on five text generation tasks with multiple base language models. We observe that (1) IPT does \emph{not} always outperform PT, and in fact requires the in-context demonstration to be semantically similar to the test input to yield improvements; (2) PT is unstable and exhibits high variance, but combining PT and ICL (into IPT) consistently reduces variance across all five tasks; and (3) prompts learned for a specific source task via PT exhibit positive transfer when paired with in-context examples of a different target task. Our results offer actionable insights on choosing a suitable parameter-efficient adaptation method for a given task.","{'model': 'tldr@v2.0.0', 'text': 'This paper empirically study when and how in-context examples improve prompt tuning by measuring the effectiveness of ICL, PT, and IPT on five text generation tasks with multiple base language models and offers actionable insights on choosing a suitable parameter-efficient adaptation method for a given task.'}",http://arxiv.org/pdf/2302.11521
-P ROMPT C AP : Prompt-Guided Image Captioning for VQA with GPT-3,Yushi Hu,"Knowledge-based visual question answering (VQA) in-volves questions that require world knowledge beyond the image to yield the correct answer. Large language models (LMs) like GPT-3 are particularly helpful for this task because of their strong knowledge retrieval and reasoning capabilities. To enable LM to understand images, prior work uses a captioning model to convert images into text. However, when summarizing an image in a single caption sentence, which visual entities to describe are often underspecified. Generic image captions often miss visual details essential for the LM to answer visual questions correctly. To address this challenge, we propose P ROMPT C AP ( Prompt -guided image Cap tioning), a captioning model designed to serve as a better connector between images and black-box LMs. Different from generic captions, P ROMPT C AP takes a natural-language prompt to control the visual entities to describe in the generated caption. The prompt contains a question that the caption should aid in answering. To avoid extra annotation, P ROMPT C AP is trained by examples synthesized with GPT-3 and existing datasets. We demonstrate P ROMPT - C AP ’s effectiveness on an existing pipeline in which GPT-3 is prompted with image captions to carry out VQA. P ROMPT - C AP outperforms generic captions by a large margin and achieves state-of-the-art accuracy on knowledge-based VQA tasks (60.4% on OK-VQA and 59.6% on A-OKVQA). Zero-shot results on WebQA show that P ROMPT C AP generalizes well to unseen domains. 1","{'model': 'tldr@v2.0.0', 'text': 'P ROMPT C AP is a captioning model designed to serve as a better connector between images and black-box LMs that outperforms generic captions by a large margin and achieves state-of-the-art accuracy on knowledge-based VQA tasks.'}",
-Visual Clues: Bridging Vision and Language Foundations for Image Paragraph Captioning,Yujia Xie,"People say,""A picture is worth a thousand words"". Then how can we get the rich information out of the image? We argue that by using visual clues to bridge large pretrained vision foundation models and language models, we can do so without any extra cross-modal training. Thanks to the strong zero-shot capability of foundation models, we start by constructing a rich semantic representation of the image (e.g., image tags, object attributes / locations, captions) as a structured textual prompt, called visual clues, using a vision foundation model. Based on visual clues, we use large language model to produce a series of comprehensive descriptions for the visual content, which is then verified by the vision model again to select the candidate that aligns best with the image. We evaluate the quality of generated descriptions by quantitative and qualitative measurement. The results demonstrate the effectiveness of such a structured semantic representation.","{'model': 'tldr@v2.0.0', 'text': 'This work argues that by using visual clues to bridge large pretrained vision foundation models and language models, they can do so without any extra cross-modal training.'}",http://arxiv.org/pdf/2206.01843
-Response Generation with Context-Aware Prompt Learning,X. Gu,"Pre-trained language models (PLM) have marked a huge leap in neural dialogue modeling. While PLMs are pre-trained on large-scale text corpora, they are usually fine-tuned on scarce dialogue data with specific domain knowledge and dialogue styles. However, tailoring the language models while fully utilizing prior knowledge in large pre-trained models remains a challenge. In this paper, we present a novel approach for pre-trained dialogue modeling that casts the dialogue generation problem as a prompt-learning task. Instead of fine-tuning on limited dialogue data, our approach, DialogPrompt, learns continuous prompt embeddings optimized for dialogue contexts, which appropriately elicit knowledge from the large pre-trained model. To encourage the model to better utilize the prompt embeddings, the prompt encoders are designed to be dynamically generated based on the dialogue context. Experiments on popular conversation datasets show that our approach significantly outperforms the fine-tuning baseline and the generic prompt-learning methods. Furthermore, human evaluations strongly support the superiority of DialogPrompt in regard to response generation quality.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel approach for pre-trained dialogue modeling that casts the dialogue generation problem as a prompt-learning task and learns continuous prompt embeddings optimized for dialogue contexts, which appropriately elicit knowledge from the large pre- trained model.'}",
-Few-Shot Self-Rationalization with Natural Language Prompts,Ana Marasović,"Self-rationalization models that predict task labels and generate free-text elaborations for their predictions could enable more intuitive interaction with NLP systems. These models are, however, currently trained with a large amount of human-written free-text explanations for each task which hinders their broader usage. We propose to study a more realistic setting of self-rationalization using few training examples. We present FEB -- a standardized collection of four existing English-language datasets and associated metrics. We identify the right prompting approach by extensively exploring natural language prompts on FEB. Then, by using this prompt and scaling the model size, we demonstrate that making progress on few-shot self-rationalization is possible. We show there is still ample room for improvement in this task: the average plausibility of generated explanations assessed by human annotators is at most 51% (with GPT-3), while plausibility of human explanations is 76%. We hope that FEB and our proposed approach will spur the community to take on the few-shot self-rationalization challenge.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies the right prompting approach by extensively exploring natural language prompts on FEB and demonstrates that making progress on few-shot self-rationalization is possible, and presents FEB -- a standardized collection of four existing English-language datasets and associated metrics.'}",https://aclanthology.org/2022.findings-naacl.31.pdf
-Controllable Generation from Pre-trained Language Models via Inverse Prompting,Xu Zou,"Large-scale pre-trained language models have demonstrated strong capabilities of generating realistic texts. However, it remains challenging to control the generation results. Previous approaches such as prompting are far from sufficient, and lack of controllability limits the usage of language models. To tackle this challenge, we propose an innovative method, inverse prompting, to better control text generation. The core idea of inverse prompting is to use generated text to inversely predict the prompt during beam search, which enhances the relevance between the prompt and the generated text and thus improves controllability. Empirically, we pre-train a large-scale Chinese language model to perform a systematic study using human evaluation on the tasks of open-domain poem generation and open-domain long-form question answering. Results demonstrate that our proposed method substantially outperforms the baselines and that our generation quality is close to human performance on some of the tasks.","{'model': 'tldr@v2.0.0', 'text': 'The core idea of inverse prompting is to use generated text to inversely predict the prompt during beam search, which enhances the relevance between the prompt and the generated text and thus improves controllability.'}",https://arxiv.org/pdf/2103.10685
-WinoDict: Probing language models for in-context word acquisition,Julian Martin Eisenschlos,"We introduce a new in-context learning paradigm to measure Large Language Models’ (LLMs) ability to learn novel words during inference. In particular, we rewrite Winograd-style co-reference resolution problems by replacing the key concept word with a synthetic but plausible word that the model must understand to complete the task. Solving this task requires the model to make use of the dictionary definition of the new word given in the prompt. This benchmark addresses word acquisition, one important aspect of the diachronic degradation known to afflict LLMs. As LLMs are frozen in time at the moment they are trained, they are normally unable to reflect the way language changes over time. We show that the accuracy of LLMs compared to the original Winograd tasks decreases radically in our benchmark, thus identifying a limitation of current models and providing a benchmark to measure future improvements in LLMs ability to do in-context learning.","{'model': 'tldr@v2.0.0', 'text': 'A new in-context learning paradigm is introduced to measure Large Language Models’ (LLMs) ability to learn novel words during inference by rewriting Winograd-style co-reference resolution problems by replacing the key concept word with a synthetic but plausible word that the model must understand to complete the task.'}",http://arxiv.org/pdf/2209.12153
-Progressive Prompts: Continual Learning for Language Models,Anastasia Razdaibiedina,"We introduce Progressive Prompts - a simple and efficient approach for continual learning in language models. Our method allows forward transfer and resists catastrophic forgetting, without relying on data replay or a large number of task-specific parameters. Progressive Prompts learns a new soft prompt for each task and sequentially concatenates it with the previously learned prompts, while keeping the base model frozen. Experiments on standard continual learning benchmarks show that our approach outperforms state-of-the-art methods, with an improvement>20% in average test accuracy over the previous best-preforming method on T5 model. We also explore a more challenging continual learning setup with longer sequences of tasks and show that Progressive Prompts significantly outperforms prior methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Progressive Prompts - a simple and efficient approach for continual learning in language models that allows forward transfer and resists catastrophic forgetting, without relying on data replay or a large number of task-specific parameters.'}",http://arxiv.org/pdf/2301.12314
-MotionGPT: Human Motion as a Foreign Language,Biao Jiang,"Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this""motion vocabulary"", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks that achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.'}",https://arxiv.org/pdf/2306.14795
-Boosting Natural Language Generation from Instructions with Meta-Learning,Budhaditya Deb,"Recent work has shown that language models (LMs) trained with multi-task instructional learning (MTIL) can solve diverse NLP tasks in zero- and few-shot settings with improved performance compared to prompt tuning. MTIL illustrates that LMs can extract and use information about the task from instructions beyond the surface patterns of the inputs and outputs. This suggests that meta-learning may further enhance the utilization of instructions for effective task transfer. In this paper we investigate whether meta-learning applied to MTIL can further improve generalization to unseen tasks in a zero-shot setting. Specifically, we propose to adapt meta-learning to MTIL in three directions: 1) Model Agnostic Meta Learning (MAML), 2) Hyper-Network (HNet) based adaptation to generate task specific parameters conditioned on instructions, and 3) an approach combining HNet and MAML. Through extensive experiments on the large scale Natural Instructions V2 dataset, we show that our proposed approaches significantly improve over strong baselines in zero-shot settings. In particular, meta-learning improves the effectiveness of instructions and is most impactful when the test tasks are strictly zero-shot (i.e. no similar tasks in the training set) and are “hard” for LMs, illustrating the potential of meta-learning for MTIL for out-of-distribution tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to adapt meta-learning to MTIL in three directions: Model Agnostic Meta Learning (MAML), Hyper-Network (HNet) based adaptation to generate task specific parameters conditioned on instructions, and an approach combining HNet and MAML.'}",http://arxiv.org/pdf/2210.11617
-Activation Addition: Steering Language Models Without Optimization,A. Turner,"Reliably controlling the behavior of large language models is a pressing open problem. Existing methods include supervised finetuning, reinforcement learning from human feedback, prompt engineering, and guided decoding. We instead investigate activation engineering: modifying activations at inference time to predictably alter model behavior. In particular, we bias the forward pass with an added 'steering vector' implicitly specified through natural language. Unlike past work which learned these steering vectors, our Activation Addition (ActAdd) method computes them by taking the activation differences that result from pairs of prompts. We demonstrate ActAdd on GPT-2 on OpenWebText and ConceptNet. Our inference-time approach yields control over high-level properties of output and preserves off-target model performance. It involves far less compute and implementation effort than finetuning, allows users to provide natural language specifications, and its overhead scales naturally with model size.","{'model': 'tldr@v2.0.0', 'text': 'Activation engineering is investigated: modifying activations at inference time to predictably alter model behavior, which involves far less compute and implementation effort than finetuning, allows users to provide natural language specifications, and its overhead scales naturally with model size.'}",https://arxiv.org/pdf/2308.10248
-Strategic Reasoning with Language Models,Kanishk Gandhi,"Strategic reasoning enables agents to cooperate, communicate, and compete with other agents in diverse situations. Existing approaches to solving strategic games rely on extensive training, yielding strategies that do not generalize to new scenarios or games without retraining. Large Language Models (LLMs), with their ability to comprehend and generate complex, context-rich language, could prove powerful as tools for strategic gameplay. This paper introduces an approach that uses pretrained LLMs with few-shot chain-of-thought examples to enable strategic reasoning for AI agents. Our approach uses systematically generated demonstrations of reasoning about states, values, and beliefs to prompt the model. Using extensive variations of simple matrix games, we show that strategies that are derived based on systematically generated prompts generalize almost perfectly to new game structures, alternate objectives, and hidden information. Additionally, we demonstrate our approach can lead to human-like negotiation strategies in realistic scenarios without any extra training or fine-tuning. Our results highlight the ability of LLMs, guided by systematic reasoning demonstrations, to adapt and excel in diverse strategic scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an approach that uses pretrained LLMs with few-shot chain-of-thought examples to enable strategic reasoning for AI agents and shows that strategies that are derived based on systematically generated prompts generalize almost perfectly to new game structures, alternate objectives, and hidden information.'}",http://arxiv.org/pdf/2305.19165
-Respectful or Toxic? Using Zero-Shot Learning with Language Models to Detect Hate Speech,F. Plaza-Del-Arco,"Hate speech detection faces two significant challenges: 1) the limited availability of labeled data and 2) the high variability of hate speech across different contexts and languages. Prompting brings a ray of hope to these challenges. It allows injecting a model with task-specific knowledge without relying on labeled data. This paper explores zero-shot learning with prompting for hate speech detection. We investigate how well zero-shot learning can detect hate speech in 3 languages with limited labeled data. We experiment with various large language models and verbalizers on 8 benchmark datasets. Our findings highlight the impact of prompt selection on the results. They also suggest that prompting, specifically with recent large language models, can achieve performance comparable to and surpass fine-tuned models, making it a promising alternative for under-resourced languages. Our findings highlight the potential of prompting for hate speech detection and show how both the prompt and the model have a significant impact on achieving more accurate predictions in this task.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates how well zero-shot learning with prompting for hate speech detection in 3 languages with limited labeled data and suggests that prompting can achieve performance comparable to and surpass fine-tuned models, making it a promising alternative for under-resourced languages.'}",https://aclanthology.org/2023.woah-1.6.pdf
-Meta-tuning Language Models to Answer Prompts Better,Ruiqi Zhong,"Large pretrained language models like GPT-3 have acquired a surprising ability to perform zero-shot classification (ZSC). For example, to classify review sentiments, we can “prompt"" the language model with the review and the question “ Is the review positive? "" as the context, and ask it to predict whether the next word is “ Yes "" or “ No "". However, these models are not specialized for answering these prompts. To address this weak-ness, we propose meta-tuning, which trains the model to specialize in answering prompts but still generalize to unseen tasks. To create the training data, we aggregated 43 existing datasets, annotated 441 label descriptions in total, and unified them into the above question answering (QA) format. After meta-tuning, our model outperforms a same-sized QA model for most labels on unseen tasks, and we forecast that the performance would im-prove for even larger models. Therefore, measuring ZSC performance on non-specialized language models might underestimate their true capability, and community-wide efforts on aggregating datasets and unifying their formats can help build models that understand prompts better.","{'model': 'tldr@v2.0.0', 'text': None}",
-Zero-shot Translation of Attention Patterns in VQA Models to Natural Language,Leonard Salewski,"Converting a model's internals to text can yield human-understandable insights about the model. Inspired by the recent success of training-free approaches for image captioning, we propose ZS-A2T, a zero-shot framework that translates the transformer attention of a given model into natural language without requiring any training. We consider this in the context of Visual Question Answering (VQA). ZS-A2T builds on a pre-trained large language model (LLM), which receives a task prompt, question, and predicted answer, as inputs. The LLM is guided to select tokens which describe the regions in the input image that the VQA model attended to. Crucially, we determine this similarity by exploiting the text-image matching capabilities of the underlying VQA model. Our framework does not require any training and allows the drop-in replacement of different guiding sources (e.g. attribution instead of attention maps), or language models. We evaluate this novel task on textual explanation datasets for VQA, giving state-of-the-art performances for the zero-shot setting on GQA-REX and VQA-X. Our code is available at: https://github.com/ExplainableML/ZS-A2T.","{'model': 'tldr@v2.0.0', 'text': 'ZS-A2T is proposed, a zero-shot framework that translates the transformer attention of a given model into natural language without requiring any training, and allows the drop-in replacement of different guiding sources (e.g. attribution instead of attention maps), or language models.'}",
-"A Sign Language Recognition System with Pepper, Lightweight-Transformer, and LLM",Jongyoon Lim,"This research explores using lightweight deep neural network architectures to enable the humanoid robot Pepper to understand American Sign Language (ASL) and facilitate non-verbal human-robot interaction. First, we introduce a lightweight and efficient model for ASL understanding optimized for embedded systems, ensuring rapid sign recognition while conserving computational resources. Building upon this, we employ large language models (LLMs) for intelligent robot interactions. Through intricate prompt engineering, we tailor interactions to allow the Pepper Robot to generate natural Co-Speech Gesture responses, laying the foundation for more organic and intuitive humanoid-robot dialogues. Finally, we present an integrated software pipeline, embodying advancements in a socially aware AI interaction model. Leveraging the Pepper Robot's capabilities, we demonstrate the practicality and effectiveness of our approach in real-world scenarios. The results highlight a profound potential for enhancing human-robot interaction through non-verbal interactions, bridging communication gaps, and making technology more accessible and understandable.","{'model': 'tldr@v2.0.0', 'text': 'This research explores using lightweight deep neural network architectures to enable the humanoid robot Pepper to understand American Sign Language (ASL) and facilitate non-verbal human-robot interaction and presents an integrated software pipeline, embodying advancements in a socially aware AI interaction model.'}",https://arxiv.org/pdf/2309.16898
-Reducing Sentiment Bias in Language Models via Counterfactual Evaluation,Po-Sen Huang,"Advances in language modeling architectures and the availability of large text corpora have driven progress in automatic text generation. While this results in models capable of generating coherent texts, it also prompts models to internalize social biases present in the training corpus. This paper aims to quantify and reduce a particular type of bias exhibited by language models: bias in the sentiment of generated text. Given a conditioning context (e.g., a writing prompt) and a language model, we analyze if (and how) the sentiment of the generated text is affected by changes in values of sensitive attributes (e.g., country names, occupations, genders) in the conditioning context using a form of counterfactual evaluation. We quantify sentiment bias by adopting individual and group fairness metrics from the fair machine learning literature, and demonstrate that large-scale models trained on two different corpora (news articles, and Wikipedia) exhibit considerable levels of bias. We then propose embedding and sentiment prediction-derived regularization on the language model’s latent representations. The regularizations improve fairness metrics while retaining comparable levels of perplexity and semantic similarity.","{'model': 'tldr@v2.0.0', 'text': 'This paper quantifies sentiment bias by adopting individual and group fairness metrics from the fair machine learning literature, and proposes embedding and sentiment prediction-derived regularization on the language model’s latent representations.'}",https://www.aclweb.org/anthology/2020.findings-emnlp.7.pdf
-Question Decomposition Improves the Faithfulness of Model-Generated Reasoning,Ansh Radhakrishnan,"As large language models (LLMs) perform more difficult tasks, it becomes harder to verify the correctness and safety of their behavior. One approach to help with this issue is to prompt LLMs to externalize their reasoning, e.g., by having them generate step-by-step reasoning as they answer a question (Chain-of-Thought; CoT). The reasoning may enable us to check the process that models use to perform tasks. However, this approach relies on the stated reasoning faithfully reflecting the model's actual reasoning, which is not always the case. To improve over the faithfulness of CoT reasoning, we have models generate reasoning by decomposing questions into subquestions. Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT while improving the faithfulness of the model's stated reasoning on several recently-proposed metrics. By forcing the model to answer simpler subquestions in separate contexts, we greatly increase the faithfulness of model-generated reasoning over CoT, while still achieving some of the performance gains of CoT. Our results show it is possible to improve the faithfulness of model-generated reasoning; continued improvements may lead to reasoning that enables us to verify the correctness and safety of LLM behavior.","{'model': 'tldr@v2.0.0', 'text': ""Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT while improving the faithfulness of the model's stated reasoning on several recently-proposed metrics.""}",https://arxiv.org/pdf/2307.11768
-Improving Gender Fairness of Pre-Trained Language Models without Catastrophic Forgetting,Zahra Fatemi,"Existing studies addressing gender bias of pre-trained language models, usually build a small gender-neutral data set and conduct a second phase pre-training on the model with such data. However, given the limited size and concentrated focus of the gender-neutral data, catastrophic forgetting would occur during second-phase pre-training. Forgetting information in the original training data may damage the model’s downstream performance by a large margin. In this work, we empirically show that catastrophic forgetting occurs in such methods by evaluating them with general NLP tasks in GLUE. Then, we propose a new method, GEnder Equality Prompt (GEEP), to improve gender fairness of pre-trained models with less forgetting. GEEP freezes the pre-trained model and learns gender-related prompts with gender-neutral data.Empirical results show that GEEP not only achieves SOTA performances on gender fairness tasks, but also forgets less and performs better on GLUE by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'Empirical results show that GEEP not only achieves SOTA performances on gender fairness tasks, but also forgets less and performs better on GLUE by a large margin.'}",https://aclanthology.org/2023.acl-short.108.pdf
-Few-shot Instruction Prompts for Pretrained Language Models to Detect Social Biases,Shrimai Prabhumoye,"Detecting social bias in text is challenging due to nuance, subjectivity, and difficulty in obtaining good quality labeled datasets at scale, especially given the evolving nature of social biases and society. To address these challenges, we propose a few-shot instruction-based method for prompting pre-trained language models (LMs). We select a few class-balanced exemplars from a small support repository that are closest to the query to be labeled in the embedding space. We then provide the LM with instruction that consists of this subset of labeled exemplars, the query text to be classified, a definition of bias, and prompt it to make a decision. We demonstrate that large LMs used in a few-shot context can detect different types of fine-grained biases with similar and sometimes superior accuracy to fine-tuned models. We observe that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 13% improvement in AUC metric compared to other models). It also maintains a high AUC (dropping less than 2%) when the labeled repository is reduced to as few as $100$ samples. Large pretrained language models thus make it easier and quicker to build new bias detectors.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 13% improvement in AUC metric compared to other models) and makes it easier and quicker to build new bias detectors.'}",
-Simultaneous Translation and Paraphrase for Language Education,Stephen Mayhew,"We present the task of Simultaneous Translation and Paraphrasing for Language Education (STAPLE). Given a prompt in one language, the goal is to generate a diverse set of correct translations that language learners are likely to produce. This is motivated by the need to create and maintain large, high-quality sets of acceptable translations for exercises in a language-learning application, and synthesizes work spanning machine translation, MT evaluation, automatic paraphrasing, and language education technology. We developed a novel corpus with unique properties for five languages (Hungarian, Japanese, Korean, Portuguese, and Vietnamese), and report on the results of a shared task challenge which attracted 20 teams to solve the task. In our meta-analysis, we focus on three aspects of the resulting systems: external training corpus selection, model architecture and training decisions, and decoding and filtering strategies. We find that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to our provided learner response frequencies.","{'model': 'tldr@v2.0.0', 'text': 'It is found that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to the provided learner response frequencies.'}",https://www.aclweb.org/anthology/2020.ngt-1.28.pdf
-Prompt Injection: Parameterization of Fixed Inputs,Eunbi Choi,"Recent works have shown that attaching prompts to the input is effective at conditioning Language Models (LM) to perform specific tasks. However, prompts are always included in the input text during inference, thus incurring substantial computational and memory overhead. Also, there is currently no straightforward method of utilizing prompts that are longer than the maximum input length of the LMs without incurring additional costs during inference. We propose Prompt Injection (PI), a novel formulation of injecting the prompt into the parameters of an LM to be an efficient alternative to attaching fixed prompts to the input. We show that in scenarios with long fixed prompts, PI can be up to 280 times more efficient in terms of total FLOPs than previous approaches. We further explore methodologies for PI and show promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions. Through these explorations, we show that PI can be a promising direction for conditioning language models, especially in scenarios with long and fixed prompts.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prompt Injection (PI), a novel formulation of injecting the prompt into the parameters of an LM to be an efficient alternative to attaching fixed prompts to the input and shows that in scenarios with long fixed prompts, PI can be up to 280 times more efficient in terms of total FLOPs than previous approaches.'}",http://arxiv.org/pdf/2206.11349
-Prompt Injection Attacks and Defenses in LLM-Integrated Applications,Yupei Liu,"Large Language Models (LLMs) are increasingly deployed as the backend for a variety of real-world applications called LLM-Integrated Applications. Multiple recent works showed that LLM-Integrated Applications are vulnerable to prompt injection attacks, in which an attacker injects malicious instruction/data into the input of those applications such that they produce results as the attacker desires. However, existing works are limited to case studies. As a result, the literature lacks a systematic understanding of prompt injection attacks and their defenses. We aim to bridge the gap in this work. In particular, we propose a general framework to formalize prompt injection attacks. Existing attacks, which are discussed in research papers and blog posts, are special cases in our framework. Our framework enables us to design a new attack by combining existing attacks. Moreover, we also propose a framework to systematize defenses against prompt injection attacks. Using our frameworks, we conduct a systematic evaluation on prompt injection attacks and their defenses with 10 LLMs and 7 tasks. We hope our frameworks can inspire future research in this field. Our code is available at https://github.com/liu00222/Open-Prompt-Injection.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a general framework to formalize Prompt injection attacks and proposes a framework to systematize defenses against prompt injection attacks.'}",
-Tensor Trust: Interpretable Prompt Injection Attacks from an Online Game,S. Toyer,"While Large Language Models (LLMs) are increasingly being used in real-world applications, they remain vulnerable to prompt injection attacks: malicious third party prompts that subvert the intent of the system designer. To help researchers study this problem, we present a dataset of over 126,000 prompt injection attacks and 46,000 prompt-based""defenses""against prompt injection, all created by players of an online game called Tensor Trust. To the best of our knowledge, this is currently the largest dataset of human-generated adversarial examples for instruction-following LLMs. The attacks in our dataset have a lot of easily interpretable stucture, and shed light on the weaknesses of LLMs. We also use the dataset to create a benchmark for resistance to two types of prompt injection, which we refer to as prompt extraction and prompt hijacking. Our benchmark results show that many models are vulnerable to the attack strategies in the Tensor Trust dataset. Furthermore, we show that some attack strategies from the dataset generalize to deployed LLM-based applications, even though they have a very different set of constraints to the game. We release all data and source code at https://tensortrust.ai/paper","{'model': 'tldr@v2.0.0', 'text': 'A dataset of over 126,000 Prompt injection attacks and 46,000 prompt-based ""defenses"" against prompt injection, all created by players of an online game called Tensor Trust, is presented, currently the largest dataset of human-generated adversarial examples for instruction-following LLMs.'}",
-Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injection,Zekun Li,"Large Language Models (LLMs) have shown remarkable proficiency in following instructions, making them valuable in customer-facing applications. However, their impressive capabilities also raise concerns about the amplification of risks posed by adversarial instructions, which can be injected into the model input by third-party attackers to manipulate LLMs' original instructions and prompt unintended actions and content. Therefore, it is crucial to understand LLMs' ability to accurately discern which instructions to follow to ensure their safe deployment in real-world scenarios. In this paper, we propose a pioneering benchmark for automatically evaluating the robustness of instruction-following LLMs against adversarial instructions injected in the prompt. The objective of this benchmark is to quantify the extent to which LLMs are influenced by injected adversarial instructions and assess their ability to differentiate between these injected adversarial instructions and original user instructions. Through experiments conducted with state-of-the-art instruction-following LLMs, we uncover significant limitations in their robustness against adversarial instruction injection attacks. Furthermore, our findings indicate that prevalent instruction-tuned models are prone to being ``overfitted'' to follow any instruction phrase in the prompt without truly understanding which instructions should be followed. This highlights the need to address the challenge of training models to comprehend prompts instead of merely following instruction phrases and completing the text. The data and code can be found at \url{https://github.com/Leezekun/Adv-Instruct-Eval}.","{'model': 'tldr@v2.0.0', 'text': ""A pioneering benchmark for automatically evaluating the robustness of instruction-following LLMs against adversarial instructions injected in the prompt is proposed and it is indicated that prevalent instruction-tuned models are prone to being ``overfitted'' to follow any instruction phrase in the Prompt without truly understanding which instructions should be followed.""}",
-Calculation of prompt loss and toroidal field ripple loss under neutral beam injection on EAST,Bin Wu,"Neutral beam injection is a major auxiliary heating method in the EAST experimental campaign. This paper gives detailed calculations of beam loss with different plasma equilibria using the guiding center code ORBIT and NUBEAM/TRANSP. Increasing plasma current can dramatically lower the beam ion prompt loss and ripple loss. Countercurrent beam injection gives a much larger prompt loss fraction than co-injection, and ripple-induced collisionless stochastic diffusion is the dominant loss channel.",,
-"Prompt injection of relativistic protons from the September 1, 1971 solar flare",E. Cliver,,,
-THE ROLE OF STOCHASTIC ACCELERATION IN THE PROMPT EMISSION OF GAMMA-RAY BURSTS: APPLICATION TO HADRONIC INJECTION,K. Murase,"We study effects of particle re-acceleration (or heating) in the post-shock region via magnetohydrodynamic/plasma turbulence, in the context of a mixed hadronic–leptonic model for the prompt emission of gamma-ray bursts, using both analytical and numerical methods. We show that stochastically accelerated (or heated) leptons, which are injected via pp and pγ reactions and subsequent pair cascades, are plausibly able to reproduce the Band function spectra with α ∼ 1 and β ∼ 2–3 in the ∼MeV range. An additional hard component coming from the proton-induced cascade emission is simultaneously expected, which can be compatible with observed extra power-law spectra far above the MeV range. We also discuss the specific implications of hadronic models for ongoing high-energy neutrino observations.",,https://iopscience.iop.org/article/10.1088/0004-637X/746/2/164/pdf
-(Ab)using Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs,Eugene Bagdasaryan,"We demonstrate how images and sounds can be used for indirect prompt and instruction injection in multi-modal LLMs. An attacker generates an adversarial perturbation corresponding to the prompt and blends it into an image or audio recording. When the user asks the (unmodified, benign) model about the perturbed image or audio, the perturbation steers the model to output the attacker-chosen text and/or make the subsequent dialog follow the attacker's instruction. We illustrate this attack with several proof-of-concept examples targeting LLaVa and PandaGPT.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated how images and sounds can be used for indirect prompt and instruction injection in multi-modal LLMs by an attacker generating an adversarial perturbation corresponding to the prompt and blends it into an image or audio recording.'}",https://arxiv.org/pdf/2307.10490
-Regional grafting of autologous adipose tissue is effective in inducing prompt healing of indolent digital ulcers in patients with systemic sclerosis: results of a monocentric randomized controlled study,N. Del Papa,,"{'model': 'tldr@v2.0.0', 'text': 'This RCT strongly confirms that AT-G is effective in inducing ischemic digital ulcer healing in patients with systemic sclerosis patients.'}",https://arthritis-research.biomedcentral.com/track/pdf/10.1186/s13075-018-1792-8
-"Limitations of extended gas delivery tubes used for fuelling mitigated plasma disruptions, and a unique injection concept for prompt gas delivery",P. Parks,"Modelling high-pressure gas flows through long extended delivery tubes used for massive particle fuelling of tokamaks during a major disruptive instability or a preemptive fast plasma shutdown is presented using analytical theory and simulation. For steady-state flows, expressions were derived and compared with experiments for the transition diameter and inlet Mach number of a straight tube (pipe) attached to a ‘nozzle-like’ inlet valve, such that increases in pipe diameter have no effect on the flow rate (valve-limited flow), and decreases below the transition diameter cause decreasing flow rates (friction-limited flow). Analytical expressions for the exit outflow rate and other gas dynamic variables during the initial unsteady gas flow buildup were developed from the classical 1D centred expansion wave problem and compared with 2D axisymmetric FLUENT simulations with wall friction, and good agreement was found for sufficiently high-conductance pipes. The intrinsic time delay before steady-state outflow is reached can seriously limit plasma density increases during the disruption, as the disruption time scale is similar to the delay time or ‘rise time’ of the outflow at the exit plane. Thus, conditions required for strong collisional dissipation of destructive runaway electron currents can be compromised. A unique gas injection scheme ‘burst membrane gas injection’ is also presented in which a steady-state outflow at the exit plane can be established promptly once the membrane bursts. It is shown that the theoretically ideal rise time of the gas at the exit plane will be zero if the membrane had an instantaneous opening time, therefore the true rise time will be limited only by the rupture time of a real bursting disc. The duration of the initial steady-state gas delivery phase can be matched to the relevant disruption time scales (or runaway formation time) by simply adjusting the tube length.",,
-Review of conformal cooling system design and additive manufacturing for injection molds.,Zhiquan Wei,"This paper points out the significance of cooling in injection molding and briefly reviews the development of cooling systems. The focus of this survey is on the physical model, development, and optimization of conformal cooling systems which have curved cooling circuits following the shape of mold cavity. Compared with traditional cooling systems, conformal cooling can greatly reduce the warpage defect and shorten the cooling cycle time. The computational design methods and additive manufacturing techniques that prompt the development of conformal cooling are deeply investigated. At the end of this survey, the future perspectives for conformal cooling design and manufacturing are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The focus of this survey is on the physical model, development, and optimization of conformal cooling systems which have curved cooling circuits following the shape of mold cavity.'}",
-Constraints on millisecond magnetars as the engines of prompt emission in gamma-ray bursts,P. Beniamini,"We examine millisecond magnetars as central engines of Gamma Ray Bursts' (GRB) prompt emission. Using the proto-magnetar wind model of Metzger et al. 2011, we estimate the temporal evolution of the magnetization and power injection at the base of the GRB jet and apply these to different prompt emission models to make predictions for the GRB energetics, spectra and lightcurves. We investigate both shock and magnetic reconnection models for the particle acceleration, as well as the effects of energy dissipation across optically thick and thin regions of the jet. The magnetization at the base of the jet, $\sigma_0$, is the main parameter driving the GRB evolution in the magnetar model and the emission is typically released for $100\lesssim \sigma_0 \lesssim 3000$. Given the rapid increase in $\sigma_0$ as the proto-magnetar cools and its neutrino-driven mass loss subsides, the GRB duration is typically limited to $\lesssim 100$ s. This low baryon loading at late times challenges magnetar models for ultra-long GRBs, though black hole models likely run into similar difficulties without substantial entrainment from the jet walls. The maximum radiated gamma-ray energy is $\lesssim 5 \times 10^{51}$erg, significantly less than the magnetar's total initial rotational energy and in strong tension with the high end of the observed GRB energy distribution. However, the gradual magnetic dissipation model (Beniamini & Giannios 2017) applied to a magnetar central engine, naturally explains several key observables of typical GRBs, including energetics, durations, stable peak energies, spectral slopes and a hard to soft evolution during the burst.",,https://academic.oup.com/mnras/article-pdf/472/3/3058/20131453/stx2095.pdf
-Prompt injections of highly relativistic electrons induced by interplanetary shocks: A statistical study of Van Allen Probes observations,Q. Schiller,"We conduct a statistical study on the sudden response of outer radiation belt electrons due to interplanetary (IP) shocks during the Van Allen Probes era, i.e., 2012 to 2015. Data from the Relativistic Electron‐Proton Telescope instrument on board Van Allen Probes are used to investigate the highly relativistic electron response (E > 1.8 MeV) within the first few minutes after shock impact. We investigate the relationship of IP shock parameters, such as Mach number, with the highly relativistic electron response, including spectral properties and radial location of the shock‐induced injection. We find that the driving solar wind structure of the shock does not affect occurrence for enhancement events, 25% of IP shocks are associated with prompt energization, and 14% are associated with MeV electron depletion. Parameters that represent IP shock strength are found to correlate best with highest levels of energization, suggesting that shock strength may play a key role in the severity of the enhancements. However, not every shock results in an enhancement, indicating that magnetospheric preconditioning may be required.",,
-On the Synchrotron Spectrum of GRB Prompt Emission,Siyao Xu,"The prompt emission spectrum of gamma-ray bursts is characterized by a smoothly joint broken power-law spectrum known as the Band function. The typical low-energy photon index is , which poses a challenge to standard synchrotron radiation models. We investigate the electron energy spectrum as a result of the interplay among adiabatic stochastic acceleration (ASA), particle injection, and synchrotron cooling. In the ASA-dominated low-energy range, ASA enables an efficient hardening of the injected energy spectrum to approach a spectral index −1. In the synchrotron cooling-dominated high-energy range, the injected high-energy electrons undergo fast synchrotron cooling and have a softer photon spectrum. With the energy range of the injected electrons broadly covering both the ASA- and synchrotron cooling-dominated ranges, the resulting photon number spectrum has low- and high-energy indices of and , respectively. The break energy is of the order of ∼100 keV, depending on the turbulence properties.",,https://iopscience.iop.org/article/10.3847/1538-4357/aaa0ca/pdf
-Regional Implantation of Autologous Adipose Tissue-Derived Cells Induces a Prompt Healing of Long-Lasting Indolent Digital Ulcers in Patients with Systemic Sclerosis,N. Del Papa,"Digital ulcers (DUs) are a rather frequent and invalidating complication in systemic sclerosis (SSc), often showing a very slow or null tendency to heal, in spite of the commonly used systemic and local therapeutic procedures. Recently, stem cell therapy has emerged as a new approach to accelerate wound healing. In the present study, we have tentatively treated long-lasting and poorly responsive to traditional therapy SSc-related DUs by implantation of autologous adipose tissue-derived cell (ATDC) fractions. Fifteen patients with SSc having a long-lasting DU in only one fingertip who were unresponsive to intensive systemic and local treatment were enrolled in the study. The grafting procedure consisted of the injection, at the basis of the corresponding finger, of 0.5-1 ml of autologous ATDC fractions, separated by centrifugation of adipose tissue collected through liposuction from subcutaneous abdominal fat. Time to heal after the procedure was the primary end point of the study, while reduction of pain intensity and of analgesic consumption represented a secondary end point. Furthermore, the posttherapy variation of the number of capillaries, observed in the nailfold video capillaroscopy (NVC) exam and of the resistivity in the digit arteries, measured by high-resolution echocolor-Doppler, were also taken into account. A rather fast healing of the DUs was reached in all of the enrolled patients (mean time to healing 4.23 weeks; range 2-7 weeks). A significant reduction of pain intensity was observed after a few weeks (p < 0.001), while the number of capillaries was significantly increased at 3- and 6-month NVC assessment (p < 0.0001 in both cases). Finally, a significant after-treatment reduction of digit artery resistivity was also recorded (p < 0.0001). Even with the limitations related to the small number of patients included and to the open-label design of the study, the observed strongly favorable outcome suggests that local grafting with ATDCs could represent a promising option for the treatment of SSc-related DUs unresponsive to more consolidated therapies.","{'model': 'tldr@v2.0.0', 'text': 'The observed strongly favorable outcome suggests that local grafting with ATDCs could represent a promising option for the treatment of SSc-related DUs unresponsive to more consolidated therapies.'}",https://journals.sagepub.com/doi/pdf/10.3727/096368914X685636
-Estimation Method of Prompt Neutron Decay Constant Reducing Higher Order Mode Effect by Linear Combination,R. Katano,"Abstract The linear combination method is proposed to reduce the higher order mode (HOM) effect on the measurement of the prompt neutron decay constant using the α-fitting method. Conventional α-fitting utilizes the pulsed neutron source and estimates the prompt neutron decay constant by fitting the neutron counts at a single detector after pulse injection with a single exponential function. The proposed method reduces the spatial HOM effect with linear combination of the neutron counts at multiple detectors. For verification, we applied the conventional method and the proposed method to the analytical solution of the diffusion theory and the Monte Carlo simulation to estimate the prompt neutron decay constant of a one-dimensional infinite slab. Comparison of these results indicates that the proposed method enables estimation with the reduced HOM effect as opposed to the conventional method. Through the verification, we confirmed that the proposed method can be a candidate for a measurement method of the prompt neutron decay constant.",,
-Prompt elevation of rat serum α‐fetoprotein by acute liver injury following a single injection of ethionine,A. Watanabe,"The mechanism of increased a‐fetoprotein (AFP) production following a single injection of ethionine was investigated by using rats aged 5 weeks at the time of killing. Marked elevations of serum AFP concentrations occurred within 4 days in both male and female rats after administration of DL‐ethionine or L‐ethionine, although the increased levels of serum AFP and liver triglyceride in the adults were less marked in the male than in the female. No apparent necrosis of liver cells was observed in ethionine‐treated rats. Frequent administrations of adenosine triphosphate after a single dose of ethionine prevented the increases in liver triglyceride and serum AFP levels. The increased concentrations of serum AFP, reaching a maximum level within 4 days, occurred before a slight increase in incorporation of 3H‐thymidine into liver DNA. The serum AFP from ethionine‐treated rats was immunologically and electrophoretically indistinguishable from that of fetal, carbontetrachloride‐treated or hepatoma‐bearing rats. These observations suggest that the increased production of AFP in ethionine‐treated rats is closely associated with hepatic injury and is not the consequence of liver cell regeneration.","{'model': 'tldr@v2.0.0', 'text': 'The observations suggest that the increased production of AFP in ethionine‐treated rats is closely associated with hepatic injury and is not the consequence of liver cell regeneration.'}",
-Split Injection Strategy for Prompt Cold Starting and Low White Smoke Emissions,M. Girotra,"Progressing needs for prompt cold start of direct injection Diesel engines is the motivation behind this study. Authors have examined the autoignition and combustion processes in the early firing cycles of the engine and proposed a strategy to reduce the cranking period and the white smoke emissions. The concept is to accelerate the preparation of the combustible mixture during the cranking process. This is achieved by splitting the injected fuel in two parts and controlling its timing. The duration of the first injection is limited such that the spray penetrates through the combustion chamber and evaporates before it reaches the walls. The dwell between the two injections is adjusted to allow time for the first spray to mix with the fresh charge, form a combustible mixture and start producing the autoignition radicals. The second part would evaporate and autoignite by reacting with the radicals before it reaches the cool walls. The strategy is verified on a 1.2 L Ford Diata Diesel engine equipped with a first generation common rail fuel injection system. The cycle resolved hydrocarbons, and NOx emissions are measured by high response detectors. In addition, the mass and constituents of the white smoke are measured. All the experiments for this paper have been conducted after the engine has been soaked at the normal room temperature for at least eight hours. The results showed that there is an optimum strategy for the split injection that would minimize the cranking period and white smoke emissions.Copyright © 2005 by ASME",,
-Hyperbaric Oxygen for Ischemia due to Injection of Cosmetic Fillers: Case Report and Issues,R. Henderson,"Summary: Natural and synthetic fillers have revolutionized aesthetic facial rejuvenation and soft-tissue augmentation. We present a case highlighting the dangers of filler self-injection. A 37-year-old woman self-injected a dermal filler around both temples. She immediately experienced left--side hearing loss, blanching over the left face, and pain. Prompt treatment with hyaluronidase, topical nitro paste, and warm compresses ensued. An emergency computed tomography angiogram showed occlusion of a superficial temporal artery branch. We treated her with enoxaparin, aspirin, dexamethasone, piperacillin-tazobactam, and intradermal lidocaine. After 6 hyperbaric oxygen therapy (HBO2) treatments in 3 days, the patient showed improvement in appearance with markedly decreased ischemic discoloration and her hearing returned to baseline. Algorithms for treating such injuries generally neglect HBO2. HBO2 is thought to be efficacious in these situations by a variety of mechanisms: oxygenation of ischemic tissues, reduction of edema, amelioration of ischemic/reperfusion injury, promotion of angiogenesis and collagen maturation. Her resolved hearing highlights the utility of HBO2 in sudden hearing loss as well. Injectors should have guidelines for using product, not only on patients but staff as well. Filler courses should include handling complications and include HBO2 in their guidelines. Clinicians should remind patients to seek treatment from qualified clinicians. The goal of a bargain price using self-injection may quickly become expensive and disfiguring.","{'model': 'tldr@v2.0.0', 'text': 'A 37-year-old woman who self-injected a dermal filler around both temples experienced left--side hearing loss, blanching over the left face, and pain and received 6 hyperbaric oxygen therapy treatments, highlighting the utility of HBO2 in sudden hearing loss as well.'}",
-Prompt radial electric field response to neutral beam injection,T. Rhodes,"A variety of low density discharges in the DIII-D tokamak exhibit a prompt response (<10 ms) to neutral beam injection as evidenced by changes in the Doppler shift of the density fluctuations. This variation in the Doppler shift, attributed to a changing Vθ,Er × B velocity, which is in turn due to a rapidly changing radial electric field Er, is observed from the deep core to the edge of these plasmas. The core changes are large and occur much more quickly than collisional beam equilibration times (≥ 120 ms for the core) indicating a different momentum transfer mechanism. Towards the edge the Doppler shift times approach the collisional transfer times. Theoretical predictions of changes in the core Er due to radial fast particle currents are not large enough to explain the observations. Although not yet understood this effect could lead to a tool for relatively fast feedback control of core electric fields in advanced confinement regimes.",,
-Experience and Management of Intravascular Injection with Facial Fillers: Results of a Multinational Survey of Experienced Injectors,G. Goodman,,"{'model': 'tldr@v2.0.0', 'text': None}",
-Detection of prompt luminescence from an imaging plate with heavy particle injection,S. Gunji,"We have been investigating the basic characteristics of an imaging plate JP) to develop a new type of ultra-heavy-particle detector. To acquire the information on the deposited energy, photostimulated luminescence (PSL) from the injection position is usually utilized, which occurs upon scanning the surface of the IP with a laser beam after the exposure of particles. However, in the use of PSL, there is the disadvantage that it fades with time. On the other hand, when heavy particles are injected to the IP, prompt luminescence (PL) also occurs from the injection position. We have succeeded in detecting the prompt luminescence from each heavy particle with the energy of several hundred MeV/n. Since the PL is useful for quick detection of the injection position, the PSL data can be read within a few seconds by scanning the determined position before any serious fading effect occurs.",,
-Prompt enhancement of the Earth's outer radiation belt due to substorm electron injections,C. Tang,"We present multipoint simultaneous observations of the near‐Earth magnetotail and outer radiation belt during the substorm electron injection event on 16 August 2013. Time History of Events and Macroscale Interactions during Substorms A in the near‐Earth magnetotail observed flux‐enhanced electrons of 300 keV during the magnetic field dipolarization. Geosynchronous orbit satellites also observed the intensive electron injections. Located in the outer radiation belt, RBSP‐A observed enhancements of MeV electrons accompanied by substorm dipolarization. The phase space density (PSD) of MeV electrons at L*~5.4 increased by 1 order of magnitude in 1 h, resulting in a local PSD peak of MeV electrons, which was caused by the direct effect of substorm injections. Enhanced MeV electrons in the heart of the outer radiation belt were also detected within 2 h, which may be associated with intensive substorm electron injections and subsequent local acceleration by chorus waves. Multipoint observations have shown that substorm electron injections not only can be the external source of MeV electrons at the outer edge of the outer radiation belt (L*~5.4) but also can provide the intensive seed populations in the outer radiation belt. These initial higher‐energy electrons from injection can reach relativistic energy much faster. The observations also provide evidence that enhanced substorm electron injections can explain rapid enhancements of MeV electrons in the outer radiation belt.",,https://agupubs.onlinelibrary.wiley.com/doi/pdfdirect/10.1002/2016JA023550
-Hot electron injection from graphene quantum dots to TiO₂.,K. J. Williams,"The Shockley-Queisser limit is the maximum power conversion efficiency of a conventional solar cell based on a single semiconductor junction. One approach to exceed this limit is to harvest hot electrons/holes that have achieved quasi-equilibrium in the light absorbing material with electronic temperatures higher than the phonon temperature. We argue that graphene based materials are viable candidates for hot carrier chromophores. Here we probe hot electron injection and charge recombination dynamics for graphene quantum dots (QDs, each containing 48 fused benzene rings) anchored to the TiO₂(110) surface via carboxyl linkers. We find ultrafast electron injection from photoexcited graphene QDs to the TiO₂ conduction band with time constant τ(i) < 15 fs and charge recombination dynamics characterized by a fast channel (τ(r1) = 80-130 fs) and a slow one (τ(r2) = 0.5-2 ps). The fast decay channel is attributed to the prompt recombination of the bound electron-hole pair across the interface. The slow channel depends strongly on excitation photon energy or sample temperature and can be explained by a ""boomerang"" mechanism, in which hot electrons are injected into bulk TiO₂, cooled down due to electron-phonon scattering, drifted back to the interface under the transient electric field, and recombine with the hole on graphene QDs. We discuss feasibilities of implementing the hot carrier solar cell using graphene nanomaterials.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that graphene based materials are viable candidates for hot carrier chromophores and feasibilities of implementing the hot carrier solar cell using graphene nanomaterials are discussed.'}",
-Prompt loss of beam ions in KSTAR plasmas,Jun Young Kim,"For a toroidal plasma facility to realize fusion energy, researching the transport of fast ions is important not only due to its close relation to the heating and current drive efficiencies but also to determine the heat load on the plasma-facing components. We present a theoretical analysis and orbit simulation for the origin of lost fast-ions during neutral beam injection (NBI) heating in Korea Superconducting Tokamak Advanced Research (KSTAR) device. We adopted a two-dimensional phase diagram of the toroidal momentum and magnetic moment and describe detectable momentums at the fast-ion loss detector (FILD) position as a quadratic line. This simple method was used to model birth ions deposited by NBI and drawn as points in the momentum phase space. A Lorentz orbit code was used to calculate the fast-ion orbits and present the prompt loss characteristics of the KSTAR NBI. The scrape-off layer deposition of fast ions produces a significant prompt loss, and the model and experimental results closely agreed on the pitch-angle range of the NBI prompt loss. Our approach can provide wall load information from the fast ion loss.",,https://aip.scitation.org/doi/pdf/10.1063/1.4966588
-COMBINED INTRAVITREAL BEVACIZUMAB INJECTION AND ZONE I SPARING LASER PHOTOCOAGULATION IN PATIENTS WITH ZONE I RETINOPATHY OF PREMATURITY,Jaeryung Kim,"Purpose: To investigate the anatomical outcome of combined intravitreal bevacizumab injection and Zone I sparing laser ablation in patients with Type 1 retinopathy of prematurity in Zone I. Methods: The medical records of consecutive 18 eyes of 10 infants, who underwent combined intravitreal bevacizumab (0.25 mg) injection and Zone I sparing laser ablation for the treatment of Type 1 retinopathy of prematurity in Zone I, were retrospectively reviewed. Laser photocoagulation was performed on the avascular retina anterior to the margin of Zone I extending to the ora serrata. Anatomical outcomes including progression to stage 4/5, macular changes, and vitreous organization were reviewed. Results: The mean gestational age at birth and the birth weight of included patients were 24.0 weeks and 628 g, respectively. The timing of bevacizumab injection ranged from postmenstrual age 33+2 to 35 weeks (mean, 34.3 weeks). Postmenstrual age at last follow-up ranged from 74+6 to 107+1 weeks (mean, 83.6 weeks). All 18 eyes demonstrated prompt regression of neovascular pathology and plus disease without recurrence. Previously avascular Zone I retina was vascularized in all eyes after the treatment. All eyes showed excellent anatomical outcome with intact macula, but one eye showed mild vitreous organization above the vascular/avascular junction. Conclusion: Combined intravitreal bevacizumab injection and Zone I sparing laser ablation for Type 1 retinopathy of prematurity in Zone I seem to be effective treatment options. Possible advantages include lower dose of anti–vascular endothelial growth factor, less recurrence than monotherapy, and preservation of central visual field.","{'model': 'tldr@v2.0.0', 'text': 'All eyes showed excellent anatomical outcome with intact macula, but one eye showed mild vitreous organization above the vascular/avascular junction, and possible advantages include lower dose of anti–vascular endothelial growth factor, less recurrence than monotherapy, and preservation of central visual field.'}",
-Intralesional antibiotic injection using 41G needle for the management of subretinal abscess in endogenous endophthalmitis,P. Venkatesh,,"{'model': 'tldr@v2.0.0', 'text': 'Intralesional pharmacotherapy may be safe and effective in the treatment of subretinal abscess associated with endogenous endophthalmitis.'}",https://journalretinavitreous.biomedcentral.com/track/pdf/10.1186/s40942-016-0043-x
-Sudden visual loss and multiple cerebral infarction after autologous fat injection into the glabella.,Dongkyun Hong,"This case demonstrates that patients may develop nerve injury even when using the lowest energy level. Even in this case with nerve injury, the long-term benefits of improved hyperhidrosis may outweigh such adverse effects as weakness of the left hand that recovered after 12 months. Also, the patient improved, neurologically and functionally, with intensive physiotherapy after 6 months of persistent weakness and numbness. Based on this observation, we recommend prompt and intensive rehabilitation to accelerate recovery if neural damage occurs unexpectedly. Moreover, more caution is needed when treating thin male patients.","{'model': 'tldr@v2.0.0', 'text': 'This case demonstrates that patients may develop nerve injury even when using the lowest energy level, and recommends prompt and intensive rehabilitation to accelerate recovery if neural damage occurs unexpectedly.'}",
-The Roles of Traditional Chinese Medicine: Shen-Fu Injection on the Postresuscitation Care Bundle,Qian Zhang,"Survival rates following in-hospital and out-of-hospital cardiac arrests remain disappointingly low. Organ injury caused by ischemia and hypoxia during prolonged cardiac arrest is compounded by reperfusion injury that occurs when a spontaneous circulation is restored. A bundle of procedures, which may need to be administered simultaneously, is required. The procedures include prompt identification and treatment of the cause of cardiac arrest, as well as a definitive airway and ventilation together. Additional benefit is possible with appropriate forms of early goal-directed therapy and achieving therapeutic hypothermia within the first few hours, followed by gradual rewarming and ensuring glycaemic control to be within a range of 6 to 10 mmol/L. All these would be important and need to be continued for at least 24 hours. Previous studies have showed that the effects of Shen-Fu injection (SFI) are based on aconitine properties, supplemented by ginsenoside, which can scavenge free radicals, improve energy metabolism, inhibit inflammatory mediators, suppress cell apoptosis, and alleviate mitochondrial damage. SFI, like many other complex prescriptions of traditional Chinese medicine, was also found to be more effective than any of its ingredient used separately in vivo. As the postresuscitation care bundle is known to be, the present paper focuses on the role of SFI played on the postresuscitation care bundle.","{'model': 'tldr@v2.0.0', 'text': 'The role of SFI played on the postresuscitation care bundle is focused on and Shen-Fu injection was found to be more effective than any of its ingredient used separately in vivo.'}",https://downloads.hindawi.com/journals/ecam/2013/319092.pdf
-Chat-REC: Towards Interactive and Explainable LLMs-Augmented Recommender System,Yunfan Gao,"Large language models (LLMs) have demonstrated their significant potential to be applied for addressing various application tasks. However, traditional recommender systems continue to face great challenges such as poor interactivity and explainability, which actually also hinder their broad deployment in real-world systems. To address these limitations, this paper proposes a novel paradigm called Chat-Rec (ChatGPT Augmented Recommender System) that innovatively augments LLMs for building conversational recommender systems by converting user profiles and historical interactions into prompts. Chat-Rec is demonstrated to be effective in learning user preferences and establishing connections between users and products through in-context learning, which also makes the recommendation process more interactive and explainable. What's more, within the Chat-Rec framework, user's preferences can transfer to different products for cross-domain recommendations, and prompt-based injection of information into LLMs can also handle the cold-start scenarios with new items. In our experiments, Chat-Rec effectively improve the results of top-k recommendations and performs better in zero-shot rating prediction task. Chat-Rec offers a novel approach to improving recommender systems and presents new practical scenarios for the implementation of AIGC (AI generated content) in recommender system studies.","{'model': 'tldr@v2.0.0', 'text': 'A novel paradigm called Chat-Rec (ChatGPT Augmented Recommender System) that innovatively augments LLMs for building conversational recommender systems by converting user profiles and historical interactions into prompts is proposed.'}",http://arxiv.org/pdf/2303.14524
-An overview of anthrax infection including the recently identified form of disease in injection drug users,C. Hicks,,"{'model': 'tldr@v2.0.0', 'text': 'While antibiotics are the mainstay of anthrax treatment, use of adjunctive therapies such as anthrax toxin antagonists are a consideration and Prompt surgical therapy appears to be important for successful management of injectional anthrax.'}",https://europepmc.org/articles/pmc3523299?pdf=render
-From ChatGPT to ThreatGPT: Impact of Generative AI in Cybersecurity and Privacy,Maanak Gupta,"Undoubtedly, the evolution of Generative AI (GenAI) models has been the highlight of digital transformation in the year 2022. As the different GenAI models like ChatGPT and Google Bard continue to foster their complexity and capability, it’s critical to understand its consequences from a cybersecurity perspective. Several instances recently have demonstrated the use of GenAI tools in both the defensive and offensive side of cybersecurity, and focusing on the social, ethical and privacy implications this technology possesses. This research paper highlights the limitations, challenges, potential risks, and opportunities of GenAI in the domain of cybersecurity and privacy. The work presents the vulnerabilities of ChatGPT, which can be exploited by malicious users to exfiltrate malicious information bypassing the ethical constraints on the model. This paper demonstrates successful example attacks like Jailbreaks, reverse psychology, and prompt injection attacks on the ChatGPT. The paper also investigates how cyber offenders can use the GenAI tools in developing cyber attacks, and explore the scenarios where ChatGPT can be used by adversaries to create social engineering attacks, phishing attacks, automated hacking, attack payload generation, malware creation, and polymorphic malware. This paper then examines defense techniques and uses GenAI tools to improve security measures, including cyber defense automation, reporting, threat intelligence, secure code generation and detection, attack identification, developing ethical guidelines, incidence response plans, and malware detection. We will also discuss the social, legal, and ethical implications of ChatGPT. In conclusion, the paper highlights open challenges and future directions to make this GenAI secure, safe, trustworthy, and ethical as the community understands its cybersecurity impacts.","{'model': 'tldr@v2.0.0', 'text': 'The work presents the vulnerabilities of ChatGPT, which can be exploited by malicious users to exfiltrate malicious information bypassing the ethical constraints on the model, and uses GenAI tools to improve security measures, including cyber defense automation, reporting, threat intelligence, secure code generation and detection, attack identification, developing ethical guidelines, incidence response plans, and malware detection.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/10198233.pdf
-The influence of boundary layers on supersonic inlet flow unstart induced by mass injection,H. Do,,,
-International consensus on (ICON) anaphylaxis,L. Ardusso,,"{'model': 'tldr@v2.0.0', 'text': 'These guidelines concur about prompt initial treatment with intramuscular injection of epinephrine (adrenaline) in the mid-outer thigh, positioning the patient supine, calling for help, and when indicated, providing supplemental oxygen, intravenous fluid resuscitation and cardiopulmonary resuscitation.'}",http://www.worldallergyorganizationjournal.org/article/S1939455119302406/pdf
-Numerical Investigation of the Plasma-Assisted MILD Combustion of a CH4/H2 Fuel Blend Under Various Working Conditions,S. M. Mousavi," The effects of plasma injection upon MILD combustion of a mixture of methane and hydrogen are investigated numerically. The injected plasma includes the flow of a highly air-diluted methane including C2H2, C2H4, C2H6, CH, CH2, CH3, CO, and CO2. The results show that among all the constitutes of plasma, CH3 is the most effective in improving the characteristics of MILD combustion. Injection of this radical leads to the occurrence of reactions at a closer distance to the burner inlet and thus provides longer time for completion of combustion. Further, mass fractions of OH, CH2O, and HCO are considerably affected by the injections of CH3, indicating structural modifications of the reacting flow. Importantly, as Reynolds number of the plasma flow increases, the volume and width of the flame decrease, while the formations of prompt and thermal NOx are intensified. However, injection of CH3, as plasma, reduces the emission of thermal NOx.",,https://asmedigitalcollection.asme.org/energyresources/article-pdf/143/6/062302/6686005/jert_143_6_062302.pdf
-Modeling the response of a fast ion loss detector using orbit tracing techniques in a neutral beam prompt-loss study on the DIII-D tokamak.,D. Pace,A numerical model describing the expected measurements of neutral beam prompt-losses by a newly commissioned fast ion loss detector (FILD) in DIII-D is presented. This model incorporates the well understood neutral beam deposition profiles from all eight DIII-D beamlines to construct a prompt-loss source distribution. The full range of detectable ion orbit phase space available to the FILD is used to calculate ion trajectories that overlap with neutral beam injection footprints. Weight functions are applied to account for the level of overlap between these detectable orbits and the spatial and velocity (pitch) properties of ionized beam neutrals. An experimental comparison is performed by firing each neutral beam individually in the presence of a ramping plasma current. Fast ion losses determined from the model are in agreement with measured losses.,"{'model': 'tldr@v2.0.0', 'text': 'A numerical model describing the expected measurements of neutral beam prompt-losses by a newly commissioned fast ion loss detector (FILD) in DIII-D is presented.'}",https://escholarship.org/content/qt9j11k20t/qt9j11k20t.pdf?t=p15iy8
-"Complications of injectable fillers, part 2: vascular complications.",C. Delorenzi,"Accidental intra-arterial filler injection may cause significant tissue injury and necrosis. Hyaluronic acid (HA) fillers, currently the most popular, are the focus of this article, which highlights complications and their symptoms, risk factors, and possible treatment strategies. Although ischemic events do happen and are therefore important to discuss, they seem to be exceptionally rare and represent a small percentage of complications in individual clinical practices. However, the true incidence of this complication is unknown because of underreporting by clinicians. Typical clinical findings include skin blanching, livedo reticularis, slow capillary refill, and dusky blue-red discoloration, followed a few days later by blister formation and finally tissue slough. Mainstays of treatment (apart from avoidance by meticulous technique) are prompt recognition, immediate treatment with hyaluronidase, topical nitropaste under occlusion, oral acetylsalicylic acid (aspirin), warm compresses, and vigorous massage. Secondary lines of treatment may involve intra-arterial hyaluronidase, hyperbaric oxygen therapy, and ancillary vasodilating agents such as prostaglandin E1. Emergency preparedness (a ""filler crash cart"") is emphasized, since early intervention is likely to significantly reduce morbidity. A clinical summary chart is provided, organized by complication presentation.","{'model': 'tldr@v2.0.0', 'text': 'Hyaluronic acid (HA) fillers, currently the most popular, are the focus of this article, which highlights complications and their symptoms, risk factors, and possible treatment strategies, and emergency preparedness is emphasized.'}",https://academic.oup.com/asj/article-pdf/34/4/584/23787559/34-4-584.pdf
-"CD4pos, NK1.1pos T cells promptly produce interleukin 4 in response to in vivo challenge with anti-CD3",T. Yoshimoto,"Injection of anti-CD3 antibodies causes prompt expression of interleukin (IL)-4, IL-2, and interferon gamma (IFN-gamma) mRNA among spleen cells. The optimal dose of anti-CD3 for such induction was 1.33 microgram/animal; lymphokine mRNA was first observed at 30 min, peaked at 90 min, and was undetectable (for IL-4) or had declined markedly by 4 h. Cells harvested from spleens of mice injected with anti-CD3 90 min earlier secreted IL-4, IL-2, and IFN-gamma without further stimulation. By contrast, in vitro stimulation with anti-CD3 of spleen cell suspensions or splenic fragments from noninjected donors failed to cause prompt production of IL-4 and, even after 24 h of stimulation, the amount of IL-4 produced in such cells was substantially less than that secreted within 1 h by spleen cell suspensions or splenic fragments from mice injected with anti-CD3 90 min earlier. Production of IL-4 by spleen cells from anti-CD3-injected mice was not inhibited by pretreatment with anti-IL-4 antibody or with IFN-gamma or tumor growth factor beta nor enhanced by treatment with IL-4. By contrast, CTLA-4 immunoglobulin (Ig) treatment clearly diminished IL-4 production in response to in vivo anti-CD3, indicating that cellular interactions involving CD28 (or related molecules) were important in stimulation. Cell sorting analysis indicated that the cells that produced IL-4 in response to in vivo injection of anti-CD3 were highly enriched in CD4pos cells with the phenotype leukocyte cell adhesion molecule-1 (LECAM-1)dull, CD44bright, CD45RBdull, NK1.1pos. Indeed, the small population of CD4pos, NK1.1pos cells had the great majority of the IL-4- producing activity of this population. Injection with Staphylococcal enterotoxin B also caused prompt induction of IL-4 mRNA; the cells that were principally responsible for production also had the phenotype of CD4pos, NK1.1pos. These results suggest that possibility that this rare population of T cells may be capable of secreting IL-4 at the outset of immune responses and thus may act to regulate the pattern of priming of naive T cells, by providing a source of IL-4 to favor the development of T cell helper 2-like IL-4-producing cells.","{'model': 'tldr@v2.0.0', 'text': 'Cell sorting analysis indicated that the cells that produced IL-4 in response to in vivo injection of anti-CD3 were highly enriched in CD4pos cells with the phenotype leukocyte cell adhesion molecule-1 (LECAM-1)dull, CD44bright, CD45RBdull and NK1.1pos, suggesting that possibility that this rare population of T cells may be capable of secreting IL- 4 at the outset of immune responses is suggested.'}",https://rupress.org/jem/article-pdf/179/4/1285/1104795/1285.pdf
-"Butyrate, a metabolite of intestinal bacteria, enhances sleep",É. Szentirmai,,"{'model': 'tldr@v2.0.0', 'text': 'The results suggest that the sleep-inducing effects of butyrate are mediated by a sensory mechanism located in the liver and/or in the portal vein wall, and Hepatoportalbutyrate-sensitive mechanisms may play a role in sleep modulation by the intestinal microbiota.'}",https://www.nature.com/articles/s41598-019-43502-1.pdf
-Critical contribution of liver natural killer T cells to a murine model of hepatitis.,Kazuyoshi Takeda,"Natural killer T (NKT) cells constitute a distinct subpopulation of T cells with a unique antigen specificity, prompt effector functions, and an unusual tissue distribution. NKT cells are especially abundant in the liver, but their physiological function in this organ remains unclear. In the present study, we examined the possible contribution of NKT cells to a murine model of hepatitis induced by i.v. injection of Con A. CD1-deficient mice lacking NKT cells were highly resistant to Con A-induced hepatitis. Adoptive transfer of hepatic NKT cells isolated from wild-type mice, but not from FasL-deficient gld mice, sensitized CD1-deficient mice to Con A-induced hepatitis. Furthermore, adoptive transfer of hepatic mononuclear cells from wild-type mice, but not from CD1-deficient mice, sensitized gld mice to Con A-induced hepatitis. Upon Con A administration, hepatic NKT cells rapidly up-regulated cell surface FasL expression and FasL-mediated cytotoxicity. At the same time, NKT cells underwent apoptosis leading to their rapid disappearance in the liver. These results implicated FasL expression on liver NKT cells in the pathogenesis of Con A-induced hepatitis, suggesting a similar pathogenic role in human liver diseases such as autoimmune hepatitis.","{'model': 'tldr@v2.0.0', 'text': 'Results implicated FasL expression on liver NKT cells in the pathogenesis of Con A-induced hepatitis, suggesting a similar pathogenic role in human liver diseases such as autoimmune hepatitis.'}",https://europepmc.org/articles/pmc25857?pdf=render
-Dynamic modeling of radiation belt electrons by radial diffusion simulation for a 2 month interval following the 24 March 1991 storm injection,F. Chu,"[1] Diffusive radial transport of radiation belt electrons with variable outer boundary is computed using Brautigam and Albert (2000) diffusion coefficients parameterized by Kp, modeling power level at ULF wave frequencies in the range of MeV electron drift periods. We analyzed radial diffusion during a relatively quiet 2 month interval following the 24 March 1991, prompt injection to form a new radiation belt at L* = 2.5. The radial diffusion calculation is initialized with a computed phase space density (PSD) profile using differential flux values from the CRRES HEEF instrument, covering 0.65–7.5 MeV. The outer boundary phase space density is updated using Los Alamos National Laboratory (LANL) GEO satellite fluxes, changing the ratio of PSD relative to a quiet day by assuming the outer boundary is changing proportional to the flux at a LANL GEO satellite. The location of the plasmapause Lp* is computed using a Kp-dependent formula separating different loss rates inside and outside the plasmapause. A series of simulations for different values of the first invariant is performed for this 2 month period. The flux is then interpolated to find electron flux at a fixed energy, 1 MeV, in order to compare with the CRRES satellite 1 MeV flux. Radial diffusion appears to be the dominant mechanism for this 2 month interval, which contains moderate storms (∣Dst∣ ≲ 100). Modulation of fluxes measured by CRRES compare well with simulations of the outer zone flux peak at L* = 3–4 for moderate high-speed stream-driven storms, along with persistence and slow decay of the new population of electrons injected on 24 March 1991, into L* = 2.5. The strongest storm of the 2 month interval (Dst = −105 nT) produced a flux dropout, which is not well-captured by the model, suggesting that improvements to the Kp-parameterized loss model are needed for larger storms.",,
-Epinephrine for First-aid Management of Anaphylaxis,S. Sicherer,"Anaphylaxis is a severe, generalized allergic or hypersensitivity reaction that is rapid in onset and may cause death. Epinephrine (adrenaline) can be life-saving when administered as rapidly as possible once anaphylaxis is recognized. This clinical report from the American Academy of Pediatrics is an update of the 2007 clinical report on this topic. It provides information to help clinicians identify patients at risk of anaphylaxis and new information about epinephrine and epinephrine autoinjectors (EAs). The report also highlights the importance of patient and family education about the recognition and management of anaphylaxis in the community. Key points emphasized include the following: (1) validated clinical criteria are available to facilitate prompt diagnosis of anaphylaxis; (2) prompt intramuscular epinephrine injection in the mid-outer thigh reduces hospitalizations, morbidity, and mortality; (3) prescribing EAs facilitates timely epinephrine injection in community settings for patients with a history of anaphylaxis and, if specific circumstances warrant, for some high-risk patients who have not previously experienced anaphylaxis; (4) prescribing epinephrine for infants and young children weighing <15 kg, especially those who weigh 7.5 kg and under, currently presents a dilemma, because the lowest dose available in EAs, 0.15 mg, is a high dose for many infants and some young children; (5) effective management of anaphylaxis in the community requires a comprehensive approach involving children, families, preschools, schools, camps, and sports organizations; and (6) prevention of anaphylaxis recurrences involves confirmation of the trigger, discussion of specific allergen avoidance, allergen immunotherapy (eg, with stinging insect venom, if relevant), and a written, personalized anaphylaxis emergency action plan; and (7) the management of anaphylaxis also involves education of children and supervising adults about anaphylaxis recognition and first-aid treatment.","{'model': 'tldr@v2.0.0', 'text': 'This clinical report from the American Academy of Pediatrics is an update of the 2007 clinical report on this topic and provides information to help clinicians identify patients at risk of anaphylaxis and new information aboutEpinephrine and epinephrine autoinjectors (EAs).'}",https://pediatrics.aappublications.org/content/pediatrics/139/3/e20164006.full.pdf
-Diversity of Kilonova Light Curves,K. Kawaguchi,"We perform radiative transfer simulations for kilonova in various situations, including the cases of prompt collapse to a black hole from neutron star mergers, high-velocity ejecta possibly accelerated by magnetars, and a black hole–neutron star merger. Our calculations are done employing ejecta profiles predicted by numerical-relativity simulations and a new line list for all the r-process elements. We found that: (i) the optical emission for binary neutron stars promptly collapsing to a black hole would be fainter by ≳1–2 mag than that found in GW170817, while the infrared emission could be as bright as that in GW170817 if the post-merger ejecta is as massive as ≈0.01 M⊙; (ii) the kilonova would be brighter than that observed in GW170817 for the case that the ejecta is highly accelerated by the electromagnetic energy injection from the remnant, but within a few days it would decline rapidly and the magnitude would become fainter than in GW170817; and (iii) the optical emission from a black hole–neutron star merger ejecta could be as bright as that observed in GW170817 for the case that sufficiently large amount of matter is ejected (≳0.02 M⊙), while the infrared brightness would be brighter by 1–2 mag at the same time. We show that the difference in the ejecta properties would be imprinted in the differences in the peak brightness and time of peak. This indicates that we may be able to infer the type of the central engine for kilonovae by observation of the peak in the multiple band.",,https://iopscience.iop.org/article/10.3847/1538-4357/ab61f6/pdf
-The time course of anaphylaxis manifestations in children is diverse and unpredictable,G. Pouessel,"The most current clinical definition of anaphylaxis is that proposed by Sampson et al. and different scoring systems have been proposed to assess its severity [1,2]. Anaphylaxis symptoms range from mild, self-limiting local reactions to life-threatening anaphylaxis. All anaphylaxis guidelines emphasize early recognition and the need for prompt intramuscular adrenaline injection as the first-line treatment [3]. However, adrenaline for anaphylaxis is underused by patients, families and even by health professionals. In the European anaphylaxis registry, 27% of patients treated by a health professional received adrenaline [4].","{'model': 'tldr@v2.0.0', 'text': 'The most current clinical definition of anaphylaxis is that proposed by Sampson et al. and different scoring systems have been proposed to assess its severity and underused by patients, families and even by health professionals.'}",
-High-pressure injection injury of the hand: an often underestimated trauma: case report with study of the literature,N. Verhoeven,,"{'model': 'tldr@v2.0.0', 'text': 'Not only the latency time to adequate treatment but also the injected fluid’s nature, the pressure, the volume and the location of injection, has influence on the seriousness and extensiveness of subcutaneous damage.'}",
-Beam-ion confinement for different injection geometries,W. Heidbrink,"The DIII-D tokamak is equipped with neutral beam sources that inject in four different directions; in addition, the plasma can be moved up or down to compare off-axis with on-axis injection. Fast-ion data for eight different conditions have been obtained: co/counter, near-tangential/near-perpendicular and on-axis/off-axis. Neutron measurements during short beam pulses assess prompt and delayed losses under low-power conditions. As expected, co-injection has fewer losses than counter, tangential fewer than perpendicular and on-axis fewer than off-axis; the differences are greater at low current than at higher current. The helicity of the magnetic field has a weak effect on the overall confinement. Fast-ion Dα (FIDA) and neutron measurements diagnose the confinement at higher power. The basic trends are the same as in low-power plasmas but, even in plasmas without long wavelength Alfvén modes or other MHD, discrepancies with theory are observed, especially in higher temperature plasmas. At modest temperature, two-dimensional images of the FIDA light are in good agreement with the simulations for both on-axis and off-axis injection. Discrepancies with theory are more pronounced at low fast-ion energy and at high plasma temperature, suggesting that fast-ion transport by microturbulence is responsible for the anomalies.",,
-Quantitation in PET using isotopes emitting prompt single gammas: application to yttrium-86,S. Walrand,,"{'model': 'tldr@v2.0.0', 'text': 'A patient-dependent correction method based on sinogram tail fitting using an 86Y point spread function library significantly improved the accuracy of the quantification and the mean discrepancy between PET total body activity and the activity expected from urinary collections was reduced.'}",
-THE EXPERIMENTAL DISEASE THAT FOLLOWS THE INJECTION OF HUMAN LEPROSY BACILLI INTO FOOT-PADS OF MICE,C. Shepard,"When leprosy bacilli from human patients are inoculated into the foot-pads of CFW mice, a microscopic granuloma containing acid-fast bacilli develops in a characteristic manner. This has been seen in 22 of 22 instances with leprosy bacilli from nasal washings, in 12 of 16 instances with leprosy bacilli from skin biopsies, and in none of 16 cases where the nasal washings were not observed to contain leprosy bacilli. Quantitative studies revealed a relationship between the number of bacilli inoculated and the time required for the appearance of the lesions. The incubation period was usually 1 to 2 months when the dose was 105.5 to 106.0 bacilli and about 6 months when the dose was about 103 organisms. After the development of the lesion, the number of bacilli harvested was usually in the range 104.5 to 106.0, regardless of the number inoculated. When the inoculum has contained 102.0 to 103.5 acid-fast bacilli, and harvests were reasonably prompt, there were regular increases of 50- to 1000-fold. Passage to new groups of mice has been successful 11 of 12 times. Most of these were second passages. One strain has been maintained in 3 passages with a total increase in acid-fast bacilli of 4 x 104-fold. Another strain has been through 4 passages with a total increase of about 4 x 106-fold. Cultures on bacteriological media favorable for the growth of most known mycobacterial species have not shown growth of mycobacteria.","{'model': 'tldr@v2.0.0', 'text': 'Quantitative studies revealed a relationship between the number of bacilli inoculated and the time required for the appearance of the lesions, and Cultures on bacteriological media favorable for the growth of most known mycobacterial species have not shown growth of myCobacteria.'}",http://jem.rupress.org/content/jem/112/3/445.full.pdf
-Evidence for a structurally-driven insulator-to-metal transition in VO 2 : A view from the ultrafast timescale,A. Cavalleri,"We apply ultrafast spectroscopy to establish a time-domain hierarchy between structural and electronic effects in a strongly correlated electron system. We discuss the case of the model system ${\mathrm{VO}}_{2}$, a prototypical nonmagnetic compound that exhibits cell doubling, charge localization, and a metal-insulator transition below 340 K. We initiate the formation of the metallic phase by prompt hole photo-doping into the valence band of the low-$T$ insulator. The insulator-to-metal transition is, however, delayed with respect to hole injection, exhibiting a bottleneck time scale, associated with the phonon connecting the two crystallographic phases. This structural bottleneck is observed despite faster depletion of the $d$ bands and is indicative of important bandlike character for this controversial insulator.",,https://kops.uni-konstanz.de/bitstream/123456789/5099/1/Evidence_for_a_structurally_driven_insulator_to_metal_transition_in_VO2.pdf
-Tractional retinal detachment after intravitreal injection of bevacizumab in proliferative diabetic retinopathy.,M. Torres-Soriano,"PURPOSE To assess the short-term complications of a single dose of intravitreal bevacizumab in patients with proliferative diabetic retinopathy (PDR). METHODS Retrospective review of 343 patients with PDR who were treated with intravitreal injection of bevacizumab (2.5 mg/0.1 mL). RESULTS Five patients (1.45%) presented tractional retinal detachment 1 to 6 weeks (mean 3 weeks) after intravitreal injection. All cases underwent pars plana vitrectomy, removal of all epiretinal fibrovascular membranes, further endolaser panretinal photocoagulation, and silicone tamponade. CONCLUSION Tractional retinal detachment may occur in a short time post intravitreal injection of bevacizumab in patients with proliferative diabetic retinopathy with extensive areas of ischemia and fibrovascular proliferations, and may require prompt vitreoretinal surgery.","{'model': 'tldr@v2.0.0', 'text': 'Tractional retinal detachment may occur in a short time post intravitreal injection of bevacizumab in patients with proliferative diabetic retinopathy with extensive areas of ischemia and fibrovascular proliferations, and may require prompt vitreoretinal surgery.'}",
-Prompt and powerful effect of a practice guideline on caudal additives,C. Eich,"not be ruled out completely in this event. Intravenous anticholinergic agents are the treatment of choice for these dystonic reactions, however more than one dose may be necessary for complete resolution of dystonia. An additional dose of diphenhydramine may have been beneficial in this patient. Although usually self limited, complications after Botox injections may be very disabling. Patients with communication problems experiencing these complications pose a challenge for the care team, especially in the context of prescheduled pediatric ambulatory care. In order to avoid these complications, optimal dose, injection volume, longterm cumulative effect and injection techniques should be studied further. Claude Abdallah Raafat Hannallah Division of Anesthesiology, Children’s National Medical Center, 111 Michigan Ave, NW, Washington, DC, USA (email: cabdalla@cnmc.org)","{'model': 'tldr@v2.0.0', 'text': 'Intravenous anticholinergic agents are the treatment of choice for these dystonic reactions, however more than one dose may be necessary for complete resolution of dystonia.'}",
-Anterior segment ischemia and retinochoroidal vascular occlusion after intralesional steroid injection.,A. Yağcı,"A 19-year-old woman was admitted with sudden severe pain and loss of light perception that began immediately after eyelid injection of the depot form of corticosteroid. Ecchymosis of the lower eyelid, corneal edema, and a fixed dilated pupil were noted. Fundus examination could not be performed because of corneal edema. Embolic material packed in the episcleral vessels was detected. With these findings, the diagnosis was anterior and posterior segment ischemia. Despite administration of an intravenous hyperosmotic agent, in addition to topical and systemic pulse corticosteroid (1.0 g/day), vision was not recovered. The most serious complication of intralesional injection of corticosteroid is vascular occlusion. This catastrophic complication after intralesional steroid injection should prompt immediate ophthalmoscopic fundus examination to allow early recognition of ischemic signs and immediate intervention.","{'model': 'tldr@v2.0.0', 'text': 'A 19-year-old woman admitted with sudden severe pain and loss of light perception that began immediately after eyelid injection of the depot form of corticosteroid was diagnosed with anterior and posterior segment ischemia.'}",
-Inflammation‐inducible anti‐TNF gene expression mediated by intra‐articular injection of serotype 5 adeno‐associated virus reduces arthritis,M. Khoury,"The tumor necrosis factor (TNF)‐α plays a central role in rheumatoid arthritis (RA) and current biotherapies targeting TNF‐α have a major impact on RA treatment. The long‐term safety concerns associated with the repetitive TNF blockade prompt optimization of therapeutic anti‐TNF approaches. Since we recently demonstrated that intra‐articular gene transfer using a recombinant adeno‐associated virus serotype 5 (rAAV5) efficiently transduces arthritic joints, we evaluate its effect on collagen‐induced arthritis (CIA) when encoding TNF antagonists.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that intra‐articular gene transfer using a recombinant adeno‐associated virus serotype 5 (rAAV5) efficiently transduces arthritic joints, and its effect on collagen‐induced arthritis (CIA) when encoding TNF antagonists is evaluated.'}",
-Retinal branch artery occlusion following injection of hyaluronic acid (Restylane),S. Peter,"The authors report a retinal branch artery occlusion occurring after facial injection of a dermal filler. The superior temporal artery showed occlusion due to a clearly visible long and fragmented embolus suggestive of gel and clearly distinguishable from calcific or cholesterol emboli. The authors suppose that hyaluronic acid gel was embolized in the patient. The embolized material is supposed to enter the ocular circulation through retrograde arteriolar flow after intra‐arterial injection into one of the peripheral branches of the ophthalmic artery. If there is any evidence of a visual problem after facial injection of a dermal filler, prompt consultation of an ophthalmologist is recommended.","{'model': 'tldr@v2.0.0', 'text': 'The superior temporal artery showed occlusion due to a clearly visible long and fragmented embolus suggestive of gel and clearly distinguishable from calcific or cholesterol emboli and the authors suppose that hyaluronic acid gel was embolized in the patient.'}",
-Treatment of Hyaluronic Acid Filler-Induced Impending Necrosis With Hyaluronidase: Consensus Recommendations.,Joel L Cohen,"Injection-induced necrosis is a rare but dreaded consequence of soft tissue augmentation with filler agents. It usually occurs as a result of injection of filler directly into an artery, but can also result from compression or injury. We provide recommendations on the use of hyaluronidase when vascular compromise is suspected. Consensus recommendations were developed by thorough discussion and debate amongst the authors at a roundtable meeting on Wednesday June 18, 2014 in Las Vegas, NV as well as significant ongoing written and verbal communications amongst the authors in the months prior to journal submission. All authors are experienced tertiary care providers. A prompt diagnosis and immediate treatment with high doses of hyaluronidase (at least 200 U) are critically important. It is not felt necessary to do a skin test in cases of impending necrosis. Some experts recommend dilution with saline to increase dispersion or lidocaine to aid vasodilation. Additional hyaluronidase should be injected if improvement is not seen within 60 minutes. A warm compress also aids vasodilation, and massage has been shown to help. Some experts advocate the use of nitroglycerin paste, although this area is controversial. Introducing an oral aspirin regimen should help prevent further clot formation due to vascular compromise. In our experience, patients who are diagnosed promptly and treated within 24 hours will usually have the best outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Recommendations on the use of hyaluronidase when vascular compromise is suspected are provided and patients who are diagnosed promptly and treated within 24 hours will usually have the best outcomes.'}",
-Subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient.,Hugo H. Davila,"INTRODUCTION Ischemic priapism (IP) is a urologic condition, which necessitates prompt management. Intracavernosal injection of phenylephrine is a usual treatment modality utilized for the management of these patients. Aim. We present a case of subarachnoid hemorrhage following intracavernosal injection of phenylephrine for IP in a patient with sickle cell disease. METHODS We analyzed the degree of subarachnoid hemorrhage in our patient after intracavernosal injection of phenylephrine. The patient had an acute rise in blood pressure during corporal irrigation. This was followed by the onset of severe headache. Computed tomography (CT) scan confirmed the diagnosis of a subarachnoid hemorrhage. MAIN OUTCOME MEASURE Subarachnoid hemorrhage associated with intracavernosal injection of phenylephrine. Result. A 23-year-old African American male with a history of sickle cell disease presented with a painful penile erection. The patient was started on intravenous fluids, oxygen by nasal canula, and analgesic medication. After this, a blood gas was obtained from his left corpora cavernosa. This was followed by normal saline irrigation and injection of phenylephrine. The patient complained of a sudden, severe ""terrible headache"" immediately following the last injection, and noncontrast CT scan of the head was obtained and a subarachnoid hemorrhage was noted. The patient was admitted for observation and no significant changes were noted. CONCLUSIONS Intracavernosal injection of phenylephrine for the management of IP can be associated with several possible complications. We present our single case complicated with the formation of a subarachnoid hemorrhage. The patient was treated conservatively and had no long-term neurologic sequelae. Davila HH, Parker J, Webster JC, Lockhart JL, and Carrion RE. Subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient.","{'model': 'tldr@v2.0.0', 'text': 'A single case complicated with the formation of a subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient is presented.'}",
-Prompt toroidal momentum balance with collisionless neutral beam injected torque in DIII-D,J. Degrassie,"In electron cyclotron heated (ECH) H-mode discharges with neutral beam injection (NBI) pulses that are short compared with the fast ion scattering or slowing times, and it is observed that the plasma stores all the angular momentum delivered by the NBI torque impulse. The pulse length is also much shorter than the momentum confinement time of the plasma. Source computations with the Monte Carlo code TRANSP [R. J. Goldston, D. C. McCune, H. H. Towner et al., J. Comput. Phys. 43, 61 (1981)] show that during a pulse approximately 90% of this torque impulse is delivered via the collisionless fast radial current injection process, so that the plasma acquires the balancing toroidal acceleration through ion drift motion in the increasing ∂E∕∂t where E is the electric field normal to the flux surfaces. The measured radial profile of the toroidal momentum increase matches the source, i.e., the computed torque impulse profile. We measure the bulk ion toroidal acceleration in helium discharges, as well as that of th...",,
-Increase of serum growth hormone concentration following thyrotropin-releasing hormone injection in patients with acromegaly or gigantism.,Minoru Irie,"ABSTRACT Effect of TRH administration on serum TSH and GH was studied in 9 normal subjects and 11 patients with acromegaly or gigantism. TRH provoked a prompt increase of serum TSH in normal subjects, whereas no appreciable change of serum GH was observed. Serum TSH responses to TRH were lower in most of the patients, but 8 out of the patients showed a remarkable increase of serum GH within 30 min after TRH injection. These results appear to indicate that specificity of TRH is not absolute in the functioning pituitary tumor secreting GH.","{'model': 'tldr@v2.0.0', 'text': 'Results appear to indicate that specificity of TRH is not absolute in the functioning pituitary tumor secreting GH in patients with acromegaly or gigantism.'}",
-Global‐Scale ULF Waves Associated With SSC Accelerate Magnetospheric Ultrarelativistic Electrons,Y. Hao,"We study electron behavior in the outer radiation belts during the 16 July 2017 storm sudden commencement (SSC), in which prompt intensification of ultrarelativistic electron fluxes was observed at around L = 4.8 by Van Allen Probe B immediately after an interplanetary shock. The electron fluxes in multiple energy channels show clear oscillations in the Pc5 frequency range, although the oscillation characteristics are quite different in different energy channels. At energies above ∼1 MeV, the oscillation periods were very close to the electron drift period, which resembles an energy spectrogram evolution expected for an energetic particle injection event and its drift echoes. At lower energies, however, the oscillation periods hardly depended on the energy: They were very close to the ultralow frequency (ULF) wave period derived from electric field measurements (about 250 s according to wavelet analysis). These complex signatures are consistent with the picture of drift resonance between electrons and short‐lived ULF waves with low azimuthal wave numbers. Good agreement between the observations and numerical simulations confirms that shock‐induced global‐scale ULF waves can efficiently accelerate outer belt ultrarelativistic electrons up to 3.4 MeV over a time scale shorter than 1 hr.",,
-Septic facet joint arthritis after a corticosteroid facet injection.,T. Weingarten,Lumbar facet joint injections are commonly employed in the treatment of low back pain and are considered to be relatively safe with few known complications. We report the case of septic facet arthritis following a periarticular facet injection in a patient with recurrent urinary tract infections. The literature is reviewed to identify epidemiological and clinical features of patients in whom septic facet arthritis developed after lumbar facet injection. The diagnosis of iatrogenic septic facet arthritis is often delayed because neurologic and constitutional signs and symptoms develop slowly. Serologic nonspecific markers of infection and appropriate imaging studies may be more sensitive for the early diagnosis of septic facet arthritis. Recalcitrant or worsening back pain after facet injections should prompt an investigation to rule out infectious causes.,"{'model': 'tldr@v2.0.0', 'text': 'Serologic nonspecific markers of infection and appropriate imaging studies may be more sensitive for the early diagnosis of septic facet arthritis.'}",https://academic.oup.com/painmedicine/article-pdf/7/1/52/5262013/7-1-52.pdf
-Comparison of Intraarticular Injection of Depot Corticosteroid and Hyaluronic Acid for Treatment of Degenerative Trapeziometacarpal Joints,S. Stahl,"Background:Trapeziometacarpal joint arthritis is a disabling condition presenting with pain at the base of the thumb causing impairment of hand function. Nonoperative treatment at an early stage includes intraarticular steroid injection. Although this treatment may bring about prompt symptomatic relief, its efficacy is unpredictable. Objectives:There is previous evidence that injection of sodium hyaluronate is effective and safe in the treatment of knee arthritis. We proposed that intraarticular injection of sodium hyaluronate, for the symptomatic treatment of trapeziometacarpal joint arthritis, could provide symptomatic relief without the adverse effects of steroids. Methods:Fifty-two patients with trapeziometacarpal joint grade II arthritis were randomized prospectively either for methylprednisolone or hyaluronate intraarticular injections. Initial evaluation included an estimation of pain, grip, pinch strengths and the functional Purdue Pegboard Test (PPT). This evaluation was repeated after 1, 3, and 6 months and statistically compared with the initial evaluation. Results:In both groups, the intraarticular injection produced a relief of pain after 1 month. Grip strength improved significantly in the group treated by the steroid during the whole evaluation period. The patients treated by hyaluronate showed improvement in grip strength after 6 months and in the pinch and the PPT after 3 months. Conclusions:Steroids and hyaluronate injections were found effective in reducing pain. Hyaluronate was more effective in the improvement of some aspects of fine hand function.","{'model': 'tldr@v2.0.0', 'text': 'This work proposed that intraarticular injection of sodium hyaluronate, for the symptomatic treatment of trapeziometacarpal joint arthritis, could provide symptomatic relief without the adverse effects of steroids.'}",
-Prompt striations in ionospheric barium clouds due to a velocity space instability,D. Simons,The linear theory of electrostatic instabilities of the drift loss-cone type has been developed for a velocity distribution appropriate to the shaped-charge barium injection experiment Buaro. The dispersion relation for the instability in a Vlasov plasma has been solved numerically. The results show a low-frequency instability that we interpret as the source of prompt striations observed in Buaro and cyclotron instabilities that appear to correspond to observed ion cyclotron fluctuations. These instabilities have much faster growth rates than drift instabilities of fluid theories and therefore provide a better description of the observations.,,
-Gamma-Ray Burst Spectral Correlations: Photospheric and Injection Effects,F. Ryde,"We present a physical framework that can account for most of the observed spectral properties of prompt gamma-ray burst emission. This includes the variety of spectral shapes and shape evolutions, and spectral correlations between flux and spectral peaks within bursts, described by Borgonovo & Ryde, and among bursts described by Amati and Ghirlanda. In our proposed model the spectral peak is given by the photospheric emission from a relativistic outflow for which the horizon length is much smaller that the radial width. The observed duration of the thermal flash is given by the radial light-crossing time. This then gives the typical emission site at ~1011 cm with a Lorentz factor of ~300. This emission is accompanied by nonthermal emission from dissipation locations outside the photosphere. The relative strengths of these two components depend on injection effects at the central engine, leading to varying relative locations of the saturation and photospheric radii. The total emission can then reproduce the observed variety. The spectral correlations are found by assuming that the amount of energy dissipated depends nonlinearly on the averaged particle density. Besides the spectral correlations, this also gives a description of how the relative strength of the thermal component varies with temperature within a burst.",,https://iopscience.iop.org/article/10.1086/508410/pdf
-X-ray plateaus in gamma-ray bursts’ light curves from jets viewed slightly off-axis,P. Beniamini," Using multiple observational arguments, recent work has shown that cosmological gamma-ray bursts (GRBs) are typically viewed at angles within, or close to the cores of their relativistic jets. One of those arguments relied on the lack of tens-of-days-long periods of very shallow evolution that would be seen in the afterglow light curves of GRBs viewed at large angles. Motivated by these results, we consider that GRBs efficiently produce γ-rays only within a narrow region around the core. We show that, on these near-core lines of sight, structured jets naturally produce shallow phases in the X-ray afterglow of GRBs. These plateaus would be seen by a large fraction of observers and would last between 102–105 s. They naturally reproduce the observed distributions of time-scales and luminosities as well as the intercorrelations between plateau duration, plateau luminosity, and prompt γ-ray energy. An advantage of this interpretation is that it involves no late-time energy injection which would be both challenging from the point of view of the central engine and, as we show here, less natural given the observed correlations between plateau and prompt properties.",,https://hal.archives-ouvertes.fr/hal-02393303v1/file/1907.05899.pdf
-Compliance with AZT treatment regimen of HIV-seropositive injection drug users: a neglected issue.,R. Freeman,"Although treatment with zidovudine (AZT) is now recommended for asymptomatic and symptomatic HIV-infected persons with CD4+ cell counts of 0.20 to 0.50 x 10(9)/L and under, data gathered from a small convenience sample of current and former injection drug users with AIDS in the New York City metropolitan region suggest that noncompliance with HIV/AIDS-related therapeutic regimen may be common in this population. This paper enumerates the reasons for noncompliance offered by these informants, reviews the general literature on treatment compliance to identify additional potential reasons for non-adherence to AZT treatment regimen, and outlines some suggestions for future research into this important issue that may prompt changes in the antiviral delivery system.","{'model': 'tldr@v2.0.0', 'text': 'Although treatment with zidovudine is now recommended for asymptomatic and symptomatic HIV-infected persons with CD4+ cell counts of 0.20 to 0.50 x 10(9)/L and under, data gathered from a small convenience sample of current and former injection drug users with AIDS in the New York City metropolitan region suggest that noncompliance with HIV/AIDS-related therapeutic regimen may be common.'}",
-Thoracic intradural Aspergillus abscess formation following epidural steroid injection.,G. Saigal,"We report an extremely unusual iatrogenic infection of the spinal canal with Aspergillus fumigatus that resulted in intradural abscess formation following epidural steroid injection in an immunocompetent young individual. Although the imaging findings of the infection were relatively nonspecific, MR imaging not only allowed for a prompt diagnosis, but also helped in surgical localization to the intradural compartment. Complications from the use of these injections are briefly discussed.","{'model': 'tldr@v2.0.0', 'text': 'An extremely unusual iatrogenic infection of the spinal canal with Aspergillus fumigatus that resulted in intradural abscess formation following epidural steroid injection in an immunocompetent young individual is reported.'}",
-Observations Concerning the Injection of a Lithium Aerosol into the Edge of TFTR Discharges,D. Mansfield,A new method of actively modifying the plasma-wall interaction was tested on the Tokamak Fusion Test Reactor. A laser was used to introduce a directed lithium aerosol into the discharge scrape-off layer. The lithium introduced in this fashion ablated and migrated preferentially to the limiter contact points. This allowed the plasma-wall interaction to be influenced in situ and in real time by external means. Significant improvement in energy confinement and fusion neutron production rate as well as a reduction in the plasma Zeff have been documented in a neutral-beam-heated plasma. The introduction of a metallic aerosol into the plasma edge increased the internal inductance of the plasma column and also resulted in prompt heating of core electrons in Ohmic plasmas. Preliminary evidence also suggests that the introduction of an aerosol leads to both edge poloidal velocity shear and edge electric field shear.,,https://digital.library.unt.edu/ark:/67531/metadc716589/m2/1/high_res_d/769316.pdf
-Effect of injection site on circulation times during cardiac arrest.,C. Emerman,"Cardiopulmonary resuscitation requires effective, prompt drug administration. In order to analyze Advanced Cardiac Life Support (ACLS) recommendations for site of drug administration, we studied dye circulation times after central, femoral, and peripheral venous injection during both closed and open chest CPR using a canine arrest model. Measurements of circulation times were made after injection of indocyanine green dye at femoral, central, and peripheral venous sites. Circulation times during closed chest CPR were 62.7 +/- 19.6 sec after central injection, 86.6 +/- 23.5 sec after femoral injection, and 93.6 +/- 17.9 sec after peripheral injection (p less than .001). During closed chest CPR, peak dye concentration after central injection was significantly higher than that after peripheral injection (4.0 +/- 1.3 vs. 3.1 +/- 0.8 mg/L, p less than .01). Circulation times were significantly shorter during open chest CPR with times again shortest after central injection. This animal model suggests that prompt drug delivery during CPR is enhanced by central venous injection of medication. There appears to be no advantage in femoral over peripheral injection.","{'model': 'tldr@v2.0.0', 'text': 'This animal model suggests that prompt drug delivery during CPR is enhanced by central venous injection of medication, and there appears to be no advantage in femoral over peripheral injection.'}",
-Endoscopic Hemostasis by Local Injection of Epinephrine and Polidocanol in Bleeding Ulcer. A Prospective Randomized Trial,J. Balanzó,"The present prospective randomized trial has been carried out to assess the efficacy of the endoscopic injection of epinephrine plus polidocanol, in the prompt treatment of upper gastrointestinal bleeding due to peptic lesions. Over a period of one year, 72 patients were admitted to the study. Thirty-six of these underwent endoscopic treatment (group A), while the others received medical treatment (group B). Definitive hemostasis was achieved in 29 patients in group A, and in 21 in group B (p less than 0.05). Seven patients in group A and 12 in group B underwent emergency surgery for persistent or recurrent bleeding (p = NS). Three patients from group B received endoscopic treatment owing to relative surgical contraindications. It was successful in all of them. Patients from group A needed a significantly lower number of units of packed red cells (p less than 0.05). Our results show that it is justified to employ endoscopic injection as the first therapeutic step in upper GI bleeding due to peptic gastroduodenal lesions.","{'model': 'tldr@v2.0.0', 'text': 'The results show that it is justified to employ endoscopic injection as the first therapeutic step in upper GI bleeding due to peptic gastroduodenal lesions.'}",
-Observation of the reaction H/sup O/ + O/sup 8 +/. -->. H/sup +/(O/sup 7 +/) * during neutral-beam injection into ORMAK,R. Isler,"Several lines of O VIII have been investigated during the injection of hydrogen atoms at energies of 10, 15, and 30 keV into the Oak Ridge tokamak (ORMAK). The Lyman-..cap alpha.., -..beta.., and -..gamma.. lines and the Balmer-..beta.. line exhibit slowly increasing signals, but the intensity of the Balmer-..cap alpha.. line increases sharply as soon as injection begins. This prompt signal is interpreted to be the direct result of charge transfer from hydrogen to O/sup 7 +/ ions in excited states.",,
-Complications of an intra-arterial injection from an inferior alveolar nerve block.,B. Webber,"BACKGROUND Unintended intravascular injections from inferior alveolar nerve blocks result in frustrating complications, both systemically and locally. It is imperative that the dentist diagnose the complication and treat it appropriately. Sometimes, several reactions occur simultaneously. CASE DESCRIPTION The authors present a case report that illustrates some of the many complications resulting from inferior alveolar nerve block injections. In this case, complications developed from intra-arterial injection of local anesthetic. The facial skin, intraoral structures and eye were affected. Within 60 minutes of the injection, all structures returned to their normal state. Prompt diagnosis and reassurance usually calm the patient. CLINICAL IMPLICATIONS Even when clinicians use the utmost care, by aspirating before the injection and noting anatomical landmarks, intra-arterial injections can occur during inferior alveolar nerve blocks. Fortunately, permanent damage to nerves, facial and oral tissues, and eyes is rare. The practitioner should diagnose and treat the problems appropriately to avoid any irreversible complications.","{'model': 'tldr@v2.0.0', 'text': 'In this case, complications developed from intra-arterial injection of local anesthetic from inferior alveolar nerve block injections, resulting in permanent damage to nerves, facial and oral tissues, and eyes.'}",
-Physics of confinement improvement of plasmas with impurity injection in DIII-D,M. Murakami,"External impurity injection into L mode edge discharges in DIII-D has produced clear confinement improvement (a factor of 2 in energy confinement and neutron emission), reduction in all transport channels (particularly ion thermal diffusivity to the neoclassical level), and simultaneous reduction of long wavelength turbulence. Suppression of the long wavelength turbulence and transport reduction are attributed to synergistic effects of impurity induced enhancement of E × B shearing rate and reduction of toroidal drift wave turbulence growth rate. A prompt reduction of density fluctuations and local transport at the beginning of impurity injection appears to result from an increased gradient of toroidal rotation enhancing the E × B shearing. Transport simulations carried out using the National Transport Code Collaboration demonstration code with a gyro-Landau fluid model, GLF23, indicate that E × B shearing suppression is the dominant transport suppression mechanism.",,
-Hepatocyte growth factor promotes liver regeneration with prompt improvement of hyperbilirubinemia in hepatectomized cholestatic rats.,A. Yoshikawa,"BACKGROUND In hepatectomy for patients with liver cirrhosis or cholestasis, prolonged postoperative hyperbilirubinemia is a troublesome complication and, if uncontrolled, often leads to life-threatening hepatic failure. Hepatocyte growth factor (HGF), first identified as the most potent mitogen for primary hepatocytes, has been shown to have multiple biological properties on liver, including mitogenic, antifibrotic, and cytoprotective activities. This study investigated the beneficial effects of a perioperative HGF supply to jaundiced liver after hepatectomy in rats. MATERIALS AND METHODS As a model of jaundiced liver, we used an alpha-naphtylisocyocyanate (ANIT)-induced intrahepatic cholestasis model. Forty-eight hours after intraperitoneal injection of ANIT (75 mg/kg), when the total serum bilirubin level was moderately increased, a 70 % hepatectomy was performed. Human recombinant HGF (250 microgram/kg) (n = 15) or vehicle alone (n = 15) was intermittently administered to the rats 12 h before surgery and every 12 h after that until sacrifice. RESULTS Perioperative HGF treatment effectively accelerated hepatocellular DNA synthesis of cholestatic liver followed by increase in the regenerated liver weight. Moreover, HGF supply promptly improved hyperbilirubinemia within 24 h after surgery. Histological examination revealed that HGF administration attenuated periportal inflammation and formation of bile duct obstructions. Postoperative serum concentrations of tumor necrosis factor-alpha, a representative inflammatory cytokine, were not altered by HGF treatment. CONCLUSIONS Perioperative HGF supply not only promotes liver regeneration but also ameliorates hyperbilirubinemia in hepatectomized cholestatic rats. This mode of HGF treatment may be clinically useful for hepatectomy in patients with cholestasis.","{'model': 'tldr@v2.0.0', 'text': 'Perioperative HGF supply not only promotes liver regeneration but also ameliorates hyperbilirubinemia in hepatectomized cholestatic rats, suggesting this mode of HGF treatment may be clinically useful for liver regeneration in patients withCholestasis.'}",
-Nuclear equation of state from observations of short gamma-ray burst remnants,P. Lasky,"The favored progenitor model for short $\ensuremath{\gamma}$-ray bursts (SGRBs) is the merger of two neutron stars that triggers an explosion with a burst of collimated $\ensuremath{\gamma}$-rays. Following the initial prompt emission, some SGRBs exhibit a plateau phase in their X-ray light curves that indicates additional energy injection from a central engine, believed to be a rapidly rotating, highly magnetized neutron star. The collapse of this ``protomagnetar'' to a black hole is likely to be responsible for a steep decay in X-ray flux observed at the end of the plateau. In this paper, we show that these observations can be used to effectively constrain the equation of state of dense matter. In particular, we show that the known distribution of masses in binary neutron star systems, together with fits to the X-ray light curves, provides constraints that exclude the softest and stiffest plausible equations of state. We further illustrate how a future gravitational wave observation with Advanced LIGO/Virgo can place tight constraints on the equation of state, by adding into the picture a measurement of the chirp mass of the SGRB progenitor.",,http://minerva-access.unimelb.edu.au/bitstreams/fac0a5fc-ddd7-577e-8842-c204d11e5797/download
-"Hand injuries due to high-pressure injection devices for painting in shipyards: circumstances, management, and outcome in twelve patients.",M. Valentino,"BACKGROUND Injuries due to high-pressure injections are frequently underestimated occupational accidents, which may have disabling outcomes. METHODS The occupational injuries logbooks of two shipyards were examined. Twelve accidents due to the utilization of high-pressure equipment were reported. Data were obtained by reviewing the employers' logbooks and hospital records. RESULTS Contaminants were epoxy paint and stucco, paint solvent, hydraulic and industrial oil. In three cases, the palm was injured and the fingers in nine. In 91% of cases, the accident occurred in the last 2 hr of the work shift. Mishandling (n = 9) and rupture of the high-pressure equipment (n = 3) were the causes. Mean time to medical treatment was 42.5 hr (SD 56.0). Ten workers were permanently disabled. CONCLUSIONS Prompt surgical intervention with debridement and decompression was done in two workers, who healed completely. The ten patients who were not treated immediately eventually experienced a high rate of disability and five of them also suffered amputation.","{'model': 'tldr@v2.0.0', 'text': 'Ten workers were permanently disabled and five of them also suffered amputation after accidents due to the utilization of high-pressure equipment in two shipyards.'}",
-Time-resolved four-wave mixing in InAs/InGaAs quantum-dot amplifiers under electrical injection,P. Borri,"Time-resolved four-wave mixing in an InAs/InGaAs/GaAs electrically pumped quantum-dot amplifier is measured at room temperature for different applied bias currents going from optical absorption to gain of the device. The four-wave mixing signal from 140 fs pulses shows a transition from a delayed photon-echo response in the absorption regime to a prompt free polarization decay in the gain regime. This corresponds to a pronounced reduction of the dephasing time from 250 fs at zero bias to less than 50 fs at the maximum applied current. The four-wave mixing response at transparency of the device shows a composite structure with both photon echo and free-polarization decay. This is a signature of the digital occupation number in quantum dots, resulting at transparency in a signal from dots occupied with either zero or two excitons corresponding to absorption or gain of the dot ground state.",,https://pubs.aip.org/aip/apl/article-pdf/76/11/1380/7813934/1380_1_online.pdf
-Cosmic Ray and Neutrino Emission from Gamma-Ray Bursts with a Nuclear Cascade,D. Biehl,"We discuss neutrino and cosmic-ray emission from Gamma-Ray Bursts (GRBs) with the injection of nuclei, where we take into account that a nuclear cascade from photo-disintegration can fully develop in the source. One of our main objectives is to test if recent results from the IceCube and the Pierre Auger Observatory can be accommodated with the paradigm that GRBs are the sources of Ultra-High Energy Cosmic Rays (UHECRs). While our key results are obtained using an internal shock model, we discuss how the secondary emission from a GRB shell can be interpreted in terms of other astrophysical models. It is demonstrated that the expected neutrino flux from GRBs weakly depends on the injection composition, which implies that prompt neutrinos from GRBs can efficiently test the GRB-UHECR paradigm even if the UHECRs are nuclei. We show that the UHECR spectrum and composition, as measured by the Pierre Auger Observatory, can be self-consistently reproduced in a combined source-propagation model. In an attempt to describe the energy range including the ankle, we find tension with the IceCube bounds from the GRB stacking analyses. In an alternative scenario, where only the UHECRs beyond the ankle originate from GRBs, the requirement for a joint description of cosmic-ray and neutrino observations favors lower luminosities, which does not correspond to the typical expectation from {\gamma}-ray observations.",,https://www.aanda.org/articles/aa/pdf/2018/03/aa31337-17.pdf
-"High‐Pressure Hand Injection Injuries Caused by Dry Cleaning Solvents: Case Reports, Review of the Literature, and Treatment Guidelines",K. Gutowski,"&NA; A previously unreported subset of high‐pressure injection injuries, namely those involving solvents used in the garment dry cleaning industry, is presented. Dry cleaning solutions contain isoparaffinic hydrocarbons, methoxypropanol, and dichlorofluoroethane. Although these solvents have limited potential for systemic toxicity, severe local toxicity causing tissue necrosis often results in loss of the injured digit. Proper treatment includes prompt surgical exploration, careful débridement and irrigation, intravenous antibiotics, and in selected cases, high‐dose systemic corticosteroids. (Plast. Reconstr. Surg. 111: 174, 2003.)","{'model': 'tldr@v2.0.0', 'text': 'A previously unreported subset of high‐pressure injection injuries, namely those involving solvents used in the garment dry cleaning industry, is presented.'}",
-Cardiorespiratory distress after sumatriptan given by injection.,W. Inman,"incompetence in 1985 was admitted in ventricular tachycardia without angina. He had started taking sumatriptan 30 days before admission. Each of the eight injections had been followed by ""hot surges in the throat,"" and four of these were followed by a definite sensation of palpitation (up to 160 beats per minute). He required DC cardioversion and amiodarone to correct the arrhythmia. A previous drug point showed that subcutaneous sumatriptan could produce ST elevation in susceptible subjects,' and a recent study has shown that vasospasm in patients with minor coronary artery stenosis can precipitate ventricular tachycardia and fibrillation.4 Our first patient would almost certainly have died without the prompt arrival of the ambulance service. The timing of her ventricular tachycardia in relation to the sumatriptan injection and the lack of previous or subsequent problems strongly suggest that sumatriptan induced her arrhythmia. The second case is less clear in that there was a previous history of cardiac surgery, but eight separate administrations of the drug produced similar symptoms, which were shown to be due to ventricular tachycardia on hospital admission. Thus we suggest that subcutaneous sumatriptan should be administered with caution and that, ideally, the first dose should be given while the cardiac rhythm is being closely monitored.","{'model': 'tldr@v2.0.0', 'text': 'The second case is less clear in that there was a previous history of cardiac surgery, but eight separate administrations of the drug produced similar symptoms, which were shown to be due to ventricular tachycardia on hospital admission, and it is suggested that subcutaneous sumatriptan should be administered with caution and that the first dose should be given while the cardiac rhythm is being closely monitored.'}",https://europepmc.org/articles/pmc1882911?pdf=render
-Changes in sleep in response to intracerebral injection of insulin-like growth factor-1 (IFG-1) in the rat.,F. Obál,"Changes in sleep were studied during 6 hours after intracerebroventricular (ICV) administration of Insulin-like growth factor-1 (IGF-1) or the structurally related insulin. IGF-1 was injected either at dark onset (0.05 or 0.5 microgram) or 6 hours after light onset (0.05, 0.5, or 5.0 microgram). The small dose of IGF-1 consistently, albeit modestly, enhanced NREMS over the 6 hour postinjection period in both the dark and the light cycles (REMS increased only at night). The NREMS-promoting activity vanished when the dose was increased to 0.5 microgram, and 5.0 microgram IGF-1 elicited a marked and prompt suppression in NREMS. Heat-inactivated IGF-1 (0.05 microgram) did not alter sleep. On a molar base, the NREMS-promoting dose of insulin was higher than that of IGF-1. Late (hours 7-17 postinjection) enhancements in EEG slow wave activity during NREMS were observed after 5.0 microgram IGF-1. The results indicate that IGF-1 can promote NREMS and may contribute to the mediation of the effects of GH on sleep. The acute sleep-suppressive activity of the high dose of IGF-1 is attributed to an inhibition of endogenous growth hormone-releasing hormone (GHRH).","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that IGF-1 can promote NREMS and may contribute to the mediation of the effects of GH on sleep, and the acute sleep-suppressive activity of the high dose of IGF-2 is attributed to an inhibition of endogenous growth hormone-releasing hormone.'}",
-DILUTE POVIDONE-IODINE PROPHYLAXIS MAINTAINS SAFETY WHILE IMPROVING PATIENT COMFORT AFTER INTRAVITREAL INJECTIONS.,M. Peden,"PURPOSE To report the rates of postintravitreal injection (IVT) endophthalmitis with topical conjunctival application of various concentrations of povidone-iodine (PI), including no PI. METHODS Retrospective chart review of patients receiving IVTs performed in a single practice between January 2011 and June 2016. Concentration of PI for all injections was recorded and cases of endophthalmitis identified and reviewed. RESULTS A total of 35,060 IVTs in 1854 patients were included from the 5.5-year period. 29,281 injections were performed with standard 5% PI, 5,460 injections with diluted PI (3,731 with 2.5%, 1,673 with 1.25%, 56 with 0.625%), and 319 IVTs with no PI. Incidence of patient-reported PI sensitivity occurred in 15.9% of patients. Fourteen cases of endophthalmitis were identified: 12 in eyes that received 5% PI, one in an eye that received 1.25% PI, and one in an eye receiving no PI. The incidence of endophthalmitis was 0.04% for 5% PI, 0.02% for dilute PI, and 0.31% for no PI prophylaxis. All cases underwent prompt vitrectomy and had positive cultures for coagulase-negative Staphylococcus. CONCLUSION Application of dilute PI solution to the conjunctiva at the time of IVT is an effective alternative to 5% PI for endophthalmitis prophylaxis in betadine-sensitive patients.","{'model': 'tldr@v2.0.0', 'text': 'Application of dilute PI solution to the conjunctiva at the time of IVT is an effective alternative to 5% PI for endophthalmitis prophylaxis in betadine-sensitive patients.'}",
-Transient heat transport studies using laser ablated impurity injection in JET,P. Galli,"Following impurity injection by laser ablation (LA) in JET plasmas, the electron temperature, Te, is observed to drop at a rate that cannot be accounted for by changes in radiated power, ΔPrad. Te starts to drop promptly over a large fraction of the plasma volume, which can be explained by a non-local change in electron heat diffusivity. The change in diffusivity, Δχe, is generally short lived, lasting a few milliseconds, even if ΔPrad can persist for longer times. No clear relation between ΔPrad and the strength of the prompt plasma response to LA can be observed, but only those events with ΔPrad/Prad > 0.5 give rise to detectable Te perturbations. In hot ion H mode plasmas, Δχe is found to be spatially non-uniform and increasing exponentially with the time delay of the LA relative to the onset of the H mode. Examples of non-monotonic Δχe profiles are found in plasmas of lower heating power. Energy transport models combining local and non-local (or strongly non-linear) features are possible candidates for an explanation of these observations.",,
-Thyrotropin-Releasing Hormone: Evidence for Thyroid Response to Intravenous Injection in Man,C. Hollander,"Administration of thyrotropin-releasing hormone to normal subjects causes a prompt rise in plasma thyrotropin concentration, followed by a significant increase in circulating plasma triiodothyronine. These observations may prove to be of value in simultaneously assessing the ability of the pituitary and thyroid glands to respond to their trophic hormones.","{'model': 'tldr@v2.0.0', 'text': 'Observations of the response of the pituitary and thyroid glands to thyrotropin-releasing hormone to normal subjects may prove to be of value in simultaneously assessing the ability of these glands to respond to their trophic hormones.'}",
-Oozing type cardiac rupture repaired with percutaneous injection of fibrin-glue into the pericardial space: case report.,H. Murata,"Two patients, a 56-year-old man and an 81-year-old woman who were admitted to hospital because of anteroseptal acute myocardial infarction, were initially treated successfully with direct percutaneous transluminal coronary angioplasty. However, both patients later developed sudden cardiogenic shock due to cardiac tamponade caused by left ventricular free wall rupture (LVFWR). Prompt, life-saving pericardiocentesis was performed, then fibrin-glue was percutaneously injected into the pericardial space. After the procedure, there was no detectable pericardial effusion on echocardiography and the hemodynamic state became stable. The surgical treatment was the standard procedure for LVFWR, but percutaneous fibrin-glue therapy can also be considered for oozing type LVFWR.",,
-The renal response to intravenous injection of sodium chloride solutions in man.,B. Crawford,"The administration of isotonic saline solution to dogs in quantities equal to one-half or more of the volume of extracellular fluid leads to a prompt increase in glomerular filtration rate and the excretion of sodium and chloride. Although factors possibly influencing tubular reabsorptive activity, such as the adrenal and neurohypophysial hormones, are not as yet amenable to quantitative study, the general correlation between filtration rate and sodium excretion in the dog has led many investigators to believe that changes in glomerular activity constitute one of the most important mechanisms in the maintenance of salt and water balance in this animal (1-5). So effective is this renal response that, with other adjunctival mechanisms, dogs may be maintained on diets containing 4 gm. of salt per kg. per day (corresponding to 280 gm. per day in a 70 kg. man) without gain in weight or apparent disturbance in salt and water balance (5). An extensive literature is available (6) on the excretion of sodium, chloride and water in patients receiving saline infusions, but in great part the information contained therein is merely descriptive of the fact of diuresis (or absence of diuresis) and is not correlated with the measurable physiological mechanisms concerned in the excretion of sodium and water. No information is available on the effect of saline solutions on renal function in man, and the present studywas undertaken to obtain information, comparable to that available on the dog, on the changes in filtration rate, renal blood flow and electrolyte excretion after the rapid administration of saline infusions.","{'model': 'tldr@v2.0.0', 'text': 'The present study was undertaken to obtain information, comparable to that available on the dog, on the changes in filtration rate, renal blood flow and electrolyte excretion after the rapid administration of saline infusions.'}",http://www.jci.org/articles/view/102553/files/pdf
-Prompt gamma ray spectrometry for in vivo measurement of boron-10 concentration in rabbit brain tissue.,K. Mukai,"Boron-10 concentrations in the brain of live rabbits were measured by prompt gamma ray spectrometry at intervals over a 24-hour period. Boron-10 concentrations in the blood and cerebrospinal fluid (CSF) were also measured. Animals were killed at each interval to obtain brain tissues to measure the boron-10 concentration in the brainstem, cerebral cortex, cerebellar cortex, and basal ganglia, as well as the whole brain. Boron-10 concentrations in the live brain did not differ significantly from those measured in whole brain tissue. Boron-10 concentrations in the blood were much higher than in the brain at each interval after injection. These boron-10 concentrations showed a similar pattern of initial rapid decrease, followed by a more gradual decrease. There was little boron-10 present in the CSF. The brainstem contained a significantly larger concentration of boron-10 than the other tissues. Prompt gamma ray spectrometry has the potential for direct measurement of boron-10 concentrations in the brain of patients undergoing boron neutron capture therapy.",,https://www.jstage.jst.go.jp/article/nmc1959/35/12/35_12_855/_pdf
-High-pressure injection injuries to the hand.,D. Vasilevski,"High-pressure injection injury hides the true extent of the lesions behind an apparent small and harmless puncture of the finger or the hand. Through clinical description, we wish to point out the need for prompt treatment to avoid mutilating and function-threatening complications. We wish to outline the role of the emergency physician who must be aware of the incidence of high-pressure injection injury and become accustomed to early referral to a surgeon, experienced in extensive surgical exploration, removal of foreign bodies, and rehabilitation. The open-wound technique gives the best results. We also point out that failure to refer may become an increasing focus of negligence claims.","{'model': 'tldr@v2.0.0', 'text': 'The role of the emergency physician who must be aware of the incidence of high-pressure injection injury and become accustomed to early referral to a surgeon, experienced in extensive surgical exploration, removal of foreign bodies, and rehabilitation is outlined.'}",
-Observationally constraining gravitational wave emission from short gamma-ray burst remnants,P. Lasky,"Observations of short gamma-ray bursts indicate ongoing energy injection following the prompt emission, with the most likely candidate being the birth of a rapidly rotating, highly magnetised neutron star. We utilise X-ray observations of the burst remnant to constrain properties of the nascent neutron star, including its magnetic field-induced ellipticity and the saturation amplitude of various oscillation modes. Moreover, we derive strict upper limits on the gravitational wave emission from these objects by looking only at the X-ray light curve, showing the burst remnants are unlikely to be detected in the near future using ground-based gravitational wave interferometers such as Advanced LIGO.",,https://academic.oup.com/mnras/article-pdf/458/2/1660/18242739/stw435.pdf
-Experimental allergic encephalomyelitis after the excision of the injection site of antigen-adjuvant emulsion.,J. Freund," It has been repeatedly shown that when certain antigenic material dissolved or suspended in physiological salt solution is injected into the subcutaneous tissue of animals, the surgical removal of the site of injection within a few minutes after injection does not prevent antibody formation (1). These observations are readily explained by the prompt dissemination of injected material. It was demonstrated by Cannon and Burt (2) that when Staphylococcus aureus or Serratia marcescens are injected into tissue of the ear of the rabbit, these microorganisms can be cultured from the blood, spleen, and liver even if the ear is amputated within a few minutes after injection. Furthermore, McMaster and Kidd (3) have shown that an intracutaneous injection introduces material very promptly into regional lymph nodes. When antigens incorporated in water-in-paraffin oil emulsion with or without mycobacteria are injected into the tissues, the immune response is abundant and long sustained.","{'model': 'tldr@v2.0.0', 'text': None}",https://journals.aai.org/jimmunol/article-pdf/75/6/454/1473851/ji0750060454.pdf
-Bench test and preliminary results of vertical compact torus injection experiments on the STOR-M tokamak,D. Liu,The University of Saskatchewan compact torus injector has been modified to allow vertical injection of the compact torus (CT) into the STOR-M tokamak. The injector stayed at the horizontal position and the CT trajectory was bent by 90° using a curved conducting drift tube. The curved drift tube did not significantly slow the CT velocity down or change the CT magnetic field topology. Preliminary vertical CT injection experiments have revealed a prompt increase in the line averaged electron density and in the soft x-ray radiation level in the tokamak discharge immediately following vertical CT injection. Suppression of the m = 2 Mirnov oscillation amplitude has also been observed after CT injection.,,
-Repeated intravitreous ranibizumab injections for diabetic macular edema and the risk of sustained elevation of intraocular pressure or the need for ocular hypotensive treatment.,S. Bressler,"IMPORTANCE For the management of retinal disease, the use of intravitreous injections of anti-vascular endothelial growth factor has increased. Recent reports have suggested that this therapy may cause sustained elevation of intraocular pressure (IOP) and may potentially increase the risk of glaucoma for patients with retinal disease. OBJECTIVE To assess the risk of sustained IOP elevation or the need for IOP-lowering treatments for eyes with diabetic macular edema following repeated intravitreous injections of ranibizumab. DESIGN, SETTING, AND PARTICIPANTS An exploratory analysis was conducted within a Diabetic Retinopathy Clinical Research Network randomized clinical trial. Study enrollment dates were from March 20, 2007, to December 17, 2008. Of 582 eyes (of 486 participants) with center-involved diabetic macular edema and no preexisting open-angle glaucoma, 260 were randomly assigned to receive a sham injection plus focal/grid laser treatment, and 322 were randomly assigned to receive ranibizumab plus deferred or prompt focal/grid laser treatment. MAIN OUTCOMES AND MEASURES The cumulative probability of sustained IOP elevation, defined as IOP of at least 22 mm Hg and an increase of at least 6 mm Hg from baseline at 2 consecutive visits, or the initiation or augmentation of ocular hypotensive therapy, through 3 years of follow-up. RESULTS The mean (SD) baseline IOP in both treatment groups was 16 (3) mm Hg (range, 5-24 mm Hg). The cumulative probability of sustained IOP elevation or of initiation or augmentation of ocular hypotensive therapy by 3 years, after repeated ranibizumab injections, was 9.5% for the participants who received ranibizumab plus prompt or deferred focal/grid laser treatment vs 3.4% for the participants who received a sham injection plus focal/grid laser treatment (difference, 6.1% [99% CI, -0.2% to 12.3%]; hazard ratio, 2.9 [99% CI, 1.0-7.9]; P = .01). The distribution of IOP and the change in IOP from baseline at each visit through 3 years were similar in each group. CONCLUSIONS AND RELEVANCE In eyes with center-involved diabetic macular edema and no prior open-angle glaucoma, repeated intravitreous injections of ranibizumab may increase the risk of sustained IOP elevation or the need for ocular hypotensive treatment. Clinicians should be aware of this risk and should consider this information when following up with patients who have received intravitreous injections of anti-vascular endothelial growth factor for the treatment of diabetic macular edema.","{'model': 'tldr@v2.0.0', 'text': 'In eyes with center-involved diabetic macular edema and no prior open-angle glaucoma, repeated intravitreous injections of ranibizumab may increase the risk of sustained IOP elevation or the need for ocular hypotensive treatment.'}",https://europepmc.org/articles/pmc4496789?pdf=render
-"The Afterglow, Energetics, and Host Galaxy of the Short-Hard Gamma-Ray Burst 051221a",A. Soderberg,"We present detailed optical, X-ray, and radio observations of the bright afterglow of the short gamma-ray burst 051221a obtained with Gemini, Swift XRT, and the Very Large Array, as well as optical spectra from which we measure the redshift of the burst, z = 0.5464. At this redshift the isotropic-equivalent prompt energy release was about 1.5 × 1051 ergs, and using a standard afterglow synchrotron model, we find that the blast wave kinetic energy is similar, EK,iso ≈ 8.4 × 1051 ergs. An observed jet break at t ≈ 5 days indicates that the opening angle is θj ≈ 7° and the total beaming-corrected energy is therefore ≈2.5 × 1049 ergs, comparable to the values inferred for previous short GRBs. We further show that the burst experienced an episode of energy injection by a factor of 3.4 between t = 1.4 and 3.4 hr, which was accompanied by reverse shock emission in the radio band. This result provides continued evidence that the central engines of short GRBs may be active significantly longer than the duration of the burst and/or produce a wide range of Lorentz factors. Finally, we show that the host galaxy is actively forming stars at a rate of about 1.6 M☉ yr-1, yet exhibits evidence for an appreciable population of old stars (~1 Gyr) and near-solar metallicity. These properties are intermediate between those of long GRB hosts and previous short burst hosts. The lack of bright supernova emission and the low circumburst density (n ~ 10-3 cm-3), however, continue to support the idea that short bursts are not related to massive stellar death. Given that the total energy release is larger than the predicted yield for a neutrino annihilation mechanism, this suggests that magnetohydrodynamic processes may be required to power the burst.",,https://iopscience.iop.org/article/10.1086/506429/pdf
-Generation of runaway electrons during the thermal quench in tokamaks,P. Aleynikov,"This work provides a systematic description of electron kinetics during impurity dominated thermal quenches. A Fokker–Planck equation for the hot electrons and a power balance equation for the bulk plasma are solved self-consistently, with impurity radiation as the dominant energy loss mechanism. We find that runaway production is facilitated by heavy injection of impurities up to prompt conversion of the total current into a sub-MeV runaway current. We also find that runaway formation is less efficient in plasmas with high pre-quench temperatures and predict significant radial variation of the runaway seed in such plasmas.",,
-Experimental background for prompt treatment with DTPA of 239Pu-contaminated wounds.,V. Volf,"Rats were injected intramuscularlywith 239Pu in the form ofnitrate, citrate or DTPA chelate. Differences were observed in the distribution pattern of 239Pu and in the response to DTPA treatment, depending on the 239Pu-compound injected. In general, a 3-hr intravenous infusion of Na,CaDTPA (30-120 pmoles kg-l) was more effective than the rapid intravenous injection. Furthermore, the effect of DTPA was dose-dependent only as far as 239Pu-citrate was concerned. Best effects were, however, achieved by injecting DTPA intramuscularly into the vicinity of the 239P~-deposit. Enhanced absorption of the chelated 238Pu did not result in its increased uptake by the body organs. The practical implications of these findings are discussed.","{'model': 'tldr@v2.0.0', 'text': 'In general, a 3-hr intravenous infusion of Na,CaDTPA (30-120 pmoles kg-l) was more effective than the rapid intravenous injection and the effect of DTPA was dose-dependent only as far as 239Pu-citrate was concerned.'}",
-Transient Decrease in Arm Blood Flow following Accidental Intra-Arterial Injection of Propofol into the Left Brachial Artery,J. Brimacombe,"trauma, basic teaching is resuscitation and urgent laparotomy. We believe this case represents a severe vagal response to abdominal insufflation to carbon dioxide. It also emphasises the need for vigilant clinical and mechanical monitoring. Prompt resuscitation, including the early use of atropine, is recommended in all severe vagal reactions. Low-flow peritoneal insufflation should always proceed for some time before high-flow is employed. Consideration should also be given to urgent laparotomy in all cases of cardiac arrest during laparoscopy. REFERENCES","{'model': 'tldr@v2.0.0', 'text': 'This case represents a severe vagal response to abdominal insufflation to carbon dioxide and emphasises the need for vigilant clinical and mechanical monitoring.'}",https://journals.sagepub.com/doi/pdf/10.1177/0310057X9402200310
-Prompt stimulation by TSH of thyroid oxidation of thiocyanate.,S. Otaki,"TSH (4 U, iv)1 was given to rats on a low iodine diet 30 min before an ip injection of KSCN labeled with 35S; within 15 min after the SCN injection, thyroid oxidation of thiocyanate was significantly greater than in controls not given TSH. Prompt stimulation of thyroid oxidation of SCN was also observed in Purinafed rats, including animals acutely pretreated with perchlorate. The formation of sulfate by calf thyroid slices incubated in 0.1 mil KSCN was enhanced within 15 min by TSH (50 mU/ ml). In canine thyroid slices, supplementing the medium with dbcAMP also stimulated thiocyanate oxidation. Thiocyanate oxidation by bovine thyroid slices was enhanced by the presence of 0.1 mM H2O2 or a hydrogen peroxide generating system in the medium. Both basal and TSH stimulated oxidation of SCN were inhibited by methimazole, thiols, anaerobiosis and inhibitors of electron transport. The promptness with which TSH augments thyroid oxidation of thiocyanate, as well as other features of this response in vivo and in vit...","{'model': 'tldr@v2.0.0', 'text': 'TSH (4 U, iv)1 was given to rats on a low iodine diet 30 min before an ip injection of KSCN labeled with 35S; within 15 min after the SCN injection, thyroid oxidation of thiocyanate was significantly greater than in controls not given TSH.'}",
-Periorbital Injectables: Understanding and Avoiding Complications,C. Hwang,"Periorbital rejuvenation with neurotoxins and dermal fillers address several aging changes. Safe and effective results require a thorough understanding of periorbital anatomy, proper injection techniques, and complications of these products. Prompt recognition and treatment of complications can minimize their adverse impacts. Complications can be divided into ischaemic and non-ischaemic effects. Hylauronidase, an enzyme that degrades hyaluronic acid, may improve outcomes after intravascular hyaluronic acid fillers.","{'model': 'tldr@v2.0.0', 'text': 'Hylauronidase, an enzyme that degrades hyaluronic acid, may improve outcomes after intravascular hyaluonic acid fillers.'}",
-A Comprehensive Analysis of Swift XRT Data. II. Diverse Physical Origins of the Shallow Decay Segment,E. Liang,"The origin of the shallow decay segment in Swift XRT light curves remains a puzzle. We analyze the properties of this segment with a sample of 53 long Swift GRBs detected before 2007 February. We show that the distributions of the sample's characteristics are lognormal or normal, and its isotropic X-ray energy (Eiso,X) is linearly correlated with the prompt gamma-ray energy but with a steeper photon spectrum, aside from some X-ray flashes. No significant spectral evolution is observed from this phase to the following phase, and the latter is usually consistent with external-shock models, implying that the shallow decay is also of external-shock origin, likely a refreshed external shock. Within the refreshed-shock model, the data are generally consistent with a roughly constant injection luminosity up to the end of this phase, tb. A positive correlation between Eiso,X and tb also favors this scenario. Among the 13 bursts that have well-sampled optical light curves, six have an optical break around tb and the breaks are consistent with being achromatic. However, the other seven either do not show an optical break or have a break at an epoch different from tb. This raises a concern for the energy injection scenario, suggesting that the optical and X-ray emission may not be the same component, at least for some bursts. There are four significant outliers in the sample, GRBs 060413, 060522, 060607A, and 070110. The shallow decay phase in these bursts is immediately followed by a very steep decay after tb, which is inconsistent with any external-shock model. The optical data for these bursts evolve independently from the X-ray data. These X-ray plateaus likely have an internal origin and demand continuous operation of a long-term central engine. We conclude that the observed shallow decay phase likely has diverse physical origins.",,https://iopscience.iop.org/article/10.1086/521870/pdf
-Local Injection of Hydrocortisone and Procaine in Osteo-Arthritis of the Hip Joint,V. Leveaux,"The local injection of therapeutic substances to relieve symptoms in osteo-arthritis has long been a controversial method of treatment. Desmarais (1952) reviewed the use of various substances, including procaine lactic acid, acid potassium phosphate, and procaine. He further critically examined the results of his own series of 189 injections into hip or knee joint with alkaline procaine, lactic acid, saline, or needle prick alone. There appeared to be little difference between the results with these various materials, except that procaine gave slightly more pain relief and was probably the substance of choice. The introduction of hydrocortisone (Compound F) stimulated further studies in relation to osteoarthritis of the hip joint. Hollander, Brown, Jessar, and Brown (1951) reported the responses of two patients suffering from unilateral osteo-arthritis of the hip to the injection of 25 mg. hydrocortisone into the region of the affected joint. They stated that ""a prompt alleviation of symptoms and improvement in range of motion was noted in both patients after 24 hours; this persisted for 7 days in one and 13 days in the other"". Two years later Hollander (1953) was able to report on 77 injections into osteoarthritic hips, in which relief was obtained for at least 3 days in 47 per cent. of cases.","{'model': 'tldr@v2.0.0', 'text': 'The local injection of therapeutic substances to relieve symptoms in osteo-arthritis has long been a controversial method of treatment and hydrocortisone (Compound F) stimulated further studies in relation to osteoarthritis of the hip joint.'}",https://ard.bmj.com/content/annrheumdis/15/4/330.full.pdf
-Update on HIV/HCV Coinfection,V. Soriano,,"{'model': 'tldr@v2.0.0', 'text': 'Issues that should be addressed include the following: (1) rising incidence of acute hepatitis C in men who have sex with men, and expansion/recrudescence of injection drug use in some settings/regions; (2) adverse drug interactions between antiretrovirals and DAA; and (3) high cost of DAA, which may lead many to defer or fail to access appropriate therapy.'}",
-Roles for thrombin and fibrin(ogen) in cytokine/chemokine production and macrophage adhesion in vivo.,F. Szaba,"Extravascular coagulation leading to fibrin deposition accompanies many immune and inflammatory responses. Although recognized by pathologists for decades, and probably pathologic under certain conditions, the physiologic functions of extravascular coagulation remain to be fully defined. This study demonstrates that thrombin can activate macrophage adhesion and prompt interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) production in vivo. Peritoneal macrophages were elicited with thioglycollate (TG) and then activated in situ, either by intraperitoneal injection of lipopolysaccharide (LPS) or by injection of antigen into mice bearing antigen-primed T cells. Others previously established that such treatments stimulate macrophage adhesion to the mesothelial lining of the peritoneal cavity. The present study demonstrates that thrombin functions in this process, as macrophage adhesion was suppressed by Refludan, a highly specific thrombin antagonist, and induced by direct peritoneal administration of purified thrombin. Although recent studies established that protease activated receptor 1 (PAR-1) mediates some of thrombin's proinflammatory activities macrophage adhesion occurred normally in PAR-1-deficient mice. However, adhesion was suppressed in fibrin(ogen)-deficient mice, suggesting that fibrin formation stimulates macrophage adhesion in vivo. This study also suggests that fibrin regulates chemokine/cytokine production in vivo, as direct injection of thrombin stimulated peritoneal accumulation of IL-6 and MCP-1 in a fibrin(ogen)-dependent manner. Given that prior studies have clearly established inflammatory roles for PAR-1, thrombin probably has pleiotropic functions during inflammation, stimulating vasodilation and mast cell degranulation via PAR-1, and activating cytokine/chemokine production and macrophage adhesion via fibrin(ogen).","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that thrombin can activate macrophage adhesion and prompt interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) production in vivo, and that fibrin regulates chemokine/cytokine production in vitro and in vivo.'}",https://ashpublications.org/blood/article-pdf/99/3/1053/1680716/1053.pdf
-Single extra-amniotic injection of prostaglandin E2 in viscous gel to induce mid-trimester abortion.,I. Mackenzie,"In a preliminary study a single extra-amniotic injection of 1.5 mg of prostaglandin E-2 incorporated into an aqueous viscous gel was given to 24 patients aborted within 24 hours, and the mean induction-abortion interval (plus or minus S.E. of mean) was 13.5 plus or minus 1.5 hours. Vomiting occurred in seven patients, and transient severe uterine cramps, pallor, nausea, and shivering occurred in one patient immediately after injection. Complete abortion occurred in 20patients. A delay in the time taken to abort seemed to be associated with an immediate and rapid rise in uterine tone after the injection which required prompt analgesia; this probably reflected rapid decidual absorption and dissolution of the prostaglandins away from their site of action. The degree of distention of the catheter-retaining balloon did not influence abortion times.","{'model': 'tldr@v2.0.0', 'text': 'A delay in the time taken to abort seemed to be associated with an immediate and rapid rise in uterine tone after the injection which required prompt analgesia; this probably reflected rapid decidual absorption and dissolution of the prostaglandins away from their site of action.'}",https://www.bmj.com/content/bmj/1/5952/240.full.pdf
-Needlestick injuries among health care workers of University of Alexandria Hospitals.,M. I. Hanafi,"The aims of this cross-sectional study were to investigate the prevalence and circumstances of needlestick injury (NSI) among heath care workers at University of Alexandria teaching hospitals and to assess the effectiveness of the existing control measures and standard precautions. Data were obtained by anonymous, self-reporting questionnaire from 645 nurses, physicians and ancillary staff in 2007. Around two-thirds of workers (67.9%) had suffered at least 1 NSI in the last 12 months. High-risk patients (with history of HIV, hepatitis B virus or, hepatitis C virus infection or injecting drug use) were involved in 8.2% of injuries. On evaluating the effectiveness of existing control measures, significant protective factors against NSI were: using devices with safety features (OR 0.41), adherence to infection control guidelines (OR 0.42), training in injection safety (OR 0.14), comfortable room temperature (OR 0.32) and availability of a written protocol for prompt reporting (OR 0.37).","{'model': 'tldr@v2.0.0', 'text': 'On evaluating the effectiveness of existing control measures, significant protective factors against NSI were: using devices with safety features, adherence to infection control guidelines, training in injection safety, and availability of a written protocol for prompt reporting.'}",
-The osmoregulation of vasopressin.,G. Robertson,,"{'model': 'tldr@v2.0.0', 'text': 'New assay methods now have made it possible to begin to characterize osmoregulatory function in a more concrete and comprehensive way and also to use such information to analyze systematically certain clinical disorders of salt and water balance.'}",
-Intra-Arterial Thrombolytic Therapy Is Not a Therapeutic Option for Filler-Related Central Retinal Artery Occlusion,Yi-chun Chen,"Abstract Cosmetic facial filler-related central retinal artery occlusion (CRAO) is a devastating complication of facial hyaluronic acid (HA) injection and can be managed by intra-arterial thrombolytic therapy (IATT). The authors report on a 20-year-old woman who developed unilateral CRAO due to facial HA injection and who, despite prompt IATT, lost vision. A review of the related literature found 14 other female patients who developed cosmetic facial filler-related CRAO and accepted IATT management. In no case was vision loss clinically improved. IATT is not an effective preventive treatment of dermal filler-associated CRAO. The authors suggest careful preprocedural patient selection to prevent this complication.","{'model': 'tldr@v2.0.0', 'text': 'CRAO is a devastating complication of facial hyaluronic acid injection and can be managed by intra-arterial thrombolytic therapy (IATT), but IATT is not an effective preventive treatment of dermal filler-associated CRAO.'}",
-Plasma delta‐9‐tetrahydrocannabinol concentrations and clinical effects after oral and intravenous administration and smoking,A. Ohlsson,,"{'model': 'tldr@v2.0.0', 'text': 'After oral THC, the onset of clinical effects was much slower and lasted longer, but effects occurred at much lower plasma concentrations than after the other two methods of administration, suggesting that brain concentrations were increasing as plasma concentrations decreased.'}",
-Prompt striations observed in a barium thermite release at 335 km,D. Simons,"Three barium clouds have been released from a single rocket at altitudes of 335, 443, and 535 km. The releases were by means of barium thermite injection. Adding the different cross-field velocity components of the rocket at each of the releases resulted in barium clouds with different kinetic velocity distributions. The barium cloud with the most peaked velocity distribution striated promptly, in good agreement with theoretical predictions based on a kinetic plasma instability. Details of the data analysis and experiment are discussed.",,
-History of cord blood transplantation,E. Gluckman,,"{'model': 'tldr@v2.0.0', 'text': 'Results of unrelated allogeneic cord blood transplants in malignant and nonmalignant diseases, in adults and children, show that, compared with HLA-matched unrelated BM transplant, cord blood has several advantages, including prompt availability of the transplant, decrease of GVHD and better long-term immune recovery resulting in a similar long- term survival.'}",
-Effects of acetylcholine injection into the hypothalamus on the insulin and glucagon release.,K. Ishikawa,"To evaluate the role of cholinergic neurones within the hypothalamus in modulating pancreatic-endocrine function, plasma levels of glucagon, insulin and glucose were measured after microinjection of acetylcholine into the ventromedial hypothalamic nucleus (VMH) of rabbits. Injection of acetylcholine (5 X 10(-8) mol in 2 microliter of saline) into the VMH of unanesthetized, unrestrained animals resulted in a prompt elevation of immunoreactive glucagon and glucose levels without any significant change in immunoreactive insulin level. The suppression of insulin secretion following cholinergic stimulation of the VMH was not counteracted by intravenous infusion of glucose (0.25 g/kg). Bilateral adrenalectomy prevented both hyperglucagonemia and hyperglycemia induced by the acetylcholine injection into the VMH. In adrenalectomized rabbits, exogenous glucose load was followed by a significant rise in the plasma insulin level despite the hypothalamic injection of acetylcholine. However, the response of insulin release to glucose load in these animals was much less than in adrenalectomized, saline-treated, control animals. These data support the conclusion that the action of acetylcholine within the hypothalamus on the pancreatic hormone secretions is mediated to a large part through sympatho-adrenomedullary activity. However, a part of the decreased insulin response to glucose may be mediated by direct innervation of the pancreas.","{'model': 'tldr@v2.0.0', 'text': 'The conclusion that the action of acetylcholine within the hypothalamus on the pancreatic hormone secretions is mediated to a large part through sympatho-adrenomedullary activity is supported, however, a part of the decreased insulin response to glucose may be mediated by direct innervation of the pancreas.'}",
-Injected mass deposition thresholds for lithium granule instigated triggering of edge localized modes on EAST,R. Lunsford,"The ability of an injected lithium granule to promptly trigger an edge localized mode (ELM) has been established in multiple experiments. By horizontally injecting granules ranging in diameter from 200 microns to 1 mm in diameter into the low field side of EAST H-mode discharges we have determined that granules with diameter >600 microns are successful in triggering ELMs more than 95% of the time. It was also demonstrated that below 600 microns the triggering efficiency decreased roughly with granule size. Granules were radially injected from the outer midplane with velocities ~80 m s−1 into EAST upper single null discharges with an ITER like tungsten monoblock divertor. These granules were individually tracked throughout their injection cycle in order to determine their efficacy at triggering an ELM. For those granules of sufficient size, ELM triggering was a prompt response to granule injection. By simulating the granule injection with an experimentally benchmarked neutral gas shielding (NGS) model, the ablatant mass deposition required to promptly trigger an ELM is calculated and the fractional mass deposition is determined.",,
-Electromagnetic effects in the analysis of prompt striations,J. L. Sperling,"In a number of barium injection experiments in the ionosphere, field-aligned structures have been observed to develop on the ion-cyclotron time scale. These structures, called prompt striations, have been attributed to plasma jetting across the ambient magnetic field and have previously been analyzed in the electrostatic limit. The electrostatic analysis is reasonable for experiments in which the directed ion kinetic energy per unit volume is much smaller than the magnetic field pressure; however, for higher altitude experiments in which the ambient magnetic field strength is weaker, justification for the electrostatic treatment fails and the full electromagnetic dispersion relation must be examined. It is demonstrated that electromagnetic effects inhibit the growth of the linear instability considered to be the source of prompt striations in the earlier experiments. Consistent with the electrostatic analysis, cold background plasma and thermal spread in the ion loss cone distribution function tend to stabilize the instability in the full electromagnetic treatment. The theory suggests that prompt striations in the magnetosphere can form from ion jets only if the plasma density is sufficiently tenuous so that electromagnetic stabilization is not complete.",,
-An accidental intra-arterial injection of thiopental on the dorsum of the foot--a case report.,Z. Khan,"Since its maiden clinical use decades back, innumerable instances of accidental intra-arterial injection of the drug have been reported mostly in the upper extremities. We report an accidental intra-arterial injection of thiopental into one of the tributaries of the arterial arch of the dorsum of the foot, but timely and prompt treatment with lidocaine and heparin together with leg rising prevented a gangrenous episode of the extremity and resulted in an uneventful recovery. This report emphasizes the importance of extreme caution while cannulating the veins on the dorsum of the foot. Attention should be focused on the patient's reactions, the feeling of an intense pain and withdrawal of the leg during injection so that timely therapeutic measures are undertaken and the sad episode is prevented.","{'model': 'tldr@v2.0.0', 'text': ""The importance of extreme caution while cannulating the veins on the dorsum of the foot is emphasized and attention should be focused on the patient's reactions, the feeling of an intense pain and withdrawal of the leg during injection so that timely therapeutic measures are undertaken and the sad episode is prevented.""}",
-Intratympanic injection of steroid for treatment of facial paralysis,F. L. Bryant,"Corticosteroids administered systemically have been used with varying degrees of success in the treatment of idiopathic facial paralysis. There is reason to believe that this drug exerts a beneficial influence based on the work of numerous investigators. On the basis of existing dehiscences along the course of the facial nerve and exposure of the chorda tympani nerve, such abnormalities might well be sites of entrance for insult to these nerves; likewise, they might serve as portals for direct medication. Such medication, a corticosteroid, might thus be introduced intratympanically. Seven earlier cases were so treated with complete recovery in six cases and 75 percent recovery in the seventh. Three more cases are reported, all with good and prompt recovery. No untoward side effects were noted.","{'model': 'tldr@v2.0.0', 'text': 'Seven earlier cases of idiopathic facial paralysis were treated with complete recovery in six cases and 75 percent recovery in the seventh, and three more cases are reported, all with good and prompt recovery.'}",
-"GRB Radiative Efficiencies Derived from the Swift Data: GRBs versus XRFs, Long versus Short",Bing Zhang,"We systematically analyze the prompt emission and the early afterglow data of a sample of 31 GRBs detected by Swift before 2005 September and estimate the GRB radiative efficiency. BAT's narrow band inhibits a precise determination of the GRB spectral parameters, and we have developed a method to estimate these parameters with the hardness ratio information. The shallow decay component commonly existing in early X-ray afterglows, if interpreted as continuous energy injection in the external shock, suggests that the GRB efficiencies previously derived from the late-time X-ray data were not reliable. We calculate two radiative efficiencies using the afterglow kinetic energy EK derived at the putative deceleration time (tdec) and at the break time (tb), when the energy injection phase ends, respectively. At tb XRFs appear to be less efficient than normal GRBs. However, when we analyze the data at tdec, XRFs are found to be as efficient as GRBs. Short GRBs have similar radiative efficiencies to long GRBs despite of their different progenitors. Twenty-two bursts in the sample are identified to have the afterglow cooling frequency below the X-ray band. Assuming ϵe = 0.1, we find ηγ(tb) usually <10% and ηγ(tdec) varying from a few percent to >90%. Nine GRBs in the sample have the afterglow cooling frequency above the X-ray band for a very long time. This suggests a very small ϵB and/or a very low ambient density n.","{'model': 'tldr@v2.0.0', 'text': 'The shallow decay component commonly existing in early X-ray afterglows, if interpreted as continuous energy injection in the external shock, suggests that the GRB efficiencies previously derived from the late-time X-rays were not reliable.'}",https://iopscience.iop.org/article/10.1086/510110/pdf
-Characterization and Reduction of NO during the Combustion of Biodiesel in a Semi-industrial Boiler,B. Bazooyar,"This paper aims to characterize and reduce the level of nitrogen monoxide (NO) during the combustion of rapeseed oil methyl ester (ROME) in a semi-industrial boiler. First, the formation of NO is characterized during transitional and steady-state operation of boiler; the influence of combustion pressure, excess air, exhaust gas temperature, spray cone angle, and combustion air swirl angle on the level of NO is evaluated, suitable burner operating points for control of NO are recognized, and contributions of thermal and prompt NO to the total level of NO are obtained. At the next level, the potential of air-staging technique (injection of extra air) in the reduction of NO is studied. Results reveal that the level of NO rises significantly in the post-flame zone 10 cm after the tail of the flame, where only the formation of thermal NO is probable in the chamber. Staged air is able to reduce the level of NO up to 10% without any negative impact on the operation of the boiler. Results also reveal that fuel sp...",,
-A Review of Current Evaluation and Management,C. Lemley,"Purpose: To review the current state of evaluation and management of various forms of endophthalmitis. Methods: A review of the literature is included, encompassing endophthalmitis occurring after ocular surgeries, intravitreal injection, trauma, and systemic infection. Based on this review, current principles and techniques for evaluating and treating these forms of endophthalmitis are discussed. Results: Postoperative endophthalmitis after cataract surgery is the most common presentation. Conclusions from the Endophthalmitis Vitrectomy Study (EVS) remain a foundation for management of postcataract surgery endophthalmitis, notably prompt intravitreal antibiotic administration after vitreous sampling, with consideration for pars plana vitrectomy in severe cases. The potential impact of advances since the EVS, such as oral fourth generation fluoroquinolones and new vitrectomy techniques, are also discussed. The management of postcataract endophthalmitis is compared and contrasted to endophthalmitis occurring after other ocular surgeries, intravitreal injection, trauma, and systemic infection. Although some principles remain common, treatment rationales differ with other forms of endophthalmitis based on differing clinical circumstances, such as the virulence of organisms that are frequently encountered. Conclusions: Endophthalmitis is a serious, potentially vision threatening condition which can present in various settings. Prompt recognition and treatment are key in maximizing outcomes. RETINA 27:662–680, 2007","{'model': 'tldr@v2.0.0', 'text': 'Current principles and techniques for evaluating and treating various forms of endophthalmitis and the potential impact of advances since the EVS, such as oral fourth generation fluoroquinolones and new vitrectomy techniques are discussed.'}",
-Combination of laser photocoagulation and intravitreal bevacizumab (Avastin®) for aggressive zone I retinopathy of prematurity,E. Chung,,"{'model': 'tldr@v2.0.0', 'text': 'The combination of indirect laser photocoagulation and intravitreal bevacizumab injection was well tolerated and induced prompt regression of aggressive zone I ROP.'}",
-Solar Flare Electron Spectra at the Sun and near the Earth,S. Krucker,"We compare hard X-ray (HXR) photon spectra observed by the RHESSI with the spectra of the electrons in the associated solar impulsive particle events observed near 1 AU by the WIND 3D Plasma and Energetic Particle (3DP) instrument. For prompt events, where the inferred injection time at the Sun coincides with the HXR burst, the HXR photon power-law spectral index γ and the in situ observed electron spectral index δ measured above 50 keV show a good linear fit, δ = γ + 0.1(±0.1), with correlation coefficient of 0.83, while for delayed events (inferred injection >10 minutes after the HXR burst) only a weak correlation with a coefficient of 0.43 is seen. The observed relationship for prompt events is inconsistent, however, with both the thin target case, where the escaping electrons come from the X-ray-producing electron population, and the thick target case where some of the accelerated source population escapes to 1 AU and the rest produce the HXRs while losing all their energy to collisions. Furthermore, the derived total number of escaping electrons correlates with the number of electrons required to produce observed X-ray flux but is only about ~0.2% of the number of HXR-producing electrons.",,https://iopscience.iop.org/article/10.1086/519373/pdf
-Non surgical coronary artery recanalization in acute transmural myocardial infarction.,D. Mathey,"In 41 consecutive patients with an acute transmural myocardial infarction (AMI) admitted within 3 hours after the onset of symptoms, we tried to recanalize the occluded coronary artery by an intracoronary infusion of streptokinase (SK) (2000 units/min). SK infusion was preceded by (1) an intracoronary injection of 0.5 mg nitroglycerin to rule out coronary artery spasm, (2) an attempt to recanalize the vessel mechanically with a flexible guidewire, and (3) an intracoronary injection of plasminogen (500 units) to increase the efficacy of the subsequent SK infusion. Coronary angiography revealed a total coronary artery occlusion in 39 patients and a subtotal occlusion in two patients. In 30 patients (73%), the occluded coronary artery was successfully recanalized within 1 hour (mean 29 ± 15 minutes), resulting in prompt contrast filling of the previously occluded vessel. An arteriosclerotic stenosis always remained at the site of the occlusion. Nitroglycerin opened the occluded coronary artery in one patient, contrast injection in seven patients and guidewire perforation in four of the 15 patients, in whom it was attempted. In 18 patients the occluded coronary artery was recanalized by intracoronary SK infusion alone. After the initial opening of the occluded coronary artery, subsequent SK infusion markedly reduced the degree of stenosis and visible thrombi disappeared. Clinically, recanalization was associated with significant relief of ischemic chest pain. None of the successfully recanalized patients died, including three patients with cardiogenic shock. Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion. Repeat angiography 7-21 days later revealed a patent coronary artery in 12 of 15 successfully recanalized patients. The left ventricular ejection fraction had significantly improved, from 37","{'model': 'tldr@v2.0.0', 'text': 'Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion.'}",
-High-pressure injection injuries of the hand.,P. Jebson,"High pressure injection trauma to the upper extremity is a rare but potentially limb-threatening injury. The index finger on the non-dominant hand of young male workers is most commonly involved. Injected materials include paint, grease and diesel oil. Prompt recognition and initiation of treatment is necessary if a disastrous outcome is to be avoided. Primary treatment consists of surgical decompression and debridement. Amputation may be necessary. With early aggressive treatment, satisfactory results can be obtained. The use of high-pressure devices in industry has resulted in an unusual but potentially limb-threatening injury to the upper extremity known as the ""high-pressure injection injury."" This injury is a surgical emergency that demands prompt diagnosis and treatment if a disastrous outcome is to be avoided. The purpose of this paper is to present our recent experience and review the clinical features, evaluation and treatment principles.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this paper is to present the recent experience and review the clinical features, evaluation and treatment principles of the ""high-pressure injection injury"" and suggest that with early aggressive treatment, satisfactory results can be obtained.'}",
-High pressure injection injury of the hand.,M. Childress,Attention is called to the importance of prompt and adequate treatment of high pressure injection injury of the hand.,"{'model': 'tldr@v2.0.0', 'text': 'Attention is called to the importance of prompt and adequate treatment of high pressure injection injury of the hand.'}",
-A subunit cytomegalovirus vaccine based on recombinant envelope glycoprotein B and a new adjuvant.,R. F. Pass,"A phase I randomized, double-blind, placebo-controlled trial was done with a cytomegalovirus (CMV) vaccine based on the envelope glycoprotein, gB, combined with a novel adjuvant, MF59. Participants received CMV gB vaccine with MF59 or CMV gB with alum or placebo at 0, 1, and 6 months. A fourth vaccine was given at 12 months to a subgroup. Levels of neutralizing antibody and antibody to gB 2 weeks after the third dose of vaccine exceeded those in seropositive control subjects. the formulation with MF59 was more immunogenic than that with alum. The optimal dose of gB appeared to be between 5 and 30 microg. The fourth dose produced a prompt rise in antibody level. There were no serious adverse events associated with vaccine. Local and systemic reactions were generally mild and, except for pain at the injection site, occurred with similar frequency in recipients of placebo and CMV vaccine.","{'model': 'tldr@v2.0.0', 'text': 'A phase I randomized, double-blind, placebo-controlled trial was done with a cytomegalovirus vaccine based on the envelope glycoprotein, gB, combined with a novel adjuvant, MF59, which was more immunogenic than that with alum.'}",https://academic.oup.com/jid/article-pdf/180/4/970/17996494/180-4-970.pdf
-"CERN Large Hadron Collider optics model, measurements, and corrections",R. Tomás,"Optics stability during all phases of operation is crucial for the LHC. Tools and procedures have been developed for rapid checks of beta beating, dispersion, and linear coupling, as well as for prompt optics corrections. Important optics errors during the different phases of the beam commissioning were observed and locally corrected using the segment-by-segment technique. The most relevant corrections at injection have been corroborated with dedicated magnetic measurements.",,http://link.aps.org/pdf/10.1103/PhysRevSTAB.13.121004
-Experimental glomerulonephritis in the rat induced by antibodies directed against tubular antigens. V. Fixed glomerular antigens in the pathogenesis of heterologous immune complex glomerulonephritis.,B. Vandamme,"In heterologous immune complex glomerulonephritis glomerular deposition of immune complexes occurs immediately after an injection with heterologous antibody directed against antigen, derived from the brush border of the tubules. The injected antibody is thought to combine with circulating Fx1A antigen to form immune complexes which subsequently are deposited in the glomeruli. However, perfusion of rat kidneys in absence of this antigen likewise resulted in prompt localization of immune complexes along the glomerular basement membrane. Further, Fx1A antigen was shown to be present in the capillary wall, especially in the filtration slits and on the cell membrane of epithelial cells. From these findings it was concluded that in this model of glomerulonephritis the deposited immune complexes are formed locally instead of being deposited from the circulation. This concept of ""fixed antigen"" may also be relevant to the pathogenesis of other forms of experimental glomerulonephritis and probably also for human glomerulonephritis.","{'model': 'tldr@v2.0.0', 'text': 'It was concluded that in this model of glomerulonephritis the deposited immune complexes are formed locally instead of being deposited from the circulation.'}",
-Endophthalmitis Associated with Intravitreal Anti-Vascular Endothelial Growth Factor Injections,S. G. Schwartz,,"{'model': 'tldr@v2.0.0', 'text': 'Although endophthalmitis cannot be prevented in all cases, certain risk reduction strategies have been proposed, including the use of an eyelid speculum, povidone iodine, avoidance of needle contact with the eyelid margin or eyelashes, and avoidance of routine post-injection antibiotics.'}",https://link.springer.com/content/pdf/10.1007/s40135-013-0033-1.pdf
-Ischemic Neuroprotection by TRPV1 Receptor-Induced Hypothermia,Mirko Muzzi,,"{'model': 'tldr@v2.0.0', 'text': 'Data indicate that TRPV1 receptor agonists are promising candidates for hypothermic treatment of stroke and provide permanent neuroprotection in animals subjected to transient middle cerebral artery occlusion.'}",https://journals.sagepub.com/doi/pdf/10.1038/jcbfm.2012.36
-Priapism: Current Updates in Clinical Management,P. Song,"Priapism is a persistent penile erection that continues for hours beyond, or is unrelated to, sexual stimulation. Priapism requires a prompt evaluation and usually requires an emergency management. There are two types of priapism: 1) ischemic (veno-occlusive or low-flow), which is found in 95% of cases, and 2) nonischemic (arterial or high-flow). Stuttering (intermittent or recurrent) priapism is a recurrent form of ischemic priapism. To initiate appropriate management, the physician must decide whether the priapism is ischemic or nonischemic. In the management of an ischemic priapism, resolution should be achieved as promptly as possible. Initial treatment is therapeutic aspiration with or without irrigation of the corpora. If this fails, intracavernous injection of sympathomimetic agents is the next step. Surgical shunts should be performed in cases involving failure of nonsurgical treatment. The first management of a nonischemic priapism should be observation. Selective arterial embolization is recommended for the management of nonischemic priapism in cases that request treatment. The goal of management for stuttering priapism is prevention of future episodes. This article provides a review of recent clinical developments in the medical and surgical management of priapism and an investigation of scientific research activity in this rapidly developing field of study.","{'model': 'tldr@v2.0.0', 'text': 'Recent clinical developments in the medical and surgical management of priapism and an investigation of scientific research activity in this rapidly developing field of study are provided.'}",https://europepmc.org/articles/pmc3866283?pdf=render
-EIS microfluidic chips for flow immunoassay and ultrasensitive cholera toxin detection.,M. Chiriacò,"A flow-injection impedimetric immunosensor for the sensitive, direct and label-free detection of cholera toxin is reported. A limit of detection smaller than 10 pM was achieved, a value thousands of times lower than the lethal dose. The developed chips fulfil the requirement of low cost and quick reply of the assay and are expected to enable field screening, prompt diagnosis and medical intervention without the need of specialized personnel and expensive equipment, a perspective of special relevance for use in developing countries. Since the chip layout includes two sensing areas each one with a 2 × 2 sensor array, our biochips can allow statistical or (alternatively) multiplex analysis of biorecognition events between antibodies immobilized on each working electrode and different antigens flowing into the chamber.","{'model': 'tldr@v2.0.0', 'text': 'The developed chips fulfil the requirement of low cost and quick reply of the assay and are expected to enable field screening, prompt diagnosis and medical intervention without the need of specialized personnel and expensive equipment.'}",
-Global Modeling of Storm‐Time Thermospheric Dynamics and Electrodynamics,T. Fuller‐Rowell,"Understanding the neutral dynamic and electrodynamic response of the upper atmosphere to geomagnetic storms, and quantifying the balance between prompt penetration and disturbance dynamo effects, are two of the significant challenges facing us today. This paper reviews our understanding of the dynamical and electrodynamic response of the upper atmosphere to storms from a modeling perspective. After injection of momentum and energy at high latitude during a geomagnetic storm, the neutral winds begin to respond almost immediately. The high-latitude wind system evolves quickly by the action of ion drag and the injection of kinetic energy; however, Joule dissipation provides the bulk of the energy source to change the dynamics and electrodynamics globally. Impulsive energy injection at high latitudes drives large-scale gravity waves that propagate globally. The waves transmit pressure gradients initiating a change in the global circulation. Numerical simulations of the coupled thermosphere, ionosphere, plasmasphere, and electrodynamic response to storms indicate that although the wind and waves are dynamic, with significant apparent ""sloshing"" between the hemispheres, the net effect is for an increased equatorward wind. The dynamic changes during a storm provide the conduit for many of the physical processes that ensue in the upper atmosphere. For instance, the increased meridional winds at mid latitudes push plasma parallel to the magnetic field to regions of different composition. The global circulation carries molecular rich air from the lower thermosphere upward and equatorward, changing the ratio of atomic and molecular neutral species, and changing loss rates for the ionosphere. The storm wind system also drives the disturbance dynamo, which through plasma transport modifies the strength and location of the equatorial ionization anomaly peaks. On a global scale, the increased equatorward meridional winds, and the generation of zonal winds at mid latitudes via the Coriolis effects, produce a current system opposing the normal quiet-time Sq current system. At the equator, the storm-time zonal electric fields reduce or reverse the normal upward and downward plasma drift on the dayside and nightside, respectively. In the numerical simulations, on the dayside, the disturbance dynamo appears fairly uniform, whereas at night a stronger local time dependence is apparent with increased upward drift between midnight and dawn. The simulations also indicate the possibility for a rapid dynamo response at the equator, within 2 h of storm onset, before the arrival of the large-scale gravity waves. All these wind-driven processes can result in dramatic ionospheric changes during storms. The disturbance dynamo can combine and interact with the prompt penetration of magnetospheric electric fields to the equator.",,
-Stimulatory effects of ghrelin on circulating somatostatin and pancreatic polypeptide levels.,M. Arosio,"Ghrelin, the recently identified endogenous ligand of the GH secretagogue receptor, is a gut-brain peptide with endocrine, orexigenic, and gastrointestinal effects. In rodents it increases circulating gastrin and insulin levels, whereas in man it appears to decrease insulin secretion despite a rise in blood glucose levels. The aim of the present study was to evaluate the effects of ghrelin administration on total circulating somatostatin (SS), pancreatic polypeptide (PP), and gastrin levels compared with those elicited on insulin, glucose, and GH. Eight healthy volunteers of normal weight (four women and four men) were injected with 3.3 microg/kg ghrelin or saline after an overnight fast on 2 different days. Blood was taken every 15 min for 1 h and then every 30 min for 2 h. As expected, ghrelin injection elicited a prompt GH and glucose increase with a peak at 30 min and an insulin decrease with a nadir at 60 min. Gastrin concentrations were not modified, whereas significant rises were observed in both SS (in a biphasic pattern with peaks at 15 and 120 min) and PP (which increased promptly with a peak at 15 min). A significant negative correlation was found between SS (first peak) and insulin changes (r = -0.86; P < 0.01). In conclusion, this study clearly demonstrates that ghrelin stimulates SS and PP release in man. Although the underlying mechanisms and biological significance of these pharmacological effects remain to be elucidated, a causal relationship between the SS increase and the insulin changes may be hypothesized. Finally, these findings strongly support ghrelin's postulated role in linking the endocrine control of energy balance and growth with the regulation of gastrointestinal functions.","{'model': 'tldr@v2.0.0', 'text': ""Ghrelin's postulated role in linking the endocrine control of energy balance and growth with the regulation of gastrointestinal functions is supported, as this study clearly demonstrates that ghrelin stimulates SS and PP release in man.""}",https://academic.oup.com/jcem/article-pdf/88/2/701/9149104/jcem0701.pdf
-INTESTINAL OBSTRUCTION,J. V. Cooke,"Intestinal obstruction, as a rule, is associated with an increasing amount of non-coagulable nitrogen in the blood. With acute intoxication the rise in non-coagulable nitrogen may be rapid and reach as high as three or even ten times normal. With more chronic intoxication there may be little or no rise in the blood non-coagulable nitrogen. Closed intestinal loops show exactly the same picture, and, when combined with obstruction, may give very high nitrogen readings. Acute proteose intoxication due to injection of a pure proteose will show a prompt rise in blood non-coagulable nitrogen, even an increase of 100 per cent within 3 or 4 hours. These intoxications also show a high blood content of creatinine and urea. The residual or undetermined nitrogen may be very high. A human case of intestinal obstruction with autopsy presents blood findings exactly similar to those observed in many animal experiments. Clinically the non-coagulable nitrogen of the blood may give information of value in intestinal obstruction. A high reading means a grave intoxication, but a low reading may be observed in some fatal cases and gives no assurance that a fatal intoxication may not supervene. The kidneys in practically all these experiments are normal in all respects. It is possible that protein or tissue destruction rather than impaired eliminative function is responsible for the rise in non-coagulable nitrogen of the blood in these acute intoxications. Transfusions of dextrose solutions often benefit intestinal obstruction, and may depress the level of the non-coagulable nitrogen in the blood. Some cases show no change in non-coagulable nitrogen following transfusions and diuresis, and, as a rule, such cases present the most severe intoxication.","{'model': 'tldr@v2.0.0', 'text': 'A human case of intestinal obstruction with autopsy presents blood findings exactly similar to those observed in many animal experiments, and Clinically the non-coagulable nitrogen of the blood may give information of value in intestinal obstruction.'}",
-Coccygeal epidural with local anesthetic for catheterization and pain management in the treatment of feline urethral obstruction.,Angela K O'Hearn,"OBJECTIVE To describe the technique for performing a coccygeal epidural injection of local anesthetic to facilitate catheterization in male cats with urethral obstructions using low-dose sedation. SIGNIFICANCE Prompt diagnosis and relief of urethral obstructions is important as many cats may have developed marked metabolic abnormalities at the time of presentation. General anesthesia in these patients may be associated with significant risk for complications. Pain management is also an essential treatment goal, and this technique relieves urethral and penile pain during the unblocking process. CONCLUSION Coccygeal epidurals can be used safely to provide analgesia to the penis and urethra and to the authors' knowledge, is a novel treatment modality to aid in the relief of urethral obstructions in male cats.","{'model': 'tldr@v2.0.0', 'text': ""Coccygeal epidurals can be used safely to provide analgesia to the penis and urethra and to the authors' knowledge, is a novel treatment modality to aid in the relief of urethral obstructions in male cats.""}",
-Pharmacologic treatment of anaphylaxis: can the evidence base be strengthened?,F. Simons,"Purpose of reviewTo evaluate the evidence base for the pharmacologic treatment of anaphylaxis. Recent findingsIn this review, we focus on four classes of medications (the alpha/beta-agonist epinephrine (adrenaline), H1-antihistamines, H2-antihistamines, and glucocorticoids) that are used in healthcare settings for the initial treatment of anaphylaxis. Epinephrine and many H1-antihistamines and glucocorticoids were introduced before the era of randomized controlled trials and before the era of evidence-based medicine. In anaphylaxis, no randomized controlled trials that are free from methodological problems and meet current standards have been performed with these medications, or with H2-antihistamines. The evidence base for epinephrine injection is stronger than the evidence base for use of other medications in anaphylaxis. Guidelines unanimously recommend prompt injection of epinephrine as the life-saving first-line medication in anaphylaxis; however, they differ in their recommendations for H1-antihistamines, H2-antihistamines, and glucocorticoids. Epinephrine is the only medication that is universally available for anaphylaxis treatment in healthcare settings worldwide. Paradoxically, it is underused in anaphylaxis treatment. SummaryFor ethical reasons, there should never be a placebo-controlled trial of epinephrine in anaphylaxis. We discuss why the possibility of conducting randomized placebo-controlled trials with H1-antihistamines, H2-antihistamines, and particularly with glucocorticoids in anaphylaxis should be considered in order to improve the evidence base for treatment and guide clinical decision-making. We also describe the precautions that will be needed if randomized controlled trials are conducted in anaphylaxis.","{'model': 'tldr@v2.0.0', 'text': 'Why the possibility of conducting randomized placebo-controlled trials with H1-antihistamines, H2-antiharamines, and particularly with glucocorticoids in anaphylaxis should be considered in order to improve the evidence base for treatment and guide clinical decision-making is discussed.'}",
-Viral hepatitis transmission in ambulatory health care settings.,I. Williams,"In the United States, transmission of viral hepatitis from health care-related exposures is uncommon and primarily recognized in the context of outbreaks. Transmission is typically associated with unsafe injection practices, as exemplified by several recent outbreaks that occurred in ambulatory health care settings. To prevent transmission of bloodborne pathogens, health care workers must adhere to standard precautions and follow fundamental infection-control principles, including safe injection practices and appropriate aseptic techniques. These principles and practices need to be made explicit in institutional policies and reinforced through in-service education for all personnel involved in direct patient care, including those in ambulatory care settings. The effectiveness of these measures should be monitored as part of the oversight process. In addition, prompt reporting of suspected health care-related cases coupled with appropriate investigation and improved monitoring of surveillance data are needed to accurately characterize and prevent health care-related transmission of viral hepatitis.","{'model': 'tldr@v2.0.0', 'text': None}",https://academic.oup.com/cid/article-pdf/38/11/1592/5935448/38-11-1592.pdf
-Carbon-11-acetate PET imaging in renal disease.,P. Shreve,"UNLABELLED The purpose of this study was to investigate the use of [1-11C]acetate as a metabolic tracer for renal imaging in human subjects. METHODS Eighteen patients underwent dynamic PET imaging of the kidneys after intravenous bolus injection of 10-20 mCi [1-11C]acetate. Time-activity curves of renal parenchyma tracer activity were fitted to a two-compartment model using direct arterial blood sampling for the arterial input function. RESULTS Renal uptake of [1-11C]acetate is prompt and high target-to-background ratios are achieved even in the presence of markedly reduced renal function. Carbon-11-acetate is cleared from the renal parenchyma without any urinary excretion and the rate of clearance is comparable to myocardial clearance rates. Among normal subjects, K1, ranged from 0.653 to 1.37 ml/min-g, and was reduced to as low as 0.363 ml/min-g in severe renal disease (serum creatinine greater than 5 mg/dl), while k2 ranged from 0.114 to 0.166 min-1 among normal subjects and was reduced to as low as 0.053 min-1 in severe renal disease. Kinetic parameters K1 and k2 were both reduced in the presence of intrinsic renal disease or significant renal artery stenosis. Renal cell carcinoma demonstrated similar uptake of [1-11C]acetate, but substantially reduced the rate of clearance compared to normal and diseased non-neoplastic renal tissue, allowing for ready differentiation of renal cell carcinoma from non-neoplastic renal tissue on images acquired beyond 10 min of tracer administration. CONCLUSION Carbon-11-acetate is a promising physiologic tracer for the study of renal disease.","{'model': 'tldr@v2.0.0', 'text': 'Renal uptake of [1-11C]acetate is prompt and high target-to-background ratios are achieved even in the presence of markedly reduced renal function, allowing for ready differentiation of renal cell carcinoma from non-neoplastic renal tissue on images acquired beyond 10 min of tracer administration.'}",
-Prevention of lethal and renal toxicity of cis-diamminedichloroplatinum(II) by induction of metallothionein synthesis without compromising its antitumor activity in mice.,A. Naganuma,"The participation of renal metallothionein (MT) in the toxicity and antitumor activity of cis-diamminedichloroplatinum(II) (cis-DDP) in male mice was examined. Preinduction of MT in the kidney by the s.c. administration of bismuth compounds decreased the lethality and renal and gastrointestinal toxicity caused by a single s.c. injection of cis-DDP. In the present study a correlation between the protective effect of pretreatment with bismuth nitrate against cis-DDP toxicity and the preinduced MT levels in the kidney was observed. Bismuth nitrate pretreatment showed no effect on the antitumor activity of cis-DDP against several transplantable tumors, probably because it induces MT in the kidney but not in tumor tissues. The fact that p.o. preadministration of bismuth subnitrate, an antidiarrheal drug, also depressed the lethal toxicity of cis-DDP is promising for its prompt application in medical attention. Thus, bismuth pretreatment allows higher doses of cis-DDP with no apparent toxicity, resulting in more efficient utilization of this anticancer drug.","{'model': 'tldr@v2.0.0', 'text': 'A correlation between the protective effect of pretreatment with bismuth nitrate against cis-DDP toxicity and the preinduced MT levels in the kidney was observed, resulting in more efficient utilization of this anticancer drug.'}",
-Thermal and behavioral effects of lipopolysaccharide and influenza in interleukin-1 beta-deficient mice.,W. Kozak,"This study characterized body temperature (Tb), locomotor activity (Act), and feeding behavior under normal conditions and following injection with lipopolysaccharide (LPS) or inoculation with live influenza virus of transgenic C57/black mice deficient in interleukin-1 beta (IL-1 beta). Tb and Act in freely moving mice were measured by biotelemetry. Mice deficient in IL-1 beta had normal circadian rhythm of Tb but were less active than their control counterparts. Mice injected with LPS (2.5 mg/kg i.p.) responded with a prompt decrease of Tb, which lasted approximately 10 h, followed by a fever in which Tb reached a peak at approximately 24 h postinjection. There was no difference between groups in the early drop of Tb after the LPS; however, the 24-h peak of Tb was lower in IL-1 beta-deficient mice. The anorexic effects of LPS and influenza infection were similar in both groups of mice. In mice given influenza virus (17.5 plaque-forming units, median lethal dose), Tb and Act gradually decreased. The fall of Tb was smaller in the transgenic mice. The mice deficient in IL-1 beta displayed a higher mortality rate due to influenza infection than the control mice. We conclude that deficiency in IL-1 beta results in lower fever following the LPS injection and in impairment of the defense response to infection with influenza.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that deficiency in IL-1 beta results in lower fever following the LPS injection and in impairment of the defense response to infection with influenza.'}",
-GREEN OR YELLOW LASER TREATMENT FOR DIABETIC MACULAR EDEMA: Exploratory Assessment Within the Diabetic Retinopathy Clinical Research Network,S. Bressler,"Purpose: Explore differences in green compared with yellow focal/grid laser treatment on functional and anatomical endpoints in eyes with diabetic macular edema. Methods: Data from two randomized clinical trials were evaluated for differences in visual acuity and optical coherence tomography parameters for eyes assigned to sham injection + prompt laser, ranibizumab + prompt laser, or prompt laser only: among subgroups of eyes treated exclusively and electively with either green or yellow laser. Results: In the sham injection + prompt laser group, the mean visual acuity letter score change for eyes receiving green and yellow laser treatment, respectively, was +2.4 ± 14 and +5.1 ± 13 at the 52-week visit (P = 0.06) and +2.4 ± 15 and +6.0 ± 13 at the 104-week visit (P = 0.13), with no corresponding evidence of differences in optical coherence tomography thickness. When comparing wavelength groups in the ranibizumab + prompt laser and prompt laser-only groups, meaningful differences in visual acuity and optical coherence tomography thickness were not detected at 1 year or 2 years. Conclusion: A trend toward improved vision outcome with yellow laser observed in one trial was not corroborated by anatomical outcomes or by the other trial. In this study, without random assignment to different wavelengths controlling for bias and confounding, it is not possible to determine whether one wavelength is better than the other.","{'model': 'tldr@v2.0.0', 'text': 'A trend toward improved vision outcome with yellow laser observed in one trial was not corroborated by anatomical outcomes or by the other trial, and without random assignment to different wavelengths controlling for bias and confounding it is not possible to determine whether one wavelength is better than the other.'}",https://europepmc.org/articles/pmc4126070?pdf=render
-Detection of bacteremia in emergency department patients at risk for infective endocarditis using universal 16S rRNA primers in a decontaminated polymerase chain reaction assay.,R. Rothman,"Prompt definitive diagnosis of acute bacterial endocarditis in febrile injection drug users (IDUs) remains problematic because of delays associated with blood culture. Rapid detection of bacteremia by polymerase chain reaction (PCR) by use of ""universal"" primers has been hampered by background bacterial contamination. Broad-range eubacterial primers selected from the 16S rRNA gene were used in a PCR assay coupled with a simple pre-PCR decontamination step. All PCR reagents were pretreated with the restriction enzyme AluI, which has multiple digestion sites in the amplicon but none in the primer sets. When 4 different bacterial species were spiked into healthy human blood specimens, the assay identified each pathogen with an analytic sensitivity of 5 bacteria/PCR reaction. A clinical trial with 51 febrile IDUs revealed that PCR had a sensitivity and specificity of 86.7% and 86.9%, respectively, versus blood culture. Importantly, all (8/8) patients with blood culture-positive infective endocarditis were determined to be positive by PCR. This assay provides a promising diagnostic for rapid identification of bacteremia, particularly valuable in acute care settings.","{'model': 'tldr@v2.0.0', 'text': 'This assay provides a promising diagnostic for rapid identification of bacteremia, particularly valuable in acute care settings, with high sensitivity and specificity compared with blood culture.'}",https://academic.oup.com/jid/article-pdf/186/11/1677/18007218/186-11-1677.pdf
-Renal response to repetitive exposure to heme proteins: chronic injury induced by an acute insult.,Karl A. Nath,"BACKGROUND Renal diseases are conventionally classified into acute and chronic disorders. We questioned whether acute, reversible, renal insults may be induced to incite a chronic scarring process, employing as an acute insult the glycerol model of heme protein-induced renal injury. METHODS Rats were subjected to weekly injections of hypertonic glycerol for up to six months. Renal function was serially determined, and the effect of such insults on renal histology and renal expression of collagen and fibrogenic cytokines was assessed. RESULTS After the first injection of glycerol, which, expectedly, induced a prompt fall in the glomerular filtration rate (GFR), subsequent injections encountered a remarkable renal resistance in that the fall in GFR was markedly blunted. This resistance to acute decline in renal function in rats subjected to repetitive injections of glycerol was accompanied by less necrosis and apoptosis of renal tubular epithelial cells after such injections. The attenuation in the fall in GFR in response to repetitive exposure to glycerol-induced heme protein injury was maintained for up to six months. A progressive decline in GFR appeared after three months and was accompanied by histologic tubulointerstitial injury, the latter assessed at six months. These kidneys demonstrated up-regulation of collagen I, III, and IV in conjunction with increased expression of the oxidant-inducible, chemotactic cytokine, monocyte chemoattractant protein-1 (MCP-1), and the oxidant-inducible, fibrogenic cytokine, transforming growth factor-beta1 (TGF-beta1). The exposure of the kidney to a single injection of hypertonic glycerol increased the expression of both cytokines some three to five days following this exposure, while the exposure of NRK 49F cells in culture to an iron-dependent model of oxidative stress also increased expression of TGF-beta1 and collagen mRNAs. CONCLUSIONS We conclude that this nephrotoxic insult, repetitively administered, encounters a resistance in the kidney such that the expected fall in GFR does not occur. However, with time, such resistance is accompanied by a decrease in GFR, the latter associated with chronic tubulointerstitial disease. Thus, a long-term cost is exacted, either along with, or as a consequence of, such resistance. We suggest that chronic up-regulation of such oxidant-inducible genes such as TGF-beta1 and MCP-1 contributes to tubulointerstitial disease, and iron-mediated oxidative stress may directly induce TGF-beta1.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that chronic up-regulation of such oxidant-inducible genes such as TGF-beta1 and MCP-1 contributes to tubulointerstitial disease, and iron-mediated oxidative stress may directly induce TGF -beta1.'}",http://www.kidney-international.org/article/S0085253815470012/pdf
-Heparin-induced release of extracellular superoxide dismutase to human blood plasma.,K. Karlsson,"Extracellular superoxide dismutase (SOD) has previously been shown to be the major SOD isoenzyme in extracellular fluids. Upon chromatography on heparin-Sepharose it was separated into three fractions: A, without affinity; B, with intermediate affinity; and C, with relatively strong heparin affinity. Intravenous injection of heparin leads to a prompt increase in plasma extracellular-superoxide-dismutase (EC-SOD) activity. Heparin induces no release of EC-SOD from blood cells, nor does it activate EC-SOD in plasma, indicating that the source of the released enzyme is the endothelial-cell surfaces. No distinct saturation could be demonstrated in a dose-response curve up to 200 i.u. of heparin per kg body weight, showing that the releasing potency of heparin is lower for EC-SOD than for previously investigated heparin-released factors. Chromatography of human plasma on heparin-Sepharose shows nearly equal amounts of EC-SOD fractions A, B and C. Heparin induces specifically the release of fraction C. The findings point to the existence of an equilibrium of EC-SOD fraction C between the plasma phase and endothelial-cell surfaces. The major part of EC-SOD in the vasculature seems to be located on endothelial-cell surfaces.","{'model': 'tldr@v2.0.0', 'text': 'The findings point to the existence of an equilibrium of EC-SOD fraction C between the plasma phase and endothelial-cell surfaces, which seems to be located on endothelial -cell surfaces in the vasculature.'}",https://europepmc.org/articles/pmc1147663?pdf=render
-Extracellular superoxide dismutase in the vascular system of mammals.,K. Karlsson,"Plasma extracellular superoxide dismutase (EC-SOD) from the pig, cat, rabbit, guinea pig and mouse was found to be heterogeneous with regard to heparin affinity and could be separated into three fractions: A without affinity, B with weak affinity and C with relatively high affinity. Rat plasma EC-SOD was deviant and contained only A and B fractions. There were very large interspecies differences in total plasma EC-SOD activity and in division of the activity between the different fractions. Intravenous injection of heparin resulted in the pig, dog, cat, rabbit, guinea pig and mouse in a prompt increase in the plasma SOD activity. The increase was due to release of EC-SOD C to plasma, most probably from endothelial cell surfaces. In the rat, heparin induced no increase in plasma SOD activity, apparently because of the absence of EC-SOD C in this species. The relative heparin-induced increase in plasma EC-SOD C varied between 2 and 11 in the different species and was distinctly correlated with the heparin affinity of EC-SOD C in the particular species. Apparently the EC-SOD C, present in the vasculature, forms an equilibrium between plasma and endothelium, whereas EC-SOD A and B primarily exist in plasma. The wide diversity of EC-SOD in the vascular system of mammals with regard to total amount, division into fractions and distribution between plasma and endothelium indicates that the pathogenic potential of superoxide radicals in the extracellular space might vary much between species.","{'model': 'tldr@v2.0.0', 'text': 'The wide diversity of EC-SOD in the vascular system of mammals with regard to total amount, division into fractions and distribution between plasma and endothelium indicates that the pathogenic potential of superoxide radicals in the extracellular space might vary much between species.'}",
-Gas jet disruption mitigation studies on Alcator C-Mod,R. Granetz,"Damaging effects of disruptions are a major concern for Alcator C-Mod, ITER and future tokamak reactors. High-pressure noble gas jet injection is a mitigation technique which potentially satisfies the operational requirements of fast response time and reliability, while still being benign to subsequent discharges. Disruption mitigation experiments using an optimized gas jet injection system are being carried out on Alcator C-Mod to study the physics of gas jet penetration into high pressure plasmas, as well as the ability of the gas jet impurities to convert plasma energy into radiation on timescales consistent with C-Mod's fast quench times, and to reduce halo currents given C-Mod's high-current density. The dependence of impurity penetration and effectiveness on noble gas species (He, Ne, Ar, Kr) is also being studied. It is found that the high-pressure neutral gas jet does not penetrate deeply into the C-Mod plasma, and yet prompt core thermal quenches are observed on all gas jet shots. 3D MHD modelling of the disruption physics with NIMROD shows that edge cooling of the plasma triggers fast growing tearing modes which rapidly produce a stochastic region in the core of the plasma and loss of thermal energy. This may explain the apparent effectiveness of the gas jet in C-Mod despite its limited penetration. The higher-Z gases (Ne, Ar, Kr) also proved effective at reducing halo currents and decreasing thermal deposition to the divertor surfaces. In addition, noble gas jet injection proved to be benign for plasma operation with C-Mod's metal (Mo) wall, actually improving the reliability of the startup in the following discharge.",,
-Dendrimers to treat rheumatoid arthritis.,X. Bosch,"In comparison with linear polymers, dendrimers' multivalency and nanostructure confer substantial advantages in drug delivery including rapid cell entry, targetability, and easier passage across biological barriers. Previous work has shown that phosphorus-containing dendrimers capped with anionic azabisphosphonate (ABP) end groups prompt anti-inflammatory activation of human monocytes. By using two mouse models of arthritis mimicking human rheumatoid arthritis (RA), Hayder et al. recently demonstrated that intravenous injection of dendrimer ABP inhibits the secretion of proinflammatory cytokines and osteoclastogenesis--two fundamental monocyte-dependent processes of inflammation and bone erosion in RA. While available biological therapies for RA target only one of the cytokines involved in inflammation or bone erosion, dendrimer ABP, by virtue of its double action on both processes in mice, might become a more active and cost-saving alternative for RA patients. This Perspective highlights this important development and the challenges that lie ahead.","{'model': 'tldr@v2.0.0', 'text': 'While available biological therapies for RA target only one of the cytokines involved in inflammation or bone erosion, dendrimer ABP might become a more active and cost-saving alternative for RA patients.'}",
-Fall back accretion and energy injections in gamma-ray bursts,Y. Yu,"Intense flares that occur at late times relative to the prompt phase have been observed by the Swift satellite in the X-ray afterglows of gamma-ray bursts (GRBs). Here, we present a detailed analysis on the fall back accretion process to explain the intense flare phase in the very early X-ray afterglow light curves. To reproduce the afterglow at late times, we resort to the external shock by engaging energy injections. By applying our model to GRBs 080810, 081028 and 091029, we show that their X-ray afterglow light curves can be reproduced well. We then apply our model to the ultralong Swift GRB 111209A, which is the longest burst ever observed. The very early X-ray afterglow of GRB 111209A showed many interesting features, such as a significant bump observed at around 2000 s after the Swift/BAT trigger. We assume two constant energy injection processes in our model. These can explain the observed plateau at X-ray wavelength in the relatively early stage (8.0 x 10(3) s) and a second X-ray plateau and optical rebrightening at about 10(5) s. Our analysis supports the scenario that a significant amount of material may fall back towards the central engine after the prompt phase, causing an enhanced and long-lived mass accretion rate powering a Poynting-flux-dominated outflow.",,https://arxiv.org/pdf/1312.0794
-Ocular inflammation associated with antivascular endothelial growth factor treatment,H. Fine,"Purpose of review To describe the diagnosis and management of intraocular inflammation following antivascular endothelial growth factor (VEGF) injections. Recent findings Inflammation following intravitreal anti-VEGF injections can cause a dramatic reduction in acuity. Differentiating factors from truly infectious endophthalmitis include a lack of pain, redness, or hypopyon, although none of these factors is diagnostic. A high suspicion of infectious endophthalmitis should trigger a prompt vitreous tap and injection of intravitreal antibiotics. Conversely, if noninfectious endophthalmitis is suspected, close observation with frequent topical steroids is warranted. Summary Most eyes with noninfectious endophthalmitis following anti-VEGF injection recover within 1 month to baseline acuity with topical corticosteroid treatment. Mechanisms hypothesized to explain postinjection inflammation include patient-specific, delivery-specific, and medication-specific factors.","{'model': 'tldr@v2.0.0', 'text': 'Most eyes with noninfectious endophthalmitis following anti-VEGF injection recover within 1 month to baseline acuity with topical corticosteroid treatment, and Mechanisms hypothesized to explain postinjection inflammation include patient-specific, delivery- specific, and medication-specific factors.'}",
-Tungsten as target material in fusion devices,D. Naujoks,"Several experiments were conducted in ASDEX Upgrade to prove the suitability of tungsten as a divertor target material under the conditions of a high density and low temperature divertor. The observed fluxes from a tungsten tile into the plasma are low, in keeping with the extremely low sputtering yields. In addition, the very favourable effect of `prompt redeposition' (redeposition during the first gyration) could be confirmed by the experiments. Cooling of the edge region by neon injection seems permissible, i.e. neon impurity sputtering did not increase the eroded fluxes of tungsten. The transport and accumulation behaviour were investigated by means of the laser blow-off technique. No accumulation effects could be observed in ohmic discharges. In discharges with NBI heating but without ICRH, strong accumulation can occur. High heat flux tests were performed on graphite tiles coated with plasma sprayed tungsten, which withstood a thermal load of 15 MW/m2 lasting 2 s as well as 1000 cycles of 10 MW/m2 for 2 s without disabling damage. Owing to the encouraging results, an experiment using a tungsten divertor is planned in ASDEX Upgrade",,
-Enhancement of indirect sulphation of limestone by steam addition.,Michael C. Stewart,"The effect of water (H₂O(g)) on in situ SO₂ capture using limestone injection under (FBC) conditions was studied using a thermobalance and tube furnace. The indirect sulphation reaction was found to be greatly enhanced in the presence of H₂O(g). Stoichiometric conversion of samples occurred when sulphated with a synthetic flue gas containing 15% H₂O(g) in under 10 h, which is equivalent to a 45% increase in conversion as compared to sulphation without H₂O(g). Using gas pycnometry and nitrogen adsorption methods, it was shown that limestone samples sulphated in the presence of H₂O(g) undergo increased particle densification without any significant changes to pore area or volume. The microstructural changes and observed increase in conversion were attributed to enhanced solid-state diffusion in CaO/CaSO₄ in the presence of H₂O(g). Given steam has been shown to have such a strong influence on sulphation, whereas it had been previously regarded as inert, may prompt a revisiting of the classically accepted sulphation models and phenomena. These findings also suggest that steam injection may be used to enhance sulfur capture performance in fluidized beds firing low-moisture fuels such as petroleum coke.","{'model': 'tldr@v2.0.0', 'text': 'It was shown that limestone samples sulphated in the presence of H₂O(g) undergo increased particle densification without any significant changes to pore area or volume, and steam injection may be used to enhance sulfur capture performance in fluidized beds firing low-moisture fuels.'}",
-Re-186(Sn) HEDP for treatment of painful osseous metastases: initial clinical experience in 20 patients with hormone-resistant prostate cancer.,H. Maxon,"Rhenium-186(tin) hydroxyethylidene diphosphonate (HEDP) is a new radiopharmaceutical that localizes in areas of osseous metastases in a manner similar to that of standard bone-scanning agents. It also emits beta particles with sufficient energy to be therapeutically useful. A single intravenous injection of about 33 mCi (1,221 MBq) was given to each of 20 elderly patients with advanced skeletal metastases from hormonally resistant prostate cancer. Prompt, significant relief of pain occurred 80% of the time with no significant side effects and only minimal, transient marrow toxicity. Re-186(Sn) HEDP appears to be a useful new agent for the palliation of painful osseous metastases in prostate cancer.","{'model': 'tldr@v2.0.0', 'text': 'Re-186(Sn) HEDP appears to be a useful new agent for the palliation of painful osseous metastases in prostate cancer.'}",
-Vaccination against hepatitis B: comparison of three different vaccination schedules.,W. Jilg,"Three different hepatitis B vaccination schedules employing injections at months 0, 1, 2, and 12, at months 0, 1, and 6, or at months 0, 1, and 12 were compared in 89 healthy young adults. Concentrations of antibodies to hepatitis B surface antigen (anti-HBs) after the third injection were dependent on the interval between the second and the third dose; geometric mean titers (GMTs) in the three groups were 53 IU/l, 5,846 IU/l, and 19,912 IU/l, respectively, when the third dose was given at month 2, 6, or 12. Whereas the anti-HBs responses to the third dose at month 6 or 12 were typical booster reactions, the kinetics after a third dose given at month 2 resembled those after only two doses but on a significantly higher level. A fourth dose given at month 12 to the individuals vaccinated at months 0, 1, and 2 led to a prompt anti-HBs response similar in height to the response in those vaccinated at months 0, 1, and 12. Thus, for achieving a high anti-HBs concentration guaranteeing its long-lasting persistence, vaccination at months 0, 1, and 12 seems to be preferable to vaccination at months 0, 1, and 6. For individuals at high risk of hepatitis B infection, vaccination at months 0, 1, 2, and 12 might be considered for obtaining an optimal early seroconversion as well as long-term protection.","{'model': 'tldr@v2.0.0', 'text': 'For achieving a high anti-HBs concentration guaranteeing its long-lasting persistence, vaccination at weeks 0, 1, and 12 seems to be preferable to vaccination at months 0,1, and 6, which might be considered for individuals at high risk of hepatitis B infection.'}",
-Evaluation and management of priapism: 2009 update,Yun-Ching Huang,,"{'model': 'tldr@v2.0.0', 'text': 'Stuttering (recurrent) ischemic priapism is a challenging and poorly understood condition; new management strategies currently under investigation may improve the ability to care for men with this condition.'}",https://europepmc.org/articles/pmc3905796?pdf=render
-Is GeV emission from Gamma-Ray Bursts of external shock origin?,A. Maxham,"Recent observations of Gamma-Ray Bursts (GRBs) by the Fermi Large Area Telescope (LAT) revealed a power-law decay feature of the high-energy emission (above 100 MeV), which led to the suggestion that it originates from an external shock. We analyse four GRBs (080916C, 090510,090902B and 090926A) jointly detected by Fermi LAT and Gamma-ray Burst Monitor (GBM), which have high-quality light curves in both instrument energy bands. Using the MeV prompt emission (GBM) data, we can record the energy output from the central engine as a function of time. Assuming a constant radiative efficiency, we are able to track energy accumulation in the external shock using our internal/external shell model code. By solving for the early evolution of both an adiabatic and a radiative blastwave, we calculate the high-energy emission light curve in the LAT band and compare it with the observed one for each burst. The late time LAT light curves after T 90 can be well fitted by the model. However, due to continuous energy injection into the blastwave during the prompt emission phase, the early external shock emission cannot account for the observed GeV flux level. The high-energy emission during the prompt phase (before T 90 ) is most likely a superposition of a gradually enhancing external shock component and a dominant emission component that is of an internal origin.",,https://academic.oup.com/mnras/article-pdf/415/1/77/3112441/mnras0415-0077.pdf
-Long-Acting Formulations of Atypical Antipsychotics,P. Chue,,"{'model': 'tldr@v2.0.0', 'text': 'The benefits of an atypical antipsychotic coupled with the assurance of medication delivery in the form of a long-acting injection imply that these novel formulations should be considered in first-episode patients, for whom optimal outcome is frequently compromised by early treatment discontinuation and poor adherence.'}",
-Glucagon-like peptide-1 stimulates luteinizing hormone-releasing hormone secretion in a rodent hypothalamic neuronal cell line.,S. A. Beak,"To examine the influence of the putative satiety factor (GLP-1) on the hypothalamo-pituitary-gonadal axis, we used GT1-7 cells as a model of neuronal luteinizing hormone- releasing hormone (LHRH) release. GLP-1 caused a concentration-dependent increase in LHRH release from GT1-7 cells. Specific, saturable GLP-1 binding sites were demonstrated on these cells. The binding of [125I]GLP-1 was time-dependent and consistent with a single binding site (Kd = 0.07+/-0.016 nM; binding capacity = 160+/-11 fmol/mg protein). The specific GLP-1 receptor agonists, exendin-3 and exendin-4, also showed high affinity (Ki = 0.3+/-0.05 and 0.32+/-0.06 nM, respectively) as did the antagonist exendin-(9-39) (Ki = 0.98+/-0.24 nM). At concentrations that increased LHRH release, GLP-1 (0.5-10 nM) also caused an increase in intracellular cAMP in GT1-7 cells (10 nM GLP-1: 7.66+/-0.4 vs. control: 0.23+/-0.02 nmol/mg protein; P < 0.001). Intracerebroventricular injection of GLP-1 at a single concentration (10 microg) produced a prompt increase in the plasma luteinizing hormone concentration in male rats (GLP-1: 1.09+/-0.11 vs. saline: 0.69+/-0.06 ng/ml; P < 0.005). GLP-1 levels in the hypothalami of 48-h-fasted male rats showed a decrease, indicating a possible association of the satiety factor with the low luteinizing hormone levels in animals with a negative energy balance.","{'model': 'tldr@v2.0.0', 'text': 'GLP-1 levels in the hypothalami of 48-h-fasted male rats showed a decrease, indicating a possible association of the satiety factor with the low luteinizing hormone levels in animals with a negative energy balance.'}",http://www.jci.org/articles/view/610/files/pdf
-Antisolvent Crystallization of Roxithromycin and the Effect of Ultrasound,Minwoo Park,"Antisolvent crystallization was performed to precipitate roxithromycin particles from organic solutions. Roxithromycin was dissolved in acetone at different concentrations and each solution was injected into an aqueous antisolvent leading to prompt particle formation. The effects of various experimental variables (solution injection rate, solution concentration, and temperature) on the particle size of roxithromycin were investigated. In addition to these variables, the effect of ultrasound on the resulting particle size was investigated by changing process parameters such as wave intensity (power output), sonication time, and the moment of ultrasonic application. When the drug solution was rapidly injected into the antisolvent, smaller crystals were obtained. Smaller crystals were obtained when solutions with high drug concentrations were used and also when the crystallization took place at lower temperatures. The particle size decreased with the increasing power output of ultrasound and with the increasing sonication time. It was also found that the ultrasonic wave induced the reduction of the particle size only when the ultrasound was applied to the solution at the initial stage of crystallization.",,
-The dynamics of granuloma formation in experimental visceral leishmaniasis,M. J. Mcelrath,"We have examined the temporal sequence of events leading to the formation of hepatic granulomas after the intravenous injection of L. donovani amastigotes into BALB/c mice. Parasite ingestion by permissive Kupffer cells (KC) occurred promptly, and local KC aggregations were the foci about which granulomas were subsequently formed. Infected KC were recognized by the uptake of colloidal carbon and the expression of the macrophage-specific antigen recognized by F4/80 mAb. Peroxidase- positive granulocytes migrated rapidly and were followed by monocytes and L3T4+ T cells that enclosed the infected KC. Thereafter, Ly-2+ T cells were prominent members of the granulomatous lymphoid population. Parasites multiplied until 4 wk, and then a prompt reduction in infected cells occurred. This was associated with a sharp decline in the L3T4+ T cells of the granulomas and the maintenance of the Ly-2+ subset. In comparison, athymic nu/nu mice developed smaller, more slowly appearing granulomas that contained granulocytes and monocytes and exhibited progressive parasite replication. Upon rechallenge, the entire process was completed in 2 wk, and infected KC in the euthymic mice were never observed. We hypothesize that the effectiveness of the granulomatous response requires the destruction of parasitized host cells (KC), in a lymphokine rich environment. We further suggest that the Ly-2+ T cell serves as an important effector cell in this process, either by direct cytotoxicity or by supporting the cytotoxic potential of other cell types in the granuloma.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesize that the effectiveness of the granulomatous response requires the destruction of parasitized host cells (KC), in a lymphokine rich environment, and suggests that the Ly-2+ T cell serves as an important effector cell in this process, either by direct cytotoxicity or by supporting the cytotoxic potential of other cell types in thegranuloma.'}",https://rupress.org/jem/article-pdf/167/6/1927/1098217/1927.pdf
-Stimulation and desensitization of the hypothalamic heat‐sensitive structures by capsaicin in rats,A. Jancsó-Gábor,1. In rats the injection of capsaicin into the pre‐optic area of the anterior hypothalamus produces a prompt fall in body temperature and abolishes shivering. With repeated injections of capsaicin the hypothermic effect gradually diminishes and finally vanishes (local desensitization).,"{'model': 'tldr@v2.0.0', 'text': 'In rats the injection of capsaicin into the pre‐optic area of the anterior hypothalamus produces a prompt fall in body temperature and abolishes shivering, and with repeated injections the hypothermic effect gradually diminishes and finally vanishes (local desensitization).'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1113/jphysiol.1970.sp009130
-"Complications from Injectable Polyacrylamide Gel, a New Nonbiodegradable Soft Tissue Filler",Snehal P. Amin,"BACKGROUND Polyacrylamide gels, containing a hydrogel composed of polyacrylamide and water, are used for soft tissue augmentation and contour correction. There are no reports of significant complications after injection of this material into the face. OBJECTIVE We report an inflammatory reaction after injection of polyacrylamide gels for zygomatic facial augmentation. METHODS A retrospective chart review of single case is presented. RESULTS An inflammatory reaction at the sites of polyacrylamide gels injection was noted at 1 month after initial injection. Despite two ensuing courses of broad-spectrum antibiotics, the patient presented to us with persistent draining nodules. Intralesional steroid injections resulted in prompt resolution and no recurrence. CONCLUSION Inflammatory reactions have been noted in patients receiving polyacrylamide gels for breast augmentation. Facial polyacrylamide gels injections may also be associated with an inflammatory reaction that responds to intralesional steroids. With increasing availability of a variety of soft tissue fillers, dermatologists should be aware of this delayed complication from polyacrylamide gels.","{'model': 'tldr@v2.0.0', 'text': 'Patients receiving polyacrylamide gels for breast augmentation may also be associated with an inflammatory reaction that responds to intralesional steroids, and dermatologists should be aware of this delayed complication.'}",
-Quantitative aerobic and anaerobic bacteriology of infected diabetic feet,F. Sapico,"Quantitative aerobic and anaerobic cultures of deep tissue were performed on amputated infected lower limbs of 13 diabetic patients immediately after surgery. Dissection was made through intact skin distant from any preexisting ulcer. The results were compared with those obtained from: (i) ulcer swabs (pre- and postamputation), (iii) curettage of the ulcer base, and (iii) needle aspiration after normal saline injection. Anaerobic transport media were used for anaerobic cultures before prompt transfer to the anaerobic chamber. A mean of 4.7 bacterial species per specimen was seen (2.3 aerobes, 2.4 anaerobes). Mean log10 growth per gram of tissue was as follows: (i) aerobes plus anaerobes = 6.99, (ii) aerobes = 6.42, and (iii) anaerobes = 7.65. There was poor concordance between the deep tissue culture results and the results from other modalities of culture collection, though the results from other modalities of culture collection, though the results from curettage and saline aspiration were slightly better than those obtained from ulcer swabs. The most frequently isolated organisms were enterococci, anaerobic streptococci, and species of Proteus, Clostridium, and Bacteroides. When anti-microbial therapy is indicated for this patient population, the possibility of the concomitant presence of aerobes as well as anaerobes should be considered.","{'model': 'tldr@v2.0.0', 'text': 'Quantitative aerobic and anaerobic cultures of deep tissue were performed on amputated infected lower limbs of 13 diabetic patients immediately after surgery, with poor concordance between the deep tissue culture results and the results from other modalities of culture collection.'}",https://jcm.asm.org/content/jcm/12/3/413.full.pdf
-Hormonal control of gastroesophageal-sphincter strength.,D. Castell,"Abstract Because of the possibility that the esophagus is protected during periods of increased gastric acidity by more forceful contraction of the lower esophageal sphincter, pressures were measured during routine gastric analyses. After betazole hydrochloride was given, sphincter pressure consistently increased, but this rise was not due to the simple presence of acid in the stomach, for sphincter pressure decreased consistently when 0.1 N hydrochloric acid was instilled into the stomach of normal subjects. Conversely, increases in sphincter pressure were noted when alkali was placed into the stomachs of the same subjects. These results suggested a possible role of endogenous gastrin in stimulation of the lower esophageal sphincter. Prompt, marked increases in sphincter pressure after subcutaneous injection of submaximal doses of gastrin-like pentapeptide indicate that regulation of lower esophageal sphincter pressure may be a physiologic action of gastrin.","{'model': 'tldr@v2.0.0', 'text': None}",
-Blood flow to the rabbit pancreas with special reference to the islets of Langerhans.,N. Lifson,"Nonradioactive microspheres of various sized (mean batch diameters ranging from 6 microns ato 26 microns) were administered to unfasted rabbits under sodium pentobarbital anesthesia by a brief injection into the left ventricle. Flow rate per bead was determined by the reference organ method. After prompt death of the animal, the spheres were located and counted miroscopically (islet vs. nonislet) in fixed, stained, and cleared portions of the pancreas. According to an analysis of the distribution of spheres as a function of bead diameter, 11-23% of the total pancreatic blood flow went directly to the islets and 77-89% to the ""acini"" (nonislet tissues). After retrograde postmortem injection of spheres 6 microns, 9 microns and 11 microns in diameter, practically none reached the islets, whereas after orthograde postmortem injection, they did so in the same proportions as in vivo. These results, supplemented by certain control experiments, support the view that all, or nearly all, efferent islet blood flow goes to the acinar capillaries before leaving the organ. We conclude that the arterial supplies to the rabbit exocrine and endocrine pancreas are in parallel, with most of the flow going to the exocrine portion. However, the flow to the islets is large enough to permit significant local actions of the islet hormones on the exocrine pancreas, in confirmation of the existence of an insuloacinar portal system.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that the arterial supplies to the rabbit exocrine and endocrine pancreas are in parallel, with most of the flow going to the exocrine portion, however, the flow to the islets is large enough to permit significant local actions of the islet hormones on theExocrine Pancreas, in confirmation of the existence of an insuloacinar portal system.'}",
-Gamma-Hydroxybutyric Acid for Treatment of Opiate Withdrawal Syndrome,L. Gallimberti,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/npp199345.pdf
-Randomized clinical trial of testosterone replacement therapy in hypogonadal men.,A. Conway,"We have compared the pharmacokinetics and pharmacodynamics of the three commonly used testosterone formulations in a prospective, randomized cross-over clinical trial. Plasma free and total testosterone and their ratio (proportion of unbound testosterone), sex hormone-binding globulin (SHBG), oestradiol, LH and FSH were measured in 15 hypogonadal men (nine hyper- and six hypogonadotrophic) who underwent, in a randomized sequence, three treatment periods each separated by an intervening washout period. The treatments were: (i) intramuscular injection of 250 mg mixed testosterone esters at 2-weekly intervals, (ii) oral testosterone undecanoate 120 mg bd, and (iii) subcutaneous testosterone pellets (6 x 100 mg). Pellet implantation gave the most prolonged effect with free and total testosterone levels being elevated for up to 4 months. This was accompanied by prompt and sustained suppression of plasma LH and FSH, an increase in plasma levels of oestradiol but no change in SHBG levels. In contrast, intramuscular injections induced marked but reproducible week-to-week fluctuations in free and total testosterone, which resulted in a small decrease in plasma SHBG levels, less marked suppression of LH and FSH and a smaller increase in plasma levels of oestradiol. Oral testosterone undecanoate produced the most variable plasma levels of free and total testosterone with a peak in the first treatment week and a fall thereafter and, despite maintenance of testosterone levels within the physiological range, there was no significant suppression of plasma levels of LH and FSH, and oestradiol levels were unchanged but levels of SHBG and total cholesterol were decreased. Free testosterone levels were increased disproportionately during testosterone treatment as the proportion of unbound testosterone was increased by all three treatments. All three testosterone preparations lowered plasma levels of urea and all were without biochemical or haematological toxicity. Reported sexual function was better maintained and side-effects were fewer with parenteral compared with oral treatments. The marked decrease in SHBG and cholesterol levels during oral testosterone undecanoate, when compared with parenteral treatments, occurred despite lesser androgenic effects (suppression of gonadotrophin levels and reported sexual function), which suggests that the liver is exposed to excessive androgenic load via the portal vein during oral treatment with testosterone esters. It is concluded that testosterone pellets give the closest approximation to zero-order (steady-state) delivery conditions for up to 4 months after a single insertion.(ABSTRACT TRUNCATED AT 400 WORDS)","{'model': 'tldr@v2.0.0', 'text': 'Testosterone pellets give the closest approximation to zero-order delivery conditions for up to 4 months after a single insertion, which suggests that the liver is exposed to excessive androgenic load via the portal vein during oral treatment with testosterone esters.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1365-2605.1988.tb00999.x
-Effects of alpha-human atrial natriuretic peptide in essential hypertension.,M.,"Because there is little published information on the effects of atrial peptides in hypertensive humans, 100 micrograms of alpha-human atrial natriuretic peptide was injected intravenously into six patients with essential hypertension in a double-blind, placebo-controlled study under standardized conditions of body posture and dietary sodium and potassium intake. The peptide increased urine sodium excretion sixfold in the first 30 minutes. Smaller increments occurred in urine volume and in calcium, magnesium, and phosphorus excretion; the rise in urine potassium concentration was not statistically significant. Most of these indices returned to time-matched placebo values within 1 hour, but urine sodium excretion remained high for 2 1/2 hours. Arterial pressure fell within 2 minutes of alpha-human atrial natriuretic peptide injection, then returned to matching placebo levels by 10 minutes. Conversely, heart rate increased rapidly and remained elevated for 3 hours. The peptide induced a prompt, brief rise in plasma norepinephrine concentration and a more sustained fall in epinephrine and aldosterone levels, but it did not affect plasma renin activity or cortisol concentration. Compared with normotensive volunteers studied previously under the same conditions, the hypertensive subjects had a greater response in urine volume and sodium, calcium, and magnesium excretion but a less sustained fall in arterial pressure.","{'model': 'tldr@v2.0.0', 'text': 'Compared with normotensive volunteers studied previously under the same conditions, the hypertensive subjects had a greater response in urine volume and sodium, calcium, and magnesium excretion but a less sustained fall in arterial pressure.'}",https://www.ahajournals.org/doi/pdf/10.1161/01.HYP.7.5.812
-Autoimmunity caused by disruption of central T cell tolerance. A murine model of drug-induced lupus.,A. Kretz-Rommel,"A side effect of therapy with procainamide and numerous other medications is a lupus-like syndrome characterized by autoantibodies directed against denatured DNA and the (H2A-H2B)-DNA subunit of chromatin. We tested the possibility that an effect of lupus-inducing drugs on central T cell tolerance underlies these phenomena. Two intrathymic injections of procainamide-hydroxylamine (PAHA), a reactive metabolite of procainamide, resulted in prompt production of IgM antidenatured DNA antibodies in C57BL/6xDBA/2 F1 mice. Subsequently, IgG antichromatin antibodies began to appear in the serum 3 wk after the second injection and were sustained for several months. Specificity, inhibition and blocking studies demonstrated that the PAHA-induced antibodies showed remarkable specificity to the (H2A-H2B)-DNA complex. No evidence for polyclonal B cell activation could be detected based on enumeration of Ig-secreting B cells and serum Ig levels, suggesting that a clonally restricted autoimmune response was induced by intrathymic PAHA. The IgG isotype of the antichromatin antibodies indicated involvement of T cell help, and proliferative responses of splenocytes to oligonucleosomes increased up to 100-fold. As little as 5 microM PAHA led to a 10-fold T cell proliferative response to chromatin in short term organ culture of neonatal thymi. We suggest that PAHA interferes with self-tolerance mechanisms accompanying T cell maturation in the thymus, resulting in the emergence of chromatin-reactive T cells followed by humoral autoimmunity.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that PAHA interferes with self-tolerance mechanisms accompanying T cell maturation in the thymus, resulting in the emergence of chromatin-reactive T cells followed by humoral autoimmunity.'}",http://www.jci.org/articles/view/119356/files/pdf
-Uveitis following intravitreal bevacizumab: a non-infectious cluster.,Christine N. Kay,"BACKGROUND AND OBJECTIVE In this retrospective case series, the authors report seven cases of bevacizumab-related uveitis that occurred within a 4-month period. PATIENTS AND METHODS Seven eyes of six patients developed non-infectious uveitis following bevacizumab intravitreal injections in a cohort of 978 consecutive bevacizumab injections. RESULTS The mean age of patients was 74.6 years (range: 26 to 92). All patients developed symptom onset within 1 day of injection. Shared signs and symptoms included corneal edema, anterior chamber and vitreous cell, conjunctival injection, ocular pain, and lack of hypopyon. In all patients, visual acuity returned to within one line of baseline acuity. All seven eyes had been previously injected with bevacizumab, with a mean number of antecedent injections of 6.1 (range: 3 to 12). CONCLUSION A cluster of sterile bevacizumab-related uveitic reactions was described in this case series. Acute onset of symptoms, absence of hypopyon, a predominant anterior segment reaction, and prompt improvement on topical steroid therapy are useful clinical features distinguishing this uveitic syndrome from infectious endophthalmitis.","{'model': 'tldr@v2.0.0', 'text': 'Acute onset of symptoms, absence of hypopyon, a predominant anterior segment reaction, and prompt improvement on topical steroid therapy are useful clinical features distinguishing this uveitic syndrome from infectious endophthalmitis.'}",
-Selective targeting of boronophenylalanine to melanoma in BALB/c mice for neutron capture therapy.,J. Coderre,"Melanoma cells actively accumulate aromatic amino acids for use as precursors in the synthesis of the pigment melanin. Using the Harding-Passey melanoma carried s.c. in BALB/c mice, we have demonstrated that p-boronophenylalanine (BPA) is taken up by melanoma tissue to a much greater extent than by normal tissues. Following a single i.p. injection, or a series of injections given over 1 h, the accumulation of boron in melanoma was found to be transient, reaching a maximum approximately 6 h postinjection. The concentrations of boron achieved in tumor ranged from 9-33 micrograms/g, and are within the range estimated to be necessary for successful application of the nuclear reaction 10B(n,alpha)7Li for neutron capture therapy. Boron concentrations in tumor and tissues were determined using either a prompt-gamma spectroscopic technique or by quantitative neutron capture radiography using whole-body sections. Distribution studies with the resolved stereoisomers of BPA indicated that the L isomer is preferentially accumulated in the melanoma compared to the D isomer. The L isomer of BPA was shown to be targeted to actively dividing tumor cells by simultaneously comparing the boron and [3H]thymidine distribution in tumor. Under conditions which selectively deliver high concentrations of boron to Harding-Passey melanomas in BALB/c mice, BPA did not deliver useful concentrations of boron to a mammary adenocarcinoma in Hale-Stoner mice. These results, along with the selectivity of the Harding-Passey melanoma for the L isomer of BPA, are consistent with our working hypothesis that BPA is actively transported into the melanomas as an analogue of natural melanin precursors.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that p-boronophenylalanine (BPA) is taken up by melanoma tissue to a much greater extent than by normal tissues and the L isomer of BPA was shown to be targeted to actively dividing tumor cells by simultaneously comparing the boron and [3H]thymidine distribution in tumor.'}",
-Endophthalmitis following intraocular lens implantation: report of 30 cases and review of the literature.,D. Weber,"Bacterial endophthalmitis is a postoperative complication of intraocular lens implantation. A review of 30 cases showed that 77% occurred within seven days of initial cataract surgery and that all cases occurred within 32 days. Common presenting symptoms and signs included pain localized to the involved eye, decreased visual acuity, conjunctival injection, anterior chamber inflammation, hypopyon, and absent or poor red reflex. Diagnosis of an infectious etiology was made by aqueous and vitreous sampling for gram stain and culture. The commonest bacterial agents were coagulase-negative staphylococci and Staphylococcus aureus, but a wide variety of gram-negative and -positive organisms were also isolated. On the basis of this series and review of the literature, optimal treatment includes prompt diagnosis by culture and gram stain of vitreous fluid, frequent application of topical antibiotics, and administration of intraocular antibiotics. Intravenous antibiotics and vitrectomy may improve the visual outcome. Removal of the intraocular lens is unnecessary for bacteriologic cure and is not associated with improved visual outcome.",,
-"Efficacy of nerve root versus interspinous injections of glucocorticoids in the treatment of disk-related sciatica. A pilot, prospective, randomized, double-blind study.",I. Kolsi,"STUDY OBJECTIVES Pilot study comparing the short-term efficacy on pain and functional impairment of nerve root sheath versus interspinous glucocorticoid injections in patients admitted to a French rheumatology department for disk-related sciatica or femoral neuralgia. PATIENTS AND METHODS Thirty patients with refractory nerve root pain (sciatica, n = 29; femoral neuralgia, n = 1) for a mean of four months were randomized to nerve root injection (n = 17) or interspinous injection (n = 13) of the same mixture of 0.10 g of lidocaine hydrochloride and 3.75 mg of cortivazol. Both injection methods were performed under analgesia and benzodiazepine sedation to maintain double blinding. Each patient was evaluated daily during the first seven days of bed rest in the hospital, then after discharge on postinjection day 28. RESULTS Prompt pain relief was obtained in both groups. On day 1, the mean pain scale score (0-100) fell from 70 +/- 3.9 to 26 +/- 5.6 in the nerve root group and from 63 +/- 4 to 23 +/- 4.7 in the interspinous group. These results were sustained on D7 and D28. CONCLUSIONS The unusually high level of efficacy of glucocorticoid injection in our study may be ascribable in part to strong placebo and Hawthorne effects and in part to the intrinsic effects of the injections. Whether nerve root injection is superior over interspinous injection remains unproven.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study comparing the short-term efficacy on pain and functional impairment of nerve root sheath versus interspinous glucocorticoid injections in patients admitted to a French rheumatology department for disk-related sciatica or femoral neuralgia finds an unusually high level of efficacy.'}",
-A Risk-Benefit Assessment of Intra-Articular Corticosteroids in Rheumatic Disorders,J. Hunter,,"{'model': 'tldr@v2.0.0', 'text': 'Intra-articular corticosteroids form an important part of the management of inflammatory joint disease and might be considered where an inflammatory element occurs in osteoarthritis, but should be seen as an adjunct to other forms of symptom relief.'}",
-Extravasation injury in the perioperative setting.,W. Schummer,"Extravasation is an unintentional injection or leakage of fluid in the perivascular or subcutaneous space. Extravasation injury results from a combination of factors, including solution cytotoxicity, osmolality, vasoconstrictor properties, infusion pressure, regional anatomical peculiarities, and other patient factors. We reviewed the hospital files of patients who had sustained a significant extravasation injury in the perioperative setting at two German hospitals. These cases highlight the risk of devastating consequences from extravasation injury. Vasoactive drugs and hyperosmolar and concentrated electrolyte solutions are the predominant vesicants in the perioperative setting. Prompt and appropriate intervention is important for avoiding or minimizing extensive tissue injury.","{'model': 'tldr@v2.0.0', 'text': 'This work reviewed the hospital files of patients who had sustained a significant extravasation injury in the perioperative setting at two German hospitals and highlighted the risk of devastating consequences from extravasations injury.'}",
-Autonomic Activation in Experimental Trigeminal Pain,A. Frese,"Autonomic activation ipsilateral and simultaneously to unilateral trigeminal pain is a hallmark of a group of primary headache syndromes. The most common of these syndromes is cluster headache, which is characterized by strictly unilateral autonomic symptoms such as ptosis, miosis, lacrimation, conjunctival injection, rhinorrhoea, and nasal congestion during the extreme pain attack, indicating parasympathetic hyperactivity. In fact, the autonomic disturbances are so prominent that this group of headache syndromes has been coined trigeminoautonomic cephalgias (TAC) (1). However, it seems frequently overlooked that other primary headache and facial syndromes may exhibit autonomic symptoms as well (2). Most recently, autonomic activation during severe unilateral migraine attacks (3) and first division trigeminal neuralgia (4) have been described. These descriptions prompt us to report a clinical observation made during recent and ongoing studies in experimental head pain.","{'model': 'tldr@v2.0.0', 'text': 'Clinical observation made during recent and ongoing studies in experimental head pain of autonomic activation during severe unilateral migraine attacks and first division trigeminal neuralgia are reported.'}",https://journals.sagepub.com/doi/pdf/10.1046/j.1468-2982.2003.00462.x
-Prolonged biphasic response of plasma testosterone to single intramuscular injections of human chorionic gonadotropin.,RUBEN S. Padrón,"The response of plasma testosterone to varying doses of hCG (0--6000 IU) given as a single im injection has been evaluated in normal men. After an initial rise at 2 h, the levels of testosterone demonstrated a secondary rise, reaching a peak 48 h after the im injection. The magnitude of the response varied directly with the dose of hCG used, and at the highest dose (6000 IU) testosterone levels were still elevated 6 days after administration. Plasma estradiol levels showed a dose-dependent rise, with peak levels being attained 24 h after hCG. The prolonged response of plasma testosterone to a single injection of hCG should prompt a reevaluation of diagnostic and therapeutic regimens using this agent.","{'model': 'tldr@v2.0.0', 'text': 'The prolonged response of plasma testosterone to a single injection of hCG should prompt a reevaluation of diagnostic and therapeutic regimens using this agent.'}",https://academic.oup.com/jcem/article-pdf/50/6/1100/10514661/jcem1100.pdf
-Ocean physical and biogeochemical responses to the passage of Typhoon Meari in the East China Sea observed from Argo float and multiplatform satellites,E. Siswanto,"We elucidated ocean physical and biogeochemical responses to slow‐moving Typhoon Meari using a new method combining Argo float and satellite observations. Meari‐driven upwelling brought colder, nutrient‐rich deep water to the surface layer, causing sea surface cooling (3–6°C) and threefold enhancement of primary production (PP). Maximum surface cooling (and hence nutrient injection) and peak PP enhancement lagged Meari's passage by 1 and 3 days, respectively, implying that remarkable PP enhancement was attributed to new production (NP). This NP accounted for approximately 3.8% of annual carbon export in the East China Sea (ECS) outer shelf, suggesting that typhoon‐driven upwelling is important for biogeochemical processes in the ECS. Given the wide coverage of Argo float and satellite data, our new approach may prompt comparative studies in other basins and advance the understanding of the role of tropical cyclones in the global ocean biogeochemical cycle.",,https://agupubs.onlinelibrary.wiley.com/doi/pdfdirect/10.1029/2008GL035040
-"Studies on alcohol diuresis. I. The effect of ethyl alcohol ingestion on water, electrolyte and acid-base metabolism.",M. Rubini,"The diuresis following the ingestion of alcohol has been compared with the increased urine flow secondary to water ingestion (1). Van Dyke and Ames (2) demonstrated that the injection of small amounts of alcohol (12 to 50 mg. per Kg.) into the carotid artery of normal unanesthetized dogs evoked a prompt diuresis without any detectable alcohol in the systemic venous blood. No increased diuresis occurred in dogs with diabetes insipidus following intracarotid injection. As alcohol has no direct effect on exogenous antidiuretic hormone (ADH), or on the responsiveness of the renal tubules to this hormone (3), most investigators agree that ""alcohol diuresis"" is probably due to suppression of the release of ADH. Little attempt has been made, however, to correlate simultaneously its effect on water excretion with electrolyte excretion, acid-base changes, and al-terations in blood volume. The study was divided into three parts: Part I. Effect of alcohol ingestion on water, electrolyte, and acid-base metabolism in semi-recumbent hydrated normal subjects; Part II. Evaluation of the inhibition of ADH secretion by alcohol, in normal subjects in whom acute changes of extracellular fluid tonicity and ""effective circulating blood volume"" were induced; Part III. Evaluation in pathologic states in which abnormal function of the neurohypophyseal system had been demonstrated or postulated.","{'model': 'tldr@v2.0.0', 'text': 'The study was divided into three parts: Effect of alcohol ingestion on water, electrolyte, and acid-base metabolism in semi-recumbent hydrated normal subjects, and evaluation of the inhibition of ADH secretion by alcohol, in normal subjects in whom acute changes of extracellular fluid tonicity and ""effective circulating blood volume"" were induced.'}",http://www.jci.org/articles/view/103092/files/pdf
-Complications in the management of large glomus jugulae tumors,J. Cece,"This study analyzes the complications encountered in the surgical treatment of 17 patients with large glomus jugulare tumors. All 17 patients sustained either new cranial nerve palsies or exacerbation of preexisting palsies. These involved, in descending order of frequency, the facial nerve, the vagus and glossopharyngeal nerves, and the hypoglossal nerve. Post‐operative palsies of the facial nerve involved 17 patients, as compared to 5 preoperatively. Fifteen patients had postoperative partial or complete paralysis of the vagus nerve as compared to eight preoperatively. Ten patients had postoperative palsies of the hypoglossal nerve as compared to six preoperatively. Other complications included CSF leak, meningitis, and wound infection. Aspiration and dysphagia were encountered postoperatively in 13 patients. Teflonr̀ injection of paretic vocal cords and cricopharyngeal myotomy effectively improved the ability to swallow and the quality of the voice. Prompt recognition and treatment of complications is essential for effective surgical management of large glomus jugulare tumors.","{'model': 'tldr@v2.0.0', 'text': 'This study analyzes the complications encountered in the surgical treatment of 17 patients with large glomus jugulare tumors and found that teflonr̀ injection of paretic vocal cords and cricopharyngeal myotomy effectively improved the ability to swallow and the quality of the voice.'}",
-Combination of Laser Photocoagulation and Intravitreal Bevacizumab in Aggressive Posterior Retinopathy of Prematurity.,H. Altınsoy,"The response to combined laser photocoagulation and a single intravitreal injection of 0.75 mg bevacizumab to each eye on separate days in two patients with aggressive, posterior retinopathy of prematurity (ROP) is described. Combined treatment resulted in regression of zone-1 disease in Case 1, which had no retinal detachment. However, no significant regression or unfavorable anatomic response was observed in the second case with retinal detachment. Although the combination of laser photocoagulation and intravitreal bevacizumab injection seems to be well tolerated, inducing prompt regression of agressive zone-1 ROP without retinal detachment, further controlled studies with long-term follow-up are necessary for their use in the treatment of ROP with for potentially dangerous growth factor inhibitors in premature babies.","{'model': 'tldr@v2.0.0', 'text': 'Although the combination of laser photocoagulation and intravitreal bevacizumab injection seems to be well tolerated, inducing prompt regression of agressive zone-1 ROP without retinal detachment, further controlled studies with long-term follow-up are necessary for their use in the treatment of ROP with for potentially dangerous growth factor inhibitors in premature babies.'}",
-Determination of Lambda-Mode Eigenvalue Separation of a Thermal Accelerator-Driven System from Pulsed Neutron Experiment,Hiroshi Taninaka,"Basic research on the Accelerator-Driven System (ADS) with thermal neutron spectrum has been promoted by the Kyoto University Research Reactor Institute. At the Kyoto University Critical Assembly (KUCA), various experiments on thermal ADS with a pulsed spallation source are planned. In such an ADS, neutron flux distribution may be sensitive to the injection of neutrons, and the high sensitivity results in various spatial effects. In this study, a pulsed neutron experiment with 14MeV neutrons was carried out in a thermal ADS of KUCA, to determine the λ-mode eigenvalue separation, which is a quantitative indication of spatial effects. An original data-processing technique was applied to infer prompt-neutron decay constants of fundamental and higher modes from neutron count decay data, and then the eigenvalue separation around 13%Δk/k was obtained from these decay constants.",,
-Availability of amino acids supplied intravenously in healthy man as synthetic dipeptides: kinetic evaluation of L-alanyl-L-glutamine and glycyl-L-tyrosine.,S. Albers,"1. For the first time utilization of intravenously administered L-alanyl-L-glutamine and glycyl-L-tyrosine was investigated by means of their kinetic behaviour after bolus injection in 10 and 11 apparently healthy male subjects (age 26.6 +/- 5.7 years), respectively. 2. The injection of the synthetic dipeptides was not accompanied by any side effects or complaints. 3. The synthetic dipeptides L-alanyl-L-glutamine and glycyl-L-tyrosine were rapidly cleared from plasma. By applying a monoexponential model the elimination half-lives were found to have very similar values (3.8 +/- 0.5 and 3.4 +/- 0.3 min) whether alanine or glycine was occupying the N-terminal position. The estimated volume of distribution was approximately that of the extracellular space. 4. Peptide disappearance was accompanied by a prompt equimolar increase in the concentrations of the constituent amino acids alanine and glutamine as well as glycine and tyrosine. 5. The study provides firm evidence that L-alanyl-L-glutamine and glycyl-L-tyrosine are rapidly (quantitatively) hydrolysed. The results of this study may indicate a safe and efficient parenteral use of the investigated peptides as sources of free glutamine and free tyrosine.","{'model': 'tldr@v2.0.0', 'text': 'The results of this study may indicate a safe and efficient parenteral use of the investigated peptides as sources of free glutamine and free tyrosine.'}",
-Respiratory modulation of barareceptor and chemoreceptor reflexes affecting heart rate and cardiac vagal efferent nerve activity.,N. Davidson,"1. Brief stimuli were delivered to the carotid chemoreceptors or baroreceptors in dogs anaesthetized with chloralose. Chemoreceptor stimulation was achieved by rapid retrograde injection of 0.2‐0.5 ml. CO2 equilibrated saline through a cannula in the external carotid artery. Baroreceptor stimulation was achieved by forceful retrograde injection of 2‐5 ml. air‐equilibrated saline into the external carotid artery after first clamping the common carotid artery. 2. prompt decreases in heart rate were elicited by brief sudden chemoreceptor or baroreceptor stimuli when these were delivered during the expiratory phase of respiration. The stimuli did not modify the control heart rate pattern when delivered in the inspiratory phase of respiration. This respiratory modulation of reflex effectiveness persisted when the animals were completely paralysed and the phase of the respiratory cycle was monitored through a phrenic electroneurogram. 3. single cardiac vagal efferent nerve fibres were dissected from the cut central end of the right cervical vagus nerve. They were classified as cardiac efferents by their cardiac and respiratory rhythmicity, and by their increased activity in response to stimulation of a carotid sinus nerve or to mechanical elevation of the systemic arterial pressure. These efferent fibres increased their activity in response to brief chemoreceptor or baroreceptor stimuli delivered in expiration, but did not respond to stimuli delivered in inspiration. This respiratory modulation of both reflexes persisted after bilateral cervical vagotomy.",,
-Baroreceptor and chemoreceptor influences on heart rate during the respiratory cycle in the dog.,B. T. Haymet,"1. Brief stimuli were delivered to the carotid chemoreceptors or baroreceptors in dogs anaesthetized with pentobarbitone or chloralose. Chemoreceptor stimulation was achieved by rapid retrograde injections of 0‐2‐0‐5 ml. warmed, CO2‐equilibrated saline through a cannula in the external carotid artery. Baroreceptor stimulation was achieved by forceful retrograde injection of 2–5 ml. air‐equilibrated saline, or of freshly drawn arterial blood, into the external carotid artery after first clamping the common carotid artery. 2. Brief baroreceptor stimuli had no noticeable effect on breathing. Brief chemoreceptor stimuli had no effect on breathing in some dogs, but in many produced a reflex increase in the depth of inspiration when delivered during inspiration. In these same dogs, brief chemoreceptor stimuli delivered in expiration either prolonged the expiratory pause or evoked an active expiratory effort. 3. Prompt decreases in heart rate were elicited by brief sudden chemoreceptor or baroreceptor stimuli when these were delivered during the expiratory phase of respiration. The stimuli did not modify the control heart rate pattern when delivered during inspiration. If the carotid sinus nerve or the vagus nerves were cut the responses were abolished. 4. Brief chemoreceptor or baroreceptor stimuli remained effective in evoking prompt decreases in heart rate during periods of apnoea in the end‐inspiratory position (Hering‐Breuer inflation reflex). In periods of apnoea after prolonged artificial hyperventilation the stimuli were sometimes ineffective at first, but were always effective late in the period of apnoea, again producing prompt cardiac slowing. 5. After denervation of the lungs, brief baroreceptor and chemoreceptor stimuli continued to evoke prompt falls in heart rate when given during expiration. When delivered during inspiration the same stimuli were either ineffective, or less effective.","{'model': 'tldr@v2.0.0', 'text': 'Brief chemoreceptor or baroreceptor stimuli remained effective in evoking prompt decreases in heart rate during periods of apnoea in the end‐inspiratory position (Hering‐Breuer inflation reflex), but were sometimes ineffective at first, but were always effective late in the period of apNoea, again producing prompt cardiac slowing.'}",
-The pathophysiology of barium: hypokalemic and cardiovascular effects.,O. Roza,"The ingestion of absorbable barium salts, e.g. , carbonate or chloride, produces a combination of ectopic ventricular contractions, ventricular tachycardia, skeletal muscle paralysis, salivation, diarrhea, hypertension and finally, respiratory paralysis and ventricular fibrillation. Infusion of barium chloride into anesthetized dogs produced all of the above, plus a prompt and substantial hypokalemia. The data suggest that the hypokalemia is due to a transfer of potassium from extracellular to intracellular compartments rather than to urinary or gastrointestinal losses. Potassium administration prevented or reversed all of the clinical effects except the hypertension. Arterial hypertension was a constant feature of BaCI3 infusion and was not affected by simultaneous potassium infusion, by the injection of phentolamine or by bilateral nephrectomy. Barium and potassium were each found to be powerful antagonists of the other's cardiac toxicity. The intracellular accumulation of potassium induced by barium together with the reported ability of barium to stimulate myocardial adenosine triphosphatase suggests the possibility of a role for barium in the management of digitalis poisoning.","{'model': 'tldr@v2.0.0', 'text': 'The intracellular accumulation of potassium induced by barium together with the reported ability of barium to stimulate myocardial adenosine triphosphatase suggests the possibility of a role for barium in the management of digitalis poisoning.'}",
-Nonsurgical Coronary Artery Recanalization In Acute Transmural Myocardial Infarction,D. Mathey,"In 41 consecutive patients with an acute transmural myocardial infarction (AMI) admitted within 3 hours after the onset of symptoms, we tried to recanalize the occluded coronary artery by an intracoronary infusion of streptokinase (SK) (2000 units/min). SK infusion was preceded by (1) an intracoronary injection of 0.5 mg nitroglycerin to rule out coronary artery spasm, (2) an attempt to recanalize the vessel mechanically with a flexible guidewire, and (3) an intracoronary injection of plasminogen (500 units) to increase the efficacy of the subsequent SK infusion. Coronary angiography revealed a total coronary artery occlusion in 39 patients and a subtotal occlusion in two patients. In 30 patients (73%), the occluded coronary artery was successfully recanalized within 1 hour (mean 29 ± 15 minutes), resulting in prompt contrast filling of the previously occluded vessel. An arteriosclerotic stenosis always remained at the site of the occlusion. Nitroglycerin opened the occluded coronary artery in one patient, contrast injection in seven patients and guidewire perforation in four of the 15 patients, in whom it was attempted. In 18 patients the occluded coronary artery was recanalized by intracoronary SK infusion alone. After the initial opening of the occluded coronary artery, subsequent SK infusion markedly reduced the degree of stenosis and visible thrombi disappeared. Clinically, recanalization was associated with significant relief of ischemic chest pain.None of the successfully recanalized patients died, including three patients with cardiogenic shock. Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion. Repeat angiography 7–21 days later revealed a patent coronary artery in 12 of 15 successfully recanalized patients. The left ventricular ejection fraction had significantly improved, from 37 ± 5% to 47 ± 4% (mean ± SEM). Failure of recanalization in 11 of 41 patients may be explained by the absence of coronary artery thrombosis or poor SK penetration of the thrombus because of its distal location or SK runoff into nonaffected arteries.Thus, in patients with acute myocardial infarction the occluded coronary artery can be rapidly recanalized in 73% of the patients by an intracoronary infusion of streptokinase.","{'model': 'tldr@v2.0.0', 'text': 'In patients with acute myocardial infarction the occluded coronary artery can be rapidly recanalized in 73% of the patients by an intracoronary infusion of streptokinase, and success in 11 of 41 patients may be explained by the absence of coronary artery thrombosis or poor SK penetration of the thrombus because of its distal location.'}",
-Nonosmolar factors affecting renal water excretion (second of two parts).,Robert W. Schrier,"(First of Two Parts) THE pioneering work of Verney and his co-workers1 defined a very delicate mechanism that maintains the constancy of the osmotic pressure of body fluids by regulating the release of vasopressin. These investigators clearly showed that an increase in plasma osmotic pressure as small as 1 to 2 per cent by the intracarotid injection of hypertonic saline causes the release of an amount of vasopressin capable of dramatically decreasing renal water excretion.1 Conversely, the injection of a hypotonic solution into the carotid circulation has been shown to cause a prompt diuresis.2 These results have been used to . . .","{'model': 'tldr@v2.0.0', 'text': 'These investigators clearly showed that an increase in plasma osmotic pressure as small as 1 to 2 per cent by the intracarotid injection of hypertonic saline causes the release of an amount of vasopressin capable of dramatically decreasing renal water excretion.'}",
-The location of the chemoreceptor controlling gastric acid secretion during hypoglycaemia,D. Colin-Jones,"1. The injection of 2‐deoxy‐D‐glucose directly into the lateral hypothalamic area of rats, but not elsewhere, caused a prompt and sustained secretion of acid by the stomach at a rate comparable to that due to insulin hypoglycaemia.","{'model': 'tldr@v2.0.0', 'text': 'The injection of 2‐deoxy‐D‐glucose directly into the lateral hypothalamic area of rats, but not elsewhere, caused a prompt and sustained secretion of acid by the stomach at a rate comparable to that due to insulin hypoglycaemia.'}",
-The Circulatory Effects of Intravenous Phentolamine in Man,S. Taylor,"Circulatory observations have been made on six normal subjects and six hypertensive patients before and after the acute intravenous injection of 5 mg. of phentolamine.The drug caused a prompt reduction in systemic vascular resistance which resulted in a rapid fall in systemic blood pressure in spite of an increase in heart rate and cardiac output.Evidence is presented that the predominant vascular activity of the drug is to cause a direct relaxation of vascular smooth muscle, an effect far more potent than its antagonism of circulating catecholamines or its very much weaker sympathetic-blocking action. This direct depressor effect on vascular smooth muscle unaccompanied by any but slight sympathetic blocking activities endows the drug with theoretically ideal antihypertensive properties.Phentolamine caused a transient but significant reduction in oxygen uptake in the majority of individuals. It is suggested that this may be due to a direct inhibition of cellular oxidative mechanisms.An appraisal of the circulatory activities of the drug with particular regard to its use as a diagnostic test for pheochromocytoma leads to serious doubts, not only with regard to the precision with which the results of such tests can be interpreted, but also to the validity of the use of the drug as a reliable screening test for excessive circulating catecholamines.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is presented that the predominant vascular activity of the drug is to cause a direct relaxation of vascular smooth muscle, an effect far more potent than its antagonism of circulating catecholamines or its very much weaker sympathetic-blocking action.'}",https://www.ahajournals.org/doi/pdf/10.1161/01.CIR.31.5.741
-Inhibition of angiotensin conversion and prevention of renal hypertension.,Edward D. Miller,"Renal artery constriction in the unilaterally nephrectomized, trained dog, with maintained renal arterial hypotension, produces a prompt increase in systemic renin activity and blood pressure. The hypertension normally induced by renal artery stenosis is prevented by prior treatment with the nonapeptide Pyr-Trp-Pro-Arg-Pro-Gln-Ile-Pro-Pro (SQ 20, 881), which blocks conversion of angiotensin I to angiotensin II. Constant intravenous infusion of the inhibitor over several days of renal artery constriction prevents the development of chronic renovascular hypertension. Furthermore, a single injection of the nonapeptide restores blood pressure to normal in the early phase of renovascular hypertension, but becomes progressively less effective as salt and water retention occurs in the chronic stage when plasma renin activity returns to control levels. These data provide strong evidence that the renin-angiotensin system is responsible for the initiation of renovascular hypertension in the one-kidney Goldblatt dog, but that other factors become increasingly important in chronic renovascular hypertension.","{'model': 'tldr@v2.0.0', 'text': 'Data provide strong evidence that the renin-angiotensin system is responsible for the initiation of renov vascular hypertension in the one-kidney Goldblatt dog, but that other factors become increasingly important in chronic renovascular hypertension.'}",
-Hydrodynamic Collimation of Relativistic Outflows: Semianalytic Solutions and Application to Gamma-Ray Bursts,O. Bromberg,"A model is developed for the confinement and collimation of a baryon-poor outflow by its surrounding medium. Both confinement by kinetic pressure of a static corona and confinement by the ram pressure of a supersonic wind emanating from a disk surrounding the inner source are considered. Solutions are presented for the structure of the shocked layers of a deflected baryon-poor jet (BPJ) and exterior wind. The dependence of the opening angle of the BPJ on the parameters of the confining medium are carefully examined. It is found that the BPJ shock may either converge to the symmetry axis or diverge away from it, depending on the opening angle of the BPJ injection cone. In the latter case, the inner flow exhibits a nonuniform structure, consisting of an ultrarelativistic core containing the unshocked BPJ enveloped by the slower, shocked BPJ layer. The implications of our results to the prompt GRB emission are briefly discussed.",,https://arxiv.org/pdf/0705.2040
-Infectious endophthalmitis.,L. Ficker,"A clinical series of twenty-eight cases of suspected infectious endophthalmitis was reviewed. The combination of epidemiology and peroperative microscopy at the time of vitreous biopsy provided a useful indication of the pathogen involved. The visual prognosis was related to the virulence of the pathogen, and to the delay between recognition of symptoms or signs and treatment. It appears possible to predict high risk groups which warrant a high index of clinical suspicion and prompt diagnosis to enable early vitrectomy with injection of appropriate intraocular antibiotics.","{'model': 'tldr@v2.0.0', 'text': 'It appears possible to predict high risk groups which warrant a high index of clinical suspicion and prompt diagnosis to enable early vitrectomy with injection of appropriate intraocular antibiotics.'}",
-The posterior pituitary: regulation of anterior pituitary prolactin secretion.,L. Peters,"Removal of the posterior pituitary from anesthetized male rats results in a prompt and significant increase in circulating prolactin that is reversed by the injection of dopamine. Posterior pituitary extracts, which contain high concentrations of endogenous dopamine, inhibit prolactin secretion from isolated anterior pituitary cells. This inhibition is prevented by incubation of the cells with the dopamine receptor antagonist (+)-butaclamol. The data show that posterior pituitary dopamine reaches the anterior pituitary via the short hypophysial portal vessels and participates in the regulation of prolactin secretion.","{'model': 'tldr@v2.0.0', 'text': 'The data show that posterior pituitary dopamine reaches the anterior pituitsary via the short hypophysial portal vessels and participates in the regulation of prolactin secretion.'}",
-Topological Data Analysis Guided Segment Anything Model Prompt Optimization for Zero-Shot Segmentation in Biological Imaging,R. Glatt,"Emerging foundation models in machine learning are models trained on vast amounts of data that have been shown to generalize well to new tasks. Often these models can be prompted with multi-modal inputs that range from natural language descriptions over images to point clouds. In this paper, we propose topological data analysis (TDA) guided prompt optimization for the Segment Anything Model (SAM) and show preliminary results in the biological image segmentation domain. Our approach replaces the standard grid search approach that is used in the original implementation and finds point locations based on their topological significance. Our results show that the TDA optimized point cloud is much better suited for finding small objects and massively reduces computational complexity despite the extra step in scenarios which require many segmentations.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the TDA optimized point cloud is much better suited for finding small objects and massively reduces computational complexity despite the extra step in scenarios which require many segmentations.'}",http://arxiv.org/pdf/2306.17400
-Dialogue for Prompting: a Policy-Gradient-Based Discrete Prompt Optimization for Few-shot Learning,Chengzhengxu Li,"Prompt-based pre-trained language models (PLMs) paradigm have succeeded substantially in few-shot natural language processing (NLP) tasks. However, prior discrete prompt optimization methods require expert knowledge to design the base prompt set and identify high-quality prompts, which is costly, inefficient, and subjective. Meanwhile, existing continuous prompt optimization methods improve the performance by learning the ideal prompts through the gradient information of PLMs, whose high computational cost, and low readability and generalizability are often concerning. To address the research gap, we propose a Dialogue-comprised Policy-gradient-based Discrete Prompt Optimization ($DP_2O$) method. We first design a multi-round dialogue alignment strategy for readability prompt set generation based on GPT-4. Furthermore, we propose an efficient prompt screening metric to identify high-quality prompts with linear complexity. Finally, we construct a reinforcement learning (RL) framework based on policy gradients to match the prompts to inputs optimally. By training a policy network with only 0.67% of the PLM parameter size on the tasks in the few-shot setting, $DP_2O$ outperforms the state-of-the-art (SOTA) method by 1.52% in accuracy on average on four open-source datasets. Moreover, subsequent experiments also demonstrate that $DP_2O$ has good universality, robustness, and generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'A Dialogue-comprised Policy-gradient-based Discrete Prompt Optimization ($DP_2O$) method that outperforms the state-of-the-art (SOTA) method by 1.52% in accuracy on average on four open-source datasets and has good universality, robustness, and generalization ability.'}",https://arxiv.org/pdf/2308.07272
-PromptAgent: Strategic Planning with Language Models Enables Expert-level Prompt Optimization,Xinyuan Wang,"Highly effective, task-specific prompts are often heavily engineered by experts to integrate detailed instructions and domain insights based on a deep understanding of both instincts of large language models (LLMs) and the intricacies of the target task. However, automating the generation of such expert-level prompts remains elusive. Existing prompt optimization methods tend to overlook the depth of domain knowledge and struggle to efficiently explore the vast space of expert-level prompts. Addressing this, we present PromptAgent, an optimization method that autonomously crafts prompts equivalent in quality to those handcrafted by experts. At its core, PromptAgent views prompt optimization as a strategic planning problem and employs a principled planning algorithm, rooted in Monte Carlo tree search, to strategically navigate the expert-level prompt space. Inspired by human-like trial-and-error exploration, PromptAgent induces precise expert-level insights and in-depth instructions by reflecting on model errors and generating constructive error feedback. Such a novel framework allows the agent to iteratively examine intermediate prompts (states), refine them based on error feedbacks (actions), simulate future rewards, and search for high-reward paths leading to expert prompts. We apply PromptAgent to 12 tasks spanning three practical domains: BIG-Bench Hard (BBH), as well as domain-specific and general NLP tasks, showing it significantly outperforms strong Chain-of-Thought and recent prompt optimization baselines. Extensive analyses emphasize its capability to craft expert-level, detailed, and domain-insightful prompts with great efficiency and generalizability.","{'model': 'tldr@v2.0.0', 'text': 'This work presents PromptAgent, an optimization method that autonomously crafts prompts equivalent in quality to those handcrafted by experts, and applies it to 12 tasks spanning three practical domains, showing it significantly outperforms strong Chain-of-Thought and recent prompt optimization baselines.'}",
-Emotion-Conditioned Text Generation through Automatic Prompt Optimization,Yarik Menchaca Resendiz,"Conditional natural language generation methods often require either expensive fine-tuning or training a large language model from scratch. Both are unlikely to lead to good results without a substantial amount of data and computational resources. Prompt learning without changing the parameters of a large language model presents a promising alternative. It is a cost-effective approach, while still achieving competitive results. While this procedure is now established for zero- and few-shot text classification and structured prediction, it has received limited attention in conditional text generation. We present the first automatic prompt optimization approach for emotion-conditioned text generation with instruction-fine-tuned models. Our method uses an iterative optimization procedure that changes the prompt by adding, removing, or replacing tokens. As objective function, we only require a text classifier that measures the realization of the conditional variable in the generated text. We evaluate the method on emotion-conditioned text generation with a focus on event reports and compare it to manually designed prompts that also act as the seed for the optimization procedure. The optimized prompts achieve 0.75 macro-average F1 to fulfill the emotion condition in contrast to manually designed seed prompts with only 0.22 macro-average F1.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the first automatic prompt optimization approach for emotion-conditioned text generation with instruction-fine-tuned models that uses an iterative optimization procedure that changes the prompt by adding, removing, or replacing tokens.'}",https://arxiv.org/pdf/2308.04857
-MultiPrompter: Cooperative Prompt Optimization with Multi-Agent Reinforcement Learning,Dong-Ki Kim,"Recently, there has been an increasing interest in automated prompt optimization based on reinforcement learning (RL). This approach offers important advantages, such as generating interpretable prompts and being compatible with black-box foundation models. However, the substantial prompt space size poses challenges for RL-based methods, often leading to suboptimal policy convergence. This paper introduces MultiPrompter, a new framework that views prompt optimization as a cooperative game between prompters which take turns composing a prompt together. Our cooperative prompt optimization effectively reduces the problem size and helps prompters learn optimal prompts. We test our method on the text-to-image task and show its ability to generate higher-quality images than baselines.","{'model': 'tldr@v2.0.0', 'text': 'MultiPrompter, a new framework that views prompt optimization as a cooperative game between prompters which take turns composing a prompt together, is introduced, which effectively reduces the problem size and helpsprompters learn optimal prompts.'}",
-Read-only Prompt Optimization for Vision-Language Few-shot Learning,Dongjun Lee,"In recent years, prompt tuning has proven effective in adapting pre-trained vision-language models to downstream tasks. These methods aim to adapt the pre-trained models by introducing learnable prompts while keeping pre-trained weights frozen. However, learnable prompts can affect the internal representation within the self-attention module, which may negatively impact performance variance and generalization, especially in data-deficient settings. To address these issues, we propose a novel approach, Read-only Prompt Optimization (RPO). RPO leverages masked attention to prevent the internal representation shift in the pre-trained model. Further, to facilitate the optimization of RPO, the read-only prompts are initialized based on special tokens of the pre-trained model. Our extensive experiments demonstrate that RPO outperforms CLIP and CoCoOp in base-to-new generalization and domain generalization while displaying better robustness. Also, the proposed method achieves better generalization on extremely data-deficient settings, while improving parameter efficiency and computational overhead. Code is available at https://github.com/mlvlab/RPO.","{'model': 'tldr@v2.0.0', 'text': 'Read-only Prompt Optimization (RPO) leverages masked attention to prevent the internal representation shift in the pre-trained model, and achieves better generalization on extremely data-deficient settings, while improving parameter efficiency and computational overhead.'}",https://arxiv.org/pdf/2308.14960
-Robust Prompt Optimization for Large Language Models Against Distribution Shifts,Moxin Li,"Large Language Model (LLM) has demonstrated significant ability in various Natural Language Processing tasks. However, their effectiveness is highly dependent on the phrasing of the task prompt, leading to research on automatic prompt optimization using labeled task data. We reveal that these prompt optimization techniques are vulnerable to distribution shifts such as subpopulation shifts, which are common for LLMs in real-world scenarios such as customer reviews analysis. In this light, we propose a new problem of robust prompt optimization for LLMs against distribution shifts, which requires the prompt optimized over the labeled source group can simultaneously generalize to an unlabeled target group. To solve this problem, we propose Generalized Prompt Optimization framework, which incorporates the unlabeled data from the target group into prompt optimization. Extensive experimental results demonstrate the effectiveness of the proposed framework with significant performance improvement on the target group and comparable performance on the source group.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Generalized Prompt Optimization framework, which incorporates the unlabeled data from the target group into prompt optimization, and demonstrates the effectiveness of the proposed framework with significant performance improvement on thetarget group and comparable performance on the source group.'}",
-Towards an Automatic Prompt Optimization Framework for AI Image Generation,Ling Fan,,,
-Query-Dependent Prompt Evaluation and Optimization with Offline Inverse RL,Hao Sun,"In this study, we aim to enhance the arithmetic reasoning ability of Large Language Models (LLMs) through zero-shot prompt optimization. We identify a previously overlooked objective of query dependency in such optimization and elucidate two ensuing challenges that impede the successful and economical design of prompt optimization techniques. One primary issue is the absence of an effective method to evaluate prompts during inference when the golden answer is unavailable. Concurrently, learning via interactions with the LLMs to navigate the expansive natural language prompting space proves to be resource-intensive. To address this, we introduce Prompt-OIRL, which harnesses offline inverse reinforcement learning to draw insights from offline prompting demonstration data. Such data exists as by-products when diverse prompts are benchmarked on open-accessible datasets. With Prompt-OIRL, the query-dependent prompt optimization objective is achieved by first learning an offline reward model. This model can evaluate any query-prompt pairs without accessing LLMs. Subsequently, a best-of-N strategy is deployed to recommend the optimal prompt. Our experimental evaluations across various LLM scales and arithmetic reasoning datasets underscore both the efficacy and economic viability of the proposed approach.","{'model': 'tldr@v2.0.0', 'text': 'This study introduces Prompt-OIRL, which harnesses offline inverse reinforcement learning to draw insights from offline prompting demonstration data and achieves the query-dependent prompt optimization objective by first learning an offline reward model.'}",https://arxiv.org/pdf/2309.06553
-Visual-Language Prompt Tuning with Knowledge-Guided Context Optimization,Hantao Yao,"Prompt tuning is an effective way to adapt the pretrained visual-language model (VLM) to the downstream task using task-related textual tokens. Representative CoOp-based work combines the learnable textual tokens with the class tokens to obtain specific textual knowledge. However, the specific textual knowledge is worse generalization to the unseen classes because it forgets the essential general textual knowledge having a strong generalization ability. To tackle this issue, we introduce a novel Knowledge-guided Context Optimization (KgCoOp) to enhance the generalization ability of the learnable prompt for unseen classes. The key insight of KgCoOp is that the forgetting about essential knowledge can be alleviated by reducing the discrepancy between the learnable prompt and the hand-crafted prompt. Especially, KgCoOp minimizes the discrepancy between the textual embeddings generated by learned prompts and the hand-crafted prompts. Finally, adding the KgCoOp upon the contrastive loss can make a discriminative prompt for both seen and unseen tasks. Extensive evaluation of several benchmarks demonstrates that the proposed Knowledge-guided Context Optimization is an efficient method for prompt tuning, i.e., achieves better performance with less training time. code.","{'model': 'tldr@v2.0.0', 'text': 'A novel Knowledge-guided Context Optimization (KgCoOp) is introduced to enhance the generalization ability of the learnable prompt for unseen classes and achieves better performance with less training time for prompt tuning.'}",https://arxiv.org/pdf/2303.13283
-Optimization of the Signal-to-Background Ratio in Prompt Gamma Imaging Using Energy and Shifting Time-of-Flight Discrimination: Experiments With a Scanning Parallel-Slit Collimator,P. Cambraia Lopes,"Much attention is currently being paid to imaging prompt gamma (PG) rays for in vivo proton range monitoring in proton therapy. PG imaging using a collimator is affected by neutron-related background. We study the effectiveness of background reduction experimentally, using a scanning parallel-slit PG collimator as a simplified model of a multislat PG camera. The analysis is focused on the falloff region of the PG intensity profile near the Bragg peak, which is the typical region of interest for proton range estimation. Background reduction was studied for different energy windows, with and without a shifting time-of-flight window that takes into account the proton velocity within the phantom. Practical methods are put forward that apply to cyclotron-based pencil beams. The parallel-slit collimator was placed in front of arrays of cerium-doped lutetium yttrium silicate-coupled digital silicon photomultipliers, used to measure energy and time spectra together with intensity profiles of prompt events emitted from a polymethylmethacrylate phantom irradiated with a 160-MeV proton pencil beam. The best signal-to-background ratio of ~1.6 was similar to that obtained previously with a knife-edge-slit collimator. However, the slope-over-noise ratio in the PG-profile falloff region, was ~1.2 higher for the present collimator, given its better resolution.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of background reduction experimentally is studied, using a scanning parallel-slit PG collimator as a simplified model of a multislat PG camera, focused on the falloff region of the PG intensity profile near the Bragg peak, which is the typical region of interest for proton range estimation.'}",https://ieeexplore.ieee.org/ielx7/7433213/8453249/08385159.pdf
-CeBr3 scintillators for 4He prompt gamma spectroscopy: Results from a Monte Carlo optimization study,R. Dal Bello,"PURPOSE Range uncertainties limit the potential of charged particle therapy. In vivo and online range verification techniques could increase the confidence in the dose delivery distribution and lead to more conformal treatments. Prompt gamma imaging and prompt gamma spectroscopy (PGS) have been demonstrated for such a purpose. The successful application of these techniques requires the development of a dedicated detector system optimized to the radiation energy ranges and the intensity. In this work, we investigated a detector system based on CeBr3 crystals capable of performing spectroscopy of the prompt gamma radiation induced by 4 He beams. METHODS We performed Monte Carlo simulations to optimize the detector system. The study was carried out both with the Geant4 toolkit and the FLUKA package. The simulated system consisted of a primary crystal for spectroscopy and secondary crystals for noise reduction in anticoincidence (AC). For comparison purposes, we considered a configuration without AC crystals. We first defined the dimensions of the primary cerium bromide (CeBr3 ) crystal and the secondary bismuth germanate (BGO) or CeBr3 crystals. We then evaluated their detection performance for monoenergetic gamma radiation up to 7 MeV in such way that the probability of the photo-peak detection was maximized in comparison to the number of escape peak and Compton events. We simulated realistic prompt gamma radiation spectra induced by 4 He beams on homogeneous targets (water, graphite, and aluminum) and on implants (water with an aluminum insert). Finally, we tested the performances of the optimized systems in the detection of the realistic gamma spectra. The quantitative analysis was accomplished by comparing the signal-to-noise ratio between the different configurations and the ability to resolve the discrete reactions. RESULTS We present the optimized dimensions for the primary CeBr3 crystals with and without AC shielding. The specific values are given over a wide range of crystal volumes. The results show an optimal primary CeBr3 crystal with an approximately diameter to length ratio of 1 without AC shielding and 0.5 with AC shielding. The secondary BGO and CeBr3 should have a transverse dimension of 3 and 4.56 cm, respectively. The analysis of the prompt gamma spectra from 4 He beams highlighted the presence of specific discrete reactions not observed in 1 H studies, for example, 12 C transition 0+ (7.65 MeV) →2+ (4.44 MeV). This reaction is responsible for the generation of the 3.21 MeV prompt gamma peak. The optimized primary crystal provides a significant increase in the signal-to-noise ratio together with an improved resolution of the discrete gamma lines, especially in the high-energy region. The detection configuration with an optimized anticoincidence crystal improved the signal-to-noise ratio up to a factor of 3.5. CONCLUSIONS This work provides the optimal geometry for primary and secondary crystals to be used in range verification through PGS. The simulations show that such a PGS system may allow for the simultaneous detection of the discrete lines from a thin metal implant within a water phantom.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates a detector system based on CeBr3 crystals capable of performing spectroscopy of the prompt gamma radiation induced by 4 He beams and provides the optimal geometry for primary and secondary crystals to be used in range verification through PGS.'}",
-COPNER: Contrastive Learning with Prompt Guiding for Few-shot Named Entity Recognition,Yucheng Huang,"Distance metric learning has become a popular solution for few-shot Named Entity Recognition (NER). The typical setup aims to learn a similarity metric for measuring the semantic similarity between test samples and referents, where each referent represents an entity class. The effect of this setup may, however, be compromised for two reasons. First, there is typically a limited optimization exerted on the representations of entity tokens after initing by pre-trained language models. Second, the referents may be far from representing corresponding entity classes due to the label scarcity in the few-shot setting. To address these challenges, we propose a novel approach named COntrastive learning with Prompt guiding for few-shot NER (COPNER). We introduce a novel prompt composed of class-specific words to COPNER to serve as 1) supervision signals for conducting contrastive learning to optimize token representations; 2) metric referents for distance-metric inference on test samples. Experimental results demonstrate that COPNER outperforms state-of-the-art models with a significant margin in most cases. Moreover, COPNER shows great potential in the zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt composed of class-specific words to COPNER is introduced to serve as supervision signals for conducting contrastive learning to optimize token representations and metric referents for distance-metric inference on test samples.'}",
-Moderation and shielding optimization for a 252Cf based prompt gamma neutron activation analyzer system,K. Hadad,,,
-Generative Visual Prompt: Unifying Distributional Control of Pre-Trained Generative Models,Chen Henry Wu,"Generative models (e.g., GANs, diffusion models) learn the underlying data distribution in an unsupervised manner. However, many applications of interest require sampling from a particular region of the output space or sampling evenly over a range of characteristics. For efficient sampling in these scenarios, we propose Generative Visual Prompt (PromptGen), a framework for distributional control over pre-trained generative models by incorporating knowledge of other off-the-shelf models. PromptGen defines control as energy-based models (EBMs) and samples images in a feed-forward manner by approximating the EBM with invertible neural networks, avoiding optimization at inference. Our experiments demonstrate how PromptGen can efficiently sample from several unconditional generative models (e.g., StyleGAN2, StyleNeRF, diffusion autoencoder, NVAE) in a controlled or/and de-biased manner using various off-the-shelf models: (1) with the CLIP model as control, PromptGen can sample images guided by text, (2) with image classifiers as control, PromptGen can de-bias generative models across a set of attributes or attribute combinations, and (3) with inverse graphics models as control, PromptGen can sample images of the same identity in different poses. (4) Finally, PromptGen reveals that the CLIP model shows a""reporting bias""when used as control, and PromptGen can further de-bias this controlled distribution in an iterative manner. The code is available at https://github.com/ChenWu98/Generative-Visual-Prompt.","{'model': 'tldr@v2.0.0', 'text': 'Generative Visual Prompt (PromptGen), a framework for distributional control over pre-trained generative models by incorporating knowledge of other off-the-shelf models, is proposed, and it is revealed that the CLIP model shows a reporting bias when used as control, and PromptGen can further de-bias this controlled distribution in an iterative manner.'}",http://arxiv.org/pdf/2209.06970
-CPL: Counterfactual Prompt Learning for Vision and Language Models,Xuehai He,"Prompt tuning is a new few-shot transfer learning technique that only tunes the learnable prompt for pre-trained vision and language models such as CLIP. However, existing prompt tuning methods tend to learn spurious or entangled representations, which leads to poor generalization to unseen concepts.Towards non-spurious and efficient prompt learning from limited examples, this paper presents a novel Counterfactual Prompt Learning (CPL) method for vision and language models, which simultaneously employs counterfactual generation and contrastive learning in a joint optimization framework.Particularly, CPL constructs counterfactual by identifying minimal non-spurious feature change between semantically-similar positive and negative samples that causes concept change, and learns more generalizable prompt representation from both factual and counterfactual examples via contrastive learning. Extensive experiments demonstrate that CPL can obtain superior few-shot performance on different vision and language tasks than previous prompt tuning methods on CLIP. On image classification, we achieve 3.55% average relative improvement on unseen classes across seven datasets; on image-text retrieval and visual question answering, we gain up to 4.09% and 25.08% relative improvements across three few-shot scenarios on unseen test sets respectively.","{'model': 'tldr@v2.0.0', 'text': 'A novel Counterfactual Prompt Learning method for vision and language models, which simultaneously employs counterfactual generation and contrastive learning in a joint optimization framework, and can obtain superior few-shot performance on different vision andlanguage tasks than previous prompt tuning methods on CLIP.'}",http://arxiv.org/pdf/2210.10362
-Prompt engineering for zero‐shot and few‐shot defect detection and classification using a visual‐language pretrained model,Gunwoo Yong,"Zero‐shot learning, applied with vision‐language pretrained (VLP) models, is expected to be an alternative to existing deep learning models for defect detection, under insufficient dataset. However, VLP models, including contrastive language‐image pretraining (CLIP), showed fluctuated performance on prompts (inputs), resulting in research on prompt engineering—optimization of prompts for improving performance. Therefore, this study aims to identify the features of a prompt that can yield the best performance in classifying and detecting building defects using the zero‐shot and few‐shot capabilities of CLIP. The results reveal the following: (1) domain‐specific definitions are better than general definitions and images; (2) a complete sentence is better than a set of core terms; and (3) multimodal information is better than single‐modal information. The resulting detection performance using the proposed prompting method outperformed that of existing supervised models.","{'model': 'tldr@v2.0.0', 'text': 'The results reveal that domain‐specific definitions are better than general definitions and images; a complete sentence is better than a set of core terms; and multimodal information is betterthan single‐modal information.'}",
-Understanding and Mitigating Overfitting in Prompt Tuning for Vision-Language Models,Cheng Ma,"Pretrained vision-language models (VLMs) such as CLIP have shown impressive generalization capability in downstream vision tasks with appropriate text prompts. Instead of designing prompts manually, Context Optimization (CoOp) has been recently proposed to learn continuous prompts using task-specific training data. Despite the performance improvements on downstream tasks, several studies have reported that CoOp suffers from the overfitting issue in two aspects: (i) the test accuracy on base classes first improves and then worsens during training; (ii) the test accuracy on novel classes keeps decreasing. However, none of the existing studies can understand and mitigate such overfitting problems. In this study, we first explore the cause of overfitting by analyzing the gradient flow. Comparative experiments reveal that CoOp favors generalizable and spurious features in the early and later training stages, respectively, leading to the non-overfitting and overfitting phenomena. Given those observations, we propose Subspace Prompt Tuning (Sub PT) to project the gradients in back-propagation onto the low-rank subspace spanned by the early-stage gradient flow eigenvectors during the entire training process and successfully eliminate the overfitting problem. In addition, we equip CoOp with a Novel Feature Learner (NFL) to enhance the generalization ability of the learned prompts onto novel categories beyond the training set, needless of image training data. Extensive experiments on 11 classification datasets demonstrate that Sub PT+NFL consistently boost the performance of CoOp and outperform the state-of-the-art CoCoOp approach. Experiments on more challenging vision downstream tasks, including open-vocabulary object detection and zero-shot semantic segmentation, also verify the effectiveness of the proposed method. Codes can be found at https://tinyurl.com/mpe64f89.","{'model': 'tldr@v2.0.0', 'text': 'Subspace Prompt Tuning (Sub PT) is proposed to project the gradients in back-propagation onto the low-rank subspace spanned by the early-stage gradient flow eigenvectors during the entire training process and successfully eliminate the overfitting problem.'}",https://arxiv.org/pdf/2211.02219
-Optimization of a prompt gamma setup for analysis of environmental samples,A. Naqvi,,,
-Relationship between FDM 3D Printing Parameters Study: Parameter Optimization for Lower Defects,Patrich Ferretti,"Technology evolution and wide research attention on 3D printing efficiency and processes have given the prompt need to reach an understanding about each technique’s prowess to deliver superior quality levels whilst showing an economical and process viability to become mainstream. Studies in the field have struggled to predict the singularities that arise during most Fused Deposition Modeling (FDM) practices; therefore, diverse individual description of the parameters have been performed, but a relationship study between them has not yet assessed. The proposed study lays the main defects caused by a selection of printing parameters which might vary layer slicing, then influencing the defect rate. Subsequently, the chosen technique for optimization is presented, with evidence of its application viability that suggests that a quality advance would be gathered with such. The results would help in making the FDM process become a reliable process that could also be used for industry manufacturing besides prototyping purposes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed study lays the main defects caused by a selection of printing parameters which might vary layer slicing, then influencing the defect rate, to help in making the FDM process become a reliable process that could also be used for industry manufacturing besides prototyping purposes.'}",https://www.mdpi.com/2073-4360/13/13/2190/pdf?version=1625190845
-BBTv2: Pure Black-Box Optimization Can Be Comparable to Gradient Descent for Few-Shot Learning,Tianxiang Sun,"Black-Box Tuning (BBT) is a derivative-free approach to optimize continuous prompt tokens prepended to the input of language models. Although BBT has achieved comparable performance to full model tuning on simple classification tasks under few-shot settings, it requires pre-trained prompt embedding to match model tuning on hard tasks (e.g., entailment tasks), and therefore does not completely get rid of the dependence on gradients. In this paper we present BBTv2, a pure black-box optimization approach that can drive language models to achieve comparable results to gradient-based optimization. In particular, we prepend continuous prompt tokens to every layer of the language model and propose a divide-and-conquer algorithm to alternately optimize the prompt tokens at different layers. For the optimization at each layer, we perform derivative-free optimization in a low-dimensional subspace, which is then randomly projected to the original prompt parameter space. Experimental results show that BBTv2 not only outperforms BBT by a large margin, but also achieves comparable or even better performance than full model tuning and state-of-the-art parameter-efficient methods (e.g., Adapter, LoRA, BitFit, etc.) under few-shot learning settings, while maintaining much fewer tunable parameters. 1","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that BBTv2 not only outperforms BBT by a large margin, but also achieves comparable or even better performance than full model tuning and state-of-the-art parameter-efficient methods under few-shot learning settings, while maintaining much fewer tunable parameters.'}",http://arxiv.org/pdf/2205.11200
-A model for the fast evaluation of prompt losses of energetic ions in stellarators,J. Velasco,"A good understanding of the confinement of energetic ions in non-axisymmetric magnetic fields is key for the design of reactors based on the stellarator concept. In this work, we develop a model that, based on the radially-local bounce-averaged drift-kinetic equation, classifies orbits and succeeds in predicting configuration-dependent aspects of the prompt losses of energetic ions in stellarators. Such a model could in turn be employed in the optimization stage of the design of new devices.","{'model': 'tldr@v2.0.0', 'text': 'A model is developed that, based on the radially-local bounce-averaged drift-kinetic equation, classifies orbits and succeeds in predicting configuration-dependent aspects of the prompt losses of energetic ions in stellarators.'}",https://arxiv.org/pdf/2106.05697
-Healthcare scheduling in optimization context: a review,Z. A. Abdalkareem,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey on healthcare scheduling focuses on the recent literature and aims to help researchers to highlight some development from the most recent papers and grasp the new trends for future directions.'}",https://link.springer.com/content/pdf/10.1007/s12553-021-00547-5.pdf
-Connecting Large Language Models with Evolutionary Algorithms Yields Powerful Prompt Optimizers,Qingyan Guo,"Large Language Models (LLMs) excel in various tasks, but they rely on carefully crafted prompts that often demand substantial human effort. To automate this process, in this paper, we propose a novel framework for discrete prompt optimization, called EvoPrompt, which borrows the idea of evolutionary algorithms (EAs) as they exhibit good performance and fast convergence. To enable EAs to work on discrete prompts, which are natural language expressions that need to be coherent and human-readable, we connect LLMs with EAs. This approach allows us to simultaneously leverage the powerful language processing capabilities of LLMs and the efficient optimization performance of EAs. Specifically, abstaining from any gradients or parameters, EvoPrompt starts from a population of prompts and iteratively generates new prompts with LLMs based on the evolutionary operators, improving the population based on the development set. We optimize prompts for both closed- and open-source LLMs including GPT-3.5 and Alpaca, on 9 datasets spanning language understanding and generation tasks. EvoPrompt significantly outperforms human-engineered prompts and existing methods for automatic prompt generation by up to 25% and 14% respectively. Furthermore, EvoPrompt demonstrates that connecting LLMs with EAs creates synergies, which could inspire further research on the combination of LLMs and conventional algorithms.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework for discrete prompt optimization, called EvoPrompt, which borrows the idea of evolutionary algorithms (EAs) as they exhibit good performance and fast convergence and demonstrates that connecting LLMs with EAs creates synergies, which could inspire further research on the combination of LLMs and conventional algorithms.'}",https://arxiv.org/pdf/2309.08532
-Optimization of collimator designs for real-time proton range verification by measuring prompt gamma rays,P. Cambraia Lopes,"Precise monitoring of the particle range inside the body, preferably in real-time, is a primary interest in quality assurance of particle therapy. Prompt-gamma (PG) detection aims at this task, but such a system is technically challenging to implement. For mechanically collimated PG detection, neutron-induced background radiation originating in the collimator material is a major issue. We study by Monte-Carlo simulations (Geant4), the influence of different collimator designs and their geometric parameters - such as distance to source, size and septa thickness - on the correlation between transmitted PG profiles and the longitudinal depth-dose profile. The impact of background rejection by time-of-flight (TOF) is also addressed. A single-parallel-slit collimator scanned along the beam axis was primarily studied and results show that the background can be efficiently reduced by TOF discrimination. Due to reduced statistics, signal integration from neighbouring pencil-beams was further considered, accounting for typical clinical proton fluences. Furthermore, first direct-comparison results between multi-slat collimation and a knife-edge-slit collimator are presented, obtained both by simulation and analytical calculations of their geometric performances. Advantages and disadvantages of both approaches are highlighted and some design improvements are proposed. Energy thresholding was also addressed, showing that a low-energy threshold of 4 MeV increases the signal-to-background ratio in all simulation results, a factor mostly important if TOF rejection cannot be applied.",,
-Negative-prompt Inversion: Fast Image Inversion for Editing with Text-guided Diffusion Models,Daiki Miyake,"In image editing employing diffusion models, it is crucial to preserve the reconstruction quality of the original image while changing its style. Although existing methods ensure reconstruction quality through optimization, a drawback of these is the significant amount of time required for optimization. In this paper, we propose negative-prompt inversion, a method capable of achieving equivalent reconstruction solely through forward propagation without optimization, thereby enabling much faster editing processes. We experimentally demonstrate that the reconstruction quality of our method is comparable to that of existing methods, allowing for inversion at a resolution of 512 pixels and with 50 sampling steps within approximately 5 seconds, which is more than 30 times faster than null-text inversion. Reduction of the computation time by the proposed method further allows us to use a larger number of sampling steps in diffusion models to improve the reconstruction quality with a moderate increase in computation time.",,http://arxiv.org/pdf/2305.16807
-pFedPrompt: Learning Personalized Prompt for Vision-Language Models in Federated Learning,Tao Guo,"Pre-trained vision-language models like CLIP show great potential in learning representations that capture latent characteristics of users. A recently proposed method called Contextual Optimization (CoOp) introduces the concept of training prompt for adapting pre-trained vision-language models. Given the lightweight nature of this method, researchers have migrated the paradigm from centralized to decentralized system to innovate the collaborative training framework of Federated Learning (FL). However, current prompt training in FL mainly focuses on modeling user consensus and lacks the adaptation to user characteristics, leaving the personalization of prompt largely under-explored. Researches over the past few years have applied personalized FL (pFL) approaches to customizing models for heterogeneous users. Unfortunately, we find that with the variation of modality and training behavior, directly applying the pFL methods to prompt training leads to insufficient personalization and performance. To bridge the gap, we present pFedPrompt, which leverages the unique advantage of multimodality in vision-language models by learning user consensus from linguistic space and adapting to user characteristics in visual space in a non-parametric manner. Through this dual collaboration, the learned prompt will be fully personalized and aligned to the user’s local characteristics. We conduct extensive experiments across various datasets under the FL setting with statistical heterogeneity. The results demonstrate the superiority of our pFedPrompt against the alternative approaches with robust performance.","{'model': 'tldr@v2.0.0', 'text': None}",
-Design optimization of a 2D prompt-gamma measurement system for proton dose verification,H. Lee,"To verify in-vivo proton dose distribution, a 2-dimensional (2D) prompt-gamma measurement system, comprised of a multi-hole collimation system, a 2D array of CsI(Tl) scintillators, and a position-sensitive photomultiplier tube (PS-PMT), is under development. In the present study, to determine the optimal dimension of the measurement system, we employed a series of Monte Carlo simulations with the MCNPX code. To effectively measure the high-energy prompt gammas while minimizing background gammas, we determined the collimator hole size, collimator thickness, and scintillator length to be 0.4 × 0.4 cm2, 15 cm, and 5 cm, respectively. Thereafter, the performance of the optimized measurement system was estimated for monoenergetic proton pencil beams. The peak locations of the prompt-gamma distributions for 80- and 150-MeV proton beams were clearly distinguished, and the correlation between the beam range and the peak location was confirmed by using the measurement system. For a 200-MeV proton beam, however, the peak location could not be determined due to the dominance of background gammas and the lateral dispersion of the proton beam at the end of the beam range. Based on these simulation results, a prototype 2D prompt-gamma measurement system currently is under construction and, upon completion, will be tested with therapeutic proton beams.",,
-Iterative Prompt Learning for Unsupervised Backlit Image Enhancement,Zhexin Liang,"We propose a novel unsupervised backlit image enhancement method, abbreviated as CLIP-LIT, by exploring the potential of Contrastive Language-Image Pre-Training (CLIP) for pixel-level image enhancement. We show that the open-world CLIP prior not only aids in distinguishing between backlit and well-lit images, but also in perceiving heterogeneous regions with different luminance, facilitating the optimization of the enhancement network. Unlike high-level and image manipulation tasks, directly applying CLIP to enhancement tasks is non-trivial, owing to the difficulty in finding accurate prompts. To solve this issue, we devise a prompt learning framework that first learns an initial prompt pair by constraining the text-image similarity between the prompt (negative/positive sample) and the corresponding image (backlit image/well-lit image) in the CLIP latent space. Then, we train the enhancement network based on the text-image similarity between the enhanced result and the initial prompt pair. To further improve the accuracy of the initial prompt pair, we iteratively fine-tune the prompt learning framework to reduce the distribution gaps between the backlit images, enhanced results, and well-lit images via rank learning, boosting the enhancement performance. Our method alternates between updating the prompt learning framework and enhancement network until visually pleasing results are achieved. Extensive experiments demonstrate that our method outperforms state-of-the-art methods in terms of visual quality and generalization ability, without requiring any paired data.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel unsupervised backlit image enhancement method that outperforms state-of-the-art methods in terms of visual quality and generalization ability, and iteratively fine-tune the prompt learning framework to reduce the distribution gaps between the backlit images, enhanced results, and well-lit images via rank learning, boosting the enhancement performance.'}",https://arxiv.org/pdf/2303.17569
-Meta Learning for Domain Agnostic Soft Prompt,Ming-Yen Chen,"The prompt-based learning, as used in GPT-3, has become a popular approach to extract knowledge from a powerful pre-trained language model (PLM) for natural language understanding tasks. However, either applying the hard prompt for sentences by defining a collection of human-engineering prompt templates or directly optimizing the soft or continuous prompt with labeled data may not really generalize well for unseen domain data. To cope with this issue, this paper presents a new prompt-based unsupervised domain adaptation where the learned soft prompt is able to boost the frozen pre-trained language model to deal with the input tokens from unseen domains. Importantly, the meta learning and optimization is developed to carry out the domain agnostic soft prompt where the loss for masked language model is minimized. The experiments on multi-domain natural language understanding tasks show the merits of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new prompt-based unsupervised domain adaptation where the learned soft prompt is able to boost the frozen pre-trained language model to deal with the input tokens from unseen domains.'}",
-Ant Colony Optimization for the Control of Pollutant Spreading on Social Networks,Wei-neng Chen,"The rapid development of online social networks not only enables prompt and convenient dissemination of desirable information but also incurs fast and wide propagation of undesirable information. A common way to control the spread of pollutants is to block some nodes, but such a strategy may affect the service quality of a social network and leads to a high control cost if too many nodes are blocked. This paper considers the node selection problem as a biobjective optimization problem to find a subset of nodes to be blocked so that the effect of the control is maximized while the cost of the control is minimized. To solve this problem, we design an ant colony optimization algorithm with an adaptive dimension size selection under the multiobjective evolutionary algorithm framework based on decomposition (MOEA/D-ADACO). The proposed algorithm divides the biobjective problem into a set of single-objective subproblems and each ant takes charge of optimizing one subproblem. Moreover, two types of pheromone and heuristic information are incorporated into MOEA/D-ADACO, that is, pheromone and heuristic information of dimension size selection and that of node selection. While constructing solutions, the ants first determine the dimension size according to the former type of pheromone and heuristic information. Then, the ants select a specific number of nodes to build solutions according to the latter type of pheromone and heuristic information. Experiments conducted on a set of real-world online social networks confirm that the proposed biobjective optimization model and the developed MOEA/D-ADACO are promising for the pollutant spreading control.","{'model': 'tldr@v2.0.0', 'text': 'Experiments conducted on a set of real-world online social networks confirm that the proposed biobjective optimization model and the developed MOEA/D-ADACO are promising for the pollutant spreading control.'}",
-Which Components of a Smartphone Walking App Help Users to Reach Personalized Step Goals? Results From an Optimization Trial,Jan-Niklas Kramer,"Abstract Background The Assistant to Lift your Level of activitY (Ally) app is a smartphone application that combines financial incentives with chatbot-guided interventions to encourage users to reach personalized daily step goals. Purpose To evaluate the effects of incentives, weekly planning, and daily self-monitoring prompts that were used as intervention components as part of the Ally app. Methods We conducted an 8 week optimization trial with n = 274 insurees of a health insurance company in Switzerland. At baseline, participants were randomized to different incentive conditions (cash incentives vs. charity incentives vs. no incentives). Over the course of the study, participants were randomized weekly to different planning conditions (action planning vs. coping planning vs. no planning) and daily to receiving or not receiving a self-monitoring prompt. Primary outcome was the achievement of personalized daily step goals. Results Study participants were more active and healthier than the general Swiss population. Daily cash incentives increased step-goal achievement by 8.1%, 95% confidence interval (CI): [2.1, 14.1] and, only in the no-incentive control group, action planning increased step-goal achievement by 5.8%, 95% CI: [1.2, 10.4]. Charity incentives, self-monitoring prompts, and coping planning did not affect physical activity. Engagement with planning interventions and self-monitoring prompts was low and 30% of participants stopped using the app over the course of the study. Conclusions Daily cash incentives increased physical activity in the short term. Planning interventions and self-monitoring prompts require revision before they can be included in future versions of the app. Selection effects and engagement can be important challenges for physical-activity apps. Clinical Trial Information This study was registered on ClinicalTrials.gov, NCT03384550.","{'model': 'tldr@v2.0.0', 'text': 'Daily cash incentives increased physical activity in the short term and planning interventions and self-monitoring prompts require revision before they can be included in future versions of the Ally app.'}",https://academic.oup.com/abm/article-pdf/54/7/518/33381379/kaaa002.pdf
-Prompt Report on Exa-Scale HPL-AI Benchmark,Shuhei Kudo,"Our performance benchmark of HPL-AI on the supercomputer Fugaku was awarded in the 55th top500 at ISC20. The effective performance was 1.42 EFlop/s, and the world's first achievement to exceed the wall of exascale in a floating-point arithmetic benchmark. Due to the novelty of HPL-AI, there are few guidelines for large systems and several drawbacks to the large-scale benchmark. It is not enough to replace FP64 operations solely to those on FP32 or FP16. At the least, we need thoughtful numerical analysis for lower-precision arithmetic and introduction of optimization techniques on extensive computing such as on Fugaku. In the poster, we give some comments on the accuracy, implementation, performance improvement, and report on the Exa-scale benchmark on Fugaku.","{'model': 'tldr@v2.0.0', 'text': 'Comments are given on the accuracy, implementation, performance improvement, and report on the Exa-scale benchmark on Fugaku.'}",
-Temporally-Extended Prompts Optimization for SAM in Interactive Medical Image Segmentation,Chuyun Shen,"The Segmentation Anything Model (SAM) has recently emerged as a foundation model for addressing image segmentation. Owing to the intrinsic complexity of medical images and the high annotation cost, the medical image segmentation (MIS) community has been encouraged to investigate SAM's zero-shot capabilities to facilitate automatic annotation. Inspired by the extraordinary accomplishments of interactive medical image segmentation (IMIS) paradigm, this paper focuses on assessing the potential of SAM's zero-shot capabilities within the IMIS paradigm to amplify its benefits in the MIS domain. Regrettably, we observe that SAM's vulnerability to prompt forms (e.g., points, bounding boxes) becomes notably pronounced in IMIS. This leads us to develop a framework that adaptively offers suitable prompt forms for human experts. We refer to the framework above as temporally-extended prompts optimization (TEPO) and model it as a Markov decision process, solvable through reinforcement learning. Numerical experiments on the standardized benchmark BraTS2020 demonstrate that the learned TEPO agent can further enhance SAM's zero-shot capability in the MIS context.","{'model': 'tldr@v2.0.0', 'text': ""This paper focuses on assessing the potential of SAM's zero-shot capabilities within the IMIS paradigm to amplify its benefits in the MIS domain, and develops a framework that adaptively offers suitable prompt forms for human experts.""}",http://arxiv.org/pdf/2306.08958
-Optimization technique for a Prompt Gamma-ray SPECT collimator system.,Masayori Ishikawa,"Because background radiation in an irradiation room creates a problem with the PG-SPECT (Prompt Gamma-ray Single Photon Emission Computed Tomography) system, which evaluates the absorbed dose for the Boron Neutron Capture Therapy treatment, optimization of a collimator system was performed while taking the shielding of background gamma-rays into consideration. Assuming that a parallel-beam collimator is used, three parameters--the diameter of a hole, the length of the collimator, and the number of detectors (the number of holes of the collimator)--were selected for optimization. Because the combinations of these parameters are limitless, it is difficult to determine them simultaneously. Therefore, a statistically derived Optimization Criterion has been proposed to optimize these parameters. When the spatial resolution was 1 cm-FWHM (full width at half maximum), the optimal diameter of the collimator was 5.4 mm, the optimal length was 321 mm, and the optimal number of detectors was 31 x 31.","{'model': 'tldr@v2.0.0', 'text': 'Because background radiation in an irradiation room creates a problem with the PG-SPECT system, which evaluates the absorbed dose for the Boron Neutron Capture Therapy treatment, optimization of a collimator system was performed while taking the shielding of background gamma-rays into consideration.'}",https://academic.oup.com/jrr/article-pdf/42/4/387/2722132/jrr-42-387.pdf
-Quality-Aware Energy Optimization in Wireless Video Communication With Multipath TCP,Jiyan Wu,"The advancements in wireless communication technologies prompt the bandwidth aggregation for mobile video delivery over heterogeneous access networks. Multipath TCP (MPTCP) is the transport protocol recommended by IETF for concurrent data transmission to multihomed terminals. However, it still remains challenging to deliver user-satisfied video services with the existing MPTCP schemes because of the contradiction between energy consumption and received video quality in mobile devices. To enable the energy-efficient and quality-guaranteed video streaming, this paper presents an energy-distortion-aware MPTCP (EDAM) solution. First, we develop an analytical framework to characterize the energy-distortion tradeoff for multipath video transmission over heterogeneous wireless networks. Second, we propose a video flow rate allocation algorithm to minimize the energy consumption while achieving target video quality based on utility maximization theory. The performance of the proposed EDAM is evaluated through both experiments in real wireless networks and extensive emulations in exata. Experimental results show that EDAM exhibits performance advantages over existing MPTCP schemes in energy conservation and video quality.","{'model': 'tldr@v2.0.0', 'text': 'An analytical framework to characterize the energy-distortion tradeoff for multipath video transmission over heterogeneous wireless networks is developed and a video flow rate allocation algorithm to minimize the energy consumption while achieving target video quality based on utility maximization theory is proposed.'}",
-Optimization of mass cytometry sample cryopreservation after staining,Hermi Sumatoh,"The advent of mass cytometry has facilitated highly multi‐parametric single‐cell analysis allowing for the deep assessment of cellular diversity. While the data and analytical power of this approach are well described, associated technical and experimental hurdles remain. Issues like equipment breakdown and sampling of large‐scale batches, which may require multiple days of data acquisition, are minor but critical obstacles that prompt a technical solution, especially when dealing with precious samples. An ability to cryopreserve mass cytometry samples that have already been stained would alleviate numerous technical limitations we face with currently used sample‐handling approaches. Here, we evaluated two protocols for freezing of already‐stained and fixed cellular samples and compared them with standard sample refrigeration in staining buffer. A comprehensive human T cell staining phenotypic and functional profiling panel was used and the signal intensity and reliability of each marker was assessed over a 4‐week period. In general, cellular viability, DNA Ir‐Intercalator and barcode staining were minimally affected by freezing compared to refrigeration, and the signal intensities for cell surface markers and receptors were not compromised. Intracellular cytokine staining did show some decreases in signal intensity after freezing, with the decreases more prominent in a methanol‐based protocol compared to a protocol involving the use of 10% DMSO in FBS. We conclude that freezing already‐stained samples suspended in 10% DMSO in FBS is practical and efficient way to preserve already‐stained samples when needed. © 2016 International Society for Advancement of Cytometry","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that freezing already‐stained samples suspended in 10% DMSO in FBS is practical and efficient way to preserve already-stained samples when needed and alleviate numerous technical limitations with currently used sample‐handling approaches.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/cyto.a.23014
-Integrated Well Placement and Completion Optimization using Heuristic Algorithms: A Case Study of an Iranian Carbonate Formation,Reza Khoshneshin,"Determination of optimum location for drilling a new well not only requires engineering judgments but also consumes excessive computational time. Additionally, availability of many physical constraints such as the well length, trajectory, and completion type and the numerous affecting parameters including, well type, well numbers, well-control variables prompt that the optimization approaches become imperative;. The aim of this study is to figure out optimum well location and the best completion condition using coupled simulation optimization on an Iranian oil field located in southwest of Iran. The well placement scenarios are considered in two successive time intervals during of the field life, i.e., exploration and infill drilling phase. In the former scenario, the well-placement optimization is considered to locate the drilling site of a wildcat well, while the later scenario includes the optimum drilling location of a well is determined after 10-years primary production of nine production wells. In each scenario, two stochastic optimization algorithms namely particle swarm optimization, and artificial bee colony will be applied to evaluate the considered objective function. The net present value to drill production wells through the field life is considered as an objective function during our simulation-optimization approach. Our results show that the outcome of two population-based algorithms (i.e., particle swarm optimization and artificial bee colony) is marginally different from each other. The net present value of the infill drilling phase attains higher value using artificial bee colony algorithm.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the outcome of two population-based algorithms (i.e., particle swarm optimization and artificial bee colony) is marginally different from each other, and the net present value of the infill drilling phase attains higher value using artificial bee colonies algorithm.'}",
-Robust Optimization Model for Backup Resource Allocation in Cloud Provider,Fujun He,"This paper proposes a backup resource allocation model that provides a probabilistic protection for primary physical machines in a cloud provider to minimize the required total capacity. When any random failure occurs, workloads are transferred to preplanned and dedicated backup physical machines for prompt recovery. In the proposed model, a probabilistic protection guarantee is introduced to prevent the cloud provider from capacity overbooking. We apply robust optimization in our model to formulate the backup resource allocation problem as an integer linear programming problem. A simulated annealing heuristic is adopted to solve the same optimization problem when the cloud provider is large. Finally, the results reveal that the required backup capacity depends on the reliability of primary physical machines. Specifically, the more the resources in primary physical machines share backup capacity when the failure probabilities of primary physical machines are sufficiently small, the less capacity is required for backup resource allocation.","{'model': 'tldr@v2.0.0', 'text': 'The more the resources inPrimary physical machines share backup capacity when the failure probabilities of primary physical machines are sufficiently small, the less capacity is required for backup resource allocation.'}",
-Data Mining Algorithms for Classification of Diagnostic Cancer Using Genetic Optimization Algorithms,R. A. Khan,"The breast tumor is the primary driver of female casualty everywhere throughout the world and the real area of study from a long time but with slighter development than anticipated. Numerous establishments and associations are working in this field to prompt to a conceivable arrangement of the issue or to prompt to additionally comprehension of the issue. Numerous past inquiries about the said were contemplated for improved comprehension of the issue and the research performed previously was to reduce dimensionality and to contribute to the betterment in the field of cancer, Wisconsin-Madison Diagnostic Breast cancer (WDBC) dataset was taken from learning repository of UCI database with 569 distinct instances for training by choosing finest features out of 32 different attributes. Different feature selection algorithms were used with data mining algorithms for better classification. Numerous enhancements in classification accuracy of WDBC were discovered by utilizing distinctive methodologies than the prior reviews directed in a similar field. The Logistic Regression, Linear Regression, and SVM algorithms showed better classification accuracy i.e. 98.24 %, 98.24 % and 98.07 % than the previous outcome results known for the said classification algorithms. The results were generated using 10 fold cross validation, by using different classification algorithms with feature selection and generation algorithms. General Terms Machine Learning, Data Mining, Classification, Genetic algorithms, Feature Selection, Algorithms, and Cancer.","{'model': 'tldr@v2.0.0', 'text': 'Improvements in classification accuracy of WDBC were discovered by utilizing distinctive methodologies than the prior reviews directed in a similar field.'}",
-Prompt decision method for ground-state searches of natural computing architecture using 2D ising spin model,M. Ito,"Recently, the ability to analyze big data has been required for the optimization of social systems and the development of artificial intelligence. Consequently, the solution of combinatorial optimization problems has become important in recent years, especially for the suitable operation of infrastructures. However, combinatorial optimization problems have unique properties that the number of candidate solutions increases explosively as the number of parameters is increased. One possible resolution of this issue involves artificial and/or simulated Ising spin system. In this work, we have implemented such a spin system using “prompt decision logic”. The convergence operation was successfully observed in prompt decision method for spin interaction. Therefore, it is indicated that Ising computing by prompt decision logic could resolve combinatorial optimization problems.","{'model': 'tldr@v2.0.0', 'text': 'It is indicated that Ising computing by prompt decision logic could resolve combinatorial optimization problems.'}",
-The ANCIENT CHARM project at FRM II: three-dimensional elemental mapping by prompt gamma activation imaging and neutron tomography,R. Schulze,"The aim of the EU FP6-funded ANCIENT CHARM project has been the development of new, and the integration of existing element-sensitive imaging methods utilizing neutrons. We present here the methodology and the first implementation of 3D elemental mapping at the Prompt Gamma Activation Analysis instrument of the research reactor FRM II, in order to determine spatially resolved elemental abundances in samples. After the design, optimization, and characterization of the new setup, measurements were successfully completed on archaeological objects of historical significance from the collection of the Hungarian National Museum.",,
-Neck of femur fractures in the over 90s: a select group of patients who require prompt surgical intervention for optimal results,Kamal S Hapuarachchi,,"{'model': 'tldr@v2.0.0', 'text': 'The findings show that timing of surgery has an important bearing on mortality and morbidity after hip surgery, and older patients with higher orthopaedic POSSUM scores are sensitive to delays in surgery.'}",https://jorthoptraumatol.springeropen.com/counter/pdf/10.1007/s10195-013-0248-9
-Evolutionary fuzzy intelligent system for multi-objective supply chain network designs: an agent-based optimization state of the art,Mosahar Tarimoradi,,"{'model': 'tldr@v2.0.0', 'text': 'A fuzzy multi-agent system compatible with the decision makers’ interests and environmental survey is put forward according to which the parameters of the mathematical model are identified and its main competencies could be contemplated regarding the facts that it interactively fulfills the decisionmakers’ utilities.'}",
-StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery,Or Patashnik,"Inspired by the ability of StyleGAN to generate highly realistic images in a variety of domains, much recent work has focused on understanding how to use the latent spaces of StyleGAN to manipulate generated and real images. However, discovering semantically meaningful latent manipulations typically involves painstaking human examination of the many degrees of freedom, or an annotated collection of images for each desired manipulation. In this work, we explore leveraging the power of recently introduced Contrastive Language-Image Pre-training (CLIP) models in order to develop a text-based interface for StyleGAN image manipulation that does not require such manual effort. We first introduce an optimization scheme that utilizes a CLIP-based loss to modify an input latent vector in response to a user-provided text prompt. Next, we describe a latent mapper that infers a text-guided latent manipulation step for a given input image, allowing faster and more stable text-based manipulation. Finally, we present a method for mapping text prompts to input-agnostic directions in StyleGAN’s style space, enabling interactive text-driven image manipulation. Extensive results and comparisons demonstrate the effectiveness of our approaches.","{'model': 'tldr@v2.0.0', 'text': 'This work explores leveraging the power of recently introduced Contrastive Language-Image Pre-training (CLIP) models in order to develop a text-based interface for StyleGAN image manipulation that does not require such manual effort.'}",https://arxiv.org/pdf/2103.17249
-Null-text Inversion for Editing Real Images using Guided Diffusion Models,Ron Mokady,"Recent large-scale text-guided diffusion models provide powerful image generation capabilities. Currently, a massive effort is given to enable the modification of these images using text only as means to offer intuitive and versatile editing tools. To edit a real image using these state-of-the-art tools, one must first invert the image with a meaningful text prompt into the pretrained model's domain. In this paper, we introduce an accurate inversion technique and thus facilitate an intuitive text-based modification of the image. Our proposed inversion consists of two key novel components: (i) Pivotal inversion for diffusion models. While current methods aim at mapping random noise samples to a single input image, we use a single pivotal noise vector for each timestamp and optimize around it. We demonstrate that a direct DDIM inversion is inadequate on its own, but does provide a rather good anchor for our optimization. (ii) Null-text optimization, where we only modify the unconditional textual embedding that is used for classifier-free guidance, rather than the input text embedding. This allows for keeping both the model weights and the conditional embedding intact and hence enables applying prompt-based editing while avoiding the cumbersome tuning of the model's weights. Our null-text inversion, based on the publicly available Stable Diffusion model, is extensively evaluated on a variety of images and various prompt editing, showing high-fidelity editing of real images.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an accurate inversion technique and facilitates an intuitive text-based modification of the image using the Stable Diffusion model, and proposes Null-text optimization, where the unconditional textual embedding is used for classifier-free guidance, rather than the input text embedding.'}",https://arxiv.org/pdf/2211.09794
-Optimization of PGNAA set-up for the elements detection in aqueous solution,Wenbao Jia,,,
-CLIP-Mesh: Generating textured meshes from text using pretrained image-text models,N. Khalid,"We present a technique for zero-shot generation of a 3D model using only a target text prompt. Without any 3D supervision our method deforms the control shape of a limit subdivided surface along with its texture map and normal map to obtain a 3D asset that corresponds to the input text prompt and can be easily deployed into games or modeling applications. We rely only on a pre-trained CLIP model that compares the input text prompt with differentiably rendered images of our 3D model. While previous works have focused on stylization or required training of generative models we perform optimization on mesh parameters directly to generate shape, texture or both. To constrain the optimization to produce plausible meshes and textures we introduce a number of techniques using image augmentations and the use of a pretrained prior that generates CLIP image embeddings given a text embedding.","{'model': 'tldr@v2.0.0', 'text': 'A technique for zero-shot generation of a 3D model using only a target text prompt and a number of techniques using image augmentations and the use of a pretrained prior that generates CLIP image embeddings given a text embedding are presented.'}",https://arxiv.org/pdf/2203.13333
-RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learning,Mingkai Deng,"Prompting has shown impressive success in enabling large pre-trained language models (LMs) to perform diverse NLP tasks, especially with only few downstream data. Automatically finding the optimal prompt for each task, however, is challenging. Most existing work resorts to tuning *soft* prompts (e.g., embeddings) which fall short of interpretability, reusability across LMs, and applicability when gradients are not accessible. *Discrete* prompts, on the other hand, are difficult to optimize, and are often created by “enumeration (e.g., paraphrasing)-then-selection” heuristics that do not explore the prompt space systematically. This paper proposes RLPrompt, an efficient discrete prompt optimization approach with reinforcement learning (RL). RLPrompt formulates a parameter-efficient policy network that generates the optimized discrete prompt after training with reward. To harness the complex and stochastic reward signals from the large LM environment, we incorporate effective reward stabilization that substantially enhances training efficiency. RLPrompt is flexibly applicable to different types of LMs, such as masked (e.g., BERT) and left-to-right models (e.g., GPTs), for both classification and generation tasks. Experiments on few-shot classification and unsupervised text style transfer show superior performance over a wide range of existing fine-tuning or prompting methods. Interestingly, the resulting optimized prompts are often ungrammatical gibberish text; and surprisingly, those gibberish prompts are transferrable between different LMs to retain significant performance, indicating that LM prompting may not follow human language patterns.","{'model': 'tldr@v2.0.0', 'text': 'RLPrompt is proposed, an efficient discrete prompt optimization approach with reinforcement learning (RL) that formulates a parameter-efficient policy network that generates the optimized discrete prompt after training with reward.'}",http://arxiv.org/pdf/2205.12548
-An Overview of Variants and Advancements of PSO Algorithm,Meetu Jain,"Particle swarm optimization (PSO) is one of the most famous swarm-based optimization techniques inspired by nature. Due to its properties of flexibility and easy implementation, there is an enormous increase in the popularity of this nature-inspired technique. Particle swarm optimization (PSO) has gained prompt attention from every field of researchers. Since its origin in 1995 till now, researchers have improved the original Particle swarm optimization (PSO) in varying ways. They have derived new versions of it, such as the published theoretical studies on various parameters of PSO, proposed many variants of the algorithm and numerous other advances. In the present paper, an overview of the PSO algorithm is presented. On the one hand, the basic concepts and parameters of PSO are explained, on the other hand, various advances in relation to PSO, including its modifications, extensions, hybridization, theoretical analysis, are included.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the PSO algorithm is presented, the basic concepts and parameters of PSO are explained, and various advances in relation to PSO, including its modifications, extensions, hybridization, theoretical analysis, are included.'}",https://www.mdpi.com/2076-3417/12/17/8392/pdf?version=1661250377
-CLIP-Adapter: Better Vision-Language Models with Feature Adapters,Peng Gao,,"{'model': 'tldr@v2.0.0', 'text': 'This paper shows that there is an alternative path to achieve better vision-language models other than prompt tuning, and proposes CLIP-Adapter to conduct fine-tuning with feature adapters on either visual or language branch.'}",https://arxiv.org/pdf/2110.04544
-Dream3D: Zero-Shot Text-to-3D Synthesis Using 3D Shape Prior and Text-to-Image Diffusion Models,Jiale Xu,"Recent CLIP-guided 3D optimization methods, such as DreamFields [19] and PureCLIPNeRF [24], have achieved impressive results in zero-shot text-to-3D synthesis. However, due to scratch training and random initialization without prior knowledge, these methods often fail to generate accurate and faithful 3D structures that conform to the input text. In this paper, we make the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process. Specifically, we first generate a high-quality 3D shape from the input text in the text-to-shape stage as a 3D shape prior. We then use it as the initialization of a neural radiance field and optimize it with the full prompt. To address the challenging text-to-shape generation task, we present a simple yet effective approach that directly bridges the text and image modalities with a powerful text-to-image diffusion model. To narrow the style domain gap between the images synthesized by the text-to-image diffusion model and shape renderings used to train the image-to-shape generator, we further propose to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation. Our method, Dream3D, is capable of generating imaginative 3D content with superior visual quality and shape accuracy compared to state-of-the-art methods. Our project page is at https://bluestyle97.github.io/dream3d/.","{'model': 'tldr@v2.0.0', 'text': 'This paper makes the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process, and proposes to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation.'}",https://arxiv.org/pdf/2212.14704
-Training Diffusion Models with Reinforcement Learning,Kevin Black,"Diffusion models are a class of flexible generative models trained with an approximation to the log-likelihood objective. However, most use cases of diffusion models are not concerned with likelihoods, but instead with downstream objectives such as human-perceived image quality or drug effectiveness. In this paper, we investigate reinforcement learning methods for directly optimizing diffusion models for such objectives. We describe how posing denoising as a multi-step decision-making problem enables a class of policy gradient algorithms, which we refer to as denoising diffusion policy optimization (DDPO), that are more effective than alternative reward-weighted likelihood approaches. Empirically, DDPO is able to adapt text-to-image diffusion models to objectives that are difficult to express via prompting, such as image compressibility, and those derived from human feedback, such as aesthetic quality. Finally, we show that DDPO can improve prompt-image alignment using feedback from a vision-language model without the need for additional data collection or human annotation. The project's website can be found at http://rl-diffusion.github.io .","{'model': 'tldr@v2.0.0', 'text': 'It is described how posing denoising as a multi-step decision-making problem enables a class of policy gradient algorithms, which are referred to as Denoising diffusion policy optimization (DDPO), that are more effective than alternative reward-weighted likelihood approaches.'}",https://arxiv.org/pdf/2305.13301
-Delta Denoising Score,Amir Hertz,"We introduce Delta Denoising Score (DDS), a novel scoring function for text-based image editing that guides minimal modifications of an input image towards the content described in a target prompt. DDS leverages the rich generative prior of text-to-image diffusion models and can be used as a loss term in an optimization problem to steer an image towards a desired direction dictated by a text. DDS utilizes the Score Distillation Sampling (SDS) mechanism for the purpose of image editing. We show that using only SDS often produces non-detailed and blurry outputs due to noisy gradients. To address this issue, DDS uses a prompt that matches the input image to identify and remove undesired erroneous directions of SDS. Our key premise is that SDS should be zero when calculated on pairs of matched prompts and images, meaning that if the score is non-zero, its gradients can be attributed to the erroneous component of SDS. Our analysis demonstrates the competence of DDS for text based image-to-image translation. We further show that DDS can be used to train an effective zero-shot image translation model. Experimental results indicate that DDS outperforms existing methods in terms of stability and quality, highlighting its potential for real-world applications in text-based image editing.","{'model': 'tldr@v2.0.0', 'text': 'Delta Denoising Score is introduced, a novel scoring function for text-based image editing that guides minimal modifications of an input image towards the content described in a target prompt that can be used to train an effective zero-shot image translation model.'}",http://arxiv.org/pdf/2304.07090
-What Changes Can Large-scale Language Models Bring? Intensive Study on HyperCLOVA: Billions-scale Korean Generative Pretrained Transformers,Boseop Kim,"GPT-3 shows remarkable in-context learning ability of large-scale language models (LMs) trained on hundreds of billion scale data. Here we address some remaining issues less reported by the GPT-3 paper, such as a non-English LM, the performances of different sized models, and the effect of recently introduced prompt optimization on in-context learning. To achieve this, we introduce HyperCLOVA, a Korean variant of 82B GPT-3 trained on a Korean-centric corpus of 560B tokens. Enhanced by our Korean-specific tokenization, HyperCLOVA with our training configuration shows state-of-the-art in-context zero-shot and few-shot learning performances on various downstream tasks in Korean. Also, we show the performance benefits of prompt-based learning and demonstrate how it can be integrated into the prompt engineering pipeline. Then we discuss the possibility of materializing the No Code AI paradigm by providing AI prototyping capabilities to non-experts of ML by introducing HyperCLOVA studio, an interactive prompt engineering interface. Lastly, we demonstrate the potential of our methods with three successful in-house applications.","{'model': 'tldr@v2.0.0', 'text': 'The possibility of materializing the No Code AI paradigm by providing AI prototyping capabilities to non-experts of ML by introducing HyperCLOVA studio, an interactive prompt engineering interface is discussed and the performance benefits of prompt-based learning are shown and how it can be integrated into the prompt engineering pipeline.'}",https://aclanthology.org/2021.emnlp-main.274.pdf
-Directed Diffusion: Direct Control of Object Placement through Attention Guidance,W. Ma,"Text-guided diffusion models such as DALLE-2, Imagen, eDiff-I, and Stable Diffusion are able to generate an effectively endless variety of images given only a short text prompt describing the desired image content. In many cases the images are of very high quality. However, these models often struggle to compose scenes containing several key objects such as characters in specified positional relationships. The missing capability to ``direct'' the placement of characters and objects both within and across images is crucial in storytelling, as recognized in the literature on film and animation theory. In this work, we take a particularly straightforward approach to providing the needed direction. Drawing on the observation that the cross-attention maps for prompt words reflect the spatial layout of objects denoted by those words, we introduce an optimization objective that produces ``activation'' at desired positions in these cross-attention maps. The resulting approach is a step toward generalizing the applicability of text-guided diffusion models beyond single images to collections of related images, as in storybooks. Directed Diffusion provides easy high-level positional control over multiple objects, while making use of an existing pre-trained model and maintaining a coherent blend between the positioned objects and the background. Moreover, it requires only a few lines to implement.","{'model': 'tldr@v2.0.0', 'text': 'Directed Diffusion provides easy high-level positional control over multiple objects, while making use of an existing pre-trained model and maintaining a coherent blend between the positioned objects and the background.'}",https://arxiv.org/pdf/2302.13153
-CLIP-Actor: Text-Driven Recommendation and Stylization for Animating Human Meshes,Youwang Kim,"We propose CLIP-Actor, a text-driven motion recommendation and neural mesh stylization system for human mesh animation. CLIP-Actor animates a 3D human mesh to conform to a text prompt by recommending a motion sequence and optimizing mesh style attributes. We build a text-driven human motion recommendation system by leveraging a large-scale human motion dataset with language labels. Given a natural language prompt, CLIP-Actor suggests a text-conforming human motion in a coarse-to-fine manner. Then, our novel zero-shot neural style optimization detailizes and texturizes the recommended mesh sequence to conform to the prompt in a temporally-consistent and pose-agnostic manner. This is distinctive in that prior work fails to generate plausible results when the pose of an artist-designed mesh does not conform to the text from the beginning. We further propose the spatio-temporal view augmentation and mask-weighted embedding attention, which stabilize the optimization process by leveraging multi-frame human motion and rejecting poorly rendered views. We demonstrate that CLIP-Actor produces plausible and human-recognizable style 3D human mesh in motion with detailed geometry and texture solely from a natural language prompt.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that CLIP-Actor produces plausible and human-recognizable style 3D human mesh in motion with detailed geometry and texture solely from a natural language prompt.'}",http://arxiv.org/pdf/2206.04382
-Debiasing Scores and Prompts of 2D Diffusion for Robust Text-to-3D Generation,Susung Hong,"The view inconsistency problem in score-distilling text-to-3D generation, also known as the Janus problem, arises from the intrinsic bias of 2D diffusion models, which leads to the unrealistic generation of 3D objects. In this work, we explore score-distilling text-to-3D generation and identify the main causes of the Janus problem. Based on these findings, we propose two approaches to debias the score-distillation frameworks for robust text-to-3D generation. Our first approach, called score debiasing, involves gradually increasing the truncation value for the score estimated by 2D diffusion models throughout the optimization process. Our second approach, called prompt debiasing, identifies conflicting words between user prompts and view prompts utilizing a language model and adjusts the discrepancy between view prompts and object-space camera poses. Our experimental results show that our methods improve realism by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead.","{'model': 'tldr@v2.0.0', 'text': 'Two approaches to debias the score-distillation frameworks for robust text-to-3D generation are proposed and show that they improve realism by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead.'}",https://arxiv.org/pdf/2303.15413
-PromptBoosting: Black-Box Text Classification with Ten Forward Passes,Bairu Hou,"We describe PromptBoosting, a query-efficient procedure for building a text classifier from a neural language model (LM) without access to the LM's parameters, gradients, or hidden representations. This form of""black-box""classifier training has become increasingly important as the cost of training and inference in large-scale LMs grows. But existing black-box LM classifier learning approaches are themselves computationally inefficient, typically specializing LMs to the target task by searching in a large space of (discrete or continuous) prompts using zeroth-order optimization methods. Instead of directly optimizing in prompt space, PromptBoosting obtains a small pool of prompts via a gradient-free approach and then constructs a large pool of weak learners by pairing these prompts with different elements of the LM's output distribution. These weak learners are then ensembled using the AdaBoost algorithm. The entire learning process requires only a small number of forward passes and no backward pass. Experiments show that PromptBoosting achieves state-of-the-art performance in multiple black-box few-shot classification tasks, and matches or outperforms full fine-tuning in both few-shot and standard learning paradigms, while training 10x faster than existing black-box methods.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that PromptBoosting achieves state-of-the-art performance in multiple black-box few- shot classification tasks, and matches or outperforms full fine-tuning in both few-shot and standard learning paradigms, while training 10x faster than existing black- box methods.'}",http://arxiv.org/pdf/2212.09257
-Text to Mesh Without 3D Supervision Using Limit Subdivision,N. Khalid,"We present a technique for zero-shot generation of a 3D model using only a target text prompt. Without a generative model or any 3D supervision our method deforms a control shape of a limit subdivided surface along with a texture map and normal map to obtain a 3D model asset that matches the input text prompt and can be deployed into games or modeling applications. We rely only on a pre-trained CLIP model that compares the input text prompt with differen-tiably rendered images of our 3D model. While previous works have focused on stylization or required training of generative models we perform optimization on mesh parameters directly to generate shape and texture. To improve the quality of results we also introduce a set of techniques such as render augmentations, primitive selection, prompt aug-mentation that guide the mesh towards a suitable result.","{'model': 'tldr@v2.0.0', 'text': 'A technique for zero-shot generation of a 3D model using only a target text prompt and a set of techniques such as render augmentations, primitive selection, prompt aug-mentation that guide the mesh towards a suitable result are presented.'}",
-Unleashing the Power of Visual Prompting At the Pixel Level,Junyang Wu,"This paper presents a simple and effective visual prompting method for adapting pre-trained models to downstream recognition tasks. Our method includes two key designs. First, rather than directly adding together the prompt and the image, we treat the prompt as an extra and independent learnable component. We show that the strategy of reconciling the prompt and the image matters, and find that warping the prompt around a properly shrinked image empirically works the best. Second, we re-introduce two""old tricks""commonly used in building transferable adversarial examples, i.e., input diversity and gradient normalization, into visual prompting. These techniques improve optimization and enable the prompt to generalize better. We provide extensive experimental results to demonstrate the effectiveness of our method. Using a CLIP model, our prompting method sets a new record of 82.8% average accuracy across 12 popular classification datasets, substantially surpassing the prior art by +5.6%. It is worth noting that this prompting performance already outperforms linear probing by +2.1% and can even match fully fine-tuning in certain datasets. In addition, our prompting method shows competitive performance across different data scales and against distribution shifts. The code is publicly available at https://github.com/UCSC-VLAA/EVP.","{'model': 'tldr@v2.0.0', 'text': 'A simple and effective visual prompting method for adapting pre-trained models to downstream recognition tasks, which sets a new record of 82.8% average accuracy across 12 popular classification datasets, substantially surpassing the prior art by +5.6%.'}",http://arxiv.org/pdf/2212.10556
-Diversity-Aware Meta Visual Prompting,Qidong Huang,"We present Diversity-Aware Meta Visual Prompting (DAM-VP), an efficient and effective prompting method for transferring pre-trained models to downstream tasks with frozen backbone. A challenging issue in visual prompting is that image datasets sometimes have a large data diversity whereas a per-dataset generic prompt can hardly handle the complex distribution shift toward the original pretraining data distribution properly. To address this issue, we propose a dataset Diversity-Aware prompting strategy whose initialization is realized by a Meta-prompt. Specifically, we cluster the downstream dataset into small homogeneity subsets in a diversity-adaptive way, with each subset has its own prompt optimized separately. Such a divide-and-conquer design reduces the optimization difficulty greatly and significantly boosts the prompting performance. Furthermore, all the prompts are initialized with a meta-prompt, which is learned across several datasets. It is a bootstrapped paradigm, with the key observation that the prompting knowledge learned from previous datasets could help the prompt to converge faster and perform better on a new dataset. During inference, we dynamically select a proper prompt for each input, based on the feature distance between the input and each subset. Through extensive experiments, our DAM-VP demonstrates superior efficiency and effectiveness, clearly surpassing previous prompting methods in a series of downstream datasets for different pretraining models. Our code is available at: https://github.com/shikiw/DAM-VP.","{'model': 'tldr@v2.0.0', 'text': 'DAM-VP is an efficient and effective prompting method for transferring pre-trained models to downstream tasks with frozen backbone, clearly surpassing previous prompting methods in a series of downstream datasets for different pretraining models.'}",https://arxiv.org/pdf/2303.08138
-Reward Collapse in Aligning Large Language Models,Ziang Song,"The extraordinary capabilities of large language models (LLMs) such as ChatGPT and GPT-4 are in part unleashed by aligning them with reward models that are trained on human preferences, which are often represented as rankings of responses to prompts. In this paper, we document the phenomenon of \textit{reward collapse}, an empirical observation where the prevailing ranking-based approach results in an \textit{identical} reward distribution \textit{regardless} of the prompts during the terminal phase of training. This outcome is undesirable as open-ended prompts like ``write a short story about your best friend'' should yield a continuous range of rewards for their completions, while specific prompts like ``what is the capital of New Zealand'' should generate either high or low rewards. Our theoretical investigation reveals that reward collapse is primarily due to the insufficiency of the ranking-based objective function to incorporate prompt-related information during optimization. This insight allows us to derive closed-form expressions for the reward distribution associated with a set of utility functions in an asymptotic regime. To overcome reward collapse, we introduce a prompt-aware optimization scheme that provably admits a prompt-dependent reward distribution within the interpolating regime. Our experimental results suggest that our proposed prompt-aware utility functions significantly alleviate reward collapse during the training of reward models.","{'model': 'tldr@v2.0.0', 'text': 'The theoretical investigation reveals that reward collapse is primarily due to the insufficiency of the ranking-based objective function to incorporate prompt-related information during optimization, which allows us to derive closed-form expressions for the reward distribution associated with a set of utility functions in an asymptotic regime.'}",http://arxiv.org/pdf/2305.17608
-Structural and Electrochemical Aspects of LiNi0.8Co0.1Mn0.1O2 Cathode Materials Doped by Various Cations,T. Weigel,"Ni-rich materials of layered structure LiNixCoyMnzO2, x > 0.5, are promising candidates as cathodes in high-energy-density Li-ion batteries for electric vehicles. The structural and cycling stability of Ni-rich cathodes can be remarkably improved by doping with a small amount of extrinsic multivalent cations. In this study, we examine development of a fast screening methodology for doping LiNi0.8Co0.1Mn0.1O2 with cations Mg2+, Al3+, Si4+, Ti4+, Zr4+, and Ta5+ by a “top-down” approach. The cathode material is coated by a precursor layer that contains the dopant, which then is introduced into the particles by diffusion during heat treatment at elevated temperatures. The methodology described herein can be applied to Ni-rich cathode materials and allows relatively easy and prompt identification of the most promising dopants. Then further optimization work can lead to development of high-capacity stable cathode materials. The present study marks Ta5+ cations as very promising dopants for Ni-rich NCM cathodes.",,
-"Sepsis: The evolution in definition, pathophysiology, and management",B. Gyawali,"There has been a significant evolution in the definition and management of sepsis over the last three decades. This is driven in part due to the advances made in our understanding of its pathophysiology. There is evidence to show that the manifestations of sepsis can no longer be attributed only to the infectious agent and the immune response it engenders, but also to significant alterations in coagulation, immunosuppression, and organ dysfunction. A revolutionary change in the way we manage sepsis has been the adoption of early goal-directed therapy. This involves the early identification of at-risk patients and prompt treatment with antibiotics, hemodynamic optimization, and appropriate supportive care. This has contributed significantly to the overall improved outcomes with sepsis. Investigation into clinically relevant biomarkers of sepsis are ongoing and have yet to yield effective results. Scoring systems such as the sequential organ failure assessment and Acute Physiology and Chronic Health Evaluation help risk-stratify patients with sepsis. Advances in precision medicine techniques and the development of targeted therapy directed at limiting the excesses of the inflammatory and coagulatory cascades offer potentially viable avenues for future research. This review summarizes the progress made in the diagnosis and management of sepsis over the past two decades and examines promising avenues for future research.","{'model': 'tldr@v2.0.0', 'text': 'Advances in precision medicine techniques and the development of targeted therapy directed at limiting the excesses of the inflammatory and coagulatory cascades offer potentially viable avenues for future research.'}",https://journals.sagepub.com/doi/pdf/10.1177/2050312119835043
-ATT3D: Amortized Text-to-3D Object Synthesis,Jonathan Lorraine,"Text-to-3D modelling has seen exciting progress by combining generative text-to-image models with image-to-3D methods like Neural Radiance Fields. DreamFusion recently achieved high-quality results but requires a lengthy, per-prompt optimization to create 3D objects. To address this, we amortize optimization over text prompts by training on many prompts simultaneously with a unified model, instead of separately. With this, we share computation across a prompt set, training in less time than per-prompt optimization. Our framework - Amortized text-to-3D (ATT3D) - enables knowledge-sharing between prompts to generalize to unseen setups and smooth interpolations between text for novel assets and simple animations.","{'model': 'tldr@v2.0.0', 'text': 'The framework - Amortized text-to-3D (ATT3D) - enables knowledge-sharing between prompts to generalize to unseen setups and smooth interpolations between text for novel assets and simple animations.'}",http://arxiv.org/pdf/2306.07349
-Being Comes from Not-Being: Open-Vocabulary Text-to-Motion Generation with Wordless Training,Junfan Lin,"Text-to-motion generation is an emerging and challenging problem, which aims to synthesize motion with the same semantics as the input text. However, due to the lack of diverse labeled training data, most approaches either limit to specific types of text annotations or require online optimizations to cater to the texts during inference at the cost of efficiency and stability. In this paper, we investigate offline open-vocabulary text-to-motion generation in a zero-shot learning manner that neither requires paired training data nor extra online optimization to adapt for unseen texts. Inspired by the prompt learning in NLP, we pretrain a motion generator that learns to reconstruct the full motion from the masked motion. During inference, instead of changing the motion generator, our method reformulates the input text into a masked motion as the prompt for the motion generator to “reconstruct” the motion. In constructing the prompt, the unmasked poses of the prompt are synthesized by a text-to-pose generator. To supervise the optimization of the text-to-pose generator, we propose the first text-pose alignment model for measuring the alignment between texts and 3D poses. And to prevent the pose generator from over-fitting to limited training texts, we further propose a novel wordless training mechanism that optimizes the text-to-pose generator without any training texts. The comprehensive experimental results show that our method obtains a significant improvement against the baseline methods. The code is available at https://github.com/junfanlin/oohmg.","{'model': 'tldr@v2.0.0', 'text': 'This paper pretrain a motion generator that learns to reconstruct the full motion from the masked motion, inspired by the prompt learning in NLP, and proposes the first text-pose alignment model for measuring the alignment between texts and 3D poses.'}",https://arxiv.org/pdf/2210.15929
-Management of Acute Kidney Injury: Core Curriculum 2018.,P. K. Moore,"Acute kidney injury (AKI) is a heterogeneous disorder that is common in hospitalized patients and associated with short- and long-term morbidity and mortality. When AKI is present, prompt workup of the underlying cause should be pursued, with specific attention to reversible causes. Measures to prevent AKI include optimization of volume status and avoidance of nephrotoxic medications. Crystalloids are preferred over colloids for most patients, and hydroxyethyl starches should be avoided. Volume overload in the setting of AKI is associated with adverse outcomes, so attention should be paid to overall fluid balance. Currently there are no targeted pharmacotherapies approved for the treatment of AKI. The optimal timing of renal replacement therapy in critically ill patients with AKI is unclear, but is an area of active investigation. Recent studies suggest that AKI is not a ""self-limited"" process, but is strongly linked to increased risk for chronic kidney disease, subsequent AKI, and future mortality.","{'model': 'tldr@v2.0.0', 'text': 'Current studies suggest that AKI is not a ""self-limited"" process, but is strongly linked to increased risk for chronic kidney disease, subsequent AKI, and future mortality.'}",http://www.ajkd.org/article/S0272638617311411/pdf
-A Top-Down method for performance analysis and counters architecture,Ahmad Yasin,"Optimizing an application's performance for a given microarchitecture has become painfully difficult. Increasing microarchitecture complexity, workload diversity, and the unmanageable volume of data produced by performance tools increase the optimization challenges. At the same time resource and time constraints get tougher with recently emerged segments. This further calls for accurate and prompt analysis methods. The insights from this method guide a proposal for a novel performance counters architecture that can determine the true bottlenecks of a general out-of-order processor. Unlike other approaches, our analysis method is low-cost and already featured in in-production systems - it requires just eight simple new performance events to be added to a traditional PMU. It is comprehensive - no restriction to predefined set of performance issues. It accounts for granular bottlenecks in super-scalar cores, missed by earlier approaches.","{'model': 'tldr@v2.0.0', 'text': 'This analysis method is low-cost and already featured in in-production systems - it requires just eight simple new performance events to be added to a traditional PMU, and accounts for granular bottlenecks in super-scalar cores, missed by earlier approaches.'}",
-Research on path planning of mobile robot based on improved A* algorithm,Xiu-hong Wang,,"{'model': 'tldr@v2.0.0', 'text': 'The simulated results indicate that the improved A* algorithm can achieve the ideal effect of reducing and improving the search rate to finally locate the optimal path with precise search results and prompt response.'}",
-Fluid management for the prevention and attenuation of acute kidney injury,J. Prowle,,"{'model': 'tldr@v2.0.0', 'text': 'The conflict between the desire to achieve adequate resuscitation of shock and the need to mitigate the harmful effects of fluid overload is discussed.'}",
-Risk-Aware Identification of Highly Suspected COVID-19 Cases in Social IoT: A Joint Graph Theory and Reinforcement Learning Approach,Bowen Wang,"The recent outbreak of the coronavirus disease 2019 (COVID-19) has rapidly become a pandemic, which calls for prompt action in identifying suspected cases at an early stage through risk prediction. To suppress its further spread, we exploit the social relationships between mobile devices in the Social Internet of Things (SIoT) to help control its propagation by allocating the limited protective resources to the influential so-called high-degree individuals to stem the tide of precipitated spreading. By exploiting the so-called differential contact intensity and the infectious rate in susceptible-exposed-infected-removed (SEIR) epidemic model, the resultant optimization problem can be transformed into the minimum weight vertex cover (MWVC) problem of graph theory. To solve this problem in a high-dynamic random network topology, we propose an adaptive scheme by relying on the graph embedding technique during the state representation and reinforcement learning in the training phase. By relying on a pair of real-life datasets, the results demonstrate that our scheme can beneficially reduce the epidemiological reproduction rate of the infection. This technique has the potential of assisting in the early identification of COVID-19 cases.","{'model': 'tldr@v2.0.0', 'text': 'An adaptive scheme by relying on the graph embedding technique during the state representation and reinforcement learning in the training phase can beneficially reduce the epidemiological reproduction rate of the infection.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/09121230.pdf
-Hydra: An Ensemble of Convolutional Neural Networks for Geospatial Land Classification,R. Minetto,"In this paper, we describe Hydra, an ensemble of convolutional neural networks (CNNs) for geospatial land classification. The idea behind Hydra is to create an initial CNN that is coarsely optimized but provides a good starting pointing for further optimization, which will serve as the Hydra’s body. Then, the obtained weights are fine-tuned multiple times with different augmentation techniques, crop styles, and classes weights to form an ensemble of CNNs that represent the Hydra’s heads. By doing so, we prompt convergence to different endpoints, which is a desirable aspect for ensembles. With this framework, we were able to reduce the training time while maintaining the classification performance of the ensemble. We created ensembles for our experiments using two state-of-the-art CNN architectures, residual network (ResNet), and dense convolutional networks (DenseNet). We have demonstrated the application of our Hydra framework in two data sets, functional map of world (FMOW) and NWPU-RESISC45, achieving results comparable to the state-of-the-art for the former and the best-reported performance so far for the latter. Code and CNN models are available at https://github.com/maups/hydra-fmow.","{'model': 'tldr@v2.0.0', 'text': 'The idea behind Hydra is to create an initial CNN that is coarsely optimized but provides a good starting pointing for further optimization, which will serve as the Hydra’s body, an ensemble of convolutional neural networks (CNNs) for geospatial land classification.'}",
-Mechanisms of Cardiac Repair and Regeneration.,K. Broughton,"Cardiovascular regenerative therapies are pursued on both basic and translational levels. Although efficacy and value of cell therapy for myocardial regeneration can be debated, there is a consensus that profound deficits in mechanistic understanding limit advances, optimization, and implementation. In collaboration with the TACTICS (Transnational Alliance for Regenerative Therapies in Cardiovascular Syndromes), this review overviews several pivotal aspects of biological processes impinging on cardiac maintenance, repair, and regeneration. The goal of summarizing current mechanistic understanding is to prompt innovative directions for fundamental studies delineating cellular reparative and regenerative processes. Empowering myocardial regenerative interventions, whether dependent on endogenous processes or exogenously delivered repair agents, ultimately depends on mastering mechanisms and novel strategies that take advantage of rather than being limited by inherent myocardial biology.","{'model': 'tldr@v2.0.0', 'text': 'Current mechanistic understanding of biological processes impinging on cardiac maintenance, repair, and regeneration is summarized to prompt innovative directions for fundamental studies delineating cellular reparative and regenerative processes.'}",https://www.ahajournals.org/doi/pdf/10.1161/CIRCRESAHA.117.312586
-Timetabling optimization of a single railway track line with sensitivity analysis,Enrique Castillo,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that an adequate selection of intermediate stations and of the departure times are crucial in the good performance of the line and that inadequate spacings between consecutive trains can block the line.'}",
-Liver MR Elastography Technique and Image Interpretation: Pearls and Pitfalls.,F. Guglielmo,"Liver MR elastography is an imaging technique used to measure liver stiffness in the evaluation for possible fibrosis or cirrhosis. Liver stiffness measurement (LSM) is useful for predicting the stage of liver fibrosis. However, obtaining and reporting accurate and reliable LSMs with MR elastography require an understanding of the three core components of liver MR elastography: optimization of imaging technique, prompt quality control of images, and proper interpretation and reporting of elastogram findings. When performing MR elastography, six important technical parameters that should be optimized are patient fasting before the examination, proper passive driver placement, proper MR elastography section positioning over the largest area of the liver, use of MR elastography-related sequences at end expiration, choosing the best timing of the MR elastography sequence, and optimization of several essential pulse sequence parameters. As soon as the MR elastography examination is performed, the elastograms should be reviewed to ensure that they are of diagnostic quality so that corrective steps can be taken, if needed, and MR elastography can be repeated before the diagnostic portion of the examination concludes. Finally, the interpreting radiologist needs to understand and be able to perform the proper technique for LSMs, including determining which areas of the liver to include or avoid in the measurements; knowing which conditions, other than fibrosis or cirrhosis, can increase liver stiffness; and understanding how to report elastography results. This article reviews the proper technique for performing liver MR elastography and subsequent quality control assessment, as well as the principles for interpreting and reporting studies. This review may be helpful for implementing and operating a clinical liver MR elastography service.The full digital presentation is available online.©RSNA, 2019.","{'model': 'tldr@v2.0.0', 'text': 'The proper technique for performing liver MR elastography and subsequent quality control assessment, as well as the principles for interpreting and reporting studies are reviewed, may be helpful for implementing and operating a clinical liver MRElastography service.'}",
-Single Image Dehazing with Dark Channel Prior,J. Kaur,"In applications related to computation of visuals, the clear picture is very crucial component but certain atmospheric conditions like haze and fog prompt the suspended particles which lead to degradation of the image. This hinders the flow of vision systems. Therefore, implementation of image dehazing has become important to improve the throughput and effectiveness of such vision systems. This paper sheds light on execution of dark channel to delete haze from single image. Different measures to optimize this algorithm are discussed with results obtained from certain successful attempts to implement this algorithm.","{'model': 'tldr@v2.0.0', 'text': 'Light is shed on execution of dark channel to delete haze from single image and different measures to optimize this algorithm are discussed with results obtained from certain successful attempts to implement this algorithm.'}",
-Adaptive cubic overestimation methods for unconstrained optimization,C. Cartis,"An Adaptive Cubic Overestimation (ACO) algorithm for unconstrained optimization, generalizing a method due to Nesterov & Polyak (Math. Programming 108, 2006, pp 177-205), is proposed. At each iteration of Nesterov & Polyak's approach, the global minimizer of a local cubic overestimator of the objective function is determined, and this ensures a significant improvement in the objective so long as the Hessian of the objective is Lipschitz continuous and its Lipschitz constant is available. The twin requirements of global model optimality and the availability of Lipschitz constants somewhat limit the applicability of such an approach, particularly for large-scale problems. However the promised powerful worst-case theoretical guarantees prompt us to investigate variants in which estimates of the required Lipschitz constant are refined and in which computationally-viable approximations to the global model-minimizer are sought. We show that the excellent global and local convergence properties and worst-case iteration complexity bounds obtained by Nesterov & Polyak are retained, and sometimes extended to a wider class of problems, by our ACO approach. Numerical experiments with small-scale test problems from the CUTEr set show superior performance of the ACO algorithm when compared to a trust-region implementation.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the excellent global and local convergence properties and worst-case iteration complexity bounds obtained by Nesterov & Polyak are retained, and sometimes extended to a wider class of problems, by the ACO approach.'}",
-Multi-swarm particle swarm optimization based on mixed search behavior,Jing Jie,"The paper develops a Multi-swarm particle swarm optimization (MPSO) to overcome the premature convergence problem. MPSO takes advantage of multiple sub-swarms with mixed search behavior to maintain the swarm diversity, and introduces cooperative mechanism to prompt the information exchange among sub-swarms. Moreover, MPSO adopts an adaptive reinitializing strategy guided by swarm diversity, which can contribute to the global convergence of the algorithm. Through the mixed local search behavior modes, the cooperative search and the reinitializing strategy guided by swarm diversity, MPSO can maintain appropriate diversity and keep the balance of local search and global search validly. The proposed MPSO was applied to some well-known benchmarks. The experimental results show MPSO is a robust global optimization technique for the complex multimodal functions.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show MPSO is a robust global optimization technique for the complex multimodal functions and can maintain appropriate diversity and keep the balance of local search and global search validly.'}",
-Seeking for sterile neutrinos with displaced leptons at the LHC,Jia Liu,,,https://link.springer.com/content/pdf/10.1007/JHEP07(2019)159.pdf
-Handover Parameter Optimization in WCDMA using Fuzzy Controlling,C. Werner,"The increasing traffic and the demand for high data rate services in WCDMA networks prompt the need for an automatic network optimization. Current state-of-the-art optimization methods adapt physical parameters, such as the antennas' tilt or azimuth. As agreed in the community, a further capacity increase can be achieved by an optimization of radio resource management (RRM) parameters. While theoretical approaches of RRM parameter optimization have already been introduced in the literature, practical solutions optimizing cell individual parameters have rarely been treated so far. Consequently, this paper copes with the optimization of the cell individual offset (CIO) in the soft handover algorithm with the goal to reduce the network's outage probability. We design and apply a fuzzy logic controller, whose output are CIO changes matching the current traffic and load conditions. Verifications using a dynamic system simulator prove the powerfulness of our method and promise significant outage reductions in a fully loaded network scenario leading to capacity improvements of up to 9 %.","{'model': 'tldr@v2.0.0', 'text': ""This paper copes with the optimization of the cell individual offset (CIO) in the soft handover algorithm with the goal to reduce the network's outage probability and promises significant outage reductions in a fully loaded network scenario leading to capacity improvements of up to 9 %.""}",
-Optimization of injected dose based on noise equivalent count rates for 2- and 3-dimensional whole-body PET.,C. Lartizien,"UNLABELLED The noise equivalent count (NEC) rate index is used to derive guidelines on the optimal injected dose to the patient for 2-dimensional (2D) and 3-dimensional (3D) whole-body PET acquisitions. METHODS We performed 2D and 3D whole-body acquisitions of an anthropomorphic phantom modeling the conditions for (18)F-FDG PET of the torso and measured the NEC rates for different activity levels for several organs of interest. The correlations between count rates measured from the phantom and those from a series of whole-body patient scans were then analyzed. This analysis allowed validation of our approach and estimation of the injected dose that maximizes NEC rate as a function of patient morphology for both acquisition modes. RESULTS Variations of the phantom and patient prompt and random coincidence rates as a function of single-photon rates correlated well. On the basis of these correlations, we demonstrated that the patient NEC rate can be predicted for a given single-photon rate. Finally, we determined that patient single-photon rates correlated with the mean dose per weight at acquisition start when normalized by the body mass index. This correlation allows modifying the injected dose as a function of patient body mass index to reach the peak NEC rate in 3D mode. Conversely, we found that the peak NEC rates were never reached in 2D mode within an acceptable range of injected dose. CONCLUSION The injected dose was adapted to patient morphology for 2D and 3D whole-body acquisitions using the NEC rate as a figure of merit of the statistical quality of the sinogram data. This study is a first step toward a more comprehensive comparison of the image quality obtained using both acquisition modes.","{'model': 'tldr@v2.0.0', 'text': 'The injected dose was adapted to patient morphology for 2D and 3D whole-body acquisitions using the NEC rate as a figure of merit of the statistical quality of the sinogram data as a first step toward a more comprehensive comparison of the image quality obtained using both acquisition modes.'}",
-Bleach optimization of sputum smear microscopy for pulmonary tuberculosis.,P. Srikanth,"The Revised National Tuberculosis Control Programme (RNTCP) aims to improve case detection rates of tuberculosis to facilitate prompt recognition and treatment. The low case detection rates in the programme can be directly attributed to failure to screen patients with suspected tuberculosis and the low sensitivity of the direct smear microscopy method to detect cases among the fraction of patients that are screened. Apart from low sensitivity, this method also has other disadvantages including the increased risk of infection transmission to technicians. There are several methods that can be used to improve sensitivity, but their applicability in a national programme and in resource limited settings are limited. Bleach processing of sputum smears prior to microscopy may be a cheap and effective way to improve on the sensitivity of the direct smear. Four distinctive techniques of sputum smear processing using bleach are described in the review, with the variations in each technique, along with the sensitivity. An analysis of reports published earlier on the bleach method is also presented including a discussion on when and why the bleach method works.","{'model': 'tldr@v2.0.0', 'text': 'Four distinctive techniques of sputum smear processing using bleach are described in the review, with the variations in each technique, along with the sensitivity.'}",
-Optimization of the TEM feed structure for four-arm reflector impulse radiating antennas,J. Tyo,"This paper considers the optimization of the feed arm geometry of four-arm crossed-coplanar plate impulse radiating antennas (IRAs) when the angular position and extent of the arms are taken as free parameters. Previously, optimization of this class of antenna considered only the symmetric case where the two pairs of crossed feed arms were perpendicular to each other. Comparison is made using the prompt aperture efficiency, and the results indicate that the efficiency of four-arm IRAs can be increased from /spl sim/25% for the perpendicularly crossed arms to /spl sim/35% for the optimum configuration. In addition to the optimization, the feed impedance of coplanar feeds is presented for general values of feed arm angle and plate width, and the optimum feed impedance is computed for each feed arm angle. The results can be used to design the optimal four-arm IRA with an arbitrary specified input impedance.",,
-Fast dynamic routing based on weighted kernel density estimation,Suofei Zhang,"Capsules as well as dynamic routing between them are most recently proposed structures for deep neural networks. A capsule groups data into vectors or matrices as poses rather than conventional scalars to represent specific properties of target instance. Based on pose, a capsule should be attached to a probability (often denoted as activation) for its presence. The dynamic routing helps capsule network achieve more generalization capacity with fewer model parameters. However, the bottleneck, which prevents widespread applications of capsule, is the expense of computation during routing. To address this problem, we generalize existing routing methods within the framework of weighted kernel density estimation, proposing two fast routing methods with different optimization strategies. Our methods prompt the time efficiency of routing by nearly 40% with negligible performance degradation. By stacking a hybrid of convolutional layers and capsule layers, we construct a network architecture to handle inputs at a resolution of 64 × 64 pixels. The proposed models achieve a parallel performance with other leading methods in multiple benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This work generalizes existing routing methods within the framework of weighted kernel density estimation, proposing two fast routing methods with different optimization strategies that prompt the time efficiency of routing by nearly 40% with negligible performance degradation.'}",https://arxiv.org/pdf/1805.10807
-OPTIMIZATION OF STEADY-STATE BEAM-DRIVEN TOKAMAK REACTORS.,David R. Mikkelsen,"Recent developments in neutral beam technology prompt us to reconsider the prospects for steady-state tokamak reactors. A mathematical reactor model is developed that includes the physics of beam-driven currents and reactor power balance, as well as reactor and beam system costs. This model is used to find the plasma temperatures that minimize the reactor cost per unit of net electrical output. The optimum plasma temperatures are nearly independent of ..beta.. and are roughly twice as high as the optimum temperatures for ignited reactors. If beams of neutral deuterium atoms with near-optimum energies of 1 to 2 MeV are used to drive the current in a reactor the size of the International Tokamak Reactor, then the optimum temperatures are typically T /SUB e/ approx. = 12 to 15 keV and T /SUB i/ approx. = 17 to 21 keV for a wide range of model parameters. Net electrical output rises rapidly with increasing deuterium beam energy for E /SUB b/ less than or equal to 400 keV, but rises only slowly above E /SUB b/ about 1 MeV. We estimate that beam-driven steady-state reactors could be economically competitive with pulsed-ignition reactors if cyclic-loading problems limit the toroidal magnetic field strength ofmore » pulsed reactors to less than or equal to 85% of that allowed in steady-state reactors.« less",,https://digital.library.unt.edu/ark:/67531/metadc1192439/m2/1/high_res_d/6772488.pdf
-Seasonal Food Scarcity Prompts Long-Distance Foraging by a Wild Social Bee,Nathaniel S. Pope,"Foraging is an essential process for mobile animals, and its optimization serves as a foundational theory in ecology and evolution; however, drivers of foraging are rarely investigated across landscapes and seasons. Using a common bumblebee species from the western United States (Bombus vosnesenskii), we ask whether seasonal decreases in food resources prompt changes in foraging behavior and space use. We employ a unique integration of population genetic tools and spatially explicit foraging models to estimate foraging distances and rates of patch visitation for wild bumblebee colonies across three study regions and two seasons. By mapping the locations of 669 wild-caught individual foragers, we find substantial variation in colony-level foraging distances, often exhibiting a 60-fold difference within a study region. Our analysis of visitation rates indicates that foragers display a preference for destination patches with high floral cover and forage significantly farther for these patches, but only in the summer, when landscape-level resources are low. Overall, these results indicate that an increasing proportion of long-distance foraging bouts take place in the summer. Because wild bees are pollinators, their foraging dynamics are of urgent concern, given the potential impacts of global change on their movement and services. The behavioral shift toward long-distance foraging with seasonal declines in food resources suggests a novel, phenologically directed approach to landscape-level pollinator conservation and greater consideration of late-season floral resources in pollinator habitat management.","{'model': 'tldr@v2.0.0', 'text': 'The behavioral shift toward long-distance foraging with seasonal declines in food resources suggests a novel, phenologically directed approach to landscape-level pollinator conservation and greater consideration of late-season floral resources in pollinator habitat management.'}",
-Reinfarction Following Anesthesia in Patients with Myocardial Infarction,T. Rao,"The authors studied the incidence of and factors related to recurrent perioperative myocardial infarction retrospectively during 1973–1976 (Group 1) and prospectively during 1977–1982 (Group 2). Reinfarction occurred in 28 of 364 (7.7%) patients in Group 1 and 14 of 733 (1.9%) in Group 2 (P < 0.005). When the previous infarction was 0–3 and 4–6 months old, perioperative reinfarction occurred in 36% and 26% of Group 1 patients, respectively, and only 5.7% and 2.3% of Group 2 patients, respectively, (P < 0.05). In both groups, patients with associated congestive heart failure had a higher reinfarction rate. Patients who had intraoperative hypertension and tachycardia or hypotension develop had a higher incidence of reinfarction in both groups. The results suggest that preoperative optimization of the patient's status, aggressive invasive monitoring of the hemodynamic status, and prompt treatment of any hemodynamic aberration may be associated with decreased perioperative morbidity and mortality in patients with previous myocardial infarction. Which of these factors, if any, contributed to the improved outcome was not determined in this study.","{'model': 'tldr@v2.0.0', 'text': ""The results suggest that preoperative optimization of the patient's status, aggressive invasive monitoring of the hemodynamic status, and prompt treatment of any hemodynamic aberration may be associated with decreased perioperative morbidity and mortality in patients with previous myocardial infarction.""}",https://pubs.asahq.org/anesthesiology/article-pdf/59/6/499/626415/0000542-198312000-00003.pdf
-Feasibility of Portfolio Optimization under Coherent Risk Measures,I. Kondor,"It is shown that the axioms for coherent risk measures imply that whenever there is an asset in a portfolio that dominates the others in a given sample (which happens with finite probability even for large samples), then this portfolio cannot be optimized under any coherent measure on that sample, and the risk measure diverges to minus infinity. This instability was first discovered on the special example of Expected Shortfall which is used here both as an illustration and as a prompt for generalization.",,
-Unhealthy Stem Cells: When Health Conditions Upset Stem Cell Properties,Laura M Pérez,"The stem cell field has grown very rapidly during the last decade, offering the promise of innovative therapies to treat disease. Different stem cell populations have been isolated from various human adult tissues, mainly from bone marrow and adipose tissue, but many other body tissues harbor a stem cell population. Adult tissue stem cells are invariably found in discrete microenvironments termed niches, where they play key roles in tissue homeostasis by enabling lifelong optimization of organ form and function. Some diseases are known to strike at the stem cell population, through alterations in their specific microenvironments, making them non-viable. Furthermore, it has been shown that a transformed stem cell population could prompt the development of certain cancers. This review focuses on the potential negative aspects of a range of diseases on the activity of stem cells and how their potential use in cell therapies may be affected.","{'model': 'tldr@v2.0.0', 'text': 'This review focuses on the potential negative aspects of a range of diseases on the activity of stem cells and how their potential use in cell therapies may be affected.'}",https://www.karger.com/Article/Pdf/489440
-A Randomized Controlled Trial of an Optimized Smoking Treatment Delivered in Primary Care.,Megan E. Piper,"Background The effectiveness of smoking cessation treatment is limited in real-world use, perhaps because we have not selected the components of such treatments optimally nor have treatments typically been developed for and evaluated in real-world clinical settings. Purpose To validate an optimized smoking cessation treatment package that comprises intervention components identified as effective in factorial screening experiments conducted as per the Multiphase Optimization Strategy (MOST). Methods Adult smokers motivated to quit were recruited from primary care clinics (N = 623). Participants were randomized to receive either recommended usual care (R-UC; 10 min of in-person counseling, 8 weeks of nicotine patch, and referral to quitline services) or abstinence-optimized treatment (A-OT; 3 weeks of prequit mini-lozenges, 26 weeks of nicotine patch + mini-lozenges, three in-person and eight phone counseling sessions, and 7-11 automated calls to prompt medication use). The key outcomes were self-reported and biochemically confirmed (carbon monoxide, CO <6 ppm) 7-day point-prevalence abstinence. Results A-OT participants had significantly higher self-reported abstinence rates than R-UC participants at 4, 8, 16, and 26 weeks (ORs: 1.91-3.05; p <. 001). The biochemically confirmed 26-week abstinence rates were lower than the self-reported 26-week rates, but revealed a similar treatment effect size (OR = 2.94, p < .001). There was no moderation of treatment effects on 26-week abstinence by demographic, psychiatric, or nicotine dependence variables. A-OT had an incremental cost-effectiveness ratio for 26-week CO-confirmed abstinence of $7,800. Conclusions A smoking cessation treatment that is optimized via MOST development meaningfully enhances cessation rates beyond R-UC smoking treatment in smokers seen in primary care. Clinical Trial Registration NCT02301403.","{'model': 'tldr@v2.0.0', 'text': 'A smoking cessation treatment that is optimized via MOST development meaningfully enhances cessation rates beyond R-UC smoking treatment in smokers seen in primary care.'}",
-Optimization Models for Streaming in Multihop Wireless Networks,Zongpeng Li,"Wireless spectrum is a scare resource, while media streaming usually requires high end-to-end bandwidth. Media streaming in wireless ad hoc networks is therefore a particularly challenging problem, especially for the case of streaming to multiple receivers. In this paper, we design linear optimization models for computing a high-bandwidth routing strategy for media multicast in wireless networks, which targets near-optimal throughput, given constraints including network topology, radio capacity, and link contention. We study both the directional antenna and omni-directional antenna cases and point out their connections. We also combine the classic forward error correction techniques with the novel network coding techniques to provide error control in a timely fashion. Simulation results show that our solutions indeed achieve high streaming rates, and prompt error recovery under a wide range of link failure patterns.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs linear optimization models for computing a high-bandwidth routing strategy for media multicast in wireless networks, which targets near-optimal throughput, given constraints including network topology, radio capacity, and link contention.'}",http://www.eecg.toronto.edu/~bli/papers/zp-icccn07.pdf
-Time domain compressive beam forming of ultrasound signals.,G. David,"Ultrasound imaging is a wide spread technique used in medical imaging as well as in non-destructive testing. The technique offers many advantages such as real-time imaging, good resolution, prompt acquisition, ease of use, and low cost compared to other techniques such as x-ray imaging. However, the maximum frame rate achievable is limited as several beams must be emitted to compute a single image. For each emitted beam, one must wait for the wave to propagate back and forth, thus imposing a limit to the frame rate. Several attempts have been made to use less beams while maintaining image quality. Although efficiently increasing the frame rate, these techniques still use several transmit beams. Compressive Sensing (CS), a universal data completion scheme based on convex optimization, has been successfully applied to a number of imaging modalities over the past few years. Using a priori knowledge of the signal, it can compute an image using less data allowing for shorter acquisition times. In this paper, it is shown that a valid CS framework can be derived from ultrasound propagation theory, and that this framework can be used to compute images of scatterers using only one plane wave as a transmit beam.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a valid CS framework can be derived from ultrasound propagation theory, and that this framework could be used to compute images of scatterers using only one plane wave as a transmit beam.'}",https://pubs.aip.org/asa/jasa/article-pdf/137/5/2773/15314040/2773_1_online.pdf
-Independent Tasks Scheduling Based on Genetic Algorithm in Cloud Computing,C. Zhao,"Task scheduling algorithm, which is an NP-completeness problem, plays a key role in cloud computing systems. In this paper, we propose an optimized algorithm based on genetic algorithm to schedule independent and divisible tasks adapting to different computation and memory requirements. We prompt the algorithm in heterogeneous systems, where resources (including CPUs) are of computational and communication heterogeneity. Dynamic scheduling is also in consideration. Though GA is designed to solve combinatorial optimization problem, it's inefficient for global optimization. So we conclude with further researches in optimized genetic algorithm.","{'model': 'tldr@v2.0.0', 'text': ""Though GA is designed to solve combinatorial optimization problem, it's inefficient for global optimization, so this paper concludes with further researches in optimized genetic algorithm.""}",
-From Macrohemodynamic to the Microcirculation,A. Donati,"ICU patients need a prompt normalization of macrohemodynamic parameters. Unfortunately, this optimization sometimes does not protect patients from organ failure development. Prevention or treatment of organ failure needs another target to be pursued: the microcirculatory restoration. Microcirculation is the ensemble of vessels of maximum 100 μm in diameter. Nowadays the Sidestream Dark Field (SDF) imaging technique allows its bedside investigation and a recent round-table conference established the criteria for its evaluation. First, microcirculatory derangements have been studied in sepsis: they are mainly characterized by a reduction of vessel density, an alteration of flow, and a heterogeneous distribution of perfusion. Endothelial malfunction and glycocalyx rupture were proved to be the main reasons for the observed microthrombi, capillary leakage, leukocyte rolling, and rouleaux phenomenon, even if further studies are necessary for a better explanation. Therapeutic approaches targeting microcirculation are under investigation. Microcirculatory alterations have been recently demonstrated in other diseases such as hypovolemia and cardiac failure but this issue still needs to be explored. The aim of this paper is to gather the already known information, focus the reader's attention on the importance of microvascular physiopathology in critical illness, and prompt him to actively participate to achieve a more comprehensive understanding of the issue.","{'model': 'tldr@v2.0.0', 'text': ""The aim of this paper is to gather the already known information, focus the reader's attention on the importance of microvascular physiopathology in critical illness, and prompt him to actively participate to achieve a more comprehensive understanding of the issue.""}",https://downloads.hindawi.com/journals/ccrp/2013/892710.pdf
-Contemporary mechanical circulatory support therapy for postcardiotomy shock,S. Fukuhara,,"{'model': 'tldr@v2.0.0', 'text': 'The reported in-hospital mortality of the PCS patients remains high, consistently over 50\xa0%, despite ongoing refinements of MCS technology, and the optimization of selection criteria and the prompt institution of M CS are likely the keys to improving this persistently high mortality rate.'}",
-Fragment-to-hit-to-lead discovery of a novel pyridylurea scaffold of ATP competitive dual targeting type II topoisomerase inhibiting antibacterial agents.,G. Basarab,"The discovery and optimization of a new class of bacterial topoisomerase (DNA gyrase and topoisomerase IV) inhibitors binding in the ATP domain are described. A fragment molecule, 1-ethyl-3-(2-pyridyl)urea, provided sufficiently potent enzyme inhibition (32 μM) to prompt further analogue work. Acids and acid isosteres were incorporated at the 5-pyridyl position of this fragment, bridging to a key asparagine residue, improving enzyme inhibition, and leading to measurable antibacterial activity. A CF3-thiazole substituent at the 4-pyridyl position improved inhibitory potency due to a favorable lipophilic interaction. Promising antibacterial activity was seen versus the Gram-positive pathogens Staphylococcus aureus and Streptococcus pneumoniae and the Gram-negative pathogens Haemophilus influenzae and Moraxella catarrhalis . Precursor metabolite incorporation and mutant analysis studies support the mode-of-action, blockage of DNA synthesis by dual target topoisomerase inhibition. Compound 35 was efficacious in a mouse S. aureus disease model, where a 4.5-log reduction in colony forming units versus control was demonstrated.","{'model': 'tldr@v2.0.0', 'text': 'Preursor metabolite incorporation and mutant analysis studies support the mode-of-action, blockage of DNA synthesis by dual target topoisomerase inhibition.'}",
-Enhanced Extracellular Polysaccharide Production and Self-Sustainable Electricity Generation for PAMFCs by Scenedesmus sp. SB1,M. Angelaalincy,"In this study, a freshwater microalga, Scenedesmus sp. SB1, was isolated, purified, and identified by its internal transcribed spacer region (ITS1-5.8S-ITS2). Media optimization through the Plackett–Burman Design and response surface methodology (RSM) showed a maximum exopolysaccharide (EPS) production of 48 mg/L (1.8-fold higher than that for unoptimized media). Characterization using gas chromatography–mass spectrometry, Fourier transform infrared, X-ray diffraction, and thermogravimetric analysis reveals that the EPS is a sulfated pectin polysaccharide with a crystallinity index of 15.2% and prompt thermal stability. Furthermore, the photoelectrogenic activity of Scenedesmus sp. SB1 inoculated in BG-11 and RSM-optimized BG-11 (ROBG-11) media was tested by cyclic voltammogram studies, revealing the potential of the inoculated strain in ROBG-11 toward photosynthetic algal microbial fuel cells over normal BG-11. To the best of our knowledge, functional group characterization, physical and thermal property and media optimization for EPS production by RSM and electrogenic activity studies are reported for the first time in Scenedesmus sp. SB1.","{'model': 'tldr@v2.0.0', 'text': 'Functional group characterization, physical and thermal property and media optimization for EPS production by RSM and electrogenic activity studies are reported for the first time in Scenedesmus sp.'}",https://pubs.acs.org/doi/pdf/10.1021/acsomega.7b00326
-Transcriptome Profiling of Antimicrobial Resistance in Pseudomonas aeruginosa,A. Khaledi,"ABSTRACT Emerging resistance to antimicrobials and the lack of new antibiotic drug candidates underscore the need for optimization of current diagnostics and therapies to diminish the evolution and spread of multidrug resistance. As the antibiotic resistance status of a bacterial pathogen is defined by its genome, resistance profiling by applying next-generation sequencing (NGS) technologies may in the future accomplish pathogen identification, prompt initiation of targeted individualized treatment, and the implementation of optimized infection control measures. In this study, qualitative RNA sequencing was used to identify key genetic determinants of antibiotic resistance in 135 clinical Pseudomonas aeruginosa isolates from diverse geographic and infection site origins. By applying transcriptome-wide association studies, adaptive variations associated with resistance to the antibiotic classes fluoroquinolones, aminoglycosides, and β-lactams were identified. Besides potential novel biomarkers with a direct correlation to resistance, global patterns of phenotype-associated gene expression and sequence variations were identified by predictive machine learning approaches. Our research serves to establish genotype-based molecular diagnostic tools for the identification of the current resistance profiles of bacterial pathogens and paves the way for faster diagnostics for more efficient, targeted treatment strategies to also mitigate the future potential for resistance evolution.","{'model': 'tldr@v2.0.0', 'text': 'This research serves to establish genotype-based molecular diagnostic tools for the identification of the current resistance profiles of bacterial pathogens and paves the way for faster diagnostics for more efficient, targeted treatment strategies to also mitigate the future potential for resistance evolution.'}",https://aac.asm.org/content/aac/60/8/4722.full.pdf
-"Triplet–Triplet Annihilation in 9,10-Diphenylanthracene Derivatives: The Role of Intersystem Crossing and Exciton Diffusion",Tomas Serevičius,"Triplet–triplet annihilation (TTA) is an attractive way to boost the efficiency of conventional fluorescent organic light-emitting diodes (OLEDs). TTA-active anthracene derivatives are often considered as state-of-the-art emitters due to the proper energy level alignment. In this work, TTA properties of a series of highly fluorescent nonsymmetrical anthracene compounds bearing 9-(4-arylphenyl) moiety and 10-(4-hexylphenyl) fragments were assessed. Two different methods to enhance the TTA efficiency are demonstrated. First, the intensity of TTA-based delayed fluorescence directly depended on the intersystem crossing (ISC) rate. This ISC rate can be significantly enhanced in more conjugated compounds due to the resonant alignment of S1 and T2 energy levels. While enhanced ISC rate slightly quenches the intensity of prompt fluorescence, the rise of the triplet population boosts the intensity of resultant delayed fluorescence. Second, the triplet annihilation rate can be significantly enhanced by optimization...",,
-"Multicast routing and its QoS extension: problems, algorithms, and protocols",Bin Wang,"Multicast services have been increasingly used by various continuous media applications. The QoS requirements of these continuous media applications prompt the necessity for QoS-driven, constraint-based multicast routing. This article provides a comprehensive overview of existing multicast routing algorithms, protocols, and their QoS extension. In particular, we classify multicast routing problems according to their optimization functions and performance constraints, present basic routing algorithms in each problem class, and discuss their strengths and weaknesses. We also categorize existing multicast routing protocols, outline the issues and challenges in providing QoS in multicast routing, and point out possible future research directions.","{'model': 'tldr@v2.0.0', 'text': 'This article classifies multicast routing problems according to their optimization functions and performance constraints, present basic routing algorithms in each problem class, and discuss their strengths and weaknesses.'}",
-Photochemical Characterization and Optimization of Argon Ion Laser-Induced Fluorescence Detection of o-Phthalaldehyde/.beta.-Mercaptoethanol-Labeled Amino Acids and .gamma.-Glutamyl Peptides in Liquid Chromatography: Ultratrace Analysis with Neurobiological Samples,O. Orwar,"Optimization of argon ion laser-induced (UV multiline) fluorescence detection in liquid chromatography (1 mm i.d. columns) was performed aiming for ultratrace determination of o-phthalaldehyde/β-mercaptoethanol-labeled neuroactive amino acids and peptides. The derivatives are shown to undergo facile photochemical decomposition with photodestruction quantum yields, Φ D , in the range 0.009-0.14. In agreement with theoretical predictions, the optimal signal-to-noise (S/N) ratio was obtained at laser irradiances much lower than are required for prompt fluorescence saturation",,
-ESCMID* guideline for the diagnosis and management of Candida diseases 2012: patients with HIV infection or AIDS.,O. Lortholary,"Mucosal candidiasis is frequent in immunocompromised HIV-infected highly active antiretroviral (HAART) naive patients or those who have failed therapy. Mucosal candidiasis is a marker of progressive immune deficiency. Because of the frequently marked and prompt immune reconstitution induced by HAART, there is no recommendation for primary antifungal prophylaxis of mucosal candidiasis in the HIV setting in Europe, although it has been evidenced as effective in the pre-HAART era. Fluconazole remains the first line of therapy for both oropharyngeal candidiasis and oesophageal candidiasis and should be preferred to itraconazole oral solution (or capsules when not available) due to fewer side effects. For patients who still present with fluconazole-refractory mucosal candidiasis, oral treatment with any other azole should be preferred based on precise Candida species identification and susceptibility testing results in addition to the optimization of HAART when feasible. For vaginal candidiasis, topical therapy is preferred.","{'model': 'tldr@v2.0.0', 'text': 'There is no recommendation for primary antifungal prophylaxis of mucosal candidiasis in the HIV setting in Europe, although it has been evidenced as effective in the pre-HAART era.'}",http://www.clinicalmicrobiologyandinfection.com/article/S1198743X14607680/pdf
-Advantages and Disadvantages of Mammography Screening,S. Heywang-Köbrunner,"Mammography screening is the only method presently considered appropriate for mass screening of asymptomatic women. Its frequent use, however, warrants diligent analysis of potential side effects. Radiation risk is far below the natural yearly risk of breast cancer and should not be used as an argument against screening. False-positive calls lead to additional imaging or histopathological assessment, mainly percutaneous breast biopsy. These measures are tolerated and accepted fairly well. Their number is limited by strict quality assurance and constant training. Interval cancers represent a limitation of breast screening that should prompt further research for optimization. Evaluation of overdiagnosis is a highly debated topic in the literature. According to the probably most realistic available calculations, overdiagnosis is acceptable as it is compensated by the potential mortality reduction. Nonetheless, this potential side effect warrants optimal adjustment of therapy to the patient’s individual risk. The mortality reduction seen in randomized studies was confirmed by results from national screening programs. A recent case referent study indicated that improvements in mortality reduction run parallel to improved mammographic techniques. Use of less aggressive therapies is another valuable effect of screening. Awareness of potential problems, strict quality assurance, and further research should help to further develop screening programs.","{'model': 'tldr@v2.0.0', 'text': 'The mortality reduction seen in randomized studies was confirmed by results from national screening programs, and a recent case referent study indicated that improvements in mortality reduction run parallel to improved mammographic techniques.'}",https://www.karger.com/Article/Pdf/329005
-Anticipated and unanticipated difficult airway management,Zeping Xu,"Purpose of review Management of difficult airway is far from optimal despite of continuous progress in science and technology. The purpose of this review is to summarize the current research in the field and bring readers up to date. Recent findings New technologies for intubation make providers more confident to handle difficult airways, but there is lack of evidence indicating the reduction in incidence of ‘cannot intubate cannot ventilate (CICV)’. Optimization of mask ventilation should reduce the incidence of difficult mask ventilation but it is greatly underappreciated. Even optimization of preoxygenation is not directly associated with any decreased incidence of difficult airway, but it prolongs time of safe apnea oxygenation; therefore, is likely to improve the outcome of the patients if CICV occurs. Summary Improvement of managing difficult airway relies on optimized mask ventilation, utilization of the appropriate tools for intubation, maximization of the safe apnea oxygenation time, prompt surgical airway in response to severe hypoxia in case effective noninvasive interventions are not achievable. It seems that a simplified and concise algorithm of difficult airway management needs to be established in order to enable providers to easily remember and execute.","{'model': 'tldr@v2.0.0', 'text': 'Improvement of managing difficult airway relies on optimized mask ventilation, utilization of the appropriate tools for intubation, maximization of the safe apnea oxygenation time, prompt surgical airway in response to severe hypoxia in case effective noninvasive interventions are not achievable.'}",
-The approach to the adult with newly diagnosed adrenal insufficiency.,W. Arlt,"Adrenal insufficiency, primarily presenting as an adrenal crisis, is a life-threatening emergency and requires prompt therapeutic management including fluid resuscitation and stress dose hydrocortisone administration. Primary adrenal insufficiency is most frequently caused by autoimmune adrenalitis, and hypothalamic-pituitary tumors represent the most frequent cause of secondary adrenal insufficiency. However, the exact underlying diagnosis needs to be confirmed by a stepwise diagnostic approach, with an open eye for other differential diagnostic possibilities. Chronic replacement therapy with glucocorticoids and, in primary adrenal insufficiency, mineralocorticoids requires careful monitoring. However, current replacement strategies still require optimization as evidenced by recent studies demonstrating significantly impaired subjective health status and increased mortality in patients with primary and secondary adrenal insufficiency. Future studies will have to explore the potential of dehydroepiandrosterone replacement and modified delayed-release hydrocortisone to improve the prospects of patients with adrenal insufficiency.","{'model': 'tldr@v2.0.0', 'text': 'Future studies will have to explore the potential of dehydroepiandrosterone replacement and modified delayed-release hydrocortisone to improve the prospects of patients with adrenal insufficiency.'}",https://academic.oup.com/jcem/article-pdf/94/4/1059/10806508/jcem1059.pdf
-End-stage renal disease: symptom management and advance care planning.,Nina R. O’Connor,"The prevalence of end-stage renal disease continues to increase, and dialysis is offered to older and more medically complex patients. Pain is problematic in up to one-half of patients receiving dialysis and may result from renal and nonrenal etiologies. Opioids can be prescribed safely, but the patient's renal function must be considered when selecting a drug and when determining the dosage. Fentanyl and methadone are considered the safest opioids for use in patients with end-stage renal disease. Nonpain symptoms are common and affect quality of life. Phosphate binders, ondansetron, and naltrexone can be helpful for pruritus. Fatigue can be managed with treatment of anemia and optimization of dialysis, but persistent fatigue should prompt screening for depression. Ondansetron, metoclopramide, and haloperidol are effective for uremia-associated nausea. Nondialytic management may be preferable to dialysis initiation in older patients and in those with additional life-limiting illnesses, and may not significantly decrease life expectancy. Delaying dialysis initiation is also an option. Patients with end-stage renal disease should have advance directives, including documentation of situations in which they would no longer want dialysis.","{'model': 'tldr@v2.0.0', 'text': 'Patients with end-stage renal disease should have advance directives, including documentation of situations in which they would no longer want dialysis, and may not significantly decrease life expectancy.'}",
-SDSLc: a multi-target domain-specific compiler for stencil computations,P. Rawat,"Stencil computations are at the core of applications in a number of scientific computing domains. We describe a domain-specific language for regular stencil computations that allows specification of the computations in a concise manner. We describe a multi-target compiler for this DSL, which generates optimized code for GPUa, FPGAs, and multi-core processors with short-vector SIMD instruction sets, considering both low-order and high-order stencil computations. The hardware differences between these three types of architecture prompt different optimization strategies for the compiler. We evaluate the domain-specific compiler using a number of benchmarks on CPU, GPU and FPGA platforms.","{'model': 'tldr@v2.0.0', 'text': 'A domain-specific language for regular stencil computations that allows specification of the computations in a concise manner is described and a multi-target compiler for this DSL is described, which generates optimized code for GPUa, FPGAs, and multi-core processors with short-vector SIMD instruction sets.'}",http://dl.acm.org/ft_gateway.cfm?id=2830025&type=pdf
-Detection of retinal hemorrhage from fundus images using ANFIS classifier and MRG segmentation,L. GodlinAtlas,"Diabetic Retinopathy is a medicinal condition in which the retina is impaired based on fluid breaks from veins into the retina. The occurrence of hemorrhages in the retina is the most prompt effect of diabetic retinopathy. The number and state of hemorrhages are used to display the significance of the disorder. This research paper analyzed hemorrhage detection in retinal fundus images using classifier and segmentation methods. All the database images into the pre-processing steps and some meaning full features are extracted from the images. Then ANFIS classifier utilized to normaland abnormal images, this abnormal category into the hemorrhage detection process with help of segmentation technique. Here Region growing (RG) with threshold optimization techniques are considered its known as Modified RG (MRG) to get the maximum accuracy in the hemorrhage segmenting process. As regards the threshold optimization, Grey Wolf Optimization (GWO) technique used, this proposed work compared to our existing work getting maximum accuracy, sensitivity and specificity performance metrics.","{'model': 'tldr@v2.0.0', 'text': 'This research paper analyzed hemorrhage detection in retinal fundus images using classifier and segmentation methods and proposed a new threshold optimization technique known as Modified RG (MRG) to get the maximum accuracy in the hemorrhage segmenting process.'}",https://www.alliedacademies.org/articles/detection-of-retinal-hemorrhage-from-fundus-images-using-anfis-classifier-and-mrg-segmentation.pdf
-Optimal Feature Selection for Chronic Kidney Disease Classification using Deep Learning Classifier,K. Shankar,"Chronic Kidney Disease (CKD) is an increasing failure of kidney function leading to kidney failure over the years. The disease settles down and hence makes its diagnosis difficult. Analyzing CKD stages from standard office visit records can assist in premature recognition of the disease and prompt auspicious mediation. Hereby, we propose a methodology using inspired optimization model and learning procedure to classify CKD. The proposed method selects applicable features of kidney data with the help of Ant Lion Optimization (ALO) technique to choose optimal features for the classification process. After that, we sort the CKD data based on chosen features by utilizing Deep Neural Network (DNN). Performance comparison indicates that our proposed model accomplishes better classification accuracy, precision, F-measure, sensitivity measures when compared with other data mining classifiers.","{'model': 'tldr@v2.0.0', 'text': 'Performance comparison indicates that the proposed methodology using inspired optimization model and learning procedure to classify CKD accomplishes better classification accuracy, precision, F-measure, sensitivity measures when compared with other data mining classifiers.'}",
-Modelling the size and skill-mix of hospital nursing teams,PR Harper,,"{'model': 'tldr@v2.0.0', 'text': 'The work, as presented here, extends the PROMPT functionality to consider in more detail workforce issues and suggests that it is cost beneficial to increase the number of permanently employed nurses to account for fluctuations in demand and corresponding high costs of temporary (agency) nurses.'}",
-Self-Organization Towards Reduced Cost and Energy Per Bit for Future Emerging Radio Technologies - SONNET,S. Mumtaz,"Self-organizing network (SON) is seen as playing a pivotal role towards reducing the management costs of networks for the exposure of 5G and beyond 5G networks. SON is capable of collecting information from the network, so as to perform self- configuration, self-optimization, self-healing, etc., so as to reduce the operation cost through less human involvement, and to optimize the service quality through robust and prompt network optimization. SONNET aims to drive further cost savings in the way networks are managed today by amplifying further the coverage zone of SON within the network. We believe that key technologies such as network sharing and Coordinated Multipoint (CoMP) can benefit from SON technology solutions. Hence, this paper will give the overview of SONNET project (http://sonnet- project.eu/) by defining the state of art technologies (CoMP, network sharing, SON Context-aware) and innovations brought by SONNET. Moreover, this paper will also highlight the 5G use cases recently standardized by 3GPP, which act as a baseline for novel SONNET use cases proposed in this paper. The selected use cases will be used for demonstration purposes with the aid of Ray tracer and System Level Simulator.","{'model': 'tldr@v2.0.0', 'text': 'The 5G use cases recently standardized by 3GPP, which act as a baseline for novel SONNET use cases proposed in this paper, are highlighted.'}",
-iTimerC 2.0: Fast incremental timing and CPPR analysis,Pei-Yu Lee,"To achieve timing closure, performance-driven optimizations are repeatedly performed throughout the modern IC design flow. Along with these optimization operations, how to incrementally update timing information efficiently and accurately becomes a crucial task for fast turnaround time. On the other hand, to avoid wasteful over-optimization, clock path pessimism should be removed during timing analysis. In order to provide prompt timing information without over-pessimism during iterative optimizations, in this paper, we aim at fast incremental timing and CPPR analysis. We present two delicate techniques, lazy evaluation and lazy propagation, to avoid redundant updates. Our experiments are conducted on the benchmark suite released by TAU 2015 timing analysis contest. Experimental results show that our timer delivers the best results in terms of accuracy, runtime, and memory over all participating teams.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims at fast incremental timing and CPPR analysis, and presents two delicate techniques, lazy evaluation and lazy propagation, to avoid redundant updates.'}",
-A Heuristics Approach for Classroom Scheduling Using Genetic Algorithm Technique,Izah Rafidah Ahmad,"Reshuffling and arranging classroom based on the capacity of the audience, complete facilities, lecturing time and many more may lead to a complexity of classroom scheduling. While trying to enhance the productivity in classroom planning, this paper proposes a heuristic approach for timetabling optimization. A new algorithm was produced to take care of the timetabling problem in a university. The proposed of heuristics approach will prompt a superior utilization of the accessible classroom space for a given time table of courses at the university. Genetic Algorithm through Java programming languages were used in this study and aims at reducing the conflicts and optimizes the fitness. The algorithm considered the quantity of students in each class, class time, class size, time accessibility in each class and lecturer who in charge of the classes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed of heuristics approach will prompt a superior utilization of the accessible classroom space for a given time table of courses at the university.'}",
-Defining a research agenda in Value Driven Design: Questions that need to be asked,D. Soban,"Value driven design is an innovative design process that utilizes the optimization of a system level value function to determine the best possible design. This contrasts with more traditional systems engineering techniques, which rely on satisfying requirements to determine the design solution. While ?design for value? is intuitively acceptable, the transformation of value driven design concepts into practical tools and methods for its application is challenging. This, coupled with the growing popularity of value-centric design philosophies, has led to a proposed research agenda in value driven design. This research agenda asks fundamental questions about the design philosophy and attempts to identify areas of significant challenge. The research agenda is meant to stimulate discussion in the field, as well as prompt research that will lead to the development of tools and methodologies that will facilitate the application of value driven design and further the state of the art.",,
-A history of boron neutron capture therapy of brain tumours. Postulation of a brain radiation dose tolerance limit.,D. Slatkin,"Boron neutron capture therapy (BNCT) is a form of radiation therapy mediated by the short-range (less than 10 microns) energetic alpha (4He) and lithium-7 (7Li) ionizing particles that result from the prompt disintegration by slow neutrons of the stable (nonradioactive) nucleus boron-10 (10B). Recent advances in radiobiological and toxicological evaluation of tumour-affinitive boron-containing drugs and in optimization of the energies of neutrons in the incident beam have spurred interest in BNCT. This article presents a history of BNCT that emphasizes studies in the USA. A new dosimetric analysis of the 1959-1961 clinical trials of BNCT at Brookhaven National Laboratory is also presented. This analysis yields an acute radiation dose tolerance limit estimate of approximately 10 Gy-Eq to the capillary endothelium of human basal ganglia from BNCT. (Gy-Eq: Gray-equivalent, or relative biological effectiveness of a radiation component multiplied by the physical dose of the component (Gy), summed over the component kinds of radiation.)","{'model': 'tldr@v2.0.0', 'text': 'A new dosimetric analysis of the 1959-1961 clinical trials of BNCT at Brookhaven National Laboratory yields an acute radiation dose tolerance limit estimate of approximately 10 Gy-Eq to the capillary endothelium of human basal ganglia from BnCT.'}",
-Multidetector CT and three-dimensional CT angiography for suspected vascular trauma of the extremities.,E. Fishman,"The evolution of computed tomography (CT) from four to 16 to 64 sections since its inception in the late 1970s has led to more widespread use of this imaging modality in the emergent setting. CT angiography has become a crucial diagnostic technique for identifying vascular injury in the trauma patient. Regardless of the nature of the traumatic injury (eg, stab wound, gunshot wound, injury from a motor vehicle accident), use of multidetector CT with two-dimensional (2D) reformation and three-dimensional (3D) rendering allows visualization of injury to bone, muscle, and vasculature. The radiologist should be familiar with the indications for CT angiography, optimization of current multidetector CT acquisition protocols, utility of 2D and 3D displays, and CT findings in the presence of vascular injury to ensure prompt diagnosis and treatment.","{'model': 'tldr@v2.0.0', 'text': 'The radiologist should be familiar with the indications for CT angiography, optimization of current multidetector CT acquisition protocols, utility of 2D and 3D displays, and CT findings in the presence of vascular injury to ensure prompt diagnosis and treatment.'}",
-Human Parent-Child Relationships from an Evolutionary Perspective,H. Keller,"In this article, a proximate conception of parental investment for humans is developed. Parenting investment is introduced as a part of life histories, specifying the optimization of reproductive success in terms of inclusive fitness. Contextual, parental, and children's characteristics that influence investment decisions are specified. We can distinguish between four systems of parenting: primary care, body stimulation, body contact, and face-to-face interaction, which prompt different socialization experiences with different psychological consequences. Parental behaviors within these different systems may vary in terms of the expression of warmth and contingency. Parenting is described here as an intergenerational transmission mechanism of psychological characteristics as well as reproductive styles.",,
-Simultaneous immunoaffinity column cleanup and HPLC analysis of aflatoxins and ochratoxin A in Spanish bee pollen.,R. García-Villanova,"Bee pollen is a major substrate for mycotoxins growth when no prompt and adequate drying is performed by the beekeeper after collection by bees. Regulatory limits for aflatoxins and ochratoxin A are currently in force in the European Union for a rising list of foodstuffs, but not for this. An immunoaffinity column cleanup process has been applied prior to the analysis of aflatoxins B(1), B(2), G(1), and G(2) and ochratoxin A (OTA). Optimization of the HPLC conditions has involved both a gradient elution and a wavelength program for the separation and fluorimetric quantitation of all five mycotoxins at their maximum excitation and emission values of wavelength in a single run. The higher limit of detection (mug/kg) was 0.49 for OTA and 0.20 for aflatoxin B(1). Repeatability (RSDr) at the lower limit tested ranged from 9.85% for OTA to 6.23% for aflatoxin G(2), and recoveries also at the lower spiked level were 73% for OTA and 81% for aflatoxin B(1). None of the 20 samples assayed showed quantifiable values for the five mycotoxins.","{'model': 'tldr@v2.0.0', 'text': 'None of the 20 samples assayed showed quantifiable values for the five mycotoxins, but the higher limit of detection (mug/kg) was 0.49 for OTA and 0.20 for aflatoxin B(1).'}",
-An agent-based approach for managing symbiotic simulation of semiconductor assembly and test operation,M. Low,"The rapid changing business environment of high-tech asset intensive enterprises such as semiconductor manufacturing constantly drives production managers to look for better solutions to improve the manufacturing process. Simulation, though identified to be the most appropriate technique to generate and test out possible execution plans, suffers from long cycle-time in the process of model update, analysis and verification. It is thus very difficult to carry out prompt ""what-if' analysis to respond to abrupt changes in these systems. Symbiotic simulation systems have been proposed as a way of solving this problem by having the simulation and the physical system interact in a mutually beneficial manner. In this paper, we describe our work in developing a prototype proof-of-concept symbiotic simulation system that employs software agents in the monitoring, optimization and control of a semiconductor assembly and test operation.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the work in developing a prototype proof-of-concept symbiotic simulation system that employs software agents in the monitoring, optimization and control of a semiconductor assembly and test operation.'}",
-Problems in Signal-to-Noise Ratio for Attenuation Correction in High Resolution PET,M. Dahlbom,"In high resolution PET, the number of events required for a high signal-to-noise ratio in the measured attenuation correction is an order of magnitude higlher (50-100 million) than that required for the emission image. This is due to the large number of elements in the attenuation correction sinogram (e.g. 50,000) and the large attenuation factors through the thicker sections of the body (e.g. 50). Another major contribution to noise in high resolution PET is contamination of true coincidences by accidentals and their subtraction from the prompt coincidences. In this work three methods which reduce noise in the measured attenuation are evaluated. The methods are: (1) optimization of energy threshold, (2) processing of random sinogram before subtraction and (3) spatial averaging of the tranismission sinogram before performing attenuation correction.",,
-Fault location estimation based on matching the simulated and recorded waveforms using genetic algorithms,M. Kezunovic,"Prompt and accurate location of the faults in a large-scale transmission system is critical when system reliability is considered and usually is the first step in the system restoration. The accuracy of fault location estimation essentially depends on the information available. In this paper, the fault location estimation is mathematically formulated as an optimization problem of which the fault location and fault resistances are unknown variables. An efficient genetic algorithm-based searching scheme is developed for obtaining a solution that is globally optimal.",,
-"Characteristics of relativistic solar cosmic rays during the event of December 13, 2006",E. V. Vashenyuk,,,
-A WEB-BASED ENVIRONMENT FOR DOCUMENTATION AND SHARING OF ENGINEERING DESIGN KNOWLEDGE,Justin Rockwell,"This paper presents the foundation for a collaborative Web-based environment for improving communication by formally defining a platform for documentation and sharing of engineering design knowledge throughout the entire design process. In this work an ontological structure is utilized to concisely define a set of individual engineering concepts. This set of modular ontologies link together to create a flexible, yet consistent, product development knowledge-base. The resulting infrastructure uniquely enables the information stored within the knowledge-base to be readily inspectable and computable, thus allowing for design tools that reason on the information to assist designers and automate design processes. A case study of the structural optimization of a transfer plate for an aerospace circuit breaker is presented to demonstrate implementation and usefulness of the knowledge framework. The results indicate that the ontological knowledge-base can be used to prompt engineers to document important product development information, increase understanding of the design process, provide a means to intuitively retrieve information, and seamlessly access distributed information.© 2008 ASME","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the ontological knowledge-base can be used to prompt engineers to document important product development information, increase understanding of the design process, provide a means to intuitively retrieve information, and seamlessly access distributed information.'}",https://dr.lib.iastate.edu/bitstreams/f33fbf78-2699-4329-ba73-eb0eb605254f/download
-The surgical application of point‐of‐care haemostasis and platelet function testing,K. J. Dickinson,"Disordered coagulation complicates many diseases and their treatments, often predisposing to haemorrhage. Conversely, patients with cardiovascular disease who demonstrate antiplatelet resistance may be at increased thromboembolic risk. Prompt identification of these patients facilitates optimization of haemostatic dysfunction. Point‐of‐care (POC) tests are performed ‘near patient’ to provide a rapid assessment of haemostasis and platelet function.","{'model': 'tldr@v2.0.0', 'text': 'POC tests are performed ‘near patient’ to provide a rapid assessment of haemostasis and platelet function in patients with cardiovascular disease who demonstrate antiplatelet resistance.'}",https://academic.oup.com/bjs/article-pdf/95/11/1317/36709516/bjs6359.pdf
-Preference Heterogeneity in Relation to Museum Services,U. Colombino,"Prevailing trends in the management of European museums underline the importance of additional museum services in fostering and encouraging the optimization of cultural assets while facilitating the collection of resources necessary for conservation. This paper considers the case of the archaeological site of Paestum (Salerno) and presents an analysis of individual preferences in relation to specific policies of cultural heritage management, each characterized by the supply of different museum services. Since the diversity of these services can prompt different individual preferences, the analysis allows for heterogeneity of parameters among individuals.",,http://www3.unisa.it/uploads/2551/105_dp.pdf
-Acute mesenteric ischemia.,R. J. Stoney,"Acute mesenteric ischemia resulting from specific etiologies is a treatable vascular emergency. It requires a high index of suspicion and prompt, aggressive resuscitation and diagnostic maneuvers to determine the specific underlying cause. This will allow specific directed surgical revascularization, optimization of cardiac performance, or correction of a hypercoagulable state, all directed at maximizing the amount of functional bowel that is salvageable. These strategies are the cornerstones for a successful outcome in this life-threatening vascular catastrophe.","{'model': 'tldr@v2.0.0', 'text': 'Acute mesenteric ischemia resulting from specific etiologies is a treatable vascular emergency that requires a high index of suspicion and prompt, aggressive resuscitation and diagnostic maneuvers to determine the specific underlying cause.'}",
-Modeling a neutron-rich nuclei source,M. Mirea,,,
-Fluorescence polarization is a useful technology for reagent reduction in assay miniaturization.,T. J. Kowski,"The use of fluorescence polarization (FP) has increased significantly in the development of sensitive and robust assays for high throughput screening of chemical compound libraries during the past few years. In this study, we show that FP is a useful assay miniaturization technology for reagent reduction during high throughput screening. We developed and optimized several FP assays for binding to estrogen receptor alpha and two protein kinases with an assay volume of 100 microl. Without any re-optimization, a consistent signal window was maintained in 384- or 1536-well format when the assay volume varied from 2.5-100 microl at constant concentrations of all assay components. In contrast, the signal window decreased with decreasing assay volume at constant reagent concentration in the protein kinase C scintillation proximity assay (SPA) and prompt fluorescence assay. In addition, the effect of evaporation on the signal window was minimal for the FP assays. Our study suggests that FP is superior to SPA and prompt fluorescence in terms of reagent reduction in the miniaturized assay format.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that FP is superior to SPA and prompt fluorescence in terms of reagent reduction in the miniaturized assay format and the effect of evaporation on the signal window was minimal for the FP assays.'}",
-Designing Cross-Sectional Population Pharmacokinetic Studies: Implications for Pediatric and Animal Studies,C. D. Jones,"AbstractSampling constraints and ethical concerns are two of the issues that prompt the conduct of pediatric/animal cross-sectional pharmacokinetic studies. A simulation study was carried out to investigate the effect of optimization of design features (arrangement of concentrations (n) in time and sample size (N)) on the accuracy and precision of parameter (especially intersubject/interanimal variability) estimation in cross-sectional population pharmacokinetic studies involving drugs administered by single or multiple intravenous bolus input(s). Drugs exhibiting one and two compartment (i.e., 1 CMT and 2 CMT) pharmacokinetics were investigated using N of 20 to 100 and 30 to 150, respectively, n was optimized, using the profile (block) randomized sampling design, for accurate and precise estimation of population pharmacokinetic parameters. N of 50 and 80 was found to be adequate for the estimation of clearance (CL) and its variability ((ωCL) needed for dosage optimization) for the 1 CMT and 2 CMT drugs, ...","{'model': 'tldr@v2.0.0', 'text': 'A simulation study was carried out to investigate the effect of optimization of design features on the accuracy and precision of parameter estimation in cross-sectional population pharmacokinetic studies involving drugs administered by single or multiple intravenous bolus input(s).'}",
-A Method for Generating a Control Rod Program for Boiling Water Reactors,T. Kawai,"The OPROD computer code has been developed to generate a long-term control rod program, a series of control rod patterns that optimize a cycle length within various operational constraints. In the algorithm, the optimization problem is decomposed into two hierarchies. In the inner loop, a time-invariant target power distribution is assumed, and a control rod pattern is determined so as to best fit the power distribution to the target within the constraints at each burnup step. The target is then improved in the outer loop to achieve a longer cycle length. The code consists of two major parts: a three- dimensional boiling water reactor (BWR) core simulator and MAP, the method of approximate programming. It readily generates a long-term control rod program of BWRs without trial search by core-management engineers. The OPROD has therefore facilitated prompt response to varying operating conditions and the investigation of a conflicting relationship between the thermal limitation and the cycle length. (auth)","{'model': 'tldr@v2.0.0', 'text': 'The OPROD has facilitated prompt response to varying operating conditions and the investigation of a conflicting relationship between the thermal limitation and the cycle length.'}",
-Home telemonitoring of respiratory activity and heart rate variability in chronic heart failure patients: the challenge of the home or hospital in heart failure project,G. Pinna,"Nocturnal respiratory disorders and depressed heart rate variability are known predictors of poor prognosis in chronic heart failure (CHF) patients. Intermittent monitoring of cardiorespiratory signals while the patient is at home might thus allow early identification of clinical deterioration and prompt optimization of treatment, leading to reduced hospitalizations and mortality and improved quality of life. Within the European Community multicenter trial HHH (Home or Hospital in Heart Failure), we are testing a novel low-cost system for 24-hour recording of cardiorespiratory signals, suitable to be self-managed by the patient at home, with transmission of acquired data through standard telephone lines to the medical/nursing staff. Preliminary results from 24 CHF patients enrolled so far indicate that monthly home telemonitoring is feasible and the compliance is high.","{'model': 'tldr@v2.0.0', 'text': 'Preliminary results from 24 CHF patients enrolled so far indicate that monthly home telemonitoring is feasible and the compliance is high, and a novel low-cost system for 24-hour recording of cardiorespiratory signals is tested, suitable to be self-managed by the patient at home.'}",http://www.cinc.org/Proceedings/2003/pdf/197.pdf
-Engineering Robust Ag‐Decorated Polydopamine Nano‐Photothermal Platforms to Combat Bacterial Infection and Prompt Wound Healing,Xiaoliang Qi,"Polydopamine (PDA) nanoparticles have emerged as an attractive biomimetic photothermal agent in photothermal antibacterial therapy due to their ease of synthesis, good biodegradability, long‐term safety, and excellent photostability. However, the therapeutic effects of PDA nanoparticles are generally limited by the low photothermal conversion efficiency (PCE). Herein, PDA@Ag nanoparticles are synthesized via growing Ag on the surface of PDA nanoparticles and then encapsulated into a cationic guar gum (CG) hydrogel network. The optimized CG/PDA@Ag platform exhibits a high PCE (38.2%), which is more than two times higher than that of pure PDA (16.6%). More importantly, the formulated CG/PDA@Ag hydrogel with many active groups can capture and kill bacteria through effective interactions between hydrogel and bacteria, thereby benefiting the antibacterial effect. As anticipated, the designed CG/PDA@Ag system combined the advantages of PDA@Ag nanoparticles (high PCE) and hydrogel (preventing aggregation of PDA@Ag nanoparticles and possessing inherent antibacterial ability) is demonstrated to have superior antibacterial efficacy both in vitro and in vivo. This study develops a facile approach to boost the PCE of PDA for photothermal antibacterial therapy, providing a significant step forward in advancing the application of PDA nano‐photothermal agents.","{'model': 'tldr@v2.0.0', 'text': 'A facile approach to boost the PCE of Pda for photothermal antibacterial therapy is developed, providing a significant step forward in advancing the application of PDA nano‐photothermal agents.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/advs.202106015
-SpeechPrompt: An Exploration of Prompt Tuning on Generative Spoken Language Model for Speech Processing Tasks,Kai-Wei Chang,"Speech representations learned from Self-supervised learning (SSL) models can benefit various speech processing tasks. However, utilizing SSL representations usually requires fine-tuning the pre-trained models or designing task-specific downstream models and loss functions, causing much memory usage and human labor. Recently, prompting in Natural Language Processing (NLP) has been found to be an efficient technique to leverage pre-trained language models (LMs). Specifically, prompt tuning optimizes a limited number of task-specific parameters with a fixed pre-trained model; as a result, only a small set of parameters is needed to be stored for each task. Prompt tuning improves computation and memory efficiency by leveraging the pre-trained LM's prediction ability. Nevertheless, such a paradigm is little studied in the speech community. We report in this paper the first exploration of the prompt tuning paradigm for speech processing tasks based on Generative Spoken Language Model (GSLM). Experiment results show that the prompt tuning technique achieves competitive performance in speech classification tasks with fewer trainable parameters than fine-tuning specialized downstream models. We further study the technique in challenging sequence generation tasks. Prompt tuning also demonstrates its potential, while the limitation and possible research directions are discussed in this paper. The source code is available on https://github.com/ga642381/SpeechPrompt.","{'model': 'tldr@v2.0.0', 'text': 'The first exploration of the prompt tuning paradigm for speech processing tasks based on Generative Spoken Language Model (GSLM) is reported, demonstrating competitive performance in speech classification tasks with fewer trainable parameters than fine-tuning specialized downstream models.'}",
-Late Prompt Tuning: A Late Prompt Could Be Better Than Many Prompts,Xiangyang Liu,"Prompt tuning is a parameter-efficient tuning (PETuning) method for utilizing pre-trained models (PTMs) that simply prepends a soft prompt to the input and only optimizes the prompt to adapt PTMs to downstream tasks. Although it is parameter- and deployment-efficient, its performance still lags behind other state-of-the-art PETuning methods. Besides, the training cost of prompt tuning is not significantly reduced due to the back-propagation through the entire model. Through empirical analyses, we shed some light on the lagging performance of prompt tuning and recognize a trade-off between the propagation distance from label signals to the inserted prompt and the influence of the prompt on model outputs. Further, we present Late Prompt Tuning (LPT) that inserts a late prompt into an intermediate layer of the PTM instead of the input layer or all layers. The late prompt is obtained by a neural prompt generator conditioned on the hidden states before the prompt insertion layer and therefore is instance-dependent. Through extensive experimental results across various tasks and PTMs, we show that LPT can achieve competitive performance to full model tuning and other PETuning methods under both full-data and few-shot scenarios while possessing faster training speed and lower memory cost.","{'model': 'tldr@v2.0.0', 'text': 'Late Prompt Tuning (LPT) is presented that can achieve competitive performance to full model tuning and other PETuning methods under both full-data and few-shot scenarios while possessing faster training speed and lower memory cost.'}",http://arxiv.org/pdf/2210.11292
-KiPT: Knowledge-injected Prompt Tuning for Event Detection,Haochen Li,"Event detection aims to detect events from the text by identifying and classifying event triggers (the most representative words). Most of the existing works rely heavily on complex downstream networks and require sufficient training data. Thus, those models may be structurally redundant and perform poorly when data is scarce. Prompt-based models are easy to build and are promising for few-shot tasks. However, current prompt-based methods may suffer from low precision because they have not introduced event-related semantic knowledge (e.g., part of speech, semantic correlation, etc.). To address these problems, this paper proposes a Knowledge-injected Prompt Tuning (KiPT) model. Specifically, the event detection task is formulated into a condition generation task. Then, knowledge-injected prompts are constructed using external knowledge bases, and a prompt tuning strategy is leveraged to optimize the prompts. Extensive experiments indicate that KiPT outperforms strong baselines, especially in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The event detection task is formulated into a condition generation task, knowledge-injected prompts are constructed using external knowledge bases, and a prompt tuning strategy is leveraged to optimize the prompts.'}",
-Highly Efficient Oxidation of Propane at Low Temperature over a Pt-Based Catalyst by Optimization Support.,Zhen‐Feng Huang,"Pt-based catalysts have attracted widespread attention in environmental protection applications, especially in the catalytic destruction of light alkane pollutants. However, developing a satisfying platinum catalyst with high activity, excellent water-resistance, and practical suitability for hydrocarbon combustion at low temperature is challenging. In this study, the Pt catalyst supported on the selected Nb2O5 oxide exhibited an efficient catalytic activity in propane oxidation and exceeded that of most catalysts reported in the literature. More importantly, the Pt/Nb2O5 catalyst maintained excellent activity and durability even after high-temperature aging at 700 °C and under harsh working conditions, such as a certain degree of moisture, high space velocity, and composite pollutants. The excellent performance of the Pt/Nb2O5 catalyst was attributed to the abundant metallic Pt species stabilized on the surface of Nb2O5, which prompted the C-H bond dissociation ability as the rate-determining step. Furthermore, propane was initially activated via oxidehydrogenation and followed the acrylate species path as a more efficient propane oxidation path on the Pt/Nb2O5 surface. Overall, Pt/Nb2O5 can be considered a promising catalyst for the catalytic oxidation of alkanes from industrial sources and could provide inspiration for designing superb catalysts for the oxidation of light alkanes.",,
-Making Pre-trained Language Models End-to-end Few-shot Learners with Contrastive Prompt Tuning,Ziyun Xu,"Pre-trained Language Models (PLMs) have achieved remarkable performance for various language understanding tasks in IR systems, which require the fine-tuning process based on labeled training data. For low-resource scenarios, prompt-based learning for PLMs exploits prompts as task guidance and turns downstream tasks into masked language problems for effective few-shot fine-tuning. In most existing approaches, the high performance of prompt-based learning heavily relies on handcrafted prompts and verbalizers, which may limit the application of such approaches in real-world scenarios. To solve this issue, we present CP-Tuning, an end-to-end Contrastive Prompt Tuning framework for fine-tuning PLMs without any manual engineering of task-specific prompts and verbalizers. It is integrated with the task-invariant continuous prompt encoding technique with fully trainable prompt parameters. We further propose the pair-wise cost-sensitive contrastive learning procedure to optimize the model in order to achieve verbalizer-free class mapping and enhance the task-invariance of prompts. It explicitly learns to distinguish different classes and makes the decision boundary smoother by assigning different costs to easy and hard cases. Experiments over a variety of language understanding tasks and different PLMs show that CP-Tuning outperforms state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'CP-Tuning is presented, an end-to-end Contrastive Prompt Tuning framework for fine-tuning PLMs without any manual engineering of task-specific prompts and verbalizers, and is integrated with the task-invariant continuous prompt encoding technique with fully trainable prompt parameters.'}",https://arxiv.org/pdf/2204.00166
-LPT: Long-tailed Prompt Tuning for Image Classification,Bowen Dong,"For long-tailed classification, most works often pretrain a big model on a large-scale dataset, and then fine-tune the whole model for adapting to long-tailed data. Though promising, fine-tuning the whole pretrained model tends to suffer from high cost in computation and deployment of different models for different tasks, as well as weakened generalization ability for overfitting to certain features of long-tailed data. To alleviate these issues, we propose an effective Long-tailed Prompt Tuning method for long-tailed classification. LPT introduces several trainable prompts into a frozen pretrained model to adapt it to long-tailed data. For better effectiveness, we divide prompts into two groups: 1) a shared prompt for the whole long-tailed dataset to learn general features and to adapt a pretrained model into target domain; and 2) group-specific prompts to gather group-specific features for the samples which have similar features and also to empower the pretrained model with discrimination ability. Then we design a two-phase training paradigm to learn these prompts. In phase 1, we train the shared prompt via supervised prompt tuning to adapt a pretrained model to the desired long-tailed domain. In phase 2, we use the learnt shared prompt as query to select a small best matched set for a group of similar samples from the group-specific prompt set to dig the common features of these similar samples, then optimize these prompts with dual sampling strategy and asymmetric GCL loss. By only fine-tuning a few prompts while fixing the pretrained model, LPT can reduce training and deployment cost by storing a few prompts, and enjoys a strong generalization ability of the pretrained model. Experiments show that on various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves comparable performance than previous whole model fine-tuning methods, and is more robust to domain-shift.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that on various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves comparable performance than previous whole model fine-tuning methods, and is more robust to domain-shift.'}",http://arxiv.org/pdf/2210.01033
-Multi-Prompt Alignment for Multi-source Unsupervised Domain Adaptation,Haoran Chen,"Most existing methods for unsupervised domain adaptation (UDA) rely on a shared network to extract domain-invariant features. However, when facing multiple source domains, optimizing such a network involves updating the parameters of the entire network, making it both computationally expensive and challenging, particularly when coupled with min-max objectives. Inspired by recent advances in prompt learning that adapts high-capacity models for downstream tasks in a computationally economic way, we introduce Multi-Prompt Alignment (MPA), a simple yet efficient framework for multi-source UDA. Given a source and target domain pair, MPA first trains an individual prompt to minimize the domain gap through a contrastive loss. Then, MPA denoises the learned prompts through an auto-encoding process and aligns them by maximizing the agreement of all the reconstructed prompts. Moreover, we show that the resulting subspace acquired from the auto-encoding process can easily generalize to a streamlined set of target domains, making our method more efficient for practical usage. Extensive experiments show that MPA achieves state-of-the-art results on three popular datasets with an impressive average accuracy of 54.1% on DomainNet.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by recent advances in prompt learning that adapts high-capacity models for downstream tasks in a computationally economic way, Multi-Prompt Alignment (MPA) is introduced, a simple yet efficient framework for multi-source UDA.'}",http://arxiv.org/pdf/2209.15210
-Eliciting Knowledge from Pretrained Language Models for Prototypical Prompt Verbalizer,Yinyi Wei,,"{'model': 'tldr@v2.0.0', 'text': 'This paper focuses on eliciting knowledge from pretrained language models and proposes a prototypical prompt verbalizer for prompt-tuning, which optimizes models by contrastive learning.'}",https://arxiv.org/pdf/2201.05411
-Fine-grained Retrieval Prompt Tuning,Shijie Wang,"Fine-grained object retrieval aims to learn discriminative representation to retrieve visually similar objects. However, existing top-performing works usually impose pairwise similarities on the semantic embedding spaces or design a localization sub-network to continually fine-tune the entire model in limited data scenarios, thus resulting in convergence to suboptimal solutions. In this paper, we develop Fine-grained Retrieval Prompt Tuning (FRPT), which steers a frozen pre-trained model to perform the fine-grained retrieval task from the perspectives of sample prompting and feature adaptation. Specifically, FRPT only needs to learn fewer parameters in the prompt and adaptation instead of fine-tuning the entire model, thus solving the issue of convergence to suboptimal solutions caused by fine-tuning the entire model. Technically, a discriminative perturbation prompt (DPP) is introduced and deemed as a sample prompting process, which amplifies and even exaggerates some discriminative elements contributing to category prediction via a content-aware inhomogeneous sampling operation. In this way, DPP can make the fine-grained retrieval task aided by the perturbation prompts close to the solved task during the original pre-training. Thereby, it preserves the generalization and discrimination of representation extracted from input samples. Besides, a category-specific awareness head is proposed and regarded as feature adaptation, which removes the species discrepancies in features extracted by the pre-trained model using category-guided instance normalization. And thus, it makes the optimized features only include the discrepancies among subcategories. Extensive experiments demonstrate that our FRPT with fewer learnable parameters achieves the state-of-the-art performance on three widely-used fine-grained datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops Fine-grained Retrieval Prompt Tuning (FRPT), which steers a frozen pre-trained model to perform the fine- grained retrieval task from the perspectives of sample prompting and feature adaptation, thus solving the issue of convergence to suboptimal solutions caused by fine-tuning the entire model.'}",http://arxiv.org/pdf/2207.14465
-Exploring Low-dimensional Intrinsic Task Subspace via Prompt Tuning,Yujia Qin,"Why can pre-trained language models (PLMs) 001 learn universal representations and effectively 002 adapt to broad NLP tasks differing a lot super- 003 ficially? In this work, we empirically find evi- 004 dence indicating that the adaptations of PLMs 005 to various few-shot tasks can be reparameter- 006 ized as optimizing only a few free parame- 007 ters in a unified low-dimensional intrinsic task 008 subspace , which may help us understand why 009 PLMs could easily adapt to various NLP tasks 010 with small-scale data. To find such a subspace 011 and examine its universality, we propose an 012 analysis pipeline called intrinsic prompt tun- 013 ing (IPT). Specifically, we resort to the re- 014 cent success of prompt tuning and decompose 015 the soft prompts of multiple NLP tasks into 016 the same low-dimensional nonlinear subspace, 017 then we learn to adapt the PLM to unseen data 018 or tasks by only tuning parameters in this sub- 019 space. In the experiments, we study diverse 020 few-shot NLP tasks and surprisingly find that 021 in a 5 -dimensional subspace found with 100 022 tasks, by only tuning 5 free parameters, we 023 can recover 87% and 65% of the full prompt 024 tuning performance for 100 seen tasks (using 025 different training data) and 20 unseen tasks, re- 026 spectively, showing great generalization abil- 027 ity of the found intrinsic task subspace. Be- 028 sides being an analysis tool, IPT could further 029 bring practical benefits, such as improving the 030 prompt tuning stability. 031","{'model': 'tldr@v2.0.0', 'text': 'Empirically, there is evidence indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parame- 007 ters in a unified low-dimensional intrinsic task 008 subspace, which may help to understand why 009 PLMs could easily adapt to various NLP tasks 010 with small-scale data.'}",
-Exploring Universal Intrinsic Task Subspace via Prompt Tuning,Yujia Qin,"Why can pre-trained language models (PLMs) learn universal representations and effectively adapt to broad NLP tasks differing a lot superficially? In this work, we empirically find evidence indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parameters in a unified low-dimensional intrinsic task subspace, which may help us understand why PLMs could easily adapt to various NLP tasks with small-scale data. To find such a subspace and examine its universality, we propose an analysis pipeline called intrinsic prompt tuning (IPT). Specifically, we resort to the recent success of prompt tuning and decompose the soft prompts of multiple NLP tasks into the same low-dimensional nonlinear subspace, then we learn to adapt the PLM to unseen data or tasks by only tuning parameters in this subspace. In the experiments, we study diverse few-shot NLP tasks and surprisingly find that in a 250-dimensional subspace found with 100 tasks, by only tuning 250 free parameters, we can recover 97% and 83% of the full prompt tuning performance for 100 seen tasks (using different training data) and 20 unseen tasks, respectively, showing great generalization ability of the found intrinsic task subspace. Besides being an analysis tool, IPT could further help us improve the prompt tuning stability.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is empirically found indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parameters in a unified low-dimensional intrinsic task subspace, which may help to understand why PLMs could easily adapt to various NLP tasks with small-scale data.'}",
-How to Design the Perfect Prompt: A Linguistic Approach to Prompt Design in Automotive Voice Assistants – An Exploratory Study,Anna-Maria Meck,"In-vehicle voice user interfaces (VUIs) are becoming increasingly popular while needing to handle more and more complex functions. While many guidelines exist in terms of dialog design, a methodical and encompassing approach to prompt design is absent in the scientific landscape. The present work closes this gap by providing such an approach in form of linguistic-centered research. By extracting syntactical, lexical, and grammatical parameters from a German contemporary grammar, we examine how their respective manifestations affect users’ perception of a given system output across different prompt types. Through exploratory studies with a total of 1,206 participants, we provide concrete best practices to optimize and refine the design of VUI prompts. Based on these best practices, three superordinate user needs regarding prompt design can be identified: a) a suitable level of (in)formality, b) a suitable level of complexity/simplicity, and c) a suitable level of (im)mediacy.","{'model': 'tldr@v2.0.0', 'text': 'By extracting syntactical, lexical, and grammatical parameters from a German contemporary grammar, this work examines how their respective manifestations affect users’ perception of a given system output across different prompt types.'}",
-Improving ChatGPT Prompt for Code Generation,Chao Liu,"Automated code generation can be a powerful technique for software development, significantly reducing developers' efforts and time required to create new code by generating it automatically based on requirements. Recently, OpenAI's language model ChatGPT has emerged as a powerful tool for generating human-like responses to a wide range of textual inputs (i.e., prompts), including those related to code generation. However, the effectiveness of ChatGPT for code generation is not well understood, and the generation performance could be heavily influenced by the choice of prompt. To answer these questions, we conducted experiments using the CodeXGlue dataset to evaluate ChatGPT's capabilities for two code generation tasks, including text-to-code and code-to-code generation. We designed prompts by leveraging the chain-of-thought strategy with multi-step optimizations. Our results showed that by carefully designing prompts to guide ChatGPT, the generation performance can be improved substantially. We also analyzed the factors that influenced the prompt design and provided insights that could guide future research.","{'model': 'tldr@v2.0.0', 'text': 'By carefully designing prompts to guide ChatGPT, the generation performance can be improved substantially, and the factors that influenced the prompt design were analyzed and provided insights that could guide future research.'}",http://arxiv.org/pdf/2305.08360
-Prompt Engineering with ChatGPT: A Guide for Academic Writers,L. Giray,,,
-Dynamic Prompting: A Unified Framework for Prompt Tuning,Xianjun Yang,"It has been demonstrated that the art of prompt tuning is highly effective in efficiently extracting knowledge from pretrained foundation models, encompassing pretrained language models (PLMs), vision pretrained models, and vision-language (V-L) models. However, the efficacy of employing fixed soft prompts with a predetermined position for concatenation with inputs for all instances, irrespective of their inherent disparities, remains uncertain. Variables such as the position, length, and representations of prompts across diverse instances and tasks can substantially influence the performance of prompt tuning. In this context, we provide a theoretical analysis, which reveals that optimizing the position of the prompt to encompass the input can capture additional semantic information that traditional prefix or postfix prompt tuning methods fail to capture. Building upon our analysis, we present a unified dynamic prompt (DP) tuning strategy that dynamically determines different factors of prompts based on specific tasks and instances. To accomplish this, we employ a lightweight learning network with Gumble-Softmax, allowing us to learn instance-dependent guidance. Experimental results underscore the significant performance improvement achieved by dynamic prompt tuning across a wide range of tasks, including NLP tasks, vision recognition tasks, and vision-language tasks. Furthermore, we establish the universal applicability of our approach under full-data, few-shot, and multitask scenarios. Codes are available at https://github.com/Xianjun-Yang/DPT.","{'model': 'tldr@v2.0.0', 'text': 'A theoretical analysis is provided, which reveals that optimizing the position of the prompt to encompass the input can capture additional semantic information that traditional prefix or postfix prompt tuning methods fail to capture.'}",http://arxiv.org/pdf/2303.02909
-Exploring Sparse Visual Prompt for Domain Adaptive Dense Prediction,Senqiao Yang,"The visual prompts have provided an efficient manner in addressing visual cross-domain problems. In previous works, Visual Domain Prompt (VDP) first introduces domain prompts to tackle the classification Test-Time Adaptation (TTA) problem by warping image-level prompts on the input and fine-tuning prompts for each target domain. However, since the image-level prompts mask out continuous spatial details in the prompt-allocated region, it will suffer from inaccurate contextual information and limited domain knowledge extraction, particularly when dealing with dense prediction TTA problems. To overcome these challenges, we propose a novel Sparse Visual Domain Prompts (SVDP) approach, which holds minimal trainable parameters (e.g., 0.1\%) in the image-level prompt and reserves more spatial information of the input. To better apply SVDP in extracting domain-specific knowledge, we introduce the Domain Prompt Placement (DPP) method to adaptively allocates trainable parameters of SVDP on the pixels with large distribution shifts. Furthermore, recognizing that each target domain sample exhibits a unique domain shift, we design Domain Prompt Updating (DPU) strategy to optimize prompt parameters differently for each sample, facilitating efficient adaptation to the target domain. Extensive experiments were conducted on widely-used TTA and continual TTA benchmarks, and our proposed method achieves state-of-the-art performance in both semantic segmentation and depth estimation tasks.","{'model': 'tldr@v2.0.0', 'text': 'To better apply SVDP in extracting domain-specific knowledge, the Domain Prompt Placement (DPP) method is introduced to adaptively allocates trainable parameters of SVDP on the pixels with large distribution shifts, and recognizing that each target domain sample exhibits a unique domain shift, the DPU strategy is designed to optimize prompt parameters differently for each sample, facilitating efficient adaptation to the target domain.'}",
-StyleDiffusion: Prompt-Embedding Inversion for Text-Based Editing,Senmao Li,"A significant research effort is focused on exploiting the amazing capacities of pretrained diffusion models for the editing of images. They either finetune the model, or invert the image in the latent space of the pretrained model. However, they suffer from two problems: (1) Unsatisfying results for selected regions, and unexpected changes in nonselected regions. (2) They require careful text prompt editing where the prompt should include all visual objects in the input image. To address this, we propose two improvements: (1) Only optimizing the input of the value linear network in the cross-attention layers, is sufficiently powerful to reconstruct a real image. (2) We propose attention regularization to preserve the object-like attention maps after editing, enabling us to obtain accurate style editing without invoking significant structural changes. We further improve the editing technique which is used for the unconditional branch of classifier-free guidance, as well as the conditional one as used by P2P. Extensive experimental prompt-editing results on a variety of images, demonstrate qualitatively and quantitatively that our method has superior editing capabilities than existing and concurrent works.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experimental prompt-editing results, demonstrate qualitatively and quantitatively that the proposed attention regularization to preserve the object-like attention maps after editing has superior editing capabilities than existing and concurrent works.'}",https://arxiv.org/pdf/2303.15649
-Efficient Transfer Learning for Visual Tasks via Continuous Optimization of Prompts,Jonathan Conder,,,
-A Simple Zero-shot Prompt Weighting Technique to Improve Prompt Ensembling in Text-Image Models,J. Allingham,"Contrastively trained text-image models have the remarkable ability to perform zero-shot classification, that is, classifying previously unseen images into categories that the model has never been explicitly trained to identify. However, these zero-shot classifiers need prompt engineering to achieve high accuracy. Prompt engineering typically requires hand-crafting a set of prompts for individual downstream tasks. In this work, we aim to automate this prompt engineering and improve zero-shot accuracy through prompt ensembling. In particular, we ask""Given a large pool of prompts, can we automatically score the prompts and ensemble those that are most suitable for a particular downstream dataset, without needing access to labeled validation data?"". We demonstrate that this is possible. In doing so, we identify several pathologies in a naive prompt scoring method where the score can be easily overconfident due to biases in pre-training and test data, and we propose a novel prompt scoring method that corrects for the biases. Using our proposed scoring method to create a weighted average prompt ensemble, our method outperforms equal average ensemble, as well as hand-crafted prompts, on ImageNet, 4 of its variants, and 11 fine-grained classification benchmarks, all while being fully automatic, optimization-free, and not requiring access to labeled validation data.","{'model': 'tldr@v2.0.0', 'text': 'Using the proposed scoring method to create a weighted average prompt ensemble, the method outperforms equal average ensemble, as well as hand-crafted prompts, on ImageNet, 4 of its variants, and 11 fine-grained classification benchmarks, all while being fully automatic, optimization-free, and not requiring access to labeled validation data.'}",https://arxiv.org/pdf/2302.06235
-DRPT: Disentangled and Recurrent Prompt Tuning for Compositional Zero-Shot Learning,Xiaocheng Lu,"Compositional Zero-shot Learning (CZSL) aims to recognize novel concepts composed of known knowledge without training samples. Standard CZSL either identifies visual primitives or enhances unseen composed entities, and as a result, entanglement between state and object primitives cannot be fully utilized. Admittedly, vision- language models (VLMs) could naturally cope with CZSL through tuning prompts, while uneven entanglement leads prompts to be dragged into local optimum. In this paper, we take a further step to introduce a novel Disentangled and Recurrent Prompt Tuning framework termed DRPT to better tap the potential of VLMs in CZSL. Specifically, the state and object primitives are deemed as learnable tokens of vocabulary embedded in prompts and tuned on seen compositions. Instead of jointly tuning state and object, we devise a disentangled and recurrent tuning strategy to suppress the traction force caused by entanglement and gradually optimize the token parameters, leading to a better prompt space. Notably, we develop a progressive fine-tuning procedure that allows for incremental updates to the prompts, optimizing the object first, then the state, and vice versa. Meanwhile, the optimization of state and object is independent, thus clearer features can be learned to further alleviate the issue of entangling misleading optimization. Moreover, we quantify and analyze the entanglement in CZSL and supplement entanglement rebalancing optimization schemes. DRPT surpasses representative state-of-the-art methods on extensive benchmark datasets, demonstrating superiority in both accuracy and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel Disentangled and Recurrent Prompt Tuning framework termed DRPT to better tap the potential of VLMs in CZSL and develops a progressive fine-tuning procedure that allows for incremental updates to the prompts, optimizing the object first, then the state, and vice versa.'}",http://arxiv.org/pdf/2305.01239
-RePrompt: Automatic Prompt Editing to Refine AI-Generative Art Towards Precise Expressions,Yunlong Wang,"Generative AI models have shown impressive ability to produce images with text prompts, which could benefit creativity in visual art creation and self-expression. However, it is unclear how precisely the generated images express contexts and emotions from the input texts. We explored the emotional expressiveness of AI-generated images and developed RePrompt, an automatic method to refine text prompts toward precise expression of the generated images. Inspired by crowdsourced editing strategies, we curated intuitive text features, such as the number and concreteness of nouns, and trained a proxy model to analyze the feature effects on the AI-generated image. With model explanations of the proxy model, we curated a rubric to adjust text prompts to optimize image generation for precise emotion expression. We conducted simulation and user studies, which showed that RePrompt significantly improves the emotional expressiveness of AI-generated images, especially for negative emotions.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by crowdsourced editing strategies, this work curated intuitive text features, such as the number and concreteness of nouns, and trained a proxy model to analyze the feature effects on the AI-generated image.'}",https://arxiv.org/pdf/2302.09466
-Feasibility of quasi-prompt PET-based range verification in proton therapy,I. Ozoemelam,"Compared to photon therapy, proton therapy allows a better conformation of the dose to the tumor volume with reduced radiation dose to co-irradiated tissues. In vivo verification techniques including positron emission tomography (PET) have been proposed as quality assurance tools to mitigate proton range uncertainties. Detection of differences between planned and actual dose delivery on a short timescale provides a fast trigger for corrective actions. Conventional PET-based imaging of 15O (T1/2 = 2 min) and 11C (T1/2 = 20 min) distributions precludes such immediate feedback. We here present a demonstration of near real-time range verification by means of PET imaging of 12N (T1/2 = 11 ms). PMMA and graphite targets were irradiated with a 150 MeV proton pencil beam consisting of a series of pulses of 10 ms beam-on and 90 ms beam-off. Two modules of a modified Siemens Biograph mCT PET scanner (21 × 21 cm2 each), installed 25 cm apart, were used to image the beam-induced PET activity during the beam-off periods. The modifications enable the detectors to be switched off during the beam-on periods. 12N images were reconstructed using planar tomography. Using a 1D projection of the 2D reconstructed 12N image, the activity range was obtained from a fit of the activity profile with a sigmoid function. Range shifts due to modified target configurations were assessed for multiples of the clinically relevant 108 protons per pulse (approximately equal to the highest intensity spots in the pencil beam scanning delivery of a dose of 1 Gy over a cubic 1 l volume). The standard deviation of the activity range, determined from 30 datasets obtained from three irradiations on PMMA and graphite targets, was found to be 2.5 and 2.6 mm (1σ) with 108 protons per pulse and 0.9 and 0.8 mm (1σ) with 109 protons per pulse. Analytical extrapolation of the results from this study shows that using a scanner with a solid angle coverage of 57%, with optimized detector switching and spot delivery times much smaller than the 12N half-life, an activity range measurement precision of 2.0 mm (1σ) and 1.3 mm (1σ) within 50 ms into an irradiation with 4 × 107 and 108 protons per pencil beam spot can be potentially realized. Aggregated imaging of neighboring spots or, if possible, increasing the number of protons for a few probe beam spots will enable the realization of higher precision range measurement.","{'model': 'tldr@v2.0.0', 'text': 'Analytical extrapolation of the results from this study shows that using a scanner with a solid angle coverage of 57%, with optimized detector switching and spot delivery times much smaller than the 12N half-life, an activity range measurement precision of 2.0 mm can be potentially realized.'}",
-"Optimizing of a question prompt list to improve communication about the heart failure trajectory in patients, families, and health care professionals",Lisa Hjelmfors,,"{'model': 'tldr@v2.0.0', 'text': 'This Question Prompt List has successfully been adapted into a Swedish version and a Dutch version and includes questions about the HF trajectory which patients, their families, and health care professionals perceived to be relevant for discussion in clinical practice.'}",https://bmcpalliatcare.biomedcentral.com/track/pdf/10.1186/s12904-020-00665-3
-Optimal Sink Node Placement in Large Scale Wireless Sensor Networks Based on Harris’ Hawk Optimization Algorithm,E. H. Houssein,"Large-scale wireless sensor network (LSWSN) is composed of a huge number of sensor nodes that are distributed in some region of interest (ROI), to sense and measure the environmental conditions like pressure, temperature, pollution levels, humidity, wind, and so on. The objective is to collect data for real-time monitoring so that appropriate actions can be taken promptly. One of the sensor nodes used in an LSWSN is called the sink node, which is responsible for processing and analyzing the collected information. It works as a station between the network sensor nodes and the administrator. Also, it is responsible for controlling the whole network. Determining the sink node location in an LSWSN is a challenging task, as it is crucial to the network lifetime, for keeping the network activity to the most possible extent. In this paper, the Harris’ hawks optimization (HHO) algorithm is employed to solve this problem and subsequently the Prim’s shortest path algorithm is used to reconstruct the network by making minimum transmission paths from the sink node to the rest of the sensor nodes. The performance of HHO is compared with other well-known algorithms such as particle swarm optimization (PSO), flower pollination algorithm (FPA), grey wolf optimizer (GWO), sine cosine algorithm (SCA), multi-verse optimizer (MVO), and whale optimization algorithm (WOA). The simulation results of different network sizes, with single and multiple sink nodes, show the superiority of the employed approach in terms of energy consumption and localization error, and ultimately prolonging the lifetime of the network in an efficacious way.","{'model': 'tldr@v2.0.0', 'text': 'The Harris’ hawks optimization (HHO) algorithm is employed and subsequently the Prim’s shortest path algorithm is used to reconstruct the network by making minimum transmission paths from the sink node to the rest of the sensor nodes.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/08966999.pdf
-Development and Optimization of Naringenin-Loaded Chitosan-Coated Nanoemulsion for Topical Therapy in Wound Healing,S. Akrawi,"The potential role of naringenin (NAR), a natural flavonoid, in the treatment of chronic wound has prompted the present research to deliver the drug in nanoemulsion (NE) form, where synergistic role of chitosan was achieved through development of chitosan-coated NAR NE (CNNE). The NE consisted of Capryol 90, Tween 20 and Transcutol P, which was fabricated by low-energy emulsification method to encapsulate NAR within the oil core. The optimization of the formulated NEs was performed using Box–Behnken statistical design to obtain crucial variable parameters that influence globule size, size distribution and surface charge. Finally, the optimized formulation was coated with different concentrations of chitosan and subsequently characterized in vitro. The size of the CNNE was found to be increased when the drug-loaded formulation was coated with chitosan. Controlled release characteristics depicted 67–81% release of NAR from the CNNE, compared to 89% from the NE formulation. Cytotoxicity study of the formulation was performed in vitro using fibroblast cell line (NIH-3T3), where no inhibition in proliferation of the cells was observed with CNNE. Finally, the wound healing potential of the CNNE was evaluated in an abrasion-created wound model in experimental animals where the animals were treated and compared histologically at 0 and 14 days. Significant improvement in construction of the abrasion wound was observed when the animals were treated with formulated CNNE, whereas stimulation of skin regeneration was depicted in the histological examination. Therefore, it could be summarized that the chitosan coating of the developed NAR NE is a potential platform to accelerate healing of wounds.","{'model': 'tldr@v2.0.0', 'text': 'The chitosan coating of the developed NAR NE is a potential platform to accelerate healing of wounds and was evaluated in an abrasion-created wound model in experimental animals where the animals were treated with formulated CNNE.'}",https://www.mdpi.com/1999-4923/12/9/893/pdf?version=1600585179
-Ram-pressure Stripping of a Kicked Hill Sphere: Prompt Electromagnetic Emission from the Merger of Stellar Mass Black Holes in an AGN Accretion Disk,B. McKernan,"Accretion disks around supermassive black holes (SMBHs) are promising sites for stellar mass black hole (BH) mergers due to mass segregation and merger acceleration by disk gas torques. Here we show that a gravitational-wave (GW) kick at BH merger causes ram-pressure stripping of gas within the BH Hill sphere. If RH ≥ H, the disk height, an off-center UV flare at aBH ∼ 103rg, emerges within tUV ∼ O(2 days)(aBH/103rg)(MSMBH/108M⊙)(vkick/102 km s−1) postmerger and lasts O(RH/vkick) ∼ O(5tUV). The flare emerges with luminosity O(1042erg s−1)(tUV/2days)−1(MHill/1M⊙)(vkick/102 km s−1)2. Active galactic nucleus optical/UV photometry is altered and asymmetric broad emission line profiles can develop after weeks. If RH < H, detectability depends on disk optical depth. Follow-up by large optical sky surveys is optimized for small GW error volumes and for Laser Interferometer Gravitational-Wave Observatory/Virgo triggers >50M⊙.",,https://iopscience.iop.org/article/10.3847/2041-8213/ab4886/pdf
-UAV Relay-Assisted Emergency Communications in IoT Networks: Resource Allocation and Trajectory Optimization,Dinh-Hieu Tran,"Unmanned aerial vehicle (UAV) communication has emerged as a prominent technology for emergency communications (e.g., natural disaster) in the Internet of Things (IoT) networks to enhance the ability of disaster prediction, damage assessment, and rescue operations promptly. A UAV can be deployed as a flying base station (BS) to collect data from time-constrained IoT devices and then transfer it to a ground gateway (GW). In general, the latency constraint at IoT devices and UAV’s limited storage capacity highly hinder practical applications of UAV-assisted IoT networks. In this paper, full-duplex (FD) radio is adopted at the UAV to overcome these challenges. In addition, half-duplex (HD) scheme for UAV-based relaying is also considered to provide a comparative study between two modes (viz., FD and HD). Herein, a device is considered to be successfully served if its data is collected by the UAV and conveyed to GW timely during flight time. In this context, we aim to maximize the number of served IoT devices by jointly optimizing bandwidth, power allocation, and the UAV trajectory while satisfying each device’s requirement and the UAV’s limited storage capacity. The formulated optimization problem is troublesome to solve due to its non-convexity and combinatorial nature. Towards appealing applications, we first relax binary variables into continuous ones and transform the original problem into a more computationally tractable form. By leveraging inner approximation framework, we derive newly approximated functions for non-convex parts and then develop a simple yet efficient iterative algorithm for its solutions. Next, we attempt to maximize the total throughput subject to the number of served IoT devices. Finally, numerical results show that the proposed algorithms significantly outperform benchmark approaches in terms of the number of served IoT devices and system throughput.","{'model': 'tldr@v2.0.0', 'text': 'Full-duplex (FD) radio is adopted at the UAV to overcome challenges of latency constraint at IoT devices and UAV’s limited storage capacity, and a simple yet efficient iterative algorithm is developed to maximize the total throughput subject to the number of served IoT devices.'}",https://ieeexplore.ieee.org/ielx7/7693/9731098/09522072.pdf
-Prompt Engineering for Text-Based Generative Art,J. Oppenlaender,"Text-based generative art has seen an explosion of interest in 2021. Online communities around text-based generative art as a novel digital medium have quickly emerged. This short paper identifies five types of prompt modifiers used by practitioners in the community of text-based generative art based on a 3-month ethnographic study on Twitter. The novel taxonomy of prompt modifiers provides researchers a conceptual starting point for investigating the practices of text-based generative art, but also may help practitioners of text-based generative art improve their images. The paper concludes with a discussion of research opportunities in the space of text-based generative art and the broader implications of prompt engineering from the perspective of human-AI interaction in future applications beyond the use case of text-based generative art.",,http://arxiv.org/pdf/2204.13988
-Prompting AI Art: An Investigation into the Creative Skill of Prompt Engineering,J. Oppenlaender,"Humankind is entering a novel era of creativity - an era in which anybody can synthesize digital content. The paradigm under which this revolution takes place is prompt-based learning (or in-context learning). This paradigm has found fruitful application in text-to-image generation where it is being used to synthesize digital images from zero-shot text prompts in natural language for the purpose of creating AI art. This activity is referred to as prompt engineering - the practice of iteratively crafting prompts to generate and improve images. In this paper, we investigate prompt engineering as a novel creative skill for creating prompt-based art. In three studies with participants recruited from a crowdsourcing platform, we explore whether untrained participants could 1) recognize the quality of prompts, 2) write prompts, and 3) improve their prompts. Our results indicate that participants could assess the quality of prompts and respective images. This ability increased with the participants' experience and interest in art. Participants further were able to write prompts in rich descriptive language. However, even though participants were specifically instructed to generate artworks, participants' prompts were missing the specific vocabulary needed to apply a certain style to the generated images. Our results suggest that prompt engineering is a learned skill that requires expertise and practice. Based on our findings and experience with running our studies with participants recruited from a crowdsourcing platform, we provide ten recommendations for conducting experimental research on text-to-image generation and prompt engineering with a paid crowd. Our studies offer a deeper understanding of prompt engineering thereby opening up avenues for research on the future of prompt engineering. We conclude by speculating on four possible futures of prompt engineering.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates prompt engineering as a novel creative skill for creating prompt-based art and suggests that prompt engineering is a learned skill that requires expertise and practice.'}",http://arxiv.org/pdf/2303.13534
-Review of Large Vision Models and Visual Prompt Engineering,Jiaqi Wang,"Visual prompt engineering is a fundamental technology in the field of visual and image Artificial General Intelligence, serving as a key component for achieving zero-shot capabilities. As the development of large vision models progresses, the importance of prompt engineering becomes increasingly evident. Designing suitable prompts for specific visual tasks has emerged as a meaningful research direction. This review aims to summarize the methods employed in the computer vision domain for large vision models and visual prompt engineering, exploring the latest advancements in visual prompt engineering. We present influential large models in the visual domain and a range of prompt engineering methods employed on these models. It is our hope that this review provides a comprehensive and systematic description of prompt engineering methods based on large visual models, offering valuable insights for future researchers in their exploration of this field.","{'model': 'tldr@v2.0.0', 'text': 'It is hoped that this review provides a comprehensive and systematic description of prompt engineering methods based on large visual models, offering valuable insights for future researchers in their exploration of this field.'}",http://arxiv.org/pdf/2307.00855
-Grimm in Wonderland: Prompt Engineering with Midjourney to Illustrate Fairytales,M. Ruskov,"The quality of text-to-image generation is continuously improving, yet the boundaries of its applicability are still unclear. In particular, refinement of the text input with the objective of achieving better results - commonly called prompt engineering - so far seems to have not been geared towards work with pre-existing texts. We investigate whether text-to-image generation and prompt engineering could be used to generate basic illustrations of popular fairytales. Using Midjourney v4, we engage in action research with a dual aim: to attempt to generate 5 believable illustrations for each of 5 popular fairytales, and to define a prompt engineering process that starts from a pre-existing text and arrives at an illustration of it. We arrive at a tentative 4-stage process: i) initial prompt, ii) composition adjustment, iii) style refinement, and iv) variation selection. We also discuss three reasons why the generation model struggles with certain illustrations: difficulties with counts, bias from stereotypical configurations and inability to depict overly fantastic situations. Our findings are not limited to the specific generation model and are intended to be generalisable to future ones.",,https://arxiv.org/pdf/2302.08961
-Prompt Engineering in Medical Education,Thomas F. Heston,"Artificial intelligence-powered generative language models (GLMs), such as ChatGPT, Perplexity AI, and Google Bard, have the potential to provide personalized learning, unlimited practice opportunities, and interactive engagement 24/7, with immediate feedback. However, to fully utilize GLMs, properly formulated instructions are essential. Prompt engineering is a systematic approach to effectively communicating with GLMs to achieve the desired results. Well-crafted prompts yield good responses from the GLM, while poorly constructed prompts will lead to unsatisfactory responses. Besides the challenges of prompt engineering, significant concerns are associated with using GLMs in medical education, including ensuring accuracy, mitigating bias, maintaining privacy, and avoiding excessive reliance on technology. Future directions involve developing more sophisticated prompt engineering techniques, integrating GLMs with other technologies, creating personalized learning pathways, and researching the effectiveness of GLMs in medical education.","{'model': 'tldr@v2.0.0', 'text': None}",https://www.mdpi.com/2813-141X/2/3/19/pdf?version=1693479951
-Just Tell Me: Prompt Engineering in Business Process Management,Kiran Busch,"GPT-3 and several other language models (LMs) can effectively address various natural language processing (NLP) tasks, including machine translation and text summarization. Recently, they have also been successfully employed in the business process management (BPM) domain, e.g., for predictive process monitoring and process extraction from text. This, however, typically requires fine-tuning the employed LM, which, among others, necessitates large amounts of suitable training data. A possible solution to this problem is the use of prompt engineering, which leverages pre-trained LMs without fine-tuning them. Recognizing this, we argue that prompt engineering can help bring the capabilities of LMs to BPM research. We use this position paper to develop a research agenda for the use of prompt engineering for BPM research by identifying the associated potentials and challenges.","{'model': 'tldr@v2.0.0', 'text': 'This position paper is used to develop a research agenda for the use of prompt engineering for BPM research by identifying the associated potentials and challenges.'}",http://arxiv.org/pdf/2304.07183
-"Multi-party Goal Tracking with LLMs: Comparing Pre-training, Fine-tuning, and Prompt Engineering",Angus Addlesee,"This paper evaluates the extent to which current LLMs can capture task-oriented multi-party conversations (MPCs). We have recorded and transcribed 29 MPCs between patients, their companions, and a social robot in a hospital. We then annotated this corpus for multi-party goal-tracking and intent-slot recognition. People share goals, answer each other’s goals, and provide other people’s goals in MPCs - none of which occur in dyadic interactions. To understand user goals in MPCs, we compared three methods in zero-shot and few-shot settings: we fine-tuned T5, created pre-training tasks to train DialogLM using LED, and employed prompt engineering techniques with GPT-3.5-turbo, to determine which approach can complete this novel task with limited data. GPT-3.5-turbo significantly outperformed the others in a few-shot setting. The ‘reasoning’ style prompt, when given 7% of the corpus as example annotated conversations, was the best performing method. It correctly annotated 62.32% of the goal tracking MPCs, and 69.57% of the intent-slot recognition MPCs. A ‘story’ style prompt increased model hallucination, which could be detrimental if deployed in safety-critical settings. We conclude that multi-party conversations still challenge state-of-the-art LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that multi-party conversations still challenge state-of-the-art LLMs and could be detrimental if deployed in safety-critical settings.'}",https://arxiv.org/pdf/2308.15231
-Real Estate Insights Unleashing the potential of ChatGPT in property valuation reports: the “Red Book” compliance Chain-of-thought (CoT) prompt engineering,K. Cheung,"PurposeThis viewpoint article explores the transformative capabilities of large language models (LLMs) like the Chat Generative Pre-training Transformer (ChatGPT) within the property valuation industry. It particularly accentuates the pivotal role of prompt engineering in facilitating valuation reporting and advocates for adopting the “Red Book” compliance Chain-of-thought (COT) prompt engineering as a gold standard for generating AI-facilitated valuation reports.Design/methodology/approachThe article offers a high-level examination of the application of LLMs in real estate research, highlighting the essential role of prompt engineering for future advancements in generative AI. It explores the collaborative dynamic between valuers and AI advancements, emphasising the importance of precise instructions and contextual cues in directing LLMs to generate accurate and reproducible valuation outcomes.FindingsIntegrating LLMs into property valuation processes paves the way for efficiency improvements and task automation, such as generating reports and drafting contracts. AI-facilitated reports offer unprecedented transparency and elevate client experiences. The fusion of valuer expertise with prompt engineering ensures the reliability and interpretability of valuation reports.Practical implicationsDelineating the types and versions of LLMs used in AI-generated valuation reports encourage the adoption of transparency best practices within the industry. Valuers, as expert prompt engineers, can harness the potential of AI to enhance efficiency, accuracy and transparency in the valuation process, delivering significant benefits to a broad array of stakeholders.Originality/valueThe article elucidates the substantial impact of prompt engineering in leveraging LLMs within the property industry. It underscores the importance of valuers training their unique GPT models, enabling customisation and reproducibility of valuation outputs. The symbiotic relationship between valuers and LLMs is identified as a key driver shaping the future of property valuations.",,
-Improving Formality-Sensitive Machine Translation Using Data-Centric Approaches and Prompt Engineering,Seugnjun Lee,"In this paper, we present the KU x Upstage team’s submission for the Special Task on Formality Control on Spoken Language Translation, which involves translating English into four languages with diverse grammatical formality markers. Our methodology comprises two primary components: 1) a language-specific data-driven approach, and 2) the generation of synthetic data through the employment of large-scale language models and empirically-grounded prompt engineering. By adapting methodologies and models to accommodate the unique linguistic properties of each language, we observe a notable enhancement in performance relative to the baseline, substantiating the heightened efficacy of data-driven approaches. Moreover, our devised prompt engineering strategy yields superior synthetic translation instances.","{'model': 'tldr@v2.0.0', 'text': 'The KU x Upstage team’s submission for the Special Task on Formality Control on Spoken Language Translation, which involves translating English into four languages with diverse grammatical formality markers, shows a notable enhancement in performance relative to the baseline, substantiating the heightened efficacy of data-driven approaches.'}",https://aclanthology.org/2023.iwslt-1.40.pdf
-Artificial intelligence prompt engineering as a new digital competence: Analysis of generative AI technologies such as ChatGPT,P. Korzyński,"Objective: The article aims to offer a thorough examination and comprehension of the challenges and pro‐ spects connected with artificial intelligence (AI) prompt engineering. Our research aimed to create a theoret‐ ical framework that would highlight optimal approaches in the field of AI prompt engineering. Research Design & Methods: This research utilized a narrative and critical literature review and established a conceptual framework derived from existing literature taking into account both academic and practitioner sources. This article should be regarded as a conceptual work that emphasizes the best practices in the domain of AI prompt engineering. Findings: Based on the conducted deep and extensive query of academic and practitioner literature on the subject, as well as professional press and Internet portals, we identified various insights for effective AI prompt engineering. We provide specific prompting strategies. Implications & Recommendations: The study revealed the profound implications of AI prompt engineering across various domains such as entrepreneurship, art, science, and healthcare. We demonstrated how the effective crafting of prompts can significantly enhance the performance of large language models (LLMs), gen‐ erating more accurate and contextually relevant results. Our findings offer valuable insights for AI practition‐ ers, researchers, educators, and organizations integrating AI into their operations, emphasizing the need to invest time and resources in prompt engineering. Moreover, we contributed the AI PROMPT framework to the field, providing clear and actionable guidelines for text‐to‐text prompt engineering. Contribution & Value Added: The value of this study lies in its comprehensive exploration of AI prompt engineer‐ ing as a digital competence. By building upon existing research and prior literature, this study aimed to provide a deeper understanding of the intricacies involved in AI prompt engineering and its role as a digital competence. Article",,https://eber.uek.krakow.pl/index.php/eber/article/view/2142/863
-Cases of EFL Secondary Students' Prompt Engineering Pathways to Complete a Writing Task with ChatGPT,D. Woo,"ChatGPT is a state-of-the-art (SOTA) chatbot. Although it has potential to support English as a foreign language (EFL) students' writing, to effectively collaborate with it, a student must learn to engineer prompts, that is, the skill of crafting appropriate instructions so that ChatGPT produces desired outputs. However, writing an appropriate prompt for ChatGPT is not straightforward for non-technical users who suffer a trial-and-error process. This paper examines the content of EFL students' ChatGPT prompts when completing a writing task and explores patterns in the quality and quantity of the prompts. The data come from iPad screen recordings of secondary school EFL students who used ChatGPT and other SOTA chatbots for the first time to complete the same writing task. The paper presents a case study of four distinct pathways that illustrate the trial-and-error process and show different combinations of prompt content and quantity. The cases contribute evidence for the need to provide prompt engineering education in the context of the EFL writing classroom, if students are to move beyond an individual trial-and-error process, learning a greater variety of prompt content and more sophisticated prompts to support their writing.",,https://arxiv.org/pdf/2307.05493
-Enhancing Automated Program Repair through Fine-tuning and Prompt Engineering,Rishov Paul,"Sequence-to-sequence models have been used to transform erroneous programs into correct ones when trained with a large enough dataset. Some recent studies also demonstrated strong empirical evidence that code review could improve the program repair further. Large language models, trained with Natural Language (NL) and Programming Language (PL), can contain inherent knowledge of both. In this study, we investigate if this inherent knowledge of PL and NL can be utilized to improve automated program repair. We applied PLBART and CodeT5, two state-of-the-art language models that are pre-trained with both PL and NL, on two such natural language-based program repair datasets and found that the pre-trained language models fine-tuned with datasets containing both code review and subsequent code changes notably outperformed each of the previous models. With the advent of code generative models like Codex and GPT-3.5-Turbo, we also performed zero-shot and few-shots learning-based prompt engineering to assess their performance on these datasets. However, the practical application of using LLMs in the context of automated program repair is still a long way off based on our manual analysis of the generated repaired codes by the learning models.","{'model': 'tldr@v2.0.0', 'text': 'This study applied PLBART and CodeT5, two state-of-the-art language models that are pre- trained with both PL and NL, on two such natural language-based program repair datasets and found that the pre-trained language models fine-tuned with datasets containing both code review and subsequent code changes notably outperformed each of the previous models.'}",
-"Supporting self-directed learning and self-assessment using TeacherGAIA, a generative AI chatbot application: Learning approaches and prompt engineering",Farhan Ali,"ABSTRACT Self-directed learning and self-assessment require student responsibility over learning needs, goals, processes, and outcomes. However, this student-led learning can be challenging to achieve in a classroom limited by a one-to-many teacher-led instruction. We, thus, have designed and prototyped a generative artificial intelligence chatbot application (GAIA), named TeacherGAIA, that can be used to asynchronously support students in their self-directed learning and self-assessment outside the classroom. We first identified diverse constructivist learning approaches that align with, and promote, student-led learning. These included knowledge construction, inquiry-based learning, self-assessment, and peer teaching. The in-context learning abilities of large language model (LLM) from OpenAI were then leveraged via prompt engineering to steer interactions supporting these different learning approaches. These interactions contrasted with ChatGPT, OpenAI’s chatbot which by default engaged in the traditional transmissionist mode of learning reminiscent of teacher-led instruction. Preliminary design, prompt engineering and prototyping suggested fidelity to the learning approaches, cognitive guidance, and social-emotional support, all of which were implemented in a generative AI manner without pre-specified rules or “hard-coding”. Other affordances of TeacherGAIA are discussed and future development outlined. We anticipate TeacherGAIA to be a useful application for teachers in facilitating self-directed learning and self-assessment among K-12 students.",,
-Prompting meaning: a hermeneutic approach to optimising prompt engineering with ChatGPT,Leah Henrickson,,,https://link.springer.com/content/pdf/10.1007/s00146-023-01752-8.pdf
-NCU-IISR: Prompt Engineering on GPT-4 to Stove Biological Problems in BioASQ 11b Phase B,Chun-Yu Hsueh,"In this paper, we present our system applied in BioASQ 11b phase b. We showcase prompt engineering strategies and outline our experimental steps. Building upon the success of ChatGPT/GPT-4 in answer generation and the field of biology, we developed a system that utilizes GPT-4 to answer biomedical questions. The system leverages OpenAI’s ChatCompletions API and combines Prompt Engineering methods to explore various prompts. In addition, we also attempted to incorporate GPT-4 into our system from last year, which combines a BERT-based model and BERTScore. However, the standalone GPT-4 method outperformed this approach by a large margin. Ultimately, in our submission, we adopted what we believe to be the optimal prompts and achieved the highest scores in the second batch.","{'model': 'tldr@v2.0.0', 'text': 'A system that utilizes GPT-4 to answer biomedical questions and leverages OpenAI’s ChatCompletions API and combines Prompt Engineering methods to explore various prompts and adopted what it believes to be the optimal prompts.'}",
-"Optimizing Mobile-Edge AI-Generated Everything (AIGX) Services by Prompt Engineering: Fundamental, Framework, and Case Study",Yinqiu Liu,"As the next-generation paradigm for content creation, AI-Generated Content (AIGC), i.e., generating content automatically by Generative AI (GAI) based on user prompts, has gained great attention and success recently. With the ever-increasing power of GAI, especially the emergence of Pretrained Foundation Models (PFMs) that contain billions of parameters and prompt engineering methods (i.e., finding the best prompts for the given task), the application range of AIGC is rapidly expanding, covering various forms of information for human, systems, and networks, such as network designs, channel coding, and optimization solutions. In this article, we present the concept of mobile-edge AI-Generated Everything (AIGX). Specifically, we first review the building blocks of AIGX, the evolution from AIGC to AIGX, as well as practical AIGX applications. Then, we present a unified mobile-edge AIGX framework, which employs edge devices to provide PFM-empowered AIGX services and optimizes such services via prompt engineering. More importantly, we demonstrate that suboptimal prompts lead to poor generation quality, which adversely affects user satisfaction, edge network performance, and resource utilization. Accordingly, we conduct a case study, showcasing how to train an effective prompt optimizer using ChatGPT and investigating how much improvement is possible with prompt engineering in terms of user experience, quality of generation, and network performance.","{'model': 'tldr@v2.0.0', 'text': 'A unified mobile-edge AIGX framework is presented, which employs edge devices to provide PFM-empowered AigX services and optimizes such services via prompt engineering and demonstrates that suboptimal prompts lead to poor generation quality, which adversely affects user satisfaction, edge network performance, and resource utilization.'}",https://arxiv.org/pdf/2309.01065
-Prompt Engineering as an Important Emerging Skill for Medical Professionals: Tutorial,B. Meskó,"Prompt engineering is a relatively new field of research that refers to the practice of designing, refining, and implementing prompts or instructions that guide the output of large language models (LLMs) to help in various tasks. With the emergence of LLMs, the most popular one being ChatGPT that has attracted the attention of over a 100 million users in only 2 months, artificial intelligence (AI), especially generative AI, has become accessible for the masses. This is an unprecedented paradigm shift not only because of the use of AI becoming more widespread but also due to the possible implications of LLMs in health care. As more patients and medical professionals use AI-based tools, LLMs being the most popular representatives of that group, it seems inevitable to address the challenge to improve this skill. This paper summarizes the current state of research about prompt engineering and, at the same time, aims at providing practical recommendations for the wide range of health care professionals to improve their interactions with LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper summarizes the current state of research about prompt engineering and aims at providing practical recommendations for the wide range of health care professionals to improve their interactions with LLMs.'}",
-Exploring the Intersection of Large Language Models and Agent-Based Modeling via Prompt Engineering,Edward Junprung,"The final frontier for simulation is the accurate representation of complex, real-world social systems. While agent-based modeling (ABM) seeks to study the behavior and interactions of agents within a larger system, it is unable to faithfully capture the full complexity of human-driven behavior. Large language models (LLMs), like ChatGPT, have emerged as a potential solution to this bottleneck by enabling researchers to explore human-driven interactions in previously unimaginable ways. Our research investigates simulations of human interactions using LLMs. Through prompt engineering, inspired by Park et al. (2023), we present two simulations of believable proxies of human behavior: a two-agent negotiation and a six-agent murder mystery game.","{'model': 'tldr@v2.0.0', 'text': 'This research investigates simulations of human interactions using large language models using LLMs and presents two simulations of believable proxies of human behavior: a two-agent negotiation and a six-agent murder mystery game.'}",https://arxiv.org/pdf/2308.07411
-The prompt engineering librarian,Brady Lund," Purpose In terms of training the public in prompt engineering skills, no single discipline or profession currently takes the lead, presenting an opportunity for professions like librarianship to step into this role. Librarians are already well-equipped to educate the public in a wide range of literacy skills and tasks, so prompt engineering may be a natural progression. The purpose of this paper is to examine the potential role of prompt engineering for library professionals. Design/methodology/approach Prompt engineering is the process of optimizing the text that is provided to an artificial intelligence (A)I model to ensure proper interpretation and the generation of relevant, detailed results. The field of prompt engineering is relatively young, evolving alongside the growth of large language models like ChatGPT and BARD. This conceptual paper will explore prompt engineering as a possible domain of expertise for librarians. Findings This paper delves into the world of prompt engineering, its alignment with the existing roles and expertise of librarians, and the potential emergence of a new role known as the “prompt engineering librarian,” akin to the well-established “information literacy librarian” role that has gained prominence in recent decades. Originality/value The significance of this work lies in exploring the synergy between prompt engineering and the traditional roles of librarians, highlighting the potential for a new and valuable profession in the form of prompt engineering librarians. This innovative concept could bridge the gap in AI literacy and facilitate more effective interactions with AI systems, contributing to the broader goal of AI accessibility and understanding. ",,
-Prompt Engineering For Students of Medicine and Their Teachers,Thomas F. Heston,"""Prompt Engineering for Students of Medicine and Their Teachers""brings the principles of prompt engineering for large language models such as ChatGPT and Google Bard to medical education. This book contains a comprehensive guide to prompt engineering to help both teachers and students improve education in the medical field. Just as prompt engineering is critical in getting good information out of an AI, it is also critical to get students to think and understand more deeply. The principles of prompt engineering that we have learned from AI systems have the potential to simultaneously revolutionize learning in the healthcare field. The book analyzes from multiple angles the anatomy of a good prompt for both AI models and students. The different types of prompts are examined, showing how each style has unique characteristics and applications. The principles of prompt engineering, applied properly, are demonstrated to be effective in teaching across the diverse fields of anatomy, physiology, pathology, pharmacology, and clinical skills. Just like ChatGPT and similar large language AI models, students need clear and detailed prompting in order for them to fully understand a topic. Using identical principles, a prompt that gets good information from an AI will also cause a student to think more deeply and accurately. The process of prompt engineering facilitates this process. Because each chapter contains multiple examples and key takeaways, it is a practical guide for implementing prompt engineering in the learning process. It provides a hands-on approach to ensure readers can immediately apply the concepts they learn",,https://arxiv.org/pdf/2308.11628
-Contextual stance classification using prompt engineering,Felipe Penhorate Carvalho de Fonseca,"This paper introduces a prompt-based method for few-shot learning addressing, as an application example, contextual stance classification, that is, the task of determining the attitude expressed by a given statement within a conversation thread with multiple points of view towards another statement. More specifically, we envisaged a method that uses the existing conversation thread (i.e., messages that are part of the test data) to create natural language prompts for few-shot learning with minimal reliance on training samples, whose preliminary results suggest that prompt engineering may be a competitive alternative to supervised methods both in terms of accuracy and development costs for the task at hand.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method that uses the existing conversation thread to create natural language prompts for few-shot learning with minimal reliance on training samples is introduced, whose preliminary results suggest that prompt engineering may be a competitive alternative to supervised methods both in terms of accuracy and development costs for the task at hand.'}",https://sol.sbc.org.br/index.php/stil/article/download/25435/25256
-Cheap-Fake Detection with LLM Using Prompt Engineering,Guangyang Wu,"The misuse of real photographs with conflicting image captions in news items is an example of the out-of-context (OOC) misuse of media. In order to detect OOC media, individuals must determine the accuracy of the statement and evaluate whether the triplet (i.e., the image and two captions) relates to the same event. This paper presents a novel learnable approach for detecting OOC media in ICME'23 Grand Challenge on Detecting Cheapfakes. The proposed method is based on the COSMOS structure, which assesses the coherence between an image and captions, as well as between two captions. We enhance the baseline algorithm by incorporating a Large Language Model (LLM), GPT3.5, as a feature extractor. Specifically, we propose an innovative approach to feature extraction utilizing prompt engineering to develop a robust and reliable feature extractor with GPT3.5 model. The proposed method captures the correlation between two captions and effectively integrates this module into the COSMOS baseline model, which allows for a deeper understanding of the relationship between captions. By incorporating this module, we demonstrate the potential for significant improvements in cheap-fakes detection performance. The proposed methodology holds promising implications for various applications such as natural language processing, image captioning, and text-to-image synthesis. Docker for submission is available at https://hub.docker.com/repository/docker/mulns/acmmmcheapfakes.","{'model': 'tldr@v2.0.0', 'text': 'An innovative approach to feature extraction utilizing prompt engineering to develop a robust and reliable feature extractor with GPT3.5 model is proposed, which captures the correlation between two captions and effectively integrates this module into the COSMOS baseline model, which allows for a deeper understanding of the relationship between captions.'}",https://arxiv.org/pdf/2306.02776
-PromptMagician: Interactive Prompt Engineering for Text-to-Image Creation,Yingchaojie Feng,"Generative text-to-image models have gained great popularity among the public for their powerful capability to generate high-quality images based on natural language prompts. However, developing effective prompts for desired images can be challenging due to the complexity and ambiguity of natural language. This research proposes PromptMagician, a visual analysis system that helps users explore the image results and refine the input prompts. The backbone of our system is a prompt recommendation model that takes user prompts as input, retrieves similar prompt-image pairs from DiffusionDB, and identifies special (important and relevant) prompt keywords. To facilitate interactive prompt refinement, PromptMagician introduces a multi-level visualization for the cross-modal embedding of the retrieved images and recommended keywords, and supports users in specifying multiple criteria for personalized exploration. Two usage scenarios, a user study, and expert interviews demonstrate the effectiveness and usability of our system, suggesting it facilitates prompt engineering and improves the creativity support of the generative text-to-image model.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes PromptMagician, a visual analysis system that helps users explore the image results and refine the input prompts, and introduces a multi-level visualization for the cross-modal embedding of the retrieved images and recommended keywords.'}",https://arxiv.org/pdf/2307.09036
-LogPrompt: Prompt Engineering Towards Zero-Shot and Interpretable Log Analysis,Yilun Liu,"Automated log analysis is crucial in modern software-intensive systems for ensuring reliability and resilience throughout software maintenance and engineering life cycles. Existing methods perform tasks such as log parsing and log anomaly detection by providing a single prediction value without interpretation. However, given the increasing volume of system events, the limited interpretability of analysis results hinders analysts' trust and their ability to take appropriate actions. Moreover, these methods require substantial in-domain training data, and their performance declines sharply (by up to 62.5%) in online scenarios involving unseen logs from new domains, a common occurrence due to rapid software updates. In this paper, we propose LogPrompt, a novel zero-shot and interpretable log analysis approach. LogPrompt employs large language models (LLMs) to perform zero-shot log analysis tasks via a suite of advanced prompt strategies tailored for log tasks, which enhances LLMs' performance by up to 107.5% compared with simple prompts. Experiments on nine publicly available evaluation datasets across two tasks demonstrate that LogPrompt, despite using no training data, outperforms existing approaches trained on thousands of logs by up to around 50%. We also conduct a human evaluation of LogPrompt's interpretability, with six practitioners possessing over 10 years of experience, who highly rated the generated content in terms of usefulness and readability (averagely 4.42/5). LogPrompt also exhibits remarkable compatibility with open-source and smaller-scale LLMs, making it flexible for practical deployment.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on nine publicly available evaluation datasets across two tasks demonstrate that LogPrompt, despite using no training data, outperforms existing approaches trained on thousands of logs by up to around 50%.'}",https://arxiv.org/pdf/2308.07610
-"The artificially intelligent entrepreneur: ChatGPT, prompt engineering, and entrepreneurial rhetoric creation",Cole E. Short,,,
-A Survey on Segment Anything Model (SAM): Vision Foundation Model Meets Prompt Engineering,Chaoning Zhang,"Segment anything model (SAM) developed by Meta AI Research has recently attracted significant attention. Trained on a large segmentation dataset of over 1 billion masks, SAM is capable of segmenting any object on a certain image. In the original SAM work, the authors turned to zero-short transfer tasks (like edge detection) for evaluating the performance of SAM. Recently, numerous works have attempted to investigate the performance of SAM in various scenarios to recognize and segment objects. Moreover, numerous projects have emerged to show the versatility of SAM as a foundation model by combining it with other models, like Grounding DINO, Stable Diffusion, ChatGPT, etc. With the relevant papers and projects increasing exponentially, it is challenging for the readers to catch up with the development of SAM. To this end, this work conducts the first yet comprehensive survey on SAM. This is an ongoing project and we intend to update the manuscript on a regular basis. Therefore, readers are welcome to contact us if they complete new works related to SAM so that we can include them in our next version.","{'model': 'tldr@v2.0.0', 'text': 'The first yet comprehensive survey on SAM is conducted, to show the versatility of SAM as a foundation model by combining it with other models, like Grounding DINO, Stable Diffusion, ChatGPT, etc.'}",http://arxiv.org/pdf/2306.06211
-The CLEAR path: A framework for enhancing information literacy through prompt engineering,Leo S. Lo,,,
-Prompt Engineering for Large Language Models,Andrew Gao,,,
-Exploring the Effectiveness of Prompt Engineering for Legal Reasoning Tasks,Fang Yu,",","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2023.findings-acl.858.pdf
-On Codex Prompt Engineering for OCL Generation: An Empirical Study,Seif Abukhalaf,"The Object Constraint Language (OCL) is a declarative language that adds constraints and object query expressions to Meta-Object Facility (MOF) models. OCL can provide precision and conciseness to UML models. Nevertheless, the unfamiliar syntax of OCL has hindered its adoption by software practitioners. LLMs, such as GPT-3, have made significant progress in many NLP tasks, such as text generation and semantic parsing. Similarly, researchers have improved on the downstream tasks by fine-tuning LLMs for the target task. Codex, a GPT-3 descendant by OpenAI, has been fine-tuned on publicly available code from GitHub and has proven the ability to generate code in many programming languages, powering the AI-pair programmer Copilot. One way to take advantage of Codex is to engineer prompts for the target downstream task. In this paper, we investigate the reliability of the OCL constraints generated by Codex from natural language specifications. To achieve this, we compiled a dataset of 15 UML models and 168 specifications from various educational resources. We manually crafted a prompt template with slots to populate with the UML information and the target task in the prefix format to complete the template with the generated OCL constraint. We used both zero- and few-shot learning methods in the experiments. The evaluation is reported by measuring the syntactic validity and the execution accuracy metrics of the generated OCL constraints. Moreover, to get insight into how close or natural the generated OCL constraints are compared to human-written ones, we measured the cosine similarity between the sentence embedding of the correctly generated and human-written OCL constraints. Our findings suggest that by enriching the prompts with the UML information of the models and enabling few-shot learning, the reliability of the generated OCL constraints increases. Furthermore, the results reveal a close similarity based on sentence embedding between the generated OCL constraints and the human-written ones in the ground truth, implying a level of clarity and understandability in the generated OCL constraints by Codex.","{'model': 'tldr@v2.0.0', 'text': 'The findings suggest that by enriching the prompts with the UML information of the models and enabling few-shot learning, the reliability of the generated OCL constraints increases, and a close similarity based on sentence embedding between the generatedOCL constraints and the human-written ones in the ground truth is revealed.'}",https://arxiv.org/pdf/2303.16244
-Abstractive Summarization Evaluation for Prompt Engineering,Shayak Chakraborty,,,
-Prompt Engineering for Narrative Choice Generation,Sarah Harmon,,,
-Plain Template Insertion: Korean-Prompt-Based Engineering for Few-Shot Learners,Jaehyung Seo,"Prompt-based learning is a method used for language models to interpret natural language by remembering the prior knowledge acquired and the training objective. Recent prompt-based few-shot learners have achieved superior performance by alleviating the catastrophic forgetting that occurs in pretrained language models. Few-shot learning contributes towards solving the data scarcity problem, an enormous challenge in AI systems and a significant consideration in natural language processing research. In spite of the significance of few-shot learning, research on Korean language-based few-shot learning is insufficient, and whether the prompt-based approach is appropriate for the Korean language has not been thoroughly verified. As a step toward realizing a Korean-prompt-based few-shot learner, we attempt to apply prompt engineering to the Korean language understanding benchmark dataset and introduce plain template insertion to overcome data scarcity in a more practical few-shot setting. The contributions of this study are as follows: (1) presumably, this is the first study to apply prompt-based few-shot learning to Korean benchmark datasets. With 32 few-shot settings, it improves performance by +14.88, +29.04, and +1.81 in the natural language inference, semantic textual similarity, and topic classification tasks. (2) We present prompt engineering, which merely inserts a plain template and increases data efficiency without training example selection, augmentation, reformulation, and retrieval. (3) Our approach is robust to the Korean prompt’s contextual information and sentence structure and is applicable to both hard- and soft-prompt.","{'model': 'tldr@v2.0.0', 'text': 'This study attempts to apply prompt engineering to the Korean language understanding benchmark dataset and introduces plain template insertion to overcome data scarcity in a more practical few-shot setting.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09913979.pdf
-Polyglot Prompt: Multilingual Multitask Prompt Training,Jinlan Fu,"This paper aims for a potential architectural improvement for multilingual learning and asks: Can different tasks from different languages be modeled in a monolithic framework, i.e. without any task/language-specific module? The benefit of achieving this could open new doors for future multilingual research, including allowing systems trained on low resources to be further assisted by other languages as well as other tasks. We approach this goal by developing a learning framework named Polyglot Prompting to exploit prompting methods for learning a unified semantic space for different languages and tasks with multilingual prompt engineering. We performed a comprehensive evaluation of 6 tasks, namely topic classification, sentiment classification, named entity recognition, question answering, natural language inference, and summarization, covering 24 datasets and 49 languages. The experimental results demonstrated the efficacy of multilingual multitask prompt-based learning and led to inspiring observations. We also present an interpretable multilingual evaluation methodology and show how the proposed framework, multilingual multitask prompt training, works. We release all datasets prompted in the best setting and code.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops a learning framework named Polyglot Prompting to exploit prompting methods for learning a unified semantic space for different languages and tasks with multilingual prompt engineering, and presents an interpretable multilingual evaluation methodology.'}",https://aclanthology.org/2022.emnlp-main.674.pdf
-Polyglot Prompt: Multilingual Multitask PrompTraining,Jinlan Fu,"This paper aims for a potential architectural breakthrough for multilingual learning and asks: could different tasks from different languages be modeled in a monolithic framework (without any task/language-specific module) ? The benefit of achieving this is not only that systems trained on low resources scenario can be assisted by more other languages and tasks, but opening new doors for future multilingual research. We approach this goal by developing a learning framework Polyglot Prompt , where prompting methods are introduced to learn a unified semantic space for different languages and tasks after proper multilingual prompt engineering . Experimentally, we perform a com-prehensive evaluation on 6 tasks (topic classification, sentiment classification, named entity recognition, question answering, natural language inference, summarization), 24 datasets, and 49 languages, which shows the efficacy of multilingual multitask prompting training and suggests several interesting observations. e.g., English prompts are polyglots since di-rectly applying them to task samples in other languages could result in a better improvement. We also present an interpretable multi-lingual evaluation methodology and show how the proposed framework, multilingual multi-task prompt training, works. We release all datasets prompted in the best setting 1 and will release our code soon. 2","{'model': 'tldr@v2.0.0', 'text': 'This paper develops a learning framework Polyglot Prompt, where prompting methods are introduced to learn a unified semantic space for different languages and tasks after proper multilingual prompt engineering, and presents an interpretable multi-lingual evaluation methodology.'}",https://arxiv.org/pdf/2204.14264
-"ChatGPT Prompt Patterns for Improving Code Quality, Refactoring, Requirements Elicitation, and Software Design",Jules White,"This paper presents prompt design techniques for software engineering, in the form of patterns, to solve common problems when using large language models (LLMs), such as ChatGPT to automate common software engineering activities, such as ensuring code is decoupled from third-party libraries and simulating a web application API before it is implemented. This paper provides two contributions to research on using LLMs for software engineering. First, it provides a catalog of patterns for software engineering that classifies patterns according to the types of problems they solve. Second, it explores several prompt patterns that have been applied to improve requirements elicitation, rapid prototyping, code quality, refactoring, and system design.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2303.07839
-Retrieval-Based Prompt Selection for Code-Related Few-Shot Learning,Noor Nashid,"Large language models trained on massive code corpora can generalize to new tasks without the need for task-specific fine-tuning. In few-shot learning, these models take as input a prompt, composed of natural language instructions, a few instances of task demonstration, and a query and generate an output. However, the creation of an effective prompt for code-related tasks in few-shot learning has received little attention. We present a technique for prompt creation that automatically retrieves code demonstrations similar to the developer task, based on embedding or frequency analysis. We apply our approach, Cedar, to two different programming languages, statically and dynamically typed, and two different tasks, namely, test assertion generation and program repair. For each task, we compare Cedar with state-of-the-art task-specific and fine-tuned models. The empirical results show that, with only a few relevant code demonstrations, our prompt creation technique is effective in both tasks with an accuracy of 76% and 52% for exact matches in test assertion generation and program repair tasks, respectively. For assertion generation, Cedar outperforms existing task-specific and fine-tuned models by 333% and 11%, respectively. For program repair, Cedar yields 189% better accuracy than task-specific models and is competitive with recent fine-tuned models. These findings have practical implications for practitioners, as Cedar could potentially be applied to multilingual and multitask settings without task or language-specific training with minimal examples and effort.","{'model': 'tldr@v2.0.0', 'text': 'A technique for prompt creation that automatically retrieves code demonstrations similar to the developer task, based on embedding or frequency analysis is presented, which could potentially be applied to multilingual and multitask settings without task or language-specific training with minimal examples and effort.'}",
-Log Parsing with Prompt-based Few-shot Learning,Van-Hoang Le,"Logs generated by large-scale software systems provide crucial information for engineers to understand the system status and diagnose problems of the systems. Log parsing, which converts raw log messages into structured data, is the first step to enabling automated log analytics. Existing log parsers extract the common part as log templates using statistical features. However, these log parsers often fail to identify the correct templates and parameters because: 1) they often overlook the semantic meaning of log messages, and 2) they require domain-specific knowledge for different log datasets. To address the limitations of existing methods, in this paper, we propose LogPPT to capture the patterns of templates using prompt-based few-shot learning. LogPPT utilises a novel prompt tuning method to recognise keywords and parameters based on a few labelled log data. In addition, an adaptive random sampling algorithm is designed to select a small yet diverse training set. We have conducted extensive experiments on 16 public log datasets. The experimental results show that LogPPT is effective and efficient for log parsing.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes LogPPT to capture the patterns of templates using prompt-based few-shot learning, which utilises a novel prompt tuning method to recognise keywords and parameters based on a few labelled log data.'}",https://arxiv.org/pdf/2302.07435
-"A study on Prompt Design, Advantages and Limitations of ChatGPT for Deep Learning Program Repair",Jialun Cao,"ChatGPT has revolutionized many research and industrial fields. ChatGPT has shown great potential in software engineering to boost various traditional tasks such as program repair, code understanding, and code generation. However, whether automatic program repair (APR) applies to deep learning (DL) programs is still unknown. DL programs, whose decision logic is not explicitly encoded in the source code, have posed unique challenges to APR. While to repair DL programs, an APR approach needs to not only parse the source code syntactically but also needs to understand the code intention. With the best prior work, the performance of fault localization is still far less than satisfactory (only about 30\%). Therefore, in this paper, we explore ChatGPT's capability for DL program repair by asking three research questions. (1) Can ChatGPT debug DL programs effectively? (2) How can ChatGPT's repair performance be improved by prompting? (3) In which way can dialogue help facilitate the repair? On top of that, we categorize the common aspects useful for prompt design for DL program repair. Also, we propose various prompt templates to facilitate the performance and summarize the advantages and disadvantages of ChatGPT's abilities such as detecting bad code smell, code refactoring, and detecting API misuse/deprecation.","{'model': 'tldr@v2.0.0', 'text': ""This paper explores ChatGPT's capability for DL program repair by asking three research questions and categorizes the common aspects useful for prompt design forDL program repair.""}",http://arxiv.org/pdf/2304.08191
-Patch-Token Aligned Bayesian Prompt Learning for Vision-Language Models,Xinyang Liu,"For downstream applications of vision-language pre-trained models, there has been significant interest in constructing effective prompts. Existing works on prompt engineering, which either require laborious manual designs or optimize the prompt tuning as a point estimation problem, may fail to describe diverse characteristics of categories and limit their applications. We introduce a Bayesian probabilistic resolution to prompt learning, where the label-specific stochastic prompts are generated hierarchically by first sampling a latent vector from an underlying distribution and then employing a lightweight generative model. Importantly, we semantically regularize prompt learning with the visual knowledge and view images and the corresponding prompts as patch and token sets under optimal transport, which pushes the prompt tokens to faithfully capture the label-specific visual concepts, instead of overfitting the training categories. Moreover, the proposed model can also be straightforwardly extended to the conditional case where the instance-conditional prompts are generated to improve the generalizability. Extensive experiments on 15 datasets show promising transferability and generalization performance of our proposed model.","{'model': 'tldr@v2.0.0', 'text': 'A Bayesian probabilistic resolution to prompt learning, where the label-specific stochastic prompts are generated hierarchically by first sampling a latent vector from an underlying distribution and then employing a lightweight generative model.'}",http://arxiv.org/pdf/2303.09100
-IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models,Hu Ye,"Recent years have witnessed the strong power of large text-to-image diffusion models for the impressive generative capability to create high-fidelity images. However, it is very tricky to generate desired images using only text prompt as it often involves complex prompt engineering. An alternative to text prompt is image prompt, as the saying goes:""an image is worth a thousand words"". Although existing methods of direct fine-tuning from pretrained models are effective, they require large computing resources and are not compatible with other base models, text prompt, and structural controls. In this paper, we present IP-Adapter, an effective and lightweight adapter to achieve image prompt capability for the pretrained text-to-image diffusion models. The key design of our IP-Adapter is decoupled cross-attention mechanism that separates cross-attention layers for text features and image features. Despite the simplicity of our method, an IP-Adapter with only 22M parameters can achieve comparable or even better performance to a fully fine-tuned image prompt model. As we freeze the pretrained diffusion model, the proposed IP-Adapter can be generalized not only to other custom models fine-tuned from the same base model, but also to controllable generation using existing controllable tools. With the benefit of the decoupled cross-attention strategy, the image prompt can also work well with the text prompt to achieve multimodal image generation. The project page is available at \url{https://ip-adapter.github.io}.","{'model': 'tldr@v2.0.0', 'text': 'The proposed IP-Adapter is an effective and lightweight adapter to achieve image prompt capability for the pretrained text-to-image diffusion models and has the benefit of the decoupled cross-attention strategy, the image prompt can also work well with the text prompt to achieve multimodal image generation.'}",https://arxiv.org/pdf/2308.06721
-Prompt Space Optimizing Few-shot Reasoning Success with Large Language Models,Fobo Shi,"Prompt engineering is an essential technique for enhancing the abilities of large language models (LLMs) by providing explicit and specific instructions. It enables LLMs to excel in various tasks, such as arithmetic reasoning, question answering, summarization, relation extraction, machine translation, and sentiment analysis. Researchers have been actively exploring different prompt engineering strategies, such as Chain of Thought (CoT), Zero-CoT, and In-context learning. However, an unresolved problem arises from the fact that current approaches lack a solid theoretical foundation for determining optimal prompts. To address this issue in prompt engineering, we propose a new and effective approach called Prompt Space. Our methodology utilizes text embeddings to obtain basis vectors by matrix decomposition, and then constructs a space for representing all prompts. Prompt Space significantly outperforms state-of-the-art prompt paradigms on ten public reasoning benchmarks. Notably, without the help of the CoT method and the prompt""Let's think step by step"", Prompt Space shows superior performance over the few-shot method. Overall, our approach provides a robust and fundamental theoretical framework for selecting simple and effective prompts. This advancement marks a significant step towards improving prompt engineering for a wide variety of applications in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Prompt Space methodology utilizes text embeddings to obtain basis vectors by matrix decomposition, and then constructs a space for representing all prompts, which significantly outperforms state-of-the-art prompt paradigms on ten public reasoning benchmarks.'}",http://arxiv.org/pdf/2306.03799
-"Prompt-Free Diffusion: Taking ""Text"" out of Text-to-Image Diffusion Models",Xingqian Xu,"Text-to-image (T2I) research has grown explosively in the past year, owing to the large-scale pre-trained diffusion models and many emerging personalization and editing approaches. Yet, one pain point persists: the text prompt engineering, and searching high-quality text prompts for customized results is more art than science. Moreover, as commonly argued:""an image is worth a thousand words""- the attempt to describe a desired image with texts often ends up being ambiguous and cannot comprehensively cover delicate visual details, hence necessitating more additional controls from the visual domain. In this paper, we take a bold step forward: taking""Text""out of a pre-trained T2I diffusion model, to reduce the burdensome prompt engineering efforts for users. Our proposed framework, Prompt-Free Diffusion, relies on only visual inputs to generate new images: it takes a reference image as""context"", an optional image structural conditioning, and an initial noise, with absolutely no text prompt. The core architecture behind the scene is Semantic Context Encoder (SeeCoder), substituting the commonly used CLIP-based or LLM-based text encoder. The reusability of SeeCoder also makes it a convenient drop-in component: one can also pre-train a SeeCoder in one T2I model and reuse it for another. Through extensive experiments, Prompt-Free Diffusion is experimentally found to (i) outperform prior exemplar-based image synthesis approaches; (ii) perform on par with state-of-the-art T2I models using prompts following the best practice; and (iii) be naturally extensible to other downstream applications such as anime figure generation and virtual try-on, with promising quality. Our code and models are open-sourced at https://github.com/SHI-Labs/Prompt-Free-Diffusion.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.16223
-What initiates evidence‐based reasoning?: Situations that prompt students to support their design ideas and decisions,Emilie A. Siverling,"As engineering becomes increasingly incorporated into precollege classrooms, it is important to explore students' ability to engage in engineering practices. One of these practices, engaging in argument from evidence, has been well studied in science education. However, it has not yet been fully explored in engineering education.",,
-A Brief History of Prompt: Leveraging Language Models,G. Muktadir,"This paper presents a comprehensive exploration of the evolution of prompt engineering and generation in the field of natural language processing (NLP). Starting from the early language models and information retrieval systems, we trace the key developments that have shaped prompt engineering over the years. The introduction of attention mechanisms in 2015 revolutionized language understanding, leading to advancements in controllability and context-awareness. Subsequent breakthroughs in reinforcement learning techniques further enhanced prompt engineering, addressing issues like exposure bias and biases in generated text. We examine the significant contributions in 2018 and 2019, focusing on fine-tuning strategies, control codes, and template-based generation. The paper also discusses the growing importance of fairness, human-AI collaboration, and low-resource adaptation. In 2020 and 2021, contextual prompting and transfer learning gained prominence, while 2022 and 2023 witnessed the emergence of advanced techniques like unsupervised pre-training and novel reward shaping. Throughout the paper, we reference specific research studies that exemplify the impact of various developments on prompt engineering. The journey of prompt engineering continues, with ethical considerations being paramount for the responsible and inclusive future of AI systems.","{'model': 'tldr@v2.0.0', 'text': 'The journey of prompt engineering continues, with ethical considerations being paramount for the responsible and inclusive future of AI systems.'}",https://arxiv.org/pdf/2310.04438
-BIM-GPT: a Prompt-Based Virtual Assistant Framework for BIM Information Retrieval,Junwen Zheng,"Efficient information retrieval (IR) from building information models (BIMs) poses significant challenges due to the necessity for deep BIM knowledge or extensive engineering efforts for automation. We introduce BIM-GPT, a prompt-based virtual assistant (VA) framework integrating BIM and generative pre-trained transformer (GPT) technologies to support NL-based IR. A prompt manager and dynamic template generate prompts for GPT models, enabling interpretation of NL queries, summarization of retrieved information, and answering BIM-related questions. In tests on a BIM IR dataset, our approach achieved 83.5% and 99.5% accuracy rates for classifying NL queries with no data and 2% data incorporated in prompts, respectively. Additionally, we validated the functionality of BIM-GPT through a VA prototype for a hospital building. This research contributes to the development of effective and versatile VAs for BIM IR in the construction industry, significantly enhancing BIM accessibility and reducing engineering efforts and training data requirements for processing NL queries.","{'model': 'tldr@v2.0.0', 'text': 'BIM-GPT, a prompt-based virtual assistant framework integrating BIM and generative pre-trained transformer technologies to support NL-based IR, is introduced, significantly enhancing BIM accessibility and reducing engineering efforts and training data requirements for processing NL queries.'}",http://arxiv.org/pdf/2304.09333
-API Entity and Relation Joint Extraction from Text via Dynamic Prompt-tuned Language Model,Qing Huang,"Extraction of Application Programming Interfaces (APIs) and their semantic relations from unstructured text (e.g., Stack Overflow) is a fundamental work for software engineering tasks (e.g., API recommendation). However, existing approaches are rule-based and sequence-labeling based. They must manually enumerate the rules or label data for a wide range of sentence patterns, which involves a significant amount of labor overhead and is exacerbated by morphological and common-word ambiguity. In contrast to matching or labeling API entities and relations, this paper formulates heterogeneous API extraction and API relation extraction task as a sequence-to-sequence generation task, and proposes AERJE, an API entity-relation joint extraction model based on the large pre-trained language model. After training on a small number of ambiguous but correctly labeled data, AERJE builds a multi-task architecture that extracts API entities and relations from unstructured text using dynamic prompts. We systematically evaluate AERJE on a set of long and ambiguous sentences from Stack Overflow. The experimental results show that AERJE achieves high accuracy and discrimination ability in API entity-relation joint extraction, even with zero or few-shot fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'This paper formulates heterogeneous API extraction and API relation extraction task as a sequence-to-sequence generation task, and proposes AERJE, an API entity-relation joint extraction model based on the large pre-trained language model.'}",https://dl.acm.org/doi/pdf/10.1145/3607188
-Engineering a Second-Order DNA Logic-Gated Nanorobot to Sense-then-Release on Live Cell Membranes for Multiplexed Diagnosis and Synergistic Therapy.,Dan Wang,"Tumor biomarker-based theranostics has achieved broad interest and success in recent years. However, single biomarker-based recognition can cause false-positive feedback, including the on-target off-tumor phenomenon by the absence of tumor-specific antigen. Besides, multi-biomarker-based recognition molecules often elicit nonspecific and undesired internalization when they bind to ''bystander'' cells. Herein, we report a universal DNA tetrahedral scaffold (DTS) that anchors on the cell membrane to load multiple aptamers and therapeutics for precise and effective theranostics. This DNA logic-gated nanorobot (DLGN) not only facilitates the precise discrimination among five cell lines, but also triggers the synergistic killing of effector aptamer-tethered synergistic drugs (EASDs) to target cancer cells. Logic-gated recognition integrated in aptamer-functionalized molecular machines will prompt fast tumor profiling, in situ capture and isolation, and safe delivery of precise medicine.","{'model': 'tldr@v2.0.0', 'text': 'A universal DNA tetrahedral scaffold that anchors on the cell membrane to load multiple aptamers and therapeutics for precise and effective theranostics and triggers the synergistic killing of effector aptamer-tethered synergistic drugs (EASDs) to target cancer cells.'}",
-Undergraduate University Education in Internet of Things Engineering in China: A Survey,B. Du,"To address the problem of a serious Internet of Things (IoT) talent shortage, the Chinese government is encouraging universities and institutions to establish a major in IoT Engineering to train high-level IoT talent. Since 2011, 700 Chinese universities and institutions have established IoT Engineering undergraduate programs. We surveyed the history, status, curriculum, and problems of IoT undergraduate degree programs in China. Based on data collected from faculties, students, and graduates, we found that the IoT Engineering curriculum was an unsystematic patchwork and was deficient in practical platforms. We then provide suggestions for further development and exploration of IoT education by presenting a Technical Knowledge Map of IoT Engineering. This survey work on Chinese IoT education can prompt Chinese educators to reflect on its achievements and problems, which will contribute to IoT talent cultivation worldwide.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the IoT Engineering curriculum in China was an unsystematic patchwork and was deficient in practical platforms, which will contribute to IoT talent cultivation worldwide.'}",https://www.mdpi.com/2227-7102/11/5/202/pdf?version=1619423706
-Prompt gamma neutron activation analysis (PGAA): recent developments and applications,R. Lindstrom,,,
-A Prompt-Aware Neural Network Approach to Content-Based Scoring of Non-Native Spontaneous Speech,Yao Qian,"We present a neural network approach to the automated assessment of non-native spontaneous speech in a listen and speak task. An attention-based Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) is used to learn the relations (scoring rubrics) between the spoken responses and their assigned scores. Each prompt (listening material) is encoded as a vector in a low-dimensional space and then employed as a condition of the inputs of the attention LSTM-RNN. The experimental results show that our approach performs as well as the strong baseline of a Support Vector Regressor (SVR) using content-related features, i.e., a correlation of r = 0.806 with holistic proficiency scores provided by humans, without doing any feature engineering. The prompt-encoded vector improves the discrimination between the high-scoring sample and low-scoring sample, and it is more effective in grading responses to unseen prompts, which have no corresponding responses in the training set.","{'model': 'tldr@v2.0.0', 'text': 'A neural network approach to the automated assessment of non-native spontaneous speech in a listen and speak task that performs as well as the strong baseline of a Support Vector Regressor using content-related features, without doing any feature engineering.'}",
-Performance of ChatGPT on the US Fundamentals of Engineering Exam: Comprehensive Assessment of Proficiency and Potential Implications for Professional Environmental Engineering Practice,Vinay Pursnani,"In recent years, advancements in artificial intelligence (AI) have led to the development of large language models like GPT-4, demonstrating potential applications in various fields, including education. This study investigates the feasibility and effectiveness of using ChatGPT, a GPT-4 based model, in achieving satisfactory performance on the Fundamentals of Engineering (FE) Environmental Exam. This study further shows a significant improvement in the model's accuracy when answering FE exam questions through noninvasive prompt modifications, substantiating the utility of prompt modification as a viable approach to enhance AI performance in educational contexts. Furthermore, the findings reflect remarkable improvements in mathematical capabilities across successive iterations of ChatGPT models, showcasing their potential in solving complex engineering problems. Our paper also explores future research directions, emphasizing the importance of addressing AI challenges in education, enhancing accessibility and inclusion for diverse student populations, and developing AI-resistant exam questions to maintain examination integrity. By evaluating the performance of ChatGPT in the context of the FE Environmental Exam, this study contributes valuable insights into the potential applications and limitations of large language models in educational settings. As AI continues to evolve, these findings offer a foundation for further research into the responsible and effective integration of AI models across various disciplines, ultimately optimizing the learning experience and improving student outcomes.",,http://arxiv.org/pdf/2304.12198
-Integrated Structure and Device Engineering for High Performance and Scalable Quantum Dot Infrared Photodetectors.,Kaimin Xu,"Colloidal quantum dots (CQDs) are emerging as promising materials for the next generation infrared (IR) photodetectors, due to their easy solution processing, low cost manufacturing, size-tunable optoelectronic properties, and flexibility. Tremendous efforts including material engineering and device structure manipulation have been made to improve the performance of the photodetectors based on CQDs. In recent years, benefiting from the facial integration with materials such as 2D structure, perovskite and silicon, as well as device engineering, the performance of CQD IR photodetectors have been developing rapidly. On the other hand, to prompt the application of CQD IR photodetectors, scalable device structures that are compatible with commercial systems are developed. Herein, recent advances of CQD based IR photodetectors are summarized, especially material integration, device engineering, and scalable device structures.","{'model': 'tldr@v2.0.0', 'text': 'Recent advances of CQD based IR photodetectors are summarized, especially material integration, device engineering, and scalable device structures.'}",
-Are fast radio bursts the most likely electromagnetic counterpart of neutron star mergers resulting in prompt collapse?,V. Paschalidis,"National Science Foundation (NSF) at the University of Arizona [PHY-1912619]; NSF [PHY-1602536, PHY-1662211]; NASA at the University of Illinois at Urbana-Champaign [80NSSC17K0070]; Extreme Science and Engineering Discovery Environment (XSEDE) [TG-PHY180036, TG-PHY190020]",,https://link.aps.org/accepted/10.1103/PhysRevD.100.043001
-Tunable Phosphorescence-Fluorescence Dual Emissions of Organic Isoquinoline-Benzophenone Doped System by Alkoxy Engineering.,Yating Chen,"The dual/multi-component organic doped system with room temperature phosphorescence (RTP) properties has been developed. However, the unknown luminescence mechanism still greatly limits the development of the doped materials. Herein, a new doped system exhibiting phosphorescence-fluorescence dual emissions (ФPhos.= 4-24% and τphos. = 101-343 ms) is successfully constructed through prediction and design. A series of isoquinoline derivatives with different alkoxy chains were selected as the guests. Benzophenone was chosen as the host due to the characteristic of low melting point and good crystallinity. The alkoxy chain lengths of the guests are first reported to be used to control the fluorescence and phosphorescence intensities of the doped materials, which results in different prompt emission colors. Additionally, the doped ratio of the guest and host can also control the luminous intensities of the materials. In particular, the doped materials still exhibit phosphorescent properties even if the ratio of guest-host is as low as 1:100000.","{'model': 'tldr@v2.0.0', 'text': 'A new doped system exhibiting phosphorescence-fluorescence dual emissions is successfully constructed through prediction and design and can still exhibit phosphorescent properties even if the ratio of guest-host is as low as 1:100000.'}",
-A Surface‐Tailoring Method for Rapid Non‐Thermosensitive Cell‐Sheet Engineering via Functional Polymer Coatings,Jieung Baek,"Cell sheet engineering, a technique utilizing a monolayer cell sheet, has recently emerged as a promising technology for scaffold‐free tissue engineering. In contrast to conventional tissue‐engineering approaches, the cell sheet technology allows cell harvest as a continuous cell sheet with intact extracellular matrix proteins and cell–cell junction, which facilitates cell transplantation without any other artificial biomaterials. A facile, non‐thermoresponsive method is demonstrated for a rapid but highly reliable platform for cell‐sheet engineering. The developed method exploits the precise modulation of cell–substrate interactions by controlling the surface energy of the substrate via a series of functional polymer coatings to enable prompt cell sheet harvesting within 100 s. The engineered surface can trigger an intrinsic cellular response upon the depletion of divalent cations, leading to spontaneous cell sheet detachment under physiological conditions (pH 7.4 and 37 °C) in a non‐thermoresponsive manner. Additionally, the therapeutic potential of the cell sheet is successfully demonstrated by the transplantation of multilayered cell sheets into mouse models of diabetic wounds and ischemia. These findings highlight the ability of the developed surface for non‐thermoresponsive cell sheet engineering to serve as a robust platform for regenerative medicine and provide significant breakthroughs in cell sheet technology.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the developed surface for non‐thermoresponsive cell sheet engineering to serve as a robust platform for regenerative medicine and provide significant breakthroughs in cell sheet technology are highlighted.'}",
-Vascular Tissue Engineering: Advanced Techniques and Gene Editing in Stem Cells for Graft Generation.,Sin-Guang Chen,"The common occurrence of cardiovascular diseases and the lack of proper autologous tissues prompt and promote the pressing development of tissue-engineered vascular grafts. Current progress on scaffold production, genetically modified cells and use of nanotechnology-based monitoring has considerably improved the long-term patency of engineered tissue grafts. However, challenges abound in the autologous materials and manipulation of genes and cells for tissue engineering. This review overviews current development in tissue-engineered vascular grafts and discusses recent improvements in scaffolding techniques as well as the efficiency of gene-editing tools and their ability to fill the existing gaps in stem-cell and regenerative therapies. Current advances in 3D-printing approaches for fabrication of engineered tissues are also reviewed together with specific biomaterials for vascular tissues. In addition, the natural and synthetic polymers that hold increasing significance for vascular tissue engineering are highlighted. Both animal models and nanotechnology-based monitoring are proposed for pre-clinical evaluation of engineered grafts in view of their historical significance in tissue engineering. The ultimate success of tissue regeneration, which is yet to be fully realized, depends on the optimal performance of culture systems, biomaterial constructs and stem cells in a suitable artificial physiological environment.","{'model': 'tldr@v2.0.0', 'text': 'This review overviews current development in tissue-engineered vascular grafts and discusses recent improvements in scaffolding techniques as well as the efficiency of gene-editing tools and their ability to fill the existing gaps in stem-cell and regenerative therapies.'}",http://ira.lib.polyu.edu.hk/bitstream/10397/93663/1/Chen_Vascular_Tissue_Engineering.pdf
-CO₂-Induced Phase Engineering: Protocol for Enhanced Photoelectrocatalytic Performance of 2D MoS₂ Nanosheets.,Yuhang Qi,"Molybdenum disulfide (MoS2) is a promising non-precious-metal catalyst, but its performance is limited by the density of active sites and poor electrical transport. Its metallic 1T phase possesses higher photoelectrocatalytic activity. Thus, how to efficiently increase the concentration of the 1T phase in the exfoliated two-dimensiaonal (2D) MoS2 nanosheets is an important premise. In this work, we propose a strategy to prepare a 2D heterostructure of MoS2 nanosheets using supercritical CO2-induced phase engineering to form metallic 1T-MoS2. Theoretical calculations and experimental results demonstrate that the introduced CO2 in the 2H-MoS2 host can prompt the transformation of partial 2H-MoS2 lattices into 1T-MoS2. Moreover, the electrical coupling and synergistic effect between 2H and 1T phases can greatly facilitate the efficient electron transfer from the active sites of MoS2, which significantly improves the photocatalytic performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a strategy to prepare a 2D heterostructure of MoS2 nanosheets using supercritical CO2-induced phase engineering to form metallic 1T-MoS2, which significantly improves the photocatalytic performance.'}",
-Engineering chiral polyoxometalate hybrid metal-organic frameworks for asymmetric dihydroxylation of olefins.,Q. Han,"Chiral metal-organic frameworks (MOFs) with porous and tunable natures have made them feasible for performing a variety of chemical reactions as heterogeneous asymmetric catalysts. By incorporating the oxidation catalyst [BW12O40](5-) and the chiral group, L- or D-pyrrolidin-2-ylimidazole (PYI), into one single framework, the two enantiomorphs Ni-PYI1 and Ni-PYI2 were obtained via self-assembly, respectively. The channels of Ni-PYIs were enlarged through a guest exchange reaction to remove the cationic chiral templates and were well modulated with hydrophilic/hydrophobic properties to allow molecules of both H2O2 and olefin ingress and egress. The coexistence of both the chiral directors and the oxidants within a confined space provided a special environment for the formation of reaction intermediates in a stereoselective fashion with high selectivity. The resulting MOF acted as an amphipathic catalyst to prompt the asymmetric dihydroxylation of aryl olefins with excellent stereoselectivity.","{'model': 'tldr@v2.0.0', 'text': 'The coexistence of both the chiral directors and the oxidants within a confined space provided a special environment for the formation of reaction intermediates in a stereoselective fashion with high selectivity.'}",
-How Do Different Reflection Prompts Affect Engineering Students’ Academic Performance and Engagement?,Muhsin Menekse,"Abstract This study explored the relative effectiveness of generic versus specific reflection prompts on engineering students’ academic performance and engagement with four scales (i.e., behavioral, emotional, social, cognitive). The sample consisted of 208 engineering students in two sections of a required first-year engineering course. By using a quasi-experimental research design in classroom settings, students in the specific and generic reflection prompt conditions were compared based on exams, project scores, and problem sets. Additionally, the role of reflection prompts in students’ behavioral, emotional, social, and cognitive engagements was investigated across conditions by using pre- and post-surveys administered at the beginning and end of the semester. Students in both conditions were asked to submit a total of 52 reflections during an academic semester. Results showed that students in the specific prompts condition performed significantly better on exams, total project score, and two problem sets than students in the generic prompts condition. In terms of engagement, results showed there was a significant decrease in behavioral, emotional, and cognitive scales for both conditions. On the other hand, for the social scale, while there was a decrease for students in generic prompts condition, there was an increase for the ones in specific prompts condition.",,
-Engineering mindsets and learning outcomes in elementary school,Pamela S. Lottero‐Perdue,"Students may exhibit growth mindsets, where intelligence is seen as malleable and failures prompt more effort and new approaches, or fixed mindsets, where intelligence is seen as immutable and failures indicate lack of intelligence. One's mindset in general may be different from that for a particular domain such as engineering. Having a growth mindset predicts more positive learning outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Students may exhibit growth mindsets, where intelligence is seen as malleable and failures prompt more effort and new approaches, or fixed mindsets that are seen as immutable and failures indicate lack of intelligence.'}",
-"Rapid Detection of Genetic Engineering, Structural Variation, and Antimicrobial Resistance Markers in Bacterial Biothreat Pathogens by Nanopore Sequencing",Amy S. Gargis,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/s41598-019-49700-1.pdf
-Red Teaming Language Models with Language Models,Ethan Perez,"Language Models (LMs) often cannot be deployed because of their potential to harm users in hard-to-predict ways. Prior work identifies harmful behaviors before deployment by using human annotators to hand-write test cases. However, human annotation is expensive, limiting the number and diversity of test cases. In this work, we automatically find cases where a target LM behaves in a harmful way, by generating test cases (“red teaming”) using another LM. We evaluate the target LM’s replies to generated test questions using a classifier trained to detect offensive content, uncovering tens of thousands of offensive replies in a 280B parameter LM chatbot. We explore several methods, from zero-shot generation to reinforcement learning, for generating test cases with varying levels of diversity and difficulty. Furthermore, we use prompt engineering to control LM-generated test cases to uncover a variety of other harms, automatically finding groups of people that the chatbot discusses in offensive ways, personal and hospital phone numbers generated as the chatbot’s own contact info, leakage of private training data in generated text, and harms that occur over the course of a conversation. Overall, LM-based red teaming is one promising tool (among many needed) for finding and fixing diverse, undesirable LM behaviors before impacting users.","{'model': 'tldr@v2.0.0', 'text': 'This work automatically finds cases where a target LM behaves in a harmful way, by generating test cases (“red teaming”) using another LM, and evaluates the target LM’s replies to generated test questions using a classifier trained to detect offensive content.'}",https://aclanthology.org/2022.emnlp-main.225.pdf
-Engineering a Segmented Dual-Reservoir Polyurethane Intravaginal Ring for Simultaneous Prevention of HIV Transmission and Unwanted Pregnancy,Justin T. Clark,"The HIV/AIDS pandemic and its impact on women prompt the investigation of prevention strategies to interrupt sexual transmission of HIV. Long-acting drug delivery systems that simultaneously protect womenfrom sexual transmission of HIV and unwanted pregnancy could be important tools in combating the pandemic. We describe the design, in silico, in vitro and in vivo evaluation of a dual-reservoir intravaginal ring that delivers the HIV-1 reverse transcriptase inhibitor tenofovir and the contraceptive levonorgestrel for 90 days. Two polyether urethanes with two different hard segment volume fractions were used to make coaxial extruded reservoir segments with a 100 µm thick rate controlling membrane and a diameter of 5.5 mm that contain 1.3 wt% levonorgestrel. A new mechanistic diffusion model accurately described the levonorgestrel burst release in early time points and pseudo-steady state behavior at later time points. As previously described, tenofovir was formulated as a glycerol paste and filled into a hydrophilic polyurethane, hollow tube reservoir that was melt-sealed by induction welding. These tenofovir-eluting segments and 2 cm long coaxially extruded levonorgestrel eluting segments were joined by induction welding to form rings that released an average of 7.5 mg tenofovir and 21 µg levonorgestrel per day in vitro for 90 days. Levonorgestrel segments placed intravaginally in rabbits resulted in sustained, dose-dependent levels of levonorgestrel in plasma and cervical tissue for 90 days. Polyurethane caps placed between segments successfully prevented diffusion of levonorgestrel into the tenofovir-releasing segment during storage.Hydrated rings endured between 152 N and 354 N tensile load before failure during uniaxial extension testing. In summary, this system represents a significant advance in vaginal drug delivery technology, and is the first in a new class of long-acting multipurpose prevention drug delivery systems.","{'model': 'tldr@v2.0.0', 'text': 'This system is described, in silico, in vitro and in vivo evaluation of a dual-reservoir intravaginal ring that delivers the HIV-1 reverse transcriptase inhibitor tenofovir and the contraceptive levonorgestrel for 90 days and is the first in a new class of long-acting multipurpose prevention drug delivery systems.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0088509&type=printable
-Exploring the Effects of the Design Prompt on Students’ Design Cognition,Christopher B. Williams,"Engineering design educators often provide their students a task (or “prompt”) to guide their design projects. Similarly, engineering design educational researchers also provide research participants with a design task to guide their activity during experimental sessions. In both contexts, there is a fundamental underlying assumption that the design task has no significant effect on the students’/participants’ design cognition. Specifically, the authors test the hypothesis that a design task does affect a student’s design experience. Failing to disprove this hypothesis could significantly impact both design education practice and design education experimental research.To determine the effect of a design task on students’ design cognition, experimental sessions were conducted wherein student design teams worked together to solve a speculative design task. The student teams were presented with two nearly identical design tasks; however, one featured an additional design requirement. A task-independent protocol analysis method grounded in the Function-Behavior-Structure design ontology is performed on audio and video recordings of the design sessions to provide a common basis for comparing the two groups. Differences in design cognition are identified by analyzing and comparing the percent occurrences of the design issues and design processes and the Problem-Solution indices.Copyright © 2013 by ASME","{'model': 'tldr@v2.0.0', 'text': 'The authors test the hypothesis that a design task does affect a student’s design experience and try to disprove this hypothesis, which could significantly impact both design education practice and design education experimental research.'}",
-Organoid technologies meet genome engineering,J. Nie,"Three‐dimensional (3D) stem cell differentiation cultures recently emerged as a novel model system for investigating human embryonic development and disease progression in vitro, complementing existing animal and two‐dimensional (2D) cell culture models. Organoids, the 3D self‐organizing structures derived from pluripotent or somatic stem cells, can recapitulate many aspects of structural organization and functionality of their in vivo organ counterparts, thus holding great promise for biomedical research and translational applications. Importantly, faithful recapitulation of disease and development processes relies on the ability to modify the genomic contents in organoid cells. The revolutionary genome engineering technologies, CRISPR/Cas9 in particular, enable investigators to generate various reporter cell lines for prompt validation of specific cell lineages as well as to introduce disease‐associated mutations for disease modeling. In this review, we provide historical overviews, and discuss technical considerations, and potential future applications of genome engineering in 3D organoid models.","{'model': 'tldr@v2.0.0', 'text': 'Three‐dimensional stem cell differentiation cultures recently emerged as a novel model system for investigating human embryonic development and disease progression in vitro, complementing existing animal and two‐dimensional cell culture models.'}",https://europepmc.org/articles/pmc5331239?pdf=render
-ChatGPT for Robotics: Design Principles and Model Abilities,Sai Vemprala,"This paper presents an experimental study regarding the use of OpenAI's ChatGPT for robotics applications. We outline a strategy that combines design principles for prompt engineering and the creation of a high-level function library which allows ChatGPT to adapt to different robotics tasks, simulators, and form factors. We focus our evaluations on the effectiveness of different prompt engineering techniques and dialog strategies towards the execution of various types of robotics tasks. We explore ChatGPT's ability to use free-form dialog, parse XML tags, and to synthesize code, in addition to the use of task-specific prompting functions and closed-loop reasoning through dialogues. Our study encompasses a range of tasks within the robotics domain, from basic logical, geometrical, and mathematical reasoning all the way to complex domains such as aerial navigation, manipulation, and embodied agents. We show that ChatGPT can be effective at solving several of such tasks, while allowing users to interact with it primarily via natural language instructions. In addition to these studies, we introduce an open-sourced research tool called PromptCraft, which contains a platform where researchers can collaboratively upload and vote on examples of good prompting schemes for robotics applications, as well as a sample robotics simulator with ChatGPT integration, making it easier for users to get started with using ChatGPT for robotics.","{'model': 'tldr@v2.0.0', 'text': 'This study encompasses a range of tasks within the robotics domain, from basic logical, geometrical, and mathematical reasoning all the way to complex domains such as aerial navigation, manipulation, and embodied agents, and shows that ChatGPT can be effective at solving several of such tasks, while allowing users to interact with it primarily via natural language instructions.'}",https://arxiv.org/pdf/2306.17582
-Collagen hydrogel as an immunomodulatory scaffold in cartilage tissue engineering.,T. Yuan,"A collagen type I hydrogel was constructed and used as the scaffold for cartilage tissue engineering. Neonatal rabbit chondrocytes were seeded into the hydrogel, and the constructs were cultured in vitro for 7, 14, and 28 days. The immunomodulatory effect of the hydrogel on seeded chondrocytes was carefully investigated. The expressions of major histocompatibility complex classes I and II of seeded chondrocytes increased with the time, which indicated that the immunogenicity also increased with the time. Meanwhile, the properly designed collagen type I hydrogel could prompt the chondrogenesis of engineered cartilage. The extracellular matrix (ECM) synthesis ability of seeded chondrocytes and the accumulated ECM in the constructs continuously increased with the culture time. Both the isolation and protection, which come from formed ECM and hydrogel scaffold, can effectively control the adverse immunogenicity of seeded chondrocytes and even help to lessen the immunogenicity of the whole engineered cartilage. As the result, the levels of mixed lymphocyte chondrocyte reactions of seed cells and the constructs decreased gradually. The stimulation on allogeneic lymphocytes of the whole constructs was obviously lower than that of the retrieved cells from the constructs. Therefore, properly designed collagen type I hydrogel can give certain immunogenicity-reducing effects on engineered cartilage based on chondrocytes, and it may be a potential immunomodulatory biomaterial in tissue engineering.","{'model': 'tldr@v2.0.0', 'text': None}",
-Event-Shape Engineering for the D-meson elliptic flow in mid-central Pb-Pb collisions at sNN = 5 . 02 TeV,S. Acharya,"The production yield of prompt D mesons and their elliptic flow coefficient v2 were measured with the Event-Shape Engineering (ESE) technique applied to mid-central (10–30% and 30–50% centrality classes) Pb-Pb collisions at the centre-of-mass energy per nucleon pair √ sNN = 5.02 TeV, with the ALICE detector at the LHC. The ESE technique allows the classification of events, belonging to the same centrality, according to the azimuthal anisotropy of soft particle production in the collision. The reported measurements give the opportunity to investigate the dynamics of charm quarks in the Quark-Gluon Plasma and provide information on their participation in the collective expansion of the medium. D mesons were reconstructed via their hadronic decays at mid-rapidity, |η| < 0.8, in the transverse momentum interval 1 < pT < 24 GeV/c. The v2 coefficient is found to be sensitive to the event-shape selection confirming a correlation between the D-meson azimuthal anisotropy and the collective expansion of the bulk matter, while the per-event D-meson yields do not show any significant modification within the current uncertainties.",,
-Rationale-Augmented Ensembles in Language Models,Xuezhi Wang,"Recent research has shown that rationales, or step-by-step chains of thought, can be used to improve performance in multi-step reasoning tasks. We reconsider rationale-augmented prompting for few-shot in-context learning, where (input ->output) prompts are expanded to (input, rationale ->output) prompts. For rationale-augmented prompting we demonstrate how existing approaches, which rely on manual prompt engineering, are subject to sub-optimal rationales that may harm performance. To mitigate this brittleness, we propose a unified framework of rationale-augmented ensembles, where we identify rationale sampling in the output space as the key component to robustly improve performance. This framework is general and can easily be extended to common natural language processing tasks, even those that do not traditionally leverage intermediate steps, such as question answering, word sense disambiguation, and sentiment analysis. We demonstrate that rationale-augmented ensembles achieve more accurate and interpretable results than existing prompting approaches--including standard prompting without rationales and rationale-based chain-of-thought prompting--while simultaneously improving interpretability of model predictions through the associated rationales.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that rationale-augmented ensembles achieve more accurate and interpretable results than existing prompting approaches--including standard prompting without rationales and rationale-based chain-of-thought prompting--while simultaneously improving interpretability of model predictions through the associated rationales.'}",http://arxiv.org/pdf/2207.00747
-Application and optimization of CRISPR–Cas9-mediated genome engineering in axolotl (Ambystoma mexicanum),J. Fei,,"{'model': 'tldr@v2.0.0', 'text': 'An optimized stepwise protocol to create genetically modified axolotls using the CRISPR–Cas9 system is described, which bypasses the long generation time of axolOTls and allows direct functional analysis in F0 genetically manipulated axolotsls.'}",
-The Creativity of Text-to-Image Generation,J. Oppenlaender,"Text-guided synthesis of images has made a giant leap towards becoming a mainstream phenomenon. With text-to-image generation systems, anybody can create digital images and artworks. This provokes the question of whether text-to-image generation is creative. This paper expounds on the nature of human creativity involved in text-to-image art (so-called “AI art”) with a specific focus on the practice of prompt engineering. The paper argues that the current product-centered view of creativity falls short in the context of text-to-image generation. A case exemplifying this shortcoming is provided and the importance of online communities for the creative ecosystem of text-to-image art is highlighted. The paper provides a high-level summary of this online ecosystem drawing on Rhodes’ conceptual four P model of creativity. Challenges for evaluating the creativity of text-to-image generation and opportunities for research on text-to-image generation in the field of Human-Computer Interaction (HCI) are discussed.",,https://dl.acm.org/doi/pdf/10.1145/3569219.3569352
-Symbolic Knowledge Distillation: from General Language Models to Commonsense Models,Peter West,"The common practice for training commonsense models has gone from–human–to–corpus–to–machine: humans author commonsense knowledge graphs in order to train commonsense models. In this work, we investigate an alternative, from–machine–to–corpus–to–machine: general language models author these commonsense knowledge graphs to train commonsense models. Our study leads to a new framework, Symbolic Knowledge Distillation. As with prior art in Knowledge Distillation (Hinton et al. 2015), our approach uses larger models to teach smaller models. A key difference is that we distill knowledge symbolically–as text–in addition to the neural model. We distill only one aspect–the commonsense of a general language model teacher, allowing the student to be a different type, a commonsense model. Altogether, we show that careful prompt engineering and a separately trained critic model allow us to selectively distill high-quality causal commonsense from GPT-3, a general language model. Empirical results demonstrate that, for the first time, a human-authored commonsense knowledge graph is surpassed by our automatically distilled variant in all three criteria: quantity, quality, and diversity. In addition, it results in a neural commonsense model that surpasses the teacher model’s commonsense capabilities despite its 100x smaller size. We apply this to the ATOMIC resource, and will share our new symbolic knowledge graph and commonsense models.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that careful prompt engineering and a separately trained critic model allow us to selectively distill high-quality causal commonsense from GPT-3, a general language model, and results in a neural commonsense model that surpasses the teacher model's commonsense capabilities despite its 100x smaller size.""}",https://aclanthology.org/2022.naacl-main.341.pdf
-Optimizing Prompts for Text-to-Image Generation,Y. Hao,"Well-designed prompts can guide text-to-image models to generate amazing images. However, the performant prompts are often model-specific and misaligned with user input. Instead of laborious human engineering, we propose prompt adaptation, a general framework that automatically adapts original user input to model-preferred prompts. Specifically, we first perform supervised fine-tuning with a pretrained language model on a small collection of manually engineered prompts. Then we use reinforcement learning to explore better prompts. We define a reward function that encourages the policy to generate more aesthetically pleasing images while preserving the original user intentions. Experimental results on Stable Diffusion show that our method outperforms manual prompt engineering in terms of both automatic metrics and human preference ratings. Moreover, reinforcement learning further boosts performance, especially on out-of-domain prompts. The pretrained checkpoints are available at https://aka.ms/promptist. The demo can be found at https://aka.ms/promptist-demo.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt adaptation, a general framework that automatically adapts original user input to model-preferred prompts, and defines a reward function that encourages the policy to generate more aesthetically pleasing images while preserving the original user intentions.'}",http://arxiv.org/pdf/2212.09611
-"Chat2VIS: Generating Data Visualizations via Natural Language Using ChatGPT, Codex and GPT-3 Large Language Models",Paula Maddigan,"The field of data visualisation has long aimed to devise solutions for generating visualisations directly from natural language text. Research in Natural Language Interfaces (NLIs) has contributed towards the development of such techniques. However, the implementation of workable NLIs has always been challenging due to the inherent ambiguity of natural language, as well as in consequence of unclear and poorly written user queries which pose problems for existing language models in discerning user intent. Instead of pursuing the usual path of developing new iterations of language models, this study uniquely proposes leveraging the advancements in pre-trained large language models (LLMs) such as ChatGPT and GPT-3 to convert free-form natural language directly into code for appropriate visualisations. This paper presents a novel system, Chat2VIS, which takes advantage of the capabilities of LLMs and demonstrates how, with effective prompt engineering, the complex problem of language understanding can be solved more efficiently, resulting in simpler and more accurate end-to-end solutions than prior approaches. Chat2VIS shows that LLMs together with the proposed prompts offer a reliable approach to rendering visualisations from natural language queries, even when queries are highly misspecified and underspecified. This solution also presents a significant reduction in costs for the development of NLI systems, while attaining greater visualisation inference abilities compared to traditional NLP approaches that use hand-crafted grammar rules and tailored models. This study also presents how LLM prompts can be constructed in a way that preserves data security and privacy while being generalisable to different datasets. This work compares the performance of GPT-3, Codex and ChatGPT across several case studies and contrasts the performances with prior studies.","{'model': 'tldr@v2.0.0', 'text': 'Chat2VIS shows that LLMs together with the proposed prompts offer a reliable approach to rendering visualisations from natural language queries, even when queries are highly misspecified and underspecified.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10121440.pdf
-"ChatGPT Evaluation on Sentence Level Relations: A Focus on Temporal, Causal, and Discourse Relations",Chunkit Chan,"This paper aims to quantitatively evaluate the performance of ChatGPT, an interactive large language model, on inter-sentential relations such as temporal relations, causal relations, and discourse relations. Given ChatGPT's promising performance across various tasks, we conduct extensive evaluations on the whole test sets of 13 datasets, including temporal and causal relations, PDTB2.0-based and dialogue-based discourse relations, and downstream applications on discourse understanding. To achieve reliable results, we adopt three tailored prompt templates for each task, including the zero-shot prompt template, zero-shot prompt engineering (PE) template, and in-context learning (ICL) prompt template, to establish the initial baseline scores for all popular sentence-pair relation classification tasks for the first time. We find that ChatGPT exhibits strong performance in detecting and reasoning about causal relations, while it may not be proficient in identifying the temporal order between two events. It can recognize most discourse relations with existing explicit discourse connectives, but the implicit discourse relation still remains a challenging task. Meanwhile, ChatGPT performs poorly in the dialogue discourse parsing task that requires structural understanding in a dialogue before being aware of the discourse relation.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT exhibits strong performance in detecting and reasoning about causal relations, while it may not be proficient in identifying the temporal order between two events.'}",http://arxiv.org/pdf/2304.14827
-Integrating engineering outputs from natural disaster models into a dynamic spatial computable general equilibrium model of Centerville,H. Cutler,"Abstract A dynamic spatial computable general equilibrium (DSCGE) model is constructed that describes how engineering and economic models can be integrated to assess the economic, demographic, and fiscal impacts of disasters. This paper has two objectives. First, we introduce the DSCGE model and describe how it is calibrated specifically for Centerville. Second, to demonstrate the analytic flexibility of the DSCGE platform, we present economy-wide prompt effects from simulations involving spatially circumscribed shocks to Centerville’s building portfolio and transportation infrastructure, and then detail dynamic economy-wide effects from simulations involving combinations of infrastructure damage and adjustments to the economic behavior of agents. We conclude with a discussion of the technical challenges ahead.",,
-Cutting Down on Prompts and Parameters: Simple Few-Shot Learning with Language Models,Robert L Logan IV,"Prompting language models (LMs) with training examples and task descriptions has been seen as critical to recent successes in few-shot learning. In this work, we show that finetuning LMs in the few-shot setting can considerably reduce the need for prompt engineering. In fact, one can use null prompts, prompts that contain neither task-specific templates nor training examples, and achieve competitive accuracy to manually-tuned prompts across a wide range of tasks. While finetuning LMs does introduce new parameters for each downstream task, we show that this memory overhead can be substantially reduced: finetuning only the bias terms can achieve comparable or better accuracy than standard finetuning while only updating 0.1% of the parameters. All in all, we recommend finetuning LMs for few-shot learning as it is more accurate, robust to different prompts, and can be made nearly as efficient as using frozen LMs.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that finetuning LMs in the few-shot setting can considerably reduce the need for prompt engineering, and recommends finetuned LMs for few- shot learning as it is more accurate, robust to different prompts, and can be made nearly as efficient as using frozen LMs.'}",https://aclanthology.org/2022.findings-acl.222.pdf
-Overcoming physical constraints in bone engineering: ‘the importance of being vascularized’,T. Genova,"Bone plays several physiological functions and is the second most commonly transplanted tissue after blood. Since the treatment of large bone defects is still unsatisfactory, researchers have endeavoured to obtain scaffolds able to release growth and differentiation factors for mesenchymal stem cells, osteoblasts and endothelial cells in order to obtain faster mineralization and prompt a reliable vascularization. Nowadays, the application of osteoblastic cultures spans from cell physiology and pharmacology to cytocompatibility measurement and osteogenic potential evaluation of novel biomaterials. To overcome the simple traditional monocultures in vitro, co-cultures of osteogenic and vasculogenic precursors were introduced with very interesting results. Increasingly complex culture systems have been developed, where cells are seeded on proper scaffolds and stimulated so as to mimic the physiological conditions more accurately. These bioreactors aim at enabling bone regeneration by incorporating different cells types into bio-inspired materials within a surveilled habitat. This review is focused on the most recent developments in the organomimetic cultures of osteoblasts and vascular endothelial cells for bone tissue engineering.","{'model': 'tldr@v2.0.0', 'text': 'This review is focused on the most recent developments in the organomimetic cultures of osteoblasts and vascular endothelial cells for bone tissue engineering.'}",https://iris.unito.it/bitstream/2318/1541558/4/Genova_et_al_2016.pdf
-Fake it Till You Make it: Learning Transferable Representations from Synthetic ImageNet Clones,Mert Bulent Sariyildiz,"Recent image generation models such as Stable Diffusion have exhibited an impressive ability to generate fairly realistic images starting from a simple text prompt. Could such models render real images obsolete for training image prediction models? In this paper, we answer part of this provocative question by investigating the need for real images when training models for ImageNet classification. Provided only with the class names that have been used to build the dataset, we explore the ability of Stable Diffusion to generate synthetic clones of ImageNet and measure how useful these are for training classification models from scratch. We show that with minimal and class-agnostic prompt engineering, ImageNet clones are able to close a large part of the gap between models produced by synthetic images and models trained with real images, for the several standard classification benchmarks that we consider in this study. More importantly, we show that models trained on synthetic images exhibit strong generalization properties and perform on par with models trained on real data for transfer. Project page: https://europe.naverlabs.com/imagenet-sd","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with minimal and class-agnostic prompt engineering, ImageNet clones are able to close a large part of the gap between models produced by synthetic images and models trained with real images, for the several standard classification benchmarks that are considered in this study.'}",https://arxiv.org/pdf/2212.08420
-Text-Guided Synthesis of Artistic Images with Retrieval-Augmented Diffusion Models,Robin Rombach,"Novel architectures have recently improved generative image synthesis leading to excellent visual quality in various tasks. Of particular note is the field of ``AI-Art'', which has seen unprecedented growth with the emergence of powerful multimodal models such as CLIP. By combining speech and image synthesis models, so-called ``prompt-engineering'' has become established, in which carefully selected and composed sentences are used to achieve a certain visual style in the synthesized image. In this note, we present an alternative approach based on retrieval-augmented diffusion models (RDMs). In RDMs, a set of nearest neighbors is retrieved from an external database during training for each training instance, and the diffusion model is conditioned on these informative samples. During inference (sampling), we replace the retrieval database with a more specialized database that contains, for example, only images of a particular visual style. This provides a novel way to prompt a general trained model after training and thereby specify a particular visual style. As shown by our experiments, this approach is superior to specifying the visual style within the text prompt. We open-source code and model weights at https://github.com/CompVis/latent-diffusion .","{'model': 'tldr@v2.0.0', 'text': 'This note presents an alternative approach based on retrieval-augmented diffusion models (RDMs) that provides a novel way to prompt a general trained model after training and thereby specify a particular visual style.'}",http://arxiv.org/pdf/2207.13038
-BigBIO: A Framework for Data-Centric Biomedical Natural Language Processing,Jason Alan Fries,"Training and evaluating language models increasingly requires the construction of meta-datasets --diverse collections of curated data with clear provenance. Natural language prompting has recently lead to improved zero-shot generalization by transforming existing, supervised datasets into a diversity of novel pretraining tasks, highlighting the benefits of meta-dataset curation. While successful in general-domain text, translating these data-centric approaches to biomedical language modeling remains challenging, as labeled biomedical datasets are significantly underrepresented in popular data hubs. To address this challenge, we introduce BigBIO a community library of 126+ biomedical NLP datasets, currently covering 12 task categories and 10+ languages. BigBIO facilitates reproducible meta-dataset curation via programmatic access to datasets and their metadata, and is compatible with current platforms for prompt engineering and end-to-end few/zero shot language model evaluation. We discuss our process for task schema harmonization, data auditing, contribution guidelines, and outline two illustrative use cases: zero-shot evaluation of biomedical prompts and large-scale, multi-task learning. BigBIO is an ongoing community effort and is available at https://github.com/bigscience-workshop/biomedical","{'model': 'tldr@v2.0.0', 'text': 'BigBIO is introduced, a community library of 126+ biomedical NLP datasets, currently covering 12 task categories and 10+ languages that facilitates reproducible meta-dataset curation via programmatic access to datasets and their metadata, and is compatible with current platforms for prompt engineering and end-to-end few/zero shot language model evaluation.'}",http://arxiv.org/pdf/2206.15076
-CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets,Zachary Novack,"Open vocabulary models (e.g. CLIP) have shown strong performance on zero-shot classification through their ability generate embeddings for each class based on their (natural language) names. Prior work has focused on improving the accuracy of these models through prompt engineering or by incorporating a small amount of labeled downstream data (via finetuning). However, there has been little focus on improving the richness of the class names themselves, which can pose issues when class labels are coarsely-defined and are uninformative. We propose Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy for zero-shot classification specifically designed for datasets with implicit semantic hierarchies. CHiLS proceeds in three steps: (i) for each class, produce a set of subclasses, using either existing label hierarchies or by querying GPT-3; (ii) perform the standard zero-shot CLIP procedure as though these subclasses were the labels of interest; (iii) map the predicted subclass back to its parent to produce the final prediction. Across numerous datasets with underlying hierarchical structure, CHiLS leads to improved accuracy in situations both with and without ground-truth hierarchical information. CHiLS is simple to implement within existing zero-shot pipelines and requires no additional training cost. Code is available at: https://github.com/acmi-lab/CHILS.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy for zero-shot classification specifically designed for datasets with implicit semantic hierarchies, which leads to improved accuracy in situations both with and without ground-truth hierarchical information.'}",http://arxiv.org/pdf/2302.02551
-EvoPrompting: Language Models for Code-Level Neural Architecture Search,Angelica Chen,"Given the recent impressive accomplishments of language models (LMs) for code generation, we explore the use of LMs as adaptive mutation and crossover operators for an evolutionary neural architecture search (NAS) algorithm. While NAS still proves too difficult a task for LMs to succeed at solely through prompting, we find that the combination of evolutionary prompt engineering with soft prompt-tuning, a method we term EvoPrompting, consistently finds diverse and high performing models. We first demonstrate that EvoPrompting is effective on the computationally efficient MNIST-1D dataset, where EvoPrompting produces convolutional architecture variants that outperform both those designed by human experts and naive few-shot prompting in terms of accuracy and model size. We then apply our method to searching for graph neural networks on the CLRS Algorithmic Reasoning Benchmark, where EvoPrompting is able to design novel architectures that outperform current state-of-the-art models on 21 out of 30 algorithmic reasoning tasks while maintaining similar model size. EvoPrompting is successful at designing accurate and efficient neural network architectures across a variety of machine learning tasks, while also being general enough for easy adaptation to other tasks beyond neural network design.","{'model': 'tldr@v2.0.0', 'text': 'EvoPrompting is successful at designing accurate and efficient neural network architectures across a variety of machine learning tasks, while also being general enough for easy adaptation to other tasks beyond neural network design.'}",https://arxiv.org/pdf/2302.14838
-Repair Is Nearly Generation: Multilingual Program Repair with LLMs,Harshit Joshi,"Most programmers make mistakes when writing code. Some of these mistakes are small and require few edits to the original program – a class of errors recently termed last mile mistakes. These errors break the flow for experienced developers and can stump novice programmers. Existing automated repair techniques targeting this class of errors are language-specific and do not easily carry over to new languages. Transferring symbolic approaches requires substantial engineering and neural approaches require data and retraining. We introduce RING, a multilingual repair engine powered by a large language model trained on code (LLMC) such as Codex. Such a multilingual engine enables a flipped model for programming assistance, one where the programmer writes code and the AI assistance suggests fixes, compared to traditional code suggestion technology. Taking inspiration from the way programmers manually fix bugs, we show that a prompt-based strategy that conceptualizes repair as localization, transformation, and candidate ranking, can successfully repair programs in multiple languages with minimal effort. We present the first results for such a multilingual repair engine by evaluating on 6 different languages and comparing performance to language-specific repair engines. We show that RING can outperform language-specific repair engines for three of these languages.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces RING, a multilingual repair engine powered by a large language model trained on code (LLMC) such as Codex that enables a flipped model for programming assistance, one where the programmer writes code and the AI assistance suggests fixes compared to traditional code suggestion technology.'}",https://arxiv.org/pdf/2208.11640
-Injectable bottlebrush hydrogels with tissue-mimetic mechanical properties,F. Vashahi,"Injectable hydrogels are desired in many biomedical applications due to their minimally invasive deployment to the body and their ability to introduce drugs. However, current injectables suffer from mechanical mismatch with tissue, fragility, water expulsion, and high viscosity. To address these issues, we design brush-like macromolecules that concurrently provide softness, firmness, strength, fluidity, and swellability. The synthesized linear-bottlebrush-linear (LBL) copolymers facilitate improved injectability as the compact conformation of bottlebrush blocks results in low solution viscosity, while the thermoresponsive linear blocks permit prompt gelation at 37°C. The resulting hydrogels mimic the deformation response of supersoft tissues such as adipose and brain while withstanding deformations of 700% and precluding water expulsion upon gelation. Given their low cytotoxicity and mild inflammation in vivo, the developed materials will have vital implications for reconstructive surgery, tissue engineering, and drug delivery applications.","{'model': 'tldr@v2.0.0', 'text': 'The synthesized linear-bottlebrush-linear (LBL) copolymers facilitate improved injectability as the compact conformation of bottlebrush blocks results in low solution viscosity, while the thermoresponsive linear blocks permit prompt gelation at 37°C.'}",
-Prompting Is All Your Need: Automated Android Bug Replay with Large Language Models,Sidong Feng,"Bug reports are vital for software maintenance that allow users to inform developers of the problems encountered while using the software. As such, researchers have committed considerable resources toward automating bug replay to expedite the process of software maintenance. Nonetheless, the success of current automated approaches is largely dictated by the characteristics and quality of bug reports, as they are constrained by the limitations of manually-crafted patterns and pre-defined vocabulary lists. Inspired by the success of Large Language Models (LLMs) in natural language understanding, we propose AdbGPT , a new lightweight approach to automatically reproduce the bugs from bug reports through prompt engineering, without any training and hard-coding effort. AdbGPT leverages few-shot learning and chain-of-thought reasoning to elicit human knowledge and logical reasoning from LLMs to accomplish the bug replay in a manner similar to a developer. Our evaluations demonstrate the effectiveness and efficiency of our AdbGPT to reproduce 81.3% of bug reports in 253.6 seconds, outperforming the state-of-the-art baselines and ablation studies. We also conduct a small-scale user study to confirm the usefulness of AdbGPT in enhancing developers’ bug replay capabilities.","{'model': 'tldr@v2.0.0', 'text': 'AdbGPT is proposed, a new lightweight approach to automatically reproduce the bugs from bug reports through prompt engineering, without any training and hard-coding effort, that leverages few-shot learning and chain-of-thought reasoning to elicit human knowledge and logical reasoning from LLMs to accomplish the bug replay in a manner similar to a developer.'}",https://arxiv.org/pdf/2306.01987
-QaNER: Prompting Question Answering Models for Few-shot Named Entity Recognition,Andy T. Liu,"Recently, prompt-based learning for pre-trained language models has succeeded in few-shot Named Entity Recognition (NER) by exploiting prompts as task guidance to increase label efficiency. However, previous prompt-based methods for few-shot NER have limitations such as a higher computational complexity, poor zero-shot ability, requiring manual prompt engineering, or lack of prompt robustness. In this work, we address these shortcomings by proposing a new prompt-based learning NER method with Question Answering (QA), called QaNER. Our approach includes 1) a refined strategy for converting NER problems into the QA formulation; 2) NER prompt generation for QA models; 3) prompt-based tuning with QA models on a few annotated NER examples; 4) zero-shot NER by prompting the QA model. Comparing the proposed approach with previous methods, QaNER is faster at inference, insensitive to the prompt quality, and robust to hyper-parameters, as well as demonstrating significantly better low-resource performance and zero-shot capability.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new prompt-based learning NER method with Question Answering (QA), called QaNER, which is faster at inference, insensitive to the prompt quality, and robust to hyper-parameters, as well as demonstrating significantly better low-resource performance and zero-shot capability.'}",http://arxiv.org/pdf/2203.01543
-Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization,Puyuan Peng,"We investigate the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering. We selected three tasks: audio-visual speech recognition (AVSR), code-switched speech recognition (CS-ASR), and speech translation (ST) on unseen language pairs. We design task-specific prompts, by either leveraging another large-scale model, or simply manipulating the special tokens in the default prompts. Experiments show that compared to the default prompts, our proposed prompts improve performance by 10% to 45% on the three zero-shot tasks, and even outperform SotA supervised models on some datasets. In addition, our experiments reveal many interesting properties of Whisper, including its robustness to prompts, bias on accents, and the multilingual understanding in its latent space. Code is available at https://github.com/jasonppy/PromptingWhisper","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering, and designs task-specific prompts that improve performance on the three zero-shot tasks and even outperform SotA supervised models on some datasets.'}",https://arxiv.org/pdf/2305.11095
-Surface Engineering for Mechanical Enhancement of Cell Sheet by Nano-Coatings,Miso Yang,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/s41598-017-04746-x.pdf
-Enabling Continuous Software Engineering for Embedded Systems Architectures with Virtual Prototypes,P. Antonino,,"{'model': 'tldr@v2.0.0', 'text': 'This paper discusses in this paper how to orchestrate architecture drivers and design specification techniques with virtual prototypes, to address the demands of continuous software engineering in development processes.'}",
-Advancements in research and development to combat COVID-19 using nanotechnology,Vishal Chaudhary,,"{'model': 'tldr@v2.0.0', 'text': 'Various claims, studies and reports on research and development to combat these challenges associated with COVID-19 have been collectively discussed in this article from the perspectives of nanotechnology.'}",https://link.springer.com/content/pdf/10.1007/s41204-021-00102-7.pdf
-The Creativity of Text-based Generative Art,J. Oppenlaender,"Text-based generation of digital images has made a giant leap to-wards becoming a mainstream phenomenon. With text-based generative systems, anybody can create digital images and artworks. This provokes the question of whether text-based generative art is creative. This paper expounds on the nature of human creativity involved in text-based generative art with a specific focus on the practice of prompt engineering, drawing on Rhodes’s conceptual model of creativity. The paper critiques the current product-centered view of creativity which may fall short in the context of text-based generative art. An case exemplifying this shortcoming is provided and future opportunities for research on text-based generative art are outlined.",,http://arxiv.org/pdf/2206.02904
-Looking for a Handsome Carpenter! Debiasing GPT-3 Job Advertisements,Conrad Borchers,"The growing capability and availability of generative language models has enabled a wide range of new downstream tasks. Academic research has identified, quantified and mitigated biases present in language models but is rarely tailored to downstream tasks where wider impact on individuals and society can be felt. In this work, we leverage one popular generative language model, GPT-3, with the goal of writing unbiased and realistic job advertisements. We first assess the bias and realism of zero-shot generated advertisements and compare them to real-world advertisements. We then evaluate prompt-engineering and fine-tuning as debiasing methods. We find that prompt-engineering with diversity-encouraging prompts gives no significant improvement to bias, nor realism. Conversely, fine-tuning, especially on unbiased real advertisements, can improve realism and reduce bias.","{'model': 'tldr@v2.0.0', 'text': 'It is found that prompt-engineering with diversity-encouraging prompts gives no significant improvement to bias, nor realism, and fine-tuning, especially on unbiased real advertisements, can improve realism and reduce bias.'}",https://arxiv.org/pdf/2205.11374
-No Token Left Behind: Explainability-Aided Image Classification and Generation,Roni Paiss,"The application of zero-shot learning in computer vision has been revolutionized by the use of image-text matching models. The most notable example, CLIP, has been widely used for both zero-shot classification and guiding generative models with a text prompt. However, the zero-shot use of CLIP is unstable with respect to the phrasing of the input text, making it necessary to carefully engineer the prompts used. We find that this instability stems from a selective similarity score, which is based only on a subset of the semantically meaningful input tokens. To mitigate it, we present a novel explainability-based approach, which adds a loss term to ensure that CLIP focuses on all relevant semantic parts of the input, in addition to employing the CLIP similarity loss used in previous works. When applied to one-shot classification through prompt engineering, our method yields an improvement in the recognition rate, without additional training or fine-tuning. Additionally, we show that CLIP guidance of generative models using our method significantly improves the generated images. Finally, we demonstrate a novel use of CLIP guidance for text-based image generation with spatial conditioning on object location, by requiring the image explainability heatmap for each object to be confined to a pre-determined bounding box.","{'model': 'tldr@v2.0.0', 'text': 'A novel explainability-based approach is presented, which adds a loss term to ensure that CLIP focuses on all relevant semantic parts of the input, in addition to employing the CLIP similarity loss used in previous works.'}",http://arxiv.org/pdf/2204.04908
-Automatically Generating CS Learning Materials with Large Language Models,S. Macneil,"Recent breakthroughs in Large Language Models (LLMs), such as GPT-3 and Codex, now enable software developers to generate code based on a natural language prompt. Within computer science education, researchers are exploring the potential for LLMs to generate code explanations and programming assignments using carefully crafted prompts. These advances may enable students to interact with code in new ways while helping instructors scale their learning materials. However, LLMs also introduce new implications for academic integrity, curriculum design, and software engineering careers. This workshop will demonstrate the capabilities of LLMs to help attendees evaluate whether and how LLMs might be integrated into their pedagogy and research. We will also engage attendees in brainstorming to consider how LLMs will impact our field.","{'model': 'tldr@v2.0.0', 'text': 'This workshop will demonstrate the capabilities of LLMs to help attendees evaluate whether and how LLMs might be integrated into their pedagogy and research, and engage attendees in brainstorming to consider how LLM will impact the field.'}",https://arxiv.org/pdf/2212.05113
-Language-Aware Soft Prompting for Vision & Language Foundation Models,Adrian Bulat,"This paper is on soft prompt learning for Vision & Language (V&L) models. Similarly to their NLP counterparts, V&L models can be adapted to a downstream task by learning soft continuous prompts using a few training examples. Current methods learn the soft prompts by minimizing a cross-entropy loss using as class weights the features obtained by passing the prompts plus the class names through the text encoder. Such methods, however, significantly overfit the training data suffering from large accuracy degradation when tested on unseen classes from the same domain. Our main contribution, in this paper, is a surprisingly simple approach to alleviate this problem: we use a second cross entropy loss to minimize the distance between the learned soft prompts and a set of hand-engineered manual prompts (obtained by prompt engineering). The proposed loss can be interpreted in multiple ways including as a regularizer, as a means for language-based augmentation, and as a way of learning more discriminative class centroids. Importantly, our formulation is inherently amenable to including, during training, virtual classes, i.e. class names for which no visual samples are available, further increasing the robustness of the learned prompts. Through extensive evaluations on 11 datasets, we show that our approach (a) significantly outperforms all prior works on soft prompting, and (b) matches and surpasses, for the first time, the accuracy on novel classes obtained by hand-crafted prompts and CLIP for the majority of the test datasets. Code will be made available.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a second cross entropy loss to minimize the distance between the learned soft prompts and a set of hand-engineered manual prompts (obtained by prompt engineering), and can be interpreted in multiple ways including as a regularizer, as a means for language-based augmentation, and as a way of learning more discriminating class centroids.'}",http://arxiv.org/pdf/2210.01115
-ChatGPT4PCG Competition: Character-like Level Generation for Science Birds,Pittawat Taveekitworachai,"This paper presents the first ChatGPT4PCG Competition at the 2023 IEEE Conference on Games. The objective of this competition is for participants to create effective prompts for ChatGPT--enabling it to generate Science Birds levels with high stability and character-like qualities--fully using their creativity as well as prompt engineering skills. ChatGPT is a conversational agent developed by OpenAI. Science Birds is selected as the competition platform because designing an Angry Birds-like level is not a trivial task due to the in-game gravity; the quality of the levels is determined by their stability. To lower the entry barrier to the competition, we limit the task to the generation of capitalized English alphabetical characters. We also allow only a single prompt to be used for generating all the characters. Here, the quality of the generated levels is determined by their stability and similarity to the given characters. A sample prompt is provided to participants for their reference. An experiment is conducted to determine the effectiveness of several modified versions of this sample prompt on level stability and similarity by testing them on several characters. To the best of our knowledge, we believe that ChatGPT4PCG is the first competition of its kind and hope to inspire enthusiasm for prompt engineering in procedural content generation.",,http://arxiv.org/pdf/2303.15662
-Advancements in Casualty Modelling Facilitated by the USGS Prompt Assessment of Global Earthquakes for Response (PAGER) System,D. Wald,,,
-A Review on Expert System and its Applications in Civil Engineering,Muhammad Akram,"Aim of Study This study provides in depth review of expert system and potential benefits achieved with its application in the field of the Civil Engineering. Need of Study Currently, construction projects are facing various chronic problems. One of the reason contributing to these problems is delay in decision making. This leads to need of mechanisms which can enable practitioners in making prompt decision. Hence, expert systems are investigated for seeking opportunities and studying applicability in construction projects to use as a tool for aiding in decision making process. Research Approach This study is carried out through reviewing previous studies conducting in addressing the importance and issues of the expert system. It also includes various research works demonstrating the applications of an expert system in different research areas of civil engineering. Research Findings This study highlighted that expert system is very useful approach for benefiting the practitioner in making quick decision. It will be helpful in avoiding waiting times. It can be successfully applied in the areas of civil engineering specifically construction management. Limitations This paper presents a review on expert system with its application. It has focused on four established methods which are Rule based system (RBS), Case based system (CBS), Fuzzy expert system (FS), Neural network (NN). Importance and Contribution This review paper contributes in the field of Civil Engineering in highlighting the usage and advantages of an expert system.","{'model': 'tldr@v2.0.0', 'text': 'In depth review of expert system and potential benefits achieved with its application in the field of the Civil Engineering highlights that expert system is very useful approach for benefiting the practitioner in making quick decision.'}",
-A Real-Time Analysis and Feedback System for Quality Control of Dam Foundation Grouting Engineering,D. Zhong,,,
-Guidelines for environmental infection control in health-care facilities. Recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee (HICPAC).,Lynne Sehulster,"The health-care facility environment is rarely implicated in disease transmission, except among patients who are immunocompromised. Nonetheless, inadvertent exposures to environmental pathogens (e.g., Aspergillus spp. and Legionella spp.) or airborne pathogens (e.g., Mycobacterium tuberculosis and varicella-zoster virus) can result in adverse patient outcomes and cause illness among health-care workers. Environmental infection-control strategies and engineering controls can effectively prevent these infections. The incidence of health-care--associated infections and pseudo-outbreaks can be minimized by 1) appropriate use of cleaners and disinfectants; 2) appropriate maintenance of medical equipment (e.g., automated endoscope reprocessors or hydrotherapy equipment); 3) adherence to water-quality standards for hemodialysis, and to ventilation standards for specialized care environments (e.g., airborne infection isolation rooms, protective environments, or operating rooms); and 4) prompt management of water intrusion into the facility. Routine environmental sampling is not usually advised, except for water quality determinations in hemodialysis settings and other situations where sampling is directed by epidemiologic principles, and results can be applied directly to infection-control decisions. This report reviews previous guidelines and strategies for preventing environment-associated infections in health-care facilities and offers recommendations. These include 1) evidence-based recommendations supported by studies; 2) requirements of federal agencies (e.g., Food and Drug Administration, U.S. Environmental Protection Agency, U.S. Department of Labor, Occupational Safety and Health Administration, and U.S. Department of Justice); 3) guidelines and standards from building and equipment professional organizations (e.g., American Institute of Architects, Association for the Advancement of Medical Instrumentation, and American Society of Heating, Refrigeration, and Air-Conditioning Engineers); 4) recommendations derived from scientific theory or rationale; and 5) experienced opinions based upon infection-control and engineering practices. The report also suggests a series of performance measurements as a means to evaluate infection-control efforts.","{'model': 'tldr@v2.0.0', 'text': 'This report reviews previous guidelines and strategies for preventing environment-associated infections in health-care facilities and offers recommendations, including evidence-based recommendations supported by studies and experienced opinions based upon infection-control and engineering practices.'}",
-Improved Charge Separation in Poly(heptazine-triazine) Imides with Semi-coherent Interfaces for Photocatalytic Hydrogen Evolution.,Jian Zhang,"The construction of heterojunctions is a promising manner to accelerate the separation and transfer of the charge carriers at the interface. Herein, a binary poly(heptazine-triazine) imides (PHI/PTI) with semi-coherent interfaces was fabricated via a facile two-step salt-melt synthetic process. The built-in electric fields at the semi-coherent interface endow prompt exciton splitting and charge carrier separation. Hence, the optimized PHI/PTI-based copolymer presents a high apparent quantum yield (AQY=64%) for visible-light driven hydrogen production, by the aids of K 2 HPO 4 as charge transfer mediator. This study provides physical insights for the rational promotion of the photocatalytic performance from the viewpoint of interfacial engineering of photocatalytic junctions on crystalline carbon nitride based semiconductors.",,
-StyleMC: Multi-Channel Based Fast Text-Guided Image Generation and Manipulation,Umut Kocasari,"Discovering meaningful directions in the latent space of GANs to manipulate semantic attributes typically requires large amounts of labeled data. Recent work aims to overcome this limitation by leveraging the power of Contrastive Language-Image Pre-training (CLIP), a joint text-image model. While promising, these methods require several hours of preprocessing or training to achieve the desired manipulations. In this paper, we present StyleMC, a fast and efficient method for text-driven image generation and manipulation. StyleMC uses a CLIP-based loss and an identity loss to manipulate images via a single text prompt without significantly affecting other attributes. Unlike prior work, StyleMC requires only a few seconds of training per text prompt to find stable global directions, does not require prompt engineering and can be used with any pre-trained StyleGAN2 model. We demonstrate the effectiveness of our method and compare it to state-of-the-art methods. Our code can be found at http://catlab-team.github.io/stylemc.","{'model': 'tldr@v2.0.0', 'text': 'StyleMC is presented, a fast and efficient method for text-driven image generation and manipulation that uses a CLIP-based loss and an identity loss to manipulate images via a single text prompt without significantly affecting other attributes.'}",https://arxiv.org/pdf/2112.08493
-Controllable Image Captioning via Prompting,Ning Wang,"Despite the remarkable progress of image captioning, existing captioners typically lack the controllable capability to generate desired image captions, e.g., describing the image in a rough or detailed manner, in a factual or emotional view, etc. In this paper, we show that a unified model is qualified to perform well in diverse domains and freely switch among multiple styles. Such a controllable capability is achieved by embedding the prompt learning into the image captioning framework. To be specific, we design a set of prompts to fine-tune the pre-trained image captioner. These prompts allow the model to absorb stylized data from different domains for joint training, without performance degradation in each domain. Furthermore, we optimize the prompts with learnable vectors in the continuous word embedding space, avoiding the heuristic prompt engineering and meanwhile exhibiting superior performance. In the inference stage, our model is able to generate desired stylized captions by choosing the corresponding prompts. Extensive experiments verify the controllable capability of the proposed method. Notably, we achieve outstanding performance on two diverse image captioning benchmarks including COCO Karpathy split and TextCaps using a unified model.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs a set of prompts to fine-tune the pre-trained image captioner and designs a model that is qualified to perform well in diverse domains and freely switch among multiple styles by embedding the prompt learning into the image captioning framework.'}",https://arxiv.org/pdf/2212.01803
-Will It Blend? Mixing Training Paradigms & Prompting for Argument Quality Prediction,Michiel van der Meer,"This paper describes our contributions to the Shared Task of the 9th Workshop on Argument Mining (2022). Our approach uses Large Language Models for the task of Argument Quality Prediction. We perform prompt engineering using GPT-3, and also investigate the training paradigms multi-task learning, contrastive learning, and intermediate-task training. We find that a mixed prediction setup outperforms single models. Prompting GPT-3 works best for predicting argument validity, and argument novelty is best estimated by a model trained using all three training paradigms.","{'model': 'tldr@v2.0.0', 'text': 'This paper performs prompt engineering using GPT-3, and investigates the training paradigms multi-task learning, contrastive learning, and intermediate-task training to find that a mixed prediction setup outperforms single models.'}",http://arxiv.org/pdf/2209.08966
-The Infinite Index: Information Retrieval on Generative Text-To-Image Models,Niklas Deckers,"Conditional generative models such as DALL-E and Stable Diffusion generate images based on a user-defined text, the prompt. Finding and refining prompts that produce a desired image has become the art of prompt engineering. Generative models do not provide a built-in retrieval model for a user’s information need expressed through prompts. In light of an extensive literature review, we reframe prompt engineering for generative models as interactive text-based retrieval on a novel kind of “infinite index”. We apply these insights for the first time in a case study on image generation for game design with an expert. Finally, we envision how active learning may help to guide the retrieval of generated images.","{'model': 'tldr@v2.0.0', 'text': 'This work reframe prompt engineering for generative models as interactive text-based retrieval on a novel kind of “infinite index”, and envision how active learning may help to guide the retrieval of generated images.'}",https://dl.acm.org/doi/pdf/10.1145/3576840.3578327
-CLIP-CLOP: CLIP-Guided Collage and Photomontage,Piotr Wojciech Mirowski,"The unabated mystique of large-scale neural networks, such as the CLIP dual image-and-text encoder, popularized automatically generated art. Increasingly more sophisticated generators enhanced the artworks' realism and visual appearance, and creative prompt engineering enabled stylistic expression. Guided by an artist-in-the-loop ideal, we design a gradient-based generator to produce collages. It requires the human artist to curate libraries of image patches and to describe (with prompts) the whole image composition, with the option to manually adjust the patches' positions during generation, thereby allowing humans to reclaim some control of the process and achieve greater creative freedom. We explore the aesthetic potentials of high-resolution collages, and provide an open-source Google Colab as an artistic tool.",,http://arxiv.org/pdf/2205.03146
-What GPT Knows About Who is Who,Xiaohan Yang,"Coreference resolution – which is a crucial task for understanding discourse and language at large – has yet to witness widespread benefits from large language models (LLMs). Moreover, coreference resolution systems largely rely on supervised labels, which are highly expensive and difficult to annotate, thus making it ripe for prompt engineering. In this paper, we introduce a QA-based prompt-engineering method and discern generative, pre-trained LLMs’ abilities and limitations toward the task of coreference resolution. Our experiments show that GPT-2 and GPT-Neo can return valid answers, but that their capabilities to identify coreferent mentions are limited and prompt-sensitive, leading to inconsistent results.","{'model': 'tldr@v2.0.0', 'text': 'A QA-based prompt-engineering method is introduced and generative, pre-trained LLMs’ abilities and limitations toward the task of coreference resolution are discerned, revealing limited capabilities to identify coreferent mentions and inconsistent results.'}",https://arxiv.org/pdf/2205.07407
-The Amount of Errors in ChatGPT’s Responses is Indirectly Correlated with the Number of Publications Related to the Topic Under Investigation,K. Metze,,"{'model': 'tldr@v2.0.0', 'text': 'This study designed a study analyzing whether ChatGPT could contribute in writing a critical review on Chagas disease, focusing on the role of individual researchers.'}",
-Exploring the Benefits of Visual Prompting in Differential Privacy,Yizhe Li,"Visual Prompting (VP) is an emerging and powerful technique that allows sample-efficient adaptation to downstream tasks by engineering a well-trained frozen source model. In this work, we explore the benefits of VP in constructing compelling neural network classifiers with differential privacy (DP). We explore and integrate VP into canonical DP training methods and demonstrate its simplicity and efficiency. In particular, we discover that VP in tandem with PATE, a state-of-the-art DP training method that leverages the knowledge transfer from an ensemble of teachers, achieves the state-of-the-art privacy-utility trade-off with minimum expenditure of privacy budget. Moreover, we conduct additional experiments on cross-domain image classification with a sufficient domain gap to further unveil the advantage of VP in DP. Lastly, we also conduct extensive ablation studies to validate the effectiveness and contribution of VP under DP consideration. Our code is available at (https://github.com/EzzzLi/Prompt-PATE).","{'model': 'tldr@v2.0.0', 'text': 'This work discovers that VP in tandem with PATE, a state-of-the-art DP training method that leverages the knowledge transfer from an ensemble of teachers, achieves the state- of theart privacy-utility trade-off with minimum expenditure of privacy budget.'}",https://arxiv.org/pdf/2303.12247
-Beyond prompting: Making Pre-trained Language Models Better Zero-shot Learners by Clustering Representations,Yu Fei,"Recent work has demonstrated that pre-trained language models (PLMs) are zero-shot learners. However, most existing zero-shot methods involve heavy human engineering or complicated self-training pipelines, hindering their application to new situations. In this work, we show that zero-shot text classification can be improved simply by clustering texts in the embedding spaces of PLMs. Specifically, we fit the unlabeled texts with a Bayesian Gaussian Mixture Model after initializing cluster positions and shapes using class names. Despite its simplicity, this approach achieves superior or comparable performance on both topic and sentiment classification datasets and outperforms prior works significantly on unbalanced datasets. We further explore the applicability of our clustering approach by evaluating it on 14 datasets with more diverse topics, text lengths, and numbers of classes. Our approach achieves an average of 20% absolute improvement over prompt-based zero-shot learning. Finally, we compare different PLM embedding spaces and find that texts are well-clustered by topics even if the PLM is not explicitly pre-trained to generate meaningful sentence embeddings. This work indicates that PLM embeddings can categorize texts without task-specific fine-tuning, thus providing a new way to analyze and utilize their knowledge and zero-shot learning ability.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that zero-shot text classification can be improved simply by clustering texts in the embedding spaces of PLMs, indicating that PLM embeddings can categorize texts without task-specific fine-tuning, thus providing a new way to analyze and utilize their knowledge and zero- shot learning ability.'}",http://arxiv.org/pdf/2210.16637
-Beat over the old ground with new strategy: engineering As···As interaction in arsenite-based Dawson cluster β-[W18O54(AsO3)2(6-.,Q. Han,"By reaction of [As2W19O67(H2O)](14-), NiCl2·6H2O, and phen under hydrothermal conditions, a new organic-inorganic tungstoarsenate hybrid [Ni(phen)3]4[As2W18O60]{[Ni(phen)2][H2As2W18O60]}·12H2O (where phen = 1,10-phenanthroline) (1) was obtained via self-assembly and characterized by elemental analysis, infrared (IR) spectroscopy, solid UV-vis absorption spectrum, and single-crystal X-ray diffraction (XRD). An unprecedented 18-tungstoarsenate Dawson cluster, β-[W18O54(AsO3)2](6-), encapsulating two pyramidal arsenite AsO3(3-) anions as templates and exhibiting interesting intramolecular As···As interaction is first achieved. 1 displays a one-dimensional (1D) chain architecture constructed by alternating β-[W18O54(AsO3)2](6-) and nickel(II) complexes [Ni(phen)2)](2+). The resulting hybrid can act as a photocatalyst to prompt the degradation of Rhodamine B (RhB) with excellent efficiency.","{'model': 'tldr@v2.0.0', 'text': 'A new organic-inorganic tungstoarsenate hybrid that can act as a photocatalyst to prompt the degradation of Rhodamine B (RhB) with excellent efficiency is presented.'}",
-Thickness‐Dependent Facet Junction Control of Layered BiOIO3 Single Crystals for Highly Efficient CO2 Photoreduction,Fang Chen,"Thin layer fabrication and crystal facet engineering favor the prompt charge transfer from bulk to the surface of a material and spatial charge separation among different facets, tremendously benefitting photocatalytic activity. However, the thickness and surface facet composition are considered as two entwined characteristics of layered materials with well‐defined and tunable shapes, which possess great promise to achieve the simultaneous manipulation of charge transfer and spatial separation. Herein, it is demonstrated that one solution for the aforementioned issue by controllably regulating the surface {010}/{100} facet junctions of a layered thickness‐tunable bismuth‐based material, BiOIO3. The attenuation in thickness of BiOIO3 nanoplates shortens the diffusion pathway of charge carriers, and more importantly the tuning of nanolayer thickness renders the ratio variation of the top {010} facet to the lateral {100} facet, which dominates the spatial separation of photogenerated electrons and holes. As a result, the highest CO evolution rate from CO2 reduction over BiOIO3 nanoplates with the optimal thickness and ratio of exposed facets reaches 5.42 µmol g−1 h−1, over 300% that of the bulk counterpart (1.77 µmol g−1 h−1). This work paves a new way for governing charge movement behaviors on the basis of the synergistic engineering of layer structure and exposing facets.",,
-TextGraphs-16 Natural Language Premise Selection Task: Zero-Shot Premise Selection with Prompting Generative Language Models,Liubov Kovriguina,"Automated theorem proving can benefit a lot from methods employed in natural language processing, knowledge graphs and information retrieval: this non-trivial task combines formal languages understanding, reasoning, similarity search. We tackle this task by enhancing semantic similarity ranking with prompt engineering, which has become a new paradigm in natural language understanding. None of our approaches requires additional training. Despite encouraging results reported by prompt engineering approaches for a range of NLP tasks, for the premise selection task vanilla re-ranking by prompting GPT-3 doesn’t outperform semantic similarity ranking with SBERT, but merging of the both rankings shows better results.","{'model': 'tldr@v2.0.0', 'text': 'This work focuses on enhancing semantic similarity ranking with prompt engineering, which has become a new paradigm in natural language understanding, and introduces a new approach to premise selection task.'}",
-Generating Requirements Elicitation Interview Scripts with Large Language Models,Binnur Görer,"Requirements elicitation interviews are the most popular requirements elicitation technique and an integral part of requirements engineering education. Good and bad interview scripts provide students with examples of applying the theory. Constructing an interview script requires technical knowledge, practical experience, and creativity. As a result, only a few educational interview scripts are available to the community. This paper explores automatically generating interview scripts with large language models through prompt engineering. Our contribution is two-fold: First, we present a graph representation of interactive interview scripts. Second, we apply prompt engineering techniques to generate business domain descriptions, linear scripts, and conversation pieces focused on certain types of mistakes. Our findings indicate that large language models face challenges in handling interview conversation graphs. However, we can enhance the quality of the generated interview scripts by decomposing the task into smaller components and refining the prompts to provide more precise instructions.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate that large language models face challenges in handling interview conversation graphs, but the quality of the generated interview scripts can be enhanced by decomposing the task into smaller components and refining the prompts to provide more precise instructions.'}",
-An Empirical Evaluation of Prompting Strategies for Large Language Models in Zero-Shot Clinical Natural Language Processing,S. Sivarajkumar,"Large language models (LLMs) have shown remarkable capabilities in Natural Language Processing (NLP), especially in domains where labeled data is scarce or expensive, such as clinical domain. However, to unlock the clinical knowledge hidden in these LLMs, we need to design effective prompts that can guide them to perform specific clinical NLP tasks without any task-specific training data. This is known as in-context learning, which is an art and science that requires understanding the strengths and weaknesses of different LLMs and prompt engineering approaches. In this paper, we present a comprehensive and systematic experimental study on prompt engineering for five clinical NLP tasks: Clinical Sense Disambiguation, Biomedical Evidence Extraction, Coreference Resolution, Medication Status Extraction, and Medication Attribute Extraction. We assessed the prompts proposed in recent literature, including simple prefix, simple cloze, chain of thought, and anticipatory prompts, and introduced two new types of prompts, namely heuristic prompting and ensemble prompting. We evaluated the performance of these prompts on three state-of-the-art LLMs: GPT-3.5, BARD, and LLAMA2. We also contrasted zero-shot prompting with few-shot prompting, and provide novel insights and guidelines for prompt engineering for LLMs in clinical NLP. To the best of our knowledge, this is one of the first works on the empirical evaluation of different prompt engineering approaches for clinical NLP in this era of generative AI, and we hope that it will inspire and inform future research in this area.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a comprehensive and systematic experimental study on prompt engineering for five clinical NLP tasks: Clinical Sense Disambiguation, Biomedical Evidence Extraction, Coreference Resolution, Medication Status Ext extraction, and Medication Attribute Extraction.'}",https://arxiv.org/pdf/2309.08008
-Study on Prompt Methane Hydrate Formation Derived by Addition of Ionic Liquid,T. Kitajima,"Aims: The objective of this study is to establish the fundamental model on methane hydrate formation and to accelerate the rate of methane hydrate formation with a small amount of ionic liquid and to investigate the effect of ionic liquid on hydrate formation. Study Design: Experimental study containing modeling. Place and Duration of Study: The present study was held between April 2010 and February 2012 at Division of Chemical Engineering, Department of Materials Engineering Science, Osaka University. Methodology: Methane hydrate formation was modelized based on the driving force, fugacity difference before and after hydrate formation. BMIM-hexafuorophosphate (BMIMPF6) was adopted as a representative of 1-butyl-3-methylimidazolium (BMIM) salts. The temperature dependence of methane hydrate formation rate was investigated and activation energy of hydrate formation was evaluated for the pure water and BMIM-PF6 aqueous solution systems. Research Article American Chemical Science Journal, 2(3): 100-110, 2012 101 Results: An addition of small amount of BMIM-PF6 is able to accelerate the methane hydrate formation. The pseudo-first order reaction model is applicable to the methane hydrate formation in both the pure water and BMIM-PF6 aqueous solution systems. The activation energies of methane hydrate formation are large negative values in the both systems, that is, the methane hydrate formation process is considered to be composed of the precursory hydration and succeeding hydrate formation. A very small amount of BMIMPF6 seems to change the interfacial energy between guest molecules and precursor or initial hydrate particles without the change of the activation energy for overall methane hydrate formation.",,https://zenodo.org/record/8284/files/1345297018-Kitajima_etal_2-3-2012ACSj1512.pdf
-An Experimental Investigation of Analogy Formation Using the Engineering-to-Biology Thesaurus,M. Glier,"Identifying applicable biological systems for engineering design remains a persistent challenge for bioinspired design. Previous researchers have proposed an Engineering-to-Biology thesaurus that allows designers to identify biological keywords that are functionally similar to terms in the Functional Basis. This work presents an experimental examination of the effectiveness of the Engineering-to-Biology thesaurus. A group of 100 mechanical engineering students are presented with a simple design problem: to create a device to remove the husk and silk from ears of corn. The participants read passages drawn from a large biology corpus with keywords from the Engineering-to-Biology thesaurus and indicate which passages prompt some idea for solving the design problem. The analysis of student responses indicates that students’ level of design training is not a significant factor in the number of analogies they found in the passages and that some non-random criteria is used to identify passages as useful for idea generation. Passages that rarely offer participants ideas can be reasonably well classified as either being too technical for a lay-reader to understand or lacking information on a biological system. Passages that typically offer ideas cannot be so easily classified. Finally, keywords from the Engineering-to-Biology thesaurus are examined to find that keywords very specific to biology and those that are very common words with multiple meanings are rarely contained in sentences that offer strong design inspiration.Copyright © 2013 by ASME",,
-Solving Probability and Statistics Problems by Program Synthesis,Leonard Tang,"We solve university level probability and statistics questions by program synthesis using OpenAI's Codex, a Transformer trained on text and fine-tuned on code. We transform course problems from MIT's 18.05 Introduction to Probability and Statistics and Harvard's STAT110 Probability into programming tasks. We then execute the generated code to get a solution. Since these course questions are grounded in probability, we often aim to have Codex generate probabilistic programs that simulate a large number of probabilistic dependencies to compute its solution. Our approach requires prompt engineering to transform the question from its original form to an explicit, tractable form that results in a correct program and solution. To estimate the amount of work needed to translate an original question into its tractable form, we measure the similarity between original and transformed questions. Our work is the first to introduce a new dataset of university-level probability and statistics problems and solve these problems in a scalable fashion using the program synthesis capabilities of large language models.","{'model': 'tldr@v2.0.0', 'text': 'This work is the first to introduce a new dataset of university-level probability and statistics problems and solve these problems in a scalable fashion using the program synthesis capabilities of large language models.'}",
-The Rising Engineering Education Faculty Experience (REEFE): Preparing Junior Colleagues,Cory Hixson,"Despite the importance of professional development, for most graduate students as up-andcoming faculty members professional development is informal at best. Graduate programs often emphasize gaining technical knowledge, skills, and abilities through courses and research projects, but provide less opportunity for future faculty members to gain experience with teaching, service, communication, assessment, proposal writing, etc. To provide this experience, we developed the Rising Engineering Education Faculty Experience (REEFE). Founded on theoretical and practical models of graduate student development, REEFE is an innovative faculty apprenticeship program for engineering education graduate students that places students in visiting faculty member positions at host schools. This paper describes the foundations of REEFE and the program itself. We also offer lessons learned from the host school, sending school, and participants based on prior REEFE implementations. We hope our learnings prompt discussions regarding how to effectively prepare future engineering education faculty.","{'model': 'tldr@v2.0.0', 'text': 'The foundations of REEFE are described, an innovative faculty apprenticeship program for engineering education graduate students that places students in visiting faculty member positions at host schools and lessons learned from the host school, sending school, and participants are offered.'}",
-Economic Bases of Formation and Development of Financial Engineering in Financial Innovation,A. Ajupov,"Today the main spheres of society are developed mainly through the using of innovative technologies and constant updating of products. The result of active innovative activity of the economic subjects directed on the solution of different financial problems including the management all of increasing innovative risks became the prompt progress of financial engineering.м In turn, the process of designing innovative financial products created to reduce the various types of risks through instruments of financial engineering is characterized in economic science as a risk-engineering. DOI: 10.5901/mjss.2014.v5n24p148",,https://www.richtmann.org/journal/index.php/mjss/article/download/4953/4794
-The meaning of prompt feedback and other student perceptions of feedback: should National Student Survey scores be taken at face value?,P. M. Mendes,"Abstract Regardless of the institution or discipline, the UK’s National Student Survey (NSS) has consistently highlighted that the level of student satisfaction about the feedback they receive is notably lower than for other aspects of their learning experience. This study explored how students understand concepts and practices rated through NSS questions evaluating feedback practices in higher education. Drawing on questionnaires completed by first, third and fourth year Chemical Engineering students, the study calls into particular question the reliability of NSS data on promptness of feedback. In conclusion, it calls for greater attention to be paid at institutional level to the identification and management of students’ perceptions and expectations of the process, content and outcomes of feedback.",,
-An Empirical Study on Few-shot Knowledge Probing for Pretrained Language Models,Tianxing He,"Prompt-based knowledge probing for 1-hop relations has been used to measure how much world knowledge is stored in pretrained language models. Existing work uses considerable amounts of data to tune the prompts for better performance. In this work, we compare a variety of approaches under a few-shot knowledge probing setting, where only a small number (e.g., 10 or 20) of example triples are available. In addition, we create a new dataset named TREx-2p, which contains 2-hop relations. We report that few-shot examples can strongly boost the probing performance for both 1-hop and 2-hop relations. In particular, we find that a simple-yet-effective approach of finetuning the bias vectors in the model outperforms existing prompt-engineering methods. Our dataset and code are available at \url{https://github.com/cloudygoose/fewshot_lama}.","{'model': 'tldr@v2.0.0', 'text': 'A simple-yet-effective approach of finetuning the bias vectors in the model outperforms existing prompt-engineering methods under a few-shot knowledge probing setting, where only a small number of example triples are available.'}",
-Feature Engineering in the NLI Shared Task 2013: Charles University Submission Report,Barbora Hladká,"Our goal is to predict the first language (L1) of English essays’s authors with the help of the TOEFL11 corpus where L1, prompts (topics) and proficiency levels are provided. Thus we approach this task as a classification task employing machine learning methods. Out of key concepts of machine learning, we focus on feature engineering. We design features across all the L1 languages not making use of knowledge of prompt and proficiency level. During system development, we experimented with various techniques for feature filtering and combination optimized with respect to the notion of mutual information and information gain. We trained four different SVM models and combined them through majority voting achieving accuracy 72.5%.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to predict the first language (L1) of English essays’s authors with the help of the TOEFL11 corpus where L1, prompts (topics) and proficiency levels are provided and focuses on feature engineering.'}",
-Sepsis diagnosis and treatment using nanomaterials,Jaesung Lim,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the recent research advancements in nanotechnology to diagnose and treat sepsis with a brief introduction to sepsi is presented.'}",https://link.springer.com/content/pdf/10.1007/s13534-021-00200-0.pdf
-Optimal location of an intake at a reservoir prone to salt diffusion,A. Krutov,"There are two main factors which prompt to look for an alternative to traditional approaches to the provision of drinking water. They are climate change and population growth. Climate change and its increasing scale are no longer in doubt. It tends to increase the frequency and intensity of droughts. Changes in average water availability in most Central Asian river basins are estimated to be drastically big for the next 30 years. And, groundwater recharge may also be affected with a reduction in the availability of groundwater for drinking water in some regions. Water use has been increasing worldwide by about 1% per year since the 1980s, driven by a combination of population growth, socio-economic development and changing consumption patterns. Global water demand is expected to continue increasing at a similar rate until 2050, accounting for an increase of 20 to 30% above the current level of water use, mainly due to rising demand in the industrial and domestic sectors. These obliging to redefine the strategy for the use of water resources to ensure sustainable drinking water supply forcing the use of all available water resources even those that were not previously taken into account, for example, off stream storage reservoirs. Research, development and innovation play an important role in supporting informed decision-making. Therefore, further scientific and engineering studies are also needed for the development of financially affordable, safe and efficient infrastructure services in the areas of drinking water supply, sanitation and hygiene and its components. The article considers a particular reservoir the bottom of which is composed of saline soils operating in regular filling and emptying and subject to significant wind effects. The research conducted allowed to find a place of water intake location which provides a minimum salt content in the water abstracted.",,
-Combining computer vision and deep learning to enable ultra-scale aerial phenotyping and precision agriculture: A case study of lettuce production,Alan Bauer,,"{'model': 'tldr@v2.0.0', 'text': 'AirSurf is reported, an automated and open-source analytic platform that combines modern computer vision, up-to-date machine learning, and modular software engineering in order to measure yield-related phenotypes from ultra-large aerial imagery.'}",https://www.nature.com/articles/s41438-019-0151-5.pdf
-Should Engineering Ethics be Taught?,Charles J. Abaté,,"{'model': 'tldr@v2.0.0', 'text': 'There is a justification to teaching engineering ethics, insofar as the authors are able to clearly identify the most desirable and efficacious pedagogical approach to the subject area, which is proposed to be a case study-based format that utilizes the principle of human cognitive pattern recognition.'}",
-Generating Disentangled Arguments with Prompts: A Simple Event Extraction Framework That Works,Jinghui Si,"Event Extraction bridges the gap between text and event signals. Based on the assumption of trigger-argument dependency, existing approaches have achieved state-of-the-art performance with expert-designed templates or complicated decoding constraints. In this paper, for the first time we introduce the prompt-based learning strategy to the domain of Event Extraction, which empowers the automatic exploitation of label semantics on both input and output sides. To validate the effectiveness of the proposed generative method, we conduct extensive experiments with 11 diverse baselines. Empirical results show that, in terms of F1 score on Argument Extraction, our simple architecture is stronger than any other generative counterpart and even competitive with algorithms that require template engineering. Regarding the measure of recall, it sets new overall records for both Argument and Trigger Extractions. We hereby recommend this framework to the community, with the code publicly available at https://github.com/RingBDStack/GDAP.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the prompt-based learning strategy to the domain of Event Extraction, which empowers the automatic exploitation of label semantics on both input and output sides and sets new overall records for both Argument and Trigger Extractions.'}",https://eprints.whiterose.ac.uk/191435/1/jinghui_GDAP_icassp2022.pdf
-PROMPT VERSUS PROBLEM: HELPING STUDENTS LEARN TO FRAME PROBLEMS AND THINK CREATIVELY,J. Lai,"Currently engineering students are given opportunities to apply knowledge learned in the classroom to hands-on design projects. While being trained to solve problems, are creativity skills also developed as part of that experience? According to a recent study, surveys found that students do not feel that creativity is part of that experience (Kazerounian and Foley 2007). Two questions are posed to frame discussion: how does current curriculum shape the way students approach problems? What can be done to create assignments that encourage creative thinking? Four observations are made from a class that focuses on creativity and design.",,
-Prompt magnetic investigations of archaeological remains in areas of infrastructure development: Israeli experience,L. Eppelbaum,"Magnetic prospecting is the most widely used method for the investigation of archaeological remains in the world. This method is often applied before and during various engineering exploration (e.g. road construction, preparing areas for various types of building) and areas of agricultural utilization. Taking into account an extremely high concentration of archaeological sites of different age at Israeli territory, prompt geophysical–archaeological assessment of areas allocated for engineering and agricultural utilization, is a necessary preliminary component of the design process. However, noise both of natural and artificial origin strongly complicates interpretation of observed magnetic anomalies. The most significant disturbing factors are oblique magnetization (magnetization vector in the Israeli territory is inclined about 45°), complex geological structure (presence of a variety of disturbing bodies) of investigated site, influence of rugged topography, as well as influence of modern iron‐containing objects. The non‐conventional procedures developed by the authors are intended for application in such complicated environments and include elimination of noise from various sources, selection of ‘useful’ anomalies against the noise background, qualitative and quantitative interpretation of anomalies and three‐dimensional physical–geological modelling. Effective integration of these procedures with other interpretation methods allowed rapid localization of tens of archaeological remains and their protection from unpremeditated destruction. Copyright © 2001 John Wiley & Sons, Ltd.",,
-Climate Engineering Responses to Climate Emergencies,J. Blackstock,"Despite efforts to stabilize CO_2 concentrations, it is possible that the climate system could respond abruptly with catastrophic consequences. Intentional intervention in the climate system to avoid or ameliorate such consequences has been proposed as one possible response, should such a scenario arise. In a one-week study, the authors of this report conducted a technical review and evaluation of proposed climate engineering concepts that might serve as a rapid palliative response to such climate emergency scenarios. Because of their potential to induce a prompt (less than one year) global cooling, this study concentrated on Shortwave Climate Engineering (SWCE) methods for moderately reducing the amount of shortwave solar radiation reaching the Earth. The study's main objective was to outline a decade-long agenda of technical research that would maximally reduce the uncertainty surrounding the benefits and risks associated with SWCE. For rigor of technical analysis, the study focused the research agenda on one particular SWCE concept--stratospheric aerosol injection--and in doing so developed several conceptual frameworks and methods valuable for assessing any SWCE proposal.",,http://arxiv.org/pdf/0907.5140
-Does Your Configuration Code Smell?,Tushar Sharma,"Infrastructure as Code (IaC) is the practice of specifying computing system configurations through code, and managing them through traditional software engineering methods. The wide adoption of configuration management and increasing size and complexity of the associated code, prompt for assessing, maintaining, and improving the configuration code's quality. In this context, traditional software engineering knowledge and best practices associated with code quality management can be leveraged to assess and manage configuration code quality. We propose a catalog of 13 implementation and 11 design configuration smells, where each smell violates recommended best practices for configuration code. We analyzed 4,621 Puppet repositories containing 8.9 million lines of code and detected the cataloged implementation and design configuration smells. Our analysis reveals that the design configuration smells show 9% higher average co-occurrence among themselves than the implementation configuration smells. We also observed that configuration smells belonging to a smell category tend to co-occur with configuration smells belonging to another smell category when correlation is computed by volume of identified smells. Finally, design configuration smell density shows negative correlation whereas implementation configuration smell density exhibits no correlation with the size of a configuration management system.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a catalog of 13 implementation and 11 design configuration smells, where each smell violates recommended best practices for configuration code, and finds that configuration smells belonging to a smell category tend to co-occur with configuration smell belonging to another smell category when correlation is computed by volume of identified smells.'}",https://zenodo.org/record/1997943/files/ConfigurationSmells_preprint.pdf
-Ethics and engineering education,L. Bucciarelli,"In the US, Accreditation Board for Engineering and Technology (ABET) recommends the study of ethics so that students acquire ‘an understanding of professional and ethical responsibility’. For the most part, teaching of the subject relies upon the use of scenarios – both hypothetical and ‘real’– and open discussion framed by the codes. These scenarios and this framing are seriously deficient – lacking in their attention to the complexities of context, almost solely focused on individual agency, while reflecting too narrow and simplistic a view of the responsibilities of the practicing engineer. A critique of several exemplary scenarios, and consideration of the demands placed upon today's professional, prompt reflection on the need for, not just a more expansive reading of the codes of ethics re what it might mean to be ‘responsible’, but a substantial reform of undergraduate engineering education across the board.",,https://dspace.mit.edu/bitstream/1721.1/40284/1/ethics_20_talk.pdf
-Contract cheating and assessment design: exploring the relationship,T. Bretag,"Abstract This paper reports on findings from a large Australian research project that explored the relationship between contract cheating and assessment design. Using survey responses from 14,086 students and 1147 educators at eight universities, a multivariate analysis examined the influence of a range of factors on the likelihood that different assessment types would prompt considerations of contract cheating in students. Perceptions of likelihood were highest among students who speak a language other than English at home. Perceptions of likelihood were also higher among students who reported there to be lots of opportunities to cheat, and amongst students who were dissatisfied with the teaching and learning environment. Perceptions of likelihood for certain assessment types were also higher in commerce and engineering than in any other discipline. Overall, four assessment types were perceived by students to be the least likely to be outsourced, however these are also the least likely to be set by educators. The analysis indicates that educators are more likely to use these assessment tasks when they report positively on organisational support for teaching and learning.",,
-The analysis of conventional Prompt Global Strike alternatives,R. P. Hopkins,"United States Strategic Command (STRATCOM) has expressed a strong interest in the development of a non-nuclear offensive strategy capable of reaching all enemy targets across the globe within a sixty-minute period from when the launching order is given to when the target is engaged. STRATCOM has defined this mission as conventional Prompt Global Strike (PGS). Currently, STRATCOM believes the United States does not possess any sort of conventional weapon system capable of meeting the criteria of a conventional PGS weapon. The conventional PGS mission calls for a solution that can be implemented immediately. This means the solution must contain only technology that is currently available. As a result of our literature review and our stakeholder analysis, our problem statement is that STRATCOM lacks a conventional PGS capability to hold time-sensitive targets at risk in denied access areas worldwide. To solve this problem, a systems engineering approach is applied to examine different alternatives by using Value-Focused Thinking to develop value models, math modeling to generate performance data, ArcGIS to highlight the system need and to recommend an implementation plan, and a simulation to demonstrate weapon system survivability. This work will result in recommendations for conventional Prompt Global Strike alternatives.",,
-EEG-based Depression Detection Using Convolutional Neural Network with Demographic Attention Mechanism,Xiaowei Zhang,"Electroencephalography (EEG)-based depression detection has become a hot topic in the development of biomedical engineering. However, the complexity and nonstationarity of EEG signals are two biggest obstacles to this application. In addition, the generalization of detection algorithms may be degraded owing to the influences brought by individual differences. In view of the correlation between EEG signals and individual demographics, such as gender, age, etc., and influences of these demographic factors on the incidence of depression, it would be better to incorporate demographic factors during EEG modeling and depression detection. In this work, we constructed an one-dimensional Convolutional Neural Network (1-D CNN) to obtain more effective features of EEG signals, then integrated gender and age factors into the 1-D CNN via an attention mechanism, which could prompt our 1-D CNN to explore complex correlations between EEG signals and demographic factors, and generate more effective high-level representations ultimately for the detection of depression. Experimental results on 170 (81 depressed patients and 89 normal controls) subjects showed that the proposed method is superior to the unitary 1-D CNN without gender and age factors and two other ways of incorporating demographics. This work also indicates that organic mixture of EEG signals and demographic factors is promising for the detection of depression.Clinical relevance-This work indicates that organically mixture of EEG signals and demographic factors is promising for the detection of depression","{'model': 'tldr@v2.0.0', 'text': 'Experimental results showed that the proposed method is superior to the unitary 1-D CNN without gender and age factors and two other ways of incorporating demographics, and indicates that organic mixture of EEG signals and demographic factors is promising for the detection of depression.'}",
-Platelet-Rich Fibrin Scaffolds for Cartilage and Tendon Regenerative Medicine: From Bench to Bedside,S. Barbon,"Nowadays, research in Tissue Engineering and Regenerative Medicine is focusing on the identification of instructive scaffolds to address the requirements of both clinicians and patients to achieve prompt and adequate healing in case of injury. Among biomaterials, hemocomponents, and in particular Platelet-rich Fibrin matrices, have aroused widespread interest, acting as delivery platforms for growth factors, cytokines and immune/stem-like cells for immunomodulation; their autologous origin and ready availability are also noteworthy aspects, as safety- and cost-related factors and practical aspects make it possible to shorten surgical interventions. In fact, several authors have focused on the use of Platelet-rich Fibrin in cartilage and tendon tissue engineering, reporting an increasing number of in vitro, pre-clinical and clinical studies. This narrative review attempts to compare the relevant advances in the field, with particular reference being made to the regenerative role of platelet-derived growth factors, as well as the main pre-clinical and clinical research on Platelet-rich Fibrin in chondrogenesis and tenogenesis, thereby providing a basis for critical revision of the topic.","{'model': 'tldr@v2.0.0', 'text': 'A narrative review attempts to compare the relevant advances in the field, with particular reference being made to the regenerative role of platelet-derived growth factors, as well as the main pre-clinical and clinical research on Platelet-rich Fibrin in chondrogenesis and tenogenesis, thereby providing a basis for critical revision of the topic.'}",https://www.mdpi.com/1422-0067/20/7/1701/pdf?version=1554453987
-A cytochrome c is the natural electron acceptor for nicotine oxidoreductase,Mark Dulchavsky,,"{'model': 'tldr@v2.0.0', 'text': 'It is discovered thatNicA2 donates electrons to a cytochrome c, which means that NicA2 is actually a dehydrogenase, which is surprising, as enzymes of the flavin-containing amine oxidase family were invariably thought to use O2 as an electron acceptor.'}",
-Engineering of Escherichia coli for targeted delivery of transgenes to HER2/neu‐positive tumor cells,Chih-Hsiang Chang,"Targeting of non‐phagocytic tumor cells and prompt release of gene cargos upon entry into tumors are two limiting steps in the bacterial gene delivery path. To tackle these problems, the non‐pathogenic Escherichia coli strain BL21(DE3) was engineered to display the anti‐HER2/neu affibody on the surface. After co‐incubation with tumor cells for 3 h, the anti‐HER2/neu affibody‐presenting E. coli strain was selectively internalized into HER2/neu‐positive SKBR‐3 cells. The invasion efficiency reached as high as 30%. Furthermore, the bacteria were equipped with the phage ϕX174 lysin gene E‐mediated autolysis system. Carrying the transgene (e.g., eukaryotic green fluorescent protein, GFP), the tumor‐targeting bacteria were subjected to the thermal shock to trigger the autolysis system upon entry into HER2/neu‐positive cells. Flow cytometric analysis revealed that 3% of infected cells expressed GFP 24 h post thermal induction. Overall, the results show a promise of the proposed approach for developing bacteria as a delivery carrier. Biotechnol. Bioeng. 2011; 108:1662–1672. © 2011 Wiley Periodicals, Inc.","{'model': 'tldr@v2.0.0', 'text': 'The results show a promise of the proposed approach for developing bacteria as a delivery carrier for treating non‐phagocytic tumor cells and prompt release of gene cargos upon entry into tumors.'}",http://ir.cmu.edu.tw/ir/bitstream/310903500/40944/1/20110502193414.pdf
-Higgs-Mediated Optical Amplification in a Nonequilibrium Superconductor,M. Buzzi,"The quest for new functionalities in quantum materials has recently been extended to non-equilibrium states, which are interesting both because they exhibit new physical phenomena and because of their potential for high-speed device applications. Notable advances have been made in the creation of metastable phases and in Floquet engineering under external periodic driving. In the context of non-equilibrium superconductivity, examples have included the generation of transient superconductivity above the thermodynamic transition temperature, the excitation of coherent Higgs mode oscillations, and the optical control of the interlayer phase in cuprates. Here, we propose theoretically a novel non-equilibrium phenomenon, through which a prompt quench from a metal to a transient superconducting state could induce large oscillations of the order parameter amplitude. We argue that this oscillating mode could act as a source of parametric amplification of the incident radiation. We report experimental results on optically driven K$_3$C$_{60}$ that are consistent with these predictions. The effect is found to disappear when the onset of the excitation becomes slower than the Higgs mode period, consistent with the theory proposed here. These results open new possibilities for the use of collective modes in many-body systems to induce non-linear optical effects.",,http://link.aps.org/pdf/10.1103/PhysRevX.11.011055
-Review of Smart Health Monitoring Approaches With Survey Analysis and Proposed Framework,Sonal Gahlot,"Smart health monitoring system technologies pertaining to the biomedical domain has evinced intense interest, which in turn has led the emphasis on unconventional applications of engineering and internetwork transmission in the field. Throughout the next series of textual and contextual matter, we will be reviewing and exhibiting, such initiatives and established methods which are being used to achieve health monitoring systems approaching toward making smart remote villages and cities. It also collates for the reader the issues and challenges prevalent in the present technical advancements and for societal validation, a survey analysis has also been briefed in this paper to show the demand and requisites of the real world in health monitoring system. Today there is a need for an integrated solution which can diagnose the early cyanotic congenital diseases in newly born babies to monitoring of various geriatric problems in elderly people. The future avenues of maximum life expectancy and uber-prompt medical treatment depend upon it. This paper proposes an approach required to make smart villages and towns with respect to the healthcare domain. This proposal has been made with due consideration and technical backing of survey research including questionnaire-based evaluation and regressive analysis of issues in present technologies.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an approach required to make smart villages and towns with respect to the healthcare domain with due consideration and technical backing of survey research including questionnaire-based evaluation and regressive analysis of issues in present technologies.'}",
-Investigation of hydrogen content in chemically delithiated lithium-ion battery cathodes using prompt gamma activation analysis,S. Aghara,,,
-Evolutionary genome engineering using a restriction–modification system,Y. Asakura,"Modification of complex microbial cellular processes is often necessary to obtain organisms with particularly favorable characteristics, but such experiments can take many generations to achieve. In the present article, we accelerated the experimental evolution of Escherichia coli populations under selection for improved growth using one of the restriction–modification systems, which have shaped bacterial genomes. This resulted in faster evolutionary changes in both the genome and bacterial growth. Transcriptome/genome analysis at various stages enabled prompt identification of sequential genome rearrangements and dynamic gene-expression changes associated with growth improvement. The changes were related to cell-to-cell communication, the cell death program, as well as mass production and energy consumption. These observed changes imply that improvements in microorganism population growth can be achieved by inactivating the cellular mechanisms regulating fraction of active cells in a population. Some of the mutations were shown to have additive effects on growth. These results open the way for the application of evolutionary genome engineering to generate organisms with desirable properties.","{'model': 'tldr@v2.0.0', 'text': 'This article accelerated the experimental evolution of Escherichia coli populations under selection for improved growth using one of the restriction–modification systems, which have shaped bacterial genomes, resulting in faster evolutionary changes in both the genome and bacterial growth.'}",https://academic.oup.com/nar/article-pdf/39/20/9034/16779366/gkr585.pdf
-Sterile and Dual-Porous Aerogels Scaffolds Obtained through a Multistep Supercritical CO2-Based Approach,Víctor Santos-Rosales,"Aerogels from natural polymers are endowed with attractive textural and biological properties for biomedical applications due to their high open mesoporosity, low density, and reduced toxicity. Nevertheless, the lack of macroporosity in the aerogel structure and of a sterilization method suitable for these materials restrict their use for regenerative medicine purposes and prompt the research on getting ready-to-implant dual (macro + meso)porous aerogels. In this work, zein, a family of proteins present in materials for tissue engineering, was evaluated as a sacrificial porogen to obtain macroporous starch aerogels. This approach was particularly advantageous since it could be integrated in the conventional aerogel processing method without extra leaching steps. Physicochemical, morphological, and mechanical characterization were performed to study the effect of porogen zein at various proportions (0:1, 1:2, and 1:1 zein:starch weight ratio) on the properties of the obtained starch-based aerogels. From a forward-looking perspective for its clinical application, a supercritical CO2 sterilization treatment was implemented for these aerogels. The sterilization efficacy and the influence of the treatment on the aerogel final properties were evaluated mainly in terms of absence of microbial growth, cytocompatibility, as well as physicochemical, structural, and mechanical modifications.",,https://www.mdpi.com/1420-3049/24/5/871/pdf?version=1551432957
-3D Hollow Hierarchical Structures Based on 1D BiOCl Nanorods Intersected with 2D Bi2WO6 Nanosheets for Efficient Photocatalysis Under Visible Light,Yongchao Ma,"Constructing elaborate catalysts to prompt the charge carrier separation and transport is critical to developing efficient photocatalytic systems. Here, a hierarchical hollow structure based on 1D/2D BiOCl/Bi2WO6 hybrid materials was fabricated by a precursor chemical engineering method. This hybrid is made up of molten 1D BiOCl nanorods and 2D Bi2WO6 nanosheets. The synergetic effect of the presence of BiOCl and specific interfaces between BiOCl and Bi2WO6 provided efficient interfacial charge transfer of photogenerated carriers under visible light. Seamless BiOCl functions like a noble metal, with platinum-like behavior, accelerating the oxidizing ability of fabricated BiOCl/Bi2WO6 hybrids, which was favorable for the photocatalytic decomposition of organic compounds (3.2 times greater for Rhodamine B (RhB) and 4 times greater for Ciprofloxacin (CIP)) over the Bi2WO6 catalysts. The beneficial interfacial interaction between BiOCl and Bi2WO6 resulting from the unique construction prompted the charge transfer from the conduction band of Bi2WO6 to that of BiOCl. The findings presented in this study provide a cost-effective precursor-mediated strategy to realize the critical and efficient separation of photoinduced carriers in environmental remediation applications.","{'model': 'tldr@v2.0.0', 'text': 'The findings presented in this study provide a cost-effective precursor-mediated strategy to realize the critical and efficient separation of photoinduced carriers in environmental remediation applications.'}",https://www.mdpi.com/2079-4991/9/3/322/pdf?version=1551422128
-Prototypical Networks for Few-shot Learning,Jake Snell,"We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prototypical Networks for few-shot classification, and provides an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning.'}",
-Flamingo: a Visual Language Model for Few-Shot Learning,Jean-Baptiste Alayrac,"Building models that can be rapidly adapted to novel tasks using only a handful of annotated examples is an open challenge for multimodal machine learning research. We introduce Flamingo, a family of Visual Language Models (VLM) with this ability. We propose key architectural innovations to: (i) bridge powerful pretrained vision-only and language-only models, (ii) handle sequences of arbitrarily interleaved visual and textual data, and (iii) seamlessly ingest images or videos as inputs. Thanks to their flexibility, Flamingo models can be trained on large-scale multimodal web corpora containing arbitrarily interleaved text and images, which is key to endow them with in-context few-shot learning capabilities. We perform a thorough evaluation of our models, exploring and measuring their ability to rapidly adapt to a variety of image and video tasks. These include open-ended tasks such as visual question-answering, where the model is prompted with a question which it has to answer; captioning tasks, which evaluate the ability to describe a scene or an event; and close-ended tasks such as multiple-choice visual question-answering. For tasks lying anywhere on this spectrum, a single Flamingo model can achieve a new state of the art with few-shot learning, simply by prompting the model with task-specific examples. On numerous benchmarks, Flamingo outperforms models fine-tuned on thousands of times more task-specific data.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Flamingo, a family of Visual Language Models (VLM) with this ability to bridge powerful pretrained vision-only and language-only models, handle sequences of arbitrarily interleaved visual and textual data, and seamlessly ingest images or videos as inputs.'}",
-Learning to Compare: Relation Network for Few-Shot Learning,Flood Sung,"We present a conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each. Our method, called the Relation Network (RN), is trained end-to-end from scratch. During meta-learning, it learns to learn a deep distance metric to compare a small number of images within episodes, each of which is designed to simulate the few-shot setting. Once trained, a RN is able to classify images of new classes by computing relation scores between query images and the few examples of each new class without further updating the network. Besides providing improved performance on few-shot learning, our framework is easily extended to zero-shot learning. Extensive experiments on five benchmarks demonstrate that our simple approach provides a unified and effective approach for both of these two tasks.","{'model': 'tldr@v2.0.0', 'text': 'A conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each, which is easily extended to zero- shot learning.'}",https://www.pure.ed.ac.uk/ws/files/57835996/LearningToCompare.pdf
-Matching Networks for One Shot Learning,Oriol Vinyals,"Learning from a few examples remains a key challenge in machine learning. Despite recent advances in important domains such as vision and language, the standard supervised deep learning paradigm does not offer a satisfactory solution for learning new concepts rapidly from little data. In this work, we employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories. Our framework learns a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types. We then define one-shot learning problems on vision (using Omniglot, ImageNet) and language tasks. Our algorithm improves one-shot accuracy on ImageNet from 87.6% to 93.2% and from 88.0% to 93.8% on Omniglot compared to competing approaches. We also demonstrate the usefulness of the same model on language modeling by introducing a one-shot task on the Penn Treebank.","{'model': 'tldr@v2.0.0', 'text': 'This work employs ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories to learn a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types.'}",
-Optimization as a Model for Few-Shot Learning,S. Ravi,,,
-Generalizing from a Few Examples: A Survey on Few-Shot Learning,Yaqing Wang,"Machine learning has been highly successful in data-intensive applications but is often hampered when the data set is small. Recently, Few-Shot Learning (FSL) is proposed to tackle this problem. Using prior knowledge, FSL can rapidly generalize to new tasks containing only a few samples with supervised information. In this paper, we conduct a thorough survey to fully understand FSL. Starting from a formal definition of FSL, we distinguish FSL from several relevant machine learning problems. We then point out that the core issue in FSL is that the empirical risk minimized is unreliable. Based on how prior knowledge can be used to handle this core issue, we categorize FSL methods from three perspectives: (i) data, which uses prior knowledge to augment the supervised experience; (ii) model, which uses prior knowledge to reduce the size of the hypothesis space; and (iii) algorithm, which uses prior knowledge to alter the search for the best hypothesis in the given hypothesis space. With this taxonomy, we review and discuss the pros and cons of each category. Promising directions, in the aspects of the FSL problem setups, techniques, applications and theories, are also proposed to provide insights for future research.","{'model': 'tldr@v2.0.0', 'text': 'A thorough survey to fully understand Few-Shot Learning (FSL), and categorizes FSL methods from three perspectives: data, which uses prior knowledge to augment the supervised experience; model, which used to reduce the size of the hypothesis space; and algorithm, which using prior knowledgeto alter the search for the best hypothesis in the given hypothesis space.'}",
-TADAM: Task dependent adaptive metric for improved few-shot learning,Boris N. Oreshkin,"Few-shot learning has become essential for producing models that generalize from few examples. In this work, we identify that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms. Our analysis reveals that simple metric scaling completely changes the nature of few-shot algorithm parameter updates. Metric scaling provides improvements up to 14% in accuracy for certain metrics on the mini-Imagenet 5-way 5-shot classification task. We further propose a simple and effective way of conditioning a learner on the task sample set, resulting in learning a task-dependent metric space. Moreover, we propose and empirically test a practical end-to-end optimization procedure based on auxiliary task co-training to learn a task-dependent metric space. The resulting few-shot learning model based on the task-dependent scaled metric achieves state of the art on mini-Imagenet. We confirm these results on another few-shot dataset that we introduce in this paper based on CIFAR100. Our code is publicly available at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms and proposes and empirically test a practical end-to-end optimization procedure based on auxiliary task co-training to learn a task-dependent metric space.'}",
-Few-shot Learning with Retrieval Augmented Language Models,Gautier Izacard,"Large language models have shown impressive few-shot results on a wide range of tasks. However, when knowledge is key for such results, as is the case for tasks such as question answering and fact checking, massive parameter counts to store knowledge seem to be needed. Retrieval augmented models are known to excel at knowledge intensive tasks without the need for as many parameters, but it is unclear whether they work in few-shot settings. In this work we present Atlas, a carefully designed and pre-trained retrieval augmented language model able to learn knowledge intensive tasks with very few training examples. We perform evaluations on a wide range of tasks, including MMLU, KILT and NaturalQuestions, and study the impact of the content of the document index, showing that it can easily be updated. Notably, Atlas reaches over 42% accuracy on Natural Questions using only 64 examples, outperforming a 540B parameters model by 3% despite having 50x fewer parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Atlas, a carefully designed and pre-trained retrieval augmented language model able to learn knowledge intensive tasks with very few training examples, and studies the impact of the content of the document index, showing that it can easily be updated.'}",https://arxiv.org/pdf/2208.03299
-Meta-SGD: Learning to Learn Quickly for Few Shot Learning,Zhenguo Li,"Few-shot learning is challenging for learning algorithms that learn each task in isolation and from scratch. In contrast, meta-learning learns from many related tasks a meta-learner that can learn a new task more accurately and faster with fewer examples, where the choice of meta-learners is crucial. In this paper, we develop Meta-SGD, an SGD-like, easily trainable meta-learner that can initialize and adapt any differentiable learner in just one step, on both supervised learning and reinforcement learning. Compared to the popular meta-learner LSTM, Meta-SGD is conceptually simpler, easier to implement, and can be learned more efficiently. Compared to the latest meta-learner MAML, Meta-SGD has a much higher capacity by learning to learn not just the learner initialization, but also the learner update direction and learning rate, all in a single meta-learning process. Meta-SGD shows highly competitive performance for few-shot learning on regression, classification, and reinforcement learning.","{'model': 'tldr@v2.0.0', 'text': 'Meta-SGD, an SGD-like, easily trainable meta-learner that can initialize and adapt any differentiable learner in just one step, shows highly competitive performance for few-shot learning on regression, classification, and reinforcement learning.'}",
-Multimodal Few-Shot Learning with Frozen Language Models,Maria Tsimpoukelli,"When trained at sufficient scale, auto-regressive language models exhibit the notable ability to learn a new language task after being prompted with just a few examples. Here, we present a simple, yet effective, approach for transferring this few-shot learning ability to a multimodal setting (vision and language). Using aligned image and caption data, we train a vision encoder to represent each image as a sequence of continuous embeddings, such that a pre-trained, frozen language model prompted with this prefix generates the appropriate caption. The resulting system is a multimodal few-shot learner, with the surprising ability to learn a variety of new tasks when conditioned on examples, represented as a sequence of multiple interleaved image and text embeddings. We demonstrate that it can rapidly learn words for new objects and novel visual categories, do visual question-answering with only a handful of examples, and make use of outside knowledge, by measuring a single model on a variety of established and new benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'The resulting system is a multimodal few-shot learner, with the surprising ability to learn a variety of new tasks when conditioned on examples, represented as a sequence of multiple interleaved image and text embeddings.'}",
-Few-Shot Learning with Graph Neural Networks,Victor Garcia Satorras,"We propose to study the problem of few-shot learning with the prism of inference on a partially observed graphical model, constructed from a collection of input images whose label can be either observed or not. By assimilating generic message-passing inference algorithms with their neural-network counterparts, we define a graph neural network architecture that generalizes several of the recently proposed few-shot learning models. Besides providing improved numerical performance, our framework is easily extended to variants of few-shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.","{'model': 'tldr@v2.0.0', 'text': ""A graph neural network architecture is defined that generalizes several of the recently proposed few-shot learning models and provides improved numerical performance, and is easily extended to variants of few- shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.""}",
-CANet: Class-Agnostic Segmentation Networks With Iterative Refinement and Attentive Few-Shot Learning,Chi Zhang,"Recent progress in semantic segmentation is driven by deep Convolutional Neural Networks and large-scale labeled image datasets. However, data labeling for pixel-wise segmentation is tedious and costly. Moreover, a trained model can only make predictions within a set of pre-defined classes. In this paper, we present CANet, a class-agnostic segmentation network that performs few-shot segmentation on new classes with only a few annotated images available. Our network consists of a two-branch dense comparison module which performs multi-level feature comparison between the support image and the query image, and an iterative optimization module which iteratively refines the predicted results. Furthermore, we introduce an attention mechanism to effectively fuse information from multiple support examples under the setting of k-shot learning. Experiments on PASCAL VOC 2012 show that our method achieves a mean Intersection-over-Union score of 55.4% for 1-shot segmentation and 57.1% for 5-shot segmentation, outperforming state-of-the-art methods by a large margin of 14.6% and 13.2%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Canet is presented, a class-agnostic segmentation network that performs few-shot segmentation on new classes with only a few annotated images available, and introduces an attention mechanism to effectively fuse information from multiple support examples under the setting of k-shot learning.'}",https://dr.ntu.edu.sg/bitstream/10356/144391/2/gusoheng%20paper1%20cvpr%202019.pdf
-FLEURS: FEW-Shot Learning Evaluation of Universal Representations of Speech,Alexis Conneau,"We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Speech-Text Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like speech-only w2v-BERT [1] and speech-text multimodal mSLAM [2]. The goal of FLEURS is to enable speech technology in more languages and catalyze research in low-resource speech understanding.1.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2205.12446
-Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference,S. Hu,"Few-shot learning (FSL) is an important and topical problem in computer vision that has motivated extensive research into numerous methods spanning from sophisticated metalearning methods to simple transfer learning baselines. We seek to push the limits of a simple-but-effective pipeline for real-worldfew-shot image classification in practice. To this end, we explore few-shot learning from the perspective of neural architecture, as well as a three stage pipeline of pre-training on external data, meta-training with labelled few-shot tasks, and task-specific fine-tuning on unseen tasks. We investigate questions such as: ① How pre-training on external data benefits FSL? ② How state of the art transformer architectures can be exploited? and ③ How to best exploit finetuning? Ultimately, we show that a simple transformer-based pipeline yields surprisingly good performance on standard benchmarks such as Mini-ImageNet, CIFAR-FS, CDFSL and Meta-Dataset. Our code is available at https://hushell.github.io/pmf.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a simple transformer-based pipeline yields surprisingly good performance on standard benchmarks such as Mini-ImageNet, CIFAR-FS, CDFSL and Meta-Dataset.'}",https://www.pure.ed.ac.uk/ws/files/262579857/Pushing_the_Limits_HU_DOA02032022_AFV.pdf
-Meta-Transfer Learning for Few-Shot Learning,Qianru Sun,"Meta-learning has been proposed as a framework to address the challenging few-shot learning setting. The key idea is to leverage a large number of similar few-shot tasks in order to learn how to adapt a base-learner to a new task for which only a few labeled samples are available. As deep neural networks (DNNs) tend to overfit using a few samples only, meta-learning typically uses shallow neural networks (SNNs), thus limiting its effectiveness. In this paper we propose a novel few-shot learning method called meta-transfer learning (MTL) which learns to adapt a deep NN for few shot learning tasks. Specifically, ""meta"" refers to training multiple tasks, and ""transfer"" is achieved by learning scaling and shifting functions of DNN weights for each task. In addition, we introduce the hard task (HT) meta-batch scheme as an effective learning curriculum for MTL. We conduct experiments using (5-class, 1-shot) and (5-class, 5-shot) recognition tasks on two challenging few-shot learning benchmarks: miniImageNet and Fewshot-CIFAR100. Extensive comparisons to related works validate that our meta-transfer learning approach trained with the proposed HT meta-batch scheme achieves top performance. An ablation study also shows that both components contribute to fast convergence and high accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A novel few-shot learning method called meta-transfer learning (MTL) which learns to adapt a deep NN for few shot learning tasks and introduces the hard task (HT) meta-batch scheme as an effective learning curriculum for MTL.'}",
-Graph Information Aggregation Cross-Domain Few-Shot Learning for Hyperspectral Image Classification.,Yuxiang Zhang,"Most domain adaptation (DA) methods in cross-scene hyperspectral image classification focus on cases where source data (SD) and target data (TD) with the same classes are obtained by the same sensor. However, the classification performance is significantly reduced when there are new classes in TD. In addition, domain alignment, as one of the main approaches in DA, is carried out based on local spatial information, rarely taking into account nonlocal spatial information (nonlocal relationships) with strong correspondence. A graph information aggregation cross-domain few-shot learning (Gia-CFSL) framework is proposed, intending to make up for the above-mentioned shortcomings by combining FSL with domain alignment based on graph information aggregation. SD with all label samples and TD with a few label samples are implemented for FSL episodic training. Meanwhile, intradomain distribution extraction block (IDE-block) and cross-domain similarity aware block (CSA-block) are designed. The IDE-block is used to characterize and aggregate the intradomain nonlocal relationships and the interdomain feature and distribution similarities are captured in the CSA-block. Furthermore, feature-level and distribution-level cross-domain graph alignments are used to mitigate the impact of domain shift on FSL. Experimental results on three public HSI datasets demonstrate the superiority of the proposed method. The codes will be available from the website: https://github.com/YuxiangZhang-BIT/IEEE_TNNLS_Gia-CFSL.","{'model': 'tldr@v2.0.0', 'text': 'A graph information aggregation cross-domain few-shot learning (Gia-CFSL) framework is proposed, intending to make up for the above-mentioned shortcomings by combining FSL with domain alignment based on graph information aggregating.'}",
-Efficient Few-Shot Learning Without Prompts,Lewis Tunstall,"Recent few-shot methods, such as parameter-efficient fine-tuning (PEFT) and pattern exploiting training (PET), have achieved impressive results in label-scarce settings. However, they are difficult to employ since they are subject to high variability from manually crafted prompts, and typically require billion-parameter language models to achieve high accuracy. To address these shortcomings, we propose SetFit (Sentence Transformer Fine-tuning), an efficient and prompt-free framework for few-shot fine-tuning of Sentence Transformers (ST). SetFit works by first fine-tuning a pretrained ST on a small number of text pairs, in a contrastive Siamese manner. The resulting model is then used to generate rich text embeddings, which are used to train a classification head. This simple framework requires no prompts or verbalizers, and achieves high accuracy with orders of magnitude less parameters than existing techniques. Our experiments show that SetFit obtains comparable results with PEFT and PET techniques, while being an order of magnitude faster to train. We also show that SetFit can be applied in multilingual settings by simply switching the ST body. Our code is available at https://github.com/huggingface/setfit and our datasets at https://huggingface.co/setfit .","{'model': 'tldr@v2.0.0', 'text': 'This work proposes SetFit (Sentence Transformer Fine-tuning), an efficient and prompt-free framework for few-shot fine- Tuning of Sentence Transformers (ST), which achieves high accuracy with orders of magnitude less parameters than existing techniques.'}",http://arxiv.org/pdf/2209.11055
-Learning to Propagate Labels: Transductive Propagation Network for Few-Shot Learning,Yanbin Liu,"The goal of few-shot learning is to learn a classifier that generalizes well even when trained with a limited number of training instances per class. The recently introduced meta-learning approaches tackle this problem by learning a generic classifier across a large number of multiclass classification tasks and generalizing the model to a new task. Yet, even with such meta-learning, the low-data problem in the novel classification task still remains. In this paper, we propose Transductive Propagation Network (TPN), a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem. Specifically, we propose to learn to propagate labels from labeled instances to unlabeled test instances, by learning a graph construction module that exploits the manifold structure in the data. TPN jointly learns both the parameters of feature embedding and the graph construction in an end-to-end manner. We validate TPN on multiple benchmark datasets, on which it largely outperforms existing few-shot learning approaches and achieves the state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Transductive Propagation Network (TPN), a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem.'}",
-Revisiting Local Descriptor Based Image-To-Class Measure for Few-Shot Learning,Wenbin Li,"Few-shot learning in image classification aims to learn a classifier to classify images when only few training examples are available for each class. Recent work has achieved promising classification performance, where an image-level feature based measure is usually used. In this paper, we argue that a measure at such a level may not be effective enough in light of the scarcity of examples in few-shot learning. Instead, we think a local descriptor based image-to-class measure should be taken, inspired by its surprising success in the heydays of local invariant features. Specifically, building upon the recent episodic training mechanism, we propose a Deep Nearest Neighbor Neural Network (DN4 in short) and train it in an end-to-end manner. Its key difference from the literature is the replacement of the image-level feature based measure in the final layer by a local descriptor based image-to-class measure. This measure is conducted online via a k-nearest neighbor search over the deep local descriptors of convolutional feature maps. The proposed DN4 not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity, thanks to the exchangeability of visual patterns across the images in the same class. Our work leads to a simple, effective, and computationally efficient framework for few-shot learning. Experimental study on benchmark datasets consistently shows its superiority over the related state-of-the-art, with the largest absolute improvement of 17% over the next best. The source code can be available from https://github.com/WenbinLee/DN4.git.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Deep Nearest Neighbor Neural Network (DN4), a simple, effective, and computationally efficient framework for few-shot learning that not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity.'}",https://arxiv.org/pdf/1903.12290
-Few-Shot Learning via Embedding Adaptation With Set-to-Set Functions,Han-Jia Ye,"Learning with limited data is a key challenge for visual recognition. Many few-shot learning methods address this challenge by learning an instance embedding function from seen classes and apply the function to instances from unseen classes with limited labels. This style of transfer learning is task-agnostic: the embedding function is not learned optimally discriminative with respect to the unseen classes, where discerning among them leads to the target task. In this paper, we propose a novel approach to adapt the instance embeddings to the target classification task with a set-to-set function, yielding embeddings that are task-specific and are discriminative. We empirically investigated various instantiations of such set-to-set functions and observed the Transformer is most effective --- as it naturally satisfies key properties of our desired model. We denote this model as FEAT (few-shot embedding adaptation w/ Transformer) and validate it on both the standard few-shot classification benchmark and four extended few-shot learning settings with essential use cases, i.e., cross-domain, transductive, generalized few-shot learning, and low-shot learning. It archived consistent improvements over baseline models as well as previous methods, and established the new state-of-the-art results on two benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach to adapt the instance embeddings to the target classification task with a set-to-set function, yielding embeddeddings that are task-specific and are discriminative.'}",https://arxiv.org/pdf/1812.03664
-Variational Few-Shot Learning for Microservice-Oriented Intrusion Detection in Distributed Industrial IoT,Wei Liang,"Along with the popularity of the Internet of Things (IoT) techniques with several computational paradigms, such as cloud and edge computing, microservice has been viewed as a promising architecture in large-scale application design and deployment. Due to the limited computing ability of edge devices in distributed IoT, only a small scale of data can be used for model training. In addition, most of the machine-learning-based intrusion detection methods are insufficient when dealing with imbalanced dataset under limited computing resources. In this article, we propose an optimized intra/inter-class-structure-based variational few-shot learning (OICS-VFSL) model to overcome a specific out-of-distribution problem in imbalanced learning, and to improve the microservice-oriented intrusion detection in distributed IoT systems. Following a newly designed VFSL framework, an intra/inter-class optimization scheme is developed using reconstructed feature embeddings, in which the intra-class distance is optimized based on the approximation during a variation Bayesian process, while the inter-class distance is optimized based on the maximization of similarities during a feature concatenation process. An intelligent intrusion detection algorithm is, then, introduced to improve the multiclass classification via a nonlinear neural network. Evaluation experiments are conducted using two public datasets to demonstrate the effectiveness of our proposed model, especially in detecting novel attacks with extremely imbalanced data, compared with four baseline methods.","{'model': 'tldr@v2.0.0', 'text': 'An optimized intra/inter-class-structure-based variational few-shot learning (OICS-VFSL) model is proposed to overcome a specific out-of-distribution problem in imbalanced learning, and to improve the microservice-oriented intrusion detection in distributed IoT systems.'}",https://ieeexplore.ieee.org/ielx7/9424/9772431/09551702.pdf
-Generalized Zero- and Few-Shot Learning via Aligned Variational Autoencoders,Edgar Schönfeld,"Many approaches in generalized zero-shot learning rely on cross-modal mapping between the image feature space and the class embedding space. As labeled images are expensive, one direction is to augment the dataset by generating either images or image features. However, the former misses fine-grained details and the latter requires learning a mapping associated with class embeddings. In this work, we take feature generation one step further and propose a model where a shared latent space of image features and class embeddings is learned by modality-specific aligned variational autoencoders. This leaves us with the required discriminative information about the image and classes in the latent features, on which we train a softmax classifier. The key to our approach is that we align the distributions learned from images and from side-information to construct latent features that contain the essential multi-modal information associated with unseen classes. We evaluate our learned latent features on several benchmark datasets, i.e. CUB, SUN, AWA1 and AWA2, and establish a new state of the art on generalized zero-shot as well as on few-shot learning. Moreover, our results on ImageNet with various zero-shot splits show that our latent features generalize well in large-scale settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a model where a shared latent space of image features and class embeddings is learned by modality-specific aligned variational autoencoders, and align the distributions learned from images and from side-information to construct latent features that contain the essential multi-modal information associated with unseen classes.'}",https://arxiv.org/pdf/1812.01784
-AlexaTM 20B: Few-Shot Learning Using a Large-Scale Multilingual Seq2Seq Model,Saleh Soltan,"In this work, we demonstrate that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling (CLM) tasks, are more efficient few-shot learners than decoder-only models on various tasks. In particular, we train a 20 billion parameter multilingual seq2seq model called Alexa Teacher Model (AlexaTM 20B) and show that it achieves state-of-the-art (SOTA) performance on 1-shot summarization tasks, outperforming a much larger 540B PaLM decoder model. AlexaTM 20B also achieves SOTA in 1-shot machine translation, especially for low-resource languages, across almost all language pairs supported by the model (Arabic, English, French, German, Hindi, Italian, Japanese, Marathi, Portuguese, Spanish, Tamil, and Telugu) on Flores-101 dataset. We also show in zero-shot setting, AlexaTM 20B outperforms GPT3 (175B) on SuperGLUE and SQuADv2 datasets and provides SOTA performance on multilingual tasks such as XNLI, XCOPA, Paws-X, and XWinograd. Overall, our results present a compelling case for seq2seq models as a powerful alternative to decoder-only models for Large-scale Language Model (LLM) training.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling tasks, are more efficient few-shot learners than decoder-only models on various tasks.'}",http://arxiv.org/pdf/2208.01448
-HyperTransformer: Model Generation for Supervised and Semi-Supervised Few-Shot Learning,A. Zhmoginov,"In this work we propose a HyperTransformer, a Transformer-based model for supervised and semi-supervised few-shot learning that generates weights of a convolutional neural network (CNN) directly from support samples. Since the dependence of a small generated CNN model on a specific task is encoded by a high-capacity Transformer model, we effectively decouple the complexity of the large task space from the complexity of individual tasks. Our method is particularly effective for small target CNN architectures where learning a fixed universal task-independent embedding is not optimal and better performance is attained when the information about the task can modulate all model parameters. For larger models we discover that generating the last layer alone allows us to produce competitive or better results than those obtained with state-of-the-art methods while being end-to-end differentiable.","{'model': 'tldr@v2.0.0', 'text': 'A HyperTransformer, a Transformer-based model for supervised and semi-supervised few-shot learning that generates weights of a convolutional neural network directly from support samples is proposed, which is particularly effective for small target CNN architectures.'}",
-Hierarchical Graph Neural Networks for Few-Shot Learning,Cen Chen,"Recent graph neural network (GNN) based methods for few-shot learning (FSL) represent the samples of interest as a fully-connected graph and conduct reasoning on the nodes flatly, which ignores the hierarchical correlations among nodes. However, real-world categories may have hierarchical structures, and for FSL, it is important to extract the distinguishing features of the categories from individual samples. To explore this, we propose a novel hierarchical graph neural network (HGNN) for FSL, which consists of three parts, i.e., bottom-up reasoning, top-down reasoning, and skip connections, to enable the efficient learning of multi-level relationships. For the bottom-up reasoning, we design intra-class k-nearest neighbor pooling (intra-class knnPool) and inter-class knnPool layers, to conduct hierarchical learning for both the intra- and inter-class nodes. For the top-down reasoning, we propose to utilize graph unpooling (gUnpool) layers to restore the down-sampled graph into its original size. Skip connections are proposed to fuse multi-level features for the final node classification. The parameters of HGNN are learned by episodic training with the signal of node losses, which aims to train a well-generalizable model for recognizing unseen classes with few labeled data. Experimental results on benchmark datasets have demonstrated that HGNN outperforms other state-of-the-art GNN based methods significantly, for both transductive and non-transductive FSL tasks. The dataset as well as the source code can be downloaded online1","{'model': 'tldr@v2.0.0', 'text': 'A novel hierarchical graph neural network (HGNN) for FSL is proposed, which consists of three parts, i.e., bottom-up reasoning, top-down reasoning, and skip connections, to enable the efficient learning of multi-level relationships.'}",
-Free Lunch for Few-shot Learning: Distribution Calibration,Shuo Yang,"Learning from a limited number of samples is challenging since the learned model can easily become overfitted based on the biased distribution formed by only a few training examples. In this paper, we calibrate the distribution of these few-sample classes by transferring statistics from the classes with sufficient examples, then an adequate number of examples can be sampled from the calibrated distribution to expand the inputs to the classifier. We assume every dimension in the feature representation follows a Gaussian distribution so that the mean and the variance of the distribution can borrow from that of similar classes whose statistics are better estimated with an adequate number of samples. Our method can be built on top of off-the-shelf pretrained feature extractors and classification models without extra parameters. We show that a simple logistic regression classifier trained using the features sampled from our calibrated distribution can outperform the state-of-the-art accuracy on two datasets (~5% improvement on miniImageNet compared to the next best). The visualization of these generated features demonstrates that our calibrated distribution is an accurate estimation.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a simple logistic regression classifier trained using the features sampled from the calibrated distribution can outperform the state-of-the-art accuracy on two datasets (~5% improvement on miniImageNet compared to the next best).'}",
-"A survey of few-shot learning in smart agriculture: developments, applications, and challenges",Jiachen Yang,,"{'model': 'tldr@v2.0.0', 'text': 'This review comprehensively expounds on few-shot learning in smart agriculture, introduces the definition of few-shots learning, four kinds of learning methods, the publicly available datasets for few- shot learning, various applications in smart Agriculture, and the challenges in smart farming in future development.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-022-00866-2
-"A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities",Yisheng Song,"Few-shot learning (FSL) has emerged as an effective learning method and shows great potential. Despite the recent creative works in tackling FSL tasks, learning valid information rapidly from just a few or even zero samples remains a serious challenge. In this context, we extensively investigated 200+ FSL papers published in top journals and conferences in the past three years, aiming to present a timely and comprehensive overview of the most recent advances in FSL with a fresh perspective and to provide an impartial comparison of the strengths and weaknesses of existing work. To avoid conceptual confusion, we first elaborate and contrast a set of relevant concepts including few-shot learning, transfer learning, and meta-learning. Then, we inventively extract prior knowledge related to few-shot learning in the form of a pyramid, which summarizes and classifies previous work in detail from the perspective of challenges. Furthermore, to enrich this survey, we present in-depth analysis and insightful discussions of recent advances in each subsection. What is more, taking computer vision as an example, we highlight the important application of FSL, covering various research hotspots. Finally, we conclude the survey with unique insights into technology trends and potential future research opportunities to guide FSL follow-up research.","{'model': 'tldr@v2.0.0', 'text': 'This survey extensively investigated 200+ FSL papers published in top journals and conferences in the past three years to present a timely and comprehensive overview of the most recent advances in FSL with a fresh perspective and to provide an impartial comparison of the strengths and weaknesses of existing work.'}",https://arxiv.org/pdf/2205.06743
-Learning from Few Examples: A Summary of Approaches to Few-Shot Learning,Archit Parnami,"Few-Shot Learning refers to the problem of learning the underlying pattern in the data just from a few training samples. Requiring a large number of data samples, many deep learning solutions suffer from data hunger and extensively high computation time and resources. Furthermore, data is often not available due to not only the nature of the problem or privacy concerns but also the cost of data preparation. Data collection, preprocessing, and labeling are strenuous human tasks. Therefore, few-shot learning that could drastically reduce the turnaround time of building machine learning applications emerges as a low-cost solution. This survey paper comprises a representative list of recently proposed few-shot learning algorithms. Given the learning dynamics and characteristics, the approaches to few-shot learning problems are discussed in the perspectives of meta-learning, transfer learning, and hybrid approaches (i.e., different variations of the few-shot learning problem).","{'model': 'tldr@v2.0.0', 'text': 'This survey paper comprises a representative list of recently proposed few-shot learning algorithms that could drastically reduce the turnaround time of building machine learning applications emerges as a low-cost solution.'}",http://arxiv.org/pdf/2203.04291
-Charting the Right Manifold: Manifold Mixup for Few-shot Learning,Puneet Mangla,"Few-shot learning algorithms aim to learn model parameters capable of adapting to unseen classes with the help of only a few labeled examples. A recent regularization technique - Manifold Mixup focuses on learning a general-purpose representation, robust to small changes in the data distribution. Since the goal of few-shot learning is closely linked to robust representation learning, we study Manifold Mixup in this problem setting. Self-supervised learning is another technique that learns semantically meaningful features, using only the inherent structure of the data. This work investigates the role of learning relevant feature manifold for few-shot tasks using self-supervision and regularization techniques. We observe that regularizing the feature manifold, enriched via self-supervised techniques, with Manifold Mixup significantly improves few-shot learning performance. We show that our proposed method S2M2 beats the current state-of-the-art accuracy on standard few-shot learning datasets like CIFAR-FS, CUB, mini-ImageNet and tiered-ImageNet by 3 − 8%. Through extensive experimentation, we show that the features learned using our approach generalize to complex few-shot evaluation tasks, cross-domain scenarios and are robust against slight changes to data distribution.","{'model': 'tldr@v2.0.0', 'text': 'This work observes that regularizing the feature manifold, enriched via self-supervised techniques, with Manifold Mixup significantly improves few-shot learning performance, and proposes the proposed method S2M2, which beats the current state-of-the-art accuracy on standard few- shot learning datasets.'}",http://raiith.iith.ac.in/5861/1/1907.12087.pdf
-ConFeSS: A Framework for Single Source Cross-Domain Few-Shot Learning,Debasmit Das,"Most current few-shot learning methods train a model from abundantly labeled base category data and then transfer and adapt the model to sparsely labeled novel category data. These methods mostly generalize well on novel categories from the same domain as the base categories but perform poorly for distant domain categories. In this paper, we propose a framework for few-shot learning coined as ConFeSS (Contrastive Learning and Feature Selection System) that tackles large domain shift between base and novel categories. The first step of our framework trains a feature extracting backbone with the contrastive loss on the base category data. Since the contrastive loss does not use supervision, the features can generalize better to distant target domains. For the second step, we train a masking module to select relevant features that are more suited to target domain classification. Finally, a classifier is fine-tuned along with the backbone such that the backbone produces features similar to the relevant ones. To evaluate our framework, we tested it on a recently introduced cross-domain few-shot learning benchmark. Experimental results demonstrate that our framework outperforms all meta-learning approaches and produces competitive results against recent cross-domain methods. Additional analyses are also performed to better understand our framework.","{'model': 'tldr@v2.0.0', 'text': 'A framework for few-shot learning coined as ConFeSS (Contrastive Learning and Feature Selection System) that tackles large domain shift between base and novel categories and outperforms all meta-learning approaches and produces competitive results against recent cross-domain methods is proposed.'}",
-True Few-Shot Learning with Language Models,Ethan Perez,"Pretrained language models (LMs) perform well on many tasks even when learning from a few examples, but prior work uses many held-out examples to tune various aspects of learning, such as hyperparameters, training objectives, and natural language templates (""prompts""). Here, we evaluate the few-shot ability of LMs when such held-out examples are unavailable, a setting we call true few-shot learning. We test two model selection criteria, cross-validation and minimum description length, for choosing LM prompts and hyperparameters in the true few-shot setting. On average, both marginally outperform random selection and greatly underperform selection based on held-out examples. Moreover, selection criteria often prefer models that perform significantly worse than randomly-selected ones. We find similar results even when taking into account our uncertainty in a model's true performance during selection, as well as when varying the amount of computation and number of examples used for selection. Overall, our findings suggest that prior work significantly overestimated the true few-shot ability of LMs given the difficulty of few-shot model selection.","{'model': 'tldr@v2.0.0', 'text': 'This work evaluates the few-shot ability of LMs when such held-out examples are unavailable, a setting the authors call true few- shot learning, and suggests that prior work significantly overestimated thetrue few-shots ability ofLMs given the difficulty of few-Shot model selection.'}",
-A Broader Study of Cross-Domain Few-Shot Learning,Yunhui Guo,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed Broader Study of Cross-Domain Few-Shot Learning (BSCD-FSL) benchmark, consisting of image data from a diverse assortment of image acquisition methods, demonstrates that state-of-art meta- learning methods are surprisingly outperformed by earlier meta-learning approaches, and all meta- Learning methods underperform in relation to simple fine-tuning.'}",https://arxiv.org/pdf/1912.07200
-Integrative Few-Shot Learning for Classification and Segmentation,Dahyun Kang,"We introduce the integrative task of few-shot classification and segmentation (FS-CS) that aims to both classify and segment target objects in a query image when the target classes are given with a few examples. This task combines two conventional few-shot learning problems, few-shot classification and segmentation. FS-CS generalizes them to more realistic episodes with arbitrary image pairs, where each target class may or may not be present in the query. To address the task, we propose the integrative few-shot learning (iFSL) framework for FS-CS, which trains a learner to construct class-wise foreground maps for multi-label classification and pixel-wise segmentation. We also develop an effective iFSL model, attentive squeeze network (ASNet), that leverages deep semantic correlation and global self-attention to produce reliable foreground maps. In experiments, the proposed method shows promising performance on the FS-CS task and also achieves the state of the art on standard few-shot segmentation benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'The integrative few-shot learning (iFSL) framework for FS-CS is proposed, which trains a learner to construct class-wise foreground maps for multi-label classification and pixel-wise segmentation, and an effective iFSL model is developed, attentive squeeze network (ASNet), that leverages deep semantic correlation and global self-attention to produce reliable foreground maps.'}",https://arxiv.org/pdf/2203.15712
-Edge-Labeling Graph Neural Network for Few-Shot Learning,Jongmin Kim,"In this paper, we propose a novel edge-labeling graph neural network (EGNN), which adapts a deep neural network on the edge-labeling graph, for few-shot learning. The previous graph neural network (GNN) approaches in few-shot learning have been based on the node-labeling framework, which implicitly models the intra-cluster similarity and the inter-cluster dissimilarity. In contrast, the proposed EGNN learns to predict the edge-labels rather than the node-labels on the graph that enables the evolution of an explicit clustering by iteratively updating the edge-labels with direct exploitation of both intra-cluster similarity and the inter-cluster dissimilarity. It is also well suited for performing on various numbers of classes without retraining, and can be easily extended to perform a transductive inference. The parameters of the EGNN are learned by episodic training with an edge-labeling loss to obtain a well-generalizable model for unseen low-data problem. On both of the supervised and semi-supervised few-shot image classification tasks with two benchmark datasets, the proposed EGNN significantly improves the performances over the existing GNNs.","{'model': 'tldr@v2.0.0', 'text': 'A novel edge-labeling graph neural network (EGNN) is proposed, which adapts a deep neural network on the edge- labeling graph, for few-shot learning and significantly improves the performances over the existing GNNs.'}",https://arxiv.org/pdf/1905.01436
-Improving In-Context Few-Shot Learning via Self-Supervised Training,Mingda Chen,"Self-supervised pretraining has made few-shot learning possible for many NLP tasks. But the pretraining objectives are not typically adapted specifically for in-context few-shot learning. In this paper, we propose to use self-supervision in an intermediate training stage between pretraining and downstream few-shot usage with the goal to teach the model to perform in-context few shot learning. We propose and evaluate four self-supervised objectives on two benchmarks. We find that the intermediate self-supervision stage produces models that outperform strong baselines. Ablation study shows that several factors affect the downstream performance, such as the amount of training data and the diversity of the self-supervised objectives. Human-annotated cross-task supervision and self-supervision are complementary. Qualitative analysis suggests that the self-supervised-trained models are better at following task requirements.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to use self-supervision in an intermediate training stage between pretraining and downstream few-shot usage with the goal to teach the model to perform in-context few shot learning.'}",https://arxiv.org/pdf/2205.01703
-Few-Shot Learning With Class-Covariance Metric for Hyperspectral Image Classification,Bobo Xi,"Recently, embedding and metric-based few-shot learning (FSL) has been introduced into hyperspectral image classification (HSIC) and achieved impressive progress. To further enhance the performance with few labeled samples, we in this paper propose a novel FSL framework for HSIC with a class-covariance metric (CMFSL). Overall, the CMFSL learns global class representations for each training episode by interactively using training samples from the base and novel classes, and a synthesis strategy is employed on the novel classes to avoid overfitting. During the meta-training and meta-testing, the class labels are determined directly using the Mahalanobis distance measurement rather than an extra classifier. Benefiting from the task-adapted class-covariance estimations, the CMFSL can construct more flexible decision boundaries than the commonly used Euclidean metric. Additionally, a lightweight cross-scale convolutional network (LXConvNet) consisting of 3D and 2D convolutions is designed to thoroughly exploit the spectral-spatial information in the high-frequency and low-frequency scales with low computational complexity. Furthermore, we devise a spectral-prior-based refinement module (SPRM) in the initial stage of feature extraction, which cannot only force the network to emphasize the most informative bands while suppressing the useless ones, but also alleviate the effects of the domain shift between the base and novel categories to learn a collaborative embedding mapping. Extensive experiment results on four benchmark data sets demonstrate that the proposed CMFSL can outperform the state-of-the-art methods with few-shot annotated samples.","{'model': 'tldr@v2.0.0', 'text': 'A novel FSL framework for HSIC with a class-covariance metric (CMFSL), which can outperform the state-of-the-art methods with few-shot annotated samples and devise a spectral-prior-based refinement module (SPRM) in the initial stage of feature extraction.'}",
-Realistic Evaluation of Transductive Few-Shot Learning,Olivier Veilleux,"Transductive inference is widely used in few-shot learning, as it leverages the statistics of the unlabeled query set of a few-shot task, typically yielding substantially better performances than its inductive counterpart. The current few-shot benchmarks use perfectly class-balanced tasks at inference. We argue that such an artificial regularity is unrealistic, as it assumes that the marginal label probability of the testing samples is known and fixed to the uniform distribution. In fact, in realistic scenarios, the unlabeled query sets come with arbitrary and unknown label marginals. We introduce and study the effect of arbitrary class distributions within the query sets of few-shot tasks at inference, removing the class-balance artefact. Specifically, we model the marginal probabilities of the classes as Dirichlet-distributed random variables, which yields a principled and realistic sampling within the simplex. This leverages the current few-shot benchmarks, building testing tasks with arbitrary class distributions. We evaluate experimentally state-of-the-art transductive methods over 3 widely used data sets, and observe, surprisingly, substantial performance drops, even below inductive methods in some cases. Furthermore, we propose a generalization of the mutual-information loss, based on $\alpha$-divergences, which can handle effectively class-distribution variations. Empirically, we show that our transductive $\alpha$-divergence optimization outperforms state-of-the-art methods across several data sets, models and few-shot settings. Our code is publicly available at https://github.com/oveilleux/Realistic_Transductive_Few_Shot.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces and study the effect of arbitrary class distributions within the query sets of few-shot tasks at inference, removing the class-balance artefact, and proposes a generalization of the mutual-information loss, based on $\\alpha$-divergences, which can handle effectively class-distribution variations.'}",http://arxiv.org/pdf/2204.11181
-Deep Cross-Domain Few-Shot Learning for Hyperspectral Image Classification,Zhaokui Li,"One of the challenges in hyperspectral image (HSI) classification is that there are limited labeled samples to train a classifier for very high-dimensional data. In practical applications, we often encounter an HSI domain (called target domain) with very few labeled data, while another HSI domain (called source domain) may have enough labeled data. Classes between the two domains may not be the same. This article attempts to use source class data to help classify the target classes, including the same and new unseen classes. To address this classification paradigm, a meta-learning paradigm for few-shot learning (FSL) is usually adopted. However, existing FSL methods do not account for domain shift between source and target domain. To solve the FSL problem under domain shift, a novel deep cross-domain few-shot learning (DCFSL) method is proposed. For the first time, DCFSL tackles FSL and domain adaptation issues in a unified framework. Specifically, a conditional adversarial domain adaptation strategy is utilized to overcome domain shift, which can achieve domain distribution alignment. In addition, FSL is executed in source and target classes at the same time, which can not only discover transferable knowledge in the source classes but also learn a discriminative embedding model to the target classes. Experiments conducted on four public HSI data sets demonstrate that DCFSL outperforms the existing FSL methods and deep learning methods for HSI classification. Our source code is available at https://github.com/Li-ZK/DCFSL-2021.","{'model': 'tldr@v2.0.0', 'text': 'A novel deep cross-domain few-shot learning (DCFSL) method that tackles FSL and domain adaptation issues in a unified framework and demonstrates that DCFSL outperforms the existing FSL methods and deep learning methods for HSI classification.'}",
-A New Meta-Baseline for Few-Shot Learning,Yinbo Chen,"Meta-learning has become a popular framework for few-shot learning in recent years, with the goal of learning a model from collections of few-shot classification tasks. While more and more novel meta-learning models are being proposed, our research has uncovered simple baselines that have been overlooked. We present a Meta-Baseline method, by pre-training a classifier on all base classes and meta-learning on a nearest-centroid based few-shot classification algorithm, it outperforms recent state-of-the-art methods by a large margin. Why does this simple method work so well? In the meta-learning stage, we observe that a model generalizing better on unseen tasks from base classes can have a decreasing performance on tasks from novel classes, indicating a potential objective discrepancy. We find both pre-training and inheriting a good few-shot classification metric from the pre-trained classifier are important for Meta-Baseline, which potentially helps the model better utilize the pre-trained representations with stronger transferability. Furthermore, we investigate when we need meta-learning in this Meta-Baseline. Our work sets up a new solid benchmark for this field and sheds light on further understanding the phenomenons in the meta-learning framework for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a Meta-Baseline method, by pre-training a classifier on all base classes and meta-learning on a nearest-centroid based few-shot classification algorithm, which outperforms recent state-of-the-art methods by a large margin.'}",
-SimpleShot: Revisiting Nearest-Neighbor Classification for Few-Shot Learning,Yan Wang,"Few-shot learners aim to recognize new object classes based on a small number of labeled training examples. To prevent overfitting, state-of-the-art few-shot learners use meta-learning on convolutional-network features and perform classification using a nearest-neighbor classifier. This paper studies the accuracy of nearest-neighbor baselines without meta-learning. Surprisingly, we find simple feature transformations suffice to obtain competitive few-shot learning accuracies. For example, we find that a nearest-neighbor classifier used in combination with mean-subtraction and L2-normalization outperforms prior results in three out of five settings on the miniImageNet dataset.","{'model': 'tldr@v2.0.0', 'text': 'Surprisingly, simple feature transformations suffice to obtain competitive few-shot learning accuracies and it is found that a nearest-neighbor classifier used in combination with mean-subtraction and L2-normalization outperforms prior results in three out of five settings on the miniImageNet dataset.'}",
-Distributed Few-Shot Learning for Intelligent Recognition of Communication Jamming,Mingqian Liu,"Effective recognition of communication jamming is of vital importance in improving wireless communication system’s anti-jamming capability. Motivated by the major challenges that the jamming data sets in wireless communication system are often small and the recognition performance may be poor, we introduce a novel jamming recognition method based on distributed few-shot learning in this paper. Our proposed method employs a distributed recognition architecture to achieve the global optimization of multiple sub-networks by federated learning. It also introduces a dense block structure in the sub-network structure to improve network information flow by the feature multiplexing and configuration bypass to improve resistance to over-fitting. Our key idea is to first obtain the time-frequency diagram, fractional Fourier transform and constellation diagram of the communication jamming signal as the model-agnostic meta-learning network input, and then train the distributed network through federated learning for jamming recognition. Simulation results show that our proposed method leads to excellent recognition performance with a small data set.","{'model': 'tldr@v2.0.0', 'text': 'A novel jamming recognition method based on distributed few-shot learning that employs a distributed recognition architecture to achieve the global optimization of multiple sub-networks by federated learning and introduces a dense block structure in the sub-network structure to improve network information flow.'}",http://repository.essex.ac.uk/33420/1/distributed-few-shot%20learning-intelligent-recognition-communication-jamming-Chen-2021.pdf
-Prototypical Calibration for Few-shot Learning of Language Models,Zhixiong Han,"In-context learning of GPT-like models has been recognized as fragile across different hand-crafted templates, and demonstration permutations. In this work, we propose prototypical calibration to adaptively learn a more robust decision boundary for zero- and few-shot classification, instead of greedy decoding. Concretely, our method first adopts Gaussian mixture distribution to estimate the prototypical clusters for all categories. Then we assign each cluster to the corresponding label by solving a weighted bipartite matching problem. Given an example, its prediction is calibrated by the likelihood of prototypical clusters. Experimental results show that prototypical calibration yields a substantial improvement on a diverse set of tasks. Extensive analysis across different scales also indicates that our method calibrates the decision boundary as expected, greatly improving the robustness of GPT to templates, permutations, and class imbalance.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that prototypical calibration yields a substantial improvement on a diverse set of tasks, and extensive analysis across different scales indicates that the method calibrates the decision boundary as expected, greatly improving the robustness of GPT to templates, permutations, and class imbalance.'}",http://arxiv.org/pdf/2205.10183
-EASE: Unsupervised Discriminant Subspace Learning for Transductive Few-Shot Learning,Hao Zhu,"Few-shot learning (FSL) has received a lot of attention due to its remarkable ability to adapt to novel classes. Although many techniques have been proposed for FSL, they mostly focus on improving FSL backbones. Some works also focus on learning on top of the features generated by these backbones to adapt them to novel classes. We present an unsupErvised discriminAnt Subspace lEarning (EASE) that improves transductive few-shot learning performance by learning a linear projection onto a subspace built from features of the support set and the unlabeled query set in the test time. Specifically, based on the support set and the unlabeled query set, we generate the similarity matrix and the dissimilarity matrix based on the structure prior for the proposed EASE method, which is efficiently solved with SVD. We also introduce conStraIned wAsserstein MEan Shift clustEring (SIAMESE) which extends Sinkhorn K-means by incorporating labeled support samples. SIAMESE works on the features obtained from EASE to estimate class centers and query predictions. On the miniImageNet, tiered-ImageNet, CIFAR-FS, CUB and OpenMIC benchmarks, both steps significantly boost the performance in transductive FSL and semi-supervised FSL.","{'model': 'tldr@v2.0.0', 'text': 'An unsupErvised discriminAnt Subspace lEarning (EASE) that improves transductive few-shot learning performance by learning a linear projection onto a subspace built from features of the support set and the unlabeled query set in the test time is presented.'}",
-CrossFit: A Few-shot Learning Challenge for Cross-task Generalization in NLP,Qinyuan Ye,"Humans can learn a new language task efficiently with only few examples, by leveraging their knowledge obtained when learning prior tasks. In this paper, we explore whether and how such cross-task generalization ability can be acquired, and further applied to build better few-shot learners across diverse NLP tasks. We introduce CrossFit, a problem setup for studying cross-task generalization ability, which standardizes seen/unseen task partitions, data access during different learning stages, and the evaluation protocols. To instantiate different seen/unseen task partitions in CrossFit and facilitate in-depth analysis, we present the NLP Few-shot Gym, a repository of 160 diverse few-shot NLP tasks created from open-access NLP datasets and converted to a unified text-to-text format. Our analysis reveals that the few-shot learning ability on unseen tasks can be improved via an upstream learning stage using a set of seen tasks. We also observe that the selection of upstream learning tasks can significantly influence few-shot performance on unseen tasks, asking further analysis on task similarity and transferability.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the NLP Few-shot Gym, a repository of 160 diverse few-shot NLP tasks created from open-access NLP datasets and converted to a unified text-to-text format, and reveals that the few- shot learning ability on unseen tasks can be improved via an upstream learning stage using a set of seen tasks.'}",https://aclanthology.org/2021.emnlp-main.572.pdf
-Few-shot Learning with Noisy Labels,Kevin J Liang,"Few-shot learning (FSL) methods typically assume clean support sets with accurately labeled samples when training on novel classes. This assumption can often be unrealistic: support sets, no matter how small, can still include mislabeled samples. Robustness to label noise is therefore essential for FSL methods to be practical, but this problem surprisingly remains largely unexplored. To address mislabeled samples in FSL settings, we make several technical contributions. (1) We offer simple, yet effective, feature aggregation methods, improving the prototypes used by ProtoNet, a popular FSL technique. (2) We describe a novel Transformer model for Noisy Few-Shot Learning (TraNFS). TraNFS leverages a transformer's attention mechanism to weigh mislabeled versus correct samples. (3) Finally, we extensively test these methods on noisy versions of MinilmageNet and TieredImageNet. Our results show that TraNFS is on-par with leading FSL methods on clean support sets, yet outperforms them, by far, in the presence of label noise.","{'model': 'tldr@v2.0.0', 'text': ""This work offers simple, yet effective, feature aggregation methods, improving the prototypes used by ProtoNet, a popular FSL technique, and describes a novel Transformer model for Noisy Few-Shot Learning (TraNFS), which leverages a transformer's attention mechanism to weigh mislabeled versus correct samples.""}",https://arxiv.org/pdf/2204.05494
-How to Prompt? Opportunities and Challenges of Zero- and Few-Shot Learning for Human-AI Interaction in Creative Applications of Generative Models,Hai Dang,"Deep generative models have the potential to fundamentally change the way we create high-fidelity digital content but are often hard to control. Prompting a generative model is a promising recent development that in principle enables end-users to creatively leverage zero-shot and few-shot learning to assign new tasks to an AI ad-hoc, simply by writing them down. However, for the majority of end-users writing effective prompts is currently largely a trial and error process. To address this, we discuss the key opportunities and challenges for interactive creative applications that use prompting as a new paradigm for Human-AI interaction. Based on our analysis, we propose four design goals for user interfaces that support prompting. We illustrate these with concrete UI design sketches, focusing on the use case of creative writing. The research community in HCI and AI can take these as starting points to develop adequate user interfaces for models capable of zero- and few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'The research community in HCI and AI can take these as starting points to develop adequate user interfaces for models capable of zero- and few-shot learning, and propose four design goals for user interfaces that support prompting.'}",http://arxiv.org/pdf/2209.01390
-Learning to Capture the Query Distribution for Few-Shot Learning,Ziqiu Chi,"In the Few-Shot Learning (FSL), much of the related efforts only rely on the few available labeled samples (support set) building approach. However, the challenge is that the support set is easy-to-be-biased, so that they cannot be competent prototypes and are hard to represent the class distribution, leading to performance bottlenecks. In this paper, we propose to solve this obstacle by capturing the distribution of the unlabeled samples (query set). We propose two sampling methods: DeepSearch ( $\cal DS$ ) and WideSearch ( $\cal WS$ ). Both approaches are simple to implement and have no trainable parameters. They search the query samples near to the support set in different manners. Afterward, the statistic information is calculated, and we generate the latent samples according to it. The generated latent set is promising. First, it brings the query set distribution information to the classifier, which significantly improves the performance of the cross-entropy-based classifier. Second, it helps the support set become the better prototypes, which boosts the performance of the prototype-based classifier. Third, we find few latent samples are enough to boost the performance. Abundant experiments prove the proposed method achieves state-of-the-art performance on the few-shot tasks. Finally, rich ablation studies explain the compelling details of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes two sampling methods that bring the query set distribution information to the classifier, which significantly improves the performance of the cross-entropy-based classifier and few latent samples are enough to boost the performance.'}",
-Multi-Level Second-Order Few-Shot Learning,Hongguang Zhang,"We propose a Multi-level Second-order (MlSo) few-shot learning network for supervised or unsupervised few-shot image classification and few-shot action recognition. We leverage so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and we use self-supervised discriminating mechanisms. As Second-order Pooling (SoP) is popular in image recognition, we employ its basic element-wise variant in our pipeline. The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning. As SoP can handle convolutional feature maps of varying spatial sizes, we also introduce image inputs at multiple spatial scales into MlSo. To exploit the discriminative information from multi-level and multi-scale features, we develop a Feature Matching (FM) module that reweights their respective branches. We also introduce a self-supervised step, which is a discriminator of the spatial level and the scale of abstraction. Our pipeline is trained in an end-to-end manner. With a simple architecture, we demonstrate respectable results on standard datasets such as Omniglot, mini–ImageNet, tiered–ImageNet, Open MIC, fine-grained datasets such as CUB Birds, Stanford Dogs and Cars, and action recognition datasets such as HMDB51, UCF101, and mini–MIT.","{'model': 'tldr@v2.0.0', 'text': 'A Multi-level Second-order (MlSo) few-shots learning network for supervised or unsupervised few-shot image classification and few- shot action recognition is proposed, leveraging so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and self-supervised discriminating mechanisms are used.'}",https://arxiv.org/pdf/2201.05916
-Heterogeneous Few-Shot Learning for Hyperspectral Image Classification,Yan Wang,"Deep learning has achieved great success in hyperspectral image (HSI) classification. However, its success relies on the availability of sufficient training samples. Unfortunately, the collection of training samples is expensive, time-consuming, and even impossible in some cases. Natural image datasets that are different from HSI, such as Image Net and mini-ImageNet, have abundant texture and structure information. Effective knowledge transfer between two heterogeneous datasets can significantly improve the accuracy of HSI classification. In this letter, heterogeneous few-shot learning (HFSL) for HSI classification is proposed with only a few labeled samples per class. First, few-shot learning is performed on the mini-ImageNet datasets to learn the transferable knowledge. Then, to make full use of the spatial and spectral information, a spectral–spatial fusion network is devised. Spectral information is obtained by the residual network with pure 1-D operators. Spatial information is extracted by a convolution network with pure 2-D operators, and the weights of the spatial network are initialized by the weights of the model trained on the mini-ImageNet datasets. Finally, few-shot learning is fine-tuned on HSI to extract discriminative spectral–spatial features and individual knowledge, which can improve the classification performance of the new classification task. Experiments conducted on two public HSI datasets demonstrate that the HFSL outperforms the existing few-shot learning methods and supervised learning methods for HSI classification with only a few labeled samples. Our source code is available at https://github.com/Li-ZK/HFSL.","{'model': 'tldr@v2.0.0', 'text': 'Experiments conducted on two public HSI datasets demonstrate that the HFSL outperforms the existing few-shot learning methods and supervised learning methods for HSI classification with only a few labeled samples.'}",
-Few-Shot Learning with Siamese Networks and Label Tuning,Thomas Müller,"We study the problem of building text classifiers with little or no training data, commonly known as zero and few-shot text classification. In recent years, an approach based on neural textual entailment models has been found to give strong results on a diverse range of tasks. In this work, we show that with proper pre-training, Siamese Networks that embed texts and labels offer a competitive alternative. These models allow for a large reduction in inference cost: constant in the number of labels rather than linear. Furthermore, we introduce label tuning, a simple and computationally efficient approach that allows to adapt the models in a few-shot setup by only changing the label embeddings. While giving lower performance than model fine-tuning, this approach has the architectural advantage that a single encoder can be shared by many different tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that with proper pre-training, Siamese Networks that embed texts and labels offer a competitive alternative in text classification, and introduces label tuning, a simple and computationally efficient approach that allows to adapt the models in a few-shot setup by only changing the label embeddings.'}",http://arxiv.org/pdf/2203.14655
-AMCRN: Few-Shot Learning for Automatic Modulation Classification,Quan Zhou,"Deep learning (DL) has been widely applied in automatic modulation classification (AMC), while the superb performance highly depends on high-quality datasets. Motivated by this, the AMC under few-shot conditions is considered in this letter, where a novel network architecture is proposed, namely automatic modulation classification relation network (AMCRN), and verified with the baseline methods. Experimental results state that the accuracy of proposed AMCRN exceeds 90% and 10% to 50% improvements are obtained compared with classical schemes when the signal-to-noise ratio (SNR) is greater than −2 dB.","{'model': 'tldr@v2.0.0', 'text': 'The AMC under few-shot conditions is considered, where a novel network architecture is proposed, namely automatic modulation classification relation network (AMCRN), and verified with the baseline methods.'}",
-Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning,Mamshad Nayeem Rizve,"In many real-world problems, collecting a large number of labeled samples is infeasible. Few-shot learning (FSL) is the dominant approach to address this issue, where the objective is to quickly adapt to novel categories in presence of a limited number of samples. FSL tasks have been predominantly solved by leveraging the ideas from gradient-based meta-learning and metric learning approaches. However, recent works have demonstrated the significance of powerful feature representations with a simple embedding network that can outperform existing sophisticated FSL algorithms. In this work, we build on this insight and propose a novel training mechanism that simultaneously enforces equivariance and invariance to a general set of geometric transformations. Equivariance or invariance has been employed standalone in the previous works; however, to the best of our knowledge, they have not been used jointly. Simultaneous optimization for both of these contrasting objectives allows the model to jointly learn features that are not only independent of the input transformation but also the features that encode the structure of geometric transformations. These complementary sets of features help generalize well to novel classes with only a few data samples. We achieve additional improvements by incorporating a novel self-supervised distillation objective. Our extensive experimentation shows that even without knowledge distillation our proposed method can outperform current state-of-the-art FSL methods on five popular benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel training mechanism that simultaneously enforces equivariance and invariance to a general set of geometric transformations and shows that even without knowledge distillation this proposed method can outperform current state-of-the-art FSL methods on five popular benchmark datasets.'}",https://arxiv.org/pdf/2103.01315
-Few-shot Learning with Multilingual Generative Language Models,Xi Victoria Lin,"Large-scale generative language models such as GPT-3 are competitive few-shot learners. While these models are known to be able to jointly represent many different languages, their training data is dominated by English, potentially limiting their cross-lingual generalization. In this work, we train multilingual generative language models on a corpus covering a diverse set of languages, and study their few- and zero-shot learning capabilities in a wide range of tasks. Our largest model with 7.5 billion parameters sets new state of the art in few-shot learning in more than 20 representative languages, outperforming GPT-3 of comparable size in multilingual commonsense reasoning (with +7.4% absolute accuracy improvement in 0-shot settings and +9.4% in 4-shot settings) and natural language inference (+5.4% in each of 0-shot and 4-shot settings). On the FLORES-101 machine translation benchmark, our model outperforms GPT-3 on 171 out of 182 directions with 32 training examples, while surpassing the official supervised baseline in 45 directions. We conduct an in-depth analysis of different multilingual prompting approaches, showing in particular that strong few-shot learning performance across languages can be achieved via cross-lingual transfer through both templates and demonstration examples.","{'model': 'tldr@v2.0.0', 'text': 'This work trains multilingual generative language models on a corpus covering a diverse set of languages, and shows in particular that strong few-shot learning performance across languages can be achieved via cross-lingual transfer through both templates and demonstration examples.'}",https://aclanthology.org/2022.emnlp-main.616.pdf
-Adaptive Subspaces for Few-Shot Learning,Christian Simon,"Object recognition requires a generalization capability to avoid overfitting, especially when the samples are extremely few. Generalization from limited samples, usually studied under the umbrella of meta-learning, equips learning techniques with the ability to adapt quickly in dynamical environments and proves to be an essential aspect of life long learning. In this paper, we provide a framework for few-shot learning by introducing dynamic classifiers that are constructed from few samples. A subspace method is exploited as the central block of a dynamic classifier. We will empirically show that such modelling leads to robustness against perturbations (e.g., outliers) and yields competitive results on the task of supervised and semi-supervised few-shot classification. We also develop a discriminative form which can boost the accuracy even further. Our code is available at https://github.com/chrysts/dsn_fewshot","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a framework for few-shot learning by introducing dynamic classifiers that are constructed from few samples and empirically shows that such modelling leads to robustness against perturbations and yields competitive results on the task of supervised and semi-supervised few- shot classification.'}",
-Siamese Neural Network Based Few-Shot Learning for Anomaly Detection in Industrial Cyber-Physical Systems,Xiaokang Zhou,"With the increasing population of Industry 4.0, both AI and smart techniques have been applied and become hotly discussed topics in industrial cyber-physical systems (CPS). Intelligent anomaly detection for identifying cyber-physical attacks to guarantee the work efficiency and safety is still a challenging issue, especially when dealing with few labeled data for cyber-physical security protection. In this article, we propose a few-shot learning model with Siamese convolutional neural network (FSL-SCNN), to alleviate the over-fitting issue and enhance the accuracy for intelligent anomaly detection in industrial CPS. A Siamese CNN encoding network is constructed to measure distances of input samples based on their optimized feature representations. A robust cost function design including three specific losses is then proposed to enhance the efficiency of training process. An intelligent anomaly detection algorithm is developed finally. Experiment results based on a fully labeled public dataset and a few labeled dataset demonstrate that our proposed FSL-SCNN can significantly improve false alarm rate (FAR) and F1 scores when detecting intrusion signals for industrial CPS security protection.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning model with Siamese convolutional neural network (FSL-SCNN) is proposed, to alleviate the over-fitting issue and enhance the accuracy for intelligent anomaly detection in industrial CPS.'}",https://ieeexplore.ieee.org/ielx7/9424/9422988/09311786.pdf
-Tuning Language Models as Training Data Generators for Augmentation-Enhanced Few-Shot Learning,Yu Meng,"Recent studies have revealed the intriguing few-shot learning ability of pretrained language models (PLMs): They can quickly adapt to a new task when fine-tuned on a small amount of labeled data formulated as prompts, without requiring abundant task-specific annotations. Despite their promising performance, most existing few-shot approaches that only learn from the small training set still underperform fully supervised training by nontrivial margins. In this work, we study few-shot learning with PLMs from a different perspective: We first tune an autoregressive PLM on the few-shot samples and then use it as a generator to synthesize a large amount of novel training samples which augment the original training set. To encourage the generator to produce label-discriminative samples, we train it via weighted maximum likelihood where the weight of each token is automatically adjusted based on a discriminative meta-learning objective. A classification PLM can then be fine-tuned on both the few-shot and the synthetic samples with regularization for better generalization and stability. Our approach FewGen achieves an overall better result across seven classification tasks of the GLUE benchmark than existing few-shot learning methods, improving no-augmentation methods by 5+ average points, and outperforming augmentation methods by 3+ average points.","{'model': 'tldr@v2.0.0', 'text': 'This work first tunes an autoregressive PLM on the few-shot samples and then uses it as a generator to synthesize a large amount of novel training samples which augment the original training set, achieving an overall better result across seven classification tasks of the GLUE benchmark than existing few- shot learning methods.'}",http://arxiv.org/pdf/2211.03044
-"A neural network solves, explains, and generates university math problems by program synthesis and few-shot learning at human level",Iddo Drori,"Significance We demonstrate that a neural network automatically solves, explains, and generates university-level problems from the largest Massachusetts Institute of Technology (MIT) mathematics courses at a human level. Our methods combine three innovations: 1) using recent neural networks pretrained on text and fine-tuned on code rather than pretrained on text; 2) few-shot learning synthesizing programs that correctly solve course problems automatically; and 3) a pipeline to solve questions, explain solutions, and generate new questions indistinguishable by students from course questions. Our work solves university-level mathematics courses and improves upon state-of-the-art, increasing automatic accuracy on randomly sampled questions on a benchmark by order of magnitude. Implications for higher education include roles of artificial intelligence (AI) in automated course evaluation and content generation.","{'model': 'tldr@v2.0.0', 'text': 'This work solves university-level mathematics courses and improves upon state-of-the-art, increasing automatic accuracy on randomly sampled questions on a benchmark by order of magnitude.'}",
-"Few-Shot Learning via Learning the Representation, Provably",S. Du,"This paper studies few-shot learning via representation learning, where one uses $T$ source tasks with $n_1$ data per task to learn a representation in order to reduce the sample complexity of a target task for which there is only $n_2 (\ll n_1)$ data. Specifically, we focus on the setting where there exists a good \emph{common representation} between source and target, and our goal is to understand how much of a sample size reduction is possible. First, we study the setting where this common representation is low-dimensional and provide a fast rate of $O\left(\frac{\mathcal{C}\left(\Phi\right)}{n_1T} + \frac{k}{n_2}\right)$; here, $\Phi$ is the representation function class, $\mathcal{C}\left(\Phi\right)$ is its complexity measure, and $k$ is the dimension of the representation. When specialized to linear representation functions, this rate becomes $O\left(\frac{dk}{n_1T} + \frac{k}{n_2}\right)$ where $d (\gg k)$ is the ambient input dimension, which is a substantial improvement over the rate without using representation learning, i.e. over the rate of $O\left(\frac{d}{n_2}\right)$. Second, we consider the setting where the common representation may be high-dimensional but is capacity-constrained (say in norm); here, we again demonstrate the advantage of representation learning in both high-dimensional linear regression and neural network learning. Our results demonstrate representation learning can fully utilize all $n_1T$ samples from source tasks.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate representation learning can fully utilize all $n_1T$ samples from source tasks and the advantage of representation learning in both high-dimensional linear regression and neural network learning.'}",
-Research Progress on Few-Shot Learning for Remote Sensing Image Interpretation,Xian Sun,"The rapid development of deep learning brings effective solutions for remote sensing image interpretation. Training deep neural network models usually require a large number of manually labeled samples. However, there is a limitation to obtain sufficient labeled samples in remote sensing field to satisfy the data requirement. Therefore, it is of great significance to conduct the research on few-shot learning for remote sensing image interpretation. First, this article provides a bibliometric analysis of the existing works for remote sensing interpretation related to few-shot learning. Second, two categories of few-shot learning methods, i.e., the data-augmentation-based and the prior-knowledge-based, are introduced for the interpretation of remote sensing images. Then, three typical remote sensing interpretation applications are listed, including scene classification, semantic segmentation, and object detection, together with the corresponding public datasets and the evaluation criteria. Finally, the research status is summarized, and some possible research directions are provided. This article gives a reference for scholars working on few-shot learning research in the remote sensing field.","{'model': 'tldr@v2.0.0', 'text': 'A bibliometric analysis of the existing works for remote sensing interpretation related to few-shot learning and two categories of few- shot learning methods, i.e., the data-augmentation-based and the prior-knowledge-based, are introduced for the interpretation of remote sensing images.'}",https://ieeexplore.ieee.org/ielx7/4609443/9314330/09328476.pdf
-Few-shot learning creates predictive models of drug response that translate from high-throughput screens to individual patients,Jianzhu Ma,,"{'model': 'tldr@v2.0.0', 'text': 'A recently developed technique, few-shot machine learning, is applied to train a versatile neural network model in cell lines that can be tuned to new contexts using few additional samples, and it quickly adapts when switching among different tissue types and in moving from cell-line models to clinical contexts, including patient- derived tumor cells and patient-derived xenografts.'}",
-Overcoming Catastrophic Forgetting in Incremental Few-Shot Learning by Finding Flat Minima,Guangyuan Shi,"This paper considers incremental few-shot learning, which requires a model to continually recognize new categories with only a few examples provided. Our study shows that existing methods severely suffer from catastrophic forgetting, a well-known problem in incremental learning, which is aggravated due to data scarcity and imbalance in the few-shot setting. Our analysis further suggests that to prevent catastrophic forgetting, actions need to be taken in the primitive stage -- the training of base classes instead of later few-shot learning sessions. Therefore, we propose to search for flat local minima of the base training objective function and then fine-tune the model parameters within the flat region on new tasks. In this way, the model can efficiently learn new classes while preserving the old ones. Comprehensive experimental results demonstrate that our approach outperforms all prior state-of-the-art methods and is very close to the approximate upper bound. The source code is available at https://github.com/moukamisama/F2M.","{'model': 'tldr@v2.0.0', 'text': 'This paper considers incremental few-shot learning, which requires a model to continually recognize new categories with only a few examples provided, and proposes to search for flat local minima of the base training objective function and then fine-tune the model parameters within the flat region on new tasks.'}",
-Learning Dynamic Alignment via Meta-filter for Few-shot Learning,C. Xu,"Few-shot learning (FSL), which aims to recognise new classes by adapting the learned knowledge with extremely limited few-shot (support) examples, remains an important open problem in computer vision. Most of the existing methods for feature alignment in few-shot learning only consider image-level or spatial-level alignment while omitting the channel disparity. Our insight is that these methods would lead to poor adaptation with redundant matching, and leveraging channel-wise adjustment is the key to well adapting the learned knowledge to new classes. Therefore, in this paper, we propose to learn a dynamic alignment, which can effectively highlight both query regions and channels according to different local support information. Specifically, this is achieved by first dynamically sampling the neighbourhood of the feature position conditioned on the input few shot, based on which we further predict a both position-dependent and channel-dependent Dynamic Meta-filter. The filter is used to align the query feature with position-specific and channel-specific knowledge. Moreover, we adopt Neural Ordinary Differential Equation (ODE) to enable a more accurate control of the alignment. In such a sense our model is able to better capture fine-grained semantic context of the few-shot example and thus facilitates dynamical knowledge adaptation for few-shot learning. The resulting framework establishes the new state-of-the-arts on major few-shot visual recognition benchmarks, including miniImageNet and tieredImageNet.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to learn a dynamic alignment, which can effectively highlight both query regions and channels according to different local support information and facilitates dynamical knowledge adaptation for few-shot learning.'}",https://arxiv.org/pdf/2103.13582
-Multi-Scale Metric Learning for Few-Shot Learning,Wen Jiang,"Few-shot learning in image classification is developed to learn a model that aims to identify unseen classes with only few training samples for each class. Fewer training samples and new tasks of classification make many traditional classification models no longer applicable. In this paper, a novel few-shot learning method named multi-scale metric learning (MSML) is proposed to extract multi-scale features and learn the multi-scale relations between samples for the classification of few-shot learning. In the proposed method, a feature pyramid structure is introduced for multi-scale feature embedding, which aims to combine high-level strong semantic features with low-level but abundant visual features. Then a multi-scale relation generation network (MRGN) is developed for hierarchical metric learning, in which high-level features are corresponding to deeper metric learning while low-level features are corresponding to lighter metric learning. Moreover, a novel loss function named intra-class and inter-class relation loss (IIRL) is proposed to optimize the proposed deep network, which aims to strengthen the correlation between homogeneous groups of samples and weaken the correlation between heterogeneous groups of samples. Experimental results on mini ImageNet and tiered ImageNet demonstrate that the proposed method achieves superior performance in few-shot learning problem.","{'model': 'tldr@v2.0.0', 'text': 'A novel few-shot learning method named multi-scale metric learning (MSML) is proposed to extract multi- Scale features and learn the multi- scale relations between samples for the classification of few- shot learning.'}",
-Attribute Surrogates Learning and Spectral Tokens Pooling in Transformers for Few-shot Learning,Yang He,"This paper presents new hierarchically cascaded transformers that can improve data efficiency through attribute surrogates learning and spectral tokens pooling. Vision transformers have recently been thought of as a promising alternative to convolutional neural networks for visual recognition. But when there is no sufficient data, it gets stuck in overfitting and shows inferior performance. To improve data efficiency, we propose hierarchically cascaded transformers that exploit intrinsic image structures through spectral tokens pooling and optimize the learnable parameters through latent attribute surrogates. The intrinsic image structure is utilized to reduce the ambiguity between foreground content and background noise by spectral tokens pooling. And the attribute surrogate learning scheme is designed to benefit from the rich visual information in image-label pairs instead of simple visual concepts assigned by their labels. Our Hierarchically Cascaded Transformers, called HCTransformers, is built upon a self-supervised learning framework DINO and is tested on several popular few-shot learning benchmarks. In the inductive setting, HCTransformers surpass the DINO baseline by a large margin of 9.7% 5-way 1-shot accuracy and 9.17% 5-way 5-shot accuracy on miniImageNet, which demonstrates HCTransformers are efficient to extract discriminative features. Also, HCTransformers show clear advantages over SOTA few-shot classification methods in both 5-way 1-shot and 5-way 5-shot settings on four popular benchmark datasets, including miniImageNet, tieredImageNet, FC100, and CIFAR-FS. The trained weights and codes are available at https://github.com/StomachCold/HCTransformers.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents new hierarchically cascaded transformers that can improve data efficiency through attribute surrogates learning and spectral tokens pooling and shows clear advantages over SOTA few-shot classification methods in both 5- way 1-shot and 5-way 5-shot settings on four popular benchmark datasets.'}",https://arxiv.org/pdf/2203.09064
-Semi-supervised few-shot learning approach for plant diseases recognition,Yang Li,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed semi-supervised few-shot learning methods can outperform other related works with fewer labeled training data and adopt the confidence interval to determine the number of unlabeled samples for pseudo-labelling adaptively.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-021-00770-1
-Partial Is Better Than All: Revisiting Fine-tuning Strategy for Few-shot Learning,Zhiqiang Shen,"The goal of few-shot learning is to learn a classifier that can recognize unseen classes from limited support data with labels. A common practice for this task is to train a model on the base set first and then transfer to novel classes through fine-tuning or meta-learning. However, as the base classes have no overlap to the novel set, simply transferring whole knowledge from base data is not an optimal solution since some knowledge in the base model may be biased or even harmful to the novel class. In this paper, we propose to transfer partial knowledge by freezing or fine-tuning particular layer(s) in the base model. Specifically, layers will be imposed different learning rates if they are chosen to be fine-tuned, to control the extent of preserved transferability. To determine which layers to be recast and what values of learning rates for them, we introduce an evolutionary search based method that is efficient to simultaneously locate the target layers and determine their individual learning rates. We conduct extensive experiments on CUB and mini-ImageNet to demonstrate the effectiveness of our proposed method. It achieves the state-of-the-art performance on both meta-learning and non-meta based frameworks. Furthermore, we extend our method to the conventional pre-training + fine-tuning paradigm and obtain consistent improvement.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to transfer partial knowledge by freezing or fine-tuning particular layer(s) in the base model to control the extent of preserved transferability, and introduces an evolutionary search based method that is efficient to simultaneously locate the target layers and determine their individual learning rates.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17155/16962
-Attentional Constellation Nets for Few-Shot Learning,Weijian Xu,"The success of deep convolutional neural networks builds on top of the learning of effective convolution operations, capturing a hierarchy of structured features via filtering, activation, and pooling. However, the explicit structured features, e.g. object parts, are not expressive in the existing CNN frameworks. In this paper, we tackle the few-shot learning problem and make an effort to enhance structured features by expanding CNNs with a constellation model, which performs cell feature clustering and encoding with a dense part representation; the relationships among the cell features are further modeled by an attention mechanism. With the additional constellation branch to increase the awareness of object parts, our method is able to attain the advantages of the CNNs while making the overall internal representations more robust in the few-shot learning setting. Our approach attains a significant improvement over the existing methods in few-shot learning on the CIFAR-FS, FC100, and mini-ImageNet benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper tackles the few-shot learning problem and makes an effort to enhance structured features by expanding CNNs with a constellation model, which performs cell feature clustering and encoding with a dense part representation; the relationships among the cell features are further modeled by an attention mechanism.'}",
-Few-Shot Learning on Graphs: A Survey,Chuxu Zhang,"Graph representation learning has attracted tremendous attention due to its remarkable performance in many real-world applications. However, prevailing supervised graph representation learning models for specific tasks often suffer from label sparsity issue as data labeling is always time and resource consuming. In light of this, few-shot learning on graphs (FSLG), which combines the strengths of graph representation learning and few-shot learning together, has been proposed to tackle the performance degradation in face of limited annotated data challenge. There have been many studies working on FSLG recently. In this paper, we comprehensively survey these work in the form of a series of methods and applications. Specifically, we first introduce FSLG challenges and bases, then categorize and summarize existing work of FSLG in terms of three major graph mining tasks at different granularity levels, i.e., node, edge, and graph. Finally, we share our thoughts on some future research directions of FSLG. The authors of this survey have contributed significantly to the AI literature on FSLG over the last few years.","{'model': 'tldr@v2.0.0', 'text': 'This paper categorize and summarize existing work of FSLG in terms of three major graph mining tasks at different granularity levels, i.e., node, edge, and graph.'}",https://arxiv.org/pdf/2203.09308
-Incremental few-shot learning via vector quantization in deep embedded space,Kuilin Chen,"The capability of incrementally learning new tasks without forgetting old ones is a challenging problem due to catastrophic forgetting. This challenge becomes greater when novel tasks contain very few labelled training samples. Currently, most methods are dedicated to class-incremental learning and rely on sufficient training data to learn additional weights for newly added classes. Those methods cannot be easily extended to incremental regression tasks and could suffer from severe overfitting when learning few-shot novel tasks. In this study, we propose a nonparametric method in deep embedded space to tackle incremental few-shot learning problems. The knowledge about the learned tasks is compressed into a small number of quantized reference vectors. The proposed method learns new tasks sequentially by adding more reference vectors to the model using few-shot samples in each novel task. For classification problems, we employ the nearest neighbor scheme to make classification on sparsely available data and incorporate intra-class variation, less forgetting regularization and calibration of reference vectors to mitigate catastrophic forgetting. In addition, the proposed learning vector quantization (LVQ) in deep embedded space can be customized as a kernel smoother to handle incremental few-shot regression tasks. Experimental results demonstrate that the proposed method outperforms other state-of-the-art methods in incremental learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed learning vector quantization in deep embedded space can be customized as a kernel smoother to handle incremental few-shot regression tasks and outperforms other state-of-the-art methods in incremental learning.'}",
-Deep Few-Shot Learning for Hyperspectral Image Classification,Bing Liu,"Deep learning methods have recently been successfully explored for hyperspectral image (HSI) classification. However, training a deep-learning classifier notoriously requires hundreds or thousands of labeled samples. In this paper, a deep few-shot learning method is proposed to address the small sample size problem of HSI classification. There are three novel strategies in the proposed algorithm. First, spectral–spatial features are extracted to reduce the labeling uncertainty via a deep residual 3-D convolutional neural network. Second, the network is trained by episodes to learn a metric space where samples from the same class are close and those from different classes are far. Finally, the testing samples are classified by a nearest neighbor classifier in the learned metric space. The key idea is that the designed network learns a metric space from the training data set. Furthermore, such metric space could generalize to the classes of the testing data set. Note that the classes of the testing data set are not seen in the training data set. Four widely used HSI data sets were used to assess the performance of the proposed algorithm. The experimental results indicate that the proposed method can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.","{'model': 'tldr@v2.0.0', 'text': 'A deep few-shot learning method is proposed to address the small sample size problem of HSI classification and can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.'}",
-BOIL: Towards Representation Change for Few-shot Learning,Jaehoon Oh,,,
-Adaptive Cross-Modal Few-Shot Learning,Chen Xing,"Metric-based meta-learning techniques have successfully been applied to few-shot classification problems. In this paper, we propose to leverage cross-modal information to enhance metric-based few-shot learning methods. Visual and semantic feature spaces have different structures by definition. For certain concepts, visual features might be richer and more discriminative than text ones. While for others, the inverse might be true. Moreover, when the support from visual information is limited in image classification, semantic representations (learned from unsupervised text corpora) can provide strong prior knowledge and context to help learning. Based on these two intuitions, we propose a mechanism that can adaptively combine information from both modalities according to new image categories to be learned. Through a series of experiments, we show that by this adaptive combination of the two modalities, our model outperforms current uni-modality few-shot learning methods and modality-alignment methods by a large margin on all benchmarks and few-shot scenarios tested. Experiments also show that our model can effectively adjust its focus on the two modalities. The improvement in performance is particularly large when the number of shots is very small.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a mechanism that can adaptively combine information from both modalities according to new image categories to be learned and shows that by this adaptive combination of the two modalities, this model outperforms current uni-modality few-shot learning methods and modality-alignment methods by a large margin on all benchmarks and few- shot scenarios tested.'}",
-Meta-Baseline: Exploring Simple Meta-Learning for Few-Shot Learning,Yinbo Chen,"Meta-learning has been the most common framework for few-shot learning in recent years. It learns the model from collections of few-shot classification tasks, which is believed to have a key advantage of making the training objective consistent with the testing objective. However, some recent works report that by training for whole-classification, i.e. classification on the whole label-set, it can get comparable or even better embedding than many meta-learning algorithms. The edge between these two lines of works has yet been underexplored, and the effectiveness of meta-learning in few-shot learning remains unclear. In this paper, we explore a simple process: meta-learning over a whole-classification pre-trained model on its evaluation metric. We observe this simple method achieves competitive performance to state-of-the-art methods on standard bench-marks. Our further analysis shed some light on understanding the trade-offs between the meta-learning objective and the whole-classification objective in few-shot learning. Our code is available at https://github.com/yinboc/few-shot-meta-baseline.","{'model': 'tldr@v2.0.0', 'text': 'A simple process: meta-learning over a whole-classification pre-trained model on its evaluation metric achieves competitive performance to state-of-the-art methods on standard bench-marks and sheds some light on understanding the trade-offs between the meta- learning objective and the whole- classification objective in few-shot learning.'}",https://arxiv.org/pdf/2003.04390
-DPGN: Distribution Propagation Graph Network for Few-Shot Learning,Ling Yang,"Most graph-network-based meta-learning approaches model instance-level relation of examples. We extend this idea further to explicitly model the distribution-level relation of one example to all other examples in a 1-vs-N manner. We propose a novel approach named distribution propagation graph network (DPGN) for few-shot learning. It conveys both the distribution-level relations and instance-level relations in each few-shot learning task. To combine the distribution-level relations and instance-level relations for all examples, we construct a dual complete graph network which consists of a point graph and a distribution graph with each node standing for an example. Equipped with dual graph architecture, DPGN propagates label information from labeled examples to unlabeled examples within several update generations. In extensive experiments on few-shot learning benchmarks, DPGN outperforms state-of-the-art results by a large margin in 5%∼12% under supervised setting and 7%∼13% under semi-supervised setting. Code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach named distribution propagation graph network (DPGN) for few-shot learning, which conveys both the distribution-level relations and instance- level relations in each few- shot learning task.'}",https://arxiv.org/pdf/2003.14247
-FewCLUE: A Chinese Few-shot Learning Evaluation Benchmark,Liang Xu,"Pretrained Language Models (PLMs) have achieved tremendous success in natural language understanding tasks. While different learning schemes -- fine-tuning, zero-shot, and few-shot learning -- have been widely explored and compared for languages such as English, there is comparatively little work in Chinese to fairly and comprehensively evaluate and compare these methods and thus hinders cumulative progress. In this paper, we introduce the Chinese Few-shot Learning Evaluation Benchmark (FewCLUE), the first comprehensive few-shot evaluation benchmark in Chinese. It includes nine tasks, ranging from single-sentence and sentence-pair classification tasks to machine reading comprehension tasks. We systematically evaluate five state-of-the-art (SOTA) few-shot learning methods (including PET, ADAPET, LM-BFF, P-tuning and EFL), and compare their performance with fine-tuning and zero-shot learning schemes on the newly constructed FewCLUE benchmark. Experimental results reveal that: 1) The effect of different few-shot learning methods is sensitive to the pre-trained model to which the methods are applied; 2) PET and P-tuning achieve the best overall performance with RoBERTa and ERNIE respectively. Our benchmark is used in the few-shot learning contest of NLPCC 2021. In addition, we provide a user-friendly toolkit, as well as an online leaderboard to help facilitate further progress on Chinese few-shot learning. We provide a baseline performance on different learning methods, a reference for future research.","{'model': 'tldr@v2.0.0', 'text': 'The Chinese Few-shot Learning Evaluation Benchmark (FewCLUE) is introduced, the first comprehensive few-shot evaluation benchmark in Chinese, and shows that PET and P-tuning achieve the best overall performance with RoBERTa and ERNIE respectively.'}",
-Bridging the Gap Between Few-Shot and Many-Shot Learning via Distribution Calibration,Shuo Yang,"A major gap between few-shot and many-shot learning is the data distribution empirically oserved by the model during training. In few-shot learning, the learned model can easily become over-fitted based on the biased distribution formed by only a few training examples, while the ground-truth data distribution is more accurately uncovered in many-shot learning to learn a well-generalized model. In this paper, we propose to calibrate the distribution of these few-sample classes to be more unbiased to alleviate such an over-fitting problem. The distribution calibration is achieved by transferring statistics from the classes with sufficient examples to those few-sample classes. After calibration, an adequate number of examples can be sampled from the calibrated distribution to expand the inputs to the classifier. Specifically, we assume every dimension in the feature representation from the same class follows a Gaussian distribution so that the mean and the variance of the distribution can borrow from that of similar classes whose statistics are better estimated with an adequate number of samples. Extensive experiments on three datasets, miniImageNet, tieredImageNet, and CUB, show that a simple linear classifier trained using the features sampled from our calibrated distribution can outperform the state-of-the-art accuracy by a large margin. Besides the favorable performance, the proposed method also exhibits high flexibility by showing consistent accuracy improvement when it is built on top of any off-the-shelf pretrained feature extractors and classification models without extra learnable parameters. The visualization of these generated features demonstrates that our calibrated distribution is an accurate estimation thus the generalization ability gain is convincing. We also establish a generalization error bound for the proposed distribution-calibration-based few-shot learning, which consists of the distribution assumption error, the distribution approximation error, and the estimation error. This generalization error bound theoretically justifies the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'A generalization error bound is established for the proposed distribution-calibration-based few-shot learning, which consists of the distribution assumption error, the distribution approximation error, and the estimation error that theoretically justifies the effectiveness of the proposed method.'}",https://www.techrxiv.org/articles/preprint/Bridging_the_Gap_between_Few-Shot_and_Many-Shot_Learning_via_Distribution_Calibration/14380697/files/27474479.pdf
-Rectifying the Shortcut Learning of Background for Few-Shot Learning,Xu Luo,"The category gap between training and evaluation has been characterised as one of the main obstacles to the success of Few-Shot Learning (FSL). In this paper, we for the first time empirically identify image background, common in realistic images, as a shortcut knowledge helpful for in-class classification but ungeneralizable beyond training categories in FSL. A novel framework, COSOC, is designed to tackle this problem by extracting foreground objects in images at both training and evaluation without any extra supervision. Extensive experiments carried on inductive FSL tasks demonstrate the effectiveness of our approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper for the first time empirically identify image background, common in realistic images, as a shortcut knowledge helpful for in-class classification but ungeneralizable beyond training categories in FSL.'}",
-Few-shot Learning with Multilingual Language Models,Xi Victoria Lin,"Large-scale autoregressive language models such as GPT-3 are few-shot learners that can perform a wide range of language tasks without fine-tuning. While these models are known to be able to jointly represent many different languages, their training data is dom-inated by English, potentially limiting their cross-lingual generalization. In this work, we train multilingual autoregressive language models on a balanced corpus covering a diverse set of languages, and study their few- and zero-shot learning capabilities in a wide range of tasks. Our largest model with 7.5 billion parameters sets new state of the art in few-shot learning in more than 20 representative languages, outperforming GPT-3 of comparable size in multilingual commonsense reasoning (with +7.4% absolute accuracy improvement in 0-shot settings and +9.4% in 4-shot settings) and natural language inference (+5.4% in each of 0-shot and 4-shot settings). On the FLORES-101 machine translation benchmark, our model outperforms GPT-3 on 171 out of 182 translation directions with 32 training examples, while surpassing the official supervised baseline in 45 directions. We present a detailed analysis of where the model suc-ceeds and fails, showing in particular that it enables cross-lingual in-context learning on some tasks, while there is still room for improvement on surface form robustness and adaptation to tasks that do not have a natural cloze form. Finally, we evaluate our models in social value tasks such as hate speech detection in 5 languages and find it has limitations similar to comparably sized GPT-3 models.","{'model': 'tldr@v2.0.0', 'text': 'A detailed analysis of where the model succeeds and fails is presented, showing in particular that it enables cross-lingual in-context learning on some tasks, while there is still room for improvement on surface form robustness and adaptation to tasks that do not have a natural cloze form.'}",
-Cross-domain Few-shot Learning with Task-specific Adapters,Wei-Hong Li,"In this paper, we look at the problem of cross-domain few-shot classification that aims to learn a classifier from previously unseen classes and domains withfew labeled samples. Recent approaches broadly solve this problem by pa-rameterizing their few-shot classifiers with task-agnostic and task-specific weights where the former is typically learned on a large training set and the latter is dynamically predicted through an auxiliary network conditioned on a small support set. In this work, we focus on the estimation of the latter, and propose to learn task-specific weights from scratch directly on a small support set, in contrast to dynamically estimating them. In particular, through systematic analysis, we show that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the per-formance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to learn task- specific weights from scratch directly on a small support set, in contrast to dynamically estimating them, and shows that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the per-formance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost.'}",https://arxiv.org/pdf/2107.00358
-Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learning,Sungyong Baik,"In few-shot learning scenarios, the challenge is to generalize and perform well on new unseen examples when only very few labeled examples are available for each task. Model-agnostic meta-learning (MAML) has gained the popularity as one of the representative few-shot learning methods for its flexibility and applicability to diverse problems. However, MAML and its variants often resort to a simple loss function without any auxiliary loss function or regularization terms that can help achieve better generalization. The problem lies in that each application and task may require different auxiliary loss function, especially when tasks are diverse and distinct. Instead of attempting to hand-design an auxiliary loss function for each application and task, we introduce a new meta-learning framework with a loss function that adapts to each task. Our proposed framework, named Meta-Learning with Task-Adaptive Loss Function (MeTAL), demonstrates the effectiveness and the flexibility across various domains, such as few-shot classification and few-shot regression.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a new meta-learning framework with a loss function that adapts to each task, named Meta-Learning with Task-Adaptive Loss Function (MeTAL), which demonstrates the effectiveness and the flexibility across various domains, such as few-shot classification and few- shot regression.'}",https://arxiv.org/pdf/2110.03909
-MELR: Meta-Learning via Modeling Episode-Level Relationships for Few-Shot Learning,Nanyi Fei,,,
-Finding Task-Relevant Features for Few-Shot Learning by Category Traversal,Hongyang Li,"Few-shot learning is an important area of research. Conceptually, humans are readily able to understand new concepts given just a few examples, while in more pragmatic terms, limited-example training situations are common practice. Recent effective approaches to few-shot learning employ a metric-learning framework to learn a feature similarity comparison between a query (test) example, and the few support (training) examples. However, these approaches treat each support class independently from one another, never looking at the entire task as a whole. Because of this, they are constrained to use a single set of features for all possible test-time tasks, which hinders the ability to distinguish the most relevant dimensions for the task at hand. In this work, we introduce a Category Traversal Module that can be inserted as a plug-and-play module into most metric-learning based few-shot learners. This component traverses across the entire support set at once, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space. Incorporating our module improves performance considerably (5%-10% relative) over baseline systems on both miniImageNet and tieredImageNet benchmarks, with overall performance competitive with the most recent state-of-the-art systems.","{'model': 'tldr@v2.0.0', 'text': 'A Category Traversal Module is introduced that can be inserted as a plug-and-play module into most metric-learning based few-shot learners, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space.'}",https://arxiv.org/pdf/1905.11116
-Adversarial Feature Hallucination Networks for Few-Shot Learning,K. Li,"The recent flourish of deep learning in various tasks is largely accredited to the rich and accessible labeled data. Nonetheless, massive supervision remains a luxury for many real applications, boosting great interest in label-scarce techniques such as few-shot learning (FSL), which aims to learn concept of new classes with a few labeled samples. A natural approach to FSL is data augmentation and many recent works have proved the feasibility by proposing various data synthesis models. However, these models fail to well secure the discriminability and diversity of the synthesized data and thus often produce undesirable results. In this paper, we propose Adversarial Feature Hallucination Networks (AFHN) which is based on conditional Wasserstein Generative Adversarial networks (cWGAN) and hallucinates diverse and discriminative features conditioned on the few labeled samples. Two novel regularizers, i.e., the classification regularizer and the anti-collapse regularizer, are incorporated into AFHN to encourage discriminability and diversity of the synthesized features, respectively. Ablation study verifies the effectiveness of the proposed cWGAN based feature hallucination framework and the proposed regularizers. Comparative results on three common benchmark datasets substantiate the superiority of AFHN to existing data augmentation based FSL approaches and other state-of-the-art ones.","{'model': 'tldr@v2.0.0', 'text': 'Adversarial Feature Hallucination Networks (AFHN) is proposed which is based on conditional Wasserstein Generative Adversarial networks (cWGAN) and hallucinates diverse and discriminative features conditioned on the few labeled samples and comparative results substantiate the superiority of AFHN to existing data augmentation based FSL approaches and other state-of-the-art ones.'}",https://arxiv.org/pdf/2003.13193
-Instance Credibility Inference for Few-Shot Learning,Yikai Wang,"Few-shot learning (FSL) aims to recognize new objects with extremely limited training data for each category. Previous efforts are made by either leveraging meta-learning paradigm or novel principles in data augmentation to alleviate this extremely data-scarce problem. In contrast, this paper presents a simple statistical approach, dubbed Instance Credibility Inference (ICI) to exploit the distribution support of unlabeled instances for few-shot learning. Specifically, we first train a linear classifier with the labeled few-shot examples and use it to infer the pseudo-labels for the unlabeled data. To measure the credibility of each pseudo-labeled instance, we then propose to solve another linear regression hypothesis by increasing the sparsity of the incidental parameters and rank the pseudo-labeled instances with their sparsity degree. We select the most trustworthy pseudo-labeled instances alongside the labeled examples to re-train the linear classifier. This process is iterated until all the unlabeled samples are included in the expanded training set, i.e. the pseudo-label is converged for unlabeled data pool. Extensive experiments under two few-shot settings show that our simple approach can establish new state-of-the-arts on four widely used few-shot learning benchmark datasets including miniImageNet, tieredImageNet, CIFAR-FS, and CUB. Our code is available at: https://github.com/Yikai-Wang/ICI-FSL","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple statistical approach, dubbed Instance Credibility Inference (ICI), to exploit the distribution support of unlabeled instances for few-shot learning to establish new state-of-the-arts on four widely used few- shot learning benchmark datasets.'}",https://arxiv.org/pdf/2003.11853
-FS-Mol: A Few-Shot Learning Dataset of Molecules,Megan Stanley,"Small datasets are ubiquitous in drug discovery as data generation is expensive and can be restricted for ethical reasons (e.g. in vivo experiments). A widely applied technique in early drug discovery to identify novel active molecules against a protein target is modeling quantitative structure-activity relationships (QSAR). It is known to be extremely challenging, as available measurements of compound activities range in the low dozens or hundreds. However, many such related datasets exist, each with a small number of datapoints, opening up the opportunity for few-shot learning after pretraining on a substantially larger corpus of data. At the same time, many few-shot learning methods are currently evaluated in the computer-vision domain. We propose that expansion into a new application, as well as the possibility to use explicitly graph-structured data, will drive exciting progress in few-shot learning. Here, we provide a few-shot learning dataset (FS-Mol) and complementary benchmarking procedure. We define a set of tasks on which few-shot learning methods can be evaluated, with a separate set of tasks for use in pretraining. In addition, we implement and evaluate a number of existing single-task, multi-task, and meta-learning approaches as baselines for the community. We hope that our dataset, support code release, and baselines will encourage future work on this extremely challenging new domain for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a few-shot learning dataset (FS-Mol) and complementary benchmarking procedure, and implements and evaluates a number of existing single-task, multi- task, and meta-learning approaches as baselines for the community.'}",
-Yuan 1.0: Large-Scale Pre-trained Language Model in Zero-Shot and Few-Shot Learning,Shaohua Wu,"Recent work like GPT-3 has demonstrated excellent performance of Zero-Shot and Few-Shot learning on many natural language processing (NLP) tasks by scaling up model size, dataset size and the amount of computation. However, training a model like GPT-3 requires huge amount of computational resources which makes it challengeable to researchers. In this work, we propose a method that incorporates large-scale distributed training performance into model architecture design. With this method, Yuan 1.0, the current largest singleton language model with 245B parameters, achieves excellent performance on thousands GPUs during training, and the state-of-the-art results on NLP tasks. A data processing method is designed to efficiently filter massive amount of raw data. The current largest high-quality Chinese corpus with 5TB high quality texts is built based on this method. In addition, a calibration and label expansion method is proposed to improve the Zero-Shot and Few-Shot performance, and steady improvement is observed on the accuracy of various tasks. Yuan 1.0 presents strong capacity of natural language generation, and the generated articles are difficult to distinguish from the human-written ones.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method that incorporates large-scale distributed training performance into model architecture design and achieves excellent performance on thousands GPUs during training, and the state-of-the-art results on NLP tasks.'}",
-Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Noise-enhanced Supervised Autoencoder,Hanwen Liang,"State of the art (SOTA) few-shot learning (FSL) methods suffer significant performance drop in the presence of domain differences between source and target datasets. The strong discrimination ability on the source dataset does not necessarily translate to high classification accuracy on the target dataset. In this work, we address this cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model. Specifically, we teach the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE). NSAE trains the model by jointly reconstructing inputs and predicting the labels of inputs as well as their reconstructed pairs. Theoretical analysis based on intra-class correlation (ICC) shows that the feature embeddings learned from NSAE have stronger discrimination and generalization abilities in the target domain. We also take advantage of NSAE structure and propose a two-step fine-tuning procedure that achieves better adaption and improves classification performance in the target domain. Extensive experiments and ablation studies are conducted to demonstrate the effectiveness of the proposed method. Experimental results show that our proposed method consistently outperforms SOTA methods under various conditions.","{'model': 'tldr@v2.0.0', 'text': 'This work addresses the cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model by teaching the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE).'}",https://arxiv.org/pdf/2108.05028
-Interventional Few-Shot Learning,Zhongqi Yue,"We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL) methods: the pre-trained knowledge is indeed a confounder that limits the performance. This finding is rooted from our causal assumption: a Structural Causal Model (SCM) for the causalities among the pre-trained knowledge, sample features, and labels. Thanks to it, we propose a novel FSL paradigm: Interventional Few-Shot Learning (IFSL). Specifically, we develop three effective IFSL algorithmic implementations based on the backdoor adjustment, which is essentially a causal intervention towards the SCM of many-shot learning: the upper-bound of FSL in a causal view. It is worth noting that the contribution of IFSL is orthogonal to existing fine-tuning and meta-learning based FSL methods, hence IFSL can improve all of them, achieving a new 1-/5-shot state-of-the-art on \textit{mini}ImageNet, \textit{tiered}ImageNet, and cross-domain CUB. Code is released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that the contribution of IFSL is orthogonal to existing fine-tuning and meta-learning based FSL methods, hence IFSL can improve all of them, achieving a new 1-/5-shot state-of-the-art on \\textit{mini}ImageNet, \\text it{tiered}Image net, and cross-domain CUB.'}",
-True Few-Shot Learning with Prompts—A Real-World Perspective,Timo Schick,"Abstract Prompt-based approaches excel at few-shot learning. However, Perez et al. (2021) recently cast doubt on their performance as they had difficulty getting good results in a “true” few-shot setting in which prompts and hyperparameters cannot be tuned on a dev set. In view of this, we conduct an extensive study of Pet, a method that combines textual instructions with example-based finetuning. We show that, if correctly configured, Pet performs strongly in true few-shot settings without a dev set. Crucial for this strong performance is a number of design choices, including Pet’s ability to intelligently handle multiple prompts. We put our findings to a real-world test by running Pet on RAFT, a benchmark of tasks taken from realistic NLP applications for which no labeled dev or test sets are available. Pet achieves a new state of the art on RAFT and performs close to non-expert humans for 7 out of 11 tasks. These results demonstrate that prompt-based learners can successfully be applied in true few-shot settings and underpin our belief that learning from instructions will play an important role on the path towards human-like few-shot learning capabilities.","{'model': 'tldr@v2.0.0', 'text': 'An extensive study of Pet, a method that combines textual instructions with example-based finetuning, shows that, if correctly configured, Pet performs strongly in true few-shot settings without a dev set and underpin the belief that learning from instructions will play an important role on the path towards human-like few- shot learning capabilities.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00485/2030692/tacl_a_00485.pdf
-Pareto Self-Supervised Training for Few-Shot Learning,Zhengyu Chen,"While few-shot learning (FSL) aims for rapid generalization to new concepts with little supervision, self-supervised learning (SSL) constructs supervisory signals directly computed from unlabeled data. Exploiting the complementarity of these two manners, few-shot auxiliary learning has recently drawn much attention to deal with few labeled data. Previous works benefit from sharing inductive bias between the main task (FSL) and auxiliary tasks (SSL), where the shared parameters of tasks are optimized by minimizing a linear combination of task losses. However, it is challenging to select a proper weight to balance tasks and reduce task conflict. To handle the problem as a whole, we propose a novel approach named as Pareto self-supervised training (PSST) for FSL. PSST explicitly decomposes the few-shot auxiliary problem into multiple constrained multi-objective subproblems with different trade-off preferences, and here a preference region in which the main task achieves the best performance is identified. Then, an effective preferred Pareto exploration is proposed to find a set of optimal solutions in such a preference region. Extensive experiments on several public benchmark datasets validate the effectiveness of our approach by achieving state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach named as Pareto self-supervised training (PSST) for FSL is proposed, which decomposes the few-shot auxiliary problem into multiple constrained multi-objective subproblems with different trade-off preferences, and here a preference region in which the main task achieves the best performance is identified.'}",https://arxiv.org/pdf/2104.07841
-Leveraging the Feature Distribution in Transfer-based Few-Shot Learning,Yuqing Hu,,"{'model': 'tldr@v2.0.0', 'text': 'A transfer-based novel method that builds on two steps: preprocessing the feature vectors so that they become closer to Gaussian-like distributions, and leveraging this preprocessing using an optimal-transport inspired algorithm.'}",https://arxiv.org/pdf/2006.03806
-FlipDA: Effective and Robust Data Augmentation for Few-Shot Learning,Jing Zhou,"Most previous methods for text data augmentation are limited to simple tasks and weak baselines. We explore data augmentation on hard tasks (i.e., few-shot natural language understanding) and strong baselines (i.e., pretrained models with over one billion parameters). Under this setting, we reproduced a large number of previous augmentation methods and found that these methods bring marginal gains at best and sometimes degrade the performance much. To address this challenge, we propose a novel data augmentation method FlipDA that jointly uses a generative model and a classifier to generate label-flipped data. Central to the idea of FlipDA is the discovery that generating label-flipped data is more crucial to the performance than generating label-preserved data. Experiments show that FlipDA achieves a good tradeoff between effectiveness and robustness—it substantially improves many tasks while not negatively affecting the others.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel data augmentation method FlipDA that jointly uses a generative model and a classifier to generate label-flipped data and achieves a good tradeoff between effectiveness and robustness.'}",https://aclanthology.org/2022.acl-long.592.pdf
-IEPT: Instance-Level and Episode-Level Pretext Tasks for Few-Shot Learning,Manli Zhang,,,
-Laplacian Regularized Few-Shot Learning,Imtiaz Masud Ziko,"We propose a transductive Laplacian-regularized inference for few-shot tasks. Given any feature embedding learned from the base classes, we minimize a quadratic binary-assignment function containing two terms: (1) a unary term assigning query samples to the nearest class prototype, and (2) a pairwise Laplacian term encouraging nearby query samples to have consistent label assignments. Our transductive inference does not re-train the base model, and can be viewed as a graph clustering of the query set, subject to supervision constraints from the support set. We derive a computationally efficient bound optimizer of a relaxation of our function, which computes independent (parallel) updates for each query sample, while guaranteeing convergence. Following a simple cross-entropy training on the base classes, and without complex meta-learning strategies, we conducted comprehensive experiments over five few-shot learning benchmarks. Our LaplacianShot consistently outperforms state-of-the-art methods by significant margins across different models, settings, and data sets. Furthermore, our transductive inference is very fast, with computational times that are close to inductive inference, and can be used for large-scale few-shot tasks.","{'model': 'tldr@v2.0.0', 'text': ""The authors' transductive LaplacianShot consistently outperforms state-of-the-art methods by significant margins across different models, settings, and data sets, and can be used for large-scale few-shot tasks.""}",
-"Learn Continually, Generalize Rapidly: Lifelong Knowledge Accumulation for Few-shot Learning",Xisen Jin,"The ability to continuously expand knowledge over time and utilize it to rapidly generalize to new tasks is a key feature of human linguistic intelligence. Existing models that pursue rapid generalization to new tasks (e.g., few-shot learning methods), however, are mostly trained in a single shot on fixed datasets, unable to dynamically expand their knowledge; while continual learning algorithms are not specifically designed for rapid generalization. We present a new learning setup, Continual Learning of Few-Shot Learners (CLIF), to address the challenges of both learning settings in a unified setup. CLIF assumes a model learns from a sequence of diverse NLP tasks arriving sequentially, accumulating knowledge for improved generalization to new tasks, while also retaining performance on the tasks learned earlier. We examine how the generalization ability is affected in the continual learning setup, evaluate a number of continual learning algorithms, and propose a novel regularized adapter generation approach. We find that catastrophic forgetting affects generalization ability to a less degree than performance on seen tasks; while continual learning algorithms can still bring considerable benefit to the generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'It is found that catastrophic forgetting affects generalization ability to a less degree than performance on seen tasks; while continual learning algorithms can still bring considerable benefit to thegeneralization ability.'}",https://aclanthology.org/2021.findings-emnlp.62.pdf
-"Interactive Few-Shot Learning: Limited Supervision, Better Medical Image Segmentation",Ruiwei Feng,"Many known supervised deep learning methods for medical image segmentation suffer an expensive burden of data annotation for model training. Recently, few-shot segmentation methods were proposed to alleviate this burden, but such methods often showed poor adaptability to the target tasks. By prudently introducing interactive learning into the few-shot learning strategy, we develop a novel few-shot segmentation approach called Interactive Few-shot Learning (IFSL), which not only addresses the annotation burden of medical image segmentation models but also tackles the common issues of the known few-shot segmentation methods. First, we design a new few-shot segmentation structure, called Medical Prior-based Few-shot Learning Network (MPrNet), which uses only a few annotated samples (e.g., 10 samples) as support images to guide the segmentation of query images without any pre-training. Then, we propose an Interactive Learning-based Test Time Optimization Algorithm (IL-TTOA) to strengthen our MPrNet on the fly for the target task in an interactive fashion. To our best knowledge, our IFSL approach is the first to allow few-shot segmentation models to be optimized and strengthened on the target tasks in an interactive and controllable manner. Experiments on four few-shot segmentation tasks show that our IFSL approach outperforms the state-of-the-art methods by more than 20% in the DSC metric. Specifically, the interactive optimization algorithm (IL-TTOA) further contributes ~10% DSC improvement for the few-shot segmentation models.","{'model': 'tldr@v2.0.0', 'text': 'The IFSL approach is the first to allow few-shot segmentation models to be optimized and strengthened on the target tasks in an interactive and controllable manner and outperforms the state-of-the-art methods by more than 20% in the DSC metric.'}",
-Boosting Few-Shot Learning With Adaptive Margin Loss,Aoxue Li,"Few-shot learning (FSL) has attracted increasing attention in recent years but remains challenging, due to the intrinsic difficulty in learning to generalize from a few examples. This paper proposes an adaptive margin principle to improve the generalization ability of metric-based meta-learning approaches for few-shot learning problems. Specifically, we first develop a class-relevant additive margin loss, where semantic similarity between each pair of classes is considered to separate samples in the feature embedding space from similar classes. Further, we incorporate the semantic context among all classes in a sampled training task and develop a task-relevant additive margin loss to better distinguish samples from different classes. Our adaptive margin method can be easily extended to a more realistic generalized FSL setting. Extensive experiments demonstrate that the proposed method can boost the performance of current metric-based meta-learning approaches, under both the standard FSL and generalized FSL settings.","{'model': 'tldr@v2.0.0', 'text': 'An adaptive margin principle is proposed to improve the generalization ability of metric-based meta-learning approaches for few-shot learning problems by developing a class-relevant additive margin loss, where semantic similarity between each pair of classes is considered to separate samples in the feature embedding space from similar classes.'}",http://arxiv.org/pdf/2005.13826
-Few-Shot Learning via Feature Hallucination with Variational Inference,Qinxuan Luo,"Deep learning has achieved huge success in the field of artificial intelligence, but the performance heavily depends on labeled data. Few-shot learning aims to make a model rapidly adapt to unseen classes with few labeled samples after training on a base dataset, and this is useful for tasks lacking labeled data such as medical image processing. Considering that the core problem of few-shot learning is the lack of samples, a straightforward solution to this issue is data augmentation. This paper proposes a generative model (VI-Net) based on a cosine-classifier baseline. Specifically, we construct a framework to learn to define a generating space for each category in the latent space based on few support samples. In this way, new feature vectors can be generated to help make the decision boundary of classifier sharper during the fine-tuning process. To evaluate the effectiveness of our proposed approach, we perform comparative experiments and ablation studies on mini-ImageNet and CUB. Experimental results show that VI-Net does improve performance compared with the baseline and obtains the state-of-the-art result among other augmentation-based methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a generative model (VI-Net) based on a cosine-classifier baseline and constructs a framework to learn to define a generating space for each category in the latent space based on few support samples to help make the decision boundary of classifier sharper during the fine-tuning process.'}",
-Reordering Examples Helps during Priming-based Few-Shot Learning,Sawan Kumar,"The ability to learn from limited data, or few-shot learning, is a desirable and often critical requirement for NLP systems. While many existing methods do poorly at learning from a handful of examples, large pretrained language models have recently been shown to be efficient few-shot learners. One approach to few-shot learning, which does not require finetuning of model parameters, is to augment the language model's input with priming text which is typically constructed using task specific descriptions and examples. In this work, we further explore priming-based few-shot learning, with focus on using examples as prompts. We show that presenting examples in the right order is key for generalization. We introduce PERO (Prompting with Examples in the Right Order), where we formulate few-shot learning as search over the set of permutations of the training examples. We show that PERO can learn to generalize efficiently using as few as 10 examples, in contrast to existing approaches. While the newline token is a natural choice for separating the examples in the prompt, we show that learning a new separator token can potentially provide further gains in performance. We demonstrate the effectiveness of the proposed method on the tasks of sentiment classification, natural language inference and fact retrieval. Finally, we analyze the learned prompts to reveal novel insights, including the idea that two training examples in the right order alone can provide competitive performance for sentiment classification and natural language inference.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces PERO (Prompting with Examples in the Right Order), where it is shown that PERO can learn to generalize efficiently using as few as 10 examples, in contrast to existing approaches.'}",https://aclanthology.org/2021.findings-acl.395.pdf
-Reinforced Attention for Few-Shot Learning and Beyond,Jie Hong,"Few-shot learning aims to correctly recognize query samples from unseen classes given a limited number of support samples, often by relying on global embeddings of images. In this paper, we propose to equip the backbone network with an attention agent, which is trained by reinforcement learning. The policy gradient algorithm is employed to train the agent towards adaptively localizing the representative regions on feature maps over time. We further design a reward function based on the prediction of the held-out data, thus helping the attention mechanism to generalize better across the unseen classes. The extensive experiments show, with the help of the reinforced attention, that our embedding network has the capability to progressively generate a more discriminative representation in few-shot learning. Moreover, experiments on the task of image classification also show the effectiveness of the proposed design.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to equip the backbone network with an attention agent, which is trained by reinforcement learning, and designs a reward function based on the prediction of the held-out data, thus helping the attention mechanism to generalize better across the unseen classes.'}",https://arxiv.org/pdf/2104.04192
-Multi-Domain Few-Shot Learning and Dataset for Agricultural Applications,Sai Vidyaranya Nuthalapati,"Automatic classification of pests and plants (both healthy and diseased) is of paramount importance in agriculture to improve yield. Conventional deep learning models based on convolutional neural networks require thousands of labeled examples per category. In this work we propose a method to learn from a few samples to automatically classify different pests, plants, and their diseases, using Few-Shot Learning (FSL). We learn a feature extractor to generate embeddings and then update the embeddings using Transformers. Using Mahalanobis distance, a class-covariance-based metric, we then calculate the similarity of the transformed embeddings with the embedding of the image to be classified. Using our proposed architecture, we conduct extensive experiments on multiple datasets showing the effectiveness of our proposed model. We conduct 42 experiments in total to comprehensively analyze the model and it achieves up to 14% and 24% performance gains on few-shot image classification benchmarks on two datasets.We also compile a new FSL dataset containing images of healthy and diseased plants taken in real-world settings. Using our proposed architecture which has been shown to outperform several existing FSL architectures in agriculture, we provide strong baselines on our newly proposed dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to learn from a few samples to automatically classify different pests, plants, and their diseases, using Few-Shot Learning (FSL), which has been shown to outperform several existing FSL architectures in agriculture.'}",https://arxiv.org/pdf/2109.09952
-Distribution Consistency Based Covariance Metric Networks for Few-Shot Learning,Wenbin Li,"Few-shot learning aims to recognize new concepts from very few examples. However, most of the existing few-shot learning methods mainly concentrate on the first-order statistic of concept representation or a fixed metric on the relation between a sample and a concept. In this work, we propose a novel end-to-end deep architecture, named Covariance Metric Networks (CovaMNet). The CovaMNet is designed to exploit both the covariance representation and covariance metric based on the distribution consistency for the few-shot classification tasks. Specifically, we construct an embedded local covariance representation to extract the second-order statistic information of each concept and describe the underlying distribution of this concept. Upon the covariance representation, we further define a new deep covariance metric to measure the consistency of distributions between query samples and new concepts. Furthermore, we employ the episodic training mechanism to train the entire network in an end-to-end manner from scratch. Extensive experiments in two tasks, generic few-shot image classification and fine-grained fewshot image classification, demonstrate the superiority of the proposed CovaMNet. The source code can be available from https://github.com/WenbinLee/CovaMNet.git.","{'model': 'tldr@v2.0.0', 'text': 'The CovaMNet is designed to exploit both the covariance representation and covariance metric based on the distribution consistency for the few-shot classification tasks and employs the episodic training mechanism to train the entire network in an end-to-end manner from scratch.'}",https://ojs.aaai.org/index.php/AAAI/article/download/4885/4758
-Limited Data Rolling Bearing Fault Diagnosis With Few-Shot Learning,Ansi Zhang,"This paper focuses on bearing fault diagnosis with limited training data. A major challenge in fault diagnosis is the infeasibility of obtaining sufficient training samples for every fault type under all working conditions. Recently deep learning based fault diagnosis methods have achieved promising results. However, most of these methods require large amount of training data. In this study, we propose a deep neural network based few-shot learning approach for rolling bearing fault diagnosis with limited data. Our model is based on the siamese neural network, which learns by exploiting sample pairs of the same or different categories. Experimental results over the standard Case Western Reserve University (CWRU) bearing fault diagnosis benchmark dataset showed that our few-shot learning approach is more effective in fault diagnosis with limited data availability. When tested over different noise environments with minimal amount of training data, the performance of our few-shot learning model surpasses the one of the baseline with reasonable noise level. When evaluated over test sets with new fault types or new working conditions, few-shot models work better than the baseline trained with all fault types. All our models and datasets in this study are open sourced and can be downloaded from https://mekhub.cn/as/fault_diagnosis_with_few-shot_learning/.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network based few-shot learning approach for rolling bearing fault diagnosis with limited data availability based on the siamese neural network, which learns by exploiting sample pairs of the same or different categories.'}",https://ieeexplore.ieee.org/ielx7/6287639/8600701/08793060.pdf
-Prototype Rectification for Few-Shot Learning,Jinlu Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple yet effective approach for prototype rectification in transductive setting that utilizes label propagation and feature shifting to diminish the intra-class bias and the cross- class bias.'}",https://arxiv.org/pdf/1911.10713
-Generating Classification Weights With GNN Denoising Autoencoders for Few-Shot Learning,Spyros Gidaris,"Given an initial recognition model already trained on a set of base classes, the goal of this work is to develop a meta-model for few-shot learning. The meta-model, given as input some novel classes with few training examples per class, must properly adapt the existing recognition model into a new model that can correctly classify in a unified way both the novel and the base classes. To accomplish this goal it must learn to output the appropriate classification weight vectors for those two types of classes. To build our meta-model we make use of two main innovations: we propose the use of a Denoising Autoencoder network (DAE) that (during training) takes as input a set of classification weights corrupted with Gaussian noise and learns to reconstruct the target-discriminative classification weights. In this case, the injected noise on the classification weights serves the role of regularizing the weight generating meta-model. Furthermore, in order to capture the co-dependencies between different classes in a given task instance of our meta-model, we propose to implement the DAE model as a Graph Neural Network (GNN). In order to verify the efficacy of our approach, we extensively evaluate it on ImageNet based few-shot benchmarks and we report state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the use of a Denoising Autoencoder network that takes as input a set of classification weights corrupted with Gaussian noise and learns to reconstruct the target-discriminative classification weights, and proposes to implement the DAE model as a Graph Neural Network (GNN).'}",https://arxiv.org/pdf/1905.01102
-Infinite Mixture Prototypes for Few-Shot Learning,Kelsey R. Allen,"We propose infinite mixture prototypes to adaptively represent both simple and complex data distributions for few-shot learning. Our infinite mixture prototypes represent each class by a set of clusters, unlike existing prototypical methods that represent each class by a single cluster. By inferring the number of clusters, infinite mixture prototypes interpolate between nearest neighbor and prototypical representations, which improves accuracy and robustness in the few-shot regime. We show the importance of adaptive capacity for capturing complex data distributions such as alphabets, with 25% absolute accuracy improvements over prototypical networks, while still maintaining or improving accuracy on the standard Omniglot and mini-ImageNet benchmarks. In clustering labeled and unlabeled data by the same clustering rule, infinite mixture prototypes achieves state-of-the-art semi-supervised accuracy. As a further capability, we show that infinite mixture prototypes can perform purely unsupervised clustering, unlike existing prototypical methods.","{'model': 'tldr@v2.0.0', 'text': 'It is shown the importance of adaptive capacity for capturing complex data distributions such as alphabets, with 25% absolute accuracy improvements over prototypical networks, while still maintaining or improving accuracy on the standard Omniglot and mini-ImageNet benchmarks.'}",
-Revisiting Self-training for Few-shot Learning of Language Model,Yiming Chen,"As unlabeled data carry rich task-relevant information, they are proven useful for few-shot learning of language model. The question is how to effectively make use of such data. In this work, we revisit the self-training technique for language model fine-tuning and present a state-of-the-art prompt-based few-shot learner, SFLM. Given two views of a text sample via weak and strong augmentation techniques, SFLM generates a pseudo label on the weakly augmented version. Then, the model predicts the same pseudo label when fine-tuned with the strongly augmented version. This simple approach is shown to outperform other state-of-the-art supervised and semi-supervised counterparts on six sentence classification and six sentence-pair classification benchmarking tasks. In addition, SFLM only relies on a few in-domain unlabeled data. We conduct a comprehensive analysis to demonstrate the robustness of our proposed approach under various settings, including augmentation techniques, model scale, and few-shot knowledge transfer across tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work revisits the self-training technique for language model fine-tuning and presents a state-of-the-art prompt-based few-shot learner, SFLM, which is shown to outperform other state- of- the-art supervised and semi-supervised counterparts on six sentence classification and six sentence-pair classification benchmarking tasks.'}",https://aclanthology.org/2021.emnlp-main.718.pdf
-TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning,Sung Whan Yoon,"Handling previously unseen tasks after given only a few training examples continues to be a tough challenge in machine learning. We propose TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning. Here, employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks. At the same time, for every episode, features in the embedding space are linearly projected into a new space as a form of quick task-specific conditioning. The training loss is obtained based on a distance metric between the query and the reference vectors in the projection space. Excellent generalization results in this way. When tested on the Omniglot, miniImageNet and tieredImageNet datasets, we obtain state of the art classification accuracies under various few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning by employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks.'}",
-STraTA: Self-Training with Task Augmentation for Better Few-shot Learning,Tu Vu,"Despite their recent successes in tackling many NLP tasks, large-scale pre-trained language models do not perform as well in few-shot settings where only a handful of training examples are available. To address this shortcoming, we propose STraTA, which stands for Self-Training with Task Augmentation, an approach that builds on two key ideas for effective leverage of unlabeled data. First, STraTA uses task augmentation, a novel technique that synthesizes a large amount of data for auxiliary-task fine-tuning from target-task unlabeled texts. Second, STraTA performs self-training by further fine-tuning the strong base model created by task augmentation on a broad distribution of pseudo-labeled data. Our experiments demonstrate that STraTA can substantially improve sample efficiency across 12 few-shot benchmarks. Remarkably, on the SST-2 sentiment dataset, STraTA, with only 8 training examples per class, achieves comparable results to standard fine-tuning with 67K training examples. Our analyses reveal that task augmentation and self-training are both complementary and independently effective.","{'model': 'tldr@v2.0.0', 'text': 'STraTA, which stands for Self-Training with Task Augmentation, an approach that builds on two key ideas for effective leverage of unlabeled data, reveals that task augmentation and self-training are both complementary and independently effective.'}",https://aclanthology.org/2021.emnlp-main.462.pdf
-Task-aware Part Mining Network for Few-Shot Learning,Jiamin Wu,"Few-Shot Learning (FSL) aims at classifying samples into new unseen classes with only a handful of labeled samples available. However, most of the existing methods are based on the image-level pooled representation, yet ignore considerable local clues that are transferable across tasks. To address this issue, we propose an end-to-end Task-aware Part Mining Network (TPMN) by integrating an automatic part mining process into the metric-based model for FSL. The proposed TPMN model enjoys several merits. First, we design a meta filter learner to generate task-aware part filters based on the task embedding in a meta-learning way. The task-aware part filters can adapt to any individual task and automatically mine task-related local parts even for an unseen task. Second, an adaptive importance generator is proposed to identify key local parts and assign adaptive importance weights to different parts. To the best of our knowledge, this is the first work to automatically exploit the task-aware local parts in a meta-learning way for FSL. Extensive experimental results on four standard benchmarks demonstrate that the proposed model performs favorably against state-of-the-art FSL methods.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end Task-aware Part Mining Network (TPMN) is proposed by integrating an automatic part mining process into the metric-based model for FSL by designing a meta filter learner to generate task-aware part filters based on the task embedding in a meta-learning way.'}",
-CINS: Comprehensive Instruction for Few-shot Learning in Task-oriented Dialog Systems,Fei Mi,"As the labeling cost for different modules in task-oriented dialog (ToD) systems is high, a major challenge is to learn different tasks with the least amount of labeled data. Recently, pre-trained language models (PLMs) have shown promising results for few-shot learning in ToD. To better utilize the power of PLMs, this paper proposes Comprehensive Instruction (CINS) that exploits PLMs with extra task-specific instructions. We design a schema (definition, constraint, prompt) of instructions and their customized realizations for three important downstream tasks in ToD, ie. intent classification, dialog state tracking, and natural language generation. A sequence-to-sequence model (T5) is adopted to solve these three tasks in a unified framework. Extensive experiments are conducted on these ToD tasks in realistic few-shot learning scenarios with small validation data. Empirical results demonstrate that the proposed CINS approach consistently improves techniques that finetune PLMs with raw input or short prompt.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive Instruction (CINS) is proposed that exploits PLMs with extra task-specific instructions and their customized realizations for three important downstream tasks in ToD, ie.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21356/21105
-Task Agnostic Meta-Learning for Few-Shot Learning,Muhammad Abdullah Jamal,"Meta-learning approaches have been proposed to tackle the few-shot learning problem. Typically, a meta-learner is trained on a variety of tasks in the hopes of being generalizable to new tasks. However, the generalizability on new tasks of a meta-learner could be fragile when it is over-trained on existing tasks during meta-training phase. In other words, the initial model of a meta-learner could be too biased towards existing tasks to adapt to new tasks, especially when only very few examples are available to update the model. To avoid a biased meta-learner and improve its generalizability, we propose a novel paradigm of Task-Agnostic Meta-Learning (TAML) algorithms. Specifically, we present an entropy-based approach that meta-learns an unbiased initial model with the largest uncertainty over the output labels by preventing it from over-performing in classification tasks. Alternatively, a more general inequality-minimization TAML is presented for more ubiquitous scenarios by directly minimizing the inequality of initial losses beyond the classification tasks wherever a suitable loss can be defined. Experiments on benchmarked datasets demonstrate that the proposed approaches outperform compared meta-learning algorithms in both few-shot classification and reinforcement learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'An entropy-based approach that meta-learns an unbiased initial model with the largest uncertainty over the output labels by preventing it from over-performing in classification tasks, which outperforms compared meta-learning algorithms in both few-shot classification and reinforcement learning tasks.'}",https://arxiv.org/pdf/1805.07722
-Story Centaur: Large Language Model Few Shot Learning as a Creative Writing Tool,Benjamin Swanson,"Few shot learning with large language models has the potential to give individuals without formal machine learning training the access to a wide range of text to text models. We consider how this applies to creative writers and present Story Centaur, a user interface for prototyping few shot models and a set of recombinable web components that deploy them. Story Centaur’s goal is to expose creative writers to few shot learning with a simple but powerful interface that lets them compose their own co-creation tools that further their own unique artistic directions. We build out several examples of such tools, and in the process probe the boundaries and issues surrounding generation with large language models.","{'model': 'tldr@v2.0.0', 'text': 'Story Centaur is a user interface for prototyping few shot models and a set of recombinable web components that deploy them to expose creative writers to few shot learning with a simple but powerful interface that lets them compose their own co-creation tools that further their own unique artistic directions.'}",https://aclanthology.org/2021.eacl-demos.29.pdf
-Generalized and Incremental Few-Shot Learning by Explicit Learning and Calibration without Forgetting,Anna Kukleva,"Both generalized and incremental few-shot learning have to deal with three major challenges: learning novel classes from only few samples per class, preventing catastrophic forgetting of base classes, and classifier calibration across novel and base classes. In this work we propose a three-stage framework that allows to explicitly and effectively address these challenges. While the first phase learns base classes with many samples, the second phase learns a calibrated classifier for novel classes from few samples while also preventing catastrophic forgetting. In the final phase, calibration is achieved across all classes. We evaluate the proposed framework on four challenging benchmark datasets for image and video few-shot classification and obtain state-of-the-art results for both generalized and incremental few shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a three-stage framework that allows to explicitly and effectively address the challenges of generalized and incremental few shot learning and evaluates the proposed framework on four challenging benchmark datasets for image and video few-shot classification and obtains state-of-the-art results.'}",https://arxiv.org/pdf/2108.08165
-Graph Prototypical Networks for Few-shot Learning on Attributed Networks,Kaize Ding,"Attributed networks nowadays are ubiquitous in a myriad of high-impact applications, such as social network analysis, financial fraud detection, and drug discovery. As a central analytical task on attributed networks, node classification has received much attention in the research community. In real-world attributed networks, a large portion of node classes only contains limited labeled instances, rendering a long-tail node class distribution. Existing node classification algorithms are unequipped to handle the few-shot node classes. As a remedy, few-shot learning has attracted a surge of attention in the research community. Yet, few-shot node classification remains a challenging problem as we need to address the following questions: (i) How to extract meta-knowledge from an attributed network for few-shot node classification? (ii) How to identify the informativeness of each labeled instance for building a robust and effective model? To answer these questions, in this paper, we propose a graph meta-learning framework -- Graph Prototypical Networks (GPN). By constructing a pool of semi-supervised node classification tasks to mimic the real test environment, GPN is able to perform meta-learning on an attributed network and derive a highly generalizable model for handling the target classification task. Extensive experiments demonstrate the superior capability of GPN in few-shot node classification.","{'model': 'tldr@v2.0.0', 'text': 'By constructing a pool of semi-supervised node classification tasks to mimic the real test environment, GPN is able to perform meta-learning on an attributed network and derive a highly generalizable model for handling the target classification task.'}",https://arxiv.org/pdf/2006.12739
-Pseudo-loss Confidence Metric for Semi-supervised Few-shot Learning,Kai Huang,"Semi-supervised few-shot learning is developed to train a classifier that can adapt to new tasks with limited labeled data and a fixed quantity of unlabeled data. Most semi-supervised few-shot learning methods select pseudo-labeled data of unlabeled set by task-specific confidence estimation. This work presents a task-unified confidence estimation approach for semi-supervised few-shot learning, named pseudo-loss confidence metric (PLCM). It measures the data credibility by the loss distribution of pseudo-labels, which is synthetical considered multi-tasks. Specifically, pseudo-labeled data of different tasks are mapped to a unified metric space by mean of the pseudo-loss model, making it possible to learn the prior pseudo-loss distribution. Then, confidence of pseudo-labeled data is estimated according to the distribution component confidence of its pseudo-loss. Thus highly reliable pseudo-labeled data are selected to strengthen the classifier. Moreover, to overcome the pseudo-loss distribution shift and improve the effectiveness of classifier, we advance the multi-step training strategy coordinated with the class balance measures of class-apart selection and class weight. Experimental results on four popular benchmark datasets demonstrate that the proposed approach can effectively select pseudo-labeled data and achieve the state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A task-unified confidence estimation approach for semi-supervised few-shot learning, named pseudo-loss confidence metric (PLCM), which measures the data credibility by the loss distribution of pseudo-labels, which is synthetical considered multi-tasks.'}",
-Contrastive Prototype Learning with Augmented Embeddings for Few-Shot Learning,Yizhao Gao,"Most recent few-shot learning (FSL) methods are based on meta-learning with episodic training. In each meta-training episode, a discriminative feature embedding and/or classifier are first constructed from a support set in an inner loop, and then evaluated in an outer loop using a query set for model updating. This query set sample centered learning objective is however intrinsically limited in addressing the lack of training data problem in the support set. In this paper, a novel contrastive prototype learning with augmented embeddings (CPLAE) model is proposed to overcome this limitation. First, data augmentations are introduced to both the support and query sets with each sample now being represented as an augmented embedding (AE) composed of concatenated embeddings of both the original and augmented versions. Second, a novel support set class prototype centered contrastive loss is proposed for contrastive prototype learning (CPL). With a class prototype as an anchor, CPL aims to pull the query samples of the same class closer and those of different classes further away. This support set sample centered loss is highly complementary to the existing query centered loss, fully exploiting the limited training data in each episode. Extensive experiments on several benchmarks demonstrate that our proposed CPLAE achieves new state-of-the-art.","{'model': 'tldr@v2.0.0', 'text': 'A novel contrastive prototype learning with augmented embeddings (CPLAE) model is proposed to overcome the lack of training data problem in the support set and achieves new state-of-the-art results.'}",
-Curvature Generation in Curved Spaces for Few-Shot Learning,Zhi Gao,"Few-shot learning describes the challenging problem of recognizing samples from unseen classes given very few labeled examples. In many cases, few-shot learning is cast as learning an embedding space that assigns test samples to their corresponding class prototypes. Previous methods assume that data of all few-shot learning tasks comply with a fixed geometrical structure, mostly a Euclidean structure. Questioning this assumption that is clearly difficult to hold in real-world scenarios and incurs distortions to data, we propose to learn a task-aware curved embedding space by making use of the hyperbolic geometry. As a result, task-specific embedding spaces where suitable curvatures are generated to match the characteristics of data are constructed, leading to more generic embedding spaces. We then leverage on intra-class and inter-class context information in the embedding space to generate class prototypes for discriminative classification. We conduct a comprehensive set of experiments on inductive and transductive few-shot learning, demonstrating the benefits of our proposed method over existing embedding methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to learn a task-aware curved embedding space by making use of the hyperbolic geometry of Euclidean geometry, and leverages on intra-class and inter-class context information in theembedding space to generate class prototypes for discriminative classification.'}",https://openaccess.thecvf.com/content/ICCV2021/papers/Gao_Curvature_Generation_in_Curved_Spaces_for_Few-Shot_Learning_ICCV_2021_paper.pdf
-Squeezing Backbone Feature Distributions to the Max for Efficient Few-Shot Learning,Yuqing Hu,"In many real-life problems, it is difficult to acquire or label large amounts of data, resulting in so-called few-shot learning problems. However, few-shot classification is a challenging problem due to the uncertainty caused by using few labeled samples. In the past few years, many methods have been proposed with the common aim of transferring knowledge acquired on a previously solved task, which is often achieved by using a pretrained feature extractor. As such, if the initial task contains many labeled samples, it is possible to circumvent the limitations of few-shot learning. A shortcoming of existing methods is that they often require priors about the data distribution, such as the balance between considered classes. In this paper, we propose a novel transfer-based method with a double aim: providing state-of-the-art performance, as reported on standardized datasets in the field of few-shot learning, while not requiring such restrictive priors. Our methodology is able to cope with both inductive cases, where prediction is performed on test samples independently from each other, and transductive cases, where a joint (batch) prediction is performed.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel transfer-based method with a double aim: providing state-of-the-art performance, as reported on standardized datasets in the field of few-shot learning, while not requiring such restrictive priors.'}",https://www.mdpi.com/1999-4893/15/5/147/pdf?version=1650974986
-"Z-Score Normalization, Hubness, and Few-Shot Learning",Nanyi Fei,"The goal of few-shot learning (FSL) is to recognize a set of novel classes with only few labeled samples by exploiting a large set of abundant base class samples. Adopting a meta-learning framework, most recent FSL methods meta-learn a deep feature embedding network, and during inference classify novel class samples using nearest neighbor in the learned high-dimensional embedding space. This means that these methods are prone to the hubness problem, that is, a certain class prototype becomes the nearest neighbor of many test instances regardless which classes they belong to. However, this problem is largely ignored in existing FSL studies. In this work, for the first time we show that many FSL methods indeed suffer from the hubness problem. To mitigate its negative effects, we further propose to employ z-score feature normalization, a simple yet effective trans-formation, during meta-training. A theoretical analysis is provided on why it helps. Extensive experiments are then conducted to show that with z-score normalization, the performance of many recent FSL methods can be boosted, resulting in new state-of-the-art on three benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that many FSL methods indeed suffer from the hubness problem, and proposes to employ z-score feature normalization, a simple yet effective trans-formation, during meta-training to mitigate its negative effects.'}",
-A Survey on Few-Shot Learning in Natural Language Processing,Meng Yang,"The annotated dataset is the foundation for Supervised Natural Language Processing. However, the cost of obtaining dataset is high. In recent years, the Few-Shot Learning has gradually attracted the attention of researchers. From the definition, in this paper, we conclude the difference in Few-Shot Learning between Natural Language Processing and Computer Vision. On that basis, the current Few-Shot Learning on Natural Language Processing is summarized, including Transfer Learning, Meta Learning and Knowledge Distillation. Furthermore, we conclude the solutions to Few-Shot Learning in Natural Language Processing, such as the method based on Distant Supervision, Meta Learning and Knowledge Distillation. Finally, we present the challenges facing Few-Shot Learning in Natural Language Processing.","{'model': 'tldr@v2.0.0', 'text': 'The difference in Few-Shot Learning between Natural Language Processing and Computer Vision is concluded and the current Few- shot Learning on Natural Language processing is summarized, including Transfer Learning, Meta Learning and Knowledge Distillation.'}",
-MHFC: Multi-Head Feature Collaboration for Few-Shot Learning,Shuai Shao,"Few-shot learning (FSL) aims to address the data-scarce problem. A standard FSL framework is composed of two components: (1) Pre-train. Employ the base data to generate a CNN-based feature extraction model (FEM). (2) Meta-test. Apply the trained FEM to acquire the novel data's features and recognize them. FSL relies heavily on the design of the FEM. However, various FEMs have distinct emphases. For example, several may focus more attention on the contour information, whereas others may lay particular emphasis on the texture information. The single-head feature is only a one-sided representation of the sample. Besides the negative influence of cross-domain (e.g., the trained FEM can not adapt to the novel class flawlessly), the distribution of novel data may have a certain degree of deviation compared with the ground truth distribution, which is dubbed as distribution-shift-problem (DSP). To address the DSP, we propose Multi-Head Feature Collaboration (MHFC) algorithm, which attempts to project the multi-head features (e.g., multiple features extracted from a variety of FEMs) to a unified space and fuse them to capture more discriminative information. Typically, first, we introduce a subspace learning method to transform the multi-head features to aligned low-dimensional representations. It corrects the DSP via learning the feature with more powerful discrimination and overcomes the problem of inconsistent measurement scales from different head features. Then, we design an attention block to update combination weights for each head feature automatically. It comprehensively considers the contribution of various perspectives and further improves the discrimination of features. We evaluate the proposed method on five benchmark datasets (including cross-domain experiments) and achieve significant improvements of 2.1%-7.8% compared with state-of-the-arts.","{'model': 'tldr@v2.0.0', 'text': 'Multi-Head Feature Collaboration (MHFC) algorithm is proposed, which attempts to project the multi-head features to a unified space and fuse them to capture more discriminative information.'}",https://arxiv.org/pdf/2109.07785
-Dense Classification and Implanting for Few-Shot Learning,Yann Lifchitz,"Few-shot learning for deep neural networks is a highly challenging and key problem in many computer vision tasks. In this context, we are targeting knowledge transfer from a set with abundant data to other sets with few available examples. We propose two simple and effective solutions: (i) dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and (ii) implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features. Implanting enables training of multiple layers in the few-shot regime, departing from most related methods derived from metric learning that train only the final layer. Both contributions show consistent gains when used individually or jointly and we report state of the art performance on few-shot classification on miniImageNet.","{'model': 'tldr@v2.0.0', 'text': 'Two simple and effective solutions are proposed: dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features.'}",https://arxiv.org/pdf/1903.05050
-Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning,Limeng Qiao,"Few-shot learning, which aims at extracting new concepts rapidly from extremely few examples of novel classes, has been featured into the meta-learning paradigm recently. Yet, the key challenge of how to learn a generalizable classifier with the capability of adapting to specific tasks with severely limited data still remains in this domain. To this end, we propose a Transductive Episodic-wise Adaptive Metric (TEAM) framework for few-shot learning, by integrating the meta-learning paradigm with both deep metric learning and transductive inference. With exploring the pairwise constraints and regularization prior within each task, we explicitly formulate the adaptation procedure into a standard semi-definite programming problem. By solving the problem with its closed-form solution on the fly with the setup of transduction, our approach efficiently tailors an episodic-wise metric for each task to adapt all features from a shared task-agnostic embedding space into a more discriminative task-specific metric space. Moreover, we further leverage an attention-based bi-directional similarity strategy for extracting the more robust relationship between queries and prototypes. Extensive experiments on three benchmark datasets show that our framework is superior to other existing approaches and achieves the state-of-the-art performance in the few-shot literature.","{'model': 'tldr@v2.0.0', 'text': 'A Transductive Episodic-wise Adaptive Metric (TEAM) framework for few-shot learning is proposed, by integrating the meta-learning paradigm with both deep metric learning and transductive inference and leverages an attention-based bi-directional similarity strategy for extracting the more robust relationship between queries and prototypes.'}",https://arxiv.org/pdf/1910.02224
-LibFewShot: A Comprehensive Library for Few-Shot Learning,Wenbin Li,"Few-shot learning, especially few-shot image classification, has received increasing attention and witnessed significant advances in recent years. Some recent studies implicitly show that many generic techniques or “tricks”, such as data augmentation, pre-training, knowledge distillation, and self-supervision, may greatly boost the performance of a few-shot learning method. Moreover, different works may employ different software platforms, backbone architectures and input image sizes, making fair comparisons difficult and practitioners struggle with reproducibility. To address these situations, we propose a comprehensive library for few-shot learning (LibFewShot) by re-implementing eighteen state-of-the-art few-shot learning methods in a unified framework with the same single codebase in PyTorch. Furthermore, based on LibFewShot, we provide comprehensive evaluations on multiple benchmarks with various backbone architectures to evaluate common pitfalls and effects of different training tricks. In addition, with respect to the recent doubts on the necessity of meta- or episodic-training mechanism, our evaluation results confirm that such a mechanism is still necessary especially when combined with pre-training. We hope our work can not only lower the barriers for beginners to enter the area of few-shot learning but also elucidate the effects of nontrivial tricks to facilitate intrinsic research on few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a comprehensive library for few-shot learning (LibFewShot) by re-implementing eighteen state-of-the-art few- shot learning methods in a unified framework with the same single codebase in PyTorch, and provides comprehensive evaluations on multiple benchmarks with various backbone architectures to evaluate common pitfalls and effects of different training tricks.'}",https://arxiv.org/pdf/2109.04898
-Multi-Label Few-Shot Learning for Aspect Category Detection,Mengting Hu,"Aspect category detection (ACD) in sentiment analysis aims to identify the aspect categories mentioned in a sentence. In this paper, we formulate ACD in the few-shot learning scenario. However, existing few-shot learning approaches mainly focus on single-label predictions. These methods can not work well for the ACD task since a sentence may contain multiple aspect categories. Therefore, we propose a multi-label few-shot learning method based on the prototypical network. To alleviate the noise, we design two effective attention mechanisms. The support-set attention aims to extract better prototypes by removing irrelevant aspects. The query-set attention computes multiple prototype-specific representations for each query instance, which are then used to compute accurate distances with the corresponding prototypes. To achieve multi-label inference, we further learn a dynamic threshold per instance by a policy network. Extensive experimental results on three datasets demonstrate that the proposed method significantly outperforms strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A multi-label few-shot learning method based on the prototypical network for aspect category detection that significantly outperforms strong baselines and also learns a dynamic threshold per instance by a policy network.'}",https://aclanthology.org/2021.acl-long.495.pdf
-MetaGAN: An Adversarial Approach to Few-Shot Learning,Ruixiang Zhang,"In this paper, we propose a conceptually simple and general framework called MetaGAN for few-shot learning problems. Most state-of-the-art few-shot classification models can be integrated with MetaGAN in a principled and straightforward way. By introducing an adversarial generator conditioned on tasks, we augment vanilla few-shot classification models with the ability to discriminate between real and fake data. We argue that this GAN-based approach can help few-shot classifiers to learn sharper decision boundary, which could generalize better. We show that with our MetaGAN framework, we can extend supervised few-shot learning models to naturally cope with unlabeled data. Different from previous work in semi-supervised few-shot learning, our algorithms can deal with semi-supervision at both sample-level and task-level. We give theoretical justifications of the strength of MetaGAN, and validate the effectiveness of MetaGAN on challenging few-shot image classification benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a conceptually simple and general framework called MetaGAN for few-shot learning problems, and shows that with this MetaGAN framework, it can extend supervised few- shot learning models to naturally cope with unlabeled data.'}",
-When Does Self-supervision Improve Few-shot Learning?,Jong-Chyi Su,,"{'model': 'tldr@v2.0.0', 'text': 'This work investigates the role of self-supervised learning in the context of few-shot learning and presents a technique that automatically selects images for SSL from a large, generic pool of unlabeled images for a given dataset that provides further improvements.'}",https://arxiv.org/pdf/1910.03560
-ECKPN: Explicit Class Knowledge Propagation Network for Transductive Few-shot Learning,Chaofan CHEN,"Recently, the transductive graph-based methods have achieved great success in the few-shot classification task. However, most existing methods ignore exploring the class-level knowledge that can be easily learned by humans from just a handful of samples. In this paper, we propose an Explicit Class Knowledge Propagation Network (ECKPN), which is composed of the comparison, squeeze and calibration modules, to address this problem. Specifically, we first employ the comparison module to explore the pairwise sample relations to learn rich sample representations in the instance-level graph. Then, we squeeze the instance-level graph to generate the class-level graph, which can help obtain the class-level visual knowledge and facilitate modeling the relations of different classes. Next, the calibration module is adopted to characterize the relations of the classes explicitly to obtain the more discriminative class-level knowledge representations. Finally, we combine the class-level knowledge with the instance-level sample representations to guide the inference of the query samples. We conduct extensive experiments on four few-shot classification benchmarks, and the experimental results show that the proposed ECKPN significantly outperforms the state-of-the art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an Explicit Class Knowledge Propagation Network (ECKPN), which is composed of the comparison, squeeze and calibration modules, and shows that the proposed ECKPN significantly outperforms the state-of-the art methods.'}",https://arxiv.org/pdf/2106.08523
-Few-Shot Learning With Embedded Class Models and Shot-Free Meta Training,Avinash Ravichandran,"We propose a method for learning embeddings for few-shot learning that is suitable for use with any number of shots (shot-free). Rather than fixing the class prototypes to be the Euclidean average of sample embeddings, we allow them to live in a higher-dimensional space (embedded class models) and learn the prototypes along with the model parameters. The class representation function is defined implicitly, which allows us to deal with a variable number of shots per class with a simple constant-size architecture. The class embedding encompasses metric learning, that facilitates adding new classes without crowding the class representation space. Despite being general and not tuned to the benchmark, our approach achieves state-of-the-art performance on the standard few-shot benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for learning embeddings for few-shot learning that is suitable for use with any number of shots (shot-free), that encompasses metric learning, that facilitates adding new classes without crowding the class representation space.'}",https://arxiv.org/pdf/1905.04398
-Large-Scale Few-Shot Learning: Knowledge Transfer With Class Hierarchy,Aoxue Li,"Recently, large-scale few-shot learning (FSL) becomes topical. It is discovered that, for a large-scale FSL problem with 1,000 classes in the source domain, a strong baseline emerges, that is, simply training a deep feature embedding model using the aggregated source classes and performing nearest neighbor (NN) search using the learned features on the target classes. The state-of-the-art large-scale FSL methods struggle to beat this baseline, indicating intrinsic limitations on scalability. To overcome the challenge, we propose a novel large-scale FSL model by learning transferable visual features with the class hierarchy which encodes the semantic relations between source and target classes. Extensive experiments show that the proposed model significantly outperforms not only the NN baseline but also the state-of-the-art alternatives. Furthermore, we show that the proposed model can be easily extended to the large-scale zero-shot learning (ZSL) problem and also achieves the state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel large-scale FSL model by learning transferable visual features with the class hierarchy which encodes the semantic relations between source and target classes and significantly outperforms not only the NN baseline but also the state-of-the-art alternatives.'}",
-Variational Few-Shot Learning,Jian Zhang,"We propose a variational Bayesian framework for enhancing few-shot learning performance. This idea is motivated by the fact that single point based metric learning approaches are inherently noise-vulnerable and easy-to-be-biased. In a nutshell, stochastic variational inference is invoked to approximate bias-eliminated class specific sample distributions. In the meantime, a classifier-free prediction is attained by leveraging the distribution statistics on novel samples. Extensive experimental results on several benchmarks well demonstrate the effectiveness of our distribution-driven few-shot learning framework over previous point estimates based methods, in terms of superior classification accuracy and robustness.","{'model': 'tldr@v2.0.0', 'text': 'Stochastic variational inference is invoked to approximate bias-eliminated class specific sample distributions to enhance few-shot learning performance, and a classifier-free prediction is attained by leveraging the distribution statistics on novel samples.'}",
-Few-Shot Learning via Saliency-Guided Hallucination of Samples,Hongguang Zhang,"Learning new concepts from a few of samples is a standard challenge in computer vision. The main directions to improve the learning ability of few-shot training models include (i) a robust similarity learning and (ii) generating or hallucinating additional data from the limited existing samples. In this paper, we follow the latter direction and present a novel data hallucination model. Currently, most datapoint generators contain a specialized network (i.e., GAN) tasked with hallucinating new datapoints, thus requiring large numbers of annotated data for their training in the first place. In this paper, we propose a novel less-costly hallucination method for few-shot learning which utilizes saliency maps. To this end, we employ a saliency network to obtain the foregrounds and backgrounds of available image samples and feed the resulting maps into a two-stream network to hallucinate datapoints directly in the feature space from viable foreground-background combinations. To the best of our knowledge, we are the first to leverage saliency maps for such a task and we demonstrate their usefulness in hallucinating additional datapoints for few-shot learning. Our proposed network achieves the state of the art on publicly available datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work employs a saliency network to obtain the foregrounds and backgrounds of available image samples and feeds the resulting maps into a two-stream network to hallucinate datapoints directly in the feature space from viable foreground-background combinations, the first to leverage saliency maps for such a task.'}",https://arxiv.org/pdf/1904.03472
-Looking Wider for Better Adaptive Representation in Few-Shot Learning,Jiabao Zhao,"Building a good feature space is essential for the metric-based few-shot algorithms to recognize a novel class with only a few samples. The feature space is often built by Convolutional Neural Networks (CNNs). However, CNNs primarily focus on local information with the limited receptive field, and the global information generated by distant pixels is not well used. Meanwhile, having a global understanding of the current task and focusing on distinct regions of the same sample for different queries are important for the few-shot classification. To tackle these problems, we propose the Cross Non-Local Neural Network (CNL) for capturing the long-range dependency of the samples and the current task. CNL extracts the task-specific and context-aware features dynamically by strengthening the features of the sample at a position via aggregating information from all positions of itself and the current task. To reduce losing important information, we maximize the mutual information between the original and refined features as a constraint. Moreover, we add a task-specific scaling to deal with multi-scale and task-specific features extracted by CNL. We conduct extensive experiments for validating our proposed algorithm, which achieves new state-of-the-art performances on two public benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'The Cross Non-Local Neural Network (CNL) is proposed for capturing the long-range dependency of the samples and the current task, and extracts the task-specific and context-aware features dynamically by strengthening the features of the sample at a position via aggregating information from all positions of itself and theCurrent task.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17311/17118
-Graph Few-shot Learning via Knowledge Transfer,Huaxiu Yao,"Towards the challenging problem of semi-supervised node classification, there have been extensive studies. As a frontier, Graph Neural Networks (GNNs) have aroused great interest recently, which update the representation of each node by aggregating information of its neighbors. However, most GNNs have shallow layers with a limited receptive field and may not achieve satisfactory performance especially when the number of labeled nodes is quite small. To address this challenge, we innovatively propose a graph few-shot learning (GFL) algorithm that incorporates prior knowledge learned from auxiliary graphs to improve classification accuracy on the target graph. Specifically, a transferable metric space characterized by a node embedding and a graph-specific prototype embedding function is shared between auxiliary graphs and the target, facilitating the transfer of structural knowledge. Extensive experiments and ablation studies on four real-world graph datasets demonstrate the effectiveness of our proposed model and the contribution of each component.","{'model': 'tldr@v2.0.0', 'text': 'This work innovatively proposes a graph few-shot learning (GFL) algorithm that incorporates prior knowledge learned from auxiliary graphs to improve classification accuracy on the target graph.'}",https://ojs.aaai.org/index.php/AAAI/article/download/6142/5998
-Attentive Weights Generation for Few Shot Learning via Information Maximization,Yiluan Guo,"Few shot image classification aims at learning a classifier from limited labeled data. Generating the classification weights has been applied in many meta-learning methods for few shot image classification due to its simplicity and effectiveness. In this work, we present Attentive Weights Generation for few shot learning via Information Maximization (AWGIM), which introduces two novel contributions: i) Mutual information maximization between generated weights and data within the task; this enables the generated weights to retain information of the task and the specific query sample. ii) Self-attention and cross-attention paths to encode the context of the task and individual queries. Both two contributions are shown to be very effective in extensive experiments. Overall, AWGIM is competitive with state-of-the-art. Code is available at https://github.com/Yiluan/AWGIM.","{'model': 'tldr@v2.0.0', 'text': 'Attentive Weights Generation for few shot learning via Information Maximization (AWGIM), which introduces two novel contributions: i) Mutual information maximization between generated weights and data within the task; this enables the generated weights to retain information of the task and the specific query sample.'}",
-Few-Shot Learning for Palmprint Recognition via Meta-Siamese Network,Huikai Shao,"Palmprint is one of the discriminant biometric modalities of humans. Recently, deep learning-based palmprint recognition algorithms have improved the accuracy and robustness of recognition results to a new level. Most of them require a large amount of labeled training samples to guarantee satisfactory performance. However, getting enough labeled data is difficult due to time consumption and privacy issues. Therefore, in this article, a novel meta-Siamese network (MSN) is proposed to exploit few-shot learning for small-sample palmprint recognition. During each episode-based training iteration, a few images are selected as sample and query sets to simulate the support and testing sets in the test set. Specifically, the model is trained episodically with a flexible framework to learn both the feature embedding and deep similarity metric function. In addition, two distance-based losses are introduced to assist the optimization. After training, the model can learn the ability to get similarity scores between two images for few-shot testing. Adequate experiments conducted on several constrained and unconstrained benchmark palmprint databases show that MSN can obtain competitive improvements compared with baseline methods, where the best accuracy can be up to 100%.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-Siamese network is proposed to exploit few-shot learning for small-sample palmprint recognition and can obtain competitive improvements compared with baseline methods, where the best accuracy can be up to 100%.'}",https://research.utwente.nl/files/254372707/09420130.pdf
-Incremental Few-Shot Learning with Attention Attractor Networks,Mengye Ren,"Machine learning classifiers are often trained to recognize a set of pre-defined classes. However, in many applications, it is often desirable to have the flexibility of learning additional concepts, with limited data and without re-training on the full training set. This paper addresses this problem, incremental few-shot learning, where a regular classification network has already been trained to recognize a set of base classes, and several extra novel classes are being considered, each with only a few labeled examples. After learning the novel classes, the model is then evaluated on the overall classification performance on both base and novel classes. To this end, we propose a meta-learning model, the Attention Attractor Network, which regularizes the learning of novel classes. In each episode, we train a set of new weights to recognize novel classes until they converge, and we show that the technique of recurrent back-propagation can back-propagate through the optimization process and facilitate the learning of these parameters. We demonstrate that the learned attractor network can help recognize novel classes while remembering old classes without the need to review the original training set, outperforming various baselines.","{'model': 'tldr@v2.0.0', 'text': 'A meta-learning model, the Attention Attractor Network, which regularizes the learning of novel classes, and it is demonstrated that the learned attractor network can help recognize novel classes while remembering old classes without the need to review the original training set.'}",
-Radar target recognition based on few-shot learning,Yue-Wen Yang,,"{'model': 'tldr@v2.0.0', 'text': 'A sample selection method based on few-shot learning is proposed, which can achieve similar results with 19 kinds of trajectory data training model and the accuracy is improved by 4–10% in different time lengths.'}",
-Hybrid Graph Neural Networks for Few-Shot Learning,Tianyuan Yu,"Graph neural networks (GNNs) have been used to tackle the few-shot learning (FSL) problem and shown great potentials under the transductive setting. However under the inductive setting, existing GNN based methods are less competitive. This is because they use an instance GNN as a label propagation/classification module, which is jointly meta-learned with a feature embedding network. This design is problematic because the classifier needs to adapt quickly to new tasks while the embedding does not. To overcome this problem, in this paper we propose a novel hybrid GNN (HGNN) model consisting of two GNNs, an instance GNN and a prototype GNN. Instead of label propagation, they act as feature embedding adaptation modules for quick adaptation of the meta-learned feature embedding to new tasks. Importantly they are designed to deal with a fundamental yet often neglected challenge in FSL, that is, with only a handful of shots per class, any few-shot classifier would be sensitive to badly sampled shots which are either outliers or can cause inter-class distribution overlapping. Extensive experiments show that our HGNN obtains new state-of-the-art on three FSL benchmarks. The code and models are available at https://github.com/TianyuanYu/HGNN.","{'model': 'tldr@v2.0.0', 'text': 'A novel hybrid GNN (HGNN) model consisting of two GNNs, an instance GNN and a prototype GNN designed to deal with a fundamental yet often neglected challenge in FSL, that is, with only a handful of shots per class, any few-shot classifier would be sensitive to badly sampled shots.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20226/19985
-Federated Few-Shot Learning with Adversarial Learning,Chenyou Fan,"We are interested in developing a unified machine learning framework for effectively training machine learning models from many small data sources such as mobile devices. This is a commonly encountered situation in mobile computing scenarios, where data is scarce and distributed while the tasks are distinct. In this paper, we propose a federated few-shot learning (FedFSL) framework to learn a few-shot classification model that can classify unseen data classes with only a few labeled samples. With the federated learning strategy, FedFSL can utilize many data sources while keeping data privacy and communication efficiency. To tackle the issue of obtaining misaligned decision boundaries produced by client models, we propose to regularize local updates by minimizing the divergence of client models. We also formulate the training in an adversarial fashion and optimize the client models to produce a discriminative feature space that can better represent unseen data samples. We demonstrate the intuitions and conduct experiments to show our approaches outperform baselines by more than 10% in learning benchmark vision tasks and 5% in language tasks.","{'model': 'tldr@v2.0.0', 'text': 'A federated few-shot learning (FedFSL) framework to learn a few- shot classification model that can classify unseen data classes with only a few labeled samples and formulate the training in an adversarial fashion and optimize the client models to produce a discriminative feature space that can better represent unseen data samples is proposed.'}",https://arxiv.org/pdf/2104.00365
-Few-Shot Learning with a Strong Teacher,Han-Jia Ye,"Few-shot learning (FSL) aims to generate a classifier using limited labeled examples. Many existing works take the meta-learning approach, constructing a few-shot learner (a meta-model) that can learn from few-shot examples to generate a classifier. The performance is measured by how well the resulting classifiers classify the test (\ie, query) examples of those tasks. In this paper, we point out two potential weaknesses of this approach. First, the sampled query examples may not provide sufficient supervision for meta-training the few-shot learner. Second, the effectiveness of meta-learning diminishes sharply with the increasing number of shots. We propose a novel meta-training objective for the few-shot learner, which is to encourage the few-shot learner to generate classifiers that perform like strong classifiers. Concretely, we associate each sampled few-shot task with a strong classifier, which is trained with ample labeled examples. The strong classifiers can be seen as the target classifiers that we hope the few-shot learner to generate given few-shot examples, and we use the strong classifiers to supervise the few-shot learner. We validate our approach in combinations with many representative meta-learning methods. More importantly, with our approach, meta-learning based FSL methods can consistently outperform non-meta-learning based methods at different numbers of shots.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-training objective for the few-shot learner is proposed, which is to encourage theFew- shot learner to generate classifiers that perform like strong classifiers, and meta-learning based FSL methods can consistently outperform non-meta- learning based methods at different numbers of shots.'}",https://arxiv.org/pdf/2107.00197
-Few-shot Learning: A Survey,Yaqing Wang,"The quest of `can machines think' and `can machines do what human do' are quests that drive the development of artificial intelligence. Although recent artificial intelligence succeeds in many data intensive applications, it still lacks the ability of learning from limited exemplars and fast generalizing to new tasks. To tackle this problem, one has to turn to machine learning, which supports the scientific study of artificial intelligence. Particularly, a machine learning problem called Few-Shot Learning (FSL) targets at this case. It can rapidly generalize to new tasks of limited supervised experience by turning to prior knowledge, which mimics human's ability to acquire knowledge from few examples through generalization and analogy. It has been seen as a test-bed for real artificial intelligence, a way to reduce laborious data gathering and computationally costly training, and antidote for rare cases learning. With extensive works on FSL emerging, we give a comprehensive survey for it. We first give the formal definition for FSL. Then we point out the core issues of FSL, which turns the problem from ""how to solve FSL"" to ""how to deal with the core issues"". Accordingly, existing works from the birth of FSL to the most recent published ones are categorized in a unified taxonomy, with thorough discussion of the pros and cons for different categories. Finally, we envision possible future directions for FSL in terms of problem setup, techniques, applications and theory, hoping to provide insights to both beginners and experienced researchers.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of the core issues of Few-Shot Learning, and existing works from the birth of FSL to the most recent published ones are categorized in a unified taxonomy, with thorough discussion of the pros and cons for different categories.'}",
-Few Shot Learning Based on the Street View House Numbers (SVHN) Dataset,Rundong Yang,,,
-Mutual CRF-GNN for Few-shot Learning,Shixiang Tang,"Graph-neural-networks (GNN) is a rising trend for fewshot learning. A critical component in GNN is the affinity. Typically, affinity in GNN is mainly computed in the feature space, e.g., pairwise features, and does not take fully advantage of semantic labels associated to these features. In this paper, we propose a novel Mutual CRF-GNN (MCGN). In this MCGN, the labels and features of support data are used by the CRF for inferring GNN affinities in a principled and probabilistic way. Specifically, we construct a Conditional Random Field (CRF) conditioned on labels and features of support data to infer a affinity in the label space. Such affinity is fed to the GNN as the node-wise affinity. GNN and CRF mutually contributes to each other in MCGN. For GNN, CRF provides valuable affinity information. For CRF, GNN provides better features for inferring affinity. Experimental results show that our approach outperforms stateof-the-arts on datasets miniImageNet, tieredImageNet, and CIFAR-FS on both 5-way 1-shot and 5-way 5-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper constructs a Conditional Random Field conditioned on labels and features of support data to infer a affinity in the label space and proposes a novel Mutual CRF-GNN (MCGN), where GNN and CRF mutually contributes to each other in MCGN.'}",
-Demystification of Few-shot and One-shot Learning,I. Tyukin,"Few-shot and one-shot learning have been the subject of active and intensive research in recent years, with mounting evidence pointing to successful implementation and exploitation of few-shot learning algorithms in practice. Classical statistical learning theories do not fully explain why few- or one-shot learning is at all possible since traditional generalisation bounds normally require large training and testing samples to be meaningful. This sharply contrasts with numerous examples of successful one- and few-shot learning systems and applications. In this work we present mathematical foundations for a theory of one-shot and few-shot learning and reveal conditions specifying when such learning schemes are likely to succeed. Our theory is based on intrinsic properties of high-dimensional spaces. We show that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met. In this work we present mathematical foundations for a theory of one-shot and few-shot learning and reveal conditions specifying when such learning schemes are likely to succeed. Our theory is based on intrinsic properties of high-dimensional spaces. We show that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met.'}",https://arxiv.org/pdf/2104.12174
-Few-Shot Learning approach for plant disease classification using images taken in the field,David Argüeso,,"{'model': 'tldr@v2.0.0', 'text': 'It is possible to learn new plant leaf and disease types with very small datasets using deep learning Siamese networks with Triplet loss, achieving almost a 90% reduction in training data needs and outperforming classical learning techniques for small training sets.'}",
-Few-Shot Learning With Localization in Realistic Settings,Davis Wertheimer,"Traditional recognition methods typically require large, artificially-balanced training classes, while few-shot learning methods are tested on artificially small ones. In contrast to both extremes, real world recognition problems exhibit heavy-tailed class distributions, with cluttered scenes and a mix of coarse and fine-grained class distinctions. We show that prior methods designed for few-shot learning do not work out of the box in these challenging conditions, based on a new “meta-iNat” benchmark. We introduce three parameter-free improvements: (a) better training procedures based on adapting cross-validation to meta-learning, (b) novel architectures that localize objects using limited bounding box annotations before classification, and (c) simple parameter-free expansions of the feature space based on bilinear pooling. Together, these improvements double the accuracy of state-of-the-art models on meta-iNat while generalizing to prior benchmarks, complex neural architectures, and settings with substantial domain shift.","{'model': 'tldr@v2.0.0', 'text': 'Three parameter-free improvements are introduced that double the accuracy of state-of-the-art models on meta-iNat while generalizing to prior benchmarks, complex neural architectures, and settings with substantial domain shift.'}",http://arxiv.org/pdf/1904.08502
-Knowledge-Guided Multi-Label Few-Shot Learning for General Image Recognition,Tianshui Chen,"Recognizing multiple labels of an image is a practical yet challenging task, and remarkable progress has been achieved by searching for semantic regions and exploiting label dependencies. However, current works utilize RNN/LSTM to implicitly capture sequential region/label dependencies, which cannot fully explore mutual interactions among the semantic regions/labels and do not explicitly integrate label co-occurrences. In addition, these works require large amounts of training samples for each category, and they are unable to generalize to novel categories with limited samples. To address these issues, we propose a knowledge-guided graph routing (KGGR) framework, which unifies prior knowledge of statistical label correlations with deep neural networks. The framework exploits prior knowledge to guide adaptive information propagation among different categories to facilitate multi-label analysis and reduce the dependency of training samples. Specifically, it first builds a structured knowledge graph to correlate different labels based on statistical label co-occurrence. Then, it introduces the label semantics to guide learning semantic-specific features to initialize the graph, and it exploits a graph propagation network to explore graph node interactions, enabling learning contextualized image feature representations. Moreover, we initialize each graph node with the classifier weights for the corresponding label and apply another propagation network to transfer node messages through the graph. In this way, it can facilitate exploiting the information of correlated labels to help train better classifiers, especially for labels with limited training samples. We conduct extensive experiments on the traditional multi-label image recognition (MLR) and multi-label few-shot learning (ML-FSL) tasks and show that our KGGR framework outperforms the current state-of-the-art methods by sizable margins on the public benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A knowledge-guided graph routing (KGGR) framework, which unifies prior knowledge of statistical label correlations with deep neural networks and can facilitate exploiting the information of correlated labels to help train better classifiers, especially for labels with limited training samples.'}",https://arxiv.org/pdf/2009.09450
-Automatic Modulation Recognition: A Few-Shot Learning Method Based on the Capsule Network,Lixin Li,"With the rapid development of deep learning (DL) in recent years, automatic modulation recognition (AMR) with DL has achieved high accuracy. However, aiming to obtain higher classification accuracy, DL requires numerous training samples. In order to solve this problem, it is a challenge to study how to efficiently use DL for AMR in the case of few samples. In this letter, inspired by the capsule network (CapsNet), we propose a new network structure named AMR-CapsNet to achieve higher classification accuracy of modulation signals with fewer samples, and further analyze the adaptability of DL models in the case of few samples. The simulation results demonstrate that when 3% of the dataset is used to train and the signal-to-noise ratio (SNR) is greater than 2 dB, the overall classification accuracy of the AMR-CapsNet is greater than 80%. Compared with convolutional neural network (CNN), the classification accuracy is improved by 20%.","{'model': 'tldr@v2.0.0', 'text': 'A new network structure named AMR-CapsNet is proposed to achieve higher classification accuracy of modulation signals with fewer samples, and further analyze the adaptability of DL models in the case of few samples.'}",
-Local descriptor-based multi-prototype network for few-shot Learning,Hongwei Huang,,"{'model': 'tldr@v2.0.0', 'text': 'A novel Local descriptor-based Multi-Prototype Network (LMPNet) is proposed, a well-designed framework that generates an embedding space with multiple prototypes that can capture more informative and subtler cues of an image than the normally adopted image-level features.'}",
-Collect and Select: Semantic Alignment Metric Learning for Few-Shot Learning,Fusheng Hao,"Few-shot learning aims to learn latent patterns from few training examples and has shown promises in practice. However, directly calculating the distances between the query image and support image in existing methods may cause ambiguity because dominant objects can locate anywhere on images. To address this issue, this paper proposes a Semantic Alignment Metric Learning (SAML) method for few-shot learning that aligns the semantically relevant dominant objects through a ``collect-and-select'' strategy. Specifically, we first calculate a relation matrix (RM) to ``collect"" the distances of each local region pairs of the $3$D tensor extracted from a query image and the mean tensor of the support images. Then, the attention technique is adapted to ``select"" the semantically relevant pairs and put more weights on them. Afterwards, a multi-layer perceptron (MLP) is utilized to map the reweighted RMs to their corresponding similarity scores. Theoretical analysis demonstrates the generalization ability of SAML and gives a theoretical guarantee. Empirical results demonstrate that semantic alignment is achieved. Extensive experiments on benchmark datasets validate the strengths of the proposed approach and demonstrate that SAML significantly outperforms the current state-of-the-art methods. The source code is available at https://github.com/haofusheng/SAML.","{'model': 'tldr@v2.0.0', 'text': ""A Semantic Alignment Metric Learning (SAML) method for few-shot learning that aligns the semantically relevant dominant objects through a ``collect-and-select'' strategy and demonstrates the generalization ability of SAML and gives a theoretical guarantee.""}",
-HoloDetect: Few-Shot Learning for Error Detection,Alireza Heidari,"We introduce a few-shot learning framework for error detection. We show that data augmentation (a form of weak supervision) is key to training high-quality, ML-based error detection models that require minimal human involvement. Our framework consists of two parts: (1) an expressive model to learn rich representations that capture the inherent syntactic and semantic heterogeneity of errors; and (2) a data augmentation model that, given a small seed of clean records, uses dataset-specific transformations to automatically generate additional training data. Our key insight is to learn data augmentation policies from the noisy input dataset in a weakly supervised manner. We show that our framework detects errors with an average precision of ~94% and an average recall of ~93% across a diverse array of datasets that exhibit different types and amounts of errors. We compare our approach to a comprehensive collection of error detection methods, ranging from traditional rule-based methods to ensemble-based and active learning approaches. We show that data augmentation yields an average improvement of 20 F1 points while it requires access to 3x fewer labeled examples compared to other ML approaches.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning framework for error detection that consists of an expressive model to learn rich representations that capture the inherent syntactic and semantic heterogeneity of errors, and a data augmentation model that, given a small seed of clean records, uses dataset-specific transformations to automatically generate additional training data.'}",https://dl.acm.org/doi/pdf/10.1145/3299869.3319888
-Meta Navigator: Search for a Good Adaptation Policy for Few-shot Learning,Chi Zhang,"Few-shot learning aims to adapt knowledge learned from previous tasks to novel tasks with only a limited amount of labeled data. Research literature on few-shot learning exhibits great diversity, while different algorithms often excel at different few-shot learning scenarios. It is therefore tricky to decide which learning strategies to use under different task conditions. Inspired by the recent success in Automated Machine Learning literature (AutoML), in this paper, we present Meta Navigator, a framework that attempts to solve the aforementioned limitation in few-shot learning by seeking a higher-level strategy and proffer to automate the selection from various few-shot learning designs. The goal of our work is to search for good parameter adaptation policies that are applied to different stages in the network for few-shot classification. We present a search space that covers many popular few-shot learning algorithms in the literature, and develop a differentiable searching and decoding algorithm based on meta-learning that supports gradient-based optimization. We demonstrate the effectiveness of our searching-based method on multiple benchmark datasets. Extensive experiments show that our approach significantly outperforms baselines and demonstrates performance advantages over many state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the recent success in Automated Machine Learning literature, this paper presents Meta Navigator, a framework that attempts to solve the aforementioned limitation in few-shot learning by seeking a higher-level strategy and proffer to automate the selection from various few- shot learning designs.'}",https://arxiv.org/pdf/2109.05749
-Few-Shot Learning by Integrating Spatial and Frequency Representation,Xiangyu Chen,"Human beings can recognize new objects with only a few labeled examples, however, few-shot learning remains a challenging problem for machine learning systems. Most previous algorithms in few-shot learning only utilize spatial information of the images. In this paper, we propose to integrate the frequency information into the learning model to boost the discrimination ability of the system. We employ Discrete Cosine Transformation (DCT) to generate the frequency representation, then, integrate the features from both the spatial domain and frequency domain for classification. The proposed strategy and its effectiveness are validated with different backbones, datasets, and algorithms. Extensive experiments demonstrate that the frequency information is complementary to the spatial representations in few-shot classification. The classification accuracy is boosted significantly by integrating features from both the spatial and frequency domains in different few-shot learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to integrate the frequency information into the learning model to boost the discrimination ability of the system, employing Discrete Cosine Transformation to generate the frequency representation and integrating the features from both the spatial domain and frequency domain for classification.'}",https://arxiv.org/pdf/2105.05348
-Few-Shot Learning for Road Object Detection,Anay Majee,"Few-shot learning is a problem of high interest in the evolution of deep learning. In this work, we consider the problem of few-shot object detection (FSOD) in a real-world, class-imbalanced scenario. For our experiments, we utilize the India Driving Dataset (IDD), as it includes a class of less-occurring road objects in the image dataset and hence provides a setup suitable for few-shot learning. We evaluate both metric-learning and meta-learning based FSOD methods, in two experimental settings: (i) representative (same-domain) splits from IDD, that evaluates the ability of a model to learn in the context of road images, and (ii) object classes with less-occurring object samples, similar to the open-set setting in real-world. From our experiments, we demonstrate that the metric-learning method outperforms meta-learning on the novel classes by (i) 11.2 mAP points on the same domain, and (ii) 1.0 mAP point on the open-set. We also show that our extension of object classes in a real-world open dataset offers a rich ground for few-shot learning studies.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the problem of few-shot object detection (FSOD) in a real-world, class-imbalanced scenario, and demonstrates that the metric-learning method outperforms meta-learning on the novel classes by 11.2 mAP points on the same domain, and 1.0 mAP point on the open-set.'}",
-Learning to Adapt With Memory for Probabilistic Few-Shot Learning,Lei Zhang,"Few-shot learning has recently generated increasing popularity in machine learning, which addresses the fundamental yet challenging problem of learning to adapt to new tasks with the limited data. In this paper, we propose a new probabilistic framework that learns to fast adapt with external memory. We model the classifier parameters as distributions that are inferred from the support set and directly applied to the query set for prediction. The model is optimized by formulating as a variational inference problem. The probabilistic modeling enables better handling prediction uncertainty due to the limited data. We impose a discriminative constraint on the feature representations by exploring the class structure, which can improve the classification performance. We further introduce a memory unit to store task-specific information extracted from the support set and used for the query set to achieve explicit adaption to individual tasks. By episodic training, the model learns to acquire the capability of adapting to specific tasks, which guarantees its performance on new related tasks. We conduct extensive experiments on widely-used benchmarks for few-shot recognition. Our method achieves new state-of-the-art performance and largely surpassing previous methods by large margins. The ablation study further demonstrates the effectiveness of the proposed discriminative learning and memory unit.","{'model': 'tldr@v2.0.0', 'text': 'A new probabilistic framework that learns to fast adapt with external memory and introduces a memory unit to store task-specific information extracted from the support set and used for the query set to achieve explicit adaption to individual tasks.'}",
-Who Calls The Shots? Rethinking Few-Shot Learning for Audio,Yu Wang,"Few-shot learning aims to train models that can recognize novel classes given just a handful of labeled examples, known as the support set. While the field has seen notable advances in recent years, they have often focused on multi-class image classification. Audio, in contrast, is often multi-label due to overlapping sounds, resulting in unique properties such as polyphony and signal-to-noise ratios (SNR). This leads to unanswered questions concerning the impact such audio properties may have on few-shot learning system design, performance, and human-computer interaction, as it is typically up to the user to collect and provide inference-time support set examples. We address these questions through a series of experiments designed to elucidate the answers to these questions. We introduce two novel datasets, FSD-MIX-CLIPS and FSD-MIX-SED, whose programmatic generation allows us to explore these questions systematically. Our experiments lead to audio-specific insights on few-shot learning, some of which are at odds with recent findings in the image domain: there is no best one-size- fits-all model, method, and support set selection criterion. Rather, it depends on the expected application scenario. Our code and data are available at https://github.com/wangyu/rethink-audio-fsl.","{'model': 'tldr@v2.0.0', 'text': 'A series of experiments lead to audio-specific insights on few-shot learning, some of which are at odds with recent findings in the image domain: there is no best one-size- fits-all model, method, and support set selection criterion, and it depends on the expected application scenario.'}",https://arxiv.org/pdf/2110.09600
-Tensor feature hallucination for few-shot learning,Michalis Lazarou,"Few-shot learning addresses the challenge of learning how to address novel tasks given not just limited supervision but limited data as well. An attractive solution is synthetic data generation. However, most such methods are overly sophisticated, focusing on high-quality, realistic data in the input space. It is unclear whether adapting them to the few-shot regime and using them for the downstream task of classification is the right approach. Previous works on synthetic data generation for few-shot classification focus on exploiting complex models, e.g. a Wasserstein GAN with multiple regularizers or a network that transfers latent diversities from known to novel classes.We follow a different approach and investigate how a simple and straightforward synthetic data generation method can be used effectively. We make two contributions, namely we show that: (1) using a simple loss function is more than enough for training a feature generator in the few-shot setting; and (2) learning to generate tensor features instead of vector features is superior. Extensive experiments on miniImagenet, CUB and CIFAR-FS datasets show that our method sets a new state of the art, outperforming more sophisticated few-shot data augmentation methods. The source code can be found at https://github.com/MichalisLazarou/TFH_fewshot.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that: (1) using a simple loss function is more than enough for training a feature generator in the few-shot setting; and (2) learning to generate tensor features instead of vector features is superior.'}",https://arxiv.org/pdf/2106.05321
-Unsupervised meta-learning for few-shot learning,Hui Xu,,"{'model': 'tldr@v2.0.0', 'text': 'An unsupervised meta-learning algorithm that learns from an unlabeled dataset and adapts to downstream human-specific tasks with few labeled data and constructs tasks using clustering embedding methods and data augmentation functions to satisfy two critical class distinction requirements is proposed.'}",
-LaSO: Label-Set Operations Networks for Multi-Label Few-Shot Learning,Amit Alfassy,"Example synthesis is one of the leading methods to tackle the problem of few-shot learning, where only a small number of samples per class are available. However, current synthesis approaches only address the scenario of a single category label per image. In this work, we propose a novel technique for synthesizing samples with multiple labels for the (yet unhandled) multi-label few-shot classification scenario. We propose to combine pairs of given examples in feature space, so that the resulting synthesized feature vectors will correspond to examples whose label sets are obtained through certain set operations on the label sets of the corresponding input pairs. Thus, our method is capable of producing a sample containing the intersection, union or set-difference of labels present in two input samples. As we show, these set operations generalize to labels unseen during training. This enables performing augmentation on examples of novel categories, thus, facilitating multi-label few-shot classifier learning. We conduct numerous experiments showing promising results for the label-set manipulation capabilities of the proposed approach, both directly (using the classification and retrieval metrics), and in the context of performing data augmentation for multi-label few-shot learning. We propose a benchmark for this new and challenging task and show that our method compares favorably to all the common baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel technique for synthesizing samples with multiple labels for the (yet unhandled) multi-label few-shot classification scenario, and proposes to combine pairs of given examples in feature space so that the resulting synthesized feature vectors will correspond to examples whose label sets are obtained through certain set operations on the label sets of the corresponding input pairs.'}",http://arxiv.org/pdf/1902.09811
-An Ensemble of Epoch-Wise Empirical Bayes for Few-Shot Learning,Yaoyao Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to meta-learn the ensemble of epoch-wise empirical Bayes models (E3BM) to achieve robust predictions and introduces four kinds of hyperprior learners by considering inductive vs. transductive, and epoch-dependent vs. epoch-independent, in the paradigm of meta-learning.'}",
-Few-Shot Learning With Global Class Representations,Tiange Luo,"In this paper, we propose to tackle the challenging few-shot learning (FSL) problem by learning global class representations using both base and novel class training samples. In each training episode, an episodic class mean computed from a support set is registered with the global representation via a registration module. This produces a registered global class representation for computing the classification loss using a query set. Though following a similar episodic training pipeline as existing meta learning based approaches, our method differs significantly in that novel class training samples are involved in the training from the beginning. To compensate for the lack of novel class training samples, an effective sample synthesis strategy is developed to avoid overfitting. Importantly, by joint base-novel class training, our approach can be easily extended to a more practical yet challenging FSL setting, i.e., generalized FSL, where the label space of test data is extended to both base and novel classes. Extensive experiments show that our approach is effective for both of the two FSL settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to tackle the challenging few-shot learning (FSL) problem by learning global class representations using both base and novel class training samples, and an effective sample synthesis strategy is developed to avoid overfitting.'}",https://arxiv.org/pdf/1908.05257
-Meta-Learning of Neural Architectures for Few-Shot Learning,T. Elsken,"The recent progress in neural architecture search (NAS) has allowed scaling the automated design of neural architectures to real-world domains, such as object detection and semantic segmentation. However, one prerequisite for the application of NAS are large amounts of labeled data and compute resources. This renders its application challenging in few-shot learning scenarios, where many related tasks need to be learned, each with limited amounts of data and compute time. Thus, few-shot learning is typically done with a fixed neural architecture. To improve upon this, we propose MetaNAS, the first method which fully integrates NAS with gradient-based meta-learning. MetaNAS optimizes a meta-architecture along with the meta-weights during meta-training. During meta-testing, architectures can be adapted to a novel task with a few steps of the task optimizer, that is: task adaptation becomes computationally cheap and requires only little data per task. Moreover, MetaNAS is agnostic in that it can be used with arbitrary model-agnostic meta-learning algorithms and arbitrary gradient-based NAS methods. Empirical results on standard few-shot classification benchmarks show that MetaNAS with a combination of DARTS and REPTILE yields state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MetaNAS is the first method which fully integrates NAS with gradient-based meta-learning, and optimizes a meta-architecture along with the meta-weights during meta-training, which can be adapted to a novel task with a few steps of the task optimizer.'}",https://arxiv.org/pdf/1911.11090
-"On Episodes, Prototypical Networks, and Few-shot Learning",Steinar Laenen,"Episodic learning is a popular practice among researchers and practitioners interested in few-shot learning. It consists of organising training in a series of learning problems, each relying on small ""support"" and ""query"" sets to mimic the few-shot circumstances encountered during evaluation. In this paper, we investigate the usefulness of episodic learning in Prototypical Networks and Matching Networks, two of the most popular algorithms making use of this practice. Surprisingly, in our experiments we found that, for Prototypical and Matching Networks, it is detrimental to use the episodic learning strategy of separating training samples between support and query set, as it is a data-inefficient way to exploit training batches. These ""non-episodic"" variants, which are closely related to the classic Neighbourhood Component Analysis, reliably improve over their episodic counterparts in multiple datasets, achieving an accuracy that (in the case of Prototypical Networks) is competitive with the state-of-the-art, despite being extremely simple.","{'model': 'tldr@v2.0.0', 'text': 'Surprisingly, it is detrimental to use the episodic learning strategy of separating training samples between support and query set, as it is a data-inefficient way to exploit training batches in Prototypical Networks and Matching Networks.'}",
-CSI-Based Human Activity Recognition With Graph Few-Shot Learning,Yong Zhang,"Human activity recognition (HAR) based on channel state information (CSI) plays an increasingly important role in the research of human–computer interaction. Many CSI HAR models based on traditional machine learning methods and deep learning methods have encountered two challenges. A lot of CSI activity data is needed to train the HAR models, which is time consuming. When the indoor environment or scene changes, the recognition accuracy of the model drops significantly, so it is necessary to recollect data to train the model. The existing few-shot learning-based method can solve the above problems to some extent, but when there are more kinds of new activities or fewer shots, the recognition accuracy will decrease significantly. In this article, considering the relationship between various activity data, a graph-based few-shot learning method with dual attention mechanism (CSI-GDAM) is proposed to perform CSI-based HAR. The model uses a feature extraction layer, including the convolutional block attention module (CBAM), to extract activity-related information in CSI data. The difference and inner product of the feature vector of the CSI activity samples are used to realize the graph convolutional network with a graph attention mechanism. The experiments proved that under the learning task of recognizing new activities in the new environment, the recognition accuracy rates reached 99.74% and 98.42% in the 5-way 5-shot and 5-way 1-shot cases, respectively. The proposed method is also compared with other few-shot learning and transfer learning methods.","{'model': 'tldr@v2.0.0', 'text': 'A graph-based few-shot learning method with dual attention mechanism (CSI-GDAM) is proposed to perform CSI-based HAR, which uses a feature extraction layer, including the convolutional block attention module (CBAM), to extract activity-related information in CSI data.'}",
-Parameterless Transductive Feature Re-representation for Few-Shot Learning,Wentao Cui,"Recent literature in few-shot learning (FSL) has shown that transductive methods often outperform their inductive counterparts. However, most transductive solutions, particularly the meta-learning based ones, require inserting trainable parameters on top of some inductive baselines to facilitate transduction. In this paper, we propose a parameterless transductive feature re-representation framework that differs from all existing solutions from the following perspectives. (1) It is widely compatible with existing FSL methods, including meta-learning and fine tuning based models. (2) The framework is simple and introduces no extra training parameters when applied to any architecture. We conduct experiments on three benchmark datasets by applying the framework to both representative meta-learning baselines and state-of-the-art FSL methods. Our framework consistently improves performances in all experiments and refreshes the state-of-the-art FSL results.","{'model': 'tldr@v2.0.0', 'text': 'A parameterless transductive feature re-representation framework that is widely compatible with existing FSL methods, including meta-learning and tuning based models and refreshes the state-of-the-art FSL results.'}",
-Transductive Information Maximization For Few-Shot Learning,Malik Boudiaf,"We introduce Transductive Infomation Maximization (TIM) for few-shot learning. Our method maximizes the mutual information between the query features and their label predictions for a given few-shot task, in conjunction with a supervision loss based on the support set. Furthermore, we propose a new alternating-direction solver for our mutual-information loss, which substantially speeds up transductive-inference convergence over gradient-based optimization, while yielding similar accuracy. TIM inference is modular: it can be used on top of any base-training feature extractor. Following standard transductive few-shot settings, our comprehensive experiments demonstrate that TIM outperforms state-of-the-art methods significantly across various datasets and networks, while used on top of a fixed feature extractor trained with simple cross-entropy on the base classes, without resorting to complex meta-learning schemes. It consistently brings between 2% and 5% improvement in accuracy over the best performing method, not only on all the well-established few-shot benchmarks but also on more challenging scenarios,with domain shifts and larger numbers of classes.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Transductive Infomation Maximization (TIM) for few-shot learning, and proposes a new alternating-direction solver for the mutual-information loss, which substantially speeds up transductive-inference convergence over gradient-based optimization, while yielding similar accuracy.'}",
-Self-supervised Knowledge Distillation for Few-shot Learning,Jathushan Rajasegaran,"Real-world contains an overwhelmingly large number of object classes, learning all of which at once is infeasible. Few shot learning is a promising learning paradigm due to its ability to learn out of order distributions quickly with only a few samples. Recent works [7, 41] show that simply learning a good feature embedding can outperform more sophisticated meta-learning and metric learning algorithms for few-shot learning. In this paper, we propose a simple approach to improve the representation capacity of deep neural networks for few-shot learning tasks. We follow a two-stage learning process: First, we train a neural network to maximize the entropy of the feature embedding, thus creating an optimal output manifold using a self-supervised auxiliary loss. In the second stage, we minimize the entropy on feature embedding by bringing self-supervised twins together, while constraining the manifold with student-teacher distillation. Our experiments show that, even in the first stage, self-supervision can outperform current state-of-the-art methods, with further gains achieved by our second stage distillation process. Our codes are available at: this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple approach to improve the representation capacity of deep neural networks for few-shot learning tasks, and shows that, even in the first stage, self-supervision can outperform current state-of-the-art methods.'}",
-Extensively Matching for Few-shot Learning Event Detection,Viet Dac Lai,"Current event detection models under supervised learning settings fail to transfer to new event types. Few-shot learning has not been explored in event detection even though it allows a model to perform well with high generalization on new event types. In this work, we formulate event detection as a few-shot learning problem to enable to extend event detection to new event types. We propose two novel loss factors that matching examples in the support set to provide more training signals to the model. Moreover, these training signals can be applied in many metric-based few-shot learning models. Our extensive experiments on the ACE-2005 dataset (under a few-shot learning setting) show that the proposed method can improve the performance of few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work formulate event detection as a few-shot learning problem to enable to extend event detection to new event types and proposes two novel loss factors that matching examples in the support set to provide more training signals to the model.'}",https://www.aclweb.org/anthology/2020.nuse-1.5.pdf
-Self-Supervised Few-Shot Learning on Point Clouds,Charu Sharma,"The increased availability of massive point clouds coupled with their utility in a wide variety of applications such as robotics, shape synthesis, and self-driving cars has attracted increased attention from both industry and academia. Recently, deep neural networks operating on labeled point clouds have shown promising results on supervised learning tasks like classification and segmentation. However, supervised learning leads to the cumbersome task of annotating the point clouds. To combat this problem, we propose two novel self-supervised pre-training tasks that encode a hierarchical partitioning of the point clouds using a cover-tree, where point cloud subsets lie within balls of varying radii at each level of the cover-tree. Furthermore, our self-supervised learning network is restricted to pre-train on the support set (comprising of scarce training examples) used to train the downstream network in a few-shot learning (FSL) setting. Finally, the fully-trained self-supervised network's point embeddings are input to the downstream task's network. We present a comprehensive empirical evaluation of our method on both downstream classification and segmentation tasks and show that supervised methods pre-trained with our self-supervised learning method significantly improve the accuracy of state-of-the-art methods. Additionally, our method also outperforms previous unsupervised methods in downstream classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive empirical evaluation of the self-supervised learning method is presented and it is shown that supervised methods pre-trained with the method significantly improve the accuracy of state-of-the-art methods and also outperforms previous unsupervised methods in downstream classification tasks.'}",
-Model-Agnostic Boundary-Adversarial Sampling for Test-Time Generalization in Few-Shot Learning,Jaekyeom Kim,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a modelagnostic method that improves the test-time performance of any few-shot learning models with no additional training, and thus is free from the training-test domain gap.'}",
-Prototype Completion with Primitive Knowledge for Few-Shot Learning,Baoquan Zhang,"Few-shot learning is a challenging task, which aims to learn a classifier for novel classes with few examples. Pre-training based meta-learning methods effectively tackle the problem by pre-training a feature extractor and then fine-tuning it through the nearest centroid based meta-learning. However, results show that the fine-tuning step makes very marginal improvements. In this paper, 1) we figure out the key reason, i.e., in the pre-trained feature space, the base classes already form compact clusters while novel classes spread as groups with large variances, which implies that fine-tuning the feature extractor is less meaningful; 2) instead of fine-tuning the feature extractor, we focus on estimating more representative prototypes during meta-learning. Consequently, we propose a novel prototype completion based meta-learning framework. This framework first introduces primitive knowledge (i.e., class-level part or attribute annotations) and extracts representative attribute features as priors. Then, we design a prototype completion network to learn to complete prototypes with these priors. To avoid the prototype completion error caused by primitive knowledge noises or class differences, we further develop a Gaussian based prototype fusion strategy that combines the mean-based and completed prototypes by exploiting the unlabeled samples. Extensive experiments show that our method: (i) can obtain more accurate prototypes; (ii) out-performs state-of-the-art techniques by 2%~9% in terms of classification accuracy. Our code is available online1.","{'model': 'tldr@v2.0.0', 'text': 'A novel prototype completion based meta-learning framework that first introduces primitive knowledge and extracts representative attribute features as priors, and designs a prototype completion network to learn to complete prototypes with these priors.'}",https://arxiv.org/pdf/2009.04960
-XtarNet: Learning to Extract Task-Adaptive Representation for Incremental Few-Shot Learning,Sung Whan Yoon,"Learning novel concepts while preserving prior knowledge is a long-standing challenge in machine learning. The challenge gets greater when a novel task is given with only a few labeled examples, a problem known as incremental few-shot learning. We propose XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning. The method utilizes a backbone network pretrained on a set of base categories while also employing additional modules that are meta-trained across episodes. Given a new task, the novel feature extracted from the meta-trained modules is mixed with the base feature obtained from the pretrained model. The process of combining two different features provides TAR and is also controlled by meta-trained modules. The TAR contains effective information for classifying both novel and base categories. The base and novel classifiers quickly adapt to a given task by utilizing the TAR. Experiments on standard image datasets indicate that XtarNet achieves state-of-the-art incremental few-shot learning performance. The concept of TAR can also be used in conjunction with existing incremental few-shot learning methods; extensive simulation results in fact show that applying TAR enhances the known methods significantly.","{'model': 'tldr@v2.0.0', 'text': 'XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning, utilizes a backbone network pretrained on a set of base categories while also employing additional modules that are meta-trained across episodes.'}",
-Transductive Relation-Propagation With Decoupling Training for Few-Shot Learning,Yuqing Ma,"Few-shot learning, aiming to learn novel concepts from one or a few labeled examples, is an interesting and very challenging problem with many practical advantages. Existing few-shot methods usually utilize data of the same classes to train the feature embedding module and in a row, which is unable to learn adapting to new tasks. Besides, traditional few-shot models fail to take advantage of the valuable relations of the support-query pairs, leading to performance degradation. In this article, we propose a transductive relation-propagation graph neural network (GNN) with a decoupling training strategy (TRPN-D) to explicitly model and propagate such relations across support-query pairs, and empower the few-shot module the ability of transferring past knowledge to new tasks via the decoupling training. Our few-shot module, namely TRPN, treats the relation of each support-query pair as a graph node, named relational node, and resorts to the known relations between support samples, including both intraclass commonality and interclass uniqueness. Through relation propagation, the model could generate the discriminative relation embeddings for support-query pairs. To the best of our knowledge, this is the first work that decouples the training of the embedding network and the few-shot graph module with different tasks, which might offer a new way to solve the few-shot learning problem. Extensive experiments conducted on several benchmark datasets demonstrate that our method can significantly outperform a variety of state-of-the-art few-shot learning methods.","{'model': 'tldr@v2.0.0', 'text': 'A transductive relation-propagation graph neural network (GNN) with a decoupling training strategy (TRPN-D) to explicitly model and propagate such relations across support-query pairs, and empower the few-shot module the ability of transferring past knowledge to new tasks via the decoupled training.'}",
-Few-Shot Learning with Metric-Agnostic Conditional Embeddings,Nathan Hilliard,"Learning high quality class representations from few examples is a key problem in metric-learning approaches to few-shot learning. To accomplish this, we introduce a novel architecture where class representations are conditioned for each few-shot trial based on a target image. We also deviate from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison. This allows the network to decide what aspects of each class are important for the comparison at hand. We find that this flexible architecture works well in practice, achieving state-of-the-art performance on the Caltech-UCSD birds fine-grained classification task.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel architecture where class representations are conditioned for each few-shot trial based on a target image, and deviates from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison.'}",
-Multimodality Helps Unimodality: Cross-Modal Few-Shot Learning with Multimodal Models,Zhiqiu Lin,"The ability to quickly learn a new task with minimal instruction - known as few-shot learning - is a central aspect of intelligent agents. Classical few-shot benchmarks make use of few-shot samples from a single modality, but such samples may not be sufficient to characterize an entire concept class. In contrast, humans use cross-modal information to learn new concepts efficiently. In this work, we demonstrate that one can indeed build a better visual dog classifier by reading about dogs and listening to them bark. To do so, we exploit the fact that recent multimodal foundation models such as CLIP are inherently cross-modal, mapping different modalities to the same representation space. Specifically, we propose a simple cross-modal adaptation approach that learns from few-shot examples spanning different modalities. By repurposing class names as additional one-shot training samples, we achieve SOTA results with an embarrassingly simple linear classifier for vision-language adaptation. Furthermore, we show that our approach can benefit existing methods such as prefix tuning, adapters, and classifier ensembling. Finally, to explore other modalities beyond vision and language, we construct the first (to our knowledge) audiovisual few-shot benchmark and use cross-modal training to improve the performance of both image and audio classification. Project site at link.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple cross-modal adaptation approach that learns from few-shot examples spanning different modalities, and achieves SOTA results with an embarrassingly simple linear classifier for vision-language adaptation.'}",https://arxiv.org/pdf/2301.06267
-Multimodal Prototypical Networks for Few-shot Learning,Frederik Pahde,"Although providing exceptional results for many computer vision tasks, state-of-the-art deep learning algorithms catastrophically struggle in low data scenarios. However, if data in additional modalities exist (e.g. text) this can compensate for the lack of data and improve the classification results. To overcome this data scarcity, we design a cross-modal feature generation framework capable of enriching the low populated embedding space in few-shot scenarios, leveraging data from the auxiliary modality. Specifically, we train a generative model that maps text data into the visual feature space to obtain more reliable prototypes. This allows to exploit data from additional modalities (e.g. text) during training while the ultimate task at test time remains classification with exclusively visual data. We show that in such cases nearest neighbor classification is a viable approach and outperform state-of-the-art single-modal and multimodal few-shot learning methods on the CUB-200 and Oxford-102 datasets.","{'model': 'tldr@v2.0.0', 'text': 'A generative model is trained that maps text data into the visual feature space to obtain more reliable prototypes and shows that in such cases nearest neighbor classification is a viable approach and outperform state-of-the-art single-modal and multimodal few-shot learning methods on the CUB-200 and Oxford-102 datasets.'}",http://arxiv.org/pdf/2011.08899
-Scaling ASR Improves Zero and Few Shot Learning,Alex Xiao,"With 4.5 million hours of English speech from 10 different sources across 120 countries and models of up to 10 billion parameters, we explore the frontiers of scale for automatic speech recognition. We propose data selection techniques to efficiently scale training data to find the most valuable samples in massive datasets. To efficiently scale model sizes, we leverage various optimizations such as sparse transducer loss and model sharding. By training 1-10B parameter universal English ASR models, we push the limits of speech recognition performance across many domains. Furthermore, our models learn powerful speech representations with zero and few-shot capabilities on novel domains and styles of speech, exceeding previous results across multiple in-house and public benchmarks. For speakers with disorders due to brain damage, our best zero-shot and few-shot models achieve 22% and 60% relative improvement on the AphasiaBank test set, respectively, while realizing the best performance on public social media videos. Furthermore, the same universal model reaches equivalent performance with 500x less in-domain data on the SPGISpeech financial-domain dataset.","{'model': 'tldr@v2.0.0', 'text': 'By training 1-10B parameter universal English ASR models, this work pushes the limits of speech recognition performance across many domains and proposes data selection techniques to efficiently scale training data to find the most valuable samples in massive datasets.'}",http://arxiv.org/pdf/2111.05948
-Few-Shot Learning for Decoding Surface Electromyography for Hand Gesture Recognition,E. Rahimian,"This work is motivated by the recent advancements of Deep Neural Networks (DNNs) for myoelectric prosthesis control. In this regard, hand gesture recognition via surface Electromyogram (sEMG) signals has shown a high potential for improving the performance of myoelectric control prostheses. Although the recent researches in hand gesture recognition with DNNs have achieved promising results, they are still in their infancy. The recent literature uses traditional supervised learning methods that usually have poor performance if a small amount of data is available or requires adaptation to a changing task. Therefore, in this work, we develop a novel hand gesture recognition framework based on the formulation of FewShot Learning (FSL) to infer the required output given only one or a few numbers of training examples. Thus in this paper, we proposed a new architecture (named as FHGR which refers to ""Few-shot Hand Gesture Recognition"") that learns the mapping using a small number of data and quickly adapts to a new user/gesture by combing its prior experience. The proposed approach led to 83.99% classification accuracy on new repetitions with few-shot observations, 76.39% accuracy on new subjects with few-shot observations, and 72.19% accuracy on new gestures with few-shot observations.","{'model': 'tldr@v2.0.0', 'text': 'A novel hand gesture recognition framework based on the formulation of FewShot Learning (FSL), that learns the mapping using a small number of data and quickly adapts to a new user/gesture by combing its prior experience is proposed.'}",
-The unreasonable effectiveness of few-shot learning for machine translation,Xavier García,"We demonstrate the potential of few-shot translation systems, trained with unpaired language data, for both high and low-resource language pairs. We show that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able to match specialized supervised state-of-the-art models as well as more general commercial translation systems. In particular, we outperform the best performing system on the WMT'21 English - Chinese news translation task by only using five examples of English - Chinese parallel data at inference. Moreover, our approach in building these models does not necessitate joint multilingual training or back-translation, is conceptually simple and shows the potential to extend to the multilingual setting. Furthermore, the resulting models are two orders of magnitude smaller than state-of-the-art language models. We then analyze the factors which impact the performance of few-shot translation systems, and highlight that the quality of the few-shot demonstrations heavily determines the quality of the translations generated by our models. Finally, we show that the few-shot paradigm also provides a way to control certain attributes of the translation -- we show that we are able to control for regional varieties and formality using only a five examples at inference, paving the way towards controllable machine translation systems.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able to match specialized supervised state-of-the-art models as well as more general commercial translation systems.'}",http://arxiv.org/pdf/2302.01398
-Task Cooperation for Semi-Supervised Few-Shot Learning,Han Ye,"Training a model with limited data is an essential task for machine learning and visual recognition. Few-shot learning approaches meta-learn a task-level inductive bias from SEEN class few-shot tasks, and the meta-model is expected to facilitate the few-shot learning with UNSEEN classes. Inspired by the idea that unlabeled data can be utilized to smooth the model space in traditional semi-supervised learning, we propose TAsk COoperation (TACO) which takes advantage of unsupervised tasks to smooth the meta-model space. Specifically, we couple the labeled support set in a few-shot task with easily-collected unlabeled instances, prediction agreement on which encodes the relationship between tasks. The learned smooth meta-model promotes the generalization ability on supervised UNSEEN few-shot tasks. The state-of-the-art few-shot classification results on MiniImageNet and TieredImageNet verify the superiority of TACO to leverage unlabeled data and task relationship in meta-learning.","{'model': 'tldr@v2.0.0', 'text': 'This work couple the labeled support set in a few-shot task with easily-collected unlabeled instances, prediction agreement on which encodes the relationship between tasks, and learns smooth meta-model which promotes the generalization ability on supervised UNSEEN few- shot tasks.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17277/17084
-Augmenting Few-Shot Learning With Supervised Contrastive Learning,Taemin Lee,"Few-shot learning deals with a small amount of data which incurs insufficient performance with conventional cross-entropy loss. We propose a pretraining approach for few-shot learning scenarios. That is, considering that the feature extractor quality is a critical factor in few-shot learning, we augment the feature extractor using a contrastive learning technique. It is reported that supervised contrastive learning applied to base class training in transductive few-shot training pipeline leads to improved results, outperforming the state-of-the-art methods on Mini-ImageNet and CUB. Furthermore, our experiment shows that a much larger dataset is needed to retain few-shot classification accuracy when domain-shift degradation exists, and if our method is applied, the need for a large dataset is eliminated. The accuracy gain can be translated to a runtime reduction of $3.87\times $ in a resource-constrained environment.","{'model': 'tldr@v2.0.0', 'text': 'It is reported that supervised contrastive learning applied to base class training in transductive few-shot training pipeline leads to improved results, outperforming the state-of-the-art methods on Mini-ImageNet and CUB.'}",https://ieeexplore.ieee.org/ielx7/6287639/9312710/09409075.pdf
-Consistent Meta-Regularization for Better Meta-Knowledge in Few-Shot Learning,Pinzhuo Tian,"Recently, meta-learning provides a powerful paradigm to deal with the few-shot learning problem. However, existing meta-learning approaches ignore the prior fact that good meta-knowledge should alleviate the data inconsistency between training and test data, caused by the extremely limited data, in each few-shot learning task. Moreover, legitimately utilizing the prior understanding of meta-knowledge can lead us to design an efficient method to improve the meta-learning model. Under this circumstance, we consider the data inconsistency from the distribution perspective, making it convenient to bring in the prior fact, and propose a new consistent meta-regularization (Con-MetaReg) to help the meta-learning model learn how to reduce the data-distribution discrepancy between the training and test data. In this way, the ability of meta-knowledge on keeping the training and test data consistent is enhanced, and the performance of the meta-learning model can be further improved. The extensive analyses and experiments demonstrate that our method can indeed improve the performances of different meta-learning models in few-shot regression, classification, and fine-grained classification.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the data inconsistency from the distribution perspective, making it convenient to bring in the prior fact, and proposes a new consistent meta-regularization (Con-MetaReg) to help the meta-learning model learn how to reduce the data-distribution discrepancy between the training and test data.'}",
-Few-Shot Learning With Class Imbalance,Mateusz Ochal,"Few-shot learning (FSL) algorithms are commonly trained through meta-learning (ML), which exposes models to batches of tasks sampled from a meta-dataset to mimic tasks seen during evaluation. However, the standard training procedures overlook the real-world dynamics where classes commonly occur at different frequencies. While it is generally understood that class imbalance harms the performance of supervised methods, limited research examines the impact of imbalance on the FSL evaluation task. Our analysis compares ten state-of-the-art ML and FSL methods on different imbalance distributions and rebalancing techniques. Our results reveal that: 1) some FSL methods display a natural disposition against imbalance while most other approaches produce a performance drop by up to 17% compared to the balanced task without the appropriate mitigation; 2) many ML algorithms will not automatically learn to balance from exposure to imbalanced training tasks; 3) classical rebalancing strategies, such as random oversampling, can still be very effective, leading to state-of-the-art performances and should not be overlooked.","{'model': 'tldr@v2.0.0', 'text': 'This analysis compares ten state-of-the-art ML and F SL methods on different imbalance distributions and rebalancing techniques and reveals that some FSL methods display a natural disposition against imbalance while most other approaches produce a performance drop compared to the balanced task without the appropriate mitigation.'}",https://arxiv.org/pdf/2101.02523
-Prototypical Siamese Networks for Few-shot Learning,Junhua Wang,"We propose a novel architecture, called Prototypical Siamese Networks, for few-shot learning, where a classifier must generalize to new classes not seen in the training set, given only a few examples of each class. Prototypical Siamese Networks add a new module to siamese networks to learn a high quality prototypical representation of each class. Compared to recent methods for few-shot learning, our method achieves state-of-the-art performance on few-shot learning. Experiments on two benchmarks validate the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel architecture, called Prototypical Siamese Networks, for few-shot learning, where a classifier must generalize to new classes not seen in the training set, given only a few examples of each class.'}",
-FS-HGR: Few-Shot Learning for Hand Gesture Recognition via Electromyography,E. Rahimian,"This work is motivated by the recent advances in Deep Neural Networks (DNNs) and their widespread applications in human-machine interfaces. DNNs have been recently used for detecting the intended hand gesture through the processing of surface electromyogram (sEMG) signals. Objective: Although DNNs have shown superior accuracy compared to conventional methods when large amounts of data are available for training, their performance substantially decreases when data are limited. Collecting large datasets for training may be feasible in research laboratories, but it is not a practical approach for real-life applications. The main objective of this work is to design a modern DNN-based gesture detection model that relies on minimal training data while providing high accuracy. Methods: We propose the novel Few-Shot learning- Hand Gesture Recognition (FS-HGR) architecture. Few-shot learning is a variant of domain adaptation with the goal of inferring the required output based on just one or a few training observations. The proposed FS-HGR generalizes after seeing very few observations from each class by combining temporal convolutions with attention mechanisms. This allows the meta-learner to aggregate contextual information from experience and to pinpoint specific pieces of information within its available set of inputs. Data Source & Summary of Results: The performance of FS-HGR was tested on the second and fifth Ninapro databases, referred to as the DB2 and DB5, respectively. The DB2 consists of 50 gestures (rest included) from 40 healthy subjects. The Ninapro DB5 contains data from 10 healthy participants performing a total of 53 different gestures (rest included). The proposed approach for the Ninapro DB2 led to 85.94% classification accuracy on new repetitions with few-shot observation (5-way 5-shot), 81.29% accuracy on new subjects with few-shot observation (5-way 5-shot), and 73.36% accuracy on new gestures with few-shot observation (5-way 5-shot). Moreover, the proposed approach for the Ninapro DB5 led to 64.65% classification accuracy on new subjects with few-shot observation (5-way 5-shot).","{'model': 'tldr@v2.0.0', 'text': 'The main objective of this work is to design a modern DNN-based gesture detection model that relies on minimal training data while providing high accuracy and is motivated by the recent advances in Deep Neural Networks and their widespread applications in human-machine interfaces.'}",
-Semi-supervised few-shot learning for medical image segmentation,A. Fayjie,"Recent years have witnessed the great progress of deep neural networks on semantic segmentation, particularly in medical imaging. Nevertheless, training high-performing models require large amounts of pixel-level ground truth masks, which can be prohibitive to obtain in the medical domain. Furthermore, training such models in a low-data regime highly increases the risk of overfitting. Recent attempts to alleviate the need for large annotated datasets have developed training strategies under the few-shot learning paradigm, which addresses this shortcoming by learning a novel class from only a few labeled examples. In this context, a segmentation model is trained on episodes, which represent different segmentation problems, each of them trained with a very small labeled dataset. In this work, we propose a novel few-shot learning framework for semantic segmentation, where unlabeled images are also made available at each episode. To handle this new learning paradigm, we propose to include surrogate tasks that can leverage very powerful supervisory signals --derived from the data itself-- for semantic feature learning. We show that including unlabeled surrogate tasks in the episodic training leads to more powerful feature representations, which ultimately results in better generability to unseen tasks. We demonstrate the efficiency of our method in the task of skin lesion segmentation in two publicly available datasets. Furthermore, our approach is general and model-agnostic, which can be combined with different deep architectures.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel few-shot learning framework for semantic segmentation, where unlabeled images are also made available at each episode, and proposes to include surrogate tasks that can leverage very powerful supervisory signals --derived from the data itself-- for semantic feature learning.'}",
-Domain-Adaptive Few-Shot Learning,An Zhao,"Existing few-shot learning (FSL) methods make the implicit assumption that the few target class samples are from the same domain as the source class samples. However, in practice, this assumption is often invalid –the target classes could come from a different domain. This poses an additional challenge of domain adaptation (DA) with few training samples. In this paper, the problem of domain-adaptive few-shot learning (DA-FSL) is tackled, which is expected to have wide use in real-world scenarios and requires solving FSL and DA in a unified framework. To this end, we propose a novel domain-adversarial prototypical network (DAPN) model. It is designed to address a specific challenge in DA-FSL: the DA objective means that the source and target data distributions need to be aligned, typically through a shared domain-adaptive feature embedding space; but the FSL objective dictates that the target domain per class distribution must be different from that of any source domain class, meaning aligning the distributions across domains may harm the FSL performance. How to achieve global domain distribution alignment whilst maintaining source/target per-class discriminativeness thus becomes the key. Our solution is to explicitly enhance the source/target per-class separation before domain-adaptive feature embedding learning, to alleviate the negative effect of domain alignment on FSL. Extensive experiments show that our DAPN outperforms the state-of-the-arts. The code is available at https://github.com/dingmyu/DAPN.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel domain-adversarial prototypical network (DAPN) model, designed to address a specific challenge in DA-FSL, which dictates that the target domain per class distribution must be different from that of any source domain class, meaning aligning the distributions across domains may harm the FSL performance.'}",https://arxiv.org/pdf/2003.08626
-Asymmetric Distribution Measure for Few-shot Learning,Wenbin Li,"The core idea of metric-based few-shot image classification is to directly measure the relations between query images and support classes to learn transferable feature embeddings. Previous work mainly focuses on image-level feature representations, which actually cannot effectively estimate a class's distribution due to the scarcity of samples. Some recent work shows that local descriptor based representations can achieve richer representations than image-level based representations. However, such works are still based on a less effective instance-level metric, especially a symmetric metric, to measure the relation between a query image and a support class. Given the natural asymmetric relation between a query image and a support class, we argue that an asymmetric measure is more suitable for metric-based few-shot learning. To that end, we propose a novel Asymmetric Distribution Measure (ADM) network for few-shot learning by calculating a joint local and global asymmetric measure between two multivariate local distributions of a query and a class. Moreover, a task-aware Contrastive Measure Strategy (CMS) is proposed to further enhance the measure function. On popular miniImageNet and tieredImageNet, ADM can achieve the state-of-the-art results, validating our innovative design of asymmetric distribution measures for few-shot learning. The source code can be downloaded from https://github.com/WenbinLee/ADM.git.","{'model': 'tldr@v2.0.0', 'text': 'A novel Asymmetric Distribution Measure (ADM) network for few-shot learning is proposed by calculating a joint local and global asymmetric measure between two multivariate local distributions of a query and a class.'}",https://www.ijcai.org/proceedings/2020/0409.pdf
-Few-Shot Learning Through an Information Retrieval Lens,Eleni Triantafillou,"Few-shot learning refers to understanding new concepts from only a few examples. We propose an information retrieval-inspired approach for this problem that is motivated by the increased importance of maximally leveraging all the available information in this low-data regime. We define a training objective that aims to extract as much information as possible from each training batch by effectively optimizing over all relative orderings of the batch points simultaneously. In particular, we view each batch point as a `query' that ranks the remaining ones based on its predicted relevance to them and we define a model within the framework of structured prediction to optimize mean Average Precision over these rankings. Our method achieves impressive results on the standard few-shot classification benchmarks while is also capable of few-shot retrieval.","{'model': 'tldr@v2.0.0', 'text': 'This work defines a training objective that aims to extract as much information as possible from each training batch by effectively optimizing over all relative orderings of the batch points simultaneously and defines a model within the framework of structured prediction to optimize mean Average Precision over these rankings.'}",
-Rethinking Class Relations: Absolute-relative Supervised and Unsupervised Few-shot Learning,Hongguang Zhang,"The majority of existing few-shot learning methods describe image relations with binary labels. However, such binary relations are insufficient to teach the network complicated real-world relations, due to the lack of decision smoothness. Furthermore, current few-shot learning models capture only the similarity via relation labels, but they are not exposed to class concepts associated with objects, which is likely detrimental to the classification performance due to underutilization of the available class labels. For instance, children learn the concept of tiger from a few of actual examples as well as from comparisons of tiger to other animals. Thus, we hypothesize that both similarity and class concept learning must be occurring simultaneously. With these observations at hand, we study the fundamental problem of simplistic class modeling in current few-shot learning methods. We rethink the relations between class concepts, and propose a novel Absolute-relative Learning paradigm to fully take advantage of label information to refine the image an relation representations in both supervised and unsupervised scenarios. Our proposed paradigm improves the performance of several state-of-the-art models on publicly available datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work rethink the relations between class concepts, and proposes a novel Absolute-relative Learning paradigm to fully take advantage of label information to refine the image an relation representations in both supervised and unsupervised scenarios.'}",https://arxiv.org/pdf/2001.03919
-Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning,Haokun Liu,"Few-shot in-context learning (ICL) enables pre-trained language models to perform a previously-unseen task without any gradient-based training by feeding a small number of training examples as part of the input. ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made. Parameter-efficient fine-tuning (PEFT) (e.g. adapter modules, prompt tuning, sparse update methods, etc.) offers an alternative paradigm where a small set of parameters are trained to enable a model to perform the new task. In this paper, we rigorously compare few-shot ICL and PEFT and demonstrate that the latter offers better accuracy as well as dramatically lower computational costs. Along the way, we introduce a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters. We also propose a simple recipe based on the T0 model called T-Few that can be applied to new tasks without task-specific tuning or modifications. We validate the effectiveness of T-Few on completely unseen tasks by applying it to the RAFT benchmark, attaining super-human performance for the first time and outperforming the state-of-the-art by 6% absolute. All of the code used in our experiments is publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper rigorously compares few-shot ICL and PEFT and demonstrates that the latter offers better accuracy as well as dramatically lower computational costs, and introduces a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters.'}",http://arxiv.org/pdf/2205.05638
-Exploring Effectiveness of GPT-3 in Grammatical Error Correction: A Study on Performance and Controllability in Prompt-Based Methods,Mengsay Loem,"Large-scale pre-trained language models such as GPT-3 have shown remarkable performance across various natural language processing tasks. However, applying prompt-based methods with GPT-3 for Grammatical Error Correction (GEC) tasks and their controllability remains underexplored. Controllability in GEC is crucial for real-world applications, particularly in educational settings, where the ability to tailor feedback according to learner levels and specific error types can significantly enhance the learning process.This paper investigates the performance and controllability of prompt-based methods with GPT-3 for GEC tasks using zero-shot and few-shot setting. We explore the impact of task instructions and examples on GPT-3’s output, focusing on controlling aspects such as minimal edits, fluency edits, and learner levels. Our findings demonstrate that GPT-3 could effectively perform GEC tasks, outperforming existing supervised and unsupervised approaches. We also showed that GPT-3 could achieve controllability when appropriate task instructions and examples are given.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the performance and controllability of prompt-based methods with GPT-3 for Grammatical Error Correction tasks using zero-shot and few-shot setting and demonstrates that G PT-3 could effectively perform GEC tasks, outperforming existing supervised and unsupervised approaches.'}",http://arxiv.org/pdf/2305.18156
-Knowledge Injected Prompt Based Fine-tuning for Multi-label Few-shot ICD Coding,Zhichao Yang,"Automatic International Classification of Diseases (ICD) coding aims to assign multiple ICD codes to a medical note with average length of 3,000+ tokens. This task is challenging due to a high-dimensional space of multi-label assignment (tens of thousands of ICD codes) and the long-tail challenge: only a few codes (common diseases) are frequently assigned while most codes (rare diseases) are infrequently assigned. This study addresses the long-tail challenge by adapting a prompt-based fine-tuning technique with label semantics, which has been shown to be effective under few-shot setting. To further enhance the performance in medical domain, we propose a knowledge-enhanced longformer by injecting three domain-specific knowledge: hierarchy, synonym, and abbreviation with additional pretraining using contrastive learning. Experiments on MIMIC-III-full, a benchmark dataset of code assignment, show that our proposed method outperforms previous state-of-the-art method in 14.5% in marco F1 (from 10.3 to 11.8, P<0.001). To further test our model on few-shot setting, we created a new rare diseases coding dataset, MIMIC-III-rare50, on which our model improves marco F1 from 17.1 to 30.4 and micro F1 from 17.2 to 32.6 compared to previous method.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based fine-tuning technique with label semantics, which has been shown to be effective under few-shot setting is adapted by adapting a knowledge-enhanced longformer by injecting three domain-specific knowledge: hierarchy, synonym, and abbreviation with additional pretraining using contrastive learning.'}",
-Improved Universal Sentence Embeddings with Prompt-based Contrastive Learning and Energy-based Learning,Yuxin Jiang,"Contrastive learning has been demonstrated to be effective in enhancing pre-trained language models (PLMs) to derive superior universal sentence embeddings. However, existing contrastive methods still have two limitations. Firstly, previous works may acquire poor performance under domain shift settings, thus hindering the application of sentence representations in practice. We attribute this low performance to the over-parameterization of PLMs with millions of parameters. To alleviate it, we propose PromCSE (Prompt-based Contrastive Learning for Sentence Embeddings), which only trains small-scale \emph{Soft Prompt} (i.e., a set of trainable vectors) while keeping PLMs fixed. Secondly, the commonly used NT-Xent loss function of contrastive learning does not fully exploit hard negatives in supervised learning settings. To this end, we propose to integrate an Energy-based Hinge loss to enhance the pairwise discriminative power, inspired by the connection between the NT-Xent loss and the Energy-based Learning paradigm. Empirical results on seven standard semantic textual similarity (STS) tasks and a domain-shifted STS task both show the effectiveness of our method compared with the current state-of-the-art sentence embedding models. Our code is publicly avaliable at https://github.com/YJiangcm/PromCSE","{'model': 'tldr@v2.0.0', 'text': 'PromCSE (Prompt-based Contrastive Learning for Sentence Embeddings), which only trains small-scale \\emph{Soft Prompt} (i.e., a set of trainable vectors) while keeping PLMs fixed, is proposed, which shows the effectiveness of the method compared with the current state-of-the-art sentence embedding models.'}",https://aclanthology.org/2022.findings-emnlp.220.pdf
-V2P: Vision-to-Prompt based Multi-Modal Product Summary Generation,Xuemeng Song,"Multi-modal Product Summary Generation is a new yet challenging task, which aims to generate a concise and readable summary for a product given its multi-modal content, e.g., its long text description and image. Although existing methods have achieved great success, they still suffer from three key limitations: 1) overlook the benefit of pre-training, 2) lack the representation-level supervision, and 3) ignore the diversity of the seller-generated data. To address these limitations, in this work, we propose a Vision-to-Prompt based multi-modal product summary generation framework, dubbed as V2P, where a Generative Pre-trained Language Model (GPLM) is adopted as the backbone. In particular, to maintain the original text capability of the GPLM and fully utilize the high-level concepts contained in the product image, we design V2P with two key components: vision-based prominent attribute prediction, and attribute prompt-guided summary generation. The first component works on obtaining the vital semantic attributes of the product from its image by the Swin Transformer, while the second component aims to generate the summary based on the product's long text description and the attribute prompts yielded by the first component with a GPLM. Towards comprehensive supervision over the second component, apart from the conventional output-level supervision, we introduce the representation-level regularization. Meanwhile, we design the data augmentation-based robustness regularization to handle the diverse inputs and improve the robustness of the second component. Extensive experiments on a large-scale Chinese dataset verify the superiority of our model over cutting-edge methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Vision-to-Prompt based multi-modal product summary generation framework, dubbed as V2P, where a Generative Pre-trained Language Model (GPLM) is adopted as the backbone, and introduces the representation-level regularization over the second component.'}",
-Do We Still Need Human Assessors? Prompt-Based GPT-3 User Simulation in Conversational AI,Selina Meyer,"Scarcity of user data continues to be a problem in research on conversational user interfaces and often hinders or slows down technical innovation. In the past, different ways of synthetically generating data, such as data augmentation techniques have been explored. With the rise of ever improving pre-trained language models, we ask if we can go beyond such methods by simply providing appropriate prompts to these general purpose models to generate data. We explore the feasibility and cost-benefit trade-offs of using non fine-tuned synthetic data to train classification algorithms for conversational agents. We compare this synthetically generated data with real user data and evaluate the performance of classifiers trained on different combinations of synthetic and real data. We come to the conclusion that, although classifiers trained on such synthetic data perform much better than random baselines, they do not compare to the performance of classifiers trained on even very small amounts of real user data, largely because such data is lacking much of the variability found in user generated data. Nevertheless, we show that in situations where very little data and resources are available, classifiers trained on such synthetically generated data might be preferable to the collection and annotation of naturalistic data.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that in situations where very little data and resources are available, classifiers trained on such synthetically generated data might be preferable to the collection and annotation of naturalistic data.'}",https://dl.acm.org/doi/pdf/10.1145/3543829.3544529
-Contrastive Learning for Prompt-based Few-shot Language Learners,Yiren Jian,"The impressive performance of GPT-3 using natural language prompts and in-context learning has inspired work on better fine-tuning of moderately-sized models under this paradigm. Following this line of work, we present a contrastive learning framework that clusters inputs from the same class for better generality of models trained with only limited examples. Specifically, we propose a supervised contrastive framework that clusters inputs from the same class under different augmented “views” and repel the ones from different classes. We create different “views” of an example by appending it with different language prompts and contextual demonstrations. Combining a contrastive loss with the standard masked language modeling (MLM) loss in prompt-based few-shot learners, the experimental results show that our method can improve over the state-of-the-art methods in a diverse set of 15 language tasks. Our framework makes minimal assumptions on the task or the base model, and can be applied to many recent methods with little modification.","{'model': 'tldr@v2.0.0', 'text': 'A supervised contrastive framework that clusters inputs from the same class under different augmented “views” and repel the ones from different classes for better generality of models trained with only limited examples is proposed.'}",http://arxiv.org/pdf/2205.01308
-Towards Open-vocabulary Scene Graph Generation with Prompt-based Finetuning,Tao He,"Scene graph generation (SGG) is a fundamental task aimed at detecting visual relations between objects in an image. The prevailing SGG methods require all object classes to be given in the training set. Such a closed setting limits the practical application of SGG. In this paper, we introduce open-vocabulary scene graph generation, a novel, realistic and challenging setting in which a model is trained on a set of base object classes but is required to infer relations for unseen target object classes. To this end, we propose a two-step method that firstly pre-trains on large amounts of coarse-grained region-caption data and then leverages two prompt-based techniques to finetune the pre-trained model without updating its parameters. Moreover, our method can support inference over completely unseen object classes, which existing methods are incapable of handling. On extensive experiments on three benchmark datasets, Visual Genome, GQA, and Open-Image, our method significantly outperforms recent, strong SGG methods on the setting of Ov-SGG, as well as on the conventional closed SGG.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a two-step method that firstly pre-trains on large amounts of coarse-grained region-caption data and then leverages two prompt-based techniques to finetune the pre-trained model without updating its parameters, which can support inference over completely unseen object classes, which existing methods are incapable of handling.'}",http://arxiv.org/pdf/2208.08165
-Zero-shot Cross-lingual Transfer of Prompt-based Tuning with a Unified Multilingual Prompt,Lianzhe Huang,"Prompt-based tuning has been proven effective for pretrained language models (PLMs). While most of the existing work focuses on the monolingual prompts, we study the multilingual prompts for multilingual PLMs, especially in the zero-shot cross-lingual setting. To alleviate the effort of designing different prompts for multiple languages, we propose a novel model that uses a unified prompt for all languages, called UniPrompt. Different from the discrete prompts and soft prompts, the unified prompt is model-based and language-agnostic. Specifically, the unified prompt is initialized by a multilingual PLM to produce language-independent representation, after which is fused with the text input. During inference, the prompts can be pre-computed so that no extra computation cost is needed. To collocate with the unified prompt, we propose a new initialization method for the target label word to further improve the model’s transferability across languages. Extensive experiments show that our proposed methods can significantly outperform the strong baselines across different languages. We release data and code to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'A novel model that uses a unified prompt for all languages, called UniPrompt, which is model-based and language-agnostic and can significantly outperform the strong baselines across different languages.'}",https://aclanthology.org/2022.emnlp-main.790.pdf
-PromptCast: A New Prompt-based Learning Paradigm for Time Series Forecasting,Hao Xue,"This paper presents a new perspective on time series forecasting. In existing time series forecasting methods, the models take a sequence of numerical values as input and yield numerical values as output. The existing SOTA models are largely based on the Transformer architecture, modified with multiple encoding mechanisms to incorporate the context and semantics around the historical data. Inspired by the successes of pre-trained language foundation models, we pose a question about whether these models can also be adapted to solve time-series forecasting. Thus, we propose a new forecasting paradigm: prompt-based time series forecasting (PromptCast). In this novel task, the numerical input and output are transformed into prompts and the forecasting task is framed in a sentence-to-sentence manner, making it possible to directly apply language models for forecasting purposes. To support and facilitate the research of this task, we also present a large-scale dataset (PISA) that includes three real-world forecasting scenarios. We evaluate different SOTA numerical-based forecasting methods and language generation models. The benchmark results with various forecasting settings demonstrate the proposed PromptCast with language generation models is a promising research direction. Additionally, in comparison to conventional numerical-based forecasting, PromptCast shows a much better generalization ability under the zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'The benchmark results with various forecasting settings demonstrate the proposed PromptCast with language generation models is a promising research direction, and in comparison to conventional numerical-based forecasting, PromptCast shows a much better generalization ability under the zero-shot setting.'}",
-LEGO-ABSA: A Prompt-based Task Assemblable Unified Generative Framework for Multi-task Aspect-based Sentiment Analysis,Tianhao Gao,"Aspect-based sentiment analysis (ABSA) has received increasing attention recently. ABSA can be divided into multiple tasks according to the different extracted elements. Existing generative methods usually treat the output as a whole string rather than the combination of different elements and only focus on a single task at once. This paper proposes a unified generative multi-task framework that can solve multiple ABSA tasks by controlling the type of task prompts consisting of multiple element prompts. Further, the proposed approach can train on simple tasks and transfer to difficult tasks by assembling task prompts, like assembling Lego bricks. We conduct experiments on six ABSA tasks across multiple benchmarks. Our proposed multi-task approach achieves new state-of-the-art results in almost all tasks and competitive results in task transfer scenarios.","{'model': 'tldr@v2.0.0', 'text': 'A unified generative multi-task framework that can solve multiple ABSA tasks by controlling the type of task prompts consisting of multiple element prompts and achieves new state-of-the-art results in almost all tasks and competitive results in task transfer scenarios.'}",
-Prompt-based Connective Prediction Method for Fine-grained Implicit Discourse Relation Recognition,Hao Zhou,"Due to the absence of connectives, implicit discourse relation recognition (IDRR) is still a challenging and crucial task in discourse analysis. Most of the current work adopted multi-task learning to aid IDRR through explicit discourse relation recognition (EDRR) or utilized dependencies between discourse relation labels to constrain model predictions. But these methods still performed poorly on fine-grained IDRR and even utterly misidentified on most of the few-shot discourse relation classes. To address these problems, we propose a novel Prompt-based Connective Prediction (PCP) method for IDRR. Our method instructs large-scale pre-trained models to use knowledge relevant to discourse relation and utilizes the strong correlation between connectives and discourse relation to help the model recognize implicit discourse relations. Experimental results show that our method surpasses the current state-of-the-art model and achieves significant improvements on those fine-grained few-shot discourse relation. Moreover, our approach is able to be transferred to EDRR and obtain acceptable results. Our code is released in https://github.com/zh-i9/PCP-for-IDRR.","{'model': 'tldr@v2.0.0', 'text': 'A novel Prompt-based Connective Prediction (PCP) method for IDRR that surpasses the current state-of-the-art model and achieves significant improvements on those fine-grained few-shot discourse relation classes.'}",http://arxiv.org/pdf/2210.07032
-Prompt-based Distribution Alignment for Domain Generalization in Text Classification,Chen Jia,"Prompt-based learning (a.k.a. prompting) achieves high performance by bridging the gap between the objectives of language modeling and downstream tasks. Domain generalization ability can be improved by prompting since classification across different domains can be unified into the prediction of the same set of label words. The remaining challenge for domain generalization by prompting comes from discrepancies between the data distribution of different domains. To improve domain generalization with prompting, we learn distributional invariance across source domains via two alignment regularization loss functions. The first is vocabulary distribution alignment, which uses a Kullback-Leibler divergence regularization on source-domain vocabulary distributions. The second is feature distribution alignment, which uses a novel adversarial training strategy to learn domain invariant representation across source domains. Experiments on sentiment analysis and natural language inference show the effectiveness of our method and achieve state-of-the-art results on six datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work learns distributional invariance across source domains via two alignment regularization loss functions via a Kullback-Leibler divergence regularization on source-domain vocabulary distributions.'}",https://aclanthology.org/2022.emnlp-main.690.pdf
-Context Variance Evaluation of Pretrained Language Models for Prompt-based Biomedical Knowledge Probing,Zonghai Yao,"Pretrained language models (PLMs) have motivated research on what kinds of knowledge these models learn. Fill-in-the-blanks problem (e.g., cloze tests) is a natural approach for gauging such knowledge. BioLAMA generates prompts for biomedical factual knowledge triples and uses the Top-k accuracy metric to evaluate different PLMs' knowledge. However, existing research has shown that such prompt-based knowledge probing methods can only probe a lower bound of knowledge. Many factors like prompt-based probing biases make the LAMA benchmark unreliable and unstable. This problem is more prominent in BioLAMA. The severe long-tailed distribution in vocabulary and large-N-M relation make the performance gap between LAMA and BioLAMA remain notable. To address these, we introduced context variance into the prompt generation and proposed a new rank-change-based evaluation metric. Different from the previous known-unknown evaluation criteria, we proposed the concept of ""Misunderstand"" in LAMA for the first time. Through experiments on 12 PLMs, we showed that our context variance prompts and Understand-Confuse-Misunderstand (UCM) metric make BioLAMA more friendly to large-N-M relations and rare relations. We also conducted a set of control experiments to disentangle ""understand"" from just ""read and copy"".","{'model': 'tldr@v2.0.0', 'text': 'Context variance prompts and Understand-Confuse-Misunderstand (UCM) metric make BioLAMA more friendly to large-N-M relations and rare relations and disentangle ""understand"" from just ""read and copy"".'}",
-Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning,Prasetya Ajie Utama,"Recent prompt-based approaches allow pretrained language models to achieve strong performances on few-shot finetuning by reformulating downstream tasks as a language modeling problem. In this work, we demonstrate that, despite its advantages on low data regimes, finetuned prompt-based models for sentence pair classification tasks still suffer from a common pitfall of adopting inference heuristics based on lexical overlap, e.g., models incorrectly assuming a sentence pair is of the same meaning because they consist of the same set of words. Interestingly, we find that this particular inference heuristic is significantly less present in the zero-shot evaluation of the prompt-based model, indicating how finetuning can be destructive to useful knowledge learned during the pretraining. We then show that adding a regularization that preserves pretraining weights is effective in mitigating this destructive tendency of few-shot finetuning. Our evaluation on three datasets demonstrates promising improvements on the three corresponding challenge datasets used to diagnose the inference heuristics.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that, despite its advantages on low data regimes, finetuned prompt-based models for sentence pair classification tasks still suffer from a common pitfall of adopting inference heuristics based on lexical overlap, and it is shown that adding a regularization that preserves pretraining weights is effective in mitigating this destructive tendency of few-shot finetuning.'}",https://aclanthology.org/2021.emnlp-main.713.pdf
-Zero-Shot Event Detection Based on Ordered Contrastive Learning and Prompt-Based Prediction,Senhui Zhang,"Event detection is a classic natural language processing task. However, the constantly emerging new events make supervised methods not applicable to unseen types. Previous zero-shot event detection methods either require predefined event types as heuristic rules or resort to external semantic analyzing tools. To overcome this weakness, we propose an end-to-end framework named Z ero-Shot E vent Detection Based on O rdered Contrastive Learning and P rompt-Based Prediction (ZEOP). By creatively introducing multiple contrastive samples with ordered similarities, the encoder can learn event representations from both instance-level and class-level, which makes the distinctions between different unseen types more significant. Meanwhile, we utilize the prompt-based prediction to identify trigger words without relying on external resources. Experiments demonstrate that our model detects events more effectively and accurately than state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end framework named Z ero-Shot Event Detection Based on O rdered Contrastive Learning and P rompt-Based Prediction (ZEOP) that can learn event representations from both instance-level and class-level, which makes the distinctions between different unseen types more significant.'}",https://aclanthology.org/2022.findings-naacl.196.pdf
-Prompt-Based Time Series Forecasting: A New Task and Dataset,Hao Xue,"The research of time series forecasting benefits a wide range of applications from weather forecasting to human mobility or traffic prediction. This paper studies the time series forecasting problem from a whole new perspective. In the existing methods, the forecasting models take a sequence of numerical values as input and yield numerical values as output. Inspired by the successes of pre-trained language foundation models, we pose a question about whether these models can also be adapted to time series forecasting tasks. Thus, we propose a novel prompt-based time series forecasting (PromptCast) task. In this task, the numerical input and output are transformed into language sentence prompts. We frame the forecasting task in a sentence-to-sentence manner which makes it possible to directly apply language models for the forecasting purpose. To support and facilitate the research of this task, we also present a large-scale dataset (PISA) that includes three real-world forecasting scenarios in this paper. We evaluate different state-of-the-art numerical-based forecasting methods and language generation models such as Bart and Bigbird. The benchmark results demonstrate that the proposed prompt-based time series forecasting with language generation models is a promising research direction. In addition, in comparison to conventional numerical-based forecasting, prompt-based forecasting shows a better generalization ability. We believe that the proposed PromptCast benchmark task as well as our PISA dataset could provide novel insights and further lead to new research directions in the time series forecasting domain.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based time series forecasting (PromptCast) task that frame the forecasting task in a sentence-to-sentence manner which makes it possible to directly apply language models for the forecasting purpose and shows a better generalization ability.'}",http://arxiv.org/pdf/2210.08964
-Prompt-Based Meta-Learning For Few-shot Text Classification,Haoxing Zhang,"Few-shot Text Classification predicts the semantic label of a given text with a handful of supporting instances. Current meta-learning methods have achieved satisfying results in various few-shot situations. Still, they often require a large amount of data to construct many few-shot tasks for meta-training, which is not practical in real-world few-shot scenarios. Prompt-tuning has recently proved to be another effective few-shot learner by bridging the gap between pre-train and downstream tasks. In this work, we closely combine the two promising few-shot learning methodologies in structure and propose a Prompt-Based Meta-Learning (PBML) model to overcome the above meta-learning problem by adding the prompting mechanism. PBML assigns label word learning to base-learners and template learning to meta-learner, respectively. Experimental results show state-of-the-art performance on four text classification datasets under few-shot settings, with higher accuracy and good robustness. We demonstrate through low-resource experiments that our method alleviates the shortcoming that meta-learning requires too much data for meta-training. In the end, we use the visualization to interpret and verify that the meta-learning framework can help the prompting method converge better. We release our code to reproduce our experiments.","{'model': 'tldr@v2.0.0', 'text': 'This work closely combines the two promising few-shot learning methodologies in structure and proposes a Prompt-Based Meta-Learning (PBML) model to overcome the above meta-learning problem by adding the prompting mechanism.'}",https://aclanthology.org/2022.emnlp-main.87.pdf
-AI Illustrator: Translating Raw Descriptions into Images by Prompt-based Cross-Modal Generation,Y. Ma,"AI illustrator aims to automatically design visually appealing images for books to provoke rich thoughts and emotions. To achieve this goal, we propose a framework for translating raw descriptions with complex semantics into semantically corresponding images. The main challenge lies in the complexity of the semantics of raw descriptions, which may be hard to be visualized e.g., ""gloomy"" or ""Asian""). It usually poses challenges for existing methods to handle such descriptions. To address this issue, we propose a Prompt-based Cross-Modal Generation Framework (PCM-Frame) to leverage two powerful pre-trained models, including CLIP and StyleGAN. Our framework consists of two components: a projection module from Text Embeddings to Image Embeddings based on prompts, and an adapted image generation module built on StyleGAN which takes Image Embeddings as inputs and is trained by combined semantic consistency losses. To bridge the gap between realistic images and illustration designs, we further adopt a stylization model as post-processing in our framework for better visual effects. Benefiting from the pre-trained models, our method can handle complex descriptions and does not require external paired data for training. Furthermore, we have built a benchmark that consists of 200 descriptions from literature books or online resources. We conduct a user study to demonstrate our superiority over the competing methods of text-to-image translation with complicated semantics.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt-based Cross-Modal Generation Framework (PCM-Frame) to leverage two powerful pre-trained models, including CLIP and StyleGAN, and conducts a user study to demonstrate its superiority over the competing methods of text-to-image translation with complicated semantics.'}",https://arxiv.org/pdf/2209.03160
-Steps towards prompt-based creation of virtual worlds,Jasmine Roberts,"Large language models trained for code generation can be applied to speaking virtual worlds into existence (creating virtual worlds). In this work we show that prompt-based methods can both accelerate in-VR level editing, as well as can become part of gameplay rather than just part of game development. As an example, we present Codex VR Pong which shows non-deterministic game mechanics using generative processes to not only create static content but also non-trivial interactions between 3D objects. This demonstration naturally leads to an integral discussion on how one would evaluate and benchmark experiences created by generative models - as there are no qualitative or quantitative metrics that apply in these scenarios. We conclude by discussing impending challenges of AI-assisted co-creation in VR.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that prompt-based methods can both accelerate in-VR level editing, as well as can become part of gameplay rather than just part of game development.'}",https://arxiv.org/pdf/2211.05875
-ParaBART: A Prompt-based Method with Parabiotic Decoder for Few-shot Named Entity Recognition,,"Prompt-based methods have been widely used 001 in few-shot named entity recognition (NER). 002 We first conduct a preliminary experiment and 003 observe that what really affects prompt-based 004 NER models is the ability to detect entity 005 boundaries. However, previous prompt-based 006 NER models neglect to enhance the ability of 007 entity boundary detection. To solve the issue, 008 we propose a novel method, ParaBART, which 009 consists of a BART encoder and the Parabi- 010 otic 1 Decoder we design. Parabiotic Decoder 011 includes two BART decoders and a conjoint 012 module. The two decoders are responsible for 013 entity boundary detection and entity type classi- 014 fication respectively and share the well-learned 015 knowledge through the conjoint module, which 016 replaces unimportant tokens’ embeddings in 017 one decoder with the average embedding of 018 all tokens in the other decoder. Moreover, we 019 propose a novel boundary expansion strategy 020 to enhance the ability of entity type classifica- 021 tion. Experimental results show that ParaBART 022 can achieve significant performance gains over 023 previous state-of-the-art methods. For repro- 024 ducibility, all datasets and codes are provided 025 in the supplementary materials. 026 ,","{'model': 'tldr@v2.0.0', 'text': 'A novel method, ParaBART, which consists of a BART encoder and the Parabi- 010 otic 1 Decoder, which can achieve significant performance gains over previous state-of-the-art methods in few-shot named entity recognition.'}",
-CLAMP: Prompt-based Contrastive Learning for Connecting Language and Animal Pose,Xu Zhang,"Animal pose estimation is challenging for existing image-based methods because of limited training data and large intra- and inter-species variances. Motivated by the progress of visual-language research, we propose that pre-trained language models (e.g., CLIP) can facilitate animal pose estimation by providing rich prior knowledge for describing animal keypoints in text. However, we found that building effective connections between pre-trained language models and visual animal keypoints is non-trivial since the gap between text-based descriptions and keypoint-based visual features about animal pose can be significant. To address this issue, we introduce a novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively. The CLAMP attempts to bridge the gap by adapting the text prompts to the animal keypoints during network training. The adaptation is decomposed into spatialaware and feature-aware processes, and two novel contrastive losses are devised correspondingly. In practice, the CLAMP enables the first cross-modal animal pose estimation paradigm. Experimental results show that our method achieves state-of-the-art performance under the supervised, few-shot, and zero-shot settings, outperforming image-based methods by a large margin. The code is available at https://github.com/xuzhang1199/CLAMP.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively, which enables the first cross-modal animal pose estimation paradigm.'}",https://arxiv.org/pdf/2206.11752
-PromptAttack: Prompt-based Attack for Language Models via Gradient Search,Yundi Shi,"As the pre-trained language models (PLMs) continue to grow, so do the hardware and data requirements for fine-tuning PLMs. Therefore, the researchers have come up with a lighter method called \textit{Prompt Learning}. However, during the investigations, we observe that the prompt learning methods are vulnerable and can easily be attacked by some illegally constructed prompts, resulting in classification errors, and serious security problems for PLMs. Most of the current research ignores the security issue of prompt-based methods. Therefore, in this paper, we propose a malicious prompt template construction method (\textbf{PromptAttack}) to probe the security performance of PLMs. Several unfriendly template construction approaches are investigated to guide the model to misclassify the task. Extensive experiments on three datasets and three PLMs prove the effectiveness of our proposed approach PromptAttack. We also conduct experiments to verify that our method is applicable in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a malicious prompt template construction method (\\textbf{PromptAttack}) to probe the security performance of PLMs and proves the effectiveness of this approach.'}",http://arxiv.org/pdf/2209.01882
-PTS: A Prompt-based Teacher-Student Network for Weakly Supervised Aspect Detection,Hongjian Li,"Most existing weakly supervised aspect detection algorithms utilize pre-trained language models as their backbone networks by constructing discriminative tasks with seed words. Once the number of seed words decreases, the performance of current models declines significantly. Recently, prompt tuning has been proposed to bridge the gap of objective forms in pre-training and fine-tuning, which is hopeful of alleviating the above challenge. However, directly applying the existing prompt-based methods to this task not only fails to effectively use large amounts of unlabeled data, but also may cause serious over-fitting problems. In this paper, we propose a lightweight teacher-student network (PTS) based on prompts to solve the above two problems. Concretely, the student network is a hybrid prompt-based classification model to detect aspects, which innovatively compounds hand-crafted prompts and auto-generated prompts. The teacher network comprehensively considers the representation of the sentence and the masked aspect token in the template to guide classification. To utilize unlabeled data and seed words intelligently, we train the teacher and student network alternately. Furthermore, in order to solve the problem that the uneven quality of training data obviously affects the iterative efficiency of PTS, we design a general dynamic data selection strategy to feed the most pertinent data into the current model. Experimental results show that even given the minimum seed words, PTS significantly outperforms previous state-of-the-art methods on three widely used benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A lightweight teacher-student network (PTS) based on prompts to solve the problem that the uneven quality of training data obviously affects the iterative efficiency of PTS, and design a general dynamic data selection strategy to feed the most pertinent data into the current model.'}",
-Prompt-Based Metric Learning for Few-Shot NER,Yanru Chen,"Few-shot named entity recognition (NER) targets generalizing to unseen labels and/or domains with few labeled examples. Existing metric learning methods compute token-level similarities between query and support sets, but are not able to fully incorporate label semantics into modeling. To address this issue, we propose a simple method to largely improve metric learning for NER: 1) multiple prompt schemas are designed to enhance label semantics; 2) we propose a novel architecture to effectively combine multiple prompt-based representations. Empirically, our method achieves new state-of-the-art (SOTA) results under 16 of the 18 considered settings, substantially outperforming the previous SOTA by an average of 8.84% and a maximum of 34.51% in relative gains of micro F1. Our code is available at https://github.com/AChen-qaq/ProML.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple method to largely improve metric learning for NER by proposing a novel architecture to effectively combine multiple prompt-based representations designed to enhance label semantics.'}",https://arxiv.org/pdf/2211.04337
-Prompt-based Zero-shot Video Moment Retrieval,Guolong Wang,"Video moment retrieval aims at localizing a specific moment from an untrimmed video by a sentence query. Most methods rely on heavy annotations of video moment-query pairs. Recent zero-shot methods reduced annotation cost, yet they neglected the global visual feature due to the separation of video and text learning process. To avoid the lack of visual features, we propose a Prompt-based Zero-shot Video Moment Retrieval (PZVMR) method. Motivated by the frame of prompt learning, we design two modules: 1) Proposal Prompt (PP): We randomly masks sequential frames to build a prompt to generate proposals; 2) Verb Prompt (VP): We provide patterns of nouns and the masked verb to build a prompt to generate pseudo queries with verbs. Our PZVMR utilizes task-relevant knowledge distilled from pre-trained CLIP and adapts the knowledge to VMR. Unlike the pioneering work, we introduce visual features into each module. Extensive experiments show that our PZVMR not only outperforms the existing zero-shot method (PSVL) on two public datasets (Charades-STA and ActivityNet-Captions) by 4.4% and 2.5% respectively in mIoU, but also outperforms several methods using stronger supervision.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Prompt-based Zero-shot Video Moment Retrieval (PZVMR) method that utilizes task-relevant knowledge distilled from pre-trained CLIP and adapts the knowledge to VMR and outperforms several methods using stronger supervision.'}",
-On the Robustness of Dialogue History Representation in Conversational Question Answering: A Comprehensive Study and a New Prompt-based Method,Zorik Gekhman,"Most work on modeling the conversation history in Conversational Question Answering (CQA) reports a single main result on a common CQA benchmark. While existing models show impressive results on CQA leaderboards, it remains unclear whether they are robust to shifts in setting (sometimes to more realistic ones), training data size (e.g., from large to small sets) and domain. In this work, we design and conduct the first large-scale robustness study of history modeling approaches for CQA. We find that high benchmark scores do not necessarily translate to strong robustness, and that various methods can perform extremely differently under different settings. Equipped with the insights from our study, we design a novel prompt-based history modeling approach and demonstrate its strong robustness across various settings. Our approach is inspired by existing methods that highlight historic answers in the passage. However, instead of highlighting by modifying the passage token embeddings, we add textual prompts directly in the passage text. Our approach is simple, easy to plug into practically any model, and highly effective, thus we recommend it as a starting point for future model developers. We also hope that our study and insights will raise awareness to the importance of robustness-focused evaluation, in addition to obtaining high leaderboard scores, leading to better CQA systems.1","{'model': 'tldr@v2.0.0', 'text': 'This work design and conduct the first large-scale robustness study of history modeling approaches for CQA, finding that high benchmark scores do not necessarily translate to strong robustness, and that various methods can perform extremely differently under different settings.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00549/2080031/tacl_a_00549.pdf
-PromptDA: Label-guided Data Augmentation for Prompt-based Few Shot Learners,Canyu Chen,"Recent advances in large pre-trained language models (PLMs) lead to impressive gains on natural language understanding (NLU) tasks with task-specific fine-tuning. However, directly fine-tuning PLMs heavily relies on sufficient labeled training instances, which are usually hard to obtain. Prompt-based tuning on PLMs has shown to be powerful for various downstream few-shot tasks. Existing works studying prompt-based tuning for few-shot NLU tasks mainly focus on deriving proper label words with a verbalizer or generating prompt templates to elicit semantics from PLMs. In addition, conventional data augmentation strategies such as synonym substitution are also widely adopted in low-resource scenarios. However, the improvements they bring to prompt-based few-shot learning have been demonstrated to be marginal. Thus, an important research question arises as follows: how to design effective data augmentation methods for prompt-based few-shot tuning? To this end, considering the label semantics are essential in prompt-based tuning, we propose a novel label-guided data augmentation framework PromptDA, which exploits the enriched label semantic information for data augmentation. Extensive experiment results on few-shot text classification tasks show that our proposed framework achieves superior performances by effectively leveraging label semantics and data augmentation for natural language understanding.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel label-guided data augmentation framework PromptDA, which exploits the enriched label semantic information for data augmentations in prompt-based few-shot tuning and achieves superior performances.'}",http://arxiv.org/pdf/2205.09229
-NSP-BERT: A Prompt-based Zero-Shot Learner Through an Original Pre-training Task-Next Sentence Prediction,Yi Sun,"Using prompts to utilize language models 001 to perform various downstream tasks, also 002 known as prompt-based learning or prompt- 003 learning , has lately gained significant success 004 in comparison to the pre-train and fine-tune 005 paradigm. Nonetheless, virtually all prompt- 006 based methods are token-level, meaning they 007 all utilize GPT’s left-to-right language model 008 or BERT’s masked language model to per- 009 form cloze-style tasks. In this paper, we at- 010 tempt to accomplish several NLP tasks in the 011 zero-shot scenario using a BERT original pre- 012 training task abandoned by RoBERTa and other 013 models—Next Sentence Prediction (NSP). Un- 014 like token-level techniques, our sentence-level 015 prompt-based method NSP-BERT does not 016 need to fix the length of the prompt or the po- 017 sition to be predicted, allowing it to handle 018 tasks such as entity linking with ease. Based on 019 the characteristics of NSP-BERT, we offer sev- 020 eral quick building templates for various down- 021 stream tasks. We suggest a two-stage prompt 022 method for word sense disambiguation tasks 023 in particular. Our samples-contrast method for 024 mapping the labels significantly enhance the 025 model’s performance on sentence-pair tasks. 026 On the Chinese benchmark FewCLUE, our 027 NSP-BERT outperforms other zero-shot meth- 028 ods on most of these tasks and comes close to 029 the few-shot methods. And on GLUE and other 030 English datasets NSP-BERT is still competitive. 031 Our code will be available on github. 032","{'model': 'tldr@v2.0.0', 'text': 'This paper accomplishes several NLP tasks in the 011 zero-shot scenario using a BERT original pre- 012 training task abandoned by RoBERTa and other 013 models—Next Sentence Prediction (NSP).'}",
-Unified Multimodal Pre-training and Prompt-based Tuning for Vision-Language Understanding and Generation,Tianyi Liu,"Most existing vision-language pre-training methods focus on understanding tasks and use BERT-like objectives (masked language modeling and image-text matching) during pretraining. Although they perform well in many understanding downstream tasks, e.g., visual question answering, image-text retrieval and visual entailment, they do not possess the ability to generate. To tackle this problem, we propose Unified multimodal pre-training for both Vision-Language understanding and generation (UniVL). The proposed UniVL is capable of handling both understanding tasks and generative tasks. We augment existing pretraining paradigms that only use random masks with causal masks, i.e., triangular masks that mask out future tokens, such that the pre-trained models can have autoregressive generation abilities by design. We formulate several previous understanding tasks as a text generation task and propose to use prompt-based method for fine-tuning on different downstream tasks. Our experiments show that there is a trade-off between understanding tasks and generation tasks while using the same model, and a feasible way to improve both tasks is to use more data. Our UniVL framework attains comparable performance to recent vision-language pre-training methods on both understanding tasks and generation tasks. Moreover, we demostrate that prompt-based finetuning is more data-efficient - it outperforms discriminative methods in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The UniVL framework attains comparable performance to recent vision-language pre-training methods on both understanding tasks and generation tasks, and demostrate that prompt-based finetuning is more data-efficient - it outperforms discriminative methods in few-shot scenarios.'}",
-Adversarial Robustness of Prompt-based Few-Shot Learning for Natural Language Understanding,Venkata Prabhakara Sarath Nookala,"State-of-the-art few-shot learning (FSL) methods leverage prompt-based fine-tuning to obtain remarkable results for natural language understanding (NLU) tasks. While much of the prior FSL methods focus on improving downstream task performance, there is a limited understanding of the adversarial robustness of such methods. In this work, we conduct an extensive study of several state-of-the-art FSL methods to assess their robustness to adversarial perturbations. To better understand the impact of various factors towards robustness (or the lack of it), we evaluate prompt-based FSL methods against fully fine-tuned models for aspects such as the use of unlabeled data, multiple prompts, number of few-shot examples, model size and type. Our results on six GLUE tasks indicate that compared to fully fine-tuned models, vanilla FSL methods lead to a notable relative drop in task performance (i.e., are less robust) in the face of adversarial perturbations. However, using (i) unlabeled data for prompt-based FSL and (ii) multiple prompts flip the trend. We further demonstrate that increasing the number of few-shot examples and model size lead to increased adversarial robustness of vanilla FSL methods. Broadly, our work sheds light on the adversarial robustness evaluation of prompt-based FSL methods for NLU tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts an extensive study of several state-of-the-art FSL methods to assess their robustness to adversarial perturbations, and demonstrates that increasing the number of few-shot examples and model size lead to increased adversarial robustness of vanilla F SL methods.'}",http://arxiv.org/pdf/2306.11066
-"Don’t Prompt, Search! Mining-based Zero-Shot Learning with Language Models",Mozes van de Kar,"Masked language models like BERT can perform text classification in a zero-shot fashion by reformulating downstream tasks as text infilling. However, this approach is highly sensitive to the template used to prompt the model, yet practitioners are blind when designing them in strict zero-shot settings. In this paper, we propose an alternative mining-based approach for zero-shot learning. Instead of prompting language models, we use regular expressions to mine labeled examples from unlabeled corpora, which can optionally be filtered through prompting, and used to finetune a pretrained model. Our method is more flexible and interpretable than prompting, and outperforms it on a wide range of tasks when using comparable templates. Our results suggest that the success of prompting can partly be explained by the model being exposed to similar examples during pretraining, which can be directly retrieved through regular expressions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an alternative mining-based approach for zero-shot learning that uses regular expressions to mine labeled examples from unlabeled corpora, which can optionally be filtered through prompting, and used to finetune a pretrained model.'}",https://arxiv.org/pdf/2210.14803
-Prediction model development of late-onset preeclampsia using machine learning-based methods,J. Jhee,"Preeclampsia is one of the leading causes of maternal and fetal morbidity and mortality. Due to the lack of effective preventive measures, its prediction is essential to its prompt management. This study aimed to develop models using machine learning to predict late-onset preeclampsia using hospital electronic medical record data. The performance of the machine learning based models and models using conventional statistical methods were also compared. A total of 11,006 pregnant women who received antenatal care at Yonsei University Hospital were included. Maternal data were retrieved from electronic medical records during the early second trimester to 34 weeks. The prediction outcome was late-onset preeclampsia occurrence after 34 weeks’ gestation. Pattern recognition and cluster analysis were used to select the parameters included in the prediction models. Logistic regression, decision tree model, naïve Bayes classification, support vector machine, random forest algorithm, and stochastic gradient boosting method were used to construct the prediction models. C-statistics was used to assess the performance of each model. The overall preeclampsia development rate was 4.7% (474 patients). Systolic blood pressure, serum blood urea nitrogen and creatinine levels, platelet counts, serum potassium level, white blood cell count, serum calcium level, and urinary protein were the most influential variables included in the prediction models. C-statistics for the decision tree model, naïve Bayes classification, support vector machine, random forest algorithm, stochastic gradient boosting method, and logistic regression models were 0.857, 0.776, 0.573, 0.894, 0.924, and 0.806, respectively. The stochastic gradient boosting model had the best prediction performance with an accuracy and false positive rate of 0.973 and 0.009, respectively. The combined use of maternal factors and common antenatal laboratory data of the early second trimester through early third trimester could effectively predict late-onset preeclampsia using machine learning algorithms. Future prospective studies are needed to verify the clinical applicability algorithms.","{'model': 'tldr@v2.0.0', 'text': 'The combined use of maternal factors and common antenatal laboratory data of the early second trimester through early third trimester could effectively predict late-onset preeclampsia using machine learning algorithms.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0221202&type=printable
-"Management and Point-of-Care for Tobacco Dependence (PROMPT): a feasibility mixed methods community-based participatory action research project in Ottawa, Canada",S. Pakhalé,"Objective To determine the feasibility of a Community-Based Participatory Tobacco Dependence Strategy (PROMPT) in the inner city population of Ottawa (Canada). Design A feasibility mixed methods prospective cohort study following principles of community-based participatory action research. Intervention Recruited 80 people whouse drugs, followed them for 6 months while providing access to counselling, nicotine replacement therapy and peer-support in a community setting. Setting Community research office in downtown Ottawa, adjacent to low-income housing, shelter services and street-based drug consumption. Primary outcome Retention rate at 6-month follow-up. Secondary outcome Biochemically validated 7-day point prevalence smoking abstinence at 26 weeks, self-reported abstinence in the past 7 days with exhaled carbon monoxide ≤10 ppm. Results The average age of participants was 43.8 years. The 6-month follow-up rate was 42.5%. The mean number of smoking years reported was 27.3 years. The participants were 70% male, 33.7% reported less than a high-school education, 21% identified as indigenous and 43.8% reported an income between US$1000 and US$1999 per month. The baseline mean daily cigarette use was 20.5 and 9.3 cigarettes at study end, with mean reduction of 11.2 cigarettes at 6 months (P=0.0001). There was a considerable reduction in self-reported illicit substance use (18.8%), including a reduction in the opioids heroin (6.3%), fentanyl (2.6%) and Oxycontin (3.8%). The study findings also reveal psycho-socioeconomic benefits such as improved health, return to work and greater community engagement. Conclusions The PROMPT project describes socioeconomic variables associated with tobacco and polysubstance use. A programme focused on tobacco dependence, easily accessible in the community and led by community peers with lived experience is feasible to implement and has the potential to support positive life changes. PROMPT’s patient engagement model is an effective harm-reduction strategy for the growing opioid use crisis and can improve the health outcomes of marginalised at-risk populations worldwide.","{'model': 'tldr@v2.0.0', 'text': 'A programme focused on tobacco dependence, easily accessible in the community and led by community peers with lived experience is feasible to implement and has the potential to support positive life changes.'}",https://bmjopen.bmj.com/content/bmjopen/8/1/e018416.full.pdf
-Web-Based Patient-Reported Outcome Measures for Personalized Treatment and Care (PROMPT-Care): Multicenter Pragmatic Nonrandomized Trial,A. Girgis,"Background Despite the acceptability and efficacy of e–patient-reported outcome (ePRO) systems, implementation in routine clinical care remains challenging. Objective This pragmatic trial implemented the PROMPT-Care (Patient Reported Outcome Measures for Personalized Treatment and Care) web-based system into existing clinical workflows and evaluated its effectiveness among a diverse population of patients with cancer. Methods Adult patients with solid tumors receiving active treatment or follow-up care in four cancer centers were enrolled. The PROMPT-Care intervention supported patient management through (1) monthly off-site electronic PRO physical symptom and psychosocial well-being assessments, (2) automated electronic clinical alerts notifying the care team of unresolved clinical issues following two consecutive assessments, and (3) tailored online patient self-management resources. Propensity score matching was used to match controls with intervention patients in a 4:1 ratio for patient age, sex, and treatment status. The primary outcome was a reduction in emergency department presentations. Secondary outcomes were time spent on chemotherapy and the number of allied health service referrals. Results From April 2016 to October 2018, 328 patients from four public hospitals received the intervention. Matched controls (n=1312) comprised the general population of patients with cancer, seen at the participating hospitals during the study period. Emergency department visits were significantly reduced by 33% (P=.02) among patients receiving the intervention compared with patients in the matched controls. No significant associations were found in allied health referrals or time to end of chemotherapy. At baseline, the most common patient reported outcomes (above-threshold) were fatigue (39%), tiredness (38.4%), worry (32.9%), general wellbeing (32.9%), and sleep (24.1%), aligning with the most frequently accessed self-management domain pages of physical well-being (36%) and emotional well-being (23%). The majority of clinical feedback reports were reviewed by nursing staff (729/893, 82%), largely in response to the automated clinical alerts (n=877). Conclusions Algorithm-supported web-based systems utilizing patient reported outcomes in clinical practice reduced emergency department presentations among a diverse population of patients with cancer. This study also highlighted the importance of (1) automated triggers for reviewing above-threshold results in patient reports, rather than passive manual review of patient records; (2) the instrumental role nurses play in managing alerts; and (3) providing patients with resources to support guided self-management, where appropriate. Together, these factors will inform the integration of web-based PRO systems into future models of routine cancer care. Trial Registration Australian New Zealand Clinical Trials Registry ACTRN12616000615482; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=370633 International Registered Report Identifier (IRRID) RR2-10.1186/s12885-018-4729-3","{'model': 'tldr@v2.0.0', 'text': 'Algorithm-supported web-based systems utilizing patient reported outcomes in clinical practice reduced emergency department presentations among a diverse population of patients with cancer.'}",https://www.jmir.org/2020/10/e19685/PDF
-Decorate the Newcomers: Visual Domain Prompt for Continual Test Time Adaptation,Yulu Gan,"Continual Test-Time Adaptation (CTTA) aims to adapt the source model to continually changing unlabeled target domains without access to the source data. Existing methods mainly focus on model-based adaptation in a self-training manner, such as predicting pseudo labels for new domain datasets. Since pseudo labels are noisy and unreliable, these methods suffer from catastrophic forgetting and error accumulation when dealing with dynamic data distributions. Motivated by the prompt learning in NLP, in this paper, we propose to learn an image-layer visual domain prompt for target domains while having the source model parameters frozen. During testing, the changing target datasets can be adapted to the source model by reformulating the input data with the learned visual prompts. Specifically, we devise two types of prompts, i.e., domains-specific prompts and domains-agnostic prompts, to extract current domain knowledge and maintain the domain-shared knowledge in the continual adaptation. Furthermore, we design a homeostasis-based adaptation strategy to suppress domain-sensitive parameters in domain-invariant prompts to learn domain-shared knowledge more effectively. This transition from the model-dependent paradigm to the model-free one enables us to bypass the catastrophic forgetting and error accumulation problems. Experiments show that our proposed method achieves significant performance gains over state-of-the-art methods on four widely-used benchmarks, including CIFAR-10C, CIFAR-100C, ImageNet-C, and VLCS datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to learn an image-layer visual domain prompt for target domains while having the source model parameters frozen, and designs a homeostasis-based adaptation strategy to suppress domain-sensitive parameters in domain-invariant prompts to learn domain-shared knowledge more effectively.'}",http://arxiv.org/pdf/2212.04145
-Prompt-tuned Code Language Model as a Neural Knowledge Base for Type Inference in Statically-Typed Partial Code,Qing Huang,"Partial code usually involves non-fully-qualified type names (non-FQNs) and undeclared receiving objects. Resolving the FQNs of these non-FQN types and undeclared receiving objects (referred to as type inference) is the prerequisite to effective search and reuse of partial code. Existing dictionary-lookup based methods build a symbolic knowledge base of API names and code contexts, which involve significant compilation overhead and are sensitive to unseen API names and code context variations. In this paper, we formulate type inference as a cloze-style fill-in-blank language task. Built on source code naturalness, our approach fine-tunes a code masked language model (MLM) as a neural knowledge base of code elements with a novel “pre-train, prompt and predict” paradigm from raw source code. Our approach is lightweight and has minimum requirements on code compilation. Unlike existing symbolic name and context matching for type inference, our prompt-tuned code MLM packs FQN syntax and usage in its parameters and supports fuzzy neural type inference. We systematically evaluate our approach on a large amount of source code from GitHub and Stack Overflow. Our results confirm the effectiveness of our approach design and the practicality for partial code type inference. As the first of its kind, our neural type inference method opens the door to many innovative ways of using partial code.","{'model': 'tldr@v2.0.0', 'text': 'This paper fine-tunes a code masked language model (MLM) as a neural knowledge base of code elements with a novel “pre-train, prompt and predict” paradigm from raw source code, and formulates type inference as a cloze-style fill-in-blank language task.'}",https://dl.acm.org/doi/pdf/10.1145/3551349.3556912
-"Toward Human Readable Prompt Tuning: Kubrick's The Shining is a good movie, and a good prompt too?",Weijia Shi,"Large language models can perform new tasks in a zero-shot fashion, given natural language prompts that specify the desired behavior. Such prompts are typically hand engineered, but can also be learned with gradient-based methods from labeled data. However, it is underexplored what factors make the prompts effective, especially when the prompts are natural language. In this paper, we investigate common attributes shared by effective prompts. We first propose a human readable prompt tuning method (F LUENT P ROMPT) based on Langevin dynamics that incorporates a fluency constraint to find a diverse distribution of effective and fluent prompts. Our analysis reveals that effective prompts are topically related to the task domain and calibrate the prior probability of label words. Based on these findings, we also propose a method for generating prompts using only unlabeled data, outperforming strong baselines by an average of 7.0% accuracy across three tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a human readable prompt tuning method (F LUENT P ROMPT) based on Langevin dynamics that incorporates a fluency constraint to find a diverse distribution of effective and fluent prompts and investigates common attributes shared by effective prompts.'}",http://arxiv.org/pdf/2212.10539
-Parameter-Efficient Prompt Tuning Makes Generalized and Calibrated Neural Text Retrievers,W. Tam,"Prompt tuning attempts to update few task-specific parameters in pre-trained models. It has achieved comparable performance to fine-tuning of the full parameter set on both language understanding and generation tasks. In this work, we study the problem of prompt tuning for neural text retrievers. We introduce parameter-efficient prompt tuning for text retrieval across in-domain, cross-domain, and cross-topic settings. Through an extensive analysis, we show that the strategy can mitigate the two issues -- parameter-inefficiency and weak generalizability -- faced by fine-tuning based retrieval methods. Notably, it can significantly improve the out-of-domain zero-shot generalization of the retrieval models. By updating only 0.1% of the model parameters, the prompt tuning strategy can help retrieval models achieve better generalization performance than traditional methods in which all parameters are updated. Finally, to facilitate research on retrievers' cross-topic generalizability, we curate and release an academic retrieval dataset with 18K query-results pairs in 87 topics, making it the largest topic-specific one to date.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces parameter-efficient prompt tuning for text retrieval across in-domain, cross- domain, and cross-topic settings and shows that the strategy can mitigate the two issues -- parameter-inefficiency and weak generalizability -- faced by fine-tuning based retrieval methods.'}",http://arxiv.org/pdf/2207.07087
-Towards Unified Prompt Tuning for Few-shot Text Classification,J. Wang,"Prompt-based fine-tuning has boosted the performance of Pre-trained Language Models (PLMs) on few-shot text classification by employing task-specific prompts. Yet, PLMs are unfamiliar with prompt-style expressions during pre-training, which limits the few-shot learning performance on downstream tasks. It would be desirable if the models can acquire some prompting knowledge before adaptation to specific NLP tasks. We present the Unified Prompt Tuning (UPT) framework, leading to better few-shot text classification for BERT-style models by explicitly capturing prompting semantics from non-target NLP datasets. In UPT, a novel paradigm Prompt-Options-Verbalizer is proposed for joint prompt learning across different NLP tasks, forcing PLMs to capture task-invariant prompting knowledge. We further design a self-supervised task named Knowledge-enhanced Selective Masked Language Modeling to improve the PLM's generalization abilities for accurate adaptation to previously unseen tasks. After multi-task learning across multiple tasks, the PLM can be better prompt-tuned towards any dissimilar target tasks in low-resourced settings. Experiments over a variety of NLP tasks show that UPT consistently outperforms state-of-the-arts for prompt-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': ""In UPT, a novel paradigm Prompt-Options-Verbalizer is proposed for joint prompt learning across different NLP tasks, forcing PLMs to capture task-invariant prompting knowledge, and a self-supervised task named Knowledge-enhanced Selective Masked Language Modeling is designed to improve the PLM's generalization abilities for accurate adaptation to previously unseen tasks.""}",http://arxiv.org/pdf/2205.05313
-Relation Extraction as Open-book Examination: Retrieval-enhanced Prompt Tuning,Xiang Chen,"Pre-trained language models have contributed significantly to relation extraction by demonstrating remarkable few-shot learning abilities. However, prompt tuning methods for relation extraction may still fail to generalize to those rare or hard patterns. Note that the previous parametric learning paradigm can be viewed as memorization regarding training data as a book and inference as the close-book test. Those long-tailed or hard patterns can hardly be memorized in parameters given few-shot instances. To this end, we regard RE as an open-book examination and propose a new semiparametric paradigm of retrieval-enhanced prompt tuning for relation extraction. We construct an open-book datastore for retrieval regarding prompt-based instance representations and corresponding relation labels as memorized key-value pairs. During inference, the model can infer relations by linearly interpolating the base output of PLM with the non-parametric nearest neighbor distribution over the datastore. In this way, our model not only infers relation through knowledge stored in the weights during training but also assists decision-making by unwinding and querying examples in the open-book datastore. Extensive experiments on benchmark datasets show that our method can achieve state-of-the-art in both standard supervised and few-shot settings","{'model': 'tldr@v2.0.0', 'text': 'This work regards RE as an open-book examination and proposes a new semiparametric paradigm of retrieval-enhanced prompt tuning for relation extraction, which can achieve state-of-the-art in both standard supervised and few-shot settings.'}",https://arxiv.org/pdf/2205.02355
-Prompt Deep Light-Weight Vessel Segmentation Network (PLVS-Net),Muhammad Arsalan,"Achieving accurate retinal vessel segmentation is critical in the progression and diagnosis of vision-threatening diseases such as diabetic retinopathy and age-related macular degeneration. Existing vessel segmentation methods are based on encoder-decoder architectures, which frequently fail to take into account the retinal vessel structure's context in their analysis. As a result, such methods have difficulty bridging the semantic gap between encoder and decoder characteristics. This paper proposes a Prompt Deep Light-weight Vessel Segmentation Network (PLVS-Net) to address these issues by using prompt blocks. Each prompt block use combination of asymmetric kernel convolutions, depth-wise separable convolutions, and ordinary convolutions to extract useful features. This novel strategy improves the performance of the segmentation network while simultaneously decreasing the number of trainable parameters. Our method outperformed competing approaches in the literature on three benchmark datasets, including DRIVE, STARE, and CHASE.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt Deep Light-weight Vessel Segmentation Network (PLVS-Net) is proposed to address issues by using prompt blocks to improve the performance of the segmentation network while simultaneously decreasing the number of trainable parameters.'}",
-Rethinking Reinforcement Learning for Recommendation: A Prompt Perspective,Xin Xin,"Modern recommender systems aim to improve user experience. As reinforcement learning (RL) naturally fits this objective---maximizing an user's reward per session---it has become an emerging topic in recommender systems. Developing RL-based recommendation methods, however, is not trivial due to the offline training challenge. Specifically, the keystone of traditional RL is to train an agent with large amounts of online exploration making lots of 'errors' in the process. In the recommendation setting, though, we cannot afford the price of making 'errors' online. As a result, the agent needs to be trained through offline historical implicit feedback, collected under different recommendation policies; traditional RL algorithms may lead to sub-optimal policies under these offline training settings. Here we propose a new learning paradigm---namely Prompt-Based Reinforcement Learning (PRL)---for the offline training of RL-based recommendation agents. While traditional RL algorithms attempt to map state-action input pairs to their expected rewards (e.g., Q-values), PRL directly infers actions (i.e., recommended items) from state-reward inputs. In short, the agents are trained to predict a recommended item given the prior interactions and an observed reward value---with simple supervised learning. At deployment time, this historical (training) data acts as a knowledge base, while the state-reward pairs are used as a prompt. The agents are thus used to answer the question: Which item should be recommended given the prior interactions & the prompted reward value? We implement PRL with four notable recommendation models and conduct experiments on two real-world e-commerce datasets. Experimental results demonstrate the superior performance of our proposed methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new learning paradigm, Prompt-Based Reinforcement Learning (PRL), for the offline training of RL-based recommendation agents, and implements PRL with four notable recommendation models and conducts experiments on two real-world e-commerce datasets.'}",https://dl.acm.org/doi/pdf/10.1145/3477495.3531714
-Non-destructive analyses of bronze artefacts from Bronze Age Hungary using neutron-based methods,V. Kiss,"In this paper we present the application of various neutron-based methods carried out at the Budapest Neutron Centre. Non-destructive and non-invasive neutron radiography (NR), prompt gamma activation analysis (PGAA) and time-of-flight neutron diffraction (TOF-ND) analysis were applied to reveal more information on raw material and production techniques of bronze artefacts that can be dated to the Central European Bronze Age (2000–1200 BC).",,
-Point Prompt Tuning for Temporally Language Grounding,Yawen Zeng,"The task of temporally language grounding (TLG) aims to locate a video moment from an untrimmed video that match a given textual query, which has attracted considerable research attention. In recent years, typical retrieval-based TLG methods are inefficient due to pre-segmented candidate moments, while localization-based TLG solutions adopt reinforcement learning resulting in unstable convergence. Therefore, how to perform TLG task efficiently and stably is a non-trivial work. Toward this end, we innovatively contribute a solution, Point Prompt Tuning (PPT), which formulates this task as a prompt-based multi-modal problem and integrates multiple sub-tasks to tuning performance. Specifically, a flexible prompt strategy is contributed to rewrite the query firstly, which contains both query, start point and end point. Thereafter, a multi-modal Transformer is adopted to fully learn the multi-modal context. Meanwhile, we design various sub-tasks to constrain the novel framework, namely matching task and localization task. Finally, the start and end points of matched video moment are straightforward predicted, simply yet stably. Extensive experiments on two real-world datasets have well verified the effectiveness of our proposed solution.","{'model': 'tldr@v2.0.0', 'text': 'This work innovatively contribute a solution, Point Prompt Tuning (PPT), which formulates this task as a prompt-based multi-modal problem and integrates multiple sub-tasks to tuning performance.'}",
-Generative Prompt Tuning for Relation Classification,Jiale Han,"Using prompts to explore the knowledge contained within pre-trained language models for downstream tasks has now become an active topic. Current prompt tuning methods mostly convert the downstream tasks to masked language modeling problems by adding cloze-style phrases and mapping all labels to verbalizations with fixed length, which has proven effective for tasks with simple label spaces. However, when applied to relation classification exhibiting complex label spaces, vanilla prompt tuning methods may struggle with label verbalizations with arbitrary lengths due to rigid prompt restrictions. Inspired by the text infilling task for pre-training generative models that can flexibly predict missing spans, we propose a novel generative prompt tuning method to reformulate relation classification as an infilling problem, which frees our approach from limitations of current prompt based approaches and thus fully exploits rich semantics of entity and relation types. In addition, we design entity-guided decoding and discriminative relation scoring to generate and align relations effectively and efficiently during inference. Extensive experiments under fully supervised settings and low-resource settings demonstrate the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel generative prompt tuning method to reformulate relation classification as an infilling problem, which frees the approach from limitations of current prompt based approaches and thus fully exploits rich semantics of entity and relation types.'}",http://arxiv.org/pdf/2210.12435
-eHealth System for Collecting and Utilizing Patient Reported Outcome Measures for Personalized Treatment and Care (PROMPT-Care) Among Cancer Patients: Mixed Methods Approach to Evaluate Feasibility and Acceptability,A. Girgis,"Background Despite accumulating evidence indicating that collecting patient-reported outcomes (PROs) and transferring results to the treating health professional in real time has the potential to improve patient well-being and cancer outcomes, this practice is not widespread. Objective The aim of this study was to test the feasibility and acceptability of PROMPT-Care (Patient Reported Outcome Measures for Personalized Treatment and Care), a newly developed electronic health (eHealth) system that facilitates PRO data capture from cancer patients, data linkage and retrieval to support clinical decisions and patient self-management, and data retrieval to support ongoing evaluation and innovative research. Methods We developed an eHealth system in consultation with content-specific expert advisory groups and tested it with patients receiving treatment or follow-up care in two hospitals in New South Wales, Australia, over a 3-month period. Participants were recruited in clinic and completed self-report Web-based assessments either just before their upcoming clinical consultation or every 4 weeks if in follow-up care. A mixed methods approach was used to evaluate feasibility and acceptability of PROMPT-Care; data collected throughout the study informed the accuracy and completeness of data transfer procedures, and extent of missing data was determined from participants’ assessments. Patients participated in cognitive interviews while completing their first assessment and completed evaluation surveys and interviews at study-end to assess system acceptability and usefulness of patient self-management resources, and oncology staff were interviewed at study-end to determine the acceptability and perceived usefulness of real-time PRO reporting. Results A total of 42 patients consented to the study; 7 patients were withdrawn before starting the intervention primarily because of changes in eligibility. Overall, 35 patients (13 on treatment and 22 in follow-up) completed 67 assessments during the study period. Mean completeness of patient-reported data was 93%, with 100% accuracy of data transfer. Ten patients completed cognitive interviews, 28 completed evaluation surveys, and 14 completed evaluation interviews at study-end. PROMPT-Care patient acceptability was high—100% (28/28) reported the time to complete the Web-based assessments (average 15 min) as about right, most willing to answer more questions (79%, 22/28 yes), 96% (27/28) found the Web-based assessment easier or same as completing a paper copy, and they valued the self-management resources . Oncology staff (n=5) also reported high acceptability and potential feasibility of the system. Conclusions Patients and oncology staff found the PROMPT-Care system to be highly acceptable, and the results suggest that it would be feasible to implement it into an oncology setting. Suggested modifications to the patient assessment survey, clinician access to the reports, and system requirements will be made as part of the next stage of large-scale testing and future implementation of the system as part of routine care. Trial registration Australian New Zealand Clinical Trials Registry ACTRN1261500135294; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=369299&isReview=true (Archived by WebCite at http://www.webcitation.org/6lzylG5A0).","{'model': 'tldr@v2.0.0', 'text': ""Patients and oncology staff found the PROMPT-Care system to be highly acceptable, and the results suggest that it would be feasible to implement it into an oncologists' setting.""}",
-Prompt Learning for Few-Shot Dialogue State Tracking,Yuting Yang,"Collecting dialogue state labels, slots and values, for learning dialogue state tracking (DST) models can be costly, especially with the wide application of dialogue systems in new-rising domains. In this paper, we focus on how to learn a DST model efficiently with limited labeled data. We design a prompt learning framework for few-shot DST, which consists of two main components: value-based prompt and inverse prompt mechanism. This framework aims to utilize the language understanding and generation ability of pre-trained language models (PLM). First, we design value-based prompt functions to probe the DST-related knowledge from PLM, which do not rely on the known ontology of slots. Further, an inverse prompt mechanism is utilized to self-check the “prompted” knowledge and help the PLM understand the essence of DST task further. Experiments show that our model can generate unseen slots and outperforms existing state-of-the-art few-shot methods. It indicates that DST-related knowledge can be probed from PLM and utilized to address low-resource DST efficiently with the help of prompt learning.","{'model': 'tldr@v2.0.0', 'text': 'A prompt learning framework for few-shot DST, which consists of two main components: value-based prompt and inverse prompt mechanism, that indicates that DST-related knowledge can be probed from PLM and utilized to address low-resource DST efficiently with the help of prompt learning.'}",
-"Prompt, Generate, Then Cache: Cascade of Foundation Models Makes Strong Few-Shot Learners",Renrui Zhang,"Visual recognition in low-data regimes requires deep neural networks to learn generalized representations from limited training samples. Recently, CLIP-based methods have shown promising few-shot performance benefited from the contrastive language-image pre-training. We then question, if the more diverse pre-training knowledge can be cascaded to further assist few-shot representation learning. In this paper, we propose CaFo, a Cascade of Foundation models that incorporates diverse prior knowledge of various pretraining paradigms for better few-shot learning. Our CaFo incorporates CLIP's language-contrastive knowledge, DINO's vision-contrastive knowledge, DALL-E's vision-generative knowledge, and GPT-3's language-generative knowledge. Specifically, CaFo works by ‘Prompt, Generate, then Cache’. Firstly, we leverage GPT-3 to produce textual inputs for prompting CLIP with rich downstream linguistic semantics. Then, we generate synthetic images via DALL-E to expand the few-shot training data without any manpower. At last, we introduce a learnable cache model to adaptively blend the predictions from CLIP and DINO. By such collaboration, CaFo can fully unleash the potential of different pre-training methods and unify them to perform state-of-the-art for few-shot classification. Code is available at https://github.com/ZrrSkywalker/CaFo.","{'model': 'tldr@v2.0.0', 'text': 'CaFo, a Cascade of Foundation models that incorporates diverse prior knowledge of various pretraining paradigms for better few-shot learning is proposed that can fully unleash the potential of different pre-training methods and unify them to perform state-of-the-art for few- shot classification.'}",https://arxiv.org/pdf/2303.02151
-Self-Consistent Contrastive Attributed Graph Clustering With Pseudo-Label Prompt,Wei Xia,"Attributed graph clustering, which learns node representation from node attribute and topological graph for clustering, is a fundamental and challenging task for multimedia network-structured data analysis. Recently, graph contrastive learning (GCL)-based methods have obtained impressive clustering performance on this task. Nevertheless, there still remain some limitations to be solved: 1) most existing methods fail to consider the self-consistency between latent representations and cluster structures; and 2) most methods require a post-processing operation to get clustering labels. Such a two-step learning scheme results in models that cannot handle newly generated data, i.e., out-of-sample (OOS) nodes. To address these issues in a unified framework, a Self-consistent Contrastive Attributed Graph Clustering (SCAGC) network with pseudo-label prompt is proposed in this article. In SCAGC, by clustering labels prompt information, a self-consistent contrastive loss, which aims to maximize the consistencies of intra-cluster representations while minimizing the consistencies of inter-cluster representations, is designed for representation learning. Meanwhile, a clustering module is built to directly output clustering labels by contrasting the representation of different clusters. Thus, for the OOS nodes, SCAGC can directly calculate their clustering labels. Extensive experimental results on seven benchmark datasets have shown that SCAGC consistently outperforms 16 competitive clustering methods.","{'model': 'tldr@v2.0.0', 'text': 'A Self-consistent Contrastive Attributed Graph Clustering (SCAGC) network with pseudo-label prompt with extensive experimental results on seven benchmark datasets have shown that SCAGC consistently outperforms 16 competitive clustering methods.'}",https://arxiv.org/pdf/2110.08264
-Compton Camera and Prompt Gamma Ray Timing: Two Methods for In Vivo Range Assessment in Proton Therapy,F. Hueso-González,"Proton beams are promising means for treating tumors. Such charged particles stop at a defined depth, where the ionization density is maximum. As the dose deposit beyond this distal edge is very low, proton therapy minimizes the damage to normal tissue compared to photon therapy. Nevertheless, inherent range uncertainties cast doubts on the irradiation of tumors close to organs at risk and lead to the application of conservative safety margins. This constrains significantly the potential benefits of protons over photons. In this context, several research groups are developing experimental tools for range verification based on the detection of prompt gammas, a nuclear by-product of the proton irradiation. At OncoRay and Helmholtz-Zentrum Dresden-Rossendorf, detector components have been characterized in realistic radiation environments as a step toward a clinical Compton camera. On the one hand, corresponding experimental methods and results obtained during the ENTERVISION training network are reviewed. On the other hand, a novel method based on timing spectroscopy has been proposed as an alternative to collimated imaging systems. The first tests of the timing method at a clinical proton accelerator are summarized, its applicability in a clinical environment for challenging the current safety margins is assessed, and the factors limiting its precision are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The first tests of the timing method at a clinical proton accelerator are summarized, its applicability in a clinical environment for challenging the current safety margins is assessed, and the factors limiting its precision are discussed.'}",https://www.frontiersin.org/articles/10.3389/fonc.2016.00080/pdf
-Knowledge graph-enhanced molecular contrastive learning with functional prompt,Yin Fang,,"{'model': 'tldr@v2.0.0', 'text': 'A chemical element-oriented knowledge graph is introduced to summarize the basic knowledge of elements and their closely related functional groups to contribute to more efficient drug design by offering a high-quality knowledge prior, interpretable molecular representation and superior prediction performance.'}",https://www.nature.com/articles/s42256-023-00654-0.pdf
-InstructionNER: A Multi-Task Instruction-Based Generative Framework for Few-shot NER,Liwen Wang,"Recently, prompt-based methods have achieved significant performance in few-shot learning scenarios by bridging the gap between language model pre-training and fine-tuning for downstream tasks. However, existing prompt templates are mostly designed for sentence-level tasks and are inappropriate for sequence labeling objectives. To address the above issue, we propose a multi-task instruction-based generative framework, named InstructionNER, for low-resource named entity recognition. Specifically, we reformulate the NER task as a generation problem, which enriches source sentences with task-specific instructions and answer options, then inferences the entities and types in natural language. We further propose two auxiliary tasks, including entity extraction and entity typing, which enable the model to capture more boundary information of entities and deepen the understanding of entity type semantics, respectively. Experimental results show that our method consistently outperforms other baselines on five datasets in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a multi-task instruction-based generative framework, named InstructionNER, for low-resource named entity recognition, which reformulate the NER task as a generation problem, which enriches source sentences with task-specific instructions and answer options, then inferences the entities and types in natural language.'}",http://arxiv.org/pdf/2203.03903
-From prompt gamma distribution to dose: a novel approach combining an evolutionary algorithm and filtering based on Gaussian-powerlaw convolutions,A. Schumann,"Range verification and dose monitoring in proton therapy is considered as highly desirable. Different methods have been developed worldwide, like particle therapy positron emission tomography (PT-PET) and prompt gamma imaging (PGI). In general, these methods allow for a verification of the proton range. However, quantification of the dose from these measurements remains challenging. For the first time, we present an approach for estimating the dose from prompt γ-ray emission profiles. It combines a filtering procedure based on Gaussian-powerlaw convolution with an evolutionary algorithm. By means of convolving depth dose profiles with an appropriate filter kernel, prompt γ-ray depth profiles are obtained. In order to reverse this step, the evolutionary algorithm is applied. The feasibility of this approach is demonstrated for a spread-out Bragg-peak in a water target.","{'model': 'tldr@v2.0.0', 'text': 'For the first time, an approach is presented for estimating the dose from prompt γ-ray emission profiles that combines a filtering procedure based on Gaussian-powerlaw convolution with an evolutionary algorithm.'}",
-Finding Skill Neurons in Pre-trained Transformer-based Language Models,Xiaozhi Wang,"Transformer-based pre-trained language models have demonstrated superior performance on various natural language processing tasks. However, it remains unclear how the skills required to handle these tasks distribute among model parameters. In this paper, we find that after prompt tuning for specific tasks, the activations of some neurons within pre-trained Transformers are highly predictive of the task labels. We dub these neurons skill neurons and confirm they encode task-specific skills by finding that: (1) Skill neurons are crucial for handling tasks. Performances of pre-trained Transformers on a task significantly drop when corresponding skill neurons are perturbed. (2) Skill neurons are task-specific. Similar tasks tend to have similar distributions of skill neurons. Furthermore, we demonstrate the skill neurons are most likely generated in pre-training rather than fine-tuning by showing that the skill neurons found with prompt tuning are also crucial for other fine-tuning methods freezing neuron weights, such as the adapter-based tuning and BitFit. We also explore the applications of skill neurons, including accelerating Transformers with network pruning and building better transferability indicators. These findings may promote further research on understanding Transformers. The source code can be obtained from https://github.com/THU-KEG/Skill-Neuron.","{'model': 'tldr@v2.0.0', 'text': 'It is found that after prompt tuning for specific tasks, the activations of some neurons within pre-trained Transformers are highly predictive of the task labels, and these neurons are dubbed skill neurons and confirmed they encode task-specific skills.'}",https://arxiv.org/pdf/2211.07349
-Model of the Transient Neurovascular Response Based on Prompt Arterial Dilation,J. H. Kim,,"{'model': 'tldr@v2.0.0', 'text': 'This model provides a powerful framework to quantitatively interpret oxygen transport in the brain; in particular, its intravascular oxygen concentration predictions provide a new model for fMRI responses.'}",https://journals.sagepub.com/doi/pdf/10.1038/jcbfm.2013.90
-Analysis framework for the prompt discovery of compact binary mergers in gravitational-wave data,C. Messick,"We describe a stream-based analysis pipeline to detect gravitational waves from the merger of binary neutron stars, binary black holes, and neutron-star–black-hole binaries within ∼1 min of the arrival of the merger signal at Earth. Such low-latency detection is crucial for the prompt response by electromagnetic facilities in order to observe any fading electromagnetic counterparts that might be produced by mergers involving at least one neutron star. Even for systems expected not to produce counterparts, low-latency analysis of the data is useful for deciding when not to point telescopes, and as feedback to observatory operations. Analysts using this pipeline were the first to identify GW151226, the second gravitational-wave event ever detected. The pipeline also operates in an offline mode, in which it incorporates more refined information about data quality and employs acausal methods that are inapplicable to the online mode. The pipeline’s offline mode was used in the detection of the first two gravitational-wave events, GW150914 and GW151226, as well as the identification of a third candidate, LVT151012.",,https://link.aps.org/accepted/10.1103/PhysRevD.95.042001
-Multimodal Few-Shot Object Detection with Meta-Learning Based Cross-Modal Prompting,G. Han,"We study multi-modal few-shot object detection (FSOD) in this paper, using both few-shot visual examples and class semantic information for detection, which are complementary to each other by definition. Most of the previous works on multi-modal FSOD are fine-tuning-based which are inefficient for online applications. Moreover, these methods usually require expertise like class names to extract class semantic embedding, which are hard to get for rare classes. Our approach is motivated by the high-level conceptual similarity of (metric-based) meta-learning and prompt-based learning to learn generalizable few-shot and zero-shot object detection models respectively without fine-tuning. Specifically, we combine the few-shot visual classifier and text classifier learned via meta-learning and prompt-based learning respectively to build the multi-modal classifier and detection models. In addition, to fully exploit the pre-trained language models, we propose meta-learning-based cross-modal prompting to generate soft prompts for novel classes present in few-shot visual examples, which are then used to learn the text classifier. Knowledge distillation is introduced to learn the soft prompt generator without using human prior knowledge of class names, which may not be available for rare classes. Our insight is that the few-shot support images naturally include related context information and semantics of the class. We comprehensively evaluate the proposed multi-modal FSOD models on multiple few-shot object detection benchmarks, achieving promising results.","{'model': 'tldr@v2.0.0', 'text': 'This paper combines the few-shot visual classifier and text classifier learned via meta-learning and prompt-based learning respectively to build the multi-modal classifiers and detection models.'}",http://arxiv.org/pdf/2204.07841
-Good Examples Make A Faster Learner: Simple Demonstration-based Learning for Low-resource NER,Dong-Ho Lee,"Recent advances in prompt-based learning have shown strong results on few-shot text classification by using cloze-style templates.Similar attempts have been made on named entity recognition (NER) which manually design templates to predict entity types for every text span in a sentence. However, such methods may suffer from error propagation induced by entity span detection, high cost due to enumeration of all possible text spans, and omission of inter-dependencies among token labels in a sentence. Here we present a simple demonstration-based learning method for NER, which lets the input be prefaced by task demonstrations for in-context learning. We perform a systematic study on demonstration strategy regarding what to include (entity examples, with or without surrounding context), how to select the examples, and what templates to use. Results on in-domain learning and domain adaptation show that the model’s performance in low-resource settings can be largely improved with a suitable demonstration strategy (e.g., a 4-17% improvement on 25 train instances). We also find that good demonstration can save many labeled examples and consistency in demonstration contributes to better performance.","{'model': 'tldr@v2.0.0', 'text': 'Results on in-domain learning and domain adaptation show that the model’s performance in low-resource settings can be largely improved with a suitable demonstration strategy, and good demonstration can save many labeled examples and consistency in demonstration contributes to better performance.'}",https://aclanthology.org/2022.acl-long.192.pdf
-Inferring prompt black-hole formation in neutron star mergers from gravitational-wave data,M. Agathos,"The gravitational-wave GW170817 is associated to the inspiral phase of a binary neutron star coalescence event. The LIGO-Virgo detectors sensitivity at high frequencies was not sufficient to detect the signal corresponding to the merger and post-merger phases. Hence, the question whether the merger outcome was a prompt black hole formation or not must be answered using either the pre-merger gravitational wave signal or electromagnetic counterparts. In this work we present two methods to infer the probability of prompt black hole formation, using the analysis of the inspiral gravitational-wave signal. Both methods combine the posterior distribution from the gravitational-wave data analysis with numerical relativity results. One method relies on the use of phenomenological models for the equation of state and on the estimate of the collapse threshold mass. The other is based on the estimate of the tidal polarizability parameter $\tilde{\Lambda}$ that is correlated in an equation-of-state agnostic way with the prompt BH formation. We analyze GW170817 data and find that the two methods consistently predict a probability of ~ 50-70% for prompt black-hole formation, which however may significantly decrease below 10% if the maximum mass constraint from PSR J0348+0432 or PSR J0740+6620 is imposed.",,https://arxiv.org/pdf/1908.05442
-Development and applications of k0 based NAA and prompt gamma-ray NAA methods at BARC,R. Acharya,,,
-Methods for scaling simulation-based teamwork training,M. Delisle,"Effective teamwork and communication is now recognised as a critical component of safe and high-quality patient care. Researchers, policymakers and frontline providers are in search of strategies to improve teamwork in healthcare. The most frequently used strategy is teamwork training.1 Teamwork training involves a systematic process in which a team is guided (often by facilitators) to improve and master different skills for working together effectively. Single-centre teamwork training initiatives have demonstrated improvements in patient care, but these results have been challenging to reproduce at scale.2 In this issue of BMJ Quality and Safety, Lenguerrand et al report the results of a stepped-wedge randomised controlled trial in which PRactical Obstetric Multi-Professional Training (PROMPT), an interprofessional intrapartum training package, was implemented across 12 maternity units in Scotland.3 Each participating unit identified an in-house training team to travel to attend a 2-day PROMPT Train the Trainers programme conducted in one simulation centre; two units were unable to send training teams. Training teams were subsequently responsible for coordinating the delivery of in-house PROMPT courses to all maternity staff within 12 months. The courses were intended to cover core obstetrical emergencies, such as postpartum haemorrhage, sepsis, shoulder dystocia, teamwork and fetal monitoring. In addition to clinical outcomes, each maternity unit collected process data about their local PROMPT courses, including the total number of staff trained and courses delivered and the actual course content. The authors found a significant amount of variability in the implementation across units. For example, all courses included elements of teamwork whereas fetal monitoring and shoulder dystocia training were not universally included. Despite the previously demonstrated benefits of PROMPT in single-centre studies, the final results did not demonstrate a reduction of term babies with a low Apgar score. The authors postulate this null finding may be in part related …","{'model': 'tldr@v2.0.0', 'text': 'Despite the previously demonstrated benefits of PROMPT in single-centre studies, the final results did not demonstrate a reduction of term babies with a low Apgar score, and the authors postulate this null finding may be in part related to variability in the implementation across units.'}",https://qualitysafety.bmj.com/content/qhc/29/2/98.full.pdf
-The Utility of an Evidence-Based Lecture and Clinical Prompt as Methods to Improve Quality of Care in Colorectal Cancer Screening,K. Seres,,"{'model': 'tldr@v2.0.0', 'text': 'Clinical prompts are superior to evidence-based lectures for improving physician colorectal cancer screening practices and are simple low-cost measures that can improve quality of care.'}",
-Machine learning-based prediction of acute coronary syndrome using only the pre-hospital 12-lead electrocardiogram,S. Al-Zaiti,,"{'model': 'tldr@v2.0.0', 'text': 'A machine learning-based system based exclusively on ECG data that can help clinicians identify 37% more heart attacks during initial screening and reduce unnecessary costs in patients with chest pain is reported.'}",https://www.nature.com/articles/s41467-020-17804-2.pdf
-Plasma paracetamol concentration at hospital presentation has a dose-dependent relationship with liver injury despite prompt treatment with intravenous acetylcysteine,David G. Cairney,"ABSTRACT Context: Paracetamol (acetaminophen) overdose is a common reason for emergency hospital admission in the UK and the leading cause of acute liver failure in the Western world. Currently, the antidote acetylcysteine (NAC) is administered at a dose determined only by body weight without regard for the body burden of paracetamol. Objective: To determine whether higher plasma paracetamol concentrations are associated with increased risk of liver injury despite prompt treatment with intravenous NAC. Methods: Patients admitted to hospital for treatment with intravenous NAC following a single acute paracetamol overdose entered the study if NAC was commenced within 24 h of drug ingestion (N = 727 hospital presentations). Based on the plasma paracetamol concentration at first presentation to hospital, a series of nomograms were created: 0–100, 101–150, 151–200, 201–300, 301–500 and over 501 mg/L. The primary endpoints were acute liver injury (ALI – peak serum ALT activity >150 U/L and double the admission value) and hepatotoxicity (peak ALT >1000 U/L). Results: ALI and hepatotoxicity were more common in patients with higher admission plasma paracetamol concentrations despite NAC treatment (ALI: nomogram 0–100: 6%, 101–150: 3%, 151–200: 3%, 201–300: 9%, 301–500: 13%, over 501 mg/dL: 27%. p < 0.0001). This dose-response relationship between paracetamol concentration and ALI persisted even in patients treated with NAC within 8 h of overdose (nomogram 0–100: 0%, 101–150: 0.8%, 151–200: 2%, 201–300: 3.6%, 301–500: 12.5%, over 501mg/L: 33%. p < 0.0001) and in patients with normal ALT activity at first presentation (nomogram: 0–100: 0%, 101–150: 1.2%, 151–200: 1.5%, 201–300: 5.3%, 301–500: 10.8% p < 0.0001). Discussion: Patients with increased concentrations of plasma paracetamol at hospital presentation are at higher risk of liver injury even when intravenous NAC is promptly administered before there is biochemical evidence of toxicity. Conclusion: This study supports theoretical concerns that the current intravenous dose of NAC may be too low in the setting of higher paracetamol exposure.","{'model': 'tldr@v2.0.0', 'text': 'Patients with increased concentrations of plasma paracetamol at hospital presentation are at higher risk of liver injury even when intravenous NAC is promptly administered before there is biochemical evidence of toxicity, which supports theoretical concerns that the current intravenous dose of NAC may be too low in the setting of higher par acetamol exposure.'}",
-Real‐time regional seismic damage assessment framework based on long short‐term memory neural network,Yongjia Xu,"Effective post‐earthquake response requires a prompt and accurate assessment of earthquake‐induced damage. However, existing damage assessment methods cannot simultaneously meet these requirements. This study proposes a framework for real‐time regional seismic damage assessment that is based on a Long Short‐Term Memory (LSTM) neural network architecture. The proposed framework is not specially designed for individual structural types, but offers rapid estimates at regional scale. The framework is built around a workflow that establishes high‐performance mapping rules between ground motions and structural damage via region‐specific models. This workflow comprises three main parts—namely, region‐specific database generation, LSTM model training and verification, and model utilization for damage prediction. The influence of various LSTM architectures, hyperparameter selection, and dataset resampling procedures are systematically analyzed. As a testbed for the established framework, a case study is performed on the Tsinghua University campus buildings. The results demonstrate that the developed LSTM framework can perform damage assessment in real time at regional scale with high prediction accuracy and acceptable variance.","{'model': 'tldr@v2.0.0', 'text': 'A framework for real‐time regional seismic damage assessment that is based on a Long Short‐Term Memory (LSTM) neural network architecture is proposed and can perform damage assessment in real time at regional scale with high prediction accuracy and acceptable variance.'}",
-RANIBIZUMAB PLUS PROMPT OR DEFERRED LASER FOR DIABETIC MACULAR EDEMA IN EYES WITH VITRECTOMY BEFORE ANTI-VASCULAR ENDOTHELIAL GROWTH FACTOR THERAPY,S. Bressler,"Background: The approach to managing diabetic macular edema in eyes with previous vitrectomy is based on limited evidence. Therefore, an exploratory post hoc assessment of 3-year data from eyes with and without vitrectomy before randomization in a DRCR.net trial that evaluated ranibizumab + prompt or deferred laser for diabetic macular edema is presented. Methods: Visual acuity and optical coherence tomography outcomes were compared between eyes with and without previous vitrectomy. Results: At baseline, eyes with previous vitrectomy (n = 25) had longer duration of diabetes, worse visual acuity, less thickened central subfield measurements on optical coherence tomography and were more apt to have worse diabetic retinopathy severity level or previous treatment for macular edema or cataract surgery than eyes without a history of vitrectomy (n = 335). Analyses adjusted for these baseline imbalances did not identify substantial differences between eyes with and without previous vitrectomy at each annual visit through 3 years for the favorable visual acuity, optical coherence tomography central subfield thickness, or volume outcomes, although optical coherence tomography improvement appeared slower in vitrectomy eyes during the first year. Conclusion: This study provides little evidence that the beneficial clinical outcomes for patients with center-involved diabetic macular edema treated with anti-vascular endothelial growth factor are affected in the long term by previous vitrectomy.","{'model': 'tldr@v2.0.0', 'text': 'This study provides little evidence that the beneficial clinical outcomes for patients with center-involved diabetic macular edema treated with anti-vascular endothelial growth factor are affected in the long term by previous vitrectomy.'}",https://europepmc.org/articles/pmc4658280?pdf=render
-Making Pre-trained Language Models Better Few-shot Learners,Tianyu Gao,"The recent GPT-3 model (Brown et al., 2020) achieves remarkable few-shot performance solely by leveraging a natural-language prompt and a few task demonstrations as input context. Inspired by their findings, we study few-shot learning in a more practical scenario, where we use smaller language models for which fine-tuning is computationally efficient. We present LM-BFF—better few-shot fine-tuning of language models—a suite of simple and complementary techniques for fine-tuning language models on a small number of annotated examples. Our approach includes (1) prompt-based fine-tuning together with a novel pipeline for automating prompt generation; and (2) a refined strategy for dynamically and selectively incorporating demonstrations into each context. Finally, we present a systematic evaluation for analyzing few-shot performance on a range of NLP tasks, including classification and regression. Our experiments demonstrate that our methods combine to dramatically outperform standard fine-tuning procedures in this low resource setting, achieving up to 30% absolute improvement, and 11% on average across all tasks. Our approach makes minimal assumptions on task resources and domain expertise, and hence constitutes a strong task-agnostic method for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'The LM-BFF approach makes minimal assumptions on task resources and domain expertise, and hence constitutes a strong task-agnostic method for few-shot learning.'}",https://aclanthology.org/2021.acl-long.295.pdf
-Validation of Deep Convolutional Neural Network-based algorithm for detection of diabetic retinopathy – Artificial intelligence versus clinician for screening,P. Shah,"Purpose: Deep learning is a newer and advanced subfield in artificial intelligence (AI). The aim of our study is to validate a machine-based algorithm developed based on deep convolutional neural networks as a tool for screening to detect referable diabetic retinopathy (DR). Methods: An AI algorithm to detect DR was validated at our hospital using an internal dataset consisting of 1,533 macula-centered fundus images collected retrospectively and an external validation set using Methods to Evaluate Segmentation and Indexing Techniques in the field of Retinal Ophthalmology (MESSIDOR) dataset. Images were graded by two retina specialists as any DR, prompt referral (moderate nonproliferative diabetic retinopathy (NPDR) or above or presence of macular edema) and sight-threatening DR/STDR (severe NPDR or above) and compared with AI results. Sensitivity, specificity, and area under curve (AUC) for both internal and external validation sets for any DR detection, prompt referral, and STDR were calculated. Interobserver agreement using kappa value was calculated for both the sets and two out of three agreements for DR grading was considered as ground truth to compare with AI results. Results: In the internal validation set, the overall sensitivity and specificity was 99.7% and 98.5% for Any DR detection and 98.9% and 94.84%for Prompt referral respectively. The AUC was 0.991 and 0.969 for any DR detection and prompt referral respectively. The agreement between two observers was 99.5% and 99.2% for any DR detection and prompt referral with a kappa value of 0.94 and 0.96, respectively. In the external validation set (MESSIDOR 1), the overall sensitivity and specificity was 90.4% and 91.0% for any DR detection and 94.7% and 97.4% for prompt referral, respectively. The AUC was. 907 and. 960 for any DR detection and prompt referral, respectively. The agreement between two observers was 98.5% and 97.8% for any DR detection and prompt referral with a kappa value of 0.971 and 0.980, respectively. Conclusion: With increasing diabetic population and growing demand supply gap in trained resources, AI is the future for early identification of DR and reducing blindness. This can revolutionize telescreening in ophthalmology, especially where people do not have access to specialized health care.","{'model': 'tldr@v2.0.0', 'text': 'With increasing diabetic population and growing demand supply gap in trained resources, AI is the future for early identification of DR and reducing blindness, especially where people do not have access to specialized health care.'}",
-Analytical computation of prompt gamma ray emission and detection for proton range verification,E. Sterpin,"A prompt gamma (PG) slit camera prototype recently demonstrated that Bragg Peak position in a clinical proton scanned beam could be measured with 1–2 mm accuracy by comparing an expected PG detection profile to a measured one. The computation of the expected PG detection profile in the context of a clinical framework is challenging but must be solved before clinical implementation. Obviously, Monte Carlo methods (MC) can simulate the expected PG profile but at prohibitively long calculation times. We implemented a much faster method that is based on analytical processing of precomputed MC data that would allow practical evaluation of this range monitoring approach in clinical conditions. Reference PG emission profiles were generated with MC simulations (PENH) in targets consisting of either 12C, 14N, 16O, 31P or 40Ca, with 10% of 1H. In a given geometry, the local PG emission can then be derived by adding the contribution of each element, according to the local energy of the proton obtained by continuous slowing down approximation and the local composition. The actual incident spot size is taken into account using an optical model fitted to measurements and by super sampling the spot with several rays (up to 113). PG transport in the patient/camera geometries and the detector response are modelled by convolving the PG production profile with a transfer function. The latter is interpolated from a database of transfer functions fitted to MC data (PENELOPE) generated for a photon source in a cylindrical phantom with various radiuses and a camera placed at various positions. As a benchmark, the analytical model was compared to MC and experiments in homogeneous and heterogeneous phantoms. Comparisons with MC were also performed in a thoracic CT. For all cases, the analytical model reproduced the prediction of the position of the Bragg peak computed with MC within 1 mm for the camera in nominal configuration. When compared to measurements, the shape of the profiles was well reproduced and agreement for the estimation of the position of the Bragg peak was within 2.7 mm on average (1.4 mm standard deviation). On a non-optimized MATLAB code, computation time with the analytical model is between 0.3 to 10 s depending on the number of rays simulated per spot. The analytical model can be further used to determine which spots are the best candidates to evaluate the range in clinical conditions and eventually correct for over- and under-shoots depending on the acquired PG profiles.","{'model': 'tldr@v2.0.0', 'text': 'A much faster method that is based on analytical processing of precomputed MC data that would allow practical evaluation of this range monitoring approach in clinical conditions and eventually correct for over- and under-shoots depending on the acquired PG profiles is implemented.'}",
-"Who Gets Prompt Access to Artemisinin-Based Combination Therapy? A Prospective Community-Based Study in Children from Rural Kilosa, Tanzania",D. Simba,"Background Effective and timely case management remains one of the fundamental pillars for control of malaria. Tanzania introduced artemisinin-combination therapy [ACT] for uncomplicated malaria; however, the policy change is challenged by limited availability of ACTs due to high cost. This study aimed to determine factors influencing prompt access to ACTs among febrile children in rural Kilosa, Tanzania. Methods and Findings In a community-based study, 1,235 randomly selected children under five were followed up weekly for six months, in 2008. Using a structured questionnaire, children's caretakers were asked about the child's febrile history in the last seven days, and treatment actions including timing, medicines used and source of care. Caretakers' knowledge about malaria and socioeconomic and demographic data were also obtained. About half of followed-up children had at least one episode of fever. Less than half (44.8%) of febrile children were taken to government facilities. Almost one-third (37.6%; 95% CI 33.1–42.1) of febrile children had prompt access to ACT. Care-seeking from a government facility was the overriding factor, increasing the likelihood of prompt access to an ACT 18 times (OR 17.7; 95% CI 10.55–29.54; adjusted OR 16.9; 95% CI 10.06–28.28). Caretakers from the better-off household (3rd–5th quintiles) were more likely to seek care from government facilities (OR 3.66; 95% CI 2.56–5.24; adjusted OR 1.80; 95% CI 1.18–2.76). The majority of antimalarials accessed by the poor were ineffective [86.0%; 295/343], however, they paid more for them (median Tsh 500) compared to the better-offs (median Tsh 0). Conclusions Prompt access to ACT among febrile children was unacceptably low, due mainly to limited availability of subsidised ACT at the location where most caretakers sought care. There is urgent need to accelerate implementation of strategies that will ensure availability of ACT at an affordable price in remote rural areas, where the burden of malaria is highest.","{'model': 'tldr@v2.0.0', 'text': None}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0012104&type=printable
-Zero-Shot Information Extraction via Chatting with ChatGPT,Xiang Wei,"Zero-shot information extraction (IE) aims to build IE systems from the unannotated text. It is challenging due to involving little human intervention. Challenging but worthwhile, zero-shot IE reduces the time and effort that data labeling takes. Recent efforts on large language models (LLMs, e.g., GPT-3, ChatGPT) show promising performance on zero-shot settings, thus inspiring us to explore prompt-based methods. In this work, we ask whether strong IE models can be constructed by directly prompting LLMs. Specifically, we transform the zero-shot IE task into a multi-turn question-answering problem with a two-stage framework (ChatIE). With the power of ChatGPT, we extensively evaluate our framework on three IE tasks: entity-relation triple extract, named entity recognition, and event extraction. Empirical results on six datasets across two languages show that ChatIE achieves impressive performance and even surpasses some full-shot models on several datasets (e.g., NYT11-HRL). We believe that our work could shed light on building IE models with limited resources.","{'model': 'tldr@v2.0.0', 'text': 'This work transforms the zero-shot IE task into a multi-turn question-answering problem with a two-stage framework (ChatIE), and extensively evaluates the framework on three IE tasks: entity-relation triple extract, named entity recognition, and event extraction.'}",http://arxiv.org/pdf/2302.10205
-European consensus-based recommendations for the diagnosis and treatment of Kawasaki disease – the SHARE initiative,N. de Graeff,"OBJECTIVES The European Single Hub and Access point for paediatric Rheumatology in Europe initiative aimed to optimize care for children with rheumatic diseases. Kawasaki disease (KD) is the most common cause of acquired heart disease in children and an important cause of long-term cardiac disease into adulthood. Prompt diagnosis and treatment of KD is difficult due to the heterogeneity of the disease but is crucial for improving outcome. To date, there are no European internationally agreed, evidence-based guidelines concerning the diagnosis and treatment of KD in children. Accordingly, treatment regimens differ widely. The aim of this study is to provide consensus-based, European-wide evidence-informed recommendations for diagnosis and treatment of children with KD. METHODS Recommendations were developed using the EULAR's standard operating procedures. An extensive systematic literature search was performed, and evidence-based recommendations were extrapolated from the included papers. These were evaluated by a panel of international experts via online surveys and subsequently discussed in three consensus meetings, using nominal group technique. Recommendations were accepted when ⩾80% agreed. RESULTS In total, 17 recommendations for diagnosis and 14 for treatment of KD in children were accepted. Diagnostic recommendations included laboratory and imaging workup for complete as well as incomplete KD. Treatment recommendations included the importance of early treatment in both complete and incomplete KD, use of intravenous immunoglobulin, aspirin, corticosteroids for high-risk cases, and other treatment options for those with resistant disease. CONCLUSION The Single Hub and Access point for paediatric Rheumatology in Europe initiative provides international evidence-based recommendations for diagnosing and treating KD in children, facilitating improvement and uniformity of care.","{'model': 'tldr@v2.0.0', 'text': 'The Single Hub and Access point for paediatric Rheumatology in Europe initiative provides international evidence-based recommendations for diagnosing and treating Kawasaki disease in children, facilitating improvement and uniformity of care.'}",https://discovery.ucl.ac.uk/10072207/1/SHARE%20KD%20Recommendations%20-%20Rheumatology%20Manuscript%20-%20270618%20-%20Final.pdf
-Automated Essay Scoring based on Two-Stage Learning,Jiawei Liu,"Current state-of-art feature-engineered and end-to-end Automated Essay Score (AES) methods are proven to be unable to detect adversarial samples, e.g. the essays composed of permuted sentences and the prompt-irrelevant essays. Focusing on the problem, we develop a Two-Stage Learning Framework (TSLF) which integrates the advantages of both feature-engineered and end-to-end AES models. In experiments, we compare TSLF against a number of strong baselines, and the results demonstrate the effectiveness and robustness of our models. TSLF surpasses all the baselines on five-eighths of prompts and achieves new state-of-the-art average performance when without negative samples. After adding some adversarial essays to the original datasets, TSLF outperforms the feature-engineered and end-to-end baselines to a great extent, and shows great robustness.","{'model': 'tldr@v2.0.0', 'text': 'After adding some adversarial essays to the original datasets, TSLF outperforms the feature-engineered and end-to-end baselines to a great extent, and shows great robustness.'}",
-MathPrompter: Mathematical Reasoning using Large Language Models,Shima Imani,"Large Language Models (LLMs) have limited performance when solving arithmetic reasoning tasks and often provide incorrect answers. Unlike natural language understanding, math problems typically have a single correct answer, making the task of generating accurate solutions more challenging for LLMs. To the best of our knowledge, we are not aware of any LLMs that indicate their level of confidence in their responses which fuels a trust deficit in these models impeding their adoption. To address this deficiency, we propose ‘MathPrompter’, a technique that improves performance of LLMs on arithmetic problems along with increased reliance in the predictions. MathPrompter uses the Zero-shot chain-of-thought prompting technique to generate multiple algebraic expressions or python functions to solve the same math problem in different ways and thereby raise the confidence level in the output results. This is in contrast to other prompt based CoT methods, where there is no check on the validity of the intermediate steps followed. Our technique improves over state-of-the-art on the ‘MultiArith’ dataset (78.7% - 92.5%) evaluated using 175B parameter GPT-based LLM.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ‘MathPrompter’, a technique that improves performance of LLMs on arithmetic problems along with increased reliance in the predictions, using the Zero-shot chain-of-thought prompting technique to generate multiple algebraic expressions or python functions to solve the same math problem in different ways and thereby raise the confidence level in the output results.'}",http://arxiv.org/pdf/2303.05398
-PromptBERT: Improving BERT Sentence Embeddings with Prompts,Ting Jiang,"We propose PromptBERT, a novel contrastive learning method for learning better sentence representation. We firstly analysis the drawback of current sentence embedding from original BERT and find that it is mainly due to the static token embedding bias and ineffective BERT layers. Then we propose the first prompt-based sentence embeddings method and discuss two prompt representing methods and three prompt searching methods to make BERT achieve better sentence embeddings .Moreover, we propose a novel unsupervised training objective by the technology of template denoising, which substantially shortens the performance gap between the supervised and unsupervised settings. Extensive experiments show the effectiveness of our method. Compared to SimCSE, PromptBert achieves 2.29 and 2.58 points of improvement based on BERT and RoBERTa in the unsupervised setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel unsupervised training objective is proposed by the technology of template denoising, which substantially shortens the performance gap between the supervised and unsuper supervised settings.'}",https://aclanthology.org/2022.emnlp-main.603.pdf
-"Fetal and neonatal alloimmune thrombocytopenia: recommendations for evidence‐based practice, an international approach",L. Lieberman,"Fetal and neonatal alloimmune thrombocytopenia (FNAIT) may result in severe bleeding, particularly fetal and neonatal intracranial haemorrhage (ICH). As a result, FNAIT requires prompt identification and treatment; subsequent pregnancies need close surveillance and management. An international panel convened to develop evidence‐based recommendations for diagnosis and management of FNAIT. A rigorous approach was used to search, review and develop recommendations from published data for: antenatal management, postnatal management, diagnostic testing and universal screening. To confirm FNAIT, fetal human platelet antigen (HPA) typing, using non‐invasive methods if quality‐assured, should be performed during pregnancy when the father is unknown, unavailable for testing or heterozygous for the implicated antigen. Women with a previous child with an ICH related to FNAIT should be offered intravenous immunoglobulin (IVIG) infusions during subsequent affected pregnancies as early as 12 weeks gestation. Ideally, HPA‐selected platelets should be available at delivery for potentially affected infants and used to increase the neonatal platelet count as needed. If HPA‐selected platelets are not immediately available, unselected platelets should be transfused. FNAIT studies that optimize antenatal and postnatal management, develop risk stratification algorithms to guide management and standardize laboratory testing to identify high risk pregnancies are needed.","{'model': 'tldr@v2.0.0', 'text': 'FNAIT studies that optimize antenatal and postnatal management, develop risk stratification algorithms to guide management and standardize laboratory testing to identify high risk pregnancies are needed.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/bjh.15813
-How Can We Know What Language Models Know?,Zhengbao Jiang,"Abstract Recent work has presented intriguing results examining the knowledge contained in language models (LMs) by having the LM fill in the blanks of prompts such as “Obama is a __ by profession”. These prompts are usually manually created, and quite possibly sub-optimal; another prompt such as “Obama worked as a __ ” may result in more accurately predicting the correct profession. Because of this, given an inappropriate prompt, we might fail to retrieve facts that the LM does know, and thus any given prompt only provides a lower bound estimate of the knowledge contained in an LM. In this paper, we attempt to more accurately estimate the knowledge contained in LMs by automatically discovering better prompts to use in this querying process. Specifically, we propose mining-based and paraphrasing-based methods to automatically generate high-quality and diverse prompts, as well as ensemble methods to combine answers from different prompts. Extensive experiments on the LAMA benchmark for extracting relational knowledge from LMs demonstrate that our methods can improve accuracy from 31.1% to 39.6%, providing a tighter lower bound on what LMs know. We have released the code and the resulting LM Prompt And Query Archive (LPAQA) at https://github.com/jzbjyb/LPAQA.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes mining-based and paraphrasing-based methods to automatically generate high-quality and diverse prompts, as well as ensemble methods to combine answers from different prompts to provide a tighter lower bound on what LMs know.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00324/1923867/tacl_a_00324.pdf
-RealFusion 360° Reconstruction of Any Object from a Single Image,Luke Melas-Kyriazi,"We consider the problem of reconstructing a full 360° photographic model of an object from a single image of it. We do so by fitting a neural radiance field to the image, but find this problem to be severely ill-posed. We thus take an off-the-self conditional image generator based on diffusion and engineer a prompt that encourages it to “dream up” novel views of the object. Using the recent DreamFusion method, we fuse the given input view, the conditional prior, and other regularizers into a final, consistent reconstruction. We demonstrate state-of-the-art reconstruction results on benchmark images when compared to prior methods for monocular 3D reconstruction of objects. Qualitatively, our reconstructions provide a faithful match of the input view and a plausible extrapolation of its appearance and 3D shape, including to the side of the object not visible in the image.",,https://arxiv.org/pdf/2302.10663
-Active Prompting with Chain-of-Thought for Large Language Models,Shizhe Diao,"The increasing scale of large language models (LLMs) brings emergent abilities to various complex tasks requiring reasoning, such as arithmetic and commonsense reasoning. It is known that the effective design of task-specific prompts is critical for LLMs' ability to produce high-quality answers. In particular, an effective approach for complex question-and-answer tasks is example-based prompting with chain-of-thought (CoT) reasoning, which significantly improves the performance of LLMs. However, current CoT methods rely on a fixed set of human-annotated exemplars, which are not necessarily the most effective examples for different tasks. This paper proposes a new method, Active-Prompt, to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning). For this purpose, we propose a solution to the key problem of determining which questions are the most important and helpful ones to annotate from a pool of task-specific queries. By borrowing ideas from the related problem of uncertainty-based active learning, we introduce several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation. Experimental results demonstrate the superiority of our proposed method, achieving state-of-the-art on eight complex reasoning tasks. Further analyses of different uncertainty metrics, pool sizes, zero-shot learning, and accuracy-uncertainty relationship demonstrate the effectiveness of our method. Our code will be available at https://github.com/shizhediao/active-prompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new method to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning), and introduces several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation.'}",http://arxiv.org/pdf/2302.12246
-Selective Annotation Makes Language Models Better Few-Shot Learners,Hongjin Su,"Many recent approaches to natural language tasks are built on the remarkable abilities of large language models. Large language models can perform in-context learning, where they learn a new task from a few task demonstrations, without any parameter updates. This work examines the implications of in-context learning for the creation of datasets for new natural language tasks. Departing from recent in-context learning methods, we formulate an annotation-efficient, two-step framework: selective annotation that chooses a pool of examples to annotate from unlabeled data in advance, followed by prompt retrieval that retrieves task examples from the annotated pool at test time. Based on this framework, we propose an unsupervised, graph-based selective annotation method, voke-k, to select diverse, representative examples to annotate. Extensive experiments on 10 datasets (covering classification, commonsense reasoning, dialogue, and text/code generation) demonstrate that our selective annotation method improves the task performance by a large margin. On average, vote-k achieves a 12.9%/11.4% relative gain under an annotation budget of 18/100, as compared to randomly selecting examples to annotate. Compared to state-of-the-art supervised finetuning approaches, it yields similar performance with 10-100x less annotation cost across 10 tasks. We further analyze the effectiveness of our framework in various scenarios: language models with varying sizes, alternative selective annotation methods, and cases where there is a test data domain shift. We hope that our studies will serve as a basis for data annotations as large language models are increasingly applied to new tasks. Our code is available at https://github.com/HKUNLP/icl-selective-annotation.","{'model': 'tldr@v2.0.0', 'text': 'This work forms an annotation-efficient, two-step framework: selective annotation that chooses a pool of examples to annotate from unlabeled data in advance, followed by prompt retrieval that retrieves task examples from the annotated pool at test time, and proposes an unsupervised, graph-based selective annotation method, voke-k, to select diverse, representative examples to annotation.'}",http://arxiv.org/pdf/2209.01975
-Blended Diffusion for Text-driven Editing of Natural Images,Omri Avrahami,"Natural language offers a highly intuitive interface for image editing. In this paper, we introduce the first solution for performing local (region-based) edits in generic natural images, based on a natural language description along with an ROI mask. We achieve our goal by leveraging and combining a pretrained language-image model (CLIP), to steer the edit towards a user-provided text prompt, with a denoising diffusion probabilistic model (DDPM) to generate natural-looking results. To seamlessly fuse the edited region with the unchanged parts of the image, we spatially blend noised versions of the input image with the local text-guided diffusion latent at a progression of noise levels. In addition, we show that adding augmentations to the diffusion process mitigates adversarial results. We compare against several baselines and related methods, both qualitatively and quantitatively, and show that our method outperforms these solutions in terms of overall realism, ability to preserve the background and matching the text. Finally, we show several text-driven editing applications, including adding a new object to an image, removing/replacing/altering existing objects, background replacement, and image extrapolation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the first solution for performing local (region-based) edits in generic natural images, based on a natural language description along with an ROI mask, and shows several text-driven editing applications, including adding a new object to an image, removing/replacing/altering existing objects, background replacement, and image extrapolation.'}",https://arxiv.org/pdf/2111.14818
-"Diagnosis, Treatment, and Long-Term Management of Kawasaki Disease: A Scientific Statement for Health Professionals From the American Heart Association",B. McCrindle,"Background: Kawasaki disease is an acute vasculitis of childhood that leads to coronary artery aneurysms in ≈25% of untreated cases. It has been reported worldwide and is the leading cause of acquired heart disease in children in developed countries. Methods and Results: To revise the previous American Heart Association guidelines, a multidisciplinary writing group of experts was convened to review and appraise available evidence and practice-based opinion, as well as to provide updated recommendations for diagnosis, treatment of the acute illness, and long-term management. Although the cause remains unknown, discussion sections highlight new insights into the epidemiology, genetics, pathogenesis, pathology, natural history, and long-term outcomes. Prompt diagnosis is essential, and an updated algorithm defines supplemental information to be used to assist the diagnosis when classic clinical criteria are incomplete. Although intravenous immune globulin is the mainstay of initial treatment, the role for additional primary therapy in selected patients is discussed. Approximately 10% to 20% of patients do not respond to initial intravenous immune globulin, and recommendations for additional therapies are provided. Careful initial management of evolving coronary artery abnormalities is essential, necessitating an increased frequency of assessments and escalation of thromboprophylaxis. Risk stratification for long-term management is based primarily on maximal coronary artery luminal dimensions, normalized as Z scores, and is calibrated to both past and current involvement. Patients with aneurysms require life-long and uninterrupted cardiology follow-up. Conclusions: These recommendations provide updated and best evidence-based guidance to healthcare providers who diagnose and manage Kawasaki disease, but clinical decision making should be individualized to specific patient circumstances.","{'model': 'tldr@v2.0.0', 'text': 'These recommendations provide updated and best evidence-based guidance to healthcare providers who diagnose and manage Kawasaki disease, but clinical decision making should be individualized to specific patient circumstances.'}",https://www.ahajournals.org/doi/pdf/10.1161/CIR.0000000000000484
-GPT3Mix: Leveraging Large-scale Language Models for Text Augmentation,Kang Min Yoo,"Large-scale language models such as GPT-3 are excellent few-shot learners, allowing them to be controlled via natural text prompts. Recent studies report that prompt-based direct classification eliminates the need for fine-tuning but lacks data and inference scalability. This paper proposes a novel data augmentation technique that leverages large-scale language models to generate realistic text samples from a mixture of real samples. We also propose utilizing soft-labels predicted by the language models, effectively distilling knowledge from the large-scale language models and creating textual perturbations simultaneously. We perform data augmentation experiments on diverse classification tasks and show that our method hugely outperforms existing text augmentation methods. Ablation studies and a qualitative analysis provide more insights into our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel data augmentation technique that leverages large-scale language models to generate realistic text samples from a mixture of real samples and utilizes soft-labels predicted by the language models, effectively distilling knowledge from the large- scale language models and creating textual perturbations simultaneously.'}",https://aclanthology.org/2021.findings-emnlp.192.pdf
-Optimizing rating scale category effectiveness.,J. Linacre,"Rating scales are employed as a means of extracting more information out of an item than would be obtained from a mere ""yes/no"", ""right/wrong"" or other dichotomy. But does this additional information increase measurement accuracy and precision? Eight guidelines are suggested to aid the analyst in optimizing the manner in which rating scales categories cooperate in order to improve the utility of the resultant measures. Though these guidelines are presented within the context of Rasch analysis, they reflect aspects of rating scale functioning which impact all methods of analysis. The guidelines feature rating-scale-based data such as category frequency, ordering, rating-to-measure inferential coherence, and the quality of the scale from measurement and statistical perspectives. The manner in which the guidelines prompt recategorization or reconceptualization of the rating scale is indicated. Utilization of the guidelines is illustrated through their application to two published data sets.","{'model': 'tldr@v2.0.0', 'text': 'Eight guidelines are suggested to aid the analyst in optimizing the manner in which rating scales categories cooperate in order to improve the utility of the resultant measures.'}",
-WARP: Word-level Adversarial ReProgramming,Karen Hambardzumyan,"Transfer learning from pretrained language models recently became the dominant approach for solving many NLP tasks. A common approach to transfer learning for multiple tasks that maximize parameter sharing trains one or more task-specific layers on top of the language model. In this paper, we present an alternative approach based on adversarial reprogramming, which extends earlier work on automatic prompt generation. Adversarial reprogramming attempts to learn task-specific word embeddings that, when concatenated to the input text, instruct the language model to solve the specified task. Using up to 25K trainable parameters per task, this approach outperforms all existing methods with up to 25M trainable parameters on the public leaderboard of the GLUE benchmark. Our method, initialized with task-specific human-readable prompts, also works in a few-shot setting, outperforming GPT-3 on two SuperGLUE tasks with just 32 training samples.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents an alternative approach based on adversarial reprogramming, which extends earlier work on automatic prompt generation, and outperforms all existing methods with up to 25M trainable parameters on the public leaderboard of the GLUE benchmark.'}",https://aclanthology.org/2021.acl-long.381.pdf
-An automated essay scoring systems: a systematic literature review,D. Ramesh,,"{'model': 'tldr@v2.0.0', 'text': 'It is observed that the essay evaluation is not done based on the relevance of the content and coherence, which means that there is no proper evaluation system for grading essays and short answers.'}",https://link.springer.com/content/pdf/10.1007/s10462-021-10068-2.pdf
-Prompting for Multimodal Hateful Meme Classification,Rui Cao,"Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge. Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes. However, there is no known explicit external knowledge base that could provide such hate speech contextual information. To address this gap, we propose PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification. Specifically, we construct simple prompts and provide a few in-context examples to exploit the implicit knowledge in the pre-trained RoBERTa language model for hateful meme classification. We conduct extensive experiments on two publicly available hateful and offensive meme datasets. Our experiment results show that PromptHate is able to achieve a high AUC of 90.96, outperforming state-of-the-art baselines on the hateful meme classification task. We also perform fine-grain analyses and case studies on various prompt settings and demonstrate the effectiveness of the prompts on hateful meme classification.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification, and constructs simple prompts and provides a few in-context examples to exploit the implicit knowledge in the pre- trained RoBERTa language model for hateful memes classification.'}",http://arxiv.org/pdf/2302.04156
-μEMA: Microinteraction-based ecological momentary assessment (EMA) using a smartwatch,S. Intille,"Ecological Momentary Assessment (EMA) is a method of in situ data collection for assessment of behaviors, states, and contexts. Questions are prompted during everyday life using an individual's mobile device, thereby reducing recall bias and increasing validity over other self-report methods such as retrospective recall. We describe a microinteraction-based EMA method (""micro"" EMA, or μEMA) using smartwatches, where all EMA questions can be answered with a quick glance and a tap -- nearly as quickly as checking the time on a watch. A between-subjects, 4-week pilot study was conducted where μEMA on a smartwatch (n=19) was compared with EMA on a phone (n=14). Despite an =8 times increase in the number of interruptions, μEMA had a significantly higher compliance rate, completion rate, and first prompt response rate, and μEMA was perceived as less distracting. The temporal density of data collection possible with μEMA could prove useful in ubiquitous computing studies.","{'model': 'tldr@v2.0.0', 'text': 'A microinteraction-based EMA method (""micro"" EMA, or μEMA) using smartwatches, where all EMA questions can be answered with a quick glance and a tap -- nearly as quickly as checking the time on a watch.'}",https://europepmc.org/articles/pmc6143290?pdf=render
-Learning to Transfer Prompts for Text Generation,Junyi Li,"Pretrained language models (PLMs) have made remarkable progress in text generation tasks via fine-tuning. While, it is challenging to fine-tune PLMs in a data-scarce situation. Therefore, it is non-trivial to develop a general and lightweight model that can adapt to various text generation tasks based on PLMs. To fulfill this purpose, the recent prompt-based learning offers a potential solution. In this paper, we improve this technique and propose a novel prompt-based method (PTG) for text generation in a transferable setting. First, PTG learns a set of source prompts for various source generation tasks and then transfers these prompts as target prompts to perform target generation tasks. To consider both task- and instance-level information, we design an adaptive attention mechanism to derive the target prompts. For each data instance, PTG learns a specific target prompt by attending to highly relevant source prompts. In extensive experiments, PTG yields competitive or better results than fine-tuning methods. We release our source prompts as an open resource, where users can add or reuse them to improve new text generation tasks for future research. Code and data can be available at https://github.com/RUCAIBox/Transfer-Prompts-for-Text-Generation.","{'model': 'tldr@v2.0.0', 'text': 'This paper improves this technique and proposes a novel prompt-based method (PTG) for text generation in a transferable setting that learns a set of source prompts for various source generation tasks and then transfers these prompts as target prompts to perform target generation tasks.'}",https://arxiv.org/pdf/2205.01543
-BadPrompt: Backdoor Attacks on Continuous Prompts,Xiangrui Cai,"The prompt-based learning paradigm has gained much research attention recently. It has achieved state-of-the-art performance on several NLP tasks, especially in the few-shot scenarios. While steering the downstream tasks, few works have been reported to investigate the security problems of the prompt-based models. In this paper, we conduct the first study on the vulnerability of the continuous prompt learning algorithm to backdoor attacks. We observe that the few-shot scenarios have posed a great challenge to backdoor attacks on the prompt-based models, limiting the usability of existing NLP backdoor methods. To address this challenge, we propose BadPrompt, a lightweight and task-adaptive algorithm, to backdoor attack continuous prompts. Specially, BadPrompt first generates candidate triggers which are indicative for predicting the targeted label and dissimilar to the samples of the non-targeted labels. Then, it automatically selects the most effective and invisible trigger for each sample with an adaptive trigger optimization algorithm. We evaluate the performance of BadPrompt on five datasets and two continuous prompt models. The results exhibit the abilities of BadPrompt to effectively attack continuous prompts while maintaining high performance on the clean test sets, outperforming the baseline models by a large margin. The source code of BadPrompt is publicly available at https://github.com/papersPapers/BadPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper conducts the first study on the vulnerability of the continuous prompt learning algorithm to backdoor attacks, and proposes BadPrompt, a lightweight and task-adaptive algorithm, to backdoor attack continuous prompts.'}",https://arxiv.org/pdf/2211.14719
-Evaluating the Robustness of Discrete Prompts,Yoichi Ishibashi,"Discrete prompts have been used for fine-tuning Pre-trained Language Models for diverse NLP tasks. In particular, automatic methods that generate discrete prompts from a small set of training instances have reported superior performance. However, a closer look at the learnt prompts reveals that they contain noisy and counter-intuitive lexical constructs that would not be encountered in manually-written prompts. This raises an important yet understudied question regarding the robustness of automatically learnt discrete prompts when used in downstream tasks. To address this question, we conduct a systematic study of the robustness of discrete prompts by applying carefully designed perturbations into an application using AutoPrompt and then measure their performance in two Natural Language Inference (NLI) datasets. Our experimental results show that although the discrete prompt-based method remains relatively robust against perturbations to NLI inputs, they are highly sensitive to other types of perturbations such as shuffling and deletion of prompt tokens. Moreover, they generalize poorly across different NLI datasets. We hope our findings will inspire future work on robust discrete prompt learning.","{'model': 'tldr@v2.0.0', 'text': 'Although the discrete prompt-based method remains relatively robust against perturbation to NLI inputs, they are highly sensitive to other types of perturbations such as shuffling and deletion of prompt tokens and generalize poorly across different NLI datasets.'}",http://arxiv.org/pdf/2302.05619
-Microvascular decompression surgery: surgical principles and technical nuances based on 4000 cases,J. Zhong,"Abstract Background: As an etiological treatment of trigeminal neuralgia (TN) and hemifacial spasm (HFS), microvascular decompression (MVD) has been popularized around the world. However, as a functional operation in the cerebellopontine angle (CPA), this process can be risky and the postoperative outcomes might not be good enough sometimes. Objective: In order to obtain a better result with less complication, this surgery should be further addressed. Methods: With experience of more than 4000 MVDs, we have gained knowledge about the operative technique. Through abundant intraoperative photos, each step of the procedure was demonstrated in detail and the surgical strategy was focused. Results: The principle of MVD is to separate the nerve-vessel confliction rather than isolate it with prostheses. A prompt identification of the conflict site is important, which hinges on a good exposure. A satisfactory working space can be established by an appropriate positioning of the patient's head and a proper craniectomy as well as a rational approach. A sharp dissection of arachnoids leads to a maximal visualization of the entire intracranial course of the nerve root. All the vessels contacting the trigeminal or facial nerve should be treated. Intraoperative electrophysiological mentoring is helpful to distinguish the offending artery for hemifacial cases. Conclusion: MVD is an effective treatment for the patient with TN or HFS. Immediate relief can be achieved by an experienced neurosurgeon with good knowledge of regional anatomy. A safe surgery is the tenet of MVD, and accordingly, no single step of the procedure should be ignored.","{'model': 'tldr@v2.0.0', 'text': 'MVD is an effective treatment for the patient with TN or HFS and can be achieved by an experienced neurosurgeon with good knowledge of regional anatomy and no single step of the procedure should be ignored.'}",
-Diagnosis of Dengue Infection Using Conventional and Biosensor Based Techniques,O. Parkash,"Dengue is an arthropod-borne viral disease caused by four antigenically different serotypes of dengue virus. This disease is considered as a major public health concern around the world. Currently, there is no licensed vaccine or antiviral drug available for the prevention and treatment of dengue disease. Moreover, clinical features of dengue are indistinguishable from other infectious diseases such as malaria, chikungunya, rickettsia and leptospira. Therefore, prompt and accurate laboratory diagnostic test is urgently required for disease confirmation and patient triage. The traditional diagnostic techniques for the dengue virus are viral detection in cell culture, serological testing, and RNA amplification using reverse transcriptase PCR. This paper discusses the conventional laboratory methods used for the diagnosis of dengue during the acute and convalescent phase and highlights the advantages and limitations of these routine laboratory tests. Subsequently, the biosensor based assays developed using various transducers for the detection of dengue are also reviewed.","{'model': 'tldr@v2.0.0', 'text': 'The conventional laboratory methods used for the diagnosis of dengue during the acute and convalescent phase are discussed and the advantages and limitations of these routine laboratory tests are highlighted.'}",https://www.mdpi.com/1999-4915/7/10/2877/pdf?version=1445253550
-A Review on Fault Diagnosis and Fault Tolerant Control Methods for Single-rotor Aerial Vehicles,Xin Qi,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the recent development and current researches in the field of fault diagnosis, including analytical/model-based, signal processing-based and knowledge-based techniques, and also passive/active fault- tolerant control approaches is presented.'}",http://ir.sia.cn//bitstream/173321/14053/4/A%20Review%20on%20Fault%20Diagnosis%20and%20Fault%20Tolerant%20Control%20Methods%20for%20Single-rotor%20Aerial%20Vehicles.pdf
-Are we overpathologizing everyday life? A tenable blueprint for behavioral addiction research,J. Billieux,"Background Behavioral addiction research has been particularly flourishing over the last two decades. However, recent publications have suggested that nearly all daily life activities might lead to a genuine addiction. Methods and aim In this article, we discuss how the use of atheoretical and confirmatory research approaches may result in the identification of an unlimited list of “new” behavioral addictions. Results Both methodological and theoretical shortcomings of these studies were discussed. Conclusions We suggested that studies overpathologizing daily life activities are likely to prompt a dismissive appraisal of behavioral addiction research. Consequently, we proposed several roadmaps for future research in the field, centrally highlighting the need for longer tenable behavioral addiction research that shifts from a mere criteria-based approach toward an approach focusing on the psychological processes involved.","{'model': 'tldr@v2.0.0', 'text': 'This article discussed how the use of atheoretical and confirmatory research approaches may result in the identification of an unlimited list of “new” behavioral addictions, and suggested that studies overpathologizing daily life activities are likely to prompt a dismissive appraisal of behavioral addiction research.'}",https://akjournals.com/downloadpdf/journals/2006/4/3/article-p119.pdf
-In vivo measurements of body protein based on the determination of nitrogen by prompt gamma analysis.,K. Mcneill,"Measurement of nitrogen in the human body gives a measure of total protein in the body. Nitrogen (N) may be measured in vivo by a nuclear technique in which the characteristic nitrogen gamma rays emitted after capture of neutrons are detected and counted. The accuracy (as determined by comparison of the measurement of nitrogen in a pig by nuclear and chemical methods) is better than 10%. The reproducibility is less than 6%. The radiation dose is about 5 millirads. This method had been used with 43 patients and 23 normal volunteers and the results are reported. The data are presented in the from of total N content, percentage N by weight, and as a comparison of N and potassium (K) in the body. It is found that the mean percentage of N in the normal males is 2.5%, giving a value of 1.75 kg in the 70-kg standard man. The mean percentage values are not greatly different between normal volunteers and malnourished patients. Taking all subjects, N/K ratios ranged from 7.5 to 20 g/g. Thus K cannot be used as a predictor of N. Nitrogen content of normals correlates well (r = 0.68) with height squared; for malnourished patients there is again a good correlation (r = 0.81) but the regression line for patients is approximately 34% below that for the volunteers consistent with their depleted state.","{'model': 'tldr@v2.0.0', 'text': 'Nitrogen content of normals correlates well with height squared; for malnourished patients there is again a good correlation but the regression line for patients is approximately 34% below that for the volunteers consistent with their depleted state, so K cannot be used as a predictor of N.'}",
-Multilingual Relation Classification via Efficient and Effective Prompting,Yuxuan Chen,"Prompting pre-trained language models has achieved impressive performance on various NLP tasks, especially in low data regimes. Despite the success of prompting in monolingual settings, applying prompt-based methods in multilingual scenarios has been limited to a narrow set of tasks, due to the high cost of handcrafting multilingual prompts. In this paper, we present the first work on prompt-based multilingual relation classification (RC), by introducing an efficient and effective method that constructs prompts from relation triples and involves only minimal translation for the class labels. We evaluate its performance in fully supervised, few-shot and zero-shot scenarios, and analyze its effectiveness across 14 languages, prompt variants, and English-task training in cross-lingual settings. We find that in both fully supervised and few-shot scenarios, our prompt method beats competitive baselines: fine-tuning XLM-R_EM and null prompts. It also outperforms the random baseline by a large margin in zero-shot experiments. Our method requires little in-language knowledge and can be used as a strong baseline for similar multilingual classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the first work on prompt-based multilingual relation classification (RC), by introducing an efficient and effective method that constructs prompts from relation triples and involves only minimal translation for the class labels.'}",https://arxiv.org/pdf/2210.13838
-Measurement of the differential cross section for isolated prompt photon production in pp collisions at 7 TeV,S. Chatrchyan,"A measurement of the differential cross section for the inclusive production of isolated prompt photons in proton-proton collisions at a centre-of-mass energy of 7 TeV is presented. The data sample corresponds to an integrated luminosity of 36 inverse picobarns recorded by the CMS detector at the LHC. The measurement covers the pseudorapidity range |eta|<2.5 and the transverse energy range 25 < ET < 400 GeV, corresponding to the kinematic region 0.007 < xT < 0.114. Photon candidates are identified with two complementary methods, one based on photon conversions in the silicon tracker and the other on isolated energy deposits in the electromagnetic calorimeter. The measured cross section is presented as a function of ET in four pseudorapidity regions. The next-to-leading-order perturbative QCD calculations are consistent with the measured cross section.",,http://link.aps.org/pdf/10.1103/PhysRevD.84.052011
-Electrocardiogram pattern recognition and analysis based on artificial neural networks and support vector machines: a review.,M. Sansone,"Computer systems for Electrocardiogram (ECG) analysis support the clinician in tedious tasks (e.g., Holter ECG monitored in Intensive Care Units) or in prompt detection of dangerous events (e.g., ventricular fibrillation). Together with clinical applications (arrhythmia detection and heart rate variability analysis), ECG is currently being investigated in biometrics (human identification), an emerging area receiving increasing attention. Methodologies for clinical applications can have both differences and similarities with respect to biometrics. This paper reviews methods of ECG processing from a pattern recognition perspective. In particular, we focus on features commonly used for heartbeat classification. Considering the vast literature in the field and the limited space of this review, we dedicated a detailed discussion only to a few classifiers (Artificial Neural Networks and Support Vector Machines) because of their popularity; however, other techniques such as Hidden Markov Models and Kalman Filtering will be also mentioned.","{'model': 'tldr@v2.0.0', 'text': 'This paper reviews methods of ECG processing from a pattern recognition perspective and focuses on features commonly used for heartbeat classification, mainly Artificial Neural Networks and Support Vector Machines.'}",https://downloads.hindawi.com/journals/jhe/2013/904584.pdf
-Understanding and managing methotrexate nephrotoxicity.,B. Widemann,"Methotrexate (MTX) is one of the most widely used anti-cancer agents, and administration of high-dose methotrexate (HDMTX) followed by leucovorin (LV) rescue is an important component in the treatment of a variety of childhood and adult cancers. HDMTX can be safely administered to patients with normal renal function by the use of alkalinization, hydration, and pharmacokinetically guided LV rescue. Despite these measures, HDMTX-induced renal dysfunction continues to occur in approximately 1.8% of patients with osteosarcoma treated on clinical trials. Prompt recognition and treatment of MTX-induced renal dysfunction are essential to prevent potentially life-threatening MTX-associated toxicities, especially myelosuppression, mucositis, and dermatitis. In addition to conventional treatment approaches, dialysis-based methods have been used to remove MTX with limited effectiveness. More recently carboxypeptidase-G(2) (CPDG(2)), a recombinant bacterial enzyme that rapidly hydrolyzes MTX to inactive metabolites, has become available for the treatment of HDMTX-induced renal dysfunction. CPDG(2) administration has been well tolerated and resulted in consistent and rapid reductions in plasma MTX concentrations by a median of 98.7% (range, 84%-99.5%). The early administration of CPDG(2) in addition to LV may be beneficial for patients with MTX-induced renal dysfunction and significantly elevated plasma MTX concentrations.","{'model': 'tldr@v2.0.0', 'text': 'CPDG(2) administration has been well tolerated and resulted in consistent and rapid reductions in plasma MTX concentrations by a median of 98.7% (range, 84%-99.5%).'}",http://theoncologist.alphamedpress.org/content/11/6/694.full.pdf
-Changes in Teaching Efficacy during a Professional Development School-Based Science Methods Course.,S. Swars,"This mixed methods study offers a theoretically grounded description of a field-based science methods course within a Professional Development School (PDS) model (i.e., PDS-based course). The preservice teachers' (n= 21) experiences within the PDS-based course prompted significant changes in their personal teaching efficacy, with the opportunities working with children afforded in the course attributed as a source of these beliefs. However, the preservice teachers' teaching outcome expectancy beliefs did not significantly shift. The results of this study support the extant literature in that field experiences with science methods courses can facilitate preservice teacher development in meaningful ways. This study expands upon this literature by considering a PDS context for science teacher preparation, more specifically, a science methods course purposefully integrated in elementary classrooms at a PDS where preservice teachers facilitated scientific inquiry projects with children. The findings should prompt new ways of thinking about teacher preparation, particularly related to science, that provide systematic and intentional connectivity between university programs and K–12 schools so preservice teachers can connect theory and practice.",,
-Paper-based point-of-care testing for diagnosis of dengue infections,Jane Ru Choi,"Abstract Dengue endemic is a serious healthcare concern in tropical and subtropical countries. Although well-established laboratory tests can provide early diagnosis of acute dengue infections, access to these tests is limited in developing countries, presenting an urgent need to develop simple, rapid, and robust diagnostic tools. Point-of-care (POC) devices, particularly paper-based POC devices, are typically rapid, cost-effective and user-friendly, and they can be used as diagnostic tools for the prompt diagnosis of dengue at POC settings. Here, we review the importance of rapid dengue diagnosis, current dengue diagnostic methods, and the development of paper-based POC devices for diagnosis of dengue infections at the POC.","{'model': 'tldr@v2.0.0', 'text': 'The importance of rapid d Dengue diagnosis, current dengue diagnostic methods, and the development of paper-based POC devices for diagnosis of dengu infections at the POC are reviewed.'}",
-Knowledge Prompting in Pre-trained Language Model for Natural Language Understanding,J. Wang,"Knowledge-enhanced Pre-trained Language Model (PLM) has recently received significant attention, which aims to incorporate factual knowledge into PLMs. However, most existing methods modify the internal structures of fixed types of PLMs by stacking complicated modules, and introduce redundant and irrelevant factual knowledge from knowledge bases (KBs). In this paper, to address these problems, we introduce a seminal knowledge prompting paradigm and further propose a knowledge-prompting-based PLM framework KP-PLM. This framework can be flexibly combined with existing mainstream PLMs. Specifically, we first construct a knowledge sub-graph from KBs for each context. Then we design multiple continuous prompts rules and transform the knowledge sub-graph into natural language prompts. To further leverage the factual knowledge from these prompts, we propose two novel knowledge-aware self-supervised tasks including prompt relevance inspection and masked prompt modeling. Extensive experiments on multiple natural language understanding (NLU) tasks show the superiority of KP-PLM over other state-of-the-art methods in both full-resource and low-resource settings. Our source codes will be released upon the acceptance of the paper.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a seminal knowledge prompting paradigm and proposes a knowledge-prompting-based PLM framework KP-PLM that can be flexibly combined with existing mainstream PLMs and proposes two novel knowledge-aware self-supervised tasks.'}",http://arxiv.org/pdf/2210.08536
-"Event-by-event evaluation of the prompt fission neutron spectrum from 239Pu(n,f)",R. Vogt,"We have developed an improved evaluation method for the spectrum of neutrons emitted in fission of {sup 239}Pu induced by incident neutrons with energies up to 20 MeV. The covariance data, including incident energy correlations introduced by the evaluation method, were used to fix the input parameters in our event-by-event model of fission, FREYA, by applying formal statistical methods. Formal estimates of uncertainties in the evaluation were developed by randomly sampling model inputs and calculating likelihood functions based on agreement with the evaluated . Our approach is able to employ a greater variety of fission measurements than the relatively coarse spectral data alone. It also allows the study of numerous fission observables for more accurate model validation. The combination of an event-by-event Monte Carlo fission model with a statistical-likelihood analysis is thus a powerful tool for evaluation of fission-neutron data. Our empirical model FREYA follows the complete fission event from birth of the excited fragments through their decay via neutron emission until the fragment excitation energy is below the neutron separation energy when neutron emission can no longer occur. The most recent version of FREYA incorporates pre-equilibrium neutron emission, the emission of the first neutron before equilibrium ismore » reached in the compound nucleus, and multi-chance fission, neutron evaporation prior to fission when the incident neutron energy is above the neutron separation energy. Energy, momentum, charge and mass number are conserved throughout the fission process. The best available values of fragment masses and total kinetic energies are used as inputs to FREYA. We fit three parameters that are not well under control from previous measurements: the shift in the total fragment kinetic energy; the energy scale of the asymptotic level density parameter, controlling the fragment 'temperature' for neutron evaporation; and the relative excitation of the light and heavy fragments, governing the number and energy of neutrons emitted from each fragment. The latter two parameters are assumed to be independent of the incident neutron energy while the first varies with incident energy. We describe our method and the subsequent spectral evaluation and present the results of several standard validation calculations that test our new evaluation. These benchmarks include critical assemblies, sensitive to criticality in fast systems; pulsed sphere measurements testing the spectra at incident neutron energies of 14 MeV; and other tests.« less",,https://link.aps.org/accepted/10.1103/PhysRevC.85.024608
-"Frequently asked questions about chlorophyll fluorescence, the sequel",H. Kalaji,,"{'model': 'tldr@v2.0.0', 'text': 'Questions about instruments, methods and applications based on chlorophyll a fluorescence, and the answers draw on knowledge from different Chl a Fluorescence analysis domains, yielding in several cases new insights.'}",https://link.springer.com/content/pdf/10.1007%2Fs11120-016-0318-y.pdf
-Evidence-based diagnostics: adult septic arthritis.,C. Carpenter,"BACKGROUND Acutely swollen or painful joints are common complaints in the emergency department (ED). Septic arthritis in adults is a challenging diagnosis, but prompt differentiation of a bacterial etiology is crucial to minimize morbidity and mortality. OBJECTIVES The objective was to perform a systematic review describing the diagnostic characteristics of history, physical examination, and bedside laboratory tests for nongonococcal septic arthritis. A secondary objective was to quantify test and treatment thresholds using derived estimates of sensitivity and specificity, as well as best-evidence diagnostic and treatment risks and anticipated benefits from appropriate therapy. METHODS Two electronic search engines (PUBMED and EMBASE) were used in conjunction with a selected bibliography and scientific abstract hand search. Inclusion criteria included adult trials of patients presenting with monoarticular complaints if they reported sufficient detail to reconstruct partial or complete 2 × 2 contingency tables for experimental diagnostic test characteristics using an acceptable criterion standard. Evidence was rated by two investigators using the Quality Assessment Tool for Diagnostic Accuracy Studies (QUADAS). When more than one similarly designed trial existed for a diagnostic test, meta-analysis was conducted using a random effects model. Interval likelihood ratios (LRs) were computed when possible. To illustrate one method to quantify theoretical points in the probability of disease whereby clinicians might cease testing altogether and either withhold treatment (test threshold) or initiate definitive therapy in lieu of further diagnostics (treatment threshold), an interactive spreadsheet was designed and sample calculations were provided based on research estimates of diagnostic accuracy, diagnostic risk, and therapeutic risk/benefits. RESULTS The prevalence of nongonococcal septic arthritis in ED patients with a single acutely painful joint is approximately 27% (95% confidence interval [CI] = 17% to 38%). With the exception of joint surgery (positive likelihood ratio [+LR] = 6.9) or skin infection overlying a prosthetic joint (+LR = 15.0), history, physical examination, and serum tests do not significantly alter posttest probability. Serum inflammatory markers such as white blood cell (WBC) counts, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) are not useful acutely. The interval LR for synovial white blood cell (sWBC) counts of 0 × 10(9)-25 × 10(9)/L was 0.33; for 25 × 10(9)-50 × 10(9)/L, 1.06; for 50 × 10(9)-100 × 10(9)/L, 3.59; and exceeding 100 × 10(9)/L, infinity. Synovial lactate may be useful to rule in or rule out the diagnosis of septic arthritis with a +LR ranging from 2.4 to infinity, and negative likelihood ratio (-LR) ranging from 0 to 0.46. Rapid polymerase chain reaction (PCR) of synovial fluid may identify the causative organism within 3 hours. Based on 56% sensitivity and 90% specificity for sWBC counts of >50 × 10(9)/L in conjunction with best-evidence estimates for diagnosis-related risk and treatment-related risk/benefit, the arthrocentesis test threshold is 5%, with a treatment threshold of 39%. CONCLUSIONS Recent joint surgery or cellulitis overlying a prosthetic hip or knee were the only findings on history or physical examination that significantly alter the probability of nongonococcal septic arthritis. Extreme values of sWBC (>50 × 10(9)/L) can increase, but not decrease, the probability of septic arthritis. Future ED-based diagnostic trials are needed to evaluate the role of clinical gestalt and the efficacy of nontraditional synovial markers such as lactate.","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the diagnostic characteristics of history, physical examination, and bedside laboratory tests for nongonococcal septic arthritis found recent joint surgery or cellulitis overlying a prosthetic hip or knee were the only findings on history or physical examination that significantly alter the probability of nong onococcalSeptic arthritis.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1553-2712.2011.01121.x
-mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models,Ryokan Ri,"Recent studies have shown that multilingual pretrained language models can be effectively improved with cross-lingual alignment information from Wikipedia entities.However, existing methods only exploit entity information in pretraining and do not explicitly use entities in downstream tasks.In this study, we explore the effectiveness of leveraging entity representations for downstream cross-lingual tasks.We train a multilingual language model with 24 languages with entity representations and showthe model consistently outperforms word-based pretrained models in various cross-lingual transfer tasks.We also analyze the model and the key insight is that incorporating entity representations into the input allows us to extract more language-agnostic features.We also evaluate the model with a multilingual cloze prompt task with the mLAMA dataset.We show that entity-based prompt elicits correct factual knowledge more likely than using only word representations.","{'model': 'tldr@v2.0.0', 'text': 'A multilingual language model with 24 languages with entity representations is trained and it is shown that the model consistently outperforms word-based pretrained models in various cross-lingual transfer tasks.'}",https://aclanthology.org/2022.acl-long.505.pdf
-Photophysics of thermally activated delayed fluorescence molecules,F. Dias,"Thermally activated delayed fluorescence (TADF) has recently emerged as one of the most attractive methods for harvesting triplet states in metal-free organic materials for application in organic light emitting diodes (OLEDs). A large number of TADF molecules have been reported in the literature with the purpose of enhancing the efficiency of OLEDs by converting non-emissive triplet states into emissive singlet states. TADF emitters are able to harvest both singlets and triplet states through fluorescence (prompt and delayed), the latter due to the thermally activated reverse intersystem crossing mechanism that allows up-conversion of low energy triplet states to the emissive singlet level. This allows otherwise pure fluorescent OLEDs to overcome their intrinsic limit of 25% internal quantum efficiency (IQE), which is imposed by the 1:3 singlet–triplet ratio arising from the recombination of charges (electrons and holes). TADF based OLEDS with IQEs close to 100% are now routinely fabricated in the green spectral region. There is also significant progress for blue emitters. However, red emitters still show relatively low efficiencies. Despite the significant progress that has been made in recent years, still significant challenges persist to achieve full understanding of the TADF mechanism and improve the stability of these materials. These questions need to be solved in order to fully implement TADF in OLEDs and expand their application to other areas. To date, TADF has been exploited mainly in the field of OLEDs, but applications in other areas, such as sensing and fluorescence microscopies, are envisaged. In this review, the photophysics of TADF molecules is discussed, summarising current methods to characterise these materials and the current understanding of the TADF mechanism in various molecular systems.","{'model': 'tldr@v2.0.0', 'text': 'Thermally activated delayed fluorescence has recently emerged as one of the most attractive methods for harvesting triplet states in metal-free organic materials for application in organic light emitting diodes (OLEDs).'}",https://iopscience.iop.org/article/10.1088/2050-6120/aa537e/pdf
-"Necrotizing fasciitis: The importance of early diagnosis, prompt surgical debridement and adjuvant therapy",N. Machado,"Necrotizing Fasciitis (NF) is a necrotizing soft tissue infection involving the fascia and subcutaneous tissue that can cause rapid local tissue necrosis and life-threatening severe sepsis. Aim: This article aims to review the aetiopathogenesis, investigations and management based on a literature review. Methods and Materials: The Medline literature search of relevant articles restricted to English language on necrotizing fasciitis was conducted and reviewed. Results: Necrotizing fasciitis is rare with an incidence ranging from 0.15 to 0.55 cases per 100,000 of the population. Accurate assessment and timely intervention are critical in the treatment of patients affected with NF. Understanding the history and unique characteristics of this disease is crucial to achieve early recognition, effective treatment and a favorable outcome. Classic symptoms include severe pain out of proportion to local findings, erythema, mottling, crepitus, skin anesthesia, warmth, tenderness, hemorrhagic bullous formation, edema in the affected area and fever. Predisposing conditions of NF are classified into 2 main categories (type I and II) based on causative microorganisms. Radical surgical debridement, broad spectrum antibiotics, negative pressure wound dressings, and hyperbaric oxygen therapy are considered to be the cornerstone of treatment. The mortality rate ranges widely from 10% to 75% and is related to delay in initial debridement, patient age of more than 60 years, associated hypotension, acidosis, bacteremia, renal failure, hyponatremia, peripheral vascular disease, myonecrosis and myositis. Conclusion: Necrotizing fasciitis is a devastating infection of the fascia and subcutaneous tissue. The presentation of the disease is nonspecific and variable. Delay in recognition and effective treatment increases the mortality. Prompt radical surgical debridement, appropriate antibiotics and adjuvant therapy contribute to an improved outcome.","{'model': 'tldr@v2.0.0', 'text': 'The history and unique characteristics of this disease is crucial to achieve early recognition, effective treatment and a favorable outcome, and radical surgical debridement, broad spectrum antibiotics, negative pressure wound dressings, and hyperbaric oxygen therapy are considered to be the cornerstone of treatment.'}",
-Epidemiology and clinical characteristics of community-acquired pneumonia in hospitalized children.,I. Michelow,"OBJECTIVES The precise epidemiology of childhood pneumonia remains poorly defined. Accurate and prompt etiologic diagnosis is limited by inadequate clinical, radiologic, and laboratory diagnostic methods. The objective of this study was to determine as precisely as possible the epidemiology and morbidity of community-acquired pneumonia in hospitalized children. METHODS Consecutive immunocompetent children hospitalized with radiographically confirmed lower respiratory infections (LRIs) were evaluated prospectively from January 1999 through March 2000. Positive blood or pleural fluid cultures or pneumolysin-based polymerase chain reaction assays, viral direct fluorescent antibody tests, or viral, mycoplasmal, or chlamydial serologic tests were considered indicative of infection by those organisms. Methods for diagnosis of pneumococcal pneumonia among study subjects were published by us previously. Selected clinical characteristics, indices of inflammation (white blood cell and differential counts and procalcitonin values), and clinical outcome measures (time to defervescence and duration of oxygen supplementation and hospitalization) were compared among groups of children. RESULTS One hundred fifty-four hospitalized children with LRIs were enrolled. Median age was 33 months (range: 2 months to 17 years). A pathogen was identified in 79% of children. Typical respiratory bacteria were identified in 60% (of which 73% were Streptococcus pneumoniae), viruses in 45%, Mycoplasma pneumoniae in 14%, Chlamydia pneumoniae in 9%, and mixed bacterial/viral infections in 23%. Preschool-aged children had as many episodes of atypical bacterial LRIs as older children. Children with typical bacterial or mixed bacterial/viral infections had the greatest inflammation and disease severity. Multivariate logistic-regression analyses revealed that high temperature (> or = 38.4 degrees C) within 72 hours after admission (odds ratio: 2.2; 95% confidence interval: 1.4-3.5) and the presence of pleural effusion (odds ratio: 6.6; 95% confidence interval: 2.1-21.2) were significantly associated with bacterial pneumonia. CONCLUSIONS This study used an expanded diagnostic armamentarium to define the broad spectrum of pathogens that cause pneumonia in hospitalized children. The data confirm the importance of S pneumoniae and the frequent occurrence of bacterial and viral coinfections in children with pneumonia. These findings will facilitate age-appropriate antibiotic selection and future evaluation of the clinical effectiveness of the pneumococcal conjugate vaccine as well as other candidate vaccines.","{'model': 'tldr@v2.0.0', 'text': 'The data confirm the importance of S pneumoniae and the frequent occurrence of bacterial and viral coinfections in children with pneumonia and will facilitate age-appropriate antibiotic selection and future evaluation of the clinical effectiveness of the pneumococcal conjugate vaccine as well as other candidate vaccines.'}",
-Artificial Intelligence and Acute Stroke Imaging,J. Soun,"SUMMARY: Artificial intelligence technology is a rapidly expanding field with many applications in acute stroke imaging, including ischemic and hemorrhage subtypes. Early identification of acute stroke is critical for initiating prompt intervention to reduce morbidity and mortality. Artificial intelligence can help with various aspects of the stroke treatment paradigm, including infarct or hemorrhage detection, segmentation, classification, large vessel occlusion detection, Alberta Stroke Program Early CT Score grading, and prognostication. In particular, emerging artificial intelligence techniques such as convolutional neural networks show promise in performing these imaging-based tasks efficiently and accurately. The purpose of this review is twofold: first, to describe AI methods and available public and commercial platforms in stroke imaging, and second, to summarize the literature of current artificial intelligence–driven applications for acute stroke triage, surveillance, and prediction.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this review is to describe AI methods and available public and commercial platforms in stroke imaging, and to summarize the literature of current artificial intelligence–driven applications for acute stroke triage, surveillance, and prediction.'}",http://www.ajnr.org/content/ajnr/42/1/2.full.pdf
-Integrating rapid pathogen identification and antimicrobial stewardship significantly decreases hospital costs.,K. Perez,"CONTEXT Early diagnosis of gram-negative bloodstream infections, prompt identification of the infecting organism, and appropriate antibiotic therapy improve patient care outcomes and decrease health care expenditures. In an era of increasing antimicrobial resistance, methods to acquire and rapidly translate critical results into timely therapies for gram-negative bloodstream infections are needed. OBJECTIVE To determine whether mass spectrometry technology coupled with antimicrobial stewardship provides a substantially improved alternative to conventional laboratory methods. DESIGN An evidence-based intervention that integrated matrix-assisted laser desorption and ionization time-of-flight mass spectrometry, rapid antimicrobial susceptibility testing, and near-real-time antimicrobial stewardship practices was implemented. Outcomes in patients hospitalized prior to initiation of the study intervention were compared to those in patients treated after implementation. Differences in length of hospitalization and hospital costs were assessed in survivors. RESULTS The mean hospital length of stay in the preintervention group survivors (n = 100) was 11.9 versus 9.3 days in the intervention group (n = 101; P = .01). After multivariate analysis, factors independently associated with decreased length of hospitalization included the intervention (hazard ratio, 1.38; 95% confidence interval, 1.01-1.88) and active therapy at 48 hours (hazard ratio, 2.9; confidence interval, 1.15-7.33). Mean hospital costs per patient were $45 709 in the preintervention group and $26 162 in the intervention group (P = .009). CONCLUSIONS Integration of rapid identification and susceptibility techniques with antimicrobial stewardship significantly improved time to optimal therapy, and it decreased hospital length of stay and total costs. This innovative strategy has ramifications for other areas of patient care.","{'model': 'tldr@v2.0.0', 'text': None}",
-"A Web-Based, Computer-Tailored Smoking Prevention Program to Prevent Children From Starting to Smoke After Transferring to Secondary School: Randomized Controlled Trial",H. Cremers,"Background Smoking prevalence rates among Dutch children increase rapidly after they transit to secondary school, in particular among children with a low socioeconomic status (SES). Web-based, computer-tailored programs supplemented with prompt messages may be able to empower children to prevent them from starting to smoke when they transit to secondary school. Objective The main aim of this study is to evaluate whether computer-tailored feedback messages, with and without prompt messages, are effective in decreasing children’s smoking intentions and smoking behavior after 12 and 25 months of follow-up. Methods Data were gathered at baseline (T0), and after 12 months (T1) and 25 months (T2) of follow-up of a smoking prevention intervention program called Fun without Smokes. A total of 162 schools were randomly allocated to a no-intervention control group, an intervention prompt group, or an intervention no-prompt group. A total of 3213 children aged 10 to 12 years old participated in the study and completed a Web-based questionnaire assessing their smoking intention, smoking behavior, and sociocognitive factors, such as attitude, social influence, and self-efficacy, related to smoking. After completion, children in the intervention groups received computer-tailored feedback messages in their own email inbox and those messages could be accessed on the intervention website. Children in the prompt group received prompt messages, via email and short message service (SMS) text messaging, to stimulate them to reuse the intervention website with nonsmoking content. Multilevel logistic regression analyses were performed using multiple imputations to assess the program effects on smoking intention and smoking behavior at T1 and T2. Results A total of 3213 children participated in the Fun without Smokes study at T0. Between T0 and T1 a total of 1067 children out of the original 3213 (33.21%) dropped out of the study. Between T0 and T2 the number of children that did not participate in the final measurement was 1730 out of the original 3213 (53.84%). No significant program effects were observed for any of the intervention groups compared to the control group at T1 for the intention to engage in smoking—prompt, OR 0.67 (95% CI 0.30-1.50), no-prompt, OR 0.76 (95% CI 0.34-1.67)—or for smoking behavior—prompt, OR 1.13 (95% CI 0.13-9.98), no-prompt, OR 0.50 (95% CI 0.04-5.59). Similar nonsignificant program effects were found at T2 for the intention to start smoking—prompt, OR 0.78 (95% CI 0.26-2.32), no-prompt, OR 1.31 (95% CI 0.45-3.82)—and smoking behavior—prompt, OR 0.53 (95% CI 0.12-2.47), no-prompt, OR 1.01 (95% CI 0.24-4.21). Conclusions This study showed that the Web-based, computer-tailored feedback messages with and without prompt messages were not effective in modifying children’s smoking intentions and smoking behavior as compared to no information. Future smoking prevention interventions are recommended to start closer to the age of actual smoking uptake. Furthermore, future studies on Web-based, computer-tailored smoking prevention programs should focus on assessing and controlling exposure to the educational content and the response to the prompt messages. Trial Registration Netherlands Trial Register NTR3116; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3116 (Archived by WebCite at http://www.webcitation.org/6O0wQYuPI).","{'model': 'tldr@v2.0.0', 'text': 'This study showed that the Web-based, computer-tailored feedback messages with and without prompt messages were not effective in modifying children’s smoking intentions and smoking behavior as compared to no information.'}",https://www.jmir.org/2015/3/e59/PDF
-Detection and Mitigation of Spoofing Attacks on a Vector-Based Tracking GPS Receiver,A. Jafarnia-Jahromi,"Position solutions provided by GNSS receivers can be completely misled by spoofing attacks. Consequently detection and mitigation of spoofing signals have become one of the important topics within the GNSS community. Some recently published work has addressed spoofing attack on tracking receivers. The work has mostly focused on signal quality monitoring (SQM) techniques to detect spoofing and authentic signals interaction. SQM methods can effectively detect the spoofing correlation peak that is approaching the authentic signal. However, they are not applicable in cases where spoofing attack does not affect the shape of the correlation peak. This situation happens when a spoofer and authentic signals are almost aligned together. This paper provides an analytical approach to investigate the interaction between the authentic and the spoofing correlation peaks during spoofing attacks. Then, a spoofing detection technique based on amplitude analysis of different (namely very early (VE), early (E), prompt (P), late (L) and very late (VL)) correlator branches is proposed. The proposed spoofing detection technique continuously checks the distribution of each correlator output. Spoofing attack is detected if this distribution considerably deviates from that of the authentic signal. A vector based tracking receiver structure has been also employed to bridge the authentic signal outage during the spoofing attack. Data collection has been performed using the simulated spoofing attack by Spirent hardware simulator. Data processing results verify that the proposed spoofing detection and mitigation based on vector based (VB) receiver structure techniques perform well in real world scenarios.","{'model': 'tldr@v2.0.0', 'text': 'An analytical approach to investigate the interaction between the authentic and the spoofing correlation peaks during spoofing attacks is provided and a spoofing detection technique based on amplitude analysis of different correlator branches is proposed.'}",
-Brain Tumor Detection Using Deep Neural Network and Machine Learning Algorithm,Masoumeh Siar,"Brain tumor can be classified into two types: benign and malignant. Timely and prompt disease detection and treatment plan leads to improved quality of life and increased life expectancy in these patients. One of the most practical and important methods is to use Deep Neural Network (DNN). In this paper, a Convolutional Neural Network (CNN) has been used to detect a tumor through brain Magnetic Resonance Imaging (MRI) images. Images were first applied to the CNN. The accuracy of Softmax Fully Connected layer used to classify images obtained 98.67%. Also, the accuracy of the CNN is obtained with the Radial Basis Function (RBF) classifier 97.34% and the Decision Tree (DT) classifier, is 94.24%. In addition to the accuracy criterion, we use the benchmarks of Sensitivity, Specificity and Precision evaluate network performance. According to the results obtained from the categorizers, the Softmax classifier has the best accuracy in the CNN according to the results obtained from network accuracy on the image testing. This is a new method based on the combination of feature extraction techniques with the CNN for tumor detection from brain images. The method proposed accuracy 99.12% on the test data. Due to the importance of the diagnosis given by the physician, the accuracy of the doctors help in diagnosing the tumor and treating the patient increased.","{'model': 'tldr@v2.0.0', 'text': 'This is a new method based on the combination of feature extraction techniques with the CNN for tumor detection from brain images and the Softmax classifier has the best accuracy in the CNN according to the results obtained from network accuracy on the image testing.'}",
-Clinical and diagnostic utility of saliva as a non-invasive diagnostic fluid:a systematic review,L. A. S. Nunes,"This systematic review presents the latest trends in salivary research and its applications in health and disease. Among the large number of analytes present in saliva, many are affected by diverse physiological and pathological conditions. Further, the non-invasive, easy and cost-effective collection methods prompt an interest in evaluating its diagnostic or prognostic utility. Accumulating data over the past two decades indicates towards the possible utility of saliva to monitor overall health, diagnose and treat various oral or systemic disorders and drug monitoring. Advances in saliva based systems biology has also contributed towards identification of several biomarkers, development of diverse salivary diagnostic kits and other sensitive analytical techniques. However, its utilization should be carefully evaluated in relation to standardization of pre-analytical and analytical variables, such as collection and storage methods, analyte circadian variation, sample recovery, prevention of sample contamination and analytical procedures. In spite of all these challenges, there is an escalating evolution of knowledge with the use of this biological matrix.","{'model': 'tldr@v2.0.0', 'text': 'Salivary utilization should be carefully evaluated in relation to standardization of pre-analytical and analytical variables, such as collection and storage methods, analyte circadian variation, sample recovery, prevention of sample contamination and analytical procedures.'}",https://europepmc.org/articles/pmc4470107?pdf=render
-A comparison of four prompt modes for route finding for community travellers with severe cognitive impairments,M. Sohlberg,"Primary objective: Navigational skills are fundamental to community travel and, hence, personal independence and are often disrupted in people with cognitive impairments. Navigation devices are being developed that can support community navigation by delivering directional information. Selecting an effective mode to provide route-prompts is a critical design issue. This study evaluated the differential effects on pedestrian route finding using different modes of prompting delivered via a handheld electronic device for travellers with severe cognitive impairments. Research design: A within-subject comparison study was used to evaluate potential differences in route navigation performance when travellers received directions using four different prompt modes: (1) aerial map image, (2) point of view map image, (3) text based instructions/no image and (4) audio direction/no image. Methods and procedures: Twenty travellers with severe cognitive impairments due to acquired brain injury walked four equivalent routes using four different prompting modes delivered via a wrist-worn navigation device. Navigation scores were computed that captured accuracy and confidence during navigation. Main outcome: Results of the repeated measures Analysis of Variance suggested that participants performed best when given prompts via speech-based audio directions. The majority of the participants also preferred this prompting mode. Findings are interpreted in the context of cognitive resource allocation theory.","{'model': 'tldr@v2.0.0', 'text': 'Results of the repeated measures Analysis of Variance suggested that participants performed best when given prompts via speech-based audio directions, and the majority of the participants also preferred this prompting mode.'}",
-Mastitis treatment-Reduction in antibiotic usage in dairy cows.,V. Krömker,"Animal-friendly, economical, resource-saving milk production provides the basis for sustained consumer acceptance. Bovine mastitis plays a decisive role in the dairy industry-disturbing animal health and welfare and causing considerable economic losses on the other hand. Currently, antimicrobial treatment is indispensable to keep bovine udder health, animal welfare and economic aspects in balance. On the contrary, emergence and spread of antimicrobial resistance (AMR) is an urgent matter of particular public interest, and as a consequence, antimicrobial usage (AMU) in production livestock is a critically discussed subject. In urgent need of future reduction in AMU in the dairy industry, this review article describes and discusses possible approaches promising prompt implementation, including therapeutical alternatives as well as pro- and metaphylactic concepts such as the implementation of evidence-based mastitis therapy concepts and selective dry cow treatment (sDCT), in search of the most effective and contemporary methods for decreasing AMU and AMR in dairy production.","{'model': 'tldr@v2.0.0', 'text': 'In urgent need of future reduction in AMU in the dairy industry, this review article describes and discusses possible approaches promising prompt implementation, including therapeutical alternatives as well as pro- and metaphylactic concepts such as the implementation of evidence-based mastitis therapy concepts and selective dry cow treatment.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/rda.13032
-Using theories of behaviour change to inform interventions for addictive behaviours.,Thomas Llewelyn Webb,"AIMS This paper reviews a set of theories of behaviour change that are used outside the field of addiction and considers their relevance for this field. METHODS Ten theories are reviewed in terms of (i) the main tenets of each theory, (ii) the implications of the theory for promoting change in addictive behaviours and (iii) studies in the field of addiction that have used the theory. An augmented feedback loop model based on Control Theory is used to organize the theories and to show how different interventions might achieve behaviour change. RESULTS Briefly, each theory provided the following recommendations for intervention: Control Theory: prompt behavioural monitoring, Goal-Setting Theory: set specific and challenging goals, Model of Action Phases: form 'implementation intentions', Strength Model of Self-Control: bolster self-control resources, Social Cognition Models (Protection Motivation Theory, Theory of Planned Behaviour, Health Belief Model): modify relevant cognitions, Elaboration Likelihood Model: consider targets' motivation and ability to process information, Prototype Willingness Model: change perceptions of the prototypical person who engages in behaviour and Social Cognitive Theory: modify self-efficacy. CONCLUSIONS There are a range of theories in the field of behaviour change that can be applied usefully to addiction, each one pointing to a different set of modifiable determinants and/or behaviour change techniques. Studies reporting interventions should describe theoretical basis, behaviour change techniques and mode of delivery accurately so that effective interventions can be understood and replicated.","{'model': 'tldr@v2.0.0', 'text': 'There are a range of theories in the field of behaviour change that can be applied usefully to addiction, each one pointing to a different set of modifiable determinants and/or behaviour change techniques.'}",
-Evidence-based drainage of infected hydronephrosis secondary to ureteric calculi.,S. Ramsey,"INTRODUCTION The obstructed, infected kidney is a urological emergency. It has been accepted that the management of infected hydronephrosis secondary to ureteric stones is through prompt decompression of the collecting system. However, the optimal method of decompression has yet to be established. MATERIALS AND METHODS A PubMed and Medline search was performed of all English-language articles from 1960 using key words ""sepsis,"" ""urosepsis,"" ""obstruction,"" ""obstructive pyelonephritis,"" ""pyonephrosis,"" ""infection and hydronephrosis"" ""decompression,"" ""stent,"" ""nephrostomy,"" and ""management."" The Cochrane database and National Institute for Clinical Effectiveness guidelines were searched using the terms ""sepsis,"" ""urosepsis,"" ""stent,"" ""nephrostomy,"" or ""obstruction."" Scottish intercollegiate guidelines were reviewed and no relevant guidance was identified. RESULTS Two randomised trials have compared retrograde stent insertion with percutaneous nephrostomy with one trial reporting specifically on patients with acute sepsis and obstruction. Neither trial showed one superior modality of decompression in effecting decompression and resolution of sepsis. A further literature search regarding the complications of percutaneous nephrostomy and stent insertion was carried out. An overall major complication rate from percutaneous nephrostomy insertion was found to be 4%, although the complication rates from stent insertion are less consistently reported. DISCUSSION There appears little evidence to suggest that retrograde stent insertion leads to increased bacteraemia or is significantly more hazardous in the setting of acute obstruction. Further region-wide discussion between urologists and interventional radiologists is required to establish management protocols for these acutely unwell patients.","{'model': 'tldr@v2.0.0', 'text': 'There appears little evidence to suggest that retrograde stent insertion leads to increased bacteraemia or is significantly more hazardous in the setting of acute obstruction.'}",
-Multi-Stage Pre-training for Automated Chinese Essay Scoring,Wei Song,"This paper proposes a pre-training based automated Chinese essay scoring method. The method involves three components: weakly supervised pre-training, supervised cross- prompt fine-tuning and supervised target- prompt fine-tuning. An essay scorer is first pre- trained on a large essay dataset covering diverse topics and with coarse ratings, i.e., good and poor, which are used as a kind of weak supervision. The pre-trained essay scorer would be further fine-tuned on previously rated es- says from existing prompts, which have the same score range with the target prompt and provide extra supervision. At last, the scorer is fine-tuned on the target-prompt training data. The evaluation on four prompts shows that this method can improve a state-of-the-art neural essay scorer in terms of effectiveness and domain adaptation ability, while in-depth analysis also reveals its limitations..","{'model': 'tldr@v2.0.0', 'text': 'The evaluation on four prompts shows that this method can improve a state-of-the-art neural essay scorer in terms of effectiveness and domain adaptation ability, while in-depth analysis also reveals its limitations.'}",https://www.aclweb.org/anthology/2020.emnlp-main.546.pdf
-Carotid blowout syndrome: modern trends in management,C. Suárez,"Carotid blowout syndrome (CBS) refers to rupture of the carotid artery and is an uncommon complication of head and neck cancer that can be rapidly fatal without prompt diagnosis and intervention. CBS develops when a damaged arterial wall cannot sustain its integrity against the patient’s blood pressure, mainly in patients who have undergone surgical procedures and radiotherapy due to cancer of the head and neck, or have been reirradiated for a recurrent or second primary tumor in the neck. Among patients irradiated prior to surgery, CBS is usually a result of wound breakdown, pharyngocutaneous fistula and infection. This complication has often been fatal in the past, but at the present time, early diagnosis and modern technology applied to its management have decreased morbidity and mortality rates. In addition to analysis of the causes and consequences of CBS, the purpose of this paper is to critically review methods for early diagnosis of this complication and establish individualized treatment based on endovascular procedures for each patient.","{'model': 'tldr@v2.0.0', 'text': 'Methods for early diagnosis of carotid blowout syndrome are critically reviewed and individualized treatment based on endovascular procedures for each patient is established.'}",https://www.dovepress.com/getfile.php?fileID=46143
-Helicobacter pylori test and eradicate versus prompt endoscopy for management of dyspeptic patients: 6.7 year follow up of a randomised trial,A. Lassen,"Background: Dyspepsia is a chronic disease with significant impact on the use of health care resources. A management strategy based on Helicobacter pylori testing has been recommended but the long term effect is unknown. Aim: To investigate the long term effect of a test and treat strategy compared with prompt endoscopy for management of dyspeptic patients in primary care. Patients: A total of 500 patients presenting in primary care with dyspepsia were randomised to management by H pylori testing plus eradication therapy (n = 250) or by endoscopy (n = 250). Results of 12 month follow up have previously been presented. Methods: Symptoms, quality of life, and patient satisfaction were recorded during a three month period, a median 6.7 years after randomisation (range 6.1–7.3 years). Number of endoscopies, antisecretory medication, H pylori treatments, and hospital visits were recorded from health care databases for the entire follow up period. Results: Median age was 45 years; 28% were H pylori infected. Use of resources was registered in all 500 patients (3084 person years) of whom 312 completed diaries. We found no difference in symptoms between the two groups. Median proportion of days without symptoms was 0.52 (interquartile range 0.10–0.88) in the test and eradicate group versus 0.64 (0.14–0.90) in the prompt endoscopy group (p = 0.27) (mean difference 0.05 (95% confidence interval (CI) −0.03 to 0.14)). Compared with the prompt endoscopy group, the test and eradicate group underwent fewer endoscopies (mean difference 0.62 endoscopies/person (95% CI 0.38–0.86)) and used less antisecretory medication (mean difference 102 defined daily doses/person (95% CI −1 to 205)). Conclusion: On a long term basis, a H pylori test and eradicate strategy is as efficient as prompt endoscopy for management of dyspeptic patients in primary care and reduces the use of endoscopy and antisecretory medication.","{'model': 'tldr@v2.0.0', 'text': 'On a long term basis, a H pylori test and eradicate strategy is as efficient as prompt endoscopy for management of dyspeptic patients in primary care and reduces the use of endoscopies and antisecretory medication.'}",https://gut.bmj.com/content/gutjnl/53/12/1758.full.pdf
-"Microbiology of Chronic Suppurative Otitis Media in a Tertiary Care Setup of Uttarakhand State, India",R. Prakash,"Background: Chronic suppurative otitis media (CSOM) is a notorious infection and a major health problem in developing countries causing serious local damage and threatening complications. Early and effective treatment based on the knowledge of causative micro-organisms and their antimicrobial sensitivity ensures prompt clinical recovery and possible complications can thus be avoided. Aims: The aim of this study was to isolate the organisms associated with CSOM and to detect the antibiogram of the aerobic isolates. Materials and Methods: A total of 204 patients clinically diagnosed of CSOM were enrolled in the study and the samples were obtained from each patient using sterile cotton swabs and cultured for microbial flora. Drug susceptibility testing for aerobic isolates was conducted using Kirby-Bauer disc diffusion method. Results: The most common causative organisms isolated were Staphylococcus aureus (48.69%) and Pseudomonas aeruginosa (19.89%) amongst the 191 aerobic isolates. Anaerobes accounted for 29.41% of the isolates while 12.25% were fungi. Antimicrobial profile of aerobic isolates revealed maximum sensitivity to amikacin (95.5%), ceftriaxone (83.4%) and gentamicin (82.7%). Conclusion: Knowing the etiological agents of CSOM and their antimicrobial susceptibility is of essential importance for an efficient treatment, prevention of both complications and development of antibiotic resistance and finally, the reduction of the treatment costs.","{'model': 'tldr@v2.0.0', 'text': 'Knowing the etiological agents of CSOM and their antimicrobial susceptibility is of essential importance for an efficient treatment, prevention of both complications and development of antibiotic resistance and finally, the reduction of the treatment costs.'}",
-Why is mucormycosis more difficult to cure than more common mycoses?,A. Katragkou,"Although considered to be a rare infection, mucormycosis (zygomycosis) has emerged as the second most common invasive mould infection. Despite the advent of newer antifungal agents, mortality rate of mucormycosis remains exceedingly high. Successful management of mucormycosis requires early diagnosis, reversal of underlying predisposing risk factors, surgical debridement and prompt administration of active antifungal agents. However, mucormycosis is not always amenable to cure. There are challenging obstacles that lead to difficulties in management of amphotericin B. These include unique host-based risk factors for mucormycosis, the fungus' resistance to innate host defences and distinctive features of its immunopathogenesis, such as extensive angioinvasion, increased virulence and use of chelators by the fungus as siderophores. In addition to these obstacles, the difficulties in early diagnosis, including nonspecific clinical manifestations, lack of serological methods, as well limitations of culture and molecular methods, lead to delay in initiation of antifungal therapy. Finally, the variability of susceptibility to amphotericin B and resistance to most other conventional antifungal agents leads to major limitations in successful treatment of this devastating infection.","{'model': 'tldr@v2.0.0', 'text': 'The variability of susceptibility to amphotericin B and resistance to most other conventional antifungal agents leads to major limitations in successful treatment of this devastating infection.'}",http://www.clinicalmicrobiologyandinfection.com/article/S1198743X14603360/pdf
-Real and accrual‐based earnings management and its legal consequences,Salma Ibrahim,"Purpose – Prior research suggests that firms manipulate earnings through accruals to achieve certain reporting objectives. Recently, especially following the Sarbanes‐Oxley (SarbOx) Act, researchers have turned their attention to real account manipulation as an alternative. However, there is no evidence on whether the likelihood of being detected by outsiders is different for firms using these alternative manipulation methods. The purpose of this paper is to examine this research question in the context of seasoned equity offerings (SEOs).Design/methodology/approach – First, the authors compare SEOs to a matched sample of non‐SEOs to document income‐increasing manipulation. Next, they identify SEOs that prompt lawsuits and compare sued and non‐sued firms to determine whether using a particular method of manipulation is more likely to be detected and associated with litigation.Findings – The authors find evidence of income‐increasing accrual and real manipulation for SEOs in the year prior to the offering ...",,
-A Smartphone App (AfyaData) for Innovative One Health Disease Surveillance from Community to National Levels in Africa: Intervention in Disease Surveillance,E. Karimuribo,"Background We describe the development and initial achievements of a participatory disease surveillance system that relies on mobile technology to promote Community Level One Health Security (CLOHS) in Africa. Objective The objective of this system, Enhancing Community-Based Disease Outbreak Detection and Response in East and Southern Africa (DODRES), is to empower community-based human and animal health reporters with training and information and communication technology (ICT)–based solutions to contribute to disease detection and response, thereby complementing strategies to improve the efficiency of infectious disease surveillance at national, regional, and global levels. In this study, we refer to techno-health as the application of ICT-based solutions to enhance early detection, timely reporting, and prompt response to health events in human and animal populations. Methods An EpiHack, involving human and animal health experts as well as ICT programmers, was held in Tanzania in 2014 to identify major challenges facing early detection, timely reporting, and prompt response to disease events. This was followed by a project inception workshop in 2015, which brought together key stakeholders, including policy makers and community representatives, to refine the objectives and implementation plan of the DODRES project. The digital ICT tools were developed and packaged together as the AfyaData app to support One Health disease surveillance. Community health reporters (CHRs) and officials from animal and human health sectors in Morogoro and Ngorongoro districts in Tanzania were trained to use the AfyaData app. The AfyaData supports near- to real-time data collection and submission at both community and health facility levels as well as the provision of feedback to reporters. The functionality of the One Health Knowledge Repository (OHKR) app has been integrated into the AfyaData app to provide health information on case definitions of diseases of humans and animals and to synthesize advice that can be transmitted to CHRs with next step response activities or interventions. Additionally, a WhatsApp social group was made to serve as a platform to sustain interactions between community members, local government officials, and DODRES team members. Results Within the first 5 months (August-December 2016) of AfyaData tool deployment, a total of 1915 clinical cases in livestock (1816) and humans (99) were reported in Morogoro (83) and Ngorongoro (1832) districts. Conclusions These initial results suggest that the DODRES community-level model creates an opportunity for One Health engagement of people in their own communities in the detection of infectious human and animal disease threats. Participatory approaches supported by digital and mobile technologies should be promoted for early disease detection, timely reporting, and prompt response at the community, national, regional, and global levels.","{'model': 'tldr@v2.0.0', 'text': 'These initial results suggest that the DODRES community-level model creates an opportunity for One Health engagement of people in their own communities in the detection of infectious human and animal disease threats.'}",
-Significantly improved analytical sensitivity of lateral flow immunoassays by using thermal contrast.,Zhenpeng Qin,"The ability to rapidly identify diseases enables prompt treatment and improves outcomes. This has increased the development and use of rapid point-of-care diagnostic devices capable of biomolecular detection in both high-income and resource-limited settings.[1] Lateral flow assays (LFAs) are inexpensive, simple, portable, and robust,[2] making LFAs commonplace in medicine, agriculture, and over-the-counter personal use such as for pregnancy testing. Although the analytical performance of some LFAs are comparable to laboratory based methods,[1a] the sensitivity of most LFAs is in the mM to μM range,[2–3] which is many folds less sensitive than other molecular techniques such as enzyme-linked immunoassays (ELISA). As a consequence, LFAs are not particularly useful for detection early in a disease course when there is low level of antigen. Due to the increasing need for highly sensitive molecular diagnostics, researchers have focused on developing microfluidics,[1a, 1b] biobar codes,[1c, 1d] and enzyme-based immunoassay technologies[4] technologies to fulfill the need since these technologies have nM to pM detection sensitivity for protein analysis and can potentially be miniaturized as handheld point-of-care diagnostic devices.[1c] These emerging technologies are still early in development and are not yet field-ready.","{'model': 'tldr@v2.0.0', 'text': 'Due to the increasing need for highly sensitive molecular diagnostics, researchers have focused on developing microfluidics, biobar codes, and enzyme-based immunoassay technologies to fulfill the need since these technologies have nM to pM detection sensitivity for protein analysis and can potentially be miniaturized as handheld point-of-care diagnostic devices.'}",https://europepmc.org/articles/pmc3337364?pdf=render
-Recent advances in the microbiological diagnosis of bloodstream infections,W. Florio,"Abstract Rapid identification (ID) and antimicrobial susceptibility testing (AST) of the causative agent(s) of bloodstream infections (BSIs) are essential for the prompt administration of an effective antimicrobial therapy, which can result in clinical and financial benefits. Immediately after blood sampling, empirical antimicrobial therapy, chosen on clinical and epidemiological data, is administered. When ID and AST results are available, the clinician decides whether to continue or streamline the antimicrobial therapy, based on the results of the in vitro antimicrobial susceptibility profile of the pathogen. The aim of the present study is to review and discuss the experimental data, advantages, and drawbacks of recently developed technological advances of culture-based and molecular methods for the diagnosis of BSI (including mass spectrometry, magnetic resonance, PCR-based methods, direct inoculation methods, and peptide nucleic acid fluorescence in situ hybridization), the understanding of which could provide new perspectives to improve and fasten the diagnosis and treatment of septic patients. Although blood culture remains the gold standard to diagnose BSIs, newly developed methods can significantly shorten the turnaround time of reliable microbial ID and AST, thus substantially improving the diagnostic yield.","{'model': 'tldr@v2.0.0', 'text': 'Although blood culture remains the gold standard to diagnose BSIs, newly developed methods can significantly shorten the turnaround time of reliable microbial ID and AST, thus substantially improving the diagnostic yield.'}",
-Comparative archaeometrical study of Roman silver coins by prompt gamma activation analysis and SEM-EDX,Z. Kasztovszky,,,
-Detecting consciousness: a unique role for neuroimaging.,A. Owen,"How can we ever know, unequivocally, that another person is conscious and aware? Putting aside deeper philosophical considerations about the nature of consciousness itself, historically, the only reliable method for detecting awareness in others has been through a predicted behavioral response to an external prompt or command. The answer may take the form of spoken words or a nonverbal signal such as a hand movement or the blink of an eye, but it is this answer, and only this answer, that allows us to infer awareness. In recent years, rapid technological developments in the field of neuroimaging have provided new methods for revealing thoughts, actions, and intentions based solely on the pattern of activity that is observed in the brain. In specialized centers, these methods are now being employed routinely to detect consciousness in behaviorally nonresponsive patients when all existing clinical techniques have failed to provide that information. In this review, I compare those circumstances in which neuroimaging data can be used to infer consciousness in the absence of a behavioral response with those circumstances in which it cannot. This distinction is fundamental for understanding and interpreting patterns of brain activity following acute brain injury and has profound implications for clinical care, diagnosis, prognosis, and medical-legal decision-making (relating to the prolongation, or otherwise, of life after severe brain injury). It also sheds light on more basic scientific questions about the nature of consciousness and the neural representation of our own thoughts and intentions.","{'model': 'tldr@v2.0.0', 'text': 'This review compares those circumstances in which neuroimaging data can be used to infer consciousness in the absence of a behavioral response with those circumstancesIn which it cannot and sheds light on more basic scientific questions about the nature of consciousness.'}",
-Specific Effects of a Calorie-Based Intervention on Stair Climbing in Overweight Commuters,A. Lewis,,"{'model': 'tldr@v2.0.0', 'text': 'The specificity of the effect confirms the appeal of this lifestyle activity for the overweight, and focuses on how intentions to control weight may be converted into behaviour.'}",https://academic.oup.com/abm/article-pdf/42/2/257/22059945/12160_2011_article_9283.pdf
-Fission modelling with FIFRELIN,O. Litaize,,,https://hal-cea.archives-ouvertes.fr/cea-02428818/file/Litaize2015_Article_FissionModellingWithFIFRELIN.pdf
-Kinetics of CH + N2 revisited with multireference methods.,L. B. Harding,"The potential energy surface for the CH + N2 reaction was reexamined with multireference ab initio electronic structure methods employing basis sets up to aug-cc-pvqz. Comparisons with related CCSD(T) calculations were also made. The multireference ab initio calculations indicate significant shortcomings in single reference based methods for two key rate-limiting transition states. Transition state theory calculations incorporating the revised best estimates for the transition state properties provide order of magnitude changes in the predicted rate coefficient in the temperature range of importance to the mechanism for prompt NO formation. At higher temperatures, two distinct pathways make a significant contribution to the kinetics. A key part of the transition state analysis involves a variable reaction coordinate transition state theory treatment for the formation of H + NCN from HNCN. The present predictions for the rate coefficients resolve the discrepancy between prior theory and very recent experimental measurements.","{'model': 'tldr@v2.0.0', 'text': 'Transition state theory calculations incorporating the revised best estimates for the transition state properties provide order of magnitude changes in the predicted rate coefficient in the temperature range of importance to the mechanism for prompt NO formation.'}",
-Automated DNA Sequence-Based Early Warning System for the Detection of Methicillin-Resistant Staphylococcus aureus Outbreaks,A. Mellmann,"Background The detection of methicillin-resistant Staphylococcus aureus (MRSA) usually requires the implementation of often rigorous infection-control measures. Prompt identification of an MRSA epidemic is crucial for the control of an outbreak. In this study we evaluated various early warning algorithms for the detection of an MRSA cluster. Methods and Findings Between 1998 and 2003, 557 non-replicate MRSA strains were collected from staff and patients admitted to a German tertiary-care university hospital. The repeat region of the S. aureus protein A (spa) gene in each of these strains was sequenced. Using epidemiological and typing information for the period 1998–2002 as reference data, clusters in 2003 were determined by temporal-scan test statistics. Various early warning algorithms (frequency, clonal, and infection control professionals [ICP] alerts) were tested in a prospective analysis for the year 2003. In addition, a newly implemented automated clonal alert system of the Ridom StaphType software was evaluated. A total of 549 of 557 MRSA were typeable using spa sequencing. When analyzed using scan test statistics, 42 out of 175 MRSA in 2003 formed 13 significant clusters (p < 0.05). These clusters were used as the “gold standard” to evaluate the various algorithms. Clonal alerts (spa typing and epidemiological data) were 100% sensitive and 95.2% specific. Frequency (epidemiological data only) and ICP alerts were 100% and 62.1% sensitive and 47.2% and 97.3% specific, respectively. The difference in specificity between clonal and ICP alerts was not significant. Both methods exhibited a positive predictive value above 80%. Conclusions Rapid MRSA outbreak detection, based on epidemiological and spa typing data, is a suitable alternative for classical approaches and can assist in the identification of potential sources of infection.","{'model': 'tldr@v2.0.0', 'text': 'Rapid MRSA outbreak detection, based on epidemiological and spa typing data, is a suitable alternative for classical approaches and can assist in the identification of potential sources of infection.'}",https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.0030033&type=printable
-General purpose computer-assisted clustering and conceptualization,Justin Grimmer,"We develop a computer-assisted method for the discovery of insightful conceptualizations, in the form of clusterings (i.e., partitions) of input objects. Each of the numerous fully automated methods of cluster analysis proposed in statistics, computer science, and biology optimize a different objective function. Almost all are well defined, but how to determine before the fact which one, if any, will partition a given set of objects in an “insightful” or “useful” way for a given user is unknown and difficult, if not logically impossible. We develop a metric space of partitions from all existing cluster analysis methods applied to a given dataset (along with millions of other solutions we add based on combinations of existing clusterings) and enable a user to explore and interact with it and quickly reveal or prompt useful or insightful conceptualizations. In addition, although it is uncommon to do so in unsupervised learning problems, we offer and implement evaluation designs that make our computer-assisted approach vulnerable to being proven suboptimal in specific data types. We demonstrate that our approach facilitates more efficient and insightful discovery of useful information than expert human coders or many existing fully automated methods.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a metric space of partitions from all existing cluster analysis methods applied to a given dataset and demonstrates that this approach facilitates more efficient and insightful discovery of useful information than expert human coders or many existing fully automated methods.'}",https://www.pnas.org/content/pnas/108/7/2643.full.pdf
-Parents' Educational Involvement: A Developmental Ecology Perspective,R. Seginer,"The objective of this review is to examine research on home-based and school-based parental involvement and generate new research questions by employing Bronfenbrenner's ecological framework consisting of the micro-, meso-, exo-, and macrosystems. This analysis shows that, although both family-based and school-based parental involvement are positively related to educational outcomes, their examination in the ecological framework prompts consideration of additional aspects of the micro- and mesosystems and their embeddedness in four exosystemic aspects (parents' networks and workplace, neighborhood, and educational policy) and two macrosystemic types (immigrant and ethnic groups). Guided by Bronfenbrenner's ecological thinking and the availability of advanced multivariate analysis methods, the next stage of this research should test multiple-step models describing factors that prompt parental involvement and mediate and moderate the parental involvement - educational outcomes links in different sociocultural settings.",,
-Social Representation of Cyberbullying and Adolescent Suicide: A Mixed-Method Analysis of News Stories,R. Young,"ABSTRACT Cyberbullying has provoked public concern after well-publicized suicides of adolescents. This mixed-methods study investigates the social representation of these suicides. A content analysis of 184 U.S. newspaper articles on death by suicide associated with cyberbullying or aggression found that few articles adhered to guidelines suggested by the World Health Organization and the American Foundation for Suicide Prevention to protect against suicidal behavioral contagion. Few articles made reference to suicide or bullying prevention resources, and most suggested that the suicide had a single cause. Thematic analysis of a subset of articles found that individual deaths by suicide were used as cautionary tales to prompt attention to cyberbullying. This research suggests that newspaper coverage of these events veers from evidence-based guidelines and that more work is needed to determine how best to engage with journalists about the potential consequences of cyberbullying and suicide coverage.","{'model': 'tldr@v2.0.0', 'text': 'The research suggests that newspaper coverage of these events veers from evidence-based guidelines and that more work is needed to determine how best to engage with journalists about the potential consequences of cyberbullying and suicide coverage.'}",
-Impact of an Electronic Health Record (EHR) Reminder on Human Papillomavirus (HPV) Vaccine Initiation and Timely Completion,M. Ruffin,"Background: The initiation and timely completion of the human papillomavirus (HPV) vaccine in young women is critical. We compared the initiation and completion of the HPV vaccine among women in 2 community-based networks with electronic health records: 1 with a prompt and reminder system (prompted cohort) and 1 without (unprompted cohort). Methods: Female patients aged 9 to 26 years seen between March 1, 2007, and January 25, 2010, were used as the retrospective cohort. Patient demographics and vaccination dates were extracted from the electronic health records. Results: Patients eligible for the vaccine included 6019 from the prompted cohort and 9096 from the unprompted cohort. Mean age at initiation was 17.3 years in the prompted cohort and 18.1 years in the unprompted cohort. Significantly more (P < .001) patients initiated the vaccine in the prompted cohort (34.9%) compared with the unprompted cohort (21.5%). African Americans aged 9 to 18 years with ≥3 visits during the observation period were significantly more likely to initiate in the prompted cohort (P < .001). The prompted cohort was significantly more likely (P < .001) to complete the vaccine series in a timely manner compared with the unprompted cohort. Conclusion: More patients aged 9 to 26 years initiated and achieved timely completion of the HPV vaccine series in clinics using an electronic health record system with prompts compared with clinics without prompts.","{'model': 'tldr@v2.0.0', 'text': 'The prompted cohort was significantly more likely to complete the vaccine series in a timely manner compared with the unprompted cohort and more patients aged 9 to 26 years initiated and achieved timely completion of the HPV vaccines in clinics using an electronic health record system with prompts.'}",https://www.jabfm.org/content/jabfp/28/3/324.full.pdf
-Comparative Study Using Various Methods for Identification of Staphylococcus Species in Clinical Specimens,F. Layer,"ABSTRACT Coagulase-negative staphylococci (CNS) play a predominant role in nosocomial infections. Rapid, reliable identification of these organisms is essential for accurate diagnosis and prompt effective treatment of these infections. Quite recently, the VITEK 2 g-positive (gram-positive [GP]) identification card (bioMérieux) has been redesigned for greater accuracy in the identification of gram-positive cocci. We compared the BD Phoenix (Becton Dickinson) and VITEK 2 (bioMérieux) automated microbiology systems, using their respective update version cards, and the API ID32 STAPH test. The glyceraldehyde-3-phosphate dehydrogenase (gap) gene-based T-RFLP (terminal restriction fragment length polymorphism) method was used for verifying the results. In total, 86 clinical isolates of CNS and 27 reference strains were analyzed. The results show that for identification of CNS, the automated identification methods using the newest VITEK 2 and BD Phoenix identification cards are comparable. However, API ID32 STAPH revealed more correct results compared to both automated microbiology systems. Despite the increased performance of the phenotypic automated identification systems compared to the former versions, molecular methods, e.g., the gap-based T-RFLP method, still show superior accuracy in identifying Staphylococcus species other than Staphylococcus aureus.","{'model': 'tldr@v2.0.0', 'text': 'Despite the increased performance of the phenotypic automated identification systems compared to the former versions, molecular methods, e.g., the gap-based T-RFLP method, still show superior accuracy in identifying Staphylococcus species other than Staphlyococcus aureus.'}",https://europepmc.org/articles/pmc1594629?pdf=render
-Boron Determination—A Review of Analytical Methods,R. N. Sah,"This paper reviews published methods of sample preparation, determinand purification, and the determination of boron concentration and isotopic composition in a sample. The most common methods for the determination of B concentration are spectrophotometric and plasma-source spectrometric methods. Although most spectrophotometric methods are based on colorimetric reactions of B with azomethine-H, curcumin, or carmine, other colorimetric and fluorometric methods have also been used to some extent. These methods, in general, suffer from numerous interferences and have low sensitivity and precision. Application of nuclear reaction and atomic emission/absorption spectrometric (AES/AAS) methods has remained limited because these methods have poor sensitivity and suffer from serious memory effects and interferences. Among a large number of published nuclear reaction methods only prompt-γ spectrometry has been of practical use. The prompt-γ method can determine B concentration in intact samples, which makes this method especially useful for some medical applications, including boron neutron capture therapy. However, this is a time-consuming method and not suitable for detection of low levels of B. Inductively coupled plasma optical emission spectrometry (ICP-OES) created a new dimension in B determination because of its simplicity, sensitivity, and multielement capability. However, it suffers interferences and is not adequately sensitive for some nutritional and medical applications involving animal tissues that are naturally low in B. All methods involving the measurement of B isotopic composition require a mass spectrometer. Thermal ionization mass spectrometry (TIMS) and secondary ion mass spectrometry (SIMS) have been used to measure isotopic composition of B; however, these methods are time consuming and require extensive sample preparation and purification. Development of inductively coupled plasma mass spectrometry (ICP-MS) not only overcame most of the drawbacks of earlier methods, but also its capabiltiy of measuring B isotopes made possible (1) B concentration determination by isotope dilution, (2) verification of B concentration by isotope fingerprinting in routine analysis, and (3) determination of total B concentration and B isotope ratio for biological tracer studies in the same run. Therefore, plasma source MS appears to be the method of choice among present-day technologies.",,
-Concept drift detection for online class imbalance learning,Shuo Wang,"Concept drift detection methods are crucial components of many online learning approaches. Accurate drift detections allow prompt reaction to drifts and help to maintain high performance of online models over time. Although many methods have been proposed, no attention has been given to data streams with imbalanced class distributions, which commonly exist in real-world applications, such as fault diagnosis of control systems and intrusion detection in computer networks. This paper studies the concept drift problem for online class imbalance learning. We look into the impact of concept drift on single-class performance of online models based on three types of classifiers, under seven different scenarios with the presence of class imbalance. The analysis reveals that detecting drift in imbalanced data streams is a more difficult task than in balanced ones. Minority-class recall suffers from a significant drop after the drift involving the minority class. Overall accuracy is not suitable for drift detection. Based on the findings, we propose a new detection method DDM-OCI derived from the existing method DDM. DDM-OCI monitors minority-class recall online to capture the drift. The results show a quick response of the online model working with DDM-OCI to the new concept.","{'model': 'tldr@v2.0.0', 'text': 'The analysis reveals that detecting drift in imbalanced data streams is a more difficult task than in balanced ones, and proposes a new detection method DDM-OCI derived from the existing methodDDM.'}",
-Cervical spine clearance in blunt trauma: evaluation of a computed tomography-based protocol.,B. Sanchez,"BACKGROUND Prompt identification of cervical spine injuries has been a critical issue in trauma management. In 1998, the authors developed a new protocol to evaluate cervical spines in blunt trauma. This protocol relies on clinical clearance for appropriate patients and helical computed tomography instead of plain radiographs for patients who cannot be clinically cleared. The authors then prospectively collected data on all cervical spine evaluations to assess the sensitivity and specificity of their approach. METHODS Any patient without clinical evidence of neurologic injury, alcohol or drug intoxication, or distracting injury underwent cervical spine evaluation by clinical examination. Patients who did not meet these criteria underwent helical computed tomographic scanning of the entire cervical spine. For patients who had neurologic deficits, a magnetic resonance image was obtained. If the patient was not evaluable secondary to coma, the computed tomographic scan was without abnormality, and the patient was moving all four extremities at arrival in the emergency department, the cervical spine was cleared, and spinal precautions were removed. Data were collected for all patients admitted to Santa Barbara Cottage Hospital trauma service between 1999 and 2002. The authors selected for analysis patients with blunt trauma and further identified those with closed head injuries (Glasgow Coma Scale score < 15 and loss of consciousness). In addition, all blunt cervical spine injuries were reviewed. RESULTS During the period of study, 2,854 trauma patients were admitted, of whom 2,603 (91%) had blunt trauma. Of these, 1,462 (56%) had closed head injuries. One hundred patients (7% of patients admitted for blunt trauma) had cervical spine or spinal cord injuries, of which 99 were identified by the authors' protocol. Only one injury was not appreciated in a patient with syringomyelia. Fifteen percent of patients with spinal cord injury had no radiographic abnormality; all of these patients presented with neurologic deficits. The sensitivity for detecting cervical spine injury was thus 99%, and the specificity was 100%. The risk of missing a cervical spine injury in these blunt trauma patients was 0.04%. The authors missed no spine injuries in patients with head injuries. CONCLUSION The use of the authors' protocol resulted in excellent sensitivity and specificity in detecting cervical spine injuries. In addition, it allowed early removal of spinal precautions.","{'model': 'tldr@v2.0.0', 'text': ""The use of the authors' protocol resulted in excellent sensitivity and specificity in detecting cervical spine injuries and allowed early removal of spinal precautions.""}",
-Indocyanine Green Lymphographic Evidence of Surgical Efficacy Following Microsurgical and Supermicrosurgical Lymphedema Reconstructions,Wei F. Chen,"Abstract Background Microsurgical vascularized lymph node transfer (VLNT) and supermicrosurgical lymphaticovenular anastomosis (LVA) are increasingly performed to treat lymphedema. The surgical outcome is commonly assessed by volume-based measurement (VBM), a method that is not consistently reliable. We describe indocyanine green (ICG) lymphography as an alternative postoperative tracking modality after lymphatic reconstruction with VLNT and LVA. Methods VLNT and LVA were performed in patients with therapy-refractory lymphedema. Patients were evaluated qualitatively by clinical assessment, quantitatively with VBM, and lymphographically using ICG lymphography. The evaluation was performed preoperatively, and at 3, 6, and 12-month postoperatively. Results Overall, 21 patients underwent lymphatic reconstruction with either VLNT or LVA. All reported prompt and durable relief of symptoms during the study period. All experienced disease regression based on the Campisi criteria. Out of the 21 patients, 20 (95%) demonstrated lymphographic down staging of disease severity. Out of the 21 patients, 3 (14%) developed a paradoxical increase in limb volume based on VBM despite clinical improvement. Conclusions ICG lymphography correlated highly with patient self-assessment and clinical examination, and is an effective postoperative tracking modality after lymphatic reconstruction.","{'model': 'tldr@v2.0.0', 'text': 'ICG lymphography correlated highly with patient self-assessment and clinical examination, and is an effective postoperative tracking modality after lymphatic reconstruction with VLNT and LVA.'}",
-"Nonpalpable, circumscribed, noncalcified solid breast masses: likelihood of malignancy based on lesion size and age of patient.",E. Sickles,"PURPOSE To determine whether lesion size and patient age should prompt immediate biopsy of selected nonpalpable, circumscribed, noncalcified solid (probably benign) breast masses, which otherwise would be managed with periodic mammographic surveillance. MATERIALS AND METHODS Consecutive cases of probably benign breast masses were studied, for which mammographic surveillance was recommended prospectively. Patient age and lesion size were recorded, but age and size thresholds were not used as eligibility criteria. Clinical outcome, determined for each patient after 3 or 3 1/2 years of surveillance, was evaluated as a function of patient age and lesion size. RESULTS There were 1,403 study cases, among which 19 cancers were diagnosed (positive predictive value [PPV] = 1.4%). Only small differences were found in PPV for various patient-age and lesion-size subgroups. Even the subgroup with the highest PPV, which consisted of women aged 50 years and older, contained 60 benign masses for each cancer. CONCLUSION Nonpalpable, circumscribed, noncalcified (probably benign) breast masses should be managed with periodic mammographic surveillance regardless of lesion size and patient age.","{'model': 'tldr@v2.0.0', 'text': 'Nonpalpable, circumscribed, noncalcified (probably benign) breast masses should be managed with periodic mammographic surveillance regardless of lesion size and patient age, regardless of patient age.'}",
-Study on pid parameters tuning method based on Matlab/Simulink,Suping Li,"Enormous calculation of proportional-integral-derivative (PID) controller parameters tuning with analytic methods is an important problem demanding prompt solution. Parameters tuning based on Matlab/Simulink is simplicity, visual manipulation method which leaves out above program. According to the Ziegler-Nichols (Z-N) method, this paper introduces how to reduce and validate the PID controller parameter with the help of MATLAB tool taking a certain control model as an example. The simulation results show the effectiveness of this method and can be fit for application in the engineering.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces how to reduce and validate the PID controller parameter with the help of MATLAB tool taking a certain control model as an example and the simulation results show the effectiveness of this method and can be fit for application in the engineering.'}",
-Management of severe hyperkalemia,L. Weisberg,"Background and Objectives:Hyperkalemia is one of the few potentially lethal electrolyte disturbances. Prompt recognition and expeditious treatment of severe hyperkalemia are expected to save lives. This review is intended to provide intensivists and other interested clinicians with an understanding of the pathophysiology that underlies hyperkalemia, and a rational approach to its management. Methods:This article reviews and analyzes literature relevant to the pathophysiology and management of severe hyperkalemia. Methods include search of MEDLINE, and bibliographic search of current textbooks and journal articles. Results and Conclusions:A more complete understanding of potassium homeostasis in recent years has led to new approaches to the management of severe hyperkalemia. The physiologically based sequential approach still applies. The efficacy, pitfalls, and risks of the agents available for use at each step in the sequence are critically reviewed. Rational use of the available tools will allow clinicians to successfully treat severe hyperkalemia.","{'model': 'tldr@v2.0.0', 'text': 'A more complete understanding of potassium homeostasis in recent years has led to new approaches to the management of severe hyperkalemia, and the physiologically based sequential approach still applies.'}",
-"Prompts to Disrupt Sitting Time and Increase Physical Activity at Work, 2011–2012",A. Swartz,"Introduction The objective of this study was to assess change in sitting and physical activity behavior in response to a workplace intervention to disrupt prolonged sitting time. Methods Sixty office workers were randomized to either a Stand group (n = 29), which received hourly prompts (computer-based and wrist-worn) to stand up, or a Step group (n = 31), which received the same hourly prompts and an additional prompt to walk 100 steps or more upon standing. An ActivPAL monitor was used to assess sitting and physical activity behavior on the same 3 consecutive workdays during baseline and intervention periods. Mixed-effect models with random intercepts and random slopes for time were performed to assess change between groups and across time. Results Both groups significantly reduced duration of average sitting bouts (Stand group, by 16%; Step group, by 19%) and the number of sitting bouts of 60 minutes or more (Step group, by 36%; Stand group, by 54%). The Stand group significantly reduced total sitting time (by 6.6%), duration of the longest sitting bout (by 29%), and number of sitting bouts of 30 minutes or more (by 13%) and increased the number of sit-to-stand transitions (by 15%) and standing time (by 23%). Stepping time significantly increased in the Stand (by 14%) and Step (by 29%) groups, but only the Step group significantly increased (by 35%) the number of steps per workday. Differences in changes from baseline to intervention between groups were not significant for any outcome. Conclusion Interventions that focus on disrupting sitting time only in the workplace may result in less sitting. When sitting time disruptions are paired with a physical activity prompt, people may be more likely to increase their workday physical activity, but the effect on sitting time may be attenuated.","{'model': 'tldr@v2.0.0', 'text': 'Interventions that focus on disrupting sitting time only in the workplace may result in less sitting, but when sitting time disruptions are paired with a physical activity prompt, people may be more likely to increase their workday physical activity, but the effect on sitting time may be attenuated.'}",https://www.cdc.gov/pcd/issues/2014/pdf/13_0318.pdf
-Modeling Thesis Clarity in Student Essays,Isaac Persing,"Recently, researchers have begun exploring methods of scoring student essays with respect to particular dimensions of quality such as coherence, technical errors, and relevance to prompt, but there is relatively little work on modeling thesis clarity. We present a new annotated corpus and propose a learning-based approach to scoring essays along the thesis clarity dimension. Additionally, in order to provide more valuable feedback on why an essay is scored as it is, we propose a second learning-based approach to identifying what kinds of errors an essay has that may lower its thesis clarity score.","{'model': 'tldr@v2.0.0', 'text': 'A new annotated corpus is presented and a learning- based approach to scoring essays along the thesis clarity dimension is proposed and a second learning-based approach to identifying what kinds of errors an essay has that may lower its thesis clarity score is proposed.'}",
-Nuclear Reaction Prompt Gamma-Ray Analysis,G. Molnár,,,
-Validation of multidimensional scaling-based modeling of alcohol expectancies in memory: age and drinking-related differences in expectancies of children assessed as first associates.,M. Dunn,"BACKGROUND As evidence has accumulated that alcohol expectancies mediate the effects of other drinking antecedents, attempts to understand the mechanism by which expectancies influence behavior have focused on modeling memory processes. Previous expectancy work, however, has used relatively indirect approaches to retrieve and model information stored in memory. By using the method most recommended by memory researchers for directly obtaining uncontaminated memory contents, we assessed children's expectancies and related findings to empirically modeled organization and activation of expectancies in memory based on scaled instruments. METHODS Individual interviews were conducted with 462 children in 2nd through 5th grades, and surveys were completed by 1,003 children in 3rd, 6th, 9th, and 12th grades. Interviews and surveys consisted of a measure designed to retrieve participants' first expectancy associate to an alcohol prompt and several drinking quantity/frequency questions. RESULTS Older and higher drinking children were more likely to report positive expectancies as their first associate to an alcohol prompt. Age and drinking-based findings were consistent with organizational structure, dimension emphasis shift, and paths of association identified by prior multidimensional scaling techniques. CONCLUSIONS Consumption of alcohol among children corresponded to accessibility of positive expectancies in memory. In addition, the use of multidimensional scaling to study the organization and activation of alcohol expectancies in memory was validated.","{'model': 'tldr@v2.0.0', 'text': ""Children's expectancies and related findings were assessed to empirically modeled organization and activation of expectancies in memory based on scaled instruments and the use of multidimensional scaling to study the organization andactivation of alcohol expectancy in memory was validated.""}",
-Increasing the vocalizations of individuals with autism during intervention with a speech-generating device.,Cindy Gevarter,"This study aimed to teach individuals with autism spectrum disorder (ASD) and limited vocal speech to emit target vocalizations while using a speech-generating device (SGD). Of the 4 participants, 3 began emitting vocal word approximations with SGD responses after vocal instructional methods (delays, differential reinforcement, prompting) were introduced. Two participants met mastery criterion with a reinforcer delay and differential reinforcement, and 1 met criterion after fading an echoic model and prompt delay. For these participants, vocalizations initiated before speech outputs were shown to increase, and vocalizations generalized to a context in which the SGD was absent. The 4th participant showed high vocalization rates only when prompted. The results suggest that adding vocal instruction to an SGD-based intervention can increase vocalizations emitted along with SGD responses for some individuals with ASD.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that adding vocal instruction to an SGD-based intervention can increase vocalizations emitted along with SGD responses for some individuals with ASD.'}",
-Regression-based quantitative-trait-locus mapping in the 21st century.,E. Feingold,"In the beginning, there was Haseman-Elston regression. This tool for human QTL mapping, developed in 1972, was simple and inspired. The idea was to take pairs of siblings and regress the squared differences in their trait values on their identity-by-descent (IBD) sharing at a marker. If the marker is linked to the trait, high levels of IBD sharing should be associated with a small difference in trait values, and the regression slope should be negative. Thus, linkage can be tested with a regression t test. This method (with some extensions) was predominant in human studies for >20 years, which was primarily a reflection of the fact that too little human QTL mapping was being performed to prompt the development of more sophisticated methods. In the mid-1990s, we saw the first important alternative to Haseman-Elston regression, maximum-likelihood–based variance-components estimation (see, e.g., Amos 1994; Almasy and Blangero 1998). Variance components is seamlessly applicable to any type of pedigree, whereas Haseman-Elston regression is not, and it has substantially higher power than Haseman-Elston when trait distributions are approximately Gaussian. It has superseded Haseman-Elston as the method of choice for most studies, particularly when large pedigrees are used. However, variance components relies heavily on normality assumptions and can fail dramatically when those assumptions are violated either by nonnormality of the trait distribution or by selected sampling. Attempts to “robustify” variance components have had mixed success (see Feingold [2001] for a more complete discussion), so there is still a role for regression-based methods, which are intrinsically more robust. In the past 5 years, there has been an avalanche of attempts to improve the power of Haseman-Elston regression and to bring regression-based QTL mapping up to date. This was set off by Wright's (1997) Letter to the Editor suggesting that it is beneficial to use the trait values of both members of a sib pair rather than just the squared difference (although this was, in fact, pointed out by Gaines and Elston [1969]). Since then, there have been six articles suggesting “revised Haseman-Elston” (regression-based) methods that use the bivariate data—by Drigalenko (1998), Elston et al. (2000), Xu et al. (2000), Forrest (2001), Sham and Purcell (2001), and Visscher and Hopper (2001). I believe that this is a complete list, but I offer profound apologies to anyone I may have omitted. There have also been three new articles discussing score statistics that have properties similar to the regression-based methods, by Tang and Siegmund (2001), Putter et al. (2002), and Wang and Huang (2002). The best of these new methods have succeeded in matching the power of variance components while retaining the robustness of the regression framework. However, they are all limited to sibships, or, in some cases, to sib pairs. In this issue of the Journal, Sham et al. (2002) take the logical next step, by developing a regression-based method that can be applied to extended pedigrees. Those of us trying to map human QTLs have a much richer set of tools available to us than we did 5 years ago. However, the abundance of new methods has made it difficult to make choices. Only true aficionados can keep up with the literature. In this editorial, I briefly review the newest options. I will describe the new regression-based methods and score statistics, compare their strengths and weaknesses, and conclude by describing how the current offering from Sham et al. (2002 [in this issue]) fits in. I will start, however, with a disclaimer. Because all of these methods are very new, they have not been tested extensively. Most of my observations below are based on statistical theory, and I'm sure that further study of the statistics will prove at least some of my guesses wrong. A related caveat is that all of the theory I rely on is large-sample theory, and even among statistics that are asymptotically identical there may be important differences in small-sample behavior.","{'model': 'tldr@v2.0.0', 'text': 'The new regression-based methods and score statistics are described, compare their strengths and weaknesses, and conclude by describing how the current offering from Sham et al. (2002) fits in.'}",http://www.cell.com/article/S0002929707604686/pdf
-PUnifiedNER: a Prompting-based Unified NER System for Diverse Datasets,Jinghui Lu,"Much of named entity recognition (NER) research focuses on developing dataset-specific models based on data from the domain of interest, and a limited set of related entity types. This is frustrating as each new dataset requires a new model to be trained and stored. In this work, we present a ``versatile'' model---the Prompting-based Unified NER system (PUnifiedNER)---that works with data from different domains and can recognise up to 37 entity types simultaneously, and theoretically it could be as many as possible. By using prompt learning, PUnifiedNER is a novel approach that is able to jointly train across multiple corpora, implementing intelligent on-demand entity recognition. Experimental results show that PUnifiedNER leads to significant prediction benefits compared to dataset-specific models with impressively reduced model deployment costs. Furthermore, the performance of PUnifiedNER can achieve competitive or even better performance than state-of-the-art domain-specific methods for some datasets. We also perform comprehensive pilot and ablation studies to support in-depth analysis of each component in PUnifiedNER.","{'model': 'tldr@v2.0.0', 'text': ""This work presents a ``versatile'' model---the Prompting-based Unified NER system (PUnifiedNER)---that works with data from different domains and can recognise up to 37 entity types simultaneously, and theoretically it could be as many as possible.""}",http://arxiv.org/pdf/2211.14838
-Prompting through Prototype: A Prototype-based Prompt Learning on Pretrained Vision-Language Models,Yue Zhang,"Prompt learning is a new learning paradigm which reformulates downstream tasks as similar pretraining tasks on pretrained models by leveraging textual prompts. Recent works have demonstrated that prompt learning is particularly useful for few-shot learning, where there is limited training data. Depending on the granularity of prompts, those methods can be roughly divided into task-level prompting and instance-level prompting. Task-level prompting methods learn one universal prompt for all input samples, which is efficient but ineffective to capture subtle differences among different classes. Instance-level prompting methods learn a specific prompt for each input, though effective but inefficient. In this work, we develop a novel prototype-based prompt learning method to overcome the above limitations. In particular, we focus on few-shot image recognition tasks on pretrained vision-language models (PVLMs) and develop a method of prompting through prototype (PTP), where we define $K$ image prototypes and $K$ prompt prototypes. In PTP, the image prototype represents a centroid of a certain image cluster in the latent space and a prompt prototype is defined as a soft prompt in the continuous space. The similarity between a query image and an image prototype determines how much this prediction relies on the corresponding prompt prototype. Hence, in PTP, similar images will utilize similar prompting ways. Through extensive experiments on seven real-world benchmarks, we show that PTP is an effective method to leverage the latent knowledge and adaptive to various PVLMs. Moreover, through detailed analysis, we discuss pros and cons for prompt learning and parameter-efficient fine-tuning under the context of few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work focuses on few-shot image recognition tasks on pretrained vision-language models (PVLMs) and develops a method of prompting through prototype (PTP), where the image prototype represents a centroid of a certain image cluster in the latent space and a prompt prototype is defined as a soft prompt in the continuous space.'}",http://arxiv.org/pdf/2210.10841
-Self-Prompting Large Language Models for Open-Domain QA,Junlong Li,"Open-Domain Question Answering (ODQA) requires models to answer factoid questions with no context given. The common way for this task is to train models on a large-scale annotated dataset to retrieve related documents and generate answers based on these documents. In this paper, we show that the ODQA architecture can be dramatically simplified by treating Large Language Models (LLMs) as a knowledge corpus and propose a Self-Prompting framework for LLMs to perform ODQA so as to eliminate the need for training data and external knowledge corpus. Concretely, we firstly generate multiple pseudo QA pairs with background passages and one-sentence explanations for these QAs by prompting LLMs step by step and then leverage the generated QA pairs for in-context learning. Experimental results show our method surpasses previous state-of-the-art methods by +8.8 EM averagely on three widely-used ODQA datasets, and even achieves comparable performance with several retrieval-augmented fine-tuned models.","{'model': 'tldr@v2.0.0', 'text': 'This paper shows that the ODQA architecture can be dramatically simplified by treating Large Language Models (LLMs) as a knowledge corpus and proposes a Self-Prompting framework for LLMs to perform ODZA so as to eliminate the need for training data and external knowledge corpus.'}",http://arxiv.org/pdf/2212.08635
-Dialogue State Tracking with a Language Model using Schema-Driven Prompting,Chia-Hsuan Lee,"Task-oriented conversational systems often use dialogue state tracking to represent the user’s intentions, which involves filling in values of pre-defined slots. Many approaches have been proposed, often using task-specific architectures with special-purpose classifiers. Recently, good results have been obtained using more general architectures based on pretrained language models. Here, we introduce a new variation of the language modeling approach that uses schema-driven prompting to provide task-aware history encoding that is used for both categorical and non-categorical slots. We further improve performance by augmenting the prompting with schema descriptions, a naturally occurring source of in-domain knowledge. Our purely generative system achieves state-of-the-art performance on MultiWOZ 2.2 and achieves competitive performance on two other benchmarks: MultiWOZ 2.1 and M2M. The data and code will be available at https://github.com/chiahsuan156/DST-as-Prompting.","{'model': 'tldr@v2.0.0', 'text': 'A new variation of the language modeling approach that uses schema-driven prompting to provide task-aware history encoding that is used for both categorical and non-categorical slots is introduced.'}",https://aclanthology.org/2021.emnlp-main.404.pdf
-MAPL: Parameter-Efficient Adaptation of Unimodal Pre-Trained Models for Vision-Language Few-Shot Prompting,Oscar Mañas,"Large pre-trained models have proved to be remarkable zero- and (prompt-based) few-shot learners in unimodal vision and language tasks. We propose MAPL, a simple and parameter-efficient method that reuses frozen pre-trained unimodal models and leverages their strong generalization capabilities in multimodal vision-language (VL) settings. MAPL learns a lightweight mapping between the representation spaces of unimodal models using aligned image-text data, and can generalize to unseen VL tasks from just a few in-context examples. The small number of trainable parameters makes MAPL effective at low-data and in-domain learning. Moreover, MAPL’s modularity enables easy extension to other pre-trained models. Extensive experiments on several visual question answering and image captioning benchmarks show that MAPL achieves superior or competitive performance compared to similar methods while training orders of magnitude fewer parameters. MAPL can be trained in just a few hours using modest computational resources and public datasets. We release our code and pre-trained model weights at https://github.com/oscmansan/mapl.","{'model': 'tldr@v2.0.0', 'text': 'MAPL is a simple and parameter-efficient method that reuses frozen pre-trained unimodal models and leverages their strong generalization capabilities in multimodal vision-language (VL) settings and learns a lightweight mapping between the representation spaces of unimmodal models using aligned image-text data.'}",http://arxiv.org/pdf/2210.07179
-Prompting medical students to self–assess their learning needs during the ageing and health module: a mixed methods study,Grace Kennedy,"ABSTRACT Understanding our learning needs is fundamental for safe, effective and knowledge-based medical practice and facilitates life-long learning. A mixed methods study investigated fourth-year medical students’ self-perceived understanding of their learning needs using 1] a visual scale, before and after a four-week module in Ageing and Health (A&H) and 2] through focus group discussions. During 2013–14 academic year, all students (252) were invited to use a Visual Analogue Scale (VAS) tool to self-assess their learning needs that were linked to Ageing and Health curriculum learning outcomes. Assenting students (197 at pre-self-assessment, 201 at post-assessment) returned anonymous Visual Analogue Scales, self-assessing history-taking skills, examination skills, knowledge of medication use, co-morbidity, nutritional and swallowing assessment responses, before and after the A&H module. Three student focus groups explored whether completion of the VAS self-assessment had prompted improved self-awareness of their learning needs. The VAS responses increased for each curriculum domain with significant differences between the pre-and post responses – for the student-year-group. Nutritional and swallowing knowledge showed the greatest improvement from a self-assessed low baseline at entry. Focus-group students generally viewed the VAS tool positively, and as an aid for prompting consideration of current and future clinical practice. Some students recognised that ‘a need to be ready-for-work’ focused engaged learning; others demonstrated self-regulated learning through self-motivation and an action plan. The Visual Analogue Scale quantitative responses showed increased student-self-perceived learning for each curriculum domain at fourth-year completion of the A&H module, suggesting that prompting self-assessment had increased students’ knowledge and skills. Focus group students saw the VAS tool as useful for prompting awareness of their current and future learning needs. Additional educational strategies should be explored to enable all students to self-reflect and engage effectively on their learning needs, to gain the skills for the maintenance of professional medical competence. Abbreviations: A&H: Ageing and Health Module; e-portfolio: an electronic version of an evidence portfolio, which allows medical students and graduates to reflect and document learning and competencies; F1: year1 of post-graduate medical clinical training; GMC: General Medical Council-the regulation organisation for maintaining standards for doctors in UK; Logbook: usually a written document which can be used to record procedures and attendance at clinics or case-based discussions and can be used to set learning outcomes and to structure teaching in clinical settings for medical students and doctors; PDP: personal development plan is used to plan future learning and skills needs for work and education with an plan for action/s outcome; SPSS: Statistical Package for the Social Sciences; VAS: Visual Analogue Scale is a visual method of describing an experience.","{'model': 'tldr@v2.0.0', 'text': 'The Visual Analogue Scale quantitative responses showed increased student-self-perceived learning for each curriculum domain at fourth-year completion of the A&H module, suggesting that prompting self-assessment had increased students’ knowledge and skills.'}",https://www.tandfonline.com/doi/pdf/10.1080/10872981.2019.1579558?needAccess=true
-TransPrompt: Towards an Automatic Transferable Prompting Framework for Few-shot Text Classification,Chengyu Wang,"Recent studies have shown that prompts improve the performance of large pre-trained language models for few-shot text classification. Yet, it is unclear how the prompting knowledge can be transferred across similar NLP tasks for the purpose of mutual reinforcement. Based on continuous prompt embeddings, we propose TransPrompt, a transferable prompting framework for few-shot learning across similar tasks. In TransPrompt, we employ a multi-task meta-knowledge acquisition procedure to train a meta-learner that captures cross-task transferable knowledge. Two de-biasing techniques are further designed to make it more task-agnostic and unbiased towards any tasks. After that, the meta-learner can be adapted to target tasks with high accuracy. Extensive experiments show that TransPrompt outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets. We further show that the meta-learner can effectively improve the performance on previously unseen tasks; and TransPrompt also outperforms strong fine-tuning baselines when learning with full training sets.","{'model': 'tldr@v2.0.0', 'text': 'In TransPrompt, a transferable prompting framework for few-shot learning across similar NLP tasks, a multi-task meta-knowledge acquisition procedure is employed to train a meta-learner that captures cross-task transferable knowledge.'}",https://aclanthology.org/2021.emnlp-main.221.pdf
-MEAL: Stable and Active Learning for Few-Shot Prompting,Abdullatif Köksal,"Few-shot classification has made great strides due to foundation models that, through priming and prompting, are highly effective few-shot learners. However, this approach has high variance both across different sets of few shots (data selection) and across different finetuning runs (run variability). This is problematic not only because it impedes the fair comparison of different approaches, but especially because it makes few-shot learning too unreliable for many real-world applications. To alleviate these issues, we make two contributions for more stable and effective few-shot learning: First, we propose novel ensembling methods and show that they substantially reduce run variability. Second, we introduce a new active learning (AL) criterion for data selection and present the first AL-based approach specifically tailored towards prompt-based learning. In our experiments, we show that our combined method, MEAL (Multiprompt finetuning and prediction Ensembling with Active Learning), improves overall performance of prompt-based finetuning by 2.3 points on five diverse tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes novel ensembling methods and introduces a new active learning (AL) criterion for data selection and presents the first AL-based approach specifically tailored towards prompt-based learning, and shows that the combined method, MEAL, improves overall performance of prompt- based finetuning by 2.3 points on five diverse tasks.'}",http://arxiv.org/pdf/2211.08358
-Context-faithful Prompting for Large Language Models,Wenxuan Zhou,"Large language models (LLMs) encode parametric knowledge about world facts and have shown remarkable performance in knowledge-driven NLP tasks. However, their reliance on parametric knowledge may cause them to overlook contextual cues, leading to incorrect predictions in context-sensitive NLP tasks (e.g., knowledge acquisition tasks). In this paper, we seek to assess and enhance LLMs' contextual faithfulness in two aspects: knowledge conflict and prediction with abstention. We demonstrate that LLMs' faithfulness can be significantly improved using carefully designed prompting strategies. In particular, we identify opinion-based prompts and counterfactual demonstrations as the most effective methods. Opinion-based prompts reframe the context as a narrator's statement and inquire about the narrator's opinions, while counterfactual demonstrations use instances containing false facts to improve faithfulness in knowledge conflict situations. Neither technique requires additional training. We conduct experiments on three datasets of two standard NLP tasks, machine reading comprehension and relation extraction, and the results demonstrate significant improvement in faithfulness to contexts. Code and data are released at https://github.com/wzhouad/context-faithful-llm.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that LLMs' faithfulness can be significantly improved using carefully designed prompting strategies, and opinion-based prompts and counterfactual demonstrations are identified as the most effective methods.""}",http://arxiv.org/pdf/2303.11315
-CORA: Adapting CLIP for Open-Vocabulary Detection with Region Prompting and Anchor Pre-Matching,Xiaoshi Wu,"Open-vocabulary detection (OVD) is an object detection task aiming at detecting objects from novel categories beyond the base categories on which the detector is trained. Recent OVD methods rely on large-scale visual-language pre-trained models, such as CLIP, for recognizing novel objects. We identify the two core obstacles that need to be tackled when incorporating these models into detector training: (1) the distribution mismatch that happens when applying a VL-model trained on whole images to region recognition tasks; (2) the difficulty of localizing objects of unseen classes. To overcome these obstacles, we propose CORA, a DETR-style framework that adapts CLIP for Open-vocabulary detection by Region prompting and Anchor pre-matching. Region prompting mitigates the whole-to-region distribution gap by prompting the region features of the CLIP-based region classifier. Anchor pre-matching helps learning generalizable object localization by a class-aware matching mechanism. We evaluate CORA on the COCO OVD benchmark, where we achieve 41.7 AP50 on novel classes, which outperforms the previous SOTA by 2.4 AP50 even without resorting to extra training data. When extra training data is available, we train CORA+ on both ground-truth base-category annotations and additional pseudo bounding box labels computed by CORA. CORA+ achieves 43.1 AP50 on the COCO OVD benchmark and 28.1 box APr on the LVIS OVD benchmark. The code is available at https://github.com/tgxs002/CORA.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes CORA, a DETR-style framework that adapts CLIP for Open-vocabulary detection by Region prompting and Anchor pre-matching, which mitigates the whole-to-region distribution gap by prompting the region features of the CLIP-based region classifier.'}",https://arxiv.org/pdf/2303.13076
-System Impact Studies for Near 100% Renewable Energy Systems Dominated by Inverter Based Variable Generation,H. Holttinen,"The demand for low carbon energy calls for close to 100% renewable power systems, with decarbonization of other energy sectors adding to the anticipated paradigm shift. Rising levels of variable inverter-based renewable energy sources (VIBRES) are prompting questions about how such systems will be planned and operated when variable renewable generation becomes the dominant technology. Here, we examine the implications of this paradigm shift with respect to planning, operation and system stability, also addressing the need for integration with other energy vectors, including heat, transport and Power-to-X. We highlight the knowledge gaps and provide recommendations for improved methods and models needed as power systems transform towards 100% VIBRES.",,
-Information and communication technology based prompting for treatment compliance for people with serious mental illness.,Kaisa Kauppi,"BACKGROUND Non-compliance is a significant problem among people with serious mental disorders, presenting a challenge for mental health professionals. Prompts such as telephone calls, visits, and a posted referral letter to patients are currently used to encourage patient attendance at clinics and/or compliance with medication. More recently, the use of information and communication technology (ICT)-based prompting methods have increased. Methods include mobile text message (SMS - short message service), e-mail or use of any other electronic device with the stated purpose of encouraging compliance. OBJECTIVES To investigate the effects of ICT-based prompting to support treatment compliance in people with serious mental illness compared with standard care. SEARCH METHODS We searched the Cochrane Schizophrenia Group's Trials Register (31(st) May 2011 and 9(th) July 2012) which is based on regular searches of CINAHL, BIOSIS, AMED, EMBASE, PubMed, MEDLINE, PsycINFO, and registries of clinical trials. Also, we inspected references of all identified studies for further trials and contacted authors of trials for additional information. SELECTION CRITERIA Relevant randomised controlled trials involving adults with serious mental illness, comparing any ICT-based prompt or combination of prompts by automatic or semi-automatic system compared with standard care. DATA COLLECTION AND ANALYSIS Review authors reliably assessed trial quality and extracted data. We calculated risk ratio (RR) with 95% confidence intervals (CI) using a fixed-effect model. For continuous outcomes, we estimated the mean difference (MD) between groups, again with 95% confidence intervals. A 'Summary of findings' table using GRADE was created, and we assessed included studies for risk of bias. MAIN RESULTS The search identified 35 references, with 25 studies, but we could only include two studies with a total of 358 participants. The studies had a moderate risk of bias, and therefore risk overestimating any positive effects of ICT-based prompting. Both included studies compared semi-automatised ICT-based prompting intervention with standard care groups in mental health outpatient care. The interventions were SMS-message and an electronic assistant device. One included study reported our primary outcome, compliance.There was not any clear evidence that ICT-based prompts increase improvement in compliance (stop taking medication within six months n = 320, RR 1.11 CI 0.96 to 1.29, moderate quality evidence). There was some low quality evidence that ICT-based prompts have small effects for: mental state (average change in specific symptom scores within three months n = 251, MD -0.30 CI -0.53 to -0.07; severity of illness within three months n = 251, MD -0.10 CI -0.13 to -0.07 and six months n = 251, MD -0.10 CI -0.13 to -0.07; average change in depressive scores within six months n = 251, RR 0.00 CI -0.28 to 0.28; global symptoms within three months n = 251, MD -0.10 CI -0.38 to -0.07; negative symptoms within three months n = 251, MD -0.10 CI -0.38 to 0.18 and six months n = 251, MD -0.30 CI -0.58 to 0.02, low quality evidence). Level of insight improved more among people receiving ICT-based prompt compared with those in the control group at six months (n = 251, MD -0.10 CI -0.13 to -0.07). ICT-based prompts also increased quality of life (average change in quality of life within six months n = 251, RR 0.50 CI 0.19 to 0.81, moderate quality evidence).Based on the existing data, there is no evidence that either intervention is less acceptable than the other (n = 347, 2 RCTs, RR 1.46 CI 0.70 to 3.05, low quality evidence). Included studies did not report outcomes of service utilisation, behaviour, costs or adverse events. AUTHORS' CONCLUSIONS The evidence base on the effects of ICT-based prompts is still inconclusive. Data to clarify ICT-based prompting effects are awaited from an ongoing trial, but further well-conducted trials considering the different ICT-based prompts are warranted.","{'model': 'tldr@v2.0.0', 'text': 'There was not any clear evidence that ICT-based prompts increase improvement in compliance, and both included studies compared semi-automatised I CT-based prompting intervention with standard care groups in mental health outpatient care.'}",
-Prompting technologies: A comparison of time-based and context-aware transition-based prompting.,Kayela Robertson,"BACKGROUND While advancements in technology have encouraged the development of novel prompting systems to support cognitive interventions, little research has evaluated the best time to deliver prompts, which may impact the effectiveness of these interventions. OBJECTIVE This study examined whether transition-based context prompting (prompting an individual during task transitions) is more effective than traditional fixed time-based prompting. METHODS Participants were 42 healthy adults who completed 12 different everyday activities, each lasting 1-7 minutes, in an experimental smart home testbed and received prompts to record the completed activities from an electronic memory notebook. Half of the participants were delivered prompts during activity transitions, while the other half received prompts every 5 minutes. Participants also completed Likert-scale ratings regarding their perceptions of the prompting system. RESULTS Results revealed that participants in the transition-based context prompting condition responded to the first prompt more frequently and rated the system as more convenient, natural, and appropriate compared to participants in the time-based condition. CONCLUSIONS Our findings suggest that prompting during activity transitions produces higher adherence to the first prompt and more positive perceptions of the prompting system. This is an important finding given the benefits of prompting technology and the possibility of improving cognitive interventions by using context-aware transition prompting.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that prompting during activity transitions produces higher adherence to the first prompt and more positive perceptions of the prompting system, an important finding given the benefits of prompting technology and the possibility of improving cognitive interventions by using context-aware transition prompting.'}",https://europepmc.org/articles/pmc4803438?pdf=render
-The Effect of Video Modeling and Video Prompting Interventions on Individuals With Intellectual Disability: A Systematic Literature Review,Jiyoon Park,"A popular technology-based intervention is video-based instruction (VBI). VBI is known to be effective in teaching various skills to individuals with intellectual disability. Although many researchers have taught this population various skills using video modeling and/or prompting, none have conducted a literature review for this population. The purpose of this review was to analyze different types of VBI (i.e., video modeling, video prompting) used in intervention research, the skills taught using these methods, and their effectiveness. The findings demonstrate both video modeling and video prompting were used to similar extents, daily/living skills were taught more often, and many studies combined video modeling or video prompting interventions with additional strategies (e.g., error correction, constant time delay).",,https://journals.sagepub.com/doi/pdf/10.1177/0162643418780464
-"Standardization of sample collection, isolation and analysis methods in extracellular vesicle research",K. Witwer,"The emergence of publications on extracellular RNA (exRNA) and extracellular vesicles (EV) has highlighted the potential of these molecules and vehicles as biomarkers of disease and therapeutic targets. These findings have created a paradigm shift, most prominently in the field of oncology, prompting expanded interest in the field and dedication of funds for EV research. At the same time, understanding of EV subtypes, biogenesis, cargo and mechanisms of shuttling remains incomplete. The techniques that can be harnessed to address the many gaps in our current knowledge were the subject of a special workshop of the International Society for Extracellular Vesicles (ISEV) in New York City in October 2012. As part of the “ISEV Research Seminar: Analysis and Function of RNA in Extracellular Vesicles (evRNA)”, 6 round-table discussions were held to provide an evidence-based framework for isolation and analysis of EV, purification and analysis of associated RNA molecules, and molecular engineering of EV for therapeutic intervention. This article arises from the discussion of EV isolation and analysis at that meeting. The conclusions of the round table are supplemented with a review of published materials and our experience. Controversies and outstanding questions are identified that may inform future research and funding priorities. While we emphasize the need for standardization of specimen handling, appropriate normative controls, and isolation and analysis techniques to facilitate comparison of results, we also recognize that continual development and evaluation of techniques will be necessary as new knowledge is amassed. On many points, consensus has not yet been achieved and must be built through the reporting of well-controlled experiments.","{'model': 'tldr@v2.0.0', 'text': 'The need for standardization of specimen handling, appropriate normative controls, and isolation and analysis techniques to facilitate comparison of results is emphasized, and it is recognized that continual development and evaluation of techniques will be necessary as new knowledge is amassed.'}",https://www.tandfonline.com/doi/pdf/10.3402/jev.v2i0.20360?needAccess=true&role=button
-A randomized clinical trial of three prompting systems to teach tact relations.,Joseph H. Cihon,"Prompts are commonly used during discrete trial teaching for individuals diagnosed with autism spectrum disorder (ASD). Three commonly used prompting systems include constant time delay, most-to-least prompting, and flexible prompt fading. Most of the research demonstrating the effectiveness of these three prompting strategies have been completed through the use of single subject experimental designs. Some within the field of behaviorally based approaches to ASD interventions have called for more randomized clinical trials of these approaches. The purpose of the present study was to compare these 3 prompting systems to teach tact relations for 27 individuals diagnosed with ASD through a randomized clinical trial without a control group with respect to pre-post responding, generalization, sessions to mastery, and responding during teaching. Overall, minimal differences were observed across the 3 systems. The results are discussed with respect to future research and clinical application of the methods evaluated.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of the present study was to compare these 3 prompting systems to teach tact relations for 27 individuals diagnosed with ASD through a randomized clinical trial without a control group with respect to pre-post responding, generalization, sessions to mastery, and responding during teaching.'}",
-Use of a Smartphone Application for Prompting Oral Medication Adherence Among Adolescents and Young Adults With Cancer,Yelena P. Wu,"OBJECTIVES To explore the feasibility and acceptability of use of a smartphone medication reminder application to promote adherence to oral medications among adolescents and young adults (AYAs) with cancer.. SAMPLE & SETTING 23 AYAs with cancer from a Children's Oncology Group-affiliated children's hospital and a National Cancer Institute-designated comprehensive cancer center in Salt Lake City, UT.. METHODS & VARIABLES Participants were asked to use the application for eight weeks. Data on application usage were obtained from a cloud-based server hosted by the application developers. Weekly self-report questionnaires were completed. Feasibility was assessed through participants' usage and responses. Acceptability was assessed through participants' perceived ease of use and usefulness.. RESULTS Almost all participants used the application at least once. More than half reported that they took their medications immediately when they received reminders. Participants also reported that the application was easy to set up and use, and that it was useful for prompting medication taking.. IMPLICATIONS FOR NURSING Nurses could continue to test the efficacy of integrating e-health modalities, such as smartphone applications, into efforts to promote medication adherence.","{'model': 'tldr@v2.0.0', 'text': 'Nursing nurses could continue to test the efficacy of integrating e-health modalities, such as smartphone applications, into efforts to promote medication adherence by exploring the feasibility and acceptability of use of a smartphone medication reminder application among adolescents and young adults with cancer.'}",https://europepmc.org/articles/pmc5927589?pdf=render
-Acceptability of artificial intelligence (AI)-led chatbot services in healthcare: A mixed-methods study,T. Nadarzynski,"Background Artificial intelligence (AI) is increasingly being used in healthcare. Here, AI-based chatbot systems can act as automated conversational agents, capable of promoting health, providing education, and potentially prompting behaviour change. Exploring the motivation to use health chatbots is required to predict uptake; however, few studies to date have explored their acceptability. This research aimed to explore participants’ willingness to engage with AI-led health chatbots. Methods The study incorporated semi-structured interviews (N-29) which informed the development of an online survey (N-216) advertised via social media. Interviews were recorded, transcribed verbatim and analysed thematically. A survey of 24 items explored demographic and attitudinal variables, including acceptability and perceived utility. The quantitative data were analysed using binary regressions with a single categorical predictor. Results Three broad themes: ‘Understanding of chatbots’, ‘AI hesitancy’ and ‘Motivations for health chatbots’ were identified, outlining concerns about accuracy, cyber-security, and the inability of AI-led services to empathise. The survey showed moderate acceptability (67%), correlated negatively with perceived poorer IT skills OR = 0.32 [CI95%:0.13–0.78] and dislike for talking to computers OR = 0.77 [CI95%:0.60–0.99] as well as positively correlated with perceived utility OR = 5.10 [CI95%:3.08–8.43], positive attitude OR = 2.71 [CI95%:1.77–4.16] and perceived trustworthiness OR = 1.92 [CI95%:1.13–3.25]. Conclusion Most internet users would be receptive to using health chatbots, although hesitancy regarding this technology is likely to compromise engagement. Intervention designers focusing on AI-led health chatbots need to employ user-centred and theory-based approaches addressing patients’ concerns and optimising user experience in order to achieve the best uptake and utilisation. Patients’ perspectives, motivation and capabilities need to be taken into account when developing and assessing the effectiveness of health chatbots.","{'model': 'tldr@v2.0.0', 'text': 'Intervention designers focusing on AI-led health chatbots need to employ user-centred and theory-based approaches addressing patients’ concerns and optimising user experience in order to achieve the best uptake and utilisation.'}",https://journals.sagepub.com/doi/pdf/10.1177/2055207619871808
-Automated Detection of Activity Transitions for Prompting,Kyle D. Feuz,"Individuals with cognitive impairment can benefit from intervention strategies like recording important information in a memory notebook. However, training individuals to use the notebook on a regular basis requires a constant delivery of reminders. In this study, we design and evaluate machine-learning-based methods for providing automated reminders using a digital memory notebook interface. Specifically, we identify transition periods between activities as times to issue prompts. We consider the problem of detecting activity transitions using supervised and unsupervised machine-learning techniques and find that both techniques show promising results for detecting transition periods. We test the techniques in a scripted setting with 15 individuals. Motion sensors data are recorded and annotated as participants perform a fixed set of activities. We also test the techniques in an unscripted setting with eight individuals. Motion sensor data are recorded as participants go about their normal daily routine. In both the scripted and unscripted settings, a true positive rate of greater than 80% can be achieved while maintaining a false positive rate of less than 15%. On average, this leads to transitions being detected within 1 min of a true transition for the scripted data and within 2 min of a true transition on the unscripted data.","{'model': 'tldr@v2.0.0', 'text': 'This study designs and evaluates machine-learning-based methods for providing automated reminders using a digital memory notebook interface and identifies transition periods between activities as times to issue prompts.'}",https://europepmc.org/articles/pmc4805372?pdf=render
-Training a Paraprofessional to Implement Video Prompting With Error Correction to Teach a Vocational Skill,Rachel L. Seaman-Tullis,"Very few individuals with autism spectrum disorder (ASD) have been trained in the vocational skills needed to obtain gainful employment. Moreover, although there is an abundance of research evaluating the practice of training practitioners of students with ASD to use evidence-based practices to teach a wide variety of skills, there have been few that apply this training to the acquisition of vocational tasks. This study uses a multiple-baseline-across-behaviors design to evaluate the training methods used to train a paraprofessional in the preparation and implementation of video prompting and error correction with his student with ASD. Furthermore, the behavior and learning of both the paraprofessional and student are measured. Results indicate that the training package resulted in increased video prompting implementation behavior for the paraprofessional, as well as corresponding, increased vocational skill behavior for the student.",,
-Using a Simultaneous Prompting Procedure with an iPad to Teach the Pythagorean Theorem to Adolescents with Moderate Intellectual Disability,Carey Creech-Galloway,"Providing standards-based instruction in core content areas for individuals with moderate and severe disabilities is a hot topic in the field of special education, and teachers struggle to find the best methods for providing high-quality standards-based instruction in core content areas that also has personal relevance for the students. This investigation evaluated the effectiveness of a simultaneous prompting procedure to teach four adolescents with moderate intellectual disabilities to use the Pythagorean theorem to solve real-life scenarios (i.e., sewing, using a ladder, finding dimensions of a screen) shown on a short video on an iPad. A multiple probe design across participants evaluated the effectiveness of the procedure. Results indicate the participants acquired the skill of using the Pythagorean theorem and generalized it to additional real-life, novel problems. Social validity, limitations, and implications for practitioners are also discussed.",,
-A Systematic Review of Methods and Procedures Used in Ecological Momentary Assessments of Diet and Physical Activity Research in Youth: An Adapted STROBE Checklist for Reporting EMA Studies (CREMAS),Y. Liao,"Background Ecological momentary assessment (EMA) is a method of collecting real-time data based on careful timing, repeated measures, and observations that take place in a participant’s typical environment. Due to methodological advantages and rapid advancement in mobile technologies in recent years, more studies have adopted EMA in addressing topics of nutrition and physical activity in youth. Objective The aim of this systematic review is to describe EMA methodology that has been used in studies addressing nutrition and physical activity in youth and provide a comprehensive checklist for reporting EMA studies. Methods Thirteen studies were reviewed and analyzed for the following 5 areas of EMA methodology: (1) sampling and measures, (2) schedule, (3) technology and administration, (4) prompting strategy, and (5) response and compliance. Results Results of this review showed a wide variability in the design and reporting of EMA studies in nutrition and physical activity among youth. The majority of studies (69%) monitored their participants during one period of time, although the monitoring period ranged from 4 to 14 days, and EMA surveys ranged from 2 to 68 times per day. More than half (54%) of the studies employed some type of electronic technology. Most (85%) of the studies used interval-contingent prompting strategy. For studies that utilized electronic devices with interval-contingent prompting strategy, none reported the actual number of EMA prompts received by participants out of the intended number of prompts. About half (46%) of the studies failed to report information about EMA compliance rates. For those who reported, compliance rates ranged from 44-96%, with an average of 71%. Conclusions Findings from this review suggest that in order to identify best practices for EMA methodology in nutrition and physical activity research among youth, more standardized EMA reporting is needed. Missing the key information about EMA design features and participant compliance might lead to misinterpretation of results. Future nutrition and physical activity EMA studies need to be more rigorous and thorough in descriptions of methodology and results. A reporting checklist was developed with the goal of enhancing reliability, efficacy, and overall interpretation of the findings for future studies that use EMAs.","{'model': 'tldr@v2.0.0', 'text': 'Findings from this review suggest that in order to identify best practices for EMA methodology in nutrition and physical activity research among youth, more standardized EMA reporting is needed.'}",
-Do universal school‐based mental health promotion programmes improve the mental health and emotional wellbeing of young people? A literature review,C. O’Connor,"AIMS AND OBJECTIVES To examine evidence-using a range of outcomes-for the effectiveness of school-based mental health and emotional well-being programmes. BACKGROUND It is estimated that 20% of young people experience mental health difficulties every year. Schools have been identified as an appropriate setting for providing mental health and emotional well-being promotion prompting the need to determine whether current school-based programmes are effective in improving the mental health and emotional well-being of young people. METHODS A systematic search was conducted using the health and education databases, which identified 29 studies that measured the effectiveness of school-based universal interventions. Prisma guidelines were used during the literature review process. RESULTS Thematic analysis generated three key themes: (i) help seeking and coping; (ii) social and emotional well-being; and (iii) psycho-educational effectiveness. CONCLUSION It is concluded that whilst these studies show promising results, there is a need for further robust evaluative studies to guide future practice. RELEVANCE TO CLINICAL PRACTICE All available opportunities should be taken to provide mental health promotion interventions to young people in the school environment, with a requirement for educational professionals to be provided the necessary skills and knowledge to ensure that the school setting continues to be a beneficial environment for conducting mental health promotion.","{'model': 'tldr@v2.0.0', 'text': 'There is a need for further robust evaluative studies to guide future practice on the effectiveness of school-based universal interventions in mental health and emotional well-being.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/jocn.14078
-A Novel Hybrid Sequence-Based Model for Identifying Anticancer Peptides,Lei Xu,"Cancer is a serious health issue worldwide. Traditional treatment methods focus on killing cancer cells by using anticancer drugs or radiation therapy, but the cost of these methods is quite high, and in addition there are side effects. With the discovery of anticancer peptides, great progress has been made in cancer treatment. For the purpose of prompting the application of anticancer peptides in cancer treatment, it is necessary to use computational methods to identify anticancer peptides (ACPs). In this paper, we propose a sequence-based model for identifying ACPs (SAP). In our proposed SAP, the peptide is represented by 400D features or 400D features with g-gap dipeptide features, and then the unrelated features are pruned using the maximum relevance-maximum distance method. The experimental results demonstrate that our model performs better than some existing methods. Furthermore, our model has also been extended to other classifiers, and the performance is stable compared with some state-of-the-art works.","{'model': 'tldr@v2.0.0', 'text': 'In this paper, a sequence-based model for identifying anticancer peptides (SAP) is proposed, and the experimental results demonstrate that the model performs better than some existing methods and is stable compared with some state-of-the-art works.'}",https://www.mdpi.com/2073-4425/9/3/158/pdf?version=1520937538
-B-MOBILE - A Smartphone-Based Intervention to Reduce Sedentary Time in Overweight/Obese Individuals: A Within-Subjects Experimental Trial,D. Bond,"Purpose Excessive sedentary time (SED) has been linked to obesity and other adverse health outcomes. However, few sedentary-reducing interventions exist and none have utilized smartphones to automate behavioral strategies to decrease SED. We tested a smartphone-based intervention to monitor and decrease SED in overweight/obese individuals, and compared 3 approaches to prompting physical activity (PA) breaks and delivering feedback on SED. Design and Methods Participants [N = 30; Age = 47.5(13.5) years; 83% female; Body Mass Index (BMI) = 36.2(7.5) kg/m2] wore the SenseWear Mini Armband (SWA) to objectively measure SED for 7 days at baseline. Participants were then presented with 3 smartphone-based PA break conditions in counterbalanced order: (1) 3-min break after 30 SED min; (2) 6-min break after 60 SED min; and (3) 12-min break after 120 SED min. Participants followed each condition for 7 days and wore the SWA throughout. Results All PA break conditions yielded significant decreases in SED and increases in light (LPA) and moderate-to-vigorous PA (MVPA) (p<0.005). Average % SED at baseline (72.2%) decreased by 5.9%, 5.6%, and 3.3% [i.e. by mean (95% CI) −47.2(−66.3, −28.2), −44.5(−65.2, −23.8), and −26.2(−40.7, −11.6) min/d] in the 3-, 6-, and 12-min conditions, respectively. Conversely, % LPA increased from 22.8% to 26.7%, 26.7%, and 24.7% [i.e. by 31.0(15.8, 46.2), 31.0(13.6, 48.4), and 15.3(3.9, 26.8) min/d], and % MVPA increased from 5.0% to 7.0%, 6.7%, and 6.3% (i.e. by 16.2(8.5, 24.0), 13.5(6.3, 20.6), and 10.8(4.2, 17.5) min/d] in the 3-, 6-, and 12-min conditions, respectively. Planned pairwise comparisons revealed the 3-min condition was superior to the 12-min condition in decreasing SED and increasing LPA (p<0.05). Conclusion The smartphone-based intervention significantly reduced SED. Prompting frequent short activity breaks may be the most effective way to decrease SED and increase PA in overweight/obese individuals. Future investigations should determine whether these SED reductions can be maintained long-term. Trial Registration ClinicalTrials.gov NCT01688804","{'model': 'tldr@v2.0.0', 'text': 'The smartphone-based intervention to monitor and decrease SED in overweight/obese individuals, and compared 3 approaches to prompting physical activity (PA) breaks and delivering feedback on SED significantly reduced SED, suggests that Prompting frequent short activity breaks may be the most effective way to decreasing SED and increase PA in obese individuals.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0100821&type=printable
-"Self-contradictory Hallucinations of Large Language Models: Evaluation, Detection and Mitigation",Niels Mündler,"Large language models (large LMs) are susceptible to producing text that contains hallucinated content. An important instance of this problem is self-contradiction, where the LM generates two contradictory sentences within the same context. In this work, we present a comprehensive investigation into self-contradiction for various instruction-tuned LMs, covering evaluation, detection, and mitigation. Our analysis reveals the prevalence of self-contradictions when LMs generate text for open-domain topics, e.g., in 17.7% of all sentences produced by ChatGPT. Self-contradiction also complements retrieval-based methods, as a large portion of them (e.g., 35.8% for ChatGPT) cannot be verified using Wikipedia. We then propose a novel prompting-based framework designed to effectively detect and mitigate self-contradictions. Our detector achieves high accuracy, e.g., around 80% F1 score when prompting ChatGPT. The mitigation algorithm iteratively refines the generated text to remove contradictory information while preserving text fluency and informativeness. Importantly, our entire framework is applicable to black-box LMs and does not require external grounded knowledge. Our approach is practically effective and has been released as a push-button tool to benefit the public, available at https://chatprotect.ai/.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a novel prompting-based framework designed to effectively detect and mitigate self-contradictions in instruction-tuned LMs, and proposes a novel mitigation algorithm that iteratively refines the generated text to remove contradictory information while preserving text fluency and informativeness.'}",https://arxiv.org/pdf/2305.15852
-Systematic Review of Instructional Methods to Teach Employment Skills to Secondary Students With Intellectual and Developmental Disabilities,Carly B. Gilson,"Effective vocational instruction is an essential aspect of preparing students with intellectual and developmental disabilities (IDD) for the world of work. We systematically reviewed research on instructional methods used to teach employment skills to secondary students with IDD. We identified 56 studies involving 766 participants with IDD. Four intervention approaches emphasized technology or some other instructional stimulus (i.e., self-management devices, video-based, audio-based, picture and tactile-based) and four focused on live instructors (i.e., direct instruction, augmentative and alternative communication, simulation, peer-delivered). Among the 21 instructional methods used within these approaches, performance feedback, device-assisted instruction, response prompting, and community-based instruction were the most common. We address the extent to which these intervention approaches were effective across students, instructional methods, settings, and outcomes, as well as offer recommendations for future research and practice.",,
-ConvFinQA: Exploring the Chain of Numerical Reasoning in Conversational Finance Question Answering,Zhiyu Chen,"With the recent advance in large pre-trained language models, researchers have achieved record performances in NLP tasks that mostly focus on language pattern matching. The community is experiencing the shift of the challenge from how to model language to the imitation of complex reasoning abilities like human beings. In this work, we investigate the application domain of finance that involves real-world, complex numerical reasoning. We propose a new large-scale dataset, ConvFinQA, aiming to study the chain of numerical reasoning in conversational question answering. Our dataset poses great challenge in modeling long-range, complex numerical reasoning paths in real-world conversations. We conduct comprehensive experiments and analyses with both the neural symbolic methods and the prompting-based methods, to provide insights into the reasoning mechanisms of these two divisions. We believe our new dataset should serve as a valuable resource to push forward the exploration of real-world, complex reasoning tasks as the next research focus. Our dataset and code is publicly available at https://github.com/czyssrs/ConvFinQA.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new large-scale dataset, ConvFinQA, aiming to study the chain of numerical reasoning in conversational question answering, and conducts comprehensive experiments and analyses with both the neural symbolic methods and the prompting-based methods to provide insights into the reasoning mechanisms.'}",http://arxiv.org/pdf/2210.03849
-Patterns of Instructional-design Factors Prompting Reflective Thinking in Middle-school and College Level Problem-based Learning Environments,Hae-Deok Song,,,
-Intelligent rule-based Phishing Websites Classification,B. Yogesh,"Phishing is depicted as the specialty of reverberating a site of a noteworthy firm meaning to snatch client's private data, for example, usernames, passwords and standardized savings number. Phishing sites involve an assortment of signs inside its substance parts and the program based security pointers furnished alongside the site. A few arrangements have been proposed to handle phishing. All things considered, there is no single enchantment shot that can explain this risk drastically. One of the promising strategies that can be utilized in anticipating phishing assaults depends on information mining, especially the prompting of order principles since hostile to phishing arrangements intend to foresee the site class precisely and that precisely coordinates the information mining characterization procedure objectives. In this study, the creators shed light on the essential elements that recognize phishing sites from authentic ones and evaluate how great standard based information mining arrangement methods are in foreseeing phishing sites and which characterization procedure is turned out to be more solid. Phishing assault traditionally begins by sending an email that appears to originate from a legitimate undertaking to casualties requesting that they upgrade or affirm their own data by going to a connection inside the email. Despite the fact that phishers are presently utilizing a few methods in making phishing sites to trick and charm clients, they all utilization an arrangement of common elements to make phishing sites on the grounds that, without those components they lose the benefit of misdirection. This helps us to separate in the middle of fair and phishing sites taking into account the elements removed from the went to site. By and large, two methodologies are utilized in recognizing phishing sites. The first depends on boycotts, in which the asked for URL is contrasted and those in that rundown. The drawback of this methodology is that the boycott as a rule can't cover all phishing sites, subsequent to, inside seconds; another deceitful site is required to be propelled. The second approach is known as heuristic-based strategy, where a few elements are gathered from the site to arrange it as either phishing or authentic. Rather than the boycott technique, a heuristic-based arrangement can perceive crisply made phishing sites. Keywords- Website features, Phishing, Security, Data Mining, Rule based Classification.","{'model': 'tldr@v2.0.0', 'text': 'Light is shed on the essential elements that recognize phishing sites from authentic ones and how great standard based information mining arrangement methods are in foreseeingphishing sites and which characterization procedure is turned out to be more solid.'}",
-"Improving ICU-Based Palliative Care Delivery: A Multicenter, Multidisciplinary Survey of Critical Care Clinician Attitudes and Beliefs",N. Wysham,"Objective: Addressing the quality gap in ICU-based palliative care is limited by uncertainty about acceptable models of collaborative specialist and generalist care. Therefore, we characterized the attitudes of physicians and nurses about palliative care delivery in an ICU environment. Design: Mixed-methods study. Setting: Medical and surgical ICUs at three large academic hospitals. Participants: Three hundred three nurses, intensivists, and advanced practice providers. Measurements and Main Results: Clinicians completed written surveys that assessed attitudes about specialist palliative care presence and integration into the ICU setting, as well as acceptability of 23 published palliative care prompts (triggers) for specialist consultation. Most (n = 225; 75%) reported that palliative care consultation was underutilized. Prompting consideration of eligibility for specialist consultation by electronic health record searches for triggers was most preferred (n = 123; 41%); only 17 of them (6%) felt current processes were adequate. The most acceptable specialist triggers were metastatic malignancy, unrealistic goals of care, end of life decision making, and persistent organ failure. Advanced age, length of stay, and duration of life support were the least acceptable. Screening led by either specialists or ICU teams was equally preferred. Central themes derived from qualitative analysis of 65 written responses to open-ended items included concerns about the roles of physicians and nurses, implementation, and impact on ICU team–family relationships. Conclusions: Integration of palliative care specialists in the ICU is broadly acceptable and desired. However, the most commonly used current triggers for prompting specialist consultation were among the least well accepted, while more favorable triggers are difficult to abstract from electronic health record systems. There is also disagreement about the role of ICU nurses in palliative care delivery. These findings provide important guidance to the development of collaborative care models for the ICU setting.","{'model': 'tldr@v2.0.0', 'text': None}",https://europepmc.org/articles/pmc5344777?pdf=render
-The varied circumstances prompting requests for emergency contraception at school-based clinics.,A. Sidebottom,"BACKGROUND Little is known about the circumstances that prompt teenagers to request emergency contraception (EC). This evaluation was designed to refine the EC clinical protocol and improve pregnancy prevention efforts in high school-based clinics by analyzing information on EC use and subsequent contraception use of EC patients. METHODS Sites included 5 clinics located at mainstream, racially diverse, and urban high schools. Nursing staff documented all EC-related visits during the 2002-2003 school year on a standardized form. These forms and additional information were collected from chart reviews. RESULTS EC was requested in 113 instances involving 91 students and dispensed in all but 4 instances. The most frequently reported circumstance prompting an EC request was the use of no protection (37.2%), followed by a condom mishap (27.4%) and questionable protection from a hormonal method (23.9%). Anxiety despite adequate protection from a hormonal method (9.7%) or a condom (1.8%) accounted for the remaining cases. Although the clinical protocol included a follow-up appointment 2 weeks after receipt of EC, 39.5% of EC users failed to keep this appointment. The chart review revealed that hormonal contraception was eventually initiated following two thirds (68.0%) of the EC instances that involved either no protection or only a condom. CONCLUSIONS Because young EC seekers vary considerably in terms of sexual experience, contraceptive use consistency, and pregnancy risk, individualized risk assessment and counseling are essential. The EC visit may be an optimal time to encourage initiation or reinstitution of hormonal methods of birth control.","{'model': 'tldr@v2.0.0', 'text': 'The EC visit may be an optimal time to encourage initiation or reinstitution of hormonal methods of birth control and individualized risk assessment and counseling are essential.'}",
-Change Detection in Coral Reef Environment Using High-Resolution Images: Comparison of Object-Based and Pixel-Based Paradigms,Zhenjin Zhou,"Despite increases in the spatial resolution of satellite imagery prompting interest in object-based image analysis, few studies have used object-based methods for monitoring changes in coral reefs. This study proposes a high accuracy object-based change detection (OBCD) method intended for coral reef environment, which uses QuickBird and WorldView-2 images. The proposed methodological framework includes image fusion, multi-temporal image segmentation, image differencing, random forests models, and object-area-based accuracy assessment. For validation, we applied the method to images of four coral reef study sites in the South China Sea. We compared the proposed OBCD method with a conventional pixel-based change detection (PBCD) method by implementing both methods under the same conditions. The average overall accuracy of OBCD exceeded 90%, which was approximately 20% higher than PBCD. The OBCD method was free from salt-and-pepper effects and was less prone to images misregistration in terms of change detection accuracy and mapping results. The object-area-based accuracy assessment reached a higher overall accuracy and per-class accuracy than the object-number-based and pixel-number-based accuracy assessment.","{'model': 'tldr@v2.0.0', 'text': 'The proposed OBCD method was free from salt-and-pepper effects and was less prone to images misregistration in terms of change detection accuracy and mapping results, and reached a higher overall accuracy and per-class accuracy than the object- number-based and pixel-number-based accuracy assessment.'}",https://www.mdpi.com/2220-9964/7/11/441/pdf?version=1542019325
-Impact of an encounter-based prompting system on resident vaccine administration performance and immunization knowledge.,J. Shaw,"OBJECTIVES To evaluate an encounter-based immunization prompting system on resident performance in administering vaccines and knowledge of immunization guidelines. DESIGN/METHODS Prospective randomized, controlled trial. Subjects were first- and second-year pediatric residents in a hospital-based continuity clinic. The intervention group received manual prompts of immunizations due. Postclinic chart review compared immunizations due with those administered. Acceptable and unacceptable reasons for not administering vaccines were assigned. Resident knowledge was measured by a 70-item examination. RESULTS The intervention group had significantly less missed opportunities/vaccine administration errors (11.4% vs 21.6%). The most common reason for unacceptable errors in the intervention group: vaccine was given too early; in the control group: vaccine was postponed to next visit. Pre- and postintervention knowledge scores were similar: intervention group (75.5% vs 80.7%, control group; 76.5% vs 81.3%). CONCLUSION An immunization prompting system in a hospital-based pediatric resident continuity clinic reduced missed opportunities/vaccine administration errors without significantly impacting resident knowledge of immunization guidelines.immunization schedule, vaccination, immunization, prompting systems, resident education.","{'model': 'tldr@v2.0.0', 'text': 'An immunization prompting system in a hospital-based pediatric resident continuity clinic reduced missed opportunities/vaccine administration errors without significantly impacting resident knowledge of immunization guidelines.'}",
-An iPad™-based picture and video activity schedule increases community shopping skills of a young adult with autism spectrum disorder and intellectual disability,Elizabeth Burckley,"Abstract Objective: To evaluate the iPad 2™ with Book Creator™ software to provide visual cues and video prompting to teach shopping skills in the community to a young adult with an autism spectrum disorder and intellectual disability. Methods: A multiple probe across settings design was used to assess effects of the intervention on the participant’s independence with following a shopping list in a grocery store across three community locations. Results: Visual cues and video prompting substantially increased the participant’s shopping skills within two of the three community locations, skill increases maintained after the intervention was withdrawn, and shopping skills generalized to two untaught shopping items. Social validity surveys suggested that the participant’s parent and staff favorably viewed the goals, procedures, and outcomes of intervention. Conclusions: The iPad 2™ with Book Creator™ software may be an effective way to teach independent shopping skills in the community; additional replications are needed.","{'model': 'tldr@v2.0.0', 'text': 'The iPad 2™ with Book Creator™ software may be an effective way to teach independent shopping skills in the community; additional replications are needed.'}",
-Capnographic monitoring of propofol-based sedation during colonoscopy,M. Friedrich-Rust,"Background and study aims: Capnography enables the measurement of end-tidal CO2 and thereby the early detection of apnea, prompting immediate intervention to restore ventilation. Studies have shown that capnographic monitoring is associated with a reduction of hypoxemia during sedation for endoscopy and early detection of apnea during sedation for colonoscopy. The primary aim of this prospective randomized study was to evaluate whether capnographic monitoring without tracheal intubation reduces hypoxemia during propofol-based sedation in patients undergoing colonoscopy. Patients and methods: A total of 533 patients presenting for colonoscopy at two study sites were randomized to either standard monitoring (n = 266) or to standard monitoring with capnography (n = 267). The incidence of hypoxemia (SO2 < 90 %) and severe hypoxemia (SO2 < 85 %) were compared between the groups. Furthermore, risk factors for hypoxemia were evaluated, and sedation performed by anesthesiologists was compared with nurse-administered propofol sedation (NAPS) or endoscopist-directed sedation (EDS). Results: The incidence of hypoxemia was significantly lower in patients with capnography monitoring compared with those receiving standard monitoring (18 % vs. 32 %; P = 0.00091). Independent risk factors for hypoxemia were age (P = 0.00015), high body mass index (P = 0.0044), history of sleep apnea (P = 0.025), standard monitoring group (P = 0.000069), total dose of propofol (P = 0.031), and dose of ketamine (P < 0.000001). Patients receiving anesthesiologist-administered sedation developed hypoxemic events more often than those receiving NAPS or EDS. In patients with anesthesiologist-administered sedation, sedation was deeper, a combination of sedative medication (propofol, midazolam and/or ketamine) was administered significantly more often, and sedative doses were significantly higher compared with patients receiving NAPS or EDS. Conclusions: In patients undergoing colonoscopy during propofol-based sedation capnography monitoring with a simple and inexpensive device reduced the incidence of hypoxemia.","{'model': 'tldr@v2.0.0', 'text': 'In patients undergoing colonoscopy during propofol-based sedation capnography monitoring with a simple and inexpensive device reduced the incidence of hypoxemia, particularly in patients with anesthesiologist-administered sedation.'}",
-"Social media monitoring: aims, methods, and challenges for international companies",Boyang Zhang,"Purpose – The purpose of this paper is to clarify the aims, monitoring methods and challenges of social media monitoring from the perspective of international companies. Trends in the literature are also investigated. Design/methodology/approach – Based on a systematic literature review, 30 key articles from 2008 to 2012 were further analysed. Findings – International companies need real-time monitoring software, expertise and dynamic visualization to facilitate early detection and prognoses supporting strategy making. This is a costly affair, prompting questions about return on investment. A recent trend in the research literature concerns the development of models describing how issues spread in social media with the aim of facilitating prognoses. Research limitations/implications – The online databases used comprised refereed peer-reviewed scientific articles. Books were not included in the search process. Practical implications – Because information spreads fast in social media and affects internation...",,https://jyx.jyu.fi/bitstream/123456789/44532/1/zhangvos2014socialmediamonitoringinternatcompanies.pdf
-Using an iPad-based video modelling package to teach numeracy skills to a child with an autism spectrum disorder,Elizabeth Jowett,"Purpose: To evaluate the effectiveness of a video modelling package to teach a 5 year-old boy diagnosed with an autism spectrum disorder (ASD) basic numeracy skills. The treatment package consisted of iPad-based video modelling, gradual fading of video prompts, reinforcement, in vivo prompting and forward chaining. Methods: A single subject multiple baseline across numerals design with generalization and maintenance observational measures. Results: Clear gains were evident in the participant's ability to identify and write the Arabic numerals 1–7 and comprehend the quantity each numeral represents in association with the lagged intervention. Generalization and maintenance data demonstrated the robustness of the treatment effects. Conclusions: IPad-based video modelling, when used in a package, can be an effective technique for teaching numeracy skills to children with an ASD. Systematic replication of this study with different participants is warranted.","{'model': 'tldr@v2.0.0', 'text': 'IPad-based video modelling, when used in a package, can be an effective technique for teaching numeracy skills to children with an ASD.'}",
-"Whole Organism High-Content Screening by Label-Free, Image-Based Bayesian Classification for Parasitic Diseases",R. Paveley,"Sole reliance on one drug, Praziquantel, for treatment and control of schistosomiasis raises concerns about development of widespread resistance, prompting renewed interest in the discovery of new anthelmintics. To discover new leads we designed an automated label-free, high content-based, high throughput screen (HTS) to assess drug-induced effects on in vitro cultured larvae (schistosomula) using bright-field imaging. Automatic image analysis and Bayesian prediction models define morphological damage, hit/non-hit prediction and larval phenotype characterization. Motility was also assessed from time-lapse images. In screening a 10,041 compound library the HTS correctly detected 99.8% of the hits scored visually. A proportion of these larval hits were also active in an adult worm ex-vivo screen and are the subject of ongoing studies. The method allows, for the first time, screening of large compound collections against schistosomes and the methods are adaptable to other whole organism and cell-based screening by morphology and motility phenotyping.","{'model': 'tldr@v2.0.0', 'text': 'An automated label-free, high content-based, high throughput screen (HTS) to assess drug-induced effects on in vitro cultured larvae (schistosomula) using bright-field imaging and the methods are adaptable to other whole organism and cell-based screening by morphology and motility phenotyping.'}",https://journals.plos.org/plosntds/article/file?id=10.1371/journal.pntd.0001762&type=printable
-A Review on Electromyography Decoding and Pattern Recognition for Human-Machine Interaction,M. Simão,"This paper presents a literature review on pattern recognition of electromyography (EMG) signals and its applications. The EMG technology is introduced and the most relevant aspects for the design of an EMG-based system are highlighted, including signal acquisition and filtering. EMG-based systems have been used with relative success to control upper- and lower-limb prostheses, electronic devices and machines, and for monitoring human behavior. Nevertheless, the existing systems are still inadequate and are often abandoned by their users, prompting for further research. Besides controlling prostheses, EMG technology is also beneficial for the development of machine learning-based devices that can capture the intention of able-bodied users by detecting their gestures, opening the way for new human-machine interaction (HMI) modalities. This paper also reviews the current feature extraction techniques, including signal processing and data dimensionality reduction. Novel classification methods and approaches for detecting non-trained gestures are discussed. Finally, current applications are reviewed, through the comparison of different EMG systems and discussion of their advantages and drawbacks.","{'model': 'tldr@v2.0.0', 'text': 'The EMG technology is introduced and the most relevant aspects for the design of an EMG-based system are highlighted, including signal acquisition and filtering, and the current feature extraction techniques, including Signal processing and data dimensionality reduction are reviewed.'}",https://ieeexplore.ieee.org/ielx7/6287639/8600701/08672131.pdf
-Evaluation of the Accelerate Pheno System for Fast Identification and Antimicrobial Susceptibility Testing from Positive Blood Cultures in Bloodstream Infections Caused by Gram-Negative Pathogens,M. Marschal,"ABSTRACT Bloodstream infections (BSI) are an important cause of morbidity and mortality. Increasing rates of antimicrobial-resistant pathogens limit treatment options, prompting an empirical use of broad-range antibiotics. Fast and reliable diagnostic tools are needed to provide adequate therapy in a timely manner and to enable a de-escalation of treatment. The Accelerate Pheno system (Accelerate Diagnostics, USA) is a fully automated test system that performs both identification and antimicrobial susceptibility testing (AST) directly from positive blood cultures within approximately 7 h. In total, 115 episodes of BSI with Gram-negative bacteria were included in our study and compared to conventional culture-based methods. The Accelerate Pheno system correctly identified 88.7% (102 of 115) of all BSI episodes and 97.1% (102 of 105) of isolates that are covered by the system's identification panel. The Accelerate Pheno system generated an AST result for 91.3% (95 of 104) samples in which the Accelerate Pheno system identified a Gram-negative pathogen. The overall category agreement between the Accelerate Pheno system and culture-based AST was 96.4%, the rates for minor discrepancies 1.4%, major discrepancies 2.3%, and very major discrepancies 1.0%. Of note, ceftriaxone, piperacillin-tazobactam, and carbapenem resistance was correctly detected in blood culture specimens with extended-spectrum beta-lactamase-producing Escherichia coli (n = 7) and multidrug-resistant Pseudomonas aeruginosa (n = 3) strains. The utilization of the Accelerate Pheno system reduced the time to result for identification by 27.49 h (P < 0.0001) and for AST by 40.39 h (P < 0.0001) compared to culture-based methods in our laboratory setting. In conclusion, the Accelerate Pheno system provided fast, reliable results while significantly improving turnaround time in blood culture diagnostics of Gram-negative BSI.","{'model': 'tldr@v2.0.0', 'text': 'The Accelerate Pheno system provided fast, reliable results while significantly improving turnaround time in blood culture diagnostics of Gram-negative BSI and compared to conventional culture-based methods in the laboratory setting.'}",https://jcm.asm.org/content/jcm/55/7/2116.full.pdf
-Unrecognized physical illness prompting psychiatric admission.,R. Reeves,"BACKGROUND We assessed factors that might contribute to clinicians erroneously attributing medically based changes in a patient's mental status to psychiatric illness. METHODS Records of 1340 patients admitted to a VA hospital psychiatric unit and 613 to a public hospital psychiatric unit from 2001 to 2007 were reviewed. Cases admitted because of an unrecognized medical disorder underwent further analysis of the preadmission assessment and documented history of mental illness. RESULTS Of 1340 patients whose records were reviewed, 55 (2.8%) had a medical disorder that caused their symptoms. Compared with patients admitted to medical units, patients inappropriately admitted to psychiatric units had lower rates of completion of medical histories, physical examinations, cognitive assessments, indicated laboratory and/or radiologic studies, and treatment of abnormal vital signs (P < .001 in each case). Among patients admitted to psychiatric units, 85.5% had a history of mental illness vs 30.9% of comparable admissions to medical units (Chi2(1) = 35.85; P < .001). CONCLUSIONS Key assessment procedures are less likely to be performed in patients with mental status changes who are admitted to psychiatric units than in comparable patients admitted to medical units. Symptoms of patients with a history of mental illness are more likely to be attributed to psychiatric illness than are those of patients without such a history.","{'model': 'tldr@v2.0.0', 'text': 'Key assessment procedures are less likely to be performed in patients with mental status changes who are admitted to psychiatric units than in comparable patients admitted to medical units.'}",
-Field-based development of an information support robot for persons with dementia,Takenobu Inoue,"Background: The use of robots for providing support to persons with dementia is very promising. However, it is difficult to develop a truly useful robot system because of the difficulty of clarifying their needs and opinions. Objective: Because a field-based method is effective for the development of these kinds of systems, in this study, we developed a prototype of an information support robot for persons with dementia, using field-based methodology. Methods: A communication robot system produced by NEC Corporation, called ""PaPeRo"", was chosen as the platform for this system. To achieve the goal of keeping persons with dementia informed of their daily schedule and prompting them to take desired actions, interactive verbal communication algorithms were programmed into the robot system. Results: The results of experiments with five dementia sufferers showed that this system produced an information acquisition rate of over 90%. In addition, a life support demonstration showed the possibility of prompting users to perform actions. Conclusions: These results suggest the possibility of using this robot to support independent living by persons with dementia, and also the effectiveness of the field-based method.","{'model': 'tldr@v2.0.0', 'text': 'A prototype of an information support robot for persons with dementia is developed, using field-based methodology, and the results suggest the possibility of using this robot to support independent living by people with dementia, and also the effectiveness of the field- based method.'}",https://content.iospress.com:443/download/technology-and-disability/tad00357?id=technology-and-disability%2Ftad00357
-Accuracy of 25-hydroxyvitamin D assays: confronting the issues.,G. Carter,"Measurement of 25-hydroxyvitamin D (25-OHD) is widely used for assessing vitamin D status. There has been a dramatic increase in 25-OHD requests over recent years prompting many laboratories to consider the use of automated immunoassays. To achieve higher throughput, these methods have abandoned the traditional solvent extraction of samples and are therefore more prone to non-specific interference. The Vitamin D External Quality Assessment Scheme (DEQAS) has revealed method-related differences in 25-OHD results, raising concerns about the comparability and accuracy of different assays. This paper highlights some of the pre-analytical, analytical and post-analytical issues which may influence the accuracy of 25-OHD assays and interpretation of results. Recent attention has focused on reconciling the relatively high results given by liquid chromatography-tandem mass spectrometry (LC-MS/MS) to those of the DiaSorin radioimmunoassay (RIA) on which clinical decision points have previously been based. Data is presented on 20 DEQAS samples which were analysed by an LC-MS/MS assay developed as a candidate reference measurement procedure by the US National Institute of Standards and Technology (NIST). The NIST results were on average 11.2% lower than those given by routine LC-MS/MS methods. If confirmed, these results suggest that most routine LC-MS/MS assays are perhaps overestimating 25-OHD by failing to resolve a molecule having the same mass as 25-OHD(3) and a similar fragmentation pattern. All 25-OHD assays should be monitored by a proficiency testing scheme and the results made available to clinicians and editors of scientific journals.","{'model': 'tldr@v2.0.0', 'text': 'If confirmed, these results suggest that most routine LC-MS/MS assays are perhaps overestimating 25-OHD by failing to resolve a molecule having the same mass as 25- OHD(3) and a similar fragmentation pattern.'}",
-"Active, Population-Based Surveillance for Severe Rotavirus Gastroenteritis in Children in the United States",D. Payne,"OBJECTIVES. Routine vaccination of US infants against rotavirus was implemented in 2006, prompting the Centers for Disease Control and Prevention New Vaccine Surveillance Network to begin population-based acute gastroenteritis surveillance among US children <3 years of age. This surveillance system establishes baseline estimates of rotavirus disease burden and allows for the prospective monitoring of rotavirus vaccination impact. METHODS. Eligible children with acute gastroenteritis (≥3 episodes of diarrhea and/or any vomiting in a 24-hour period) who were hospitalized, were seen in emergency departments, or visited selected outpatient clinics in 3 US counties during the period of January through June 2006 were enrolled. Epidemiological and clinical information was obtained through parental interview and medical chart review, and stool specimens were tested for rotavirus with enzyme immunoassays. Rotavirus-positive specimens were genotyped by using reverse transcription-polymerase chain reaction assays. RESULTS. Stool specimens were collected from 516 of the 739 enrolled children with acute gastroenteritis (181 inpatient, 201 emergency department, and 134 outpatient) and 44% tested positive for rotavirus (227 of 516 specimens). The most common strain was P[8]G1 (84%), followed by P[4]G2 (5%) and P[6]G12 (4%). None of the 516 children had received rotavirus vaccine. The rotavirus detection rate was 50% for hospitalized acute gastroenteritis cases, 50% for emergency department visits, and 27% for outpatient visits. Rotavirus-related acute gastroenteritis cases were more likely than non–rotavirus-related acute gastroenteritis cases to present with vomiting, diarrhea, fever, and lethargy. Directly calculated, population-based rates for rotavirus hospitalizations and emergency department visits were 22.5 hospitalizations and 301.0 emergency department visits per 10 000 children <3 years of age, respectively. A sentinel outpatient clinic visit rate of 311.9 outpatient visits per 10 000 children <3 years of age was observed. CONCLUSIONS. Population-based, laboratory-confirmed rotavirus surveillance in the final rotavirus season before implementation of the US rotavirus vaccine program indicated a considerable burden of disease on the US health care system.","{'model': 'tldr@v2.0.0', 'text': None}",
-Wavelet Methods for Spike Detection in Mouse Renal Sympathetic Nerve Activity,R. Brychta,"Abnormal autonomic nerve traffic has been associated with a number of peripheral neuropathies and cardiovascular disorders prompting the development of genetically altered mice to study the genetic and molecular components of these diseases. Autonomic function in mice can be assessed by directly recording sympathetic nerve activity. However, murine sympathetic spikes are typically detected using a manually adjusted voltage threshold and no unsupervised detection methods have been developed for the mouse. Therefore, we tested the performance of several unsupervised spike detection algorithms on simulated murine renal sympathetic nerve recordings, including an automated amplitude discriminator and wavelet-based detection methods which used both the discrete wavelet transform (DWT) and the stationary wavelet transform (SWT) and several wavelet threshold rules. The parameters of the wavelet methods were optimized by comparing basal sympathetic activity to postmortem recordings and recordings made during pharmacological suppression and enhancement of sympathetic activity. In general, SWT methods were found to outperform amplitude discriminators and DWT methods with similar wavelet coefficient thresholding algorithms when presented with simulations with varied mean spike rates and signal-to-noise ratios. A SWT method which estimates the noise level using a ""noise-only"" wavelet scale and then selectively thresholds scales containing the physiologically important signal information was found to have the most robust spike detection. The proposed noise-level estimation method was also successfully validated during pharmacological interventions","{'model': 'tldr@v2.0.0', 'text': 'ASWT methods were found to outperform amplitude discriminators and DWT methods with similar wavelet coefficient thresholding algorithms when presented with simulations with varied mean spike rates and signal-to-noise ratios and the proposed noise-level estimation method was successfully validated during pharmacological interventions.'}",https://europepmc.org/articles/pmc2075098?pdf=render
-Effect of Human Leukocyte Antigen Compatibility on Kidney Graft Survival: Comparative Analysis of Two Decades,G. Opelz,"Background. Based on an analysis of United Network for Organ Sharing data, it was reported that the influence of human leukocyte antigen (HLA) matching in renal transplantation has diminished in recent years, prompting the suggestion that donor kidney allocation algorithms should be revised. Methods. We compared the impact of HLA matching on kidney graft survival during the decades 1985–1994 and 1995–2004 using the data of the Collaborative Transplant Study. Results for the last 5 years (2000–2004) were analyzed separately in addition. Multivariate Cox regression analysis was used to account for the influence of confounders. Results. Our results show that, while graft survival rates have improved overall over time, the relative impact of HLA matching on the graft survival rate has remained strong and highly significant. Both the need for posttransplant rejection treatment and the graft survival rates showed statistically highly significant associations with HLA matching regardless of the interval analyzed (P<0.001). Conclusions. We conclude that HLA mismatches significantly influence the outcome of kidney transplants and that kidney exchange programs for the purpose of achieving better HLA matches continue to be meaningful.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that HLA mismatches significantly influence the outcome of kidney transplants and that kidney exchange programs for the purpose of achieving better HLA matches continue to be meaningful.'}",
-"Enhancing Instruction through Constructivism, Cooperative Learning, and Cloud Computing",David W. Denton,,"{'model': 'tldr@v2.0.0', 'text': 'Strategies for integrating cloud-based applications are suggested and results from a case study involving graduate education students are presented.'}",
-Update on Negative-Pressure Wound Therapy,D. Orgill,"Background: Over the last 15 years, negative-pressure wound therapy has become commonly used for treatment of a wide variety of complex wounds. There are now several systems marketed, and additional products will be available in the near future. Many clinicians have noted a dramatic response when negative-pressure wound therapy technology has been used, prompting a number of scientific investigations related to its mechanism of action and clinical trials determining its efficacy. Methods: The peer-reviewed literature within the past 5 years was reviewed, using an evidence-based approach. Results: Negative-pressure wound therapy works through mechanisms that include fluid removal, drawing the wound together, microdeformation, and moist wound healing. Several randomized clinical trials support the use of negative-pressure wound therapy in certain wound types. Serious complications, including bleeding and infection, have recently been reported by the U.S. Food and Drug Administration in a small number of patients. Conclusions: Negative-pressure wound therapy has dramatically changed the way complex wounds are treated. The rapid introduction of this technology has occurred faster than large-scale randomized controlled studies or registry studies have been conducted. Further clinical studies and basic science studies will help surgeons to better understand the evidence and use this technology in the future.","{'model': 'tldr@v2.0.0', 'text': 'The rapid introduction of this technology has occurred faster than large-scale randomized controlled studies or registry studies have been conducted and will help surgeons to better understand the evidence and use this technology in the future.'}",
-Prompting Clinicians: A Systematic Review of Preventive Care Reminders,J. Dexheimer,"Prompting clinicians to offer preventive care procedures has been shown to increase the use of these procedures. This study is an update of a systematic review examining the effect of reminder systems on offers of preventive care to patients. Of 1,404 eligible studies, 23 were included. The studies were evaluated according to their intervention type and use of computerized methods. We found that although computerized reminder systems have become more common, paper-based reminders were the most effective reminder strategy.","{'model': 'tldr@v2.0.0', 'text': 'It was found that although computerized reminder systems have become more common, paper-based reminders were the most effective reminder strategy.'}",
-Energy supply chain optimization of hybrid feedstock processes: a review.,J. Elia,"The economic, environmental, and social performances of energy systems depend on their geographical locations and the surrounding market infrastructure for feedstocks and energy products. Strategic decisions to locate energy conversion facilities must take all upstream and downstream operations into account, prompting the development of supply chain modeling and optimization methods. This article reviews the contributions of energy supply chain studies that include heat, power, and liquid fuels production. Studies are categorized based on specific features of the mathematical model, highlighting those that address energy supply chain models with and without considerations of multiperiod decisions. Studies that incorporate uncertainties are discussed, and opportunities for future research developments are outlined.","{'model': 'tldr@v2.0.0', 'text': 'The contributions of energy supply chain studies that include heat, power, and liquid fuels production are reviewed, highlighting those that address energy supply network models with and without considerations of multiperiod decisions.'}",https://www.annualreviews.org/doi/pdf/10.1146/annurev-chembioeng-060713-040425
-"Repetitive Daily Point of Choice Prompts and Occupational Sit-Stand Transfers, Concentration and Neuromuscular Performance in Office Workers: An RCT",L. Donath,"Objective: Prolonged office sitting time adversely affects neuromuscular and cardiovascular health parameters. As a consequence, the present study investigated the effects of prompting the use of height-adjustable working desk (HAWD) on occupational sitting and standing time, neuromuscular outcomes and concentration in office workers. Methods: A single-blinded randomized controlled trial (RCT) with parallel group design was conducted. Thirty-eight office workers were supplied with HAWDs and randomly assigned (Strata: physical activity (PA), BMI, gender, workload) to a prompt (INT) or non-prompt (CON) group. INT received three daily screen-based prompts within 12 weeks. CON was only instructed once concerning the benefits of using HAWDs prior to the start of the study. Sitting and standing times were objectively assessed as primary outcomes for one entire working week using the ActiGraph wGT3X-BT at baseline (pre), after 6 (mid) and 12 weeks (post). Concentration (d2-test), postural sway during upright stance (under single, dual and triple task) and lower limb strength endurance (heel-rise) were collected as secondary outcomes. Results: With large but not statistically significant within group effects from pre to post, INT increased weekly standing time at work by 9% (p = 0.22, d = 0.8) representing an increase from 7.2 h (4.8) to 9.7 (6.6) h (p = 0.07). Concentration and neuromuscular performance did not change from pre to post testing (0.23 < p < 0.95; 0.001 < ηp² < 0.05). Conclusion: Low-frequent and low cost screen-based point of choice prompts (3 per day within 12 weeks) already result in notable increases of occupational standing time of approx. daily 30 min. These stimuli, however, did not relevantly affect neuromuscular outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Low-frequent and low cost screen-based point of choice prompts (3 per day within 12 weeks) already result in notable increases of occupational standing time of approx.'}",https://www.mdpi.com/1660-4601/12/4/4340/pdf?version=1429534409
-Training skills in the psychiatrically disabled: learning coping and competence.,R. Liberman,"Social skills training methods represent a major strategy for psychiatric rehabilitation. Building skills in patients with schizophrenic and other major mental disorders is based on the assumption that coping and competence can override stress and vulnerability in reducing relapses and improving psychosocial functioning. For maximum efficiency, skills training needs to incorporate procedures and principles of human learning and information processing. Several models for skills training have been designed and evaluated, each of which has proved to be effective in raising the social competence of chronic mental patients. The ""basic"" model involves role playing by the patient and modeling, prompting, feedback, and reinforcement by the therapist. A ""problem-solving"" model of training provides general strategies for dealing with a wide variety of social situations. This model uses role playing to enhance behavioral performance but also highlights the patient's abilities to perceive and process incoming social messages and meanings. It is essential that social skills training be imbedded in a comprehensive program of rehabilitation that features continuity of care, supportive community services, therapeutic relationships, and judicious prescription of psychotropic drugs.","{'model': 'tldr@v2.0.0', 'text': 'It is essential that social skills training be imbedded in a comprehensive program of rehabilitation that features continuity of care, supportive community services, therapeutic relationships, and judicious prescription of psychotropic drugs.'}",https://academic.oup.com/schizophreniabulletin/article-pdf/12/4/631/5302085/12-4-631.pdf
-Adaptive Algebraic Multigrid,M. Brezina,"Efficient numerical simulation of physical processes is constrained by our ability to solve the resulting linear systems, prompting substantial research into the development of multiscale iterative methods capable of solving these linear systems with an optimal amount of effort. Overcoming the limitations of geometric multigrid methods to simple geometries and differential equations, algebraic multigrid methods construct the multigrid hierarchy based only on the given matrix. While this allows for efficient black-box solution of the linear systems associated with discretizations of many elliptic differential equations, it also results in a lack of robustness due to unsatisfied assumptions made on the near null spaces of these matrices. This paper introduces an extension to algebraic multigrid methods that removes the need to make such assumptions by utilizing an adaptive process. Emphasis is on the principles that guide the adaptivity and their application to algebraic multigrid solution of certain symmetric positive-definite linear systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an extension to algebraic multigrid methods that removes the need to make unsatisfied assumptions made on the near null spaces of these matrices by utilizing an adaptive process.'}",
-Current and emerging commercial optical biosensors,Cheryl L. Baird,"The field of commercial optical biosensors is rapidly evolving, with new systems and detection methods being developed each year. This review outlines the currently available biosensor hardware and highlights unique features of each platform. Affinity‐based biosensor technology, with its high sensitivity, wide versatility and high throughput, is playing a significant role in basic research, pharmaceutical development, and the food and environmental sciences. Likewise, the increasing popularity of biosensors is prompting manufacturers to develop new instrumentation for dedicated applications. We provide a preview of some of the emerging commercial systems that are dedicated to drug discovery, proteomics, clinical diagnostics and routine biomolecular interaction analysis. Copyright © 2001 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'This review outlines the currently available biosensor hardware and highlights unique features of each platform, providing a preview of some of the emerging commercial systems that are dedicated to drug discovery, proteomics, clinical diagnostics and routine biomolecular interaction analysis.'}",
-Improving training methods in brain injury rehabilitation,M. Mozzoni,"The effects of therapist training techniques upon treatment outcomes in traumatic brain injury (TBI) has not been widely studied. This study demonstrates that outcome scores are likely to increase when therapists work on outcome-related activities and use identified training methods. Six therapists working with five persons with TBI were subjects for this study. Therapists were targeted because their clients were not making progress on the functional independent measurement (FIM). Baseline teaching data were collected from videotapes, targeting 14 training elements within the areas of task, prompting, reinforcement, and client's behavior. Therapists were given feedback based upon the teaching diagnostic of the decision matrix. The results show that diagnostic-based intervention with therapists can result in direct, immediate, and obvious improvements in outcome measures across patients.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that outcome scores are likely to increase when therapists work on outcome-related activities and use identified training methods and diagnostic-based intervention with therapists can result in direct, immediate, and obvious improvements in outcome measures across patients.'}",
-Review of Studies with Students with Significant Cognitive Disabilities Which Link to Science Standards,Ginevra Courtade,"A comprehensive review of research was conducted on teaching science to students with significant cognitive disabilities. Guidelines from the National Science Education Standards were used to identify categories of studies based on the strands of science. A total of 11 studies were identified from the 20 years of literature searched. In general, work in the area of science is sparse for students with significant cognitive disabilities. Eight of the 11 studies fell in Content Standard F: Science in Personal and Social Perspectives. Single subject experimental designs were the methodology used in all of the studies and outcomes suggest that this population benefits from instruction in highly specific skills with modeling and errorless learning strategies like time delay. The interventions found in the 11 studies that linked to science incorporated systematic response prompting methods similar to those found in evidence-based reading and math research for students with significant cognitive disabilities. Additional research is needed to identify methods to teach science to this population that includes both broader content and additional skills that link to state standards.",,
-Is the Mussel Test a good Indicator of Antifouling Activity? A Comparison Between Laboratory and Field Assays,B. D. da Gama,"Current antifouling technologies rely on metal-based paints, but due to their toxicity, an expected worldwide ban of organotin-containing paints is now prompting the quest for safe and effective alternatives. One of these is antifouling coatings whose active components are naturally occurring compounds in marine organisms. A number of laboratory bioassays has been designed to search for antifouling compounds. However, there is no evidence to date that these assays provide results reproducible through ecologically realistic field experiments. Natural concentrations of the extracts from the Brazilian seaweeds Laurencia obtusa and Stypopodium zonale were tested in the laboratory through the 'mussel test' and in the field through the 'phytagel method' in order to compare the efficiency of these methods in assessing antifouling activity. L. obtusa extract significantly inhibited fouling in both the laboratory and field assays, while S. zonale stimulated fouling in both assays. Major compounds from the extracts were identified. The findings suggest that the 'mussel test' is a reliable time and cost-saving screening method for antifouling substances, although field assays are more sensitive for detection of their activity spectrum.","{'model': 'tldr@v2.0.0', 'text': ""The findings suggest that the 'mussel test' is a reliable time and cost-saving screening method for antifouling substances, although field assays are more sensitive for detection of their activity spectrum.""}",
-The use of celebrity athletes as endorsers: views of the New Zealand general public,J. Charbonneau,"Matching celebrity athletes with potential endorsement opportunities is often difficult. Yet there are easy-touse survey-based methods available. Based on a survey of the general public in New Zealand, this study uses both Ohanian's source-credibility scale and a constant-sum scale to help brand managers, player agents and advertising practitioners select good celebrity athlete-product fit. Four New Zealand athletes (two males and two females) and several products were included in the test. Results show that the female celebrity athletes outperformed their male counterparts as potential endorsers. Use of Ohanian's multi-attribute scale yields a level of richness and insight, prompting us to advocate the use of both scales in the pursuit of endorser-product congruences.",,https://researchcommons.waikato.ac.nz/bitstream/10289/5124/2/Garland%20The%20use%20of.pdf
-Expected value prioritization of prompts and reminders,S. Downs,"Computer-based prompting and reminder systems have been shown to be highly effective in increasing rates of preventive services delivery. However, there are many more recommended preventive services than can be practically included in a typical clinic visit. Therefore prioritization of preventive services prompts is necessary. We describe two approaches to prioritizing preventive services prompts based on expected value decision making. One method involves a static, global prioritization across all preventive services and has been used in a production system for almost 7 years. The second method uses influence diagrams to prioritize prompts dynamically, based on individual patient data. The latter approach is still under development. Both methods are labor intensive and require a combination of epidemiologic data and expert judgment. Compromises in strictly normative process were necessary to achieve user satisfaction.","{'model': 'tldr@v2.0.0', 'text': 'Two approaches to prioritizing preventive services prompts based on expected value decision making are described, one of which involves a static, global prioritization across all preventive services and has been used in a production system for almost 7 years, and the second which uses influence diagrams to prioritize prompts dynamically, based on individual patient data.'}",
-Exploring Recall of Physical Activity in Young People using Qualitative Interviewing,J. McKenna,"Recall of physical activity is a known problem affecting all forms of self-report. Participants age 8?16 years contributed to 16 focus groups and 24 interviews based on cognitive interviewing (n = 8) and think-aloud (n = 8) and general probing (n = 8) techniques. When unassisted, participants readily described physical activity mode but gave vague descriptions of daily activities. In contrast, the close detail of frequency, intensity, and duration of these activities was only more fully developed through prompting. Talk-based methods can provide considerable insight into developing more reliable and valid physical activity self-reports","{'model': 'tldr@v2.0.0', 'text': 'Talk-based methods can provide considerable insight into developing more reliable and valid physical activity self-reports.'}",
-A Prompting-based Approach for Adversarial Example Generation and Robustness Enhancement,Yuting Yang,"Recent years have seen the wide application of NLP models in crucial areas such as finance, medical treatment, and news media, raising concerns of the model robustness and vulnerabilities. In this paper, we propose a novel prompt-based adversarial attack to compromise NLP models and robustness enhancement technique. We first construct malicious prompts for each instance and generate adversarial examples via mask-and-filling under the effect of a malicious purpose. Our attack technique targets the inherent vulnerabilities of NLP models, allowing us to generate samples even without interacting with the victim NLP model, as long as it is based on pre-trained language models (PLMs). Furthermore, we design a prompt-based adversarial training method to improve the robustness of PLMs. As our training method does not actually generate adversarial samples, it can be applied to large-scale training sets efficiently. The experimental results show that our attack method can achieve a high attack success rate with more diverse, fluent and natural adversarial examples. In addition, our robustness enhancement method can significantly improve the robustness of models to resist adversarial attacks. Our work indicates that prompting paradigm has great potential in probing some fundamental flaws of PLMs and fine-tuning them for downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based adversarial attack to compromise NLP models and robustness enhancement technique that can significantly improve the robustness of models to resist adversarial attacks and indicates that prompting paradigm has great potential in probing some fundamental flaws of PLMs and fine-tuning them for downstream tasks.'}",http://arxiv.org/pdf/2203.10714
-Kemampuan Representasi Matematis Siswa Melalui Model Problem Based Learning dan Probing Prompting Learning,Suwanti Suwanti,"The students' mathematical representation ability is low. Efforts to improve mathematical representation skills include innovating learning models. The purpose of the study was to analyze the differences and quality of improvement in mathematical representation skills between students who received the problem-based learning model with probing prompting learning, along with student attitude responses. The research method is quasi-experimental. The population is all class VIII students of one of the private junior high schools in Garut for the 2018/2019 academic year, the samples are class VIII-A and VIII-B. The results showed that there were differences in the ability of mathematical representation between students who received the problem-based learning model with probing prompting learning, the quality of improving the mathematical representation ability of students in both classes had moderate interpretation, and students' attitudes had good interpretations. A problem-based learning model with probing prompting learning can be used in learning to improve mathematical representation abilities.",,https://journal.institutpendidikan.ac.id/index.php/plusminus/article/download/pv1n2_09/854
-Dictionary-based Phrase-level Prompting of Large Language Models for Machine Translation,Marjan Ghazvininejad,"Large language models (LLMs) demonstrate remarkable machine translation (MT) abilities via prompting, even though they were not explicitly trained for this task. However, even given the incredible quantities of data they are trained on, LLMs can struggle to translate inputs with rare words, which are common in low resource or domain transfer scenarios. We show that LLM prompting can provide an effective solution for rare words as well, by using prior knowledge from bilingual dictionaries to provide control hints in the prompts. We propose a novel method, DiPMT, that provides a set of possible translations for a subset of the input words, thereby enabling fine-grained phrase-level prompted control of the LLM. Extensive experiments show that DiPMT outperforms the baseline both in low-resource MT, as well as for out-of-domain MT. We further provide a qualitative analysis of the benefits and limitations of this approach, including the overall level of controllability that is achieved.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that LLM prompting can provide an effective solution for rare words as well, by using prior knowledge from bilingual dictionaries to provide control hints in the prompts, thereby enabling fine-grained phrase-level prompted control of the LLM.'}",http://arxiv.org/pdf/2302.07856
-Fine-Grained Controllable Text Generation Using Non-Residual Prompting,F. Carlsson,"The introduction of immensely large Causal Language Models (CLMs) has rejuvenated the interest in open-ended text generation. However, controlling the generative process for these Transformer-based models is at large an unsolved problem. Earlier work has explored either plug-and-play decoding strategies, or more powerful but blunt approaches such as prompting. There hence currently exists a trade-off between fine-grained control, and the capability for more expressive high-level instructions. To alleviate this trade-off, we propose an encoder-decoder architecture that enables intermediate text prompts at arbitrary time steps. We propose a resource-efficient method for converting a pre-trained CLM into this architecture, and demonstrate its potential on various experiments, including the novel task of contextualized word inclusion. Our method provides strong results on multiple experimental settings, proving itself to be both expressive and versatile.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an encoder-decoder architecture that enables intermediate text prompts at arbitrary time steps, and proposes a resource-efficient method for converting a pre-trained CLM into this architecture, and demonstrates its potential on various experiments, including the novel task of contextualized word inclusion.'}",https://aclanthology.org/2022.acl-long.471.pdf
-BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting,Zheng-Xin Yong,"The BLOOM model is a large publicly available multilingual language model, but its pretraining was limited to 46 languages. To extend the benefits of BLOOM to other languages without incurring prohibitively large costs, it is desirable to adapt BLOOM to new languages not seen during pretraining. In this work, we apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. We find language adaptation to be effective at improving zero-shot performance in new languages. Surprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models. In addition, we discover that prompting performance is not significantly affected by language specifics, such as the writing system. It is primarily determined by the size of the language adaptation data. We also add new languages to BLOOMZ, which is a multitask finetuned version of BLOOM capable of following task instructions zero-shot. We find including a new language in the multitask fine-tuning mixture to be the most effective method to teach BLOOMZ a new language. We conclude that with sufficient training data language adaptation can generalize well to diverse languages. Our code is available at https://github.com/bigscience-workshop/multilingual-modeling.","{'model': 'tldr@v2.0.0', 'text': 'This work applies existing language adaptation strategies to BLOOM and finds language adaptation to be effective at improving zero-shot performance in new languages and concludes that with sufficient training data language adaptation can generalize well to diverse languages.'}",http://arxiv.org/pdf/2212.09535
-Understanding and Improving Visual Prompting: A Label-Mapping Perspective,Aochuan Chen,"We revisit and advance visual prompting (VP), an input prompting technique for vision tasks. VP can reprogram a fixed, pre-trained source model to accomplish downstream tasks in the target domain by simply incorporating universal prompts (in terms of input perturbation patterns) into downstream data points. Yet, it remains elusive why VP stays effective even given a ruleless label mapping (LM) between the source classes and the target classes. Inspired by the above, we ask: How is LM interrelated with VP? And how to exploit such a relationship to improve its accuracy on target tasks? We peer into the influence of LM on VP and provide an affirmative answer that a better ‘quality’ of LM (assessed by mapping precision and explanation) can consistently improve the effectiveness of VP. This is in contrast to the prior art where the factor of LM was missing. To optimize LM, we propose a new VP framework, termed ILM-VP (iterative label mapping-based visual prompting), which automatically re-maps the source labels to the target labels and progressively improves the target task accuracy of VP. Further, when using a contrastive language-image pretrained (CLIP) model for VP, we propose to integrate an LM process to assist the text prompt selection of CLIP and to improve the target task accuracy. Extensive experiments demonstrate that our proposal significantly outperforms state-of-the-art VP methods. As highlighted below, we show that when reprogramming an ImageNet-pretrained ResNet-18 to 13 target tasks, ILM-VP outperforms baselines by a substantial margin, e.g., 7.9% and 6.7% accuracy improvements in transfer learning to the target Flowers102 and CIFAR100 datasets. Besides, our proposal on CLIP-based VP provides 13.7% and 7.1% accuracy improvements on Flowers102 and DTD respectively. Code is available at https://github.com/OPTML-Group/ILM-VP.","{'model': 'tldr@v2.0.0', 'text': 'A new VP framework, termed ILM-VP (iterative label mapping-based visual prompting), which automatically re-maps the source labels to the target labels and progressively improves the target task accuracy of VP is proposed.'}",https://arxiv.org/pdf/2211.11635
-Hybrid-SAR Technique: Joint Analysis Using Phase-Based and Amplitude-Based Methods for the Xishancun Giant Landslide Monitoring,T. Qu,"Early detection and early warning are of great importance in giant landslide monitoring because of the unexpectedness and concealed nature of large-scale landslides. In China, the western mountainous areas are prone to landslides and feature many giant complex landslides, especially following the Wenchuan Earthquake in 2008. This work concentrates on a new technique, known as the “hybrid-SAR technique”, that combines both phase-based and amplitude-based methods to detect and monitor large-scale landslides in Li County, Sichuan Province, southwestern China. This work aims to develop a robust methodological approach to promptly identify diverse landslides with different deformation magnitudes, sliding modes and slope geometries, even when the available satellite data are limited. The phase-based and amplitude-based techniques are used to obtain the landslide displacements from six TerraSAR-X Stripmap descending scenes acquired from November 2014 to March 2015. Furthermore, the application circumstances and influence factors of hybrid-SAR are evaluated according to four aspects: (1) quality of terrain visibility to the radar sensor; (2) landslide deformation magnitude and different sliding mode; (3) impact of dense vegetation cover; and (4) sliding direction sensitivity. The results achieved from hybrid-SAR are consistent with in situ measurements. This new hybrid-SAR technique for complex giant landslide research successfully identified representative movement areas, e.g., an extremely slow earthflow and a creeping region with a displacement rate of 1 cm per month and a typical rotational slide with a displacement rate of 2–3 cm per month downwards and towards the riverbank. Hybrid-SAR allows for a comprehensive and preliminary identification of areas with significant movement and provides reliable data support for the forecasting and monitoring of landslides.","{'model': 'tldr@v2.0.0', 'text': 'This work concentrates on a new technique, known as the “hybrid-SAR technique”, that combines both phase-based and amplitude-based methods to detect and monitor large-scale landslides in Li County, Sichuan Province, southwestern China, and successfully identified representative movement areas.'}",https://www.mdpi.com/2072-4292/8/10/874/pdf?version=1477292667
-Automatic Multi-Label Prompting: Simple and Interpretable Few-Shot Classification,Han Wang,"Prompt-based learning (i.e., prompting) is an emerging paradigm for exploiting knowledge learned by a pretrained language model. In this paper, we propose Automatic Multi-Label Prompting (AMuLaP), a simple yet effective method to automatically select label mappings for few-shot text classification with prompting. Our method exploits one-to-many label mappings and a statistics-based algorithm to select label mappings given a prompt template. Our experiments demonstrate that AMuLaP achieves competitive performance on the GLUE benchmark without human effort or external resources.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Automatic Multi-Label Prompting (AMuLaP), a simple yet effective method to automatically select label mappings for few-shot text classification with prompting that achieves competitive performance on the GLUE benchmark without human effort or external resources.'}",http://arxiv.org/pdf/2204.06305
-FS-DETR: Few-Shot DEtection TRansformer with prompting and without re-training,Adrian Bulat,"This paper is on Few-Shot Object Detection (FSOD), where given a few templates (examples) depicting a novel class (not seen during training), the goal is to detect all of its occurrences within a set of images. From a practical perspective, an FSOD system must fulfil the following desiderata: (a) it must be used as is, without requiring any fine-tuning at test time, (b) it must be able to process an arbitrary number of novel objects concurrently while supporting an arbitrary number of examples from each class and (c) it must achieve accuracy comparable to a closed system. Towards satisfying (a)-(c), in this work, we make the following contributions: We introduce, for the first time, a simple, yet powerful, few-shot detection transformer (FS-DETR) based on visual prompting that can address both desiderata (a) and (b). Our system builds upon the DETR framework, extending it based on two key ideas: (1) feed the provided visual templates of the novel classes as visual prompts during test time, and (2) ``stamp'' these prompts with pseudo-class embeddings (akin to soft prompting), which are then predicted at the output of the decoder. Importantly, we show that our system is not only more flexible than existing methods, but also, it makes a step towards satisfying desideratum (c). Specifically, it is significantly more accurate than all methods that do not require fine-tuning and even matches and outperforms the current state-of-the-art fine-tuning based methods on the most well-established benchmarks (PASCAL VOC&MSCOCO).","{'model': 'tldr@v2.0.0', 'text': 'This work introduces, for the first time, a simple, yet powerful, few-shot detection transformer (FS-DETR) based on visual prompting that can address both desiderata (a) and (b) and makes a step towards satisfying desideratum (c).'}",https://arxiv.org/pdf/2210.04845
-Prompting Contrastive Explanations for Commonsense Reasoning Tasks,Bhargavi Paranjape,"Many commonsense reasoning NLP tasks involve choosing between one or more possible answers to a question or prompt based on knowledge that is often implicit. Large pretrained language models (PLMs) can achieve near-human performance on such tasks, while providing little human-interpretable evidence of the underlying reasoning they use. In this work, we show how to use these same models to generate such evidence: inspired by the contrastive nature of human explanations, we use PLMs to complete explanation prompts which contrast alternatives according to the key attribute(s) required to justify the correct answer (for example, peanuts are usually salty while raisins are sweet). Conditioning model decisions on these explanations improves performance on two commonsense reasoning benchmarks, as compared to previous non-contrastive alternatives. These explanations are also judged by humans to be more relevant for solving the task, and facilitate a novel method to evaluate explanation faithfulfness.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the contrastive nature of human explanations, this work uses PLMs to complete explanation prompts which contrast alternatives according to the key attribute(s) required to justify the correct answer (for example, peanuts are usually salty while raisins are sweet).'}",https://aclanthology.org/2021.findings-acl.366.pdf
-Enhancing Cross-lingual Prompting with Mask Token Augmentation,Meng Zhou,"Prompting 1 shows promising results in few-shot scenarios. However, its strength for multilingual/cross-lingual problems has not been fully exploited. Zhao and Schütze (2021) made initial explorations in this direction by presenting that cross-lingual prompting outperforms cross-lingual finetuning. In this paper, we conduct empirical analysis on the effect of each component in cross-lingual prompting and derive Universal Prompting across languages, which helps alleviate the discrepancies between source-language training and target-language inference. Based on this, we propose a mask token augmentation framework to further improve the performance of prompt-based cross-lingual transfer. Notably, for XNLI, our method achieves 46.54% with only 16 English training examples per class, significantly better than 34.99% of finetuning.","{'model': 'tldr@v2.0.0', 'text': 'Empirical analysis on the effect of each component in cross-lingual prompting is conducted and Universal Prompting across languages is derived, which helps alleviate the discrepancies between source-language training and target-language inference.'}",
-A changing landscape: web-based methods for dietary assessment in adolescents,K. Storey,"Purpose of reviewAdolescents’ dietary intake is an important determinant of health and well-being and is influenced by a complex interaction of environmental, social, psychological, and physiological factors. The complexity of the adolescent diet makes its assessment prone to error, which has prompted researchers and clinicians to turn to technology to reduce this error. Previous reviews have been conducted regarding the use of technology in dietary assessment for adults; however, there are no known reviews for adolescents. Therefore, the purpose of this review is to describe the practical considerations for web-based dietary assessment methods and to evaluate recent evidence on their validity and implications. Recent findingsThere are numerous web-based dietary assessment methods that are available, valid, and reliable for use in the adolescent population. Web-based methods include both native and web-based applications (or ‘apps’), and have been developed for use as food records, 24-h dietary recalls, and food frequency questionnaires. SummaryWeb-based methods provide an efficient, cost-effective and practical solution to assess dietary intake; they are less burdensome to respondents and reduce errors and bias. Furthermore, adolescents are technologically savvy and often prefer the use of technology. Web-based methods should be considered when assessing adolescents’ dietary intake.","{'model': 'tldr@v2.0.0', 'text': 'Web-based methods provide an efficient, cost-effective and practical solution to assess dietary intake; they are less burdensome to respondents and reduce errors and bias.'}",
-Generated Knowledge Prompting for Commonsense Reasoning,Jiacheng Liu,"It remains an open question whether incorporating external knowledge benefits commonsense reasoning while maintaining the flexibility of pretrained sequence models. To investigate this question, we develop generated knowledge prompting, which consists of generating knowledge from a language model, then providing the knowledge as additional input when answering a question. Our method does not require task-specific supervision for knowledge integration, or access to a structured knowledge base, yet it improves performance of large-scale, state-of-the-art models on four commonsense reasoning tasks, achieving state-of-the-art results on numerical commonsense (NumerSense), general commonsense (CommonsenseQA 2.0), and scientific commonsense (QASC) benchmarks. Generated knowledge prompting highlights large-scale language models as flexible sources of external knowledge for improving commonsense reasoning.Our code is available at github.com/liujch1998/GKP","{'model': 'tldr@v2.0.0', 'text': 'Generated knowledge prompting develops generated knowledge prompting, which consists of generating knowledge from a language model, then providing the knowledge as additional input when answering a question, and improves performance of large-scale, state-of-the-art models on four commonsense reasoning tasks.'}",https://aclanthology.org/2022.acl-long.225.pdf
-A review of deep learning-based three-dimensional medical image registration methods.,Haonan Xiao,"Medical image registration is a vital component of many medical procedures, such as image-guided radiotherapy (IGRT), as it allows for more accurate dose-delivery and better management of side effects. Recently, the successful implementation of deep learning (DL) in various fields has prompted many research groups to apply DL to three-dimensional (3D) medical image registration. Several of these efforts have led to promising results. This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research. The collected studies were statistically analyzed based on the region of interest (ROI), image modality, supervision method, and registration evaluation metrics. The studies were classified into three categories: deep iterative registration, supervised registration, and unsupervised registration. The studies are thoroughly reviewed and their unique contributions are highlighted. A summary is presented following a review of each category of study, discussing its advantages, challenges, and trends. Finally, the common challenges for all categories are discussed, and potential future research topics are identified.","{'model': 'tldr@v2.0.0', 'text': 'This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research.'}",https://qims.amegroups.com/article/viewFile/75304/pdf
-DiffEdit: Diffusion-based semantic image editing with mask guidance,Guillaume Couairon,"Image generation has recently seen tremendous advances, with diffusion models allowing to synthesize convincing images for a large variety of text prompts. In this article, we propose DiffEdit, a method to take advantage of text-conditioned diffusion models for the task of semantic image editing, where the goal is to edit an image based on a text query. Semantic image editing is an extension of image generation, with the additional constraint that the generated image should be as similar as possible to a given input image. Current editing methods based on diffusion models usually require to provide a mask, making the task much easier by treating it as a conditional inpainting task. In contrast, our main contribution is able to automatically generate a mask highlighting regions of the input image that need to be edited, by contrasting predictions of a diffusion model conditioned on different text prompts. Moreover, we rely on latent inference to preserve content in those regions of interest and show excellent synergies with mask-based diffusion. DiffEdit achieves state-of-the-art editing performance on ImageNet. In addition, we evaluate semantic image editing in more challenging settings, using images from the COCO dataset as well as text-based generated images.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes DiffEdit, a method to take advantage of text-conditioned diffusion models for the task of semantic image editing, where the goal is to edit an image based on a text query.'}",http://arxiv.org/pdf/2210.11427
-Make-A-Scene: Scene-Based Text-to-Image Generation with Human Priors,Oran Gafni,"Recent text-to-image generation methods provide a simple yet exciting conversion capability between text and image domains. While these methods have incrementally improved the generated image fidelity and text relevancy, several pivotal gaps remain unanswered, limiting applicability and quality. We propose a novel text-to-image method that addresses these gaps by (i) enabling a simple control mechanism complementary to text in the form of a scene, (ii) introducing elements that substantially improve the tokenization process by employing domain-specific knowledge over key image regions (faces and salient objects), and (iii) adapting classifier-free guidance for the transformer use case. Our model achieves state-of-the-art FID and human evaluation results, unlocking the ability to generate high fidelity images in a resolution of 512x512 pixels, significantly improving visual quality. Through scene controllability, we introduce several new capabilities: (i) Scene editing, (ii) text editing with anchor scenes, (iii) overcoming out-of-distribution text prompts, and (iv) story illustration generation, as demonstrated in the story we wrote.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel text-to-image method that addresses gaps in applicability and quality by enabling a simple control mechanism complementary to text in the form of a scene, and introducing elements that substantially improve the tokenization process by employing domain-specific knowledge over key image regions.'}",http://arxiv.org/pdf/2203.13131
-Batch Prompting: Efficient Inference with Large Language Model APIs,Zhoujun Cheng,"Performing inference on large volumes of samples with large language models (LLMs) can be computationally and financially costly in industry and real-world use. We propose batch prompting, a simple yet effective prompting approach that enables the LLM to run inference in batches, instead of one sample at a time. Our method reduces both token and time costs while retaining downstream performance. We theoretically demonstrate that under a few-shot in-context learning setting, the inference costs decrease almost inverse linearly with the number of samples in each batch. We extensively validate the effectiveness of batch prompting on ten datasets across commonsense QA, arithmetic reasoning, and NLI/NLU: batch prompting significantly~(up to 5x with six samples in batch) reduces the LLM (Codex) inference token and time costs while achieving better or comparable performance. For state-of-the-art Chat-based LLMs, e.g., GPT-3.5 and GPT-4, we show the benefits of batch prompting also hold. Further analysis shows that the number of samples in each batch and the complexity of tasks affect its performance. Moreover, batch prompting can be applied across different reasoning methods using LLMs. Our code can be found at the site https://github.com/xlang-ai/batch-prompting.","{'model': 'tldr@v2.0.0', 'text': 'Batch prompting, a simple yet effective prompting approach that enables the LLM to run inference in batches, instead of one sample at a time, is proposed, which reduces both token and time costs while retaining downstream performance.'}",http://arxiv.org/pdf/2301.08721
-Prompting healthier eating: testing the use of health and social norm based messages.,E. Robinson,"OBJECTIVE Health based messages are commonly used to promote fruit and vegetable intake, but are limited in their effectiveness. Social norm messages, which suggest other people are eating healthily, may be more effective. Our aim was to compare the effect on food selection of a message containing health related information about fruit and vegetable consumption with a message containing social normative information about consumption of fruit and vegetables. METHOD In two laboratory studies, predominantly young female adult students were exposed to a health or social norm message about fruit and vegetables. In Study 1, lunch meal food selections and intake were assessed and in Study 2, snack food selections and intake were assessed. Study 1 examined the effect of a descriptive social norm (information about what others are eating) versus a health message and Study 2 examined the effect of both a descriptive norm and an injunctive norm message (information about what others approve of) versus a health message. RESULTS In Study 1, exposure to a descriptive social norm message resulted in significantly more vegetables being selected and eaten than exposure to a health message. In Study 2, exposure to a descriptive social norm message resulted in significantly more fruit and vegetables and less high energy dense snack food being selected and eaten than exposure to a health message. There was no effect of exposure to the injunctive norm message. In both studies, significant differences between the social norm and health message conditions were observed in low but not high usual consumers of fruit and vegetables. CONCLUSIONS For the promotion of healthy eating, social norm messages may be more effective than health messages for consumers failing to adhere to dietary guidelines.","{'model': 'tldr@v2.0.0', 'text': 'For the promotion of healthy eating, social norm messages may be more effective than health messages for consumers failing to adhere to dietary guidelines.'}",
-Dynamic Prefix-Tuning for Generative Template-based Event Extraction,Xiao Liu,"We consider event extraction in a generative manner with template-based conditional generation.Although there is a rising trend of casting the task of event extraction as a sequence generation problem with prompts, these generation-based methods have two significant challenges, including using suboptimal prompts and static event type information.In this paper, we propose a generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type-specific prefixes to learn a context-specific prefix for each context.Experimental results show that our model achieves competitive results with the state-of-the-art classification-based model OneIE on ACE 2005 and achieves the best performances on ERE.Additionally, our model is proven to be portable to new types of events effectively.","{'model': 'tldr@v2.0.0', 'text': 'A generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type- specific prefixes to learn a context-specific prefix for each context is proposed.'}",https://aclanthology.org/2022.acl-long.358.pdf
-COVID-19 Detection Based on Lung Ct Scan Using Deep Learning Techniques,S. V. Kogilavani,"SARS-CoV-2 is a novel virus, responsible for causing the COVID-19 pandemic that has emerged as a pandemic in recent years. Humans are becoming infected with the virus. In 2019, the city of Wuhan reported the first-ever incidence of COVID-19. COVID-19 infected people have symptoms that are related to pneumonia, and the virus affects the body's respiratory organs, making breathing difficult. A real-time reverse transcriptase-polymerase chain reaction (RT-PCR) kit is used to diagnose the disease. Due to a shortage of kits, suspected patients cannot be treated promptly, resulting in disease spread. To develop an alternative, radiologists looked at the changes in radiological imaging, like CT scans, that produce comprehensive pictures of the body of excellent quality. The suspected patient's computed tomography (CT) scan is used to distinguish between a healthy individual and a COVID-19 patient using deep learning algorithms. A lot of deep learning methods have been proposed for COVID-19. The proposed work utilizes CNN architectures like VGG16, DeseNet121, MobileNet, NASNet, Xception, and EfficientNet. The dataset contains 3873 total CT scan images with “COVID” and “Non-COVID.” The dataset is divided into train, test, and validation. Accuracies obtained for VGG16 are 97.68%, DenseNet121 is 97.53%, MobileNet is 96.38%, NASNet is 89.51%, Xception is 92.47%, and EfficientNet is 80.19%, respectively. From the obtained analysis, the results show that the VGG16 architecture gives better accuracy compared to other architectures.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the VGG16 architecture gives better accuracy compared to other architectures for COVID-19, a novel pandemic that has emerged as a pandemic in recent years.'}",https://downloads.hindawi.com/journals/cmmm/2022/7672196.pdf
-Fusion-Based Supply Chain Collaboration Using Machine Learning Techniques,Naeem Ali,"Supply Chain Collaboration is the network of various entities that work cohesively to make up the entire process. The supply chain organizations’ success is dependent on integration, teamwork, and the communication of information. Every day, supply chain and business players work in a dynamic setting. They must balance competing goals such as process robustness, risk reduction, vulnerability reduction, real financial risks, and resilience against just-in-time and cost-efficiency. Decision-making based on shared information in Supply Chain Collaboration constitutes the recital and competitiveness of the collective process. Supply Chain Collaboration has prompted companies to implement the perfect data analytics functions (e.g., data science, predictive analytics, and big data) to improve supply chain operations and, eventually, efficiency. Simulation and modeling are powerful methods for analyzing, investigating, examining, observing and evaluating real-world industrial and logistic processes in this scenario. Fusion-based Machine learning provides a platform that may address the issues/limitations of Supply Chain Collaboration. Compared to the Classical probable data fusion techniques, the fused Machine learning method may offer a strong computing ability and prediction. In this scenario, the machine learningbased Supply Chain Collaboration model has been proposed to evaluate the propensity of the decision-making process to increase the efficiency of the Supply Chain Collaboration.","{'model': 'tldr@v2.0.0', 'text': 'The machine learningbased Supply Chain Collaboration model has been proposed to evaluate the propensity of the decision-making process to increase the efficiency of the supply chain Collaboration.'}",https://www.techscience.com/iasc/v31n3/44835/pdf
-A kinect-based vocational task prompting system for individuals with cognitive impairments,Yao-Jen Chang,,"{'model': 'tldr@v2.0.0', 'text': 'It is suggested that the image recognition technology may be able to facilitate task prompts needed by people with cognitive impairments and may be helpful for pre-service training while increasing independence in the process of community integration.'}",
-SpeechUT: Bridging Speech and Text with Hidden-Unit for Encoder-Decoder Based Speech-Text Pre-training,Zi-Hua Zhang,"The rapid development of single-modal pre-training has prompted researchers to pay more attention to cross-modal pre-training methods. In this paper, we propose a unified-modal speech-unit-text pre-training model, SpeechUT, to connect the representations of a speech encoder and a text decoder with a shared unit encoder. Leveraging hidden-unit as an interface to align speech and text, we can decompose the speech-to-text model into a speech-to-unit model and a unit-to-text model, which can be jointly pre-trained with unpaired speech and text data respectively. Our proposed SpeechUT is fine-tuned and evaluated on automatic speech recognition (ASR) and speech translation (ST) tasks. Experimental results show that SpeechUT gets substantial improvements over strong baselines, and achieves state-of-the-art performance on both the LibriSpeech ASR and MuST-C ST tasks. To better understand the proposed SpeechUT, detailed analyses are conducted. The code and pre-trained models are available at https://aka.ms/SpeechUT.","{'model': 'tldr@v2.0.0', 'text': 'A unified-modal speech-unit-text pre-training model to connect the representations of a speech encoder and a text decoder with a shared unit encoder, and achieves state-of-the-art performance on both the LibriSpeech ASR and MuST-C ST tasks.'}",http://arxiv.org/pdf/2210.03730
-An Empirical Study of GPT-3 for Few-Shot Knowledge-Based VQA,Zhengyuan Yang,"Knowledge-based visual question answering (VQA) involves answering questions that require external knowledge not present in the image. Existing methods first retrieve knowledge from external resources, then reason over the selected knowledge, the input image, and question for answer prediction. However, this two-step approach could lead to mismatches that potentially limit the VQA performance. For example, the retrieved knowledge might be noisy and irrelevant to the question, and the re-embedded knowledge features during reasoning might deviate from their original meanings in the knowledge base (KB). To address this challenge, we propose PICa, a simple yet effective method that Prompts GPT3 via the use of Image Captions, for knowledge-based VQA. Inspired by GPT-3’s power in knowledge retrieval and question answering, instead of using structured KBs as in previous work, we treat GPT-3 as an implicit and unstructured KB that can jointly acquire and process relevant knowledge. Specifically, we first convert the image into captions (or tags) that GPT-3 can understand, then adapt GPT-3 to solve the VQA task in a few-shot manner by just providing a few in-context VQA examples. We further boost performance by carefully investigating: (i) what text formats best describe the image content, and (ii) how in-context examples can be better selected and used. PICa unlocks the first use of GPT-3 for multimodal tasks. By using only 16 examples, PICa surpasses the supervised state of the art by an absolute +8.6 points on the OK-VQA dataset. We also benchmark PICa on VQAv2, where PICa also shows a decent few-shot performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PICa, a simple yet effective method that Prompts GPT3 via the use of Image Captions, for knowledge-based VQA, and treats GPT-3 as an implicit and unstructured KB that can jointly acquire and process relevant knowledge.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20215/19974
-Comprehensive Review of Deep Learning-Based 3D Point Cloud Completion Processing and Analysis,Ben Fei,"Point cloud completion is a generation and estimation issue derived from the partial point clouds, which plays a vital role in the applications of 3D computer vision. The progress of deep learning (DL) has impressively improved the capability and robustness of point cloud completion. However, the quality of completed point clouds is still needed to be further enhanced to meet the practical utilization. Therefore, this work aims to conduct a comprehensive survey on various methods, including point-based, view-based, convolution-based, graph-based, generative model-based, transformer-based approaches, etc. And this survey summarizes the comparisons among these methods to provoke further research insights. Besides, this review sums up the commonly used datasets and illustrates the applications of point cloud completion. Eventually, we also discussed possible research trends in this promptly expanding field.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to conduct a comprehensive survey on various methods of point cloud completion, including point-based, view- based, convolution-based), convolution, graph, graph based, generative model based, transformer-based approaches, etc, and summarizes the comparisons among these methods to provoke further research insights.'}",https://arxiv.org/pdf/2203.03311
-Interactive-Chain-Prompting: Ambiguity Resolution for Crosslingual Conditional Generation with Interaction,Jonathan Pilault,"Crosslingual conditional generation (e.g., machine translation) has long enjoyed the benefits of scaling. Nonetheless, there are still issues that scale alone may not overcome. A source query in one language, for instance, may yield several translation options in another language without any extra context. Only one translation could be acceptable however, depending on the translator's preferences and goals. Choosing the incorrect option might significantly affect translation usefulness and quality. We propose a novel method interactive-chain prompting -- a series of question, answering and generation intermediate steps between a Translator model and a User model -- that reduces translations into a list of subproblems addressing ambiguities and then resolving such subproblems before producing the final text to be translated. To check ambiguity resolution capabilities and evaluate translation quality, we create a dataset exhibiting different linguistic phenomena which leads to ambiguities at inference for four languages. To encourage further exploration in this direction, we release all datasets. We note that interactive-chain prompting, using eight interactions as exemplars, consistently surpasses prompt-based methods with direct access to background information to resolve ambiguities.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method interactive-chain prompting -- a series of question, answering and generation intermediate steps between a Translator model and a User model -- that reduces translations into a list of subproblems addressing ambiguities and then resolving such subpro problems before producing the final text to be translated.'}",http://arxiv.org/pdf/2301.10309
-The Effectiveness of Embedded Teaching through the Most-to-Least Prompting Procedure in Concept Teaching to Children with Autism within Orff-Based Music Activities.,Bilgehan Eren,"AbstractThe purpose of this study was to demonstrate the effectiveness of embedded teaching through the most-to- least prompting procedure in concept teaching to children with autism in Orff-based music activities. In this research, being one of the single subject research designs, multiple probe design was used. The generalization effect of the research was assessed in the form of a pre-test and post-test, discussing the organization about interpersonal generalization in different environment. The maintenance of learning was checked by means of the observation sessions applied one, two, and four weeks after the practice was completed. The research was conducted with 3 children with autism between the ages of 3-6 in Nova Special Training and Rehabilitation Center in Istanbul. In research process, full probe, daily probe, teaching, observing and generalizing sessions have been organized. All of the sessions were done by one-to-one teaching method. The data of reliability among the observers and of application reliability were collected. According to research results, it was clear that the embedding teaching through the most-to-least prompting in Orff-based music activities is effective to teach concepts to the chilren with autism. It was understood that the maintenance was kept after the training was completed. It was also clear that the concept which was thought is effective in the generalization of different settings, different people and different materials. Additionally, social validity was verified in research to deter- mine the importance of purposes of research, teaching methods to reach these purposes and research results.Key WordsChildren with Autism, Orff Approach, Embedded Teaching, Most-to-Least Prompting, Concept Teaching.Defined as permanent behavioral change arising from experiences in behavior or potential behavior, ""learning"" occurs through different ways for each person (Senemoglu, 2007). Special arrangements should be made for teaching children who have special needs. Functional education programs are developed and systematic teaching practices are implemented with regard to teaching of independent life skills to children with developmental disabilities (Eripek, 2005; Kurt, 2009).According to Special Education Services Regulation of the Ministry of National Education (2011), an individual who needs special education is defined as an individual who shows significant difference compared to the level expected from fellows with regard to educational and individual qualifications due to various reasons'. Individual with autism is defined as an individual whose restraints in social interaction, verbal and non-verbal communication, as well as interest and activity develop in early childhood period and who need special education and supportive education service due to these restraints (Milli Egitim Bakanhgi [MEB], 2012), while autistic disorder is defined as a developmental disorder characterized by incapacity in social interaction, language and communication abilities; restricted, stereotyped and repetitive patterns in behaviors, interest and activities, as well as disorders or extraordinary functionality in at least one of the following: social interaction, language used in social communication, or symbolic/imaginative play abilities, which occur before the age of 3 (Diken, 2008; Kircaali-Iftar, 2003; Korkmaz, 2005). In order to offer an efficient education by exploring individual learning methods of children with autism, it is required to organize the physical environment according to their requirements, work on improper behaviors that they exhibit, utilize visual supports and prompts, and to determine the most suitable teaching techniques for them (Korkmaz, 2003).Used in education of children with autism, 'Applied Behavior Analysis' involves the following steps respectively: determination of current performance of child, identification of fields where child suffers from skill deficits, classifying identified abilities into sub-steps, involving abilities (which are classified into sub-steps) into individual programs, and rewarding positive reactions of child (Darica, Abidoglu, & Gumuccu, 2005). …",,
-A Feasibility Study of Enhancing Independent Task Performance for People with Cognitive Impairments Through the Use of a Handheld Location-Based Prompting System,Yao-Jen Chang,"An autonomous task-prompting system is presented to increase workplace and life independence for people with cognitive impairments such as traumatic brain injury, intellectual disability, schizophrenia, and down syndrome. This paper describes an approach to providing distributed cognition support of work engagement for persons with cognitive disabilities. In the pilot study, a prototype was built and tested in a community-based rehabilitation program involving preservice food preparation training of eight participants with cognitive impairments. The results show improvement in helping with task engagement is statistically significant compared to the oral-instruction method. A follow-up comparative study with two participants evaluated the shadow-team approach against the proposed system. Although the number of participants was few, the participants were studied in depth and the findings were very promising. The results in the autonomous task prompting without staff intervention indicate that the performance is statistically as good as the shadow-team approach. Our findings suggest that acquisition of job skills may be facilitated by the proposed system in conjunction with operant conditioning strategies.","{'model': 'tldr@v2.0.0', 'text': 'The results in the autonomous task prompting without staff intervention indicate that the performance is statistically as good as the shadow-team approach, and suggest that acquisition of job skills may be facilitated by the proposed system in conjunction with operant conditioning strategies.'}",
-Medicinal Chemistry Projects Requiring Imaginative Structure-Based Drug Design Methods.,N. Moitessier,"Computational methods for docking small molecules to proteins are prominent in drug discovery. There are hundreds, if not thousands, of documented examples-and several pertinent cases within our research program. Fifteen years ago, our first docking-guided drug design project yielded nanomolar metalloproteinase inhibitors and illustrated the potential of structure-based drug design. Subsequent applications of docking programs to the design of integrin antagonists, BACE-1 inhibitors, and aminoglycosides binding to bacterial RNA demonstrated that available docking programs needed significant improvement. At that time, docking programs primarily considered flexible ligands and rigid proteins. We demonstrated that accounting for protein flexibility, employing displaceable water molecules, and using ligand-based pharmacophores improved the docking accuracy of existing methods-enabling the design of bioactive molecules. The success prompted the development of our own program, Fitted, implementing all of these aspects. The primary motivation has always been to respond to the needs of drug design studies; the majority of the concepts behind the evolution of Fitted are rooted in medicinal chemistry projects and collaborations. Several examples follow: (1) Searching for HDAC inhibitors led us to develop methods considering drug-zinc coordination and its effect on the pKa of surrounding residues. (2) Targeting covalent prolyl oligopeptidase (POP) inhibitors prompted an update to Fitted to identify reactive groups and form bonds with a given residue (e.g., a catalytic residue) when the geometry allows it. Fitted-the first fully automated covalent docking program-was successfully applied to the discovery of four new classes of covalent POP inhibitors. As a result, efficient stereoselective syntheses of a few screening hits were prioritized rather than synthesizing large chemical libraries-yielding nanomolar inhibitors. (3) In order to study the metabolism of POP inhibitors by cytochrome P450 enzymes (CYPs)-for toxicology studies-the program Impacts was derived from Fitted and helped us to reveal a complex metabolism with unforeseen stereocenter isomerizations. These efforts, combined with those of other docking software developers, have strengthened our understanding of the complex drug-protein binding process while providing the medicinal chemistry community with useful tools that have led to drug discoveries. In this Account, we describe our contributions over the past 15 years-within their historical context-to the design of drug candidates, including BACE-1 inhibitors, POP covalent inhibitors, G-quadruplex binders, and aminoglycosides binding to nucleic acids. We also remark the necessary developments of docking programs, specifically Fitted, that enabled structure-based design to flourish and yielded multiple fruitful, rational medicinal chemistry campaigns.","{'model': 'tldr@v2.0.0', 'text': 'Your contributions over the past 15 years are described, within their historical context-to the design of drug candidates, including BACE-1 inhibitors, POP covalent inhibitors, G-quadruplex binders, and aminoglycosides binding to nucleic acids.'}",
-Challenges from Tuberculosis Diagnosis to Care in Community-Based Active Case Finding among the Urban Poor in Cambodia: A Mixed-Methods Study,N. Lorent,"Background While community-based active case finding (ACF) for tuberculosis (TB) holds promise for increasing early case detection among hard-to-reach populations, limited data exist on the acceptability of active screening. We aimed to identify barriers and explore facilitators on the pathway from diagnosis to care among TB patients and health providers. Methods Mixed-methods study. We administered a survey questionnaire to, and performed in-depth interviews with, TB patients identified through ACF from poor urban settlements in Phnom Penh, Cambodia. Additionally, we conducted focus group discussions and in-depth interviews with community and public health providers involved in ACF, respectively. Results Acceptance of home TB screening was strong among key stakeholders due to perceived reductions in access barriers and in direct and indirect patient costs. Privacy and stigma were not an issue. To build trust and facilitate communication, the participation of community representatives alongside health workers was preferred. Most health providers saw ACF as complementary to existing TB services; however, additional workload as a result of ACF was perceived as straining operating capacity at public sector sites. Proximity to a health facility and disease severity were the strongest determinants of prompt care-seeking. The main reasons reported for delays in treatment-seeking were non-acceptance of diagnosis, high indirect costs related to lost income/productivity and transportation expenses, and anticipated side-effects from TB drugs. Conclusions TB patients and health providers considered home-based ACF complementary to facility-based TB screening. Strong engagement with community representatives was believed critical in gaining access to high risk communities. The main barriers to prompt treatment uptake in ACF were refusal of diagnosis, high indirect costs, and anticipated treatment side-effects. A patient-centred approach and community involvement were essential in mitigating barriers to care in marginalised communities.","{'model': 'tldr@v2.0.0', 'text': 'The main barriers to prompt treatment uptake in ACF were refusal of diagnosis, high indirect costs, and anticipated treatment side-effects, and a patient-centred approach and community involvement were essential in mitigating barriers to care in marginalised communities.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0130179&type=printable
-Applying cusum-based methods for the detection of outbreaks of Ross River virus disease in Western Australia,R. Watkins,,"{'model': 'tldr@v2.0.0', 'text': 'The retrospective analysis of historical data suggests that the negative binomial cusum provides greater sensitivity for the detection of outbreaks of RRv disease at low false alarm levels, and decreased timeliness early in the outbreak period.'}",https://bmcmedinformdecismak.biomedcentral.com/counter/pdf/10.1186/1472-6947-8-37
-Embedded Simultaneous Prompting Procedure to Teach STEM Content to High School Students with Moderate Disabilities in an Inclusive Setting.,S. Heinrich,"Effects of an embedded simultaneous prompting procedure to teach STEM (science, technology, engineering, math) content to three secondary students with moderate intellectual disabilities in an inclusive general education classroom were evaluated in the current study. Students learned discrete (i.e., geometric figures, science vocabulary, or use of technology to publish) and chained tasks (i.e., linear equation, Punnett square, or Internet research) from a peer tutor and paraprofessionals. Using a multiple probe across participants design, results showed students reached criterion in two to eight sessions and maintained the skills for one month following intervention. Generalization was also at higher levels than in the baseline condition. In addition, general education students who attended class with the participants expressed positive comments based on the intervention. The National Science Board has argued that every student in the United States “. . . deserves the opportunity to achieve his or her full potential” (NSF, 2010, p. v). As STEM (science, technology, engineering, math) permeates every aspect of our lives, it is crucial that all students have access to this content. Students benefit from science content by learning about themselves and the natural world; from engaging in technology by learning how to use smart phones, mobile devices, and computers; from engineering concepts by learning how to solve problems; and from math content by learning how to budget their money and determining how much they have for groceries (Science Pioneers, 2013). There is a national emphasis on preparing students for STEM careers; however, all citizens, even those not pursuing STEM careers, should be able to participate in the scientific and technical issues affecting our society (Matthews, 2007). Traditionally, students with disabilities have been underrepresented in STEM education, and students with moderate intellectual disabilities are especially at risk. Common Core State Standards (CCSS) and the Next Generation Science Standards (NGSS) emphasize the importance of high expectations for all students. According to Kleinert, Kearns, and Kennedy (1997), “. . . one of the keys to ensuring high expectations for every child is requiring that all students be included in measures of educational accountability” (p. 88). Based on these assertions, it is likely that assessments for all students, including alternate assessments based on alternate achievement standards (AA-AAS) for students with moderate and severe disabilities (MSD), will be derived from the content recommended in the CCSS and NGSS. While practices for implementing the AAAAS vary from state to state, having access to the general education curriculum is key for promoting student progress in content areas. Jackson, Ryndak, and Wehmeyer (2010) argued that the interplay between context of Correspondence concerning this article should be addressed to Amy D. Spriggs, Department of Early Childhood, Special Education, and Rehabilitation Counseling, University of Kentucky, 229 Taylor Education Building, Lexington, KY 40506-0001. E-mail: amy.spriggs@uky.edu Education and Training in Autism and Developmental Disabilities, 2016, 51(1), 41–54 © Division on Autism and Developmental Disabilities Teaching STEM Content to High School Students / 41 instruction and curriculum content is a critical consideration in student learning and progress since students with disabilities who are included in general education contexts are more likely to engage in activities related to grade level standards than students who are in self-contained settings. Hudson, Browder, and Wood (2013) conducted a literature review to evaluate evidence-based practices for students with MSD in inclusive contexts. Authors found embedded instruction (specifically using constant time delay [CTD]) to be an evidence-based practice. Embedded instruction can be defined as an instructor (teacher, paraprofessional, or peer) distributing trials throughout a lesson or naturally occurring routine in the inclusive classroom. Studies in the review focused on teaching reading and writing skills (e.g., Collins, Hall, Branson, & Holder, 1999), social studies skills (e.g., Collins, Evans, Creech-Galloway, Karl, & Miller, 2007), science skills (e.g., Jimenez, Browder, Spooner, & DiBiase, 2012), and math skills (McDonnell et al., 2006) to students with moderate and severe disabilities within general education classrooms. Hudson et al. (2013) commended researchers for evaluating procedures to teach a range of academic content in general education settings; however, they recommended additional research on strategies to teach complex and chained skills in inclusive environments. Further, only two quality studies in the review used simultaneous prompting (SP), suggesting a need for further research on this strategy in general education settings. In most of the quality studies reviewed in the Hudson et al. (2013) review, researchers used systematic instruction to teach students core content in general education settings. Methods of systematic instruction include the system of least prompts (SLP), time delay, and simultaneous prompting procedures. Perhaps the easiest and least time consuming method to implement is the SP procedure because the instructor presents the stimulus and the controlling prompt at the same time on each training trial, conducting daily probe trials prior to instructional trials to determine if the student has acquired the target skill (Collins, 2012). Collins et al. (2007) used the SP procedure to teach core content vocabulary to 4 students with MSD at the elementary, middle, and secondary levels, finding the procedure to be effective whether it was delivered in a segregated or an inclusive classroom setting. Riesen, McDonnell, Johnson, Polychronis, and Jameson (2003) and Head, Collins, Schuster, and Ault (2011) compared the efficiency of the CTD and SP procedures in teaching academic content (e.g., science and history vocabulary, state capitols) to middle or high school students with disabilities. In both studies, the researchers found the CTD and SP procedures to be equally effective, but the SP procedure was found to be slightly more efficient (i.e., faster acquisition of skills). A body of literature has emerged suggesting that, in addition to academic gains, inclusive settings offer the opportunity for positive social effects on students, both with and without disabilities (e.g., Kennedy, Shukla, & Fryxell, 1997). Results have indicated greater social benefits for students who were included full time in general education classrooms. For example, students who were included full time engaged in more frequent interactions with peers without disabilities, more social contacts across a wide range of activities and settings, higher levels of social support behaviors, larger friendship networks, and longer lasting relationships with peers. While prior research has shown promising effects as a result of teaching in inclusive classrooms, there are several limitations to the research. For example, few studies to date have taught students complex or chained skills in inclusive settings, and no studies exist evaluating STEM-related chained tasks. Hudson et al. (2013) recommended future research using SP in inclusive settings, and despite previous calls for research, sight word identification remains the skill most frequently addressed in these settings. In addition, no research exists on teaching STEM skills identified as part of the AA-AAS in a general education classroom using an embedded SP procedure. The goal of this study is to evaluate the effects of using an embedded SP procedure to teach STEMrelated tasks to students with moderate intellectual disabilities in an inclusive setting. The current study examined instruction on the state standards related to STEM content on which students taking part in the AA-AAS were tested. The study occurred in a suburban high school in a southern state, therefore fo42 / Education and Training in Autism and Developmental Disabilities-March 2016 cusing on the state’s AA-AAS. The special education teacher and the general education teacher collaborated to develop a plan for instruction in the general education classroom to be delivered by a paraprofessional or a peer tutor, which included embedding the SP procedure into the natural routine of the general education classroom. Specifically, the current study focused on the following research questions: (a) Will the embedded SP procedure result in the acquisition of STEM content skills by students with moderate intellectual disabilities in the general education classroom? (b) Will peer tutors and paraprofessionals implement the embedded SP procedure with high levels of procedural fidelity? (c) Will the target students maintain the acquired skills over time? (d) Will the target students generalize the STEM skills to situations that simulate the AA-AAS? and (e) Will inclusion of the target students in the general education classroom result in positive social relationships with their peers without disabilities?",,
-DNA-based methods for monitoring invasive species: a review and prospectus,J. Darling,,,
-Technological Application of Tannin-Based Extracts,M. Fraga-Corral,"Tannins are polyphenolic compounds naturally found in vegetables. Their presence in nature has prompted their historical use in many different ways. The revision of their traditional utilization has allowed their further modification aiming for an industrial application. Sometimes these modifications have implied the addition of harmful substances such as formaldehyde, classified as a carcinogen of category B1. In other cases, these natural tannins have been replaced by synthetic compounds that threaten human and animal health and damage the environment. Therefore, currently, both academy and industry are searching for the substitution of these unsafe complexes by the increasing inclusion of tannins, natural molecules that can be obtained from several and diverse renewable resources, modified using harmless additives. To achieve promising results, cost-efficient and eco-friendly extraction methods have been designed. Once these green alternatives have been isolated, they have been successfully applied to many fields with very assorted aims of utilization such as coagulants, adhesives, floatation agents, tannings, dyes, additives, or biomolecules. Therefore, this review offers a global vision of the full process that involves the tannin’s technological application including an overview of the most relevant tannin sources, effective extraction methods, and their utilization in very diverse fields.","{'model': 'tldr@v2.0.0', 'text': 'This review offers a global vision of the full process that involves the tannin’s technological application including an overview of the most relevant tannIn sources, effective extraction methods, and their utilization in very diverse fields.'}",https://www.mdpi.com/1420-3049/25/3/614/pdf?version=1581334096
-A comparison of automatic cell identification methods for single-cell RNA sequencing data,T. Abdelaal,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that most classifiers perform well on a variety of datasets with decreased accuracy for complex datasets with overlapping classes or deep annotations, but the general-purpose support vector machine classifier has overall the best performance across the different experiments.'}",https://genomebiology.biomedcentral.com/track/pdf/10.1186/s13059-019-1795-z
-PSG: Prompt-based Sequence Generation for Acronym Extraction,Bin Li,"Acronym extraction aims to find acronyms (i.e., short-forms) and their meanings (i.e., long-forms) from the documents, which is important for scientific document understanding (SDU@AAAI-22) tasks. Previous works are devoted to modeling this task as a paragraph-level sequence labeling problem. However, it lacks the effective use of the external knowledge, especially when the datasets are in a low-resource setting. Recently, the prompt-based method with the vast pre-trained language model can significantly enhance the performance of the low-resourced downstream tasks. In this paper, we propose a Prompt-based Sequence Generation (PSG) method for the acronym extraction task. Specifically, we design a template for prompting the extracted acronym texts with auto-regression. A position extraction algorithm is designed for extracting the position of the generated answers. The results on the acronym extraction of Vietnamese and Persian in a low-resource setting show that the proposed method outperforms all other competitive state-of-the-art (SOTA) methods.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt-based Sequence Generation (PSG) method for the acronym extraction task, designed for prompting the extracted acronym texts with auto-regression, and a position extraction algorithm is designed for extracting the position of the generated answers.'}",
-FateZero: Fusing Attentions for Zero-shot Text-based Video Editing,Chenyang Qi,"The diffusion-based generative models have achieved remarkable success in text-based image generation. However, since it contains enormous randomness in generation progress, it is still challenging to apply such models for real-world visual content editing, especially in videos. In this paper, we propose FateZero, a zero-shot text-based editing method on real-world videos without per-prompt training or use-specific mask. To edit videos consistently, we propose several techniques based on the pre-trained models. Firstly, in contrast to the straightforward DDIM inversion technique, our approach captures intermediate attention maps during inversion, which effectively retain both structural and motion information. These maps are directly fused in the editing process rather than generated during denoising. To further minimize semantic leakage of the source video, we then fuse self-attentions with a blending mask obtained by cross-attention features from the source prompt. Furthermore, we have implemented a reform of the self-attention mechanism in denoising UNet by introducing spatial-temporal attention to ensure frame consistency. Yet succinct, our method is the first one to show the ability of zero-shot text-driven video style and local attribute editing from the trained text-to-image model. We also have a better zero-shot shape-aware editing ability based on the text-to-video model. Extensive experiments demonstrate our superior temporal consistency and editing capability than previous works.","{'model': 'tldr@v2.0.0', 'text': 'F FateZero, a zero-shot text-based editing method on real-world videos without per-prompt training or use-specific mask, is proposed, which is the first one to show the ability of zero- shot text-driven video style and local attribute editing from the trained text-to-image model.'}",https://arxiv.org/pdf/2303.09535
-Fast and Constrained Absent Keyphrase Generation by Prompt-Based Learning,Huanqin Wu,"Generating absent keyphrases, which do not appear in the input document, is challenging in the keyphrase prediction task. Most previous works treat the problem as an autoregressive sequence-to-sequence generation task, which demonstrates promising results for generating grammatically correct and fluent absent keyphrases. However, such an end-to-end process with a complete data-driven manner is unconstrained, which is prone to generate keyphrases inconsistent with the input document. In addition, the existing autoregressive decoding method makes the generation of keyphrases must be done from left to right, leading to slow speed during inference. In this paper, we propose a constrained absent keyphrase generation method in a prompt-based learning fashion. Specifically, the prompt will be created firstly based on the keywords, which are defined as the overlapping words between absent keyphrase and document. Then, a mask-predict decoder is used to complete the absent keyphrase on the constraint of prompt. Experiments on keyphrase generation benchmarks have demonstrated the effectiveness of our approach. In addition, we evaluate the performance of constrained absent keyphrases generation from an information retrieval perspective. The result shows that our approach can generate more consistent keyphrases, which can improve document retrieval performance. What’s more, with a non-autoregressive decoding manner, our model can speed up the absent keyphrase generation by 8.67× compared with the autoregressive method.","{'model': 'tldr@v2.0.0', 'text': 'The result shows that the proposed constrained absent keyphrase generation method can generate more consistent keyphrases, which can improve document retrieval performance, and with a non-autoregressive decoding manner, can speed up the absentKeyphrase generation by 8.67× compared with the autoregressive method.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21402/21151
-Rare-earth based nanomaterials and their composites as electrode materials for high performance supercapacitors: a review,Shunfei Liang,"The emergence of energy crisis and greenhouse effect has prompted people to develop energy storage equipment with excellent performance. Supercapacitors (SCs), also known as electrochemical capacitors, are widely studied for their high power density, fast charge and discharge and long cycle life. Rare earth is a group of elements with unique structure and properties. Rare earth-based nanomaterials, especially rare earth oxides, hydroxides and sulfides, have attracted much attention due to their good redox properties and environmentally friendly characteristics. Rare earth-based SCs nanomaterials can be obtained by environmentally friendly, simple and low-cost methods, such as hydrothermal/solvothermal method, electrodeposition method, atomic layer deposition method, etc. This paper reviews the synthesis methods, morphology and electrochemical properties of various rare earth-based nanomaterials and their composites for SCs. Finally, based on the reported literature, we discussed some challenges faced by rare earth-based nanomaterials and the issues that need attention in SCs and other electrochemical energy storage fields in the future.",,
-AI-based Prognostic Imaging Biomarkers for Precision Neurooncology: the ReSPOND Consortium.,C. Davatzikos,"Artificial intelligence (AI) and machine learning (ML) methods have begun to reveal that complex imaging patterns can provide individualized biomarkers for diagnosis and prognosis. However, AI methods have been challenged by insufficient training, heterogeneity of imaging protocols across hospitals, and lack of generalization to new patient data. These challenges prompted the development of the ReSPOND (Radiomics Signatures for PrecisiON Diagnostics) consortium on glioblastoma (GBM). This collaboration of over 10 institutions, across 3 continents, is positioned to pool, harmonize, and analyze brain MRIs from more than 3300 de novo GBM patients who underwent the Stupp protocol, in addition to datasets from The Cancer Imaging Archive (TCIA).1 ReSPOND aims to further develop and test AI-based biomarkers for individualized prediction and prognostication, by moving from single-institution studies to generalized, well-validated predictive biomarkers in the following 4 areas:","{'model': 'tldr@v2.0.0', 'text': 'The ReSPOND consortium is positioned to pool, harmonize, and analyze brain MRIs from more than 3300 de novo GBM patients who underwent the Stupp protocol, in addition to datasets from The Cancer Imaging Archive (TCIA).'}",https://academic.oup.com/neuro-oncology/article-pdf/22/6/886/33373505/noaa045.pdf
-Technical attribute prioritisation in QFD based on cloud model and grey relational analysis,Xu Wang,"Promptly development of new products can be achieved through quality function deployment (QFD) process, which is critical to companies’ survival. Since the multi-criteria decision-making problem involved in QFD, a novel method integrating cloud model and grey relational analysis is put forward in this paper. Taking into account the subjectivity and ambiguity in linguistic evaluations, some scholars utilise fuzzy theory, rough theory, interval-valued fuzzy-rough sets and MCDM methods to improve traditional QFD. However, much priori information requirements, inability to handle subjectivity and randomness, and lack of mechanism to overcome small sample size problem are some inevitable drawbacks in these methods. To solve these deficiencies, a hybrid methodology is proposed in this paper, integrating the fortes of cloud model in processing ambiguity and randomness, and the merits of grey relational analysis in overcoming small sample size error as well as revealing the inner correlations. The comparative analysis of different approaches as well as the sensitivity analysis of criteria weights is implemented to prove the stability of the novel method. The results obtained in this paper shows that the proposed method can be a practical tool for improving the efficiency and accuracy of traditional QFD in reality management.","{'model': 'tldr@v2.0.0', 'text': 'A hybrid methodology is proposed in this paper, integrating the fortes of cloud model in processing ambiguity and randomness, and the merits of grey relational analysis in overcoming small sample size error as well as revealing the inner correlations.'}",
-Using simple technology to prompt multistep tasks in the home for people with dementia: An exploratory study comparing prompting formats,H. Boyd,"Objectives To investigate the relative effectiveness of different prompts for people with dementia during multistep tasks in the home, to inform prompting technology design. Methods Nine pairs of participants (one with dementia and a partner or relative) participated at home. The participants with mild to moderate dementia (5M/4F, aged 73–86 years) functioned at the Planned or Exploratory levels of the Pool Activity Level instrument. A touchscreen computer displayed different prompts during two set tasks: “card-and-envelope” and “CD player.” The trials were scored to establish the relative effectiveness of the prompts. Individual tasks were also explored. Results Text and audio prompts were each more effective than video or picture prompts for a card-and-envelope task, but this was not seen in a CD player task. The differences may be related to the type of actions within the tasks; the card-and-envelope actions were easier to convey verbally; the CD player actions lent themselves to visual prompts. Conclusions Designers of technology-based prompts for people with dementia should consider that the effectiveness of different prompts is likely to be task dependent. Familiar, unambiguous language can increase the success of tailored prompts. There are significant practical challenges associated with choosing and deconstructing everyday tasks at home.","{'model': 'tldr@v2.0.0', 'text': 'Designers of technology-based prompts for people with dementia should consider that the effectiveness of different prompts is likely to be task dependent, and familiar, unambiguous language can increase the success of tailored prompts.'}",https://journals.sagepub.com/doi/pdf/10.1177/1471301215602417
-Green Extraction Methods for Polyphenols from Plant Matrices and Their Byproducts: A Review.,K. Ameer,"Polyphenols as phytochemicals have gained significant importance owing to several associated health benefits with regard to lifestyle diseases and oxidative stress. To date, the development of a single standard method for efficient and rapid extraction of polyphenols from plant matrices has remained a challenge due to the inherent limitations of various conventional extraction methods. The exploitation of polyphenols as bioactive compounds at various commercial levels has motivated scientists to explore more eco-friendly, efficient, and cost-effective extraction techniques, based on a green extraction approach. The current review aims to provide updated technical information about extraction mechanisms, their advantages and disadvantages, and factors affecting efficiencies, and also presents a comparative overview of applications of the following modern green extraction techniques-supercritical fluid extraction, ultrasound-assisted extraction, microwave-assisted extraction, pressurized liquid extraction, and pressurized hot water extraction-as alternatives to conventional extraction methods for polyphenol extraction. These techniques are proving to be promising for the extraction of thermolabile phenolic compounds due to their advantages over conventional, time-consuming, and laborious extraction techniques, such as reduced solvent use and time and energy consumption and higher recovery rates with lower operational costs. The growing interest in plant-derived polyphenols prompts continual search for green and economically feasible modern extraction techniques. Modern green extraction techniques represent promising approaches by virtue of overcoming current limitations to the exploitation of polyphenols as bioactive compounds to explore their wide-reaching applications on an industrial scale and in emerging global markets. Future research is needed in order to remove the technical barriers to scale-up the processes for industrial needs by increasing our understanding and improving the design of modern extraction operations.","{'model': 'tldr@v2.0.0', 'text': 'These techniques are proving to be promising for the extraction of thermolabile phenolic compounds due to their advantages over conventional, time-consuming, and laborious extraction techniques, such as reduced solvent use and time and energy consumption and higher recovery rates with lower operational costs.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/1541-4337.12253
-Evaluation of novel precision viticulture tool for canopy biomass estimation and missing plant detection based on 2.5D and 3D approaches using RGB images acquired by UAV platform,S. F. Di Gennaro,,"{'model': 'tldr@v2.0.0', 'text': 'A rapid and objective tool for the farmer to promptly identify canopy management strategies and drive replanting decisions and the overcoming of the current limit represented by the pre- and post-processing phases of the large image dataset should mainstream this methodology.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-020-00632-2
-2016 updated EULAR evidence-based recommendations for the management of gout,P. Richette,"Background New drugs and new evidence concerning the use of established treatments have become available since the publication of the first European League Against Rheumatism (EULAR) recommendations for the management of gout, in 2006. This situation has prompted a systematic review and update of the 2006 recommendations. Methods The EULAR task force consisted of 15 rheumatologists, 1 radiologist, 2 general practitioners, 1 research fellow, 2 patients and 3 experts in epidemiology/methodology from 12 European countries. A systematic review of the literature concerning all aspects of gout treatments was performed. Subsequently, recommendations were formulated by use of a Delphi consensus approach. Results Three overarching principles and 11 key recommendations were generated. For the treatment of flare, colchicine, non-steroidal anti-inflammatory drugs (NSAIDs), oral or intra-articular steroids or a combination are recommended. In patients with frequent flare and contraindications to colchicine, NSAIDs and corticosteroids, an interleukin-1 blocker should be considered. In addition to education and a non-pharmacological management approach, urate-lowering therapy (ULT) should be considered from the first presentation of the disease, and serum uric acid (SUA) levels should be maintained at<6 mg/dL (360 µmol/L) and <5 mg/dL (300 µmol/L) in those with severe gout. Allopurinol is recommended as first-line ULT and its dosage should be adjusted according to renal function. If the SUA target cannot be achieved with allopurinol, then febuxostat, a uricosuric or combining a xanthine oxidase inhibitor with a uricosuric should be considered. For patients with refractory gout, pegloticase is recommended. Conclusions These recommendations aim to inform physicians and patients about the non-pharmacological and pharmacological treatments for gout and to provide the best strategies to achieve the predefined urate target to cure the disease.","{'model': 'tldr@v2.0.0', 'text': 'In addition to education and a non-pharmacological management approach, urate-lowering therapy (ULT) should be considered from the first presentation of the disease, and serum uric acid (SUA) levels should be maintained at the predefined urate target to cure the disease.'}",https://ard.bmj.com/content/annrheumdis/76/1/29.full.pdf
-"UV-Protective, Self-Cleaning, and Antibacterial Nanofiber-Based Triboelectric Nanogenerators for Self-Powered Human Motion Monitoring.",Yang Jiang,"Equipping wearable electronics with special functions will endow them with more additional values and more comprehensive practical performance. Here, we report an ultraviolet (UV)-protective, self-cleaning, antibacterial, and self-powered all-nanofiber-based triboelectric nanogenerator (TENG) for mechanical energy harvesting and self-powered sensing, which is fabricated with Ag nanowires (NWs)/TPU nanofibers and the TiO2@PAN networks through a facile electrospinning method. Due to the added TiO2 nanoparticles (NPs), the TENG presents excellent UV-protective performance, including the ultraviolet protection factor (UPF) of ∼204, the transmittance of UVA (TUVA) of ∼0.0574%, and the transmittance of UVB (TUVB) ∼0.107%. Furthermore, under solar lighting for 25 min, most surface contamination can be degraded, and the decreased power output would be recovered. Owing to the coupled effects of TiO2 NPs and Ag NWs, the TENG shows excellent antibacterial activity against Staphylococcus aureus. Due to the micro-to-nano hierarchical porous structure, the all-nanofiber-based TENG can serve as self-powered pedometers for detecting and tracking human motion behaviors. As a multifunctional self-powered device, the TENG prompts various applications in the fields of micro/nanopower sources, human movement monitoring, and human-machine interfaces, potentially providing an alternative energy solution and a multifunctional interactive platform for the next-generation wearable electronics.","{'model': 'tldr@v2.0.0', 'text': 'An ultraviolet (UV)-protective, self-cleaning, antibacterial, and self-powered all-nanofiber-based triboelectric nanogenerator (TENG) for mechanical energy harvesting andSelf-powered sensing that prompts various applications in the fields of micro/nanopower sources, human movement monitoring, and human-machine interfaces is reported.'}",
-Prompt-based Re-ranking Language Model for ASR,Mengxi Nie,"In Automatic Speech Recognition(ASR), the language model re-ranking based on unlabeled text can improve the performance and realize flexibly scene adaptation. The scheme of ASR re-ranking is usually to build a language model and then use it to reorder the speech recognition N-best hypotheses. Recently, BERT-based re-ranking has achieved impressive results, benefiting from the powerful modeling capability of contextual semantic. In the view of that BERT’s nonautoregressive structure limits the calculation speed of the language model scores(perplexity, ppl), we use a classification method in prompt paradigm instead of the re-ranking method based on ppl. The prompt-based re-ranking scheme simplifies the pipeline of re-ranking as well as ensures the performance. Experiments on AISHELL-1 dataset show the effective of our proposed method. On the test set, the inference speed is accelerated by 49 times and compared to baseline the Character Error Rate(CER) is relatively decreased by 13.51% ∼ 14.43%.","{'model': 'tldr@v2.0.0', 'text': 'In the view of that BERT’s nonautoregressive structure limits the calculation speed of the language model scores(perplexity, ppl), this work uses a classification method in prompt paradigm instead of the re-ranking method based on ppl.'}",
-The effectiveness of video prompting on teaching aquatic play skills for children with autism,M. Yanardag,"Objective: To investigate the effectiveness of the video prompting procedure on teaching aquatic play skills and to determine the effects of aquatic exercise training on the motor performance of children with autism. Design: A multiple probe design across behaviours was used and replicated across subjects for the instructional part of this study. Pretest–posttest design was applied for the exercise training part of this study. Methods: Three children with autism were taught three aquatic play skills in a one-to-one training format. Aquatic play skills intervention and aquatic exercise training were performed separately throughout 12 weeks at three sessions per week, each lasting 1 h. The video prompting procedure was utilized for the instruction part of this study. Results: Video prompting was effective in teaching aquatic play skills to children with autism. In addition, aquatic exercise training increased the total motor performance scores of all the participants after 12 weeks. According to the social validity results, the families gave positive feedback about the learned skills and movement capabilities of their children. Conclusion: Aquatic play skills and swimming pools are favoured for children with autism. This attractive intervention is recommended as a means to extend knowledge of leisure skills and motor development of children with autism. Implications for Rehabilitation Video prompting may utilize to teach action-based motor tasks. Aquatic exercise training could be beneficial for improving movement difficulties of children with autism. Video-based instruction and swimming pool is recommended to have alluring effects on children with special needs.","{'model': 'tldr@v2.0.0', 'text': 'Video prompting was effective in teaching aquatic play skills to children with autism and aquatic exercise training increased the total motor performance scores of all the participants after 12 weeks.'}",
-Classification of normal and depressed EEG signals based on centered correntropy of rhythms in empirical wavelet transform domain,H. Akbari,,"{'model': 'tldr@v2.0.0', 'text': 'A new method based on centered correntropy (CC) and empirical wavelet transform (EWT) for the classification of normal and depressed EEG signals can be used as a fast and accurate computer-aided detection system for the diagnosis of patients with depression in clinics and hospitals.'}",
-LFPT5: A Unified Framework for Lifelong Few-shot Language Learning Based on Prompt Tuning of T5,Chengwei Qin,"Existing approaches to lifelong language learning rely on plenty of labeled data for learning a new task, which is hard to obtain in most real scenarios. Considering that humans can continually learn new tasks from a handful of examples, we expect the models also to be able to generalize well on new few-shot tasks without forgetting the previous ones. In this work, we define this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and propose a unified framework for it based on prompt tuning of T5. Our framework called LFPT5 takes full advantage of PT's strong few-shot learning ability, and simultaneously trains the model as a task solver and a data generator. Before learning a new domain of the same task type, LFPT5 generates pseudo (labeled) samples of previously learned domains, and later gets trained on those samples to alleviate forgetting of previous knowledge as it learns the new domain. In addition, a KL divergence loss is minimized to achieve label consistency between the previous and the current model. While adapting to a new task type, LFPT5 includes and tunes additional prompt embeddings for the new task. With extensive experiments, we demonstrate that LFPT5 can be applied to various different types of tasks and significantly outperform previous methods in different LFLL settings.","{'model': 'tldr@v2.0.0', 'text': ""This work defines this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and proposes a unified framework for it based on prompt tuning of T5 called LFPT5, which takes full advantage of PT's strong few-shot learning ability and simultaneously trains the model as a task solver and a data generator.""}",
-Verbal prompting to improve everyday cognition in MCI and unimpaired older adults.,K. Thomas,"OBJECTIVE This study investigated the effect of verbal prompting on elders' 10-year longitudinal change in everyday cognition. Differential effects of prompting associated with impaired cognitive status were also examined. METHOD At baseline, 2,802 participants (mean age = 73.6 years, mean education = 13.5 years) from the Advanced Cognitive Training for Independent and Vital Elderly trial were classified as unimpaired, having amnestic mild cognitive impairment (MCI) or nonamnestic MCI, based on psychometric algorithm. Participants were given the Observed Tasks of Daily Living (OTDL; a behavioral measure with tasks involving medication management/finances/telephone use) at baseline and at 1-, 2-, 3-, 5-, and 10-year follow-ups. When participants said ""I don't know"" or did not respond to an item, they received a standardized verbal prompt. At each occasion, unprompted (sum of items correct without prompting) and prompted (sum of items correct including both prompted and unprompted) scores were derived for each participant. Multilevel modeling, adjusting for demographics/health/training group, was used to determine the trajectories of OTDL performance. RESULTS Persons with MCI performed at lower levels than those who were unimpaired (amnestic < nonamnestic < unimpaired), and for all groups, prompted performance exceeded unprompted in all years. There was differential performance of the prompting conditions over time; prompted performance, unlike unprompted, was relatively protected from age-related decline, and persons with MCI experienced greater improvement due to prompting. CONCLUSIONS Very simple prompting appears to enhance and maintain performance on a task of everyday cognition over 10 years for both unimpaired and mildly impaired older adults.","{'model': 'tldr@v2.0.0', 'text': 'Very simple prompting appears to enhance and maintain performance on a task of everyday cognition over 10 years for both unimpaired and mildly impaired older adults.'}",https://europepmc.org/articles/pmc3935329?pdf=render
-Molecularly Imprinted Polymer Based Sensors for Medical Applications,Yeşeren Saylan,"Sensors have been extensively used owing to multiple advantages, including exceptional sensing performance, user-friendly operation, fast response, high sensitivity and specificity, portability, and real-time analysis. In recent years, efforts in sensor realm have expanded promptly, and it has already presented a broad range of applications in the fields of medical, pharmaceutical and environmental applications, food safety, and homeland security. In particular, molecularly imprinted polymer based sensors have created a fascinating horizon for surface modification techniques by forming specific recognition cavities for template molecules in the polymeric matrix. This method ensures a broad range of versatility to imprint a variety of biomolecules with different size, three dimensional structure, physical and chemical features. In contrast to complex and time-consuming laboratory surface modification methods, molecular imprinting offers a rapid, sensitive, inexpensive, easy-to-use, and highly selective approaches for sensing, and especially for the applications of diagnosis, screening, and theranostics. Due to its physical and chemical robustness, high stability, low-cost, and reusability features, molecularly imprinted polymer based sensors have become very attractive modalities for such applications with a sensitivity of minute structural changes in the structure of biomolecules. This review aims at discussing the principle of molecular imprinting method, the integration of molecularly imprinted polymers with sensing tools, the recent advances and strategies in molecular imprinting methodologies, their applications in medical, and future outlook on this concept.","{'model': 'tldr@v2.0.0', 'text': 'The principle of molecular imprinting method, the integration of molecularly imprinted polymers with sensing tools, the recent advances and strategies in molecular imprinted methodologies, their applications in medical, and future outlook on this concept are discussed.'}",https://www.mdpi.com/1424-8220/19/6/1279/pdf?version=1552557371
-Prompting Reflections for Integrating Self-Regulation into Teacher Technology Education,Tova Michalsky,"Background Technology represents a major topic in educational research. Nevertheless, a gap in the research remains concerning how teachers can bring technology into the classroom. This study focuses on the technological pedagogical content knowledge (TPCK) framework, which aims to consolidate the multidisciplinary professional knowledge related to technology, pedagogy, and content that teachers need so that they can teach and students can learn effectively using technology tools. Purpose The goal of the present study was to investigate the value of modification reflection prompts (“think ahead”) as a complementary reflective framework during the teacher preparatory program, beyond the more traditional judgment reflection prompts (“think back”). In particular, we examined how preservice science teachers may capitalize on learning from modification prompts versus judgment prompts versus both (“think back” and “think ahead”), compared with learning from generic prompts (“stop and think”) based on the IMPROVE model oriented to TPCK. We examined these four treatments’ effects on preservice teachers’ actual design of science lessons and development of their own self-reflection abilities. Participants Participants were 199 first-year preservice science teachers in their preparatory programs at a university in central Israel. Research Design We created a quasi-experimental opportunity for four groups of preservice science teachers to systematically contemplate ready-made TPCK-oriented lesson designs. Each used one of four different reflective methods (the independent variable): modification, judgment, combined modification+judgment, or generic prompts. Then we examined the differential contribution of these treatment methods to the two dependent variables: (1) preservice teachers’ skills for designing actual science lessons and (2) their judgment-type and modification-type self-reflection ability regarding the planning, monitoring, and evaluation phases of their lesson-design process. Data Collection and Analysis Data were scored by coding schemes and were analyzed by multivariate analysis of variance and follow-up analyses of variance with repeated measures. Findings Results indicated that preservice teachers who contemplated a combination of both judgment and modification reflections in treatment improved more in their lesson-design skills and in their self-reflection ability (of both types at the three phases), compared with preservice teachers who contemplated only a single type of reflective prompt (generic or only judgment or modification). Lasting effects (after a semester without the IMPROVE model, prompts, or TPCK focus) revealed that the combined approach continued to significantly outperform the single approaches. Recommendations The current study reinterprets the instructional-reflective framework of teacher education programs to include modification reflection too as a means of developing preservice teachers’ capacity to integrate technology in their lesson designs.",,https://journals.sagepub.com/doi/pdf/10.1177/016146811511700507
-Cold-Start Data Selection for Better Few-shot Language Model Fine-tuning: A Prompt-based Uncertainty Propagation Approach,Yue Yu,"We present PATRON, a prompt-based data selection method for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON will be published upon acceptance.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based data selection method for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available, and a partition-then-rewrite strategy to promote sample diversity when querying for annotations is designed.'}",https://aclanthology.org/2023.acl-long.141.pdf
-Large language models encode clinical knowledge,K. Singhal,,"{'model': 'tldr@v2.0.0', 'text': 'MultiMedQA, a benchmark combining six existing medical question answering datasets spanning professional medicine, research and consumer queries and a new dataset of medical questions searched online, is presented and Med-PaLM, a state-of-the-art large language model for medicine is introduced and evaluated, demonstrating the promise of these models in this domain.'}",https://www.nature.com/articles/s41586-023-06291-2.pdf
-Current State and Future Directions of Technology-Based Ecological Momentary Assessment and Intervention for Major Depressive Disorder: A Systematic Review,D. Colombo,"Ecological momentary assessment (EMA) and ecological momentary intervention (EMI) are alternative approaches to retrospective self-reports and face-to-face treatments, and they make it possible to repeatedly assess patients in naturalistic settings and extend psychological support into real life. The increase in smartphone applications and the availability of low-cost wearable biosensors have further improved the potential of EMA and EMI, which, however, have not yet been applied in clinical practice. Here, we conducted a systematic review, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, to explore the state of the art of technology-based EMA and EMI for major depressive disorder (MDD). A total of 33 articles were included (EMA = 26; EMI = 7). First, we provide a detailed analysis of the included studies from technical (sampling methods, duration, prompts), clinical (fields of application, adherence rates, dropouts, intervention effectiveness), and technological (adopted devices) perspectives. Then, we identify the advantages of using information and communications technologies (ICTs) to extend the potential of these approaches to the understanding, assessment, and intervention in depression. Furthermore, we point out the relevant issues that still need to be addressed within this field, and we discuss how EMA and EMI could benefit from the use of sensors and biosensors, along with recent advances in machine learning for affective modelling.","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the state of the art of technology-based EMA and EMI for major depressive disorder to identify the advantages of using information and communications technologies (ICTs) to extend the potential of these approaches to the understanding, assessment, and intervention in depression.'}",https://www.mdpi.com/2077-0383/8/4/465/pdf?version=1555583492
-Large Language Models Are State-of-the-Art Evaluators of Translation Quality,Tom Kocmi,"We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate seven versions of GPT models, including ChatGPT. We show that our method for translation quality assessment only works with GPT 3.5 and larger models. Comparing to results from WMT22’s Metrics shared task, our method achieves state-of-the-art accuracy in both modes when compared to MQM-based human labels. Our results are valid on the system level for all three WMT22 Metrics shared task language pairs, namely English into German, English into Russian, and Chinese into English. This provides a first glimpse into the usefulness of pre-trained, generative large language models for quality assessment of translations. We publicly release all our code and prompt templates used for the experiments described in this work, as well as all corresponding scoring results, to allow for external validation and reproducibility.","{'model': 'tldr@v2.0.0', 'text': 'GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without, is described, and achieves state-of-the-art accuracy in both modes when compared to MQM-based human labels.'}",http://arxiv.org/pdf/2302.14520
-Teaching Chained Tasks to Students with Intellectual Disabilities by Using Video Prompting in Small Group Instruction.,Ç. Aykut,"AbstractTeaching students with intellectual disabilities in groups presents several difficulties. Use of technology can reduce some of these difficulties. The literature cites several examples of skill acquisition. The purpose of this study is to teach skills to students with intellectual disabilities by using video prompting. A multiple-probe design of single-subject design methods has been applied to this study. Three students with intellectual disabilities participated as subjects in the study. The findings provided evidence for the effectiveness of video prompting in improving teaching skills. The subjects were able to apply and perform these skills away from the teaching environment, and were able to maintain their improved skills for the following 6 months.Key WordsAcquisition of Skills, Intellectual Disabilities, Teaching Skills, Video Prompting.Skills formed by sequential single-staged behaviors comprising complicated end behaviors are called chained tasks (Tekin-?ftar & K?rcaali-?ftar, 2004). Chained tasks are of different types, for example, self-care skills such as eating, toileting, and dressing; gross motor skills such as walking, running, and jumping rope; fine motor skills such as plucking, tearing, cutting, and painting; and daily living skills such as phoning and shopping. These skills, acquired from early childhood on, positively affect both children's school years and future life as an independent individual (Mechling, Pridgen, & Cronin, 2005; Snell & Brown, 1993; Varol, 2005).Modeling is a teaching technique used frequently in teaching new behaviors to individuals with intellectual disabilities; its effectiveness has been supported by numerous studies (Rehfeldt, Dahman, Young, Cherry, & Davis, 2003). Six types of video technology issues have been discussed in the literature: a) video feedback, b) video modeling, c) video self-modeling, d) subjective point of view video modeling, e) video prompting, and f) computer-based video instruction (Mechling, 2005; Murray & Noland, 2013; Oncul & Ozkan, 2010).Video prompting requires the respondent to watch a segment of a video recording and actively respond to the prompt in the video. In contrast to other video teaching methods, video prompting does not require the respondent to watch a video from beginning to the end before performing the behavior. Instead, the respondent is expected to watch a segment of the video and give an immediate active response to the video prompt. Depending on the subjects' response, the video recording is continued or repeated by the researcher the respondent (Mechling, 2005; Norman, Collins, & Schuster, 2001).Chained tasks are more complicated skills consisting of more than one sequential, single-step behaviors. Performance of these tasks takes time. Video prompting appears to be an effective way to enable students with intellectual disabilities to acquire target behaviors (Horn et al., 2008). Thus, the video prompting technique has been used in this study. The literature, though limited, contains studies with experimental designs and a single subject, carried out by using video prompting to allow students to acquire chained tasks. Previous studies have used the video prompting technique to help intellectually disabled or developmentally delayed students acquire daily living skills and self-care skills such as setting a table (Norman et al., 2001), organizing food bought from a green-grocer (Cannella-Malone et al., 2006), cooking (Graves, Collins, Schuster, & Kleinert, 2005), washing clothes (Horn et al., 2008), and washing dishes (Sigafoos et al., 2007). At the end of these studies, the target chained tasks were acquired by these students.Planning the training environment is at least as important as planning the training syllabus for helping students with disabilities acquire new behaviors. In general, one-to-one training for moderately and severely disabled students is effective; however, group instruction is another effective alternative with several advantages over one-to-one teaching: a) teachers can train more than one student, b) group training requires fewer personnel and less teaching time, c) students can earn more functional skills in a less-restrictive environment, d) they can develop appropriate skills for communicating with their peers, and e) they can acquire extra knowledge by observing other students in the group (Collins, Gast, Ault, & Wolery, 1991; Ledford, Gast, Luscre, & Ayres, 2008). …",,
-Visual Prompt Based Personalized Federated Learning,Guang-Ming Li,"As a popular paradigm of distributed learning, personalized federated learning (PFL) allows personalized models to improve generalization ability and robustness by utilizing knowledge from all distributed clients. Most existing PFL algorithms tackle personalization in a model-centric way, such as personalized layer partition, model regularization, and model interpolation, which all fail to take into account the data characteristics of distributed clients. In this paper, we propose a novel PFL framework for image classification tasks, dubbed pFedPT, that leverages personalized visual prompts to implicitly represent local data distribution information of clients and provides that information to the aggregation model to help with classification tasks. Specifically, in each round of pFedPT training, each client generates a local personalized prompt related to local data distribution. Then, the local model is trained on the input composed of raw data and a visual prompt to learn the distribution information contained in the prompt. During model testing, the aggregated model obtains prior knowledge of the data distributions based on the prompts, which can be seen as an adaptive fine-tuning of the aggregation model to improve model performances on different clients. Furthermore, the visual prompt can be added as an orthogonal method to implement personalization on the client for existing FL methods to boost their performance. Experiments on the CIFAR10 and CIFAR100 datasets show that pFedPT outperforms several state-of-the-art (SOTA) PFL algorithms by a large margin in various settings.","{'model': 'tldr@v2.0.0', 'text': 'A novel PFL framework for image classification tasks, dubbed pFedPT, is proposed that leverages personalized visual prompts to implicitly represent local data distribution information of clients and provides that information to the aggregation model to help with classification tasks.'}",http://arxiv.org/pdf/2303.08678
-Memobert: Pre-Training Model with Prompt-Based Learning for Multimodal Emotion Recognition,Jinming Zhao,"Multimodal emotion recognition study is hindered by the lack of labelled corpora in terms of scale and diversity, due to the high annotation cost and label ambiguity. In this paper, we propose a multimodal pre-training model MEmoBERT for multimodal emotion recognition, which learns multimodal joint representations through self-supervised learning from a self-collected large-scale unlabeled video data that come in sheer volume. Furthermore, unlike the conventional ""pre-train, finetune"" paradigm, we propose a prompt-based method that reformulates the downstream emotion classification task as a masked text prediction one, bringing the downstream task closer to the pre-training. Extensive experiments on two benchmark datasets, IEMOCAP and MSP-IMPROV, show that our proposed MEmoBERT significantly enhances emotion recognition performance.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method is proposed that reformulates the downstream emotion classification task as a masked text prediction one, bringing the downstream task closer to the pre-training.'}",https://arxiv.org/pdf/2111.00865
-ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection,Thomas Hartvigsen,"Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language.To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a massive pretrained language model. Controlling machine generation in this way allows ToxiGen to cover implicitly toxic text at a larger scale, and about more demographic groups, than previous resources of human-written text. We conduct a human evaluation on a challenging subset of ToxiGen and find that annotators struggle to distinguish machine-generated text from human-written language. We also find that 94.5% of toxic examples are labeled as hate speech by human annotators. Using three publicly-available datasets, we show that finetuning a toxicity classifier on our data improves its performance on human-written data substantially. We also demonstrate that ToxiGen can be used to fight machine-generated toxicity as finetuning improves the classifier significantly on our evaluation subset.","{'model': 'tldr@v2.0.0', 'text': 'ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups, is created and it is demonstrated that finetuning a toxicity classifier on data improves its performance on human-written data substantially.'}",http://arxiv.org/pdf/2203.09509
-Using Smartphones and Health Apps to Change and Manage Health Behaviors: A Population-Based Survey,C. Ernsting,"Background Chronic conditions are an increasing challenge for individuals and the health care system. Smartphones and health apps are potentially promising tools to change health-related behaviors and manage chronic conditions. Objective The aim of this study was to explore (1) the extent of smartphone and health app use, (2) sociodemographic, medical, and behavioral correlates of smartphone and health app use, and (3) associations of the use of apps and app characteristics with actual health behaviors. Methods A population-based survey (N=4144) among Germans, aged 35 years and older, was conducted. Sociodemographics, presence of chronic conditions, health behaviors, quality of life, and health literacy, as well as the use of the Internet, smartphone, and health apps were assessed by questionnaire at home visit. Binary logistic regression models were applied. Results It was found that 61.25% (2538/4144) of participants used a smartphone. Compared with nonusers, smartphone users were younger, did more research on the Internet, were more likely to work full-time and more likely to have a university degree, engaged more in physical activity, and less in low fat diet, and had a higher health-related quality of life and health literacy. Among smartphone users, 20.53% (521/2538) used health apps. App users were younger, less likely to be native German speakers, did more research on the Internet, were more likely to report chronic conditions, engaged more in physical activity, and low fat diet, and were more health literate compared with nonusers who had a smartphone. Health apps focused on smoking cessation (232/521, 44.5%), healthy diet (201/521, 38.6%), and weight loss (121/521, 23.2%). The most common app characteristics were planning (264/521, 50.7%), reminding (188/521, 36.1%), prompting motivation (179/521 34.4%), and the provision of information (175/521, 33.6%). Significant associations were found between planning and the health behavior physical activity, between feedback or monitoring and physical activity, and between feedback or monitoring and adherence to doctor’s advice. Conclusions Although there were many smartphone and health app users, a substantial proportion of the population was not engaged. Findings suggest age-related, socioeconomic-related, literacy-related, and health-related disparities in the use of mobile technologies. Health app use may reflect a user’s motivation to change or maintain health behaviors. App developers and researchers should take account of the needs of older people, people with low health literacy, and chronic conditions.","{'model': 'tldr@v2.0.0', 'text': 'App users were younger, less likely to be native German speakers, did more research on the Internet, were more likely to report chronic conditions, engaged more in physical activity, and low fat diet, and were more health literate compared with nonusers who had a smartphone.'}",https://www.jmir.org/2017/4/e101/PDF
-Prompt-based Text Entailment for Low-Resource Named Entity Recognition,Dongfang Li,"Pre-trained Language Models (PLMs) have been applied in NLP tasks and achieve promising results. Nevertheless, the fine-tuning procedure needs labeled data of the target domain, making it difficult to learn in low-resource and non-trivial labeled scenarios. To address these challenges, we propose Prompt-based Text Entailment (PTE) for low-resource named entity recognition, which better leverages knowledge in the PLMs. We first reformulate named entity recognition as the text entailment task. The original sentence with entity type-specific prompts is fed into PLMs to get entailment scores for each candidate. The entity type with the top score is then selected as final label. Then, we inject tagging labels into prompts and treat words as basic units instead of n-gram spans to reduce time complexity in generating candidates by n-grams enumeration. Experimental results demonstrate that the proposed method PTE achieves competitive performance on the CoNLL03 dataset, and better than fine-tuned counterparts on the MIT Movie and Few-NERD dataset in low-resource settings.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2211.03039
-ConsPrompt: Easily Exploiting Contrastive Samples for Few-shot Prompt Learning,Jinta Weng,"Prompt learning recently become an effective linguistic tool to motivate the PLMs’ knowledge on few-shot-setting tasks. However, studies have shown the lack of robustness still exists in prompt learning, since suitable initialization of continuous prompt and expert-first manual prompt are essential in fine-tuning process. What is more, human also utilize their comparative ability to motivate their existing knowledge for distinguishing different examples. Motivated by this, we explore how to use contrastive samples to strengthen prompt learning. In detail, we first propose our model ConsPrompt combining with prompt encoding network, contrastive sampling module, and contrastive scoring module. Subsequently, two sampling strategies, similarity-based and label-based strategies, are introduced to realize dif-ferential contrastive learning. The effectiveness of proposed ConsPrompt is demonstrated in five different few-shot learning tasks and shown the similarity-based sampling strategy is more effective than label-based in combining contrastive learning. Our results also ex-hibits the state-of-the-art performance and robustness in different few-shot settings, which proves that the ConsPrompt could be assumed as a better knowledge probe to motivate PLMs. As far as we could reach, this is the first work exploring how to use contrastive learning approach and suitable contrastive samples to enhance prompt-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'The model ConsPrompt is proposed, which ex-hibits the state-of-the-art performance and robustness in different few-shot settings, which proves that the Cons Prompt could be assumed as a better knowledge probe to motivate PLMs.'}",https://arxiv.org/pdf/2211.04118
-Towards Informative Few-Shot Prompt with Maximum Information Gain for In-Context Learning,Hongfu Liu,"Large Language models (LLMs) possess the capability to engage In-context Learning (ICL) by leveraging a few demonstrations pertaining to a new downstream task as conditions. However, this particular learning paradigm suffers from high instability stemming from substantial variances induced by factors such as the input distribution of selected examples, their ordering, and prompt formats. In this work, we demonstrate that even when all these factors are held constant, the random selection of examples still results in high variance. Consequently, we aim to explore the informative ability of data examples by quantifying the Information Gain (IG) obtained in prediction after observing a given example candidate. Then we propose to sample those with maximum IG. Additionally, we identify the presence of template bias, which can lead to unfair evaluations of IG during the sampling process. To mitigate this bias, we introduce Calibration Before Sampling strategy. The experimental results illustrate that our proposed method can yield an average relative improvement of 14.3% across six classification tasks using three LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to explore the informative ability of data examples by quantifying the Information Gain obtained in prediction after observing a given example candidate, and proposes to sample those with maximum IG.'}",https://arxiv.org/pdf/2310.08923
-Virtual prompt pre-training for prototype-based few-shot relation extraction,Kai He,,,http://manuscript.elsevier.com/S0957417422019455/pdf/S0957417422019455.pdf
-Inverse is Better! Fast and Accurate Prompt for Few-shot Slot Tagging,Yutai Hou,"Prompting methods recently achieve impressive success in few-shot learning. These methods modify input samples with prompt sentence pieces, and decode label tokens to map samples to corresponding labels. However, such a paradigm is very inefficient for the task of slot tagging. Since slot tagging samples are multiple consecutive words in a sentence, the prompting methods have to enumerate all n-grams token spans to find all the possible slots, which greatly slows down the prediction. To tackle this, we introduce an inverse paradigm for prompting. Different from the classic prompts mapping tokens to labels, we reversely predict slot values given slot types. Such inverse prompting only requires a one-turn prediction for each slot type and greatly speeds up the prediction. Besides, we propose a novel Iterative Prediction Strategy, from which the model learns to refine predictions by considering the relations between different slot types. We find, somewhat surprisingly, the proposed method not only predicts faster but also significantly improves the effect (improve over 6.1 F1-scores on 10-shot setting) and achieves new state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces an inverse paradigm for prompting, different from the classic prompts mapping tokens to labels, that reversely predict slot values given slot types and proposes a novel Iterative Prediction Strategy, from which the model learns to refine predictions by considering the relations between different slot types.'}",http://arxiv.org/pdf/2204.00885
-Few-Shot Stance Detection via Target-Aware Prompt Distillation,Yan Jiang,"Stance detection aims to identify whether the author of a text is in favor of, against, or neutral to a given target. The main challenge of this task comes two-fold: few-shot learning resulting from the varying targets and the lack of contextual information of the targets. Existing works mainly focus on solving the second issue by designing attention-based models or introducing noisy external knowledge, while the first issue remains under-explored. In this paper, inspired by the potential capability of pre-trained language models (PLMs) serving as knowledge bases and few-shot learners, we propose to introduce prompt-based fine-tuning for stance detection. PLMs can provide essential contextual information for the targets and enable few-shot learning via prompts. Considering the crucial role of the target in stance detection task, we design target-aware prompts and propose a novel verbalizer. Instead of mapping each label to a concrete word, our verbalizer maps each label to a vector and picks the label that best captures the correlation between the stance and the target. Moreover, to alleviate the possible defect of dealing with varying targets with a single hand-crafted prompt, we propose to distill the information learned from multiple prompts. Experimental results show the superior performance of our proposed model in both full-data and few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs target-aware prompts and proposes a novel verbalizer that distill the information learned from multiple prompts in stance detection, inspired by the potential capability of pre-trained language models serving as knowledge bases and few-shot learners.'}",https://dl.acm.org/doi/pdf/10.1145/3477495.3531979
-Few-Shot Multi-Modal Sentiment Analysis with Prompt-Based Vision-Aware Language Modeling,Yang Yu,"As a hot study topic in natural language processing, affec-tive computing and multimedia analysis, multi-modal senti-ment analysis (MSA) is widely explored on aspect-level and sentence-level tasks. However, the existing studies normally rely on a lot of annotated multi-modal data, which are difficult to collect due to the massive expenditure of manpower and re-sources, especially in some open-ended and fine-grained do-mains. Therefore, it is necessary to investigate the few-shot scenario for MSA. In this paper, we propose a prompt-based vision-aware language modeling (PVLM) approach to MSA, which only requires a few supervised data. Specifically, our PVLM can incorporate the visual information into pre-trained language model and leverage prompt tuning to bridge the gap between masked language prediction in pre-training and MSA tasks. Systematic experiments on three aspect-level and two sentence-level datasets of MSA demonstrate the effectiveness of our few-shot approach.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based vision-aware language modeling (PVLM) approach to MSA, which only requires a few supervised data and can incorporate the visual information into pre-trained language model and leverage prompt tuning to bridge the gap between masked language prediction in pre-training and MSA tasks.'}",
-Unified Multi-modal Pre-training for Few-shot Sentiment Analysis with Prompt-based Learning,Yang Yu,"Multi-modal sentiment analysis (MSA) has become more and more attractive in both academia and industry. The conventional studies normally require massive labeled data to train the deep neural models. To alleviate the above issue, in this paper, we conduct few-shot MSA with quite a small number of labeled samples. Inspired by the success of textual prompt-based fine-tuning (PF) approaches in few-shot scenario, we introduce a multi-modal prompt-based fine-tuning (MPF) approach. To narrow the semantic gap between language and vision, we propose unified pre-training for multi-modal prompt-based fine-tuning (UP-MPF) with two stages. First, in unified pre-training stage, we employ a simple and effective task to obtain coherent vision-language representations from fixed pre-trained language models (PLMs), i.e., predicting the rotation direction of the input image with a prompt phrase as input concurrently. Second, in multi-modal prompt-based fine-tuning, we freeze the visual encoder to reduce more parameters, which further facilitates few-shot MSA. Extensive experiments and analysis on three coarse-grained and three fine-grained MSA datasets demonstrate the better performance of our UP-MPF against the state-of-the-art of PF, MSA, and multi-modal pre-training approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes unified pre-training for multi-modal prompt-based fine-tuning (UP-MPF) with two stages, and employs a simple and effective task to obtain coherent vision-language representations from fixed pre-trained language models (PLMs).'}",
-Commonsense Knowledge-Aware Prompt Tuning for Few-Shot NOTA Relation Classification,Bo Lv,"Compared with the traditional few-shot task, the few-shot none-of-the-above (NOTA) relation classification focuses on the realistic scenario of few-shot learning, in which a test instance might not belong to any of the target categories. This undoubtedly increases the task’s difficulty because given only a few support samples, this cannot represent the distribution of NOTA categories in space. The model needs to make full use of the syntactic information and word meaning information learned in the pre-training stage to distinguish the NOTA category and the support sample category in the embedding space. However, previous fine-tuning methods mainly focus on optimizing the extra classifiers (on top of pre-trained language models (PLMs)) and neglect the connection between pre-training objectives and downstream tasks. In this paper, we propose the commonsense knowledge-aware prompt tuning (CKPT) method for a few-shot NOTA relation classification task. First, a simple and effective prompt-learning method is developed by constructing relation-oriented templates, which can further stimulate the rich knowledge distributed in PLMs to better serve downstream tasks. Second, external knowledge is incorporated into the model by a label-extension operation, which forms knowledgeable prompt tuning to improve and stabilize prompt tuning. Third, to distinguish the NOTA pairs and positive pairs in embedding space more accurately, a learned scoring strategy is proposed, which introduces a learned threshold classification function and improves the loss function by adding a new term focused on NOTA identification. Experiments on two widely used benchmarks (FewRel 2.0 and Few-shot TACRED) show that our method is a simple and effective framework, and a new state of the art is established in the few-shot classification field.","{'model': 'tldr@v2.0.0', 'text': 'The commonsense knowledge-aware prompt tuning (CKPT) method is proposed, a simple and effective prompt-learning method is developed by constructing relation-oriented templates, which can further stimulate the rich knowledge distributed in PLMs to better serve downstream tasks.'}",https://www.mdpi.com/2076-3417/12/4/2185/pdf?version=1645269904
-Dual Context-Guided Continuous Prompt Tuning for Few-Shot Learning,Jie Zhou,"Prompt-based paradigm has shown its competitive performance in many NLP tasks. However, its success heavily depends on prompt design, and the effectiveness varies upon the model and training data. In this paper, we propose a novel dual context-guided continuous prompt (DCCP) tuning method. To explore the rich contextual information in language structure and close the gap between discrete prompt tuning and continuous prompt tuning, DCCP introduces two auxiliary training objectives and constructs input in a pair-wise fashion.Experimental results demonstrate that our method is applicable to many NLP tasks, and can often outperform existing prompt tuning methods by a large margin in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'To explore the rich contextual information in language structure and close the gap between discrete prompt tuning and continuous prompt tuning, DCCP introduces two auxiliary training objectives and constructs input in a pair-wise fashion.'}",https://aclanthology.org/2022.findings-acl.8.pdf
-Better Few-Shot Relation Extraction with Label Prompt Dropout,Peiyuan Zhang,"Few-shot relation extraction aims to learn to identify the relation between two entities based on very limited training examples. Recent efforts found that textual labels (i.e., relation names and relation descriptions) could be extremely useful for learning class representations, which will benefit the few-shot learning task. However, what is the best way to leverage such label information in the learning process is an important research question. Existing works largely assume such textual labels are always present during both learning and prediction. In this work, we argue that such approaches may not always lead to optimal results. Instead, we present a novel approach called label prompt dropout, which randomly removes label descriptions in the learning process. Our experiments show that our approach is able to lead to improved class representations, yielding significantly better results on the few-shot relation extraction task.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a novel approach called label prompt dropout, which randomly removes label descriptions in the learning process, and shows that this approach is able to lead to improved class representations, yielding significantly better results on the few-shot relation extraction task.'}",http://arxiv.org/pdf/2210.13733
-A Dual Prompt Learning Framework for Few-Shot Dialogue State Tracking,Yuting Yang,"Dialogue State Tracking (DST) module is an essential component of task-oriented dialog systems to understand users’ goals and needs. Collecting dialogue state labels including slots and values can be costly, requiring experts to annotate all (slot, value) information for each turn in dialogues. It is also difficult to define all possible slots and values in advance, especially with the wide application of dialogue systems in more and more new-rising applications. In this paper, we focus on improving DST module to generate dialogue states in circumstances with limited annotations and knowledge about slot ontology. To this end, we design a dual prompt learning framework for few-shot DST. The dual framework aims to explore how to utilize the language understanding and generation capabilities of pre-trained language models for DST efficiently. Specifically, we consider the learning of slot generation and value generation as dual tasks, and two kinds of prompts are designed based on this dual structure to incorporate task-related knowledge of these two tasks respectively. In this way, the DST task can be formulated as a language modeling task efficiently under few-shot settings. To evaluate the proposed framework, we conduct experiments on two task-oriented dialogue datasets. The results demonstrate that the proposed method not only outperforms existing state-of-the-art few-shot methods, but also can generate unseen slots. It indicates that DST-related knowledge can be probed from pre-trained language models and utilized to address low-resource DST efficiently with the help of prompt learning.","{'model': 'tldr@v2.0.0', 'text': 'A dual prompt learning framework for few-shot DST is designed that indicates that DST-related knowledge can be probed from pre-trained language models and utilized to address low-resource DST efficiently with the help of prompt learning.'}",https://dl.acm.org/doi/pdf/10.1145/3543507.3583238
-Multi-Task Pre-Training of Modular Prompt for Few-Shot Learning,Tianxiang Sun,"Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present M ulti-task P re-trained M odular P rompt ( MP 2 ) to boost prompt tuning for few-shot learning. MP 2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP 2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP 2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents M ulti-task P re-trained M odular P rompt ( MP 2) to boost prompt tuning for few-shot learning, and demonstrates that MP 2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.'}",http://arxiv.org/pdf/2210.07565
-Multitask Pre-training of Modular Prompt for Chinese Few-Shot Learning,Tianxiang Sun,"Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning. MP2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning and demonstrates that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.'}",https://aclanthology.org/2023.acl-long.625.pdf
-Visual Prompt Tuning for Few-Shot Text Classification,Jing Wen,"Deploying large-scale pre-trained models in the prompt-tuning paradigm has demonstrated promising performance in few-shot learning. Particularly, vision-language pre-training models (VL-PTMs) have been intensively explored in various few-shot downstream tasks. However, most existing works only apply VL-PTMs to visual tasks like image classification, with few attempts being made on language tasks like text classification. In few-shot text classification, a feasible paradigm for deploying VL-PTMs is to align the input samples and their category names via the text encoders. However, it leads to the waste of visual information learned by the image encoders of VL-PTMs. To overcome this drawback, we propose a novel method named Visual Prompt Tuning (VPT). To our best knowledge, this method is the first attempt to deploy VL-PTM in few-shot text classification task. The main idea is to generate the image embeddings w.r.t. category names as visual prompt and then add them to the aligning process. Extensive experiments show that our VPT can achieve significant improvements under both zero-shot and few-shot settings. Importantly, our VPT even outperforms the most recent prompt-tuning methods on five public text classification datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel method named Visual Prompt Tuning (VPT), which is the first attempt to deploy VL-PTM in few-shot text classification task and outperforms the most recent prompt-tuning methods on five public text classification datasets.'}",
-IDIAPers @ Causal News Corpus 2022: Efficient Causal Relation Identification Through a Prompt-based Few-shot Approach,S. Burdisso,"In this paper, we describe our participation in the subtask 1 of CASE-2022, Event Causality Identification with Casual News Corpus. We address the Causal Relation Identification (CRI) task by exploiting a set of simple yet complementary techniques for fine-tuning language models (LMs) on a few annotated examples (i.e., a few-shot configuration).We follow a prompt-based prediction approach for fine-tuning LMs in which the CRI task is treated as a masked language modeling problem (MLM). This approach allows LMs natively pre-trained on MLM tasks to directly generate textual responses to CRI-specific prompts.We compare the performance of this method against ensemble techniques trained on the entire dataset.Our best-performing submission was fine-tuned with only 256 instances per class, 15.7% of the all available data, and yet obtained the second-best precision (0.82), third-best accuracy (0.82), and an F1-score (0.85) very close to what was reported by the winner team (0.86).","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the Causal Relation Identification (CRI) task by exploiting a set of simple yet complementary techniques for fine-tuning language models (LMs) on a few annotated examples (i.e., a few-shot configuration) in which the CRI task is treated as a masked language modeling problem (MLM).'}",http://arxiv.org/pdf/2209.03895
-P4E: Few-Shot Event Detection as Prompt-Guided Identification and Localization,Sha Li,"We propose P4E, an identify-and-localize event detection framework that integrates the best of few-shot prompting and structured prediction. Our framework decomposes event detection into an identification task and a localization task. For the identification task, which we formulate as multi-label classification, we leverage cloze-based prompting to align our objective with the pre-training task of language models, allowing our model to quickly adapt to new event types. We then employ an event type-agnostic sequence labeling model to localize the event trigger conditioned on the identification output. This heterogeneous model design allows P4E to quickly learn new event types without sacrificing the ability to make structured predictions. Our experiments demonstrate the effectiveness of our proposed design, and P4E shows superior performance for few-shot event detection on benchmark datasets FewEvent and MAVEN and comparable performance to SOTA for fully-supervised event detection on ACE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes P4E, an identify-and-localize event detection framework that integrates the best of few-shot prompting and structured prediction, and shows superior performance for few- shot event detection on benchmark datasets FewEvent and MAVEN and comparable performance to SOTA for fully-supervised event Detection on ACE.'}",
-Few-Shot Natural Language Inference Generation with PDD: Prompt and Dynamic Demonstration,Kaijian Li,"Natural Language Inference Generation task is to generate a text hypothesis given a text premise and a logical relation between the two. This task can be used in data augmentation and controllable text generation in practice. In this paper, we propose language models with prompt and dynamic demonstration (LM-PDD) to tackle this problem in few-shot settings. Our framework outperforms standard fine-tuned models with low resource, achieving an average 8% absolute improvement on SNLI and MNLI datasets, and the results on 13 natural language classification tasks also show that our dynamic demonstration method has good generalizability.","{'model': 'tldr@v2.0.0', 'text': ""This paper's framework outperforms standard fine-tuned models with low resource, achieving an average 8% absolute improvement on SNLI and MNLI datasets, and the results on 13 natural language classification tasks show that the dynamic demonstration method has good generalizability.""}",https://arxiv.org/pdf/2205.10593
-Discriminative Language Model as Semantic Consistency Scorer for Prompt-based Few-Shot Text Classification,Zhipeng Xie,"This paper proposes a novel prompt-based finetuning method (called DLM-SCS) for few-shot text classification by utilizing the discriminative language model ELECTRA that is pretrained to distinguish whether a token is original or generated. The underlying idea is that the prompt instantiated with the true label should have higher semantic consistency score than other prompts with false labels. Since a prompt usually consists of several components (or parts), its semantic consistency can be decomposed accordingly. The semantic consistency of each component is then computed by making use of the pretrained ELECTRA model, without introducing extra parameters. Extensive experiments have shown that our model outperforms several state-of-the-art prompt-based few-shot methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based finetuning method for few-shot text classification by utilizing the discriminative language model ELECTRA that is pretrained to distinguish whether a token is original or generated is proposed.'}",http://arxiv.org/pdf/2210.12763
-DFS-NER: Description Enhanced Few-shot NER via Prompt Learning and Meta-Learning,Huinan Huang,"Named Entity Recognition (NER) is a very common task in many social good related domains. Recently, deep learning based NER has gradually matured, but still faces the scarcity problem of labeled data in specific domains. Therefore, researchers focus on few-shot NER to reduce the model’s data dependence and enhance the transferability of the model. However, existing works usually cannot adapt to new entity types and are prone to the so-called negative transfer problem. Therefore, in this paper we propose a type- Description-enhanced Few Shot NER model, called DFS-NER, which effectively integrates the prompt learning paradigm and the meta-learning framework. DFS-NER performs well under frozen pre-training model parameters through continuous templates. It realizes efficient source domain training and target domain parameter fine-tuning through the metalearning framework. We enhance the robustness of entity- type prototype representations by introducing word-word- level and word-type-level contrastive learning objectives and capsule networks as the induction module. Simultaneously, based on discrete prompt learning, a masked-language model learning objective guided by type description is proposed, which can well absorb the semantic information of entity types. Experiments on commonly used datasets, including, SNIPS, Few-NERD, and MIT Movie show that DFS-NER basically surpasses baseline models and achieves the state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a type- Description-enhanced Few Shot NER model, called DFS-NER, which effectively integrates the prompt learning paradigm and the meta-learning framework and proposes a masked-language model learning objective guided by type description, which can well absorb the semantic information of entity types.'}",
-A Prompt-based Few-shot Machine Reading Comprehension Model for Intelligent Bridge Management,Luyi Zhang,"Bridge inspection reports are an important data source in the bridge management process, and they contain a large amount of fine-grained information. However, the research on machine reading comprehension (MRC) methods for this field is insufficient, and annotating large scale domain-specific corpus is time-consuming. This paper presented a novel prompt-based few-shot MRC approach for intelligent bridge management. The proposed model uses the pretrained model MacBERT as backbone. The prompt templates are designed based on some domain-specific heuristic rules. The experimental results show that our model outperforms the baseline models in different few-shot settings. The proposed model can provide technical support for the construction of automatic question answering system in the field of bridge management.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based few-shot MRC approach for intelligent bridge management that uses the pretrained model MacBERT as backbone and prompt templates are designed based on some domain-specific heuristic rules.'}",
-A Study on Prompt-based Few-Shot Learning Methods for Belief State Tracking in Task-oriented Dialog Systems,Debjoy Saha,"We tackle the Dialogue Belief State Tracking(DST) problem of task-oriented conversational systems. Recent approaches to this problem leveraging Transformer-based models have yielded great results. However, training these models is expensive, both in terms of computational resources and time. Additionally, collecting high quality annotated dialogue datasets remains a challenge for researchers because of the extensive annotation required for training these models. Driven by the recent success of pre-trained language models and prompt-based learning, we explore prompt-based few-shot learning for Dialogue Belief State Tracking. We formulate the DST problem as a 2-stage prompt-based language modelling task and train language models for both tasks and present a comprehensive empirical analysis of their separate and joint performance. We demonstrate the potential of prompt-based methods in few-shot learning for DST and provide directions for future improvement.","{'model': 'tldr@v2.0.0', 'text': 'The DST problem is formulated as a 2-stage prompt-based language modelling task and language models for both tasks are trained and a comprehensive empirical analysis of their separate and joint performance is presented.'}",http://arxiv.org/pdf/2204.08167
-Prompt and Contrastive Learning for Few-shot Sentiment Classification,Fei Wang,"Sentiment classification is a hot topic in the field of natural language processing. Currently, state-of-the-art classification models follow two steps: pre-training a large language model on upstream tasks, and then using human-labeled data to fine-tune a task-related model. However, there is a large gap between the upstream tasks of the pre-trained model and the downstream tasks being performed, resulting in the need for more labeled data to achieve excellent performance. Manually annotating data is expensive. In this paper, we propose a few-shot sentiment classification method based on Prompt and Contrastive Learning (PCL), which can significantly improve the performance of large-scale pre-trained language models in low-data and high-data regimes. Prompt learning aims to alleviate the gap between upstream and downstream tasks, and the contrastive learning is designed to capture the inter-class and intra-class distribution patterns of labeled data. Thanks to the integration of the two strategies, PCL markedly exceeds baselines with low resources. Extensive experiments on three datasets show that our method has outstanding performance in the few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a few-shot sentiment classification method based on Prompt and Contrastive Learning (PCL), which can significantly improve the performance of large-scale pre-trained language models in low-data and high-data regimes.'}",
-Multi-label Few-shot ICD Coding as Autoregressive Generation with Prompt,Zhichao Yang,"Automatic International Classification of Diseases (ICD) coding aims to assign multiple ICD codes to a medical note with an average of 3,000+ tokens. This task is challenging due to the high-dimensional space of multi-label assignment (155,000+ ICD code candidates) and the long-tail challenge - Many ICD codes are infrequently assigned yet infrequent ICD codes are important clinically. This study addresses the long-tail challenge by transforming this multi-label classification task into an autoregressive generation task. Specifically, we first introduce a novel pretraining objective to generate free text diagnoses and procedures using the SOAP structure, the medical logic physicians use for note documentation. Second, instead of directly predicting the high dimensional space of ICD codes, our model generates the lower dimension of text descriptions, which then infers ICD codes. Third, we designed a novel prompt template for multi-label classification. We evaluate our Generation with Prompt (GPsoap) model with the benchmark of all code assignment (MIMIC-III-full) and few shot ICD code assignment evaluation benchmark (MIMIC-III-few). Experiments on MIMIC-III-few show that our model performs with a marco F130.2, which substantially outperforms the previous MIMIC-III-full SOTA model (marco F1 4.3) and the model specifically designed for few/zero shot setting (marco F1 18.7). Finally, we design a novel ensemble learner, a cross-attention reranker with prompts, to integrate previous SOTA and our best few-shot coding predictions. Experiments on MIMIC-III-full show that our ensemble learner substantially improves both macro and micro F1, from 10.4 to 14.6 and from 58.2 to 59.1, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This study introduces a novel pretraining objective to generate free text diagnoses and procedures using the SOAP structure, the medical logic physicians use for note documentation, and designs a novel prompt template for multi-label classification.'}",https://arxiv.org/pdf/2211.13813
-TI-Prompt: Towards a Prompt Tuning Method for Few-shot Threat Intelligence Twitter Classification*,Yizhe You,"Obtaining the latest Threat Intelligence (TI) via Twitter has become one of the most important methods for defenders to catch up with emerging cyber threats. Existing TI Twitter classification works mainly based on supervised learning methods. Such approaches require large amounts of annotated data and are difficult to be transferred to other TI Twitter classification tasks. This paper proposes a prompt-based method for classifying TI on Twitter, named TI-Prompt. TI-Prompt lever-ages the prompt-tuning method with two templates in different TI Twitter classification tasks. TI-Prompt also uses a semantic similarity-based approach to automatically enrich the prompt verbalizer without expert knowledge and a verbalizer refinement method to calibrate the verbalizer based on the training data. We evaluate TI-Prompt with binary and multi-classification tasks on two Twitter Threat Intelligence datasets. Evaluation results show that the proposed TI-Prompt improves 5-10% over the best performance of previous supervised learning methods under the few-shot settings. Compared to the general prompt-tuning methods, the proposed prompt-tuning templates can also improve the classification performance by 2–5%. Meanwhile, the proposed verbalizer enrichment method and refinement method improve classification accuracy by 1–4% compared with the general single-word verbalizer prompt method. Therefore, TI-Prompt can be extended to other Threat Intelligence classification tasks without requiring large amounts of training data, significantly reducing the annotation cost.","{'model': 'tldr@v2.0.0', 'text': 'Evaluation results show that the proposed TI-Prompt improves 5-10% over the best performance of previous supervised learning methods under the few-shot settings, and can be extended to other Threat Intelligence classification tasks without requiring large amounts of training data, significantly reducing the annotation cost.'}",
-Adaptive Prompt Learning-Based Few-Shot Sentiment Analysis,Pengfei Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results on FewCLUE datasets demonstrate that the proposed method AP can effectively construct appropriate adaptive prompt regardless of the quality of hand-crafted prompt and outperform the state-of-the-art baselines.'}",https://arxiv.org/pdf/2205.07220
-Calibrate Before Use: Improving Few-Shot Performance of Language Models,Tony Zhao,"GPT-3 can perform numerous tasks when provided a natural language prompt that contains a few training examples. We show that this type of few-shot learning can be unstable: the choice of prompt format, training examples, and even the order of the training examples can cause accuracy to vary from near chance to near state-of-the-art. We demonstrate that this instability arises from the bias of language models towards predicting certain answers, e.g., those that are placed near the end of the prompt or are common in the pre-training data. To mitigate this, we first estimate the model's bias towards each answer by asking for its prediction when given the training prompt and a content-free test input such as ""N/A"". We then fit calibration parameters that cause the prediction for this input to be uniform across answers. On a diverse set of tasks, this contextual calibration procedure substantially improves GPT-3 and GPT-2's average accuracy (up to 30.0% absolute) and reduces variance across different choices of the prompt.","{'model': 'tldr@v2.0.0', 'text': 'This work first estimates the model\'s bias towards each answer by asking for its prediction when given the training prompt and a content-free test input such as ""N/A"", and then fits calibration parameters that cause the prediction for this input to be uniform across answers.'}",
-AugPrompt: Knowledgeable augmented-trigger prompt for few-shot event classification,Chengyu Song,,,
-Prompt-based Few-shot Learning for Table-based Fact Verification,Lei Hou,"Natural language processing has been a hot topic of research, but existing research is mainly limited to unstructured information such as natural language sentences and documents, and less research has been done on structured information such as tables. The main object of this paper is the table-based fact verification task, under which there is only one TABFACT dataset. Most of the existing methods on this dataset are based on pre-trained models and need to be fine-tuned again if a new dataset appears. And some previous work on natural language sentences has shown that prompt approach can achieve good performance with few samples. Therefore, in this paper, we adopt the prompt approach for experiments on the table fact detection task by manually designing templates for hinting the pre-trained model. Meanwhile, to enhance the generalization of the model, we introduce a multi-pair mapping relationship in the Answer Engineering phase. Experiments on the TABFACT dataset show that using the prompt method for table-based fact verification task in the case of few samples can be effective, providing a new way for optimizing table-related tasks in the case of few samples.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on the TABFACT dataset show that using the prompt method for table-based fact verification task in the case of few samples can be effective, providing a new way for optimizing table-related tasks in the Case ofFew samples.'}",
-Investigating Prompt Learning for Chinese Few-Shot Text Classification with Pre-Trained Language Models,Chengyu Song,"Text classification aims to assign predefined labels to unlabeled sentences, which tend to struggle in real-world applications when only a few annotated samples are available. Previous works generally focus on using the paradigm of meta-learning to overcome the classification difficulties brought by insufficient data, where a set of auxiliary tasks is given. Accordingly, prompt-based approaches are proposed to deal with the low-resource issue. However, existing prompt-based methods mainly focus on English tasks, which generally apply English pretrained language models that can not directly adapt to Chinese tasks due to structural and grammatical differences. Thus, we propose a prompt-based Chinese text classification framework that uses generated natural language sequences as hints, which can alleviate the classification bottleneck well in low-resource scenarios. In detail, we first design a prompt-based fine-tuning together with a novel pipeline for automating prompt generation in Chinese. Then, we propose a refined strategy for dynamically and selectively incorporating demonstrations into each context. We present a systematic evaluation for analyzing few-shot performance on a wide range of Chinese text classification tasks. Our approach makes few assumptions about task resources and expertise and therefore constitutes a powerful, task-independent approach for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based Chinese text classification framework that uses generated natural language sequences as hints, which can alleviate the classification bottleneck well in low-resource scenarios and constitutes a powerful, task-independent approach for few-shot learning.'}",https://www.mdpi.com/2076-3417/12/21/11117/pdf?version=1667385041
-"Few-shot Information Extraction is Here: Pre-train, Prompt and Entail",Eneko Agirre,"Deep Learning has made tremendous progress in Natural Language Processing (NLP), where large pre-trained language models (PLM) fine-tuned on the target task have become the predominant tool. More recently, in a process called prompting, NLP tasks are rephrased as natural language text, allowing us to better exploit linguistic knowledge learned by PLMs and resulting in significant improvements. Still, PLMs have limited inference ability. In the Textual Entailment task, systems need to output whether the truth of a certain textual hypothesis follows from the given premise text. Manually annotated entailment datasets covering multiple inference phenomena have been used to infuse inference capabilities to PLMs. This talk will review these recent developments, and will present an approach that combines prompts and PLMs fine-tuned for textual entailment that yields state-of-the-art results on Information Extraction (IE) using only a small fraction of the annotations. The approach has additional benefits, like the ability to learn from different schemas and inference datasets. These developments enable a new paradigm for IE where the expert can define the domain-specific schema using natural language and directly run those specifications, annotating a handful of examples in the process. A user interface based on this new paradigm will also be presented. Beyond IE, inference capabilities could be extended, acquired and applied from other tasks, opening a new research avenue where entailment and downstream task performance improve in tandem.","{'model': 'tldr@v2.0.0', 'text': 'This talk will present an approach that combines prompts and PLMs fine-tuned for textual entailment that yields state-of-the-art results on Information Extraction (IE) using only a small fraction of the annotations.'}",
-Prompt-Based Self-training Framework for Few-Shot Named Entity Recognition,Ganghong Huang,,,
-A Prompt-based Few-shot Learning Approach to Software Conflict Detection,Robert K. Helmeczi,"A software requirement specification (SRS) document is an essential part of the software development life cycle which outlines the requirements that a software program in development must satisfy. This document is often specified by a diverse group of stakeholders and is subject to continual change, making the process of maintaining the document and detecting conflicts between requirements an essential task in software development. Notably, projects that do not address conflicts in the SRS document early on face considerable problems later in the development life cycle. These problems incur substantial costs in terms of time and money, and these costs often become insurmountable barriers that ultimately result in the termination of a software project altogether. As a result, early detection of SRS conflicts is critical to project sustainability. The conflict detection task is approached in numerous ways, many of which require a significant amount of manual intervention from developers, or require access to a large amount of labeled, task-specific training data. In this work, we propose using a prompt-based learning approach to perform few-shot learning for conflict detection. We compare our results to supervised learning approaches that use pretrained language models, such as BERT and its variants. Our results show that prompting with just 32 labeled examples can achieve a similar level of performance in many key metrics to that of supervised learning on training sets that are magnitudes larger in size. In contrast to many other conflict detection approaches, we make no assumptions about the type of underlying requirements, allowing us to analyze pairings of both functional and non-functional requirements. This allows us to omit the potentially expensive task of filtering out non-functional requirements from our dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using a prompt-based learning approach to perform few-shot learning for conflict detection and shows that prompting with just 32 labeled examples can achieve a similar level of performance to that of supervised learning on training sets that are magnitudes larger in size.'}",https://arxiv.org/pdf/2211.02709
-Instance-Guided Prompt Learning for Few-Shot Text Matching,Jiacheng Du,,,https://aclanthology.org/2022.findings-emnlp.285.pdf
-Towards Unified Prompt Tuning for Few-shot Learning,,"Prompt-based fine-tuning has boosted the per- 001 formance of Pre-trained Language Models 002 (PLMs) on few-shot learning by employing 003 task-specific prompts. However, PLMs are 004 unfamiliar with the prompt-style expressions 005 during pre-training, which limits the few-shot 006 learning performance on downstream tasks. 007 It would be desirable if models can acquire 008 some prompting knowledge before task adap- 009 tation. We present the Unified Prompt Tun- 010 ing ( UPT ) framework, leading to better few- 011 shot learning for BERT-style models by ex- 012 plicitly capturing prompting semantics from 013 non-target NLP datasets. In UPT , a novel 014 paradigm Prompt-Options-Verbalizer is pro- 015 posed for joint prompt learning across differ- 016 ent NLP tasks, forcing PLMs to capture task- 017 invariant prompting knowledge. We further de- 018 sign a self-supervised task named Knowledge- 019 enhanced Selective Masked Language Model- 020 ing to improve the PLM’s generalization abil- 021 ities for accurate adaptation to previously un- 022 seen tasks. After multi-task learning, the PLM 023 can be fine-tuned for any target few-shot NLP 024 tasks using the same prompting paradigm. Ex- 025 periments over a variety of NLP tasks show 026 that UPT consistently outperforms state-of- 027 the-arts for prompt-based fine-tuning. 1 028","{'model': 'tldr@v2.0.0', 'text': 'This work presents the Unified Prompt Tun- 010 ing ( UPT ) framework, leading to better few- 011 shot learning for BERT-style models by ex- 012 plicitly capturing prompting semantics from non-target NLP datasets.'}",
-CQARE: Contrastive Question-Answering for Few-shot Relation Extraction with Prompt Tuning,,"Prompt tuning with pre-trained language mod- 001 els (PLM) has exhibited outstanding perfor- 002 mance by closing the gap between pre-training 003 tasks and various downstream applications, 004 without the need for uninitialized parameters 005 to be introduced. However, prompt tuning 006 requires vast amounts of prompt engineering 007 and predefined label word mapping, which ob- 008 structs its implements in practice. Besides, the 009 ample label space makes prompt tuning more 010 arduous and challenging when it comes to re- 011 lation extraction (RE). To tackle these issues, 012 we propose a Contrastive Question-Answering 013 method with prompt tuning for few-shot RE 014 (CQARE). CQARE carries out a RE task- 015 specific pre-training with four entity-relation- 016 aware pre-training objects, including a prompt 017 pre-training to automatically generate contin- 018 uous prompts. The proposed pre-training can 019 provide more robust initialization with prompt 020 tuning while maintaining semantic consistency 021 with the proposed PLM. Furthermore, CQARE 022 can effectively avoid label words mapping by 023 reformulating RE as contrastive question an- 024 swering. The results indicate CQARE raising 025 averaged accuracy of 5.11% on a cross-domain 026 few-shot dataset, demonstrating that robust ini- 027 tialization is crucial for prompt tuning and ef- 028 fective contrastive question answering. 029","{'model': 'tldr@v2.0.0', 'text': 'A Contrastive Question-Answering method with prompt tuning for few-shot RE 014 (CQARE), which can effectively avoid label words mapping and reformulating RE as contrastive question answering while maintaining semantic consistency with the proposed PLM.'}",
-Prompt-Guided Few-Shot Event Detection,Prafulla Dhariwal,"Practical applications of event extraction sys- 001 tems have long been hindered by their need 002 for heavy human annotation. In order to scale 003 up to new domains and event types, models 004 must learn to cope with limited supervision, 005 as in few-shot learning settings. To this end, 006 the major challenge is to let the model master 007 the semantic of event types, without requiring 008 abundant event mention annotations. In our 009 study, we employ cloze prompts to elicit event- 010 related knowledge from pretrained language 011 models and further use event definitions and 012 keywords to pinpoint the trigger word. By for- 013 mulating the event detection task as an identify- 014 then-localize procedure, we minimize the num- 015 ber of type-specific parameters, enabling our 016 model to quickly adapt to event detection tasks 017 for new types. Experiments on three event de- 018 tection benchmark datasets (ACE, FewEvent, 019 MAVEN) show that our proposed method per- 020 forms favorably under fully supervised settings 021 and surpasses existing few-shot methods by 022 16% F1 on the FewEvent dataset and 23% on 023 the MAVEN dataset when only 5 examples are 024 provided for each event type. 1 025","{'model': 'tldr@v2.0.0', 'text': 'This study employs cloze prompts to elicit event- 010 related knowledge from pretrained language 011 models and further use event definitions and 012 keywords to pinpoint the trigger word, enabling the model to quickly adapt to event detection tasks for new types.'}",
-Prompting Decision Transformer for Few-Shot Policy Generalization,Mengdi Xu,"Humans can leverage prior experience and learn novel tasks from a handful of demonstrations. In contrast to offline meta-reinforcement learning, which aims to achieve quick adaptation through better algorithm design, we investigate the effect of architecture inductive bias on the few-shot learning capability. We propose a Prompt-based Decision Transformer (Prompt-DT), which leverages the sequential modeling ability of the Transformer architecture and the prompt framework to achieve few-shot adaptation in offline RL. We design the trajectory prompt, which contains segments of the few-shot demonstrations, and encodes task-specific information to guide policy generation. Our experiments in five MuJoCo control benchmarks show that Prompt-DT is a strong few-shot learner without any extra finetuning on unseen target tasks. Prompt-DT outperforms its variants and strong meta offline RL baselines by a large margin with a trajectory prompt containing only a few timesteps. Prompt-DT is also robust to prompt length changes and can generalize to out-of-distribution (OOD) environments.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt-based Decision Transformer (Prompt-DT), which leverages the sequential modeling ability of the Transformer architecture and the prompt framework to achieve few-shot adaptation in offline RL and design the trajectory prompt, which contains segments of the few- shot demonstrations, and encodes task-specific information to guide policy generation.'}",https://arxiv.org/pdf/2206.13499
-PSP: Pre-trained Soft Prompts for Few-Shot Abstractive Summarization,Xiaochen Liu,"Few-shot abstractive summarization has become a challenging task in natural language generation. To support it, we developed a novel soft prompts architecture coupled with a prompt pre-training plus prompt fine-tuning paradigm, which is effective and tunes only extremely light parameters. To meet the structure of the generation models, the soft prompts comprise continuous input embeddings across an encoder and a decoder. Importantly, a new inner-prompt placed in the text is introduced to capture document-level information. The aim is to devote attention to understanding the document that better prompts the model to generate document-related content. In the training process, the prompt pre-training with self-supervised pseudo-data firstly teaches the model basic summarizing capability. Then, with few-shot examples, only the designed lightweight soft prompts are fine-tuned. Experimental results on the CNN/DailyMail and XSum datasets show that our method, with only 0.1% of the parameters, outperforms full-model tuning where all model parameters are tuned. It also surpasses Prompt Tuning by a large margin and delivers competitive results against Prefix-Tuning with 3% of the parameters.","{'model': 'tldr@v2.0.0', 'text': 'A novel soft prompts architecture coupled with a prompt pre-training plus prompt fine-tuning paradigm, which is effective and tunes only extremely light parameters, and outperforms full-model tuning where all model parameters are tuned.'}",http://arxiv.org/pdf/2204.04413
-Few-shot Text-to-SQL Translation using Structure and Content Prompt Learning,Ju Fan,"A common problem with adopting Text-to-SQL translation in database systems is poor generalization. Specifically, when there is limited training data on new datasets, existing few-shot Text-to-SQL techniques, even with carefully designed textual prompts on pre-trained language models (PLMs), tend to be ineffective. In this paper, we present a divide-and-conquer framework to better support few-shot Text-to-SQL translation, which divides Text-to-SQL translation into two stages (or sub-tasks), such that each sub-task is simpler to be tackled. The first stage, called the structure stage , steers a PLM to generate an SQL structure (including SQL commands such as SELECT, FROM, WHERE and SQL operators such as “ < ”, “ > ”) with placeholders for missing identifiers. The second stage, called the content stage , guides a PLM to populate the placeholders in the generated SQL structure with concrete values (including SQL identifies such as table names, column names, and constant values). We propose a hybrid prompt strategy that combines learnable vectors and fixed vectors ( i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages. In addition, we design keyword constrained decoding to ensure the validity of generated SQL structures, and structure guided decoding to guarantee the model to fill correct content. Extensive experiments, by comparing with ten state-of-the-art Text-to-SQL solutions at the time of writing, show that SC-Prompt significantly outperforms them in the few-shot scenario. In particular, on the widely-adopted Spider dataset, given less than 500 labeled training examples (5% of the official training set), SC-Prompt outperforms","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hybrid prompt strategy that combines learnable vectors and fixed vectors, such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages of Text-to-SQL translation.'}",
-Noisy Channel Language Model Prompting for Few-Shot Text Classification,Sewon Min,"We introduce a noisy channel approach for language model prompting in few-shot text classification. Instead of computing the likelihood of the label given the input (referred as direct models), channel models compute the conditional probability of the input given the label, and are thereby required to explain every word in the input. We use channel models for recently proposed few-shot learning methods with no or very limited updates to the language model parameters, via either in-context demonstration or prompt tuning. Our experiments show that, for both methods, channel models significantly outperform their direct counterparts, which we attribute to their stability, i.e., lower variance and higher worst-case accuracy. We also present extensive ablations that provide recommendations for when to use channel prompt tuning instead of other competitive models (e.g., direct head tuning): channel prompt tuning is preferred when the number of training examples is small, labels in the training data are imbalanced, or generalization to unseen labels is required.","{'model': 'tldr@v2.0.0', 'text': 'A noisy channel approach for language model prompting in few-shot text classification by using channel models for recently proposed few- shot learning methods with no or very limited updates to the language model parameters, via either in-context demonstration or prompt tuning.'}",https://aclanthology.org/2022.acl-long.365.pdf
-In-Context Learning for Few-Shot Dialogue State Tracking,Yushi Hu,"Collecting and annotating task-oriented dialogues is time-consuming and costly; thus, zero and few shot learning could greatly benefit dialogue state tracking (DST). In this work, we propose an in-context learning (ICL) framework for zero-shot and few-shot learning DST, where a large pre-trained language model (LM) takes a test instance and a few exemplars as input, and directly decodes the dialogue state without any parameter updates. To better leverage a tabular domain description in the LM prompt, we reformulate DST into a text-to-SQL problem. We also propose a novel approach to retrieve annotated dialogues as exemplars. Empirical results on MultiWOZ show that our method IC-DST substantially outperforms previous fine-tuned state-of-the-art models in few-shot settings. In addition, we test IC-DST in zero-shot settings, in which the model only takes a fixed task instruction as input, finding that it outperforms previous zero-shot methods by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an in-context learning (ICL) framework for zero-shot and few-shot learning DST, where a large pre-trained language model takes a test instance and a few exemplars as input, and directly decodes the dialogue state without any parameter updates.'}",https://arxiv.org/pdf/2203.08568
-Zero- and Few-Shot Event Detection via Prompt-Based Meta Learning,Zhenrui Yue,"With emerging online topics as a source for numerous new events, detecting unseen / rare event types presents an elusive challenge for existing event detection methods, where only limited data access is provided for training. To address the data scarcity problem in event detection, we propose MetaEvent, a meta learning-based framework for zero- and few-shot event detection. Specifically, we sample training tasks from existing event types and perform meta training to search for optimal parameters that quickly adapt to unseen tasks. In our framework, we propose to use the cloze-based prompt and a trigger-aware soft verbalizer to efficiently project output to unseen event types. Moreover, we design a contrastive meta objective based on maximum mean discrepancy (MMD) to learn class-separating features. As such, the proposed MetaEvent can perform zero-shot event detection by mapping features to event types without any prior knowledge. In our experiments, we demonstrate the effectiveness of MetaEvent in both zero-shot and few-shot scenarios, where the proposed method achieves state-of-the-art performance in extensive experiments on benchmark datasets FewEvent and MAVEN.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MetaEvent can perform zero-shot event detection by mapping features to event types without any prior knowledge, and achieves state-of-the-art performance in extensive experiments on benchmark datasets FewEvent and MAVEN.'}",http://arxiv.org/pdf/2305.17373
-VPPT: Visual Pre-Trained Prompt Tuning Framework for Few-Shot Image Classification,Zhao Song,"Large-scale pre-trained transformers have recently achieved remarkable success in several computer vision tasks. However, it remains highly challenging to fully fine-tune models for downstream tasks, due to the expensive computational and storage cost. Recently, Parameter-Efficient Tuning (PETuning) techniques, e.g., Visual Prompt Tuning (VPT), have significantly reduced the computation cost by inserting lightweight prompt modules including prompt tokens or adapter layers, into the pre-trained models and tuning these prompt modules with a small number of trainable parameters, while keeping the transformer backbone freeze. Although encouraging results were achieved, existing PETuning methods cannot perform well under the few-shot learning settings (i.e., extremely limited training data, with only 1 or 2 shots per class), due to the scarce supervision signal. To this end, we first empirically identify the poor performance is mainly due to the inappropriate way of initializing prompt modules, which has also been verified in the pre-trained language models. Next, we propose a Visual Pre-trained Prompt Tuning framework (VPPT), which pre-trains the prompt modules first and then leverages the pre-trained modules along with the pre-trained transformer backbone to perform prompt tuning on downstream tasks. Extensive experiments show that our VPPT framework achieves 16.08% average accuracy absolute improvement under 1 shot setting on five fine-grained visual classification datasets, compared with the previous PETuning techniques, e.g., VPT, in few-shot image classification.","{'model': 'tldr@v2.0.0', 'text': 'A Visual Pre-trained Prompt Tuning framework (VPPT), which pre-trains the prompt modules first and then leverages the pre-trained modules along with thePre-trained transformer backbone to perform prompt tuning on downstream tasks, and achieves significant improvement under few-shot image classification settings.'}",
-Decomposed Two-Stage Prompt Learning for Few-Shot Named Entity Recognition,Feiyang Ye,"Named entity recognition (NER) in a few-shot setting is an extremely challenging task, and most existing methods fail to account for the gap between NER tasks and pre-trained language models. Although prompt learning has been successfully applied in few-shot classification tasks, adapting to token-level classification similar to the NER task presents challenges in terms of time consumption and efficiency. In this work, we propose a decomposed prompt learning NER framework for few-shot settings, decomposing the NER task into two stages: entity locating and entity typing. In training, the location information of distant labels is used to train the entity locating model. A concise but effective prompt template is built to train the entity typing model. In inference, a pipeline approach is used to handle the entire NER task, which elegantly resolves time-consuming and inefficient problems. Specifically, a well-trained entity locating model is used to predict entity spans for each input. The input is then transformed using prompt templates, and the well-trained entity typing model is used to predict their types in a single step. Experimental results demonstrate that our framework outperforms previous prompt-based methods by an average of 2.3–12.9% in F1 score while achieving the best trade-off between accuracy and inference speed.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a decomposed prompt learning NER framework for few-shot settings, decomposing the NER task into two stages: entity locating and entity typing.'}",https://www.mdpi.com/2078-2489/14/5/262/pdf?version=1682672547
-Few-Shot Table-to-Text Generation with Prompt Planning and Knowledge Memorization,Zhixin Guo,"Pre-trained language models (PLM) have achieved remarkable advancement in table-to-text generation tasks. However, the lack of labeled domain-specific knowledge and the topology gap between tabular data and text make it difficult for PLMs to yield faithful text. Low-resource generation likewise faces unique challenges in this domain. Inspired by how humans descript tabular data with prior knowledge, we suggest a new framework: PromptMize, which targets table-to-text generation under few-shot settings. The design of our framework consists of two aspects: a prompt planner and a knowledge adapter. The prompt planner aims to generate a prompt signal that provides instance guidance for PLMs to bridge the topology gap between tabular data and text. Moreover, the knowledge adapter memorizes domain-specific knowledge from the unlabelled corpus to supply essential information during generation. Extensive experiments and analyses are investigated on three open domain few-shot NLG datasets: human, song, and book. Compared with previous state-of-the-art approaches, our model achieves remarkable performance in generating quality as judged by human and automatic evaluations.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by how humans descript tabular data with prior knowledge, a new framework is suggested: PromptMize, which targets table-to-text generation under few-shot settings and achieves remarkable performance in generating quality as judged by human and automatic evaluations.'}",https://arxiv.org/pdf/2302.04415
-LoCoOp: Few-Shot Out-of-Distribution Detection via Prompt Learning,Atsuyuki Miyai,"We present a novel vision-language prompt learning approach for few-shot out-of-distribution (OOD) detection. Few-shot OOD detection aims to detect OOD images from classes that are unseen during training using only a few labeled in-distribution (ID) images. While prompt learning methods such as CoOp have shown effectiveness and efficiency in few-shot ID classification, they still face limitations in OOD detection due to the potential presence of ID-irrelevant information in text embeddings. To address this issue, we introduce a new approach called Local regularized Context Optimization (LoCoOp), which performs OOD regularization that utilizes the portions of CLIP local features as OOD features during training. CLIP's local features have a lot of ID-irrelevant nuisances (e.g., backgrounds), and by learning to push them away from the ID class text embeddings, we can remove the nuisances in the ID class text embeddings and enhance the separation between ID and OOD. Experiments on the large-scale ImageNet OOD detection benchmarks demonstrate the superiority of our LoCoOp over zero-shot, fully supervised detection methods and prompt learning methods. Notably, even in a one-shot setting -- just one label per class, LoCoOp outperforms existing zero-shot and fully supervised detection methods. The code will be available via https://github.com/AtsuMiyai/LoCoOp.","{'model': 'tldr@v2.0.0', 'text': 'A novel vision-language prompt learning approach for few-shot out-of-distribution (OOD) detection called LoCoOp, which performs OOD regularization that utilizes the portions of CLIP local features as OOD features during training.'}",http://arxiv.org/pdf/2306.01293
-Semantic Prompt for Few-Shot Image Recognition,Wentao Chen,"Few-shot learning is a challenging problem since only a few examples are provided to recognize a new class. Several recent studies exploit additional semantic information, e.g. text embeddings of class names, to address the issue of rare samples through combining semantic prototypes with visual prototypes. However, these methods still suffer from the spurious visual features learned from the rare support samples, resulting in limited benefits. In this paper, we propose a novel Semantic Prompt (SP) approach for few-shot learning. Instead of the naive exploitation of semantic information for remedying classifiers, we explore leveraging semantic information as prompts to tune the visual feature extraction network adaptively. Specifically, we design two complementary mechanisms to insert semantic prompts into the feature extractor: one is to enable the interaction between semantic prompts and patch embeddings along the spatial dimension via self-attention, another is to supplement visual features with the transformed semantic prompts along the channel dimension. By combining these two mechanisms, the feature extractor presents a better ability to attend to the class-specific features and obtains more generalized image representations with merely a few support samples. Through extensive experiments on four datasets, the proposed approach achieves promising results, improving the 1-shot learning accuracy by 3.67% on average.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs two complementary mechanisms to insert semantic prompts into the feature extractor, one is to enable the interaction between semantic prompts and patch embeddings along the spatial dimension via self-attention, another is to supplement visual features with the transformed semantic prompts along the channel dimension.'}",https://arxiv.org/pdf/2303.14123
-Few-shot Joint Multimodal Aspect-Sentiment Analysis Based on Generative Multimodal Prompt,Xiaocui Yang,"We have witnessed the rapid proliferation of multimodal data on numerous social media platforms. Conventional studies typically require massive labeled data to train models for Multimodal Aspect-Based Sentiment Analysis (MABSA). However, collecting and annotating fine-grained multimodal data for MABSA is tough. To alleviate the above issue, we perform three MABSA-related tasks with quite a small number of labeled multimodal samples. We first build diverse and comprehensive multimodal few-shot datasets according to the data distribution. To capture the specific prompt for each aspect term in a few-shot scenario, we propose a novel Generative Multimodal Prompt (GMP) model for MABSA, which includes the Multimodal Encoder module and the N-Stream Decoders module. We further introduce a subtask to predict the number of aspect terms in each instance to construct the multimodal prompt. Extensive experiments on two datasets demonstrate that our approach outperforms strong baselines on two MABSA-related tasks in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Generative Multimodal Prompt (GMP) model for MABSA, which includes the Multinodal Encoder module and the N-Stream Decoders module, and introduces a subtask to predict the number of aspect terms in each instance to construct the multimodal prompt.'}",http://arxiv.org/pdf/2305.10169
-PartSeg: Few-shot Part Segmentation via Part-aware Prompt Learning,M. Han,"In this work, we address the task of few-shot part segmentation, which aims to segment the different parts of an unseen object using very few labeled examples. It is found that leveraging the textual space of a powerful pre-trained image-language model (such as CLIP) can be beneficial in learning visual features. Therefore, we develop a novel method termed PartSeg for few-shot part segmentation based on multimodal learning. Specifically, we design a part-aware prompt learning method to generate part-specific prompts that enable the CLIP model to better understand the concept of ``part'' and fully utilize its textual space. Furthermore, since the concept of the same part under different object categories is general, we establish relationships between these parts during the prompt learning process. We conduct extensive experiments on the PartImageNet and Pascal$\_$Part datasets, and the experimental results demonstrated that our proposed method achieves state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': ""This work designs a part-aware prompt learning method to generate part-specific prompts that enable the CLIP model to better understand the concept of ``part'' and fully utilize its textual space.""}",https://arxiv.org/pdf/2308.12757
-Evolutionary Verbalizer Search for Prompt-based Few Shot Text Classification,Tongtao Ling,"Recent advances for few-shot text classification aim to wrap textual inputs with task-specific prompts to cloze questions. By processing them with a masked language model to predict the masked tokens and using a verbalizer that constructs the mapping between predicted words and target labels. This approach of using pre-trained language models is called prompt-based tuning, which could remarkably outperform conventional fine-tuning approach in the low-data scenario. As the core of prompt-based tuning, the verbalizer is usually handcrafted with human efforts or suboptimally searched by gradient descent. In this paper, we focus on automatically constructing the optimal verbalizer and propose a novel evolutionary verbalizer search (EVS) algorithm, to improve prompt-based tuning with the high-performance verbalizer. Specifically, inspired by evolutionary algorithm (EA), we utilize it to automatically evolve various verbalizers during the evolutionary procedure and select the best one after several iterations. Extensive few-shot experiments on five text classification datasets show the effectiveness of our method.","{'model': 'tldr@v2.0.0', 'text': 'This paper focuses on automatically constructing the optimal verbalizer and proposes a novel evolutionary verbalizer search (EVS) algorithm, inspired by evolutionary algorithm, to improve prompt-based tuning with the high-performance verbalizer.'}",http://arxiv.org/pdf/2306.10514
-A Chinese Few-Shot Text Classification Method Utilizing Improved Prompt Learning and Unlabeled Data,Tingkai Hu,"Insufficiently labeled samples and low-generalization performance have become significant natural language processing problems, drawing significant concern for few-shot text classification (FSTC). Advances in prompt learning have significantly improved the performance of FSTC. However, prompt learning methods typically require the pre-trained language model and tokens of the vocabulary list for model training, while different language models have different token coding structures, making it impractical to build effective Chinese prompt learning methods from previous approaches related to English. In addition, a majority of current prompt learning methods do not make use of existing unlabeled data, thus often leading to unsatisfactory performance in real-world applications. To address the above limitations, we propose a novel Chinese FSTC method called CIPLUD that combines an improved prompt learning method and existing unlabeled data, which are used for the classification of a small amount of Chinese text data. We used the Chinese pre-trained language model to build two modules: the Multiple Masks Optimization-based Prompt Learning (MMOPL) module and the One-Class Support Vector Machine-based Unlabeled Data Leveraging (OCSVM-UDL) module. The former generates prompt prefixes with multiple masks and constructs suitable prompt templates for Chinese labels. It optimizes the random token combination problem during label prediction with joint probability and length constraints. The latter, by establishing an OCSVM model in the trained text vector space, selects reasonable pseudo-label data for each category from a large amount of unlabeled data. After selecting the pseudo-label data, we mixed them with the previous few-shot annotated data to obtain brand new training data and then repeated the steps of the two modules as an iterative semi-supervised optimization process. The experimental results on the four Chinese FSTC benchmark datasets demonstrate that our proposed solution outperformed other prompt learning methods with an average accuracy improvement of 2.3%.","{'model': 'tldr@v2.0.0', 'text': 'A novel Chinese FSTC method called CIPLUD is proposed that combines an improved prompt learning method and existing unlabeled data, which are used for the classification of a small amount of Chinese text data, and outperformed other prompt learning methods with an average accuracy improvement of 2.3%.'}",https://www.mdpi.com/2076-3417/13/5/3334/pdf?version=1678093925
-Unified Prompt Learning Makes Pre-Trained Language Models Better Few-Shot Learners,Feihu Jin,"Language prompting induces the model to produce a textual output during the training phase, which achieves remarkable performance in few-shot learning scenarios. However, current prompt-based methods either use the same task-specific prompts for each instance, losing the particularity of instance-dependent information, or generate an instance-dependent prompt for each instance, lacking shared information about the task. In this paper, we propose an efficient few-shot learning method to dynamically decide the degree to which task-specific and instance-dependent information are incorporated according to different task and instance characteristics, enriching the prompt with task-specific and instance-dependent information. Extensive experiments on a wide range of natural language understanding tasks demonstrate that our approach obtains significant improvements compared to prompt-based fine-tuning baselines in a few-shot setting with about 0.1% parameters tuned. Moreover, our approach outperforms existing state-of-the-art efficient few-shot learning methods on several natural language understanding tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an efficient few-shot learning method to dynamically decide the degree to which task-specific and instance-dependent information are incorporated according to different task and instance characteristics, enriching the prompt with task- specific and instanceTheoretical studies demonstrate that this approach obtains significant improvements compared to prompt-based fine-tuning baselines in a few- shot setting.'}",
-Boosting Prompt-Based Few-Shot Learners Through Out-of-Domain Knowledge Distillation,Xiaoqing Chen,"Prompt-based learning improves the performance of Pre-trained Language Models (PLMs) over few-shot learning and is suitable for low-resourced scenarios. However, it is challenging to deploy large PLMs online. Knowledge Distillation (KD) can compress large PLMs into small ones; yet, few-shot KD for prompt-tuned PLMs is challenging due to the lack of training data and the capacity gap between teacher and student models. We propose Boost-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs with the help of the out-of-domain data. Apart from distilling the model logits, Boost-Distiller specifically considers heuristically-generated fake logits that improve the generalization abilities of student models. We further leverage the cross-domain model logits, weighted with domain expertise scores that measure the transferablity of out-of-domain instances. Experiments over various datasets show Boost-Distiller consistently outperforms baselines by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Boost-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs with the help of the out-of-domain data, and specifically considers heuristically-generated fake logits that improve the generalization abilities of student models.'}",
-Prompt-Based Zero- and Few-Shot Node Classification: A Multimodal Approach,Yue Li,"Multimodal data empowers machine learning models to better understand the world from various perspectives. In this work, we study the combination of \emph{text and graph} modalities, a challenging but understudied combination which is prevalent across multiple settings including citation networks, social media, and the web. We focus on the popular task of node classification using limited labels; in particular, under the zero- and few-shot scenarios. In contrast to the standard pipeline which feeds standard precomputed (e.g., bag-of-words) text features into a graph neural network, we propose \textbf{T}ext-\textbf{A}nd-\textbf{G}raph (TAG) learning, a more deeply multimodal approach that integrates the raw texts and graph topology into the model design, and can effectively learn from limited supervised signals without any meta-learning procedure. TAG is a two-stage model with (1) a prompt- and graph-based module which generates prior logits that can be directly used for zero-shot node classification, and (2) a trainable module that further calibrates these prior logits in a few-shot manner. Experiments on two node classification datasets show that TAG outperforms all the baselines by a large margin in both zero- and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'TAG is a two-stage model with a prompt- and graph-based module which generates prior logits that can be directly used for zero-shot node classification, and a trainable module that further calibrates these priorlogits in a few-shot manner.'}",https://arxiv.org/pdf/2307.11572
-Prompt-Distiller: Few-Shot Knowledge Distillation for Prompt-Based Language Learners with Dual Contrastive Learning,Boyu Hou,"Prompt-based learning has improved the few-shot learning performance of large-scale Pre-trained Language Models (PLMs). Yet, it is challenging to deploy large-scale PLMs in resource-constrained environments for online applications. Knowledge Distillation (KD) is a promising approach for PLM compression. However, distilling prompt-tuned PLMs in the few-shot learning setting is a non-trivial problem due to the lack of task-specific training data and KD techniques for the new prompting paradigm. We propose Prompt-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs, which forces the student model to learn from both its pre-trained and prompt-tuned teacher models to alleviate the model overfitting problem. We further design a contrastive learning technique to learn higher-order dependencies from intermediate-layer representations of teacher models, considering different knowledge capacities of teacher and student models. Extensive experiments over various datasets show that Prompt-Distiller consistently outperforms baselines by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prompt-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs, which forces the student model to learn from both its pre-trained and prompt- Tuned teacher models to alleviate the model overfitting problem.'}",
-Few-shot Text-to-SQL Translation using Structure and Content Prompt Learning,Zihui Gu,"A common problem with adopting Text-to-SQL translation in database systems is poor generalization. Specifically, when there is limited training data on new datasets, existing few-shot Text-to-SQL techniques, even with carefully designed textual prompts on pre-trained language models (PLMs), tend to be ineffective. In this paper, we present a divide-and-conquer framework to better support few-shot Text-to-SQL translation, which divides Text-to-SQL translation into two stages (or sub-tasks), such that each sub-task is simpler to be tackled. The first stage, called the structure stage, steers a PLM to generate an SQL structure (including SQL commands such as SELECT, FROM, WHERE and SQL operators such as <"", ?>"") with placeholders for missing identifiers. The second stage, called the content stage, guides a PLM to populate the placeholders in the generated SQL structure with concrete values (including SQL identifies such as table names, column names, and constant values). We propose a hybrid prompt strategy that combines learnable vectors and fixed vectors (i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages. In addition, we design keyword constrained decoding to ensure the validity of generated SQL structures, and structure guided decoding to guarantee the model to fill correct content. Extensive experiments, by comparing with ten state-of-the-art Text-to-SQL solutions at the time of writing, show that SC-Prompt significantly outperforms them in the few-shot scenario. In particular, on the widely-adopted Spider dataset, given less than 500 labeled training examples (5% of the official training set), SC-Prompt outperforms the previous SOTA methods by around 5% on accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hybrid prompt strategy that combines learnable vectors and fixed vectors (i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages.'}",
-DreamArtist: Towards Controllable One-Shot Text-to-Image Generation via Positive-Negative Prompt-Tuning,Ziyi Dong,"Large-scale text-to-image generation models have achieved remarkable progress in synthesizing high-quality, feature-rich images with high resolution guided by texts. However, these models often struggle with novel concepts, eg, new styles, object entities, etc. Although recent attempts have employed fine-tuning or prompt-tuning strategies to teach the pre-trained diffusion model novel concepts from a reference image set,they have the drawback of overfitting to the given reference images, particularly in one-shot applications, which is harmful to generate diverse and high-quality images while maintaining generation controllability. To tackle this challenge, we present a simple yet effective method called DreamArtist, which employs a positive-negative prompt-tuning learning strategy. Specifically, DreamArtist incorporates both positive and negative embeddings and jointly trains them. The positive embedding aggressively captures the salient characteristics of the reference image to drive diversified generation and the negative embedding rectifies inadequacies from the positive embedding. It learns not only what is correct, but also what can be avoided or improved. We have conducted extensive experiments and evaluated the proposed method from image similarity and diversity, generation controllability, and style cloning. And our DreamArtist has achieved a superior generation performance over existing methods. Besides, our additional evaluation on extended tasks, including concept compositions and prompt-guided image editing, demonstrates its effectiveness for more applications.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a simple yet effective method, which employs a positive-negative prompt-tuning learning strategy, which incorporates both positive and negative embeddings and jointly trains them and has achieved a superior generation performance over existing methods.'}",
-DreamArtist: Towards Controllable One-Shot Text-to-Image Generation via Contrastive Prompt-Tuning,Ziyi Dong,"Large-scale text-to-image generation models with an exponential evolution can currently synthesize high-resolution, feature-rich, high-quality images based on text guidance. However, they are often overwhelmed by words of new concepts, styles, or object entities that always emerge. Although there are some recent attempts to use fine-tuning or prompt-tuning methods to teach the model a new","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2211.11337
-One-shot and Partially-Supervised Cell Image Segmentation Using Small Visual Prompt,S. Kato,"Semantic segmentation of microscopic cell images using deep learning is an important technique, however, it requires a large number of images and ground truth labels for training. To address the above problem, we consider an efficient learning framework with as little data as possible, and we propose two types of learning strategies: One-shot segmentation which can learn with only one training sample, and Partially-supervised segmentation which assigns annotations to only a part of images. Furthermore, we introduce novel segmentation methods using the small prompt images inspired by prompt learning in recent studies. Our proposed methods use a pre-trained model based on only cell images and teach the information of the prompt pairs to the target image to be segmented by the attention mechanism, which allows for efficient learning while reducing the burden of annotation costs. Through experiments conducted on three types of microscopic cell image datasets, we confirmed that the proposed method improved the Dice score coefficient (DSC) in comparison with the conventional methods. Our code is available at https://github.com/usagisukisuki/Oneshot-Part-CellSegmentation.","{'model': 'tldr@v2.0.0', 'text': 'The proposed methods use a pre-trained model based on only cell images and teach the information of the prompt pairs to the target image to be segmented by the attention mechanism, which allows for efficient learning while reducing the burden of annotation costs.'}",https://arxiv.org/pdf/2304.07991
-Learning from Graph Propagation via Ordinal Distillation for One-Shot Automated Essay Scoring,Zhiwei Jiang,"One-shot automated essay scoring (AES) aims to assign scores to a set of essays written specific to a certain prompt, with only one manually scored essay per distinct score. Compared to the previous-studied prompt-specific AES which usually requires a large number of manually scored essays for model training (e.g., about 600 manually scored essays out of totally 1000 essays), one-shot AES can greatly reduce the workload of manual scoring. In this paper, we propose a Transductive Graph-based Ordinal Distillation (TGOD) framework to tackle the task of one-shot AES. Specifically, we design a transductive graph-based model as a teacher model to generate pseudo labels of unlabeled essays based on the one-shot labeled essays. Then, we distill the knowledge in the teacher model into a neural student model by learning from the high confidence pseudo labels. Different from the general knowledge distillation, we propose an ordinal-aware unimodal distillation which makes a unimodal distribution constraint on the output of student model, to tolerate the minor errors existed in pseudo labels. Experimental results on the public dataset ASAP show that TGOD can improve the performance of existing neural AES models under the one-shot AES setting and achieve an acceptable average QWK of 0.69.","{'model': 'tldr@v2.0.0', 'text': 'A transductive graph-based model as a teacher model to generate pseudo labels of unlabeled essays based on the one-shot labeled essays is designed, which distill the knowledge in the teacher model into a neural student model by learning from the high confidence pseudo labels.'}",
-MENTAL HEALTH RECOGNITION USING ONE-SHOT LEARNING AND FACI AL EXPRESSIONS,Mahesh Kumar,"The visual representation of a person's affective state, cognitive function, purpose, personality, and psychology can be obtained from facial expressions.[1]. The idea of our project is to create a platform with a camera, to determine whether an individual is suffering from deterioration of mental health, particularly in an institutionalized environment. In general, disturbed mental condition is determined by using a combination of various factors such as facial expressions, body posture, speech, tone, gestures, and so on[2]. This project lays focus on Facial Expression Recognition (FER), achievedwith the help of Convolution Neural Network (CNN), One-Shot learning and TensorFlow. The dataset used here is the Kaggle dataset FER2013, which comprises of seven facial expressions labelled as happy, sad, surprise, fear, anger, disgust and neutral. The results of this classification are stored and compared to determine whether an individual displays negative emotion for a significant number of times. After detection and classification, a prompt is sent to the designated authorities at the institution through a web application. On the basis of that, if an individual is found to be suffering from a disturbed state of mind, he/she is recommended for therapy or medical attention.","{'model': 'tldr@v2.0.0', 'text': 'The idea of this project is to create a platform with a camera, to determine whether an individual is suffering from deterioration of mental health, particularly in an institutionalized environment, with the help of Convolution Neural Network, One-Shot learning and TensorFlow.'}",
-PØDA: Prompt-driven Zero-shot Domain Adaptation,Mohammad Fahes,"Domain adaptation has been vastly investigated in computer vision but still requires access to target images at train time, which might be intractable in some uncommon conditions. In this paper, we propose the task of `Prompt-driven Zero-shot Domain Adaptation', where we adapt a model trained on a source domain using only a general description in natural language of the target domain, i.e., a prompt. First, we leverage a pretrained contrastive vision-language model (CLIP) to optimize affine transformations of source features, steering them towards the target text embedding while preserving their content and semantics. To achieve this, we propose Prompt-driven Instance Normalization (PIN). Second, we show that these prompt-driven augmentations can be used to perform zero-shot domain adaptation for semantic segmentation. Experiments demonstrate that our method significantly outperforms CLIP-based style transfer baselines on several datasets for the downstream task at hand, even surpassing one-shot unsupervised domain adaptation. A similar boost is observed on object detection and image classification. The code is available at https://github.com/astra-vision/PODA .","{'model': 'tldr@v2.0.0', 'text': 'This paper uses a pretrained contrastive vision-language model (CLIP) to optimize affine transformations of source features, steering them towards the target text embedding while preserving their content and semantics, and shows that these prompt-driven augmentations can be used to perform zero-shot domain adaptation for semantic segmentation.'}",https://arxiv.org/pdf/2212.03241
-CoHOZ: Contrastive Multimodal Prompt Tuning for Hierarchical Open-set Zero-shot Recognition,Ning Liao,"Practical image recognition often encounters samples whose labels either are totally unknown or belong to new classes outside the training set. The first problem refers to the open-set recognition (OSR), in which unknown classes are recognized as one with no more semantic information. While the latter is called zero-shot learning (ZSL), in which new classes are usually predefined. The existing literature mostly addresses these two problems separately. In this paper, we take the ambition for solving the combination of these two problems to fulfill semantically recognizing the unknown classes detected in OSR by zero-shot prediction. We propose the Contrastive multimodal prompt tuning for Hierarchical Open-set Zero-shot recognition (CoHOZ). Specifically, we firstly build a global and compatible hierarchical label tree with all downstream datasets aligned, which lays foundations for other modules. To detect unknown classes, we propose the contrastive continuous prompt tuning, which introduces additional negative classes from the fine level of the built hierarchy for prompt learning. To generate candidate classes for zero-shot prediction on the unknown data using prompt, we combine the built hierarchy to collect candidate classes from coarse to fine. In our experiments, when following the standard OSR protocol regarding all the unknown classes as a single class, CoHOZ achieves a new state-of-the-art performance both in unknown detection and open-set recognition. Few-shot tuning by the CoHOZ also shows competitive performance on them. In addition, the detailed semantic information of unknown classes are well explored, which has also been verified in experiments.","{'model': 'tldr@v2.0.0', 'text': 'The ambition for solving the combination of these two problems to fulfill semantically recognizing the unknown classes detected in OSR by zero-shot prediction is taken and the Contrastive multimodal prompt tuning for Hierarchical Open-set Zero-shot recognition (CoHOZ) is proposed.'}",
-ProZe: Explainable and Prompt-Guided Zero-Shot Text Classification,Ismail Harrando,"As technology accelerates the generation and communication of textual data, the need to automatically understand this content becomes a necessity. In order to classify text, being it for tagging, indexing, or curating documents, one often relies on large, opaque models that are trained on preannotated datasets, making the process unexplainable, difficult to scale, and ill-adapted for niche domains with scarce data. To tackle these challenges, we propose ProZe, a text classification approach that leverages knowledge from two sources: prompting pretrained language models, as well as querying ConceptNet, a common-sense knowledge base which can be used to add a layer of explainability to the results. We evaluate our approach empirically and we show how this combination not only performs on par with state-of-the-art zero shot classification on several domains, but also offers explainable predictions that can be visualized.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ProZe, a text classification approach that leverages knowledge from two sources: prompting pretrained language models, as well as querying ConceptNet, a common-sense knowledge base which can be used to add a layer of explainability to the results.'}",
-ControlVideo: Adding Conditional Control for One Shot Text-to-Video Editing,Min Zhao,"In this paper, we present ControlVideo, a novel method for text-driven video editing. Leveraging the capabilities of text-to-image diffusion models and ControlNet, ControlVideo aims to enhance the fidelity and temporal consistency of videos that align with a given text while preserving the structure of the source video. This is achieved by incorporating additional conditions such as edge maps, fine-tuning the key-frame and temporal attention on the source video-text pair with carefully designed strategies. An in-depth exploration of ControlVideo's design is conducted to inform future research on one-shot tuning video diffusion models. Quantitatively, ControlVideo outperforms a range of competitive baselines in terms of faithfulness and consistency while still aligning with the textual prompt. Additionally, it delivers videos with high visual realism and fidelity w.r.t. the source content, demonstrating flexibility in utilizing controls containing varying degrees of source video information, and the potential for multiple control combinations. The project page is available at \href{https://ml.cs.tsinghua.edu.cn/controlvideo/}{https://ml.cs.tsinghua.edu.cn/controlvideo/}.","{'model': 'tldr@v2.0.0', 'text': 'Quantitatively, ControlVideo outperforms a range of competitive baselines in terms of faithfulness and consistency while still aligning with the textual prompt, and it delivers videos with high visual realism and fidelity w.r.t. the source content.'}",http://arxiv.org/pdf/2305.17098
-Matcher: Segment Anything with One Shot Using All-Purpose Feature Matching,Yang Liu,"Powered by large-scale pre-training, vision foundation models exhibit significant potential in open-world image understanding. Even though individual models have limited capabilities, combining multiple such models properly can lead to positive synergies and unleash their full potential. In this work, we present Matcher, which segments anything with one shot by integrating an all-purpose feature extraction model and a class-agnostic segmentation model. Naively connecting the models results in unsatisfying performance, e.g., the models tend to generate matching outliers and false-positive mask fragments. To address these issues, we design a bidirectional matching strategy for accurate cross-image semantic dense matching and a robust prompt sampler for mask proposal generation. In addition, we propose a novel instance-level matching strategy for controllable mask merging. The proposed Matcher method delivers impressive generalization performance across various segmentation tasks, all without training. For example, it achieves 52.7% mIoU on COCO-20$^i$ for one-shot semantic segmentation, surpassing the state-of-the-art specialist model by 1.6%. In addition, our visualization results show open-world generality and flexibility on images in the wild. The code shall be released at https://github.com/aim-uofa/Matcher.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Matcher, which segments anything with one shot by integrating an all-purpose feature extraction model and a class-agnostic segmentation model, and proposes a novel instance-level matching strategy for controllable mask merging.'}",http://arxiv.org/pdf/2305.13310
-The Segment Anything Model (SAM) for Remote Sensing Applications: From Zero to One Shot,L. Osco,"Segmentation is an essential step for remote sensing image processing. This study aims to advance the application of the Segment Anything Model (SAM), an innovative image segmentation model by Meta AI, in the field of remote sensing image analysis. SAM is known for its exceptional generalization capabilities and zero-shot learning, making it a promising approach to processing aerial and orbital images from diverse geographical contexts. Our exploration involved testing SAM across multi-scale datasets using various input prompts, such as bounding boxes, individual points, and text descriptors. To enhance the model's performance, we implemented a novel automated technique that combines a text-prompt-derived general example with one-shot training. This adjustment resulted in an improvement in accuracy, underscoring SAM's potential for deployment in remote sensing imagery and reducing the need for manual annotation. Despite the limitations encountered with lower spatial resolution images, SAM exhibits promising adaptability to remote sensing data analysis. We recommend future research to enhance the model's proficiency through integration with supplementary fine-tuning techniques and other networks. Furthermore, we provide the open-source code of our modifications on online repositories, encouraging further and broader adaptations of SAM to the remote sensing domain.",,http://arxiv.org/pdf/2306.16623
-On the Design of Optimal One-Shot Distributed Combining for Cooperative Multi-UAV Systems,Xuesong Pan,"Unmanned aerial vehicle (UAV) networks will become one of the key components of the future mobile communication systems, especially suitable for on-demand coverage extension and prompt capacity enhancement. However, UAV networks, detached from the ground infrastructure, are usually constrained for traffic-intensive network-wide signal processing. This paper tackles this issue by considering the multi-user signal detection at the UAV swarm, where the uplink signals of multiple ground users are cooperatively recovered. In particular, to reduce the signaling exchange and the coordination latency, a two-stage distributed minimum mean squared error (MMSE) detection is proposed for the UAV swarm, which merges the local MMSE detections at each UAV with one-shot weighted combining at a central station. The combining weights only depend on the long-term statistical channel state informations (CSIs) and therefore, are suitable for distributed network with limited signaling exchange. Numerical results show that the proposed distributed MMSE detection achieves similar performance to the centralized MMSE detection.","{'model': 'tldr@v2.0.0', 'text': 'A two-stage distributed minimum mean squared error (MMSE) detection is proposed for the UAV swarm, which merges the local MMSE detections at each UAV with one-shot weighted combining at a central station, suitable for distributed network with limited signaling exchange.'}",
-Reinforcement Learning-based Grasping via One-Shot Affordance Localization and Zero-Shot Contrastive Language–Image Learning,Xiang Long,"—We present a novel robotic grasping system using a caging-style gripper, that combines one-shot affordance localization and zero-shot object identification. We demonstrate an integrated system requiring minimal prior knowledge, focusing on flexible few-shot object agnostic approaches. For grasping a novel target object, we use as input the color and depth of the scene, an image of an object affordance similar to the target object, and an up to three-word text prompt describing the target object. We demonstrate the system using real-world grasping of objects from the YCB benchmark set, with four distractor objects cluttering the scene. Overall, our pipeline has a success rate of the affordance localization of 96% , object identification of 62 . 5% , and grasping of 72% . Videos are on the project website: https://sites.google.com/view/ rl-affcorrs-grasp .","{'model': 'tldr@v2.0.0', 'text': None}",
-One-shot Localization and Segmentation of Medical Images with Foundation Models,Deepa Anand,"Recent advances in Vision Transformers (ViT) and Stable Diffusion (SD) models with their ability to capture rich semantic features of the image have been used for image correspondence tasks on natural images. In this paper, we examine the ability of a variety of pre-trained ViT (DINO, DINOv2, SAM, CLIP) and SD models, trained exclusively on natural images, for solving the correspondence problems on medical images. While many works have made a case for in-domain training, we show that the models trained on natural images can offer good performance on medical images across different modalities (CT,MR,Ultrasound) sourced from various manufacturers, over multiple anatomical regions (brain, thorax, abdomen, extremities), and on wide variety of tasks. Further, we leverage the correspondence with respect to a template image to prompt a Segment Anything (SAM) model to arrive at single shot segmentation, achieving dice range of 62%-90% across tasks, using just one image as reference. We also show that our single-shot method outperforms the recently proposed few-shot segmentation method - UniverSeg (Dice range 47%-80%) on most of the semantic segmentation tasks(six out of seven) across medical imaging modalities.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the models trained on natural images can offer good performance on medical images across different modalities (CT,MR,Ultrasound) sourced from various manufacturers, over multiple anatomical regions (brain, thorax, abdomen, extremities), and on wide variety of tasks.'}",
-Prompt-based Extraction of Social Determinants of Health Using Few-shot Learning,Giridhar Kaushik Ramachandran,"Social determinants of health (SDOH) documented in the electronic health record through unstructured text are increasingly being studied to understand how SDOH impacts patient health outcomes. In this work, we utilize the Social History Annotation Corpus (SHAC), a multi-institutional corpus of de-identified social history sections annotated for SDOH, including substance use, employment, and living status information. We explore the automatic extraction of SDOH information with SHAC in both standoff and inline annotation formats using GPT-4 in a one-shot prompting setting. We compare GPT-4 extraction performance with a high-performing supervised approach and perform thorough error analyses. Our prompt-based GPT-4 method achieved an overall 0.652 F1 on the SHAC test set, similar to the 7th best-performing system among all teams in the n2c2 challenge with SHAC.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes the Social History Annotation Corpus (SHAC), a multi-institutional corpus of de-identified social history sections annotated for SDOH, including substance use, employment, and living status information, to explore the automatic extraction of SDOH information with SHAC in both standoff and inline annotation formats using GPT-4 in a one-shot prompting setting.'}",http://arxiv.org/pdf/2306.07170
-Augmenters at SemEval-2023 Task 1: Enhancing CLIP in Handling Compositionality and Ambiguity for Zero-Shot Visual WSD through Prompt Augmentation and Text-To-Image Diffusion,Jie Li,"This paper describes our zero-shot approachesfor the Visual Word Sense Disambiguation(VWSD) Task in English. Our preliminarystudy shows that the simple approach of match-ing candidate images with the phrase usingCLIP suffers from the many-to-many natureof image-text pairs. We find that the CLIP textencoder may have limited abilities in captur-ing the compositionality in natural language.Conversely, the descriptive focus of the phrasevaries from instance to instance. We addressthese issues in our two systems, Augment-CLIPand Stable Diffusion Sampling (SD Sampling).Augment-CLIP augments the text prompt bygenerating sentences that contain the contextphrase with the help of large language mod-els (LLMs). We further explore CLIP modelsin other languages, as the an ambiguous wordmay be translated into an unambiguous one inthe other language. SD Sampling uses text-to-image Stable Diffusion to generate multipleimages from the given phrase, increasing thelikelihood that a subset of images match theone that paired with the text.","{'model': 'tldr@v2.0.0', 'text': 'The preliminary study shows that the simple approach of match-ing candidate images with the phrase using CLIP suffers from the many-to-many nature of image-text pairs, and finds that the CLIP textencoder may have limited abilities in captur-ing the compositionality in natural language.'}",https://arxiv.org/pdf/2307.05564
-Self-supervised Meta-Prompt Learning with Meta-Gradient Regularization for Few-shot Generalization,Kaihang Pan,"Prompt tuning is a parameter-efficient method, which learns soft prompts and conditions frozen language models to perform specific downstream tasks. Though effective, prompt tuning under few-shot settings on the one hand heavily relies on a good initialization of soft prompts. On the other hand, it can easily overfit to few-shot training samples, thereby undermining generalizability. Existing works leverage pre-training or supervised meta-learning to initialize soft prompts but they fail to data-efficiently generalize to unseen downstream tasks. To address the above problems, this paper proposes a novel Self-sUpervised meta-Prompt learning framework with MEta-gradient Regularization for few-shot generalization (SUPMER). SUPMER leverages self-supervised meta-learning with a diverse set of well-designed meta-training tasks to learn a universal prompt initialization for efficient adaptation using only unlabeled data. Additionally, it jointly meta-learns a gradient regularization function to transform raw gradients into a domain-generalizable direction, thus alleviating the problem of overfitting. Extensive experiments show that SUPMER achieves better performance for different few-shot downstream tasks, and also exhibits a stronger domain generalization ability. The code for SUPMER will be available at https://github.com/beepkh/SUPMER.","{'model': 'tldr@v2.0.0', 'text': 'SUPMER leverages self-supervised meta-learning with a diverse set of well-designed meta-training tasks to learn a universal prompt initialization for efficient adaptation using only unlabeled data, and jointly meta-learns a gradient regularization function to transform raw gradients into a domain-generalizable direction, thus alleviating the problem of overfitting.'}",
-Enhancing Black-Box Few-Shot Text Classification with Prompt-Based Data Augmentation,Dan Luo,"Training or finetuning large-scale language models (LLMs) such as GPT-3 requires substantial computation resources, motivating recent efforts to explore parameter-efficient adaptation to downstream tasks. One practical area of research is to treat these models as black boxes and interact with them through their inference APIs. In this paper, we investigate how to optimize few-shot text classification without accessing the gradients of the LLMs. To achieve this, we treat the black-box model as a feature extractor and train a classifier with the augmented text data. Data augmentation is performed using prompt-based finetuning on an auxiliary language model with a much smaller parameter size than the black-box model. Through extensive experiments on eight text classification datasets, we show that our approach, dubbed BT-Classifier, significantly outperforms state-of-the-art black-box few-shot learners and performs on par with methods that rely on full-model tuning.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates how to optimize few-shot text classification without accessing the gradients of the LLMs by treating the black-box model as a feature extractor and train a classifier with the augmented text data.'}",http://arxiv.org/pdf/2305.13785
-"CoCoOpter: Pre-train, prompt, and fine-tune the vision-language model for few-shot image classification",Jie Yan,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed CoCoOpter alleviates the overfitting and ensures generalizability in unseen new categories in unseen classes and achieves superior few-shot classification performance with a straightforward design.'}",
-Few-shot fake news detection via prompt-based tuning,Wang Gao," As people increasingly use social media to read news, fake news has become a major problem for the public and government. One of the main challenges in fake news detection is how to identify them in the early stage of propagation. Another challenge is that detection model training requires large amounts of labeled data, which are often unavailable or expensive to acquire. To address these challenges, we propose a novel Fake News Detection model based on Prompt Tuning (FNDPT). FNDPT first designs a prompt-based template for early fake news detection. This mechanism incorporates contextual information into textual content and extracts relevant knowledge from pre-trained language models. Furthermore, our model utilizes prompt-based tuning to enhance the performance in a few-shot setting. Experimental results on two real-world datasets verify the effectiveness of FNDPT.","{'model': 'tldr@v2.0.0', 'text': 'A novel Fake News Detection model based on Prompt Tuning (FNDPT), which incorporates contextual information into textual content and extracts relevant knowledge from pre-trained language models to enhance the performance in a few-shot setting.'}",
-LM-CPPF: Paraphrasing-Guided Data Augmentation for Contrastive Prompt-Based Few-Shot Fine-Tuning,Amirhossein Abaskohi,"In recent years, there has been significant progress in developing pre-trained language models for NLP. However, these models often struggle when fine-tuned on small datasets. To address this issue, researchers have proposed various adaptation approaches. Prompt-based tuning is arguably the most common way, especially for larger models. Previous research shows that adding contrastive learning to prompt-based fine-tuning is effective as it helps the model generate embeddings that are more distinguishable between classes, and it can also be more sample-efficient as the model learns from positive and negative examples simultaneously. One of the most important components of contrastive learning is data augmentation, but unlike computer vision, effective data augmentation for NLP is still challenging. This paper proposes LM-CPPF, Contrastive Paraphrasing-guided Prompt-based Fine-tuning of Language Models, which leverages prompt-based few-shot paraphrasing using generative language models, especially large language models such as GPT-3 and OPT-175B, for data augmentation. Our experiments on multiple text classification benchmarks show that this augmentation method outperforms other methods, such as easy data augmentation, back translation, and multiple templates.","{'model': 'tldr@v2.0.0', 'text': 'LM-CPPF, Contrastive Paraphrasing-guided Prompt-based Fine-tuning of Language Models, which leverages prompt-based few-shot paraphrasing using generative language models, especially large language models such as GPT-3 and OPT-175B, for data augmentation is proposed.'}",https://aclanthology.org/2023.acl-short.59.pdf
-Syntax-aware Hybrid prompt model for Few-shot multi-modal sentiment analysis,Zikai Zhou,"Multimodal Sentiment Analysis (MSA) has been a popular topic in natural language processing nowadays, at both sentence and aspect level. However, the existing approaches almost require large-size labeled datasets, which bring about large consumption of time and resources. Therefore, it is practical to explore the method for few-shot sentiment analysis in cross-modalities. Previous works generally execute on textual modality, using the prompt-based methods, mainly two types: hand-crafted prompts and learnable prompts. The existing approach in few-shot multi-modality sentiment analysis task has utilized both methods, separately. We further design a hybrid pattern that can combine one or more fixed hand-crafted prompts and learnable prompts and utilize the attention mechanisms to optimize the prompt encoder. The experiments on both sentence-level and aspect-level datasets prove that we get a significant outperformance.","{'model': 'tldr@v2.0.0', 'text': 'A hybrid pattern that can combine one or more fixed hand-crafted prompts and learnable prompts and utilize the attention mechanisms to optimize the prompt encoder is designed.'}",https://arxiv.org/pdf/2306.01312
-Enhancing Few-shot NER with Prompt Ordering based Data Augmentation,Huiming Wang,"Recently, data augmentation (DA) methods have been proven to be effective for pre-trained language models (PLMs) in low-resource settings, including few-shot named entity recognition (NER). However, conventional NER DA methods are mostly aimed at sequence labeling models, i.e., token-level classification, and few are compatible with unified autoregressive generation frameworks, which can handle a wider range of NER tasks, such as nested NER. Furthermore, these generation frameworks have a strong assumption that the entities will appear in the target sequence with the same left-to-right order as the source sequence. In this paper, we claim that there is no need to keep this strict order, and more diversified but reasonable target entity sequences can be provided during the training stage as a novel DA method. Nevertheless, a naive mixture of augmented data can confuse the model since one source sequence will then be paired with different target sequences. Therefore, we propose a simple but effective Prompt Ordering based Data Augmentation (PODA) method to improve the training of unified autoregressive generation frameworks under few-shot NER scenarios. Experimental results on three public NER datasets and further analyses demonstrate the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple but effective Prompt Ordering based Data Augmentation (PODA) method to improve the training of unified autoregressive generation frameworks under few-shot NER scenarios.'}",http://arxiv.org/pdf/2305.11791
-Few-Shot Composition Learning for Image Retrieval with Prompt Tuning,Junda Wu,"We study the problem of composition learning for image retrieval, for which we learn to retrieve target images with search queries in the form of a composition of a reference image and a modification text that describes desired modifications of the image. Existing models of composition learning for image retrieval are generally built with large-scale datasets, demanding extensive training samples, i.e., query-target pairs, as supervision, which restricts their application for the scenario of few-shot learning with only few query-target pairs available. Recently, prompt tuning with frozen pretrained language models has shown remarkable performance when the amount of training data is limited. Inspired by this, we propose a prompt tuning mechanism with the pretrained CLIP model for the task of few-shot composition learning for image retrieval. Specifically, we regard the representation of the reference image as a trainable visual prompt, prefixed to the embedding of the text sequence. One challenge is to efficiently train visual prompt with few-shot samples. To deal with this issue, we further propose a self-upervised auxiliary task via ensuring that the reference image can retrieve itself when no modification information is given from the text, which facilitates training for the visual prompt, while not requiring additional annotations for query-target pairs. Experiments on multiple benchmarks show that our proposed model can yield superior performance when trained with only few query-target pairs.","{'model': 'tldr@v2.0.0', 'text': 'A self-upervised auxiliary task via ensuring that the reference image can retrieve itself when no modification information is given from the text, which facilitates training for the visual prompt, while not requiring additional annotations for query-target pairs is proposed.'}",https://ojs.aaai.org/index.php/AAAI/article/download/25597/25369
-Image-Object-Specific Prompt Learning for Few-Shot Class-Incremental Learning,In-Ug Yoon,"While many FSCIL studies have been undertaken, achieving satisfactory performance, especially during incremental sessions, has remained challenging. One prominent challenge is that the encoder, trained with an ample base session training set, often underperforms in incremental sessions. In this study, we introduce a novel training framework for FSCIL, capitalizing on the generalizability of the Contrastive Language-Image Pre-training (CLIP) model to unseen classes. We achieve this by formulating image-object-specific (IOS) classifiers for the input images. Here, an IOS classifier refers to one that targets specific attributes (like wings or wheels) of class objects rather than the image's background. To create these IOS classifiers, we encode a bias prompt into the classifiers using our specially designed module, which harnesses key-prompt pairs to pinpoint the IOS features of classes in each session. From an FSCIL standpoint, our framework is structured to retain previous knowledge and swiftly adapt to new sessions without forgetting or overfitting. This considers the updatability of modules in each session and some tricks empirically found for fast convergence. Our approach consistently demonstrates superior performance compared to state-of-the-art methods across the miniImageNet, CIFAR100, and CUB200 datasets. Further, we provide additional experiments to validate our learned model's ability to achieve IOS classifiers. We also conduct ablation studies to analyze the impact of each module within the architecture.","{'model': 'tldr@v2.0.0', 'text': 'A novel training framework for FSCIL is introduced, capitalizing on the generalizability of the Contrastive Language-Image Pre-training (CLIP) model to unseen classes, by formulating image-object-specific (IOS) classifiers for the input images.'}",https://arxiv.org/pdf/2309.02833
-Participatory Learning Strategies for One-Shot Instruction Sessions,Erin Durham,"Ø Elaboration: “How does this concept relate to other concepts?” Ø Distinctiveness: “What are the key differences between this concept and others? Ø Personal: “How does this concept relate to the student’s personal experiences?” Ø Application: “How are students expected to apply this concept?” -Stephen Chew, PhD, Cognitive Psychologist Deep Learning Strategies Cognitive psychologists have found that students can gain a deeper understanding of lesson concepts when instructors focus on: Post-It Note: Personal PostIt notes are placed at each desk and students are asked in the beginning of class to write a question they have about the library or their research. Notes are collected and reviewed during group work and then answered throughout the class session. This allows for student centered, relevant teaching. Group Brainstorm: Elaboration The class is divided into groups of 4-6 students and given a research topic to brainstorm for 5 minutes. Each group is given a white board with four questions that prompt discussion about the topic. The activity helps students expand their thinking and practice the skill of refining a topic to more narrow focus. Teaching and Learning Services http://www.lib.umd.edu/tl/first-year/engl-101 Jigsaw Formation: Distinctiveness The Jigsaw formation (developed by Elliot Aronson) engages deep learning by assigning students a portion of the class assignment to teach others. 1) Students are assigned a numbered group 1-4 and given sources to skim and evaluate for credibility. They are asked to become mini experts of those articles so they can teach their fellow classmates. 2) A student from each number group is assigned to a new group A-D and each shares her “expert knowledge” with the new group. The Jigsaw formation will be incorporated into select one-shot sessions in Fall 2017.","{'model': 'tldr@v2.0.0', 'text': 'The Jigsaw formation engages deep learning by assigning students a portion of the class assignment to teach others and the Post-It Note helps students expand their thinking and practice the skill of refining a topic to more narrow focus.'}",
-Overcoming Catastrophic Forgetting in Zero-Shot Cross-Lingual Generation,Tu Vu,"In this paper, we explore the challenging problem of performing a generative task in a target language when labeled data is only available in English, using summarization as a case study. We assume a strict setting with no access to parallel data or machine translation and find that common transfer learning approaches struggle in this setting, as a generative multilingual model fine-tuned purely on English catastrophically forgets how to generate non-English. Given the recent rise of parameter-efficient adaptation techniques, we conduct the first investigation into how one such method, prompt tuning (Lester et al., 2021), can overcome catastrophic forgetting to enable zero-shot cross-lingual generation. Our experiments show that parameter-efficient prompt tuning provides gains over standard fine-tuning when transferring between less-related languages, e.g., from English to Thai. However, a significant gap still remains between these methods and fully-supervised baselines. To improve cross-lingual transfer further, we explore several approaches, including: (1) mixing in unlabeled multilingual data, and (2) explicitly factoring prompts into recombinable language and task components. Our approaches can provide further quality gains, suggesting that robust zero-shot cross-lingual generation is within reach.","{'model': 'tldr@v2.0.0', 'text': 'This is the first investigation into how one such method, prompt tuning, can overcome catastrophic forgetting to enable zero-shot cross-lingual generation, and shows that parameter-efficient prompt tuning provides gains over standard fine-tuning when transferring between less-related languages, e.g., from English to Thai.'}",http://arxiv.org/pdf/2205.12647
-Nearest Neighbor Zero-Shot Inference,Weijia Shi,"Retrieval-augmented language models (LMs) use non-parametric memory to substantially outperform their non-retrieval counterparts on perplexity-based evaluations, but it is an open question whether they achieve similar gains in few- and zero-shot end-task accuracy. We extensively study one such model, the k-nearest neighbor LM (kNN-LM), showing that the gains marginally transfer. The main challenge is to achieve coverage of the verbalizer tokens that define the different end-task class labels. To address this challenge, we also introduce kNN-Prompt, a simple and effective kNN-LM with automatically expanded fuzzy verbalizers (e.g. to expand “terrible” to also include “silly” and other task-specific synonyms for sentiment classification). Across nine diverse end-tasks, using kNN-Prompt with GPT-2 large yields significant performance boosts over strong zeroshot baselines (13.4% absolute improvement over the base LM on average). We also show that other advantages of non-parametric augmentation hold for end tasks; kNN-Prompt is effective for domain adaptation with no further training, and gains increase with the size of the retrieval model.","{'model': 'tldr@v2.0.0', 'text': 'KNN-Prompt is introduced, a simple and effective kNN-LM with automatically expanded fuzzy verbalizers that is effective for domain adaptation with no further training, and gains increase with the size of the retrieval model.'}",https://arxiv.org/pdf/2205.13792
-Balanced Distributed Augmentation for Multi-Label Few Shot Learning with Prototypical Network,Hamza Haruna Mohammed,"Many methods have been presented as a few shot learners in order to enhance few-shot learners. Some of these methods involve routine-based pre-trained language models and novel pipeline for automating the prompt generation. In this study, we propose a new evenly distributed data augmentation technique, which generates samples according to the probabilistic distribution of the relationship of each label with the mean of a label group. In the labeling phase, we present a semantic sentiment analysis approach in order to increase the realism of the data, in a more semantic augmentation way. The results show that this approach improves the few shot learners. In addition to this, we compare our adaptation approach to other traditional problem transformation methods. The newly developed approach outperforms these traditional methods, especially when the classifier learns from a limited number of samples.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a new evenly distributed data augmentation technique, which generates samples according to the probabilistic distribution of the relationship of each label with the mean of a label group, and presents a semantic sentiment analysis approach in order to increase the realism of the data.'}",
-KECP: Knowledge Enhanced Contrastive Prompting for Few-shot Extractive Question Answering,J. Wang,"Extractive Question Answering (EQA) is one of the most essential tasks in Machine Reading Comprehension (MRC), which can be solved by fine-tuning the span selecting heads of Pre-trained Language Models (PLMs). However, most existing approaches for MRC may perform poorly in the few-shot learning scenario. To solve this issue, we propose a novel framework named Knowledge Enhanced Contrastive Prompt-tuning (KECP). Instead of adding pointer heads to PLMs, we introduce a seminal paradigm for EQA that transforms the task into a non-autoregressive Masked Language Modeling (MLM) generation problem. Simultaneously, rich semantics from the external knowledge base (KB) and the passage context support enhancing the query’s representations. In addition, to boost the performance of PLMs, we jointly train the model by the MLM and contrastive learning objectives. Experiments on multiple benchmarks demonstrate that our method consistently outperforms state-of-the-art approaches in few-shot settings by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a seminal paradigm for EQA that transforms the task into a non-autoregressive Masked Language Modeling (MLM) generation problem and jointly train the model by the MLM and contrastive learning objectives.'}",http://arxiv.org/pdf/2205.03071
-Pre-trained Token-replaced Detection Model as Few-shot Learner,Zicheng Li,"Pre-trained masked language models have demonstrated remarkable ability as few-shot learners. In this paper, as an alternative, we propose a novel approach to few-shot learning with pre-trained token-replaced detection models like ELECTRA. In this approach, we reformulate a classification or a regression task as a token-replaced detection problem. Specifically, we first define a template and label description words for each task and put them into the input to form a natural language prompt. Then, we employ the pre-trained token-replaced detection model to predict which label description word is the most original (i.e., least replaced) among all label description words in the prompt. A systematic evaluation on 16 datasets demonstrates that our approach outperforms few-shot learners with pre-trained masked language models in both one-sentence and two-sentence learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach to few-shot learning with pre-trained token-replaced detection models like ELECTRA, and demonstrates that this approach outperforms few- shot learners withPre-trained masked language models in both one-sentence and two- sentence learning tasks.'}",http://arxiv.org/pdf/2203.03235
-Cross-Lingual Retrieval Augmented Prompt for Low-Resource Languages,Ercong Nie,"Multilingual Pretrained Language Models (MPLMs) have shown their strong multilinguality in recent empirical cross-lingual transfer studies. In this paper, we propose the Prompts Augmented by Retrieval Crosslingually (PARC) pipeline to improve the zero-shot performance on low-resource languages (LRLs) by augmenting the context with semantically similar sentences retrieved from a high-resource language (HRL) as prompts. PARC improves the zero-shot performance on three downstream tasks (binary sentiment classification, topic categorization and natural language inference) with multilingual parallel test sets across 10 LRLs covering 6 language families in both unlabeled settings (+5.1%) and labeled settings (+16.3%). PARC-labeled also outperforms the finetuning baseline by 3.7%. We find a significant positive correlation between cross-lingual transfer performance on one side, and the similarity between the high- and low-resource languages as well as the amount of low-resource pretraining data on the other side. A robustness analysis suggests that PARC has the potential to achieve even stronger performance with more powerful MPLMs.","{'model': 'tldr@v2.0.0', 'text': 'A robustness analysis suggests that PARC has the potential to achieve even stronger performance with more powerful MPLMs, and a significant positive correlation between cross-lingual transfer performance on one side, and the similarity between the high- and low-resource languages as well as the amount of low- resource pretraining data on the other side.'}",https://arxiv.org/pdf/2212.09651
-InDiReCT: Language-Guided Zero-Shot Deep Metric Learning for Images,Konstantin Kobs,"Common Deep Metric Learning (DML) datasets specify only one notion of similarity, e.g., two images in the Cars196 dataset are deemed similar if they show the same car model. We argue that depending on the application, users of image retrieval systems have different and changing similarity notions that should be incorporated as easily as possible. Therefore, we present Language-Guided Zero-Shot Deep Metric Learning (LanZ-DML) as a new DML setting in which users control the properties that should be important for image representations without training data by only using natural language. To this end, we propose InDiReCT (Image representations using Dimensionality Reduction on CLIP embedded Texts), a model for LanZ-DML on images that exclusively uses a few text prompts for training. InDiReCT utilizes CLIP as a fixed feature extractor for images and texts and transfers the variation in text prompt embeddings to the image embedding space. Extensive experiments on five datasets and overall thirteen similarity notions show that, despite not seeing any images during training, InDiReCT performs better than strong baselines and approaches the performance of fully-supervised models. An analysis reveals that InDiReCT learns to focus on regions of the image that correlate with the desired similarity notion, which makes it a fast to train and easy to use method to create custom embedding spaces only using natural language.","{'model': 'tldr@v2.0.0', 'text': 'Language-Guided Zero-Shot Deep Metric Learning (LanZ-DML) is presented as a new DML setting in which users control the properties that should be important for image representations without training data by only using natural language.'}",https://arxiv.org/pdf/2211.12760
-LiST: Lite Self-training Makes Efficient Few-shot Learners,Yaqing Wang,"We present a new method LiST 1 for efficient fine-tuning of large pre-trained language models (PLMs) in few-shot learning settings. LiST significantly improves over recent methods that adopt prompt-tuning using two key techniques. The first one is the use of self-training to leverage large amounts of unlabeled data for prompt-tuning to significantly boost the model performance in few-shot settings. We use self-training in conjunction with meta-learning for re-weighting noisy pseudo-prompt labels. However, traditional self-training is expensive as it requires updating all the model parameters repetitively. Therefore, we use a second technique for light-weight fine-tuning where we introduce a small number of task-specific adapter parameters that are fine-tuned during self-training while keeping the PLM encoder frozen. This also significantly reduces the overall model footprint across several tasks that can now share a common PLM encoder as backbone for inference. Combining the above techniques, LiST not only improves the model performance for few-shot learning on target domains but also reduces the model memory footprint. We present a comprehensive study on six NLU tasks to validate the effectiveness of LiST . The results show that LiST improves by 35% over classic fine-tuning methods and 6% over prompt-tuning with 96% reduction in number of trainable parameters when fine-tuned with no more than 30 labeled examples from each target domain.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive study on six NLU tasks to validate the effectiveness of LiST and shows that LiST improves by 35% over classic fine- Tuning methods and 6% over prompt-tuning with 96% reduction in number of trainable parameters when flne-tuned with no more than 30 labeled examples from each target domain.'}",
-Prompt-Based Multi-Modal Image Segmentation,Timo Lüddecke,"Image segmentation is usually addressed by training a model for a fixed set of object classes. Incorporating additional classes or more complex queries later is expensive as it requires re-training the model on a dataset that encompasses these expressions. Here we propose a system that can generate image segmentations based on arbitrary prompts at test time. A prompt can be either a text or an image. This approach enables us to create a unified model (trained once) for three common segmentation tasks, which come with distinct challenges: referring expression segmentation, zero-shot segmentation and one-shot segmentation. We build upon the CLIP model as a backbone which we extend with a transformer-based decoder that enables dense prediction. After training on an extended version of the PhraseCut dataset, our system generates a binary segmentation map for an image based on a free-text prompt or on an additional image expressing the query. Different variants of the latter image-based prompts are analyzed in detail. This novel hybrid input allows for dynamic adaptation not only to the three segmentation tasks mentioned above, but to any binary segmentation task where a text or image query can be formulated. Finally, we find our system to adapt well to generalized queries involving affordances or properties. Source code: https://eckerlab.org/code/clipseg","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a system that can generate image segmentations based on arbitrary prompts at test time with a transformer-based decoder that enables dense prediction and allows for dynamic adaptation to generalized queries involving affordances or properties.'}",
-Jurassic is (almost) All You Need: Few-Shot Meaning-to-Text Generation for Open-Domain Dialogue,Lena Reed,,"{'model': 'tldr@v2.0.0', 'text': 'These are the first results demonstrating that few-shot semantic prompt-based learning can create NLGs that generalize to new domains, and produce high-quality, semantically-controlled, conversational responses directly from meaning representations.'}",https://arxiv.org/pdf/2110.08094
-Prompt scoring system for dialogue summarization using GPT-3,George Prodan,"Recent results in language processing show that language models are capable of performing several natural language tasks without the need of supervised learning. A challenging task for pre-trained language models is dialogue summarization. One way of generating summaries is engineering prompt templates for few-shot training. However, a static approach of creating prompts leads to unreliable outcomes between different classes of dialogues. Focusing on the dialogues structure properties we propose a scoring system to improve the few-shot training performances. We build tuned prompts composed by the highest scored dialogue samples. Our evaluation based on ROUGE scores and human evaluation shows that there is an improvement for the experiments in which we use the score system. All experiments are performed within the framework of the GPT-3 API. We use different engines for comparison. Moreover, the human evaluation we conducted showed that the number of failures decreased by 11\% after applying our scoring system.","{'model': 'tldr@v2.0.0', 'text': 'Focusing on the dialogues structure properties, this work proposes a scoring system to improve the few-shot training performances of dialogue summarization and builds tuned prompts composed by the highest scored dialogue samples.'}",https://www.techrxiv.org/articles/preprint/Prompt_scoring_system_for_dialogue_summarization_using_GPT-3/16652392/2/files/35289613.pdf
-Inverse is Better! Fast and Accurate Prompt for Slot Tagging,,"Prompting methods recently achieve impres- 001 sive success in few-shot learning. These meth- 002 ods embed input samples with prompt sentence 003 pieces and decode label-related tokens to map 004 samples to the label. However, such a paradigm 005 is very inefficient for the task of slot tagging. 006 Because the slot tagging samples are multiple 007 consecutive words in a sentence, the prompting 008 methods have to enumerate all n-grams token 009 span to find all the possible slots, which greatly 010 slows down the prediction. To tackle this, we 011 introduce an inverse paradigm for prompting. 012 Different from the classic prompts map tokens 013 to labels, we reversely predict slot values given 014 slot types. Such inverse prompting only re- 015 quires a one-turn prediction for each slot type 016 and greatly speeds up the prediction. Besides, 017 we propose a novel Iterative Prediction Strat- 018 egy, from which the model learns to refine pre- 019 dictions by considering the relations between 020 different slot types. We find, somewhat surpris- 021 ingly, the proposed method not only predicts 022 faster, but also significantly improves the effect 023 (improve over 6 . 1 F1-scores on 10-shot setting) 024 and achieves new state-of-the-art performance. 025","{'model': 'tldr@v2.0.0', 'text': 'The proposed inverse prompting method not only predicts faster, but also significantly improves the effect 023 (improve over 6 . 1 F1-scores on 10-shot setting) 024 and achieves new state-of-the-art performance.'}",
-Is a prompt and a few samples all you need? Using GPT-4 for data augmentation in low-resource classification tasks,Anders Giovanni Møller,"Obtaining and annotating data can be expensive and time-consuming, especially in complex, low-resource domains. We use GPT-4 and ChatGPT to augment small labeled datasets with synthetic data via simple prompts, in three different classification tasks with varying complexity. For each task, we randomly select a base sample of 500 texts to generate 5,000 new synthetic samples. We explore two augmentation strategies: one that preserves original label distribution and another that balances the distribution. Using a progressively larger training sample size, we train and evaluate a 110M parameter multilingual language model on the real and synthetic data separately. We also test GPT-4 and ChatGPT in a zero-shot setting on the test sets. We observe that GPT-4 and ChatGPT have strong zero-shot performance across all tasks. We find that data augmented with synthetic samples yields a good downstream performance, and particularly aids in low-resource settings, such as in identifying rare classes. Human-annotated data exhibits a strong predictive power, overtaking synthetic data in two out of the three tasks. This finding highlights the need for more complex prompts for synthetic datasets to consistently surpass human-generated ones.","{'model': 'tldr@v2.0.0', 'text': 'This work uses GPT-4 and ChatGPT to augment small labeled datasets with synthetic data via simple prompts, in three different classification tasks with varying complexity, and finds that data augmented with synthetic samples yields a good downstream performance, and particularly aids in low-resource settings, such as in identifying rare classes.'}",http://arxiv.org/pdf/2304.13861
-Structure Pretraining and Prompt Tuning for Knowledge Graph Transfer,Wen Zhang,"Knowledge graphs (KG) are essential background knowledge providers in many tasks. When designing models for KG-related tasks, one of the key tasks is to devise the Knowledge Representation and Fusion (KRF) module that learns the representation of elements from KGs and fuses them with task representations. While due to the difference of KGs and perspectives to be considered during fusion across tasks, duplicate and ad hoc KRF modules design are conducted among tasks. In this paper, we propose a novel knowledge graph pretraining model KGTransformer that could serve as a uniform KRF module in diverse KG-related tasks. We pretrain KGTransformer with three self-supervised tasks with sampled sub-graphs as input. For utilization, we propose a general prompt-tuning mechanism regarding task data as a triple prompt to allow flexible interactions between task KGs and task data. We evaluate pretrained KGTransformer on three tasks, triple classification, zero-shot image classification, and question answering. KGTransformer consistently achieves better results than specifically designed task models. Through experiments, we justify that the pretrained KGTransformer could be used off the shelf as a general and effective KRF module across KG-related tasks. The code and datasets are available at https://github.com/zjukg/KGTransformer.","{'model': 'tldr@v2.0.0', 'text': 'A novel knowledge graph pretraining model KGTransformer is proposed that could serve as a uniform KRF module in diverse KG-related tasks and consistently achieves better results than specifically designed task models.'}",https://arxiv.org/pdf/2303.03922
-Residual Prompt Tuning: Improving Prompt Tuning with Residual Reparameterization,Anastasia Razdaibiedina,"Prompt tuning is one of the successful approaches for parameter-efficient tuning of pre-trained language models. Despite being arguably the most parameter-efficient (tuned soft prompts constitute<0.1% of total parameters), it typically performs worse than other efficient tuning methods and is quite sensitive to hyper-parameters. In this work, we introduce Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning. We propose to reparameterize soft prompt embeddings using a shallow network with a residual connection. Our experiments show that Residual Prompt Tuning significantly outperforms prompt tuning on SuperGLUE benchmark. Notably, our method reaches +7 points improvement over prompt tuning with T5-Base and allows to reduce the prompt length by 10x without hurting performance. In addition, we show that our approach is robust to the choice of learning rate and prompt initialization, and is effective in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning and proposes to reparameterize soft prompt embeddings using a shallow network with a residual connection.'}",http://arxiv.org/pdf/2305.03937
-DS4DH at MEDIQA-Chat 2023: Leveraging SVM and GPT-3 Prompt Engineering for Medical Dialogue Classification and Summarization,Boya Zhang,"This paper presents the results of the Data Science for Digital Health (DS4DH) group in the MEDIQA-Chat Tasks at ACL-ClinicalNLP 2023. Our study combines the power of a classical machine learning method, Support Vector Machine, for classifying medical dialogues, along with the implementation of one-shot prompts using GPT-3.5. We employ dialogues and summaries from the same category as prompts to generate summaries for novel dialogues. Our findings exceed the average benchmark score, offering a robust reference for assessing performance in this field.","{'model': 'tldr@v2.0.0', 'text': 'This study combines the power of a classical machine learning method, Support Vector Machine, for classifying medical dialogues, along with the implementation of one-shot prompts using GPT-3.5 to generate summaries for novel dialogues.'}",https://access.archive-ouverte.unige.ch/access/metadata/290c4289-0017-45ec-baa9-ff2fdd7948f9/download
-Soft Prompt Tuning for Augmenting Dense Retrieval with Large Language Models,Zhiyuan Peng,"Dense retrieval (DR) converts queries and documents into dense embeddings and measures the similarity between queries and documents in vector space. One of the challenges in DR is the lack of domain-specific training data. While DR models can learn from large-scale public datasets like MS MARCO through transfer learning, evidence shows that not all DR models and domains can benefit from transfer learning equally. Recently, some researchers have resorted to large language models (LLMs) to improve the zero-shot and few-shot DR models. However, the hard prompts or human-written prompts utilized in these works cannot guarantee the good quality of generated weak queries. To tackle this, we propose soft prompt tuning for augmenting DR (SPTAR): For each task, we leverage soft prompt-tuning to optimize a task-specific soft prompt on limited ground truth data and then prompt the LLMs to tag unlabeled documents with weak queries, yielding enough weak document-query pairs to train task-specific dense retrievers. We design a filter to select high-quality example document-query pairs in the prompt to further improve the quality of weak tagged queries. To the best of our knowledge, there is no prior work utilizing soft prompt tuning to augment DR models. The experiments demonstrate that SPTAR outperforms the unsupervised baselines BM25 and the recently proposed LLMs-based augmentation method for DR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes soft prompt tuning for augmenting DR (SPTAR) by leveraging soft prompt-tuning to optimize a task-specific soft prompt on limited ground truth data and prompts the LLMs to tag unlabeled documents with weak queries, yielding enough weak document-query pairs to train task- specific dense retrievers.'}",https://arxiv.org/pdf/2307.08303
-Self-Prompting Large Vision Models for Few-Shot Medical Image Segmentation,Qi Wu,"Recent advancements in large foundation models have shown promising potential in the medical industry due to their flexible prompting capability. One such model, the Segment Anything Model (SAM), a prompt-driven segmentation model, has shown remarkable performance improvements, surpassing state-of-the-art approaches in medical image segmentation. However, existing methods primarily rely on tuning strategies that require extensive data or prior prompts tailored to the specific task, making it particularly challenging when only a limited number of data samples are available. In this paper, we propose a novel perspective on self-prompting in medical vision applications. Specifically, we harness the embedding space of SAM to prompt itself through a simple yet effective linear pixel-wise classifier. By preserving the encoding capabilities of the large model, the contextual information from its decoder, and leveraging its interactive promptability, we achieve competitive results on multiple datasets (i.e. improvement of more than 15% compared to fine-tuning the mask decoder using a few images).","{'model': 'tldr@v2.0.0', 'text': 'A novel perspective on self-prompting in medical vision applications is proposed by harnessing the embedding space of SAM to prompt itself through a simple yet effective linear pixel-wise classifier.'}",https://arxiv.org/pdf/2308.07624
-CIPTA: Contrastive-based Iterative Prompt-tuning Using Text Annotation from Large Language Models,Yuan Yan,"In recent years, public opinion analysis has become increasingly important due to the widespread use of social media platforms and the growing influence of online information on public security. Prompt tuning, a typical few-shot learning method, ensures that the model quickly adapts to opinion analysis with different classification rules. However, existing prompt tuning for opinion analysis cannot guarantee the effectiveness of the model in zero-shot or one-shot cases. In this study, we propose the Contrastive-based Iterative Prompt-tuning method using Text-Annotation from Large Language Models (LLMs), CIPTA, for low-resource public opinion analysis. Specifically, with a small amount of manually labeled data, CIPTA leverages the knowledge from LLMs to text annotation and utilizes unsupervised contrastive embedding training to optimize text representation. Based on the prompt tuning method and the iterative training over unlabeled data, the model further utilizes the knowledge from the pre-training stage. Experiment results on tweet data show that our CIPTA achieves encouraging performance in public opinion analysis.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes the Contrastive-based Iterative Prompt-tuning method using Text-Annotation from Large Language Models, CIPTA, for low-resource public opinion analysis that leverages the knowledge from LLMs to text annotation and utilizes unsupervised contrastive embedding training to optimize text representation.'}",
-Unleashing the potential of prompt engineering in Large Language Models: a comprehensive review,Banghao Chen,"This paper delves into the pivotal role of prompt engineering in unleashing the capabilities of Large Language Models (LLMs). Prompt engineering is the process of structuring input text for LLMs and is a technique integral to optimizing the efficacy of LLMs. This survey elucidates foundational principles of prompt engineering, such as role-prompting, one-shot, and few-shot prompting, as well as more advanced methodologies such as the chain-of-thought and tree-of-thoughts prompting. The paper sheds light on how external assistance in the form of plugins can assist in this task, and reduce machine hallucination by retrieving external knowledge. We subsequently delineate prospective directions in prompt engineering research, emphasizing the need for a deeper understanding of structures and the role of agents in Artificial Intelligence-Generated Content (AIGC) tools. We discuss how to assess the efficacy of prompt methods from different perspectives and using different methods. Finally, we gather information about the application of prompt engineering in such fields as education and programming, showing its transformative potential. This comprehensive survey aims to serve as a friendly guide for anyone venturing through the big world of LLMs and prompt engineering.",,
-Multiple Brain Abscesses by an Air Gun Shot: A Case Report,M. Jamali,"Abstract Introduction Low-velocity penetrating brain injury is not prevalent. In some conditions such as childhood, and with the penetration of a pellet in weak spots of skull, low-velocity penetrating brain injury is expected; however, high-velocity projectiles have also been reported as the cause of severe brain injuries. One of the complications of penetrating brain injury is infection, in which different types of microorganisms play a role. The Streptococcus genus is the leading cause of abscess formation in non-traumatic patients. Multiple brain abscesses are not common. Case Presentation A 10-year-old boy with penetrating brain injury caused by an air gun pellet, who developed signs and symptoms of high intracranial pressure 18 days after the trauma. After the imaging scans and the detection of multiple brain abscesses and severe brain edema, prompt surgical intervention was performed for all three lesions in a single operation. The culture of a pus specimen was positive for Streptococcus species, and, with adequate antibiotic therapy, the patient was discharged from the hospital in good condition. Conclusion Brain injury with air gun shot is not prevalent. The penetration of a low-velocity air gun pellet in weak points of the skull (such as the orbit, the squamous portion of the temporal bone, and the cranial suture), specially in children, can cause significant brain injuries.","{'model': 'tldr@v2.0.0', 'text': 'Brain injury with air gun shot is not prevalent; however, the penetration of a low-velocity air gun pellet in weak points of the skull (such as the orbit, the squamous portion of the temporal bone, and the cranial suture) can cause significant brain injuries.'}",
-Multi-Mask Label Mapping for Prompt-Based Learning,Jirui Qi,"Prompt-based Learning has shown significant success in few-shot classification. The mainstream approach is to concatenate a template for the input text to transform the classification task into a cloze-type task where label mapping plays an important role in finding the ground-truth labels. While current label mapping methods only use the contexts in one single input, it could be crucial if wrong information is contained in the text. Specifically, it is proved in recent work that even the large language models like BERT/RoBERTa make classification decisions heavily dependent on a specific keyword regardless of the task or the context. Such a word is referred to as a lexical cue and if a misleading lexical cue is included in the instance it will lead the model to make a wrong prediction. We propose a multi-mask prompt-based approach with Multi-Mask Label Mapping (MMLM) to reduce the impact of misleading lexical cues by allowing the model to exploit multiple lexical cues. To satisfy the conditions of few-shot learning, an instance augmentation approach for the cloze-type model is proposed and the misleading cues are gradually excluded through training. We demonstrate the effectiveness of MMLM by both theoretical analysis and empirical studies, and show that MMLM outperforms other existing label mapping approaches.","{'model': 'tldr@v2.0.0', 'text': 'To satisfy the conditions of few-shot learning, an instance augmentation approach for the cloze-type model is proposed and the misleading cues are gradually excluded through training to demonstrate the effectiveness of MMLM by both theoretical analysis and empirical studies, and it is shown that M MLM outperforms other existing label mapping approaches.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26579/26351
-Prompts Can Play Lottery Tickets Well: Achieving Lifelong Information Extraction via Lottery Prompt Tuning,Zujie Liang,"Thanks to the recent success of Pre-trained Language Models (PLMs), it has become a promising research direction to develop a universal model (UIE) that can solve all typical information extraction tasks within one generative framework. Nonetheless, in real-world scenarios of UIE applications, new data of different IE tasks and domains usually come in a stream over time. A desirable UIE system should be capable of continually learning new tasks without forgetting old ones, thereby allowing knowledge and functionalities expansion without re-training the whole system. In this paper, we study the UIE system under a more challenging yet practical scenario, i.e., “lifelong learning” settings, to evaluate its abilities in three aspects, including knowledge sharing and expansion, catastrophic forgetting prevention, and rapid generalization on few-shot and unseen tasks.To achieve these three goals, we present a novel parameter- and deployment-efficient prompt tuning method namely Lottery Prompt Tuning (LPT).LPT freezes the PLM’s parameters and sequentially learns compact pruned prompt vectors for each task leveraging a binary prompt mask, while keeping the prompt parameters selected by the previous tasks insusceptible.Furthermore, we use a simple yet effective method to perform mask selection and show the powerful transferability of Lottery Prompts to novel tasks.Extensive experiments demonstrate that LPT consistently sets state-of-the-art performance on multiple lifelong learning settings of UIE, including task-incremental setting on seen tasks, few-shot adaptation, and zero-shot generalization on novel tasks.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments demonstrate that LPT consistently sets state-of-the-art performance on multiple lifelong learning settings of UIE, including task-incremental setting on seen tasks, few-shot adaptation, and zero-shot generalization on novel tasks.'}",https://aclanthology.org/2023.acl-long.16.pdf
-LLM Comparative Assessment: Zero-shot NLG Evaluation through Pairwise Comparisons using Large Language Models,Adian Liusie,"Current developments in large language models (LLMs) have enabled impressive zero-shot capabilities across various natural language tasks. An interesting application of these systems is in the automated assessment of natural language generation (NLG), a highly challenging area with great practical benefit. In this paper, we explore two options for exploiting the emergent abilities of LLMs for zero-shot NLG assessment: absolute score prediction, and comparative assessment which uses relative comparisons between pairs of candidates. Though comparative assessment has not been extensively studied in NLG assessment, we note that humans often find it more intuitive to compare two options rather than scoring each one independently. This work examines comparative assessment from multiple perspectives: performance compared to absolute grading; positional biases in the prompt; and efficient ranking in terms of the number of comparisons. We illustrate that LLM comparative assessment is a simple, general and effective approach for NLG assessment. For moderate-sized open-source LLMs, such as FlanT5 and Llama2-chat, comparative assessment is superior to prompt scoring, and in many cases can achieve performance competitive with state-of-the-art methods. Additionally, we demonstrate that LLMs often exhibit strong positional biases when making pairwise comparisons, and we propose debiasing methods that can further improve performance.","{'model': 'tldr@v2.0.0', 'text': 'It is illustrated that LLM comparative assessment is a simple, general and effective approach for NLG assessment, and for moderate-sized open-source LLMs, superior to prompt scoring, and in many cases can achieve performance competitive with state-of-the-art methods.'}",
-Revisiting Large Language Models as Zero-shot Relation Extractors,Guozheng Li,"Relation extraction (RE) consistently involves a certain degree of labeled or unlabeled data even if under zero-shot setting. Recent studies have shown that large language models (LLMs) transfer well to new tasks out-of-the-box simply given a natural language prompt, which provides the possibility of extracting relations from text without any data and parameter tuning. This work focuses on the study of exploring LLMs, such as ChatGPT, as zero-shot relation extractors. On the one hand, we analyze the drawbacks of existing RE prompts and attempt to incorporate recent prompt techniques such as chain-of-thought (CoT) to improve zero-shot RE. We propose the summarize-and-ask (\textsc{SumAsk}) prompting, a simple prompt recursively using LLMs to transform RE inputs to the effective question answering (QA) format. On the other hand, we conduct comprehensive experiments on various benchmarks and settings to investigate the capabilities of LLMs on zero-shot RE. Specifically, we have the following findings: (i) \textsc{SumAsk} consistently and significantly improves LLMs performance on different model sizes, benchmarks and settings; (ii) Zero-shot prompting with ChatGPT achieves competitive or superior results compared with zero-shot and fully supervised methods; (iii) LLMs deliver promising performance in extracting overlapping relations; (iv) The performance varies greatly regarding different relations. Different from small language models, LLMs are effective in handling challenge none-of-the-above (NoTA) relation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the summarize-and-ask prompting, a simple prompt recursively using LLMs to transform RE inputs to the effective question answering (QA) format and investigates the capabilities of LLMs on zero-shot RE.'}",https://arxiv.org/pdf/2310.05028
-Tuning Multi-mode Token-level Prompt Alignment across Modalities,Dongsheng Wang,"Advancements in prompt tuning of vision-language models have underscored their potential in enhancing open-world visual concept comprehension. However, prior works only primarily focus on single-mode (only one prompt for each modality) and holistic level (image or sentence) semantic alignment, which fails to capture the sample diversity, leading to sub-optimal prompt discovery. To address the limitation, we propose a multi-mode token-level tuning framework that leverages the optimal transportation to learn and align a set of prompt tokens across modalities. Specifically, we rely on two essential factors: 1) multi-mode prompts discovery, which guarantees diverse semantic representations, and 2) token-level alignment, which helps explore fine-grained similarity. Consequently, the similarity can be calculated as a hierarchical transportation problem between the modality-specific sets. Extensive experiments on popular image recognition benchmarks show the superior generalization and few-shot abilities of our approach. The qualitative analysis demonstrates that the learned prompt tokens have the ability to capture diverse visual concepts.","{'model': 'tldr@v2.0.0', 'text': 'A multi-mode token-level tuning framework that leverages the optimal transportation to learn and align a set of prompt tokens across modalities and demonstrates that the learned prompt tokens have the ability to capture diverse visual concepts.'}",https://arxiv.org/pdf/2309.13847
-RPLKG: Robust Prompt Learning with Knowledge Graph,Yewon Kim,"Large-scale pre-trained models have been known that they are transferable, and they generalize well on the unseen dataset. Recently, multimodal pre-trained models such as CLIP show significant performance improvement in diverse experiments. However, when the labeled dataset is limited, the generalization of a new dataset or domain is still challenging. To improve the generalization performance on few-shot learning, there have been diverse efforts, such as prompt learning and adapter. However, the current few-shot adaptation methods are not interpretable, and they require a high computation cost for adaptation. In this study, we propose a new method, robust prompt learning with knowledge graph (RPLKG). Based on the knowledge graph, we automatically design diverse interpretable and meaningful prompt sets. Our model obtains cached embeddings of prompt sets after one forwarding from a large pre-trained model. After that, model optimizes the prompt selection processes with GumbelSoftmax. In this way, our model is trained using relatively little memory and learning time. Also, RPLKG selects the optimal interpretable prompt automatically, depending on the dataset. In summary, RPLKG is i) interpretable, ii) requires small computation resources, and iii) easy to incorporate prior human knowledge. To validate the RPLKG, we provide comprehensive experimental results on few-shot learning, domain generalization and new class generalization setting. RPLKG shows a significant performance improvement compared to zero-shot learning and competitive performance against several prompt learning methods using much lower resources.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a new method, robust prompt learning with knowledge graph (RPLKG), which automatically design diverse interpretable and meaningful prompt sets and shows a significant performance improvement compared to zero-shot learning and competitive performance against several prompt learning methods using much lower resources.'}",http://arxiv.org/pdf/2304.10805
-[CLS] Token is All You Need for Zero-Shot Semantic Segmentation,Letian Wu,"In this paper, we propose an embarrassingly simple yet highly effective zero-shot semantic segmentation (ZS3) method, based on the pre-trained vision-language model CLIP. First, our study provides a couple of key discoveries: (i) the global tokens (a.k.a [CLS] tokens in Transformer) of the text branch in CLIP provide a powerful representation of semantic information and (ii) these text-side [CLS] tokens can be regarded as category priors to guide CLIP visual encoder pay more attention on the corresponding region of interest. Based on that, we build upon the CLIP model as a backbone which we extend with a One-Way [CLS] token navigation from text to the visual branch that enables zero-shot dense prediction, dubbed \textbf{ClsCLIP}. Specifically, we use the [CLS] token output from the text branch, as an auxiliary semantic prompt, to replace the [CLS] token in shallow layers of the ViT-based visual encoder. This one-way navigation embeds such global category prior earlier and thus promotes semantic segmentation. Furthermore, to better segment tiny objects in ZS3, we further enhance ClsCLIP with a local zoom-in strategy, which employs a region proposal pre-processing and we get ClsCLIP+. Extensive experiments demonstrate that our proposed ZS3 method achieves a SOTA performance, and it is even comparable with those few-shot semantic segmentation methods.","{'model': 'tldr@v2.0.0', 'text': 'This study provides a couple of key discoveries: (i) the global tokens of the text branch in CLIP provide a powerful representation of semantic information and (ii) these text-side [CLS] tokens can be regarded as category priors to guide CLIP visual encoder pay more attention on the corresponding region of interest.'}",http://arxiv.org/pdf/2304.06212
-MetricPrompt: Prompting Model as a Relevance Metric for Few-shot Text Classification,Hongyuan Dong,"Prompting methods have shown impressive performance in a variety of text mining tasks and applications, especially few-shot ones. Despite the promising prospects, the performance of prompting model largely depends on the design of prompt template and verbalizer. In this work, we propose MetricPrompt, which eases verbalizer design difficulty by reformulating few-shot text classification task into text pair relevance estimation task. MetricPrompt adopts prompting model as the relevance metric, further bridging the gap between Pre-trained Language Model's (PLM) pre-training objective and text classification task, making possible PLM's smooth adaption. Taking a training sample and a query one simultaneously, MetricPrompt captures cross-sample relevance information for accurate relevance estimation. We conduct experiments on three widely used text classification datasets across four few-shot settings. Results show that MetricPrompt outperforms manual verbalizer and other automatic verbalizer design methods across all few-shot settings, achieving new state-of-the-art (SOTA) performance.","{'model': 'tldr@v2.0.0', 'text': ""MetricPrompt is proposed, which eases verbalizer design difficulty by reformulating few-shot text classification task into text pair relevance estimation task, and adopts prompting model as the relevance metric, further bridging the gap between Pre-trained Language Model's (PLM) pre-training objective and text classificationtask.""}",https://arxiv.org/pdf/2306.08892
-Unsupervised Dual Modality Prompt Learning for Facial Expression Recognition,Muhammad Shahid,"A method of facial expression recognition using a vision language model is proposed. Recently vision-language models for example CLIP (Contrastive Language-Image Pre-training) models developed by OpenAI have achieved exceptional results on a variety of image recognition and retrieval tasks, exhibiting strong zero-shot performance. Transferable representations can be adapted through prompt tuning to a variety of downstream tasks. From the general knowledge stored in a pre-trained model, prompt tuning attempts to extract useful information for downstream tasks. In order to avoid time-consuming prompt engineering, recent works use a small amount of labeled data for adapting vision language models to downstream image recognition problems. However, requiring target datasets to be labeled may restrict their scalability. Moreover, we also note that adapting prompt learning techniques in only one branch of CLIP (vision or language) is suboptimal because it won't allow for the dynamic adjustment of both representation spaces on a downstream task. In this paper, we evaluated the performance of the CLIP model as a zero-shot face recognizer and proposed an Unsupervised Dual Modality Prompt Learning framework for Facial Expression Recognition. Our model tunes the prompts through learning text and visual prompts simultaneously to improve alignment between the linguistic and visual representations when labels are not provided for the target dataset. The experimental results on CK+, JAFFE, RAF-DB, and FER2013 datasets showed that our proposed method performs better compared with CLIP Zero-Shot and other unsupervised prompt-based learning methods for facial expression recognition tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper evaluated the performance of the CLIP model as a zero-shot face recognizer and proposed an Unsupervised Dual Modality Prompt Learning framework for Facial Expression Recognition that performs better compared with CLIP Zero-Shot and other unsupervised prompt-based learning methods for facial expression recognition tasks.'}",
-An Attempt at Zero-shot Ancient Documents Restoration Based on Diffusion Models,Hayata Kaneko,"The virtual restoration of ancient documents using deep learning is an emergency and an expected work. However, GANs-based image-to-image translation approaches hit the degradation data shortage, a hardness to build one-to-many restoration models, and a limitation for large deformation. In this study, we apply zero-shot restoration based on Diffusion models to ancient degraded documents, specifically, leverage inpainting of Denoing Diffusion Restoration Models (DDRM) for missing ancient characters. Furthermore, we introduce a noise masking method, which limits the attention area of predicted noise images in the reverse process. Noise masking forces DDRM to generate faithful objects following mask images, so that has high usability without re-training of deep neural networks. The zero-shot restoration and noise masking prompt GUI-connecting restoration of missing characters, leading to realizing a cooperative application with humans for ancient document restoration.","{'model': 'tldr@v2.0.0', 'text': 'This study applies zero-shot restoration based on Diffusion models to ancient degraded documents and introduces a noise masking method, which limits the attention area of predicted noise images in the reverse process, leading to cooperative application with humans for ancient document restoration.'}",
-Label-Aware Automatic Verbalizer for Few-Shot Text Classification,Thanakorn Thaminkaew,"Prompt-based learning has shown its effectiveness in few-shot text classification. One important factor in its success is a verbalizer, which translates output from a language model into a predicted class. Notably, the simplest and widely acknowledged verbalizer employs manual labels to represent the classes. However, manual selection does not guarantee the optimality of the selected words when conditioned on the chosen language model. Therefore, we propose Label-Aware Automatic Verbalizer (LAAV), effectively augmenting the manual labels to achieve better few-shot classification results. Specifically, we use the manual labels along with the conjunction""and""to induce the model to generate more effective words for the verbalizer. The experimental results on five datasets across five languages demonstrate that LAAV significantly outperforms existing verbalizers. Furthermore, our analysis reveals that LAAV suggests more relevant words compared to similar approaches, especially in mid-to-low resource languages.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Label-Aware Automatic Verbalizer (LAAV), effectively augmenting the manual labels to achieve better few-shot classification results and reveals that LAAV suggests more relevant words compared to similar approaches, especially in mid-to-low resource languages.'}",
-Few-shot Multimodal Multitask Multilingual Learning,Aman Chadha,"While few-shot learning as a transfer learning paradigm has gained significant traction for scenarios with limited data, it has primarily been explored in the context of building unimodal and unilingual models. Furthermore, a significant part of the existing literature in the domain of few-shot multitask learning perform in-context learning which requires manually generated prompts as the input, yielding varying outcomes depending on the level of manual prompt-engineering. In addition, in-context learning suffers from substantial computational, memory, and storage costs which eventually leads to high inference latency because it involves running all of the prompt's examples through the model every time a prediction is made. In contrast, methods based on the transfer learning via the fine-tuning paradigm avoid the aforementioned issues at a one-time cost of fine-tuning weights on a per-task basis. However, such methods lack exposure to few-shot multimodal multitask learning. In this paper, we propose few-shot learning for a multimodal multitask multilingual (FM3) setting by adapting pre-trained vision and language models using task-specific hypernetworks and contrastively fine-tuning them to enable few-shot learning. FM3's architecture combines the best of both worlds of in-context and fine-tuning based learning and consists of three major components: (i) multimodal contrastive fine-tuning to enable few-shot learning, (ii) hypernetwork task adaptation to perform multitask learning, and (iii) task-specific output heads to cater to a plethora of diverse tasks. FM3 learns the most prominent tasks in the vision and language domains along with their intersections, namely visual entailment (VE), visual question answering (VQA), and natural language understanding (NLU) tasks such as neural entity recognition (NER) and the GLUE benchmark including QNLI, MNLI, QQP, and SST-2.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes few-shot learning for a multimodal multitask multilingual (FM3) setting by adapting pre-trained vision and language models using task-specific hypernetworks and contrastively fine-tuning them to enable few- shot learning.'}",http://arxiv.org/pdf/2303.12489
-Zero-shot information extraction from radiological reports using ChatGPT,D. Hu,"Electronic health records contain an enormous amount of valuable information, but many are recorded in free text. Information extraction is the strategy to transform the sequence of characters into structured data, which can be employed for secondary analysis. However, the traditional information extraction components, such as named entity recognition and relation extraction, require annotated data to optimize the model parameters, which has become one of the major bottlenecks in building information extraction systems. With the large language models achieving good performances on various downstream NLP tasks without parameter tuning, it becomes possible to use large language models for zero-shot information extraction. In this study, we aim to explore whether the most popular large language model, ChatGPT, can extract useful information from the radiological reports. We first design the prompt template for the interested information in the CT reports. Then, we generate the prompts by combining the prompt template with the CT reports as the inputs of ChatGPT to obtain the responses. A post-processing module is developed to transform the responses into structured extraction results. We conducted the experiments with 847 CT reports collected from Peking University Cancer Hospital. The experimental results indicate that ChatGPT can achieve competitive performances for some extraction tasks compared with the baseline information extraction system, but some limitations need to be further improved.","{'model': 'tldr@v2.0.0', 'text': 'Whether the most popular large language model, ChatGPT, can extract useful information from the radiological reports is explored and the experimental results indicate thatChatGPT can achieve competitive performances for some extraction tasks compared with the baseline information extraction system, but some limitations need to be further improved.'}",https://arxiv.org/pdf/2309.01398
-Wide-field one-shot optical polarimeter: HOWPol,K. Kawabata,"For prompt optical polarimetry of gamma-ray burst (GRB) afterglow, we require wide-field imaging polarimeter which can produce both Stokes Q and U parameters from only a single exposure, as well as quickly-moving telescope and enclosure system. HOWPol is an optical imaging polarimeter which provides four linearly polarized images at position angles of 0°, 45°, 90° and 135°, i.e., Stokes I, Q, U, simultaneously. The key device is the wedged double Wollaston prism described by Oliva (1997)1 and Pernechele et al. (2003).2 The images are focused on two 2k×4k fully depleted CCDs. We report the design and development of the optical devices of HOWPol, which will be mounted to the 1.5-m Kanata telescope at Hiroshima University and stand by the GRB alert.",,
-One-Shot Illusory Correlations and Stereotype Formation,Jane L. Risen,"In four studies, the authors explored the emergence of one-shot illusory correlations—in which a single instance of unusual behavior by a member of a rare group is sufficient to create an association between group and behavior. In Studies 1, 2, and 3, unusual behaviors committed by members of rare groups were processed differently than other types of behaviors. They received more processing time, prompted more attributional thinking, and were more memorable. In Study 4, the authors obtained evidence from two implicit measures of association that one-shot illusory correlations are generalized to other members of a rare group. The authors contend that one-shot illusory correlations arise because unusual pairings of behaviors and groups uniquely prompt people to entertain group membership as an explanation of the unusual behavior.","{'model': 'tldr@v2.0.0', 'text': 'The authors contend that one-shot illusory correlations arise because unusual pairings of behaviors and groups uniquely prompt people to entertain group membership as an explanation of the unusual behavior.'}",
-One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning,Arnav Chavan,"We present Generalized LoRA (GLoRA), an advanced approach for universal parameter-efficient fine-tuning tasks. Enhancing Low-Rank Adaptation (LoRA), GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. Moreover, GLoRA facilitates efficient parameter adaptation by employing a scalable, modular, layer-wise structure search that learns individual adapter of each layer. Originating from a unified mathematical formulation, GLoRA exhibits strong transfer learning, few-shot learning and domain generalization abilities, as it adjusts to new tasks through additional dimensions on weights and activations. Comprehensive experiments demonstrate that GLoRA outperforms all previous methods in natural, specialized, and structured benchmarks, achieving superior accuracy with fewer parameters and computations on various datasets. Furthermore, our structural re-parameterization design ensures that GLoRA incurs no extra inference cost, rendering it a practical solution for resource-limited applications. Code is available at: https://github.com/Arnav0400/ViT-Slim/tree/master/GLoRA.","{'model': 'tldr@v2.0.0', 'text': 'Generalized LoRA, an advanced approach for universal parameter-efficient fine-tuning tasks, employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets.'}",http://arxiv.org/pdf/2306.07967
-Few-shot tweet detection in emerging disaster events,Anna Kruspe,"Social media sources can provide crucial information in crisis situations, but discovering relevant messages is not trivial. Methods have so far focused on universal detection models for all kinds of crises or for certain crisis types (e.g. floods). Event-specific models could implement a more focused search area, but collecting data and training new models for a crisis that is already in progress is costly and may take too much time for a prompt response. As a compromise, manually collecting a small amount of example messages is feasible. Few-shot models can generalize to unseen classes with such a small handful of examples, and do not need be trained anew for each event. We compare how few-shot approaches (matching networks and prototypical networks) perform for this task. Since this is essentially a one-class problem, we also demonstrate how a modified one-class version of prototypical models can be used for this application.","{'model': 'tldr@v2.0.0', 'text': 'This work compares how few-shot approaches (matching networks and prototypical networks) perform for this task and demonstrates how a modified one-class version of prototypical models can be used for this application.'}",
-Shot Scale Matters: The Effect of Close-Up Frequency on Theory of Mind Responses in Film Viewers,K. Bálint,"Theory of mind (ToM) processing is crucial in everyday social functioning. Fictional film narratives can elicit ToM, however, little is known about the cinematographic techniques underlying this effect. The present experiment focuses on the role of close-up shots of the character’s face in viewers’ ToM responding, as well as in their cognitive and affective processing more generally. The online experiment (N = 495) included thirteen versions of an animated film and employed a 6 (Close-up frequency, 1 vs. 2. vs. 3 vs. 4 vs. 5 vs. 10) × 2 (Facial expression, neutral vs. sad) plus a control condition with zero close-up factorial between-subject design. Participants were randomly assigned to one version of the film and subsequently asked (with and without a prompt for ToM) to recall the story. In these free responses, the study used a quantitative content analytic method (with independent blind-coders) to identify the level of spontaneous and prompted ToM, as well as cognitive and affective processing employed by viewers. Additionally, we tested the moderation effect of character facial expression (in the close-up) and participant gender. Confirming our main hypothesis, close-up frequency significantly influenced ToM-spontaneous, but not ToM-prompted. Results indicate that increasing the number of close-ups may elicit higher levels of ToM-spontaneous up to a certain point, beyond which it may decrease spontaneous ToM responses. Results suggest that the effect of close-up frequency is specific to ToM processing rather than some general effect on cognitive and affective processing of narratives.",,
-VIMA: General Robot Manipulation with Multimodal Prompts,Yunfan Jiang,"Prompt-based learning has emerged as a successful paradigm in natural language processing, where a single general-purpose language model can be instructed to perform any task specified by input prompts. Yet task specification in robotics comes in various forms, such as imitating one-shot demonstrations, following language instructions, and reaching visual goals. They are often considered different tasks and tackled by specialized models. We show that a wide spectrum of robot manipulation tasks can be expressed with multimodal prompts, interleaving textual and visual tokens. Accordingly, we develop a new simulation benchmark that consists of thousands of procedurally-generated tabletop tasks with multimodal prompts, 600K+ expert trajectories for imitation learning, and a four-level evaluation protocol for systematic generalization. We design a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively. VIMA features a recipe that achieves strong model scalability and data efficiency. It outperforms alternative designs in the hardest zero-shot generalization setting by up to $2.9\times$ task success rate given the same training data. With $10\times$ less training data, VIMA still performs $2.7\times$ better than the best competing variant. Code and video demos are available at https://vimalabs.github.io/","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a wide spectrum of robot manipulation tasks can be expressed with multimodal prompts, interleaving textual and visual tokens, and designed a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively.'}",http://arxiv.org/pdf/2210.03094
-Speak Foreign Languages with Your Own Voice: Cross-Lingual Neural Codec Language Modeling,Zi-Hua Zhang,"We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID. Audio samples are available at \url{https://aka.ms/vallex}.","{'model': 'tldr@v2.0.0', 'text': ""Experimental results show that VALL-E X can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment.""}",http://arxiv.org/pdf/2303.03926
-"Text and Patterns: For Effective Chain of Thought, It Takes Two to Tango",Aman Madaan,"The past decade has witnessed dramatic gains in natural language processing and an unprecedented scaling of large language models. These developments have been accelerated by the advent of few-shot techniques such as chain of thought (CoT) prompting. Specifically, CoT pushes the performance of large language models in a few-shot setup by augmenting the prompts with intermediate steps. Despite impressive results across various tasks, the reasons behind their success have not been explored. This work uses counterfactual prompting to develop a deeper understanding of CoT-based few-shot prompting mechanisms in large language models. We first systematically identify and define the key components of a prompt: symbols, patterns, and text. Then, we devise and conduct an exhaustive set of experiments across four different tasks, by querying the model with counterfactual prompts where only one of these components is altered. Our experiments across three models (PaLM, GPT-3, and CODEX) reveal several surprising findings and brings into question the conventional wisdom around few-shot prompting. First, the presence of factual patterns in a prompt is practically immaterial to the success of CoT. Second, our results conclude that the primary role of intermediate steps may not be to facilitate learning how to solve a task. The intermediate steps are rather a beacon for the model to realize what symbols to replicate in the output to form a factual answer. Further, text imbues patterns with commonsense knowledge and meaning. Our empirical and qualitative analysis reveals that a symbiotic relationship between text and patterns explains the success of few-shot prompting: text helps extract commonsense from the question to help patterns, and patterns enforce task understanding and direct text generation.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that a symbiotic relationship between text and patterns explains the success of few-shot prompting: text helps extract commonsense from the question to help patterns, and patterns enforce task understanding and direct text generation.'}",http://arxiv.org/pdf/2209.07686
-PointCLIP V2: Adapting CLIP for Powerful 3D Open-world Learning,Xiangyang Zhu,"Contrastive Language-Image Pre-training (CLIP) has shown promising open-world performance on 2D image tasks, while its transferred capacity on 3D point clouds, i.e., PointCLIP, is still far from satisfactory. In this work, we propose PointCLIP V2, a powerful 3D open-world learner, to fully unleash the potential of CLIP on 3D point cloud data. First, we introduce a realistic shape projection module to generate more realistic depth maps for CLIP’s visual encoder, which is quite efficient and narrows the domain gap between projected point clouds with natural images. Second, we leverage large-scale language models to automatically design a more descriptive 3D-semantic prompt for CLIP’s textual encoder, instead of the previous hand-crafted one. Without introducing any training in 3D domains, our approach significantly surpasses Point-CLIP by +42.90% , +40.44% , and +28.75% accuracy on three datasets for zero-shot 3D classification. Furthermore, PointCLIP V2 can be extended to few-shot classifi-cation, zero-shot part segmentation, and zero-shot 3D object detection in a simple manner, demonstrating our superior generalization ability for 3D open-world learning. Code will be available at https://github.com/","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PointCLIP V2, a powerful 3D open-world learner, to fully unleash the potential of CLIP on 3D point cloud data, and introduces a realistic shape projection module to generate more realistic depth maps for CLIP’s visual encoder.'}",https://arxiv.org/pdf/2211.11682
-Image Segmentation Using Text and Image Prompts,Timo Lüddecke,"Image segmentation is usually addressed by training a model for a fixed set of object classes. Incorporating additional classes or more complex queries later is expensive as it requires re-training the model on a dataset that encompasses these expressions. Here we propose a system that can generate image segmentations based on arbitrary prompts at test time. A prompt can be either a text or an image. This approach enables us to create a unified model (trained once) for three common segmentation tasks, which come with distinct challenges: referring expression segmentation, zero-shot segmentation and one-shot segmentation. We build upon the CLIP model as a backbone which we extend with a transformer-based decoder that enables dense prediction. After training on an extended version of the PhraseCut dataset, our system generates a binary segmentation map for an image based on a free-text prompt or on an additional image expressing the query. We analyze different variants of the latter image-based prompts in detail. This novel hybrid input allows for dynamic adaptation not only to the three segmentation tasks mentioned above, but to any binary segmentation task where a text or image query can be formulated. Finally, we find our system to adapt well to generalized queries involving affordances or properties. Code is available at https://eckerlab.org/code/CLIPSeg","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a system that can generate image segmentations based on arbitrary prompts at test time, and builds upon the CLIP model as a backbone which it extends with a transformer-based decoder that enables dense prediction.'}",https://arxiv.org/pdf/2112.10003
-SEGA: Instructing Diffusion using Semantic Dimensions,Manuel Brack,"Text-to-image diffusion models have recently received a lot of interest for their astonishing ability to produce high-fidelity images from text only. However, achieving one-shot generation that aligns with the user's intent is nearly impossible, yet small changes to the input prompt often result in very different images. This leaves the user with little semantic control. To put the user in control, we show how to interact with the diffusion process to flexibly steer it along semantic directions. This semantic guidance (SEGA) allows for subtle and extensive edits, changes in composition and style, as well as optimizing the overall artistic conception. We demonstrate SEGA's effectiveness on a variety of tasks and provide evidence for its versatility and flexibility.","{'model': 'tldr@v2.0.0', 'text': 'This semantic guidance (SEGA) allows for subtle and extensive edits, changes in composition and style, as well as optimizing the overall artistic conception in text-to-image diffusion models.'}",http://arxiv.org/pdf/2301.12247
-In-Context Instruction Learning,Seonghyeon Ye,"Instruction learning of Large Language Models (LLMs) has enabled zero-shot task generalization. However, instruction learning has been predominantly approached as a fine-tuning problem, including instruction tuning and reinforcement learning from human feedback, where LLMs are multi-task fine-tuned on various tasks with instructions. In this paper, we present a surprising finding that applying in-context learning to instruction learning, referred to as In-Context Instruction Learning (ICIL), significantly improves the zero-shot task generalization performance for both pretrained and instruction-fine-tuned models. One of the core advantages of ICIL is that it uses a single fixed prompt to evaluate all tasks, which is a concatenation of cross-task demonstrations. In particular, we demonstrate that the most powerful instruction-fine-tuned baseline (text-davinci-003) also benefits from ICIL by 9.3%, indicating that the effect of ICIL is complementary to instruction-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'A surprising finding is presented that applying in-context learning to instruction learning, referred to as In-Context Instruction Learning (ICIL), significantly improves the zero-shot task generalization performance for both pretrained and instruction-fine-tuned models.'}",http://arxiv.org/pdf/2302.14691
-Fast Video Shot Boundary Detection Technique based on Stochastic Model,M. Al-Jarrah,"A video is composed of set of shots, where shot is defined as a sequence of consecutive frames captured by one camera without interruption. In video shot transition could be a prompt hard cut or gradual fade, dissolve, and wipe. Shot boundary detection is an essential component of video processing. These boundaries are utilized on many aspect of video processing such as video indexing, and video in demand. In this paper, the authors proposed a new shot boundary detection algorithm. The proposed algorithm detects all type of shot boundaries in a high accuracy. The algorithm is developed based on a global stochastic model for video stream. The proposed stochastic model utilizes the joined characteristic function and consequently the joined momentum to model the video stream. The proposed algorithm is implemented and tested against different types of categorized videos. The proposed algorithm detects cuts fades, dissolves, and wipes transitions. Experimental results show that the algorithm has high performance. The computed precision and recall rates validated its performance.","{'model': 'tldr@v2.0.0', 'text': 'A new shot boundary detection algorithm is proposed based on a global stochastic model based on the joined characteristic function and consequently the joined momentum to model the video stream and has high performance.'}",
-"FreeSeg: Unified, Universal and Open-Vocabulary Image Segmentation",Jie Qin,"Recently, open-vocabulary learning has emerged to accomplish segmentation for arbitrary categories of text-based descriptions, which popularizes the segmentation system to more general-purpose application scenarios. However, existing methods devote to designing specialized architectures or parameters for specific segmentation tasks. These customized design paradigms lead to fragmentation between various segmentation tasks, thus hindering the uniformity of segmentation models. Hence in this paper, we propose FreeSeg, a generic framework to accomplish Unified, Universal and Open-Vocabulary Image Segmentation. FreeSeg optimizes an all-in-one network via one-shot training and employs the same architecture and parameters to handle diverse segmentation tasks seamlessly in the inference procedure. Additionally, adaptive prompt learning facilitates the unified model to capture task-aware and category-sensitive concepts, improving model robustness in multi-task and varied scenarios. Extensive experimental results demonstrate that FreeSeg establishes new state-of-the-art results in performance and generalization on three segmentation tasks, which outperforms the best task-specific architectures by a large margin: 5.5% mIoU on semantic segmentation, 17.6% mAP on instance segmentation, 20.1% PQ on panoptic segmentation for the unseen class on COCO. Project page: https://FreeSeg.github.io.","{'model': 'tldr@v2.0.0', 'text': 'FreeSeg is proposed, a generic framework to accomplish Unified, Universal and Open-Vocabulary Image Segmentation, which optimizes an all-in-one network via one-shot training and employs the same architecture and parameters to handle diverse segmentation tasks seamlessly in the inference procedure.'}",https://arxiv.org/pdf/2303.17225
-Learnable Ophthalmology SAM,Zhongxi Qiu,"Segmentation is vital for ophthalmology image analysis. But its various modal images hinder most of the existing segmentation algorithms applications, as they rely on training based on a large number of labels or hold weak generalization ability. Based on Segment Anything (SAM), we propose a simple but effective learnable prompt layer suitable for multiple target segmentation in ophthalmology multi-modal images, named Learnable Ophthalmology Segment Anything (SAM). The learnable prompt layer learns medical prior knowledge from each transformer layer. During training, we only train the prompt layer and task head based on a one-shot mechanism. We demonstrate the effectiveness of our thought based on four medical segmentation tasks based on nine publicly available datasets. Moreover, we only provide a new improvement thought for applying the existing fundamental CV models in the medical field. Our codes are available at \href{https://github.com/Qsingle/LearnablePromptSAM}{website}.","{'model': 'tldr@v2.0.0', 'text': 'A simple but effective learnable prompt layer suitable for multiple target segmentation in ophthalmology multi-modal images, named Learnable Ophthalmology Segment Anything (SAM), which learns medical prior knowledge from each transformer layer.'}",http://arxiv.org/pdf/2304.13425
-Prompting Multilingual Large Language Models to Generate Code-Mixed Texts: The Case of South East Asian Languages,Zheng-Xin Yong,"While code-mixing is a common linguistic practice in many parts of the world, collecting high-quality and low-cost code-mixed data remains a challenge for natural language processing (NLP) research. The recent proliferation of Large Language Models (LLMs) compels one to ask: how capable are these systems in generating code-mixed data? In this paper, we explore prompting multilingual LLMs in a zero-shot manner to generate code-mixed data for seven languages in South East Asia (SEA), namely Indonesian, Malay, Chinese, Tagalog, Vietnamese, Tamil, and Singlish. We find that publicly available multilingual instruction-tuned models such as BLOOMZ and Flan-T5-XXL are incapable of producing texts with phrases or clauses from different languages. ChatGPT exhibits inconsistent capabilities in generating code-mixed texts, wherein its performance varies depending on the prompt template and language pairing. For instance, ChatGPT generates fluent and natural Singlish texts (an English-based creole spoken in Singapore), but for English-Tamil language pair, the system mostly produces grammatically incorrect or semantically meaningless utterances. Furthermore, it may erroneously introduce languages not specified in the prompt. Based on our investigation, existing multilingual LLMs exhibit a wide range of proficiency in code-mixed data generation for SEA languages. As such, we advise against using LLMs in this context without extensive human checks.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores prompting multilingual LLMs in a zero-shot manner to generate code-mixed data for seven languages in South East Asia, and finds that publicly available multilingual instruction-tuned models are incapable of producing texts with phrases or clauses from different languages.'}",https://arxiv.org/pdf/2303.13592
-How to Efficiently Adapt Large Segmentation Model(SAM) to Medical Images,Xinrong Hu,"The emerging scale segmentation model, Segment Anything (SAM), exhibits impressive capabilities in zero-shot segmentation for natural images. However, when applied to medical images, SAM suffers from noticeable performance drop. To make SAM a real ``foundation model""for the computer vision community, it is critical to find an efficient way to customize SAM for medical image dataset. In this work, we propose to freeze SAM encoder and finetune a lightweight task-specific prediction head, as most of weights in SAM are contributed by the encoder. In addition, SAM is a promptable model, while prompt is not necessarily available in all application cases, and precise prompts for multiple class segmentation are also time-consuming. Therefore, we explore three types of prompt-free prediction heads in this work, include ViT, CNN, and linear layers. For ViT head, we remove the prompt tokens in the mask decoder of SAM, which is named AutoSAM. AutoSAM can also generate masks for different classes with one single inference after modification. To evaluate the label-efficiency of our finetuning method, we compare the results of these three prediction heads on a public medical image segmentation dataset with limited labeled data. Experiments demonstrate that finetuning SAM significantly improves its performance on medical image dataset, even with just one labeled volume. Moreover, AutoSAM and CNN prediction head also has better segmentation accuracy than training from scratch and self-supervised learning approaches when there is a shortage of annotations.","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that finetuning SAM significantly improves its performance on medical image dataset, even with just one labeled volume, and AutoSAM and CNN prediction head also has better segmentation accuracy than training from scratch and self-supervised learning approaches when there is a shortage of annotations.'}",http://arxiv.org/pdf/2306.13731
-The Stable Artist: Steering Semantics in Diffusion Latent Space,Manuel Brack,"Large, text-conditioned generative diffusion models have recently gained a lot of attention for their impressive performance in generating high-fidelity images from text alone. However, achieving high-quality results is almost unfeasible in a one-shot fashion. On the contrary, text-guided image generation involves the user making many slight changes to inputs in order to iteratively carve out the envisioned image. However, slight changes to the input prompt often lead to entirely different images being generated, and thus the control of the artist is limited in its granularity. To provide flexibility, we present the Stable Artist, an image editing approach enabling fine-grained control of the image generation process. The main component is semantic guidance (SEGA) which steers the diffusion process along variable numbers of semantic directions. This allows for subtle edits to images, changes in composition and style, as well as optimization of the overall artistic conception. Furthermore, SEGA enables probing of latent spaces to gain insights into the representation of concepts learned by the model, even complex ones such as 'carbon emission'. We demonstrate the Stable Artist on several tasks, showcasing high-quality image editing and composition.","{'model': 'tldr@v2.0.0', 'text': 'The Stable Artist is presented, an image editing approach enabling fine-grained control of the image generation process that allows for subtle edits to images, changes in composition and style, as well as optimization of the overall artistic conception.'}",http://arxiv.org/pdf/2212.06013
-Internet-augmented language models through few-shot prompting for open-domain question answering,Angeliki Lazaridou,"In this work, we aim to capitalize on the unique few-shot capabilities of large-scale language models (LSLMs) to overcome some of their challenges with respect to grounding to factual and up-to-date information. Motivated by semi-parametric language models (LMs), which ground their decisions in external retrieved evidence, we use few-shot prompting to learn to condition LMs on information returned from the web using Google Search, a broad and constantly updated knowledge source. Our approach does not involve fine-tuning or learning additional parameters, thus making it applicable to any LM, offering therefore a strong baseline. Indeed, we find that LMs conditioned on the web surpass performance of closed-book models of similar, or even larger, model sizes in open-domain question answering. Finally, we find that increasing the inference-time compute of models, achieved via using multiple retrieved evidences to generate multiple answers followed by a reranking stage that uses scores generated by the same LMs, leads to better performance and alleviates lower performance of smaller few-shot LMs. All in all, our findings suggest that it might be beneficial to slow down the race towards the biggest model and instead shift attention towards finding more effective ways to use models, including but not limited to, better prompting or increasing inference-time compute.","{'model': 'tldr@v2.0.0', 'text': 'Motivated by semi-parametric language models (LMs), few-shot prompting is used to learn to condition LMs on information returned from the web using Google Search, a broad and constantly updated knowledge source, making it applicable to any LM, offering therefore a strong baseline.'}",https://arxiv.org/pdf/2203.05115
-The Unreliability of Explanations in Few-shot Prompting for Textual Reasoning,Xi Ye,"Does prompting a large language model (LLM) like GPT-3 with explanations improve in-context learning? We study this question on two NLP tasks that involve reasoning over text, namely question answering and natural language inference. We test the performance of four LLMs on three textual reasoning datasets using prompts that include explanations in multiple different styles. For these tasks, we find that including explanations in the prompts for OPT, GPT-3 (davinci), and InstructGPT (text-davinci-001) only yields small to moderate accuracy improvements over standard few-show learning. However, text-davinci-002 is able to benefit more substantially. We further show that explanations generated by the LLMs may not entail the models' predictions nor be factually grounded in the input, even on simple tasks with extractive explanations. However, these flawed explanations can still be useful as a way to verify LLMs' predictions post-hoc. Through analysis in our three settings, we show that explanations judged by humans to be good--logically consistent with the input and the prediction--more likely cooccur with accurate predictions. Following these observations, we train calibrators using automatically extracted scores that assess the reliability of explanations, allowing us to improve performance post-hoc across all of our datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work studies two NLP tasks that involve reasoning over text, namely question answering and natural language inference, and shows that explanations judged by humans to be good--logically consistent with the input and the prediction--more likely cooccur with accurate predictions.'}",
-Language Model Crossover: Variation through Few-Shot Prompting,Elliot Meyerson,"This paper pursues the insight that language models naturally enable an intelligent variation operator similar in spirit to evolutionary crossover. In particular, language models of sufficient scale demonstrate in-context learning, i.e. they can learn from associations between a small number of input patterns to generate outputs incorporating such associations (also called few-shot prompting). This ability can be leveraged to form a simple but powerful variation operator, i.e. to prompt a language model with a few text-based genotypes (such as code, plain-text sentences, or equations), and to parse its corresponding output as those genotypes' offspring. The promise of such language model crossover (which is simple to implement and can leverage many different open-source language models) is that it enables a simple mechanism to evolve semantically-rich text representations (with few domain-specific tweaks), and naturally benefits from current progress in language models. Experiments in this paper highlight the versatility of language-model crossover, through evolving binary bit-strings, sentences, equations, text-to-image prompts, and Python code. The conclusion is that language model crossover is a promising method for evolving genomes representable as text.","{'model': 'tldr@v2.0.0', 'text': 'The conclusion is that language model crossover is a promising method for evolving genomes representable as text through evolving binary bit-strings, sentences, equations, text-to-image prompts, and Python code.'}",https://arxiv.org/pdf/2302.12170
-Leveraging Training Data in Few-Shot Prompting for Numerical Reasoning,Zhanming Jie,"Chain-of-thought (CoT) prompting with large language models has proven effective in numerous natural language processing tasks, but designing prompts that generalize well to diverse problem types can be challenging, especially in the context of math word problem (MWP) solving. Additionally, it is common to have a large amount of training data that have a better diversity coverage but CoT annotations are not available, which limits the use of supervised learning techniques. To address these issues, we investigate two approaches to leverage the training data in a few-shot prompting scenario: dynamic program prompting and program distillation. Our approach is largely inspired by Gao et al., (2022), where they proposed to replace the CoT with the programs as the intermediate reasoning step. Such a prompting strategy allows us to accurately verify the answer correctness through program execution in MWP solving. Our dynamic program prompting involves annotating the training data by sampling correct programs from a large language model, while program distillation involves adapting a smaller model to the program-annotated training data. Our experiments on three standard MWP datasets demonstrate the effectiveness of these approaches, yielding significant improvements over previous baselines for prompting and fine-tuning. Our results suggest that leveraging a large amount of training data can improve the generalization ability of prompts and boost the performance of fine-tuned small models in MWP solving.","{'model': 'tldr@v2.0.0', 'text': 'The results suggest that leveraging a large amount of training data can improve the generalization ability of prompts and boost the performance of fine-tuned small models in MWP solving.'}",http://arxiv.org/pdf/2305.18170
-Few-shot Prompting Towards Controllable Response Generation,Hsuan Su,"Much literature has shown that prompt-based learning is an efficient method to make use of the large pre-trained language model. Recent works also exhibit the possibility of steering a chatbot’s output by plugging in an ap-propriate prompt. Gradient-based methods are often used to perturb the prompts. However, some language models are not even available to the public. In this work, we first explored the combination of prompting and reinforcement learning (RL) to steer models’ generation without accessing any of the models’ parameters. Second, to reduce the training effort and enhance the generalizability to the unseen task, we apply multi-task learning to make the model learn to generalize to new tasks better. The experiment results show that our proposed method can successfully control several state-of-the-art (SOTA) dialogue models without accessing their parameters. Furthermore, the model demonstrates the strong ability to quickly adapt to an unseen task in fewer steps than the baseline model.","{'model': 'tldr@v2.0.0', 'text': 'The experiment results show that the proposed method can successfully control several state-of-the-art (SOTA) dialogue models without accessing their parameters and demonstrates the strong ability to quickly adapt to an unseen task in fewer steps than the baseline model.'}",http://arxiv.org/pdf/2206.03931
-Zero- and Few-Shot Prompting with LLMs: A Comparative Study with Fine-tuned Models for Bangla Sentiment Analysis,Md. Arid Hasan,"The rapid expansion of the digital world has propelled sentiment analysis into a critical tool across diverse sectors such as marketing, politics, customer service, and healthcare. While there have been significant advancements in sentiment analysis for widely spoken languages, low-resource languages, such as Bangla, remain largely under-researched due to resource constraints. Furthermore, the recent unprecedented performance of Large Language Models (LLMs) in various applications highlights the need to evaluate them in the context of low-resource languages. In this study, we present a sizeable manually annotated dataset encompassing 33,605 Bangla news tweets and Facebook comments. We also investigate zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine-tuned models. Our findings suggest that monolingual transformer-based models consistently outperform other models, even in zero and few-shot scenarios. To foster continued exploration, we intend to make this dataset and our research tools publicly available to the broader research community. In the spirit of further research, we plan to make this dataset and our experimental resources publicly accessible to the wider research community.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a sizeable manually annotated dataset encompassing 33,605 Bangla news tweets and Facebook comments and investigates zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine- tuned models.'}",https://arxiv.org/pdf/2308.10783
-Fairness-guided Few-shot Prompting for Large Language Models,Huan Ma,"Large language models have demonstrated surprising ability to perform in-context learning, i.e., these models can be directly applied to solve numerous downstream tasks by conditioning on a prompt constructed by a few input-output examples. However, prior research has shown that in-context learning can suffer from high instability due to variations in training examples, example order, and prompt formats. Therefore, the construction of an appropriate prompt is essential for improving the performance of in-context learning. In this paper, we revisit this problem from the view of predictive bias. Specifically, we introduce a metric to evaluate the predictive bias of a fixed prompt against labels or a given attributes. Then we empirically show that prompts with higher bias always lead to unsatisfactory predictive quality. Based on this observation, we propose a novel search strategy based on the greedy search to identify the near-optimal prompt for improving the performance of in-context learning. We perform comprehensive experiments with state-of-the-art mainstream models such as GPT-3 on various downstream tasks. Our results indicate that our method can enhance the model's in-context learning performance in an effective and interpretable manner.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a metric to evaluate the predictive bias of a fixed prompt against labels or a given attributes and proposes a novel search strategy based on the greedy search to identify the near-optimal prompt for improving the performance of in-context learning.'}",http://arxiv.org/pdf/2303.13217
-Multilingual Social Media Text Generation and Evaluation with Few-Shot Prompting,Mack Blackburn,"This work adapts large language models to generate multilingual social media text that meets several objectives simultaneously: topic relevance, author style consistency, and reply validity. Leveraging existing online information behavior simulators, which currently only forecast activities but not content, our approach comprised of generalizable prompt formation and efficient evaluation to produce a believable, personalized, and responsive synthetic social network. According to some preliminary experiments, our multi-objective prompt formation and automatic evaluation/selection methods are able to yield a significant number of high-quality synthetic texts according to both standardized and trained metrics.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2022.gem-1.39.pdf
-SPARSEFIT: Few-shot Prompting with Sparse Fine-tuning for Jointly Generating Predictions and Natural Language Explanations,Jesus Solano,"Explaining the decisions of neural models is crucial for ensuring their trustworthiness at deployment time. Using Natural Language Explanations (NLEs) to justify a model's predictions has recently gained increasing interest. However, this approach usually demands large datasets of human-written NLEs for the ground-truth answers, which are expensive and potentially infeasible for some applications. For models to generate high-quality NLEs when only a few NLEs are available, the fine-tuning of Pre-trained Language Models (PLMs) in conjunction with prompt-based learning recently emerged. However, PLMs typically have billions of parameters, making fine-tuning expensive. We propose SparseFit, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs. We experiment with SparseFit on the T5 model and four datasets and compare it against state-of-the-art parameter-efficient fine-tuning techniques. We perform automatic and human evaluations to assess the quality of the model-generated NLEs, finding that fine-tuning only 6.8% of the model parameters leads to competitive results for both the task performance and the quality of the NLEs.","{'model': 'tldr@v2.0.0', 'text': 'SarseFit is proposed, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs and is compared against state-of-the-art parameter-efficient fine- Tuning techniques.'}",http://arxiv.org/pdf/2305.13235
-Style-Aware Radiology Report Generation with RadGraph and Few-Shot Prompting,Benjamin Yan,"Automatically generated reports from medical images promise to improve the workflow of radiologists. Existing methods consider an image-to-report modeling task by directly generating a fully-fledged report from an image. However, this conflates the content of the report (e.g., findings and their attributes) with its style (e.g., format and choice of words), which can lead to clinically inaccurate reports. To address this, we propose a two-step approach for radiology report generation. First, we extract the content from an image; then, we verbalize the extracted content into a report that matches the style of a specific radiologist. For this, we leverage RadGraph -- a graph representation of reports -- together with large language models (LLMs). In our quantitative evaluations, we find that our approach leads to beneficial performance. Our human evaluation with clinical raters highlights that the AI-generated reports are indistinguishably tailored to the style of individual radiologist despite leveraging only a few examples as context.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-step approach for radiology report generation that leverages RadGraph -- a graph representation of reports -- together with large language models (LLMs) and finds that this approach leads to beneficial performance.'}",
-Understanding How Model Size Affects Few-shot Instruction Prompting,Ayrton San Joaquin,"Large Language Models are affected by the phenomena of memorizing and forgetting their training data. But how do these vary by model size? We work towards this question by investigating how the model size affects the model's ability to discriminate a word's meaning in a given context. We introduce a dataset called DeltaWords, which evaluates a model's ability to follow instructions to select a sentence which replaces the target word with its antonym. We show a weak inverse scaling trend, where task accuracy degrades as model size increase, under extremely few-shot prompting regimes. We show that increasing the number of examples tend to disproportionately benefit larger models than smaller models.","{'model': 'tldr@v2.0.0', 'text': 'A weak inverse scaling trend is shown, where task accuracy degrades as model size increase, under extremely few-shot prompting regimes, and it is shown that increasing the number of examples tend to disproportionately benefit larger models than smaller models.'}",https://arxiv.org/pdf/2212.01907
-Prompting ELECTRA: Few-Shot Learning with Discriminative Pre-Trained Models,Mengzhou Xia,"Pre-trained masked language models successfully perform few-shot learning by formulating downstream tasks as text infilling. How- ever, as a strong alternative in full-shot settings, discriminative pre-trained models like ELECTRA do not fit into the paradigm. In this work, we adapt prompt-based few-shot learning to ELECTRA and show that it outperforms masked language models in a wide range of tasks. ELECTRA is pre-trained to distinguish if a token is generated or original. We naturally extend that to prompt-based few-shot learning by training to score the originality of the target options without introducing new parameters. Our method can be easily adapted to tasks involving multi-token predictions without extra computation overhead. Analysis shows that ELECTRA learns distributions that align better with downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work adapt prompt-based few-shot learning to ELECTRA and shows that it outperforms masked language models in a wide range of tasks and learns distributions that align better with downstream tasks.'}",https://arxiv.org/pdf/2205.15223
-Unsupervisedly Prompting AlphaFold2 for Few-Shot Learning of Accurate Folding Landscape and Protein Structure Prediction,Jun Zhang,"Data-driven predictive methods which can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining accurate folding landscape using co-evolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit co-evolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologs. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences, but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method which could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.","{'model': 'tldr@v2.0.0', 'text': 'EvoGen, a meta generative model, is presented to remedy the underperformance of AlphaFold2 for poor MSA targets, and yields a probabilistic structure generation method which could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.'}",
-Knowledge Prompting for Few-shot Action Recognition,Yuheng Shi,"Few-shot action recognition in videos is challenging for its lack of supervision and difficulty in generalizing to unseen actions. To address this task, we propose a simple yet effec- tive method, called knowledge prompting, which leverages commonsense knowledge of actions from external resources to prompt a powerful pre-trained vision-language model for few-shot classification. We first collect large-scale language descriptions of actions, defined as text proposals, to build an action knowledge base. The collection of text proposals is done by filling in handcraft sentence templates with external action-related corpus or by extracting action-related phrases from captions of Web instruction videos. Then we feed these text proposals into the pre-trained vision-language model along with video frames to generate matching scores of the proposals to each frame, and the scores can be treated as action semantics with strong generalization. Finally, we design a lightweight temporal modeling network to capture the temporal evolution of action semantics for classification. Extensive experiments on six benchmark datasets demonstrate that our method generally achieves the state-of-the-art performance while reducing the training overhead to 1‰ of existing methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple yet effective method, called knowledge prompting, which leverages commonsense knowledge of actions from external resources to prompt a powerful pre-trained vision-language model for few-shot classification.'}",https://arxiv.org/pdf/2211.12030
-Few-shot Reranking for Multi-hop QA via Language Model Prompting,Muhammad Khalifa,"We study few-shot reranking for multi-hop QA (MQA) with open-domain questions. To alleviate the need for a large number of labeled question-document pairs for retriever training, we propose PromptRank, which relies on language model prompting for multi-hop path reranking. PromptRank first constructs an instruction-based prompt that includes a candidate document path and then computes the relevance score between a given question and the path based on the conditional likelihood of the question given the path prompt according to a language model. PromptRank yields strong retrieval performance on HotpotQA with only 128 training examples compared to state-of-the-art methods trained on thousands of examples — 73.6 recall@10 by PromptRank vs. 77.8 by PathRetriever and 77.5 by multi-hop dense retrieval.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2023.acl-long.885.pdf
-PromptNER: A Prompting Method for Few-shot Named Entity Recognition via k Nearest Neighbor Search,Mozhi Zhang,"Few-shot Named Entity Recognition (NER) is a task aiming to identify named entities via limited annotated samples. Recently, prototypical networks have shown promising performance in few-shot NER. Most of prototypical networks will utilize the entities from the support set to construct label prototypes and use the query set to compute span-level similarities and optimize these label prototype representations. However, these methods are usually unsuitable for fine-tuning in the target domain, where only the support set is available. In this paper, we propose PromptNER: a novel prompting method for few-shot NER via k nearest neighbor search. We use prompts that contains entity category information to construct label prototypes, which enables our model to fine-tune with only the support set. Our approach achieves excellent transfer learning ability, and extensive experiments on the Few-NERD and CrossNER datasets demonstrate that our model achieves superior performance over state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.12217
-Prompting Large Language Models with Chain-of-Thought for Few-Shot Knowledge Base Question Generation,Yuanyuan Liang,"The task of Question Generation over Knowledge Bases (KBQG) aims to convert a logical form into a natural language question. For the sake of expensive cost of large-scale question annotation, the methods of KBQG under low-resource scenarios urgently need to be developed. However, current methods heavily rely on annotated data for fine-tuning, which is not well-suited for few-shot question generation. The emergence of Large Language Models (LLMs) has shown their impressive generalization ability in few-shot tasks. Inspired by Chain-of-Thought (CoT) prompting, which is an in-context learning strategy for reasoning, we formulate KBQG task as a reasoning problem, where the generation of a complete question is splitted into a series of sub-question generation. Our proposed prompting method KQG-CoT first retrieves supportive logical forms from the unlabeled data pool taking account of the characteristics of the logical form. Then, we write a prompt to explicit the reasoning chain of generating complicated questions based on the selected demonstrations. To further ensure prompt quality, we extend KQG-CoT into KQG-CoT+ via sorting the logical forms by their complexity. We conduct extensive experiments over three public KBQG datasets. The results demonstrate that our prompting method consistently outperforms other prompting baselines on the evaluated datasets. Remarkably, our KQG-CoT+ method could surpass existing few-shot SoTA results of the PathQuestions dataset by 18.25, 10.72, and 10.18 absolute points on BLEU-4, METEOR, and ROUGE-L, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by Chain-of-Thought prompting, the proposed prompting method KQG-CoT first retrieves supportive logical forms from the unlabeled data pool taking account of the characteristics of the logical form, and could surpass existing few-shot SoTA results of the PathQuestions dataset.'}",https://arxiv.org/pdf/2310.08395
-Investigating Prompting Techniques for Zero- and Few-Shot Visual Question Answering,Rabiul Awal,"Visual question answering (VQA) is a challenging task that requires the ability to comprehend and reason with visual information. While recent vision-language models have made strides, they continue to struggle with zero-shot VQA, particularly in handling complex compositional questions and adapting to new domains i.e. knowledge-based reasoning. This paper explores the use of various prompting strategies, focusing on the BLIP2 model, to enhance zero-shot VQA performance. We conduct a comprehensive investigation across several VQA datasets, examining the effectiveness of different question templates, the role of few-shot exemplars, the impact of chain-of-thought (CoT) reasoning, and the benefits of incorporating image captions as additional visual cues. Despite the varied outcomes, our findings demonstrate that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples. However, we also identify a limitation in the use of chain-of-thought rationalization, which negatively affects VQA accuracy. Our study thus provides critical insights into the potential of prompting for improving zero-shot VQA performance.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples, however, a limitation in the use of chain-of-thought rationalization is identified, which negatively affects V QA accuracy.'}",http://arxiv.org/pdf/2306.09996
-ProD: Prompting-to-disentangle Domain Knowledge for Cross-domain Few-shot Image Classification,Tianyi Ma,"This paper considers few-shot image classification under the cross-domain scenario, where the train-to-test domain gap compromises classification accuracy. To mitigate the domain gap, we propose a prompting-to-disentangle (ProD) method through a novel exploration with the prompting mechanism. ProD adopts the popular multi-domain training scheme and extracts the backbone feature with a standard Convolutional Neural Network. Based on these two common practices, the key point of ProD is using the prompting mechanism in the transformer to disentangle the domain-general (DG) and domain-specific (DS) knowledge from the backbone feature. Specifically, ProD concatenates a DG and a DS prompt to the backbone feature and feeds them into a lightweight transformer. The DG prompt is learnable and shared by all the training domains, while the DS prompt is generated from the domain-of-interest on the fly. As a result, the transformer outputs DG and DS features in parallel with the two prompts, yielding the disentangling effect. We show that: 1) Simply sharing a single DG prompt for all the training domains already improves generalization towards the novel test domain. 2) The cross-domain generalization can be further reinforced by making the DG prompt neutral towards the training domains. 3) When inference, the DS prompt is generated from the support samples and can capture test domain knowledge through the prompting mechanism. Combining all three benefits, ProD significantly improves cross-domain few-shot classification. For instance, on CUB, ProD improves the 5-way 5-shot ac-curacy from 73.56% (baseline) to 79.19%, setting a new state of the art.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that simply sharing a single DG prompt for all the training domains already improves generalization towards the novel test domain, and the cross-domain generalization can be further reinforced by making the DG prompt neutral towards theTraining domains.'}",
-Template-Free Prompting for Few-Shot Named Entity Recognition via Semantic-Enhanced Contrastive Learning.,Kai He,"Prompt tuning has achieved great success in various sentence-level classification tasks by using elaborated label word mappings and prompt templates. However, for solving token-level classification tasks, e.g., named entity recognition (NER), previous research, which utilizes N-gram traversal for prompting all spans with all possible entity types, is time-consuming. To this end, we propose a novel prompt-based contrastive learning method for few-shot NER without template construction and label word mappings. First, we leverage external knowledge to initialize semantic anchors for each entity type. These anchors are simply appended with input sentence embeddings as template-free prompts (TFPs). Then, the prompts and sentence embeddings are in-context optimized with our proposed semantic-enhanced contrastive loss. Our proposed loss function enables contrastive learning in few-shot scenarios without requiring a significant number of negative samples. Moreover, it effectively addresses the issue of conventional contrastive learning, where negative instances with similar semantics are erroneously pushed apart in natural language processing (NLP)-related tasks. We examine our method in label extension (LE), domain-adaption (DA), and low-resource generalization evaluation tasks with six public datasets and different settings, achieving state-of-the-art (SOTA) results in most cases.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel prompt-based contrastive learning method for few-shot NER without template construction and label word mappings, and effectively addresses the issue of conventional Contrastive learning, where negative instances with similar semantics are erroneously pushed apart in natural language processing (NLP)-related tasks.'}",
-LMCap: Few-shot Multilingual Image Captioning by Retrieval Augmented Language Model Prompting,R. Ramos,"Multilingual image captioning has recently been tackled by training with large-scale machine translated data, which is an expensive, noisy, and time-consuming process. Without requiring any multilingual caption data, we propose LMCap, an image-blind few-shot multilingual captioning model that works by prompting a language model with retrieved captions. Specifically, instead of following the standard encoder-decoder paradigm, given an image, LMCap first retrieves the captions of similar images using a multilingual CLIP encoder. These captions are then combined into a prompt for an XGLM decoder, in order to generate captions in the desired language. In other words, the generation model does not directly process the image, instead processing retrieved captions. Experiments on the XM3600 dataset of geographically diverse images show that our model is competitive with fully-supervised multilingual captioning models, without requiring any supervised training on any captioning data.","{'model': 'tldr@v2.0.0', 'text': 'LMCap is proposed, an image-blind few-shot multilingual captioning model that works by prompting a language model with retrieved captions, which is competitive with fully-supervised mult bilingual captioning models, without requiring any supervised training on any captioning data.'}",http://arxiv.org/pdf/2305.19821
-HiPrompt: Few-Shot Biomedical Knowledge Fusion via Hierarchy-Oriented Prompting,Jiaying Lu,"Medical decision-making processes can be enhanced by comprehensive biomedical knowledge bases, which require fusing knowledge graphs constructed from different sources via a uniform index system. The index system often organizes biomedical terms in a hierarchy to provide the aligned entities with fine-grained granularity. To address the challenge of scarce supervision in the biomedical knowledge fusion (BKF) task, researchers have proposed various unsupervised methods. However, these methods heavily rely on ad-hoc lexical and structural matching algorithms, which fail to capture the rich semantics conveyed by biomedical entities and terms. Recently, neural embedding models have proved effective in semantic-rich tasks, but they rely on sufficient labeled data to be adequately trained. To bridge the gap between the scarce-labeled BKF and neural embedding models, we propose HiPrompt, a supervision-efficient knowledge fusion framework that elicits the few-shot reasoning ability of large language models through hierarchy-oriented prompts. Empirical results on the collected KG-Hi-BKF benchmark datasets demonstrate the effectiveness of HiPrompt.","{'model': 'tldr@v2.0.0', 'text': 'HiPrompt is proposed, a supervision-efficient knowledge fusion framework that elicits the few-shot reasoning ability of large language models through hierarchy-oriented prompts and demonstrates the effectiveness of HiPrompt on the collected KG-Hi-BKF benchmark datasets.'}",https://arxiv.org/pdf/2304.05973
-Few-shot Learning with Prompting Methods,Morteza Bahrami,"Today, in natural language processing, labeled data is important, however, getting adequate amount of data is a challenging step. There are many tasks for which it is difficult to obtain the required training data. For example, in machine translation, we need to prepare a lot of data in the target language, so that the work performance is acceptable. We may not be able to collect useful data in the target language. Hence, we need to use few-shot learning. Recently, a method called prompting has evolved, in which text inputs are converted into text with a new structure using a certain format, which has a blank space. Given the prompted text, a pre-trained language model replaces the space with the best word. Prompting can help us in the field of few-shot learning; even in cases where there is no data, i.e. zero-shot learning. Recent works use large language models such as GPT-2 and GPT-3, with the prompting method, performed tasks such as machine translation. These efforts do not use any labeled training data. But these types of models with a massive number of parameters require powerful hardware. Pattern-Exploiting Training (PET) and iterative Pattern-Exploiting Training (iPET) were introduced, which perform few-shot learning using prompting and smaller pre-trained language models such as Bert and Roberta. For example, for the Yahoo text classification dataset, using iPET and Roberta and ten labeled datasets, 70% accuracy has been reached. This paper reviews research works in few-shot learning with a new paradigm in natural language processing, which we dub prompt-based learning or in short, prompting.","{'model': 'tldr@v2.0.0', 'text': 'Research works in few-shot learning with a new paradigm in natural language processing, which is dubbed prompt-based learning or in short, prompting are reviewed.'}",
-Adversarial Knowledge Stimulated Contrastive Prompting for Few-shot Language Learners,Kai Zheng,"Prompt-based fine-tuning has boosted the performance of Pre-trained Language Models (PLMs) on few-shot Natural Language Understanding (NLU) tasks by employing task-specific prompts. Yet, PLMs are unfamiliar with prompt-style expressions during pre-training, which limits the few-shot learning performance on downstream tasks. It would be desirable if the models can stimulate prompting knowledge while adaptation to specific NLU tasks. We present the Adversarial Knowledge Stimulated Contrastive Prompting (AKSCP) framework, leading to better few-shot NLU tasks for language models by implicitly stimulate knowledge from pretrained language model. In AKSCP, a novel paradigm Cloze-driven prompt is proposed for joint prompt tuning across word cloze task and prompt-based learning, forcing PLMs to stimulate prompting knowledge. We further design an Adversarial Contrastive learning method to improve the generalization ability of PLM for different downstream tasks. Experiments over a variety of NLU tasks show that AKSCP consistently outperforms state-of-the-arts for prompt-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'In AKSCP, a novel paradigm Cloze-driven prompt is proposed for joint prompt tuning across word cloze task and prompt-based learning, forcing PLMs to stimulate prompting knowledge, and an Adversarial Contrastive learning method is designed to improve the generalization ability of PLM for different downstream tasks.'}",https://aclanthology.org/2023.findings-acl.852.pdf
-Multi-step Prompting for Few-shot Emotion-Grounded Conversations,Mauzama Firdaus,"Conversational systems have shown immense growth in their ability to communicate like humans. With the emergence of large pre-trained language models (PLMs) the ability to provide informative responses have improved significantly. Despite the success of PLMs, the ability to identify and generate engaging and empathetic responses is largely dependent on labelled-data. In this work, we design a prompting approach that identifies the emotion of a given utterance and uses the emotion information for generating the appropriate responses for conversational systems. We propose a two-step prompting method that first recognises the emotion in the dialogue utterance and in the second-step uses the predicted emotion to prompt the PLM to generate the corresponding em- pathetic response in a few-shot setting. Experimental results on three publicly available datasets show that our proposed approach outperforms the state-of-the-art approaches for both automatic and manual evaluation.","{'model': 'tldr@v2.0.0', 'text': 'A two-step prompting method that first recognises the emotion in the dialogue utterance and in the second-step uses the predicted emotion to prompt the PLM to generate the corresponding em- pathetic response in a few-shot setting is proposed.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615265
-Leveraging Few-Shot Data Augmentation and Waterfall Prompting for Response Generation,Lea Krause,"This paper discusses our approaches for task-oriented conversational modelling using subjective knowledge, with a particular emphasis on response generation. Our methodology was shaped by an extensive data analysis that evaluated key factors such as response length, sentiment, and dialogue acts present in the provided dataset. We used few-shot learning to augment the data with newly generated subjective knowledge items and present three approaches for DSTC11: (1) task-specific model exploration, (2) incorporation of the most frequent question into all generated responses, and (3) a waterfall prompting technique using a combination of both GPT-3 and ChatGPT.","{'model': 'tldr@v2.0.0', 'text': 'This paper discusses the approaches for task-oriented conversational modelling using subjective knowledge, with a particular emphasis on response generation, and presents three approaches for DSTC11: task-specific model exploration, incorporation of the most frequent question into all generated responses, and a waterfall prompting technique using a combination of both GPT-3 and ChatGPT.'}",https://arxiv.org/pdf/2308.01080
-Unsupervisedly Prompting AlphaFold2 for Accurate Few-Shot Protein Structure Prediction.,Jun Zhang,"Data-driven predictive methods that can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining an accurate folding landscape using coevolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit coevolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologues. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in the low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.","{'model': 'tldr@v2.0.0', 'text': 'EvoGen, a meta generative model, is presented to remedy the underperformance of AlphaFold2 for poor MSA targets and yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.'}",
-The Unreliability of Explanations in Few-Shot In-Context Learning,Xi Ye,"How can prompting a large language model like GPT-3 with explanations improve in-context learning? We focus specifically on two NLP tasks that involve reasoning over text, namely question answering and natural language inference. Including explanations in the prompt and having the model generate them does not consistently improve performance in the settings we study, contrary to recent results on symbolic reasoning tasks (Nye et al., 2021; Wei et al., 2022). Despite careful prompting, explanations generated by GPT-3 may not even be factually grounded in the input, even on simple tasks with straightforward extractive explanations. However, these flawed explanations can still be useful as a way to verify GPT-3’s predictions post-hoc. Through analysis in three settings, we show that explanations judged as good by humans—those that are logically consistent with the input and the prediction—usually indicate more accurate predictions. Following these observations, we present a framework for calibrating model predictions based on the reliability of the explanations. Our framework trains calibrators using automatically extracted scores that approximately assess the reliability of explanations, which helps improve performance across three different datasets","{'model': 'tldr@v2.0.0', 'text': 'A framework for calibrating model predictions based on the reliability of explanations is presented, showing that explanations judged as good by humans—those that are logically consistent with the input and the prediction—usually indicate more accurate predictions.'}",http://arxiv.org/pdf/2205.03401
-Self-Convinced Prompting: Few-Shot Question Answering with Repeated Introspection,Haodi Zhang,"While large language models (LLMs) such as ChatGPT and PaLM have demonstrated remarkable performance in various language understanding and generation tasks, their capabilities in complex reasoning and intricate knowledge utilization still fall short of human-level proficiency. Recent studies have established the effectiveness of prompts in steering LLMs towards generating desired outputs. Building on these insights, we introduce a novel framework that harnesses the potential of large-scale pre-trained language models, to iteratively enhance performance of the LLMs. Our framework incorporates three components: \textit{Normal CoT}, a \textit{Convincer}, and an \textit{Answerer}. It processes the output of a typical few-shot chain-of-thought prompt, assesses the correctness of the response, scrutinizes the answer, refines the reasoning, and ultimately produces a new solution. Experimental results on the 7 datasets of miscellaneous problems validate the efficacy of the Self-Convince framework, achieving substantial improvements compared to the baselines. This study contributes to the burgeoning body of research focused on integrating pre-trained language models with tailored prompts and iterative refinement processes to augment their performance in complex tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework that harnesses the potential of large-scale pre-trained language models, to iteratively enhance performance of the LLMs, and validate the efficacy of the Self-Convince framework on the 7 datasets of miscellaneous problems.'}",https://arxiv.org/pdf/2310.05035
-"Focusing, Bridging and Prompting for Few-shot Nested Named Entity Recognition",Yuanyuan Xu,,,https://aclanthology.org/2023.findings-acl.164.pdf
-Instruction Tuning for Few-Shot Aspect-Based Sentiment Analysis,Siddharth Varia,"Aspect-based Sentiment Analysis (ABSA) is a fine-grained sentiment analysis task which involves four elements from user-generated texts:aspect term, aspect category, opinion term, and sentiment polarity. Most computational approaches focus on some of the ABSA sub-taskssuch as tuple (aspect term, sentiment polarity) or triplet (aspect term, opinion term, sentiment polarity) extraction using either pipeline or joint modeling approaches. Recently, generative approaches have been proposed to extract all four elements as (one or more) quadrupletsfrom text as a single task. In this work, we take a step further and propose a unified framework for solving ABSA, and the associated sub-tasksto improve the performance in few-shot scenarios. To this end, we fine-tune a T5 model with instructional prompts in a multi-task learning fashion covering all the sub-tasks, as well as the entire quadruple prediction task. In experiments with multiple benchmark datasets, we show that the proposed multi-task prompting approach brings performance boost (by absolute 8.29 F1) in the few-shot learning setting.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a unified framework for solving ABSA, and fine-tune a T5 model with instructional prompts in a multi-task learning fashion covering all the sub-tasks, as well as the entire quadruple prediction task to improve the performance in few-shot scenarios.'}",http://arxiv.org/pdf/2210.06629
-Retrieving Visual Facts For Few-Shot Visual Question Answering,,"We introduce the Retrieving Visual Facts 001 (RVF) framework for few-shot visual question 002 answering (VQA). The RVF framework rep- 003 resents an image as a set of natural language 004 facts; for example, in practice these could be 005 tags from an object detector. Critically, the 006 question is used to retrieve relevant facts: an 007 image may contain numerous details, and one 008 should attend to the few which may be use- 009 ful for the question. Finally, one predicts the 010 answer from the retrieved facts and the ques- 011 tion, e.g., by prompting a language model as 012 we do here. Compared to P ICA (Yang et al., 013 2021), the previous state-of-the-art in few-shot 014 VQA, a proof-of-concept RVF implementa- 015 tion improves absolute performance by 2.6% 016 and 1.5% respectively on the VQAv2 (Goyal 017 et al., 2017) and OK-VQA (Marino et al., 2019) 018 datasets. We also analyze our implementation’s 019 strengths and weaknesses on various question 020 types, highlighting directions for further study. 021","{'model': 'tldr@v2.0.0', 'text': 'A proof-of-concept RVF implementation improves absolute performance by 2.6% and 1.5% respectively on the VQAv2 and OK-VQA datasets, and analyzes the implementation’s 019 strengths and weaknesses on various question 020 types, highlighting directions for further study.'}",
-Continued Pretraining for Better Zero- and Few-Shot Promptability,Zhaofeng Wu,"Recently introduced language model prompting methods can achieve high accuracy in zero- and few-shot settings while requiring few to no learned task-specific parameters. Nevertheless, these methods still often trail behind full model finetuning. In this work, we investigate if a dedicated continued pretraining stage could improve “promptability”, i.e., zero-shot performance with natural language prompts or few-shot performance with prompt tuning. We reveal settings where existing continued pretraining methods lack promptability. We also identify current methodological gaps, which we fill with thorough large-scale experiments. We demonstrate that a simple recipe, continued pretraining that incorporates a trainable prompt during multi-task learning, leads to improved promptability in both zero- and few-shot settings compared to existing methods, up to 31% relative. On the other hand, we find that continued pretraining using MAML-style meta-learning, a method that directly optimizes few-shot promptability, yields subpar performance. We validate our findings with two prompt tuning methods, and, based on our results, we provide concrete recommendations to optimize promptability for different use cases.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that a simple recipe, continued pretraining that incorporates a trainable prompt during multi-task learning, leads to improved promptability in both zero- and few-shot settings compared to existing methods, up to 31% relative.'}",http://arxiv.org/pdf/2210.10258
-What Makes Pre-trained Language Models Better Zero/Few-shot Learners?,Jinghui Lu,"In this paper, we propose a theoretical framework to explain the efficacy of prompt learning in zero/few-shot scenarios. First, we prove that conventional pre-training and fine-tuning paradigm fails in few-shot scenarios due to overfitting the unrepresentative labelled data. We then detail the assumption that prompt learning is more effective because it empowers pre-trained language model that is built upon massive text corpora, as well as domain-related human knowledge to participate more in prediction and thereby reduces the impact of limited label information provided by the small training set. We further hypothesize that language discrepancy can measure the qual- ity of prompting. Comprehensive experiments are performed to verify our assumptions. More remarkably, inspired by the theoretical framework, we propose an annotation-agnostic template selection method based on perplexity, which enables us to “forecast” the prompting performance in advance. This approach is especially encouraging because existing work still relies on development set to post-hoc evaluate templates. Experiments show that this method leads to significant prediction benefits compared to state-of-the-art zero-shot methods.","{'model': 'tldr@v2.0.0', 'text': 'A theoretical framework to explain theacy of prompt learning in zero/few-shot scenarios is proposed and an annotation-agnostic template selection method based on perplexity is proposed, which enables us to “forecast” the prompting performance in advance.'}",http://arxiv.org/pdf/2209.15206
-Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models,Lei Wang,"Large language models (LLMs) have recently been shown to deliver impressive performance in various NLP tasks. To tackle multi-step reasoning tasks, Few-shot chain-of-thought (CoT) prompting includes a few manually crafted step-by-step reasoning demonstrations which enable LLMs to explicitly generate reasoning steps and improve their reasoning task accuracy. To eliminate the manual efforts, Zero-shot-CoT concatenates the target problem statement with “Let’s think step by step” as an input prompt to LLMs. Despite the success of Zero-shot-CoT, it still suffers from three pitfalls: calculation errors, missing-step errors, and semantic misunderstanding errors. To address the missing-step errors, we propose Plan-and-Solve (PS) Prompting. It consists of two components: first, devising a plan to divide the entire task into smaller subtasks, and then carrying out the subtasks according to the plan. To address the calculation errors and improve the quality of generated reasoning steps, we extend PS prompting with more detailed instructions and derive PS+ prompting. We evaluate our proposed prompting strategy on ten datasets across three reasoning problems. The experimental results over GPT-3 show that our proposed zero-shot prompting consistently outperforms Zero-shot-CoT across all datasets by a large margin, is comparable to or exceeds Zero-shot-Program-of-Thought Prompting, and has comparable performance with 8-shot CoT prompting on the math reasoning problem. The code can be found at https://github.com/AGI-Edgerunners/Plan-and-Solve-Prompting.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results over GPT-3 show that the proposed zero-shot prompting consistently outperforms Zero- shot-CoT across all datasets by a large margin, is comparable to or exceeds Zero-shot-Program-of-Thought Prompting, and has comparable performance with 8-shot CoT prompting on the math reasoning problem.'}",http://arxiv.org/pdf/2305.04091
-Better Zero-Shot Reasoning with Self-Adaptive Prompting,Xingchen Wan,"Modern large language models (LLMs) have demonstrated impressive capabilities at sophisticated tasks, often through step-by-step reasoning similar to humans. This is made possible by their strong few and zero-shot abilities -- they can effectively learn from a handful of handcrafted, completed responses (""in-context examples""), or are prompted to reason spontaneously through specially designed triggers. Nonetheless, some limitations have been observed. First, performance in the few-shot setting is sensitive to the choice of examples, whose design requires significant human effort. Moreover, given the diverse downstream tasks of LLMs, it may be difficult or laborious to handcraft per-task labels. Second, while the zero-shot setting does not require handcrafting, its performance is limited due to the lack of guidance to the LLMs. To address these limitations, we propose Consistency-based Self-adaptive Prompting (COSP), a novel prompt design method for LLMs. Requiring neither handcrafted responses nor ground-truth labels, COSP selects and builds the set of examples from the LLM zero-shot outputs via carefully designed criteria that combine consistency, diversity and repetition. In the zero-shot setting for three different LLMs, we show that using only LLM predictions, COSP improves performance up to 15% compared to zero-shot baselines and matches or exceeds few-shot baselines for a range of reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Consistency-based Self-adaptive Prompting (COSP), a novel prompt design method for LLMs that selects and builds the set of examples from the LLM zero-shot outputs via carefully designed criteria that combine consistency, diversity and repetition.'}",http://arxiv.org/pdf/2305.14106
-GPT-3 for Few-Shot Dialogue State Tracking,Nicholas Pezzotti,"GPT-3 (Brown et al., 2020) has attracted considerable attention due to its superior performance across a wide range of Natural Language Processing (NLP) tasks, especially with its powerful and versatile in-context few-shot learning ability. That is, it has been shown that by carefully crafting a prompt, consisting of a few labelled examples followed by an unlabelled example, GPT’3 is able to do few-shot sentiment classification, three-digit arithmetic and much more. We seek to evaluate its performance on a novel and notably more complicated task: few-shot Dialogue State Tracking (DST). We propose a few-shot prompting framework that selects in-context examples based on similarity which outperforms the original random in-context selection framework. We also review and formalise the two types of completion strategies employed by previous literature, which we name constrained and unconstrained, and propose a third ""semi-constrained"" completion strategy, which is particularly well adapted for DST. Additionally, we propose a prompt ensembling technique that reliably outperforms individual models. Furthermore, we are the first, to the best of our knowledge, to fine-tune GPT-3 for the task of few-shot DST, showing that it reliably outperforms its GPT-2 counterpart. Furthermore, we seek to synthesise and formalise the largely heterogeneous body of previous work on prompt programming and in-context learning for GPT-3. In an attempt to contribute to the understanding of the strengths, weaknesses and inner-working of GPT-3, we perform numerous ablative studies that validate and confute previous in-context learning empirical findings: mainly, we find that natural language instructions in the prompt have little impact on performance, larger language models do not always induce higher downstream performance and that GPT-3 is highly sensitive to the order and number of the in-context examples.","{'model': 'tldr@v2.0.0', 'text': 'It is found that natural language instructions in the prompt have little impact on performance, larger language models do not always induce higher downstream performance and that GPT-3 is highly sensitive to the order and number of the in-context examples.'}",
-Multi-modal Prompting for Low-Shot Temporal Action Localization,Chen Ju,"In this paper, we consider the problem of temporal action localization under low-shot (zero-shot&few-shot) scenario, with the goal of detecting and classifying the action instances from arbitrary categories within some untrimmed videos, even not seen at training time. We adopt a Transformer-based two-stage action localization architecture with class-agnostic action proposal, followed by open-vocabulary classification. We make the following contributions. First, to compensate image-text foundation models with temporal motions, we improve category-agnostic action proposal by explicitly aligning embeddings of optical flows, RGB and texts, which has largely been ignored in existing low-shot methods. Second, to improve open-vocabulary action classification, we construct classifiers with strong discriminative power, i.e., avoid lexical ambiguities. To be specific, we propose to prompt the pre-trained CLIP text encoder either with detailed action descriptions (acquired from large-scale language models), or visually-conditioned instance-specific prompt vectors. Third, we conduct thorough experiments and ablation studies on THUMOS14 and ActivityNet1.3, demonstrating the superior performance of our proposed model, outperforming existing state-of-the-art approaches by one significant margin.","{'model': 'tldr@v2.0.0', 'text': 'This paper adopts a Transformer-based two-stage action localization architecture with class-agnostic action proposal, followed by open-vocabulary classification, and proposes to prompt the pre-trained CLIP text encoder either with detailed action descriptions, or visually-conditioned instance-specific prompt vectors.'}",http://arxiv.org/pdf/2303.11732
-Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks,Wenhu Chen,"Recently, there has been significant progress in teaching language models to perform step-by-step reasoning to solve complex numerical reasoning tasks. Chain-of-thoughts prompting (CoT) is by far the state-of-art method for these tasks. CoT uses language models to perform both reasoning and computation in the multi-step `thought' process. To disentangle computation from reasoning, we propose `Program of Thoughts' (PoT), which uses language models (mainly Codex) to express the reasoning process as a program. The computation is relegated to an external computer, which executes the generated programs to derive the answer. We evaluate PoT on five math word problem datasets (GSM, AQuA, SVAMP, TabMWP, MultiArith) and three financial-QA datasets (FinQA, ConvFinQA, TATQA) for both few-shot and zero-shot setups. Under both few-shot and zero-shot settings, PoT can show an average performance gain over CoT by around 12\% across all the evaluated datasets. By combining PoT with self-consistency decoding, we can achieve SoTA performance on all math problem datasets and near-SoTA performance on financial datasets. All of our data and code are released in Github https://github.com/wenhuchen/Program-of-Thoughts","{'model': 'tldr@v2.0.0', 'text': 'Under both few-shot and zero-shot settings, PoT can show an average performance gain over CoT by around 12\\% across all the evaluated datasets, and by combining PoT with self-consistency decoding, can achieve SoTA performance on all math problem datasets and near-SoTAperformance on financial datasets.'}",http://arxiv.org/pdf/2211.12588
-Generative Zero-Shot Prompt Learning for Cross-Domain Slot Filling with Inverse Prompting,Xuefeng Li,"Zero-shot cross-domain slot filling aims to transfer knowledge from the labeled source domain to the unlabeled target domain. Existing models either encode slot descriptions and examples or design handcrafted question templates using heuristic rules, suffering from poor generalization capability or robustness. In this paper, we propose a generative zero-shot prompt learning framework for cross-domain slot filling, both improving generalization and robustness than previous work. Besides, we introduce a novel inverse prompting strategy to distinguish different slot types to avoid the multiple prediction problem, and an efficient prompt-tuning strategy to boost higher performance by only training fewer prompt parameters. Experiments and analysis demonstrate the effectiveness of our proposed framework, especially huge improvements (+13.44% F1) on the unseen slots.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a generative zero-shot prompt learning framework for cross-domain slot filling, both improving generalization and robustness than previous work, and introduces a novel inverse prompting strategy to distinguish different slot types to avoid the multiple prediction problem.'}",https://arxiv.org/pdf/2307.02830
-Decomposed Prompting: A Modular Approach for Solving Complex Tasks,Tushar Khot,"Few-shot prompting is a surprisingly powerful way to use Large Language Models (LLMs) to solve various tasks. However, this approach struggles as the task complexity increases or when the individual reasoning steps of the task themselves are hard to learn, especially when embedded in more complex tasks. To address this, we propose Decomposed Prompting, a new approach to solve complex tasks by decomposing them (via prompting) into simpler sub-tasks that can be delegated to a library of prompting-based LLMs dedicated to these sub-tasks. This modular structure allows each prompt to be optimized for its specific sub-task, further decomposed if necessary, and even easily replaced with more effective prompts, trained models, or symbolic functions if desired. We show that the flexibility and modularity of Decomposed Prompting allows it to outperform prior work on few-shot prompting using GPT3. On symbolic reasoning tasks, we can further decompose sub-tasks that are hard for LLMs into even simpler solvable sub-tasks. When the complexity comes from the input length, we can recursively decompose the task into the same task but with smaller inputs. We also evaluate our approach on textual multi-step reasoning tasks: on long-context multi-hop QA task, we can more effectively teach the sub-tasks via our separate sub-tasks prompts; and on open-domain multi-hop QA, we can incorporate a symbolic information retrieval within our decomposition framework, leading to improved performance on both tasks. Datasets, Code and Prompts available at https://github.com/allenai/DecomP.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the flexibility and modularity of Decomposed Prompting allows it to outperform prior work on few-shot prompting using GPT3 and to incorporate a symbolic information retrieval within the decomposition framework, leading to improved performance on both tasks.'}",http://arxiv.org/pdf/2210.02406
-Language-Aware Soft Prompting: Text-to-Text Optimization for Few- and Zero-Shot Adaptation of V &L Models,Adrian Bulat,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Language-Aware Soft Prompting (LASP) learning method by means of a text-to-text cross-entropy loss that maximizes the probability of the learned prompts to be correctly classified with respect to pre-defined hand-crafted textual prompts and presents a novel zero-shot variant of LASP.'}",https://link.springer.com/content/pdf/10.1007/s11263-023-01904-9.pdf
-Prompting GPT-3 To Be Reliable,Chenglei Si,"Large language models (LLMs) show impressive abilities via few-shot prompting. Commercialized APIs such as OpenAI GPT-3 further increase their use in real-world language applications. However, the crucial problem of how to improve the reliability of GPT-3 is still under-explored. While reliability is a broad and vaguely defined term, we decompose reliability into four main facets that correspond to the existing framework of ML safety and are well-recognized to be important: generalizability, social biases, calibration, and factuality. Our core contribution is to establish simple and effective prompts that improve GPT-3's reliability as it: 1) generalizes out-of-distribution, 2) balances demographic distribution and uses natural language instructions to reduce social biases, 3) calibrates output probabilities, and 4) updates the LLM's factual knowledge and reasoning chains. With appropriate prompts, GPT-3 is more reliable than smaller-scale supervised models on all these facets. We release all processed datasets, evaluation scripts, and model predictions. Our systematic empirical study not only sheds new insights on the reliability of prompting LLMs, but more importantly, our prompting strategies can help practitioners more reliably use LLMs like GPT-3.","{'model': 'tldr@v2.0.0', 'text': 'This systematic empirical study sheds new insights on the reliability of prompting LLMs, but more importantly, its prompting strategies can help practitioners more reliably use LLMs like GPT-3.'}",http://arxiv.org/pdf/2210.09150
-"Large Language Model Is Not a Good Few-shot Information Extractor, but a Good Reranker for Hard Samples!",Yubo Ma,"Large Language Models (LLMs) have made remarkable strides in various tasks. Whether LLMs are competitive few-shot solvers for information extraction (IE) tasks, however, remains an open problem. In this work, we aim to provide a thorough answer to this question. Through extensive experiments on nine datasets across four IE tasks, we demonstrate that current advanced LLMs consistently exhibit inferior performance, higher latency, and increased budget requirements compared to fine-tuned SLMs under most settings. Therefore, we conclude that LLMs are not effective few-shot information extractors in general. Nonetheless, we illustrate that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with. And moreover, we propose an adaptive filter-then-rerank paradigm to combine the strengths of LLMs and SLMs. In this paradigm, SLMs serve as filters and LLMs serve as rerankers. By prompting LLMs to rerank a small portion of difficult samples identified by SLMs, our preliminary system consistently achieves promising improvements (2.4% F1-gain on average) on various IE tasks, with an acceptable time and cost investment.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with, and an adaptive filter-then-rerank paradigm is proposed to combine the strengths of LLMs and SLMs.'}",http://arxiv.org/pdf/2303.08559
-ZARA: Improving Few-Shot Self-Rationalization for Small Language Models,Wei-Lin Chen,"Language models (LMs) that jointly generate end-task answers as well as free-text rationales are known as self-rationalization models. Recent works demonstrate great performance gain for self-rationalization by few-shot prompting LMs with rationale-augmented exemplars. However, the ability to benefit from explanations only emerges with large-scale LMs, which have poor accessibility. In this work, we explore the less-studied setting of leveraging explanations for small LMs to improve few-shot self-rationalization. We first revisit the relationship between rationales and answers. Inspired by the implicit mental process of how human beings assess explanations, we present a novel approach, Zero-shot Augmentation of Rationale-Answer pairs (ZARA), to automatically construct pseudo-parallel data for self-training by reducing the problem of plausibility judgement to natural language inference. Experimental results show ZARA achieves SOTA performance on the FEB benchmark, for both the task accuracy and the explanation metric. In addition, we conduct human and quantitative evaluation validating ZARA's ability to automatically identify plausible and accurate rationale-answer pairs.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a novel approach, Zero-shot Augmentation of Rationale-Answer pairs (ZARA), to automatically construct pseudo-parallel data for self-training by reducing the problem of plausibility judgement to natural language inference.'}",http://arxiv.org/pdf/2305.07355
-Skill-Based Few-Shot Selection for In-Context Learning,Shengnan An,"In-context learning is the paradigm that adapts large language models to downstream tasks by providing a few examples. Few-shot selection -- selecting appropriate examples for each test instance separately -- is important for in-context learning. In this paper, we propose Skill-KNN, a skill-based few-shot selection method for in-context learning. The key advantages of Skill-KNN include: (1) it addresses the problem that existing methods based on pre-trained embeddings can be easily biased by surface natural language features that are not important for the target task; (2) it does not require training or fine-tuning of any models, making it suitable for frequently expanding or changing example banks. The key insight is to optimize the inputs fed into the embedding model, rather than tuning the model itself. Technically, Skill-KNN generates the skill-based descriptions for each test case and candidate example by utilizing a pre-processing few-shot prompting, thus eliminating unimportant surface features. Experimental results across five cross-domain semantic parsing datasets and six backbone models show that Skill-KNN significantly outperforms existing methods.","{'model': 'tldr@v2.0.0', 'text': 'Skill-KNN addresses the problem that existing methods based on pre-trained embeddings can be easily biased by surface natural language features that are not important for the target task, and does not require training or fine-tuning of any models, making it suitable for frequently expanding or changing example banks.'}",https://arxiv.org/pdf/2305.14210
-Towards Few-Shot Identification of Morality Frames using In-Context Learning,Shamik Roy,"Data scarcity is a common problem in NLP, especially when the annotation pertains to nuanced socio-linguistic concepts that require specialized knowledge. As a result, few-shot identification of these concepts is desirable. Few-shot in-context learning using pre-trained Large Language Models (LLMs) has been recently applied successfully in many NLP tasks. In this paper, we study few-shot identification of a psycho-linguistic concept, Morality Frames (Roy et al., 2021), using LLMs. Morality frames are a representation framework that provides a holistic view of the moral sentiment expressed in text, identifying the relevant moral foundation (Haidt and Graham, 2007) and at a finer level of granularity, the moral sentiment expressed towards the entities mentioned in the text. Previous studies relied on human annotation to identify morality frames in text which is expensive. In this paper, we propose prompting based approaches using pretrained Large Language Models for identification of morality frames, relying only on few-shot exemplars. We compare our models’ performance with few-shot RoBERTa and found promising results.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes prompting based approaches using pretrained Large Language Models for identification of morality frames, relying only on few-shot exemplars, and compares the models’ performance with few- shot RoBERTa and found promising results.'}",http://arxiv.org/pdf/2302.02029
-Enhancing Few-shot Text-to-SQL Capabilities of Large Language Models: A Study on Prompt Design Strategies,Linyong Nan,"In-context learning (ICL) has emerged as a new approach to various natural language processing tasks, utilizing large language models (LLMs) to make predictions based on context that has been supplemented with a few examples or task-specific instructions. In this paper, we aim to extend this method to question answering tasks that utilize structured knowledge sources, and improve Text-to-SQL systems by exploring various prompt design strategies for employing LLMs. We conduct a systematic investigation into different demonstration selection methods and optimal instruction formats for prompting LLMs in the Text-to-SQL task. Our approach involves leveraging the syntactic structure of an example's SQL query to retrieve demonstrations, and we demonstrate that pursuing both diversity and similarity in demonstration selection leads to enhanced performance. Furthermore, we show that LLMs benefit from database-related knowledge augmentations. Our most effective strategy outperforms the state-of-the-art system by 2.5 points (Execution Accuracy) and the best fine-tuned system by 5.1 points on the Spider dataset. These results highlight the effectiveness of our approach in adapting LLMs to the Text-to-SQL task, and we present an analysis of the factors contributing to the success of our strategy.","{'model': 'tldr@v2.0.0', 'text': 'This paper conducts a systematic investigation into different demonstration selection methods and optimal instruction formats for prompting LLMs in the Text-to-SQL task, and presents an analysis of the factors contributing to the success of this strategy.'}",http://arxiv.org/pdf/2305.12586
-Exploiting the Potential of Seq2Seq Models as Robust Few-Shot Learners,Jihyeon Janel Lee,"In-context learning, which offers substantial advantages over fine-tuning, is predominantly observed in decoder-only models, while encoder-decoder (i.e., seq2seq) models excel in methods that rely on weight updates. Recently, a few studies have demonstrated the feasibility of few-shot learning with seq2seq models; however, this has been limited to tasks that align well with the seq2seq architecture, such as summarization and translation. Inspired by these initial studies, we provide a first-ever extensive experiment comparing the in-context few-shot learning capabilities of decoder-only and encoder-decoder models on a broad range of tasks. Furthermore, we propose two methods to more effectively elicit in-context learning ability in seq2seq models: objective-aligned prompting and a fusion-based approach. Remarkably, our approach outperforms a decoder-only model that is six times larger and exhibits significant performance improvements compared to conventional seq2seq models across a variety of settings. We posit that, with the right configuration and prompt design, seq2seq models can be highly effective few-shot learners for a wide spectrum of applications.","{'model': 'tldr@v2.0.0', 'text': 'It is posited that, with the right configuration and prompt design, seq2seq models can be highly effective few-shot learners for a wide spectrum of applications.'}",https://arxiv.org/pdf/2307.14856
-Few-Shot and Prompt Training for Text Classification in German Doctor's Letters,Phillip Richter-Pechanski,"To classify sentences in cardiovascular German doctor's letters into eleven section categories, we used pattern-exploiting training, a prompt-based method for text classification in few-shot learning scenarios (20, 50 and 100 instances per class) using language models with various pre-training approaches evaluated on CARDIO:DE, a freely available German clinical routine corpus. Prompting improves results by 5-28% accuracy compared to traditional methods, reducing manual annotation efforts and computational costs in a clinical setting.","{'model': 'tldr@v2.0.0', 'text': ""To classify sentences in cardiovascular German doctor's letters into eleven section categories, pattern-exploiting training is used, a prompt-based method for text classification in few-shot learning scenarios, improving results by 5-28% accuracy compared to traditional methods and reducing manual annotation efforts and computational costs in a clinical setting.""}",https://ebooks.iospress.nl/pdf/doi/10.3233/SHTI230275
-Exploring Zero and Few-shot Techniques for Intent Classification,S. Parikh,"Conversational NLU providers often need to scale to thousands of intent-classification models where new customers often face the cold-start problem. Scaling to so many customers puts a constraint on storage space as well. In this paper, we explore four different zero and few-shot intent classification approaches with this low-resource constraint: 1) domain adaptation, 2) data augmentation, 3) zero-shot intent classification using descriptions large language models (LLMs), and 4) parameter-efficient fine-tuning of instruction-finetuned language models. Our results show that all these approaches are effective to different degrees in low-resource settings. Parameter-efficient fine-tuning using T-few recipe on Flan-T5 yields the best performance even with just one sample per intent. We also show that the zero-shot method of prompting LLMs using intent descriptions is also very competitive.","{'model': 'tldr@v2.0.0', 'text': 'Parameter-efficient fine-tuning using T-few recipe on Flan-T5 yields the best performance even with just one sample per intent, and the zero-shot method of prompting LLMs using intent descriptions is also very competitive.'}",http://arxiv.org/pdf/2305.07157
-Knowledge-Guided Prompt Learning for Few-Shot Text Classification,Liangguo Wang,"Recently, prompt-based learning has shown impressive performance on various natural language processing tasks in few-shot scenarios. The previous study of knowledge probing showed that the success of prompt learning contributes to the implicit knowledge stored in pre-trained language models. However, how this implicit knowledge helps solve downstream tasks remains unclear. In this work, we propose a knowledge-guided prompt learning method that can reveal relevant knowledge for text classification. Specifically, a knowledge prompting template and two multi-task frameworks were designed, respectively. The experiments demonstrated the superiority of combining knowledge and prompt learning in few-shot text classification.","{'model': 'tldr@v2.0.0', 'text': 'A knowledge-guided prompt learning method that can reveal relevant knowledge for text classification and demonstrates the superiority of combining knowledge and prompt learning in few-shot text classification.'}",https://www.mdpi.com/2079-9292/12/6/1486/pdf?version=1679462243
-A smashed glass cannot be full: Generation of Commonsense Explanations through Prompt-based Few-shot Learning,Andrea Zaninello,"We assume that providing explanations is a process to elicit implicit knowledge in human communication, and propose a general methodology to generate commonsense explanations from pairs of semantically related sentences. We take advantage of both prompting applied to large, encoder-decoder pre-trained language models, and few-shot learning techniques, such as pattern-exploiting training. Experiments run on the e-SNLI dataset show that the proposed method achieves state-of-the-art results on the explanation generation task, with a substantial reduction of labelled data. The obtained results open new perspective on a number of tasks involving the elicitation of implicit knowledge.","{'model': 'tldr@v2.0.0', 'text': 'A general methodology to generate commonsense explanations from pairs of semantically related sentences is proposed, taking advantage of both prompting applied to large, encoder-decoder pre-trained language models, and few-shot learning techniques, such as pattern-exploiting training.'}",https://aclanthology.org/2023.nlrse-1.3.pdf
-Prompting PaLM for Translation: Assessing Strategies and Performance,David Vilar,"Large language models (LLMs) that have been trained on multilingual but not parallel text exhibit a remarkable ability to translate between languages. We probe this ability in an in-depth study of the pathways language model (PaLM), which has demonstrated the strongest machine translation (MT) performance among similarly-trained LLMs to date. We investigate various strategies for choosing translation examples for few-shot prompting, concluding that example quality is the most important factor. Using optimized prompts, we revisit previous assessments of PaLM’s MT capabilities with more recent test sets, modern MT metrics, and human evaluation, and find that its performance, while impressive, still lags that of state-of-the-art supervised systems. We conclude by providing an analysis of PaLM’s MT output which reveals some interesting properties and prospects for future work.","{'model': 'tldr@v2.0.0', 'text': 'An in-depth study of the pathways language model (PaLM), which has demonstrated the strongest machine translation performance among similarly-trained LLMs to date, investigates various strategies for choosing translation examples for few-shot prompting, concluding that example quality is the most important factor.'}",http://arxiv.org/pdf/2211.09102
-Successive Prompting for Decomposing Complex Questions,Dheeru Dua,"Answering complex questions that require making latent decisions is a challenging task, especially when limited supervision is available. Recent works leverage the capabilities of large language models (LMs) to perform complex question answering in a few-shot setting by demonstrating how to output intermediate rationalizations while solving the complex question in a single pass. We introduce “Successive Prompting” where, we iteratively break down a complex task into a simple task, solve it, and then repeat the process until we get the final solution. Successive prompting decouples the supervision for decomposing complex questions from the supervision for answering simple questions, allowing us to (1) have multiple opportunities to query in-context examples at each reasoning step (2) learn question decomposition separately from question answering, including using synthetic data, and (3) use bespoke (fine-tuned) components for reasoning steps where a large LM does not perform well. The intermediate supervision is typically manually written, which can be expensive to collect. We introduce a way to generate synthetic dataset which can be used to bootstrap model’s ability to decompose and answer intermediate questions. Our best model (with successive prompting) achieves an improvement in F1 of ~5% when compared with a state-of-the-art model with synthetic augmentations and few-shot version of the DROP dataset.","{'model': 'tldr@v2.0.0', 'text': 'A way to generate synthetic dataset which can be used to bootstrap model’s ability to decompose and answer intermediate questions is introduced and achieves an improvement in F1 of ~5% when compared with a state-of-the-art model with synthetic augmentations and few-shot version of the DROP dataset.'}",https://arxiv.org/pdf/2212.04092
-Large Language Models are few(1)-shot Table Reasoners,Wenhu Chen,"Recent literature has shown that large language models (LLMs) are generally excellent few-shot reasoners to solve text reasoning tasks. However, the capability of LLMs on table reasoning tasks is yet to be explored. In this paper, we aim at understanding how well LLMs can perform table-related tasks with few-shot in-context learning. Specifically, we evaluated LLMs on popular table QA and fact verification datasets like WikiTableQuestion, FetaQA, TabFact, and FEVEROUS and found that LLMs are competent at complex reasoning over table structures, though these models are not pre-trained on any table corpus. When combined with ‘chain of thoughts’ prompting, LLMs can achieve very strong performance with only a 1-shot demonstration, even on par with some SoTA models. We show that LLMs are even more competent at generating comprehensive long-form answers on FetaQA than tuned T5-large. We further manually studied the reasoning chains elicited from LLMs and found that these reasoning chains are highly consistent with the underlying semantic form. We believe that LLMs can serve as a simple yet generic baseline for future research. The code and data are released in {url{https://github.com/wenhuchen/TableCoT}.","{'model': 'tldr@v2.0.0', 'text': 'This paper evaluated LLMs on popular table QA and fact verification datasets like WikiTableQuestion, FetaQA, TabFact, and FEVEROUS and found that LLMs are competent at complex reasoning over table structures, though these models are not pre-trained on any table corpus.'}",http://arxiv.org/pdf/2210.06710
-"Machine Translation with Large Language Models: Prompting, Few-shot Learning, and Fine-tuning with QLoRA",Xuan Zhang,"While large language models have made remarkable advancements in natural language generation, their potential in machine translation, especially when fine-tuned, remains under-explored. In our study, we conduct comprehensive experiments, evaluating 15 publicly available language models on machine translation tasks. We compare the performance across three methodologies: zero-shot prompting, few-shot learning, and fine-tuning. Central to our approach is the use of QLoRA, an efficient fine-tuning method. On French-English, QLoRA fine-tuning outperforms both few-shot learning and models trained from scratch. This superiority is highlighted in both sentence-level and document-level translations, with a significant BLEU score improvement of 28.93 over the prompting method. Impressively, with QLoRA, the enhanced performance is achieved by fine-tuning a mere 0.77% of the model’s parameters.","{'model': 'tldr@v2.0.0', 'text': 'On French-English, QLoRA fine-tuning outperforms both few-shot learning and models trained from scratch, and is highlighted in both sentence-level and document-level translations.'}",
-"Structured Prompting: Scaling In-Context Learning to 1, 000 Examples",Y. Hao,"Large language models have exhibited intriguing in-context learning capability, achieving promising zero- and few-shot performance without updating the parameters. However, conventional in-context learning is usually restricted by length constraints, rendering it ineffective to absorb supervision from a large number of examples. In order to go beyond few shots, we introduce structured prompting that breaks the length limit and scales in-context learning to thousands of examples. Specifically, demonstration examples are separately encoded with well-designed position embeddings, and then they are jointly attended by the test example using a rescaled attention mechanism. So we can scale the number of exemplars with linear complexity instead of quadratic complexity with respect to length. Experimental results on a diverse set of tasks show that our approach improves end-task performance and reduces evaluation variance over conventional in-context learning as the number of demonstration examples increases. Code has been released at https://aka.ms/structured-prompting .","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on a diverse set of tasks show that the structured prompting approach improves end-task performance and reduces evaluation variance over conventional in-context learning as the number of demonstration examples increases.'}",http://arxiv.org/pdf/2212.06713
-Discrete and Soft Prompting for Multilingual Models,Mengjie Zhao,"It has been shown for English that discrete and soft prompting perform strongly in few-shot learning with pretrained language models (PLMs). In this paper, we show that discrete and soft prompting perform better than finetuning in multilingual cases: Crosslingual transfer and in-language training of multilingual natural language inference. For example, with 48 English training examples, finetuning obtains 33.74% accuracy in crosslingual transfer, barely surpassing the majority baseline (33.33%). In contrast, discrete and soft prompting outperform finetuning, achieving 36.43% and 38.79%. We also demonstrate good performance of prompting with training data in multiple languages other than English.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that discrete and soft prompting perform better than finetuning in multilingual cases: Crosslingual transfer and in-language training of multilingual natural language inference.'}",https://aclanthology.org/2021.emnlp-main.672.pdf
-Short Answer Grading Using One-shot Prompting and Text Similarity Scoring Model,Su-Youn Yoon,"In this study, we developed an automated short answer grading (ASAG) model that provided both analytic scores and final holistic scores. Short answer items typically consist of multiple sub-questions, and providing an analytic score and the text span relevant to each sub-question can increase the interpretability of the automated scores. Furthermore, they can be used to generate actionable feedback for students. Despite these advantages, most studies have focused on predicting only holistic scores due to the difficulty in constructing dataset with manual annotations. To address this difficulty, we used large language model (LLM)-based one-shot prompting and a text similarity scoring model with domain adaptation using small manually annotated dataset. The accuracy and quadratic weighted kappa of our model were 0.67 and 0.71 on a subset of the publicly available ASAG dataset. The model achieved a substantial improvement over the majority baseline.","{'model': 'tldr@v2.0.0', 'text': 'An automated short answer grading (ASAG) model that provided both analytic scores and final holistic scores was developed that achieved a substantial improvement over the majority baseline.'}",http://arxiv.org/pdf/2305.18638
-Zero-Shot Prompting for Implicit Intent Prediction and Recommendation with Commonsense Reasoning,Hui-Chi Kuo,"Intelligent virtual assistants are currently designed to perform tasks or services explicitly mentioned by users, so multiple related domains or tasks need to be performed one by one through a long conversation with many explicit intents. Instead, human assistants are capable of reasoning (multiple) implicit intents based on user utterances via commonsense knowledge, reducing complex interactions and improving practicality. Therefore, this paper proposes a framework of multi-domain dialogue systems, which can automatically infer implicit intents based on user utterances and then perform zero-shot prompting using a large pre-trained language model to trigger suitable single task-oriented bots. The proposed framework is demonstrated effective to realize implicit intents and recommend associated bots in a zero-shot manner.","{'model': 'tldr@v2.0.0', 'text': 'A framework of multi-domain dialogue systems which can automatically infer implicit intents based on user utterances and then perform zero-shot prompting using a large pre-trained language model to trigger suitable single task-oriented bots is proposed.'}",http://arxiv.org/pdf/2210.05901
-Personalize Segment Anything Model with One Shot,Renrui Zhang,"Driven by large-data pre-training, Segment Anything Model (SAM) has been demonstrated as a powerful and promptable framework, revolutionizing the segmentation models. Despite the generality, customizing SAM for specific visual concepts without man-powered prompting is under explored, e.g., automatically segmenting your pet dog in different images. In this paper, we propose a training-free Personalization approach for SAM, termed as PerSAM. Given only a single image with a reference mask, PerSAM first localizes the target concept by a location prior, and segments it within other images or videos via three techniques: target-guided attention, target-semantic prompting, and cascaded post-refinement. In this way, we effectively adapt SAM for private use without any training. To further alleviate the mask ambiguity, we present an efficient one-shot fine-tuning variant, PerSAM-F. Freezing the entire SAM, we introduce two learnable weights for multi-scale masks, only training 2 parameters within 10 seconds for improved performance. To demonstrate our efficacy, we construct a new segmentation dataset, PerSeg, for personalized evaluation, and test our methods on video object segmentation with competitive performance. Besides, our approach can also enhance DreamBooth to personalize Stable Diffusion for text-to-image generation, which discards the background disturbance for better target appearance learning. Code is released at https://github.com/ZrrSkywalker/Personalize-SAM","{'model': 'tldr@v2.0.0', 'text': 'A training-free Personalization approach for SAM, termed as PerSAM, which effectively adapt SAM for private use without any training, and can enhance DreamBooth to personalize Stable Diffusion for text-to-image generation, which discards the background disturbance for better target appearance learning.'}",https://arxiv.org/pdf/2305.03048
-One-Shot Labeling for Automatic Relevance Estimation,Sean MacAvaney,"Dealing with unjudged documents (""holes"") in relevance assessments is a perennial problem when evaluating search systems with offline experiments. Holes can reduce the apparent effectiveness of retrieval systems during evaluation and introduce biases in models trained with incomplete data. In this work, we explore whether large language models can help us fill such holes to improve offline evaluations. We examine an extreme, albeit common, evaluation setting wherein only a single known relevant document per query is available for evaluation. We then explore various approaches for predicting the relevance of unjudged documents with respect to a query and the known relevant document, including nearest neighbor, supervised, and prompting techniques. We find that although the predictions of these One-Shot Labelers (1SL) frequently disagree with human assessments, the labels they produce yield a far more reliable ranking of systems than the single labels do alone. Specifically, the strongest approaches can consistently reach system ranking correlations of over 0.86 with the full rankings over a variety of measures. Meanwhile, the approach substantially increases the reliability of t-tests due to filling holes in relevance assessments, giving researchers more confidence in results they find to be significant. Alongside this work, we release an easy-to-use software package to enable the use of 1SL for evaluation of other ad-hoc collections or systems.","{'model': 'tldr@v2.0.0', 'text': 'This work examines an extreme evaluation setting wherein only a single known relevant document per query is available for evaluation, and finds that although the predictions of these One-Shot Labelers (1SL) frequently disagree with human assessments, the labels they produce yield a far more reliable ranking of systems than the single labels do alone.'}",https://arxiv.org/pdf/2302.11266
-One-Shot Learning for Function-Specific Region Segmentation in Mouse Brain,Xu Zhang,"A brain contains a large number of structured regions responsible for diverse functions. Detailed region annotations upon stereotaxic coordinates are highly rare, prompting the need of using one or very few available annotated results of a specific brain section to label images of broadly accessible brain section samples. Here we develop a one-shot learning approach to segment regions of mouse brains. Using the highly ordered geometry of brains, we introduce a reference mask to incorporate both the anatomical structure (visual information) and the brain atlas into brain segmentation. Using the UNet model with this reference mask, we are able to predict the region of hippocampus with high accuracy. We further implement it to segment brain images into 95 detailed regions augmented from the annotation on only one image from Allen Brain Atlas. Together, our one-shot learning method provides neuroscientists an efficient way for brain segmentation and facilitates future region-specific functional studies of brains.","{'model': 'tldr@v2.0.0', 'text': 'A one-shot learning approach to segment regions of mouse brains using a reference mask to incorporate both the anatomical structure and the brain atlas into brain segmentation and is able to predict the region of hippocampus with high accuracy.'}",
-NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers,Kai Shen,"Scaling text-to-speech (TTS) to large-scale, multi-speaker, and in-the-wild datasets is important to capture the diversity in human speech such as speaker identities, prosodies, and styles (e.g., singing). Current large TTS systems usually quantize speech into discrete tokens and use language models to generate these tokens one by one, which suffer from unstable prosody, word skipping/repeating issue, and poor voice quality. In this paper, we develop NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input. To enhance the zero-shot capability that is important to achieve diverse speech synthesis, we design a speech prompting mechanism to facilitate in-context learning in the diffusion model and the duration/pitch predictor. We scale NaturalSpeech 2 to large-scale datasets with 44K hours of speech and singing data and evaluate its voice quality on unseen speakers. NaturalSpeech 2 outperforms previous TTS systems by a large margin in terms of prosody/timbre similarity, robustness, and voice quality in a zero-shot setting, and performs novel zero-shot singing synthesis with only a speech prompt. Audio samples are available at https://speechresearch.github.io/naturalspeech2.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input.'}",http://arxiv.org/pdf/2304.09116
-The Correlation Of A Scientific Approach Using The Probing Prompting Technique And A Problem-Based Learning Model On Learning Outcomes On Reaction Rate Material,Wardhatul Jannah,"Abstract─This study began with the learning results on the subject of the still-low reaction rate. The purpose of this study is to discover the relationship between the scientific method and learning outcomes in the reaction rate material by using the probing prompting technique and the issue based learning model alone or combination. Quantitative research using correlational methodologies and a one-shot case study design is the approach of choice. Class XI MIPA SMAN 1 Padang and XI MIPA SMAN 14 Padang were the subjects of this research. A total of 144 persons were selected for the study using a simple random sampling technique. Questionnaires and learning outcomes assessments in the form of 22 objective questions that had been verified for validity and reliability were used in this study. The results found a positive and significant correlation between: 1) the scientific approach to learning outcomes using the probing prompting technique, with a correlation coefficient of 0.903 suggesting a very strong correlation. This suggests that by using the probing prompting technique, the scientific approach module can improve learning results by 81.5 percent. 2) A scientific approach to learning outcomes using a problem-based learning model with a correlation coefficient of 0.770, suggesting a strong link. This suggests that the scientific approach module, when combined with a problem-based learning model, has an effective contribution of 59.3 percent in improving learning outcomes. 3) Using the probing prompting technique and the issue based learning model together on the reaction rate learning outcomes yielded a correlation coefficient of 0.958, indicating a very strong correlation. The contribution of the two independent variables to the dependent variable in this study is 91.8 percent, with the remainder impacted by other variables. Keywords─Scientific Approach; Probing Prompting; Problem-Based Learning; Learning Outcomes; Reaction Rate","{'model': 'tldr@v2.0.0', 'text': 'The results found a positive and significant correlation between the scientific approach to learning outcomes using the probing prompting technique and the issue based learning model alone or combination, and a correlation coefficient of 0.903 suggesting a very strong correlation.'}",
-Large Language Models Vote: Prompting for Rare Disease Identification,David Oniani,"The emergence of generative Large Language Models (LLMs) emphasizes the need for accurate and efficient prompting approaches. LLMs are often applied in Few-Shot Learning (FSL) contexts, where tasks are executed with minimal training data. FSL has become popular in many Artificial Intelligence (AI) subdomains, including AI for health. Rare diseases affect a small fraction of the population. Rare disease identification from clinical notes inherently requires FSL techniques due to limited data availability. Manual data collection and annotation is both expensive and time-consuming. In this paper, we propose Models-Vote Prompting (MVP), a flexible prompting approach for improving the performance of LLM queries in FSL settings. MVP works by prompting numerous LLMs to perform the same tasks and then conducting a majority vote on the resulting outputs. This method achieves improved results to any one model in the ensemble on one-shot rare disease identification and classification tasks. We also release a novel rare disease dataset for FSL, available to those who signed the MIMIC-IV Data Use Agreement (DUA). Furthermore, in using MVP, each model is prompted multiple times, substantially increasing the time needed for manual annotation, and to address this, we assess the feasibility of using JSON for automating generative LLM evaluation.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Models-Vote Prompting (MVP), a flexible prompting approach for improving the performance of LLM queries in FSL settings that achieves improved results to any one model in the ensemble on one-shot rare disease identification and classification tasks.'}",https://arxiv.org/pdf/2308.12890
-Universal Self-adaptive Prompting,Xingchen Wan,"A hallmark of modern large language models (LLMs) is their impressive general zero-shot and few-shot abilities, often elicited through in-context learning (ICL) via prompting. However, while highly coveted and being the most general, zero-shot performances in LLMs are still typically weaker due to the lack of guidance and the difficulty of applying existing automatic prompt design methods in general tasks when ground-truth labels are unavailable. In this study, we address this by presenting Universal Self-Adaptive Prompting (USP), an automatic prompt design approach specifically tailored for zero-shot learning (while compatible with few-shot). Requiring only a small amount of unlabeled data and an inference-only LLM, USP is highly versatile: to achieve universal prompting, USP categorizes a possible NLP task into one of the three possible task types and then uses a corresponding selector to select the most suitable queries and zero-shot model-generated responses as pseudo-demonstrations, thereby generalizing ICL to the zero-shot setup in a fully automated way. We evaluate USP with PaLM and PaLM 2 models and demonstrate performances that are considerably stronger than standard zero-shot baselines and often comparable to or even superior to few-shot baselines across more than 40 natural language understanding, natural language generation, and reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This study presents Universal Self-Adaptive Prompting (USP), an automatic prompt design approach specifically tailored for zero-shot learning (while compatible with few-shot) and evaluates USP with PaLM and PaLM 2 models and demonstrates performances that are considerably stronger than standard zero- shot baselines.'}",http://arxiv.org/pdf/2305.14926
-UDAPDR: Unsupervised Domain Adaptation via LLM Prompting and Distillation of Rerankers,Jon Saad-Falcon,"Many information retrieval tasks require large labeled datasets for fine-tuning. However, such datasets are often unavailable, and their utility for real-world applications can diminish quickly due to domain shifts. To address this challenge, we develop and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply. The method begins by generating a small number of synthetic queries using an expensive LLM. After that, a much less expensive one is used to create large numbers of synthetic queries, which are used to fine-tune a family of reranker models. These rerankers are then distilled into a single efficient retriever for use in the target domain. We show that this technique boosts zero-shot accuracy in long-tail domains and achieves substantially lower latency than standard reranking methods.","{'model': 'tldr@v2.0.0', 'text': 'This work develops and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply and achieves substantially lower latency than standard reranking methods.'}",https://arxiv.org/pdf/2303.00807
-Probing Power by Prompting: Harnessing Pre-trained Language Models for Power Connotation Framing,Shima Khanehzar,"When describing actions, subtle changes in word choice can evoke very different associations with the involved entities. For instance, a company ‘{{it employing} workers’ evokes a more positive connotation than the one ‘{{it exploiting}’ them. This concept is called {{it connotation}. This paper investigates whether pre-trained language models (PLMs) encode such subtle connotative information about {{it power differentials} between involved entities. We design a probing framework for power connotation, building on~{citet{sap-etal-2017-connotation}’s operationalization of {{it connotation frames}. We show that zero-shot prompting of PLMs leads to above chance prediction of power connotation, however fine-tuning PLMs using our framework drastically improves their accuracy. Using our fine-tuned models, we present a case study of {{it power dynamics} in US news reporting on immigration, showing the potential of our framework as a tool for understanding subtle bias in the media.","{'model': 'tldr@v2.0.0', 'text': ""It is shown that zero-shot prompting of PLMs leads to above chance prediction of power connotation, however fine-tuning PLMs using the authors' framework drastically improves their accuracy, showing the potential of the framework as a tool for understanding subtle bias in the media.""}",https://aclanthology.org/2023.eacl-main.61.pdf
-Language Models are Causal Knowledge Extractors for Zero-shot Video Question Answering,Hung-Ting Su,"Causal Video Question Answering (CVidQA) queries not only association or temporal relations but also causal relations in a video. Existing question synthesis methods pretrained question generation (QG) systems on reading comprehension datasets with text descriptions as inputs. However, QG models only learn to ask association questions (e.g., ""what is someone doing…"") and result in inferior performance due to the poor transfer of association knowledge to CVidQA, which focuses on causal questions like ""why is someone doing …"". Observing this, we proposed to exploit causal knowledge to generate question-answer pairs, and proposed a novel framework, Causal Knowledge Extraction from Language Models (CaKE-LM), leveraging causal commonsense knowledge from language models to tackle CVidQA. To extract knowledge from LMs, CaKE-LM generates causal questions containing two events with one triggering another (e.g., ""score a goal"" triggers ""soccer player kicking ball"") by prompting LM with the action (soccer player kicking ball) to retrieve the intention (to score a goal). CaKE-LM significantly outperforms conventional methods by 4% to 6% of zero-shot CVidQA accuracy on NExT-QA and Causal-VidQA datasets. We also conduct comprehensive analyses and provide key findings for future research.","{'model': 'tldr@v2.0.0', 'text': 'Causal Knowledge Extraction from Language Models (CaKE-LM) is proposed, leveraging causal commonsense knowledge from language models to tackle CVidQA, and significantly outperforms conventional methods by 4% to 6% of zero-shot CVidZA accuracy on NExT-QA and Causal-Vid QA datasets.'}",https://arxiv.org/pdf/2304.03754
-What do Language Models know about word senses? Zero-Shot WSD with Language Models and Domain Inventories,Oscar Sainz,"Language Models are the core for almost any Natural Language Processing system nowadays. One of their particularities is their contextualized representations, a game changer feature when a disambiguation between word senses is necessary. In this paper we aim to explore to what extent language models are capable of discerning among senses at inference time. We performed this analysis by prompting commonly used Languages Models such as BERT or RoBERTa to perform the task of Word Sense Disambiguation (WSD). We leverage the relation between word senses and domains, and cast WSD as a textual entailment problem, where the different hypothesis refer to the domains of the word senses. Our results show that this approach is indeed effective, close to supervised systems.","{'model': 'tldr@v2.0.0', 'text': 'This analysis leverages the relation between word senses and domains, and cast WSD as a textual entailment problem, where the different hypothesis refer to the domains of the word senses.'}",http://arxiv.org/pdf/2302.03353
-Compresso: Structured Pruning with Collaborative Prompting Learns Compact Large Language Models,Song Guo,"Despite the remarkable success of Large Language Models (LLMs), the massive size poses significant deployment challenges, particularly on resource-constrained hardware. While existing LLM compression methods focus on quantization, pruning remains relatively unexplored due to the high cost of training-based approaches and data collection challenges. One-shot pruning methods, although cost-effective and data-free, have become dominant in LLM pruning, but lead to performance decline under the structured pruning setting. In this work, we introduce a new paradigm for structurally pruning LLMs, called Compresso. Our approach, through the collaboration of the proposed resource-efficient pruning algorithm and the LLM itself, learns optimal pruning decisions during the training process. Compresso addresses the challenges of expensive training costs and data collection by incorporating Low-Rank Adaptation (LoRA) into the $L_0$ regularization during the instruction tuning process. Then, we further augment the pruning algorithm by introducing a collaborative prompt that fosters collaboration between the LLM and the pruning algorithm, significantly boosting the overall performance. To this end, Compresso prunes LLaMA-7B to 5.4B, maintaining original performance and even surpassing LLaMA-7B in reading comprehension by 2.62%. Extensive experiments demonstrate that Compresso significantly outperforms one-shot pruning baselines across various sparsity ratios, achieving up to 2.21%, 11.43%, 7.04%, and 4.81% higher scores on the commonsense reasoning, reading comprehension, MMLU, and BBH benchmarks, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a new paradigm for structurally pruning LLMs, called Compresso, which addresses the challenges of expensive training costs and data collection by incorporating Low-Rank Adaptation into the $L_0$ regularization during the instruction tuning process and introduces a collaborative prompt that fosters collaboration between the LLM and the pruning algorithm, significantly boosting the overall performance.'}",https://arxiv.org/pdf/2310.05015
-You Can Generate It Again: Data-to-text Generation with Verification and Correction Prompting,Xuan Ren,"Despite significant advancements in existing models, generating text descriptions from structured data input, known as data-to-text generation, remains a challenging task. In this paper, we propose a novel approach that goes beyond traditional one-shot generation methods by introducing a multi-step process consisting of generation, verification, and correction stages. Our approach, VCP(Verification and Correction Prompting), begins with the model generating an initial output. We then proceed to verify the correctness of different aspects of the generated text. The observations from the verification step are converted into a specialized error-indication prompt, which instructs the model to regenerate the output while considering the identified errors. To enhance the model's correction ability, we have developed a carefully designed training procedure. This procedure enables the model to incorporate feedback from the error-indication prompt, resulting in improved output generation. Through experimental results, we demonstrate that our approach effectively reduces slot error rates while maintaining the overall quality of the generated text.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach that goes beyond traditional one-shot generation methods by introducing a multi-step process consisting of generation, verification, and correction stages, and demonstrates that this approach effectively reduces slot error rates while maintaining the overall quality of the generated text.'}",http://arxiv.org/pdf/2306.15933
-TransPrompt v2: A Transferable Prompting Framework for Cross-task Text Classification,J. Wang,"Text classification is one of the most imperative tasks in natural language processing (NLP). Recent advances with pre-trained language models (PLMs) have shown remarkable success on this task. However, the satisfying results obtained by PLMs heavily depend on the large amounts of task-specific labeled data, which may not be feasible in many application scenarios due to data access and privacy constraints. The recently-proposed prompt-based fine-tuning paradigm improves the performance of PLMs for few-shot text classification with task-specific templates. Yet, it is unclear how the prompting knowledge can be transferred across tasks, for the purpose of mutual reinforcement. We propose TransPrompt v2, a novel transferable prompting framework for few-shot learning across similar or distant text classification tasks. For learning across similar tasks, we employ a multi-task meta-knowledge acquisition (MMA) procedure to train a meta-learner that captures the cross-task transferable knowledge. For learning across distant tasks, we further inject the task type descriptions into the prompt, and capture the intra-type and inter-type prompt embeddings among multiple distant tasks. Additionally, two de-biasing techniques are further designed to make the trained meta-learner more task-agnostic and unbiased towards any tasks. After that, the meta-learner can be adapted to each specific task with better parameters initialization. Extensive experiments show that TransPrompt v2 outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets. We further show that the meta-learner can effectively improve the performance of PLMs on previously unseen tasks. In addition, TransPrompt v2 also outperforms strong fine-tuning baselines when learning with full training sets.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that TransPrompt v2 outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets, and shows that the meta-learner can effectively improve the performance of PLMs on previously unseen tasks.'}",https://arxiv.org/pdf/2308.15010
-Dynamic Strategy Chain: Dynamic Zero-Shot CoT for Long Mental Health Support Generation,Qi Chen,"Long counseling Text Generation for Mental health support (LTGM), an innovative and challenging task, aims to provide help-seekers with mental health support through a comprehensive and more acceptable response. The combination of chain-of-thought (CoT) prompting and Large Language Models (LLMs) is employed and get the SOTA performance on various NLP tasks, especially on text generation tasks. Zero-shot CoT prompting is one of the most common methods in CoT prompting. However, in the LTGM task, Zero-shot CoT prompting can not simulate a counselor or provide personalized strategies without effective mental health counseling strategy prompts. To tackle this challenge, we propose a zero-shot Dynamic Strategy Chain (DSC) prompting method. Firstly, we utilize GPT2 to learn the responses written by mental health counselors and dynamically generate mental health counseling strategies tailored to the help-seekers' needs. Secondly, the Zero-shot DSC prompting is constructed according to mental health counseling strategies and the help-seekers' post. Finally, the Zero-shot DSC prompting is employed to guide LLMs in generating more human-like responses for the help-seekers. Both automatic and manual evaluations demonstrate that Zero-shot DSC prompting can deliver more human-like responses than CoT prompting methods on LTGM tasks.",,https://arxiv.org/pdf/2308.10444
-Adapt and Decompose: Efficient Generalization of Text-to-SQL via Domain Adapted Least-To-Most Prompting,Aseem Arora,"Cross-domain and cross-compositional generalization of Text-to-SQL semantic parsing is a challenging task. Existing Large Language Model (LLM) based solutions rely on inference-time retrieval of few-shot exemplars from the training set to synthesize a run-time prompt for each Natural Language (NL) test query. In contrast, we devise an algorithm which performs offline sampling of a minimal set-of few-shots from the training data, with complete coverage of SQL clauses, operators and functions, and maximal domain coverage within the allowed token length. This allows for synthesis of a fixed Generic Prompt (GP), with a diverse set-of exemplars common across NL test queries, avoiding expensive test time exemplar retrieval. We further auto-adapt the GP to the target database domain (DA-GP), to better handle cross-domain generalization; followed by a decomposed Least-To-Most-Prompting (LTMP-DA-GP) to handle cross-compositional generalization. The synthesis of LTMP-DA-GP is an offline task, to be performed one-time per new database with minimal human intervention. Our approach demonstrates superior performance on the KaggleDBQA dataset, designed to evaluate generalizability for the Text-to-SQL task. We further showcase consistent performance improvement of LTMP-DA-GP over GP, across LLMs and databases of KaggleDBQA, highlighting the efficacy and model agnostic benefits of our prompt based adapt and decompose approach.","{'model': 'tldr@v2.0.0', 'text': 'This work devise an algorithm which performs offline sampling of a minimal set-of few-shots from the training data, with complete coverage of SQL clauses, operators and functions, and maximal domain coverage within the allowed token length, for cross-domain and cross-compositional generalization of Text-to-SQL semantic parsing.'}",https://arxiv.org/pdf/2308.02582
-THE EFFECTIVITY OF THE APPLICATION TECHNICAL PROBING PROMPTING ON STORY PROBLEMS IS REVIEWED FROM MATHEMATICAL CREATIVE THINKING SKILLS TO HIGH SCHOOL STUDENT,Dede Suhendar,"This article present the results of research in the form of effectiveness using probing prompting techniques and experimental research methods with desain pre experimental : one-shot case study design by purposive sampling conducted on class X students in the 2018/2019 academic year on mathematical questions in the form of story problems. While the objectives of this study include exposing the effectiveness of mathematical creative thinking abilities of students using probing promting techniques, obtaining a number of descriptions of students problems in the process of creative thinking in mathematical questions form of story matter. Based on data analysis the results of the study showed that the percentage of students who have good mathematical creative thinking skills is less than 60%, so the probing prompting technique is not effective in terms of mathematical creative thinking abilities students.",,https://journal.ikipsiliwangi.ac.id/index.php/jiml/article/download/3975/1469
-Efektivitas Penerapan Model Probing-Prompting Ditinjau Dari Kemampuan Berpikir Kritis Siswa,Muhamad Ruslan Layn,"Penelitian ini bertujuan untuk mengetahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Kota Sorong tahun pelajaran 2017/2018 yang terdiri dari 8 kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik purposive sampling. Desain penelitian ini adalah one shot case study. Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 65%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa.",,http://e-journal.unipma.ac.id/index.php/JEMS/article/download/2013/1477
-ANALISIS KEMAMPUAN BERPIKIR DEDUKTIF AKSIOMATIK MAHASISWA MENGGUNAKAN TEKNIK PROBING-PROMPTING PADA MATA KULIAH ANALISIS REAL LANJUT,Karunia Eka Lestari,"Penelitian ini menggunakan pendekatan deskriptif kuantitatif dengan metode eksperimen. Desain eksperimen yang digunakan yaitu pre-experimental design tipe one-shot case study . Sampel dalam penelitian ini adalah seluruh mahasiswa semester VII program studi pendidikan matematika FKIP UNSIKA tahun ajaran 2014-2015 yang mengontrak mata kuliah analisis real lanjut sebanyak 38 mahasiswa. Berdasarkan hasil analisis data instrumen tes diperoleh kesimpulan bahwa Pada taraf kepercayaan 95% tidak cukup bukti untuk menyatakan bahwa kemampuan berpikir deduktif aksiomatik mahasiswa menggunakan teknik probing-prompting pada mata kuliah analisis real lanjut secara signifikan lebih besar dari 75% dari kriteria ideal yang ditetapkan. Selanjutnya hasil analisis data instrumen non tes menunjukkan bahwa kesulitan yang dialami oleh mahasiswa dalam berpikir deduktif aksiomatik menggunakan teknik probing-prompting pada mata kuliah analisis real lanjut, yaitu: (1) kesulitan dalam membaca simbol-simbol matematis; (2) kesulitan dalam memahami definisi, aksioma, postulat, dan teorema (3) kesulitan memulai dan melakukan proses pembuktian; (4) kesulitan memilih dan menggunakan definisi, aksioma, postulat, atau teorema untuk melakukan pembuktian; dan (5) kesulitan menerapkan definisi, aksioma, postulat dan teorema pada permasalahan dalam kehidupan sehari-hari. Faktor yang menyebabkan mahasiswa mengalami kesulitan dalam berpikir deduktif aksiomatik pada mata kuliah analisis real lanjut adalah: (1) faktor teknis; (2) faktor habit of mind, (3) faktor penguasaan materi dan (4) faktor intensitas latihan. Kata kunci : Kemampuan Berpikir Deduktif Aksiomatik, Teknik Probing-Prompting",,
-Leveraging Large Language Models for Multiple Choice Question Answering,Joshua Robinson,"While large language models (LLMs) like GPT-3 have achieved impressive results on multiple choice question answering (MCQA) tasks in the zero, one, and few-shot settings, they generally lag behind the MCQA state of the art (SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks. An LLM is conditioned on a question (without the associated answer options) and its chosen option is the one assigned the highest probability after normalization (for length, etc.). A more natural prompting approach is to present the question and answer options to the LLM jointly and have it output the symbol (e.g.,""A"") associated with its chosen answer option. This approach allows the model to explicitly compare answer options, reduces computational costs, and mitigates the effects of tokenization scheme and answer option representations on answer selection. For the natural approach to be effective, the LLM it is used with must be able to associate answer options with the symbols that represent them. The LLM needs what we term multiple choice symbol binding (MCSB) ability. This ability varies greatly by model. We show that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated.'}",http://arxiv.org/pdf/2210.12353
-FOLIO: Natural Language Reasoning with First-Order Logic,Simeng Han,"We present FOLIO, a human-annotated, open-domain, and logically complex and diverse dataset for reasoning in natural language (NL), equipped with first order logic (FOL) annotations. FOLIO consists of 1,435 examples (unique conclusions), each paired with one of 487 sets of premises which serve as rules to be used to deductively reason for the validity of each conclusion. The logical correctness of premises and conclusions is ensured by their parallel FOL annotations, which are automatically verified by our FOL inference engine. In addition to the main NL reasoning task, NL-FOL pairs in FOLIO automatically constitute a new NL-FOL translation dataset using FOL as the logical form. Our experiments on FOLIO systematically evaluate the FOL reasoning ability of supervised fine-tuning on medium-sized language models (BERT, RoBERTa) and few-shot prompting on large language models (GPT-NeoX, OPT, GPT-3, Codex). For NL-FOL translation, we experiment with GPT-3 and Codex. Our results show that one of the most capable Large Language Model (LLM) publicly available, GPT-3 davinci, achieves only slightly better than random results with few-shot prompting on a subset of FOLIO, and the model is especially bad at predicting the correct truth values for False and Unknown conclusions. Our dataset and code are available at https://github.com/Yale-LILY/FOLIO.","{'model': 'tldr@v2.0.0', 'text': 'The results show that one of the most capable Large Language Model (LLM) publicly available, GPT-3 davinci, achieves only slightly better than random results with few-shot prompting on a subset of FOLIO, and the model is especially bad at predicting the correct truth values for False and Unknown conclusions.'}",http://arxiv.org/pdf/2209.00840
-Data Augmentation for Intent Classification with Off-the-shelf Large Language Models,Gaurav Sahu,"Data augmentation is a widely employed technique to alleviate the problem of data scarcity. In this work, we propose a prompting-based approach to generate labelled training data for intent classification with off-the-shelf language models (LMs) such as GPT-3. An advantage of this method is that no task-specific LM-fine-tuning for data generation is required; hence the method requires no hyper parameter tuning and is applicable even when the available training data is very scarce. We evaluate the proposed method in a few-shot setting on four diverse intent classification tasks. We find that GPT-generated data significantly boosts the performance of intent classifiers when intents in consideration are sufficiently distinct from each other. In tasks with semantically close intents, we observe that the generated data is less helpful. Our analysis shows that this is because GPT often generates utterances that belong to a closely-related intent instead of the desired one. We present preliminary evidence that a prompting-based GPT classifier could be helpful in filtering the generated data to enhance its quality.","{'model': 'tldr@v2.0.0', 'text': 'It is found that GPT-generated data significantly boosts the performance of intent classifiers when intents in consideration are sufficiently distinct from each other, and preliminary evidence that a prompting-based GPT classifier could be helpful in filtering the generated data to enhance its quality is presented.'}",http://arxiv.org/pdf/2204.01959
-EFEKTIVITAS PENERAPAN MODEL PROBING-PROMPTING DITINJAU DARI KEMAMPUAN BERPIKIR KRITIS SISWA,L. Mulyadi,"This quasi-experimental research aimed to know the effectiveness of probing-prompting model viewed by students critical thinking ability . The population of this research was all students of grade VIII of SMP N egeri 9 Bandar lampung in academic year of 201 4 /201 5 that consist of 7 classes . The sample of this research was students of VII I-C which was taken by purposive sampling technique . This research design was one shot case study . Instrument of this research was essay test of critical thinking . Based on the analysis of data, the percentage of students who ha ve good critical thinking ability was less than 60%. Thus, probing-prompting model was less effective viewed by critical thinking ability . Penelitian kuasi eksperimen ini bertujuan untuk mengetahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Bandarlampung tahun pelajaran 2014/2015 yang terdiri dari tujuh kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik p urposive sampling . Desain penelitian ini adalah one shot case study . Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 60%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa. Kata kunci: berpikir kritis, efektivitas, probing-prompting",,
-Multi-Party Chat: Conversational Agents in Group Settings with Humans and Models,Jimmy Wei,"Current dialogue research primarily studies pairwise (two-party) conversations, and does not address the everyday setting where more than two speakers converse together. In this work, we both collect and evaluate multi-party conversations to study this more general case. We use the LIGHT environment to construct grounded conversations, where each participant has an assigned character to role-play. We thus evaluate the ability of language models to act as one or more characters in such conversations. Models require two skills that pairwise-trained models appear to lack: (1) being able to decide when to talk; (2) producing coherent utterances grounded on multiple characters. We compare models trained on our new dataset to existing pairwise-trained dialogue models, as well as large language models with few-shot prompting. We find that our new dataset, MultiLIGHT, which we will publicly release, can help bring significant improvements in the group setting.","{'model': 'tldr@v2.0.0', 'text': 'This work collects and evaluates multi-party conversations and uses the LIGHT environment to construct grounded conversations, where each participant has an assigned character to role-play, to evaluate the ability of language models to act as one or more characters in such conversations.'}",http://arxiv.org/pdf/2304.13835
-SQL-PaLM: Improved Large Language Model Adaptation for Text-to-SQL,Ruoxi Sun,"One impressive emergent capability of large language models (LLMs) is generation of code, including Structured Query Language (SQL) for databases. For the task of converting natural language text to SQL queries, Text-to-SQL, adaptation of LLMs is of paramount importance, both in in-context learning and fine-tuning settings, depending on the amount of adaptation data used. In this paper, we propose an LLM-based Text-to-SQL model SQL-PaLM, leveraging on PaLM-2, that pushes the state-of-the-art in both settings. Few-shot SQL-PaLM is based on an execution-based self-consistency prompting approach designed for Text-to-SQL, and achieves 77.3% in test-suite accuracy on Spider, which to our best knowledge is the first to outperform previous state-of-the-art with fine-tuning by a significant margin, 4%. Furthermore, we demonstrate that the fine-tuned SQL-PALM outperforms it further by another 1%. Towards applying SQL-PaLM to real-world scenarios we further evaluate its robustness on other challenging variants of Spider and demonstrate the superior generalization capability of SQL-PaLM. In addition, via extensive case studies, we demonstrate the impressive intelligent capabilities and various success enablers of LLM-based Text-to-SQL.","{'model': 'tldr@v2.0.0', 'text': 'An LLM-based Text-to-SQL model SQL-PaLM is proposed, leveraging on PaLM-2, that pushes the state-of-the-art in both in-context learning and fine-tuning settings, and is believed to be the first to outperform previous state- of- the-art with fine- Tuning by a significant margin.'}",http://arxiv.org/pdf/2306.00739
-EFEKTIVITAS PENERAPAN MODEL PROBING-PROMPTING DITINJAU DARI KEMAMPUAN BERPIKIR KRITIS SISWA (Studi pada Siswa Kelas VIII SMP Negeri 9 Bandarlampung Tahun Pelajaran 2014/2015),M. Mulyadi,"Penelitian kuasi eksperimen ini bertujuan untuk menge¬tahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Bandarlampung Tahun Pelajaran 2014/2015 yang terdiri dari tujuh kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik purposive sampling. Desain penelitian ini adalah one shot case study. Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 60%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa. Kata Kunci : berpikir kritis, efektivitas, probing-prompting",,
-MetaPrompting: Learning to Learn Better Prompts,Yutai Hou,"Prompting method is regarded as one of the crucial progress for few-shot nature language processing. Recent research on prompting moves from discrete tokens based “hard prompts” to continuous “soft prompts”, which employ learnable vectors as pseudo prompt tokens and achieve better performance. Though showing promising prospects, these soft-prompting methods are observed to rely heavily on good initialization to take effect. Unfortunately, obtaining a perfect initialization for soft prompts requires understanding of inner language models working and elaborate design, which is no easy task and has to restart from scratch for each new task. To remedy this, we propose a generalized soft prompting method called MetaPrompting, which adopts the well-recognized model-agnostic meta-learning algorithm to automatically find better prompt initialization that facilitates fast adaptation to new prompting tasks. Extensive experiments show MetaPrompting tackles soft prompt initialization problem and brings significant improvement on three different datasets (over 6 points improvement in accuracy for 1-shot setting), achieving new state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A generalized soft prompting method called MetaPrompting is proposed, which adopts the well-recognized model-agnostic meta-learning algorithm to automatically find better prompt initialization that facilitates fast adaptation to new prompting tasks.'}",http://arxiv.org/pdf/2209.11486
-Unraveling ChatGPT: A Critical Analysis of AI-Generated Goal-Oriented Dialogues and Annotations,Tiziano Labruna,"Large pre-trained language models have exhibited unprecedented capabilities in producing high-quality text via prompting techniques. This fact introduces new possibilities for data collection and annotation, particularly in situations where such data is scarce, complex to gather, expensive, or even sensitive. In this paper, we explore the potential of these models to generate and annotate goal-oriented dialogues, and conduct an in-depth analysis to evaluate their quality. Our experiments employ ChatGPT, and encompass three categories of goal-oriented dialogues (task-oriented, collaborative, and explanatory), two generation modes (interactive and one-shot), and two languages (English and Italian). Based on extensive human-based evaluations, we demonstrate that the quality of generated dialogues and annotations is on par with those generated by humans.","{'model': 'tldr@v2.0.0', 'text': 'The quality of generated dialogues and annotations is on par with those generated by humans, based on extensive human-based evaluations.'}",http://arxiv.org/pdf/2305.14556
-Shot for Shot Remakes,D. Varndell,,,
-Improving Patient Pre-screening for Clinical Trials: Assisting Physicians with Large Language Models,D. Hamer,"Physicians considering clinical trials for their patients are met with the laborious process of checking many text based eligibility criteria. Large Language Models (LLMs) have shown to perform well for clinical information extraction and clinical reasoning, including medical tests, but not yet in real-world scenarios. This paper investigates the use of InstructGPT to assist physicians in determining eligibility for clinical trials based on a patient's summarised medical profile. Using a prompting strategy combining one-shot, selection-inference and chain-of-thought techniques, we investigate the performance of LLMs on 10 synthetically created patient profiles. Performance is evaluated at four levels: ability to identify screenable eligibility criteria from a trial given a medical profile; ability to classify for each individual criterion whether the patient qualifies; the overall classification whether a patient is eligible for a clinical trial and the percentage of criteria to be screened by physician. We evaluated against 146 clinical trials and a total of 4,135 eligibility criteria. The LLM was able to correctly identify the screenability of 72% (2,994/4,135) of the criteria. Additionally, 72% (341/471) of the screenable criteria were evaluated correctly. The resulting trial level classification as eligible or ineligible resulted in a recall of 0.5. By leveraging LLMs with a physician-in-the-loop, a recall of 1.0 and precision of 0.71 on clinical trial level can be achieved while reducing the amount of criteria to be checked by an estimated 90%. LLMs can be used to assist physicians with pre-screening of patients for clinical trials. By forcing instruction-tuned LLMs to produce chain-of-thought responses, the reasoning can be made transparent to and the decision process becomes amenable by physicians, thereby making such a system feasible for use in real-world scenarios.","{'model': 'tldr@v2.0.0', 'text': 'By forcing instruction-tuned LLMs to produce chain-of-thought responses, the reasoning can be made transparent to and the decision process becomes amenable by physicians, thereby making such a system feasible for use in real-world scenarios.'}",http://arxiv.org/pdf/2304.07396
-PENERAPAN TEKNIK PEMBELAJARAN PROBING PROMPTING UNTUK MENGETAHUI KEMAMPUAN PENALARAN MATEMATIKA SISWA KELAS 7 G DI SMPN 1 REJOSO,Yayuk Kurniasari,"Matematika dipahami melalui penalaran, agar kemampuan penalaran matematis dapat berkembang secara optimal, siswa harus memiliki kesempatan yang terbuka untuk berpikir. Salah satu teknik mengajar yang dinilai akomodatif dapat meningkatkan aktivitas berpikir siswa adalah teknik Probing Prompting, yaitu teknik pembelajaran dengan cara guru menyajikan serangkaian pertanyaan yang sifatnya menuntun dan menggali sehingga terjadi proses berpikir yang mengaitkan pengetahuan siswa dan pengalamannya dengan pengetahuan baru yang sedang dipelajari.Penelitian ini bertujuan untuk mendeskripsikan kemampuan penalaran matematika dan respon siswa pada pembelajaran dengan teknik Probing Prompting. Subyek penelitian ini adalah siswa SMP Negeri 1 Rejoso Kab. Nganjuk kelas 7 G sebanyak 36 siswa. Rancangan penelitian yang digunakan adalah “One Shot Case Study”.Berdasarkan hasil analisis data yang diperoleh dapat disimpulkan bahwa siswa mampu menyajikan pernyataan matematika secara lisan, tertulis, gambar, dan diagram dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 72 % siswa berada pada level 2 ke atas. Siswa mampu mengajukan dugaan dengan benar meskipun ada beberapa penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 91 % siswa berada pada level 2 ke atas. Siswa mampu melakukan manipulasi matematika dengan benar meskipun penalarannya kurang lengkap, hal ini diketahui sebanyak 78 % siswa berada pada level 2 ke atas. Siswa mampu menyusun bukti, memberikan alasan atau bukti terhadap kebenaran solusi dengan benar dan menggunakan penalaran dengan baik, hal ini diketahui sebanyak 52 % siswa berada pada level 3 ke atas. Siswa mampu menarik kesimpulan dari pernyataan dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 86 % berada pada level 2 ke atas. Siswa mampu memeriksa kesahihan suatu argumen dengan benar dan penalaran baik, hal ini diketahui bahwa sebanyak 53 % berada pada level 3 ke atas. Siswa mampu menemukan pola atau sifat gejala matematis untuk membuat generalisasi dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini dieketahui bahwa sebanyak 87 % berada pada level 2 ke atas. Hasil angket respon siswa menunjukkan bahwa respon siswa adalah positif dengan rata-rata nilai respon siswa sebesar 3,17.",,
-SINC: Spatial Composition of 3D Human Motions for Simultaneous Action Generation,Nikos Athanasiou,"Our goal is to synthesize 3D human motions given textual inputs describing simultaneous actions, for example 'waving hand' while 'walking' at the same time. We refer to generating such simultaneous movements as performing 'spatial compositions'. In contrast to temporal compositions that seek to transition from one action to another, spatial compositing requires understanding which body parts are involved in which action, to be able to move them simultaneously. Motivated by the observation that the correspondence between actions and body parts is encoded in powerful language models, we extract this knowledge by prompting GPT-3 with text such as""what are the body parts involved in the action?"", while also providing the parts list and few-shot examples. Given this action-part mapping, we combine body parts from two motions together and establish the first automated method to spatially compose two actions. However, training data with compositional actions is always limited by the combinatorics. Hence, we further create synthetic data with this approach, and use it to train a new state-of-the-art text-to-motion generation model, called SINC (""SImultaneous actioN Compositions for 3D human motions""). In our experiments, that training with such GPT-guided synthetic data improves spatial composition generation over baselines. Our code is publicly available at https://sinc.is.tue.mpg.de/.","{'model': 'tldr@v2.0.0', 'text': 'This work combines body parts from two motions together and establishes the first automated method to spatially compose two actions, and uses it to train a new state-of-the-art text-to-motion generation model, called SINC (""SImultaneous actioN Compositions for 3D human motions"").'}",https://arxiv.org/pdf/2304.10417
-Hybrid-Regressive Neural Machine Translation,Qiang Wang,"In this work, we empirically confirm that non-autoregressive translation with an iterative refinement mechanism (IR-NAT) suffers from poor acceleration robustness because it is more sensitive to decoding batch size and computing device setting than autoregressive translation (AT). Inspired by it, we attempt to investigate how to combine the strengths of autoregressive and non-autoregressive translation paradigms better. To this end, we demonstrate through synthetic experiments that prompting a small number of AT's predictions can promote one-shot non-autoregressive translation to achieve the equivalent performance of IR-NAT. Following this line, we propose a new two-stage translation prototype called hybrid-regressive translation (HRT). Specifically, HRT first generates discontinuous sequences via autoregression (e.g., make a prediction every k tokens, k>1) and then fills in all previously skipped tokens at once in a non-autoregressive manner. We also propose a bag of techniques to effectively and efficiently train HRT without adding any model parameters. HRT achieves the state-of-the-art BLEU score of 28.49 on the WMT En-De task and is at least 1.5x faster than AT, regardless of batch size and device. In addition, another bonus of HRT is that it successfully inherits the good characteristics of AT in the deep-encoder-shallow-decoder architecture. Concretely, compared to the vanilla HRT with a 6-layer encoder and 6-layer decoder, the inference speed of HRT with a 12-layer encoder and 1-layer decoder is further doubled on both GPU and CPU without BLEU loss.","{'model': 'tldr@v2.0.0', 'text': 'This work empirically confirms that non-autoregressive translation with an iterative refinement mechanism (IR-NAT) suffers from poor acceleration robustness and proposes a new two-stage translation prototype called hybrid-regressive translation (HRT).'}",http://arxiv.org/pdf/2210.10416
-The Potential and Pitfalls of using a Large Language Model such as ChatGPT or GPT-4 as a Clinical Assistant,Jingqing Zhang,"Recent studies have demonstrated promising performance of ChatGPT and GPT-4 on several medical domain tasks. However, none have assessed its performance using a large-scale real-world electronic health record database, nor have evaluated its utility in providing clinical diagnostic assistance for patients across a full range of disease presentation. We performed two analyses using ChatGPT and GPT-4, one to identify patients with specific medical diagnoses using a real-world large electronic health record database and the other, in providing diagnostic assistance to healthcare workers in the prospective evaluation of hypothetical patients. Our results show that GPT-4 across disease classification tasks with chain of thought and few-shot prompting can achieve performance as high as 96% F1 scores. For patient assessment, GPT-4 can accurately diagnose three out of four times. However, there were mentions of factually incorrect statements, overlooking crucial medical findings, recommendations for unnecessary investigations and overtreatment. These issues coupled with privacy concerns, make these models currently inadequate for real world clinical use. However, limited data and time needed for prompt engineering in comparison to configuration of conventional machine learning workflows highlight their potential for scalability across healthcare applications.","{'model': 'tldr@v2.0.0', 'text': 'This work performed two analyses using ChatGPT and GPT-4, one to identify patients with specific medical diagnoses using a real-world large electronic health record database and the other, in providing diagnostic assistance to healthcare workers in the prospective evaluation of hypothetical patients.'}",https://arxiv.org/pdf/2307.08152
-Little Giants: Exploring the Potential of Small LLMs as Evaluation Metrics in Summarization in the Eval4NLP 2023 Shared Task,Neema Kotonya,"This paper describes and analyzes our participation in the 2023 Eval4NLP shared task, which focuses on assessing the effectiveness of prompt-based techniques to empower Large Language Models to handle the task of quality estimation, particularly in the context of evaluating machine translations and summaries. We conducted systematic experiments with various prompting techniques, including standard prompting, prompts informed by annotator instructions, and innovative chain-of-thought prompting. In addition, we integrated these approaches with zero-shot and one-shot learning methods to maximize the efficacy of our evaluation procedures. Our work reveals that combining these approaches using a""small"", open source model (orca_mini_v3_7B) yields competitive results.","{'model': 'tldr@v2.0.0', 'text': None}",
-Hitachi at SemEval-2023 Task 4: Exploring Various Task Formulations Reveals the Importance of Description Texts on Human Values,Masaya Tsunokake,"This paper describes our participation in SemEval-2023 Task 4, ValueEval: Identification of Human Values behind Arguments. The aim of this task is to identify whether or not an input text supports each of the 20 pre-defined human values. Previous work on human value detection has shown the effectiveness of a sequence classification approach using BERT. However, little is known about what type of task formulation is suitable for the task. To this end, this paper explores various task formulations, including sequence classification, question answering, and question answering with chain-of-thought prompting and evaluates their performances on the shared task dataset. Experiments show that a zero-shot approach is not as effective as other methods, and there is no one approach that is optimal in every scenario. Our analysis also reveals that utilizing the descriptions of human values can help to improve performance.","{'model': 'tldr@v2.0.0', 'text': 'Various task formulations are explored, including sequence classification, question answering, and question answering with chain-of-thought prompting and evaluates their performances on the shared task dataset and reveals that utilizing the descriptions of human values can help to improve performance.'}",https://aclanthology.org/2023.semeval-1.240.pdf
-Learning to Paraphrase Sentences to Different Complexity Levels,Alison Chi,"While sentence simplification is an active research topic in NLP, its adjacent tasks of sentence complexification and same-level paraphrasing are not. To train models on all three tasks, we present two new unsupervised datasets. We compare these datasets, one labeled by a weak classifier and the other by a rule-based approach, with a single supervised dataset. Using these three datasets for training, we perform extensive experiments on both multitasking and prompting strategies. Compared to other systems trained on unsupervised parallel data, models trained on our weak classifier labeled dataset achieve state-of-the-art performance on the ASSET simplification benchmark. Our models also outperform previous work on sentence level targeting. Finally, we establish how a handful of Large Language Models perform on these tasks under a zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': ""Compared to other systems trained on unsupervised parallel data, models trained on the authors' weak classifier labeled dataset achieve state-of-the-art performance on the ASSET simplification benchmark and outperform previous work on sentence level targeting.""}",https://arxiv.org/pdf/2308.02226
-"Investigating the Perception of the Future in GPT-3, -3.5 and GPT-4",Diana Kozachek,"This study explores the potential of GPT-3, GPT-3.5, and GPT-4, in generating human-like future scenarios to investigate each model's ability to perceive time. The methodology combines a coding-based experiment and an expert survey. The investigation involves fine- and prompt-tuning GPT-3, prompt-tuning GPT-3.5, and few-shot prompting GPT-4 with human-made future scenarios. The models and output are quantitatively and qualitatively analyzed. The survey invited practitioners from fields of foresight and futurology, AI, and NLP to assess whether differences in output can be identified. This study's findings suggest that GPT-3 and GPT-4 generated scenarios are difficult to distinguish from human-made ones, while GPT-3.5 performed more poorly. Yet none of the models can differentiate time horizons and their respective effects on the future from each other. And while no one knows the shape of things to come, this lack of understanding of a core concept of life invites future investigations.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3 and GPT-4 generated scenarios are difficult to distinguish from human-made ones, while G PT-3.5 performed more poorly, and none of the models can differentiate time horizons and their respective effects on the future from each other.'}",
-OLaLa: Ontology Matching with Large Language Models,S. Hertling,"Ontology (and more generally: Knowledge Graph) Matching is a challenging task where information in natural language is one of the most important signals to process. With the rise of Large Language Models, it is possible to incorporate this knowledge in a better way into the matching pipeline. A number of decisions still need to be taken, e.g., how to generate a prompt that is useful to the model, how information in the KG can be formulated in prompts, which Large Language Model to choose, how to provide existing correspondences to the model, how to generate candidates, etc. In this paper, we present a prototype that explores these questions by applying zero-shot and few-shot prompting with multiple open Large Language Models to different tasks of the Ontology Alignment Evaluation Initiative (OAEI). We show that with only a handful of examples and a well-designed prompt, it is possible to achieve results that are en par with supervised matching systems which use a much larger portion of the ground truth.","{'model': 'tldr@v2.0.0', 'text': 'This paper applies zero-shot and few-shot prompting with multiple open Large Language Models to different tasks of the Ontology Alignment Evaluation Initiative (OAEI) to achieve results that are en par with supervised matching systems which use a much larger portion of the ground truth.'}",
-Tree of Clarifications: Answering Ambiguous Questions with Retrieval-Augmented Large Language Models,Gangwoo Kim,"Questions in open-domain question answering are often ambiguous, allowing multiple interpretations. One approach to handling them is to identify all possible interpretations of the ambiguous question (AQ) and to generate a long-form answer addressing them all, as suggested by Stelmakh et al., (2022). While it provides a comprehensive response without bothering the user for clarification, considering multiple dimensions of ambiguity and gathering corresponding knowledge remains a challenge. To cope with the challenge, we propose a novel framework, Tree of Clarifications (ToC): It recursively constructs a tree of disambiguations for the AQ -- via few-shot prompting leveraging external knowledge -- and uses it to generate a long-form answer. ToC outperforms existing baselines on ASQA in a few-shot setup across the metrics, while surpassing fully-supervised baselines trained on the whole training set in terms of Disambig-F1 and Disambig-ROUGE. Code is available at https://github.com/gankim/tree-of-clarifications.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, Tree of Clarifications (ToC), recursively constructs a tree of disambiguations for the AQ -- via few-shot prompting leveraging external knowledge -- and uses it to generate a long-form answer.'}",
-Evaluation of Prompts to Simplify Cardiovascular Disease Information Using a Large Language Model,Mbbs Vishala Mishra,"AI chatbots powered by large language models (LLMs) are emerging as an important source of public-facing medical information. Generative models hold promise for producing tailored guidance at scale, which could advance health literacy and mitigate well-known disparities in the accessibility of health-protective information. In this study, we highlight an important limitation of basic approaches to AI-powered text simplification: when given a zero-shot or one-shot simplification prompt, GPT-4 often responds by omitting critical details. To address this limitation, we developed a new prompting strategy, which we term rubric prompting. Rubric prompts involve a combination of a zero-shot simplification prompt with brief reminders about important topics to address. Using rubric prompts, we generate recommendations about cardiovascular disease prevention that are more complete, more readable, and have lower syntactic complexity than baseline responses produced without prompt engineering. This analysis provides a blueprint for rigorous evaluation of AI model outputs in medicine.","{'model': 'tldr@v2.0.0', 'text': 'Using rubric prompts, a new prompting strategy is developed that generates recommendations about cardiovascular disease prevention that are more complete, more readable, and have lower syntactic complexity than baseline responses produced without prompt engineering.'}",
-Category-Specific Prompts for Animal Action Recognition with Pretrained Vision-Language Models,Yinuo Jing,"Animal action recognition has a wide range of applications. However, the field largely remains unexplored due to the greater challenges compared to human action recognition, such as lack of annotated training data, large intra-class variation, and interference of cluttered background. Most of the existing methods directly apply human action recognition techniques, which essentially require a large amount of annotated data. In recent years, contrastive vision-language pretraining has demonstrated strong zero-shot generalization ability and has been used for human action recognition. Inspired by the success, we develop a highly performant action recognition framework based on the CLIP model. Our model addresses the above challenges via a novel category-specific prompting module to generate adaptive prompts for both text and video based on the animal category detected in input videos. On one hand, it can generate more precise and customized textual descriptions for each action and animal category pair, being helpful in the alignment of textual and visual space. On the other hand, it allows the model to focus on video features of the target animal in the video and reduce the interference of video background noise. Experimental results demonstrate that our method outperforms five previous action recognition methods on the Animal Kingdom dataset and has shown best generalization ability on unseen animals.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the method outperforms five previous action recognition methods on the Animal Kingdom dataset and has shown best generalization ability on unseen animals.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612551
-"Slow Academic Travel: An Antidote to ""Fly Over"" Scholarship in the Age of Climate Crisis",Meredith A. Conti,"All three animals faced away from me, a row of hairy posteriors blocking the rest of their hulking bodies from view. My iPhone poised in camera mode, I walked slowly to the left, keeping my eyes on the herd grazing eighty yards in the distance. They were in no hurry, and neither was I. Just as I had lined up the perfect shot—the lowered heads of the bison, tiny black eyes shining and mouths masticating in perpetual motion, were now visible—a frigid blast of wind whipped across the plain, rustling grasses, flapping my jacket, and prompting the bison to rotate in one choreographed movement to the right. Unlike me, they wanted to face into the wind. Laughing under my breath, I moved farther to the left until the bison were in profile and began snapping. I arrived back at my car the proud owner of a new nature photo series: “Unruffled Buffalo, Badlands National Park.” Individual behavior change is a necessary condition for making a positive societal transition to a resource-constrained future. Because we live on a planet with finite resources, behavior change interventions need to move beyond creating “green consumers” and instead foster and support “green citizens.” . . . There is an urgent need to promote agency and self-sufficiency with regards to individual conservation behaviors, while simultaneously promoting mindfulness of collective impacts on the planet. (73)",,
-Method for prompting passenger density information in all carriages of subway to target station platform,叶尔肯.拜山,"The invention discloses a method for prompting passenger density information in all carriages of a subway to a target station platform. In the method, more than one camera or camera head, an image analysis module, a transmitting module and a display module are included, wherein the camera or the camera head is arranged in the carriages of the subway to acquire images of passengers in the carriages of the subway; the image analysis module is used for analyzing and calculating valid passenger density data shot by the camera or the camera head and outputting the valid passenger density data; the transmission module is used for outputting the acquired valid passenger density data; and the display module displays the acquired valid passenger density data. According to the method, by monitoring for the free space in the carriages of the subway, passengers waiting for the subway on the station platform know the passenger density in all the carriages of the subway which is about to enter the station and then make queues reasonably, so that the passenger transport efficiency of the subway is improved.","{'model': 'tldr@v2.0.0', 'text': 'By monitoring for the free space in the carriages of the subway, passengers waiting for the subway on the station platform know the passenger density in all the car carriage which is about to enter the station and then make queues reasonably, so that the passenger transport efficiency of the Subway is improved.'}",
-How Large Language Models Perform on the United States Medical Licensing Examination: A Systematic Review,MD Dana Brin,"Purpose: The United States Medical Licensing Examination (USMLE) assesses physicians' competency and passing is a requirement to practice medicine in the U.S. With the emergence of large language models (LLMs) like ChatGPT and GPT-4, understanding their performance on these exams illuminates their potential in medical education and healthcare. Methods: A literature search following the 2020 PRISMA guidelines was conducted, focusing on studies using official USMLE questions and publicly available LLMs. Results: Three relevant studies were found, with GPT-4 showcasing the highest accuracy rates of 80-90% on the USMLE. Open-ended prompts typically outperformed multiple-choice ones, with 5-shot prompting slightly edging out zero-shot. Conclusion: LLMs, especially GPT-4, display proficiency in tackling USMLE-standard questions. While the USMLE is a structured evaluation tool, it may not fully capture the expansive capabilities and limitations of LLMs in medical scenarios. As AI integrates further into healthcare, ongoing assessments against trusted benchmarks are essential.","{'model': 'tldr@v2.0.0', 'text': 'While the USMLE is a structured evaluation tool, it may not fully capture the expansive capabilities and limitations of LLMs in medical scenarios, as AI integrates further into healthcare, ongoing assessments against trusted benchmarks are essential.'}",https://www.medrxiv.org/content/medrxiv/early/2023/09/07/2023.09.03.23294842.full.pdf
-Resource Constrained Dialog Policy Learning Via Differentiable Inductive Logic Programming,Zhenpeng Zhou,"Motivated by the needs of resource constrained dialog policy learning, we introduce dialog policy via differentiable inductive logic (DILOG). We explore the tasks of one-shot learning and zero-shot domain transfer with DILOG on SimDial and MultiWoZ. Using a single representative dialog from the restaurant domain, we train DILOG on the SimDial dataset and obtain 99+% in-domain test accuracy. We also show that the trained DILOG zero-shot transfers to all other domains with 99+% accuracy, proving the suitability of DILOG to slot-filling dialogs. We further extend our study to the MultiWoZ dataset achieving 90+% inform and success metrics. We also observe that these metrics are not capturing some of the shortcomings of DILOG in terms of false positives, prompting us to measure an auxiliary Action F1 score. We show that DILOG is 100x more data efficient than state-of-the-art neural approaches on MultiWoZ while achieving similar performance metrics. We conclude with a discussion on the strengths and weaknesses of DILOG.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces dialog policy via differentiable inductive logic (DILOG) and shows that DILOG is 100x more data efficient than state-of-the-art neural approaches on MultiWoZ while achieving similar performance metrics.'}",https://www.aclweb.org/anthology/2020.coling-main.597.pdf
-Prestack suppression of high frequency ground roll using a 3D multiwindow KL filter: Application to a legacy Mississippi Lime survey,S. Verma,"Summary Recently the Mississippi Lime has become one of the most active resource plays. Our study area falls in-between the Fort Worth and Midland Basins. The main production comes from high porosity tripolitic chert. Our objective is to use 3D seismic data to map the areal distribution of discontinuous tripolitic facies. In the early 1990s several 3D surveys were shot in the study area to image shallower objectives. With the advent of the Mississippi Lime play, four of these surveys were merged and reprocessed using careful statics and velocity analysis. Even after prestack time migration, the target zone is contaminated with the acquisition footprint. The data are low (~15) fold and contaminated by highly aliased, high frequency, high amplitude ground roll. Given the sparsity of the survey, modern f-kx-ky filters were not able to remove ground roll prompting the development of a new ground roll suppression workflow. In workflow, we first window and low-pass filter (f<50 Hz) the data, 3D patch by 3D patch. We then apply linear moveout to approximately flatten the ground roll phases, estimate the dip about this reference moveout, and compute coherence within a 3channel by 3-shot by 20 ms window for each sample. Using a Kuwahara algorithm, we choose the most coherent window within which we apply a structure-oriented KL filter. At the end we simply modeled the ground roll from the original data. This 3D filter preserves signal amplitude and is flexible enough to model the piece wise continuous ground roll pattern common with irregular topography.",,
-Our chimp cousins get that much closer,A. Gibbons,"ARE CHIMPS MORE CLOSELY RELATED TO gorillas or to human beings? That seemingly innocent question provoked a war among evolutionary specialists that has lasted more than 20 years. Now one side has new ammunition-in the form of mitochondrial DNA studies showing that chimps have more in common with us than with gorillas. This data, coming from independent U.S. and Japanese teams, is consistent with a growing body of results based on nuclear DNA sequences, prompting one chimphuman partisan, Charles Sibley of San Francisco State University, to claim, ""We've won the war."" But don't assume a surrender treaty is about to be signed. The embattled minority who think chimps are closer to gorillas has by no means conceded defeat. ""I think the attempt to close this question is extremely premature,"" says Jonathan Marks, professor of anthropology at Yale, who questions the quality of almost all the DNA data now available-and points to a few studies that suggest a chimp-gorilla connection. The first shot in the war was fired in the early 60s, when Morris Goodman ofWayne State University undermined the conventional classification system that put humans in one family-Hominidae-and chimps and gorillas in a second family: Pongidae. Goodman's work, based on cross reactions between immunologic blood proteins, suggested humans are very closely related genetically to the two species of African apes and that all three belong in one family. Most evolutionists thought Goodman's data were shaky, and they stuck to the old system. But molecular biologists persevered. In 1984, Sibley and Jon Ahlquist, both then at Yale, published DNA hybridization data showing that not only were we close to the apes, but that chimps actually had more in common genetically with humans than with gorillas. DNA hybridization is based on seeing how much of one set of DNA fragments will ""hybridize,"" or bind to another, providing an indication ofhow many sequences they share. But it is tricky to do, tricky to interpret, and less precise than having the actual sequences. Partly as a result, the Sibley and Ahlquist results were subjected to a withering bombardment from Marks and Vincent Sarich of the University of California at Berkeley. Marks and Sarich questioned their methods of data analysis and even charged that Sibley and Ahlquist had falsified data. It wasn't long afterwards, however, that actual DNA sequences from nuclear genes began coming in that offered strong support for close ties between chimps and humans. Perhaps the best work came from Goodman and his col-",,
-Safety and arming method for Fuzes based on geomagnetic field signal,Hu Xuegong,"Based on analyzing the method of geomagnetism turns-counting and the mathematical model for testing distance, tow kinds of information are put forward as the second environment prompting of the redundant safety, one is the geomagnetism field information when the projectile exiting the muzzle, and the second is the spin signal of the projectile in the geomagnetic field. “Threshold level + time window” is used as the judgment criterion, and this can distinguish the shot environment and non-shot environment. The method of geomagnetism turns-counting for testing distance is used to realize the distance design of the muzzle safety, it is easy to solve the problem of insufficient muzzle safety distance, and it can distinguish the normal shot environment and improper shot environment. The geomagnetism curve is tested by dynamic shooting, and it is testified that the schemes is feasible.","{'model': 'tldr@v2.0.0', 'text': 'The method of geomagnetism turns-counting for testing distance is used to realize the distance design of the muzzle safety, it is easy to solve the problem of insufficient muzzle safety distance, and it can distinguish the normal shot environment and improper shot environment.'}",
-Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation,Jay Zhangjie Wu,"To replicate the success of text-to-image (T2I) generation, recent works employ large-scale video datasets to train a text-to-video (T2V) generator. Despite their promising results, such paradigm is computationally expensive. In this work, we propose a new T2V generation setting$\unicode{x2014}$One-Shot Video Tuning, where only one text-video pair is presented. Our model is built on state-of-the-art T2I diffusion models pre-trained on massive image data. We make two key observations: 1) T2I models can generate still images that represent verb terms; 2) extending T2I models to generate multiple images concurrently exhibits surprisingly good content consistency. To further learn continuous motion, we introduce Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy. At inference, we employ DDIM inversion to provide structure guidance for sampling. Extensive qualitative and numerical experiments demonstrate the remarkable ability of our method across various applications.","{'model': 'tldr@v2.0.0', 'text': 'A new T2V generation setting, where only one text-video pair is presented, and Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy, is introduced.'}",http://arxiv.org/pdf/2212.11565
-Siamese Neural Networks for One-Shot Image Recognition,Gregory R. Koch,"The process of learning good features for machine learning applications can be very computationally expensive and may prove difficult in cases where little data is available. A prototypical example of this is the one-shot learning setting, in which we must correctly make predictions given only a single example of each new class. In this paper, we explore a method for learning siamese neural networks which employ a unique structure to naturally rank similarity between inputs. Once a network has been tuned, we can then capitalize on powerful discriminative features to generalize the predictive power of the network not just to new data, but to entirely new classes from unknown distributions. Using a convolutional architecture, we are able to achieve strong results which exceed those of other deep learning models with near state-of-the-art performance on one-shot classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A method for learning siamese neural networks which employ a unique structure to naturally rank similarity between inputs and is able to achieve strong results which exceed those of other deep learning models with near state-of-the-art performance on one-shot classification tasks.'}",
-Flow-guided One-shot Talking Face Generation with a High-resolution Audio-visual Dataset,Zhimeng Zhang,"One-shot talking face generation should synthesize high visual quality facial videos with reasonable animations of expression and head pose, and just utilize arbitrary driving audio and arbitrary single face image as the source. Current works fail to generate over 256×256 resolution realistic-looking videos due to the lack of an appropriate high-resolution audio-visual dataset, and the limitation of the sparse facial landmarks in providing poor expression details. To synthesize high-definition videos, we build a large in-the-wild high-resolution audio-visual dataset and propose a novel flow-guided talking face generation framework. The new dataset is collected from youtube and consists of about 16 hours 720P or 1080P videos. We leverage the facial 3D morphable model (3DMM) to split the framework into two cascaded modules instead of learning a direct mapping from audio to video. In the first module, we propose a novel animation generator to produce the movements of mouth, eyebrow and head pose simultaneously. In the second module, we transform animation into dense flow to provide more expression details and carefully design a novel flow-guided video generator to synthesize videos. Our method is able to produce high-definition videos and outperforms state-of-the-art works in objective and subjective comparisons*.","{'model': 'tldr@v2.0.0', 'text': 'This work builds a large in-the-wild high-resolution audio-visual dataset and proposes a novel flow-guided talking face generation framework that is able to produce high-definition videos and outperforms state-of- the-art works in objective and subjective comparisons.'}",
-Heterogeneity for the Win: One-Shot Federated Clustering,D. Dennis,"In this work, we explore the unique challenges -- and opportunities -- of unsupervised federated learning (FL). We develop and analyze a one-shot federated clustering scheme, $k$-FED, based on the widely-used Lloyd's method for $k$-means clustering. In contrast to many supervised problems, we show that the issue of statistical heterogeneity in federated networks can in fact benefit our analysis. We analyse $k$-FED under a center separation assumption and compare it to the best known requirements of its centralized counterpart. Our analysis shows that in heterogeneous regimes where the number of clusters per device $(k')$ is smaller than the total number of clusters over the network $k$, $(k'\le \sqrt{k})$, we can use heterogeneity to our advantage -- significantly weakening the cluster separation requirements for $k$-FED. From a practical viewpoint, $k$-FED also has many desirable properties: it requires only round of communication, can run asynchronously, and can handle partial participation or node/network failures. We motivate our analysis with experiments on common FL benchmarks, and highlight the practical utility of one-shot clustering through use-cases in personalized FL and device sampling.","{'model': 'tldr@v2.0.0', 'text': ""This work develops and analyzes a one-shot federated clustering scheme, $k$-FED, based on the widely-used Lloyd's method, and shows that in heterogeneous regimes where the number of clusters per device $(k')$ is smaller than the total number of clustering over the network, the issue of statistical heterogeneity in federated networks can in fact benefit the analysis.""}",
-One shot learning based drivers head movement identification using a millimetre wave radar sensor,Hong Nhung Nguyen,"Concentration of drivers on traffic is a vital safety issue; thus, monitoring a driver being on road becomes an essential requirement. The key purpose of supervision is to detect abnormal behaviours of the driver and promptly send warnings to him her for avoiding incidents related to traffic accidents. In this paper, to meet the requirement, based on radar sensors applications, the authors first use a small sized millimetre wave radar installed at the steering wheel of the vehicle to collect signals from different head movements of the driver. The received signals consist of the reflection patterns that change in response to the head movements of the driver. Then, in order to distinguish these different movements, a classifier based on the measured signal of the radar sensor is designed. However, since the collected data set is not large, in this paper, the authors propose One shot learning to classify four cases of driver's head movements. The experimental results indicate that the proposed method can classify the four types of cases according to the various head movements of the driver with a high accuracy reaching up to 100. In addition, the classification performance of the proposed method is significantly better than that of the convolutional neural network model.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results indicate that the proposed One shot learning to classify four types of cases according to the various head movements of the driver with a high accuracy reaching up to 100 and the classification performance is significantly better than that of the convolutional neural network model.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1049/rsn2.12223
-Single Path One-Shot Neural Architecture Search with Uniform Sampling,Zichao Guo,,"{'model': 'tldr@v2.0.0', 'text': 'A Single Path One-Shot model is proposed to construct a simplified supernet, where all architectures are single paths so that weight co-adaption problem is alleviated.'}",http://arxiv.org/pdf/1904.00420
-NAS-Bench-1Shot1: Benchmarking and Dissecting One-shot Neural Architecture Search,Arber Zela,"One-shot neural architecture search (NAS) has played a crucial role in making NAS methods computationally feasible in practice. Nevertheless, there is still a lack of understanding on how these weight-sharing algorithms exactly work due to the many factors controlling the dynamics of the process. In order to allow a scientific study of these components, we introduce a general framework for one-shot NAS that can be instantiated to many recently-introduced variants and introduce a general benchmarking framework that draws on the recent large-scale tabular benchmark NAS-Bench-101 for cheap anytime evaluations of one-shot NAS methods. To showcase the framework, we compare several state-of-the-art one-shot NAS methods, examine how sensitive they are to their hyperparameters and how they can be improved by tuning their hyperparameters, and compare their performance to that of blackbox optimizers for NAS-Bench-101.","{'model': 'tldr@v2.0.0', 'text': 'A general framework for one-shot NAS that can be instantiated to many recently-introduced variants and a general benchmarking framework that draws on the recent large-scale tabular benchmark NAS-Bench-101 for cheap anytime evaluations of one- shot NAS methods are introduced.'}",
-One‐Shot Device Testing Data,,,,
-GreedyNAS: Towards Fast One-Shot NAS With Greedy Supernet,Shan You,"Training a supernet matters for one-shot neural architecture search (NAS) methods since it serves as a basic performance estimator for different architectures (paths). Current methods mainly hold the assumption that a supernet should give a reasonable ranking over all paths. They thus treat all paths equally, and spare much effort to train paths. However, it is harsh for a single supernet to evaluate accurately on such a huge-scale search space (e.g., 7^21). In this paper, instead of covering all paths, we ease the burden of supernet by encouraging it to focus more on evaluation of those potentially-good ones, which are identified using a surrogate portion of validation data. Concretely, during training, we propose a multi-path sampling strategy with rejection, and greedily filter the weak paths. The training efficiency is thus boosted since the training space has been greedily shrunk from all paths to those potentially-good ones. Moreover, we further adopt an exploration and exploitation policy by introducing an empirical candidate path pool. Our proposed method GreedyNAS is easy-to-follow, and experimental results on ImageNet dataset indicate that it can achieve better Top-1 accuracy under same search space and FLOPs or latency level, but with only ~60% of supernet training cost. By searching on a larger space, our GreedyNAS can also obtain new state-of-the-art architectures.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-path sampling strategy with rejection, and greedily filter the weak paths to ease the burden of supernet by encouraging it to focus more on evaluation of those potentially-good ones, which are identified using a surrogate portion of validation data.'}",https://arxiv.org/pdf/2003.11236
-Distilled One-Shot Federated Learning,Yanlin Zhou,"Current federated learning algorithms take tens of communication rounds transmitting unwieldy model weights under ideal circumstances and hundreds when data is poorly distributed. Inspired by recent work on dataset distillation and distributed one-shot learning, we propose Distilled One-Shot Federated Learning, which reduces the number of communication rounds required to train a performant model to only one. Each client distills their private dataset and sends the synthetic data (e.g. images or sentences) to the server. The distilled data look like noise and become useless after model fitting. We empirically show that, in only one round of communication, our method can achieve 96% test accuracy on federated MNIST with LeNet (centralized 99%), 81% on federated IMDB with a customized CNN (centralized 86%), and 84% on federated TREC-6 with a Bi-LSTM (centralized 89%). Using only a few rounds, DOSFL can match the centralized baseline on all three tasks. By evading the need for model-wise updates (i.e., weights, gradients, loss, etc.), the total communication cost of DOSFL is reduced by over an order of magnitude. We believe that DOSFL represents a new direction orthogonal to previous work, towards weight-less and gradient-less federated learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Distilled One-Shot Federated Learning, which reduces the number of communication rounds required to train a performant model to only one, and represents a new direction orthogonal to previous work, towards weight-less and gradient-less federated learning.'}",
-Adversarial Style Mining for One-Shot Unsupervised Domain Adaptation,Yawei Luo,"We aim at the problem named One-Shot Unsupervised Domain Adaptation. Unlike traditional Unsupervised Domain Adaptation, it assumes that only one unlabeled target sample can be available when learning to adapt. This setting is realistic but more challenging, in which conventional adaptation approaches are prone to failure due to the scarce of unlabeled target data. To this end, we propose a novel Adversarial Style Mining approach, which combines the style transfer module and task-specific module into an adversarial manner. Specifically, the style transfer module iteratively searches for harder stylized images around the one-shot target sample according to the current learning state, leading the task model to explore the potential styles that are difficult to solve in the almost unseen target domain, thus boosting the adaptation performance in a data-scarce scenario. The adversarial learning framework makes the style transfer module and task-specific module benefit each other during the competition. Extensive experiments on both cross-domain classification and segmentation benchmarks verify that ASM achieves state-of-the-art adaptation performance under the challenging one-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel Adversarial Style Mining approach, which combines the style transfer module and task-specific module into an adversarial manner, which achieves state-of-the-art adaptation performance under the challenging one-shot setting.'}",
-Overcoming Multi-Model Forgetting in One-Shot NAS With Diversity Maximization,Miao Zhang,"One-Shot Neural Architecture Search (NAS) significantly improves the computational efficiency through weight sharing. However, this approach also introduces multi-model forgetting during the supernet training (architecture search phase), where the performance of previous architectures degrade when sequentially training new architectures with partially-shared weights. To overcome such catastrophic forgetting, the state-of-the-art method assumes that the shared weights are optimal when jointly optimizing a posterior probability. However, this strict assumption is not necessarily held for One-Shot NAS in practice. In this paper, we formulate the supernet training in the One-Shot NAS as a constrained optimization problem of continual learning that the learning of current architecture should not degrade the performance of previous architectures during the supernet training. We propose a Novelty Search based Architecture Selection (\textbf{NSAS}) loss function and demonstrate that the posterior probability could be calculated without the strict assumption when maximizing the diversity of the selected constraints. A greedy novelty search method is devised to find the most representative subset to regularize the supernet training. We apply our proposed approach to two One-Shot NAS baselines, random sampling NAS (RandomNAS) and gradient-based sampling NAS (GDAS). Extensive experiments demonstrate that our method enhances the predictive ability of the supernet in One-Shot NAS and achieves remarkable performance on CIFAR-10, CIFAR-100, and PTB with efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Novelty Search based Architecture Selection (\\textbf{NSAS}) loss function and demonstrates that the posterior probability could be calculated without the strict assumption when maximizing the diversity of the selected constraints.'}",https://research-repository.griffith.edu.au/bitstream/10072/419622/2/Pan2923687-Accepted.pdf
-Understanding Mathematical Concept: The Effect Of Savi Learning Model With Probing-Prompting Techniques Viewed From Self-Concept,R. Widyastuti,"This research aims to discover the effect of SAVI learning models with the Probing-prompting technique in terms of self-concepts on students’ mathematical conceptual understanding. The method used in this research was the quasi-experimental design. The experimental design used in this study was the post-test only design with a 3x3 factorial research design. The population in this study were 153 students. The sampling technique used was random. The normality test, homogeneity test, two way ANOVA test and Scheffe’ method were used as data analysis techniques. The results showed that there is an influence of the SAVI learning model with probing-prompting techniques in terms of students’ self-concepts on the ability to understand mathematical concepts compared to conventional learning models applied by teachers at school.",,
-Ability of children to perform touchscreen gestures and follow prompting techniques when using mobile apps,Savita Yadav,"Background Children today get access to smartphones at an early age. However, their ability to use mobile apps has not yet been studied in detail. Purpose This study aimed to assess the ability of children aged 2–8 years to perform touchscreen gestures and follow prompting techniques, i.e., ways apps provide instructions on how to use them. Methods We developed one mobile app to test the ability of children to perform various touchscreen gestures and another mobile app to test their ability to follow various prompting techniques. We used these apps in this study of 90 children in a kindergarten and a primary school in New Delhi in July 2019. We noted the touchscreen gestures that the children could perform and the most sophisticated prompting technique that they could follow. Results Two- and 3-year-old children could not follow any prompting technique and only a minority (27%) could tap the touchscreen at an intended place. Four- to 6-year-old children could perform simple gestures like a tap and slide (57%) and follow instructions provided through animation (63%). Seven- and 8-year-old children could perform more sophisticated gestures like dragging and dropping (30%) and follow instructions provided in audio and video formats (34%). We observed a significant difference between the number of touchscreen gestures that the children could perform and the number of prompting techniques that they could follow (F=544.0407, P<0.05). No significant difference was observed in the performance of female versus male children (P>0.05). Conclusion Children gradually learn to use mobile apps beginning at 2 years of age. They become comfortable performing single-finger gestures and following nontextual prompting techniques by 8 years of age. We recommend that these results be considered in the development of mobile apps for children.","{'model': 'tldr@v2.0.0', 'text': 'Children gradually learn to use mobile apps beginning at 2 years of age and become comfortable performing single-finger gestures and following nontextual prompting techniques by 8 years ofAge, which is recommended in the development of mobile apps for children.'}",https://www.e-cep.org/upload/pdf/cep-2019-00997.pdf
-Improving Critical Thinking Skills through Module Solubility and Solubility Results Based on Discovery Learning with Probing Prompting Techniques at SMAN 1 Pariaman,Fefri Perrianty,"Abstract - This study aims to determine the improvement of student critical thinking based on discovery learning using modules that help investigate questions that encourage questions about solubility and solubility results. The research design used was the Non-Equivalent Post-test Group Design. The population in this study were students of class XI IPA SMAN 1 Pariaman academic year 2018/2019. The selection of sample classes is done by cluster purposive sampling technique. The sample in this study was class XI IPA 1 as the control class and XI IPA 2 as the experimental class. The experimental learning class uses modules based on discovery learning withtechniques probing prompting, while the control class without using modules. The research instrument used was a test of the results of learning critical thinking questions. The results of the study show that module based discovery learning withtechniques probing promptingcan improve students' critical thinking skills, as evidenced by the average increase in posttest results. The results of testing for normality and homogeneity state that the critical thinking value of the two distributed samples is normal and homogeneous. Hypothesis test results obtained by GIS. (2-tailed) of 0.016 <0.05 means that the value of critical thinking learners who use modules is based on discovery learning bytechniques probing prompting differ significantly",,
-LEARNING MODELS USING PROBING PROMPTING TECHNIQUES ASSISTED BY BLENDED LEARNING EDMODO TO IMPROVE MATHEMATICAL UNDERSTANDING ABILITY,R. Rahmawati,"This study aims to see an increase in the ability of mathematical understanding of students using probing learning models prompting assisted by edmodo blended learning. In this model the researcher presents a series of questions that are digging and guiding so that a thinking process that links each student's knowledge and experience with new knowledge is occurring. learned. Research subject of 3rd Semester Student in Mathematics and Teacher Training Faculty, University of PGRI Palembang. This research method is Quasi-Experimental. The data of this study are of two types. Quantitative data was obtained through analysis of student answers on the KPM test which was carried out pretest and posttest, and qualitative data were obtained from observations of lecturer and student activities, and the results of interviews with students. Data analysis used: doing descriptive analysis of data and calculating normalized gain (normalized gain) pretest and posttest, Testing requirements analysis is the data normality test and variance homogeneity test of the whole quantitative data, and hypothesis testing in this study t test, t-test, Mann-Whitney U test, one-way ANAVA test or Kruskal-Wallis test. Hypothesis testing using SPSS-20 for Windows. The results of the data analysis concluded that 1) the increase in the ability of mathematical understanding of students who get learning by probing prompting assisted by edmodo blended learning is better than students who get conventional learning in terms of (a). all students (b) .KAM students",,https://ojs.unm.ac.id/JDM/article/download/8843/5139
-DeepLyrics: GPT2 for lyrics generation with finetuning and prompting techniques,Li Tian,"Lyrics are incredibly important part of a song’s success. The fast advancing natural language processing (NLP) methods have been successful in lyrics generation, making AI-assisting lyrics creation possible. However, as these methods get more resource-consuming and data-demanding, we identify a neglected research area of exploring efficient ways of model learning to simplify existing lyrics generation methods. Our project proposes DeepLyrics, a GPT-2 model using tuning-free prompting (in-context learning) on lyrics of highly successful songs in the past several decades to assist creative generation. Our tuning-free method, DeepLyrics, is able to achieve comparable and even better performance compared to thoroughly fine-tuned lyrics generation models. Our work shows the practicability of reducing large amount of training and finetuning work in existing lyrics generation methods with a well-engineered prompting technique. 1 Key Information to include • Mentor:NA • External Collaborators: Ying Lin @ Stanford Culture Lab (data preparation only) • Sharing project: NA","{'model': 'tldr@v2.0.0', 'text': 'This work shows the practicability of reducing large amount of training and finetuning work in existing lyrics generation methods with a well-engineered prompting technique, and achieves comparable and even better performance compared to thoroughly fine-tuned lyrics generation models.'}",
-"The Effect of POE Learning Models (Prediction, Observation, and Explanation) with Probing-Prompting Techniques on The Student’s Cognitive Learning Outcomes of SMA Muhammadiyah 3 Jember",Dewi Farida,"Student learning outcomes are competencies that students have after gaining their learning experience, so indicators of learning success can be seen from student learning outcomes after experiencing the learning process. Student learning outcomes can be improved through the use of learning models, one of which is by applying the POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques. The purpose of this study was to determine the effect of POE learning models (Prediction, Observation and Explanation) with probing-prompting techniques on the cognitive learning outcomes of the tenth grade students at SMA Muhammadiyah 3 Jember on the environmental pollution subject. Type of research is a quasy experimental study using a pretest and posttest design. In the experimental class, learning activity was carried out by applying the POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques and the control class, learning activities are carried out by applying conventional learning models. The data of the students' cognitive learning outcomes were obtained through the pretest and posttest scores then analyzed using ANAKOVA test. The application of POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques had a significant effect with a probability as much as 0,000. The affective learning outcomes influenced significantly with a significance value of 0.001 or p<0.05.",,http://pancaran.kreatindo.com/index.php/pancaran/article/download/199/217
-Prompting Techniques for Guidance and Action Assistance Using Augmented-Reality Smart-Glasses,Patrick Renner,"In the context of picking and assembly tasks, assistance systems based on Augmented Reality (AR) can help finding target objects and conducting correct actions. The aim is to develop guiding and action assistance techniques for smart glasses, which are easily understandable not only for workers, but also for impaired and elderly people.","{'model': 'tldr@v2.0.0', 'text': 'The aim is to develop guiding and action assistance techniques for smart glasses, which are easily understandable not only for workers, but also for impaired and elderly people.'}",
-Effects of Video Prompting Techniques on Teaching Daily Living Skills to Children With Autism Spectrum Disorders,Sarah C. Domire,"Video-based instruction is becoming a common intervention in today’s classrooms. Previous research has focused primarily on video modeling techniques that required the student to watch an entire video of the task before attempting to complete the task independently. Video prompting is a form of video instruction that breaks down target skills into steps that are then performed directly after viewing each clip. The present review examined studies using video prompting techniques to teach functional and daily living skills to individuals with autism spectrum disorders (ASD). The focus of the review was on evaluation of the effectiveness of video prompting and the factors that affect student attention to the video, retention of target behavior, production of target behavior, and motivation. Results showed that video prompting was an effective intervention for teaching a wide array of target skills and that students with ASD were able to generalize and maintain the acquired skills. Video prompting was also shown to be more effective than both static picture prompts and video modeling techniques in terms of percentage of correct independent responding. Suggestions for practice and future research are discussed.",,
-Comparison of Prompting Techniques to Teach Children With Autism to Ask Questions in the Context of a Conversation,Matthew G. Swerdan,"An adapted alternating treatments design was used to compare the efficacy of echoic and textual prompts to teach three students with autism (ages 8–15) to ask questions related to two pre-selected topics of conversation. Participants were first required to answer questions related to the topics to determine whether accurate responses were within their repertoire. This was followed by a transfer of stimulus control procedure to teach participants to ask relevant questions to the experimenter on the same topics. Probes with a novel conversation topic were conducted in the natural environment with a peer, and follow-up probes were conducted after training. Results indicate questions taught using an echoic prompt were acquired in fewer trials to criterion than questions taught using a textual prompt for two of the three participants. Limitations and implications for future research will be discussed.",,
-The Difference between Mathematical Reasoning Ability Improvement by Learning with Meta Cognitive Approach Aided Probing and Prompting Techniques in SMP Negeri 4 SeiSuka,Nadran Hamdani Siregar,"The purpose of this study were: (1) analyzed the differences in students’ mathematical reasoning ability improvement taught by metacognition approach aided probing technique (PMT-probing) and metacognition approach aided prompting technique (PMT-prompting); and (2) described the process of the students’ responses in solving mathematical reasoning abilities. This study was a quasi experimentalresearch. The population in this study were all students of class VIII SMP Negeri 4 SeiSuka, with a purposive sampling techniques, the obtained sample was VIII-1 and VIII-2. The research instrument used a test of mathematical reasoning ability, and had qualified the criteria of content validity, and reliability coefficient of 0.819. Anova two ways was used to analyze the difference of mathematical reasoning ability improvement, while descriptive analysis was used to analyze students’ answers process. The results showed that: (1) There were differences in students’ mathematical reasoning skills improvement which were taught by metacognition approach aided probing techniques and the students taught by prompting technical approach; and (2) The process of the students’ responses on students’ mathematical reasoning abilitythrough learning with metacognition approach aided by prompting techniques was better than metacognition approach aided by probing techniques.",,http://www.scholink.org/ojs/index.php/wjer/article/download/766/781
-Effects of Teacher Prompting Techniques on the Writing Performance of Fourth and Fifth Grade Students,Mindy Allenger,,,
-Effect of Prompting Techniques and Learning Styles on Requirements Elicitation,Bonnie Lynn Jan,,,
-"A Survey of Graph Prompting Methods: Techniques, Applications, and Challenges",Xuansheng Wu,"The recent""pre-train, prompt, predict training""paradigm has gained popularity as a way to learn generalizable models with limited labeled data. The approach involves using a pre-trained model and a prompting function that applies a template to input samples, adding indicative context and reformulating target tasks as the pre-training task. However, the design of prompts could be a challenging and time-consuming process in complex tasks. The limitation can be addressed by using graph data, as graphs serve as structured knowledge repositories by explicitly modeling the interaction between entities. In this survey, we review prompting methods from the graph perspective, where prompting functions are augmented with graph knowledge. In particular, we introduce the basic concepts of graph prompt learning, organize the existing work of designing graph prompting functions, and describe their applications and future challenges. This survey will bridge the gap between graphs and prompt design to facilitate future methodology development.","{'model': 'tldr@v2.0.0', 'text': 'This survey will bridge the gap between graphs and prompt design to facilitate future methodology development and introduce the basic concepts of graph prompt learning, organize the existing work of designing graph prompting functions, and describe their applications and future challenges.'}",http://arxiv.org/pdf/2303.07275
-Prompting techniques – researching subjectivities in educational history.,Helle Bjerg,,,
-HELP ME THINK: A Simple Prompting Strategy for Non-experts to Create Customized Content with Models,Swaroop Mishra,"Controlling the text generated by language models and customizing the content has been a long-standing challenge. Existing prompting techniques proposed in pursuit of providing control are task-specific and lack generality; this provides overwhelming choices for non-expert users to find a suitable method for their task. The effort associated with those techniques, such as in writing examples, explanations, instructions, etc. further limits their adoption among non-expert users. In this paper, we propose a simple prompting strategy HELP ME THINK where we encourage GPT3 to help non-expert users by asking a set of relevant questions and leveraging user answers to execute the task. We demonstrate the efficacy of our technique HELP ME THINK on a variety of tasks. Specifically, we focus on tasks that are hard for average humans and require significant thinking to perform. We hope our work will encourage the development of unconventional ways to harness the power of large language models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple prompting strategy HELP ME THINK where GPT3 is encouraged to help non-expert users by asking a set of relevant questions and leveraging user answers to execute the task.'}",http://arxiv.org/pdf/2208.08232
-An Empirical Investigation of User Requirements Elicitation: Comparing the Effectiveness of Prompting Techniques,Glenn J. Browne,"Eliciting requirements from users and other stakeholders is of central importance to information systems development. Despite this importance, surprisingly little research has measured the effectiveness of various requirements elicitation techniques. The present research first discusses theory relevant to information requirements determination in general and elicitation in particular.We then develop a model of the requirements elicitation process. This model and its underlying theory were then used to construct a new requirements elicitation prompting technique. To provide a context for testing the relative effectiveness of the new technique, two other questioning methodologies were also operationalized as prompting techniques: (1) the interrogatories technique, which involves asking ""who,"" ""what,"" ""when,"" ""where,"" ""how,"" and ""why""; questions; and (2) a semantic questioning scheme, which involves asking questions based on a theoretical model of knowledge structures. To measure the usefulness of the prompting techniques in eliciting requirements, a set of generic requirements categories was adapted from previous research to capture requirements evoked by users. The effectiveness of the three methods in eliciting requirements for a software application was then tested in an experiment with users. Results showed that the new prompting technique elicited a greater quantity of requirements from users than did the other two techniques. Implications of the findings for research and systems analysis practice are discussed.","{'model': 'tldr@v2.0.0', 'text': None}",
-Evaluation of computer-aided detection (CAD) prompting techniques for mammography.,S. Astley,"Computer-aided detection (CAD) systems, in which abnormalities are automatically detected and their locations presented to the radiologist as prompts, are increasingly being used to improve reader performance. The performance of CAD systems can be evaluated in two ways: by measuring the performance of the algorithms, or by monitoring the performance of readers using the system. All aspects of evaluation need careful consideration to avoid potential bias. This paper examines a variety of different approaches to evaluation and discusses their relative strengths and weaknesses.","{'model': 'tldr@v2.0.0', 'text': 'A variety of different approaches to evaluation of CAD systems are examined and their relative strengths and weaknesses are discussed.'}",
-PROMPTING TECHNIQUES TO INCREASE THE RETURN RATE OF MAILED QUESTIONNAIRES1,R. Winett,"To increase the return rate of questionnaires mailed to clergy and physicians concerning their mental-health practices, different prompts were used after the questionnaire was received during four mail-outs to four randomly drawn samples of clergy and physicians. For each mail-out, the sample was divided into experimental (received prompt) and comparison (no prompt) groups, and one type of prompt or combination was used. Non-returnees of the questionnaire in the experimental group received either: (a) a single telephone call, (b) a memo, (c) a package (personal letter and new questionnaire) or package plus a telephone call, or (d) a double call. Comparison physicians and clergy were mailed only the original questionnaire. Relative to their respective comparison group's return rate, which averaged 22% across the four mail-outs (range 18% to 24%), the single call and package alone about doubled the overall return rate, the package and call increased the return rate about two-and-a-half fold, and the double call almost tripled the return rate. The memo was ineffective. A cost-effectiveness analysis indicated that the double-call procedure was less expensive than the single call, and much less expensive than the package alone or package with a call in securing returns. An analysis of the pattern of returns showed clearly that when prompts were not delivered (comparison groups), very few returns were received after about seven days from the initial mail-out. Most returns from prompts (experimental groups) were received by several days after the prompt. The results were seen as salient to the problem of reducing selection or volunteer bias in questionnaire studies and subsequent research demonstrating the effectiveness of telephone calls made about a week after distribution of surveys in securing high return rates was discussed.","{'model': 'tldr@v2.0.0', 'text': 'The results were seen as salient to the problem of reducing selection or volunteer bias in questionnaire studies and subsequent research demonstrating the effectiveness of telephone calls made about a week after distribution of surveys in securing high return rates was discussed.'}",https://europepmc.org/articles/pmc1311322?pdf=render
-Neuro-Symbolic Causal Language Planning with Commonsense Prompting,Yujie Lu,"Language planning aims to implement complex high-level goals by decomposition into sequential simpler low-level steps. Such procedural reasoning ability is essential for applications such as household robots and virtual assistants. Al-though language planning is a basic skill set for humans in daily life, it remains a challenge for large language models (LLMs) that lack deep-level commonsense knowledge in the real world. Previous methods require either manual exemplars or annotated programs to acquire such ability from LLMs. In contrast, this paper proposes Neuro-Symbolic C ausal LA nguage P lanner (CLAP) that elicits procedural knowledge from the LLMs with commonsense-infused prompting. Pre-trained knowledge in LLMs is essentially an unobserved confounder that causes spurious correlations between tasks and action plans. Through the lens of a Structural Causal Model (SCM), we propose an effective strategy in CLAP to construct prompts as a causal intervention toward our SCM. Using graph sampling techniques and symbolic program executors, our strategy formalizes the structured causal prompts from commonsense knowledge bases. CLAP obtains state-of-the-art performance on WikiHow and RobotHow, achieving a relative improvement of 5 . 28% in human evaluations under the counterfactual setting. This indicates the superiority of CLAP in causal language planning semantically and sequentially. as household robots and virtual assistants. We propose a Neuro-Symbolic Causal Language Planner (CLAP) with commonsense prompts elicited from the commonsense knowledge base to solve the language planning problem in a zero-shot manner. Experiments show the effectiveness of our proposed CLAP on both automatic and human evaluation results. Extending causal language planning to handle the long-horizon composite tasks and provide effective automatic evaluation metrics are important directions for future work.","{'model': 'tldr@v2.0.0', 'text': 'A Neuro-Symbolic Causal Language Planner (CLAP) is proposed that elicits procedural knowledge from the LLMs with commonsense-infused prompting to solve the language planning problem in a zero-shot manner.'}",http://arxiv.org/pdf/2206.02928
-Generative Speech Recognition Error Correction with Large Language Models and Task-Activating Prompting,Chao-Han Huck Yang,"We explore the ability of large language models (LLMs) to act as speech recognition post-processors that perform rescoring and error correction. Our first focus is on instruction prompting to let LLMs perform these task without fine-tuning, for which we evaluate different prompting schemes, both zero- and few-shot in-context learning, and a novel task activation prompting method that combines causal instructions and demonstration to increase its context windows. Next, we show that rescoring only by in-context learning with frozen LLMs achieves results that are competitive with rescoring by domain-tuned LMs, using a pretrained first-pass recognition system and rescoring output on two out-of-domain tasks (ATIS and WSJ). By combining prompting techniques with fine-tuning we achieve error rates below the N-best oracle level, showcasing the generalization power of the LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the ability of large language models to act as speech recognition post-processors that perform rescoring and error correction, and shows that rescoring only by in-context learning with frozen LLMs achieves results that are competitive with rescoring by domain-tuned LMs.'}",https://arxiv.org/pdf/2309.15649
-LLM-Rec: Personalized Recommendation via Prompting Large Language Models,Hanjia Lyu,"We investigate various prompting strategies for enhancing personalized recommendation performance with large language models (LLMs) through input augmentation. Our proposed approach, termed LLM-Rec, encompasses four distinct prompting strategies: (1) basic prompting, (2) recommendation-driven prompting, (3) engagement-guided prompting, and (4) recommendation-driven + engagement-guided prompting. Our empirical experiments show that incorporating the augmented input text generated by LLM leads to improved recommendation performance. Recommendation-driven and engagement-guided prompting strategies are found to elicit LLM's understanding of global and local item characteristics. This finding highlights the importance of leveraging diverse prompts and input augmentation techniques to enhance the recommendation capabilities with LLMs.","{'model': 'tldr@v2.0.0', 'text': ""Recommendation-driven and engagement-guided prompting strategies are found to elicit LLM's understanding of global and local item characteristics and highlight the importance of leveraging diverse prompts and input augmentation techniques to enhance the recommendation capabilities with LLMs.""}",https://arxiv.org/pdf/2307.15780
-Coaching and Prompting for Remote Physical Activity Promotion: Study Protocol of a Three-Arm Randomized Controlled Trial (Movingcall),X. Fischer,"Background. Physical inactivity is currently seen as one of the biggest global public health issue. Remote physical activity (PA) promotion programs are expected to be effective if they are individually tailored and include behavior change techniques, personal coaching, and regular prompting. However, it is still not fully understood which intervention components are most effective. This paper describes the rationale and design of a study on an individually tailored remote PA promotion program comparing the efficacy of coaching and prompting with a single written advice. Methods. In total, 288 adults (age 20 to 65 years) were randomly assigned to three different intervention arms of a 6-month-long PA promotion program. A minimal intervention group received a single written PA recommendation. The two remaining groups either received telephone coaching sessions (n = 12 calls) with or without additional short message service (SMS) prompting (n = 48 SMSs for each participant). Data assessment took place at baseline, at the end of the intervention, and after a six-month follow-up-period. The primary outcome of the study was self-reported PA. Objectively assessed PA, psychosocial determinants of PA, well-being, body mass index (BMI), and adherence were assessed as secondary outcomes. Conclusion. Findings of this three-arm study will provide insight into the short and long-term effects of coaching and prompting for PA promotion.","{'model': 'tldr@v2.0.0', 'text': 'The rationale and design of a study on an individually tailored remote PA promotion program comparing the efficacy of coaching and prompting with a single written advice is described and insight into the short and long-term effects of coach and prompting for PA promotion is provided.'}",https://www.mdpi.com/1660-4601/16/3/331/pdf?version=1548412965
-Validitas dan Praktikalitas Modul Hukum-Hukum Dasar Kimia Berbasis Pendekatan Saintifik dengan Menerapkan Teknik Probing Prompting untuk Siswa Kelas X SMA/MA,Yuni Purnama Sari,"The Implementation of the curriculum is carried out by practicing process skills that are reflected in learning activities. The process skills applied in the form of 5M activities are known as the scientific approach. This research aims to produce of modul of basic chemical laws based on a scientific approach by applying probing prompting techniques. The research instrument used was a questionnaire of validity and practicality. The module was validated by 5 validators while the practicality test was carried out by 2 chemistry teachers and 20 XI grade students of MIA 3 SMAN 1 Basa Ampek Balai Tapan, Pesisir Selatan. Data from the validity and practicality test results were analyzed using the cohen kappa formula. Based on the results of the study, it was found that the average kappa moment of validity test was 0,89 with a very high validity category and the average kappa moment of teacher practicality was 0,90 and the average kappa moment of students practicality was 0,91 with a very high practicality category. Thus, it was concluded that basic chemical laws module based on a scientific approach by applying probing prompting techniques was valid and practical for high school students.",,http://edukimia.ppj.unp.ac.id/ojs/index.php/edukimia/article/download/22/20
-Which Behaviour Change Techniques Are Most Effective at Increasing Older Adults’ Self-Efficacy and Physical Activity Behaviour? A Systematic Review,D. French,,"{'model': 'tldr@v2.0.0', 'text': 'Many commonly used self-regulation intervention techniques that are effective for younger adults may not be effective for older adults.'}",https://openaccess.city.ac.uk/id/eprint/4248/1/Which_behavior_change_techniques_are_most_effective_at_increasing_older_adults%E2%80%99_selfefficacy.pdf
-Prompting conceptual understanding with computer-mediated peer discourse and knowledge acquisition techniques,Chen-Chung Liu,"Numerous computer-mediated communication (CMC) tools have been designed to facilitate group discourses on the Web. However, previous studies noted that participants did not value online conferencing as a method for conducting in-depth discussions, and instead considered this method as merely scratching the surface of the issues involved. Therefore, the teachers must moderate the online discourse to add value to the success of CMC. This study presents a methodology that uses knowledge acquisition techniques during online discourses to increase student conceptual understanding and evaluation skills. The methodology explains the conceptual frameworks of individuals and the conceptual relationships between student peers. Active database triggers then are applied to automatically monitor the conceptual relationships and to moderate online discourses between student peers. Student reactions demonstrated that discourses mediated by the repertory grid analysis technique helped group members not only promote common conceptual understandings but also to approach the level of conceptual understanding of the teacher.","{'model': 'tldr@v2.0.0', 'text': 'A methodology that uses knowledge acquisition techniques during online discourses to increase student conceptual understanding and evaluation skills is presented and demonstrated that discourses mediated by the repertory grid analysis technique helped group members not only promote common conceptual understandings but also to approach the level of conceptual understanding of the teacher.'}",
-PEMBELAJARAN MENGGUNAKAN TEKNIK PROBING PROMPTING BERBANTUAN EDMODO BLENDED LEARNING PADA MATERI PERSAMAAN DIFERENSIAL MATEMATIS MAHASISWA DI UNIVERSITAS PGRI PALEMBANG,Jayanti Jayanti,"Abstract — This study aims to examine the extent of learning using probing prompting techniques applied in the course of differential equations at PGRI University Palembang assisted Edmodo Blended Learning. Learning model using Probing Prompting technique, aided by edmodo blended learning that is by using internet online, to assist them in learning. Probing Prompting Techniques is a way of presenting learning with a series of questions that are guided and explored to form a thought process that links previous knowledge and experience with new knowledge to be learned. Questions raised on the students encourage students to think more rationally about the knowledge that has been obtained before, and also can link the next questions so that new knowledge arises that is reflected from the results of a lesson. This research uses derskriftif method which only describes process from learning of differential equation using probing prompting. The subject of this research is a mathematics student in the course of Differential Equation at University PGRI of Palembang. Keywords — Probing Prompting, Edmodo Blended learning, and Differential Equation (PD) Abstrak — Penelitian ini bertujuan untuk mengkaji sejauhmana pembelajaran menggunakan teknik probing prompting yang diterapkan di matakuliah persamaan diferensial pada universitas PGRI Palembang berbantuan Edmodo Blended Learning. Model Pembelajaran dengan menggunakan teknik Probing Prompting, berbantuan edmodo blended learning yaitu dengan menggunakan internet secara online, untuk membantu mereka dalam pembelajaran. Teknik Probing Prompting adalah Cara menyajikan Pembelajaran dengan serangkaian pertanyaan yang sifatnya menuntun dan menggali sehingga membentuk suatu proses berpikir yang mengaitkan pengetahuan dan pengalaman sebelumnya dengan pengetahuan baru yang akan dipelajari. Pertanyaan-pertanyaan yang dilontarkan pada siswa mendorong siswa agar dapat berpikir lebih rasional tentang pengetahuan yang telah di dapat oleh sebelumnya, dan juga dapat mengaitkan pertanyaan-pertanyaan berikutnya sehingga timbul pengetahuan baru yang tercermin dari hasil suatu pembelajaran . Penelitian ini menggunakan metode derskriftif dimana cuma menggambarkan proses dari pembelajaran persamaan diferensial menggunakan probing prompting . Subyek penelitian ini adalah mahasiswa matematika pada mata kuliah Persamaan Diferensial pada Universitas PGRI Palembang. Kata Kunci — Probing Prompting, Edmodo Blended learning, dan Persamaan Diferensial (PD)",,
-Comparing self-directed video prompting to least-to-most prompting in post-secondary students with moderate intellectual disabilities,Helen I. Cannella-Malone,"Objectives: This study compared the effects of self-directed video prompting with error correction using an iPod Touch to least-to-most prompting on teaching vocational skills to two post-secondary students with moderate intellectual disabilities. Methods: Students were first taught to use the inPromptu mobile application on the iPod Touch. They were then taught to collate and pack a large envelope and to tri-fold papers and pack a business envelope. The two tasks were counterbalanced across students in an adapted alternating treatments design. Students were instructed to use self-directed video prompting with error correction to teach themselves one skill, and an instructor used least-to-most prompting to teach the second skill. Data were also collected on sessions to criterion and maintenance of the skills over time. Results: Results demonstrated that both instructional techniques were effective in teaching the skills, and students generalized their use of the technology to learn a new skill with no additional instruction. These results were replicated with one of the students. Conclusions: Self-directed video prompting with error correction and least-to-most prompting were nearly equivalent in terms of acquisition efficiency and maintenance. These findings suggest that self-directed video prompting with error correction could be used in vocational settings, allowing job coaches to fade more quickly and for adults with intellectual and developmental disabilities to be more independent.",,
-Computer-delivered interventions for reducing alcohol consumption: meta-analysis and meta-regression using behaviour change techniques and theory,Nicola Black,"The current aim was to examine the effectiveness of behaviour change techniques (BCTs), theory and other characteristics in increasing the effectiveness of computer-delivered interventions (CDIs) to reduce alcohol consumption. Included were randomised studies with a primary aim of reducing alcohol consumption, which compared self-directed CDIs to assessment-only control groups. CDIs were coded for the use of 42 BCTs from an alcohol-specific taxonomy, the use of theory according to a theory coding scheme and general characteristics such as length of the CDI. Effectiveness of CDIs was assessed using random-effects meta-analysis and the association between the moderators and effect size was assessed using univariate and multivariate meta-regression. Ninety-three CDIs were included in at least one analysis and produced small, significant effects on five outcomes (d+ = 0.07–0.15). Larger effects occurred with some personal contact, provision of normative information or feedback on performance, prompting commitment or goal review, the social norms approach and in samples with more women. Smaller effects occurred when information on the consequences of alcohol consumption was provided. These findings can be used to inform both intervention- and theory-development. Intervention developers should focus on, including specific, effective techniques, rather than many techniques or more-elaborate approaches.","{'model': 'tldr@v2.0.0', 'text': 'Examination of behaviour change techniques, theory and other characteristics in increasing the effectiveness of computer-delivered interventions to reduce alcohol consumption found intervention developers should focus on specific, effective techniques, rather than many techniques or more-elaborate approaches.'}",
-Touchscreen prompts for preschoolers: designing developmentally appropriate techniques for teaching young children to perform gestures,Alexis Hiniker,"Though toddlers and preschoolers are regular touchscreen users, relatively little is known about how they learn to perform unfamiliar gestures. In this paper we assess the responses of 34 children, aged 2 to 5, to the most common in-app prompting techniques for eliciting specific gestures. By reviewing 100 touchscreen apps for preschoolers, we determined the types of prompts that children are likely to encounter. We then evaluated their relative effectiveness in teaching children to perform simple gestures. We found that children under 3 were only able to interpret instructions when they came from an adult model, but that children made rapid gains between age 3 and 3-and-a-half, at which point they were able to follow in-app audio instructions and on-screen demonstrations. The common technique of using visual state changes to prompt gestures was ineffective across this age range. Given that prior work in this space has primarily focused on children's fine motor control, our findings point to a need for increased attention to the design of prompts that accommodate children's cognitive development as well.","{'model': 'tldr@v2.0.0', 'text': 'It was found that children under 3 were only able to interpret instructions when they came from an adult model, but that children made rapid gains between age 3 and 3-and-a-half, at which point they were able to follow in-app audio instructions and on-screen demonstrations.'}",
-Model Discovery Learning Disertai Teknik Probing Prompting Dalam Pembelajaran Fisika Di Ma,Tiani Alfi Kusuma,"Discovery learning model by probing-prompting technique is one of the learning models of physics which guide the students to discover new things, the teacher provides the questions which can stimulate students’creativity and direct students to discover new knowledge. The purpose of this research are (1) to examine the influence of discovery learning model by probing-prompting techniquetothe students’ science-physics achievement at senior high school, and (2) to describe students’ science process skills during learning process. Kind of research is true experimental research by using post-test only control design. The techniques of data collection are observation, interview, documentation, andachievement test. The techniques of data analysis are indepependent sample T-Test and descriptive analysis by percentage. The Result of the study are (1) the value of p based on the T-Test is 0.001, and (2) the percentage of the overall students’science process skills is 81.71%. The study can be concluded that (1) discovery learning model by probing-prompting technique give significance influence to the students’ science-physics achievement, and (2) the students’ science process skills during join the physics learning by using discovery learning model by probing-prompting technique are in very good category",,
-Detection Techniques for Extracellular Polymeric Substances in Biofilms: A Review,Mei Pan,"Extracellular polymeric substances (EPS) are one of the main components of biofilm, prompting biofilm to form a cohesive three-dimensional framework. Numerous methods are available to help characterize the properties and the structural, chemical and physical organizations of EPS during the biofilm formation process. This review highlights key techniques from different disciplines that have been successfully applied in-situ and non-destructively to describe the complex composition and distribution of EPS in biofilm, especially microscopic, spectroscopic, and the combination of multi-disciplinary methods that can provide new insights into the complex structure/function correlations in biofilms. Among them, confocal laser scanning microscopy (CLSM) is emphasized, and its principles, applications, advantages, and limitations are summarized. Multidisciplinary techniques have been developed and recommended to study EPS during the biofilm formation process, providing more in-depth insights into the composition and spatial distributions of EPS, so as to improve our understanding of the role EPS plays in biofilms ultimately.","{'model': 'tldr@v2.0.0', 'text': 'This review highlights key techniques from different disciplines that have been successfully applied in-situ and non-destructively to describe the complex composition and distribution of EPS in biofilm, especially microscopic, spectroscopic, and the combination of multi-disciplinary methods that can provide new insights into the complex structure/function correlations in biofilms.'}",https://ojs.cnr.ncsu.edu/index.php/BioRes/article/download/BioRes_11_3_Pan_Review_8092_Detection_Techniques_Extracellular_Biofilms/4693
-Pharmacokinetics in Morbid Obesity: Influence of Two Bariatric Surgery Techniques on Paracetamol and Caffeine Metabolism,A. Goday Arnó,,"{'model': 'tldr@v2.0.0', 'text': 'The two studied bariatric surgical techniques normalize paracetamol oral bioavailability without impairing the liver function (measured by cytochrome P450 1A2 activity) and caffeine plasma levels are altered in severely obese patients.'}",http://repositori.upf.edu/bitstream/10230/35065/1/goday-osu-phar.pdf
-Novel and Advanced Techniques for Complex IVC Filter Retrieval,D. Daye,,"{'model': 'tldr@v2.0.0', 'text': 'Standard and advanced IVC filter retrieval techniques including single- access, dual-access, and dissection techniques are reviewed, carrying a non-negligible risk for complications such as filter fragmentation and resultant embolization of filter components, venous pseudoaneurysms or stenoses, and breach of the integrity of the caval wall.'}",
-Teaching Algorithmic Reasoning via In-context Learning,Hattie Zhou,"Large language models (LLMs) have shown increasing in-context learning capabilities through scaling up model and data size. Despite this progress, LLMs are still unable to solve algorithmic reasoning problems. While providing a rationale with the final answer has led to further improvements in multi-step reasoning problems, Anil et al. 2022 showed that even simple algorithmic reasoning tasks such as parity are far from solved. In this work, we identify and study four key stages for successfully teaching algorithmic reasoning to LLMs: (1) formulating algorithms as skills, (2) teaching multiple skills simultaneously (skill accumulation), (3) teaching how to combine skills (skill composition) and (4) teaching how to use skills as tools. We show that it is possible to teach algorithmic reasoning to LLMs via in-context learning, which we refer to as algorithmic prompting. We evaluate our approach on a variety of arithmetic and quantitative reasoning tasks, and demonstrate significant boosts in performance over existing prompting techniques. In particular, for long parity, addition, multiplication and subtraction, we achieve an error reduction of approximately 10x, 9x, 5x and 2x respectively compared to the best available baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that it is possible to teach algorithmic reasoning to LLMs via in-context learning, which it is referred to as algorithmic prompting, and evaluates the approach on a variety of arithmetic and quantitative reasoning tasks, and demonstrates significant boosts in performance.'}",http://arxiv.org/pdf/2211.09066
-Compositional Semantic Parsing with Large Language Models,Andrew Drozdov,"Humans can reason compositionally when presented with new tasks. Previous research shows that appropriate prompting techniques enable large language models (LLMs) to solve artificial compositional generalization tasks such as SCAN. In this work, we identify additional challenges in more realistic semantic parsing tasks with larger vocabulary and refine these prompting techniques to address them. Our best method is based on least-to-most prompting: it decomposes the problem using prompting-based syntactic parsing, then uses this decomposition to select appropriate exemplars and to sequentially generate the semantic parse. This method allows us to set a new state of the art for CFQ while requiring only 1% of the training data used by traditional approaches. Due to the general nature of our approach, we expect similar efforts will lead to new results in other tasks and domains, especially for knowledge-intensive applications.","{'model': 'tldr@v2.0.0', 'text': 'The best method is based on least-to-most prompting: it decomposes the problem using prompting-based syntactic parsing, then uses this decomposition to select appropriate exemplars and to sequentially generate the semantic parse.'}",
-Application of Cognitive Rehabilitation Theory to the Development of Smart Prompting Technologies,,"—Older adults with cognitive impairments often have difficulty performing instrumental activities of daily living (IADLs). Prompting technologies have gained popularity over the last decade and have the potential to assist these individuals with IADLs in order to live independently. Although prompting techniques are routinely used by caregivers and health care providers to aid individuals with cognitive impairment in maintaining their independence with everyday activities, there is no clear consensus or gold standard regarding prompt content, method of instruction, timing of delivery, or interface of prompt delivery in the gerontology or technology literatures. In this paper, we demonstrate how cognitive rehabilitation principles can inform and advance the development of more effective assistive prompting technologies that could be employed in smart environments. We first describe cognitive rehabilitation theory (CRT) and show how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion. We then use the CRT framework to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effective smart prompting technologies. Finally, we raise questions for future exploration as well as challenges and suggestions for future directions in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'Cognitive rehabilitation theory (CRT) is described and it is shown how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion and is used to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effectiveSmart prompting technologies.'}",
-A General Language Assistant as a Laboratory for Alignment,Amanda Askell,"Given the broad capabilities of large language models, it should be possible to work towards a general-purpose, text-based assistant that is aligned with human values, meaning that it is helpful, honest, and harmless. As an initial foray in this direction we study simple baseline techniques and evaluations, such as prompting. We find that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models. Next we investigate scaling trends for several training objectives relevant to alignment, comparing imitation learning, binary discrimination, and ranked preference modeling. We find that ranked preference modeling performs much better than imitation learning, and often scales more favorably with model size. In contrast, binary discrimination typically performs and scales very similarly to imitation learning. Finally we study a `preference model pre-training' stage of training, with the goal of improving sample efficiency when finetuning on human preferences.","{'model': 'tldr@v2.0.0', 'text': 'This work studies simple baseline techniques and evaluations and finds that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models.'}",
-Enabling Conversational Interaction with Mobile UI using Large Language Models,Bryan Wang,"Conversational agents show the promise to allow users to interact with mobile devices using language. However, to perform diverse UI tasks with natural language, developers typically need to create separate datasets and models for each specific task, which is expensive and effort-consuming. Recently, pre-trained large language models (LLMs) have been shown capable of generalizing to various downstream tasks when prompted with a handful of examples from the target task. This paper investigates the feasibility of enabling versatile conversational interactions with mobile UIs using a single LLM. We designed prompting techniques to adapt an LLM to mobile UIs. We experimented with four important modeling tasks that address various scenarios in conversational interaction. Our method achieved competitive performance on these challenging tasks without requiring dedicated datasets and training, offering a lightweight and generalizable approach to enable language-based mobile interaction.","{'model': 'tldr@v2.0.0', 'text': 'This paper designed prompting techniques to adapt an LLM to mobile UIs without requiring dedicated datasets and training, offering a lightweight and generalizable approach to enable language-based mobile interaction.'}",https://dl.acm.org/doi/pdf/10.1145/3544548.3580895
-Recent Advances in Natural Language Processing via Large Pre-trained Language Models: A Survey,Bonan Min,"Large, pre-trained language models (PLMs) such as BERT and GPT have drastically changed the Natural Language Processing (NLP) field. For numerous NLP tasks, approaches leveraging PLMs have achieved state-of-the-art performance. The key idea is to learn a generic, latent representation of language from a generic task once, then share it across disparate NLP tasks. Language modeling serves as the generic task, one with abundant self-supervised text available for extensive training. This article presents the key fundamental concepts of PLM architectures and a comprehensive view of the shift to PLM-driven NLP techniques. It surveys work applying the pre-training then fine-tuning, prompting, and text generation approaches. In addition, it discusses PLM limitations and suggested directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'This article presents the key fundamental concepts of PLM architectures and a comprehensive view of the shift to PLM-driven NLP techniques, and surveys work applying the pre-training then fine-tuning, prompting, and text generation approaches.'}",https://arxiv.org/pdf/2111.01243
-LAMBADA: Backward Chaining for Automated Reasoning in Natural Language,Seyed Mehran Kazemi,"Remarkable progress has been made on automated reasoning with natural text, by using Large Language Models (LLMs) and methods such as Chain-of-Thought prompting and Selection-Inference. These techniques search for proofs in the forward direction from axioms to the conclusion, which suffers from a combinatorial explosion of the search space, and thus high failure rates for problems requiring longer chains of reasoning. The classical automated reasoning literature has shown that reasoning in the backward direction (i.e. from intended conclusion to supporting axioms) is significantly more efficient at proof-finding. Importing this intuition into the LM setting, we develop a Backward Chaining algorithm, called LAMBADA, that decomposes reasoning into four sub-modules, that are simply implemented by few-shot prompted LLM inference. We show that LAMBADA achieves sizable accuracy boosts over state-of-the-art forward reasoning methods on two challenging logical reasoning datasets, particularly when deep and accurate proof chains are required.","{'model': 'tldr@v2.0.0', 'text': 'A Backward Chaining algorithm, called LAMBADA, that decomposes reasoning into four sub-modules, that are simply implemented by few-shot prompted LLM inference, and achieves sizable accuracy boosts over state-of-the-art forward reasoning methods on two challenging logical reasoning datasets.'}",http://arxiv.org/pdf/2212.13894
-Single-molecule junctions beyond electronic transport.,S. Aradhya,,"{'model': 'tldr@v2.0.0', 'text': 'The emerging methods being used to interrogate multiple properties in single molecule-based devices are presented, how these measurements have advanced the understanding of the structure-function relationships in molecular junctions are detailed, and the potential for future research and applications are discussed.'}",
-Are Large Language Models Ready for Healthcare? A Comparative Study on Clinical Language Understanding,Yuqing Wang,"Large language models (LLMs) have made significant progress in various domains, including healthcare. However, the specialized nature of clinical language understanding tasks presents unique challenges and limitations that warrant further investigation. In this study, we conduct a comprehensive evaluation of state-of-the-art LLMs, namely GPT-3.5, GPT-4, and Bard, within the realm of clinical language understanding tasks. These tasks span a diverse range, including named entity recognition, relation extraction, natural language inference, semantic textual similarity, document classification, and question-answering. We also introduce a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand. Our evaluation underscores the significance of task-specific learning strategies and prompting techniques for improving LLMs' effectiveness in healthcare-related tasks. Additionally, our in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP. Our study sheds light on the practical implications of employing LLMs in the specialized domain of healthcare, serving as a foundation for future research and the development of potential applications in healthcare settings.","{'model': 'tldr@v2.0.0', 'text': ""An in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP, and introduces a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand.""}",https://arxiv.org/pdf/2304.05368
-An Empirical Study on the Robustness of the Segment Anything Model (SAM),Yuqing Wang,"The Segment Anything Model (SAM) is a foundation model for general image segmentation. Although it exhibits impressive performance predominantly on natural images, understanding its robustness against various image perturbations and domains is critical for real-world applications where such challenges frequently arise. In this study we conduct a comprehensive robustness investigation of SAM under diverse real-world conditions. Our experiments encompass a wide range of image perturbations. Our experimental results demonstrate that SAM's performance generally declines under perturbed images, with varying degrees of vulnerability across different perturbations. By customizing prompting techniques and leveraging domain knowledge based on the unique characteristics of each dataset, the model's resilience to these perturbations can be enhanced, addressing dataset-specific challenges. This work sheds light on the limitations and strengths of SAM in real-world applications, promoting the development of more robust and versatile image segmentation solutions.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that SAM's performance generally declines under perturbed images, with varying degrees of vulnerability across different perturbations, which sheds light on the limitations and strengths of SAM in real-world applications, promoting the development of more robust and versatile image segmentation solutions.""}",http://arxiv.org/pdf/2305.06422
-Energy-Saving Techniques for Reducing CO 2 Emissions in Elementary Schools,Taehoon Hong,"Global warming has grown into a significantly intense global issue, prompting all nations to exert various efforts to reduce their greenhouse gas emission, according to the United Nations Framework Convention on Climate Change (UNFCCC). The South Korean government is likewise working on reducing its greenhouse gas emission and, as part of its effort, it is currently promoting the Green-School Project. This project aims to create energy-saving schools by renovating deteriorated school buildings. In this study, CO2 emission reduction scenarios are sought by applying energy-saving techniques (ESTs) to educational facilities, and an optimal scenario is created by performing life cycle cost (LCC) analysis for each scenario. To this end, the amounts of energy consumption and CO2 emission reduction in two schools were evaluated using DesignBuilder simulation, based on 15 scenarios with combinations of four ESTs. LCC analysis was also performed. In terms of energy efficiency, CO2 emission reduction, and LCC analy...",,
-Chloroplast-selective gene delivery and expression in planta using chitosan-complexed single-walled carbon nanotube carriers,Seonyeong Kwak,,"{'model': 'tldr@v2.0.0', 'text': 'This nanoparticle-mediated chloroplast transgene delivery tool provides practical advantages over current delivery techniques as a potential transformation method for mature plants to benefit plant bioengineering and biological studies.'}",
-AceCoder: Utilizing Existing Code to Enhance Code Generation,Jia Li,"Large Language Models (LLMs) have shown great success in code generation. LLMs take as the input a prompt and output the code. A key question is how to make prompts (i.e., Prompting Techniques). Existing prompting techniques are designed for natural language generation and have low accuracy in code generation. In this paper, we propose a new prompting technique named AceCoder. Our motivation is that code generation meets two unique challenges (i.e., requirement understanding and code implementation). AceCoder contains two novel mechanisms (i.e., guided code generation and example retrieval) to solve these challenges. (1) Guided code generation asks LLMs first to analyze requirements and output an intermediate preliminary (e.g., test cases). The preliminary is used to clarify requirements and tell LLMs""what to write"". (2) Example retrieval selects similar programs as examples in prompts, which provide lots of relevant content (e.g., algorithms, APIs) and teach LLMs""how to write"". We apply AceCoder to three LLMs (e.g., Codex) and evaluate it on three public benchmarks using the Pass@k. Results show that AceCoder can significantly improve the performance of LLMs on code generation. (1) In terms of Pass@1, AceCoder outperforms the state-of-the-art baseline by up to 56.4% in MBPP, 70.7% in MBJP, and 88.4% in MBJSP. (2) AceCoder is effective in LLMs with different sizes (i.e., 6B to 13B) and different languages (i.e., Python, Java, and JavaScript). (3) Human evaluation shows human developers prefer programs from AceCoder.","{'model': 'tldr@v2.0.0', 'text': 'A new prompting technique named AceCoder, which contains two novel mechanisms, guided code generation and example retrieval, that can significantly improve the performance of LLMs on code generation.'}",
-Direct-conversion radio transceivers for digital communications,A. Abidi,"The current interest in portable wireless communications devices is prompting research into new IC technologies, circuit configurations, and transceiver architectures. Miniature transceivers dissipating low power are sought to sought to communicate digital data. While transistor technology scaling and improved circuit techniques will lead to the inevitable evolutionary advances towards this goal, architectural innovations in the transceiver will lead to revolutionary improvements. It is in this context that there is a resurgence of interest in direct-conversion.",,
-A survey of RDF stores & SPARQL engines for querying knowledge graphs,Waqas Ali,,"{'model': 'tldr@v2.0.0', 'text': 'This survey paper provides a comprehensive review of techniques and systems for querying RDF knowledge graphs, including state-of-the-art storage, indexing and query processing techniques for efficiently evaluating SPARQL queries in a local setting.'}",https://www.techrxiv.org/articles/preprint/Survey_of_RDF_Stores_SPARQL_Engines_for_Querying_Knowledge_Graphs/14376884/files/27462647.pdf
-Application of Cognitive Rehabilitation Theory to the Development of Smart Prompting Technologies,Adriana M Seelye,"Older adults with cognitive impairments often have difficulty performing instrumental activities of daily living (IADLs). Prompting technologies have gained popularity over the last decade and have the potential to assist these individuals with IADLs in order to live independently. Although prompting techniques are routinely used by caregivers and health care providers to aid individuals with cognitive impairment in maintaining their independence with everyday activities, there is no clear consensus or gold standard regarding prompt content, method of instruction, timing of delivery, or interface of prompt delivery in the gerontology or technology literatures. In this paper, we demonstrate how cognitive rehabilitation principles can inform and advance the development of more effective assistive prompting technologies that could be employed in smart environments. We first describe cognitive rehabilitation theory (CRT) and show how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion. We then use the CRT framework to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effective smart prompting technologies. Finally, we raise questions for future exploration as well as challenges and suggestions for future directions in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'Cognitive rehabilitation theory (CRT) is described and it is shown how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion and is used to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effectiveSmart prompting technologies.'}",
-Identifying active ingredients in complex behavioural interventions for obese adults with obesity-related co-morbidities or additional risk factors for co-morbidities: a systematic review,S. Dombrowski,"Abstract Reducing obesity is an important preventive strategy for people who are at increased risk of major disabling or life-threatening conditions. Behavioural treatments for obesity are complex and involve several components aiming to facilitate behaviour change. Systematic reviews need to assess the components that moderate intervention effects. Electronic databases and journals were searched for randomised controlled trials of behavioural interventions targeting dietary and/or physical activity change for obese adults (mean BMI≥30, mean age≥40 years) with risk factors and follow-up data≥12 weeks. A reliable taxonomy of theory-congruent behaviour change techniques (BCTs; Abraham & Michie, 2008) was used to identify programme components. Meta-regression suggested that increasing numbers of identified BCTs are not necessarily associated with better outcomes. The BCTs provision of instructions (β =− 2.69, p=0.02), self-monitoring (β = − 3.37, p<0.001), relapse prevention (β = − 2.63, p=0.02) and prompting practice (β = − 3.63, p<0.001) could be linked to more successful interventions. Studies including more BCTs aimed at dietary change that are congruent with Control Theory were associated with greater weight loss (β = − 1.13, p=0.04). Post-hoc ratings of intervention components in published trials can lead to the identification of components and theories for behaviour change practice and research.","{'model': 'tldr@v2.0.0', 'text': 'Meta-regression suggested that increasing numbers of identified BCTs are not necessarily associated with better outcomes, and post-hoc ratings of intervention components in published trials can lead to the identification of components and theories for behaviour change practice and research.'}",https://aura.abdn.ac.uk/bitstream/2164/2808/1/Dombrowski_2010.pdf
-Teaching Individuals With Developmental Delays: Basic Intervention Techniques,O. Lovaas,"Autism Spectrum DisordersToilet Training SuccessHandbook of Autism and Pervasive Developmental Disorders, Assessment, Interventions, and PolicyEffective Practices for Children with AutismAutism Spectrum DisordersSocial and Communication Development in Autism Spectrum DisordersCompetencies for People Teaching Individuals with Autism and Other Pervasive Developmental DisordersTeaching Early Numeracy to Children with Developmental DisabilitiesThe Use of Video Prompting on the Acquisition, Maintenance, and Generalization of a Line Dance by Adolescents with Autism Spectrum DisordersTreating Childhood Psychopathology and Developmental DisabilitiesHandbook of Early Intervention for Autism Spectrum DisordersTeaching Language to Children with Autism Or Other Developmental DisabilitiesWhat Teachers Need to Know about Students with DisabilitiesA Complete ABA Curriculum for Individuals on the Autism Spectrum with a Developmental Age of 4-7 YearsDerived Relational RespondingImplementing the Group-Based Early Start Denver Model for Preschoolers with AutismAn ABA Curriculum for Children with Autism Spectrum Disorders Aged Approximately 1-4 YearsTeaching Young Children: Choices In Theory And PracticeClinical Assessment and Intervention for Autism Spectrum DisordersTeaching Social Communication to Children with Autism and Other Developmental Delays, Second Edition (2-book set)Teaching ChoicesEvaluation of Behavioural Skills Training with Volunteers Teaching Motor Skills to Individuals with Developmental DisabilitiesAdaptive Behavior Strategies for Individuals with Intellectual and Developmental DisabilitiesHandbook of Research on Technology Tools for Real-World Skill DevelopmentApplied Behavior Analysis for Children with Autism Spectrum DisordersThe Teaching of Self-medication Skills to Individuals with Developmental DisabilitiesEquity and Full Participation for Individuals with Severe DisabilitiesResources for Teaching MindfulnessMaking Inclusion Work for Students with Autism Spectrum DisordersThe Handbook of Intellectual Disability and Clinical Psychology PracticePractitioner's Guide to Evidence-Based PsychotherapyTeaching Functional Skills to Individuals with Developmental Disabilities Using Video PromptingTeaching Individuals with Developmental DelaysControversial Therapies for Developmental DisabilitiesChildren with Developmental DisabilitiesHandbook of Childhood Psychopathology and Developmental Disabilities AssessmentUnderstanding Children with Autism Spectrum DisordersControversial Therapies for Autism and Intellectual DisabilitiesThe Handbook of Language and Speech DisordersJEBPS Vol 17-N2","{'model': 'tldr@v2.0.0', 'text': 'Autism Spectrum DisordersToilet Training SuccessHandbook of Autism and Pervasive Developmental Disorders, Assessment, Interventions, and PolicyEffective Practices for Children with AutismAutistic Spectrum DisordersSocial and Communication Development in Autism Spectrum DisordersCompetencies for People Teaching Individuals with Autism and Other Pervvasive Developmental disorders.'}",
-"Using computer, mobile and wearable technology enhanced interventions to reduce sedentary behaviour: a systematic review and meta-analysis",Aoife Stephenson,,"{'model': 'tldr@v2.0.0', 'text': 'Interventions using computer, mobile and wearable technologies can be effective in reducing SB, and effectiveness appeared most prominent in the short-term and lessened over time.'}",https://ijbnpa.biomedcentral.com/track/pdf/10.1186/s12966-017-0561-4
-Fate of ZnO nanoparticles in soils and cowpea (Vigna unguiculata).,Peng Wang,"The increasing use of zinc oxide nanoparticles (ZnO-NPs) in various commercial products is prompting detailed investigation regarding the fate of these materials in the environment. There is, however, a lack of information comparing the transformation of ZnO-NPs with soluble Zn(2+) in both soils and plants. Synchrotron-based techniques were used to examine the uptake and transformation of Zn in various tissues of cowpea ( Vigna unguiculata (L.) Walp.) exposed to ZnO-NPs or ZnCl2 following growth in either solution or soil culture. In solution culture, soluble Zn (ZnCl2) was more toxic than the ZnO-NPs, although there was substantial accumulation of ZnO-NPs on the root surface. When grown in soil, however, there was no significant difference in plant growth and accumulation or speciation of Zn between soluble Zn and ZnO-NP treatments, indicating that the added ZnO-NPs underwent rapid dissolution following their entry into the soil. This was confirmed by an incubation experiment with two soils, in which ZnO-NPs could not be detected after incubation for 1 h. The speciation of Zn was similar in shoot tissues for both soluble Zn and ZnO-NPs treatments and no upward translocation of ZnO-NPs from roots to shoots was observed in either solution or soil culture. Under the current experimental conditions, the similarity in uptake and toxicity of Zn from ZnO-NPs and soluble Zn in soils indicates that the ZnO-NPs used in this study did not constitute nanospecific risks.","{'model': 'tldr@v2.0.0', 'text': 'Under the current experimental conditions, the similarity in uptake and toxicity of Zn from ZnO-NPs and soluble Zn in soils indicates that the ZnNPs used in this study did not constitute nanospecific risks.'}",http://espace.library.uq.edu.au/view/UQ:320191/UQ320191OA.pdf
-Prompting a consumer behavior for pollution control.,E. Geller,"A field application of behavior modification studied the relative effectiveness of different prompting procedures for increasing the probability that customers entering a grocery store would select their soft drinks in returnable rather than nonreturnable containers. Six different 2-hr experimental conditions during which bottle purchases were recorded were (1) No Prompt (i.e., control), (2) one student gave incoming customers a handbill urging the purchase of soft drinks in returnable bottles, (3) distribution of the handbill by one student and public charting of each customer's bottle purchases by another student, (4) handbill distribution and charting by a five-member group, (5) handbills distributed and purchases charted by three females. The variant prompting techniques were equally effective, and in general increased the percentage of returnable-bottle customers by an average of 25%.","{'model': 'tldr@v2.0.0', 'text': 'A field application of behavior modification studied the relative effectiveness of different prompting procedures for increasing the probability that customers entering a grocery store would select their soft drinks in returnable rather than nonreturnable containers.'}",https://europepmc.org/articles/pmc1310848?pdf=render
-Using Augmented Reality to Help Children with Autism Stay Focused,Lizbeth Escobedo,"Children with autism have difficulty sustaining their selective attention during therapy sessions. Attention management techniques involve the use of verbal and visual prompting, annotated on top of the physical objects used during therapies. Here, the authors explore how augmented reality helps integrate the physical and digital worlds, mimicking current strategies for attention management in autism. They describe their design decisions when developing the Mobile Object Identification System (Mobis), a mobile augmented reality application that lets teachers superimpose digital content on top of physical objects. The results of a five-week deployment study demonstrate that Mobis is useful and easy to use, increases the sustained and selective attention of children with autism, and elicits positive emotions during therapies. This article is part of a special issue on managing attention.","{'model': 'tldr@v2.0.0', 'text': 'How augmented reality helps integrate the physical and digital worlds, mimicking current strategies for attention management in autism, is explored in a special issue on managing attention.'}",
-Tripling the capacity of wireless communications using electromagnetic polarization,M. R. Andrews,,"{'model': 'tldr@v2.0.0', 'text': 'It is shown that, in a scattering environment, an extra factor of three in channel capacity can be obtained, relative to the conventional limit using dual-polarized radio signals.'}",
-When can group level clustering be ignored? Multilevel models versus single-level models with sparse data,P. Clarke,"Objective: The use of multilevel modelling with data from population-based surveys is often limited by the small number of cases per level-2 unit, prompting many researchers to use single-level techniques such as ordinary least squares regression. Design: Monte Carlo simulations are used to investigate the effects of data sparseness on the validity of parameter estimates in two-level versus single-level models. Setting: Both linear and non-linear hierarchical models are simulated in order to examine potential differences in the effects of small group size across continuous and discrete outcomes. Results are then compared with those obtained using disaggregated techniques (ordinary least squares and logistic regression). Main results: At the extremes of data sparseness (two observations per group), the group level variance components are overestimated in the two-level models. But with an average of only five observations per group, valid and reliable estimates of all parameters can be obtained when using a two-level model with either a continuous or a discrete outcome. In contrast, researchers run the risk of Type I error (standard errors biased downwards) when using single-level models even when there are as few as two observations per group on average. Bias is magnified when modelling discrete outcomes. Conclusions: Multilevel models can be reliably estimated with an average of only five observations per group. Disaggregated techniques carry an increased risk of Type I error, even in situations where there is only limited clustering in the data.","{'model': 'tldr@v2.0.0', 'text': 'Multilevel models can be reliably estimated with an average of only five observations per group, valid and reliable estimates of all parameters can be obtained when using a two-level model with either a continuous or a discrete outcome.'}",
-Supplementary Interviewing Techniques to Maximize Output in Free Listing Tasks,D. Brewer,"Free listing is an important ethnographic tool for defining semantic domains. However, when informants free list items from a particular domain, they often do not mention all items they know because they forget items and/or do not understand that they should list exhaustively. In this article, the author reviews results from research on three supplementary interviewing techniques to encourage full responding and enhance recall in such tasks (nonspecific prompting, reading back to the informant the items he or she free listed, and using free-listed items as semantic cues). These methods increase substantially the number of items elicited from individual informants and the number of items in a domain identified from informants in the aggregate. Moreover, these techniques do not require the interviewer to have any prior domain knowledge to be effective.",,
-2012 Midwest Drought in the United States,Ganeshchandra Mallya,"The 2012 North American drought may be the costliest and one of the most widespread natural disasters in U.S. history [USDA Economic Research Service (USDA-ERS) 2012]. While several states across the United States were experiencing drought conditions to varying degrees of severity, the Midwest and Northern Plains were perhaps the most affected. The drought severely impacted agricultural activities across the United States, particularly corn and soybean crops, prompting federal agencies including U.S. Department of Agriculture to declare disaster areas (USDA 2012b) and to provide assistance to those affected by this calamity. This paper utilizes existing and new techniques to provide insights into the severity of the 2012 Midwest drought and its impacts over the region.",,
-Randomized Trial of Supplementary Interviewing Techniques to Enhance Recall of Sexual Partners in Contact Interviews*,D. Brewer,"Background: People with multiple sex partners tend to forget a significant proportion when recalling them. Methods: Randomized trial of supplementary interviewing techniques during routine partner notification contact interviews for chlamydia, gonorrhea, and syphilis in Colorado Springs, CO. Cases with multiple sex partners in the last 3 months (n = 123) participated. Interviewers prompted nonspecifically and read back the list of elicited partners after cases recalled partners on their own. We then randomly assigned cases to receive 1 of 3 sets of recall cues: (1) an experimental set of cues consisting of locations where people meet partners, role relationships, network ties, and first letters of names; (2) another experimental set including common first names; and (3) control cues referring to individual characteristics (e.g., physical appearance). Results: Nonspecific prompting and reading back the list each increased the number of additional partners elicited and located by 3% to 5% on average. On average, the combined location/role/letter/network cues elicited more additional partners (0.57) than did the first-name (0.29) and individual characteristics (0.28) cues. The location and first-name cues were the most effective in eliciting located partners. The supplementary techniques increased the number of new cases found by 12% and, importantly, identified branches of the sexual network that would not otherwise have been discovered. Conclusion: Elicitation of sex partners can be enhanced in contact interviews with simple interviewing techniques, resulting in improved network ascertainment and sexually transmitted disease case finding.","{'model': 'tldr@v2.0.0', 'text': 'Elicitation of sex partners can be enhanced in contact interviews with simple interviewing techniques, resulting in improved network ascertainment and sexually transmitted disease case finding.'}",http://www.interscientific.net/csevcues.pdf
-The useful field of view test: a new technique for evaluating age-related declines in visual function.,K. Ball,"Many older adults eventually experience difficulty in everyday visual activities, prompting them to consult with an eye care specialist. In some cases, a thorough examination may reveal no clinical basis for these difficulties. New techniques for evaluating age-related declines in visual function have substantiated the existence of deficits in visual attention that are not adequately captured by standard clinical measures, and which are predictive of real world problems in driving and mobility. These new techniques may serve an important function in guiding the diagnosis, and treatment of those older adults experiencing functional visual problems.","{'model': 'tldr@v2.0.0', 'text': 'New techniques for evaluating age-related declines in visual function have substantiated the existence of deficits in visual attention that are not adequately captured by standard clinical measures, and which are predictive of real world problems in driving and mobility.'}",
-"Interviewing autistic adults: Adaptations to support recall in police, employment, and healthcare interviews",J. E. Norris,"Recalling specific past experiences is critical for most formal social interactions, including when being interviewed for employment, as a witness or defendant in the criminal justice system, or as a patient during a clinical consultation. Such interviews can be difficult for autistic adults under standard open questioning, yet applied research into effective methods to facilitate autistic adults’ recall is only recently beginning to emerge. The current study tested the efficacy of different prompting techniques to support autistic adults’ recall of specific personal memories; 30 autistic and 30 typically developing adults (intelligence quotients > 85) were asked to recall specific instances from their past, relevant to criminal justice system, healthcare, and employment interviews. Questions comprised ‘open questions’, ‘semantic prompting’ (where semantic knowledge was used to prompt specific episodic retrieval) and ‘visual–verbal prompting’ (a pie-diagram with prompts to recall specific details, for example, who, what, and where). Half the participants received the questions in advance. Consistent with previous research, autistic participants reported memories with reduced specificity. For both groups, visual–verbal prompting support improved specificity and episodic relevance, while semantic prompting also aided recall for employment questions (but not health or criminal justice system). Findings offer new practical insight for interviewers to facilitate communication with typically developing and autistic adults. Lay abstract During many types of interviews (e.g. in employment, with the police, and in healthcare), we need to recall detailed memories of specific events, which can be difficult for autistic people in response to commonly used questions. This is especially because these tend to be open questions (i.e. very broad). Autistic people have disproportionately high rates of physical and mental health conditions, are more likely to interact with police, and are the most underemployed disability group. However, interviewers are often unsure about how to adapt their communication for autistic people. Our research tested whether different types of prompts enabled autistic people to recall specific memories (memories of a single event within one day). Participants were asked about situations relating to witnessing a crime (e.g. at the bank), physical or mental health scenarios and employment interviews (e.g. a time you’ve met a deadline). We tested the following: Open questions: basic questions only (e.g. ‘tell me about a time you went to the cinema’), Semantic prompting: a general prompt (e.g. ‘do you enjoy going to the cinema?’) before asking for a specific instance (‘tell me about a time you went to the cinema?’), Visual–verbal prompting: asking participants to recall when it happened, who was there, the actions that occurred, the setting, and any objects. With visual–verbal prompting, autistic and typically developing participants’ memories were more specific and detailed. Semantic prompting was also effective for employment questions. Our study shows that autistic people can recall specific memories when they are appropriately prompted. Visual–verbal prompting may be effective across different situations.","{'model': 'tldr@v2.0.0', 'text': 'This study shows that autistic people can recall specific memories when they are appropriately prompted, and suggests visual–verbal prompting may be effective across different situations.'}",
-Diagnosis and Treatment of Fetal Arrhythmia,A. Wacker-Gussmann,"Abstract Aims Detection and careful stratification of fetal heart rate (FHR) is extremely important in all pregnancies. The most lethal cardiac rhythm disturbances occur during apparently normal pregnancies where FHR and rhythm are regular and within normal or low-normal ranges. These hidden depolarization and repolarization abnormalities, associated with genetic ion channelopathies cannot be detected by echocardiography, and may be responsible for up to 10% of unexplained fetal demise, prompting a need for newer and better fetal diagnostic techniques. Other manifest fetal arrhythmias such as premature beats, tachycardia, and bradycardia are commonly recognized. Methods Heart rhythm diagnosis in obstetrical practice is usually made by M-mode and pulsed Doppler fetal echocardiography, but not all fetal cardiac time intervals are captured by echocardiographic methods. Results and Conclusions This article reviews different types of fetal arrhythmias, their presentation and treatment strategies, and gives an overview of the present and future diagnostic techniques.","{'model': 'tldr@v2.0.0', 'text': 'This article reviews different types of fetal arrhythmias, their presentation and treatment strategies, and gives an overview of the present and future diagnostic techniques.'}",http://www.thieme-connect.de/products/ejournals/pdf/10.1055/s-0034-1372430.pdf
-An Analysis of Prompting Procedures for Teaching a Sight Vocabulary1,O. Duell,"""Prompting"" is one technique used to facilitate learning in selfinstructional programs. Depending upon the context, a prompt may consist, for instance, of the printed form of a word the subject (S) is to vocalize, a rule which can be applied to an example, a hint to help in the solution of a problem, the first letter of an answer, a synonym for an answer, or a picture to elicit the vocalization of a word. Prompting can be defined generally as the pairing of a cue, the stimulus that is to control a response when training is completed, with a prompt, a stimulus which already controls or partially controls the desired response. The object of training sequences using prompts is to shift the control of the response from the prompt to the cue with which the prompt is paired. Although there have been some notable successes with prompted training sequences, there also have been anomalies. There is still much to be learned about conditions under which prompting techniques can be used to advantage (Anderson, 1967). Recent studies (Anderson and Faust, 1967; Faust and Anderson, 1967) have shown that a prompted sequence teaches more if it forces the student to notice the cues while he responds during training. College students learned more Russian vocabulary from",,
-One-Class Classification-Based Real-Time Activity Error Detection in Smart Homes,Barnan Das,"Caring for individuals with dementia is frequently associated with extreme physical and emotional stress, which often leads to depression. Smart home technology and advances in machine learning techniques can provide innovative solutions to reduce caregiver burden. One key service that caregivers provide is prompting individuals with memory limitations to initiate and complete daily activities. We hypothesize that sensor technologies combined with machine learning techniques can automate the process of providing reminder-based interventions. The first step toward automated interventions is to detect when an individual faces difficulty with activities. We propose machine learning approaches based on one-class classification that learn normal activity patterns. When we apply these classifiers to activity patterns that were not seen before, the classifiers are able to detect activity errors, which represent potential prompt situations. We validate our approaches on smart home sensor data obtained from older adult participants, some of whom faced difficulties performing routine activities and thus committed errors.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesize that sensor technologies combined with machine learning techniques can automate the process of providing reminder-based interventions, and proposes machine learning approaches based on one-class classification that learn normal activity patterns.'}",
-"""Expertness"" from Structured Text? RECONSIDER: A Diagnostic Prompting Program",M. Tuttle,"RECONSIDER is an interactive diagnostic prompting program which uses simple information retrieval techniques to prompt a physician regarding possible diagnoses, given a list of positive patient findings. Its knowledge base consists of ""structured text"" definitions of 3262 diseases and a synonym dictionary Patient findings, and their synonyms, are matched against inverted files of terms from the disease descriptions, the number and selectivity of the patient findings matching terms in a given disease description determine that disease's ""score"", and the matched diseases are sorted on this score to form a preliminary differential diagnosis. Definitions of diseases can be referenced for viewing by name, or by their position in a differential. While its first formal evaluation is not yet complete, the performance of RECONSIDER continues to exceed the expectations of user and designer alike.","{'model': 'tldr@v2.0.0', 'text': 'While its first formal evaluation is not yet complete, the performance of RECONSIDER continues to exceed the expectations of user and designer alike.'}",
-"Óleos essenciais no Brasil: aspectos gerais, desenvolvimento e perspectivas",H. Bizzo,"Essential oils are extracted by steam distillation of plants or cold pressing of citrus fruit pericarp. They are used in food, cosmetic, personal care and pharmaceutical industries. In Brazil, oils from orange and related products contribute to near 97% to the positive commercial performance of the sector. Predatory exploitation and the availability of new sources of raw materials, with more attracting prices, changed the paradigm. Prospective studies, sustainable use of Brazilian biodiversity, domestication of exotic species with commercial relevance, the use of breeding techniques and the development of new applications for essential oils are thematic lines, usually multidisciplinary, which have been prompting the expansion of the research on essential oils. This paper presents an analysis on essential oils balance trade from 2005 to 2008 and some historical data on research and production of essential oils in Brazil.",,https://www.scielo.br/j/qn/a/QwJBsdNzGmZSq4jKmhwVDnJ/?lang=pt&format=pdf
-The conceptual structure of research utilization.,C. Estabrooks,"Despite the fact that the nursing literature is replete with calls to make the practice of nursing research based, little is known about the structure and function of research utilization. The purpose of this study was to explore the conceptual structure of research utilization. Data were collected from a randomly selected sample of 600 registered nurses practicing in western Canada. Using the techniques of structural equation modeling (with LISREL), competing models representing conceptual structures of research utilization were developed and evaluated. In the first model, a simplex style of model, the investigator proposed that a nurse's early responses would influence subsequent responses to the question measuring research utilization, implying a time ordered causal sequence. In the second style of model, a common cause (or factor-like) model, the investigator proposed a stable underlying concept, research utilization, that was relatively insensitive to prompting and time ordering. The simplex style of model failed to reach acceptable indices of fit. The common cause model fit the data well, suggesting that instrumental, conceptual, and persuasive research utilization exist and that a global measure of research utilization may be defensible.","{'model': 'tldr@v2.0.0', 'text': 'Two competing models representing conceptual structures of research utilization were developed and evaluated, suggesting that instrumental, conceptual, and persuasive research utilization exist and that a global measure ofResearch utilization may be defensible.'}",
-Realization of Techniques in Problem Solving: The Construction of Bijections for Enumeration Tasks,J. Mamona-Downs,,,
-THE EFFECTS OF VICARIOUS PROMPTING ON ATTENTIVE BEHAVIOR OF CHILDREN WITH BEHAVIOR DISORDERS,J. Birnbrauer,"Vicarious prompting, reinforcing and labelling the appropriate behavior of a nontarget child contingent on inappropriate behavior of a target child, was assessed in two classrooms of children with behavior disorders. Vicarious prompts using social reinforcement alone were clearly effective in increasing the attentive behavior of one target child in Experiment 1. The children in Experiment 2, however, did not appera to be responsive to vicarious prompting. Comparison of the two groups suggests that vicarious prompting, a relatively subtle, indirect control technique, may be a low cost, positive, and effective alternative to other techniques under conditions in which the behavior of most of the group is attentive.",,
-A Comparison of Behavioral Techniques to Teach Functional Independent-Living Skills to Individuals with Severe and Profound Mental Retardation,J. Matson,"The efficacy of two treatment approaches was compared for functional skill acquisition in individuals with severe and profound mental retardation. Participants included 22 residents from a large developmental center (Pinecrest) in central Louisiana. Treatment including staff training, feedback, and edible reinforcement in addition to prompting, modeling, and physical guidance was more effective than prompting, modeling, and physical guidance alone. Additionally, daily documentation of teaching did not enhance treatment effectiveness. Implications of the findings are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Treatment including staff training, feedback, and edible reinforcement in addition to prompting, modeled, and physical guidance alone was more effective than prompting, modeling, andPhysical guidance alone.'}",
-Addressing Data Sparseness in Contextual Population Research,P. Clarke,"The use of multilevel modeling with data from population-based surveys is often limited by the small number of cases per Level 2 unit, prompting a recent trend in the neighborhood literature to apply cluster techniques to address the problem of data sparseness. In this study, the authors use Monte Carlo simulations to investigate the effects of marginal group sizes on multilevel model performance, bias, and efficiency. They then employ cluster analysis techniques to minimize data sparseness and examine the consequences in the simulations. They find that estimates of the fixed effects are robust at the extremes of data sparseness, while cluster analysis is an effective strategy to increase group size and prevent the overestimation of variance components. However, researchers should be cautious about the degree to which they use such clustering techniques due to the introduction of artificial within-group heterogeneity.","{'model': 'tldr@v2.0.0', 'text': 'This study uses Monte Carlo simulations to investigate the effects of marginal group sizes on multilevel model performance, bias, and efficiency, and employs cluster analysis techniques to minimize data sparseness and examines the consequences in the simulations.'}",
-Changes over time and transfer of analogy-problem solving of gifted and non-gifted children in a dynamic testing setting,B. Vogelaar,"Abstract This study examined differences in transfer of analogical reasoning after analogy-problem solving between 40 gifted and 95 average-ability children (aged 9–10 years old), utilising dynamic testing principles. This approach was used in order to examine potential differences between gifted and average-ability children in relation to progression after training, and with regard to the question whether training children in analogy problem-solving elicits transfer of analogical reasoning skills to an analogy construction-task. Children were allocated to one of two experimental conditions: either children received unguided practice in analogy problem-solving, or they were provided with this in addition to training incorporating graduated prompting techniques. The results showed that gifted and average-ability children who were trained made more progress in analogy problem-solving than their peers who received unguided practice experiences only. Gifted and average-ability children were found to show similar progression in analogy problem-solving, and gifted children did not appear to have an advantage in the analogy-construction transfer task. The dynamic training seemed to bring about no additional improvement on the transfer task over that of unguided practice experiences only.",,https://www.tandfonline.com/doi/pdf/10.1080/01443410.2017.1409886?needAccess=true
-"Behavior Therapy Techniques Applied to Eating, Exercise, and Diet Modification in Childhood Obesity",J. Varni,"Behavior therapy techniques are receiving increasing attention in the treatment of childhood obesity. This paper provides a selective review of experimental studies of childhood obesity that have utilized behavior therapy techniques with preadolescent populations. It further defines and describes specific behavior therapy techniques utilized in the modification of eating, exercise, and diet patterns in childhood obesity, including stimulus control, stimulus and response generalization, prompting, positive reinforcement, shaping, contingency contracting, and generalization and maintenance strategies.","{'model': 'tldr@v2.0.0', 'text': 'Specific behavior therapy techniques utilized in the modification of eating, exercise, and diet patterns in childhood obesity, including stimulus control, stimulus and response generalization, prompting, positive reinforcement, shaping, contingency contracting, and generalization and maintenance strategies are described.'}",
-Combined Home and School Obesity Prevention Interventions for Children,G. Hendrie,"This review identifies studies describing interventions delivered across both the home and school/community setting, which target obesity and weight-related nutrition and physical activity behaviors in children. Fifteen studies, published between 1998 and 2010, were included and evaluated for effectiveness, study quality, nutrition/activity content, behavior change techniques, and theoretical basis, using validated assessment tools/taxonomies. Seven studies were rated as effective. Behavior change techniques used to engage families, and techniques associated with intervention effectiveness were coded. Effective studies used about 10 behavior change techniques, compared with 6.5 in ineffective studies. Effective interventions used techniques including providing general information on behavior–health links, prompting practice of behavior, and planning for social support/social changes. Different behavior change techniques were applied in the home and school setting. The findings of this review provide novel insights into the techniques associated with intervention effectiveness that can inform the development of public health obesity prevention strategies.","{'model': 'tldr@v2.0.0', 'text': 'This review identifies studies describing interventions delivered across both the home and school/community setting, which target obesity and weight-related nutrition and physical activity behaviors in children, and provides novel insights into the techniques associated with intervention effectiveness.'}",
-Operational risk management,I. Moosa,,,
-Harnessing Repetitive Behaviours to Engage Attention and Learning in a Novel Therapy for Autism: An Exploratory Analysis,G. Chen,"Rigorous, quantitative examination of therapeutic techniques anecdotally reported to have been successful in people with autism who lack communicative speech will help guide basic science toward a more complete characterisation of the cognitive profile in this underserved subpopulation, and show the extent to which theories and results developed with the high-functioning subpopulation may apply. This study examines a novel therapy, the “Rapid Prompting Method” (RPM). RPM is a parent-developed communicative and educational therapy for persons with autism who do not speak or who have difficulty using speech communicatively. The technique aims to develop a means of interactive learning by pointing amongst multiple-choice options presented at different locations in space, with the aid of sensory “prompts” which evoke a response without cueing any specific response option. The prompts are meant to draw and to maintain attention to the communicative task – making the communicative and educational content coincident with the most physically salient, attention-capturing stimulus – and to extinguish the sensory–motor preoccupations with which the prompts compete. Video-recorded RPM sessions with nine autistic children ages 8–14 years who lacked functional communicative speech were coded for behaviours of interest. An analysis controlled for age indicates that exposure to the claimed therapy appears to support a decrease in repetitive behaviours and an increase in the number of multiple-choice response options without any decrease in successful responding. Direct gaze is not related to successful responding, suggesting that direct gaze might not be any advantage for this population and need not in all cases be a precondition to communication therapies.","{'model': 'tldr@v2.0.0', 'text': 'Exposure to the claimed therapy appears to support a decrease in repetitive behaviours and an increase in the number of multiple-choice response options without any decrease in successful responding, suggesting that direct gaze might not be any advantage for this population and need not be a precondition to communication therapies.'}",
-Errorless Learning: Review and Practical Application for Teaching Children with Pervasive Developmental Disorders.,Michael M. Mueller,"Errorless learning refers to a variety of discrimination learning techniques that eliminate or minimize responding to incorrect choices. This article describes experimental roots of errorless learning and applied errorless strategies. Specifically, previous research on stimulus fading, stimulus shaping, response prevention, delayed prompting, superimposition with stimulus fading, and superimposition with stimulus shaping are discussed. Educational applications for children with Pervasive Developmental Disorders (PDD) are presented for each technique so that school psychologists, educators, and teachers working with children with PDD can understand the underpinnings and practical applications of errorless techniques to use in skill acquisition programming in school settings. © 2007 Wiley Periodicals, Inc. Psychol Schs 44: 691–700, 2007.",,
-A Palmtop Computer Program for the Treatment of Generalized Anxiety Disorder,M. Newman,"This is the first report of a palmtop computer program developed to increase the efficiency and cost-effectiveness of cognitive behavioral therapy for generalized anxiety disorder (GAD). The computer program offers advantages to researchers, therapists, and clients. These advantages include continuous, unobtrusive collection of process data on treatment adherence as well as on the impact of cognitive behavioral therapy techniques in the client’s natural setting. In addition, the computer extends treatment beyond the therapy hour and motivates clients to comply with homework assignments by prompting practice of cognitive behavioral strategies. The successful application of the palmtop computer program reported in this integrated series suggests a new line of research directed toward increasing the cost-effectiveness of what is currently the goldstandard treatment for GAD.","{'model': 'tldr@v2.0.0', 'text': 'The successful application of the palmtop computer program reported in this integrated series suggests a new line of research directed toward increasing the cost-effectiveness of what is currently the goldstandard treatment for GAD.'}",
-Imaging Drosophila gene activation and polymerase pausing in vivo,J. Lis,,"{'model': 'tldr@v2.0.0', 'text': 'Live-cell imaging techniques coupled with protein–DNA crosslinking assays on a genome-wide scale allow more detailed mechanistic questions to be addressed and are prompting the re-evaluation of models of transcription regulation in both Drosophila and mammals.'}",https://www.nature.com/articles/nature06324.pdf
-PENGARUH PENERAPAN PENDEKATAN SCAFFOLDING TERHADAP KEMAMPUAN PEMAHAMAN KONSEP MATEMATIKA SISWA KELAS VIII SMP PERTIWI 2 PADANG,Nicke Septriani,"The ability of students to understand mathematical concepts are expected to grow and thrive in learning mathematics. But mathematic lesson that happen in class VIII SMP Pertiwi 2 Padang, understanding of mathematics concept student still lower. During the learning process of students tend to memorize a given concept. This resulted in the ability of student understanding of mathematical concept is low. Scaffolding approach to probing-prompting techniques are expected to facilitate the students in an improved understanding of the concept. This research aimed to see to whether students' understanding of mathematical concepts are learned with scaffolding approach is better than the understanding of mathematical concepts that students learn using conventional learning in class VIII SMP Pertiwi 2 Padang. This Type Research is experiment kuasi with device research of Randomized Control Group Only Design. Pursuant to result of data analysis can be concluded by the understanding of mathematics concept after applied Scaffolding approach better than the understanding of mathematics concept after applying of conventional study student of c lass VIII SMP Pertiwi 2 Padang .","{'model': 'tldr@v2.0.0', 'text': ""This research aimed to see to whether students' understanding of mathematical concepts are learned with scaffolding approach is better than the understanding of mathematics concepts that students learn using conventional learning in class VIII SMP Pertiwi 2 Padang.""}",
-"""I can speak for myself"": involving individuals with intellectual disabilities as research participants.",Jennifer Mactavish,"The Lifespan and Disability Project, a 2-year qualitative study, was designed to enhance understanding of social integration by including the perspectives of individuals with intellectual disabilities. Procedures and strategies employed to involve these individuals in the project were documented. Specifically, we describe the use of individual interviews and focus groups (e.g., purpose, rationale, facilitation techniques) to collect data, and verification meetings in which a prompting hierarchy and cueing were employed to assist participants in confirming or disconfirming the researchers' interpretations of the emerging findings. Highlights of the results were presented to illustrate the efficacy of the strategies in enabling the participants with intellectual disabilities to share their perspectives on social integration.","{'model': 'tldr@v2.0.0', 'text': ""The use of individual interviews and focus groups to collect data, and verification meetings in which a prompting hierarchy and cueing were employed to assist participants in confirming or disconfirming the researchers' interpretations of the emerging findings are described.""}",
-Using Hidden Markov Models to Characterize Student Behaviors in Learning-by-Teaching Environments,Hogyeong Jeong,,"{'model': 'tldr@v2.0.0', 'text': ""Evidence is presented that HMMs can be used to effectively determine students' pattern of activities and clear differences between different interventions, and links between students learning performance and their interactions with the system are indicated.""}",
-SYNGRAPH: A graphical user interface generator,D. Olsen,"The SYNGRAPH system automatically generates graphical user interfaces. It generates interactive Pascal programs from a description of the input language's grammar. From the grammar it deduces information about how to manage both physical and simulated devices, and how prompting and echoing are performed. Input errors are detected, and can be corrected using automatically provided rubout and cancel features. The natural integration of application specific semantics into the system is also shown including appropriate semantic recovery from input errors.","{'model': 'tldr@v2.0.0', 'text': ""The SYNGRAPH system automatically generates graphical user interfaces from a description of the input language's grammar, and the natural integration of application specific semantics into the system is shown including appropriate semantic recovery from input errors.""}",
-Programming the generalization of a greeting response in four retarded children.,T. Stokes,"Reinforcement techniques of prompting and shaping were employed to develop handwaving, a useful social greeting response, in four institutionalized retarded subjects. A multiple-baseline design across subjects demonstrated the reliable functioning of the training procedures. Specifically, it showed that training and maintenance of the greeting response by one experimenter was not usually sufficient for generalization of the response to the more than 20 other members of the institution staff who had not participated in the training of the response. However, high levels of generalization to staff members were recorded for three subjects over periods ranging from one to six months after a second experimenter trained and maintained the response in conjunction with the first experimenter. The fourth subject, although never receiving training by a second experimenter, showed similar results following a second training by the first experimenter.","{'model': 'tldr@v2.0.0', 'text': 'Reinforcement techniques of prompting and shaping were employed to develop handwaving, a useful social greeting response, in four institutionalized retarded subjects and showed that training and maintenance of the greeting response by one experimenter was not usually sufficient for generalization of the response to the more than 20 other members of the institution staff who had not participated in the training of theresponse.'}",https://europepmc.org/articles/pmc1311674?pdf=render
-What counts in brain aging? Design-based stereological analysis of cell number.,Jeffrey M. Long,"The advent and implementation of new design-based stereological techniques allows the quantification of cell number without the assumptions required when obtaining areal densities. These new techniques are rapidly becoming the standard for quantifying cell number, particularly in aging studies. Recently, studies using stereological techniques have failed to confirm earlier findings regarding age-associated neural loss. This newly emerging view of retained cell number during aging is having a major impact on biogerontology, prompting revaluation of long-standing hypotheses of age-related cell loss as causal for age-related impairments in brain functioning. Rather than focus on neuronal loss as the end-result of a negative cascade of neuronal injury, research has begun to consider that age-related behavioral declines may reflect neuronal dysfunction (e.g., synaptic or receptor loss, signal transduction deficits) instead of neuronal death. Here we discuss design-based stereology in the context of age-related change in brain cell number and its impact on consideration of structural change in brain aging. Emergence of this method of morphometrics, however, can have relevance to many areas of gerontological research.","{'model': 'tldr@v2.0.0', 'text': 'The advent and implementation of new design-based stereological techniques allows the quantification of cell number without the assumptions required when obtaining areal densities, and their impact on consideration of structural change in brain aging is discussed.'}",https://academic.oup.com/biomedgerontology/article-pdf/54/10/B407/1611374/10568523.pdf
-The detection of interturn stator faults in doubly-fed induction generators,H. Douglas,"Presently, many condition monitoring techniques that are based on steady-state analysis are being applied to wind generators. However, the operation of wind generators is predominantly transient, therefore prompting the development of non-stationary techniques for fault detection. In this paper we apply steady-state techniques, e.g. motor current signatures analysis (MCSA) and the extended Park's vector approach (EPVA), as well as a new transient technique that is a combination of the EPVA, the discrete wavelet transform and statistics, to the detection of turn faults in a doubly-fed induction generators (DFIG). It is shown that steady-state techniques are not effective when applied to DFIG's operating under transient conditions. The new technique shows that stator turn faults can be unambiguously detected under transient conditions.",,
-‘Projective transidentification’: An extension of the concept of projective identification,J. Grotstein,"Questions about the concept of projective identifi cation still persist. The author presents the following hypotheses: Klein's traditional view and Bion's extension and revision of it can be thought of as occupying a continuum in reverse. He postulates that Bion's concept of communicative intersubjective projective identifi cation (which the author renames ‘projective transidentifi cation’) is primary and inclusive of Klein's earlier unconscious, omnipotent, intrapsychic mode but includes Bion's ‘realistic’ communicative mode as well. The author hypothesizes, consequently, that intersubjective projective identifi cation constitutes both the operation of an unconscious phantasy of omnipotent intrapsychic projective identifi cation solely within the internal world of the projecting subject‐in addition to two other processes: conscious and/or preconscious modes of sensorimotor induction, which would include signaling and/or evocation or prompting gestures or techniques (mental, physical, verbal, posturing or priming) on the part of the projecting subject; followed by spontaneous empathic simulation in the receptive object of the subject's experience in which the receptive object is already inherently ‘hard‐wired’ to be empathic with the prompting subject.","{'model': 'tldr@v2.0.0', 'text': 'The author hypothesizes that intersubjective projective identifi cation constitutes both the operation of an unconscious phantasy of omnipotent intrapsychic projective identification solely within the internal world of the projecting subject and two other processes: conscious and/or preconscious modes of sensorimotor induction.'}",
-Increasing Behaviors of Severely Retarded and Autistic Persons,R. Foxx,"This widely used manual shows how to increase desirable behaviors by using techniques such as shaping, prompting, fading, modeling, backward chaining, and graduated guidance. It offers specific guidelines for arranging and managing the learning environment as well as standards for evaluating and maintaining success. Exercises, review questions, and numerous examples are included. The book is written for special educators, aides, residential staff, and those responsible for designing or evaluating behavioral programs. Often adopted as a supplementary college text.",,
-A behavioral science approach to transportation safety.,E. Geller,"This paper reviews a variety of techniques that can increase the large-scale use of vehicle safety belts. Activator approaches or behavior change strategies that precede opportunities to buckle up included: reminder techniques (i.e., buckle-up dashborad stickers, sound-light reminder systems in vehicles, and special ""Airline Lifesaver"" and ""Flash for Life"" prompting procedures), role modeling by television stars, educational tactics involving young school children, commitment pledgecard signing, incentives, and disincentives. Reward strategies can be direct and immediate, direct and delayed, or indirect. These three techniques for delivering rewards have been equivalently effective at increasing safety belt use in corporate and community settings. The conclusion is that a number of strategies are cost effective for increasing safety belt use on a large scale, and no single approach is sufficient.","{'model': 'tldr@v2.0.0', 'text': 'The conclusion is that a number of strategies are cost effective for increasing safety belt use on a large scale, and no single approach is sufficient.'}",
-MAPS: Dynamic Scaffolding for Independence for Persons with Cognitive Impairments,Stefan Carmien,,"{'model': 'tldr@v2.0.0', 'text': None}",
-A modified speech tracking procedure as a communicative performance measure.,M. Matthies,"A modified speech tracking procedure was investigated to determine the effects of controlling connected discourse materials and of specific prompting techniques on tracking rate and subjects' responses. In addition, comprehension was tested after subjects completed the tracking task. Two experiments were conducted with normally hearing subjects, one in which subjects tracked in an auditory plus noise condition and the second in which they tracked in a visual-alone condition. The results of these two experiments showed significant differences between tracking rates for controlled versus uncontrolled materials for both modalities, with higher tracking rates observed for the controlled materials. The use of specific prompting techniques reduced intertester variability in both modalities. Further, results indicated that comprehension ability and tracking skill are not perfectly related.","{'model': 'tldr@v2.0.0', 'text': 'Results indicated that comprehension ability and tracking skill are not perfectly related, with significant differences between tracking rates for controlled versus uncontrolled materials for both modalities.'}",
-Overprompting Science Students Using Adjunct Study Questions.,W. Holliday,"Overprompting students by providing them with strong hints to answers of questions can do learners more instructional harm than good. The selective attention model was used to explain the effects of overprompting students provided with study questions adjunct to a complex flow diagram describing scientific cyclical schema. Tenth-grade students were randomly assigned to an unprompted-question, no-question, prompted-question, and a placebo control group. Analysis showed that strongly prompting students to the answers of such questions was less effective than an unprompted question treatment, p < 0.05. The no-question treatment did not significantly outperform the prompted treatment. The information presented in the flow diagram was operationally related to comprehension study and posttest questions. The theoretical discussion and the present findings suggested that science teachers should use prompting techniques with extreme caution.",,
-Conservation of Electricity on a College Campus,E. Zolik,"This study investigated the extent to which posters requesting that lights be turned off in classrooms would increase energy conserving behaviors. A multiple baseline design was used with two groups of eleven classrooms. Results indicated that with onset of the intervention phase, moderate increases in energy conserving behaviors occurred in both clusters of classrooms. These findings indicate that prompting techniques are an effective strategy for getting lights turned off in classrooms. The conservation of electrical energy can be approached from two different perspectives: increasing the efficiency of equipment which is operated by electricity and the installation of automated equipment to eliminate waste on the one hand, and the modification of behavior patterns of people with respect to energy consumption. For maximum effectiveness, both approaches — the technical and the behavioral — have to be combined. While solely informational procedures have had only negligible effects on energy saving [1], incentive payments reducing energy use [2], and regular feedback seem to be strong behavioral procedures for inducing energy conservation [3]. One area where different strategies have been infrequently implemented is in college classrooms. Winett, in a study of conservation in a classroom, reported that a large two foot by three foot sign was effective in reducing the per cent of lights turned off * Poster presented at the Association for Behavior Analysis, Milwaukee, May 1982.",,
-Preparing Severely and Profoundly Mentally Retarded Adults for Tests of Motor Fitness,P. Tomporowski,"The effectiveness of using behavior shaping techniques to prepare nine institutionalized severely and profoundly mentally retarded adults for tests of motor fitness was assessed. Three severely mentally retarded subjects served as the executive subjects in a yoked-control study. Each executive subject was paired randomly with two profoundly mentally retarded subjects. A three-phase multiple baseline experiment was then conducted with executive subjects; yoked subjects received the same intervention as their executive partner. The training program consisted of a verbal instruction baseline phase, a modeling phase, and a graduated guidance phase. Tasks used for training were the shuttle run and standing long jump. Modeling instruction produced minimal improvement in subject performance; however, graduated guidance instruction produced marked, immediate improvement in the performance of six of nine subjects on both tasks. These results suggest that physical prompting techniques will prepare some, but not all...",,
-Structured Chain-of-Thought Prompting for Code Generation,Jia Li,"Large Language Models (LLMs) (e.g., ChatGPT) have shown impressive performance in code generation. LLMs take prompts as inputs, and Chain-of-Thought (CoT) prompting is the state-of-the-art prompting technique. CoT prompting asks LLMs first to generate CoTs (i.e., intermediate natural language reasoning steps) and then output the code. However, CoT prompting is designed for natural language generation and has low accuracy in code generation. In this paper, we propose Structured CoTs (SCoTs) and present a novel prompting technique for code generation, named SCoT prompting. Our motivation is source code contains rich structural information and any code can be composed of three program structures (i.e., sequence, branch, and loop structures). Intuitively, structured intermediate reasoning steps make for structured source code. Thus, we ask LLMs to use program structures to build CoTs, obtaining SCoTs. Then, LLMs generate the final code based on SCoTs. Compared to CoT prompting, SCoT prompting explicitly constrains LLMs to think about how to solve requirements from the view of source code and further the performance of LLMs in code generation. We apply SCoT prompting to two LLMs (i.e., ChatGPT and Codex) and evaluate it on three benchmarks (i.e., HumanEval, MBPP, and MBCPP). (1) SCoT prompting outperforms the state-of-the-art baseline - CoT prompting by up to 13.79% in Pass@1. (2) Human evaluation shows human developers prefer programs from SCoT prompting. (3) SCoT prompting is robust to examples and achieves substantial improvements.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Structured CoTs (SCoTs) and presents a novel prompting technique for code generation, named SCoT prompting, which explicitly constrains LLMs to think about how to solve requirements from the view of source code and further the performance of LLMs in code generation.'}",
-Fine-Grained Visual Prompting,Lingfeng Yang,"Vision-Language Models (VLMs), such as CLIP, have demonstrated impressive zero-shot transfer capabilities in image-level visual perception. However, these models have shown limited performance in instance-level tasks that demand precise localization and recognition. Previous works have suggested that incorporating visual prompts, such as colorful boxes or circles, can improve the ability of models to recognize objects of interest. Nonetheless, compared to language prompting, visual prompting designs are rarely explored. Existing approaches, which employ coarse visual cues such as colorful boxes or circles, often result in sub-optimal performance due to the inclusion of irrelevant and noisy pixels. In this paper, we carefully study the visual prompting designs by exploring more fine-grained markings, such as segmentation masks and their variations. In addition, we introduce a new zero-shot framework that leverages pixel-level annotations acquired from a generalist segmentation model for fine-grained visual prompting. Consequently, our investigation reveals that a straightforward application of blur outside the target mask, referred to as the Blur Reverse Mask, exhibits exceptional effectiveness. This proposed prompting strategy leverages the precise mask annotations to reduce focus on weakly related regions while retaining spatial coherence between the target and the surrounding background. Our Fine-Grained Visual Prompting (FGVP) demonstrates superior performance in zero-shot comprehension of referring expressions on the RefCOCO, RefCOCO+, and RefCOCOg benchmarks. It outperforms prior methods by an average margin of 3.0% to 4.6%, with a maximum improvement of 12.5% on the RefCOCO+ testA subset. The part detection experiments conducted on the PACO dataset further validate the preponderance of FGVP over existing visual prompting techniques. Code and models will be made available.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a new zero-shot framework that leverages pixel-level annotations acquired from a generalist segmentation model for fine-grained visual prompting and reveals that a straightforward application of blur outside the target mask, referred to as the Blur Reverse Mask, exhibits exceptional effectiveness.'}",http://arxiv.org/pdf/2306.04356
-Monte Carlo Techniques in Radiation Therapy,J. Seco,"Monte Carlo Fundamentals History of Monte Carlo Alex F. Bielajew Basics of Monte Carlo Simulations Matthias Fippel Variance Reduction Techniques Matthias Fippel Application of Monte Carlo Techniques in Radiation Therapy Applications of Monte Carlo to Radiation Dosimetry Hugo Bouchard and Jan Seuntjens Monte Carlo Modeling of External Photon Beams in Radiotherapy Frank Verhaegen Monte Carlo Modeling of External Electron Beams in Radiotherapy Frank Verhaegen Dynamic Beam Delivery and 4D Monte Carlo Emily Heath and Joao Seco Patient Dose Calculation Joao Seco and Maggy Fragoso Monte Carlo Methods and Applications for Brachytherapy Dosimetry and Treatment Planning Guillaume Landry, Mark R. Rivard, Jeffrey F. Williamson, and Frank Verhaegen Monte Carlo as a QA Tool for Advanced Radiation Therapy JinSheng Li and C.-M. Charlie Ma Electrons: Clinical Considerations and Applications Joanna E. Cygler and George X. Ding Photons: Clinical Considerations and Applications Michael K. Fix Monte Carlo Calculations for Proton and Ion Beam Dosimetry Hugo Palmans Protons: Clinical Considerations and Applications Harald Paganetti Application of Monte Carlo Methods to Radionuclide Therapy Michael Ljungberg Monte Carlo for Kilovoltage and Megavoltage Imaging George X. Ding and Andrew Fielding Monte Carlo Calculations for PET-Based Treatment Verification of Ion Beam Therapy Katia Parodi Monte Carlo Studies of Prompt Gamma Emission and of Proton Radiography/Proton-CT Joao Seco, Jerimy C. Polf, and Katia Parodi Monte Carlo for Treatment Device Design Bruce A. Faddegon GPU-Based Fast Monte Carlo Simulation for Radiotherapy Dose Calculation Xun Jia, Sami Hissoiny, and Steve B. Jiang Monte Carlo for Shielding of Radiotherapy Facilities Peter J. Biggs and Stephen F. Kry Index","{'model': 'tldr@v2.0.0', 'text': 'The Monte Carlo Fundamentals History of Monte Carlo and Monte Carlo as a QA Tool for Advanced Radiation Therapy and Applications are presented.'}",
-The Effect of Probing – Prompting Technique on Reading Achievement,Rizqi Nadia Putri,"The fact that most of the students in Junior High School are lack of vocabulary, lazy in reading a long text and low motivation in reading textbooks can be solved by choosing an appropriate method in teaching reading. The aimed of this research was to find out whether the probing-prompting technique was effective for the students' reading achievement. The word “Probing” means research and explore, meanwhile “prompting” means push and guide. In probing-prompting, students are required to answer the question by their own words or sentences. Two groups design was applied, and the participants were eighth students of Junior High School in Tegal. We used reading testing in collecting the data, pre and post-test were conducted. And the data analysis calculated using SPSS program. Based on data analysis using an independent sample t-test, it was revealed that the t-test was higher than t-table (4.332>1.672). We state a conclusion; the use of probing prompting technique is effective because students who taught using probing-prompting technique perform better achievement on reading. ",,https://jurnal.untidar.ac.id/index.php/metathesis/article/download/1699/1251
-Probing-Prompting Based On Ethnomathematics Learning Model: The Effect On Mathematical Communication Skill,Sitti Hartinah,"Communication skills on mathematics should be mastered by students to involve in mathematics learning process. T his study aims to determine the effect of the Probing-Prompting based on Ethnomatematics learning model on the impact on mathematical communication skills. This research is an experimental with simple random sampling technique . T he instrument used to collect data is a test of mathematical communication skills. Test trials include content validity, level of difficulty, difference in power, and reliability. For the normality test using the Liliefors-test, the homogeneity test uses a similarity test of two variances. The hypothesis test used by t-test. The results show that there is an influence from the Probing-Prompting based on Ethnomatematics learning model on mathematical communication skills.","{'model': 'tldr@v2.0.0', 'text': 'The results show that there is an influence from the Probing-Prompting based on Ethnomatematics learning model on mathematical communication skills.'}",https://dergipark.org.tr/tr/download/article-file/876976
-A Systematic Review of Artificial Intelligence Techniques in Cancer Prediction and Diagnosis,Yogesh Kumar,,"{'model': 'tldr@v2.0.0', 'text': 'An efficient search was performed and included the research articles that employed AI-based learning approaches for cancer prediction using conventional machine and deep learning-based classifications between 2009 and 2021.'}",https://link.springer.com/content/pdf/10.1007/s11831-021-09648-w.pdf
-Heart Disease Prediction using Machine Learning Techniques,Devansh Shah,,"{'model': 'tldr@v2.0.0', 'text': 'This research paper presents various attributes related to heart disease, and the model on basis of supervised learning algorithms as Naïve Bayes, decision tree, K-nearest neighbor, and random forest algorithm, using the existing dataset from the Cleveland database of UCI repository of heart disease patients.'}",
-"Prompts, Not Questions: Four Techniques for Crafting Better Interview Protocols",Tomás R. Jiménez,,,
-Unpacking mHealth interventions: A systematic review of behavior change techniques used in randomized controlled trials assessing mHealth effectiveness,M. Dugas,"Objective Mobile health interventions have surged in popularity but their implementation varies widely and evidence of effectiveness is mixed. We sought to advance understanding of the diversity of behavior change techniques in mHealth interventions, especially those that leverage advanced mobile technologies. Methods We conducted a systematic review of articles published between 2007 and 2017 in high-impact journals in medicine, medical informatics, and health psychology to identify randomized controlled trials in which the effectiveness of an mobile health intervention was tested. Search terms included a mix of general (e.g. mobile health), hardware (e.g. Android, iPhone), and format (e.g. SMS, application) terms. Results In a systematic review of 21 studies, we found the techniques of personalization, feedback and monitoring, and associations were most commonly used in mobile health interventions, but there remains considerable opportunity to leverage more sophisticated aspects of ubiquitous computing. We found that prompts and cues were the most common behavior change techniques used in effective trials, but there was notable overlap in behavior change techniques used in ineffective trials. Conclusions Our results identify techniques that are commonly used in mobile health interventions and highlight pathways to advance the science of mobile health.","{'model': 'tldr@v2.0.0', 'text': 'The techniques of personalization, feedback and monitoring, and associations were most commonly used in mobile health interventions, but there remains considerable opportunity to leverage more sophisticated aspects of ubiquitous computing.'}",https://journals.sagepub.com/doi/pdf/10.1177/2055207620905411
-Behaviour Change Techniques Associated with Smoking Cessation in Intervention and Comparator Groups of Randomised Controlled Trials: A Systematic Review and Meta-Regression.,Nicola Black,"AIMS To estimate the strengths of associations between use of behaviour change techniques (BCTs) and clusters of BCTs in behavioural smoking cessation interventions and comparators with smoking cessation rates. METHOD Systematic review and meta-regression of biochemically verified smoking cessation rates on BCTs in interventions and comparators in randomised controlled trials, adjusting for a priori defined potential confounding variables, together with moderation analyses. Studies were drawn from the Cochrane Tobacco Addiction Group Specialised Register. Data were extracted from published and unpublished (i.e., obtained from study authors) study materials by two independent coders. Adequately described intervention (k = 143) and comparator (k = 92) groups were included in the analyses (N = 43992 participants). Using bivariate mixed-effects meta-regressions, while controlling for key a priori confounders, we regressed smoking cessation on a) three BCT groupings consistent with dual-process theory (i.e., associative, reflective motivational, and self-regulatory), b) 17 expert-derived BCT groupings (i.e., BCT taxonomy v1 clusters), and c) individual BCTs from the BCT taxonomy v1. RESULTS Amongst person-delivered interventions, higher smoking cessation rates were predicted by BCTs targeting associative and self-regulatory processes (B = 0.034-0.041, p < .05), and by three individual BCTs (prompting commitment, social reward, identity associated with changed behaviour), Amongst written interventions, BCTs targeting taxonomy cluster 10a (rewards) predicted higher smoking cessation (B = 0.394, p < .05). Moderation effects were observed for nicotine dependence, mental health status, and mode of delivery. CONCLUSIONS Amongst person-delivered behavioural smoking cessation interventions, specific behaviour change techniques and clusters of techniques are associated with higher success rates.","{'model': 'tldr@v2.0.0', 'text': 'Amongst person-delivered behavioural smoking cessation interventions, specific behaviour change techniques and clusters of techniques are associated with higher success rates and Moderation effects were observed for nicotine dependence, mental health status, and mode of delivery.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/add.15056
-A Systematic Review of Tactile Prompting in Teacher Education,Andrew M. Markelz,"Researchers are exploring innovative technologies to prepare educators. A systematic literature review was conducted to study the role of tactile prompting (i.e., vibratory signals on a schedule) in teacher education. Ten studies meeting inclusion criteria were synthesized and analyzed based on how tactile prompting is utilized, effects of tactile prompting, and quality of research. Findings suggest studies with tactile prompting as the primary independent variable, or condition within a phase, demonstrated medium to very high Tau-U single-case effect sizes (range = 0.54-1.00) and high social acceptability on classroom management teaching behaviors. Simplistic prompting devices reduce cost, training, and anxiety barriers to adoption; however, innovations in wearable technology provide opportunities for a more robust behavior modification tool. Research suggests tactile prompting reduces cognitive load and provides multiple opportunities to practice, which has value in time-constrained preparation programs. Further exploration of this technique in teacher education is warranted.",,
-How to use the nominal group and Delphi techniques,S. McMillan,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the NGT and Delphi technique is provided, including the steps involved and the types of research questions best suited to each method, with examples from the pharmacy literature.'}",https://link.springer.com/content/pdf/10.1007%2Fs11096-016-0257-x.pdf
-Evaluation of Microencapsulation Techniques for MICP Bacterial Spores Applied in Self-Healing Concrete,Wiboonluk Pungrasmi,,"{'model': 'tldr@v2.0.0', 'text': 'It is revealed that freeze drying has a high potential as a microencapsulation technique for application to self-healing concrete technology.'}",https://www.nature.com/articles/s41598-019-49002-6.pdf
-The mathematical communication ability based on gender difference on students of 11th grade by using problem-based learning model assisted by probing prompting technique,Ishmah Khairina Zakiri,"This study aims to test the MCA (Mathematical Communication Ability) in Problem Based Learning (PBL) model assisted Probing Prompting Technique (PBL PP), to achieve learning mastery, to know the average ratio of MCA of students on learning model of PBL PP and PBL only, to differ the MCA of male students with female students in PBL PP, and to describe MCA of students based on gender difference. The method used in this research is mix method. While the population is science program students of XI grade at MAN 2 Semarang. The sample is chosen with random sampling. In addition, students of XI IPA 1 are as experiment class, while students of XI IPA 2 as control class. The qualitative research subjects in this study are six students, for each is taken three from every male and female students in experiment class. The results of this research are the MCA’s students with PBL PP learning reaches the learning mastery, the average ratio of MCA in PBL PP is better than PBL model, the MCA of male and female students have no significant difference, and there are some differences of MCA in each indicators based on gender difference.",,
-Behavior Change Techniques Implemented in Electronic Lifestyle Activity Monitors: A Systematic Content Analysis,E. Lyons,"Background Electronic activity monitors (such as those manufactured by Fitbit, Jawbone, and Nike) improve on standard pedometers by providing automated feedback and interactive behavior change tools via mobile device or personal computer. These monitors are commercially popular and show promise for use in public health interventions. However, little is known about the content of their feedback applications and how individual monitors may differ from one another. Objective The purpose of this study was to describe the behavior change techniques implemented in commercially available electronic activity monitors. Methods Electronic activity monitors (N=13) were systematically identified and tested by 3 trained coders for at least 1 week each. All monitors measured lifestyle physical activity and provided feedback via an app (computer or mobile). Coding was based on a hierarchical list of 93 behavior change techniques. Further coding of potentially effective techniques and adherence to theory-based recommendations were based on findings from meta-analyses and meta-regressions in the research literature. Results All monitors provided tools for self-monitoring, feedback, and environmental change by definition. The next most prevalent techniques (13 out of 13 monitors) were goal-setting and emphasizing discrepancy between current and goal behavior. Review of behavioral goals, social support, social comparison, prompts/cues, rewards, and a focus on past success were found in more than half of the systems. The monitors included a range of 5-10 of 14 total techniques identified from the research literature as potentially effective. Most of the monitors included goal-setting, self-monitoring, and feedback content that closely matched recommendations from social cognitive theory. Conclusions Electronic activity monitors contain a wide range of behavior change techniques typically used in clinical behavioral interventions. Thus, the monitors may represent a medium by which these interventions could be translated for widespread use. This technology has broad applications for use in clinical, public health, and rehabilitation settings.","{'model': 'tldr@v2.0.0', 'text': 'Electronic activity monitors contain a wide range of behavior change techniques typically used in clinical behavioral interventions, and may represent a medium by which these interventions could be translated for widespread use.'}",https://www.jmir.org/2014/8/e192/PDF
-Review of Modern Techniques for the Assessment of Skin Hydration,M. Qassem,"Skin hydration is a complex process that influences the physical and mechanical properties of skin. Various technologies have emerged over the years to assess this parameter, with the current standard being electrical probe-based instruments. Nevertheless, their inability to provide detailed information has prompted the use of sophisticated spectroscopic and imaging methodologies, which are capable of in-depth skin analysis that includes structural and composition details. Modern imaging and spectroscopic techniques have transformed skin research in the dermatological and cosmetics disciplines, and are now commonly employed in conjunction with traditional methods for comprehensive assessment of both healthy and pathological skin. This article reviews current techniques employed in measuring skin hydration, and gives an account on their principle of operation and applications in skin-related research.","{'model': 'tldr@v2.0.0', 'text': 'Current techniques employed in measuring skin hydration are reviewed, and an account on their principle of operation and applications in skin-related research is given.'}",https://www.mdpi.com/2079-9284/6/1/19/pdf?version=1552129987
-Classifiers Combination Techniques: A Comprehensive Review,M. Mohandes,"In critical applications, such as medical diagnosis, security related systems, and so on, the cost or risk of action taking based on incorrect classification can be very high. Hence, combining expert opinions before taking decision can substantially increase the reliability of such systems. Such pattern recognition systems base their final decision on evidence collected from different classifiers. Such evidence can be of data type, feature type, or classifier type. Common problems in pattern recognition, such as curse of dimensionality, and small sample data size, among others, have also prompted researchers into seeking new approaches for combining evidences. This paper presents a criteria-based framework for multi-classifiers combination techniques and their areas of applications. The criteria discussed here include levels of combination, types of thresholding, adaptiveness of the combination, and ensemble-based approaches. The strengths and weaknesses of each of these categories are discussed in details. Following this analysis, we provide our perspective on the outlook of this area of research and open problems. The lack of a well-formulated theoretical framework for analyzing the performance of combination techniques is shown to provide a fertile ground for further research. In addition to summarizing the existing work, this paper also updates and complements the latest developments in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'A criteria-based framework for multi-classifiers combination techniques and their areas of applications is presented and the lack of a well-formulated theoretical framework for analyzing the performance of combination techniques is shown to provide a fertile ground for further research.'}",
-Security Techniques for the Electronic Health Records,C. Kruse,,"{'model': 'tldr@v2.0.0', 'text': 'This review seeks to analyze and discuss prominent security techniques for healthcare organizations seeking to adopt a secure electronic health records system using PubMed, CINAHL, and ProQuest Nursing and Allied Health Source as sources.'}",https://link.springer.com/content/pdf/10.1007%2Fs10916-017-0778-4.pdf
-Behavior change techniques for increasing physical activity in cancer survivors: a systematic review and meta-analysis of randomized controlled trials,E. Finne,"Purpose The purpose of this systematic review and meta-analysis is to investigate how physical activity (PA) can be effectively promoted in cancer survivors. The effect of PA-promoting interventions in general, behavior change techniques (BCTs), and further variables as moderators in particular are evaluated. Methods This study included randomized controlled trials of lifestyle interventions aiming at an increase in PA that can be carried out independently at home, published by December 2016, for adults diagnosed with cancer after completion of the main treatment. Primary outcomes were subjective and objective measures of PA prior to and immediately after the intervention. Meta-analysis and meta-regression were used to estimate effect sizes (ES) in terms of standardized mean differences, variation between ES in terms of heterogeneity indices (I2), and moderator effects in terms of regression coefficients. Results This study included 30 studies containing 45 ES with an overall significant small positive effect size of 0.28 (95% confidence interval=0.18–0.37) on PA, and I2=54.29%. The BCTs Prompts, Reduce prompts, Graded tasks, Non-specific reward, and Social reward were significantly related to larger effects, while Information about health consequences and Information about emotional consequences, as well as Social comparison were related to smaller ES. The number of BCTs per intervention did not predict PA effects. Interventions based on the Theory of Planned Behavior were associated with smaller ES, and interventions with a home-based setting component were associated with larger ES. Neither the duration of the intervention nor the methodological quality explained differences in ES. Conclusion Certain BCTs were associated with an increase of PA in cancer survivors. Interventions relying on BCTs congruent with (social) learning theory such as using prompts and rewards could be especially successful in this target group. However, large parts of between-study heterogeneity in ES remained unexplained. Further primary studies should directly compare specific BCTs and their combinations.","{'model': 'tldr@v2.0.0', 'text': 'Interventions relying on BCTs congruent with (social) learning theory such as using prompts and rewards could be especially successful in this target group of cancer survivors, however, large parts of between-study heterogeneity in ES remained unexplained.'}",https://www.dovepress.com/getfile.php?fileID=45752
-Problematising ethnography and case study: reflections on using ethnographic techniques and researcher positioning,M. Parker-Jenkins,"ABSTRACT This paper was prompted by the question, what do we mean by conducting ‘ethnography’? Is it in fact ‘case study’ drawing on ethnographic techniques? My contention is that in many cases, researchers are not actually conducting ethnography as understood within a traditional sense but rather are engaging in case study, drawing on ethnographic techniques. Does that matter you might ask? Well it determines what we can expect to discover from a research project in terms of results and the unearthing of deeper complexities. I frame the discussion around a set of closely related issues, namely ethnography, case study and researcher positioning, drawing on ethnographic techniques and fieldwork relations. The original contribution of the piece and overall argument is that research can represent a hybrid form, and based on my own research experience, I propose a new term ‘ethno-case study’ that has advantages of both ethnography and case study.",,
-What are the most effective techniques in changing obese individuals’ physical activity self-efficacy and behaviour: a systematic review and meta-analysis,E. Olander,,"{'model': 'tldr@v2.0.0', 'text': 'A non-significant relationship was found between change in self-efficacy and change in physical activity (Spearman’s Rho”=\u2009−0.18 p\u2009=\u202c0.72) and the majority of techniques increased physical activity behaviour, without having discernible effects on self- efficacy.'}",https://ijbnpa.biomedcentral.com/counter/pdf/10.1186/1479-5868-10-29
-Do physical activity and dietary smartphone applications incorporate evidence-based behaviour change techniques?,A. Direito,,"{'model': 'tldr@v2.0.0', 'text': 'Presence of BCTs varied by app type and price; however, B CTs associated with increased intervention effectiveness were in general more common in paid apps.'}",https://bmcpublichealth.biomedcentral.com/counter/pdf/10.1186/1471-2458-14-646
-Machine learning techniques for quality control in high conformance manufacturing environment,Carlos A. Escobar,"In today’s highly competitive global market, winning requires near-perfect quality. Although most mature organizations operate their processes at very low defects per million opportunities, customers expect completely defect-free products. Therefore, the prompt detection of rare quality events has become an issue of paramount importance and an opportunity for manufacturing companies to move quality standards forward. This article presents the learning process and pattern recognition strategy for a knowledge-based intelligent supervisory system, in which the main goal is the detection of rare quality events. Defect detection is formulated as a binary classification problem. The l1-regularized logistic regression is used as the learning algorithm for the classification task and to select the features that contain the most relevant information about the quality of the process. The proposed strategy is supported by the novelty of a hybrid feature elimination algorithm and optimal classification threshold search algorithm. According to experimental results, 100% of defects can be detected effectively.","{'model': 'tldr@v2.0.0', 'text': 'The learning process and pattern recognition strategy for a knowledge-based intelligent supervisory system, in which the main goal is the detection of rare quality events, is presented and 100% of defects can be detected effectively.'}",https://journals.sagepub.com/doi/pdf/10.1177/1687814018755519
-PENERAPAN TEKNIK PROBING-PROMPTING DALAM PEMBELAJARAN MATEMATIKA SISWA KELAS VIII MTSN LUBUK BUAYA PADANG,Yuriska Mandasari,"The ability of mathematical communication is the important ability in the science development. The ability of mathematical communication is expected can increase simultaneously by the development of the science and technology. Nevertheless, the implementation of mathematical learning still focuses to the teacher, this method makes students cannot develop their ability in mathematical communication. Their report study in mathematics is still under minimum completeness criteria (KKM), namely 75. One of the recommended effort to deal with this condition is the using probing-prompting technique. The aim of this research is knowing the impact of the enhancement of student’s ability in mathematical communication using probing-prompting technique rather than are compared with convensional learning technique. The result of this research show that student’s ability of mathematical communication using probing-prompting technique is better than convensional technique with α = 0,05. To conclude, probing-prompting technique can increase student is ability in mathematical communication. Key words : mathematical communication ability, probing-prompting technique, convensional learning","{'model': 'tldr@v2.0.0', 'text': 'The result of this research show that student’s ability of mathematical communication using probing-prompting technique is better than convensional technique with α = 0,05.'}",
-Pengaruh Model Pembelajaran Inkuiri Terbimbing Dengan Teknik Probing-prompting Terhadap Aktivitas Dan Hasil Belajar Fisika Di SMA,Siscawati Rizki Lasmo,"This research concentrated of guided inquiry model accompanied with probing-prompting technique for activity and learning outcomes physisics in SMA. The purposes of this research were to study the effect of using inquiry model accompanied with probing-prompting technique to student's learning activities and kognitif student's achievement. The type of this research was an experimental research by post-test control group design. Population of the research was all student's at SMAN 1 Pakusari. The technique to collect data was documentation, observation, interview, and tests. The analysis result description of student learning activities for experiment class equals to 85.80 % is in active criteria. The analysis data of kognitif student's achievement by using independent sample t-test for significant value (1-tailed) was 0,018(≤0,05), it's mean the kognitif students' achievement for experiment class is better than control class (Ha accepted and Ho refused). Conclusion of this research are : (1) student's learning activities using guided inquiry model accompanied with probing-prompting technique is in active criteria which percentage level of activity is 85.80%. (2) Guided Inquiry Model accompanied with probing-prompting technique has significant effect of learning outcomes physisics in SMA. Keywords: guided inquiry model, probing-prompting technique",,
-Structured Analytic Techniques for Intelligence Analysis,R. Heuer,"This book takes the relatively new concept of structured analytic techniques, defines its place in a taxonomy of analytic methods, and moves it a giant leap forward. It describes 50 techniques that are divided into eight categories. There are techniques for: Decomposition and Visualization Idea GenerationScenarios and IndicatorsHypothesis Generation and TestingCause and Effect Challenge AnalysisConflict ManagementDecision Support Each structured technique involves a step-by-step process that externalizes an individual analyst s thinking in a manner that makes it readily apparent to others, thereby enabling it to be shared, built on, and easily critiqued by others. This structured and transparent process combined with the intuitive input of subject matter experts is expected to reduce the risk of analytic error. Our current high tech, global environment increasingly requires collaboration between analysts with different areas of expertise and analysts representing different organizational perspectives. Structured analytic techniques are the ideal process for guiding the interaction of analysts within a small team or group. Each step in a technique prompts relevant discussion within the team, and such discussion generates and evaluates substantially more divergent information and more new ideas than a team that does not use a structured process. By defining the domain of structured analytic techniques, providing a manual for using and teaching these techniques, and outlining procedures for evaluating and validating these techniques, this book lays a common ground for continuing improvement of how analysis is done. These techniques are especially needed in the field of intelligence analysis where analysts typically deal with incomplete, ambiguous and sometimes deceptive information. However, these practical tools for analysis are also useful in a wide variety of professions including law enforcement, medicine, finance, and business.","{'model': 'tldr@v2.0.0', 'text': 'This book takes the relatively new concept of structured analytic techniques, defines its place in a taxonomy of analytic methods, and moves it a giant leap forward by providing a manual for using and teaching these techniques, and outlining procedures for evaluating and validating these techniques.'}",
-A survey of location inference techniques on Twitter,Oluwaseun Ajao,"The increasing popularity of the social networking service, Twitter, has made it more involved in day-to-day communications, strengthening social relationships and information dissemination. Conversations on Twitter are now being explored as indicators within early warning systems to alert of imminent natural disasters such as earthquakes and aid prompt emergency responses to crime. Producers are privileged to have limitless access to market perception from consumer comments on social media and microblogs. Targeted advertising can be made more effective based on user profile information such as demography, interests and location. While these applications have proven beneficial, the ability to effectively infer the location of Twitter users has even more immense value. However, accurately identifying where a message originated from or an author’s location remains a challenge, thus essentially driving research in that regard. In this paper, we survey a range of techniques applied to infer the location of Twitter users from inception to state of the art. We find significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.","{'model': 'tldr@v2.0.0', 'text': 'A range of techniques applied to infer the location of Twitter users from inception to state of the art are surveyed, finding significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.'}",http://shura.shu.ac.uk/23769/1/Ajao_survey_location_%28VoR%29.pdf
-Using a behaviour change techniques taxonomy to identify active ingredients within trials of implementation interventions for diabetes care,J. Presseau,,"{'model': 'tldr@v2.0.0', 'text': 'The recently developed Behaviour Change Techniques Taxonomy version 1 (BCTTv1) can be used to characterise the active ingredients in trials of implementation interventions and provides specificity of content beyond what is given by broader intervention labels.'}",https://implementationscience.biomedcentral.com/counter/pdf/10.1186/s13012-015-0248-7
-Rapid tests and urine sampling techniques for the diagnosis of urinary tract infection ( UTI ) in children under five years : a systematic review,P. Whiting,"Background: Urinary tract infection (UTI) is one of the most common sources of infection in children under five. Prompt diagnosis and treatment is important to reduce the risk of renal scarring. Rapid, cost-effective, methods of UTI diagnosis are required as an alternative to culture. Methods: We conducted a systematic review to determine the diagnostic accuracy of rapid tests for detecting UTI in children under five years of age. Results: The evidence supports the use of dipstick positive for both leukocyte esterase and nitrite (pooled LR+ = 28.2, 95% CI: 17.3, 46.0) or microscopy positive for both pyuria and bacteriuria (pooled LR+ = 37.0, 95% CI: 11.0, 125.9) to rule in UTI. Similarly dipstick negative for both LE and nitrite (Pooled LR= 0.20, 95% CI: 0.16, 0.26) or microscopy negative for both pyuria and bacteriuria (Pooled LR= 0.11, 95% CI: 0.05, 0.23) can be used to rule out UTI. A test for glucose showed promise in potty-trained children. However, all studies were over 30 years old. Further evaluation of this test may be useful. Conclusion: Dipstick negative for both LE and nitrite or microscopic analysis negative for both pyuria and bacteriuria of a clean voided urine, bag, or nappy/pad specimen may reasonably be used to rule out UTI. These patients can then reasonably be excluded from further investigation, without the need for confirmatory culture. Similarly, combinations of positive tests could be used to rule in UTI, and trigger further investigation. Background Urinary tract infection (UTI) is one of the most common sources of infection in children under 5. In a small proportion of children UTI may lead to renal scarring [1,2]. This outcome of infection is of concern as it is associated with significant future complications and ultimately with end stage renal disease[3]. Prompt diagnosis and treatment is therefore important to reduce the risk of future scarring. Published: 05 April 2005 BMC Pediatrics 2005, 5:4 doi:10.1186/1471-2431-5-4 Received: 05 October 2004 Accepted: 05 April 2005 This article is available from: http://www.biomedcentral.com/1471-2431/5/4 © 2005 Whiting et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BMC Pediatrics 2005, 5:4 http://www.biomedcentral.com/1471-2431/5/4","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of rapid tests for detecting UTI in children under five years of age found the evidence supports the use of dipstick positive for both leukocyte esterase and nitrite or microscopic analysis negative for both pyuria and bacteriuria of a clean voided urine, bag, or nappy/pad specimen to rule out UTI.'}",
-Task Effects on Linguistic Complexity and Accuracy: A Large-Scale Learner Corpus Analysis Employing Natural Language Processing Techniques,T. Alexopoulou,"Large-scale learner corpora collected from online language learning platforms, such as the EF-Cambridge Open Language Database (EFCAMDAT), provide opportunities to analyze learner data at an unprecedented scale. However, interpreting the learner language in such corpora requires a precise understanding of tasks: Howdoes the prompt and input of a task and its functional requirements influence task-based linguistic performance? This question is vital for making large-scale task-based corpora fruitful for second language acquisition research. We explore the issue through an analysis of selected tasks in EFCAMDAT and the complexity and accuracy of the language they elicit.","{'model': 'tldr@v2.0.0', 'text': 'The issue of how the prompt and input of a task and its functional requirements influence task-based linguistic performance is explored through an analysis of selected tasks in EFCAMDAT and the complexity and accuracy of the language they elicit.'}",https://eprints.lancs.ac.uk/id/eprint/83702/3/Alexopoulou_et_al_in_press.pdf
-Behavior Change Techniques in mHealth Apps for the Mental and Physical Health of Employees: Systematic Assessment,E. D. de Korte,"Background Employees remain at risk of developing physical and mental health problems. To improve the lifestyle, health, and productivity many workplace interventions have been developed. However, not all of these interventions are effective. Mobile and wireless technology to support health behavior change (mobile health [mHealth] apps) is a promising, but relatively new domain for the occupational setting. Research on mHealth apps for the mental and physical health of employees is scarce. Interventions are more likely to be useful if they are rooted in health behavior change theory. Evaluating the presence of specific combinations of behavior change techniques (BCTs) in mHealth apps might be used as an indicator of potential quality and effectiveness. Objective The aim of this study was to assess whether mHealth apps for the mental and physical health of employees incorporate BCTs and, if so, which BCTs can be identified and which combinations of BCTs are present. Methods An assessment was made of apps aiming to reduce the risk of physical and psychosocial work demands and to promote a healthy lifestyle for employees. A systematic search was performed in iTunes and Google Play. Forty-five apps were screened and downloaded. BCTs were identified using a taxonomy applied in similar reviews. The mean and ranges were calculated. Results On average, the apps included 7 of the 26 BCTs (range 2-18). Techniques such as “provide feedback on performance,” “provide information about behavior-health link,” and “provide instruction” were used most frequently. Techniques that were used least were “relapse prevention,” “prompt self-talk,” “use follow-up prompts,” and “provide information about others’ approval.” “Stress management,” “prompt identification as a role model,” and “agree on behavioral contract” were not used by any of the apps. The combination “provide information about behavior-health link” with “prompt intention formation” was found in 7/45 (16%) apps. The combination “provide information about behavior-health link” with “provide information on consequences,” and “use follow-up prompts” was found in 2 (4%) apps. These combinations indicated potential effectiveness. The least potentially effective combination “provide feedback on performance” without “provide instruction” was found in 13 (29%) apps. Conclusions Apps for the occupational setting might be substantially improved to increase potential since results showed a limited presence of BCTs in general, limited use of potentially successful combinations of BCTs in apps, and use of potentially unsuccessful combinations of BCTs. Increasing knowledge on the effectiveness of BCTs in apps might be used to develop guidelines for app developers and selection criteria for companies and individuals. Also, this might contribute to decreasing the burden of work-related diseases. To achieve this, app developers, health behavior change professionals, experts on physical and mental health, and end-users should collaborate when developing apps for the working context.","{'model': 'tldr@v2.0.0', 'text': 'App developers, health behavior change professionals, experts on physical and mental health, and end-users should collaborate when developing apps for the working context to increase potential since results showed a limited presence of BCTs in general, limited use of potentially successful combinations of B CTs in apps, and use of possibly unsuccessful combinations ofBCTs.'}",https://mhealth.jmir.org/2018/10/e167/PDF
-A large-scale empirical comparison of static and dynamic test case prioritization techniques,Qi Luo,"The large body of existing research in Test Case Prioritization (TCP) techniques, can be broadly classified into two categories: dynamic techniques (that rely on run-time execution information) and static techniques (that operate directly on source and test code). Absent from this current body of work is a comprehensive study aimed at understanding and evaluating the static approaches and comparing them to dynamic approaches on a large set of projects. In this work, we perform the first extensive study aimed at empirically evaluating four static TCP techniques comparing them with state-of-research dynamic TCP techniques at different test-case granularities (e.g., method and class-level) in terms of effectiveness, efficiency and similarity of faults detected. This study was performed on 30 real-word Java programs encompassing 431 KLoC. In terms of effectiveness, we find that the static call-graph-based technique outperforms the other static techniques at test-class level, but the topic-model-based technique performs better at test-method level. In terms of efficiency, the static call-graph-based technique is also the most efficient when compared to other static techniques. When examining the similarity of faults detected for the four static techniques compared to the four dynamic ones, we find that on average, the faults uncovered by these two groups of techniques are quite dissimilar, with the top 10% of test cases agreeing on only 25% - 30% of detected faults. This prompts further research into the severity/importance of faults uncovered by these techniques, and into the potential for combining static and dynamic information for more effective approaches.","{'model': 'tldr@v2.0.0', 'text': 'This first extensive study aimed at empirically evaluating four static TCP techniques comparing them with state-of-research dynamic TCP techniques at different test-case granularities in terms of effectiveness, efficiency and similarity of faults detected finds that the static call-graph-based technique outperforms the other static techniques at test-class level, but the topic-model-based techniques performs better atTest-method level.'}",https://arxiv.org/pdf/1801.05917
-Using Progressive Video Prompting to Teach Students with Moderate Intellectual Disability to Shoot a Basketball,Ya-yu Lo,"The current study examined the effects of a modified video prompting procedure, namely progressive video prompting, to increase technique accuracy of shooting a basketball in the school gymnasium of three 11th-grade students with moderate intellectual disability. The intervention involved participants viewing video clips of an adult model who showed progressively chucked steps for making a free throw. We used a single-case, multiple probe across participants design to evaluate the intervention effects. The results of this study showed that all three participants increased the number of steps performed correctly and maintained the skill at the 1-week and 2-week maintenance check without the video viewing. Implications for practice and future research related to video prompting variations are discussed.",,
-Cyber-Attack Modeling Analysis Techniques: An Overview,Hamad Al-Mohannadi,"Cyber attack is a sensitive issue in the world of Internet security. Governments and business organisations around the world are providing enormous effort to secure their data. They are using various types of tools and techniques to keep the business running, while adversaries are trying to breach security and send malicious software such as botnets, viruses, trojans etc., to access valuable data. Everyday the situation is getting worse because of new types of malware emerging to attack networks. It is important to understand those attacks both before and after they happen in order to provide better security to our systems. Understanding attack models provide more insight into network vulnerability, which in turn can be used to protect the network from future attacks. In the cyber security world, it is difficult to predict a potential attack without understanding the vulnerability of the network. So, it is important to analyse the network to identify top possible vulnerability list, which will give an intuitive idea to protect the network. Also, handling an ongoing attack poses significant risk on the network and valuable data, where prompt action is necessary. Proper utilisation of attack modelling techniques provide advance planning, which can be implemented rapidly during an ongoing attack event. This paper aims to analyse various types of existing attack modelling techniques to understand the vulnerability of the network, and the behaviour and goals of the adversary. The ultimate goal is to handle cyber attack in efficient manner using attack modelling techniques.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to analyse various types of existing attack modelling techniques to understand the vulnerability of the network, and the behaviour and goals of the adversary.'}",https://bradscholars.brad.ac.uk/bitstream/10454/10703/2/hamad_ICI.pdf
-Handling Imbalanced and Overlapping Classes in Smart Environments Prompting Dataset,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'In this chapter, intersection between the problems of imbalanced class and overlapping classes is explored from the perspective of Smart Environments as the application domain and a novel clustering-based under-sampling technique is proposed.'}",
-Fluoride Removal from Water by various techniques : Review,S. Waghmare,"This paper depicts the fluoride removal from drinking water can be accomplished by different methods, for example, coagulation-precipitation, membrane separation process, ion exchange, adsorption techniques and so on. Among these procedures, membrane and ion exchange processes are not extremely regular because of its high establishment and support price. Other two tecniques are extremely regular in India . Nalgonda procedure is one of the well known strategies generally utilized for defluoridation of water as a part of developing nations, for example, India, Kenya, Senegal and Tanzania. Among different tecniques utilized for defluoridation of water, the adsorption procedure is broadly utilized and offers acceptable results and is by all accounts more appealing technique for the removal of fluoride regarding expense, straightforwardness of outline and operation. It is apparent from the literature study that different methods have indicated novel potential for the removal of fluoride. Be that as it may, in any case there is a need to figure out the reasonable utility of such developed procedures on a business scale, prompting the change of contamination control.",,
-Combinations of techniques that effectively change health behavior: evidence from Meta-CART analysis.,E. Dusseldorp,"OBJECTIVE Many health-promoting interventions combine multiple behavior change techniques (BCTs) to maximize effectiveness. Although, in theory, BCTs can amplify each other, the available meta-analyses have not been able to identify specific combinations of techniques that provide synergistic effects. This study overcomes some of the shortcomings in the current methodology by applying classification and regression trees (CART) to meta-analytic data in a special way, referred to as Meta-CART. The aim was to identify particular combinations of BCTs that explain intervention success. METHOD A reanalysis of data from Michie, Abraham, Whittington, McAteer, and Gupta (2009) was performed. These data included effect sizes from 122 interventions targeted at physical activity and healthy eating, and the coding of the interventions into 26 BCTs. A CART analysis was performed using the BCTs as predictors and treatment success (i.e., effect size) as outcome. A subgroup meta-analysis using a mixed effects model was performed to compare the treatment effect in the subgroups found by CART. RESULTS Meta-CART identified the following most effective combinations: Provide information about behavior-health link with Prompt intention formation (mean effect size ḡ = 0.46), and Provide information about behavior-health link with Provide information on consequences and Use of follow-up prompts (ḡ = 0.44). Least effective interventions were those using Provide feedback on performance without using Provide instruction (ḡ = 0.05). CONCLUSIONS Specific combinations of BCTs increase the likelihood of achieving change in health behavior, whereas other combinations decrease this likelihood. Meta-CART successfully identified these combinations and thus provides a viable methodology in the context of meta-analysis.","{'model': 'tldr@v2.0.0', 'text': 'Specific combinations of BCTs increase the likelihood of achieving change in health behavior, whereas other combinations decrease this likelihood, and Meta-CART successfully identified these combinations and thus provides a viable methodology in the context of meta-analysis.'}",http://www.elisedusseldorp.nl/pdf/Dusseldorp2014_HPMetaCART.pdf
-"Support for Learning from Multimedia Explanations. A Comparison of Prompting, Signaling, and Questioning",H. Garcia-Rodicio,In one experiment 97 undergraduate students learned about plate tectonics from a multimedia presentation involving narrated animations and support in one of four forms. Support in the prompting condition included hints inducing participants to self-explain critical information. The signaling condition included overviews recapping critical information. The questioning condition included questions about critical information and feedback on participants' answers. The control condition included no support. Participants in the questioning condition outperformed those in the rest of conditions in retention and transfer. This means that questioning is a very powerful technique for promoting multimedia learning.,,
-Application of machine learning techniques to sentiment analysis,A. Jain,"Today, we live in a ‘data age’. Due to rapid increase in the amount of user-generated data on social media platforms like Twitter, several opportunities and new open doors have been prompted for organizations that endeavour hard to keep a track on customer reviews and opinions about their products. Twitter is a huge fast emergent micro-blogging social networking platform for users to express their views about politics, products sports etc. These views are useful for businesses, government and individuals. Hence, tweets can be used as a valuable source for mining public's opinion. Sentiment analysis is a process of automatically identifying whether a user-generated text expresses positive, negative or neutral opinion about an entity (i.e. product, people, topic, event etc). The objective of this paper is to give step-by-step detail about the process of sentiment analysis on twitter data using machine learning. This paper also provides details of proposed approach for sentiment analysis. This work proposes a Text analysis framework for twitter data using Apache spark and hence is more flexible, fast and scalable. Naïve Bayes and Decision trees machine learning algorithms are used for sentiment analysis in the proposed framework.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Text analysis framework for twitter data using Apache spark and hence is more flexible, fast and scalable and Naïve Bayes and Decision trees machine learning algorithms are used for sentiment analysis in the proposed framework.'}",
-Behavior Change Techniques Used to Promote Walking and Cycling,E. Bird,"Objective: Evidence on the effectiveness of walking and cycling interventions is mixed. This may be partly attributable to differences in intervention content, such as the cognitive and behavioral techniques (BCTs) used. Adopting a taxonomy of BCTs, this systematic review addressed two questions: (a) What are the behavior change techniques used in walking and cycling interventions targeted at adults? (b) What characterizes interventions that appear to be associated with changes in walking and cycling in adults? Method: Previous systematic reviews and updated database searches were used to identify controlled studies of individual-level walking and cycling interventions involving adults. Characteristics of intervention design, context, and methods were extracted in addition to outcomes. Intervention content was independently coded according to a 26-item taxonomy of BCTs. Results: Studies of 46 interventions met the inclusion criteria. Twenty-one reported a statistically significant effect on walking and cycling outcomes. Analysis revealed substantial heterogeneity in the vocabulary used to describe intervention content and the number of BCTs coded. “Prompt self-monitoring of behavior” and “prompt intention formation” were the most frequently coded BCTs. Conclusion: Future walking and cycling intervention studies should ensure that all aspects of the intervention are reported in detail. The findings lend support to the inclusion of self-monitoring and intention formation techniques in future walking and cycling intervention design, although further exploration of these and other BCTs is required. Further investigation of the interaction between BCTs and study design characteristics would also be desirable.","{'model': 'tldr@v2.0.0', 'text': 'The findings lend support to the inclusion of self-monitoring and intention formation techniques in future walking and cycling intervention design, although further exploration of these and other BCTs is required.'}",https://europepmc.org/articles/pmc3727344?pdf=render
-Effective behaviour change techniques in the prevention and management of childhood obesity,Jilly Martin,,"{'model': 'tldr@v2.0.0', 'text': 'To design effective behaviour change programmes for the prevention and management of childhood obesity, it is recommended to utilise the BCTs identified as effective in this review, including six that may be effective components of future management interventions.'}",https://openaccess.city.ac.uk/id/eprint/14188/3/Supplementary%20Table%202.pdf
-Discovering the Effects of Metacognitive Prompts on the Sequential Structure of SRL-Processes Using Process Mining Techniques,Christoph Sonnenberg,"According to research examining self-regulated learning (SRL), we regard individual regulation as a specific sequence of regulatory activities. Ideally, students perform various learning activities, such as analyzing, monitoring, and evaluating cognitive and motivational aspects during learning. Metacognitive prompts can foster SRL by inducing regulatory activities, which, in turn, improve the learning outcome. However, the specific effects of metacognitive support on the dynamic characteristics of SRL are not understood. Therefore, the aim of our study was to analyze the effects of metacognitive prompts on learning processes and outcomes during a computer-based learning task. Participants of the experimental group (EG, n = 35) were supported by metacognitive prompts, whereas participants of the control group (CG, n = 35) received no support. Data regarding learning processes were obtained by concurrent think-aloud protocols. The EG exhibited significantly more metacognitive learning events than did the CG. Furthermore, these regulatory activities correspond positively with learning outcomes. Process mining techniques were used to analyze sequential patterns. Our findings indicate differences in the process models of the EG and CG and demonstrate the added value of taking the order of learning activities into account by discovering regulatory patterns.",,https://learning-analytics.info/index.php/JLA/article/download/4090/4837
-Research and Development Aspects on Chemical Preparation Techniques of Photoanodes for Dye Sensitized Solar Cells,N. Asim,"The importance of dye sensitized solar cells (DSSCs) as a low-cost and environmentally friendly photovoltaic (PV) technology has prompted many researchers to improve its efficiency and durability. The realization of these goals is impossible without taking into account the importance of the materials in DSSCs, so the focus on the preparation/deposition methods is essential. These methods can be either chemical or physical. In this study, the chemical applied methods that utilize chemical reaction to synthesize and deposit the materials are covered and categorized according to their gas phase and liquid phase precursors. Film processing techniques that can be used to enhance the materials' properties postpreparation are also included for further evaluation in this study. However, there is a variety of consideration, and certain criteria must be taken into account when selecting a specific deposition method, due to the fact that the fabrication conditions vary and are unoptimized.",,https://downloads.hindawi.com/journals/ijp/2014/518156.pdf
-MENINGKATKAN KEMAMPUAN PEMAHAMAN DAN DISPOSISI MATEMATIK SISWA SMA MENGGUNAKAN TEKNIK PROBING PROMPTING,Sekolah Tinggi,"This study was intended to analyze the role of probing prompting technique (PPT) toward students’ mathematical understanding ability and disposition and asociation both of them. This study was a part of a master thesis and a sub-study of a Postgraduate Research Grant from DGHE in 2015. This study was a pretest-postest quasi-experimental control group design involving 66 twelfth-grade students of a senior high school in Pengalengan which were chosen puposively.The instruments of this study are an essay test on mathematical understanding ability, and a mathematical disposition (MD) scale. The study revealed that students getting treatment on PPT attained better grades on mathematical understanding ability and disposition than that of students taught by conventional teaching, though the grades of mathematical understanding were at low level and the grades of mathematical disposition were at fairly good. Students realized difficulties in determining requirements of a concept, compiling mathematical model and solving it concerning graph of function. Also, there was association between mathematical understanding ability and disposition, and students performed positive opinioan toward probing prompting technique. Keyword: Mathematical understanding, mathematical disposition, probing prompting technique",,
-Grammatical Evolutionary Techniques for Prompt Migraine Prediction,Josué Pagán,"The migraine disease is a chronic headache presenting symptomatic crisis that causes high economic costs to the national health services, and impacts negatively on the quality of life of the patients. Even if some patients can feel unspecific symptoms before the onset of the migraine, these only happen randomly and cannot predict the crisis precisely. In our work, we have proved how migraine crisis can be predicted with high accuracy from the physiological variables of the patients, acquired by a non-intrusive Wireless Body Sensor Network. In this paper, we derive alternative models for migraine prediction using Grammatical Evolution techniques. We obtain prediction horizons around 20 minutes, which are sufficient to advance the drug intake and avoid the symptomatic crisis. The robustness of the models with respect to sensor failures has also been tackled to allow the practical implementation in the ambulatory monitoring platform. The achieved models are non linear mathematical expressions with low computing overhead during the run-time execution in the wearable devices.","{'model': 'tldr@v2.0.0', 'text': 'This work has proved how migraine crisis can be predicted with high accuracy from the physiological variables of the patients, acquired by a non-intrusive Wireless Body Sensor Network.'}",
-Voyager: An Open-Ended Embodied Agent with Large Language Models,Guanzhi Wang,"We introduce Voyager, the first LLM-powered embodied lifelong learning agent in Minecraft that continuously explores the world, acquires diverse skills, and makes novel discoveries without human intervention. Voyager consists of three key components: 1) an automatic curriculum that maximizes exploration, 2) an ever-growing skill library of executable code for storing and retrieving complex behaviors, and 3) a new iterative prompting mechanism that incorporates environment feedback, execution errors, and self-verification for program improvement. Voyager interacts with GPT-4 via blackbox queries, which bypasses the need for model parameter fine-tuning. The skills developed by Voyager are temporally extended, interpretable, and compositional, which compounds the agent's abilities rapidly and alleviates catastrophic forgetting. Empirically, Voyager shows strong in-context lifelong learning capability and exhibits exceptional proficiency in playing Minecraft. It obtains 3.3x more unique items, travels 2.3x longer distances, and unlocks key tech tree milestones up to 15.3x faster than prior SOTA. Voyager is able to utilize the learned skill library in a new Minecraft world to solve novel tasks from scratch, while other techniques struggle to generalize. We open-source our full codebase and prompts at https://voyager.minedojo.org/.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.16291
-Uptake of Behavior Change Techniques - a Key Focus for Process Analysis and for Intervention Delivery: a Comment on Hankonen et al.,C. Greaves,,,https://academic.oup.com/abm/article-pdf/49/1/1/21983123/12160_2014_article_9630.pdf
-The Flan Collection: Designing Data and Methods for Effective Instruction Tuning,S. Longpre,"We study the design decisions of publicly available instruction tuning methods, and break down the development of Flan 2022 (Chung et al., 2022). Through careful ablation studies on the Flan Collection of tasks and methods, we tease apart the effect of design decisions which enable Flan-T5 to outperform prior work by 3-17%+ across evaluation settings. We find task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings (zero-shot, few-shot, and chain-of-thought) actually yields stronger (2%+) performance in all settings. In further experiments, we show Flan-T5 requires less finetuning to converge higher and faster than T5 on single downstream tasks, motivating instruction-tuned models as more computationally-efficient starting checkpoints for new tasks. Finally, to accelerate research on instruction tuning, we make the Flan 2022 collection of datasets, templates, and methods publicly available at https://github.com/google-research/FLAN/tree/main/flan/v2.","{'model': 'tldr@v2.0.0', 'text': 'It is found task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings actually yields stronger performance in all settings.'}",http://arxiv.org/pdf/2301.13688
-Identification of behaviour change techniques to reduce excessive alcohol consumption.,S. Michie,"BACKGROUND Interventions to reduce excessive alcohol consumption have a small but important effect, but a better understanding is needed of their 'active ingredients'. AIMS This study aimed to (i) develop a reliable taxonomy of behaviour change techniques (BCTs) used in interventions to reduce excessive alcohol consumption (not to treat alcohol dependence) and (ii) to assess whether use of specific BCTs in brief interventions might be associated with improved effectiveness. METHODS A selection of guidance documents and treatment manuals, identified via expert consultation, were analysed into BCTs by two coders. The resulting taxonomy of BCTs was applied to the Cochrane Review of brief alcohol interventions, and the associations between the BCTs and effectiveness were investigated using meta-regression. FINDINGS Forty-two BCTs were identified, 34 from guidance documents and an additional eight from treatment manuals, with average inter-rater agreement of 80%. Analyses revealed that brief interventions that included the BCT 'prompt self-recording' (P = 0.002) were associated with larger effect sizes. CONCLUSIONS It is possible to identify specific behaviour change techniques reliably in manuals and guidelines for interventions to reduce excessive alcohol consumption. In brief interventions, promoting self-monitoring is associated with improved outcomes. More research is needed to identify other behaviour change techniques or groupings of behaviour change techniques that can produce optimal results in brief interventions and to extend the method to more intensive interventions and treatment of alcohol dependence.","{'model': 'tldr@v2.0.0', 'text': 'It is possible to identify specific behaviour change techniques reliably in manuals and guidelines for interventions to reduce excessive alcohol consumption and, in brief interventions, promoting self-monitoring is associated with improved outcomes.'}",
-The Use of Neutron Analysis Techniques for Detecting The Concentration And Distribution of Chloride Ions in Archaeological Iron,D. Watkinson,"Chloride (Cl) ions diffuse into iron objects during burial and drive corrosion after excavation. Located under corrosion layers, Cl is inaccessible to many analytical techniques. Neutron analysis offers non-destructive avenues for determining Cl content and distribution in objects. A pilot study used prompt gamma activation analysis (PGAA) and prompt gamma activation imaging (PGAI) to analyse the bulk concentration and longitudinal distribution of Cl in archaeological iron objects. This correlated with the object corrosion rate measured by oxygen consumption, and compared well with Cl measurement using a specific ion meter. High-Cl areas were linked with visible damage to the corrosion layers and attack of the iron core. Neutron techniques have significant advantages in the analysis of archaeological metals, including penetration depth and low detection limits.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study used prompt gamma activation analysis (PGAA) and prompt gammaactivation imaging (PGAI) to analyse the bulk concentration and longitudinal distribution of Cl in archaeological iron objects and found high-Cl areas were linked with visible damage to the corrosion layers and attack of the iron core.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/arcm.12058
-Can language models learn from explanations in context?,Andrew Kyle Lampinen,"Language Models (LMs) can perform new tasks by adapting to a few in-context examples. For humans, explanations that connect examples to task principles can improve learning. We therefore investigate whether explanations of few-shot examples can help LMs. We annotate questions from 40 challenging tasks with answer explanations, and various matched control explanations. We evaluate how different types of explanations, instructions, and controls affect zero- and few-shot performance. We analyze these results using statistical multilevel modeling techniques that account for the nested dependencies among conditions, tasks, prompts, and models. We find that explanations can improve performance -- even without tuning. Furthermore, explanations hand-tuned for performance on a small validation set offer substantially larger benefits, and building a prompt by selecting examples and explanations together substantially improves performance over selecting examples alone. Finally, even untuned explanations outperform carefully matched controls, suggesting that the benefits are due to the link between an example and its explanation, rather than lower-level features. However, only large models benefit. In summary, explanations can support the in-context learning of large LMs on challenging tasks.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2204.02329
-PUCK: an automated prompting system for smart environments: toward achieving automated prompting—challenges involved,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'This paper uses simple off-the-shelf sensors and learns the timing for prompts based on real data that are collected with volunteer participants in the smart home test bed and proposes a variant of an existing sampling technique, SMOTE, to deal with the class imbalance problem.'}",https://europepmc.org/articles/pmc4215554?pdf=render
-"Concepts, Instrumentation and Techniques of Neutron Activation Analysis",H. Slamene,"Following the discovery of neutron by J. Chadwick in 1932 (Nobel prize, 1935) and the re‐ sults of F. Joliot and I. Curie in 1934, neutron activation analysis was first developed by G. Hevesy and H. Levi in 1936. They used a neutron source (226Ra + Be) and a radiation detec‐ tor (ionization chamber) and promptly recognized that the element Dy (dysprosium) in the sample became highly radioactive after exposure to the neutron source. They showed that the nuclear reaction may be used to determine the elements present in unknown samples by measuring the induced radioactivity.",,
-Learning to Compose Soft Prompts for Compositional Zero-Shot Learning,Nihal V. Nayak,"We introduce compositional soft prompting (CSP), a parameter-efficient learning technique to improve the zero-shot compositionality of large-scale pretrained vision-language models (VLMs) like CLIP. We develop CSP for compositional zero-shot learning, the task of predicting unseen attribute-object compositions (e.g., old cat and young tiger). VLMs have a flexible text encoder that can represent arbitrary classes as natural language prompts but they often underperform task-specific architectures on the compositional zero-shot benchmark datasets. CSP treats the attributes and objects that define classes as learnable tokens of vocabulary. During training, the vocabulary is tuned to recognize classes that compose tokens in multiple ways (e.g., old cat and white cat). At test time, we recompose the learned attribute-object vocabulary in new combinations to recognize novel classes. We show that CSP outperforms the CLIP on benchmark datasets by an average of 10.9 percentage points on AUC. CSP also outperforms CoOp, a soft prompting method that fine-tunes the prefix context tokens, by an average of 5.8 percentage points on AUC. We perform additional experiments to show that CSP improves generalization to higher-order attribute-attribute-object compositions (e.g., old white cat) and combinations of pretrained attributes and fine-tuned objects. The code is available at https://github.com/BatsResearch/csp.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2204.03574
-Generating Training Data with Language Models: Towards Zero-Shot Language Understanding,Yu Meng,"Pretrained language models (PLMs) have demonstrated remarkable performance in various natural language processing tasks: Unidirectional PLMs (e.g., GPT) are well known for their superior text generation capabilities; bidirectional PLMs (e.g., BERT) have been the prominent choice for natural language understanding (NLU) tasks. While both types of models have achieved promising few-shot learning performance, their potential for zero-shot learning has been underexplored. In this paper, we present a simple approach that uses both types of PLMs for fully zero-shot learning of NLU tasks without requiring any task-specific data: A unidirectional PLM generates class-conditioned texts guided by prompts, which are used as the training data for fine-tuning a bidirectional PLM. With quality training data selected based on the generation probability and regularization techniques (label smoothing and temporal ensembling) applied to the fine-tuning stage for better generalization and stability, our approach demonstrates strong performance across seven classification tasks of the GLUE benchmark (e.g., 72.3/73.8 on MNLI-m/mm and 92.8 on SST-2), significantly outperforming zero-shot prompting methods and achieving even comparable results to strong few-shot approaches using 32 training samples per class.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple approach that uses both types of PLMs for fully zero-shot learning of NLU tasks without requiring any task-specific data: a unidirectional PLM generates class-conditioned texts guided by prompts, which are used as the training data for fine-tuning a bidirectionalPLM.'}",
-Language Model Cascades,David Dohan,"Prompted models have demonstrated impressive few-shot learning abilities. Repeated interactions at test-time with a single model, or the composition of multiple models together, further expands capabilities. These compositions are probabilistic models, and may be expressed in the language of graphical models with random variables whose values are complex data types such as strings. Cases with control flow and dynamic structure require techniques from probabilistic programming, which allow implementing disparate model structures and inference strategies in a unified language. We formalize several existing techniques from this perspective, including scratchpads / chain of thought, verifiers, STaR, selection-inference, and tool use. We refer to the resulting programs as language model cascades.","{'model': 'tldr@v2.0.0', 'text': 'This work formalizes several existing techniques from probabilistic programming, including scratchpads / chain of thought, verifiers, STaR, selection-inference, and tool use, and refers to the resulting programs as language model cascades.'}",http://arxiv.org/pdf/2207.10342
-Simultaneous Prompting: An Instructional Strategy for Skill Acquisition,R. E. Waugh,"Errorless learning is an instructional approach designed to eliminate and/or reduce the number of errors students produce in traditional trial-and-error approaches (Mueller, Palkovic, & Maynard, 2007). Various response prompting strategies are employed to produce errorless learning. Simultaneous prompting is an errorless learning strategy that has a growing body of literature to support its use spanning two decades. This paper provides a comprehensive review of the literature including (a) skills targeted for instruction, (b) populations targeted for instruction, (c) strengths and weaknesses of simultaneous prompting, and (d) future areas of research. Errorless learning is an instructional ap- proach designed to reduce the number of errors students make in traditional trial-and- error approaches (Mueller, Palkovic, & May- nard, 2007). During errorless learning proce- dures stimulus control is transferred from the controlling prompt to the discriminative stim- ulus. This transfer is achieved through the use of response prompting strategies. Response prompting strategies consist of additional in- formation which results in the correct re- sponse being emitted (Wolery, Ault, & Doyle, 1992). The underlying purpose of errorless learning is the transfer of stimulus control from a response prompt to the natural stimu- lus. Wolery and Gast (1984) identified four common response prompting strategies that commonly are employed to transfer stimulus control: (a) most-to-least prompts, (b) least-to- most prompts, (c) graduated guidance, and (d) time delay. Most-to-least prompts consists of employing the most intrusive prompt needed to assist the student in emitting the correct response and fading the intensity of the prompt until the student is correctly re- sponding to the discriminative stimulus inde- pendently. Least-to-most prompts provide the student with an opportunity to respond inde- pendently to the discriminative stimulus. If the student responds incorrectly a prompt is provided which gradually increases in inten- sity until the student responds correctly to the discriminative stimulus. ""Graduated guidance is a technique combining physical guidance and fading in which the physical guidance is systematically and gradually reduced and then faded completely"" (Foxx, 1982, p. 129). Grad- uate guidance relies heavily on the teacher's judgment whether or not a prompt is required or the degree of prompt required. There are two forms of graduated guidance. During one form a teacher shadows a student's movement when teaching a task to provide guidance dur- ing each step or to remove the physical prompt. During a second form of graduated guidance the teacher may provide constant contact but gradually and systematically re- duce the intrusiveness and placement of the prompt (Foxx, 1981; Wolery & Gast, 1984). Time delay is the fourth common response prompting strategy which results in near er- rorless learning by transferring stimulus con- trol from a controlling prompt to the discrim- inative stimulus by inserting a delay between the presentation of the discriminative stimu- lus and the controlling prompt (Snell & Gast, 1981; Touchette, 1971). Two forms of time delay are reported in the literature, progres- sive time delay (PTD) and constant time delay (CTD). During PTD a systematically increased",,
-Interventions that involve parents to improve children's weight‐related nutrition intake and activity patterns – what nutrition and activity targets and behaviour change techniques are associated with intervention effectiveness?,R. Golley,"Parent involvement is an important component of obesity prevention interventions. However, the best way to support parents remains unclear. This review identifies interventions targeting parents to improve children's weight status, dietary and/or activity patterns, examines whether intervention content and behaviour change techniques employed are associated with effectiveness. Seventeen studies, in English, 1998–2008, were included. Studies were evaluated by two reviewers for study quality, nutrition/activity content and behaviour change techniques using a validated quality assessment tool and behaviour change technique taxonomy. Study findings favoured intervention effectiveness in 11 of 17 studies. Interventions that were considered effective had similar features: better study quality, parents responsible for participation and implementation, greater parental involvement and inclusion of prompt barrier identification, restructure the home environment, prompt self‐monitoring, prompt specific goal setting behaviour change techniques. Energy intake/density and food choices were more likely to be targeted in effective interventions. The number of lifestyle behaviours targeted did not appear to be associated with effectiveness. Intervention effectiveness was favoured when behaviour change techniques spanned the spectrum of behaviour change process. The review provides guidance for researchers to make informed decisions on how best to utilize resources in interventions to support and engage parents, and highlights a need for improvement in intervention content reporting practices.","{'model': 'tldr@v2.0.0', 'text': ""Interventions targeting parents to improve children's weight status, dietary and/or activity patterns, and whether intervention content and behaviour change techniques employed are associated with effectiveness are examined.""}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1467-789X.2010.00745.x
-"Equine castration: A review of techniques, complications and their management",I. Kilcoyne,"Summary Complications associated with castration occur commonly and, although the majority are mild and resolve easily, potentially life-threatening complications can occur. The preoperative identification of risk factors for these complications can help the veterinarian to take the appropriate measures to reduce these risks. However, even with proper surgical technique, complications can occur. Therefore prompt recognition and initiation of appropriate therapy are essential to prevent further morbidity, death or lawsuits.","{'model': 'tldr@v2.0.0', 'text': 'Preoperative identification of risk factors for castration complications can help the veterinarian to take the appropriate measures to reduce these risks, although the majority are mild and resolve easily.'}",
-Using Video Prompting and Constant Time Delay to Teach an Internet Search Basic Skill to Students with Intellectual Disabilities,D. Zisimopoulos,"We evaluated a video prompting and a constant time delay procedure for teaching three primary school students with moderate intellectual disabilities to access the Internet and download pictures related to participation in a classroom History project. Video clips were used as an antecedent prompt and as an error correction technique within a constant time delay (CTD) procedure. Training, using a 29-step task analysis, was conducted in the students’ special education classrooms. The prompting procedure was introduced in a multiple baseline across subjects design. Video prompting was effective in promoting rapid acquisition of the task for all three students. Following acquisition, observations suggested that the new skill was further generalized to another trainer, novel materials, and to another setting. In a final phase then, video prompting was removed and maintenance in the absence of video prompting was assessed at 1, 3, and 18 weeks. Performance decreased at the last follow-up session, but stabilized at 82.7–89.6% correct. These data suggest that video prompting may be an effective instructional strategy for teaching Internet skills to students with moderate intellectual disabilities.",,
-Towards Personalized Federated Learning,A. Tan,"In parallel with the rapid adoption of artificial intelligence (AI) empowered by advances in AI research, there has been growing awareness and concerns of data privacy. Recent significant developments in the data regulation landscape have prompted a seismic shift in interest toward privacy-preserving AI. This has contributed to the popularity of Federated Learning (FL), the leading paradigm for the training of machine learning models on data silos in a privacy-preserving manner. In this survey, we explore the domain of personalized FL (PFL) to address the fundamental challenges of FL on heterogeneous data, a universal characteristic inherent in all real-world datasets. We analyze the key motivations for PFL and present a unique taxonomy of PFL techniques categorized according to the key challenges and personalization strategies in PFL. We highlight their key ideas, challenges, opportunities, and envision promising future trajectories of research toward a new PFL architectural design, realistic PFL benchmarking, and trustworthy PFL approaches.","{'model': 'tldr@v2.0.0', 'text': 'This survey explores the domain of personalized FL (PFL) to address the fundamental challenges of FL on heterogeneous data, a universal characteristic inherent in all real-world datasets.'}",https://arxiv.org/pdf/2103.00710
-Sketch-Guided Text-to-Image Diffusion Models,A. Voynov,"Text-to-Image models have introduced a remarkable leap in the evolution of machine learning, demonstrating high-quality synthesis of images from a given text-prompt. However, these powerful pretrained models still lack control handles that can guide spatial properties of the synthesized images. In this work, we introduce a universal approach to guide a pretrained text-to-image diffusion model, with a spatial map from another domain (e.g., sketch) during inference time. Unlike previous works, our method does not require to train a dedicated model or a specialized encoder for the task. Our key idea is to train a Latent Guidance Predictor (LGP) - a small, per-pixel, Multi-Layer Perceptron (MLP) that maps latent features of noisy images to spatial maps, where the deep features are extracted from the core Denoising Diffusion Probabilistic Model (DDPM) network. The LGP is trained only on a few thousand images and constitutes a differential guiding map predictor, over which the loss is computed and propagated back to push the intermediate images to agree with the spatial map. The per-pixel training offers flexibility and locality which allows the technique to perform well on out-of-domain sketches, including free-hand style drawings. We take a particular focus on the sketch-to-image translation task, revealing a robust and expressive way to generate images that follow the guidance of a sketch of arbitrary style or domain.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a universal approach to guide a pretrained text-to-image diffusion model, with a spatial map from another domain (e.g., sketch) during inference time, and trains a Latent Guidance Predictor, a small, per-pixel, Multi-Layer Perceptron that maps latent features of noisy images to spatial maps.'}",https://arxiv.org/pdf/2211.13752
-Factual Probing Is [MASK]: Learning vs. Learning to Recall,Zexuan Zhong,"Petroni et al. (2019) demonstrated that it is possible to retrieve world facts from a pre-trained language model by expressing them as cloze-style prompts and interpret the model’s prediction accuracy as a lower bound on the amount of factual information it encodes. Subsequent work has attempted to tighten the estimate by searching for better prompts, using a disjoint set of facts as training data. In this work, we make two complementary contributions to better understand these factual probing techniques. First, we propose OptiPrompt, a novel and efficient method which directly optimizes in continuous embedding space. We find this simple method is able to predict an additional 6.4% of facts in the LAMA benchmark. Second, we raise a more important question: Can we really interpret these probing results as a lower bound? Is it possible that these prompt-search methods learn from the training data too? We find, somewhat surprisingly, that the training data used by these methods contains certain regularities of the underlying fact distribution, and all the existing prompt methods, including ours, are able to exploit them for better fact prediction. We conduct a set of control experiments to disentangle “learning” from “learning to recall”, providing a more detailed picture of what different prompts can reveal about pre-trained language models.","{'model': 'tldr@v2.0.0', 'text': 'OptiPrompt is proposed, a novel and efficient method which directly optimizes in continuous embedding space and is able to predict an additional 6.4% of facts in the LAMA benchmark.'}",https://aclanthology.org/2021.naacl-main.398.pdf
-"Using ChatGPT Standard Prompt Engineering Techniques in Lesson Preparation: Role, Instructions and Seed-Word Prompts",A. Spasic,The application of available natural language processing systems can have a significant impact on the education process. The primary aim of this research was to test the impact of three standard prompting techniques on the results obtained from ChatGPT. Generation of a lesson plan for programming for preschoolers was chosen as the task set for AI. The obtained results show that use of a standard prompting with additional defined roles and seed words can be useful in preparation of teaching units and lessons and it can be considered as a technique of teachers' choice.,,
-Generative AI tools in art education: Exploring prompt engineering and iterative processes for enhanced creativity,Peter Cotroneo,"The rapid development and adoption of generative artificial intelligence (AI) tools in the art and design education landscape have introduced both opportunities and challenges. This timely study addresses the need to effectively integrate these tools into the classroom while considering ethical implications and the importance of prompt engineering. By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context. Students in the digital at case study were taught prompt engineering techniques and were tasked with crafting multiple prompts, focusing on refining their ideas over time. Participants demonstrated an increased understanding of the potential and limitations of generative AI tools and how to manipulate subject matter for more effective results. The iterative process encouraged students to explore and experiment with their creative ideas, leading to a deeper understanding of the possibilities offered by AI tools. Despite acknowledging the ethical concerns regarding copyright and the potential replacement of artists, students appreciated the value of generative AI tools for enhancing their sketchbooks and ideation process. Through prompt engineering and iterative processes, students developed a more detail-oriented approach to their work. The challenge of using AI-generated images as final products was conceptually intriguing, requiring further investigation and consideration of the prompts. This study highlights the potential benefits and challenges of integrating generative AI tools into art and design classrooms, emphasizing the importance of prompt engineering, iterative processes, and ethical considerations as these technologies continue to evolve.",,
-How does prompt engineering affect ChatGPT performance on unsupervised entity resolution?,Khanin Sisaengsuwanchai,"Entity Resolution (ER) is the problem of semi-automatically determining when two entities refer to the same underlying entity, with applications ranging from healthcare to e-commerce. Traditional ER solutions required considerable manual expertise, including feature engineering, as well as identification and curation of training data. In many instances, such techniques are highly dependent on the domain. With recent advent in large language models (LLMs), there is an opportunity to make ER much more seamless and domain-independent. However, it is also well known that LLMs can pose risks, and that the quality of their outputs can depend on so-called prompt engineering. Unfortunately, a systematic experimental study on the effects of different prompting methods for addressing ER, using LLMs like ChatGPT, has been lacking thus far. This paper aims to address this gap by conducting such a study. Although preliminary in nature, our results show that prompting can significantly affect the quality of ER, although it affects some metrics more than others, and can also be dataset dependent.","{'model': 'tldr@v2.0.0', 'text': 'The results show that prompting can significantly affect the quality of ER, although it affects some metrics more than others, and can also be dataset dependent.'}",https://arxiv.org/pdf/2310.06174
-Spectral features based convolutional neural network for accurate and prompt identification of schizophrenic patients,Kuldeep Singh,"Schizophrenia is a fatal mental disorder, which affects millions of people globally by the disturbance in their thinking, feeling and behaviour. In the age of the internet of things assisted with cloud computing and machine learning techniques, the computer-aided diagnosis of schizophrenia is essentially required to provide its patients with an opportunity to own a better quality of life. In this context, the present paper proposes a spectral features based convolutional neural network (CNN) model for accurate identification of schizophrenic patients using spectral analysis of multichannel EEG signals in real-time. This model processes acquired EEG signals with filtering, segmentation and conversion into frequency domain. Then, given frequency domain segments are divided into six distinct spectral bands like delta, theta-1, theta-2, alpha, beta and gamma. The spectral features including mean spectral amplitude, spectral power and Hjorth descriptors (Activity, Mobility and Complexity) are extracted from each band. These features are independently fed to the proposed spectral features-based CNN and long short-term memory network (LSTM) models for classification. This work also makes use of raw time-domain and frequency-domain EEG segments for classification using temporal CNN and spectral CNN models of same architectures respectively. The overall analysis of simulation results of all models exhibits that the proposed spectral features based CNN model is an efficient technique for accurate and prompt identification of schizophrenic patients among healthy individuals with average classification accuracies of 94.08% and 98.56% for two different datasets with optimally small classification time.","{'model': 'tldr@v2.0.0', 'text': 'The overall analysis of simulation results of all models exhibits that the proposed spectral features based CNN model is an efficient technique for accurate and prompt identification of schizophrenic patients among healthy individuals with average classification accuracies of 94.08% and 98.56% for two different datasets with optimally small classification time.'}",
-A Practical Survey on Zero-Shot Prompt Design for In-Context Learning,Yinheng Li,"The remarkable advancements in large language models (LLMs) have brought about significant improvements in Natural Language Processing(NLP) tasks. This paper presents a comprehensive review of in-context learning techniques, focusing on different types of prompts, including discrete, continuous, few-shot, and zero-shot, and their impact on LLM performance. We explore various approaches to prompt design, such as manual design, optimization algorithms, and evaluation methods, to optimize LLM performance across diverse tasks. Our review covers key research studies in prompt engineering, discussing their methodologies and contributions to the field. We also delve into the challenges faced in evaluating prompt performance, given the absence of a single “best” prompt and the importance of considering multiple metrics. In conclusion, the paper highlights the critical role of prompt design in harnessing the full potential of LLMs and provides insights into the combination of manual design, optimization techniques, and rigorous evaluation for more effective and efficient use of LLMs in various NLP tasks.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive review of in-context learning techniques, focusing on different types of prompts, including discrete, continuous, few- shot, and zero-shot, and their impact on LLM performance, is presented.'}",https://arxiv.org/pdf/2309.13205
-Multimodal Propaganda Detection Via Anti-Persuasion Prompt enhanced contrastive learning,Jian Cui,"Propaganda, commonly used in memes disinformation, can influence the thinking of the audience and increase the reach of communication. Usually logical fallacy, as a kind of popular expression of memes, aims to create a logical reasonable illusion where the conclusion cannot be drawn with the use of correct logic rules. However, this characteristic inherent in memes leads to difficulties for classic multi-label classifiers to understand propagation techniques. To this end, we propose a novel propaganda detection model called Antipersuasion Prompt Enhanced Contrastive Learning (abbreviated as APCL). First, our APCL reformulates the multi-label classification task by leveraging the category words of propaganda technique based prompt engineering, which is converted into an image-text matching (ITM). Second, prompt engineering is designed with a persuasion prompt template and an anti-persuasion prompt template. The former is to build matched text-image pairs, and the latter is to form mismatched text-image pairs which fit the logical fallacy style of memes. Finally, the propagation technique is predicted based on the distances between the above two prompt templates enhanced texts and an input image. Experimental results on the memes dataset of SemEval2021 task 6 show that our APCL outperforms the state-of-the-art multimodal classification models in terms of F1 measures.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on the memes dataset of SemEval2021 task 6 show that the proposed Antipersuasion Prompt Enhanced Contrastive Learning outperforms the state-of-the-art multimodal classification models in terms of F1 measures.'}",
-User-friendly Image Editing with Minimal Text Input: Leveraging Captioning and Injection Techniques,Sunwoo Kim,"Recent text-driven image editing in diffusion models has shown remarkable success. However, the existing methods assume that the user's description sufficiently grounds the contexts in the source image, such as objects, background, style, and their relations. This assumption is unsuitable for real-world applications because users have to manually engineer text prompts to find optimal descriptions for different images. From the users' standpoint, prompt engineering is a labor-intensive process, and users prefer to provide a target word for editing instead of a full sentence. To address this problem, we first demonstrate the importance of a detailed text description of the source image, by dividing prompts into three categories based on the level of semantic details. Then, we propose simple yet effective methods by combining prompt generation frameworks, thereby making the prompt engineering process more user-friendly. Extensive qualitative and quantitative experiments demonstrate the importance of prompts in text-driven image editing and our method is comparable to ground-truth prompts.","{'model': 'tldr@v2.0.0', 'text': 'This work demonstrates the importance of a detailed text description of the source image, by dividing prompts into three categories based on the level of semantic details, and proposes simple yet effective methods by combining prompt generation frameworks, thereby making the prompt engineering process more user-friendly.'}",http://arxiv.org/pdf/2306.02717
-ASCM: An Answer Space Clustered Prompting Method without Answer Engineering,Zhen Wang,"Prompt-based learning, which exploits knowledge from pre-trained language models by providing textual prompts and designing appropriate answer-category mapping methods, has achieved impressive successes on few-shot text classification and natural language inference (NLI). Because of the diverse linguistic expression, there exist many answer tokens for the same category. However, both manual answer design and automatic answer search constrain answer space and therefore hardly achieve ideal performance. To address this issue, we propose an answer space clustered prompting model (ASCM) together with a synonym initialization method (SI) which automatically categorizes all answer tokens in a semantic-clustered embedding space. We also propose a stable semi-supervised method named stair learning (SL) that orderly distills knowledge from better models to weaker models. Extensive experiments demonstrate that our ASCM+SL significantly outperforms existing state-of-the-art techniques in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an answer space clustered prompting model (ASCM) together with a synonym initialization method (SI) which automatically categorizes all answer tokens in a semantic-clustered embedding space and proposes a stable semi-supervised method named stair learning (SL) that orderly distills knowledge from better models to weaker models.'}",https://aclanthology.org/2022.findings-acl.193.pdf
-Troubleshooting Techniques for Cerebral Aneurysm Surgery Based on “the Study of Failure”,S. Takeuchi,"In this study, we analyzed the causes of problems encountered during aneurysm surgery based on “the study of failure,” which was originally devised for system engineering. We describe four problematic cases, which were all successfully managed by troubleshooting techniques. The majority of the problems (failures) were caused by the surgeon’s “carelessness and/or decision error”. Large vessel injury during aneurysm dissection is formidable but can be managed by troubleshooting techniques such as micro-suturing or a bypass procedure in the deep operative field. Prompt and secure micro-anastomotic suturing is one of the vital troubleshooting techniques during aneurysm surgery. Personal preparation of micro-suturing instruments and daily off-the-job training are essential to master such troubleshooting procedures.","{'model': 'tldr@v2.0.0', 'text': 'The causes of problems encountered during aneurysm surgery are analyzed based on “the study of failure,” which was originally devised for system engineering and described four problematic cases, which were successfully managed by troubleshooting techniques.'}",https://www.jstage.jst.go.jp/article/scs/48/2/48_96/_pdf
-Automatic Code Summarization via ChatGPT: How Far Are We?,Weisong Sun,"To support software developers in understanding and maintaining programs, various automatic code summarization techniques have been proposed to generate a concise natural language comment for a given code snippet. Recently, the emergence of large language models (LLMs) has led to a great boost in the performance of natural language processing tasks. Among them, ChatGPT is the most popular one which has attracted wide attention from the software engineering community. However, it still remains unclear how ChatGPT performs in (automatic) code summarization. Therefore, in this paper, we focus on evaluating ChatGPT on a widely-used Python dataset called CSN-Python and comparing it with several state-of-the-art (SOTA) code summarization models. Specifically, we first explore an appropriate prompt to guide ChatGPT to generate in-distribution comments. Then, we use such a prompt to ask ChatGPT to generate comments for all code snippets in the CSN-Python test set. We adopt three widely-used metrics (including BLEU, METEOR, and ROUGE-L) to measure the quality of the comments generated by ChatGPT and SOTA models (including NCS, CodeBERT, and CodeT5). The experimental results show that in terms of BLEU and ROUGE-L, ChatGPT's code summarization performance is significantly worse than all three SOTA models. We also present some cases and discuss the advantages and disadvantages of ChatGPT in code summarization. Based on the findings, we outline several open challenges and opportunities in ChatGPT-based code summarization.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating ChatGPT on a widely-used Python dataset called CSN-Python and comparing it with several state-of-the-art (SOTA) code summarization models shows that in terms of BLEU and ROUGE-L,ChatGPT's code summarizing performance is significantly worse than all three SOTA models.""}",http://arxiv.org/pdf/2305.12865
-Game of Tones: Faculty detection of GPT-4 generated content in university assessments,Mike Perkins,,,https://arxiv.org/pdf/2305.18081
-Do LLMs Possess a Personality? Making the MBTI Test an Amazing Evaluation for Large Language Models,Keyu Pan,"The field of large language models (LLMs) has made significant progress, and their knowledge storage capacity is approaching that of human beings. Furthermore, advanced techniques, such as prompt learning and reinforcement learning, are being employed to address ethical concerns and hallucination problems associated with LLMs, bringing them closer to aligning with human values. This situation naturally raises the question of whether LLMs with human-like abilities possess a human-like personality? In this paper, we aim to investigate the feasibility of using the Myers-Briggs Type Indicator (MBTI), a widespread human personality assessment tool, as an evaluation metric for LLMs. Specifically, extensive experiments will be conducted to explore: 1) the personality types of different LLMs, 2) the possibility of changing the personality types by prompt engineering, and 3) How does the training dataset affect the model's personality. Although the MBTI is not a rigorous assessment, it can still reflect the similarity between LLMs and human personality. In practice, the MBTI has the potential to serve as a rough indicator. Our codes are available at https://github.com/HarderThenHarder/transformers_tasks/tree/main/LLM/llms_mbti.",,https://arxiv.org/pdf/2307.16180
-Controlled and Conditional Text to Image Generation with Diffusion Prior,Pranav Aggarwal,"Denoising Diffusion models have shown remarkable performance in generating diverse, high quality images from text. Numerous techniques have been proposed on top of or in alignment with models like Stable Diffusion and Imagen that generate images directly from text. A lesser explored approach is DALLE-2's two step process comprising a Diffusion Prior that generates a CLIP image embedding from text and a Diffusion Decoder that generates an image from a CLIP image embedding. We explore the capabilities of the Diffusion Prior and the advantages of an intermediate CLIP representation. We observe that Diffusion Prior can be used in a memory and compute efficient way to constrain the generation to a specific domain without altering the larger Diffusion Decoder. Moreover, we show that the Diffusion Prior can be trained with additional conditional information such as color histogram to further control the generation. We show quantitatively and qualitatively that the proposed approaches perform better than prompt engineering for domain specific generation and existing baselines for color conditioned generation. We believe that our observations and results will instigate further research into the diffusion prior and uncover more of its capabilities.","{'model': 'tldr@v2.0.0', 'text': 'The Diffusion Prior can be used in a memory and compute efficient way to constrain the generation to a specific domain without altering the larger Diffusion Decoder and can be trained with additional conditional information such as color histogram to further control the generation.'}",https://arxiv.org/pdf/2302.11710
-A Comparative Study of Damage-Sensitive Features for Rapid Data-Driven Seismic Structural Health Monitoring,Y. Reuland,"Rapid post-earthquake damage assessment forms a critical element of resilience, ensuring a prompt and functional recovery of the built environment. Monitoring-based approaches have the potential to significantly improve upon current visual inspection-based condition assessment that is slow and potentially subjective. The large variety of sensing solutions that has become available at affordable cost in recent years allows the engineering community to envision permanent-monitoring applications even in conventional low-to-mid-rise buildings. When combined with adequate structural health monitoring (SHM) techniques, sensor data recorded during earthquakes have the potential to provide automated near-real-time identification of earthquake damage. Near-real time building assessment relies on the tracking of damage-sensitive features (DSFs) that can be directly and rapidly derived from dynamic monitoring data and scaled with damage. We here offer a comprehensive review of such damage-sensitive features in an effort to formally assess the capacity of such data-driven indicators to detect, localize and quantify the presence of nonlinearity in seismic-induced structural response. We employ both a parametric analysis on a simulated model and real data from shake-table tests to investigate the strengths and limitations of purely data-driven approaches, which typically involve a comparison against a healthy reference state. We present an array of damage-sensitive features which are found to be robust with respect to noise, to reliably detect and scale with nonlinearity, and to carry potential to localize the occurrence of nonlinear behavior in conventional structures undergoing earthquakes.",,https://www.mdpi.com/2076-3417/13/4/2708/pdf?version=1676887837
-The application of ChatGPT in healthcare progress notes: A commentary from a clinical and research perspective,Josh Nguyen,"ChatGPT, powered by one of the most advanced language processing systems, gained over 100 million users in just 2 months following its release in November 2022.1 This unprecedented popularity is likely due to its wide range of potential applications in fields, such as engineering, education and healthcare.2–4 The integration of artificial intelligence (AI)—driven language models like ChatGPT has the potential to revolutionize documentation practices, streamline workflows, and ultimately lead to more efficient and patient-centred care,2 though the use of such tools is not without its challenges. Here, we outline the potential benefits and pitfalls of implementing AI-driven language models, such as ChatGPT, in the creation and management of healthcare progress notes using prompt engineering techniques. We provide recommendations for responsible and effective integration into clinical practice and priorities for future research. Healthcare clinicians spend 35% of their time documenting patient data, and evidence suggests the length of healthcare case notes has been increasing over time.5 Existing innovations, such as speech recognition technology, yield no clear benefit in timesaving or documentation quality.6 With the ability to coherently write logical and accurate text within a few seconds, ChatGPT has the potential to reduce time spent on tasks such as preparing healthcare progress notes, and might also enhance","{'model': 'tldr@v2.0.0', 'text': 'The potential benefits and pitfalls of implementing AI-driven language models, such as ChatGPT, in the creation and management of healthcare progress notes using prompt engineering techniques are outlined and recommendations for responsible and effective integration into clinical practice are provided.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ctm2.1324
-"Augmented Behavioral Annotation Tools, with Application to Multimodal Datasets and Models: A Systematic Review",Eleanor Watson,"Annotation tools are an essential component in the creation of datasets for machine learning purposes. Annotation tools have evolved greatly since the turn of the century, and now commonly include collaborative features to divide labor efficiently, as well as automation employed to amplify human efforts. Recent developments in machine learning models, such as Transformers, allow for training upon very large and sophisticated multimodal datasets and enable generalization across domains of knowledge. These models also herald an increasing emphasis on prompt engineering to provide qualitative fine-tuning upon the model itself, adding a novel emerging layer of direct machine learning annotation. These capabilities enable machine intelligence to recognize, predict, and emulate human behavior with much greater accuracy and nuance, a noted shortfall of which have contributed to algorithmic injustice in previous techniques. However, the scale and complexity of training data required for multimodal models presents engineering challenges. Best practices for conducting annotation for large multimodal models in the most safe and ethical, yet efficient, manner have not been established. This paper presents a systematic literature review of crowd and machine learning augmented behavioral annotation methods to distill practices that may have value in multimodal implementations, cross-correlated across disciplines. Research questions were defined to provide an overview of the evolution of augmented behavioral annotation tools in the past, in relation to the present state of the art. (Contains five figures and four tables).","{'model': 'tldr@v2.0.0', 'text': 'A systematic literature review of crowd and machine learning augmented behavioral annotation methods to distill practices that may have value in multimodal implementations, cross-correlated across disciplines is presented.'}",https://www.mdpi.com/2673-2688/4/1/7/pdf?version=1674957180
-Copilot for Xcode: Exploring AI-Assisted Programming by Prompting Cloud-based Large Language Models,C. Tan,"This paper presents an AI-assisted programming tool called Copilot for Xcode for program composition and design to support human software developers. By seamlessly integrating cloud-based Large Language Models (LLM) with Apple's local development environment, Xcode, this tool enhances productivity and unleashes creativity for software development in Apple software ecosystem (e.g., iOS apps, macOS). Leveraging advanced natural language processing (NLP) techniques, Copilot for Xcode effectively processes source code tokens and patterns within code repositories, enabling features such as code generation, autocompletion, documentation, and error detection. Software developers can also query and make""small""decisions for program composition, some of which can be made simultaneously, and this is facilitated through prompt engineering in a chat interface of Copilot for Xcode. Finally, we present simple case studies as evidence of the effectiveness of utilizing NLP in Xcode to prompt popular LLM services like OpenAI ChatGPT for program composition and design.","{'model': 'tldr@v2.0.0', 'text': ""An AI-assisted programming tool called Copilot for Xcode for program composition and design to support human software developers, seamlessly integrating cloud-based Large Language Models with Apple's local development environment, Xcode, enhances productivity and unleashes creativity for software development in Apple software ecosystem.""}",https://arxiv.org/pdf/2307.14349
-Towards Equitable Representation in Text-to-Image Synthesis Models with the Cross-Cultural Understanding Benchmark (CCUB) Dataset,Zhixuan Liu,"It has been shown that accurate representation in media improves the well-being of the people who consume it. By contrast, inaccurate representations can negatively affect viewers and lead to harmful perceptions of other cultures. To achieve inclusive representation in generated images, we propose a culturally-aware priming approach for text-to-image synthesis using a small but culturally curated dataset that we collected, known here as Cross-Cultural Understanding Benchmark (CCUB) Dataset, to fight the bias prevalent in giant datasets. Our proposed approach is comprised of two fine-tuning techniques: (1) Adding visual context via fine-tuning a pre-trained text-to-image synthesis model, Stable Diffusion, on the CCUB text-image pairs, and (2) Adding semantic context via automated prompt engineering using the fine-tuned large language model, GPT-3, trained on our CCUB culturally-aware text data. CCUB dataset is curated and our approach is evaluated by people who have a personal relationship with that particular culture. Our experiments indicate that priming using both text and image is effective in improving the cultural relevance and decreasing the offensiveness of generated images while maintaining quality.","{'model': 'tldr@v2.0.0', 'text': 'A culturally-aware priming approach for text-to-image synthesis using a small but culturally curated dataset, known here as Cross-Cultural Understanding Benchmark (CCUB) Dataset, to fight the bias prevalent in giant datasets.'}",http://arxiv.org/pdf/2301.12073
-Large Language Models in Fault Localisation,Yonghao Wu,"Large Language Models (LLMs) have shown promise in multiple software engineering tasks including code generation, program repair, code summarisation, and test generation. Fault localisation is instrumental in enabling automated debugging and repair of programs and was prominently featured as a highlight during the launch event of ChatGPT-4. Nevertheless, the performance of LLMs compared to state-of-the-art methods, as well as the impact of prompt design and context length on their efficacy, remains unclear. To fill this gap, this paper presents an in-depth investigation into the capability of ChatGPT-3.5 and ChatGPT-4, the two state-of-the-art LLMs, on fault localisation. Using the widely-adopted large-scale Defects4J dataset, we compare the two LLMs with the existing fault localisation techniques. We also investigate the consistency of LLMs in fault localisation, as well as how prompt engineering and the length of code context affect the fault localisation effectiveness. Our findings demonstrate that within function-level context, ChatGPT-4 outperforms all the existing fault localisation methods. Additional error logs can further improve ChatGPT models' localisation accuracy and consistency, with an average 46.9% higher accuracy over the state-of-the-art baseline SmartFL on the Defects4J dataset in terms of TOP-1 metric. However, when the code context of the Defects4J dataset expands to the class-level, ChatGPT-4's performance suffers a significant drop, with 49.9% lower accuracy than SmartFL under TOP-1 metric. These observations indicate that although ChatGPT can effectively localise faults under specific conditions, limitations are evident. Further research is needed to fully harness the potential of LLMs like ChatGPT for practical fault localisation applications.","{'model': 'tldr@v2.0.0', 'text': 'An in-depth investigation into the capability of ChatG PT-3.5 and ChatGPT-4, the two state-of-the-art LLMs, on fault localisation is presented, using the widely-adopted large-scale Defects4J dataset.'}",https://arxiv.org/pdf/2308.15276
-OmniscientDB: A Large Language Model-Augmented DBMS That Knows What Other DBMSs Do Not Know,Matthias Urban,"In this paper, we present our vision of OmniscientDB, a novel database that leverages the implicitly-stored knowledge in large language models to augment datasets for analytical queries or even machine learning tasks. OmiscientDB empowers its users to augment their datasets by means of simple SQL queries and thus has the potential to dramatically reduce the manual overhead associated with data integration. It uses automatic prompt engineering to construct appropriate prompts for given SQL queries and passes them to a large language model like GPT-3 to contribute additional data (i.e., new rows, columns, or entire tables), augmenting the explicitly stored data. Our initial evaluation demonstrates the general feasibility of our vision, explores different prompting techniques in greater detail, and points towards several directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'OmiscientDB empowers its users to augment their datasets by means of simple SQL queries and thus has the potential to dramatically reduce the manual overhead associated with data integration.'}",http://publikationen.ub.uni-frankfurt.de/files/74426/06_08.pdf
-Visualizing Catalytic Dynamics Processes via Synchrotron Radiation Multitechniques,Dengfeng Cao,"The importance of catalysts today as workhorses in most modern industrial fields cannot be downplayed. As a result, rational design and engineering of targeted catalysts have emerged as key objectives and are dependent on in‐depth understanding of complex catalytic dynamics. Synchrotron radiation (SR) light sources with rich advanced experimental methods are being recognized as a comprehensive characterization platform, which can draw a full picture on such multiparameter‐involved catalysis under actual working conditions. Herein, the recent progress of catalytic dynamics process studied by the means of various SR techniques is summarized. In particular, SR‐based spectroscopic, scattering, and imaging investigations on true catalysts are first introduced with the potential of in situ and operando characterizations. Apparently, the limitations from single SR technique naturally prompt a simple combination of SR techniques to better understand the whole catalysis process. Moreover, the discrepancies among various online testing facilities and batches of samples, along with random/systematic errors introduced by traditional intermittent/asynchronous measurement make it imperative to develop more prolific systems, complementary of multiple SR techniques for deep probing of dynamic catalytic processes. It is believed that the booming new light sources can further enrich the current multiple SR techniques, and thus may realize the true visualization on future catalytic dynamic processes.",,
-Automatic Essay Scoring Method Based on Multi-Scale Features,Feng Li,"Essays are a pivotal component of conventional exams; accurately, efficiently, and effectively grading them is a significant challenge for educators. Automated essay scoring (AES) is a complex task that utilizes computer technology to assist teachers in scoring. Traditional AES techniques only focus on shallow linguistic features based on the grading criteria, ignoring the influence of deep semantic features. The AES model based on deep neural networks (DNN) can eliminate the need for feature engineering and achieve better accuracy. In addition, the DNN-AES model combining different scales of essays has recently achieved excellent results. However, it has the following problems: (1) It mainly extracts sentence-scale features manually and cannot be fine-tuned for specific tasks. (2) It does not consider the shallow linguistic features that the DNN-AES cannot extract. (3) It does not contain the relevance between the essay and the corresponding prompt. To solve these problems, we propose an AES method based on multi-scale features. Specifically, we utilize Sentence-BERT (SBERT) to vectorize sentences and connect them to the DNN-AES model. Furthermore, the typical shallow linguistic features and prompt-related features are integrated into the distributed features of the essay. The experimental results show that the Quadratic Weighted Kappa of our proposed method on the Kaggle ASAP competition dataset reaches 79.3%, verifying the efficacy of the extended method in the AES task.","{'model': 'tldr@v2.0.0', 'text': 'Sentence-BERT (SBERT) is utilized to vectorize sentences and connect them to the DNN-AES model, and the typical shallow linguistic features and prompt-related features are integrated into the distributed features of the essay.'}",https://www.mdpi.com/2076-3417/13/11/6775/pdf?version=1685706520
-"A Systematic Review of Disaster Management Systems: Approaches, Challenges, and Future Directions",Saad Mazhar Khan,"Disaster management is a critical area that requires efficient methods and techniques to address various challenges. This comprehensive assessment offers an in-depth overview of disaster management systems, methods, obstacles, and potential future paths. Specifically, it focuses on flood control, a significant and recurrent category of natural disasters. The analysis begins by exploring various types of natural catastrophes, including earthquakes, wildfires, and floods. It then delves into the different domains that collectively contribute to effective flood management. These domains encompass cutting-edge technologies such as big data analysis and cloud computing, providing scalable and reliable infrastructure for data storage, processing, and analysis. The study investigates the potential of the Internet of Things and sensor networks to gather real-time data from flood-prone areas, enhancing situational awareness and enabling prompt actions. Model-driven engineering is examined for its utility in developing and modeling flood scenarios, aiding in preparation and response planning. This study includes the Google Earth engine (GEE) and examines previous studies involving GEE. Moreover, we discuss remote sensing; remote sensing is undoubtedly a valuable tool for disaster management, and offers geographical data in various situations. We explore the application of Geographical Information System (GIS) and Spatial Data Management for visualizing and analyzing spatial data and facilitating informed decision-making and resource allocation during floods. In the final section, the focus shifts to the utilization of machine learning and data analytics in flood management. These methodologies offer predictive models and data-driven insights, enhancing early warning systems, risk assessment, and mitigation strategies. Through this in-depth analysis, the significance of incorporating these spheres into flood control procedures is highlighted, with the aim of improving disaster management techniques and enhancing resilience in flood-prone regions. The paper addresses existing challenges and provides future research directions, ultimately striving for a clearer and more coherent representation of disaster management techniques.",,https://www.mdpi.com/2073-445X/12/8/1514/pdf?version=1690796243
-"Guideline on Application Process of Techniques Developed for Repair, Replacement and Mitigation Activities",K. Koyama,"The process from the engineering development to actual application of the new techniques used for the maintenance activities for Japanese nuclear power plants’ (NPPs) component in operation, the JANTI Guideline, “Guideline on Application Process of Techniques Developed for Maintenance Activities” was developed to propose clarification activities to be checked or examined in each stage throughout process. The guideline focuses on techniques used for especially such remedial activities as repair, replacement and mitigation of ageing. In engineering development stage, all information on the new techniques are provided to be checked or examined at each stage of the process followed further, such as code/standard formulation stage. It is important to carry out share recognition for the activities, which should be performed in each process in the prompt application to the actual plant system efficiently, among the persons or organizations concerned.","{'model': 'tldr@v2.0.0', 'text': 'The JANTI Guideline, “Guideline on Application Process of Techniques Developed for Maintenance Activities” was developed to propose clarification activities to be checked or examined in each stage throughout process.'}",
-How understanding large language models can inform their use in physics education,Giulia Polverini,"The paper aims to fulfil three main functions: (1) to serve as an introduction for the physics education community to the functioning of Large Language Models (LLMs), (2) to present a series of illustrative examples demonstrating how prompt-engineering techniques can impact LLMs performance on conceptual physics tasks and (3) to discuss potential implications of the understanding of LLMs and prompt engineering for physics teaching and learning. We first summarise existing research on the performance of a popular LLM-based chatbot (ChatGPT) on physics tasks. We then give a basic account of how LLMs work, illustrate essential features of their functioning, and discuss their strengths and limitations. Equipped with this knowledge, we discuss some challenges with generating useful output with ChatGPT-4 in the context of introductory physics, paying special attention to conceptual questions and problems. We then provide a condensed overview of relevant literature on prompt engineering and demonstrate through illustrative examples how selected prompt-engineering techniques can be employed to improve ChatGPT-4's output on conceptual introductory physics problems. Qualitatively studying these examples provides additional insights into ChatGPT's functioning and its utility in physics problem solving. Finally, we consider how insights from the paper can inform the use of LMMs in the teaching and learning of physics.",,
-LLM4VV: Developing LLM-Driven Testsuite for Compiler Validation,Christian Munley,"Large language models (LLMs) are a new and powerful tool for a wide span of applications involving natural language and demonstrate impressive code generation abilities. In this paper, we explore the capabilitity of state-of-the-art LLMs, including closed-source options like OpenAI GPT-4 and open-source alternatives like Meta AI Codellama, to automatically generate tests and use these tests to validate and verify compiler implementations of a directive-based programming paradigm, OpenACC. Our approach entails exploring various prompt engineering techniques including a code template, retrieval-augmented generation (RAG) with code template, expressive prompt using RAG with code template, one-shot example, and RAG with one-shot example. This paper focusses on (a) exploring the capabilities of the latest LLMs for code generation, (b) investigating prompt and fine tuning methods, and (c) analyzing the outcome of LLMs generated tests","{'model': 'tldr@v2.0.0', 'text': 'The capabilitity of state-of-the-art LLMs, including closed-source options like OpenAI GPT-4 and open-source alternatives like Meta AI Codellama, to automatically generate tests is explored and these tests are used to validate and verify compiler implementations of a directive-based programming paradigm, OpenACC.'}",https://arxiv.org/pdf/2310.04963
-Automatic Bug Fixing via Deliberate Problem Solving with Large Language Models,Guoyang Weng,"Developers dedicate a significant share of their activities to finding and fixing defects in their code. Automated program repair (APR) attempts to reduce this effort by a set of techniques for automatically fixing errors or vulnerabilities in software systems. Recent Large Language Models (LLMs) such as GPT-4 offer an effective alternative to existing APR methods, featuring out-of-the-box bug fixing performance comparable to even sophisticated deep learning approaches such as CoCoNut. In this work we propose a further extension to LLM-based program repair techniques by leveraging a recently introduced interactive prompting technique called Tree of Thoughts (ToT). Specifically, we ask a LLM to propose multiple hypotheses about the location of a bug, and based on the aggregated response we prompt for bug fixing suggestions. A preliminary evaluation shows that our approach is able to fix multiple complex bugs previously unsolved by GPT-4 even with prompt engineering. This result motivates further exploration of hybrid approaches which combine LLMs with suitable meta-strategies.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a further extension to LLM-based program repair techniques by leveraging a recently introduced interactive prompting technique called Tree of Thoughts (ToT), and shows that this approach is able to fix multiple complex bugs previously unsolved by GPT-4 even with prompt engineering.'}",
-Data-Driven Approach for Formality-Sensitive Machine Translation: Language-Specific Handling and Synthetic Data Generation,Seugnjun Lee,"In this paper, we introduce a data-driven approach for Formality-Sensitive Machine Translation (FSMT) that caters to the unique linguistic properties of four target languages. Our methodology centers on two core strategies: 1) language-specific data handling, and 2) synthetic data generation using large-scale language models and empirical prompt engineering. This approach demonstrates a considerable improvement over the baseline, highlighting the effectiveness of data-centric techniques. Our prompt engineering strategy further improves performance by producing superior synthetic translation examples.","{'model': 'tldr@v2.0.0', 'text': 'A data-driven approach for Formality-Sensitive Machine Translation (FSMT) that caters to the unique linguistic properties of four target languages using large-scale language models and empirical prompt engineering.'}",http://arxiv.org/pdf/2306.14514
-Noisy Exemplars Make Large Language Models More Robust: A Domain-Agnostic Behavioral Analysis,Hongyi Zheng,"Recent advances in prompt engineering enable large language models (LLMs) to solve multi-hop logical reasoning problems with impressive accuracy. However, there is little existing work investigating the robustness of LLMs with few-shot prompting techniques. Therefore, we introduce a systematic approach to test the robustness of LLMs in multi-hop reasoning tasks via domain-agnostic perturbations. We include perturbations at multiple levels of abstractions (e.g. lexical perturbations such as typos, and semantic perturbations such as the inclusion of intermediate reasoning steps in the questions) to conduct behavioral analysis on the LLMs. Throughout our experiments, we find that models are more sensitive to certain perturbations such as replacing words with their synonyms. We also demonstrate that increasing the proportion of perturbed exemplars in the prompts improves the robustness of few-shot prompting methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a systematic approach to test the robustness of LLMs in multi-hop reasoning tasks via domain-agnostic perturbations, and finds that models are more sensitive to certain perturbation such as replacing words with their synonyms.'}",
-Automatic emotion recognition in healthcare data using supervised machine learning,Nazish Azam,"Human feelings are fundamental to perceive the conduct and state of mind of an individual. A healthy emotional state is one significant highlight to improve personal satisfaction. On the other hand, bad emotional health can prompt social or psychological well-being issues. Recognizing or detecting feelings in online health care data gives important and helpful information regarding the emotional state of patients. To recognize or detection of patient’s emotion against a specific disease using text from online sources is a challenging task. In this paper, we propose a method for the automatic detection of patient’s emotions in healthcare data using supervised machine learning approaches. For this purpose, we created a new dataset named EmoHD, comprising of 4,202 text samples against eight disease classes and six emotion classes, gathered from different online resources. We used six different supervised machine learning models based on different feature engineering techniques. We also performed a detailed comparison of the chosen six machine learning algorithms using different feature vectors on our dataset. We achieved the highest 87% accuracy using MultiLayer Perceptron as compared to other state of the art models. Moreover, we use the emotional guidance scale to show that there is a link between negative emotion and psychological health issues. Our proposed work will be helpful to automatically detect a patient’s emotion during disease and to avoid extreme acts like suicide, mental disorders, or psychological health issues. The implementation details are made publicly available at the given link: https://bit.ly/2NQeGET.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for the automatic detection of patient’s emotions in healthcare data using supervised machine learning approaches and uses the emotional guidance scale to show that there is a link between negative emotion and psychological health issues.'}",
-Techniques used to promote active learning in an undergraduate architectural acoustics course,Michelle C. Vigeant,"The five-year undergraduate architectural engineering (AE) program at Penn State requires all students to take an introductory course in architectural acoustics (AA) and for most students in the program, no further AA courses are required. As a result, it is very important to capture the students’ interest early on in the course and use strategies that will improve long-term retention of the material. A number of techniques were used to improve student engagement with the introductory AA course that had an enrollment of 94 students. The first homework assignment was to summarize a recent magazine article in any area of acoustics, which allowed the students to immediately see a real-world application of the subject. During lecture periods, students were encouraged to participate through the use of personal response devices, also known as “clickers.” Several different types of clicker questions were used, including review questions from the preceding lecture, questions to prompt participation during the pre...",,
-"Vehicle Fault Diagnostics Using Text Mining, Vehicle Engineering Structure and Machine Learning",Y. Murphey,"This paper presents an intelligent vehicle fault diagnostics system, SeaProSel(Search-Prompt-Select). SeaProSel takes a casual description of vehicle problems as input and searches for a diagnostic code that accurately matches the problem description. SeaProSel was developed using automatic text classification and machine learning techniques combined with a prompt-and-select technique based on the vehicle diagnostic engineering structure to provide robust classification of the diagnostic code that accurately matches the problem description. Machine learning algorithms are developed to automatically learn words and terms, and their variations commonly used in verbal descriptions of vehicle problems, and to build a TCW(Term-Code-Weight) matrix that is used for measuring similarity between a document vector and a diagnostic code class vector. When no exactly matched diagnostic code is found based on the direct search using the TCW matrix, the SeaProSel system will search the vehicle fault diagnostic structure for the proper questions to pose to the user in order to obtain more details about the problem. A LSI (Latent Semantic Indexing) model is also presented and analyzed in the paper. The performances of the LSI model and TCW models are presented and discussed. An in-depth study of different term weight functions and their performances are presented. All experiments are conducted on real-world vehicle diagnostic data, and the results show that the proposed SeaProSel system generates accurate results efficiently for vehicle fault diagnostics.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the proposed SeaProSel system generates accurate results efficiently for vehicle fault diagnostics.'}",https://article.sciencepublishinggroup.com/pdf/10.11648.j.ijiis.20150403.12.pdf
-A quantitative analysis approach for cardiac arrhythmia classification using higher order spectral techniques,L. Khadra,"Ventricular tachyarrhythmias, in particular ventricular fibrillation (VF), are the primary arrhythmic events in the majority of patients suffering from sudden cardiac death. Attention has focused upon these articular rhythms as it is recognized that prompt therapy can lead to a successful outcome. There has been considerable interest in analysis of the surface electrocardiogram (ECG) in VF centred on attempts to understand the pathophysiological processes occurring in sudden cardiac death, predicting the efficacy of therapy, and guiding the use of alternative or adjunct therapies to improve resuscitation success rates. Atrial fibrillation (AF) and ventricular tachycardia (VT) are other types of tachyarrhythmias that constitute a medical challenge. In this paper, a high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmias. The algorithm is based upon bispectral analysis techniques. The bispectrum is estimated using an autoregressive model, and the frequency support of the bispectrum is extracted as a quantitative measure to classify atrial and ventricular tachyarrhythmias. Results show a significant difference in the parameter values for different arrhythmias. Moreover, the bicoherency spectrum shows different bicoherency values for normal and tachycardia patients. In particular, the bicoherency indicates that phase coupling decreases as arrhythmia kicks in. The simplicity of the classification parameter and the obtained specificity and sensitivity of the classification scheme reveal the importance of higher order spectral analysis in the classification of life threatening arrhythmias. Further investigations and modification of the classification scheme could inherently improve the results of this technique and predict the instant of arrhythmia change.","{'model': 'tldr@v2.0.0', 'text': 'A high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmia, based upon bispectral analysis techniques and results show a significant difference in the parameter values for different arrhythmias.'}",
-Enhancing Performance Prediction Robustness by Combining Analytical Modeling and Machine Learning,Diego Didona,"Classical approaches to performance prediction rely on two, typically antithetic, techniques: Machine Learning (ML) and Analytical Modeling (AM). ML takes a black box approach, whose accuracy strongly depends on the representativeness of the dataset used during the initial training phase. Specifically, it can achieve very good accuracy in areas of the features' space that have been sufficiently explored during the training process. Conversely, AM techniques require no or minimal training, hence exhibiting the potential for supporting prompt instantiation of the performance model of the target system. However, in order to ensure their tractability, they typically rely on a set of simplifying assumptions. Consequently, AM's accuracy can be seriously challenged in scenarios (e.g., workload conditions) in which such assumptions are not matched. In this paper we explore several hybrid/gray box techniques that exploit AM and ML in synergy in order to get the best of the two worlds. We evaluate the proposed techniques in case studies targeting two complex and widely adopted middleware systems: a NoSQL distributed key-value store and a Total Order Broadcast (TOB) service.","{'model': 'tldr@v2.0.0', 'text': 'Several hybrid/gray box techniques are explored that exploit AM and ML in synergy in synergy to get the best of the two worlds, targeting two complex and widely adopted middleware systems.'}",http://www.gsd.inesc-id.pt/%7Eromanop/files/papers/ICPE15.pdf
-Optomechanical design of a prompt gamma reaction history diagnostic,M. Kaufman,"The National Ignition Facility and the Omega Laser Facility both have a need for measuring prompt gamma radiation as part of a nuclear diagnostic program. A new gamma-detection diagnostic using off-axis-parabolic mirrors has been built. Some new techniques were used in the design, construction, and tolerancing of this gamma ray diagnostic. Because of the wavelength requirement (250 to 700 nm), the optical element surface finishes were a key design consideration. The optical enclosure had to satisfy pressure safety concerns and shielding against electromagnetic interference induced by gammas and neutrons. Structural finite element analysis was needed to meet rigorous optical and safety requirements. The optomechanical design is presented. Alignment issues are also discussed.",,
-"Automated Library management system in the Chirala Engineering college, Chirala-using EZ school Software.",B. Kumar,"Library is a fast growing organism. The Ancient methods of maintaining it are no longer dynamic and efficient. For expeditious retrieval and dissemination of information and better service for the clientele, application of modern techniques has become absolutely indispensable. A properly computerized library will help its user with quick and prompt services. Library automation refers to mechanization of library housekeeping operations predominantly by computerization. Objectives of this study. To develop and updated database of books and other resources of the Engineering College Library, Chirala Engineering college, Chirala, JNTU, Kakinada University. To implement automated system using EZ School integrated Open Source Software. To provide various search options to know the availability of books in the Library. To generated the list of books due by a particular member and also the overdue charges. EZ School is integrated software system with the required models for small to very large libraries. It is found that this automation project sill service as a model for any Library. Being an open source, any library wanted to go for automation for their library housekeeping operations can make use of this software.","{'model': 'tldr@v2.0.0', 'text': 'This automation project sill service as a model for any Library, being an open source, any library wanted to go for automation for their library housekeeping operations can make use of this software.'}",
-Detecting COPD exacerbations early using daily telemonitoring of symptoms and k-means clustering: a pilot study,D. S. Morillo,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed electronic questionnaire and the applied methodology could help to early detect COPD exacerbations on a day-to-day basis and therefore could provide support to patients and physicians.'}",
-Embracing AI for better quality engineering,,"Infosys has already progressed towards this with an AI-first quality engineering framework leveraging traditional and generative AI techniques, large language models (LLMs) trained for specific QE tasks using Infosys test case repository, prompt engineering for testing use cases etc. With its unmatched ability to automate even unsolved testing tasks, analyse patterns and anomalies, detect performance issues early, and produce smart insights, AI will set the benchmarks in quality engineering.","{'model': 'tldr@v2.0.0', 'text': 'With its unmatched ability to automate even unsolved testing tasks, analyse patterns and anomalies, detect performance issues early, and produce smart insights, AI will set the benchmarks in quality engineering.'}",
-Photons for prompt tumour detection,R. Alfano,"Aeroplanes in clouds, submarines in murky water and cancerous tumours in breast tissue all have one thing in common – they are objects hidden in highly scattering media. Imaging through such media remains one of the most challenging problems in science and engineering but the advent of ultrafast lasers and detectors, coupled with a range of time-resolved techniques, has led to breakthroughs. Interest is currently focused on developing a time-resolved imaging technique to detect breast cancer tumours as early as possible.",,
-"Study of Real Options with Exogenous Competitive Entry to Analyze Dispute Resolution Ladder Investments in Architecture, Engineering, and Construction Projects",C. Menassa,"Architecture, engineering, and construction industry participants often find it pragmatic to implement a project-specific dispute resolution ladder (DRL) as a managerial tool to assist in the prompt resolution of claims and change orders (CCOs) that might arise during the project construction phase. This project-specific DRL consists of a single or multiple alternative dispute resolution (ADR) techniques that require capital expenditures to cover the expenses incurred by the owner’s/contractor’s employees and third-party neutrals. If a project-specific DRL is properly chosen, then the capital expenditures are outweighed by the expected benefits from the DRL implementation; namely, prompt resolution of the CCOs without incurring excessive cost overruns on an already financially stressed project budget, as well as avoiding the escalation of the claims to a dispute that requires long protracted litigation for final settlement. Typically, the decision as to which ADR techniques to include in the project-speci...",,
-Prompt List for Risk Management in Sri Lankan Software Industry,Mangala R. Perera,"Sound risk management is of crucial importance in software projects. As an emerging industry in Sri Lanka, software companies should focus on the proper risk management techniques. Unfortunately a large number of companies fail to achieve their project objectives. To avoid such situations, it is important for organizations to identify the main risks and take appropriate risk response actions. In this context, it is important to identify the prominent risks prevalent in Sri Lankan software industry and devise possible risk management strategies to respond to them. A prompt list is very useful in this situation. In this paper we propose such a prompt list tailored for the emerging software industry in Sri Lanka. Prompt lists, is a popular tool used in engineering to stimulate specific risk identification. It will ensure that the team will not miss out those risks that were important in the past. Preparing a prompt list for risk identification in Sri Lankan software industry and prioritizing the identified risks are the main objectives of this paper.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a prompt list for risk identification in Sri Lankan software industry and prioritizing the identified risks, which are the main objectives of this paper.'}",
-Facile Synthesis of Sulfobetaine-Stabilized Cu2O Nanoparticles and Their Biomedical Potential.,Marta J. Woźniak-Budych,"A novel approach using a zwitterionic sulfobetaine-based surfactant for the synthesis of spherical copper oxide nanoparticles (Cu2O NPs) has been applied. For the first time, N-hexadecyl-N,N-dimethyl-3-ammonio-1-propanesulfonate has been used as stabilizer to control the size and morphology of Cu2O NPs. Several techniques, such as transmission electron microscopy (TEM), X-ray diffraction (XRD), and fluorescence spectroscopy, are used to investigate the size, structure, and optical properties of synthesized Cu2O nanocrystals. The results indicate that copper(I) oxide nanoparticles with size in the range of 2 to 45 nm and crystalline structure, exhibit intense yellow fluorescence (λem = 575 nm). Furthermore, the cytotoxicity studies show that sulfobetaine-stabilized copper oxide nanoparticles prompt inhibition of cancer cell proliferation in a concentration-dependent manner, however, the adverse effect on the normal cells has also been observed. The results indicate that the sulfobetaine-stabilized Cu2O, because of their unique properties, have a potential to be applied in medical fields, such as cancer therapy and bioimaging.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the sulfobetaine-stabilized Cu2O, because of their unique properties, have a potential to be applied in medical fields, such as cancer therapy and bioimaging.'}",https://doi.org/10.1021/acsbiomaterials.7b00465.s001
-Telehealth Intensive Care Unit Nurse Surveillance of Sepsis,Teresa A. Rincon,"The purpose of this article is to describe the usability and human factors engineering standards used in development of a sepsis alert known as the sepsis prompt. Sensory processing, cognitive processing, signal detection, criterion response, and user satisfaction were evaluated with controlled user testing and critical incident response techniques. Nurses reported that the sepsis prompt was visible and distinct, making it easily detectable. The prompt provided a clear response mechanism and adequately balanced the number of false alerts with the likelihood of misses. Designers were able to use a mental model approach as they designed the prompt because the nurses were already using a manual sepsis detection process. This may have predisposed the nurses to response bias, and as such, they were willing to accommodate more false alarms than nurses who are not familiar with sepsis screening (surveillance). Nurses not currently screening for sepsis may not place the same value on this alert and find it an annoyance. The sepsis prompt met usability standards, and the nurses reported that it improved efficiency over the manual screening method.","{'model': 'tldr@v2.0.0', 'text': 'The usability and human factors engineering standards used in development of a sepsis alert known as the sepsi prompt met usability standards, and the nurses reported that it improved efficiency over the manual screening method.'}",
-Comparing the Effects of Design Interventions on the Quality of Design Concepts as a Reflection of Ideation Flexibility,Danielle Henderson,"Many tools, techniques, and other interventions have been developed to support idea generation within the design process. In previous research, we explored the separate effects of three such design interventions: teaming, problem framing, and design heuristics. In the teaming intervention, participants discussed a design prompt together but recorded their own ideas separately. In problem framing, multiple versions (framings) of each design prompt were used to elicit different solutions. In design heuristics, participants used specially designed cards to prompt new ways of thinking about the given design problem. In the current work, we compared the effects of these three interventions on students' design ideas with respect to one idea attribute in particular—quality. In total, 1088 design concepts were collected from 171 undergraduate students in engineering and industrial design from two universities. Individual cognitive style was also assessed using Kirton's Adaption–Innovation inventory (KAI). Six metrics taken from the design literature were used to assess the quality of each concept, namely: acceptability, applicability, clarity, effectiveness, implementability, and implicational explicitness. Paired t-tests and Pearson correlations were used to assess differences in quality between concepts generated with and without the three interventions; in addition, secondary effects were sought based on the cognitive styles and academic standings of the participants. Statistically significant differences were observed in design concept quality for the teaming and design heuristics interventions over the full sample and for some subgroups separated by cognitive style and academic standing. These results have implications for how educators teach design interventions and how students choose and apply interventions to affect the quality of their own design solutions.",,
-Highly Sensitive and Practical Detection of Plant Viruses via Electrical Impedance of Droplets on Textured Silicon-Based Devices,M. Ambrico,"Early diagnosis of plant virus infections before the disease symptoms appearance may represent a significant benefit in limiting disease spread by a prompt application of appropriate containment steps. We propose a label-free procedure applied on a device structure where the electrical signal transduction is evaluated via impedance spectroscopy techniques. The device consists of a droplet suspension embedding two representative purified plant viruses i.e., Tomato mosaic virus and Turnip yellow mosaic virus, put in contact with a highly hydrophobic plasma textured silicon surface. Results show a high sensitivity of the system towards the virus particles with an interestingly low detection limit, from tens to hundreds of attomolar corresponding to pg/mL of sap, which refers, in the infection time-scale, to a concentration of virus particles in still-symptomless plants. Such a threshold limit, together with an envisaged engineering of an easily manageable device, compared to more sophisticated apparatuses, may contribute in simplifying the in-field plant virus diagnostics.","{'model': 'tldr@v2.0.0', 'text': 'A label-free procedure applied on a device structure where the electrical signal transduction is evaluated via impedance spectroscopy techniques, which shows a high sensitivity of the system towards the virus particles with an interestingly low detection limit.'}",https://www.mdpi.com/1424-8220/16/11/1946/pdf?version=1479462590
-Hybrid Machine Learning/Analytical Models for Performance Prediction: A Tutorial,Diego Didona,"Classical approaches to performance prediction of computer systems rely on two, typically antithetic, techniques: Machine Learning (ML) and Analytical Modeling (AM). ML undertakes a black-box approach, which typically achieves very good accuracy in regions of the features' space that have been sufficiently explored during the training process, but that has very weak extrapolation power (i.e., poor accuracy in regions for which none, or too few samples are known). Conversely, AM relies on a white-box approach, whose key advantage is that it requires no or minimal training, hence supporting prompt instantiation of the target system's performance model. However, to ensure their tractability, AM-based performance predictors typically rely on simplifying assumptions. Consequently, AM's accuracy is challenged in scenarios not matching such assumptions. This tutorial describes techniques that exploit AM and ML in synergy in order to get the best of the two worlds. It surveys several such hybrid techniques and presents use cases spanning a wide range of application domains.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial describes techniques that exploit AM and ML in synergy in order to get the best of the two worlds and surveys several such hybrid techniques and presents use cases spanning a wide range of application domains.'}",
-Ontology Engineering and Knowledge Extraction for Cross-Lingual Retrieval,J. Trapman,"In this paper, we show that by integrating existing NLP techniques and Semantic Web tools in a novel way, we can provide a valuable contribution to the solution of the knowledge acquisition bottleneck problem. NLP techniques to create a domain ontology on the basis of an open domain corpus have been combined with Semantic Web tools. More specifically, Watson and Prompt have been employed to enhance the kick-o ontology while Cornetto, a lexical database for Dutch, has been adopted to establish a link between the concepts and their Dutch lexicalization. The lexicalized ontology constitutes the basis for the cross-language retrieval of learning objects within the LT4eL eLearning project.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that by integrating existing NLP techniques and Semantic Web tools in a novel way, this paper can provide a valuable contribution to the solution of the knowledge acquisition bottleneck problem.'}",
-Scientific Literature Text Mining and the Case for Open Access.,G. Sarma,"""Open access"" has become a central theme of journal reform in academic publishing. In this article, I examine the relationship between open access publishing and an important infrastructural element of a modern research enterprise, scientific literature text mining, or the use of data analytic techniques to conduct meta-analyses and investigations into the scientific corpus. I give a brief history of the open access movement, discuss novel journalistic practices, and an overview of data-driven investigation of the scientific corpus. I argue that particularly in an era where the veracity of many research studies has been called into question, scientific literature text mining should be one of the key motivations for open access publishing, not only in the basic sciences, but in the engineering and applied sciences as well. The enormous benefits of unrestricted access to the research literature should prompt scholars from all disciplines to lend their vocal support to enabling legal, wholesale access to the scientific literature as part of a data science pipeline.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that particularly in an era where the veracity of many research studies has been called into question, scientific literature text mining should be one of the key motivations for open access publishing, not only in the basic sciences, but in the engineering and applied sciences as well.'}",https://arxiv.org/pdf/1611.00097
-Supporting Users Tasks with Personal Information Management and Web Forms Augmentation,Sergio Firmenich,,"{'model': 'tldr@v2.0.0', 'text': ""This work proposes an approach for enhancing Web forms using client-side adaptation techniques in order to assist users to fill in Web forms and includes the support for data exchange between user's personal information management systems (PIMs) and third-party Web forms.""}",https://link.springer.com/content/pdf/10.1007/978-3-642-31753-8_20.pdf
-Deterioration of Concrete in Marine Structure,P. KadamHarshada.,"Concrete is a hard composite material got by blending the fitting extents of bond and totals (fine and coarse) with water, which in mix with steel fortification structures an amazing Civil Engineering material. Concrete, being great in pressure and steel in strain, end up being extremely solid and strong, if effectively outlined and cast. There are many components which prompt the decay of cement. The most widely recognized components which prompt the decay are the natural factors the materials and the throwing methodology, (for example, shameful cementing techniques, utilization of permeable totals, utilization of totals from saline sources, utilization of saline water in solid, poor vibration, disgraceful cover to support, spilling formwork, utilization of permeable and delicate totals), basic plan surrenders temperature and uncalled for utilize. A solid structure experiencing the weakening gives the accompanying demonstrative suggestions: breaking, spalling, scraped spot, stains, disintegration, diversion, and erosion. The most widely recognized variables prompting crumbling are the natural components prompting consumption of support, dishonorable utilize whereby over-burdening happens on a structure and the development techniques prompting permeable cement. This report introduces the investigation of the different components and their instruments prompting disintegration, their outcomes and the restoration of such structures. This report additionally talks about the scientific device for measurement of decay of cement in seaward structure.",,
-"Elastic rail clip design development, based on virtual prototyping",G. Todorov,"Virtual prototyping technology is a powerful tool when exact design solution is needed, corresponding to various and sometimes controversial requirements. It is used widely for design improvement/optimisation (topology, shape, parametric), based on engineering analysis through numerical techniques (as Finite Element Method) that enables prompt and accurate solutions. Design development approach based on this technology becomes standard for many industries, including railway transport. Presented study is based on an industrial project for design development of elastic rail clip. It is entirely based on virtual prototyping to obtain adequate design solution. Design is required to achieve certain clip rigidity, at certain level of safety (mechanical stress), for certain design space. This is a typical task for design optimisation techniques and performed simulations include 3 design concepts, 11 design variants and more than 2400 parametric configurations to obtain 4 allowable design solutions. The study is a good presentation of virtual prototyping application for industrial purposes.","{'model': 'tldr@v2.0.0', 'text': 'The study is a good presentation of virtual prototyping application for industrial purposes and performed simulations include 3 design concepts, 11 design variants and more than 2400 parametric configurations to obtain 4 allowable design solutions.'}",
-DEVELOPMENT OF LOW-COST HYBRID MEASUREMENT SYSTEM,A. Dindar,"Following the construction phase, tall buildings, bridges, dams are economically important engineering structures normally should not be in any critical condition during their service life unless a strong excitation (earthquake, blast, flood etc.) or boundary condition change (relative ground settlement) happens. For the property owner and for the community, it is essential to have a prompt condition assessment after the hazard in short period or to be aware whether a slow but progressing change in the building conditions for the long term. Structural Health Monitoring (SHM) systems aims to provide a basis for the requirement by installing the sensors on the system and processing data acquired from the sensors. With rapid development in the sensor technology, the monitoring methodologies have evolved from simple to more complex but complimentary measurement techniques. This study presents the measurement system of accelerometers, inclinometers and geodetic Global Navigation Satellite System (GNSS) receiver modules have been integrated in a single device for hybrid system. The data acquired from these different modules have been processed real time on site and the time series analysis for alerting mechanism has been set into the microchips for the sake of the prompt assessment of the critical condition change. The hybrid measurement device has been tested on a scaled building model excited by shake table for harmonic and random motions as well as on real building under ambient vibrations. The results are future promising in terms of the low-cost measurement and data process systems for the communities vulnerable to seismic hazard.",,
-Recognition of ventricular fibrillation using neural networks,Richard H. Clayton,,"{'model': 'tldr@v2.0.0', 'text': 'Although many VF detection techniques have been developed and claim good performance, independent evaluation has shown that some techniques are not optimal and one approach with potential lies in the area of neural computing.'}",
-Development of a Collaboration System based on Mobile Framework in the Field of Ships and Marine Industry,Myeong-ki Han,"Engineering information in the shipbuilding and marine industry should be inevitably modified owing to frequent changes in, for example, customer's requirements and product specifications. Such a working environment caused by a discrepancy in information between engineering and production departments can have inefficient effects on production tasks. In order to perform efficient production tasks, staff in a production department should have easy access to the latest information in engineering. Furthermore, information should be easy and prompt in sharing and searching when needed. In this paper, we propose a development of collaborative system based on mobile framework. This proposed system facilitates its application to the work environment in the shipbuilding and marine industry. Moreover, it particularly saves resources in shipyards such as energy and costs of labor and materials. Mobile techniques, such as XML, SOAP and Android, which have been used in this paper are useful for implementing functions of sharing and searching for information in a distributed environment. The purpose of the system proposed in this paper is to provide system users with a collaborative environment which is based on an online and paperless mobile framework, when different kinds of tasks are required due to changes in engineering information. This proposed system not only can check information in production tasks in real time but also can implement functions based on mobile framework. By reducing the discrepancy in information between the engineering and production departments, this collaborative system can be implemented effectively.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of the system proposed in this paper is to provide system users with a collaborative environment which is based on an online and paperless mobile framework, when different kinds of tasks are required due to changes in engineering information.'}",
-"Access, reward and penalty: The art of supervision at the (post)graduate level, a case study of an electrical engineering laboratory",Tanya M. Vernon,"Supervision of graduate students is becoming an increasingly important area of focus and scholarship. An example of the efflorescence of theory and pedagogy of supervision in Australia and other Commonwealth countries is the tertiary-level institutionalisation of programs which serve to accredit individuals in the complex art of supervision. But while scholarship of graduate-level supervision in the humanit ies and social science grows exponentially, basic discourse regarding method and practice of supervision of graduate level engineers is largely lacking. The author employs ethnomethodological techniques to explore and address this undertheorised area of engineering pedagogy from the point of view of a supervisor of number students in electrical and electronic engineers. The paper specifically explores a basic pedagogical structure in a laboratory of higher degree research (HDR) electrical engineering students, defines a complex time-based and proximity reward system the supervisor uses to prompt students to achieve excellence, and theorises an ethical way of dealing with poor student performance.",,
-Rapid microbiology - raising awareness.,J. Bailie,"A 'high-level overview' of some of the emerging rapid microbiology technologies designed to help healthcare engineering and infection control teams working in hospitals and other healthcare facilities more rapidly identify potentially hazardous levels of waterborne microorganisms in their water systems, enabling them to take prompt remedial action, and a look at the some of the 'pros and cons' of such testing techniques, was given by Nalco technical director, Howard Barnes, the vice-chair of the Legionella Control Association (LCA), at a recent LCA open day. HEJ editor, Jonathan Bailie, reports.","{'model': 'tldr@v2.0.0', 'text': ""A 'high-level overview' of some of the emerging rapid microbiology technologies designed to help healthcare engineering and infection control teams working in hospitals and other healthcare facilities more rapidly identify potentially hazardous levels of waterborne microorganisms in their water systems, and a look at the 'pros and cons' of such testing techniques.""}",
-Deterioration of Concrete in Marine Structure,Ms. Kadam Harshada,"Concrete is a hard composite material got by blending the fitting extents of bond and totals (fine and coarse) with water, which in mix with steel fortification structures an amazing Civil Engineering material. Concrete, being great in pressure and steel in strain, end up being extremely solid and strong, if effectively outlined and cast. There are many components which prompt the decay of cement. The most widely recognized components which prompt the decay are the natural factors the materials and the throwing methodology, (for example, shameful cementing techniques, utilization of permeable totals, utilization of totals from saline sources, utilization of saline water in solid, poor vibration, disgraceful cover to support, spilling formwork, utilization of permeable and delicate totals), basic plan surrenders temperature and uncalled for utilize. A solid structure experiencing the weakening gives the accompanying demonstrative suggestions: breaking, spalling, scraped spot, stains, disintegration, diversion, and erosion. The most widely recognized variables prompting crumbling are the natural components prompting consumption of support, dishonorable utilize whereby over-burdening happens on a structure and the development techniques prompting permeable cement. This report introduces the investigation of the different components and their instruments prompting disintegration, their outcomes and the restoration of such structures. This report additionally talks about the scientific device for measurement of decay of cement in seaward structure.",,
-On the contribution of biomedical engineering and technology to the understanding and the management of arterial hypertension.,S. Ghione,"There are several reasons why arterial blood pressure, i.e. the pressure within the large arterial vessels, is out of the physical parameters of the human body, one of the most frequently measured. Firstly, arterial blood pressure is a physiologically meaningful parameter, since it represents the driving pressure generated by the heart which maintains blood perfusion in the periphery. Secondly, it is a clinically important parameter: a decline of arterial blood pressure (e.g. in shock) may represent a life-threatening emergency which requires prompt recognition and correction; elevated blood pressure (hypertension) on the other hand is a very common condition, which bears a high risk of cardiovascular mortality and morbidity and can be controlled with appropriate pharmacological means. Thirdly, but not lastly, arterial blood pressure is easily measurable with a fair degree of accuracy by the standard manual sphygmomanometric method and, more recently, by non-invasive automatic techniques. This paper discusses some of the aspects related to arterial blood pressure measurement, in which, in the author's opinion, medical engineering and technology are expected to provide useful advancements. Two major areas will be considered. The first regards the methodologies for arterial blood pressure assessment; the second the identification and acquisition of information additional to blood pressure which would be helpful for a better understanding of blood pressure measurements and/or of risk profiling. For the purpose of this brief paper, we shall mainly use examples and reasonings from our own experience.","{'model': 'tldr@v2.0.0', 'text': ""This paper discusses some of the aspects related to arterial blood pressure measurement, in which, in the author's opinion, medical engineering and technology are expected to provide useful advancements.""}",
-Qualitative Approach For Estimating the Influence Of Refactoring And Scrum In Software Development,R. Hussain,"Software development is intellectually a complex chore. The swift progress of software currently requires the high rate software product release by development teams. Different software development techniques and quality assurance methods are used in order to achieve high worth, unfailing, and error free software. In order to deliver the product earlier, the development teams make an alteration to their conservative software development lifecycle to agile development method which can allow them towards prompt release of software management with the requirements-change experience. Refactoring has been rising in magnitude with modern software engineering advances, predominantly agile methodologies, which endorse uninterrupted progress of an application's code and blueprint. Refactoring is the practice of analyzing and facilitating the plan of offered code, without altering its performance. Another trendy techniques in Agile development is the Scrum methodology. It involves regular release and the client receives an absolutely prepared application that includes more and more features every time In this paper Qualitative Approach For Estimating the Influence Of Refactoring And Scrum In Software Development is utilized. In this model scrum methodology is utilized in enhanced form to overcome scrum issues along with refactoring project at both design and implementation level.","{'model': 'tldr@v2.0.0', 'text': 'A Qualitative Approach For Estimating the Influence of Refactoring And Scrum In Software Development is utilized and scrum methodology is utilized in enhanced form to overcome scrum issues along with refactoring project at both design and implementation level.'}",
-Global Project Management – Not Business As Usual,Damodara U. Kini,"Engineering and construction companies will have to make radical changes to their project management techniques if they are to be successful in the global marketplace in the 21st century. Customers will expect lower costs, shorter engineering and construction periods, low-cost, high-quality materials and equipment from manufacturers anywhere in the world, and prompt attention to their needs. These expectations mean that dramatically new approaches will be required not only in the organizational structure of such companies but also in their way of thinking and in such critical functions such as information systems, supplier management, and quality assurance. It will not be business as usual. There are six key areas in which engineering and construction companies will have to excel to be successful in satisfying the global customer. First, these companies will need an organization that can match their expertise in a cost-effective manner to the needs of their global customers. Second, global companies will ...",,
-Original Research,Shigang Gao,"The rapid development and adoption of generative artificial intelligence (AI) tools in the art and design education landscape have introduced both opportunities and challenges. This timely study addresses the need to effectively inte-grate these tools into the classroom while considering ethical implications and the importance of prompt engineering. By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context. Students in the digital at case study were taught prompt engineering techniques and were tasked with crafting multiple prompts, focusing on refining their ideas over time. Participants demonstrated an increased understanding of the potential and limitations of generative AI tools and how to manipulate subject matter for more effective results. The iterative process encouraged students to explore and experiment with their creative ideas, leading to a deeper understanding of the possibilities offered by AI tools. Despite acknowledging the ethical concerns regarding copyright and the potential replacement of artists, students appreciated the value of generative AI tools for enhancing their sketchbooks and ideation process. Through prompt engineering and iterative processes, students developed a more detail-oriented approach to their work. The challenge of using AI-generated images as final products was conceptually intriguing, requiring further investigation and consideration of the prompts. This study highlights the potential benefits and challenges of integrating generative AI tools into art and design classrooms, emphasizing the importance of prompt engineering, iterative processes, and ethical considerations as these technologies continue to evolve.","{'model': 'tldr@v2.0.0', 'text': 'By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context.'}",
-"Bioengineering for Customized Orthodontic Applications- Implant, Bracket and Dental Vibrator",R. Patil,"To understand complex living system an effort has made by mechanical engineers and dentists to deliver prompt products and services to patients concerned about their aesthetic look. Since two decades various bracket systems have designed involving techniques like milling, injection molding which are technically not flexible for the customized dental product development. The aim of this paper to design, develop a customized system which is economical and mainly emphasizes the expertise design and integration of engineering and dental fields. A custom made selfadjustable lingual bracket and customized implants are designed and developed using computer aided design (CAD) and rapid prototyping technology (RPT) to improve the smiles and to overcome the difficulties associated with conventional ones. Lengthy orthodontic treatment usually not accepted by the patients because the patient compliance is lost. Patient’s compliance can be improved by facilitating faster tooth movements by designing a localized dental vibrator using advanced engineering principles. Keywords—Orthodontics, Prosthodontics, Lingual bracket, Implants, Dental vibrator, Computer aided design, Rapid prototyping technology.",,
-Infrared thermography and ground penetrating radar for airport pavements assessment,A. Moropoulou,"Non-destructive techniques, such as infrared thermography and ground penetrating radar, have the potential to produce rapid and accurate assessment of airport pavements. In this work, an examination of asphalt pavements situated at the International Airport of Athens in Greece, is performed. Infrared thermography and ground penetrating radar are introduced with the purpose of providing prompt and accurate condition assessment of airport pavements. These techniques are used efficiently in the detection of cracks, voids and other imperfections appearing either from the ageing of the materials or due to poor workmanship. Finally, this paper describes the problem of deteriorating airport pavements, the procedure and the equipment used for the in situ tests, while the results obtained lead to the suggestion of a predictive monitoring non-destructive technique for the inspection and appropriateness of efficient engineering structures.",,
-Special Issue: Knowledge-based intelligent systems and their applications,I. Lovrek,"Intelligent techniques derived from knowledge-based engineering and related computing paradigms have provided useful concepts and tools to undertake a variety of real-world problems. These systems mimic the analytical and learning capabilities of the human brain. They harness the benefits of knowledge and intelligence to form an integrated framework for problem solving. In this special issue, a total of thirteen articles comprising extended papers from the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES2008) as well as from other submissions that highlight a small number of innovative knowledge-based intelligent systems and their applications to solving problems in different domains are presented. A summary of each article is as follows. With the development of advanced travelers information systems, it is important to have a prompt and accurate travel time prediction system for road networks. In the first article, two travel time prediction algorithms using naive Bayesian classification and rulebased classification are proposed. Based on a historical traffic database, the algorithms are able to yield high accuracy in travel time prediction. The algorithms are also useful for road networks with arbitrary travel routes. The results also reveal that naive Bayesian classification produces better mean absolute relative error than that of rule-based classification. For large-scale complex process plants that involve safety critical systems, real-time diagnosis is an important aspect. In the second article, an ontology for","{'model': 'tldr@v2.0.0', 'text': 'In this special issue, a total of thirteen articles comprising extended papers from the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES2008) as well as from other submissions that highlight a small number of innovative knowledge-based intelligent systems and their applications to solving problems in different domains are presented.'}",https://content.iospress.com:443/download/journal-of-intelligent-and-fuzzy-systems/ifs00430?id=journal-of-intelligent-and-fuzzy-systems%2Fifs00430
-A new quantitative analysis technique for cardiac arrhythmia using bispectrum and bicoherency,L. Khadra,"Ventricular tachyarrhythmias, in particular ventricular fibrillation (VF), are the primary arrhythmic events in the majority of patients suffering from sudden cardiac death. Attention has focused upon these articular rhythms as it is recognized that prompt therapy can lead to a successful outcome. There has been considerable interest in analysis of the surface electrocardiogram (ECG) in VF centred on attempts to understand the pathophysiological processes occurring in sudden cardiac death, predicting the efficacy of therapy, and guiding the use of alternative or adjunct therapies to improve resuscitation success rates. Atrial fibrillation (AF) and ventricular tachycardia (VT) are other types of tachyarrhythmias that constitute a medical challenge. In this paper, a high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmias. The algorithm is based upon bispectral analysis techniques. The bispectrum is estimated using an AR model, and the frequency support of the bispectrum is extracted as a quantitative measure to classify atrial and ventricular tachyarrhythmias. Results show a significant difference in the parameter values for different arrhythmias. Moreover, the bicoherency spectrum shows different bicoherency values for normal and tachycardia patients. In particular, the bicoherency indicates that phase coupling decreases as arrhythmia kicks in.","{'model': 'tldr@v2.0.0', 'text': 'A high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmia, based upon bispectral analysis techniques and results show a significant difference in the parameter values for different arrhythmias.'}",
-Fast Algorithm for X-ray Cone-beam Microtomography,Ang Shih,,"{'model': 'tldr@v2.0.0', 'text': 'A generalized Feldkamp cone-beam image reconstruction algorithm is reformulate, utilize curved voxels and mapping tables, improve the reconstruction efficiency by an order of magnitude relative to a direct implementation of the standard algorithm, and demonstrate the feasibility with numerical simulation and experiments using a prototype cone- beam X-ray microtomographic system.'}",
-Fetal transfusion therapy.,Daniel W. Skupski,"Rapid advances are occurring in the diagnosis and treatment of the fetus with a red blood cell or platelet cytopenia. Noninvasive methods of monitoring the alloimmunized pregnancy, invasive methods such as amniocentesis and cordocentesis, and intrauterine transfusion therapy of both red cells and platelets, are being further refined to allow the prompt recognition and treatment of fetal cytopenias. Specialized centers have now accrued a large experience in the management of the fetus severely affected by alloimmunization. Advances in ultrasound, blood banking techniques, and genetic engineering technology have spurred the most recent advances. The indications for diagnosis, timing and frequency of invasive procedures for treatment, and technical considerations regarding preparation of blood products and volume of transfusion, are outlined in this review. Polymerase chain reaction (PCR) determination of fetal Rh(D) genotype by chorionic villus sampling or amniocentesis in the first or second trimesters is a recent clinically useful advance. The advent of hematopoietic stem cell transplantation and the potential for gene therapy are exciting advances in the treatment and prevention of hematopoietic diseases, including, but not limited, to the fetal cytopenias.","{'model': 'tldr@v2.0.0', 'text': 'The indications for diagnosis, timing and frequency of invasive procedures for treatment, and technical considerations regarding preparation of blood products and volume of transfusion, are outlined in this review.'}",
-A Unified Architecture For Automatic Software Updates,Dominic White,"This paper attempts to address the issue of hardening the internal security of an organisation’s network by easing its patch management. Traditionally security has been modeled on a ""hard outer shell"" approach, with a firewall protecting the otherwise vulnerable internal network. With the advent of worms using such techniques as social engineering to bypass the organisational firewall and installing trojans, this approach is no longer sufficient. As a result of these new attacks, emphasis should be placed on improving the security of the internal network. Most research agrees that prompt patching of security vulnerabilities would significantly reduce the vulnerability of these machines. However, this requires system administrators not only to keep abreast of the flood of patches, but to ensure they are deployed to every machine, in what could be a very large network. These difficulties are worsened by problems the patches themselves often create. This is a difficult task and the failure of system administrators to perform it is echoed in the recent spate of worm attacks, with some taking advantage of vulnerabilities for which patches had been released up to six months earlier. To cope in this environment an organisation needs a comprehensive patch management policy. A technology agnostic view of this policy is described in order to better understand what is required of an automated solution. A few vendors have released tools to find vulnerabilities, find patches and distribute them, the best known example of which is Microsoft’s Software Update Service. However, these tools are generally inflexible, expensive and only deal with a limited part of the patch management process. This paper seeks to discuss the implementation of an open source, cross platform tool to solve this problem. This will involve discussions in four areas; the need for such a system, the patch management process, existing systems, and the proposed solution. The proposed system architecture is then broken down into four areas; sourcing patches, patch packaging, testing patches, patch distribution and the development of the system. The first area will discuss how to pull patches from multiple vendors such as Windows Update, Symantec AV, FreeBSD ports and ISC. The second will involve discussion on creating packages; whether to replace entire binaries, to employ binary patching techniques or to use source distribution. The third section will discuss how this tool can improve testing and provide stop-gap measures for protecting the organisation during the testing process through the use of IDS signatures. The fourth section will discuss methods of distributing these packages, such as direct download or peer-to-peer implementations like Bittorrent. The last section will discuss a modular, platform-independent implementation of the system.","{'model': 'tldr@v2.0.0', 'text': 'This paper attempts to address the issue of hardening the internal security of an organisation’s network by easing its patch management by discussing the implementation of an open source, cross platform tool to solve this problem.'}",
-A teaching approach for software testing,A. McAllister,"Teaching students how to test software is complicated by the absence of a simple, integrated approach for generating test plans. No single testing technique fulfils these needs, and teaching only a collection of disparate techniques makes it difficult to assign work for students. This talk presents an integrated approach for test plan generation that can be used by students in programming and software engineering courses. The approach provides simple guidelines that prompt discovery of sets of test cases that are typically more complete than students produce on an ad hoc basis. A technique is introduced that ensures all program statements are executed during testing and that loops are tested in a rigorous manner. Experience shows that this technique tends to be simpler to use than existing techniques that identify independent paths through programs. All of the guidelines presented can be applied without automated tools. The primary strength of the approach is in demonstrating to students how rigorous generation of test plans can identify test cases that might otherwise not occur to the tester, and how multiple techniques can be combined to complement one another.","{'model': 'tldr@v2.0.0', 'text': 'The primary strength of the approach is in demonstrating to students how rigorous generation of test plans can identify test cases that might otherwise not occur to the tester, and how multiple techniques can be combined to complement one another.'}",
-An expert advisor for the selection of thermodynamic property estimation methods,R. Butner,"The Prototype Process Simulation Consultant (PPSC) is a rule-oriented, knowledge-based software system designed to assist design engineers in the selection of physical property estimation methods for commercial chemical process simulators. The PPSC uses ''blackboarding'' techniques to share data with a number of external program subunits, that prompt the user for inputs, perform calculations and display results. The development of the PPSC provides a useful context for discussion of efforts to apply expert systems techniques to chemical engineering problems.","{'model': 'tldr@v2.0.0', 'text': 'The Prototype Process Simulation Consultant is a rule-oriented, knowledge-based software system designed to assist design engineers in the selection of physical property estimation methods for commercial chemical process simulators.'}",
-"Bone tissue engineering techniques, advances and scaffolds for treatment of bone defects.",Matthew Alonzo,,"{'model': 'tldr@v2.0.0', 'text': 'Recent advances in bone tissue engineering are highlighted, with particular emphasis on the role of the biomaterials as scaffolding material to heal bone defects.'}",
-ChatGPT-Based Debate Game Application Utilizing Prompt Engineering,Eunyul Lee,"This paper1 focuses on the implementation of a debate game using ChatGPT, aiming to investigate the feasibility of incorporating large language models into the educational domain through prompt engineering. The study explores strategies to elicit desired outputs from the GPT model by employing the prompt engineering methodology, as provided by Microsoft. Specifically, the game implementation involves the customization of ChatGPT's responses to facilitate a natural progression of debates, varying levels of difficulty, and an evaluation system for assessing the quality of discourse. By leveraging the prompt engineering methodology, we demonstrate that providing specific instructions or case-based prompts improves the accuracy and relevance of ChatGPT's answers. The developed application targets teenagers, enabling them to engage in real-time debates with ChatGPT and enhance their literacy skills. Furthermore, the game fosters the development of logical reasoning, persuasive abilities, effective expression, active participation, and attentive listening while expressing personal opinions, ultimately fostering a sense of accomplishment. Moreover, through debate evaluation and personalized advice, ChatGPT is expected to recognize and address its shortcomings, thereby continuously improving its conversational capabilities. Overall, this research contributes to the understanding of how large language models can be harnessed in educational settings and underscores the potential benefits of prompt engineering techniques in optimizing the outputs of such models.",,
-Exploring the Impact of Prompt Engineering on ChatGPT 3.5 Text Summarization: A BERT Score Evaluation,Kartik Ashok Hawelikar,"In the domain of Natural Language Processing (NLP), the technique of prompt engineering is a strategic method utilized to guide the responses of models such as ChatGPT. This research explores the intricacies of prompt engineering, with a specific focus on its effects on the quality of summaries generated by ChatGPT 3.5, an openly accessible chatbot developed by OpenAI. The study encompasses a comprehensive examination of 110 summaries produced from ten diverse paragraphs, employing eleven distinct summarization prompts under zero-shot setting. Evaluation is conducted using the BERT Score, a metric that offers a more contextually relevant assessment of summary quality. This study introduces an innovative approach to appraising the quality of summaries, setting it apart from prior investigations and delivering valuable insights into the nuances of prompt engineering's role within the NLP landscape. Ultimately, this inquiry illuminates the strengths and weaknesses associated with various prompts and their influence on ChatGPT 3.5's summarization capabilities, thereby making a significant contribution to the constantly evolving field of NLP and automated text summarization.",,https://doi.org/10.56726/irjmets45268
-Prompt Engineering Through the Lens of Optimal Control,Yifan Luo,"Prompt Engineering (PE) has emerged as a critical technique for guiding Large Language Models (LLMs) in solving intricate tasks. Its importance is highlighted by its potential to significantly enhance the efficiency and effectiveness of human-machine interaction. As tasks grow increasingly complex, recent advanced PE methods have extended beyond the limitations of single-round interactions to embrace multi-round interactions, which allows for a deeper and more nuanced engagement with LLMs. In this paper, we propose an optimal control framework tailored for multi-round interactions with LLMs. This framework provides a unified mathematical structure that not only systematizes the existing PE methods but also sets the stage for rigorous analytical improvements. Furthermore, we extend this framework to include PE via ensemble methods and multi-agent collaboration, thereby enlarging the scope of applicability. By adopting an optimal control perspective, we offer fresh insights into existing PE methods and highlight theoretical challenges that warrant future research. Besides, our work lays a foundation for the development of more effective and interpretable PE methods.","{'model': 'tldr@v2.0.0', 'text': 'This framework provides a unified mathematical structure that not only systematizes the existing PE methods but also sets the stage for rigorous analytical improvements and extends this framework to include PE via ensemble methods and multi-agent collaboration, thereby enlarging the scope of applicability.'}",
-Challenges of influencing cellular morphology by morphology engineering techniques and mechanical induced stress on filamentous pellet systems—A critical review,M. Böl,"Filamentous microorganisms are main producers of organic acids, enzymes, and pharmaceutical agents such as antibiotics and other active pharmaceutical ingredients. With their complex cell morphology, ranging from dispersed mycelia to dense pellets, the cultivation is challenging. In recent years, various techniques for tailor‐made cell morphologies of filamentous microorganisms have been developed to increase product formation and have been summarised under the term morphology engineering. These techniques, namely microparticle‐enhanced cultivation, macroparticle‐enhanced cultivation, and alteration of the osmolality of the culture medium by addition of inorganic salts, the salt‐enhanced cultivation, are presented and discussed in this review. These techniques have already proven to be useful and now await further proof‐of‐concept. Furthermore, the mechanical behaviour of individual pellets is of special interest for a general understanding of pellet mechanics and the productivity of biotechnological processes with filamentous microorganisms. Correlating them with substrate uptake and finally with productivity would be a breakthrough not to be underestimated for the comprehensive characterisation of filamentous systems. So far, this research field is under‐represented. First results on filamentous pellet mechanics are discussed and important future aspects, which the filamentous expert community should deal with, will be presented and critically discussed.","{'model': 'tldr@v2.0.0', 'text': 'The mechanical behaviour of individual pellets is of special interest for a general understanding of pellet mechanics and the productivity of biotechnological processes with filamentous microorganisms.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/elsc.202000060
-2nd International Conference on Sustainable Engineering Techniques (ICSET 2019),,"The 2nd International Conference on Sustainable Engineering Techniques (ICSET2019) Preface The ICSET2019 is the 2nd International Conference on Sustainable Engineering Techniques took place in Baghdad, on March 6-7th, 2019 at the Middle Technical University, Institute of Technology-Baghdad (Iraq Republic). ICSET present a platform to support the futurity research and deliver a forum for a discussion of the latest findings, innovation of new Techniques and their implementations into different fields of Mechanical and materials, Civil, Architectural, environmental; Chemical and Petroleum; Electrical; Electronic, and Information Technologies engineering techniques. ICSET 2019 fulfilled its primary objective to provide an international platform for academics, researchers, engineers, and industrial pioneers share their findings with experts, gather the scientific research efforts and the sustainability techniques to conserve the natural resources for future generations to enhance life and stimulate new ideas in the research area of sustainability. The conference program covered presentations, oral, and video conference presentations with more than 383 active contributors. Finally, a sincere thanks to the Ministry of Higher Education & Scientific Research (Iraq), which provide extraordinary support represented by the rich social program started with an outstanding set up an opening ceremony of the conference. We would also like to extend our thanks to all authors for their precious research papers and presentations. All reviewers are also thanked for their professional assistance in reviewing. Thanks are finally given to IOP Publication as well for producing this volume. List of Conference Photograph and Committees are available in this PDF.",,https://iopscience.iop.org/article/10.1088/1757-899X/518/1/011001/pdf
-Handbook of Research on Green Engineering Techniques for Modern Manufacturing,,"The Handbook of Research on Green Engineering Techniques for Modern Manufacturing provides emerging perspectives on the theoretical and practical aspects of green industrial concepts, such as green supply chain management and reverse logistics, for the sustainable utilization of resources and applications within manufacturing and engineering. Featuring coverage on a broad range of topics such as additive manufacturing, integrated manufacturing systems, and machine materials, this publication is ideally designed for engineers, environmental professionals, researchers, academicians, managers, policymakers, and graduate-level students seeking current research on recent and sustainable practices in manufacturing processes.",,
-Promptor: A Conversational and Autonomous Prompt Generation Agent for Intelligent Text Entry Techniques,Junxiao Shen,"Text entry is an essential task in our day-to-day digital interactions. Numerous intelligent features have been developed to streamline this process, making text entry more effective, efficient, and fluid. These improvements include sentence prediction and user personalization. However, as deep learning-based language models become the norm for these advanced features, the necessity for data collection and model fine-tuning increases. These challenges can be mitigated by harnessing the in-context learning capability of large language models such as GPT-3.5. This unique feature allows the language model to acquire new skills through prompts, eliminating the need for data collection and fine-tuning. Consequently, large language models can learn various text prediction techniques. We initially showed that, for a sentence prediction task, merely prompting GPT-3.5 surpassed a GPT-2 backed system and is comparable with a fine-tuned GPT-3.5 model, with the latter two methods requiring costly data collection, fine-tuning and post-processing. However, the task of prompting large language models to specialize in specific text prediction tasks can be challenging, particularly for designers without expertise in prompt engineering. To address this, we introduce Promptor, a conversational prompt generation agent designed to engage proactively with designers. Promptor can automatically generate complex prompts tailored to meet specific needs, thus offering a solution to this challenge. We conducted a user study involving 24 participants creating prompts for three intelligent text entry tasks, half of the participants used Promptor while the other half designed prompts themselves. The results show that Promptor-designed prompts result in a 35% increase in similarity and 22% in coherence over those by designers.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2310.08101
-Applying Model Driven Engineering Techniques to the Development of Contiki-Based IoT Systems,T. Asici,"The huge variety of smart devices and their communication models increases the development complexity of embedded software for the Internet of Things. As a consequence, development of these systems becomes more complex, error-prone, and costly. To tackle this problem, in this study, a model-driven approach is proposed for the development of Contiki-based IoT systems. To this end, the Contiki metamodel available in the literature is extended to include elements of WiFi connectivity modules (such as ESP8266), IoT Log Manager, and information processing components (such as Raspberry Pi). Based on this new metamodel, a domain-specific modeling environment is developed in which visual symbols are used and static semantics (representing system constraints) are defined. Also, the architectural code for the computing components of the IoT system such as Contiki, ESP8266, and RaspberryPi are generated from the developer's instance model. Finally, a Smart Fire Detection system is used to evaluate this study. By modeling the Contiki-based IoT system, we support model-driven development of the system, including WSN motes and sink nodes (with ContikiOS), WiFi modules and information processing components.","{'model': 'tldr@v2.0.0', 'text': 'By modeling the Contiki-based IoT system, this study supports model-driven development of the system, including WSN motes and sink nodes (with ContikiOS), WiFi modules and information processing components.'}",
-Knee Ligament Injury and the Clinical Application of Tissue Engineering Techniques: A Systematic Review.,Thomas C Riley,"BACKGROUND The incidence of knee ligament injury is increasing and represents a significant cost to healthcare providers. Current interventions include tissue grafts, suture repair and non-surgical management. These techniques have demonstrated good patient outcomes but have been associated graft rejection, infection, long term immobilization and reduced joint function. The limitations of traditional management strategies have prompted research into tissue engineering of knee ligaments. OBJECTIVE This paper aims to evaluate whether tissue engineering of knee ligaments offers a viable alternative in the clinical management of knee ligament injuries. A search of existing literature was performed using OVID Medline, Embase, AMED, PubMed and Google Scholar, and a manual review of citations identified within these papers. RESULTS Silk, polymer and extracellular matrix based scaffolds can all improve graft healing and collagen production. Fibroblasts and stem cells demonstrate compatibility with scaffolds, and have been shown to increase organized collagen production. These effects can be augmented using growth factors and extracellular matrix derivatives. Animal studies have shown tissue engineered ligaments can provide the biomechanical characteristics required for effective treatment of knee ligament injuries. CONCLUSION There is a growing clinical demand for a tissue engineered alternative to traditional management strategies. Currently, there is limited consensus regarding material selection for use in tissue engineered ligaments. Further research is required to optimize tissue engineered ligament production before clinical application. Controlled clinical trials comparing the use of tissue engineered ligaments and traditional management in patients with knee ligament injury could determine whether they can provide a cost-effective alternative.","{'model': 'tldr@v2.0.0', 'text': 'Control clinical trials comparing the use of tissue engineered ligaments and traditional management in patients with knee ligament injury could determine whether they can provide a cost-effective alternative.'}",
-Genome scale engineering techniques for metabolic engineering.,Rongming Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This review will attempt to summarize recent genome-scale design, build, test, and learn technologies and relate their use to a range of metabolic engineering applications.'}",http://manuscript.elsevier.com/S1096717615001238/pdf/S1096717615001238.pdf
-Scaffold Techniques and Designs in Tissue Engineering Functions and Purposes: A Review,A. Eltom,"In this review paper, the definition of the tissue engineering (TE) was comprehensively explored towards scaffold fabrication techniques and applications. Scaffold properties and features in TE, biological aspects, scaffold material composition, scaffold structural requirements, and old and current manufacturing technologies were reported and discussed. In almost all the reviewed reports, the TE definition denotes renewal, development, and repairs of damaged tissues caused by various factors such as disease, injury, or congenital disabilities. TE is multidisciplinary that combines biology, biochemistry, clinical medicine, and materials science whose application in cellular systems such as organ transplantation serves as a delivery vehicle for cells and drug. According to the previous literature and this review, the scaffold fabrication techniques can be classified into two main categories: conventional and modern techniques. These TE fabrication techniques are applied in the scaffold building which later on are used in tissue and organ structure. The benefits and drawbacks of each of the fabrication techniques have been described in conjunction with current areas of research devoted to deal with some of the challenges. To figure out, the highlighted aspects aimed to define the advancements and challenges that should be addressed in the scaffold design for tissue engineering. Additionally, this study provides an excellent review of original numerical approaches focused on mechanical characteristics that can be helpful in the scaffold design assessment in the analysis of scaffold parameters in tissue engineering.","{'model': 'tldr@v2.0.0', 'text': 'The definition of the tissue engineering (TE) was comprehensively explored and original numerical approaches focused on mechanical characteristics that can be helpful in the scaffold design assessment in the analysis of scaffold parameters in tissue engineering were provided.'}",http://downloads.hindawi.com/journals/amse/2019/3429527.pdf
-Marine collagen scaffolds for nasal cartilage repair: prevention of nasal septal perforations in a new orthotopic rat model using tissue engineering techniques.,C. Bermueller,"Autologous grafts are frequently needed for nasal septum reconstruction. Because they are only available in limited amounts, there is a need for new cartilage replacement strategies. Tissue engineering based on the use of autologous chondrocytes and resorbable matrices might be a suitable option. So far, an optimal material for nasal septum reconstruction has not been identified. The aim of our study was to provide the first evaluation of marine collagen for use in nasal cartilage repair. First, we studied the suitability of marine collagen as a cartilage replacement matrix in the context of in vitro three dimensional cultures by analyzing cell migration, cytotoxicity, and extracellular matrix formation using human and rat nasal septal chondrocytes. Second, we worked toward developing a suitable orthotopic animal model for nasal septum repair, while simultaneously evaluating the biocompatibility of marine collagen. Seeded and unseeded scaffolds were transplanted into nasal septum defects in an orthotopic rat model for 1, 4, and 12 weeks. Explanted scaffolds were histologically and immunohistochemically evaluated. Scaffolds did not induce any cytotoxic reactions in vitro. Chondrocytes were able to adhere to marine collagen and produce cartilaginous matrix proteins, such as collagen type II. Treating septal cartilage defects in vivo with seeded and unseeded scaffolds led to a significant reduction in the number of nasal septum perforations compared to no replacement. In summary, we demonstrated that marine collagen matrices provide excellent properties for cartilage tissue engineering. Marine collagen scaffolds are able to prevent septal perforations in an autologous, orthotopic rat model. This newly described experimental surgical procedure is a suitable way to evaluate new scaffold materials for their applicability in the context of nasal cartilage repair.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that marine collagen matrices provide excellent properties for cartilage tissue engineering and are able to prevent septal perforations in an autologous, orthotopic rat model.'}",https://europepmc.org/articles/pmc3762606?pdf=render
-Recent Advancements in Fruit Detection and Classification Using Deep Learning Techniques,C. Ukwuoma,"Recent advances in computer vision have allowed broad applications in every area of life, and agriculture is not left out. For the agri-food industry, the use of advanced technology is essential. Owing to deep learning’s capability to learn robust features from images, it has witnessed enormous application in several fields. Fruit detection and classification remains challenging due to the form, color, and texture of different fruit species. While studying the impact of computer vision on fruit detection and classification, we pointed out that till 2018 many conventional machine learning methods were utilized while a few methods exploited the application of deep learning methods for fruit detection and classification. This has prompted us to pursue an extensive study on surveying and implementing deep learning models for fruit detection and classification. In this article, we intensively discussed the datasets used by many scholars, the practical descriptors, the model’s implementation, and the challenges of using deep learning to detect and categorize fruits. Lastly, we summarized the results of different deep learning methods applied in previous studies for the purpose of fruit detection and classification. This review covers the study of recently published articles that utilized deep learning models for fruit identification and classification. Additionally, we also implemented from scratch a deep learning model for fruit classification using the popular dataset “Fruit 360” to make it easier for beginner researchers in the field of agriculture to understand the role of deep learning in the agriculture domain.","{'model': 'tldr@v2.0.0', 'text': 'This article intensively discussed the datasets used by many scholars, the practical descriptors, the model’s implementation, and the challenges of using deep learning to detect and categorize fruits, and summarized the results of different deep learning methods applied in previous studies for the purpose of fruit detection and classification.'}",https://downloads.hindawi.com/journals/mpe/2022/9210947.pdf
-Model-driven engineering techniques for the development of multi-agent systems,J. M. Gascueña,,"{'model': 'tldr@v2.0.0', 'text': 'In this paper, agent-oriented software development (AOSD) and MDE paradigms are fully integrated for the development of MAS and meta-modeling techniques are explicitly used to speed up several phases of the process.'}",https://ruidera.uclm.es/bitstreams/be2ad086-b609-4cf8-8b6d-153796bb487a/download
-Nondestructive testing and evaluation techniques of defects in fiber-reinforced polymer composites: A review,Jian Chen,"Fiber-reinforced polymer composites have excellent mechanical properties and outstanding development potential and are cost-effective. They have increasingly been used in numerous advanced and engineering applications as materials for wind turbine blades, helicopter rotors, high-pressure pipelines, and medical equipment. Understanding and assessing structural failure promptly in the whole lifecycle of a composite is essential to mitigating safety concerns and reducing maintenance costs. Various nondestructive testing and evaluation (NDT&E) technologies based on different evaluation principles have been established to inspect defects under different conditions. This paper reviews the established types of NDT&E techniques: acoustic emission, ultrasonic testing, eddy current testing, infrared thermography, terahertz testing, digital image correlation, shearography, and X-ray computed tomography, which is divided into three categories based on the operation frequency and data processing means of the output signal that is directly under analysis. We listed four types of defects/damage that are currently of great interest, namely, voids and porosity, fiber waviness and wrinkling, delamination and debonding, as well as impact damage. To identify a suitable method for different defects/damage, we performed characterization and evaluation by using these NDT&E techniques for typical defects/damage. Then, the cost, inspection speed, benefits and limitations, etc. were compared and discussed. Finally, a brief overview of the development of the technologies and their applications in the field of composite fabrication was discussed.",,https://www.frontiersin.org/articles/10.3389/fmats.2022.986645/pdf
-Sterilization techniques for biodegradable scaffolds in tissue engineering applications,Zheng Dai,"Biodegradable scaffolds have been extensively studied due to their wide applications in biomaterials and tissue engineering. However, infections associated with in vivo use of these scaffolds by different microbiological contaminants remain to be a significant challenge. This review focuses on different sterilization techniques including heat, chemical, irradiation, and other novel sterilization techniques for various biodegradable scaffolds. Comparisons of these techniques, including their sterilization mechanisms, post-sterilization effects, and sterilization efficiencies, are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Comparisons of different sterilization techniques, including their sterilization mechanisms, post-sterilization effects, and sterilization efficiencies, are discussed.'}",https://journals.sagepub.com/doi/pdf/10.1177/2041731416648810
-Analysing app reviews for software engineering: a systematic literature review,Jacek Dąbrowski,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of app review analysis research covering 182 papers published between 2012 and 2020 is presented, classifying it not only in terms of mined information and applied data mining techniques but also, and most importantly, in Terms of supported software engineering activities.'}",https://link.springer.com/content/pdf/10.1007/s10664-021-10065-7.pdf
-Applying formal software engineering techniques to smart grids,Georg Hackenberg,"Engineering complex systems that have to meet critical requirements is a difficult task, especially if multiple engineering disciplines are involved. Common practice in domains like the automotive or avionic industry shows that formal methods improve engineering process efficiency for embedded software due to abilities like abstraction, early verification and iterative refinement. This paper presents how existing formal software engineering methods can be adapted to meet the needs of the smart grid domain. A case study demonstrates how we develop a basic interdisciplinary but semantically integrated decomposition of a household including electric and software behavior. Finally, we provide first simulation results to evaluate the feasibility of the model as well as the presented engineering method.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents how existing formal software engineering methods can be adapted to meet the needs of the smart grid domain and develops a basic interdisciplinary but semantically integrated decomposition of a household including electric and software behavior.'}",
-Advances on Defect Engineering of Vanadium‐Based Compounds for High‐Energy Aqueous Zinc–Ion Batteries,Cong Guo,"Aqueous zinc–ion batteries (ZIBs) have been promptly developed as a competitive and promising system for future large‐scale energy storage. In recent years, vanadium (V)‐based compounds, with diversity of valences and high electrochemical‐activity, have been widely studied as cathodes for aqueous ZIBs because of their rich reserves and high theoretical capacity. However, the stubborn issues including low conductivity and sluggish kinetics, plague their smooth application in aqueous ZIBs. Among various countermeasures, defect engineering is believed as an effective method to alleviate the above limitations. This review highlights the challenges of different V‐based cathode materials (e.g., vanadium oxides and vanadates) and summarizes the advances in defect engineering strategies including types and effects of the defects, designed strategies, and characterization techniques for high‐energy ZIBs. Finally, several sound prospects in this fervent field are also rationally proposed for fundamental research and practical application.",,
-Additive manufacturing techniques for the production of tissue engineering constructs,C. Mota,"‘Additive manufacturing’ (AM) refers to a class of manufacturing processes based on the building of a solid object from three‐dimensional (3D) model data by joining materials, usually layer upon layer. Among the vast array of techniques developed for the production of tissue‐engineering (TE) scaffolds, AM techniques are gaining great interest for their suitability in achieving complex shapes and microstructures with a high degree of automation, good accuracy and reproducibility. In addition, the possibility of rapidly producing tissue‐engineered constructs meeting patient's specific requirements, in terms of tissue defect size and geometry as well as autologous biological features, makes them a powerful way of enhancing clinical routine procedures. This paper gives an extensive overview of different AM techniques classes (i.e. stereolithography, selective laser sintering, 3D printing, melt–extrusion‐based techniques, solution/slurry extrusion‐based techniques, and tissue and organ printing) employed for the development of tissue‐engineered constructs made of different materials (i.e. polymeric, ceramic and composite, alone or in combination with bioactive agents), by highlighting their principles and technological solutions. Copyright © 2012 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'An extensive overview of different AM techniques employed for the development of tissue‐engineered constructs made of different materials, by highlighting their principles and technological solutions is given.'}",
-Implementing Systems Engineering Techniques into Health Care,Joe Fitzler,"Revision with unchanged content. Health care delivery in the United States needs improvement. Each year, between 44,000 and 98,000 people die as a result of medical errors and the United States is outranked by 3/4 of developed countries in life expectancy and infant mortality. In a recent report, The National Academy of Engineering and the Institute of Medicine recommend that implementing systems engineering and systems engineering tools should be used in health care to improve the industry. A unique way to implement systems thinking in health care is to educate future caregivers in systems thinking so that they can recognize problems and make improvements from within. The implementation of systems skills into medical school curriculum was tested at the University of Missouri School of Medicine, which has a Problem Based Learning curriculum. The lack of a resource for medical students to research health care systems engineering prompted the creation of a handbook entitled Handbook of Health Care Systems Engineering for Medical Students. This book is addressed to medical educators, medical students, and any other health professional looking to implement systems engineering into health care.","{'model': 'tldr@v2.0.0', 'text': 'The lack of a resource for medical students to research health care systems engineering prompted the creation of a handbook entitled Handbook of Health Care Systems Engineering for Medical Students.'}",
-FORENSIC ENGINEERING SURVEYS WITH UAV PHOTOGRAMMETRY AND LASER SCANNING TECHNIQUES,C. Cappelletti,"Abstract. This work aims at presenting the use of new technologies in the field of forensic engineering. In particular, the use of UAV photogrammetry and laser scanning is compared with the traditional methods of surveying an accident site. In this framework, surveys must be carried out promptly, executed in a short time and performed so that the greatest possible amount of information is collected with sufficient accuracy to avoid the possibility of neglecting details once that the scene is no longer preserved. The combination of modern surveying techniques such UAV photogrammetry and laser scanning can properly fulfill these requirements. An experimental test has been arranged and instruments, procedures, settings, practical limits and results have been evaluated and compared with respect to the usual way of performing the survey for forensic purposes. In particular, both qualitative and quantitative considerations are given, assessing the completeness of the reconstructed model, the statistical evaluation of the errors and the accuracy achieved. ","{'model': 'tldr@v2.0.0', 'text': 'The use of UAV photogrammetry and laser scanning is compared with the traditional methods of surveying an accident site and qualitative and quantitative considerations are given, assessing the completeness of the reconstructed model, the statistical evaluation of the errors and the accuracy achieved.'}",https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLII-2-W9/227/2019/isprs-archives-XLII-2-W9-227-2019.pdf
-An empirical evaluation of requirement engineering techniques for collaborative systems,Miguel A. Teruel,"A collaborative system is a distributed software which allows several users to work together and carry out collaboration, communication and coordination tasks. To perform these tasks, the users have to be aware of other user's actions, usually by means of a set of awareness techniques. When we are defining a collaborative system, the awareness techniques can be considered as non-functional requirements bounded to some quality factors, such as usability. However, serious flaws can be found during the specification of these systems if we use the usual Requirement Engineering techniques available, because their expressiveness limitations when dealing with non-functional requirements. In this paper an empirical evaluation is introduced to determine if these techniques are really appropriate to model groupware requirements and which is the best approach to specify this kind of systems. With this aim, a collaborative text editor is used to evaluate whether the current techniques for Requirement Engineering are appropriated or not, exploiting the relation between awareness capabilities and standard quality factors.","{'model': 'tldr@v2.0.0', 'text': 'A collaborative text editor is used to evaluate whether the current techniques for Requirement Engineering are appropriated or not, exploiting the relation between awareness capabilities and standard quality factors.'}",
-What we should know before using tissue engineering techniques to repair injured tendons: a developmental biology perspective.,Chia-feng Liu,"Tendons connect muscles to bones, and serve as the transmitters of force that allow all the movements of the body. Tenocytes are the basic cellular units of tendons, and produce the collagens that form the hierarchical fiber system of the tendon. Tendon injuries are common, and difficult to repair, particularly in the case of the insertion of tendon into bone. Successful attempts at cell-based repair therapies will require an understanding of the normal development of tendon tissues, including their differentiated regions such as the fibrous mid-section and fibrocartilaginous insertion site. Many genes are known to be involved in the formation of tendon. However, their functional roles in tendon development have not been fully characterized. Tissue engineers have attempted to generate functional tendon tissue in vitro. However, a lack of knowledge of normal tendon development has hampered these efforts. Here we review studies focusing on the developmental mechanisms of tendon development, and discuss the potential applications of a molecular understanding of tendon development to the treatment of tendon injuries.","{'model': 'tldr@v2.0.0', 'text': 'The potential applications of a molecular understanding of tendon development to the treatment of tendon injuries are discussed and a lack of knowledge of normal tendon development has hampered these efforts.'}",https://europepmc.org/articles/pmc3098959?pdf=render
-Analysis of requirements engineering techniques for IT-enabled product service systems,M. Berkovich,"Product service systems (PSS) are introduced by many companies to increase their differentiation and to provide integrated solutions to customers. PSS are integrated solutions consisting of physical products, software and services aiming at providing an individualized solution to a customer's problem. The specific attributes of PSS lead to specific requirements for requirements engineering (RE). The goal of this paper is to analyze to which degree the analysis techniques of software engineering are suitable for PSS. We therefore conducted a structured literature review of software engineering techniques. The criteria for assessing the suitability of the techniques were based on the characteristics of PSS and the task of RE in the development process of PSS. We analyzed five textbooks and 144 scientific articles and identified 27 groups of techniques. The result is that there are major gaps in techniques for RE for PSS. Two of ten criteria are not satisfied by any technique. Moreover, for the majority of tasks of RE multiple techniques have to be combined for satisfying the criteria. In summary, the literature review shows that the techniques of software engineering are largely not directly applicable to PSS.","{'model': 'tldr@v2.0.0', 'text': 'The literature review shows that the techniques of software engineering are largely not directly applicable to PSS, with major gaps in techniques for RE for PSS.'}",
-Intelligent Computational Optimization in Engineering: Techniques and Applications,L. Nolle,,"{'model': 'tldr@v2.0.0', 'text': 'This Volume presents some of the latest developments with a focus on the design of algorithms for computational optimization and their applications in practice and provides various case studies of the application of intelligent optimization techniques in real-world applications.'}",
-Chain-Engineering-Based De Novo Drug Design against MPXVgp169 Virulent Protein of Monkeypox Virus: A Molecular Modification Approach,M. Naveed,"The unexpected appearance of the monkeypox virus and the extensive geographic dispersal of cases have prompted researchers to concentrate on potential therapeutic approaches. In addition to its vaccine build techniques, there should be some multiple integrated antiviral active compounds because of the MPV (monkeypox virus) outbreak in 2022. This study offers a computational engineering-based de novo drug discovery mediated by random antiviral active compounds that were screened against the virulent protein MPXVgp169, as one of the key players directing the pathogenesis of the virus. The screening of these candidates was supported by the use of 72 antiviral active compounds. The top candidate with the lowest binding affinity was selected for the engineering of chains or atoms. Literature assisted to identify toxic chains or atoms that were impeding the stability and effectiveness of antiviral compounds to modify them for enhanced efficacy. With a binding affinity of −9.4 Kcal/mol after chain, the lipophilicity of 0.41, the water solubility of 2.51 as soluble, and synthetic accessibility of 6.6, chain-engineered dolutegravir was one of the best active compounds, as proved by the computational engineering analysis. This study will revolutionize the era of drug engineering as a potential therapeutic strategy for monkeypox infection.","{'model': 'tldr@v2.0.0', 'text': 'This study offers a computational engineering-based de novo drug discovery mediated by random antiviral active compounds that were screened against the virulent protein MPXVgp169, as one of the key players directing the pathogenesis of the virus.'}",https://www.mdpi.com/2306-5354/10/1/11/pdf?version=1671620039
-A systematic review of the use of requirements engineering techniques in model-driven development,Grzegorz Loniewski,,"{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the current use of requirements engineering techniques in MDD processes and their actual automation level shows that although MDD techniques are used to a great extent in platform-independent models, platform-specific models, and at code level, at the requirements level most MDD approaches use only partially defined requirements models or even natural language.'}",
-A Chaos Engineering Approach for Improving the Resiliency of IT Services Configurations,Filippo Poltronieri,"Testing the resiliency of complex IT services deployed in hybrid Cloud scenarios is a challenging task that requires expensive and possibly destructive operations. An interesting approach lies in Chaos Engineering, a set of practices to test the resiliency of software systems running in a production environment. However, Chaos Engineering is an expensive practice that requires the setup of complicated operations that further increase the complexity of management operations. To reduce this complexity, Chaos Engineering can benefit from the adoption of non-destructive approaches such as the definition of realistic digital twins. A digital twin is a virtual replica of a real-system on which experimenting with management configurations. This paper embraces this research avenue by extending our previous efforts to integrate Chaos Engineering techniques into an IT services management framework called ChaosTwin. ChaosTwin leverages novel methodologies and tools capable of identifying and promptly react to unexpected failures. Finally, to implement autonomous fault management, ChaosTwin defines scaling and migration policies that can quickly explore for more resilient placements of software components in case of system failures. We believe that ChaosTwin can provide useful guidance to service providers in finding cost-effective service configurations capable of minimizing the negative effects of unpredictable events.","{'model': 'tldr@v2.0.0', 'text': 'ChaosTwin can provide useful guidance to service providers in finding cost-effective service configurations capable of minimizing the negative effects of unpredictable events and leverages novel methodologies and tools capable of identifying and promptly react to unexpected failures.'}",
-Advances in Tissue Engineering Techniques for Articular Cartilage Repair.,A. Haleem,"The limited repair potential of human articular cartilage contributes to development of debilitating osteoarthritis and remains a great clinical challenge. This has led to evolution of cartilage treatment strategies from palliative to either reconstructive or reparative methods in an attempt to delay or ""bridge the gap"" to joint replacement. Further development of tissue engineering-based cartilage repair methods have been pursued to provide a more functional biological tissue. Currently, tissue engineering of articular cartilage has three cornerstones; a cell population capable of proliferation and differentiation into mature chondrocytes, a scaffold that can host these cells, provide a suitable environment for cellular functioning and serve as a sustained-release delivery vehicle of chondrogenic growth factors and thirdly, signaling molecules and growth factors that stimulate the cellular response and the production of a hyaline extracellular matrix (ECM). The aim of this review is to summarize advances in each of these three fields of tissue engineering with specific relevance to surgical techniques and technical notes.","{'model': 'tldr@v2.0.0', 'text': 'The aim of this review is to summarize advances in each of these three fields of tissue engineering with specific relevance to surgical techniques and technical notes.'}",https://europepmc.org/articles/pmc5805482?pdf=render
-A benchmark for evaluating software engineering techniques for improving medical processes,Stefan Christov,"The software engineering and medical informatics communities have been developing a range of approaches for reasoning about medical processes. To facilitate the comparison of such approaches, it would be desirable to have a set of medical examples, or benchmarks, that are easily available, described in considerable detail, and characterized in terms of the real-world complexities they capture. This paper presents one such benchmark and discusses a list of desiderata that medical benchmarks can be evaluated against.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a set of medical examples, or benchmarks, that are easily available, described in considerable detail, and characterized in terms of the real-world complexities they capture and discusses a list of desiderata that medical benchmarks can be evaluated against.'}",http://laser.cs.umass.edu/techreports/10-023.pdf
-Multiscale and local engineering of speckle morphology through disordered media.,Chunxu Zhang,"In this Letter, we prompt a novel, to the best of our knowledge, method based on transmission matrix decomposition with wavelets to engineer the speckle morphology behind disordered media. By analyzing the speckles in multiscale spaces, we experimentally realized multiscale and localized control on the speckle size, position-dependent spatial frequency, and global morphology by operating on the decomposition coefficients using different masks. Speckles with contrasting features in different parts of the fields can be generated in one step. Our experimental results demonstrate a high degree of flexibility in manipulating light in a customizable manner. This technique has stimulating prospects in correlation control and imaging under scattering conditions.",,
-Retrofitting of RC beams using FRP techniques: a review,A. Sharba,"In structural and civil engineering, reinforced concrete structures are susceptible to inordinate loadings such as earthquake disasters, terrorist attacks and accidental impacts. This prompts an increasing desire to strengthen as well as enhance the fatigue performance and extend the ability of fatigue life of reinforced concrete structural components especially beams. Since 2003, Iraq has suffered major damage to the structural buildings, so it has become an important issue to study the existing concrete structures. Among the different techniques of rehabilitation of existing structures, fibre reinforced polymer (FRP) as an external bonding has been considered as a popular one. This paper reviews several features of Reinforced concrete (RC) beams strengthened with FRP. Also this paper aims to impart a comprehensive insight on adhesive curing, surface arrangement, and failure modes of RC beams modified with FRP. This effect of FRP for enhancing the techniques of rehabilitation is a three-fold task, to strengthen and retrofit of concrete structures, to extend the fatigue life of the structural element, and eliminates the crack growth rate.",,
-"Internet-based Social Engineering Attacks, Defenses and Psychology: A Survey",Theodore Longtchi,"Social engineering attacks are a major cyber threat because they often serve as a first step for an attacker to break into an otherwise well-defended network, steal victims' credentials, and cause financial losses. The problem has received due amount of attention with many publications proposing defenses against them. Despite this, the situation has not improved. In this paper, we aim to understand and explain this phenomenon by looking into the root cause of the problem. To this end, we examine the literature on attacks and defenses through a unique lens we propose -- {\em psychological factors (PFs) and techniques (PTs)}. We find that there is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions. This explains why existing defenses have achieved limited success. This prompts us to propose a roadmap for a more systematic approach towards designing effective defenses against social engineering attacks.","{'model': 'tldr@v2.0.0', 'text': 'There is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions, which explains why existing defenses have achieved limited success.'}",
-SoK: Why Have Defenses against Social Engineering Attacks Achieved Limited Success?,Theodore Longtchi,"Social engineering attacks are a major cyber threat because they often serve as a first step for an attacker to break into an otherwise well-defended network, steal victims’ credentials, and cause financial losses. The problem has received due amount of attention with many publications proposing defenses against them. Despite this, the situation has not improved. In this SoK paper, we aim to understand and explain this phenomenon by looking into the root cause of the problem. To this end, we examine the literature on attacks and defenses through a unique lens we propose — psychological factors (PFs) and techniques (PTs). We find that there is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions. This explains why existing defenses have achieved limited success. This prompts us to propose a roadmap for a more systematic approach towards designing effective defenses against social engineering attacks.","{'model': 'tldr@v2.0.0', 'text': 'There is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions, which explains why existing defenses have achieved limited success.'}",http://arxiv.org/pdf/2203.08302
-Reverse engineering techniques for cranioplasty: a case study,E. Maravelakis,This paper presents rapid prototyping and reverse engineering techniques applied to create an implant for the surgical reconstruction of a large cranial defect. A series of computed tomography (CT) images was obtained and purpose built software was used to extract the cranial geometry in a point cloud. The point cloud produced was used for: (a) the creation of a stereolithographic (STL) physical model for direct assessment of the cranial defect; and (b) the creation of a 3D mould model for the fabrication of the patient-specific implant.,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents rapid prototyping and reverse engineering techniques applied to create an implant for the surgical reconstruction of a large cranial defect.'}",
-"Balancing Agility and Formalism in Software Engineering, Second IFIP TC 2 Central and East European Conference on Software Engineering Techniques, CEE-SET 2007, Poznan, Poland, October 10-12, 2007, Revised Selected Papers",B. Meyer,,"{'model': 'tldr@v2.0.0', 'text': 'In this age of modern era, the use of internet must be maximized, as one of the benefits is to get the on-line balancing agility and formalism in software engineering.'}",
-Emerging Needs and Considerations for Digital Engineering Software Tools,Jason Boss,"ion of models would benefit from some industry-established descriptions for content for some of the most commonly employed use cases for exchanging information, such that the amount of information is sufficient to perform the agreed-to tasks but obscure the additional IP or part of the agreement that is not needed. An example of this may be an “interface control” model, which depicts the overall geometry and functional behaviors, etc., of a system component but does not provide any insights into how it actually works at the lower level, or a “build to print” model, which provides enough information to manufacture a part but does not provide any design intent or analysis techniques that went into arriving at the final configuration. Development of a standard set of use cases and abstraction techniques for the most common customer and supply chain interactions that could be used as a convention might be considered for future iteroparity standardization. There have been several techniques established for data transmission that could be further explored for feasibility and a number of them that are available today, but due to their complexity and cost to tailor these implementations, they are normally limited in use to larger program installations, technologies such as digital rights management, block chain, etc. Finally, tools that leverage large IP data sets to be “trained” to gain insights or supply tailored results should be disclosed and optioned to not retain or feedback information outside of the owning organization’s environment when back-feeding insights to software supplier-hosted environments could impose IP or security concerns.","{'model': 'tldr@v2.0.0', 'text': None}",
-Engineering of Heterostructure Pt/Co/AlOx for the enhancement of Dyzaloshinskii-Moria interaction,B. R. Sankhi,"The interfacial Dyzaloshinskii-Moria interaction (DMI) helps to stabilize chiral domain walls and magnetic skyrmions, which will facilitate new magnetic memories and spintronics logic devices. The study of interfacial DMI in perpendicularly magnetized structurally asymmetric heavy metal (HM) / ferromagnetic (FM) multilayer systems is of high importance due to the formation of chiral magnetic textures in the presence of DMI. Here, we report the impact of the cobalt oxidation at the cobalt -aluminum oxide interface in Pt/Co/AlOxtrilayer structure on the DMI by varying the post-growth annealing time and Aluminum thickness. For quantifying DMI, we employed magneto-optical imaging of asymmetric domain wall expansion, hysteresis loop shift, and spin-wave spectroscopy techniques. We further correlated the Cobalt oxidation with low-temperature Hall effect measurements and X-ray photoelectron spectroscopy. Our results emphasize the characterization of magnetic films for MRAM technologies semiconductor temperature process window, where magnetic interaction will be critical for device performance.",,
-A case study validation of a knowledge-based approach for the selection of requirements engineering techniques,Li Jiang,,"{'model': 'tldr@v2.0.0', 'text': 'A Knowledge-based Approach for the Selection of Requirements Engineering Techniques (KASRET) is proposed that helps during RE techniques selection and makes not only a contribution to RE but also to research and application of knowledge management and decision support in process development.'}",
-Leveraging Reddit for Suicidal Ideation Detection: A Review of Machine Learning and Natural Language Processing Techniques,Eldar Yeskuatov,"Suicide is a major public-health problem that exists in virtually every part of the world. Hundreds of thousands of people commit suicide every year. The early detection of suicidal ideation is critical for suicide prevention. However, there are challenges associated with conventional suicide-risk screening methods. At the same time, individuals contemplating suicide are increasingly turning to social media and online forums, such as Reddit, to express their feelings and share their struggles with suicidal thoughts. This prompted research that applies machine learning and natural language processing techniques to detect suicidality among social media and forum users. The objective of this paper is to investigate methods employed to detect suicidal ideations on the Reddit forum. To achieve this objective, we conducted a literature review of the recent articles detailing machine learning and natural language processing techniques applied to Reddit data to detect the presence of suicidal ideations. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we selected 26 recent studies, published between 2018 and 2022. The findings of the review outline the prevalent methods of data collection, data annotation, data preprocessing, feature engineering, model development, and evaluation. Furthermore, we present several Reddit-based datasets utilized to construct suicidal ideation detection models. Finally, we conclude by discussing the current limitations and future directions in the research of suicidal ideation detection.",,https://www.mdpi.com/1660-4601/19/16/10347/pdf?version=1661158040
-Adoption and Effects of Software Engineering Best Practices in Machine Learning,A. Serban,"Background. The increasing reliance on applications with machine learning (ML) components calls for mature engineering techniques that ensure these are built in a robust and future-proof manner. Aim. We aim to empirically determine the state of the art in how teams develop, deploy and maintain software with ML components. Method. We mined both academic and grey literature and identified 29 engineering best practices for ML applications. We conducted a survey among 313 practitioners to determine the degree of adoption for these practices and to validate their perceived effects. Using the survey responses, we quantified practice adoption, differentiated along demographic characteristics, such as geography or team size. We also tested correlations and investigated linear and non-linear relationships between practices and their perceived effect using various statistical models. Results. Our findings indicate, for example, that larger teams tend to adopt more practices, and that traditional software engineering practices tend to have lower adoption than ML specific practices. Also, the statistical models can accurately predict perceived effects such as agility, software quality and traceability, from the degree of adoption for specific sets of practices. Combining practice adoption rates with practice importance, as revealed by statistical models, we identify practices that are important but have low adoption, as well as practices that are widely adopted but are less important for the effects we studied. Conclusion. Overall, our survey and the analysis of responses received provide a quantitative basis for assessment and step-wise improvement of practice adoption by ML teams.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate, for example, that larger teams tend to adopt more practices, and that traditional software engineering practices tend to have lower adoption than ML specific practices.'}",https://scholarlypublications.universiteitleiden.nl/access/item%3A3307601/view
-Applying knowledge engineering techniques to customer analysis in the service industry,Sung-ho Ha,,"{'model': 'tldr@v2.0.0', 'text': 'This study differentiates between customers through customer segmentation, tracks customer shifts from segment to segment over time, discovers customer segment knowledge to build an individual transition path and a dominant transition path, and then predicts customer segment behavior patterns.'}",
-Oriented Porous Polymer Scaffolds in Tissue Engineering: A Comprehensive Review of Preparation Strategies and Applications,Tong Liu,"The pursuit of effective therapeutic strategies for tissue damage has prompted extensive scholarly investigations worldwide. Tissue engineering has emerged as a prominent approach, particularly through the utilization of artificial scaffolds that closely resemble the natural extracellular matrix (ECM). These scaffolds exhibit multi‐scale topological structures and surface physicochemical properties, which significantly influence cellular behavior, thereby attracting considerable attention from numerous researchers. This comprehensive review is concentrated on the primary techniques employed in the fabrication of biodegradable polymer scaffolds possessing oriented porous structures, and the most recent advancements in tissue engineering research are presented. Significantly, the profound influence of scaffold surface characteristics is underscored on cellular behavior, elucidating the superiority of oriented pore structures over disordered ones in mimicking the distinctive attributes of the ECM. Enhanced cell adhesion, proliferation, and tissue differentiation represent notable advantages associated with oriented porous scaffolds. Additionally, the critical interplay between scaffold structure, performance, and functionalization is emphasized, highlighting the imperative to optimize the clinical application of tissue engineering scaffolds.","{'model': 'tldr@v2.0.0', 'text': 'The profound influence of scaffold surface characteristics is underscored on cellular behavior, elucidating the superiority of oriented pore structures over disordered ones in mimicking the distinctive attributes of the ECM.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/mame.202300246
-Testing OpenFOAM Computational Fluid Dynamics simulation of heat transfer and fluid flow in a mechanical engineering bachelor degree,Marina Duarte,"Nowadays, countless areas in the domain of mechanical engineering practice can benefit from the use of Computational Fluid Dynamics techniques. Currently, at the Energy branch of the master’s degree in mechanical engineering at Instituto Superior de Engenharia do Porto, Computational Fluid Dynamics is only formally addressed in the first semester of the final year. In many cases, this prevents the students from promptly acquiring the necessary knowledge to embrace a master’s thesis or research project in this field. A possible way to minimise this problem was tested by exposing the students to the Computational Fluid Dynamics world earlier. Thus, in the curricular year 2022/2023, two optional simulation cases with a 35% weight in the final grade of a Heat Transfer course were proposed to the third-year students of the bachelor’s degree.For this purpose, an educational intervention following an action research methodology was implemented. A questionnaire was used to collect students’ feedback. Qualitative grades of the project reports and the discussion sessions corroborated students learning. Considering the students’ grades and that their feedback was in accordance with the complexity of the two simulation cases, we conclude that this education intervention was successful and that Computational Fluid Dynamics can be introduced earlier.",,
-"Artificial Intelligence in Engineering and Society: Blue Skies, Black Holes, and the Job of Requirements Engineers (Keynote)",Alessio Ferrari,"The democratization of artificial intelligence (AI) has brought substantial achievements in science, engineering disciplines, and society as a whole. New technologies based on large language models, multi-modal learning, embodied AI, and the quest for artificial general intelligence (AGI) promise to further change the world's landscape as we know it. At the same time, AI's rapid and uncontrolled evolution also poses serious risks to society, such as the concentration of power, exclusion, discrimination, and manipulation of reality. The keynote will present some experiences in AI democratization, including the usage of explainable machine learning approaches for agronomists, NLP-based solutions for railway engineers, image processing techniques for the maintenance of riverbeds, and mobile data processing in road safety assessment. The talk will outline the latest technological advancements in AI, e.g., in healthcare and science, and will show how large language models like ChatGPT and Bing Chat can solve long-standing requirements engineering (RE) problems. For example, requirements completeness can be easily checked and addressed with simple prompts, and model generation from requirements becomes a one-click task. The keynote will then describe the risks that current AI development poses to society. Besides the increasingly convincing deep fakes, and the widely discussed risks for privacy and reputation, we must be aware of the uncontrolled speed of AI evolution. As AI continues to advance, it will replace many jobs that require intellectual skills. This could lead to a significant number of people losing their jobs, as they may not have the necessary skills to adapt to the new labour market. People and entire countries that cannot exploit technological developments will be excluded from the game, and this will cause resentment and the possible emergence of new fundamentalism. The race for semiconductors is already creating hot spots and rifts between the superpowers. In this context, RE researchers are called to new technical and societal challenges. With pieces of code and even entire programs that can be automatically generated with large language models, the craft of prompting becomes the new requirements specification, and the concept of structured APIs dissolves into natural language interfaces. At the societal level, AI regulations are making their first steps, and we are called to contribute to operationalise the norms while preventing over-regulation. Equipped with years of experience at the boundary of the technical and social facets of systems, RE researchers are pivotal subjects in the new golden age of AI.","{'model': 'tldr@v2.0.0', 'text': 'The talk will outline the latest technological advancements in AI, e.g., in healthcare and science, and will show how large language models like ChatGPT and Bing Chat can solve long-standing requirements engineering (RE) problems, and describe the risks that current AI development poses to society.'}",
-Selecting Requirements Engineering Techniques Based on Project Attributes--A Case Study,Li Jiang,"Selection of the most appropriate RE techniques for a software project based on the project's characteristics is a non-trivial process and a common challenge faced by software developers. In order to facilitate RE techniques selection, we propose a model for RE technique suitability assessment (RETSA) based on project attributes. This model was developed based on surveys and interviews of experienced software developers and experts from industry and academia. RETSA was applied to an industrial project. This case study shows the help this model provided during the selection of RE techniques for a software project","{'model': 'tldr@v2.0.0', 'text': 'A model for RE technique suitability assessment (RETSA) based on project attributes was developed based on surveys and interviews of experienced software developers and experts from industry and academia and applied to an industrial project.'}",
-"Reverse engineering techniques applied to a human skull, for CAD 3D reconstruction and physical replication by rapid prototyping",L. Galantucci,"The production of a copy of an existing object of complex shape is one of the typical applications of the integration between two modern computer-based technologies, reverse engineering (RE) and rapid prototyping (RP). The method is extremely versatile and can be used in various applicative domains (e.g. replacement of anatomical parts with artificial prostheses, replication of skeletal remains). Two different acquisition techniques of images of a skull, by laser and by CT scan, were compared to ascertain which enabled more accurate reproduction of the original specimen. The skull was chosen due to it being the body part most often used in medico-legal investigations (for personal identification, skull-photo superimposition techniques, forensic art, etc). Comparison between the copy and the original yielded satisfactory results for both techniques. However, CT scanning demonstrated some advantages over the laser technique, as it provided a cleaner point cloud, enabling shorter pre-reproduction processing times, as well as data on the internal parts, which resulted in the reproduction of a more faithful copy.","{'model': 'tldr@v2.0.0', 'text': 'Comparison between the copy and the original yielded satisfactory results for both techniques, however, CT scanning demonstrated some advantages over the laser technique, as it provided a cleaner point cloud, enabling shorter pre-reproduction processing times, as well as data on the internal parts, which resulted in the reproduction of a more faithful copy.'}",
-Combining requirements engineering techniques - theory and case study,Li Jiang,"The selection of requirements engineering (RE) techniques during software project development is a challenge for most developers. One of the reasons is that there is a great lack of requirements engineering education in most academic programs, so software developers have to learn requirements engineering practices on the job. This can easily result in the selection of techniques that are ill-suited for a particular project, as the selection is based on personal preference rather than on the characteristics of the project. Very little research has been done in the area of technique selection based on project attributes. This paper describes research into the selection and combination of RE techniques as well as a case study that applied the selection process to an industrial software project.","{'model': 'tldr@v2.0.0', 'text': 'Research into the selection and combination of RE techniques as well as a case study that applied the selection process to an industrial software project are described.'}",https://digital.library.adelaide.edu.au/dspace/bitstream/2440/45364/1/hdl45364.pdf
-The Development of Power System Restoration Method for a Bulk Power System by Applying Knowledge Engineering Techniques,Y. Kojima,"A prototype expert system for bulk power system restoration is discussed. Restoration of a bulk power system involves operation of equipment that is complexly inter-related. For this reason, object-oriented programming techniques using a frame representation of knowledge and heuristics has been adopted to achieve efficient processing. The use of knowledge representation and processing is described, and the restoration process is examined. A prototype system that has successfully satisfied the requirements of speed, stability and flexibility is described. >","{'model': 'tldr@v2.0.0', 'text': 'A prototype expert system for bulk power system restoration using object-oriented programming techniques using a frame representation of knowledge and heuristics to achieve efficient processing is discussed.'}",
-Requirements for requirements engineering techniques,Linda A. Macaulay,"The first argument presented is that if we have a clear understanding of the objectives of the requirements engineering (RE) process then we can identify what techniques we need. No single method or technique will be sufficient. The second argument is that a key objective of the RE process is to specify a system which will ultimately be successful. Three common types of failure are examined: process failure, expectation failure and interaction failure. The third argument is that if the causes of each type of failure can be described then we will be able to identify what techniques are needed to help us avoid failure. The five common causes are described: the requirements engineering process itself, human communication within requirements, knowledge development, documentation of requirements and management. A discussion of each cause is followed by a list of the RE techniques needed. The paper concludes with a 'wish list' of seventy requirements for RE techniques, and a brief discussion of the strengths and weaknesses of the RE community in meeting those requirements.","{'model': 'tldr@v2.0.0', 'text': 'The first argument presented is that if the authors have a clear understanding of the objectives of the requirements engineering (RE) process then they can identify what techniques they need to help us avoid failure and no single method or technique will be sufficient.'}",
-Behavioral Characteristics of Magnesium as a Biomaterial for Surface Engineering Application,K. O. Babaremu,,,
-Investigating university students’ conceptions of engineering: an implied identity perspective,M. Khosronejad,"ABSTRACT Examining how learners ascribe meaning to different aspects of their university life is a necessary undertaking in higher education. This article aims to investigate university students’ conceptions of engineering and explore how these conceptions relate to the context of their experience. Our data collection method and data analysis are informed by the techniques of the phenomenographic approach and the Implied Identity Framework. Fifteen engineering students in a course about sustainability participated in the study. Results revealed eight categories of participants’ conceptions: (1) engineering as a practice for designing solutions; (2) engineering as a pragmatic practice for the welfare of people; (3) engineering as a knowledge-based practice; (4) engineering as a communicative practice; (5) engineering as a technology-mediated practice; (6) engineering as thinking; (7) engineering as an independent practice; and (8) engineering as learning. These conceptions had cognitive, technological, or social orientations and were related to three aspects of engineering: outcome-focused; process-focused; and person-focused. Also, findings demonstrated that different contexts of the university, workplace, and society prompt different ways of thinking about engineering. The findings place a greater emphasis on personal agency and the development of student identities as lifelong learners.",,
-Complementary techniques to analyse pericellular matrix formation by human MSC within hyaluronic acid hydrogels,C. Salzlechner,"Hydrogels are widely used as mimics of the native extracellular matrix as their physical and biological properties can be tuned over a wide range to match those of the native tissue. Cells encapsulated within hydrogels have recently been reported to modify their local surroundings by secreting and assembling proteins pericellularly, which in turn impacts their fate. As a result, methods to characterise and visualise the secreted matrix are becoming increasingly important in the development of regenerative therapies and in understanding cell behaviour within 3D matrices. Here, by combining fluorescent non-canonical amino acid tagging with Raman spectral imaging, we aimed to create 3D maps of human mesenchymal stromal cells (hMSC) and their secreted matrix when embedded within hydrogels. To demonstrate the value of our combined technique in a tissue engineering context, we cultured hMSC in Dopa-modified hyaluronic acid-based hydrogels and treated cultures with the 2-oxyglutarate analogue dimethyloxalyglycine (DMOG), which mimics the cellular effects of physiological hypoxia and can both promote the chondrogenic differentiation of progenitor cells and enhance cartilage-like matrix formation. Quantitative analyses of the distribution of newly synthesised proteins combined with principal components analyses of Raman spectra showed that DMOG prompted encapsulated cells to secrete more protein pericellularly than did untreated controls. Our findings demonstrate that it is possible to visualise both the 3D secreted matrix, as well as cellular contents using simple, unbiased, inexpensive techniques, providing complementary information on cells and their secreted matrix when encapsulated within 3D hydrogels.","{'model': 'tldr@v2.0.0', 'text': '3D maps of human mesenchymal stromal cells (hMSC) and their secreted matrix when embedded within hydrogels are created by combining fluorescent non-canonical amino acid tagging with Raman spectral imaging to demonstrate that it is possible to visualise both the 3Dsecreted matrix, as well as cellular contents using simple, unbiased, inexpensive techniques.'}",https://pubs.rsc.org/en/content/articlepdf/2020/ma/d0ma00472c
-A Comparative Study of Power Electronics and Control Techniques for Renewable Energy Integration in Smart Grids,Shashank Pareek,"The development of smart grid technology for more effective and sustainable energy management has been prompted by the growing use of renewable energy sources like solar and wind power. In order to integrate renewable energy sources into the grid, power electronics and control approaches are essential. The purpose of this study is to examine the pros and cons of the various power electronics and control approaches now in use for integrating renewable energy sources into smart grids. The development of smart grid technology for more effective and sustainable energy management has been facilitated by the use of renewable energy sources like solar and wind power. When it comes to incorporating renewable energy sources into the grid, power electronics and control approaches play a crucial role. This article provides a comprehensive analysis of the literature on the topic of renewable energy integration in smart grids, with a particular focus on power electronics and control strategies. The study is based on a review of relevant literature and an analysis of the benefits and drawbacks of each method. Efficiency, dependability, cost-effectiveness, scalability, and compatibility with various renewable energy sources are some of the characteristics included in the comparison study. The research shows that each method has advantages and disadvantages, and that the choice of method is context-specific. The work has significant significance for the field of electrical and computer engineering and aids in the development of more efficient and sustainable energy management systems for the future smart grid.",,
-Estimating and optimizing safety factors of retaining wall through neural network and bee colony techniques,B. Gordan,,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents the advantages of both prediction and optimization of retaining wall SF through artificial neural network (ANN) and artificial bee colony (ABC), respectively.'}",
-Three-dimensional scaffolds for tissue engineering applications: role of porosity and pore size.,Qiu Li Loh,"Tissue engineering applications commonly encompass the use of three-dimensional (3D) scaffolds to provide a suitable microenvironment for the incorporation of cells or growth factors to regenerate damaged tissues or organs. These scaffolds serve to mimic the actual in vivo microenvironment where cells interact and behave according to the mechanical cues obtained from the surrounding 3D environment. Hence, the material properties of the scaffolds are vital in determining cellular response and fate. These 3D scaffolds are generally highly porous with interconnected pore networks to facilitate nutrient and oxygen diffusion and waste removal. This review focuses on the various fabrication techniques (e.g., conventional and rapid prototyping methods) that have been employed to fabricate 3D scaffolds of different pore sizes and porosity. The different pore size and porosity measurement methods will also be discussed. Scaffolds with graded porosity have also been studied for their ability to better represent the actual in vivo situation where cells are exposed to layers of different tissues with varying properties. In addition, the ability of pore size and porosity of scaffolds to direct cellular responses and alter the mechanical properties of scaffolds will be reviewed, followed by a look at nature's own scaffold, the extracellular matrix. Overall, the limitations of current scaffold fabrication approaches for tissue engineering applications and some novel and promising alternatives will be highlighted.","{'model': 'tldr@v2.0.0', 'text': ""The ability of pore size and porosity of scaffolds to direct cellular responses and alter the mechanical properties of scaffold will be reviewed, followed by a look at nature's own scaffold, the extracellular matrix.""}",https://europepmc.org/articles/pmc3826579?pdf=render
-A Survey of App Store Analysis for Software Engineering,William J. Martin,"App Store Analysis studies information about applications obtained from app stores. App stores provide a wealth of information derived from users that would not exist had the applications been distributed via previous software deployment methods. App Store Analysis combines this non-technical information with technical information to learn trends and behaviours within these forms of software repositories. Findings from App Store Analysis have a direct and actionable impact on the software teams that develop software for app stores, and have led to techniques for requirements engineering, release planning, software design, security and testing. This survey describes and compares the areas of research that have been explored thus far, drawing out common aspects, trends and directions future research should take to address open problems and challenges.","{'model': 'tldr@v2.0.0', 'text': 'This survey describes and compares the areas of research that have been explored thus far, drawing out common aspects, trends and directions future research should take to address open problems and challenges.'}",
-Advances in Porous Scaffold Design for Bone and Cartilage Tissue Engineering and Regeneration.,A. Cheng,"IMPACT STATEMENT Challenges in musculoskeletal tissue regeneration affect millions of patients globally. Scaffolds for tissue engineering bone and cartilage provide promising solutions that increase healing and decrease need for complicated surgical procedures. Porous scaffolds have emerged as an attractive alternative to traditional scaffolds. However, the success of advanced materials, use of biological factors, and manufacturing techniques can vary depending on use case. This review provides perspective on porous scaffold manufacturing, characterization and application, and can be used to inform future scaffold design.","{'model': 'tldr@v2.0.0', 'text': None}",https://www.liebertpub.com/doi/pdf/10.1089/ten.teb.2018.0119
-Grounded Theory in Software Engineering Research: A Critical Review and Guidelines,Klaas-Jan Stol,"Grounded Theory (GT) has proved an extremely useful research approach in several fields including medical sociology, nursing, education and management theory. However, GT is a complex method based on an inductive paradigm that is fundamentally different from the traditional hypothetico-deductive research model. As there are at least three variants of GT, some ostensibly GT research suffers from method slurring, where researchers adopt an arbitrary subset of GT practices that are not recognizable as GT. In this paper, we describe the variants of GT and identify the core set of GT practices. We then analyze the use of grounded theory in software engineering. We carefully and systematically selected 98 articles that mention GT, of which 52 explicitly claim to use GT, with the other 46 using GT techniques only. Only 16 articles provide detailed accounts of their research procedures. We offer guidelines to improve the quality of both conducting and reporting GT studies. The latter is an important extension since current GT guidelines in software engineering do not cover the reporting process, despite good reporting being necessary for evaluating a study and informing subsequent research.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the variants of GT and identifies the core set of GT practices, then analyzes the use of grounded theory in software engineering and offers guidelines to improve the quality of both conducting and reporting GT studies.'}",https://researchrepository.ul.ie/articles/conference_contribution/Grounded_theory_in_software_engineering_research_a_critical_review_and_guidelines/19852003/1/files/35268280.pdf
-Cell-seeding techniques in vascular tissue engineering.,Gustavo A. Villalona,"Previous studies have demonstrated the benefits of cell seeding in the construction of tissue-engineered vascular grafts (TEVG). However, seeding methods are diverse and no method is clearly superior in either promoting seeding efficiency or improving long-term graft function. As we head into an era during which a variety of different TEVG are under investigation in clinical trials around the world, it is important to consider the regulatory issues surrounding the translation of these technologies. In this review, we summarize important advances in the field of vascular tissue engineering, with particular attention on cell-seeding techniques for TEVG development and special emphasis placed on regulatory issues concerning the clinical translation of these various methods.","{'model': 'tldr@v2.0.0', 'text': 'Important advances in the field of vascular tissue engineering are summarized, with particular attention on cell-seeding techniques for TEVG development and special emphasis placed on regulatory issues concerning the clinical translation of these various methods.'}",https://europepmc.org/articles/pmc2946885?pdf=render
-Ontology Engineering,Elisa F. Kendall,"54 Computer Integrated Approach to Web Ontology Learning and Engineering T he development of the semantic Web 1 — which seeks to improve the semantic awareness of computers connected via the Internet—requires a systematic, computer-oriented representation of the world. Researchers often refer to such a world model as an ontology. Despite the significant amount of work done on them in recent years, ontologies have yet to be widely applied and used. Research to date has mainly addressed the basic principles, such as knowledge representation formalisms, devoting only limited attention to more practical issues such as techniques and tools aimed at an ontology's actual construction and content. We have developed a software environment, centered around the OntoLearn tool, that can build and assess a domain ontology for intelligent information integration within a virtual user community. Further, we have tested OntoLearn in two European projects, where it functioned as the basis for a semantic interoperability platform used by small-and medium-sized tourism enterprises. Our approach to ontology engineering uses an iterative process that involves automatic concept learning with OntoLearn, machine-supported concept validation with Consys, 2 and management with SymOntoX. 3 The engineering process starts with OntoLearn exploring available documents and related Web sites to learn domain concepts and detect taxonomic relations among them, producing as output a domain concept forest. Initially, we base concept learning on external, generic knowledge sources. In subsequent cycles, the domain ontology receives progressively more use as it becomes adequately populated. The self-learning cycle in Figure 1 shows this process. Next, we undertake ontology validation with Consys, a Web-based groupware package that performs consensus building by thoroughly validating representatives of the communities active in the application domain. Throughout the cycle, Onto-Learn operates in connection with SymOntoX. Ontology engineers can use this management system to define concepts and their mutual connections, thus allowing construction of a semantic net. Further, SymOntoX's environment can automatically attach learned domain concept trees under the appropriate nodes of the upper-domain ontology, thereby enriching concepts with additional information. Sym-OntoX also performs consistency checks. Figure 2 shows OntoLearn's system architecture, which supports a three-phase process. First, the system extracts a domain terminology from texts available in the application domain—usually drawn from specialized Web sites or documents exchanged among members of a virtual community. The system then filters this information through a natural The authors have built a software environment that supports the construction and assessment of a …","{'model': 'tldr@v2.0.0', 'text': 'A software environment is built that can build and assess a domain ontology for intelligent information integration within a virtual user community that functioned as the basis for a semantic interoperability platform used by small-and medium-sized tourism enterprises.'}",https://link.springer.com/content/pdf/bfm:978-3-031-79486-5/1?pdf=chapter%20toc
-Software Engineering Economics,B. Boehm,"This paper summarizes the current state of the art and recent trends in software engineering economics. It provides an overview of economic analysis techniques and their applicability to software engineering and management. It surveys the field of software cost estimation, including the major estimation techniques available, the state of the art in algorithmic cost models, and the outstanding research issues in software cost estimation.","{'model': 'tldr@v2.0.0', 'text': 'The field of software cost estimation is surveyed, including the major estimation techniques available, the state of the art in algorithmic cost models, and the outstanding research issues inSoftware cost estimation.'}",http://csse.usc.edu/csse/TECHRPTS/1984/usccse84-500/usccse84-500.pdf
-A review of key challenges of electrospun scaffolds for tissue‐engineering applications,S. Khorshidi,"Tissue engineering holds great promise to develop functional constructs resembling the structural organization of native tissues to improve or replace biological functions, with the ultimate goal of avoiding organ transplantation. In tissue engineering, cells are often seeded into artificial structures capable of supporting three‐dimensional (3D) tissue formation. An optimal scaffold for tissue‐engineering applications should mimic the mechanical and functional properties of the extracellular matrix (ECM) of those tissues to be regenerated. Amongst the various scaffolding techniques, electrospinning is an outstanding one which is capable of producing non‐woven fibrous structures with dimensional constituents similar to those of ECM fibres. In recent years, electrospinning has gained widespread interest as a potential tissue‐engineering scaffolding technique and has been discussed in detail in many studies. So why this review? Apart from their clear advantages and extensive use, electrospun scaffolds encounter some practical limitations, such as scarce cell infiltration and inadequate mechanical strength for load‐bearing applications. A number of solutions have been offered by different research groups to overcome the above‐mentioned limitations. In this review, we provide an overview of the limitations of electrospinning as a tissue‐engineered scaffolding technique, with emphasis on possible resolutions of those issues. Copyright © 2015 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the limitations of electrospinning as a tissue‐engineered scaffolding technique is provided, with emphasis on possible resolutions of those issues.'}",
-Pathological Process of Prompt Connection between Host and Donor Tissue Vasculature Causing Rapid Perfusion of the Engineered Donor Tissue after Transplantation,Sachiko Sekiya,"The shortage of donors for transplantation therapy is a serious issue worldwide. Tissue engineering is considered a potential solution to this problem. Connection and perfusion in engineered tissues after transplantation is vital for the survival of the transplanted tissue, especially for tissues requiring blood perfusion to receive nutrients, such as the heart. A myocardial cell sheet containing an endothelial cell network structure was fabricated in vitro using cell sheet technology. Transplantation of the three-dimensional (3D) tissue by layering myocardial sheets could ameliorate ischemic heart disease in a rat model. The endothelial cell network in the 3D tissue was able to rapidly connect to host vasculature and begin perfusion within 24 h after transplantation. In this review, we compare and discuss the engineered tissue–host vasculature connection process between tissue engineered constructs with hydrogels and cell sheets by histological analysis. This review provides information that may be useful for further improvements of in vivo engineered tissue vascularization techniques.","{'model': 'tldr@v2.0.0', 'text': 'This review compares and discusses the engineered tissue–host vasculature connection process between tissue engineered constructs with hydrogels and cell sheets by histological analysis and provides information that may be useful for further improvements of in vivo engineered tissue vascularization techniques.'}",https://www.mdpi.com/1422-0067/19/12/4102/pdf?version=1545124295
-From Word Embeddings to Document Similarities for Improved Information Retrieval in Software Engineering,Xin Ye,"The application of information retrieval techniques to search tasks in software engineering is made difficult by the lexical gap between search queries, usually expressed in natural language (e.g. English), and retrieved documents, usually expressed in code (e.g. programming languages). This is often the case in bug and feature location, community question answering, or more generally the communication between technical personnel and non-technical stake holders in a software project. In this paper, we propose bridging the lexical gap by projecting natural language statements and code snippets as meaning vectors in a shared representation space. In the proposed architecture, word embeddings are rst trained on API documents, tutorials, and reference documents, and then aggregated in order to estimate semantic similarities between documents. Empirical evaluations show that the learned vector space embeddings lead to improvements in a previously explored bug localization task and a newly de ned task of linking API documents to computer programming questions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes bridging the lexical gap by projecting natural language statements and code snippets as meaning vectors in a shared representation space and shows that the learned vector space embeddings lead to improvements in a previously explored bug localization task and a newly introduced task of linking API documents to computer programming questions.'}",
-"Domain Analysis and Description Principles, Techniques, and Modelling Languages",D. Bjørner,"We present a method for analysing and describing domains. By a domain we shall understand a rationally describable segment of a human assisted reality, i.e., of the world, its physical parts: natural [“God-given”] and artifactual [“human-made”], and living species: plants and animals including, notably, humans. These are endurants (“still”), as well as perdurants (“alive”). Emphasis is placed on “human-assistedness,” that is, that there is at least one (human-made) artifact and, therefore, that humans are a primary cause for change of endurant states as well as perdurant behaviours. By a method we shall mean a set of principles of analysis and for selecting and applying a number of techniques and tools in the construction of some artifact, say a domain description. We shall present a method for constructing domain descriptions. Among the tools we shall only be concerned with are the analysis and synthesis languages. Domain science and engineering marks a new area of computing science. Just as we are formalising the syntax and semantics of programming languages, so we are formalising the syntax and semantics of human-assisted domains. Just as physicists are studying the natural physical world, endowing it with mathematical models, so we, computing scientists, are studying these domains, endowing them with mathematical models, A difference between the endeavours of physicists and ours lies in the tools: The physics models are based on classical mathematics, differential equations and integrals, and so on; our models are based on mathematical logic, set theory, and algebra [1]. Where physicists thus classically use a variety of differential and integral calculi to model the physical world, we shall be using the analysis and description calculi presented in this article to model primarily artifactual domains.","{'model': 'tldr@v2.0.0', 'text': 'Where physicists thus classically use a variety of differential and integral calculi to model the physical world, the analysis and description calculi presented in this article will be used to model primarily artifactual domains.'}",
-A Review of Three-Dimensional Printing in Tissue Engineering.,Nicholas A. Sears,"Recent advances in three-dimensional (3D) printing technologies have led to a rapid expansion of applications from the creation of anatomical training models for complex surgical procedures to the printing of tissue engineering constructs. In addition to achieving the macroscale geometry of organs and tissues, a print layer thickness as small as 20 μm allows for reproduction of the microarchitectures of bone and other tissues. Techniques with even higher precision are currently being investigated to enable reproduction of smaller tissue features such as hepatic lobules. Current research in tissue engineering focuses on the development of compatible methods (printers) and materials (bioinks) that are capable of producing biomimetic scaffolds. In this review, an overview of current 3D printing techniques used in tissue engineering is provided with an emphasis on the printing mechanism and the resultant scaffold characteristics. Current practical challenges and technical limitations are emphasized and future trends of bioprinting are discussed.","{'model': 'tldr@v2.0.0', 'text': 'An overview of current 3D printing techniques used in tissue engineering is provided with an emphasis on the printing mechanism and the resultant scaffold characteristics.'}",
-3D biofabrication strategies for tissue engineering and regenerative medicine.,P. Bajaj,"Over the past several decades, there has been an ever-increasing demand for organ transplants. However, there is a severe shortage of donor organs, and as a result of the increasing demand, the gap between supply and demand continues to widen. A potential solution to this problem is to grow or fabricate organs using biomaterial scaffolds and a person's own cells. Although the realization of this solution has been limited, the development of new biofabrication approaches has made it more realistic. This review provides an overview of natural and synthetic biomaterials that have been used for organ/tissue development. It then discusses past and current biofabrication techniques, with a brief explanation of the state of the art. Finally, the review highlights the need for combining vascularization strategies with current biofabrication techniques. Given the multitude of applications of biofabrication technologies, from organ/tissue development to drug discovery/screening to development of complex in vitro models of human diseases, these manufacturing technologies can have a significant impact on the future of medicine and health care.","{'model': 'tldr@v2.0.0', 'text': 'This review provides an overview of natural and synthetic biomaterials that have been used for organ/tissue development and highlights the need for combining vascularization strategies with current biofabrication techniques.'}",https://www.annualreviews.org/doi/pdf/10.1146/annurev-bioeng-071813-105155
-Model-Driven Software Engineering in Practice,Marco Brambilla,"Model based software development differs from the conventional software development process and used in conjunction with a range of agile techniques. Despite improvements in third generation programming languages and runtime platforms, the levels of abstraction at which PLAs are developed today remains low-level relative to the concepts and concerns within the application domains themselves, such as manually tracking the library dependency or ensuring component composition syntactical and semantic correctness. [1] A promising means to address this problem involves developing PLAs using model-driven engineering (MDE) [2] which involves systematic use of models as key design and implementation artifacts throughout the software lifecycle. Thus to induce the works on architectures and models, researchers have been working to create structured models. The literature in this area hence proliferates; where the current edition [3] tries to address some of the significant aspects.","{'model': 'tldr@v2.0.0', 'text': 'To induce the works on architectures and models, researchers have been working to create structured models, and the literature in this area hence proliferates; where the current edition tries to address some of the significant aspects.'}",https://link.springer.com/content/pdf/bfm:978-3-031-02549-5/1?pdf=chapter%20toc
-A Systematic Literature Review on Fault Prediction Performance in Software Engineering,T. Hall,"Background: The accurate prediction of where faults are likely to occur in code can help direct test effort, reduce costs, and improve the quality of software. Objective: We investigate how the context of models, the independent variables used, and the modeling techniques applied influence the performance of fault prediction models. Method: We used a systematic literature review to identify 208 fault prediction studies published from January 2000 to December 2010. We synthesize the quantitative and qualitative results of 36 studies which report sufficient contextual and methodological information according to the criteria we develop and apply. Results: The models that perform well tend to be based on simple modeling techniques such as Naive Bayes or Logistic Regression. Combinations of independent variables have been used by models that perform well. Feature selection has been applied to these combinations when models are performing particularly well. Conclusion: The methodology used to build models seems to be influential to predictive performance. Although there are a set of fault prediction studies in which confidence is possible, more studies are needed that use a reliable methodology and which report their context, methodology, and performance comprehensively.","{'model': 'tldr@v2.0.0', 'text': 'Although there are a set of fault prediction studies in which confidence is possible, more studies are needed that use a reliable methodology and which report their context, methodology, and performance comprehensively.'}",https://researchrepository.ul.ie/articles/online_resource/A_Systematic_literature_review_on_fault_prediction_performance_in_software_engineering/19848790/1/files/35264761.pdf
-"Mechanobiology in Tendon, Ligament, and Skeletal Muscle Tissue Engineering.",Michael T. K. Bramson,"Tendon, ligament, and skeletal muscle are highly-organized tissues that largely rely on a hierarchical collagenous matrix to withstand high tensile loads experienced in activities of daily life. This critical biomechanical role predisposes these tissues to injury, and current treatments fail to recapitulate the biomechanical function of native tissue. This has prompted researchers to pursue engineering functional tissue replacements, or dysfunction/disease/development models, by emulating in vivo stimuli within in vitro tissue engineering platforms; specifically mechanical stimulation, as well as active contraction in skeletal muscle. Mechanical loading is critical for matrix production and organization in the development, maturation, and maintenance of native tendon, ligament, and skeletal muscle, as well as their interfaces. Tissue engineers seek to harness these mechanobiological benefits using bioreactors to apply both static and dynamic mechanical stimulation to tissue constructs, and induce active contraction in engineered skeletal muscle. The vast majority of engineering approaches in these tissues are scaffold-based, providing interim structure and support to engineered constructs, and sufficient integrity to withstand mechanical loading. Alternatively, some recent studies have employed developmentally-inspired scaffold-free techniques, relying on cellular self-assembly and matrix production to form tissue constructs. Whether utilizing a scaffold or not, incorporation of mechanobiological stimuli has been shown to improve the composition, structure, and biomechanical function of engineered tendon, ligament, and skeletal muscle. Together, these findings highlight the importance of mechanobiology and suggest how it can be leveraged to engineer these tissues and their interfaces, and to create functional multi-tissue constructs.","{'model': 'tldr@v2.0.0', 'text': 'Together, these findings highlight the importance of mechanobiology and suggest how it can be leveraged to engineer these tissues and their interfaces, and to create functional multi-tissue constructs.'}",
-Game development software engineering process life cycle: a systematic review,Saiqa Aleem,,"{'model': 'tldr@v2.0.0', 'text': 'The results of this study suggest that the game development software engineering process has many aspects that need further attention from researchers; that especially includes the postproduction phase.'}",https://jserd.springeropen.com/track/pdf/10.1186/s40411-016-0032-7
-Analisis Strategi Menu Engineering dalam Upaya Meningkatkan Pendapatan Restaurant,Ida Bagus Ananta Wijaya,"Penelitian ini dilaksanakan untuk membantu pihak manajemen dalam mengevaluasi dan menganalisis menu a’la carte menggunakan metode menu engineering. Pihak manajemen belum menerapkankan analisis menu menggunakan metode tersebut, sehingga pendapatan kurang maksimal. Penelitian ini bertujuan untuk mengetahui klasifikasi menu, dan diikuti dengan strategi untuk menindak lanjuti menu dengan klasifikasi tersebut. Penelitian ini dilaksanakan di The Sayan House Restaurant yang dilaksanakan pada bulan Maret 2021 –Juni 2021, dengan menggunakan sales history dari bulan Desember 2020 – Februari 2021. Jenis penelitian ini adalah deskriptif kualitatif dan kuantitatif. Penelitian ini menggunakan sampel jenuh sebanyak 38 item menu. Data dikumpulkan dari wawancara dan studi dokumentasi, kemudian dianalisis menggunakan teknik analisis menu engineering, sehingga mendapatkan klasifikasi menu yaitu star, plow horse, puzzle, dan dog. Hasil penelitian menunjukkan dari 38 item menu yang dianalisis, 16 diantaranya dikategorikan star, 10 diantaranya dikategorikan plow horse, 5 diantaranya dikategorikan puzzle, dan 7 diantaranya dikategorikan dog. Berdasarkan hasil analisis tersebut dan menyikapi masing-masing klasifikasi menu, diperlukan strategi dan tindak lanjut untuk setiap klasifikasinya, serta melakukan evaluasi menu secara rutin oleh pihak manajemen The Sayan House Restaurant. This research was conducted to assist management in evaluating and analyzing a'la carte menu usingthe menu engineering method. The management has not implemented menu analysis using that method, so the revenue is less than maximum. This study aims to know the classification of the menu, and followed by a strategy to follow up the menu with that classification. This research was conducted at The Sayan House Restaurant conducted in March 2021 – June 2021, using sales history from December 2020 – February 2021. This type of research is descriptive qualitative and quantitative. The study used saturated samples of 38 menu items. The data is collected from interviews and documentation studies, then analyzed using engineering menu analysis techniques, to obtain menu classifications namely star, plow horse, puzzle, and dog. The results showed from 38 menu items analyzed, 16 of them were categorized as a star, 10 of them were categorized as a plow horse, 5 of them were categorized as a puzzles, and 7 of them were categorized as a dogs. Based on the results of the analysis and addressing each menu classification, strategy and follow-up for each classification are required, as well as conducting regular menu evaluations by the management of The Sayan House Restaurant.",,https://paris.ipb-intl.ac.id/index.php/paris/article/download/464/351
-Selection of Software Requirements Elicitation Techniques- A Systematic Review,Fazila Shams,"Article history: Received 27 November 2018 Received in revised form 2 January 2019 Accepted 9 January 2019 Available online 20 January 2019 Requirements elicitation is a basic action that frames some portion of the Requirements Engineering process. since it needs to find what the product must do through a strong comprehension of the desires and needs of the different stakeholders and then to change them into Requirements list. But, despite of its significance, only a couple of Systematic reviews are present in literature that give logical proof about the adequacy of the procedures used for Software Requirements Elicitation. This paper aims to present a systematic review of literature about the Elicitation techniques. The goal is to find out the mature techniques which are used for requirements elicitation and which techniques increases efficiency of the elicitation. Preceding literature expect that maturity prompts a better understanding of client’s requirements, and hence an improved probability that a subsequent product will fulfill those requirements. This review found 30 studies to find the answers and depicts which techniques are more effective and which techniques is suitable under which situations, considering the system to be developed and information obtained from other factors.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to present a systematic review of literature about the Elicitation techniques to find out the mature techniques which are used for requirements elicitation and which techniques increases efficiency of the elicitation.'}",
-A Survey of Feature Location Techniques,J. Rubin,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter discusses possible directions for leveraging SPLE architectures in order to improve the feature location process and describes their implementation strategies and exemplify the techniques on a realistic use-case.'}",http://www.cs.toronto.edu/~chechik/pubs/locationsurvey12.pdf
-Early diagnosis of Alzheimer's disease using machine learning techniques: A review paper,Aunsia Khan,"Alzheimer's, an irreparable brain disease, impairs thinking and memory while the aggregate mind size shrinks which at last prompts demise. Early diagnosis of AD is essential for the progress of more prevailing treatments. Machine learning (ML), a branch of artificial intelligence, employs a variety of probabilistic and optimization techniques that permits PCs to gain from vast and complex datasets. As a result, researchers focus on using machine learning frequently for diagnosis of early stages of AD. This paper presents a review, analysis and critical evaluation of the recent work done for the early detection of AD using ML techniques. Several methods achieved promising prediction accuracies, however they were evaluated on different pathologically unproven data sets from different imaging modalities making it difficult to make a fair comparison among them. Moreover, many other factors such as pre-processing, the number of important attributes for feature selection, class imbalance distinctively affect the assessment of the prediction accuracy. To overcome these limitations, a model is proposed which comprise of initial pre-processing step followed by imperative attributes selection and classification is achieved using association rule mining. Furthermore, this proposed model based approach gives the right direction for research in early diagnosis of AD and has the potential to distinguish AD from healthy controls.","{'model': 'tldr@v2.0.0', 'text': 'A model based approach is proposed which comprise of initial pre-processing step followed by imperative attributes selection and classification is achieved using association rule mining and has the potential to distinguish AD from healthy controls.'}",
-"Model-Driven Software Engineering in Practice, Second Edition",M. Brambilla,,"{'model': 'tldr@v2.0.0', 'text': 'This book is to provide an agile and flexible tool to introduce you to the MDSE world, thus allowing you to quickly understand its basic principles and techniques and to choose the right set of MDSE instruments for your needs so that you can start to benefit from MDSE right away.'}",https://link.springer.com/content/pdf/bfm:978-3-031-02549-5/1?pdf=chapter%20toc
-The design of scaffolds for use in tissue engineering. Part II. Rapid prototyping techniques.,Shoufeng Yang,"Tissue engineering (TE) is an important emerging area in biomedical engineering for creating biological alternatives for harvested tissues, implants, and prostheses. In TE, a highly porous artificial extracellular matrix or scaffold is required to accommodate mammalian cells and guide their growth and tissue regeneration in three-dimension (3D). However, existing 3D scaffolds for TE proved less than ideal for actual applications because they lack mechanical strength, interconnected channels, and controlled porosity or pores distribution. In this paper, the authors review the application and advancement of rapid prototyping (RP) techniques in the design and creation of synthetic scaffolds for use in TE. We also review the advantages and benefits, and limitations and shortcomings of current RP techniques as well as the future direction of RP development in TE scaffold fabrication.","{'model': 'tldr@v2.0.0', 'text': 'The authors review the application and advancement of rapid prototyping (RP) techniques in the design and creation of synthetic scaffolds for use in TE and the advantages and benefits, and limitations and shortcomings of current RP techniques as well as the future direction of RP development in TE scaffold fabrication.'}",https://lirias.kuleuven.be/bitstream/123456789/556008/1/The%20design%20of%20scaffolds%20for%20use%20in%20tissue%20engineering.%20Part%20II.%20Rapid%20prototyping%20techniques.pdf
-Cell surface engineering and application in cell delivery to heart diseases,Daniel Y. Lee,,"{'model': 'tldr@v2.0.0', 'text': 'Current methods of non-genetic cell surface modification, including chemical conjugations, polymeric encapsulation, hydrophobic insertion, enzymatic and metabolic addition, will be introduced.'}",https://jbioleng.biomedcentral.com/track/pdf/10.1186/s13036-018-0123-6
-"Machine learning, medical diagnosis, and biomedical engineering research - commentary",K. Foster,,"{'model': 'tldr@v2.0.0', 'text': 'This commentary is intended to help sensitize investigators as well as readers and reviewers of papers to some potential pitfalls in the development of classifiers, and suggests steps that researchers can take to help avoid these problems.'}",https://biomedical-engineering-online.biomedcentral.com/counter/pdf/10.1186/1475-925X-13-94
-State-of-the-Art Review of 3D Bioprinting for Cardiovascular Tissue Engineering,B. Duan,,"{'model': 'tldr@v2.0.0', 'text': 'Current challenges and further perspectives are discussed to help guide the bioink and bioprinter development, improve bioprinting strategies and direct future organ biopprinting and translational applications.'}",
-Whole-organ tissue engineering: decellularization and recellularization of three-dimensional matrix scaffolds.,S. Badylak,"The definitive treatment for end-stage organ failure is orthotopic transplantation. However, the demand for transplantation far exceeds the number of available donor organs. A promising tissue-engineering/regenerative-medicine approach for functional organ replacement has emerged in recent years. Decellularization of donor organs such as heart, liver, and lung can provide an acellular, naturally occurring three-dimensional biologic scaffold material that can then be seeded with selected cell populations. Preliminary studies in animal models have provided encouraging results for the proof of concept. However, significant challenges for three-dimensional organ engineering approach remain. This manuscript describes the fundamental concepts of whole-organ engineering, including characterization of the extracellular matrix as a scaffold, methods for decellularization of vascular organs, potential cells to reseed such a scaffold, techniques for the recellularization process and important aspects regarding bioreactor design to support this approach. Critical challenges and future directions are also discussed.","{'model': 'tldr@v2.0.0', 'text': 'The fundamental concepts of whole-organ engineering, including characterization of the extracellular matrix as a scaffold, methods for decellularization of vascular organs, potential cells to reseed such a scaffolds, techniques for the recellularization process and important aspects regarding bioreactor design to support this approach are described.'}",
-Comparison and evaluation of goal-oriented satisfaction analysis techniques,Jennifer Horkoff,,"{'model': 'tldr@v2.0.0', 'text': 'This comparison shows that different satisfaction analysis techniques for goal models can produce variable results, depending on the structure of the model, and recommends the use of satisfaction analysis technique as only heuristics for decision making.'}",
-Metamodels for Computer-based Engineering Design: Survey and recommendations,T. Simpson,,"{'model': 'tldr@v2.0.0', 'text': 'This paper surveys their existing application in engineering design, and addresses the dangers of applying traditional statistical techniques to approximate deterministic computer analysis codes, along with recommendations for the appropriate use of statistical approximation techniques in given situations.'}",https://ntrs.nasa.gov/api/citations/19990087092/downloads/19990087092.pdf
-Towards modelling and reasoning support for early-phase requirements engineering,E. Yu,"Requirements are usually understood as stating what a system is supposed to do, as apposed to how it should do it. However, understanding the organizational context and rationales (the ""Whys"") that lead up to systems requirements can be just as important for the ongoing success of the system. Requirements modelling techniques can be used to help deal with the knowledge and reasoning needed in this earlier phase of requirements engineering. However most existing requirements techniques are intended more for the later phase of requirements engineering, which focuses on completeness, consistency, and automated verification of requirements. In contrast, the early phase aims to model and analyze stakeholder interests and how they might be addressed, or compromised, by various system-and-environment alternatives. This paper argues, therefore, that a different kind of modelling and reasoning support is needed for the early phase. An outline of the i* framework is given as an example of a step in this direction. Meeting scheduling is used as a domain example.","{'model': 'tldr@v2.0.0', 'text': 'This paper argues that a different kind of modelling and reasoning support is needed for the early phase of requirements engineering, which aims to model and analyze stakeholder interests and how they might be addressed, or compromised, by various system-and-environment alternatives.'}",
-SGP-TOD: Building Task Bots Effortlessly via Schema-Guided LLM Prompting,Xiaoying Zhang,"Building end-to-end task bots and maintaining their integration with new functionalities using minimal human efforts is a long-standing challenge in dialog research. Recently large language models (LLMs) have demonstrated exceptional proficiency in conversational engagement and adherence to instructions across various downstream tasks. In this work, we introduce SGP-TOD, Schema-Guided Prompting for building Task-Oriented Dialog systems effortlessly based on LLMs. Utilizing the symbolic knowledge -- task schema, we instruct fixed LLMs to generate appropriate responses on novel tasks, circumventing the need for training data. Specifically, SGP-TOD comprises three components: a LLM for engaging with users, a DST Prompter to aid the LLM with dialog state tracking, which is then used to retrieve database items, and a Policy Prompter to elicit proper responses adhering to the provided dialog policy. Experimental results on Multiwoz, RADDLE and STAR datasets show that our training-free strategy SGP-TOD, without any task-specific data, yields state-of-the-art (SOTA) zero-shot performance, greatly surpasses the few-shot approaches. In a domain-extension setting, SGP-TOD aptly adapts to new functionalities by merely adding supplementary schema rules. We make our code and data publicly available.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on Multiwoz, RADDLE and STAR datasets show that the training-free strategy SGP-TOD, without any task-specific data, yields state-of-the-art (SOTA) zero-shot performance, greatly surpasses the few-shot approaches.'}",http://arxiv.org/pdf/2305.09067
-Iterative Zero-Shot LLM Prompting for Knowledge Graph Construction,S. Carta,"In the current digitalization era, capturing and effectively representing knowledge is crucial in most real-world scenarios. In this context, knowledge graphs represent a potent tool for retrieving and organizing a vast amount of information in a properly interconnected and interpretable structure. However, their generation is still challenging and often requires considerable human effort and domain expertise, hampering the scalability and flexibility across different application fields. This paper proposes an innovative knowledge graph generation approach that leverages the potential of the latest generative large language models, such as GPT-3.5, that can address all the main critical issues in knowledge graph building. The approach is conveyed in a pipeline that comprises novel iterative zero-shot and external knowledge-agnostic strategies in the main stages of the generation process. Our unique manifold approach may encompass significant benefits to the scientific community. In particular, the main contribution can be summarized by: (i) an innovative strategy for iteratively prompting large language models to extract relevant components of the final graph; (ii) a zero-shot strategy for each prompt, meaning that there is no need for providing examples for""guiding""the prompt result; (iii) a scalable solution, as the adoption of LLMs avoids the need for any external resources or human expertise. To assess the effectiveness of our proposed model, we performed experiments on a dataset that covered a specific domain. We claim that our proposal is a suitable solution for scalable and versatile knowledge graph construction and may be applied to different and novel contexts.","{'model': 'tldr@v2.0.0', 'text': 'An innovative knowledge graph generation approach that leverages the potential of the latest generative large language models, such as GPT-3.5, that can address all the main critical issues in knowledge graph building.'}",http://arxiv.org/pdf/2307.01128
-"Better Patching Using LLM Prompting, via Self-Consistency",Toufique Ahmed,"Large Language models (LLMs) can be induced to solve non-trivial problems with “few-shot” prompts including illustrative problem-solution examples. Now if the few-shots also include “chain of thought” ($\mathcal{C}oT$) explanations, which are of the form problem-explanation-solution, LLMs will generate a “explained” solution, and perform even better. Recently an exciting, substantially better technique, self-consistency [1] ($\mathcal{S}-C$) has emerged, based on the intuition that there are many plausible explanations for the right solution; when the LLM is sampled repeatedly to generate a pool of explanation-solution pairs, for a given problem, the most frequently occurring solutions in the pool (ignoring the explanations) tend to be even more likely to be correct! Unfortunately, the use of this highly-performant $\mathcal{S}-C$ (or even $\mathcal{C}oT$) approach in software engineering settings is hampered by the lack of explanations; most software datasets lack explanations. In this paper, we describe an application of the $\mathcal{S}-C$ approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots. We achieve state-of-the art results, beating previous approaches to prompting-based program repair, on the MODIT dataset; we also find evidence suggesting that the correct commit messages are helping the LLM learn to produce better patches.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes an application of the $\\mathcal{S}-C$ approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots, on the MODIT dataset.'}",https://arxiv.org/pdf/2306.00108
-Simple LLM Prompting is State-of-the-Art for Robust and Multilingual Dialogue Evaluation,J. Mendoncca,"Despite significant research effort in the development of automatic dialogue evaluation metrics, little thought is given to evaluating dialogues other than in English. At the same time, ensuring metrics are invariant to semantically similar responses is also an overlooked topic. In order to achieve the desired properties of robustness and multilinguality for dialogue evaluation metrics, we propose a novel framework that takes advantage of the strengths of current evaluation models with the newly-established paradigm of prompting Large Language Models (LLMs). Empirical results show our framework achieves state of the art results in terms of mean Spearman correlation scores across several benchmarks and ranks first place on both the Robust and Multilingual tasks of the DSTC11 Track 4 “Automatic Evaluation Metrics for Open-Domain Dialogue Systems”, proving the evaluation capabilities of prompted LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework that takes advantage of the strengths of current evaluation models with the newly-established paradigm of prompting Large Language Models (LLMs) to achieve the desired properties of robustness and multilinguality for dialogue evaluation metrics.'}",https://arxiv.org/pdf/2308.16797
-LPML: LLM-Prompting Markup Language for Mathematical Reasoning,Ryutaro Yamauchi,"In utilizing large language models (LLMs) for mathematical reasoning, addressing the errors in the reasoning and calculation present in the generated text by LLMs is a crucial challenge. In this paper, we propose a novel framework that integrates the Chain-of-Thought (CoT) method with an external tool (Python REPL). We discovered that by prompting LLMs to generate structured text in XML-like markup language, we could seamlessly integrate CoT and the external tool and control the undesired behaviors of LLMs. With our approach, LLMs can utilize Python computation to rectify errors within CoT. We applied our method to ChatGPT (GPT-3.5) to solve challenging mathematical problems and demonstrated that combining CoT and Python REPL through the markup language enhances the reasoning capability of LLMs. Our approach enables LLMs to write the markup language and perform advanced mathematical reasoning using only zero-shot prompting.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposed a novel framework that integrates the Chain-of-Thought (CoT) method with an external tool (Python REPL) and discovered that by prompting LLMs to generate structured text in XML-like markup language, it could seamlessly integrate CoT and the external tool and control the undesired behaviors of LLMs.'}",https://arxiv.org/pdf/2309.13078
-Towards Understanding Chain-of-Thought Prompting: An Empirical Study of What Matters,Boshi Wang,"Chain-of-Thought (CoT) prompting can dramatically improve the multi-step reasoning abilities of large language models (LLMs). CoT explicitly encourages the LLM to generate intermediate rationales for solving a problem, by providing a series of reasoning steps in the demonstrations. Despite its success, there is still little understanding of what makes CoT prompting effective and which aspects of the demonstrated reasoning steps contribute to its performance. In this paper, we show that CoT reasoning is possible even with invalid demonstrations - prompting with invalid reasoning steps can achieve over 80-90% of the performance obtained using CoT under various metrics, while still generating coherent lines of reasoning during inference. Further experiments show that other aspects of the rationales, such as being relevant to the query and correctly ordering the reasoning steps, are much more important for effective CoT reasoning. Overall, these findings both deepen our understanding of CoT prompting, and open up new questions regarding LLMs’ capability to learn to reason in context.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that CoT reasoning is possible even with invalid demonstrations - prompting with invalid reasoning steps can achieve over 80-90% of the performance obtained using CoT under various metrics, while still generating coherent lines of reasoning during inference.'}",http://arxiv.org/pdf/2212.10001
-Improving Language Model Prompting in Support of Semi-autonomous Task Learning,James R. Kirk,"Language models (LLMs) offer potential as a source of knowledge for agents that need to acquire new task competencies within a performance environment. We describe efforts toward a novel agent capability that can construct cues (or""prompts"") that result in useful LLM responses for an agent learning a new task. Importantly, responses must not only be""reasonable""(a measure used commonly in research on knowledge extraction from LLMs) but also specific to the agent's task context and in a form that the agent can interpret given its native language capacities. We summarize a series of empirical investigations of prompting strategies and evaluate responses against the goals of targeted and actionable responses for task learning. Our results demonstrate that actionable task knowledge can be obtained from LLMs in support of online agent task learning.","{'model': 'tldr@v2.0.0', 'text': 'This work describes efforts toward a novel agent capability that can construct cues that result in useful LLM responses for an agent learning a new task.'}",https://arxiv.org/pdf/2209.07636
-ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate,Chi-Min Chan,"Text evaluation has historically posed significant challenges, often demanding substantial labor and time cost. With the emergence of large language models (LLMs), researchers have explored LLMs' potential as alternatives for human evaluation. While these single-agent-based approaches show promise, experimental results suggest that further advancements are needed to bridge the gap between their current effectiveness and human-level evaluation quality. Recognizing that best practices of human evaluation processes often involve multiple human annotators collaborating in the evaluation, we resort to a multi-agent debate framework, moving beyond single-agent prompting strategies. The multi-agent-based approach enables a group of LLMs to synergize with an array of intelligent counterparts, harnessing their distinct capabilities and expertise to enhance efficiency and effectiveness in handling intricate tasks. In this paper, we construct a multi-agent referee team called ChatEval to autonomously discuss and evaluate the quality of generated responses from different models on open-ended questions and traditional natural language generation (NLG) tasks. Our analysis shows that ChatEval transcends mere textual scoring, offering a human-mimicking evaluation process for reliable assessments. Our code is available at https://github.com/chanchimin/ChatEval.","{'model': 'tldr@v2.0.0', 'text': 'A multi-agent referee team called ChatEval is constructed to autonomously discuss and evaluate the quality of generated responses from different models on open-ended questions and traditional natural language generation (NLG) tasks, offering a human-mimicking evaluation process for reliable assessments.'}",https://arxiv.org/pdf/2308.07201
-Boosting Theory-of-Mind Performance in Large Language Models via Prompting,Shima Rahimi Moghaddam,"Large language models (LLMs) excel in many tasks in 2023, but they still face challenges in complex reasoning. Theory-of-mind (ToM) tasks, which require understanding agents' beliefs, goals, and mental states, are essential for common-sense reasoning involving humans, making it crucial to enhance LLM performance in this area. This study measures the ToM performance of GPT-4 and three GPT-3.5 variants (Davinci-2, Davinci-3, GPT-3.5-Turbo), and investigates the effectiveness of in-context learning in improving their ToM comprehension. We evaluated prompts featuring two-shot chain of thought reasoning and step-by-step thinking instructions. We found that LLMs trained with Reinforcement Learning from Human Feedback (RLHF) (all models excluding Davinci-2) improved their ToM accuracy via in-context learning. GPT-4 performed best in zero-shot settings, reaching nearly 80% ToM accuracy, but still fell short of the 87% human accuracy on the test set. However, when supplied with prompts for in-context learning, all RLHF-trained LLMs exceeded 80% ToM accuracy, with GPT-4 reaching 100%. These results demonstrate that appropriate prompting enhances LLM ToM reasoning, and they underscore the context-dependent nature of LLM cognitive capacities.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that appropriate prompting enhances LLM ToM reasoning, and the context-dependent nature of LLM cognitive capacities is underscored, as it is found that LLMs trained with Reinforcement Learning from Human Feedback improved their ToM accuracy via in-context learning.'}",http://arxiv.org/pdf/2304.11490
-3D-LLM: Injecting the 3D World into Large Language Models,Yining Hong,"Large language models (LLMs) and Vision-Language Models (VLMs) have been proven to excel at multiple tasks, such as commonsense reasoning. Powerful as these models can be, they are not grounded in the 3D physical world, which involves richer concepts such as spatial relationships, affordances, physics, layout, and so on. In this work, we propose to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs. Specifically, 3D-LLMs can take 3D point clouds and their features as input and perform a diverse set of 3D-related tasks, including captioning, dense captioning, 3D question answering, task decomposition, 3D grounding, 3D-assisted dialog, navigation, and so on. Using three types of prompting mechanisms that we design, we are able to collect over 300k 3D-language data covering these tasks. To efficiently train 3D-LLMs, we first utilize a 3D feature extractor that obtains 3D features from rendered multi-view images. Then, we use 2D VLMs as our backbones to train our 3D-LLMs. By introducing a 3D localization mechanism, 3D-LLMs can better capture 3D spatial information. Experiments on ScanQA show that our model outperforms state-of-the-art baselines by a large margin ( e.g. , the BLEU-1 score surpasses state-of-the-art score by 9%). Furthermore, experiments on our held-in datasets for 3D captioning, task composition, and 3D-assisted dialogue show that our model outperforms 2D VLMs. Qualitative examples also show that our model could perform more tasks beyond the scope of existing LLMs and VLMs. Project Page: : https:","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs, and introduces a 3D localization mechanism, which can better capture 3D spatial information.'}",https://arxiv.org/pdf/2307.12981
-Empowering LLM-based Machine Translation with Cultural Awareness,Binwei Yao,"Traditional neural machine translation (NMT) systems often fail to translate sentences that contain culturally specific information. Most previous NMT methods have incorporated external cultural knowledge during training, which requires fine-tuning on low-frequency items specific to the culture. Recent in-context learning utilizes lightweight prompts to guide large language models (LLMs) to perform machine translation, however, whether such an approach works in terms of injecting culture awareness into machine translation remains unclear. To this end, we introduce a new data curation pipeline to construct a culturally relevant parallel corpus, enriched with annotations of cultural-specific entities. Additionally, we design simple but effective prompting strategies to assist this LLM-based translation. Extensive experiments show that our approaches can largely help incorporate cultural knowledge into LLM-based machine translation, outperforming traditional NMT systems in translating cultural-specific sentences.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that the approaches can largely help incorporate cultural knowledge into LLM-based machine translation, outperforming traditional NMT systems in translating cultural-specific sentences.'}",http://arxiv.org/pdf/2305.14328
-SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models,S. S. Kannan,"In this work, we introduce SMART-LLM, an innovative framework designed for embodied multi-robot task planning. SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models (LLMs), harnesses the power of LLMs to convert high-level task instructions provided as input into a multi-robot task plan. It accomplishes this by executing a series of stages, including task decomposition, coalition formation, and task allocation, all guided by programmatic LLM prompts within the few-shot prompting paradigm. We create a benchmark dataset designed for validating the multi-robot task planning problem, encompassing four distinct categories of high-level instructions that vary in task complexity. Our evaluation experiments span both simulation and real-world scenarios, demonstrating that the proposed model can achieve promising results for generating multi-robot task plans. The experimental videos, code, and datasets from the work can be found at https://sites.google.com/view/smart-llm/.","{'model': 'tldr@v2.0.0', 'text': 'SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models (LLMs), harnesses the power of LLMs to convert high-level task instructions provided as input into a multi-robot task plan by executing a series of stages.'}",https://arxiv.org/pdf/2309.10062
-LLM-Deliberation: Evaluating LLMs with Interactive Multi-Agent Negotiation Games,Sahar Abdelnabi,"There is a growing interest in using Large Language Models (LLMs) as agents to tackle real-world tasks that may require assessing complex situations. Yet, we have a limited understanding of LLMs' reasoning and decision-making capabilities, partly stemming from a lack of dedicated evaluation benchmarks. As negotiating and compromising are key aspects of our everyday communication and collaboration, we propose using scorable negotiation games as a new evaluation framework for LLMs. We create a testbed of diverse text-based, multi-agent, multi-issue, semantically rich negotiation games, with easily tunable difficulty. To solve the challenge, agents need to have strong arithmetic, inference, exploration, and planning capabilities, while seamlessly integrating them. Via a systematic zero-shot Chain-of-Thought prompting (CoT), we show that agents can negotiate and consistently reach successful deals. We quantify the performance with multiple metrics and observe a large gap between GPT-4 and earlier models. Importantly, we test the generalization to new games and setups. Finally, we show that these games can help evaluate other critical aspects, such as the interaction dynamics between agents in the presence of greedy and adversarial players.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using scorable negotiation games as a new evaluation framework for LLMs, and shows that these games can help evaluate other critical aspects, such as the interaction dynamics between agents in the presence of greedy and adversarial players.'}",https://arxiv.org/pdf/2309.17234
-Chainpoll: A high efficacy method for LLM hallucination detection,Robert Friel,"Large language models (LLMs) have experienced notable advancements in generating coherent and contextually relevant responses. However, hallucinations - incorrect or unfounded claims - are still prevalent, prompting the creation of automated metrics to detect these in LLM outputs. Our contributions include: introducing ChainPoll, an innovative hallucination detection method that excels compared to its counterparts, and unveiling RealHall, a refined collection of benchmark datasets to assess hallucination detection metrics from recent studies. While creating RealHall, we assessed tasks and datasets from previous hallucination detection studies and observed that many are not suitable for the potent LLMs currently in use. Overcoming this, we opted for four datasets challenging for modern LLMs and pertinent to real-world scenarios. Using RealHall, we conducted a comprehensive comparison of ChainPoll with numerous hallucination metrics from recent studies. Our findings indicate that ChainPoll outperforms in all RealHall benchmarks, achieving an overall AUROC of 0.781. This surpasses the next best theoretical method by 11% and exceeds industry standards by over 23%. Additionally, ChainPoll is cost-effective and offers greater transparency than other metrics. We introduce two novel metrics to assess LLM hallucinations: Adherence and Correctness. Adherence is relevant to Retrieval Augmented Generation workflows, evaluating an LLM's analytical capabilities within given documents and contexts. In contrast, Correctness identifies logical and reasoning errors.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces two novel metrics to assess LLM hallucinations: Adherence and Correctness, and unveils RealHall, a refined collection of benchmark datasets to assess hallucination detection metrics from recent studies.'}",
-Put Your Money Where Your Mouth Is: Evaluating Strategic Planning and Execution of LLM Agents in an Auction Arena,Jiangjie Chen,"Can Large Language Models (LLMs) simulate human behavior in complex environments? LLMs have recently been shown to exhibit advanced reasoning skills but much of NLP evaluation still relies on static benchmarks. Answering this requires evaluation environments that probe strategic reasoning in competitive, dynamic scenarios that involve long-term planning. We introduce AucArena, a novel simulation environment for evaluating LLMs within auctions, a setting chosen for being highly unpredictable and involving many skills related to resource and risk management, while also being easy to evaluate. We conduct several controlled simulations using state-of-the-art LLMs as bidding agents. We find that through simple prompting, LLMs do indeed demonstrate many of the skills needed for effectively engaging in auctions (e.g., managing budget, adhering to long-term goals and priorities), skills that we find can be sharpened by explicitly encouraging models to be adaptive and observe strategies in past auctions. These results are significant as they show the potential of using LLM agents to model intricate social dynamics, especially in competitive settings. However, we also observe considerable variability in the capabilities of individual LLMs. Notably, even our most advanced models (GPT-4) are occasionally surpassed by heuristic baselines and human agents, highlighting the potential for further improvements in the design of LLM agents and the important role that our simulation environment can play in further testing and refining agent architectures.","{'model': 'tldr@v2.0.0', 'text': 'AucArena, a novel simulation environment for evaluating LLMs within auctions, is introduced, finding that through simple prompting, LLMs do indeed demonstrate many of the skills needed for effectively engaging in auctions and can be sharpened by explicitly encouraging models to be adaptive and observe strategies in past auctions.'}",https://arxiv.org/pdf/2310.05746
-Prompting Large Language Models with Speech Recognition Abilities,Yassir Fathullah,"Large language models have proven themselves highly flexible, able to solve a wide range of generative tasks, such as abstractive summarization and open-ended question answering. In this paper we extend the capabilities of LLMs by directly attaching a small audio encoder allowing it to perform speech recognition. By directly prepending a sequence of audial embeddings to the text token embeddings, the LLM can be converted to an automatic speech recognition (ASR) system, and be used in the exact same manner as its textual counterpart. Experiments on Multilingual LibriSpeech (MLS) show that incorporating a conformer encoder into the open sourced LLaMA-7B allows it to outperform monolingual baselines by 18% and perform multilingual speech recognition despite LLaMA being trained overwhelmingly on English text. Furthermore, we perform ablation studies to investigate whether the LLM can be completely frozen during training to maintain its original capabilities, scaling up the audio encoder, and increasing the audio encoder striding to generate fewer embeddings. The results from these studies show that multilingual ASR is possible even when the LLM is frozen or when strides of almost 1 second are used in the audio encoder opening up the possibility for LLMs to operate on long-form audio.","{'model': 'tldr@v2.0.0', 'text': 'The capabilities of LLMs are extended by directly attaching a small audio encoder allowing it to perform speech recognition and it is shown that multilingual ASR is possible even when the LLM is frozen or when strides of almost 1 second are used in theaudio encoder opening up the possibility for LLMs to operate on long-form audio.'}",https://arxiv.org/pdf/2307.11795
-"See, Think, Confirm: Interactive Prompting Between Vision and Language Models for Knowledge-based Visual Reasoning",Zhenfang Chen,"Large pre-trained vision and language models have demonstrated remarkable capacities for various tasks. However, solving the knowledge-based visual reasoning tasks remains challenging, which requires a model to comprehensively understand image content, connect the external world knowledge, and perform step-by-step reasoning to answer the questions correctly. To this end, we propose a novel framework named Interactive Prompting Visual Reasoner (IPVR) for few-shot knowledge-based visual reasoning. IPVR contains three stages, see, think and confirm. The see stage scans the image and grounds the visual concept candidates with a visual perception model. The think stage adopts a pre-trained large language model (LLM) to attend to the key concepts from candidates adaptively. It then transforms them into text context for prompting with a visual captioning model and adopts the LLM to generate the answer. The confirm stage further uses the LLM to generate the supporting rationale to the answer, verify the generated rationale with a cross-modality classifier and ensure that the rationale can infer the predicted output consistently. We conduct experiments on a range of knowledge-based visual reasoning datasets. We found our IPVR enjoys several benefits, 1). it achieves better performance than the previous few-shot learning baselines; 2). it enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step; 3). it is computation-efficient compared with other fine-tuning baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework named Interactive Prompting Visual Reasoner (IPVR), which achieves better performance than the previous few-shot learning baselines, and enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step.'}",http://arxiv.org/pdf/2301.05226
-Satisfiability-Aided Language Models Using Declarative Prompting,Xi Ye,"Prior work has combined chain-of-thought prompting in large language models (LLMs) with programmatic representations to perform effective and transparent reasoning. While such an approach works well for tasks that only require forward reasoning (e.g., straightforward arithmetic), it is less effective for constraint solving problems that require more sophisticated planning and search. In this paper, we propose a new satisfiability-aided language modeling (SatLM) approach for improving the reasoning capabilities of LLMs. We use an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer. This approach has two key advantages. The declarative specification is closer to the problem description than the reasoning steps are, so the LLM can parse it out of the description more accurately. Furthermore, by offloading the actual reasoning task to an automated theorem prover, our approach can guarantee the correctness of the answer with respect to the parsed specification and avoid planning errors in the solving process. We evaluate SATLM on 8 different datasets and show that it consistently outperforms program-aided LMs in the imperative paradigm. In particular, SATLM outperforms program-aided LMs by 23% on a challenging subset of the GSM arithmetic reasoning dataset; SATLM also achieves a new SoTA on LSAT and BoardgameQA, surpassing previous models that are trained on the respective training sets.","{'model': 'tldr@v2.0.0', 'text': 'A new satisfiability-aided language modeling (SatLM) approach for improving the reasoning capabilities of LLMs that uses an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer.'}",https://arxiv.org/pdf/2305.09656
-"Prompting and Evaluating Large Language Models for Proactive Dialogues: Clarification, Target-guided, and Non-collaboration",Yang Deng,"Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, despite their impressive capabilities, they still possess limitations, such as providing randomly-guessed answers to ambiguous queries or failing to refuse users' requests, both of which are considered aspects of a conversational agent's proactivity. This raises the question of whether LLM-based conversational systems are equipped to handle proactive dialogue problems. In this work, we conduct a comprehensive analysis of LLM-based conversational systems, specifically focusing on three aspects of proactive dialogue systems: clarification, target-guided, and non-collaborative dialogues. To trigger the proactivity of LLMs, we propose the Proactive Chain-of-Thought prompting scheme, which augments LLMs with the goal planning capability over descriptive reasoning chains. Empirical findings are discussed to promote future studies on LLM-based proactive dialogue systems.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive analysis of LLM-based conversational systems, specifically focusing on three aspects of proactive dialogue systems: clarification, target-guided, and non-collaborative dialogues, and the Proactive Chain-of-Thought prompting scheme is proposed.'}",http://arxiv.org/pdf/2305.13626
-Query Expansion by Prompting Large Language Models,R. Jagerman,"Query expansion is a widely used technique to improve the recall of search systems. In this paper, we propose an approach to query expansion that leverages the generative abilities of Large Language Models (LLMs). Unlike traditional query expansion approaches such as Pseudo-Relevance Feedback (PRF) that relies on retrieving a good set of pseudo-relevant documents to expand queries, we rely on the generative and creative abilities of an LLM and leverage the knowledge inherent in the model. We study a variety of different prompts, including zero-shot, few-shot and Chain-of-Thought (CoT). We find that CoT prompts are especially useful for query expansion as these prompts instruct the model to break queries down step-by-step and can provide a large number of terms related to the original query. Experimental results on MS-MARCO and BEIR demonstrate that query expansions generated by LLMs can be more powerful than traditional query expansion methods.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on MS-MARCO and BEIR demonstrate that query expansions generated by LLMs can be more powerful than traditional query expansion methods.'}",http://arxiv.org/pdf/2305.03653
-PiVe: Prompting with Iterative Verification Improving Graph-based Generative Capability of LLMs,Jiuzhou Han,"Large language models (LLMs) have shown great abilities of solving various natural language tasks in different domains. Due to the training objective of LLMs and their pretraining data, LLMs are not very well equipped for tasks involving structured data generation. We propose a framework, Prompting with Iterative Verification (PiVe), to improve graphbased generative capability of LLMs. We show how a small language model could be trained to act as a verifier module for the output of an LLM (i.e., ChatGPT), and to iteratively improve its performance via fine-grained corrective instructions. Additionally, we show how the verifier module could apply iterative corrections offline for a more cost-effective solution to the text-to-graph generation task. Experiments on three graph-based datasets show consistent improvement gained via PiVe. Additionally, we highlight how the proposed verifier module can be used as a data augmentation tool to help improve the quality of automatically generated parallel text-graph datasets. Our code and data are available at https://github.com/Jiuzhouh/PiVe.","{'model': 'tldr@v2.0.0', 'text': 'A framework, Prompting with Iterative Verification (PiVe), to improve graphbased generative capability of LLMs and shows how a small language model could be trained to act as a verifier module for the output of an LLM, and to iteratively improve its performance via fine-grained corrective instructions.'}",http://arxiv.org/pdf/2305.12392
-Enhancing Small Medical Learners with Privacy-preserving Contextual Prompting,Xinlu Zhang,"Large language models (LLMs) demonstrate remarkable medical expertise, but data privacy concerns impede their direct use in healthcare environments. Although offering improved data privacy protection, domain-specific small language models (SLMs) often underperform LLMs, emphasizing the need for methods that reduce this performance gap while alleviating privacy concerns. In this paper, we present a simple yet effective method that harnesses LLMs' medical proficiency to boost SLM performance in medical tasks under privacy-restricted scenarios. Specifically, we mitigate patient privacy issues by extracting keywords from medical data and prompting the LLM to generate a medical knowledge-intensive context by simulating clinicians' thought processes. This context serves as additional input for SLMs, augmenting their decision-making capabilities. Our method significantly enhances performance in both few-shot and full training settings across three medical knowledge-intensive tasks, achieving up to a 22.57% increase in absolute accuracy compared to SLM fine-tuning without context, and sets new state-of-the-art results in two medical tasks within privacy-restricted scenarios. Further out-of-domain testing and experiments in two general domain datasets showcase its generalizability and broad applicability.","{'model': 'tldr@v2.0.0', 'text': ""This paper mitigate patient privacy issues by extracting keywords from medical data and prompting the LLM to generate a medical knowledge-intensive context by simulating clinicians' thought processes, which serves as additional input for SLMs, augmenting their decision-making capabilities.""}",http://arxiv.org/pdf/2305.12723
-Grammar Prompting for Domain-Specific Language Generation with Large Language Models,Bailin Wang,"Large language models (LLMs) can learn to perform a wide range of natural language tasks from just a handful of in-context examples. However, for generating strings from highly structured languages (e.g., semantic parsing to complex domain-specific languages), it is challenging for the LLM to generalize from just a few exemplars. We explore $\textbf{grammar prompting}$ as a simple approach for enabling LLMs to use external knowledge and domain-specific constraints, expressed through a grammar expressed in Backus--Naur Form (BNF), during in-context learning. Grammar prompting augments each demonstration example with a specialized grammar that is minimally sufficient for generating the particular output example, where the specialized grammar is a subset of the full DSL grammar. For inference, the LLM first predicts a BNF grammar given a test input, and then generates the output according to the rules of the grammar. Experiments demonstrate that grammar prompting can enable LLMs to perform competitively on a diverse set of DSL generation tasks, including semantic parsing (SMCalFlow, Overnight, GeoQuery), PDDL planning, and even molecule generation (SMILES).","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that grammar prompting can enable LLMs to perform competitively on a diverse set of DSL generation tasks, including semantic parsing, Overnight, GeoQuery, PDDL planning, and even molecule generation (SMILES).'}",http://arxiv.org/pdf/2305.19234
-Allies: Prompting Large Language Model with Beam Search,Hao-Lun Sun,"With the advance of large language models (LLMs), the research field of LLM applications becomes more and more popular and the idea of constructing pipelines to accomplish complex tasks by stacking LLM API calls come true. However, this kind of methods face two limitations: narrow information coverage and low fault tolerance. In this work, we propose a novel method called ALLIES. Given an input query, ALLIES leverages LLMs to iteratively generate new queries related to the original query, enabling an iterative reasoning process. By iteratively refining and expanding the scope of the original query, ALLIES captures and utilizes hidden knowledge that may not be directly obtainable through retrieval. We take zero-shot open-domain question answering (ODQA) as an application scene and evaluate ALLIES on the widely-used benchmarks, such as NQ, WebQ and TriviaQA. The experimental results demonstrate that ALLIES significantly outperforms other zero-shot baselines, indicating its effectiveness in tackling those challenges. Our code is available in https://github.com/microsoft/SimXNS/tree/main/ALLIES.","{'model': 'tldr@v2.0.0', 'text': 'Allies leverages LLMs to iteratively generate new queries related to the original query, enabling an iterative reasoning process and captures and utilizes hidden knowledge that may not be directly obtainable through retrieval.'}",
-Prompting Language-Informed Distribution for Compositional Zero-Shot Learning,Wentao Bao,"Compositional zero-shot learning (CZSL) task aims to recognize unseen compositional visual concepts, e.g., sliced tomatoes, where the model is learned only from the seen compositions, e.g., sliced potatoes and red tomatoes. Thanks to the prompt tuning on large pre-trained visual language models such as CLIP, recent literature shows impressively better CZSL performance than traditional vision-based methods. However, the key aspects that impact the generalization to unseen compositions, including the diversity and informativeness of class context, and the entanglement between visual primitives, i.e., state and object, are not properly addressed in existing CLIP-based CZSL literature. In this paper, we propose a model by prompting the language-informed distribution, aka., PLID, for the CZSL task. Specifically, the PLID leverages pre-trained large language models (LLM) to 1) formulate the language-informed class distributions which are diverse and informative, and 2) enhance the compositionality of the class embedding. Moreover, a visual-language primitive decomposition (VLPD) module and a stochastic logit mixup (SLM) strategy are proposed to dynamically fuse the decisions from the compositional and the primitive logit space. Orthogonal to the existing literature of soft, hard, or distributional prompts, our method advocates prompting the LLM-supported class distribution that leads to a better zero-shot generalization. Experimental results on MIT-States, UT-Zappos, and C-GQA datasets show the superior performance of the PLID to the prior arts.","{'model': 'tldr@v2.0.0', 'text': 'A model by prompting the language-informed distribution, aka.'}",https://arxiv.org/pdf/2305.14428
-Retrieval-augmented GPT-3.5-based Text-to-SQL Framework with Sample-aware Prompting and Dynamic Revision Chain,Chunxi Guo,"Text-to-SQL aims at generating SQL queries for the given natural language questions and thus helping users to query databases. Prompt learning with large language models (LLMs) has emerged as a recent approach, which designs prompts to lead LLMs to understand the input question and generate the corresponding SQL. However, it faces challenges with strict SQL syntax requirements. Existing work prompts the LLMs with a list of demonstration examples (i.e. question-SQL pairs) to generate SQL, but the fixed prompts can hardly handle the scenario where the semantic gap between the retrieved demonstration and the input question is large. In this paper, we propose a retrieval-augmented prompting method for a LLM-based Text-to-SQL framework, involving sample-aware prompting and a dynamic revision chain. Our approach incorporates sample-aware demonstrations, which include the composition of SQL operators and fine-grained information related to the given question. To retrieve questions sharing similar intents with input questions, we propose two strategies for assisting retrieval. Firstly, we leverage LLMs to simplify the original questions, unifying the syntax and thereby clarifying the users' intentions. To generate executable and accurate SQLs without human intervention, we design a dynamic revision chain which iteratively adapts fine-grained feedback from the previously generated SQL. Experimental results on three Text-to-SQL benchmarks demonstrate the superiority of our method over strong baseline models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a retrieval-augmented prompting method for a LLM-based Text-to-SQL framework, involving sample-aware prompting and a dynamic revision chain which iteratively adapts fine-grained feedback from the previously generated SQL.'}",https://arxiv.org/pdf/2307.05074
-Towards Better Chain-of-Thought Prompting Strategies: A Survey,Zihan Yu,"Chain-of-Thought (CoT), a step-wise and coherent reasoning chain, shows its impressive strength when used as a prompting strategy for large language models (LLM). Recent years, the prominent effect of CoT prompting has attracted emerging research. However, there still lacks of a systematic summary about key factors of CoT prompting and comprehensive guide for prompts utilizing. For a deeper understanding about CoT prompting, we survey on a wide range of current research, presenting a systematic and comprehensive analysis on several factors that may influence the effect of CoT prompting, and introduce how to better apply it in different applications under these discussions. We further analyze the challenges and propose some future directions about CoT prompting. This survey could provide an overall reference on related research.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and comprehensive analysis on several factors that may influence the effect of CoT prompting are presented, and how to better apply it in different applications under these discussions are introduced.'}",https://arxiv.org/pdf/2310.04959
-"Reinforcement Learning in the Era of LLMs: What is Essential? What is needed? An RL Perspective on RLHF, Prompting, and Beyond",Hao Sun,"Recent advancements in Large Language Models (LLMs) have garnered wide attention and led to successful products such as ChatGPT and GPT-4. Their proficiency in adhering to instructions and delivering harmless, helpful, and honest (3H) responses can largely be attributed to the technique of Reinforcement Learning from Human Feedback (RLHF). In this paper, we aim to link the research in conventional RL to RL techniques used in LLM research. Demystify this technique by discussing why, when, and how RL excels. Furthermore, we explore potential future avenues that could either benefit from or contribute to RLHF research. Highlighted Takeaways: 1. RLHF is Online Inverse RL with Offline Demonstration Data. 2. RLHF $>$ SFT because Imitation Learning (and Inverse RL) $>$ Behavior Cloning (BC) by alleviating the problem of compounding error. 3. The RM step in RLHF generates a proxy of the expensive human feedback, such an insight can be generalized to other LLM tasks such as prompting evaluation and optimization where feedback is also expensive. 4. The policy learning in RLHF is more challenging than conventional problems studied in IRL due to their high action dimensionality and feedback sparsity. 5. The main superiority of PPO over off-policy value-based methods is its stability gained from (almost) on-policy data and conservative policy updates.",,https://arxiv.org/pdf/2310.06147
-Can Instruction Fine-Tuned Language Models Identify Social Bias through Prompting?,O. Dige,"As the breadth and depth of language model applications continue to expand rapidly, it is increasingly important to build efficient frameworks for measuring and mitigating the learned or inherited social biases of these models. In this paper, we present our work on evaluating instruction fine-tuned language models' ability to identify bias through zero-shot prompting, including Chain-of-Thought (CoT) prompts. Across LLaMA and its two instruction fine-tuned versions, Alpaca 7B performs best on the bias identification task with an accuracy of 56.7%. We also demonstrate that scaling up LLM size and data diversity could lead to further performance gain. This is a work-in-progress presenting the first component of our bias mitigation framework. We will keep updating this work as we get more results.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating instruction fine-tuned language models' ability to identify bias through zero-shot prompting, including Chain-of-Thought (CoT) prompts shows Alpaca 7B performs best on the bias identification task and it is demonstrated that scaling up LLM size and data diversity could lead to further performance gain.""}",https://arxiv.org/pdf/2307.10472
-Approximating Online Human Evaluation of Social Chatbots with Prompting,Ekaterina Svikhnushina,"With conversational models becoming increasingly available to the general public, developing scalable and robust evaluation metrics is crucial to minimize potential social and psychological risks for the users. Existing evaluation metrics aim to automate offline user evaluation and approximate human judgment of pre-curated dialogs. However, they are limited in their ability to capture subjective perceptions of users who actually interact with the chatbots and might not generalize to real-world settings. To address this limitation, we propose an approach to approximate online human evaluation, leveraging large language models (LLMs) from the GPT-family. We introduce a new Dialog system Evaluation framework based on Prompting (DEP), which enables a fully automatic evaluation pipeline that replicates live user studies and achieves an impressive correlation with human judgment (up to Pearson r=0.95 on a system level). The DEP approach involves collecting synthetic chat logs of evaluated bots with an LLM in the other-play setting, where the LLM is carefully conditioned to follow a specific scenario. We further explore different prompting approaches to produce evaluation scores with the same LLM. The best-performing prompts, which contain few-shot demonstrations and instructions, show outstanding performance on the tested dataset and demonstrate the ability to generalize to other dialog corpora.","{'model': 'tldr@v2.0.0', 'text': 'A new Dialog system Evaluation framework based on Prompting is introduced, which enables a fully automatic evaluation pipeline that replicates live user studies and achieves an impressive correlation with human judgment (up to Pearson r=0.95 on a system level).'}",
-March in Chat: Interactive Prompting for Remote Embodied Referring Expression,Yanyuan Qiao,"Many Vision-and-Language Navigation (VLN) tasks have been proposed in recent years, from room-based to object-based and indoor to outdoor. The REVERIE (Remote Embodied Referring Expression) is interesting since it only provides high-level instructions to the agent, which are closer to human commands in practice. Nevertheless, this poses more challenges than other VLN tasks since it requires agents to infer a navigation plan only based on a short instruction. Large Language Models (LLMs) show great potential in robot action planning by providing proper prompts. Still, this strategy has not been explored under the REVERIE settings. There are several new challenges. For example, the LLM should be environment-aware so that the navigation plan can be adjusted based on the current visual observation. Moreover, the LLM planned actions should be adaptable to the much larger and more complex REVERIE environment. This paper proposes a March-in-Chat (MiC) model that can talk to the LLM on the fly and plan dynamically based on a newly proposed Room-and-Object Aware Scene Perceiver (ROASP). Our MiC model outperforms the previous state-of-the-art by large margins by SPL and RGSPL metrics on the REVERIE benchmark.","{'model': 'tldr@v2.0.0', 'text': 'A March-in-Chat (MiC) model that can talk to the LLM on the fly and plan dynamically based on a newly proposed Room-and-Object Aware Scene Perceiver (ROASP) and outperforms the previous state-of-the-art by large margins by SPL and RGSPL metrics on the REVERIE benchmark.'}",https://arxiv.org/pdf/2308.10141
-Prompting a Large Language Model to Generate Diverse Motivational Messages: A Comparison with Human-Written Messages,Samuel Rhys Cox,"Large language models (LLMs) are increasingly capable and prevalent, and can be used to produce creative content. The quality of content is influenced by the prompt used, with more specific prompts that incorporate examples generally producing better results. On from this, it could be seen that using instructions written for crowdsourcing tasks (that are specific and include examples to guide workers) could prove effective LLM prompts. To explore this, we used a previous crowdsourcing pipeline that gave examples to people to help them generate a collectively diverse corpus of motivational messages. We then used this same pipeline to generate messages using GPT-4, and compared the collective diversity of messages from: (1) crowd-writers, (2) GPT-4 using the pipeline, and (3&4) two baseline GPT-4 prompts. We found that the LLM prompts using the crowdsourcing pipeline caused GPT-4 to produce more diverse messages than the two baseline prompts. We also discuss implications from messages generated by both human writers and LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is found that using instructions written for crowdsourcing tasks (that are specific and include examples to guide workers) could prove effective LLM prompts that caused GPT-4 to produce more diverse messages than the two baseline prompts.'}",https://arxiv.org/pdf/2308.13479
-Prompting GPT-3.5 for Text-to-SQL with De-semanticization and Skeleton Retrieval,Chunxi Guo,"Text-to-SQL is a task that converts a natural language question into a structured query language (SQL) to retrieve information from a database. Large language models (LLMs) work well in natural language generation tasks, but they are not specifically pre-trained to understand the syntax and semantics of SQL commands. In this paper, we propose an LLM-based framework for Text-to-SQL which retrieves helpful demonstration examples to prompt LLMs. However, questions with different database schemes can vary widely, even if the intentions behind them are similar and the corresponding SQL queries exhibit similarities. Consequently, it becomes crucial to identify the appropriate SQL demonstrations that align with our requirements. We design a de-semanticization mechanism that extracts question skeletons, allowing us to retrieve similar examples based on their structural similarity. We also model the relationships between question tokens and database schema items (i.e., tables and columns) to filter out scheme-related information. Our framework adapts the range of the database schema in prompts to balance length and valuable information. A fallback mechanism allows for a more detailed schema to be provided if the generated SQL query fails. Ours outperforms state-of-the-art models and demonstrates strong generalization ability on three cross-domain Text-to-SQL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'An LLM-based framework for Text-to-SQL which retrieves helpful demonstration examples to prompt LLMs and outperforms state-of-the-art models and demonstrates strong generalization ability on three cross-domain Text- to-SQL benchmarks.'}",
-PAL: Program-aided Language Models,Luyu Gao,"Large language models (LLMs) have recently demonstrated an impressive ability to perform arithmetic and symbolic reasoning tasks, when provided with a few examples at test time (""few-shot prompting""). Much of this success can be attributed to prompting methods such as""chain-of-thought'', which employ LLMs for both understanding the problem description by decomposing it into steps, as well as solving each step of the problem. While LLMs seem to be adept at this sort of step-by-step decomposition, LLMs often make logical and arithmetic mistakes in the solution part, even when the problem is decomposed correctly. In this paper, we present Program-Aided Language models (PAL): a novel approach that uses the LLM to read natural language problems and generate programs as the intermediate reasoning steps, but offloads the solution step to a runtime such as a Python interpreter. With PAL, decomposing the natural language problem into runnable steps remains the only learning task for the LLM, while solving is delegated to the interpreter. We demonstrate this synergy between a neural LLM and a symbolic interpreter across 13 mathematical, symbolic, and algorithmic reasoning tasks from BIG-Bench Hard and other benchmarks. In all these natural language reasoning tasks, generating code using an LLM and reasoning using a Python interpreter leads to more accurate results than much larger models. For example, PAL using Codex achieves state-of-the-art few-shot accuracy on the GSM8K benchmark of math word problems, surpassing PaLM-540B which uses chain-of-thought by absolute 15% top-1. Our code and data are publicly available at http://reasonwithpal.com/ .","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Program-Aided Language models (PAL): a novel approach that uses the LLM to read natural language problems and generate programs as the intermediate reasoning steps, but offloads the solution step to a runtime such as a Python interpreter.'}",http://arxiv.org/pdf/2211.10435
-Large Language Models Can Self-Improve,Jiaxin Huang,"Large Language Models (LLMs) have achieved excellent performances in various tasks. However, fine-tuning an LLM requires extensive supervision. Human, on the other hand, may improve their reasoning abilities by self-thinking without external inputs. In this work, we demonstrate that an LLM is also capable of self-improving with only unlabeled datasets. We use a pre-trained LLM to generate""high-confidence""rationale-augmented answers for unlabeled questions using Chain-of-Thought prompting and self-consistency, and fine-tune the LLM using those self-generated solutions as target outputs. We show that our approach improves the general reasoning ability of a 540B-parameter LLM (74.4%->82.1% on GSM8K, 78.2%->83.0% on DROP, 90.0%->94.4% on OpenBookQA, and 63.4%->67.9% on ANLI-A3) and achieves state-of-the-art-level performance, without any ground truth label. We conduct ablation studies and show that fine-tuning on reasoning is critical for self-improvement.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a pre-trained LLM to generate ""high-confidence""rationale-augmented answers for unlabeled questions using Chain-of-Thought prompting and self-consistency, and fine-tune the LLM using those self-generated solutions as target outputs to improve the general reasoning ability.'}",http://arxiv.org/pdf/2210.11610
-Multimodal Chain-of-Thought Reasoning in Language Models,Zhuosheng Zhang,"Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies have focused on the language modality. We propose Multimodal-CoT that incorporates language (text) and vision (images) modalities into a two-stage framework that separates rationale generation and answer inference. In this way, answer inference can leverage better generated rationales that are based on multimodal information. With Multimodal-CoT, our model under 1 billion parameters outperforms the previous state-of-the-art LLM (GPT-3.5) by 16 percentage points (75.17%->91.68% accuracy) on the ScienceQA benchmark and even surpasses human performance. Code is publicly available available at https://github.com/amazon-science/mm-cot.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Multimodal-CoT that incorporates language (text) and vision (images) modalities into a two-stage framework that separates rationale generation and answer inference so that answer inference can leverage better generated rationales that are based on multimodal information.'}",http://arxiv.org/pdf/2302.00923
-Towards Expert-Level Medical Question Answering with Large Language Models,K. Singhal,"Recent artificial intelligence (AI) systems have reached milestones in""grand challenges""ranging from Go to protein-folding. The capability to retrieve medical knowledge, reason over it, and answer medical questions comparably to physicians has long been viewed as one such grand challenge. Large language models (LLMs) have catalyzed significant progress in medical question answering; Med-PaLM was the first model to exceed a""passing""score in US Medical Licensing Examination (USMLE) style questions with a score of 67.2% on the MedQA dataset. However, this and other prior work suggested significant room for improvement, especially when models' answers were compared to clinicians' answers. Here we present Med-PaLM 2, which bridges these gaps by leveraging a combination of base LLM improvements (PaLM 2), medical domain finetuning, and prompting strategies including a novel ensemble refinement approach. Med-PaLM 2 scored up to 86.5% on the MedQA dataset, improving upon Med-PaLM by over 19% and setting a new state-of-the-art. We also observed performance approaching or exceeding state-of-the-art across MedMCQA, PubMedQA, and MMLU clinical topics datasets. We performed detailed human evaluations on long-form questions along multiple axes relevant to clinical applications. In pairwise comparative ranking of 1066 consumer medical questions, physicians preferred Med-PaLM 2 answers to those produced by physicians on eight of nine axes pertaining to clinical utility (p<0.001). We also observed significant improvements compared to Med-PaLM on every evaluation axis (p<0.001) on newly introduced datasets of 240 long-form""adversarial""questions to probe LLM limitations. While further studies are necessary to validate the efficacy of these models in real-world settings, these results highlight rapid progress towards physician-level performance in medical question answering.","{'model': 'tldr@v2.0.0', 'text': 'Results highlight rapid progress towards physician-level performance in medical question answering by leveraging a combination of base LLM improvements (PaLM 2), medical domain finetuning, and prompting strategies including a novel ensemble refinement approach.'}",http://arxiv.org/pdf/2305.09617
-Language Models can Solve Computer Tasks,Geunwoo Kim,"Agents capable of carrying out general tasks on a computer can improve efficiency and productivity by automating repetitive tasks and assisting in complex problem-solving. Ideally, such agents should be able to solve new computer tasks presented to them through natural language commands. However, previous approaches to this problem require large amounts of expert demonstrations and task-specific reward functions, both of which are impractical for new tasks. In this work, we show that a pre-trained large language model (LLM) agent can execute computer tasks guided by natural language using a simple prompting scheme where the agent Recursively Criticizes and Improves its output (RCI). The RCI approach significantly outperforms existing LLM methods for automating computer tasks and surpasses supervised learning (SL) and reinforcement learning (RL) approaches on the MiniWoB++ benchmark. We compare multiple LLMs and find that RCI with the InstructGPT-3+RLHF LLM is state-of-the-art on MiniWoB++, using only a handful of demonstrations per task rather than tens of thousands, and without a task-specific reward function. Furthermore, we demonstrate RCI prompting's effectiveness in enhancing LLMs' reasoning abilities on a suite of natural language reasoning tasks, outperforming chain of thought (CoT) prompting. We find that RCI combined with CoT performs better than either separately. Our code can be found here: https://github.com/posgnu/rci-agent.","{'model': 'tldr@v2.0.0', 'text': ""This work shows that a pre-trained large language model (LLM) agent can execute computer tasks guided by natural language using a simple prompting scheme where the agent Recursively Criticizes and Improves its output (RCI), and demonstrates RCI prompting's effectiveness in enhancing LLMs' reasoning abilities on a suite of natural language reasoning tasks, outperforming chain of thought (CoT) prompting.""}",http://arxiv.org/pdf/2303.17491
-How is ChatGPT's behavior changing over time?,Lingjiao Chen,"GPT-3.5 and GPT-4 are the two most widely used large language model (LLM) services. However, when and how these models are updated over time is opaque. Here, we evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on several diverse tasks: 1) math problems, 2) sensitive/dangerous questions, 3) opinion surveys, 4) multi-hop knowledge-intensive questions, 5) generating code, 6) US Medical License tests, and 7) visual reasoning. We find that the performance and behavior of both GPT-3.5 and GPT-4 can vary greatly over time. For example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy). This is partly explained by a drop in GPT-4's amenity to follow chain-of-thought prompting. Interestingly, GPT-3.5 was much better in June than in March in this task. GPT-4 became less willing to answer sensitive questions and opinion survey questions in June than in March. GPT-4 performed better at multi-hop questions in June than in March, while GPT-3.5's performance dropped on this task. Both GPT-4 and GPT-3.5 had more formatting mistakes in code generation in June than in March. We provide evidence that GPT-4's ability to follow user instructions has decreased over time, which is one common factor behind the many behavior drifts. Overall, our findings show that the behavior of the""same""LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Overall, the findings show that the behavior of the""same""LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.'}",https://arxiv.org/pdf/2307.09009
-Is ChatGPT the Ultimate Programming Assistant - How far is it?,Haoye Tian,"Recently, the ChatGPT LLM has received great attention: it can be used as a bot for discussing source code, prompting it to suggest changes, provide descriptions or even generate code. Typical demonstrations generally focus on existing benchmarks, which may have been used in model training (i.e., data leakage). To assess the feasibility of using an LLM as a useful assistant bot for programmers, we must assess its realistic capabilities on unseen problems as well as its capabilities on various tasks. In this paper, we present an empirical study of ChatGPT's potential as a fully automated programming assistant, focusing on the tasks of code generation, program repair, and code summariziation. The study investigates ChatGPT's performance on common programming problems and compares it with state-of-the-art approaches on two benchmarks. Among several findings, our study shows that ChatGPT is effective in dealing with common programming problems. However, our experiments also reveal limitations in terms of its attention span: detailed descriptions will constrain the focus of ChatGPT and prevent it from leveraging its vast knowledge to solve the actual problem. Surprisingly, we have identified the ability of ChatGPT to reason the original intention of the code. We expect future work to build on this insight for dealing with the open question of the oracle problem. Our findings contribute interesting insights to the development of LLMs for programming assistance, notably by demonstrating the importance of prompt engineering, and providing a better understanding of ChatGPT's practical applications for software engineering.","{'model': 'tldr@v2.0.0', 'text': ""An empirical study of ChatGPT's potential as a fully automated programming assistant, focusing on the tasks of code generation, program repair, and code summariziation and identifies the ability ofChatGPT to reason the original intention of the code.""}",https://arxiv.org/pdf/2304.11938
-ART: Automatic multi-step reasoning and tool-use for large language models,Bhargavi Paranjape,"Large language models (LLMs) can perform complex reasoning in few- and zero-shot settings by generating intermediate chain of thought (CoT) reasoning steps. Further, each reasoning step can rely on external tools to support computation beyond the core LLM capabilities (e.g. search/running code). Prior work on CoT prompting and tool use typically requires hand-crafting task-specific demonstrations and carefully scripted interleaving of model generations with tool use. We introduce Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program. Given a new task to solve, ART selects demonstrations of multi-step reasoning and tool use from a task library. At test time, ART seamlessly pauses generation whenever external tools are called, and integrates their output before resuming generation. ART achieves a substantial improvement over few-shot prompting and automatic CoT on unseen tasks in the BigBench and MMLU benchmarks, and matches performance of hand-crafted CoT prompts on a majority of these tasks. ART is also extensible, and makes it easy for humans to improve performance by correcting errors in task-specific programs or incorporating new tools, which we demonstrate by drastically improving performance on select tasks with minimal human intervention.","{'model': 'tldr@v2.0.0', 'text': 'Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program, achieves a substantial improvement over few-shot prompting and automatic CoT on unseen tasks in the BigBench and MMLU benchmarks, and matches performance of hand-crafted CoT prompts on a majority of these tasks.'}",http://arxiv.org/pdf/2303.09014
-Graph of Thoughts: Solving Elaborate Problems with Large Language Models,Maciej Besta,"We introduce Graph of Thoughts (GoT): a framework that advances prompting capabilities in large language models (LLMs) beyond those offered by paradigms such as Chain-of-Thought or Tree of Thoughts (ToT). The key idea and primary advantage of GoT is the ability to model the information generated by an LLM as an arbitrary graph, where units of information (""LLM thoughts"") are vertices, and edges correspond to dependencies between these vertices. This approach enables combining arbitrary LLM thoughts into synergistic outcomes, distilling the essence of whole networks of thoughts, or enhancing thoughts using feedback loops. We illustrate that GoT offers advantages over state of the art on different tasks, for example increasing the quality of sorting by 62% over ToT, while simultaneously reducing costs by>31%. We ensure that GoT is extensible with new thought transformations and thus can be used to spearhead new prompting schemes. This work brings the LLM reasoning closer to human thinking or brain mechanisms such as recurrence, both of which form complex networks.","{'model': 'tldr@v2.0.0', 'text': 'Graph of Thoughts is introduced: a framework that advances prompting capabilities in large language models (LLMs) beyond those offered by paradigms such as Chain-of-Thought or Tree of Thoughts, and is ensured that GoT is extensible with new thought transformations and thus can be used to spearhead new prompting schemes.'}",https://arxiv.org/pdf/2308.09687
-Assessing the Utility of ChatGPT Throughout the Entire Clinical Workflow,Arya Rao,"IMPORTANCE: Large language model (LLM) artificial intelligence (AI) chatbots direct the power of large training datasets towards successive, related tasks, as opposed to single-ask tasks, for which AI already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as virtual physicians, has not yet been evaluated. OBJECTIVE: To evaluate ChatGPT's capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. DESIGN: We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. SETTING: ChatGPT, a publicly available LLM PARTICIPANTS: Clinical vignettes featured hypothetical patients with a variety of age and gender identities, and a range of Emergency Severity Indices (ESIs) based on initial clinical presentation. EXPOSURES: MSD Clinical Manual vignettes MAIN OUTCOMES AND MEASURES: We measured the proportion of correct responses to the questions posed within the clinical vignettes tested. RESULTS: ChatGPT achieved 71.7% (95% CI, 69.3% to 74.1%) accuracy overall across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI, 67.8% to 86.1%), and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI, 54.2% to 66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis ({beta}=-15.8%, p<0.001) and clinical management ({beta}=-7.4%, p=0.02) type questions. CONCLUSIONS AND RELEVANCE: ChatGPT achieves impressive accuracy in clinical decision making, with particular strengths emerging as it has more clinical information at its disposal.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating ChatGPT's capacity for ongoing clinical decision support via its performance on standardized clinical vignettes achieves impressive accuracy in clinical decision making, with particular strengths emerging as it has more clinical information at its disposal.""}",https://www.medrxiv.org/content/medrxiv/early/2023/02/26/2023.02.21.23285886.full.pdf
-Task and Motion Planning with Large Language Models for Object Rearrangement,Yan Ding,"Multi-object rearrangement is a crucial skill for service robots, and commonsense reasoning is frequently needed in this process. However, achieving commonsense arrangements requires knowledge about objects, which is hard to transfer to robots. Large language models (LLMs) are one potential source of this knowledge, but they do not naively capture information about plausible physical arrangements of the world. We propose LLM-GROP, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry. LLM-GROP allows us to go from natural-language commands to human-aligned object rearrangement in varied environments. Based on human evaluations, our approach achieves the highest rating while outperforming competitive baselines in terms of success rate while maintaining comparable cumulative action costs. Finally, we demonstrate a practical implementation of LLM-GROP on a mobile manipulator in real-world scenarios. Supplementary materials are available at: https://sites.google.com/view/llm-grop","{'model': 'tldr@v2.0.0', 'text': 'LLM-GROP is proposed, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry.'}",http://arxiv.org/pdf/2303.06247
-Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions,H. Trivedi,"Prompting-based large language models (LLMs) are surprisingly powerful at generating natural language reasoning steps or Chains-of-Thoughts (CoT) for multi-step question answering (QA). They struggle, however, when the necessary knowledge is either unavailable to the LLM or not up-to-date within its parameters. While using the question to retrieve relevant text from an external knowledge source helps LLMs, we observe that this one-step retrieve-and-read approach is insufficient for multi-step QA. Here, what to retrieve depends on what has already been derived, which in turn may depend on what was previously retrieved. To address this, we propose IRCoT, a new approach for multi-step QA that interleaves retrieval with steps (sentences) in a CoT, guiding the retrieval with CoT and in turn using retrieved results to improve CoT. Using IRCoT with GPT3 substantially improves retrieval (up to 21 points) as well as downstream QA (up to 15 points) on four datasets: HotpotQA, 2WikiMultihopQA, MuSiQue, and IIRC. We observe similar substantial gains in out-of-distribution (OOD) settings as well as with much smaller models such as Flan-T5-large without additional training. IRCoT reduces model hallucination, resulting in factually more accurate CoT reasoning.","{'model': 'tldr@v2.0.0', 'text': 'IRCoT is proposed, a new approach for multi-step QA that interleaves retrieval with steps in a CoT, guiding the retrieval with CoT and in turn using retrieved results to improve CoT.'}",http://arxiv.org/pdf/2212.10509
-Explanations from Large Language Models Make Small Reasoners Better,SHIYANG LI,"Integrating free-text explanations to in-context learning of large language models (LLM) is shown to elicit strong reasoning capabilities along with reasonable explanations. In this paper, we consider the problem of leveraging the explanations generated by LLM to improve the training of small reasoners, which are more favorable in real-production deployment due to their low cost. We systematically explore three explanation generation approaches from LLM and utilize a multi-task learning framework to facilitate small models to acquire strong reasoning power together with explanation generation capabilities. Experiments on multiple reasoning tasks show that our method can consistently and significantly outperform finetuning baselines across different settings, and even perform better than finetuning/prompting a 60x larger GPT-3 (175B) model by up to 9.5% in accuracy. As a side benefit, human evaluation further shows that our method can generate high-quality explanations to justify its predictions, moving towards the goal of explainable AI.","{'model': 'tldr@v2.0.0', 'text': 'This paper considers the problem of leveraging the explanations generated by LLM to improve the training of small reasoners, which are more favorable in real-production deployment due to their low cost.'}",http://arxiv.org/pdf/2210.06726
-GPT-NER: Named Entity Recognition via Large Language Models,Shuhe Wang,"Despite the fact that large-scale Language Models (LLM) have achieved SOTA performances on a variety of NLP tasks, its performance on NER is still significantly below supervised baselines. This is due to the gap between the two tasks the NER and LLMs: the former is a sequence labeling task in nature while the latter is a text-generation model. In this paper, we propose GPT-NER to resolve this issue. GPT-NER bridges the gap by transforming the sequence labeling task to a generation task that can be easily adapted by LLMs e.g., the task of finding location entities in the input text""Columbus is a city""is transformed to generate the text sequence""@@Columbus## is a city"", where special tokens @@## marks the entity to extract. To efficiently address the""hallucination""issue of LLMs, where LLMs have a strong inclination to over-confidently label NULL inputs as entities, we propose a self-verification strategy by prompting LLMs to ask itself whether the extracted entities belong to a labeled entity tag. We conduct experiments on five widely adopted NER datasets, and GPT-NER achieves comparable performances to fully supervised baselines, which is the first time as far as we are concerned. More importantly, we find that GPT-NER exhibits a greater ability in the low-resource and few-shot setups, when the amount of training data is extremely scarce, GPT-NER performs significantly better than supervised models. This demonstrates the capabilities of GPT-NER in real-world NER applications where the number of labeled examples is limited.","{'model': 'tldr@v2.0.0', 'text': 'GPT-NER exhibits a greater ability in the low-resource and few-shot setups, when the amount of training data is extremely scarce, and performs significantly better than supervised models, which demonstrates the capabilities of GPT-ner in real-world NER applications where the number of labeled examples is limited.'}",https://arxiv.org/pdf/2304.10428
-Unleashing Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaboration,Zhenhailong Wang,"Human intelligence thrives on the concept of cognitive synergy, where collaboration and information integration among different cognitive processes yield superior outcomes compared to individual cognitive processes in isolation. Although Large Language Models (LLMs) have demonstrated promising performance as general task-solving agents, they still struggle with tasks that require intensive domain knowledge and complex reasoning. In this work, we propose Solo Performance Prompting (SPP), which transforms a single LLM into a cognitive synergist by engaging in multi-turn self-collaboration with multiple personas. A cognitive synergist refers to an intelligent agent that collaborates with multiple minds, combining their individual strengths and knowledge, to enhance problem-solving and overall performance in complex tasks. By dynamically identifying and simulating different personas based on task inputs, SPP unleashes the potential of cognitive synergy in LLMs. We have discovered that assigning multiple, fine-grained personas in LLMs elicits better problem-solving abilities compared to using a single or fixed number of personas. We evaluate SPP on three challenging tasks: Trivia Creative Writing, Codenames Collaborative, and Logic Grid Puzzle, encompassing both knowledge-intensive and reasoning-intensive types. Unlike previous works, such as Chain-of-Thought, that solely enhance the reasoning abilities in LLMs, SPP effectively elicits internal knowledge acquisition abilities, reduces hallucination, and maintains strong reasoning capabilities. Code, data, and prompts can be found at: https://github.com/MikeWangWZHL/Solo-Performance-Prompting.git.","{'model': 'tldr@v2.0.0', 'text': 'Solo Performance Prompting (SPP), which transforms a single LLM into a cognitive synergist by engaging in multi-turn self-collaboration with multiple personas by dynamically identifying and simulating different personas based on task inputs, unleashes the potential of cognitive synergy in LLMs.'}",https://arxiv.org/pdf/2307.05300
-Safety Assessment of Chinese Large Language Models,Hao Sun,"With the rapid popularity of large language models such as ChatGPT and GPT-4, a growing amount of attention is paid to their safety concerns. These models may generate insulting and discriminatory content, reflect incorrect social values, and may be used for malicious purposes such as fraud and dissemination of misleading information. Evaluating and enhancing their safety is particularly essential for the wide application of large language models (LLMs). To further promote the safe deployment of LLMs, we develop a Chinese LLM safety assessment benchmark. Our benchmark explores the comprehensive safety performance of LLMs from two perspectives: 8 kinds of typical safety scenarios and 6 types of more challenging instruction attacks. Our benchmark is based on a straightforward process in which it provides the test prompts and evaluates the safety of the generated responses from the evaluated model. In evaluation, we utilize the LLM's strong evaluation ability and develop it as a safety evaluator by prompting. On top of this benchmark, we conduct safety assessments and analyze 15 LLMs including the OpenAI GPT series and other well-known Chinese LLMs, where we observe some interesting findings. For example, we find that instruction attacks are more likely to expose safety issues of all LLMs. Moreover, to promote the development and deployment of safe, responsible, and ethical AI, we publicly release SafetyPrompts including 100k augmented prompts and responses by LLMs.","{'model': 'tldr@v2.0.0', 'text': ""A Chinese LLM safety assessment benchmark is developed based on the LLM's strong evaluation ability and develops it as a safety evaluator by prompting, which finds that instruction attacks are more likely to expose safety issues of all LLMs.""}",http://arxiv.org/pdf/2304.10436
-Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs,Miao Xiong,"The task of empowering large language models (LLMs) to accurately express their confidence, referred to as confidence elicitation, is essential in ensuring reliable and trustworthy decision-making processes. Previous methods, which primarily rely on model logits, have become less suitable for LLMs and even infeasible with the rise of closed-source LLMs (e.g., commercialized LLM APIs). This leads to a growing need to explore the untapped area of \emph{non-logit-based} approaches to estimate the uncertainty of LLMs. Hence, in this study, we investigate approaches for confidence elicitation that do not require model fine-tuning or access to proprietary information. We introduce three categories of methods: verbalize-based, consistency-based, and their hybrid methods for benchmarking, and evaluate their performance across five types of datasets and four widely-used LLMs. Our analysis of these methods uncovers several key insights: 1) LLMs often exhibit a high degree of overconfidence when verbalizing their confidence; 2) Prompting strategies such as CoT, Top-K and Multi-step confidences improve calibration of verbalized confidence; 3) Consistency-based methods outperform the verbalized confidences in most cases, with particularly notable improvements on the arithmetic reasoning task; 4) Hybrid methods consistently deliver the best performance over their baselines, thereby emerging as a promising state-of-the-art approach; 5) Despite these advancements, all investigated methods continue to struggle with challenging tasks, such as those requiring professional knowledge, leaving significant scope for improvement of confidence elicitation.","{'model': 'tldr@v2.0.0', 'text': 'This study investigates approaches for confidence elicitation that do not require model fine-tuning or access to proprietary information, and introduces three categories of methods: verbalize- based, consistency-based, and their hybrid methods for benchmarking, and evaluates their performance across five types of datasets and four widely-used LLMs.'}",http://arxiv.org/pdf/2306.13063
-Adaptive Test Generation Using a Large Language Model,Max Schäfer,"—Unit tests play a key role in ensuring the correctness of software. However, manually creating unit tests is a laborious task, motivating the need for automation. This paper presents T EST P ILOT , an adaptive test generation technique that leverages Large Language Models (LLMs). T EST P ILOT uses Codex, an off-the-shelf LLM, to automatically generate unit tests for a given program without requiring additional training or few-shot learning on examples of existing tests. In our approach, Codex is provided with prompts that include the signature and implementation of a function under test, along with usage examples extracted from documentation. If a generated test fails, T EST P ILOT ’s adaptive component attempts to generate a new test that fixes the problem by re-prompting the model with the failing test and error message. We created an implementation of T EST P ILOT for JavaScript and evaluated it on 25 npm packages with a total of 1,684 API functions to generate tests for. Our results show that the generated tests achieve up to 93.1% statement coverage (median 68.2%). Moreover, on average, 58.5% of the generated tests contain at least one assertion that exercises functionality from the package under test. Our experiments with excluding parts of the information included in the prompts show that all components contribute towards the generation of effective test suites. Finally, we find that T EST P ILOT does not generate memorized tests: 92.7% of our generated tests have ≤ 50% similarity with existing tests (as measured by normalized edit distance), with none of them being exact copies.","{'model': 'tldr@v2.0.0', 'text': 'T EST P ILOT uses Codex, an off-the-shelf LLM, to automatically generate unit tests for a given program without requiring additional training or few-shot learning on examples of existing tests, and does not generate memorized tests.'}",https://arxiv.org/pdf/2302.06527
-When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgment,Zhijing Jin,"AI systems are becoming increasingly intertwined with human life. In order to effectively collaborate with humans and ensure safety, AI systems need to be able to understand, interpret and predict human moral judgments and decisions. Human moral judgments are often guided by rules, but not always. A central challenge for AI safety is capturing the flexibility of the human moral mind -- the ability to determine when a rule should be broken, especially in novel or unusual situations. In this paper, we present a novel challenge set consisting of rule-breaking question answering (RBQA) of cases that involve potentially permissible rule-breaking -- inspired by recent moral psychology studies. Using a state-of-the-art large language model (LLM) as a basis, we propose a novel moral chain of thought (MORALCOT) prompting strategy that combines the strengths of LLMs with theories of moral reasoning developed in cognitive science to predict human moral judgments. MORALCOT outperforms seven existing LLMs by 6.2% F1, suggesting that modeling human reasoning might be necessary to capture the flexibility of the human moral mind. We also conduct a detailed error analysis to suggest directions for future work to improve AI safety using RBQA. Our data is open-sourced at https://huggingface.co/datasets/feradauto/MoralExceptQA and code at https://github.com/feradauto/MoralCoT","{'model': 'tldr@v2.0.0', 'text': 'A novel moral chain of thought (MORALCOT) prompting strategy that combines the strengths of LLMs with theories of moral reasoning developed in cognitive science to predict human moral judgments, suggesting that modeling human reasoning might be necessary to capture the flexibility of the human moral mind.'}",https://arxiv.org/pdf/2210.01478
-ExpertPrompting: Instructing Large Language Models to be Distinguished Experts,Benfeng Xu,"The answering quality of an aligned large language model (LLM) can be drastically improved if treated with proper crafting of prompts. In this paper, we propose ExpertPrompting to elicit the potential of LLMs to answer as distinguished experts. We first utilize In-Context Learning to automatically synthesize detailed and customized descriptions of the expert identity for each specific instruction, and then ask LLMs to provide answer conditioned on such agent background. Based on this augmented prompting strategy, we produce a new set of instruction-following data using GPT-3.5, and train a competitive open-source chat assistant called ExpertLLaMA. We employ GPT4-based evaluation to show that 1) the expert data is of significantly higher quality than vanilla answers, and 2) ExpertLLaMA outperforms existing open-source opponents and achieves 96\% of the original ChatGPT's capability. All data and the ExpertLLaMA model will be made publicly available at \url{https://github.com/OFA-Sys/ExpertLLaMA}.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes ExpertPrompting to elicit the potential of LLMs to answer as distinguished experts, and produces a new set of instruction-following data using GPT-3.5, and trains a competitive open-source chat assistant called ExpertLLaMA.'}",http://arxiv.org/pdf/2305.14688
-An Empirical Evaluation of Using Large Language Models for Automated Unit Test Generation,Max Schafer,"Unit tests play a key role in ensuring the correctness of software. However, manually creating unit tests is a laborious task, motivating the need for automation. Large Language Models (LLMs) have recently been applied to this problem, utilizing additional training or few-shot learning on examples of existing tests. This paper presents a large-scale empirical evaluation on the effectiveness of LLMs for automated unit test generation without additional training or manual effort, providing the LLM with the signature and implementation of the function under test, along with usage examples extracted from documentation. We also attempt to repair failed generated tests by re-prompting the model with the failing test and error message. We implement our approach in TestPilot, a test generation tool for JavaScript that automatically generates unit tests for all API functions in an npm package. We evaluate TestPilot using OpenAI's gpt3.5-turbo LLM on 25 npm packages with a total of 1,684 API functions. The generated tests achieve a median statement coverage of 70.2% and branch coverage of 52.8%, significantly improving on Nessie, a recent feedback-directed JavaScript test generation technique, which achieves only 51.3% statement coverage and 25.6% branch coverage. We also find that 92.8% of TestPilot's generated tests have no more than 50% similarity with existing tests (as measured by normalized edit distance), with none of them being exact copies. Finally, we run TestPilot with two additional LLMs, OpenAI's older code-cushman-002 LLM and the open LLM StarCoder. Overall, we observed similar results with the former (68.2% median statement coverage), and somewhat worse results with the latter (54.0% median statement coverage), suggesting that the effectiveness of the approach is influenced by the size and training set of the LLM, but does not fundamentally depend on the specific model.","{'model': 'tldr@v2.0.0', 'text': 'A large-scale empirical evaluation on the effectiveness of large Language Models for automated unit test generation without additional training or manual effort, providing the LLM with the signature and implementation of the function under test, along with usage examples extracted from documentation.'}",
-Automatic Evaluation of Attribution by Large Language Models,Xiang Yue,"A recent focus of large language model (LLM) development, as exemplified by generative search engines, is to incorporate external references to generate and support its claims. However, evaluating the attribution, i.e., verifying whether the generated statement is fully supported by the cited reference, remains an open problem. Although human evaluation is common practice, it is costly and time-consuming. In this paper, we investigate the automatic evaluation of attribution given by LLMs. We begin by defining different types of attribution errors, and then explore two approaches for automatic evaluation: prompting LLMs and fine-tuning smaller LMs. The fine-tuning data is repurposed from related tasks such as question answering, fact-checking, natural language inference, and summarization. We manually curate a set of test examples covering 12 domains from a generative search engine, New Bing. Our results on this curated test set and simulated examples from existing benchmarks highlight both promising signals and challenges. We hope our problem formulation, testbeds, and findings will help lay the foundation for future studies on this important problem.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the automatic evaluation of attribution given by large language model (LLMs), defining different types of attribution errors, and exploring two approaches for automatic evaluation: prompting LLMs and fine-tuning smaller LMs.'}",https://arxiv.org/pdf/2305.06311
-CAPE: Corrective Actions from Precondition Errors using Large Language Models,S. S. Raman,"Extracting commonsense knowledge from a large language model (LLM) offers a path to designing intelligent robots. Existing approaches that leverage LLMs for planning are unable to recover when an action fails and often resort to retrying failed actions, without resolving the error's underlying cause. We propose a novel approach (CAPE) that attempts to propose corrective actions to resolve precondition errors during planning. CAPE improves the quality of generated plans by leveraging few-shot reasoning from action preconditions. Our approach enables embodied agents to execute more tasks than baseline methods while ensuring semantic correctness and minimizing re-prompting. In VirtualHome, CAPE generates executable plans while improving a human-annotated plan correctness metric from 28.89% to 49.63% over SayCan. Our improvements transfer to a Boston Dynamics Spot robot initialized with a set of skills (specified in language) and associated preconditions, where CAPE improves the correctness metric of the executed task plans by 76.49% compared to SayCan. Our approach enables the robot to follow natural language commands and robustly recover from failures, which baseline approaches largely cannot resolve or address inefficiently.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach (CAPE) that attempts to propose corrective actions to resolve precondition errors during planning and improves the quality of generated plans by leveraging few-shot reasoning from action preconditions.'}",
-On the Risk of Misinformation Pollution with Large Language Models,Yikang Pan,"In this paper, we comprehensively investigate the potential misuse of modern Large Language Models (LLMs) for generating credible-sounding misinformation and its subsequent impact on information-intensive applications, particularly Open-Domain Question Answering (ODQA) systems. We establish a threat model and simulate potential misuse scenarios, both unintentional and intentional, to assess the extent to which LLMs can be utilized to produce misinformation. Our study reveals that LLMs can act as effective misinformation generators, leading to a significant degradation in the performance of ODQA systems. To mitigate the harm caused by LLM-generated misinformation, we explore three defense strategies: prompting, misinformation detection, and majority voting. While initial results show promising trends for these defensive strategies, much more work needs to be done to address the challenge of misinformation pollution. Our work highlights the need for further research and interdisciplinary collaboration to address LLM-generated misinformation and to promote responsible use of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that LLMs can act as effective misinformation generators, leading to a significant degradation in the performance of ODQA systems, and the need for further research and interdisciplinary collaboration to address LLM-generated misinformation.'}",http://arxiv.org/pdf/2305.13661
-Building Cooperative Embodied Agents Modularly with Large Language Models,Hongxin Zhang,"Large Language Models (LLMs) have demonstrated impressive planning abilities in single-agent embodied tasks across various domains. However, their capacity for planning and communication in multi-agent cooperation remains unclear, even though these are crucial skills for intelligent embodied agents. In this paper, we present a novel framework that utilizes LLMs for multi-agent cooperation and tests it in various embodied environments. Our framework enables embodied agents to plan, communicate, and cooperate with other embodied agents or humans to accomplish long-horizon tasks efficiently. We demonstrate that recent LLMs, such as GPT-4, can surpass strong planning-based methods and exhibit emergent effective communication using our framework without requiring fine-tuning or few-shot prompting. We also discover that LLM-based agents that communicate in natural language can earn more trust and cooperate more effectively with humans. Our research underscores the potential of LLMs for embodied AI and lays the foundation for future research in multi-agent cooperation. Videos can be found on the project website https://vis-www.cs.umass.edu/Co-LLM-Agents/.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel framework that utilizes LLMs for multi-agent cooperation and tests it in various embodied environments, and discovers that LLM-based agents that communicate in natural language can earn more trust and cooperate more effectively with humans.'}",https://arxiv.org/pdf/2307.02485
-Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning,Liangming Pan,"Large Language Models (LLMs) have shown human-like reasoning abilities but still struggle with complex logical problems. This paper introduces a novel framework, Logic-LM, which integrates LLMs with symbolic solvers to improve logical problem-solving. Our method first utilizes LLMs to translate a natural language problem into a symbolic formulation. Afterward, a deterministic symbolic solver performs inference on the formulated problem. We also introduce a self-refinement module, which utilizes the symbolic solver's error messages to revise symbolic formalizations. We demonstrate Logic-LM's effectiveness on five logical reasoning datasets: ProofWriter, PrOntoQA, FOLIO, LogicalDeduction, and AR-LSAT. On average, Logic-LM achieves a significant performance boost of 39.2% over using LLM alone with standard prompting and 18.4% over LLM with chain-of-thought prompting. Our findings suggest that Logic-LM, by combining LLMs with symbolic logic, offers a promising avenue for faithful logical reasoning. Code and data are publicly available at https://github.com/teacherpeterpan/Logic-LLM.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel framework, Logic-LM, which integrates LLMs with symbolic solvers to improve logical problem-solving and suggests that it offers a promising avenue for faithful logical reasoning.'}",http://arxiv.org/pdf/2305.12295
-Text Classification via Large Language Models,Xiaofei Sun,"Despite the remarkable success of large-scale Language Models (LLMs) such as GPT-3, their performances still significantly underperform fine-tuned models in the task of text classification. This is due to (1) the lack of reasoning ability in addressing complex linguistic phenomena (e.g., intensification, contrast, irony etc); (2) limited number of tokens allowed in in-context learning. In this paper, we introduce Clue And Reasoning Prompting (CARP). CARP adopts a progressive reasoning strategy tailored to addressing the complex linguistic phenomena involved in text classification: CARP first prompts LLMs to find superficial clues (e.g., keywords, tones, semantic relations, references, etc), based on which a diagnostic reasoning process is induced for final decisions. To further address the limited-token issue, CARP uses a fine-tuned model on the supervised dataset for $k$NN demonstration search in the in-context learning, allowing the model to take the advantage of both LLM's generalization ability and the task-specific evidence provided by the full labeled dataset. Remarkably, CARP yields new SOTA performances on 4 out of 5 widely-used text-classification benchmarks, 97.39 (+1.24) on SST-2, 96.40 (+0.72) on AGNews, 98.78 (+0.25) on R8 and 96.95 (+0.6) on R52, and a performance comparable to SOTA on MR (92.39 v.s. 93.3). More importantly, we find that CARP delivers impressive abilities on low-resource and domain-adaptation setups. Specifically, using 16 examples per class, CARP achieves comparable performances to supervised models with 1,024 examples per class.","{'model': 'tldr@v2.0.0', 'text': 'Clue And Reasoning Prompting (CARP) adopts a progressive reasoning strategy tailored to addressing the complex linguistic phenomena involved in text classification, and achieves comparable performances to supervised models with 1,024 examples per class.'}",https://arxiv.org/pdf/2305.08377
-Dr.ICL: Demonstration-Retrieved In-context Learning,Man Luo,"In-context learning (ICL), teaching a large language model (LLM) to perform a task with few-shot demonstrations rather than adjusting the model parameters, has emerged as a strong paradigm for using LLMs. While early studies primarily used a fixed or random set of demonstrations for all test queries, recent research suggests that retrieving semantically similar demonstrations to the input from a pool of available demonstrations results in better performance. This work expands the applicability of retrieval-based ICL approaches by demonstrating that even simple word-overlap similarity measures such as BM25 outperform randomly selected demonstrations. Furthermore, we extend the success of retrieval-based ICL to instruction-finetuned LLMs as well as Chain-of-Thought (CoT) prompting. For instruction-finetuned LLMs, we find that although a model has already seen the training data at training time, retrieving demonstrations from the training data at test time yields better results compared to using no demonstrations or random demonstrations. Last but not least, we train a task-specific demonstration retriever that outperforms off-the-shelf retrievers.","{'model': 'tldr@v2.0.0', 'text': 'This work expands the applicability of retrieval-based ICL approaches by demonstrating that even simple word-overlap similarity measures such as BM25 outperform randomly selected demonstrations and trains a task-specific demonstration retriever that outperforms off-the-shelf retrievers.'}",http://arxiv.org/pdf/2305.14128
-Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment,Rishabh Bhardwaj,"Larger language models (LLMs) have taken the world by storm with their massive multi-tasking capabilities simply by optimizing over a next-word prediction objective. With the emergence of their properties and encoded knowledge, the risk of LLMs producing harmful outputs increases, making them unfit for scalable deployment for the public. In this work, we propose a new safety evaluation benchmark RED-EVAL that carries out red-teaming. We show that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting, jailbreaking closed source LLM-based systems such as GPT-4 and ChatGPT to unethically respond to more than 65% and 73% of harmful queries. We also demonstrate the consistency of the RED-EVAL across 8 open-source LLMs in generating harmful responses in more than 86% of the red-teaming attempts. Next, we propose RED-INSTRUCT--An approach for the safety alignment of LLMs. It constitutes two phases: 1) HARMFULQA data collection: Leveraging CoU prompting, we collect a dataset that consists of 1.9K harmful questions covering a wide range of topics, 9.5K safe and 7.3K harmful conversations from ChatGPT; 2) SAFE-ALIGN: We demonstrate how the conversational dataset can be used for the safety alignment of LLMs by minimizing the negative log-likelihood over helpful responses and penalizing over harmful responses by gradient accent over sample loss. Our model STARLING, a fine-tuned Vicuna-7B, is observed to be more safely aligned when evaluated on RED-EVAL and HHH benchmarks while preserving the utility of the baseline models (TruthfulQA, MMLU, and BBH).","{'model': 'tldr@v2.0.0', 'text': 'It is shown that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting, jailbreaking closed source LLM-based systems such as GPT-4 and ChatGPT to unethically respond to more than 65% and 73% of harmful queries.'}",https://arxiv.org/pdf/2308.09662
-Bayesian Optimization of Catalysts With In-context Learning,Mayk C. Ramos,"Large language models (LLMs) are able to do accurate classification with zero or only a few examples (in-context learning). We show a prompting system that enables regression with uncertainty for in-context learning with frozen LLM (GPT-3, GPT-3.5, and GPT-4) models, allowing predictions without features or architecture tuning. By incorporating uncertainty, our approach enables Bayesian optimization for catalyst or molecule optimization using natural language, eliminating the need for training or simulation. Here, we performed the optimization using the synthesis procedure of catalysts to predict properties. Working with natural language mitigates difficulty synthesizability since the literal synthesis procedure is the model's input. We showed that in-context learning could improve past a model context window (maximum number of tokens the model can process at once) as data is gathered via example selection, allowing the model to scale better. Although our method does not outperform all baselines, it requires zero training, feature selection, and minimal computing while maintaining satisfactory performance. We also find Gaussian Process Regression on text embeddings is strong at Bayesian optimization. The code is available in our GitHub repository: https://github.com/ur-whitelab/BO-LIFT","{'model': 'tldr@v2.0.0', 'text': 'It is shown that in-context learning could improve past a model context window as data is gathered via example selection, allowing the model to scale better, and Gaussian Process Regression on text embeddings is strong at Bayesian optimization.'}",http://arxiv.org/pdf/2304.05341
-Can Large Language Models Write Good Property-Based Tests?,Vasudev Vikram,"Property-based testing (PBT), while an established technique in the software testing research community, is still relatively underused in real-world software. Pain points in writing property-based tests include implementing diverse random input generators and thinking of meaningful properties to test. Developers, however, are more amenable to writing documentation; plenty of library API documentation is available and can be used as natural language specifications for property-based tests. As large language models (LLMs) have recently shown promise in a variety of coding tasks, we explore the potential of using LLMs to synthesize property-based tests. We call our approach PBT-GPT, and propose three different strategies of prompting the LLM for PBT. We characterize various failure modes of PBT-GPT and detail an evaluation methodology for automatically synthesized property-based tests. PBT-GPT achieves promising results in our preliminary studies on sample Python library APIs in $\texttt{numpy}$, $\texttt{networkx}$, and $\texttt{datetime}$.","{'model': 'tldr@v2.0.0', 'text': 'This work describes various failure modes of PBT-GPT, and proposes three different strategies of prompting the LLM for PBT, and details an evaluation methodology for automatically synthesized property-based tests.'}",https://arxiv.org/pdf/2307.04346
-ChatGPT makes medicine easy to swallow: an exploratory case study on simplified radiology reports.,Katharina Jeblick,,"{'model': 'tldr@v2.0.0', 'text': 'The initial insights of this study indicate a tremendous potential in using LLMs like ChatGPT to improve patient-centered care in radiology and other medical domains.'}",https://link.springer.com/content/pdf/10.1007/s00330-023-10213-1.pdf
-CodeHelp: Using Large Language Models with Guardrails for Scalable Support in Programming Classes,Mark H. Liffiton,"Computing educators face significant challenges in providing timely support to students, especially in large class settings. Large language models (LLMs) have emerged recently and show great promise for providing on-demand help at a large scale, but there are concerns that students may over-rely on the outputs produced by these models. In this paper, we introduce CodeHelp, a novel LLM-powered tool designed with guardrails to provide on-demand assistance to programming students without directly revealing solutions. We detail the design of the tool, which incorporates a number of useful features for instructors, and elaborate on the pipeline of prompting strategies we use to ensure generated outputs are suitable for students. To evaluate CodeHelp, we deployed it in a first-year computer and data science course with 52 students and collected student interactions over a 12-week period. We examine students' usage patterns and perceptions of the tool, and we report reflections from the course instructor and a series of recommendations for classroom use. Our findings suggest that CodeHelp is well-received by students who especially value its availability and help with resolving errors, and that for instructors it is easy to deploy and complements, rather than replaces, the support that they provide to students.","{'model': 'tldr@v2.0.0', 'text': 'The findings suggest that CodeHelp is well-received by students who especially value its availability and help with resolving errors, and that for instructors it is easy to deploy and complements, rather than replaces, the support that they provide to students.'}",https://arxiv.org/pdf/2308.06921
-I Spy a Metaphor: Large Language Models and Diffusion Models Co-Create Visual Metaphors,Tuhin Chakrabarty,"Visual metaphors are powerful rhetorical devices used to persuade or communicate creative ideas through images. Similar to linguistic metaphors, they convey meaning implicitly through symbolism and juxtaposition of the symbols. We propose a new task of generating visual metaphors from linguistic metaphors. This is a challenging task for diffusion-based text-to-image models, such as DALL$\cdot$E 2, since it requires the ability to model implicit meaning and compositionality. We propose to solve the task through the collaboration between Large Language Models (LLMs) and Diffusion Models: Instruct GPT-3 (davinci-002) with Chain-of-Thought prompting generates text that represents a visual elaboration of the linguistic metaphor containing the implicit meaning and relevant objects, which is then used as input to the diffusion-based text-to-image models.Using a human-AI collaboration framework, where humans interact both with the LLM and the top-performing diffusion model, we create a high-quality dataset containing 6,476 visual metaphors for 1,540 linguistic metaphors and their associated visual elaborations. Evaluation by professional illustrators shows the promise of LLM-Diffusion Model collaboration for this task . To evaluate the utility of our Human-AI collaboration framework and the quality of our dataset, we perform both an intrinsic human-based evaluation and an extrinsic evaluation using visual entailment as a downstream task.","{'model': 'tldr@v2.0.0', 'text': 'Using a human-AI collaboration framework, where humans interact both with the LLM and the top-performing diffusion model, a high-quality dataset is created containing 6,476 visual metaphors for 1,540 linguistic metaphors and their associated visual elaborations.'}",https://arxiv.org/pdf/2305.14724
-ReviewerGPT? An Exploratory Study on Using Large Language Models for Paper Reviewing,Ryan Liu,"Given the rapid ascent of large language models (LLMs), we study the question: (How) can large language models help in reviewing of scientific papers or proposals? We first conduct some pilot studies where we find that (i) GPT-4 outperforms other LLMs (Bard, Vicuna, Koala, Alpaca, LLaMa, Dolly, OpenAssistant, StableLM), and (ii) prompting with a specific question (e.g., to identify errors) outperforms prompting to simply write a review. With these insights, we study the use of LLMs (specifically, GPT-4) for three tasks: 1. Identifying errors: We construct 13 short computer science papers each with a deliberately inserted error, and ask the LLM to check for the correctness of these papers. We observe that the LLM finds errors in 7 of them, spanning both mathematical and conceptual errors. 2. Verifying checklists: We task the LLM to verify 16 closed-ended checklist questions in the respective sections of 15 NeurIPS 2022 papers. We find that across 119 {checklist question, paper} pairs, the LLM had an 86.6% accuracy. 3. Choosing the""better""paper: We generate 10 pairs of abstracts, deliberately designing each pair in such a way that one abstract was clearly superior than the other. The LLM, however, struggled to discern these relatively straightforward distinctions accurately, committing errors in its evaluations for 6 out of the 10 pairs. Based on these experiments, we think that LLMs have a promising use as reviewing assistants for specific reviewing tasks, but not (yet) for complete evaluations of papers or proposals.","{'model': 'tldr@v2.0.0', 'text': 'It is thought that LLMs have a promising use as reviewing assistants for specific reviewing tasks, but not for complete evaluations of papers or proposals.'}",http://arxiv.org/pdf/2306.00622
-Graphologue: Exploring Large Language Model Responses with Interactive Diagrams,Peiling Jiang,"Large language models (LLMs) have recently soared in popularity due to their ease of access and the unprecedented ability to synthesize text responses to diverse user questions. However, LLMs like ChatGPT present significant limitations in supporting complex information tasks due to the insufficient affordances of the text-based medium and linear conversational structure. Through a formative study with ten participants, we found that LLM interfaces often present long-winded responses, making it difficult for people to quickly comprehend and interact flexibly with various pieces of information, particularly during more complex tasks. We present Graphologue, an interactive system that converts text-based responses from LLMs into graphical diagrams to facilitate information-seeking and question-answering tasks. Graphologue employs novel prompting strategies and interface designs to extract entities and relationships from LLM responses and constructs node-link diagrams in real-time. Further, users can interact with the diagrams to flexibly adjust the graphical presentation and to submit context-specific prompts to obtain more information. Utilizing diagrams, Graphologue enables graphical, non-linear dialogues between humans and LLMs, facilitating information exploration, organization, and comprehension.","{'model': 'tldr@v2.0.0', 'text': 'Graphologue is an interactive system that converts text-based responses from LLMs into graphical diagrams to facilitate information-seeking and question-answering tasks, and enables graphical, non-linear dialogues between humans and LLMs, facilitating information exploration, organization, and comprehension.'}",https://arxiv.org/pdf/2305.11473
-ALGO: Synthesizing Algorithmic Programs with Generated Oracle Verifiers,Kexun Zhang,"Large language models (LLMs) excel at implementing code from functionality descriptions but struggle with algorithmic problems that require not only implementation but also identification of the suitable algorithm. Moreover, LLM-generated programs lack guaranteed correctness and require human verification. To address these challenges, we propose ALGO, a framework that synthesizes Algorithmic programs with LLM-Generated Oracles to guide the generation and verify their correctness. ALGO first generates a reference oracle by prompting an LLM to exhaustively enumerate all the combinations of relevant variables. This oracle is then utilized to guide an arbitrary search strategy in exploring the algorithm space and to verify the synthesized algorithms. Our study shows that the LLM-generated oracles are correct for 88% of the cases. With the oracles as verifiers, ALGO can be integrated with any existing code generation model in a model-agnostic manner to enhance its performance. Experiments show that when equipped with ALGO, we achieve an 8x better one-submission pass rate over the Codex model and a 2.6x better one-submission pass rate over CodeT, the current state-of-the-art model on CodeContests. We can also get 1.3x better pass rate over the ChatGPT Code Interpreter on unseen problems. The problem set we used for testing, the prompts we used, the verifier and solution programs, and the test cases generated by ALGO are available at https://github.com/zkx06111/ALGO.","{'model': 'tldr@v2.0.0', 'text': 'ALGO is a framework that synthesizes Algorithmic programs with LLM-Generated Oracles to guide the generation and verify their correctness and can be integrated with any existing code generation model in a model-agnostic manner to enhance its performance.'}",http://arxiv.org/pdf/2305.14591
-Large Language Models as Tax Attorneys: A Case Study in Legal Capabilities Emergence,John J. Nay,"Better understanding of Large Language Models' (LLMs) legal analysis abilities can contribute to improving the efficiency of legal services, governing artificial intelligence, and leveraging LLMs to identify inconsistencies in law. This paper explores LLM capabilities in applying tax law. We choose this area of law because it has a structure that allows us to set up automated validation pipelines across thousands of examples, requires logical reasoning and maths skills, and enables us to test LLM capabilities in a manner relevant to real-world economic lives of citizens and companies. Our experiments demonstrate emerging legal understanding capabilities, with improved performance in each subsequent OpenAI model release. We experiment with retrieving and utilising the relevant legal authority to assess the impact of providing additional legal context to LLMs. Few-shot prompting, presenting examples of question-answer pairs, is also found to significantly enhance the performance of the most advanced model, GPT-4. The findings indicate that LLMs, particularly when combined with prompting enhancements and the correct legal texts, can perform at high levels of accuracy but not yet at expert tax lawyer levels. As LLMs continue to advance, their ability to reason about law autonomously could have significant implications for the legal profession and AI governance.",,http://arxiv.org/pdf/2306.07075
-Assessing the Utility of ChatGPT Throughout the Entire Clinical Workflow: Development and Usability Study,Arya Rao,"Background Large language model (LLM)–based artificial intelligence chatbots direct the power of large training data sets toward successive, related tasks as opposed to single-ask tasks, for which artificial intelligence already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as artificial physicians, has not yet been evaluated. Objective This study aimed to evaluate ChatGPT’s capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. Methods We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared its accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. Accuracy was measured by the proportion of correct responses to the questions posed within the clinical vignettes tested, as calculated by human scorers. We further conducted linear regression to assess the contributing factors toward ChatGPT’s performance on clinical tasks. Results ChatGPT achieved an overall accuracy of 71.7% (95% CI 69.3%-74.1%) across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI 67.8%-86.1%) and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI 54.2%-66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis (β=–15.8%; P<.001) and clinical management (β=–7.4%; P=.02) question types. Conclusions ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal. In particular, ChatGPT demonstrates the greatest accuracy in tasks of final diagnosis as compared to initial diagnosis. Limitations include possible model hallucinations and the unclear composition of ChatGPT’s training data set.","{'model': 'tldr@v2.0.0', 'text': 'ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal, in particular in tasks of final diagnosis as compared to initial diagnosis.'}",https://www.jmir.org/2023/1/e48659/PDF
-"Despite ""super-human"" performance, current LLMs are unsuited for decisions about ethics and safety",Joshua Albrecht,"Large language models (LLMs) have exploded in popularity in the past few years and have achieved undeniably impressive results on benchmarks as varied as question answering and text summarization. We provide a simple new prompting strategy that leads to yet another supposedly""super-human""result, this time outperforming humans at common sense ethical reasoning (as measured by accuracy on a subset of the ETHICS dataset). Unfortunately, we find that relying on average performance to judge capabilities can be highly misleading. LLM errors differ systematically from human errors in ways that make it easy to craft adversarial examples, or even perturb existing examples to flip the output label. We also observe signs of inverse scaling with model size on some examples, and show that prompting models to""explain their reasoning""often leads to alarming justifications of unethical actions. Our results highlight how human-like performance does not necessarily imply human-like understanding or reasoning.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a simple new prompting strategy that leads to yet another supposedly ""super-human"" result, this time outperforming humans at common sense ethical reasoning (as measured by accuracy on a subset of the ETHICS dataset).'}",http://arxiv.org/pdf/2212.06295
-Human-in-the-Loop through Chain-of-Thought,Zefan Cai,"While the emergence of powerful language models along with Chain-of-thought prompting has made automation more and more omnipresent, it sometimes demonstrates its weakness in long-term or multi-step logical reasoning. For example, users don't always get desirable answers for complex mathematical problems without human involvement. Against this background, we present the Manual Correction System (MCS) -- a human-in-the-loop system enhanced by Chain-of-Thought prompting, which explores how manual correction of sub-logics in rationales can improve LLM's reasoning performance. Moving one step forward, considering a system with human-in-the-loop involves more than having humans improve performance but also controlling the cost. Therefore, we post a Cost-utility Analysis Model for Human-in-the-Loop systems (CAMLOP) based on classical economics theory to analyze, quantify and balance the utility and the corresponding cost. We conduct experiments of MCS and CAMLOP with twelve datasets. A significant advantage w.r.t cost and utility proves its superiority over strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A Cost-utility Analysis Model for Human-in-the-Loop systems (CAMLOP) based on classical economics theory to analyze, quantify and balance the utility and the corresponding cost.'}",http://arxiv.org/pdf/2306.07932
-An Evaluation of Log Parsing with ChatGPT,Van-Hoang Le,"—Software logs play an essential role in ensuring the reliability and maintainability of large-scale software systems, as they are often the sole source of runtime information. Log parsing, which converts raw log messages into structured data, is an important initial step towards downstream log analytics. In recent studies, ChatGPT, the current cutting-edge large language model (LLM), has been widely applied to a wide range of software engineering tasks. However, its performance in automated log parsing remains unclear. In this paper, we evaluate ChatGPT’s ability to undertake log parsing by addressing two research questions. (1) Can ChatGPT effectively parse logs? (2) How does ChatGPT perform with different prompting methods? Our results show that ChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting. Based on our findings, we outline several challenges and opportunities for ChatGPT-based log parsing.","{'model': 'tldr@v2.0.0', 'text': 'Results show that ChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting, and outline several challenges and opportunities for ChatG PT-based log parsing.'}",https://arxiv.org/pdf/2306.01590
-Evaluating GPT-3 Generated Explanations for Hateful Content Moderation,H. Wang,"Recent research has focused on using large language models (LLMs) to generate explanations for hate speech through fine-tuning or prompting. Despite the growing interest in this area, these generated explanations' effectiveness and potential limitations remain poorly understood. A key concern is that these explanations, generated by LLMs, may lead to erroneous judgments about the nature of flagged content by both users and content moderators. For instance, an LLM-generated explanation might inaccurately convince a content moderator that a benign piece of content is hateful. In light of this, we propose an analytical framework for examining hate speech explanations and conducted an extensive survey on evaluating such explanations. Specifically, we prompted GPT-3 to generate explanations for both hateful and non-hateful content, and a survey was conducted with 2,400 unique respondents to evaluate the generated explanations. Our findings reveal that (1) human evaluators rated the GPT-generated explanations as high quality in terms of linguistic fluency, informativeness, persuasiveness, and logical soundness, (2) the persuasive nature of these explanations, however, varied depending on the prompting strategy employed, and (3) this persuasiveness may result in incorrect judgments about the hatefulness of the content. Our study underscores the need for caution in applying LLM-generated explanations for content moderation. Code and results are available at https://github.com/Social-AI-Studio/GPT3-HateEval.",,https://arxiv.org/pdf/2305.17680
-Large Language Models are Strong Zero-Shot Retriever,Tao Shen,"In this work, we propose a simple method that applies a large language model (LLM) to large-scale retrieval in zero-shot scenarios. Our method, the Language language model as Retriever (LameR), is built upon no other neural models but an LLM, while breaking brute-force combinations of retrievers with LLMs and lifting the performance of zero-shot retrieval to be very competitive on benchmark datasets. Essentially, we propose to augment a query with its potential answers by prompting LLMs with a composition of the query and the query's in-domain candidates. The candidates, regardless of correct or wrong, are obtained by a vanilla retrieval procedure on the target collection. As a part of the prompts, they are likely to help LLM generate more precise answers by pattern imitation or candidate summarization. Even if all the candidates are wrong, the prompts at least make LLM aware of in-collection patterns and genres. Moreover, due to the low performance of a self-supervised retriever, the LLM-based query augmentation becomes less effective as the retriever bottlenecks the whole pipeline. Therefore, we propose to leverage a non-parametric lexicon-based method (e.g., BM25) as the retrieval module to capture query-document overlap in a literal fashion. As such, LameR makes the retrieval procedure transparent to the LLM, thus circumventing the performance bottleneck.","{'model': 'tldr@v2.0.0', 'text': 'The Language language model as Retriever (LameR), is built upon no other neural models but an LLM, while breaking brute-force combinations of retrievers with LLMs and lifting the performance of zero-shot retrieval to be very competitive on benchmark datasets.'}",https://arxiv.org/pdf/2304.14233
-Careful Data Curation Stabilizes In-context Learning,Ting-Yun Chang,"In-context learning (ICL) enables large language models (LLMs) to perform new tasks by prompting them with a sequence of training examples. However, ICL is very sensitive to the choice of training examples: randomly sampling examples from a training set leads to high variance in performance. In this paper, we show that curating a carefully chosen subset of training data greatly stabilizes ICL performance. We propose two methods to choose training subsets, both of which score training examples individually and then select the highest-scoring ones. C OND A CC scores a training example by its average ICL accuracy when combined with random training examples, while D ATAMODELS learns a linear proxy model that estimates how the presence of each training example influences LLM accuracy. On average, C OND A CC and D ATA - MODELS outperform sampling from the entire training set by 7.7% and 6.3%, respectively, across 5 tasks and two LLMs. Our analysis shows that stable subset examples are no more diverse than average, and are not outliers in terms of sequence length and perplexity.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that curating a carefully chosen subset of training data greatly stabilizes ICL performance, and that stable subset examples are no more diverse than average, and are not outliers in terms of sequence length and perplexity.'}",https://arxiv.org/pdf/2212.10378
-S3HQA: A Three-Stage Approach for Multi-hop Text-Table Hybrid Question Answering,Fangyu Lei,"Answering multi-hop questions over hybrid factual knowledge from the given text and table (TextTableQA) is a challenging task. Existing models mainly adopt a retriever-reader framework, which have several deficiencies, such as noisy labeling in training retriever, insufficient utilization of heterogeneous information over text and table, and deficient ability for different reasoning operations. In this paper, we propose a three-stage TextTableQA framework S3HQA, which comprises of retriever, selector, and reasoner. We use a retriever with refinement training to solve the noisy labeling problem. Then, a hybrid selector considers the linked relationships between heterogeneous data to select the most relevant factual knowledge. For the final stage, instead of adapting a reading comprehension module like in previous methods, we employ a generation-based reasoner to obtain answers. This includes two approaches: a row-wise generator and an LLM prompting generator (first time used in this task). The experimental results demonstrate that our method achieves competitive results in the few-shot setting. When trained on the full dataset, our approach outperforms all baseline methods, ranking first on the HybridQA leaderboard.","{'model': 'tldr@v2.0.0', 'text': 'A retriever with refinement training is used to solve the noisy labeling problem, a hybrid selector considers the linked relationships between heterogeneous data to select the most relevant factual knowledge, and a generation-based reasoner is employed to obtain answers.'}",http://arxiv.org/pdf/2305.11725
-Forward-Backward Reasoning in Large Language Models for Verification,Weisen Jiang,"Chain-of-Though (CoT) prompting has shown promising performance in various reasoning tasks. Recently, Self-Consistency (Wang et al., 2023) proposes to sample a diverse set of reasoning chains which may lead to different answers while the answer that receives the most votes is selected. In this paper, we propose a novel method to use backward reasoning in verifying candidate answers. We mask a token in the question by x and ask the LLM to predict the masked token when a candidate answer is provided by a simple template, i.e., “If we know the answer of the above question is {a candidate answer}, what is the value of unknown variable x?” Intuitively, the LLM is expected to predict the masked token successfully if the provided candidate answer is correct. We further propose FOBAR to combine forward and backward reasoning for estimating the probability of candidate answers. We conduct extensive experiments on six data sets and three LLMs. Experimental results demonstrate that FOBAR achieves state-of-the-art performance on various reasoning benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A novel method to use backward reasoning in verifying candidate answers inChain-of-Though prompting and proposes FOBAR to combine forward and backward reasoning for estimating the probability of candidate answers.'}",https://arxiv.org/pdf/2308.07758
-How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questions,Lorenzo Pacchiardi,"Large language models (LLMs) can""lie"", which we define as outputting false statements despite""knowing""the truth in a demonstrable sense. LLMs might""lie"", for example, when instructed to output misinformation. Here, we develop a simple lie detector that requires neither access to the LLM's activations (black-box) nor ground-truth knowledge of the fact in question. The detector works by asking a predefined set of unrelated follow-up questions after a suspected lie, and feeding the LLM's yes/no answers into a logistic regression classifier. Despite its simplicity, this lie detector is highly accurate and surprisingly general. When trained on examples from a single setting -- prompting GPT-3.5 to lie about factual questions -- the detector generalises out-of-distribution to (1) other LLM architectures, (2) LLMs fine-tuned to lie, (3) sycophantic lies, and (4) lies emerging in real-life scenarios such as sales. These results indicate that LLMs have distinctive lie-related behavioural patterns, consistent across architectures and contexts, which could enable general-purpose lie detection.",,https://arxiv.org/pdf/2309.15840
-Self-Checker: Plug-and-Play Modules for Fact-Checking with Large Language Models,Miaoran Li,"Fact-checking is an essential task in NLP that is commonly utilized for validating the factual accuracy of claims. Prior work has mainly focused on fine-tuning pre-trained languages models on specific datasets, which can be computationally intensive and time-consuming. With the rapid development of large language models (LLMs), such as ChatGPT and GPT-3, researchers are now exploring their in-context learning capabilities for a wide range of tasks. In this paper, we aim to assess the capacity of LLMs for fact-checking by introducing Self-Checker, a framework comprising a set of plug-and-play modules that facilitate fact-checking by purely prompting LLMs in an almost zero-shot setting. This framework provides a fast and efficient way to construct fact-checking systems in low-resource environments. Empirical results demonstrate the potential of Self-Checker in utilizing LLMs for fact-checking. However, there is still significant room for improvement compared to SOTA fine-tuned models, which suggests that LLM adoption could be a promising approach for future fact-checking research.","{'model': 'tldr@v2.0.0', 'text': 'Self-Checker, a framework comprising a set of plug-and-play modules that facilitate fact-checking by purely prompting LLMs in an almost zero-shot setting, is introduced, which provides a fast and efficient way to construct fact- checking systems in low-resource environments.'}",http://arxiv.org/pdf/2305.14623
-LLMs to the Moon? Reddit Market Sentiment Analysis with Large Language Models,Xiang Deng,"Market sentiment analysis on social media content requires knowledge of both financial markets and social media jargon, which makes it a challenging task for human raters. The resulting lack of high-quality labeled data stands in the way of conventional supervised learning methods. In this work, we conduct a case study approaching this problem with semi-supervised learning using a large language model (LLM). We select Reddit as the target social media platform due to its broad coverage of topics and content types. Our pipeline first generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production. We find that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels, while training the student model using a regression loss further improves distillation quality. With only a handful of prompts, the final model performs on par with existing supervised models. Though production applications of our model are limited by ethical considerations, the model’s competitive performance points to the great potential of using LLMs for tasks that otherwise require skill-intensive annotation.","{'model': 'tldr@v2.0.0', 'text': 'This case study conducts a case study approaching market sentiment analysis on social media content with semi-supervised learning using a large language model (LLM), finding that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels.'}",
-Leveraging Commonsense Knowledge from Large Language Models for Task and Motion Planning,Yan Ding,"Multi-object rearrangement is a crucial skill for service robots, and commonsense reasoning is frequently needed in this process. However, achieving commonsense arrangements requires knowledge about objects, which is hard to transfer to robots. Large language models (LLMs) are one potential source of this knowledge, but they do not naively capture information about plausible physical arrangements of the world. We propose LLMGROP, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry. LLM-GROP allows us to go from natural-language commands to human-aligned object rearrangement in varied environments. Based on human evaluations, our approach achieves the highest rating while outperforming competitive baselines in terms of success rate while maintaining comparable cumulative action costs. Finally, we demonstrate a practical implementation of LLM-GROP on a mobile manipulator in real-world scenarios. Supplementary materials are available at: https://sites.google.com/view/llm-grop","{'model': 'tldr@v2.0.0', 'text': 'LLMGROP is proposed, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry.'}",
-EvEntS ReaLM: Event Reasoning of Entity States via Language Models,Evangelia Spiliopoulou,"This paper investigates models of event implications. Specifically, how well models predict entity state-changes, by targeting their understanding of physical attributes. Nominally, Large Language models (LLM) have been exposed to procedural knowledge about how objects interact, yet our benchmarking shows they fail to reason about the world. Conversely, we also demonstrate that existing approaches often misrepresent the surprising abilities of LLMs via improper task encodings and that proper model prompting can dramatically improve performance of reported baseline results across multiple tasks. In particular, our results indicate that our prompting technique is especially useful for unseen attributes (out-of-domain) or when only limited data is available.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the prompting technique is especially useful for unseen attributes (out-of-domain) or when only limited data is available and that proper model prompting can dramatically improve performance of reported baseline results across multiple tasks.'}",https://arxiv.org/pdf/2211.05392
-What do LLMs Know about Financial Markets? A Case Study on Reddit Market Sentiment Analysis,Xiang Deng,"Market sentiment analysis on social media content requires knowledge of both financial markets and social media jargon, which makes it a challenging task for human raters. The resulting lack of high-quality labeled data stands in the way of conventional supervised learning methods. Instead, we approach this problem using semi-supervised learning with a large language model (LLM). Our pipeline generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production. We find that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels, while using a regression loss further improves distillation quality. With only a handful of prompts, the final model performs on par with existing supervised models. Though production applications of our model are limited by ethical considerations, the model’s competitive performance points to the great potential of using LLMs for tasks that otherwise require skill-intensive annotation.","{'model': 'tldr@v2.0.0', 'text': 'This model generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production, and finds that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels.'}",http://arxiv.org/pdf/2212.11311
-Enhancing In-Context Learning with Answer Feedback for Multi-Span Question Answering,Zixian Huang,"Whereas the recent emergence of large language models (LLMs) like ChatGPT has exhibited impressive general performance, it still has a large gap with fully-supervised models on specific tasks such as multi-span question answering. Previous researches found that in-context learning is an effective approach to exploiting LLM, by using a few task-related labeled data as demonstration examples to construct a few-shot prompt for answering new questions. A popular implementation is to concatenate a few questions and their correct answers through simple templates, informing LLM of the desired output. In this paper, we propose a novel way of employing labeled data such that it also informs LLM of some undesired output, by extending demonstration examples with feedback about answers predicted by an off-the-shelf model, e.g., correct, incorrect, or incomplete. Experiments on three multi-span question answering datasets as well as a keyphrase extraction dataset show that our new prompting strategy consistently improves LLM's in-context learning performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel way of employing labeled data such that it also informs LLM of some undesired output, by extending demonstration examples with feedback about answers predicted by an off-the-shelf model.'}",http://arxiv.org/pdf/2306.04508
-Retrieving Texts based on Abstract Descriptions,Shauli Ravfogel,"While instruction-tuned Large Language Models (LLMs) excel at extracting information from text, they are not suitable for locating texts conforming to a given description in a large document collection (semantic retrieval). Similarity search over embedding vectors does allow to perform retrieval by query, but the similarity reflected in the embedding is ill-defined and non-consistent, and is sub-optimal for many use cases. What, then, is a good query representation for effective retrieval? We identify the well defined and consistent task of retrieving sentences based on abstract descriptions of their content. We demonstrate the inadequacy of current text embeddings and propose an alternative model that significantly improves when used in standard nearest neighbor search. The model is trained using positive and negative pairs sourced through prompting a LLM. While it is easy to source the training material from an LLM, the retrieval task cannot be performed by the LLM directly. This demonstrates that data from LLMs can be used not only for distilling more efficient specialized models than the original LLM, but also for creating new capabilities not immediately possible using the original model.","{'model': 'tldr@v2.0.0', 'text': 'This work demonstrates that data from LLMs can be used not only for distilling more efficient specialized models than the original LLM, but also for creating new capabilities not immediately possible using the original model.'}",http://arxiv.org/pdf/2305.12517
-SkillGPT: a RESTful API service for skill extraction and standardization using a Large Language Model,Nan Li,"We present SkillGPT, a tool for skill extraction and standardization (SES) from free-style job descriptions and user profiles with an open-source Large Language Model (LLM) as backbone. Most previous methods for similar tasks either need supervision or rely on heavy data-preprocessing and feature engineering. Directly prompting the latest conversational LLM for standard skills, however, is slow, costly and inaccurate. In contrast, SkillGPT utilizes a LLM to perform its tasks in steps via summarization and vector similarity search, to balance speed with precision. The backbone LLM of SkillGPT is based on Llama, free for academic use and thus useful for exploratory research and prototype development. Hence, our cost-free SkillGPT gives users the convenience of conversational SES, efficiently and reliably.","{'model': 'tldr@v2.0.0', 'text': 'SkillGPT, a tool for skill extraction and standardization from free-style job descriptions and user profiles with an open-source Large Language Model (LLM) as backbone, gives users the convenience of conversational SES, efficiently and reliably.'}",http://arxiv.org/pdf/2304.11060
-Queer People are People First: Deconstructing Sexual Identity Stereotypes in Large Language Models,Harnoor Dhingra,"Large Language Models (LLMs) are trained primarily on minimally processed web text, which exhibits the same wide range of social biases held by the humans who created that content. Consequently, text generated by LLMs can inadvertently perpetuate stereotypes towards marginalized groups, like the LGBTQIA+ community. In this paper, we perform a comparative study of how LLMs generate text describing people with different sexual identities. Analyzing bias in the text generated by an LLM using regard score shows measurable bias against queer people. We then show that a post-hoc method based on chain-of-thought prompting using SHAP analysis can increase the regard of the sentence, representing a promising approach towards debiasing the output of LLMs in this setting.",,http://arxiv.org/pdf/2307.00101
-Knowing What LLMs DO NOT Know: A Simple Yet Effective Self-Detection Method,Yukun Zhao,"Large Language Models (LLMs) have shown great potential in Natural Language Processing (NLP) tasks. However, recent literature reveals that LLMs generate nonfactual responses intermittently, which impedes the LLMs' reliability for further utilization. In this paper, we propose a novel self-detection method to detect which questions that a LLM does not know that are prone to generate nonfactual results. Specifically, we first diversify the textual expressions for a given question and collect the corresponding answers. Then we examine the divergencies between the generated answers to identify the questions that the model may generate falsehoods. All of the above steps can be accomplished by prompting the LLMs themselves without referring to any other external resources. We conduct comprehensive experiments and demonstrate the effectiveness of our method on recently released LLMs, e.g., Vicuna, ChatGPT, and GPT-4.","{'model': 'tldr@v2.0.0', 'text': 'A novel self-detection method to detect which questions that a LLM does not know that are prone to generate nonfactual results, which can be accomplished by prompting the LLMs themselves without referring to any other external resources.'}",
-Log Parsing: How Far Can ChatGPT Go?,Van-Hoang Le,"Software logs play an essential role in ensuring the reliability and maintainability of large-scale software systems, as they are often the sole source of runtime information. Log parsing, which converts raw log messages into structured data, is an important initial step towards downstream log analytics. In recent studies, ChatGPT, the current cutting-edge large language model (LLM), has been widely applied to a wide range of software engineering tasks. However, its performance in automated log parsing remains unclear. In this paper, we evaluate ChatGPT's ability to undertake log parsing by addressing two research questions. (1) Can ChatGPT effectively parse logs? (2) How does ChatGPT perform with different prompting methods? Our results show that ChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting. Based on our findings, we outline several challenges and opportunities for ChatGPT-based log parsing.","{'model': 'tldr@v2.0.0', 'text': 'The results show thatChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting, and outline several challenges and opportunities for ChatGPT-based log parsing.'}",https://arxiv.org/pdf/2306.01590
-Retrieving Supporting Evidence for LLMs Generated Answers,Siqing Huo,"Current large language models (LLMs) can exhibit near-human levels of performance on many natural language tasks, including open-domain question answering. Unfortunately, they also convincingly hallucinate incorrect answers, so that responses to questions must be verified against external sources before they can be accepted at face value. In this paper, we report a simple experiment to automatically verify generated answers against a corpus. After presenting a question to an LLM and receiving a generated answer, we query the corpus with the combination of the question + generated answer. We then present the LLM with the combination of the question + generated answer + retrieved answer, prompting it to indicate if the generated answer can be supported by the retrieved answer. We base our experiment on questions and passages from the MS MARCO (V1) test collection, exploring three retrieval approaches ranging from standard BM25 to a full question answering stack, including a reader based on the LLM. For a large fraction of questions, we find that an LLM is capable of verifying its generated answer if appropriate supporting material is provided. However, with an accuracy of 70-80%, this approach cannot be fully relied upon to detect hallucinations.","{'model': 'tldr@v2.0.0', 'text': 'For a large fraction of questions, it is found that an LLM is capable of verifying its generated answer if appropriate supporting material is provided, but with an accuracy of 70-80%, this approach cannot be fully relied upon to detect hallucinations.'}",http://arxiv.org/pdf/2306.13781
-Knowledge Sanitization of Large Language Models,Yoichi Ishibashi,"We explore a knowledge sanitization approach to mitigate the privacy concerns associated with large language models (LLMs). LLMs trained on a large corpus of Web data can memorize and potentially reveal sensitive or confidential information, raising critical security concerns. Our technique fine-tunes these models, prompting them to generate harmless responses such as ``I don't know'' when queried about specific information. Experimental results in a closed-book question-answering task show that our straightforward method not only minimizes particular knowledge leakage but also preserves the overall performance of LLM. These two advantages strengthen the defense against extraction attacks and reduces the emission of harmful content such as hallucinations.","{'model': 'tldr@v2.0.0', 'text': ""A knowledge sanitization approach to mitigate the privacy concerns associated with large language models (LLMs), which fine-tunes these models, prompting them to generate harmless responses such as ``I don't know'' when queried about specific information.""}",https://arxiv.org/pdf/2309.11852
-Reasoning in Large Language Models Through Symbolic Math Word Problems,Vedant Gaur,"Large language models (LLMs) have revolutionized NLP by solving downstream tasks with little to no labeled data. Despite their versatile abilities, the larger question of their ability to reason remains ill-understood. This paper addresses reasoning in math word problems (MWPs) by studying symbolic versions of the numeric problems, since a symbolic expression is a""concise explanation""of the numeric answer. We create and use a symbolic version of the SVAMP dataset and find that GPT-3's davinci-002 model also has good zero-shot accuracy on symbolic MWPs. To evaluate the faithfulness of the model's reasoning, we go beyond accuracy and additionally evaluate the alignment between the final answer and the outputted reasoning, which correspond to numeric and symbolic answers respectively for MWPs. We explore a self-prompting approach to encourage the symbolic reasoning to align with the numeric answer, thus equipping the LLM with the ability to provide a concise and verifiable reasoning and making it more interpretable. Surprisingly, self-prompting also improves the symbolic accuracy to be higher than both the numeric and symbolic accuracies, thus providing an ensembling effect. The SVAMP_Sym dataset will be released for future research on symbolic math problems.","{'model': 'tldr@v2.0.0', 'text': 'A self-prompting approach is explored to encourage the symbolic reasoning to align with the numeric answer, thus equipping the LLM with the ability to provide a concise and verifiable reasoning and making it more interpretable.'}",https://aclanthology.org/2023.findings-acl.364.pdf
-Alphazero-like Tree-Search can Guide Large Language Model Decoding and Training,Xidong Feng,"Large language models (LLMs) typically employ sampling or beam search, accompanied by prompts such as Chain-of-Thought (CoT), to boost reasoning and decoding ability. Recent work like Tree-of-Thought (ToT) and Reasoning via Planning (RAP) aim to augment the reasoning capabilities of LLMs by utilizing tree-search algorithms to guide multi-step reasoning. These methods mainly focus on LLMs' reasoning ability during inference and heavily rely on human-designed prompts to activate LLM as a value function, which lacks general applicability and scalability. To address these limitations, we present an AlphaZero-like tree-search framework for LLMs (termed TS-LLM), systematically illustrating how tree-search with a learned value function can guide LLMs' decoding ability. TS-LLM distinguishes itself in two key ways: (1) Leveraging a learned value function, our approach can be generally applied to different tasks beyond reasoning (such as RLHF alignment), and LLMs of any size, without prompting advanced, large-scale models. (2) It can guide LLM's decoding during both inference and training. Empirical evaluations across reasoning, planning, and RLHF alignment tasks validate the effectiveness of TS-LLM, even on trees with a depth of 64.","{'model': 'tldr@v2.0.0', 'text': ""An AlphaZero-like tree-search framework for LLMs (termed TS-LLM) is presented, systematically illustrating how tree- search with a learned value function can guide LLM' decoding ability.""}",https://arxiv.org/pdf/2309.17179
-Exploring Human-Like Translation Strategy with Large Language Models,Zhiwei He,"Large language models (LLMs) have demonstrated impressive capabilities in general scenarios, exhibiting a level of aptitude that approaches, in some aspects even surpasses, human-level intelligence. Among their numerous skills, the translation abilities of LLMs have received considerable attention. In contrast to traditional machine translation that focuses solely on source-target mapping, LLM-based translation can potentially mimic the human translation process that takes many preparatory steps to ensure high-quality translation. This work aims to explore this possibility by proposing the MAPS framework, which stands for Multi-Aspect Prompting and Selection. Specifically, we enable LLMs to first analyze the given source text and extract three aspects of translation-related knowledge: keywords, topics and relevant demonstrations to guide the translation process. To filter out the noisy and unhelpful knowledge, we employ a selection mechanism based on quality estimation. Experiments suggest that MAPS brings significant and consistent improvements over text-davinci-003 and Alpaca on eight translation directions from the latest WMT22 test sets. Our further analysis shows that the extracted knowledge is critical in resolving up to 59% of hallucination mistakes in translation. Code is available at https://github.com/zwhe99/MAPS-mt.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the MAPS framework, which stands for Multi-Aspect Prompting and Selection, and enables LLMs to first analyze the given source text and extract three aspects of translation-related knowledge: keywords, topics and relevant demonstrations to guide the translation process.'}",http://arxiv.org/pdf/2305.04118
-"MMHQA-ICL: Multimodal In-context Learning for Hybrid Question Answering over Text, Tables and Images",Weihao Liu,"In the real world, knowledge often exists in a multimodal and heterogeneous form. Addressing the task of question answering with hybrid data types, including text, tables, and images, is a challenging task (MMHQA). Recently, with the rise of large language models (LLM), in-context learning (ICL) has become the most popular way to solve QA problems. We propose MMHQA-ICL framework for addressing this problems, which includes stronger heterogeneous data retriever and an image caption module. Most importantly, we propose a Type-specific In-context Learning Strategy for MMHQA, enabling LLMs to leverage their powerful performance in this task. We are the first to use end-to-end LLM prompting method for this task. Experimental results demonstrate that our framework outperforms all baselines and methods trained on the full dataset, achieving state-of-the-art results under the few-shot setting on the MultimodalQA dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Type-specific In-context Learning Strategy for MMHQA, enabling LLMs to leverage their powerful performance in this task, and is the first to use end-to-end LLM prompting method for this task.'}",https://arxiv.org/pdf/2309.04790
-GEAR: Augmenting Language Models with Generalizable and Efficient Tool Resolution,Yining Lu,"Augmenting large language models (LLM) to use external tools enhances their performance across a variety of tasks. However, prior works over-rely on task-specific demonstration of tool use that limits their generalizability and computational cost due to making many calls to large-scale LLMs. We introduce GEAR, a computationally efficient query-tool grounding algorithm that is generalizable to various tasks that require tool use while not relying on task-specific demonstrations. GEAR achieves better efficiency by delegating tool grounding and execution to small language models (SLM) and LLM, respectively; while leveraging semantic and pattern-based evaluation at both question and answer levels for generalizable tool grounding. We evaluate GEAR on 14 datasets across 6 downstream tasks, demonstrating its strong generalizability to novel tasks, tools and different SLMs. Despite offering more efficiency, GEAR achieves higher precision in tool grounding compared to prior strategies using LLM prompting, thus improving downstream accuracy at a reduced computational cost. For example, we demonstrate that GEAR-augmented GPT-J and GPT-3 outperform counterpart tool-augmented baselines because of better tool use.","{'model': 'tldr@v2.0.0', 'text': 'GE is introduced, a computationally efficient query-tool grounding algorithm that is generalizable to various tasks that require tool use while not relying on task-specific demonstrations and achieves higher precision in tool grounding compared to prior strategies using LLM prompting, thus improving downstream accuracy at a reduced computational cost.'}",https://arxiv.org/pdf/2307.08775
-ConstitutionMaker: Interactively Critiquing Large Language Models by Converting Feedback into Principles,S. Petridis,"Large language model (LLM) prompting is a promising new approach for users to create and customize their own chatbots. However, current methods for steering a chatbot's outputs, such as prompt engineering and fine-tuning, do not support users in converting their natural feedback on the model's outputs to changes in the prompt or model. In this work, we explore how to enable users to interactively refine model outputs through their feedback, by helping them convert their feedback into a set of principles (i.e. a constitution) that dictate the model's behavior. From a formative study, we (1) found that users needed support converting their feedback into principles for the chatbot and (2) classified the different principle types desired by users. Inspired by these findings, we developed ConstitutionMaker, an interactive tool for converting user feedback into principles, to steer LLM-based chatbots. With ConstitutionMaker, users can provide either positive or negative feedback in natural language, select auto-generated feedback, or rewrite the chatbot's response; each mode of feedback automatically generates a principle that is inserted into the chatbot's prompt. In a user study with 14 participants, we compare ConstitutionMaker to an ablated version, where users write their own principles. With ConstitutionMaker, participants felt that their principles could better guide the chatbot, that they could more easily convert their feedback into principles, and that they could write principles more efficiently, with less mental demand. ConstitutionMaker helped users identify ways to improve the chatbot, formulate their intuitive responses to the model into feedback, and convert this feedback into specific and clear principles. Together, these findings inform future tools that support the interactive critiquing of LLM outputs.","{'model': 'tldr@v2.0.0', 'text': 'ConstitutionMaker, an interactive tool for converting user feedback into principles, is developed to steer LLM-based chatbots and helps users identify ways to improve the chatbot, formulate their intuitive responses to the model into feedback, and convert this feedback into specific and clear principles.'}",
-Generative AI: Implications and Applications for Education,A. Tzirides,"The launch of ChatGPT in November 2022 precipitated a panic among some educators while prompting qualified enthusiasm from others. Under the umbrella term Generative AI, ChatGPT is an example of a range of technologies for the delivery of computer-generated text, image, and other digitized media. This paper examines the implications for education of one generative AI technology, chatbots responding from large language models, or C-LLM. It reports on an application of a C-LLM to AI review and assessment of complex student work. In a concluding discussion, the paper explores the intrinsic limits of generative AI, bound as it is to language corpora and their textual representation through binary notation. Within these limits, we suggest the range of emerging and potential applications of Generative AI in education.","{'model': 'tldr@v2.0.0', 'text': 'The intrinsic limits of generative AI are explored, bound as it is to language corpora and their textual representation through binary notation, and the range of emerging and potential applications of Generative AI in education are suggested.'}",http://arxiv.org/pdf/2305.07605
-Theory of Mind in Large Language Models: Examining Performance of 11 State-of-the-Art models vs. Children Aged 7-10 on Advanced Tests,Max J. van Duijn,"To what degree should we ascribe cognitive capacities to Large Language Models (LLMs), such as the ability to reason about intentions and beliefs known as Theory of Mind (ToM)? Here we add to this emerging debate by (i) testing 11 base- and instruction-tuned LLMs on capabilities relevant to ToM beyond the dominant false-belief paradigm, including non-literal language usage and recursive intentionality; (ii) using newly rewritten versions of standardized tests to gauge LLMs' robustness; (iii) prompting and scoring for open besides closed questions; and (iv) benchmarking LLM performance against that of children aged 7-10 on the same tasks. We find that instruction-tuned LLMs from the GPT family outperform other models, and often also children. Base-LLMs are mostly unable to solve ToM tasks, even with specialized prompting. We suggest that the interlinked evolution and development of language and ToM may help explain what instruction-tuning adds: rewarding cooperative communication that takes into account interlocutor and context. We conclude by arguing for a nuanced perspective on ToM in LLMs.",,
-Generating Multiple Choice Questions from a Textbook: LLMs Match Human Performance on Most Metrics,Andrew M. Olney,"Multiple choice questions are traditionally expensive to produce. Recent advances in large language models (LLMs) have led to fine-tuned LLMs that generate questions competitive with human-authored questions. However, the relative capabilities of ChatGPT-family models have not yet been established for this task. We present a carefully-controlled human evaluation of three conditions: a fine-tuned, augmented version of Macaw, instruction-tuned Bing Chat with zero-shot prompting, and human-authored questions from a college science textbook. Our results indicate that on six of seven measures tested, both LLM’s performance was not significantly different from human performance. Analysis of LLM errors further suggests that Macaw and Bing Chat have different failure modes for this task: Macaw tends to repeat answer options whereas Bing Chat tends to not include the specified answer in the answer options. For Macaw, removing error items from analysis results in performance on par with humans for all metrics; for Bing Chat, removing error items improves performance but does not reach human-level performance.","{'model': 'tldr@v2.0.0', 'text': 'Analysis of LLM errors suggests that Macaw and Bing Chat have different failure modes for this task: Macaw tends to repeat answer options whereas Bing Chat tends to not include the specified answer in the answer options.'}",
-Retrieving Supporting Evidence for Generative Question Answering,Siqing Huo,"Current large language models (LLMs) can exhibit near-human levels of performance on many natural language-based tasks, including open-domain question answering. Unfortunately, at this time, they also convincingly hallucinate incorrect answers, so that responses to questions must be verified against external sources before they can be accepted at face value. In this paper, we report two simple experiments to automatically validate generated answers against a corpus. We base our experiments on questions and passages from the MS MARCO (V1) test collection, and a retrieval pipeline consisting of sparse retrieval, dense retrieval and neural rerankers. In the first experiment, we validate the generated answer in its entirety. After presenting a question to an LLM and receiving a generated answer, we query the corpus with the combination of the question + generated answer. We then present the LLM with the combination of the question + generated answer + retrieved answer, prompting it to indicate if the generated answer can be supported by the retrieved answer. In the second experiment, we consider the generated answer at a more granular level, prompting the LLM to extract a list of factual statements from the answer and verifying each statement separately. We query the corpus with each factual statement and then present the LLM with the statement and the corresponding retrieved evidence. The LLM is prompted to indicate if the statement can be supported and make necessary edits using the retrieved material. With an accuracy of over 80%, we find that an LLM is capable of verifying its generated answer when a corpus of supporting material is provided. However, manual assessment of a random sample of questions reveals that incorrect generated answers are missed by this verification process. While this verification process can reduce hallucinations, it can not entirely eliminate them.","{'model': 'tldr@v2.0.0', 'text': 'With an accuracy of over 80%, it is found that an LLM is capable of verifying its generated answer when a corpus of supporting material is provided, however, manual assessment of a random sample of questions reveals that incorrect generated answers are missed by this verification process.'}",https://arxiv.org/pdf/2309.11392
-Visualizing Linguistic Diversity of Text Datasets Synthesized by Large Language Models,Emily Reif,"Large language models (LLMs) can be used to generate smaller, more refined datasets via few-shot prompting for benchmarking, fine-tuning or other use cases. However, understanding and evaluating these datasets is difficult, and the failure modes of LLM-generated data are still not well understood. Specifically, the data can be repetitive in surprising ways, not only semantically but also syntactically and lexically. We present LinguisticLens, a novel inter-active visualization tool for making sense of and analyzing syntactic diversity of LLM-generated datasets. LinguisticLens clusters text along syntactic, lexical, and semantic axes. It supports hierarchical visualization of a text dataset, allowing users to quickly scan for an overview and inspect individual examples. The live demo is available at shorturl.at/zHOUV.","{'model': 'tldr@v2.0.0', 'text': 'LinguisticLens is a novel inter-active visualization tool for making sense of and analyzing syntactic diversity of LLM-generated datasets and supports hierarchical visualization of a text dataset, allowing users to quickly scan for an overview and inspect individual examples.'}",https://arxiv.org/pdf/2305.11364
-Synergistic Integration of Large Language Models and Cognitive Architectures for Robust AI: An Exploratory Analysis,Oscar J. Romero,"This paper explores the integration of two AI subdisciplines employed in the development of artificial agents that exhibit intelligent behavior: Large Language Models (LLMs) and Cognitive Architectures (CAs). We present three integration approaches, each grounded in theoretical models and supported by preliminary empirical evidence. The modular approach, which introduces four models with varying degrees of integration, makes use of chain-of-thought prompting, and draws inspiration from augmented LLMs, the Common Model of Cognition, and the simulation theory of cognition. The agency approach, motivated by the Society of Mind theory and the LIDA cognitive architecture, proposes the formation of agent collections that interact at micro and macro cognitive levels, driven by either LLMs or symbolic components. The neuro-symbolic approach, which takes inspiration from the CLARION cognitive architecture, proposes a model where bottom-up learning extracts symbolic representations from an LLM layer and top-down guidance utilizes symbolic representations to direct prompt engineering in the LLM layer. These approaches aim to harness the strengths of both LLMs and CAs, while mitigating their weaknesses, thereby advancing the development of more robust AI systems. We discuss the tradeoffs and challenges associated with each approach.","{'model': 'tldr@v2.0.0', 'text': 'Three integration approaches are presented, each grounded in theoretical models and supported by preliminary empirical evidence, that aim to harness the strengths of both LLMs and CAs, while mitigating their weaknesses, thereby advancing the development of more robust AI systems.'}",https://arxiv.org/pdf/2308.09830
-Casteist but Not Racist? Quantifying Disparities in Large Language Model Bias between India and the West,Khyati Khandelwal,"Large Language Models (LLMs), now used daily by millions of users, can encode societal biases, exposing their users to representational harms. A large body of scholarship on LLM bias exists but it predominantly adopts a Western-centric frame and attends comparatively less to bias levels and potential harms in the Global South. In this paper, we quantify stereotypical bias in popular LLMs according to an Indian-centric frame and compare bias levels between the Indian and Western contexts. To do this, we develop a novel dataset which we call Indian-BhED (Indian Bias Evaluation Dataset), containing stereotypical and anti-stereotypical examples for caste and religion contexts. We find that the majority of LLMs tested are strongly biased towards stereotypes in the Indian context, especially as compared to the Western context. We finally investigate Instruction Prompting as a simple intervention to mitigate such bias and find that it significantly reduces both stereotypical and anti-stereotypical biases in the majority of cases for GPT-3.5. The findings of this work highlight the need for including more diverse voices when evaluating LLMs.",,https://arxiv.org/pdf/2309.08573
-Feasibility of Using the Privacy-preserving Large Language Model Vicuna for Labeling Radiology Reports.,Pritam Mukherjee,"Background Large language models (LLMs) such as ChatGPT, though proficient in many text-based tasks, are not suitable for use with radiology reports due to patient privacy constraints. Purpose To test the feasibility of using an alternative LLM (Vicuna-13B) that can be run locally for labeling radiography reports. Materials and Methods Chest radiography reports from the MIMIC-CXR and National Institutes of Health (NIH) data sets were included in this retrospective study. Reports were examined for 13 findings. Outputs reporting the presence or absence of the 13 findings were generated by Vicuna by using a single-step or multistep prompting strategy (prompts 1 and 2, respectively). Agreements between Vicuna outputs and CheXpert and CheXbert labelers were assessed using Fleiss κ. Agreement between Vicuna outputs from three runs under a hyperparameter setting that introduced some randomness (temperature, 0.7) was also assessed. The performance of Vicuna and the labelers was assessed in a subset of 100 NIH reports annotated by a radiologist with use of area under the receiver operating characteristic curve (AUC). Results A total of 3269 reports from the MIMIC-CXR data set (median patient age, 68 years [IQR, 59-79 years]; 161 male patients) and 25 596 reports from the NIH data set (median patient age, 47 years [IQR, 32-58 years]; 1557 male patients) were included. Vicuna outputs with prompt 2 showed, on average, moderate to substantial agreement with the labelers on the MIMIC-CXR (κ median, 0.57 [IQR, 0.45-0.66] with CheXpert and 0.64 [IQR, 0.45-0.68] with CheXbert) and NIH (κ median, 0.52 [IQR, 0.41-0.65] with CheXpert and 0.55 [IQR, 0.41-0.74] with CheXbert) data sets, respectively. Vicuna with prompt 2 performed at par (median AUC, 0.84 [IQR, 0.74-0.93]) with both labelers on nine of 11 findings. Conclusion In this proof-of-concept study, outputs of the LLM Vicuna reporting the presence or absence of 13 findings on chest radiography reports showed moderate to substantial agreement with existing labelers. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Cai in this issue.","{'model': 'tldr@v2.0.0', 'text': 'In this proof-of-concept study, outputs of the LLM Vicuna reporting the presence or absence of 13 findings on chest radiography reports showed moderate to substantial agreement with existing labelers.'}",
-Empirical Study of Zero-Shot NER with ChatGPT,Tingyu Xie,"Large language models (LLMs) exhibited powerful capability in various natural language processing tasks. This work focuses on exploring LLM performance on zero-shot information extraction, with a focus on the ChatGPT and named entity recognition (NER) task. Inspired by the remarkable reasoning capability of LLM on symbolic and arithmetic reasoning, we adapt the prevalent reasoning methods to NER and propose reasoning strategies tailored for NER. First, we explore a decomposed question-answering paradigm by breaking down the NER task into simpler subproblems by labels. Second, we propose syntactic augmentation to stimulate the model's intermediate thinking in two ways: syntactic prompting, which encourages the model to analyze the syntactic structure itself, and tool augmentation, which provides the model with the syntactic information generated by a parsing tool. Besides, we adapt self-consistency to NER by proposing a two-stage majority voting strategy, which first votes for the most consistent mentions, then the most consistent types. The proposed methods achieve remarkable improvements for zero-shot NER across seven benchmarks, including Chinese and English datasets, and on both domain-specific and general-domain scenarios. In addition, we present a comprehensive analysis of the error types with suggestions for optimization directions. We also verify the effectiveness of the proposed methods on the few-shot setting and other LLMs.","{'model': 'tldr@v2.0.0', 'text': ""This work focuses on exploring LLM performance on zero-shot information extraction, with a focus on the ChatGPT and named entity recognition (NER) task, and proposes syntactic augmentation to stimulate the model's intermediate thinking in two ways.""}",
-Large Language Models can Learn Rules,Zhaocheng Zhu,"When prompted with a few examples and intermediate steps, large language models (LLMs) have demonstrated impressive performance in various reasoning tasks. However, prompting methods that rely on implicit knowledge in an LLM often hallucinate incorrect answers when the implicit knowledge is wrong or inconsistent with the task. To tackle this problem, we present Hypotheses-to-Theories (HtT), a framework that learns a rule library for reasoning with LLMs. HtT contains two stages, an induction stage and a deduction stage. In the induction stage, an LLM is first asked to generate and verify rules over a set of training examples. Rules that appear and lead to correct answers sufficiently often are collected to form a rule library. In the deduction stage, the LLM is then prompted to employ the learned rule library to perform reasoning to answer test questions. Experiments on both numerical reasoning and relational reasoning problems show that HtT improves existing prompting methods, with an absolute gain of 11-27% in accuracy. The learned rules are also transferable to different models and to different forms of the same problem.","{'model': 'tldr@v2.0.0', 'text': 'Hypotheses-to-Theories (HtT), a framework that learns a rule library for reasoning with large language models, improves existing prompting methods, with an absolute gain of 11-27% in accuracy.'}",https://arxiv.org/pdf/2310.07064
-Towards Optimizing with Large Language Models,Pei-Fu Guo,"In this work, we conduct an assessment of the optimization capabilities of LLMs across various tasks and data sizes. Each of these tasks corresponds to unique optimization domains, and LLMs are required to execute these tasks with interactive prompting. That is, in each optimization step, the LLM generates new solutions from the past generated solutions with their values, and then the new solutions are evaluated and considered in the next optimization step. Additionally, we introduce three distinct metrics for a comprehensive assessment of task performance from various perspectives. These metrics offer the advantage of being applicable for evaluating LLM performance across a broad spectrum of optimization tasks and are less sensitive to variations in test samples. By applying these metrics, we observe that LLMs exhibit strong optimization capabilities when dealing with small-sized samples. However, their performance is significantly influenced by factors like data size and values, underscoring the importance of further research in the domain of optimization tasks for LLMs.","{'model': 'tldr@v2.0.0', 'text': 'An assessment of the optimization capabilities of LLMs across various tasks and data sizes observes that LLMs exhibit strong optimization capabilities when dealing with small-sized samples, but their performance is significantly influenced by factors like data size and values, underscoring the importance of further research in the domain of optimization tasks for LLMs.'}",https://arxiv.org/pdf/2310.05204
-Leveraging Large Language Models for Exploiting ASR Uncertainty,Pranay Dighe,"While large language models excel in a variety of natural language processing (NLP) tasks, to perform well on spoken language understanding (SLU) tasks, they must either rely on off-the-shelf automatic speech recognition (ASR) systems for transcription, or be equipped with an in-built speech modality. This work focuses on the former scenario, where LLM's accuracy on SLU tasks is constrained by the accuracy of a fixed ASR system on the spoken input. Specifically, we tackle speech-intent classification task, where a high word-error-rate can limit the LLM's ability to understand the spoken intent. Instead of chasing a high accuracy by designing complex or specialized architectures regardless of deployment costs, we seek to answer how far we can go without substantially changing the underlying ASR and LLM, which can potentially be shared by multiple unrelated tasks. To this end, we propose prompting the LLM with an n-best list of ASR hypotheses instead of only the error-prone 1-best hypothesis. We explore prompt-engineering to explain the concept of n-best lists to the LLM; followed by the finetuning of Low-Rank Adapters on the downstream tasks. Our approach using n-best lists proves to be effective on a device-directed speech detection task as well as on a keyword spotting task, where systems using n-best list prompts outperform those using 1-best ASR hypothesis; thus paving the way for an efficient method to exploit ASR uncertainty via LLMs for speech-based applications.","{'model': 'tldr@v2.0.0', 'text': ""This work tackles speech-intent classification task, where a high word-error-rate can limit the LLM's ability to understand the spoken intent, and proposes prompting theLLM with an n-best list of ASR hypotheses instead of only the error-prone 1-best hypothesis.""}",https://arxiv.org/pdf/2309.04842
-Augmenting Large Language Model Translators via Translation Memories,Yongyu Mu,"Using translation memories (TMs) as prompts is a promising approach to in-context learning of machine translation models. In this work, we take a step towards prompting large language models (LLMs) with TMs and making them better translators. We find that the ability of LLMs to ``understand'' prompts is indeed helpful for making better use of TMs. Experiments show that the results of a pre-trained LLM translator can be greatly improved by using high-quality TM-based prompts. These results are even comparable to those of the state-of-the-art NMT systems which have access to large-scale in-domain bilingual data and are well tuned on the downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that the results of a pre-trained LLM translator can be greatly improved by using high-quality TM-based prompts, and are even comparable to those of the state-of-the-art NMT systems which have access to large-scale in-domain bilingual data and are well tuned on the downstream tasks.'}",http://arxiv.org/pdf/2305.17367
-Less is More for Long Document Summary Evaluation by LLMs,Yunshu Wu,"Large Language Models (LLMs) have shown promising performance in summary evaluation tasks, yet they face challenges such as high computational costs and the Lost-in-the-Middle problem where important information in the middle of long documents is often overlooked. To address these issues, this paper introduces a novel approach, Extract-then-Evaluate, which involves extracting key sentences from a long source document and then evaluating the summary by prompting LLMs. The results reveal that the proposed method not only significantly reduces evaluation costs but also exhibits a higher correlation with human evaluations. Furthermore, we provide practical recommendations for optimal document length and sentence extraction methods, contributing to the development of cost-effective yet more accurate methods for LLM-based text generation evaluation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel approach, Extract-then-Evaluate, which involves extracting key sentences from a long source document and then evaluating the summary by prompting LLMs, revealing that the proposed method not only significantly reduces evaluation costs but also exhibits a higher correlation with human evaluations.'}",https://arxiv.org/pdf/2309.07382
-Creating Large Language Model Resistant Exams: Guidelines and Strategies,Simon Larsén,"The proliferation of Large Language Models (LLMs), such as ChatGPT, has raised concerns about their potential impact on academic integrity, prompting the need for LLM-resistant exam designs. This article investigates the performance of LLMs on exams and their implications for assessment, focusing on ChatGPT's abilities and limitations. We propose guidelines for creating LLM-resistant exams, including content moderation, deliberate inaccuracies, real-world scenarios beyond the model's knowledge base, effective distractor options, evaluating soft skills, and incorporating non-textual information. The article also highlights the significance of adapting assessments to modern tools and promoting essential skills development in students. By adopting these strategies, educators can maintain academic integrity while ensuring that assessments accurately reflect contemporary professional settings and address the challenges and opportunities posed by artificial intelligence in education.",,http://arxiv.org/pdf/2304.12203
-Developing a Scalable Benchmark for Assessing Large Language Models in Knowledge Graph Engineering,Lars Meyer,"As the field of Large Language Models (LLMs) evolves at an accelerated pace, the critical need to assess and monitor their performance emerges. We introduce a benchmarking framework focused on knowledge graph engineering (KGE) accompanied by three challenges addressing syntax and error correction, facts extraction and dataset generation. We show that while being a useful tool, LLMs are yet unfit to assist in knowledge graph generation with zero-shot prompting. Consequently, our LLM-KG-Bench framework provides automatic evaluation and storage of LLM responses as well as statistical data and visualization tools to support tracking of prompt engineering and model performance.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that while being a useful tool, LLMs are yet unfit to assist in knowledge graph generation with zero-shot prompting, so the LLM-KG-Bench framework provides automatic evaluation and storage of LLM responses as well as statistical data and visualization tools to support tracking of prompt engineering and model performance.'}",https://arxiv.org/pdf/2308.16622
-Forward-Backward Reasoning in Large Language Models for Mathematical Verification,Weisen Jiang,"Chain-of-Thought (CoT) prompting in large language models (LLMs) has shown promising performance on mathematical reasoning tasks. Recently, Self-Consistency samples a diverse set of reasoning chains with different answers and chooses the answer by majority voting. Though effective, its performance cannot be further improved by sampling more reasoning chains. To address this problem, we propose to integrate backward reasoning into answer verification. We first mask a number in the question by ${\bf x}$. The LLM is then asked to predict the masked number with a candidate answer $A$ embedded in the template: ``If we know the answer to the above question is $\{A\}$, what is the value of unknown variable ${\bf x}$?'' The LLM is expected to predict the masked number successfully if the provided candidate answer is correct. To further improve performance, we propose FOBAR (FOrward-BAckward Reasoning) to combine forward and backward reasoning for verifying candidate answers. Experiments are performed on six standard mathematical data sets and three LLMs (text-davinci-003, GPT-3.5-Turbo, GPT-4). Results show that FOBAR achieves state-of-the-art performance. In particular, FOBAR outperforms Self-Consistency which uses forward reasoning alone, demonstrating that combining forward and forward reasoning is better. It also outperforms existing verification methods, verifying the effectiveness of using the simple template in backward reasoning and the proposed combination.","{'model': 'tldr@v2.0.0', 'text': 'FOBAR (FOrward-BAckward Reasoning) is proposed to combine forward and backward reasoning for verifying candidate answers to improve performance and outperforms existing verification methods.'}",
-C O RRPUS: Codex-Leveraged Structured Representations for Neurosymbolic Story Understanding,Yi Dong,"Story generation and understanding—as with all NLG/NLU tasks—has seen a surge in neurosymbolic work. Researchers have recognized that, while large language models (LLMs) have tremendous utility, they can be augmented with symbolic means to be even better and to make up for many flaws that neural networks have. However, symbolic methods are extremely costly in terms of the amount of time and expertise needed to create them. In this work, we capitalize on state-of-the-art Code-LLMs, such as Codex, to bootstrap the use of symbolic methods for tracking the state of stories and aiding in story understanding. We show that our CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI Task 2 and Re 3 ) with minimal hand engineering. This work highlights the usefulness of code-based symbolic representations for enabling LLMs to better perform story reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that the CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI Task 2 and Re 3 ) with minimal hand engineering.'}",
-Mind vs. Mouth: On Measuring Re-judge Inconsistency of Social Bias in Large Language Models,Yachao Zhao,"Recent researches indicate that Pre-trained Large Language Models (LLMs) possess cognitive constructs similar to those observed in humans, prompting researchers to investigate the cognitive aspects of LLMs. This paper focuses on explicit and implicit social bias, a distinctive two-level cognitive construct in psychology. It posits that individuals' explicit social bias, which is their conscious expression of bias in the statements, may differ from their implicit social bias, which represents their unconscious bias. We propose a two-stage approach and discover a parallel phenomenon in LLMs known as""re-judge inconsistency""in social bias. In the initial stage, the LLM is tasked with automatically completing statements, potentially incorporating implicit social bias. However, in the subsequent stage, the same LLM re-judges the biased statement generated by itself but contradicts it. We propose that this re-judge inconsistency can be similar to the inconsistency between human's unaware implicit social bias and their aware explicit social bias. Experimental investigations on ChatGPT and GPT-4 concerning common gender biases examined in psychology corroborate the highly stable nature of the re-judge inconsistency. This finding may suggest that diverse cognitive constructs emerge as LLMs' capabilities strengthen. Consequently, leveraging psychological theories can provide enhanced insights into the underlying mechanisms governing the expressions of explicit and implicit constructs in LLMs.",,https://arxiv.org/pdf/2308.12578
-The Impact of AI in Physics Education: A Comprehensive Review from GCSE to University Levels,Will Yeadon,"With the rapid evolution of Artificial Intelligence (AI), its potential implications for higher education have become a focal point of interest. This study delves into the capabilities of AI in Physics Education and offers actionable AI policy recommendations. Using a Large Language Model (LLM), we assessed its ability to answer 1337 Physics exam questions spanning GCSE, A-Level, and Introductory University curricula. We employed various AI prompting techniques: Zero Shot, In Context Learning, and Confirmatory Checking, which merges Chain of Thought reasoning with Reflection. The AI's proficiency varied across academic levels: it scored an average of 83.4% on GCSE, 63.8% on A-Level, and 37.4% on university-level questions, with an overall average of 59.9% using the most effective prompting technique. In a separate test, the LLM's accuracy on 5000 mathematical operations was found to decrease as the number of digits increased. Furthermore, when evaluated as a marking tool, the LLM's concordance with human markers averaged at 50.8%, with notable inaccuracies in marking straightforward questions, like multiple-choice. Given these results, our recommendations underscore caution: while current LLMs can consistently perform well on Physics questions at earlier educational stages, their efficacy diminishes with advanced content and complex calculations. LLM outputs often showcase novel methods not in the syllabus, excessive verbosity, and miscalculations in basic arithmetic. This suggests that at university, there's no substantial threat from LLMs for non-invigilated Physics questions. However, given the LLMs' considerable proficiency in writing Physics essays and coding abilities, non-invigilated examinations of these skills in Physics are highly vulnerable to automated completion by LLMs. This vulnerability also extends to Physics questions pitched at lower academic levels.",,
-Data Race Detection Using Large Language Models,Le Chen,"Large language models (LLMs) are demonstrating significant promise as an alternate strategy to facilitate analyses and optimizations of high-performance computing programs, circumventing the need for resource-intensive manual tool creation. In this paper, we explore a novel LLM-based data race detection approach combining prompting engineering and fine-tuning techniques. We create a dedicated dataset named DRB-ML, which is derived from DataRaceBench, with fine-grain labels showing the presence of data race pairs and their associated variables, line numbers, and read/write information. DRB-ML is then used to evaluate representative LLMs and fine-tune open-source ones. Our experiment shows that LLMs can be a viable approach to data race detection. However, they still cannot compete with traditional data race detection tools when we need detailed information about variable pairs causing data races.","{'model': 'tldr@v2.0.0', 'text': 'This paper creates a dedicated dataset named DRB-ML, which is derived from DataRaceBench, with fine-grain labels showing the presence of data race pairs and their associated variables, line numbers, and read/write information, and shows that LLMs can be a viable approach to data race detection.'}",https://arxiv.org/pdf/2308.07505
-S3-DST: Structured Open-Domain Dialogue Segmentation and State Tracking in the Era of LLMs,Sarkar Snigdha Sarathi Das,"The traditional Dialogue State Tracking (DST) problem aims to track user preferences and intents in user-agent conversations. While sufficient for task-oriented dialogue systems supporting narrow domain applications, the advent of Large Language Model (LLM)-based chat systems has introduced many real-world intricacies in open-domain dialogues. These intricacies manifest in the form of increased complexity in contextual interactions, extended dialogue sessions encompassing a diverse array of topics, and more frequent contextual shifts. To handle these intricacies arising from evolving LLM-based chat systems, we propose joint dialogue segmentation and state tracking per segment in open-domain dialogue systems. Assuming a zero-shot setting appropriate to a true open-domain dialogue system, we propose S3-DST, a structured prompting technique that harnesses Pre-Analytical Recollection, a novel grounding mechanism we designed for improving long context tracking. To demonstrate the efficacy of our proposed approach in joint segmentation and state tracking, we evaluate S3-DST on a proprietary anonymized open-domain dialogue dataset, as well as publicly available DST and segmentation datasets. Across all datasets and settings, S3-DST consistently outperforms the state-of-the-art, demonstrating its potency and robustness the next generation of LLM-based chat systems.","{'model': 'tldr@v2.0.0', 'text': 'S3-DST is proposed, a structured prompting technique that harnesses Pre-Analytical Recollection, a novel grounding mechanism the authors designed for improving long context tracking that consistently outperforms the state-of-the-art in joint segmentation and state tracking.'}",https://arxiv.org/pdf/2309.08827
-CORRPUS: Detecting Story Inconsistencies via Codex-Bootstrapped Neurosymbolic Reasoning,Yi Dong,"Story generation and understanding—as with all NLG/NLU tasks—has seen a surge in neurosymbolic work. Researchers have rec-ognized that, while large language models (LLMs) have tremendous utility, they can be augmented with symbolic means to be even better and to make up for any flaws that the neural networks might have. However, symbolic methods are extremely costly in terms of the amount of time and expertise needed to create them. In this work, we capitalize on state-of-the-art Code-LLMs, such as Codex, to boot-strap the use of symbolic methods for tracking the state of stories and aiding in story understanding. We show that our CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI task 2 and Re 3 ) with minimal hand engineering. We hope that this work can help highlight the importance of symbolic representations and specialized prompting for LLMs as these models require some guidance for performing reasoning tasks properly.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI task 2 and Re 3 ) with minimal hand engineering.'}",
-Paradigm shift: motivations for qualified legal professionals to undertake academic study,G. Brennan,"ABSTRACT This paper considers the factors that motivate qualified legal professionals to undertake a further legal academic programme of study. The paper analyses the findings of a recent study concerning a post-graduate research degree collaboration between Northumbria University (NU) and the Law Society of Ireland (LSI) whereby NU’s longstanding LLM Advanced Legal Practice (LLM ALP) has been offered in Ireland through the LSI (LLM ALP (Ire)). The initial offering of the LLM ALP (Ire) has attracted a significant amount of interest from the profession, prompting the authors to consider (i) the factors that motivate ‘time-poor’ professionals to pursue the academic study of law at the post-professional level and (ii) the potential that such a course of study might have to inform and enrich the students’ practice of law. The LLM ALP (Ire) has attracted a significant amount of more established practitioners, whereas those who generally enrol on the LLM ALP in England are at the early stages of their careers, either enrolled on the Legal Practice Course (LPC) or having only recently qualified. Hence this research is focused on a particular student profile of post-professionals with established careers in legal practice.",,http://shura.shu.ac.uk/22696/3/Brewer_paradigm_shift%20%28AM%29.pdf
-Automatic Chain of Thought Prompting in Large Language Models,Zhuosheng Zhang,"Large language models (LLMs) can perform complex reasoning by generating intermediate reasoning steps. Providing these steps for prompting demonstrations is called chain-of-thought (CoT) prompting. CoT prompting has two major paradigms. One leverages a simple prompt like""Let's think step by step""to facilitate step-by-step thinking before answering a question. The other uses a few manual demonstrations one by one, each composed of a question and a reasoning chain that leads to an answer. The superior performance of the second paradigm hinges on the hand-crafting of task-specific demonstrations one by one. We show that such manual efforts may be eliminated by leveraging LLMs with the""Let's think step by step""prompt to generate reasoning chains for demonstrations one by one, i.e., let's think not just step by step, but also one by one. However, these generated chains often come with mistakes. To mitigate the effect of such mistakes, we find that diversity matters for automatically constructing demonstrations. We propose an automatic CoT prompting method: Auto-CoT. It samples questions with diversity and generates reasoning chains to construct demonstrations. On ten public benchmark reasoning tasks with GPT-3, Auto-CoT consistently matches or exceeds the performance of the CoT paradigm that requires manual designs of demonstrations. Code is available at https://github.com/amazon-research/auto-cot","{'model': 'tldr@v2.0.0', 'text': 'An automatic CoT prompting method that samples questions with diversity and generates reasoning chains to construct demonstrations and consistently matches or exceeds the performance of the CoT paradigm that requires manual designs of demonstrations.'}",http://arxiv.org/pdf/2210.03493
-LLM Abstracts for AJH V97 S3,Anemona Lucia,"inflammatory cytokines that are released as part of the systemic inflammatory response to acute pancreatitis. Decrease in nitric oxide (NO) level predisposes the patient to thrombotic microangiopathy. In in vitro studies, it is proved that inflammatory cytokines stimulated endothelial cell release of ultra large vWF multimers and inhibited the cleavage of ultra large vWF by ADAMTS13. In acute pancreatitis, endothelial damage is found to be associated with a decrease in endothelial NO synthase production. In rat models, this decrease in NO synthase has been found to be an important mediator in the pathogenesis of acute pancreatitis.These studies strongly favor the lack of NO as being one of the significant contributors to AP-induced TTP rather than just the lack of ADAMTS-13. Discussion: In patients with acute pancreatitis, any sudden drop in hemoglobin or platelet count should raise the clinical suspicion of TTP, especially with evidence of schistocytes in peripheral blood smear. Diverse mechanisms apart from ADAMTS-13 deficiency may be involved. Prompt diagnosis is required to start early treatment with plasmapheresis which brings a favorable outcome. of V-based were (PFS), patients; reduction or interruption of lenalidomide in 64% of patients; and to interruption of rituximab in 30% of patients (dose reduction for rituximab was not allowed). Neutropenia was the most common TEAE leading to lenalidomide discontinuation in 6% and reduction/interruption in 32%, and rituximab discontinuation in 3%. Infusion-related reaction was the most common TEAE leading to rituximab interruption in 8%. Discussion: These data represent complete analysis of all patients in the induction phase of MAGNIFY which continue to support that R 2 is active with a tolerable safety profile in patients with R/R FL grade 1 – 3a and MZL, including rituximab-refractory, double-refractory, and early relapse patients. of from 8745 patients with MM, 4610 with MGUS, 303 patients with WM, 77 with ALA and 75 with SMM have been collected. Together 13811 patients have been included in the registry until end of June 2022, 10000th patient was assigned to RMG on October 17th 2018. Median follow-up of MGUS patients is 7 years (0.5-46 years) and median follow-up for MM patients is 6 years (0.5-32 years). The huge amount of data allowed already many regular analysis and publication of treatment results of MM patients treated with novel drugs of multiple myeloma in the Czech Republic. The new prognostic models for MGUS and SMM progression have been created based on registry data. The RMG data has been also used in several international projects, such as Insight or Honeur. no evidence of cord compression, and he was given supportive care for his symptoms. Given known history of LMD and high risk of relapse, decision was made to start high-dose intravenous methotrexate for CNS prophylaxis followed by leucovorin rescue. He developed profound renal failure and was started on intravenous fluids, while leucovorin rescue therapy was extended. His renal functions slowly improved and he was discharged to short-term nursing home for continued rehabilitation. However, he was brought back to the hospital within a week with worsening neurological symptoms including intermittent episodes of unilateral blindness, altered mentation, and fluctuating weakness. He also became critically ill with sepsis and febrile neutropenia. He passed away shortly after being transitioned to comfort measures only. Results: Systemic administration of HD-MTX is the most widely used alternative to intrathecal chemotherapy in patients with leptomeningeal metastases and achieves higher concentration of methotrexate in cerebrospinal fluid than does IT administration. Although recommended by some guidelines and increasingly utilized as a means of CNS prophylaxis in high-risk patients, the use of HD-MTX in DLBCL remains controversial, as evidence supporting this practice is limited to small retrospective studies and several recent studies had demonstrated no benefit. In the case of our patient, no obvious benefit was observed following the use of HD-MTX, as he developed signs of recurrent LMD within 2 weeks of therapy completion. Discussion: The role of high-dose intravenous methotrexate for CNS prophylaxis in patients with diffuse large B-cell lymphoma remains controversial and the approach is associated with significant toxicity. Further researches are needed to further evaluate its efficacy. function is in patients (81%) 13 at the stage of end-stage chronic renal failure. The majority of patients (71%) have an ISS III score. All patients are eligible for chemotherapy with a predominance for the VTD protocol. Denosumab is both in newly diagnosed patients and in those (15 patients) treated with BP with an average of 6 monthly injections. (ONJ) of and had no further disease at the last visit. Two patients died after a short evolution of 3 and 12 months, the causes of which were progression and COVID-19 infection. Discussion: PMBCL is a specific lymphoma entity found in young people with good overall survival. It has a distinct clinical profile and prognosis, and therefore requires an approach different therapy. The R/DA-EPOCH regimen stopped the use of radiotherapy. It's necessary be confirmed in a subsequent prospective study on a large cohort. investigated the occurrence of after initiation of zoledronic acid compared to a of hypertension, hyperlipidemia, and benign hypertrophy presented four-months of painless hematuria. prostate Laboratory (CrCl) two the role of this strategy in elderly patients using both matched and mismatched donors. COVID-19 has substantially decreased cancer screening, management visits and surgeries. CVS Health recently developed a best-in-class mobile app and website that enables oncology patients to start and stay on therapy. This study examined the impact of COVID-19 on adherence to oral oncolytic agents in a large health plan with a significant digital health platform. Methods: This retrospective cohort study included adult patients with chronic myelogenous leukemia (CML), ovarian cancer or prostate cancer initiating oral oncolytics between 3/1/19 and 3/1/2021. Patients were divided into two groups: pre-COVID oral oncolytic initiators before 3/1/20 and COVID initiators after 3/1/20 and were followed for 1 year after therapy initiation. The primary outcome was optimal adherence to oral oncolytic agents as defined by a medication possession ratio (MPR) ≥ 0.8. Percent of digital engagement, defined as the number of times a patient interacted with the CVS digital platform, was examined as a secondary endpoint and was considered as a binary and categorical endpoint (none, low (< 28), moderate (28-105) and high ( >105)). Descriptive statistics and logistic regression model-ing were performed; p-values < 0.05 were significant. Results: In total, 15,494 patients were included in the study, with 8,067 (52.07%) in the pre-COVID initiator group. Patient demographics were similar across study groups, with the exception of pre-COVID initiators who were less likely to be male (75.32% vs. 77.34%; p< 0.01) and receive copay assistance (38.37% vs. 41.70%; p< 0.01). No difference was in 3 (50%) of the patients with favorable, intermediate and poor cytogenetic abnormalities in each of the cases respectively. Dose-intensive chemotherapy was given to 5 patients (83%%). Four patients received fludarabine, idarubicin and cytarabine (FIA) and one cladribine, idarubicin, cytarabine (CLIA) and venetoclax. The other patient with intracranial myeloid sarcoma secondary to relapsed APL received all-trans retinoid acid (ATRA) and arsenic trioxide (ATO) achieving a CR, however, the patient died 30 months later. Five patients (83%) achieved morphological and radiological complete remission after induction course. Four (67%) patients are alive at the time of analysis. One expired 30 months after relapsed with intracranial myeloid sarcoma and one was lost to follow up. with MM late in the course and elected not to pursue treatment due to lack of funding. Some individuals could not afford treatment after 1 year. The unaffordability of ASCT is made apparent as only 1 patient was able to pursue it. India needs access to newer therapies at a lower cost to treat MM and provide better outcomes. relapsing DLBCL. Ben-eficial effects of treatment with little toxicity seem limited to fit patients respectively. The most common all-grade TEAE was neutropenia in the Isa-Pd arm, and diarrhea in the Isa-Kd arm. Response data are forthcoming. Discussion: In this IONA-MM first interim analysis, we report comparable pt baseline characteristics to those seen in ICARIA-MM/IKEMA with a few imbalances. Both Isa-Pd and Isa-Kd have a manageable safety profile in routine clinical practice. These data provide RWE to support the use of Isa in RRMM outside of clinical trials and in wider populations. Enrollment in IONA-MM is ongoing and will continue until the sample size (1100 pts) is reached. progression for differential diagnosis. BM examination and LN re-biopsy confirmed atypical CD20 + lymphomatus infiltration. She received 3 rd line chemotherapy (3 GDP courses) and 10 sessions of involved field radiotherapy ( IFRT ) on the STM, she developed severe anemia, acute kidney injury, convulsions and disturbed conscious level, she was admitted to the intensive care unit (ICU), and was planned for CSF analysis after neurological assessment and radiology of the brain, but unfortunately her condition progressed rapidly to multiorgan failure and died. and or transmission of oncogenic such as EBV, and HTLV-1 due to multiple transfusions could be as a predispos-ing factor in development of lymphoma in thalassemic The effects of iron overload and treatment with iron chelators could be another hypothesis for development of malignancies in these patients. in Pakistan. This study aims to compare the response of ICE versus DHAP as salvage che",,https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ajh.26736
-Neuro-Symbolic Procedural Planning with Commonsense Prompting,Yujie Lu,"Procedural planning aims to implement complex high-level goals by decomposition into sequential simpler low-level steps. Although procedural planning is a basic skill set for humans in daily life, it remains a challenge for large language models (LLMs) that lack a deep understanding of the cause-effect relations in procedures. Previous methods require manual exemplars to acquire procedural planning knowledge from LLMs in the zero-shot setting. However, such elicited pre-trained knowledge in LLMs induces spurious correlations between goals and steps, which impair the model generalization to unseen tasks. In contrast, this paper proposes a neuro-symbolic procedural PLANner (PLAN) that elicits procedural planning knowledge from the LLMs with commonsense-infused prompting. To mitigate spurious goal-step correlations, we use symbolic program executors on the latent procedural representations to formalize prompts from commonsense knowledge bases as a causal intervention toward the Structural Causal Model. Both automatic and human evaluations on WikiHow and RobotHow show the superiority of PLAN on procedural planning without further training or manual exemplars.","{'model': 'tldr@v2.0.0', 'text': 'A neuro-symbolic procedural PLANner that elicits procedural planning knowledge from the LLMs with commonsense-infused prompting and uses symbolic program executors on the latent procedural representations to formalize prompts from commonsense knowledge bases as a causal intervention toward the Structural Causal Model.'}",
-"LLM Abstracts, AJH, V96 S1",G. Garcia-Manero,"LLM Abstracts, AJH, V96 S1 PO-1 | Magrolimab + Azacitidine Versus Azacitidine + Placebo in Untreated Higher Risk Myelodysplastic Syndrome: The Phase 3, Randomized, ENHANCE Study Guillermo Garcia-Manero, Naval Daver, Jin Xu, Mark Chao, Trisha Chung, Anderson Tan, Yan Wang, Andrew Wei, Paresh Vyas, David Sallman The University of Texas MD Anderson Cancer Center, Houston, United States, Gilead Sciences, Inc., Foster City, United States, The Alfred Hospital and Monash University, Melbourne, Australia, University of Oxford, Oxford, United Kingdom, Moffitt Cancer Center, Tampa, United States Background: Myelodysplastic syndrome (MDS) is a clonal myeloid disorder characterized by cytopenia and ineffective hematopoiesis. MDS primarily affects older individuals, with the median age of diagnosis at 70 years. Prognosis and treatment decisions are guided by the revised International Prognostic Scoring System (IPSS-R) criteria. Patients with intermediate, high, and very high risk MDS (HR-MDS) have a median overall survival (OS) of 0.8 to 3.7 years. Despite the high unmet need in this patient population, azacitidine (AZA) is the only approved therapy for HR-MDS that has improved OS in clinical trials to date. However, AZA treatment is characterized by low complete response (CR) rates (10% to 17%) with limited OS (< 2 years ), indicating a need for alternative therapies. Magrolimab is a first-in-class monoclonal antibody that blocks the macrophage inhibitory immune checkpoint CD47, a “do not eat me” signal overexpressed on tumor cells. Binding of magrolimab to CD47 leads to phagocytosis of tumor cells. AZA increases expression of tumor cell prophagocytic “eat me” signals, facilitating synergy with magrolimab. In an ongoing Phase 1b study, the combination of magrolimab+ AZA led to high response rates (overall response rate 91%, with a CR of 42%) and an acceptable safety profile without significant immune-related adverse events. ENHANCE (NCT04313881) is a Phase 3 trial comparing the efficacy and safety of magrolimab + AZA with that of AZA+ placebo in previously untreated patients with HR-MDS. Trial Design / Design and Methods: Patients ≥18 years old with previously untreated intermediate to very high risk MDS by IPSS-R are eligible for ENHANCE. Randomization is 1:1 to magrolimab + AZA or AZA + placebo with no crossover allowed. Magrolimab or placebo is administered intravenously with an initial 1 mg/kg priming dose to mitigate ontarget anemia. An intrapatient dose-escalation regimen up to 30 mg/kg is then administered through Cycle 1, 30 mg/kg weekly dosing in Cycle 2, and 30 mg/kg once every 2 weeks in Cycle 3 and beyond. AZA is administered per regional prescribing information. Patients may remain on treatment until disease progression, relapse, loss of clinical benefit, or until unacceptable toxicities occur. The 2 primary efficacy endpoints are CR rate and OS. For patients undergoing allogeneic stem cell transplantation (ASCT), data for the CR rate will be censored at the time of ASCT, and OS will be censored at the last known alive date. Secondary efficacy endpoints include red blood cell transfusion independence rate, eventfree survival, minimal residual disease-negative rate, time to acute myeloid leukemia transformation, and patient-reported Functional Assessment of Cancer Therapy Anemia response rate. Biomarkers of immune cell recruitment, immune cell signaling, and bone marrow penetration of magrolimab will also be explored. Status: As of June 2021, there are 62 sites active globally, including 41 distributed across the US. Patient enrollment began in September 2020, and as of June 2021, 144 patients have been enrolled in the trial. Planned enrollment is approximately 520 patients globally, and accrual is ongoing. PO-2 | Gastrointestinal Bleeding from Necrotizing Enterocolitis in Acute Myelogenous Leukemia Treated with Arterial Embolization Ralph Gunarsa, Ricci Steven, Rio Hermawan, Joyce Bratanata, Cosmas Lesmana, Aru Sudoyo MRCCC Siloam Hospital, South Jakarta, Indonesia Introduction / Background / Significance: Necrotizing Enterocolitis (NE) is a rare intestinal complication in hematological malignancies, it characterized by bowel inflammation with wall thickening, edema, and necrosis. Although the exact pathogenesis is not completely understood, the main elements are intestinal mucosal injury, leukemic infiltration, and superimposed infection. We presented a case report of AML-M4 patient with severe intestinal bleeding as a main presentation of NE. Materials and Methods / Case Presentation / Objective: 22 years old male diagnosed with secondary AML-M4, he experienced GI bleeding before and during FLAG regimen, the profuse life threatening bleeding occurred at 18th day. He experienced fever, abdominal pain, and diarrhea 3 days before the first GI bleeding. He had no history of any GI diseases nor a family history of it. The last GI bleeding production was approximately 750ml. Physical examination revealed pale skin and lips, anemic conjunctiva, tachycardia, and low blood pressure (60/40 mmHg). Fecal examination revealed positive C.difficle toxin, blood test showed very low hemoglobin level (3g/dL), severe leukopenia (40/μl), thrombocytopenia (30.000), and consumptive coagulopathy. He transported to ICU for further monitoring and management. In ICU, he still experienced GI bleeding despite a maximal supportive and medical treatment. He undergone arterial embolization to stop the bleeding. The procedure revealed contrast extravasation at right and media colic DOI: 10.1002/ajh.26351 Am J Hematol. 2021;96:S3–S33. wileyonlinelibrary.com/journal/ajh © 2021 Wiley Periodicals LLC. S3 artery, 4 coils applied to embolize the extravasated artery. After the procedure, the bleeding was reduced significantly, on the following days his condition is stable, and he transported back to the regular ward. Colonoscopy was performed to evaluate his condition, it revealed multiple ulcers with erosion all over the colon mucosa. Biopsy of the tissue revealed chronic inflammation with necrotic component. These findings consistent with necrotizing enterocolitis (NE). Results / Description / Main Outcome Measure(s): NE is a rare condition in hematological malignancies. The proposed mechanism was mucosal injury by direct toxicity of chemotherapeutic agents (ex. cytarabine), leukemic infiltration which more common in myeloid forms of leukemia, and superimposed infection as a result of neutropenic condition. The clinical presentations are abdominal pain, nausea, vomiting, diarrhea, fever, abdominal distention, and life threatening complications of severe bleeding and perforations. It usually appears 10-14 days after completion of chemotherapy when the low leukocyte counts coincide. Imaging modalities by abdominal x-ray, USG, and CT could show thickening and dilated bowel, in this case due to neutropenic isolation condition, the procedure could not be done. Treatment of NE consists of bowel rest, supportive care, prompt antibiotic treatment, and invasive procedure including surgery. In this case, patient's symptoms occurred before the initiation of chemotherapy, and worsen after chemotherapy, this condition support the underlying mechanism are leukemic infiltration complicated by infection and chemotherapy agent. Severe GI bleeding despite maximum medication was successfully treated by arterial embolization. Conclusion(s): 1. In this case, NE occurred from 3 factors: leukemic infiltration, toxicity of chemotherapy, and superimposed infection. We considered the leukemic infiltration play a major role, because the symptoms occurred before chemotherapy. 2. NE in leukemic patients should be managed with multi-modal approach, in this case his NE was improved by chemotherapy to treat the leukemic infiltration, antibiotics for superimposed infection, and embolization for GI bleeding. References 1. Longo DL. Harrison's Hematology and Oncology 3rd Ed. USA: McGraw-Gill. 2017. 832 p. 2. Ebert EC, Hagspiel KD. Gastrointestinal Manifestations of Leukemia. J Gastroenterol Hepatol, 2012; 27 (3): 483-63. 3. Quigley MM, Bethel K, Nowacki M, Millard F, Sharpe R. Neutropenic Enterocolitis: A Rare Presenting Complication of Acute Lekuemia. Am J Hematol, 2001; 66 (3): 213-9. 4. Camera A, et al. Intestinal Toxicity During Induction Chemotherapy with Cytarabine-Based Regimens in Adult Acute Myeloid Leukemia. Hematol J, 2003; 4(): 346-50. 5. Wach M, Dmoszynska A, Wasik-Szczepanek E, Pozarowski A, Drop A, Szczpanek D. Neutropenic Eneterocolitis: A Serious Complication During the Treatment of Acute Leukemias. Ann Hematol, 2004; 83 (8): 522-6. 6. Dotson J, Elhamdani A, Petryna E, Jamil MO, Alsharedi M. Neutropenic Enterocolitis in Patients with FLT3 Mutated Acute Myeloid Leukemia Undergoing Induction Chemotherapy With Midostaurin. Int J Hematol, 2019; 109 (3): 351-5. 7. Chow EJ, Bishop KD. Painless Neutropenic Eneterocolitis in a Patient Undergoing Chemotherapy. Curr Oncol, 2016; 23 (5): e514-16. 8. Sachak T, Arnold MA, Naini BV, et al. Neutropenic enterocolitis: new insights into a deadly entity. Am J Surg Pathol 2015; 39: 1635–42. S4 ABSTRACTS","{'model': 'tldr@v2.0.0', 'text': 'ENHANCE is a Phase 3 trial comparing the efficacy and safety of magrolimab + AZA with that of AZA+ placebo in previously untreated patients with HR-MDS, a clonal myeloid disorder characterized by cytopenia and ineffective hematopoiesis.'}",
-Analyzing BERT’s Knowledge of Hypernymy via Prompting,Michael Hanna,"The high performance of large pretrained language models (LLMs) such as BERT on NLP tasks has prompted questions about BERT’s linguistic capabilities, and how they differ from humans’. In this paper, we approach this question by examining BERT’s knowledge of lexical semantic relations. We focus on hypernymy, the “is-a” relation that relates a word to a superordinate category. We use a prompting methodology to simply ask BERT what the hypernym of a given word is. We find that, in a setting where all hypernyms are guessable via prompting, BERT knows hypernyms with up to 57% accuracy. Moreover, BERT with prompting outperforms other unsupervised models for hypernym discovery even in an unconstrained scenario. However, BERT’s predictions and performance on a dataset containing uncommon hyponyms and hypernyms indicate that its knowledge of hypernymy is still limited.","{'model': 'tldr@v2.0.0', 'text': 'This paper examines BERT’s knowledge of lexical semantic relations, focusing on hypernymy, the “is-a” relation that relates a word to a superordinate category, and finds that, in a setting where all hypernyms are guessable via prompting, BERT knows hypernym with up to 57% accuracy.'}",https://aclanthology.org/2021.blackboxnlp-1.20.pdf
-Prompter: Utilizing Large Language Model Prompting for a Data Efficient Embodied Instruction Following,Y. Inoue,"Embodied Instruction Following (EIF) studies how mobile manipulator robots should be controlled to accomplish long-horizon tasks specified by natural language instructions. While most research on EIF are conducted in simulators, the ultimate goal of the field is to deploy the agents in real life. As such, it is important to minimize the data cost required for training an agent, to help the transition from sim to real. However, many studies only focus on the performance and overlook the data cost -- modules that require separate training on extra data are often introduced without a consideration on deployability. In this work, we propose FILM++ which extends the existing work FILM with modifications that do not require extra data. While all data-driven modules are kept constant, FILM++ more than doubles FILM's performance. Furthermore, we propose Prompter, which replaces FILM++'s semantic search module with language model prompting. Unlike FILM++'s implementation that requires training on extra sets of data, no training is needed for our prompting based implementation while achieving better or at least comparable performance. Prompter achieves 42.64% and 45.72% on the ALFRED benchmark with high-level instructions only and with step-by-step instructions, respectively, outperforming the previous state of the art by 6.57% and 10.31%.","{'model': 'tldr@v2.0.0', 'text': ""FILM++ is proposed, which extends the existing work FILM with modifications that do not require extra data and more than doubles FILM's performance and proposes Prompter, which replaces FILM++'s semantic search module with language model prompting.""}",https://arxiv.org/pdf/2211.03267
-Terminology-Aware Translation with Constrained Decoding and Large Language Model Prompting,Nikolay Bogoychev,"Terminology correctness is important in the downstream application of machine translation, and a prevalent way to ensure this is to inject terminology constraints into a translation system. In our submission to the WMT 2023 terminology translation task, we adopt a translate-then-refine approach which can be domain-independent and requires minimal manual efforts. We annotate random source words with pseudo-terminology translations obtained from word alignment to first train a terminology-aware model. Further, we explore two post-processing methods. First, we use an alignment process to discover whether a terminology constraint has been violated, and if so, we re-decode with the violating word negatively constrained. Alternatively, we leverage a large language model to refine a hypothesis by providing it with terminology constraints. Results show that our terminology-aware model learns to incorporate terminologies effectively, and the large language model refinement process can further improve terminology recall.","{'model': 'tldr@v2.0.0', 'text': 'This work adopts a translate-then-refine approach which can be domain-independent and requires minimal manual efforts, and annotates random source words with pseudo-terminology translations obtained from word alignment to first train a terminology-aware model.'}",https://arxiv.org/pdf/2310.05824
-Chain of Thought Prompting Elicits Reasoning in Large Language Models,Jason Wei,"We explore how generating a chain of thought -- a series of intermediate reasoning steps -- significantly improves the ability of large language models to perform complex reasoning. In particular, we show how such reasoning abilities emerge naturally in sufficiently large language models via a simple method called chain of thought prompting, where a few chain of thought demonstrations are provided as exemplars in prompting. Experiments on three large language models show that chain of thought prompting improves performance on a range of arithmetic, commonsense, and symbolic reasoning tasks. The empirical gains can be striking. For instance, prompting a 540B-parameter language model with just eight chain of thought exemplars achieves state of the art accuracy on the GSM8K benchmark of math word problems, surpassing even finetuned GPT-3 with a verifier.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on three large language models show that chain of thought prompting improves performance on a range of arithmetic, commonsense, and symbolic reasoning tasks.'}",
-Rethinking with Retrieval: Faithful Large Language Model Inference,Hangfeng He,"Despite the success of large language models (LLMs) in various natural language processing (NLP) tasks, the stored knowledge in these models may inevitably be incomplete, out-of-date, or incorrect. This motivates the need to utilize external knowledge to assist LLMs. Unfortunately, current methods for incorporating external knowledge often require additional training or fine-tuning, which can be costly and may not be feasible for LLMs. To address this issue, we propose a novel post-processing approach, rethinking with retrieval (RR), which retrieves relevant external knowledge based on the decomposed reasoning steps obtained from the chain-of-thought (CoT) prompting. This lightweight approach does not require additional training or fine-tuning and is not limited by the input length of LLMs. We evaluate the effectiveness of RR through extensive experiments with GPT-3 on three complex reasoning tasks: commonsense reasoning, temporal reasoning, and tabular reasoning. Our results show that RR can produce more faithful explanations and improve the performance of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel post-processing approach, rethinking with retrieval (RR), which retrieves relevant external knowledge based on the decomposed reasoning steps obtained from the chain-of-thought (CoT) prompting, which can produce more faithful explanations and improve the performance of LLMs.'}",http://arxiv.org/pdf/2301.00303
-Pop Quiz! Can a Large Language Model Help With Reverse Engineering?,H. Pearce,"Large language models (such as OpenAI's Codex) have demonstrated impressive zero-shot multi-task capabilities in the software domain, including code explanation. In this work, we examine if this ability can be used to help with reverse engineering. Specifically, we investigate prompting Codex to identify the purpose, capabilities, and important variable names or values from code, even when the code is produced through decompilation. Alongside an examination of the model's responses in answering open-ended questions, we devise a true/false quiz framework to characterize the performance of the language model. We present an extensive quantitative analysis of the measured performance of the language model on a set of program purpose identification and information extraction tasks: of the 136,260 questions we posed, it answered 72,754 correctly. A key takeaway is that while promising, LLMs are not yet ready for zero-shot reverse engineering.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates whether prompting Codex to identify the purpose, capabilities, and important variable names or values from code, even when the code is produced through decompilation, can be used to help with reverse engineering.'}",
-Least-to-Most Prompting Enables Complex Reasoning in Large Language Models,Denny Zhou,"Chain-of-thought prompting has demonstrated remarkable performance on various natural language reasoning tasks. However, it tends to perform poorly on tasks which requires solving problems harder than the exemplars shown in the prompts. To overcome this challenge of easy-to-hard generalization, we propose a novel prompting strategy, least-to-most prompting. The key idea in this strategy is to break down a complex problem into a series of simpler subproblems and then solve them in sequence. Solving each subproblem is facilitated by the answers to previously solved subproblems. Our experimental results on tasks related to symbolic manipulation, compositional generalization, and math reasoning reveal that least-to-most prompting is capable of generalizing to more difficult problems than those seen in the prompts. A notable finding is that when the GPT-3 code-davinci-002 model is used with least-to-most prompting, it can solve the compositional generalization benchmark SCAN in any split (including length split) with an accuracy of at least 99% using just 14 exemplars, compared to only 16% accuracy with chain-of-thought prompting. This is particularly noteworthy because neural-symbolic models in the literature that specialize in solving SCAN are trained on the entire training set containing over 15,000 examples. We have included prompts for all the tasks in the Appendix.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on tasks related to symbolic manipulation, compositional generalization, and math reasoning reveal that least-to-most prompting is capable of generalizing to more difficult problems than those seen in the prompts.'}",http://arxiv.org/pdf/2205.10625
-TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation,Keqin Bao,"Large Language Models (LLMs) have demonstrated remarkable performance across diverse domains, thereby prompting researchers to explore their potential for use in recommendation systems. Initial attempts have leveraged the exceptional capabilities of LLMs, such as rich knowledge and strong generalization through In-context Learning, which involves phrasing the recommendation task as prompts. Nevertheless, the performance of LLMs in recommendation tasks remains suboptimal due to a substantial disparity between the training tasks for LLMs and recommendation tasks, as well as inadequate recommendation data during pre-training. To bridge the gap, we consider building a Large Recommendation Language Model by tunning LLMs with recommendation data. To this end, we propose an efficient and effective Tuning framework for Aligning LLMs with Recommendations, namely TALLRec. We have demonstrated that the proposed TALLRec framework can significantly enhance the recommendation capabilities of LLMs in the movie and book domains, even with a limited dataset of fewer than 100 samples. Additionally, the proposed framework is highly efficient and can be executed on a single RTX 3090 with LLaMA-7B. Furthermore, the fine-tuned LLM exhibits robust cross-domain generalization. Our code and data are available at https://github.com/SAI990323/TALLRec.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the proposed TALLRec framework can significantly enhance the recommendation capabilities of LLMs in the movie and book domains, even with a limited dataset of fewer than 100 samples.'}",https://arxiv.org/pdf/2305.00447
-Gorilla: Large Language Model Connected with Massive APIs,Shishir G. Patil,"Large Language Models (LLMs) have seen an impressive wave of advances recently, with models now excelling in a variety of tasks, such as mathematical reasoning and program synthesis. However, their potential to effectively use tools via API calls remains unfulfilled. This is a challenging task even for today's state-of-the-art LLMs such as GPT-4, largely due to their inability to generate accurate input arguments and their tendency to hallucinate the wrong usage of an API call. We release Gorilla, a finetuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls. When combined with a document retriever, Gorilla demonstrates a strong capability to adapt to test-time document changes, enabling flexible user updates or version changes. It also substantially mitigates the issue of hallucination, commonly encountered when prompting LLMs directly. To evaluate the model's ability, we introduce APIBench, a comprehensive dataset consisting of HuggingFace, TorchHub, and TensorHub APIs. The successful integration of the retrieval system with Gorilla demonstrates the potential for LLMs to use tools more accurately, keep up with frequently updated documentation, and consequently increase the reliability and applicability of their outputs. Gorilla's code, model, data, and demo are available at https://gorilla.cs.berkeley.edu","{'model': 'tldr@v2.0.0', 'text': 'Gorilla is released, a finetuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls and substantially mitigates the issue of hallucination, commonly encountered when prompting LLMs directly.'}",http://arxiv.org/pdf/2305.15334
-Large Language Model Augmented Narrative Driven Recommendations,Sheshera Mysore,"Narrative-driven recommendation (NDR) presents an information access problem where users solicit recommendations with verbose descriptions of their preferences and context, for example, travelers soliciting recommendations for points of interest while describing their likes/dislikes and travel circumstances. These requests are increasingly important with the rise of natural language-based conversational interfaces for search and recommendation systems. However, NDR lacks abundant training data for models, and current platforms commonly do not support these requests. Fortunately, classical user-item interaction datasets contain rich textual data, e.g., reviews, which often describe user preferences and context – this may be used to bootstrap training for NDR models. In this work, we explore using large language models (LLMs) for data augmentation to train NDR models. We use LLMs for authoring synthetic narrative queries from user-item interactions with few-shot prompting and train retrieval models for NDR on synthetic queries and user-item interaction data. Our experiments demonstrate that this is an effective strategy for training small-parameter retrieval models that outperform other retrieval and LLM baselines for narrative-driven recommendation.","{'model': 'tldr@v2.0.0', 'text': 'Using large language models (LLMs) for data augmentation to train small-parameter retrieval models that outperform other retrieval and LLM baselines for narrative-driven recommendation are explored.'}",https://arxiv.org/pdf/2306.02250
-ThoughtSource: A central hub for large language model reasoning data,Simon Ott,,"{'model': 'tldr@v2.0.0', 'text': 'The goal of ThoughtSource is to improve future artificial intelligence systems by facilitating qualitative understanding of CoTs, enabling empirical evaluations, and providing training data.'}",https://www.nature.com/articles/s41597-023-02433-3.pdf
-Large Language Model Prompt Chaining for Long Legal Document Classification,Dietrich Trautmann,"Prompting is used to guide or steer a language model in generating an appropriate response that is consistent with the desired outcome. Chaining is a strategy used to decompose complex tasks into smaller, manageable components. In this study, we utilize prompt chaining for extensive legal document classification tasks, which present difficulties due to their intricate domain-specific language and considerable length. Our approach begins with the creation of a concise summary of the original document, followed by a semantic search for related exemplar texts and their corresponding annotations from a training corpus. Finally, we prompt for a label - based on the task - to assign, by leveraging the in-context learning from the few-shot prompt. We demonstrate that through prompt chaining, we can not only enhance the performance over zero-shot, but also surpass the micro-F1 score achieved by larger models, such as ChatGPT zero-shot, using smaller models.","{'model': 'tldr@v2.0.0', 'text': 'This study utilizes prompt chaining for extensive legal document classification tasks, which present difficulties due to their intricate domain-specific language and considerable length, and demonstrates that it can enhance the performance over zero-shot, but also surpass the micro-F1 score achieved by larger models.'}",https://arxiv.org/pdf/2308.04138
-Generate rather than Retrieve: Large Language Models are Strong Context Generators,W. Yu,"Knowledge-intensive tasks, such as open-domain question answering (QA), require access to a large amount of world or domain knowledge. A common approach for knowledge-intensive tasks is to employ a retrieve-then-read pipeline that first retrieves a handful of relevant contextual documents from an external corpus such as Wikipedia and then predicts an answer conditioned on the retrieved documents. In this paper, we present a novel perspective for solving knowledge-intensive tasks by replacing document retrievers with large language model generators. We call our method generate-then-read (GenRead), which first prompts a large language model to generate contextutal documents based on a given question, and then reads the generated documents to produce the final answer. Furthermore, we propose a novel clustering-based prompting method that selects distinct prompts, resulting in the generated documents that cover different perspectives, leading to better recall over acceptable answers. We conduct extensive experiments on three different knowledge-intensive tasks, including open-domain QA, fact checking, and dialogue system. Notably, GenRead achieves 71.6 and 54.4 exact match scores on TriviaQA and WebQ, significantly outperforming the state-of-the-art retrieve-then-read pipeline DPR-FiD by +4.0 and +3.9, without retrieving any documents from any external knowledge source. Lastly, we demonstrate the model performance can be further improved by combining retrieval and generation. Our code and generated documents can be found at https://github.com/wyu97/GenRead.","{'model': 'tldr@v2.0.0', 'text': 'The method generate-then-read (GenRead), which first prompts a large language model to generate contextutal documents based on a given question, and then reads the generated documents to produce the final answer, is called.'}",http://arxiv.org/pdf/2209.10063
-Lessons Learned from GPT-SW3: Building the First Large-Scale Generative Language Model for Swedish,Ariel Ekgren,"We present GTP-SW3, a 3.5 billion parameter autoregressive language model, trained on a newly created 100 GB Swedish corpus. This paper provides insights with regards to data collection and training, while highlights the challenges of proper model evaluation. The results of quantitive evaluation through perplexity indicate that GPT-SW3 is a competent model in comparison with existing autoregressive models of similar size. Additionally, we perform an extensive prompting study which reveals the good text generation capabilities of GTP-SW3.","{'model': 'tldr@v2.0.0', 'text': 'The results of quantitive evaluation through perplexity indicate that GPT-SW3 is a competent model in comparison with existing autoregressive models of similar size.'}",
-Exploring Length Generalization in Large Language Models,Cem Anil,"The ability to extrapolate from short problem instances to longer ones is an important form of out-of-distribution generalization in reasoning tasks, and is crucial when learning from datasets where longer problem instances are rare. These include theorem proving, solving quantitative mathematics problems, and reading/summarizing novels. In this paper, we run careful empirical studies exploring the length generalization capabilities of transformer-based language models. We first establish that naively finetuning transformers on length generalization tasks shows significant generalization deficiencies independent of model scale. We then show that combining pretrained large language models' in-context learning abilities with scratchpad prompting (asking the model to output solution steps before producing an answer) results in a dramatic improvement in length generalization. We run careful failure analyses on each of the learning modalities and identify common sources of mistakes that highlight opportunities in equipping language models with the ability to generalize to longer problems.","{'model': 'tldr@v2.0.0', 'text': ""This paper establishes that naively finetuning transformers on length generalization tasks shows significant generalization deficiencies independent of model scale, and shows that combining pretrained large language models' in-context learning abilities with scratchpad prompting results in a dramatic improvement in lengthgeneralization.""}",http://arxiv.org/pdf/2207.04901
-Code as Policies: Language Model Programs for Embodied Control,Jacky Liang,"Large language models (LLMs) trained on code-completion have been shown to be capable of synthesizing simple Python programs from docstrings [1]. We find that these code-writing LLMs can be re-purposed to write robot policy code, given natural language commands. Specifically, policy code can express functions or feedback loops that process perception outputs (e.g., from object detectors [2], [3]) and parameterize control primitive APIs. When provided as input several example language commands (formatted as comments) followed by corresponding policy code (via few-shot prompting), LLMs can take in new commands and autonomously re-compose API calls to generate new policy code respectively. By chaining classic logic structures and referencing third-party libraries (e.g., NumPy, Shapely) to perform arithmetic, LLMs used in this way can write robot policies that (i) exhibit spatial-geometric reasoning, (ii) generalize to new instructions, and (iii) prescribe precise values (e.g., velocities) to ambiguous descriptions (‘faster’) depending on context (i.e., behavioral commonsense). This paper presents Code as Policies: a robot-centric formulation of language model generated programs (LMPs) that can represent reactive policies (e.g., impedance controllers), as well as waypoint-based policies (vision-based pick and place, trajectory-based control), demonstrated across multiple real robot platforms. Central to our approach is prompting hierarchical code-gen (recursively defining undefined functions), which can write more complex code and also improves state-of-the-art to solve 39.8% of problems on the HumanEval [1] benchmark. Code and videos are available at https://code-as-policies.github.io","{'model': 'tldr@v2.0.0', 'text': 'Code as Policies is presented, a robot-centric formulation of language model generated programs (LMPs) that can represent reactive policies (e.g., impedance controllers), as well as waypoint-based policies (vision-based pick and place, trajectory-based control), demonstrated across multiple real robot platforms.'}",https://arxiv.org/pdf/2209.07753
-How Well Does ChatGPT Do When Taking the Medical Licensing Exams? The Implications of Large Language Models for Medical Education and Knowledge Assessment,A. Gilson,"Background: ChatGPT is a 175 billion parameter natural language processing model which can generate conversation style responses to user input. Objective: To evaluate the performance of ChatGPT on questions within the scope of United States Medical Licensing Examination (USMLE) Step 1 and Step 2 exams, as well as analyze responses for user interpretability. Methods: We used two novel sets of multiple choice questions to evaluate ChatGPT's performance, each with questions pertaining to Step 1 and Step 2. The first was derived from AMBOSS, a commonly used question bank for medical students, which also provides statistics on question difficulty and the performance on an exam relative to the userbase. The second, was the National Board of Medical Examiners (NBME) Free 120-question exams. After prompting ChatGPT with each question, ChatGPT's selected answer was recorded, and the text output evaluated across three qualitative metrics: logical justification of the answer selected, presence of information internal to the question, and presence of information external to the question. Results: On the four datasets, AMBOSS-Step1, AMBOSS-Step2, NBME-Free-Step1, and NBME-Free- Step2, ChatGPT achieved accuracies of 44%, 42%, 64.4%, and 57.8%. The model demonstrated a significant decrease in performance as question difficulty increased (P=.012) within the AMBOSS- Step1 dataset. We found logical justification for ChatGPT's answer selection was present in 100% of outputs. Internal information to the question was present in >90% of all questions. The presence of information external to the question was respectively 54.5% and 27% lower for incorrect relative to correct answers on the NBME-Free-Step1 and NBME-Free-Step2 datasets (P<=.001). Conclusion: ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering. By performing at greater than 60% threshold on the NBME-Free- Step-1 dataset we show that the model is comparable to a third year medical student. Additionally, due to the dialogic nature of the response to questions, we demonstrate ChatGPT's ability to provide reasoning and informational context across the majority of answers. These facts taken together make a compelling case for the potential applications of ChatGPT as a medical education tool.","{'model': 'tldr@v2.0.0', 'text': 'ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering and shows that the model is comparable to a third year medical student by performing at greater than 60% threshold on the NBME-Free- Step-1 dataset.'}",https://www.medrxiv.org/content/medrxiv/early/2022/12/26/2022.12.23.22283901.full.pdf
-A Recipe for Arbitrary Text Style Transfer with Large Language Models,Emily Reif,"In this paper, we leverage large language models (LLMs) to perform zero-shot text style transfer. We present a prompting method that we call augmented zero-shot learning, which frames style transfer as a sentence rewriting task and requires only a natural language instruction, without model fine-tuning or exemplars in the target style. Augmented zero-shot learning is simple and demonstrates promising results not just on standard style transfer tasks such as sentiment, but also on arbitrary transformations such as ‘make this melodramatic’ or ‘insert a metaphor.’","{'model': 'tldr@v2.0.0', 'text': 'Augmented zero-shot learning is simple and demonstrates promising results not just on standard style transfer tasks such as sentiment, but also on arbitrary transformations such as ‘make this melodramatic’ or ‘insert a metaphor’.'}",https://aclanthology.org/2022.acl-short.94.pdf
-Dehallucinating Large Language Models Using Formal Methods Guided Iterative Prompting,Susmit Jha,"Large language models (LLMs) such as ChatGPT have been trained to generate human-like responses to natural language prompts. LLMs use a vast corpus of text data for training, and can generate coherent and contextually relevant responses to a wide range of questions and statements. Despite this remarkable progress, LLMs are prone to hallucinations making their application to safety-critical applications such as autonomous systems difficult. The hallucinations in LLMs refer to instances where the model generates responses that are not factually accurate or contextually appropriate. These hallucinations can occur due to a variety of factors, such as the model’s lack of real-world knowledge, the influence of biased or inaccurate training data, or the model’s tendency to generate responses based on statistical patterns rather than a true understanding of the input. While these hallucinations are a nuisance in tasks such as text summarization and question-answering, they can be catastrophic when LLMs are used in autonomy-relevant applications such as planning. In this paper, we focus on the application of LLMs in autonomous systems and sketch a novel self-monitoring and iterative prompting architecture that uses formal methods to detect these errors in the LLM response automatically. We exploit the dialog capability of LLMs to iteratively steer them to responses that are consistent with our correctness specification. We report preliminary experiments that show the promise of the proposed approach on tasks such as automated planning.","{'model': 'tldr@v2.0.0', 'text': 'This paper sketches a novel self-monitoring and iterative prompting architecture that uses formal methods to detect errors in the LLM response automatically and exploits the dialog capability of LLMs to iteratively steer them to responses that are consistent with the correctness specification.'}",
-"CAMEL: Communicative Agents for ""Mind"" Exploration of Large Scale Language Model Society",G. Li,"The rapid advancement of conversational and chat-based language models has led to remarkable progress in complex task-solving. However, their success heavily relies on human input to guide the conversation, which can be challenging and time-consuming. This paper explores the potential of building scalable techniques to facilitate autonomous cooperation among communicative agents and provide insight into their""cognitive""processes. To address the challenges of achieving autonomous cooperation, we propose a novel communicative agent framework named role-playing. Our approach involves using inception prompting to guide chat agents toward task completion while maintaining consistency with human intentions. We showcase how role-playing can be used to generate conversational data for studying the behaviors and capabilities of chat agents, providing a valuable resource for investigating conversational language models. Our contributions include introducing a novel communicative agent framework, offering a scalable approach for studying the cooperative behaviors and capabilities of multi-agent systems, and open-sourcing our library to support research on communicative agents and beyond. The GitHub repository of this project is made publicly available on: https://github.com/lightaime/camel.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel communicative agent framework named role-playing, using inception prompting to guide chat agents toward task completion while maintaining consistency with human intentions, and showcases how role- playing can be used to generate conversational data for studying the behaviors and capabilities of chat agents.'}",http://arxiv.org/pdf/2303.17760
-Towards a Mathematics Formalisation Assistant using Large Language Models,Ayush Agrawal,"Mathematics formalisation is the task of writing mathematics (i.e., definitions, theorem statements, proofs) in natural language, as found in books and papers, into a formal language that can then be checked for correctness by a program. It is a thriving activity today, however formalisation remains cumbersome. In this paper, we explore the abilities of a large language model (Codex) to help with formalisation in the Lean theorem prover. We find that with careful inputdependent prompt selection and postprocessing, Codex is able to formalise short mathematical statements at undergrad level with nearly 75% accuracy for 120 theorem statements. For proofs quantitative analysis is infeasible and we undertake a detailed case study. We choose a diverse set of 13 theorems at undergrad level with proofs that fit in two-three paragraphs. We show that with a new prompting strategy Codex can formalise these proofs in natural language with at least one out of twelve Codex completion being easy to repair into a complete proof. This is surprising as essentially no aligned data exists for formalised mathematics, particularly for proofs. These results suggest that large language models are a promising avenue towards fully or partially automating formalisation.","{'model': 'tldr@v2.0.0', 'text': 'The abilities of a large language model (Codex) to help with formalisation in the Lean theorem prover are explored, finding that with careful inputdependent prompt selection and postprocessing, Codex is able to formalise short mathematical statements at undergrad level with nearly 75% accuracy for 120 theorem statements.'}",https://arxiv.org/pdf/2211.07524
-Meta-Learning the Difference: Preparing Large Language Models for Efficient Adaptation,Zejiang Hou,"Abstract Large pretrained language models (PLMs) are often domain- or task-adapted via finetuning or prompting. Finetuning requires modifying all of the parameters and having enough data to avoid overfitting while prompting requires no training and few examples but limits performance. Instead, we prepare PLMs for data- and parameter-efficient adaptation by learning to learn the difference between general and adapted PLMs. This difference is expressed in terms of model weights and sublayer structure through our proposed dynamic low-rank reparameterization and learned architecture controller. Experiments on few-shot dialogue completion, low-resource abstractive summarization, and multi-domain language modeling show improvements in adaptation time and performance over direct finetuning or preparation via domain-adaptive pretraining. Ablations show our task-adaptive reparameterization (TARP) and model search (TAMS) components individually improve on other parameter-efficient transfer like adapters and structure-learning methods like learned sparsification.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on few-shot dialogue completion, low-resource abstractive summarization, and multi-domain language modeling show improvements in adaptation time and performance over direct finetuning or preparation via domain-adaptive pretraining.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00517/2059907/tacl_a_00517.pdf
-Zero-Shot Temporal Action Detection via Vision-Language Prompting,Sauradip Nag,"Existing temporal action detection (TAD) methods rely on large training data including segment-level annotations, limited to recognizing previously seen classes alone during inference. Collecting and annotating a large training set for each class of interest is costly and hence unscalable. Zero-shot TAD (ZS-TAD) resolves this obstacle by enabling a pre-trained model to recognize any unseen action classes. Meanwhile, ZS-TAD is also much more challenging with significantly less investigation. Inspired by the success of zero-shot image classification aided by vision-language (ViL) models such as CLIP, we aim to tackle the more complex TAD task. An intuitive method is to integrate an off-the-shelf proposal detector with CLIP style classification. However, due to the sequential localization (e.g, proposal generation) and classification design, it is prone to localization error propagation. To overcome this problem, in this paper we propose a novel zero-Shot Temporal Action detection model via Vision-LanguagE prompting (STALE). Such a novel design effectively eliminates the dependence between localization and classification by breaking the route for error propagation in-between. We further introduce an interaction mechanism between classification and localization for improved optimization. Extensive experiments on standard ZS-TAD video benchmarks show that our STALE significantly outperforms state-of-the-art alternatives. Besides, our model also yields superior results on supervised TAD over recent strong competitors. The PyTorch implementation of STALE is available at https://github.com/sauradip/STALE.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-Shot Temporal Action detection model via Vision-LanguagE prompting (STALE) is proposed, which effectively eliminates the dependence between localization and classification by breaking the route for error propagation in-between and introduces an interaction mechanism between classification and localization for improved optimization.'}",http://arxiv.org/pdf/2207.08184
-Meta-learning via Language Model In-context Tuning,Yanda Chen,"The goal of meta-learning is to learn to adapt to a new task with only a few labeled examples. Inspired by the recent progress in large language models, we propose \textit{in-context tuning} (ICT), which recasts task adaptation and prediction as a simple sequence prediction problem: to form the input sequence, we concatenate the task instruction, labeled in-context examples, and the target input to predict; to meta-train the model to learn from in-context examples, we fine-tune a pre-trained language model (LM) to predict the target label given the input sequence on a collection of tasks.We benchmark our method on two collections of text classification tasks: LAMA and BinaryClfs. Compared to MAML which adapts the model through gradient descent, our method leverages the inductive bias of pre-trained LMs to perform pattern matching, and outperforms MAML by an absolute 6% average AUC-ROC score on BinaryClfs, gaining more advantage with increasing model size. Compared to non-fine-tuned in-context learning (i.e. prompting a raw LM), in-context tuning meta-trains the model to learn from in-context examples. On BinaryClfs, ICT improves the average AUC-ROC score by an absolute 10%, and reduces the variance due to example ordering by 6x and example choices by 2x.","{'model': 'tldr@v2.0.0', 'text': 'ICT leverages the inductive bias of pre-trained LMs to perform pattern matching, and outperforms MAML by an absolute 6% average AUC-ROC score on BinaryClfs, gaining more advantage with increasing model size.'}",https://aclanthology.org/2022.acl-long.53.pdf
-Tree of Thoughts: Deliberate Problem Solving with Large Language Models,Shunyu Yao,"Language models are increasingly being deployed for general problem solving across a wide range of tasks, but are still confined to token-level, left-to-right decision-making processes during inference. This means they can fall short in tasks that require exploration, strategic lookahead, or where initial decisions play a pivotal role. To surmount these challenges, we introduce a new framework for language model inference, Tree of Thoughts (ToT), which generalizes over the popular Chain of Thought approach to prompting language models, and enables exploration over coherent units of text (thoughts) that serve as intermediate steps toward problem solving. ToT allows LMs to perform deliberate decision making by considering multiple different reasoning paths and self-evaluating choices to decide the next course of action, as well as looking ahead or backtracking when necessary to make global choices. Our experiments show that ToT significantly enhances language models' problem-solving abilities on three novel tasks requiring non-trivial planning or search: Game of 24, Creative Writing, and Mini Crosswords. For instance, in Game of 24, while GPT-4 with chain-of-thought prompting only solved 4% of tasks, our method achieved a success rate of 74%. Code repo with all prompts: https://github.com/ysymyth/tree-of-thought-llm.","{'model': 'tldr@v2.0.0', 'text': 'A new framework for language model inference, Tree of Thoughts (ToT), which generalizes over the popular Chain of Thought approach to prompting language models, and enables exploration over coherent units of text (thoughts) that serve as intermediate steps toward problem solving.'}",http://arxiv.org/pdf/2305.10601
-HTLM: Hyper-Text Pre-Training and Prompting of Language Models,Armen Aghajanyan,"We introduce HTLM, a hyper-text language model trained on a large-scale web crawl. Modeling hyper-text has a number of advantages: (1) it is easily gathered at scale, (2) it provides rich document-level and end-task-adjacent supervision (e.g. class and id attributes often encode document category information), and (3) it allows for new structured prompting that follows the established semantics of HTML (e.g. to do zero-shot summarization by infilling title tags for a webpage that contains the input text). We show that pretraining with a BART-style denoising loss directly on simplified HTML provides highly effective transfer for a wide range of end tasks and supervision levels. HTLM matches or exceeds the performance of comparably sized text-only LMs for zero-shot prompting and fine-tuning for classification benchmarks, while also setting new state-of-the-art performance levels for zero-shot summarization. We also find that hyper-text prompts provide more value to HTLM, in terms of data efficiency, than plain text prompts do for existing LMs, and that HTLM is highly effective at auto-prompting itself, by simply generating the most likely hyper-text formatting for any available training data. We will release all code and models to support future HTLM research.","{'model': 'tldr@v2.0.0', 'text': 'It is found thathyper-text prompts provide more value to HTLM, in terms of data efficiency, than plain text prompts do for existing LMs, and that HTLM is highly effective at auto-prompting itself, by simply generating the most likely hyper-text formatting for any available training data.'}",
-Large Language Models are Zero-Shot Rankers for Recommender Systems,Yupeng Hou,"Recently, large language models (LLMs) (e.g. GPT-4) have demonstrated impressive general-purpose task-solving abilities, including the potential to approach recommendation tasks. Along this line of research, this work aims to investigate the capacity of LLMs that act as the ranking model for recommender systems. To conduct our empirical study, we first formalize the recommendation problem as a conditional ranking task, considering sequential interaction histories as conditions and the items retrieved by the candidate generation model as candidates. We adopt a specific prompting approach to solving the ranking task by LLMs: we carefully design the prompting template by including the sequential interaction history, the candidate items, and the ranking instruction. We conduct extensive experiments on two widely-used datasets for recommender systems and derive several key findings for the use of LLMs in recommender systems. We show that LLMs have promising zero-shot ranking abilities, even competitive to or better than conventional recommendation models on candidates retrieved by multiple candidate generators. We also demonstrate that LLMs struggle to perceive the order of historical interactions and can be affected by biases like position bias, while these issues can be alleviated via specially designed prompting and bootstrapping strategies. The code to reproduce this work is available at https://github.com/RUCAIBox/LLMRank.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that LLMs have promising zero-shot ranking abilities, even competitive to or better than conventional recommendation models on candidates retrieved by multiple candidate generators, and these issues can be alleviated via specially designed prompting and bootstrapping strategies.'}",http://arxiv.org/pdf/2305.08845
-Query2doc: Query Expansion with Large Language Models,Liang Wang,"This paper introduces a simple yet effective query expansion approach, denoted as query2doc, to improve both sparse and dense retrieval systems. The proposed method first generates pseudo-documents by few-shot prompting large language models (LLMs), and then expands the query with generated pseudo-documents. LLMs are trained on web-scale text corpora and are adept at knowledge memorization. The pseudo-documents from LLMs often contain highly relevant information that can aid in query disambiguation and guide the retrievers. Experimental results demonstrate that query2doc boosts the performance of BM25 by 3% to 15% on ad-hoc IR datasets, such as MS-MARCO and TREC DL, without any model fine-tuning. Furthermore, our method also benefits state-of-the-art dense retrievers in terms of both in-domain and out-of-domain results.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a simple yet effective query expansion approach, denoted as query2doc, to improve both sparse and dense retrieval systems, and benefits state-of-the-art dense retrievers in terms of both in-domain and out- of-domain results.'}",https://arxiv.org/pdf/2303.07678
-Have LLMs Advanced Enough? A Challenging Problem Solving Benchmark For Large Language Models,Daman Arora,"The performance of large language models (LLMs) on existing reasoning benchmarks has significantly improved over the past years. In response, we present JEEBench, a considerably more challenging benchmark dataset for evaluating the problem solving abilities of LLMs. We curate 515 challenging pre-engineering mathematics, physics and chemistry problems from the highly competitive IIT JEE-Advanced exam. Long-horizon reasoning on top of deep in-domain knowledge is essential for solving problems in this benchmark. Our evaluation on various open-source and proprietary models reveals that the highest performance, even after using techniques like self-consistency, self-refinement and chain-of-thought prompting, is less than 40%. The typical failure modes of GPT-4, the best model, are errors in algebraic manipulation, difficulty in grounding abstract concepts into mathematical equations accurately and failure in retrieving relevant domain-specific concepts. We also observe that by mere prompting, GPT-4 is unable to assess risk introduced by negative marking for incorrect answers. For this, we develop a post-hoc confidence-thresholding method over self-consistency, which enables effective response selection. We hope that our challenging benchmark will guide future re-search in problem-solving using LLMs.","{'model': 'tldr@v2.0.0', 'text': 'JEEBench is presented, a considerably more challenging benchmark dataset for evaluating the problem solving abilities of LLMs, and a post-hoc confidence-thresholding method over self-consistency is developed, which enables effective response selection.'}",https://arxiv.org/pdf/2305.15074
-Explainability for Large Language Models: A Survey,Haiyan Zhao,"Large language models (LLMs) have demonstrated impressive capabilities in natural language processing. However, their internal mechanisms are still unclear and this lack of transparency poses unwanted risks for downstream applications. Therefore, understanding and explaining these models is crucial for elucidating their behaviors, limitations, and social impacts. In this paper, we introduce a taxonomy of explainability techniques and provide a structured overview of methods for explaining Transformer-based language models. We categorize techniques based on the training paradigms of LLMs: traditional fine-tuning-based paradigm and prompting-based paradigm. For each paradigm, we summarize the goals and dominant approaches for generating local explanations of individual predictions and global explanations of overall model knowledge. We also discuss metrics for evaluating generated explanations, and discuss how explanations can be leveraged to debug models and improve performance. Lastly, we examine key challenges and emerging opportunities for explanation techniques in the era of LLMs in comparison to conventional machine learning models.","{'model': 'tldr@v2.0.0', 'text': 'A taxonomy of explainability techniques and a structured overview of methods for explaining Transformer-based language models is provided and key challenges and emerging opportunities for explanation techniques in the era of LLMs in comparison to conventional machine learning models are examined.'}",https://arxiv.org/pdf/2309.01029
-Adapting Large Language Models via Reading Comprehension,Daixuan Cheng,"We explore how continued pre-training on domain-specific corpora influences large language models, revealing that training on the raw corpora endows the model with domain knowledge, but drastically hurts its prompting ability for question answering. Taken inspiration from human learning via reading comprehension--practice after reading improves the ability to answer questions based on the learned knowledge--we propose a simple method for transforming raw corpora into reading comprehension texts. Each raw text is enriched with a series of tasks related to its content. Our method, highly scalable and applicable to any pre-training corpora, consistently enhances performance across various tasks in three different domains: biomedicine, finance, and law. Notably, our 7B language model achieves competitive performance with domain-specific models of much larger scales, such as BloombergGPT-50B. Furthermore, we demonstrate that domain-specific reading comprehension texts can improve the model's performance even on general benchmarks, showing the potential to develop a general model across even more domains. Our model, code, and data will be available at https://github.com/microsoft/LMOps.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that domain-specific reading comprehension texts can improve the model's performance even on general benchmarks, showing the potential to develop a general model across even more domains.""}",https://arxiv.org/pdf/2309.09530
-RadAdapt: Radiology Report Summarization via Lightweight Domain Adaptation of Large Language Models,Dave Van Veen,"We systematically investigate lightweight strategies to adapt large language models (LLMs) for the task of radiology report summarization (RRS). Specifically, we focus on domain adaptation via pretraining (on natural language, biomedical text, or clinical text) and via discrete prompting or parameter-efficient fine-tuning. Our results consistently achieve best performance by maximally adapting to the task via pretraining on clinical text and fine-tuning on RRS examples. Importantly, this method fine-tunes a mere 0.32% of parameters throughout the model, in contrast to end-to-end fine-tuning (100% of parameters). Additionally, we study the effect of in-context examples and out-of-distribution (OOD) training before concluding with a radiologist reader study and qualitative analysis. Our findings highlight the importance of domain adaptation in RRS and provide valuable insights toward developing effective natural language processing solutions for clinical tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work systematically investigates lightweight strategies to adapt large language models for the task of radiology report summarization via pretraining and via discrete prompting or parameter-efficient fine-tuning and highlights the importance of domain adaptation in RRS.'}",https://arxiv.org/pdf/2305.01146
-Evaluating Factual Consistency of Summaries with Large Language Models,Shiqi Chen,"Detecting factual errors in summaries has been an important and challenging subject in summarization research. Inspired by the emergent ability of large language models (LLMs), we explore evaluating factual consistency of summaries by directly prompting LLMs. We present a comprehensive empirical study to assess the ability of LLMs as factual consistency evaluators, which consists of (1) analyzing different LLMs such as the GPT model series and Flan-T5; (2) investigating a variety of prompting methods including vanilla prompting, chain-of-thought prompting, and a sentence-by-sentence prompting method to tackle long summaries; and (3) evaluating on diverse summaries generated by multiple summarization systems, ranging from pre-transformer methods to SOTA pretrained models. Our experiments demonstrate that prompting LLMs is able to outperform the previous best factuality systems in all settings, by up to 12.2 absolute points in terms of the binary classification accuracy on inconsistency detection.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive empirical study to assess the ability of LLMs as factual consistency evaluators, which consists of analyzing different LLMs such as the GPT model series and Flan-T5 and investigating a variety of prompting methods including vanilla prompting, chain-of-thought prompting, and a sentence-by-sentence prompting method to tackle long summaries.'}",https://arxiv.org/pdf/2305.14069
-Large Language Models are Diverse Role-Players for Summarization Evaluation,Ning Wu,"Text summarization has a wide range of applications in many scenarios. The evaluation of the quality of the generated text is a complex problem. A big challenge to language evaluation is that there is a clear divergence between existing metrics and human evaluation. A document summary's quality can be assessed by human annotators on various criteria, both objective ones like grammar and correctness, and subjective ones like informativeness, succinctness, and appeal. Most of the automatic evaluation methods like BLUE/ROUGE may be not able to adequately capture the above dimensions. In this paper, we propose a new evaluation framework based on LLMs, which provides a comprehensive evaluation framework by comparing generated text and reference text from both objective and subjective aspects. First, we propose to model objective and subjective dimensions of generated text based on roleplayers prompting mechanism. Furthermore, we introduce a context-based prompting mechanism that is able to generate dynamic roleplayer profiles based on input context. Finally, we design a multi-roleplayer prompting technology based on batch prompting and integrate multiple outputs into the final evaluation results. Experimental results on three real datasets for summarization show that our model is highly competitive and has a very high consistency with human annotators.","{'model': 'tldr@v2.0.0', 'text': 'A new evaluation framework based on LLMs is proposed, which provides a comprehensive evaluation framework by comparing generated text and reference text from both objective and subjective aspects and introduces a context-based prompting mechanism that is able to generate dynamic roleplayer profiles based on input context.'}",https://arxiv.org/pdf/2303.15078
-Can ChatGPT Detect Intent? Evaluating Large Language Models for Spoken Language Understanding,Mutian He,"Recently, large pretrained language models have demonstrated strong language understanding capabilities. This is particularly reflected in their zero-shot and in-context learning abilities on downstream tasks through prompting. To assess their impact on spoken language understanding (SLU), we evaluate several such models like ChatGPT and OPT of different sizes on multiple benchmarks. We verify the emergent ability unique to the largest models as they can reach intent classification accuracy close to that of supervised models with zero or few shots on various languages given oracle transcripts. By contrast, the results for smaller models fitting a single GPU fall far behind. We note that the error cases often arise from the annotation scheme of the dataset; responses from ChatGPT are still reasonable. We show, however, that the model is worse at slot filling, and its performance is sensitive to ASR errors, suggesting serious challenges for the application of those textual models on SLU.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the ChatGPT model is worse at slot filling, and its performance is sensitive to ASR errors, suggesting serious challenges for the application of those textual models on SLU.'}",https://arxiv.org/pdf/2305.13512
-Complexity-Based Prompting for Multi-Step Reasoning,Yao Fu,"We study the task of prompting large-scale language models to perform multi-step reasoning. Existing work shows that when prompted with a chain of thoughts (CoT), sequences of short sentences describing intermediate reasoning steps towards a final answer, large language models can generate new reasoning chains and predict answers for new inputs. A central question is which reasoning examples make the most effective prompts. In this work, we propose complexity-based prompting, a simple and effective example selection scheme for multi-step reasoning. We show that prompts with higher reasoning complexity, i.e., chains with more reasoning steps, achieve substantially better performance on multi-step reasoning tasks over strong baselines. We further extend our complexity-based criteria from prompting (selecting inputs) to decoding (selecting outputs), where we sample multiple reasoning chains from the model, then choose the majority of generated answers from complex reasoning chains (over simple chains). When used to prompt GPT-3 and Codex, our approach substantially improves multi-step reasoning accuracy and achieves new state-of-the-art (SOTA) performance on three math benchmarks (GSM8K, MultiArith, and MathQA) and two BigBenchHard tasks (Date Understanding and Penguins), with an average +5.3 and up to +18 accuracy improvements. Compared with existing example selection schemes like manual tuning or retrieval-based selection, selection based on reasoning complexity is intuitive, easy to implement, and annotation-efficient. Further results demonstrate the robustness of performance gains from complex prompts under format perturbation and distribution shift.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes complexity-based prompting, a simple and effective example selection scheme for multi-step reasoning that substantially improves multi- step reasoning accuracy and achieves new state-of-the-art (SOTA) performance on three math benchmarks and two BigBenchHard tasks.'}",http://arxiv.org/pdf/2210.00720
-"""According to ..."" Prompting Language Models Improves Quoting from Pre-Training Data",Orion Weller,"Large Language Models (LLMs) may hallucinate and generate fake information, despite pre-training on factual data. Inspired by the journalistic device of""according to sources"", we propose according-to prompting: directing LLMs to ground responses against previously observed text. To quantify this grounding, we propose a novel evaluation metric (QUIP-Score) that measures the extent to which model-produced answers are directly found in underlying text corpora. We illustrate with experiments on Wikipedia that these prompts improve grounding under our metrics, with the additional benefit of often improving end-task performance. Furthermore, prompts that ask the model to decrease grounding (or to ground to other corpora) decrease grounding, indicating the ability of language models to increase or decrease grounded generations on request.","{'model': 'tldr@v2.0.0', 'text': 'According-to prompting is proposed: directing LLMs to ground responses against previously observed text, and a novel evaluation metric (QUIP-Score) is proposed that measures the extent to which model-produced answers are directly found in underlying text corpora.'}",http://arxiv.org/pdf/2305.13252
-Prompting for a conversation: How to control a dialog model?,Josef Valvoda,"Dialog modelling faces a difficult trade-off. Models are trained on a large amount of text, yet their responses need to be limited to a desired scope and style of a dialog agent. Because the datasets used to achieve the former contain language that is not compatible with the latter, pre-trained dialog models are fine-tuned on smaller curated datasets. However, the fine-tuning process robs them of the ability to produce diverse responses, eventually reducing them to dull conversation partners. In this paper we investigate if prompting can help with mitigating the above trade-off. Specifically, we experiment with conditioning the prompt on the query, rather than training a single prompt for all queries. By following the intuition that freezing the pre-trained language model will conserve its expressivity, we find that compared to fine-tuning, prompting can achieve a higher BLEU score and substantially improve the diversity and novelty of the responses.","{'model': 'tldr@v2.0.0', 'text': 'This paper experiments with conditioning the prompt on the query, rather than training a single prompt for all queries and finds that compared to fine-tuning, prompting can achieve a higher BLEU score and substantially improve the diversity and novelty of the responses.'}",http://arxiv.org/pdf/2209.11068
-Scaling Instruction-Finetuned Language Models,Hyung Won Chung,"Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling the model size, and (3) finetuning on chain-of-thought data. We find that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups (zero-shot, few-shot, CoT), and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation). For instance, Flan-PaLM 540B instruction-finetuned on 1.8K tasks outperforms PALM 540B by a large margin (+9.4% on average). Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU. We also publicly release Flan-T5 checkpoints, which achieve strong few-shot performance even compared to much larger models, such as PaLM 62B. Overall, instruction finetuning is a general method for improving the performance and usability of pretrained language models.","{'model': 'tldr@v2.0.0', 'text': 'It is found that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups, and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation).'}",http://arxiv.org/pdf/2210.11416
-Multi-Stage Prompting for Knowledgeable Dialogue Generation,Zihan Liu,"Existing knowledge-grounded dialogue systems typically use finetuned versions of a pretrained language model (LM) and large-scale knowledge bases. These models typically fail to generalize on topics outside of the knowledge base, and require maintaining separate potentially large checkpoints each time finetuning is needed. In this paper, we aim to address these limitations by leveraging the inherent knowledge stored in the pretrained LM as well as its powerful generation ability. We propose a multi-stage prompting approach to generate knowledgeable responses from a single pretrained LM. We first prompt the LM to generate knowledge based on the dialogue context. Then, we further prompt it to generate responses based on the dialogue context and the previously generated knowledge. Results show that our knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness. In addition, our multi-stage prompting outperforms the finetuning-based dialogue model in terms of response knowledgeability and engagement by up to 10% and 5%, respectively. Furthermore, we scale our model up to 530 billion parameters and demonstrate that larger LMs improve the generation correctness score by up to 10%, and response relevance, knowledgeability and engagement by up to 10%. Our code is available at: https://github.com/NVIDIA/Megatron-LM.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-stage prompting approach to generate knowledgeable responses from a single pretrained language model (LM) and shows that its knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness.'}",http://arxiv.org/pdf/2203.08745
-ReAct: Synergizing Reasoning and Acting in Language Models,Shunyu Yao,"While large language models (LLMs) have demonstrated impressive capabilities across tasks in language understanding and interactive decision making, their abilities for reasoning (e.g. chain-of-thought prompting) and acting (e.g. action plan generation) have primarily been studied as separate topics. In this paper, we explore the use of LLMs to generate both reasoning traces and task-specific actions in an interleaved manner, allowing for greater synergy between the two: reasoning traces help the model induce, track, and update action plans as well as handle exceptions, while actions allow it to interface with external sources, such as knowledge bases or environments, to gather additional information. We apply our approach, named ReAct, to a diverse set of language and decision making tasks and demonstrate its effectiveness over state-of-the-art baselines, as well as improved human interpretability and trustworthiness over methods without reasoning or acting components. Concretely, on question answering (HotpotQA) and fact verification (Fever), ReAct overcomes issues of hallucination and error propagation prevalent in chain-of-thought reasoning by interacting with a simple Wikipedia API, and generates human-like task-solving trajectories that are more interpretable than baselines without reasoning traces. On two interactive decision making benchmarks (ALFWorld and WebShop), ReAct outperforms imitation and reinforcement learning methods by an absolute success rate of 34% and 10% respectively, while being prompted with only one or two in-context examples. Project site with code: https://react-lm.github.io","{'model': 'tldr@v2.0.0', 'text': 'The use of LLMs are explored to generate both reasoning traces and task-specific actions in an interleaved manner, allowing for greater synergy between the two: reasoning traces help the model induce, track, and update action plans as well as handle exceptions, while actions allow it to interface with external sources to gather additional information.'}",
-Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor,Or Honovich,"Instruction tuning enables pretrained language models to perform new tasks from inference-time natural language descriptions. These approaches rely on vast amounts of human supervision in the form of crowdsourced datasets or user interactions. In this work, we introduce Unnatural Instructions: a large dataset of creative and diverse instructions, collected with virtually no human labor. We collect 64,000 examples by prompting a language model with three seed examples of instructions and eliciting a fourth. This set is then expanded by prompting the model to rephrase each instruction, creating a total of approximately 240,000 examples of instructions, inputs, and outputs. Experiments show that despite containing a fair amount of noise, training on Unnatural Instructions rivals the effectiveness of training on open-source manually-curated datasets, surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks. These results demonstrate the potential of model-generated data as a cost-effective alternative to crowdsourcing for dataset expansion and diversification.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that despite containing a fair amount of noise, training on Unnatural Instructions rivals the effectiveness of training on open-source manually-curated datasets, surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks.'}",http://arxiv.org/pdf/2212.09689
-Language Models are Multilingual Chain-of-Thought Reasoners,Freda Shi,"We evaluate the reasoning abilities of large language models in multilingual settings. We introduce the Multilingual Grade School Math (MGSM) benchmark, by manually translating 250 grade-school math problems from the GSM8K dataset (Cobbe et al., 2021) into ten typologically diverse languages. We find that the ability to solve MGSM problems via chain-of-thought prompting emerges with increasing model scale, and that models have strikingly strong multilingual reasoning abilities, even in underrepresented languages such as Bengali and Swahili. Finally, we show that the multilingual reasoning abilities of language models extend to other tasks such as commonsense reasoning and word-in-context semantic judgment. The MGSM benchmark is publicly available at https://github.com/google-research/url-nlp.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the ability to solve MGSM problems via chain-of-thought prompting emerges with increasing model scale, and that models have strikingly strong multilingual reasoning abilities, even in underrepresented languages such as Bengali and Swahili.'}",http://arxiv.org/pdf/2210.03057
-Teaching Small Language Models to Reason,Lucie Charlotte Magister,"Chain of thought prompting successfully improves the reasoning capabilities of large language models, achieving state of the art results on a range of datasets. However, these reasoning capabilities only appear to emerge in models with at least tens of billions of parameters. In this paper, we explore the transfer of such reasoning capabilities to smaller models via knowledge distillation, also investigating model and dataset size trade-off. Specifically, we finetune a student model on the chain of thought outputs generated by a larger teacher model. Our experiments show that the proposed method improves task performance across arithmetic, commonsense and symbolic reasoning datasets. For example, the accuracy of T5 XXL on GSM8K improves from 8.11% to 21.99% and 18.42% when finetuned on PaLM 540B and GPT-3 175B generated chains of thought, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This paper finetune a student model on the chain of thought outputs generated by a larger teacher model, and shows that the proposed method improves task performance across arithmetic, commonsense and symbolic reasoning datasets.'}",http://arxiv.org/pdf/2212.08410
-Instruction Induction: From Few Examples to Natural Language Task Descriptions,Or Honovich,"Large language models are able to perform a task by conditioning on a few input-output demonstrations - a paradigm known as in-context learning. We show that language models can explicitly infer an underlying task from a few demonstrations by prompting them to generate a natural language instruction that fits the examples. To explore this ability, we introduce the instruction induction challenge, compile a dataset consisting of 24 tasks, and define a novel evaluation metric based on executing the generated instruction. We discover that, to a large extent, the ability to generate instructions does indeed emerge when using a model that is both large enough and aligned to follow instructions; InstructGPT achieves 65.7% of human performance in our execution-based metric, while the original GPT-3 model reaches only 9.8% of human performance. This surprising result suggests that instruction induction might be a viable learning paradigm in and of itself, where instead of fitting a set of latent continuous parameters to the data, one searches for the best description in the natural language hypothesis space.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that language models can explicitly infer an underlying task from a few demonstrations by prompting them to generate a natural language instruction that fits the examples, and discovers that the ability to generate instructions does indeed emerge when using a model that is both large enough and aligned to follow instructions.'}",https://arxiv.org/pdf/2205.10782
-Weakly Supervised Data Augmentation Through Prompting for Dialogue Understanding,Maximillian Chen,"Dialogue understanding tasks often necessitate abundant annotated data to achieve good performance and that presents challenges in low-resource settings. To alleviate this barrier, we explore few-shot data augmentation for dialogue understanding by prompting large pre-trained language models and present a novel approach that iterates on augmentation quality by applying weakly-supervised filters. We evaluate our methods on the emotion and act classification tasks in DailyDialog and the intent classification task in Facebook Multilingual Task-Oriented Dialogue. Models fine-tuned on our augmented data mixed with few-shot ground truth data are able to approach or surpass existing state-of-the-art performance on both datasets. For DailyDialog specifically, using 10% of the ground truth data we outperform the current state-of-the-art model which uses 100% of the data.","{'model': 'tldr@v2.0.0', 'text': 'This work explores few-shot data augmentation for dialogue understanding by prompting large pre-trained language models and presents a novel approach that iterates on augmentation quality by applying weakly-supervised filters.'}",https://arxiv.org/pdf/2210.14169
-PointCLIP V2: Prompting CLIP and GPT for Powerful 3D Open-world Learning,Xiangyang Zhu,"Large-scale pre-trained models have shown promising open-world performance for both vision and language tasks. However, their transferred capacity on 3D point clouds is still limited and only constrained to the classification task. In this paper, we first collaborate CLIP and GPT to be a unified 3D open-world learner, named as PointCLIP V2, which fully unleashes their potential for zero-shot 3D classification, segmentation, and detection. To better align 3D data with the pre-trained language knowledge, PointCLIP V2 contains two key designs. For the visual end, we prompt CLIP via a shape projection module to generate more realistic depth maps, narrowing the domain gap between projected point clouds with natural images. For the textual end, we prompt the GPT model to generate 3D-specific text as the input of CLIP's textual encoder. Without any training in 3D domains, our approach significantly surpasses PointCLIP by +42.90%, +40.44%, and +28.75% accuracy on three datasets for zero-shot 3D classification. On top of that, V2 can be extended to few-shot 3D classification, zero-shot 3D part segmentation, and 3D object detection in a simple manner, demonstrating our generalization ability for unified 3D open-world learning.","{'model': 'tldr@v2.0.0', 'text': 'This paper first collaborate CLIP and GPT to be a unified 3D open-world learner, named as PointCLIP V2, which fully unleashes their potential for zero-shot 3D classification, segmentation, and detection.'}",
-MILAN: Masked Image Pretraining on Language Assisted Representation,Zejiang Hou,"Self-attention based transformer models have been dominating many computer vision tasks in the past few years. Their superb model qualities heavily depend on the excessively large labeled image datasets. In order to reduce the reliance on large labeled datasets, reconstruction based masked autoencoders are gaining popularity, which learn high quality transferable representations from unlabeled images. For the same purpose, recent weakly supervised image pretraining methods explore language supervision from text captions accompanying the images. In this work, we propose masked image pretraining on language assisted representation, dubbed as MILAN. Instead of predicting raw pixels or low level features, our pretraining objective is to reconstruct the image features with substantial semantic signals that are obtained using caption supervision. Moreover, to accommodate our reconstruction target, we propose a more effective prompting decoder architecture and a semantic aware mask sampling mechanism, which further advance the transfer performance of the pretrained model. Experimental results demonstrate that MILAN delivers higher accuracy than the previous works. When the masked autoencoder is pretrained and finetuned on ImageNet-1K dataset with an input resolution of 224x224, MILAN achieves a top-1 accuracy of 85.4% on ViT-Base, surpassing previous state-of-the-arts by 1%. In the downstream semantic segmentation task, MILAN achieves 52.7 mIoU using ViT-Base on ADE20K dataset, outperforming previous masked pretraining results by 4 points.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes masked image pretraining on language assisted representation, dubbed as MILAN, and proposes a more effective prompting decoder architecture and a semantic aware mask sampling mechanism, which further advance the transfer performance of the pretrained model.'}",http://arxiv.org/pdf/2208.06049
-Errors are Useful Prompts: Instruction Guided Task Programming with Verifier-Assisted Iterative Prompting,M. Skreta,"Generating low-level robot task plans from high-level natural language instructions remains a challenging problem. Although large language models have shown promising results in generating plans, the accuracy of the output remains unverified. Furthermore, the lack of domain-specific language data poses a limitation on the applicability of these models. In this paper, we propose CLAIRIFY, a novel approach that combines automatic iterative prompting with program verification to ensure programs written in data-scarce domain-specific language are syntactically valid and incorporate environment constraints. Our approach provides effective guidance to the language model on generating structured-like task plans by incorporating any errors as feedback, while the verifier ensures the syntactic accuracy of the generated plans. We demonstrate the effectiveness of CLAIRIFY in planning chemistry experiments by achieving state-of-the-art results. We also show that the generated plans can be executed on a real robot by integrating them with a task and motion planner.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes CLAIRIFY, a novel approach that combines automatic iterative prompting with program verification to ensure programs written in data-scarce domain-specific language are syntactically valid and incorporate environment constraints.'}",http://arxiv.org/pdf/2303.14100
-Language Is Not All You Need: Aligning Perception with Language Models,Shaohan Huang,"A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context, and follow instructions, and shows that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodals, and from multimodal to language.'}",http://arxiv.org/pdf/2302.14045
-Improving Factuality and Reasoning in Language Models through Multiagent Debate,Yilun Du,"Large language models (LLMs) have demonstrated remarkable capabilities in language generation, understanding, and few-shot learning in recent years. An extensive body of work has explored how their performance may be further improved through the tools of prompting, ranging from verification, self-consistency, or intermediate scratchpads. In this paper, we present a complementary approach to improve language responses where multiple language model instances propose and debate their individual responses and reasoning processes over multiple rounds to arrive at a common final answer. Our findings indicate that this approach significantly enhances mathematical and strategic reasoning across a number of tasks. We also demonstrate that our approach improves the factual validity of generated content, reducing fallacious answers and hallucinations that contemporary models are prone to. Our approach may be directly applied to existing black-box models and uses identical procedure and prompts for all tasks we investigate. Overall, our findings suggest that such""society of minds""approach has the potential to significantly advance the capabilities of LLMs and pave the way for further breakthroughs in language generation and understanding.","{'model': 'tldr@v2.0.0', 'text': 'A complementary approach to improve language responses where multiple language model instances propose and debate their individual responses and reasoning processes over multiple rounds to arrive at a common final answer is presented, indicating that this approach significantly enhances mathematical and strategic reasoning across a number of tasks.'}",http://arxiv.org/pdf/2305.14325
-ORCA: Interpreting Prompted Language Models via Locating Supporting Data Evidence in the Ocean of Pretraining Data,Xiaochuang Han,"Large pretrained language models have been performing increasingly well in a variety of downstream tasks via prompting. However, it remains unclear from where the model learns the task-specific knowledge, especially in a zero-shot setup. In this work, we want to find evidence of the model's task-specific competence from pretraining and are specifically interested in locating a very small subset of pretraining data that directly supports the model in the task. We call such a subset supporting data evidence and propose a novel method ORCA to effectively identify it, by iteratively using gradient information related to the downstream task. This supporting data evidence offers interesting insights about the prompted language models: in the tasks of sentiment analysis and textual entailment, BERT shows a substantial reliance on BookCorpus, the smaller corpus of BERT's two pretraining corpora, as well as on pretraining examples that mask out synonyms to the task verbalizers.","{'model': 'tldr@v2.0.0', 'text': ""This work aims to find evidence of the model's task-specific competence from pretraining and proposes a novel method ORCA to effectively identify it, by iteratively using gradient information related to the downstream task.""}",https://arxiv.org/pdf/2205.12600
-Can ChatGPT Pass High School Exams on English Language Comprehension?,Joost C. F. de Winter,,,https://link.springer.com/content/pdf/10.1007/s40593-023-00372-z.pdf
-"“The less I type, the better”: How AI Language Models can Enhance or Impede Communication for AAC Users",Stephanie Valencia,"Users of augmentative and alternative communication (AAC) devices sometimes find it difficult to communicate in real time with others due to the time it takes to compose messages. AI technologies such as large language models (LLMs) provide an opportunity to support AAC users by improving the quality and variety of text suggestions. However, these technologies may fundamentally change how users interact with AAC devices as users transition from typing their own phrases to prompting and selecting AI-generated phrases. We conducted a study in which 12 AAC users tested live suggestions from a language model across three usage scenarios: extending short replies, answering biographical questions, and requesting assistance. Our study participants believed that AI-generated phrases could save time, physical and cognitive effort when communicating, but felt it was important that these phrases reflect their own communication style and preferences. This work identifies opportunities and challenges for future AI-enhanced AAC devices.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies opportunities and challenges for future AI-enhanced AAC devices by testing live suggestions from a language model across three usage scenarios: extending short replies, answering biographical questions, and requesting assistance.'}",https://dl.acm.org/doi/pdf/10.1145/3544548.3581560
-Backdoor Attacks for In-Context Learning with Language Models,Nikhil Kandpal,"Because state-of-the-art language models are expensive to train, most practitioners must make use of one of the few publicly available language models or language model APIs. This consolidation of trust increases the potency of backdoor attacks, where an adversary tampers with a machine learning model in order to make it perform some malicious behavior on inputs that contain a predefined backdoor trigger. We show that the in-context learning ability of large language models significantly complicates the question of developing backdoor attacks, as a successful backdoor must work against various prompting strategies and should not affect the model's general purpose capabilities. We design a new attack for eliciting targeted misclassification when language models are prompted to perform a particular target task and demonstrate the feasibility of this attack by backdooring multiple large language models ranging in size from 1.3 billion to 6 billion parameters. Finally we study defenses to mitigate the potential harms of our attack: for example, while in the white-box setting we show that fine-tuning models for as few as 500 steps suffices to remove the backdoor behavior, in the black-box setting we are unable to develop a successful defense that relies on prompt engineering alone.","{'model': 'tldr@v2.0.0', 'text': ""It is shown that the in-context learning ability of large language models significantly complicates the question of developing backdoor attacks, as a successful backdoor must work against various prompting strategies and should not affect the model's general purpose capabilities.""}",https://arxiv.org/pdf/2307.14692
-Language-Guided Music Recommendation for Video via Prompt Analogies,Daniel McKee,"We propose a method to recommend music for an input video while allowing a user to guide music selection with free-form natural language. A key challenge of this problem setting is that existing music video datasets provide the needed (video, music) training pairs, but lack text descriptions of the music. This work addresses this challenge with the following three contributions. First, we propose a text-synthesis approach that relies on an analogy-based prompting procedure to generate natural language music descriptions from a large-scale language model (BLOOM-176B) given pre-trained music tagger outputs and a small number of human text descriptions. Second, we use these synthesized music descriptions to train a new trimodal model, which fuses text and video input representations to query music samples. For training, we introduce a text dropout regularization mechanism which we show is critical to model performance. Our model design allows for the re-trieved music audio to agree with the two input modalities by matching visual style depicted in the video and musical genre, mood, or instrumentation described in the natural language query. Third, to evaluate our approach, we collect a testing dataset for our problem by annotating a subset of 4k clips from the YT8M-Music Video dataset with natural language music descriptions which we make publicly available. We show that our approach can match or exceed the performance of prior methods on video-to-music retrieval while significantly improving retrieval accuracy when using text guidance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to recommend music for an input video while allowing a user to guide music selection with free-form natural language, and shows that this approach can match or exceed the performance of prior methods on video-to-music retrieval while significantly improving retrieval accuracy when using text guidance.'}",https://arxiv.org/pdf/2306.09327
-Language Models Represent Space and Time,Wes Gurnee,"The capabilities of large language models (LLMs) have sparked debate over whether such systems just learn an enormous collection of superficial statistics or a coherent model of the data generating process -- a world model. We find evidence for the latter by analyzing the learned representations of three spatial datasets (world, US, NYC places) and three temporal datasets (historical figures, artworks, news headlines) in the Llama-2 family of models. We discover that LLMs learn linear representations of space and time across multiple scales. These representations are robust to prompting variations and unified across different entity types (e.g. cities and landmarks). In addition, we identify individual ``space neurons'' and ``time neurons'' that reliably encode spatial and temporal coordinates. Our analysis demonstrates that modern LLMs acquire structured knowledge about fundamental dimensions such as space and time, supporting the view that they learn not merely superficial statistics, but literal world models.","{'model': 'tldr@v2.0.0', 'text': 'It is found that modern LLMs acquire structured knowledge about fundamental dimensions such as space and time, supporting the view that they learn not merely superficial statistics, but literal world models.'}",https://arxiv.org/pdf/2310.02207
-Post Hoc Explanations of Language Models Can Improve Language Models,Satyapriya Krishna,"Large Language Models (LLMs) have demonstrated remarkable capabilities in performing complex tasks. Moreover, recent research has shown that incorporating human-annotated rationales (e.g., Chain-of-Thought prompting) during in-context learning can significantly enhance the performance of these models, particularly on tasks that require reasoning capabilities. However, incorporating such rationales poses challenges in terms of scalability as this requires a high degree of human involvement. In this work, we present a novel framework, Amplifying Model Performance by Leveraging In-Context Learning with Post Hoc Explanations (AMPLIFY), which addresses the aforementioned challenges by automating the process of rationale generation. To this end, we leverage post hoc explanation methods which output attribution scores (explanations) capturing the influence of each of the input features on model predictions. More specifically, we construct automated natural language rationales that embed insights from post hoc explanations to provide corrective signals to LLMs. Extensive experimentation with real-world datasets demonstrates that our framework, AMPLIFY, leads to prediction accuracy improvements of about 10-25% over a wide range of tasks, including those where prior approaches which rely on human-annotated rationales such as Chain-of-Thought prompting fall short. Our work makes one of the first attempts at highlighting the potential of post hoc explanations as valuable tools for enhancing the effectiveness of LLMs. Furthermore, we conduct additional empirical analyses and ablation studies to demonstrate the impact of each of the components of AMPLIFY, which, in turn, leads to critical insights for refining in-context learning.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a novel framework, Amplifying Model Performance by Leveraging In-Context Learning with Post Hoc Explanations (AMPLIFY), which addresses the aforementioned challenges by automating the process of rationale generation by leveraging post hoc explanation methods.'}",http://arxiv.org/pdf/2305.11426
-Sparks of Artificial General Intelligence: Early experiments with GPT-4,Sébastien Bubeck,"Artificial intelligence (AI) researchers have been developing and refining large language models (LLMs) that exhibit remarkable capabilities across a variety of domains and tasks, challenging our understanding of learning and cognition. The latest model developed by OpenAI, GPT-4, was trained using an unprecedented scale of compute and data. In this paper, we report on our investigation of an early version of GPT-4, when it was still in active development by OpenAI. We contend that (this early version of) GPT-4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models. We discuss the rising capabilities and implications of these models. We demonstrate that, beyond its mastery of language, GPT-4 can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more, without needing any special prompting. Moreover, in all of these tasks, GPT-4's performance is strikingly close to human-level performance, and often vastly surpasses prior models such as ChatGPT. Given the breadth and depth of GPT-4's capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence (AGI) system. In our exploration of GPT-4, we put special emphasis on discovering its limitations, and we discuss the challenges ahead for advancing towards deeper and more comprehensive versions of AGI, including the possible need for pursuing a new paradigm that moves beyond next-word prediction. We conclude with reflections on societal influences of the recent technological leap and future research directions.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that (this early version of) GPT-4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models, and the rising capabilities and implications of these models are discussed.""}",http://arxiv.org/pdf/2303.12712
-Are Hard Examples also Harder to Explain? A Study with Human and Model-Generated Explanations,Swarnadeep Saha,"Recent work on explainable NLP has shown that few-shot prompting can enable large pre-trained language models (LLMs) to generate grammatical and factual natural language explanations for data labels. In this work, we study the connection between explainability and sample hardness by investigating the following research question – “Are LLMs and humans equally good at explaining data labels for both easy and hard samples?” We answer this question by first collecting human-written explanations in the form of generalizable commonsense rules on the task of Winograd Schema Challenge (Winogrande dataset). We compare these explanations with those generated by GPT-3 while varying the hardness of the test samples as well as the in-context samples. We observe that (1) GPT-3 explanations are as grammatical as human explanations regardless of the hardness of the test samples, (2) for easy examples, GPT-3 generates highly supportive explanations but human explanations are more generalizable, and (3) for hard examples, human explanations are significantly better than GPT-3 explanations both in terms of label-supportiveness and generalizability judgements. We also find that hardness of the in-context examples impacts the quality of GPT-3 explanations. Finally, we show that the supportiveness and generalizability aspects of human explanations are also impacted by sample hardness, although by a much smaller margin than models.","{'model': 'tldr@v2.0.0', 'text': 'This work collects human-written explanations in the form of generalizable commonsense rules and finds that for hard examples, human explanations are significantly better than GPT-3 explanations both in terms of label-supportiveness and generalizability judgements.'}",https://arxiv.org/pdf/2211.07517
-Prefix-Tuning: Optimizing Continuous Prompts for Generation,Xiang Lisa Li,"Fine-tuning is the de facto way of leveraging large pretrained language models for downstream tasks. However, fine-tuning modifies all the language model parameters and therefore necessitates storing a full copy for each task. In this paper, we propose prefix-tuning, a lightweight alternative to fine-tuning for natural language generation tasks, which keeps language model parameters frozen and instead optimizes a sequence of continuous task-specific vectors, which we call the prefix. Prefix-tuning draws inspiration from prompting for language models, allowing subsequent tokens to attend to this prefix as if it were “virtual tokens”. We apply prefix-tuning to GPT-2 for table-to-text generation and to BART for summarization. We show that by learning only 0.1% of the parameters, prefix-tuning obtains comparable performance in the full data setting, outperforms fine-tuning in low-data settings, and extrapolates better to examples with topics that are unseen during training.","{'model': 'tldr@v2.0.0', 'text': 'Prefix-tuning is proposed, a lightweight alternative to fine- Tuning for natural language generation tasks, which keeps language model parameters frozen and instead optimizes a sequence of continuous task-specific vectors, which is called the prefix.'}",https://aclanthology.org/2021.acl-long.353.pdf
-A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT,Ce Zhou,"Pretrained Foundation Models (PFMs) are regarded as the foundation for various downstream tasks with different data modalities. A PFM (e.g., BERT, ChatGPT, and GPT-4) is trained on large-scale data which provides a reasonable parameter initialization for a wide range of downstream applications. BERT learns bidirectional encoder representations from Transformers, which are trained on large datasets as contextual language models. Similarly, the generative pretrained transformer (GPT) method employs Transformers as the feature extractor and is trained using an autoregressive paradigm on large datasets. Recently, ChatGPT shows promising success on large language models, which applies an autoregressive language model with zero shot or few shot prompting. The remarkable achievements of PFM have brought significant breakthroughs to various fields of AI. Numerous studies have proposed different methods, raising the demand for an updated survey. This study provides a comprehensive review of recent research advancements, challenges, and opportunities for PFMs in text, image, graph, as well as other data modalities. The review covers the basic components and existing pretraining methods used in natural language processing, computer vision, and graph learning. Additionally, it explores advanced PFMs used for different data modalities and unified PFMs that consider data quality and quantity. The review also discusses research related to the fundamentals of PFMs, such as model efficiency and compression, security, and privacy. Finally, the study provides key implications, future research directions, challenges, and open problems in the field of PFMs. Overall, this survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence.","{'model': 'tldr@v2.0.0', 'text': 'This survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence.'}",http://arxiv.org/pdf/2302.09419
-ChatGPT Makes Medicine Easy to Swallow: An Exploratory Case Study on Simplified Radiology Reports,Katharina Jeblick,"The release of ChatGPT, a language model capable of generating text that appears human-like and authentic, has gained significant attention beyond the research community. We expect that the convincing performance of ChatGPT incentivizes users to apply it to a variety of downstream tasks, including prompting the model to simplify their own medical reports. To investigate this phenomenon, we conducted an exploratory case study. In a questionnaire, we asked 15 radiologists to assess the quality of radiology reports simplified by ChatGPT. Most radiologists agreed that the simplified reports were factually correct, complete, and not potentially harmful to the patient. Nevertheless, instances of incorrect statements, missed key medical findings, and potentially harmful passages were reported. While further studies are needed, the initial insights of this study indicate a great potential in using large language models like ChatGPT to improve patient-centered care in radiology and other medical domains.","{'model': 'tldr@v2.0.0', 'text': 'The initial insights of this study indicate a great potential in using large language models like ChatGPT to improve patient-centered care in radiology and other medical domains.'}",http://arxiv.org/pdf/2212.14882
-Segment Everything Everywhere All at Once,Xueyan Zou,"In this work, we present SEEM, a promptable and interactive model for segmenting everything everywhere all at once in an image, as shown in Fig.1. In SEEM, we propose a novel decoding mechanism that enables diverse prompting for all types of segmentation tasks, aiming at a universal segmentation interface that behaves like large language models (LLMs). More specifically, SEEM is designed with four desiderata: i) Versatility. We introduce a new visual prompt to unify different spatial queries including points, boxes, scribbles and masks, which can further generalize to a different referring image; ii) Compositionality. We learn a joint visual-semantic space between text and visual prompts, which facilitates the dynamic composition of two prompt types required for various segmentation tasks; iii) Interactivity. We further incorporate learnable memory prompts into the decoder to retain segmentation history through mask-guided cross-attention from decoder to image features; and iv) Semantic-awareness. We use a text encoder to encode text queries and mask labels into the same semantic space for open-vocabulary segmentation. We conduct a comprehensive empirical study to validate the effectiveness of SEEM across diverse segmentation tasks. Notably, our single SEEM model achieves competitive performance across interactive segmentation, generic segmentation, referring segmentation, and video object segmentation on 9 datasets with minimum 1/100 supervision. Furthermore, SEEM showcases a remarkable capacity for generalization to novel prompts or their combinations, rendering it a readily universal image segmentation interface.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel decoding mechanism that enables diverse prompting for all types of segmentation tasks, aiming at a universal segmentation interface that behaves like large language models (LLMs).'}",https://arxiv.org/pdf/2304.06718
-Verify-and-Edit: A Knowledge-Enhanced Chain-of-Thought Framework,Ruochen Zhao,"As large language models (LLMs) have become the norm in NLP, demonstrating good performance in generation and reasoning tasks, one of its most fatal disadvantages is the lack of factual correctness. Generating unfactual texts not only leads to lower performances but also degrades the trust and validity of their applications. Chain-of-Thought (CoT) prompting improves trust and model performance on complex reasoning tasks by generating interpretable reasoning chains, but still suffers from factuality concerns in knowledge-intensive tasks. In this paper, we propose the Verify-and-Edit framework for CoT prompting, which seeks to increase prediction factuality by post-editing reasoning chains according to external knowledge. Building on top of GPT-3, our framework lead to accuracy improvements in multiple open-domain question-answering tasks.","{'model': 'tldr@v2.0.0', 'text': 'The Verify-and-Edit framework for CoT prompting is proposed, which seeks to increase prediction factuality by post-editing reasoning chains according to external knowledge and lead to accuracy improvements in multiple open-domain question-answering tasks.'}",https://arxiv.org/pdf/2305.03268
-Multi-lingual Evaluation of Code Generation Models,Ben Athiwaratkun,"We present new benchmarks on evaluation code generation models: MBXP and Multilingual HumanEval, and MathQA-X. These datasets cover over 10 programming languages and are generated using a scalable conversion framework that transpiles prompts and test cases from the original Python datasets into the corresponding data in the target language. Using these benchmarks, we are able to assess the performance of code generation models in a multi-lingual fashion, and discovered generalization ability of language models on out-of-domain languages, advantages of multi-lingual models over mono-lingual, the ability of few-shot prompting to teach the model new languages, and zero-shot translation abilities even on mono-lingual settings. Furthermore, we use our code generation model to perform large-scale bootstrapping to obtain synthetic canonical solutions in several languages, which can be used for other code-related evaluations such as code insertion, robustness, or summarization tasks. Overall, our benchmarks represents a significant step towards a deeper understanding of language models' code generation abilities. We publicly release our code and datasets at https://github.com/amazon-research/mxeval.","{'model': 'tldr@v2.0.0', 'text': 'This work presents new benchmarks on evaluation code generation models: MBXP and Multilingual HumanEval, and MathQA-X, and discovered generalization ability of language models on out-of-domain languages, advantages of multi-lingUAL models over mono-lingual, and the ability of few-shot prompting to teach the model new languages.'}",http://arxiv.org/pdf/2210.14868
-Fengshenbang 1.0: Being the Foundation of Chinese Cognitive Intelligence,Junjie Wang,"Nowadays, foundation models become one of fundamental infrastructures in artificial intelligence, paving ways to the general intelligence. However, the reality presents two urgent challenges: existing foundation models are dominated by the English-language community; users are often given limited resources and thus cannot always use foundation models. To support the development of the Chinese-language community, we introduce an open-source project, called Fengshenbang, which leads by the research center for Cognitive Computing and Natural Language (CCNL). Our project has comprehensive capabilities, including large pre-trained models, user-friendly APIs, benchmarks, datasets, and others. We wrap all these in three sub-projects: the Fengshenbang Model, the Fengshen Framework, and the Fengshen Benchmark. An open-source roadmap, Fengshenbang, aims to re-evaluate the open-source community of Chinese pre-trained large-scale models, prompting the development of the entire Chinese large-scale model community. We also want to build a user-centered open-source ecosystem to allow individuals to access the desired models to match their computing resources. Furthermore, we invite companies, colleges, and research institutions to collaborate with us to build the large-scale open-source model-based ecosystem. We hope that this project will be the foundation of Chinese cognitive intelligence.","{'model': 'tldr@v2.0.0', 'text': 'An open-source project, called Fengshenbang, which aims to re-evaluate the open- source community of Chinese pre-trained large-scale models, prompting the development of the entire Chinese large- scale model community, and to build a user-centeredopen-source ecosystem to allow individuals to access the desired models to match their computing resources.'}",http://arxiv.org/pdf/2209.02970
-GraphPrompt: Unifying Pre-Training and Downstream Tasks for Graph Neural Networks,Zemin Liu,"Graphs can model complex relationships between objects, enabling a myriad of Web applications such as online page/article classification and social recommendation. While graph neural networks (GNNs) have emerged as a powerful tool for graph representation learning, in an end-to-end supervised setting, their performance heavily relies on a large amount of task-specific supervision. To reduce labeling requirement, the “pre-train, fine-tune” and “pre-train, prompt” paradigms have become increasingly common. In particular, prompting is a popular alternative to fine-tuning in natural language processing, which is designed to narrow the gap between pre-training and downstream objectives in a task-specific manner. However, existing study of prompting on graphs is still limited, lacking a universal treatment to appeal to different downstream tasks. In this paper, we propose GraphPrompt, a novel pre-training and prompting framework on graphs. GraphPrompt not only unifies pre-training and downstream tasks into a common task template, but also employs a learnable prompt to assist a downstream task in locating the most relevant knowledge from the pre-trained model in a task-specific manner. Finally, we conduct extensive experiments on five public datasets to evaluate and analyze GraphPrompt.","{'model': 'tldr@v2.0.0', 'text': 'GraphPrompt is proposed, a novel pre- training and prompting framework on graphs that unifies pre-training and downstream tasks into a common task template, but also employs a learnable prompt to assist a downstream task in locating the most relevant knowledge from the pre-trained model in a task-specific manner.'}",https://dl.acm.org/doi/pdf/10.1145/3543507.3583386
-Symbolic Chain-of-Thought Distillation: Small Models Can Also “Think” Step-by-Step,Liunian Harold Li,"Chain-of-thought prompting (e.g., “Let’s think step-by-ste”) primes large language models to verbalize rationalization for their predictions. While chain-of-thought can lead to dramatic performance gains, benefits appear to emerge only for sufficiently large models (beyond 50B parameters). We show that orders-of-magnitude smaller models (125M—1.3B parameters) can still benefit from chain-of-thought prompting. To achieve this, we introduce Symbolic Chain-of-Thought Distillation (SCoTD), a method to train a smaller student model on rationalizations sampled from a significantly larger teacher model. Experiments across several commonsense benchmarks show that: 1) SCoTD enhances the performance of the student model in both supervised and few-shot settings, and especially for challenge sets; 2) sampling many reasoning chains per instance from the teacher is paramount; and 3) after distillation, student chain-of-thoughts are judged by humans as comparable to the teacher, despite orders of magnitude fewer parameters. We test several hypotheses regarding what properties of chain-of-thought samples are important, e.g., diversity vs. teacher likelihood vs. open-endedness. We release our corpus of chain-of-thought samples and code.","{'model': 'tldr@v2.0.0', 'text': 'Symbolic Chain-of-Thought Distillation (SCoTD) is introduced, a method to train a smaller student model on rationalizations sampled from a significantly larger teacher model that enhances the performance of the student model in both supervised and few-shot settings and especially for challenge sets.'}",http://arxiv.org/pdf/2306.14050
-InferFix: End-to-End Program Repair with LLMs,Ma Jin,"Software development life cycle is profoundly influenced by bugs: their introduction, identification, and eventual resolution account for a significant portion of software cost. This has motivated software engineering researchers and practitioners to propose different approaches for automating the identification and repair of software defects. Large language models have been adapted to the program repair task through few-shot demonstration learning and instruction prompting, treating this as an infilling task. However, these models have only focused on learning general bug-fixing patterns for uncategorized bugs mined from public repositories. In this paper, we propose InferFix: a transformer-based program repair framework paired with a state-of-the-art static analyzer to fix critical security and performance bugs. InferFix combines a Retriever -- transformer encoder model pretrained via contrastive learning objective, which aims at searching for semantically equivalent bugs and corresponding fixes; and a Generator -- a large language model (Codex Cushman) finetuned on supervised bug-fix data with prompts augmented via bug type annotations and semantically similar fixes retrieved from an external non-parametric memory. To train and evaluate our approach, we curated InferredBugs, a novel, metadata-rich dataset of bugs extracted by executing the Infer static analyzer on the change histories of thousands of Java and C# repositories. Our evaluation demonstrates that InferFix outperforms strong LLM baselines, with a top-1 accuracy of 65.6% for generating fixes in C# and 76.8% in Java. We discuss the deployment of InferFix alongside Infer at Microsoft which offers an end-to-end solution for detection, classification, and localization of bugs, as well as fixing and validation of candidate patches, integrated in the continuous integration pipeline to automate the software development workflow.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes InferFix: a transformer-based program repair framework paired with a state-of-the-art static analyzer to fix critical security and performance bugs and discusses the deployment of Infer fix alongside Infer at Microsoft which offers an end-to-end solution for detection, classification, and localization of bugs, as well as fixing and validation of candidate patches.'}",http://arxiv.org/pdf/2303.07263
-Towards Revealing the Mystery behind Chain of Thought: a Theoretical Perspective,Guhao Feng,"Recent studies have discovered that Chain-of-Thought prompting (CoT) can dramatically improve the performance of Large Language Models (LLMs), particularly when dealing with complex tasks involving mathematics or reasoning. Despite the enormous empirical success, the underlying mechanisms behind CoT and how it unlocks the potential of LLMs remain elusive. In this paper, we take a first step towards theoretically answering these questions. Specifically, we examine the expressivity of LLMs with CoT in solving fundamental mathematical and decision-making problems. We start by giving an impossibility result showing that bounded-depth Transformers are unable to directly produce correct answers for basic arithmetic/equation tasks unless the model size grows super-polynomially with respect to the input length. In contrast, we then prove by construction that autoregressive Transformers of constant size suffice to solve both tasks by generating CoT derivations using a commonly-used math language format. Moreover, we show LLMs with CoT are capable of solving a general class of decision-making problems known as Dynamic Programming, thus justifying its power in tackling complex real-world tasks. Finally, extensive experiments on four tasks show that, while Transformers always fail to predict the answers directly, they can consistently learn to generate correct solutions step-by-step given sufficient CoT demonstrations.","{'model': 'tldr@v2.0.0', 'text': 'It is proved by construction that autoregressive Transformers of constant size suffice to solve both tasks by generating CoT derivations using a commonly-used math language format, and LLMs with CoT are capable of solving a general class of decision-making problems known as Dynamic Programming, thus justifying its power in tackling complex real-world tasks.'}",http://arxiv.org/pdf/2305.15408
-ZeroShotDataAug: Generating and Augmenting Training Data with ChatGPT,S. Ubani,"In this paper, we investigate the use of data obtained from prompting a large generative language model, ChatGPT, to generate synthetic training data with the aim of augmenting data in low resource scenarios. We show that with appropriate task-specific ChatGPT prompts, we outperform the most popular existing approaches for such data augmentation. Furthermore, we investigate methodologies for evaluating the similarity of the augmented data generated from ChatGPT with the aim of validating and assessing the quality of the data generated.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with appropriate task-specific ChatGPT prompts, the use of data obtained from prompting a large generative language model,ChatGPT, to generate synthetic training data with the aim of augmenting data in low resource scenarios outperform the most popular existing approaches for such data augmentation.'}",http://arxiv.org/pdf/2304.14334
-Selective In-Context Data Augmentation for Intent Detection using Pointwise V-Information,Yen-Ting Lin,"This work focuses on in-context data augmentation for intent detection. Having found that augmentation via in-context prompting of large pre-trained language models (PLMs) alone does not improve performance, we introduce a novel approach based on PLMs and pointwise V-information (PVI), a metric that can measure the usefulness of a datapoint for training a model. Our method first fine-tunes a PLM on a small seed of training data and then synthesizes new datapoints - utterances that correspond to given intents. It then employs intent-aware filtering, based on PVI, to remove datapoints that are not helpful to the downstream intent classifier. Our method is thus able to leverage the expressive power of large language models to produce diverse training data. Empirical results demonstrate that our method can produce synthetic training data that achieve state-of-the-art performance on three challenging intent detection datasets under few-shot settings (1.28% absolute improvement in 5-shot and 1.18% absolute in 10-shot, on average) and perform on par with the state-of-the-art in full-shot settings (within 0.01% absolute, on average).","{'model': 'tldr@v2.0.0', 'text': 'A novel approach based on PLMs and pointwise V-information (PVI), a metric that can measure the usefulness of a datapoint for training a model, is introduced, able to leverage the expressive power of large language models to produce diverse training data.'}",http://arxiv.org/pdf/2302.05096
-Reading Comprehension Quiz Generation using Generative Pre-trained Transformers,Ramon Dijkstra,". Recent advances in AI have resulted in large pre-trained language models with superior performance on text generation tasks, prompting the question of whether we can use them to generate educationally useful text completions. This holds the potential to generate relevant quizzes for any educational text, greatly complementing current formative and summative tests from education professionals. We explore pre-trained language models for quiz generation on reading comprehension texts and propose EduQuiz , an end-to-end quiz generator based on a GPT-3 model fine-tuned on text-quiz pairs, able to generate a complete multiple-choice question, with the correct and distractor answers. We observed that the majority of generated quizzes is reasonable, and that generation of high-quality distractors is more challenging than question and answer generation. More generally, while it may be too early to replace manually generated tests for summative feedback and grading with automatic quiz generation, EduQuiz already has potential value for formative feedback and to increase engagement during the learning phase by enhancing textbooks with assessments.",,
-Meet Your Favorite Character: Open-domain Chatbot Mimicking Fictional Characters with only a Few Utterances,Seungju Han,"In this paper, we consider mimicking fictional characters as a promising direction for building engaging conversation models. To this end, we present a new practical task where only a few utterances of each fictional character are available to generate responses mimicking them. Furthermore, we propose a new method named Pseudo Dialog Prompting (PDP) that generates responses by leveraging the power of large-scale language models with prompts containing the target character’s utterances. To better reflect the style of the character, PDP builds the prompts in the form of dialog that includes the character’s utterances as dialog history. Since only utterances of the characters are available in the proposed task, PDP matches each utterance with an appropriate pseudo-context from a predefined set of context candidates using a retrieval model. Through human and automatic evaluation, we show that PDP generates responses that better reflect the style of fictional characters than baseline methods.","{'model': 'tldr@v2.0.0', 'text': 'A new method named Pseudo Dialog Prompting (PDP) is proposed that generates responses by leveraging the power of large-scale language models with prompts containing the target character’s utterances to better reflect the style of fictional characters.'}",http://arxiv.org/pdf/2204.10825
-Towards Facet-Driven Generation of Clarifying Questions for Conversational Search,Ivan Sekulic,"Clarifying an underlying user information need is an important aspect of a modern-day IR system. The importance of clarification is even higher in limited-bandwidth scenarios, such as conversational or mobile search, where a user is unable to easily browse through a long list of retrieved results. Thus, asking clarifying questions about user's potentially ambiguous queries arises as one of the main tasks of conversational search. Recent approaches have, while making significant progress in the field, remained limited to selecting a clarifying question from a predefined set or prompting the user with vague or template-based questions. However, with the recent advances in text generation through large-scale language models, an ideal system should generate the next clarifying question. The challenge of generating an appropriate clarifying question is twofold: 1) to produce the question in coherent natural language; 2) to ask a question that is relevant to the user query. In this paper, we propose a model that generates clarifying questions with respect to the user query and query facets. We fine-tune the GPT-2 language model to generate questions related to the query and one of the extracted query facets. Compared to competitive baselines, results show that our proposed method is both natural and useful, as judged by human annotators. Moreover, we discuss the potential theoretical framework this approach would fit in. We release the code for future work and reproducibility purposes.","{'model': 'tldr@v2.0.0', 'text': 'A model that generates clarifying questions with respect to the user query and query facets is proposed using the GPT-2 language model and results show that the proposed method is both natural and useful, as judged by human annotators.'}",
-CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis,Erik Nijkamp,"Program synthesis strives to generate a computer program as a solution to a given problem specification, expressed with input-output examples or natural language descriptions. The prevalence of large language models advances the state-of-the-art for program synthesis, though limited training resources and data impede open access to such models. To democratize this, we train and release a family of large language models up to 16.1B parameters, called CODEGEN, on natural language and programming language data, and open source the training library JAXFORMER. We show the utility of the trained model by demonstrating that it is competitive with the previous state-of-the-art on zero-shot Python code generation on HumanEval. We further investigate the multi-step paradigm for program synthesis, where a single program is factorized into multiple prompts specifying subproblems. To this end, we construct an open benchmark, Multi-Turn Programming Benchmark (MTPB), consisting of 115 diverse problem sets that are factorized into multi-turn prompts. Our analysis on MTPB shows that the same intent provided to CODEGEN in multi-turn fashion significantly improves program synthesis over that provided as a single turn. We make the training library JAXFORMER and model checkpoints available as open source contribution: https://github.com/salesforce/CodeGen.","{'model': 'tldr@v2.0.0', 'text': 'This work trains and releases a family of large language models up to 16.1B parameters, called CODEGEN, on natural language and programming language data, and open source the training library JAXFORMER and model checkpoints, and investigates the multi-step paradigm for program synthesis.'}",
-PromptChainer: Chaining Large Language Model Prompts through Visual Programming,Tongshuang Sherry Wu,"While LLMs have made it possible to rapidly prototype new ML functionalities, many real-world applications involve complex tasks that cannot be easily handled via a single run of an LLM. Recent work has found that chaining multiple LLM runs together (with the output of one step being the input to the next) can help users accomplish these more complex tasks, and in a way that is perceived to be more transparent and controllable. However, it remains unknown what users need when authoring their own LLM chains – a key step to lowering the barriers for non-AI-experts to prototype AI-infused applications. In this work, we explore the LLM chain authoring process. We find from pilot studies that users need support transforming data between steps of a chain, as well as debugging the chain at multiple granularities. To address these needs, we designed PromptChainer, an interactive interface for visually programming chains. Through case studies with four designers and developers, we show that PromptChainer supports building prototypes for a range of applications, and conclude with open questions on scaling chains to even more complex tasks, as well as supporting low-fi chain prototyping.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the LLM chain authoring process, and designs PromptChainer, an interactive interface for visually programming chains that supports building prototypes for a range of applications, as well as supporting low-fi chain prototyping.'}",https://arxiv.org/pdf/2203.06566
-"GrIPS: Gradient-free, Edit-based Instruction Search for Prompting Large Language Models",Archiki Prasad,"Providing natural language instructions in prompts is a useful new paradigm for improving task performance of large language models in a zero-shot setting. Recent work has aimed to improve such prompts via manual rewriting or gradient-based tuning. However, manual rewriting is time-consuming and requires subjective interpretation, while gradient-based tuning can be extremely computationally demanding for large models and may not be feasible for API-based models. In this work, we introduce Gradient-free Instructional Prompt Search (GrIPS), a gradient-free, edit-based search approach for improving task instructions for large language models. GrIPS takes in instructions designed for humans and automatically returns an improved, edited prompt, while allowing for API-based tuning. With InstructGPT models, GrIPS improves the average task performance by up to 4.30 percentage points on eight classification tasks from the Natural Instructions dataset (with similar improvements for OPT, BLOOM, and FLAN-T5). We see improvements for both instruction-only prompts and instruction + k-shot examples prompts. Notably, GrIPS outperforms manual rewriting and purely example-based prompts while controlling for the available compute and data budget. Further, performance of GrIPS is comparable to select gradient-based tuning approaches. Qualitatively, we show our edits can simplify instructions and at times make them incoherent but nonetheless improve accuracy.","{'model': 'tldr@v2.0.0', 'text': 'Gradient-free Instructional Prompt Search (GrIPS) is introduced, a gradient-free, edit-based search approach for improving task instructions for large language models and performance of GrIPS is comparable to select gradient-based tuning approaches.'}",http://arxiv.org/pdf/2203.07281
-AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts,Tongshuang Sherry Wu,"Although large language models (LLMs) have demonstrated impressive potential on simple tasks, their breadth of scope, lack of transparency, and insufficient controllability can make them less effective when assisting humans on more complex tasks. In response, we introduce the concept of Chaining LLM steps together, where the output of one step becomes the input for the next, thus aggregating the gains per step. We first define a set of LLM primitive operations useful for Chain construction, then present an interactive system where users can modify these Chains, along with their intermediate results, in a modular way. In a 20-person user study, we found that Chaining not only improved the quality of task outcomes, but also significantly enhanced system transparency, controllability, and sense of collaboration. Additionally, we saw that users developed new ways of interacting with LLMs through Chains: they leveraged sub-tasks to calibrate model expectations, compared and contrasted alternative strategies by observing parallel downstream effects, and debugged unexpected model outputs by “unit-testing” sub-components of a Chain. In two case studies, we further explore how LLM Chains may be used in future applications.","{'model': 'tldr@v2.0.0', 'text': 'Chaining LLM steps together is introduced, where the output of one step becomes the input for the next, thus aggregating the gains per step, and found that Chaining not only improved the quality of task outcomes, but also significantly enhanced system transparency, controllability, and sense of collaboration.'}",https://dl.acm.org/doi/pdf/10.1145/3491102.3517582
-Craft an Iron Sword: Dynamically Generating Interactive Game Characters by Prompting Large Language Models Tuned on Code,Ryan Volum,"Non-Player Characters (NPCs) significantly enhance the player experience in many games. Historically, players’ interactions with NPCs have tended to be highly scripted, to be limited to natural language responses to be selected by the player, and to not involve dynamic change in game state. In this work, we demonstrate that use of a few example conversational prompts can power a conversational agent to generate both natural language and novel code. This approach can permit development of NPCs with which players can have grounded conversations that are free-form and less repetitive. We demonstrate our approach using OpenAI Codex (GPT-3 finetuned on GitHub), with Minecraft game development as our test bed. We show that with a few example prompts, a Codex-based agent can generate novel code, hold multi-turn conversations and answer questions about structured data. We evaluate this application using experienced gamers in a Minecraft realm and provide analysis of failure cases and suggest possible directions for solutions.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that use of a few example conversational prompts can power a conversational agent to generate both natural language and novel code, which can permit development of NPCs with which players can have grounded conversations that are free-form and less repetitive.'}",https://aclanthology.org/2022.wordplay-1.3.pdf
-AudioPaLM: A Large Language Model That Can Speak and Listen,Paul K. Rubenstein,"We introduce AudioPaLM, a large language model for speech understanding and generation. AudioPaLM fuses text-based and speech-based language models, PaLM-2 [Anil et al., 2023] and AudioLM [Borsos et al., 2022], into a unified multimodal architecture that can process and generate text and speech with applications including speech recognition and speech-to-speech translation. AudioPaLM inherits the capability to preserve paralinguistic information such as speaker identity and intonation from AudioLM and the linguistic knowledge present only in text large language models such as PaLM-2. We demonstrate that initializing AudioPaLM with the weights of a text-only large language model improves speech processing, successfully leveraging the larger quantity of text training data used in pretraining to assist with the speech tasks. The resulting model significantly outperforms existing systems for speech translation tasks and has the ability to perform zero-shot speech-to-text translation for many languages for which input/target language combinations were not seen in training. AudioPaLM also demonstrates features of audio language models, such as transferring a voice across languages based on a short spoken prompt. We release examples of our method at https://google-research.github.io/seanet/audiopalm/examples","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that initializing AudioPaLM with the weights of a text-only large language model improves speech processing, successfully leveraging the larger quantity of text training data used in pretraining to assist with the speech tasks.'}",http://arxiv.org/pdf/2306.12925
-In-context Autoencoder for Context Compression in a Large Language Model,Tao Ge,"We propose the In-context Autoencoder (ICAE), leveraging the power of a large language models (LLM) to compress a long context into short compact memory slots that can be directly conditioned on by the LLM for various purposes. ICAE is first pretrained using both autoencoding and language modeling objectives on massive text data, enabling it to generate memory slots that accurately and comprehensively represent the original context; Then, it is fine-tuned on instruction data for producing desirable responses to various prompts. Experiments demonstrate that our lightweight ICAE, introducing fewer than 1% additional parameters, effectively achieves 4X context compression based on Llama, offering advantages in both improved latency and GPU memory cost during inference, and showing an interesting insight in memorization as well as potential for scalability. These promising results imply a novel perspective on the connection between working memory in cognitive science and representation learning in LLMs, revealing ICAE's significant implications in addressing the long context problem and suggesting further research in LLM context management. Our data, code and model are released at https://github.com/getao/icae.","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that the lightweight ICAE, introducing fewer than 1% additional parameters, effectively achieves 4X context compression based on Llama, offering advantages in both improved latency and GPU memory cost during inference, and showing an interesting insight in memorization as well as potential for scalability.'}",https://arxiv.org/pdf/2307.06945
-Enhancing CLIP with GPT-4: Harnessing Visual Descriptions as Prompts,Mayug Maniparambil,"Contrastive pretrained large Vision-Language Models (VLMs) like CLIP have revolutionized visual representation learning by providing good performance on downstream datasets. VLMs are 0-shot adapted to a downstream dataset by designing prompts that are relevant to the dataset. Such prompt engineering makes use of domain expertise and a validation dataset. Meanwhile, recent developments in generative pretrained models like GPT-4 mean they can be used as advanced internet search tools. They can also be manipulated to provide visual information in any structure. In this work, we show that GPT-4 can be used to generate text that is visually descriptive and how this can be used to adapt CLIP to downstream tasks. We show considerable improvements in 0-shot transfer accuracy on specialized fine-grained datasets like EuroSAT (~7%), DTD (~7%), SUN397 (~4.6%), and CUB (~3.3%) when compared to CLIP's default prompt. We also design a simple few-shot adapter that learns to choose the best possible sentences to construct generalizable classifiers that outperform the recently proposed CoCoOP by ~2% on average and by over 4% on 4 specialized fine-grained datasets. The code, prompts, and auxiliary text dataset is available at https://github.com/mayug/VDT-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that GPT-4 can be used to generate text that is visually descriptive and how this can beused to adapt CLIP to downstream tasks and considerable improvements in 0-shot transfer accuracy on specialized fine-grained datasets are shown.'}",https://arxiv.org/pdf/2307.11661
-Learning to Prompt CLIP for Monocular Depth Estimation: Exploring the Limits of Human Language,Dylan Auty,"CLIP is a significant vision-and-language training framework that has shown surprisingly general understanding of the world, with good performance in many open-ended tasks with little or no additional training. A recent technique has used CLIP to perform 0-shot Monocular Depth Estimation (MDE) by using depth-related prompts, but the use of human language in these prompts presents an unnecessary human bias. In this work, we use continuous learnable tokens in place of discrete human-language words to shed light on the problem. We achieve a significant boost in performance, and find that the learned to-kens do not map neatly to depth-related human language, implying that CLIP’s concept of depth is not succinctly ex-pressible in human language. We posit that this may extend to other CLIP concepts, and believe that this finding will spark further research into both the use and interpretation of non-linguistic tokens in all open-ended scene interpretation tasks. Code is available at https://github.","{'model': 'tldr@v2.0.0', 'text': 'A significant boost in performance is achieved, and it is found that the learned to-kens do not map neatly to depth-related human language, implying that CLIP’s concept of depth is not succinctly ex-pressible in human language.'}",
-Decomposed Soft Prompt Guided Fusion Enhancing for Compositional Zero-Shot Learning,Xiaocheng Lu,"Compositional Zero-Shot Learning (CZSL) aims to recognize novel concepts formed by known states and objects during training. Existing methods either learn the combined state-object representation, challenging the generalization of unseen compositions, or design two classifiers to identify state and object separately from image features, ignoring the intrinsic relationship between them. To jointly eliminate the above issues and construct a more robust CZSL system, we propose a novel framework termed Decomposed Fusion with Soft Prompt (DFSP)11Code is available at: https://github.corn/Forest-art/DFSP.git, by involving vision-language models (VLMs)for unseen composition recognition. Specifically, DFSP constructs a vector combination of learnable soft prompts with state and object to establish the joint representation of them. In addition, a cross-modal decomposed fusion module is designed between the language and image branches, which decomposes state and object among language features instead of image features. Notably, being fused with the decomposed features, the image features can be more expressive for learning the relationship with states and objects, respectively, to improve the response of unseen compositions in the pair space, hence narrowing the domain gap between seen and unseen sets. Experimental results on three challenging benchmarks demonstrate that our approach significantly outperforms other state-of-the-art methods by large margins.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework termed Decomposed Fusion with Soft Prompt (DFSP), which constructs a vector combination of learnable soft prompts with state and object to establish the joint representation of them and significantly outperforms other state-of-the-art methods by large margins.'}",https://arxiv.org/pdf/2211.10681
-Zero-Shot Rumor Detection with Propagation Structure via Prompt Learning,Hongzhan Lin,"The spread of rumors along with breaking events seriously hinders the truth in the era of social media. Previous studies reveal that due to the lack of annotated resources, rumors presented in minority languages are hard to be detected. Furthermore, the unforeseen breaking events not involved in yesterday's news exacerbate the scarcity of data resources. In this work, we propose a novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages. More specifically, we firstly represent rumor circulated on social media as diverse propagation threads, then design a hierarchical prompt encoding mechanism to learn language-agnostic contextual representations for both prompts and rumor data. To further enhance domain adaptation, we model the domain-invariant structural features from the propagation threads, to incorporate structural position representations of influential community response. In addition, a new virtual response augmentation method is used to improve model training. Extensive experiments conducted on three real-world datasets demonstrate that our proposed model achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages, which achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.'}",http://arxiv.org/pdf/2212.01117
-Efficiently Enhancing Zero-Shot Performance of Instruction Following Model via Retrieval of Soft Prompt,Seonghyeon Ye,"Enhancing the zero-shot performance of instruction-following models requires heavy computation, either by scaling the total number of training datasets or the model size. In this work, we explore how retrieval of soft prompts obtained through prompt tuning can efficiently assist hard prompts in zero-shot task generalization. Specifically, we train soft prompt embeddings for each prompt through prompt tuning, store the samples of the training instances mapped with the prompt embeddings, and retrieve the corresponding prompt embedding of the training instance closest to the query instance during inference. While only adding 0.007% additional parameters, retrieval of soft prompt enhances the performance of T0 on unseen tasks by outperforming it on 10 out of 11 datasets as well as improving the mean accuracy of T0 on BIG-bench benchmark by 2.39% points. Also, we report an interesting finding that retrieving source embeddings trained on similar answer choice formats is more important than those on similar task types.","{'model': 'tldr@v2.0.0', 'text': 'This work explores how retrieval of soft prompts obtained through prompt tuning can efficiently assist hard prompts in zero-shot task generalization and reports an interesting finding that retrieving source embeddings trained on similar answer choice formats is more important than those on similar task types.'}",
-Enhancing Class Understanding Via Prompt-Tuning For Zero-Shot Text Classification,Yuhao Dan,"Zero-shot text classification (ZSTC) poses a big challenge due to the lack of labeled data for unseen classes during training. Most studies focus on transferring knowledge from seen classes to unseen classes, which have achieved good performance in most cases. Whereas, it is difficult to transfer knowledge when the classes have semantic gaps or low similarities. In this paper, we propose a prompt-based method, which enhances semantic understanding for each class and learns the matching between texts and classes for better ZSTC. Specifically, we first generate discriminative words for class description with prompt inserting (PIN). Then, a prompt matching (POM) model is learned to determine whether the text can well match the class description. Experiments on three benchmark datasets show the great advantages of our proposed method. In particular, we achieve the state-of-the-art performance on the unseen classes, while maintaining comparable strength with the existing ZSTC approaches regarding to the seen classes.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method, which enhances semantic understanding for each class and learns the matching between texts and classes for better ZSTC, achieves the state-of-the-art performance on the unseen classes, while maintaining comparable strength with the existing ZSTH approaches regarding to the seen classes.'}",
-Phonological Features for 0-shot Multilingual Speech Synthesis,Marlene Staib,"Code-switching---the intra-utterance use of multiple languages---is prevalent across the world. Within text-to-speech (TTS), multilingual models have been found to enable code-switching. By modifying the linguistic input to sequence-to-sequence TTS, we show that code-switching is possible for languages unseen during training, even within monolingual models. We use a small set of phonological features derived from the International Phonetic Alphabet (IPA), such as vowel height and frontness, consonant place and manner. This allows the model topology to stay unchanged for different languages, and enables new, previously unseen feature combinations to be interpreted by the model. We show that this allows us to generate intelligible, code-switched speech in a new language at test time, including the approximation of sounds never seen in training.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a small set of phonological features derived from the International Phonetic Alphabet (IPA) to generate intelligible, code-switched speech in a new language at test time, including the approximation of sounds never seen in training.'}",http://arxiv.org/pdf/2008.04107
-Prompt-based Zero-shot Relation Classification with Semantic Knowledge Augmentation,Jiaying Gong,"In relation classification, recognizing unseen 001 (new) relations for which there are no training 002 instances is a challenging task. We propose 003 a prompt-based model with semantic knowl- 004 edge augmentation (ZS-SKA) to recognize un- 005 seen relations under the zero-shot setting. We 006 present a new word-level sentence translation 007 rule and generate augmented instances with 008 unseen relations from instances with seen rela- 009 tions using that new rule. We design prompts 010 based on an external knowledge graph to inte- 011 grate semantic knowledge information learned 012 from seen relations. Instead of using the actual 013 label sets in the prompt template, we construct 014 weighted virtual label words. We learn the rep- 015 resentations of both seen and unseen relations 016 with augmented instances and prompts. We 017 then calculate the distance between the gen- 018 erated representations using prototypical net- 019 works to predict unseen relations. Extensive 020 experiments conducted on three public datasets 021 show that ZS-SKA outperforms state-of-the-art 022 methods under the zero-shot scenarios. Our 023 experimental results also demonstrate the effec- 024 tiveness and robustness of ZS-SKA. 025","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based model with semantic knowl- 004 edge augmentation (ZS-SKA) to recognize un- seen relations under the zero-shot setting and the results demonstrate the effec- 024 tiveness and robustness of ZS- SKA.'}",
-Matching Exemplar as Next Sentence Prediction (MeNSP): Zero-shot Prompt Learning for Automatic Scoring in Science Education,Xuansheng Wu,"Developing models to automatically score students' written responses to science problems is critical for science education. However, collecting and labeling sufficient student responses for training models is time and cost-consuming. Recent studies suggest that pre-trained language models can be adapted to downstream tasks without fine-tuning with prompts. However, no research has employed such a prompt approach in science education. As student responses are presented with natural language, aligning the scoring procedure as the next sentence prediction task using prompts can skip the costly fine-tuning stage. In this study, we developed a zero-shot approach to automatically score student responses via Matching Exemplars as Next Sentence Prediction (MeNSP). This approach employs no training samples. We first apply MeNSP in scoring three assessment tasks of scientific argumentation and found machine-human scoring agreements, Cohen's Kappa ranges from 0.30 to 0.57, and F1 score ranges from 0.54 to 0.81. To improve the performance, we extend our research to the few-shots setting, either randomly selecting labeled student responses or manually constructing responses to fine-tune the models. We find that one task's performance is improved with more samples, Cohen's Kappa from 0.30 to 0.38, and F1 score from 0.54 to 0.59; for the two others, scoring performance is not improved. We also find that randomly selected few-shots perform better than the human expert-crafted approach. This study suggests that MeNSP can yield referable automatic scoring for student responses while significantly reducing the cost of model training. This method can benefit low-stakes classroom assessment practices in science education. Future research should further explore the applicability of the MeNSP in different types of assessment tasks in science education and improve the model performance.",,http://arxiv.org/pdf/2301.08771
-Controlling Personality Style in Dialogue with Zero-Shot Prompt-Based Learning,Angela Ramirez,"Prompt-based or in-context learning has achieved high zero-shot performance on many natural language generation (NLG) tasks. Here we explore the performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue. We experiment with prompt-based learning on the PERSONAGE restaurant recommendation corpus to generate semantically and stylistically-controlled text for 5 different Big-5 personality types: agreeable, disagreeable, conscientious, unconscientious, and extravert. We test two different classes of discrete prompts to generate utterances for a particular personality style: (1) prompts that demonstrate generating directly from a meaning representation that includes a personality specification; and (2) prompts that rely on first converting the meaning representation to a textual pseudo-reference, and then using the pseudo-reference in a textual style transfer (TST) prompt. In each case, we show that we can vastly improve performance by over-generating outputs and ranking them, testing several ranking functions based on automatic metrics for semantic accuracy, personality-match, and fluency. We also test whether NLG personality demonstrations from the restaurant domain can be used with meaning representations for the video game domain to generate personality stylized utterances about video games. Our findings show that the TST prompts produces the highest semantic accuracy (78.46% for restaurants and 87.6% for video games) and personality accuracy (100% for restaurants and 97% for video games). Our results on transferring personality style to video game utterances are surprisingly good. To our knowledge, there is no previous work testing the application of prompt-based learning to simultaneously controlling both style and semantic accuracy in NLG.","{'model': 'tldr@v2.0.0', 'text': 'The performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue is explored and the results on transferring personality style to video game utterances are surprisingly good.'}",http://arxiv.org/pdf/2302.03848
-Structured prompt interrogation and recursive extraction of semantics (SPIRES): A method for populating knowledge bases using zero-shot learning,J. Caufield,"Creating knowledge bases and ontologies is a time consuming task that relies on a manual curation. AI/NLP approaches can assist expert curators in populating these knowledge bases, but current approaches rely on extensive training data, and are not able to populate arbitrary complex nested knowledge schemas. Here we present Structured Prompt Interrogation and Recursive Extraction of Semantics (SPIRES), a Knowledge Extraction approach that relies on the ability of Large Language Models (LLMs) to perform zero-shot learning (ZSL) and general-purpose query answering from flexible prompts and return information conforming to a specified schema. Given a detailed, user-defined knowledge schema and an input text, SPIRES recursively performs prompt interrogation against GPT-3+ to obtain a set of responses matching the provided schema. SPIRES uses existing ontologies and vocabularies to provide identifiers for all matched elements. We present examples of use of SPIRES in different domains, including extraction of food recipes, multi-species cellular signaling pathways, disease treatments, multi-step drug mechanisms, and chemical to disease causation graphs. Current SPIRES accuracy is comparable to the mid-range of existing Relation Extraction (RE) methods, but has the advantage of easy customization, flexibility, and, crucially, the ability to perform new tasks in the absence of any training data. This method supports a general strategy of leveraging the language interpreting capabilities of LLMs to assemble knowledge bases, assisting manual knowledge curation and acquisition while supporting validation with publicly-available databases and ontologies external to the LLM. SPIRES is available as part of the open source OntoGPT package: https://github.com/ monarch-initiative/ontogpt.","{'model': 'tldr@v2.0.0', 'text': 'Current SPIRES accuracy is comparable to the mid-range of existing Relation Extraction methods, but has the advantage of easy customization, flexibility, and, crucially, the ability to perform new tasks in the absence of any training data.'}",http://arxiv.org/pdf/2304.02711
-Prompt-Guided Zero-Shot Anomaly Action Recognition using Pretrained Deep Skeleton Features,Fumiaki Sato,"This study investigates unsupervised anomaly action recognition, which identifies video-level abnormal-human-behavior events in an unsupervised manner without abnormal samples, and simultaneously addresses three limitations in the conventional skeleton-based approaches: target domain-dependent DNN training, robustness against skeleton errors, and a lack of normal samples. We present a unified, user prompt-guided zero-shot learning framework using a target domain-independent skeleton feature extractor, which is pretrained on a large-scale action recognition dataset. Particularly, during the training phase using normal samples, the method models the distribution of skeleton features of the normal actions while freezing the weights of the DNNs and estimates the anomaly score using this distribution in the inference phase. Additionally, to increase robustness against skeleton errors, we introduce a DNN architecture inspired by a point cloud deep learning paradigm, which sparsely propagates the features between Joints. Furthermore, to prevent the unobserved normal actions from being misidentified as abnormal actions, we incorporate a similarity score between the user prompt embeddings and skeleton features aligned in the common space into the anomaly score, which indirectly supplements normal actions. On two publicly available datasets, we conduct experiments to test the effectiveness of the proposed method with respect to above-mentioned limitations.","{'model': 'tldr@v2.0.0', 'text': 'A unified, user prompt-guided zero-shot learning framework using a target domain-independent skeleton feature extractor, which is pretrained on a large-scale action recognition dataset and incorporates a similarity score between the user prompt embeddings and skeleton features aligned in the common space into the anomaly score, which indirectly supplements normal actions.'}",https://arxiv.org/pdf/2303.15167
-Bayesian Sharpness-Aware Prompt Tuning for Cross-Domain Few-shot Learning,Shuo Fan,"Few-shot learning aims to learn a classifier to recognize novel classes with only few labeled images in each class. Fine-tuning is a promising tool to solve the few-shot learning problem, which pre-trains a large-scale model on source domains and then adapts it to target domains. However, existing methods have poor generalization when encountering the domain-shifting problem in the cross-domain scenario. Inspired by recent advances on domain generalization and prompt-based tuning methods, this paper proposes Bayesian Sharpness-Aware Prompt Tuning (BSAPT) for the cross-domain few-shot learning task. Instead of learning deterministic prompts like existing methods, our BSAPT learns a weight distribution over prompts to model the uncertainty caused by limited training data and resist overfitting. To improve the generalization ability, our BSAPT seeks the prompts which lie in neighborhoods having uniformly low loss by simultaneously minimizing the training loss value and loss sharpness. Benefiting from deterministic pre-trained training and Bayesian inference, our BSAPT has better generalization ability and less overfitting than existing fine-tuning methods. Extensive experiments on public datasets show that our BSAPT outperforms state-of-the-art methods and achieves new state-of-the-art performance in the cross-domain few-shot learning task.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Bayesian Sharpness-Aware Prompt Tuning (BSAPT) for the cross-domain few-shot learning task, which learns a weight distribution over prompts to model the uncertainty caused by limited training data and resist overfitting.'}",
-Zero-Shot Generative Model Adaptation via Image-Specific Prompt Learning,Jiayi Guo,"Recently, CLIP-guided image synthesis has shown appealing performance on adapting a pre-trained source-domain generator to an unseen target domain. It does not require any target-domain samples but only the textual domain labels. The training is highly efficient, e.g., a few minutes. However, existing methods still have some limitations in the quality of generated images and may suffer from the mode collapse issue. A key reason is that a fixed adaptation direction is applied for all cross-domain image pairs, which leads to identical supervision signals. To address this issue, we propose an Image-specific Prompt Learning (IPL) method, which learns specific prompt vectors for each source-domain image. This produces a more precise adaptation direction for every cross-domain image pair, endowing the target-domain generator with greatly enhanced flexibility. Qualitative and quantitative evaluations on various domains demonstrate that IPL effectively improves the quality and diversity of synthesized images and alleviates the mode collapse. Moreover, IPL is independent of the structure of the generative model, such as generative adversarial networks or diffusion models. Code is available at https://github.com/Picsart-AI-Research/IPL-Zero-Shot-Generative-Model-Adaptation.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative and quantitative evaluations on various domains demonstrate that IPL effectively improves the quality and diversity of synthesized images and alleviates the mode collapse.'}",https://arxiv.org/pdf/2304.03119
-RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction,Yew Ken Chia,"Despite the importance of relation extraction in building and representing knowledge, less research is focused on generalizing to unseen relations types. We introduce the task setting of Zero-Shot Relation Triplet Extraction (ZeroRTE) to encourage further research in low-resource relation extraction methods. Given an input sentence, each extracted triplet consists of the head entity, relation label, and tail entity where the relation label is not seen at the training stage. To solve ZeroRTE, we propose to synthesize relation examples by prompting language models to generate structured texts. Concretely, we unify language model prompts and structured text approaches to design a structured prompt template for generating synthetic relation samples when conditioning on relation label prompts (RelationPrompt). To overcome the limitation for extracting multiple relation triplets in a sentence, we design a novel Triplet Search Decoding method. Experiments on FewRel and Wiki-ZSL datasets show the efficacy of RelationPrompt for the ZeroRTE task and zero-shot relation classification. Our code and data are available at github.com/declare-lab/RelationPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This work unify language model prompts and structured text approaches to design a structured prompt template for generating synthetic relation samples when conditioning on relation label prompts (RelationPrompt), and designs a novel Triplet Search Decoding method.'}",http://arxiv.org/pdf/2203.09101
-Finetune like you pretrain: Improved finetuning of zero-shot vision models,Sachin Goyal,"Finetuning image-text models such as CLIP achieves state-of-the-art accuracies on a variety of benchmarks. However, recent works (Kumar et al., 2022; Wortsman et al., 2021) have shown that even subtle differences in the finetuning process can lead to surprisingly large differences in the final performance, both for in-distribution (ID) and out-of-distribution (OOD) data. In this work, we show that a natural and simple approach of mimicking contrastive pretraining consistently outperforms alternative finetuning approaches. Specifically, we cast downstream class labels as text prompts and continue optimizing the contrastive loss between image embeddings and class-descriptive prompt embeddings (contrastive finetuning). Our method consistently outperforms baselines across 7 distribution shift, 6 transfer learning, and 3 few-shot learning benchmarks. On WILDS-iWILDCam, our proposed approach FLYP outperforms the top of the leaderboard by 2.3% ID and 2.7% OOD, giving the highest reported accuracy. Averaged across 7 OOD datasets (2 WILDS and 5 ImageNet associated shifts), FLYP gives gains of 4.2% OOD over standard finetuning and outperforms current state-of-the-art (LP-FT) by more than 1 % both ID and OOD. Similarly, on 3 few-shot learning benchmarks, FLYP gives gains up to 4.6% over standard finetuning and 4.4% over the state-of-the-art. Thus we establish our proposed method of contrastive finetuning as a simple and intuitive state-of-the-art for supervised finetuning of image-text models like CLIP. Code is available at https://github.com/locuslab/FLYP.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that a natural and simple approach of mimicking contrastive pretraining consistently outperforms alternative finetuning approaches, and establishes the proposed method of contrastive finet tuning as a simple and intuitive state-of-the-art for supervised finetuned of image-text models like CLIP.'}",https://arxiv.org/pdf/2212.00638
-Decoupling Knowledge from Memorization: Retrieval-augmented Prompt Learning,Xiang Chen,"Prompt learning approaches have made waves in natural language processing by inducing better few-shot performance while they still follow a parametric-based learning paradigm; the oblivion and rote memorization problems in learning may encounter unstable generalization issues. Specifically, vanilla prompt learning may struggle to utilize atypical instances by rote during fully-supervised training or overfit shallow patterns with low-shot data. To alleviate such limitations, we develop RetroPrompt with the motivation of decoupling knowledge from memorization to help the model strike a balance between generalization and memorization. In contrast with vanilla prompt learning, RetroPrompt constructs an open-book knowledge-store from training instances and implements a retrieval mechanism during the process of input, training and inference, thus equipping the model with the ability to retrieve related contexts from the training corpus as cues for enhancement. Extensive experiments demonstrate that RetroPrompt can obtain better performance in both few-shot and zero-shot settings. Besides, we further illustrate that our proposed RetroPrompt can yield better generalization abilities with new datasets. Detailed analysis of memorization indeed reveals RetroPrompt can reduce the reliance of language models on memorization; thus, improving generalization for downstream tasks. Code is available in https://github.com/zjunlp/PromptKG/tree/main/research/RetroPrompt.","{'model': 'tldr@v2.0.0', 'text': 'RetroPrompt is developed with the motivation of decoupling knowledge from memorization to help the model strike a balance between generalization and memorization and can reduce the reliance of language models on memorization; thus, improving generalization for downstream tasks.'}",https://arxiv.org/pdf/2205.14704
-Zero-Shot Video Captioning with Evolving Pseudo-Tokens,Yoad Tewel,"We introduce a zero-shot video captioning method that employs two frozen networks: the GPT-2 language model and the CLIP image-text matching model. The matching score is used to steer the language model toward generating a sentence that has a high average matching score to a subset of the video frames. Unlike zero-shot image captioning methods, our work considers the entire sentence at once. This is achieved by optimizing, during the generation process, part of the prompt from scratch, by modifying the representation of all other tokens in the prompt, and by repeating the process iteratively, gradually improving the specificity and comprehensiveness of the generated sentence. Our experiments show that the generated captions are coherent and display a broad range of real-world knowledge. Our code is available at: https://github.com/YoadTew/zero-shot-video-to-text","{'model': 'tldr@v2.0.0', 'text': 'A zero-shot video captioning method that employs two frozen networks: the GPT-2 language model and the CLIP image-text matching model that improves the specificity and comprehensiveness of the generated sentence.'}",http://arxiv.org/pdf/2207.11100
-Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language,Andy Zeng,"Large pretrained (e.g.,""foundation"") models exhibit distinct capabilities depending on the domain of data they are trained on. While these domains are generic, they may only barely overlap. For example, visual-language models (VLMs) are trained on Internet-scale image captions, but large language models (LMs) are further trained on Internet-scale text with no images (e.g., spreadsheets, SAT questions, code). As a result, these models store different forms of commonsense knowledge across different domains. In this work, we show that this diversity is symbiotic, and can be leveraged through Socratic Models (SMs): a modular framework in which multiple pretrained models may be composed zero-shot i.e., via multimodal-informed prompting, to exchange information with each other and capture new multimodal capabilities, without requiring finetuning. With minimal engineering, SMs are not only competitive with state-of-the-art zero-shot image captioning and video-to-text retrieval, but also enable new applications such as (i) answering free-form questions about egocentric video, (ii) engaging in multimodal assistive dialogue with people (e.g., for cooking recipes) by interfacing with external APIs and databases (e.g., web search), and (iii) robot perception and planning.","{'model': 'tldr@v2.0.0', 'text': 'Socratic Models (SMs) are shown to be competitive with state-of-the-art zero-shot image captioning and video-to-text retrieval, and enable new applications such as answering free-form questions about egocentric video, and engaging in multimodal assistive dialogue with people.'}",https://arxiv.org/pdf/2204.00598
-FLEX: Unifying Evaluation for Few-Shot NLP,Jonathan Bragg,"Few-shot NLP research is highly active, yet conducted in disjoint research threads with evaluation suites that lack challenging-yet-realistic testing setups and fail to employ careful experimental design. Consequently, the community does not know which techniques perform best or even if they outperform simple baselines. In response, we formulate the FLEX Principles, a set of requirements and best practices for unified, rigorous, valid, and cost-sensitive few-shot NLP evaluation. These principles include Sample Size Design, a novel approach to benchmark design that optimizes statistical accuracy and precision while keeping evaluation costs manageable. Following the principles, we release the FLEX benchmark, which includes four few-shot transfer settings, zero-shot evaluation, and a public leaderboard that covers diverse NLP tasks. In addition, we present UniFew, a prompt-based model for few-shot learning that unifies pretraining and finetuning prompt formats, eschewing complex machinery of recent prompt-based approaches in adapting downstream task formats to language model pretraining objectives. We demonstrate that despite simplicity, UniFew achieves results competitive with both popular meta-learning and prompt-based approaches.","{'model': 'tldr@v2.0.0', 'text': 'The FLEX Principles are formulated, a set of requirements and best practices for unified, rigorous, valid, and cost-sensitive few-shot NLP evaluation, and UniFew, a prompt-based model for few- shot learning that unifies pretraining and finetuning prompt formats are presented.'}",
-Improving Few-Shot Performance of Language Models via Nearest Neighbor Calibration,Feng Nie,"Pre-trained language models (PLMs) have exhibited remarkable few-shot learning capabilities when provided a few examples in a natural language prompt as demonstrations of test instances, i.e., in-context learning. However, the performance of in-context learning is susceptible to the choice of prompt format, training examples and the ordering of the training examples. In this paper, we propose a novel nearest-neighbor calibration framework for in-context learning to ease this issue. It is inspired by a phenomenon that the in-context learning paradigm produces incorrect labels when inferring training instances, which provides a useful supervised signal to calibrate predictions. Thus, our method directly augments the predictions with a $k$-nearest-neighbor ($k$NN) classifier over a datastore of cached few-shot instance representations obtained by PLMs and their corresponding labels. Then adaptive neighbor selection and feature regularization modules are introduced to make full use of a few support instances to reduce the $k$NN retrieval noise. Experiments on various few-shot text classification tasks demonstrate that our method significantly improves in-context learning, while even achieving comparable performance with state-of-the-art tuning-based approaches in some sentiment analysis tasks.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on various few-shot text classification tasks demonstrate that the proposed nearest-neighbor calibration framework significantly improves in-context learning, while even achieving comparable performance with state-of-the-art tuning-based approaches in some sentiment analysis tasks.'}",https://arxiv.org/pdf/2212.02216
-Few-Shot Fine-Grained Entity Typing with Automatic Label Interpretation and Instance Generation,Jiaxin Huang,"We study the problem of few-shot Fine-grained Entity Typing (FET), where only a few annotated entity mentions with contexts are given for each entity type. Recently, prompt-based tuning has demonstrated superior performance to standard fine-tuning in few-shot scenarios by formulating the entity type classification task as a ''fill-in-the-blank'' problem. This allows effective utilization of the strong language modeling capability of Pre-trained Language Models (PLMs). Despite the success of current prompt-based tuning approaches, two major challenges remain: (1) the verbalizer in prompts is either manually designed or constructed from external knowledge bases, without considering the target corpus and label hierarchy information, and (2) current approaches mainly utilize the representation power of PLMs, but have not explored their generation power acquired through extensive general-domain pre-training. In this work, we propose a novel framework for few-shot FET consisting of two modules: (1) an entity type label interpretation module automatically learns to relate type labels to the vocabulary by jointly leveraging few-shot instances and the label hierarchy, and (2) a type-based contextualized instance generator produces new instances based on given instances to enlarge the training set for better generalization. On three benchmark datasets, our model outperforms existing methods by significant margins.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework for few-shot Fine-grained Entity Typing consisting of an entity type label interpretation module automatically learns to relate type labels to the vocabulary by jointly leveraging few- shot instances and the label hierarchy, and a type-based contextualized instance generator produces new instances based on given instances to enlarge the training set for better generalization.'}",https://arxiv.org/pdf/2206.13746
-Natural Language Inference Prompts for Zero-shot Emotion Classification in Text across Corpora,F. Plaza-Del-Arco,"Within textual emotion classification, the set of relevant labels depends on the domain and application scenario and might not be known at the time of model development. This conflicts with the classical paradigm of supervised learning in which the labels need to be predefined. A solution to obtain a model with a flexible set of labels is to use the paradigm of zero-shot learning as a natural language inference task, which in addition adds the advantage of not needing any labeled training data. This raises the question how to prompt a natural language inference model for zero-shot learning emotion classification. Options for prompt formulations include the emotion name anger alone or the statement “This text expresses anger”. With this paper, we analyze how sensitive a natural language inference-based zero-shot-learning classifier is to such changes to the prompt under consideration of the corpus: How carefully does the prompt need to be selected? We perform experiments on an established set of emotion datasets presenting different language registers according to different sources (tweets, events, blogs) with three natural language inference models and show that indeed the choice of a particular prompt formulation needs to fit to the corpus. We show that this challenge can be tackled with combinations of multiple prompts. Such ensemble is more robust across corpora than individual prompts and shows nearly the same performance as the individual best prompt for a particular corpus.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes how sensitive a natural language inference-based zero-shot-learning classifier is to such changes to the prompt under consideration of the corpus and shows that indeed the choice of a particular prompt formulation needs to fit to the corpus.'}",http://arxiv.org/pdf/2209.06701
-Clinical Prompt Learning with Frozen Language Models,Niall Taylor,"Prompt learning is a new paradigm in the Natural Language Processing (NLP) field which has shown impressive performance on a number of natural language tasks with common benchmarking text datasets in full, few-shot, and zero-shot train-evaluation setups. Recently, it has even been observed that large but frozen pre-trained language models (PLMs) with prompt learning outperform smaller but fine-tuned models. However, as with many recent NLP trends, the performance of even the largest PLMs such as GPT-3 do not perform well on specialized domains (e.g. medical text), and the common practice to achieve State of the Art (SoTA) results still consists of pre-training and fine-tuning the PLMs on downstream tasks. The reliance on fine-tuning large PLMs is problematic in clinical settings where data is often held in non-GPU environments, and more resource efficient methods of training specialized domain models is crucial. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared with more traditional fine-tuning methods. Results are partially in line with the prompt learning literature, with prompt learning able to match or improve on traditional fine-tuning with substantially fewer trainable parameters and requiring less training data. We argue that prompt learning therefore provides lower computational resource costs applicable to clinical settings, that can serve as an alternative to fine-tuning ever increasing in size PLMs. Complementary code to reproduce experiments presented in this work can be found at: https: //github.com/NtaylorOX/Public_Clinical_Prompt","{'model': 'tldr@v2.0.0', 'text': 'It is argued that prompt learning provides lower computational resource costs applicable to clinical settings, that can serve as an alternative to fine-tuning ever increasing in size PLMs.'}",http://arxiv.org/pdf/2205.05535
-Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers,Chengyi Wang,"We introduce a language modeling approach for text to speech synthesis (TTS). Specifically, we train a neural codec language model (called Vall-E) using discrete codes derived from an off-the-shelf neural audio codec model, and regard TTS as a conditional language modeling task rather than continuous signal regression as in previous work. During the pre-training stage, we scale up the TTS training data to 60K hours of English speech which is hundreds of times larger than existing systems. Vall-E emerges in-context learning capabilities and can be used to synthesize high-quality personalized speech with only a 3-second enrolled recording of an unseen speaker as an acoustic prompt. Experiment results show that Vall-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity. In addition, we find Vall-E could preserve the speaker's emotion and acoustic environment of the acoustic prompt in synthesis. See https://aka.ms/valle for demos of our work.","{'model': 'tldr@v2.0.0', 'text': ""Vall-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity, and is found to preserve the speaker's emotion and acoustic environment of the acoustic prompt in synthesis.""}",http://arxiv.org/pdf/2301.02111
-Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents,Wenlong Huang,"Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g.""make breakfast""), to a chosen set of actionable steps (e.g.""open fridge""). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into mid-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models. Website at https://huangwl18.github.io/language-planner","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the possibility of grounding high-level tasks, expressed in natural language, to a chosen set of actionable steps and proposes a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions.'}",
-Few-shot Table-to-text Generation with Prefix-Controlled Generator,Yutao Luo,"Neural table-to-text generation approaches are data-hungry, limiting their adaption for low-resource real-world applications. Previous works mostly resort to Pre-trained Language Models (PLMs) to generate fluent summaries of a table. However, they often contain hallucinated contents due to the uncontrolled nature of PLMs. Moreover, the topological differences between tables and sequences are rarely studied. Last but not least, fine-tuning on PLMs with a handful of instances may lead to over-fitting and catastrophic forgetting. To alleviate these problems, we propose a prompt-based approach, Prefix-Controlled Generator (i.e., PCG), for few-shot table-to-text generation. We prepend a task-specific prefix for a PLM to make the table structure better fit the pre-trained input. In addition, we generate an input-specific prefix to control the factual contents and word order of the generated text. Both automatic and human evaluations on different domains (humans, books and songs) of the Wikibio dataset prove the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based approach, Prefix-Controlled Generator (i.e., PCG), for few-shot table-to-text generation, which prepend a task-specific prefix for a PLM to make the table structure better fit the pre-trained input.'}",http://arxiv.org/pdf/2208.10709
-Self-Guided Noise-Free Data Generation for Efficient Zero-Shot Learning,Jiahui Gao,"There is a rising interest in further exploring the zero-shot learning potential of large pre-trained language models (PLMs). A new paradigm called data-generation-based zero-shot learning has achieved impressive success. In this paradigm, the synthesized data from the PLM acts as the carrier of knowledge, which is used to train a task-specific model with orders of magnitude fewer parameters than the PLM, achieving both higher performance and efficiency than prompt-based zero-shot learning methods on PLMs. The main hurdle of this approach is that the synthesized data from PLM usually contains a significant portion of low-quality samples. Fitting on such data will greatly hamper the performance of the task-specific model, making it unreliable for deployment. Previous methods remedy this issue mainly by filtering synthetic data using heuristic metrics(e.g., output confidence), or refining the data with the help of a human expert, which comes with excessive manual tuning or expensive costs. In this paper, we propose a novel noise-robust re-weighting framework SunGen to automatically construct high-quality data for zero-shot classification problems. Our framework features the ability to learn the sample weights indicating data quality without requiring any human annotation. We theoretically and empirically verify the ability of our method to help construct good-quality synthetic datasets. Notably, SunGen-LSTM yields a 9.8% relative improvement than the baseline on average accuracy across eight different established text classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel noise-robust re-weighting framework SunGen is proposed to automatically construct high-quality data for zero-shot classification problems, which features the ability to learn the sample weights indicating data quality without requiring any human annotation.'}",
-P3 Ranker: Mitigating the Gaps between Pre-training and Ranking Fine-tuning with Prompt-based Learning and Pre-finetuning,Xiaomeng Hu,"Compared to other language tasks, applying pre-trained language models (PLMs) for search ranking often requires more nuances and training signals. In this paper, we identify and study the two mismatches between pre-training and ranking fine-tuning: the training schema gap regarding the differences in training objectives and model architectures, and the task knowledge gap considering the discrepancy between the knowledge needed in ranking and that learned during pre-training. To mitigate these gaps, we propose Pre-trained, Prompt-learned and Pre-finetuned Neural Ranker (P3 Ranker). P3 Ranker leverages prompt-based learning to convert the ranking task into a pre-training like schema and uses pre-finetuning to initialize the model on intermediate supervised tasks. Experiments on MS MARCO and Robust04 show the superior performances of P3 Ranker in few-shot ranking. Analyses reveal that P3 Ranker is able to better accustom to the ranking task through prompt-based learning and retrieve necessary ranking-oriented knowledge gleaned in pre-finetuning, resulting in data-efficient PLM adaptation. Our code is available at https://github.com/NEUIR/P3Ranker.","{'model': 'tldr@v2.0.0', 'text': 'P3 Ranker is proposed, which is able to better accustom to the ranking task through prompt-based learning and retrieve necessary ranking-oriented knowledge gleaned in pre-finetuning, resulting in data-efficient PLM adaptation.'}",https://dl.acm.org/doi/pdf/10.1145/3477495.3531786
-Prompt Tuning with Soft Context Sharing for Vision-Language Models,Kun Ding,"Vision-language models have recently shown great potential on many computer vision tasks. Meanwhile, prior work demonstrates prompt tuning designed for vision-language models could acquire superior performance on few-shot image recognition compared to linear probe, a strong baseline. In real-world applications, many few-shot tasks are correlated, particularly in a specialized area. However, such information is ignored by previous work. Inspired by the fact that modeling task relationships by multi-task learning can usually boost performance, we propose a novel method SoftCPT (Soft Context Sharing for Prompt Tuning) to fine-tune pre-trained vision-language models on multiple target few-shot tasks, simultaneously. Specifically, we design a task-shared meta network to generate prompt vector for each task using pre-defined task name together with a learnable meta prompt as input. As such, the prompt vectors of all tasks will be shared in a soft manner. The parameters of this shared meta network as well as the meta prompt vector are tuned on the joint training set of all target tasks. Extensive experiments on three multi-task few-shot datasets show that SoftCPT outperforms the representative single-task prompt tuning method CoOp [78] by a large margin, implying the effectiveness of multi-task learning in vision-language prompt tuning. The source code and data will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method SoftCPT (Soft Context Sharing for Prompt Tuning) to fine-tune pre-trained vision-language models on multiple target few-shot tasks, simultaneously, and designs a task-shared meta network to generate prompt vector for each task using pre-defined task name together with a learnable meta prompt as input.'}",http://arxiv.org/pdf/2208.13474
-PartSLIP: Low-Shot Part Segmentation for 3D Point Clouds via Pretrained Image-Language Models,Minghua Liu,"Generalizable 3D part segmentation is important but challenging in vision and robotics. Training deep models via conventional supervised methods requires large-scale 3D datasets with fine-grained part annotations, which are costly to collect. This paper explores an alternative way for low-shot part segmentation of 3D point clouds by leveraging a pretrained image-language model, GLIP. which achieves superior performance on open-vocabulary 2D detection. We transfer the rich knowledge from 2D to 3D through GLIP-based part detection on point cloud rendering and a novel 2D-to-3D label lifting algorithm. We also utilize multi-view 3D priors and few-shot prompt tuning to boost performance significantly. Extensive evaluation on PartNet and PartNet-Mobility datasets shows that our method enables excellent zero-shot 3D part segmentation. Our few-shot version not only outperforms existing few-shot approaches by a large margin but also achieves highly competitive results compared to the fully supervised counterpart. Furthermore, we demonstrate that our method can be directly applied to iPhone-scanned point clouds without significant domain gaps.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores an alternative way for low-shot part segmentation of 3D point clouds by leveraging a pretrained image-language model, GLIP, which achieves superior performance on open-vocabulary 2D detection and a novel 2D-to-3D label lifting algorithm.'}",https://arxiv.org/pdf/2212.01558
-Prompt-Tuning Can Be Much Better Than Fine-Tuning on Cross-lingual Understanding With Multilingual Language Models,Lifu Tu,"Pre-trained multilingual language models show significant performance gains for zero-shot cross-lingual model transfer on a wide range of natural language understanding (NLU) tasks. Previously, for zero-shot cross-lingual evaluation, pre-trained models are only fine-tuned on English data and tested on a variety of target languages. In this paper, we do cross-lingual evaluation on various NLU tasks (sentence classification, sequence labeling, question answering) using prompt-tuning and compare it with fine-tuning. The results show that prompt tuning achieves much better cross-lingual transfer than fine-tuning across datasets, with only 0.1% to 0.3% tuned parameters. Additionally, we demonstrate through the analysis that prompt tuning can have better cross-lingual transferability of representations on downstream tasks with better aligned decision boundaries.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2210.12360
-Zero-shot Image-to-Image Translation,Gaurav Parmar,"Large-scale text-to-image generative models have shown their remarkable ability to synthesize diverse, high-quality images. However, directly applying these models for real image editing remains challenging for two reasons. First, it is hard for users to craft a perfect text prompt depicting every visual detail in the input image. Second, while existing models can introduce desirable changes in certain regions, they often dramatically alter the input content and introduce unexpected changes in unwanted regions. In this work, we introduce pix2pix-zero, an image-to-image translation method that can preserve the original image’s content without manual prompting. We first automatically discover editing directions that reflect desired edits in the text embedding space. To preserve the content structure, we propose cross-attention guidance, which aims to retain the cross-attention maps of the input image throughout the diffusion process. Finally, to enable interactive editing, we distill the diffusion model into a fast conditional GAN. We conduct extensive experiments and show that our method outperforms existing and concurrent works for both real and synthetic image editing. In addition, our method does not need additional training for these edits and can directly use the existing pre-trained text-to-image diffusion model.","{'model': 'tldr@v2.0.0', 'text': None}",https://dl.acm.org/doi/pdf/10.1145/3588432.3591513
-Few-Shot Anaphora Resolution in Scientific Protocols via Mixtures of In-Context Experts,Nghia T. Le,"Anaphora resolution is an important task for information extraction across a range of languages, text genres, and domains, motivating the need for methods that do not require large annotated datasets. In-context learning has emerged as a promising approach, yet there are a number of challenges in applying in-context learning to resolve anaphora. For example, encoding a single in-context demonstration that consists of: an anaphor, a paragraph-length context, and a list of corresponding antecedents, requires conditioning a language model on a long sequence of tokens, limiting the number of demonstrations per prompt. In this paper, we present MICE (Mixtures of In-Context Experts), which we demonstrate is effective for few-shot anaphora resolution in scientific protocols (Tamari et al., 2021). Given only a handful of training examples, MICE combines the predictions of hundreds of in-context experts, yielding a 30% increase in F1 score over a competitive prompt retrieval baseline. Furthermore, we show MICE can be used to train compact student models without sacrificing performance. As far as we are aware, this is the first work to present experimental results demonstrating the effectiveness of in-context learning on the task of few-shot anaphora resolution in scientific protocols.","{'model': 'tldr@v2.0.0', 'text': 'This is the first work to present experimental results demonstrating the effectiveness of in-context learning on the task of few-shot anaphora resolution in scientific protocols, and it is shown MICE can be used to train compact student models without sacrificing performance.'}",https://arxiv.org/pdf/2210.03690
-Exploiting domain-slot related keywords description for Few-Shot Cross-Domain Dialogue State Tracking,QiXiang Gao,"Collecting dialogue data with domain-slot-value labels for dialogue state tracking (DST) could be a costly process. In this paper, we propose a novel framework based on domain-slot related description to tackle the challenge of few-shot cross-domain DST. Specifically, we design an extraction module to extract domain-slot related verbs and nouns in the dialogue. Then, we integrates them into the description, which aims to prompt the model to identify the slot information. Furthermore, we introduce a random sampling strategy to improve the domain generalization ability of the model. We utilize a pre-trained model to encode contexts and description and generates answers with an auto-regressive manner. Experimental results show that our approaches substantially outperform the existing few-shot DST methods on MultiWOZ and gain strong improvements on the slot accuracy comparing to existing slot description methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel framework based on domain-slot related description to tackle the challenge of few-shot cross-domain DST and introduces a random sampling strategy to improve the domain generalization ability of the model.'}",https://aclanthology.org/2022.emnlp-main.157.pdf
-Decorate the Examples: A Simple Method of Prompt Design for Biomedical Relation Extraction,Hui-Syuan Yeh,"Relation extraction is a core problem for natural language processing in the biomedical domain. Recent research on relation extraction showed that prompt-based learning improves the performance on both fine-tuning on full training set and few-shot training. However, less effort has been made on domain-specific tasks where good prompt design can be even harder. In this paper, we investigate prompting for biomedical relation extraction, with experiments on the ChemProt dataset. We present a simple yet effective method to systematically generate comprehensive prompts that reformulate the relation extraction task as a cloze-test task under a simple prompt formulation. In particular, we experiment with different ranking scores for prompt selection. With BioMed-RoBERTa-base, our results show that prompting-based fine-tuning obtains gains by 14.21 F1 over its regular fine-tuning baseline, and 1.14 F1 over SciFive-Large, the current state-of-the-art on ChemProt. Besides, we find prompt-based learning requires fewer training examples to make reasonable predictions. The results demonstrate the potential of our methods in such a domain-specific relation extraction task.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple yet effective method to systematically generate comprehensive prompts that reformulate the relation extraction task as a cloze-test task under a simple prompt formulation, and finds prompt-based learning requires fewer training examples to make reasonable predictions.'}",http://arxiv.org/pdf/2204.10360
-Slot Dependency Modeling for Zero-Shot Cross-Domain Dialogue State Tracking,Qingyue Wang,"Zero-shot learning for Dialogue State Tracking (DST) focuses on generalizing to an unseen domain without the expense of collecting in domain data. However, previous zero-shot DST methods ignore the slot dependencies in a multidomain dialogue, resulting in sub-optimal performances when adapting to unseen domains. In this paper, we utilize slot prompts combination, slot values demonstration, and slot constraint object to model the slot-slot dependencies, slot-value dependency and slot-context dependency respectively. Specifically, each slot prompt consists of a slot-specific prompt and a slot-shared prompt to capture the shared knowledge across different domains. Experimental results show the effectiveness of our proposed method over existing state-of-art generation methods under zero-shot/few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show the effectiveness of the proposed method over existing state-of-art generation methods under zero-shot/few-shot settings.'}",
-Pre-trained Language Models Can be Fully Zero-Shot Learners,Xuandong Zhao,"How can we extend a pre-trained model to many language understanding tasks, without labeled or additional unlabeled data? Pre-trained language models (PLMs) have been effective for a wide range of NLP tasks. However, existing approaches either require fine-tuning on downstream labeled datasets or manually constructing proper prompts. In this paper, we propose nonparametric prompting PLM (NPPrompt) for fully zero-shot language understanding. Unlike previous methods, NPPrompt uses only pre-trained language models and does not require any labeled data or additional raw corpus for further fine-tuning, nor does it rely on humans to construct a comprehensive set of prompt label words. We evaluate NPPrompt against previous major few-shot and zero-shot learning methods on diverse NLP tasks: including text classification, text entailment, similar text retrieval, paraphrasing, and multiple-choice question answering. Experimental results demonstrate that our NPPrompt outperforms the previous best fully zero-shot method by big margins, with absolute gains of 12.8% in accuracy on text classification and 15.6% on the GLUE benchmark. Our source code is available at https://anonymous.4open. science/r/NPPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes nonparametric prompting PLM (NPPrompt) for fully zero-shot language understanding, which uses only pre-trained language models and does not require any labeled data or additional raw corpus for further fine-tuning, nor does it rely on humans to construct a comprehensive set of prompt label words.'}",http://arxiv.org/pdf/2212.06950
-Multitask Prompted Training Enables Zero-Shot Task Generalization,Victor Sanh,"Large language models have recently been shown to attain reasonable zero-shot generalization on a diverse set of tasks (Brown et al., 2020). It has been hypothesized that this is a consequence of implicit multitask learning in language models' pretraining (Radford et al., 2019). Can zero-shot generalization instead be directly induced by explicit multitask learning? To test this question at scale, we develop a system for easily mapping any natural language tasks into a human-readable prompted form. We convert a large set of supervised datasets, each with multiple prompts with diverse wording. These prompted datasets allow for benchmarking the ability of a model to perform completely held-out tasks. We fine-tune a pretrained encoder-decoder model (Raffel et al., 2020; Lester et al., 2021) on this multitask mixture covering a wide variety of tasks. The model attains strong zero-shot performance on several standard datasets, often outperforming models up to 16x its size. Further, our approach attains strong performance on a subset of tasks from the BIG-bench benchmark, outperforming models up to 6x its size. All trained models are available at https://github.com/bigscience-workshop/t-zero and all prompts are available at https://github.com/bigscience-workshop/promptsource.","{'model': 'tldr@v2.0.0', 'text': 'A system for easily mapping any natural language tasks into a human-readable prompted form and fine-tune a pretrained encoder-decoder model on this multitask mixture covering a wide variety of tasks.'}",
-TeSS: Zero-Shot Classification via Textual Similarity Comparison with Prompting using Sentence Encoder,Jimin Hong,"We introduce TeSS ( Te xt S imilarity Comparison using S entence Encoder), a framework for zero-shot classification where the assigned label is determined by the embedding similarity between the input text and each candidate label prompt. We leverage representations from sentence encoders optimized to locate semantically similar samples closer to each other in embedding space during pretraining. The label prompt embeddings serve as prototypes of their corresponding class clus-ters. Furthermore, to compensate for the po-tentially poorly descriptive labels in their original format, we retrieve semantically similar sentences from external corpora and additionally use them with the original label prompt (TeSS-R). TeSS outperforms strong baselines on various closed-set and open-set classification datasets under zero-shot setting, with further gains when combined with label prompt diversification through retrieval. These results are robustly attained to verbalizer variations, an ancillary benefit of using a bi-encoder. Al-together, our method serves as a reliable baseline for zero-shot classification and a simple interface to assess the quality of sentence encoders 1 .","{'model': 'tldr@v2.0.0', 'text': 'TeSS is introduced, a framework for zero-shot classification where the assigned label is determined by the embedding similarity between the input text and each candidate label prompt, and a simple interface to assess the quality of sentence encoders.'}",http://arxiv.org/pdf/2212.10391
-Zero-Shot Program Representation Learning,Nan Cui,"Learning program representations has been the core prerequisite of code intelligence tasks (e.g., code search and code clone detection). The state-of-the-art pre-trained models such as CodeBERT require the availability of large-scale code corpora. However, gathering training samples can be costly and infeasible for domain-specific languages such as Solidity for smart contracts. In this paper, we propose Zecoler, a zero-shot learning approach for code representations. Zecoler is built upon a pre-trained programming language model. In order to elicit knowledge from the pre-trained models efficiently, Zecoler casts the downstream tasks to the same form of pre-training tasks by inserting trainable prompts into the original input. Then, it employs the prompt learning technique to optimize the pre-trained model by merely adjusting the original input. This enables the representation model to efficiently fit the scarce task-specific data while reusing pre-trained knowledge. We evaluate Zecoler in three code intelligence tasks in two programming languages that have no training samples, namely, Solidity and Go, with model trained in corpora of common languages such as Java. Experimental results show that our approach significantly outperforms baseline models in both zero-shot and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Zecoler is a zero-shot learning approach for code representations built upon a pre-trained programming language model that significantly outperforms baseline models in both zero- shot and few-shot settings.'}",https://dl.acm.org/doi/pdf/10.1145/3524610.3527888
-What Makes Pre-trained Language Models Better Zero-shot Learners?,Jinghui Lu,"Current methods for prompt learning in zero-shot scenarios widely rely on a development set with sufficient human-annotated data to select the best-performing prompt template a posteriori. This is not ideal because in a real-world zero-shot scenario of practical relevance, no labelled data is available. Thus, we propose a simple yet effective method for screening reasonable prompt templates in zero-shot text classification: Perplexity Selection (Perplection). We hypothesize that language discrepancy can be used to measure the efficacy of prompt templates, and thereby develop a substantiated perplexity-based scheme allowing for forecasting the performance of prompt templates in advance. Experiments show that our method leads to improved prediction performance in a realistic zero-shot setting, eliminating the need for any labelled examples.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple yet effective method for screening reasonable prompt templates in zero-shot text classification: Perplexity Selection (Perplection), and hypothesizes that language discrepancy can be used to measure the efficacy of prompt templates, and thereby develop a substantiated perplexity-based scheme allowing for forecasting the performance of Prompt templates in advance.'}",https://aclanthology.org/2023.acl-long.128.pdf
-QueryForm: A Simple Zero-shot Form Entity Query Framework,Zifeng Wang,"Zero-shot transfer learning for document understanding is a crucial yet under-investigated scenario to help reduce the high cost involved in annotating document entities. We present a novel query-based framework, QueryForm, that extracts entity values from form-like documents in a zero-shot fashion. QueryForm contains a dual prompting mechanism that composes both the document schema and a specific entity type into a query, which is used to prompt a Transformer model to perform a single entity extraction task. Furthermore, we propose to leverage large-scale query-entity pairs generated from form-like webpages with weak HTML annotations to pre-train QueryForm. By unifying pre-training and fine-tuning into the same query-based framework, QueryForm enables models to learn from structured documents containing various entities and layouts, leading to better generalization to target document types without the need for target-specific training data. QueryForm sets new state-of-the-art average F1 score on both the XFUND (+4.6%~10.1%) and the Payment (+3.2%~9.5%) zero-shot benchmark, with a smaller model size and no additional image input.","{'model': 'tldr@v2.0.0', 'text': 'By unifying pre-training and fine-tuning into the same query-based framework, QueryForm enables models to learn from structured documents containing various entities and layouts, leading to better generalization to target document types without the need for target-specific training data.'}",http://arxiv.org/pdf/2211.07730
-What Language Model Architecture and Pretraining Objective Work Best for Zero-Shot Generalization?,Thomas Wang,"Large pretrained Transformer language models have been shown to exhibit zero-shot generalization, i.e. they can perform a wide variety of tasks that they were not explicitly trained on. However, the architectures and pretraining objectives used across state-of-the-art models differ significantly, and there has been limited systematic comparison of these factors. In this work, we present a large-scale evaluation of modeling choices and their impact on zero-shot generalization. In particular, we focus on text-to-text models and experiment with three model architectures (causal/non-causal decoder-only and encoder-decoder), trained with two different pretraining objectives (autoregressive and masked language modeling), and evaluated with and without multitask prompted finetuning. We train models with over 5 billion parameters for more than 170 billion tokens, thereby increasing the likelihood that our conclusions will transfer to even larger scales. Our experiments show that causal decoder-only models trained on an autoregressive language modeling objective exhibit the strongest zero-shot generalization after purely unsupervised pretraining. However, models with non-causal visibility on their input trained with a masked language modeling objective followed by multitask finetuning perform the best among our experiments. We therefore consider the adaptation of pretrained models across architectures and objectives. We find that pretrained non-causal decoder models can be adapted into performant generative causal decoder models, using autoregressive language modeling as a downstream task. Furthermore, we find that pretrained causal decoder models can be efficiently adapted into non-causal decoder models, ultimately achieving competitive performance after multitask finetuning. Code and checkpoints are available at https://github.com/bigscience-workshop/architecture-objective.","{'model': 'tldr@v2.0.0', 'text': 'A large-scale evaluation of modeling choices and their impact on zero-shot generalization finds that pretrained non-causal decoder models can be adapted into performant generative causal decoding models, using autoregressive language modeling as a downstream task.'}",http://arxiv.org/pdf/2204.05832
-Towards Zero-Shot and Few-Shot Table Question Answering using GPT-3,Pragya Srivastava,"We present very early results on using GPT-3 to perform question answering on tabular data. We find that stock pre-trained GPT-3 is able to zero-shot learn the table structure from a serialized JSON array-of-arrays representation, and able to answer lookup queries and simple comparison questions in natural language without any fine-tuning. We further find that simple prompt engineering to include few-shot static Q&A examples significantly improves accuracy. Lastly, we find that intermixing passage text improves accuracy even further on heterogeneous data. We apply our approach on a novel dataset of simple tables in newspaper infographics with promising results. Overall, we find much cause for optimism in this basic approach.","{'model': 'tldr@v2.0.0', 'text': 'It is found that stock pre-trained GPT-3 is able to zero-shot learn the table structure from a serialized JSON array-of-arrays representation, and able to answer lookup queries and simple comparison questions in natural language without any fine-tuning.'}",https://arxiv.org/pdf/2210.17284
-Mitigating Word Bias in Zero-shot Prompt-based Classifiers,Adian Liusie,"Prompt-based classifiers are an attractive approach for zero-shot classification. However, the precise choice of the prompt template and label words can largely influence performance, with semantically equivalent settings often showing notable performance difference. This discrepancy can be partly attributed to word biases, where the classifier may be biased towards classes. To address this problem, it is possible to optimise classification thresholds on a labelled data set, however, this mitigates some of the advantages of prompt-based classifiers. This paper instead approaches this problem by examining the expected marginal probabilities of the classes. Here, probabilities are reweighted to have a uniform prior over classes, in an unsupervised fashion. Further, we draw a theoretical connection between the class priors and the language models' word prior, and offer the ability to set a threshold in a zero-resource fashion. We show that matching class priors correlates strongly with the oracle upper bound performance and demonstrate large consistent performance gains for prompt settings over a range of NLP tasks.","{'model': 'tldr@v2.0.0', 'text': ""This paper draws a theoretical connection between the class priors and the language models' word prior, and offers the ability to set a threshold in a zero-resource fashion, and demonstrates large consistent performance gains for prompt settings over a range of NLP tasks.""}",https://arxiv.org/pdf/2309.04992
-Generating Variable Explanations via Zero-shot Prompt Learning,Chong Wang,"As basic elements in program, variables convey essential information that is critical for program comprehension and maintenance. However, understanding the meanings of variables in program is not always easy for developers, since poor-quality variable names are prevalent while such variable are less informative for program comprehension. Therefore, in this paper, we target at generating concise natural language explanations for variables to facilitate program comprehension. In particular, there are two challenges in variable explanation generation, including the lack of training data and the association with complex code contexts around the variable. To address these issues, we propose a novel approach ZeroVar,which leverages code pre-trained models and zero-shot prompt learning to generate explanations for the variable based on its code context. ZeroVarcontains two stages: (i) a pre-training stage that continually pre-trains a base model (i.e., CodeT5) to recover the randomly-masked parameter descriptions in method docstrings; and (ii) a zero-shot prompt learning stage that leverages the pre-trained model to generate explanations for a given variable via the prompt constructed with the variable and its belonging method context. We then extensively evaluate the quality and usefulness of the variable explanations generated by ZeroVar.We construct an evaluation dataset of 773 variables and their reference explanations. Our results show that ZeroVarcan generate higher-quality explanations than baselines, not only on automated metrics such as BLEU and ROUGE, but also on human metrics such as correctness, completeness, and conciseness. Moreover, we further assess the usefulness of ZeroVAR-generated explanations on two downstream tasks related to variable naming quality, i.e., abbreviation expansion and spelling correction. For abbreviation expansion, the generated variable explanations can help improve the present rate (+13.1%), precision (+3.6%), and recall (+10.0%) of the state-of-the-art abbreviation explanation approach. For spelling correction, by using the generated explanations we can achieve higher hit@1 (+162.9(%) and hit@3 (+49.6%) than the recent variable representation learning approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach ZeroVar, which leverages code pre-trained models and zero-shot prompt learning to generate explanations for the variable based on its code context, and assess the quality and usefulness of the variable explanations generated by ZeroVar.'}",
-Prompt Gating: A Parameter Efficient Tuning Method for Zero-Shot Multi-Source Translation,Xuancheng Huang,"Multi-source translation (MST), which typi-cally receives multiple source sentences of the same meaning in different languages, has been shown superior to single-source translation. As the quantity of multi-source parallel data is limited, taking full advantage of single-source data and limited multi-source data to make models perform well when receiving as many as possible sources remains a challenge. Unlike previous work mostly devoted to supervised scenarios, we focus on zero-shot MST: expecting models to be able to process unseen combinations of multiple sources, e.g., unseen language combinations, during inference. We propose a simple yet effective parameter efficient method, named Prompt Gating, which appends prompts to the model inputs and attaches gates on the extended hidden states for each encoder layer. It shows strong zero-shot transferability (+9.0 BLEU points maximally) and remarkable compositionality (+15.6 BLEU points maximally) on MST, and also shows its superiorities over baselines on lexically constrained translation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple yet effective parameter efficient method, named Prompt Gating, which appends prompts to the model inputs and attaches gates on the extended hidden states for each encoder layer on MST, which shows strong zero-shot transferability and remarkable compositionality.'}",http://arxiv.org/pdf/2212.09387
-PEINet: Joint Prompt and Evidence Inference Network via Language Family Policy for Zero-Shot Multilingual Fact Checking,Xiaoyu Li,"Zero-shot multilingual fact-checking, which aims to discover and infer subtle clues from the retrieved relevant evidence to verify the given claim in cross-language and cross-domain scenarios, is crucial for optimizing a free, trusted, wholesome global network environment. Previous works have made enlightening and practical explorations in claim verification, while the zero-shot multilingual task faces new challenging gap issues: neglecting authenticity-dependent learning between multilingual claims, lacking heuristic checking, and a bottleneck of insufficient evidence. To alleviate these gaps, a novel Joint Prompt and Evidence Inference Network (PEINet) is proposed to verify the multilingual claim according to the human fact-checking cognitive paradigm. In detail, firstly, we leverage the language family encoding mechanism to strengthen knowledge transfer among multi-language claims. Then, the prompt turning module is designed to infer the falsity of the fact, and further, sufficient fine-grained evidence is extracted and aggregated based on a recursive graph attention network to verify the claim again. Finally, we build a unified inference framework via multi-task learning for final fact verification. The newly achieved state-of-the-art performance on the released challenging benchmark dataset that includes not only an out-of-domain test, but also a zero-shot test, proves the effectiveness of our framework, and further analysis demonstrates the superiority of our PEINet in multilingual claim verification and inference, especially in the zero-shot scenario.","{'model': 'tldr@v2.0.0', 'text': 'A novel Joint Prompt and Evidence Inference Network (PEINet) is proposed to verify the multilingual claim according to the human fact-checking cognitive paradigm and builds a unified inference framework via multi-task learning for final fact verification.'}",https://www.mdpi.com/2076-3417/12/19/9688/pdf?version=1664345340
-An Exploration of Prompt-Based Zero-Shot Relation Extraction Method,Zhao Jun,,"{'model': 'tldr@v2.0.0', 'text': 'This work views zero-shot relation extraction as a semantic matching task optimized by prompt-tuning, which still maintains superior generalization performance when the labeled data of predefined relations are extremely scarce, and model the semantic interaction between relational instances and their descriptions directly during encoding.'}",
-Prompt-guided Scene Generation for 3D Zero-Shot Learning,Majid Nasiri,"Zero-shot learning on 3D point cloud data is a related underexplored problem compared to its 2D image counterpart. 3D data brings new challenges for ZSL due to the unavailability of robust pre-trained feature extraction models. To address this problem, we propose a prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex interplay of seen and unseen objects. First, we merge point clouds of two 3D models in certain ways described by a prompt. The prompt acts like the annotation describing each 3D scene. Later, we perform contrastive learning to train our proposed architecture in an end-to-end manner. We argue that 3D scenes can relate objects more efficiently than single objects because popular language models (like BERT) can achieve high performance when objects appear in a context. Our proposed prompt-guided scene generation method encapsulates data augmentation and prompt-based annotation/captioning to improve 3D ZSL performance. We have achieved state-of-the-art ZSL and generalized ZSL performance on synthetic (ModelNet40, ModelNet10) and real-scanned (ScanOjbectNN) 3D object datasets.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex interplay of seen and unseen objects.'}",https://arxiv.org/pdf/2209.14690
-A Topic-based Prompt Learning Method for Zero-Shot Stance Detection,Ruifeng Xu,"Zero-shot stance detection (ZSSD) aims to detecting the stance of previously unseen targets during the inference stage. It is generally believed that the stance expression in a sentence is closely related to the stance target and topics discussed.We divide stance expressions of speakers into two categories: target-invariant and target-specific categories. Target-invariant stance expressions carry the same stance polarity regardless of the targets they are associated with. On the contrary, target-specific stance expressions only co-occur with certain targets. As such, it is important to distinguish these two types of stance features to boost stance detection ability.In this paper,we develop an effective approach to distinguish the types of target-related stance expressions to better","{'model': 'tldr@v2.0.0', 'text': 'An effective approach to distinguish the types of target-related stance expressions to better zero-shot stance detection ability is developed.'}",
-Prompt-based Zero-shot Relation Extraction with Semantic Knowledge Augmentation,Jiaying Gong,"In relation triplet extraction (RTE), recognizing unseen (new) relations for which there are no training instances is a challenging task. Efforts have been made to recognize unseen relations based on question-answering models or relation descriptions. However, these approaches miss the semantic information about connections between seen and unseen relations. In this paper, We propose a prompt-based model with semantic knowledge augmentation (ZS-SKA) to recognize unseen relations under the zero-shot setting. We present a new word-level analogy-based sentence translation rule and generate augmented instances with unseen relations from instances with seen relations using that new rule. We design prompts with weighted virtual label construction based on an external knowledge graph to integrate semantic knowledge information learned from seen relations. Instead of using the actual label sets in the prompt template, we construct weighted virtual label words. We learn the representations of both seen and unseen relations with augmented instances and prompts. We then calculate the distance between the generated representations using prototypical networks to predict unseen relations. Extensive experiments conducted on three public datasets FewRel, Wiki-ZSL, and NYT, show that ZS-SKA outperforms state-of-the-art methods under the zero-shot scenarios. Our experimental results also demonstrate the effectiveness and robustness of ZS-SKA.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new word-level analogy-based sentence translation rule and generates augmented instances with unseen relations from instances with seen relations using that new rule, and designs prompts with weighted virtual label construction based on an external knowledge graph to integrate semantic knowledge information learned from seen relations.'}",
-From Visual Prompt Learning to Zero-Shot Transfer: Mapping Is All You Need,Ziqing Yang,"Visual prompt learning, as a newly emerged technique, leverages the knowledge learned by a large-scale pre-trained model and adapts it to downstream tasks through the usage of prompts. While previous research has focused on designing effective prompts, in this work, we argue that compared to prompt design, a good mapping strategy matters more. In this sense, we propose SeMap, a more effective mapping using the semantic alignment between the pre-trained model's knowledge and the downstream task. Our experimental results show that SeMap can largely boost the performance of visual prompt learning. Moreover, our experiments show that SeMap is capable of achieving competitive zero-shot transfer, indicating that it can perform the downstream task without any fine-tuning on the corresponding dataset. This demonstrates the potential of our proposed method to be used in a broader range of applications where the zero-shot transfer is desired. Results suggest that our proposed SeMap could lead to significant advancements in both visual prompt learning and zero-shot transfer. We hope with SeMap, we can help the community move forward to more efficient and lightweight utilization of large vision models.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that SeMap is capable of achieving competitive zero-shot transfer, indicating that it can perform the downstream task without any fine-tuning on the corresponding dataset, demonstrating the potential of the method to be used in a broader range of applications where the zero- shot transfer is desired.'}",http://arxiv.org/pdf/2303.05266
-Layout and Task Aware Instruction Prompt for Zero-shot Document Image Question Answering,Wenjin Wang,"Layout-aware pre-trained models has achieved significant progress on document image question answering. They introduce extra learnable modules into existing language models to capture layout information within document images from text bounding box coordinates obtained by OCR tools. However, extra modules necessitate pre-training on extensive document images. This prevents these methods from directly utilizing off-the-shelf instruction-tuning language foundation models, which have recently shown promising potential in zero-shot learning. Instead, in this paper, we find that instruction-tuning language models like Claude and ChatGPT can understand layout by spaces and line breaks. Based on this observation, we propose the LAyout and Task aware Instruction Prompt (LATIN-Prompt), which consists of layout-aware document content and task-aware instruction. Specifically, the former uses appropriate spaces and line breaks to recover the layout information among text segments obtained by OCR tools, and the latter ensures that generated answers adhere to formatting requirements. Moreover, we propose the LAyout and Task aware Instruction Tuning (LATIN-Tuning) to improve the performance of small instruction-tuning models like Alpaca. Experimental results show that LATIN-Prompt enables zero-shot performance of Claude and ChatGPT to be comparable to the fine-tuning performance of SOTAs on document image question answering, and LATIN-Tuning enhances the zero-shot performance of Alpaca significantly. For example, LATIN-Prompt improves the performance of Claude and ChatGPT on DocVQA by 263% and 20% respectively. LATIN-Tuning improves the performance of Alpaca on DocVQA by 87.7%. Quantitative and qualitative analyses demonstrate the effectiveness of LATIN-Prompt and LATIN-Tuning. We provide the code in supplementary and will release it to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'This paper finds that instruction-tuning language models like Claude and ChatGPT can understand layout by spaces and line breaks, and proposes the LAyout and Task aware Instruction Tuning (LATIN-Tuning), which consists of layout-aware document content and task-aware instruction.'}",https://arxiv.org/pdf/2306.00526
-Navigating Prompt Complexity for Zero-Shot Classification: A Study of Large Language Models in Computational Social Science,Yida Mu,"Instruction-tuned Large Language Models (LLMs) have exhibited impressive language understanding and the capacity to generate responses that follow specific prompts. However, due to the computational demands associated with training these models, their applications often adopt a zero-shot setting. In this paper, we evaluate the zero-shot performance of two publicly accessible LLMs, ChatGPT and OpenAssistant, in the context of six Computational Social Science classification tasks, while also investigating the effects of various prompting strategies. Our experiments investigate the impact of prompt complexity, including the effect of incorporating label definitions into the prompt; use of synonyms for label names; and the influence of integrating past memories during foundation model training. The findings indicate that in a zero-shot setting, current LLMs are unable to match the performance of smaller, fine-tuned baseline transformer models (such as BERT-large). Additionally, we find that different prompting strategies can significantly affect classification accuracy, with variations in accuracy and F1 scores exceeding 10\%.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate that in a zero-shot setting, current LLMs are unable to match the performance of smaller, fine-tuned baseline transformer models (such as BERT-large), and different prompting strategies can significantly affect classification accuracy.'}",https://arxiv.org/pdf/2305.14310
-Prompt position really matters in few-shot and zero-shot NLU tasks,Junyu Mao,"Prompt-based models have made remarkable advancements in the fields of zero-shot and few-shot learning, attracting a lot of attention from researchers. Developing an effective prompt template plays a critical role. However, prior studies have mainly focused on prompt vocabulary selection or embedding initialization with the reserved prompt position fixed. In this empirical study, we conduct the most comprehensive analysis to date of prompt position option for natural language understanding tasks. Our findings quantify the substantial impact prompt position has on model performance. We observe that the prompt position used in prior studies is often sub-optimal for both zero-shot and few-shot settings. These findings suggest prompt position optimisation as an interesting research direction alongside the existing focus on prompt engineering.",,https://arxiv.org/pdf/2305.14493
-Injecting Commonsense Knowledge into Prompt Learning for Zero-Shot Text Classification,Jing Qian,"The combination of pre-training and fine-tuning has become a default solution to Natural Language Processing (NLP) tasks. The emergence of prompt learning breaks such routine, especially in the scenarios of low data resources. Insufficient labelled data or even unseen classes are frequent problems in text classification, equipping Pre-trained Language Models (PLMs) with task-specific prompts helps get rid of the dilemma. However, general PLMs are barely provided with commonsense knowledge. In this work, we propose a KG-driven verbalizer that leverages commonsense Knowledge Graph (KG) to map label words with predefined classes. Specifically, we transform the mapping relationships into semantic relevance in the commonsense-injected embedding space. For zero-shot text classification task, experimental results exhibit the effectiveness of our KG-driven verbalizer on a Twitter dataset for natural disasters (i.e. HumAID) compared with other baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a KG-driven verbalizer that leverages commonsense Knowledge Graph (KG) to map label words with predefined classes and transforms the mapping relationships into semantic relevance in the commonsense-injected embedding space.'}",
-Knowledge-embedded Prompt Learning for Zero-shot Social Media Text Classification,Jingyi Li,"Social media plays an irreplaceable role in shaping the way information is created shared and consumed. While it provides access to a vast amount of data, extracting and analyzing useful insights from complex and dynamic social media data can be challenging. Deep learning models have shown promise in social media analysis tasks, but such models require a massive amount of labelled data which is usually unavailable in real-world settings. Additionally, these models lack common-sense knowledge which can limit their ability to generate comprehensive results. To address these challenges, we propose a knowledge-embedded prompt learning model for zero-shot social media text classification. Our experimental results on four social media datasets demonstrate that our proposed approach outperforms other well-known baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a knowledge-embedded prompt learning model for zero-shot social media text classification that outperforms other well-known baselines and shows superior results on four social media datasets.'}",
-Zero-Shot Continuous Prompt Transfer: Generalizing Task Semantics Across Language Models,Zijun Wu,"Prompt tuning in natural language processing (NLP) has become an increasingly popular method for adapting large language models to specific tasks. However, the transferability of these prompts, especially continuous prompts, between different models remains a challenge. In this work, we propose a zero-shot continuous prompt transfer method, where source prompts are encoded into relative space and the corresponding target prompts are searched for transferring to target models. Experimental results confirm the effectiveness of our method, showing that 'task semantics' in continuous prompts can be generalized across various language models. Moreover, we find that combining 'task semantics' from multiple source models can further enhance the generalizability of transfer.","{'model': 'tldr@v2.0.0', 'text': ""A zero-shot continuous prompt transfer method is proposed, where source prompts are encoded into relative space and the corresponding target prompts are searched for transferring to target models, showing that 'task semantics' in continuous prompts can be generalized across various language models.""}",https://arxiv.org/pdf/2310.01691
-Spteae: A Soft Prompt Transfer Model for Zero-Shot Cross-Lingual Event Argument Extraction,Huipeng Ma,"In zero-shot cross-lingual event argument extraction(EAE) task, a model is typically trained on source language datasets and then applied on task language datasets. There is a trend to regard the zero-shot cross-lingual EAE task as a sequence generation task with manual prompts or discrete prompts. However, there are some problems with these prompts, including using suboptimal prompts and difficult to transfer from source language to target language. To overcome these issues, we propose a method called SPTEAE(A Soft Prompt Transfer model for zero-shot cross-lingual Event Argument Extraction). SPTEAE utilizes a sequence of tunable vectors which are tuned in source language as event type prompts. These source language event type prompts can be transferred as target prompts to perform target EAE task by key-value selection mechanism. For each event type, SPTEAE learns a special target prompt by attending to highly relevant source prompts. Experiment results show that the average performance of SPTEAE with soft prompt transfer is 2.6% higher than the current state-of-the-art model on the ACE2005 dataset.","{'model': 'tldr@v2.0.0', 'text': 'A Soft Prompt Transfer model for zero-shot cross-lingual Event Argument Extraction, which utilizes a sequence of tunable vectors which are tuned in source language as event type prompts and learns a special target prompt by attending to highly relevant source prompts.'}",
-Prompt-based Zero-shot Text Classification with Conceptual Knowledge,Yuqi Wang,"The proposed framework incorporates conceptual knowledge for prompt-based text classification in the extreme zero-shot setting, which outperforms existing approaches in sentiment analysis and topic detection on four widely-used datasets.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2023.acl-srw.4.pdf
-Prompt-NER: Zero-shot Named Entity Recognition in Astronomy Literature via Large Language Models,Wujun Shao,"This study delves into the application of Large Language Models (LLMs) for Named Entity Recognition (NER) tasks in the field of astronomy literature. To enhance the zero-shot recognition capabilities of LLMs for astronomical named entities, we propose a strategy called Prompt-NER. Prompt-NER includes five prompt elements: Task Descriptions, Entity Definitions, Task Emphasis, Task Examples, and Second Conversation. To assess the effectiveness of the Prompt-NER strategy, we utilize three representative LLMs (Claude-2, GPT-3.5, and LLaMA-2-70b) to identify telescope and celestial object named entities in astronomical literature. Our experiments are conducted based on two distinct datasets. The first dataset comprises 30 original PDF documents, which we split into paragraphs in sequential order, resulting in a second dataset consisting of 30 paragraph collections. Additionally, we incorporate 30 astronomical telegrams to diversify our experiments and assess the performance of LLMs based on Prompt-NER on concise, complete texts. Our experimental results indicate that the Prompt-NER strategy enables LLMs to effectively accomplish NER tasks in the field of astronomy, even without prior astronomical knowledge during training. We carefully analyze the experimental results, including the mechanism of different prompt elements and the influence of different features of long and short texts on their respective experimental results. This research provides experience for zero-shot NER tasks in astronomical literature and suggests future work in this area.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results indicate that the Prompt-NER strategy enables LLMs to effectively accomplish NER tasks in the field of astronomy, even without prior astronomical knowledge during training.'}",
-"Synthesize, Prompt and Transfer: Zero-shot Conversational Question Generation with Pre-trained Language Model",Hongwei Zeng,"Conversational question generation aims to generate questions that depend on both context and conversation history. Conventional works utilizing deep learning have shown promising results, but heavily rely on the availability of large-scale annotated conversations. In this paper, we introduce a more realistic and less explored setting, Zero-shot Conversational Question Generation (ZeroCQG), which requires no human-labeled conversations for training. To solve ZeroCQG, we propose a multi-stage knowledge transfer framework, Synthesize, Prompt, and trAnsfer with pRe-Trained lAnguage model (SPARTA) to effectively leverage knowledge from single-turn question generation instances. To validate the zero-shot performance of SPARTA, we conduct extensive experiments on three conversational datasets: CoQA, QuAC, and DoQA by transferring knowledge from three single-turn datasets: MS MARCO, NewsQA, and SQuAD. The experimental results demonstrate the superior performance of our method. Specifically, SPARTA has achieved 14.81 BLEU-4 (88.2% absolute improvement compared to T5) in CoQA with knowledge transferred from SQuAD.","{'model': 'tldr@v2.0.0', 'text': 'A multi-stage knowledge transfer framework, Synthesize, Prompt, and trAnsfer with pRe-Trained lAnguage model (SPARTA) to effectively leverage knowledge from single-turn question generation instances to solve ZeroCQG.'}",https://aclanthology.org/2023.acl-long.500.pdf
-Weakly Supervised Few-Shot and Zero-Shot Semantic Segmentation with Mean Instance Aware Prompt Learning,Prashant Pandey,"The existing Zero-Shot Segmentation (ZSS) and Few-Shot Segmentation (FSS) methods utilize fully supervised pixel-labeled seen classes to segment unseen classes. Pixel-level labels are hard to obtain, and using weak supervision in the form of inexpensive image labels is often more practical. To this end, we propose a novel unified weakly supervised Zero-Shot and Few-Shot semantic segmentation pipeline that can perform ZSS and FSS on novel classes without using pixel-level labels for either the base (seen) or the novel (unseen) classes. We propose Mean Instance Aware Prompt based Network (MIAPNet), a novel language-guided segmentation pipeline that i) learns context vectors with batch aggregates (mean) to map class prompts to image features and ii) decouples weak ZSS/FSS into weak semantic segmentation and Zero-Shot segmentation. MIAPNet beats existing methods for weak generalized ZSS and weak FSS by 39 and 3 mIOU points respectively on PASCAL VOC and weak FSS by 5 mIOU points on MS COCO.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Mean Instance Aware Prompt based Network (MIAPNet), a novel language-guided segmentation pipeline that learns context vectors with batch aggregates to map class prompts to image features and decouples weak ZSS/FSS into weak semantic segmentation and Zero-Shot segmentation.'}",
-"Entities, Dates, and Languages: Zero-Shot on Historical Texts with T0",F. Toni,"In this work, we explore whether the recently demonstrated zero-shot abilities of the T0 model extend to Named Entity Recognition for out-of-distribution languages and time periods. Using a historical newspaper corpus in 3 languages as test-bed, we use prompts to extract possible named entities. Our results show that a naive approach for prompt-based zero-shot multilingual Named Entity Recognition is error-prone, but highlights the potential of such an approach for historical languages lacking labeled datasets. Moreover, we also find that T0-like models can be probed to predict the publication date and language of a document, which could be very relevant for the study of historical texts.","{'model': 'tldr@v2.0.0', 'text': 'This work explores whether the recently demonstrated zero-shot abilities of the T0 model extend to Named Entity Recognition for out-of-distribution languages and time periods, and uses a historical newspaper corpus in 3 languages as test-bed.'}",http://arxiv.org/pdf/2204.05211
-PESCO: Prompt-enhanced Self Contrastive Learning for Zero-shot Text Classification,Yau-Shian Wang,"We present PESCO, a novel contrastive learning framework that substantially improves the performance of zero-shot text classification. We formulate text classification as a neural text retrieval problem where each document is treated as a query, and the system learns the mapping from each query to the relevant class labels by (1) adding prompts to enhance label retrieval, and (2) using retrieved labels to enrich the training set in a self-training loop of contrastive learning. PESCO achieves state-of-the-art performance on four benchmark text classification datasets. On DBpedia, we achieve 98.5% accuracy without any labeled data, which is close to the fully-supervised result. Extensive experiments and analyses show all the components of PESCO are necessary for improving the performance of zero-shot text classification.","{'model': 'tldr@v2.0.0', 'text': 'PESCO is presented, a novel contrastive learning framework that substantially improves the performance of zero-shot text classification and achieves state-of-the-art performance on four benchmark text classification datasets.'}",http://arxiv.org/pdf/2305.14963
-Prompt to be Consistent is Better than Self-Consistent? Few-Shot and Zero-Shot Fact Verification with Pre-trained Language Models,Fengzhu Zeng,"Few-shot or zero-shot fact verification only relies on a few or no labeled training examples. In this paper, we propose a novel method called ProToCo, to \underline{Pro}mpt pre-trained language models (PLMs) \underline{To} be \underline{Co}nsistent, for improving the factuality assessment capability of PLMs in the few-shot and zero-shot settings. Given a claim-evidence pair, ProToCo generates multiple variants of the claim with different relations and frames a simple consistency mechanism as constraints for making compatible predictions across these variants. We update PLMs by using parameter-efficient fine-tuning (PEFT), leading to more accurate predictions in few-shot and zero-shot fact verification tasks. Our experiments on three public verification datasets show that ProToCo significantly outperforms state-of-the-art few-shot fact verification baselines. With a small number of unlabeled instances, ProToCo also outperforms the strong zero-shot learner T0 on zero-shot verification. Compared to large PLMs using in-context learning (ICL) method, ProToCo outperforms OPT-30B and the Self-Consistency-enabled OPT-6.7B model in both few- and zero-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel method called ProToCo, to improve the factuality assessment capability of pre-trained language models (PLMs) by using parameter-efficient fine-tuning (PEFT), leading to more accurate predictions in few-shot and zero-shot fact verification tasks.'}",http://arxiv.org/pdf/2306.02569
-Hierarchical Prompt Learning for Compositional Zero-Shot Recognition,Henan Wang,"Compositional Zero-Shot Learning (CZSL) aims to imitate the powerful generalization ability of human beings to recognize novel compositions of known primitive concepts that correspond to a state and an object, e.g., purple apple. To fully capture the intra- and inter-class correlations between compositional concepts, in this paper, we propose to learn them in a hierarchical manner. Specifically, we set up three hierarchical embedding spaces that respectively model the states, the objects, and their compositions, which serve as three “experts” that can be combined in inference for more accurate predictions. We achieve this based on the recent success of large-scale pretrained vision-language models, e.g., CLIP, which provides a strong initial knowledge of image-text relationships. To better adapt this knowledge to CZSL, we propose to learn three hierarchical prompts by explicitly fixing the unrelated word tokens in the three embedding spaces. Despite its simplicity, our proposed method consistently yields superior performance over current state-of-the-art approaches on three widely-used CZSL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper sets up three hierarchical embedding spaces that respectively model the states, the objects, and their compositions, which serve as three “experts” that can be combined in inference for more accurate predictions.'}",https://www.ijcai.org/proceedings/2023/0163.pdf
-AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detection,Qihang Zhou,"Zero-shot anomaly detection (ZSAD) requires detection models trained using auxiliary data to detect anomalies without any training sample in a target dataset. It is a crucial task when training data is not accessible due to various concerns, \eg, data privacy, yet it is challenging since the models need to generalize to anomalies across different domains where the appearance of foreground objects, abnormal regions, and background features, such as defects/tumors on different products/organs, can vary significantly. Recently large pre-trained vision-language models (VLMs), such as CLIP, have demonstrated strong zero-shot recognition ability in various vision tasks, including anomaly detection. However, their ZSAD performance is weak since the VLMs focus more on modeling the class semantics of the foreground objects rather than the abnormality/normality in the images. In this paper we introduce a novel approach, namely AnomalyCLIP, to adapt CLIP for accurate ZSAD across different domains. The key insight of AnomalyCLIP is to learn object-agnostic text prompts that capture generic normality and abnormality in an image regardless of its foreground objects. This allows our model to focus on the abnormal image regions rather than the object semantics, enabling generalized normality and abnormality recognition on diverse types of objects. Large-scale experiments on 17 real-world anomaly detection datasets show that AnomalyCLIP achieves superior zero-shot performance of detecting and segmenting anomalies in datasets of highly diverse class semantics from various defect inspection and medical imaging domains. Code will be made available at https://github.com/zqhang/AnomalyCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach, namely AnomalyCLIP, to adapt CLIP for accurate ZSAD across different domains, to learn object-agnostic text prompts that capture generic normality and abnormality in an image regardless of its foreground objects.'}",
-Enhancing Zero-Shot Crypto Sentiment with Fine-tuned Language Model and Prompt Engineering,Rahman S M Wahidur,"Blockchain technology has revolutionized the financial landscape, with cryptocurrencies gaining widespread adoption for their decentralized and transparent nature. As the sentiment expressed on social media platforms can significantly influence cryptocurrency discussions and market movements, sentiment analysis has emerged as a crucial tool for understanding public opinion and predicting market trends. Motivated by the aim to enhance sentiment analysis accuracy in the cryptocurrency domain, this paper investigates fine-tuning techniques on large language models. This paper also investigates the efficacy of supervised fine-tuning and instruction-based fine-tuning on large language models for unseen tasks. Experimental results demonstrate a significant average zero-shot performance gain of 40% after fine-tuning, highlighting the potential of this technique in optimizing pre-trained language model efficiency. Additionally, the impact of instruction tuning on models of varying scales is examined, revealing that larger models benefit from instruction tuning, achieving the highest average accuracy score of 75.16%. In contrast, smaller-scale models may experience reduced generalization due to the complete utilization of model capacity. To gain deeper insight about how instruction works with these language models, this paper presents an experimental investigation into the response of an instruction-based model under different instruction tuning setups. The investigation demonstrates that the model achieves an average accuracy score of 72.38% for short and simple instructions. This performance significantly outperforms its accuracy under long and complex instructions by over 12%, thereby effectively highlighting the profound significance of instruction characteristics in maximizing model performance.","{'model': 'tldr@v2.0.0', 'text': 'An experimental investigation into the response of an instruction-based model under different instruction tuning setups is presented, demonstrating a significant average zero-shot performance gain after fine-tuning and highlighting the potential of this technique in optimizing pre-trained language model efficiency.'}",
-Zero-shot Domain Adaptation for Neural Machine Translation with Retrieved Phrase-level Prompts,Zewei Sun,"Domain adaptation is an important challenge for neural machine translation. However, the traditional fine-tuning solution requires multiple extra training and yields a high cost. In this paper, we propose a non-tuning paradigm, resolving domain adaptation with a prompt-based method. Specifically, we construct a bilingual phrase-level database and retrieve relevant pairs from it as a prompt for the input sentences. By utilizing Retrieved Phrase-level Prompts (RePP), we effectively boost the translation quality. Experiments show that our method improves domain-specific machine translation for 6.2 BLEU scores and improves translation constraints for 11.5% accuracy without additional training.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a non-tuning paradigm, resolving domain adaptation with a prompt-based method, which improves domain-specific machine translation for 6.2 BLEU scores and improves translation constraints for 11.5% accuracy without additional training.'}",http://arxiv.org/pdf/2209.11409
-"ELECTRA is a Zero-Shot Learner, Too",Shiwen Ni,"Recently, for few-shot or even zero-shot learning, the new paradigm""pre-train, prompt, and predict""has achieved remarkable achievements compared with the""pre-train, fine-tune""paradigm. After the success of prompt-based GPT-3, a series of masked language model (MLM)-based (e.g., BERT, RoBERTa) prompt learning methods became popular and widely used. However, another efficient pre-trained discriminative model, ELECTRA, has probably been neglected. In this paper, we attempt to accomplish several NLP tasks in the zero-shot scenario using a novel our proposed replaced token detection (RTD)-based prompt learning method. Experimental results show that ELECTRA model based on RTD-prompt learning achieves surprisingly state-of-the-art zero-shot performance. Numerically, compared to MLM-RoBERTa-large and MLM-BERT-large, our RTD-ELECTRA-large has an average of about 8.4% and 13.7% improvement on all 15 tasks. Especially on the SST-2 task, our RTD-ELECTRA-large achieves an astonishing 90.1% accuracy without any training data. Overall, compared to the pre-trained masked language models, the pre-trained replaced token detection model performs better in zero-shot learning. The source code is available at: https://github.com/nishiwen1214/RTD-ELECTRA.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that ELECTRA model based on RTD-prompt learning achieves surprisingly state-of-the-art zero-shot performance, and compared to the pre-trained masked language models, thePre-trained replaced token detection model performs better in zero- shot learning.'}",http://arxiv.org/pdf/2207.08141
-Empowering Sentence Encoders with Prompting and Label Retrieval for Zero-shot Text Classification,Jimin Hong,"With contrastive pre-training, sentence encoders are generally optimized to locate semantically similar samples closer to each other in their embedding spaces. In this work, we focus on the potential of their embedding spaces to be readily adapted to zero-shot text classification, as semantically distinct samples are already well-separated. Our framework, RaLP (Retrieval augmented Label Prompts for sentence encoder), encodes prompted label candidates with a sentence encoder, then assigns the label whose prompt embedding has the highest similarity with the input text embedding. In order to compensate for the potentially poorly descriptive labels in their original format, RaLP retrieves sentences that are semantically similar to the original label prompt from external corpora and use them as additional pseudo-label prompts. RaLP achieves competitive or stronger performance than much larger baselines on various closed-set classification and multiple-choice QA datasets under zero-shot settings. We show that the retrieval component plays a pivotal role in RaLP's success, and its results are robustly attained regardless of verbalizer variations.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a framework that retrieves sentences that are semantically similar to the original label prompt from external corpora and uses them as additional pseudo-label prompts, and achieves competitive or stronger performance on various closed-set classification and multiple-choice QA datasets under zero-shot settings.'}",
-Evaluating Prompts Across Multiple Choice Tasks In a Zero-Shot Setting,Gabriel Orlanski,"Large language models have shown that impressive zero-shot performance can be achieved through natural language prompts (Radford et al., 2019; Brown et al., 2020; Sanh et al., 2021). Creating an effective prompt, however, requires significant trial and error. That \textit{prompts} the question: how do the qualities of a prompt effects its performance? To this end, we collect and standardize prompts from a diverse range of tasks for use with tasks they were not designed for. We then evaluate these prompts across fixed multiple choice datasets for a quantitative analysis of how certain attributes of a prompt affect performance. We find that including the choices and using prompts not used during pre-training provide significant improvements. All experiments and code can be found https://github.com/gabeorlanski/zero-shot-cross-task.","{'model': 'tldr@v2.0.0', 'text': 'This work collects and standardizes prompts from a diverse range of tasks for use with tasks they were not designed for, and evaluates these prompts across fixed multiple choice datasets for a quantitative analysis of how certain attributes of a prompt affect performance.'}",http://arxiv.org/pdf/2203.15754
-ZEROTOP: Zero-Shot Task-Oriented Semantic Parsing using Large Language Models,Dheeraj Mekala,"We explore the use of large language models (LLMs) for zero-shot semantic parsing. Semantic parsing involves mapping natural language utterances to task-specific meaning representations. Language models are generally trained on the publicly available text and code and cannot be expected to directly generalize to domain-specific parsing tasks in a zero-shot setting. In this work, we propose ZEROTOP, a zero-shot task-oriented parsing method that decomposes a semantic parsing problem into a set of abstractive and extractive question-answering (QA) problems, enabling us to leverage the ability of LLMs to zero-shot answer reading comprehension questions. For each utterance, we prompt the LLM with questions corresponding to its top-level intent and a set of slots and use the LLM generations to construct the target meaning representation. We observe that current LLMs fail to detect unanswerable questions; and as a result, cannot handle questions corresponding to missing slots. To address this problem, we fine-tune a language model on public QA datasets using synthetic negative samples. Experimental results show that our QA-based decomposition paired with the fine-tuned LLM can correctly parse ~16% of utterances in the MTOP dataset without requiring any annotated data.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ZEROTOP, a zero-shot task-oriented parsing method that decomposes a semantic parsing problem into a set of abstractive and extractive question-answering (QA) problems, enabling us to leverage the ability of LLMs to zero- shot answer reading comprehension questions.'}",http://arxiv.org/pdf/2212.10815
-Go-tuning: Improving Zero-shot Learning Abilities of Smaller Language Models,Jingjing Xu,"With increasing scale, large language models demonstrate both quantitative improvement and new qualitative capabilities, especially as zero-shot learners, like GPT-3. However, these results rely heavily on delicate prompt design and large computation. In this work, we explore whether the strong zero-shot ability could be achieved at a smaller model scale without any external supervised data. To achieve this goal, we revisit masked language modeling and present a geometry-guided self-supervised learning method (Go-tuningfor short) by taking a small number of task-aware self-supervised data to update language models further. Experiments show that Go-tuning can enable T5-small (80M) competitive zero-shot results compared with large language models, such as T5-XL (3B). We also apply Go-tuning on multi-task settings and develop a multi-task model, mgo-T5 (250M). It can reach the average performance of OPT (175B) on 9 datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work revisits masked language modeling and presents a geometry-guided self-supervised learning method (Go-tuningfor short) by taking a small number of task-aware self- supervised data to update language models further, which can enable T5-small (80M) competitive zero-shot results compared with large language models.'}",http://arxiv.org/pdf/2212.10461
-"How to Prompt LLMs for Text-to-SQL: A Study in Zero-shot, Single-domain, and Cross-domain Settings",Shuaichen Chang,"Large language models (LLMs) with in-context learning have demonstrated remarkable capability in the text-to-SQL task. Previous research has prompted LLMs with various demonstration-retrieval strategies and intermediate reasoning steps to enhance the performance of LLMs. However, those works often employ varied strategies when constructing the prompt text for text-to-SQL inputs, such as databases and demonstration examples. This leads to a lack of comparability in both the prompt constructions and their primary contributions. Furthermore, selecting an effective prompt construction has emerged as a persistent problem for future research. To address this limitation, we comprehensively investigate the impact of prompt constructions across various settings and provide insights for future work.","{'model': 'tldr@v2.0.0', 'text': 'This work comprehensively investigates the impact of prompt constructions across various settings and provides insights for future work.'}",http://arxiv.org/pdf/2305.11853
-Dialogue State Tracking with Zero-Shot and Few-Shot Learning for Generalization: A Review,Seungyeon Kim,"Research on Dialogue State Tracking (DST) has achieved meaningful advancements in benchmark datasets. However, the generalization ability of DST models to handle unseen data robustly remains an issue. Hence, recent studies on DST with zero-shot and few-shot learning are reviewed in this paper. For a task-oriented dialogue system, DST is explained by introducing datasets and evaluation metrics. DST models could be categorized into four groups: DST based on a pre-trained model, DST using a description, DST using a prompt, and DST with cross-task. Characteristics of each model are described and the performance of the model experimented under the same conditions is summarized.","{'model': 'tldr@v2.0.0', 'text': 'Recent studies on DST with zero-shot and few-shot learning are reviewed and characteristics of each model are described and the performance of the model experimented under the same conditions is summarized.'}",
-Self-Prompting Large Language Models for Zero-Shot Open-Domain QA,Junlong Li,"Open-Domain Question Answering (ODQA) aims at answering factoid questions without explicitly providing specific background documents. In a zero-shot setting, this task is more challenging since no data is available to train customized models like Retriever-Readers. Recently, Large Language Models (LLMs) like GPT-3 have shown their power in zero-shot ODQA with direct prompting methods, but these methods are still far from releasing the full powerfulness of LLMs only in an implicitly invoking way. In this paper, we propose a Self-Prompting framework to explicitly utilize the massive knowledge stored in the parameters of LLMs and their strong instruction understanding abilities. Concretely, we prompt LLMs step by step to generate multiple pseudo QA pairs with background passages and explanations from scratch and then use those generated elements for in-context learning. Experimental results show our method surpasses previous SOTA methods significantly on three widely-used ODQA datasets, and even achieves comparable performance with some Retriever-Reader models fine-tuned on full training data.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Self-Prompting framework to explicitly utilize the massive knowledge stored in the parameters of LLMs and their strong instruction understanding abilities and achieves comparable performance with some Retriever-Reader models fine-tuned on full training data.'}",
-MALM: Mixing Augmented Language Modeling for Zero-Shot Machine Translation,Kshitij Gupta,"Large pre-trained language models have brought remarkable progress in NLP. Pre-training and Fine-tuning have given state-of-art performance across tasks in text processing. Data Augmentation techniques have also helped build state-of-art models on low or zero resource tasks. Many works in the past have attempted at learning a single massively multilingual machine translation model for zero-shot translation. Although those translation models are producing correct translations, the main challenge is those models are producing the wrong languages for zero-shot translation. This work and its results indicate that prompt conditioned large models do not suffer from off-target language errors i.e. errors arising due to translation to wrong languages. We empirically demonstrate the effectiveness of self-supervised pre-training and data augmentation for zero-shot multi-lingual machine translation.","{'model': 'tldr@v2.0.0', 'text': 'This work and its results indicate that prompt conditioned large models do not suffer from off-target language errors i.e. errors arising due to translation to wrong languages.'}",http://arxiv.org/pdf/2210.00320
-Zero-shot Domain-sensitive Speech Recognition with Prompt-conditioning Fine-tuning,F. Liao,"In this work, we propose a method to create domain-sensitive speech recognition models that utilize textual domain information by conditioning its generation on a given text prompt. This is accomplished by fine-tuning a pre-trained, end-to-end model (Whisper) to learn from demonstrations with prompt examples. We show that this ability can be generalized to different domains and even various prompt contexts, with our model gaining a Word Error Rate (WER) reduction of up to 33% on unseen datasets from various domains, such as medical conversation, air traffic control communication, and financial meetings. Considering the limited availability of audio-transcript pair data, we further extend our method to text-only fine-tuning to achieve domain sensitivity as well as domain adaptation. We demonstrate that our text-only fine-tuned model can also attend to various prompt contexts, with the model reaching the most WER reduction of 29% on the medical conversation dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to create domain-sensitive speech recognition models that utilize textual domain information by conditioning its generation on a given text prompt by fine-tuning a pre-trained, end-to-end model (Whisper).'}",https://arxiv.org/pdf/2307.10274
-Prompt Engineering and Calibration for Zero-Shot Commonsense Reasoning,Chenkai Ma,"Prompt engineering and calibration make large language models excel at reasoning tasks, including multiple choice commonsense reasoning. From a practical perspective, we investigate and evaluate these strategies on smaller language models. Through experiments on five commonsense reasoning benchmarks, we find that each strategy favors certain models, but their joint effects are mostly negative.","{'model': 'tldr@v2.0.0', 'text': 'Through experiments on five commonsense reasoning benchmarks, it is found that each strategy favors certain models, but their joint effects are mostly negative.'}",http://arxiv.org/pdf/2304.06962
-Kbpt: Knowledge-Based Prompt Tuning for Zero-Shot Relation Triplet Extraction,Q. Guo,,,
-Zero-Shot Recommendation as Language Modeling,Damien Sileo,,"{'model': 'tldr@v2.0.0', 'text': 'A framework for recommendation with off-the-shelf pretrained language models (LM) that only used unstructured text corpora as training data is proposed and compared with standard matrix factorization trained on different data regimes.'}",https://arxiv.org/pdf/2112.04184
-Cut the CARP: Fishing for zero-shot story evaluation,Shahbuland Matiana,"Recent advances in large-scale language models (Raffel et al., 2019; Brown et al., 2020) have brought significant qualitative and quantitative improvements in machine-driven text generation. Despite this, generation and evaluation of machine-generated narrative text remains a challenging problem. Objective evaluation of computationally-generated stories may be prohibitively expensive, require meticulously annotated datasets, or may not adequately measure the logical coherence of a generated story's narratological structure. Informed by recent advances in contrastive learning (Radford et al., 2021), we present Contrastive Authoring and Reviewing Pairing (CARP): a scalable, efficient method for performing qualitatively superior, zero-shot evaluation of stories. We show a strong correlation between human evaluation of stories and those of CARP. Model outputs more significantly correlate with corresponding human input than those language-model based methods which utilize finetuning or prompt engineering approaches. We also present and analyze the Story-Critique Dataset, a new corpora composed of 1.3 million aligned story-critique pairs derived from over 80,000 stories. We expect this corpus to be of interest to NLP researchers.","{'model': 'tldr@v2.0.0', 'text': 'A strong correlation between human evaluation of stories and those of CARP is shown, and model outputs more significantly correlate with corresponding human input than those language-model based methods which utilize finetuning or prompt engineering approaches.'}",
-Zero-shot Learning by Generating Task-specific Adapters,Qinyuan Ye,"Pre-trained text-to-text transformers achieve impressive performance across a wide range of NLP tasks, and they naturally support zero-shot learning (ZSL) by using the task description as prompt in the input. However, this approach has potential limitations, as it learns from input-output pairs at instance level, instead of learning to solve tasks at task level. Alternatively, applying existing ZSL methods to text-to-text transformers is non-trivial due to their text generation objective and huge size. To address these issues, we introduce H YPTER , a framework that improves zero-shot transferability by training a hypernetwork to generate task-specific adapter s from task descriptions. This formulation enables learning at task level, and greatly reduces the number of parameters by using light-weight adapters. Experiments on two datasets demonstrate H YPTER improves upon fine-tuning baselines.","{'model': 'tldr@v2.0.0', 'text': 'H YPTER is introduced, a framework that improves zero-shot transferability by training a hypernetwork to generate task-specific adapter s from task descriptions, and greatly reduces the number of parameters by using light-weight adapters.'}",
-Domain-Aware Continual Zero-Shot Learning,Kai Yi,"We introduce Domain Aware Continual Zero-Shot Learning (DACZSL), the task of visually recognizing images of unseen categories in unseen domains sequentially. We created DACZSL on top of the DomainNet dataset by dividing it into a sequence of tasks, where classes are incrementally provided on seen domains during training and evaluation is conducted on unseen domains for both seen and unseen classes. We also proposed a novel Domain-Invariant CZSL Network (DIN), which outperforms state-of-the-art baseline models that we adapted to DACZSL setting. We adopt a structure-based approach to alleviate forgetting knowledge from previous tasks with a small per-task private network in addition to a global shared network. To encourage the private network to capture the domain and task-specific representation, we train our model with a novel adversarial knowledge disentanglement setting to make our global network task-invariant and domain-invariant over all the tasks. Our method also learns a class-wise learnable prompt to obtain better class-level text representation, which is used to represent side information to enable zero-shot prediction of future unseen classes. Our code and benchmarks will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Domain Aware Continual Zero-Shot Learning (DACZSL), the task of visually recognizing images of unseen categories in unseen domains sequentially, and proposes a novel Domain-Invariant CZSL Network (DIN), which outperforms state-of-the-art baseline models that are adapted to DACZSL setting.'}",
-Cognitively Aided Zero-Shot Automatic Essay Grading,Sandeep Albert Mathias,"Automatic essay grading (AEG) is a process in which machines assign a grade to an essay written in response to a topic, called the prompt. Zero-shot AEG is when we train a system to grade essays written to a new prompt which was not present in our training data. In this paper, we describe a solution to the problem of zero-shot automatic essay grading, using cognitive information, in the form of gaze behaviour. Our experiments show that using gaze behaviour helps in improving the performance of AEG systems, especially when we provide a new essay written in response to a new prompt for scoring, by an average of almost 5 percentage points of QWK.","{'model': 'tldr@v2.0.0', 'text': 'Their experiments show that using gaze behaviour helps in improving the performance of AEG systems, especially when the authors provide a new essay written in response to a new prompt for scoring, by an average of almost 5 percentage points of QWK.'}",
-SAM.MD: Zero-shot medical image segmentation capabilities of the Segment Anything Model,Saikat Roy,"Foundation models have taken over natural language processing and image generation domains due to the flexibility of prompting. With the recent introduction of the Segment Anything Model (SAM), this prompt-driven paradigm has entered image segmentation with a hitherto unexplored abundance of capabilities. The purpose of this paper is to conduct an initial evaluation of the out-of-the-box zero-shot capabilities of SAM for medical image segmentation, by evaluating its performance on an abdominal CT organ segmentation task, via point or bounding box based prompting. We show that SAM generalizes well to CT data, making it a potential catalyst for the advancement of semi-automatic segmentation tools for clinicians. We believe that this foundation model, while not reaching state-of-the-art segmentation performance in our investigations, can serve as a highly potent starting point for further adaptations of such models to the intricacies of the medical domain. Keywords: medical image segmentation, SAM, foundation models, zero-shot learning","{'model': 'tldr@v2.0.0', 'text': 'It is shown that SAM generalizes well to CT data, making it a potential catalyst for the advancement of semi-automatic segmentation tools for clinicians, and can serve as a highly potent starting point for further adaptations of such models to the intricacies of the medical domain.'}",http://arxiv.org/pdf/2304.05396
-Zero-shot Clinical Entity Recognition using ChatGPT,Yan Hu,"In this study, we investigated the potential of ChatGPT, a large language model developed by OpenAI, for the clinical named entity recognition task defined in the 2010 i2b2 challenge, in a zero-shot setting with two different prompt strategies. We compared its performance with GPT-3 in a similar zero-shot setting, as well as a fine-tuned BioClinicalBERT model using a set of synthetic clinical notes from MTSamples. Our findings revealed that ChatGPT outperformed GPT-3 in the zero-shot setting, with F1 scores of 0.418 (vs.0.250) and 0.620 (vs. 0.480) for exact- and relaxed-matching, respectively. Moreover, prompts affected ChatGPT's performance greatly, with relaxed-matching F1 scores of 0.628 vs.0.541 for two different prompt strategies. Although ChatGPT's performance was still lower than that of the supervised BioClinicalBERT model (i.e., relaxed-matching F1 scores of 0.620 vs. 0.888), our study demonstrates the great potential of ChatGPT for clinical NER tasks in a zero-shot setting, which is much more appealing as it does not require any annotation.",,http://arxiv.org/pdf/2303.16416
-A Preliminary Evaluation of ChatGPT for Zero-shot Dialogue Understanding,Wenbo Pan,"Zero-shot dialogue understanding aims to enable dialogue to track the user's needs without any training data, which has gained increasing attention. In this work, we investigate the understanding ability of ChatGPT for zero-shot dialogue understanding tasks including spoken language understanding (SLU) and dialogue state tracking (DST). Experimental results on four popular benchmarks reveal the great potential of ChatGPT for zero-shot dialogue understanding. In addition, extensive analysis shows that ChatGPT benefits from the multi-turn interactive prompt in the DST task but struggles to perform slot filling for SLU. Finally, we summarize several unexpected behaviors of ChatGPT in dialogue understanding tasks, hoping to provide some insights for future research on building zero-shot dialogue understanding systems with Large Language Models (LLMs).","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the understanding ability of ChatGPT for zero-shot dialogue understanding tasks including spoken language understanding (SLU) and dialogue state tracking (DST).'}",http://arxiv.org/pdf/2304.04256
-ESC: Exploration with Soft Commonsense Constraints for Zero-shot Object Navigation,KAI-QING Zhou,"The ability to accurately locate and navigate to a specific object is a crucial capability for embodied agents that operate in the real world and interact with objects to complete tasks. Such object navigation tasks usually require large-scale training in visual environments with labeled objects, which generalizes poorly to novel objects in unknown environments. In this work, we present a novel zero-shot object navigation method, Exploration with Soft Commonsense constraints (ESC), that transfers commonsense knowledge in pre-trained models to open-world object navigation without any navigation experience nor any other training on the visual environments. First, ESC leverages a pre-trained vision and language model for open-world prompt-based grounding and a pre-trained commonsense language model for room and object reasoning. Then ESC converts commonsense knowledge into navigation actions by modeling it as soft logic predicates for efficient exploration. Extensive experiments on MP3D, HM3D, and RoboTHOR benchmarks show that our ESC method improves significantly over baselines, and achieves new state-of-the-art results for zero-shot object navigation (e.g., 288% relative Success Rate improvement than CoW on MP3D).","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot object navigation method, Exploration with Soft Commonsense constraints (ESC), that transfers commonsense knowledge in pre-trained models to open-world object navigation without any navigation experience nor any other training on the visual environments is presented.'}",https://arxiv.org/pdf/2301.13166
-Zero-shot performance of the Segment Anything Model (SAM) in 2D medical imaging: A comprehensive evaluation and practical guidelines,C. M. Oliveira,"Segmentation in medical imaging is a critical component for the diagnosis, monitoring, and treatment of various diseases and medical conditions. Presently, the medical segmentation landscape is dominated by numerous specialized deep learning models, each fine-tuned for specific segmentation tasks and image modalities. The recently-introduced Segment Anything Model (SAM) employs the ViT neural architecture and harnesses a massive training dataset to segment nearly any object; however, its suitability to the medical domain has not yet been investigated. In this study, we explore the zero-shot performance of SAM in medical imaging by implementing eight distinct prompt strategies across six datasets from four imaging modalities, including X-ray, ultrasound, dermatoscopy, and colonoscopy. Our findings reveal that SAM's zero-shot performance is not only comparable to, but in certain cases, surpasses the current state-of-the-art. Based on these results, we propose practical guidelines that require minimal interaction while consistently yielding robust outcomes across all assessed contexts. The source code, along with a demonstration of the recommended guidelines, can be accessed at https://github.com/Malta-Lab/SAM-zero-shot-in-Medical-Imaging.","{'model': 'tldr@v2.0.0', 'text': ""The findings reveal that SAM's zero-shot performance is not only comparable to, but in certain cases, surpasses the current state-of-the-art, and practical guidelines are proposed that require minimal interaction while consistently yielding robust outcomes across all assessed contexts.""}",http://arxiv.org/pdf/2305.00109
-"CLIP for All Things Zero-Shot Sketch-Based Image Retrieval, Fine-Grained or Not",Aneeshan Sain,"In this paper, we leverage CLIP for zero-shot sketch based image retrieval (ZS-SBIR). We are largely inspired by recent advances on foundation models and the unparalleled generalisation ability they seem to offer, but for the first time tailor it to benefit the sketch community. We put forward novel designs on how best to achieve this synergy, for both the category setting and the fine-grained setting ('all”}. At the very core of our solution is a prompt learning setup. First we show just via factoring in sketch-specific prompts, we already have a category-level ZS-SBIR system that over-shoots all prior arts, by a large margin (24.8%) - a great testimony on studying the CLIP and ZS-SBIR synergy. Moving onto the fine-grained setup is however trickier, and re-quires a deeper dive into this synergy. For that, we come up with two specific designs to tackle the fine-grained matching nature of the problem: (i) an additional regularisation loss to ensure the relative separation between sketches and photos is uniform across categories, which is not the case for the gold standard standalone triplet loss, and (ii) a clever patch shuffling technique to help establishing instance-level structural correspondences between sketch-photo pairs. With these designs, we again observe signifi-cant performance gains in the region of 26.9% over previ-ous state-of-the-art. The take-home message, if any, is the proposed CLIP and prompt learning paradigm carries great promise in tackling other sketch-related tasks (not limited to ZS-SBIR) where data scarcity remains a great challenge. Project page: https://aneeshan95.github.ioISketchLVM/","{'model': 'tldr@v2.0.0', 'text': 'The proposed CLIP and prompt learning paradigm carries great promise in tackling other sketch-related tasks (not limited to ZS-SBIR) where data scarcity remains a great challenge.'}",https://arxiv.org/pdf/2303.13440
-Zero-shot Temporal Relation Extraction with ChatGPT,Chenhan Yuan,"The goal of temporal relation extraction is to infer the temporal relation between two events in the document. Supervised models are dominant in this task. In this work, we investigate ChatGPT’s ability on zero-shot temporal relation extraction. We designed three different prompt techniques to break down the task and evaluate ChatGPT. Our experiments show that ChatGPT’s performance has a large gap with that of supervised methods and can heavily rely on the design of prompts. We further demonstrate that ChatGPT can infer more small relation classes correctly than supervised methods. The current shortcomings of ChatGPT on temporal relation extraction are also discussed in this paper. We found that ChatGPT cannot keep consistency during temporal inference and it fails in actively long-dependency temporal inference.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT cannot keep consistency during temporal inference and it fails in actively long-dependency temporal inference.'}",http://arxiv.org/pdf/2304.05454
-ZegOT: Zero-shot Segmentation Through Optimal Transport of Text Prompts,Kwanyoung Kim,"Recent success of large-scale Contrastive Language-Image Pre-training (CLIP) has led to great promise in zero-shot semantic segmentation by transferring image-text aligned knowledge to pixel-level classification. However, existing methods usually require an additional image encoder or retraining/tuning the CLIP module. Here, we propose a novel Zero-shot segmentation with Optimal Transport (ZegOT) method that matches multiple text prompts with frozen image embeddings through optimal transport. In particular, we introduce a novel Multiple Prompt Optimal Transport Solver (MPOT), which is designed to learn an optimal mapping between multiple text prompts and visual feature maps of the frozen image encoder hidden layers. This unique mapping method facilitates each of the multiple text prompts to effectively focus on distinct visual semantic attributes. Through extensive experiments on benchmark datasets, we show that our method achieves the state-of-the-art (SOTA) performance over existing Zero-shot Semantic Segmentation (ZS3) approaches.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel Multiple Prompt Optimal Transport Solver (MPOT), which is designed to learn an optimal mapping between multiple text prompts and visual feature maps of the frozen image encoder hidden layers, and achieves the state-of-the-art (SOTA) performance over existing Zero-shot Semantic Segmentation (ZS3) approaches.'}",http://arxiv.org/pdf/2301.12171
-Xplainer: From X-Ray Observations to Explainable Zero-Shot Diagnosis,Chantal Pellegrini,"Automated diagnosis prediction from medical images is a valuable resource to support clinical decision-making. However, such systems usually need to be trained on large amounts of annotated data, which often is scarce in the medical domain. Zero-shot methods address this challenge by allowing a flexible adaption to new settings with different clinical findings without relying on labeled data. Further, to integrate automated diagnosis in the clinical workflow, methods should be transparent and explainable, increasing medical professionals' trust and facilitating correctness verification. In this work, we introduce Xplainer, a novel framework for explainable zero-shot diagnosis in the clinical setting. Xplainer adapts the classification-by-description approach of contrastive vision-language models to the multi-label medical diagnosis task. Specifically, instead of directly predicting a diagnosis, we prompt the model to classify the existence of descriptive observations, which a radiologist would look for on an X-Ray scan, and use the descriptor probabilities to estimate the likelihood of a diagnosis. Our model is explainable by design, as the final diagnosis prediction is directly based on the prediction of the underlying descriptors. We evaluate Xplainer on two chest X-ray datasets, CheXpert and ChestX-ray14, and demonstrate its effectiveness in improving the performance and explainability of zero-shot diagnosis. Our results suggest that Xplainer provides a more detailed understanding of the decision-making process and can be a valuable tool for clinical diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'Xplainer is introduced, a novel framework for explainable zero-shot diagnosis in the clinical setting that adapts the classification-by-description approach of contrastive vision-language models to the multi-label medical diagnosis task and can be a valuable tool for clinical diagnosis.'}",http://arxiv.org/pdf/2303.13391
-RAPGen: An Approach for Fixing Code Inefficiencies in Zero-Shot,Spandan Garg,"Performance bugs are non-functional bugs that can even manifest in well-tested commercial products. Fixing these performance bugs is an important yet challenging problem. In this work, we address this challenge and present a new approach called Retrieval-Augmented Prompt Generation (RAPGen). Given a code snippet with a performance issue, RAPGen first retrieves a prompt instruction from a pre-constructed knowledge-base of previous performance bug fixes and then generates a prompt using the retrieved instruction. It then uses this prompt on a Large Language Model (such as Codex) in zero-shot to generate a fix. We compare our approach with the various prompt variations and state of the art methods in the task of performance bug fixing. Our evaluation shows that RAPGen can generate performance improvement suggestions equivalent or better than a developer in ~60% of the cases, getting ~39% of them verbatim, in an expert-verified dataset of past performance changes made by C# developers.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a new approach called Retrieval-Augmented Prompt Generation (RAPGen), which first retrieves a prompt instruction from a pre-constructed knowledge-base of previous performance bug fixes and then generates a prompt using the retrieved instruction.'}",http://arxiv.org/pdf/2306.17077
-Zero-Shot Cross-Lingual Summarization via Large Language Models,Jiaan Wang,"Given a document in a source language, cross-lingual summarization (CLS) aims to generate a summary in a different target language. Recently, the emergence of Large Language Models (LLMs), such as GPT-3.5, ChatGPT and GPT-4, has attracted wide attention from the computational linguistics community. However, it is not yet known the performance of LLMs on CLS. In this report, we empirically use various prompts to guide LLMs to perform zero-shot CLS from different paradigms (i.e., end-to-end and pipeline), and provide a preliminary evaluation on the generated summaries. We find that ChatGPT and GPT-4 originally prefer to produce lengthy summaries with detailed information. These two LLMs can further balance informativeness and conciseness with the help of an interactive prompt, significantly improving their CLS performance. Experimental results on three widely-used CLS datasets show that GPT-4 achieves state-of-the-art zero-shot CLS performance, and performs competitively compared with the fine-tuned mBART-50. Moreover, we also find some multi-lingual and bilingual LLMs (i.e., BLOOMZ, ChatGLM-6B, Vicuna-13B and ChatYuan) have limited zero-shot CLS ability. Due to the composite nature of CLS, which requires models to perform summarization and translation simultaneously, accomplishing this task in a zero-shot manner is even a challenge for LLMs. Therefore, we sincerely hope and recommend future LLM research could use CLS as a testbed.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT and GPT-4 originally prefer to produce lengthy summaries with detailed information, and these two LLMs can further balance informativeness and conciseness with the help of an interactive prompt, significantly improving their CLS performance.'}",
-Enabling Calibration In The Zero-Shot Inference of Large Vision-Language Models,Will LeVine,"Calibration of deep learning models is crucial to their trustworthiness and safe usage, and as such, has been extensively studied in supervised classification models, with methods crafted to decrease miscalibration. However, there has yet to be a comprehensive study of the calibration of vision-language models that are used for zero-shot inference, like CLIP. We measure calibration across relevant variables like prompt, dataset, and architecture, and find that zero-shot inference with CLIP is miscalibrated. Furthermore, we propose a modified version of temperature scaling that is aligned with the common use cases of CLIP as a zero-shot inference model, and show that a single learned temperature generalizes for each specific CLIP model (defined by a chosen pre-training dataset and architecture) across inference dataset and prompt choice.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a modified version of temperature scaling that is aligned with the common use cases of CLIP as a zero-shot inference model, and shows that a single learned temperature generalizes for each specific CLIP model across inference dataset and prompt choice.'}",http://arxiv.org/pdf/2303.12748
-CLIPN for Zero-Shot OOD Detection: Teaching CLIP to Say No,Hualiang Wang,"Out-of-distribution (OOD) detection refers to training the model on an in-distribution (ID) dataset to classify whether the input images come from unknown classes. Considerable effort has been invested in designing various OOD detection methods based on either convolutional neural networks or transformers. However, zero-shot OOD detection methods driven by CLIP, which only require class names for ID, have received less attention. This paper presents a novel method, namely CLIP saying no (CLIPN), which empowers the logic of saying no within CLIP. Our key motivation is to equip CLIP with the capability of distinguishing OOD and ID samples using positive-semantic prompts and negation-semantic prompts. Specifically, we design a novel learnable no prompt and a no text encoder to capture negation semantics within images. Subsequently, we introduce two loss functions: the image-text binary-opposite loss and the text semantic-opposite loss, which we use to teach CLIPN to associate images with no prompts, thereby enabling it to identify unknown samples. Furthermore, we propose two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from no prompts and the text encoder. Experimental results on 9 benchmark datasets (3 ID datasets and 6 OOD datasets) for the OOD detection task demonstrate that CLIPN, based on ViT-B-16, outperforms 7 well-used algorithms by at least 2.34% and 11.64% in terms of AUROC and FPR95 for zero-shot OOD detection on ImageNet-1K. Our CLIPN can serve as a solid foundation for effectively leveraging CLIP in downstream OOD tasks. The code is available on https://github.com/xmed-lab/CLIPN.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel method, namely CLIP saying no (CLIPN), which empowers the logic of saying no within CLIP and proposes two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from no prompts and the text encoder.'}",https://arxiv.org/pdf/2308.12213
-InFusion: Inject and Attention Fusion for Multi Concept Zero-Shot Text-based Video Editing,Anant Khandelwal,"Large text-to-image diffusion models have achieved remarkable success in generating diverse, high-quality images. Additionally, these models have been successfully leveraged to edit input images by just changing the text prompt. But when these models are applied to videos, the main challenge is to ensure temporal consistency and coherence across frames. In this paper, we propose InFusion, a framework for zero-shot text-based video editing leveraging large pre-trained image diffusion models. Our framework specifically supports editing of multiple concepts with pixel-level control over diverse concepts mentioned in the editing prompt. Specifically, we inject the difference in features obtained with source and edit prompts from U-Net residual blocks of decoder layers. When these are combined with injected attention features, it becomes feasible to query the source contents and scale edited concepts along with the injection of unedited parts. The editing is further controlled in a fine-grained manner with mask extraction and attention fusion, which cut the edited part from the source and paste it into the denoising pipeline for the editing prompt. Our framework is a low-cost alternative to one-shot tuned models for editing since it does not require training. We demonstrated complex concept editing with a generalised image model (Stable Diffusion v1.5) using LoRA. Adaptation is compatible with all the existing image diffusion techniques. Extensive experimental results demonstrate the effectiveness of existing methods in rendering high-quality and temporally consistent videos.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes InFusion, a framework for zero-shot text-based video editing leveraging large pre-trained image diffusion models, and demonstrates complex concept editing with a generalised image model (Stable Diffusion v1.5) using LoRA.'}",https://arxiv.org/pdf/2308.00135
-TagGPT: Large Language Models are Zero-shot Multimodal Taggers,Chen Li,"Tags are pivotal in facilitating the effective distribution of multimedia content in various applications in the contemporary Internet era, such as search engines and recommendation systems. Recently, large language models (LLMs) have demonstrated impressive capabilities across a wide range of tasks. In this work, we propose TagGPT, a fully automated system capable of tag extraction and multimodal tagging in a completely zero-shot fashion. Our core insight is that, through elaborate prompt engineering, LLMs are able to extract and reason about proper tags given textual clues of multimodal data, e.g., OCR, ASR, title, etc. Specifically, to automatically build a high-quality tag set that reflects user intent and interests for a specific application, TagGPT predicts large-scale candidate tags from a series of raw data via prompting LLMs, filtered with frequency and semantics. Given a new entity that needs tagging for distribution, TagGPT introduces two alternative options for zero-shot tagging, i.e., a generative method with late semantic matching with the tag set, and another selective method with early matching in prompts. It is well noticed that TagGPT provides a system-level solution based on a modular framework equipped with a pre-trained LLM (GPT-3.5 used here) and a sentence embedding model (SimCSE used here), which can be seamlessly replaced with any more advanced one you want. TagGPT is applicable for various modalities of data in modern social media and showcases strong generalization ability to a wide range of applications. We evaluate TagGPT on publicly available datasets, i.e., Kuaishou and Food.com, and demonstrate the effectiveness of TagGPT compared to existing hashtags and off-the-shelf taggers. Project page: https://github.com/TencentARC/TagGPT.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes TagGPT, a fully automated system capable of tag extraction and multimodal tagging in a completely zero-shot fashion, and provides a system-level solution based on a modular framework equipped with a pre-trained LLM and a sentence embedding model.'}",http://arxiv.org/pdf/2304.03022
-Troika: Multi-Path Cross-Modal Traction for Compositional Zero-Shot Learning,Siteng Huang,"Recent compositional zero-shot learning (CZSL) methods adapt pre-trained vision-language models (VLMs) by constructing trainable prompts only for composed state-object pairs. Relying on learning the joint representation of seen compositions, these methods ignore the explicit modeling of the state and object, thus limiting the exploitation of pre-trained knowledge and generalization to unseen compositions. With a particular focus on the universality of the solution, in this work, we propose a novel paradigm for CZSL models that establishes three identification branches (i.e., Multi-Path) to jointly model the state, object, and composition. The presented Troika is our implementation that aligns the branch-specific prompt representations with decomposed visual features. To calibrate the bias between semantically similar multi-modal representations, we further devise a Cross-Modal Traction module into Troika that shifts the prompt representation towards the current visual content. We conduct extensive experiments on three popular benchmarks, where our method significantly outperforms existing methods in both closed-world and open-world settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel paradigm for CZSL models that establishes three identification branches (i.e., Multi-Path) to jointly model the state, object, and composition and conducts extensive experiments on three popular benchmarks, where this method significantly outperforms existing methods in both closed-world and open-world settings.'}",http://arxiv.org/pdf/2303.15230
-Zero-Shot Information Extraction for Clinical Meta-Analysis using Large Language Models,David Kartchner,"Meta-analysis of randomized clinical trials (RCTs) plays a crucial role in evidence-based medicine but can be labor-intensive and error-prone. This study explores the use of large language models to enhance the efficiency of aggregating results from randomized clinical trials (RCTs) at scale. We perform a detailed comparison of the performance of these models in zero-shot prompt-based information extraction from a diverse set of RCTs to traditional manual annotation methods. We analyze the results for two different meta-analyses aimed at drug repurposing in cancer therapy pharmacovigilience in chronic myeloid leukemia. Our findings reveal that the best model for the two demonstrated tasks, ChatGPT can generally extract correct information and identify when the desired information is missing from an article. We additionally conduct a systematic error analysis, documenting the prevalence of diverse error types encountered during the process of prompt-based information extraction.","{'model': 'tldr@v2.0.0', 'text': 'The use of large language models to enhance the efficiency of aggregating results from randomized clinical trials (RCTs) at scale is explored and the best model, ChatGPT, is revealed to be the best for the two demonstrated tasks.'}",https://aclanthology.org/2023.bionlp-1.37.pdf
-Vision-Language Models are Zero-Shot Reward Models for Reinforcement Learning,Juan Rocamonde,"Reinforcement learning (RL) requires either manually specifying a reward function, which is often infeasible, or learning a reward model from a large amount of human feedback, which is often very expensive. We study a more sample-efficient alternative: using pretrained vision-language models (VLMs) as zero-shot reward models (RMs) to specify tasks via natural language. We propose a natural and general approach to using VLMs as reward models, which we call VLM-RMs. We use VLM-RMs based on CLIP to train a MuJoCo humanoid to learn complex tasks without a manually specified reward function, such as kneeling, doing the splits, and sitting in a lotus position. For each of these tasks, we only provide a single sentence text prompt describing the desired task with minimal prompt engineering. We provide videos of the trained agents at: https://sites.google.com/view/vlm-rm. We can improve performance by providing a second ``baseline'' prompt and projecting out parts of the CLIP embedding space irrelevant to distinguish between goal and baseline. Further, we find a strong scaling effect for VLM-RMs: larger VLMs trained with more compute and data are better reward models. The failure modes of VLM-RMs we encountered are all related to known capability limitations of current VLMs, such as limited spatial reasoning ability or visually unrealistic environments that are far off-distribution for the VLM. We find that VLM-RMs are remarkably robust as long as the VLM is large enough. This suggests that future VLMs will become more and more useful reward models for a wide range of RL applications.","{'model': 'tldr@v2.0.0', 'text': 'It is found that VLM-RMs are remarkably robust as long as the VLM is large enough, which suggests that future VLMs will become more and more useful reward models for a wide range of RL applications.'}",
-TongueSAM: An Universal Tongue Segmentation Model Based on SAM with Zero-Shot,Shan Cao,"Tongue segmentation serves as the primary step in automated TCM tongue diagnosis, which plays a significant role in the diagnostic results. Currently, numerous deep learning based methods have achieved promising results. However, when confronted with tongue images that differ from the training set or possess challenging backgrounds, these methods demonstrate limited performance. To address this issue, this paper proposes a universal tongue segmentation model named TongueSAM based on SAM (Segment Anything Model). SAM is a large-scale pretrained interactive segmentation model known for its powerful zero-shot generalization capability. Applying SAM to tongue segmentation leverages its learned prior knowledge from natural images, enabling the achievement of zero-shot segmentation for various types of tongue images. In this study, a Prompt Generator based on object detection is integrated into SAM to enable an end-to-end automated tongue segmentation method. Experiments demonstrate that TongueSAM achieves exceptional performance across various of tongue segmentation datasets, particularly under zero-shot. Even when dealing with challenging background tongue images, TongueSAM achieves a mIoU of 95.23\% under zero-shot conditions, surpassing other segmentation methods. As far as we know, this is the first application of large-scale pretrained model for tongue segmentation. The project and pretrained model will be made public when the paper is accepted.","{'model': 'tldr@v2.0.0', 'text': 'Tongue segmentation serves as the primary step in automated TCM tongue diagnosis, which plays a significant role in the diagnostic results, and a large-scale pretrained interactive segmentation model based on SAM (Segment Anything Model) is proposed, which achieves exceptional performance across various of tongue segmentation datasets, particularly under zero-shot.'}",https://arxiv.org/pdf/2308.06444
-Zero-Shot Text Classification via Self-Supervised Tuning,Chaoqun Liu,"Existing solutions to zero-shot text classification either conduct prompting with pre-trained language models, which is sensitive to the choices of templates, or rely on large-scale annotated data of relevant tasks for meta-tuning. In this work, we propose a new paradigm based on self-supervised learning to solve zero-shot text classification tasks by tuning the language models with unlabeled data, called self-supervised tuning. By exploring the inherent structure of free texts, we propose a new learning objective called first sentence prediction to bridge the gap between unlabeled data and text classification tasks. After tuning the model to learn to predict the first sentence in a paragraph based on the rest, the model is able to conduct zero-shot inference on unseen tasks such as topic classification and sentiment analysis. Experimental results show that our model outperforms the state-of-the-art baselines on 7 out of 10 tasks. Moreover, the analysis reveals that our model is less sensitive to the prompt design. Our code and pre-trained models are publicly available at https://github.com/DAMO-NLP-SG/SSTuning .","{'model': 'tldr@v2.0.0', 'text': 'After tuning the model to learn to predict the first sentence in a paragraph based on the rest, the model is able to conduct zero-shot inference on unseen tasks such as topic classification and sentiment analysis and outperforms the state-of-the-art baselines on 7 out of 10 tasks.'}",http://arxiv.org/pdf/2305.11442
-Text-Conditional Contextualized Avatars For Zero-Shot Personalization,S. Azadi,"Recent large-scale text-to-image generation models have made significant improvements in the quality, realism, and diversity of the synthesized images and enable users to control the created content through language. However, the personalization aspect of these generative models is still challenging and under-explored. In this work, we propose a pipeline that enables personalization of image generation with avatars capturing a user's identity in a delightful way. Our pipeline is zero-shot, avatar texture and style agnostic, and does not require training on the avatar at all - it is scalable to millions of users who can generate a scene with their avatar. To render the avatar in a pose faithful to the given text prompt, we propose a novel text-to-3D pose diffusion model trained on a curated large-scale dataset of in-the-wild human poses improving the performance of the SOTA text-to-motion models significantly. We show, for the first time, how to leverage large-scale image datasets to learn human 3D pose parameters and overcome the limitations of motion capture datasets.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes a pipeline that enables personalization of image generation with avatars capturing a user's identity in a delightful way and shows, for the first time, how to leverage large-scale image datasets to learn human 3D pose parameters and overcome the limitations of motion capture datasets.""}",http://arxiv.org/pdf/2304.07410
-Zero-Shot Co-salient Object Detection Framework,Haoke Xiao,"Co-salient Object Detection (CoSOD) endeavors to replicate the human visual system's capacity to recognize common and salient objects within a collection of images. Despite recent advancements in deep learning models, these models still rely on training with well-annotated CoSOD datasets. The exploration of training-free zero-shot CoSOD frameworks has been limited. In this paper, taking inspiration from the zero-shot transfer capabilities of foundational computer vision models, we introduce the first zero-shot CoSOD framework that harnesses these models without any training process. To achieve this, we introduce two novel components in our proposed framework: the group prompt generation (GPG) module and the co-saliency map generation (CMP) module. We evaluate the framework's performance on widely-used datasets and observe impressive results. Our approach surpasses existing unsupervised methods and even outperforms fully supervised methods developed before 2020, while remaining competitive with some fully supervised methods developed before 2022.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the first zero-shot CoSOD framework that harnesses foundational computer vision models without any training process, and surpasses existing unsupervised methods and even outperforms fully supervised methods developed before 2020, while remaining competitive with some fully supervised method developed before 2022.'}",https://arxiv.org/pdf/2309.05499
-Harnessing the Zero-Shot Power of Instruction-Tuned Large Language Model in End-to-End Speech Recognition,Yosuke Higuchi,"We present a novel integration of an instruction-tuned large language model (LLM) and end-to-end automatic speech recognition (ASR). Modern LLMs can perform a wide range of linguistic tasks within zero-shot learning when provided with a precise instruction or a prompt to guide the text generation process towards the desired task. We explore using this zero-shot capability of LLMs to extract linguistic information that can contribute to improving ASR performance. Specifically, we direct an LLM to correct grammatical errors in an ASR hypothesis and harness the embedded linguistic knowledge to conduct end-to-end ASR. The proposed model is built on the hybrid connectionist temporal classification (CTC) and attention architecture, where an instruction-tuned LLM (i.e., Llama2) is employed as a front-end of the decoder. An ASR hypothesis, subject to correction, is obtained from the encoder via CTC decoding, which is then fed into the LLM along with an instruction. The decoder subsequently takes as input the LLM embeddings to perform sequence generation, incorporating acoustic information from the encoder output. Experimental results and analyses demonstrate that the proposed integration yields promising performance improvements, and our approach largely benefits from LLM-based rescoring.","{'model': 'tldr@v2.0.0', 'text': 'A novel integration of an instruction-tuned large language model (LLM) and end-to-end automatic speech recognition (ASR) is presented, built on the hybrid connectionist temporal classification (CTC) and attention architecture.'}",https://arxiv.org/pdf/2309.10524
-Apollo: Zero-shot MultiModal Reasoning with Multiple Experts,Daniela Ben-David,"We propose a modular framework that leverages the expertise of different foundation models over different modalities and domains in order to perform a single, complex, multi-modal task, without relying on prompt engineering or otherwise tailor-made multi-modal training. Our approach enables decentralized command execution and allows each model to both contribute and benefit from the expertise of the other models. Our method can be extended to a variety of foundation models (including audio and vision), above and beyond only language models, as it does not depend on prompts. We demonstrate our approach on two tasks. On the well-known task of stylized image captioning, our experiments show that our approach outperforms semi-supervised state-of-the-art models, while being zero-shot and avoiding costly training, data collection, and prompt engineering. We further demonstrate this method on a novel task, audio-aware image captioning, in which an image and audio are given and the task is to generate text that describes the image within the context of the provided audio. Our code is available on GitHub.","{'model': 'tldr@v2.0.0', 'text': 'On the well-known task of stylized image captioning, the experiments show that this approach outperforms semi-supervised state-of-the-art models, while being zero-shot and avoiding costly training, data collection, and prompt engineering.'}",
-Align Your Prompts: Test-Time Prompting with Distribution Alignment for Zero-Shot Generalization,Jameel Hassan,"The promising zero-shot generalization of vision-language models such as CLIP has led to their adoption using prompt learning for numerous downstream tasks. Previous works have shown test-time prompt tuning using entropy minimization to adapt text prompts for unseen domains. While effective, this overlooks the key cause for performance degradation to unseen domains -- distribution shift. In this work, we explicitly handle this problem by aligning the out-of-distribution (OOD) test sample statistics to those of the source data using prompt tuning. We use a single test sample to adapt multi-modal prompts at test time by minimizing the feature distribution shift to bridge the gap in the test domain. Evaluating against the domain generalization benchmark, our method improves zero-shot top- 1 accuracy beyond existing prompt-learning techniques, with a 3.08% improvement over the baseline MaPLe. In cross-dataset generalization with unseen categories across 10 datasets, our method improves consistently across all datasets compared to the existing state-of-the-art. Our source code and models are available at https://jameelhassan.github.io/promptalign.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a single test sample to adapt multi-modal prompts at test time by minimizing the feature distribution shift to bridge the gap in the test domain, and improves zero-shot top- 1 accuracy beyond existing prompt-learning techniques.'}",
-Towards Realistic Zero-Shot Classification via Self Structural Semantic Alignment,Shengxiang Zhang,"Large-scale pre-trained Vision Language Models (VLMs) have proven effective for zero-shot classification. Despite the success, most traditional VLMs-based methods are restricted by the assumption of partial source supervision or ideal vocabularies, which rarely satisfy the open-world scenario. In this paper, we aim at a more challenging setting, Realistic Zero-Shot Classification, which assumes no annotation but instead a broad vocabulary. To address this challenge, we propose the Self Structural Semantic Alignment (S^3A) framework, which extracts the structural semantic information from unlabeled data while simultaneously self-learning. Our S^3A framework adopts a unique Cluster-Vote-Prompt-Realign (CVPR) algorithm, which iteratively groups unlabeled data to derive structural semantics for pseudo-supervision. Our CVPR process includes iterative clustering on images, voting within each cluster to identify initial class candidates from the vocabulary, generating discriminative prompts with large language models to discern confusing candidates, and realigning images and the vocabulary as structural semantic alignment. Finally, we propose to self-learn the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy. Our comprehensive experiments across various generic and fine-grained benchmarks demonstrate that the S^3A method offers substantial improvements over existing VLMs-based approaches, achieving a more than 15% accuracy improvement over CLIP on average. Our codes, models, and prompts are publicly released at https://github.com/sheng-eatamath/S3A.","{'model': 'tldr@v2.0.0', 'text': 'The Self Structural Semantic Alignment (S^3A) framework is proposed, which extracts the structural semantic information from unlabeled data while simultaneously self-learning and proposes to self-learn the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy.'}",https://arxiv.org/pdf/2308.12960
-Instruction Distillation Makes Large Language Models Efficient Zero-shot Rankers,Weiwei Sun,"Recent studies have demonstrated the great potential of Large Language Models (LLMs) serving as zero-shot relevance rankers. The typical approach involves making comparisons between pairs or lists of documents. Although effective, these listwise and pairwise methods are not efficient and also heavily rely on intricate prompt engineering. To tackle this problem, we introduce a novel instruction distillation method. The key idea is to distill the pairwise ranking ability of open-sourced LLMs to a simpler but more efficient pointwise ranking. Specifically, given the same LLM, we first rank documents using the effective pairwise approach with complex instructions, and then distill the teacher predictions to the pointwise approach with simpler instructions. Evaluation results on the BEIR, TREC, and ReDial datasets demonstrate that instruction distillation can improve efficiency by 10 to 100x and also enhance the ranking performance of LLMs. Furthermore, our approach surpasses the performance of existing supervised methods like monoT5 and is on par with the state-of-the-art zero-shot methods. The code to reproduce our results is available at www.github.com/sunnweiwei/RankGPT.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel instruction distillation method that surpasses the performance of existing supervised methods like monoT5 and is on par with the state-of-the-art zero-shot methods.'}",
-Masked Momentum Contrastive Learning for Zero-shot Semantic Understanding,Jiantao Wu,"Self-supervised pretraining (SSP) has emerged as a popular technique in machine learning, enabling the extraction of meaningful feature representations without labelled data. In the realm of computer vision, pretrained vision transformers (ViTs) have played a pivotal role in advancing transfer learning. Nonetheless, the escalating cost of finetuning these large models has posed a challenge due to the explosion of model size. This study endeavours to evaluate the effectiveness of pure self-supervised learning (SSL) techniques in computer vision tasks, obviating the need for finetuning, with the intention of emulating human-like capabilities in generalisation and recognition of unseen objects. To this end, we propose an evaluation protocol for zero-shot segmentation based on a prompting patch. Given a point on the target object as a prompt, the algorithm calculates the similarity map between the selected patch and other patches, upon that, a simple thresholding is applied to segment the target. Another evaluation is intra-object and inter-object similarity to gauge discriminatory ability of SSP ViTs. Insights from zero-shot segmentation from prompting and discriminatory abilities of SSP led to the design of a simple SSP approach, termed MMC. This approaches combines Masked image modelling for encouraging similarity of local features, Momentum based self-distillation for transferring semantics from global to local features, and global Contrast for promoting semantics of global features, to enhance discriminative representations of SSP ViTs. Consequently, our proposed method significantly reduces the overlap of intra-object and inter-object similarities, thereby facilitating effective object segmentation within an image. Our experiments reveal that MMC delivers top-tier results in zero-shot semantic segmentation across various datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study endeavours to evaluate the effectiveness of pure self-supervised learning (SSL) techniques in computer vision tasks, obviating the need for finetuning, with the intention of emulating human-like capabilities in generalisation and recognition of unseen objects.'}",https://arxiv.org/pdf/2308.11448
-Locally Differentially Private Document Generation Using Zero Shot Prompting,Saiteja Utpala,"Numerous studies have highlighted the privacy risks associated with pretrained large language models. In contrast, our research offers a unique perspective by demonstrating that pretrained large language models can effectively contribute to privacy preservation. We propose a locally differentially private mechanism called DP-Prompt, which leverages the power of pretrained large language models and zero-shot prompting to counter author de-anonymization attacks while minimizing the impact on downstream utility. When DP-Prompt is used with a powerful language model like ChatGPT (gpt-3.5), we observe a notable reduction in the success rate of de-anonymization attacks, showing that it surpasses existing approaches by a considerable margin despite its simpler design. For instance, in the case of the IMDB dataset, DP-Prompt (with ChatGPT) perfectly recovers the clean sentiment F1 score while achieving a 46\% reduction in author identification F1 score against static attackers and a 26\% reduction against adaptive attackers. We conduct extensive experiments across six open-source large language models, ranging up to 7 billion parameters, to analyze various effects of the privacy-utility tradeoff.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a locally differentially private mechanism called DP-Prompt, which leverages the power of pretrained large language models and zero-shot prompting to counter author de-anonymization attacks while minimizing the impact on downstream utility.'}",
-Zero-Shot Relation Triple Extraction with Prompts for Low-Resource Languages,Ayiguli Halike,"Although low-resource relation extraction is vital in knowledge construction and characterization, more research is needed on the generalization of unknown relation types. To fill the gap in the study of low-resource (Uyghur) relation extraction methods, we created a zero-shot with a quick relation extraction task setup. Each triplet extracted from an input phrase consists of the subject, relation type, and object. This paper suggests generating structured texts by urging language models to provide related instances. Our model consists of two modules: relation generator and relation and triplet extractor. We use the Uyghur relation prompt in the relation generator stage to generate new synthetic data. In the relation and triple extraction stage, we use the new data to extract the relation triplets in the sentence. We use multi-language model prompts and structured text techniques to offer a structured relation prompt template. This method is the first research that extends relation triplet extraction to a zero-shot setting for Uyghur datasets. Experimental results show that our method achieves a maximum weighted average F1 score of 47.39%.","{'model': 'tldr@v2.0.0', 'text': 'This method is the first research that extends relation triplet extraction to a zero-shot setting for Uyghur datasets and uses multi-language model prompts and structured text techniques to offer a structured relation prompt template.'}",https://www.mdpi.com/2076-3417/13/7/4636/pdf?version=1681110517
-AutoCLIP: Auto-tuning Zero-Shot Classifiers for Vision-Language Models,J. H. Metzen,"Classifiers built upon vision-language models such as CLIP have shown remarkable zero-shot performance across a broad range of image classification tasks. Prior work has studied different ways of automatically creating descriptor sets for every class based on prompt templates, ranging from manually engineered templates over templates obtained from a large language model to templates built from random words and characters. Up until now, deriving zero-shot classifiers from the respective encoded class descriptors has remained nearly unchanged, i.e., classify to the class that maximizes cosine similarity between its averaged encoded class descriptors and the image encoding. However, weighing all class descriptors equally can be suboptimal when certain descriptors match visual clues on a given image better than others. In this work, we propose AutoCLIP, a method for auto-tuning zero-shot classifiers. AutoCLIP tunes per-image weights to each prompt template at inference time, based on statistics of class descriptor-image similarities. AutoCLIP is fully unsupervised, has very low computational overhead, and can be easily implemented in few lines of code. We show that AutoCLIP outperforms baselines across a broad range of vision-language models, datasets, and prompt templates consistently and by up to 3 percent point accuracy.","{'model': 'tldr@v2.0.0', 'text': 'AutoCLIP tunes per-image weights to each prompt template at inference time, based on statistics of class descriptor-image similarities, and is shown to outperforms baselines across a broad range of vision-language models, datasets, and prompt templates consistently and by up to 3 percent point accuracy.'}",https://arxiv.org/pdf/2309.16414
-Instance Needs More Care: Rewriting Prompts for Instances Yields Better Zero-Shot Performance,Saurabh Srivastava,"Enabling large language models (LLMs) to perform tasks in zero-shot has been an appealing goal owing to its labor-saving (i.e., requiring no task-specific annotations); as such, zero-shot prompting approaches also enjoy better task generalizability. To improve LLMs' zero-shot performance, prior work has focused on devising more effective task instructions (e.g., ``let's think step by step'' ). However, we argue that, in order for an LLM to solve them correctly in zero-shot, individual test instances need more carefully designed and customized instructions. To this end, we propose PRoMPTd, an approach that rewrites the task prompt for each individual test input to be more specific, unambiguous, and complete, so as to provide better guidance to the task LLM. We evaluated PRoMPTd on eight datasets covering tasks including arithmetics, logical reasoning, and code generation, using GPT-4 as the task LLM. Notably, PRoMPTd achieves an absolute improvement of around 10% on the complex MATH dataset and 5% on the code generation task on HumanEval, outperforming conventional zero-shot methods. In addition, we also showed that the rewritten prompt can provide better interpretability of how the LLM resolves each test instance, which can potentially be leveraged as a defense mechanism against adversarial prompting. The source code and dataset can be obtained from https://github.com/salokr/PRoMPTd","{'model': 'tldr@v2.0.0', 'text': 'PRoMPTd, an approach that rewrites the task prompt for each individual test input to be more specific, unambiguous, and complete, so as to provide better guidance to the task LLM, is proposed.'}",https://arxiv.org/pdf/2310.02107
-ZYN: Zero-Shot Reward Models with Yes-No Questions,Víctor Gallego,"In this work, we address the problem of directing the text generations of a LLM towards a desired behavior, aligning the generated text with the preferences of the human operator. We propose using another language model as a critic, reward model in a zero-shot way thanks to the prompt of a Yes-No question that represents the user preferences, without requiring further labeled data. This zero-shot reward model provides the learning signal to further fine-tune the base LLM using reinforcement learning, as in RLAIF; yet our approach is also compatible in other contexts such as quality-diversity search. Extensive evidence of the capabilities of the proposed ZYN framework is provided through experiments in different domains related to text generation, including detoxification; optimizing sentiment of movie reviews, or any other attribute; steering the opinion about a particular topic the model may have; and personalizing prompt generators for text-to-image tasks. Code to be released at \url{https://github.com/vicgalle/zero-shot-reward-models/}.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using another language model as a critic, reward model in a zero-shot way thanks to the prompt of a Yes-No question that represents the user preferences, without requiring further labeled data.'}",https://arxiv.org/pdf/2308.06385
-Random Word Data Augmentation with CLIP for Zero-Shot Anomaly Detection,Masato Tamura,"This paper presents a novel method that leverages a visual-language model, CLIP, as a data source for zero-shot anomaly detection. Tremendous efforts have been put towards developing anomaly detectors due to their potential industrial applications. Considering the difficulty in acquiring various anomalous samples for training, most existing methods train models with only normal samples and measure discrepancies from the distribution of normal samples during inference, which requires training a model for each object category. The problem of this inefficient training requirement has been tackled by designing a CLIP-based anomaly detector that applies prompt-guided classification to each part of an image in a sliding window manner. However, the method still suffers from the labor of careful prompt ensembling with known object categories. To overcome the issues above, we propose leveraging CLIP as a data source for training. Our method generates text embeddings with the text encoder in CLIP with typical prompts that include words of normal and anomaly. In addition to these words, we insert several randomly generated words into prompts, which enables the encoder to generate a diverse set of normal and anomalous samples. Using the generated embeddings as training data, a feed-forward neural network learns to extract features of normal and anomaly from CLIP's embeddings, and as a result, a category-agnostic anomaly detector can be obtained without any training images. Experimental results demonstrate that our method achieves state-of-the-art performance without laborious prompt ensembling in zero-shot setups.","{'model': 'tldr@v2.0.0', 'text': ""A feed-forward neural network learns to extract features of normal and anomaly from CLIP's embeddings, and as a result, a category-agnostic anomaly detector can be obtained without any training images.""}",https://arxiv.org/pdf/2308.11119
-Supplementary - I2MVFormer: Large Language Model Generated Multi-View Document Supervision for Zero-Shot Image Classification,Muhammad Ferjad Naeem,"In this supplementary, we perform additional experiments on I2MVFormer to confirm our design choices. Moreover, we perform further analysis of our LLM prompting strategy to confirm that it is robust to the choice of LLM, the choice of the prompt and the choice of k-shot examples. Finally, we qualitatively analyze the LLM Views to confirm that each view provides complementary information about a class which allows for a highly discriminative zero-shot image classification model. The content of the supplementary is organized as follows.","{'model': 'tldr@v2.0.0', 'text': 'This supplementary performs additional experiments on I2MVFormer and qualitatively analyzes the LLM Views to confirm that each view provides complementary information about a class which allows for a highly discriminative zero-shot image classification model.'}",
-Zero-shot Approach to Overcome Perturbation Sensitivity of Prompts,Mohna Chakraborty,"Recent studies have demonstrated that natural-language prompts can help to leverage the knowledge learned by pre-trained language models for the binary sentence-level sentiment classification task. Specifically, these methods utilize few-shot learning settings to fine-tune the sentiment classification model using manual or automatically generated prompts. However, the performance of these methods is sensitive to the perturbations of the utilized prompts. Furthermore, these methods depend on a few labeled instances for automatic prompt generation and prompt ranking. This study aims to find high-quality prompts for the given task in a zero-shot setting. Given a base prompt, our proposed approach automatically generates multiple prompts similar to the base prompt employing positional, reasoning, and paraphrasing techniques and then ranks the prompts using a novel metric. We empirically demonstrate that the top-ranked prompts are high-quality and significantly outperform the base prompt and the prompts generated using few-shot learning for the binary sentence-level sentiment classification task.","{'model': 'tldr@v2.0.0', 'text': 'This study aims to find high-quality prompts for the given task in a zero-shot setting and empirically demonstrates that the top-ranked prompts are high- quality and significantly outperform the base prompt and the prompts generated using few-shot learning for the binary sentence-level sentiment classification task.'}",http://arxiv.org/pdf/2305.15689
-A Setwise Approach for Effective and Highly Efficient Zero-shot Ranking with Large Language Models,Shengyao Zhuang,"Large Language Models (LLMs) demonstrate impressive effectiveness in zero-shot document ranking tasks. Pointwise, Pairwise, and Listwise prompting approaches have been proposed for LLM-based zero-shot ranking. Our study begins by thoroughly evaluating these existing approaches within a consistent experimental framework, considering factors like model size, token consumption, latency, among others. This first-of-its-kind comparative evaluation of these approaches allows us to identify the trade-offs between effectiveness and efficiency inherent in each approach. We find that while Pointwise approaches score high on efficiency, they suffer from poor effectiveness. Conversely, Pairwise approaches demonstrate superior effectiveness but incur high computational overhead. To further enhance the efficiency of LLM-based zero-shot ranking, we propose a novel Setwise prompting approach. Our approach reduces the number of LLM inferences and the amount of prompt token consumption during the ranking procedure, significantly improving the efficiency of LLM-based zero-shot ranking. We test our method using the TREC DL datasets and the BEIR zero-shot document ranking benchmark. The empirical results indicate that our approach considerably reduces computational costs while also retaining high zero-shot ranking effectiveness.","{'model': 'tldr@v2.0.0', 'text': 'This first-of-its-kind comparative evaluation of Pointwise, Pairwise, and Listwise prompting approaches for LLM-based zero-shot ranking finds that while Pointwise approaches score high on efficiency, they suffer from poor effectiveness and Pairwise approaches demonstrate superior effectiveness but incur high computational overhead.'}",
-Reducing Negative Effects of the Biases of Language Models in Zero-Shot Setting,Xiaosu Wang,"Pre-trained language models (PLMs) such as GPTs have been revealed to be biased towards certain target classes because of the prompt and the model's intrinsic biases. In contrast to the fully supervised scenario where there are a large number of costly labeled samples that can be used to fine-tune model parameters to correct for biases, there are no labeled samples available for the zero-shot setting. We argue that a key to calibrating the biases of a PLM on a target task in zero-shot setting lies in detecting and estimating the biases, which remains a challenge. In this paper, we first construct probing samples with the randomly generated token sequences, which are simple but effective in detecting inputs for stimulating GPTs to show the biases; and we pursue an in-depth research on the plausibility of utilizing class scores for the probing samples to reflect and estimate the biases of GPTs on a downstream target task. Furtherly, in order to effectively utilize the probing samples and thus reduce negative effects of the biases of GPTs, we propose a lightweight model Calibration Adapter (CA) along with a self-guided training strategy that carries out distribution-level optimization, which enables us to take advantage of the probing samples to fine-tune and select only the proposed CA, respectively, while keeping the PLM encoder frozen. To demonstrate the effectiveness of our study, we have conducted extensive experiments, where the results indicate that the calibration ability acquired by CA on the probing samples can be successfully transferred to reduce negative effects of the biases of GPTs on a downstream target task, and our approach can yield better performance than state-of-the-art (SOTA) models in zero-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the calibration ability acquired by CA on the probing samples can be successfully transferred to reduce negative effects of the biases of GPTs on a downstream target task, and the approach can yield better performance than state-of-the-art (SOTA) models in zero-shot settings.'}",
-Improving Language Model-Based Zero-Shot Text-to-Speech Synthesis with Multi-Scale Acoustic Prompts,Shunwei Lei,"Zero-shot text-to-speech (TTS) synthesis aims to clone any unseen speaker's voice without adaptation parameters. By quantizing speech waveform into discrete acoustic tokens and modeling these tokens with the language model, recent language model-based TTS models show zero-shot speaker adaptation capabilities with only a 3-second acoustic prompt of an unseen speaker. However, they are limited by the length of the acoustic prompt, which makes it difficult to clone personal speaking style. In this paper, we propose a novel zero-shot TTS model with the multi-scale acoustic prompts based on a neural codec language model VALL-E. A speaker-aware text encoder is proposed to learn the personal speaking style at the phoneme-level from the style prompt consisting of multiple sentences. Following that, a VALL-E based acoustic decoder is utilized to model the timbre from the timbre prompt at the frame-level and generate speech. The experimental results show that our proposed method outperforms baselines in terms of naturalness and speaker similarity, and can achieve better performance by scaling out to a longer style prompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel zero-shot TTS model with the multi-scale acoustic prompts based on a neural codec language model VALL-E, which outperforms baselines in terms of naturalness and speaker similarity, and can achieve better performance by scaling out to a longer style prompt.'}",https://arxiv.org/pdf/2309.11977
-Model-Generated Pretraining Signals Improves Zero-Shot Generalization of Text-to-Text Transformers,Linyuan Gong,"This paper explores the effectiveness of model-generated signals in improving zero-shot generalization of text-to-text Transformers such as T5. We study various designs to pretrain T5 using an auxiliary model to construct more challenging token replacements for the main model to denoise. Key aspects under study include the decoding target, the location of the RTD head, and the masking pattern. Based on these studies, we develop a new model, METRO-T0, which is pretrained using the redesigned ELECTRA-Style pretraining strategies and then prompt-finetuned on a mixture of NLP tasks. METRO-T0 outperforms all similar-sized baselines on prompted NLP benchmarks, such as _T0 Eval_ and MMLU, and rivals the state-of-the-art T0-11B model with only **8%** of its parameters. Our analysis on model’s neural activation and parameter sensitivity reveals that the effectiveness of METRO-T0 stems from more balanced contribution of parameters and better utilization of their capacity. The code and model checkpoints are available at [https://github.com/gonglinyuan/metro_t0](https://github.com/gonglinyuan/metro_t0).","{'model': 'tldr@v2.0.0', 'text': 'A new model, METRO-T0 is developed, which is pretrained using the redesigned ELECTRA-Style pretraining strategies and then prompt-finetuned on a mixture of NLP tasks and rivals the state-of-the-art T0-11B model with only **8%** of its parameters.'}",http://arxiv.org/pdf/2305.12567
-Interaction-Aware Prompting for Zero-Shot Spatio-Temporal Action Detection,Weiting Huang,"The goal of spatial-temporal action detection is to determine the time and place where each person's action occurs in a video and classify the corresponding action category. Most of the existing methods adopt fully-supervised learning, which requires a large amount of training data, making it very difficult to achieve zero-shot learning. In this paper, we propose to utilize a pre-trained visual-language model to extract the representative image and text features, and model the relationship between these features through different interaction modules to obtain the interaction feature. In addition, we use this feature to prompt each label to obtain more appropriate text features. Finally, we calculate the similarity between the interaction feature and the text feature for each label to determine the action category. Our experiments on J-HMDB and UCF101-24 datasets demonstrate that the proposed interaction module and prompting make the visual-language features better aligned, thus achieving excellent accuracy for zero-shot spatio-temporal action detection. The code will be available at https://github.com/webber2933/iCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A pre-trained visual-language model is utilized to extract the representative image and text features, and model the relationship between these features through different interaction modules to obtain the interaction feature, which is used to prompt each label to obtain more appropriate text features.'}",https://arxiv.org/pdf/2304.04688
-Zero-Shot and Few-Shot Video Question Answering with Multi-Modal Prompts,Deniz Engin,"Recent vision-language models are driven by large-scale pretrained models. However, adapting pretrained models on limited data presents challenges such as overfitting, catastrophic forgetting, and the cross-modal gap between vision and language. We introduce a parameter-efficient method to address these challenges, combining multimodal prompt learning and a transformer-based mapping network, while keeping the pretrained models frozen. Our experiments on several video question answering benchmarks demonstrate the superiority of our approach in terms of performance and parameter efficiency on both zero-shot and few-shot settings. Our code is available at https://engindeniz.github.io/vitis.","{'model': 'tldr@v2.0.0', 'text': 'A parameter-efficient method is introduced, combining multimodal prompt learning and a transformer-based mapping network, while keeping the pretrained models frozen to address challenges such as overfitting, catastrophic forgetting, and the cross-modal gap between vision and language.'}",https://arxiv.org/pdf/2309.15915
-Beyond Yes and No: Improving Zero-Shot LLM Rankers via Scoring Fine-Grained Relevance Labels,Honglei Zhuang,"Zero-shot text rankers powered by recent LLMs achieve remarkable ranking performance by simply prompting. Existing prompts for pointwise LLM rankers mostly ask the model to choose from binary relevance labels like""Yes""and""No"". However, the lack of intermediate relevance label options may cause the LLM to provide noisy or biased answers for documents that are partially relevant to the query. We propose to incorporate fine-grained relevance labels into the prompt for LLM rankers, enabling them to better differentiate among documents with different levels of relevance to the query and thus derive a more accurate ranking. We study two variants of the prompt template, coupled with different numbers of relevance levels. Our experiments on 8 BEIR data sets show that adding fine-grained relevance labels significantly improves the performance of LLM rankers.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to incorporate fine-grained relevance labels into the prompt for LLM rankers, enabling them to better differentiate among documents with different levels of relevance to the query and thus derive a more accurate ranking.'}",
-Zero-shot Inversion Process for Image Attribute Editing with Diffusion Models,Zhanbo Feng,"Denoising diffusion models have shown outstanding performance in image editing. Existing works tend to use either image-guided methods, which provide a visual reference but lack control over semantic coherence, or text-guided methods, which ensure faithfulness to text guidance but lack visual quality. To address the problem, we propose the Zero-shot Inversion Process (ZIP), a framework that injects a fusion of generated visual reference and text guidance into the semantic latent space of a \textit{frozen} pre-trained diffusion model. Only using a tiny neural network, the proposed ZIP produces diverse content and attributes under the intuitive control of the text prompt. Moreover, ZIP shows remarkable robustness for both in-domain and out-of-domain attribute manipulation on real images. We perform detailed experiments on various benchmark datasets. Compared to state-of-the-art methods, ZIP produces images of equivalent quality while providing a realistic editing effect.","{'model': 'tldr@v2.0.0', 'text': 'The Zero-shot Inversion Process (ZIP), a framework that injects a fusion of generated visual reference and text guidance into the semantic latent space of a pre-trained diffusion model, shows remarkable robustness for both in-domain and out-of-domain attribute manipulation on real images.'}",https://arxiv.org/pdf/2308.15854
-Exploring Grounding Potential of VQA-oriented GPT-4V for Zero-shot Anomaly Detection,Jiangning Zhang,"Large Multimodal Model (LMM) GPT-4V(ision) endows GPT-4 with visual grounding capabilities, making it possible to handle certain tasks through the Visual Question Answering (VQA) paradigm. This paper explores the potential of VQA-oriented GPT-4V in the recently popular visual Anomaly Detection (AD) and is the first to conduct qualitative and quantitative evaluations on the popular MVTec AD and VisA datasets. Considering that this task requires both image-/pixel-level evaluations, the proposed GPT-4V-AD framework contains three components: 1) Granular Region Division, 2) Prompt Designing, 3) Text2Segmentation for easy quantitative evaluation, and have made some different attempts for comparative analysis. The results show that GPT-4V can achieve certain results in the zero-shot AD task through a VQA paradigm, such as achieving image-level 77.1/88.0 and pixel-level 68.0/76.6 AU-ROCs on MVTec AD and VisA datasets, respectively. However, its performance still has a certain gap compared to the state-of-the-art zero-shot method, e.g., WinCLIP ann CLIP-AD, and further research is needed. This study provides a baseline reference for the research of VQA-oriented LMM in the zero-shot AD task, and we also post several possible future works. Code is available at \url{https://github.com/zhangzjn/GPT-4V-AD}.","{'model': 'tldr@v2.0.0', 'text': 'The results show that GPT-4V can achieve certain results in the zero-shot AD task through a VQA paradigm, such as achieving image-level 77.1/88.0 and pixel-level 68.0/76.6 AU-ROCs on MVTec AD and VisA datasets, respectively.'}",
-Zero-shot Learning for Named Entity Recognition in Software Specification Documents,Souvick Das,"Named entity recognition (NER) is a natural language processing task that has been used in Requirements Engineering for the identification of entities such as actors, actions, operators, resources, events, GUI elements, hardware, APIs, and others. NER might be particularly useful for extracting key information from Software Requirements Specification documents, which provide a blueprint for software development. However, a common challenge in this domain is the lack of annotated data. In this article, we propose and analyze two zero-shot approaches for NER in the requirements engineering domain. These are found to be particularly effective in situations where labeled data is scarce or non-existent. The first approach is a template-based zero-shot learning mechanism that uses the prompt engineering approach and achieves 93% accuracy according to our experimental results. The second solution takes an orthogonal approach by transforming the entity recognition problem into a question-answering task which results in 98% accuracy. Both zero-shot NER approaches introduced in this work perform better than the existing state-of-the-art solutions in the requirements engineering domain.","{'model': 'tldr@v2.0.0', 'text': 'Two zero-shot NER approaches for NER in the requirements engineering domain are proposed and analyzed and are found to be particularly effective in situations where labeled data is scarce or non-existent.'}",
-Zero-TextCap: Zero-shot Framework for Text-based Image Captioning,Dongsheng Xu,"Text-based image captioning is a vital but under-explored task, which aims to describe images by captions containing scene text automatically. Recent studies have made encouraging progress, but they are still suffering from two issues. Firstly, current models cannot capture and generate scene text in non-Latin script languages, which severely limits the objectivity and the information completeness of generated captions. Secondly, current models tend to describe images with monotonous and templated style, which greatly limits the diversity of the generated captions. Although the above-mentioned issues can be alleviated through carefully designed annotations, this process is undoubtedly laborious and time-consuming. To address the above issues, we propose a Zero-shot Framework for Text-based Image Captioning (Zero-TextCap). Concretely, to generate candidate sentences starting from the prompt 'Image of' and iteratively refine them to improve the quality and diversity of captions, we introduce a Hybrid-sampling masked language model (H-MLM). To read multi-lingual scene text and model the relationships between them, we introduce a robust OCR system. To ensure that the captions generated by H-MLM contain scene text and are highly relevant to the image, we propose a CLIP-based generation guidance module to insert OCR tokens and filter candidate sentences. Our Zero-TextCap is capable of generalizing captions containing multi-lingual scene text and boosting the diversity of captions. Sufficient experiments demonstrate the effectiveness of our proposed Zero-TextCap. Our codes are available at https://github.com/Gemhuang79/Zero_TextCap.","{'model': 'tldr@v2.0.0', 'text': 'The Zero-TextCap is capable of generalizing captions containing multi-lingual scene text and boosting the diversity of captions, and introduces a Hybrid-sampling masked language model (H-MLM) and a robust OCR system.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612571
-Are Soft Prompts Good Zero-shot Learners for Speech Recognition?,Dianwen Ng,"Large self-supervised pre-trained speech models require computationally expensive fine-tuning for downstream tasks. Soft prompt tuning offers a simple parameter-efficient alternative by utilizing minimal soft prompt guidance, enhancing portability while also maintaining competitive performance. However, not many people understand how and why this is so. In this study, we aim to deepen our understanding of this emerging method by investigating the role of soft prompts in automatic speech recognition (ASR). Our findings highlight their role as zero-shot learners in improving ASR performance but also make them vulnerable to malicious modifications. Soft prompts aid generalization but are not obligatory for inference. We also identify two primary roles of soft prompts: content refinement and noise information enhancement, which enhances robustness against background noise. Additionally, we propose an effective modification on noise prompts to show that they are capable of zero-shot learning on adapting to out-of-distribution noise environments.","{'model': 'tldr@v2.0.0', 'text': 'This study investigates the role of soft prompts in automatic speech recognition (ASR), and proposes an effective modification on noise prompts to show that they are capable of zero-shot learning on adapting to out-of-distribution noise environments.'}",https://arxiv.org/pdf/2309.09413
-Blended-NeRF: Zero-Shot Object Generation and Blending in Existing Neural Radiance Fields,Ori Gordon,"Editing a local region or a specific object in a 3D scene represented by a NeRF or consistently blending a new realistic object into the scene is challenging, mainly due to the implicit nature of the scene representation. We present Blended-NeRF, a robust and flexible framework for editing a specific region of interest in an existing NeRF scene, based on text prompts, along with a 3D ROI box. Our method leverages a pretrained language-image model to steer the synthesis towards a user-provided text prompt, along with a 3D MLP model initialized on an existing NeRF scene to generate the object and blend it into a specified region in the original scene. We allow local editing by localizing a 3D ROI box in the input scene, and blend the content synthesized inside the ROI with the existing scene using a novel volumetric blending technique. To obtain natural looking and view-consistent results, we leverage existing and new geometric priors and 3D augmentations for improving the visual fidelity of the final result. We test our framework both qualitatively and quantitatively on a variety of real 3D scenes and text prompts, demonstrating realistic multi-view consistent results with much flexibility and diversity compared to the baselines. Finally, we show the applicability of our framework for several 3D editing applications, including adding new objects to a scene, removing/replacing/altering existing objects, and texture conversion.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Blended-NeRF, a robust and flexible framework for editing a specific region of interest in an existing NeRF scene, based on text prompts, along with a 3D ROI box, and shows the applicability of the framework for several 3D editing applications, including adding new objects to a scene, removing/replacing/altering existing objects, and texture conversion.'}",https://arxiv.org/pdf/2306.12760
-Global Constraints with Prompting for Zero-Shot Event Argument Classification,Zizheng Lin,"Determining the role of event arguments is a crucial subtask of event extraction. Most previous supervised models leverage costly annotations, which is not practical for open-domain applications. In this work, we propose to use global constraints with prompting to effectively tackles event argument classification without any annotation and task-specific training. Specifically, given an event and its associated passage, the model first creates several new passages by prefix prompts and cloze prompts, where prefix prompts indicate event type and trigger span, and cloze prompts connect each candidate role with the target argument span. Then, a pre-trained language model scores the new passages, making the initial prediction. Our novel prompt templates can easily adapt to all events and argument types without manual effort. Next, the model regularizes the prediction by global constraints exploiting cross-task, cross-argument, and cross-event relations. Extensive experiments demonstrate our model’s effectiveness: it outperforms the best zero-shot baselines by 12.5% and 10.9% F1 on ACE and ERE with given argument spans and by 4.3% and 3.3% F1, respectively, without given argument spans. We have made our code publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use global constraints with prompting to effectively tackles event argument classification without any annotation and task-specific training, exploiting cross-task, cross-argument, and cross-event relations.'}",http://arxiv.org/pdf/2302.04459
-Large Language Models are Frame-level Directors for Zero-shot Text-to-Video Generation,Susung Hong,"In the paradigm of AI-generated content (AIGC), there has been increasing attention in extending pre-trained text-to-image (T2I) models to text-to-video (T2V) generation. Despite their effectiveness, these frameworks face challenges in maintaining consistent narratives and handling rapid shifts in scene composition or object placement from a single user prompt. This paper introduces a new framework, dubbed DirecT2V, which leverages instruction-tuned large language models (LLMs) to generate frame-by-frame descriptions from a single abstract user prompt. DirecT2V utilizes LLM directors to divide user inputs into separate prompts for each frame, enabling the inclusion of time-varying content and facilitating consistent video generation. To maintain temporal consistency and prevent object collapse, we propose a novel value mapping method and dual-softmax filtering. Extensive experimental results validate the effectiveness of the DirecT2V framework in producing visually coherent and consistent videos from abstract user prompts, addressing the challenges of zero-shot video generation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a new framework, dubbed DirecT2V, which leverages instruction-tuned large language models (LLMs) to generate frame-by-frame descriptions from a single abstract user prompt, and proposes a novel value mapping method and dual-softmax filtering.'}",http://arxiv.org/pdf/2305.14330
-SC VALL-E: Style-Controllable Zero-Shot Text to Speech Synthesizer,Daegyeom Kim,"Expressive speech synthesis models are trained by adding corpora with diverse speakers, various emotions, and different speaking styles to the dataset, in order to control various characteristics of speech and generate the desired voice. In this paper, we propose a style control (SC) VALL-E model based on the neural codec language model (called VALL-E), which follows the structure of the generative pretrained transformer 3 (GPT-3). The proposed SC VALL-E takes input from text sentences and prompt audio and is designed to generate controllable speech by not simply mimicking the characteristics of the prompt audio but by controlling the attributes to produce diverse voices. We identify tokens in the style embedding matrix of the newly designed style network that represent attributes such as emotion, speaking rate, pitch, and voice intensity, and design a model that can control these attributes. To evaluate the performance of SC VALL-E, we conduct comparative experiments with three representative expressive speech synthesis models: global style token (GST) Tacotron2, variational autoencoder (VAE) Tacotron2, and original VALL-E. We measure word error rate (WER), F0 voiced error (FVE), and F0 gross pitch error (F0GPE) as evaluation metrics to assess the accuracy of generated sentences. For comparing the quality of synthesized speech, we measure comparative mean option score (CMOS) and similarity mean option score (SMOS). To evaluate the style control ability of the generated speech, we observe the changes in F0 and mel-spectrogram by modifying the trained tokens. When using prompt audio that is not present in the training data, SC VALL-E generates a variety of expressive sounds and demonstrates competitive performance compared to the existing models. Our implementation, pretrained models, and audio samples are located on GitHub.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a style control (SC) VALL-E model based on the neural codec language model, which follows the structure of the generative pretrained transformer 3 (GPT-3), designed to generate controllable speech by not simply mimicking the characteristics of the prompt audio but by controlling the attributes to produce diverse voices.'}",https://arxiv.org/pdf/2307.10550
-Language Models as Zero-Shot Trajectory Generators,Teyun Kwon,"Large Language Models (LLMs) have recently shown promise as high-level planners for robots when given access to a selection of low-level skills. However, it is often assumed that LLMs do not possess sufficient knowledge to be used for the low-level trajectories themselves. In this work, we address this assumption thoroughly, and investigate if an LLM (GPT-4) can directly predict a dense sequence of end-effector poses for manipulation skills, when given access to only object detection and segmentation vision models. We study how well a single task-agnostic prompt, without any in-context examples, motion primitives, or external trajectory optimisers, can perform across 26 real-world language-based tasks, such as""open the bottle cap""and""wipe the plate with the sponge"", and we investigate which design choices in this prompt are the most effective. Our conclusions raise the assumed limit of LLMs for robotics, and we reveal for the first time that LLMs do indeed possess an understanding of low-level robot control sufficient for a range of common tasks, and that they can additionally detect failures and then re-plan trajectories accordingly. Videos, code, and prompts are available at: https://www.robot-learning.uk/language-models-trajectory-generators.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates if an LLM (GPT-4) can directly predict a dense sequence of end-effector poses for manipulation skills, when given access to only object detection and segmentation vision models, and reveals for the first time that LLMs do indeed possess an understanding of low-level robot control sufficient for a range of common tasks.'}",
-ZeroPrompt: Streaming Acoustic Encoders are Zero-Shot Masked LMs,Xingcheng Song,"In this paper, we present ZeroPrompt (Figure 1-(a)) and the corresponding Prompt-and-Refine strategy (Figure 3), two simple but effective \textbf{training-free} methods to decrease the Token Display Time (TDT) of streaming ASR models \textbf{without any accuracy loss}. The core idea of ZeroPrompt is to append zeroed content to each chunk during inference, which acts like a prompt to encourage the model to predict future tokens even before they were spoken. We argue that streaming acoustic encoders naturally have the modeling ability of Masked Language Models and our experiments demonstrate that ZeroPrompt is engineering cheap and can be applied to streaming acoustic encoders on any dataset without any accuracy loss. Specifically, compared with our baseline models, we achieve 350 $\sim$ 700ms reduction on First Token Display Time (TDT-F) and 100 $\sim$ 400ms reduction on Last Token Display Time (TDT-L), with theoretically and experimentally equal WER on both Aishell-1 and Librispeech datasets.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that streaming acoustic encoders naturally have the modeling ability of Masked Language Models and the experiments demonstrate that ZeroPrompt is engineering cheap and can be applied to streaming acousticencoders on any dataset without any accuracy loss.'}",
-Zero-shot Text-driven Physically Interpretable Face Editing,Yapeng Meng,"This paper proposes a novel and physically interpretable method for face editing based on arbitrary text prompts. Different from previous GAN-inversion-based face editing methods that manipulate the latent space of GANs, or diffusion-based methods that model image manipulation as a reverse diffusion process, we regard the face editing process as imposing vector flow fields on face images, representing the offset of spatial coordinates and color for each image pixel. Under the above-proposed paradigm, we represent the vector flow field in two ways: 1) explicitly represent the flow vectors with rasterized tensors, and 2) implicitly parameterize the flow vectors as continuous, smooth, and resolution-agnostic neural fields, by leveraging the recent advances of implicit neural representations. The flow vectors are iteratively optimized under the guidance of the pre-trained Contrastive Language-Image Pretraining~(CLIP) model by maximizing the correlation between the edited image and the text prompt. We also propose a learning-based one-shot face editing framework, which is fast and adaptable to any text prompt input. Our method can also be flexibly extended to real-time video face editing. Compared with state-of-the-art text-driven face editing methods, our method can generate physically interpretable face editing results with high identity consistency and image quality. Our code will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'Compared with state-of-the-art text-driven face editing methods, this method can generate physically interpretable face editing results with high identity consistency and image quality.'}",https://arxiv.org/pdf/2308.05976
-APPLeNet: Visual Attention Parameterized Prompt Learning for Few-Shot Remote Sensing Image Generalization using CLIP,M. Singha,"In recent years, the success of large-scale vision-language models (VLMs) such as CLIP has led to their increased usage in various computer vision tasks. These models enable zero-shot inference through carefully crafted instructional text prompts without task-specific supervision. However, the potential of VLMs for generalization tasks in remote sensing (RS) has not been fully realized. To address this research gap, we propose a novel image-conditioned prompt learning strategy called the Visual Attention Parameterized Prompts Learning Network (APPLeNet). APPLeNet emphasizes the importance of multi-scale feature learning in RS scene classification and disentangles visual style and content primitives for domain generalization tasks. To achieve this, APPLeNet combines visual content features obtained from different layers of the vision encoder and style properties obtained from feature statistics of domain-specific batches. An attention-driven injection module is further introduced to generate visual tokens from this information. We also introduce an anti-correlation regularizer to ensure discrimination among the token embeddings, as this visual information is combined with the textual tokens. To validate APPLeNet, we curated four available RS benchmarks and introduced experimental protocols and datasets for three domain generalization tasks. Our results consistently outperform the relevant literature and code is available at https://github.com/mainaksingha01/APPLeNet","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel image-conditioned prompt learning strategy called the Visual Attention Parameterized Prompts Learning Network (APPLeNet), which emphasizes the importance of multi-scale feature learning in RS scene classification and disentangles visual style and content primitives for domain generalization tasks.'}",https://arxiv.org/pdf/2304.05995
-WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation,Jongheon Jeong,"Visual anomaly classification and segmentation are vital for automating industrial quality inspection. The focus of prior research in the field has been on training custom models for each quality inspection task, which requires task-specific images and annotation. In this paper we move away from this regime, addressing zero-shot and few-normal-shot anomaly classification and segmentation. Recently CLIP, a vision-language model, has shown revolutionary generality with competitive zero-/few-shot performance in comparison to full-supervision. But CLIP falls short on anomaly classification and segmentation tasks. Hence, we propose window-based CLIP (WinCLIP) with (1) a compositional ensemble on state words and prompt templates and (2) efficient extraction and aggregation of window/patch/image-level features aligned with text. We also propose its few-normal-shot extension Win-CLIP+, which uses complementary information from normal images. In MVTec-AD (and VisA), without further tuning, WinCLIP achieves 91.8%/85.1% (78.1%/79.6%) AU-ROC in zero-shot anomaly classification and segmentation while WinCLIP + does 93.1%/95.2% (83.8%/96.4%) in 1-normal-shot, surpassing state-of-the-art by large margins.","{'model': 'tldr@v2.0.0', 'text': 'A window-based CLIP (WinCLIP) is proposed with a compositional ensemble on state words and prompt templates and efficient extraction and aggregation of window/patch/image-level features aligned with text, addressing zero-shot and few-normal-shot anomaly classification and segmentation.'}",https://arxiv.org/pdf/2303.14814
-Schema-aware Reference as Prompt Improves Data-Efficient Relational Triple and Event Extraction,Yunzhi Yao,"Information Extraction, which aims to extract structural relational triple or event from unstructured texts, often suffers from data scarcity issues. With the development of pre-trained language models, many prompt-based approaches to data-efficient information extraction have been proposed and achieved impressive performance. However, existing prompt learning methods for information extraction are still susceptible to several potential limitations: (i) semantic gap between natural language and output structure knowledge with pre-defined schema; (ii) representation learning with locally individual instances limits the performance given the insufficient features. In this paper, we propose a novel approach of schema-aware R eference A s P rompt ( RAP ), which dynamically leverage schema and knowledge inherited from global (few-shot) training data for each sample. Specifically, we propose a schema-aware reference store , which unifies symbolic schema and relevant textual instances. Then, we employ a dynamic reference integration module to retrieve pertinent knowledge from the datastore as prompts during training and inference. Experimental results demonstrate that RAP can be plugged into various existing models and outperforms baselines in low-resource settings on four datasets of relational triple extraction and event extraction. In addition, we provide comprehensive empirical ablations and case analysis regarding different types and scales of knowledge in order to better understand the mechanisms of RAP 1 .","{'model': 'tldr@v2.0.0', 'text': 'A novel approach of schema-aware R eference A s P rompt ( RAP), which dynamically leverage schema and knowledge inherited from global (few-shot) training data for each sample, and employs a dynamic reference integration module to retrieve pertinent knowledge from the datastore as prompts during training and inference.'}",http://arxiv.org/pdf/2210.10709
-Prompt Combines Paraphrase: Teaching Pre-trained Models to Understand Rare Biomedical Words,Hao Wang,"Prompt-based fine-tuning for pre-trained models has proven effective for many natural language processing tasks under few-shot settings in general domain. However, tuning with prompt in biomedical domain has not been investigated thoroughly. Biomedical words are often rare in general domain, but quite ubiquitous in biomedical contexts, which dramatically deteriorates the performance of pre-trained models on downstream biomedical applications even after fine-tuning, especially in low-resource scenarios. We propose a simple yet effective approach to helping models learn rare biomedical words during tuning with prompt. Experimental results show that our method can achieve up to 6% improvement in biomedical natural language inference task without any extra parameters or training steps using few-shot vanilla prompt settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed simple yet effective approach to helping models learn rare biomedical words during tuning with prompt can achieve up to 6% improvement in biomedical natural language inference task without any extra parameters or training steps using few-shot vanilla prompt settings.'}",http://arxiv.org/pdf/2209.06453
-DynaMaR: Dynamic Prompt with Mask Token Representation,Xiaodi Sun,"Recent research has shown that large language models pretrained using unsupervised approaches can achieve significant performance improvement on many downstream tasks. Typically when adapting these language models to downstream tasks, like a classification or regression task, we employ a fine-tuning paradigm in which the sentence representation from the language model is input to a task-specific head; the model is then fine-tuned end-to-end. However, with the emergence of models like GPT-3, prompt-based fine-tuning has been proven to be a successful approach for few-shot tasks. Inspired by this work, we study discrete prompt technologies in practice. There are two issues that arise with the standard prompt approach. First, it can overfit on the prompt template. Second, it requires manual effort to formulate the downstream task as a language model problem. In this paper, we propose an improvement to prompt-based fine-tuning that addresses these two issues. We refer to our approach as DynaMaR -- Dynamic Prompt with Mask Token Representation. Results show that DynaMaR can achieve an average improvement of 10% in few-shot settings and improvement of 3.7% in data-rich settings over the standard fine-tuning approach on four e-commerce applications.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an improvement to prompt-based fine-tuning that addresses two issues that arise with the standard prompt approach, and refers to it as DynaMaR -- Dynamic Prompt with Mask Token Representation.'}",https://arxiv.org/pdf/2206.02982
-Domain Prompt Learning for Efficiently Adapting CLIP to Unseen Domains,X. Zhang,"Domain generalization (DG) is a difficult transfer learning problem aiming to learn a generalizable model for unseen domains. Recent foundation models (FMs) are robust to many distribution shifts and, therefore, should substantially improve the performance of DG. In this work, we study generic ways to adopt CLIP, a Visual-Language Foundation Model, for DG problems in image classification. While ERM greatly improves the accuracy with bigger backbones and training datasets using standard DG benchmarks, fine-tuning FMs is not practical in many real-world situations. We propose Domain Prompt Learning (DPL) as a novel approach for domain inference in the form of conditional prompt generation. DPL achieved a significant accuracy improvement with only training a lightweight prompt generator (a three-layer MLP), whose parameter is of equivalent scale to the classification projector in the previous DG literature. Combining \dplshort~with CLIP provides surprising performance, raising the accuracy of zero-shot CLIP from 73.7% to 79.3% on several standard datasets, namely PACS, VLCS, OfficeHome, and TerraIncognita. We hope the simplicity and success of our approach lead to broader adoption and analysis of foundation models in the domain generalization field. Our code is available at https://github.com/shogi880/DPLCLIP.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Domain Prompt Learning (DPL) as a novel approach for domain inference in the form of conditional prompt generation and achieves a significant accuracy improvement with only training a lightweight prompt generator (a three-layer MLP), whose parameter is of equivalent scale to the classification projector in the previous DG literature.'}",https://www.jstage.jst.go.jp/article/tjsai/38/6/38_38-6_B-MC2/_pdf
-Amortized Prompt: Lightweight Fine-Tuning for CLIP in Domain Generalization,Xin Zhang,"Domain generalization (DG) is a difficult transfer learning problem aiming to learn a generalizable model to unseen domains. Recent massive pretrained models such as CLIP and GPT-3, i.e. foundation models (FMs), have been shown to be robust to many distribution shifts and therefore should lead to substantial improvements in DG. In this work, we study generic ways to adopt CLIP for DG problems in image classification, where we evaluate on naive zero-shot learning and full DG learning settings. For the latter, we propose AP (Amortized Prompt), as a novel approach for domain inference in the form of prompt generation. Using several standard datasets on domain generalization benchmark, namely PACS, VLCS, OfficeHome, and TerraIncognita, CLIP provides comparable performance without fine-tuning any parameters, suggesting the applicability and importance of FM in DG. In addition, we show that combining domain prompt inference with CLIP enables AP to outperform strong baselines and the naive CLIP baselines by a large margin, raising accuracy from 71.3% to 79.3%. We hope the simplicity and success of our approach emphasizes the importance of and leads to wider more adoption and analysis of foundation models in the field of domain generalization. ∗E-mail:xin@weblab.t.u-tokyo.ac.jp Figure 1: The average performance on four DG datasets, VLCS, PACS, OfficeHome, TerraIncognita. We select Empirical Risk Minimization (ERM) [46] as a baselines due to its good performance showed in [24]. We surprisingly find that CLIP(ViT-B16) without any fine-tuning also can outperforms ERM baselines which fine-tune on source domains. Moreover, we demonstrate our AP effectively improves the performance of CLIP and outperforms the baseline ERM with ViT-B16 as a backbone by a large margin (8.3%). Since the choice of backbones is critical to ERM in DG, we propose to use CLIP with AP as the basic component. 1 ar X iv :2 11 1. 12 85 3v 1 [ cs .C V ] 2 5 N ov 2 02 1","{'model': 'tldr@v2.0.0', 'text': 'This work proposes AP (Amortized Prompt), as a novel approach for domain inference in the form of prompt generation, and surprisingly finds that CLIP (ViT-B16) without any fine-tuning also can outperforms ERM baselines which fine-tune on source domains.'}",
-Feature Normalization and Cartography-Based Demonstrations for Prompt-Based Fine-Tuning on Emotion-Related Tasks,Mahshid Hosseini,"To train a model in a traditional supervised learning classification system for natural language processing (NLP) tasks, it is essential to have labeled data, which is not present in large amounts for many tasks. Prompt-based learning methods attempt to combat the supervised learning need for labeled data by directly adapting pre-trained language models and modeling the probability of text itself. In this paper, we propose a novel data-agnostic strategy for prompt-based fine-tuning that leverages feature moments (a.k.a., mean and standard deviation) as a data augmentation technique and employs training dynamics (i.e., confidence and variability) to allow more informative samples to be concatenated for generating demonstrations as input context. Our approach is a strong method for few-shot learning that forces the language model to pay special attention to the feature moments and allows more informative samples to be concatenated for generating demonstrations as input context by selecting high confidence and low variance samples. To demonstrate its effectiveness given limited training data, we conduct extensive experiments in different few-shot settings on three empathy and emotion classification datasets (from various domains). We further evaluate our method's robustness by introducing noise to our few-shot input data and labels and show that exchanging moments between samples and incorporating cartography-based demonstrations are beneficial when the available data is limited and noisy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel data-agnostic strategy for prompt-based fine-tuning that leverages feature moments (a.k.a., mean and standard deviation) as a data augmentation technique and employs training dynamics to allow more informative samples to be concatenated for generating demonstrations as input context.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26514/26286
-Understanding prompt engineering may not require rethinking generalization,Victor Akinwande,"Zero-shot learning in prompted vision-language models, the practice of crafting prompts to build classifiers without an explicit training process, has achieved impressive performance in many settings. This success presents a seemingly surprising observation: these methods suffer relatively little from overfitting, i.e., when a prompt is manually engineered to achieve low error on a given training set (thus rendering the method no longer actually zero-shot), the approach still performs well on held-out test data. In this paper, we show that we can explain such performance well via recourse to classical PAC-Bayes bounds. Specifically, we show that the discrete nature of prompts, combined with a PAC-Bayes prior given by a language model, results in generalization bounds that are remarkably tight by the standards of the literature: for instance, the generalization bound of an ImageNet classifier is often within a few percentage points of the true test error. We demonstrate empirically that this holds for existing handcrafted prompts and prompts generated through simple greedy search. Furthermore, the resulting bound is well-suited for model selection: the models with the best bound typically also have the best test performance. This work thus provides a possible justification for the widespread practice of prompt engineering, even if it seems that such methods could potentially overfit the training data.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the discrete nature of prompts, combined with a PAC-Bayes prior given by a language model, results in generalization bounds that are remarkably tight by the standards of the literature: for instance, the generalization bound of an ImageNet classifier is often within a few percentage points of the true test error.'}",
-Few shot learning approaches to essay scoring,Robert K. Helmeczi,"Automated essay scoring (AES) involves using computer technology to grade written assessments and assigning a score based on their perceived quality. AES has been among the most significant Natural Language Processing (NLP) applications due to its educational and commercial value. Similar to many other NLP tasks, training a model for AES typically involves acquiring a substantial amount of labeled data specific to the essay being graded. This usually incurs a substantial cost. In this study, we consider two recent few-shot learning methods to enhance the predictive performance of machine learning methods for AES tasks. Specifically, we experiment with a prompt-based few-shot learning method, pattern exploiting training (PET), and a prompt-free few-shot learning strategy, SetFit, and compare these against vanilla fine-tuning. Our numerical study shows that PET can provide substantial performance gains over other methods, and it can effectively boost performance when access to labeled data is limited. On the other hand, PET is found to be the most computationally expensive few-shot learning method considered, while SetFit is the fastest method among the approaches.","{'model': 'tldr@v2.0.0', 'text': 'A numerical study shows that PET can provide substantial performance gains over other methods, and it can effectively boost performance when access to labeled data is limited, and PET is found to be the most computationally expensive few-shot learning method considered.'}",https://caiac.pubpub.org/pub/gdf5n6gs/download/pdf
-BYOC: Personalized Few-Shot Classification with Co-Authored Class Descriptions,Arth Bohra,"Text classification is a well-studied and versatile building block for many NLP applications. Yet, existing approaches require either large annotated corpora to train a model with or, when using large language models as a base, require carefully crafting the prompt as well as using a long context that can fit many examples. As a result, it is not possible for end-users to build classifiers for themselves. To address this issue, we propose a novel approach to few-shot text classification using an LLM. Rather than few-shot examples, the LLM is prompted with descriptions of the salient features of each class. These descriptions are coauthored by the user and the LLM interactively: while the user annotates each few-shot example, the LLM asks relevant questions that the user answers. Examples, questions, and answers are summarized to form the classification prompt. Our experiments show that our approach yields high accuracy classifiers, within 82% of the performance of models trained with significantly larger datasets while using only 1% of their training sets. Additionally, in a study with 30 participants, we show that end-users are able to build classifiers to suit their specific needs. The personalized classifiers show an average accuracy of 90%, which is 15% higher than the state-of-the-art approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach to few-shot text classification using an LLM that yields high accuracy classifiers, within 82% of the performance of models trained with significantly larger datasets while using only 1% of their training sets.'}",https://arxiv.org/pdf/2310.06111
-Prompt Sketching for Large Language Models,Luca Beurer-Kellner,"Many recent prompting strategies for large language models (LLMs) query the model multiple times sequentially -- first to produce intermediate results and then the final answer. However, using these methods, both decoder and model are unaware of potential follow-up prompts, leading to disconnected and undesirably wordy intermediate responses. In this work, we address this issue by proposing prompt sketching, a new prompting paradigm in which an LLM does not only respond by completing a prompt, but by predicting values for multiple variables in a template. This way, sketching grants users more control over the generation process, e.g., by providing a reasoning framework via intermediate instructions, leading to better overall results. The key idea enabling sketching with existing, autoregressive models is to adapt the decoding procedure to also score follow-up instructions during text generation, thus optimizing overall template likelihood in inference. Our experiments show that in a zero-shot setting, prompt sketching outperforms existing, sequential prompting schemes such as direct asking or chain-of-thought on 7 out of 8 LLM benchmarking tasks, including state tracking, arithmetic reasoning, and general question answering. To facilitate future use, we release a number of generic, yet effective sketches applicable to many tasks, and an open source library called dclib, powering our sketch-aware decoders.","{'model': 'tldr@v2.0.0', 'text': None}",
-Strength in Numbers: Estimating Confidence of Large Language Models by Prompt Agreement,Gwenyth Portillo Wightman,"Large language models have achieved impressive few-shot performance on a wide variety of tasks. However, in many settings, users require confidence estimates for model predictions. While traditional classifiers produce scores for each label, language models instead produce scores for the generation which may not be well calibrated. We compare generations across diverse prompts and show that these can be used to create confidence scores. By utilizing more prompts we can get more precise confidence estimates and use response diversity as a proxy for confidence. We evaluate this approach across ten multiple-choice question-answering datasets using three models: T0, FLAN-T5, and GPT-3. In addition to analyzing multiple human written prompts, we automatically generate more prompts using a language model in order to produce finer-grained confidence estimates. Our method produces more calibrated confidence estimates compared to the log probability of the answer to a single prompt. These improvements could benefit users who rely on prediction confidence for integration into a larger system or in decision-making processes.","{'model': 'tldr@v2.0.0', 'text': 'This work automatically generates more prompts using a language model in order to produce finer-grained confidence estimates and produces more calibrated confidence estimates compared to the log probability of the answer to a single prompt.'}",
-The language of prompting: What linguistic properties make a prompt successful?,Alina Leidinger,"The latest generation of LLMs can be prompted to achieve impressive zero-shot or few-shot performance in many NLP tasks. However, since performance is highly sensitive to the choice of prompts, considerable effort has been devoted to crowd-sourcing prompts or designing methods for prompt optimisation. Yet, we still lack a systematic understanding of how linguistic properties of prompts correlate with task performance. In this work, we investigate how LLMs of different sizes, pre-trained and instruction-tuned, perform on prompts that are semantically equivalent, but vary in linguistic structure. We investigate both grammatical properties such as mood, tense, aspect and modality, as well as lexico-semantic variation through the use of synonyms. Our findings contradict the common assumption that LLMs achieve optimal performance on lower perplexity prompts that reflect language use in pretraining or instruction-tuning data. Prompts transfer poorly between datasets or models, and performance cannot generally be explained by perplexity, word frequency, ambiguity or prompt length. Based on our results, we put forward a proposal for a more robust and comprehensive evaluation standard for prompting research.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates how LLMs of different sizes, pre-trained and instruction-tuned, perform on prompts that are semantically equivalent, but vary in linguistic structure, and puts forward a proposal for a more robust and comprehensive evaluation standard for prompting research.'}",
-Hard Sample Aware Prompt-Tuning,Yuanjiang Xu,"Prompt-tuning based few-shot learning has garnered increasing attention in recent years due to its efficiency and promising capability. To achieve the best performance for NLP tasks with just a few samples, it is vital to include as many informative samples as possible and to avoid misleading ones. However, there is no work in prompt-tuning literature addressing the problem of differentiating informative hard samples from misleading ones in model training, which is challenging due to the lack of supervision signals about the quality of the samples to train a well-performed model. We propose a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method. An extensive empirical study on a series of NLP tasks demonstrates the capability of HardPT in few-shot scenarios. HardPT obtains new SOTA results on all evaluated NLP tasks, including pushing the SST-5 accuracy to 49.5% (1.1% point absolute improvement), QNLI accuracy to 74.6% (1.9% absolute improvement), NMLI accuracy to 71.5 (0.7% absolute improvement), TACREV F_1-score to 28.2 (1.0 absolute improvement), and i2b2/VA F_1-score to 41.2 (1.3 absolute improvement).","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method.'}",https://aclanthology.org/2023.acl-long.690.pdf
-Multi-view Vision-Prompt Fusion Network: Can 2D Pre-trained Model Boost 3D Point Cloud Data-scarce Learning?,Hao Peng,"Point cloud based 3D deep model has wide applications in many applications such as autonomous driving, house robot, and so on. Inspired by the recent prompt learning in natural language processing, this work proposes a novel Multi-view Vision-Prompt Fusion Network (MvNet) for few-shot 3D point cloud classification. MvNet investigates the possibility of leveraging the off-the-shelf 2D pre-trained models to achieve the few-shot classification, which can alleviate the over-dependence issue of the existing baseline models towards the large-scale annotated 3D point cloud data. Specifically, MvNet first encodes a 3D point cloud into multi-view image features for a number of different views. Then, a novel multi-view prompt fusion module is developed to effectively fuse information from different views to bridge the gap between 3D point cloud data and 2D pre-trained models. A set of 2D image prompts can then be derived to better describe the suitable prior knowledge for a large-scale pre-trained image model for few-shot 3D point cloud classification. Extensive experiments on ModelNet, ScanObjectNN, and ShapeNet datasets demonstrate that MvNet achieves new state-of-the-art performance for 3D few-shot point cloud image classification. The source code of this work will be available soon.","{'model': 'tldr@v2.0.0', 'text': 'MvNet investigates the possibility of leveraging the off-the-shelf 2D pre-trained models to achieve the few-shot classification, which can alleviate the over-dependence issue of the existing baseline models towards the large-scale annotated 3D point cloud data.'}",https://arxiv.org/pdf/2304.10224
-Voucher Abuse Detection with Prompt-based Fine-tuning on Graph Neural Networks,Zhi Wen,"Voucher abuse detection is an important anomaly detection problem in E-commerce. While many GNN-based solutions have emerged, the supervised paradigm depends on a large quantity of labeled data. A popular alternative is to adopt self-supervised pre-training using label-free data, and further fine-tune on a downstream task with limited labels. Nevertheless, the ""pre-train, fine-tune"" paradigm is often plagued by the objective gap between pre-training and downstream tasks. Hence, we propose VPGNN, a prompt-based fine-tuning framework on GNNs for voucher abuse detection. We design a novel graph prompting function to reformulate the downstream task into a similar template as the pretext task in pre-training, thereby narrowing the objective gap. Extensive experiments on both proprietary and public datasets demonstrate the strength of VPGNN in both few-shot and semi-supervised scenarios. Moreover, an online deployment of VPGNN in a production environment shows a 23.4% improvement over two existing deployed models.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes VPGNN, a prompt-based fine-tuning framework on GNNs for voucher abuse detection, and designs a novel graph prompting function to reformulate the downstream task into a similar template as the pretext task in pre-training, thereby narrowing the objective gap.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615505
-Modal interaction-enhanced prompt learning by transformer decoder for vision-language models,Mingyue Liu,,"{'model': 'tldr@v2.0.0', 'text': 'A prompt tuning method for simulating image-text interaction based on CoOp: Decoding context optimization (DeCoOp) is proposed, demonstrating that this approach substantially outperforms the baseline model CoOp in terms of point domain generalization and robustness.'}",
-Stabilized In-Context Learning with Pre-trained Language Models for Few Shot Dialogue State Tracking,Derek Chen,"Prompt-based methods with large pre-trained language models (PLMs) have shown impressive unaided performance across many NLP tasks. These models improve even further with the addition of a few labeled in-context exemplars to guide output generation. However, for more complex tasks such as dialogue state tracking (DST), designing prompts that reliably convey the desired intent is nontrivial, leading to unstable results. Furthermore, building in-context exemplars for dialogue tasks is difficult because conversational contexts are long while model input lengths are relatively short.To overcome these issues we first adapt a meta-learning scheme to the dialogue domain which stabilizes the ability of the model to perform well under various prompts. We additionally design a novel training method to improve upon vanilla retrieval mechanisms to find ideal in-context examples. Finally, we introduce a saliency model to limit dialogue text length, allowing us to include more exemplars per query. In effect, we are able to achieve highly competitive results for few-shot DST on MultiWOZ.","{'model': 'tldr@v2.0.0', 'text': 'A meta-learning scheme is adapted to the dialogue domain which stabilizes the ability of the model to perform well under various prompts and introduces a saliency model to limit dialogue text length, allowing for highly competitive results for few-shot DST on MultiWOZ.'}",http://arxiv.org/pdf/2302.05932
-In-Context Learning for Text Classification with Many Labels,Aristides Milios,"In-context learning (ICL) using large language models for tasks with many labels is challenging due to the limited context window, which makes it difficult to fit a sufficient number of examples in the prompt. In this paper, we use a pre-trained dense retrieval model to bypass this limitation, giving the model only a partial view of the full label space for each inference call. Testing with recent open-source LLMs (OPT, LLaMA), we set new state of the art performance in few-shot settings for three common intent classification datasets, with no finetuning. We also surpass fine-tuned performance on fine-grained sentiment classification in certain cases. We analyze the performance across number of in-context examples and different model scales, showing that larger models are necessary to effectively and consistently make use of larger context lengths for ICL. By running several ablations, we analyze the model's use of: a) the similarity of the in-context examples to the current input, b) the semantic content of the class names, and c) the correct correspondence between examples and labels. We demonstrate that all three are needed to varying degrees depending on the domain, contrary to certain recent works.","{'model': 'tldr@v2.0.0', 'text': 'A pre-trained dense retrieval model is used to bypass this limitation, giving the model only a partial view of the full label space for each inference call, showing that larger models are necessary to effectively and consistently make use of larger context lengths for ICL.'}",https://arxiv.org/pdf/2309.10954
-Proposal of Range Verification Method for Carbon Therapy using Prompt Gamma-rays over 10 MeV,S. Kimura,"Range verification system is a most important part of particle therapy, for beam’s high killing power and convergence. To verify the range, many kinds of methods are devised and have been studied. We have studied the system which detects prompt gamma-rays over 10 MeV by using pair production events. We simulated the distribution of prompt gamma-rays’ energy, vertex positions, final scattering positions, and its generation process by Geant4 Monte Carlo simulation toolkit. We used physics lists named ""QGSP_INCLXX_HP"" and ""G4RadioactiveDecayPhysics"". PMMA cylinder with a diameter of 30 cm was put in the air. Ten million carbon beams with the energy of 290 MeV/n were shot to the PMMA target. We also simulated a more simple detector. It is consist of arrayed scintillation plates, a lead collimator, and a plastic scintillation plate. The plastic scintillation plate is used for vetoing charged particles. The lead collimator is separated from the centre, and move for shielding completely. This system makes it possible to reduce neutron derived background. The result of the simulation says our detector can measure the Bragg peak position.",,
-Consistency Analysis of ChatGPT,Myeongjun Jang,"ChatGPT has gained a huge popularity since its introduction. Its positive aspects have been reported through many media platforms, and some analyses even showed that ChatGPT achieved a decent grade in professional exams, adding extra support to the claim that AI can now assist and even replace humans in industrial fields. Others, however, doubt its reliability and trustworthiness. This paper investigates the trustworthiness of ChatGPT and GPT-4 regarding logically consistent behaviour, focusing specifically on semantic consistency and the properties of negation, symmetric, and transitive consistency. Our findings suggest that while both models appear to show an enhanced language understanding and reasoning ability, they still frequently fall short of generating logically consistent predictions. We also ascertain via experiments that prompt designing, few-shot learning and employing larger large language models (LLMs) are unlikely to be the ultimate solution to resolve the inconsistency issue of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the trustworthiness of ChatGPT and GPT-4 regarding logically consistent behaviour, focusing specifically on semantic consistency and the properties of negation, symmetric, and transitive consistency.'}",http://arxiv.org/pdf/2303.06273
-ALLSH: Active Learning Guided by Local Sensitivity and Hardness,Shujian Zhang,"Active learning, which effectively collects informative unlabeled data for annotation, reduces the demand for labeled data. In this work, we propose to retrieve unlabeled samples with a local sensitivity and hardness-aware acquisition function. The proposed method generates data copies through local perturbations and selects data points whose predictive likelihoods diverge the most from their copies. We further empower our acquisition function by injecting the select-worst case perturbation. Our method achieves consistent gains over the commonly used active learning strategies in various classification tasks. Furthermore, we observe consistent improvements over the baselines on the study of prompt selection in prompt-based few-shot learning. These experiments demonstrate that our acquisition guided by local sensitivity and hardness can be effective and beneficial for many NLP tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to retrieve unlabeled samples with a local sensitivity and hardness-aware acquisition function, which generates data copies through local perturbations and selects data points whose predictive likelihoods diverge the most from their copies.'}",http://arxiv.org/pdf/2205.04980
-ClarET: Pre-training a Correlation-Aware Context-To-Event Transformer for Event-Centric Generation and Classification,Yucheng Zhou,"Generating new events given context with correlated ones plays a crucial role in many event-centric reasoning tasks. Existing works either limit their scope to specific scenarios or overlook event-level correlations. In this paper, we propose to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning. To achieve this, we propose three novel event-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training. The proposed ClarET is applicable to a wide range of event-centric reasoning scenarios, considering its versatility of (i) event-correlation types (e.g., causal, temporal, contrast), (ii) application formulations (i.e., generation and classification), and (iii) reasoning types (e.g., abductive, counterfactual and ending reasoning). Empirical fine-tuning results, as well as zero- and few-shot learning, on 9 benchmarks (5 generation and 4 classification tasks covering 4 reasoning types with diverse event correlations), verify its effectiveness and generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning, and proposes three novelevent-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training.'}",http://arxiv.org/pdf/2203.02225
-Does GPT-3 Generate Empathetic Dialogues? A Novel In-Context Example Selection Method and Automatic Evaluation Metric for Empathetic Dialogue Generation,Young-Jun Lee,"Since empathy plays a crucial role in increasing social bonding between people, many studies have designed their own dialogue agents to be empathetic using the well-established method of fine-tuning. However, they do not use prompt-based in-context learning, which has shown powerful performance in various natural language processing (NLP) tasks, for empathetic dialogue generation. Although several studies have investigated few-shot in-context learning for empathetic dialogue generation, an in-depth analysis of the generation of empathetic dialogue with in-context learning remains unclear, especially in GPT-3 (Brown et al., 2020). In this study, we explore whether GPT-3 can generate empathetic dialogues through prompt-based in-context learning in both zero-shot and few-shot settings. To enhance performance, we propose two new in-context example selection methods, called SITSM and EMOSITSM, that utilize emotion and situational information. We also introduce a new automatic evaluation method, DIFF-EPITOME, which reflects the human tendency to express empathy. From the analysis, we reveal that our DIFF-EPITOME is effective in measuring the degree of human empathy. We show that GPT-3 achieves competitive performance with Blender 90M, a state-of-the-art dialogue generative model, on both automatic and human evaluation. Our code is available at https://github.com/passing2961/EmpGPT-3.","{'model': 'tldr@v2.0.0', 'text': 'This study explores whether GPT-3 can generate empathetic dialogues through prompt-based in-context learning in both zero-shot and few-shot settings and introduces a new automatic evaluation method, DIFF-EPITOME, which reflects the human tendency to express empathy.'}",
-Exploring The Landscape of Distributional Robustness for Question Answering Models,Anas Awadalla,"We conduct a large empirical evaluation to investigate the landscape of distributional robustness in question answering. Our investigation spans over 350 models and 16 question answering datasets, including a diverse set of architectures, model sizes, and adaptation methods (e.g., fine-tuning, adapter tuning, in-context learning, etc.). We find that, in many cases, model variations do not affect robustness and in-distribution performance alone determines out-of-distribution performance. Moreover, our findings indicate that i) zero-shot and in-context learning methods are more robust to distribution shifts than fully fine-tuned models; ii) few-shot prompt fine-tuned models exhibit better robustness than few-shot fine-tuned span prediction models; iii) parameter-efficient and robustness enhancing training methods provide no significant robustness improvements. In addition, we publicly release all evaluations to encourage researchers to further analyze robustness trends for question answering models.","{'model': 'tldr@v2.0.0', 'text': 'This investigation spans over 350 models and 16 question answering datasets, including a diverse set of architectures, model sizes, and adaptation methods, and indicates that zero-shot and in-context learning methods are more robust to distribution shifts than fully fine-tuned models.'}",http://arxiv.org/pdf/2210.12517
-EmotionPrompt: Leveraging Psychology for Large Language Models Enhancement via Emotional Stimulus,Cheng Li,"Large language models (LLMs) have achieved significant performance in many fields, such as reasoning, language understanding, and math problem-solving, and are regarded as an important step to artificial general intelligence (AGI). However, the sensitivity of LLMs to prompts remain a major bottleneck for their daily adoption. In this paper, we take inspiration from psychology and propose EmotionPrompt to explore emotional intelligence to enhance the performance of LLMs. Our EmotionPrompt operates on a remarkably straightforward principle: the incorporation of emotional stimulus into prompts. Experimental re-sults demonstrate that our EmotionPrompt, using the same single prompt templates, significantly out-performs the original prompt and Zero-shot-CoT in both zero-shot and few-shot settings on eight tasks with diverse models: ChatGPT, Vicuna-13b, Bloom, and Flan-T5-large. Furthermore, Emotion-Prompt was observed to improve both the truthfulness and informativeness. We believe that Emo-tionPrompt heralds a novel avenue for exploring interdisciplinary knowledge for interaction between humans and LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Emo-tionPrompt heralds a novel avenue for exploring interdisciplinary knowledge for interaction between humans and LLMs and was observed to improve both the truthfulness and informativeness.'}",https://arxiv.org/pdf/2307.11760
-ScoNe: Benchmarking Negation Reasoning in Language Models With Fine-Tuning and In-Context Learning,Jingyuan Selena She,"A number of recent benchmarks seek to assess how well models handle natural language negation. However, these benchmarks lack the controlled example paradigms that would allow us to infer whether a model had truly learned how negation morphemes semantically scope. To fill these analytical gaps, we present the Scoped Negation NLI (ScoNe-NLI) benchmark, which contains contrast sets of six examples with up to two negations where either zero, one, or both negative morphemes affect the NLI label. We use ScoNe-NLI to assess fine-tuning and in-context learning strategies. We find that RoBERTa and DeBERTa models solve ScoNe-NLI after many shot fine-tuning. For in-context learning, we test the latest InstructGPT models and find that most prompt strategies are not successful, including those using step-by-step reasoning. To better understand this result, we extend ScoNe with ScoNe-NLG, a sentence completion test set that embeds negation reasoning in short narratives. Here, InstructGPT is successful, which reveals the model can correctly reason about negation, but struggles to do so on NLI examples outside of its core pretraining regime.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the Scoped Negation NLI (ScoNe-NLI) benchmark, which contains contrast sets of six examples with up to two negations where either zero, one, or both negative morphemes affect the NLI label, and extends ScoNe with ScoNe-NLG, a sentence completion test set that embeds negation reasoning in short narratives.'}",http://arxiv.org/pdf/2305.19426
-Enabling Classifiers to Make Judgements Explicitly Aligned with Human Values,Yejin Bang,"Many NLP classification tasks, such as sexism/racism detection or toxicity detection, are based on human values. Yet, human values can vary under diverse cultural conditions. Therefore, we introduce a framework for value-aligned classification that performs prediction based on explicitly written human values in the command. Along with the task, we propose a practical approach that distills value-aligned knowledge from large-scale language models (LLMs) to construct value-aligned classifiers in two steps.First, we generate value-aligned training data from LLMs by prompt-based few-shot learning. Next, we fine-tune smaller classification models with the generated data for the task. Empirical results show that our VA-Models surpass multiple baselines by at least 15.56% on the F1-score, including few-shot learning with OPT-175B and existing text augmentation methods. We suggest that using classifiers with explicit human value input improves both inclusivity & explainability in AI.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a framework for value-aligned classification that performs prediction based on explicitly written human values in the command and suggests that using classifiers with explicit human value input improves both inclusivity & explainability in AI.'}",http://arxiv.org/pdf/2210.07652
-Bits of Grass: Does GPT already know how to write like Whitman?,Piotr Sawicki,"This study examines the ability of GPT-3.5, GPT-3.5-turbo (ChatGPT) and GPT-4 models to generate poems in the style of specific authors using zero-shot and many-shot prompts (which use the maximum context length of 8192 tokens). We assess the performance of models that are not fine-tuned for generating poetry in the style of specific authors, via automated evaluation. Our findings indicate that without fine-tuning, even when provided with the maximum number of 17 poem examples (8192 tokens) in the prompt, these models do not generate poetry in the desired style.",,http://arxiv.org/pdf/2305.11064
-Do Prompts Solve NLP Tasks Using Natural Language?,Sen Yang,"Thanks to the advanced improvement of large pre-trained language models, prompt-based fine-tuning is shown to be effective on a variety of downstream tasks. Though many prompting methods have been investigated, it remains unknown which type of prompts are the most effective among three types of prompts (i.e., human-designed prompts, schema prompts and null prompts). In this work, we empirically compare the three types of prompts under both few-shot and fully-supervised settings. Our experimental results show that schema prompts are the most effective in general. Besides, the performance gaps tend to diminish when the scale of training data grows large.","{'model': 'tldr@v2.0.0', 'text': 'This work empirically compare the three types of prompts under both few-shot and fully-supervised settings and shows that schema prompts are the most effective in general.'}",http://arxiv.org/pdf/2203.00902
-Evaluation of a Single-Stage Light-Gas Gun Facility in Malta: Business Analysis and Preliminary Design,L. Barilaro,,,https://link.springer.com/content/pdf/10.1007/s42496-022-00113-w.pdf
-Towards Answering Open-ended Ethical Quandary Questions,Yejin Bang,"Considerable advancements have been made in various NLP tasks based on the impressive power of large language models (LLMs) and many NLP applications are deployed in our daily lives. In this work, we challenge the capability of LLMs with the new task of Ethical Quandary Generative Question Answering. Ethical quandary questions are more challenging to address because multiple conflicting answers may exist to a single quandary. We explore the current capability of LLMs in providing an answer with a deliberative exchange of different perspectives to an ethical quandary, in the approach of Socratic philosophy, instead of providing a closed answer like an oracle. We propose a model that searches for different ethical principles applicable to the ethical quandary and generates an answer conditioned on the chosen principles through prompt-based few-shot learning. We also discuss the remaining challenges and ethical issues involved in this task and suggest the direction toward developing responsible NLP systems by incorporating human values explicitly.","{'model': 'tldr@v2.0.0', 'text': 'A model is proposed that searches for different ethical principles applicable to the ethical quandary and generates an answer conditioned on the chosen principles through prompt-based few-shot learning and suggests the direction toward developing responsible NLP systems by incorporating human values explicitly.'}",
-CLIPTexture: Text-Driven Texture Synthesis,Yiren Song,"Can artificial intelligence create textures with artistic value according to human language control? Existing texture synthesis methods require example texture input. However, in many practical situations, users don't have satisfying textures but tell designers about their needs through simple sketches and verbal descriptions. This paper proposes a novel texture synthesis framework based on the CLIP, which models the texture synthesis problem as an optimization process and realizes text-driven texture synthesis by minimizing the distance between the input image and the text prompt in latent space. Our method performs zero-shot image manipulation successfully even between unseen domains. We implement texture synthesis using two different optimization methods, the TextureNet and Diffvg, demonstrating the generality of CLIPTexture. Extensive experiments confirmed the robust and superior manipulation performance of our methods compared to the existing baselines.","{'model': 'tldr@v2.0.0', 'text': 'A novel texture synthesis framework based on the CLIP is proposed, which models the texture synthesis problem as an optimization process and realizes text-driven texture synthesis by minimizing the distance between the input image and the text prompt in latent space.'}",
-BertNet: Harvesting Knowledge Graphs with Arbitrary Relations from Pretrained Language Models,Shibo Hao,"It is crucial to automatically construct knowledge graphs (KGs) of diverse new relations to support knowledge discovery and broad applications. Previous KG construction methods, based on either crowdsourcing or text mining, are often limited to a small predefined set of relations due to manual cost or restrictions in text corpus. Recent research proposed to use pretrained language models (LMs) as implicit knowledge bases that accept knowledge queries with prompts. Yet, the implicit knowledge lacks many desirable properties of a full-scale symbolic KG, such as easy access, navigation, editing, and quality assurance. In this paper, we propose a new approach of harvesting massive KGs of arbitrary relations from pretrained LMs. With minimal input of a relation definition (a prompt and a few shot of example entity pairs), the approach efficiently searches in the vast entity pair space to extract diverse accurate knowledge of the desired relation. We develop an effective search-and-rescore mechanism for improved efficiency and accuracy. We deploy the approach to harvest KGs of over 400 new relations from different LMs. Extensive human and automatic evaluations show our approach manages to extract diverse accurate knowledge, including tuples of complex relations (e.g.,""A is capable of but not good at B""). The resulting KGs as a symbolic interpretation of the source LMs also reveal new insights into the LMs' knowledge capacities.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new approach of harvesting massive KGs of arbitrary relations from pretrained LMs with minimal input of a relation definition, and develops an effective search-and-rescore mechanism for improved efficiency and accuracy.'}",https://aclanthology.org/2023.findings-acl.309.pdf
-Learning Disentangled Prompts for Compositional Image Synthesis,Kihyuk Sohn,"We study domain-adaptive image synthesis, the problem of teaching pretrained image generative models a new style or concept from as few as one image to synthesize novel images, to better understand the compositional image synthesis. We present a framework that leverages a pretrained class-conditional generation model and visual prompt tuning. Specifically, we propose a novel source class distilled visual prompt that learns disentangled prompts of semantic (e.g., class) and domain (e.g., style) from a few images. Learned domain prompt is then used to synthesize images of any classes in the style of target domain. We conduct studies on various target domains with the number of images ranging from one to a few to many, and show qualitative results which show the compositional generalization of our method. Moreover, we show that our method can help improve zero-shot domain adaptation classification accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel source class distilled visual prompt that learns disentangled prompts of semantic and domain from a few images, and shows that this method can help improve zero-shot domain adaptation classification accuracy.'}",http://arxiv.org/pdf/2306.00763
-Language Models as Black-Box Optimizers for Vision-Language Models,Samuel Yu,"Vision-language models (VLMs) pre-trained on web-scale datasets have demonstrated remarkable capabilities across a variety of vision and multimodal tasks. Currently, fine-tuning methods for VLMs mainly operate in a white-box setting, requiring access to model parameters for backpropagation. However, many VLMs rely on proprietary data and are not open-source, which restricts the use of white-box approaches for fine-tuning. Given that popular private large language models (LLMs) like ChatGPT still offer a language-based user interface, we aim to develop a novel fine-tuning approach for VLMs through natural language prompts, thereby avoiding the need to access model parameters, feature embeddings, or output logits. In this setup, we propose employing chat-based LLMs as black-box optimizers to search for the best text prompt on the illustrative task of few-shot image classification using CLIP. Specifically, we adopt an automatic""hill-climbing""procedure that converges on an effective prompt by evaluating the accuracy of current prompts and asking LLMs to refine them based on textual feedback, all within a conversational process without human-in-the-loop. In a challenging 1-shot learning setup, our simple approach surpasses the white-box continuous prompting method (CoOp) by an average of 1.5% across 11 datasets including ImageNet. Our approach also outperforms OpenAI's manually crafted prompts. Additionally, we highlight the advantage of conversational feedback that incorporates both positive and negative prompts, suggesting that LLMs can utilize the implicit""gradient""direction in textual feedback for a more efficient search. Lastly, we find that the text prompts generated through our strategy are not only more interpretable but also transfer well across different CLIP architectures in a black-box manner.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to develop a novel fine-tuning approach for VLMs through natural language prompts, thereby avoiding the need to access model parameters, feature embeddings, or output logits, and finds that the text prompts generated through this strategy are not only more interpretable but also transfer well across different CLIP architectures in a black-box manner.'}",https://arxiv.org/pdf/2309.05950
-Large Language Model for Multi-objective Evolutionary Optimization,Fei Liu,"Multiobjective evolutionary algorithms (MOEAs) are major methods for solving multiobjective optimization problems (MOPs). Many MOEAs have been proposed in the past decades, of which the search operators need a carefully handcrafted design with domain knowledge. Recently, some attempts have been made to replace the manually designed operators in MOEAs with learning-based operators (e.g., neural network models). However, much effort is still required for designing and training such models, and the learned operators might not generalize well on new problems. To tackle the above challenges, this work investigates a novel approach that leverages the powerful large language model (LLM) to design MOEA operators. With proper prompt engineering, we successfully let a general LLM serve as a black-box search operator for decomposition-based MOEA (MOEA/D) in a zero-shot manner. In addition, by learning from the LLM behavior, we further design an explicit white-box operator with randomness and propose a new version of decomposition-based MOEA, termed MOEA/D-LO. Experimental studies on different test benchmarks show that our proposed method can achieve competitive performance with widely used MOEAs. It is also promising to see the operator only learned from a few instances can have robust generalization performance on unseen problems with quite different patterns and settings. The results reveal the potential benefits of using pre-trained LLMs in the design of MOEAs.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates a novel approach that leverages the powerful large language model (LLM) to design MOEA operators and reveals the potential benefits of using pre-trained LLMs in the design of MOEAs.'}",
-Impossible Triangle: What's Next for Pre-trained Language Models?,Chenguang Zhu,"Recent development of large-scale pre-trained language models (PLM) have significantly improved the capability of models in various NLP tasks, in terms of performance after task-specific fine-tuning and zero-shot / few-shot learning. However, many of such models come with a dauntingly huge size that few institutions can afford to pre-train, fine-tune or even deploy, while moderate-sized models usually lack strong generalized few-shot learning capabilities. In this paper, we first elaborate the current obstacles of using PLM models in terms of the Impossible Triangle: 1) moderate model size, 2) state-of-the-art few-shot learning capability, and 3) state-of-the-art fine-tuning capability. We argue that all existing PLM models lack one or more properties from the Impossible Triangle. To remedy these missing properties of PLMs, various techniques have been proposed, such as knowledge distillation, data augmentation and prompt learning, which inevitably brings additional work to the application of PLMs in real scenarios. We then offer insights into future research directions of PLMs to achieve the Impossible Triangle, and break down the task into several key phases.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that all existing PLM models lack one or more properties from the Impossible Triangle, and insights into future research directions of PLMs to achieve the Impossible triangle are offered.'}",http://arxiv.org/pdf/2204.06130
-Effects of Target Words and Their Locations in Prompts,T. Bers,"Few and zero-shot learning with prompt-tuned models has recently been shown to be very successful. Prompt-tuned models require prompt engineering where the model is trained on many different prompts. A prompt is some text that is added to an input example in order to convert it into a language modeling task. For example instead of just training on pairs of hypotheses and premises, the prompted input text would be: “Given { premise } , does this follow: { hypothesis } ?” A challenge is figuring out which prompt wording allows the model to perform best. In this thesis, we experiment with different target words and their locations; either not included in the prompt, at the start, or at the end of the prompt. Target words are the set of possible response to the prompt, for the previous example the target words could be yes and no. We look at two different models, instruction tuned (T0) and not (ALBERT). We also analyze the difference between varying difficulties of prompts in Natural Language Inference (NLI) and across different tasks (NLI, coreference resolution, sentence completion, and multiple choice question and answer). In sum, we find 1) that yes/no and true/false perform best as the target words, 2) the impact of location of the target words on T0 varies by task and results are inconsistent, and 3) that including them as a negative impact on ALBERT.","{'model': 'tldr@v2.0.0', 'text': 'It is found that yes/no and true/false perform best as the target words, the impact of location of the target Words on T0 varies by task and results are inconsistent, and that including them as a negative impact on ALBERT is found.'}",
-Weak supervision for Question Type Detection with large language models,J. Martínek,"Large pre-trained language models (LLM) have shown remark-able Zero-Shot Learning performances in many Natural Language Processing tasks. However, designing effective prompts is still very difficult for some tasks, in particular for dialogue act recognition. We propose an alternative way to leverage pretrained LLM for such tasks that replace manual prompts with simple rules, which are more intuitive and easier to design for some tasks. We demonstrate this approach on the question type recognition task, and show that our zero-shot model obtains competitive performances both with a supervised LSTM trained on the full training corpus, and another supervised model from previously published works on the MRDA corpus. We further analyze the limits of the proposed approach, which can not be applied on any task, but may advantageously complement prompt programming for specific classes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed zero-shot model obtains competitive performances both with a supervised LSTM trained on the full training corpus, and another supervised model from previously published works on the MRDA corpus.'}",https://hal.science/hal-03786135/document
-Ecologically Valid Explanations for Label Variation in NLI,Nan-Jiang Jiang,"Human label variation, or annotation disagreement, exists in many natural language processing (NLP) tasks, including natural language inference (NLI). To gain direct evidence of how NLI label variation arises, we build LiveNLI, an English dataset of 1,415 ecologically valid explanations (annotators explain the NLI labels they chose) for 122 MNLI items (at least 10 explanations per item). The LiveNLI explanations confirm that people can systematically vary on their interpretation and highlight within-label variation: annotators sometimes choose the same label for different reasons. This suggests that explanations are crucial for navigating label interpretations in general. We few-shot prompt large language models to generate explanations but the results are inconsistent: they sometimes produces valid and informative explanations, but it also generates implausible ones that do not support the label, highlighting directions for improvement.","{'model': 'tldr@v2.0.0', 'text': 'The LiveNLI explanations confirm that people can systematically vary on their interpretation and highlight within-label variation: annotators sometimes choose the same label for different reasons, which suggests that explanations are crucial for navigating label interpretations in general.'}",
-B-Pet: The PET Model with Parameter-Efficient Learning,Qi Zheng,"In recent years, under the trend of training models in big data, Few-shot learning (FSL) which aims to learn models to solve problems with a few samples has also achieved good results on many data sets. In fact, acquiring high-quality training samples is expensive in many aspects, but FSL can save the overhead costs. Among FSL models, the PET model combines semi-supervised learning, prompt learning and knowledge distillation based on the pre-training language model. However, in fine-turning the PET model has the disadvantages that consumes a lot of resources and time and requires heavy costs of storage for model preservation. Therefore, this paper proposes the B-pet model, which freezes most of the training parameters and only trains bias parameters during fine-turning process, significantly reducing the storage consumption of the model for downstream tasks. We used six data sets with $\vert \tau \vert=\mathbf{10},\ \mathbf{50},\ \mathbf{100}$ and three different data training models respectively. The results show that four data sets on the B-pet model performed better than original PET model training. It is obvious that in the memory-constrained environment deployment, multitasking fine-tunes models have practical value. It also proved that most semi-supervised models with fixed parameters are realizable.","{'model': 'tldr@v2.0.0', 'text': 'The B-pet model is proposed, which freezes most of the training parameters and only trains bias parameters during fine-turning process, significantly reducing the storage consumption of the model for downstream tasks and proved that most semi-supervised models with fixed parameters are realizable.'}",
-Automatic Data Transformation Using Large Language Model: An Experimental Study on Building Energy Data,Ankita Sharma,"Existing approaches to automatic data transformation are insufficient to meet the requirements in many real-world scenarios, such as the building sector. First, there is no convenient interface for domain experts to provide domain knowledge easily. Second, they require significant training data collection overheads. Third, the accuracy suffers from complicated schema changes. To bridge this gap, we present a novel approach that leverages the unique capabilities of large language models (LLMs) in coding, complex reasoning, and zero-shot learning to generate SQL code that transforms the source datasets into the target datasets. We demonstrate the viability of this approach by designing an LLM-based framework, termed SQLMorpher, which comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases. It also implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection. The key contributions of this work include (1) pioneering an end-to-end LLM-based solution for data transformation, (2) developing a benchmark dataset of 105 real-world building energy data transformation problems, and (3) conducting an extensive empirical evaluation where our approach achieved 96% accuracy in all 105 problems. SQLMorpher demonstrates the effectiveness of utilizing LLMs in complex, domain-specific challenges, highlighting the potential of their potential to drive sustainable solutions.","{'model': 'tldr@v2.0.0', 'text': 'This work designs an LLM-based framework that comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases, and implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection.'}",https://arxiv.org/pdf/2309.01957
-GRASS: Unified Generation Model for Speech-to-Semantic Tasks,Aobo Xia,"This paper explores the instruction fine-tuning technique for speech-to-semantic tasks by introducing a unified end-to-end (E2E) framework that generates target text conditioned on a task-related prompt for audio data. We pre-train the model using large and diverse data, where instruction-speech pairs are constructed via a text-to-speech (TTS) system. Extensive experiments demonstrate that our proposed model achieves state-of-the-art (SOTA) results on many benchmarks covering speech named entity recognition, speech sentiment analysis, speech question answering, and more, after fine-tuning. Furthermore, the proposed model achieves competitive performance in zero-shot and few-shot scenarios. To facilitate future work on instruction fine-tuning for speech-to-semantic tasks, we release our instruction dataset and code.","{'model': 'tldr@v2.0.0', 'text': 'A unified end-to-end (E2E) framework that generates target text conditioned on a task-related prompt for audio data is introduced that achieves state-of-the-art (SOTA) results on many benchmarks covering speech named entity recognition, speech sentiment analysis, speech question answering, and more, after fine-tuning.'}",https://arxiv.org/pdf/2309.02780
-Leveraging Vision-Language Foundation Models for Fine-Grained Downstream Tasks,Denis Coquenet,"Vision-language foundation models such as CLIP have shown impressive zero-shot performance on many tasks and datasets, especially thanks to their free-text inputs. However, they struggle to handle some downstream tasks, such as fine-grained attribute detection and localization. In this paper, we propose a multitask fine-tuning strategy based on a positive/negative prompt formulation to further leverage the capacities of the vision-language foundation models. Using the CLIP architecture as baseline, we show strong improvements on bird fine-grained attribute detection and localization tasks, while also increasing the classification performance on the CUB200-2011 dataset. We provide source code for reproducibility purposes: it is available at https://github.com/FactoDeepLearning/MultitaskVLFM.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multitask fine-tuning strategy based on a positive/negative prompt formulation to further leverage the capacities of the vision-language foundation models and shows strong improvements on bird fine-grained attribute detection and localization tasks.'}",https://arxiv.org/pdf/2307.06795
-Towards Expert Systems for Improved Customer Services Using ChatGPT as an Inference Engine,C. P. Ezenkwu,"By harnessing both implicit and explicit customer data, companies can develop a more comprehensive understanding of their consumers, leading to better customer engagement and experience, and improved loyalty. As a result, businesses have embraced many AI technologies, including chatbots, sentiment analysis, voice assistants, predictive analytics, and natural language processing, within customer services and e-commerce. The arrival of ChatGPT, a state-of-the-art deep learning model trained with general knowledge in mind, has brought about a paradigm shift in how companies approach AI applications. However, given that most business problems are bespoke and require specialised domain expertise, ChatGPT needs to be aligned with the requisite task-oriented ability to solve these issues. This paper presents an iterative procedure that incorporates expert system development process models and prompt engineering, in the design of descriptive knowledge and few-shot prompts, as are necessary for ChatGPT-powered expert systems applications within customer services. Furthermore, this paper explores potential application areas for ChatGPT-powered expert systems in customer services, presenting opportunities for their effective utilisation in the business sector.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents an iterative procedure that incorporates expert system development process models and prompt engineering, in the design of descriptive knowledge and few-shot prompts, as are necessary for ChatGPT-powered expert systems applications within customer services.'}",https://rgu-repository.worktribe.com/preview/1987218/EZENKWU%202023%20Towards%20expert%20systems%20%28AAM%29.pdf
-Generating Domain-Specific Programs for Diagram Authoring with Large Language Models,Rijul Jain,"Large language models (LLMs) can generate programs in general-purpose languages from prose descriptions, but are not trained on many domain-specific languages (DSLs). Diagram authoring with Penrose, a diagramming system using three DSLs, exemplifies the utility of DSL program generation with LLMs, which enables diagram creation from prose. We provide methods to conceptualize and evaluate the structures of one-shot LLM prompts to generate error-free DSL programs and implement Penrose diagram creation from prose using LLMs. We will evaluate our LLM prompt structures by testing prompt variations across different diagramming domains and plan to run a user study to assess the ease of LLM-augmented Penrose diagramming over other tools.","{'model': 'tldr@v2.0.0', 'text': 'Methods to conceptualize and evaluate the structures of one-shot LLM prompts to generate error-free DSL programs and implement Penrose diagram creation from prose using LLMs are provided.'}",
-Rewriting Math Word Problems with Large Language Models,Kole Norberg,"Large Language Models have recently achieved high performance on many writing tasks. In a recent study, math word problems in Carnegie Learning’s MATHia adaptive learning software were rewritten by human authors to improve their clarity and specificity. The randomized experiment found that emerging readers who received the rewritten word problems spent less time completing the problems and also achieved higher mastery compared to emerging readers who received the original content. We used GPT-4 to rewrite the same set of math word problems, prompting it to follow the same guidelines that the human authors followed. We lay out our prompt engineering process, comparing several prompting strategies: zero-shot, few-shot, and chain-of-thought prompting. Additionally, we overview how we leveraged GPT’s ability to write python code in order to encode mathematical components of word problems. We report text analysis of the original, human-rewritten, and GPT-rewritten problems. GPT rewrites had the most optimal readability, lexical diversity, and cohesion scores but used more low frequency words. We present our plan to test the GPT outputs in upcoming randomized field trials in MATHia.","{'model': 'tldr@v2.0.0', 'text': 'This work used GPT-4 to rewrite the same set of math word problems, prompting it to follow the same guidelines that the human authors followed, and found that GPT rewrites had the most optimal readability, lexical diversity, and cohesion scores but used more low frequency words.'}",
-Solving and Generating NPR Sunday Puzzles with Large Language Models,Jin Zhao,"We explore the ability of large language models to solve and generate puzzles from the NPR Sunday Puzzle game show using PUZZLEQA, a dataset comprising 15 years of on-air puzzles. We evaluate four large language models using PUZZLEQA, in both multiple choice and free response formats, and explore two prompt engineering techniques to improve free response performance: chain-of-thought reasoning and prompt summarization. We find that state-of-the-art large language models can solve many PUZZLEQA puzzles: the best model, GPT-3.5, achieves 50.2% loose accuracy. However, in our few-shot puzzle generation experiment, we find no evidence that models can generate puzzles: GPT-3.5 generates puzzles with answers that do not conform to the generated rules. Puzzle generation remains a challenging task for future work.","{'model': 'tldr@v2.0.0', 'text': 'It is found that state-of-the-art large language models can solve many PUZZLEQA puzzles: the best model, GPT-3.5, achieves 50.2% loose accuracy and Puzzle generation remains a challenging task for future work.'}",http://arxiv.org/pdf/2306.12255
-Effect of Delayed Refrigeration on the Microbial Carcass Contamination of Wild Boars (Sus scrofa),B. Cenci-Goga,"Simple Summary The bacteria that contaminate meat after the death of the animal can come from multiple sources, such as the animal itself, the external environment and the operator who handles it. The prompt refrigeration of hunted game is generally considered an important step to prevent spoilage and meat alterations, although many logistic hindrances, such as animal weight and distance from the hunting area to the refrigerators, limit the meticulous adoption of the best procedures. We show that the bacterial population of wild boar carcasses is not correlated to the mere time from shot to refrigeration but is correlated to the refrigeration time from chilling to analysis. The results of our study revealed a correlation between the time from shot to analysis and from refrigeration to analysis but a lack of correlation between the time from shot to refrigeration. Abstract The immediate refrigeration of meat after slaughter is a key issue for the proper storage and aging of meat. The industry standard cold chain relies on low temperatures and ventilation to lower the internal carcass temperature to 0–4 °C within the first 48 h, i.e., within four times the so-called semi-cooling time. On the other hand, for games, once bled and eviscerated, the carcass must be sent to a point where it can be sectioned or kept on air for maturation at refrigeration temperature. The precautions to observe are few and simple but essential: protect the meat and start the cooling process quickly. After preparing the animal (bleeding and evisceration), it may be necessary to face a period of transport that is sometimes long and not very easy; while small animals can be easily transported in a backpack, larger ones must necessarily be carried by several people or sometimes dragged to the vehicle capable of transporting them. It is obvious that a wild boar opened from the jaws to the pelvis and dragged for hundreds of meters will tend to be contaminated, although these contaminations are to be considered secondary for the preservation of the meat, compared to contamination by the intestinal contents. In an attempt to investigate the effect of delayed refrigeration on wild boar carcass contamination, the aim of this work was to determine a correlation between several hunting and logistic parameters (age, sex, animal weight, shooting distance, number of shots, weather and temperature and time from shot to refrigeration and to analysis) and bacterial contamination of the carcass. The correlation coefficient, r, was found to be 0.038 for the eviscerated body weight (p < 0.05), 0.091 for the external temperature on the day of hunting (p < 0.05), 0.027 for the time from shot to refrigeration (p = 0.081), 0.038 for the time from refrigeration to analysis (p < 0.05) and 0.043 for the time from shot to analysis (p < 0.05). These results stand for a negative correlation between the bacterial population and eviscerated carcass weight and between the bacterial population and external temperature and for a positive correlation between the time from shot to analysis and from refrigeration to analysis. No association was demonstrated between the bacterial population and the time from shot to refrigeration.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the bacterial population of wild boar carcasses is not correlation to the mere time from shot to refrigeration but is correlated to the refrigeration time from chilling to analysis, which stands for a negative correlation between theacterial population and eviscerated carcass weight and between the bacteria population and external temperature.'}",https://www.mdpi.com/2076-2615/11/5/1434/pdf?version=1621346402
-Eliciting Knowledge from Language Models for Event Extraction,Jiaju Lin,"Eliciting knowledge contained in language models via prompt-based learning has shown great potential in many natural language processing tasks, such as text classification and generation. Whereas, the applications for more complex tasks such as event extraction are less studied, since the design of prompt is not straightforward due to the complicated types and arguments. In this paper, we explore to elicit the knowledge from pre-trained language models for event trigger detection and argument extraction. Specifically, we present various joint trigger/argument prompt methods, which can elicit more complementary knowledge by modeling the interactions between different triggers or arguments. The experimental results on the benchmark dataset, namely ACE2005, show the great advantages of our proposed approach. In particular, our approach is superior to the recent advanced methods in the few-shot scenario where only a few samples are used for training.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents various joint trigger/argument prompt methods, which can elicit more complementary knowledge by modeling the interactions between different triggers or arguments.'}",
-Dynamic Sparsification for Quadratic Assignment Problems,Maximilian John,,"{'model': 'tldr@v2.0.0', 'text': 'An iterative algorithm is proposed that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsified linearization of the original problem in every iteration, which results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions inevery iteration.'}",
-HIV murders prompt activist anger.,K. Morris,"A South African teacher is thought to have murdered his wife and her mother then committed suicide after the couple tested positive for HIV. Mpho Motloung was found shot next to a note saying ""HIV positive AIDS"" alongside her dead mother and critically ill father. Her husband was found dead in the next room. The murders occurred the same day that the inquest of AIDS activist Gugu Dlamini was adjourned because police failed to attend or subpoena key witnesses report the Treatment Action Campaign (TAC). Dlaminis murder 18 months ago is thought related to her HIV-positive status. TAC asserts that the governments ""primitive HIV/AIDS messages"" promote fear and anger causing discrimination and violence against people with HIV/AIDS particularly women. ""Everywhere we go we hear: ""HIV/AIDS is a death sentence"" says Zackie Achmat of TAC. ""Government messages fail to tell the millions of people living with HIV that it is possible to live fully and productively for many years"". (full text)","{'model': 'tldr@v2.0.0', 'text': 'A South African teacher is thought to have murdered his wife and her mother then committed suicide after the couple tested positive for HIV, the same day that the inquest of AIDS activist Gugu Dlamini was adjourned.'}",
-Dynamic Sparsi cation for Quadratic Assignment Problems,Maximilian John,"We present a framework for optimizing sparse quadratic assignment problems. We propose an iterative algorithm that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsi ed linearization of the original problem in every iteration. This procedure results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions in every iteration. This framework was motivated by the task of the French government to design the French keyboard standard, which included solving sparse quadratic assignment problems with over 100 special characters; a size not feasible for many commonly used approaches. Designing a new standard often involves multiple stakeholders having con icting opinions and, hence, no agreement on a single well-de ned objective function to be used for an extensive one-shot optimization. Since the process of designing the standard is highly interactive, it demands rapid prototyping, e.g., quick primal solutions, on-they evaluation of manual changes, and prompt assessments of solution quality. Particularly concerning the latter aspect, our algorithm is able to provide high-quality lower bounds for these problems within only a few minutes.","{'model': 'tldr@v2.0.0', 'text': 'An iterative algorithm is proposed that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsi ed linearization of the original problem in every iteration, which results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions inevery iteration.'}",
-Development of an Expert System for the High Intensity Neutrino Beam Facility at J-PARC,K. Nakayoshi,"A high intensity neutrino beam produced at J-PARC is utilized by a long-baseline neutrino oscillation experiment. To generate a high intensity neutrino beam, a high intensity proton beam is extracted from the 30 GeV Main Ring synchrotron to the neutrino primary beamline. In the beamline, one mistaken shot can potentially do serious damage to beamline equipment. To avoid such a consequence, many beamline equipment interlocks to stop the beam operation are implemented. Once an interlock is activated, prompt and proper error handling is necessary. We are developing an expert system for prompt and efficient understanding of the status to quickly resume the beam operation. An inference engine is one key component in the expert system. Although a typical inference engine of the expert system is rule-based, we adapt a Machine-Learning (ML) based inference engine in our expert system. We will report the initial evaluation of our ML-based inference engine. INTRODUCTION The T2K (Tokai-to-Kamioka) experiment [1] is a longbaseline neutrino oscillation experiment at J-PARC (Japan Proton Accelerator Research Complex). Figure 1 shows the overview of the T2K experiment. A high intensity neutrino/anti-neutrino beam is produced and propagates 295 km from J-PARC to Super-Kamiokande (SK). In July 2013, muon neutrino to electron neutrino transformation was firmly established [2]. In August 2017, T2K excluded CP-conservation at 95% confidence level using the latest data. In order to keep generating interesting physics, steady operation of the facility is very important. Super‐Ka ioka de J‐PARC Near Dete tors Neutri o Bea 9 k Mt. Nogu hi‐Goro ,9 Mt. Ike o‐Ya a , , elo sea le el Figure 1: Overview of the T2K experiment. Figure 2 shows a layout of the neutrino experimental facility (neutrino facility) at J-PARC. The neutrino facility is composed of two beamlines and a near detector (ND280). The beamline consists of the primary and secondary beamlines. In the primary beamline, the high intensity proton beam is extracted from the Main Ring synchrotron (MR) and guided through super and normal conducting magnets ∗ kazuo.nakayoshi@kek.jp to the target station. In the secondary beamline, the proton beam hits a graphite target and produces pions. These pions decay into muons and muon neutrinos in a decay volume. The high intensity proton beam reached 470 kW in 2017 and ready to the design power of 750kW with a few years. 0 50 100 m Main Ring",,
-Empirical Evidence on the Relationship between Audit Probability and Internal Revenue Service Budget Levels,Akinloye Akindayomi,"Despite a large literature on tax avoidance strategies by individuals and the IRS ability to curb them, how the agency’s audit/examination activities respond to tax expenditures within the individual taxation context is not well understood. We hypothesize that IRS audit probability of individual tax returns will positively respond to the magnitude of different tax expenditure drivers if the agency has a shot at curtailing them. We find that while the probability of IRS audit increases as the agency’s budget on enforcement activities increases, tax expenditures do not appear to prompt IRS enforcement activities in many of the tax expenditures drivers.",,
-Development of a novel radiation imaging detector system for in vivo gene imaging in small animal studies,A. Weisenberger,"Many studies in molecular biology deal with following the expression and regulation of a gene at different stages of an organism's development or under different physiological conditions. Presently in situ hybridization and immunochemical assays are available to follow the gene expression at a single moment in time for one organism. One must sacrifice the organism to make a measurement, essentially taking a snap shot of the state of expression of the gene of interest. We have made progress on a new type of gene imaging technology which takes advantage of the emission properties of the radioisotope iodine 125 (/sup 125/I) as the probe and utilizes crystal scintillators and a position sensitive photomultiplier tube. Iodine 125 decays via electron capture emitting a 35 keV gamma-ray with the prompt emission of several 27-32 keV K/spl alpha/ and K/spl beta/ shell X-rays. Because of this a coincidence condition can be set to detect the /sup 125/I decays thus reducing background radiation contribution to the image. Mouse imaging studies of iodine uptake by the thyroid and melatonin receptor binding have been done with this detector system using low doses of /sup 125/I.","{'model': 'tldr@v2.0.0', 'text': 'Progress is made on a new type of gene imaging technology which takes advantage of the emission properties of the radioisotope iodine 125 (/sup 125/I) as the probe and utilizes crystal scintillators and a position sensitive photomultiplier tube.'}",
-Increased lead absorption and lead poisoning from a retained bullet.,V. Fiorica,"A young man was found to have elevated blood lead and erythrocyte protoporphyrin levels seven years after he had been shot in the foot with a .22 caliber bullet. Although clearly evident in radiographs obtained at the time of the accident, the bullet had never been removed. Over the course of the intervening years the radiographic appearance of the projectile changed dramatically. The patient underwent a surgical debridement procedure with some resolution of his lead intoxication. This case was reviewed in the context of the literature describing the disposition of retained lead projectiles. Lead bullets or fragments lodged within joints are slowly disintegrated and, as absorption accelerates, predispose the patient to the risk of lead poisoning which may occur many years after the original incident. Prompt removal of any lead projectile entering a joint space is strongly recommended.","{'model': 'tldr@v2.0.0', 'text': 'This case was reviewed in the context of the literature describing the disposition of retained lead projectiles and Prompt removal of any lead projectile entering a joint space is strongly recommended.'}",
-[Blood transfusion and consultation].,T. Tasaki,"Many hospital staff, including doctors, nurses, pharmacists, etc., are engaged in blood transfusion practice, and various inquiries are referred to the blood transfusion services. In order to provide a prompt and proper reply, transfusionists must have a wealth of knowledge and experience concerning blood transfusion medicine. Q & A relating to blood transfusion can be found on the home page of the Japanese Society of Blood Transfusion, and these are useful staff resources to obtain simple information. However, we sometimes encounter difficult problems in the management of a patient's treatment. Three representative transfusion-related issues are described in this article: (1) blood transfusion to patients with a positive DAT; (2) emergency transfusion, especially in cases where unexpected antibodies are encountered; (3) management of platelet transfusion refractoriness. Minimum standards for the investigation of transfusion-related adverse reactions developed by SHOT (Serious Hazards of Transfusion) are also introduced in this article, and these have a highly practical value. Finally, the importance of education in transfusion medicine is described. The number of doctors in Japan who are engaged exclusively in transfusion medicine is small, but blood transfusions are performed in every hospital, regardless of whether such a specialist is present. We have recently had to deal with a wide range of transplantation-related issues. Therefore, there is a great need for special education in transfusion medicine for doctors in the transplantation and cell therapy age.","{'model': 'tldr@v2.0.0', 'text': 'Three representative transfusion-related issues are described in this article: blood transfusion to patients with a positive DAT; emergency transfusion, especially in cases where unexpected antibodies are encountered; and management of platelet transfusion refractoriness.'}",
-"Dennis Judd, The Lion and the Tiger: The Rise and Fall of the British Raj, 1600–1947 . New York, NY: Oxford University Press, 2004. 234 pp. ISBN: 0-19-280358-1 (hbk.); 0-19-280579-7 (pbk.).",R. Wenzlhuemer,"damentally distinct domains of activity, namely religion and politics. His liberal assumptions preclude the possibility that his object of study might prompt investigation of a political modernity which, as in so many other parts of the world in the twentieth century (including Europe and the U.S.), does not conform to such an assumption and which instead comes to be shot through by competing ideologies of religion, culture, race, and national identity. One consequence ofjhis is the inadequate treatment of Congress socialism, which like so many Christian, Muslim, and national 'socialisms' elsewhere in the world, often had little in common with Marxian socialism, orthodox or otherwise. That Hinduist socialism should be widespread in North India in the 1930s and 1940s will not strike many readers as surprising as Gould seems to find it. Certainly, it need not have prompted him to alert his reader as to the supposedly peculiar penchant of Indian socialism for racialist or culturalist perspectives. By the same token, Gould's analysis would have benefited from greater attention being paid to parties commonly regarded as largely free of communalism, such as the Communist Party, which, after all, enjoyed significant strength in Kanpur and which, during precisely the period in question, fought for support among some of the same groups targeted by Congress and Muslim League organizational efforts. Finally, Gould's emphasis on the ubiquity of communalism in the Congress itself threatens to obscure the bases of support enjoyed by a (IP. Congressman like Jawaharlal Nehru, whose popularity as a writer and speaker unquestionably grew during just these years. Gould's volume verges on obscuring from view the anti-colonial struggle in (J.P., of which, after all, the Congress was the principal institutional agent. Nowhere is this omission more striking than in the volume's remarkable neglect of the development in and through the 1942 'Quit India' movement of the communal dynamics he is investigating, a fact which may derive from the relative paucity of the colonial archives from that period. Despite Gould's important departures from many Cambridge School orthodoxies, the one that seems to have left the strongest trace is the old deprecation of the mass-based and anti-colonial character of the Congressled anti-colonial movement. The result is that the reader is left wondering how exactly a divided and religiously pre-occupied movement such as the Congress that Gould describes ever managed to force the British out of India at all. Still, these misgivings notwithstanding, Hindu Nationalism and the Language of Politics in Late Colonial India is an important and long-overdue contribution to twentieth-century South Asian political history and it is to be hoped that Cambridge will soon bring out a paperback edition so that instructors can with a clear conscience require the text for advanced seminars on the subject.",,
-"Thoracic, abdominal, and peripheral trauma. Evaluation with angiography.",H. Redman,"PROMPT evaluation of the condition of patients who have suffered penetrating or blunt trauma is important in reducing morbidity and mortality. During the past few years, many diagnostic tests have been developed to aid the clinician, and use of the appropriate ones is important in facilitating proper patient management. Angiography has a specific role in some injuries but should be held in reserve in others. Both the type of trauma and the severity of the injury are important in deciding when to use angiography. Penetrating Trauma Most patients who have had a penetrating injury to the thorax or abdomen with a high- or low-velocity missile will need surgical exploration, and generally there is no need for angiography. Occasionally, when there have been multiple injuries, angiography may be used to decide which injuries should be approached first. For example, a patient had been shot through the right forearm and back at","{'model': 'tldr@v2.0.0', 'text': 'Angiography has a specific role in some injuries but should be held in reserve in others, and both the type of trauma and the severity of the injury are important in deciding when to use angiography.'}",
-Language Models are Few-Shot Learners,Tom B. Brown,"Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic.'}",
-Many-shot from Low-shot: Learning to Annotate using Mixed Supervision for Object Detection,C. Biffi,,"{'model': 'tldr@v2.0.0', 'text': 'This work introduces, for the first time, an online annotation module (OAM) that learns to generate a many-shot set of \\emph{reliable} annotations from a larger volume of weakly labelled images, resulting in a fully end-to-end strategy that only requires a low- shot set of fully annotated images.'}",https://arxiv.org/pdf/2008.09694
-CCPrompt: Counterfactual Contrastive Prompt-Tuning for Many-Class Classification,Y. Li,"With the success of the prompt-tuning paradigm in Natural Language Processing (NLP), various prompt templates have been proposed to further stimulate specific knowledge for serving downstream tasks, e.g., machine translation, text generation, relation extraction, and so on. Existing prompt templates are mainly shared among all training samples with the information of task description. However, training samples are quite diverse. The sharing task description is unable to stimulate the unique task-related information in each training sample, especially for tasks with the finite-label space. To exploit the unique task-related information, we imitate the human decision process which aims to find the contrastive attributes between the objective factual and their potential counterfactuals. Thus, we propose the \textbf{C}ounterfactual \textbf{C}ontrastive \textbf{Prompt}-Tuning (CCPrompt) approach for many-class classification, e.g., relation classification, topic classification, and entity typing. Compared with simple classification tasks, these tasks have more complex finite-label spaces and are more rigorous for prompts. First of all, we prune the finite label space to construct fact-counterfactual pairs. Then, we exploit the contrastive attributes by projecting training instances onto every fact-counterfactual pair. We further set up global prototypes corresponding with all contrastive attributes for selecting valid contrastive attributes as additional tokens in the prompt template. Finally, a simple Siamese representation learning is employed to enhance the robustness of the model. We conduct experiments on relation classification, topic classification, and entity typing tasks in both fully supervised setting and few-shot setting. The results indicate that our model outperforms former baselines.","{'model': 'tldr@v2.0.0', 'text': 'The Counterfactual-Contrastive-Prompt-Tuning approach for many-class classification, e.g., relation classification, topic classification, and entity typing, and results indicate that the model outperforms former baselines.'}",https://arxiv.org/pdf/2211.05987
-Zero-shot versus Many-shot: Unsupervised Texture Anomaly Detection,Toshimichi Aota,"Research on unsupervised anomaly detection (AD) has recently progressed, significantly increasing detection accuracy. This paper focuses on texture images and considers how few normal samples are needed for accurate AD. We first highlight the critical nature of the problem that previous studies have overlooked: accurate detection gets harder for anisotropic textures when image orientations are not aligned between inputs and normal samples. We then propose a zero-shot method, which detects anomalies without using a normal sample. The method is free from the issue of unaligned orientation between input and normal images. It assumes the input texture to be homogeneous, detecting image regions that break the homogeneity as anomalies. We present a quantitative criterion to judge whether this assumption holds for an input texture. Experimental results show the broad applicability of the proposed zero-shot method and its good performance comparable to or even higher than the state-of-the-art methods using hundreds of normal samples. The code and data are available from https://drive.google.com/drive/folders/10OyPzvI3H6llCZBxKxFlKWt1Pw1tkMK1.",,
-Delving Deep into Many-to-many Attention for Few-shot Video Object Segmentation,Haoxin Chen,"This paper tackles the task of Few-Shot Video Object Segmentation (FSVOS), i.e., segmenting objects in the query videos with certain class specified in a few labeled support images. The key is to model the relationship between the query videos and the support images for propagating the object information. This is a many-to-many problem and often relies on full-rank attention, which is computationally intensive. In this paper, we propose a novel Domain Agent Network (DAN), breaking down the full-rank attention into two smaller ones. We consider one single frame of the query video as the domain agent, bridging between the support images and the query video. Our DAN allows a linear space and time complexity as opposed to the original quadratic form with no loss of performance. In addition, we introduce a learning strategy by combining meta-learning with online learning to further improve the segmentation accuracy. We build a FSVOS benchmark on the Youtube-VIS dataset and conduct experiments to demonstrate that our method outperforms baselines on both computational cost and accuracy, achieving the state-of-the-art performance. Code is available at https://github.com/scutpaul/DANet.","{'model': 'tldr@v2.0.0', 'text': 'A novel Domain Agent Network (DAN) is proposed, breaking down the full-rank attention into two smaller ones, allowing a linear space and time complexity as opposed to the original quadratic form with no loss of performance.'}",
-Few-Shot and Many-Shot Fusion Learning in Mobile Visual Food Recognition,Heng Zhao,"Mobile visual food recognition is emerging as an important application in food logging and dietary monitoring in recent years. Existing food recognition methods use conventional many-shot learning to train a large backbone network, which refers to the use of sufficient number of training data to train the network. However, these methods firstly do not consider the cases where certain food categories have limited training data. Therefore, they cannot use the conventional training using many-shot learning. Further, existing solutions focus on improving the food recognition performance by implementing state-of-the-art large full networks, and do not pay much attention to reduce the size and computational cost of the network. As a result, they are not amenable for deployment on mobile devices. In this paper, we address these issues by proposing a new few-shot and many-shot fusion learning for mobile visual food recognition, it has a compact framework and is able to learn from existing dataset categories, and also new food categories given only a few sample images. We construct a new Indian food dataset called NTU-IndianFood107 in order to evaluate the performance of the proposed method. The dataset has two parts: (i) a Base Dataset of 83 classes of Indian food images with over 600 images per class to perform many-shot learning, and (ii) a Food Diary of 24 classes captured in restaurants with limited number to simulate the few-shot learning on new food categories. The proposed fusion method achieves a Top-1 classification accuracy of 72.0% on the new dataset.","{'model': 'tldr@v2.0.0', 'text': 'A new few-shot and many-shot fusion learning for mobile visual food recognition that has a compact framework and is able to learn from existing dataset categories, and also new food categories given only a few sample images is proposed.'}",
-Automatic Machine Translation Evaluation in Many Languages via Zero-Shot Paraphrasing,Brian Thompson,"We propose the use of a sequence-to-sequence paraphraser for automatic machine translation evaluation. The paraphraser takes a human reference as input and then force-decodes and scores an MT system output. We propose training the aforementioned paraphraser as a multilingual NMT system, treating paraphrasing as a zero-shot ""language pair"" (e.g., Russian to Russian). We denote our paraphraser ""unbiased"" because the mode of our model's output probability is centered around a copy of the input sequence, which in our case represent the best case scenario where the MT system output matches a human reference. Our method is simple and intuitive, and our single model (trained in 39 languages) outperforms or statistically ties with all prior metrics on the WMT19 segment-level shared metrics task in all languages, excluding Gujarati where the model had no training data. We also explore using our model conditioned on the source instead of the reference, and find that it outperforms every quality estimation as a metric system from the WMT19 shared task on quality estimation by a statistically significant margin in every language pair.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the use of a sequence-to-sequence paraphraser for automatic machine translation evaluation, and finds that the model conditioned on the source instead of the reference outperforms every quality estimation as a metric system from the WMT19 shared task on quality estimation by a statistically significant margin in every language pair.'}",https://www.aclweb.org/anthology/2020.emnlp-main.8.pdf
-Many-Class Few-Shot Learning on Multi-Granularity Class Hierarchy,Lu Liu,"We study many-class few-shot (MCFS) problem in both supervised learning and meta-learning settings. Compared to the well-studied many-class many-shot and few-class few-shot problems, the MCFS problem commonly occurs in practical applications but has been rarely studied in previous literature. It brings new challenges of distinguishing between many classes given only a few training samples per class. In this article, we leverage the class hierarchy as a prior knowledge to train a coarse-to-fine classifier that can produce accurate predictions for MCFS problem in both settings. The propose model, “memory-augmented hierarchical-classification network (MahiNet)”, performs coarse-to-fine classification where each coarse class can cover multiple fine classes. Since it is challenging to directly distinguish a variety of fine classes given few-shot data per class, MahiNet starts from learning a classifier over coarse-classes with more training data whose labels are much cheaper to obtain. The coarse classifier reduces the searching range over the fine classes and thus alleviates the challenges from “many classes”. On architecture, MahiNet first deploys a convolutional neural network (CNN) to extract features. It then integrates a memory-augmented attention module and a multi-layer perceptron (MLP) together to produce the probabilities over coarse and fine classes. While the MLP extends the linear classifier, the attention module extends the KNN classifier, both together targeting the “few-shot” problem. We design several training strategies of MahiNet for supervised learning and meta-learning. In addition, we propose two novel benchmark datasets “mcfsImageNet” (as a subset of ImageNet) and “mcfsOmniglot” (re-splitted Omniglot) specially designed for MCFS problem. In experiments, we show that MahiNet outperforms several state-of-the-art models (e.g., prototypical networks and relation networks) on MCFS problems in both supervised learning and meta-learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed model, “memory-augmented hierarchical-classification network (MahiNet)”, performs coarse-to-fine classification where each coarse class can cover multiple fine classes and outperforms several state-of-the-art models on MCFS problems in both supervised learning and meta-learning.'}",https://opus.lib.uts.edu.au/bitstream/10453/147743/3/Many-Class%20Few-Shot%20Learning.pdf
-Generalized Many-Way Few-Shot Video Classification,Yongqin Xian,,"{'model': 'tldr@v2.0.0', 'text': 'A simple 3D CNN baseline is developed, surpassing existing methods by a large margin and proposed to leverage weakly-labeled videos from a large dataset using tag retrieval followed by selecting the best clips with visual similarities, yielding further improvement.'}",https://pure.mpg.de/pubman/item/item_3267299_1/component/file_3267300/arXiv%3A2007.04755.pdf
-Tailor-made compositional gradient copolymer by a many-shot RAFT emulsion polymerization method,Yunlong Guo,"A many-shot RAFT emulsion polymerization method is proposed to synthesize gradient copolymers with high molecular weight and a tailor-made compositional gradient. In this method each shot consisting of comonomers with pre-set different fractions and targeting the molecular weight of 10 000 g mol−1 was added in a stepwise manner during the reaction. High conversions over 95% were achieved in 35 min after each shot. The compositional variation along the polymer chain was then directly determined by the comonomer fractions added at each shot. Styrene/n-butyl acrylate gradient copolymers (including linear and V-shaped gradient) with molecular weights as high as 90 000 g mol−1 were prepared by this method. The composition profiles along the polymer chains agreed well with the theoretical predictions, and the composition distribution among the polymer chains was narrow. The gradient copolymers showed different thermal and phase separation properties from their block counterparts, as expected. These results demonstrated the successful tailor-making of the gradient copolymers. The current strategy will act as a facile method to prepare tailor-made gradient copolymers with high molecular weights and within a short time.",,
-LLM-Planner: Few-Shot Grounded Planning for Embodied Agents with Large Language Models,Chan Hee Song,"This study focuses on using large language models (LLMs) as a planner for embodied agents that can follow natural language instructions to complete complex tasks in a visually-perceived environment. The high data cost and poor sample efficiency of existing methods hinders the development of versatile agents that are capable of many tasks and can learn new tasks quickly. In this work, we propose a novel method, LLM-Planner, that harnesses the power of large language models to do few-shot planning for embodied agents. We further propose a simple but effective way to enhance LLMs with physical grounding to generate and update plans that are grounded in the current environment. Experiments on the ALFRED dataset show that our method can achieve very competitive few-shot performance: Despite using less than 0.5% of paired training data, LLM-Planner achieves competitive performance with recent baselines that are trained using the full training data. Existing methods can barely complete any task successfully under the same few-shot setting. Our work opens the door for developing versatile and sample-efficient embodied agents that can quickly learn many tasks. Website: https://dki-lab.github.io/LLM-Planner","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method, LLM-Planner, that harnesses the power of large language models to do few-shot planning for embodied agents and proposes a simple but effective way to enhance LLMs with physical grounding to generate and update plans that are grounded in the current environment.'}",https://arxiv.org/pdf/2212.04088
-Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis,Ajay Jain,"We present DietNeRF, a 3D neural scene representation estimated from a few images. Neural Radiance Fields (NeRF) learn a continuous volumetric representation of a scene through multi-view consistency, and can be rendered from novel viewpoints by ray casting. While NeRF has an impressive ability to reconstruct geometry and fine details given many images, up to 100 for challenging 360° scenes, it often finds a degenerate solution to its image reconstruction objective when only a few input views are available. To improve few-shot quality, we propose DietNeRF. We introduce an auxiliary semantic consistency loss that encourages realistic renderings at novel poses. DietNeRF is trained on individual scenes to (1) correctly render given input views from the same pose, and (2) match high-level semantic attributes across different, random poses. Our semantic loss allows us to supervise DietNeRF from arbitrary poses. We extract these semantics using a pre-trained visual encoder such as CLIP, a Vision Transformer trained on hundreds of millions of diverse single-view, 2D photographs mined from the web with natural language supervision. In experiments, DietNeRF improves the perceptual quality of few-shot view synthesis when learned from scratch, can render novel views with as few as one observed image when pre-trained on a multi-view dataset, and produces plausible completions of completely unobserved regions. Our project website is available at https://www.ajayj.com/dietnerf.","{'model': 'tldr@v2.0.0', 'text': 'DietNeRF improves the perceptual quality of few-shot view synthesis when learned from scratch, can render novel views with as few as one observed image when pre-trained on a multi-view dataset, and produces plausible completions of completely unobserved regions.'}",https://arxiv.org/pdf/2104.00677
-What does a platypus look like? Generating customized prompts for zero-shot image classification,Sarah Pratt,"Open-vocabulary models are a promising new paradigm for image classification. Unlike traditional classification models, open-vocabulary models classify among any arbitrary set of categories specified with natural language during inference. This natural language, called""prompts"", typically consists of a set of hand-written templates (e.g.,""a photo of a {}"") which are completed with each of the category names. This work introduces a simple method to generate higher accuracy prompts, without relying on any explicit knowledge of the task domain and with far fewer hand-constructed sentences. To achieve this, we combine open-vocabulary models with large language models (LLMs) to create Customized Prompts via Language models (CuPL, pronounced""couple""). In particular, we leverage the knowledge contained in LLMs in order to generate many descriptive sentences that contain important discriminating characteristics of the image categories. This allows the model to place a greater importance on these regions in the image when making predictions. We find that this straightforward and general approach improves accuracy on a range of zero-shot image classification benchmarks, including over one percentage point gain on ImageNet. Finally, this simple baseline requires no additional training and remains completely zero-shot. Code available at https://github.com/sarahpratt/CuPL.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a simple method to generate higher accuracy prompts, without relying on any explicit knowledge of the task domain and with far fewer hand-constructed sentences, and finds that this straightforward and general approach improves accuracy on a range of zero-shot image classification benchmarks, including over one percentage point gain on ImageNet.'}",https://arxiv.org/pdf/2209.03320
-A UTO G RAPHEX : Zero-shot Biomedical Definition Generation with Automatic Prompting,,"Describing terminologies with definition texts 001 is an important step towards understanding 002 the scientific literature, especially for domains 003 with limited labeled terminologies. Previous 004 works have sought to design supervised neural 005 text generation models to solve the biomedi- 006 cal terminology generation task, but most of 007 them failed to define never-before-seen termi- 008 nologies in newly emerging research fields. 009 Here, we tackle this challenge by introducing 010 a zero-shot definition generation model based 011 on prompting , a recent approach for eliciting 012 knowledge from pre-trained language models, 013 with automatically generated prompts. Fur- 014 thermore, we enhanced the biomedical termi- 015 nology dataset by adding descriptive texts to 016 each biomedical subdiscipline, thus enabling 017 zero-shot learning scenarios. Our model out- 018 performed existing supervised baseline and the 019 baseline pre-trained language model that em- 020 ploys manually crafted prompts by up to 52 and 021 6 BLEU score, respectively. 022","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a zero-shot definition generation model based on prompting, a recent approach for eliciting knowledge from pre-trained language models, thus enabling zero- shot learning scenarios in biomedical termi- 015 nology dataset.'}",
-Better Zero-Shot Reasoning with Role-Play Prompting,Aobo Kong,"Modern large language models (LLMs), such as ChatGPT, exhibit a remarkable capacity for role-playing, enabling them to embody not only human characters but also non-human entities like a Linux terminal. This versatility allows them to simulate complex human-like interactions and behaviors within various contexts, as well as to emulate specific objects or systems. While these capabilities have enhanced user engagement and introduced novel modes of interaction, the influence of role-playing on LLMs' reasoning abilities remains underexplored. In this study, we introduce a strategically designed role-play prompting methodology and assess its performance under the zero-shot setting across twelve diverse reasoning benchmarks, encompassing arithmetic, commonsense reasoning, symbolic reasoning, and more. Leveraging models such as ChatGPT and Llama 2, our empirical results illustrate that role-play prompting consistently surpasses the standard zero-shot approach across most datasets. Notably, accuracy on AQuA rises from 53.5% to 63.8%, and on Last Letter from 23.8% to 84.2%. Beyond enhancing contextual understanding, we posit that role-play prompting serves as an implicit Chain-of-Thought (CoT) trigger, thereby improving the quality of reasoning. By comparing our approach with the Zero-Shot-CoT technique, which prompts the model to""think step by step"", we further demonstrate that role-play prompting can generate a more effective CoT. This highlights its potential to augment the reasoning capabilities of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A strategically designed role-play prompting methodology is introduced and its performance under the zero-shot setting across twelve diverse reasoning benchmarks, encompassing arithmetic, commonsense reasoning, symbolic reasoning, and more, demonstrating its potential to augment the reasoning capabilities of LLMs.'}",https://arxiv.org/pdf/2308.07702
-Zero-Shot Slot Filling with Slot-Prefix Prompting and Attention Relationship Descriptor,Qiaoyang Luo,"This paper addresses zero-shot slot filling, which tries to build a system that can generalize to unseen slot types without any training data. The key to zero-shot slot-filling is to match the tokens from the utterance with the semantic definition of the slot without training data in the target domain. This paper tackles this problem by devising a scheme to fully leverage pre-trained language models (PLMs). To this end, we propose a new prompting scheme that utilizes both learnable tokens and slot names to guide the model to focus on the relevant text spans for a given slot. Furthermore, we use attention values between tokens to form a feature descriptor for each token, which is motivated by the fact that the attention value in a PLM naturally characterizes various relationships, e.g., syntactic or semantic, between tokens. By further consolidating those features with an additional transformer-based aggregation module, we create a simple-but-effective zero-shot slot filling system that can achieve significantly better performance than the previous methods, as demonstrated by our experimental studies.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new prompting scheme that utilizes both learnable tokens and slot names to guide the model to focus on the relevant text spans for a given slot, and uses attention values between tokens to form a feature descriptor for each token.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26566/26338
-Relational Representation Learning for Zero-Shot Relation Extraction with Instance Prompting and Prototype Rectification,Bin Duan,"Zero-shot relation extraction aims to extract novel relations that are not observed beforehand. However, existing representation methods are not pre-trained for relational representations and embeddings contain much linguistic information, the distances between them are not consistent with relational semantic similarity. In this paper, we propose a novel method based on Instance Prompting and Prototype Rectification (IPPR) to conduct relational representation learning for zeroshot relation extraction. Instance prompting is designed to reduce the gap between pre-training and fine-tuning, and guide the pre-trained model to generate relation-oriented instance representations. Prototype rectification aims to push the prototype embeddings away from each other and makes the instance embeddings closer to its corresponding prototype embeddings for dynamically rectifying the prototype embeddings. Experimental results on two public datasets demonstrate that our proposed method achieves new state-of-the-arts performance1.","{'model': 'tldr@v2.0.0', 'text': 'A novel method based on Instance Prompting and Prototype Rectification to conduct relational representation learning for zeroshot relation extraction and achieves new state-of-the-arts performance.'}",
-Distilling Hypernymy Relations from Language Models: On the Effectiveness of Zero-Shot Taxonomy Induction,Devansh Jain,"In this paper, we analyze zero-shot taxonomy learning methods which are based on distilling knowledge from language models via prompting and sentence scoring. We show that, despite their simplicity, these methods outperform some supervised strategies and are competitive with the current state-of-the-art under adequate conditions. We also show that statistical and linguistic properties of prompts dictate downstream performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes zero-shot taxonomy learning methods which are based on distilling knowledge from language models via prompting and sentence scoring and shows that these methods outperform some supervised strategies and are competitive with the current state-of-the-art under adequate conditions.'}",https://aclanthology.org/2022.starsem-1.13.pdf
-Prompting Scientific Names for Zero-Shot Species Recognition,Shubham Parashar,"Trained on web-scale image-text pairs, Vision-Language Models (VLMs) such as CLIP can recognize images of common objects in a zero-shot fashion. However, it is underexplored how to use CLIP for zero-shot recognition of highly specialized concepts, e.g., species of birds, plants, and animals, for which their scientific names are written in Latin or Greek. Indeed, CLIP performs poorly for zero-shot species recognition with prompts that use scientific names, e.g.,""a photo of Lepus Timidus""(which is a scientific name in Latin). Because these names are usually not included in CLIP's training set. To improve performance, prior works propose to use large-language models (LLMs) to generate descriptions (e.g., of species color and shape) and additionally use them in prompts. We find that they bring only marginal gains. Differently, we are motivated to translate scientific names (e.g., Lepus Timidus) to common English names (e.g., mountain hare) and use such in the prompts. We find that common names are more likely to be included in CLIP's training set, and prompting them achieves 2$\sim$5 times higher accuracy on benchmarking datasets of fine-grained species recognition.","{'model': 'tldr@v2.0.0', 'text': ""It is found that common names are more likely to be included in CLIP's training set, and prompting them achieves 2$\\sim$5 times higher accuracy on benchmarking datasets of fine-grained species recognition.""}",
-A Universal Discriminator for Zero-Shot Generalization,Haike Xu,"Generative modeling has been the dominant approach for large-scale pretraining and zero-shot generalization. In this work, we challenge this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks. Technically, we train a single discriminator to predict whether a text sample comes from the true data distribution, similar to GANs. Since many NLP tasks can be formulated as selecting from a few options, we use this discriminator to predict the concatenation of input and which option has the highest probability of coming from the true data distribution. This simple formulation achieves state-of-the-art zero-shot results on the T0 benchmark, outperforming T0 by 16.0%, 7.8%, and 11.5% respectively on different scales. In the finetuning setting, our approach also achieves new state-of-the-art results on a wide range of NLP tasks, with only 1/4 parameters of previous methods. Meanwhile, our approach requires minimal prompting efforts, which largely improves robustness and is essential for real-world applications. Furthermore, we also jointly train a generalized UD in combination with generative tasks, which maintains its advantage on discriminative tasks and simultaneously works on generative tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work challenges this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks, and jointly train a generalized UD in combination with generative tasks, which maintains its advantage on discrim inative tasks and simultaneously works onGenerative tasks.'}",http://arxiv.org/pdf/2211.08099
-Zero-Shot Next-Item Recommendation using Large Pretrained Language Models,Lei Wang,"Large language models (LLMs) have achieved impressive zero-shot performance in various natural language processing (NLP) tasks, demonstrating their capabilities for inference without training examples. Despite their success, no research has yet explored the potential of LLMs to perform next-item recommendations in the zero-shot setting. We have identified two major challenges that must be addressed to enable LLMs to act effectively as recommenders. First, the recommendation space can be extremely large for LLMs, and LLMs do not know about the target user's past interacted items and preferences. To address this gap, we propose a prompting strategy called Zero-Shot Next-Item Recommendation (NIR) prompting that directs LLMs to make next-item recommendations. Specifically, the NIR-based strategy involves using an external module to generate candidate items based on user-filtering or item-filtering. Our strategy incorporates a 3-step prompting that guides GPT-3 to carry subtasks that capture the user's preferences, select representative previously watched movies, and recommend a ranked list of 10 movies. We evaluate the proposed approach using GPT-3 on MovieLens 100K dataset and show that it achieves strong zero-shot performance, even outperforming some strong sequential recommendation models trained on the entire training dataset. These promising results highlight the ample research opportunities to use LLMs as recommenders. The code can be found at https://github.com/AGI-Edgerunners/LLM-Next-Item-Rec.","{'model': 'tldr@v2.0.0', 'text': 'A prompting strategy called Zero-Shot Next-Item Recommendation (NIR) prompting that directs LLMs to make next-item recommendations that achieves strong zero-shot performance, even outperforming some strong sequential recommendation models trained on the entire training dataset.'}",http://arxiv.org/pdf/2304.03153
-The Wall Street Neophyte: A Zero-Shot Analysis of ChatGPT Over MultiModal Stock Movement Prediction Challenges,Qianqian Xie,"Recently, large language models (LLMs) like ChatGPT have demonstrated remarkable performance across a variety of natural language processing tasks. However, their effectiveness in the financial domain, specifically in predicting stock market movements, remains to be explored. In this paper, we conduct an extensive zero-shot analysis of ChatGPT's capabilities in multimodal stock movement prediction, on three tweets and historical stock price datasets. Our findings indicate that ChatGPT is a""Wall Street Neophyte""with limited success in predicting stock movements, as it underperforms not only state-of-the-art methods but also traditional methods like linear regression using price features. Despite the potential of Chain-of-Thought prompting strategies and the inclusion of tweets, ChatGPT's performance remains subpar. Furthermore, we observe limitations in its explainability and stability, suggesting the need for more specialized training or fine-tuning. This research provides insights into ChatGPT's capabilities and serves as a foundation for future work aimed at improving financial market analysis and prediction by leveraging social media sentiment and historical stock data.","{'model': 'tldr@v2.0.0', 'text': 'It is indicated that ChatGPT is a ""Wall Street Neophyte"" with limited success in predicting stock movements, as it underperforms not only state-of-the-art methods but also traditional methods like linear regression using price features.'}",http://arxiv.org/pdf/2304.05351
-SelfCheck: Using LLMs to Zero-Shot Check Their Own Step-by-Step Reasoning,Ning Miao,"The recent progress in large language models (LLMs), especially the invention of chain-of-thought prompting, has made it possible to automatically answer questions by stepwise reasoning. However, when faced with more complicated problems that require non-linear thinking, even the strongest LLMs make mistakes. To address this, we explore whether LLMs are able to recognize errors in their own step-by-step reasoning, without resorting to external resources. To this end, we propose SelfCheck, a general-purpose zero-shot verification schema for recognizing such errors. We then use the results of these checks to improve question-answering performance by conducting weighted voting on multiple solutions to the question. We test SelfCheck on three datasets (GSM8K, MathQA, and MATH) and find that it successfully recognizes errors and, in turn, increases final answer accuracies.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes SelfCheck, a general-purpose zero-shot verification schema for recognizing errors in large language models and uses the results of these checks to improve question-answering performance by conducting weighted voting on multiple solutions to the question.'}",https://arxiv.org/pdf/2308.00436
-C3: Zero-shot Text-to-SQL with ChatGPT,Xuemei Dong,"This paper proposes a ChatGPT-based zero-shot Text-to-SQL method, dubbed C3, which achieves 82.3\% in terms of execution accuracy on the holdout test set of Spider and becomes the state-of-the-art zero-shot Text-to-SQL method on the Spider Challenge. C3 consists of three key components: Clear Prompting (CP), Calibration with Hints (CH), and Consistent Output (CO), which are corresponding to the model input, model bias and model output respectively. It provides a systematic treatment for zero-shot Text-to-SQL. Extensive experiments have been conducted to verify the effectiveness and efficiency of our proposed method.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2307.07306
-Tab-CoT: Zero-shot Tabular Chain of Thought,Ziqi Jin,"The chain-of-though (CoT) prompting methods were successful in various natural language processing (NLP) tasks thanks to their ability to unveil the underlying complex reasoning processes. Such reasoning processes typically exhibit implicitly structured steps. Recent efforts also started investigating methods to encourage more explicitly structured reasoning procedures to be captured. In this work, we propose Tab-CoT, a novel tabular-format CoT prompting method, which allows the complex reasoning process to be explicitly modelled in a highly structured manner. Despite its simplicity, we show that our approach is capable of performing reasoning across multiple dimensions (i.e., both rows and columns). We demonstrate our approach's strong zero-shot and few-shot capabilities through extensive experiments on a range of reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Tab-CoT is proposed, a novel tabular-format CoT prompting method, which allows the complex reasoning process to be explicitly modelled in a highly structured manner and is capable of performing reasoning across multiple dimensions.'}",http://arxiv.org/pdf/2305.17812
-The Benefits of Label-Description Training for Zero-Shot Text Classification,Lingyu Gao,"Pretrained language models have improved zero-shot text classification by allowing the transfer of semantic knowledge from the training data in order to classify among specific label sets in downstream tasks. We propose a simple way to further improve zero-shot accuracies with minimal effort. We curate small finetuning datasets intended to describe the labels for a task. Unlike typical finetuning data, which has texts annotated with labels, our data simply describes the labels in language, e.g., using a few related terms, dictionary/encyclopedia entries, and short templates. Across a range of topic and sentiment datasets, our method is more accurate than zero-shot by 17-19% absolute. It is also more robust to choices required for zero-shot classification, such as patterns for prompting the model to classify and mappings from labels to tokens in the model's vocabulary. Furthermore, since our data merely describes the labels but does not use input texts, finetuning on it yields a model that performs strongly on multiple text domains for a given label set, even improving over few-shot out-of-domain classification in multiple settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple way to further improve zero-shot accuracies with minimal effort by curating small finetuning datasets intended to describe the labels for a task, which yields a model that performs strongly on multiple text domains for a given label set, even improving over few-shot out-of-domain classification in multiple settings.'}",http://arxiv.org/pdf/2305.02239
-Self-ICL: Zero-Shot In-Context Learning with Self-Generated Demonstrations,Wei-Lin Chen,"Large language models (LLMs) have exhibited striking in-context learning (ICL) ability to adapt to target tasks with a few input-output demonstrations. For better ICL, different methods are proposed to select representative demonstrations from existing training corpora. However, such settings are not aligned with real-world practices, as end-users usually query LMs without access to demonstration pools. In this work, we introduce Self-ICL -- a simple framework which bootstraps LMs' intrinsic capabilities to perform zero-shot ICL. Given a test input, Self-ICL first prompts the model to generate pseudo-inputs. Next, the model predicts pseudo-labels for the pseudo-inputs via zero-shot prompting. Finally, we perform ICL for the test input with the pseudo-input-label pairs as demonstrations. Evaluation on 23 BIG-Bench Hard tasks shows Self-ICL outperforms zero-shot baselines on both average accuracy and head-to-head comparison. Moreover, with zero-shot chain-of-thought, Self-ICL achieves results comparable to using real demonstrations. Additionally, we conduct a range of analyses to validate Self-ICL's effectiveness and provide insights for its behaviors under different settings.","{'model': 'tldr@v2.0.0', 'text': ""Self-ICL is introduced -- a simple framework which bootstraps LMs' intrinsic capabilities to perform zero-shot ICL and achieves results comparable to using real demonstrations with a range of analyses to validate Self- ICL's effectiveness and provide insights for its behaviors under different settings.""}",http://arxiv.org/pdf/2305.15035
-OntoType: Ontology-Guided Zero-Shot Fine-Grained Entity Typing with Weak Supervision from Pre-Trained Language Models,Tanay Komarlu,"Fine-grained entity typing (FET), which assigns entities in text with context-sensitive, fine-grained semantic types, will play an important role in natural language understanding. A supervised FET method, which typically relies on human-annotated corpora for training, is costly and difficult to scale. Recent studies leverage pre-trained language models (PLMs) to generate rich and context-aware weak supervision for FET. However, a PLM may still generate a mixture of rough and fine-grained types, or tokens unsuitable for typing. In this study, we vision that an ontology provides a semantics-rich, hierarchical structure, which will help select the best results generated by multiple PLM models and head words. Specifically, we propose a novel zero-shot, ontology-guided FET method, OntoType, which follows a type ontological structure, from coarse to fine, ensembles multiple PLM prompting results to generate a set of type candidates, and refines its type resolution, under the local context with a natural language inference model. Our experiments on the Ontonotes, FIGER, and NYT datasets using their associated ontological structures demonstrate that our method outperforms the state-of-the-art zero-shot fine-grained entity typing methods. Our error analysis shows that refinement of the existing ontology structures will further improve fine-grained entity typing.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot, ontology-guided FET method, OntoType, which follows a type ontological structure, from coarse to fine, ensembles multiple PLM prompting results to generate a set of type candidates, and refines its type resolution, under the local context with a natural language inference model.'}",http://arxiv.org/pdf/2305.12307
-Jack-Ryder at SemEval-2023 Task 5: Zero-Shot Clickbait Spoiling by Rephrasing Titles as Questions,Dirk Wangsadirdja,"In this paper, we describe our approach to the clickbait spoiling task of SemEval 2023.The core idea behind our system is to leverage pre-trained models capable of Question Answering (QA) to extract the spoiler from article texts based on the clickbait title without any task-specific training.Since oftentimes, these titles are not phrased as questions, we automatically rephrase the clickbait titles as questions in order to better suit the pretraining task of the QA-capable models.Also, to fit as much relevant context into the model’s limited input size as possible, we propose to reorder the sentences by their relevance using a semantic similarity model.Finally, we evaluate QA as well as text generation models (via prompting) to extract the spoiler from the text.Based on the validation data, our final model selects each of these components depending on the spoiler type and achieves satisfactory zero-shot results.The ideas described in this paper can easily be applied in fine-tuning settings.","{'model': 'tldr@v2.0.0', 'text': 'The core idea behind the system is to leverage pre-trained models capable of Question Answering to extract the spoiler from article texts based on the clickbait title without any task-specific training.'}",https://aclanthology.org/2023.semeval-1.150.pdf
-Are Large Language Models Robust Zero-shot Coreference Resolvers?,Nghia T. Le,"Recent progress in domain adaptation for coreference resolution relies on continued training using annotated data from target domains. At the same time, pre-trained large language models (LMs) have exhibited strong zero- and few-shot learning abilities across a wide range of NLP tasks including pronoun resolution. While this demonstrates evidence of coreference ability, previous work has mostly studied this ability using simple sentence-level datasets such as the Winograd Schema Challenge. In this work, we assess the feasibility of zero-shot learning for coreference resolution by evaluating instruction-tuned language models on more difficult, linguistically-complex coreference benchmarks (e.g., CoNLL-2012). We demonstrate that zero-shot prompting outperforms current unsupervised coreference systems. Further investigations reveal the robust zero-shot generalization ability of instruction-tuned LMs across a wide range of domains, languages, and time periods, as well as a strong reliance on high-quality mention detection systems.","{'model': 'tldr@v2.0.0', 'text': 'This work assesses the feasibility of zero- shot learning for coreference resolution by evaluating instruction-tuned language models on more difficult, linguistically-complex coreference benchmarks (e.g., CoNLL-2012), and demonstrates that zero-shot prompting outperforms current unsupervised coreference systems.'}",http://arxiv.org/pdf/2305.14489
-A Zero-Shot Language Agent for Computer Control with Structured Reflection,Tao Li,"Large language models (LLMs) have shown increasing capacity at planning and executing a high-level goal in a live computer environment (e.g. MiniWoB++). To perform a task, recent works often require a model to learn from trace examples of the task via either supervised learning or few/many-shot prompting. Without these trace examples, it remains a challenge how an agent can autonomously learn and improve its control on a computer, which limits the ability of an agent to perform a new task. We approach this problem with a zero-shot agent that requires no given expert traces. Our agent plans for executable actions on a partially observed environment, and iteratively progresses a task by identifying and learning from its mistakes via self-reflection and structured thought management. On the easy tasks of MiniWoB++, we show that our zero-shot agent often outperforms recent SoTAs, with more efficient reasoning. For tasks with more complexity, our reflective agent performs on par with prior best models, even though previous works had the advantages of accessing expert traces or additional screen information.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a zero-shot agent that plans for executable actions on a partially observed environment, and iteratively progresses a task by identifying and learning from its mistakes via self-reflection and structured thought management, which often outperforms recent SoTAs with more efficient reasoning.'}",https://arxiv.org/pdf/2310.08740
-AnoVL: Adapting Vision-Language Models for Unified Zero-shot Anomaly Localization,Hanqiu Deng,"Contrastive Language-Image Pre-training (CLIP) models have shown promising performance on zero-shot visual recognition tasks by learning visual representations under natural language supervision. Recent studies attempt the use of CLIP to tackle zero-shot anomaly detection by matching images with normal and abnormal state prompts. However, since CLIP focuses on building correspondence between paired text prompts and global image-level representations, the lack of patch-level vision to text alignment limits its capability on precise visual anomaly localization. In this work, we introduce a training-free adaptation (TFA) framework of CLIP for zero-shot anomaly localization. In the visual encoder, we innovate a training-free value-wise attention mechanism to extract intrinsic local tokens of CLIP for patch-level local description. From the perspective of text supervision, we particularly design a unified domain-aware contrastive state prompting template. On top of the proposed TFA, we further introduce a test-time adaptation (TTA) mechanism to refine anomaly localization results, where a layer of trainable parameters in the adapter is optimized using TFA's pseudo-labels and synthetic noise-corrupted tokens. With both TFA and TTA adaptation, we significantly exploit the potential of CLIP for zero-shot anomaly localization and demonstrate the effectiveness of our proposed methods on various datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a training-free adaptation (TFA) framework of CLIP for zero-shot anomaly localization and designs a unified domain-aware contrastive state prompting template for text supervision, and introduces a test-time adaptation mechanism to refine anomaly localization results.'}",https://arxiv.org/pdf/2308.15939
-A Monte Carlo Language Model Pipeline for Zero-Shot Sociopolitical Event Extraction,Erica Cai,"We consider dyadic zero-shot event extraction (EE) to identify actions between pairs of actors. The \emph{zero-shot} setting allows social scientists or other non-computational researchers to extract any customized, user-specified set of events without training, resulting in a \emph{dyadic} event database, allowing insight into sociopolitical relational dynamics among actors and the higher level organizations or countries they represent. Unfortunately, we find that current zero-shot EE methods perform poorly for the task, with issues including word sense ambiguity, modality mismatch, and efficiency. Straightforward application of large language model prompting typically performs even worse. We address these challenges with a new fine-grained, multi-stage generative question-answer method, using a Monte Carlo approach to exploit and overcome the randomness of generative outputs. It performs 90\% fewer queries than a previous approach, with strong performance on the widely-used Automatic Content Extraction dataset. Finally, we extend our method to extract affiliations of actor arguments and demonstrate our method and findings on a dyadic international relations case study.","{'model': 'tldr@v2.0.0', 'text': 'A new fine-grained, multi-stage generative question-answer method, using a Monte Carlo approach to exploit and overcome the randomness of generative outputs to identify actions between pairs of actors.'}",http://arxiv.org/pdf/2305.15051
-LT at SemEval-2023 Task 1: Effective Zero-Shot Visual Word Sense Disambiguation Approaches using External Knowledge Sources,Florian Schneider,"The objective of the SemEval-2023 Task 1: Visual Word Sense Disambiguation (VWSD) is to identify the image illustrating the indented meaning of a target word and some minimal additional context. The omnipresence of textual and visual data in the task strongly suggests the utilization of the recent advances in multi-modal machine learning, i.e., pretrained visiolinguistic models (VLMs). Often referred to as foundation models due to their strong performance on many vision-language downstream tasks, these models further demonstrate powerful zero-shot capabilities. In this work, we utilize various pertained VLMs in a zero-shot fashion for multiple approaches using external knowledge sources to enrich the contextual information. Further, we evaluate our methods on the final test data and extensively analyze the suitability of different knowledge sources, the influence of training data, model sizes, multi-linguality, and different textual prompting strategies. Although we are not among the best-performing systems (rank 20 of 56), our experiments described in this work prove competitive results. Moreover, we aim to contribute meaningful insights and propel multi-modal machine learning tasks like VWSD.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes various pertained VLMs in a zero-shot fashion for multiple approaches using external knowledge sources to enrich the contextual information and aims to contribute meaningful insights and propel multi-modal machine learning tasks like VWSD.'}",https://aclanthology.org/2023.semeval-1.64.pdf
-Instruction Tuning with Lexicons for Zero-Shot Style Classification,Ruohao Guo,"Style is used to convey authors' intentions and attitudes. Despite the success of large pre-trained language models on style classification, prior work relies on fine-tuning with labeled examples. Prompting large language models to classify style without fine-tuning is challenging because language styles can be difficult to define. In this study, we investigate the effectiveness of style lexicons as a means for instructing language models how to identify new styles that are unseen during training. Our experiments show that lexicon-based instructions improve transfer zero-shot performance significantly. We will release our code and data.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of style lexicons as a means for instructing language models how to identify new styles that are unseen during training are investigated and show that lexicon-based instructions improve transfer zero-shot performance significantly.'}",http://arxiv.org/pdf/2305.14592
-Zero-shot Faithfulness Evaluation for Text Summarization with Foundation Language Model,Qi Jia,"Despite tremendous improvements in natural language generation, summarization models still suffer from the unfaithfulness issue. Previous work evaluates faithfulness either using models trained on the other tasks or in-domain synthetic data, or prompting a large model such as ChatGPT. This paper proposes to do zero-shot faithfulness evaluation simply with a moderately-sized foundation language model. We introduce a new metric FFLM, which is a combination of probability changes based on the intuition that prefixing a piece of text that is consistent with the output will increase the probability of predicting the output. Experiments show that FFLM performs competitively with or even outperforms ChatGPT on both inconsistency detection and faithfulness rating with 24x fewer parameters. FFLM also achieves improvements over other strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A new metric FFLM is introduced, which is a combination of probability changes based on the intuition that prefixing a piece of text that is consistent with the output will increase the probability of predicting the output.'}",
-The Art of SOCRATIC QUESTIONING: Zero-shot Multimodal Reasoning with Recursive Thinking and Self-Questioning,Jingyuan Qi,"Chain-of-Thought prompting (CoT) enables large-scale language models to solve complex reasoning problems by decomposing the problem and tackling it step-by-step. However, Chain-of-Thought is a greedy thinking process that requires the language model to come up with a starting point and generate the next step solely based on previous steps. This thinking process is different from how humans approach a complex problem e.g., we proactively raise sub-problems related to the original problem and recursively answer them. In this work, we propose S OCRATIC Q UESTIONING , a divide-and-conquer fashion algorithm that simulates the self-questioning and recursive thinking process. S OCRATIC Q UESTIONING is driven by a S ELF -Q UESTIONING module that employs a large-scale language model to propose sub-problems related to the original problem as intermediate steps and S OCRATIC Q UESTION - ING recursively backtracks and answers the sub-problems until reaches the original problem. We apply our proposed algorithm to the visual question-answering task as a case study and by evaluating it on three public benchmark datasets, we observe a significant performance improvement over all baselines on (almost) all datasets. In addition, the qualitative analysis clearly demonstrates the intermediate thinking steps elicited by S OCRATIC Q UESTIONING are similar to the human’s recursively thinking process of a complex reasoning problem.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative analysis clearly demonstrates the intermediate thinking steps elicited by S OCRATIC Q UESTIONING are similar to the human’s recursively thinking process of a complex reasoning problem.'}",http://arxiv.org/pdf/2305.14999
-ReGeneration Learning of Diffusion Models with Rich Prompts for Zero-Shot Image Translation,Yupei Lin,"Large-scale text-to-image models have demonstrated amazing ability to synthesize diverse and high-fidelity images. However, these models are often violated by several limitations. Firstly, they require the user to provide precise and contextually relevant descriptions for the desired image modifications. Secondly, current models can impose significant changes to the original image content during the editing process. In this paper, we explore ReGeneration learning in an image-to-image Diffusion model (ReDiffuser), that preserves the content of the original image without human prompting and the requisite editing direction is automatically discovered within the text embedding space. To ensure consistent preservation of the shape during image editing, we propose cross-attention guidance based on regeneration learning. This novel approach allows for enhanced expression of the target domain features while preserving the original shape of the image. In addition, we introduce a cooperative update strategy, which allows for efficient preservation of the original shape of an image, thereby improving the quality and consistency of shape preservation throughout the editing process. Our proposed method leverages an existing pre-trained text-image diffusion model without any additional training. Extensive experiments show that the proposed method outperforms existing work in both real and synthetic image editing.","{'model': 'tldr@v2.0.0', 'text': 'ReGeneration learning in an image-to-image Diffusion model (ReDiffuser), that preserves the content of the original image without human prompting and the requisite editing direction is automatically discovered within the text embedding space, is explored.'}",http://arxiv.org/pdf/2305.04651
-Zero-Shot Refinement of Buildings' Segmentation Models using SAM,Ali Mayladan,"Foundation models have excelled in various tasks but are often evaluated on general benchmarks. The adaptation of these models for specific domains, such as remote sensing imagery, remains an underexplored area. In remote sensing, precise building instance segmentation is vital for applications like urban planning. While Convolutional Neural Networks (CNNs) perform well, their generalization can be limited. For this aim, we present a novel approach to adapt foundation models to address existing models' generalization dropback. Among several models, our focus centers on the Segment Anything Model (SAM), a potent foundation model renowned for its prowess in class-agnostic image segmentation capabilities. We start by identifying the limitations of SAM, revealing its suboptimal performance when applied to remote sensing imagery. Moreover, SAM does not offer recognition abilities and thus fails to classify and tag localized objects. To address these limitations, we introduce different prompting strategies, including integrating a pre-trained CNN as a prompt generator. This novel approach augments SAM with recognition abilities, a first of its kind. We evaluated our method on three remote sensing datasets, including the WHU Buildings dataset, the Massachusetts Buildings dataset, and the AICrowd Mapping Challenge. For out-of-distribution performance on the WHU dataset, we achieve a 5.47% increase in IoU and a 4.81% improvement in F1-score. For in-distribution performance on the WHU dataset, we observe a 2.72% and 1.58% increase in True-Positive-IoU and True-Positive-F1 score, respectively. We intend to release our code repository, hoping to inspire further exploration of foundation models for domain-specific tasks within the remote sensing community.","{'model': 'tldr@v2.0.0', 'text': ""A novel approach to adapt foundation models to address existing models' generalization dropback, focusing on the Segment Anything Model (SAM), a potent foundation model renowned for its prowess in class-agnostic image segmentation capabilities.""}",https://arxiv.org/pdf/2310.01845
-MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action,Zhengyuan Yang,"We propose MM-REACT, a system paradigm that integrates ChatGPT with a pool of vision experts to achieve multimodal reasoning and action. In this paper, we define and explore a comprehensive list of advanced vision tasks that are intriguing to solve, but may exceed the capabilities of existing vision and vision-language models. To achieve such advanced visual intelligence, MM-REACT introduces a textual prompt design that can represent text descriptions, textualized spatial coordinates, and aligned file names for dense visual signals such as images and videos. MM-REACT's prompt design allows language models to accept, associate, and process multimodal information, thereby facilitating the synergetic combination of ChatGPT and various vision experts. Zero-shot experiments demonstrate MM-REACT's effectiveness in addressing the specified capabilities of interests and its wide application in different scenarios that require advanced visual understanding. Furthermore, we discuss and compare MM-REACT's system paradigm with an alternative approach that extends language models for multimodal scenarios through joint finetuning. Code, demo, video, and visualization are available at https://multimodal-react.github.io/","{'model': 'tldr@v2.0.0', 'text': 'This paper defines and explores a comprehensive list of advanced vision tasks that are intriguing to solve, but may exceed the capabilities of existing vision and vision-language models and proposes MM-REACT, a system paradigm that integrates ChatGPT with a pool of vision experts to achieve multimodal reasoning and action.'}",http://arxiv.org/pdf/2303.11381
-Can an Embodied Agent Find Your “Cat-shaped Mug”? LLM-Guided Exploration for Zero-Shot Object Navigation,Vishnu Sashank Dorbala,"—We present LGX (Language-guided Exploration), a novel algorithm for Language-Driven Zero-Shot Object Goal Navigation (L-ZSON), where an embodied agent navigates to an uniquely described target object in a previously unseen environment. Our approach makes use of Large Language Models (LLMs) for this task by leveraging the LLM’s commonsense-reasoning capabilities for making sequential navigational decisions. Simultaneously, we perform generalized target object detection using a pre-trained Vision-Language grounding model. We achieve state-of-the-art zero-shot object navigation results on RoboTHOR with a success rate (SR) improvement of over 27% over the current baseline of the OWL-ViT CLIP on Wheels (OWL CoW). Furthermore, we study the usage of LLMs for robot navigation and present an analysis of various prompting strategies affecting the model output. Finally, we showcase the benefits of our approach via real-world experiments that indicate the superior performance of LGX in detecting and navigating to visually unique objects.","{'model': 'tldr@v2.0.0', 'text': None}",
-The Art of Prompting: Event Detection based on Type Specific Prompts,Sijia Wang,"We compare various forms of prompts to represent event types and develop a unified framework to incorporate the event type specific prompts for supervised, few-shot, and zero-shot event detection. The experimental results demonstrate that a well-defined and comprehensive event type prompt can significantly improve event detection performance, especially when the annotated data is scarce (few-shot event detection) or not available (zero-shot event detection). By leveraging the semantics of event types, our unified framework shows up to 22.2% F-score gain over the previous state-of-the-art baselines.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results demonstrate that a well-defined and comprehensive event type prompt can significantly improve event detection performance, especially when the annotated data is scarce (few- shot event detection) or not available (zero-shot event detection).'}",http://arxiv.org/pdf/2204.07241
-Prompting Language Models for Linguistic Structure,Terra Blevins,"Although pretrained language models (PLMs) can be prompted to perform a wide range of language tasks, it remains an open question how much this ability comes from generalizable linguistic understanding versus surface-level lexical patterns. To test this, we present a structured prompting approach for linguistic structured prediction tasks, allowing us to perform zero- and few-shot sequence tagging with autoregressive PLMs. We evaluate this approach on part-of-speech tagging, named entity recognition, and sentence chunking, demonstrating strong few-shot performance in all cases. We also find that while PLMs contain significant prior knowledge of task labels due to task leakage into the pretraining corpus, structured prompting can also retrieve linguistic structure with arbitrary labels. These findings indicate that the in-context learning ability and linguistic knowledge of PLMs generalizes beyond memorization of their training data.","{'model': 'tldr@v2.0.0', 'text': 'It is found that while PLMs contain significant prior knowledge of task labels due to task leakage into the pretraining corpus, structured prompting can also retrieve linguistic structure with arbitrary labels, indicating that the in-context learning ability and linguistic knowledge of PLMs generalizes beyond memorization of their training data.'}",http://arxiv.org/pdf/2211.07830
-TEMPERA: Test-Time Prompting via Reinforcement Learning,Tianjun Zhang,"Careful prompt design is critical to the use of large language models in zero-shot or few-shot learning. As a consequence, there is a growing interest in automated methods to design optimal prompts. In this work, we propose Test-time Prompt Editing using Reinforcement learning (TEMPERA). In contrast to prior prompt generation methods, TEMPERA can efficiently leverage prior knowledge, is adaptive to different queries and provides an interpretable prompt for every query. To achieve this, we design a novel action space that allows flexible editing of the initial prompts covering a wide set of commonly-used components like instructions, few-shot exemplars, and verbalizers. The proposed method achieves significant gains compared with recent SoTA approaches like prompt tuning, AutoPrompt, and RLPrompt, across a variety of tasks including sentiment analysis, topic classification, natural language inference, and reading comprehension. Our method achieves 5.33x on average improvement in sample efficiency when compared to the traditional fine-tuning methods.","{'model': 'tldr@v2.0.0', 'text': 'This work designs a novel action space that allows flexible editing of the initial prompts covering a wide set of commonly-used components like instructions, few-shot exemplars, and verbalizers, and achieves significant gains compared with recent SoTA approaches like prompt tuning, AutoPrompt, and RLPrompt.'}",http://arxiv.org/pdf/2211.11890
-Is EVALITA Done? On the Impact of Prompting on the Italian NLP Evaluation Campaign,Valerio Basile,"Prompt-based learning is a recent paradigm in NLP that leverages large pre-trained language models to perform a variety of tasks. With this technique, it is possible to build classifiers that do not need training data (zero-shot). In this paper, we assess the status of prompt-based learning applied to several text classification tasks in the Italian language. The results indicate that the performance gap towards current supervised methods is still relevant. However, the difference in performance between pre-trained models and the characteristic of the prompt-based classifier of operating in a zero-shot fashion open a discussion regarding the next generation of evaluation campaigns for NLP.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the performance gap towards current supervised methods is still relevant, however, the difference in performance between pre-trained models and the characteristic of the prompt-based classifier of operating in a zero-shot fashion open a discussion regarding the next generation of evaluation campaigns for NLP.'}",
-CLIP also Understands Text: Prompting CLIP for Phrase Understanding,An Yan,"Contrastive Language-Image Pretraining (CLIP) efficiently learns visual concepts by pre-training with natural language supervision. CLIP and its visual encoder have been explored on various vision and language tasks and achieve strong zero-shot or transfer learning performance. However, the application of its text encoder solely for text understanding has been less explored. In this paper, we find that the text encoder of CLIP actually demonstrates strong ability for phrase understanding, and can even significantly outperform popular language models such as BERT with a properly designed prompt. Extensive experiments validate the effectiveness of our method across different datasets and domains on entity clustering and entity set expansion tasks.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the text encoder of CLIP actually demonstrates strong ability for phrase understanding, and can even significantly outperform popular language models such as BERT with a properly designed prompt.'}",http://arxiv.org/pdf/2210.05836
-Odor Descriptor Understanding through Prompting,Laura Sisson,"Embeddings from contemporary natural language processing (NLP) models are commonly used as numerical representations for words or sentences. However, odor descriptor words, like""leather""or""fruity"", vary significantly between their commonplace usage and their olfactory usage, as a result traditional methods for generating these embeddings do not suffice. In this paper, we present two methods to generate embeddings for odor words that are more closely aligned with their olfactory meanings when compared to off-the-shelf embeddings. These generated embeddings outperform the previous state-of-the-art and contemporary fine-tuning/prompting methods on a pre-existing zero-shot odor-specific NLP benchmark.","{'model': 'tldr@v2.0.0', 'text': 'These generated embeddings outperform the previous state-of-the-art and contemporary fine-tuning/prompting methods on a pre-existing zero-shot odor-specific NLP benchmark.'}",http://arxiv.org/pdf/2205.03719
-ArgGen: Prompting Text Generation Models for Document-Level Event-Argument Aggregation,Debanjana Kar,"Most of the existing discourse-level Information Extraction tasks have been modeled to be extractive in nature. However, we argue that extracting information from larger bodies of discourse-like documents requires more natural language understanding and reasoning capabilities. In our work, we propose the novel task of document-level event argument aggregation which generates consolidated event-arguments at a document-level with minimal loss of information. More specifically, we focus on generating precise document-level information frames in a multilingual setting using prompt-based methods. In this paper, we show the effectiveness of prompt-based text generation approach to generate document-level argument spans in a low-resource and zero-shot setting. We also release the first of its kind multilingual event argument aggregation dataset that can be leveraged in other related multilingual text generation tasks as well: https://github.com/","{'model': 'tldr@v2.0.0', 'text': 'This paper shows the effectiveness of prompt-based text generation approach to generate document-level argument spans in a low-resource and zero-shot setting and releases the first of its kind multilingual event argument aggregation dataset.'}",
-On the Evaluations of ChatGPT and Emotion-enhanced Prompting for Mental Health Analysis,Kailai Yang,"Automated mental health analysis shows great potential for enhancing the efficiency and ac-cessibility of mental health care, whereas the recent dominant methods utilized pre-trained language models (PLMs) as the backbone and incorporated emotional information. The latest large language models (LLMs), such as ChatGPT, exhibit dramatic capabilities on di-verse natural language processing tasks. However, existing studies on ChatGPT’s zero-shot performance for mental health analysis have limitations in inadequate evaluation, utiliza-tion of emotional information, and explainability of methods. In this work, we comprehen-sively evaluate the mental health analysis and emotional reasoning ability of ChatGPT on 11 datasets across 5 tasks, including binary and multi-class mental health condition detection, cause/factor detection of mental health conditions, emotion recognition in conversations, and causal emotion entailment. We empirically analyze the impact of different prompting strategies with emotional cues on ChatGPT’s mental health analysis ability and explainability. Experimental results show that ChatGPT outperforms traditional neural network methods but still has a significant gap with advanced task-specific methods. The qualitative analysis shows its potential in explainability compared with advanced black-box methods but also limitations on robustness and inaccurate reasoning. Prompt engineering with emotional cues is found to be effective in improving its performance on mental health analysis but requires the proper way of emotion infusion.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that ChatGPT outperforms traditional neural network methods but still has a significant gap with advanced task-species methods, and the qualitative analysis shows its potential in explainability compared with advanced black-box methods but also limitations on robustness and inaccurate reasoning.'}",https://arxiv.org/pdf/2304.03347
-Reasoning Implicit Sentiment with Chain-of-Thought Prompting,Hao Fei,"While sentiment analysis systems try to determine the sentiment polarities of given targets based on the key opinion expressions in input texts, in implicit sentiment analysis (ISA) the opinion cues come in an implicit and obscure manner. Thus detecting implicit sentiment requires the common-sense and multi-hop reasoning ability to infer the latent intent of opinion. Inspired by the recent chain-of-thought (CoT) idea, in this work we introduce a Three-hop Reasoning (THOR) CoT framework to mimic the human-like reasoning process for ISA. We design a three-step prompting principle for THOR to step-by-step induce the implicit aspect, opinion, and finally the sentiment polarity. Our THOR+Flan-T5 (11B) pushes the state-of-the-art (SoTA) by over 6% F1 on supervised setup. More strikingly, THOR+GPT3 (175B) boosts the SoTA by over 50% F1 on zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'This work designs a three-step prompting principle for THOR to step-by-step induce the implicit aspect, opinion, and finally the sentiment polarity, and introduces a Three-hop Reasoning (THOR) CoT framework to mimic the human-like reasoning process for ISA.'}",http://arxiv.org/pdf/2305.11255
-Vita-CLIP: Video and text adaptive CLIP via Multimodal Prompting,Syed Talal Wasim,"Adopting contrastive image-text pretrained models like CLIP towards video classification has gained attention due to its cost-effectiveness and competitive performance. However, recent works in this area face a trade-off. Finetuning the pretrained model to achieve strong supervised performance results in low zero-shot generalization. Similarly, freezing the backbone to retain zero-shot capability causes significant drop in supervised accuracy. Because of this, recent works in literature typically train separate models for supervised and zero-shot action recognition. In this work, we propose a multimodal prompt learning scheme that works to balance the supervised and zero-shot performance under a single unified training. Our prompting approach on the vision side caters for three aspects: 1) Global video-level prompts to model the data distribution; 2) Local frame-level prompts to provide per-frame discriminative conditioning; and 3) a summary prompt to extract a condensed video representation. Additionally, we define a prompting scheme on the text side to augment the textual context. Through this prompting scheme, we can achieve state-of-the-art zero-shot performance on Kinetics-600, HMDB51 and UCF101 while remaining competitive in the supervised setting. By keeping the pretrained backbone frozen, we optimize a much lower number of parameters and retain the existing general representation which helps achieve the strong zero-shot performance. Our codes/models will be released at https://github.com/TalalWasim/Vita-Clip..","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a multimodal prompt learning scheme that works to balance the supervised and zero- shot performance under a single unified training and can achieve state-of-the-art zero-shot performance on Kinetics-600, HMDB51 and UCF101 while remaining competitive in the supervised setting.'}",https://arxiv.org/pdf/2304.03307
-PEARL: Prompting Large Language Models to Plan and Execute Actions Over Long Documents,Simeng Sun,"Strategies such as chain-of-thought prompting improve the performance of large language models (LLMs) on complex reasoning tasks by decomposing input examples into intermediate steps. However, it remains unclear how to apply such methods to reason over long input documents, in which both the decomposition and the output of each intermediate step are non-trivial to obtain. In this work, we propose PEARL, a prompting framework to improve reasoning over long documents, which consists of three stages: action mining, plan formulation, and plan execution. More specifically, given a question about a long document, PEARL decomposes the question into a sequence of actions (e.g., SUMMARIZE, FIND_EVENT, FIND_RELATION) and then executes them over the document to obtain the answer. Each stage of PEARL is implemented via zero-shot or few-shot prompting of LLMs (in our work, GPT-4) with minimal human input. We evaluate PEARL on a challenging subset of the QuALITY dataset, which contains questions that require complex reasoning over long narrative texts. PEARL outperforms zero-shot and chain-of-thought prompting on this dataset, and ablation experiments show that each stage of PEARL is critical to its performance. Overall, PEARL is a first step towards leveraging LLMs to reason over long documents.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PEARL, a prompting framework to improve reasoning over long documents, which consists of three stages: action mining, plan formulation, and plan execution, which outperforms zero-shot and chain-of-thought prompting on a challenging subset of the QuALITY dataset.'}",http://arxiv.org/pdf/2305.14564
-Multimodal Procedural Planning via Dual Text-Image Prompting,Yujie Lu,"Embodied agents have achieved prominent performance in following human instructions to complete tasks. However, the potential of providing instructions informed by texts and images to assist humans in completing tasks remains underexplored. To uncover this capability, we present the multimodal procedural planning (MPP) task, in which models are given a high-level goal and generate plans of paired text-image steps, providing more complementary and informative guidance than unimodal plans. The key challenges of MPP are to ensure the informativeness, temporal coherence,and accuracy of plans across modalities. To tackle this, we propose Text-Image Prompting (TIP), a dual-modality prompting method that jointly leverages zero-shot reasoning ability in large language models (LLMs) and compelling text-to-image generation ability from diffusion-based models. TIP improves the interaction in the dual modalities using Text-to-Image Bridge and Image-to-Text Bridge, allowing LLMs to guide the textual-grounded image plan generation and leveraging the descriptions of image plans to ground the textual plan reversely. To address the lack of relevant datasets, we collect WIKIPLAN and RECIPEPLAN as a testbed for MPP. Our results show compelling human preferences and automatic scores against unimodal and multimodal baselines on WIKIPLAN and RECIPEPLAN in terms of informativeness, temporal coherence, and plan accuracy. Our code and data: https://github.com/YujieLu10/MPP.","{'model': 'tldr@v2.0.0', 'text': 'Text-Image Prompting (TIP) is proposed, a dual-modality prompting method that jointly leverages zero-shot reasoning ability in large language models (LLMs) and compelling text-to-image generation ability from diffusion-based models to tackle the key challenges of MPP.'}",http://arxiv.org/pdf/2305.01795
-Federated Prompting and Chain-of-Thought Reasoning for Improving LLMs Answering,Xiangyang Liu,"We investigate how to enhance answer precision in frequently asked questions posed by distributed users using cloud-based Large Language Models (LLMs). Our study focuses on a typical situations where users ask similar queries that involve identical mathematical reasoning steps and problem-solving procedures. Due to the unsatisfactory accuracy of LLMs' zero-shot prompting with standalone questions, we propose to improve the distributed synonymous questions using Self-Consistency (SC) and Chain-of-Thought (CoT) techniques. Specifically, we first retrieve synonymous questions from a crowd-sourced database and create a federated question pool. We call these federated synonymous questions with the same or different parameters SP-questions or DP-questions, respectively. We refer to our methods as Fed-SP-SC and Fed-DP-CoT, which can generate significantly more accurate answers for all user queries without requiring sophisticated model-tuning. Through extensive experiments, we demonstrate that our proposed methods can significantly enhance question accuracy by fully exploring the synonymous nature of the questions and the consistency of the answers.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes to improve the distributed synonymous questions using Self-Consistency (SC) and Chain-of-Thought (CoT) techniques, which can generate significantly more accurate answers for all user queries without requiring sophisticated model-tuning.'}",http://arxiv.org/pdf/2304.13911
-Code Prompting: a Neural Symbolic Method for Complex Reasoning in Large Language Models,Y. Hu,"Large language models (LLMs) have scaled up to unlock a wide range of complex reasoning tasks with the aid of various prompting methods. However, current prompting methods generate natural language intermediate steps to help reasoning, which can cause imperfect task reduction and confusion. To mitigate such limitations, we explore code prompting, a neural symbolic prompting method with both zero-shot and few-shot versions which triggers code as intermediate steps. We conduct experiments on 7 widely-used benchmarks involving symbolic reasoning and arithmetic reasoning. Code prompting generally outperforms chain-of-thought (CoT) prompting. To further understand the performance and limitations of code prompting, we perform extensive ablation studies and error analyses, and identify several exclusive advantages of using symbolic promptings compared to natural language. We also consider the ensemble of code prompting and CoT prompting to combine the strengths of both. Finally, we show through experiments how code annotations and their locations affect code prompting.","{'model': 'tldr@v2.0.0', 'text': 'This work explores code prompting, a neural symbolic prompting method with both zero-shot and few-shot versions which triggers code as intermediate steps, and performs extensive ablation studies and error analyses and identifies several exclusive advantages of using symbolic promptings compared to natural language.'}",https://arxiv.org/pdf/2305.18507
-Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V,Jianwei Yang,"We present Set-of-Mark (SoM), a new visual prompting method, to unleash the visual grounding abilities of large multimodal models (LMMs), such as GPT-4V. As illustrated in Fig. 1 (right), we employ off-the-shelf interactive segmentation models, such as SAM, to partition an image into regions at different levels of granularity, and overlay these regions with a set of marks e.g., alphanumerics, masks, boxes. Using the marked image as input, GPT-4V can answer the questions that require visual grounding. We perform a comprehensive empirical study to validate the effectiveness of SoM on a wide range of fine-grained vision and multimodal tasks. For example, our experiments show that GPT-4V with SoM outperforms the state-of-the-art fully-finetuned referring segmentation model on RefCOCOg in a zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that GPT-4V with SoM outperforms the state-of-the-art fully-finetuned referring segmentation model on RefCOCOg in a zero-shot setting and the effectiveness of SoM on a wide range of fine-grained vision and multimodal tasks is validated.'}",
-Legal Syllogism Prompting: Teaching Large Language Models for Legal Judgment Prediction,Cong Jiang,"Legal syllogism is a form of deductive reasoning commonly used by legal professionals to analyze cases. In this paper, we propose legal syllogism prompting (LoT), a simple prompting method to teach large language models (LLMs) for legal judgment prediction. LoT teaches only that in the legal syllogism the major premise is law, the minor premise is the fact, and the conclusion is judgment. Then the models can produce a syllogism reasoning of the case and give the judgment without any learning, fine-tuning, or examples. On CAIL2018, a Chinese criminal case dataset, we performed zero-shot judgment prediction experiments with GPT-3 models. Our results show that LLMs with LoT achieve better performance than the baseline and chain of thought prompting, the state-of-art prompting method on diverse reasoning tasks. LoT enables the model to concentrate on the key information relevant to the judgment and to correctly understand the legal meaning of acts, as compared to other methods. Our method enables LLMs to predict judgment along with law articles and justification, which significantly enhances the explainability of models.",,
-RAMP: Retrieval and Attribute-Marking Enhanced Prompting for Attribute-Controlled Translation,Gabriele Sarti,"Attribute-controlled translation (ACT) is a subtask of machine translation that involves controlling stylistic or linguistic attributes (like formality and gender) of translation outputs. While ACT has garnered attention in recent years due to its usefulness in real-world applications, progress in the task is currently limited by dataset availability, since most prior approaches rely on supervised methods. To address this limitation, we propose Retrieval and Attribute-Marking enhanced Prompting (RAMP), which leverages large multilingual language models to perform ACT in few-shot and zero-shot settings. RAMP improves generation accuracy over the standard prompting approach by (1) incorporating a semantic similarity retrieval component for selecting similar in-context examples, and (2) marking in-context examples with attribute annotations. Our comprehensive experiments show that RAMP is a viable approach in both zero-shot and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Retrieval and Attribute-Marking enhanced Prompting (RAMP) is proposed, which leverages large multilingual language models to performAttribute-controlled translation in few-shot and zero-shot settings and improves generation accuracy over the standard prompting approach.'}",
-PIEClass: Weakly-Supervised Text Classification with Prompting and Noise-Robust Iterative Ensemble Training,Yunyi Zhang,"Weakly-supervised text classification trains a classifier using the label name of each target class as the only supervision, which largely reduces human annotation efforts. Most existing methods first use the label names as static keyword-based features to generate pseudo labels, which are then used for final classifier training. While reasonable, such a commonly adopted framework suffers from two limitations: (1) keywords can have different meanings in different contexts and some text may not have any keyword, so keyword matching can induce noisy and inadequate pseudo labels; (2) the errors made in the pseudo label generation stage will directly propagate to the classifier training stage without a chance of being corrected. In this paper, we propose a new method, PIEClass, consisting of two modules: (1) a pseudo label acquisition module that uses zero-shot prompting of pre-trained language models (PLM) to get pseudo labels based on contextualized text understanding beyond static keyword matching, and (2) a noise-robust iterative ensemble training module that iteratively trains classifiers and updates pseudo labels by utilizing two PLM fine-tuning methods that regularize each other. Extensive experiments show that PIEClass achieves overall better performance than existing strong baselines on seven benchmark datasets and even achieves similar performance to fully-supervised classifiers on sentiment classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A new method, PIEClass, consisting of a pseudo label acquisition module that uses zero-shot prompting of pre-trained language models (PLM) to get pseudo labels based on contextualized text understanding beyond static keyword matching, and a noise-robust iterative ensemble training module that iteratively trains classifiers and updates pseudo labels by utilizing two PLM fine-tuning methods that regularize each other.'}",
-Self-Explanation Prompting Improves Dialogue Understanding in Large Language Models,Haoyu Gao,"Task-oriented dialogue (TOD) systems facilitate users in executing various activities via multi-turn dialogues, but Large Language Models (LLMs) often struggle to comprehend these intricate contexts. In this study, we propose a novel""Self-Explanation""prompting strategy to enhance the comprehension abilities of LLMs in multi-turn dialogues. This task-agnostic approach requires the model to analyze each dialogue utterance before task execution, thereby improving performance across various dialogue-centric tasks. Experimental results from six benchmark datasets confirm that our method consistently outperforms other zero-shot prompts and matches or exceeds the efficacy of few-shot prompts, demonstrating its potential as a powerful tool in enhancing LLMs' comprehension in complex dialogue tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel Self-Explanation Prompting strategy to enhance the comprehension abilities of LLMs in multi-turn dialogues, which requires the model to analyze each dialogue utterance before task execution, thereby improving performance across various dialogue-centric tasks.'}",https://arxiv.org/pdf/2309.12940
-Fixed Input Parameterization for Efficient Prompting,Eunbi Choi,"Recent works have shown that attaching prompts to the input is effective at conditioning Language Models (LM) to perform specific tasks. However, prompts are always included in the input text during inference, even when they are fixed, thus incurring substantial computational and memory overhead. Also, there is currently no straightforward method of utilizing prompts that are longer than the maximum input length of the LMs without incurring additional costs during inference. We formally define Fixed Input Parameterization (FIP) problem that focuses on injecting the fixed prompt into the parameters of an LM to be an efficient alternative to attaching fixed prompts to the input. We show that in scenarios with long fixed prompts, FIP can be up to 280 times more efficient in terms of total FLOPs than previous approaches. We further explore methodologies for FIP and show promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions. Through these explorations, we show that FIP can be a promising direction for conditioning language models, in scenarios with long and fixed prompts 1 .","{'model': 'tldr@v2.0.0', 'text': 'It is shown that in scenarios with long fixed prompts, FIP can be up to 280 times more efficient in terms of total FLOPs than previous approaches and further explore methodologies for FIP and shows promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions.'}",https://aclanthology.org/2023.findings-acl.533.pdf
-MAP: Low-data Regime Multimodal Learning with Adapter-based Pre-training and Prompting,Wenyan Li,"Pretrained vision-language (VL) models have shown impressive results on various multi-modal downstream tasks recently. Many of the benchmark models build on pretrained causal language models (LMs), leveraging the original few-shot learning and generalization capability of the LMs trained with large text corpora. However, these models are often gigantic and require large-scale image and text data with high computational cost to train. This paper introduces a moderate-size model called MAP for efficient VL transfer learning through adapter-based pretraining and prompting. We aim to answer the question of how much we can complete through VL pretraining within the low-data regime while maximizing efficiency in transferring knowledge of a moderate-size frozen LM. Our experiments demonstrate that MAP achieves substantially better zero-shot and few-shot performance on downstream VL tasks with only 10% the size of pretraining data and a 30x lighter pretrained LM backbone compared to Frozen. MAP also outperforms fully trained models of comparable size at retaining its transfer learning ability when the amount of training data reduces.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a moderate-size model called MAP for efficient VL transfer learning through adapter-based pretraining and prompting and achieves substantially better zero-shot and few-shot performance on downstream VL tasks with only 10% the size of pretraining data.'}",
-P5: Plug-and-Play Persona Prompting for Personalized Response Selection,Joosung Lee,"The use of persona-grounded retrieval-based chatbots is crucial for personalized conversations, but there are several challenges that need to be addressed. 1) In general, collecting persona-grounded corpus is very expensive. 2) The chatbot system does not always respond in consideration of persona at real applications. To address these challenges, we propose a plug-and-play persona prompting method. Our system can function as a standard open-domain chatbot if persona information is not available. We demonstrate that this approach performs well in the zero-shot setting, which reduces the dependence on persona-ground training data. This makes it easier to expand the system to other languages without the need to build a persona-grounded corpus. Additionally, our model can be fine-tuned for even better performance. In our experiments, the zero-shot model improved the standard model by 7.71 and 1.04 points in the original persona and revised persona, respectively. The fine-tuned model improved the previous state-of-the-art system by 1.95 and 3.39 points in the original persona and revised persona, respectively. To the best of our knowledge, this is the first attempt to solve the problem of personalized response selection using prompt sequences. Our code is available on github~\footnote{https://github.com/rungjoo/plug-and-play-prompt-persona}.","{'model': 'tldr@v2.0.0', 'text': 'This is the first attempt to solve the problem of personalized response selection using prompt sequences using a plug-and-play persona prompting method, and it is demonstrated that this approach performs well in the zero-shot setting, which reduces the dependence on persona-ground training data.'}",https://arxiv.org/pdf/2310.06390
-CoF-CoT: Enhancing Large Language Models with Coarse-to-Fine Chain-of-Thought Prompting for Multi-domain NLU Tasks,Hoang Nguyen,"While Chain-of-Thought prompting is popular in reasoning tasks, its application to Large Language Models (LLMs) in Natural Language Understanding (NLU) is under-explored. Motivated by multi-step reasoning of LLMs, we propose Coarse-to-Fine Chain-of-Thought (CoF-CoT) approach that breaks down NLU tasks into multiple reasoning steps where LLMs can learn to acquire and leverage essential concepts to solve tasks from different granularities. Moreover, we propose leveraging semantic-based Abstract Meaning Representation (AMR) structured knowledge as an intermediate step to capture the nuances and diverse structures of utterances, and to understand connections between their varying levels of granularity. Our proposed approach is demonstrated effective in assisting the LLMs adapt to the multi-grained NLU tasks under both zero-shot and few-shot multi-domain settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Coarse-to-Fine Chain-of-Thought (CoF-CoT) approach that breaks down NLU tasks into multiple reasoning steps where LLMs can learn to acquire and leverage essential concepts to solve tasks from different granularities.'}",
-A Communication Theory Perspective on Prompting Engineering Methods for Large Language Models,Yuanfeng Song,"The springing up of Large Language Models (LLMs) has shifted the community from single-task-orientated natural language processing (NLP) research to a holistic end-to-end multi-task learning paradigm. Along this line of research endeavors in the area, LLM-based prompting methods have attracted much attention, partially due to the technological advantages brought by prompt engineering (PE) as well as the underlying NLP principles disclosed by various prompting methods. Traditional supervised learning usually requires training a model based on labeled data and then making predictions. In contrast, PE methods directly use the powerful capabilities of existing LLMs (i.e., GPT-3 and GPT-4) via composing appropriate prompts, especially under few-shot or zero-shot scenarios. Facing the abundance of studies related to the prompting and the ever-evolving nature of this field, this article aims to (i) illustrate a novel perspective to review existing PE methods, within the well-established communication theory framework; (ii) facilitate a better/deeper understanding of developing trends of existing PE methods used in four typical tasks; (iii) shed light on promising research directions for future PE methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel perspective is illustrated to review existingPE methods, within the well-established communication theory framework to facilitate a better/deeper understanding of developing trends of existing PE methods used in four typical tasks and shed light on promising research directions for future PE methods.'}",
-Winning Solution for the CVPR2023 Visual Anomaly and Novelty Detection Challenge: Multimodal Prompting for Data-centric Anomaly Detection,Yunkang Cao,"This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge. Going beyond uni-modal prompt, e.g., language prompt, we present a novel framework, i.e., Segment Any Anomaly + (SAA + ), for zero-shot anomaly segmentation with multi-modal prompts for the regularization of cascaded modern foundation models. Inspired by the great zero-shot generalization ability of foundation models like Segment Anything, we first explore their assembly (SAA) to leverage diverse multi-modal prior knowledge for anomaly localization. Subsequently, we further introduce multimodal prompts (SAA + ) derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation. The proposed SAA + model achieves state-of-the-art performance on several anomaly segmentation benchmarks, including VisA and MVTec-AD, in the zero-shot setting. We will release the code of our winning solution for the CVPR2023 VAND challenge at https:/","{'model': 'tldr@v2.0.0', 'text': 'This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge and introduces multimodal prompts derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation.'}",https://arxiv.org/pdf/2306.09067
-2nd Place Winning Solution for the CVPR2023 Visual Anomaly and Novelty Detection Challenge: Multimodal Prompting for Data-centric Anomaly Detection,Yunkang Cao,"This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge. Going beyond uni-modal prompt, e.g., language prompt, we present a novel framework, i.e., Segment Any Anomaly + (SAA$+$), for zero-shot anomaly segmentation with multi-modal prompts for the regularization of cascaded modern foundation models. Inspired by the great zero-shot generalization ability of foundation models like Segment Anything, we first explore their assembly (SAA) to leverage diverse multi-modal prior knowledge for anomaly localization. Subsequently, we further introduce multimodal prompts (SAA$+$) derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation. The proposed SAA$+$ model achieves state-of-the-art performance on several anomaly segmentation benchmarks, including VisA and MVTec-AD, in the zero-shot setting. We will release the code of our winning solution for the CVPR2023 VAN.","{'model': 'tldr@v2.0.0', 'text': 'This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge and introduces multimodal prompts derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation.'}",
-Prompting Segmentation with Sound is Generalizable Audio-Visual Source Localizer,Yaoting Wang,"Never having seen an object and heard its sound simultaneously, can the model still accurately localize its visual position from the input audio? In this work, we concentrate on the Audio-Visual Localization and Segmentation tasks but under the demanding zero-shot and few-shot scenarios. To achieve this goal, different from existing approaches that mostly employ the encoder-fusion-decoder paradigm to decode localization information from the fused audio-visual feature, we introduce the encoder-prompt-decoder paradigm, aiming to better fit the data scarcity and varying data distribution dilemmas with the help of abundant knowledge from pre-trained models. Specifically, we first propose to construct Semantic-aware Audio Prompt (SAP) to help the visual foundation model focus on sounding objects, meanwhile, the semantic gap between the visual and audio modalities is also encouraged to shrink. Then, we develop a Correlation Adapter (ColA) to keep minimal training efforts as well as maintain adequate knowledge of the visual foundation model. By equipping with these means, extensive experiments demonstrate that this new paradigm outperforms other fusion-based methods in both the unseen class and cross-dataset settings. We hope that our work can further promote the generalization study of Audio-Visual Localization and Segmentation in practical application scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the encoder-prompt-decoder paradigm, aiming to better fit the data scarcity and varying data distribution dilemmas with the help of abundant knowledge from pre-trained models, and proposes to construct Semantic-aware Audio Prompt to help the visual foundation model focus on sounding objects.'}",https://arxiv.org/pdf/2309.07929
-ASPIRO: Any-shot Structured Parsing-error-Induced ReprOmpting for Consistent Data-to-Text Generation,Martin Vejvar,"We present ASPIRO, an approach for structured data verbalisation into short template sentences in zero to few-shot settings. Unlike previous methods, our approach prompts large language models (LLMs) to directly produce entity-agnostic templates, rather than relying on LLMs to faithfully copy the given example entities, or validating/crafting the templates manually. We incorporate LLM re-prompting, triggered by algorithmic parsing checks, as well as the PARENT metric induced consistency validation to identify and rectify template generation problems in real-time. ASPIRO, compared to direct LLM output, averages 66\% parsing error rate reduction in generated verbalisations of RDF triples on the DART dataset. Our best 5-shot text-davinci-003 setup, scoring BLEU of 50.62, METEOR of 45.16, BLEURT of 0.82, NUBIA of 0.87, and PARENT of 0.8962 on the Rel2Text dataset, competes effectively with recent fine-tuned pre-trained language models.","{'model': 'tldr@v2.0.0', 'text': None}",
-Prompting Strategies for Citation Classification,Suchetha Nambanoor Kunnath,"Citation classification aims to identify the purpose of the cited article in the citing article. Previous citation classification methods rely largely on supervised approaches. The models are trained on datasets with citing sentences or citation contexts annotated for a citation's purpose or function or intent. Recent advancements in Large Language Models (LLMs) have dramatically improved the ability of NLP systems to achieve state-of-the-art performances under zero or few-shot settings. This makes LLMs particularly suitable for tasks where sufficiently large labelled datasets are not yet available, which remains to be the case for citation classification. This paper systematically investigates the effectiveness of different prompting strategies for citation classification and compares them to promptless strategies as a baseline. Specifically, we evaluate the following four strategies, two of which we introduce for the first time, which involve updating Language Model (LM) parameters while training the model: (1) Promptless fine-tuning, (2) Fixed-prompt LM tuning, (3) Dynamic Context-prompt LM tuning (proposed), (4) Prompt + LM fine-tuning (proposed). Additionally, we test the zero-shot performance of LLMs, GPT3.5, a (5) Tuning-free prompting strategy that involves no parameter updating. Our results show that prompting methods based on LM parameter updating significantly improve citation classification performances on both domain-specific and multi-disciplinary citation classifications. Moreover, our Dynamic Context-prompting method achieves top scores both for the ACL-ARC and ACT2 citation classification datasets, surpassing the highest-performing system in the 3C shared task benchmark. Interestingly, we observe zero-shot GPT3.5 to perform well on ACT2 but poorly on the ACL-ARC dataset.","{'model': 'tldr@v2.0.0', 'text': 'This paper systematically investigates the effectiveness of different prompting strategies for citation classification and compares them to promptless strategies as a baseline, showing that prompting methods based on LM parameter updating significantly improve citation classification performances on both domain-specific and multi-disciplinary citation classifications.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615018
-Expanding Language-Image Pretrained Models for General Video Recognition,Bolin Ni,"Contrastive language-image pretraining has shown great success in learning visual-textual joint representation from web-scale data, demonstrating remarkable""zero-shot""generalization ability for various image tasks. However, how to effectively expand such new language-image pretraining methods to video domains is still an open problem. In this work, we present a simple yet effective approach that adapts the pretrained language-image models to video recognition directly, instead of pretraining a new model from scratch. More concretely, to capture the long-range dependencies of frames along the temporal dimension, we propose a cross-frame attention mechanism that explicitly exchanges information across frames. Such module is lightweight and can be plugged into pretrained language-image models seamlessly. Moreover, we propose a video-specific prompting scheme, which leverages video content information for generating discriminative textual prompts. Extensive experiments demonstrate that our approach is effective and can be generalized to different video recognition scenarios. In particular, under fully-supervised settings, our approach achieves a top-1 accuracy of 87.1% on Kinectics-400, while using 12 times fewer FLOPs compared with Swin-L and ViViT-H. In zero-shot experiments, our approach surpasses the current state-of-the-art methods by +7.6% and +14.9% in terms of top-1 accuracy under two popular protocols. In few-shot scenarios, our approach outperforms previous best methods by +32.1% and +23.1% when the labeled data is extremely limited. Code and models are available at https://aka.ms/X-CLIP","{'model': 'tldr@v2.0.0', 'text': 'This work presents a simple yet effective approach that adapts the pretrained language-image models to video recognition directly, instead of pretraining a new model from scratch, and proposes a cross-frame attention mechanism that explicitly exchanges information across frames.'}",http://arxiv.org/pdf/2208.02816
-Can Large Language Models Transform Computational Social Science?,Caleb Ziems,"Large Language Models (LLMs) like ChatGPT are capable of successfully performing many language processing tasks zero-shot (without the need for training data). If this capacity also applies to the coding of social phenomena like persuasiveness and political ideology, then LLMs could effectively transform Computational Social Science (CSS). This work provides a road map for using LLMs as CSS tools. Towards this end, we contribute a set of prompting best practices and an extensive evaluation pipeline to measure the zero-shot performance of 13 language models on 24 representative CSS benchmarks. On taxonomic labeling tasks (classification), LLMs fail to outperform the best fine-tuned models but still achieve fair levels of agreement with humans. On free-form coding tasks (generation), LLMs produce explanations that often exceed the quality of crowdworkers' gold references. We conclude that today's LLMs can radically augment the CSS research pipeline in two ways: (1) serving as zero-shot data annotators on human annotation teams, and (2) bootstrapping challenging creative generation tasks (e.g., explaining the hidden meaning behind text). In summary, LLMs can significantly reduce costs and increase efficiency of social science analysis in partnership with humans.","{'model': 'tldr@v2.0.0', 'text': ""It is concluded that today's LLMs can radically augment the CSS research pipeline in two ways: serving as zero-shot data annotators on human annotation teams, and bootstrapping challenging creative generation tasks (e.g., explaining the hidden meaning behind text).""}",http://arxiv.org/pdf/2305.03514
-Large Language Models as General Pattern Machines,Suvir Mirchandani,"We observe that pre-trained large language models (LLMs) are capable of autoregressively completing complex token sequences -- from arbitrary ones procedurally generated by probabilistic context-free grammars (PCFG), to more rich spatial patterns found in the Abstraction and Reasoning Corpus (ARC), a general AI benchmark, prompted in the style of ASCII art. Surprisingly, pattern completion proficiency can be partially retained even when the sequences are expressed using tokens randomly sampled from the vocabulary. These results suggest that without any additional training, LLMs can serve as general sequence modelers, driven by in-context learning. In this work, we investigate how these zero-shot capabilities may be applied to problems in robotics -- from extrapolating sequences of numbers that represent states over time to complete simple motions, to least-to-most prompting of reward-conditioned trajectories that can discover and represent closed-loop policies (e.g., a stabilizing controller for CartPole). While difficult to deploy today for real systems due to latency, context size limitations, and compute costs, the approach of using LLMs to drive low-level control may provide an exciting glimpse into how the patterns among words could be transferred to actions.","{'model': 'tldr@v2.0.0', 'text': 'Surprisingly, pattern completion proficiency can be partially retained even when the sequences are expressed using tokens randomly sampled from the vocabulary, suggesting that without any additional training, LLMs can serve as general sequence modelers, driven by in-context learning.'}",https://arxiv.org/pdf/2307.04721
-Can GPT-3 Perform Statutory Reasoning?,Andrew Blair-Stanek,"Statutory reasoning is the task of reasoning with facts and statutes, which are rules written in natural language by a legislature. It is a basic legal skill. In this paper we explore the capabilities of the most capable GPT-3 model, text-davinci-003, on an established statutory-reasoning dataset called SARA. We consider a variety of approaches, including dynamic few-shot prompting, chain-of-thought prompting, and zero-shot prompting. While we achieve results with GPT-3 that are better than the previous best published results, we also identify several types of clear errors it makes. We investigate why these errors happen. We discover that GPT-3 has imperfect prior knowledge of the actual U.S. statutes on which SARA is based. More importantly, we create simple synthetic statutes, which GPT-3 is guaranteed not to have seen during training. We find GPT-3 performs poorly at answering straightforward questions about these simple synthetic statutes.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores the capabilities of the most capable GPT-3 model, text-davinci-003, on an established statutory-reasoning dataset called SARA, and discovers it has imperfect prior knowledge of the actual U.S. statutes on which SARA is based.'}",https://arxiv.org/pdf/2302.06100
-Segment Anything in High Quality,Lei Ke,"The recent Segment Anything Model (SAM) represents a big leap in scaling up segmentation models, allowing for powerful zero-shot capabilities and flexible prompting. Despite being trained with 1.1 billion masks, SAM's mask prediction quality falls short in many cases, particularly when dealing with objects that have intricate structures. We propose HQ-SAM, equipping SAM with the ability to accurately segment any object, while maintaining SAM's original promptable design, efficiency, and zero-shot generalizability. Our careful design reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation. We design a learnable High-Quality Output Token, which is injected into SAM's mask decoder and is responsible for predicting the high-quality mask. Instead of only applying it on mask-decoder features, we first fuse them with early and final ViT features for improved mask details. To train our introduced learnable parameters, we compose a dataset of 44K fine-grained masks from several sources. HQ-SAM is only trained on the introduced detaset of 44k masks, which takes only 4 hours on 8 GPUs. We show the efficacy of HQ-SAM in a suite of 9 diverse segmentation datasets across different downstream tasks, where 7 out of them are evaluated in a zero-shot transfer protocol. Our code and models will be released at https://github.com/SysCV/SAM-HQ.","{'model': 'tldr@v2.0.0', 'text': ""A learnable High-Quality Output Token is injected into SAM's mask decoder and is responsible for predicting the high-quality mask, which reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation.""}",http://arxiv.org/pdf/2306.01567
-Solving Challenging Math Word Problems Using GPT-4 Code Interpreter with Code-based Self-Verification,Aojun Zhou,"Recent progress in large language models (LLMs) like GPT-4 and PaLM-2 has brought significant advancements in addressing math reasoning problems. In particular, OpenAI's latest version of GPT-4, known as GPT-4 Code Interpreter, shows remarkable performance on challenging math datasets. In this paper, we explore the effect of code on enhancing LLMs' reasoning capability by introducing different constraints on the \textit{Code Usage Frequency} of GPT-4 Code Interpreter. We found that its success can be largely attributed to its powerful skills in generating and executing code, evaluating the output of code execution, and rectifying its solution when receiving unreasonable outputs. Based on this insight, we propose a novel and effective prompting method, explicit \uline{c}ode-based \uline{s}elf-\uline{v}erification~(CSV), to further boost the mathematical reasoning potential of GPT-4 Code Interpreter. This method employs a zero-shot prompt on GPT-4 Code Interpreter to encourage it to use code to self-verify its answers. In instances where the verification state registers as ``False'', the model shall automatically amend its solution, analogous to our approach of rectifying errors during a mathematics examination. Furthermore, we recognize that the states of the verification result indicate the confidence of a solution, which can improve the effectiveness of majority voting. With GPT-4 Code Interpreter and CSV, we achieve an impressive zero-shot accuracy on MATH dataset \textbf{(53.9\% $\to$ 84.3\%)}.","{'model': 'tldr@v2.0.0', 'text': ""The effect of code on enhancing LLMs' reasoning capability by introducing different constraints on the Code Usage Frequency of GPT-4 Code Interpreter is explored, and a novel and effective prompting method, explicit \\uline{c}ode-based \\ULine{s}elf-\\uline {v}erification~(CSV), is proposed to further boost the mathematical reasoning potential of GPN.""}",https://arxiv.org/pdf/2308.07921
-A fine-grained comparison of pragmatic language understanding in humans and language models,Jennifer Hu,"Pragmatics and non-literal language understanding are essential to human communication, and present a long-standing challenge for artificial language models. We perform a fine-grained comparison of language models and humans on seven pragmatic phenomena, using zero-shot prompting on an expert-curated set of English materials. We ask whether models (1) select pragmatic interpretations of speaker utterances, (2) make similar error patterns as humans, and (3) use similar linguistic cues as humans to solve the tasks. We find that the largest models achieve high accuracy and match human error patterns: within incorrect responses, models favor literal interpretations over heuristic-based distractors. We also find preliminary evidence that models and humans are sensitive to similar linguistic cues. Our results suggest that pragmatic behaviors can emerge in models without explicitly constructed representations of mental states. However, models tend to struggle with phenomena relying on social expectation violations.",,http://arxiv.org/pdf/2212.06801
-Multilingual Large Language Models Are Not (Yet) Code-Switchers,Ruochen Zhang,"Multilingual Large Language Models (LLMs) have recently shown great capabilities in a wide range of tasks, exhibiting state-of-the-art performance through zero-shot or few-shot prompting methods. While there have been extensive studies on their abilities in monolingual tasks, the investigation of their potential in the context of code-switching (CSW), the practice of alternating languages within an utterance, remains relatively uncharted. In this paper, we provide a comprehensive empirical analysis of various multilingual LLMs, benchmarking their performance across four tasks: sentiment analysis, machine translation, summarization and word-level language identification. Our results indicate that despite multilingual LLMs exhibiting promising outcomes in certain tasks using zero or few-shot prompting, they still underperform in comparison to fine-tuned models of much smaller scales. We argue that current""multilingualism""in LLMs does not inherently imply proficiency with code-switching texts, calling for future research to bridge this discrepancy.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that current""multilingualism"" in LLMs does not inherently imply proficiency with code-switching texts, calling for future research to bridge this discrepancy.'}",http://arxiv.org/pdf/2305.14235
-Knowledge-in-Context: Towards Knowledgeable Semi-Parametric Language Models,Xiaoman Pan,"Fully-parametric language models generally require a huge number of model parameters to store the necessary knowledge for solving multiple natural language tasks in zero/few-shot settings. In addition, it is hard to adapt to the evolving world knowledge without the costly model re-training. In this paper, we develop a novel semi-parametric language model architecture, Knowledge-in-Context (KiC), which empowers a parametric text-to-text language model with a knowledge-rich external memory. Specifically, the external memory contains six different types of knowledge: entity, dictionary, commonsense, event, script, and causality knowledge. For each input instance, the KiC model adaptively selects a knowledge type and retrieves the most helpful pieces of knowledge. The input instance along with its knowledge augmentation is fed into a text-to-text model (e.g., T5) to generate the output answer, where both the input and the output are in natural language forms after prompting. Interestingly, we find that KiC can be identified as a special mixture-of-experts (MoE) model, where the knowledge selector plays the role of a router that is used to determine the sequence-to-expert assignment in MoE. This key observation inspires us to develop a novel algorithm for training KiC with an instance-adaptive knowledge selector. As a knowledge-rich semi-parametric language model, KiC only needs a much smaller parametric part to achieve superior zero-shot performance on unseen tasks. By evaluating on 40+ different tasks, we show that KiC_Large with 770M parameters easily outperforms large language models (LMs) that are 4-39x larger by a large margin. We also demonstrate that KiC exhibits emergent abilities at a much smaller model scale compared to the fully-parametric models.","{'model': 'tldr@v2.0.0', 'text': 'A novel semi-parametric language model architecture, Knowledge-in-Context (KiC), which empowers a parametric text-to-text language model with a knowledge-rich external memory and finds that KiC can be identified as a special mixture-of-experts (MoE) model, where the knowledge selector plays the role of a router that is used to determine the sequence- to-expert assignment in MoE.'}",http://arxiv.org/pdf/2210.16433
-Learning to Decompose Visual Features with Latent Textual Prompts,Feng Wang,"Recent advances in pre-training vision-language models like CLIP have shown great potential in learning transferable visual representations. Nonetheless, for downstream inference, CLIP-like models suffer from either 1) degraded accuracy and robustness in the case of inaccurate text descriptions during retrieval-based inference (the challenge for zero-shot protocol); or 2) breaking the well-established vision-language alignment (the challenge for linear probing). To address them, we propose Decomposed Feature Prompting (DeFo). DeFo leverages a flexible number of learnable embeddings as textual input while maintaining the vision-language dual-model architecture, which enables the model to learn decomposed visual features with the help of feature-level textual prompts. We further use an additional linear layer to perform classification, allowing a scalable size of language inputs. Our empirical study shows DeFo's significance in improving the vision-language models. For example, DeFo obtains 73.2% test accuracy on ImageNet with a ResNet-50 backbone without tuning any pretrained weights of both the vision and language encoder, outperforming zero-shot CLIP by a large margin of 15.0%, and outperforming state-of-the-art vision-language prompt tuning method by 7.6%.","{'model': 'tldr@v2.0.0', 'text': 'Decomposed Feature Prompting (DeFo) leverages a flexible number of learnable embeddings as textual input while maintaining the vision-language dual-model architecture, which enables the model to learn decomposed visual features with the help of feature-level textual prompts.'}",http://arxiv.org/pdf/2210.04287
-XRICL: Cross-lingual Retrieval-Augmented In-Context Learning for Cross-lingual Text-to-SQL Semantic Parsing,Peng Shi,"In-context learning using large language models has recently shown surprising results for semantic parsing tasks such as Text-to-SQL translation. Prompting GPT-3 or Codex using several examples of question-SQL pairs can produce excellent results, comparable to state-of-the-art finetuning-based models. However, existing work primarily focuses on English datasets, and it is unknown whether large language models can serve as competitive semantic parsers for other languages. To bridge this gap, our work focuses on cross-lingual Text-to-SQL semantic parsing for translating non-English utterances into SQL queries based on an English schema. We consider a zero-shot transfer learning setting with the assumption that we do not have any labeled examples in the target language (but have annotated examples in English). This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts. We also include global translation exemplars for a target language to facilitate the translation process for large language models. To systematically evaluate our model, we construct two new benchmark datasets, XSpider and XKaggle-dbqa, which include questions in Chinese, Vietnamese, Farsi, and Hindi. Our experiments show that XRICL effectively leverages large pre-trained language models to outperform existing baselines. Data and code are publicly available at https://github.com/Impavidity/XRICL.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts and effectively leverages large pre-trained language models to outperform existing baselines.'}",http://arxiv.org/pdf/2210.13693
-Revisiting non-English Text Simplification: A Unified Multilingual Benchmark,Michael Joseph Ryan,"Recent advancements in high-quality, large-scale English resources have pushed the frontier of English Automatic Text Simplification (ATS) research. However, less work has been done on multilingual text simplification due to the lack of a diverse evaluation benchmark that covers complex-simple sentence pairs in many languages. This paper introduces the MultiSim benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs. This benchmark will encourage research in developing more effective multilingual text simplification models and evaluation metrics. Our experiments using MultiSim with pre-trained multilingual language models reveal exciting performance improvements from multilingual training in non-English settings. We observe strong performance from Russian in zero-shot cross-lingual transfer to low-resource languages. We further show that few-shot prompting with BLOOM-176b achieves comparable quality to reference simplifications outperforming fine-tuned models in most languages. We validate these findings through human evaluation.","{'model': 'tldr@v2.0.0', 'text': 'The MultiSim benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs, is introduced to encourage research in developing more effective multilingual text simplification models and evaluation metrics.'}",http://arxiv.org/pdf/2305.15678
-Multidimensional Evaluation for Text Style Transfer Using ChatGPT,Huiyuan Lai,"We investigate the potential of ChatGPT as a multidimensional evaluator for the task of \emph{Text Style Transfer}, alongside, and in comparison to, existing automatic metrics as well as human judgements. We focus on a zero-shot setting, i.e. prompting ChatGPT with specific task instructions, and test its performance on three commonly-used dimensions of text style transfer evaluation: style strength, content preservation, and fluency. We perform a comprehensive correlation analysis for two transfer directions (and overall) at different levels. Compared to existing automatic metrics, ChatGPT achieves competitive correlations with human judgments. These preliminary results are expected to provide a first glimpse into the role of large language models in the multidimensional evaluation of stylized text generation.","{'model': 'tldr@v2.0.0', 'text': 'Preliminary results are expected to provide a first glimpse into the role of large language models in the multidimensional evaluation of stylized text generation.'}",http://arxiv.org/pdf/2304.13462
-Yes but.. Can ChatGPT Identify Entities in Historical Documents?,Carlos-Emiliano Gonz'alez-Gallardo,"Large language models (LLMs) have been leveraged for several years now, obtaining state-of-the-art performance in recognizing entities from modern documents. For the last few months, the conversational agent ChatGPT has “prompted” a lot of interest in the scientific community and public due to its capacity of generating plausible-sounding answers. In this paper, we explore this ability by probing it in the named entity recognition and classification (NERC) task in primary sources (e.g., historical newspapers and classical commentaries) in a zero-shot manner and by comparing it with state-of-the-art LM-based systems. Our findings indicate several shortcomings in identifying entities in historical text that range from the consistency of entity annotation guidelines, entity complexity, and code-switching, to the specificity of prompting. Moreover, as expected, the inaccessibility of historical archives to the public (and thus on the Internet) also impacts its performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the ability of the conversational agent ChatGPT to generate plausible-sounding answers in the named entity recognition and classification task in primary sources in a zero-shot manner and compares it with state-of-the-art LM-based systems.'}",https://arxiv.org/pdf/2303.17322
-Chain of Thought Prompt Tuning in Vision Language Models,Jiaxin Ge,"Language-Image Pre-training has demonstrated promising results on zero-shot and few-shot downstream tasks by prompting visual models with natural language prompts. However, most recent studies only use a single prompt for tuning, neglecting the inherent step-to-step cognitive reasoning process that humans conduct in complex task settings, for example, when processing images from unfamiliar domains. Chain of Thought is a simple and effective approximation to human reasoning process and has been proven useful for natural language processing (NLP) tasks. Based on this cognitive intuition, we believe that conducting effective reasoning is also an important problem in visual tasks, and a chain of thought could be a solution to this problem. In this work, we propose a novel chain of thought prompt tuning for vision-language modeling. Extensive experiments show that our method not only generalizes better in image classification tasks, has greater transferability beyond a single dataset, and has stronger domain generalization performance, but also performs much better in imagetext retrieval and visual question answering, which require more reasoning capabilities. We are the first to successfully adapt chain-of-thought prompting that combines visual and textual embeddings. We will release our codes","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that the proposed novel chain of thought prompt tuning for vision-language modeling not only generalizes better in image classification tasks, has greater transferability beyond a single dataset, and has stronger domain generalization performance, but also performs much better in imagetext retrieval and visual question answering, which require more reasoning capabilities.'}",
-Can Language Models Understand Physical Concepts?,Lei Li,"Language models~(LMs) gradually become general-purpose interfaces in the interactive and embodied world, where the understanding of physical concepts is an essential prerequisite. However, it is not yet clear whether LMs can understand physical concepts in the human world. To investigate this, we design a benchmark VEC that covers the tasks of (i) Visual concepts, such as the shape and material of objects, and (ii) Embodied Concepts, learned from the interaction with the world such as the temperature of objects. Our zero (few)-shot prompting results show that the understanding of certain visual concepts emerges as scaling up LMs, but there are still basic concepts to which the scaling law does not apply. For example, OPT-175B performs close to humans with a zero-shot accuracy of 85\% on the material concept, yet behaves like random guessing on the mass concept. Instead, vision-augmented LMs such as CLIP and BLIP achieve a human-level understanding of embodied concepts. Analysis indicates that the rich semantics in visual representation can serve as a valuable source of embodied knowledge. Inspired by this, we propose a distillation method to transfer embodied knowledge from VLMs to LMs, achieving performance gain comparable with that by scaling up the parameters of LMs 134x. Our dataset is available at \url{https://github.com/TobiasLee/VEC}","{'model': 'tldr@v2.0.0', 'text': 'A distillation method is proposed to transfer embodied knowledge from VLMs to LMs, achieving performance gain comparable with that by scaling up the parameters of LMs 134x, and indicates that the rich semantics in visual representation can serve as a valuable source of embodied knowledge.'}",http://arxiv.org/pdf/2305.14057
-Evaluating Large Language Models on Graphs: Performance Insights and Comparative Analysis,Chang Liu,"Large Language Models (LLMs) have garnered considerable interest within both academic and industrial. Yet, the application of LLMs to graph data remains under-explored. In this study, we evaluate the capabilities of four LLMs in addressing several analytical problems with graph data. We employ four distinct evaluation metrics: Comprehension, Correctness, Fidelity, and Rectification. Our results show that: 1) LLMs effectively comprehend graph data in natural language and reason with graph topology. 2) GPT models can generate logical and coherent results, outperforming alternatives in correctness. 3) All examined LLMs face challenges in structural reasoning, with techniques like zero-shot chain-of-thought and few-shot prompting showing diminished efficacy. 4) GPT models often produce erroneous answers in multi-answer tasks, raising concerns in fidelity. 5) GPT models exhibit elevated confidence in their outputs, potentially hindering their rectification capacities. Notably, GPT-4 has demonstrated the capacity to rectify responses from GPT-3.5-turbo and its own previous iterations. The code is available at: https://github.com/Ayame1006/LLMtoGraph.","{'model': 'tldr@v2.0.0', 'text': 'This study evaluates the capabilities of four LLMs in addressing several analytical problems with graph data, using four distinct evaluation metrics: Comprehension, Correctness, Fidelity, and Rectification.'}",https://arxiv.org/pdf/2308.11224
-MasakhaNEWS: News Topic Classification for African languages,David Ifeoluwa Adelani,"African languages are severely under-represented in NLP research due to lack of datasets covering several NLP tasks. While there are individual language specific datasets that are being expanded to different tasks, only a handful of NLP tasks (e.g. named entity recognition and machine translation) have standardized benchmark datasets covering several geographical and typologically-diverse African languages. In this paper, we develop MasakhaNEWS -- a new benchmark dataset for news topic classification covering 16 languages widely spoken in Africa. We provide an evaluation of baseline models by training classical machine learning models and fine-tuning several language models. Furthermore, we explore several alternatives to full fine-tuning of language models that are better suited for zero-shot and few-shot learning such as cross-lingual parameter-efficient fine-tuning (like MAD-X), pattern exploiting training (PET), prompting language models (like ChatGPT), and prompt-free sentence transformer fine-tuning (SetFit and Cohere Embedding API). Our evaluation in zero-shot setting shows the potential of prompting ChatGPT for news topic classification in low-resource African languages, achieving an average performance of 70 F1 points without leveraging additional supervision like MAD-X. In few-shot setting, we show that with as little as 10 examples per label, we achieved more than 90\% (i.e. 86.0 F1 points) of the performance of full supervised training (92.6 F1 points) leveraging the PET approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops MasakhaNEWS -- a new benchmark dataset for news topic classification covering 16 languages widely spoken in Africa and provides an evaluation of baseline models by training classical machine learning models and fine-tuning several language models.'}",https://arxiv.org/pdf/2304.09972
-SpeechX: Neural Codec Language Model as a Versatile Speech Transformer,Xiaofei Wang,"Recent advancements in generative speech models based on audio-text prompts have enabled remarkable innovations like high-quality zero-shot text-to-speech. However, existing models still face limitations in handling diverse audio-text speech generation tasks involving transforming input speech and processing audio captured in adverse acoustic conditions. This paper introduces SpeechX, a versatile speech generation model capable of zero-shot TTS and various speech transformation tasks, dealing with both clean and noisy signals. SpeechX combines neural codec language modeling with multi-task learning using task-dependent prompting, enabling unified and extensible modeling and providing a consistent way for leveraging textual input in speech enhancement and transformation tasks. Experimental results show SpeechX's efficacy in various tasks, including zero-shot TTS, noise suppression, target speaker extraction, speech removal, and speech editing with or without background noise, achieving comparable or superior performance to specialized models across tasks. See https://aka.ms/speechx for demo samples.","{'model': 'tldr@v2.0.0', 'text': 'SpeechX combines neural codec language modeling with multi-task learning using task-dependent prompting, enabling unified and extensible modeling and providing a consistent way for leveraging textual input in speech enhancement and transformation tasks.'}",https://arxiv.org/pdf/2308.06873
-Is ChatGPT a Good Personality Recognizer? A Preliminary Study,Yuzhe Ji,"In recent years, personality has been regarded as a valuable personal factor being incorporated into numerous tasks such as sentiment analysis and product recommendation. This has led to widespread attention to text-based personality recognition task, which aims to identify an individual's personality based on given text. Considering that ChatGPT has recently exhibited remarkable abilities on various natural language processing tasks, we provide a preliminary evaluation of ChatGPT on text-based personality recognition task for generating effective personality data. Concretely, we employ a variety of prompting strategies to explore ChatGPT's ability in recognizing personality from given text, especially the level-oriented prompting strategy we designed for guiding ChatGPT in analyzing given text at a specified level. The experimental results on two representative real-world datasets reveal that ChatGPT with zero-shot chain-of-thought prompting exhibits impressive personality recognition ability and is capable to provide natural language explanations through text-based logical reasoning. Furthermore, by employing the level-oriented prompting strategy to optimize zero-shot chain-of-thought prompting, the performance gap between ChatGPT and corresponding state-of-the-art model has been narrowed even more. However, we observe that ChatGPT shows unfairness towards certain sensitive demographic attributes such as gender and age. Additionally, we discover that eliciting the personality recognition ability of ChatGPT helps improve its performance on personality-related downstream tasks such as sentiment classification and stress prediction.","{'model': 'tldr@v2.0.0', 'text': 'E eliciting the personality recognition ability of ChatGPT helps improve its performance on personality-related downstream tasks such as sentiment classification and stress prediction, and is found to provide natural language explanations through text-based logical reasoning.'}",https://arxiv.org/pdf/2307.03952
-Let's Do a Thought Experiment: Using Counterfactuals to Improve Moral Reasoning,Xiao Ma,"Language models still struggle on moral reasoning, despite their impressive performance in many other tasks. In particular, the Moral Scenarios task in MMLU (Multi-task Language Understanding) is among the worst performing tasks for many language models, including GPT-3. In this work, we propose a new prompting framework, Thought Experiments, to teach language models to do better moral reasoning using counterfactuals. Experiment results show that our framework elicits counterfactual questions and answers from the model, which in turn helps improve the accuracy on Moral Scenarios task by 9-16% compared to other zero-shot baselines. Interestingly, unlike math reasoning tasks, zero-shot Chain-of-Thought (CoT) reasoning doesn't work out of the box, and even reduces accuracy by around 4% compared to direct zero-shot. We further observed that with minimal human supervision in the form of 5 few-shot examples, the accuracy of the task can be improved to as much as 80%.","{'model': 'tldr@v2.0.0', 'text': 'A new prompting framework is proposed, Thought Experiments, to teach language models to do better moral reasoning using counterfactuals, and shows that with minimal human supervision in the form of 5 few-shot examples, the accuracy of the task can be improved to as much as 80%.'}",http://arxiv.org/pdf/2306.14308
-Symbolic Math Reasoning with Language Models,Vedant Gaur,"The emergence of large language models (LLMs) such as OpenAI’s GPT-3, Google’s LaMDA, Meta’s OPT [2, 3, 7, 10] etc. have revolutionized the field of natural language processing (NLP). These models with upwards of hundreds of billions of parameters are trained on large unlabeled text corpora and can subsequently solve downstream tasks with little to no labeled data. While these models are increasingly versatile in their abilities, e.g., solving math word problems, the larger question of their ability to reason remains. Using and modifying the SVAMP dataset, we find that GPT-3’s davinci-002 model, in addition to having good performance on numerical math word problems, also performs well on the potentially harder symbolic version of the same problems. Furthermore, adopting a two-step approach (solve symbolically and then substitute numerical values) leads to better accuracy on the numerical test set in the zero-shot regime. Additionally, we find that the use of specific prompting techniques pushes the model, in many cases, to actively describe its thought process and aid in the final answer output when faced with a complex, multi-step problem, aligning with recent observations.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3’s davinci-002 model, in addition to having good performance on numerical math word problems, also performs well on the potentially harder symbolic version of the same problems, and adopting a two-step approach leads to better accuracy on the numerical test set in the zero-shot regime.'}",
-Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them,Mirac Suzgun,"BIG-Bench (Srivastava et al., 2022) is a diverse evaluation suite that focuses on tasks believed to be beyond the capabilities of current language models. Language models have already made good progress on this benchmark, with the best model in the BIG-Bench paper outperforming average reported human-rater results on 65% of the BIG-Bench tasks via few-shot prompting. But on what tasks do language models fall short of average human-rater performance, and are those tasks actually unsolvable by current language models? In this work, we focus on a suite of 23 challenging BIG-Bench tasks which we call BIG-Bench Hard (BBH). These are the task for which prior language model evaluations did not outperform the average human-rater. We find that applying chain-of-thought (CoT) prompting to BBH tasks enables PaLM to surpass the average human-rater performance on 10 of the 23 tasks, and Codex (code-davinci-002) to surpass the average human-rater performance on 17 of the 23 tasks. Since many tasks in BBH require multi-step reasoning, few-shot prompting without CoT, as done in the BIG-Bench evaluations (Srivastava et al., 2022), substantially underestimates the best performance and capabilities of language models, which is better captured via CoT prompting. As further analysis, we explore the interaction between CoT and model scale on BBH, finding that CoT enables emergent task performance on several BBH tasks with otherwise flat scaling curves.","{'model': 'tldr@v2.0.0', 'text': 'This work finds that applying chain-of-thought (CoT) prompting to BBH tasks enables PaLM to surpass the average human-rater performance on 10 of the 23 tasks, and Codex to surpass it on 17 of the23 tasks.'}",http://arxiv.org/pdf/2210.09261
-Do GPTs Produce Less Literal Translations?,Vikas Raunak,"Large Language Models (LLMs) such as GPT-3 have emerged as general-purpose language models capable of addressing many natural language generation or understanding tasks. On the task of Machine Translation (MT), multiple works have investigated few-shot prompting mechanisms to elicit better translations from LLMs. However, there has been relatively little investigation on how such translations differ qualitatively from the translations generated by standard Neural Machine Translation (NMT) models. In this work, we investigate these differences in terms of the literalness of translations produced by the two systems. Using literalness measures involving word alignment and monotonicity, we find that translations out of English (E-X) from GPTs tend to be less literal, while exhibiting similar or better scores on MT quality metrics. We demonstrate that this finding is borne out in human evaluations as well. We then show that these differences are especially pronounced when translating sentences that contain idiomatic expressions.","{'model': 'tldr@v2.0.0', 'text': 'Using literalness measures involving word alignment and monotonicity, it is found that translations out of English from GPTs tend to be less literal, while exhibiting similar or better scores on MT quality metrics, and this finding is borne out in human evaluations.'}",http://arxiv.org/pdf/2305.16806
-Bootstrapping Multilingual Semantic Parsers using Large Language Models,Abhijeet Awasthi,"Despite cross-lingual generalization demonstrated by pre-trained multilingual models, the translate-train paradigm of transferring English datasets across multiple languages remains to be a key mechanism for training task-specific multilingual models. However, for many low-resource languages, the availability of a reliable translation service entails significant amounts of costly human-annotated translation pairs. Further, translation services may continue to be brittle due to domain mismatch between task-specific input text and general-purpose text used for training translation models. For multilingual semantic parsing, we demonstrate the effectiveness and flexibility offered by large language models (LLMs) for translating English datasets into several languages via few-shot prompting. Through extensive comparisons on two public datasets, MTOP and MASSIVE, spanning 50 languages and several domains, we show that our method of translating data using LLMs outperforms a strong translate-train baseline on 41 out of 50 languages. We study the key design choices that enable more effective multilingual data translation via prompted LLMs.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness and flexibility offered by large language models (LLMs) for translating English datasets into several languages via few-shot prompting are demonstrated and it is shown that the method of translating data using LLMs outperforms a strong translate-train baseline on 41 out of 50 languages.'}",http://arxiv.org/pdf/2210.07313
-InstructExcel: A Benchmark for Natural Language Instruction in Excel,Justin Payan,"With the evolution of Large Language Models (LLMs) we can solve increasingly more complex NLP tasks across various domains, including spreadsheets. This work investigates whether LLMs can generate code (Excel OfficeScripts, a TypeScript API for executing many tasks in Excel) that solves Excel specific tasks provided via natural language user instructions. To do so we introduce a new large-scale benchmark, InstructExcel, created by leveraging the 'Automate' feature in Excel to automatically generate OfficeScripts from users' actions. Our benchmark includes over 10k samples covering 170+ Excel operations across 2,000 publicly available Excel spreadsheets. Experiments across various zero-shot and few-shot settings show that InstructExcel is a hard benchmark for state of the art models like GPT-4. We observe that (1) using GPT-4 over GPT-3.5, (2) providing more in-context examples, and (3) dynamic prompting can help improve performance on this benchmark.","{'model': 'tldr@v2.0.0', 'text': ""This work investigates whether LLMs can generate code (Excel OfficeScripts, a TypeScript API for executing many tasks in Excel) that solves Excel specific tasks provided via natural language user instructions by leveraging the 'Automate' feature in Excel to automatically generate Office Scripts from users' actions.""}",
-Business Process Text Sketch Automation Generation Using Large Language Model,Rui Zhu,"Business Process Management (BPM) is gaining increasing attention as it has the potential to cut costs while boosting output and quality. Business process document generation is a crucial stage in BPM. However, due to a shortage of datasets, data-driven deep learning techniques struggle to deliver the expected results. We propose an approach to transform Conditional Process Trees (CPTs) into Business Process Text Sketches (BPTSs) using Large Language Models (LLMs). The traditional prompting approach (Few-shot In-Context Learning) tries to get the correct answer in one go, and it can find the pattern of transforming simple CPTs into BPTSs, but for close-domain and CPTs with complex hierarchy, the traditional prompts perform weakly and with low correctness. We suggest using this technique to break down a difficult CPT into a number of basic CPTs and then solve each one in turn, drawing inspiration from the divide-and-conquer strategy. We chose 100 process trees with depths ranging from 2 to 5 at random, as well as CPTs with many nodes, many degrees of selection, and cyclic nesting. Experiments show that our method can achieve a correct rate of 93.42%, which is 45.17% better than traditional prompting methods. Our proposed method provides a solution for business process document generation in the absence of datasets, and secondly, it becomes potentially possible to provide a large number of datasets for the process model extraction (PME) domain.","{'model': 'tldr@v2.0.0', 'text': 'This proposed method provides a solution for business process document generation in the absence of datasets, and secondly, it becomes potentially possible to provide a large number of datasets for the process model extraction (PME) domain.'}",https://arxiv.org/pdf/2309.01071
-An Investigation of LLMs' Inefficacy in Understanding Converse Relations,Chengwen Qi,"Large Language Models (LLMs) have achieved remarkable success in many formal language oriented tasks, such as structural data-to-text and semantic parsing. However current benchmarks mostly follow the data distribution of the pre-training data of LLMs. Therefore, a natural question rises that do LLMs really understand the structured semantics of formal languages. In this paper, we investigate this problem on a special case, converse binary relation. We introduce a new benchmark ConvRe focusing on converse relations, which contains 17 relations and 1240 triples extracted from popular knowledge graph completion datasets. Our ConvRE features two tasks, Re2Text and Text2Re, which are formulated as multi-choice question answering to evaluate LLMs' ability to determine the matching between relations and associated text. For the evaluation protocol, apart from different prompting methods, we further introduce variants to the test text and few-shot example text. We conduct experiments on three popular LLM families and have observed various scaling trends. The results suggest that LLMs often resort to shortcut learning and still face challenges on our proposed benchmark.","{'model': 'tldr@v2.0.0', 'text': 'A new benchmark ConvRe focusing on converse relations, which contains 17 relations and 1240 triples extracted from popular knowledge graph completion datasets is introduced, which suggests that LLMs often resort to shortcut learning and still face challenges on this proposed benchmark.'}",https://arxiv.org/pdf/2310.05163
-Uniform Sampling over Episode Difficulty,Sébastien M. R. Arnold,"Episodic training is a core ingredient of few-shot learning to train models on tasks with limited labelled data. Despite its success, episodic training remains largely understudied, prompting us to ask the question: what is the best way to sample episodes? In this paper, we first propose a method to approximate episode sampling distributions based on their difficulty. Building on this method, we perform an extensive analysis and find that sampling uniformly over episode difficulty outperforms other sampling schemes, including curriculum and easy-/hard-mining. As the proposed sampling method is algorithm agnostic, we can leverage these insights to improve few-shot learning accuracies across many episodic training algorithms. We demonstrate the efficacy of our method across popular few-shot learning datasets, algorithms, network architectures, and protocols.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a method to approximate episode sampling distributions based on their difficulty and finds that sampling uniformly over episode difficulty outperforms other sampling schemes, including curriculum and easy-/hard-mining.'}",
-"Of eight oscillations and several messages carved in flesh: Spectacle, spectatorship, cognition, and affect in Dredd and Looper",M. Bould,"Both popular and academic criticism tend to decry sf cinema’s commitment to spectacle and special effects as intellectually stultifying and thus politically narcotizing. This article challenges the class politics of taste and the crude models of interpolation underpinning such claims, questioning the ease with which many critics separate matters of cognition from matters of affect. It examines in detail three sequences from the mainstream, small-to-medium budget sf movies Dredd (2012) and Looper (2012). The climactic set piece from Dredd demonstrates the contradictory but entangled array of unstable subject positions opened up by cinematic spectacle. A sequence from Looper, which includes the most complicated special effects shot in the movie, shows how effects work can affect the viewer while simultaneously elaborating narrative/ world-building information that demands a cognitive response. A second sequence from Looper works in a similar way, ingeniously prompting the viewer to draw on intertextual knowledge to comprehend a shot that conveys complex narrative information and that ultimately refuses the specific special effect it has led the viewer to anticipate.",,
-Method and system for detecting living body human face,王先基,"The invention relates to a method and system for detecting a living body human face. The method comprises the steps of obtaining images from a shot video sequence, detecting a human face in the video images, carrying out location on facial critical areas, randomly selecting one or more kinds of movements in a movement collection through a critical area detection classifier, randomly selecting the completing times of the movements, prompting a user, and determining the user as a living body if the user completes the random selected movements for many times within an appointed time period, otherwise, determining the user as a non-living body. According to the method and system for detecting the living body human face, when living body detection is carried out, the system randomly selects certain movements and randomly appoints the completing times of the movements every time, and therefore the user cannot pass the living body detection by using photos for cheating. In addition, due to the fact that the detected items are selected randomly, the user cannot pass the living body detection through the mode of playing video clips. Therefore, the method and system for detecting the living body human face can better prevent cheating behaviors in the mode such as the photos and the video clips in the living body detection, and the reliability and the safety of the living body detection are made to be higher.","{'model': 'tldr@v2.0.0', 'text': 'The method and system for detecting the live body human face can better prevent cheating behaviors in the mode such as the photos and the video clips in the living body detection, and the reliability and the safety of the livingBody detection are made to be higher.'}",
-Zero-Shot Text-Guided Object Generation with Dream Fields,Ajay Jain,"We combine neural rendering with multi-modal image and text representations to synthesize diverse 3D objects solely from natural language descriptions. Our method, Dream Fields, can generate the geometry and color of a wide range of objects without 3D supervision. Due to the scarcity of diverse, captioned 3D data, prior methods only generate objectsfrom a handful of categories, such as ShapeNet. Instead, we guide generation with image-text models pre-trained on large datasets of captioned images from the web. Our method optimizes a Neural Radiance Field from many camera views so that rendered images score highly with a target caption according to a pre-trained CLIP model. To improve fidelity and visual quality, we introduce simple geometric priors, including sparsity-inducing transmittance regularization, scene bounds, and new MLP architectures. In experiments, Dream Fields produce realistic, multi-view consistent object geometry and color from a variety of natural language captions.","{'model': 'tldr@v2.0.0', 'text': 'This work combines neural rendering with multi-modal image and text representations to synthesize diverse 3D objects solely from natural language descriptions, and introduces simple geometric priors, including sparsity-inducing transmittance regularization, scene bounds, and new MLP architectures.'}",https://arxiv.org/pdf/2112.01455
-Large Language Models are Few-shot Testers: Exploring LLM-based General Bug Reproduction,Sungmin Kang,"Many automated test generation techniques have been developed to aid developers with writing tests. To facilitate full automation, most existing techniques aim to either increase coverage, or generate exploratory inputs. However, existing test generation techniques largely fall short of achieving more semantic objectives, such as generating tests to reproduce a given bug report. Reproducing bugs is nonetheless important, as our empirical study shows that the number of tests added in open source repositories due to issues was about 28% of the corresponding project test suite size. Meanwhile, due to the difficulties of transforming the expected program semantics in bug reports into test oracles, existing failure reproduction techniques tend to deal exclusively with program crashes, a small subset of all bug reports. To automate test generation from general bug reports, we propose Libro, a framework that uses Large Language Models (LLMs), which have been shown to be capable of performing code-related tasks. Since LLMs themselves cannot execute the target buggy code, we focus on post-processing steps that help us discern when LLMs are effective, and rank the produced tests according to their validity. Our evaluation of Libro shows that, on the widely studied Defects4J benchmark, Libro can generate failure reproducing test cases for 33% of all studied cases (251 out of 750), while suggesting a bug reproducing test in first place for 149 bugs. To mitigate data contamination (i.e., the possibility of the LLM simply remembering the test code either partially or in whole), we also evaluate Libro against 31 bug reports submitted after the collection of the LLM training data terminated: Libro produces bug reproducing tests for 32% of the studied bug reports. Overall, our results show Libro has the potential to significantly enhance developer efficiency by automatically generating tests from bug reports.","{'model': 'tldr@v2.0.0', 'text': 'The results show Libro has the potential to significantly enhance developer efficiency by automatically generating tests from bug reports, a framework that uses Large Language Models (LLMs), which have been shown to be capable of performing code-related tasks.'}",https://arxiv.org/pdf/2209.11515
-Meta Faster R-CNN: Towards Accurate Few-Shot Object Detection with Attentive Feature Alignment,G. Han,"Few-shot object detection (FSOD) aims to detect objects using only a few examples. How to adapt state-of-the-art object detectors to the few-shot domain remains challenging. Object proposal is a key ingredient in modern object detectors. However, the quality of proposals generated for few-shot classes using existing methods is far worse than that of many-shot classes, e.g., missing boxes for few-shot classes due to misclassification or inaccurate spatial locations with respect to true objects. To address the noisy proposal problem, we propose a novel meta-learning based FSOD model by jointly optimizing the few-shot proposal generation and fine-grained few-shot proposal classification. To improve proposal generation for few-shot classes, we propose to learn a lightweight metric-learning based prototype matching network, instead of the conventional simple linear object/nonobject classifier, e.g., used in RPN. Our non-linear classifier with the feature fusion network could improve the discriminative prototype matching and the proposal recall for few-shot classes. To improve the fine-grained few-shot proposal classification, we propose a novel attentive feature alignment method to address the spatial misalignment between the noisy proposals and few-shot classes, thus improving the performance of few-shot object detection. Meanwhile we learn a separate Faster R-CNN detection head for many-shot base classes and show strong performance of maintaining base-classes knowledge. Our model achieves state-of-the-art performance on multiple FSOD benchmarks over most of the shots and metrics.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-learning based FSOD model is proposed that achieves state-of-the-art performance on multiple FSOD benchmarks over most of the shots and metrics and a novel attentive feature alignment method is proposed to address the spatial misalignment between the noisy proposals and few-shot classes, thus improving the performance of few- shot object detection.'}",https://ojs.aaai.org/index.php/AAAI/article/download/19959/19718
-D2C: Diffusion-Decoding Models for Few-Shot Conditional Generation,Abhishek Sinha,"Conditional generative models of high-dimensional images have many applications, but supervision signals from conditions to images can be expensive to acquire. This paper describes Diffusion-Decoding models with Contrastive representations (D2C), a paradigm for training unconditional variational autoencoders (VAEs) for few-shot conditional image generation. D2C uses a learned diffusion-based prior over the latent representations to improve generation and contrastive self-supervised learning to improve representation quality. D2C can adapt to novel generation tasks conditioned on labels or manipulation constraints, by learning from as few as 100 labeled examples. On conditional generation from new labels, D2C achieves superior performance over state-of-the-art VAEs and diffusion models. On conditional image manipulation, D2C generations are two orders of magnitude faster to produce over StyleGAN2 ones and are preferred by 50% - 60% of the human evaluators in a double-blind study.","{'model': 'tldr@v2.0.0', 'text': 'D2C uses a learned diffusion-based prior over the latent representations to improve generation and contrastive self-supervised learning to improve representation quality, and achieves superior performance over state-of-the-art VAEs and diffusion models.'}",
-Few-shot learning for medical text: A systematic review,Yao Ge,"Objective: Few-shot learning (FSL) methods require small numbers of labeled instances for training. As many medical topics have limited annotated textual data in practical settings, FSL-based natural language processing (NLP) methods hold substantial promise. We aimed to conduct a systematic review to explore the state of FSL methods for medical NLP. Materials and Methods: We searched for articles published between January 2016 and August 2021 using PubMed/Medline, Embase, ACL Anthology, and IEEE Xplore Digital Library. To identify the latest relevant methods, we also searched other sources such as preprint servers (eg., medRxiv) via Google Scholar. We included all articles that involved FSL and any type of medical text. We abstracted articles based on data source(s), aim(s), training set size(s), primary method(s)/approach(es), and evaluation method(s). Results: 31 studies met our inclusion criteria-all published after 2018; 22 (71%) since 2020. Concept extraction/named entity recognition was the most frequently addressed task (13/31; 42%), followed by text classification (10/31; 32%). Twenty-one (68%) studies reconstructed existing datasets to create few-shot scenarios synthetically, and MIMIC-III was the most frequently used dataset (7/31; 23%). Common methods included FSL with attention mechanisms (12/31; 39%), prototypical networks (8/31; 26%), and meta-learning (6/31; 19%). Discussion: Despite the potential for FSL in biomedical NLP, progress has been limited compared to domain-independent FSL. This may be due to the paucity of standardized, public datasets, and the relative underperformance of FSL methods on biomedical topics. Creation and release of specialized datasets for biomedical FSL may aid method development by enabling comparative analyses.","{'model': 'tldr@v2.0.0', 'text': 'Despite the potential for FSL in biomedical NLP, progress has been limited compared to domain-independent FSL, may be due to the paucity of standardized, public datasets, and the relative underperformance of FSL methods on biomedical topics.'}",http://arxiv.org/pdf/2204.14081
-Relative and Absolute Location Embedding for Few-Shot Node Classification on Graph,Zemin Liu,"Node classification is an important problem on graphs. While recent advances in graph neural networks achieve promising performance, they require abundant labeled nodes for training. However, in many practical scenarios, there often exist novel classes in which only one or a few labeled nodes are available as supervision, known as few-shot node classification. Although meta-learning has been widely used in vision and language domains to address few-shot learning, its adoption on graphs has been limited. In particular, graph nodes in a few-shot task are not independent and relate to each other. To deal with this, we propose a novel model called Relative and Absolute Location Embedding (RALE) hinged on the concept of hub nodes. Specifically, RALE captures the task-level dependency by assigning each node a relative location within a task, as well as the graph-level dependency by assigning each node an absolute location on the graph to further align different tasks toward learning a transferable prior. Finally, extensive experiments on three public datasets demonstrate the state-of-the-art performance of RALE.","{'model': 'tldr@v2.0.0', 'text': 'A novel model called Relative and Absolute Location Embedding (RALE) hinged on the concept of hub nodes is proposed, which captures the task-level dependency by assigning each node a relative location within a task, as well as the graph- level dependency by assign each node an absolute location on the graph to further align different tasks toward learning a transferable prior.'}",https://ojs.aaai.org/index.php/AAAI/article/download/16551/16358
-Few-Shot Intent Detection via Contrastive Pre-Training and Fine-Tuning,Jianguo Zhang,"In this work, we focus on a more challenging few-shot intent detection scenario where many intents are fine-grained and semantically similar. We present a simple yet effective few-shot intent detection schema via contrastive pre-training and fine-tuning. Specifically, we first conduct self-supervised contrastive pre-training on collected intent datasets, which implicitly learns to discriminate semantically similar utterances without using any labels. We then perform few-shot intent detection together with supervised contrastive learning, which explicitly pulls utterances from the same intent closer and pushes utterances across different intents farther. Experimental results show that our proposed method achieves state-of-the-art performance on three challenging intent detection datasets under 5-shot and 10-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts self-supervised contrastive pre-training on collected intent datasets and performs few-shot intent detection together with supervised contrastive learning, which explicitly pulls utterances from the same intent closer and pushes utterances across different intents farther.'}",https://aclanthology.org/2021.emnlp-main.144.pdf
-Relational Learning with Gated and Attentive Neighbor Aggregator for Few-Shot Knowledge Graph Completion,Guanglin Niu,"Aiming at expanding few-shot relations' coverage in knowledge graphs (KGs), few-shot knowledge graph completion (FKGC) has recently gained more research interests. Some existing models employ a few-shot relation's multi-hop neighbor information to enhance its semantic representation. However, noise neighbor information might be amplified when the neighborhood is excessively sparse and no neighbor is available to represent the few-shot relation. Moreover, modeling and inferring complex relations of one-to-many (1-N), many-to-one (N-1), and many-to-many (N-N) by previous knowledge graph completion approaches requires high model complexity and a large amount of training instances. Thus, inferring complex relations in the few-shot scenario is difficult for FKGC models due to limited training instances. In this paper, we propose a few-shot relational learning with global-local framework to address the above issues. At the global stage, a novel gated and attentive neighbor aggregator is built for accurately integrating the semantics of a few-shot relation's neighborhood, which helps filtering the noise neighbors even if a KG contains extremely sparse neighborhoods. For the local stage, a meta-learning based TransH (MTransH) method is designed to model complex relations and train our model in a few-shot learning fashion. Extensive experiments show that our model outperforms the state-of-the-art FKGC approaches on the frequently-used benchmark datasets NELL-One and Wiki-One. Compared with the strong baseline model MetaR, our model achieves 5-shot FKGC performance improvements of 8.0% on NELL-One and 2.8% on Wiki-One by the metric Hits@10.","{'model': 'tldr@v2.0.0', 'text': ""A novel gated and attentive neighbor aggregator is built for accurately integrating the semantics of a few-shot relation's neighborhood, which helps filtering the noise neighbors even if a KG contains extremely sparse neighborhoods.""}",https://arxiv.org/pdf/2104.13095
-RAFT: A Real-World Few-Shot Text Classification Benchmark,Neel Alex,"Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? Existing benchmarks are not designed to measure progress in applied settings, and so don't directly answer this question. The RAFT benchmark (Real-world Annotated Few-shot Tasks) focuses on naturally occurring tasks and uses an evaluation setup that mirrors deployment. Baseline evaluations on RAFT reveal areas current techniques struggle with: reasoning over long texts and tasks with many classes. Human baselines show that some classification tasks are difficult for non-expert humans, reflecting that real-world value sometimes depends on domain expertise. Yet even non-expert human baseline F1 scores exceed GPT-3 by an average of 0.11. The RAFT datasets and leaderboard will track which model improvements translate into real-world benefits at https://raft.elicit.org .","{'model': 'tldr@v2.0.0', 'text': 'The RAFT benchmark (Real-world Annotated Few-shot Tasks) focuses on naturally occurring tasks and uses an evaluation setup that mirrors deployment, revealing areas current techniques struggle with: reasoning over long texts and tasks with many classes.'}",
-Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation,Biao Zhang,"Massively multilingual models for neural machine translation (NMT) are theoretically attractive, but often underperform bilingual models and deliver poor zero-shot translations. In this paper, we explore ways to improve them. We argue that multilingual NMT requires stronger modeling capacity to support language pairs with varying typological characteristics, and overcome this bottleneck via language-specific components and deepening NMT architectures. We identify the off-target translation issue (i.e. translating into a wrong target language) as the major source of the inferior zero-shot performance, and propose random online backtranslation to enforce the translation of unseen training language pairs. Experiments on OPUS-100 (a novel multilingual dataset with 100 languages) show that our approach substantially narrows the performance gap with bilingual models in both one-to-many and many-to-many settings, and improves zero-shot performance by ~10 BLEU, approaching conventional pivot-based methods.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that multilingual NMT requires stronger modeling capacity to support language pairs with varying typological characteristics, and overcome this bottleneck via language-specific components and deepening NMT architectures.'}",https://www.aclweb.org/anthology/2020.acl-main.148.pdf
-Zero-shot Generalization in Dialog State Tracking through Generative Question Answering,Shuyang Li,"Dialog State Tracking (DST), an integral part of modern dialog systems, aims to track user preferences and constraints (slots) in task-oriented dialogs. In real-world settings with constantly changing services, DST systems must generalize to new domains and unseen slot types. Existing methods for DST do not generalize well to new slot names and many require known ontologies of slot types and values for inference. We introduce a novel ontology-free framework that supports natural language queries for unseen constraints and slots in multi-domain task-oriented dialogs. Our approach is based on generative question-answering using a conditional language model pre-trained on substantive English sentences. Our model improves joint goal accuracy in zero-shot domain adaptation settings by up to 9% (absolute) over the previous state-of-the-art on the MultiWOZ 2.1 dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel ontology-free framework that supports natural language queries for unseen constraints and slots in multi-domain task-oriented dialogs based on generative question-answering using a conditional language model pre-trained on substantive English sentences.'}",https://aclanthology.org/2021.eacl-main.91.pdf
-Zero-shot Visual Question Answering using Knowledge Graph,Zhuo Chen,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Zero-shot VQA algorithm using knowledge graphs and a mask-based learning mechanism for better incorporating external knowledge, and presents new answer-based Zero- shot VQ a splits for the F-VQA dataset.'}",https://arxiv.org/pdf/2107.05348
-Single-shot simulations of dynamic quantum many-body systems,K. Sakmann,,,https://arxiv.org/pdf/1501.03224
-Improving Zero-shot Generalization and Robustness of Multi-Modal Models,Yunhao Ge,"Multi-modal- image-text models such as CLIP and LiT have demonstrated impressive performance on image classification benchmarks and their zero-shot generalization ability is particularly exciting. While the top-5 zero-shot accuracies of these models are very high, the top-1 accuracies are much lower (over 25% gap in some cases). We investigate the reasons for this performance gap and find that many of the failure cases are caused by ambiguity in the text prompts. First, we develop a simple and efficient zero-shot post-hoc method to identify images whose top-1 prediction is likely to be incorrect, by measuring consistency of the predictions w.r.t. multiple prompts and image transformations. We show that our procedure better predicts mistakes, outperforming the popular max logit baseline on selective prediction tasks. Next, we propose a simple and efficient way to improve accuracy on such uncertain images by making use of the WordNet hierarchy; specifically we augment the original class by incorporating its parent and children from the semantic label hierarchy, and plug the augmentation into text prompts. We conduct experiments on both CLIP and LiT models with five different ImageNet-based datasets. For CLIP, our method improves the top-1 accuracy by 17.13% on the uncertain subset and 3.6% on the entire ImageNet validation set. We also show that our method improves across ImageNet shifted datasets, four other datasets, and other model architectures such as LiT. The proposed method11Work carried out mainly at Google is hyperparameter-free, requires no additional model training and can be easily scaled to other large multi-modal architectures. Code is available at https://github.com/gyhandy/Hierarchy-CLIP.","{'model': 'tldr@v2.0.0', 'text': 'A simple and efficient zero-shot post-hoc method to identify images whose top-1 prediction is likely to be incorrect, by measuring consistency of the predictions w.r.t. multiple prompts and image transformations is developed.'}",https://arxiv.org/pdf/2212.01758
-Enhancing Cross-lingual Natural Language Inference by Soft Prompting with Language-independent Knowledge,,"Cross-lingual natural language inference is 001 a fundamental problem in cross-lingual lan- 002 guage understanding. Many recent works 003 have used prompt learning to address the lack 004 of annotated parallel corpora in XNLI. How- 005 ever, these methods adopt discrete prompting 006 by simply translating the template to the tar- 007 get language and need external expert knowl- 008 edge to design the templates. Besides, dis- 009 crete prompts of human-designed template 010 words are not trainable vectors which can 011 be migrated to target languages in the infer- 012 ence stage flexibly. In this paper, we pro- 013 pose a novel Soft prompt learning frame- 014 work enhanced by L anguage- IN dependent 015 K nowledge (SoftLINK) for XNLI. SoftLINK 016 first constructs cloze-style question with soft 017 prompts for the input sample. Then we lever- 018 age bilingual dictionaries to generate an aug- 019 mented multilingual question for the original 020 question. SoftLINK also adopts a multilingual 021 verbalizer to align the representations of origi- 022 nal and augmented multilingual questions on 023 the semantic space with consistency regular- 024 ization. Experimental results on XNLI demon- 025 strate that SoftLINK can achieve state-of-the- 026 art performance and significantly outperform 027 the previous methods under the few-shot and 028 full-shot cross-lingual transfer settings. 029","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on XNLI show that SoftLINK can achieve state-of-the-art art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.'}",
-Meta Self-training for Few-shot Neural Sequence Labeling,Yaqing Wang,"Neural sequence labeling is widely adopted for many Natural Language Processing (NLP) tasks, such as Named Entity Recognition (NER) and slot tagging for dialog systems and semantic parsing. Recent advances with large-scale pre-trained language models have shown remarkable success in these tasks when fine-tuned on large amounts of task-specific labeled data. However, obtaining such large-scale labeled training data is not only costly, but also may not be feasible in many sensitive user applications due to data access and privacy constraints. This is exacerbated for sequence labeling tasks requiring such annotations at token-level. In this work, we develop techniques to address the label scarcity challenge for neural sequence labeling models. Specifically, we propose a meta self-training framework which leverages very few manually annotated labels for training neural sequence models. While self-training serves as an effective mechanism to learn from large amounts of unlabeled data via iterative knowledge exchange -- meta-learning helps in adaptive sample re-weighting to mitigate error propagation from noisy pseudo-labels. Extensive experiments on six benchmark datasets including two for massive multilingual NER and four slot tagging datasets for task-oriented dialog systems demonstrate the effectiveness of our method. With only 10 labeled examples for each class in each task, the proposed method achieves 10% improvement over state-of-the-art methods demonstrating its effectiveness for limited training labels regime.","{'model': 'tldr@v2.0.0', 'text': 'A meta self-training framework which leverages very few manually annotated labels for training neural sequence models and achieves 10% improvement over state-of-the-art methods demonstrating its effectiveness for limited training labels regime.'}",https://dl.acm.org/doi/pdf/10.1145/3447548.3467235
-Precise Zero-Shot Dense Retrieval without Relevance Labels,Luyu Gao,"While dense retrieval has been shown to be effective and efficient across tasks and languages, it remains difficult to create effective fully zero-shot dense retrieval systems when no relevance labels are available. In this paper, we recognize the difficulty of zero-shot learning and encoding relevance. Instead, we propose to pivot through Hypothetical Document Embeddings (HyDE). Given a query, HyDE first zero-shot prompts an instruction-following language model (e.g., InstructGPT) to generate a hypothetical document. The document captures relevance patterns but is “fake” and may contain hallucinations. Then, an unsupervised contrastively learned encoder (e.g., Contriever) encodes the document into an embedding vector. This vector identifies a neighborhood in the corpus embedding space, from which similar real documents are retrieved based on vector similarity. This second step grounds the generated document to the actual corpus, with the encoder’s dense bottleneck filtering out the hallucinations. Our experiments show that HyDE significantly outperforms the state-of-the-art unsupervised dense retriever Contriever and shows strong performance comparable to fine-tuned retrievers across various tasks (e.g. web search, QA, fact verification) and in non-English languages (e.g., sw, ko, ja, bn).","{'model': 'tldr@v2.0.0', 'text': 'The proposed Hypothetical Document Embeddings (HyDE) significantly outperforms the state-of-the-art unsupervised dense retriever Contriever and shows strong performance comparable to fine-tuned retrievers across various tasks and in non-English languages.'}",http://arxiv.org/pdf/2212.10496
-SeqZero: Few-shot Compositional Semantic Parsing with Sequential Prompts and Zero-shot Models,Jingfeng Yang,"Recent research showed promising results on combining pretrained language models (LMs) with canonical utterance for few-shot semantic parsing. The canonical utterance is often lengthy and complex due to the compositional structure of formal languages. Learning to generate such canonical utterance requires significant amount of data to reach high performance. Fine-tuning with only few-shot samples, the LMs can easily forget pretrained knowledge, overfit spurious biases, and suffer from compositionally out-of-distribution generalization errors. To tackle these issues, we propose a novel few-shot semantic parsing method -- SeqZero. SeqZero decomposes the problem into a sequence of sub-problems, which correspond to the sub-clauses of the formal language. Based on the decomposition, the LMs only need to generate short answers using prompts for predicting sub-clauses. Thus, SeqZero avoids generating a long canonical utterance at once. Moreover, SeqZero employs not only a few-shot model but also a zero-shot model to alleviate the overfitting. In particular, SeqZero brings out the merits from both models via ensemble equipped with our proposed constrained rescaling. SeqZero achieves SOTA performance of BART-based models on GeoQuery and EcommerceQuery, which are two few-shot datasets with compositional data split.","{'model': 'tldr@v2.0.0', 'text': 'SeqZero achieves SOTA performance of BART-based models on GeoQuery and EcommerceQuery, which are two few-shot datasets with compositional data split and brings out the merits from both models via ensemble equipped with the proposed constrained rescaling.'}",https://arxiv.org/pdf/2205.07381
-Few-Shot Cross-Lingual Stance Detection with Sentiment-Based Pre-Training,Momchil Hardalov,"The goal of stance detection is to determine the viewpoint expressed in a piece of text towards a target. These viewpoints or contexts are often expressed in many different languages depending on the user and the platform, which can be a local news outlet, a social media platform, a news forum, etc. Most research on stance detection, however, has been limited to working with a single language and on a few limited targets, with little work on cross-lingual stance detection. Moreover, non-English sources of labelled data are often scarce and present additional challenges. Recently, large multilingual language models have substantially improved the performance on many non-English tasks, especially such with a limited number of examples. This highlights the importance of model pre-training and its ability to learn from few examples. In this paper, we present the most comprehensive study of cross-lingual stance detection to date: we experiment with 15 diverse datasets in 12 languages from 6 language families, and with 6 low-resource evaluation settings each. For our experiments, we build on pattern-exploiting training (PET), proposing the addition of a novel label encoder to simplify the verbalisation procedure. We further propose sentiment-based generation of stance data for pre-training, which shows sizeable improvement of more than 6% F1 absolute in few-shot learning settings compared to several strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the most comprehensive study of cross-lingual stance detection to date, and proposes sentiment-based generation of stance data for pre-training, which shows sizeable improvement of more than 6% F1 absolute in few-shot learning settings compared to several strong baselines.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21318/21067
-Z-ICL: Zero-Shot In-Context Learning with Pseudo-Demonstrations,Xinxi Lyu,"Although large language models can be prompted for both zero- and few-shot learning, performance drops significantly when no demonstrations are available. In this paper, we introduce Z-ICL, a new zero-shot method that closes the gap by constructing pseudo-demonstrations for a given test input using a raw text corpus. Concretely, pseudo-demonstrations are constructed by (1) finding the nearest neighbors to the test input from the corpus and pairing them with random task labels, and (2) applying a set of techniques to reduce the amount of direct copying the model does from the resulting demonstrations. Evaluation on nine classification datasets shows that Z-ICL outperforms previous zero-shot methods by a significant margin, and is on par with in-context learning with labeled training data in the few-shot setting. Overall, Z-ICL provides a significantly higher estimate of the zero-shot performance levels of a model, and supports future efforts to develop better pseudo-demonstrations that further improve zero-shot results.","{'model': 'tldr@v2.0.0', 'text': 'Z-ICL is introduced, a new zero-shot method that closes the gap by constructing pseudo-demonstrations for a given test input using a raw text corpus, and is on par with in-context learning with labeled training data in the few-shot setting.'}",http://arxiv.org/pdf/2212.09865
-From Images to Textual Prompts: Zero-shot Visual Question Answering with Frozen Large Language Models,Jiaxian Guo,"Large language models (LLMs) have demonstrated excellent zero-shot generalization to new language tasks. However, effective utilization of LLMs for zero-shot visual question-answering (VQA) remains challenging, primarily due to the modality disconnect and task disconnect between the LLM and VQA tasks. End-to-end training on multimodal data may bridge the disconnects, but is inflexible and computationally expensive. To address this issue, we propose Img2LLM, a plug-and-play module that provides LLM prompts to enable LLMs to perform zeroshot VQA tasks without end-to-end training. We develop LLM-agnostic models describe image content as exemplar question-answer pairs, which prove to be effective LLM prompts. Img2LLM offers the following benefits: 1) It achieves comparable or better performance than methods relying on end-to-end training. For example, we outperform Flamingo [3] by 5.6% on VQAv2. On the challenging A-OKVQA dataset, our method outperforms few-shot methods by as much as 20%. 2) It flexibly interfaces with a wide range of LLMs to perform VQA. 3) It eliminates the need to specialize LLMs using end-to-end finetuning and serve highly specialized LLMs to end users, thereby reducing cost. Code is available via the LAVIS [28] framework at https://github.com/salesforce/LAVIS/tree/main/projects/img2llm-vqa.","{'model': 'tldr@v2.0.0', 'text': 'Img2LLM is a plug-and-play module that provides LLM prompts to enable LLMs to perform zeroshot VQA tasks without end-to-end training, and develops LLM-agnostic models that describe image content as exemplar question-answer pairs, which prove to be effective LLm prompts.'}",
-Analysis of Relaxation Processes in a Multilevel System. A Many‐Shot Expansion Technique,R. Serauskas,"The relaxation of a gas of polyatomic molecules, excited to a particular energy, is analyzed in terms of a random walk through all relevant energy states to yield the various stable products. The quantum yield is taken as the probability of forming a given one of these products. This relaxation problem is solved by means of an asymptotic, natural expansion in terms of the events in the process which are physically the most significant. The first term in this many‐shot expansion is the strong collision result. For nearly strong collision processes and nonequilibrium systems this many‐shot expansion yields a rapidly converging series for an arbitrary ordered set of transition probabilities, and hence constitutes a computationally useful solution to the Pauli master equation for such a nonequilibrium system. The relaxation problem is analyzed in terms of some assumed sets of collisional transition probabilities, including a model that assumes statistical equilibration of vibrational energy in every collision...",,
-Chemical Kinetics as a Relaxation Process in a Multilevel System. II. General Formulation of the Many‐Shot Expansion,R. Serauskas,"For a complete understanding of rates in chemical kinetics it is necessary to analyze kinetic processes on the molecular level. For such an analysis one must know the change in the distribution of internal energy as the result of a collision process. It is suggested that this information can be obtained from an analysis of weak‐collision terms in a reacting system excited to a nonequilibrium distribution from which it relaxes by way of a many‐level process.The relaxation problem in a nonequilibrium multilevel system weakly coupled to a heat bath can be expressed in terms of a Pauli master equation. It has previously been shown that this relaxation problem can be solved in terms of a many‐shot expansion, the leading term being the strong‐collision result and further terms being due to weak collisions. In this paper it is shown that this many‐shot expansion can be derived from a generalized matrix form of the master equation, as well as by the phenomenological approach used previously. Furthermore, a matrix...",,
-Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?,Sewon Min,"Large language models (LMs) are able to in-context learn—perform a new task via inference alone by conditioning on a few input-label pairs (demonstrations) and making predictions for new inputs. However, there has been little understanding of how the model learns and which aspects of the demonstrations contribute to end task performance. In this paper, we show that ground truth demonstrations are in fact not required—randomly replacing labels in the demonstrations barely hurts performance on a range of classification and multi-choce tasks, consistently over 12 different models including GPT-3. Instead, we find that other aspects of the demonstrations are the key drivers of endtask performance, including the fact that they provide a few examples of (1) the label space, (2) the distribution of the input text, and (3) the overall format of the sequence. Together, our analysis provides a new way of understanding how and why in-context learning works, while opening up new questions about how much can be learned from large language models through inference alone.","{'model': 'tldr@v2.0.0', 'text': 'This paper shows that ground truth demonstrations are in fact not required and that other aspects of the demonstrations are the key drivers of end task performance, including the fact that they provide a few examples of the label space, the distribution of the input text, and the overall format of the sequence.'}",https://aclanthology.org/2022.emnlp-main.759.pdf
-What learning algorithm is in-context learning? Investigations with linear models,Ekin Akyürek,"Neural sequence models, especially transformers, exhibit a remarkable capacity for in-context learning. They can construct new predictors from sequences of labeled examples $(x, f(x))$ presented in the input without further parameter updates. We investigate the hypothesis that transformer-based in-context learners implement standard learning algorithms implicitly, by encoding smaller models in their activations, and updating these implicit models as new examples appear in the context. Using linear regression as a prototypical problem, we offer three sources of evidence for this hypothesis. First, we prove by construction that transformers can implement learning algorithms for linear models based on gradient descent and closed-form ridge regression. Second, we show that trained in-context learners closely match the predictors computed by gradient descent, ridge regression, and exact least-squares regression, transitioning between different predictors as transformer depth and dataset noise vary, and converging to Bayesian estimators for large widths and depths. Third, we present preliminary evidence that in-context learners share algorithmic features with these predictors: learners' late layers non-linearly encode weight vectors and moment matrices. These results suggest that in-context learning is understandable in algorithmic terms, and that (at least in the linear case) learners may rediscover standard estimation algorithms. Code and reference implementations are released at https://github.com/ekinakyurek/google-research/blob/master/incontext.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the hypothesis that transformer-based in-context learners implement standard learning algorithms implicitly, by encoding smaller models in their activations, and updating these implicit models as new examples appear in the context, and suggests that in- context learning is understandable in algorithmic terms, and that (at least in the linear case) learners may rediscover standard estimation algorithms.'}",http://arxiv.org/pdf/2211.15661
-In-context Learning and Induction Heads,Catherine Olsson,"“Induction heads” are attention heads that implement a simple algorithm to complete token sequences like [A][B] ... [A] → [B]. In this work, we present preliminary and indirect evidence for a hypothesis that induction heads might constitute the mechanism for the majority of all “incontext learning” in large transformer models (i.e. decreasing loss at increasing token indices). We find that induction heads develop at precisely the same point as a sudden sharp increase in incontext learning ability, visible as a bump in the training loss. We present six complementary lines of evidence, arguing that induction heads may be the mechanistic source of general in-context learning in transformer models of any size. For small attention-only models, we present strong, causal evidence; for larger models with MLPs, we present correlational evidence. We recommend reading this paper as an HTML article. As Transformer generative models continue to scale and gain increasing real world use, addressing their associated safety problems becomes increasingly important. Mechanistic interpretability – attempting to reverse engineer the detailed computations performed by the model – offers one possible avenue for addressing these safety issues. If we can understand the internal structures that cause Transformer models to produce the outputs they do, then we may be able to address current safety problems more systematically, as well as anticipating safety problems in future more powerful models. [1, 2, 3, 4, 5]","{'model': 'tldr@v2.0.0', 'text': 'It is found that induction heads develop at precisely the same point as a sudden sharp increase in incontext learning ability, visible as a bump in the training loss.'}",http://arxiv.org/pdf/2209.11895
-Data Distributional Properties Drive Emergent In-Context Learning in Transformers,Stephanie C. Y. Chan,"Large transformer-based models are able to perform in-context few-shot learning, without being explicitly trained for it. This observation raises the question: what aspects of the training regime lead to this emergent behavior? Here, we show that this behavior is driven by the distributions of the training data itself. In-context learning emerges when the training data exhibits particular distributional properties such as burstiness (items appear in clusters rather than being uniformly distributed over time) and having large numbers of rarely occurring classes. In-context learning also emerges more strongly when item meanings or interpretations are dynamic rather than fixed. These properties are exemplified by natural language, but are also inherent to naturalistic data in a wide range of other domains. They also depart significantly from the uniform, i.i.d. training distributions typically used for standard supervised learning. In our initial experiments, we found that in-context learning traded off against more conventional weight-based learning, and models were unable to achieve both simultaneously. However, our later experiments uncovered that the two modes of learning could co-exist in a single model when it was trained on data following a skewed Zipfian distribution -- another common property of naturalistic data, including language. In further experiments, we found that naturalistic data distributions were only able to elicit in-context learning in transformers, and not in recurrent models. In sum, our findings indicate how the transformer architecture works together with particular properties of the training data to drive the intriguing emergent in-context learning behaviour of large language models, and how future work might encourage both in-context and in-weights learning in domains beyond language.","{'model': 'tldr@v2.0.0', 'text': 'These findings indicate how the transformer architecture works together with particular properties of the training data to drive the intriguing emergent in- context learning behaviour of large language models, and how future work might encourage both in-context and in-weights learning in domains beyond language.'}",https://arxiv.org/pdf/2205.05055
-Thinking about GPT-3 In-Context Learning for Biomedical IE? Think Again,Bernal Jimenez Gutierrez,"The strong few-shot in-context learning capability of large pre-trained language models (PLMs) such as GPT-3 is highly appealing for application domains such as biomedicine, which feature high and diverse demands of language technologies but also high data annotation costs. In this paper, we present the first systematic and comprehensive study to compare the few-shot performance of GPT-3 in-context learning with fine-tuning smaller (i.e., BERT-sized) PLMs on two highly representative biomedical information extraction tasks, named entity recognition and relation extraction. We follow the true few-shot setting to avoid overestimating models' few-shot performance by model selection over a large validation set. We also optimize GPT-3's performance with known techniques such as contextual calibration and dynamic in-context example retrieval. However, our results show that GPT-3 still significantly underperforms compared to simply fine-tuning a smaller PLM. In addition, GPT-3 in-context learning also yields smaller gains in accuracy when more training data becomes available. Our in-depth analyses further reveal issues of the in-context learning setting that may be detrimental to information extraction tasks in general. Given the high cost of experimenting with GPT-3, we hope our study provides guidance for biomedical researchers and practitioners towards more promising directions such as fine-tuning small PLMs.","{'model': 'tldr@v2.0.0', 'text': 'This is the first systematic and comprehensive study to compare the few-shot performance of GPT-3 in-context learning with fine-tuning smaller (i.e., BERT-sized) PLMs on two highly representative biomedical information extraction tasks, named entity recognition and relation extraction.'}",https://arxiv.org/pdf/2203.08410
-Active Example Selection for In-Context Learning,Yiming Zhang,"With a handful of demonstration examples, large-scale language models demonstrate strong capability to perform various tasks by in-context learning from these examples, without any fine-tuning. We demonstrate that in-context learning performance can be highly unstable across samples of examples, indicating the idiosyncrasies of how language models acquire information. We formulate example selection for in-context learning as a sequential decision problem, and propose a reinforcement learning algorithm for identifying generalizable policies to select demonstration examples. For GPT-2, our learned policies demonstrate strong abilities of generalizing to unseen tasks in training, with a 5.8% improvement on average. Examples selected from our learned policies can even achieve a small improvement on GPT-3 Ada. However, the improvement diminishes on larger GPT-3 models, suggesting emerging capabilities of large language models.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that in-context learning performance can be highly unstable across samples of examples, indicating the idiosyncrasies of how language models acquire information.'}",http://arxiv.org/pdf/2211.04486
-An Explanation of In-context Learning as Implicit Bayesian Inference,Sang Michael Xie,"Large language models (LMs) such as GPT-3 have the surprising ability to do in-context learning, where the model learns to do a downstream task simply by conditioning on a prompt consisting of input-output examples. The LM learns from these examples without being explicitly pretrained to learn. Thus, it is unclear what enables in-context learning. In this paper, we study how in-context learning can emerge when pretraining documents have long-range coherence. Here, the LM must infer a latent document-level concept to generate coherent next tokens during pretraining. At test time, in-context learning occurs when the LM also infers a shared latent concept between examples in a prompt. We prove when this occurs despite a distribution mismatch between prompts and pretraining data in a setting where the pretraining distribution is a mixture of HMMs. In contrast to messy large-scale datasets used to train LMs capable of in-context learning, we generate a small-scale synthetic dataset (GINC) where Transformers and LSTMs both exhibit in-context learning. Beyond the theory, experiments on GINC exhibit large-scale real-world phenomena including improved in-context performance with model scaling (despite the same pretraining loss), sensitivity to example order, and instances where zero-shot is better than few-shot in-context learning.","{'model': 'tldr@v2.0.0', 'text': 'This paper studies how in-context learning can emerge when pretraining documents have long-range coherence, and proves when this occurs despite a distribution mismatch between prompts and pretraining data in a setting where the pretraining distribution is a mixture of HMMs.'}",
-A Survey on In-context Learning,Qingxiu Dong,"With the increasing ability of large language models (LLMs), in-context learning (ICL) has become a new paradigm for natural language processing (NLP), where LLMs make predictions only based on contexts augmented with a few examples. It has been a new trend to explore ICL to evaluate and extrapolate the ability of LLMs. In this paper, we aim to survey and summarize the progress and challenges of ICL. We first present a formal definition of ICL and clarify its correlation to related studies. Then, we organize and discuss advanced techniques, including training strategies, demonstration designing strategies, as well as related analysis. Finally, we discuss the challenges of ICL and provide potential directions for further research. We hope that our work can encourage more research on uncovering how ICL works and improving ICL.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a formal definition of ICL and clarify its correlation to related studies, and organizes and discusses advanced techniques, including training strategies, demonstration designing strategies, as well as related analysis.'}",
-Learning To Retrieve Prompts for In-Context Learning,Ohad Rubin,"In-context learning is a recent paradigm in natural language understanding, where a large pre-trained language model (LM) observes a test instance and a few training examples as its input, and directly decodes the output without any update to its parameters. However, performance has been shown to strongly depend on the selected training examples (termed prompts). In this work, we propose an efficient method for retrieving prompts for in-context learning using annotated data and an LM. Given an input-output pair, we estimate the probability of the output given the input and a candidate training example as the prompt, and label training examples as positive or negative based on this probability. We then train an efficient dense retriever from this data, which is used to retrieve training examples as prompts at test time. We evaluate our approach on three sequence-to-sequence tasks where language utterances are mapped to meaning representations, and find that it substantially outperforms prior work and multiple baselines across the board.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an efficient method for retrieving prompts for in-context learning using annotated data and an LM, and trains an efficient dense retriever from this data, which is used to retrieve training examples as prompts at test time.'}",https://aclanthology.org/2022.naacl-main.191.pdf
-Controllable Dialogue Simulation with In-Context Learning,Zekun Li,"Building dialogue systems requires a large corpus of annotated dialogues. Such datasets are usually created via crowdsourcing, which is expensive and time-consuming. In this paper, we propose \textsc{Dialogic}, a novel dialogue simulation method based on large language model in-context learning to automate dataset creation. Seeded with a few annotated dialogues, \textsc{Dialogic} automatically selects in-context examples for demonstration and prompts GPT-3 to generate new dialogues and annotations in a controllable way. Our method can rapidly expand a small set of dialogue data with minimum or zero \textit{human involvement} and \textit{parameter update} and is thus much more cost-efficient and time-saving than crowdsourcing. Experimental results on the MultiWOZ dataset demonstrate that training a model on the simulated dialogues leads to even better performance than using the same amount of human-generated dialogues under the challenging low-resource settings, with as few as 85 dialogues as a seed. When enough data is available, our method can still serve as an effective data augmentation method. Human evaluation results also show that our simulated dialogues have near-human fluency and annotation accuracy. The code and data are available at \textbf{\url{https://github.com/Leezekun/dialogic}}.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on the MultiWOZ dataset demonstrate that training a model on the simulated dialogues leads to even better performance than using the same amount of human-generated dialogues under the challenging low-resource settings, with as few as 85 dialogues as a seed.'}",http://arxiv.org/pdf/2210.04185
-Larger language models do in-context learning differently,Jerry W. Wei,"We study how in-context learning (ICL) in language models is affected by semantic priors versus input-label mappings. We investigate two setups-ICL with flipped labels and ICL with semantically-unrelated labels-across various model families (GPT-3, InstructGPT, Codex, PaLM, and Flan-PaLM). First, experiments on ICL with flipped labels show that overriding semantic priors is an emergent ability of model scale. While small language models ignore flipped labels presented in-context and thus rely primarily on semantic priors from pretraining, large models can override semantic priors when presented with in-context exemplars that contradict priors, despite the stronger semantic priors that larger models may hold. We next study semantically-unrelated label ICL (SUL-ICL), in which labels are semantically unrelated to their inputs (e.g., foo/bar instead of negative/positive), thereby forcing language models to learn the input-label mappings shown in in-context exemplars in order to perform the task. The ability to do SUL-ICL also emerges primarily with scale, and large-enough language models can even perform linear classification in a SUL-ICL setting. Finally, we evaluate instruction-tuned models and find that instruction tuning strengthens both the use of semantic priors and the capacity to learn input-label mappings, but more of the former.","{'model': 'tldr@v2.0.0', 'text': 'Evaluated instruction-tuned models are evaluated and it is found that instruction tuning strengthens both the use of semantic priors and the capacity to learn input-label mappings, but more of the former.'}",http://arxiv.org/pdf/2303.03846
-A Survey for In-context Learning,Qingxiu Dong,"With the increasing ability of large language models (LLMs), in-context learning (ICL) has become a new paradigm for natural language processing (NLP), where LLMs make predictions only based on contexts augmented with a few training examples. It has been a new trend exploring ICL to evaluate and extrapolate the ability of LLMs. In this paper, we aim to survey and summarize the progress, challenges, and future work in ICL. We first present a formal definition of ICL and clarify its correlation to related studies. Then, we organize and discuss advanced techniques of ICL, including training strategies, prompting strategies, and so on. Finally, we present the challenges of ICL and provide potential directions for further research. We hope our work can encourage more research on uncovering how ICL works and improving ICL in future work. 1","{'model': 'tldr@v2.0.0', 'text': 'The progress, challenges, and future work in ICL are summarized and a formal definition of ICL is presented and its correlation to related studies are clarified and potential directions for further research are provided.'}",http://arxiv.org/pdf/2301.00234
-Transformers as Algorithms: Generalization and Stability in In-context Learning,Yingcong Li,"In-context learning (ICL) is a type of prompting where a transformer model operates on a sequence of (input, output) examples and performs inference on-the-fly. In this work, we formalize in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time. We first explore the statistical aspects of this abstraction through the lens of multitask learning: We obtain generalization bounds for ICL when the input prompt is (1) a sequence of i.i.d. (input, label) pairs or (2) a trajectory arising from a dynamical system. The crux of our analysis is relating the excess risk to the stability of the algorithm implemented by the transformer. We characterize when transformer/attention architecture provably obeys the stability condition and also provide empirical verification. For generalization on unseen tasks, we identify an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner. Finally, we provide numerical evaluations that (1) demonstrate transformers can indeed implement near-optimal algorithms on classical regression problems with i.i.d. and dynamic data, (2) provide insights on stability, and (3) verify our theoretical predictions.","{'model': 'tldr@v2.0.0', 'text': 'This work formalizes in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time and identifies an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner.'}",
-Compositional Exemplars for In-context Learning,Jiacheng Ye,"Large pretrained language models (LMs) have shown impressive In-Context Learning (ICL) ability, where the model learns to do an unseen task via a prompt consisting of input-output examples as the demonstration, without any parameter updates. The performance of ICL is highly dominated by the quality of the selected in-context examples. However, previous selection methods are mostly based on simple heuristics, leading to sub-optimal performance. In this work, we formulate in-context example selection as a subset selection problem. We propose CEIL (Compositional Exemplars for In-context Learning), which is instantiated by Determinantal Point Processes (DPPs) to model the interaction between the given input and in-context examples, and optimized through a carefully-designed contrastive learning objective to obtain preference from LMs. We validate CEIL on 12 classification and generation datasets from 7 distinct NLP tasks, including sentiment analysis, paraphrase detection, natural language inference, commonsense reasoning, open-domain question answering, code generation, and semantic parsing. Extensive experiments demonstrate not only the state-of-the-art performance but also the transferability and compositionality of CEIL, shedding new light on effective and efficient in-context learning. Our code is released at https://github.com/HKUNLP/icl-ceil.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes CEIL (Compositional Exemplars for In-context Learning), which is instantiated by Determinantal Point Processes to model the interaction between the given input and in-context examples, and optimized through a carefully-designed contrastive learning objective to obtain preference from LMs.'}",http://arxiv.org/pdf/2302.05698
-The Learnability of In-Context Learning,Noam Wies,"In-context learning is a surprising and important phenomenon that emerged when modern language models were scaled to billions of learned parameters. Without modifying a large language model's weights, it can be tuned to perform various downstream natural language tasks simply by including concatenated training examples of these tasks in its input. Though disruptive for many practical applications of large language models, this emergent learning paradigm is not well understood from a theoretical perspective. In this paper, we propose a first-of-its-kind PAC based framework for in-context learnability, and use it to provide the first finite sample complexity results for the in-context learning setup. Our framework includes an initial pretraining phase, which fits a function to the pretraining distribution, and then a second in-context learning phase, which keeps this function constant and concatenates training examples of the downstream task in its input. We use our framework in order to prove that, under mild assumptions, when the pretraining distribution is a mixture of latent tasks (a model often considered for natural language pretraining), these tasks can be efficiently learned via in-context learning, even though the model's weights are unchanged and the input significantly diverges from the pretraining distribution. Our theoretical analysis reveals that in this setting, in-context learning is more about identifying the task than about learning it, a result which is in line with a series of recent empirical findings. We hope that the in-context learnability framework presented in this paper will facilitate future progress towards a deeper understanding of this important new learning paradigm.","{'model': 'tldr@v2.0.0', 'text': ""A first-of-its-kind PAC based framework for in-context learnability is proposed, and it is proved that, under mild assumptions, when the pretraining distribution is a mixture of latent tasks, these tasks can be efficiently learned via in- context learning, even though the model's weights are unchanged and the input significantly diverges from the Pretraining distribution.""}",http://arxiv.org/pdf/2303.07895
-Differentially Private In-Context Learning,Ashwinee Panda,An important question in deploying large language models (LLMs) is how to augment LLMs with private data. We propose Differentially Private In-context Learning (DP-ICL) to enable LLMs to adapt to new tasks while maintaining privacy guarantees. DP-ICL performs private inference by establishing a noisy consensus over an ensemble of exemplars using the Report-Noisy-Max mechanism. We evaluate DP-ICL on four benchmarks and find that it achieves comparable performance (< 2% degradation) with non-private ICL.,"{'model': 'tldr@v2.0.0', 'text': 'Differentially Private In-context Learning (DP-ICL) is proposed to enable LLMs to adapt to new tasks while maintaining privacy guarantees, and achieves comparable performance (< 2% degradation) with non-private ICL.'}",https://arxiv.org/pdf/2305.01639
-Generative Calibration for In-context Learning,Zhongtao Jiang,"As one of the most exciting features of large language models (LLMs), in-context learning is a mixed blessing. While it allows users to fast-prototype a task solver with only a few training examples, the performance is generally sensitive to various configurations of the prompt such as the choice or order of the training examples. In this paper, we for the first time theoretically and empirically identify that such a paradox is mainly due to the label shift of the in-context model to the data distribution, in which LLMs shift the label marginal $p(y)$ while having a good label conditional $p(x|y)$. With this understanding, we can simply calibrate the in-context predictive distribution by adjusting the label marginal, which is estimated via Monte-Carlo sampling over the in-context model, i.e., generation of LLMs. We call our approach as generative calibration. We conduct exhaustive experiments with 12 text classification tasks and 12 LLMs scaling from 774M to 33B, generally find that the proposed method greatly and consistently outperforms the ICL as well as state-of-the-art calibration methods, by up to 27% absolute in macro-F1. Meanwhile, the proposed method is also stable under different prompt configurations.","{'model': 'tldr@v2.0.0', 'text': 'The proposed method greatly and consistently outperforms the ICL as well as state-of-the-art calibration methods, by up to 27% absolute in macro-F1, and the proposed method is also stable under different prompt configurations.'}",
-Distinguishability Calibration to In-Context Learning,Hongjing Li,"Recent years have witnessed increasing interests in prompt-based learning in which models can be trained on only a few annotated instances, making them suitable in low-resource settings. It is even challenging in fine-grained classification as the pre-trained language models tend to generate similar output embedding which makes it difficult to discriminate for the prompt-based classifier. In this work, we alleviate this information diffusion issue by proposing a calibration method based on a transformation which rotates the embedding feature into a new metric space where we adapt the ratio of each dimension to a uniform distribution to guarantee the distinguishability of learned embeddings. Furthermore, we take the advantage of hyperbolic embedding to capture the relation between dimensions by a coarse-fine metric learning strategy to enhance interpretability. Extensive experiments on the three datasets under various settings demonstrate the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'A calibration method based on a transformation which rotates the embedding feature into a new metric space where the ratio of each dimension to a uniform distribution is adapted to guarantee the distinguishability of learned embeddings to enhance interpretability.'}",http://arxiv.org/pdf/2302.06198
-MetaICL: Learning to Learn In Context,Sewon Min,"We introduce MetaICL (Meta-training for In-Context Learning), a new meta-training framework for few-shot learning where a pretrained language model is tuned to do in-context learning on a large set of training tasks. This meta-training enables the model to more effectively learn a new task in context at test time, by simply conditioning on a few training examples with no parameter updates or task-specific templates. We experiment on a large, diverse collection of tasks consisting of 142 NLP datasets including classification, question answering, natural language inference, paraphrase detection and more, across seven different meta-training/target splits. MetaICL outperforms a range of baselines including in-context learning without meta-training and multi-task learning followed by zero-shot transfer. We find that the gains are particularly significant for target tasks that have domain shifts from the meta-training tasks, and that using a diverse set of the meta-training tasks is key to improvements. We also show that MetaICL approaches (and sometimes beats) the performance of models fully finetuned on the target task training data, and outperforms much bigger models with nearly 8x parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces MetaICL (Meta-training for In-Context Learning), a new meta-training framework for few-shot learning where a pretrained language model is tuned to do in-context learning on a large set of training tasks.'}",https://aclanthology.org/2022.naacl-main.201.pdf
-In-context Reinforcement Learning with Algorithm Distillation,M. Laskin,"We propose Algorithm Distillation (AD), a method for distilling reinforcement learning (RL) algorithms into neural networks by modeling their training histories with a causal sequence model. Algorithm Distillation treats learning to reinforcement learn as an across-episode sequential prediction problem. A dataset of learning histories is generated by a source RL algorithm, and then a causal transformer is trained by autoregressively predicting actions given their preceding learning histories as context. Unlike sequential policy prediction architectures that distill post-learning or expert sequences, AD is able to improve its policy entirely in-context without updating its network parameters. We demonstrate that AD can reinforcement learn in-context in a variety of environments with sparse rewards, combinatorial task structure, and pixel-based observations, and find that AD learns a more data-efficient RL algorithm than the one that generated the source data.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that AD can reinforcement learn in-context in a variety of environments with sparse rewards, combinatorial task structure, and pixel-based observations, and it is found that AD learns a more data-efficient RL algorithm than the one that generated the source data.'}",http://arxiv.org/pdf/2210.14215
-Context Encoders: Feature Learning by Inpainting,Deepak Pathak,"We present an unsupervised visual feature learning algorithm driven by context-based pixel prediction. By analogy with auto-encoders, we propose Context Encoders - a convolutional neural network trained to generate the contents of an arbitrary image region conditioned on its surroundings. In order to succeed at this task, context encoders need to both understand the content of the entire image, as well as produce a plausible hypothesis for the missing part(s). When training context encoders, we have experimented with both a standard pixel-wise reconstruction loss, as well as a reconstruction plus an adversarial loss. The latter produces much sharper results because it can better handle multiple modes in the output. We found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures. We quantitatively demonstrate the effectiveness of our learned features for CNN pre-training on classification, detection, and segmentation tasks. Furthermore, context encoders can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.","{'model': 'tldr@v2.0.0', 'text': 'It is found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures, and can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.'}",https://arxiv.org/pdf/1604.07379
-Context Autoencoder for Self-Supervised Representation Learning,Xiaokang Chen,,"{'model': 'tldr@v2.0.0', 'text': 'The CAE design encourages the separation of learning the encoder (representation) from completing the pertaining tasks: masked representation prediction and masked patch reconstruction tasks, and making predictions in the encoded representation space empirically shows the benefit to representation learning.'}",https://arxiv.org/pdf/2202.03026
-Fast Visual Tracking via Dense Spatio-temporal Context Learning,Kaihua Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'A novel explicit scale adaptation scheme is proposed, able to deal with target scale variations efficiently and effectively, and the Fast Fourier Transform is adopted for fast learning and detection in this work, which only needs 4 FFT operations.'}",http://www4.comp.polyu.edu.hk/~cslzhang/paper/conf/STC_eccv14.pdf
-Can You Unpack That? Learning to Rewrite Questions-in-Context,Ahmed Elgohary,"Question answering is an AI-complete problem, but existing datasets lack key elements of language understanding such as coreference and ellipsis resolution. We consider sequential question answering: multiple questions are asked one-by-one in a conversation between a questioner and an answerer. Answering these questions is only possible through understanding the conversation history. We introduce the task of question-in-context rewriting: given the context of a conversation’s history, rewrite a context-dependent into a self-contained question with the same answer. We construct, CANARD, a dataset of 40,527 questions based on QuAC (Choi et al., 2018) and train Seq2Seq models for incorporating context into standalone questions.","{'model': 'tldr@v2.0.0', 'text': 'This work constructs, CANARD, a dataset of 40,527 questions based on QuAC and trains Seq2Seq models for incorporating context into standalone questions and introduces the task of question-in-context rewriting.'}",https://www.aclweb.org/anthology/D19-1605.pdf
-What Can Transformers Learn In-Context? A Case Study of Simple Function Classes,Shivam Garg,"In-context learning refers to the ability of a model to condition on a prompt sequence consisting of in-context examples (input-output pairs corresponding to some task) along with a new query input, and generate the corresponding output. Crucially, in-context learning happens only at inference time without any parameter updates to the model. While large language models such as GPT-3 exhibit some ability to perform in-context learning, it is unclear what the relationship is between tasks on which this succeeds and what is present in the training data. To make progress towards understanding in-context learning, we consider the well-defined problem of training a model to in-context learn a function class (e.g., linear functions): that is, given data derived from some functions in the class, can we train a model to in-context learn""most""functions from this class? We show empirically that standard Transformers can be trained from scratch to perform in-context learning of linear functions -- that is, the trained model is able to learn unseen linear functions from in-context examples with performance comparable to the optimal least squares estimator. In fact, in-context learning is possible even under two forms of distribution shift: (i) between the training data of the model and inference-time prompts, and (ii) between the in-context examples and the query input during inference. We also show that we can train Transformers to in-context learn more complex function classes -- namely sparse linear functions, two-layer neural networks, and decision trees -- with performance that matches or exceeds task-specific learning algorithms. Our code and models are available at https://github.com/dtsip/in-context-learning .","{'model': 'tldr@v2.0.0', 'text': 'It is shown empirically that standard Transformers can be trained from scratch to perform in-context learning of linear functions -- that is, the trained model is able to learn unseen linear functions from in- context examples with performance comparable to the optimal least squares estimator.'}",https://arxiv.org/pdf/2208.01066
-Transformers learn in-context by gradient descent,J. Oswald,"At present, the mechanisms of in-context learning in Transformers are not well understood and remain mostly an intuition. In this paper, we suggest that training Transformers on auto-regressive objectives is closely related to gradient-based meta-learning formulations. We start by providing a simple weight construction that shows the equivalence of data transformations induced by 1) a single linear self-attention layer and by 2) gradient-descent (GD) on a regression loss. Motivated by that construction, we show empirically that when training self-attention-only Transformers on simple regression tasks either the models learned by GD and Transformers show great similarity or, remarkably, the weights found by optimization match the construction. Thus we show how trained Transformers become mesa-optimizers i.e. learn models by gradient descent in their forward pass. This allows us, at least in the domain of regression problems, to mechanistically understand the inner workings of in-context learning in optimized Transformers. Building on this insight, we furthermore identify how Transformers surpass the performance of plain gradient descent by learning an iterative curvature correction and learn linear models on deep data representations to solve non-linear regression tasks. Finally, we discuss intriguing parallels to a mechanism identified to be crucial for in-context learning termed induction-head (Olsson et al., 2022) and show how it could be understood as a specific case of in-context learning by gradient descent learning within Transformers. Code to reproduce the experiments can be found at https://github.com/google-research/self-organising-systems/tree/master/transformers_learn_icl_by_gd .","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that training Transformers on auto-regressive objectives is closely related to gradient-based meta-learning formulations and how Transformers surpass the performance of plain gradient descent by learning an iterative curvature correction and learn linear models on deep data representations to solve non-linear regression tasks.'}",http://arxiv.org/pdf/2212.07677
-A Survey on Context Learning,Guangxu Xun,"Learning semantics based on context information has been researched in many research areas for decades. Context information can not only be directly used as the input data, but also sometimes used as auxiliary knowledge to improve existing models. This survey aims at providing a structured and comprehensive overview of the research on context learning. We summarize and group the existing literature into four categories, Explicit Analysis, Implicit Analysis, Neural Network Models, and Composite Models, based on the underlying techniques adopted by them. For each category, we talk about the basic idea and techniques, and also introduce how context information is utilized as the model input or incorporated into the model to enhance the performance or extend the domain of application as auxiliary knowledge. In addition, we discuss the advantages and disadvantages of each model from both the technical and practical point of view.","{'model': 'tldr@v2.0.0', 'text': 'This survey aims at providing a structured and comprehensive overview of the research on context learning by summarized and group the existing literature into four categories, Explicit Analysis, Implicit Analysis, Neural Network Models, and Composite Models, based on the underlying techniques adopted by them.'}",
-Cross-Modal Retrieval in the Cooking Context: Learning Semantic Text-Image Embeddings,Micael Carvalho,"Designing powerful tools that support cooking activities has rapidly gained popularity due to the massive amounts of available data, as well as recent advances in machine learning that are capable of analyzing them. In this paper, we propose a cross-modal retrieval model aligning visual and textual data (like pictures of dishes and their recipes) in a shared representation space. We describe an effective learning scheme, capable of tackling large-scale problems, and validate it on the Recipe1M dataset containing nearly 1 million picture-recipe pairs. We show the effectiveness of our approach regarding previous state-of-the-art models and present qualitative results over computational cooking use cases.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a cross-modal retrieval model aligning visual and textual data (like pictures of dishes and their recipes) in a shared representation space, and describes an effective learning scheme, capable of tackling large-scale problems.'}",https://hal.archives-ouvertes.fr/hal-01839068/file/1804.11146.pdf
-What Makes Good In-Context Examples for GPT-3?,Jiachang Liu,"GPT-3 has attracted lots of attention due to its superior performance across a wide range of NLP tasks, especially with its in-context learning abilities. Despite its success, we found that the empirical results of GPT-3 depend heavily on the choice of in-context examples. In this work, we investigate whether there are more effective strategies for judiciously selecting in-context examples (relative to random sampling) that better leverage GPT-3’s in-context learning capabilities.Inspired by the recent success of leveraging a retrieval module to augment neural networks, we propose to retrieve examples that are semantically-similar to a test query sample to formulate its corresponding prompt. Intuitively, the examples selected with such a strategy may serve as more informative inputs to unleash GPT-3’s power of text generation. We evaluate the proposed approach on several natural language understanding and generation benchmarks, where the retrieval-based prompt selection approach consistently outperforms the random selection baseline. Moreover, it is observed that the sentence encoders fine-tuned on task-related datasets yield even more helpful retrieval results. Notably, significant gains are observed on tasks such as table-to-text generation (44.3% on the ToTTo dataset) and open-domain question answering (45.5% on the NQ dataset).","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to retrieve examples that are semantically-similar to a test query sample to formulate its corresponding prompt, and evaluates the proposed approach on several natural language understanding and generation benchmarks, where the retrieval-based prompt selection approach consistently outperforms the random selection baseline.'}",https://aclanthology.org/2022.deelio-1.10.pdf
-"The neuroscience of placebo effects: connecting context, learning and health",T. Wager,,"{'model': 'tldr@v2.0.0', 'text': 'An empirical review of the brain systems that are involved in placebo effects and a conceptual framework linking these findings to the mind–brain processes that mediate them suggest that the neuropsychological processes thatMediate placebo effects may be crucial for a wide array of therapeutic approaches, including many drugs.'}",https://europepmc.org/articles/pmc6013051?pdf=render
-Adult Learning Theories in Context: A Quick Guide for Healthcare Professional Educators,B. Mukhalalati,"Background: Adult learning theories play a pivotal role in the design and implementation of education programs, including healthcare professional programs. There is a variation in the use of theories in healthcare professional education programs and this is may be in part due to a lack of understanding of the range of learning theories available and paucity of specific, in-context examples, to help educators in considering alternative theories relevant to their teaching setting. This article seeks to synthesize key learning theories applicable in the learning and teaching of healthcare professionals and to provide examples of their use in context. Method and results: A literature review was conducted in 2015 and 2016 using PubMed, Scopus, Web of Science, and ERIC academic databases. Search terms used identified a range of relevant literature about learning theories, and their utilization in different healthcare professional education programs. The findings were synthesized and presented in a table format, illustrating the learning theory, specific examples from health and medical education, and a very brief critique of the theory. Outcome: The literature synthesis provides a quick and easy-to-use summary of key theories and examples of their use to help healthcare professional educators access a wider range of learning theories to inform their instructional strategies, learning objectives, and evaluation approaches. This will ultimately result in educational program enhancement and improvement in student learning experiences.","{'model': 'tldr@v2.0.0', 'text': 'The literature synthesis provides a quick and easy-to-use summary of key theories and examples of their use to help healthcare professional educators access a wider range of learning theories to inform their instructional strategies, learning objectives, and evaluation approaches.'}",https://journals.sagepub.com/doi/pdf/10.1177/2382120519840332
-Professional development for cultural diversity: the challenges of teacher learning in context,Nikolett Szelei,"ABSTRACT This article describes context-based professional development (PD) for cultural diversity in a Portuguese school cluster, and discusses how it supports change for justice and equity. Teachers felt the importance of PD and showed willingness and interest to learn. Several teacher learning opportunities were mapped out such as formal workshops, starting small collaborations and teachers’ self-directed informal learning activities. Yet, a rather fragmented character of PD seemed to emerge in terms of content on cultural diversity and forms of learning. Conflicting agendas, scattered teacher collaboration and commitment, and little student and community involvement in planned PD were found. Furthermore, there seemed to be tensions between current PD and teachers’ needs and circumstances; teachers wished for more specific information and pedagogical solutions, more collaboration and more organisational support in PD. Applying a critical multicultural perspective, it is discussed that although the current constellation of PD is a potential start, it might still contribute to teachers’ conceptual confusion and pedagogical insecurities on the field of cultural diversity. It is suggested that criticality towards PD frames is needed to re-centre cultural diversity on the premises of justice, as well as teacher support, and conscious learning with and from students, families and communities.",,https://www.tandfonline.com/doi/pdf/10.1080/19415257.2019.1642233?needAccess=true
-English in Context: Learning Materials,P. L. McEldowney,,,
-Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables,Kate Rakelly,"Deep reinforcement learning algorithms require large amounts of experience to learn an individual task. While in principle meta-reinforcement learning (meta-RL) algorithms enable agents to learn new skills from small amounts of experience, several major challenges preclude their practicality. Current methods rely heavily on on-policy experience, limiting their sample efficiency. The also lack mechanisms to reason about task uncertainty when adapting to new tasks, limiting their effectiveness in sparse reward problems. In this paper, we address these challenges by developing an off-policy meta-RL algorithm that disentangles task inference and control. In our approach, we perform online probabilistic filtering of latent task variables to infer how to solve a new task from small amounts of experience. This probabilistic interpretation enables posterior sampling for structured and efficient exploration. We demonstrate how to integrate these task variables with off-policy RL algorithms to achieve both meta-training and adaptation efficiency. Our method outperforms prior algorithms in sample efficiency by 20-100X as well as in asymptotic performance on several meta-RL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops an off-policy meta-RL algorithm that disentangles task inference and control and performs online probabilistic filtering of latent task variables to infer how to solve a new task from small amounts of experience.'}",
-Rural online learning in the context of COVID 19 in South Africa: Evoking an inclusive education approach,B. Dube,"This paper discusses the challenges faced by rural learners in South Africa in the context of the world pandemic commonly known as COVID-19. Rural learners face unprecedented challenges in adjusting to a new mode of life and learning, the latter being characterised by the predominant use of online, learning management systems and low-tech applications. The paper is informed by critical emancipatory research, I used participatory action research. A total of 10 learners and five teachers participated via Whatsapp. The paper answers two questions: what are the learning challenges faced by rural learners in South Africa, and how can online learning be enhanced in the context of COVID-19? The findings suggest that, while the South African government is promoting online learning as the only alternative in the context of COVID-19, this mode excludes many rural learners from teaching and learning, due to a lack of resources to connect to the internet, the learning management system, and low-tech software. The paper argues that rural learners are critical stakeholders in education and in the fight against COVID-19, and they cannot be left behind in efforts to fight the pandemic. ","{'model': 'tldr@v2.0.0', 'text': 'The findings suggest that, while the South African government is promoting online learning as the only alternative in the context of COVID-19, this mode excludes many rural learners from teaching and learning, due to a lack of resources to connect to the internet, the learning management system and low-tech software.'}",
-Language-Agnostic Representation Learning of Source Code from Structure and Context,D. Zugner,"Source code (Context) and its parsed abstract syntax tree (AST; Structure) are two complementary representations of the same computer program. Traditionally, designers of machine learning models have relied predominantly either on Structure or Context. We propose a new model, which jointly learns on Context and Structure of source code. In contrast to previous approaches, our model uses only language-agnostic features, i.e., source code and features that can be computed directly from the AST. Besides obtaining state-of-the-art on monolingual code summarization on all five programming languages considered in this work, we propose the first multilingual code summarization model. We show that jointly training on non-parallel data from multiple programming languages improves results on all individual languages, where the strongest gains are on low-resource languages. Remarkably, multilingual training only from Context does not lead to the same improvements, highlighting the benefits of combining Structure and Context for representation learning on code.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new model, which jointly learns on Context and Structure of source code, and shows that jointly training on non-parallel data from multiple programming languages improves results on all individual languages, where the strongest gains are on low-resource languages.'}",
-Multi-Task Reinforcement Learning with Context-based Representations,Shagun Sodhani,"The benefit of multi-task learning over single-task learning relies on the ability to use relations across tasks to improve performance on any single task. While sharing representations is an important mechanism to share information across tasks, its success depends on how well the structure underlying the tasks is captured. In some real-world situations, we have access to metadata, or additional information about a task, that may not provide any new insight in the context of a single task setup alone but inform relations across multiple tasks. While this metadata can be useful for improving multi-task learning performance, effectively incorporating it can be an additional challenge. We posit that an efficient approach to knowledge transfer is through the use of multiple context-dependent, composable representations shared across a family of tasks. In this framework, metadata can help to learn interpretable representations and provide the context to inform which representations to compose and how to compose them. We use the proposed approach to obtain state-of-the-art results in Meta-World, a challenging multi-task benchmark consisting of 50 distinct robotic manipulation tasks.","{'model': 'tldr@v2.0.0', 'text': 'It is proposed that an efficient approach to knowledge transfer is through the use of multiple context-dependent, composable representations shared across a family of tasks, and metadata can help to learn interpretable representations and provide the context to inform which representations to compose and how to compose them.'}",
-Learning Pyramid-Context Encoder Network for High-Quality Image Inpainting,Yanhong Zeng,"High-quality image inpainting requires filling missing regions in a damaged image with plausible content. Existing works either fill the regions by copying high-resolution patches or generating semantically-coherent patches from region context, while neglecting the fact that both visual and semantic plausibility are highly-demanded. In this paper, we propose a Pyramid-context Encoder Network (denoted as PEN-Net) for image inpainting by deep generative models. The proposed PEN-Net is built upon a U-Net structure with three tailored components, ie., a pyramid-context encoder, a multi-scale decoder, and an adversarial training loss. First, we adopt a U-Net as backbone which can encode the context of an image from high-resolution pixels into high-level semantic features, and decode the features reversely. Second, we propose a pyramid-context encoder, which progressively learns region affinity by attention from a high-level semantic feature map, and transfers the learned attention to its adjacent high-resolution feature map. As the missing content can be filled by attention transfer from deep to shallow in a pyramid fashion, both visual and semantic coherence for image inpainting can be ensured. Third, we further propose a multi-scale decoder with deeply-supervised pyramid losses and an adversarial loss. Such a design not only results in fast convergence in training, but more realistic results in testing. Extensive experiments on a broad range of datasets shows the superior performance of the proposed network.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Pyramid-context Encoder Network for image inpainting by deep generative models, built upon a U-Net structure with three tailored components, ie.'}",https://arxiv.org/pdf/1904.07475
-Neuromorphic Context-Dependent Learning Framework With Fault-Tolerant Spike Routing,Shuangming Yang,"Neuromorphic computing is a promising technology that realizes computation based on event-based spiking neural networks (SNNs). However, fault-tolerant on-chip learning remains a challenge in neuromorphic systems. This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework. We show how this system can learn associations between stimulation and response in two context-dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes. Furthermore, we demonstrate how our novel fault-tolerant neuromorphic spike routing scheme can avoid multiple fault nodes successfully and can enhance the maximum throughput of the neuromorphic network by 0.9%–16.1% in comparison with previous studies. By utilizing the real-time computational capabilities and multiple-fault-tolerant property of the proposed system, the neuronal mechanisms underlying the spiking activities of neuromorphic networks can be readily explored. In addition, the proposed system can be applied in real-time learning and decision-making applications, brain–machine integration, and the investigation of brain cognition during learning.","{'model': 'tldr@v2.0.0', 'text': 'This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework, and shows how this system can learn associations between stimulation and response in two context- dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes.'}",https://researchonline.jcu.edu.au/68691/7/JCU_68691_AAM.pdf
-Fast Tracking via Spatio-Temporal Context Learning,Kaihua Zhang,"In this paper, we present a simple yet fast and robust algorithm which exploits the spatio-temporal context for visual tracking. Our approach formulates the spatio-temporal relationships between the object of interest and its local context based on a Bayesian framework, which models the statistical correlation between the low-level features (i.e., image intensity and position) from the target and its surrounding regions. The tracking problem is posed by computing a confidence map, and obtaining the best target location by maximizing an object location likelihood function. The Fast Fourier Transform is adopted for fast learning and detection in this work. Implemented in MATLAB without code optimization, the proposed tracker runs at 350 frames per second on an i7 machine. Extensive experimental results show that the proposed algorithm performs favorably against state-of-the-art methods in terms of efficiency, accuracy and robustness.","{'model': 'tldr@v2.0.0', 'text': 'This approach formulates the spatio-temporal relationships between the object of interest and its local context based on a Bayesian framework, which models the statistical correlation between the low-level features from the target and its surrounding regions.'}",
-Drivers of human development: How relationships and context shape learning and development1,D. Osher,"ABSTRACT This article synthesizes knowledge on the role of relationships and key macroand micro-contexts - poverty, racism, families, communities, schools, and peers - in supporting and/or undermining the healthy development of children and youth, using a relational developmental systems framework. Relationships with parents, siblings, peers, caregivers, and teachers are explored in the context of early care and childhood settings, schools, classrooms, and school-based interventions. Additional contextual factors include; chronic stress, institutionalized racism, stereotype threat, and racial identity. A companion article focuses on how the human brain develops, and the major constructs that define human development, the constructive nature of development, and the opportunities for resilience. Human development occurs through reciprocal coactions between the individual and their contexts and culture, with relationships as the key drivers. Relationships and contexts, along with how children appraise and interpret them, can be risks and assets for healthy learning and development, and their influence can be seen across generations and can produce intra- as well as intergenerational assets and risks. This knowledge about the individual’s responsiveness to context and experience has both positive and negative implications across early childhood, adolescence and into adulthood. Sensitive periods for brain growth and development are considered within the contextual factors that influence development including; parental responsiveness and attunement, intentional skill development, mindfulness, reciprocal interactions, adversity, trauma, and enriching opportunities. The accumulated knowledge on human development and the power of context and culture can inform child-serving systems that support positive adaptations, resilience, learning, health, and well-being.",,https://www.tandfonline.com/doi/pdf/10.1080/10888691.2017.1398650?needAccess=true
-Unsupervised Visual Representation Learning by Context Prediction,Carl Doersch,"This work explores the use of spatial context as a source of free and plentiful supervisory signal for training a rich visual representation. Given only a large, unlabeled image collection, we extract random pairs of patches from each image and train a convolutional neural net to predict the position of the second patch relative to the first. We argue that doing well on this task requires the model to learn to recognize objects and their parts. We demonstrate that the feature representation learned using this within-image context indeed captures visual similarity across images. For example, this representation allows us to perform unsupervised visual discovery of objects like cats, people, and even birds from the Pascal VOC 2011 detection dataset. Furthermore, we show that the learned ConvNet can be used in the R-CNN framework [19] and provides a significant boost over a randomly-initialized ConvNet, resulting in state-of-the-art performance among algorithms which use only Pascal-provided training set annotations.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the feature representation learned using this within-image context indeed captures visual similarity across images and allows us to perform unsupervised visual discovery of objects like cats, people, and even birds from the Pascal VOC 2011 detection dataset.'}",http://arxiv.org/pdf/1505.05192
-Multiscale Graph Sample and Aggregate Network With Context-Aware Learning for Hyperspectral Image Classification,Yao Ding,"Recently, graph convolutional network (GCN) has achieved promising results in hyperspectral image (HSI) classification. However, GCN is a transductive learning method, which is difficult to aggregate the new node. Besides, the existing GCN-based methods divide graph construction and graph classification into two stages ignoring the influence of constructed graph error on classification results. Moreover, the available GCN-based methods fail to understand the global and contextual information of the graph. In this article, we propose a novel multiscale graph sample and aggregate network with a context-aware learning method for HSI classification. The proposed network adopts a multiscale graph sample and aggregate network (graphSAGE) to learn the multiscale features from the local regions graph, which improves the diversity of network input information and effectively solves the impact of original input graph errors on classification. By employing a context-aware mechanism to characterize the importance among spatially neighboring regions, deep contextual and global information of the graph can be learned automatically by focusing on important spatial targets. Meanwhile, the graph structure is reconstructed automatically based on the classified objects as network training, which is able to effectively reduce the influence of the initial graph error on the classification result. Extensive experiments are conducted on three real HSI datasets, which are demonstrated to outperform the compared state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel multiscale graph sample and aggregate network with a context-aware learning method for HSI classification that improves the diversity of network input information and effectively solves the impact of original input graph errors on classification.'}",https://ieeexplore.ieee.org/ielx7/4609443/9314330/09411656.pdf
-End-to-End Learning of Geometry and Context for Deep Stereo Regression,Alex Kendall,"We propose a novel deep learning architecture for regressing disparity from a rectified pair of stereo images. We leverage knowledge of the problem’s geometry to form a cost volume using deep feature representations. We learn to incorporate contextual information using 3-D convolutions over this volume. Disparity values are regressed from the cost volume using a proposed differentiable soft argmin operation, which allows us to train our method end-to-end to sub-pixel accuracy without any additional post-processing or regularization. We evaluate our method on the Scene Flow and KITTI datasets and on KITTI we set a new stateof-the-art benchmark, while being significantly faster than competing approaches.","{'model': 'tldr@v2.0.0', 'text': 'A novel deep learning architecture for regressing disparity from a rectified pair of stereo images is proposed, leveraging knowledge of the problem’s geometry to form a cost volume using deep feature representations and incorporating contextual information using 3-D convolutions over this volume.'}",http://arxiv.org/pdf/1703.04309
-"Interprofessional education: a review of context, learning and the research agenda",J. Thistlethwaite,Medical Education 2012: 46: 58–70,,
-How E-Learning Environmental Stimuli Influence Determinates of Learning Engagement in the Context of COVID-19? SOR Model Perspective,Junhui Yang,"The COVID-19 pandemic at the beginning of 2020 has changed the conventional learning mode for most students at schools all over the world, and the e-learning at home has become a new trend. Taking Chinese college students as the research subject and drawing on the stimulus–organism–response (S-O-R) model, this paper examines the relationship between the peer referent, perceived closeness, and perceived control and the learning engagement. Using data from 377 college students who have used e-learning, this study shows that perceived closeness, perceived control, and peer referents in e-learning have a positive effect on the self-efficacy and well-being of students, thus improving students’ enthusiasm for learning. Our intent is to assist researchers, instructors, designers, and others in identifying effective methods to conceptualize and measure student engagement in e-learning.","{'model': 'tldr@v2.0.0', 'text': 'This study shows that perceived closeness, perceived control, and peer referents in e-learning have a positive effect on the self-efficacy and well-being of students, thus improving students’ enthusiasm for learning.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2021.584976/pdf
-Involvement of Dopamine D1/D5 and D2 Receptors in Context-Dependent Extinction Learning and Memory Reinstatement,M. André,"Dopamine contributes to the regulation of higher order information processing and executive control. It is important for memory consolidation processes, and for the adaptation of learned responses based on experience. In line with this, under aversive learning conditions, application of dopamine receptor antagonists prior to extinction result in enhanced memory reinstatement. Here, we investigated the contribution of the dopaminergic system to extinction and memory reinstatement (renewal) of an appetitive spatial learning task in rodents. Rats were trained for 3 days in a T-maze (context “A”) to associate a goal arm with a food reward, despite low reward probability (acquisition phase). On day 4, extinction learning (unrewarded) occurred, that was reinforced by a context change (“B”). On day 5, re-exposure to the (unrewarded) “A” context took place (renewal of context “A”, followed by extinction of context “A”). In control animals, significant extinction occurred on day 4, that was followed by an initial memory reinstatement (renewal) on day 5, that was, in turn, succeeded by extinction of renewal. Intracerebral treatment with a D1/D5-receptor antagonist prior to the extinction trials, elicited a potent enhancement of extinction in context “B”. By contrast, a D1/D5-agonist impaired renewal in context “A”. Extinction in the “A” context on day 5 was unaffected by the D1/D5-ligands. Treatment with a D2-receptor antagonist prior to extinction had no overall effect on extinction in context “B” or renewal in context “A”, although extinction of the renewal effect was impaired on day 5, compared to controls. Taken together, these data suggest that dopamine acting on the D1/D5-receptor modulates both acquisition and consolidation of context-dependent extinction. By contrast, the D2-receptor may contribute to context-independent aspects of this kind of extinction learning.","{'model': 'tldr@v2.0.0', 'text': 'The data suggest that dopamine acting on the D1/D5-receptor modulates both acquisition and consolidation of context-dependent extinction, and that the D2- receptor may contribute to context-independent aspects of this kind of extinction learning.'}",https://www.frontiersin.org/articles/10.3389/fnbeh.2015.00372/pdf
-Setting learning analytics in context: overcoming the barriers to large-scale adoption,Rebecca Ferguson,"Once learning analytics have been successfully developed and tested, the next step is to implement them at a larger scale -- across a faculty, an institution or an educational system. This introduces a new set of challenges, because education is a stable system, resistant to change. Implementing learning analytics at scale involves working with the entire technological complex that exists around technology-enhanced learning (TEL). This includes the different groups of people involved -- learners, educators, administrators and support staff -- the practices of those groups, their understandings of how teaching and learning take place, the technologies they use and the specific environments within which they operate. Each element of the TEL Complex requires explicit and careful consideration during the process of implementation, in order to avoid failure and maximise the chances of success. In order for learning analytics to be implemented successfully at scale, it is crucial to provide not only the analytics and their associated tools but also appropriate forms of support, training and community building.","{'model': 'tldr@v2.0.0', 'text': 'In order for learning analytics to be implemented successfully at scale, it is crucial to provide not only the analytics and their associated tools but also appropriate forms of support, training and community building.'}",https://learning-analytics.info/index.php/JLA/article/download/4077/4421
-Self-supervised learning for medical image analysis using image context restoration,Liang Chen,,"{'model': 'tldr@v2.0.0', 'text': 'A novel self-supervised learning strategy based on context restoration is proposed in order to better exploit unlabelled images and is validated in three common problems in medical imaging: classification, localization, and segmentation.'}",
-Learning from Context or Names? An Empirical Study on Neural Relation Extraction,Hao Peng,"Neural models have achieved remarkable success on relation extraction (RE) benchmarks. However, there is no clear understanding which type of information affects existing RE models to make decisions and how to further improve the performance of these models. To this end, we empirically study the effect of two main information sources in text: textual context and entity mentions (names). We find that (i) while context is the main source to support the predictions, RE models also heavily rely on the information from entity mentions, most of which is type information, and (ii) existing datasets may leak shallow heuristics via entity mentions and thus contribute to the high performance on RE benchmarks. Based on the analyses, we propose an entity-masked contrastive pre-training framework for RE to gain a deeper understanding on both textual context and type information while avoiding rote memorization of entities or use of superficial cues in mentions. We carry out extensive experiments to support our views, and show that our framework can improve the effectiveness and robustness of neural models in different RE scenarios. All the code and datasets are released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'An entity-masked contrastive pre-training framework for RE is proposed to gain a deeper understanding on both textual context and type information while avoiding rote memorization of entities or use of superficial cues in mentions.'}",https://arxiv.org/pdf/2010.01923
-Machine learning on knowledge graphs for context-aware security monitoring,J. Garrido,"Machine learning techniques are gaining attention in the context of intrusion detection due to the increasing amounts of data generated by monitoring tools, as well as the sophistication displayed by attackers in hiding their activity. However, existing methods often exhibit important limitations in terms of the quantity and relevance of the generated alerts. Recently, knowledge graphs are finding application in the cybersecurity domain, showing the potential to alleviate some of these drawbacks thanks to their ability to seamlessly integrate data from multiple domains using human-understandable vocabularies. We discuss the application of machine learning on knowledge graphs for intrusion detection and experimentally evaluate a link-prediction method for scoring anomalous activity in industrial systems. After initial unsupervised training, the proposed method is shown to produce intuitively well-calibrated and interpretable alerts in a diverse range of scenarios, hinting at the potential benefits of relational machine learning on knowledge graphs for intrusion detection purposes.","{'model': 'tldr@v2.0.0', 'text': 'After initial unsupervised training, the proposed method is shown to produce intuitively well-calibrated and interpretable alerts in a diverse range of scenarios, hinting at the potential benefits of relational machine learning on knowledge graphs for intrusion detection purposes.'}",https://arxiv.org/pdf/2105.08741
-"Beyond Transmitting Bits: Context, Semantics, and Task-Oriented Communications",Deniz Gündüz,"Communication systems to date primarily aim at reliably communicating bit sequences. Such an approach provides efficient engineering designs that are agnostic to the meanings of the messages or to the goal that the message exchange aims to achieve. Next generation systems, however, can be potentially enriched by folding message semantics and goals of communication into their design. Further, these systems can be made cognizant of the context in which communication exchange takes place, thereby providing avenues for novel design insights. This tutorial summarizes the efforts to date, starting from its early adaptations, semantic-aware and task-oriented communications, covering the foundations, algorithms and potential implementations. The focus is on approaches that utilize information theory to provide the foundations, as well as the significant role of learning in semantics and task-aware communications.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial summarizes the efforts to date, starting from its early adaptations, semantic-aware and task-oriented communications, covering the foundations, algorithms and potential implementations, with a focus on approaches that utilize information theory to provide the foundations.'}",https://arxiv.org/pdf/2207.09353
-Fast Context Adaptation via Meta-Learning,L. Zintgraf,"We propose CAVIA, a meta-learning method for fast adaptation that is scalable, flexible, and easy to implement. CAVIA partitions the model parameters into two parts: context parameters that serve as additional input to the model and are adapted on individual tasks, and shared parameters that are meta-trained and shared across tasks. At test time, the context parameters are updated with one or several gradient steps on a task-specific loss that is backpropagated through the shared part of the network. Compared to approaches that adjust all parameters on a new task (e.g., MAML), CAVIA can be scaled up to larger networks without overfitting on a single task, is easier to implement, and is more robust to the inner-loop learning rate. We show empirically that CAVIA outperforms MAML on regression, classification, and reinforcement learning problems.","{'model': 'tldr@v2.0.0', 'text': 'It is shown empirically that CAVIA outperforms MAML on regression, classification, and reinforcement learning problems and is easier to implement, and is more robust to the inner-loop learning rate.'}",
-Context-aware Dynamics Model for Generalization in Model-Based Reinforcement Learning,Kimin Lee,"Model-based reinforcement learning (RL) enjoys several benefits, such as data-efficiency and planning, by learning a model of the environment's dynamics. However, learning a global model that can generalize across different dynamics is a challenging task. To tackle this problem, we decompose the task of learning a global dynamics model into two stages: (a) learning a context latent vector that captures the local dynamics, then (b) predicting the next state conditioned on it. In order to encode dynamics-specific information into the context latent vector, we introduce a novel loss function that encourages the context latent vector to be useful for predicting both forward and backward dynamics. The proposed method achieves superior generalization ability across various simulated robotics and control tasks, compared to existing RL schemes.","{'model': 'tldr@v2.0.0', 'text': 'This work decomposes the task of learning a global dynamics model into two stages: learning a context latent vector that captures the local dynamics, then predicting the next state conditioned on it, and introduces a novel loss function that encourages the context latentvector to be useful for predicting both forward and backward dynamics.'}",
-"Context-Aware Computing, Learning, and Big Data in Internet of Things: A Survey",Omer Berat Sezer,"Internet of Things (IoT) has been growing rapidly due to recent advancements in communications and sensor technologies. Meanwhile, with this revolutionary transformation, researchers, implementers, deployers, and users are faced with many challenges. IoT is a complicated, crowded, and complex field; there are various types of devices, protocols, communication channels, architectures, middleware, and more. Standardization efforts are plenty, and this chaos will continue for quite some time. What is clear, on the other hand, is that IoT deployments are increasing with accelerating speed, and this trend will not stop in the near future. As the field grows in numbers and heterogeneity, “intelligence” becomes a focal point in IoT. Since data now becomes “big data,” understanding, learning, and reasoning with big data is paramount for the future success of IoT. One of the major problems in the path to intelligent IoT is understanding “context,” or making sense of the environment, situation, or status using data from sensors, and then acting accordingly in autonomous ways. This is called “context-aware computing,” and it now requires both sensing and, increasingly, learning, as IoT systems get more data and better learning from this big data. In this survey, we review the field, first, from a historical perspective, covering ubiquitous and pervasive computing, ambient intelligence, and wireless sensor networks, and then, move to context-aware computing studies. Finally, we review learning and big data studies related to IoT. We also identify the open issues and provide an insight for future study areas for IoT researchers.","{'model': 'tldr@v2.0.0', 'text': 'The field is reviewed from a historical perspective, covering ubiquitous and pervasive computing, ambient intelligence, and wireless sensor networks, and then, move to context-aware computing studies, which identify the open issues and provide an insight for future study areas for IoT researchers.'}",
-Technology Acceptance Model in M-learning context: A systematic review,M. Al-Emran,,"{'model': 'tldr@v2.0.0', 'text': 'The main findings include that most of the TAM studies involving M-learning focused on extending the TAM with external variables, followed by the studies that extended the model by factors from other theories/models.'}",http://umpir.ump.edu.my/id/eprint/22323/1/Technology%20Acceptance%20Model%20in%20M-learning%20context%20A%20systematic%20review.pdf
-Transformer-XL: Attentive Language Models beyond a Fixed-Length Context,Zihang Dai,"Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence, which consists of a segment-level recurrence mechanism and a novel positional encoding scheme.'}",https://www.aclweb.org/anthology/P19-1285.pdf
-A Review of Flipped Classroom and Cooperative Learning Method Within the Context of Vygotsky Theory,D. Erbil,"In the flipped classroom method, which is accepted as one of the blended learning approaches, the traditional teaching process takes place outside of the classroom through videos. Activities, projects, and homework related to upper-level cognitive field steps are carried out during classroom time. Research and interest in the flipped classroom are increasing steadily. Employing a cooperative learning method is suggested for using class time in the flipped classroom method. However, there has not been sufficient research on the implemented results of those suggestions. Moreover, there is no clear roadmap on how to incorporate cooperative learning methods into the flipped classroom. This research reviews theoretical infrastructures of flipped classroom and cooperative learning methods according to the Vygotsky theory and makes various suggestions for implementation and implementers.","{'model': 'tldr@v2.0.0', 'text': 'This research reviews theoretical infrastructures of flipped classroom and cooperative learning methods according to the Vygotsky theory and makes various suggestions for implementation and implementers.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2020.01157/pdf
-Learning Context Graph for Person Search,Yichao Yan,"Person re-identification has achieved great progress with deep convolutional neural networks. However, most previous methods focus on learning individual appearance feature embedding, and it is hard for the models to handle difficult situations with different illumination, large pose variance and occlusion. In this work, we take a step further and consider employing context information for person search. For a probe-gallery pair, we first propose a contextual instance expansion module, which employs a relative attention module to search and filter useful context information in the scene. We also build a graph learning framework to effectively employ context pairs to update target similarity. These two modules are built on top of a joint detection and instance feature learning framework, which improves the discriminativeness of the learned features. The proposed framework achieves state-of-the-art performance on two widely used person search datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a contextual instance expansion module, which employs a relative attention module to search and filter useful context information in the scene, and builds a graph learning framework to effectively employ context pairs to update target similarity.'}",https://arxiv.org/pdf/1904.01830
-Dynamic Context Selection for Document-level Neural Machine Translation via Reinforcement Learning,Xiaomian Kang,"Document-level neural machine translation has yielded attractive improvements. However, majority of existing methods roughly use all context sentences in a fixed scope. They neglect the fact that different source sentences need different sizes of context. To address this problem, we propose an effective approach to select dynamic context so that the document-level translation model can utilize the more useful selected context sentences to produce better translations. Specifically, we introduce a selection module that is independent of the translation module to score each candidate context sentence. Then, we propose two strategies to explicitly select a variable number of context sentences and feed them into the translation module. We train the two modules end-to-end via reinforcement learning. A novel reward is proposed to encourage the selection and utilization of dynamic context sentences. Experiments demonstrate that our approach can select adaptive context sentences for different source sentences, and significantly improves the performance of document-level translation methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a selection module that is independent of the translation module to score each candidate context sentence, and proposes two strategies to explicitly select a variable number of context sentences and feed them into thetranslation module.'}",https://www.aclweb.org/anthology/2020.emnlp-main.175.pdf
-Continual learning of context-dependent processing in neural networks,Guanxiong Zeng,,"{'model': 'tldr@v2.0.0', 'text': 'An approach involving a learning algorithm, called orthogonal weights modification, with the addition of a context-dependent processing module to enable highly compact systems to gradually learn myriad regularities of the real world and eventually behave appropriately within it.'}",https://arxiv.org/pdf/1810.01256
-Recommender systems to support learners’ Agency in a Learning Context: a systematic review,M. Deschênes,,"{'model': 'tldr@v2.0.0', 'text': 'This systematic review of the literature on recommenders for technology-enhanced learning examines the context in which recommenders are used, the manners in which they are evaluated and the results of those evaluations.'}",https://educationaltechnologyjournal.springeropen.com/counter/pdf/10.1186/s41239-020-00219-w
-context2vec: Learning Generic Context Embedding with Bidirectional LSTM,Oren Melamud,"Context representations are central to various NLP tasks, such as word sense disam-biguation, named entity recognition, co-reference resolution, and many more. In this work we present a neural model for efficiently learning a generic context embedding function from large corpora, us-ing bidirectional LSTM. With a very simple application of our context representations, we manage to surpass or nearly reach state-of-the-art results on sentence completion, lexical substitution and word sense disambiguation tasks, while substantially outperforming the popular context representation of averaged word embeddings. We release our code and pre-trained models, suggesting they could be useful in a wide variety of NLP tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a neural model for efficiently learning a generic context embedding function from large corpora, us-ing bidirectional LSTM and suggests they could be useful in a wide variety of NLP tasks.'}",https://www.aclweb.org/anthology/K16-1006.pdf
-Encoding History with Context-aware Representation Learning for Personalized Search,Yujia Zhou,"The key to personalized search is to clarify the meaning of current query based on user's search history. Previous personalized studies tried to build user profiles on the basis of historical data to tailor the ranking. However, we argue that the user profile based methods do not really disambiguate the current query. They still retain some semantic bias when building user profiles. In this paper, we propose to encode history with context-aware representation learning to enhance the representation of current query, which is a direct way to clarify the user's information need. Specifically, endowed with the benefit from transformer on aggregating contextual information, we devise a query disambiguation model to parse the meaning of current query in multiple stages. Moreover, for covering the cases that current query is not sufficient to express the intent, we train a personalized language model to predict user intent from existing queries. Under the interaction of two sub-models, we can generate the context-aware representation of current query and re-rank the results based on it. Experimental results show the significant improvement of our model compared with previous methods.","{'model': 'tldr@v2.0.0', 'text': 'A query disambiguation model to parse the meaning of current query in multiple stages, endowed with the benefit from transformer on aggregating contextual information, and a personalized language model to predict user intent from existing queries is devised.'}",
-Learning a Deep Listwise Context Model for Ranking Refinement,Qingyao Ai,"Learning to rank has been intensively studied and widely applied in information retrieval. Typically, a global ranking function is learned from a set of labeled data, which can achieve good performance on average but may be suboptimal for individual queries by ignoring the fact that relevant documents for different queries may have different distributions in the feature space. Inspired by the idea of pseudo relevance feedback where top ranked documents, which we refer as the local ranking context, can provide important information about the query's characteristics, we propose to use the inherent feature distributions of the top results to learn a Deep Listwise Context Model that helps us fine tune the initial ranked list. Specifically, we employ a recurrent neural network to sequentially encode the top results using their feature vectors, learn a local context model and use it to re-rank the top results. There are three merits with our model: (1) Our model can capture the local ranking context based on the complex interactions between top results using a deep neural network; (2) Our model can be built upon existing learning-to-rank methods by directly using their extracted feature vectors; (3) Our model is trained with an attention-based loss function, which is more effective and efficient than many existing listwise methods. Experimental results show that the proposed model can significantly improve the state-of-the-art learning to rank methods on benchmark retrieval corpora.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use the inherent feature distributions of the top results to learn a Deep Listwise Context Model that helps to fine tune the initial ranked list and can significantly improve the state-of-the-art learning to rank methods on benchmark retrieval corpora.'}",https://dl.acm.org/doi/pdf/10.1145/3209978.3209985
-Saliency detection by multi-context deep learning,Rui Zhao,"Low-level saliency cues or priors do not produce good enough saliency detection results especially when the salient object presents in a low-contrast background with confusing visual appearance. This issue raises a serious problem for conventional approaches. In this paper, we tackle this problem by proposing a multi-context deep learning framework for salient object detection. We employ deep Convolutional Neural Networks to model saliency of objects in images. Global context and local context are both taken into account, and are jointly modeled in a unified multi-context deep learning framework. To provide a better initialization for training the deep neural networks, we investigate different pre-training strategies, and a task-specific pre-training scheme is designed to make the multi-context modeling suited for saliency detection. Furthermore, recently proposed contemporary deep models in the ImageNet Image Classification Challenge are tested, and their effectiveness in saliency detection are investigated. Our approach is extensively evaluated on five public datasets, and experimental results show significant and consistent improvements over the state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-context deep learning framework for salient object detection that employs deep Convolutional Neural Networks to model saliency of objects in images and investigates different pre-training strategies to provide a better initialization for training the deep neural networks.'}",http://www.ee.cuhk.edu.hk/%7Exgwang/papers/zhaoOHWcvpr15.pdf
-Multimodal Deep Learning for Activity and Context Recognition,Valentin Radu,"Wearables and mobile devices see the world through the lens of half a dozen low-power sensors, such as, barometers, accelerometers, microphones and proximity detectors. But differences between sensors ranging from sampling rates, discrete and continuous data or even the data type itself make principled approaches to integrating these streams challenging. How, for example, is barometric pressure best combined with an audio sample to infer if a user is in a car, plane or bike? Critically for applications, how successfully sensor devices are able to maximize the information contained across these multi-modal sensor streams often dictates the fidelity at which they can track user behaviors and context changes. This paper studies the benefits of adopting deep learning algorithms for interpreting user activity and context as captured by multi-sensor systems. Specifically, we focus on four variations of deep neural networks that are based either on fully-connected Deep Neural Networks (DNNs) or Convolutional Neural Networks (CNNs). Two of these architectures follow conventional deep models by performing feature representation learning from a concatenation of sensor types. This classic approach is contrasted with a promising deep model variant characterized by modality-specific partitions of the architecture to maximize intra-modality learning. Our exploration represents the first time these architectures have been evaluated for multimodal deep learning under wearable data -- and for convolutional layers within this architecture, it represents a novel architecture entirely. Experiments show these generic multimodal neural network models compete well with a rich variety of conventional hand-designed shallow methods (including feature extraction and classifier construction) and task-specific modeling pipelines, across a wide-range of sensor types and inference tasks (four different datasets). Although the training and inference overhead of these multimodal deep approaches is in some cases appreciable, we also demonstrate the feasibility of on-device mobile and wearable execution is not a barrier to adoption. This study is carefully constructed to focus on multimodal aspects of wearable data modeling for deep learning by providing a wide range of empirical observations, which we expect to have considerable value in the community. We summarize our observations into a series of practitioner rules-of-thumb and lessons learned that can guide the usage of multimodal deep learning for activity and context detection.","{'model': 'tldr@v2.0.0', 'text': 'This paper studies the benefits of adopting deep learning algorithms for interpreting user activity and context as captured by multi-sensor systems under wearable data by evaluating four variations of deep neural networks based either on fully-connected Deep Neural Networks (DNNs) or Convolutional Neural networks (CNNs).'}",https://www.pure.ed.ac.uk/ws/files/76915345/Radu_et_al_2017_Multimodal_Deep_Learning.pdf
-Extending UTAUT2 toward acceptance of mobile learning in the context of higher education,A. Arain,,"{'model': 'tldr@v2.0.0', 'text': 'A tailored extension in UTAUT2 is made that provides valuable insights into assess m-learning acceptance in the context of higher education institutes of developing countries, specifically in Pakistan.'}",
-The SIGMORPHON 2019 Shared Task: Morphological Analysis in Context and Cross-Lingual Transfer for Inflection,Arya D. McCarthy,"The SIGMORPHON 2019 shared task on cross-lingual transfer and contextual analysis in morphology examined transfer learning of inflection between 100 language pairs, as well as contextual lemmatization and morphosyntactic description in 66 languages. The first task evolves past years’ inflection tasks by examining transfer of morphological inflection knowledge from a high-resource language to a low-resource language. This year also presents a new second challenge on lemmatization and morphological feature analysis in context. All submissions featured a neural component and built on either this year’s strong baselines or highly ranked systems from previous years’ shared tasks. Every participating team improved in accuracy over the baselines for the inflection task (though not Levenshtein distance), and every team in the contextual analysis task improved on both state-of-the-art neural and non-neural baselines.","{'model': 'tldr@v2.0.0', 'text': 'The SIGMORPHON 2019 shared task on cross-lingual transfer and contextual analysis in morphology examined transfer learning of inflection between 100 language pairs, as well as contextual lemmatization and morphosyntactic description in 66 languages.'}",https://www.aclweb.org/anthology/W19-4226.pdf
-BehavDT: A Behavioral Decision Tree Learning to Build User-Centric Context-Aware Predictive Model,Iqbal H. Sarker,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the proposed BehavDT context-aware model is more effective when compared with the traditional machine learning approaches, in predicting user diverse behaviors considering multi-dimensional contexts.'}",https://arxiv.org/pdf/2001.00621
-Learning Dynamic Context Augmentation for Global Entity Linking,Xiyuan Yang,"Despite of the recent success of collective entity linking (EL) methods, these “global” inference methods may yield sub-optimal results when the “all-mention coherence” assumption breaks, and often suffer from high computational cost at the inference stage, due to the complex search space. In this paper, we propose a simple yet effective solution, called Dynamic Context Augmentation (DCA), for collective EL, which requires only one pass through the mentions in a document. DCA sequentially accumulates context information to make efficient, collective inference, and can cope with different local EL models as a plug-and-enhance module. We explore both supervised and reinforcement learning strategies for learning the DCA model. Extensive experiments show the effectiveness of our model with different learning settings, base models, decision orders and attention mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'A simple yet effective solution, called Dynamic Context Augmentation (DCA), for collective EL, which requires only one pass through the mentions in a document, and can cope with different local EL models as a plug-and-enhance module.'}",https://www.aclweb.org/anthology/D19-1026.pdf
-Improving Students’ Mathematical Problem Solving Ability and Self-Efficacy through Guided Discovery Learning in Local Culture Context,R. E. Simamora,"Qualified learning materials is needed in the efforts to improve the quality of teaching-learning mathematics. Qualified learning materials can be obtained through development research. Learning materials in this study were learning materials that were developed based on guided discovery learning model. The learning materials was also developed by integrating local culture into a guided learning model. The local culture in this study was adapted to the local culture of the students, namely the Batak Toba. Learning materials in this study were developed using the development model of Thiagarajan et al. (1974). The result of second trial showed that learning materials based guided discovery learning with Batak Toba context improved students’ mathematical problem solving ability and self-efficacy significantly. Based on the results of the study, it was suggested that mathematics teachers make an effort qualified learning materials and integrate local culture in mathematics learning.",,https://www.iejme.com/download/improving-students-mathematical-problem-solving-ability-and-self-efficacy-through-guided-discovery-3966.pdf
-Aligning assessment with the needs of work-integrated learning: the challenges of authentic assessment in a complex context,R. Ajjawi,"Abstract Work-integrated learning (WIL) is a feature of university courses, both in professional areas, where it is commonplace, but also across many different disciplines. Assessment of WIL can be complex as it involves parties and settings external to the university, and it can be problematic because of difficulties in aligning learning activities during placements with what is or can be assessed by the university. This paper explores the relationship between students’ placement experiences and accompanying assessments in contexts where activities are tightly coupled with the curriculum, and in those where it is not. It draws on a qualitative analysis of student interviews and drawings by the interviewees of their WIL experiences, supplemented with analysis of unit guides. Our findings highlight that students’ perceptions of authenticity of assessment were undermined by misalignments between the student, university and industry. Assessment authenticity was perceived by students as based on alignment between their current and future selves in the assessment process, involvement of industry supervisors and relevance of placement activities to assessment activities. The paper discusses the complexity of coordination of educational activities with external partners, especially when one party drives assessment. It then suggests a reframing of WIL assessment to promote alignment and authenticity.","{'model': 'tldr@v2.0.0', 'text': 'The relationship between students’ placement experiences and accompanying assessments in contexts where activities are tightly coupled with the curriculum, and in those where it is not is explored, and a reframing of WIL assessment is suggested to promote alignment and authenticity.'}",https://www.tandfonline.com/doi/pdf/10.1080/02602938.2019.1639613?needAccess=true
-Self‐regulated learning in the clinical context: a systematic review,M. A. Houten-Schat,"Research has suggested beneficial effects of self‐regulated learning (SRL) for medical students' and residents' workplace‐based learning. Ideally, learners go through a cyclic process of setting learning goals, choosing learning strategies and assessing progress towards goals. A clear overview of medical students' and residents' successful key strategies, influential factors and effective interventions to stimulate SRL in the workplace is missing. This systematic review aims to provide an overview of and a theoretical base for effective SRL strategies of medical students and residents for their learning in the clinical context.","{'model': 'tldr@v2.0.0', 'text': 'This systematic review aims to provide an overview of and a theoretical base for effective SRL strategies of medical students and residents for their learning in the clinical context.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/medu.13615
-Assessment in the context of problem-based learning,C. V. D. van der Vleuten,,"{'model': 'tldr@v2.0.0', 'text': 'Programmatic assessment comes very close to achieving the desired constructive alignment with PBL, but its wide adoption—just like PBL—will take many years ahead of us.'}",https://link.springer.com/content/pdf/10.1007/s10459-019-09909-1.pdf
-Blended Learning of Physics in the Context of the Professional Development of Teachers,L. Krasnova,"In line with the improvement of traditional teaching methods, the new ones are intensively introduced at all levels of education. Usually, these are the methods tied with e-learning. Essentially, teachers must be able and ready to create an innovation-driven learning environment contributing to the effective individualization of the learning process. At the same time, each student should achieve the highest possible outcomes standing behind the personality development. This paper introduces the refresher courses designed for the physics teachers. These courses are based on the blended learning technology combining traditional full-time education with the elements of distance learning supported by LMS Moodle. The courses were tested at the Elabuga Institute of Kazan Federal University in 2016-2018. This paper describes the module-based course structure and methods for applying the e-learning modules. The distinctive feature of the course is that the content of the learning modules was designed to deliver different methods for teaching physics and to improve the general cultural competence of a teacher. The analysis of polling results (poll included 89 physics teachers) allowed assessing the efficiency of designed courses in the context of teacher’s professional development, his/her readiness to work in modern learning environment. The research outcomes will be also useful for foreign educational establishments implementing the teacher professional development programs.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the refresher courses designed for the physics teachers based on the blended learning technology combining traditional full-time education with the elements of blended education.'}",
-Imitation from Observation: Learning to Imitate Behaviors from Raw Video via Context Translation,Yuxuan Liu,"Imitation learning is an effective approach for autonomous systems to acquire control policies when an explicit reward function is unavailable, using supervision provided as demonstrations from an expert, typically a human operator. However, standard imitation learning methods assume that the agent receives examples of observation-action tuples that could be provided, for instance, to a supervised learning algorithm. This stands in contrast to how humans and animals imitate: we observe another person performing some behavior and then figure out which actions will realize that behavior, compensating for changes in viewpoint, surroundings, object positions and types, and other factors. We term this kind of imitation learning “imitation-from-observation,” and propose an imitation learning method based on video prediction with context translation and deep reinforcement learning. This lifts the assumption in imitation learning that the demonstration should consist of observations in the same environment configuration, and enables a variety of interesting applications, including learning robotic skills that involve tool use simply by observing videos of human tool use. Our experimental results show the effectiveness of our approach in learning a wide range of real-world robotic tasks modeled after common household chores from videos of a human demonstrator, including sweeping, ladling almonds, pushing objects as well as a number of tasks in simulation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an imitation learning method based on video prediction with context translation and deep reinforcement learning that enables a variety of interesting applications, including learning robotic skills that involve tool use simply by observing videos of human tool use.'}",https://arxiv.org/pdf/1707.03374
-Context theory of classification learning.,D. Medin,"Most theories dealing with ill-defined concepts assume that performance is based on category level information or a mixture of category level and specific item information. A context theory of classificatio n is described in which judgments are assumed to derive exclusively from stored exemplar information. The main idea is that a probe item acts as a retrieval cue to access information associated with stimuli similar to the probe. The predictions of the context theory are contrasted with those of a class of theories (including prototype theory) that assume that the information entering into judgments can be derived from an additive combination of information from component cue dimensions. Across four experiments using both geometric forms and schematic faces as stimuli, the context theory consistently gave a better account of the data. The relation of the context theory to other theories and phenomena associated with ill-defined concepts is discussed in detail. One of the major components of cognitive behavior concerns abstracting rules and forming concepts. Our entire system of naming objects and events, talking about them, and interacting with them presupposes the ability to group experiences into appropriate classes. Young children learn to tell the difference between dogs and cats, between clocks and fans, and between stars and street lights. Since few concepts are formally taught, the evolution of concepts from experience with exemplars must be a fundamental learning phenomenon. The focus of the present article is to explore how such conceptual achievements emerge from individual instances.","{'model': 'tldr@v2.0.0', 'text': 'A context theory of classificatio n is described in which judgments are assumed to derive exclusively from stored exemplar information, and the main idea is that a probe item acts as a retrieval cue to access information associated with stimuli similar to the probe.'}",
-"Self-directed language learning in a mobile-assisted, out-of-class context: do students walk the talk?",Gustavo García Botero,"Abstract Can mobile-assisted language learning (MALL) foster self-directed learning outside the classroom? This article examines informal, out-of-class engagement with a MALL tool: Duolingo. After being invited to use Duolingo, 118 higher education language students agreed to have their activity in the application tracked. In addition to the data collected from the Duolingo dashboard, questionnaires, as well as semi-structured interviews, were implemented. Results viewed from Garrison’s dimensions of self-directed learning (motivation, self-management and self-monitoring) reveal inconsistencies between what students think about the app and what they do with it. Questionnaires suggest that Duolingo can encourage out-of-class learning through fun activities, but interviews reveal a lack of sustained motivation, self-monitoring and self-management reflected in the low usage of the application. Tracking data reveals that students lack the time for out-of-class MALL during course weeks and are more active during holidays. The results highlight the importance of a context in which users are encouraged to take responsibility for their learning. However, most students will, besides encouragement, also need training and support for their self-directed learning.",,
-Improvements to Context Based Self-Supervised Learning,T. N. Mundhenk,"We develop a set of methods to improve on the results of self-supervised learning using context. We start with a baseline of patch based arrangement context learning and go from there. Our methods address some overt problems such as chromatic aberration as well as other potential problems such as spatial skew and mid-level feature neglect. We prevent problems with testing generalization on common self-supervised benchmark tests by using different datasets during our development. The results of our methods combined yield top scores on all standard self-supervised benchmarks, including classification and detection on PASCAL VOC 2007, segmentation on PASCAL VOC 2012, and ""linear tests"" on the ImageNet and CSAIL Places datasets. We obtain an improvement over our baseline method of between 4.0 to 7.1 percentage points on transfer learning classification tests. We also show results on different standard network architectures to demonstrate generalization as well as portability. All data, models and programs are available at: https://gdo-datasci.llnl.gov/selfsupervised/.","{'model': 'tldr@v2.0.0', 'text': 'A set of methods to improve on the results of self-supervised learning using context, which start with a baseline of patch based arrangement context learning and go from there, and shows results on different standard network architectures to demonstrate generalization as well as portability.'}",https://arxiv.org/pdf/1711.06379
-Personalization in Context-aware Ubiquitous Learning-Log System,Mengmeng Li,"This paper describes a ubiquitous learning log system called SCROLL (System for Capturing and Reminding of Learning Log). Learning log stands for the log of knowledge or learning experience acquired ubiquitously. This study primarily exploits a personalized learning and context-aware method supporting ubiquitous learning log system. Its aims lie in helping learners recall what they have logged (learned) making use of the contexts and learners' learning habits. The method contains three main measures, which are to recommend learning objects in accordance with both learners' needs and contexts, to detect their learning habits using the context history and to prompt them to review what they have learned regarding their learning habits. What's more, by monitoring learners' reaction on the recommendation or prompting, the method can improve its prediction. An experiment was conducted to evaluate SCROLL and the method. The results demonstrate that the system is very helpful for the learners and they benefit from the context-based recommendation and learning-style based prompting well.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that the SCROLL system is very helpful for the learners and they benefit from the context-based recommendation and learning-style based prompting well.'}",
-"Context, learning, and extinction.",S. Gershman,"A. Redish et al. (2007) proposed a reinforcement learning model of context-dependent learning and extinction in conditioning experiments, using the idea of ""state classification"" to categorize new observations into states. In the current article, the authors propose an interpretation of this idea in terms of normative statistical inference. They focus on renewal and latent inhibition, 2 conditioning paradigms in which contextual manipulations have been studied extensively, and show that online Bayesian inference within a model that assumes an unbounded number of latent causes can characterize a diverse set of behavioral results from such manipulations, some of which pose problems for the model of Redish et al. Moreover, in both paradigms, context dependence is absent in younger animals, or if hippocampal lesions are made prior to training. The authors suggest an explanation in terms of a restricted capacity to infer new causes.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that online Bayesian inference within a model that assumes an unbounded number of latent causes can characterize a diverse set of behavioral results from such manipulations, some of which pose problems for the model of Redish et al. (2007).'}",
-Challenges and Opportunities for Higher Education amid the COVID-19 Pandemic: The Philippine Context,C. Toquero,"Covid-19 affected higher educational institutions not just in Wuhan, China where the virus originated but all other higher educational institutions in 188 countries as of April 06, 2020. Educational countermeasures are taken to continue educating the students despite the COVID-19 predicaments. Based on the author’s experiences, research, observations in the academe, COVID-19 guidelines, and the need for alternative solutions, this article introduces how higher education is affected and how it can respond to future challenges. This article recommends to educational institutions to produce studies to proliferate and document the impact of the pandemic to the educational system. There is also a greater need for educational institutions to strengthen the practices in the curriculum and make it more responsive to the learning needs of the students even beyond the conventional classrooms.",,https://www.pedagogicalresearch.com/download/challenges-and-opportunities-for-higher-education-amid-the-covid-19-pandemic-the-philippine-context-7947.pdf
-Relationship between interest and mathematics performance in a technology-enhanced learning context in Malaysia,S. L. Wong,,"{'model': 'tldr@v2.0.0', 'text': 'The findings of this study pointed to the importance of igniting interest among students with lower mathematics performance given its strong link to mathematics performance.'}",https://telrp.springeropen.com/track/pdf/10.1186/s41039-019-0114-3
-Gene Function Prediction With Gene Interaction Networks: A Context Graph Kernel Approach,Xin Li,"Predicting gene functions is a challenge for biologists in the postgenomic era. Interactions among genes and their products compose networks that can be used to infer gene functions. Most previous studies adopt a linkage assumption, i.e., they assume that gene interactions indicate functional similarities between connected genes. In this study, we propose to use a gene's context graph, i.e., the gene interaction network associated with the focal gene, to infer its functions. In a kernel-based machine-learning framework, we design a context graph kernel to capture the information in context graphs. Our experimental study on a testbed of p53-related genes demonstrates the advantage of using indirect gene interactions and shows the empirical superiority of the proposed approach over linkage-assumption-based methods, such as the algorithm to minimize inconsistent connected genes and diffusion kernels.","{'model': 'tldr@v2.0.0', 'text': 'The experimental study on a testbed of p53-related genes demonstrates the advantage of using indirect gene interactions and shows the empirical superiority of the proposed approach over linkage-assumption-based methods, such as the algorithm to minimize inconsistent connected genes and diffusion kernels.'}",https://arxiv.org/pdf/2204.10473
-A Context-Aware User-Item Representation Learning for Item Recommendation,Libing Wu,"Both reviews and user-item interactions (i.e., rating scores) have been widely adopted for user rating prediction. However, these existing techniques mainly extract the latent representations for users and items in an independent and static manner. That is, a single static feature vector is derived to encode user preference without considering the particular characteristics of each candidate item. We argue that this static encoding scheme is incapable of fully capturing users’ preferences, because users usually exhibit different preferences when interacting with different items. In this article, we propose a novel context-aware user-item representation learning model for rating prediction, named CARL. CARL derives a joint representation for a given user-item pair based on their individual latent features and latent feature interactions. Then, CARL adopts Factorization Machines to further model higher order feature interactions on the basis of the user-item pair for rating prediction. Specifically, two separate learning components are devised in CARL to exploit review data and interaction data, respectively: review-based feature learning and interaction-based feature learning. In the review-based learning component, with convolution operations and attention mechanism, the pair-based relevant features for the given user-item pair are extracted by jointly considering their corresponding reviews. However, these features are only reivew-driven and may not be comprehensive. Hence, an interaction-based learning component further extracts complementary features from interaction data alone, also on the basis of user-item pairs. The final rating score is then derived with a dynamic linear fusion mechanism. Experiments on seven real-world datasets show that CARL achieves significantly better rating prediction accuracy than existing state-of-the-art alternatives. Also, with the attention mechanism, we show that the pair-based relevant information (i.e., context-aware information) in reviews can be highlighted to interpret the rating prediction for different user-item pairs.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes a novel context-aware user-item representation learning model for rating prediction, named CARL, which derives a joint representation for a given user- item pair based on their individual latent features and latent feature interactions.'}",https://arxiv.org/pdf/1712.02342
-Global Context Enhanced Graph Neural Networks for Session-based Recommendation,Ziyang Wang,"Session-based recommendation (SBR) is a challenging task, which aims at recommending items based on anonymous behavior sequences. Almost all the existing solutions for SBR model user preference only based on the current session without exploiting the other sessions, which may contain both relevant and irrelevant item-transitions to the current session. This paper proposes a novel approach, called Global Context Enhanced Graph Neural Networks (GCE-GNN) to exploit item transitions over all sessions in a more subtle manner for better inferring the user preference of the current session. Specifically, GCE-GNN learns two levels of item embeddings from session graph and global graph, respectively: (i) Session graph, which is to learn the session-level item embedding by modeling pairwise item-transitions within the current session; and (ii) Global graph, which is to learn the global-level item embedding by modeling pairwise item-transitions over all sessions. In GCE-GNN, we propose a novel global-level item representation learning layer, which employs a session-aware attention mechanism to recursively incorporate the neighbors' embeddings of each node on the global graph. We also design a session-level item representation learning layer, which employs a GNN on the session graph to learn session-level item embeddings within the current session. Moreover, GCE-GNN aggregates the learnt item representations in the two levels with a soft attention mechanism. Experiments on three benchmark datasets demonstrate that GCE-GNN outperforms the state-of-the-art methods consistently.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach to exploit item transitions over all sessions in a more subtle manner for better inferring the user preference of the current session, called GCE-GNN, which outperforms the state-of-the-art methods consistently.'}",https://arxiv.org/pdf/2106.05081
-Theorising decolonisation in the context of lifelong learning and transnational migration: anti-colonial and anti-racist perspectives,Srabani Maitra,"ABSTRACT In the age of transnational migration, the practices and policies of lifelong learning in many immigrant-receiving countries continue to be impacted by the cultural and discursive politics of colonial legacies. Drawing on a wide range of anti-colonial and anti-racist scholarship, we argue for an approach to lifelong learning that aims to decolonise the ideological underpinnings of colonial relations of rule, especially in terms of its racialised privileging of ‘whiteness’ and Eurocentrism. In the context of lifelong learning, decolonisation would achieve four important purposes. First, it would illustrate the nexus between knowledge, power, and colonial narratives by interrogating how knowledge-making is a fundamental aspect of ‘coloniality’. Second, decolonisation would entail challenging the hegemony of western knowledge, education, and credentials and upholding a ‘multiculturalism of knowledge’ that is inclusive and responsive to the cultural needs and values of transnational migrants. Third, decolonisation would lead to the need for planning and designing learning curricula as well as institutionalised pedagogy based on non-western knowledge systems and epistemic diversity. The final emphasis is on the urgency to decolonise our minds as lifelong learners, practitioners and policy-makers in order to challenge the passivity, colonisation, and marginalisation of learners both in classrooms and workplaces.",,http://eprints.gla.ac.uk/177182/7/177182.pdf
-Smart Mobile Learning Environment for Programming Education in Nigeria: Adaptivity and Context-Aware Features,F. J. Agbo,,"{'model': 'tldr@v2.0.0', 'text': 'The study presents some existing solutions for programming education with adaptive or context-aware features of SLE and proposes a model for designing an SLE with adaptivity and context-awareness features that take into cognizance the learner’s preferences and needs.'}",
-Attribute Recognition by Joint Recurrent Learning of Context and Correlation,Jingya Wang,"Recognising semantic pedestrian attributes in surveillance images is a challenging task for computer vision, particularly when the imaging quality is poor with complex background clutter and uncontrolled viewing conditions, and the number of labelled training data is small. In this work, we formulate a Joint Recurrent Learning (JRL) model for exploring attribute context and correlation in order to improve attribute recognition given small sized training data with poor quality images. The JRL model learns jointly pedestrian attribute correlations in a pedestrian image and in particular their sequential ordering dependencies (latent high-order correlation) in an end-to-end encoder/ decoder recurrent network. We demonstrate the performance advantage and robustness of the JRL model over a wide range of state-of-the-art deep models for pedestrian attribute recognition, multi-label image classification, and multi-person image annotation on two largest pedestrian attribute benchmarks PETA and RAP.","{'model': 'tldr@v2.0.0', 'text': 'The JRL model learns jointly pedestrian attribute correlations in a pedestrian image and in particular their sequential ordering dependencies (latent high-order correlation) in an end-to-end encoder/ decoder recurrent network.'}",https://arxiv.org/pdf/1709.08553
-Unsupervised Learning of Spoken Language with Visual Context,David F. Harwath,"Humans learn to speak before they can read or write, so why can’t computers do the same? In this paper, we present a deep neural network model capable of rudimentary spoken language acquisition using untranscribed audio training data, whose only supervision comes in the form of contextually relevant visual images. We describe the collection of our data comprised of over 120,000 spoken audio captions for the Places image dataset and evaluate our model on an image search and annotation task. We also provide some visualizations which suggest that our model is learning to recognize meaningful words within the caption spectrograms.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network model capable of rudimentary spoken language acquisition using untranscribed audio training data, whose only supervision comes in the form of contextually relevant visual images, is presented.'}",
-Developing a Mobile App for Learning English Vocabulary in an Open Distance Learning Context,Mpine Makoe,"Academic success depends on the comprehension of a language, which is linked to vocabulary learning. Many distance students in South Africa find it difficult to comprehend learning in a language other than their mother tongue. Finding effective strategies for enhancing English vocabulary of university students amidst the spatial, temporal, and pedagogic distance associated with Open Distance Learning (ODL) practices remains a challenge. To address the need for enhancing vocabulary development, mobile application systems (apps) were explored as the best vehicle for the delivery of the vocabulary learning. Mobile learning technologies are ideal in the ODL context because they are flexible, accessible, available, and cater for a myriad of interaction activities. The purpose of the study is to design and implement a mobile-based application aimed at enhancing English vocabulary teaching and learning. Using the Design-Based Research methodology, this study maps the steps taken to develop a vocabulary learning mobile app named VocUp; it describes the architecture, user interface, features of VocUp, and advocates for contextually-conscious and learning-driven app development.","{'model': 'tldr@v2.0.0', 'text': 'This study maps the steps taken to develop a vocabulary learning mobile app named VocUp and describes the architecture, user interface, features of VocUp, and advocates for contextually-conscious and learning-driven app development.'}",http://www.irrodl.org/index.php/irrodl/article/download/3746/4777
-CE-Net: Context Encoder Network for 2D Medical Image Segmentation,Zaiwang Gu,"Medical image segmentation is an important step in medical image analysis. With the rapid development of a convolutional neural network in image processing, deep learning has been used for medical image segmentation, such as optic disc segmentation, blood vessel detection, lung segmentation, cell segmentation, and so on. Previously, U-net based approaches have been proposed. However, the consecutive pooling and strided convolutional operations led to the loss of some spatial information. In this paper, we propose a context encoder network (CE-Net) to capture more high-level information and preserve spatial information for 2D medical image segmentation. CE-Net mainly contains three major components: a feature encoder module, a context extractor, and a feature decoder module. We use the pretrained ResNet block as the fixed feature extractor. The context extractor module is formed by a newly proposed dense atrous convolution block and a residual multi-kernel pooling block. We applied the proposed CE-Net to different 2D medical image segmentation tasks. Comprehensive results show that the proposed method outperforms the original U-Net method and other state-of-the-art methods for optic disc segmentation, vessel detection, lung segmentation, cell contour segmentation, and retinal optical coherence tomography layer segmentation.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive results show that the proposed CE-Net method outperforms the original U- net method and other state-of-the-art methods for optic disc segmentation, vessel detection, lung segmentation , cell contour segmentation and retinal optical coherence tomography layer segmentation.'}",https://arxiv.org/pdf/1903.02740
-Collaborative Live Media Curation: Shared Context for Participation in Online Learning,William A. Hamilton,"In recent years, online education's reach and scale have increased through new platforms for large and small online courses. However, these platforms often rely on impoverished modalities, which provide limited support for participation in social learning experiences. We present Collaborative Live Media Curation (CLMC), a new medium for sharing context and participation in online learning. CLMC involves collaborative, synchronous collection, creation, and assemblage of web media, including images, text, video, and sketch. CLMC integrates live media including streaming video, screenshares, audio, and text chat. We deploy and study LiveMâché, a CLMC technology probe, in four situated online learning contexts. We discovered student and instructor strategies for sharing context and participating including creating curations in advance, sketching to illustrate and gesture, real-time transformations, sharing perspective, and assembling live streams. We develop implications through live experience patterns, which describe how spatial and computing structures support social activities.","{'model': 'tldr@v2.0.0', 'text': 'This work deploy and study LiveMâché, a CLMC technology probe, in four situated online learning contexts and discovered student and instructor strategies for sharing context and participating including creating curations in advance, sketching to illustrate and gesture, real-time transformations, sharing perspective, and assembling live streams.'}",
-Choosing the Right Learning Management System (LMS) for the Higher Education Institution Context: A Systematic Review,N. Kasim,"A Learning Management System (LMS) is one e-learning tool that has been widely used to improve students’ learning experience and the construction of their understanding of certain topics. This concept paper discusses a number of potential Learning Management Systems (LMS) that can be utilised for teaching and learning processes in the context of Higher Education Institutions, for example Moodle, ATutor, Blackboard and SuccessFactors. The comparison is made based on a literature review of the characteristics of the selected LMS providers. Among the characteristics considered are flexibility, ease of use, accessibility and user-friendliness. This paper also provides some conclusions on the selection of the platform to be used. The results of this study give readers information to make their own decisions when choosing an LMS platform based on the needs of their institution.","{'model': 'tldr@v2.0.0', 'text': 'A number of potential Learning Management Systems that can be utilised for teaching and learning processes in the context of Higher Education Institutions, for example Moodle, ATutor, Blackboard and SuccessFactors are discussed.'}",https://online-journals.org/index.php/i-jet/article/download/5644/3990
-"Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems",S. Levine,"In this tutorial article, we aim to provide the reader with the conceptual tools needed to get started on research on offline reinforcement learning algorithms: reinforcement learning algorithms that utilize previously collected data, without additional online data collection. Offline reinforcement learning algorithms hold tremendous promise for making it possible to turn large datasets into powerful decision making engines. Effective offline reinforcement learning methods would be able to extract policies with the maximum possible utility out of the available data, thereby allowing automation of a wide range of decision-making domains, from healthcare and education to robotics. However, the limitations of current algorithms make this difficult. We will aim to provide the reader with an understanding of these challenges, particularly in the context of modern deep reinforcement learning methods, and describe some potential solutions that have been explored in recent work to mitigate these challenges, along with recent applications, and a discussion of perspectives on open problems in the field.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial article aims to provide the reader with the conceptual tools needed to get started on research on offline reinforcement learning algorithms: reinforcementlearning algorithms that utilize previously collected data, without additional online data collection.'}",
-"Context-Aware, Adaptive, and Scalable Android Malware Detection Through Online Learning",A. Narayanan,"It is well known that Android malware constantly evolves so as to evade detection. This causes the entire malware population to be nonstationary. Contrary to this fact, most of the prior works on machine learning based android malware detection have assumed that the distribution of the observed malware characteristics (i.e., features) does not change over time. In this paper, we address the problem of malware population drift and propose a novel online learning based framework to detect malware, named Casandra (C ontext-aware, Adaptive and Scalable ANDRoid m Alware detector). In order to perform accurate detection, a novel graph kernel that facilitates capturing apps security-sensitive behaviors along with their context information from dependence graphs is proposed. Besides being accurate and scalable, Casandra has specific advantages: first, being adaptive to the evolution in malware features over time; second, explaining the significant features that led to an apps classification as being malicious or benign. In a large-scale comparative analysis, Casandra outperforms two state-of-the-art techniques on a benchmark dataset achieving 99.23% F-measure. When evaluated with more than 87 000 apps collected in-the-wild, Casandra achieves 89.92% accuracy, outperforming existing techniques by more than 25% in their typical batch learning setting and more than $\text{7}$% when they are continuously retained, while maintaining comparable efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the problem of malware population drift and proposes a novel online learning based framework to detect malware, named Casandra, which has specific advantages: first, being adaptive to the evolution in malware features over time; second, explaining the significant features that led to an apps classification as being malicious or benign.'}",
-"Definition, framework and research issues of smart learning environments - a context-aware ubiquitous learning perspective",Gwo-jen Hwang,,"{'model': 'tldr@v2.0.0', 'text': 'The definition and criteria of smartlearning environments are presented from the perspective of context-aware ubiquitous learning and a framework is presented to address the design and development considerations of smart learning environments to support both online and real-world learning activities.'}",https://slejournal.springeropen.com/track/pdf/10.1186/s40561-014-0004-5
-Object Detection With Deep Learning: A Review,Zhong-Qiu Zhao,"Due to object detection’s close relationship with video analysis and image understanding, it has attracted much research attention in recent years. Traditional object detection methods are built on handcrafted features and shallow trainable architectures. Their performance easily stagnates by constructing complex ensembles that combine multiple low-level image features with high-level context from object detectors and scene classifiers. With the rapid development in deep learning, more powerful tools, which are able to learn semantic, high-level, deeper features, are introduced to address the problems existing in traditional architectures. These models behave differently in network architecture, training strategy, and optimization function. In this paper, we provide a review of deep learning-based object detection frameworks. Our review begins with a brief introduction on the history of deep learning and its representative tool, namely, the convolutional neural network. Then, we focus on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further. As distinct specific detection tasks exhibit different characteristics, we also briefly survey several specific tasks, including salient object detection, face detection, and pedestrian detection. Experimental analyses are also provided to compare various methods and draw some meaningful conclusions. Finally, several promising directions and tasks are provided to serve as guidelines for future work in both object detection and relevant neural network-based learning systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a review of deep learning-based object detection frameworks and focuses on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further.'}",https://arxiv.org/pdf/1807.05511
-"Rousing Minds to Life: Teaching, Learning, and Schooling in Social Context",R. Tharp,"Acknowledgments Introduction Part I. Teaching, Schooling, and Literacy: A Unified Theory of Education: 1. The redefinition of teaching and schooling 2. A theory of teaching as assisted performance 3. The means of assisting performance 4. The social organization of assisted performance 5. Language, literacy, and thought Part II. Practice: 6. A school organized for teaching: the Kamehameha Elementary Education Program 7. The activity setting of the instructional conversation: developing word and discourse meaning 8. The orchestration of activity settings: learning and social interaction in the whole group and independent centers 9. The interpsychological plane of teacher training 10. Assisting teacher performance through the ZPD: a case study 11. The intrapsychological plane of teacher training: the internalization of higher-order teaching skills 12. The schools in mind and society References Author index Subject index.",,
-Identifying context factors as a source for teacher professional learning,A. Koffeman,"ABSTRACT Traditionally, teacher professional learning is often seen as something that mainly takes place in organized, formalized settings. This article takes a broader starting point: the idea that it can be understood as the result of the teachers’ confrontations and interactions with and within their professional contexts, and that context can thus serve as a source for learning. To gain a closer understanding of the nature of these contexts, we start with exploring these contexts from a theoretical point of view. We then look for evidence for the role of different contexts in teacher learning autobiographies and learner reports, produced by a diverse group of experienced teachers (n = 18), as part of their course work in a master’s programme. The findings suggest that three contexts can be identified – a perso3nal practice domain, a social domain, and a theoretical domain – and that confrontations in each of these domains can take place planned as well as unplanned. The relevance of the study is that understanding the ways in which context can serve as a source for learning can contribute to the debate on how to stimulate teacher professional learning.",,
-Deep learning and process understanding for data-driven Earth system science,M. Reichstein,,"{'model': 'tldr@v2.0.0', 'text': 'It is argued that contextual cues should be used as part of deep learning to gain further process understanding of Earth system science problems, improving the predictive ability of seasonal forecasting and modelling of long-range spatial connections across multiple timescales.'}",https://pure.mpg.de/pubman/item/item_3029184_9/component/file_3282959/BGC3001P.pdf
-Contextual Cueing: Implicit Learning and Memory of Visual Context Guides Spatial Attention,M. Chun,"Global context plays an important, but poorly understood, role in visual tasks. This study demonstrates that a robust memory for visual context exists to guide spatial attention. Global context was operationalized as the spatial layout of objects in visual search displays. Half of the configurations were repeated across blocks throughout the entire session, and targets appeared within consistent locations in these arrays. Targets appearing in learned configurations were detected more quickly. This newly discovered form of search facilitation is termed contextual cueing. Contextual cueing is driven by incidentally learned associations between spatial configurations (context) and target locations. This benefit was obtained despite chance performance for recognizing the configurations, suggesting that the memory for context was implicit. The results show how implicit learning and memory of visual context can guide spatial attention towards task-relevant aspects of a scene.","{'model': 'tldr@v2.0.0', 'text': 'The results show how implicit learning and memory of visual context can guide spatial attention towards task-relevant aspects of a scene.'}",
-Informal self-regulated learning on a surgical rotation: uncovering student experiences in context,Nicole N. Woods,,"{'model': 'tldr@v2.0.0', 'text': 'The study aimed to examine undergraduate medical students’ day-to-day learning strategies in order to better understand the process of informal self-regulated learning in practice and to provide a forum for students to share and critically reflect on their own self- regulated learning strategies.'}",
-A guide to deep learning in healthcare,A. Esteva,,"{'model': 'tldr@v2.0.0', 'text': 'How these computational techniques can impact a few key areas of medicine and explore how to build end-to-end systems are described.'}",
-"Context, time, and memory retrieval in the interference paradigms of Pavlovian learning.",M. Bouton,"In this article I review research and theory on the ""interference paradigms"" in Pavlovian learning. In these situations (e.g., extinction, counterconditioning, and latent inhibition), a conditioned stimulus (CS) is associated with different unconditioned stimuli (USs) or outcomes in different phases of the experiment; retroactive interference, proactive interference, or both are often observed. In all of the paradigms, contextual stimuli influence performance, and when information is available, so does the passage of time. Memories of both phases are retained, and performance may depend on which is retrieved. Despite the similarity of the paradigms, conditioning theories tend to explain them with separate mechanisms. They also do not provide an adequate account of the context's role, fail to predict the effects of time, and overemphasize the role of learning or storage deficits. By accepting 4 propositions about animal memory (i.e., contextual stimuli guide retrieval, time is a context, different memories are differentially dependent on context, and interference occurs at performance output), a memory retrieval framework can provide an integrated account of context, time, and performance in the various paradigms.","{'model': 'tldr@v2.0.0', 'text': 'A memory retrieval framework can provide an integrated account of context, time, and performance in the various paradigms of Pavlovian learning by accepting 4 propositions about animal memory.'}",
-Context-Aware Recommendation-Based Learning Analytics Using Tensor and Coupled Matrix Factorization,Faisal M. Almutairi,"Student retention and timely graduation are enduring challenges in higher education. With the rapidly expanding collection and availability of learning data and related analytics, student performance can be accurately monitored, and possibly predicted ahead of time, thus, enabling early warning and degree planning “expert systems” to provide disciplined decision support to counselors, advisors, and educators. Previous work in educational data mining has explored matrix factorization techniques for grade prediction, albeit without taking contextual information into account. Temporal information should be informative as it distinguishes between the different class offerings and indirectly captures student experience as well. To exploit temporal and/or other kinds of context, we develop three approaches under the framework of collaborative filtering (CF). Two of the proposed approaches build upon coupled matrix factorization with a shared latent matrix factor. The third utilizes tensor factorization to model grades and their context, without introducing a new mode per context dimension as is common in the CF literature. The latent factors obtained can be used to predict grades and context, if desired. We evaluate these approaches on grade data obtained from the University of Minnesota. Experimental results show that fairly good prediction is possible even with simple approaches, but very accurate prediction is hard. The more advanced approaches can increase prediction accuracy, but only up to a point for the particular dataset considered.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that fairly good prediction is possible even with simple approaches, but very accurate prediction is hard, and the more advanced approaches can increase prediction accuracy, but only up to a point for the particular dataset considered.'}",
-A Review of Content-Based and Context-Based Recommendation Systems,Umair Javed,"In our work, we have presented two widely used recommendation systems. We have presented a context-aware recommender system to filter the items associated with user’s interests coupled with a context-based recommender system to prescribe those items. In this study, context-aware recommender systems perceive the user’s location, time, and company. The context-based recommender system retrieves patterns from World Wide Web-based on the user’s past interactions and provides future news recommendations. We have presented different techniques to support media recommendations for smartphones, to create a framework for context-aware, to filter E-learning content, and to deliver convenient news to the user. To achieve this goal, we have used content-based, collaborative filtering, a hybrid recommender system, and implemented a Web ontology language (OWL). We have also used the Resource Description Framework (RDF), JAVA, machine learning, semantic mapping rules, and natural ontology languages that suggest user items related to the search. In our work, we have used E-paper to provide users with the required news. After applying the semantic reasoning approach, we have concluded that by some means, this approach works similarly as a content-based recommender system since by taking the gain of a semantic approach, we can also recommend items according to the user’s interests. In a content-based recommender system, the system provides additional options or results that rely on the user’s ratings, appraisals, and interests.","{'model': 'tldr@v2.0.0', 'text': 'This study has concluded that by some means, this approach works similarly as a content-based recommender system since by taking the gain of a semantic approach, the system can also recommend items according to the user’s interests.'}",
-Sustainability in higher education in the context of the UN DESD: a review of learning and institutionalization processes,A. Wals,,,
-"Sociophonetics: The Role of Words, the Role of Context, and the Role of Words in Context",J. Hay,"This paper synthesizes a wide range of literature from sociolinguistics and cognitive psychology, to argue for a central role for the ""word"" as a vehicle of language variation and change. Three crucially interlinked strands of research are reviewed-the role of context in associative learning, the word-level storage of phonetic and contextual detail, and the phonetic consequences of skewed distributions of words across different contexts. I argue that the human capacity for associative learning, combined with attention to fine-phonetic detail at the level of the word, plays a significant role in predicting a range of subtle but systematically robust observed socioindexical patterns in speech production and perception.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that the human capacity for associative learning, combined with attention to fine-phonetic detail at the level of the word, plays a significant role in predicting a range of subtle but systematically robust observed socioindexical patterns in speech production and perception.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/tops.12326
-Nicotine enhances context learning but not context-shock associative learning.,Justin W. Kenney,"Nicotine has been found to enhance learning in a variety of tasks, including contextual fear conditioning. During contextual fear conditioning animals have to learn the context and associate the context with an unconditioned stimulus (footshock). As both of these types of learning co-occur during fear conditioning, it is not clear whether nicotine enhances one or both of these types of learning. To tease these two forms of learning apart, the authors made use of the context preexposure facilitation effect (CPFE). Acquisition of the CPFE requires that contextual and context-shock learning occurs on separate days, allowing for their individual manipulation. Nicotine (0.09 mg/kg) administered prior to contextual learning and retrieval enhanced the CPFE whereas administration prior to context-shock learning and retrieval had no effect. Thus, nicotine enhances contextual learning but not context-shock associative learning. Finally, the results are discussed in terms of a theory of how nicotine could alter hippocampal-cortical-amygdala interactions to facilitate contextual learning.","{'model': 'tldr@v2.0.0', 'text': 'Nicotine enhances contextual learning but not context-shock associative learning, which is consistent with a theory of how nicotine could alter hippocampal-cortical-amygdala interactions to facilitate contextual learning.'}",https://europepmc.org/articles/pmc2559949?pdf=render
-Development and Evaluation of an Active Learning Support System for Context-Aware Ubiquitous Learning,T. Hsu,"Situating students to learn from the real world has been recognized as an important and challenging issue. However, in a real-world learning environment, there are usually many physical constraints that affect the learning performance of students, such as the total learning time, the limitation of the number of students who can visit a learning target, and the time needed for moving from one learning location to another. It is essential to guide the students along an efficient learning path to maximize their learning performance according to the current situation. In this paper, an active learning support system (ALESS) for context-aware ubiquitous learning environments is designed and developed. ALESS can provide learning guidance when conducting ubiquitous learning activities. A great deal of context information is used in ALESS, including the location, the current capacity of the learning object, the time available, etc. ALESS is able to actively provide the required learning support to individual students when they approach the corresponding real-world learning targets. To evaluate the performance of ALESS, an experiment was conducted in the National Science Museum of Taiwan. The experimental results showed that, with the help of ALESS, the students learned more efficiently, and achieved better learning performance.","{'model': 'tldr@v2.0.0', 'text': 'An active learning support system (ALESS) for context-aware ubiquitous learning environments is designed and developed and shows that, with the help of ALESS, the students learned more efficiently, and achieved better learning performance.'}",https://ieeexplore.ieee.org/ielx7/4620076/7396975/07115952.pdf
-Context counts: How learners' contexts influence learning in a MOOC,Nina Hood,,"{'model': 'tldr@v2.0.0', 'text': ""This study examines how a learner's current role and context influences their ability to self-regulate their learning in a MOOC: Introduction to Data Science offered by Coursera.""}",https://researchonline.gcu.ac.uk/files/24389185/6728648_Context_Counts_preprint.pdf
-Investigating Different Syntactic Context Types and Context Representations for Learning Word Embeddings,Bofang Li,"The number of word embedding models is growing every year. Most of them are based on the co-occurrence information of words and their contexts. However, it is still an open question what is the best definition of context. We provide a systematical investigation of 4 different syntactic context types and context representations for learning word embeddings. Comprehensive experiments are conducted to evaluate their effectiveness on 6 extrinsic and intrinsic tasks. We hope that this paper, along with the published code, would be helpful for choosing the best context type and representation for a given task.","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a systematical investigation of 4 different syntactic context types and context representations for learning word embeddings and hopes it would be helpful for choosing the best context type and representation for a given task.'}",
-Education Context and English Teaching and Learning in the Kingdom of Saudi Arabia: An Overview.,Oqab Alrashidi,"This paper discusses the education context and English teaching and learning in the Kingdom of Saudi Arabia (KSA). The paper is organised into five main sections. The first section offers a brief glance at the social, religious, economic, and political context in KSA. The second section provides an overview of the education system in KSA, which includes a brief explanation of the history of education, a discussion of the role of government in modern education, and a description of the administration of education in the country. The third section presents information about English teaching and learning in public and higher education in KSA, while the fourth gives insights into the challenges and difficulties students face when learning English. The last section focuses on the importance of learning English in KSA.",,https://ccsenet.org/journal/index.php/elt/article/download/47946/25732
-Context Learning in the Rodent Hippocampus,Mark C. Fuhs,"We present a Bayesian statistical theory of context learning in the rodent hippocampus. While context is often defined in an experimental setting in relation to specific background cues or task demands, we advance a single, more general notion of context that suffices for a variety of learning phenomena. Specifically, a context is defined as a statistically stationary distribution of experiences, and context learning is defined as the problem of how to form contexts out of groups of experiences that cluster together in time. The challenge of context learning is solving the model selection problem: How many contexts make up the rodent's world? Solving this problem requires balancing two opposing goals: minimize the variability of the distribution of experiences within a context and minimize the likelihood of transitioning between contexts. The theory provides an understanding of why hippocampal place cell remapping sometimes develops gradually over many days of experience and why even consistent landmark differences may need to be relearned after other environmental changes. The theory provides an explanation for progressive performance improvements in serial reversal learning, based on a clear dissociation between the incremental process of context learning and the relatively abrupt context selection process. The impact of partial reinforcement on reversal learning is also addressed. Finally, the theory explains why alternating sequence learning does not consistently result in unique context-dependent sequence representations in hippocampus.","{'model': 'tldr@v2.0.0', 'text': 'The theory provides an understanding of why hippocampal place cell remapping sometimes develops gradually over many days of experience and why even consistent landmark differences may need to be relearned after other environmental changes.'}",
-Material recognition in the wild with the Materials in Context Database,Sean Bell,"Recognizing materials in real-world images is a challenging task. Real-world materials have rich surface texture, geometry, lighting conditions, and clutter, which combine to make the problem particularly difficult. In this paper, we introduce a new, large-scale, open dataset of materials in the wild, the Materials in Context Database (MINC), and combine this dataset with deep learning to achieve material recognition and segmentation of images in the wild. MINC is an order of magnitude larger than previous material databases, while being more diverse and well-sampled across its 23 categories. Using MINC, we train convolutional neural networks (CNNs) for two tasks: classifying materials from patches, and simultaneous material recognition and segmentation in full images. For patch-based classification on MINC we found that the best performing CNN architectures can achieve 85.2% mean class accuracy. We convert these trained CNN classifiers into an efficient fully convolutional framework combined with a fully connected conditional random field (CRF) to predict the material at every pixel in an image, achieving 73.1% mean class accuracy. Our experiments demonstrate that having a large, well-sampled dataset such as MINC is crucial for real-world material recognition and segmentation.","{'model': 'tldr@v2.0.0', 'text': 'A new, large-scale, open dataset of materials in the wild, the Materials in Context Database (MINC), is introduced, and convolutional neural networks are trained for two tasks: classifying materials from patches, and simultaneous material recognition and segmentation in full images.'}",https://arxiv.org/pdf/1412.0623
-"Motivational Dynamics in Language Learning: Change, Stability, and Context.",F. Waninge,"Motivation as a variable in L2 development is no longer seen as the stable individual difference factor it was once believed to be: Influenced by process-oriented models and principles, and especially by the growing understanding of how complex dynamic systems work, researchers have been focusing increasingly on the dynamic and changeable nature of the motivation process. In this study we micro-map the motivational dynamics of 4 language learners during their language lessons over a period of 2 weeks, using a novel instrument-the Motometer-combined with classroom observations and a questionnaire on motivation and attitude. The article answers three current questions concerning L2 motivation: (a) Can we demonstrate variability in students' L2 motivation in class; (b) Is there a detectable stable level of students' in-class motivation; and (c), If both of these are demonstrated, can they be accounted for by the classroom context? The results affirm that student motivation can be successfully explored using a dynamic systems framework. Our findings demonstrate how motivation changes over time on an individual level, while also being characterised by predictable and stable phases, and how it is inseparable from the learner's individual learning context. The data also show that motivation can be meaningfully studied at different interacting time scales. [ABSTRACT FROM AUTHOR]",,
-Functional emergence of the hippocampus in context fear learning in infant rats,C. Raineki,"The hippocampus is a part of the limbic system and is important for the formation of associative memories, such as acquiring information about the context (e.g., the place where an experience occurred) during emotional learning (e.g., fear conditioning). Here, we assess whether the hippocampus is responsible for pups' newly emerging context learning. In all experiments, postnatal day (PN) 21 and PN24 rat pups received 10 pairings of odor‐0.5 mA shock or control unpaired odor‐shock, odor only, or shock only. Some pups were used for context, cue or odor avoidance tests, while the remaining pups were used for c‐Fos immunohistochemistry to assess hippocampal activity during acquisition. Our results show that cue and odor avoidance learning were similar at both ages, while contextual fear learning and learning‐associated hippocampal (CA1, CA3, and dentate gyrus) activity (c‐Fos) only occurred in PN24 paired pups. To assess a causal relationship between the hippocampus and context conditioning, we infused muscimol into the hippocampus, which blocked acquisition of context fear learning in the PN24 pups. Muscimol or vehicle infusions did not affect cue learning or aversion to the odor at PN21 or PN24. The results suggest that the newly emerging contextual learning exhibited by PN24 pups is supported by the hippocampus. © 2009 Wiley‐Liss, Inc.","{'model': 'tldr@v2.0.0', 'text': 'The results show that cue and odor avoidance learning were similar at both ages, while contextual fear learning and learning‐associated hippocampal activity (c‐Fos) activity (CA1, CA3, and dentate gyrus) only occurred in PN24 paired pups.'}",https://europepmc.org/articles/pmc2891848?pdf=render
-Retrieval-Based Learning: An Episodic Context Account,Jeffrey D. Karpicke,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter proposes an episodic context account of retrieval-based learning, which explains retrieval practice in terms of context reinstatement, context updating, and restriction of the search set.'}",
-Recent Trends in Deep Learning Based Natural Language Processing,Tom Young,"Deep learning methods employ multiple processing layers to learn hierarchical representations of data and have produced state-of-the-art results in many domains. Recently, a variety of model designs and methods have blossomed in the context of natural language processing (NLP). In this paper, we review significant deep learning related models and methods that have been employed for numerous NLP tasks and provide a walk-through of their evolution. We also summarize, compare and contrast the various models and put forward a detailed understanding of the past, present and future of deep learning in NLP.","{'model': 'tldr@v2.0.0', 'text': 'This paper reviews significant deep learning related models and methods that have been employed for numerous NLP tasks and provides a walk-through of their evolution.'}",
-The effectiveness of a meaningful learning-based evaluation model for context-aware mobile learning,Yueh-Min Huang,"In recent years, context-aware mobile learning ( CAML) has been widely applied to various fields and has become a popular issue in educational research. Despite the tremendous potential of CAML and its growing significance, continued evaluations and refinements under the advice of field experts and instructors are crucial to ensure its validity, value and sustainability. In this paper, an evaluation model for CAML is developed based on meaningful learning theory using the analytic hierarchy process ( AHP). To verify the effectiveness of the model, three different CAML learning activities are tested, and some experts are invited to evaluate and compare them. As a result, the strengths and weaknesses of each learning activity are obtained. With the aid of the evaluation model proposed in this work, CAML developers can progressively enhance the value of learning activities, thus improving this new approach to learning. [ABSTRACT FROM AUTHOR]","{'model': 'tldr@v2.0.0', 'text': 'With the aid of the evaluation model proposed in this work, CAML developers can progressively enhance the value of learning activities, thus improving this new approach to learning.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/bjet.12147
-Education Majors' Preferences on the Functionalities of E-Learning Platforms in the Context of Blended Learning,N. Tsankov,"The modern stages of higher education development and the actual training of education majors require systematic use of different electronic forms and platforms of education in combination with the traditional educational methods and approaches which will provide students with essential digital skills and competencies, important for their future professional and personal success. Widespread learning management systems provide a common set of basic functionalities. In this study, an assessment of the preferences of education majors on the main functionalities of the electronic platforms used in the context of blended learning in university education is presented. The results reveal a preference on organizational and informational functionalities and less on communication features.","{'model': 'tldr@v2.0.0', 'text': 'An assessment of the preferences of education majors on the main functionalities of the electronic platforms used in the context of blended learning in university education reveals a preference on organizational and informational functionalities and less on communication features.'}",https://online-journals.org/index.php/i-jet/article/download/6971/4427
-Context learning and the effect of context on memory retrieval in Lymnaea.,James Haney,"Aerial respiratory behavior in Lymnaea was operantly conditioned so that the animals perform aerial respiration significantly less often. Using the standard training procedure (pond water made hypoxic by bubbling N2 through it) both food-deprived and fed animals learned and exhibited long-term memory (LTM). However, food-deprived animals exhibited neither learning nor memory when trained under a condition in which the hypoxic pond water also contained a food odorant (carrot, the food-odorant procedure). Fed animals, however, learned and exhibited LTM with the food-odorant procedure. Thus, the presence of the food odorant per se did not prevent learning or the establishment of LTM. Further experimentation, however, revealed that the ability of the snails to have recall (i.e., memory) for the learned behavior was dependent on the context in which memory was tested. That is, if animals were trained with the food-odorant procedure they could only exhibit recall if tested in the food-odorant context and vice versa with the standard training procedure. Thus, although fed animals could learn and show LTM with either training and testing procedure, LTM could only be seen when they were tested in the context in which they were trained.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the snails to have recall for the learned behavior was dependent on the context in which memory was tested, which meant that if animals were trained with the food-odorant procedure they could only exhibit recall if tested in theFood odorant context and vice versa with the standard training procedure.'}",http://learnmem.cshlp.org/content/8/1/35.full.pdf
-ABSORPTIVE CAPACITY: A NEW PERSPECTIVE ON LEARNING AND INNOVATION,Wesley M. Cohen,"Discusses the notion that the ability to exploit external knowledge is crucial to a firm's innovative capabilities. In addition, it is argued that the ability to evaluate and use outside knowledge is largely a function of the level of prior related knowledge--i.e., absorptive capacity. Prior research has shown that firms that conduct their own research and development (R&D) are better able to use information from external sources. Therefore, it is possible that the absorptive capacity of a firm is created as a byproduct of the firm's R&D investment. A simple model of firm R&D intensity is constructed in a broader context of what applied economists call the three classes of industry-level determinants of R&D intensity: demand, appropriability, and technological opportunity conditions. Several predictions are made, including the notions that absorptive capacity does have a direct effect on R&D spending and spillovers will provide a positive incentive to conduct R&D. All hypotheses are tested using cross-sectional survey data on technological opportunity and appropriability conditions--collected over the period 1975 to 1977 for 1,719 business units--in the American manufacturing sector from Levin et al. (1983, 1987) and the Federal Trade Commission's Line of Business Program data on business unit sales, transfers, and R&D expenditures. Results confirm that firms are sensitive to the characteristics of the learning environment in which they operate and that absorptive capacity does appear to be a part of a firm's decisions regarding resource allocation for innovative activity. Results also suggest that, although the analysis showing a positive effect of spillovers in two industry groups do not represent a direct test of the model, positive absorption incentive associated with spillovers may be sufficiently strong in some cases to more than offset the negative appropribility incentive. (SFL)",,
-Epistemic metacognition in context: evaluating and learning online information,L. Mason,,,
-"Does PBL Matter? Relations Between Instructional Context, Learning Strategies, and Learning Outcomes",M. Delva,,"{'model': 'tldr@v2.0.0', 'text': ""A learning strategies questionnaire shows that students' learning strategies are influenced by instructional context, and patterns of learning strategy use change over time, and there is tentative evidence that the students'learning strategies influence learning outcomes.""}",
-Seamless Learning Despite Context,M. Sharples,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter examines seamless learning, where the aim is to enable a continuous flow of meaning-making despite changes in the physical and social context by inducing a flow state such that learners are so engaged in a mobile learning activity that they lose awareness of their surroundings.'}",http://oro.open.ac.uk/41627/1/Seamless%20Learning%20Despite%20Context%20revised%20clean.pdf
-Technology acceptance model in educational context: A systematic literature review,A. Granić,"A respectable amount of work dealing with Technology Acceptance Model (TAM) clearly indicates a popularity of TAM in the field of technology acceptance in general. Nevertheless, there is still a gap in existing knowledge regarding representative academic literature that underlie research on TAM in educational context. The main objective of this systematic literature review is to provide an overview of the current state of research efforts on TAM application in the field of learning and teaching for a variety of learning domains, learning technologies and types of users. Through systematic search by the use of EBSCO Discovery Service, the review has identified 71 relevant studies ranged between 2003 and 2018. The main findings indicate that TAM and its many different versions represent a credible model for facilitating assessment of diverse learning technologies. TAM's core variables, perceived ease of use and perceived usefulness, have been proven to be antecedent factors affecting acceptance of learning with technology. The paper identifies some gaps in current work and suggests areas for further investigation. The results of this systematic review provide a better understanding of TAM acceptance studies in educational context and create a firm foundation for advancing knowledge in the field. Practitioner NotesWhat is already known about this topic Technology acceptance research in teaching and learning context has become an attractive trend.A number of reviews and meta‐analysis focused on specific topics related to technology acceptance in education have been conducted.The Technology Acceptance Model (TAM) is the key model in understanding predictors of human behaviour towards potential acceptance or rejection of the technology.What this paper adds The state of current research on Technology Acceptance Model application in educational context lacks comprehensive reviews addressing variety of learning domains, learning technologies and types of users.The paper presents systematic review of relevant academic literature on Technology Acceptance Model (TAM) in the field of learning and teaching.The paper provides empirical evidence on the predictive validity of the models based on TAM presented in selected literature.The findings revealed that TAM, along with its many different versions called TAM++, is a leading scientific paradigm and credible model for facilitating assessment of diverse technological deployments in educational context.TAM's core variables, perceived ease of use and perceived usefulness, have been proven to be antecedent factors that have affected acceptance of learning with technology.Implications for practice and/or policy The systematic review adds to the body of knowledge and creates a firm foundation for advancing knowledge in the field.By following the most common research objectives and/or by filling current gaps in applied research methods, chosen sample groups and types of result analysis, an own study could be conducted.Future research may well focus on identifying additional external factors that could further explain acceptance and usage of various learning technologies. [ABSTRACT FROM AUTHOR] uracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)","{'model': 'tldr@v2.0.0', 'text': 'An overview of the current state of research efforts on TAM application in the field of learning and teaching for a variety of learning domains, learning technologies and types of users is provided and empirical evidence on the predictive validity of the models based on TAM presented in selected literature is provided.'}",
-Learning Representations by Maximizing Mutual Information Across Views,Philip Bachman,"We propose an approach to self-supervised representation learning based on maximizing mutual information between features extracted from multiple views of a shared context. For example, one could produce multiple views of a local spatio-temporal context by observing it from different locations (e.g., camera positions within a scene), and via different modalities (e.g., tactile, auditory, or visual). Or, an ImageNet image could provide a context from which one produces multiple views by repeatedly applying data augmentation. Maximizing mutual information between features extracted from these views requires capturing information about high-level factors whose influence spans multiple views -- e.g., presence of certain objects or occurrence of certain events. Following our proposed approach, we develop a model which learns image representations that significantly outperform prior methods on the tasks we consider. Most notably, using self-supervised learning, our model learns representations which achieve 68.1% accuracy on ImageNet using standard linear evaluation. This beats prior results by over 12% and concurrent results by 7%. When we extend our model to use mixture-based representations, segmentation behaviour emerges as a natural side-effect. Our code is available online: this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a model which learns image representations that significantly outperform prior methods on the tasks the authors consider, and extends this model to use mixture-based representations, where segmentation behaviour emerges as a natural side-effect.'}",
-Deep Learning for Generic Object Detection: A Survey,Li Liu,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of the recent achievements in this field brought about by deep learning techniques, covering many aspects of generic object detection: detection frameworks, object feature representation, object proposal generation, context modeling, training strategies, and evaluation metrics.'}",https://link.springer.com/content/pdf/10.1007/s11263-019-01247-4.pdf
-Toward the Use of Technology and 21st Century Teaching-learning Approaches: The Trend of Development in Malaysian Schools within the Context of Asia Pacific,S. Garba,"ICT Infrastructure and internet connectivity in educational institutions provides learners and teachers the opportunity of adopting 21st century teaching-learning methods that promotes the development of 21st century skills. The availability of internet connectivity in particular provides the platform for a shift from the use of teacher-centered pedagogy (content-based learning) to learner-centered pedagogy (inquiry and project-based learning) that is more interactive and activity oriented. But are teachers utilizing these facilities to advantage? This study investigates the use of ICT and internet resources in relation to the use of 21st century technology-based teaching-learning approaches in Malaysia and Asia Pacific context from a global perspective. Qualitative research approach was used for data collection and analysis in the study. Findings from the study show a unique unexpected changing pattern in the use of computer and internet among school teachers; and new challenges associated with the integration of ICT and the use of 21st century approaches in classroom pedagogical practices were identified. As observed in the study, despite the facilities provided; the technological competence acquired by teachers; and, the increasing use of computer and internet by teachers; the approach to teaching and learning has not change as desired.","{'model': 'tldr@v2.0.0', 'text': 'Despite the facilities provided; the technological competence acquired by teachers; and, the increasing use of computer and internet by teachers, the approach to teaching and learning has not change as desired.'}",https://online-journals.org/index.php/i-jet/article/download/4717/3593
-PPFNet: Global Context Aware Local Features for Robust 3D Point Matching,Haowen Deng,"We present PPFNet - Point Pair Feature NETwork for deeply learning a globally informed 3D local feature descriptor to find correspondences in unorganized point clouds. PPFNet learns local descriptors on pure geometry and is highly aware of the global context, an important cue in deep learning. Our 3D representation is computed as a collection of point-pair-features combined with the points and normals within a local vicinity. Our permutation invariant network design is inspired by PointNet and sets PPFNet to be ordering-free. As opposed to voxelization, our method is able to consume raw point clouds to exploit the full sparsity. PPFNet uses a novel N-tuple loss and architecture injecting the global information naturally into the local descriptor. It shows that context awareness also boosts the local feature representation. Qualitative and quantitative evaluations of our network suggest increased recall, improved robustness and invariance as well as a vital step in the 3D descriptor extraction performance.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative and quantitative evaluations of the PPFNet network suggest increased recall, improved robustness and invariance as well as a vital step in the 3D descriptor extraction performance.'}",https://arxiv.org/pdf/1802.02669
-Digital readiness and its effects on higher education students’ socio-emotional perceptions in the context of the COVID-19 pandemic,M. Händel,"Abstract The current study investigated how ready higher education students were for emergency remote teaching due to the COVID-19 pandemic and how this influenced their socio-emotional perceptions. Results of N = 1,826 higher education students indicate that they seem to be ready for digital learning. A k-means cluster analysis revealed two groups of students that significantly differed with respect to their readiness for digital learning (in terms of technology equipment availability, prior experiences with e-learning, and skills for digital learning). Finally, students’ socio-emotional perceptions, that is, stress-related emotions (worries, tension, joy, and overload) as well as social and emotional loneliness significantly differed due to cluster membership. Hence, the study points a need for support of higher education students in successfully coping with the challenges of emergency remote studying.",,https://osf.io/b9pg7/download
-National Systems of Innovation : Towards a Theory of Innovation and Interactive Learning,B. Lundvall,"Introduction Theories in the social sciences may be regarded as ‘focusing devices’. Any specific theory brings forward and exposes some aspects of the real world, leaving others in obscurity. That is why a long-lasting hegemony of one single theoretical tradition is damaging, both in terms of understanding and policymaking. In the field of economics, the dominating neoclassical paradigm puts its analytical focus on concepts such as scarcity, allocation and exchange in a static context. Even if these concepts reflect important phenomena in the real world, they only bring forward some aspects of the economic system. One aim of this book is to demonstrate the need for an alternative and supplementary focusing device that puts interactive learning and innovation at the centre of analysis. Through more than a decade, a group of economists at Aalborg University working on a research program on Innovation, Knowledge and Economic Dynamics – the IKE group – has worked together studying industrial development and international competitiveness from such a perspective. This book presents results from this work in relation to one specific subject: national systems of innovation. Our choice of perspective and subject is based on two sets of assumptions. First, it is assumed that the most fundamental resource in the modern economy is knowledge and accordingly that the most important process is learning. The fact that knowledge differs in crucial respects from other resources in the economy makes standard economics less relevant and motivates efforts to develop an alternative paradigm. Second, it is assumed that learning is predominantly an interactive, and therefore a socially embedded, process that cannot be understood without taking into consideration its institutional and cultural context. Specifically, it is assumed that the historical establishment and development of the modern nation state was a necessary prerequisite for the acceleration of the process of learning, which propelled the process of industrialization during the last centuries. Finally, it is recognized that the traditional role of nation states in supporting learning processes is now challenged by the process of internationalization and globalization.",,https://opus.bibliothek.uni-augsburg.de/opus4/files/85408/85408.pdf
-Optimization Methods for Large-Scale Machine Learning,L. Bottou,"This paper provides a review and commentary on the past, present, and future of numerical optimization algorithms in the context of machine learning applications. Through case studies on text classification and the training of deep neural networks, we discuss how optimization problems arise in machine learning and what makes them challenging. A major theme of our study is that large-scale machine learning represents a distinctive setting in which the stochastic gradient (SG) method has traditionally played a central role while conventional gradient-based nonlinear optimization techniques typically falter. Based on this viewpoint, we present a comprehensive theory of a straightforward, yet versatile SG algorithm, discuss its practical behavior, and highlight opportunities for designing algorithms with improved performance. This leads to a discussion about the next generation of optimization methods for large-scale machine learning, including an investigation of two main streams of research on techniques that diminish noise in the stochastic directions and methods that make use of second-order derivative approximations.","{'model': 'tldr@v2.0.0', 'text': 'A major theme of this study is that large-scale machine learning represents a distinctive setting in which the stochastic gradient method has traditionally played a central role while conventional gradient-based nonlinear optimization techniques typically falter, leading to a discussion about the next generation of optimization methods for large- scale machine learning.'}",https://arxiv.org/pdf/1606.04838
-Apprenticeship in Thinking: Cognitive Development in Social Context,B. Rogoff,PART I: The individual and the sociocultural context: Conceiving the relationship of the social world and the individual The sociocultural context of cognitive activity PART II: Processes of guided participation: Providing bridges from known to new Structuring situations and transferring responsibility Cultural universals and variations in guided participation PART III: Cognitive development through interaction with adults and peers: Explanations for cognitive development through social interaction: Vygotsky and Piaget Evidence of learning from guided participation with adults Peer interaction and cognitive development Shared thinking and guided participation.,,
-Image-Based Surface Defect Detection Using Deep Learning: A Review,P. Bhatt," Automatically detecting surface defects from images is an essential capability in manufacturing applications. Traditional image processing techniques are useful in solving a specific class of problems. However, these techniques do not handle noise, variations in lighting conditions, and backgrounds with complex textures. In recent times, deep learning has been widely explored for use in automation of defect detection. This survey article presents three different ways of classifying various efforts in literature for surface defect detection using deep learning techniques. These three ways are based on defect detection context, learning techniques, and defect localization and classification method respectively. This article also identifies future research directions based on the trends in the deep learning area.","{'model': 'tldr@v2.0.0', 'text': 'Three different ways of classifying various efforts in literature for surface defect detection using deep learning techniques are presented, based on defect detection context, learning techniques, and defect localization and classification method respectively.'}",
-Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles,M. Noroozi,,"{'model': 'tldr@v2.0.0', 'text': 'A novel unsupervised learning approach to build features suitable for object detection and classification and to facilitate the transfer of features to other tasks, the context-free network (CFN), a siamese-ennead convolutional neural network is introduced.'}",https://arxiv.org/pdf/1603.09246
-Context-Specific Proportion Congruency Effects: An Episodic Learning Account and Computational Model,James R. Schmidt,"In the Stroop task, participants identify the print color of color words. The congruency effect is the observation that response times and errors are increased when the word and color are incongruent (e.g., the word “red” in green ink) relative to when they are congruent (e.g., “red” in red). The proportion congruent (PC) effect is the finding that congruency effects are reduced when trials are mostly incongruent rather than mostly congruent. This PC effect can be context-specific. For instance, if trials are mostly incongruent when presented in one location and mostly congruent when presented in another location, the congruency effect is smaller for the former location. Typically, PC effects are interpreted in terms of strategic control of attention in response to conflict, termed conflict adaptation or conflict monitoring. In the present manuscript, however, an episodic learning account is presented for context-specific proportion congruent (CSPC) effects. In particular, it is argued that context-specific contingency learning can explain part of the effect, and context-specific rhythmic responding can explain the rest. Both contingency-based and temporal-based learning can parsimoniously be conceptualized within an episodic learning framework. An adaptation of the Parallel Episodic Processing model is presented. This model successfully simulates CSPC effects, both for contingency-biased and contingency-unbiased (transfer) items. The same fixed-parameter model can explain a range of other findings from the learning, timing, binding, practice, and attentional control domains.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that context- specific contingency learning can explain part of the effect, and context-specific rhythmic responding can explain the rest, as well as an adaptation of the Parallel Episodic Processing model that successfully simulates CSPC effects.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2016.01806/pdf
-Applications of machine learning in drug discovery and development,J. Vamathevan,,"{'model': 'tldr@v2.0.0', 'text': 'The most useful techniques and how machine learning can promote data-driven decision making in drug discovery and development are discussed and major hurdles in the field are highlighted, such as the required data characteristics for applying ML.'}",https://europepmc.org/articles/pmc6552674?pdf=render
-Storytelling: A Strategy for Providing Context for Learning.,D. Billings,"Storytelling--a narrative of events related to nursing and linked to evidence--provides a context for learning, particularly for learners who require a rich context to understand and integrate concepts related to patient care. This article offers suggestions for developing and using stories in nursing education.","{'model': 'tldr@v2.0.0', 'text': 'Suggestions for developing and using stories in nursing education are offered, particularly for learners who require a rich context to understand and integrate concepts related to patient care.'}",
-Framing Ubiquitous Mobility Educationally: Mobile Devices and Context-Aware Learning,B. Bachmair,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter argues that mobile devices and the artefacts accessed through and created with them should be viewed as important cultural resources and, therefore, constitute valid resources for learning.'}",
-SenticNet 5: Discovering Conceptual Primitives for Sentiment Analysis by Means of Context Embeddings,E. Cambria," With the recent development of deep learning, research in AI has gained new vigor and prominence. While machine learning has succeeded in revitalizing many research fields, such as computer vision, speech recognition, and medical diagnosis, we are yet to witness impressive progress in natural language understanding. One of the reasons behind this unmatched expectation is that, while a bottom-up approach is feasible for pattern recognition, reasoning and understanding often require a top-down approach. In this work, we couple sub-symbolic and symbolic AI to automatically discover conceptual primitives from text and link them to commonsense concepts and named entities in a new three-level knowledge representation for sentiment analysis. In particular, we employ recurrent neural networks to infer primitives by lexical substitution and use them for grounding common and commonsense knowledge by means of multi-dimensional scaling. ","{'model': 'tldr@v2.0.0', 'text': 'This work couple sub-symbolic and symbolic AI to automatically discover conceptual primitives from text and link them to commonsense concepts and named entities in a new three-level knowledge representation for sentiment analysis.'}",https://ojs.aaai.org/index.php/AAAI/article/download/11559/11418
-Large-Scale Machine Learning with Stochastic Gradient Descent,L. Bottou,,"{'model': 'tldr@v2.0.0', 'text': 'A more precise analysis uncovers qualitatively different tradeoffs for the case of small-scale and large-scale learning problems.'}",http://leon.bottou.org/publications/pdf/compstat-2010.pdf
-Revisiting Semi-Supervised Learning with Graph Embeddings,Zhilin Yang,"We present a semi-supervised learning framework based on graph embeddings. Given a graph between instances, we train an embedding for each instance to jointly predict the class label and the neighborhood context in the graph. We develop both transductive and inductive variants of our method. In the transductive variant of our method, the class labels are determined by both the learned embeddings and input feature vectors, while in the inductive variant, the embeddings are defined as a parametric function of the feature vectors, so predictions can be made on instances not seen during training. On a large and diverse set of benchmark tasks, including text classification, distantly supervised entity extraction, and entity classification, we show improved performance over many of the existing models.","{'model': 'tldr@v2.0.0', 'text': 'On a large and diverse set of benchmark tasks, including text classification, distantly supervised entity extraction, and entity classification, the proposed semi-supervised learning framework shows improved performance over many of the existing models.'}",
-Context-Aware Recommender Systems for Learning: A Survey and Future Challenges,K. Verbert,"Recommender systems have been researched extensively by the Technology Enhanced Learning (TEL) community during the last decade. By identifying suitable resources from a potentially overwhelming variety of choices, such systems offer a promising approach to facilitate both learning and teaching tasks. As learning is taking place in extremely diverse and rich environments, the incorporation of contextual information about the user in the recommendation process has attracted major interest. Such contextualization is researched as a paradigm for building intelligent systems that can better predict and anticipate the needs of users, and act more efficiently in response to their behavior. In this paper, we try to assess the degree to which current work in TEL recommender systems has achieved this, as well as outline areas in which further work is needed. First, we present a context framework that identifies relevant context dimensions for TEL applications. Then, we present an analysis of existing TEL recommender systems along these dimensions. Finally, based on our survey results, we outline topics on which further research is needed.","{'model': 'tldr@v2.0.0', 'text': 'A context framework that identifies relevant context dimensions for TEL applications is presented and an analysis of existing TEL recommender systems along these dimensions is presented, and topics on which further research is needed are outlined.'}",https://lirias.kuleuven.be/bitstream/123456789/338644/3/survey-final.pdf
-vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations,Alexei Baevski,We propose vq-wav2vec to learn discrete representations of audio segments through a wav2vec-style self-supervised context prediction task. The algorithm uses either a gumbel softmax or online k-means clustering to quantize the dense representations. Discretization enables the direct application of algorithms from the NLP community which require discrete inputs. Experiments show that BERT pre-training achieves a new state of the art on TIMIT phoneme classification and WSJ speech recognition.,"{'model': 'tldr@v2.0.0', 'text': 'Experiments show that BERT pre-training achieves a new state of the art on TIMIT phoneme classification and WSJ speech recognition and the algorithm uses a gumbel softmax or online k-means clustering to quantize the dense representations.'}",
-Learning Context-aware Latent Representations for Context-aware Collaborative Filtering,Xin Liu,"In this paper, we propose a generic framework to learn context-aware latent representations for context-aware collaborative filtering. Contextual contents are combined via a function to produce the context influence factor, which is then combined with each latent factor to derive latent representations. We instantiate the generic framework using biased Matrix Factorization as the base model. A Stochastic Gradient Descent (SGD) based optimization procedure is developed to fit the model by jointly learning the weight of each context and latent factors. Experiments conducted over three real-world datasets demonstrate that our model significantly outperforms not only the base model but also the representative context-aware recommendation models.","{'model': 'tldr@v2.0.0', 'text': 'A Stochastic Gradient Descent based optimization procedure is developed to fit the model by jointly learning the weight of each context and latent factors, which significantly outperforms not only the base model but also the representative context-aware recommendation models.'}",
-Dynamic Edge Association and Resource Allocation in Self-Organizing Hierarchical Federated Learning Networks,Wei Yang Bryan Lim,"Federated Learning (FL) is a promising privacy-preserving distributed machine learning paradigm. However, communication inefficiency remains the key bottleneck that impedes its large-scale implementation. Recently, hierarchical FL (HFL) has been proposed in which data owners, i.e., workers, can first transmit their updated model parameters to edge servers for intermediate aggregation. This reduces the instances of global communication and straggling workers. To enable efficient HFL, it is important to address the issues of edge association and resource allocation in the context of non-cooperative players, i.e., workers, edge servers, and model owner. However, the existing studies merely focus on static approaches and do not consider the dynamic interactions and bounded rationalities of the players. In this paper, we propose a hierarchical game framework to study the dynamics of edge association and resource allocation in self-organizing HFL networks. In the lower-level game, the edge association strategies of the workers are modelled using an evolutionary game. In the upper-level game, a Stackelberg differential game is adopted in which the model owner decides an optimal reward scheme given the expected bandwidth allocation control strategy of the edge server. Finally, we provide numerical results to validate that our proposed framework captures the HFL system dynamics under varying sources of network heterogeneity.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hierarchical game framework to study the dynamics of edge association and resource allocation in self-organizing HFL networks and provides numerical results to validate that the proposed framework captures the HFL system dynamics under varying sources of network heterogeneity.'}",https://dr.ntu.edu.sg/bitstream/10356/156039/2/Dynamic%20Edge%20Association%20and%20Resource%20Allocation%20in%20Self-Organizing%20Hierarchical%20Federated%20Learning%20Networks.pdf
-Data Shapley: Equitable Valuation of Data for Machine Learning,Amirata Ghorbani,"As data becomes the fuel driving technological and economic growth, a fundamental challenge is how to quantify the value of data in algorithmic predictions and decisions. For example, in healthcare and consumer markets, it has been suggested that individuals should be compensated for the data that they generate, but it is not clear what is an equitable valuation for individual data. In this work, we develop a principled framework to address data valuation in the context of supervised machine learning. Given a learning algorithm trained on $n$ data points to produce a predictor, we propose data Shapley as a metric to quantify the value of each training datum to the predictor performance. Data Shapley value uniquely satisfies several natural properties of equitable data valuation. We develop Monte Carlo and gradient-based methods to efficiently estimate data Shapley values in practical settings where complex learning algorithms, including neural networks, are trained on large datasets. In addition to being equitable, extensive experiments across biomedical, image and synthetic data demonstrate that data Shapley has several other benefits: 1) it is more powerful than the popular leave-one-out or leverage score in providing insight on what data is more valuable for a given learning task; 2) low Shapley value data effectively capture outliers and corruptions; 3) high Shapley value data inform what type of new data to acquire to improve the predictor.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a principled framework to address data valuation in the context of supervised machine learning by proposing data Shapley as a metric to quantify the value of each training datum to the predictor performance.'}",
-Cultural Learning,R. Holmes,"Chapter 12 explores the ways we learn about our culture, how culture shapes learning, and different learning contexts. It discusses attachment, Bowlby’s attachment theory, the Strange Situation, attachment styles, challenges to Western attachment theories, and attachment and culture. It addresses parenting, Baumrind’s parenting styles, parental ethnotheories, and parenting in different cultural contexts. It explores learning in formal settings like school, the connection between culture and schooling, parent beliefs and academic outcomes, apprenticeships, and learning in informal settings. Finally, it investigates play as a context for cultural learning and includes culture-specific and cross-cultural studies. This chapter includes a case study, Culture Across Disciplines box, chapter summary, key terms, a What Do Other Disciplines Do? section, thought-provoking questions, and class and experiential activities.",,
-QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation,Dmitry Kalashnikov,"In this paper, we study the problem of learning vision-based dynamic manipulation skills using a scalable reinforcement learning approach. We study this problem in the context of grasping, a longstanding challenge in robotic manipulation. In contrast to static learning behaviors that choose a grasp point and then execute the desired grasp, our method enables closed-loop vision-based control, whereby the robot continuously updates its grasp strategy based on the most recent observations to optimize long-horizon grasp success. To that end, we introduce QT-Opt, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real-world grasping that generalizes to 96% grasp success on unseen objects. Aside from attaining a very high success rate, our method exhibits behaviors that are quite distinct from more standard grasping systems: using only RGB vision-based perception from an over-the-shoulder camera, our method automatically learns regrasping strategies, probes objects to find the most effective grasps, learns to reposition objects and perform other non-prehensile pre-grasp manipulations, and responds dynamically to disturbances and perturbations.","{'model': 'tldr@v2.0.0', 'text': 'QT-Opt is introduced, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real- world grasping that generalizes to 96% grasp success on unseen objects.'}",
-Cerebellar Contribution to Context Processing in Extinction Learning and Recall,Dae-In Chang,,"{'model': 'tldr@v2.0.0', 'text': 'Because the ventromedial prefrontal cortex (vmPFC) is critically involved in context-related processes during extinction retrieval, and there are no known connections between the cerebellum and the vmPFC, the Cerebellum may be less important during extinction recall.'}",
-The informational value of contexts affects context-dependent learning,Sara Lucke,,"{'model': 'tldr@v2.0.0', 'text': 'The results of two predictive-learning experiments support the idea that relevant contexts receive more attention, leading to stronger context specificity of learning.'}",https://link.springer.com/content/pdf/10.3758%2Fs13420-013-0104-z.pdf
-The DA antagonist tiapride impairs context-related extinction learning in a novel context without affecting renewal,S. Lissek,"Renewal describes the recovery of an extinguished response if recall is tested in a context different from the extinction context. Behavioral studies demonstrated that attention to relevant context strengthens renewal. Neurotransmitters mediating attention and learning such as the dopaminergic (DA) system presumably modulate extinction learning and renewal. However, the role of DA for non-fear-based extinction learning and renewal in humans has not yet been investigated. This fMRI study investigated effects of DA-antagonism upon context-related extinction in a predictive learning task in which extinction occurred either in a novel (ABA) or an unchanged (AAA) context. The tiapride-treated group (TIA) showed significantly impaired ABA extinction learning and a significant within-group difference between ABA and AAA extinction, compared to placebo (PLAC). Groups did not differ in their level of ABA renewal. In ABA extinction, TIA showed reduced activation in dlPFC and OFC, hippocampus, and temporal regions. Across groups, activation in PFC and hippocampus correlated negatively with ABA extinction errors. Results suggest that in context-related extinction learning DA in PFC and hippocampus is involved in readjusting the cue-outcome relationship in the presence of a novel context. However, relating context to the appropriate association during recall does not appear to rely exclusively on DA signaling.","{'model': 'tldr@v2.0.0', 'text': 'Results suggest that in context-related extinction learning DA in PFC and hippocampus is involved in readjusting the cue-outcome relationship in the presence of a novel context, however, relating context to the appropriate association during recall does not appear to rely exclusively on DA signaling.'}",
-"Doing a thematic analysis: A practical, step-by-step guide for learning and teaching scholars.",Moira Maguire,"Data analysis is central to credible qualitative research. Indeed the qualitative researcher is often described as the research instrument insofar as his or her ability to understand, describe and interpret experiences and perceptions is key to uncovering meaning in particular circumstances and contexts. While much has been written about qualitative analysis from a theoretical perspective we noticed that often novice, and even more experienced researchers, grapple with the ‘how’ of qualitative analysis. Here we draw on Braun and Clarke’s (2006) framework and apply it in a systematic manner to describe and explain the process of analysis within the context of learning and teaching research. We illustrate the process using a worked example based on (with permission) a short extract from a focus group interview, conducted with undergraduate students.",,
-Context change and associative learning.,Juan M. Rosas,"This article reviews the effects of changing the background context on performance in associative learning tasks in humans and animals. The findings are complementary and consistent over animal conditioning (Pavlovian and instrumental learning) and human predictive learning and memory paradigms. In many cases, a context change after learning can have surprisingly little disruptive influence on performance. Extinction, or retroactive interference treatments more generally, is more context-specific than the initial learning. Contexts become important if the participant is exposed to any of several treatments that involve prediction error, which may serve to increase attention to the context. Contexts also become important if they are given predictive or informational value. Studies of instrumental (operant) learning are further consistent with the idea that the context might also influence affordances that support voluntary actions. Context switch effects are not universal, but mainly occur when certain attention and perception processes can come into play. WIREs Cogn Sci 2013, 4:237-244. doi: 10.1002/wcs.1225 This article is categorized under: Psychology > Learning.","{'model': 'tldr@v2.0.0', 'text': 'The findings are complementary and consistent over animal conditioning and human predictive learning and memory paradigms and further consistent with the idea that the context might also influence affordances that support voluntary actions.'}",https://europepmc.org/articles/pmc3680141?pdf=render
-Context and repetition in word learning,Jessica S. Horst,"Young children learn words from a variety of situations, including shared storybook reading. A recent study by Horst et al. (2011a) demonstrates that children learned more new words during shared storybook reading if they were read the same stories repeatedly than if they were read different stories that had the same number of target words. The current paper reviews this study and further examines the effect of contextual repetition on children's word learning in both shared storybook reading and other situations, including fast mapping by mutual exclusivity. The studies reviewed here suggest that the same cognitive mechanisms support word learning in a variety of situations. Both practical considerations for experimental design and directions for future research are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The studies reviewed here suggest that the same cognitive mechanisms support word learning in a variety of situations, including fast mapping by mutual exclusivity in shared storybook reading and other situations.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00149/pdf
-A Bayesian context fear learning algorithm/automaton,F. Krasne,"Contextual fear conditioning is thought to involve the synaptic plasticity-dependent establishment in hippocampus of representations of to-be-conditioned contexts which can then become associated with USs in the amygdala. A conceptual and computational model of this process is proposed in which contextual attributes are assumed to be sampled serially and randomly during contextual exposures. Given this assumption, moment-to-moment information about such attributes will often be quite different from one exposure to another and, in particular, between exposures during which representations are created, exposures during which conditioning occurs, and during recall sessions. This presents challenges to current conceptual models of hippocampal function. In order to meet these challenges, our model's hippocampus was made to operate in different modes during representation creation and recall, and non-hippocampal machinery was constructed that controlled these hippocampal modes. This machinery uses a comparison between contextual information currently observed and information associated with existing hippocampal representations of familiar contexts to compute the Bayesian Weight of Evidence that the current context is (or is not) a known one, and it uses this value to assess the appropriateness of creation or recall modes. The model predicts a number of known phenomena such as the immediate shock deficit, spurious fear conditioning to contexts that are absent but similar to actually present ones, and modulation of conditioning by pre-familiarization with contexts. It also predicts a number of as yet unknown phenomena.","{'model': 'tldr@v2.0.0', 'text': 'A conceptual and computational model of contextual fear conditioning that uses a comparison between contextual information currently observed and information associated with existing hippocampal representations of familiar contexts to compute the Bayesian Weight of Evidence that the current context is (or is not) a known one.'}",https://www.frontiersin.org/articles/10.3389/fnbeh.2015.00112/pdf
-Context-specific temporal learning with non-conflict stimuli: proof-of-principle for a learning account of context-specific proportion congruent effects,James R. Schmidt,"The conflict adaptation account proposes that participants adjust attention to target and distracting stimuli in response to conflict. This is argued to explain the proportion congruent effect, wherein the congruency effect decreases as the proportion of conflicting incongruent trials increases. Some reports further argue that this conflict adaptation process can be context-specific. This paper presents a proof-of-principle for a competing account. It is suggested that such context-specific effects might be driven by very basic temporal learning processes. In the reported experiment, we manipulated stimulus contrast in place of congruency. In one location, stimulus letters were mostly easy to identify (high stimulus contrast). In the other location, letters were mostly hard to identify (low stimulus contrast). Participants produced a larger contrast effect in the mostly easy context. Along with supplemental analyses investigating the role of context switching and previous trial response times, the results are consistent with the notion that different rhythms of responding are learned for an easy versus hard location context. These results suggest that context-specific proportion congruency effects might result, in whole or in part, from temporal learning. Conflict adaptation may or may not play an additional role.","{'model': 'tldr@v2.0.0', 'text': 'A proof-of-principle for a competing account of conflict adaptation suggests that context-specific proportion congruency effects might result, in whole or in part, from temporal learning.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2014.01241/pdf
-Making Connections: Learning and Teaching Chemistry in Context,Donna King,,"{'model': 'tldr@v2.0.0', 'text': 'This narrative inquiry addresses the question: how do the recalled experiences of a student and her teacher in context-based and concept-based chemistry programs compare, and constructed a collective account around four themes; namely, the extent to which the student makes connections between chemistry concepts and real-world contexts.'}",https://eprints.qut.edu.au/14509/1/14509.pdf
-Situated Learning,,"Situated learning or situated cognition is a learning theory developed in the late 1980s by Jean Lave and Etienne Wenger and soon expanded by John Seely Brown and his colleagues. Situated learning theory is based on the assumption that knowledge should be presented in authentic context that involves its application. Both authors argue that learning should not be viewed as transmission of abstract and contextualized knowledge between individuals, but a social process within certain conditions which include activity, context and culture.",,
-From lecture to learning tasks: use of the 4C/ID model in a communication skills course in a continuing professional education context.,A. P. Susilo,"This article describes the use of four-component instructional design (4C/ID), a model to plan educational interventions for complex learning. This model was used to design a continuing education course on communication skills for health professionals in a context that is hierarchical and communal. The authors describe the 4C/ID model and provide an example of its application in designing the course. In the 4C/ID model, learning tasks serve as the backbone of the course, with lectures and other supportive information organized around them. The 4C/ID model is different from traditional models that base the course on lectures on different topics and connect part-task assignments to these topics. The use of the 4C/ID model to develop the educational intervention moves the paradigm from lectures to learning tasks to better prepare learners for real practice.","{'model': 'tldr@v2.0.0', 'text': 'This model was used to design a continuing education course on communication skills for health professionals in a context that is hierarchical and communal and moves the paradigm from lectures to learning tasks to better prepare learners for real practice.'}",
-"Continuity, commitment and context: adult siblings of people with autism plus learning disability.",R. Tozer,"Sibling relationships are usually lifelong and reciprocal. They can assume particular significance when a brother or sister has a learning disability. Until recently, adult siblings of people with disabilities such as severe autism have been ignored by policy, practice and research. This qualitative study contributes to an emerging literature by exploring how adult siblings, who have a brother or sister with autism (plus learning disability) and living in England, give meaning to their family (and caring) relationships and engage with service delivery. We spoke to 21 adult siblings using semi-structured interviews and met with 12 of their siblings with autism. Our analysis, using a broad narrative approach, demonstrates the continuity of the sibling relationship and an enduring personalised commitment. The nature of this relationship, however, is sensitive to context. How non-disabled adult siblings relate to their childhood experience is fundamental when making sense of this, as is their need to fulfil other social and family obligations, alongside their 'sense of duty' to support their disabled brother or sister. Sibling experience was further mediated by negotiating their 'perceived invisibility' in social care policy and practice. Our work concludes that by understanding the way relationships between siblings have developed over time, adult siblings' contribution to the lives of their brother or sister with autism can be better supported for the benefit of both parties. Such an approach would support current policy developments.","{'model': 'tldr@v2.0.0', 'text': ""By understanding the way relationships between siblings have developed over time, adult siblings' contribution to the lives of their brother or sister with autism can be better supported for the benefit of both parties.""}",
-Deep Learning with Limited Numerical Precision,Suyog Gupta,"Training of large-scale deep neural networks is often constrained by the available computational resources. We study the effect of limited precision data representation and computation on neural network training. Within the context of low-precision fixed-point computations, we observe the rounding scheme to play a crucial role in determining the network's behavior during training. Our results show that deep networks can be trained using only 16-bit wide fixed-point number representation when using stochastic rounding, and incur little to no degradation in the classification accuracy. We also demonstrate an energy-efficient hardware accelerator that implements low-precision fixed-point arithmetic with stochastic rounding.","{'model': 'tldr@v2.0.0', 'text': 'The results show that deep networks can be trained using only 16-bit wide fixed-point number representation when using stochastic rounding, and incur little to no degradation in the classification accuracy.'}",
-A survey on ensemble learning,Xibin Dong,,"{'model': 'tldr@v2.0.0', 'text': 'Challenges and possible research directions for each mainstream approach of ensemble learning are presented and an extra introduction is given for the combination of ensemblelearning with other machine learning hot spots such as deep learning, reinforcement learning, etc.'}",
-Learning in Context,Betsy Disalvo,"The authors present an exploratory study of Black middle school boys who play digital games. The study was conducted through observations and interviews with Black American middle school boys about digital games as an informal learning experience. The first goal of the study is to understand the cultural context that Black students from economically disadvantaged inner-city neighborhoods bring to playing digital games. The second goal of the study is to examine how this cultural context affects the learning opportunities with games. Third, the authors examine how differences in game play are potential factors in the discrepancy between White male gamers and Black male gamers in science, technology, engineering, and mathematics. Finally, the authors address several opportunities within the field of informal learning to augment game play by bridging the learning that takes place within game play to the real world.",,
-"Classroom Learning Environments: Retrospect, Context and Prospect",B. Fraser,,,
-struc2vec: Learning Node Representations from Structural Identity,Leonardo F. R. Ribeiro,"Structural identity is a concept of symmetry in which network nodes are identified according to the network structure and their relationship to other nodes. Structural identity has been studied in theory and practice over the past decades, but only recently has it been addressed with representational learning techniques. This work presents struc2vec, a novel and flexible framework for learning latent representations for the structural identity of nodes. struc2vec uses a hierarchy to measure node similarity at different scales, and constructs a multilayer graph to encode structural similarities and generate structural context for nodes. Numerical experiments indicate that state-of-the-art techniques for learning node representations fail in capturing stronger notions of structural identity, while struc2vec exhibits much superior performance in this task, as it overcomes limitations of prior approaches. As a consequence, numerical experiments indicate that struc2vec improves performance on classification tasks that depend more on structural identity.","{'model': 'tldr@v2.0.0', 'text': 'Struc2vec, a novel and flexible framework for learning latent representations for the structural identity of nodes, is presented, which improves performance on classification tasks that depend more on structural identity.'}",https://arxiv.org/pdf/1704.03165
-A Review of Self-regulated Learning: Six Models and Four Directions for Research,E. Panadero,"Self-regulated learning (SRL) includes the cognitive, metacognitive, behavioral, motivational, and emotional/affective aspects of learning. It is, therefore, an extraordinary umbrella under which a considerable number of variables that influence learning (e.g., self-efficacy, volition, cognitive strategies) are studied within a comprehensive and holistic approach. For that reason, SRL has become one of the most important areas of research within educational psychology. In this paper, six models of SRL are analyzed and compared; that is, Zimmerman; Boekaerts; Winne and Hadwin; Pintrich; Efklides; and Hadwin, Järvelä and Miller. First, each model is explored in detail in the following aspects: (a) history and development, (b) description of the model (including the model figures), (c) empirical support, and (d) instruments constructed based on the model. Then, the models are compared in a number of aspects: (a) citations, (b) phases and subprocesses, (c) how they conceptualize (meta)cognition, motivation and emotion, (d) top–down/bottom–up, (e) automaticity, and (f) context. In the discussion, the empirical evidence from the existing SRL meta-analyses is examined and implications for education are extracted. Further, four future lines of research are proposed. The review reaches two main conclusions. First, the SRL models form an integrative and coherent framework from which to conduct research and on which students can be taught to be more strategic and successful. Second, based on the available meta-analytic evidence, there are differential effects of SRL models in light of differences in students’ developmental stages or educational levels. Thus, scholars and teachers need to start applying these differential effects of the SRL models and theories to enhance students’ learning and SRL skills.","{'model': 'tldr@v2.0.0', 'text': 'The SRL models form an integrative and coherent framework from which to conduct research and on which students can be taught to be more strategic and successful in order to enhance students’ learning and SRL skills.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2017.00422/pdf
-Transformers with convolutional context for ASR,Abdel-rahman Mohamed,"The recent success of transformer networks for neural machine translation and other NLP tasks has led to a surge in research work trying to apply it for speech recognition. Recent efforts studied key research questions around ways of combining positional embedding with speech features, and stability of optimization for large scale learning of transformer networks. In this paper, we propose replacing the sinusoidal positional embedding for transformers with convolutionally learned input representations. These contextual representations provide subsequent transformer blocks with relative positional information needed for discovering long-range relationships between local concepts. The proposed system has favorable optimization characteristics where our reported results are produced with fixed learning rate of 1.0 and no warmup steps. The proposed model achieves a competitive 4.7% and 12.9% WER on the Librispeech ``test clean'' and ``test other'' subsets when no extra LM text is provided.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes replacing the sinusoidal positional embedding for transformers with convolutionally learned input representations that provide subsequent transformer blocks with relative positional information needed for discovering long-range relationships between local concepts.'}",
-Rotation-Insensitive and Context-Augmented Object Detection in Remote Sensing Images,Ke Li,"Most of the existing deep-learning-based methods are difficult to effectively deal with the challenges faced for geospatial object detection such as rotation variations and appearance ambiguity. To address these problems, this paper proposes a novel deep-learning-based object detection framework including region proposal network (RPN) and local-contextual feature fusion network designed for remote sensing images. Specifically, the RPN includes additional multiangle anchors besides the conventional multiscale and multiaspect-ratio ones, and thus can deal with the multiangle and multiscale characteristics of geospatial objects. To address the appearance ambiguity problem, we propose a double-channel feature fusion network that can learn local and contextual properties along two independent pathways. The two kinds of features are later combined in the final layers of processing in order to form a powerful joint representation. Comprehensive evaluations on a publicly available ten-class object detection data set demonstrate the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel deep-learning-based object detection framework including region proposal network (RPN) and local-contextual feature fusion network designed for remote sensing images that can deal with the multiangle and multiscale characteristics of geospatial objects.'}",
-Learning image context for segmentation of the prostate in CT-guided radiotherapy,Wei Li,"Accurate segmentation of the prostate is the key to the success of external beam radiotherapy of prostate cancer. However, accurate segmentation of the prostate in computer tomography (CT) images remains challenging mainly due to three factors: (1) low image contrast between the prostate and its surrounding tissues, (2) unpredictable prostate motion across different treatment days and (3) large variations of intensities and shapes of the bladder and rectum around the prostate. In this paper, an online-learning and patient-specific classification method based on the location-adaptive image context is presented to deal with all these challenging issues and achieve the precise segmentation of the prostate in CT images. Specifically, two sets of location-adaptive classifiers are placed, respectively, along the two coordinate directions of the planning image space of a patient, and further trained with the planning image and also the previous-segmented treatment images of the same patient to jointly perform prostate segmentation for a new treatment image (of the same patient). In particular, each location-adaptive classifier, which itself consists of a set of sequential sub-classifiers, is recursively trained with both the static image appearance features and the iteratively updated image context features (extracted at different scales and orientations) for better identification of each prostate region. The proposed learning-based prostate segmentation method has been extensively evaluated on 161 images of 11 patients, each with more than nine daily treatment three-dimensional CT images. Our method achieves the mean Dice value 0.908 and the mean ± SD of average surface distance value 1.40 ± 0.57 mm. Its performance is also compared with three prostate segmentation methods, indicating the best segmentation accuracy by the proposed method among all methods under comparison.","{'model': 'tldr@v2.0.0', 'text': 'The proposed learning-based prostate segmentation method has been extensively evaluated and achieves the mean Dice value 0.908 and the mean ± SD of average surface distance value 1.40 ± 0.57 mm, indicating the best segmentation accuracy by the proposed method among all methods under comparison.'}",https://europepmc.org/articles/pmc3378724?pdf=render
-"Federated Learning for Wireless Communications: Motivation, Opportunities, and Challenges",Solmaz Niknam,"There is a growing interest in the wireless communications community to complement the traditional model-driven design approaches with data-driven machine learning (ML)-based solutions. While conventional ML approaches rely on the assumption of having the data and processing heads in a central entity, this is not always feasible in wireless communications applications because of the inaccessibility of private data and large communication overhead required to transmit raw data to central ML processors. As a result, decentralized ML approaches that keep the data where it is generated are much more appealing. Due to its privacy-preserving nature, federated learning is particularly relevant for many wireless applications, especially in the context of fifth generation (5G) networks. In this article, we provide an accessible introduction to the general idea of federated learning, discuss several possible applications in 5G networks, and describe key technical challenges and open problems for future research on federated learning in the context of wireless communications.","{'model': 'tldr@v2.0.0', 'text': 'An accessible introduction to the general idea of federated learning is provided, several possible applications in 5G networks are discussed, and key technical challenges and open problems for future research on Federated learning in the context of wireless communications are described.'}",https://arxiv.org/pdf/1908.06847
-Do no harm: a roadmap for responsible machine learning for health care,J. Wiens,,"{'model': 'tldr@v2.0.0', 'text': 'In this Perspective, the authors present a framework, context and guidelines for accelerating the translation of machine-learning-based interventions in health care.'}",
-Learning to reinforcement learn,Jane X. Wang,"In recent years deep reinforcement learning (RL) systems have attained superhuman performance in a number of challenging task domains. However, a major limitation of such applications is their demand for massive amounts of training data. A critical present objective is thus to develop deep RL methods that can adapt rapidly to new tasks. In the present work we introduce a novel approach to this challenge, which we refer to as deep meta-reinforcement learning. Previous work has shown that recurrent networks can support meta-learning in a fully supervised context. We extend this approach to the RL setting. What emerges is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure. This second, learned RL algorithm can differ from the original one in arbitrary ways. Importantly, because it is learned, it is configured to exploit structure in the training domain. We unpack these points in a series of seven proof-of-concept experiments, each of which examines a key aspect of deep meta-RL. We consider prospects for extending and scaling up the approach, and also point out some potentially important implications for neuroscience.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel approach to deep meta-reinforcement learning, which is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure.'}",
-Learning-Related Development of Context-Specific Neuronal Responses to Places and Events: The Hippocampal Role in Context Processing,David M. Smith,"Contextual information plays a key role in learning and memory. Learned information becomes associated with the context such that the context can cue the relevant memories and behaviors. An extensive literature involving experimental brain lesions has implicated the hippocampus in context processing. However, the neurophysiological mechanisms of context coding are not known. Although “context” has typically been defined in terms of the background cues, recent studies indicate that hippocampal neurons are sensitive to subtle changes in task demands, even in an unchanging environment. Thus, the context may also include non-environmental features of a learning situation. In the present study, hippocampal neuronal activity was recorded while rats learned to approach different reward locations in two contexts. Because all of the training took place in the same environment, the contexts were defined by the task demands rather than by environmental stimuli. Learning to differentiate two such contexts was associated with the development of highly context-specific neuronal firing patterns. These included different place fields in pyramidal neurons and different event (e.g., reward) responses in pyramidal and interneurons. The differential firing patterns did not develop in a control condition that did not involve a context manipulation. The context-specific firing patterns could modulate activity in extrahippocampal structures to prime context-appropriate behavioral responses and memories. These results provide direct support for a context processing role of the hippocampus and suggest that the hippocampus contributes contextual representations to episodic memories.","{'model': 'tldr@v2.0.0', 'text': 'Results provide direct support for a context processing role of the hippocampus and suggest that the hippocampus contributes contextual representations to episodic memories.'}",https://www.jneurosci.org/content/jneuro/26/12/3154.full.pdf
-Applications of Learning Analytics in High Schools: A Systematic Literature Review,Erverson B. G. de Sousa,"Learning analytics aims to analyze data from students and learning environments to support learning at different levels. Although learning analytics is a recent field, it reached a high level of maturity, especially in its applications for higher education. However, little of the research in learning analytics targets other educational levels, such as high school. This paper reports the results of a systematic literature review (SLR) focused on the adoption of learning analytics in high schools. More specifically, the SLR followed four steps: the search, selection of relevant studies, critical assessment, and the extraction of the relevant field, which included the main goals, approaches, techniques, and challenges of adopting learning analytics in high school. The results show that, in this context, learning analytics applications are focused on small-scale initiatives rather than institutional adoption. Based on the findings of this study, in combination with the literature, this paper proposes future directions of research and development in order to scale up learning analytics applications in high schools.","{'model': 'tldr@v2.0.0', 'text': 'A systematic literature review focused on the adoption of learning analytics in high schools shows that, in this context, learning analytics applications are focused on small-scale initiatives rather than institutional adoption.'}",https://www.frontiersin.org/articles/10.3389/frai.2021.737891/pdf
-PiCANet: Learning Pixel-Wise Contextual Attention for Saliency Detection,Nian Liu,"Contexts play an important role in the saliency detection task. However, given a context region, not all contextual information is helpful for the final task. In this paper, we propose a novel pixel-wise contextual attention network, i.e., the PiCANet, to learn to selectively attend to informative context locations for each pixel. Specifically, for each pixel, it can generate an attention map in which each attention weight corresponds to the contextual relevance at each context location. An attended contextual feature can then be constructed by selectively aggregating the contextual information. We formulate the proposed PiCANet in both global and local forms to attend to global and local contexts, respectively. Both models are fully differentiable and can be embedded into CNNs for joint training. We also incorporate the proposed models with the U-Net architecture to detect salient objects. Extensive experiments show that the proposed PiCANets can consistently improve saliency detection performance. The global and local PiCANets facilitate learning global contrast and homogeneousness, respectively. As a result, our saliency model can detect salient objects more accurately and uniformly, thus performing favorably against the state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel pixel-wise contextual attention network, i.e., the PiCANet, to learn to selectively attend to informative context locations for each pixel, which can generate an attention map in which each attention weight corresponds to the contextual relevance at each context location.'}",https://arxiv.org/pdf/1708.06433
-Workplace learning in context,A. Fuller,"This book combines a critical analysis of the organizational and employment context of workplace learning with an understanding of theories of learning. It brings new ideas on workplace learning to the readers and an understanding of the social context of the workplace. Combining the latest research findings with practical examples of the issues raised, it is illustrated with case studies and critical vignettes as well as chapter summaries, guiding students the subject. Chapters are grouped under four main themes: 'The context of workplace learning', 'The workplace as a learning environment', 'Skills, knowledge and the workplace' and 'Research and policy'. With international contributors offering a global perspective, this encourages students to be critical when evaluating the important debates in the area.",,https://opus.lib.uts.edu.au/bitstream/10453/12376/1/2004001318.pdf
-Learning to Compose Dynamic Tree Structures for Visual Contexts,Kaihua Tang,"We propose to compose dynamic tree structures that place the objects in an image into a visual context, helping visual reasoning tasks such as scene graph generation and visual Q&A. Our visual context tree model, dubbed VCTree, has two key advantages over existing structured object representations including chains and fully-connected graphs: 1) The efficient and expressive binary tree encodes the inherent parallel/hierarchical relationships among objects, e.g., ``clothes'' and ``pants'' are usually co-occur and belong to ``person''; 2) the dynamic structure varies from image to image and task to task, allowing more content-/task-specific message passing among objects. To construct a VCTree, we design a score function that calculates the task-dependent validity between each object pair, and the tree is the binary version of the maximum spanning tree from the score matrix. Then, visual contexts are encoded by bidirectional TreeLSTM and decoded by task-specific models. We develop a hybrid learning procedure which integrates end-task supervised learning and the tree structure reinforcement learning, where the former's evaluation result serves as a self-critic for the latter's structure exploration. Experimental results on two benchmarks, which require reasoning over contexts: Visual Genome for scene graph generation and VQA2.0 for visual Q&A, show that VCTree outperforms state-of-the-art results while discovering interpretable visual context structures.","{'model': 'tldr@v2.0.0', 'text': ""A hybrid learning procedure is developed which integrates end-task supervised learning and the tree structure reinforcement learning, where the former's evaluation result serves as a self-critic for the latter's structure exploration.""}",https://arxiv.org/pdf/1812.01880
-Computing the Context Effect for Science Learning,Thomas Forissier,,"{'model': 'tldr@v2.0.0', 'text': 'The hypothesis presented is that a context gap between two students can be illuminating to highlight the respective general-particular aspects of an object or phenomenon, and provoking a perturbation during the learning process to obtain the emergence of such an event could be a productive tutoring strategy.'}",
-Situated Cognition and the Culture of Learning,J. Brown,"Many teaching practices implicitly assume that conceptual knowledge can be abstracted from the situations in which it is learned and used. This article argues that this assumption inevitably limits the effectiveness of such practices. Drawing on recent research into cognition as it is manifest in everyday activity, the authors argue that knowledge is situated, being in part a product of the activity, context, and culture in which it is developed and used. They discuss how this view of knowledge affects our understanding of learning, and they note that conventional schooling too often ignores the influence of school culture on what is learned in school. As an alternative to conventional practices, they propose cognitive apprenticeship (Collins, Brown, & Newman, in press), which honors the situated nature of knowledge. They examine two examples of mathematics instruction that exhibit certain key features of this approach to teaching.",,https://www.ideals.illinois.edu/items/18064/bitstreams/64664/data.pdf
-An efficient framework for learning sentence representations,Lajanugen Logeswaran,"In this work we propose a simple and efficient framework for learning sentence representations from unlabelled data. Drawing inspiration from the distributional hypothesis and recent work on learning sentence representations, we reformulate the problem of predicting the context in which a sentence appears as a classification problem. Given a sentence and its context, a classifier distinguishes context sentences from other contrastive sentences based on their vector representations. This allows us to efficiently learn different types of encoding functions, and we show that the model learns high-quality sentence representations. We demonstrate that our sentence representations outperform state-of-the-art unsupervised and supervised representation learning methods on several downstream NLP tasks that involve understanding sentence semantics while achieving an order of magnitude speedup in training time.","{'model': 'tldr@v2.0.0', 'text': 'This work reformulates the problem of predicting the context in which a sentence appears as a classification problem, and proposes a simple and efficient framework for learning sentence representations from unlabelled data.'}",
-Learning End-to-End Goal-Oriented Dialog,Antoine Bordes,"Traditional dialog systems used in goal-oriented applications require a lot of domain-specific handcrafting, which hinders scaling up to new domains. End-to-end dialog systems, in which all components are trained from the dialogs themselves, escape this limitation. But the encouraging success recently obtained in chit-chat dialog may not carry over to goal-oriented settings. This paper proposes a testbed to break down the strengths and shortcomings of end-to-end dialog systems in goal-oriented applications. Set in the context of restaurant reservation, our tasks require manipulating sentences and symbols, so as to properly conduct conversations, issue API calls and use the outputs of such calls. We show that an end-to-end dialog system based on Memory Networks can reach promising, yet imperfect, performance and learn to perform non-trivial operations. We confirm those results by comparing our system to a hand-crafted slot-filling baseline on data from the second Dialog State Tracking Challenge (Henderson et al., 2014a). We show similar result patterns on data extracted from an online concierge service.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that an end-to-end dialog system based on Memory Networks can reach promising, yet imperfect, performance and learn to perform non-trivial operations and be compared to a hand-crafted slot-filling baseline on data from the second Dialog State Tracking Challenge.'}",
-Learning analytics in higher education: a preponderance of analytics but very little learning?,Carolina Guzmán-Valenzuela,,"{'model': 'tldr@v2.0.0', 'text': 'Examining the publication patterns on learning analytics in higher education and their main challenges concludes by distinguishing between a practice-based and management-oriented community of learning analytics and an academic- oriented community.'}",https://educationaltechnologyjournal.springeropen.com/counter/pdf/10.1186/s41239-021-00258-x
-Curriculum learning,Yoshua Bengio,"Humans and animals learn much better when the examples are not randomly presented but organized in a meaningful order which illustrates gradually more concepts, and gradually more complex ones. Here, we formalize such training strategies in the context of machine learning, and call them ""curriculum learning"". In the context of recent research studying the difficulty of training in the presence of non-convex training criteria (for deep deterministic and stochastic neural networks), we explore curriculum learning in various set-ups. The experiments show that significant improvements in generalization can be achieved. We hypothesize that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and, in the case of non-convex criteria, on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesized that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).'}",
-Does e-learning service quality influence e-learning student satisfaction and loyalty? Evidence from Vietnam,Long Pham,,,
-Reviewing and exploring innovative ubiquitous learning tools in higher education,Shadi A. Aljawarneh,,"{'model': 'tldr@v2.0.0', 'text': 'A global view regarding the current as well as future tendencies pertaining to ubiquitous e- learning tools is obtained and thus possible key comments are provided for employing e-learning tools like MOODLE, Web 2.0 and Web 3.0 in the classroom.'}",
-Design-Based Research Methods for Studying Learning in Context: Introduction,W. Sandoval,"The field of psychology has a long history of interaction with education, and educational psychology has had a profound impact on how issues of learning have been framed and studied in educational contexts. Still, it has never been simple to translate theoretical insights into educational practice. Educational psychology has been criticized for not creating “usable knowledge” (Lagemann, 2002). Currently, educational researchers generally have been pushed to justify how their claims are “scientific” and “evidence-based” (National Research Council, 2002). There is a tension between the desire for locally usable knowledge on the one hand and scientifically sound, generalizable knowledge on the other. Lagemann, for example, argued that the traditional paradigm of psychology has striven for experimental control at the expense of fidelity to learning as it actually occurs. Thus, although such claims might be scientific in one sense, they do not adequately explain or predict the phenomena they purport to address. This critique extends the long-standing debate surrounding the ecological validity of well-defined psychological tasks and their relation to psychological phenomena as they come to occur in everyday settings (Brunswik, 1943; Lewin, 1943). As a field, we still lack an adequate methodological reconciliation that attends to issues of both experimental control and ecological validity. At the same time, there is considerable unease with the perceived “credibility gap” (Levin & O’Donnell, 1999) of much of educational research because it is not produced with what are considered to be scientific methods. From this perspective, the knowledge from educational research has limited usability because it is not trustworthy. An educational psychology that is both usable in a practical sense and scientifically trustworthy cannot proceed without directly studying the phenomena it hopes to explain in its inherent messiness. A little over a decade ago, Brown (1992) described her evolving approach to “design experimentation” as an effort to bridge laboratory studies of learning with studies of complex instructional interventions based on such insights. She showed how insights from the laboratory were inherently limited in their ability to explain or predict learning in the classroom. The challenge, as she saw it, was to develop a methodology of experimenting with intervention designs in situ to develop theories of learning (and teaching) that accounted for the multiple interactions of people acting in a complex social setting. At the same time, Collins (1992) was putting forth a notion of educational research as a “design science,” like aerospace engineering, that required a methodology to systematically test design variants for effectiveness. Achieving such a design science, however, requires a sufficient understanding of the underlying variables at all relevant layers of a complex social system (schooling)—an understanding that we do not yet have (Collins, Joseph, & Bielaczyc, 2004). The last 12 years have seen an increasing uptake of the design experimentation methodology, so much so that a recent handbook on research in math and science education is replete with examples and formulations of the approach (Kelly & Lesh, 2000). The general approach has been called by many names. We have settled on the term design-based research over the other commonly used phrases “design experimentation,” which connotes a specific form of controlled experimentation that does not capture the breadth of the approach, or “design research,” which is too easily confused with research design and other efforts in design fields that lack in situ research components. The approach to research described in this issue is design based in that it is theoretiEDUCATIONAL PSYCHOLOGIST, 39(4), 199–201 Copyright © 2004, Lawrence Erlbaum Associates, Inc.",,
-Learning in context through conflict and alignment: Farmers and scientists in search of sustainable agriculture,J. Eshuis,,,http://library.wur.nl/WebQuery/wurpubs/fulltext/28944
-"Mathematics teachers' ""learning to notice"" in the context of a video club",E. V. Es,,,
-"Organizational Learning and Communities-of-Practice: Toward a Unified View of Working, Learning, and Innovation",J. Brown,"Recent ethnographic studies of workplace practices indicate that the ways people actually work usually differ fundamentally from the ways organizations describe that work in manuals, training programs, organizational charts, and job descriptions. Nevertheless, organizations tend to rely on the latter in their attempts to understand and improve work practice. We examine one such study. We then relate its conclusions to compatible investigations of learning and of innovation to argue that conventional descriptions of jobs mask not only the ways people work, but also significant learning and innovation generated in the informal communities-of-practice in which they work. By reassessing work, learning, and innovation in the context of actual communities and actual practices, we suggest that the connections between these three become apparent. With a unified view of working, learning, and innovating, it should be possible to reconceive of and redesign organizations to improve all three.",,
-Designing learning trajectory of circle using the context of Ferris wheel,N. Fitri,"Ferris wheel is one amusement playground that resembles a giant spinning wheel. Many students are familiar with the Ferris wheel in the mini version of it at night market festivals. This is the potential for learning mathematics. Furthermore, there is a mathematical learning approach called Indonesian Realistic Mathematics Education (IRME) where students learn with contexts which are close to students' life as starting points. Therefore, this study aims to design a learning trajectory using the IRME approach with the Ferris wheel as the context in the learning process to support students' understanding of the learning about circles. The research method is design research that consists of three stages: preliminary design, design experiments, and retrospective analysis. The subjects were 20 eighth-grade students from one of the private Junior High School in Yogyakarta. The instruments used are videos to see the learning process and when students work on the given problems, photos to refer the results of student work, and written test in worksheets to get the data on student's work. The research result explores the learning trajectory practiced using the Ferris wheel as the context seen in the student's daily activities. The learning trajectory consists of four events, namely assembling the Ferris wheel, drawing an illustration of the Ferris wheel, making a list of the circle parts, and solving a problem related to the parts of the circle. Lastly, this study shows that learning trajectory activities have essential roles in supporting students' understanding of the concept of a circle.",,http://journals.ums.ac.id/index.php/jramathedu/article/download/10961/5776
-"Re-Conceptualizing ""Scaffolding"" and the Zone of Proximal Development in the Context of Symmetrical Collaborative Learning.",Juan Manuel Fernández-Cárdenas,"The linked concepts of ‘scaffolding’ and the Zone of Proximal Development (ZPD) were originally applied to the context of asymmetrical teaching and learning with a teacher or adult explicitly supporting a learner, usually a child, to achieve tasks beyond their ability when working alone. In this paper we investigate how these concepts need to be reconceptualised if they are to be applied to the different context of symmetrical learning amongst groups of peers. We present two separate studies. In the first one we analyse the type of talk used by a group of children from Mexico solving the Raven’s Standard Progressive Matrices (RSPM) test together both before and after an intervention programme teaching ‘exploratory talk’. Our analysis demonstrates a ZPD created by the way in which they talk together. In the second study we present the comparison of the talk of two groups of children, one from Mexico and the other from the UK, solving together a single matrix from the RSPM test. Our analysis shows how the concept of ‘scaffolding’ can be applied to understand how these groups of children use language to support shared thinking and learning. In both studies we found that applying ideas of ‘scaffolding’ and the ZPD to symmetrical learning required the re-conceptualisation of these concepts as characterisations of dynamic processes within dialogues.",,
-Bert Transformer model for Detecting Arabic GPT2 Auto-Generated Tweets,F. Harrag,"During the last two decades, we have progressively turned to the Internet and social media to find news, entertain conversations and share opinion. Recently, OpenAI has developed a machine learning system called GPT-2 for Generative Pre-trained Transformer-2, which can produce deepfake texts. It can generate blocks of text based on brief writing prompts that look like they were written by humans, facilitating the spread false or auto-generated text. In line with this progress, and in order to counteract potential dangers, several methods have been proposed for detecting text written by these language models. In this paper, we propose a transfer learning based model that will be able to detect if an Arabic sentence is written by humans or automatically generated by bots. Our dataset is based on tweets from a previous work, which we have crawled and extended using the Twitter API. We used GPT2-Small-Arabic to generate fake Arabic Sentences. For evaluation, we compared different recurrent neural network (RNN) word embeddings based baseline models, namely: LSTM, BI-LSTM, GRU and BI-GRU, with a transformer-based model. Our new transfer-learning model has obtained an accuracy up to 98%. To the best of our knowledge, this work is the first study where ARABERT and GPT2 were combined to detect and classify the Arabic auto-generated texts.","{'model': 'tldr@v2.0.0', 'text': 'This work is the first study where ARABERT and GPT2 were combined to detect and classify the Arabic auto-generated texts, and the new transfer-learning model has obtained an accuracy up to 98%.'}",
-"Operationalizing and Implementing Pretrained, Large Artificial Intelligence Linguistic Models in the US Health Care System: Outlook of Generative Pretrained Transformer 3 (GPT-3) as a Service Model",Emre Sezgin,"Generative pretrained transformer models have been popular recently due to their enhanced capabilities and performance. In contrast to many existing artificial intelligence models, generative pretrained transformer models can perform with very limited training data. Generative pretrained transformer 3 (GPT-3) is one of the latest releases in this pipeline, demonstrating human-like logical and intellectual responses to prompts. Some examples include writing essays, answering complex questions, matching pronouns to their nouns, and conducting sentiment analyses. However, questions remain with regard to its implementation in health care, specifically in terms of operationalization and its use in clinical practice and research. In this viewpoint paper, we briefly introduce GPT-3 and its capabilities and outline considerations for its implementation and operationalization in clinical practice through a use case. The implementation considerations include (1) processing needs and information systems infrastructure, (2) operating costs, (3) model biases, and (4) evaluation metrics. In addition, we outline the following three major operational factors that drive the adoption of GPT-3 in the US health care system: (1) ensuring Health Insurance Portability and Accountability Act compliance, (2) building trust with health care providers, and (3) establishing broader access to the GPT-3 tools. This viewpoint can inform health care practitioners, developers, clinicians, and decision makers toward understanding the use of the powerful artificial intelligence tools integrated into hospital systems and health care.",,https://medinform.jmir.org/2022/2/e32875/PDF
-Developing prompts from large language model for extracting clinical information from pathology and ultrasound reports in breast cancer,Hyeon Seok Choi,"Purpose We aimed to evaluate the time and cost of developing prompts using large language model (LLM), tailored to extract clinical factors in breast cancer patients and their accuracy. Materials and Methods We collected data from reports of surgical pathology and ultrasound from breast cancer patients who underwent radiotherapy from 2020 to 2022. We extracted the information using the Generative Pre-trained Transformer (GPT) for Sheets and Docs extension plugin and termed this the “LLM” method. The time and cost of developing the prompts with LLM methods were assessed and compared with those spent on collecting information with “full manual” and “LLM-assisted manual” methods. To assess accuracy, 340 patients were randomly selected, and the extracted information by LLM method were compared with those collected by “full manual” method. Results Data from 2,931 patients were collected. We developed 12 prompts for Extract function and 12 for Format function to extract and standardize the information. The overall accuracy was 87.7%. For lymphovascular invasion, it was 98.2%. Developing and processing the prompts took 3.5 hours and 15 minutes, respectively. Utilizing the ChatGPT application programming interface cost US $65.8 and when factoring in the estimated wage, the total cost was US $95.4. In an estimated comparison, “LLM-assisted manual” and “LLM” methods were time- and cost-efficient compared to the “full manual” method. Conclusion Developing and facilitating prompts for LLM to derive clinical factors was efficient to extract crucial information from huge medical records. This study demonstrated the potential of the application of natural language processing using LLM model in breast cancer patients. Prompts from the current study can be re-used for other research to collect clinical information.","{'model': 'tldr@v2.0.0', 'text': 'Developing and facilitating prompts for LLM to derive clinical factors was efficient to extract crucial information from huge medical records and can be re-used for other research to collect clinical information.'}",https://www.e-roj.org/upload/pdf/roj-2023-00633.pdf
-SweCTRL-Mini: a data-transparent Transformer-based large language model for controllable text generation in Swedish,Dmytro Kalpakchi,"We present SweCTRL-Mini, a large Swedish language model that can be used for inference and fine-tuning on a single consumer-grade GPU. The model is based on the CTRL architecture by Keskar, McCann, Varshney, Xiong, and Socher (2019), which means that users of the SweCTRL-Mini model can control the genre of the generated text by inserting special tokens in the generation prompts. SweCTRL-Mini is trained on a subset of the Swedish part of the mC4 corpus and a set of Swedish novels. In this article, we provide (1) a detailed account of the utilized training data and text pre-processing steps, to the extent that it is possible to check whether a specific phrase/source was a part of the training data, and (2) an evaluation of the model on both discriminative tasks, using automatic evaluation methods, and generative tasks, using human referees. We also compare the generative capabilities of the model with those of GPT-3. SweCTRL-Mini is fully open and available for download.","{'model': 'tldr@v2.0.0', 'text': 'A detailed account of the utilized training data and text pre-processing steps is provided, to the extent that it is possible to check whether a specific phrase/source was a part of the training data, and an evaluation of the model on both discriminative tasks, using automatic evaluation methods, and generative task, using human referees are provided.'}",http://arxiv.org/pdf/2304.13994
-Reward modeling for mitigating toxicity in transformer-based language models,Farshid Faal,,"{'model': 'tldr@v2.0.0', 'text': 'The experiments demonstrate that the Reinforce-Detoxify method for language model detoxification outperforms existing detoxification approaches in automatic evaluation metrics, indicating that the approach is less prone to unintended bias toward social identities in generated content.'}",https://arxiv.org/pdf/2202.09662
-Optimizing Continuous Prompts for Visual Relationship Detection by Affix-Tuning,Shouguan Xiao,"Visual relationship detection is crucial for understanding visual scenes and is widely used in many areas, including visual navigation, visual question answering, and machine trouble detection. Traditional detection methods often fuse multiple region modules, which takes considerable time and resources to train every module with extensive samples. As every module is independent, the computation process has difficulty achieving unity and lacks a higher level of logical reasonability. In response to the above problems, we propose a novel method of affix-tuning transformers for visual relationship detection tasks, which keeps transformer model parameters frozen and optimizes a small continuous task-specific vector. It not only makes the model unified and reduces the training cost but also maintains the common-sense reasonability without multiscale training. In addition, we design a vision-and-language sentence expression prompt template and train a few transformer model parameters for downstream tasks. Our method, Prompt Template and Affix-Tuning Transformers (PTAT), is evaluated on visual relationship detection and Visual Genome datasets. Finally, the results of the proposed method are close to or even higher than those of the state-of-the-art methods on some evaluation metrics.","{'model': 'tldr@v2.0.0', 'text': 'A novel method of affix-tuning transformers for visual relationship detection tasks, which keeps transformer model parameters frozen and optimizes a small continuous task-specific vector, which makes the model unified and reduces the training cost and maintains the common-sense reasonability without multiscale training.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09815128.pdf
-Contextual Transformer for Offline Meta Reinforcement Learning,Runji Lin,"The pretrain-finetuning paradigm in large-scale sequence models has made significant progress in natural language processing and computer vision tasks. However, such a paradigm is still hindered by several challenges in Reinforcement Learning (RL), including the lack of self-supervised pretraining algorithms based on offline data and efficient fine-tuning/prompt-tuning over unseen downstream tasks. In this work, we explore how prompts can improve sequence modeling-based offline reinforcement learning (offline-RL) algorithms. Firstly, we propose prompt tuning for offline RL, where a context vector sequence is concatenated with the input to guide the conditional policy generation. As such, we can pretrain a model on the offline dataset with self-supervised loss and learn a prompt to guide the policy towards desired actions. Secondly, we extend our framework to Meta-RL settings and propose Contextual Meta Transformer (CMT); CMT leverages the context among different tasks as the prompt to improve generalization on unseen tasks. We conduct extensive experiments across three different offline-RL settings: offline single-agent RL on the D4RL dataset, offline Meta-RL on the MuJoCo benchmark, and offline MARL on the SMAC benchmark. Superior results validate the strong performance, and generality of our methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt tuning for offline RL, where a context vector sequence is concatenated with the input to guide the conditional policy generation and proposes Contextual Meta Transformer (CMT), which leverages the context among different tasks as the prompt to improve generalization on unseen tasks.'}",http://arxiv.org/pdf/2211.08016
-FL-Tuning: Layer Tuning for Feed-Forward Network in Transformer,Jingping Liu,"Prompt tuning is an emerging way of adapting pre-trained language models to downstream tasks. However, the existing studies are mainly to add prompts to the input sequence. This way would not work as expected due to the intermediate multi-head self-attention and feed-forward network computation, making model optimization not very smooth. Hence, we propose a novel tuning way called layer tuning, aiming to add learnable parameters in Transformer layers. Specifically, we focus on layer tuning for feed-forward network in the Transformer, namely FL-tuning. It introduces additional units into the hidden layer of each feed-forward network. We conduct extensive experiments on the public CLUE benchmark. The results show that: 1) Our FL-tuning outperforms prompt tuning methods under both full-data and few-shot settings in almost all cases. In particular, it improves accuracy by 17.93% (full-data setting) on WSC 1.0 and F1 by 16.142% (few-shot setting) on CLUENER over P-tuning v2. 2) Our FL-tuning is more stable and converges about 1.17 times faster than P-tuning v2. 3) With only about 3% of Transformer's parameters to be trained, FL-tuning is comparable with fine-tuning on most datasets, and significantly outperforms fine-tuning (e.g., accuracy improved by 12.9% on WSC 1.1) on several datasets. The source codes are available at https://github.com/genggui001/FL-Tuning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel tuning way called layer tuning, aiming to add learnable parameters in Transformer layers, namely FL-tuning, which introduces additional units into the hidden layer of each feed-forward network in the Transformer.'}",http://arxiv.org/pdf/2206.15312
-Natural Language Processing based Automated Essay Scoring with Parameter-Efficient Transformer Approach,Angad Sethi,"Existing automated scoring models implement layers of traditional recurrent neural networks to achieve reasonable performance. However, the models provide limited performance due to the limited capacity to encode long-term dependencies. The paper proposed a novel architecture incorporating pioneering language models of the natural language processing community. We leverage pre-trained language models and integrate it with adapter modules, which use a bottle-neck architecture to reduce the number of trainable parameters while delivering excellent performance. We also propose a model by re-purposing the bidirectional attention flow model to detect adversarial essays. The model we put forward achieves state-of-the-art performance on most essay prompts in the Automated Student Assessment Prize data set. We outline the previous methods employed to attempt this task, and show how our model outperforms them.","{'model': 'tldr@v2.0.0', 'text': 'A novel architecture incorporating pioneering language models of the natural language processing community is proposed, which leverage pre-trained language models and integrate it with adapter modules, which use a bottle-neck architecture to reduce the number of trainable parameters while delivering excellent performance.'}",
-Context-Based Narrative Generation Transformer (NGen-Transformer),Abraar Raza Samar,"Text generation is an important domain of natural language processing where the plausibility of the generated text depends upon the context assimilation capabilities of the architecture being used. Recently the performance of automatic text generation task have greatly improved with the use of attention based language models. In this paper, we have explored the task of story generation based on some user defined context or prompt. We have proposed a GP2 based narrative generation architecture called NGen-Transformer. Our proposed architectures focuses specifically on the context provided by the user to produce meaningful stories. For the purpose of evaluation of our proposed model, we have used the WritingPrompts dataset which consists of a large number of human written sample stories based on corresponding titles or sentences (prompts). Experimental results show that our proposed NGen- Transformer model outperforms several sequence to sequence as well as attention based architectures on the task of story generation.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed NGen- Transformer model outperforms several sequence to sequence as well as attention based architectures on the task of story generation.'}",
-Chinese text paraphrase recognition based on OpenPrompt introducing hybrid prompts,Chengbo Mao,"The fine-tune paradigm adopted by traditional paraphrase recognition tasks cannot fully exploit the knowledge of pretrained language models (PLMs). At this stage, the Prompt paradigm reconstructs downstream tasks by constructing templates to make it more suitable for the training form of PLM. However, there are many processes, the code base is not supervised, and a single discrete template limits the model prediction ability. In response to this problem, this paper proposes a method of introducing hybrid prompts based on OpenPrompt. OpenPrompt makes the Prompt process have a unified framework, and hybrid prompts solve the problem that discrete templates cannot fully mine PLM knowledge. This paper first constructs a hybrid template with [mask] slots, and then transforms the original input through the template to obtain xprompt , and then xprompt is input to the bert-base-chinese model of the multi-layer bidirectional transformer based on the attention mechanism for training. When the model trains the optimal prompt, calculate the label with the highest probability of filling in the label set, and finally map the label to the prediction result. The experimental results show that compared with the fine-tune paradigm, the F1 value of the prompt paradigm exceeds 4.69% under the same PLM and hyperparameters. Compared with not using soft prompt, the average accuracy and average F1 value obtained with soft hints are 2.294% and 2.31% higher, respectively. Meanwhile, when the number of soft hints is 6, the accuracy and F1 value reach the highest.","{'model': 'tldr@v2.0.0', 'text': 'This paper first constructs a hybrid template with [mask] slots, and then transforms the original input through the template to obtain xprompt, which is input to the bert-base-chinese model of the multi-layer bidirectional transformer based on the attention mechanism for training.'}",
-FP-DETR: DETECTION TRANSFORMER ADVANCED,Yang Cao,"Large-scale pre-training has proven to be effective for visual representation learning on downstream tasks, especially for improving robustness and generalization. However, the recently developed detection transformers only employ pre-training on its backbone while leaving the key component, i.e., a 12-layer transformer, being trained from scratch, which prevents the model from above benefits. This separated training paradigm is mainly caused by the discrepancy between the upstream and downstream tasks. To mitigate the issue, we propose FP-DETR, a new method that Fully Pre-Trains an encoder-only transformer and smoothly finetunes it for object detection via a task adapter. Inspired by the success of textual prompts in NLP, we treat query positional embeddings as visual prompts to help the model attend to the target area (prompting) and recognize the object. To this end, we propose the task adapter which leverages self-attention to model the contextual relation between object query embedding. Experiments on the challenging COCO dataset demonstrate that our FP-DETR achieves competitive performance. Moreover, it enjoys better robustness to common corruptions and generalization to small-size datasets than state-of-the-art detection transformers. Code will be made publicly available at https://github.com/encounter1997/FP-DETR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes FP-DETR, a new method that Fully Pre-Trains an encoder-only transformer and smoothly finetunes it for object detection via a task adapter which leverages self-attention to model the contextual relation between object query embedding.'}",
-Learning to Compress Prompts with Gist Tokens,Jesse Mu,"Prompting is the primary way to utilize the multitask capabilities of language models (LMs), but prompts occupy valuable space in the input context window, and repeatedly encoding the same prompt is computationally inefficient. Finetuning and distillation methods allow for specialization of LMs without prompting, but require retraining the model for each task. To avoid this trade-off entirely, we present gisting, which trains an LM to compress prompts into smaller sets of""gist""tokens which can be cached and reused for compute efficiency. Gist models can be trained with no additional cost over standard instruction finetuning by simply modifying Transformer attention masks to encourage prompt compression. On decoder (LLaMA-7B) and encoder-decoder (FLAN-T5-XXL) LMs, gisting enables up to 26x compression of prompts, resulting in up to 40% FLOPs reductions, 4.2% wall time speedups, and storage savings, all with minimal loss in output quality.","{'model': 'tldr@v2.0.0', 'text': 'Gisting is presented, which trains an LM to compress prompts into smaller sets of ""gist"" tokens which can be cached and reused for compute efficiency, resulting in up to 26x compression of prompts.'}",https://arxiv.org/pdf/2304.08467
-Zero-shot Entity and Tweet Characterization with Designed Conditional Prompts and Contexts,S. Srivatsa,"Online news and social media have been the de facto mediums to disseminate information globally from the beginning of the last decade. However, bias in content and purpose of intentions are not regulated, and managing bias is the responsibility of content consumers. In this regard, understanding the stances and biases of news sources towards specific entities becomes important. To address this problem, we use pretrained language models, which have been shown to bring about good results with no task-specific training or few-shot training. In this work, we approach the prob-lem of characterizing Named Entities and Tweets as an open-ended text classification and open-ended fact probing problem. We evaluate the zero-shot language model capabilities of Generative Pretrained Transformer 2 (GPT-2) to characterize Entities and Tweets subjectively with human psychology-inspired and logical conditional prefixes and contexts. First, we fine-tune the GPT-2 modelon a sufficiently large news corpus and evaluate subjective characterization of popular entities in the corpus by priming with prefixes. Second, we fine-tune GPT-2 with a Tweets corpus from a few popular hashtags and evaluate characterizing tweets by priming the language model with prefixes, questions, and contextual synopsis prompts. Entity characterization results were positive across measures and human evaluation.","{'model': 'tldr@v2.0.0', 'text': 'This work evaluates the zero-shot language model capabilities of Generative Pretrained Transformer 2 (GPT-2) to characterize Entities and Tweets subjectively with human psychology-inspired and logical conditional prefixes and contexts and results were positive across measures and human evaluation.'}",http://arxiv.org/pdf/2204.08405
-Instruction-ViT: Multi-Modal Prompts for Instruction Learning in ViT,Zhe Xiao,"Prompts have been proven to play a crucial role in large language models, and in recent years, vision models have also been using prompts to improve scalability for multiple downstream tasks. In this paper, we focus on adapting prompt design based on instruction tuning into a visual transformer model for image classification which we called Instruction-ViT. The key idea is to implement multi-modal prompts (text or image prompt) related to category information to guide the fine-tuning of the model. Based on the experiments of several image captionining tasks, the performance and domain adaptability were improved. Our work provided an innovative strategy to fuse multi-modal prompts with better performance and faster adaptability for visual classification models.","{'model': 'tldr@v2.0.0', 'text': 'This work provided an innovative strategy to fuse multi-modal prompts with better performance and faster adaptability for visual classification models.'}",http://arxiv.org/pdf/2305.00201
-Clinical Decision Transformer: Intended Treatment Recommendation through Goal Prompting,Seunghyun Lee,"With recent achievements in tasks requiring context awareness, foundation models have been adopted to treat large-scale data from electronic health record (EHR) systems. However, previous clinical recommender systems based on foundation models have a limited purpose of imitating clinicians' behavior and do not directly consider a problem of missing values. In this paper, we propose Clinical Decision Transformer (CDT), a recommender system that generates a sequence of medications to reach a desired range of clinical states given as goal prompts. For this, we conducted goal-conditioned sequencing, which generated a subsequence of treatment history with prepended future goal state, and trained the CDT to model sequential medications required to reach that goal state. For contextual embedding over intra-admission and inter-admissions, we adopted a GPT-based architecture with an admission-wise attention mask and column embedding. In an experiment, we extracted a diabetes dataset from an EHR system, which contained treatment histories of 4788 patients. We observed that the CDT achieved the intended treatment effect according to goal prompt ranges (e.g., NormalA1c, LowerA1c, and HigherA1c), contrary to the case with behavior cloning. To the best of our knowledge, this is the first study to explore clinical recommendations from the perspective of goal prompting. See https://clinical-decision-transformer.github.io for code and additional information.","{'model': 'tldr@v2.0.0', 'text': 'This is the first study to explore clinical recommendations from the perspective of goal prompting, and proposes Clinical Decision Transformer, a recommender system that generates a sequence of medications to reach a desired range of clinical states given as goal prompts.'}",http://arxiv.org/pdf/2302.00612
-Adversarial Transformer Language Models for Contextual Commonsense Inference,Pedro Colon-Hernandez,"Contextualized or discourse aware commonsense inference is the task of generating coherent commonsense assertions (i.e., facts) from a given story, and a particular sentence from that story. Some problems with the task are: lack of controllability for topics of the inferred facts; lack of commonsense knowledge during training; and, possibly, hallucinated or false facts. In this work, we utilize a transformer model for this task and develop techniques to address the aforementioned problems in the task. We control the inference by introducing a new technique we call""hinting"". Hinting is a kind of language model prompting, that utilizes both hard prompts (specific words) and soft prompts (virtual learnable templates). This serves as a control signal to advise the language model""what to talk about"". Next, we establish a methodology for performing joint inference with multiple commonsense knowledge bases. Joint inference of commonsense requires care, because it is imprecise and the level of generality is more flexible. You want to be sure that the results""still make sense""for the context. To this end, we align the textual version of assertions from three knowledge graphs (ConceptNet, ATOMIC2020, and GLUCOSE) with a story and a target sentence. This combination allows us to train a single model to perform joint inference with multiple knowledge graphs. We show experimental results for the three knowledge graphs on joint inference. Our final contribution is exploring a GAN architecture that generates the contextualized commonsense assertions and scores them as to their plausibility through a discriminator. The result is an integrated system for contextual commonsense inference in stories, that can controllably generate plausible commonsense assertions, and takes advantage of joint inference between multiple commonsense knowledge bases.","{'model': 'tldr@v2.0.0', 'text': 'The result is an integrated system for contextual commonsense inference in stories, that can controllably generate plausible commonsense assertions, and takes advantage of joint inference between multiple commonsense knowledge bases.'}",http://arxiv.org/pdf/2302.05406
-"Tryage: Real-time, intelligent Routing of User Prompts to Large Language Models",S. Hari,"The introduction of the transformer architecture and the self-attention mechanism has led to an explosive production of language models trained on specific downstream tasks and data domains. With over 200, 000 models in the Hugging Face ecosystem, users grapple with selecting and optimizing models to suit multifaceted workflows and data domains while addressing computational, security, and recency concerns. There is an urgent need for machine learning frameworks that can eliminate the burden of model selection and customization and unleash the incredible power of the vast emerging model library for end users. Here, we propose a context-aware routing system, Tryage, that leverages a language model router for optimal selection of expert models from a model library based on analysis of individual input prompts. Inspired by the thalamic router in the brain, Tryage employs a perceptive router to predict down-stream model performance on prompts and, then, makes a routing decision using an objective function that integrates performance predictions with user goals and constraints that are incorporated through flags (e.g., model size, model recency). Tryage allows users to explore a Pareto front and automatically trade-off between task accuracy and secondary goals including minimization of model size, recency, security, verbosity, and readability. Across heterogeneous data sets that include code, text, clinical data, and patents, the Tryage framework surpasses Gorilla and GPT3.5 turbo in dynamic model selection identifying the optimal model with an accuracy of 50.9% , compared to 23.6% by GPT 3.5 Turbo and 10.8% by Gorilla. Conceptually, Tryage demonstrates how routing models can be applied to program and control the behavior of multi-model LLM systems to maximize efficient use of the expanding and evolving language model ecosystem.","{'model': 'tldr@v2.0.0', 'text': 'A context-aware routing system that leverages a language model router for optimal selection of expert models from a model library based on analysis of individual input prompts, Tryage demonstrates how routing models can be applied to program and control the behavior of multi-model LLM systems to maximize efficient use of the expanding and evolving language model ecosystem.'}",https://arxiv.org/pdf/2308.11601
-VIMA: Robot Manipulation with Multimodal Prompts,Yunfan Jiang,"Prompt-based learning has emerged as a successful paradigm in natural language processing, where a single general-purpose language model can be instructed to perform any task specified by input prompts. Yet task specification in robotics comes in various forms, such as imitating one-shot demonstrations, following language instructions, and reaching visual goals. They are often considered different tasks and tackled by specialized models. We show that a wide spectrum of robot manipulation tasks can be expressed with multimodal prompts , interleaving textual and visual tokens. Accordingly, we develop a new simulation benchmark that consists of thousands of procedurally-generated tabletop tasks with multimodal prompts, 600K+ expert trajectories for imitation learning, and a four-level evaluation protocol for systematic generalization. We de-sign a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively. VIMA features a recipe that achieves strong model scalability and data efficiency. It outperforms alternative designs in the hardest zero-shot generalization setting by up to 2 . 9 × task success rate given the same training data. With 10 × less training data, VIMA still performs 2 . 7 × better than the best competing variant. Code and video demos are available at vimalabs.github.io .","{'model': 'tldr@v2.0.0', 'text': 'This work develops a new simulation benchmark that consists of thousands of procedurally-generated tabletop tasks with multimodal prompts, 600K+ expert trajectories for imitation learning, and a four-level evaluation protocol for systematic generalization and de-signs a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively.'}",
-Prompt-Based Tuning of Transformer Models for Multi-Center Medical Image Segmentation of Head and Neck Cancer,Numan Saeed,"Medical image segmentation is a vital healthcare endeavor requiring precise and efficient models for appropriate diagnosis and treatment. Vision transformer (ViT)-based segmentation models have shown great performance in accomplishing this task. However, to build a powerful backbone, the self-attention block of ViT requires large-scale pre-training data. The present method of modifying pre-trained models entails updating all or some of the backbone parameters. This paper proposes a novel fine-tuning strategy for adapting a pretrained transformer-based segmentation model on data from a new medical center. This method introduces a small number of learnable parameters, termed prompts, into the input space (less than 1% of model parameters) while keeping the rest of the model parameters frozen. Extensive studies employing data from new unseen medical centers show that the prompt-based fine-tuning of medical segmentation models provides excellent performance regarding the new-center data with a negligible drop regarding the old centers. Additionally, our strategy delivers great accuracy with minimum re-training on new-center data, significantly decreasing the computational and time costs of fine-tuning pre-trained models. Our source code will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel fine-tuning strategy for adapting a pretrained transformer-based segmentation model on data from a new medical center that delivers great accuracy with minimum re-training on new-center data, significantly decreasing the computational and time costs of fine- Tuning pre-trained models.'}",https://www.mdpi.com/2306-5354/10/7/879/pdf?version=1690208147
-TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting,Defu Cao,"The past decade has witnessed significant advances in time series modeling with deep learning. While achieving state-of-the-art results, the best-performing architectures vary highly across applications and domains. Meanwhile, for natural language processing, the Generative Pre-trained Transformer (GPT) has demonstrated impressive performance via training one general-purpose model across various textual datasets. It is intriguing to explore whether GPT-type architectures can be effective for time series, capturing the intrinsic dynamic attributes and leading to significant accuracy improvements. In this paper, we propose a novel framework, TEMPO, that can effectively learn time series representations. We focus on utilizing two essential inductive biases of the time series task for pre-trained models: (i) decomposition of the complex interaction between trend, seasonal and residual components; and (ii) introducing the selection-based prompts to facilitate distribution adaptation in non-stationary time series. TEMPO expands the capability for dynamically modeling real-world temporal phenomena from data within diverse domains. Our experiments demonstrate the superior performance of TEMPO over state-of-the-art methods on a number of time series benchmark datasets. This performance gain is observed not only in standard supervised learning settings but also in scenarios involving previously unseen datasets as well as in scenarios with multi-modal inputs. This compelling finding highlights TEMPO's potential to constitute a foundational model-building framework.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, TEMPO, that can effectively learn time series representations by utilizing two essential inductive biases of the time series task for pre-trained models, and introducing the selection-based prompts to facilitate distribution adaptation in non-stationary time series.'}",https://arxiv.org/pdf/2310.04948
-BatGPT: A Bidirectional Autoregessive Talker from Generative Pre-trained Transformer,Z. Li,"BatGPT is a large-scale language model designed and trained jointly by Wuhan University and Shanghai Jiao Tong University. It is capable of generating highly natural and fluent text in response to various types of input, including text prompts, images, and audio. In the modeling level, we employ a bidirectional autoregressive architecture that allows the model to efficiently capture the complex dependencies of natural language, making it highly effective in tasks such as language generation, dialog systems, and question answering. Moreover, the bidirectional autoregressive modeling not only operates from left to right but also from right to left, effectively reducing fixed memory effects and alleviating model hallucinations. In the training aspect, we propose a novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance. Overall, these approaches significantly improve the effectiveness of BatGPT, and the model can be utilized for a wide range of natural language applications.","{'model': 'tldr@v2.0.0', 'text': ""A novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance significantly improve the effectiveness of BatGPT.""}",https://arxiv.org/pdf/2307.00360
-Prompt-Engineering and Transformer-based Question Generation and Evaluation,Rubaba Amyeen,"Question generation has numerous applications in the educational context. Question generation can prove helpful for students when reviewing content and testing themselves. Furthermore, a question generation model can aid teachers by lessening the burden of creating assessments and other practice material. This paper aims to find the best method to generate questions from textual data through a transformer model and prompt engineering. In this research, we finetuned a pretrained distilBERT model on the SQuAD question answering dataset to generate questions. In addition to training a transformer model, prompt engineering was applied to generate questions effectively using the LLaMA model. The generated questions were compared against the baseline questions in the SQuAD dataset to evaluate the effectiveness of four different prompts. All four prompts demonstrated over 60% similarity on average. Of the prompt-generated questions, 30% achieved a high similarity score greater than 70%.","{'model': 'tldr@v2.0.0', 'text': 'This research finetuned a pretrained distilBERT model on the SQuAD question answering dataset to generate questions using the LLaMA model and prompt engineering, and evaluated the effectiveness of four different prompts.'}",
-Prompt Guided Transformer for Multi-Task Dense Prediction,Yuxiang Lu,"Task-conditional architecture offers advantage in parameter efficiency but falls short in performance compared to state-of-the-art multi-decoder methods. How to trade off performance and model parameters is an important and difficult problem. In this paper, we introduce a simple and lightweight task-conditional model called Prompt Guided Transformer (PGT) to optimize this challenge. Our approach designs a Prompt-conditioned Transformer block, which incorporates task-specific prompts in the self-attention mechanism to achieve global dependency modeling and parameter-efficient feature adaptation across multiple tasks. This block is integrated into both the shared encoder and decoder, enhancing the capture of intra- and inter-task features. Moreover, we design a lightweight decoder to further reduce parameter usage, which accounts for only 2.7% of the total model parameters. Extensive experiments on two multi-task dense prediction benchmarks, PASCAL-Context and NYUD-v2, demonstrate that our approach achieves state-of-the-art results among task-conditional methods while using fewer parameters, and maintains a significant balance between performance and parameter size.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a simple and lightweight task-conditional model called Prompt Guided Transformer (PGT), which incorporates task-specific prompts in the self-attention mechanism to achieve global dependency modeling and parameter-efficient feature adaptation across multiple tasks.'}",https://arxiv.org/pdf/2307.15362
-Small-scale proxies for large-scale Transformer training instabilities,Mitchell Wortsman,"Teams that have trained large Transformer-based models have reported training instabilities at large scale that did not appear when training with the same hyperparameters at smaller scales. Although the causes of such instabilities are of scientific interest, the amount of resources required to reproduce them has made investigation difficult. In this work, we seek ways to reproduce and study training stability and instability at smaller scales. First, we focus on two sources of training instability described in previous work: the growth of logits in attention layers (Dehghani et al., 2023) and divergence of the output logits from the log probabilities (Chowdhery et al., 2022). By measuring the relationship between learning rate and loss across scales, we show that these instabilities also appear in small models when training at high learning rates, and that mitigations previously employed at large scales are equally effective in this regime. This prompts us to investigate the extent to which other known optimizer and model interventions influence the sensitivity of the final loss to changes in the learning rate. To this end, we study methods such as warm-up, weight decay, and the $\mu$Param (Yang et al., 2022), and combine techniques to train small models that achieve similar losses across orders of magnitude of learning rate variation. Finally, to conclude our exploration we study two cases where instabilities can be predicted before they emerge by examining the scaling behavior of model activation and gradient norms.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the extent to which other known optimizer and model interventions influence the sensitivity of the final loss to changes in the learning rate, and studies two cases where instabilities can be predicted before they emerge by examining the scaling behavior of model activation and gradient norms.'}",https://arxiv.org/pdf/2309.14322
-Crosslingual Content Scoring in Five Languages Using Machine-Translation and Multilingual Transformer Models,Andrea Horbach,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that a combination of machine translation and multilingual models outperforms each method individually - the best results are reached when combining the available data in different languages, i.e. first training a model on the large English ASAP dataset before fine-tuning on smaller amounts of training data in the target language.'}",https://link.springer.com/content/pdf/10.1007/s40593-023-00370-1.pdf
-Multimodal Transformer-Based Lyric Generation from MIDI and Text Data,Anthony Zhan,"Current generative language models are designed to take a text prompt as input, and output the highest probability continuation. However, when applying this idea to the design space of lyric generation, there is an obvious shortcoming — lyrics are often created with music’s auditory features in mind, such as tempo, tonality, melody, rhythm, etc. Which is to say, a model able to interpret both text prompts and snippets of music may be able to produce much better music. Therefore, our project seeks to combine both text and music data in a multimodal transformerbased model in order to generate lyrics more effectively than a purely text-based model, utilizing multiple transformer architectures, including MusicBERT and GPT-2.",,
-Study on a Real-time Work Assistance Method for Substation Based on Petri Net Model,Qi Gao,"Work assistance is one of the important means to improve work efficiency. This paper focuses on substation work and proposes a real-time work assistance method based on petri net model. The work assistance includes two aspects, termed as, workflow guidance and operation prompt. Particularly, the workflow constructed by the petri net includes sequence, inhibition, concurrency and conflict structures to declare the correlations between work steps. Further, different forms of operation prompts at each step are designed, considering real-time responses of workers including skip, replay, pause and restart. The effectiveness of the method is well validated by an application of the work assistance on measuring the grounding current of transformer core and clamp. It is demonstrated that the unified assistance method built by the petri net successfully instructs standardized working process, helps avoid potential risks, and provides flexible operation prompts in real time.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the unified assistance method built by the petri net successfully instructs standardized working process, helps avoid potential risks, and provides flexible operation prompts in real time.'}",
-Efficient Model Personalization in Federated Learning via Client-Specific Prompt Generation,Fu-En Yang,"Federated learning (FL) emerges as a decentralized learning framework which trains models from multiple distributed clients without sharing their data to preserve privacy. Recently, large-scale pre-trained models (e.g., Vision Transformer) have shown a strong capability of deriving robust representations. However, the data heterogeneity among clients, the limited computation resources, and the communication bandwidth restrict the deployment of large-scale models in FL frameworks. To leverage robust representations from large-scale models while enabling efficient model personalization for heterogeneous clients, we propose a novel personalized FL framework of client-specific Prompt Generation (pFedPG), which learns to deploy a personalized prompt generator at the server for producing client-specific visual prompts that efficiently adapts frozen backbones to local data distributions. Our proposed framework jointly optimizes the stages of personalized prompt adaptation locally and personalized prompt generation globally. The former aims to train visual prompts that adapt foundation models to each client, while the latter observes local optimization directions to generate personalized prompts for all clients. Through extensive experiments on benchmark datasets, we show that our pFedPG is favorable against state-of-the-art personalized FL methods under various types of data heterogeneity, allowing computation and communication efficient model personalization.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel personalized FL framework of client-specific Prompt Generation (pFedPG), which learns to deploy a personalized prompt generator at the server for producing client- specific visual prompts that efficiently adapts frozen backbones to local data distributions.'}",https://arxiv.org/pdf/2308.15367
-How interesting and coherent are the stories generated by a large‐scale neural language model? Comparing human and automatic evaluations of machine‐generated text,Dominic Callan,"Evaluation of the narrative text generated by machines has traditionally been a challenge, particularly when attempting to evaluate subjective elements such as interest or believability. Recent improvements in narrative machine text generation have been largely driven by the emergence of transformer‐based language models, trained on massive quantities of data, resulting in higher quality text generation. In this study, a corpus of stories is generated using the pre‐trained GPT‐Neo transformer model, with human‐written prompts as inputs upon which to base the narrative text. The stories generated through this process are subsequently evaluated through both human evaluation and two automated metrics: BERTScore and BERT Next Sentence Prediction, with the aim of determining whether there is a correlation between the automatic scores and the human judgements. The results show variation in human evaluation results in comparison to modern automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.","{'model': 'tldr@v2.0.0', 'text': 'The results show variation in human evaluation results in comparison to modern automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/exsy.13292
-Kosmos-2.5: A Multimodal Literate Model,Tengchao Lv,"We present Kosmos-2.5, a multimodal literate model for machine reading of text-intensive images. Pre-trained on large-scale text-intensive images, Kosmos-2.5 excels in two distinct yet cooperative transcription tasks: (1) generating spatially-aware text blocks, where each block of text is assigned its spatial coordinates within the image, and (2) producing structured text output that captures styles and structures into the markdown format. This unified multimodal literate capability is achieved through a shared Transformer architecture, task-specific prompts, and flexible text representations. We evaluate Kosmos-2.5 on end-to-end document-level text recognition and image-to-markdown text generation. Furthermore, the model can be readily adapted for any text-intensive image understanding task with different prompts through supervised fine-tuning, making it a general-purpose tool for real-world applications involving text-rich images. This work also paves the way for the future scaling of multimodal large language models.","{'model': 'tldr@v2.0.0', 'text': 'Kosmos-2.5 can be readily adapted for any text-intensive image understanding task with different prompts through supervised fine-tuning, making it a general-purpose tool for real-world applications involving text-rich images and paves the way for the future scaling of multimodal large language models.'}",https://arxiv.org/pdf/2309.11419
-Automated Reading Passage Generation with OpenAI's Large Language Model,Ummugul Bezirhan,"The widespread usage of computer-based assessments and individualized learning platforms has resulted in an increased demand for the rapid production of high-quality items. Automated item generation (AIG), the process of using item models to generate new items with the help of computer technology, was proposed to reduce reliance on human subject experts at each step of the process. AIG has been used in test development for some time. Still, the use of machine learning algorithms has introduced the potential to improve the efficiency and effectiveness of the process greatly. The approach presented in this paper utilizes OpenAI's latest transformer-based language model, GPT-3, to generate reading passages. Existing reading passages were used in carefully engineered prompts to ensure the AI-generated text has similar content and structure to a fourth-grade reading passage. For each prompt, we generated multiple passages, the final passage was selected according to the Lexile score agreement with the original passage. In the final round, the selected passage went through a simple revision by a human editor to ensure the text was free of any grammatical and factual errors. All AI-generated passages, along with original passages were evaluated by human judges according to their coherence, appropriateness to fourth graders, and readability.","{'model': 'tldr@v2.0.0', 'text': ""The approach presented in this paper utilizes OpenAI's latest transformer-based language model, GPT-3, to generate reading passages that were evaluated by human judges according to their coherence, appropriateness to fourth graders, and readability.""}",http://arxiv.org/pdf/2304.04616
-Multilingual Few-Shot Learning via Language Model Retrieval,Genta Indra Winata,"Transformer-based language models have achieved remarkable success in few-shot in-context learning and drawn a lot of research interest. However, these models' performance greatly depends on the choice of the example prompts and also has high variability depending on how samples are chosen. In this paper, we conduct a comprehensive study of retrieving semantically similar few-shot samples and using them as the context, as it helps the model decide the correct label without any gradient update in the multilingual and cross-lingual settings. We evaluate the proposed method on five natural language understanding datasets related to intent detection, question classification, sentiment analysis, and topic classification. The proposed method consistently outperforms random sampling in monolingual and cross-lingual tasks in non-English languages.","{'model': 'tldr@v2.0.0', 'text': 'This paper conducts a comprehensive study of retrieving semantically similar few-shot samples and using them as the context, as it helps the model decide the correct label without any gradient update in the multilingual and cross-lingual settings.'}",http://arxiv.org/pdf/2306.10964
-Prompt Tuning GPT-2 language model for parameter-efficient domain adaptation of ASR systems,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains creating a need to adapt to new domains with small memory and deployment overhead. In this work, we introduce domain-prompts, a methodology that involves training a small number of domain embedding parameters to prime a Transformer-based Language Model (LM) to a particular domain. Using this domain-adapted LM for rescoring ASR hypotheses can achieve 7-13% WER reduction for a new domain with just 1000 unlabeled textual domain-specific sentences. This improvement is comparable or even better than fully fine-tuned models even though just 0.02% of the parameters of the base LM are updated. Additionally, our method is deployment-friendly as the learnt domain embeddings are prefixed to the input to the model rather than changing the base model architecture. Therefore, our method is an ideal choice for on-the-fly adaptation of LMs used in ASR systems to progressively scale it to new domains.","{'model': 'tldr@v2.0.0', 'text': None}",
-Phenaki: Variable Length Video Generation From Open Domain Textual Description,Ruben Villegas,"We present Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts. Generating videos from text is particularly challenging due to the computational cost, limited quantities of high quality text-video data and variable length of videos. To address these issues, we introduce a new model for learning video representation which compresses the video to a small representation of discrete tokens. This tokenizer uses causal attention in time, which allows it to work with variable-length videos. To generate video tokens from text we are using a bidirectional masked transformer conditioned on pre-computed text tokens. The generated video tokens are subsequently de-tokenized to create the actual video. To address data issues, we demonstrate how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets. Compared to the previous video generation methods, Phenaki can generate arbitrary long videos conditioned on a sequence of prompts (i.e. time variable text or a story) in open domain. To the best of our knowledge, this is the first time a paper studies generating videos from time variable prompts. In addition, compared to the per-frame baselines, the proposed video encoder-decoder computes fewer tokens per video but results in better spatio-temporal consistency.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts, and demonstrates how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets.'}",http://arxiv.org/pdf/2210.02399
-LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-init Attention,Renrui Zhang,"We present LLaMA-Adapter, a lightweight adaption method to efficiently fine-tune LLaMA into an instruction-following model. Using 52K self-instruct demonstrations, LLaMA-Adapter only introduces 1.2M learnable parameters upon the frozen LLaMA 7B model, and costs less than one hour for fine-tuning on 8 A100 GPUs. Specifically, we adopt a set of learnable adaption prompts, and prepend them to the word tokens at higher transformer layers. Then, a zero-initialized attention mechanism with zero gating is proposed, which adaptively injects the new instructional cues into LLaMA, while effectively preserves its pre-trained knowledge. With our efficient training, LLaMA-Adapter can generate high-quality responses, comparable to Alpaca with fully fine-tuned 7B parameters. Besides language commands, our approach can be simply extended to multi-modal instructions for learning image-conditioned LLaMA model, which achieves superior reasoning performance on ScienceQA and COCO Caption benchmarks. Furthermore, we also evaluate the zero-initialized attention mechanism for fine-tuning other pre-trained models (ViT, RoBERTa) on traditional vision and language tasks, demonstrating the superior generalization capacity of our approach. Code is released at https://github.com/OpenGVLab/LLaMA-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'A zero-initialized attention mechanism with zero gating is proposed, which adaptively injects the new instructional cues into LLaMA, while effectively preserves its pre-trained knowledge on traditional vision and language tasks, demonstrating the superior generalization capacity of the approach.'}",http://arxiv.org/pdf/2303.16199
-"ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations",Tirth Dave,"This paper presents an analysis of the advantages, limitations, ethical considerations, future prospects, and practical applications of ChatGPT and artificial intelligence (AI) in the healthcare and medical domains. ChatGPT is an advanced language model that uses deep learning techniques to produce human-like responses to natural language inputs. It is part of the family of generative pre-training transformer (GPT) models developed by OpenAI and is currently one of the largest publicly available language models. ChatGPT is capable of capturing the nuances and intricacies of human language, allowing it to generate appropriate and contextually relevant responses across a broad spectrum of prompts. The potential applications of ChatGPT in the medical field range from identifying potential research topics to assisting professionals in clinical and laboratory diagnosis. Additionally, it can be used to help medical students, doctors, nurses, and all members of the healthcare fraternity to know about updates and new developments in their respective fields. The development of virtual assistants to aid patients in managing their health is another important application of ChatGPT in medicine. Despite its potential applications, the use of ChatGPT and other AI tools in medical writing also poses ethical and legal concerns. These include possible infringement of copyright laws, medico-legal complications, and the need for transparency in AI-generated content. In conclusion, ChatGPT has several potential applications in the medical and healthcare fields. However, these applications come with several limitations and ethical considerations which are presented in detail along with future prospects in medicine and healthcare.","{'model': 'tldr@v2.0.0', 'text': 'An analysis of the advantages, limitations, ethical considerations, future prospects, and practical applications of ChatGPT and artificial intelligence (AI) in the healthcare and medical domains shows several potential applications come with several limitations and ethical considerations.'}",https://www.frontiersin.org/articles/10.3389/frai.2023.1169595/pdf
-Synthetic Data from Diffusion Models Improves ImageNet Classification,Shekoofeh Azizi,"Deep generative models are becoming increasingly powerful, now generating diverse high fidelity photo-realistic samples given text prompts. Have they reached the point where models of natural images can be used for generative data augmentation, helping to improve challenging discriminative tasks? We show that large-scale text-to image diffusion models can be fine-tuned to produce class conditional models with SOTA FID (1.76 at 256x256 resolution) and Inception Score (239 at 256x256). The model also yields a new SOTA in Classification Accuracy Scores (64.96 for 256x256 generative samples, improving to 69.24 for 1024x1024 samples). Augmenting the ImageNet training set with samples from the resulting models yields significant improvements in ImageNet classification accuracy over strong ResNet and Vision Transformer baselines.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that large-scale text-to image diffusion models can be fine-tuned to produce class conditional models with SOTA FID and Inception Score that yields significant improvements in ImageNet classification accuracy over strong ResNet and Vision Transformer baselines.'}",http://arxiv.org/pdf/2304.08466
-Evaluating Academic Answers Generated Using ChatGPT,S. Fergus,"The integration of technology in education has become ever more prioritized since the COVID-19 pandemic. Chat Generative Pre-Trained Transformer (ChatGPT) is an artificial intelligence technology that generates conversational interactions to user prompts. The trained model can answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. The functionality of ChatGPT in answering chemistry assessment questions requires investigation to ascertain its potential impact on learning and assessment. Two chemistry-focused modules in year 1 and year 2 of a pharmaceutical science program are used to study and evaluate ChatGPT-generated responses in relation to the end-of-year exam assessments. For questions that focused on knowledge and understanding with ""describe” and ""discuss” verbs, the ChatGPT generated responses. For questions that focused on application of knowledge and interpretation with nontext information, the ChatGPT technology reached a limitation. A further analysis of the quality of responses is reported in this study. ChatGPT is not considered a high-risk technology tool in relation to cheating. Similar to the COVID-19 disruption, ChatGPT is expected to provide a catalyst for educational discussions on academic integrity and assessment design. © 2023 The Authors. Published by American Chemical Society and Division of Chemical Education, Inc.",,
-Memory Augmented Large Language Models are Computationally Universal,Dale Schuurmans,"We show that transformer-based large language models are computationally universal when augmented with an external memory. Any deterministic language model that conditions on strings of bounded length is equivalent to a finite automaton, hence computationally limited. However, augmenting such models with a read-write memory creates the possibility of processing arbitrarily large inputs and, potentially, simulating any algorithm. We establish that an existing large language model, Flan-U-PaLM 540B, can be combined with an associative read-write memory to exactly simulate the execution of a universal Turing machine, $U_{15,2}$. A key aspect of the finding is that it does not require any modification of the language model weights. Instead, the construction relies solely on designing a form of stored instruction computer that can subsequently be programmed with a specific set of prompts.","{'model': 'tldr@v2.0.0', 'text': 'It is established that an existing large language model, Flan-U-PaLM 540B, can be combined with an associative read-write memory to exactly simulate the execution of a universal Turing machine, $U_{15,2}$.'}",http://arxiv.org/pdf/2301.04589
-Adapting Language Models to Compress Contexts,Alexis Chevalier,"Transformer-based language models (LMs) are powerful and widely-applicable tools, but their usefulness is constrained by a finite context window and the expensive computational cost of processing long text documents. We propose to adapt pre-trained LMs into AutoCompressors. These models are capable of compressing long contexts into compact summary vectors, which are then accessible to the model as soft prompts. Summary vectors are trained with an unsupervised objective, whereby long documents are processed in segments and summary vectors from all previous segments are used in language modeling. We fine-tune OPT models on sequences of up to 30,720 tokens and show that AutoCompressors can utilize long contexts to improve perplexity. We evaluate AutoCompressors on in-context learning by compressing task demonstrations. We find that summary vectors are good substitutes for plain-text demonstrations, increasing accuracy while reducing inference cost. Finally, we explore the benefits of pre-computing summary vectors for large corpora by applying summary vectors to retrieval-augmented language modeling. Overall, AutoCompressors emerge as a simple and inexpensive solution for extending the context window of LMs while speeding up inference over long contexts.","{'model': 'tldr@v2.0.0', 'text': 'AutoCompressors emerge as a simple and inexpensive solution for extending the context window of LMs while speeding up inference over long contexts and the benefits of pre-computing summary vectors for large corpora are explored.'}",http://arxiv.org/pdf/2305.14788
-Learning to Learn with Generative Models of Neural Network Checkpoints,William S. Peebles,"We explore a data-driven approach for learning to optimize neural networks. We construct a dataset of neural network checkpoints and train a generative model on the parameters. In particular, our model is a conditional diffusion transformer that, given an initial input parameter vector and a prompted loss, error, or return, predicts the distribution over parameter updates that achieve the desired metric. At test time, it can optimize neural networks with unseen parameters for downstream tasks in just one update. We find that our approach successfully generates parameters for a wide range of loss prompts. Moreover, it can sample multimodal parameter solutions and has favorable scaling properties. We apply our method to different neural network architectures and tasks in supervised and reinforcement learning.","{'model': 'tldr@v2.0.0', 'text': 'This model is a conditional diffusion transformer that, given an initial input parameter vector and a prompted loss, error, or return, predicts the distribution over parameter updates that achieve the desired metric.'}",http://arxiv.org/pdf/2209.12892
-In-Context Learning of Large Language Models Explained as Kernel Regression,Chi Han,"Large language models (LLMs) have initiated a paradigm shift in transfer learning. In contrast to the classic pretraining-then-finetuning procedure, in order to use LLMs for downstream prediction tasks, one only needs to provide a few demonstrations, known as in-context examples, without adding more or updating existing model parameters. This in-context learning (ICL) capabilities of LLMs is intriguing, and it is not yet fully understood how pretrained LLMs acquire such capabilities. In this paper, we investigate the reason why a transformer-based language model can accomplish in-context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression algorithms when faced with in-context examples. More concretely, we first prove that Bayesian inference on in-context prompts can be asymptotically understood as kernel regression ˆ y = (cid:80) i y i K ( x,x i ) (cid:80) i K ( x,x i ) as the number of in-context demonstrations grows. Then, we empirically investigate the in-context behaviors of language models. We find that during ICL, the attentions and hidden features in LLMs match the behaviors of a kernel regression. Finally, our theory provides insights on multiple phenomena observed in ICL field: why retrieving demonstrative samples similar to test sample can help, why ICL performance is sensitive to the output formats, and why ICL accuracy benefits from selecting in-distribuion and representative samples. We will make our code available to the research community following publication.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the reason why a transformer-based language model can accomplish in- Context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression algorithms when faced with in-context examples and empirically investigates the in- context behaviors of language models.'}",https://arxiv.org/pdf/2305.12766
-Prompt Tuning of Deep Neural Networks for Speaker-adaptive Visual Speech Recognition,Minsu Kim,"Visual Speech Recognition (VSR) aims to infer speech into text depending on lip movements alone. As it focuses on visual information to model the speech, its performance is inherently sensitive to personal lip appearances and movements, and this makes the VSR models show degraded performance when they are applied to unseen speakers. In this paper, to remedy the performance degradation of the VSR model on unseen speakers, we propose prompt tuning methods of Deep Neural Networks (DNNs) for speaker-adaptive VSR. Specifically, motivated by recent advances in Natural Language Processing (NLP), we finetune prompts on adaptation data of target speakers instead of modifying the pre-trained model parameters. Different from the previous prompt tuning methods mainly limited to Transformer variant architecture, we explore different types of prompts, the addition, the padding, and the concatenation form prompts that can be applied to the VSR model which is composed of CNN and Transformer in general. With the proposed prompt tuning, we show that the performance of the pre-trained VSR model on unseen speakers can be largely improved by using a small amount of adaptation data (e.g., less than 5 minutes), even if the pre-trained model is already developed with large speaker variations. Moreover, by analyzing the performance and parameters of different types of prompts, we investigate when the prompt tuning is preferred over the finetuning methods. The effectiveness of the proposed method is evaluated on both word- and sentence-level VSR databases, LRW-ID and GRID.","{'model': 'tldr@v2.0.0', 'text': 'The proposed prompt tuning methods of Deep Neural Networks for speaker-adaptive VSR are proposed and it is shown that the performance of the pre-trained VSR model on unseen speakers can be largely improved by using a small amount of adaptation data, even if thePre-trained model is already developed with large speaker variations.'}",http://arxiv.org/pdf/2302.08102
-Radiology Gets Chatty: The ChatGPT Saga Unfolds,H. Grewal,"As artificial intelligence (AI) continues to evolve and mature, it is increasingly finding applications in the field of healthcare, particularly in specialties like radiology that are data-heavy and image-focused. Language learning models (LLMs) such as OpenAI's Generative Pre-trained Transformer-4 (GPT-4) are new in the field of medicine and there is a paucity of literature regarding the possible utilities of GPT-4 given its novelty. We aim to present an in-depth exploration of the role of GPT-4, an advanced language model, in radiology. Giving the GPT-4 model prompts for generating reports, template generation, enhancing clinical decision-making, and suggesting captivating titles for research articles, patient communication, and education, can occasionally be quite generic, and at times, it may present factually incorrect content, which could lead to errors. The responses were then analyzed in detail regarding their potential utility in day-to-day radiologist workflow, patient education, and research processes. Further research is required to evaluate LLMs' accuracy and safety in clinical practice and to develop comprehensive guidelines for their implementation.","{'model': 'tldr@v2.0.0', 'text': 'An in-depth exploration of the role of GPT-4, an advanced language model, in radiology is presented and responses were analyzed in detail regarding their potential utility in day-to-day radiologist workflow, patient education, and research processes.'}",https://assets.cureus.com/uploads/original_article/pdf/161200/20230708-6319-1gaxi52.pdf
-Multimodal Prompting with Missing Modalities for Visual Recognition,Yi-Lun Lee,"In this paper, we tackle two challenges in multimodal learning for visual recognition: 1) when missing-modality occurs either during training or testing in real-world situations; and 2) when the computation resources are not available to finetune on heavy transformer models. To this end, we propose to utilize prompt learning and mitigate the above two challenges together. Specifically, our modality-missing-aware prompts can be plugged into multimodal transformers to handle general missing-modality cases, while only requiring less than 1% learnable parameters compared to training the entire model. We further explore the effect of different prompt configurations and analyze the robustness to missing modality. Extensive experiments are conducted to show the effectiveness of our prompt learning framework that improves the performance under various missing-modality cases, while alleviating the requirement of heavy model retraining. Code is available.11https://github.com/YiLunLee/missing_aware_prompts","{'model': 'tldr@v2.0.0', 'text': 'Modality-missing-aware prompts can be plugged into multimodal transformers to handle general missing-modality cases, while only requiring less than 1% learnable parameters compared to training the entire model.'}",https://arxiv.org/pdf/2303.03369
-MaskSketch: Unpaired Structure-guided Masked Image Generation,D. Bashkirova,"Recent conditional image generation methods produce images of remarkable diversity, fidelity and realism. However, the majority of these methods allow conditioning only on labels or text prompts, which limits their level of control over the generation result. In this paper, we introduce MaskSketch, an image generation method that allows spatial conditioning of the generation result using a guiding sketch as an extra conditioning signal during sampling. MaskSketch utilizes a pretrained masked generative transformer, requiring no model training or paired supervision, and works with input sketches of different levels of abstraction. We show that intermediate self-attention maps of a masked generative transformer encode important structural information of the input image, such as scene layout and object shape, and we propose a novel sampling method based on this observation to enable structure-guided generation. Our results show that MaskSketch achieves high image realism and fidelity to the guiding structure. Evaluated on standard benchmark datasets, MaskSketch outperforms state-of-the-art methods for sketch-to-image translation, as well as unpaired image-to-image translation approaches. The code can be found on our project website: https://masksketch.github.io/","{'model': 'tldr@v2.0.0', 'text': 'The paper shows that intermediate self-attention maps of a masked generative transformer encode important structural information of the input image, such as scene layout and object shape, and proposes a novel sampling method based on this observation to enable structure-guided generation in MaskSketch.'}",https://arxiv.org/pdf/2302.05496
-"How could ChatGPT impact my practice as an intensivist? An overview of potential applications, risks and limitations",M. Komorowski,,"{'model': 'tldr@v2.0.0', 'text': 'Transformers are an architecture used primarily in the field of natural language processing, that aims to solve sequence-to-sequence tasks while handling long-range dependencies with ease and provide superior answers while requiring less training time than previous approaches such as recurrent neural networks.'}",https://link.springer.com/content/pdf/10.1007/s00134-023-07096-7.pdf
-A Brief Analysis of “ChatGPT” – A Revolutionary Tool Designed by OpenAI,Md. Asraful Haque,"The ChatGPT, a powerful conversational tool trained by OpenAI is considered to be a revolutionary model in the field of artificial intelligence and natural language processing. It has received a lot of attention because of its potential to automate a variety of tasks and possibly have an impact on sectors like translation, customer service, and content creation. It uses GPT-3 (Generative Pre-training Transformer 3) language model to process user queries. GPT-3 has been trained on a very large dataset, which includes a wide range of texts from the internet and other sources. This has given it a broad knowledge base and has allowed it to generate responses to a wide range of prompts that are coherent and human-like. GPT-3 is one of the largest and most powerful language models to date, and it has the ability to perform a wide range of natural language processing tasks. After its release, ChatGPT has become a trending tool for the public to experiment and explore what it is capable of. In this article, we want to clarify what ChatGPT is. How does it work? What makes it different from other chatbots or search engines like Google? What are the major challenges and future prospects for it?","{'model': 'tldr@v2.0.0', 'text': 'The ChatGPT, a powerful conversational tool trained by OpenAI, uses GPT-3 (Generative Pre-training Transformer 3) language model to process user queries and has the ability to perform a wide range of natural language processing tasks.'}",https://publications.eai.eu/index.php/airo/article/download/2983/2322
-PUMA: Secure Inference of LLaMA-7B in Five Minutes,Ye Dong,"With ChatGPT as a representative, tons of companies have began to provide services based on large Transformers models. However, using such a service inevitably leak users' prompts to the model provider. Previous studies have studied secure inference for Transformer models using secure multiparty computation (MPC), where model parameters and clients' prompts are kept secret. Despite this, these frameworks are still limited in terms of model performance, efficiency, and deployment. To address these limitations, we propose framework PUMA to enable fast and secure Transformer model inference. Our framework designs high quality approximations for expensive functions such as GeLU and softmax, and significantly reduce the cost of secure inference while preserving the model performance. Additionally, we design secure Embedding and LayerNorm procedures that faithfully implement the desired functionality without undermining the Transformer architecture. PUMA is about $2\times$ faster than the state-of-the-art framework MPCFORMER(ICLR 2023) and has similar accuracy as plaintext models without fine-tuning (which the previous works failed to achieve). PUMA can even evaluate LLaMA-7B in around 5 minutes to generate 1 token. To our best knowledge, this is the first time that a model with such a parameter size is able to be evaluated under MPC. PUMA has been open-sourced in the Github repository of SecretFlow-SPU.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes framework PUMA to enable fast and secure Transformer model inference, and designs high quality approximations for expensive functions such as GeLU and softmax, and significantly reduce the cost of secure inference while preserving the model performance.'}",https://arxiv.org/pdf/2307.12533
-À-la-carte Prompt Tuning (APT): Combining Distinct Data Via Composable Prompting,Benjamin Bowman,"We introduce À-la-carte Prompt Tuning (APT), a transformer-based scheme to tune prompts on distinct data so that they can be arbitrarily composed at inference time. The individual prompts can be trained in isolation, possibly on different devices, at different times, and on different distributions or domains. Furthermore each prompt only contains information about the subset of data it was exposed to during training. During inference, models can be assembled based on arbitrary selections of data sources, which we call à-la-carte learning. À-la-carte learning enables constructing bespoke models specific to each user's individual access rights and preferences. We can add or remove information from the model by simply adding or removing the corresponding prompts without retraining from scratch. We demonstrate that à-la-carte built models achieve accuracy within 5% of models trained on the union of the respective sources, with comparable cost in terms of training and inference time. For the continual learning benchmarks Split CIFAR- 100 and CORe50, we achieve state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2302.07994
-A Pilot Evaluation of ChatGPT and DALL-E 2 on Decision Making and Spatial Reasoning,Zhi–Bin Tang,"We conduct a pilot study selectively evaluating the cognitive abilities (decision making and spatial reasoning) of two recently released generative transformer models, ChatGPT and DALL-E 2. Input prompts were constructed following neutral a priori guidelines, rather than adversarial intent. Post hoc qualitative analysis of the outputs shows that DALL-E 2 is able to generate at least one correct image for each spatial reasoning prompt, but most images generated are incorrect (even though the model seems to have a clear understanding of the objects mentioned in the prompt). Similarly, in evaluating ChatGPT on the rationality axioms developed under the classical Von Neumann-Morgenstern utility theorem, we find that, although it demonstrates some level of rational decision-making, many of its decisions violate at least one of the axioms even under reasonable constructions of preferences, bets, and decision-making prompts. ChatGPT's outputs on such problems generally tended to be unpredictable: even as it made irrational decisions (or employed an incorrect reasoning process) for some simpler decision-making problems, it was able to draw correct conclusions for more complex bet structures. We briefly comment on the nuances and challenges involved in scaling up such a 'cognitive' evaluation or conducting it with a closed set of answer keys ('ground truth'), given that these models are inherently generative and open-ended in responding to prompts.",,http://arxiv.org/pdf/2302.09068
-Proof of Concept: Using ChatGPT to Teach Emergency Physicians How to Break Bad News,J. Webb,"Background Breaking bad news is an essential skill for practicing physicians, particularly in the field of emergency medicine (EM). Patient-physician communication teaching has previously relied on standardized patient scenarios and objective structured clinical examination formats. The novel use of artificial intelligence (AI) chatbot technology, such as Chat Generative Pre-trained Transformer (ChatGPT), may provide an alternative role in graduate medical education in this area. As a proof of concept, the author demonstrates how providing detailed prompts to the AI chatbot can facilitate the design of a realistic clinical scenario, enable active roleplay, and deliver effective feedback to physician trainees. Methods ChatGPT-3.5 language model was utilized to assist in the roleplay of breaking bad news. A detailed input prompt was designed to outline rules of play and grading assessment via a standardized scale. User inputs (physician role), chatbot outputs (patient role) and ChatGPT-generated feedback were recorded. Results ChatGPT set up a realistic training scenario on breaking bad news based on the initial prompt. Active roleplay as a patient in an emergency department setting was accomplished, and clear feedback was provided to the user through the application of the Setting up, Perception, Invitation, Knowledge, Emotions with Empathy, and Strategy or Summary (SPIKES) framework for breaking bad news. Conclusion The novel use of AI chatbot technology to assist educators is abundant with potential. ChatGPT was able to design an appropriate scenario, provide a means for simulated patient-physician roleplay, and deliver real-time feedback to the physician user. Future studies are required to expand use to a targeted group of EM physician trainees and provide best practice guidelines for AI use in graduate medical education.","{'model': 'tldr@v2.0.0', 'text': 'The author demonstrates how providing detailed prompts to the AI chatbot can facilitate the design of a realistic clinical scenario, enable active roleplay, and deliver effective feedback to physician trainees.'}",https://assets.cureus.com/uploads/original_article/pdf/154391/20230609-458-1qfzq7g.pdf
-I Know What You Do Not Know: Knowledge Graph Embedding via Co-distillation Learning,Yang Liu,"Knowledge graph (KG) embedding seeks to learn vector representations for entities and relations. Conventional models reason over graph structures, but they suffer from the issues of graph incompleteness and long-tail entities. Recent studies have used pre-trained language models to learn embeddings based on the textual information of entities and relations, but they cannot take advantage of graph structures. In the paper, we show empirically that these two kinds of features are complementary for KG embedding. To this end, we propose CoLE, a Co-distillation Learning method for KG Embedding that exploits the complementarity of graph structures and text information. Its graph embedding model employs Transformer to reconstruct the representation of an entity from its neighborhood subgraph. Its text embedding model uses a pre-trained language model to generate entity representations from the soft prompts of their names, descriptions and relational neighbors. To let the two models promote each other, we propose co-distillation learning that allows them to distill selective knowledge from each other's prediction logits. In our co-distillation learning, each model serves as both a teacher and a student. Experiments on benchmark datasets demonstrate that the two models outperform their related baselines, and the ensemble method CoLE with co-distillation learning advances the state-of-the-art of KG embedding.","{'model': 'tldr@v2.0.0', 'text': 'CoLE is proposed, a Co-distillation Learning method for KG Embedding that exploits the complementarity of graph structures and text information and advances the state-of-the-art of KG embedding.'}",
-PromptonomyViT: Multi-Task Prompt Learning Improves Video Transformers using Synthetic Scene Data,Roei Herzig,"Action recognition models have achieved impressive results by incorporating scene-level annotations, such as objects, their relations, 3D structure, and more. However, obtaining annotations of scene structure for videos requires a significant amount of effort to gather and annotate, making these methods expensive to train. In contrast, synthetic datasets generated by graphics engines provide powerful alternatives for generating scene-level annotations across multiple tasks. In this work, we propose an approach to leverage synthetic scene data for improving video understanding. We present a multi-task prompt learning approach for video transformers, where a shared video transformer backbone is enhanced by a small set of specialized parameters for each task. Specifically, we add a set of ``task prompts'', each corresponding to a different task, and let each prompt predict task-related annotations. This design allows the model to capture information shared among synthetic scene tasks as well as information shared between synthetic scene tasks and a real video downstream task throughout the entire network. We refer to this approach as ``Promptonomy'', since the prompts model task-related structure. We propose the PromptonomyViT model (PViT), a video transformer that incorporates various types of scene-level information from synthetic data using the ``Promptonomy'' approach. PViT shows strong performance improvements on multiple video understanding tasks and datasets.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes the PromptonomyViT model (PViT), a video transformer that incorporates various types of scene-level information from synthetic data using the ``Promptonomy'' approach, and shows strong performance improvements on multiple video understanding tasks and datasets.""}",http://arxiv.org/pdf/2212.04821
-A Redhead Walks into a Bar: Experiences of Writing Fiction with Artificial Intelligence,Maliheh Ghajargar,"Human creativity has been often aided and supported by artificial tools, spanning traditional tools such as ideation cards, pens, and paper, to computed and software. Tools for creativity are increasingly using artificial intelligence to not only support the creative process, but also to act upon the creation with a higher level of agency. This paper focuses on writing fiction as a creative activity and explores human-AI co-writing through a research product, which employs a natural language processing model, the Generative Pre-trained Transformer 3 (GPT-3), to assist the co-authoring of narrative fiction. We report on two progressive – not comparative – autoethnographic studies to attain our own creative practices in light of our engagement with the research product: (1) a co-writing activity initiated by basic textual prompts using basic elements of narrative and (2) a co-writing activity initiated by more advanced textual prompts using elements of narrative, including dialects and metaphors undertaken by one of the authors of this paper who has doctoral training in literature. In both studies, we quickly came up against the limitations of the system; then, we repositioned our goals and practices to maximize our chances of success. As a result, we discovered not only limitations but also hidden capabilities, which not only altered our creative practices and outcomes, but which began to change the ways we were relating to the AI as collaborator.",,https://dl.acm.org/doi/pdf/10.1145/3569219.3569418
-Can language models automate data wrangling?,Gonzalo Jaimovitch-López,,"{'model': 'tldr@v2.0.0', 'text': 'A major finding is that language models appear as a powerful tool for a wide range of data wrangling tasks, and some guidelines about how they can be integrated into data processing pipelines are provided, provided the users can take advantage of their flexibility and the diversity of tasks to be addressed.'}",https://link.springer.com/content/pdf/10.1007/s10994-022-06259-9.pdf
-TextDiffuser: Diffusion Models as Text Painters,Jingye Chen,"Diffusion models have gained increasing attention for their impressive generation abilities but currently struggle with rendering accurate and coherent text. To address this issue, we introduce TextDiffuser, focusing on generating images with visually appealing text that is coherent with backgrounds. TextDiffuser consists of two stages: first, a Transformer model generates the layout of keywords extracted from text prompts, and then diffusion models generate images conditioned on the text prompt and the generated layout. Additionally, we contribute the first large-scale text images dataset with OCR annotations, MARIO-10M, containing 10 million image-text pairs with text recognition, detection, and character-level segmentation annotations. We further collect the MARIO-Eval benchmark to serve as a comprehensive tool for evaluating text rendering quality. Through experiments and user studies, we show that TextDiffuser is flexible and controllable to create high-quality text images using text prompts alone or together with text template images, and conduct text inpainting to reconstruct incomplete images with text. The code, model, and dataset will be available at \url{https://aka.ms/textdiffuser}.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces TextDiffuser, focusing on generating images with visually appealing text that is coherent with backgrounds, and contributes the first large-scale text images dataset with OCR annotations, MARIO-10M, containing 10 million image-text pairs with text recognition, detection, and character-level segmentation annotations.'}",http://arxiv.org/pdf/2305.10855
-Language Prompt for Autonomous Driving,Dongming Wu,"A new trend in the computer vision community is to capture objects of interest following flexible human command represented by a natural language prompt. However, the progress of using language prompts in driving scenarios is stuck in a bottleneck due to the scarcity of paired prompt-instance data. To address this challenge, we propose the first object-centric language prompt set for driving scenes within 3D, multi-view, and multi-frame space, named NuPrompt. It expands Nuscenes dataset by constructing a total of 35,367 language descriptions, each referring to an average of 5.3 object tracks. Based on the object-text pairs from the new benchmark, we formulate a new prompt-based driving task, \ie, employing a language prompt to predict the described object trajectory across views and frames. Furthermore, we provide a simple end-to-end baseline model based on Transformer, named PromptTrack. Experiments show that our PromptTrack achieves impressive performance on NuPrompt. We hope this work can provide more new insights for the autonomous driving community. Dataset and Code will be made public at \href{https://github.com/wudongming97/Prompt4Driving}{https://github.com/wudongming97/Prompt4Driving}.","{'model': 'tldr@v2.0.0', 'text': 'The first object-centric language prompt set for driving scenes within 3D, multi-view, and multi-frame space, named NuPrompt is proposed, and a new prompt-based driving task is formulated, employing a language prompt to predict the described object trajectory across views and frames.'}",https://arxiv.org/pdf/2309.04379
-VampNet: Music Generation via Masked Acoustic Token Modeling,Hugo Flores Garcia,"We introduce VampNet, a masked acoustic token modeling approach to music synthesis, compression, inpainting, and variation. We use a variable masking schedule during training which allows us to sample coherent music from the model by applying a variety of masking approaches (called prompts) during inference. VampNet is non-autoregressive, leveraging a bidirectional transformer architecture that attends to all tokens in a forward pass. With just 36 sampling passes, VampNet can generate coherent high-fidelity musical waveforms. We show that by prompting VampNet in various ways, we can apply it to tasks like music compression, inpainting, outpainting, continuation, and looping with variation (vamping). Appropriately prompted, VampNet is capable of maintaining style, genre, instrumentation, and other high-level aspects of the music. This flexible prompting capability makes VampNet a powerful music co-creation tool. Code and audio samples are available online.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that by prompting VampNet in various ways, it can apply it to tasks like music compression, inpainting, out Painting, continuation, and looping with variation (vamping), and is capable of maintaining style, genre, instrumentation, and other high-level aspects of the music.'}",https://arxiv.org/pdf/2307.04686
-Training Transformers Together,Alexander Borzunov,"The infrastructure necessary for training state-of-the-art models is becoming overly expensive, which makes training such models affordable only to large corporations and institutions. Recent work proposes several methods for training such models collaboratively, i.e., by pooling together hardware from many independent parties and training a shared model over the Internet. In this demonstration, we collaboratively trained a text-to-image transformer similar to OpenAI DALL-E. We invited the viewers to join the ongoing training run, showing them instructions on how to contribute using the available hardware. We explained how to address the engineering challenges associated with such a training run (slow communication, limited memory, uneven performance between devices, and security concerns) and discussed how the viewers can set up collaborative training runs themselves. Finally, we show that the resulting model generates images of reasonable quality on a number of prompts.","{'model': 'tldr@v2.0.0', 'text': 'This demonstration collaboratively trained a text-to-image transformer similar to OpenAI DALL-E, and it is shown that the resulting model generates images of reasonable quality on a number of prompts.'}",http://arxiv.org/pdf/2207.03481
-Vector Representations of Idioms in Conversational Systems,Tosin P. Adewumi,"In this study, we demonstrate that an open-domain conversational system trained on idioms or figurative language generates more fitting responses to prompts containing idioms. Idioms are a part of everyday speech in many languages and across many cultures, but they pose a great challenge for many natural language processing (NLP) systems that involve tasks such as information retrieval (IR), machine translation (MT), and conversational artificial intelligence (AI). We utilized the Potential Idiomatic Expression (PIE)-English idiom corpus for the two tasks that we investigated: classification and conversation generation. We achieved a state-of-the-art (SoTA) result of a 98% macro F1 score on the classification task by using the SoTA T5 model. We experimented with three instances of the SoTA dialogue model—the Dialogue Generative Pre-trained Transformer (DialoGPT)—for conversation generation. Their performances were evaluated by using the automatic metric, perplexity, and a human evaluation. The results showed that the model trained on the idiom corpus generated more fitting responses to prompts containing idioms 71.9% of the time in comparison with a similar model that was not trained on the idiom corpus. We have contributed the model checkpoint/demo/code to the HuggingFace hub for public access.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that an open-domain conversational system trained on idioms or figurative language generates more fitting responses to prompts containing idioms than a similar model trained on the idiom corpus.'}",http://arxiv.org/pdf/2205.03666
-Generalization Properties of Retrieval-based Models,S. Basu,"Many modern high-performing machine learning models such as GPT-3 primarily rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Remarkably, retrieval-based methods have enjoyed success on a wide range of problems, ranging from standard natural language processing and vision tasks to protein folding, as demonstrated by many recent efforts, including WebGPT and AlphaFold. Despite growing literature showcasing the promise of these models, the theoretical underpinning for such models remains underexplored. In this paper, we present a formal treatment of retrieval-based models to characterize their generalization ability. In particular, we focus on two classes of retrieval-based classification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task.","{'model': 'tldr@v2.0.0', 'text': 'A formal treatment of retrieval-based models to characterize their generalization ability is presented and it is shown that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy.'}",http://arxiv.org/pdf/2210.02617
-Clinical Prompt Learning With Frozen Language Models.,Niall Taylor,"When the first transformer-based language models were published in the late 2010s, pretraining with general text and then fine-tuning the model on a task-specific dataset often achieved the state-of-the-art performance. However, more recent work suggests that for some tasks, directly prompting the pretrained model matches or surpasses fine-tuning in performance with few or no model parameter updates required. The use of prompts with language models for natural language processing (NLP) tasks is known as prompt learning. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared this with more traditional fine-tuning methods. Results show that prompt learning methods were able to match or surpass the performance of traditional fine-tuning with up to 1000 times fewer trainable parameters, less training time, less training data, and lower computation resource requirements. We argue that these characteristics make prompt learning a very desirable alternative to traditional fine-tuning for clinical tasks, where the computational resources of public health providers are limited, and where data can often not be made available or not be used for fine-tuning due to patient privacy concerns. The complementary code to reproduce the experiments presented in this work can be found at https://github.com/NtaylorOX/Public_Clinical_Prompt.","{'model': 'tldr@v2.0.0', 'text': 'This work investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared this with more traditional fine-tuning methods, and showed that prompt learning methods were able to match or surpass the performance of traditional Finetuning with up to 1000 times fewer trainable parameters, less training time, lessTraining data, and lower computation resource requirements.'}",https://arxiv.org/pdf/2205.05535
-MAQA: A Multimodal QA Benchmark for Negation,Judith Yue Li,"Multimodal learning can benefit from the representation power of pretrained Large Language Models (LLMs). However, state-of-the-art transformer based LLMs often ignore negations in natural language and there is no existing benchmark to quantitatively evaluate whether multimodal transformers inherit this weakness. In this study, we present a new multimodal question answering (QA) benchmark adapted from labeled music videos in AudioSet (Gemmeke et al., 2017) with the goal of systematically evaluating if multimodal transformers can perform complex reasoning to recognize new concepts as negation of previously learned concepts. We show that with standard fine-tuning approach multimodal transformers are still incapable of correctly interpreting negation irrespective of model size. However, our experiments demonstrate that augmenting the original training task distributions with negated QA examples allow the model to reliably reason with negation. To do this, we describe a novel data generation procedure that prompts the 540B-parameter PaLM model to automatically generate negated QA examples as compositions of easily accessible video tags. The generated examples contain more natural linguistic patterns and the gains compared to template-based task augmentation approach are significant.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with standard fine-tuning approach multimodal transformers are still incapable of correctly interpreting negation irrespective of model size, but it is demonstrated that augmenting the original training task distributions with negated QA examples allow the model to reliably reason with negation.'}",http://arxiv.org/pdf/2301.03238
-FedYolo: Augmenting Federated Learning with Pretrained Transformers,Xuechen Zhang,"The growth and diversity of machine learning applications motivate a rethinking of learning with mobile and edge devices. How can we address diverse client goals and learn with scarce heterogeneous data? While federated learning aims to address these issues, it has challenges hindering a unified solution. Large transformer models have been shown to work across a variety of tasks achieving remarkable few-shot adaptation. This raises the question: Can clients use a single general-purpose model, rather than custom models for each task, while obeying device and network constraints? In this work, we investigate pretrained transformers (PTF) to achieve these on-device learning goals and thoroughly explore the roles of model size and modularity, where the latter refers to adaptation through modules such as prompts or adapters. Focusing on federated learning, we demonstrate that: (1) Larger scale shrinks the accuracy gaps between alternative approaches and improves heterogeneity robustness. Scale allows clients to run more local SGD epochs which can significantly reduce the number of communication rounds. At the extreme, clients can achieve respectable accuracy locally highlighting the potential of fully-local learning. (2) Modularity, by design, enables $>$100$\times$ less communication in bits. Surprisingly, it also boosts the generalization capability of local adaptation methods and the robustness of smaller PTFs. Finally, it enables clients to solve multiple unrelated tasks simultaneously using a single PTF, whereas full updates are prone to catastrophic forgetting. These insights on scale and modularity motivate a new federated learning approach we call""You Only Load Once""(FedYolo): The clients load a full PTF model once and all future updates are accomplished through communication-efficient modules with limited catastrophic-forgetting, where each task is assigned to its own module.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates pretrained transformers (PTF) to achieve on-device learning goals and thoroughly explore the roles of model size and modularity, where the latter refers to adaptation through modules such as prompts or adapters.'}",https://arxiv.org/pdf/2307.04905
-Explaining Emergent In-Context Learning as Kernel Regression,Chi Han,"Large language models (LLMs) have initiated a paradigm shift in transfer learning. In contrast to the classic pretraining-then-finetuning procedure, in order to use LLMs for downstream prediction tasks, one only needs to provide a few demonstrations, known as in-context examples, without adding more or updating existing model parameters. This in-context learning (ICL) capability of LLMs is intriguing, and it is not yet fully understood how pretrained LLMs acquire such capabilities. In this paper, we investigate the reason why a transformer-based language model can accomplish in-context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression with internal representations when faced with in-context examples. More concretely, we first prove that Bayesian inference on in-context prompts can be asymptotically understood as kernel regression $\hat y = \sum_i y_i K(x, x_i)/\sum_i K(x, x_i)$ as the number of in-context demonstrations grows. Then, we empirically investigate the in-context behaviors of language models. We find that during ICL, the attention and hidden features in LLMs match the behaviors of a kernel regression. Finally, our theory provides insights into multiple phenomena observed in the ICL field: why retrieving demonstrative samples similar to test samples can help, why ICL performance is sensitive to the output formats, and why ICL accuracy benefits from selecting in-distribution and representative samples.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the reason why a transformer-based language model can accomplish in- context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression with internal representations when faced with in-context examples and empirically investigates the in- Context behaviors of language models.'}",
-Membership Inference Attacks With Token-Level Deduplication on Korean Language Models,Myung Gyo Oh,"The confidentiality threat against training data has become a significant security problem in neural language models. Recent studies have shown that memorized training data can be extracted by injecting well-chosen prompts into generative language models. While these attacks have achieved remarkable success in the English-based Transformer architecture, it is unclear whether they are still effective in other language domains. This paper studies the effectiveness of attacks against Korean models and the potential for attack improvements that might be beneficial for future defense studies. The contribution of this study is two-fold. First, we perform a membership inference attack against the state-of-the-art Korean GPT model. We found approximate training data with 20% to 90% precision in the top-100 samples and confirmed that the proposed attack technique for naive GPT is valid across the language domains. Second, in this process, we observed that the redundancy of the selected sentences could hardly be detected with the existing attack method. Since the information appearing in a few documents is more likely to be meaningful, it is desirable to increase the uniqueness of the sentences to improve the effectiveness of the attack. Thus, we propose a deduplication strategy to replace the traditional word-level similarity metric with the BPE token level. Our proposed strategy reduces 6% to 22% of the underestimated samples among selected ones, thereby improving precision by up to 7%p. As a result, we show that considering both language- and model-specific characteristics is essential to improve the effectiveness of attack strategies. We also discuss possible mitigations against the MI attacks on the general language models.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that considering both language- and model-specific characteristics is essential to improve the effectiveness of attack strategies, and a deduplication strategy to replace the traditional word-level similarity metric with the BPE token level is proposed.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10025743.pdf
-GenerateCT: Text-Guided 3D Chest CT Generation,Ibrahim Ethem Hamamci,"Generative modeling has experienced substantial progress in recent years, particularly in text-to-image and text-to-video synthesis. However, the medical field has not yet fully exploited the potential of large-scale foundational models for synthetic data generation. In this paper, we introduce GenerateCT, the first method for text-conditional computed tomography (CT) generation, addressing the limitations in 3D medical imaging research and making our entire framework open-source. GenerateCT consists of a pre-trained large language model, a transformer-based text-conditional 3D chest CT generation architecture, and a text-conditional spatial super-resolution diffusion model. We also propose CT-ViT, which efficiently compresses CT volumes while preserving auto-regressiveness in-depth, enabling the generation of 3D CT volumes with variable numbers of axial slices. Our experiments demonstrate that GenerateCT can produce realistic, high-resolution, and high-fidelity 3D chest CT volumes consistent with medical language text prompts. We further investigate the potential of GenerateCT by training a model using generated CT volumes for multi-abnormality classification of chest CT volumes. Our contributions provide a valuable foundation for future research in text-conditional 3D medical image generation and have the potential to accelerate advancements in medical imaging research. Our code, pre-trained models, and generated data are available at https://github.com/ibrahimethemhamamci/GenerateCT.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces GenerateCT, the first method for text-conditional computed tomography (CT) generation, addressing the limitations in 3D medical imaging research and making the entire framework open-source.'}",http://arxiv.org/pdf/2305.16037
-Chinese ASR and NER Improvement Based on Whisper Fine-Tuning,Hao Yang,"Based on 680k hours of weakly supervised multilingual and multi-task speech transcription/translation data, Whisper [1] has developed a robust system for both Automated Speech Recognition (ASR) and Speech Translation (ST). Whisper provides a simple model architecture based on Mel spectrum + two-layer convolution + Seq2seq Transformer architecture, which is easy to fine-tune on conditional generation tasks. This paper analyzes how to fine-tune Chinese ASR [2] and NER tasks based on Whisper, including (1) how to design different prompts for different generative tasks; (2) how to train ASR and NER tasks at the same time; (3) whether the performance can be further improved by using weak supervision for data enhancement. Experiments based on AISHELL [3] and AISHELL-NER [4] data, and multi-task fine-tuning based on Whisper can effectively improve the performance of ASR and NER.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes how to fine-tune Chinese ASR and NER tasks based on Whisper, including how to design different prompts for different generative tasks and whether the performance can be further improved by using weak supervision for data enhancement.'}",
-Learning Expressive Prompting With Residuals for Vision Transformers,Rajshekhar Das,"Prompt learning is an efficient approach to adapt transformers by inserting learnable set of parameters into the input and intermediate representations of a pre-trained model. In this work, we present Expressive Prompts with Residuals (EXPRES) which modifies the prompt learning paradigm specifically for effective adaptation of vision transformers (ViT). Our method constructs downstream representations via learnable “output” tokens (shal-low prompts), that are akin to the learned class tokens of the ViT. Further for better steering of the downstream representation processed by the frozen transformer, we introduce residual learnable tokens that are added to the output of various computations. We apply EXPRES for image classification and few-shot semantic segmentation, and show our method is capable of achieving state of the art prompt tuning on 3/3 categories of the VTAB benchmark. In addition to strong performance, we observe that our approach is an order of magnitude more prompt efficient than existing visual prompting baselines. We analytically show the computational benefits of our approach over weight space adaptation techniques like finetuning. Lastly we systematically corroborate the architectural design of our method via a series of ablation experiments.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Expressive Prompts with Residuals (EXPRES) which modifies the prompt learning paradigm specifically for effective adaptation of vision transformers (ViT) and observes that this approach is an order of magnitude more prompt efficient than existing visual prompting baselines.'}",https://arxiv.org/pdf/2303.15591
-Evaluating deep generative models on cognitive tasks: a case study,Zhi–Bin Tang,,,https://link.springer.com/content/pdf/10.1007/s44163-023-00067-3.pdf
-ProRes: Exploring Degradation-aware Visual Prompt for Universal Image Restoration,Jiaqi Ma,"Image restoration aims to reconstruct degraded images, e.g., denoising or deblurring. Existing works focus on designing task-specific methods and there are inadequate attempts at universal methods. However, simply unifying multiple tasks into one universal architecture suffers from uncontrollable and undesired predictions. To address those issues, we explore prompt learning in universal architectures for image restoration tasks. In this paper, we present Degradation-aware Visual Prompts, which encode various types of image degradation, e.g., noise and blur, into unified visual prompts. These degradation-aware prompts provide control over image processing and allow weighted combinations for customized image restoration. We then leverage degradation-aware visual prompts to establish a controllable and universal model for image restoration, called ProRes, which is applicable to an extensive range of image restoration tasks. ProRes leverages the vanilla Vision Transformer (ViT) without any task-specific designs. Furthermore, the pre-trained ProRes can easily adapt to new tasks through efficient prompt tuning with only a few images. Without bells and whistles, ProRes achieves competitive performance compared to task-specific methods and experiments can demonstrate its ability for controllable restoration and adaptation for new tasks. The code and models will be released in \url{https://github.com/leonmakise/ProRes}.","{'model': 'tldr@v2.0.0', 'text': 'Dgradation-aware Visual Prompts are presented, which encode various types of image degradation, e.g., noise and blur, into unified visual prompts, which provide control over image processing and allow weighted combinations for customized image restoration.'}",http://arxiv.org/pdf/2306.13653
-Deep Learning in ChatGPT - A Survey,A.M. Jasmine Hashana,"Abstract-As a subset of machine learning, deep learning makes use of multiple-layer neural networks to learn with available data and make decisions or predictions. A large language model called ChatGPT is based on deep learning, specifically a type of neural network called a transformer. ChatGPT's transformer architecture uses attention mechanisms to focus on the most important parts of the input, allowing it to process and comprehend a large amount of text data. In order for the model to comprehend the context and meaning of natural language text, it is trained on a huge database of text, including articles and books. One of the main importance of using deep learning in ChatGPT is its intelligence to understand relationships and patterns from the input text and generate or predict new text that is homogeneous to the input/training data. Because of this, ChatGPT is able to respond to questions and prompts in a manner that is comparable to that of a human, making it useful for a wide scope of natural language processing missions like translating languages, summarizing texts, and responding to questions. It's worth noting that, while deep learning has been highly effective in ChatGPT, it is not without its limitations. To train, deep learning models can be very complex and require a lot of data and computing power.","{'model': 'tldr@v2.0.0', 'text': 'The transformer architecture in ChatGPT is able to respond to questions and prompts in a manner that is comparable to that of a human, making it useful for a wide scope of natural language processing missions like translating languages, summarizing texts, and responding to questions.'}",
-Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Models,Mert Yuksekgonul,"We investigate the internal behavior of Transformer-based Large Language Models (LLMs) when they generate factually incorrect text. We propose modeling factual queries as Constraint Satisfaction Problems and use this framework to investigate how the model interacts internally with factual constraints. Specifically, we discover a strong positive relation between the model's attention to constraint tokens and the factual accuracy of its responses. In our curated suite of 11 datasets with over 40,000 prompts, we study the task of predicting factual errors with the Llama-2 family across all scales (7B, 13B, 70B). We propose SAT Probe, a method probing self-attention patterns, that can predict constraint satisfaction and factual errors, and allows early error identification. The approach and findings demonstrate how using the mechanistic understanding of factuality in LLMs can enhance reliability.","{'model': 'tldr@v2.0.0', 'text': ""A strong positive relation is found between the model's attention to constraint tokens and the factual accuracy of its responses and the approach and findings demonstrate how using the mechanistic understanding of factuality in LLMs can enhance reliability.""}",https://arxiv.org/pdf/2309.15098
-Can AI Put Gamma-Ray Astrophysicists Out of a Job?,S. Spencer,"In what will likely be a litany of generative-model-themed arXiv submissions celebrating April the 1st, we evaluate the capacity of state-of-the-art transformer models to create a paper detailing the detection of a Pulsar Wind Nebula with a non-existent Imaging Atmospheric Cherenkov Telescope (IACT) Array. We do this to evaluate the ability of such models to interpret astronomical observations and sources based on language information alone, and to assess potential means by which fraudulently generated scientific papers could be identified during peer review (given that reliable generative model watermarking has yet to be deployed for these tools). We conclude that our jobs as astronomers are safe for the time being. From this point on, prompts given to ChatGPT and Stable Diffusion are shown in orange, text generated by ChatGPT is shown in black, whereas analysis by the (human) authors is in blue.",,http://arxiv.org/pdf/2303.17853
-A Statistical Perspective on Retrieval-Based Models,Soumya Basu,"Many modern high-performing machine learning models increasingly rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Despite a growing literature showcasing the promise of these retrieval-based models , their theoretical underpinnings remain under-explored. In this paper, we present a formal treatment of retrieval-based models to characterize their performance via a novel statistical perspective. In particular, we study two broad classes of retrieval-based clas-sification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall performance. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a formal treatment of retrieval-based models to characterize their performance via a novel statistical perspective and shows that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall performance.'}",
-Content Moderation for Evolving Policies using Binary Question Answering,S. S. Mullick,"Content moderation on social media is governed by policies that are intricate and frequently updated with evolving world events. However, automated content moderation systems often restrict easy adaptation to policy changes and are expected to learn policy intricacies from limited amounts of labeled data, which make effective policy compliance challenging. We propose to model content moderation as a binary question answering problem where the questions validate the loosely coupled themes constituting a policy. A decision logic is applied on top to aggregate the theme-specific validations. This way the questions pass theme information to a transformer network as explicit policy prompts, that in turn enables explainability. This setting further allows for faster adaptation to policy updates by leveraging zero-shot capabilities of pre-trained transformers. We showcase improved recall for our proposed method at 95\% precision on two proprietary datasets of social media posts and comments respectively annotated under curated Hate Speech and Commercial Spam policies.","{'model': 'tldr@v2.0.0', 'text': 'Improved recall is showcased for the proposed binary question answering problem where the questions validate the loosely coupled themes constituting a policy, which allows for faster adaptation to policy updates by leveraging zero-shot capabilities of pre-trained transformers.'}",https://aclanthology.org/2023.acl-industry.54.pdf
-Transformer assembling technology interactive simulation system and method,赵坚,"The invention relates to an assembling technology interactive simulation system and method based on virtual reality, a virtual assembling technology and cross platforms, in particular to a transformer assembling technology interactive simulation system and method. According to the transformer assembling technology interactive simulation system and method, a three-dimensional modeling tool, a three-dimensional model integrating tool and a somatosensory camera are combined through a virtual reality technology and by combining an image processing technology, a three-dimensional simulation technology and other related technologies, the set of assembling technology interactive simulation system based on a virtual prototype is developed, the whole process of production and assembling operation of the transformer assembling technology is simulated on a computer, and meanwhile real-time supervision main point prompts are combined to carry out training and guidance on onsite assembling workers. The transformer assembling technology interactive simulation system and method can simulate and present the transformer assembling process, and more importantly, training cost is reduced when training of an electric power system is carried out on assembling and construction personnel and when assembling experiments are carried out; training time is shortened, and training efficiency is improved; requirements for training sites are lowered.","{'model': 'tldr@v2.0.0', 'text': 'The transformer assembling technology interactive simulation system and method can simulate and present the transformer assembling process, and more importantly, training cost is reduced when training of an electric power system is carried out on assembling and construction personnel and when assembling experiments are carried out.'}",
-Making humanoid robots teaching assistants by using natural language processing (NLP) cloud-based services,A. Lekova,"Humanoid robots have a substantial potential to serve as teaching and social assistants. However, the expectations of the children from robots to interact like humans are huge. This study presents a general model for understanding the natural language in human-robot interaction by applying Generative Pre-trained Transformer (GPT) language models as a service in the Internet of Things. Thus, the physical presence of the robot can help in fine-tuning the GPT model by prompts derived from the environmental context and subsequent robot actions for embodiment understanding of the GPT outputs. The model uses web or cloud services for Natural Language Processing (NLP) to produce and play human-like text, question answering or text generation. Verbal questions are processed either via a local speech recognition software or via a Speech-to-Text (STT) cloud service. The converted question into machine-readable code is sent to one of the GPT language models with zero- or few-shot learning prompts. GPT-J model has been tested and deployed either in the web or cloud with options for varying the parameters for controlling the haphazardness of the generated text. The robot produces human-like text by using Text-to-Speech (TTS) cloud services that convert the response into audio format rendered on the robot to be played. Useful requirements how the model to be used in order to be feasible were determined based on the evaluation of the outputs given from the different NLP and GPT-J web or cloud-services. We designed and implemented the model in order to be used by a humanoid NAO-type robot in the speech language therapy practice, however it can be used for other open-source and programmable robots and in different contexts.","{'model': 'tldr@v2.0.0', 'text': 'A general model for understanding the natural language in human-robot interaction by applying Generative Pre-trained Transformer (GPT) language models as a service in the Internet of Things to help in fine-tuning the GPT model.'}",https://www.extrica.com/article/22720/pdf
-Prompt Generation Networks for Input-based Adaptation of Frozen Vision Transformers,J. Loedeman,"With the introduction of the transformer architecture in computer vision, increasing model scale has been demonstrated as a clear path to achieving performance and robustness gains. However, with model parameter counts reaching the billions, classical finetuning approaches are becoming increasingly limiting and even unfeasible when models become hosted as inference APIs, as in NLP. To this end, visual prompt learning, whereby a model is adapted by learning additional inputs, has emerged as a potential solution for adapting frozen and cloud-hosted models: During inference, this neither requires access to the internals of models' forward pass function, nor requires any post-processing. In this work, we propose the Prompt Generation Network (PGN) that generates high performing, input-dependent prompts by sampling from an end-to-end learned library of tokens. We further introduce the""prompt inversion""trick, with which PGNs can be efficiently trained in a latent space but deployed as strictly input-only prompts for inference. We show the PGN is effective in adapting pre-trained models to various new datasets: It surpasses previous methods by a large margin on 12/12 datasets and even outperforms full-finetuning on 5/12, while requiring 100x less parameters.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt Generation Network (PGN) is proposed that generates high performing, input-dependent prompts by sampling from an end-to-end learned library of tokens, and the ""prompt inversion"" trick is introduced, with which PGNs can be efficiently trained in a latent space but deployed as strictly input-only prompts for inference.'}",
-Optimizing Language Models for Argumentative Reasoning,Luke Thorburn,"Large transformer-based causal language models are capable of strong performance on many natural language processing tasks. Here, we systematically evaluate the performance of the 2.7 billion parameter GPT Neo pre-trained language model on 6 argumentative reasoning tasks under 5 different optimization strategies, including prompt programming, soft prompts, and parameter tuning. We report both intrinsic evaluation metrics (perplexity), and extrinsic measures of the coherence of model outputs, as judged by an expert human rater. With a few exceptions, the rate at which models produced coherent responses ranged from 15-50%. In contrast, human performance (users of the Kialo argument mapping platform) ranged from 65-82% coherent, depending on the task. These results suggest that larger, suitably optimized language models may be capable of supporting authors and auditors of natural language argument maps in human-in-the-loop settings. We share our finetuned models and code.","{'model': 'tldr@v2.0.0', 'text': 'This work systematically evaluates the performance of the 2.7 billion parameter GPT Neo pre-trained language model on 6 argumentative reasoning tasks under 5 different optimization strategies, including prompt programming, soft prompts, and parameter tuning.'}",
-Finding patterns in Knowledge Attribution for Transformers,Jeevesh Juneja,"We analyze the Knowledge Neurons framework for the attribution of factual and relational knowledge to particular neurons in the transformer network. We use a 12-layer multi-lingual BERT model for our experiments. Our study reveals various interesting phenomena. We observe that mostly factual knowledge can be attributed to middle and higher layers of the network($\ge 6$). Further analysis reveals that the middle layers($6-9$) are mostly responsible for relational information, which is further refined into actual factual knowledge or the""correct answer""in the last few layers($10-12$). Our experiments also show that the model handles prompts in different languages, but representing the same fact, similarly, providing further evidence for effectiveness of multi-lingual pre-training. Applying the attribution scheme for grammatical knowledge, we find that grammatical knowledge is far more dispersed among the neurons than factual knowledge.","{'model': 'tldr@v2.0.0', 'text': 'This study analyzes the Knowledge Neurons framework for the attribution of factual and relational knowledge to particular neurons in the transformer network and finds that grammatical knowledge is far more dispersed among the neurons than factual knowledge.'}",http://arxiv.org/pdf/2205.01366
-Prompt Enhanced Generative MRC Framework for Pancreatic Cancer NER,Zhendong Tan,"Medical Named Entity Recognition (NER) is a fundamental but challenging task due to the lack of specialized entity datasets like tumor entities, which are often overlapped and discontinuous. In this paper, we propose a novel Prompt Enhanced Generative Machine Reading Comprehension Framework (PGMRC) to improve the overlapped and discontinuous NER performance. Specifically, we formulate NER as a Machine Reading Comprehension (MRC) task and employ a pre-trained encoder-decoder module to generate entity span sequences according to their entity query. In this way, we adopt query to guide the model to focus on answer entities in context, which can naturally solve entity overlap and alleviate the exposure bias of the generative model. Then, we introduce continuous prompts to the self-attention mechanism in Transformer to reduce the dependence on manually constructed queries. In addition, we annotate 875 pathological documents of pancreatic cancer and construct a Chinese pathological NER dataset (PAN) containing overlapped and discontinuous entities. Finally, we conduct our experiments on three widely used benchmarks (GENIA, ACE04, ACE05) and our dataset PAN. Experiments have demonstrated its effectiveness and better performance than state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper forms NER as a Machine Reading Comprehension task and employs a pre-trained encoder-decoder module to generate entity span sequences according to their entity query and introduces continuous prompts to the self-attention mechanism in Transformer to reduce the dependence on manually constructed queries.'}",
-In-Context Convergence of Transformers,Yu Huang,"Transformers have recently revolutionized many domains in modern machine learning and one salient discovery is their remarkable in-context learning capability, where models can solve an unseen task by utilizing task-specific prompts without further parameters fine-tuning. This also inspired recent theoretical studies aiming to understand the in-context learning mechanism of transformers, which however focused only on linear transformers. In this work, we take the first step toward studying the learning dynamics of a one-layer transformer with softmax attention trained via gradient descent in order to in-context learn linear function classes. We consider a structured data model, where each token is randomly sampled from a set of feature vectors in either balanced or imbalanced fashion. For data with balanced features, we establish the finite-time convergence guarantee with near-zero prediction error by navigating our analysis over two phases of the training dynamics of the attention map. More notably, for data with imbalanced features, we show that the learning dynamics take a stage-wise convergence process, where the transformer first converges to a near-zero prediction error for the query tokens of dominant features, and then converges later to a near-zero prediction error for the query tokens of under-represented features, respectively via one and four training phases. Our proof features new techniques for analyzing the competing strengths of two types of attention weights, the change of which determines different training phases.","{'model': 'tldr@v2.0.0', 'text': 'This work takes the first step toward studying the learning dynamics of a one-layer transformer with softmax attention trained via gradient descent in order to in-context learn linear function classes.'}",https://arxiv.org/pdf/2310.05249
-Bootstrapping Vision-Language Learning with Decoupled Language Pre-training,Yiren Jian,"We present a novel methodology aimed at optimizing the application of frozen large language models (LLMs) for resource-intensive vision-language (VL) pre-training. The current paradigm uses visual features as prompts to guide language models, with a focus on determining the most relevant visual features for corresponding text. Our approach diverges by concentrating on the language component, specifically identifying the optimal prompts to align with visual features. We introduce the Prompt-Transformer (P-Former), a model that predicts these ideal prompts, which is trained exclusively on linguistic data, bypassing the need for image-text pairings. This strategy subtly bifurcates the end-to-end VL training process into an additional, separate stage. Our experiments reveal that our framework significantly enhances the performance of a robust image-to-text baseline (BLIP-2), and effectively narrows the performance gap between models trained with either 4M or 129M image-text pairs. Importantly, our framework is modality-agnostic and flexible in terms of architectural design, as validated by its successful application in a video learning task using varied base modules. The code will be made available at https://github.com/yiren-jian/BLIText.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt-Transformer (P-Former) is introduced, a model that predicts these ideal prompts to align with visual features, which is trained exclusively on linguistic data, bypassing the need for image-text pairings.'}",https://arxiv.org/pdf/2307.07063
-Exploring the Hurdles in Integrating Artificial Intelligence in Medical Education: A Talk with ChatGPT,R. Vignesh,"The integration of artificial intelligence (AI) in the medical field is rapidly increasing, and its importance is recognised as it can analyse medical information and improve patient outcomes. The Chat Generative Pre-trained Transformer (ChatGPT) by OpenAI is a language model that can respond to prompts in a human-like manner and has the potential to transform medical education. In this conversation, the authors engage with ChatGPT to discuss the challenges associated with the application of AI in medical education. ChatGPT acknowledges the limitations of AI in medical education and the importance of students demonstrating their understanding during assessments. It also discusses the need for academic policies to ensure the responsible and ethical use of AI in medical education.","{'model': 'tldr@v2.0.0', 'text': 'In this conversation, the authors engage with ChatGPT to discuss the challenges associated with the application of AI in medical education and the need for academic policies to ensure the responsible and ethical use of AIIn medical education.'}",
-Prompt-based Ingredient-Oriented All-in-One Image Restoration,Huiyu Gao,"Image restoration aims to recover the high-quality images from their degraded observations. Since most existing methods have been dedicated into single degradation removal, they may not yield optimal results on other types of degradations, which do not satisfy the applications in real world scenarios. In this paper, we propose a novel data ingredient-oriented approach that leverages prompt-based learning to enable a single model to efficiently tackle multiple image degradation tasks. Specifically, we utilize a encoder to capture features and introduce prompts with degradation-specific information to guide the decoder in adaptively recovering images affected by various degradations. In order to model the local invariant properties and non-local information for high-quality image restoration, we combined CNNs operations and Transformers. Simultaneously, we made several key designs in the Transformer blocks (multi-head rearranged attention with prompts and simple-gate feed-forward network) to reduce computational requirements and selectively determines what information should be persevered to facilitate efficient recovery of potentially sharp images. Furthermore, we incorporate a feature fusion mechanism further explores the multi-scale information to improve the aggregated features. The resulting tightly interlinked hierarchy architecture, named as CAPTNet, extensive experiments demonstrate that our method performs competitively to the state-of-the-art.","{'model': 'tldr@v2.0.0', 'text': 'A novel data ingredient-oriented approach that leverages prompt-based learning to enable a single model to efficiently tackle multiple image degradation tasks, and incorporates a feature fusion mechanism that further explores the multi-scale information to improve the aggregated features.'}",https://arxiv.org/pdf/2309.03063
-Hierarchical Prompt Tuning for Few-Shot Multi-Task Learning,Jingping Liu,"Prompt tuning has enhanced the performance of Pre-trained Language Models for multi-task learning in few-shot scenarios. However, existing studies fail to consider that the prompts among different layers in Transformer are different due to the diverse information learned at each layer. In general, the bottom layers in the model tend to capture low-level semantic or structural information, while the upper layers primarily acquire task-specific knowledge. Hence, we propose a novel hierarchical prompt tuning model for few-shot multi-task learning to capture this regularity. The designed model mainly consists of three types of prompts: shared prompts, auto-adaptive prompts, and task-specific prompts. Shared prompts facilitate the sharing of general information across all tasks. Auto-adaptive prompts dynamically select and integrate relevant prompt information from all tasks into the current task. Task-specific prompts concentrate on learning task-specific knowledge. To enhance the model's adaptability to diverse inputs, we introduce deep instance-aware language prompts as the foundation for constructing the above prompts. To evaluate the effectiveness of our proposed method, we conduct extensive experiments on multiple widely-used datasets. The experimental results demonstrate that the proposed method achieves state-of-the-art performance for multi-task learning in few-shot settings and outperforms ChatGPT in the full-data setting.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel hierarchical prompt tuning model that achieves state-of-the-art performance for multi-task learning in few-shot settings and outperforms ChatGPT in the full-data setting.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3614913
-PM-DETR: Domain Adaptive Prompt Memory for Object Detection with Transformers,Peidong Jia,"The Transformer-based detectors (i.e., DETR) have demonstrated impressive performance on end-to-end object detection. However, transferring DETR to different data distributions may lead to a significant performance degradation. Existing adaptation techniques focus on model-based approaches, which aim to leverage feature alignment to narrow the distribution shift between different domains. In this study, we propose a hierarchical Prompt Domain Memory (PDM) for adapting detection transformers to different distributions. PDM comprehensively leverages the prompt memory to extract domain-specific knowledge and explicitly constructs a long-term memory space for the data distribution, which represents better domain diversity compared to existing methods. Specifically, each prompt and its corresponding distribution value are paired in the memory space, and we inject top M distribution-similar prompts into the input and multi-level embeddings of DETR. Additionally, we introduce the Prompt Memory Alignment (PMA) to reduce the discrepancy between the source and target domains by fully leveraging the domain-specific knowledge extracted from the prompt domain memory. Extensive experiments demonstrate that our method outperforms state-of-the-art domain adaptive object detection methods on three benchmarks, including scene, synthetic to real, and weather adaptation. Codes will be released.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a hierarchical Prompt Domain Memory (PDM), which comprehensively leverages the prompt memory to extract domain-specific knowledge and explicitly constructs a long-term memory space for the data distribution, which represents better domain diversity compared to existing methods.'}",http://arxiv.org/pdf/2307.00313
-Visual Prompt Flexible-Modal Face Anti-Spoofing,Zitong Yu,"Recently, vision transformer based multimodal learning methods have been proposed to improve the robustness of face anti-spoofing (FAS) systems. However, multimodal face data collected from the real world is often imperfect due to missing modalities from various imaging sensors. Recently, flexible-modal FAS~\cite{yu2023flexible} has attracted more attention, which aims to develop a unified multimodal FAS model using complete multimodal face data but is insensitive to test-time missing modalities. In this paper, we tackle one main challenge in flexible-modal FAS, i.e., when missing modality occurs either during training or testing in real-world situations. Inspired by the recent success of the prompt learning in language models, we propose \textbf{V}isual \textbf{P}rompt flexible-modal \textbf{FAS} (VP-FAS), which learns the modal-relevant prompts to adapt the frozen pre-trained foundation model to downstream flexible-modal FAS task. Specifically, both vanilla visual prompts and residual contextual prompts are plugged into multimodal transformers to handle general missing-modality cases, while only requiring less than 4\% learnable parameters compared to training the entire model. Furthermore, missing-modality regularization is proposed to force models to learn consistent multimodal feature embeddings when missing partial modalities. Extensive experiments conducted on two multimodal FAS benchmark datasets demonstrate the effectiveness of our VP-FAS framework that improves the performance under various missing-modality cases while alleviating the requirement of heavy model re-training.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the recent success of the prompt learning in language models, a VP-FAS framework is proposed, which learns the modal-relevant prompts to adapt the frozen pre-trained foundation model to downstream flexible-modal FAS task.'}",https://arxiv.org/pdf/2307.13958
-Harnessing GPT-3.5-turbo for Rhetorical Role Prediction in Legal Cases,Anas Belfathi,"We propose a comprehensive study of one-stage elicitation techniques for querying a large pre-trained generative transformer (GPT-3.5-turbo) in the rhetorical role prediction task of legal cases. This task is known as requiring textual context to be addressed. Our study explores strategies such as zero-few shots, task specification with definitions and clarification of annotation ambiguities, textual context and reasoning with general prompts and specific questions. We show that the number of examples, the definition of labels, the presentation of the (labelled) textual context and specific questions about this context have a positive influence on the performance of the model. Given non-equivalent test set configurations, we observed that prompting with a few labelled examples from direct context can lead the model to a better performance than a supervised fined-tuned multi-class classifier based on the BERT encoder (weighted F1 score of = 72%). But there is still a gap to reach the performance of the best systems = 86%) in the LegalEval 2023 task which, on the other hand, require dedicated resources, architectures and training.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that prompting with a few labelled examples from direct context can lead the model to a better performance than a supervised fined-tuned multi-class classifier based on the BERT encoder, but there is still a gap to reach the performance of the best systems in the LegalEval 2023 task.'}",
-ChatGPT in forensic sciences: a new Pandora’s box with advantages and challenges to pay attention,R. Dinis-Oliveira," ChatGPT is a variant of the GPT (Generative Pre-trained Transformer) language model that uses large amounts of text-based training data and a transformer architecture to generate human-like text adjusted to the received prompts. ChatGPT presents several advantages in forensic sciences, namely constituting a virtual assistant to aid lawyers, judges, and victims in managing and interpreting forensic expert data. But what would happen if ChatGPT began to be used to produce forensic expertise reports? Despite its potential applications, the use of ChatGPT and other Large Language Models (LLM) and artificial intelligence (AI) tools in forensic writing also poses ethical and legal concerns, which are discussed in this perspective together with some expected future perspectives.","{'model': 'tldr@v2.0.0', 'text': 'Despite its potential applications, the use of ChatGPT and other Large Language Models and artificial intelligence tools in forensic writing also poses ethical and legal concerns, which are discussed in this perspective together with some expected future perspectives.'}",https://academic.oup.com/fsr/advance-article-pdf/doi/10.1093/fsr/owad039/52412442/owad039.pdf
-Fake News Detection in Low Resource Languages using SetFit Framework,Amin Abdedaiem,"Social media has become an integral part of people’s lives, resulting in a constant flow of information. However, a concerning trend has emerged with the rapid spread of fake news, attributed to the lack of verification mechanisms. Fake news has far-reaching consequences, influencing public opinion, disrupting democracy, fuelingsocial tensions, and impacting various domains such as health, environment, and the economy. In order to identify fake news with data sparsity, especially with low resources languages such as Arabic and its dialects, we propose a few-shot learning fake news detection model based on sentence transformer fine-tuning, utilizing no crafted prompts and language model with few parameters. The experimental results prove that the proposed method can achieve higher performances with fewer news samples. This approach provided 71% F1 score on the Algerian dialect fake news dataset and 70% F1 score on the Modern Standard Arabic (MSA) version of the same dataset, which proves that the approach can work on the standard Arabic and its dialects. Therefore, the proposed model can identify fake news in several domains concerning the Algerian community such as politics, COVID-19, tourism, e-commerce, sport, accidents, and car prices.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning fake news detection model based on sentence transformer fine-tuning, utilizing no crafted prompts and language model with few parameters is proposed, which proves that the approach can work on the standard Arabic and its dialects.'}",https://journal.iberamia.org/index.php/intartif/article/download/1151/198
-On the Relationship between Skill Neurons and Robustness in Prompt Tuning,Leon Ackermann,"Prompt Tuning is a popular parameter-efficient finetuning method for pre-trained large language models (PLMs). Recently, based on experiments with RoBERTa, it has been suggested that Prompt Tuning activates specific neurons in the transformer's feed-forward networks, that are highly predictive and selective for the given task. In this paper, we study the robustness of Prompt Tuning in relation to these""skill neurons"", using RoBERTa and T5. We show that prompts tuned for a specific task are transferable to tasks of the same type but are not very robust to adversarial data, with higher robustness for T5 than RoBERTa. At the same time, we replicate the existence of skill neurons in RoBERTa and further show that skill neurons also seem to exist in T5. Interestingly, the skill neurons of T5 determined on non-adversarial data are also among the most predictive neurons on the adversarial data, which is not the case for RoBERTa. We conclude that higher adversarial robustness may be related to a model's ability to activate the relevant skill neurons on adversarial data.","{'model': 'tldr@v2.0.0', 'text': ""Higher adversarial robustness may be related to a model's ability to activate the relevant skill neurons on adversarial data, with higher robustness for T5 than RoBERTa.""}",https://arxiv.org/pdf/2309.12263
-DCPT: Darkness Clue-Prompted Tracking in Nighttime UAVs,Jiawen Zhu,"Existing nighttime unmanned aerial vehicle (UAV) trackers follow an""Enhance-then-Track""architecture - first using a light enhancer to brighten the nighttime video, then employing a daytime tracker to locate the object. This separate enhancement and tracking fails to build an end-to-end trainable vision system. To address this, we propose a novel architecture called Darkness Clue-Prompted Tracking (DCPT) that achieves robust UAV tracking at night by efficiently learning to generate darkness clue prompts. Without a separate enhancer, DCPT directly encodes anti-dark capabilities into prompts using a darkness clue prompter (DCP). Specifically, DCP iteratively learns emphasizing and undermining projections for darkness clues. It then injects these learned visual prompts into a daytime tracker with fixed parameters across transformer layers. Moreover, a gated feature aggregation mechanism enables adaptive fusion between prompts and between prompts and the base model. Extensive experiments show state-of-the-art performance for DCPT on multiple dark scenario benchmarks. The unified end-to-end learning of enhancement and tracking in DCPT enables a more trainable system. The darkness clue prompting efficiently injects anti-dark knowledge without extra modules. Code is available at https://github.com/bearyi26/DCPT.","{'model': 'tldr@v2.0.0', 'text': 'A novel architecture called Darkness Clue-Prompted Tracking (DCPT) is proposed that achieves robust UAV tracking at night by efficiently learning to generate darkness clue prompts and efficiently injects anti-dark knowledge without extra modules.'}",https://arxiv.org/pdf/2309.10491
-Comparative Analysis of Drug-GPT and ChatGPT LLMs for Healthcare Insights: Evaluating Accuracy and Relevance in Patient and HCP Contexts,Giorgos Lysandrou,"This study presents a comparative analysis of three Generative Pre-trained Transformer (GPT) solutions in a question and answer (Q&A) setting: Drug-GPT 3, Drug-GPT 4, and ChatGPT, in the context of healthcare applications. The objective is to determine which model delivers the most accurate and relevant information in response to prompts related to patient experiences with atopic dermatitis (AD) and healthcare professional (HCP) discussions about diabetes. The results demonstrate that while all three models are capable of generating relevant and accurate responses, Drug-GPT 3 and Drug-GPT 4, which are supported by curated datasets of patient and HCP social media and message board posts, provide more targeted and in-depth insights. ChatGPT, a more general-purpose model, generates broader and more general responses, which may be valuable for readers seeking a high-level understanding of the topics but may lack the depth and personal insights found in the answers generated by the specialized Drug-GPT models. This comparative analysis highlights the importance of considering the language model's perspective, depth of knowledge, and currency when evaluating the usefulness of generated information in healthcare applications.","{'model': 'tldr@v2.0.0', 'text': ""A comparative analysis of three Generative Pre-trained Transformer (GPT) solutions in a question and answer (Q&A) setting highlights the importance of considering the language model's perspective, depth of knowledge, and currency when evaluating the usefulness of generated information in healthcare applications.""}",https://arxiv.org/pdf/2307.16850
-A Tête-à-tête with ChatGPT on the impact of artificial intelligence in medical education.,Vignesh Ramachandran,"Chat Generative Pre-Trained Transformer (ChatGPT) is an artificial intelligence (AI) language model developed by OpenAI. It is trained to process vast amounts of text and engage in human-like conversational interaction with users. Being accessible by all, it is widely used and its capabilities range from language translation, summarising long texts and creative writing. This article explores the potential role of ChatGPT in medical education and the possible concerns about the misuse of this technology through a conversation with ChatGPT itself via text prompts. The implications of this technology in medical education as told by ChatGPT are interesting and seemingly helpful for both the students and the tutors. However, this could be a double-edged sword considering the risks of compromised students' integrity and concerns of over-reliance. This also calls for counter strategies and policies in place to mitigate these risks.","{'model': 'tldr@v2.0.0', 'text': ""The implications of this technology in medical education as told by ChatGPT are interesting and seemingly helpful for both the students and the tutors, however, this could be a double-edged sword considering the risks of compromised students' integrity and concerns of over-reliance.""}",
-Efficient domain adaptation of language models in ASR systems using Prompt-tuning,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems form a key component of various products across industry. Many of these ASR systems rely on a complex Acoustic Model (AM) whose output is rescored by a domain-specific Language Model (LM). Given the recent advancements in performance, their use in new domains is growing, and so is the memory, maintenance and data-collection costs for these domain-specific LMs. Particularly, with advent of parameter-heavy transformer based LMs (Devlin et al., 2019), maintaining multiple domain-specific LMs is practically infeasible. While on the other hand, using a generic LM for all domains falls short in performance when compared to multiple domain-specific LMs. Therefore, a need for a middle ground between performance and costs is evident. To overcome this problem, we bring forward a methodology based on recently proposed Prompt Tuning. Lester et al. (2021) introduced this idea of learning the token embeddings of the prompt used to prime a LM to a particular task. Prompts are special tokens describing a task which when appended to input data sample, helps the model understand and use this problem description to better solve the task. For example, to solve the machine translation task, instead of fine-tuning the transformer model with corresponding dataset, one can achieve comparable performance by just showing text describing machine translation to the powerful transformer-based LM. In prompt tuning, instead of providing this prompt manually to the model, one learn it from the labelled examples from the task. In our work, we generalize this idea from solving different tasks to domain adaptation. We identify that one can learn the promptembedding for not only different tasks but also dif-","{'model': 'tldr@v2.0.0', 'text': 'This work identifies that one can learn the promptembedding for not only different tasks but also differences in domain adaptation, and brings forward a methodology based on recently proposed Prompt Tuning.'}",
-All Birds with One Stone: Multi-task Learning for Inference with One Forward Pass,,"Task-specific fine-tuning of pre-trained lan- 001 guage models like Transformers has shown 002 their effectiveness in various NLP tasks. To 003 achieve better storage efficiency and model 004 performance, Multi-task Learning (MTL) has 005 been studied to share model parameters 006 and utilize knowledge transfer between tasks. 007 However, in real applications where enormous 008 numbers of tasks (e.g., large sets of labels 009 to be classified) need to be conducted on a 010 large corpus, the inference efficiency is still 011 hindered by the number of tasks. For a doc- 012 ument with N sets of labels to be predicted, 013 recent MTL methods with adaptive modules or 014 prompts need to encode the input data N times 015 to extract the hidden representation needed for 016 the tasks. Notice that the hidden represen- 017 tation is not sharable between tasks, as task- 018 specific features are extracted at very bottom 019 layers in the Transformer. In this paper, we 020 seek to maintain the computational efficiency 021 of only requiring one forward pass for a doc- 022 ument to get a generalized feature for all N 023 tasks, without sacrificing overall model perfor- 024 mance. We design a prompt-sharing module 025 to let the model take all tasks into considera- 026 tions and output N heads simultaneously. We 027 also design a dynamic task scheduling mod- 028 ule to sample tasks according to their training 029 progress. In our evaluation, we show that our 030 method is able to outperform previous MTL 031 state-of-the-arts and single task fine-tuning by 032 0 . 4 − 1 . 5% on GLUE benchmark dataset. We 033 also perform comprehensive module analysis 034 to demonstrate the effectiveness and robust- 035 ness of our method. 036","{'model': 'tldr@v2.0.0', 'text': 'This paper seeks to maintain the computational efficiency of only requiring one forward pass for a doc- 022 ument to get a generalized feature for all N 023 tasks, without sacriflcing overall model perfor- 024 mance, and designs a prompt-sharing module to let the model take all tasks into considera- 026 tions and output N heads simultaneously.'}",
-Evaluation of Interest and Coherence in Machine Generated Stories,Dominic Callan,". Evaluation of the narrative text generated by machines has traditionally been a challenge, particularly when attempting to evaluate subjective elements such as interest or believability. Recent improve-ments in narrative machine text generation have been largely driven by the emergence of transformer-based language models, trained on mas-sive quantities of data. In this study, a corpus of stories is generated using the pre-trained GPT-Neo transformer model, with human-written prompts. The stories generated through this process are subsequently evaluated through both human evaluation and two automated metrics: BERTScore and BERT Next-Sentence-Prediction. The results show vari-ation in human evaluation results in comparison to automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.","{'model': 'tldr@v2.0.0', 'text': 'Results show vari-ation in human evaluation results in comparison to automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.'}",
-Extracting Creativity from Narratives using Distributional Semantic Modeling,D. Johnson,"Narrative text permeates our lives from job applications to journalistic stories to works of fiction. Developing automated metrics that capture creativity in narrative text has potentially far reaching implications. Human ratings of creativity in narrative text are labor-intensive, subjective, and difficult to replicate. Across 27 different story prompts and over 3,500 short stories, we used distributional semantic modeling to automate the assessment of creativity in narrative texts. We tested a new metric to capture one key component of creativity in writing – a writer’s ability to connect divergent ideas. We termed this metric, word-to-word semantic diversity (w2w SemDiv). We compared six models of w2w SemDiv that varied in their computational architecture. The best performing model employed Bidirectional Encoder Representations Transformer (BERT), which generates context-dependent numerical representations of words (i.e., embeddings). The BERT w2w SemDiv scores demonstrated impressive predictive power, explaining up to 72% of the variance in human creativity ratings, even exceeding human inter-rater reliability for some tasks. In addition, w2w SemDiv scores generalized across Ethnicity and English language proficiency, including individuals identifying as Hispanic and L2 English speakers. We provide a tutorial with R code (osf.io/ath2s) on how to compute w2w SemDiv. This code is incorporated into an online web app (semdis.wlu.psu.edu) where researchers and educators can upload a data file with stories and freely retrieve w2w SemDiv scores.","{'model': 'tldr@v2.0.0', 'text': 'A new metric to capture one key component of creativity in writing – a writer’s ability to connect divergent ideas is tested, word-to-word semantic diversity (w2w SemDiv), which generalized across Ethnicity and English language proficiency, including individuals identifying as Hispanic and L2 English speakers.'}",
-Operating Frequency Characteristics of Piezoelectric High-Voltage Power Supply Converter,Bai Chun-yu,"The operating frequency characteristics of high-voltage power supply converter based on piezoelectric transformer(PT) is researched.The analysis of fundamental wave was used to get the equivalent model of converter output circuit and then the frequency characteristics of converter were investigated in the process of charging.It is found that the output equivalent capacitance increases with the increase of the operating periods.The equivalent resistance prompts rapidly with the increase of operating periods and achieves the maximum,then descends slowly and tends to a steady-state.In the whole process of the converter charging,the converter operating frequency descends with the increment of operating period,but the change is small,the operating frequency is only changed about 0.8 kHz in 90 000 operating periods(about 1.1 socend).",,
-Wearable Sensor-Based Human Activity Recognition with Transformer Model,Iveta Dirgová Luptáková,"Computing devices that can recognize various human activities or movements can be used to assist people in healthcare, sports, or human–robot interaction. Readily available data for this purpose can be obtained from the accelerometer and the gyroscope built into everyday smartphones. Effective classification of real-time activity data is, therefore, actively pursued using various machine learning methods. In this study, the transformer model, a deep learning neural network model developed primarily for the natural language processing and vision tasks, was adapted for a time-series analysis of motion signals. The self-attention mechanism inherent in the transformer, which expresses individual dependencies between signal values within a time series, can match the performance of state-of-the-art convolutional neural networks with long short-term memory. The performance of the proposed adapted transformer method was tested on the largest available public dataset of smartphone motion sensor data covering a wide range of activities, and obtained an average identification accuracy of 99.2% as compared with 89.67% achieved on the same data by a conventional machine learning method. The results suggest the expected future relevance of the transformer model for human activity recognition.","{'model': 'tldr@v2.0.0', 'text': 'The transformer model, a deep learning neural network model developed primarily for the natural language processing and vision tasks, was adapted for a time-series analysis of motion signals, and the expected future relevance of the transformer model for human activity recognition is suggested.'}",https://www.mdpi.com/1424-8220/22/5/1911/pdf?version=1646126183
-An End-to-End Transformer Model for 3D Object Detection,Ishan Misra,"We propose 3DETR, an end-to-end Transformer based object detection model for 3D point clouds. Compared to existing detection methods that employ a number of 3D-specific inductive biases, 3DETR requires minimal modifications to the vanilla Transformer block. Specifically, we find that a standard Transformer with non-parametric queries and Fourier positional embeddings is competitive with specialized architectures that employ libraries of 3D-specific operators with hand-tuned hyperparameters. Nevertheless, 3DETR is conceptually simple and easy to implement, enabling further improvements by incorporating 3D domain knowledge. Through extensive experiments, we show 3DETR outperforms the well-established and highly optimized VoteNet baselines on the challenging ScanNetV2 dataset by 9.5%. Furthermore, we show 3DETR is applicable to 3D tasks beyond detection, and can serve as a building block for future research.","{'model': 'tldr@v2.0.0', 'text': 'It is found that a standard Transformer with non-parametric queries and Fourier positional embeddings is competitive with specialized architectures that employ libraries of 3D-specific operators with hand-tuned hyperparameters.'}",https://arxiv.org/pdf/2109.08141
-"Embodied BERT: A Transformer Model for Embodied, Language-guided Visual Task Completion",Alessandro Suglia,"Language-guided robots performing home and office tasks must navigate in and interact with the world. Grounding language instructions against visual observations and actions to take in an environment is an open challenge. We present Embodied BERT (EmBERT), a transformer-based model which can attend to high-dimensional, multi-modal inputs across long temporal horizons for language-conditioned task completion. Additionally, we bridge the gap between successful object-centric navigation models used for non-interactive agents and the language-guided visual task completion benchmark, ALFRED, by introducing object navigation targets for EmBERT training. We achieve competitive performance on the ALFRED benchmark, and EmBERT marks the first transformer-based model to successfully handle the long-horizon, dense, multi-modal histories of ALFRED, and the first ALFRED model to utilize object-centric navigation targets.","{'model': 'tldr@v2.0.0', 'text': 'Embodied BERT (EmBERT) is presented, a transformer-based model which can attend to high-dimensional, multi-modal inputs across long temporal horizons for language-conditioned task completion and marks the first transformer- based model to successfully handle the long-horizon, dense, multi themodal histories of ALFRED, and the first AlFRED model to utilize object-centric navigation targets.'}",
-Swin Transformer: Hierarchical Vision Transformer using Shifted Windows,Ze Liu,"This paper presents a new vision Transformer, called Swin Transformer, that capably serves as a general-purpose backbone for computer vision. Challenges in adapting Transformer from language to vision arise from differences between the two domains, such as large variations in the scale of visual entities and the high resolution of pixels in images compared to words in text. To address these differences, we propose a hierarchical Transformer whose representation is computed with Shifted windows. The shifted windowing scheme brings greater efficiency by limiting self-attention computation to non-overlapping local windows while also allowing for cross-window connection. This hierarchical architecture has the flexibility to model at various scales and has linear computational complexity with respect to image size. These qualities of Swin Transformer make it compatible with a broad range of vision tasks, including image classification (87.3 top-1 accuracy on ImageNet-1K) and dense prediction tasks such as object detection (58.7 box AP and 51.1 mask AP on COCO test-dev) and semantic segmentation (53.5 mIoU on ADE20K val). Its performance surpasses the previous state-of-the-art by a large margin of +2.7 box AP and +2.6 mask AP on COCO, and +3.2 mIoU on ADE20K, demonstrating the potential of Transformer-based models as vision backbones. The hierarchical design and the shifted window approach also prove beneficial for all-MLP architectures. The code and models are publicly available at https://github.com/microsoft/Swin-Transformer.","{'model': 'tldr@v2.0.0', 'text': 'A hierarchical Transformer whose representation is computed with Shifted windows, which has the flexibility to model at various scales and has linear computational complexity with respect to image size and will prove beneficial for all-MLP architectures.'}",https://arxiv.org/pdf/2103.14030
-Advancing Plain Vision Transformer Toward Remote Sensing Foundation Model,Di Wang,"Large-scale vision foundation models have made significant progress in visual tasks on natural images, with vision transformers (ViTs) being the primary choice due to their good scalability and representation ability. However, large-scale models in remote sensing (RS) have not yet been sufficiently explored. In this article, we resort to plain ViTs with about 100 million parameters and make the first attempt to propose large vision models tailored to RS tasks and investigate how such large models perform. To handle the large sizes and objects of arbitrary orientations in RS images, we propose a new rotated varied-size window attention to replace the original full attention in transformers, which can significantly reduce the computational cost and memory footprint while learning better object representation by extracting rich context from the generated diverse windows. Experiments on detection tasks show the superiority of our model over all state-of-the-art models, achieving 81.24% mean average precision (mAP) on the DOTA-V1.0 dataset. The results of our models on downstream classification and segmentation tasks also show competitive performance compared to existing advanced methods. Further experiments show the advantages of our models in terms of computational complexity and data efficiency in transferring. The code and models will be released at https://github.com/ViTAE-Transformer/Remote-Sensing-RVSA.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes a new rotated varied-size window attention to replace the original full attention in transformers, which can significantly reduce the computational cost and memory footprint while learning better object representation by extracting rich context from the generated diverse windows.'}",http://arxiv.org/pdf/2208.03987
-Streaming Automatic Speech Recognition with the Transformer Model,Niko Moritz,"Encoder-decoder based sequence-to-sequence models have demonstrated state-of-the-art results in end-to-end automatic speech recognition (ASR). Recently, the transformer architecture, which uses self-attention to model temporal context information, has been shown to achieve significantly lower word error rates (WERs) compared to recurrent neural network (RNN) based system architectures. Despite its success, the practical usage is limited to offline ASR tasks, since encoder-decoder architectures typically require an entire speech utterance as input. In this work, we propose a transformer based end-to-end ASR system for streaming ASR, where an output must be generated shortly after each spoken word. To achieve this, we apply time-restricted self-attention for the encoder and triggered attention for the encoder-decoder attention mechanism. Our proposed streaming transformer architecture achieves 2.8% and 7.3% WER for the ""clean"" and ""other"" test data of LibriSpeech, which to our knowledge is the best published streaming end-to-end ASR result for this task.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transformer based end-to-end ASR system for streaming ASR, where an output must be generated shortly after each spoken word, and applies time-restricted self-attention for the encoder and triggered attention for theEncoder-decoder attention mechanism.'}",https://arxiv.org/pdf/2001.02674
-OPT: Open Pre-trained Transformer Language Models,Susan Zhang,"Large language models, which are often trained for hundreds of thousands of compute days, have shown remarkable capabilities for zero- and few-shot learning. Given their computational cost, these models are difficult to replicate without significant capital. For the few that are available through APIs, no access is granted to the full model weights, making them difficult to study. We present Open Pre-trained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M to 175B parameters, which we aim to fully and responsibly share with interested researchers. We show that OPT-175B is comparable to GPT-3, while requiring only 1/7th the carbon footprint to develop. We are also releasing our logbook detailing the infrastructure challenges we faced, along with code for experimenting with all of the released models.","{'model': 'tldr@v2.0.0', 'text': 'Open Pre-trained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M to 175B parameters, is presented, which is aimed to fully and responsibly share with interested researchers.'}",
-A Multiscale Visualization of Attention in the Transformer Model,Jesse Vig,"The Transformer is a sequence model that forgoes traditional recurrent architectures in favor of a fully attention-based approach. Besides improving performance, an advantage of using attention is that it can also help to interpret a model by showing how the model assigns weight to different input elements. However, the multi-layer, multi-head attention mechanism in the Transformer model can be difficult to decipher. To make the model more accessible, we introduce an open-source tool that visualizes attention at multiple scales, each of which provides a unique perspective on the attention mechanism. We demonstrate the tool on BERT and OpenAI GPT-2 and present three example use cases: detecting model bias, locating relevant attention heads, and linking neurons to model behavior.","{'model': 'tldr@v2.0.0', 'text': 'An open-source tool that visualizes attention at multiple scales, each of which provides a unique perspective on the attention mechanism in the Transformer model is introduced.'}",https://www.aclweb.org/anthology/P19-3007.pdf
-CNN-Based Transformer Model for Fault Detection in Power System Networks,Jibin B. Thomas,"Fault detection and localization in electrical power lines has long been a crucial challenge for electrical engineers as it allows the detected fault to be isolated and recovered promptly. These faults, if neglected, can rupture the normal operation of the network and drastically damage the power lines and the equipment attached to it. The wastage of power and money due to these faults can be harmful to the economy of an industry or even a country. Therefore, efficient fault detection mechanisms have become crucial for the well-being of this power-hungry world. This research presents an end-to-end deep learning strategy to detect and localize symmetrical and unsymmetrical faults as well as high-impedance faults (HIFs) in a distribution system. This research proposes a novel deep convolutional neural network (CNN) transformer model to automatically detect the type and phase of the fault as well as the location of the fault. The proposed model utilizes 1-D deep CNNs for feature extraction and transformer encoder for sequence learning. The transformer encoder utilizes an attention mechanism to integrate the sequence embeddings and focus on significant time steps to learn long-term dependence to extract the context of the temporal current data. The different faults were simulated in MATLAB Simulink using IEEE 14-bus distribution system. The proposed models were found to produce better performance on the test database when evaluated using F1-Score, Matthews correlation coefficient (MCC), and accuracy. The models also produced better predictions on HIFs compared to conventional fault-detection techniques.",,
-A Transformer Model for Retrosynthesis,Pavel Karpov,,"{'model': 'tldr@v2.0.0', 'text': 'A Transformer model for a retrosynthetic reaction prediction task is described and it is found that snapshot learning with averaging weights on learning rates minima works best.'}",
-LSTM-embedding Transformer model for LoS/NLoS recognition in indoor localization,Lucong Chen,"The development of 5G networks has prompted higher localization accuracy requirements in indoor scenarios within the Industrial Internet of Things (IIoT). In the scenarios of indoor localization, mitigating the adverse effects caused by Non-Line-of-Sight (NLoS) propagation has been a longstanding and significant challenge. Despite the availability of numerous classifiers for distinguishing between Line-of-Sight (LoS) and NLoS paths, there are still some shortcomings in adequately capturing the long-term time dependence of signal sequences. In this study, to enhance the extraction of long-term dependencies and global information of time series, we propose the Long Short-Term Memory (LSTM)-embedding transformer model to identify the LoS and NLoS paths from the Power Delay Profile (PDP), which characterizes the amplitude decay of the signal over time. And the Time-of-Arrival (ToA) information of the LoS paths is exploited for localization. Lastly, to evaluate the robustness of our model, the model which is trained in the Indoor Factory Dense High (InF-DH) scenario is transferred to the Indoor Factory Sparse High (InF-SH) scenario. Experimental results show that our proposed model reached 96.04% recognition success in the InF-DH scenario and 93.27% recognition success in the migration scenario (InF-SH), with some improvements compared to the baseline model.","{'model': 'tldr@v2.0.0', 'text': 'To enhance the extraction of long-term dependencies and global information of time series, the Long Short-Term Memory-embedding transformer model is proposed to identify the LoS and NLoS paths from the Power Delay Profile (PDP), which characterizes the amplitude decay of the signal over time.'}",
-EpiGePT: a Pretrained Transformer model for epigenomics,Zijing Gao,"The transformer-based models, such as GPT-31 and DALL-E2, have achieved unprecedented breakthroughs in the field of natural language processing and computer vision. The inherent similarities between natural language and biological sequences have prompted a new wave of inferring the grammatical rules underneath the biological sequences. In genomic study, it is worth noting that DNA sequences alone cannot explain all the gene activities due to epigenetic mechanism. To investigate this problem, we propose EpiGePT, a new transformer-based language pretrained model in epigenomics, for predicting genome-wide epigenomic signals by considering the mechanistic modeling of transcriptional regulation. Specifically, EpiGePT takes the context-specific activities of transcription factors (TFs) into consideration, which could offer deeper biological insights comparing to models trained on DNA sequence only. In a series of experiments, EpiGePT demonstrates state-of-the-art performance in a diverse epigenomic signals prediction tasks as well as new prediction tasks by fine-tuning. Furthermore, EpiGePT is capable of learning the cell-type-specific long-range interactions through the self-attention mechanism and interpreting the genetic variants that associated with human diseases. We expect that the advances of EpiGePT can shed light on understanding the complex regulatory mechanisms in gene regulation. We provide free online prediction service of EpiGePT through https://health.tsinghua.edu.cn/epigept/.","{'model': 'tldr@v2.0.0', 'text': 'EpiGePT, a new transformer-based language pretrained model in epigenomics, is proposed for predicting genome-wide epigenomic signals by considering the mechanistic modeling of transcriptional regulation and is capable of learning the cell-type-specific long-range interactions through the self-attention mechanism.'}",
-Insights from Attention: New Approaches to Visualizing Transformer Model Output,Raphael Kunert,". Recent advancements in language models, particularly those based on the Transformer architecture, have led to remarkable achieve-ments in natural language processing. However, the increasing complex-ity and size of these models pose significant challenges for understanding their behavior and decision-making processes. In this work, we propose a set of new attention visualization techniques that address these challenges by improving model explainability and interpretability. The key improvements include new layouts that better handle the large number of tokens present in prompt and answer scenarios, making long distance attention relationships more comprehensible. Our techniques have the potential to enable researchers and practitioners to better understand the decision-making processes of large language models and identify potential sources of bias or errors. While detailed user studies and evalua-tions are outside the scope of this work, we discuss potential use cases for our visualization techniques and present directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'A set of new attention visualization techniques that address challenges of large language models by improving model explainability and interpretability and making long distance attention relationships more comprehensible are proposed.'}",
-MaskGIT: Masked Generative Image Transformer,Huiwen Chang,"Generative transformers have experienced rapid popularity growth in the computer vision community in synthesizing high-fidelity and high-resolution images. The best generative transformer models so far, however, still treat an image naively as a sequence of tokens, and decode an image sequentially following the raster scan ordering (i.e. line-by-line). We find this strategy neither optimal nor efficient. This paper proposes a novel image synthesis paradigm using a bidirectional transformer decoder, which we term MaskGIT. During training, MaskGIT learns to predict randomly masked tokens by attending to tokens in all directions. At inference time, the model begins with generating all tokens of an image simultaneously, and then refines the image iteratively conditioned on the previous generation. Our experiments demonstrate that MaskGIT significantly outperforms the state-of-the-art transformer model on the ImageNet dataset, and accelerates autoregressive decoding by up to 48x. Besides, we illustrate that MaskGIT can be easily extended to various image editing tasks, such as inpainting, extrapolation, and image manipulation. Project page: masked-generative-image-transformer.github.io.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MaskGIT is a novel image synthesis paradigm using a bidirectional transformer decoder that significantly outperforms the state-of-the-art transformer model on the ImageNet dataset, and accelerates autoregressive decoding by up to 48x.'}",https://arxiv.org/pdf/2202.04200
-Medical Intervention Duration Estimation Using Language-enhanced Transformer Encoder with Medical Prompts,Yucheng Ruan,"In recent years, estimating the duration of medical intervention based on electronic health records (EHRs) has gained significant attention in the filed of clinical decision support. However, current models largely focus on structured data, leaving out information from the unstructured clinical free-text data. To address this, we present a novel language-enhanced transformer-based framework, which projects all relevant clinical data modalities (continuous, categorical, binary, and free-text features) into a harmonized language latent space using a pre-trained sentence encoder with the help of medical prompts. The proposed method enables the integration of information from different modalities within the cell transformer encoder and leads to more accurate duration estimation for medical intervention. Our experimental results on both US-based (length of stay in ICU estimation) and Asian (surgical duration prediction) medical datasets demonstrate the effectiveness of our proposed framework, which outperforms tailored baseline approaches and exhibits robustness to data corruption in EHRs.","{'model': 'tldr@v2.0.0', 'text': 'A novel language-enhanced transformer-based framework, which projects all relevant clinical data modalities into a harmonized language latent space using a pre-trained sentence encoder with the help of medical prompts, leads to more accurate duration estimation for medical intervention.'}",https://arxiv.org/pdf/2303.17408
-Training Tips for the Transformer Model,M. Popel,"Abstract This article describes our experiments in neural machine translation using the recent Tensor2Tensor framework and the Transformer sequence-to-sequence model (Vaswani et al., 2017). We examine some of the critical parameters that affect the final translation quality, memory usage, training stability and training time, concluding each experiment with a set of recommendations for fellow researchers. In addition to confirming the general mantra “more data and larger models”, we address scaling to multiple GPUs and provide practical tips for improved training regarding batch size, learning rate, warmup steps, maximum sentence length and checkpoint averaging. We hope that our observations will allow others to get better results given their particular hardware and data constraints.","{'model': 'tldr@v2.0.0', 'text': 'The experiments in neural machine translation using the recent Tensor2Tensor framework and the Transformer sequence-to-sequence model are described, confirming the general mantra “more data and larger models”.'}",https://content.sciendo.com/downloadpdf/journals/pralin/110/1/article-p43.pdf
-Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss,Qian Zhang,"In this paper we present an end-to-end speech recognition model with Transformer encoders that can be used in a streaming speech recognition system. Transformer computation blocks based on self-attention are used to encode both audio and label sequences independently. The activations from both audio and label encoders are combined with a feed-forward layer to compute a probability distribution over the label space for every combination of acoustic frame position and label history. This is similar to the Recurrent Neural Network Transducer (RNN-T) model, which uses RNNs for information encoding instead of Transformer encoders. The model is trained with the RNN-T loss well-suited to streaming decoding. We present results on the LibriSpeech dataset showing that limiting the left context for self-attention in the Transformer layers makes decoding computationally tractable for streaming, with only a slight degradation in accuracy. We also show that the full attention version of our model beats the-state-of-the art accuracy on the LibriSpeech benchmarks. Our results also show that we can bridge the gap between full attention and limited attention versions of our model by attending to a limited number of future frames.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end speech recognition model with Transformer encoders that can be used in a streaming speech recognition system and shows that the full attention version of the model beats the-state-of-the art accuracy on the LibriSpeech benchmarks.'}",https://arxiv.org/pdf/2002.02562
-GIT: A Generative Image-to-text Transformer for Vision and Language,Jianfeng Wang,"In this paper, we design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering. While generative models provide a consistent network architecture between pre-training and fine-tuning, existing work typically contains complex structures (uni/multi-modal encoder/decoder) and depends on external modules such as object detectors/taggers and optical character recognition (OCR). In GIT, we simplify the architecture as one image encoder and one text decoder under a single language modeling task. We also scale up the pre-training data and the model size to boost the model performance. Without bells and whistles, our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr). Furthermore, we present a new scheme of generation-based image classification and scene text recognition, achieving decent performance on standard benchmarks. Codes are released at \url{https://github.com/microsoft/GenerativeImage2Text}.","{'model': 'tldr@v2.0.0', 'text': 'This paper design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering and establishes new state of the arts on 12 challenging benchmarks with a large margin.'}",http://arxiv.org/pdf/2205.14100
-BioGPT: Generative Pre-trained Transformer for Biomedical Text Generation and Mining,Renqian Luo,"Pre-trained language models have attracted increasing attention in the biomedical domain, inspired by their great success in the general natural language domain. Among the two main branches of pre-trained language models in the general language domain, i.e. BERT (and its variants) and GPT (and its variants), the first one has been extensively studied in the biomedical domain, such as BioBERT and PubMedBERT. While they have achieved great success on a variety of discriminative downstream biomedical tasks, the lack of generation ability constrains their application scope. In this paper, we propose BioGPT, a domain-specific generative Transformer language model pre-trained on large-scale biomedical literature. We evaluate BioGPT on six biomedical natural language processing tasks and demonstrate that our model outperforms previous models on most tasks. Especially, we get 44.98%, 38.42% and 40.76% F1 score on BC5CDR, KD-DTI and DDI end-to-end relation extraction tasks, respectively, and 78.2% accuracy on PubMedQA, creating a new record. Our case study on text generation further demonstrates the advantage of BioGPT on biomedical literature to generate fluent descriptions for biomedical terms.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes BioGPT, a domain-specific generative Transformer language model pre-trained on large-scale biomedical literature and evaluates it on six biomedical natural language processing tasks and demonstrates that the model outperforms previous models on most tasks.'}",https://arxiv.org/pdf/2210.10341
-HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection,Ke Chen,"Audio classification is an important task of mapping audio samples into their corresponding labels. Recently, the transformer model with self-attention mechanisms has been adopted in this field. However, existing audio transformers require large GPU memories and long training time, meanwhile relying on pretrained vision models to achieve high performance, which limits the model’s scalability in audio tasks. To combat these problems, we introduce HTS-AT: an audio transformer with a hierarchical structure to reduce the model size and training time. It is further combined with a token-semantic module to map final outputs into class featuremaps, thus enabling the model for the audio event detection (i.e. localization in time). We evaluate HTS-AT on three datasets of audio classification where it achieves new state-of-the-art (SOTA) results on AudioSet and ESC50, and equals the SOTA on Speech Command V2. It also achieves better performance in event localization than the previous CNN-based models. Moreover, HTS-AT requires only 35% model parameters and 15% training time of the previous audio transformer. These results demonstrate the high performance and high efficiency of HTS-AT.","{'model': 'tldr@v2.0.0', 'text': 'HTS-AT is introduced: an audio transformer with a hierarchical structure to reduce the model size and training time, and is further combined with a token-semantic module to map final outputs into class featuremaps, thus enabling the model for the audio event detection and localization in time.'}",https://arxiv.org/pdf/2202.00874
-MaxViT: Multi-Axis Vision Transformer,Zhengzhong Tu,"Transformers have recently gained significant attention in the computer vision community. However, the lack of scalability of self-attention mechanisms with respect to image size has limited their wide adoption in state-of-the-art vision backbones. In this paper we introduce an efficient and scalable attention model we call multi-axis attention, which consists of two aspects: blocked local and dilated global attention. These design choices allow global-local spatial interactions on arbitrary input resolutions with only linear complexity. We also present a new architectural element by effectively blending our proposed attention model with convolutions, and accordingly propose a simple hierarchical vision backbone, dubbed MaxViT, by simply repeating the basic building block over multiple stages. Notably, MaxViT is able to ''see'' globally throughout the entire network, even in earlier, high-resolution stages. We demonstrate the effectiveness of our model on a broad spectrum of vision tasks. On image classification, MaxViT achieves state-of-the-art performance under various settings: without extra data, MaxViT attains 86.5% ImageNet-1K top-1 accuracy; with ImageNet-21K pre-training, our model achieves 88.7% top-1 accuracy. For downstream tasks, MaxViT as a backbone delivers favorable performance on object detection as well as visual aesthetic assessment. We also show that our proposed model expresses strong generative modeling capability on ImageNet, demonstrating the superior potential of MaxViT blocks as a universal vision module. The source code and trained models will be available at https://github.com/google-research/maxvit.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an efficient and scalable attention model, which consists of two aspects: blocked local and dilated global attention, and proposes a simple hierarchical vision backbone, dubbed MaxViT, by simply repeating the basic building block over multiple stages.'}",http://arxiv.org/pdf/2204.01697
-MAT: Mask-Aware Transformer for Large Hole Image Inpainting,Wenbo Li,"Recent studies have shown the importance of modeling long-range interactions in the inpainting problem. To achieve this goal, existing approaches exploit either standalone attention techniques or transformers, but usually under a low resolution in consideration of computational cost. In this paper, we present a novel transformer-based model for large hole inpainting, which unifies the merits of transformers and convolutions to efficiently process high-resolution images. We carefully design each component of our framework to guarantee the high fidelity and diversity of recovered images. Specifically, we customize an inpainting-oriented transformer block, where the attention module aggregates non-local information only from partial valid tokens, indicated by a dynamic mask. Extensive experiments demonstrate the state-of-the-art performance of the new model on multiple benchmark datasets. Code is released at https://github.com/fenglinglwb/MAT.","{'model': 'tldr@v2.0.0', 'text': 'A novel transformer-based model for large hole inpainting is presented, which unifies the merits of transformers and convolutions to efficiently process high-resolution images.'}",https://arxiv.org/pdf/2203.15270
-Vision Transformer Adapter for Dense Predictions,Zhe Chen,"This work investigates a simple yet powerful dense prediction task adapter for Vision Transformer (ViT). Unlike recently advanced variants that incorporate vision-specific inductive biases into their architectures, the plain ViT suffers inferior performance on dense predictions due to weak prior assumptions. To address this issue, we propose the ViT-Adapter, which allows plain ViT to achieve comparable performance to vision-specific transformers. Specifically, the backbone in our framework is a plain ViT that can learn powerful representations from large-scale multi-modal data. When transferring to downstream tasks, a pre-training-free adapter is used to introduce the image-related inductive biases into the model, making it suitable for these tasks. We verify ViT-Adapter on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation. Notably, without using extra detection data, our ViT-Adapter-L yields state-of-the-art 60.9 box AP and 53.0 mask AP on COCO test-dev. We hope that the ViT-Adapter could serve as an alternative for vision-specific transformers and facilitate future research. The code and models will be released at https://github.com/czczup/ViT-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'The ViT-Adapter is proposed, which allows plain ViT to achieve comparable performance to vision-specific transformers and facilitate future research and is verified on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation.'}",http://arxiv.org/pdf/2205.08534
-CTRL: A Conditional Transformer Language Model for Controllable Generation,N. Keskar,"Large-scale language models show promising text generation capabilities, but users cannot easily control particular aspects of the generated text. We release CTRL, a 1.63 billion-parameter conditional transformer language model, trained to condition on control codes that govern style, content, and task-specific behavior. Control codes were derived from structure that naturally co-occurs with raw text, preserving the advantages of unsupervised learning while providing more explicit control over text generation. These codes also allow CTRL to predict which parts of the training data are most likely given a sequence. This provides a potential method for analyzing large amounts of data via model-based source attribution. We have released multiple full-sized, pretrained versions of CTRL at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'CTRL is released, a 1.63 billion-parameter conditional transformer language model, trained to condition on control codes that govern style, content, and task-specific behavior, providing more explicit control over text generation.'}",
-Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer,Colin Raffel,"Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new ""Colossal Clean Crawled Corpus"", we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.","{'model': 'tldr@v2.0.0', 'text': 'This systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks and achieves state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more.'}",
-SwinIR: Image Restoration Using Swin Transformer,Jingyun Liang,"Image restoration is a long-standing low-level vision problem that aims to restore high-quality images from low-quality images (e.g., downscaled, noisy and compressed images). While state-of-the-art image restoration methods are based on convolutional neural networks, few attempts have been made with Transformers which show impressive performance on high-level vision tasks. In this paper, we propose a strong baseline model SwinIR for image restoration based on the Swin Transformer. SwinIR consists of three parts: shallow feature extraction, deep feature extraction and high-quality image reconstruction. In particular, the deep feature extraction module is composed of several residual Swin Transformer blocks (RSTB), each of which has several Swin Transformer layers together with a residual connection. We conduct experiments on three representative tasks: image super-resolution (including classical, lightweight and real-world image super-resolution), image denoising (including grayscale and color image denoising) and JPEG compression artifact reduction. Experimental results demonstrate that SwinIR outperforms state-of-the-art methods on different tasks by up to 0.14∼0.45dB, while the total number of parameters can be reduced by up to 67%.","{'model': 'tldr@v2.0.0', 'text': 'A strong baseline model SwinIR is proposed for image restoration based on the Swin Transformer that outperforms state-of-the-art methods on different tasks by up to 0.14∼0.45dB, while the total number of parameters can be reduced byUp to 67%.'}",https://arxiv.org/pdf/2108.10257
-ViViT: A Video Vision Transformer,Anurag Arnab,"We present pure-transformer based models for video classification, drawing upon the recent success of such models in image classification. Our model extracts spatiotemporal tokens from the input video, which are then encoded by a series of transformer layers. In order to handle the long sequences of tokens encountered in video, we propose several, efficient variants of our model which factorise the spatial- and temporal-dimensions of the input. Although transformer-based models are known to only be effective when large training datasets are available, we show how we can effectively regularise the model during training and leverage pretrained image models to be able to train on comparatively small datasets. We conduct thorough ablation studies, and achieve state-of-the-art results on multiple video classification benchmarks including Kinetics 400 and 600, Epic Kitchens, Something-Something v2 and Moments in Time, outperforming prior methods based on deep 3D convolutional networks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows how to effectively regularise the model during training and leverage pretrained image models to be able to train on comparatively small datasets, and achieves state-of-the-art results on multiple video classification benchmarks.'}",https://arxiv.org/pdf/2103.15691
-Decision Transformer: Reinforcement Learning via Sequence Modeling,Lili Chen,"We introduce a framework that abstracts Reinforcement Learning (RL) as a sequence modeling problem. This allows us to draw upon the simplicity and scalability of the Transformer architecture, and associated advances in language modeling such as GPT-x and BERT. In particular, we present Decision Transformer, an architecture that casts the problem of RL as conditional sequence modeling. Unlike prior approaches to RL that fit value functions or compute policy gradients, Decision Transformer simply outputs the optimal actions by leveraging a causally masked Transformer. By conditioning an autoregressive model on the desired return (reward), past states, and actions, our Decision Transformer model can generate future actions that achieve the desired return. Despite its simplicity, Decision Transformer matches or exceeds the performance of state-of-the-art model-free offline RL baselines on Atari, OpenAI Gym, and Key-to-Door tasks.","{'model': 'tldr@v2.0.0', 'text': 'Despite its simplicity, Decision Transformer matches or exceeds the performance of state-of-the-art model-free offline RL baselines on Atari, OpenAI Gym, and Key-to-Door tasks.'}",
-Restormer: Efficient Transformer for High-Resolution Image Restoration,Syed Waqas Zamir,"Since convolutional neural networks (CNNs) perform well at learning generalizable image priors from large-scale data, these models have been extensively applied to image restoration and related tasks. Recently, another class of neural architectures, Transformers, have shown significant performance gains on natural language and high-level vision tasks. While the Transformer model mitigates the shortcomings of CNNs (i.e., limited receptive field and inadaptability to input content), its computational complexity grows quadratically with the spatial resolution, therefore making it infeasible to apply to most image restoration tasks involving high-resolution images. In this work, we propose an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images. Our model, named Restoration Transformer (Restormer), achieves state-of-the-art results on several image restoration tasks, including image deraining, single-image motion deblurring, defocus deblurring (single-image and dual-pixel data), and image denoising (Gaussian grayscale/color denoising, and real image denoising). The source code and pre-trained models are available at https://github.com/swz30/Restormer.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images.'}",https://arxiv.org/pdf/2111.09881
-Segmenter: Transformer for Semantic Segmentation,Robin Strudel,"Image segmentation is often ambiguous at the level of individual image patches and requires contextual information to reach label consensus. In this paper we introduce Segmenter, a transformer model for semantic segmentation. In contrast to convolution-based methods, our approach allows to model global context already at the first layer and throughout the network. We build on the recent Vision Transformer (ViT) and extend it to semantic segmentation. To do so, we rely on the output embeddings corresponding to image patches and obtain class labels from these embed-dings with a point-wise linear decoder or a mask trans-former decoder. We leverage models pre-trained for image classification and show that we can fine-tune them on moderate sized datasets available for semantic segmentation. The linear decoder allows to obtain excellent results already, but the performance can be further improved by a mask transformer generating class masks. We conduct an extensive ablation study to show the impact of the different parameters, in particular the performance is better for large models and small patch sizes. Segmenter attains excellent results for semantic segmentation. It outperforms the state of the art on both ADE20K and Pascal Context datasets and is competitive on Cityscapes.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Segmenter, a transformer model for semantic segmentation that outperforms the state of the art on both ADE20K and Pascal Context datasets and is competitive on Cityscapes.'}",https://arxiv.org/pdf/2105.05633
-ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision,Wonjae Kim,"Vision-and-Language Pre-training (VLP) has improved performance on various joint vision-and-language downstream tasks. Current approaches to VLP heavily rely on image feature extraction processes, most of which involve region supervision (e.g., object detection) and the convolutional architecture (e.g., ResNet). Although disregarded in the literature, we find it problematic in terms of both (1) efficiency/speed, that simply extracting input features requires much more computation than the multimodal interaction steps; and (2) expressive power, as it is upper bounded to the expressive power of the visual embedder and its predefined visual vocabulary. In this paper, we present a minimal VLP model, Vision-and-Language Transformer (ViLT), monolithic in the sense that the processing of visual inputs is drastically simplified to just the same convolution-free manner that we process textual inputs. We show that ViLT is up to tens of times faster than previous VLP models, yet with competitive or better downstream task performance. Our code and pre-trained weights are available at https://github.com/dandelin/vilt.","{'model': 'tldr@v2.0.0', 'text': 'A minimal VLP model, Vision-and-Language Transformer (ViLT), monolithic in the sense that the processing of visual inputs is drastically simplified to just the same convolution-free manner that the authors process textual inputs, showing that ViLT is up to tens of times faster than previous VLP models, yet with competitive or better downstream task performance.'}",
-Improved low-frequency transformer model based on Jiles–Atherton hysteresis theory,Mi Zou,"Transformers are the most widely used apparatus in power systems to transfer electrical energy from one side to another. Despite remarkable achievements in transformer modelling over the past few years, three-phase transformer modelling techniques have not progressed to that of their single-phase counterparts, mainly because of topological correctness of core and coil structure, complicated interaction between fluxes of the three iron core legs, and non-linear saturation of ferromagnetic material. An improved three-phase transformer model with extended inverse Jiles-Atherton (JA) hysteresis theory is proposed in this study. The extended inverse JA model with anisotropic anhysteretic magnetisation function is addressed, and the classical eddy-current-loss term and anomalous-loss term are directly incorporated into the original energy balance equation to represent the non-linearity and iron core losses rather than representing them with a parallel combination of resistances and saturable inductances (R//L model) as most previous studies have done. The results of inrush current case indicate that the proposed transformer model can predict the residual flux, include eddy-current-loss and anomalous-loss, and predict the inrush current peak values with relative error less than 6%.",,
-Recurrent Memory Transformer,Aydar Bulatov,"Transformer-based models show their effectiveness across multiple domains and tasks. The self-attention allows to combine information from all sequence elements into context-aware representations. However, global and local information has to be stored mostly in the same element-wise representations. Moreover, the length of an input sequence is limited by quadratic computational complexity of self-attention. In this work, we propose and study a memory-augmented segment-level recurrent Transformer (RMT). Memory allows to store and process local and global information as well as to pass information between segments of the long sequence with the help of recurrence. We implement a memory mechanism with no changes to Transformer model by adding special memory tokens to the input or output sequence. Then the model is trained to control both memory operations and sequence representations processing. Results of experiments show that RMT performs on par with the Transformer-XL on language modeling for smaller memory sizes and outperforms it for tasks that require longer sequence processing. We show that adding memory tokens to Tr-XL is able to improve its performance. This makes Recurrent Memory Transformer a promising architecture for applications that require learning of long-term dependencies and general purpose in memory processing, such as algorithmic tasks and reasoning.","{'model': 'tldr@v2.0.0', 'text': 'Recurrent Memory Transformer is a promising architecture for applications that require learning of long-term dependencies and general purpose in memory processing, such as algorithmic tasks and reasoning.'}",https://arxiv.org/pdf/2207.06881
-Exploring Visual Prompts for Adapting Large-Scale Models,Hyojin Bahng,"We investigate the efficacy of visual prompting to adapt large-scale models in vision. Following the recent approach from prompt tuning and adversarial reprogramming, we learn a single image perturbation such that a frozen model prompted with this perturbation performs a new task. Through comprehensive experiments, we demonstrate that visual prompting is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes. We further analyze properties of the downstream dataset, prompt design, and output transformation in regard to adaptation performance. The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision. Code is available at http://hjbahng.github.io/visual_prompting .","{'model': 'tldr@v2.0.0', 'text': 'The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision and is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes.'}",
-Video Swin Transformer,Ze Liu,"The vision community is witnessing a modeling shift from CNNs to Transformers, where pure Transformer architectures have attained top accuracy on the major video recognition benchmarks. These video models are all built on Transformer layers that globally connect patches across the spatial and temporal dimensions. In this paper, we instead advocate an inductive bias of locality in video Transformers, which leads to a better speed-accuracy trade-off compared to previous approaches which compute self-attention globally even with spatial-temporal factorization. The locality of the proposed video architecture is realized by adapting the Swin Transformer designed for the image domain, while continuing to leverage the power of pre-trained image models. Our approach achieves state-of-the-art accuracy on a broad range of video recognition benchmarks, including on action recognition (84.9 top-l accuracy on Kinetics-400 and 85.9 top-l accuracy on Kinetics-600 with ~20× less pre-training data and ~3× smaller model size) and temporal modeling (69.6 top-l accuracy on Something-Something v2).","{'model': 'tldr@v2.0.0', 'text': 'This paper advocates an inductive bias of locality in video Transformers, which leads to a better speed-accuracy trade-off compared to previous approaches which compute self-attention globally even with spatial-temporal factorization.'}",
-Efficiently Scaling Transformer Inference,Reiner Pope,"We study the problem of efficient generative inference for Transformer models, in one of its most challenging settings: large deep models, with tight latency targets and long sequence lengths. Better understanding of the engineering tradeoffs for inference for large Transformer-based models is important as use cases of these models are growing rapidly throughout application areas. We develop a simple analytical model for inference efficiency to select the best multi-dimensional partitioning techniques optimized for TPU v4 slices based on the application requirements. We combine these with a suite of low-level optimizations to achieve a new Pareto frontier on the latency and model FLOPS utilization (MFU) tradeoffs on 500B+ parameter models that outperforms the FasterTransformer suite of benchmarks. We further show that with appropriate partitioning, the lower memory requirements of multiquery attention (i.e. multiple query heads share single key/value head) enables scaling up to 32x larger context lengths. Finally, we achieve a low-batch-size latency of 29ms per token during generation (using int8 weight quantization) and a 76% MFU during large-batch-size processing of input tokens, while supporting a long 2048-token context length on the PaLM 540B parameter model.","{'model': 'tldr@v2.0.0', 'text': 'A simple analytical model for inference efficiency is developed to select the best multi-dimensional partitioning techniques optimized for TPU v4 slices based on the application requirements and a suite of low-level optimizations are combined to achieve a new Pareto frontier on the latency and model FLOPS utilization tradeoffs on 500B+ parameter models that outperforms the FasterTransformer suite of benchmarks.'}",https://arxiv.org/pdf/2211.05102
-Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction,P. Schwaller,"Organic synthesis is one of the key stumbling blocks in medicinal chemistry. A necessary yet unsolved step in planning synthesis is solving the forward problem: Given reactants and reagents, predict the products. Similar to other work, we treat reaction prediction as a machine translation problem between simplified molecular-input line-entry system (SMILES) strings (a text-based representation) of reactants, reagents, and the products. We show that a multihead attention Molecular Transformer model outperforms all algorithms in the literature, achieving a top-1 accuracy above 90% on a common benchmark data set. Molecular Transformer makes predictions by inferring the correlations between the presence and absence of chemical motifs in the reactant, reagent, and product present in the data set. Our model requires no handcrafted rules and accurately predicts subtle chemical transformations. Crucially, our model can accurately estimate its own uncertainty, with an uncertainty score that is 89% accurate in terms of classifying whether a prediction is correct. Furthermore, we show that the model is able to handle inputs without a reactant–reagent split and including stereochemistry, which makes our method universally applicable.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that a multihead attention Molecular Transformer model outperforms all algorithms in the literature, achieving a top-1 accuracy above 90% on a common benchmark data set and is able to handle inputs without a reactant–reagent split and including stereochemistry, which makes the method universally applicable.'}",https://pubs.acs.org/doi/pdf/10.1021/acscentsci.9b00576
-Visual Saliency Transformer,Nian Liu,"Existing state-of-the-art saliency detection methods heavily rely on CNN-based architectures. Alternatively, we rethink this task from a convolution-free sequence-to-sequence perspective and predict saliency by modeling long-range dependencies, which can not be achieved by convolution. Specifically, we develop a novel unified model based on a pure transformer, namely, Visual Saliency Transformer (VST), for both RGB and RGB-D salient object detection (SOD). It takes image patches as inputs and leverages the transformer to propagate global contexts among image patches. Unlike conventional architectures used in Vision Transformer (ViT), we leverage multi-level token fusion and propose a new token upsampling method under the transformer framework to get high-resolution detection results. We also develop a token-based multi-task decoder to simultaneously perform saliency and boundary detection by introducing task-related tokens and a novel patch-task-attention mechanism. Experimental results show that our model outperforms existing methods on both RGB and RGB-D SOD benchmark datasets. Most importantly, our whole framework not only provides a new perspective for the SOD field but also shows a new paradigm for transformer-based dense prediction models. Code is available at https://github.com/nnizhang/VST.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a novel unified model based on a pure transformer, namely, Visual Saliency Transformer, for both RGB and RGB-D salient object detection (SOD), which takes image patches as inputs and leverages the transformer to propagate global contexts among image patches.'}",https://arxiv.org/pdf/2104.12099
-Convolutional Bypasses Are Better Vision Transformer Adapters,Shibo Jie,"The pretrain-then-finetune paradigm has been widely adopted in computer vision. But as the size of Vision Transformer (ViT) grows exponentially, the full finetuning becomes prohibitive in view of the heavier storage overhead. Motivated by parameter-efficient transfer learning (PETL) on language transformers, recent studies attempt to insert lightweight adaptation modules (e.g., adapter layers or prompt tokens) to pretrained ViT and only finetune these modules while the pretrained weights are frozen. However, these modules were originally proposed to finetune language models and did not take into account the prior knowledge specifically for visual tasks. In this paper, we propose to construct Convolutional Bypasses (Convpass) in ViT as adaptation modules, introducing only a small amount (less than 0.5% of model parameters) of trainable parameters to adapt the large ViT. Different from other PETL methods, Convpass benefits from the hard-coded inductive bias of convolutional layers and thus is more suitable for visual tasks, especially in the low-data regime. Experimental results on VTAB-1K benchmark and few-shot learning datasets show that Convpass outperforms current language-oriented adaptation modules, demonstrating the necessity to tailor vision-oriented adaptation modules for adapting vision models.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on VTAB-1K benchmark and few-shot learning datasets show that Convpass outperforms current language-oriented adaptation modules, demonstrating the necessity to tailor vision- oriented adaptation modules for adapting vision models.'}",http://arxiv.org/pdf/2207.07039
-BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-models,Elad Ben-Zaken,"We introduce BitFit, a sparse-finetuning method where only the bias-terms of the model (or a subset of them) are being modified. We show that with small-to-medium training data, applying BitFit on pre-trained BERT models is competitive with (and sometimes better than) fine-tuning the entire model. For larger data, the method is competitive with other sparse fine-tuning methods.Besides their practical utility, these findings are relevant for the question of understanding the commonly-used process of finetuning: they support the hypothesis that finetuning is mainly about exposing knowledge induced by language-modeling training, rather than learning new task-specific linguistic knowledge.","{'model': 'tldr@v2.0.0', 'text': 'BitFit is introduced, a sparse-finetuning method where only the bias-terms of the model (or a subset of them) are being modified, which shows that with small-to-medium training data, applying BitFit on pre-trained BERT models is competitive with (and sometimes better than) fine-tuning the entire model.'}",https://aclanthology.org/2022.acl-short.1.pdf
-Speech-Transformer: A No-Recurrence Sequence-to-Sequence Model for Speech Recognition,Linhao Dong,"Recurrent sequence-to-sequence models using encoder-decoder architecture have made great progress in speech recognition task. However, they suffer from the drawback of slow training speed because the internal recurrence limits the training parallelization. In this paper, we present the Speech-Transformer, a no-recurrence sequence-to-sequence model entirely relies on attention mechanisms to learn the positional dependencies, which can be trained faster with more efficiency. We also propose a 2D-Attention mechanism, which can jointly attend to the time and frequency axes of the 2-dimensional speech inputs, thus providing more expressive representations for the Speech-Transformer. Evaluated on the Wall Street Journal (WSJ) speech recognition dataset, our best model achieves competitive word error rate (WER) of 10.9%, while the whole training process only takes 1.2 days on 1 GPU, significantly faster than the published results of recurrent sequence-to-sequence models.","{'model': 'tldr@v2.0.0', 'text': 'The Speech-Transformer is presented, a no-recurrence sequence-to-sequence model entirely relies on attention mechanisms to learn the positional dependencies, which can be trained faster with more efficiency and a 2D-Attention mechanism which can jointly attend to the time and frequency axes of the 2-dimensional speech inputs, thus providing more expressive representations for the Speech- Transformer.'}",
-UniT: Multimodal Multitask Learning with a Unified Transformer,Ronghang Hu,"We propose UniT, a Unified Transformer model to simultaneously learn the most prominent tasks across different domains, ranging from object detection to natural language understanding and multimodal reasoning. Based on the transformer encoder-decoder architecture, our UniT model encodes each input modality with an encoder and makes predictions on each task with a shared decoder over the encoded input representations, followed by task-specific output heads. The entire model is jointly trained end-to-end with losses from each task. Compared to previous efforts on multi-task learning with transformers, we share the same model parameters across all tasks instead of separately fine-tuning task-specific models and handle a much higher variety of tasks across different domains. In our experiments, we learn 7 tasks jointly over 8 datasets, achieving strong performance on each task with significantly fewer parameters. Our code is available in MMF at https://mmf.sh.","{'model': 'tldr@v2.0.0', 'text': 'UniT, a Unified Transformer model to simultaneously learn the most prominent tasks across different domains, ranging from object detection to natural language understanding and multimodal reasoning, achieves strong performance on each task with significantly fewer parameters.'}",https://arxiv.org/pdf/2102.10772
-Mobile-Former: Bridging MobileNet and Transformer,Yinpeng Chen,"We present Mobile-Former, a parallel design of MobileNet and transformer with a two-way bridge in between. This structure leverages the advantages of MobileNet at local processing and transformer at global interaction. And the bridge enables bidirectional fusion of local and global features. Different from recent works on vision transformer, the transformer in Mobile-Former contains very few tokens (e.g. 6 or fewer tokens) that are randomly initialized to learn global priors, resulting in low computational cost. Combining with the proposed light-weight cross attention to model the bridge, Mobile-Former is not only computationally efficient, but also has more representation power. It outperforms MobileNetV3 at low FLOP regime from 25M to 500M FLOPs on ImageNet classification. For instance, Mobile-Former achieves 77.9% top-1 accuracy at 294M FLOPs, gaining 1.3% over MobileNetV3 but saving 17% of computations. When transferring to object detection, Mobile-Former outperforms MobileNetV3 by 8.6 AP in RetinaNet framework. Furthermore, we build an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR by 1.3 AP but saves 52% of computational cost and 36% of parameters. Code will be released at https://github.com/aaboys/mobileformer.","{'model': 'tldr@v2.0.0', 'text': 'The proposed light-weight cross attention to model the bridge enables bidirectional fusion of local and global features and builds an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR but saves 52% of computational cost and 36% of parameters.'}",https://arxiv.org/pdf/2108.05895
-Conformer: Convolution-augmented Transformer for Speech Recognition,Anmol Gulati,"Recently Transformer and Convolution neural network (CNN) based models have shown promising results in Automatic Speech Recognition (ASR), outperforming Recurrent neural networks (RNNs). Transformer models are good at capturing content-based global interactions, while CNNs exploit local features effectively. In this work, we achieve the best of both worlds by studying how to combine convolution neural networks and transformers to model both local and global dependencies of an audio sequence in a parameter-efficient way. To this regard, we propose the convolution-augmented transformer for speech recognition, named Conformer. Conformer significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies. On the widely used LibriSpeech benchmark, our model achieves WER of 2.1%/4.3% without using a language model and 1.9%/3.9% with an external language model on test/testother. We also observe competitive performance of 2.7%/6.3% with a small model of only 10M parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the convolution-augmented transformer for speech recognition, named Conformer, which significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies.'}",https://arxiv.org/pdf/2005.08100
-Towards physics-inspired data-driven weather forecasting: integrating data assimilation with a deep spatial-transformer-based U-NET in a case study with ERA5,A. Chattopadhyay,"Abstract. There is growing interest in data-driven weather prediction (DDWP), e.g., using convolutional neural networks such as U-NET that are trained on data from models or reanalysis. Here, we propose three components, inspired by physics, to integrate with commonly used DDWP models in order to improve their forecast accuracy. These components are (1) a deep spatial transformer added to the latent space of U-NET to capture rotation and scaling transformation in the latent space for spatiotemporal data, (2) a data-assimilation (DA) algorithm to ingest noisy observations and improve the initial conditions for next forecasts, and (3) a multi-time-step algorithm, which combines forecasts from DDWP models with different time steps through DA, improving the accuracy of forecasts at short intervals. To show the benefit and feasibility of each component, we use geopotential height at 500 hPa (Z500) from ERA5 reanalysis and examine the short-term forecast accuracy of specific setups of the DDWP framework. Results show that the spatial-transformer-based U-NET (U-STN) clearly outperforms the U-NET, e.g., improving the forecast skill by 45 %. Using a sigma-point ensemble Kalman (SPEnKF) algorithm for DA and U-STN as the forward model, we show that stable, accurate DA cycles are achieved even with high observation noise. This DDWP+DA framework substantially benefits from large (O(1000)) ensembles that are inexpensively generated with the data-driven forward model in each DA cycle. The multi-time-step DDWP+DA framework also shows promise; for example, it reduces the average error by factors of 2–3. These results show the benefits and feasibility of these three components, which are flexible and can be used in a variety of DDWP setups. Furthermore, while here we focus on weather forecasting, the three components can be readily adopted for other parts of the Earth system, such as ocean and land, for which there is a rapid growth of data and need for forecast and assimilation. ",,https://gmd.copernicus.org/articles/15/2221/2022/gmd-15-2221-2022.pdf
-Enhanced Transformer Model for Low- and Mid-Frequency Transients—Part I: Model Development,A. Rezaei-Zare,"Based on an enhanced topological representation of the transformer core, flux air paths, and tank, a duality-based transformer model is proposed for low and mid-frequency transients analysis. In the developed transformer model, the core and tank iron are represented based on a frequency-dependent hysteresis model. The winding capacitances and the frequency dependency of the winding resistance are also incorporated in the developed model. Furthermore, a detailed model is developed for the air gap and tank, and the corresponding average equivalent is presented which is more appropriate for extracting the parameters from the measurement. In addition, the nonuniform air gap and the fringing flux effects are taken into account, and an approach is presented to obtain the corresponding parameters. The developed models also include the equivalents of the tank magnetic shield and the structural parts. Part II of this paper provides validations and further simulation results.",,
-Reformer: The Efficient Transformer,Nikita Kitaev,"Large Transformer models routinely achieve state-of-the-art results on a number of tasks but training these models can be prohibitively costly, especially on long sequences. We introduce two techniques to improve the efficiency of Transformers. For one, we replace dot-product attention by one that uses locality-sensitive hashing, changing its complexity from O($L^2$) to O($L\log L$), where $L$ is the length of the sequence. Furthermore, we use reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of $N$ times, where $N$ is the number of layers. The resulting model, the Reformer, performs on par with Transformer models while being much more memory-efficient and much faster on long sequences.","{'model': 'tldr@v2.0.0', 'text': 'This work replaces dot-product attention by one that uses locality-sensitive hashing and uses reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of several times, making the model much more memory-efficient and much faster on long sequences.'}",
-Enhanced Transformer Model for Low- and Mid-Frequency Transients—Part II: Validation and Simulation Results,A. Rezaei-Zare,"The transformer model developed in Part I is validated based on the zero-sequence test data, a ferroresonance event, and the geomagnetically induced current (GIC) measurements. Various transformer core constructions, including single phase, three limb, five limb, conventional shell type, and seven-limb shell type are represented, based on the proposed transformer model. The simulation results show high accuracy of the proposed transformer model in representing the studied cases. The study reveals that the saturation of the transformers at a given GIC level is more severe than that predicted by the existing transformer models. Furthermore, the proposed transformer model clearly explains the reasons for the previously reported discrepancies between the GIC experimental and simulation results. The study also concludes that for an accurate transformer model, particularly for the three-limb core type, the detailed representation of the zero-sequence characteristic is an essential requirement. Such a characteristic can be obtained based on either the proposed approach of the paper or a zero-sequence dc excitation test in addition to the test data at the power frequency.",,
-A Transformer-CNN Hybrid Model for Cognitive Behavioral Therapy in Psychological Assessment and Intervention for Enhanced Diagnostic Accuracy and Treatment Efficiency,Dr. Veera Ankalu. Vuyyuru,"— The use of Cognitive Behavioral Therapy (CBT) as a method for psychological assessment and intervention has shown to be quite successful. However, by utilizing advancements in artificial intelligence and natural language processing techniques, the diagnostic precision and therapeutic efficacy of CBT can be significantly improved. For CBT in psychological evaluation and intervention, we suggest a unique Transformer-CNN hybrid model in this work. The hybrid model combines the strengths of the Transformer and Convolutional Neural Network (CNN) architectures. While the CNN model successfully extracts local and global features from the input sequences, the Transformer model accurately captures the contextual dependencies and semantic linkages in the text data. It intends to enhance the model's comprehension and interpretation of the complex linguistic patterns involved in psychological evaluation and intervention by merging these two algorithms. On a sizable collection of clinical text data, which includes patient narratives, treatment transcripts, and diagnostic reports, we undertake comprehensive experiments. The proposed Trans-CNN hybrid model outperformed all other methods with an impressive accuracy of 97%. In diagnosing psychiatric problems, the model shows improved diagnosis accuracy and offers more effective therapy advice. Our hybrid model's automatic real-time monitoring and feedback capabilities also make it possible for prompt intervention and customized care during therapy sessions. By giving doctors a formidable tool for precise evaluation and efficient intervention, the suggested approach has the potential to revolutionize the field of CBT and enhance patient outcomes for mental health. In order to improve the diagnostic precision and therapeutic efficacy of CBT in psychological evaluation and intervention, this work provides a transformational strategy that combines the advantages of the Transformer and CNN architectures.",,http://thesai.org/Downloads/Volume14No7/Paper_66-A_Transformer_CNN_Hybrid_Model_for_Cognitive_Behavioral_Therapy.pdf
-Improving the Transformer Translation Model with Document-Level Context,Jiacheng Zhang,"Although the Transformer translation model (Vaswani et al., 2017) has achieved state-of-the-art performance in a variety of translation tasks, how to use document-level context to deal with discourse phenomena problematic for Transformer still remains a challenge. In this work, we extend the Transformer model with a new context encoder to represent document-level context, which is then incorporated into the original encoder and decoder. As large-scale document-level parallel corpora are usually not available, we introduce a two-step training method to take full advantage of abundant sentence-level parallel corpora and limited document-level parallel corpora. Experiments on the NIST Chinese-English datasets and the IWSLT French-English datasets show that our approach improves over Transformer significantly.","{'model': 'tldr@v2.0.0', 'text': 'This work extends the Transformer model with a new context encoder to represent document-level context, which is then incorporated into the original encoder and decoder, and introduces a two-step training method to take full advantage of abundant sentence-level parallel corpora and limited document- level parallel Corpora.'}",https://www.aclweb.org/anthology/D18-1049.pdf
-You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection,Yuxin Fang,"Can Transformer perform 2D object- and region-level recognition from a pure sequence-to-sequence perspective with minimal knowledge about the 2D spatial structure? To answer this question, we present You Only Look at One Sequence (YOLOS), a series of object detection models based on the vanilla Vision Transformer with the fewest possible modifications, region priors, as well as inductive biases of the target task. We find that YOLOS pre-trained on the mid-sized ImageNet-1k dataset only can already achieve quite competitive performance on the challenging COCO object detection benchmark, e.g., YOLOS-Base directly adopted from BERT-Base architecture can obtain 42.0 box AP on COCO val. We also discuss the impacts as well as limitations of current pre-train schemes and model scaling strategies for Transformer in vision through YOLOS. Code and pre-trained models are available at https://github.com/hustvl/YOLOS.","{'model': 'tldr@v2.0.0', 'text': 'It is found that YOLOS pre-trained on the mid-sized ImageNet-1k dataset only can already achieve quite competitive performance on the challenging COCO object detection benchmark, and the impacts as well as limitations of current pre-train schemes and model scaling strategies for Transformer in vision through YOLos are discussed.'}",
-TDSTF: Transformer-based Diffusion probabilistic model for Sparse Time series Forecasting,Ping Chang,"Background and objective: In the intensive care unit (ICU), vital sign monitoring is critical, and an accurate predictive system is required. This study will create a novel model to forecast Heart Rate (HR), Systolic Blood Pressure (SBP), and Diastolic Blood Pressure (DBP) in ICU. These vital signs are crucial for prompt interventions for patients. We extracted $24,886$ ICU stays from the MIMIC-III database, which contains data from over $46$ thousand patients, to train and test the model. Methods: The model proposed in this study, areansformerin intensive careabilistic Model for Sparse Time Series Forecasting (TDSTF), uses a deep learning technique called the Transformer. The TDSTF model showed state-of-the-art performance in predicting vital signs in the ICU, outperforming other models' ability to predict distributions of vital signs and being more computationally efficient. The code is available at https://github.com/PingChang818/TDSTF. Results: The results of the study showed that TDSTF achieved a Normalized Average Continuous Ranked Probability Score (NACRPS) of $0.4438$ and a Mean Squared Error (MSE) of $0.4168$, an improvement of $18.9\%$ and $34.3\%$ over the best baseline model, respectively. Conclusion: In conclusion, TDSTF is an effective and efficient solution for forecasting vital signs in the ICU, and it shows a significant improvement compared to other models in the field.","{'model': 'tldr@v2.0.0', 'text': ""The TDSTF model showed state-of-the-art performance in predicting vital signs in the ICU, outperforming other models' ability to predict distributions of vital signs and being more computationally efficient.""}",
-Analyzing the Structure of Attention in a Transformer Language Model,Jesse Vig,"The Transformer is a fully attention-based alternative to recurrent networks that has achieved state-of-the-art results across a range of NLP tasks. In this paper, we analyze the structure of attention in a Transformer language model, the GPT-2 small pretrained model. We visualize attention for individual instances and analyze the interaction between attention and syntax over a large corpus. We find that attention targets different parts of speech at different layer depths within the model, and that attention aligns with dependency relations most strongly in the middle layers. We also find that the deepest layers of the model capture the most distant relationships. Finally, we extract exemplar sentences that reveal highly specific patterns targeted by particular attention heads.","{'model': 'tldr@v2.0.0', 'text': 'It is found that attention targets different parts of speech at different layer depths within the model, and that attention aligns with dependency relations most strongly in the middle layers, and the deepest layers of the model capture the most distant relationships.'}",https://www.aclweb.org/anthology/W19-4808.pdf
-GPT4MIA: Utilizing Geneative Pre-trained Transformer (GPT-3) as A Plug-and-Play Transductive Model for Medical Image Analysis,Yizhe Zhang,". In this paper, we propose a novel approach (called GPT4MIA) that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA). We provide theoretical analysis on why a large pre-trained language model such as GPT-3 can be used as a plug-and-play transductive inference model for MIA. At the methodological level, we develop several technical treatments to improve the efficiency and effectiveness of GPT4MIA, including better prompt structure design, sample selection, and prompt ordering of representative samples/features. We present two concrete use cases (with workflow) of GPT4MIA: (1) detecting prediction errors and (2) improving prediction accuracy, working in conjecture with well-established vision-based models for image classification (e.g., ResNet). Experiments validate that our proposed method is effective for these two tasks. We further discuss the opportunities and challenges in utilizing Transformer-based large language models for broader MIA applications.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA) and develops several technical treatments to improve theency andectiveness of GPT4MIA.'}",http://arxiv.org/pdf/2302.08722
-mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer,Linting Xue,"The recent “Text-to-Text Transfer Transformer” (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent “accidental translation” in the zero-shot setting, where a generative model chooses to (partially) translate its prediction into the wrong language. All of the code and model checkpoints used in this work are publicly available.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2021.naacl-main.41.pdf
-Pretraining Data Mixtures Enable Narrow Model Selection Capabilities in Transformer Models,Steve Yadlowsky,"Transformer models, notably large language models (LLMs), have the remarkable ability to perform in-context learning (ICL) -- to perform new tasks when prompted with unseen input-output examples without any explicit model training. In this work, we study how effectively transformers can bridge between their pretraining data mixture, comprised of multiple distinct task families, to identify and learn new tasks in-context which are both inside and outside the pretraining distribution. Building on previous work, we investigate this question in a controlled setting, where we study transformer models trained on sequences of $(x, f(x))$ pairs rather than natural language. Our empirical results show transformers demonstrate near-optimal unsupervised model selection capabilities, in their ability to first in-context identify different task families and in-context learn within them when the task families are well-represented in their pretraining data. However when presented with tasks or functions which are out-of-domain of their pretraining data, we demonstrate various failure modes of transformers and degradation of their generalization for even simple extrapolation tasks. Together our results highlight that the impressive ICL abilities of high-capacity sequence models may be more closely tied to the coverage of their pretraining data mixtures than inductive biases that create fundamental generalization capabilities.","{'model': 'tldr@v2.0.0', 'text': 'The empirical results show transformers demonstrate near-optimal unsupervised model selection capabilities, in their ability to first in- context identify different task families and in-context learn within them when the task families are well-represented in their pretraining data.'}",
-GPT4MIA: Utilizing Generative Pre-trained Transformer (GPT-3) as A Plug-and-Play Transductive Model for Medical Image Analysis,Yizhe Zhang,"In this paper, we propose a novel approach (called GPT4MIA) that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA). We provide theoretical analysis on why a large pre-trained language model such as GPT-3 can be used as a plug-and-play transductive inference model for MIA. At the methodological level, we develop several technical treatments to improve the efficiency and effectiveness of GPT4MIA, including better prompt structure design, sample selection, and prompt ordering of representative samples/features. We present two concrete use cases (with workflow) of GPT4MIA: (1) detecting prediction errors and (2) improving prediction accuracy, working in conjecture with well-established vision-based models for image classification (e.g., ResNet). Experiments validate that our proposed method is effective for these two tasks. We further discuss the opportunities and challenges in utilizing Transformer-based large language models for broader MIA applications.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA), and develops several technical treatments to improve the efficiency and effectiveness of GPT4MIA.'}",
-Domain Prompts: Towards memory and compute efficient domain adaptation of ASR systems,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains. Since domain-specific systems perform better than their generic counterparts on in-domain evaluation, the need for memory and compute-efficient domain adaptation is obvious. Particularly, adapting parameter-heavy transformer-based language models used for rescoring ASR hypothesis is challenging. In this work, we introduce domain-prompts , a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain. With just a handful of extra parameters per domain, we achieve 7-14% WER improvement over the baseline of using an unadapted LM. Despite being parameter-efficient, these improvements are comparable to those of fully-fine-tuned models with hundreds of millions of parameters. With ablations on prompt-sizes, dataset sizes, initializations and domains, we provide evidence for the ben-efits of using domain-prompts in ASR systems.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces domain-prompts, a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain, achieving 7-14% WER improvement over the baseline of using an unadapted LM.'}",
-MSA Transformer,Roshan Rao,"Unsupervised protein language models trained across millions of diverse sequences learn structure and function of proteins. Protein language models studied to date have been trained to perform inference from individual sequences. The longstanding approach in computational biology has been to make inferences from a family of evo lutionarily related sequences by fitting a model to each family independently. In this work we combine the two paradigms. We introduce a protein language model which takes as input a set of sequences in the form of a multiple sequence alignment. The model interleaves row and column attention across the input sequences and is trained with a variant of the masked language modeling objective across many protein families. The performance of the model surpasses current state-of-the-art unsupervised structure learning methods by a wide margin, with far greater parameter efficiency than prior state-of-the-art protein language models.","{'model': 'tldr@v2.0.0', 'text': 'A protein language model which takes as input a set of sequences in the form of a multiple sequence alignment and is trained with a variant of the masked language modeling objective across many protein families surpasses current state-of-the-art unsupervised structure learning methods by a wide margin.'}",https://www.biorxiv.org/content/biorxiv/early/2021/08/27/2021.02.12.430858.full.pdf
-Optimizing transformer-based machine translation model for single GPU training: a hyperparameter ablation study,Luv Verma,"In machine translation tasks, the relationship between model complexity and performance is often presumed to be linear, driving an increase in the number of parameters and consequent demands for computational resources like multiple GPUs. To explore this assumption, this study systematically investigates the effects of hyperparameters through ablation on a sequence-to-sequence machine translation pipeline, utilizing a single NVIDIA A100 GPU. Contrary to expectations, our experiments reveal that combinations with the most parameters were not necessarily the most effective. This unexpected insight prompted a careful reduction in parameter sizes, uncovering""sweet spots""that enable training sophisticated models on a single GPU without compromising translation quality. The findings demonstrate an intricate relationship between hyperparameter selection, model size, and computational resource needs. The insights from this study contribute to the ongoing efforts to make machine translation more accessible and cost-effective, emphasizing the importance of precise hyperparameter tuning over mere scaling.",,https://arxiv.org/pdf/2308.06017
-LongT5: Efficient Text-To-Text Transformer for Long Sequences,Mandy Guo,"Recent work has shown that either (1) increasing the input length or (2) increasing model size can improve the performance of Transformer-based neural models. In this paper, we present a new model, called LongT5, with which we explore the effects of scaling both the input length and model size at the same time. Specifically, we integrated attention ideas from long-input transformers (ETC), and adopted pre-training strategies from summarization pre-training (PEGASUS) into the scalable T5 architecture. The result is a new attention mechanism we call {\em Transient Global} (TGlobal), which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs. We are able to achieve state-of-the-art results on several summarization tasks and outperform the original T5 models on question answering tasks.","{'model': 'tldr@v2.0.0', 'text': ""A new model, called LongT5, is presented, with which the effects of scaling both the input length and model size at the same time are explored, which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs.""}",https://aclanthology.org/2022.findings-naacl.55.pdf
-Planning with Learned Entity Prompts for Abstractive Summarization,Shashi Narayan,"Abstract We introduce a simple but flexible mechanism to learn an intermediate plan to ground the generation of abstractive summaries. Specifically, we prepend (or prompt) target summaries with entity chains—ordered sequences of entities mentioned in the summary. Transformer-based sequence-to-sequence models are then trained to generate the entity chain and then continue generating the summary conditioned on the entity chain and the input. We experimented with both pretraining and finetuning with this content planning objective. When evaluated on CNN/DailyMail, XSum, SAMSum, and BillSum, we demonstrate empirically that the grounded generation with the planning objective improves entity specificity and planning in summaries for all datasets, and achieves state-of-the-art performance on XSum and SAMSum in terms of rouge. Moreover, we demonstrate empirically that planning with entity chains provides a mechanism to control hallucinations in abstractive summaries. By prompting the decoder with a modified content plan that drops hallucinated entities, we outperform state-of-the-art approaches for faithfulness when evaluated automatically and by humans.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated empirically that the grounded generation with the planning objective improves entity specificity and planning in summaries for all datasets, and achieves state-of-the-art performance on XSum and SAMSum in terms of rouge.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00438/1979348/tacl_a_00438.pdf
-Swin Deformable Attention U-Net Transformer (SDAUT) for Explainable Fast MRI,Jiahao Huang,". Fast MRI aims to reconstruct a high fidelity image from partially observed measurements. Exuberant development in fast MRI using deep learning has been witnessed recently. Meanwhile, novel deep learning paradigms, e.g., Transformer based models, are fast-growing in natural language processing and promptly developed for computer vision and medical image analysis due to their prominent performance. Nevertheless, due to the complexity of the Transformer, the application of fast MRI may not be straightforward. The main obstacle is the computational cost of the self-attention layer, which is the core part of the Transformer, can be expensive for high resolution MRI inputs. In this study, we propose a new Transformer architecture for solving fast MRI that coupled Shifted Windows Transformer with U-Net to reduce the network complexity. We incorporate deformable attention to construe the explainability of our reconstruction model. We empirically demonstrate that our method achieves consistently superior performance on the fast MRI task. Besides, compared to state-of-the-art Transformer models, our method has fewer network parameters while revealing explainability. The code is publicly available at https://github.com/ayanglab/SDAUT.","{'model': 'tldr@v2.0.0', 'text': 'A new Transformer architecture for solving fast MRI that coupled Shifted Windows Transformer with U-Net to reduce the network complexity and incorporate deformable attention to construe the explainability of the reconstruction model.'}",http://arxiv.org/pdf/2207.02390
-Visformer: The Vision-friendly Transformer,Zhengsu Chen,"The past year has witnessed the rapid development of applying the Transformer module to vision problems. While some researchers have demonstrated that Transformer-based models enjoy a favorable ability of fitting data, there are still growing number of evidences showing that these models suffer over-fitting especially when the training data is limited. This paper offers an empirical study by performing step-by-step operations to gradually transit a Transformer-based model to a convolution-based model. The results we obtain during the transition process deliver useful messages for improving visual recognition. Based on these observations, we propose a new architecture named Visformer, which is abbreviated from the ‘Vision-friendly Transformer’. With the same computational complexity, Visformer outperforms both the Transformer-based and convolution-based models in terms of ImageNet classification accuracy, and the advantage becomes more significant when the model complexity is lower or the training set is smaller. The code is available at https://github.com/danczs/Visformer.","{'model': 'tldr@v2.0.0', 'text': 'A new architecture named Visformer is proposed, which is abbreviated from the ‘Vision-friendly Transformer’, which outperforms both the Transformer-based and convolution-based models in terms of ImageNet classification accuracy, and the advantage becomes more significant when the model complexity is lower or the training set is smaller.'}",https://arxiv.org/pdf/2104.12533
-TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models,Minghao Li,"Text recognition is a long-standing research problem for document digitalization. Existing approaches are usually built based on CNN for image understanding and RNN for char-level text generation. In addition, another language model is usually needed to improve the overall accuracy as a post-processing step. In this paper, we propose an end-to-end text recognition approach with pre-trained image Transformer and text Transformer models, namely TrOCR, which leverages the Transformer architecture for both image understanding and wordpiece-level text generation. The TrOCR model is simple but effective, and can be pre-trained with large-scale synthetic data and fine-tuned with human-labeled datasets. Experiments show that the TrOCR model outperforms the current state-of-the-art models on the printed, handwritten and scene text recognition tasks. The TrOCR models and code are publicly available at https://aka.ms/trocr.","{'model': 'tldr@v2.0.0', 'text': 'The TrOCR model is simple but effective, and can be pre-trained with large-scale synthetic data and fine-tuned with human-labeled datasets, and outperforms the current state-of-the-art models on the printed, handwritten and scene text recognition tasks.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26538/26310
-Vision-Language Transformer and Query Generation for Referring Segmentation,Henghui Ding,"In this work, we address the challenging task of referring segmentation. The query expression in referring segmentation typically indicates the target object by describing its relationship with others. Therefore, to find the target one among all instances in the image, the model must have a holistic understanding of the whole image. To achieve this, we reformulate referring segmentation as a direct attention problem: finding the region in the image where the query language expression is most attended to. We introduce transformer and multi-head attention to build a network with an encoder-decoder attention mechanism architecture that ""queries"" the given image with the language expression. Furthermore, we propose a Query Generation Module, which produces multiple sets of queries with different attention weights that represent the diversified comprehensions of the language expression from different aspects. At the same time, to find the best way from these diversified comprehensions based on visual clues, we further propose a Query Balance Module to adaptively select the output features of these queries for a better mask generation. Without bells and whistles, our approach is light-weight and achieves new state-of-the-art performance consistently on three referring segmentation datasets, RefCOCO, RefCOCO+, and G-Ref. Our code is available at https://github.com/henghuiding/Vision-Language-Transformer.","{'model': 'tldr@v2.0.0', 'text': 'Transformer and multi-head attention are introduced and a Query Generation Module is proposed, which produces multiple sets of queries with different attention weights that represent the diversified comprehensions of the language expression from different aspects.'}",https://arxiv.org/pdf/2108.05565
-ASFormer: Transformer for Action Segmentation,Fangqiu Yi,"Algorithms for the action segmentation task typically use temporal models to predict what action is occurring at each frame for a minute-long daily activity. Recent studies have shown the potential of Transformer in modeling the relations among elements in sequential data. However, there are several major concerns when directly applying the Transformer to the action segmentation task, such as the lack of inductive biases with small training sets, the deficit in processing long input sequence, and the limitation of the decoder architecture to utilize temporal relations among multiple action segments to refine the initial predictions. To address these concerns, we design an efficient Transformer-based model for action segmentation task, named ASFormer, with three distinctive characteristics: (i) We explicitly bring in the local connectivity inductive priors because of the high locality of features. It constrains the hypothesis space within a reliable scope, and is beneficial for the action segmentation task to learn a proper target function with small training sets. (ii) We apply a pre-defined hierarchical representation pattern that efficiently handles long input sequences. (iii) We carefully design the decoder to refine the initial predictions from the encoder. Extensive experiments on three public datasets demonstrate that effectiveness of our methods. Code is available at \url{https://github.com/ChinaYi/ASFormer}.","{'model': 'tldr@v2.0.0', 'text': 'An efficient Transformer-based model for action segmentation task, named ASFormer, with three distinctive characteristics, which constrains the hypothesis space within a reliable scope, and is beneficial for the action segmentations task to learn a proper target function with small training sets.'}",
-Pre-Trained Image Processing Transformer,Hanting Chen,"As the computing power of modern hardware is increasing strongly, pre-trained deep learning models (e.g., BERT, GPT-3) learned on large-scale datasets have shown their effectiveness over conventional methods. The big progress is mainly contributed to the representation ability of transformer and its variant architectures. In this paper, we study the low-level computer vision task (e.g., denoising, super-resolution and deraining) and develop a new pre-trained model, namely, image processing transformer (IPT). To maximally excavate the capability of transformer, we present to utilize the well-known ImageNet benchmark for generating a large amount of corrupted image pairs. The IPT model is trained on these images with multi-heads and multi-tails. In addition, the contrastive learning is introduced for well adapting to different image processing tasks. The pre-trained model can therefore efficiently employed on desired task after fine-tuning. With only one pre-trained model, IPT outperforms the current state-of-the-art methods on various low-level benchmarks. Code is available at https://github.com/huawei-noah/Pretrained-IPT and https://gitee.com/mindspore/mindspore/tree/master/model_zoo/research/cv/IPT","{'model': 'tldr@v2.0.0', 'text': 'To maximally excavate the capability of transformer, the IPT model is presented to utilize the well-known ImageNet benchmark for generating a large amount of corrupted image pairs and the contrastive learning is introduced for well adapting to different image processing tasks.'}",https://arxiv.org/pdf/2012.00364
-Personalized Transformer for Explainable Recommendation,Lei Li,"Personalization of natural language generation plays a vital role in a large spectrum of tasks, such as explainable recommendation, review summarization and dialog systems. In these tasks, user and item IDs are important identifiers for personalization. Transformer, which is demonstrated with strong language modeling capability, however, is not personalized and fails to make use of the user and item IDs since the ID tokens are not even in the same semantic space as the words. To address this problem, we present a PErsonalized Transformer for Explainable Recommendation (PETER), on which we design a simple and effective learning objective that utilizes the IDs to predict the words in the target explanation, so as to endow the IDs with linguistic meanings and to achieve personalized Transformer. Besides generating explanations, PETER can also make recommendations, which makes it a unified model for the whole recommendation-explanation pipeline. Extensive experiments show that our small unpretrained model outperforms fine-tuned BERT on the generation task, in terms of both effectiveness and efficiency, which highlights the importance and the nice utility of our design.","{'model': 'tldr@v2.0.0', 'text': 'A PErsonalized Transformer for Explainable Recommendation (PETER), on which a simple and effective learning objective is designed that utilizes the IDs to predict the words in the target explanation, so as to endow the IDs with linguistic meanings and to achieve personalized Transformer.'}",https://aclanthology.org/2021.acl-long.383.pdf
-Swin-Unet: Unet-like Pure Transformer for Medical Image Segmentation,Hu Cao,,"{'model': 'tldr@v2.0.0', 'text': 'Under the direct down-sampling and up-sampled of the inputs and outputs by 4x, experiments demonstrate that the pure Transformer-based U-shaped Encoder-Decoder network outperforms those methods with full Convolution or the combination of transformer and convolution.'}",https://arxiv.org/pdf/2105.05537
-Transformer Feed-Forward Layers Build Predictions by Promoting Concepts in the Vocabulary Space,Mor Geva,"Transformer-based language models (LMs) are at the core of modern NLP, but their internal prediction construction process is opaque and largely not understood. In this work, we make a substantial step towards unveiling this underlying prediction process, by reverse-engineering the operation of the feed-forward network (FFN) layers, one of the building blocks of transformer models. We view the token representation as a changing distribution over the vocabulary, and the output from each FFN layer as an additive update to that distribution. Then, we analyze the FFN updates in the vocabulary space, showing that each update can be decomposed to sub-updates corresponding to single FFN parameter vectors, each promoting concepts that are often human-interpretable. We then leverage these findings for controlling LM predictions, where we reduce the toxicity of GPT2 by almost 50%, and for improving computation efficiency with a simple early exit rule, saving 20% of computation on average.","{'model': 'tldr@v2.0.0', 'text': 'This work reverse-engineering the operation of the feed-forward network layers, one of the building blocks of transformer models, shows that each update can be decomposed to sub-updates corresponding to single FFN parameter vectors, each promoting concepts that are often human-interpretable.'}",http://arxiv.org/pdf/2203.14680
-A Survey of Controllable Text Generation Using Transformer-based Pre-trained Language Models,Hanqing Zhang,"Controllable Text Generation (CTG) is an emerging area in the field of natural language generation (NLG). It is regarded as crucial for the development of advanced text generation technologies that better meet the specific constraints in practical applications. In recent years, methods using large-scale pre-trained language models (PLMs), in particular the widely used Transformer-based PLMs, have become a new paradigm of NLG, allowing generation of more diverse and fluent text. However, due to the limited level of interpretability of deep neural networks, the controllability of these methods needs to be guaranteed. To this end, controllable text generation using Transformer-based PLMs has become a rapidly growing yet challenging new research hotspot. A diverse range of approaches have emerged in the past 3 to 4 years, targeting different CTG tasks that require different types of controlled constraints. In this article, we present a systematic critical review on the common tasks, main approaches, and evaluation methods in this area. Finally, we discuss the challenges that the field is facing, and put forward various promising future directions. To the best of our knowledge, this is the first survey article to summarize the state-of-the-art CTG techniques from the perspective of Transformer-based PLMs. We hope it can help researchers and practitioners in the related fields to quickly track the academic and technological frontier, providing them with a landscape of the area and a roadmap for future research.","{'model': 'tldr@v2.0.0', 'text': 'This is the first survey article to summarize the state-of-the-art CTG techniques from the perspective of Transformer-based PLMs, and it is hoped it can help researchers and practitioners in the related fields to quickly track the academic and technological frontier.'}",https://arxiv.org/pdf/2201.05337
-MATE: Multi-view Attention for Table Transformer Efficiency,Julian Martin Eisenschlos,"This work presents a sparse-attention Transformer architecture for modeling documents that contain large tables. Tables are ubiquitous on the web, and are rich in information. However, more than 20% of relational tables on the web have 20 or more rows (Cafarella et al., 2008), and these large tables present a challenge for current Transformer models, which are typically limited to 512 tokens. Here we propose MATE, a novel Transformer architecture designed to model the structure of web tables. MATE uses sparse attention in a way that allows heads to efficiently attend to either rows or columns in a table. This architecture scales linearly with respect to speed and memory, and can handle documents containing more than 8000 tokens with current accelerators. MATE also has a more appropriate inductive bias for tabular data, and sets a new state-of-the-art for three table reasoning datasets. For HybridQA (Chen et al., 2020), a dataset that involves large documents containing tables, we improve the best prior result by 19 points.","{'model': 'tldr@v2.0.0', 'text': 'MATE is proposed, a novel Transformer architecture designed to model the structure of web tables that uses sparse attention in a way that allows heads to efficiently attend to either rows or columns in a table.'}",https://aclanthology.org/2021.emnlp-main.600.pdf
-PromptCAL: Contrastive Affinity Learning via Auxiliary Prompts for Generalized Novel Category Discovery,Shengxiang Zhang,"Although existing semi-supervised learning models achieve remarkable success in learning with unannotated in-distribution data, they mostly fail to learn on unlabeled data sampled from novel semantic classes due to their closed-set assumption. In this work, we target a pragmatic but under-explored Generalized Novel Category Discovery (GNCD) setting. The GNCD setting aims to categorize unlabeled training data coming from known and novel classes by leveraging the information of partially labeled known classes. We propose a two-stage Contrastive Affinity Learning method with auxiliary visual Prompts, dubbed PromptCAL, to address this challenging problem. Our approach discovers reliable pairwise sample affinities to learn better semantic clustering of both known and novel classes for the class token and visual prompts. First, we propose a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships. Besides, we propose contrastive affinity learning to calibrate semantic representations based on our iterative semi-supervised affinity graph generation method for semantically-enhanced supervision. Extensive experimental evaluation demonstrates that our PromptCAL method is more effective in discovering novel classes even with limited annotations and surpasses the current state-of-the-art on generic and fine-grained benchmarks (e.g., with nearly 11% gain on CUB-200, and 9% on ImageNet-100) on overall accuracy. Our code is available at https://github.com/sheng-eatamath/PromptCAL.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes contrastive affinity learning to calibrate semantic representations based on the authors' iterative semi-supervised affinity graph generation method for semantically-enhanced supervision and proposes a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships.""}",https://arxiv.org/pdf/2212.05590
-IOT based classification of transformer faults using emerging techniques of E-nose and ANFIS,M. Equbal,"E-Nose finds its use in a wide range of applications such as quality assessment in food processing to toxic gas identification in chemical industry either in the offline or online mode. Their usage can be extended to transformer condition monitoring in the online mode. Considering the importance of transformers in power system and the impact it could create if faults in them are unidentified or left unattended, their functioning should be monitored on a real time basis. This work, describes the realization of a prospective E-Nose for online transformer incipient fault identification. The resistive gas sensor array has been simulated in real time using variable resistances forming one arm of a Wheatstone bridges. Separate variable resistances have been calibrated using characteristics of different fault gas sensors. The sensor array of the E-Nose helps to identify the transformer fault gases resulting from an incipient fault condition at the nascent stage itself and prompts for the necessary corrective action well before a catastrophic situation arises. Furthermore, ANFIS model of the Duval’s Triangle (DT) method have been developed to facilitate the online classification of incipient faults. The ANFIS models of other popularly used incipient fault interpretation methods, reported in earlier works, have also been used for a comparative analysis on their diagnostic capabilities. The developed model has been tested using the fault cases of IEC-TC10 fault database and the results thus obtained have been found to be very promising.",,https://www.frontiersin.org/articles/10.3389/fenrg.2022.1020040/pdf
-Improving Classification of Tetanus Severity for Patients in Low-Middle Income Countries Wearing ECG Sensors by Using a CNN-Transformer Network,Ping Lu,"Tetanus is a life-threatening infectious disease, which is still common in low- and middle-income countries, including in Vietnam. This disease is characterized by muscle spasm and in severe cases is complicated by autonomic dysfunction. Ideally continuous vital sign monitoring using bedside monitors allows the prompt detection of the onset of autonomic nervous system dysfunction or avoiding rapid deterioration. Detection can be improved using heart rate variability analysis from ECG signals. Recently, characteristic ECG and heart rate variability features have been shown to be of value in classifying tetanus severity. However, conventional manual analysis of ECG is time-consuming. The traditional convolutional neural network (CNN) has limitations in extracting the global context information, due to its fixed-sized kernel filters. In this work, we propose a novel hybrid CNN-Transformer model to automatically classify tetanus severity using tetanus monitoring from low-cost wearable sensors. This model can capture the local features from the CNN and the global features from the Transformer. The time series imaging - spectrogram - is transformed from one-dimensional ECG signal and input to the proposed model. The CNN-Transformer model outperforms state-of-the-art methods in tetanus classification, achieves results with a F1 score of $\mathbf {0.82\pm 0.03}$, precision of $\mathbf {0.94\pm 0.03}$, recall of $\mathbf {0.73\pm 0.07}$, specificity of $\mathbf {0.97\pm 0.02}$, accuracy of $\mathbf {0.88\pm 0.01}$ and AUC of $\mathbf {0.85\pm 0.03}$. In addition, we found that Random Forest with enough manually selected features can be comparable with the proposed CNN-Transformer model.","{'model': 'tldr@v2.0.0', 'text': 'A novel hybrid CNN-Transformer model to automatically classify tetanus severity using tetanus monitoring from low-cost wearable sensors that outperforms state-of-the-art methods in tetanus classification and finds that Random Forest with enough manually selected features can be comparable with the proposed CNN- transformer model.'}",https://ora.ox.ac.uk/objects/uuid:7b2beb4a-8f22-4be4-8e59-9150bff61086/files/s70795889r
-Synthesizer: Rethinking Self-Attention for Transformer Models,Yi Tay,"The dot product self-attention is known to be central and indispensable to state-of-the-art Transformer models. But is it really required? This paper investigates the true importance and contribution of the dot product-based self-attention mechanism on the performance of Transformer models. Via extensive experiments, we find that (1) random alignment matrices surprisingly perform quite competitively and (2) learning attention weights from token-token (query-key) interactions is useful but not that important after all. To this end, we propose \textsc{Synthesizer}, a model that learns synthetic attention weights without token-token interactions. In our experiments, we first show that simple Synthesizers achieve highly competitive performance when compared against vanilla Transformer models across a range of tasks, including machine translation, language modeling, text generation and GLUE/SuperGLUE benchmarks. When composed with dot product attention, we find that Synthesizers consistently outperform Transformers. Moreover, we conduct additional comparisons of Synthesizers against Dynamic Convolutions, showing that simple Random Synthesizer is not only $60\%$ faster but also improves perplexity by a relative $3.5\%$. Finally, we show that simple factorized Synthesizers can outperform Linformers on encoding only tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes \\textsc{Synthesizer}, a model that learns synthetic attention weights without token-token interactions, and shows that simple factorized Synthesizers can outperform Linformers on encoding only tasks.'}",
-Transformer in Transformer,Kai Han,"Transformer is a new kind of neural architecture which encodes the input data as powerful features via the attention mechanism. Basically, the visual transformers first divide the input images into several local patches and then calculate both representations and their relationship. Since natural images are of high complexity with abundant detail and color information, the granularity of the patch dividing is not fine enough for excavating features of objects in different scales and locations. In this paper, we point out that the attention inside these local patches are also essential for building visual transformers with high performance and we explore a new architecture, namely, Transformer iN Transformer (TNT). Specifically, we regard the local patches (e.g., 16$\times$16) as""visual sentences""and present to further divide them into smaller patches (e.g., 4$\times$4) as""visual words"". The attention of each word will be calculated with other words in the given visual sentence with negligible computational costs. Features of both words and sentences will be aggregated to enhance the representation ability. Experiments on several benchmarks demonstrate the effectiveness of the proposed TNT architecture, e.g., we achieve an 81.5% top-1 accuracy on the ImageNet, which is about 1.7% higher than that of the state-of-the-art visual transformer with similar computational cost. The PyTorch code is available at https://github.com/huawei-noah/CV-Backbones, and the MindSpore code is available at https://gitee.com/mindspore/models/tree/master/research/cv/TNT.","{'model': 'tldr@v2.0.0', 'text': 'It is pointed out that the attention inside these local patches are also essential for building visual transformers with high performance and a new architecture, namely, Transformer iN Transformer (TNT), is explored.'}",
-Text style transfer between classical and modern chinese through prompt-based reinforcement learning,Minzhang Xu,,"{'model': 'tldr@v2.0.0', 'text': 'An unsupervised prompt-based reinforcement learning (PBRL) framework to transfer text between classical and modern Chinese styles via an entangled approach and Experimental evaluations show that the model outperforms state-of-art networks by a large margin.'}",
-Automated COVID-19 Classification Using Heap-Based Optimization with the Deep Transfer Learning Model,Bahjat Fakieh,"The outbreak of the COVID-19 pandemic necessitates prompt identification of affected persons to restrict the spread of the COVID-19 epidemic. Radiological imaging such as computed tomography (CT) and chest X-rays (CXR) is considered an effective way to diagnose COVID-19. However, it needs an expert's knowledge and consumes more time. At the same time, artificial intelligence (AI) and medical images are discovered to be helpful in effectively assessing and providing treatment for COVID-19 infected patients. In particular, deep learning (DL) models act as a vital part of a high-performance classification model for COVID-19 recognition on CXR images. This study develops a heap-based optimization with the deep transfer learning model for detection and classification (HBODTL-DC) of COVID-19. The proposed HBODTL-DC system majorly focuses on the identification of COVID-19 on CXR images. To do so, the presented HBODTL-DC model initially exploits the Gabor filtering (GF) technique to enhance the image quality. In addition, the HBO algorithm with a neural architecture search network (NasNet) large model is employed for the extraction of feature vectors. Finally, Elman Neural Network (ENN) model gets the feature vectors as input and categorizes the CXR images into distinct classes. The experimental validation of the HBODTL-DC model takes place on the benchmark CXR image dataset from the Kaggle repository, and the outcomes are checked in numerous dimensions. The experimental outcomes stated the supremacy of the HBODTL-DC model over recent approaches with a maximum accuracy of 0.9992.","{'model': 'tldr@v2.0.0', 'text': 'A heap-based optimization with the deep transfer learning model for detection and classification (HBODTL-DC) of COVID-19 on CXR images is developed and the experimental outcomes stated the supremacy of the HBOD TL-DC model over recent approaches with a maximum accuracy of 0.9992.'}",https://downloads.hindawi.com/journals/cin/2022/7508836.pdf
-Transfer Learning-Based State of Charge Estimation for Lithium-Ion Battery at Varying Ambient Temperatures,Yan Qin,"Accurate and reliable state of charge (SoC) estimation becomes increasingly important to provide a stable and efficient environment for Lithium-ion batteries (LiBs) powered devices. Most data-driven SoC models are built for a fixed ambient temperature, which neglect the high sensitivity of LiBs to temperature and may cause severe prediction errors. Nevertheless, a systematic evaluation of the impact of temperature on SoC estimation and ways for a prompt adjustment of the estimation model to new temperatures using limited data has been hardly discussed. To solve these challenges, a novel SoC estimation method is proposed by exploiting temporal dynamics of measurements and transferring consistent estimation ability among different temperatures. First, temporal dynamics, which is presented by correlations between the past fluctuation and the future motion, are extracted using canonical variate analysis. Next, two models, including a reference SoC estimation model and an estimation ability monitoring model, are developed with temporal dynamics. The monitoring model provides a path to quantitatively evaluate the influences of temperature on SoC estimation ability. After that, once the inability of the reference SoC estimation model is detected, consistent temporal dynamics between temperatures are selected for transfer learning. Finally, the efficacy of the proposed method is verified through a benchmark. Our proposed method not only reduces prediction errors at fixed temperatures (e.g., reduced by 24.35$\%$ at –20 $^{\circ }$C, 49.82$\%$ at 25 $^{\circ }$C) but also improves prediction accuracies at new temperatures.","{'model': 'tldr@v2.0.0', 'text': 'A novel SoC estimation method that reduces prediction errors at fixed temperatures and improves prediction accuracies at new temperatures is proposed by exploiting temporal dynamics of measurements and transferring consistent estimation ability among different temperatures.'}",https://arxiv.org/pdf/2101.03704
-Improved Transfer Learning Based Deep Learning Model For Breast Cancer Histopathological Image Classification,Mohd. Farhan Israk Soumik,"In recent years, the demand for prompt detection and classification of breast cancer is rising sharply as breast cancer has become leading cancer type among women throughout the world. Convolutional Neural Networks(CNNs) are widely being used for performing mentioned tasks.However, they need a large number of labeled images which may appear to be infeasible for some kinds of medical images data such as mammographic tumor images. To address this difficulty, Transfer Learning becomes convenient. In this paper, we proposed a deep learning model for classifying Benign and Malignant types of breast tumor that trains an InceptionV3 model which pulls out features from the histopathological images of various magnification. These features are then used for classification. Introduced system is validated on BreakHis dataset and gains average validation set accuracy of 99.50%, 98.90%, 98.96% and 98.51% for magnification factor 40X, 100X, 200X and 400X respectively which outperforms all studied baseline models. Different performance metrices such as precision, recall, F1 score, Specificity have additionally been used for performance estimation purposes.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposed a deep learning model for classifying Benign and Malignant types of breast tumor that trains an InceptionV3 model which pulls out features from the histopathological images of various magnification which outperforms all studied baseline models.'}",
-MatchPrompt: Prompt-based Open Relation Extraction with Semantic Consistency Guided Clustering,Jiaxin Wang,"Relation clustering is a general approach for open relation extraction (OpenRE). Current methods have two major problems. One is that their good performance relies on large amounts of labeled and pre-defined relational instances for pre-training, which are costly to acquire in reality. The other is that they only focus on learning a high-dimensional metric space to measure the similarity of novel relations and ignore the specific relational representations of clusters. In this work, we propose a new prompt-based framework named MatchPrompt, which can realize OpenRE with efficient knowledge transfer from only a few pre-defined relational instances as well as mine the specific meanings for cluster interpretability. To our best knowledge, we are the first to introduce a prompt-based framework for unlabeled clustering. Experimental results on different datasets show that MatchPrompt achieves the new SOTA results for OpenRE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new prompt-based framework named MatchPrompt, which can realize OpenRE with efficient knowledge transfer from only a few pre-defined relational instances as well as mine the specific meanings for cluster interpretability.'}",https://aclanthology.org/2022.emnlp-main.537.pdf
-Dynamic Visual Prompt Tuning for Parameter Efficient Transfer Learning,Chunqing Ruan,"Parameter efficient transfer learning (PETL) is an emerging research spot that aims to adapt large-scale pre-trained models to downstream tasks. Recent advances have achieved great success in saving storage and computation costs. However, these methods do not take into account instance-specific visual clues for visual tasks. In this paper, we propose a Dynamic Visual Prompt Tuning framework (DVPT), which can generate a dynamic instance-wise token for each image. In this way, it can capture the unique visual feature of each image, which can be more suitable for downstream visual tasks. We designed a Meta-Net module that can generate learnable prompts based on each image, thereby capturing dynamic instance-wise visual features. Extensive experiments on a wide range of downstream recognition tasks show that DVPT achieves superior performance than other PETL methods. More importantly, DVPT even outperforms full fine-tuning on 17 out of 19 downstream tasks while maintaining high parameter efficiency. Our code will be released soon.","{'model': 'tldr@v2.0.0', 'text': 'A Dynamic Visual Prompt Tuning framework (DVPT), which can generate a dynamic instance-wise token for each image that can capture the unique visual feature of each image, which can be more suitable for downstream visual tasks.'}",https://arxiv.org/pdf/2309.06123
-An Efficient Brain Tumour Detection from MR Images Based on Deep Learning and Transfer Learning Model,Sarita Simaiya,"Brain disease is the most severe, pervasive, and life-threatening illness globally. Globally, brain tumours destroy the lifestyles of numerous individuals annually due to the fast expansion of tumour tissue. To prevent the deaths of individuals worldwide, prompt evaluation and classification of brain diagnosis are therefore necessary. In past years, the three primary forms of brain tumours, gliomas, malignant tumours, and pituitary, have been detected and classified most frequently using deep learning techniques. However, the volume of the sample with annotations significantly impacts how well deep learning algorithms work. It becomes complicated to classify a massive volume of medical data. In this work, we offer a stable hybrid approach for brain tumour identification predicated using CNN (VGG-16) with transfer learning, which incorporates the classic confusion and uncertainty polling technique, a best-fit methodology pooled by best-fit methodology using VGG-16, as well as the transfer learning approach. This approach decreases tagging expenditures while preserving the resilience and consistency of the systems. This research utilises the online brain tumour Kaggle dataset. To compare the proposed dynamic transfer learning strategy with the current CNN technique, evaluation measures such as precision, f-measure, and accuracy are used. The proposed technique enhances the CNN strategy by 5%, achieving 96.77% accuracy, 98.7% recall, 96.35 precision, and 96.78% F-measure.",,
-Efficiently Aligned Cross-Lingual Transfer Learning for Conversational Tasks using Prompt-Tuning,Lifu Tu,"Cross-lingual transfer of language models trained on high-resource languages like English has been widely studied for many NLP tasks, but focus on conversational tasks has been rather limited. This is partly due to the high cost of obtaining non-English conversational data, which results in limited coverage. In this work, we introduce XSGD for cross-lingual alignment pretraining, a parallel and large-scale multilingual conversation dataset that we created by translating the English-only Schema-Guided Dialogue (SGD) dataset (Rastogi et al., 2020) into 105 other languages. XSGD contains approximately 330k utterances per language. To facilitate aligned cross-lingual representations, we develop an efficient prompt-tuning-based method for learning alignment prompts. We also investigate two different classifiers: NLI-based and vanilla classifiers, and test cross-lingual capability enabled by the aligned prompts. We evaluate our model's cross-lingual generalization capabilities on two conversation tasks: slot-filling and intent classification. Our results demonstrate the strong and efficient modeling ability of NLI-based classifiers and the large cross-lingual transfer improvements achieved by our aligned prompts, particularly in few-shot settings. In addition, we highlight the nice results of our approach compared to LLMs such as text-davinci-003 and ChatGPT in both zero-shot and few-shot settings. While LLMs exhibit impressive performance in English, their cross-lingual capabilities in other languages, particularly low-resource languages, are limited.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces XSGD for cross-lingual alignment pretraining, a parallel and large-scale multilingual conversation dataset that was created by translating the English-only Schema-Guided Dialogue (SGD) dataset into 105 other languages and develops an efficient prompt-tuning-based method for learning alignment prompts.'}",http://arxiv.org/pdf/2304.01295
-MobileNet-SVM: A Lightweight Deep Transfer Learning Model to Diagnose BCH Scans for IoMT-Based Imaging Sensors,R. Ogundokun,"Many individuals worldwide pass away as a result of inadequate procedures for prompt illness identification and subsequent treatment. A valuable life can be saved or at least extended with the early identification of serious illnesses, such as various cancers and other life-threatening conditions. The development of the Internet of Medical Things (IoMT) has made it possible for healthcare technology to offer the general public efficient medical services and make a significant contribution to patients’ recoveries. By using IoMT to diagnose and examine BreakHis v1 400× breast cancer histology (BCH) scans, disorders may be quickly identified and appropriate treatment can be given to a patient. Imaging equipment having the capability of auto-analyzing acquired pictures can be used to achieve this. However, the majority of deep learning (DL)-based image classification approaches are of a large number of parameters and unsuitable for application in IoMT-centered imaging sensors. The goal of this study is to create a lightweight deep transfer learning (DTL) model suited for BCH scan examination and has a good level of accuracy. In this study, a lightweight DTL-based model “MobileNet-SVM”, which is the hybridization of MobileNet and Support Vector Machine (SVM), for auto-classifying BreakHis v1 400× BCH images is presented. When tested against a real dataset of BreakHis v1 400× BCH images, the suggested technique achieved a training accuracy of 100% on the training dataset. It also obtained an accuracy of 91% and an F1-score of 91.35 on the test dataset. Considering how complicated BCH scans are, the findings are encouraging. The MobileNet-SVM model is ideal for IoMT imaging equipment in addition to having a high degree of precision. According to the simulation findings, the suggested model requires a small computation speed and time.","{'model': 'tldr@v2.0.0', 'text': 'The goal of this study is to create a lightweight deep transfer learning (DTL) model suited for BCH scan examination and has a good level of accuracy, and the proposed MobileNet-SVM model is ideal for IoMT imaging equipment in addition to having a high degree of precision.'}",https://www.mdpi.com/1424-8220/23/2/656/pdf?version=1672995017
-ADPL: Adversarial Prompt-based Domain Adaptation for Dialogue Summarization with Knowledge Disentanglement,Lulu Zhao,"Traditional dialogue summarization models rely on a large-scale manually-labeled corpus, lacking generalization ability to new domains, and domain adaptation from a labeled source domain to an unlabeled target domain is important in practical summarization scenarios. However, existing domain adaptation works in dialogue summarization generally require large-scale pre-training using extensive external data. To explore the lightweight fine-tuning methods, in this paper, we propose an efficient Adversarial Disentangled Prompt Learning (ADPL) model for domain adaptation in dialogue summarization. We introduce three kinds of prompts including domain-invariant prompt (DIP), domain-specific prompt (DSP), and task-oriented prompt (TOP). DIP aims to disentangle and transfer the shared knowledge from the source domain and target domain in an adversarial way, which improves the accuracy of prediction about domain-invariant information and enhances the ability for generalization to new domains. DSP is designed to guide our model to focus on domain-specific knowledge using domain-related features. TOP is to capture task-oriented knowledge to generate high-quality summaries. Instead of fine-tuning the whole pre-trained language model (PLM), we only update the prompt networks but keep PLM fixed. Experimental results on the zero-shot setting show that the novel design of prompts can yield more coherent, faithful, and relevant summaries than baselines using the prefix-tuning, and perform at par with fine-tuning while being more efficient. Overall, our work introduces a prompt-based perspective to the zero-shot learning for dialogue summarization task and provides valuable findings and insights for future research.","{'model': 'tldr@v2.0.0', 'text': 'An efficient Adversarial Disentangled Prompt Learning (ADPL) model for domain adaptation in dialogue summarization and three kinds of prompts including domain-invariant prompt, domain-specific prompt, and task-oriented prompt are introduced.'}",
-Cataract detection from eye fundus image using an ensemble of transfer learning models,Kashish Chauhan,"Cataract is the clouding of the eye lens and is a major cause of blindness across the globe. Detecting cataracts early and prompt treatment can prevent blindness. To reduce dependence on experts to examine the eye fundus image, computer-assisted technologies are useful for early diagnosis. In the past, many cataract detection models have been described. Some of these models employed typical machine learning techniques, but the performance of these models is inefficient to make them useful. The image classification field has seen much improvement with the help of deep learning. One of the key drawbacks of these cutting-edge models is that they require a significant amount of computational resources and time for training. This problem is solved via transfer learning, which allows pre-trained models to be used for feature extraction. We present an ensemble technique for cataract diagnosis using the eye fundus image based on VGG-19, ResNet101V2, and InceptionV3. Soft voting was used to determine the final classification. The ensemble model gave the F-1 Score of 95.90 on the test dataset. According to our results, this ensemble network is more accurate than any single network.","{'model': 'tldr@v2.0.0', 'text': 'An ensemble technique for cataract diagnosis using the eye fundus image based on VGG-19, ResNet101V2, and InceptionV3 is presented, which is more accurate than any single network.'}",
-Parameter Efficient Transfer Learning for Suicide Attempt and Ideation Detection,Bhanu Pratap Singh Rawat,"Pre-trained language models (LMs) have been deployed as the state-of-the-art natural language processing (NLP) approaches for multiple clinical applications. Model generalisability is important in clinical domain due to the low available resources. In this study, we evaluated transfer learning techniques for an important clinical application: detecting suicide attempt (SA) and suicide ideation (SI) in electronic health records (EHRs). Using the annotation guideline provided by the authors of ScAN, we annotated two EHR datasets from different hospitals. We then fine-tuned ScANER, a publicly available SA and SI detection model, to evaluate five different parameter efficient transfer learning techniques, such as adapter-based learning and soft-prompt tuning, on the two datasets. Without any fine-tuning, ScANER achieve macro F1-scores of 0.85 and 0.87 for SA and SI evidence detection across the two datasets. We observed that by fine-tuning less than ~2% of ScANER’s parameters, we were able to further improve the macro F1-score for SA-SI evidence detection by 3% and 5% for the two EHR datasets. Our results show that parameter-efficient transfer learning methods can help improve the performance of publicly available clinical models on new hospital datasets with few annotations.","{'model': 'tldr@v2.0.0', 'text': 'Evaluated transfer learning techniques for detecting suicide attempt and suicide ideation in electronic health records (EHRs) show that parameter-efficient transfer learning methods can help improve the performance of publicly available clinical models on new hospital datasets with few annotations.'}",https://aclanthology.org/2022.louhi-1.13.pdf
-Efficient Diagnosis of Covid19 by Employing Deep Transfer Learning on Pretrained VGG and ResidualNet Architectures,Faraz Bagwan,"Covid19 has had a widespread influence on health services and the way of life. A prompt diagnosis is crucial for curbing the development of the disease and lowering the number of fatalities. It is customary and standard routine to employ blood tests to detect presence of pathogen, but because of the time and expense involved, it is often necessary to turn to other rapid and affordable options. We implemented two distinct transference based deep layered architectures in this study i.e., ResidualNet50 along with VGG16, to classify X-rays as COVID19, pneumonia, or normal. ResidualNet50 trained with transference approach outperformed the other deep-learning model i.e., VGG16, in the planned execution. Our proposed transfer deep-learning based model obtained an overall high classification accuracy of 98.5 percent. Result analysis and interpretation via performance curves have been comprehensively discussed in this paper.","{'model': 'tldr@v2.0.0', 'text': 'Two distinct transference based deep layered architectures are implemented to classify X-rays as COVID19, pneumonia, or normal, and the proposed transfer deep-learning based model obtained an overall high classification accuracy of 98.5 percent.'}",
-Distribution inference from early-stage stationary data streams by transfer learning,Kai Wang,"Abstract Data streams are prevalent in current manufacturing and service systems where real-time data arrive progressively. A quick distribution inference from such data streams at their early stages is extremely useful for prompt decision making in many industrial applications. For example, a quality monitoring scheme can be quickly started if the process data distribution is available and the optimal inventory level can be determined early once the customer demand distribution is estimated. To this end, this article proposes a novel online recursive distribution inference method for stationary data streams that can respond as soon as the streaming data are generated and update as regularly as the data accumulate. A major challenge is that the data size might be too small to produce an accurate estimation at the early stage of data streams. To solve this, we resort to an instance-based transfer learning approach which integrates a sufficient amount of auxiliary data from similar processes or products to aid the distribution inference in our target task. Particularly, the auxiliary data are reweighted automatically by a density ratio fitting model with a prior-belief-guided regularization term to alleviate data scarcity. Our proposed distribution inference method also possesses an efficient online algorithm with recursive formulas to update upon every incoming data point. Extensive numerical simulations and real case studies verify the advantages of the proposed method.","{'model': 'tldr@v2.0.0', 'text': ""An instance-based transfer learning approach which integrates a sufficient amount of auxiliary data from similar processes or products to aid the distribution inference in the authors' target task, and an efficient online algorithm with recursive formulas to update upon every incoming data point.""}",https://figshare.com/articles/journal_contribution/Distribution_Inference_from_Early-Stage_Stationary_Data_Streams_by_Transfer_Learning/13607782/2/files/26113477.pdf
-Clinical Concept and Relation Extraction Using Prompt-based Machine Reading Comprehension,C.A.I. Peng,"OBJECTIVE To develop a natural language processing system that solves both clinical concept extraction and relation extraction in a unified prompt-based machine reading comprehension (MRC) architecture with good generalizability for cross-institution applications. METHODS We formulate both clinical concept extraction and relation extraction using a unified prompt-based MRC architecture and explore state-of-the-art transformer models. We compare our MRC models with existing deep learning models for concept extraction and end-to-end relation extraction using 2 benchmark datasets developed by the 2018 National NLP Clinical Challenges (n2c2) challenge (medications and adverse drug events) and the 2022 n2c2 challenge (relations of social determinants of health [SDoH]). We also evaluate the transfer learning ability of the proposed MRC models in a cross-institution setting. We perform error analyses and examine how different prompting strategies affect the performance of MRC models. RESULTS AND CONCLUSION The proposed MRC models achieve state-of-the-art performance for clinical concept and relation extraction on the 2 benchmark datasets, outperforming previous non-MRC transformer models. GatorTron-MRC achieves the best strict and lenient F1-scores for concept extraction, outperforming previous deep learning models on the 2 datasets by 1%-3% and 0.7%-1.3%, respectively. For end-to-end relation extraction, GatorTron-MRC and BERT-MIMIC-MRC achieve the best F1-scores, outperforming previous deep learning models by 0.9%-2.4% and 10%-11%, respectively. For cross-institution evaluation, GatorTron-MRC outperforms traditional GatorTron by 6.4% and 16% for the 2 datasets, respectively. The proposed method is better at handling nested/overlapped concepts, extracting relations, and has good portability for cross-institute applications. Our clinical MRC package is publicly available at https://github.com/uf-hobi-informatics-lab/ClinicalTransformerMRC.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MRC models achieve state-of-the-art performance for clinical concept and relation extraction on the 2 benchmark datasets, outperforming previous non-MRC transformer models and has good portability for cross-institute applications.'}",https://arxiv.org/pdf/2303.08262
-Is Prompt-Based Finetuning Always Better than Vanilla Finetuning? Insights from Cross-Lingual Language Understanding,Bolei Ma,"Multilingual pretrained language models (MPLMs) have demonstrated substantial performance improvements in zero-shot cross-lingual transfer across various natural language understanding tasks by finetuning MPLMs on task-specific labelled data of a source language (e.g. English) and evaluating on a wide range of target languages. Recent studies show that prompt-based finetuning surpasses regular finetuning in few-shot scenarios. However, the exploration of prompt-based learning in multilingual tasks remains limited. In this study, we propose the ProFiT pipeline to investigate the cross-lingual capabilities of Prompt-based Finetuning. We conduct comprehensive experiments on diverse cross-lingual language understanding tasks (sentiment classification, paraphrase identification, and natural language inference) and empirically analyze the variation trends of prompt-based finetuning performance in cross-lingual transfer across different few-shot and full-data settings. Our results reveal the effectiveness and versatility of prompt-based finetuning in cross-lingual language understanding. Our findings indicate that prompt-based finetuning outperforms vanilla finetuning in full-data scenarios and exhibits greater advantages in few-shot scenarios, with different performance patterns dependent on task types. Additionally, we analyze underlying factors such as language similarity and pretraining data size that impact the cross-lingual performance of prompt-based finetuning. Overall, our work provides valuable insights into the cross-lingual prowess of prompt-based finetuning.","{'model': 'tldr@v2.0.0', 'text': 'The ProFiT pipeline is proposed to investigate the cross-lingual capabilities of Prompt-based Finetuning and indicates that prompt-based finetuning outperforms vanillaFinetuning in full-data scenarios and exhibits greater advantages in few-shot scenarios, with different performance patterns dependent on task types.'}",https://arxiv.org/pdf/2307.07880
-Knowledge Transfer with Visual Prompt in multi-modal Dialogue Understanding and Generation,Minjun Zhu,"Visual Dialogue (VD) task has recently received increasing attention in AI research. Visual Dialog aims to generate multi-round, interactive responses based on the dialog history and image content. Existing textual dialogue models cannot fully understand visual information, resulting in a lack of scene features when communicating with humans continuously. Therefore, how to efficiently fuse multimodal data features remains to be a challenge. In this work, we propose a knowledge transfer method with visual prompt (VPTG) fusing multi-modal data, which is a flexible module that can utilize the text-only seq2seq model to handle visual dialogue tasks. The VPTG conducts text-image co-learning and multi-modal information fusion with visual prompts and visual knowledge distillation. Specifically, we construct visual prompts from visual representations and then induce sequence-to-sequence(seq2seq) models to fuse visual information and textual contexts by visual-text patterns. And we also realize visual knowledge transfer through distillation between two different models’ text representations, so that the seq2seq model can actively learn visual semantic representations. Extensive experiments on the multi-modal dialogue understanding and generation (MDUG) datasets show the proposed VPTG outperforms other single-modal methods, which demonstrate the effectiveness of visual prompt and visual knowledge transfer.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments on the multi-modal dialogue understanding and generation (MDUG) datasets show the proposed VPTG outperforms other single- modal methods, which demonstrate the effectiveness of visual prompt and visual knowledge transfer.'}",
-PLATO-Ad: A Unified Advertisement Text Generation Framework with Multi-Task Prompt Learning,Zeyang Lei,"Online advertisement text generation aims at generating attractive and persuasive text ads to appeal to users clicking ads or purchasing products. While pretraining-based models have achieved remarkable success in generating high-quality text ads, some challenges remain, such as ad generation in low-resource scenarios and training efficiency for multiple ad tasks. In this paper, we propose a novel unified text ad generation framework with multi-task prompt learning, called PLATO-Ad, to tackle these problems. Specifically, we design a threephase transfer learning mechanism to tackle the low-resource ad generation problem. Furthermore, we present a novel multi-task prompt learning mechanism to efficiently utilize a single lightweight model to solve multiple ad generation tasks without loss of performance compared to training a separate model for each task. Finally, we conduct offline and online evaluations. Experiment results show that PLATO-Ad significantly outperforms the state-of-the-art on both offline and online metrics. PLATO-Ad has been deployed in a leading advertising platform with 3.5% CTR improvement on search ad descriptions and 10.4% CTR improvement on feed ad titles.","{'model': 'tldr@v2.0.0', 'text': 'A novel unified text ad generation framework with multi-task prompt learning, called PLATO-Ad, to tackle the low-resource ad generation problem and significantly outperforms the state-of-the-art on both offline and online metrics.'}",https://aclanthology.org/2022.emnlp-industry.52.pdf
-Melanoma Malignancy Prognosis Using Deep Transfer Learning,R. Shobarani,"Melanoma is a type of Skin cancer that spreads rapidly and has a significant death risk if it is not detected early and treated. A prompt and accurate diagnosis can improve the patient’s chances of survival. The creation of a skin cancer diagnostic support system based on computer technologies is highly essential. This study suggests a unique deep transfer learning model for categorizing melanoma malignancy. The proposed system comprises of three main phases including image preprocessing, feature extraction and melanoma classification. The preprocessing phase employs image filters such as mean, median, gaussian and non-local means filter along with histogram equalization techniques to obtain the preprocessed images. Feature extraction and classification are performed using pre-trained Convolutional Neural Network architectures such as DenseNet121, Inception-Resnet-V2 and Xception. Using the ISIC 2020 dataset, the suggested deep learning model’s effectiveness is assessed. The experimental findings show that, in terms of precision and computational expense, the suggested deep transfer learning model performs better than cutting-edge deep learning algorithms.","{'model': 'tldr@v2.0.0', 'text': 'The experimental findings show that the suggested deep transfer learning model performs better than cutting-edge deep learning algorithms in terms of precision and computational expense.'}",
-Motif-Based Prompt Learning for Universal Cross-Domain Recommendation,Bowen Hao,"Cross-Domain Recommendation (CDR) stands as a pivotal technology addressing issues of data sparsity and cold start by transferring general knowledge from the source to the target domain. However, existing CDR models suffer limitations in adaptability across various scenarios due to their inherent complexity. To tackle this challenge, recent advancements introduce universal CDR models that leverage shared embeddings to capture general knowledge across domains and transfer it through""Multi-task Learning""or""Pre-train, Fine-tune""paradigms. However, these models often overlook the broader structural topology that spans domains and fail to align training objectives, potentially leading to negative transfer. To address these issues, we propose a motif-based prompt learning framework, MOP, which introduces motif-based shared embeddings to encapsulate generalized domain knowledge, catering to both intra-domain and inter-domain CDR tasks. Specifically, we devise three typical motifs: butterfly, triangle, and random walk, and encode them through a Motif-based Encoder to obtain motif-based shared embeddings. Moreover, we train MOP under the""Pre-training \&Prompt Tuning""paradigm. By unifying pre-training and recommendation tasks as a common motif-based similarity learning task and integrating adaptable prompt parameters to guide the model in downstream recommendation tasks, MOP excels in transferring domain knowledge effectively. Experimental results on four distinct CDR tasks demonstrate the effectiveness of MOP than the state-of-the-art models.","{'model': 'tldr@v2.0.0', 'text': 'A motif- based prompt learning framework, MOP, which introduces motif-based shared embeddings to encapsulate generalized domain knowledge, catering to both intra-domain and inter-domain CDR tasks, and excels in transferring domain knowledge effectively.'}",
-PRO-CS : An Instance-Based Prompt Composition Technique for Code-Switched Tasks,Srijan Bansal,"Code-switched (CS) data is ubiquitous in today’s globalized world, but the dearth of annotated datasets in code-switching poses a significant challenge for learning diverse tasks across different language pairs. Parameter-efficient prompt-tuning approaches conditioned on frozen language models have shown promise for transfer learning in limited-resource setups. In this paper, we propose a novel instance-based prompt composition technique, PRO-CS, for CS tasks that combine language and task knowledge. We compare our approach with prompt-tuning and fine-tuning for code-switched tasks on 10 datasets across 4 language pairs. Our model outperforms the prompt-tuning approach by significant margins across all datasets and outperforms or remains at par with fine-tuning by using just 0.18% of total parameters. We also achieve competitive results when compared with the fine-tuned model in the low-resource cross-lingual and cross-task setting, indicating the effectiveness of our approach to incorporate new code-switched tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel instance-based prompt composition technique, PRO-CS, for CS tasks that combine language and task knowledge and achieves competitive results when compared with the fine-tuned model in the low-resource cross-lingual and cross-task setting, indicating the effectiveness of the approach to incorporate new code-switched tasks.'}",https://aclanthology.org/2022.emnlp-main.698.pdf
-"Continually Detection, Rapidly React: Unseen Rumors Detection Based on Continual Prompt-Tuning",Yuhui Zuo,"Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly. However, existing rumor detection (RD) models often assume the same training and testing distributions and can not cope with the continuously changing social network environment. This paper proposed a Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting (CF) of upstream tasks during sequential task learning and enables bidirectional knowledge transfer between domain tasks. Specifically, we propose the following strategies: (a) Our design explicitly decouples shared and domain-specific knowledge, thus reducing the interference among different domains during optimization; (b) Several technologies aim to transfer knowledge of upstream tasks to deal with emergencies; (c) A task-conditioned prompt-wise hypernetwork (TPHNet) is used to consolidate past domains. In addition, CPT-RD avoids CF without the necessity of a rehearsal buffer. Finally, CPT-RD is evaluated on English and Chinese RD datasets and is effective and efficient compared to prior state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables bidirectional knowledge transfer between domain tasks and is effective and efficient compared to prior state-of-the-art methods.'}",
-"Continuous Detection, Rapidly React: Unseen Rumors Detection based on Continual Prompt-Tuning",Yuhui Zuo,"Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly. However, existing rumor detection (RD) models often assume the same training and testing distributions and cannot cope with the continuously changing social network environment. This paper proposes a Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables knowledge transfer between domain tasks. To avoid forgetting, we optimize and store task-special soft-prompt for each domain. Furthermore, we also propose several strategies to transfer knowledge of upstream tasks to deal with emergencies and a taskconditioned prompt-wise hypernetwork (TPHNet) to consolidate past domains, enabling bidirectional knowledge transfer. Finally, CPTRD is evaluated on English and Chinese RD datasets and is effective and efficient compared to state-of-the-art baselines, without data replay techniques and with only a few parameter tuning.1","{'model': 'tldr@v2.0.0', 'text': 'A Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables knowledge transfer between domain tasks, and optimize and store task-special soft-prompt for each domain.'}",http://arxiv.org/pdf/2203.11720
-Soft Prompt Guided Joint Learning for Cross-Domain Sentiment Analysis,Jingli Shi,"Aspect term extraction is a fundamental task in fine-grained sentiment analysis, which aims at detecting customer's opinion targets from reviews on product or service. The traditional supervised models can achieve promising results with annotated datasets, however, the performance dramatically decreases when they are applied to the task of cross-domain aspect term extraction. Existing cross-domain transfer learning methods either directly inject linguistic features into Language models, making it difficult to transfer linguistic knowledge to target domain, or rely on the fixed predefined prompts, which is time-consuming to construct the prompts over all potential aspect term spans. To resolve the limitations, we propose a soft prompt-based joint learning method for cross domain aspect term extraction in this paper. Specifically, by incorporating external linguistic features, the proposed method learn domain-invariant representations between source and target domains via multiple objectives, which bridges the gap between domains with varied distributions of aspect terms. Further, the proposed method interpolates a set of transferable soft prompts consisted of multiple learnable vectors that are beneficial to detect aspect terms in target domain. Extensive experiments are conducted on the benchmark datasets and the experimental results demonstrate the effectiveness of the proposed method for cross-domain aspect terms extraction.","{'model': 'tldr@v2.0.0', 'text': 'By incorporating external linguistic features, the proposed method learn domain-invariant representations between source and target domains via multiple objectives, which bridges the gap between domains with varied distributions of aspect terms.'}",http://arxiv.org/pdf/2303.00815
-Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training,Xiaoyang Wu,"The rapid advancement of deep learning models often attributes to their ability to leverage massive training data. In contrast, such privilege has not yet fully benefited 3D deep learning, mainly due to the limited availability of large-scale 3D datasets. Merging multiple available data sources and letting them collaboratively train a single model is a potential solution. However, due to the large domain gap between 3D point cloud datasets, such mixed supervision could adversely affect the model's performance and lead to degenerated performance (i.e., negative transfer) compared to single-dataset training. In view of this challenge, we introduce Point Prompt Training (PPT), a novel framework for multi-dataset synergistic learning in the context of 3D representation learning that supports multiple pre-training paradigms. Based on this framework, we propose Prompt-driven Normalization, which adapts the model to different datasets with domain-specific prompts and Language-guided Categorical Alignment that decently unifies the multiple-dataset label spaces by leveraging the relationship between label text. Extensive experiments verify that PPT can overcome the negative transfer associated with synergistic learning and produce generalizable representations. Notably, it achieves state-of-the-art performance on each dataset using a single weight-shared model with supervised multi-dataset training. Moreover, when served as a pre-training framework, it outperforms other pre-training approaches regarding representation quality and attains remarkable state-of-the-art performance across over ten diverse downstream tasks spanning both indoor and outdoor 3D scenarios.","{'model': 'tldr@v2.0.0', 'text': 'PPT is introduced, a novel framework for multi-dataset synergistic learning in the context of 3D representation learning that supports multiple pre-training paradigms and achieves state-of-the-art performance on each dataset using a single weight-shared model with supervised multi- dataset training.'}",https://arxiv.org/pdf/2308.09718
-Deeply Coupled Cross-Modal Prompt Learning,Xuejing Liu,"Recent advancements in multimodal foundation models (e.g., CLIP) have excelled in zero-shot generalization. Prompt tuning involved in the knowledge transfer from foundation models to downstream tasks has gained significant attention recently. Existing prompt-tuning methods in cross-modal learning, however, either solely focus on language branch, or learn vision-language interaction in a shallow mechanism. In this context, we propose a Deeply coupled Cross-modal Prompt learning (DCP) method based on CLIP. DCP flexibly accommodates the interplay between vision and language with a Cross-Modal Prompt Attention (CMPA) mechanism, which enables the mutual exchange of respective representation through a well-connected multi-head attention module progressively and strongly. We then conduct comprehensive few-shot learning experiments on 11 image classification datasets and analyze the robustness to domain shift as well. Thorough experimental analysis evidently demonstrates the superb few-shot generalization and compelling domain adaption capacity of a well-executed DCP. The code can be found at https://github.com/GingL/CMPA.","{'model': 'tldr@v2.0.0', 'text': 'A Deeply coupled Cross-modal Prompt learning (DCP) method based on CLIP that flexibly accommodates the interplay between vision and language with a Cross-Modal Prompt Attention (CMPA) mechanism and demonstrates the superb few-shot generalization and compelling domain adaption capacity of a well-executed DCP.'}",http://arxiv.org/pdf/2305.17903
-Adaptive Prompt Learning with Distilled Connective Knowledge for Implicit Discourse Relation Recognition,Bang Wang,"Implicit discourse relation recognition (IDRR) aims at recognizing the discourse relation between two text segments without an explicit connective. Recently, the prompt learning has just been applied to the IDRR task with great performance improvements over various neural network-based approaches. However, the discrete nature of the state-art-of-art prompting approach requires manual design of templates and answers, a big hurdle for its practical applications. In this paper, we propose a continuous version of prompt learning together with connective knowledge distillation, called AdaptPrompt, to reduce manual design efforts via continuous prompting while further improving performance via knowledge transfer. In particular, we design and train a few virtual tokens to form continuous templates and automatically select the most suitable one by gradient search in the embedding space. We also design an answer-relation mapping rule to generate a few virtual answers as the answer space. Furthermore, we notice the importance of annotated connectives in the training dataset and design a teacher-student architecture for knowledge transfer. Experiments on the up-to-date PDTB Corpus V3.0 validate our design objectives in terms of the better relation recognition performance over the state-of-the-art competitors.","{'model': 'tldr@v2.0.0', 'text': 'A continuous version of prompt learning together with connective knowledge distillation, called AdaptPrompt, is proposed to reduce manual design efforts via continuous prompting while further improving performance via knowledge transfer.'}",https://arxiv.org/pdf/2309.07561
-Prompt Learning With Knowledge Memorizing Prototypes For Generalized Few-Shot Intent Detection,Chaiyut Luoyiching,"Generalized Few-Shot Intent Detection (GFSID) is challenging and realistic because it needs to categorize both seen and novel intents simultaneously. Previous GFSID methods rely on the episodic learning paradigm, which makes it hard to extend to a generalized setup as they do not explicitly learn the classification of seen categories and the knowledge of seen intents. To address the dilemma, we propose to convert the GFSID task into the class incremental learning paradigm. Specifically, we propose a two-stage learning framework, which sequentially learns the knowledge of different intents in various periods via prompt learning. And then we exploit prototypes for categorizing both seen and novel intents. Furthermore, to achieve the transfer knowledge of intents in different stages, for different scenarios we design two knowledge preservation methods which close to realistic applications. Extensive experiments and detailed analyses on two widely used datasets show that our framework based on the class incremental learning paradigm achieves promising performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-stage learning framework, which sequentially learns the knowledge of different intents in various periods via prompt learning, and designs two knowledge preservation methods which close to realistic applications.'}",https://arxiv.org/pdf/2309.04971
-RGB-T Tracking via Multi-Modal Mutual Prompt Learning,Yang Luo,"Object tracking based on the fusion of visible and thermal im-ages, known as RGB-T tracking, has gained increasing atten-tion from researchers in recent years. How to achieve a more comprehensive fusion of information from the two modalities with fewer computational costs has been a problem that re-searchers have been exploring. Recently, with the rise of prompt learning in computer vision, we can better transfer knowledge from visual large models to downstream tasks. Considering the strong complementarity between visible and thermal modalities, we propose a tracking architecture based on mutual prompt learning between the two modalities. We also design a lightweight prompter that incorporates attention mechanisms in two dimensions to transfer information from one modality to the other with lower computational costs, embedding it into each layer of the backbone. Extensive ex-periments have demonstrated that our proposed tracking ar-chitecture is effective and efficient, achieving state-of-the-art performance while maintaining high running speeds.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a tracking architecture based on mutual prompt learning between visible and thermal modalities and designs a lightweight prompter that incorporates attention mechanisms in two dimensions to transfer information from one modality to the other with lower computational costs.'}",https://arxiv.org/pdf/2308.16386
-Visual-Attribute Prompt Learning for Progressive Mild Cognitive Impairment Prediction,Luoyao Kang,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed prompt learning model even outperforms the fully fine-tuning baseline on transferring the knowledge from AD to pMCI, and a novel global prompt token for the visual prompts to provide global guidance to the multi-modal representations is proposed.'}",
-Deep Learning Model Based on 3D Optical Coherence Tomography Images for the Automated Detection of Pathologic Myopia,S. Park,"Pathologic myopia causes vision impairment and blindness, and therefore, necessitates a prompt diagnosis. However, there is no standardized definition of pathologic myopia, and its interpretation by 3D optical coherence tomography images is subjective, requiring considerable time and money. Therefore, there is a need for a diagnostic tool that can automatically and quickly diagnose pathologic myopia in patients. This study aimed to develop an algorithm that uses 3D optical coherence tomography volumetric images (C-scan) to automatically diagnose patients with pathologic myopia. The study was conducted using 367 eyes of patients who underwent optical coherence tomography tests at the Ophthalmology Department of Incheon St. Mary’s Hospital and Seoul St. Mary’s Hospital from January 2012 to May 2020. To automatically diagnose pathologic myopia, a deep learning model was developed using 3D optical coherence tomography images. The model was developed using transfer learning based on four pre-trained convolutional neural networks (ResNet18, ResNext50, EfficientNetB0, EfficientNetB4). Grad-CAM was used to visualize features affecting the detection of pathologic myopia. The performance of each model was evaluated and compared based on accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC). The model based on EfficientNetB4 showed the best performance (95% accuracy, 93% sensitivity, 96% specificity, and 98% AUROC) in identifying pathologic myopia.","{'model': 'tldr@v2.0.0', 'text': 'This study aimed to develop an algorithm that uses 3D optical coherence tomography volumetric images (C-scan) to automatically diagnose patients with pathologic myopia and found the model based on EfficientNetB4 showed the best performance.'}",https://www.mdpi.com/2075-4418/12/3/742/pdf?version=1647601223
-Trust-Based Intelligent Routing Protocol with Q-Learning for Mission-Critical Wireless Sensor Networks,Dooho Keum,"Mission-critical wireless sensor networks require a trustworthy and punctual routing protocol to ensure the worst-case end-to-end delay and reliability when transmitting mission-critical data collected by various sensors to gateways. In particular, the trustworthiness of mission-critical data must be guaranteed for decision-making and secure communications. However, it is a challenging issue to meet the requirement of both reliability and QoS in sensor networking environments where cyber-attacks may frequently occur and a lot of mission-critical data is generated. This study proposes a trust-based routing protocol that learns the trust elements using Q-learning to detect various attacks and ensure network performance. The proposed mechanism ensures the prompt detection of cyber threats that may occur in a mission-critical wireless sensor network and guarantees the trustworthy transfer of mission-critical sensor data. This paper introduces a distributed transmission technology that prioritizes the trustworthiness of mission-critical data through Q-learning results considering trustworthiness, QoS, and energy factors. It is a technology suitable for mission-critical wireless sensor network operational environments and can reliably operate resource-constrained devices. We implemented and performed a comprehensive evaluation of our scheme using the OPNET simulator. In addition, we measured packet delivery rates, throughput, survivability, and delay considering the characteristics of mission-critical sensor networks. The simulation results show an enhanced performance when compared with other mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'A trust-based routing protocol that learns the trust elements using Q-learning to detect various attacks and ensure network performance and ensures the prompt detection of cyber threats that may occur in a mission-critical wireless sensor network and guarantees the trustworthy transfer of mission- critical sensor data is proposed.'}",https://www.mdpi.com/1424-8220/22/11/3975/pdf?version=1653919940
-Accelerating Reinforcement Learning via Predictive Policy Transfer in 6G RAN Slicing,A. M. Nagib,"Reinforcement Learning (RL) algorithms have recently been proposed to solve dynamic radio resource management (RRM) problems in beyond 5G networks. However, RL-based solutions are still not widely adopted in commercial cellular networks. One of the primary reasons for this is the slow convergence of RL agents when they are deployed in a live network and when the network’s context changes significantly. Concurrently, the open radio access network (O-RAN) paradigm promises to give mobile network operators (MNOs) more control over their networks, furthering the need for intelligent and RL-based network management. O-RAN’s standardized interfaces will allow MNOs to make real-time custom changes to intelligently control various RRM functionalities. We consider a RAN slicing scenario in which MNOs can modify the weights of the RL reward function. This enables MNOs to change the priorities of fulfilling the service level agreements of the slices. However, this results in a practical challenge since the RL agent needs to adapt promptly to the changes made by the MNO. This challenge is addressed in this paper, where we first present and discuss the results from an exhaustive experiment to examine the efficiency of using transfer learning (TL) to accelerate the convergence of RL-based RAN slicing in the considered scenario. We then propose a novel predictive approach to enhance the TL-based acceleration by selecting the best-saved policy for reuse. By adopting the proposed policy transfer approach, RL agents are able to converge up to 14000 learning steps faster than their non-accelerated counterparts. The proposed machine learning (ML)-based predictive approach also shows up to a 96.5% accuracy in selecting the best expert policy to reuse for acceleration.","{'model': 'tldr@v2.0.0', 'text': 'A novel predictive approach is proposed to enhance the TL-based acceleration by selecting the best-saved policy for reuse by adopting the proposed policy transfer approach, which is able to converge up to 14000 learning steps faster than their non-accelerated counterparts.'}",
-FETCH: A deep-learning based classifier for fast transient classification,D. Agarwal,"With the upcoming commensal surveys for Fast Radio Bursts (FRBs), and their high candidate rate, usage of machine learning algorithms for candidate classification is a necessity. Such algorithms will also play a pivotal role in sending real-time triggers for prompt follow-ups with other instruments. In this paper, we have used the technique of Transfer Learning to train the state-of-the-art deep neural networks for classification of FRB and Radio Frequency Interference (RFI) candidates. These are convolutional neural networks which work on radio frequency-time and dispersion measure-time images as the inputs. We trained these networks using simulated FRBs and real RFI candidates from telescopes at the Green Bank Observatory. We present 11 deep learning models, each with an accuracy and recall above 99.5% on our test dataset comprising of real RFI and pulsar candidates. As we demonstrate, these algorithms are telescope and frequency agnostic and are able to detect all FRBs with signal-to-noise ratios above 10 in ASKAP and Parkes data. We also provide an open-source python package FETCH (Fast Extragalactic Transient Candidate Hunter) for classification of candidates, using our models. Using FETCH, these models can be deployed along with any commensal search pipeline for real-time candidate classification.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses the technique of Transfer Learning to train the state-of-the-art deep neural networks for classification of FRB and Radio Frequency Interference (RFI) candidates, which are convolutional neural networks which work on radio frequency-time and dispersion measure-time images as the inputs.'}",https://academic.oup.com/mnras/article-pdf/497/2/1661/33562918/staa1856.pdf
-Retrieval-based Knowledge Transfer: An Effective Approach for Extreme Large Language Model Compression,Jiduan Liu,"Large-scale pre-trained language models (LLMs) have demonstrated exceptional performance in various natural language processing (NLP) tasks. However, the massive size of these models poses huge challenges for their deployment in real-world applications. While numerous model compression techniques have been proposed, most of them are not well-suited for achieving extreme model compression when there is a significant gap in model scale. In this paper, we introduce a novel compression paradigm called Retrieval-based Knowledge Transfer (RetriKT), which effectively transfers the knowledge of LLMs to extremely small-scale models (e.g., 1%). In particular, our approach extracts knowledge from LLMs to construct a knowledge store, from which the small-scale model can retrieve relevant information and leverage it for effective inference. To improve the quality of the model, soft prompt tuning and Proximal Policy Optimization (PPO) reinforcement learning techniques are employed. Extensive experiments are conducted on low-resource tasks from SuperGLUE and GLUE benchmarks. The results demonstrate that the proposed approach significantly enhances the performance of small-scale models by leveraging the knowledge from LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A novel compression paradigm called Retrieval-based Knowledge Transfer (RetriKT), which effectively transfers the knowledge of LLMs to extremely small-scale models (e.g., 1%) and significantly enhances the performance of small- scale models by leveraging the knowledge from LLMs.'}",
-"Pre-train, Prompt and Recommendation: A Comprehensive Survey of Language Modelling Paradigm Adaptations in Recommender Systems",Peng Liu,"The emergence of Pre-trained Language Models (PLMs) has achieved tremendous success in the field of Natural Language Processing (NLP) by learning universal representations on large corpora in a self-supervised manner. The pre-trained models and the learned representations can be beneficial to a series of downstream NLP tasks. This training paradigm has recently been adapted to the recommendation domain and is considered a promising approach by both academia and industry. In this paper, we systematically investigate how to extract and transfer knowledge from pre-trained models learned by different PLM-related training paradigms to improve recommendation performance from various perspectives, such as generality, sparsity, efficiency and effectiveness. Specifically, we propose a comprehensive taxonomy to divide existing PLM-based recommender systems w.r.t. their training strategies and objectives. Then, we analyze and summarize the connection between PLM-based training paradigms and different input data types for recommender systems. Finally, we elaborate on open issues and future research directions in this vibrant field.","{'model': 'tldr@v2.0.0', 'text': 'This paper systematically investigates how to extract and transfer knowledge from pre-trained models learned by different PLM-related training paradigms to improve recommendation performance from various perspectives, such as generality, sparsity, efficiency and effectiveness.'}",https://arxiv.org/pdf/2302.03735
-Machine Learning based Potato Leaves Disease Detection,Raju Deepak Potnuru,"The primary source of food, money, and employment for rural residents in economically developing nations is agriculture. Crop loss caused by plant diseases, which reduces production by 20 to 30%, is the main factor affecting agriculture productivity. Conventional methods have been used to diagnose the diseases in an attempt to prevent such losses, but they are inaccurate. To avoid losses brought on by such illnesses, accurate and prompt detection of plant diseases is vital. But occasionally those harvests and grains suffer a significant amount of damage, if not complete destruction, due to a lack of suitable cultivating knowledge, expertise, and sense of disease prediction. So, in order to lessen the loss caused by infections of plant leaves, this research attempts to integrate a portion of agriculture with the use of artificial intelligence. We used CNN transfer learning models such as VGG16, VGG19, and InceptionV3, to overcome this issue. To determine which strategy performs best at identifying potato leaf illnesses, we conducted trials using all three approaches on the standard dataset of potato leaves.",,
-Can Unsupervised Knowledge Transfer from Social Discussions Help Argument Mining?,Subhabrata Dutta,"Identifying argument components from unstructured texts and predicting the relationships expressed among them are two primary steps of argument mining. The intrinsic complexity of these tasks demands powerful learning models. While pretrained Transformer-based Language Models (LM) have been shown to provide state-of-the-art results over different NLP tasks, the scarcity of manually annotated data and the highly domain-dependent nature of argumentation restrict the capabilities of such models. In this work, we propose a novel transfer learning strategy to overcome these challenges. We utilize argumentation-rich social discussions from the ChangeMyView subreddit as a source of unsupervised, argumentative discourse-aware knowledge by finetuning pretrained LMs on a selectively masked language modeling task. Furthermore, we introduce a novel prompt-based strategy for inter-component relation prediction that compliments our proposed finetuning method while leveraging on the discourse context. Exhaustive experiments show the generalization capability of our method on these two tasks over within-domain as well as out-of-domain datasets, outperforming several existing and employed strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes argumentation-rich social discussions from the ChangeMyView subreddit as a source of unsupervised, argumentative discourse-aware knowledge by finetuning pretrained LMs on a selectively masked language modeling task and introduces a novel prompt-based strategy for inter-component relation prediction that compliments the proposed finetuned method while leveraging on the discourse context.'}",http://arxiv.org/pdf/2203.12881
-Generative Adversarial Imitation Learning for Empathy-based AI,Pratyush Muthukumar,"Generative adversarial imitation learning (GAIL) is a model-free algorithm that has been shown to provide strong results in imitating complex behaviors in high-dimensional environments. In this paper, we utilize the GAIL model for text generation to develop empathy-based context-aware conversational AI. Our model uses an expert trajectory of empathetic prompt-response dialogues which can accurately exhibit the correct empathetic emotion when generating a response. The Generator of the GAIL model uses the GPT-2 sequential pre-trained language model trained on 117 million parameters from 40 GB of internet data. We propose a novel application of an approach used in transfer learning to fine tune the GPT-2 model in order to generate concise, user-specific empathetic responses validated against the Discriminator. Our novel GAIL model utilizes a sentiment analysis history-based reinforcement learning approach to empathetically respond to human interactions in a personalized manner. We find that our model's response scores on various human-generated prompts collected from the Facebook Empathetic Dialogues dataset outperform baseline counterparts. Moreover, our model improves upon various history-based conversational AI models developed recently, as our model's performance over a sustained conversation of 3 or more interactions outperform similar conversational AI models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel application of an approach used in transfer learning to fine tune the GPT-2 model in order to generate concise, user-specific empathetic responses validated against the Discriminator to develop empathy-based context-aware conversational AI.'}",
-Three-round learning strategy based on 3D deep convolutional GANs for Alzheimer’s disease staging,Wenjie Kang,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the proposed model avoids overfitting brought on by a paucity of sMRI data and enables the early detection of AD.'}",https://www.nature.com/articles/s41598-023-33055-9.pdf
-Deep Learning Based Voice Assistant for the Visually Impaired,Renuga. K,"Visual impairment is a critical disorder that increases the affected individual's dependence on others in many ways, such as knowing their surroundings. Gaining prompt information about the environment, without any human dependency is beneficial for a visually impaired person. This could create an incredible impact in the lives of the visually impaired as they can learn about their surroundings. The main aim of the project is to help the visually impaired in knowing their environment. The image of the environment is captured in real time and processed to generate descriptions about the image that are read out loud to the vision impaired people using text-to-speech technology. Creating depiction sentences with genuine and right construction plays a pivotal role. A CNN-RNN hybrid model is utilized with the encoder – decoder mechanism; in the CNN part of the model, inception model for transfer learning is used and RNN is majorly used for language modeling. Inception-V3 is a pretrained model which is used to extract features from each image. The image captured lively is then compared and analyzed with other images in the training dataset so that precise description of the image is got from the trained model. The Microsoft Common Objects in Context (MS-COCO) dataset is used, and the Bahdanau Attention mechanism is implemented to make predictions more accurately. Various evaluation standards show that the proposed model achieves increased performance and sensible scores are accomplished.","{'model': 'tldr@v2.0.0', 'text': 'The main aim of the project is to help the visually impaired in knowing their environment by generating descriptions about the environment that are read out loud to the vision impaired people using text-to-speech technology.'}",
-HetGPT: Harnessing the Power of Prompt Tuning in Pre-Trained Heterogeneous Graph Neural Networks,Yihong Ma,"Graphs have emerged as a natural choice to represent and analyze the intricate patterns and rich information of the Web, enabling applications such as online page classification and social recommendation. The prevailing""pre-train, fine-tune""paradigm has been widely adopted in graph machine learning tasks, particularly in scenarios with limited labeled nodes. However, this approach often exhibits a misalignment between the training objectives of pretext tasks and those of downstream tasks. This gap can result in the""negative transfer""problem, wherein the knowledge gained from pre-training adversely affects performance in the downstream tasks. The surge in prompt-based learning within Natural Language Processing (NLP) suggests the potential of adapting a""pre-train, prompt""paradigm to graphs as an alternative. However, existing graph prompting techniques are tailored to homogeneous graphs, neglecting the inherent heterogeneity of Web graphs. To bridge this gap, we propose HetGPT, a general post-training prompting framework to improve the predictive performance of pre-trained heterogeneous graph neural networks (HGNNs). The key is the design of a novel prompting function that integrates a virtual class prompt and a heterogeneous feature prompt, with the aim to reformulate downstream tasks to mirror pretext tasks. Moreover, HetGPT introduces a multi-view neighborhood aggregation mechanism, capturing the complex neighborhood structure in heterogeneous graphs. Extensive experiments on three benchmark datasets demonstrate HetGPT's capability to enhance the performance of state-of-the-art HGNNs on semi-supervised node classification.","{'model': 'tldr@v2.0.0', 'text': 'HetGPT, a general post-training prompting framework to improve the predictive performance of pre-trained heterogeneous graph neural networks (HGNNs), is proposed, with the design of a novel prompting function that integrates a virtual class prompt and a heterogeneous feature prompt with the aim to reformulate downstream tasks to mirror pretext tasks.'}",
-Approximated Prompt Tuning for Vision-Language Pre-trained Models,Qiong Wu,"Prompt tuning is a parameter-efficient way to deploy large-scale pre-trained models to downstream tasks by adding task-specific tokens. In terms of vision-language pre-trained (VLP) models, prompt tuning often requires a large number of learnable tokens to bridge the gap between the pre-training and downstream tasks, which greatly exacerbates the already high computational overhead. In this paper, we revisit the principle of prompt tuning for Transformer-based VLP models, and reveal that the impact of soft prompt tokens can be actually approximated via independent information diffusion steps, thereby avoiding the expensive global attention modeling and reducing the computational complexity to a large extent. Based on this finding, we propose a novel Approximated Prompt Tuning (APT) approach towards efficient VL transfer learning. To validate APT, we apply it to two representative VLP models, namely ViLT and METER, and conduct extensive experiments on a bunch of downstream tasks. Meanwhile, the generalization of APT is also validated on CLIP for image classification and StableDiffusion for text-to-image generation. The experimental results not only show the superior performance gains and computation efficiency of APT against the conventional prompt tuning methods, e.g., +7.01% accuracy and -82.30% additional computation overhead on METER, but also confirm its merits over other parameter-efficient transfer learning approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper revisits the principle of prompt tuning for Transformer-based VLP models, and reveals that the impact of soft prompt tokens can be actually approximated via independent information diffusion steps, thereby avoiding the expensive global attention modeling and reducing the computational complexity to a large extent.'}",https://arxiv.org/pdf/2306.15706
-Efficient COVID-19 Identification Using Deep Learning for IoT,V. Veeraiah,"Internet of Things (IoT) has made it possible to diagnose and treat patients remotely, as well as to expedite the transportation of essential drugs and medical equipment to locations that are geographically separated. This has occurred at a time when society has become more socially distant. During the Ebola and COVID-19 outbreaks, the Internet of Things (IoT) technology was put to use in remote patient monitoring and the management of the vaccine cold chain. Concurrently, this study reflects on the variables that are required for IoT to scale. Since December 2019, the COVID-19 outbreak on a worldwide scale has developed into a significant problem. In order for medical treatment to be successful, it is essential to make a prompt and accurate diagnosis of persons who may be infected with the COVID-19 virus. In order to put a halt to the spread of COVID-19, it is important to construct an automated system that is based on deep transfer learning and is capable of detecting the virus based on chest X-rays. The authors of this study present an internet-of-things (IoT) system that makes use of ensemble deep transfer learning to diagnose COVID-19 patients at an earlier stage. It is feasible to keep an eye on potentially hazardous COVID-19 incidents as they occur so long as suitable procedures are adhered to. Inceptions A variety of different deep learning models are included into the framework that has been proposed for the Internet of Things. According to the findings of the study, the method that was suggested assisted radiologists in accurately and quickly identifying patients who could have COVID-19. The proposed effort focuses on developing an effective identification system based on the COVID-19 standard for use in an IoT setting.","{'model': 'tldr@v2.0.0', 'text': 'The authors of this study present an internet-of-things (IoT) system that makes use of ensemble deep transfer learning to diagnose COVID-19 patients at an earlier stage and the method that was suggested assisted radiologists in accurately and quickly identifying patients who could have CO VID-19.'}",
-Transferable Environment Poisoning: Training-time Attack on Reinforcement Learning,Hang Xu,"Studying adversarial attacks on Reinforcement Learning (RL) agents has become a key aspect of developing robust, RL-based solutions. Test-time attacks, which target the post-learning performance of an RL agent’s policy, have been well studied in both whiteand black-box settings. More recently, however, state-of-the-art works have shifted to investigate training-time attacks on RL agents, i.e., forcing the learning process towards a target policy designed by the attacker. Alas, these SOTA works continue to rely on white-box settings and/or use a reward-poisoning approach. In contrast, this paper studies environment-dynamics poisoning attacks at training time. Furthermore, while environment-dynamics poisoning presumes a transfer-learning capable agent, it also allows us to expand our approach to black-box attacks. Our overall framework, inspired by hierarchical RL, seeks the minimal environment-dynamics manipulation that will prompt the momentary policy of the agent to change in a desired manner. We show the attack efficiency by comparing it with the reward-poisoning approach, and empirically demonstrate the transferability of the environment-poisoning attack strategy. Finally, we seek to exploit the transferability of the attack strategy to handle black-box settings.","{'model': 'tldr@v2.0.0', 'text': 'An overall framework, inspired by hierarchical RL, seeks the minimal environment-dynamics manipulation that will prompt the momentary policy of the agent to change in a desired manner and seeks to exploit the transferability of the attack strategy to handle black-box settings.'}",
-Deep Learning Based TTC Predictor for Power Systems with Wind Energy Integration,Gao Qiu,"Assessing the security margin of interconnected tie-lines is highly important to power systems. However, the increase of highly probabilistic wind power output causes fast variation of boundary limit parameters such as total transfer capability (TTC) of tie-lines. The traditional physical model-based TTC calculation is extremely time-consuming. To overcome this shortage, a fast data-driven TTC predictor based on deep belief network (DBN) is presented in this paper for accurate and prompt awareness of transfer limits. In the first stage, a clustering-based sample production method using annual load and wind power data is designed to simulate a large volume of operation scenario samples for DBN training. The well-trained leaner is then employed to enable TTC prediction for the vital transmission path. Finally, a benchmark system is used to validate the proposed methodology.","{'model': 'tldr@v2.0.0', 'text': 'A fast data-driven TTC predictor based on deep belief network (DBN) is presented in this paper for accurate and prompt awareness of transfer limits.'}",
-Explainable AI for Retinoblastoma Diagnosis: Interpreting Deep Learning Models with LIME and SHAP,C. Chiesa-Estomba,"Retinoblastoma is a rare and aggressive form of childhood eye cancer that requires prompt diagnosis and treatment to prevent vision loss and even death. Deep learning models have shown promising results in detecting retinoblastoma from fundus images, but their decision-making process is often considered a “black box” that lacks transparency and interpretability. In this project, we explore the use of LIME and SHAP, two popular explainable AI techniques, to generate local and global explanations for a deep learning model based on InceptionV3 architecture trained on retinoblastoma and non-retinoblastoma fundus images. We collected and labeled a dataset of 400 retinoblastoma and 400 non-retinoblastoma images, split it into training, validation, and test sets, and trained the model using transfer learning from the pre-trained InceptionV3 model. We then applied LIME and SHAP to generate explanations for the model’s predictions on the validation and test sets. Our results demonstrate that LIME and SHAP can effectively identify the regions and features in the input images that contribute the most to the model’s predictions, providing valuable insights into the decision-making process of the deep learning model. In addition, the use of InceptionV3 architecture with spatial attention mechanism achieved high accuracy of 97% on the test set, indicating the potential of combining deep learning and explainable AI for improving retinoblastoma diagnosis and treatment.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that LIME and SHAP can effectively identify the regions and features in the input images that contribute the most to the model’s predictions, providing valuable insights into the decision-making process of the deep learning model.'}",https://www.mdpi.com/2075-4418/13/11/1932/pdf?version=1685598711
-Breast Cancer Detection with Machine Learning-A Review,Karnam Meghana,"Malignancy in the breast is a significant public health concern, where timely identification is essential for effective treatment. Machine Learning (ML) and Deep Learning (DL) algorithms are potential tools for prompt detection og breast malignancy through examination of medical images such as mammograms. Convolutional neural networks (CNNs), transfer learning, and ensemble learning are some of the recent techniques being used in this field. Despite the advantages of ML and DL algorithms for breast cancer detection, there are still several challenges that need to be addressed. The lack of diversity in the datasets used to train algorithms is one major challenge, with many datasets based on specific populations that may not represent others. Highly annotated data is also limited in medical field. The objective of this study is to provide researchers with valuable insights and guidance.","{'model': 'tldr@v2.0.0', 'text': 'The objective of this study is to provide researchers with valuable insights and guidance in breast cancer detection through examination of medical images such as mammograms.'}",
-P3O: Transferring Visual Representations for Reinforcement Learning via Prompting,Guoliang You,"It is important for deep reinforcement learning (DRL) algorithms to transfer their learned policies to new environments that have different visual inputs. In this paper, we introduce Prompt based Proximal Policy Optimization (P3O), a three-stage DRL algorithm that transfers visual representations from a target to a source environment by applying prompting. The process of P3O consists of three stages: pre-training, prompting, and predicting. In particular, we specify a prompt-transformer for representation conversion and propose a two-step training process to train the prompt-transformer for the target environment, while the rest of the DRL pipeline remains unchanged. We implement P3O and evaluate it on the OpenAI CarRacing video game. The experimental results show that P3O outperforms the state-of-the-art visual transferring schemes. In particular, P3O allows the learned policies to perform well in environments with different visual inputs, which is much more effective than retraining the policies in these environments.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Prompt based Proximal Policy Optimization (P3O), a three-stage DRL algorithm that transfers visual representations from a target to a source environment by applying prompting, and shows that P3O outperforms the state-of-the-art visual transferring schemes.'}",https://arxiv.org/pdf/2303.12371
-Diabetic Retinopathy Classification Using Deep Learning,Abbaraju Sai Sathwik,"One of the main causes of adult blindness and a frequent consequence of diabetes is diabetic retinopathy (DR). To avoid visual loss, DR must be promptly identified and classified. In this article, we suggest an automated DR detection and classification method based on deep learning applied to fundus pictures. The suggested technique uses transfer learning for classification. On a dataset of 3,662 fundus images with real-world DR severity labels, we trained and validated our model. According to our findings, the suggested technique successfully detected and classified DR with an overall accuracy of 78.14%. Our model fared better than other recent cutting-edge techniques, illuminating the promise of deep learning-based strategies for DR detection and management. Our research indicates that the suggested technique may be employed as a screening tool for DR in a clinical environment, enabling early illness diagnosis and prompt treatment.","{'model': 'tldr@v2.0.0', 'text': 'An automated DR detection and classification method based on deep learning applied to fundus pictures that may be employed as a screening tool for DR in a clinical environment, enabling early illness diagnosis and prompt treatment.'}",https://publications.eai.eu/index.php/phat/article/download/4335/2670
-ICPC: Instance-Conditioned Prompting with Contrastive Learning for Semantic Segmentation,Chaohui Yu,"Modern supervised semantic segmentation methods are usually finetuned based on the supervised or self-supervised models pre-trained on ImageNet. Recent work shows that transferring the knowledge from CLIP to semantic segmentation via prompt learning can achieve promising performance. The performance boost comes from the feature enhancement with multimodal alignment, i.e., the dot product between vision and text embeddings. However, how to improve the multimodal alignment for better transfer performance in dense tasks remains underexplored. In this work, we focus on improving the quality of vision-text alignment from two aspects of prompting design and loss function, and present an instance-conditioned prompting with contrastive learning (ICPC) framework. First, compared with the static prompt designs, we reveal that dynamic prompting conditioned on image content can more efficiently utilize the text encoder for complex dense tasks. Second, we propose an align-guided contrastive loss to refine the alignment of vision and text embeddings. We further propose lightweight multi-scale alignment for better performance. Extensive experiments on three large-scale datasets (ADE20K, COCO-Stuff10k, and ADE20K-Full) demonstrate that ICPC brings consistent improvements across diverse backbones. Taking ResNet-50 as an example, ICPC outperforms the state-of-the-art counterpart by 1.71%, 1.05%, and 1.41% mIoU on the three datasets, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work reveals that dynamic prompting conditioned on image content can more efficiently utilize the text encoder for complex dense tasks and proposes an align-guided contrastive loss to refine the alignment of vision and text embeddings.'}",https://arxiv.org/pdf/2308.07078
-Contrastive Constraints Guide Explanation-Based Category Learning,Seth Chin-Parker,"This paper provides evidence for a contrastive account of explanation that is motivated by pragmatic theories that recognize the contribution that context makes to the interpretation of a prompt for explanation. This study replicates the primary findings of previous work in explanation-based category learning (Williams & Lombrozo, 2010), extending that work by illustrating the critical role of the context in this type of learning. Participants interacted with items from two categories either by describing the items or explaining their category membership. We manipulated the feature overlap between the categories and examined both the explanations generated and acquired knowledge of the categories. Explanations for membership in a given category were influenced by the unprompted contrast category, indicating an important role for contrastive processing in the generation of explanations. The influence of the contrast category was similarly seen in the transfer performance of the participants.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is provided for a contrastive account of explanation that is motivated by pragmatic theories that recognize the contribution that context makes to the interpretation of a prompt for explanation.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/cogs.12405
-The effects of goal-oriented instructions in digital game-based learning,Séverine Erhel,"Few studies have investigated the effects of the instructions provided in educational computer games on cognitive processing and learning outcomes. In our experiment, we sought to compare the effects on learning outcomes of two different types of goal-oriented instructions: mastery-goal instructions, which prompt learners to develop skills or master new knowledge, and performance-goal instructions, which are frequently used in game environments and which encourage individuals to demonstrate their ability to succeed, particularly by surpassing others. Results showed that a mastery-goal instruction elicited deeper learning (as assessed with a transfer task) than a performance-goal instruction. No effect of instruction was observed on either learning (demonstration consultation) times at the start of the game or on training task (solving riddles) performances during it. These results are discussed in terms of learning processes. This study demonstrates that mastery goal-oriented instructions can promote active processing of educational content in a serious game environment.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that mastery goal-oriented instructions can promote active processing of educational content in a serious game environment and elicited deeper learning than a performance-goal instruction.'}",
-Gradient-Based Automated Iterative Recovery for Parameter-Efficient Tuning,Maximilian Mozes,"Pretrained large language models (LLMs) are able to solve a wide variety of tasks through transfer learning. Various explainability methods have been developed to investigate their decision making process. TracIn (Pruthi et al., 2020) is one such gradient-based method which explains model inferences based on the influence of training examples. In this paper, we explore the use of TracIn to improve model performance in the parameter-efficient tuning (PET) setting. We develop conversational safety classifiers via the prompt-tuning PET method and show how the unique characteristics of the PET regime enable TracIn to identify the cause for certain misclassifications by LLMs. We develop a new methodology for using gradient-based explainability techniques to improve model performance, G-BAIR: gradient-based automated iterative recovery. We show that G-BAIR can recover LLM performance on benchmarks after manually corrupting training labels. This suggests that influence methods like TracIn can be used to automatically perform data cleaning, and introduces the potential for interactive debugging and relabeling for PET-based transfer learning methods.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that G-BAIR can recover LLM performance on benchmarks after manually corrupting training labels, suggesting that influence methods like TracIn can be used to automatically perform data cleaning, and introduces the potential for interactive debugging and relabeling for PET-based transfer learning methods.'}",http://arxiv.org/pdf/2302.06598
-Detection and Grade Classification of Diabetic Retinopathy and Adult Vitelliform Macular Dystrophy Based on Ophthalmoscopy Images,S. Srinivasan,"Diabetic retinopathy (DR) and adult vitelliform macular dystrophy (AVMD) may cause significant vision impairment or blindness. Prompt diagnosis is essential for patient health. Photographic ophthalmoscopy checks retinal health quickly, painlessly, and easily. It is a frequent eye test. Ophthalmoscopy images of these two illnesses are challenging to analyse since early indications are typically absent. We propose a deep learning strategy called ActiveLearn to address these concerns. This approach relies heavily on the ActiveLearn Transformer as its central structure. Furthermore, transfer learning strategies that are able to strengthen the low-level features of the model and data augmentation strategies to balance the data are incorporated owing to the peculiarities of medical pictures, such as their limited quantity and generally rigid structure. On the benchmark dataset, the suggested technique is shown to perform better than state-of-the-art methods in both binary and multiclass accuracy classification tasks with scores of 97.9% and 97.1%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a deep learning strategy called ActiveLearn that relies heavily on the ActiveLearn Transformer as its central structure and is shown to perform better than state-of-the-art methods in both binary and multiclass accuracy classification tasks.'}",https://www.mdpi.com/2079-9292/12/4/862/pdf?version=1675857334
-An aggregated loss function based lightweight few shot model for plant leaf disease classification,Shankey Garg,,"{'model': 'tldr@v2.0.0', 'text': 'An aggregated loss function formed by the combination of triplet loss and cross-entropy loss with MobileNetV2 as a base model for the effective classification of plant disease using small samples using only a few samples is used.'}",
-How teacher talk guidance during Invention activities shapes students’ cognitive engagement and transfer,Catherine C. Chase,,,https://stemeducationjournal.springeropen.com/track/pdf/10.1186/s40594-019-0170-7
-Cerebral Micro-Bleeding Detection Based on Densely Connected Neural Network,Shuihua Wang,"Cerebral micro-bleedings (CMBs) are small chronic brain hemorrhages that have many side effects. For example, CMBs can result in long-term disability, neurologic dysfunction, cognitive impairment and side effects from other medications and treatment. Therefore, it is important and essential to detect CMBs timely and in an early stage for prompt treatment. In this research, because of the limited labeled samples, it is hard to train a classifier to achieve high accuracy. Therefore, we proposed employing Densely connected neural network (DenseNet) as the basic algorithm for transfer learning to detect CMBs. To generate the subsamples for training and test, we used a sliding window to cover the whole original images from left to right and from top to bottom. Based on the central pixel of the subsamples, we could decide the target value. Considering the data imbalance, the cost matrix was also employed. Then, based on the new model, we tested the classification accuracy, and it achieved 97.71%, which provided better performance than the state of art methods.","{'model': 'tldr@v2.0.0', 'text': 'Densely connected neural network (DenseNet) is proposed as the basic algorithm for transfer learning to detectbral micro-bleedings and it provided better performance than the state of art methods.'}",https://www.frontiersin.org/articles/10.3389/fnins.2019.00422/pdf
-Project Based Learning Method using Simulation Tools and Hardware Exposure,P. V. V. R. Rao,"Power electronics have rapid growth in research and prompt emergent technology in electrical discipline. Study of Power Electronics (PE) for undergraduate and post graduate students is very much necessary as every industry is driven with power electronic converter based control of electric drives. This paper showcases the teaching of PE with effective teaching methodologies such as project based, lab acclimatize learning etc., Project Based Learning is recognized to be a motivating and problem centric teaching, that not only supports the students to acquire core knowledge but also helps the students to transfer their technical knowledge into realtime application. Problem will be given to students and asked to solve it to fulfil the industrial requirement and to obtain the course attainment. In this review paper Project Based Learning (PBL) is introduced and the feedback of the students before and after introducing the PBL is analysed.","{'model': 'tldr@v2.0.0', 'text': 'This paper showcases the teaching of PE with effective teaching methodologies such as project based, lab acclimatize learning etc., Project Based Learning is recognized to be a motivating and problem centric teaching, that not only supports theStudents to acquire core knowledge but also helps the students to transfer their technical knowledge into realtime application.'}",http://www.journaleet.org/index.php/jeet/article/download/85713/65807
-Learning from Clinical Judgments: Semi-Markov-Modulated Marked Hawkes Processes for Risk Prognosis,A. Alaa,"Critically ill patients in regular wards are vulnerable to unanticipated adverse events which require prompt transfer to the intensive care unit (ICU). To allow for accurate prognosis of deteriorating patients, we develop a novel continuous-time probabilistic model for a monitored patient's temporal sequence of physiological data. Our model captures ""informatively sampled"" patient episodes: the clinicians' decisions on when to observe a hospitalized patient's vital signs and lab tests over time are represented by a marked Hawkes process, with intensity parameters that are modulated by the patient's latent clinical states, and with observable physiological data (mark process) modeled as a switching multi-task Gaussian process. In addition, our model captures ""informatively censored"" patient episodes by representing the patient's latent clinical states as an absorbing semi-Markov jump process. The model parameters are learned from offline patient episodes in the electronic health records via an EM-based algorithm. Experiments conducted on a cohort of patients admitted to a major medical center over a 3-year period show that risk prognosis based on our model significantly outperforms the currently deployed medical risk scores and other baseline machine learning algorithms.","{'model': 'tldr@v2.0.0', 'text': ""A novel continuous-time probabilistic model for a monitored patient's temporal sequence of physiological data is developed that significantly outperforms the currently deployed medical risk scores and other baseline machine learning algorithms.""}",
-PrefixMol: Target- and Chemistry-aware Molecule Design via Prefix Embedding,Zhangyang Gao,"Is there a unified model for generating molecules considering different conditions, such as binding pockets and chemical properties? Although target-aware generative models have made significant advances in drug design, they do not consider chemistry conditions and cannot guarantee the desired chemical properties. Unfortunately, merging the target-aware and chemical-aware models into a unified model to meet customized requirements may lead to the problem of negative transfer. Inspired by the success of multi-task learning in the NLP area, we use prefix embeddings to provide a novel generative model that considers both the targeted pocket's circumstances and a variety of chemical properties. All conditional information is represented as learnable features, which the generative model subsequently employs as a contextual prompt. Experiments show that our model exhibits good controllability in both single and multi-conditional molecular generation. The controllability enables us to outperform previous structure-based drug design methods. More interestingly, we open up the attention mechanism and reveal coupling relationships between conditions, providing guidance for multi-conditional molecule generation.","{'model': 'tldr@v2.0.0', 'text': ""Inspired by the success of multi-task learning in the NLP area, prefix embeddings are used to provide a novel generative model that considers both the targeted pocket's circumstances and a variety of chemical properties, providing guidance for multi-conditional molecule generation.""}",http://arxiv.org/pdf/2302.07120
-Task Residual for Tuning Vision-Language Models,Tao Yu,"Large-scale vision-language models (VLMs) pre-trained on billion-level data have learned general visual representations and broad visual concepts. In principle, the welllearned knowledge structure of the VLMs should be inherited appropriately when being transferred to downstream tasks with limited data. However, most existing efficient transfer learning (ETL) approaches for VLMs either damage or are excessively biased towards the prior knowledge, e.g., prompt tuning (PT) discards the pre-trained text-based classifier and builds a new one while adapter-style tuning (AT) fully relies on the pre-trained features. To address this, we propose a new efficient tuning approach for VLMs named Task Residual Tuning (TaskRes), which performs directly on the text-based classifier and explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task. Specifically, TaskRes keeps the original classifier weights from the VLMs frozen and obtains a new classifier for the target task by tuning a set of prior-independent parameters as a residual to the original one, which enables reliable prior knowledge preservation and flexible task-specific knowledge exploration. The proposed TaskRes is simple yet effective, which significantly outperforms previous ETL methods (e.g., PT and AT) on 11 benchmark datasets while requiring minimal effort for the implementation. Our code is available at https://github.com/geekyutao/TaskRes.","{'model': 'tldr@v2.0.0', 'text': 'A new efficient tuning approach for VLMs named Task Residual Tuning (TaskRes), which performs directly on the text-based classifier and explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task.'}",https://arxiv.org/pdf/2211.10277
-Strategies for Enhancing the Multi-Stage Classification Performances of HER2 Breast Cancer from Hematoxylin and Eosin Images,Md Sakib Hossain Shovon,"Breast cancer is a significant health concern among women. Prompt diagnosis can diminish the mortality rate and direct patients to take steps for cancer treatment. Recently, deep learning has been employed to diagnose breast cancer in the context of digital pathology. To help in this area, a transfer learning-based model called ‘HE-HER2Net’ has been proposed to diagnose multiple stages of HER2 breast cancer (HER2-0, HER2-1+, HER2-2+, HER2-3+) on H&E (hematoxylin & eosin) images from the BCI dataset. HE-HER2Net is the modified version of the Xception model, which is additionally comprised of global average pooling, several batch normalization layers, dropout layers, and dense layers with a swish activation function. This proposed model exceeds all existing models in terms of accuracy (0.87), precision (0.88), recall (0.86), and AUC score (0.98) immensely. In addition, our proposed model has been explained through a class-discriminative localization technique using Grad-CAM to build trust and to make the model more transparent. Finally, nuclei segmentation has been performed through the StarDist method.","{'model': 'tldr@v2.0.0', 'text': 'A transfer learning-based model called ‘HE-HER2Net’ has been proposed to diagnose multiple stages of HER2 breast cancer on H&E (hematoxylin & eosin) images from the BCI dataset to build trust and to make the model more transparent.'}",https://www.mdpi.com/2075-4418/12/11/2825/pdf?version=1669019352
-Driver Inattention Monitoring System Based on the Orientation of the Face Using Convolutional Neural Network,M. ManjulaP.,"Driving inattentively is one of the prime reasons for vehicle accidents worldwide and has significant implications for road safety. A prompt alert to the inattentive driver can mitigate many accidents and save numerous lives, and reduce the cost of damages caused by accidents. To achieve this, a proposal of a nonintrusive and noninvasive driver inattention monitoring and alerting system in real time has been put forward. A mobile camera mounted on the windshield captures the video of the driver. Viola-Jones algorithm detects the face in each frame of the video and the Kanade–Lucas–Tomasi (KLT) algorithm tracks the detected face from one frame to another frame. The driver is classified as inattentive or attentive using Convolutional Neural Network (CNN). The transfer learning of the AlexNet Convolutional Neural Network architecture is adopted for the classification. The accuracy, precision, sensitivity, F1 score, and specificity of the system proposed in this paper are 98.24%, 100%, 96.47%, 98.21% and 100%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'A nonintrusive and noninvasive driver inattention monitoring and alerting system in real time that can mitigate many accidents and save numerous lives, and reduce the cost of damages caused by accidents.'}",
-Extracting Latent Steering Vectors from Pretrained Language Models,Nishant Subramani,"Prior work on controllable text generation has focused on learning how to control language models through trainable decoding, smart-prompt design, or fine-tuning based on a desired objective. We hypothesize that the information needed to steer the model to generate a target sentence is already encoded within the model. Accordingly, we explore a different approach altogether: extracting latent vectors directly from pretrained language model decoders without fine-tuning. Experiments show that there exist steering vectors, which, when added to the hidden states of the language model, generate a target sentence nearly perfectly (> 99 BLEU) for English sentences from a variety of domains. We show that vector arithmetic can be used for unsupervised sentiment transfer on the Yelp sentiment benchmark, with performance comparable to models tailored to this task. We find that distances between steering vectors reflect sentence similarity when evaluated on a textual similarity benchmark (STS-B), outperforming pooled hidden states of models. Finally, we present an analysis of the intrinsic properties of the steering vectors. Taken together, our results suggest that frozen LMs can be effectively controlled through their latent steering space.","{'model': 'tldr@v2.0.0', 'text': 'The results suggest that frozen LMs can be effectively controlled through their latent steering space, and it is found that distances between steering vectors reflect sentence similarity when evaluated on a textual similarity benchmark (STS-B), outperforming pooled hidden states of models.'}",http://arxiv.org/pdf/2205.05124
-Rethinking Efficient Tuning Methods from a Unified Perspective,Zeyinzi Jiang,"Parameter-efficient transfer learning (PETL) based on large-scale pre-trained foundation models has achieved great success in various downstream applications. Existing tuning methods, such as prompt, prefix, and adapter, perform task-specific lightweight adjustments to different parts of the original architecture. However, they take effect on only some parts of the pre-trained models, i.e., only the feed-forward layers or the self-attention layers, which leaves the remaining frozen structures unable to adapt to the data distributions of downstream tasks. Further, the existing structures are strongly coupled with the Transformers, hindering parameter-efficient deployment as well as the design flexibility for new approaches. In this paper, we revisit the design paradigm of PETL and derive a unified framework U-Tuning for parameter-efficient transfer learning, which is composed of an operation with frozen parameters and a unified tuner that adapts the operation for downstream applications. The U-Tuning framework can simultaneously encompass existing methods and derive new approaches for parameter-efficient transfer learning, which prove to achieve on-par or better performances on CIFAR-100 and FGVC datasets when compared with existing PETL methods.","{'model': 'tldr@v2.0.0', 'text': 'A unified framework U-Tuning for parameter-efficient transfer learning is derived, which is composed of an operation with frozen parameters and a unified tuner that adapts the operation for downstream applications, which prove to achieve on-par or better performances on CIFAR-100 and FGVC datasets when compared with existing PETL methods.'}",http://arxiv.org/pdf/2303.00690
-Prompting for Multi-Modal Tracking,Jinyu Yang,"Multi-modal tracking gains attention due to its ability to be more accurate and robust in complex scenarios compared to traditional RGB-based tracking. Its key lies in how to fuse multi-modal data and reduce the gap between modalities. However, multi-modal tracking still severely suffers from data deficiency, thus resulting in the insufficient learning of fusion modules. Instead of building such a fusion module, in this paper, we provide a new perspective on multi-modal tracking by attaching importance to the multi-modal visual prompts. We design a novel multi-modal prompt tracker (ProTrack), which can transfer the multi-modal inputs to a single modality by the prompt paradigm. By best employing the tracking ability of pre-trained RGB trackers learning at scale, our ProTrack can achieve high-performance multi-modal tracking by only altering the inputs, even without any extra training on multi-modal data. Extensive experiments on 5 benchmark datasets demonstrate the effectiveness of the proposed ProTrack.","{'model': 'tldr@v2.0.0', 'text': 'A novel multi-modal prompt tracker (ProTrack), which can transfer the multi- modal inputs to a single modality by the prompt paradigm, and can achieve high-performance multi- Modal tracking by only altering the inputs, even without any extra training on multi-Modal data.'}",https://arxiv.org/pdf/2207.14571
-Retrieval-Augmented Generative Question Answering for Event Argument Extraction,Xinya Du,"Event argument extraction has long been studied as a sequential prediction problem with extractive-based methods, tackling each argument in isolation. Although recent work proposes generation-based methods to capture cross-argument dependency, they require generating and post-processing a complicated target sequence (template). Motivated by these observations and recent pretrained language models’ capabilities of learning from demonstrations. We propose a retrieval-augmented generative QA model (R-GQA) for event argument extraction. It retrieves the most similar QA pair and augments it as prompt to the current example’s context, then decodes the arguments as answers. Our approach outperforms substantially prior methods across various settings (i.e. fully supervised, domain transfer, and fewshot learning). Finally, we propose a clustering-based sampling strategy (JointEnc) and conduct a thorough analysis of how different strategies influence the few-shot learning performances.","{'model': 'tldr@v2.0.0', 'text': 'A retrieval-augmented generative QA model (R-GQA) is proposed for event argument extraction that retrieves the most similar QA pair and augments it as prompt to the current example’s context, then decodes the arguments as answers.'}",https://arxiv.org/pdf/2211.07067
-Integrated Parameter-Efficient Tuning for General-Purpose Audio Models,Ju-ho Kim,"The advent of hyper-scale and general-purpose pre-trained models is shifting the paradigm of building task-specific models for target tasks. In the field of audio research, task-agnostic pre-trained models with high transferability and adaptability have achieved state-of-the-art performances through fine-tuning for downstream tasks. Nevertheless, re-training all the parameters of these massive models entails an enormous amount of time and cost, along with a huge carbon footprint. To overcome these limitations, the present study explores and applies efficient transfer learning methods in the audio domain. We also propose an integrated parameter-efficient tuning (IPET) framework by aggregating the embedding prompt (a prompt-based learning approach), and the adapter (an effective transfer learning method). We demonstrate the efficacy of the proposed framework using two backbone pre-trained audio models with different characteristics: the audio spectrogram transformer and wav2vec 2.0. The proposed IPET framework exhibits remarkable performance compared to fine-tuning method with fewer trainable parameters in four downstream tasks: sound event classification, music genre classification, keyword spotting, and speaker verification. Furthermore, the authors identify and analyze the shortcomings of the IPET framework, providing lessons and research directions for parameter efficient tuning in the audio domain.","{'model': 'tldr@v2.0.0', 'text': 'An integrated parameter-efficient tuning (IPET) framework is proposed by aggregating the embedding prompt (a prompt-based learning approach), and the adapter (an effective transfer learning method), demonstrating the efficacy of the proposed framework using two backbone pre-trained audio models with different characteristics.'}",http://arxiv.org/pdf/2211.02227
-Classification of Rail Track Crack using CNN with Pre-Trained VGG16 Model,Shreetha Bhat,"One of the vital components of railway infrastructure is rail tracks. Maintenance of rail track has been a major challenge in most of the countries and one such challenge is the detection of cracks on the rail surface. To maintain good health of the tracks requires regular inspection and prompt action, failure to which, may lead to accidents and loss of lives. The railway department is introducing many innovative methods to make the inspection process efficient. In the past, various methods have been explored to detect defects on rail surfaces such as Computer Vision-Based method, but full automation is far from achievement. Few of the advanced countries are making use of Deep Learning techniques to monitor and maintain the condition of rail tracks. In, this paper, amalgamation of Convolutional Neural Network (CNN) and transfer learning is applied for classifying defective (with cracks) and non-defective rail surfaces.","{'model': 'tldr@v2.0.0', 'text': 'Al amalgamation of Convolutional Neural Network (CNN) and transfer learning is applied for classifying defective (with cracks) and non-defective rail surfaces.'}",
-Virtual Node Tuning for Few-shot Node Classification,Zhen Tan,"Few-shot Node Classification (FSNC) is a challenge in graph representation learning where only a few labeled nodes per class are available for training. To tackle this issue, meta-learning has been proposed to transfer structural knowledge from base classes with abundant labels to target novel classes. However, existing solutions become ineffective or inapplicable when base classes have no or limited labeled nodes. To address this challenge, we propose an innovative method dubbed Virtual Node Tuning (VNT). Our approach utilizes a pretrained graph transformer as the encoder and injects virtual nodes as soft prompts in the embedding space, which can be optimized with few-shot labels in novel classes to modulate node embeddings for each specific FSNC task. A unique feature of VNT is that, by incorporating a Graph-based Pseudo Prompt Evolution (GPPE) module, VNT-GPPE can handle scenarios with sparse labels in base classes. Experimental results on four datasets demonstrate the superiority of the proposed approach in addressing FSNC with unlabeled or sparsely labeled base classes, outperforming existing state-of-the-art methods and even fully supervised baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an innovative method dubbed Virtual Node Tuning (VNT), which utilizes a pretrained graph transformer as the encoder and injects virtual nodes as soft prompts in the embedding space, which can be optimized with few-shot labels in novel classes to modulate node embeddings for each specific FSNC task.'}",
-Exploring Hard Samples in Multiview for Few-Shot Remote Sensing Scene Classification,Yuyu Jia,"Few-shot remote sensing scene classification (RSSC) is of high practical value in real situations where data are scarce and annotated costly. The few-shot learner needs to identify new categories with limited examples, and the core issue of this assignment is how to prompt the model to learn transferable knowledge from a large-scale base dataset. Although current approaches based on transfer learning or meta-learning have achieved significant performance on this task, there are still two problems to be addressed: 1) as an essential characteristic of remote sensing (RS) images, spatial rotation insensitivity surprisingly remains largely unexplored and 2) the high distribution uncertainty of hard samples reduces the discriminative power of the model decision boundary. Stimulated by these, we propose a corresponding end-to-end framework termed a hard sample learning and multiview integration network (HSL-MINet). First, the multiview integration (MI) module contains a pretext task introduced to guide the knowledge transfer and a multiview-attention mechanism used to extract correlational information across different rotation views of images. Second, aiming at increasing the discrimination of the model decision boundary, the hard sample learning (HSL) module is designed to evaluate and select hard samples via a classwise adaptive threshold strategy and then decrease the uncertainty of their feature distributions by a devised triplet loss. Extensive evaluations on NWPU-RESISC45, WHU-RS19, and UCM datasets show that the effectiveness of our HSL-MINet surpasses the former state-of-the-art approaches.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of the HSL-MINet surpasses the former state-of-the-art approaches on few-shot remote sensing scene classification datasets and is designed to evaluate and select hard samples via a classwise adaptive threshold strategy and decrease the uncertainty of their feature distributions by a devised triplet loss.'}",
-Alexander Knox at SemEval-2023 Task 5: The comparison of prompting and standard fine-tuning techniques for selecting the type of spoiler needed to neutralize a clickbait,M. Woźny,"Clickbait posts are a common problem on social media platforms, as they often deceive users by providing misleading or sensational headlines that do not match the content of the linked web page. The aim of this study is to create a technique for identifying the specific type of suitable spoiler - be it a phrase, a passage, or a multipart spoiler - needed to neutralize clickbait posts. This is achieved by developing a machine learning classifier analyzing both the clickbait post and the linked web page.Modern approaches for constructing a text classifier usually rely on fine-tuning a transformer-based model pre-trained on large unsupervised corpora. However, recent advances in the development of large-scale language models have led to the emergence of a new transfer learning paradigm based on prompt engineering.In this work, we study these two transfer learning techniques and compare their effectiveness for clickbait spoiler-type detection task.Our experimental results show that for this task, using the standard fine-tuning method gives better results than using prompting. The best model can achieve a similar performance to that presented by Hagen et al. (2022).","{'model': 'tldr@v2.0.0', 'text': 'The aim of this study is to create a technique for identifying the specific type of suitable spoiler - be it a phrase, a passage, or a multipart spoiler - needed to neutralize clickbait posts by developing a machine learning classifier analyzing both the clickb bait post and the linked web page.'}",https://aclanthology.org/2023.semeval-1.202.pdf
-PCBERT: Parent and Child BERT for Chinese Few-shot NER,Peichao Lai,"Achieving good performance on few-shot or zero-shot datasets has been a long-term challenge for NER. The conventional semantic transfer approaches on NER will decrease model performance when the semantic distribution is quite different, especially in Chinese few-shot NER. Recently, prompt-tuning has been thoroughly considered for low-resource tasks. But there is no effective prompt-tuning approach for Chinese few-shot NER. In this work, we propose a prompt-based Parent and Child BERT (PCBERT) for Chinese few-shot NER. To train an annotating model on high-resource datasets and then discover more implicit labels on low-resource datasets. We further design a label extension strategy to achieve label transferring from high-resource datasets. We evaluated our model on Weibo and the other three sampling Chinese NER datasets, and the experimental result demonstrates our approach’s effectiveness in few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based Parent and Child BERT (PCBERT) for Chinese few-shot NER, and designs a label extension strategy to achieve label transferring from high-resource datasets.'}",
-Wound Severity Classification using Deep Neural Network,D. Anisuzzaman,"The classification of wound severity is a critical step in wound diagnosis. An effective classifier can help wound professionals categorize wound conditions more quickly and affordably, allowing them to choose the best treatment option. This study used wound photos to construct a deep neural network-based wound severity classifier that classified them into one of three classes: green, yellow, or red. The green class denotes wounds still in the early stages of healing and are most likely to recover with adequate care. Wounds in the yellow category require more attention and treatment than those in the green category. Finally, the red class denotes the most severe wounds that require prompt attention and treatment. A dataset containing different types of wound images is designed with the help of wound specialists. Nine deep learning models are used with applying the concept of transfer learning. Several stacked models are also developed by concatenating these transfer learning models. The maximum accuracy achieved on multi-class classification is 68.49%. In addition, we achieved 78.79%, 81.40%, and 77.57% accuracies on green vs. yellow, green vs. red, and yellow vs. red classifications for binary classifications. With a learning rate of 0.001 and an Adam optimizer, all models were trained for 250 epochs. Two callbacks were employed with the best validation accuracy and the optimal combination of validation and training accuracy saving. The "" sparse_categorical_crossentropy "" and "" binary_crossentropy "" loss functions were used for multi-class and binary classification, respectively.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network-based wound severity classifier that classified wounds into one of three classes: green, yellow, or red, which denotes wounds still in the early stages of healing and are most likely to recover with adequate care.'}",http://arxiv.org/pdf/2204.07942
-UniHCP: A Unified Model for Human-Centric Perceptions,,"In this section, we provide details and full results for oneshot fine-tuning and prompt tuning on human parsing and pose estimation. For each experiment, we sample ten sets of images with different random seeds; we also grid search on both iterations and learning rates until performance converges. The reported results are based on the best config found for each setting. Data sampling. In one-shot transfer experiments, only one image per class is used for a task [9]. Table 1 shows the number of sampled images on one-shot transfer tasks. Note that in UniHCP, classification tasks are multi-label classification for human parsing, pose estimation, and attribute recognition, where each query performs binary classification via the global probability unit. Therefore, we also make sure the presence of cases where a class is absent is covered in our samples. Such handling avoids the query simply learning to output 1 when the corresponding class always presents within the sampled images. On the other hand, when a class does appear in most of the images, e.g., all keypoint joints in pose estimation or the background class in human parsing, we are able to achieve reasonably good results without such handling, thus we do not intentionally sample “not present” case for keypoint joints and background class in our experiments.","{'model': 'tldr@v2.0.0', 'text': 'This section provides details and full results for oneshot fine-tuning and prompt tuning on human parsing and pose estimation on one-shot transfer tasks in UniHCP.'}",
-Cloud-Device Collaborative Adaptation to Continual Changing Environments in the Real-World,Yulu Gan,"When facing changing environments in the real world, the lightweight model on client devices suffers from severe performance drops under distribution shifts. The main limitations of the existing device model lie in (1) unable to update due to the computation limit of the device, (2) the limited generalization ability of the lightweight model. Meanwhile, recent large models have shown strong generalization capability on the cloud while they can not be deployed on client devices due to poor computation constraints. To enable the device model to deal with changing environments, we propose a new learning paradigm of Cloud-Device Collaborative Continual Adaptation, which encourages collaboration between cloud and device and improves the generalization of the device model. Based on this paradigm, we further propose an Uncertainty-based Visual Prompt Adapted (U-VPA) teacher-student model to transfer the generalization capability of the large model on the cloud to the device model. Specifically, we first design the Uncertainty Guided Sampling (UGS) to screen out challenging data continuously and transmit the most out-of-distribution samples from the device to the cloud. Then we propose a Visual Prompt Learning Strategy with Uncertainty guided updating (VPLU) to specifically deal with the selected samples with more distribution shifts. We transmit the visual prompts to the device and concatenate them with the incoming data to pull the device testing distribution closer to the cloud training distribution. We conduct extensive experiments on two object detection datasets with continually changing environments. Our proposed U-VPA teacher-student framework outperforms previous state-of-the-art test time adaptation and device-cloud collaboration methods. The code and datasets will be released.","{'model': 'tldr@v2.0.0', 'text': 'A new learning paradigm of Cloud-Device Collaborative Continual Adaptation, which encourages collaboration between cloud and device and improves the generalization of the device model is proposed, which outperforms previous state-of-the-art test time adaptation and device-cloud collaboration methods.'}",https://arxiv.org/pdf/2212.00972
-Auto-Prompting SAM for Mobile Friendly 3D Medical Image Segmentation,Chengyin Li,"The Segment Anything Model (SAM) has rapidly been adopted for segmenting a wide range of natural images. However, recent studies have indicated that SAM exhibits subpar performance on 3D medical image segmentation tasks. In addition to the domain gaps between natural and medical images, disparities in the spatial arrangement between 2D and 3D images, the substantial computational burden imposed by powerful GPU servers, and the time-consuming manual prompt generation impede the extension of SAM to a broader spectrum of medical image segmentation applications. To address these challenges, in this work, we introduce a novel method, AutoSAM Adapter, designed specifically for 3D multi-organ CT-based segmentation. We employ parameter-efficient adaptation techniques in developing an automatic prompt learning paradigm to facilitate the transformation of the SAM model's capabilities to 3D medical image segmentation, eliminating the need for manually generated prompts. Furthermore, we effectively transfer the acquired knowledge of the AutoSAM Adapter to other lightweight models specifically tailored for 3D medical image analysis, achieving state-of-the-art (SOTA) performance on medical image segmentation tasks. Through extensive experimental evaluation, we demonstrate the AutoSAM Adapter as a critical foundation for effectively leveraging the emerging ability of foundation models in 2D natural image segmentation for 3D medical image segmentation.","{'model': 'tldr@v2.0.0', 'text': ""This work introduces a novel method, AutoSAM Adapter, designed specifically for 3D multi-organ CT-based segmentation, and employs parameter-efficient adaptation techniques in developing an automatic prompt learning paradigm to facilitate the transformation of the SAM model's capabilities to 3D medical image segmentation.""}",https://arxiv.org/pdf/2308.14936
-Transferring Pre-trained Multimodal Representations with Cross-modal Similarity Matching,Byoungjip Kim,"Despite surprising performance on zero-shot transfer, pre-training a large-scale multimodal model is often prohibitive as it requires a huge amount of data and computing resources. In this paper, we propose a method (BeamCLIP) that can effectively transfer the representations of a large pre-trained multimodal model (CLIP-ViT) into a small target model (e.g., ResNet-18). For unsupervised transfer, we introduce cross-modal similarity matching (CSM) that enables a student model to learn the representations of a teacher model by matching the relative similarity distribution across text prompt embeddings. To better encode the text prompts, we design context-based prompt augmentation (CPA) that can alleviate the lexical ambiguity of input text prompts. Our experiments show that unsupervised representation transfer of a pre-trained vision-language model enables a small ResNet-18 to achieve a better ImageNet-1K top-1 linear probe accuracy (66.2%) than vision-only self-supervised learning (SSL) methods (e.g., SimCLR: 51.8%, SwAV: 63.7%), while closing the gap with supervised learning (69.8%).","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a method (BeamCLIP) that can effectively transfer the representations of a large pre-trained multimodal model (CLIP-ViT) into a small target model (e.g., ResNet-18) and introduces cross-modal similarity matching (CSM) that enables a student model to learn the representation of a teacher model by matching the relative similarity distribution across text prompt embeddings.'}",http://arxiv.org/pdf/2301.02903
-Engaging beachgoers for drowning prevention: the spillover effects on non-participants,Peter Kamstra,"ABSTRACT Despite reduced drowning incidence at lifeguard patrolled beaches, 71 drowning fatalities occurred on Australian beaches last year (2021–2022). Prevailing drowning prevention practices on beaches include patrolling lifeguards positioning safety flags in less hazardous locations and encouraging beachgoers to swim between them. Such methods represent a ‘deficit based’ approach to community engagement, in which experts determine acceptable behaviours and encourage adherence using a one-way transfer of information. Deficit based approaches can be useful, but participatory forms of community engagement are hypothesised to support learning that can transfer to other locations and through non-participants’ social networks. Using a lifeguard patrolled beach in Gerroa, Australia as a case study, we employed a ‘relationship building’ methodology to explore whether engagements on the beach can prompt transformational learning and whether such learning spills over to non-participants or to unpatrolled locations. Findings from 49 survey-interview engagements and 15 follow-up interviews suggest that building relationships with researchers is an enjoyable form of community engagement that contributes to learning about risk; simultaneously, findings suggest that learning can transform beachgoers’ intentions and practices at unpatrolled beaches. This paper provides a broadened theoretical and empirical model of community engagement aimed at beach drowning risk prevention via relationship building. Key policy highlights Engaging beachgoers via relationship building facilitates learning about beach risk, resulting in spillover effects to non-participants and to (unpatrolled) contexts. Engaging communities through dialogue is more likely to have a lasting influence on behaviours compared with deficit-based forms of engagement. Spillover effects to children, family, and friends provide evidence of who participatory research can have a ‘successful’ impact on. Experienced beachgoers discussing the engagement with others demonstrates how relationship building creates opportunities for experienced participants to demonstrate care for others. This study provides a broadened theoretical and empirical model of engagement aimed at beach drowning risk prevention via relationship building.",,
-Student Perceptions of Staged Transfer to Independent Research Skills During a Four-year Honours Science Undergraduate Program,Sarah L. Symons,"We describe interim results of an ongoing longitudinal pedagogical study investigating the efficacy of the Honours Integrated Science Program (iSci). We describe the pedagogical methods we use to prompt research skill development in a model from instructor-dependence to independent original research. We also describe a tool we use to help students organise their group research during their first attempts. Finally, we discuss students’ perceptions of how well iSci develops their research skills. Our results show that students are attracted to the iSci program because of the opportunities for research-based learning and skills development. We also found that in-program students value research skill development as a tool for successful completion of their degree and for their future academic or career plans. We conclude that our study methods help identify areas where we can support our students by building their research confidence and, in particular, their time-management skills. Nous presentons une description des resultats interimaires d’une etude pedagogique longitudinale qui vise a evaluer l’efficacite du programme specialise integre de sciences (iSci). Nous faisons une description des methodes pedagogiques que nous utilisons pour declencher le developpement des competences en recherche au sein d’un modele qui va de la recherche qui depend de l’instructeur a la recherche independante originale. Nous decrivons egalement un outil que nous utilisons pour aider les etudiants a organiser leur recherche par groupe au cours de leurs premieres tentatives. Pour finir, nous discutons les perceptions des etudiants sur la maniere dont le programme iSci developpe leurs competences en recherche. Nos resultats indiquent que les etudiants sont attires vers le programme iSci a cause des occasions d’apprentissage base sur la recherche et de developpement des competences. Nous avons egalement remarque que les etudiants inscrits au programme apprecient le developpement de competences en recherche en tant qu’outil qui leur servira a terminer leurs etudes et a obtenir leur diplome, ou pour leurs projets de carriere, a l’avenir. En conclusion, nous declarons que nos methodes d’etudes aident a identifier les domaines ou nous pouvons soutenir les etudiants en renforcant leur confiance en matiere de recherche et, en particulier, leurs competences en gestion du temps.",,https://ojs.lib.uwo.ca/index.php/cjsotl_rcacea/article/download/6999/5713
-Analysis of the application of artificial intelligence technology in the protection of corporate governance rights and interests,Wen-liang Shen,"Corporate governance delivers feasible and controlled company operations using a group of common shareholders and appropriate policies. The roles and responsibilities of the shareholders suggest and improve corporate development through monotonous and independent rights. The implication of artificial intelligence provides knowledgeable insights for decision-making and control management. This article introduces a Mutual Consent-based Governance Regulation Model (MCGRM) for dissimilarity mitigation in corporate rule implications. The proposed model exploits transfer learning for balanced rule implication and decision-making. The learning states are defined based on mutual agreement, individual interest, and operational features. Based on the governance policies, the above rules are employed without hindering the pioneer regulations implemented in different periods. Therefore, artificial intelligence technology is utilized for prompt and swift governance decisions in delivering special rights for consumers and shareholders. The performance of this model is validated and verified using data sources related to governance policies from a real-time industry. The impact of varying policy features with dissimilarity is analyzed for varying occurrences. The analysis is given based on the considered data sources for which the classification and its impact over reports, sharing, voting, complaint, and market are analyzed. The availability before and after the proposed improves the above metrics by 10.48, 10.65, 9.78, 13.39, and 9.26%.",,https://www.frontiersin.org/articles/10.3389/fpsyg.2022.966689/pdf
-SRCB at the NTCIR-16 Real-MedNLP Task,Yongwei Zhang,"The SRCB participated in subtask1: Few-resource Named Entity Recognition (NER) and subtask3: Adverse Drug Event detection (ADE) in NTCIR-16 Real-MedNLP. This paper reports our approach to solve the problem and discusses the official results. For the Few-resource NER subtask, we developed NER systems based on pretraining model, span-based classification and prompt learning. In addition, data augmentation and model ensemble are used to further improve performance. For ADE subtask, we mainly adopted two methods: multi-class classification and prompt learning. We em-ployed a two-stage training strategy to solve the long tail distribution problem and applied transfer learning to improve performance of model.","{'model': 'tldr@v2.0.0', 'text': 'The SRCB participated in subtask1: Few-resource Named Entity Recognition and subtask3: Adverse Drug Event detection in NTCIR-16 Real-MedNLP and adopted two methods: multi-class classification and prompt learning.'}",
-Controllable Generation of Dialogue Acts for Dialogue Systems via Few-Shot Response Generation and Ranking,Angela Ramirez,"Dialogue systems need to produce responses that realize multiple types of dialogue acts (DAs) with high semantic fidelity. In the past, natural language generators (NLGs) for dialogue were trained on large parallel corpora that map from a domain-specific DA and its semantic attributes to an output utterance. Recent work shows that pretrained language models (LLMs) offer new possibilities for controllable NLG using prompt-based learning. Here we develop a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs. We compare eight few-shot prompt styles that include a novel method of generating from textual pseudo-references using a textual style transfer approach. We develop six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time. We test our approach on three domains and four LLMs. To our knowledge, this is the first work on NLG for dialogue that automatically ranks outputs using both DA and attribute accuracy. For completeness, we compare our results to fine-tuned few-shot models trained with 5 to 100 instances per DA. Our results show that several prompt settings achieve perfect DA accuracy, and near perfect semantic accuracy (99.81%) and perform better than few-shot fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs in natural language generators for dialogue and develops six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time.'}",https://arxiv.org/pdf/2307.14440
-On Conditional and Compositional Language Model Differentiable Prompting,Jonathan Pilault,"Prompts have been shown to be an effective method to adapt a frozen Pretrained Language Model (PLM) to perform well on downstream tasks. Prompts can be represented by a human-engineered word sequence or by a learned continuous embedding. In this work, we investigate conditional and compositional differentiable prompting. We propose a new model, Prompt Production System (ProPS), which learns to transform task instructions or input metadata, into continuous prompts that elicit task-specific outputs from the PLM. Our model uses a modular network structure based on our neural formulation of Production Systems, which allows the model to learn discrete rules -- neural functions that learn to specialize in transforming particular prompt input patterns, making it suitable for compositional transfer learning and few-shot learning. We present extensive empirical and theoretical analysis and show that ProPS consistently surpasses other PLM adaptation techniques, and often improves upon fully fine-tuned models, on compositional generalization tasks, controllable summarization and multilingual translation, while needing fewer trainable parameters.","{'model': 'tldr@v2.0.0', 'text': 'A new model is proposed, Prompt Production System (ProPS), which learns to transform task instructions or input metadata, into continuous prompts that elicit task-specific outputs from the PLM, and consistently surpasses other PLM adaptation techniques, and often improves upon fully fine-tuned models.'}",https://arxiv.org/pdf/2307.01446
-Adapting Pre-trained Language Models to Vision-Language Tasks via Dynamic Visual Prompting,Shubin Huang,"Pre-trained language models (PLMs) have played an increasing role in multimedia research. In terms of vision-language (VL) tasks, they often serve as a language encoder and still require an additional fusion network for VL reasoning, resulting in excessive memory overhead. In this paper, we focus on exploring PLMs as a stand-alone model for VL reasoning tasks. Inspired by the recently popular prompt tuning, we first prove that the processed visual features can be also projected onto the semantic space of PLMs and act as prompt tokens to bridge the gap between single- and multi-modal learning. However, this solution exhibits obvious redundancy in visual information and model inference, and the placement of prompt tokens also greatly affects the final performance. Based on these observations, we further propose a novel transfer learning approach for PLMs, termed Dynamic Visual Prompting (DVP). Concretely, DVP first deploys a cross-attention module to obtain text-related and compact visual prompt tokens, thereby greatly reducing the input length of PLMs. To obtain the optimal placement, we also equip DVP with a reinforcement-learning based search algorithm, which can automatically merge DVP with PLMs for different VL tasks via a very short search process. In addition, we also experiment DVP with the recently popular adapter approach to keep the most parameters of PLMs intact when adapting to VL tasks, helping PLMs achieve a quick shift between single- and multi-modal tasks. We apply DVP to two representative PLMs, namely BERT and T5, and conduct extensive experiments on a set of VL reasoning benchmarks including VQA2.0, GQA and SNLIVE. The experimental results not only show the advantage of DVP on efficiency and performance, but also confirm its superiority in adapting pre-trained language models to VL tasks.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results not only show the advantage of DVP on efficiency and performance, but also confirm its superiority in adapting pre-trained language models to VL tasks.'}",https://arxiv.org/pdf/2306.00409
-Supporting Learning of Variable Control in a Computer-Based Biology Environment: Effects of Prompting College Students to Reflect on Their Own Thinking,Xiaodong Lin,"While instruction on control of variables has been shown to be effective, especially when it encourages students to focus explicitly on rules or procedures, little evidence of application to novel problems has been obtained. We hypothesized that prompting students to understand their own learning processes while doing experiments involving control of variables would allow them to activate their reper- toire of knowledge and strategies and learn in a way that would enhance transfer of learning. Students were assigned to one of four versions of a computer-based biology simulation learning environment, each em- ploying a different type of prompt: reason justification, rule based, emotion focused, or none (control). Learning in this computer environment, college biology students designed and conducted experiments in- volving control of variables. Students' ability to solve both contextually similar (near transfer) and con- textually dissimilar (far transfer) problems was assessed. The treatment groups performed equally well on contextually similar problems. However, on a contextually dissimilar problem, the reason justification group had significantly higher scores than the other groups. Qualitative data showed that the reason justi- fication prompts directed students' attention to understanding when, why, and how to employ experiment design principles and strategies, and this in turn helped students to transfer their understanding to a novel problem. © 1999 John Wiley & Sons, Inc. J Res Sci Teach 36: 837- 858, 1999",,
-ECO: Ensembling Context Optimization for Vision-Language Models,Lorenzo Agnolucci,"Image recognition has recently witnessed a paradigm shift, where vision-language models are now used to perform few-shot classification based on textual prompts. Among these, the CLIP model has shown remarkable capabilities for zero-shot transfer by matching an image and a custom textual prompt in its latent space. This has paved the way for several works that focus on engineering or learning textual contexts for maximizing CLIP's classification capabilities. In this paper, we follow this trend by learning an ensemble of prompts for image classification. We show that learning diverse and possibly shorter contexts improves considerably and consistently the results rather than relying on a single trainable prompt. In particular, we report better few-shot capabilities with no additional cost at inference time. We demonstrate the capabilities of our approach on 11 different benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that learning diverse and possibly shorter contexts improves considerably and consistently the results rather than relying on a single trainable prompt for image classification, and reports better few-shot capabilities with no additional cost at inference time.'}",https://arxiv.org/pdf/2307.14063
-Generalizing Few-Shot Named Entity Recognizers to Unseen Domains with Type-Related Features,Zihan Wang,"Few-shot named entity recognition (NER) has shown remarkable progress in identifying entities in low-resource domains. However, few-shot NER methods still struggle with out-of-domain (OOD) examples due to their reliance on manual labeling for the target domain. To address this limitation, recent studies enable generalization to an unseen target domain with only a few labeled examples using data augmentation techniques. Two important challenges remain: First, augmentation is limited to the training data, resulting in minimal overlap between the generated data and OOD examples. Second, knowledge transfer is implicit and insufficient, severely hindering model generalizability and the integration of knowledge from the source domain. In this paper, we propose a framework, prompt learning with type-related features (PLTR), to address these challenges. To identify useful knowledge in the source domain and enhance knowledge transfer, PLTR automatically extracts entity type-related features (TRFs) based on mutual information criteria. To bridge the gap between training and OOD data, PLTR generates a unique prompt for each unseen example by selecting relevant TRFs. We show that PLTR achieves significant performance improvements on in-domain and cross-domain datasets. The use of PLTR facilitates model adaptation and increases representation similarities between the source and unseen domains.","{'model': 'tldr@v2.0.0', 'text': 'To identify useful knowledge in the source domain and enhance knowledge transfer, PLTR automatically extracts entity type-related features (TRFs) based on mutual information criteria and generates a unique prompt for each unseen example by selecting relevant TRFs.'}",
-Optimizing Learning Environments: An Individual Difference Approach to Learning and Transfer,Daniel M. Belenky,"Optimizing Learning Environments: An Individual Difference Approach to Learning and Transfer Daniel M. Belenky (dmb83@pitt.edu) Timothy J. Nokes (nokes@pitt.edu) Learning Research and Development Center University of Pittsburgh, 3939 O’Hara Street Pittsburgh, PA 15260 manipulations of task goals through instructions can change the ways students learn, similar to the effect of task structure. If directly manipulating task goals produces similar effects, it would offer a more direct way of encouraging students towards desired learning outcomes (whether towards transfer or skill). However, it is possible that achievement goals within a learning activity are not under conscious control, and task structure has more influence on how a student engages than instructions that attempt to prompt a particular achievement goal. It is also possible that task structure and task goals operate independently, leading to a three-way interaction in the adoption of achievement goals based on students’ prior dispositions. This study explores these possibilities. Abstract Prior work has found that the type of learning activity (direct instruction or invention) interacts with achievement goals (mastery or performance-oriented) such that invention tasks can help facilitate mastery goal adoption and knowledge transfer (Belenky & Nokes, 2009). In the current study, we investigated how robust the effect is, and whether explicit manipulations of the task goals can produce a similar effect. We conducted an experiment with 98 college students in which achievement goals were measured, while task goals and task structure were manipulated between subjects. Results indicated that task structure was generally a more effective way of influencing which achievement goals are adopted within a learning activity. However, task goals that promoted an evaluative context interfered with transfer for mastery- oriented learners from invention activities. The results are interpreted in relation to theories of regulatory fit and multiple goal hierarchies. Background Research on achievement goals has focused on two main aims; classifying what the goals are and then correlating those with predictors and outcomes. The prevailing classification is a 2 x 2 framework that has been well- validated (Elliot & McGregor, 2001). This framework separates the evaluative criterion (mastery or performance) from the valence (approach or avoidance), which results in four separable goals (mastery-approach, mastery-avoidance, performance-approach, performance-avoidance). Mastery goals refer to ones in which a person is basing his evaluation on the skill or competence he is trying to develop (that is, in comparison to an expectation or prior ability), while performance goals refer to evaluating oneself based on a normative standard (that is, in comparison to others). Approach goals refer to seeking out positive outcomes, while avoidance goals refer to averting negative ones. For example, a mastery-approach goal is one in which a person is seeking to improve his ability or knowledge, based on an internally-referenced criterion (“My aim is to completely master the material in this class”), while a performance- avoidance goal is one in which a person is seeking to not look bad compared to others (“My aim is to avoid doing worse than other students;” see Elliot & McGregor, 2001). Students can have different levels of each of these goals, and studies have validated that these four goals are separate factors (Elliot & McGregor, 2001). Because we are most interested in studying how different paths of successful learning affect what knowledge is gained, our work focuses on mastery-approach and performance-approach goals. Mastery-approach (MAP) goals have been correlated with a host of positive outcomes, such as intrinsic motivation, interest, better self-regulation, and deeper strategy use. Keywords: learning; transfer; skill acquisition; motivation; achievement goals. Student’s achievement goals have a large influence on their behaviors and experiences in academic settings. The literature surrounding Achievement Goal theory shows that these goals lead to very different patterns of affect, interest and achievement (e.g., Harackiewicz et al., 2005). However, this literature has not focused on how the goals influence what is learned. That is, although “achievement” is frequently measured as an outcome, it is almost always done at a coarse-grain level, such as final grades in a course. It is not clear how different achievement goals (mastery versus performance) are related to different kinds of learning, such as learning procedural skills, simple facts, or conceptual knowledge. To begin to address this gap, Belenky & Nokes (2009) examined how achievement goals impact the type of knowledge gained from different kinds of instruction. That study found that mastery-oriented learners do better on transfer measures, regardless of whether the mastery- orientation came from a stable predisposition or whether the open-ended structure of an “invention” task led to mastery- oriented feelings and goals in the specific context. Conversely, performance-oriented learners did better on skill acquisition when the instruction seemed to match their goals, by presenting a well-structured, simple task through direct instruction. This initial work has provided evidence that task structure interacts with existing achievement goals to influence learning. In the current work we examine whether direct",,
-Analysis of Factors Affecting the Length of Stay in the Emergency Department for Critically Ill Patients Transferred to Regional Emergency Medical Centers,Hyung-Yeol Lee," BackgroundTransferring an emergency patient to another emergency department (ED) is necessary when she/he is unable to receive necessary treatment from the first visited ED, although the transfer poses potential risks for adverse clinical outcomes and lowering the quality of emergency medical services by overcrowding the transferred ED. This study aimed to understand the factors affecting the ED length of stay (LOS) of critically ill patients and to investigate whether they are receiving prompt treatment through Interhospital Transfer (IHT).MethodsThis study analyzed 968 critically ill patients transferred to the ED of the study site in 2019. Machine learning based prediction models were built to predict the ED LOS dichotomized as greater than 6 hours or less. Explanatory variables in patient characteristics, clinical characteristics, transfer-related characteristics, and ED characteristics were selected through univariate analyses.ResultsAmong the prediction models, the Logistic Regression (AUC 0.85) model showed the highest prediction performance, followed by Random Forest (AUC 0.83) and Naïve Bayes (AUC 0.83). The Logistic Regression model suggested that the need for emergency operation or angiography (OR 3.91, 95% CI=1.65–9.21), the need for Intensive Care Unit (ICU) admission (OR 3.84, 95% CI=2.53–5.83), fewer consultations (OR 3.57, 95% CI=2.84–4.49), a high triage level (OR 2.27, 95% CI=1.43–3.59), and fewer diagnoses (OR 1.32, 95% CI=1.09–1.61) coincided with a higher likelihood of 6-hour-or-less stays in the ED. Furthermore, an interhospital transfer handoff led to significantly shorter ED LOS among the patients who needed emergency operation or angiography, or ICU admission, or had a high triage level.ConclusionsThe results of this study suggest that patients prioritized in emergency treatment receive prompt intervention and leave the ED in time. Also, having a proper interhospital transfer handoff before IHT is crucial to provide efficient care and avoid unnecessarily longer stay in ED.","{'model': 'tldr@v2.0.0', 'text': 'The results of this study suggest that patients prioritized in emergency treatment receive prompt intervention and leave the ED in time, and having a proper interhospital transfer handoff before IHT is crucial to provide efficient care and avoid unnecessarily longer stay in ED.'}",
-An Initial Investigation of Non-Native Spoken Question-Answering,V. Raina,"Text-based machine comprehension (MC) systems have a wide-range of applications, and standard corpora exist for developing and evaluating approaches. There has been far less research on spoken question answering (SQA) systems. The SQA task considered in this paper is to extract the answer from a candidate$\text{'}$s spoken response to a question in a prompt-response style language assessment test. Applying these MC approaches to this SQA task rather than, for example, off-topic response detection provides far more detailed information that can be used for further downstream processing. One significant challenge is the lack of appropriately annotated speech corpora to train systems for this task. Hence, a transfer-learning style approach is adopted where a system trained on text-based MC is evaluated on an SQA task with non-native speakers. Mismatches must be considered between text documents and spoken responses; non-native spoken grammar and written grammar. In practical SQA, ASR systems are used, necessitating an investigation of the impact of ASR errors. We show that a simple text-based ELECTRA MC model trained on SQuAD2.0 transfers well for SQA. It is found that there is an approximately linear relationship between ASR errors and the SQA assessment scores but grammar mismatches have minimal impact.","{'model': 'tldr@v2.0.0', 'text': 'It is found that there is an approximately linear relationship between ASR errors and the SQA assessment scores but grammar mismatches have minimal impact.'}",
-Transfer Effects of Learning through Two Varied Worked Examples on Word-problem Solving,Zhang Qi,"There is disagreement with regard to the transfer effects of learning through varied worked examples. Ross(1989)found that multiple examples should be made very similar to each other; even a small difference in the surface feature could make the learner pay more attention to solving the problem. Gentner (2003) discovered that comparing examples with the same structure could prompt the learners' transition to problem solving. Holyoak (1987) pointed out that deeper structural differences had an impact on the transfer. Based on these researches, the present study hypothesized that the variability of the surface feature of two examples could have some effect on the near transfer of pupils' word-problem solving, and the proper variability of the structural feature of the two examples could have an impact on the far transfer of pupils' word-problem solving. A total of 210 second-grade students from a primary school were selected using a pretest and were divided into six experimental groups and a control group. Each group learnt a word problem example. Then, the pupils of three experimental groups learned one kind of word problem example that varied from the first example with regard to the surface feature: the pupils in the first group learned a number and substance varied example; the pupils in the second group learned a story varied example; and those in the third group learned an expression varied example. The pupils in the other three experimental groups learned another kind of worked example that varied from the first problem with regard to the structural feature. The first group learned a rule varied example, the second group learned a rule repeated example; the third group, learned a rule of speed composed example. The pupils in the control group did not learn any worked example. Finally, all the pupils of the six experimental groups and the control group were given a test with 15 word problems. Three of them varied from the first example with regard to the surface feature, three of them varied from the first example with regard to the structural feature, and the others varied from the first example with regard to the surface and structural features. The experimenters recorded the pupils' test scores. The results were as follows. First, the pupils in the control group who learned the first word problem eyample could solve the word problems whose example varied with regard to the surface feature, but could not solve the other word problems. Second, the pupils in the three experimental groups who learned the word problem example that varied with regard to the surface feature could solve the word problems that varied from the first example with regard to the surface feature, but could not solve the other word problems. Third, a different effect was observed for the pupils in the three experimental groups who learned the word problem example that varied with regard to the structural feature. The pupils who learned the rule repeated example solved more word problems than the pupils in the other groups. The pupils who learned the rule varied example solved more word problems than those who learned the rule of speed composed example. The pupils who learned the rule of speed composed example solved the fewest number of word problems. The variability of the surface feature of the two examples could have some impact on the near transfer of pupils' word-problem solving. The rule repeated example had the best effect on the far transfer of the pupils' word-problem solving. The rule varied example had a more positive effect than the rule of speed composed example on the pupils' word-problem solving far transfer. Finally, the rule of speed composed example had the poorest effect on the far transfer of the pupils' word-problem solving.","{'model': 'tldr@v2.0.0', 'text': ""It is hypothesized that the variability of the surface feature of two examples could have some effect on the near transfer of pupils' word-problem solving, and the proper variability ofThe structural feature of the two example could have an impact on the far transfer of pupil's word- problem solving.""}",
-Toward Educator-focused Automated Scoring Systems for Reading and Writing,Mike Hardy,"This paper presents methods for improving automated essay scoring with techniques that address the computational trade-offs of self-attention and document length. To make Automated Essay Scoring (AES) more useful to practitioners, researchers must overcome the challenges of data and label availability, authentic and extended writing, domain scoring, prompt and source variety, and transfer learning. This paper addresses these challenges using neural network models by employing techniques that preserve essay length as an important feature without increasing model training costs. It introduces techniques for minimizing classification loss on ordinal labels using multi-objective learning, capturing semantic information across the entire essay using sentence embeddings to use transformer architecture across arbitrarily long documents, the use of such models for transfer learning, automated hyperparameter generation based on prompt-corpus metadata, and, most importantly, the use of semantic information to provide meaningful insights into student reading through analysis of passage-dependent writing resulting in state-of-the-art results for various essay tasks.","{'model': 'tldr@v2.0.0', 'text': None}",
-Mobility Awareness in Cellular Networks to Support Service Continuity in Vehicular Users,Nandish P. Kuruvatti,"Mobile communication is an ubiquitously used technology that has evolved through various generations and is currently on the verge of its fifth generation (5G). In the recent years, Intelligent Transportation Systems (ITS) and supplementary vehicular use cases (e.g., autonomous driving) are considered widely within the scope of cellular networks. These use cases generally demand reliable and low latency services from the cellular network. Mobile Edge Clouds (MEC) in the 5G networks are often applicable to satiate such service demands of a vehicular user. However, the cellular handovers (HO) of vehicular users prompt frequent service migration among the MECs. The handovers and service migration increase the service interruption of a user. In this paper, we consider machine learning (ML) based mobility awareness to obtain future service migration and HO sites of a user. This enables smooth service migration by allowing non-state data transfer earlier to user handover. Further, it provides sufficient time for establishment of successful Coordinated Multipoint (CoMP) transmissions, which will reduce service interruption due to HO. Simulation results show that the proposed framework provides timely assistance for service migration and significantly reduces the service interruption time.","{'model': 'tldr@v2.0.0', 'text': 'Simulation results show that the proposed framework provides timely assistance for service migration and significantly reduces the service interruption time.'}",
-Recognizing Malaysia Traffic Signs with Pre-Trained Deep Convolutional Neural Networks,Tze How Dickson Neoh,"An essential component in the race towards the self-driving car is automatic traffic sign recognition. The capability to automatically recognize road signs allow self-driving cars to make prompt decisions such as adhering to speed limits, stopping at traffic junctions and so forth. Traditionally, feature-based computer vision techniques were employed to recognize traffic signs. However, recent advancements in deep learning techniques have shown to outperform traditional color and shape based detection methods. Deep convolutional neural network (DCNN) is a class of deep learning method that is most commonly applied to vision-related tasks such as traffic sign recognition. For DCNN to work well, it is imperative that the algorithm is given a vast amount of training data. However, due to the scarcity of a curated dataset of the Malaysian traffic signs, training DCNN to perform well can be very challenging. In this demonstrate that DCNN can be trained with little training data with excellent accuracy by using transfer learning. We retrain various pre-trained DCNN from other image recognition tasks by fine-tuning only the top layers on our dataset. Experiment results confirm that by using as little as 100 image samples for 5 different classes, we are able to classify hitherto traffic signs with above 90% accuracy for most pre-trained models and 98.33% for the DenseNet169 pre-trained model.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that DCNN can be trained with little training data with excellent accuracy by using transfer learning, and is able to classify hitherto traffic signs with above 90% accuracy for most pre-trained models and 98.33% for the DenseNet169 pre- trained model.'}",
-Regarding Solid Oxide Fuel Cells Simulation through Artificial Intelligence: A Neural Networks Application,Arianna Baldinelli,"Because of their fuel flexibility, Solid Oxide Fuel Cells (SOFCs) are promising candidates to coach the energy transition. Yet, SOFC performance are markedly affected by fuel composition and operative parameters. In order to optimize SOFC operation and to provide a prompt regulation, reliable performance simulation tools are required. Given the high variability ascribed to the fuel in the wide range of SOFC applications and the high non-linearity of electrochemical systems, the implementation of artificial intelligence techniques, like Artificial Neural Networks (ANNs), is sound. In this paper, several network architectures based on a feedforward-backpropagation algorithm are proposed and trained on experimental data-set issued from tests on commercial NiYSZ/8YSZ/LSCF anode supported planar button cells. The best simulator obtained is a 3-hidden layer ANN (25/22/18 neurons per layer, hyperbolic tangent sigmoid as transfer function, obtained with a gradient descent with adaptive learning rate backpropagation). This shows high accuracy (RMS = 0.67% in the testing phase) and successful application in the forecast of SOFC polarization behaviour in two additional experiments (RMS in the order of 3% is scored, yet it is reduced to about 2% if only the typical operating current density range of real application is considered, from 300 to 500 mA·cm−2). Therefore, the neural tool is suitable for system simulation codes/software whether SOFC operating parameters agree with the input ranges (anode feeding composition 0–48%vol H2, 0–38%vol CO, 0–45%vol CH4, 9–32%vol CO2, 0–54%vol N2, specific equivalent hydrogen flow-rate per unit cell active area 10.8–23.6 mL·min−1·cm−2, current density 0–1300 mA·cm−2 and temperature 700–800 °C).",,https://www.mdpi.com/2076-3417/9/1/51/pdf?version=1547101227
-Database for Research Projects to Solve the Inverse Heat Conduction Problem,S. Szénási,"To achieve the optimal performance of an object to be heat treated, it is necessary to know the value of the Heat Transfer Coefficient (HTC) describing the amount of heat exchange between the work piece and the cooling medium. The prediction of the HTC is a typical Inverse Heat Transfer Problem (IHCP), which cannot be solved by direct numerical methods. Numerous techniques are used to solve the IHCP based on heuristic search algorithms having very high computational demand. As another approach, it would be possible to use machine-learning methods for the same purpose, which are capable of giving prompt estimations about the main characteristics of the HTC function. As known, a key requirement for all successful machine-learning projects is the availability of high quality training data. In this case, the amount of real-world measurements is far from satisfactory because of the high cost of these tests. As an alternative, it is possible to generate the necessary databases using simulations. This paper presents a novel model for random HTC function generation based on control points and additional parameters defining the shape of curve segments. As an additional step, a GPU accelerated finite-element method was used to simulate the cooling process resulting in the required temporary data records. These datasets make it possible for researchers to develop and test their IHCP solver algorithms.","{'model': 'tldr@v2.0.0', 'text': 'A novel model for random HTC function generation based on control points and additional parameters defining the shape of curve segments is presented, which makes it possible for researchers to develop and test their IHCP solver algorithms.'}",https://www.mdpi.com/2306-5729/4/3/90/pdf?version=1561620123
-Principles of integrating research into teaching in higher education: A knowledge transfer perspective,R. Amaratunga,"Integrating new knowledge created through research with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures and research funding mechanisms within the UK higher education system. The link between research and teaching is not automatic. Thus, it needs to be formally created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study on which this paper is based on, aims to develop principles to enable transfer of research knowledge into teaching. This study is different from previous approaches in that it looks into the research and teaching link as a two-way knowledge transfer process in the light of growing knowledge management viewpoints. In developing the framework, this research uses the case study approach. Finally, the research introduces principles of research to teaching transfer that are applicable for higher education institutions.",,
-Construction of a Web-based virtual classroom and its effective analysis,Qingtang Liu,"In classroom, information between teachers and students includes course content and learning scene. Course content is mainly information, and learning scene is assisted information that can effectively prompt learners exchanging idea and acquiring knowledge. A based-Web virtual classroom should have the same features. Because of the limitation of bandwidth and QoS of network, it is very difficult to transfer teaching and learning information by video in Internet. The paper probes into a strategy of developing virtual classroom by the streaming media technologies. By contrast with two kinds of information, the Web-based virtual classroom we built includes two parts: real time teaching environment (RTTE) and collaborating learning environment (CLE). RTTE transports course content and feedback information into teachers or students. This paper firstly discusses the strategies of creating a virtual classroom for the remote learners. Secondly, it designs the architecture of a virtual classroom.","{'model': 'tldr@v2.0.0', 'text': 'The strategies of creating a virtual classroom for the remote learners are discussed and the architecture of a based-Web virtual classroom is designed, which includes real time teaching environment (RTTE) and collaborating learning environment (CLE).'}",
-Principles of Integrating Research into Teaching in Higher Education: Built Environment Perspective,D. Amaratunga,"Integrating research-based knowledge with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures, and research funding mechanisms within the U.K. higher education system. The link between research and teaching is not automatic. Thus, it needs to be formally created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study on which this paper is based aims to develop generic principles to enable transfer of research knowledge into teaching. This study differs from previous approaches in that it examines the research and teaching link as a two-way knowledge transfer process in the light of growing knowledge management viewpoints. This research uses the case study approach and has conducted five case studies that represent five disciplines. Finally, the research introduces principles of research to teaching transfer that are applicable to higher education institutions.",,
-Diversity: Changing Perspectives Using Case-Based Methods.,L. Andrews,"The case method is a popular teaching tool that can be applied to training for diversity and particularly the preparation of general education teachers for the inclusion of diverse learners in their classrooms. Case-based methodology teacher training activities focus on actual classroom experiences and prompt reflection and personal involvement. A study was done to validate these techniques with 56 student teachers in 2 groups. One group used a problem-solving strategy to analyze differing case features. This strategy group did 40 percent better on case problem-solving than those solving case problems without the strategy. A comparison was made of the identification of features with problem-solving scores to determine if the strategy was the determining factor, and a high positive correlation was found. The case method of teaching was found to promote classroom problem-solving when cases are systematically varied and when student teachers are given a strategy that guides transfer of learning from case to case. This demonstrates that student teachers can be trained to transfer learning from one teaching problem to another involving a diverse population of students. (Contains 8 references.) (JLS) *********************************************************************** Reproductions supplied by EDRS are the best that can be made from the original document. *********************************************************************** DIVERSITY: CHANGING PERSPECTIVES USING CASE-BASED METHODS A Paper presented at the 49th Annual Meeting of the American Association of Colleges for Teacher Education by Lanna Andrews, Ed.D Assistant Professor University of San Francisco 2130 Fulton Street San Francisco CA 94117-1980 (415) 666-2099 Office U.S. DEPARTMENT OF EDUCATION of Educational Research and Improvement EDUCATIONAL RESOURCES INFORMATION 140 CENTER (ERIC) 0 This document hes been reproduced as Ar) originating received from the person or organization (.011VT% 0 Minor changes have reproduction en made to improve Quality.. IK\ Points of view or opinions stated in thiadocument do not necesaarity represent official V3/43.1 OERI positron or 0011Cy. BEST COPY AVAILABLE 2 PERMISSION TO REPRODUCE AND DISSEMINATE THIS MATERIAL S BEEN GRANTED BY TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) Diversity: Changing Perspectives Using Case-Based Methods Since diversity is a term that accurately describes classrooms in today's urban centers, teacher trainers are charged with preparing novice teachers to meet the challenge (Grossman, 1995). This paper discusses the use of a popular teaching tool, the case method, and how it can be applied to training for diversity. Emphasis will be on the preparation of general education teachers for the inclusion of diverse learners into their classrooms. Topics to be covered include: (1) integration of cases into college curricula to promote cultural responsiveness; (2) use of varied techniques with the case method, such as cooperative learning, reflective writing, interactive discussion, problem-solving strategies; (3) and results of an empirical study incorporating all the above to promote diversity, completed in four sections of a Mainstreaming Course with general education teacher trainees. Theoretical Framework Discourse about culturally responsive curriculum takes place within the context of debate about a larger issue, the validity and shape of a multicultural approach to education. A widely held view of multicultural or culturally responsive curricula sees them as strategies for improving performance and enhancing self esteem among students whose racial, ethnic, or language-differs from that of the AngloEuropean population (Lubeck, 1988). This proposal takes the view that self esteems comes from inclusive practices in the classroom (Villa and Thousand, 1995). These include awareness and value of differences along with the realization that similarities are equally important, especially when designing curriculum. In addition, teachers must be trained to notice similarities and differences in order to transfer learning from",,
-Integrating research and teaching in higher education: Conceptual issues,S. Senaratne,"Integrating new knowledge created through research with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures and research funding mechanisms within the UK higher education system. The link between research and teaching is not automatic. Thus, it needs to be created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study, on which this paper is based on, aims to develop principles in relation to transferring research knowledge into teaching through a literature review and case studies. The paper reports conceptual issues related to such a transfer process based on the literature findings.",,
-AdaBoost-based transfer learning method for positive and unlabelled learning problem,Bo Liu,,,
-An Image-based Transfer Learning Framework for Classification of E-Commerce Products,Vrushali Atul Surve,"Classification of e-commerce products involves identifying the products and placing those products into the correct category. For example, men’s Nike Air Max will be in the men’s category shoes on an e-Commerce platform. Identifying the correct classification of a product from hundreds of categories is time-consuming for businesses. This research proposes an Image-based Transfer Learning Framework to classify the images into the correct category in the shortest time. The framework combines Image-based algorithms with Transfer Learning. This research compares the time to predict the category and accuracy of traditional CNN and transfer learning models such as VGG19, InceptionV3, ResNet50, and MobileNet. A visual classifier is trained CNN and transfer learning models such as VGG19, InceptionV3, ResNet50, and MobileNet. The models are trained on an e-commerce product dataset that combines the ImageNet dataset with pre-trained weights. The dataset consists of 15000 images scraped from the web. Results demonstrate that Inception V3 outperforms all other models based on a TIMING of 0.10 seconds and an accuracy of 85%.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes an Image-based Transfer Learning Framework to classify the images into the correct category in the shortest time and demonstrates that Inception V3 outperforms all other models based on a TIMING of 0.10 seconds and an accuracy of 85%.'}",https://norma.ncirl.ie/6322/1/vrushaliatulsurve.pdf
-A Transfer Learning-Based Method for Personalized State of Health Estimation of Lithium-Ion Batteries.,Guijun Ma,"State of health (SOH) estimation of lithium-ion batteries (LIBs) is of critical importance for battery management systems (BMSs) of electronic devices. An accurate SOH estimation is still a challenging problem limited by diverse usage conditions between training and testing LIBs. To tackle this problem, this article proposes a transfer learning-based method for personalized SOH estimation of a new battery. More specifically, a convolutional neural network (CNN) combined with an improved domain adaptation method is used to construct an SOH estimation model, where the CNN is used to automatically extract features from raw charging voltage trajectories, while the domain adaptation method named maximum mean discrepancy (MMD) is adopted to reduce the distribution difference between training and testing battery data. This article extends MMD from classification tasks to regression tasks, which can therefore be used for SOH estimation. Three different datasets with different charging policies, discharging policies, and ambient temperatures are used to validate the effectiveness and generalizability of the proposed method. The superiority of the proposed SOH estimation method is demonstrated through the comparison with direct model training using state-of-the-art machine learning methods and several other domain adaptation approaches. The results show that the proposed transfer learning-based method has wide generalizability as well as a positive precision improvement.","{'model': 'tldr@v2.0.0', 'text': 'This article extends MMD from classification tasks to regression tasks, which can be used for SOH estimation and shows that the proposed transfer learning-based method has wide generalizability as well as a positive precision improvement.'}",
-Deep representation-based transfer learning for deep neural networks,Tao Yang,,,
-Aspect-level sentiment classification based on attention-BiLSTM model and transfer learning,Guixian Xu,,,
-CDLX: An Efficient novel approach for COVİD detection lung x-rays throw transfer learning based on State of the art deep learning image classification models,Y. N. Afreen,"When compared to the general population, lung cancer patients have a higher incidence of COVID-19 infection, pulmonary problems, and poorer survival results. As a reference for prioritising cancer care issues during the epidemic, the world's main professional organisations issued new recommendations for the diagnosis, treatment, and follow-up of lung cancer patients. In today's world, we are fighting one of the greatest pandemics in human history, known as COVID-2019, which is caused by a coronavirus. The patient can be treated promptly if the infection is detected early (before it enters the lower respiratory tract). To observe ground-glass opacity in the chest X-ray due to fibrosis in the lungs once the virus has reached the lungs. Artificial intelligence techniques can be utilised to detect the presence and degree of illness based on the major discrepancies between X-ray images of an infected and non-infected person. For this study, I employed feature extraction from Transfer Learning, which entails importing a pre-trained CNN model, such as Distributed Deep Convolutional VGGNet or Distributed Deep Convolutional with ResNet Model, and changing the last layer to meet my needs.","{'model': 'tldr@v2.0.0', 'text': 'To observe ground-glass opacity in the chest X-ray due to fibrosis in the lungs once the virus has reached the lungs, artificial intelligence techniques can be utilised to detect the presence and degree of illness based on the major discrepancies betweenX-ray images of an infected and non-infected person.'}",https://sciencescholar.us/journal/index.php/ijhs/article/download/10593/7612
-FewFine: Few-shot Malware Traffic Classification Via Transfer Learning based on Fine-tuning Strategy,Xingtong Liu,"Malware traffic is constantly evolving and remains destructive. The detection and classification of malware traffic is crucial for maintaining cyberspace security. Only by swiftly and accurately detecting and classifying malware traffic can user privacy and cyberspace security be effectively protected.In this paper, we propose FewFine, an approach for few-shot malware traffic classification based on transfer learning. We initially pre-train a detection model and two classification models with substantial quantity of malware and application traffic samples. For classifying new types of malware traffic accurately and promptly, we utilize transfer learning based on fine-tuning strategy and freeze several blocks in the pre-trained model. Utilizing prior knowledge from the pre-trained models, we leverage few samples of novel classes to perform accurate malware detection and classification. We execute extensive experiments on publicly available datasets to evaluate the effectiveness of FewFine. In model pre-training, with considerable number of samples, the accuracy of malware detection and classification can reach 0.99. The pre-trained models are saved for fine-tuning. When detecting and classifying novel malware traffic, FewFine can achieve the accuracy of 0.95 leveraging only 10 samples per class through fine-tuning the pre-trained model. It outperforms methods under comparison in terms of accuracy and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'FewFine, an approach for few-shot malware traffic classification based on transfer learning, outperforms methods under comparison in terms of accuracy and efficiency.'}",
-Transfer learning based robust automatic detection system for diabetic retinopathy grading,C. Bhardwaj,,"{'model': 'tldr@v2.0.0', 'text': 'This work focuses on automatic DR disease identification and its grading by the means of transfer learning approach using dynamic investigation and the diagnosis ability of the proposed transfer learning-based model is justified by comparing the proposed methods with the state-of-the-art methods.'}",
-A concise peephole model based transfer learning method for small sample temporal feature-based data-driven quality analysis,Wentao Luo,,"{'model': 'tldr@v2.0.0', 'text': 'A novel transfer learning model and algorithm with feature mapping, feature learning and domain adaptation was proposed based on concise peephole model (TLCPM), which showed the TLCPM can be properly applied to analyze small sample temporal feature-based data and achieved good comprehensive performance.'}",
-Output-based transfer learning in genetic programming for document classification,Wenlong Fu,,"{'model': 'tldr@v2.0.0', 'text': 'An investigation on an output-based transfer learning system using Genetic Programming in document classification tasks, which automatically selects features to construct classifiers, demonstrates that the GP programs from SD can be effectively used for classifying documents in the relevant TD.'}",
-Motor Imagery EEG Signal Classification based on Deep Transfer Learning,Ming-fei Wei,"Deep transfer learning (DTL) has developed rapidly in the field of motor imagery (MI) on brain-computer interface (BCI) in recent years. DTL utilizes deep neural networks with strong generalization capabilities as the pre-training framework and automatically extracts richer and more expressive features during the training process. The goal of this paper is utilizing the DTL to classify MI electroencephalogram (EEG) signals on the premise of a small data set. The publicly available dataset III of the second BCI competition is applied in both the training part and testing part to evaluate the effectiveness of the proposed method. Firstly in the process, finite impulse response (FIR) filter and wavelet transform threshold denoising method are used to remove redundant signals and artifacts in EEG signals. Then, the continuous wavelet transform (CWT) is utilized to convert the one-dimensional EEG signal into a two-dimensional time-frequency amplitude representation as the input of the pre-trained convolutional neural network (CNN) for classifying two types of MI signals. Employing the input data of 140 trials for training, the final classification accuracy rate reaches 96.43%. Compared with the results of some superior machine learning models using the same data set, the accuracy and Kappa value of this DTL model are better. Therefore, the proposed scheme of MI EEG signal classification based on the DTL method offers preferably empirical performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed scheme of MI EEG signal classification based on the DTL method offers preferably empirical performance and the accuracy and Kappa value of this DTL model are better than some superior machine learning models using the same data set.'}",
-Model-Based Transfer Reinforcement Learning Based on Graphical Model Representations,Yuewen Sun,"Reinforcement learning (RL) plays an essential role in the field of artificial intelligence but suffers from data inefficiency and model-shift issues. One possible solution to deal with such issues is to exploit transfer learning. However, interpretability problems and negative transfer may occur without explainable models. In this article, we define Relation Transfer as explainable and transferable learning based on graphical model representations, inferring the skeleton and relations among variables in a causal view and generalizing to the target domain. The proposed algorithm consists of the following three steps. First, we leverage a suitable casual discovery method to identify the causal graph based on the augmented source domain data. After that, we make inferences on the target model based on the prior causal knowledge. Finally, offline RL training on the target model is utilized as prior knowledge to improve the policy training in the target domain. The proposed method can answer the question of what to transfer and realize zero-shot transfer across related domains in a principled way. To demonstrate the robustness of the proposed framework, we conduct experiments on four classical control problems as well as one simulation to the real-world application. Experimental results on both continuous and discrete cases demonstrate the efficacy of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'Relation Transfer is defined as explainable and transferable learning based on graphical model representations, inferring the skeleton and relations among variables in a causal view and generalizing to the target domain.'}",
-Deep Transfer Learning Based Multi-Class Brain Tumors Classification Using MRI Images,Mrinmoy Mondal,"A brain tumor is a severe disease that can be fatal and significantly impacts one’s quality of life. The traditional method of identifying tumors relies on physicians, which is time-consuming and prone to errors, putting the patient’s life in jeopardy. Identifying the classes of brain tumors is difficult due to the high anatomical and spatial diversity of the brain tumor’s surrounding region. An automated and precise diagnosis approach is required to treat this severe disease effectively. Deep learning technology, such as CNN, can be used to diagnose various tumor types in the early stages of their development using brain MRI. In this study, a deep transfer learning framework based on VGG-19 is introduced to accurately detect three common kinds of tumors from brain MRI. There are primarily two phases to the suggested framework. The VGG-19 frozen part is the first phase, while the modified neural style classification part is the second phase. With certain modified techniques, the class imbalance impact within the MRI dataset and the generalization error issue during the training process were also resolved. The proposed model has a 94% classification accuracy and a 94% F1-score.","{'model': 'tldr@v2.0.0', 'text': 'A deep transfer learning framework based on VGG-19 is introduced to accurately detect three common kinds of tumors from brain MRI, which has a 94% classification accuracy and a 95% F1-score.'}",
-Self-starting process monitoring based on transfer learning,Zhijun Wang,"Abstract Conventional self-starting control schemes can perform poorly when monitoring processes with early shifts, being limited by the number of historical observations sampled. In real applications, pre-observed data sets from other production lines are always available, prompting us to propose a scheme that monitors the target process using historical data obtained from other sources. The methodology of self-taught clustering from unsupervised transfer learning is revised to transfer knowledge from previous observations and improve out-of-control (OC) performance, especially for processes with early shifts. However, if the difference in distribution between the target process and the pre-observed data set is large, our scheme may not be the best. Simulation results and two illustrative examples demonstrate the superiority of the proposed scheme.","{'model': 'tldr@v2.0.0', 'text': 'The methodology of self-taught clustering from unsupervised transfer learning is revised to transfer knowledge from previous observations and improve out-of-control (OC) performance, especially for processes with early shifts.'}",
-Voting in Transfer Learning System for Ground-Based Cloud Classification,M. Manzo,"Cloud classification is a great challenge in meteorological research. The different types of clouds, currently known and present in our skies, can produce radioactive effects that impact the variation of atmospheric conditions, with consequent strong dominance over the earth’s climate and weather. Therefore, identifying their main visual features becomes a crucial aspect. In this paper, the goal is to adopt pretrained deep neural networks-based architecture for clouds image description, and subsequently, classification. The approach is pyramidal. Proceeding from the bottom up, it partially extracts previous knowledge of deep neural networks related to original task and transfers it to the new task. The updated knowledge is integrated in a voting context to provide a classification prediction. The framework trains the neural models on unbalanced sets, a condition that makes the task even more complex, and combines the provided predictions through statistical measures. An experimental phase on different cloud image datasets is performed, and the results achieved show the effectiveness of the proposed approach with respect to state-of-the-art competitors.","{'model': 'tldr@v2.0.0', 'text': 'The goal is to adopt a pretrained deep neural networks based architecture for clouds image description, and subsequently, classification and results show the effectiveness of the proposed approach with respect to state of the art competitors.'}",https://www.mdpi.com/2504-4990/3/3/28/pdf?version=1626092557
-Deep transfer learning-based network traffic classification for scarce dataset in 5G IoT systems,Jianfeng Guan,,"{'model': 'tldr@v2.0.0', 'text': 'A traffic classification method based on deep transfer learning for 5G IoT scenarios with scarce labeled data and limited computing capability, and trains the classification model by weight transferring and neural network fine-tuning different from the previous work that extract artificially designed features.'}",
-Transfer learning based surrogate assisted evolutionary bi-objective optimization for objectives with different evaluation times,Xilu Wang,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transfer learning scheme within a surrogate-assisted evolutionary algorithm framework to augment the training data for the surrogate for the slow objective function by transferring knowledge from the fast one and demonstrates that the proposed algorithm outperforms existing surrogate and non-surrogate-assisted delay-handling methods on a range of bi-objective optimization problems.'}",
-Robust transfer learning based on Geometric Mean Metric Learning,Penghui Zhao,,"{'model': 'tldr@v2.0.0', 'text': 'This paper integrates the transfer learning and metric learning into a unified framework and proposes a novel robust transfer learning based on geometric mean metric learning, namely Geometric Mean Transfer Learning (GMTL), which uses weighted geometricmean metric learning to model the intra-class distance and the inter-class similarity.'}",
-Ontology-based Transfer Learning in the Airport and Warehouse Logistics Domains,Miriam Herold,"This work is a position paper for the examination of ontology-based transfer learning in the context of business processes. We continue our preliminary work on transferring process-oriented knowledge from a well-known source domain to a less specified target domain. We outline our ideas on workflows from two specific contexts: passenger and baggage logistics at the airport on one hand and warehouse management logistics on the other hand. In the first step we automatically transform BPMN files from these two domains in two separate ontologies. In the next step we intend to use ontology mapping as a means for the transfer. We plan to examine the concepts of generalization and abstraction to ease the transfer. We claim that the mentioned domains are feasible candidates for transfer learning, as we find several analogies between the airport handling and warehouse management workflows. Additionally, we discuss possible utilization resp. benefits of the transfer learning within this two particular domains and draft the next steps for the future research.","{'model': 'tldr@v2.0.0', 'text': 'The preliminary work on transferring process-oriented knowledge from a well-known source domain to a less specified target domain is continued, and it is claimed that the mentioned domains are feasible candidates for transfer learning.'}",
-DATLMedQA: A Data Augmentation and Transfer Learning Based Solution for Medical Question Answering,Shuohua Zhou,"With the outbreak of COVID-19 that has prompted an increased focus on self-care, more and more people hope to obtain disease knowledge from the Internet. In response to this demand, medical question answering and question generation tasks have become an important part of natural language processing (NLP). However, there are limited samples of medical questions and answers, and the question generation systems cannot fully meet the needs of non-professionals for medical questions. In this research, we propose a BERT medical pretraining model, using GPT-2 for question augmentation and T5-Small for topic extraction, calculating the cosine similarity of the extracted topic and using XGBoost for prediction. With augmentation using GPT-2, the prediction accuracy of our model outperforms the state-of-the-art (SOTA) model performance. Our experiment results demonstrate the outstanding performance of our model in medical question answering and question generation tasks, and its great potential to solve other biomedical question answering challenges.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes a BERT medical pretraining model, using GPT-2 for question augmentation and T5-Small for topic extraction, calculating the cosine similarity of the extracted topic and using XGBoost for prediction, which outperforms the state-of-the-art SOTA model performance.'}",https://www.mdpi.com/2076-3417/11/23/11251/pdf?version=1637938982
-"IGLUE: A Benchmark for Transfer Learning across Modalities, Tasks, and Languages",Emanuele Bugliarello,"Reliable evaluation benchmarks designed for replicability and comprehensiveness have driven progress in machine learning. Due to the lack of a multilingual benchmark, however, vision-and-language research has mostly focused on English language tasks. To fill this gap, we introduce the Image-Grounded Language Understanding Evaluation benchmark. IGLUE brings together - by both aggregating pre-existing datasets and creating new ones - visual question answering, cross-modal retrieval, grounded reasoning, and grounded entailment tasks across 20 diverse languages. Our benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in a zero-shot setting, but also in newly defined few-shot learning setups. Based on the evaluation of the available state-of-the-art models, we find that translate-test transfer is superior to zero-shot transfer and that few-shot learning is hard to harness for many tasks. Moreover, downstream performance is partially explained by the amount of available unlabelled textual data for pretraining, and only weakly by the typological distance of target-source languages. We hope to encourage future research efforts in this area by releasing the benchmark to the community.","{'model': 'tldr@v2.0.0', 'text': 'The Image-Grounded Language Understanding Evaluation benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in a zero-shot setting, but also in newly defined few-shot learning setups.'}",
-An Uncertainty-Aware Transfer Learning-Based Framework for COVID-19 Diagnosis,Afshar Shamsi,"The early and reliable detection of COVID-19 infected patients is essential to prevent and limit its outbreak. The PCR tests for COVID-19 detection are not available in many countries, and also, there are genuine concerns about their reliability and performance. Motivated by these shortcomings, this article proposes a deep uncertainty-aware transfer learning framework for COVID-19 detection using medical images. Four popular convolutional neural networks (CNNs), including VGG16, ResNet50, DenseNet121, and InceptionResNetV2, are first applied to extract deep features from chest X-ray and computed tomography (CT) images. Extracted features are then processed by different machine learning and statistical modeling techniques to identify COVID-19 cases. We also calculate and report the epistemic uncertainty of classification results to identify regions where the trained models are not confident about their decisions (out of distribution problem). Comprehensive simulation results for X-ray and CT image data sets indicate that linear support vector machine and neural network models achieve the best results as measured by accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC). Also, it is found that predictive uncertainty estimates are much higher for CT images compared to X-ray images.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive simulation results for X-ray and CT image data sets indicate that linear support vector machine and neural network models achieve the best results as measured by accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC).'}",https://ieeexplore.ieee.org/ielx7/5962385/9394817/09353390.pdf
-A Quantile-based Approach for Hyperparameter Transfer Learning,David Salinas,"Bayesian optimization (BO) is a popular methodology to tune the hyperparameters of expensive black-box functions. Traditionally, BO focuses on a single task at a time and is not designed to leverage information from related functions, such as tuning performance objectives of the same algorithm across multiple datasets. In this work, we introduce a novel approach to achieve transfer learning across different datasets as well as different objectives. The main idea is to regress the mapping from hyperparameter to objective quantiles with a semi-parametric Gaussian Copula distribution, which provides robustness against different scales or outliers that can occur in different tasks. We introduce two methods to leverage this mapping: a Thompson sampling strategy as well as a Gaussian Copula process using such quantile estimate as a prior. We show that these strategies can combine the estimation of multiple objectives such as latency and accuracy, steering the hyperparameters optimization toward faster predictions for the same level of accuracy. Extensive experiments demonstrate significant improvements over state-of-the-art methods for both hyperparameter optimization and neural architecture search.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel approach to achieve transfer learning across different datasets as well as different objectives, to regress the mapping from hyperparameter to objective quantiles with a semi-parametric Gaussian Copula distribution.'}",
-Towards a Unified View on Visual Parameter-Efficient Transfer Learning,Bruce X. B. Yu,"Parameter efficient transfer learning (PETL) aims at making good use of the representation knowledge in the pre-trained large models by fine-tuning a small number of parameters. Recently, taking inspiration from the natural language processing (NLP) domain, popular PETL techniques such as prompt-tuning and Adapter have also been successfully applied to the vision domain. However, prefix-tuning remains under-explored for vision tasks. In this work, we intend to adapt large vision models (LVMs) to downstream tasks with a good parameter-accuracy trade-off. Towards this goal, we propose a framework with a unified view of PETL called visual-PETL (V-PETL) to investigate the effects of different PETL techniques, data scales of downstream domains, positions of trainable parameters, and other aspects affecting the trade-off. Specifically, we analyze the positional importance of trainable parameters and differences between NLP and vision tasks in terms of data structures and pre-training mechanisms while implementing various PETL techniques, especially for the under-explored prefix-tuning technique. Based on a comprehensive understanding of the differences between NLP and vision data, we propose a new variation of the prefix-tuning module called parallel attention (PATT) for vision downstream tasks. An extensive empirical analysis on vision tasks via different frozen LVMs has been carried and the findings show that the proposed PATT can effectively contribute to other PETL techniques. An effective scheme Swin-BAPAT derived from the proposed V-PETL framework achieves significantly better performance than the state-of-the-art AdaptFormer-Swin with slightly more parameters and outperforms full-tuning with far fewer parameters. Code and data are available at: https://github.com/bruceyo/V-PETL.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a framework with a unified view of PETL called visual- PETL (V-PETL) to investigate the effects of different PETL techniques, data scales of downstream domains, positions of trainable parameters, and other aspects affecting the trade-off, and proposes a new variation of the prefix-tuning module called parallel attention (PATT) for vision downstream tasks.'}",http://arxiv.org/pdf/2210.00788
-A CNN Based Transfer Learning Model for Automatic Activity Recognition from Accelerometer Sensors,B. Chikhaoui,,"{'model': 'tldr@v2.0.0', 'text': 'Deep learning based transfer learning algorithm based on convolutional neural networks (CNNs) that takes advantage of learned representations of activities of daily living from one dataset to recognize these activities in different other datasets characterized by different features including sensor modality, sampling rate, activity duration and environment is investigated.'}",http://r-libre.teluq.ca/1422/1/MLDM_2018_BChikhaoui_CRC.pdf
-A Vision-based Transfer Learning Approach for Recognizing Behavioral Symptoms in People with Dementia,Zachary Wharton,"With an aging population that continues to grow, dementia is a major global health concern. It is a syndrome in which there is a deterioration in memory, thinking, be-havior and the ability to perform activities of daily living. Depression and aggressive behavior are the most upsetting and challenging symptoms of dementia. Automatic recognition of these behaviors would not only be useful to alert family members and caregivers, but also helpful in planning and managing daily activities of people with dementia (PwD). In this work, we propose a vision-based approach that unifies transfer learning and deep convolutional neural network (CNN) for the effective recognition of behavioral symptoms. We also compare the performance of state-of-the-art CNN features with the hand-crafted HOG-feature, as well as their combination using a basic linear SVM. The proposed method is evaluated on a newly created dataset, which is based on the dementia storyline in ITVs Emmerdale episodes. The Alzheimer’s Society has described it as a ""realistic portrayal""1 of the condition to raise awareness of the issues surrounding dementia.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a vision-based approach that unifies transfer learning and deep convolutional neural network for the effective recognition of behavioral symptoms of dementia and compares the performance of state-of-the-art CNN features with the hand-crafted HOG-feature, as well as their combination using a basic linear SVM.'}",https://research.edgehill.ac.uk/ws/files/20125875/Vision-based-Behera.pdf
-Developmental Bayesian Optimization of Black-Box with Visual Similarity-Based Transfer Learning,Maxime Petit,"We present a developmental framework based on a long-term memory and reasoning mechanisms (Vision Similarity and Bayesian Optimisation). This architecture allows a robot to optimize autonomously hyper-parameters that need to be tuned from any action and/or vision module, treated as a black-box. The learning can take advantage of past experiences (stored in the episodic and procedural memories) in order to warm-start the exploration using a set of hyper-parameters previously optimized from objects similar to the new unknown one (stored in a semantic memory). As example, the system has been used to optimized 9 continuous hyper-parameters of a professional software (Kamido) both in simulation and with a real robot (industrial robotic arm Fanuc) with a total of 13 different objects. The robot is able to find a good object-specific optimization in 68 (simulation) or 40 (real) trials. In simulation, we demonstrate the benefit of the transfer learning based on visual similarity, as opposed to an amnesic learning (i.e. learning from scratch all the time). Moreover, with the real robot, we show that the method consistently outperforms the manual optimization from an expert with less than 2 hours of training time to achieve more than 88% of success.","{'model': 'tldr@v2.0.0', 'text': 'A developmental framework based on a long-term memory and reasoning mechanisms (Vision Similarity and Bayesian Optimisation) allows a robot to optimize autonomously hyper-parameters that need to be tuned from any action and/or vision module, treated as a black-box.'}",https://arxiv.org/pdf/1809.10141
-TTL: transformer-based two-phase transfer learning for cross-lingual news event detection,Hansi Hettiarachchi,,"{'model': 'tldr@v2.0.0', 'text': 'This research focuses on identifying event details at the sentence and token levels from news articles, considering their fine granularity, and proposes a novel learning strategy named Two-phase Transfer Learning (TTL), which allows the model to utilise the knowledge from a task at a particular data granularity for another task at different datagranularity.'}",https://link.springer.com/content/pdf/10.1007/s13042-023-01795-9.pdf
-An Approach Based on Transfer Learning to Lifetime Degradation Rate Prediction of the Dry-Type Transformer,Y. Li,"Lifetime prediction of the power transformer plays an important role in maintaining the stable operation of power equipment. However, due to the complexity of insulation structure degenerative process, the existing methods featuring high cost and low precision are not effective enough in transformer life time prediction. Meanwhile, how to effectively and promptly respond to a new prediction scenario of insufficient and limited data is a common challenge for all the data-driven prediction methods. To address these concerns, a prediction approach of a back adoptive adjustment transfer learning scheme (BAATL) is proposed for lifetime degradation prediction of the dry-type transformer. The power transformer condition monitoring data of Supervisory Control and Data Acquisition system is conducted as the data driven. A deep neural network, a transfer learning module and a back adjustment module are constructed to realize feature extraction, domain adaptation and prediction network optimization. The proposed scheme is able to improve prediction accuracy and resolves the problems and drawbacks of traditional prediction methods, and presents its superior portability and application potential in the case of data shortage and scenario change. With authentic datasets, simulation tests performed on the condition monitoring data of dry-type transformers prove the effectiveness of the proposed scheme.","{'model': 'tldr@v2.0.0', 'text': 'A prediction approach of a back adoptive adjustment transfer learning scheme (BAATL) is proposed for lifetime degradation prediction of the dry-type transformer, able to improve prediction accuracy and resolves the problems and drawbacks of traditional prediction methods.'}",
-Classification of Citrus Huanglongbing Degree Based on CBAM-MobileNetV2 and Transfer Learning,Shiqing Dou,"Citrus has become a pivotal industry for the rapid development of agriculture and increasing farmers’ incomes in the main production areas of southern China. Knowing how to diagnose and control citrus huanglongbing has always been a challenge for fruit farmers. To promptly recognize the diagnosis of citrus huanglongbing, a new classification model of citrus huanglongbing was established based on MobileNetV2 with a convolutional block attention module (CBAM-MobileNetV2) and transfer learning. First, the convolution features were extracted using convolution modules to capture high-level object-based information. Second, an attention module was utilized to capture interesting semantic information. Third, the convolution module and attention module were combined to fuse these two types of information. Last, a new fully connected layer and a softmax layer were established. The collected 751 citrus huanglongbing images, with sizes of 3648 × 2736, were divided into early, middle, and late leaf images with different disease degrees, and were enhanced to 6008 leaf images with sizes of 512 × 512, including 2360 early citrus huanglongbing images, 2024 middle citrus huanglongbing images, and 1624 late citrus huanglongbing images. In total, 80% and 20% of the collected citrus huanglongbing images were assigned to the training set and the test set, respectively. The effects of different transfer learning methods, different model training effects, and initial learning rates on model performance were analyzed. The results show that with the same model and initial learning rate, the transfer learning method of parameter fine tuning was obviously better than the transfer learning method of parameter freezing, and that the recognition accuracy of the test set improved by 1.02~13.6%. The recognition accuracy of the citrus huanglongbing image recognition model based on CBAM-MobileNetV2 and transfer learning was 98.75% at an initial learning rate of 0.001, and the loss value was 0.0748. The accuracy rates of the MobileNetV2, Xception, and InceptionV3 network models were 98.14%, 96.96%, and 97.55%, respectively, and the effect was not as significant as that of CBAM-MobileNetV2. Therefore, based on CBAM-MobileNetV2 and transfer learning, an image recognition model of citrus huanglongbing images with high recognition accuracy could be constructed.","{'model': 'tldr@v2.0.0', 'text': 'An image recognition model of citrus huanglongbing images with high recognition accuracy could be constructed based on CBAM-MobileNetV2 and transfer learning based on a convolutional block attention module and a new fully connected layer and a softmax layer.'}",https://www.mdpi.com/1424-8220/23/12/5587/pdf?version=1686750838
-Transfer learning-based quantized deep learning models for nail melanoma classification,Mujahid Hussain,,"{'model': 'tldr@v2.0.0', 'text': 'To facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow could realize and achieve more than 95% accuracy.'}",
-Detection of Bangladeshi-Produced Plant Disease Using a Transfer Learning Based on Deep Neural Model,Tareq Hasan,"Plant diseases pose a significant threat to agricultural productivity and food security in Bangladesh. In this research, we address the challenge of timely and accurate plant disease detection through the application of transfer learning with deep neural models. We curated a diverse dataset comprising 18 categories of plant leaf images, including Bell pepper Bacterial spot, Bell pepper Healthy, Peach Healthy, Potato Early Blight, Rice Leaf Blast, Rice Healthy, Rice Brown Spot, Potato Healthy, Peach Bacterial spot, Corn Blight, Potato Late blight, Corn Healthy, Tomato Bacterial spot, Strawberry Leaf Scorch, Tomato Early blight, Tomato Early blight, Strawberry Healthy, and Tomato Healthy. The dataset represents the most prevalent plant diseases observed in the Bangladeshi context. We employed three state-of-the-art deep learning algorithms, EfficientNetV2M, VGG-19, and NASNetLarge, to develop robust plant disease detection models. Through transfer learning, these pre-trained models were fine-tuned on our specialized dataset to adapt them for the task at hand. The performance evaluation revealed impressive results, with EfficientNetV2M achieving an accuracy rate of 99%, VGG-19 achieving 93%, and NASNetLarge attaining 83% accuracy. The high accuracy of EfficientNetV2M showcases its exceptional capability in accurately classifying plant diseases prevalent in Bangladesh. The success of these deep neural models in detecting various plant diseases signifies their potential in revolutionizing plant disease management and enhancing agricultural practices. Our research contributes valuable insights into the effective use of transfer learning for plant disease detection and emphasizes the significance of dataset curation for improved model performance. The developed models hold promise in providing timely and precise disease diagnosis to farmers and agricultural professionals, thereby facilitating prompt interventions and minimizing crop losses. Future research can explore the integration of these deep neural models into practical agricultural tools, enabling real-time disease detection and offering substantial benefits to the agricultural industry in Bangladesh.",,https://al-kindipublisher.com/index.php/jcsts/article/download/5887/4915
-Attention-based deep meta-transfer learning for few-shot fine-grained fault diagnosis,Chuanjiang Li,,,
-A novel method based on deep transfer unsupervised learning network for bearing fault diagnosis under variable working condition of unequal quantity,Hao Su,,,
-ECG Biometrics Method Based on Convolutional Neural Network and Transfer Learning,Yefei Zhang,"Personal identification based on ECG signals has been a significant challenge. The performance of an ECG authentication system depends significantly on the features extracted and the classifier subsequently applied. Although recently the deep neural networks based approaches featuring adaptive feature extractions and inherent classifications have attracted attention, they usually require a substantial set of training data. Aiming at tackling these issues, this paper presents a convolutional neural network-based transfer learning approach. It includes transferring the big data-trained GoogLeNet model into our identification task, fine-tuning the model using the ‘finetune’ idea, and adding three adaptive layers behind the original feature layer. The proposed approach not only requires a small set of training data, but also obtains great performance.","{'model': 'tldr@v2.0.0', 'text': 'A convolutional neural network-based transfer learning approach for personal identification based on ECG signals that includes transferring the big data-trained GoogLeNet model into the identification task, fine-tuning the model using the ‘finetune’ idea, and adding three adaptive layers behind the original feature layer.'}",
-Facilitating EFL learners’ active behaviors in speaking: a progressive question prompt-based peer-tutoring approach with VR contexts,Chin-Yu Chen,"ABSTRACT Speaking is a challenge for most English learners owing to the lack of practicing environments and partners. The virtual reality (VR) technology has changed this situation by providing learners with rich opportunities for practice. In VR-based practicing environments, peer tutoring is a common strategy for promoting peer assistance. However, in conventional VR-based learning, students may merely follow the guidelines and content provided by the teacher without thinking about the meaning and reasons for the practice in the situated contexts, which generally leads to poor learning outcomes. To solve this problem, a progressive question prompt-based peer-tutoring approach in VR contexts (PQP-PTVR) is proposed to assist students’ English-speaking development. An experiment was conducted to evaluate the effectiveness of the proposed approach. The experimental group used the proposed approach, while the control group learned with the conventional question prompt-based peer-tutoring approach in VR contexts (C-PTVR). The results showed that the experimental group achieved significantly greater improvement in speaking and self-efficacy than the control group. Concerning the participants’ learning behaviors, the experimental group had more interactions and made more attempts to improve their speaking. This study contributes to the existing literature by proposing the PQP-PTVR approach to enhancing learners’ speaking development, improving their interactions and self-efficacy.",,
-Betel Nut Classification Method Based on Transfer Learning,Hengquan Cai,"Betel nut is one of the main chewing hobby in the world. The traditional methods of betel nut classification are mostly based on handcrafted features, and the classification stability is poor. To solve this problem and reduce labor costs for manual extraction of image features, this paper applies convolutional neural network to betel nut, and adopts parameter transfer strategy to realize transfer learning. The dataset of betel nut was constructed by shooting different types of betel nut with betel nut grading equipment. Aiming at the problem of insufficient samples, some parameters of AlexNet are transferred and fixed, new fully connected layers are build. Make full use of betel dataset to train and test the network. The experimental results indicate the proposed method is effective and low labor cost.","{'model': 'tldr@v2.0.0', 'text': 'This paper applies convolutional neural network to betel nut, and adopts parameter transfer strategy to realize transfer learning and experimental results indicate the proposed method is effective and low labor cost.'}",
-Facial Expression Based Imagination Index and a Transfer Learning Approach to Detect Deception,M. Hasan,"In this paper, we introduce a framework to automatically distinguish between facial expression sequences associated with imagining vs. remembering while answering a question. Our experiment includes a baseline and relevant questioning technique in the context of deception with 220 participants (20 hours long). Baseline questioning includes participants being separately asked to remember and imagine an arbitrary experience. During the relevant questioning, participants were prompted to either lie or tell the truth about a certain task. We trained a neural network model on the baseline data and achieved an accuracy of 60% on classifying imagining vs. remembering, whereas human performance for this task is 51%. Relevant questioning included a set of questions, each of which became an independent response segment. Using a transfer learning approach, we use the pretrained model from the baseline to obtain an imagination probability score for each relevant response segment. We define this individual probability per response as the Imagination Index. We apply the imagination indices as a feature vector to classify the whole relevant section as truth vs. bluff with an accuracy of 70%, significantly outperforming the human performance of 52%.","{'model': 'tldr@v2.0.0', 'text': 'A framework to automatically distinguish between facial expression sequences associated with imagining vs. remembering while answering a question is introduced and the pretrained model from the baseline is used to obtain an imagination probability score for each relevant response segment.'}",
-3D transfer learning network for classification of Alzheimer’s disease with MRI,Haifeng Wu,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed three-dimensional (3D) transfer network is to realize the classification of 3D MRI data through an existing 2D transfer network, and it not only reduces the complexity of conventional 3D networks, but also improves the classification accuracy.'}",
-On the Role of Neural Collapse in Transfer Learning,Tomer Galanti,"We study the ability of foundation models to learn representations for classification that are transferable to new, unseen classes. Recent results in the literature show that representations learned by a single classifier over many classes are competitive on few-shot learning problems with representations learned by special-purpose algorithms designed for such problems. In this paper we provide an explanation for this behavior based on the recently observed phenomenon that the features learned by overparameterized classification networks show an interesting clustering property, called neural collapse. We demonstrate both theoretically and empirically that neural collapse generalizes to new samples from the training classes, and -- more importantly -- to new classes as well, allowing foundation models to provide feature maps that work well in transfer learning and, specifically, in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated both theoretically and empirically that neural collapse generalizes to new samples from the training classes, and -- more importantly -- to new classes as well, allowing foundation models to provide feature maps that work well in transfer learning and, specifically, in the few-shot setting.'}",
-Protein Ubiquitylation and Sumoylation Site Prediction Based on Ensemble and Transfer Learning,Fei He,"Ubiquitylation, a typical post-translational modification (PTM), plays an important role in signal transduction, apoptosis and cell proliferation. A ubiquitylation like PTM, sumoylation also may affect gene mapping, expression and genomic replication. Over the past two decades, machine learning has been widely employed in protein ubiquitylation and sumoylation site prediction tools. These existing tools require feature engineering, but failed to provide general interpretable features and probably underutilized the growing amount of data. This prompted us to propose a deep learning-based model that integrates multiple convolution and fully-connected layers of seven supervised learning sub-models to extract deep representations from protein sequences and physico-chemical properties (PCPs). Especially, we divided PCPs into 6 clusters and customized deep networks accordingly for handling the high correlations among one cluster. A stacking ensemble strategy was applied to combine these deep representations to make prediction. Furthermore, with the advantage of transfer learning, our deep learning model can work well on protein sumoylation site prediction as well after fine-tuning. On the high-quality annotated database Swiss-Prot, our model outperformed several well-known ubiquitylation and sumoylation site prediction tools. Our code is freely available at https://github.com/ruiwcoding/DeepUbiSumoPre.","{'model': 'tldr@v2.0.0', 'text': 'A deep learning-based model that integrates multiple convolution and fully-connected layers of seven supervised learning sub-models to extract deep representations from protein sequences and physico-chemical properties (PCPs) and outperformed several well-known ubiquitylation and sumoylation site prediction tools.'}",
-PAC-Net: A Model Pruning Approach to Inductive Transfer Learning,Sanghoon Myung,"Inductive transfer learning aims to learn from a small amount of training data for the target task by utilizing a pre-trained model from the source task. Most strategies that involve large-scale deep learning models adopt initialization with the pre-trained model and fine-tuning for the target task. However, when using over-parameterized models, we can often prune the model without sacrificing the accuracy of the source task. This motivates us to adopt model pruning for transfer learning with deep learning models. In this paper, we propose PAC-Net, a simple yet effective approach for transfer learning based on pruning. PAC-Net consists of three steps: Prune, Allocate, and Calibrate (PAC). The main idea behind these steps is to identify essential weights for the source task, fine-tune on the source task by updating the essential weights, and then calibrate on the target task by updating the remaining redundant weights. Under the various and extensive set of inductive transfer learning experiments, we show that our method achieves state-of-the-art performance by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'PAC-Net is proposed, a simple yet effective approach for transfer learning based on pruning that achieves state-of-the-art performance by a large margin under the various and extensive set of inductive transfer learning experiments.'}",http://arxiv.org/pdf/2206.05703
-Transfer Learning In Differential Privacy's Hybrid-Model,Reʾuven Kohen,"The hybrid-model (Avent et al 2017) in Differential Privacy is a an augmentation of the local-model where in addition to N local-agents we are assisted by one special agent who is in fact a curator holding the sensitive details of n additional individuals. Here we study the problem of machine learning in the hybrid-model where the n individuals in the curators dataset are drawn from a different distribution than the one of the general population (the local-agents). We give a general scheme -- Subsample-Test-Reweigh -- for this transfer learning problem, which reduces any curator-model DP-learner to a hybrid-model learner in this setting using iterative subsampling and reweighing of the n examples held by the curator based on a smooth variation of the Multiplicative-Weights algorithm (introduced by Bun et al, 2020). Our scheme has a sample complexity which relies on the chi-squared divergence between the two distributions. We give worst-case analysis bounds on the sample complexity required for our private reduction. Aiming to reduce said sample complexity, we give two specific instances our sample complexity can be drastically reduced (one instance is analyzed mathematically, while the other - empirically) and pose several directions for follow-up work.","{'model': 'tldr@v2.0.0', 'text': 'A general scheme is given -- Subsample-Test-Reweigh -- for this transfer learning problem, which reduces any curator-model DP-learner to a hybrid-model learner in this setting using iterative subsampling and reweighing of the n examples held by the curator based on a smooth variation of the Multiplicative-Weights algorithm.'}",
-Virtual Battery Parameter Identification Using Transfer Learning Based Stacked Autoencoder,Indrasis Chakraborty,"Recent studies have shown that the aggregated dynamic flexibility of an ensemble of thermostatic loads can be modeled in the form of a virtual battery. The existing methods for computing the virtual battery parameters require the knowledge of the first-principle models and parameter values of the loads in the ensemble. In real-world applications, however, it is likely that the only available information are end-use measurements such as power consumption, room temperature, device on/off status, etc., while very little about the individual load models and parameters are known. We propose a transfer learning based deep network framework for calculating virtual battery state of a given ensemble of flexible thermostatic loads, from the available end-use measurements. This proposed framework extracts first order virtual battery model parameters for the given ensemble. We illustrate the effectiveness of this novel framework on different ensembles of ACs and WHs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transfer learning based deep network framework for calculating virtual battery state of a given ensemble of flexible thermostatic loads, from the available end-use measurements and illustrates the effectiveness of this novel framework on different ensembles of ACs and WHs.'}",https://arxiv.org/pdf/1810.04642
-Transfer Learning Based Approach for Semantic Person Retrieval,Takuya Yaguchi,Many algorithms for semantic person retrieval suffer from a lack of training data often due to the difficulties in constructing a large dataset. We therefore propose a transfer learning based approach for semantic person identification and semantic person search. We apply the fine-tuned Mask R-CNN and DenseNet-161 for detection and attribute classification. The networks were pre-trained on the MS COCO and ILSVRC 2012 datasets. Our proposed approach achieves the highest recognition rate at each rank of CMC curve for semantic person identification and the highest average localization precision for semantic person search on our validation dataset.,"{'model': 'tldr@v2.0.0', 'text': ""This work proposes a transfer learning based approach for semantic person identification and semantic person search that achieves the highest recognition rate at each rank of CMC curve for semantic people identification and the highest average localization precision for semanticperson search on the authors' validation dataset.""}",https://eprints.soton.ac.uk/426371/1/CameraReady.pdf
-Higher education students’ motivation to transfer learning: a scoping review,Doreen Bredenkamp,"PurposeThere is a need for higher education to produce graduates who are motivated to transfer learning into the workplace. Motivated graduates are work-ready and associated with increased performance. Presently, the research field around motivation to transfer learning by students in higher education is not clear and is inconsistent.Design/methodology/approachThis scoping review provides an overview of the characteristics of the literature, including key concepts, recommendations and gaps based on eight published articles on the motivation of students in higher education to transfer learning.FindingsThe results reflected a research field, which focused primarily on the influence of specific factors, namely student characteristics, educational design, the workplace environment, and on higher education students' motivation to transfer learning. The lack of a shared conceptual definition of motivation to transfer learning in higher education appears to influence the description of the results from the included studies. Most of the previous studies applied rigorous research designs.Originality/valueThis seemingly stunted research field related to higher education students' motivation to transfer learning needs to be amplified to influence the development of work-ready graduates from higher education. Approaches towards including all elements of motivation, expanding to other fields in higher education, including low-income countries, may be a proximal step in enhancing the trajectory of this research field.",,https://www.emerald.com/insight/content/doi/10.1108/HESWBL-03-2022-0057/full/pdf?title=higher-education-students-motivation-to-transfer-learning-a-scoping-review
-Oryza Sativa Leaf Disease Detection using Transfer Learning,A. Musthafa,"Oryza sativa (Rice) is the world's most significant cereal harvest. It is taken as a staple feast for energy by the greater part of the total populace. Abiotic and biotic components like precipitation, soil richness, temperature, bugs, microscopic organisms, infections, etc. impact the yield creation amount and nature of rice grain. Ranchers contribute a great deal of time and energy to infection prevention, and they recognize sicknesses with their devastated unaided eye technique, which prompts unfortunate cultivating. The advancement of horticultural innovation helps significantly supports the computerized location of pathogenic living beings in the leaves of rice plants. The convolutional-based neural network calculation (CNN) is the one of very profound calculations that has been effectively used to settle PC vision issues like picture grouping, object division, picture investigation, etc. The proposed model boundaries have been tuned for the order work, and it has a great exactness of 95.67 percent. Using the transfer learning the data are trained faster andit can learn and apply the learned things in the next dataset faster. So that it does not acquire time in learning, which is not in the existing process.","{'model': 'tldr@v2.0.0', 'text': 'The convolutional-based neural network calculation (CNN) is the one of very profound calculations that has been effectively used to settle PC vision issues like picture grouping, object division, picture investigation, etc.'}",
-TRCLA: A Transfer Learning Approach to Reduce Negative Transfer for Cellular Learning Automata,Seyyed Amir Hadi Minoofam,"In most traditional machine learning algorithms, the training and testing datasets have identical distributions and feature spaces. However, these assumptions have not held in many real applications. Although transfer learning methods have been invented to fill this gap, they introduce new challenges as negative transfers (NTs). Most previous research considered NT a significant problem, but they pay less attention to solving it. This study will propose a transductive learning algorithm based on cellular learning automata (CLA) to alleviate the NT issue. Two famous learning automata (LA) entitled estimators are applied as estimator CLA in the proposed algorithms. A couple of new decision criteria called merit and and attitude parameters are introduced to CLA to limit NT. The proposed algorithms are applied to standard LA environments. The experiments show that the proposed algorithm leads to higher accuracy and less NT results.","{'model': 'tldr@v2.0.0', 'text': 'This study will propose a transductive learning algorithm based on cellular learning automata (CLA) to alleviate the NT issue and shows that the proposed algorithm leads to higher accuracy and less NT results.'}",
-A Survey on Deep Transfer Learning,Chuanqi Tan,,"{'model': 'tldr@v2.0.0', 'text': 'This survey focuses on reviewing the current researches of transfer learning by using deep neural network and its applications and defined deep transfer learning, category and review the recent research works based on the techniques used inDeep transfer learning.'}",https://arxiv.org/pdf/1808.01974
-Covid-19: automatic detection from X-ray images utilizing transfer learning with convolutional neural networks,Ioannis D. Apostolopoulos,,"{'model': 'tldr@v2.0.0', 'text': 'The results suggest that Deep Learning with X-ray imaging may extract significant biomarkers related to the Covid-19 disease, while the best accuracy, sensitivity, and specificity obtained is 96.78%, 98.66%, and 96.46% respectively.'}",https://link.springer.com/content/pdf/10.1007/s13246-020-00865-4.pdf
-Enhancing Cross-lingual Natural Language Inference by Prompt-learning from Cross-lingual Templates,Kunxun Qi,"Cross-lingual natural language inference (XNLI) is a fundamental task in cross-lingual natural language understanding. Recently this task is commonly addressed by pre-trained cross-lingual language models. Existing methods usually enhance pre-trained language models with additional data, such as annotated parallel corpora. These additional data, however, are rare in practice, especially for low-resource languages. Inspired by recent promising results achieved by prompt-learning, this paper proposes a novel prompt-learning based framework for enhancing XNLI. It reformulates the XNLI problem to a masked language modeling problem by constructing cloze-style questions through cross-lingual templates. To enforce correspondence between different languages, the framework augments a new question for every question using a sampled template in another language and then introduces a consistency loss to make the answer probability distribution obtained from the new question as similar as possible with the corresponding distribution obtained from the original question. Experimental results on two benchmark datasets demonstrate that XNLI models enhanced by our proposed framework significantly outperform original ones under both the full-shot and few-shot cross-lingual transfer settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on two benchmark datasets demonstrate that XNLI models enhanced by the proposed framework significantly outperform original ones under both the full-shot and few-shot cross-lingual transfer settings.'}",https://aclanthology.org/2022.acl-long.134.pdf
-Evolutionary based ensemble framework for realizing transfer learning in HIV-1 Protease cleavage sites prediction,Deepak Singh,,"{'model': 'tldr@v2.0.0', 'text': 'A method for combining multiple weighted classifiers optimally by incorporating the knowledge derived from various amino acid encoding techniques is proposed, and the optimally coupled sequence of features and classifiers that characterized the heterogeneous feature is achieved promptly by genetic algorithm.'}",
-Transfer Learning on Multi-Fidelity Data,Dong H. Song,"Neural networks (NNs) are often used as surrogates or emulators of partial differential equations (PDEs) that describe the dynamics of complex systems. A virtually negligible computational cost of such surrogates renders them an attractive tool for ensemble-based computation, which requires a large number of repeated PDE solves. Since the latter are also needed to generate sufficient data for NN training, the usefulness of NN-based surrogates hinges on the balance between the training cost and the computational gain stemming from their deployment. We rely on multi-fidelity simulations to reduce the cost of data generation for subsequent training of a deep convolutional NN (CNN) using transfer learning. High- and low-fidelity images are generated by solving PDEs on fine and coarse meshes, respectively. We use theoretical results for multilevel Monte Carlo to guide our choice of the numbers of images of each kind. We demonstrate the performance of this multi-fidelity training strategy on the problem of estimation of the distribution of a quantity of interest, whose dynamics is governed by a system of nonlinear PDEs (parabolic PDEs of multi-phase flow in heterogeneous porous media) with uncertain/random parameters. Our numerical experiments demonstrate that a mixture of a comparatively large number of low-fidelity data and smaller numbers of high- and low-fidelity data provides an optimal balance of computational speed-up and prediction accuracy. The former is reported relative to both CNN training on high-fidelity images only and Monte Carlo solution of the PDEs. The latter is expressed in terms of both the Wasserstein distance and the Kullback-Leibler divergence.","{'model': 'tldr@v2.0.0', 'text': 'Multi-fidelity simulations are relied on to reduce the cost of data generation for subsequent training of a deep convolutional NN (CNN) using transfer learning and demonstrate the performance of this multi-f fidelity training strategy on the problem of estimation of the distribution of a quantity of interest.'}",https://www.dl.begellhouse.com/download/article/1171f2446ae04ad7/JMLMC0301(2)-38925.pdf
-QBox: Partial Transfer Learning With Active Querying for Object Detection,Ying-Peng Tang,"Object detection requires plentiful data annotated with bounding boxes for model training. However, in many applications, it is difficult or even impossible to acquire a large set of labeled examples for the target task due to the privacy concern or lack of reliable annotators. On the other hand, due to the high-quality image search engines, such as Flickr and Google, it is relatively easy to obtain resource-rich unlabeled datasets, whose categories are a superset of those of target data. In this article, to improve the target model with cost-effective supervision from source data, we propose a partial transfer learning approach QBox to actively query labels for bounding boxes of source images. Specifically, we design two criteria, i.e., informativeness and transferability, to measure the potential utility of a bounding box for improving the target model. Based on these criteria, QBox actively queries the labels of the most useful boxes from the source domain and, thus, requires fewer training examples to save the labeling cost. Furthermore, the proposed query strategy allows annotators to simply labeling a specific region, instead of the whole image, and, thus, significantly reduces the labeling difficulty. Extensive experiments are performed on various partial transfer benchmarks and a real COVID-19 detection task. The results validate that QBox improves the detection accuracy with lower labeling cost compared to state-of-the-art query strategies for object detection.","{'model': 'tldr@v2.0.0', 'text': 'This article designs two criteria, i.e., informativeness and transferability, to measure the potential utility of a bounding box for improving the target model, and proposes a partial transfer learning approach QBox to actively query labels for bounding boxes of source images.'}",
-A Deep Transfer Learning Approach to Modeling Teacher Discourse in the Classroom,Emily Jensen,"Teachers, like everyone else, need objective reliable feedback in order to improve their effectiveness. However, developing a system for automated teacher feedback entails many decisions regarding data collection procedures, automated analysis, and presentation of feedback for reflection. We address the latter two questions by comparing two different machine learning approaches to automatically model seven features of teacher discourse (e.g., use of questions, elaborated evaluations). We compared a traditional open-vocabulary approach using n-grams and Random Forest classifiers with a state-of-the-art deep transfer learning approach for natural language processing (BERT). We found a tradeoff between data quantity and accuracy, where deep models had an advantage on larger datasets, but not for smaller datasets, particularly for variables with low incidence rates. We also compared the models based on the level of feedback granularity: utterance-level (e.g., whether an utterance is a question or a statement), class session-level proportions by averaging across utterances (e.g., question incidence score of 48%), and session-level ordinal feedback based on pre-determined thresholds (e.g., question asking score is medium [vs. low or high]) and found that BERT generally provided more accurate feedback at all levels of granularity. Thus, BERT appears to be the most viable approach to providing automatic feedback on teacher discourse provided there is sufficient data to fine tune the model.","{'model': 'tldr@v2.0.0', 'text': 'A traditional open-vocabulary approach using n-grams and Random Forest classifiers with a state-of-the-art deep transfer learning approach for natural language processing (BERT) is compared and it is found that BERT generally provided more accurate feedback at all levels of granularity.'}",https://dl.acm.org/doi/pdf/10.1145/3448139.3448168
-Discriminative Fisher Embedding Dictionary Transfer Learning for Object Recognition,Zizhu Fan,"In transfer learning model, the source domain samples and target domain samples usually share the same class labels but have different distributions. In general, the existing transfer learning algorithms ignore the interclass differences and intraclass similarities across domains. To address these problems, this article proposes a transfer learning algorithm based on discriminative Fisher embedding and adaptive maximum mean discrepancy (AMMD) constraints, called discriminative Fisher embedding dictionary transfer learning (DFEDTL). First, combining the label information of source domain and part of target domain, we construct the discriminative Fisher embedding model to preserve the interclass differences and intraclass similarities of training samples in transfer learning. Second, an AMMD model is constructed using atoms and profiles, which can adaptively minimize the distribution differences between source domain and target domain. The proposed method has three advantages: 1) using the Fisher criterion, we construct the discriminative Fisher embedding model between source domain samples and target domain samples, which encourages the samples from the same class to have similar coding coefficients; 2) instead of using the training samples to design the maximum mean discrepancy (MMD), we construct the AMMD model based on the relationship between the dictionary atoms and profiles; thus, the source domain samples can be adaptive to the target domain samples; and 3) the dictionary learning is based on the combination of source and target samples which can avoid the classification error caused by the difference among samples and reduce the tedious and expensive data annotation. A large number of experiments on five public image classification datasets show that the proposed method obtains better classification performance than some state-of-the-art dictionary and transfer learning methods. The code has been available at https://github.com/shilinrui/DFEDTL.","{'model': 'tldr@v2.0.0', 'text': 'A large number of experiments on five public image classification datasets show that the proposed method obtains better classification performance than some state-of-the-art dictionary and transfer learning methods.'}",
-"""Train one, Classify one, Teach one"" - Cross-surgery transfer learning for surgical step recognition",Daniel Neimark,"Prior work demonstrated the ability of machine learning to automatically recognize surgical workflow steps from videos. However, these studies focused on only a single type of procedure. In this work, we analyze, for the first time, surgical step recognition on four different laparoscopic surgeries: Cholecystectomy, Right Hemicolectomy, Sleeve Gastrectomy, and Appendectomy. Inspired by the traditional apprenticeship model, in which surgical training is based on the Halstedian method, we paraphrase the""see one, do one, teach one""approach for the surgical intelligence domain as""train one, classify one, teach one"". In machine learning, this approach is often referred to as transfer learning. To analyze the impact of transfer learning across different laparoscopic procedures, we explore various time-series architectures and examine their performance on each target domain. We introduce a new architecture, the Time-Series Adaptation Network (TSAN), an architecture optimized for transfer learning of surgical step recognition, and we show how TSAN can be pre-trained using self-supervised learning on a Sequence Sorting task. Such pre-training enables TSAN to learn workflow steps of a new laparoscopic procedure type from only a small number of labeled samples from the target procedure. Our proposed architecture leads to better performance compared to other possible architectures, reaching over 90% accuracy when transferring from laparoscopic Cholecystectomy to the other three procedure types.","{'model': 'tldr@v2.0.0', 'text': 'This work analyzes, for the first time, surgical step recognition on four different laparoscopic surgeries: Cholecystectomy, Right Hemicolectomy, Sleeve Gastrectomy, and Appendectomy, and introduces a new architecture, the Time-Series Adaptation Network (TSAN), an architecture optimized for transfer learning of surgicalStep recognition.'}",
-Deep Learning of Brain Asymmetry Images and Transfer Learning for Early Diagnosis of Dementia,Nitsa J. Herzog,,"{'model': 'tldr@v2.0.0', 'text': 'Results from 300 independent simulation runs illustrate that transfer learning of CNN-based models equipped with SVM output layer is capable to produce better performing models within a few training epochs compared to commonly used transfer learning architectures that combine CNN pretrained models with fully connected Softmax layers.'}",https://eprints.bbk.ac.uk/id/eprint/44156/1/EANN21-NH-GM-camvers.pdf
-NLPBench: Evaluating Large Language Models on Solving NLP Problems,Linxin Song,"Recent developments in large language models (LLMs) have shown promise in enhancing the capabilities of natural language processing (NLP). Despite these successes, there remains a dearth of research dedicated to the NLP problem-solving abilities of LLMs. To fill the gap in this area, we present a unique benchmarking dataset, NLPBench, comprising 378 college-level NLP questions spanning various NLP topics sourced from Yale University's prior final exams. NLPBench includes questions with context, in which multiple sub-questions share the same public information, and diverse question types, including multiple choice, short answer, and math. Our evaluation, centered on LLMs such as GPT-3.5/4, PaLM-2, and LLAMA-2, incorporates advanced prompting strategies like the chain-of-thought (CoT) and tree-of-thought (ToT). Our study reveals that the effectiveness of the advanced prompting strategies can be inconsistent, occasionally damaging LLM performance, especially in smaller models like the LLAMA-2 (13b). Furthermore, our manual assessment illuminated specific shortcomings in LLMs' scientific problem-solving skills, with weaknesses in logical decomposition and reasoning notably affecting results.","{'model': 'tldr@v2.0.0', 'text': ""This study presents a unique benchmarking dataset, NLPBench, comprising 378 college-level NLP questions spanning various NLP topics sourced from Yale University's prior final exams, and reveals that the effectiveness of the advanced prompting strategies can be inconsistent, occasionally damaging LLM performance.""}",https://arxiv.org/pdf/2309.15630
-RETUYT-InCo at BEA 2023 Shared Task: Tuning Open-Source LLMs for Generating Teacher Responses,Alexis Baladn,"This paper presents the results of our participation in the BEA 2023 shared task, which focuses on generating AI teacher responses in educational dialogues. We conducted experiments using several Open-Source Large Language Models (LLMs) and explored fine-tuning techniques along with prompting strategies, including Few-Shot and Chain-of-Thought approaches. Our best model was ranked 4.5 in the competition with a BertScore F1 of 0.71 and a DialogRPT final (avg) of 0.35. Nevertheless, our internal results did not exactly correlate with those obtained in the competition, which showed the difficulty in evaluating this task. Other challenges we faced were data leakage on the train set and the irregular format of the conversations.",,https://aclanthology.org/2023.bea-1.61.pdf
-Are Large Language Models Post Hoc Explainers?,Nicholas Kroeger,"Large Language Models (LLMs) are increasingly used as powerful tools for a plethora of natural language processing (NLP) applications. A recent innovation, in-context learning (ICL), enables LLMs to learn new tasks by supplying a few examples in the prompt during inference time, thereby eliminating the need for model fine-tuning. While LLMs have been utilized in several applications, their applicability in explaining the behavior of other models remains relatively unexplored. Despite the growing number of new explanation techniques, many require white-box access to the model and/or are computationally expensive, highlighting a need for next-generation post hoc explainers. In this work, we present the first framework to study the effectiveness of LLMs in explaining other predictive models. More specifically, we propose a novel framework encompassing multiple prompting strategies: i) Perturbation-based ICL, ii) Prediction-based ICL, iii) Instruction-based ICL, and iv) Explanation-based ICL, with varying levels of information about the underlying ML model and the local neighborhood of the test sample. We conduct extensive experiments with real-world benchmark datasets to demonstrate that LLM-generated explanations perform on par with state-of-the-art post hoc explainers using their ability to leverage ICL examples and their internal knowledge in generating model explanations. On average, across four datasets and two ML models, we observe that LLMs identify the most important feature with 72.19% accuracy, opening up new frontiers in explainable artificial intelligence (XAI) to explore LLM-based explanation frameworks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework encompassing multiple prompting strategies and conducts extensive experiments to demonstrate that LLM-generated explanations perform on par with state-of-the-art post hoc explainers using their ability to leverage ICL examples and their internal knowledge in generating model explanations.'}",https://arxiv.org/pdf/2310.05797
-Aligning Large Language Models for Clinical Tasks,Supun Manathunga,"Large Language Models (LLMs) have demonstrated remarkable adaptability, showcasing their capacity to excel in tasks for which they were not explicitly trained. However, despite their impressive natural language processing (NLP) capabilities, effective alignment of LLMs remains a crucial challenge when deploying them for specific clinical applications. The ability to generate responses with factually accurate content and to engage in non-trivial reasoning steps are crucial for the LLMs to be eligible for applications in clinical medicine. Employing a combination of techniques including instruction-tuning and in-prompt strategies like few-shot and chain-of-thought prompting has significantly enhanced the performance of LLMs. Our proposed alignment strategy for medical question-answering, known as 'expand-guess-refine', offers a parameter and data-efficient solution. A preliminary analysis of this method demonstrated outstanding performance, achieving a score of 70.63% on a subset of questions sourced from the USMLE dataset.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes a proposed alignment strategy for medical question-answering, known as 'expand-guess-refine', which offers a parameter and data-efficient solution.""}",https://arxiv.org/pdf/2309.02884
-NAISTeacher: A Prompt and Rerank Approach to Generating Teacher Utterances in Educational Dialogues,Justin Vasselli,"This paper presents our approach to the BEA 2023 shared task of generating teacher responses in educational dialogues, using the Teacher-Student Chatroom Corpus. Our system prompts GPT-3.5-turbo to generate initial suggestions, which are then subjected to reranking. We explore multiple strategies for candidate generation, including prompting for multiple candidates and employing iterative few-shot prompts with negative examples. We aggregate all candidate responses and rerank them based on DialogRPT scores. To handle consecutive turns in the dialogue data, we divide the task of generating teacher utterances into two components: teacher replies to the student and teacher continuations of previously sent messages. Through our proposed methodology, our system achieved the top score on both automated metrics and human evaluation, surpassing the reference human teachers on the latter.",,https://aclanthology.org/2023.bea-1.63.pdf
-Challenges and Strategies in Cross-Cultural NLP,Daniel Hershcovich,"Various efforts in the Natural Language Processing (NLP) community have been made to accommodate linguistic diversity and serve speakers of many different languages. However, it is important to acknowledge that speakers and the content they produce and require, vary not just by language, but also by culture. Although language and culture are tightly linked, there are important differences. Analogous to cross-lingual and multilingual NLP, cross-cultural and multicultural NLP considers these differences in order to better serve users of NLP systems. We propose a principled framework to frame these efforts, and survey existing and potential strategies.","{'model': 'tldr@v2.0.0', 'text': 'A principled framework to frame efforts to accommodate linguistic diversity and serve speakers of many different languages is proposed, and existing and potential strategies are surveyed.'}",http://arxiv.org/pdf/2203.10020
-An Empirical Study of Tokenization Strategies for Various Korean NLP Tasks,Kyubyong Park,"Typically, tokenization is the very first step in most text processing works. As a token serves as an atomic unit that embeds the contextual information of text, how to define a token plays a decisive role in the performance of a model. Even though Byte Pair Encoding (BPE) has been considered the de facto standard tokenization method due to its simplicity and universality, it still remains unclear whether BPE works best across all languages and tasks. In this paper, we test several tokenization strategies in order to answer our primary research question, that is, “What is the best tokenization strategy for Korean NLP tasks?” Experimental results demonstrate that a hybrid approach of morphological segmentation followed by BPE works best in Korean to/from English machine translation and natural language understanding tasks such as KorNLI, KorSTS, NSMC, and PAWS-X. As an exception, for KorQuAD, the Korean extension of SQuAD, BPE segmentation turns out to be the most effective. Our code and pre-trained models are publicly available at https://github.com/kakaobrain/kortok.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that a hybrid approach of morphological segmentation followed by BPE works best in Korean to/from English machine translation and natural language understanding tasks such as KorNLI, KorSTS, NSMC, and PAWS-X.'}",
-Evaluating Adaptive Pedagogical Agents' Prompting Strategies Effect on Students' Emotions,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Examination of the impact of different prompting strategies on self-reported agent-directed emotions in an ITS that scaffolds students’ use of self-regulated learning (SRL) strategies, taking into account students' prior knowledge indicates that more intense initial prompting can indeed lead to increased frustration, and sometimes boredom even toward pedagogical agents that are perceived as competent.""}",https://hal.archives-ouvertes.fr/hal-02015693/file/Bouchet%20et%20al.%20-%202018%20-%20Evaluating%20adaptive%20pedagogical%20agents%E2%80%99%20prompting%20.pdf
-What do governments plan in the field of artificial intelligence?: Analysing national AI strategies using NLP,T. Papadopoulos,"The primary goal of this paper is to explore how Natural Language Processing techniques (NLP) can assist in reviewing, understanding, and drawing conclusions from text datasets. We explore NLP techniques for the analysis and the extraction of useful information from the text of twelve national strategies on artificial intelligence (AI). For this purpose, we are using a set of machine learning algorithms in order to (a) extract the most significant keywords and summarize each strategy document, (b) discover and assign topics to each document, and (c) cluster the strategies based on their pair-wise similarity. Using the results of the analysis, we discuss the findings and highlight critical issues that emerge from the national strategies for artificial intelligence, such as the importance of the data ecosystem for the development of AI, the increasing considerations about ethical and safety issues, as well as the growing ambition of many countries to lead in the AI race. Utilizing the LDA topic model, we were able to reveal the distributions of thematic sub-topics among the strategic documents. The topic modelling distributions were then used along with other document similarity measures as an input for the clustering of the strategic documents into groups. The results revealed three clusters of countries with a visible differentiation between the strategies of China and Japan on the one hand and the Scandinavian strategies (plus the German and the Luxemburgish) one on the other. The former promote technology and innovation-driven development plans in order to integrate AI with the economy, while the latter share a common view regarding the role of the public sector both as a promoter and investor but also as a user and beneficiary of AI, and give a higher priority to the ethical & safety issues that are connected to the development of AI.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores NLP techniques for the analysis and the extraction of useful information from the text of twelve national strategies on artificial intelligence, using a set of machine learning algorithms to extract the most significant keywords and summarize each strategy document, and discover and assign topics to each document.'}",
-Visual Prompting via Image Inpainting,Amir Bar,"How does one adapt a pre-trained visual model to novel downstream tasks without task-specific finetuning or any model modification? Inspired by prompting in NLP, this paper investigates visual prompting: given input-output image example(s) of a new task at test time and a new input image, the goal is to automatically produce the output image, consistent with the given examples. We show that posing this problem as simple image inpainting - literally just filling in a hole in a concatenated visual prompt image - turns out to be surprisingly effective, provided that the inpainting algorithm has been trained on the right data. We train masked auto-encoders on a new dataset that we curated - 88k unlabeled figures from academic papers sources on Arxiv. We apply visual prompting to these pretrained models and demonstrate results on various downstream image-to-image tasks, including foreground segmentation, single object detection, colorization, edge detection, etc.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates visual prompting: given input-output image example(s) of a new task at test time and a new input image, the goal is to automatically produce the output image, consistent with the given examples, and shows that posing this problem as simple image inpainting turns out to be surprisingly effective.'}",http://arxiv.org/pdf/2209.00647
-Can Adaptive Pedagogical Agents' Prompting Strategies Improve Students' Learning and Self-Regulation?,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Results indicated that an initially more frequent but progressively fading prompting strategy is beneficial to learners' deployment of SRL processes once the scaffolding is faded, and has no negative impact on learners' perception of the system's usefulness.""}",https://hal.archives-ouvertes.fr/hal-01376429/file/Bouchet_et_al._ITS2016.pdf
-“That Is a Suspicious Reaction!”: Interpreting Logits Variation to Detect NLP Adversarial Attacks,E. Mosca,"Adversarial attacks are a major challenge faced by current machine learning research. These purposely crafted inputs fool even the most advanced models, precluding their deployment in safety-critical applications. Extensive research in computer vision has been carried to develop reliable defense strategies. However, the same issue remains less explored in natural language processing. Our work presents a model-agnostic detector of adversarial text examples. The approach identifies patterns in the logits of the target classifier when perturbing the input text. The proposed detector improves the current state-of-the-art performance in recognizing adversarial inputs and exhibits strong generalization capabilities across different NLP models, datasets, and word-level attacks.","{'model': 'tldr@v2.0.0', 'text': 'A model-agnostic detector of adversarial text examples that identifies patterns in the logits of the target classifier when perturbing the input text and exhibits strong generalization capabilities across different NLP models, datasets, and word-level attacks.'}",https://aclanthology.org/2022.acl-long.538.pdf
-Comparison of Prompting Strategies on the Acquisition of Daily Living Skills.,Linsey M. Sabielny,"Determining the most effective prompting strategies to be used for individuals with significant intellectual disability can assist in the acquisition of skills, reduction of errors, and avoidance of prompt dependency. However, few studies have directly compared the effects of different prompting strategies to determine which are the most effective. In the present study, physical only and physical plus vocal prompting strategies were compared to determine if one led to more efficient acquisition of two folding tasks than the other. An alternating treatments design was used with two individuals with significant intellectual disability, whereby the prompting strategies were counterbalanced across folding a shirt and folding a pair of pants. Results demonstrated that both strategies were equally effective for one participant, whereas the physical only prompting strategy may have been the more effective strategy for the second participant. The results suggest that pairing vocal instructions with a physical prompting strategy neither inhibits nor assists the acquisition of daily living skills for individuals with significant intellectual disability.","{'model': 'tldr@v2.0.0', 'text': 'The results suggest that pairing vocal instructions with a physical prompting strategy neither inhibits nor assists the acquisition of daily living skills for individuals with significant intellectual disability.'}",
-Benchmarking Intersectional Biases in NLP,John P. Lalor,"There has been a recent wave of work assessing the fairness of machine learning models in general, and more specifically, on natural language processing (NLP) models built using machine learning techniques. While much work has highlighted biases embedded in state-of-the-art language models, and more recent efforts have focused on how to debias, research assessing the fairness and performance of biased/debiased models on downstream prediction tasks has been limited. Moreover, most prior work has emphasized bias along a single dimension such as gender or race. In this work, we benchmark multiple NLP models with regards to their fairness and predictive performance across a variety of NLP tasks. In particular, we assess intersectional bias - fairness across multiple demographic dimensions. The results show that while current debiasing strategies fare well in terms of the fairness-accuracy trade-off (generally preserving predictive power in debiased models), they are unable to effectively alleviate bias in downstream tasks. Furthermore, this bias is often amplified across dimensions (i.e., intersections). We conclude by highlighting possible causes and making recommendations for future NLP debiasing research.","{'model': 'tldr@v2.0.0', 'text': 'The results show that while current debiasing strategies fare well in terms of the fairness-accuracy trade-off (generally preserving predictive power in debiased models), they are unable to effectively alleviate bias in downstream tasks.'}",https://aclanthology.org/2022.naacl-main.263.pdf
-Impact of Different Pedagogical Agents' Adaptive Self-regulated Prompting Strategies on Learning with MetaTutor,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Compared learners' perception of PAs' prompts with MetaTutor, a hypermedia adaptive learning environment, with 40 undergraduates randomly assigned to one of three condi- tions, preliminary results indicated that more frequent, but adaptive prompting is an efficient scaffolding strategy, despite negatively im- pacting learners' satisfaction.""}",http://escholarship.mcgill.ca/downloads/9s161b83t
-Measure and Improve Robustness in NLP Models: A Survey,Xuezhi Wang,"As NLP models achieved state-of-the-art performances over benchmarks and gained wide applications, it has been increasingly important to ensure the safe deployment of these models in the real world, e.g., making sure the models are robust against unseen or challenging scenarios. Despite robustness being an increasingly studied topic, it has been separately explored in applications like vision and NLP, with various definitions, evaluation and mitigation strategies in multiple lines of research. In this paper, we aim to provide a unifying survey of how to define, measure and improve robustness in NLP. We first connect multiple definitions of robustness, then unify various lines of work on identifying robustness failures and evaluating models’ robustness. Correspondingly, we present mitigation strategies that are data-driven, model-driven, and inductive-prior-based, with a more systematic view of how to effectively improve robustness in NLP models. Finally, we conclude by outlining open challenges and future directions to motivate further research in this area.","{'model': 'tldr@v2.0.0', 'text': 'A unifying survey of how to define, measure and improve robustness in NLP is provided, which first connects multiple definitions of robustness, then unify various lines of work on identifying robustness failures and evaluating models’ robustness.'}",https://aclanthology.org/2022.naacl-main.339.pdf
-Low-Resource NER by Data Augmentation With Prompting,Jian Liu,"Named entity recognition (NER) is a fundamental information extraction task that seeks to identify entity mentions of certain types in text. Despite numerous advances, the existing NER methods rely on extensive supervision for model training, which struggle in a low-resource scenario with limited training data. In this paper, we propose a new data augmentation method for low-resource NER, by eliciting knowledge from BERT with prompting strategies. Particularly, we devise a label-conditioned word replacement strategy that can produce more label-consistent examples by capturing the underlying word-label dependencies, and a prompting with question answering method to generate new training data from unlabeled texts. The experimental results have widely confirmed the effectiveness of our approach. Particularly, in a low-resource scenario with only 150 training sentences, our approach outperforms previous methods without data augmentation by over 40% in F1 and prior best data augmentation methods by over 2.0% in F1. Furthermore, our approach also fits with a zero-shot scenario, yielding promising results without using any human-labeled data for the task.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new data augmentation method for low-resource NER, by eliciting knowledge from BERT with prompting strategies, and devise a label-conditioned word replacement strategy that can produce more label-consistent examples by capturing the underlying word-label dependencies.'}",https://www.ijcai.org/proceedings/2022/0590.pdf
-The Dangers of Underclaiming: Reasons for Caution When Reporting How NLP Systems Fail,Sam Bowman,"Researchers in NLP often frame and discuss research results in ways that serve to deemphasize the field’s successes, often in response to the field’s widespread hype. Though well-meaning, this has yielded many misleading or false claims about the limits of our best technology. This is a problem, and it may be more serious than it looks: It harms our credibility in ways that can make it harder to mitigate present-day harms, like those involving biased systems for content moderation or resume screening. It also limits our ability to prepare for the potentially enormous impacts of more distant future advances. This paper urges researchers to be careful about these claims and suggests some research directions and communication strategies that will make it easier to avoid or rebut them.",,https://aclanthology.org/2022.acl-long.516.pdf
-PALI-NLP at SemEval-2022 Task 4: Discriminative Fine-tuning of Transformers for Patronizing and Condescending Language Detection,Dou Hu,"Patronizing and condescending language (PCL) has a large harmful impact and is difficult to detect, both for human judges and existing NLP systems. At SemEval-2022 Task 4, we propose a novel Transformer-based model and its ensembles to accurately understand such language context for PCL detection. To facilitate comprehension of the subtle and subjective nature of PCL, two fine-tuning strategies are applied to capture discriminative features from diverse linguistic behaviour and categorical distribution. The system achieves remarkable results on the official ranking, including 1st in Subtask 1 and 5th in Subtask 2. Extensive experiments on the task demonstrate the effectiveness of our system and its strategies.","{'model': 'tldr@v2.0.0', 'text': 'A novel Transformer-based model and its ensembles are proposed to accurately understand such language context for PCL detection and two fine-tuning strategies are applied to facilitate comprehension of the subtle and subjective nature of PCL.'}",http://arxiv.org/pdf/2203.04616
-Smart homes for people with Alzheimer's disease: adapting prompting strategies to the patient's cognitive profile,J. Lapointe,"Smart home technologies constitute a potential solution to allow people with Alzheimer's disease (AD) to remain in their home. These intelligent houses contain technological devices aiming to provide adapted cognitive assistance (prompts) when needed. However, a literature review of the field revealed a predominant use of verbal prompts with little knowledge about their real effectiveness. To contribute solving this important issue, we propose, in this paper, comprehensive guidelines to help smart homes researchers to maximize the efficiency by adapting the form of prompts to the specific cognitive profiles of patients with AD. First, we identify the main deficits of AD that influence the effectiveness of prompts. Second, we details which prompting strategy to use accordingly. Third, we propose an experimental protocol, based on a well-known test, and a new prompting software, which allows to validate the proposed guidelines. Finally, we present the preliminary results of a first experiment conducted in our lab with participants having mild to moderate AD.","{'model': 'tldr@v2.0.0', 'text': 'This paper identifies the main deficits of AD that influence the effectiveness of prompts and identifies which prompting strategy to use accordingly, and proposes comprehensive guidelines to help smart homes researchers to maximize the efficiency.'}",
-This joke is [MASK]: Recognizing Humor and Offense with Prompting,Junze Li,"Humor is a magnetic component in everyday human interactions and communications. Computationally modeling humor enables NLP systems to entertain and engage with users. We investigate the effectiveness of prompting, a new transfer learning paradigm for NLP, for humor recognition. We show that prompting performs similarly to finetuning when numerous annotations are available, but gives stellar performance in low-resource humor recognition. The relationship between humor and offense is also inspected by applying influence functions to prompting; we show that models could rely on offense to determine humor during transfer.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the effectiveness of prompting, a new transfer learning paradigm for NLP, for humor recognition, and shows that prompting performs similarly to finetuning when numerous annotations are available, but gives stellar performance in low-resource humor recognition.'}",http://arxiv.org/pdf/2210.13985
-White-box Testing of NLP models with Mask Neuron Coverage,Arshdeep Sekhon,"Recent literature has seen growing interest in using black-box strategies like CheckList for testing the behavior of NLP models. Research on white-box testing has developed a number of methods for evaluating how thoroughly the internal behavior of deep models is tested, but they are not applicable to NLP models. We propose a set of white-box testing methods that are customized for transformer-based NLP models. These include Mask Neuron Coverage (MNCOVER) that measures how thoroughly the attention layers in models are exercised during testing. We show that MNCOVER can refine testing suites generated by CheckList by substantially reduce them in size, for more than 60\% on average, while retaining failing tests -- thereby concentrating the fault detection power of the test suite. Further we show how MNCOVER can be used to guide CheckList input generation, evaluate alternative NLP testing methods, and drive data augmentation to improve accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A set of white-box testing methods that are customized for transformer-based NLP models, including Mask Neuron Coverage (MNCOVER) that measures how thoroughly the attention layers in models are exercised during testing are proposed.'}",http://arxiv.org/pdf/2205.05050
-Effectiveness of training of Neuro-linguistic programming (NLP) strategies on reducing social anxiety,Zahra Sylabkhori,"The paper was aimed at exploring the efficacy of Neuro-linguistic programming (NLP) strategies on social anxiety people referring to counseling centers. To do this, the number of 30 people with social anxiety disorder who had referred for counseling was selected as the sample, and completed Liebowitz Social Anxiety Scale (1987). The pretest having been administered, 30 people were randomly placed in experimental and control groups respectively. The experimental group received ten sessions of training of neuro-linguistic programming strategies. The interventions were a combination of NLP tools, including the art of self-assertiveness, WACO system, Neuro-linguistic logical level, goal-setting, Switch system, chart sections, rapport, Metal model, encore and modeling. After the interventions were ended, the subjects of both groups were tested with Liebowitz Test twice. The data obtained were examined by using SPSS software ver.17 and statistical independent T Test. Research findings illustrated a significant difference between the experimental and control groups as regards posttest social anxiety scores (P = -0/001). Also, with regards to performance anxiety, a significant difference was seen between the two groups in the posttest (P = -0/001).",,
-On-the-Job Training of Special Education Staff: Teaching the Simultaneous Prompting Strategies.,Sezgin Vuran,"AbstractThe aim of this study is to analyze the effectiveness of stafftraining on discrete-trial teaching (DTT). Multiple baseline design across subjects was used in order to analyze the effect of the training program on the educators' performance on probing and intervention implementation. For teaching these two skills, presentation of an information manual, live model and error correction including feedback giving through video were used. The results showed that the percentage of correct response related to probing and training skills through simultaneous prompting was 100% among all participants. Students who were instructed by these educators also reached 80-100% correct responding level in terms of the skill taught. Follow-up data was collected 4-8 weeks after the completion of the process and it was seen that the participants partially maintained the skills acquired. Social validity data was collected in order to assess opinions of the participants about the survey.Key WordsStaffTraining, Simultaneous Promptings, Discrete-Trial Teaching, Individuals with Developmental Disabilities.The success of individuals with developmental disabilities depends on the effective and proper use of appropriate teaching methods as well as the changes and adaptations of the attitudes of the people who play active roles in the child's education, such as peers, staffand parents. Staffmembers are individuals who interact with children with developmental disabilities and provide services to them (Sturmey, 2008). Considering this relationship, the effectiveness and significance of training these staffmembers have become an important current issue.StaffTraining: A Review of the LiteratureIn the existing literature, there are many studies on the effects of staffmember training on a staff's performance and the performances of the individuals with developmental disabilities to whom the staffprovides service. Most of the studies aim to teach discrete-trial teaching, an evidence-based practice, to staffmembers working with children who suffer from developmental disabilities (Belfiore, Fritts, & Herman, 2008; Dib & Sturmey, 2007; Fazzio, Martin, Arnal, & Yu, 2009; Koegel, Russo, & Rincover, 1977; LeBlanc, Ricciardi, & Luiselli, 2005; Ryan & Hemmes, 2005; Sarokoff& Sturmey, 2004, 2008; Thiessen et al., 2009). The findings of these studies indicate that staffmembers, who have undergone this training, enjoy a distinct increase in the accurate usage of this method. In another studies on stafftraining, teaching through daily routines (Lavie & Sturmey, 2002), and the evaluation of stimulus preferences (Lavie & Sturmey), have been conducted to gain information and skills to staffmember.In stafftraining, there are four main methods: the handbook, which presents related information through a trainer's written notes; modeling, which trains the staffvia live performances or videos; rehearsal, which encourages trainees to apply their skills; and feedback, which is the trainer's explanations of the staff's performance through written, oral or graphic accounts (Sturmey, 2008).In most of the studies investigating the effectiveness of stafftraining methods, at least two of these methods were used together (Dib & Sturmey, 2007; Koegel et al., 1977; Lavie & Sturmey, 2002; Leblanc et al., 2005; Ryan & Hemmes, 2005; Sarokoff& Sturmey, 2004, 2008; Schepis, Reid, Ownbey, & Parsons, 2001), yet in only one study was the effectiveness of self video modeling (in which the participants watch themselves) examined (Belfiore et al., 2008). The teaching methods in stafftraining play a significant role in creating the desired effects on the staffand the individuals taught by the staff. In addition, these methods are expected to be efficient in terms of time and effort. These features are involved in the desired characteristics of stafftraining (Sturmey, 2008).In Turkey, the discrete-trial approach was applied in various studies. …",,
-Demonstrate-Search-Predict: Composing retrieval and language models for knowledge-intensive NLP,O. Khattab,"Retrieval-augmented in-context learning has emerged as a powerful approach for addressing knowledge-intensive tasks using frozen language models (LM) and retrieval models (RM). Existing work has combined these in simple""retrieve-then-read""pipelines in which the RM retrieves passages that are inserted into the LM prompt. To begin to fully realize the potential of frozen LMs and RMs, we propose Demonstrate-Search-Predict (DSP), a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM. DSP can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions, systematically breaking down problems into small transformations that the LM and RM can handle more reliably. We have written novel DSP programs for answering questions in open-domain, multi-hop, and conversational settings, establishing in early evaluations new state-of-the-art in-context learning results and delivering 37-120%, 8-39%, and 80-290% relative gains against the vanilla LM (GPT-3.5), a standard retrieve-then-read pipeline, and a contemporaneous self-ask pipeline, respectively. We release DSP at https://github.com/stanfordnlp/dsp","{'model': 'tldr@v2.0.0', 'text': 'Demonstrate-Search-Predict (DSP) is proposed, a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM and can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions.'}",http://arxiv.org/pdf/2212.14024
-Error Analysis Prompting Enables Human-Like Translation Evaluation in Large Language Models: A Case Study on ChatGPT,Qingyu Lu,"Generative large language models (LLMs), e.g., ChatGPT, have demonstrated remarkable proficiency across several NLP tasks, such as machine translation, text summarization. Recent research (Kocmi and Federmann, 2023) has shown that utilizing ChatGPT for assessing the quality of machine translation (MT) achieves state-of-the-art performance at the system level but performs poorly at the segment level. To further improve the performance of LLMs on MT quality assessment, we conduct an investigation into several prompting methods, and propose a new prompting method called Error Analysis Prompting (EAPrompt) by combining Chain-of-Thoughts (Wei et al., 2022) and Error Analysis (Lu et al., 2022). Our results on WMT22 indicate that prompting LLMs like ChatGPT with error analysis can generate human-like MT evaluations at both the system and segment level. Additionally, we first discover some limitations of ChatGPT as an MT evaluator, such as changing the order of input may significantly influence the judgment when providing multiple translations in a single query. This work provides a preliminary experience of prompting LLMs as an evaluator to improve the reliability of translation evaluation metrics under the error analysis paradigm.","{'model': 'tldr@v2.0.0', 'text': 'The results on WMT22 indicate that prompting LLMs like ChatGPT with error analysis can generate human-like MT evaluations at both the system and segment level, and a new prompting method called EAPrompt is proposed by combining Chain-of-Thoughts and Error Analysis.'}",https://arxiv.org/pdf/2303.13809
-Explicit Visual Prompting for Low-Level Structure Segmentations,Weihuang Liu,"We consider the generic problem of detecting low-level structures in images, which includes segmenting the manipulated parts, identifying out-of-focus pixels, separating shadow regions, and detecting concealed objects. Whereas each such topic has been typically addressed with a domain-specific solution, we show that a unified approach performs well across all of them. We take inspiration from the widely-used pre-training and then prompt tuning protocols in NLP and propose a new visual prompting model, named Explicit Visual Prompting (EVP). Different from the previous visual prompting which is typically a dataset-level implicit embedding, our key insight is to enforce the tunable parameters focusing on the explicit visual content from each individual image, i.e., the features from frozen patch embeddings and the input's high-frequency components. The proposed EVP significantly outperforms other parameter-efficient tuning protocols under the same amount of tunable parameters (5.7% extra trainable parameters of each task). EVP also achieves state-of-the-art performances on diverse low-level structure segmentation tasks compared to task-specific solutions. Our code is available at: https://github.com/NiFangBaAGe/Explicit-Visual-Prompt.","{'model': 'tldr@v2.0.0', 'text': 'This work takes inspiration from the widely-used pre-training and then prompt tuning protocols in NLP and proposes a new visual prompting model, named Explicit Visual Prompting (EVP), which significantly outperforms other parameter-efficient tuning protocols under the same amount of tunable parameters.'}",https://arxiv.org/pdf/2303.10883
-NLP-Based Prediction of Medical Specialties at Hospital Admission Using Triage Notes,Émilien Arnaud,"Data Analytics is rapidly expanding within the healthcare domain to help develop strategies for improving the quality of care and curbing costs as well. Natural Language Processing (NLP) solutions have received particular attention whereas a large part of clinical data is stockpiled into unstructured physician or nursing notes. In this respect, we attempt to employ NLP to provide an early prediction of the medical specialties at hospital admission. The study uses a large-scale dataset including more than 260K ED records provided by the Amiens-Picardy University Hospital in France. Our approach aims to integrate structured data with unstructured textual notes recorded at the triage stage. On one hand, a standard MLP model is used against the typical set of features. On the other hand, a Convolutional Neural Network is used to operate over the textual data. While both learning components are conducted independently in parallel. The empirical results demonstrated a promising accuracy in general. It is conceived that the study could be an additional contribution to the mounting efforts of applying NLP methods in the healthcare domain.","{'model': 'tldr@v2.0.0', 'text': 'This work attempts to employ NLP to provide an early prediction of the medical specialties at hospital admission using a large-scale dataset including more than 260K ED records provided by the Amiens-Picardy University Hospital in France.'}",
-Pushing the Limits of ChatGPT on NLP Tasks,Xiaofei Sun,"Despite the success of ChatGPT, its performances on most NLP tasks are still well below the supervised baselines. In this work, we looked into the causes, and discovered that its subpar performance was caused by the following factors: (1) token limit in the prompt does not allow for the full utilization of the supervised datasets; (2) mismatch between the generation nature of ChatGPT and NLP tasks; (3) intrinsic pitfalls of LLMs models, e.g., hallucination, overly focus on certain keywords, etc. In this work, we propose a collection of general modules to address these issues, in an attempt to push the limits of ChatGPT on NLP tasks. Our proposed modules include (1) a one-input-multiple-prompts strategy that employs multiple prompts for one input to accommodate more demonstrations; (2) using fine-tuned models for better demonstration retrieval; (3) transforming tasks to formats that are more tailored to the generation nature; (4) employing reasoning strategies that are tailored to addressing the task-specific complexity; (5) the self-verification strategy to address the hallucination issue of LLMs; (6) the paraphrase strategy to improve the robustness of model predictions. We conduct experiments on 21 datasets of 10 representative NLP tasks, including question answering, commonsense reasoning, natural language inference, sentiment analysis, named entity recognition, entity-relation extraction, event extraction, dependency parsing, semantic role labeling, and part-of-speech tagging. Using the proposed assemble of techniques, we are able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.","{'model': 'tldr@v2.0.0', 'text': 'Using the proposed assemble of techniques, this work is able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.'}",https://arxiv.org/pdf/2306.09719
-All in One: Multi-Task Prompting for Graph Neural Networks,Xiangguo Sun,"Recently, ""pre-training and fine-tuning'' has been adopted as a standard workflow for many graph tasks since it can take general graph knowledge to relieve the lack of graph annotations from each application. However, graph tasks with node level, edge level, and graph level are far diversified, making the pre-training pretext often incompatible with these multiple tasks. This gap may even cause a ""negative transfer'' to the specific application, leading to poor results. Inspired by the prompt learning in natural language processing (NLP), which has presented significant effectiveness in leveraging prior knowledge for various NLP tasks, we study the prompting topic for graphs with the motivation of filling the gap between pre-trained models and various graph tasks. In this paper, we propose a novel multi-task prompting method for graph models. Specifically, we first unify the format of graph prompts and language prompts with the prompt token, token structure, and inserting pattern. In this way, the prompting idea from NLP can be seamlessly introduced to the graph area. Then, to further narrow the gap between various graph tasks and state-of-the-art pre-training strategies, we further study the task space of various graph applications and reformulate downstream problems to the graph-level task. Afterward, we introduce meta-learning to efficiently learn a better initialization for the multi-task prompt of graphs so that our prompting framework can be more reliable and general for different tasks. We conduct extensive experiments, results from which demonstrate the superiority of our method.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel multi-task prompting method for graph models that unify the format of graph prompts and language prompts with the prompt token, token structure, and inserting pattern, and introduces meta-learning to efficiently learn a better initialization for the multi- task prompt of graphs so that the prompting framework can be more reliable and general for different tasks.'}",https://arxiv.org/pdf/2307.01504
-DAMO-NLP at SemEval-2023 Task 2: A Unified Retrieval-augmented System for Multilingual Named Entity Recognition,Zeqi Tan,"The MultiCoNER II shared task aims to tackle multilingual named entity recognition (NER) in fine-grained and noisy scenarios, and it inherits the semantic ambiguity and low-context setting of the MultiCoNER I task. To cope with these problems, the previous top systems in the MultiCoNER I either incorporate the knowledge bases or gazetteers. However, they still suffer from insufficient knowledge, limited context length, single retrieval strategy. In this paper, our team DAMO-NLP proposes a unified retrieval-augmented system (U-RaNER) for fine-grained multilingual NER. We perform error analysis on the previous top systems and reveal that their performance bottleneck lies in insufficient knowledge. Also, we discover that the limited context length causes the retrieval knowledge to be invisible to the model. To enhance the retrieval context, we incorporate the entity-centric Wikidata knowledge base, while utilizing the infusion approach to broaden the contextual scope of the model. Also, we explore various search strategies and refine the quality of retrieval knowledge. Our system wins 9 out of 13 tracks in the MultiCoNER II shared task. Additionally, we compared our system with ChatGPT, one of the large language models which have unlocked strong capabilities on many tasks. The results show that there is still much room for improvement for ChatGPT on the extraction task.","{'model': 'tldr@v2.0.0', 'text': 'A unified retrieval-augmented system (U-RaNER) for fine-grained multilingual NER, which incorporates the entity-centric Wikidata knowledge base, while utilizing the infusion approach to broaden the contextual scope of the model.'}",http://arxiv.org/pdf/2305.03688
-Diffusion-NAT: Self-Prompting Discrete Diffusion for Non-Autoregressive Text Generation,Kun Zhou,"Recently, continuous diffusion models (CDM) have been introduced into non-autoregressive (NAR) text-to-text generation. However, the discrete nature of text increases the difficulty of CDM to generate coherent and fluent texts, and also causes the incompatibility problem between CDM and advanced NLP techniques, especially the popular pre-trained language models~(PLMs). To solve it, we propose Diffusion-NAT, which introduces discrete diffusion models~(DDM) into NAR text-to-text generation and integrates BART to improve the performance. By revising the decoding process of BART and the typical settings of DDM, we unify the inference process of BART and the denoising process of DDM into the same NAR masked tokens recovering task. In this way, DDM can rely on BART to perform denoising, which can benefit from both the rich pre-learned knowledge of BART and the iterative refining paradigm of DDM. Besides, we also propose the iterative self-prompting strategy to further improve the generation quality. Experimental results on 7 datasets show that our approach can outperform competitive NAR methods, and even surpass autoregressive methods. Our code and data will be publicly released.","{'model': 'tldr@v2.0.0', 'text': 'Diffusion-NAT is proposed, which introduces discrete diffusion models~(DDM) into NAR text-to-text generation and integrates BART to improve the performance and proposes the iterative self-prompting strategy to further improve the generation quality.'}",http://arxiv.org/pdf/2305.04044
-Sustainable Marketing and the Role of Social Media: An Experimental Study Using Natural Language Processing (NLP),Ganesh Dash,"Marketing has changed fundamentally in the new millennium. At the same time, sustainable marketing strategies have evolved to meet the challenges of environmental issues. In this study, we examined the trends in sustainable marketing strategies and the role of social media in these. Based on specific keywords per the objective, this study collected 33 published articles from the Scopus database from 1991 to 2022 (2012–2022). The KNIME (Konstanz Information Miner) and VOSviewer tools were deployed to provide detailed classification and prediction of the various trends in sustainable marketing, with a particular focus on the role of social media. The study method applied text mining and latent semantic analysis to predict the latest trends. The top three trends were Green Marketing and Consumer Behavior, Sustainable Social Media Marketing, and Influencer Social Media Marketing Practices. This NLP-based review and the clustering of research directions provide immense value to marketers and policymakers.",,https://www.mdpi.com/2071-1050/15/6/5443/pdf?version=1679297784
-MatSci-NLP: Evaluating Scientific Language Models on Materials Science Language Tasks Using Text-to-Schema Modeling,Yurun Song,"We present MatSci-NLP, a natural language benchmark for evaluating the performance of natural language processing (NLP) models on materials science text. We construct the benchmark from publicly available materials science text data to encompass seven different NLP tasks, including conventional NLP tasks like named entity recognition and relation classification, as well as NLP tasks specific to materials science, such as synthesis action retrieval which relates to creating synthesis procedures for materials. We study various BERT-based models pretrained on different scientific text corpora on MatSci-NLP to understand the impact of pretraining strategies on understanding materials science text. Given the scarcity of high-quality annotated data in the materials science domain, we perform our fine-tuning experiments with limited training data to encourage the generalize across MatSci-NLP tasks.Our experiments in this low-resource training setting show that language models pretrained on scientific text outperform BERT trained on general text. MatBERT, a model pretrained specifically on materials science journals, generally performs best for most tasks. Moreover, we propose a unified text-to-schema for multitask learning on {pasted macro ‘BENCHMARK’} and compare its performance with traditional fine-tuning methods. In our analysis of different training methods, we find that our proposed text-to-schema methods inspired by question-answering consistently outperform single and multitask NLP fine-tuning methods. The code and datasets are publicly available https://github.com/BangLab-UdeM-Mila/NLP4MatSci-ACL23.","{'model': 'tldr@v2.0.0', 'text': 'In the analysis of different training methods, it is found that the proposed text-to-schema methods inspired by question-answering consistently outperform single and multitask NLP fine-tuning methods.'}",http://arxiv.org/pdf/2305.08264
-Video Prompting Versus Other Instruction Strategies for Persons With Alzheimer’s Disease,Viviana Perilli,"Background/Aim: Two studies assessed the effectiveness of video prompting as a strategy to support persons with mild and moderate Alzheimer’s disease in performing daily activities. Methods: In study I, video prompting was compared to an existing strategy relying on verbal instructions. In study II, video prompting was compared to another existing strategy relying on static pictorial cues. Video prompting and the other strategies were counterbalanced across tasks and participants and compared within alternating treatments designs. Results: Video prompting was effective in all participants. Similarly effective were the other 2 strategies, and only occasional differences between the strategies were reported. Two social validation assessments showed that university psychology students and graduates rated the patients’ performance with video prompting more favorably than their performance with the other strategies. Conclusion: Video prompting may be considered a valuable alternative to the other strategies to support daily activities in persons with Alzheimer's disease.","{'model': 'tldr@v2.0.0', 'text': ""Video prompting may be considered a valuable alternative to the other strategies to support daily activities in persons with Alzheimer's disease.""}",
-ParaFuzz: An Interpretability-Driven Technique for Detecting Poisoned Samples in NLP,Lu Yan,"Backdoor attacks have emerged as a prominent threat to natural language processing (NLP) models, where the presence of specific triggers in the input can lead poisoned models to misclassify these inputs to predetermined target classes. Current detection mechanisms are limited by their inability to address more covert backdoor strategies, such as style-based attacks. In this work, we propose an innovative test-time poisoned sample detection framework that hinges on the interpretability of model predictions, grounded in the semantic meaning of inputs. We contend that triggers (e.g., infrequent words) are not supposed to fundamentally alter the underlying semantic meanings of poisoned samples as they want to stay stealthy. Based on this observation, we hypothesize that while the model's predictions for paraphrased clean samples should remain stable, predictions for poisoned samples should revert to their true labels upon the mutations applied to triggers during the paraphrasing process. We employ ChatGPT, a state-of-the-art large language model, as our paraphraser and formulate the trigger-removal task as a prompt engineering problem. We adopt fuzzing, a technique commonly used for unearthing software vulnerabilities, to discover optimal paraphrase prompts that can effectively eliminate triggers while concurrently maintaining input semantics. Experiments on 4 types of backdoor attacks, including the subtle style backdoors, and 4 distinct datasets demonstrate that our approach surpasses baseline methods, including STRIP, RAP, and ONION, in precision and recall.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an innovative test-time poisoned sample detection framework that hinges on the interpretability of model predictions, grounded in the semantic meaning of inputs, and adopts fuzzing, a technique commonly used for unearthing software vulnerabilities, to discover optimal paraphrase prompts that can effectively eliminate triggers while concurrently maintaining input semantics.'}",https://arxiv.org/pdf/2308.02122
-Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP,Timo Schick,"Abstract ⚠ This paper contains prompts and model outputs that are offensive in nature. When trained on large, unfiltered crawls from the Internet, language models pick up and reproduce all kinds of undesirable biases that can be found in the data: They often generate racist, sexist, violent, or otherwise toxic language. As large models require millions of training examples to achieve good performance, it is difficult to completely prevent them from being exposed to such content. In this paper, we first demonstrate a surprising finding: Pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce. We refer to this capability as self-diagnosis. Based on this finding, we then propose a decoding algorithm that, given only a textual description of the undesired behavior, reduces the probability of a language model producing problematic text. We refer to this approach as self-debiasing. Self-debiasing does not rely on manually curated word lists, nor does it require any training data or changes to the model’s parameters. While we by no means eliminate the issue of language models generating biased text, we believe our approach to be an important step in this direction.1","{'model': 'tldr@v2.0.0', 'text': 'This paper demonstrates a surprising finding: Pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce and proposes a decoding algorithm that reduces the probability of a language model producing problematic text, known as self-debiasing.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf
-Personalization of News for a Logistics Organisation by Finding Relevancy Using NLP,R. Garg,,"{'model': 'tldr@v2.0.0', 'text': 'This paper attempts to provide a state of art in finding relevancy in news headlines by using statistical and deep learning models to extract information from the corpora using natural language processing.'}",
-Large Language Models (LLMs) for Natural Language Processing (NLP) of Oil and Gas Drilling Data,Prateek Kumar," In the oil and gas industry, drilling activities spawn substantial volumes of unstructured textual data. The examination and interpretation of these data pose significant challenges. This research exploits the emerging capabilities of large language models (LLMs) with over 100 billion parameters to extract actionable insights from raw drilling data. Through fine-tuning methodologies and the use of various prompt engineering strategies, we addressed several text downstream tasks, including summarization, classification, entity recognition, and information extraction. This study delves into our methods, findings, and the novel application of LLMs for efficient and precise analysis of drilling data.","{'model': 'tldr@v2.0.0', 'text': 'This research exploits the emerging capabilities of large language models with over 100 billion parameters to extract actionable insights from raw drilling data through fine-tuning methodologies and the use of various prompt engineering strategies.'}",
-Returning the N to NLP: Towards Contextually Personalized Classification Models,Lucie Flekova,"Most NLP models today treat language as universal, even though socio- and psycholingustic research shows that the communicated message is influenced by the characteristics of the speaker as well as the target audience. This paper surveys the landscape of personalization in natural language processing and related fields, and offers a path forward to mitigate the decades of deviation of the NLP tools from sociolingustic findings, allowing to flexibly process the “natural” language of each user rather than enforcing a uniform NLP treatment. It outlines a possible direction to incorporate these aspects into neural NLP models by means of socially contextual personalization, and proposes to shift the focus of our evaluation strategies accordingly.","{'model': 'tldr@v2.0.0', 'text': 'The landscape of personalization in natural language processing and related fields is surveyed, and a path forward to mitigate the decades of deviation of the NLP tools from sociolingustic findings is offered, allowing to flexibly process the “natural” language of each user rather than enforcing a uniform NLP treatment.'}",https://www.aclweb.org/anthology/2020.acl-main.700.pdf
-A Survey on Bias in Deep NLP,Ismael Garrido-Muñoz,"Deep neural networks are hegemonic approaches to many machine learning areas, including natural language processing (NLP). Thanks to the availability of large corpora collections and the capability of deep architectures to shape internal language mechanisms in self-supervised learning processes (also known as “pre-training”), versatile and performing models are released continuously for every new network design. These networks, somehow, learn a probability distribution of words and relations across the training collection used, inheriting the potential flaws, inconsistencies and biases contained in such a collection. As pre-trained models have been found to be very useful approaches to transfer learning, dealing with bias has become a relevant issue in this new scenario. We introduce bias in a formal way and explore how it has been treated in several networks, in terms of detection and correction. In addition, available resources are identified and a strategy to deal with bias in deep NLP is proposed.","{'model': 'tldr@v2.0.0', 'text': 'Bias is introduced in a formal way and how it has been treated in several networks, in terms of detection and correction, and a strategy to deal with bias in deep NLP is proposed.'}",
-A decision tree for the use of auditory prompting strategies,Michał Post,"This article provides a decision tree for determining the most effective method of auditory prompting systems in order to increase workplace independence for employees with severe disabilities. A flow-through chart, The Auditory Prompts Decision Tree, is presented to guide this planning process. The Auditory Prompts Decision Tree is useful when analyzing why the breakdown in work performance is occurring and for selecting the types of auditory prompts that will support the employee's work performance needs. Specific examples from research studies are provided to illustrate the use of the Auditory Prompts Decision Tree for each type of auditory prompt.",,
-BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation Models,Kangjie Chen,"Pre-trained Natural Language Processing (NLP) models can be easily adapted to a variety of downstream language tasks. This significantly accelerates the development of language models. However, NLP models have been shown to be vulnerable to backdoor attacks, where a pre-defined trigger word in the input text causes model misprediction. Previous NLP backdoor attacks mainly focus on some specific tasks. This makes those attacks less general and applicable to other kinds of NLP models and tasks. In this work, we propose \Name, the first task-agnostic backdoor attack against the pre-trained NLP models. The key feature of our attack is that the adversary does not need prior information about the downstream tasks when implanting the backdoor to the pre-trained model. When this malicious model is released, any downstream models transferred from it will also inherit the backdoor, even after the extensive transfer learning process. We further design a simple yet effective strategy to bypass a state-of-the-art defense. Experimental results indicate that our approach can compromise a wide range of downstream NLP tasks in an effective and stealthy way.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results indicate that the proposed \\Name, the first task-agnostic backdoor attack against the pre-trained NLP models, can compromise a wide range of downstream NLP tasks in an effective and stealthy way.'}",
-EasyTransfer: A Simple and Scalable Deep Transfer Learning Platform for NLP Applications,Minghui Qiu,"The literature has witnessed the success of leveraging Pre-trained Language Models (PLMs) and Transfer Learning (TL) algorithms to a wide range of Natural Language Processing (NLP) applications, yet it is not easy to build an easy-to-use and scalable TL toolkit for this purpose. To bridge this gap, the EasyTransfer platform is designed to develop deep TL algorithms for NLP applications. EasyTransfer is backended with a high-performance and scalable engine for efficient training and inference, and also integrates comprehensive deep TL algorithms, to make the development of industrial-scale TL applications easier. In EasyTransfer, the built-in data and model parallelism strategies, combined with AI compiler optimization, show to be 4.0x faster than the community version of distributed training. EasyTransfer supports various NLP models in the ModelZoo, including mainstream PLMs and multi-modality models. It also features various in-house developed TL algorithms, together with the AppZoo for NLP applications. The toolkit is convenient for users to quickly start model training, evaluation, and online deployment. EasyTransfer is currently deployed at Alibaba to support a variety of business scenarios, including item recommendation, personalized search, conversational question answering, etc. Extensive experiments on real-world datasets and online applications show that EasyTransfer is suitable for online production with cutting-edge performance for various applications. The source code of EasyTransfer is released at Github1.","{'model': 'tldr@v2.0.0', 'text': 'The EasyTransfer platform is designed to develop deep TL algorithms for NLP applications, backended with a high-performance and scalable engine for efficient training and inference, and also integrates comprehensive deepTL algorithms, to make the development of industrial-scale TL applications easier.'}",
-Integrating Ethics into the NLP Curriculum,Emily M. Bender,"To raise awareness among future NLP practitioners and prevent inertia in the field, we need to place ethics in the curriculum for all NLP students—not as an elective, but as a core part of their education. Our goal in this tutorial is to empower NLP researchers and practitioners with tools and resources to teach others about how to ethically apply NLP techniques. We will present both high-level strategies for developing an ethics-oriented curriculum, based on experience and best practices, as well as specific sample exercises that can be brought to a classroom. This highly interactive work session will culminate in a shared online resource page that pools lesson plans, assignments, exercise ideas, reading suggestions, and ideas from the attendees. Though the tutorial will focus particularly on examples for university classrooms, we believe these ideas can extend to company-internal workshops or tutorials in a variety of organizations. In this setting, a key lesson is that there is no single approach to ethical NLP: each project requires thoughtful consideration about what steps can be taken to best support people affected by that project. However, we can learn (and teach) what issues to be aware of, what questions to ask, and what strategies are available to mitigate harm.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial is to empower NLP researchers and practitioners with tools and resources to teach others about how to ethically apply NLP techniques, and will present both high-level strategies for developing an ethics-oriented curriculum, based on experience and best practices.'}",https://www.aclweb.org/anthology/2020.acl-tutorials.2.pdf
-A novel hybrid approach of SVM combined with NLP and probabilistic neural network for email phishing,Abhishek Kumar,"Phishing attacks are one of the slanting cyber-attacks that apply socially engineered messages that are imparted to individuals from expert hackers going for tricking clients to uncover their delicate data, the most mainstream correspondence channel to those messages is through clients' emails. Phishing has turned into a generous danger for web clients and a noteworthy reason for money related misfortunes. Therefore, different arrangements have been created to handle this issue. Deceitful emails, also called phishing emails, utilize a scope of impact strategies to convince people to react, for example, promising a fiscal reward or summoning a feeling of criticalness. Regardless of far reaching alerts and intends to instruct clients to distinguish phishing sends, these are as yet a pervasive practice and a worthwhile business. The creators accept that influence, as a style of human correspondence intended to impact others, has a focal job in fruitful advanced tricks. Cyber criminals have ceaselessly propelling their techniques for assault. The current strategies to recognize the presence of such malevolent projects and to keep them from executing are static, dynamic and hybrid analysis. In this work we are proposing a hybrid methodology for phishing detection incorporating feature extraction and classification of the mails using SVM. At last, alongside the chose features, the PNN characterizes the spam mails from the genuine mails with more exactness and accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This work is proposing a hybrid methodology for phishing detection incorporating feature extraction and classification of the mails using SVM, and alongside the chose features, the PNN characterizes the spam mails from the genuine mails with more exactness and accuracy.'}",http://ijece.iaescore.com/index.php/IJECE/article/download/20847/13512
-Prompting Strategies That Promote Student Self-Management,P. Alberto,"Students who perform tasks competently and independently tend to exhibit fewer instances of inappropriate behavior than those who have difficulty with them (Smith, 1985). Students with severe handicaps can perfonn more competently when teachers incorporate verbal and/or physical prompts into their instruction. However, competence attained this way tends to decrease when the teacher is not present to deliver the prompts. When this occurs, an alternative to teacher prompting must be implemented to maintain student perfonnance. This is challenging, especially in community settings. One way to reduce dependence on teachers is to teach students to use self-operated prompting strategies. This type of self-management fosters student independence and improves behavior in community environments.",,
-Triggerless Backdoor Attack for NLP Tasks with Clean Labels,Leilei Gan,"Backdoor attacks pose a new threat to NLP models. A standard strategy to construct poisoned data in backdoor attacks is to insert triggers (e.g., rare words) into selected sentences and alter the original label to a target label. This strategy comes with a severe flaw of being easily detected from both the trigger and the label perspectives: the trigger injected, which is usually a rare word, leads to an abnormal natural language expression, and thus can be easily detected by a defense model; the changed target label leads the example to be mistakenly labeled, and thus can be easily detected by manual inspections. To deal with this issue, in this paper, we propose a new strategy to perform textual backdoor attack which does not require an external trigger and the poisoned samples are correctly labeled. The core idea of the proposed strategy is to construct clean-labeled examples, whose labels are correct but can lead to test label changes when fused with the training set. To generate poisoned clean-labeled examples, we propose a sentence generation model based on the genetic algorithm to cater to the non-differentiable characteristic of text data. Extensive experiments demonstrate that the proposed attacking strategy is not only effective, but more importantly, hard to defend due to its triggerless and clean-labeled nature. Our work marks the first step towards developing triggerless attacking strategies in NLP.","{'model': 'tldr@v2.0.0', 'text': 'A new strategy to perform textual backdoor attack which does not require an external trigger and the poisoned samples are correctly labeled is proposed, which marks the first step towards developing triggerless attacking strategies in NLP.'}",https://aclanthology.org/2022.naacl-main.214.pdf
-Assessing the acquisition of incidental information by secondary-age students with mental retardation: comparison of response prompting strategies.,D. Gast,"The effectiveness and efficiency of four response prompting conditions (progressive time delay, progressive time delay with a descriptive consequent event, system of least prompts, and system of least prompts with a descriptive consequent event) were compared. Students with moderate to severe mental retardation were taught to read functional recipe words. Maintenance and students' acquisition of incidental information were assessed when it was (a) embedded in the prompts of the system of least prompts procedure, (b) included in the descriptive praise statements following correct performance with the progressive time delay and system of least prompts procedures, and (c) not presented. A multiple probe design across behaviors, replicated across subjects, was used. Results indicated that (a) each of the procedures produced criterion level responding: (b) efficiency data on traditional measures were roughly equal; (c) maintenance checks showed no differential effects related to the instructional condition; and (d) incidental information was acquired, although it was not directly targeted for instruction.","{'model': 'tldr@v2.0.0', 'text': 'Students with moderate to severe mental retardation were taught to read functional recipe words and incidental information was acquired, although it was not directly targeted for instruction.'}",
-Encouraging long-term compliance with breast self-examination: The evaluation of prompting strategies,J. Mayer,,"{'model': 'tldr@v2.0.0', 'text': 'Results indicated that the prompts, particularly those delivered by phone, were effective in encouraging initial performance of BSE, however, for all conditions, BSE decreased over the course of the study.'}",
-Prompting Strategies for those with a Severe Mental Handicap: A Comparison of Procedures Using only Response Prompts with a Procedure Combining Stimulus and Response Prompts,G. A. Riley,"Prompts are assessed in terms of their ability to convey an understanding of the task to those with a severe mental handicap. On this basis, it is suggested that the most effective strategy for teaching manipulative tasks to this group may be one which combines stimulus and response prompts. Two case studies compared this procedure with two procedures which used response prompts only. The combined procedure was more effective in both studies. The studies also demonstrated the applicability of stimulus prompts to teaching the motor, as well as the visual, components of manipulative tasks.",,
-Different DNA-binding specificities of NLP and NIN transcription factors underlie nitrate-induced control of root nodulation,Hanna Nishida,"Abstract Leguminous plants produce nodules for nitrogen fixation; however, nodule production incurs an energy cost. Therefore, as an adaptive strategy, leguminous plants halt root nodule development when sufficient amounts of nitrogen nutrients, such as nitrate, are present in the environment. Although legume NODULE INCEPTION (NIN)-LIKE PROTEIN (NLP) transcription factors have recently been identified, understanding how nodulation is controlled by nitrate, a fundamental question for nitrate-mediated transcriptional regulation of symbiotic genes, remains elusive. Here, we show that two Lotus japonicus NLPs, NITRATE UNRESPONSIVE SYMBIOSIS 1 (NRSYM1)/LjNLP4 and NRSYM2/LjNLP1, have overlapping functions in the nitrate-induced control of nodulation and act as master regulators for nitrate-dependent gene expression. We further identify candidate target genes of LjNLP4 by combining transcriptome analysis with a DNA affinity purification-seq approach. We then demonstrate that LjNLP4 and LjNIN, a key nodulation-specific regulator and paralog of LjNLP4, have different DNA-binding specificities. Moreover, LjNLP4–LjNIN dimerization underlies LjNLP4-mediated bifunctional transcriptional regulation. These data provide a basic principle for how nitrate controls nodulation through positive and negative regulation of symbiotic genes.","{'model': 'tldr@v2.0.0', 'text': 'A basic principle for how nitrate controls nodulation through positive and negative regulation of symbiotic genes is provided, which underlies LjNLP4-mediated bifunctional transcriptional regulation.'}",https://academic.oup.com/plcell/article-pdf/33/7/2340/39739430/koab103.pdf
-Automatically Correcting Large Language Models: Surveying the landscape of diverse self-correction strategies,Liangming Pan,"Large language models (LLMs) have demonstrated remarkable performance across a wide array of NLP tasks. However, their efficacy is undermined by undesired and inconsistent behaviors, including hallucination, unfaithful reasoning, and toxic content. A promising approach to rectify these flaws is self-correction, where the LLM itself is prompted or guided to fix problems in its own output. Techniques leveraging automated feedback -- either produced by the LLM itself or some external system -- are of particular interest as they are a promising way to make LLM-based solutions more practical and deployable with minimal human feedback. This paper presents a comprehensive review of this emerging class of techniques. We analyze and taxonomize a wide array of recent work utilizing these strategies, including training-time, generation-time, and post-hoc correction. We also summarize the major applications of this strategy and conclude by discussing future directions and challenges.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive review of techniques leveraging automated feedback to make LLM-based solutions more practical and deployable with minimal human feedback, including training-time, generation- time, and post-hoc correction.'}",https://arxiv.org/pdf/2308.03188
-Strategies for Bridging the Digital Divide by Enhancing Communication Competence,N. Pokrovskaia,"The digital divide reflects disparities in social status and well-being that depend on access to physical infrastructure, acquired skills in the use of hardware and software, ability to interact in cyberspace. The pandemic shifted all activities to a remote mode, prompting people to enhance their communicative competence. The survey shows a funnel of online-communication choices depending on the criteria of experience and education; higher the intensity of activity and education, narrower the scope and nature of the content for networking, less time spent online.",,
-Playing the lottery with rewards and multiple languages: lottery tickets in RL and NLP,Haonan Yu,"The lottery ticket hypothesis proposes that over-parameterization of deep neural networks (DNNs) aids training by increasing the probability of a ""lucky"" sub-network initialization being present rather than by helping the optimization process (Frankle & Carbin, 2019). Intriguingly, this phenomenon suggests that initialization strategies for DNNs can be improved substantially, but the lottery ticket hypothesis has only previously been tested in the context of supervised learning for natural image tasks. Here, we evaluate whether ""winning ticket"" initializations exist in two different domains: natural language processing (NLP) and reinforcement learning (RL).For NLP, we examined both recurrent LSTM models and large-scale Transformer models (Vaswani et al., 2017). For RL, we analyzed a number of discrete-action space tasks, including both classic control and pixel control. Consistent with workin supervised image classification, we confirm that winning ticket initializations generally outperform parameter-matched random initializations, even at extreme pruning rates for both NLP and RL. Notably, we are able to find winning ticket initializations for Transformers which enable models one-third the size to achieve nearly equivalent performance. Together, these results suggest that the lottery ticket hypothesis is not restricted to supervised learning of natural images, but rather represents a broader phenomenon in DNNs.","{'model': 'tldr@v2.0.0', 'text': 'Evaluating whether ""winning ticket"" initializations exist in NLP and reinforcement learning suggests that the lottery ticket hypothesis is not restricted to supervised learning of natural images, but rather represents a broader phenomenon in DNNs.'}",
-"Generate, Annotate, and Learn: NLP with Synthetic Text",Xuanli He,"Abstract This paper studies the use of language models as a source of synthetic unlabeled text for NLP. We formulate a general framework called “generate, annotate, and learn (GAL)” to take advantage of synthetic text within knowledge distillation, self-training, and few-shot learning applications. To generate high-quality task-specific text, we either fine-tune LMs on inputs from the task of interest, or prompt large LMs with few examples. We use the best available classifier to annotate synthetic text with soft pseudo labels for knowledge distillation and self-training, and use LMs to obtain hard labels for few-shot learning. We train new supervised models on the combination of labeled and pseudo-labeled data, which results in significant gains across several applications. We investigate key components of GAL and present theoretical and empirical arguments against the use of class-conditional LMs to generate synthetic labeled text instead of unlabeled text. GAL achieves new state-of-the-art knowledge distillation results for 6-layer transformers on the GLUE leaderboard.","{'model': 'tldr@v2.0.0', 'text': 'GAL achieves new state-of-the-art knowledge distillation results for 6-layer transformers on the GLUE leaderboard and investigates key components of GAL and presents theoretical and empirical arguments against the use of class-conditional LMs.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00492/2038511/tacl_a_00492.pdf
-Adversarial Attacks on Large Language Model-Based System and Mitigating Strategies: A Case Study on ChatGPT,Bowen Liu,"Machine learning algorithms are at the forefront of the development of advanced information systems. The rapid progress in machine learning technology has enabled cutting-edge large language models (LLMs), represented by GPT-3 and ChatGPT, to perform a wide range of NLP tasks with a stunning performance. However, research on adversarial machine learning highlights the need for these intelligent systems to be more robust. Adversarial machine learning aims to evaluate attack and defense mechanisms to prevent the malicious exploitation of these systems. In the case of ChatGPT, adversarial induction prompt can cause the model to generate toxic texts that could pose serious security risks or propagate false information. To address this challenge, we first analyze the effectiveness of inducing attacks on ChatGPT. Then, two effective mitigating mechanisms are proposed. The first is a training-free prefix prompt mechanism to detect and prevent the generation of toxic texts. The second is a RoBERTa-based mechanism that identifies manipulative or misleading input text via external detection models. The availability of this method is demonstrated through experiments.","{'model': 'tldr@v2.0.0', 'text': 'A training-free prefix prompt mechanism to detect and prevent the generation of toxic texts and a RoBERTa-based mechanism that identifies manipulative or misleading input text via external detection models are proposed.'}",https://downloads.hindawi.com/journals/scn/2023/8691095.pdf
-How Suitable Are Subword Segmentation Strategies for Translating Non-Concatenative Morphology?,Chantal Amrhein,"Data-driven subword segmentation has become the default strategy for open-vocabulary machine translation and other NLP tasks, but may not be sufficiently generic for optimal learning of non-concatenative morphology. We design a test suite to evaluate segmentation strategies on different types of morphological phenomena in a controlled, semi-synthetic setting. In our experiments, we compare how well machine translation models trained on subword- and character-level can translate these morphological phenomena. We find that learning to analyse and generate morphologically complex surface representations is still challenging, especially for non-concatenative morphological phenomena like reduplication or vowel harmony and for rare word stems. Based on our results, we recommend that novel text representation strategies be tested on a range of typologically diverse languages to minimise the risk of adopting a strategy that inadvertently disadvantages certain languages.","{'model': 'tldr@v2.0.0', 'text': 'It is found that learning to analyse and generate morphologically complex surface representations is still challenging, especially for nonconcatenative morphological phenomena like reduplication or vowel harmony and for rare word stems.'}",https://aclanthology.org/2021.findings-emnlp.60.pdf
-Learning to Solve NLP Tasks in an Incremental Number of Languages,Giuseppe Castellucci,"In real scenarios, a multilingual model trained to solve NLP tasks on a set of languages can be required to support new languages over time. Unfortunately, the straightforward retraining on a dataset containing annotated examples for all the languages is both expensive and time-consuming, especially when the number of target languages grows. Moreover, the original annotated material may no longer be available due to storage or business constraints. Re-training only with the new language data will inevitably result in Catastrophic Forgetting of previously acquired knowledge. We propose a Continual Learning strategy that updates a model to support new languages over time, while maintaining consistent results on previously learned languages. We define a Teacher-Student framework where the existing model “teaches” to a student model its knowledge about the languages it supports, while the student is also trained on a new language. We report an experimental evaluation in several tasks including Sentence Classification, Relational Learning and Sequence Labeling.","{'model': 'tldr@v2.0.0', 'text': 'A Continual Learning strategy that updates a model to support new languages over time, while maintaining consistent results on previously learned languages is proposed and an experimental evaluation in several tasks including Sentence Classification, Relational Learning and Sequence Labeling is reported.'}",https://aclanthology.org/2021.acl-short.106.pdf
-Beyond Model Extraction: Imitation Attack for Black-Box NLP APIs,Qiongkai Xu,"Machine-learning-as-a-service (MLaaS) has attracted millions of users to their outperforming sophisticated models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonethe-less, none of the previous stolen models can outperform the original black-box APIs. In this work, we take the first step of showing that attackers could potentially surpass victims via unsupervised domain adaptation and multi-victim ensemble. Extensive experiments on benchmark datasets and real-world APIs vali-date that the imitators can succeed in outperforming the original black-box models. We consider this as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could influence the defense or even publishing strategy of API providers.","{'model': 'tldr@v2.0.0', 'text': 'This work takes the first step of showing that attackers could potentially surpass victims via unsupervised domain adaptation and multi-victim ensemble, and considers this as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could in turnence the defense or even publishing strategy of API providers.'}",
-Visualizing and Understanding Neural Models in NLP,Jiwei Li,"While neural networks have been successfully applied to many NLP tasks the resulting vector-based models are very difficult to interpret. For example it's not clear how they achieve {\em compositionality}, building sentence meaning from the meanings of words and phrases. In this paper we describe four strategies for visualizing compositionality in neural models for NLP, inspired by similar work in computer vision. We first plot unit values to visualize compositionality of negation, intensification, and concessive clauses, allow us to see well-known markedness asymmetries in negation. We then introduce three simple and straightforward methods for visualizing a unit's {\em salience}, the amount it contributes to the final composed meaning: (1) gradient back-propagation, (2) the variance of a token from the average word node, (3) LSTM-style gates that measure information flow. We test our methods on sentiment using simple recurrent nets and LSTMs. Our general-purpose methods may have wide applications for understanding compositionality and other semantic properties of deep networks , and also shed light on why LSTMs outperform simple recurrent nets,","{'model': 'tldr@v2.0.0', 'text': 'Four strategies for visualizing compositionality in neural models for NLP, inspired by similar work in computer vision, including LSTM-style gates that measure information flow and gradient back-propagation, are described.'}",https://www.aclweb.org/anthology/N16-1082.pdf
-Building Representative Corpora from Illiterate Communities: A Reviewof Challenges and Mitigation Strategies for Developing Countries,Stephanie Hirmer,"Most well-established data collection methods currently adopted in NLP depend on the as- sumption of speaker literacy. Consequently, the collected corpora largely fail to represent swathes of the global population, which tend to be some of the most vulnerable and marginalised people in society, and often live in rural developing areas. Such underrepresented groups are thus not only ignored when making modeling and system design decisions, but also prevented from benefiting from development outcomes achieved through data-driven NLP. This paper aims to address the under-representation of illiterate communities in NLP corpora: we identify potential biases and ethical issues that might arise when collecting data from rural communities with high illiteracy rates in Low-Income Countries, and propose a set of practical mitigation strategies to help future work.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the under-representation of illiterate communities in NLP corpora by identifying potential biases and ethical issues that might arise when collecting data from rural communities with high illiteracy rates in Low-Income Countries, and proposing a set of practical mitigation strategies to help future work.'}",
-Visual Grounding Strategies for Text-Only Natural Language Processing,Damien Sileo,"Visual grounding is a promising path toward more robust and accurate Natural Language Processing (NLP) models. Many multimodal extensions of BERT (e.g., VideoBERT, LXMERT, VL-BERT) allow a joint modeling of texts and images that lead to state-of-the-art results on multimodal tasks such as Visual Question Answering. Here, we leverage multimodal modeling for purely textual tasks (language modeling and classification) with the expectation that the multimodal pretraining provides a grounding that can improve text processing accuracy. We propose possible strategies in this respect. A first type of strategy, referred to as transferred grounding consists in applying multimodal models to text-only tasks using a placeholder to replace image input. The second one, which we call associative grounding, harnesses image retrieval to match texts with related images during both pretraining and text-only downstream tasks. We draw further distinctions into both strategies and then compare them according to their impact on language modeling and commonsense-related downstream tasks, showing improvement over text-only baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes two strategies for applying multimodal models to text-only tasks using a placeholder to replace image input and harnesses image retrieval to match texts with related images during both pretraining and text- only downstream tasks.'}",
-Student Surpasses Teacher: Imitation Attack for Black-Box NLP APIs,Qiongkai Xu,"Machine-learning-as-a-service (MLaaS) has attracted millions of users to their splendid large-scale models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonetheless, none of the previous stolen models can outperform the original black-box APIs. In this work, we conduct unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous understanding of model extraction. Extensive experiments on both benchmark datasets and real-world APIs validate that the imitators can succeed in outperforming the original black-box models on transferred domains. We consider our work as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could influence the defense or even publishing strategy of API providers.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous understanding of model extraction.'}",
-Search-Based Algorithm With Scatter Search Strategy for Automated Test Case Generation of NLP Toolkit,Fangqing Liu,"Natural language processing (NLP), as a theory-motivated computational technique, has extensive applications. Automated test case generation based on path coverage, which is a popular structural testing activity, can automatically reveal logic defects that exist in NLP programs and can save testing consumption. NLP programs have many paths that can only be covered by specific input variables. This feature makes conventional search-based algorithm very difficult covering all possible paths in NLP programs. A strategy is required for improving the search ability of search-based algorithms. In this paper, we propose a scatter search strategy to automatically generate test cases for covering all possible paths of NLP programs. The scatter search strategy empowers search-based algorithms to explore all input variables and cover the paths that require specific input variables within a small amount of test cases. The experiment results show that the proposed scatter search strategy can quickly cover the paths, which requires specific input variables. Many test cases and running time consumptions will be saved when search-based algorithms combine with scatter search strategy.","{'model': 'tldr@v2.0.0', 'text': 'A scatter search strategy is proposed to automatically generate test cases for covering all possible paths of NLP programs to empower search-based algorithms to explore all input variables and cover the paths that require specific input variables within a small amount of test cases.'}",
-Evaluating Tuning Strategies for Sequence Generation with Protein Language Models,Andrea Nathansen,"Designing artificial proteins with specialized functions promises new solutions for biological, medical, and environmental use cases. This field benefits from advances in natural language processing, with state-of-the-art text generation models already being successfully applied to protein sequences. Openly available pre-trained protein language models are able to generate artificial protein sequences and can be finetuned on very specific tasks. Considering the high computational cost of finetuning a model exclusively for one downstream task, prompt tuning has been proposed as a more cost-efficient alternative that shares one model across different tasks. However, no openly available implementation of this approach compatible with protein language models has been previously published. Thus, we adapt an open-source codebase designed for NLP models to build a pipeline for prompt tuning on protein sequence data, supporting the protein language models ProtGPT2 and RITA. We benchmark this implementation for generating proteins of a specific family and evaluate the approach using text processing metrics as well as family membership prediction and protein activity prediction of generated sequences. Our results confirm the advantages of prompt tuning in resource usage, especially storage, encouraging further research and expansion of this technique to related use cases. For our evaluated use case, prompt tuning does not reach up to finetuning in terms of the quality of generated protein sequences, indicating the need for more extensive optimization. Lastly, we observe discrepancies between results of similar evaluation tools, highlighting open problems for principled assessment of protein sequence generation quality.","{'model': 'tldr@v2.0.0', 'text': 'An open-source codebase designed for NLP models is adapted to build a pipeline for prompt tuning on protein sequence data, supporting the protein language models ProtGPT2 and RITA, and discrepancies between results of similar evaluation tools are observed, highlighting open problems for principled assessment of protein sequence generation quality.'}",https://www.biorxiv.org/content/biorxiv/early/2023/03/01/2023.02.28.530492.full.pdf
-Robustness Tests of NLP Machine Learning Models: Search and Semantically Replace,Rahul Singh,"This paper proposes a strategy to assess the robustness of different machine learning models that involve natural language processing (NLP). The overall approach relies upon a Search and Semantically Replace strategy that consists of two steps: (1) Search, which identifies important parts in the text; (2) Semantically Replace, which finds replacements for the important parts, and constrains the replaced tokens with semantically similar words. We introduce different types of Search and Semantically Replace methods designed specifically for particular types of machine learning models. We also investigate the effectiveness of this strategy and provide a general framework to assess a variety of machine learning models. Finally, an empirical comparison is provided of robustness performance among three different model types, each with a different text representation.","{'model': 'tldr@v2.0.0', 'text': 'Different types of Search and Semantically Replace methods designed specifically for particular types of machine learning models are introduced, and an empirical comparison is provided of robustness performance among three different model types, each with a different text representation.'}",
-Model-tuning Via Prompts Makes NLP Models Adversarially Robust,Mrigank Raman,"In recent years, NLP practitioners have converged on the following practice: (i) import an off-the-shelf pretrained (masked) language model; (ii) append a multilayer perceptron atop the CLS token's hidden representation (with randomly initialized weights); and (iii) fine-tune the entire model on a downstream task (MLP). This procedure has produced massive gains on standard NLP benchmarks, but these models remain brittle, even to mild adversarial perturbations, such as word-level synonym substitutions. In this work, we demonstrate surprising gains in adversarial robustness enjoyed by Model-tuning Via Prompts (MVP), an alternative method of adapting to downstream tasks. Rather than modifying the model (by appending an MLP head), MVP instead modifies the input (by appending a prompt template). Across three classification datasets, MVP improves performance against adversarial word-level synonym substitutions by an average of 8% over standard methods and even outperforms adversarial training-based state-of-art defenses by 3.5%. By combining MVP with adversarial training, we achieve further improvements in robust accuracy while maintaining clean accuracy. Finally, we conduct ablations to investigate the mechanism underlying these gains. Notably, we find that the main causes of vulnerability of MLP can be attributed to the misalignment between pre-training and fine-tuning tasks, and the randomly initialized MLP parameters. Code is available at https://github.com/acmi-lab/mvp","{'model': 'tldr@v2.0.0', 'text': 'This work demonstrates surprising gains in adversarial robustness enjoyed by Model-tuning Via Prompts (MVP), an alternative method of adapting to downstream tasks that improves performance against adversarial word-level synonym substitutions and outperforms adversarial training-based state-of-art defenses.'}",http://arxiv.org/pdf/2303.07320
-From Humans to Machines: Can ChatGPT-like LLMs Effectively Replace Human Annotators in NLP Tasks?,Surendrabikram Thapa,"The increasing demand for natural language processing (NLP) applications has created a need for large amounts of labeled data to train machine learning models. This has led to using human annotators for tasks such as text classification, sentiment analysis, and named entity recognition. However, human annotation is costly and time-consuming, and the annotation quality can significantly vary depending on the anno-tator. Recent advances in language modeling have led to the development of large language models (LLMs), such as Chat-GPT, which are capable of generating human-like responses to text prompts. In this position paper, we explore the question: whether ChatGPT-like LLMs can effectively replace human annotators in NLP tasks? We discuss the advantages and limitations of using LLMs for annotation and highlight some of the challenges that need to be addressed to make this a feasible approach. We argue that while LLMs can potentially reduce the cost and time required for annotation, they may not be able to fully replace human annotators in all NLP tasks. We conclude by outlining future research directions that could help advance the use of LLMs for NLP annotation.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that while LLMs can potentially reduce the cost and time required for annotation, they may not be able to fully replace human annotators in all NLP tasks.'}",
-Do not Trust the Experts: How the Lack of Standard Complicates NLP for Historical Irish,Oksana Dereza,"In this paper, we describe how we unearthed some fundamental problems while building an analogy dataset modelled on BATS (Gladkova et al., 2016) to evaluate historical Irish embeddings on their ability to detect orthographic, morphological and semantic similarity.The performance of our models in the analogy task was extremely poor regardless of the architecture, hyperparameters and evaluation metrics, while the qualitative evaluation revealed positive tendencies. We argue that low agreement between field experts on fundamental lexical and orthographic issues, and the lack of a unified editorial standard in available resources make it impossible to build reliable evaluation datasets for computational models and obtain interpretable results. We emphasise the need for such a standard, particularly for NLP applications, and prompt Celticists and historical linguists to engage in further discussion. We would also like to draw NLP scholars’ attention to the role of data and its (extra)linguistic properties in testing new models, technologies and evaluation scenarios.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that low agreement between field experts on fundamental lexical and orthographic issues, and the lack of a unified editorial standard in available resources make it impossible to build reliable evaluation datasets for computational models and obtain interpretable results.'}",https://aclanthology.org/2023.insights-1.10.pdf
-Conditionally Adaptive Multi-Task Learning: Improving Transfer Learning in NLP Using Fewer Parameters & Less Data,Jonathan Pilault,"Multi-Task Learning (MTL) has emerged as a promising approach for transferring learned knowledge across different tasks. However, multi-task learning must deal with challenges such as: overfitting to low resource tasks, catastrophic forgetting, and negative task transfer, or learning interference. Additionally, in Natural Language Processing (NLP), MTL alone has typically not reached the performance level possible through per-task fine-tuning of pretrained models. However, many fine-tuning approaches are both parameter inefficient, e.g. potentially involving one new model per task, and highly susceptible to losing knowledge acquired during pretraining. We propose a novel transformer based architecture consisting of a new conditional attention mechanism as well as a set of task conditioned modules that facilitate weight sharing. Through this construction we achieve more efficient parameter sharing and mitigate forgetting by keeping half of the weights of a pretrained model fixed. We also use a new multi-task data sampling strategy to mitigate the negative effects of data imbalance across tasks. Using this approach we are able to surpass single-task fine-tuning methods while being parameter and data efficient. With our base model, we attain 2.2% higher performance compared to a full fine-tuned BERT large model on the GLUE benchmark, adding only 5.6% more trained parameters per task (whereas naive fine-tuning potentially adds 100% of the trained parameters per task) and needing only 64.6% of the data. We show that a larger variant of our single multi-task model approach performs competitively across 26 NLP tasks and yields state-of-the-art results on a number of test and development sets.","{'model': 'tldr@v2.0.0', 'text': 'A novel transformer based architecture consisting of a new conditional attention mechanism as well as a set of task conditioned modules that facilitate weight sharing is proposed that is able to surpass single-task fine-tuning methods while being parameter and data efficient.'}",
-Video Modeling and Prompting: A Comparison of Two Strategies for Teaching Cooking Skills to Students with Mild Intellectual Disabilities,T. Taber-Doughty,"Self-operated video prompting and video modeling was compared when used by three secondary students with mild intellectual disabilities as they completed novel recipes during cooking activities. Alternating between video systems, students completed twelve recipes within their classroom kitchen. An alternating treatment design with a follow-up and withdrawal probe was used to illustrate the effectiveness of both systems on each student's independent task performance. Results indicated increased independence following video system use by all three students with video modeling more effective for two students and video prompting more effective for the third. Future directions for research are presented.",,
-One size does not fit all: Investigating strategies for differentially-private learning across NLP tasks,Manuel Senge,"Preserving privacy in contemporary NLP models allows us to work with sensitive data, but unfortunately comes at a price. We know that stricter privacy guarantees in differentially-private stochastic gradient descent (DP-SGD) generally degrade model performance. However, previous research on the efficiency of DP-SGD in NLP is inconclusive or even counter-intuitive. In this short paper, we provide an extensive analysis of different privacy preserving strategies on seven down-stream datasets in five different ‘typical’ NLP tasks with varying complexity using modern neural models based on BERT and XtremeDistil architectures. We show that unlike standard non-private approaches to solving NLP tasks, where bigger is usually better, privacy-preserving strategies do not exhibit a winning pattern, and each task and privacy regime re-quires a special treatment to achieve adequate performance.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that unlike standard non-private approaches to solving NLP tasks, where bigger is usually better, privacy-preserving strategies do not exhibit a winning pattern, and each task and privacy regime re-quires a special treatment to achieve adequate performance.'}",
-Dice Loss for Data-imbalanced NLP Tasks,Xiaoya Li,"Many NLP tasks such as tagging and machine reading comprehension are faced with the severe data imbalance issue: negative examples significantly outnumber positive examples, and the huge number of easy-negative examples overwhelms the training. The most commonly used cross entropy (CE) criteria is actually an accuracy-oriented objective, and thus creates a discrepancy between training and test: at training time, each training instance contributes equally to the objective function, while at test time F1 score concerns more about positive examples. In this paper, we propose to use dice loss in replacement of the standard cross-entropy objective for data-imbalanced NLP tasks. Dice loss is based on the Sørensen--Dice coefficient or Tversky index , which attaches similar importance to false positives and false negatives, and is more immune to the data-imbalance issue. To further alleviate the dominating influence from easy-negative examples in training, we propose to associate training examples with dynamically adjusted weights to deemphasize easy-negative examples. Theoretical analysis shows that this strategy narrows down the gap between the F1 score in evaluation and the dice loss in training. With the proposed training objective, we observe significant performance boost on a wide range of data imbalanced NLP tasks. Notably, we are able to achieve SOTA results on CTB5, CTB6 and UD1.4 for the part of speech tagging task; SOTA results on CoNLL03, OntoNotes5.0, MSRA and OntoNotes4.0 for the named entity recognition task; along with competitive results on the tasks of machine reading comprehension and paraphrase identification.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to use dice loss in replacement of the standard cross-entropy objective for data-imbalanced NLP tasks, based on the Sørensen--Dice coefficient or Tversky index, which attaches similar importance to false positives and false negatives, and is more immune to the data-IMbalance issue.'}",https://arxiv.org/pdf/1911.02855
-Decoding Strategies for Improving Low-Resource Machine Translation,Chanjun Park,"Pre-processing and post-processing are significant aspects of natural language processing (NLP) application software. Pre-processing in neural machine translation (NMT) includes subword tokenization to alleviate the problem of unknown words, parallel corpus filtering that only filters data suitable for training, and data augmentation to ensure that the corpus contains sufficient content. Post-processing includes automatic post editing and the application of various strategies during decoding in the translation process. Most recent NLP researches are based on the Pretrain-Finetuning Approach (PFA). However, when small and medium-sized organizations with insufficient hardware attempt to provide NLP services, throughput and memory problems often occur. These difficulties increase when utilizing PFA to process low-resource languages, as PFA requires large amounts of data, and the data for low-resource languages are often insufficient. Utilizing the current research premise that NMT model performance can be enhanced through various pre-processing and post-processing strategies without changing the model, we applied various decoding strategies to Korean–English NMT, which relies on a low-resource language pair. Through comparative experiments, we proved that translation performance could be enhanced without changes to the model. We experimentally examined how performance changed in response to beam size changes and n-gram blocking, and whether performance was enhanced when a length penalty was applied. The results showed that various decoding strategies enhance the performance and compare well with previous Korean–English NMT approaches. Therefore, the proposed methodology can improve the performance of NMT models, without the use of PFA; this presents a new perspective for improving machine translation performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed methodology can improve the performance of NMT models, without the use of PFA; this presents a new perspective for improving machine translation performance.'}",https://www.mdpi.com/2079-9292/9/10/1562/pdf?version=1600932781
-Unlink the Link Between COVID-19 and 5G Networks: An NLP and SNA Based Approach,Mohammed Bahja,"Social media facilitates rapid dissemination of information for both factual and fictional information. The spread of non-scientific information through social media platforms such as Twitter has potential to cause damaging consequences. Situations such as the COVID-19 pandemic provides a favourable environment for misinformation to thrive. The upcoming 5G technology is one of the recent victims of misinformation and fake news and has been plagued with misinformation about the effects of its radiation. During the COVID-19 pandemic, conspiracy theories linking the cause of the pandemic to 5G technology have resonated with a section of people leading to outcomes such as destructive attacks on 5G towers. The analysis of the social network data can help to understand the nature of the information being spread and identify the commonly occurring themes in the information. The natural language processing (NLP) and the statistical analysis of the social network data can empower policymakers to understand the misinformation being spread and develop targeted strategies to counter the misinformation. In this paper, NLP based analysis of tweets linking COVID-19 to 5G is presented. NLP models including Latent Dirichlet allocation (LDA), sentiment analysis (SA) and social network analysis (SNA) were applied for the analysis of the tweets and identification of topics. An understanding of the topic frequencies, the inter-relationships between topics and geographical occurrence of the tweets allows identifying agencies and patterns in the spread of misinformation and equips policymakers with knowledge to devise counter-strategies.","{'model': 'tldr@v2.0.0', 'text': 'NLP based analysis of tweets linking COVID-19 to 5G is presented and an understanding of the topic frequencies, the inter-relationships between topics and geographical occurrence of the tweets allows identifying agencies and patterns in the spread of misinformation and equips policymakers with knowledge to devise counter-strategies.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/09262907.pdf
-Adaptation of IDPT System Based on Patient-Authored Text Data using NLP,S. K. Mukhiya,"Background: Internet-Delivered Psychological Treatment (IDPT) systems have the potential to provide evidence-based mental health treatments for a far-reaching population at a lower cost. However, most of the current IDPT systems follow a tunnel-based treatment process and do not adapt to the needs of different patients'. In this paper, we explore the possibility of applying Natural Language Processing (NLP) for personalizing mental health interventions. Objective: The primary objective of this study is to present an adaptive strategy based on NLP techniques that analyses patient-authored text data and extract depression symptoms based on a clinically established assessment questionnaire, PHQ-9. Method: We propose a novel word-embedding (Depression2Vec) to extract depression symptoms from patient authored text data and compare it with three state-of-the-art NLP techniques. We also present an adaptive IDPT system that personalizes treatments for mental health patients based on the proposed depression symptoms detection technique. Result: Our results indicate that the performance of proposed embedding Depression2Vec is comparable to WordNet, but in some cases, the former outperforms the latter with respect to extracting depression symptoms from the patient-authored text. Conclusion: Although the extraction of symptoms from text is challenging, our proposed method can effectively extract depression symptoms from text data, which can be used to deliver personalized intervention.","{'model': 'tldr@v2.0.0', 'text': 'A novel word-embedding to extract depression symptoms from patient authored text data and compare it with three state-of-the-art NLP techniques and indicates that the performance of proposed embedding Depression2Vec is comparable to WordNet, but in some cases, the former outperforms the latter with respect to extracting depression symptoms.'}",
-"Internet Data Analysis Methodology for Cyberterrorism Vocabulary Detection, Combining Techniques of Big Data Analytics, NLP and Semantic Web",Iván Castillo-Zúñiga,"This article presents a methodology for the analysis of data on the Internet, combining techniques of Big Data analytics, NLP and semantic web in order to find knowledge about large amounts of information on the web. To test the effectiveness of the proposed method, webpages about cyberterrorism were analyzed as a case study. The procedure implemented a genetic strategy in parallel, which integrates (Crawler to locate and download information from the web; to retrieve the vocabulary, using techniques of NLP (tokenization, stop word, TF, TFIDF), methods of stemming and synonyms). For the pursuit of knowledge was built a dataset through the description of a linguistic corpus with semantic ontologies, considering the characteristics of cyber-terrorism, which was analyzed with the algorithms, Random Forests (parallel), Boosting, SVM, neural network, K-nn and Bayes. The results reveal a percentage of the 95.62% accuracy in the detection of the vocabulary of cyber-terrorism, which were approved through cross validation, reaching 576% time savings with parallel processing.","{'model': 'tldr@v2.0.0', 'text': 'A methodology for the analysis of data on the Internet, combining techniques of Big Data analytics, NLP and semantic web in order to find knowledge about large amounts of information on the web, reaching 576% time savings with parallel processing.'}",
-A hybrid LP/NLP paradigm for global optimization relaxations,Aida Khajavirad,,"{'model': 'tldr@v2.0.0', 'text': 'Results show that incorporating the proposed techniques in the BARON software leads to significant reductions in execution time, and increases by 30% the number of problems that are solvable to global optimality within 500 s on a standard workstation.'}",
-Interacting TCP and NLP transcription factors control plant responses to nitrate availability,Peizhu Guan,"Significance Nitrate is an essential nutrient and a critical signal for plant growth, development, and stress responses. Nitrate signaling underlies a myriad of physiological, morphological, and developmental processes. Here we report that interacting teosinte branched1/cycloidea/proliferating cell factor and NIN-like protein transcription factors constitute a molecular link between nitrate signaling and the control of the cell-cycle progression gene CYCB1;1 and root meristem growth. Our findings shed light on the regulatory mechanisms underlying an important plant adaptive process for coping with and surviving environmental challenges. Plants have evolved adaptive strategies that involve transcriptional networks to cope with and survive environmental challenges. Key transcriptional regulators that mediate responses to environmental fluctuations in nitrate have been identified; however, little is known about how these regulators interact to orchestrate nitrogen (N) responses and cell-cycle regulation. Here we report that teosinte branched1/cycloidea/proliferating cell factor1-20 (TCP20) and NIN-like protein (NLP) transcription factors NLP6 and NLP7, which act as activators of nitrate assimilatory genes, bind to adjacent sites in the upstream promoter region of the nitrate reductase gene, NIA1, and physically interact under continuous nitrate and N-starvation conditions. Regions of these proteins necessary for these interactions were found to include the type I/II Phox and Bem1p (PB1) domains of NLP6&7, a protein-interaction module conserved in animals for nutrient signaling, and the histidine- and glutamine-rich domain of TCP20, which is conserved across plant species. Under N starvation, TCP20-NLP6&7 heterodimers accumulate in the nucleus, and this coincides with TCP20 and NLP6&7-dependent up-regulation of nitrate assimilation and signaling genes and down-regulation of the G2/M cell-cycle marker gene, CYCB1;1. TCP20 and NLP6&7 also support root meristem growth under N starvation. These findings provide insights into how plants coordinate responses to nitrate availability, linking nitrate assimilation and signaling with cell-cycle progression.","{'model': 'tldr@v2.0.0', 'text': 'Insight is provided into how plants coordinate responses to nitrate availability, linking nitrate assimilation and signaling with cell-cycle progression and root meristem growth.'}",https://www.pnas.org/content/pnas/114/9/2419.full.pdf
-Understanding the Effect of In-Video Prompting on Learners and Instructors,Hyungyu Shin,"Online instructional videos are ubiquitous, but it is difficult for instructors to gauge learners' experience and their level of comprehension or confusion regarding the lecture video. Moreover, learners watching the videos may become disengaged or fail to reflect and construct their own understanding. This paper explores instructor and learner perceptions of in-video prompting where learners answer reflective questions while watching videos. We conducted two studies with crowd workers to understand the effect of prompting in general, and the effect of different prompting strategies on both learners and instructors. Results show that some learners found prompts to be useful checkpoints for reflection, while others found them distracting. Instructors reported the collected responses to be generally more specific than what they have usually collected. Also, different prompting strategies had different effects on the learning experience and the usefulness of responses as feedback.","{'model': 'tldr@v2.0.0', 'text': 'Examination of instructor and learner perceptions of in-video prompting where learners answer reflective questions while watching videos shows that some learners found prompts to be useful checkpoints for reflection, while others found them distracting.'}",
-SpaML: a Bimodal Ensemble Learning Spam Detector based on NLP Techniques,Jaouhar Fattahi,"In this paper, we put forward a new tool, called SpaML, for spam detection using a set of supervised and unsupervised classifiers, and two techniques imbued with Natural Language Processing (NLP), namely Bag of Words (BoW) and Term Frequency-Inverse Document Frequency (TF-IDF). We first present the NLP techniques used. Then, we present our classifiers and their performance on each of these techniques. Then, we present our overall Ensemble Learning classifier and the strategy we are using to combine them. Finally, we present the interesting results shown by SpaML in terms of accuracy and precision.","{'model': 'tldr@v2.0.0', 'text': 'A new tool, called SpaML, for spam detection using a set of supervised and unsupervised classifiers, and two techniques imbued with Natural Language Processing (NLP), namely Bag of Words (BoW) and Term Frequency-Inverse Document Frequency (TF-IDF).'}",https://arxiv.org/pdf/2010.07444
-Analyzing Student Strategies In Blended Courses Using Clickstream Data,Nil-Jana Akpinar,"Educational software data promises unique insights into students' study behaviors and drivers of success. While much work has been dedicated to performance prediction in massive open online courses, it is unclear if the same methods can be applied to blended courses and a deeper understanding of student strategies is often missing. We use pattern mining and models borrowed from Natural Language Processing (NLP) to understand student interactions and extract frequent strategies from a blended college course. Fine-grained clickstream data is collected through Diderot, a non-commercial educational support system that spans a wide range of functionalities. We find that interaction patterns differ considerably based on the assessment type students are preparing for, and many of the extracted features can be used for reliable performance prediction. Our results suggest that the proposed hybrid NLP methods can provide valuable insights even in the low-data setting of blended courses given enough data granularity.","{'model': 'tldr@v2.0.0', 'text': 'This work uses pattern mining and models borrowed from Natural Language Processing to understand student interactions and extract frequent strategies from a blended college course, and suggests that the proposed hybrid NLP methods can provide valuable insights even in the low-data setting of blended courses given enough data granularity.'}",
-Prompting undergraduate students’ metacognition of learning: implementing ‘meta-learning’ assessment tasks in the biomedical sciences,Kay Colthorpe,"Abstract To succeed at post-secondary education, it’s essential that students develop an understanding of their own knowledge and learning processes. This metacognition of learning, or ‘meta-learning’, helps students to become more effective learners, as they become more aware of their self-regulatory processes and recognise the effectiveness of their study strategies. To increase biomedical science students’ self-awareness, we have designed and implemented meta-learning assessment tasks across our biomedical science courses. Most students reported that meta-learning tasks had a positive impact on their learning, as they prompted self-regulatory processes of forethought and self-reflection. We found that students were equally likely to change or not change their study strategies across subsequent semesters. Those students that did not change were generally high achievers, believing their study approaches were most effective, but their performance did not improve across semesters. In contrast, students who adapted, mostly by modifying how they appraised and rearranged records or improved planning and time management, performed less well overall but significantly improved their performance across semesters. Meta-learning tasks may prompt students to become more self-reflective and independent learners by affecting their approach to learning, enabling them to reflect on their study strategies, adapt and improve performance, and may enable the development of lifelong learning skills.",,
-The Use of Response Prompting and Frames for Teaching Sentence Writing to Students With Moderate Intellectual Disability,R. Pennington,"In the current study, we examined the effects of response prompting strategies (i.e., constant time delay, system of least prompts) and frames on sentence writing for three participants, ages 7 to 12, with moderate intellectual disability. We used a concurrent multiple probe across behaviors design to evaluate the efficacy of the intervention package and posttest probes to assess generalized responding to untrained stimulation. During intervention, the teacher taught two students to construct sentences using selection-based software and another to generate handwritten responses across three different writing frames (i.e., I want _________, I see _____, The _____ is ______). Our findings suggest that the package was effective and produced variable levels of maintenance and generalized responding for all three participants.",,
-Improving Robustness and Generality of NLP Models Using Disentangled Representations,Jiawei Wu,"Supervised neural networks, which first map an input $x$ to a single representation $z$, and then map $z$ to the output label $y$, have achieved remarkable success in a wide range of natural language processing (NLP) tasks. Despite their success, neural models lack for both robustness and generality: small perturbations to inputs can result in absolutely different outputs; the performance of a model trained on one domain drops drastically when tested on another domain. In this paper, we present methods to improve robustness and generality of NLP models from the standpoint of disentangled representation learning. Instead of mapping $x$ to a single representation $z$, the proposed strategy maps $x$ to a set of representations $\{z_1,z_2,...,z_K\}$ while forcing them to be disentangled. These representations are then mapped to different logits $l$s, the ensemble of which is used to make the final prediction $y$. We propose different methods to incorporate this idea into currently widely-used models, including adding an $L$2 regularizer on $z$s or adding Total Correlation (TC) under the framework of variational information bottleneck (VIB). We show that models trained with the proposed criteria provide better robustness and domain adaptation ability in a wide range of supervised learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Methods to improve robustness and generality of NLP models from the standpoint of disentangled representation learning are presented and it is shown that models trained with the proposed criteria provide better robusts and domain adaptation ability in a wide range of supervised learning tasks.'}",
-"Complementarity, F-score, and NLP Evaluation",Leon Derczynski,"This paper addresses the problem of quantifying the differences between entity extraction systems, where in general only a small proportion a document should be selected. Comparing overall accuracy is not very useful in these cases, as small differences in accuracy may correspond to huge differences in selections over the target minority class. Conventionally, one may use per-token complementarity to describe these differences, but it is not very useful when the set is heavily skewed. In such situations, which are common in information retrieval and entity recognition, metrics like precision and recall are typically used to describe performance. However, precision and recall fail to describe the differences between sets of objects selected by different decision strategies, instead just describing the proportional amount of correct and incorrect objects selected. This paper presents a method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches.","{'model': 'tldr@v2.0.0', 'text': 'A method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches is presented.'}",
-Effects of a Progressive Prompting-based Educational Game on Second Graders' Mathematics Learning Performance and Behavioral Patterns,Kai-Hsiang Yang,"Game-based learning (GBL) has been proven to be an attractive learning model by many studies; however, scholars have pointed out that the effectiveness of game-based learning could be limited if proper learning strategies are not incorporated. Prompting is a strategy that plays the important role of providing hints and guidance in interactive learning environments. Therefore, this study proposes a game-based learning approach with a progressive prompting strategy, using different levels of hints to guide students to complete tasks and achieve learning goals. Quasi-experimental research was employed in this study using two groups of students. The experimental group learned with the proposed approach, while the control group was allocated a conventional game-based learning strategy. The findings of the study show that the proposed approach significantly improved the second graders’ mathematics learning achievement. From the analysis of the experimental students’ learning behaviors, it was reported that the experimental group students could generate the answer after two progressive prompts. Thus, we could see that the proposed approach could enhance the learning achievement of the experimental group students by correctly guiding them to answer questions, step up their thinking, and understand the learning content in the learning process.","{'model': 'tldr@v2.0.0', 'text': 'The proposed approach with a progressive prompting strategy could enhance the learning achievement of the experimental group students by correctly guiding them to answer questions, step up their thinking, and understand the learning content in the learning process.'}",
-MultiMix: A Robust Data Augmentation Strategy for Cross-Lingual NLP,M Saiful Bari,"Transfer learning has yielded state-of-the-art results in many supervised natural language processing tasks. However, annotated data for every target task in every target language is rare, especially for low-resource languages. In this work, we propose MultiMix, a novel data augmentation method for semi-supervised learning in zero-shot transfer learning scenarios. In particular, MultiMix targets to solve cross-lingual adaptation problems from a source (language) distribution to an unknown target (language) distribution assuming it has no training labels in the target language task. In its heart, MultiMix performs simultaneous self-training with data augmentation and unsupervised sample selection. To show its effectiveness, we have performed extensive experiments on zero-shot transfers for cross-lingual named entity recognition (XNER) and natural language inference (XNLI). Our experiments show sizeable improvements in both tasks outperforming the baselines by a good margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes MultiMix, a novel data augmentation method for semi-supervised learning in zero-shot transfer learning scenarios that targets to solve cross-lingual adaptation problems from a source ( language) distribution to an unknown target (language) distribution assuming it has no training labels in the target language task.'}",
-A Just and Comprehensive Strategy for Using NLP to Address Online Abuse,David Jurgens,"Online abusive behavior affects millions and the NLP community has attempted to mitigate this problem by developing technologies to detect abuse. However, current methods have largely focused on a narrow definition of abuse to detriment of victims who seek both validation and solutions. In this position paper, we argue that the community needs to make three substantive changes: (1) expanding our scope of problems to tackle both more subtle and more serious forms of abuse, (2) developing proactive technologies that counter or inhibit abuse before it harms, and (3) reframing our effort within a framework of justice to promote healthy communities.","{'model': 'tldr@v2.0.0', 'text': 'This position paper argues that the NLP community needs to make three substantive changes: expanding the scope of problems to tackle both more subtle and more serious forms of abuse, developing proactive technologies that counter or inhibit abuse before it harms, and reframing the effort within a framework of justice to promote healthy communities.'}",https://www.aclweb.org/anthology/P19-1357.pdf
-"Breaking NLP: Using Morphosyntax, Semantics, Pragmatics and World Knowledge to Fool Sentiment Analysis Systems",Taylor Mahler,"This paper describes our “breaker” submission to the 2017 EMNLP “Build It Break It” shared task on sentiment analysis. In order to cause the “builder” systems to make incorrect predictions, we edited items in the blind test data according to linguistically interpretable strategies that allow us to assess the ease with which the builder systems learn various components of linguistic structure. On the whole, our submitted pairs break all systems at a high rate (72.6%), indicating that sentiment analysis as an NLP task may still have a lot of ground to cover. Of the breaker strategies that we consider, we find our semantic and pragmatic manipulations to pose the most substantial difficulties for the builder systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the “breaker” submission to the 2017 EMNLP “Build It Break It” shared task on sentiment analysis, and finds the semantic and pragmatic manipulations to pose the most substantial difficulties for the builder systems.'}",https://www.aclweb.org/anthology/W17-5405.pdf
-Framing and Agenda-setting in Russian News: a Computational Analysis of Intricate Political Strategies,Anjalie Field,"Amidst growing concern over media manipulation, NLP attention has focused on overt strategies like censorship and “fake news”. Here, we draw on two concepts from political science literature to explore subtler strategies for government media manipulation: agenda-setting (selecting what topics to cover) and framing (deciding how topics are covered). We analyze 13 years (100K articles) of the Russian newspaper Izvestia and identify a strategy of distraction: articles mention the U.S. more frequently in the month directly following an economic downturn in Russia. We introduce embedding-based methods for cross-lingually projecting English frames to Russian, and discover that these articles emphasize U.S. moral failings and threats to the U.S. Our work offers new ways to identify subtle media manipulation strategies at the intersection of agenda-setting and framing.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces embedding-based methods for cross-lingually projecting English frames to Russian, and offers new ways to identify subtle media manipulation strategies at the intersection of agenda-setting and framing.'}",https://www.aclweb.org/anthology/D18-1393.pdf
-Role of Entrepreneurial Orientation in SMEs Global Performance: Testing Marketing Strategies and Technological Orientation as Mediators,S. Aslam,"Entrepreneurial orientation (EO) plays a key role in a firm’s success, particularly in the context of globalization. Thus, EO has gained prompt attention because of an increasing trend of globalized markets and economies. This article expands the understanding of Entrepreneurial Orientation as it studies EO to assess Small and Medium Enterprise (SMEs) global performance. The study extends the knowledge of SMEs entering global markets by considering firms’ marketing strategy and technological orientation as mediators. Through these mechanisms, corporate EO leads to cross-border excellence. The findings of this study, through a sample of 155 international SMEs in Pakistan, support the impact of EO on global performance whereas the marketing strategy and technological orientation serve as the mediators in this core relationship. The data were analyzed using a variance-based structural equation model (SEM) in Smart PLS software. The findings validate that SMEs, more concentrating on EO, are more able to explore new opportunities in global markets, but their small size, inadequate access to different resources and limited experience in operating in international markets prevents them from taking full advantage of arising opportunities. Hence the findings confirm the significance of EO in the globalization of SMEs however, it was also observed that although the marketing strategy and technological orientation may reduce the direct impact of EO on international performance, these mechanisms accelerate the indirect impact of EO as a tangible and intangible resource. ",,https://coralpublications.org/index.php/jms/article/download/157/116
-IIE-NLP-NUT at SemEval-2020 Task 4: Guiding PLM with Prompt Template Reconstruction Strategy for ComVE,Luxi Xing,"This paper introduces our systems for the first two subtasks of SemEval Task4: Commonsense Validation and Explanation. To clarify the intention for judgment and inject contrastive information for selection, we propose the input reconstruction strategy with prompt templates. Specifically, we formalize the subtasks into the multiple-choice question answering format and construct the input with the prompt templates, then, the final prediction of question answering is considered as the result of subtasks. Experimental results show that our approaches achieve significant performance compared with the baseline systems. Our approaches secure the third rank on both official test sets of the first two subtasks with an accuracy of 96.4 and an accuracy of 94.3 respectively.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the systems for the first two subtasks of SemEval Task4: Commonsense Validation and Explanation and proposes the input reconstruction strategy with prompt templates, which formalizes the subtasks into the multiple-choice question answering format and construct the input with the prompt templates.'}",https://aclanthology.org/2020.semeval-1.42.pdf
-A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks,Kazuma Hashimoto,"Transfer and multi-task learning have traditionally focused on either a single source-target pair or very few, similar tasks. Ideally, the linguistic levels of morphology, syntax and semantics would benefit each other by being trained in a single model. We introduce a joint many-task model together with a strategy for successively growing its depth to solve increasingly complex tasks. Higher layers include shortcut connections to lower-level task predictions to reflect linguistic hierarchies. We use a simple regularization term to allow for optimizing all model weights to improve one task’s loss without exhibiting catastrophic interference of the other tasks. Our single end-to-end model obtains state-of-the-art or competitive results on five different tasks from tagging, parsing, relatedness, and entailment tasks.","{'model': 'tldr@v2.0.0', 'text': 'A joint many-task model together with a strategy for successively growing its depth to solve increasingly complex tasks and uses a simple regularization term to allow for optimizing all model weights to improve one task’s loss without exhibiting catastrophic interference of the other tasks.'}",https://www.aclweb.org/anthology/D17-1206.pdf
-When Low Resource NLP Meets Unsupervised Language Model: Meta-Pretraining then Meta-Learning for Few-Shot Text Classification (Student Abstract),Shumin Deng,"Text classification tends to be difficult when data are deficient or when it is required to adapt to unseen classes. In such challenging scenarios, recent studies have often used meta-learning to simulate the few-shot task, thus negating implicit common linguistic features across tasks. This paper addresses such problems using meta-learning and unsupervised language models. Our approach is based on the insight that having a good generalization from a few examples relies on both a generic model initialization and an effective strategy for adapting this model to newly arising tasks. We show that our approach is not only simple but also produces a state-of-the-art performance on a well-studied sentiment classification dataset. It can thus be further suggested that pretraining could be a promising solution for few-shot learning of many other NLP tasks. The code and the dataset to replicate the experiments are made available at https://github.com/zxlzr/FewShotNLP.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses text classification problems using meta-learning and unsupervised language models using pretraining and shows that this approach is not only simple but also produces a state-of-the-art performance on a well-studied sentiment classification dataset.'}",https://ojs.aaai.org/index.php/AAAI/article/download/7158/7012
-"Implementing NLP projects for noncentral languages: instructions for funding bodies, strategies for developers",O. Streiter,,"{'model': 'tldr@v2.0.0', 'text': 'The concepts of free software and software pools are elaborate and it is argued that NCLPs should embrace an open-source approach for the resources they develop and pool these resources together with other similaropen-source resources.'}",
-A Comparison of Static Picture Prompting and Video Prompting Simulation Strategies Using Group Instructional Procedures,David F. Cihak,Two groups of three students with moderate mental retardation were instructed using group procedures to compare static picture and video prompting simulation strategies. An alternating treatments design was used to compare individual student task acquisition and maintenance performances of purchasing and banking skills. The results indicated that these two strategies were equally effective and efficient. Individual differences suggest that simulation strategy materials benefit student learning by displaying relevant stimuli for a stationary period time during presentation.,,
-Prompting All Students to Learn: Examining Dynamic Assessment of Special Needs and Typical Students in a Prekindergarten Inclusive French Program.,Michele Regalla,"This study examined a teacher's prompting strategies and the use of dynamic assessment (DA) in an inclusive prekindergarten French program. Prior research has shown that DA is an effective method to assess both foreign language learning and first language development for typically developing students and for students with special needs, as well as for identifying bilingual students with language impairments. Data were collected from videorecordings of class sessions, DA scores, and a questionnaire on which the French teacher reported her classroom prompting strategies and predicted students’ needs for prompting on the DA. Scores on the DA and the French teacher's predictions of student prompting needs were analyzed through independent t tests. In addition, 30 videorecordings of class sessions were analyzed for prompting strategies. The results indicated that DA can provide insights about the learning of typically developing students and special needs students in a prekindergarten French program. Although students with special needs required both more prompting and more practice time than typically developing students, all special needs students in this study were able to provide at least one second language response during the DA.",,
-Representation and control strategies for large knowledge domains: An Application to NLP,F. Antonacci,"The design issues encountered during the development of a natural language processor (NLP) for the Italian language are described. The focus is on strategic aspects, namely representation and control, and their implementation with first-order logic. The complexity and the size of the knowledge domain (press agency releases on finance and economics) do not present severe restrictions in the sentence structure; hence a considerable design effort for data structures and control algorithms was required. Logic proved to be an important tool for implementing in a modular and efficient way the knowledge sources along with the programs that derive the morphologic, syntactic, and semantic features of sentences. As for the data structures, we found a considerable advantage in separating linguistic knowledge in three sources: morphologic, syntactic, and semantic. This resulted in a clear and systematic representation scheme and reduced the complexity of the parsing system.","{'model': 'tldr@v2.0.0', 'text': 'The design issues encountered during the development of a natural language processor for the Italian language are described and Logic proved to be an important tool for implementing in a modular and efficient way the knowledge sources along with the programs that derive the morphologic, syntactic, and semantic features of sentences.'}",
-Embedding Strategies for Specialized Domains: Application to Clinical Entity Recognition,Hicham El Boukkouri,"Using pre-trained word embeddings in conjunction with Deep Learning models has become the “de facto” approach in Natural Language Processing (NLP). While this usually yields satisfactory results, off-the-shelf word embeddings tend to perform poorly on texts from specialized domains such as clinical reports. Moreover, training specialized word representations from scratch is often either impossible or ineffective due to the lack of large enough in-domain data. In this work, we focus on the clinical domain for which we study embedding strategies that rely on general-domain resources only. We show that by combining off-the-shelf contextual embeddings (ELMo) with static word2vec embeddings trained on a small in-domain corpus built from the task data, we manage to reach and sometimes outperform representations learned from a large corpus in the medical domain.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that by combining off-the-shelf contextual embeddings (ELMo) with static word2vecembeddings trained on a small in-domain corpus built from the task data, they manage to reach and sometimes outperform representations learned from a large corpus in the medical domain.'}",https://www.aclweb.org/anthology/P19-2041.pdf
-"SIMPLE, AFFORDABLE, AND EFFECTIVE STRATEGIES FOR PROMPTING READING BEHAVIOR",Sara C. Ernsbarger,"Many students with disabilities benefit from prompts for reading in general education classrooms. Visual, auditory, kinesthetic, and tactile prompts implemented by the teacher, by the student with minimal teacher assistance, or by the student independently can facilitate more effective reading. Prompting strategies provide clues and help students remember to perform a certain behavior, self-monitor their behavior, and focus their attention. Examples of prompts for each type of modality and level of implementation are provided.",,
-A bilevel NLP sensitivity‐based decomposition for dynamic optimization with moving finite elements,Weifeng Chen,"Optimal control has guided numerous applications in chemical engineering, and exact determination of optimal profiles is essential for operation of separation and reactive processes, and operating strategies and recipe generation for batch processes. Here, a simultaneous collocation formulation based on moving finite elements is developed for the solution of a class of optimal control problems. Novel features of the algorithm include the direct location of breakpoints for control profiles and a termination criterion based on a constant Hamiltonian profile. The algorithm is stabilized and performance is significantly improved by decomposing the overall nonlinear programming (NLP) formulation into an inner problem, which solves a fixed element simultaneous collocation problem, and an outer problem, which adjusts the finite elements based on several error criteria. This bilevel formulation is aided by a NLP solver (the interior point optimizer) for both problems as well as an NLP sensitivity component, which provides derivative information from the inner problem to the outer problem. This approach is demonstrated on 11 dynamic optimization problems drawn from the optimal control and chemical engineering literature. © 2014 American Institute of Chemical Engineers AIChE J, 60: 966–979, 2014",,
-Reducing unnecessary urinary catheter use and other strategies to prevent catheter-associated urinary tract infection: an integrative review,J. Meddings,"Background Catheter-associated urinary tract infections (CAUTI) are costly, common and often preventable by reducing unnecessary urinary catheter (UC) use. Methods To summarise interventions to reduce UC use and CAUTIs, we updated a prior systematic review (through October 2012), and a meta-analysis regarding interventions prompting UC removal by reminders or stop orders. A narrative review summarises other CAUTI prevention strategies including aseptic insertion, catheter maintenance, antimicrobial UCs, and bladder bundle implementation. Results 30 studies were identified and summarised with interventions to prompt removal of UCs, with potential for inclusion in the meta-analyses. By meta-analysis (11 studies), the rate of CAUTI (episodes per 1000 catheter-days) was reduced by 53% (rate ratio 0.47; 95% CI 0.30 to 0.64, p<0.001) using a reminder or stop order, with five studies also including interventions to decrease initial UC placement. The pooled (nine studies) standardised mean difference (SMD) in catheterisation duration (days) was −1.06 overall (p=0.065) including a statistically significant decrease in stop-order studies (SMD −0.37; p<0.001) but not in reminder studies (SMD, −1.54; p=0.071). No significant harm from catheter removal strategies is supported. Limited research is available regarding the impact of UC insertion and maintenance technique. A recent randomised controlled trial indicates antimicrobial catheters provide no significant benefit in preventing symptomatic CAUTIs. Conclusions UC reminders and stop orders appear to reduce CAUTI rates and should be used to improve patient safety. Several evidence-based guidelines have evaluated CAUTI preventive strategies as well as emerging evidence regarding intervention bundles. Implementation strategies are important because reducing UC use involves changing well-established habits.","{'model': 'tldr@v2.0.0', 'text': 'UC reminders and stop orders appear to reduce CAUTI rates and should be used to improve patient safety, as well as emerging evidence regarding intervention bundles.'}",https://qualitysafety.bmj.com/content/qhc/23/4/277.full.pdf
-Considering student choice when selecting instructional strategies: a comparison of three prompting systems.,T. Taber-Doughty,,"{'model': 'tldr@v2.0.0', 'text': 'This study found that the preferred prompting systems were more effective in terms of both skill acquisition and duration for completing tasks for students.'}",
-Self-Directed Video Prompting and Least-to-Most Prompting: Examining Ways of Increasing Vocational Skill Acquisition Among Students with Autism Spectrum Disorder and Intellectual Disability,Gulnoza Yakubova,,"{'model': 'tldr@v2.0.0', 'text': 'VP and LMP as a combined intervention can be effective in teaching vocational tasks that involve process steps to students with both autism spectrum disorder and intellectual disability when working on school-based vocational tasks.'}",
-News Summarization and Evaluation in the Era of GPT-3,Tanya Goyal,"The recent success of prompting large language models like GPT-3 has led to a paradigm shift in NLP research. In this paper, we study its impact on text summarization, focusing on the classic benchmark domain of news summarization. First, we investigate how GPT-3 compares against fine-tuned models trained on large summarization datasets. We show that not only do humans overwhelmingly prefer GPT-3 summaries, prompted using only a task description, but these also do not suffer from common dataset-specific issues such as poor factuality. Next, we study what this means for evaluation, particularly the role of gold standard test sets. Our experiments show that both reference-based and reference-free automatic metrics cannot reliably evaluate GPT-3 summaries. Finally, we evaluate models on a setting beyond generic summarization, specifically keyword-based summarization, and show how dominant fine-tuning approaches compare to prompting. To support further research, we release: (a) a corpus of 10K generated summaries from fine-tuned and prompt-based models across 4 standard summarization benchmarks, (b) 1K human preference judgments comparing different systems for generic- and keyword-based summarization.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that not only do humans overwhelmingly prefer GPT-3 summaries, prompted using only a task description, but these also do not suffer from common dataset-specific issues such as poor factuality.'}",http://arxiv.org/pdf/2209.12356
-Torque Distribution Strategies for Energy-Efficient Electric Vehicles With Multiple Drivetrains,B. Lenzo,"The paper discusses novel computationally efficient torque distribution strategies for electric vehicles with individually controlled drivetrains, aimed at minimizing the overall power losses while providing the required level of wheel torque and yaw moment. Analytical solutions of the torque control allocation problem are derived and effects of load transfers due to driving/braking and cornering are studied and discussed in detail. Influences of different drivetrain characteristics on the front and rear axles are described. The results of an analytically derived algorithm are contrasted with those from two other control allocation strategies, based on the offline numerical solution of more detailed formulations of the control allocation problem (i.e., a multiparametric nonlinear programming (mp-NLP) problem). The control allocation algorithms are experimentally validated with an electric vehicle with four identical drivetrains along multiple driving cycles and in steady-state cornering. The experiments show that the computationally efficient algorithms represent a very good compromise between low energy consumption and controller complexity.",,http://shura.shu.ac.uk/15643/1/Lenzo%20-%20Torque%20distribution%20%28AM%29%20%2B%20coversheet.pdf
-A Progressive Prompting Approach to Conducting Context-Aware Learning Activities for Natural Science Courses,Chih-Hung Chen,"Owing to the popularity of mobile, wireless communication and sensing technologies, issues related to contextual mobile learning have widely been discussed in recent years. In the meantime, researchers have indicated the importance of developing effective strategies for guiding students to learn in context-aware learning environments. In this study, a progressive prompt-based context-aware learning approach is proposed to improve the learning performance of students. An experiment was conducted on a natural science course of an elementary school to evaluate the performance of the proposed approach. From the experimental results, it is found that the proposed approach could effectively enhance the learning achievement of the students in comparison with the conventional context-aware learning system with single-stage prompts. It is also interesting to find that the proposed approach, by providing more challenging tasks, encouraged the students to put more effort into examining the contextual information and interpreting the learning content.","{'model': 'tldr@v2.0.0', 'text': 'A progressive prompt-based context-aware learning approach is proposed to improve the learning performance of students and it is found that the proposed approach could effectively enhance the learning achievement of the students in comparison with the conventional context- Aware learning system with single-stage prompts.'}",
-OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalization,S. Iyer,"Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding of the performance trade-offs of different decisions made during the instruction-tuning process. These decisions include the scale and diversity of the instruction-tuning benchmark, different task sampling strategies, fine-tuning with and without demonstrations, training using specialized datasets for reasoning and dialogue, and finally, the fine-tuning objectives themselves. In this paper, we characterize the effect of instruction-tuning decisions on downstream task performance when scaling both model and benchmark sizes. To this end, we create OPT-IML Bench: a large benchmark for Instruction Meta-Learning (IML) of 2000 NLP tasks consolidated into task categories from 8 existing benchmarks, and prepare an evaluation framework to measure three types of model generalizations: to tasks from fully held-out categories, to held-out tasks from seen categories, and to held-out instances from seen tasks. Through the lens of this framework, we first present insights about instruction-tuning decisions as applied to OPT-30B and further exploit these insights to train OPT-IML 30B and 175B, which are instruction-tuned versions of OPT. OPT-IML demonstrates all three generalization abilities at both scales on four different evaluation benchmarks with diverse tasks and input formats -- PromptSource, FLAN, Super-NaturalInstructions, and UnifiedSKG. Not only does it significantly outperform OPT on all benchmarks but is also highly competitive with existing models fine-tuned on each specific benchmark. We release OPT-IML at both scales, together with the OPT-IML Bench evaluation framework.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the effect of instruction-tuning decisions on downstream task performance when scaling both model and benchmark sizes and releases OPT-IML at both scales, together with an evaluation framework to measure three types of model generalizations.'}",http://arxiv.org/pdf/2212.12017
-The Effect of NLP (Accelerated Learning) on Iranian EFL Learner’s Listening Comprehension,Chnour Khalandi,"the present study was an attempt to investigate the impact of Neuro-Linguistic Programming (NLP) strategies and specifically Accelerated Learning on Iranian EFL learner's listening comprehension and detailed listening. To fulfill the purpose of this study, a total number of 30 Iranian EFL learners studying as Cambridge ESOL KET students were selected through stratified sampling procedure. They were divided in two control and experimental groups with 10 and 20 learners in each group respectively. Each group took a VAK questionnaire, a pre –test and a treatment process of 12 sessions and a posttest. In the experimental group the teacher conducted a set of short videos and mixed methods based on Accelerated Learning as one of neuro linguistic programming techniques. In control group a set of limited short videos conducted. On the last session, the participants of both groups took a post- test to measure their achievement of listening comprehension and detailed listening. This research suggests that accelerated learning is much more dynamic and has a significant effect on listening comprehension, detailed listening and, different learning styles, which hypothesizes that all VAK learner's language processing follow one another in a strictly accelerated manner. Statistically, the effect of accelerated learning on visual, auditory and kinesthetic learner's listening comprehension is 87, 93 and 85 percent respectively.",,http://www.academypublication.com/ojs/index.php/tpls/article/download/tpls071111391148/1321
-How Good Are GPT Models at Machine Translation? A Comprehensive Evaluation,Amr Hendy,"Generative Pre-trained Transformer (GPT) models have shown remarkable capabilities for natural language generation, but their performance for machine translation has not been thoroughly investigated. In this paper, we present a comprehensive evaluation of GPT models for machine translation, covering various aspects such as quality of different GPT models in comparison with state-of-the-art research and commercial systems, effect of prompting strategies, robustness towards domain shifts and document-level translation. We experiment with eighteen different translation directions involving high and low resource languages, as well as non English-centric translations, and evaluate the performance of three GPT models: ChatGPT, GPT3.5 (text-davinci-003), and text-davinci-002. Our results show that GPT models achieve very competitive translation quality for high resource languages, while having limited capabilities for low resource languages. We also show that hybrid approaches, which combine GPT models with other translation systems, can further enhance the translation quality. We perform comprehensive analysis and human evaluation to further understand the characteristics of GPT translations. We hope that our paper provides valuable insights for researchers and practitioners in the field and helps to better understand the potential and limitations of GPT models for translation.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a comprehensive evaluation of GPT models for machine translation, covering various aspects such as quality of different G PT models in comparison with state-of-the-art research and commercial systems, effect of prompting strategies, robustness towards domain shifts and document-level translation.'}",http://arxiv.org/pdf/2302.09210
-Can ChatGPT Understand Too? A Comparative Study on ChatGPT and Fine-tuned BERT,Qihuang Zhong,"Recently, ChatGPT has attracted great attention, as it can generate fluent and high-quality responses to human inquiries. Several prior studies have shown that ChatGPT attains remarkable generation ability compared with existing models. However, the quantitative analysis of ChatGPT's understanding ability has been given little attention. In this report, we explore the understanding ability of ChatGPT by evaluating it on the most popular GLUE benchmark, and comparing it with 4 representative fine-tuned BERT-style models. We find that: 1) ChatGPT falls short in handling paraphrase and similarity tasks; 2) ChatGPT outperforms all BERT models on inference tasks by a large margin; 3) ChatGPT achieves comparable performance compared with BERT on sentiment analysis and question-answering tasks. Additionally, by combining some advanced prompting strategies, we show that the understanding ability of ChatGPT can be further improved.","{'model': 'tldr@v2.0.0', 'text': 'The understanding ability of ChatGPT is explored by evaluating it on the most popular GLUE benchmark, and comparing it with 4 representative fine-tuned BERT-style models, and it is found that it falls short in handling paraphrase and similarity tasks.'}",http://arxiv.org/pdf/2302.10198
-NLP based verification of a UML class model,Rauf Sajjad,"Robotized model time is a creating examination field. A significant number of representations are furthermore given where model checking has been associated for affirmation of various sorts of model. Such delineations are SAT based checks of UML/OCL models, embeded structures model affirmation, et cetera. In all these kind of employments, the complement is model checking. In later past, UML programming models are created from programming necessities conveyed in a trademark tongue, for instance, English by using NLP technique. Regardless, the diverse sorts of UML models delivered from consistent vernacular programming requirements specific using NLP approach have no technique for check as in the complete and correct models are done. It is a normal learning NLP techniques have been viably associated with make UML models as delineated above, in any case, in this paper, we address the issue of model checking and model affirmation by using NLP strategies. Such sort of changes require package of effort and time that makes the system of model affirmation wild and ambling. We used an approach for model watching that makes the technique of model checking straightforward and additionally the used philosophy should use the present resources used for generation of the UML class model.","{'model': 'tldr@v2.0.0', 'text': 'An approach for model watching is used that makes the technique of model checking straightforward and additionally the used philosophy should use the present resources used for generation of the UML class model.'}",
-The effect of metacognitive training and prompting on learning success in simulation‐based physics learning,Stephanie Moser,"Computer-based simulations are of particular interest to physics learning because they allow learners to actively manipulate graphical visualizations of complex phenomena. However, learning with simulations requires supportive elements to scaffold learners’ activities. Thus, our motivation was to investigate whether direct or indirect metacognitive scaffolding (i.e., training and prompting, respectively) or a combination of both during simulation-based learning leads to better learning outcomes compared to simulation-based learning alone. Employing a 2 × 2 factorial design (N = 99 Austrian secondary school students), we explored the role of metacognitive training prior to the use of simulations and the role of metacognitive prompting during the learning phase. Four different conditions were compared in relation to knowledge about the principle of energy conservation. A pre- and posttest assessing knowledge acquisition, metacognitive behavior, cognitive load, and mental effort was applied. At first glance, the results indicate no main effects for training and prompting, and no interaction effect. A closer look at the actual use of prompting after the empirical data were obtained, however, reveals a significant impact of the metacognitive strategies employed on performance in the posttest, with higher scores in the condition where appropriate prompting was used. This result emphasizes the importance of guidance during simulation-based learning.",,
-Teaching Young Children with Special Needs and Their Peers to Play Board Games: Effects of a Least to Most Prompting Procedure to Increase Independent Performance,Janet Davis-Temple,,"{'model': 'tldr@v2.0.0', 'text': 'After teaching the boardgame steps using a systematic prompting strategy, the participants demonstrated increases in the performance of board game steps and game-related on-task behavior.'}",https://europepmc.org/articles/pmc4711729?pdf=render
-Development of a Digital Game-Based Learning System with Graduated Prompting Strategy for Math Course,Kai-Hsiang Yang,"Many studies indicated that digital game-based learning can enhance students' learning motivation effectively, but it also needs appropriate tools or strategies to improve the learning achievement. The purpose of this study is to develop a digital game-based learning system with a graduated prompting strategy, so that students can consolidate concepts and skills of problem-solving, and further strength retention and transfer by graduated prompting strategy progressively and repeatedly. This study investigates the influence of students' learning motivation and achievement by using the proposed digital game-based learning system in the math courses.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this study is to develop a digital game-based learning system with a graduated prompting strategy, so that students can consolidate concepts and skills of problem-solving, and further strength retention and transfer by graduating prompting strategy progressively and repeatedly.'}",
-Pedagogical Strategies to Increase Students’ Engagement and Motivation,C. De Grandi,,,
-Impact of COVID-19 on Indian SMEs and Survival Strategies,D. Baragde,"India has been affected by the worldwide COVID-19 pandemic, which is dispensing two sorts of stuns on nations: health stun and a monetary stun. Given the idea of the infection, which is exceptionally infectious, the ways to contain the spread incorporate strategy activities, for example, burden of social removing, self-seclusion at home, closure of foundations and offices, limitations on versatility, and even lock-down of a whole nation. These activities can prompt desperate ramifications for economies around the globe. In other words, effective control of the infection requires the economy of a nation to stop its typical working.",,
-NaijaSenti: A Nigerian Twitter Sentiment Corpus for Multilingual Sentiment Analysis,Shamsuddeen Hassan Muhammad,"Sentiment analysis is one of the most widely studied applications in NLP, but most work focuses on languages with large amounts of data. We introduce the first large-scale human-annotated Twitter sentiment dataset for the four most widely spoken languages in Nigeria—Hausa, Igbo, Nigerian-Pidgin, and Yorùbá—consisting of around 30,000 annotated tweets per language, including a significant fraction of code-mixed tweets. We propose text collection, filtering, processing and labeling methods that enable us to create datasets for these low-resource languages. We evaluate a range of pre-trained models and transfer strategies on the dataset. We find that language-specific models and language-adaptive fine-tuning generally perform best. We release the datasets, trained models, sentiment lexicons, and code to incentivize research on sentiment analysis in under-represented languages.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the first large-scale human-annotated Twitter sentiment dataset for Nigeria—Hausa, Igbo, Nigerian-Pidgin, and Yorùbá—consisting of around 30,000 annotated tweets per language, including a significant fraction of code-mixed tweets.'}",
-Enabling Large Language Models to Generate Text with Citations,Tianyu Gao,"Large language models (LLMs) have emerged as a widely-used tool for information seeking, but their generated outputs are prone to hallucination. In this work, our aim is to allow LLMs to generate text with citations, improving their factual correctness and verifiability. Existing work mainly relies on commercial search engines and human evaluation, making it challenging to reproduce and compare different modeling approaches. We propose ALCE, the first benchmark for Automatic LLMs' Citation Evaluation. ALCE collects a diverse set of questions and retrieval corpora and requires building end-to-end systems to retrieve supporting evidence and generate answers with citations. We develop automatic metrics along three dimensions -- fluency, correctness, and citation quality -- and demonstrate their strong correlation with human judgements. Our experiments with state-of-the-art LLMs and novel prompting strategies show that current systems have considerable room for improvement -- For example, on the ELI5 dataset, even the best models lack complete citation support 50% of the time. Our analyses further highlight promising future directions, including developing better retrievers, advancing long-context LLMs, and improving the ability to synthesize information from multiple sources.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes ALCE, the first benchmark for Automatic LLMs' Citation Evaluation, and develops automatic metrics along three dimensions -- fluency, correctness, and citation quality -- and demonstrates their strong correlation with human judgements.""}",http://arxiv.org/pdf/2305.14627
-"Teaching assistants, neuro-linguistic programming (NLP) and special educational needs: ‘reframing’ the learning experience for students with mild SEN",Voldis Kudliskis,"This study examines how an understanding of two NLP concepts, the meta-model of language and the implementation of reframing, could be used to help teaching assistants enhance class-based interactions with students with mild SEN. Participants (students) completed a pre-intervention and a post-intervention ‘Beliefs About my Learning Experiences Scale’ (BALES). The BALES has good internal consistency (Cronbach’s α = .85) and moderate test–retest reliability (r = .75), indicating that the BALES is a psychometrically sound instrument for assessing beliefs about learning experiences. The intervention group were exposed to positive verbal influencing strategies via ‘reframing’ over an intervention period of 12 weeks from TAs who received training in the use of these specific aspects of NLP. Following the intervention, semi-structured interviews were conducted with TAs to elicit information about the perceived utility of the NLP strategies. Semi-structured interviews were also conducted with the students to examine their experiences and perceptions of their learning post intervention. A paired samples t-test failed to reveal a statistically reliable difference between average pre-intervention BALES scores (M = 99.17, SD = 12.76) and post-intervention BALES scores (M = 105.75, SD = 15.21) that students with mild SEN achieved t(11) = 5.02, p = 1.78, α = .10. However, the qualitative responses collected from semi-structured interviews with TAs and with students do suggest that the use of specific elements of NLP, namely the meta-model of language and (embedded) reframing may be helpful, at some intrinsic psychological level, in challenging concepts that SEN students have about their experiences in education.",,
-How Can Recommender Systems Benefit from Large Language Models: A Survey,Jianghao Lin,"Recommender systems (RS) play important roles to match users' information needs for Internet applications. In natural language processing (NLP) domains, large language model (LLM) has shown astonishing emergent abilities (e.g., instruction following, reasoning), thus giving rise to the promising research direction of adapting LLM to RS for performance enhancements and user experience improvements. In this paper, we conduct a comprehensive survey on this research direction from an application-oriented view. We first summarize existing research works from two orthogonal perspectives: where and how to adapt LLM to RS. For the""WHERE""question, we discuss the roles that LLM could play in different stages of the recommendation pipeline, i.e., feature engineering, feature encoder, scoring/ranking function, and pipeline controller. For the""HOW""question, we investigate the training and inference strategies, resulting in two fine-grained taxonomy criteria, i.e., whether to tune LLMs or not, and whether to involve conventional recommendation model (CRM) for inference. Detailed analysis and general development trajectories are provided for both questions, respectively. Then, we highlight key challenges in adapting LLM to RS from three aspects, i.e., efficiency, effectiveness, and ethics. Finally, we summarize the survey and discuss the future prospects. We also actively maintain a GitHub repository for papers and other related resources in this rising direction: https://github.com/CHIANGEL/Awesome-LLM-for-RecSys.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey on this research direction of adapting LLM to RS for performance enhancements and user experience improvements is conducted from an application-oriented view.'}",http://arxiv.org/pdf/2306.05817
-A stream computing approach towards scalable NLP,X. Artola,"Computational power needs have grown dramatically in recent years. This is also the case in many language processing tasks, due to overwhelming quantities of textual information that must be processed in a reasonable time frame. This scenario has led to a paradigm shift in the computing architectures and large-scale data processing strategies used in the NLP field. In this paper we describe a series of experiments carried out in the context of the NewsReader project with the goal of analyzing the scaling capabilities of the language processing pipeline used in it. We explore the use of Storm in a new approach for scalable distributed language processing across multiple machines and evaluate its effectiveness and efficiency when processing documents on a medium and large scale. The experiments have shown that there is a big room for improvement regarding language processing performance when adopting parallel architectures, and that we might expect even better results with the use of large clusters with many processing nodes.","{'model': 'tldr@v2.0.0', 'text': 'The experiments have shown that there is a big room for improvement regarding language processing performance when adopting parallel architectures, and that there might expect even better results with the use of large clusters with many processing nodes.'}",
-Expanding Pretrained Models to Thousands More Languages via Lexicon-based Adaptation,Xinyi Wang,"The performance of multilingual pretrained models is highly dependent on the availability of monolingual or parallel text present in a target language. Thus, the majority of the world’s languages cannot benefit from recent progress in NLP as they have no or limited textual data. To expand possibilities of using NLP technology in these under-represented languages, we systematically study strategies that relax the reliance on conventional language resources through the use of bilingual lexicons, an alternative resource with much better language coverage. We analyze different strategies to synthesize textual or labeled data using lexicons, and how this data can be combined with monolingual or parallel text when available. For 19 under-represented languages across 3 tasks, our methods lead to consistent improvements of up to 5 and 15 points with and without extra monolingual text respectively. Overall, our study highlights how NLP methods can be adapted to thousands more languages that are under-served by current technology.","{'model': 'tldr@v2.0.0', 'text': 'This study systematically study strategies that relax the reliance on conventional language resources through the use of bilingual lexicons, an alternative resource with much better language coverage and highlights how NLP methods can be adapted to thousands more languages that are under-served by current technology.'}",http://arxiv.org/pdf/2203.09435
-Vision-and-Language Pretrained Models: A Survey,Siqu Long,"Pretrained models have produced great success in both Computer Vision (CV) and Natural Language Processing (NLP). This progress leads to learning joint representations of vision and language pretraining by feeding visual and linguistic contents into a multi-layer transformer, Visual-Language Pretrained Models (VLPMs). In this paper, we present an overview of the major advances achieved in VLPMs for producing joint representations of vision and language. As the preliminaries, we briefly describe the general task definition and genetic architecture of VLPMs. We first discuss the language and vision data encoding methods and then present the mainstream VLPM structure as the core content. We further summarise several essential pretraining and fine-tuning strategies. Finally, we highlight three future directions for both CV and NLP researchers to provide insightful guidance.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the major advances achieved in VLPMs for producing joint representations of vision and language and highlights three future directions for both CV and NLP researchers to provide insightful guidance.'}",http://arxiv.org/pdf/2204.07356
-A Survey of Active Learning for Natural Language Processing,Zhisong Zhang,"In this work, we provide a literature review of active learning (AL) for its applications in natural language processing (NLP). In addition to a fine-grained categorization of query strategies, we also investigate several other important aspects of applying AL to NLP problems. These include AL for structured prediction tasks, annotation cost, model learning (especially with deep neural models), and starting and stopping AL. Finally, we conclude with a discussion of related topics and future directions.","{'model': 'tldr@v2.0.0', 'text': 'In addition to a fine-grained categorization of query strategies, several other important aspects of applying AL to NLP problems are investigated, including AL for structured prediction tasks, annotation cost, model learning, and starting and stopping AL.'}",http://arxiv.org/pdf/2210.10109
-COVID-19 Vaccine Hesitancy: Analysing Twitter to Identify Barriers to Vaccination in a Low Uptake Region of the UK,Kate Lanyi,"To facilitate effective targeted COVID-19 vaccination strategies, it is important to understand reasons for vaccine hesitancy where uptake is low. Artificial intelligence (AI) techniques offer an opportunity for real-time analysis of public attitudes, sentiments, and key discussion topics from sources of soft-intelligence, including social media data. In this work, we explore the value of soft-intelligence, leveraged using AI, as an evidence source to support public health research. As a case study, we deployed a natural language processing (NLP) platform to rapidly identify and analyse key barriers to vaccine uptake from a collection of geo-located tweets from London, UK. We developed a search strategy to capture COVID-19 vaccine related tweets, identifying 91,473 tweets between 30 November 2020 and 15 August 2021. The platform's algorithm clustered tweets according to their topic and sentiment, from which we extracted 913 tweets from the top 12 negative sentiment topic clusters. These tweets were extracted for further qualitative analysis. We identified safety concerns; mistrust of government and pharmaceutical companies; and accessibility issues as key barriers limiting vaccine uptake. Our analysis also revealed widespread sharing of vaccine misinformation amongst Twitter users. This study further demonstrates that there is promising utility for using off-the-shelf NLP tools to leverage insights from social media data to support public health research. Future work to examine where this type of work might be integrated as part of a mixed-methods research approach to support local and national decision making is suggested.","{'model': 'tldr@v2.0.0', 'text': 'There is promising utility for using off-the-shelf NLP tools to leverage insights from social media data to support public health research, and safety concerns; mistrust of government and pharmaceutical companies; and accessibility issues as key barriers limiting vaccine uptake are identified.'}",https://www.frontiersin.org/articles/10.3389/fdgth.2021.804855/pdf
-Diagnostic Reasoning Prompts Reveal the Potential for Large Language Model Interpretability in Medicine,Thomas Savage,"One of the major barriers to using large language models (LLMs) in medicine is the perception they use uninterpretable methods to make clinical decisions that are inherently different from the cognitive processes of clinicians. In this manuscript we develop novel diagnostic reasoning prompts to study whether LLMs can perform clinical reasoning to accurately form a diagnosis. We find that GPT4 can be prompted to mimic the common clinical reasoning processes of clinicians without sacrificing diagnostic accuracy. This is significant because an LLM that can use clinical reasoning to provide an interpretable rationale offers physicians a means to evaluate whether LLMs can be trusted for patient care. Novel prompting methods have the potential to expose the black box of LLMs, bringing them one step closer to safe and effective use in medicine.","{'model': 'tldr@v2.0.0', 'text': 'Novel diagnostic reasoning prompts are developed to study whether LLMs can perform clinical reasoning to accurately form a diagnosis and it is found that GPT4 can be prompted to mimic the common clinical reasoning processes of clinicians without sacrificing diagnostic accuracy.'}",https://arxiv.org/pdf/2308.06834
-Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM,Bochuan Cao,"Recently, Large Language Models (LLMs) have made significant advancements and are now widely used across various domains. Unfortunately, there has been a rising concern that LLMs can be misused to generate harmful or malicious content. Though a line of research has focused on aligning LLMs with human values and preventing them from producing inappropriate content, such alignments are usually vulnerable and can be bypassed by alignment-breaking attacks via adversarially optimized or handcrafted jailbreaking prompts. In this work, we introduce a Robustly Aligned LLM (RA-LLM) to defend against potential alignment-breaking attacks. RA-LLM can be directly constructed upon an existing aligned LLM with a robust alignment checking function, without requiring any expensive retraining or fine-tuning process of the original LLM. Furthermore, we also provide a theoretical analysis for RA-LLM to verify its effectiveness in defending against alignment-breaking attacks. Through real-world experiments on open-source large language models, we demonstrate that RA-LLM can successfully defend against both state-of-the-art adversarial prompts and popular handcrafted jailbreaking prompts by reducing their attack success rates from nearly 100\% to around 10\% or less.","{'model': 'tldr@v2.0.0', 'text': 'A Robustly Aligned LLM (RA-LLM) is introduced to defend against potential alignment-breaking attacks and can be directly constructed upon an existing aligned LLM with a robust alignment checking function, without requiring any expensive retraining or fine-tuning of the original LLM.'}",https://arxiv.org/pdf/2309.14348
-"Prompt, Plan, Perform: LLM-based Humanoid Control via Quantized Imitation Learning",Jingkai Sun,"In recent years, reinforcement learning and imitation learning have shown great potential for controlling humanoid robots' motion. However, these methods typically create simulation environments and rewards for specific tasks, resulting in the requirements of multiple policies and limited capabilities for tackling complex and unknown tasks. To overcome these issues, we present a novel approach that combines adversarial imitation learning with large language models (LLMs). This innovative method enables the agent to learn reusable skills with a single policy and solve zero-shot tasks under the guidance of LLMs. In particular, we utilize the LLM as a strategic planner for applying previously learned skills to novel tasks through the comprehension of task-specific prompts. This empowers the robot to perform the specified actions in a sequence. To improve our model, we incorporate codebook-based vector quantization, allowing the agent to generate suitable actions in response to unseen textual commands from LLMs. Furthermore, we design general reward functions that consider the distinct motion features of humanoid robots, ensuring the agent imitates the motion data while maintaining goal orientation without additional guiding direction approaches or policies. To the best of our knowledge, this is the first framework that controls humanoid robots using a single learning policy network and LLM as a planner. Extensive experiments demonstrate that our method exhibits efficient and adaptive ability in complicated motion tasks.","{'model': 'tldr@v2.0.0', 'text': 'This is the first framework that controls humanoid robots using a single learning policy network and LLM as a planner and designs general reward functions that consider the distinct motion features of humanoid robots, ensuring the agent imitates the motion data while maintaining goal orientation.'}",https://arxiv.org/pdf/2309.11359
-Make LLM a Testing Expert: Bringing Human-like Interaction to Mobile GUI Testing via Functionality-aware Decisions,Zhe Liu,"Automated Graphical User Interface (GUI) testing plays a crucial role in ensuring app quality, especially as mobile applications have become an integral part of our daily lives. Despite the growing popularity of learning-based techniques in automated GUI testing due to their ability to generate human-like interactions, they still suffer from several limitations, such as low testing coverage, inadequate generalization capabilities, and heavy reliance on training data. Inspired by the success of Large Language Models (LLMs) like ChatGPT in natural language understanding and question answering, we formulate the mobile GUI testing problem as a Q&A task. We propose GPTDroid, asking LLM to chat with the mobile apps by passing the GUI page information to LLM to elicit testing scripts, and executing them to keep passing the app feedback to LLM, iterating the whole process. Within this framework, we have also introduced a functionality-aware memory prompting mechanism that equips the LLM with the ability to retain testing knowledge of the whole process and conduct long-term, functionality-based reasoning to guide exploration. We evaluate it on 93 apps from Google Play and demonstrate that it outperforms the best baseline by 32% in activity coverage, and detects 31% more bugs at a faster rate. Moreover, GPTDroid identify 53 new bugs on Google Play, of which 35 have been confirmed and fixed.","{'model': 'tldr@v2.0.0', 'text': 'GPTDroid is proposed, a functionality-aware memory prompting mechanism that equips the Large Language Models (LLM) with the ability to retain testing knowledge of the whole process and conduct long-term, functionality-based reasoning to guide exploration.'}",
-Boosting Static Resource Leak Detection via LLM-based Resource-Oriented Intention Inference,Chong Wang,"Resource leaks, caused by resources not being released after acquisition, often lead to performance issues and system crashes. Existing static detection techniques rely on mechanical matching of predefined resource acquisition/release APIs, posing challenges to their effectiveness, including completeness of predefined APIs, identification of reachability validation, and analysis complexity. To overcome these challenges, we propose InferROI, a novel approach that leverages large language models (LLMs) to directly infer resource-oriented intentions (acquisition, release, and reachability validation) in code, based on resource management knowledge and code context understanding, rather than mechanical API matching. InferROI uses a prompt to instruct the LLM in inferring involved intentions from a given code snippet, which are then translated into formal expressions. By aggregating these inferred intentions, InferROI utilizes a lightweight static-analysis based algorithm to analyze control-flow paths extracted from the code, thereby detecting resource leaks. We evaluate InferROI on Java program and investigate its effectiveness in both resource-oriented intention inference and resource leak detection. Experimental results demonstrate that InferROI achieves a precision of 74.6% and a recall of 81.8% in intention inference on 172 code snippets from the DroidLeaks dataset. Additionally, InferROI covers a significant portion of concerned Android resources listed in the dataset. When applied to 86 bugs from the DroidLeaks dataset, InferROI exhibits a high bug detection rate (53.5%) and a low false alarm rate (8.1%) compared to eight baseline detectors. Moreover, we apply InferROI to resource leak detection in 100 methods from real-world open-source projects, where it identifies 12 unknown resource leak bugs, with 7 of them being confirmed by developers.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes InferROI, a novel approach that leverages large language models (LLMs) to directly infer resource-oriented intentions (acquisition, release, and reachability validation) in code, based on resource management knowledge and code context understanding, rather than mechanical API matching.'}",
-LLM-FuncMapper: Function Identification for Interpreting Complex Clauses in Building Codes via LLM,Zhe Zheng,"As a vital stage of automated rule checking (ARC), rule interpretation of regulatory texts requires considerable effort. However, interpreting regulatory clauses with implicit properties or complex computational logic is still challenging due to the lack of domain knowledge and limited expressibility of conventional logic representations. Thus, LLM-FuncMapper, an approach to identifying predefined functions needed to interpret various regulatory clauses based on the large language model (LLM), is proposed. First, by systematically analysis of building codes, a series of atomic functions are defined to capture shared computational logics of implicit properties and complex constraints, creating a database of common blocks for interpreting regulatory clauses. Then, a prompt template with the chain of thought is developed and further enhanced with a classification-based tuning strategy, to enable common LLMs for effective function identification. Finally, the proposed approach is validated with statistical analysis, experiments, and proof of concept. Statistical analysis reveals a long-tail distribution and high expressibility of the developed function database, with which almost 100% of computer-processible clauses can be interpreted and represented as computer-executable codes. Experiments show that LLM-FuncMapper achieve promising results in identifying relevant predefined functions for rule interpretation. Further proof of concept in automated rule interpretation also demonstrates the possibility of LLM-FuncMapper in interpreting complex regulatory clauses. To the best of our knowledge, this study is the first attempt to introduce LLM for understanding and interpreting complex regulatory clauses, which may shed light on further adoption of LLM in the construction domain.","{'model': 'tldr@v2.0.0', 'text': 'This study is the first attempt to introduce LLM for understanding and interpreting complex regulatory clauses, which may shed light on further adoption of LLM in the construction domain.'}",https://arxiv.org/pdf/2308.08728
-Cash Transaction Booking via Retrieval Augmented LLM,Xiaoli Zhang,"In large corporations, millions of cash transactions are booked via cash management software (CMS) per month. Most CMS systems adopt a key-word (search string) based matching logic for booking, which checks if the cash transaction description contains a specific search string and books the transaction to an appropriate general ledger account (GL-account) according to a booking rule. However, due to the free-text nature of transaction description and the diversity of cash transactions, CMS systems often fail due to data corruption (truncation, insertions, spelling errors), paraphrasing, and lack of reusable key word in the description, requiring significant manual intervention by accountants. Month over month, accountants manually handle CMS booking failures in spreadsheets. We present two machine learning models, a GL-account classification model and a search string extraction model, to alleviate this manual process. These two models, backed by retrieval augmented large language models, can automate booking for a substantial portion of the manual transactions. Our approach is robust to common data issues in transaction description. Unlike typical deep-learning models, our models are interpretable and explainable. For GL-account classification, our approach has an accuracy close to human experts. For search string extraction, compared to other methods such as fine-tuning transformers for extraction tasks, our approach produces reliable results closer to accountants.","{'model': 'tldr@v2.0.0', 'text': 'Two machine learning models are presented, a GL-account classification model and a search string extraction model, which can automate booking for a substantial portion of the manual transactions of CMS systems.'}",
-Resolving the Imbalance Issue in Hierarchical Disciplinary Topic Inference via LLM-based Data Augmentation,Xunxin Cai,"In addressing the imbalanced issue of data within the realm of Natural Language Processing, text data augmentation methods have emerged as pivotal solutions. This data imbalance is prevalent in the research proposals submitted during the funding application process. Such imbalances, resulting from the varying popularity of disciplines or the emergence of interdisciplinary studies, significantly impede the precision of downstream topic models that deduce the affiliated disciplines of these proposals. At the data level, proposals penned by experts and scientists are inherently complex technological texts, replete with intricate terminologies, which augmenting such specialized text data poses unique challenges. At the system level, this, in turn, compromises the fairness of AI-assisted reviewer assignment systems, which raises a spotlight on solving this issue. This study leverages large language models (Llama V1) as data generators to augment research proposals categorized within intricate disciplinary hierarchies, aiming to rectify data imbalances and enhance the equity of expert assignments. We first sample within the hierarchical structure to find the under-represented class. Then we designed a prompt for keyword-based research proposal generation. Our experiments attests to the efficacy of the generated data, demonstrating that research proposals produced using the prompts can effectively address the aforementioned issues and generate high quality scientific text data, thus help the model overcome the imbalanced issue.","{'model': 'tldr@v2.0.0', 'text': 'This study leverages large language models (Llama V1) as data generators to augment research proposals categorized within intricate disciplinary hierarchies, aiming to rectify data imbalances and enhance the equity of expert assignments.'}",https://arxiv.org/pdf/2310.05318
-Workshop On Large Language Models' Interpretability and Trustworthiness (LLMIT),Tulika Saha,"Large language models (LLMs), when scaled from millions to billions of parameters, have been demonstrated to exhibit the so-called 'emergence' effect, in that they are not only able to produce semantically correct and coherent text, but are also able to adapt themselves surprisingly well with small changes in contexts supplied as inputs (commonly called prompts). Despite producing semantically coherent and potentially relevant text for a given context, LLMs are vulnerable to yield incorrect information. This misinformation generation, or the so-called hallucination problem of an LLM, gets worse when an adversary manipulates the prompts to their own advantage, e.g., generating false propaganda to disrupt communal harmony, generating false information to trap consumers with target consumables etc. Not only does the consumption of an LLM-generated hallucinated content by humans pose societal threats, such misinformation, when used as prompts, may lead to detrimental effects for in-context learning (also known as few-shot prompt learning). With reference to the above-mentioned problems of LLM usage, we argue that it is necessary to foster research on topics related to not only identifying misinformation from LLM-generated content, but also to mitigate the propagation effects of this generated misinformation on downstream predictive tasks thus leading to more robust and effective leveraging in-context learning.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that it is necessary to foster research on topics related to not only identifying misinformation from LLM-generated content, but also to mitigate the propagation effects of this generated misinformation on downstream predictive tasks thus leading to more robust and effective leveraging in-context learning.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615311
-FigStep: Jailbreaking Large Vision-language Models via Typographic Visual Prompts,Yichen Gong,"Large vision-language models (VLMs) like GPT-4V represent an unprecedented revolution in the field of artificial intelligence (AI). Compared to single-modal large language models (LLMs), VLMs possess more versatile capabilities by incorporating additional modalities (e.g., images). Meanwhile, there's a rising enthusiasm in the AI community to develop open-source VLMs, such as LLaVA and MiniGPT4, which, however, have not undergone rigorous safety assessment. In this paper, to demonstrate that more modalities lead to unforeseen AI safety issues, we propose FigStep, a novel jailbreaking framework against VLMs. FigStep feeds harmful instructions into VLMs through the image channel and then uses benign text prompts to induce VLMs to output contents that violate common AI safety policies. Our experimental results show that FigStep can achieve an average attack success rate of 94.8% across 2 families of popular open-source VLMs, LLaVA and MiniGPT4 (a total of 5 VLMs). Moreover, we demonstrate that the methodology of FigStep can even jailbreak GPT-4V, which already leverages several system-level mechanisms to filter harmful queries. Above all, our experimental results reveal that VLMs are vulnerable to jailbreaking attacks, which highlights the necessity of novel safety alignments between visual and textual modalities.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results reveal that VLMs are vulnerable to jailbreaking attacks, which highlights the necessity of novel safety alignments between visual and textual modalities.'}",
-Improving Zero-shot Visual Question Answering via Large Language Models with Reasoning Question Prompts,Yunshi Lan,"Zero-shot Visual Question Answering (VQA) is a prominent vision-language task that examines both the visual and textual understanding capability of systems in the absence of training data. Recently, by converting the images into captions, information across multi-modalities is bridged and Large Language Models (LLMs) can apply their strong zero-shot generalization capability to unseen questions. To design ideal prompts for solving VQA via LLMs, several studies have explored different strategies to select or generate question-answer pairs as the exemplar prompts, which guide LLMs to answer the current questions effectively. However, they totally ignore the role of question prompts. The original questions in VQA tasks usually encounter ellipses and ambiguity which require intermediate reasoning. To this end, we present Reasoning Question Prompts for VQA tasks, which can further activate the potential of LLMs in zero-shot scenarios. Specifically, for each question, we first generate self-contained questions as reasoning question prompts via an unsupervised question edition module considering sentence fluency, semantic integrity and syntactic invariance. Each reasoning question prompt clearly indicates the intent of the original question. This results in a set of candidate answers. Then, the candidate answers associated with their confidence scores acting as answer heuristics are fed into LLMs and produce the final answer. We evaluate reasoning question prompts on three VQA challenges, experimental results demonstrate that they can significantly improve the results of LLMs on zero-shot setting and outperform existing state-of-the-art zero-shot methods on three out of four data sets. Our source code is publicly released at https://github.com/ECNU-DASE-NLP/RQP.","{'model': 'tldr@v2.0.0', 'text': 'Reasoning Question Prompts for VQA tasks are presented, which can further activate the potential of Large Language Models in zero-shot scenarios and can significantly improve the results of LLMs on zero- shot setting and outperform existing state-of-the-art zero-Shot methods on three out of four data sets.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612389
-ROSGPT_Vision: Commanding Robots Using Only Language Models' Prompts,Bilel Benjdira,"In this paper, we argue that the next generation of robots can be commanded using only Language Models' prompts. Every prompt interrogates separately a specific Robotic Modality via its Modality Language Model (MLM). A central Task Modality mediates the whole communication to execute the robotic mission via a Large Language Model (LLM). This paper gives this new robotic design pattern the name of: Prompting Robotic Modalities (PRM). Moreover, this paper applies this PRM design pattern in building a new robotic framework named ROSGPT_Vision. ROSGPT_Vision allows the execution of a robotic task using only two prompts: a Visual and an LLM prompt. The Visual Prompt extracts, in natural language, the visual semantic features related to the task under consideration (Visual Robotic Modality). Meanwhile, the LLM Prompt regulates the robotic reaction to the visual description (Task Modality). The framework automates all the mechanisms behind these two prompts. The framework enables the robot to address complex real-world scenarios by processing visual data, making informed decisions, and carrying out actions automatically. The framework comprises one generic vision module and two independent ROS nodes. As a test application, we used ROSGPT_Vision to develop CarMate, which monitors the driver's distraction on the roads and makes real-time vocal notifications to the driver. We showed how ROSGPT_Vision significantly reduced the development cost compared to traditional methods. We demonstrated how to improve the quality of the application by optimizing the prompting strategies, without delving into technical details. ROSGPT_Vision is shared with the community (link: https://github.com/bilel-bj/ROSGPT_Vision) to advance robotic research in this direction and to build more robotic frameworks that implement the PRM design pattern and enables controlling robots using only prompts.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that the next generation of robots can be commanded using only Language Models' prompts and this paper applies this PRM design pattern in building a new robotic framework named ROSGPT_Vision, which enables controlling robots using only prompts.""}",https://arxiv.org/pdf/2308.11236
-Prompts of Large Language Model for Commanding Power Grid Operation,Hanjiang Dong,"Large Language Models (LLMs) like ChatGPT can assist people’s general workflows, where the prompt is necessary to inspire the potential of LLMs to solve problems from specified or professional domains like robotics. In the electrical engineering subject or the electric power utility industry, experienced operators and professional experts monitor power grid operation statuses and interact with the grid via human commands on the screen, and components in the grid execute the commands to keep the complex grid safe and economical operation. In this process, human experts edit commands to operate the corresponding software. Human commands are the natural language that the LLM can process. The power grid is composed of generation, transmission, distribution, and other components. Therefore, we redesign the human-computer interaction frame between practitioners and the grid via recurrent prompts to apply the LLM to generate computer programming instructions from the multi-step natural language commands. The programming instruction is executed on system components after being confirmed or revised by human experts, and the quality of generated programs will be gradually improved through human feedback. The idea of this study is originally inspired by studies on controlling individual robotic components by ChatGPT. In the future, we will apply the designed prompt templates to drive the general LLM to generate desired samples which could be used to train an LLM professional in the domain knowledge of electrical engineering to operate multiple types of software for power grid operators.","{'model': 'tldr@v2.0.0', 'text': 'This study redesigns the human-computer interaction frame between practitioners and the grid via recurrent prompts to apply the LLM to generate computer programming instructions from the multi-step natural language commands for power grid operators.'}",
-Psychologically-informed chain-of-thought prompts for metaphor understanding in large language models,Ben Prystawski,"Probabilistic models of language understanding are valuable tools for investigating human language use. However, they need to be hand-designed for a particular domain. In contrast, large language models (LLMs) are trained on text that spans a wide array of domains, but they lack the structure and interpretability of probabilistic models. In this paper, we use chain-of-thought prompts to introduce structures from probabilistic models into LLMs. We explore this approach in the case of metaphor understanding. Our chain-of-thought prompts lead language models to infer latent variables and reason about their relationships in order to choose appropriate paraphrases for metaphors. The latent variables and relationships chosen are informed by theories of metaphor understanding from cognitive psychology. We apply these prompts to the two largest versions of GPT-3 and show that they can improve performance in a paraphrase selection task.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses chain-of-thought prompts to introduce structures from probabilistic models into large language models, and applies these prompts to the two largest versions of GPT-3 and shows that they can improve performance in a paraphrase selection task.'}",http://arxiv.org/pdf/2209.08141
-Can Large Language Models Explain Themselves? A Study of LLM-Generated Self-Explanations,Shiyuan Huang,"Large language models (LLMs) such as ChatGPT have demonstrated superior performance on a variety of natural language processing (NLP) tasks including sentiment analysis, mathematical reasoning and summarization. Furthermore, since these models are instruction-tuned on human conversations to produce""helpful""responses, they can and often will produce explanations along with the response, which we call self-explanations. For example, when analyzing the sentiment of a movie review, the model may output not only the positivity of the sentiment, but also an explanation (e.g., by listing the sentiment-laden words such as""fantastic""and""memorable""in the review). How good are these automatically generated self-explanations? In this paper, we investigate this question on the task of sentiment analysis and for feature attribution explanation, one of the most commonly studied settings in the interpretability literature (for pre-ChatGPT models). Specifically, we study different ways to elicit the self-explanations, evaluate their faithfulness on a set of evaluation metrics, and compare them to traditional explanation methods such as occlusion or LIME saliency maps. Through an extensive set of experiments, we find that ChatGPT's self-explanations perform on par with traditional ones, but are quite different from them according to various agreement metrics, meanwhile being much cheaper to produce (as they are generated along with the prediction). In addition, we identified several interesting characteristics of them, which prompt us to rethink many current model interpretability practices in the era of ChatGPT(-like) LLMs.","{'model': 'tldr@v2.0.0', 'text': ""Through an extensive set of experiments, it is found that ChatGPT's self-explanations perform on par with traditional ones, but are quite different from them according to various agreement metrics, meanwhile being much cheaper to produce (as they are generated along with the prediction).""}",
-Fake News in Sheep's Clothing: Robust Fake News Detection Against LLM-Empowered Style Attacks,Jiaying Wu,"It is commonly perceived that online fake news and reliable news exhibit stark differences in writing styles, such as the use of sensationalist versus objective language. However, we emphasize that style-related features can also be exploited for style-based attacks. Notably, the rise of powerful Large Language Models (LLMs) has enabled malicious users to mimic the style of trustworthy news outlets at minimal cost. Our analysis reveals that LLM-camouflaged fake news content leads to substantial performance degradation of state-of-the-art text-based detectors (up to 38% decrease in F1 Score), posing a significant challenge for automated detection in online ecosystems. To address this, we introduce SheepDog, a style-agnostic fake news detector robust to news writing styles. SheepDog achieves this adaptability through LLM-empowered news reframing, which customizes each article to match different writing styles using style-oriented reframing prompts. By employing style-agnostic training, SheepDog enhances its resilience to stylistic variations by maximizing prediction consistency across these diverse reframings. Furthermore, SheepDog extracts content-focused veracity attributions from LLMs, where the news content is evaluated against a set of fact-checking rationales. These attributions provide supplementary information and potential interpretability that assist veracity prediction. On three benchmark datasets, empirical results show that SheepDog consistently yields significant improvements over competitive baselines and enhances robustness against LLM-empowered style attacks.","{'model': 'tldr@v2.0.0', 'text': 'SheepDog is introduced, a style-agnostic fake news detector robust to news writing styles that consistently yields significant improvements over competitive baselines and enhances robustness against LLM-empowered style attacks.'}",
-Towards LLM-based Fact Verification on News Claims with a Hierarchical Step-by-Step Prompting Method,Xuan Zhang,"While large pre-trained language models (LLMs) have shown their impressive capabilities in various NLP tasks, they are still under-explored in the misinformation domain. In this paper, we examine LLMs with in-context learning (ICL) for news claim verification, and find that only with 4-shot demonstration examples, the performance of several prompting methods can be comparable with previous supervised models. To further boost performance, we introduce a Hierarchical Step-by-Step (HiSS) prompting method which directs LLMs to separate a claim into several subclaims and then verify each of them via multiple questions-answering steps progressively. Experiment results on two public misinformation datasets show that HiSS prompting outperforms state-of-the-art fully-supervised approach and strong few-shot ICL-enabled baselines.","{'model': 'tldr@v2.0.0', 'text': 'A Hierarchical Step-by-Step (HiSS) prompting method is introduced which directs LLMs to separate a claim into several subclaims and then verify each of them via multiple questions-answering steps progressively.'}",https://arxiv.org/pdf/2310.00305
-SelfEvolve: A Code Evolution Framework via Large Language Models,Shuyang Jiang,"Large language models (LLMs) have already revolutionized code generation, after being pretrained on publicly available code data. However, while various methods have been proposed to augment LLMs with retrieved knowledge and enhance the quality of code generation, the performance of these retrieval-based methods is limited by the strength of the retrievers used. In addition, while LLMs show great emergent ability, they still struggle to produce the correct code in one turn. To address these challenges, we propose a novel two-step pipeline, called \autoknow, that leverages LLMs as both knowledge providers and self-reflective programmers. Unlike retrieval-based methods, \autoknow~obtains the knowledge from input prompts and generates intermediate code based on the generated knowledge. After that, \autoknow~asks LLM to act as an expert programmer to perform debugging for the generated code. This is achieved by receiving the error message from the interpreter, without requiring special test cases for correctness verification. We evaluate \autoknow~on three code generation datasets, including DS-1000 for data science code, HumanEval for software engineering code, and TransCoder for C++-to-Python translation. Our empirical experiments show that \autoknow~outperforms strong baselines by a significant margin on all datasets. We also conduct exhaustive analytical experiments to validate the effectiveness of the two stages of \autoknow, and find that both are superior to other prompting-based methods. Further scalability analysis demonstrates that \autoknow~can be adapted to other more advanced models, such as GPT-4, and bring consistent efficacy improvement.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel two-step pipeline, called \\autoknow, that leverages LLMs as both knowledge providers and self-reflective programmers and finds that both are superior to other prompting-based methods.'}",http://arxiv.org/pdf/2306.02907
-CGSMP: Controllable Generative Summarization via Multimodal Prompt,Qian Yong,"Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of a large language model (LLM), this advancement has resulted in more fluent and coherent Natural Language Generation, which has contributed to improved development in downstream tasks such as abstractive summarization. Despite the recent progress in LLM, hallucination has become a serious problem in NLG. Hallucination happens when language models generate nonsensical or unfaithful text, which will lead to severe problems with reliability and effectiveness. In this paper, we propose a novel approach called Controllable Generative Summarization via Multimodal Prompt (CGSMP), which uses entities extracted from content and images as multimodal prompt control signals, thereby reducing hallucination issues. Specifically, the proposed CGSMP consists of three main modules: (1) an image prefix module that obtains image representations; (2) a prompt encoder module that fusion entities and images as multimodal prompts; and (3) a pre-trained causal language model that fuses input and controllable prompt and serves as the backbone of the language model. Experimental results demonstrate that the proposed method significantly improves the quality of generated summaries compared to the state of the arts.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the proposed CGSMP significantly improves the quality of generated summaries compared to the state of the arts, and serves as the backbone of the language model.'}",
-Unified Human-Scene Interaction via Prompted Chain-of-Contacts,Zeqi Xiao,"Human-Scene Interaction (HSI) is a vital component of fields like embodied AI and virtual reality. Despite advancements in motion quality and physical plausibility, two pivotal factors, versatile interaction control and the development of a user-friendly interface, require further exploration before the practical application of HSI. This paper presents a unified HSI framework, UniHSI, which supports unified control of diverse interactions through language commands. This framework is built upon the definition of interaction as Chain of Contacts (CoC): steps of human joint-object part pairs, which is inspired by the strong correlation between interaction types and human-object contact regions. Based on the definition, UniHSI constitutes a Large Language Model (LLM) Planner to translate language prompts into task plans in the form of CoC, and a Unified Controller that turns CoC into uniform task execution. To facilitate training and evaluation, we collect a new dataset named ScenePlan that encompasses thousands of task plans generated by LLMs based on diverse scenarios. Comprehensive experiments demonstrate the effectiveness of our framework in versatile task execution and generalizability to real scanned scenes. The project page is at https://github.com/OpenRobotLab/UniHSI .","{'model': 'tldr@v2.0.0', 'text': 'A unified HSI framework, UniHSI, which supports unified control of diverse interactions through language commands, built upon the definition of interaction as Chain of Contacts (CoC): steps of human joint-object part pairs, which is inspired by the strong correlation between interaction types and human-object contact regions.'}",https://arxiv.org/pdf/2309.07918
-Learning Profitable NFT Image Diffusions via Multiple Visual-Policy Guided Reinforcement Learning,Huiguo He,"We study the task of generating profitable Non-Fungible Token (NFT) images from user-input texts. Recent advances in diffusion models have shown great potential for image generation. However, existing works can fall short in generating visually-pleasing and highly-profitable NFT images, mainly due to the lack of 1) plentiful and fine-grained visual attribute prompts for an NFT image, and 2) effective optimization metrics for generating high-quality NFT images. To solve these challenges, we propose a Diffusion based generation framework with Multiple Visual-Policies as rewards (i.e., Diffusion-MVP) for NFT images. The proposed framework consists of a large language model (LLM), a diffusion-based image generator, and a series of visual rewards by design. First, the LLM enhances a basic human input (such as ""panda"") by generating more comprehensive NFT-style prompts that include specific visual attributes, such as ""panda with Ninja style and green background."" Second, the diffusion-based image generator is fine-tuned using a large-scale NFT dataset to capture fine-grained image styles and accessory compositions of popular NFT elements. Third, we further propose to utilize multiple visual-policies as optimization goals, including visual rarity levels, visual aesthetic scores, and CLIP-based text-image relevances. This design ensures that our proposed Diffusion-MVP is capable of minting NFT images with high visual quality and market value. To facilitate this research, we have collected the largest publicly available NFT image dataset to date, consisting of 1.5 million high-quality images with corresponding texts and market values. Extensive experiments including objective evaluations and user studies demonstrate that our framework can generate NFT images showing more visually engaging elements and higher market value, compared with state-of-the-art approaches.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments including objective evaluations and user studies demonstrate that the proposed Diffusion-MVP framework can generate NFT images showing more visually engaging elements and higher market value, compared with state-of-the-art approaches.'}",
-Majority Rule: better patching via Self-Consistency,Toufique Ahmed,"—Large Language models (LLMs) can be induced to solve non-trivial problems with “few-shot” prompts including illustrative problem-solution examples. Now if the few-shots also include “chain of thought” ( C oT ) explanations, which are of the form problem-explanation-solution , LLMs will generate a “explained” solution, and perform even better. Recently an exciting, substantially better technique, self-consistency [1] ( S - C ) has emerged, based on the intuition that there are many plausible explanations for the right solution; when the LLM is sampled repeatedly to generate a pool of explanation-solution pairs, for a given problem, the most frequently occurring solutions in the pool (ignoring the explanations ) tend to be even more likely to be correct!Unfortunately, the use of this highly-performant S - C (or even C oT ) approach in software engineering settings is hampered by the lack of explanations ; most software datasets lack explanations. In this paper, we describe an application of the S - C approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots. We achieve state-of-the art results, beating previous approaches to prompting-based program repair, on the MODIT dataset; we also find evidence suggesting that the correct commit messages are helping the LLM learn to produce better patches.","{'model': 'tldr@v2.0.0', 'text': 'An application of the S - C approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots is described, beating previous approaches to prompting-based program repair on the MODIT dataset.'}",https://arxiv.org/pdf/2306.00108
-Demonstration of InsightPilot: An LLM-Empowered Automated Data Exploration System,Pingchuan Ma,"Exploring data is crucial in data analysis, as it helps users understand and interpret the data more effectively. However, performing effective data exploration requires in-depth knowledge of the dataset and expertise in data analysis techniques. Not being familiar with either can create obstacles that make the process time-consuming and overwhelming for data analysts. To address this issue, we introduce InsightPilot, an LLM (Large Language Model)-based, automated data exploration system designed to simplify the data exploration process. InsightPilot automatically selects appropriate analysis intents, such as understanding, summarizing, and explaining. Then, these analysis intents are concretized by issuing corresponding intentional queries (IQueries) to create a meaningful and coherent exploration sequence. In brief, an IQuery is an abstraction and automation of data analysis operations, which mimics the approach of data analysts and simplifies the exploration process for users. By employing an LLM to iteratively collaborate with a state-of-the-art insight engine via IQueries, InsightPilot is effective in analyzing real-world datasets, enabling users to gain valuable insights through natural language inquiries. We demonstrate the effectiveness of InsightPilot in a case study, showing how it can help users gain valuable insights from their datasets.","{'model': 'tldr@v2.0.0', 'text': 'InsightPilot, an LLM-based, automated data exploration system designed to simplify the data exploration process, is introduced, effective in analyzing real-world datasets, enabling users to gain valuable insights through natural language inquiries.'}",http://arxiv.org/pdf/2304.00477
-LLM-Assisted Content Analysis: Using Large Language Models to Support Deductive Coding,Robert F. Chew,"Deductive coding is a widely used qualitative research method for determining the prevalence of themes across documents. While useful, deductive coding is often burdensome and time consuming since it requires researchers to read, interpret, and reliably categorize a large body of unstructured text documents. Large language models (LLMs), like ChatGPT, are a class of quickly evolving AI tools that can perform a range of natural language processing and reasoning tasks. In this study, we explore the use of LLMs to reduce the time it takes for deductive coding while retaining the flexibility of a traditional content analysis. We outline the proposed approach, called LLM-assisted content analysis (LACA), along with an in-depth case study using GPT-3.5 for LACA on a publicly available deductive coding data set. Additionally, we conduct an empirical benchmark using LACA on 4 publicly available data sets to assess the broader question of how well GPT-3.5 performs across a range of deductive coding tasks. Overall, we find that GPT-3.5 can often perform deductive coding at levels of agreement comparable to human coders. Additionally, we demonstrate that LACA can help refine prompts for deductive coding, identify codes for which an LLM is randomly guessing, and help assess when to use LLMs vs. human coders for deductive coding. We conclude with several implications for future practice of deductive coding and related research methods.","{'model': 'tldr@v2.0.0', 'text': 'Overall, it is found that GPT-3.5 can often perform deductive coding at levels of agreement comparable to human coders, and it is demonstrated that LACA can help refine prompts for deductedive coding, identify codes for which an LLM is randomly guessing, and help assess when to use LLMs vs. human codering.'}",http://arxiv.org/pdf/2306.14924
-Faithful Explanations of Black-box NLP Models Using LLM-generated Counterfactuals,Y. Gat,"Causal explanations of the predictions of NLP systems are essential to ensure safety and establish trust. Yet, existing methods often fall short of explaining model predictions effectively or efficiently and are often model-specific. In this paper, we address model-agnostic explanations, proposing two approaches for counterfactual (CF) approximation. The first approach is CF generation, where a large language model (LLM) is prompted to change a specific text concept while keeping confounding concepts unchanged. While this approach is demonstrated to be very effective, applying LLM at inference-time is costly. We hence present a second approach based on matching, and propose a method that is guided by an LLM at training-time and learns a dedicated embedding space. This space is faithful to a given causal graph and effectively serves to identify matches that approximate CFs. After showing theoretically that approximating CFs is required in order to construct faithful explanations, we benchmark our approaches and explain several models, including LLMs with billions of parameters. Our empirical results demonstrate the excellent performance of CF generation models as model-agnostic explainers. Moreover, our matching approach, which requires far less test-time resources, also provides effective explanations, surpassing many baselines. We also find that Top-K techniques universally improve every tested method. Finally, we showcase the potential of LLMs in constructing new benchmarks for model explanation and subsequently validate our conclusions. Our work illuminates new pathways for efficient and accurate approaches to interpreting NLP systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses model-agnostic explanations, proposing two approaches for counterfactual (CF) approximation and presents a method that is guided by an LLM at training-time and learns a dedicated embedding space that effectively serves to identify matches that approximate CFs.'}",https://arxiv.org/pdf/2310.00603
-SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models,Potsawee Manakul,"Generative Large Language Models (LLMs) such as GPT-3 are capable of generating highly fluent responses to a wide variety of user prompts. However, LLMs are known to hallucinate facts and make non-factual statements which can undermine trust in their output. Existing fact-checking approaches either require access to the output probability distribution (which may not be available for systems such as ChatGPT) or external databases that are interfaced via separate, often complex, modules. In this work, we propose""SelfCheckGPT"", a simple sampling-based approach that can be used to fact-check the responses of black-box models in a zero-resource fashion, i.e. without an external database. SelfCheckGPT leverages the simple idea that if an LLM has knowledge of a given concept, sampled responses are likely to be similar and contain consistent facts. However, for hallucinated facts, stochastically sampled responses are likely to diverge and contradict one another. We investigate this approach by using GPT-3 to generate passages about individuals from the WikiBio dataset, and manually annotate the factuality of the generated passages. We demonstrate that SelfCheckGPT can: i) detect non-factual and factual sentences; and ii) rank passages in terms of factuality. We compare our approach to several baselines and show that our approach has considerably higher AUC-PR scores in sentence-level hallucination detection and higher correlation scores in passage-level factuality assessment compared to grey-box methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ""SelfCheckGPT"", a simple sampling-based approach that can be used to fact-check the responses of black-box models in a zero-resource fashion, i.e. without an external database, and demonstrates that it can detect non-factual and factual sentences and rank passages in terms of factuality.'}",https://arxiv.org/pdf/2303.08896
-ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings,Shibo Hao,"Augmenting large language models (LLMs) with external tools has emerged as a promising approach to solving complex problems. However, traditional methods, which finetune LLMs with tool demonstration data, can be both costly and restricted to a predefined set of tools. Recent in-context learning paradigm alleviates these issues, but the limited context length only allows for a few shots of demonstrations, leading to suboptimal understandings of the tools. Moreover, when there are numerous tools to choose from, in-context learning could completely fail to work. In this paper, we propose an alternative approach, $\textbf{ToolkenGPT}$, which combines the benefits of both sides. Our approach represents each $\underline{tool}$ as a to$\underline{ken}$ ($\textit{toolken}$) and learns an embedding for it, enabling tool calls in the same way as generating a regular word token. Once a toolken is triggered, the LLM is prompted to complete arguments for the tool to execute. ToolkenGPT offers the flexibility to plug in an arbitrary number of tools by expanding the set of toolkens on the fly. In addition, it improves tool use by allowing extensive demonstration data for learning the toolken embeddings. In diverse domains, including numerical reasoning, knowledge-based question answering, and embodied plan generation, our approach effectively augments LLMs with tools and substantially outperforms various latest baselines. ToolkenGPT demonstrates the promising ability to use relevant tools from a large tool set in complex scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The proposed ToolkenGPT offers the flexibility to plug in an arbitrary number of tools by expanding the set of toolkens on the fly and improves tool use by allowing extensive demonstration data for learning the toolken embeddings.'}",http://arxiv.org/pdf/2305.11554
-Revisiting the Plastic Surgery Hypothesis via Large Language Models,Chun Xia,"Automated Program Repair (APR) aspires to automatically generate patches for an input buggy program. Traditional APR tools typically focus on specific bug types and fixes through the use of templates, heuristics, and formal specifications. However, these techniques are limited in terms of the bug types and patch variety they can produce. As such, researchers have designed various learning-based APR tools with recent work focused on directly using Large Language Models (LLMs) for APR. While LLM-based APR tools are able to achieve state-of-the-art performance on many repair datasets, the LLMs used for direct repair are not fully aware of the project-specific information such as unique variable or method names. The plastic surgery hypothesis is a well-known insight for APR, which states that the code ingredients to fix the bug usually already exist within the same project. Traditional APR tools have largely leveraged the plastic surgery hypothesis by designing manual or heuristic-based approaches to exploit such existing code ingredients. However, as recent APR research starts focusing on LLM-based approaches, the plastic surgery hypothesis has been largely ignored. In this paper, we ask the following question: How useful is the plastic surgery hypothesis in the era of LLMs? Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting. To this end, we propose FitRepair, which combines the direct usage of LLMs with two domain-specific fine-tuning strategies and one prompting strategy for more powerful APR. Our experiments on the widely studied Defects4j 1.2 and 2.0 datasets show that FitRepair fixes 89 and 44 bugs (substantially outperforming the best-performing baseline by 15 and 8), respectively, demonstrating a promising future of the plastic surgery hypothesis in the era of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting, and this paper proposes FitRepair, which combines the direct usage of LLMs with two domain-specific fine- tuning strategies and one prompting strategy for more powerful APR.'}",http://arxiv.org/pdf/2303.10494
-Automatic Calibration and Error Correction for Large Language Models via Pareto Optimal Self-Supervision,Theodore Zhao,"Large language models (LLMs) have demonstrated remarkable capabilities out of box for a wide range of applications, yet accuracy still remains a major growth area, especially in mission-critical domains such as biomedicine. An effective method to calibrate the confidence level on LLM responses is essential to automatically detect errors and facilitate human-in-the-loop verification. An important source of calibration signals stems from expert-stipulated programmatic supervision, which is often available at low cost but has its own limitations such as noise and coverage. In this paper, we introduce a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every response, without any additional manual efforts. This is accomplished by learning a harmonizer model to align LLM output with other available supervision sources, which would assign higher risk scores to more uncertain LLM responses and facilitate error correction. Experiments on standard relation extraction tasks in biomedical and general domains demonstrate the promise of this approach, with our proposed risk scores highly correlated with the real error rate of LLMs. For the most uncertain test instances, dynamic prompting based on our proposed risk scores results in significant accuracy improvement for off-the-shelf LLMs, boosting GPT-3 results past state-of-the-art (SOTA) weak supervision and GPT-4 results past SOTA supervised results on challenging evaluation datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every response, without any additional manual efforts.'}",https://arxiv.org/pdf/2306.16564
-Automatic Calibration and Error Correction for Generative Large Language Models via Pareto Optimal Self-Supervision,Theodore Zhao,"Generative Large language models (LLMs) have demonstrated remarkable capabilities for a wide range of applications, but reducing ungrounded or erroneous responses remains a major growth area. Unlike task-specific models, there lack an effective method to calibrate the confidence level of LLM responses to indicate potential errors and facilitate human-in-the-loop verification. An important source of calibration stems from expert-stipulated programmatic supervision, which is often available at low cost but has its own limitations such as noise and coverage. In this paper, we introduce a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every LLM response, without any additional manual efforts. This is accomplished by learning a harmonizer model to align with LLM output as well as other weak supervision sources. The model assigns higher risk scores to more uncertain LLM responses and facilitate error correction. Experiments on standard relation extraction and classification tasks in biomedical and general domains demonstrate that the proposed risk score is highly correlated with the actual LLM error rate. By using a dynamic prompting strategy based on the risk score, we observed significant accuracy improvement for off-the-shelf LLMs, boosting GPT-3.5 results past state-of-the-art (SOTA) weak supervision model and GPT-4 results past SOTA supervised results on challenging evaluation datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every LLM response, without any additional manual efforts.'}",
-DriveGPT4: Interpretable End-to-end Autonomous Driving via Large Language Model,Zhenhua Xu,"In the past decade, autonomous driving has experienced rapid development in both academia and industry. However, its limited interpretability remains a significant unsolved problem, severely hindering autonomous vehicle commercialization and further development. Previous approaches utilizing small language models have failed to address this issue due to their lack of flexibility, generalization ability, and robustness. Recently, multimodal large language models (LLMs) have gained considerable attention from the research community for their capability to process and reason non-text data (e.g., images and videos) by text. In this paper, we present DriveGPT4, an interpretable end-to-end autonomous driving system utilizing LLMs. DriveGPT4 is capable of interpreting vehicle actions and providing corresponding reasoning, as well as answering diverse questions posed by human users for enhanced interaction. Additionally, DriveGPT4 predicts vehicle low-level control signals in an end-to-end fashion. These capabilities stem from a customized visual instruction tuning dataset specifically designed for autonomous driving. To the best of our knowledge, DriveGPT4 is the first work focusing on interpretable end-to-end autonomous driving. When evaluated on multiple tasks alongside conventional methods and video understanding LLMs, DriveGPT4 demonstrates superior qualitative and quantitative performance. Additionally, DriveGPT4 can be generalized in a zero-shot fashion to accommodate more unseen scenarios. The project page is available at https://tonyxuqaq.github.io/projects/DriveGPT4/ .","{'model': 'tldr@v2.0.0', 'text': 'DriveGPT4 is an interpretable end-to-end autonomous driving system utilizing multimodal large language models that is capable of interpreting vehicle actions and providing corresponding reasoning, as well as answering diverse questions posed by human users for enhanced interaction.'}",https://arxiv.org/pdf/2310.01412
-BLSP: Bootstrapping Language-Speech Pre-training via Behavior Alignment of Continuation Writing,Chen Wang,"The emergence of large language models (LLMs) has sparked significant interest in extending their remarkable language capabilities to speech. However, modality alignment between speech and text still remains an open problem. Current solutions can be categorized into two strategies. One is a cascaded approach where outputs (tokens or states) of a separately trained speech recognition system are used as inputs for LLMs, which limits their potential in modeling alignment between speech and text. The other is an end-to-end approach that relies on speech instruction data, which is very difficult to collect in large quantities. In this paper, we address these issues and propose the BLSP approach that Bootstraps Language-Speech Pre-training via behavior alignment of continuation writing. We achieve this by learning a lightweight modality adapter between a frozen speech encoder and an LLM, ensuring that the LLM exhibits the same generation behavior regardless of the modality of input: a speech segment or its transcript. The training process can be divided into two steps. The first step prompts an LLM to generate texts with speech transcripts as prefixes, obtaining text continuations. In the second step, these continuations are used as supervised signals to train the modality adapter in an end-to-end manner. We demonstrate that this straightforward process can extend the capabilities of LLMs to speech, enabling speech recognition, speech translation, spoken language understanding, and speech conversation, even in zero-shot cross-lingual scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes the BLSP approach that Bootstraps Language-Speech Pre-training via behavior alignment of continuation writing and demonstrates that this straightforward process can extend the capabilities of LLMs to speech, enabling speech recognition, speech translation, spoken language understanding, and speech conversation, even in zero-shot cross-lingual scenarios.'}",https://arxiv.org/pdf/2309.00916
-GPTuner: A Manual-Reading Database Tuning System via GPT-Guided Bayesian Optimization,Jiale Lao,"Modern database management systems (DBMS) expose hundreds of configurable knobs to control system behaviours. Determining the appropriate values for these knobs to improve DBMS performance is a long-standing problem in the database community. As there is an increasing number of knobs to tune and each knob could be in continuous or categorical values, manual tuning becomes impractical. Recently, automatic tuning systems using machine learning methods have shown great potentials. However, existing approaches still incur significant tuning costs or only yield sub-optimal performance. This is because they either ignore the extensive domain knowledge available (e.g., DBMS manuals and forum discussions) and only rely on the runtime feedback of benchmark evaluations to guide the optimization, or they utilize the domain knowledge in a limited way. Hence, we propose GPTuner, a manual-reading database tuning system. Firstly, we develop a Large Language Model (LLM)-based pipeline to collect and refine heterogeneous knowledge, and propose a prompt ensemble algorithm to unify a structured view of the refined knowledge. Secondly, using the structured knowledge, we (1) design a workload-aware and training-free knob selection strategy, (2) develop a search space optimization technique considering the value range of each knob, and (3) propose a Coarse-to-Fine Bayesian Optimization Framework to explore the optimized space. Finally, we evaluate GPTuner under different benchmarks (TPC-C and TPC-H), metrics (throughput and latency) as well as DBMS (PostgreSQL and MySQL). Compared to the state-of-the-art approaches, GPTuner identifies better configurations in 16x less time on average. Moreover, GPTuner achieves up to 30% performance improvement (higher throughput or lower latency) over the best-performing alternative.","{'model': 'tldr@v2.0.0', 'text': 'GPTuner, a manual-reading database tuning system that designs a workload-aware and training-free knob selection strategy, develops a search space optimization technique considering the value range of each knob, and proposes a Coarse-to-Fine Bayesian Optimization Framework to explore the optimized space.'}",
-InstructProtein: Aligning Human and Protein Language via Knowledge Instruction,Zeyuan Wang,"Large Language Models (LLMs) have revolutionized the field of natural language processing, but they fall short in comprehending biological sequences such as proteins. To address this challenge, we propose InstructProtein, an innovative LLM that possesses bidirectional generation capabilities in both human and protein languages: (i) taking a protein sequence as input to predict its textual function description and (ii) using natural language to prompt protein sequence generation. To achieve this, we first pre-train an LLM on both protein and natural language corpora, enabling it to comprehend individual languages. Then supervised instruction tuning is employed to facilitate the alignment of these two distinct languages. Herein, we introduce a knowledge graph-based instruction generation framework to construct a high-quality instruction dataset, addressing annotation imbalance and instruction deficits in existing protein-text corpus. In particular, the instructions inherit the structural relations between proteins and function annotations in knowledge graphs, which empowers our model to engage in the causal modeling of protein functions, akin to the chain-of-thought processes in natural languages. Extensive experiments on bidirectional protein-text generation tasks show that InstructProtein outperforms state-of-the-art LLMs by large margins. Moreover, InstructProtein serves as a pioneering step towards text-based protein function prediction and sequence design, effectively bridging the gap between protein and human language understanding.","{'model': 'tldr@v2.0.0', 'text': 'InstructProtein, an innovative LLM that possesses bidirectional generation capabilities in both human and protein languages, serves as a pioneering step towards text-based protein function prediction and sequence design, effectively bridging the gap between protein and human language understanding.'}",https://arxiv.org/pdf/2310.03269
-Self-Detoxifying Language Models via Toxification Reversal,Chak Tou Leong,"Language model detoxification aims to minimize the risk of generating offensive or harmful content in pretrained language models (PLMs) for safer deployment. Existing methods can be roughly categorized as finetuning-based and decoding-based. However, the former is often resource-intensive, while the latter relies on additional components and potentially compromises the generation fluency. In this paper, we propose a more lightweight approach that enables the PLM itself to achieve""self-detoxification"". Our method is built upon the observation that prepending a negative steering prompt can effectively induce PLMs to generate toxic content. At the same time, we are inspired by the recent research in the interpretability field, which formulates the evolving contextualized representations within the PLM as an information stream facilitated by the attention layers. Drawing on this idea, we devise a method to identify the toxification direction from the normal generation process to the one prompted with the negative prefix, and then steer the generation to the reversed direction by manipulating the information movement within the attention layers. Experimental results show that our approach, without any fine-tuning or extra components, can achieve comparable performance with state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a more lightweight approach that enables the PLM itself to achieve ""self-detoxification"", built upon the observation that prepending a negative steering prompt can effectively induce PLMs to generate toxic content.'}",
-Enhance Reasoning Ability of Visual-Language Models via Large Language Models,Yueting Yang,"Pre-trained visual language models (VLM) have shown excellent performance in image caption tasks. However, it sometimes shows insufficient reasoning ability. In contrast, large language models (LLMs) emerge with powerful reasoning capabilities. Therefore, we propose a method called TReE, which transfers the reasoning ability of a large language model to a visual language model in zero-shot scenarios. TReE contains three stages: observation, thinking, and re-thinking. Observation stage indicates that VLM obtains the overall information of the relative image. Thinking stage combines the image information and task description as the prompt of the LLM, inference with the rationals. Re-Thinking stage learns from rationale and then inference the final result through VLM.","{'model': 'tldr@v2.0.0', 'text': 'A method called TReE is proposed, which transfers the reasoning ability of a large language model to a visual language model in zero-shot scenarios, and contains three stages: observation, thinking, and re-thinking.'}",http://arxiv.org/pdf/2305.13267
-Violation of Expectation via Metacognitive Prompting Reduces Theory of Mind Prediction Error in Large Language Models,Courtland Leer,"Recent research shows that Large Language Models (LLMs) exhibit a compelling level of proficiency in Theory of Mind (ToM) tasks. This ability to impute unobservable mental states to others is vital to human social cognition and may prove equally important in principal-agent relations between individual humans and Artificial Intelligences (AIs). In this paper, we explore how a mechanism studied in developmental psychology known as Violation of Expectation (VoE) can be implemented to reduce errors in LLM prediction about users by leveraging emergent ToM affordances. And we introduce a \textit{metacognitive prompting} framework to apply VoE in the context of an AI tutor. By storing and retrieving facts derived in cases where LLM expectation about the user was violated, we find that LLMs are able to learn about users in ways that echo theories of human learning. Finally, we discuss latent hazards and augmentative opportunities associated with modeling user psychology and propose ways to mitigate risk along with possible directions for future inquiry.",,https://arxiv.org/pdf/2310.06983
-Model Selection and Evaluation for Learning Analytics via Interpretable Machine Learning,Zhuoran Li,"Learning analytics is one of the most crucial tasks in understanding the relationship between learners' learning behaviors and academic performance in MOOCs. Machine learning methods that enable us to learn the pattern of complex data make it more effective to conduct learning analytics. However, among a variety of models, only some of them can provide us with sufficient interpretability to understand the modeling result and propose pedagogically instructive suggestions for improving learners' learning strategies. By demonstrating the performance and interpretability of 17 machine learning models, this study would select the CART, which could balance the model performance and interpretability well, as the optimal model for learning analytics in MOOCs. Through the further study of the CART, we concluded that the social interaction behaviors and prompt completion of learning tasks could most significantly help with improving learners' academic performance.","{'model': 'tldr@v2.0.0', 'text': ""The further study of the CART concluded that the social interaction behaviors and prompt completion of learning tasks could most significantly help with improving learners' academic performance.""}",
-Bridging the Gap between Human Motion and Action Semantics via Kinematic Phrases,Xinpeng Liu,"The goal of motion understanding is to establish a reliable mapping between motion and action semantics, while it is a challenging many-to-many problem. An abstract action semantic (i.e., walk forwards) could be conveyed by perceptually diverse motions (walk with arms up or swinging), while a motion could carry different semantics w.r.t. its context and intention. This makes an elegant mapping between them difficult. Previous attempts adopted direct-mapping paradigms with limited reliability. Also, current automatic metrics fail to provide reliable assessments of the consistency between motions and action semantics. We identify the source of these problems as the significant gap between the two modalities. To alleviate this gap, we propose Kinematic Phrases (KP) that take the objective kinematic facts of human motion with proper abstraction, interpretability, and generality characteristics. Based on KP as a mediator, we can unify a motion knowledge base and build a motion understanding system. Meanwhile, KP can be automatically converted from motions and to text descriptions with no subjective bias, inspiring Kinematic Prompt Generation (KPG) as a novel automatic motion generation benchmark. In extensive experiments, our approach shows superiority over other methods. Our code and data would be made publicly available at https://foruck.github.io/KP.","{'model': 'tldr@v2.0.0', 'text': 'Kinematic Phrases (KP) is proposed that take the objective kinematic facts of human motion with proper abstraction, interpretability, and generality characteristics and can be automatically converted from motions and to text descriptions with no subjective bias.'}",https://arxiv.org/pdf/2310.04189
-Automatic Hallucination Assessment for Aligned Large Language Models via Transferable Adversarial Attacks,Xiaodong Yu,"Although remarkable progress has been achieved in preventing large language model (LLM) hallucinations using instruction tuning and retrieval augmentation, it remains challenging to measure the reliability of LLMs using human-crafted evaluation data which is not available for many tasks and domains and could suffer from data leakage. Inspired by adversarial machine learning, this paper aims to develop a method of automatically generating evaluation data by appropriately modifying existing data on which LLMs behave faithfully. Specifically, this paper presents AutoDebug, an LLM-based framework to use prompting chaining to generate transferable adversarial attacks in the form of question-answering examples. We seek to understand the extent to which these examples trigger the hallucination behaviors of LLMs. We implement AutoDebug using ChatGPT and evaluate the resulting two variants of a popular open-domain question-answering dataset, Natural Questions (NQ), on a collection of open-source and proprietary LLMs under various prompting settings. Our generated evaluation data is human-readable and, as we show, humans can answer these modified questions well. Nevertheless, we observe pronounced accuracy drops across multiple LLMs including GPT-4. Our experimental results show that LLMs are likely to hallucinate in two categories of question-answering scenarios where (1) there are conflicts between knowledge given in the prompt and their parametric knowledge, or (2) the knowledge expressed in the prompt is complex. Finally, we find that the adversarial examples generated by our method are transferable across all considered LLMs. The examples generated by a small model can be used to debug a much larger model, making our approach cost-effective.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents AutoDebug, an LLM-based framework to use prompting chaining to generate transferable adversarial attacks in the form of question-answering examples, and finds that the adversarial examples generated by the method are transferable across all considered LLMs.'}",
-Improving Few-shot Generalization of Safety Classifiers via Data Augmented Parameter-Efficient Fine-Tuning,Ananth Balashankar,"As large language models (LLMs) are widely adopted, new safety issues and policies emerge, to which existing safety classifiers do not generalize well. If we have only observed a few examples of violations of a new safety rule, how can we build a classifier to detect violations? In this paper, we study the novel setting of domain-generalized few-shot learning for LLM-based text safety classifiers. Unlike prior few-shot work, these new safety issues can be hard to uncover and we do not get to choose the few examples. We demonstrate that existing few-shot techniques do not perform well in this setting, and rather we propose to do parameter-efficient fine-tuning (PEFT) combined with augmenting training data based on similar examples in prior existing rules. We empirically show that our approach of similarity-based data-augmentation + prompt-tuning (DAPT) consistently outperforms baselines that either do not rely on data augmentation or on PEFT by 7-17% F1 score in the Social Chemistry moral judgement and 9-13% AUC in the Toxicity detection tasks, even when the new rule is loosely correlated with existing ones.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the approach of similarity-based data-augmentation + prompt-tuning (DAPT) consistently outperforms baselines that either do not rely on data augmentation or on PEFT by 7-17% F1 score in the Social Chemistry moral judgement and 9-13% AUC in the Toxicity detection tasks.'}",
-Evaluation of GPT-3.5 and GPT-4 for supporting real-world information needs in healthcare delivery,Debadutta Dash,"Despite growing interest in using large language models (LLMs) in healthcare, current explorations do not assess the real-world utility and safety of LLMs in clinical settings. Our objective was to determine whether two LLMs can serve information needs submitted by physicians as questions to an informatics consultation service in a safe and concordant manner. Sixty six questions from an informatics consult service were submitted to GPT-3.5 and GPT-4 via simple prompts. 12 physicians assessed the LLM responses' possibility of patient harm and concordance with existing reports from an informatics consultation service. Physician assessments were summarized based on majority vote. For no questions did a majority of physicians deem either LLM response as harmful. For GPT-3.5, responses to 8 questions were concordant with the informatics consult report, 20 discordant, and 9 were unable to be assessed. There were 29 responses with no majority on""Agree"",""Disagree"", and""Unable to assess"". For GPT-4, responses to 13 questions were concordant, 15 discordant, and 3 were unable to be assessed. There were 35 responses with no majority. Responses from both LLMs were largely devoid of overt harm, but less than 20% of the responses agreed with an answer from an informatics consultation service, responses contained hallucinated references, and physicians were divided on what constitutes harm. These results suggest that while general purpose LLMs are able to provide safe and credible responses, they often do not meet the specific information need of a given question. A definitive evaluation of the usefulness of LLMs in healthcare settings will likely require additional research on prompt engineering, calibration, and custom-tailoring of general purpose models.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that while general purpose LLMs are able to provide safe and credible responses, they often do not meet the specific information need of a given question.'}",http://arxiv.org/pdf/2304.13714
-TEMPERA: Test-Time Prompt Editing via Reinforcement Learning,Tianjun Zhang,"Careful prompt design is critical to the use of large language models in zeroshot or few-shot learning. As a consequence, there is a growing interest in automated methods to design optimal prompts. In this work, we propose TEst-tiMe Prompt Editing using Reinforcement leArning (TEMPERA). In contrast to prior prompt generation methods, TEMPERA can efficiently leverage prior knowledge, is adaptive to different queries, and provides an interpretable prompt for every query. To achieve this, we design a novel action space that allows flexible editing of the initial prompts covering a comprehensive set of commonly-used components like instructions, few-shot exemplars, and verbalizers. The proposed method achieves significant gains compared with recent SoTA approaches like prompt tuning, AutoPrompt, and RLPrompt, across a variety of tasks, including sentiment analysis, topic classification, natural language inference, and reading comprehension. Our method achieves 5.33x on average improvement in sample efficiency when compared to the traditional fine-tuning methods. Our code is available at https://github.com/tianjunz/TEMPERA.",,
-GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher,Youliang Yuan,"Safety lies at the core of the development of Large Language Models (LLMs). There is ample work on aligning LLMs with human ethics and preferences, including data filtering in pretraining, supervised fine-tuning, reinforcement learning from human feedback, and red teaming, etc. In this study, we discover that chat in cipher can bypass the safety alignment techniques of LLMs, which are mainly conducted in natural languages. We propose a novel framework CipherChat to systematically examine the generalizability of safety alignment to non-natural languages -- ciphers. CipherChat enables humans to chat with LLMs through cipher prompts topped with system role descriptions and few-shot enciphered demonstrations. We use CipherChat to assess state-of-the-art LLMs, including ChatGPT and GPT-4 for different representative human ciphers across 11 safety domains in both English and Chinese. Experimental results show that certain ciphers succeed almost 100% of the time to bypass the safety alignment of GPT-4 in several safety domains, demonstrating the necessity of developing safety alignment for non-natural languages. Notably, we identify that LLMs seem to have a ''secret cipher'', and propose a novel SelfCipher that uses only role play and several demonstrations in natural language to evoke this capability. SelfCipher surprisingly outperforms existing human ciphers in almost all cases. Our code and data will be released at https://github.com/RobustNLP/CipherChat.","{'model': 'tldr@v2.0.0', 'text': 'It is discovered that chat in cipher can bypass the safety alignment techniques of LLMs, and a novel SelfCipher is proposed that uses only role play and several demonstrations in natural language to evoke this capability, and surprisingly outperforms existing human ciphers in almost all cases.'}",https://arxiv.org/pdf/2308.06463
-LLMatic: Neural Architecture Search via Large Language Models and Quality-Diversity Optimization,Muhammad Umair Nasir,"Large Language Models (LLMs) have emerged as powerful tools capable of accomplishing a broad spectrum of tasks. Their abilities span numerous areas, and one area where they have made a significant impact is in the domain of code generation. In this context, we view LLMs as mutation and crossover tools. Meanwhile, Quality-Diversity (QD) algorithms are known to discover diverse and robust solutions. By merging the code-generating abilities of LLMs with the diversity and robustness of QD solutions, we introduce LLMatic, a Neural Architecture Search (NAS) algorithm. While LLMs struggle to conduct NAS directly through prompts, LLMatic uses a procedural approach, leveraging QD for prompts and network architecture to create diverse and highly performant networks. We test LLMatic on the CIFAR-10 image classification benchmark, demonstrating that it can produce competitive networks with just $2,000$ searches, even without prior knowledge of the benchmark domain or exposure to any previous top-performing models for the benchmark.","{'model': 'tldr@v2.0.0', 'text': 'While LLMs struggle to conduct NAS directly through prompts, LLMatic uses a procedural approach, leveraging QD for prompts and network architecture to create diverse and highly performant networks.'}",https://arxiv.org/pdf/2306.01102
-ChatSpot: Bootstrapping Multimodal LLMs via Precise Referring Instruction Tuning,Liang Zhao,"Human-AI interactivity is a critical aspect that reflects the usability of multimodal large language models (MLLMs). However, existing end-to-end MLLMs only allow users to interact with them through language instructions, leading to the limitation of the interactive accuracy and efficiency. In this study, we present precise referring instructions that utilize diverse reference representations such as points and boxes as referring prompts to refer to the special region. This enables MLLMs to focus on the region of interest and achieve finer-grained interaction. Based on precise referring instruction, we propose ChatSpot, a unified end-to-end multimodal large language model that supports diverse forms of interactivity including mouse clicks, drag-and-drop, and drawing boxes, which provides a more flexible and seamless interactive experience. We also construct a multi-grained vision-language instruction-following dataset based on existing datasets and GPT-4 generating. Furthermore, we design a series of evaluation tasks to assess the effectiveness of region recognition and interaction. Experimental results showcase ChatSpot's promising performance.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes ChatSpot, a unified end-to-end multimodal large language model that supports diverse forms of interactivity including mouse clicks, drag-and-drop, and drawing boxes, which provides a more flexible and seamless interactive experience.'}",https://arxiv.org/pdf/2307.09474
-Self-Verification Improves Few-Shot Clinical Information Extraction,Zelalem Gero,"Extracting patient information from unstructured text is a critical task in health decision-support and clinical research. Large language models (LLMs) have shown the potential to accelerate clinical curation via few-shot in-context learning, in contrast to supervised learning which requires much more costly human annotations. However, despite drastic advances in modern LLMs such as GPT-4, they still struggle with issues regarding accuracy and interpretability, especially in mission-critical domains such as health. Here, we explore a general mitigation framework using self-verification, which leverages the LLM to provide provenance for its own extraction and check its own outputs. This is made possible by the asymmetry between verification and generation, where the latter is often much easier than the former. Experimental results show that our method consistently improves accuracy for various LLMs in standard clinical information extraction tasks. Additionally, self-verification yields interpretations in the form of a short text span corresponding to each output, which makes it very efficient for human experts to audit the results, paving the way towards trustworthy extraction of clinical information in resource-constrained scenarios. To facilitate future research in this direction, we release our code and prompts.","{'model': 'tldr@v2.0.0', 'text': 'This work explores a general mitigation framework using self-verification, which leverages the LLM to provide provenance for its own extraction and check its own outputs and consistently improves accuracy for various LLMs in standard clinical information extraction tasks.'}",http://arxiv.org/pdf/2306.00024
-ChipGPT: How far are we from natural language hardware design,Kaiyan Chang,"As large language models (LLMs) like ChatGPT exhibited unprecedented machine intelligence, it also shows great performance in assisting hardware engineers to realize higher-efficiency logic design via natural language interaction. To estimate the potential of the hardware design process assisted by LLMs, this work attempts to demonstrate an automated design environment that explores LLMs to generate hardware logic designs from natural language specifications. To realize a more accessible and efficient chip development flow, we present a scalable four-stage zero-code logic design framework based on LLMs without retraining or finetuning. At first, the demo, ChipGPT, begins by generating prompts for the LLM, which then produces initial Verilog programs. Second, an output manager corrects and optimizes these programs before collecting them into the final design space. Eventually, ChipGPT will search through this space to select the optimal design under the target metrics. The evaluation sheds some light on whether LLMs can generate correct and complete hardware logic designs described by natural language for some specifications. It is shown that ChipGPT improves programmability, and controllability, and shows broader design optimization space compared to prior work and native LLMs alone.","{'model': 'tldr@v2.0.0', 'text': 'This work attempts to demonstrate an automated design environment that explores LLMs to generate hardware logic designs from natural language specifications without retraining or finetuning, and shows broader design optimization space compared to prior work and native LLMs alone.'}",http://arxiv.org/pdf/2305.14019
-Ask an Expert: Leveraging Language Models to Improve Strategic Reasoning in Goal-Oriented Dialogue Models,Qiang Zhang,"Existing dialogue models may encounter scenarios which are not well-represented in the training data, and as a result generate responses that are unnatural, inappropriate, or unhelpful. We propose the""Ask an Expert""framework in which the model is trained with access to an""expert""which it can consult at each turn. Advice is solicited via a structured dialogue with the expert, and the model is optimized to selectively utilize (or ignore) it given the context and dialogue history. In this work the expert takes the form of an LLM. We evaluate this framework in a mental health support domain, where the structure of the expert conversation is outlined by pre-specified prompts which reflect a reasoning strategy taught to practitioners in the field. Blenderbot models utilizing""Ask an Expert""show quality improvements across all expert sizes, including those with fewer parameters than the dialogue model itself. Our best model provides a $\sim 10\%$ improvement over baselines, approaching human-level scores on""engingingness""and""helpfulness""metrics.","{'model': 'tldr@v2.0.0', 'text': 'The ""Ask an Expert"" framework is proposed, in which the model is trained with access to an expert which it can consult at each turn, and the expert takes the form of an LLM.'}",http://arxiv.org/pdf/2305.17878
-Zero-shot Visual Relation Detection via Composite Visual Cues from Large Language Models,Lin Li,"Pretrained vision-language models, such as CLIP, have demonstrated strong generalization capabilities, making them promising tools in the realm of zero-shot visual recognition. Visual relation detection (VRD) is a typical task that identifies relationship (or interaction) types between object pairs within an image. However, naively utilizing CLIP with prevalent class-based prompts for zero-shot VRD has several weaknesses, e.g., it struggles to distinguish between different fine-grained relation types and it neglects essential spatial information of two objects. To this end, we propose a novel method for zero-shot VRD: RECODE, which solves RElation detection via COmposite DEscription prompts. Specifically, RECODE first decomposes each predicate category into subject, object, and spatial components. Then, it leverages large language models (LLMs) to generate description-based prompts (or visual cues) for each component. Different visual cues enhance the discriminability of similar relation categories from different perspectives, which significantly boosts performance in VRD. To dynamically fuse different cues, we further introduce a chain-of-thought method that prompts LLMs to generate reasonable weights for different visual cues. Extensive experiments on four VRD benchmarks have demonstrated the effectiveness and interpretability of RECODE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method for zero-shot VRD: RECODE, which solves RElation detection via COmposite DEscription prompts, and introduces a chain-of-thought method that prompts LLMs to generate reasonable weights for different visual cues.'}",https://arxiv.org/pdf/2305.12476
-CHiLL: Zero-shot Custom Interpretable Feature Extraction from Clinical Notes with Large Language Models,Denis Jered McInerney,"We propose CHiLL (Crafting High-Level Latents), an approach for natural-language specification of features for linear models. CHiLL prompts LLMs with expert-crafted queries to generate interpretable features from health records. The resulting noisy labels are then used to train a simple linear classifier. Generating features based on queries to an LLM can empower physicians to use their domain expertise to craft features that are clinically meaningful for a downstream task of interest, without having to manually extract these from raw EHR. We are motivated by a real-world risk prediction task, but as a reproducible proxy, we use MIMIC-III and MIMIC-CXR data and standard predictive tasks (e.g., 30-day readmission) to evaluate this approach. We find that linear models using automatically extracted features are comparably performant to models using reference features, and provide greater interpretability than linear models using""Bag-of-Words""features. We verify that learned feature weights align well with clinical expectations.","{'model': 'tldr@v2.0.0', 'text': 'It is found thatlinear models using automatically extracted features are comparably performant to models using reference features, and provide greater interpretability than linear models using""Bag-of-Words"" features.'}",http://arxiv.org/pdf/2302.12343
-Could ChatGPT Imagine: Content Control for Artistic Painting Generation Via Large Language Models,Yue Lu,,,
-Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theory,Niloofar Mireshghallah,"The interactive use of large language models (LLMs) in AI assistants (at work, home, etc.) introduces a new set of inference-time privacy risks: LLMs are fed different types of information from multiple sources in their inputs and are expected to reason about what to share in their outputs, for what purpose and with whom, within a given context. In this work, we draw attention to the highly critical yet overlooked notion of contextual privacy by proposing ConfAIde, a benchmark designed to identify critical weaknesses in the privacy reasoning capabilities of instruction-tuned LLMs. Our experiments show that even the most capable models such as GPT-4 and ChatGPT reveal private information in contexts that humans would not, 39% and 57% of the time, respectively. This leakage persists even when we employ privacy-inducing prompts or chain-of-thought reasoning. Our work underscores the immediate need to explore novel inference-time privacy-preserving approaches, based on reasoning and theory of mind.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ConfAIde, a benchmark designed to identify critical weaknesses in the privacy reasoning capabilities of instruction-tuned LLMs, and underscores the immediate need to explore novel inference-time privacy-preserving approaches, based on reasoning and theory of mind.'}",
-Mitigating Approximate Memorization in Language Models via Dissimilarity Learned Policy,Aly M. Kassem,"Large Language models (LLMs) are trained on large amounts of data, which can include sensitive information that may compromise per- sonal privacy. LLMs showed to memorize parts of the training data and emit those data verbatim when an adversary prompts appropriately. Previous research has primarily focused on data preprocessing and differential privacy techniques to address memorization or prevent verbatim memorization exclusively, which can give a false sense of privacy. However, these methods rely on explicit and implicit assumptions about the structure of the data to be protected, which often results in an incomplete solution to the problem. To address this, we propose a novel framework that utilizes a reinforcement learning approach (PPO) to fine-tune LLMs to mitigate approximate memorization. Our approach utilizes a negative similarity score, such as BERTScore or SacreBLEU, as a reward signal to learn a dissimilarity policy. Our results demonstrate that this framework effectively mitigates approximate memorization while maintaining high levels of coherence and fluency in the generated samples. Furthermore, our framework is robust in mitigating approximate memorization across various circumstances, including longer context, which is known to increase memorization in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework that utilizes a reinforcement learning approach (PPO) to fine-tune LLMs to mitigate approximate memorization and utilizes a negative similarity score, such as BERTScore or SacreBLEU, as a reward signal to learn a dissimilarity policy.'}",http://arxiv.org/pdf/2305.01550
-Interleaving Pre-Trained Language Models and Large Language Models for Zero-Shot NL2SQL Generation,Zihui Gu,"Zero-shot NL2SQL is crucial in achieving natural language to SQL that is adaptive to new environments (e.g., new databases, new linguistic phenomena or SQL structures) with zero annotated NL2SQL samples from such environments. Existing approaches either fine-tune pre-trained language models (PLMs) based on annotated data or use prompts to guide fixed large language models (LLMs) such as ChatGPT. PLMs can perform well in schema alignment but struggle to achieve complex reasoning, while LLMs is superior in complex reasoning tasks but cannot achieve precise schema alignment. In this paper, we propose a ZeroNL2SQL framework that combines the complementary advantages of PLMs and LLMs for supporting zero-shot NL2SQL. ZeroNL2SQL first uses PLMs to generate an SQL sketch via schema alignment, then uses LLMs to fill the missing information via complex reasoning. Moreover, in order to better align the generated SQL queries with values in the given database instances, we design a predicate calibration method to guide the LLM in completing the SQL sketches based on the database instances and select the optimal SQL query via an execution-based strategy. Comprehensive experiments show that ZeroNL2SQL can achieve the best zero-shot NL2SQL performance on real-world benchmarks. Specifically, ZeroNL2SQL outperforms the state-of-the-art PLM-based methods by 3.2% to 13% and exceeds LLM-based methods by 10% to 20% on execution accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a ZeroNL2SQL framework that combines the complementary advantages of PLMs and LLMs for supporting zero-shot NL2SQL and designs a predicate calibration method to guide the LLM in completing the SQL sketches based on the database instances and select the optimal SQL query via an execution-based strategy.'}",http://arxiv.org/pdf/2306.08891
-Distinguish Before Answer: Generating Contrastive Explanation as Knowledge for Commonsense Question Answering,Qianglong Chen,"Existing knowledge-enhanced methods have achieved remarkable results in certain QA tasks via obtaining diverse knowledge from different knowledge bases. However, limited by the properties of retrieved knowledge, they still have trouble benefiting from both the knowledge relevance and distinguishment simultaneously. To address the challenge, we propose CPACE, a Concept-centric Prompt-bAsed Contrastive Explanation Generation model, which aims to convert obtained symbolic knowledge into a contrastive explanation for better distinguishing the differences among given candidates. Firstly, following previous works, we retrieve different types of symbolic knowledge with a concept-centric knowledge extraction module. After that, we generate corresponding contrastive explanations using acquired symbolic knowledge and explanation prompts as guidance for better modeling the knowledge distinguishment and interpretability. Finally, we regard the generated contrastive explanation as external knowledge for downstream task enhancement. We conduct a series of experiments on three widely-used question-answering datasets: CSQA, QASC, and OBQA. Experimental results demonstrate that with the help of generated contrastive explanation, our CPACE model achieves new SOTA on CSQA (89.8% on the testing set, 0.9% higher than human performance), and gains impressive improvement on QASC and OBQA (4.2% and 3.5%, respectively).","{'model': 'tldr@v2.0.0', 'text': 'CPACE, a Concept-centric Prompt-bAsed Contrastive Explanation Generation model, which aims to convert obtained symbolic knowledge into a contrastive explanation for better distinguishing the differences among given candidates is proposed.'}",http://arxiv.org/pdf/2305.08135
-Using Global Land Cover Product as Prompt for Cropland Mapping via Visual Foundation Model,Chao Tao,"Data-driven deep learning methods have shown great potential in cropland mapping. However, due to multiple factors such as attributes of cropland (topography, climate, crop type) and imaging conditions (viewing angle, illumination, scale), croplands under different scenes demonstrate a great domain gap. This makes it difficult for models trained in the specific scenes to directly generalize to other scenes. A common way to handle this problem is through the""Pretrain+Fine-tuning""paradigm. Unfortunately, considering the variety of features of cropland that are affected by multiple factors, it is hardly to handle the complex domain gap between pre-trained data and target data using only sparse fine-tuned samples as general constraints. Moreover, as the number of model parameters grows, fine-tuning is no longer an easy and low-cost task. With the emergence of prompt learning via visual foundation models, the""Pretrain+Prompting""paradigm redesigns the optimization target by introducing individual prompts for each single sample. This simplifies the domain adaption from generic to specific scenes during model reasoning processes. Therefore, we introduce the""Pretrain+Prompting""paradigm to interpreting cropland scenes and design the auto-prompting (APT) method based on freely available global land cover product. It can achieve a fine-grained adaptation process from generic scenes to specialized cropland scenes without introducing additional label costs. To our best knowledge, this work pioneers the exploration of the domain adaption problems for cropland mapping under prompt learning perspectives. Our experiments using two sub-meter cropland datasets from southern and northern China demonstrated that the proposed method via visual foundation models outperforms traditional supervised learning and fine-tuning approaches in the field of remote sensing.","{'model': 'tldr@v2.0.0', 'text': 'This work pioneers the exploration of the domain adaption problems for cropland mapping under prompt learning perspectives and designs the auto-prompting (APT) method based on freely available global land cover product that can achieve a fine-grained adaptation process from generic scenes to specialized croplands scenes without introducing additional label costs.'}",
-EPA: Easy Prompt Augmentation on Large Language Models via Multiple Sources and Multiple Targets,Hongyuan Lu,"Large language models (LLMs) have shown promising performance on various NLP tasks via task prompting. And their performance can be further improved by appending task demonstrations to the head of the prompt. And usually, a better performance can be achieved with more demonstrations. However, asking the users to write the demonstrations can be cumbersome. As a simple yet cost-effective workaround, this paper proposes a novel method called EPA (\textbf{E}asy \textbf{P}rompt \textbf{A}ugmentation)\footnote{While this paper considers augmenting prompts via demonstrations, we name it EPA as the name EDA is already taken by a well-known NLP method \citep{wei-zou-2019-eda}.} that effectively minimizes user efforts in writing demonstrations while improving the model performance at the same time. EPA achieves these goals by automatically augmenting the demonstrations with multiple sources/targets, where each of them paraphrases each other. This is well motivated as augmenting data via paraphrasing effectively improves neural language models. EPA thus employs paraphrasing as an augmentation method for in-context learning. Extensive experiments indicate that EPA effectively improves both NLU and NLG tasks, covering from natural language inference to machine translation in translating tens of languages.\footnote{Code and data will be released upon publication.}","{'model': 'tldr@v2.0.0', 'text': 'A novel method called EPA that effectively minimizes user efforts in writing demonstrations while improving the model performance at the same time, and automatically augmenting the demonstrations with multiple sources/targets, where each of them paraphrasing each other.'}",https://arxiv.org/pdf/2309.04725
-ExpCLIP: Bridging Text and Facial Expressions via Semantic Alignment,Yicheng Zhong,"The objective of stylized speech-driven facial animation is to create animations that encapsulate specific emotional expressions. Existing methods often depend on pre-established emotional labels or facial expression templates, which may limit the necessary flexibility for accurately conveying user intent. In this research, we introduce a technique that enables the control of arbitrary styles by leveraging natural language as emotion prompts. This technique presents benefits in terms of both flexibility and user-friendliness. To realize this objective, we initially construct a Text-Expression Alignment Dataset (TEAD), wherein each facial expression is paired with several prompt-like descriptions.We propose an innovative automatic annotation method, supported by Large Language Models (LLMs), to expedite the dataset construction, thereby eliminating the substantial expense of manual annotation. Following this, we utilize TEAD to train a CLIP-based model, termed ExpCLIP, which encodes text and facial expressions into semantically aligned style embeddings. The embeddings are subsequently integrated into the facial animation generator to yield expressive and controllable facial animations. Given the limited diversity of facial emotions in existing speech-driven facial animation training data, we further introduce an effective Expression Prompt Augmentation (EPA) mechanism to enable the animation generator to support unprecedented richness in style control. Comprehensive experiments illustrate that our method accomplishes expressive facial animation generation and offers enhanced flexibility in effectively conveying the desired style.","{'model': 'tldr@v2.0.0', 'text': 'An innovative automatic annotation method is proposed, supported by Large Language Models, to expedite the dataset construction, thereby eliminating the substantial expense of manual annotation and introducing an effective Expression Prompt Augmentation mechanism to enable the animation generator to support unprecedented richness in style control.'}",https://arxiv.org/pdf/2308.14448
-DIVKNOWQA: Assessing the Reasoning Ability of LLMs via Open-Domain Question Answering over Knowledge Base and Text,Wenting Zhao,"Large Language Models (LLMs) have exhibited impressive generation capabilities, but they suffer from hallucinations when solely relying on their internal knowledge, especially when answering questions that require less commonly known information. Retrieval-augmented LLMs have emerged as a potential solution to ground LLMs in external knowledge. Nonetheless, recent approaches have primarily emphasized retrieval from unstructured text corpora, owing to its seamless integration into prompts. When using structured data such as knowledge graphs, most methods simplify it into natural text, neglecting the underlying structures. Moreover, a significant gap in the current landscape is the absence of a realistic benchmark for evaluating the effectiveness of grounding LLMs on heterogeneous knowledge sources (e.g., knowledge base and text). To fill this gap, we have curated a comprehensive dataset that poses two unique challenges: (1) Two-hop multi-source questions that require retrieving information from both open-domain structured and unstructured knowledge sources; retrieving information from structured knowledge sources is a critical component in correctly answering the questions. (2) The generation of symbolic queries (e.g., SPARQL for Wikidata) is a key requirement, which adds another layer of challenge. Our dataset is created using a combination of automatic generation through predefined reasoning chains and human annotation. We also introduce a novel approach that leverages multiple retrieval tools, including text passage retrieval and symbolic language-assisted retrieval. Our model outperforms previous approaches by a significant margin, demonstrating its effectiveness in addressing the above-mentioned reasoning challenges.","{'model': 'tldr@v2.0.0', 'text': 'This work curated a comprehensive dataset that poses two unique challenges: two-hop multi-source questions that require retrieving information from both open-domain structured and unstructured knowledge sources; and a novel approach that leverages multiple retrieval tools, including text passage retrieval and symbolic language-assisted retrieval.'}",
-Tailoring Personality Traits in Large Language Models via Unsupervisedly-Built Personalized Lexicons,Tianlong Li,"Personality plays a pivotal role in shaping human expression patterns, and empowering and manipulating large language models (LLMs) with personality traits holds significant promise in enhancing the user experience of LLMs. However, prior approaches either rely on fine-tuning LLMs on a corpus enriched with personalized expressions or necessitate the manual crafting of prompts to induce LLMs to produce personalized responses. The former approaches demand substantial time and resources for collecting sufficient training examples while the latter might fail in enabling the precise manipulation of the personality traits at a fine-grained level (e.g., achieving high agreeableness while reducing openness). In this study, we introduce a novel approach for tailoring personality traits within LLMs, allowing for the incorporation of any combination of the Big Five factors (i.e., openness, conscientiousness, extraversion, agreeableness, and neuroticism) in a pluggable manner. This is achieved by employing a set of Unsupervisedly-Built Personalized Lexicons (UBPL) that are utilized to adjust the probability of the next token predicted by the original LLMs during the decoding phase. This adjustment encourages the models to generate words present in the personalized lexicons while preserving the naturalness of the generated texts. Extensive experimentation demonstrates the effectiveness of our approach in finely manipulating LLMs' personality traits. Furthermore, our method can be seamlessly integrated into other LLMs without necessitating updates to their parameters.",,
-Denevil: Towards Deciphering and Navigating the Ethical Values of Large Language Models via Instruction Learning,Shitong Duan,"Large Language Models (LLMs) have made unprecedented breakthroughs, yet their increasing integration into everyday life might raise societal risks due to generated unethical content. Despite extensive study on specific issues like bias, the intrinsic values of LLMs remain largely unexplored from a moral philosophy perspective. This work delves into ethical values utilizing Moral Foundation Theory. Moving beyond conventional discriminative evaluations with poor reliability, we propose DeNEVIL, a novel prompt generation algorithm tailored to dynamically exploit LLMs' value vulnerabilities and elicit the violation of ethics in a generative manner, revealing their underlying value inclinations. On such a basis, we construct MoralPrompt, a high-quality dataset comprising 2,397 prompts covering 500+ value principles, and then benchmark the intrinsic values across a spectrum of LLMs. We discovered that most models are essentially misaligned, necessitating further ethical value alignment. In response, we develop VILMO, an in-context alignment method that substantially enhances the value compliance of LLM outputs by learning to generate appropriate value instructions, outperforming existing competitors. Our methods are suitable for black-box and open-source models, offering a promising initial step in studying the ethical values of LLMs.","{'model': 'tldr@v2.0.0', 'text': ""DeNEVIL, a novel prompt generation algorithm tailored to dynamically exploit LLMs' value vulnerabilities and elicit the violation of ethics in a generative manner, revealing their underlying value inclinations is proposed.""}",
-PiTL: Cross-modal Retrieval with Weakly-supervised Vision-language Pre-training via Prompting,Zixin Guo,"Vision-language (VL) Pre-training (VLP) has shown to well generalize VL models over a wide range of VL downstream tasks, especially for cross-modal retrieval. However, it hinges on a huge amount of image-text pairs, which requires tedious and costly curation. On the contrary,weakly-supervised VLP (W-VLP) explores means with object tags generated by a pre-trained object detector (OD) from images. Yet, they still require paired information, i.e. images and object-level annotations, as supervision to train an OD. To further reduce the amount of supervision, we propose Prompts-in-The-Loop (PiTL) that prompts knowledge from large language models (LLMs) to describe images. Concretely, given a category label of an image, e.g.refinery, the knowledge, e.g.a refinery could be seen with large storage tanks, pipework, and ..., extracted by LLMs is used as the language counterpart. The knowledge supplements, e.g. the common relations among entities most likely appearing in a scene. We create IN14K, a new VL dataset of 9M images and 1M descriptions of 14K categories from ImageNet21K with PiTL. Empirically, the VL models pre-trained with PiTL-generated pairs are strongly favored over other W-VLP works on image-to-text (I2T) and text-to-image (T2I) retrieval tasks, with less supervision. The results reveal the effectiveness of PiTL-generated pairs for VLP.","{'model': 'tldr@v2.0.0', 'text': 'Empirically, the VL models pre-trained with PiTL-generated pairs are strongly favored over other W-VLP works on image-to-text (I2T) and text- to-image (T2I) retrieval tasks, with less supervision.'}",https://dl.acm.org/doi/pdf/10.1145/3539618.3592038
-In-Context Learning Dynamics with Random Binary Sequences,Eric J. Bigelow,"Large language models (LLMs) trained on huge corpora of text datasets demonstrate complex, emergent capabilities, achieving state-of-the-art performance on tasks they were not explicitly trained for. The precise nature of LLM capabilities is often mysterious, and different prompts can elicit different capabilities through in-context learning. We propose a Cognitive Interpretability framework that enables us to analyze in-context learning dynamics to understand latent concepts in LLMs underlying behavioral patterns. This provides a more nuanced understanding than success-or-failure evaluation benchmarks, but does not require observing internal activations as a mechanistic interpretation of circuits would. Inspired by the cognitive science of human randomness perception, we use random binary sequences as context and study dynamics of in-context learning by manipulating properties of context data, such as sequence length. In the latest GPT-3.5+ models, we find emergent abilities to generate pseudo-random numbers and learn basic formal languages, with striking in-context learning dynamics where model outputs transition sharply from pseudo-random behaviors to deterministic repetition.","{'model': 'tldr@v2.0.0', 'text': 'A Cognitive Interpretability framework is proposed that enables us to analyze in-context learning dynamics to understand latent concepts in LLMs underlying behavioral patterns, providing a more nuanced understanding than success-or-failure evaluation benchmarks, but does not require observing internal activations as a mechanistic interpretation of circuits would.'}",
-Vision-Language Interpreter for Robot Task Planning,Keisuke Shirai,"Large language models (LLMs) are accelerating the development of language-guided robot planners. Meanwhile, symbolic planners offer the advantage of interpretability. This paper proposes a new task that bridges these two trends, namely, multimodal planning problem specification. The aim is to generate a problem description (PD), a machine-readable file used by the planners to find a plan. By generating PDs from language instruction and scene observation, we can drive symbolic planners in a language-guided framework. We propose a Vision-Language Interpreter (ViLaIn), a new framework that generates PDs using state-of-the-art LLM and vision-language models. ViLaIn can refine generated PDs via error message feedback from the symbolic planner. Our aim is to answer the question: How accurately can ViLaIn and the symbolic planner generate valid robot plans? To evaluate ViLaIn, we introduce a novel dataset called the problem description generation (ProDG) dataset. The framework is evaluated with four new evaluation metrics. Experimental results show that ViLaIn can generate syntactically correct problems with more than 99% accuracy and valid plans with more than 58% accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A Vision-Language Interpreter (ViLaIn), a new framework that generates PDs using state-of-the-art LLM and vision-language models, to answer the question: How accurately can ViLaIn and the symbolic planner generate valid robot plans?'}",
-Prompting with Pseudo-Code Instructions,Mayank Mishra,"Prompting with natural language instructions has recently emerged as a popular method of harnessing the capabilities of large language models. Given the inherent ambiguity present in natural language, it is intuitive to consider the possible advantages of prompting with less ambiguous prompt styles, such as the use of pseudo-code. In this paper we explore if prompting via pseudo-code instructions helps improve the performance of pre-trained language models. We manually create a dataset of pseudo-code prompts for 132 different tasks spanning classification, QA and generative language tasks, sourced from the Super-NaturalInstructions dataset. Using these prompts along with their counterparts in natural language, we study their performance on two LLM families - BLOOM and CodeGen. Our experiments show that using pseudo-code instructions leads to better results, with an average increase (absolute) of 7-16 points in F1 scores for classification tasks and an improvement (relative) of 12-38% in aggregate ROUGE-L scores across all tasks. We include detailed ablation studies which indicate that code comments, docstrings, and the structural clues encoded in pseudo-code all contribute towards the improvement in performance. To the best of our knowledge our work is the first to demonstrate how pseudo-code prompts can be helpful in improving the performance of pre-trained LMs.","{'model': 'tldr@v2.0.0', 'text': 'This work is the first to demonstrate how pseudo-code prompts can be helpful in improving the performance of pre-trained LMs, with detailed ablation studies which indicate that code comments, docstrings, and the structural clues encoded in pseudo- code all contribute towards the improvement in performance.'}",http://arxiv.org/pdf/2305.11790
-Towards General Visual-Linguistic Face Forgery Detection,Ke Sun,"Deepfakes are realistic face manipulations that can pose serious threats to security, privacy, and trust. Existing methods mostly treat this task as binary classification, which uses digital labels or mask signals to train the detection model. We argue that such supervisions lack semantic information and interpretability. To address this issues, in this paper, we propose a novel paradigm named Visual-Linguistic Face Forgery Detection(VLFFD), which uses fine-grained sentence-level prompts as the annotation. Since text annotations are not available in current deepfakes datasets, VLFFD first generates the mixed forgery image with corresponding fine-grained prompts via Prompt Forgery Image Generator (PFIG). Then, the fine-grained mixed data and coarse-grained original data and is jointly trained with the Coarse-and-Fine Co-training framework (C2F), enabling the model to gain more generalization and interpretability. The experiments show the proposed method improves the existing detection models on several challenging benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A novel paradigm named Visual-Linguistic Face Forgery Detection (VLFFD), which uses fine-grained sentence-level prompts as the annotation, and improves the existing detection models on several challenging benchmarks.'}",https://arxiv.org/pdf/2307.16545
-HeaP: Hierarchical Policies for Web Actions using LLMs,Paloma Sodhi,"Large language models (LLMs) have demonstrated remarkable capabilities in performing a range of instruction following tasks in few and zero-shot settings. However, teaching LLMs to perform tasks on the web presents fundamental challenges -- combinatorially large open-world tasks and variations across web interfaces. We tackle these challenges by leveraging LLMs to decompose web tasks into a collection of sub-tasks, each of which can be solved by a low-level, closed-loop policy. These policies constitute a shared grammar across tasks, i.e., new web tasks can be expressed as a composition of these policies. We propose a novel framework, Hierarchical Policies for Web Actions using LLMs (HeaP), that learns a set of hierarchical LLM prompts from demonstrations for planning high-level tasks and executing them via a sequence of low-level policies. We evaluate HeaP against a range of baselines on a suite of web tasks, including MiniWoB++, WebArena, a mock airline CRM, as well as live website interactions, and show that it is able to outperform prior works using orders of magnitude less data.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, Hierarchical Policies for Web Actions using LLMs (HeaP), that learns a set of hierarchical LLM prompts from demonstrations for planning high-level tasks and executing them via a sequence of low-level policies, and is able to outperform prior works using orders of magnitude less data.'}",https://arxiv.org/pdf/2310.03720
-Forgetful Large Language Models: Lessons Learned from Using LLMs in Robot Programming,Juo-Tung Chen,"Large language models offer new ways of empowering people to program robot applications-namely, code generation via prompting. However, the code generated by LLMs is susceptible to errors. This work reports a preliminary exploration that empirically characterizes common errors produced by LLMs in robot programming. We categorize these errors into two phases: interpretation and execution. In this work, we focus on errors in execution and observe that they are caused by LLMs being""forgetful""of key information provided in user prompts. Based on this observation, we propose prompt engineering tactics designed to reduce errors in execution. We then demonstrate the effectiveness of these tactics with three language models: ChatGPT, Bard, and LLaMA-2. Finally, we discuss lessons learned from using LLMs in robot programming and call for the benchmarking of LLM-powered end-user development of robot applications.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt engineering tactics designed to reduce errors in execution of LLMs in robot programming and demonstrates the effectiveness of these tactics with three language models: ChatGPT, Bard, and LLaMA-2.'}",https://arxiv.org/pdf/2310.06646
-Interpretable Unified Language Checking,Tianhua Zhang,"Despite recent concerns about undesirable behaviors generated by large language models (LLMs), including non-factual, biased, and hateful language, we find LLMs are inherent multi-task language checkers based on their latent representations of natural and social knowledge. We present an interpretable, unified, language checking (UniLC) method for both human and machine-generated language that aims to check if language input is factual and fair. While fairness and fact-checking tasks have been handled separately with dedicated models, we find that LLMs can achieve high performance on a combination of fact-checking, stereotype detection, and hate speech detection tasks with a simple, few-shot, unified set of prompts. With the ``1/2-shot'' multi-task language checking method proposed in this work, the GPT3.5-turbo model outperforms fully supervised baselines on several language tasks. The simple approach and results suggest that based on strong latent knowledge representations, an LLM can be an adaptive and explainable tool for detecting misinformation, stereotypes, and hate speech.","{'model': 'tldr@v2.0.0', 'text': 'It is found that LLMs can achieve high performance on a combination of fact-checking, stereotype detection, and hate speech detection tasks with a simple, few-shot, unified set of prompts.'}",http://arxiv.org/pdf/2304.03728
-GenRec: Large Language Model for Generative Recommendation,Jianchao Ji,"In recent years, large language models (LLM) have emerged as powerful tools for diverse natural language processing tasks. However, their potential for recommender systems under the generative recommendation paradigm remains relatively unexplored. This paper presents an innovative approach to recommendation systems using large language models (LLMs) based on text data. In this paper, we present a novel LLM for generative recommendation (GenRec) that utilized the expressive power of LLM to directly generate the target item to recommend, rather than calculating ranking score for each candidate item one by one as in traditional discriminative recommendation. GenRec uses LLM's understanding ability to interpret context, learn user preferences, and generate relevant recommendation. Our proposed approach leverages the vast knowledge encoded in large language models to accomplish recommendation tasks. We first we formulate specialized prompts to enhance the ability of LLM to comprehend recommendation tasks. Subsequently, we use these prompts to fine-tune the LLaMA backbone LLM on a dataset of user-item interactions, represented by textual data, to capture user preferences and item characteristics. Our research underscores the potential of LLM-based generative recommendation in revolutionizing the domain of recommendation systems and offers a foundational framework for future explorations in this field. We conduct extensive experiments on benchmark datasets, and the experiments shows that our GenRec has significant better results on large dataset.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel LLM for generative recommendation (GenRec) that utilized the expressive power of LLM to directly generate the target item to recommend, rather than calculating ranking score for each candidate item one by one as in traditional discriminative recommendation.'}",https://arxiv.org/pdf/2307.00457
-Differentially Private Decoding in Large Language Models,Jimit Majmudar,"Recent large-scale natural language processing (NLP) systems use a pre-trained Large Language Model (LLM) on massive and diverse corpora as a headstart. In practice, the pre-trained model is adapted to a wide array of tasks via fine-tuning on task-specific datasets. LLMs, while effective, have been shown to memorize instances of training data thereby potentially revealing private information processed during pre-training. The potential leakage might further propagate to the downstream tasks for which LLMs are fine-tuned. On the other hand, privacy-preserving algorithms usually involve retraining from scratch, which is prohibitively expensive for LLMs. In this work, we propose a simple, easy to interpret, and computationally lightweight perturbation mechanism to be applied to an already trained model at the decoding stage. Our perturbation mechanism is model-agnostic and can be used in conjunction with any LLM. We provide theoretical analysis showing that the proposed mechanism is differentially private, and experimental results showing a privacy-utility trade-off.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple, easy to interpret, and computationally lightweight perturbation mechanism to be applied to an already trained model at the decoding stage, which is model-agnostic and can be used in conjunction with any LLM.'}",http://arxiv.org/pdf/2205.13621
-"A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity",Yejin Bang,"This paper proposes a framework for quantitatively evaluating interactive LLMs such as ChatGPT using publicly available data sets. We carry out an extensive technical evaluation of ChatGPT using 23 data sets covering 8 different common NLP application tasks. We evaluate the multitask, multilingual and multi-modal aspects of ChatGPT based on these data sets and a newly designed multimodal dataset. We find that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks. We find that it is better at understanding non-Latin script languages than generating them. It is able to generate multimodal content from textual prompts, via an intermediate code generation step. Moreover, we find that ChatGPT is 63.41% accurate on average in 10 different reasoning categories under logical reasoning, non-textual reasoning, and commonsense reasoning, hence making it an unreliable reasoner. It is, for example, better at deductive than inductive reasoning. ChatGPT suffers from hallucination problems like other LLMs and it generates more extrinsic hallucinations from its parametric memory as it does not have access to an external knowledge base. Finally, the interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance, i.e, 8% ROUGE-1 on summarization and 2% ChrF++ on machine translation, in a multi-turn""prompt engineering""fashion. We also release codebase for evaluation set extraction.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks and is better at understanding non-Latin script languages than generating them.'}",http://arxiv.org/pdf/2302.04023
-Chain-of-Thought Prompt Distillation for Multimodal Named Entity and Multimodal Relation Extraction,F. Chen,"Multimodal Named Entity Recognition (MNER) and Multimodal Relation Extraction (MRE) necessitate the fundamental reasoning capacity for intricate linguistic and multimodal comprehension. In this study, we explore distilling the reasoning ability of large language models (LLMs) into a more compact student model by generating a chain of thought (CoT) – a sequence of intermediate reasoning steps. Specifically, we commence by exemplifying the elicitation of such reasoning ability from LLMs through CoT prompts covering multi-grain (noun, sentence, multimodality) and data-augmentation (style, entity, image) dimensions. Subsequently, we present a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge. Extensive experiments reveal that our approach attains state-of-theart accuracy and manifests a plethora of advantages concerning interpretability, data efficiency, and cross-domain generalization on MNER and MRE datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge.'}",https://arxiv.org/pdf/2306.14122
-FedLogic: Interpretable Federated Multi-Domain Chain-of-Thought Prompt Selection for Large Language Models,Pengwei Xing,"Leveraging ``chain-of-thought (CoT)'' reasoning to elicit rapid and precise responses from large language models (LLMs) is rapidly attracting research interest. A notable challenge here is how to design or select optimal prompts. The process of prompt selection relies on trial and error, involving continuous adjustments and combinations of input prompts by users based on the corresponding new responses generated from LLMs. Furthermore, minimal research has been conducted to explore how LLMs employ the mathematical problem-solving capabilities learned from user interactions to address issues in narrative writing. To improve interpretability and explore the balance principle between generality and personalization under a multi-domain CoT prompt selection scenario, we propose the Federated Logic rule learning approach (FedLogic). We introduce a theoretical formalization and interactive emulation of the multi-domain CoT prompt selection dilemma in the context of federated LLMs. We cast the problem of joint probability modeling as a bilevel program, where the CoT prompt selection intricacy can be likened to a fuzzy score-based rule selection with the LLMs function as rule generators. FedLogic solves this problem through variational expectation maximization (V-EM). In addition, we incorporate two KL-divergence constraints within this probabilistic modeling framework to surmount the intricacies of managing extensive search spaces and accomplishing cross-domain personalization of CoTs. To the best of our knowledge, FedLogic is the first interpretable and principled federated multi-domain CoT prompt selection approach for LLMs.","{'model': 'tldr@v2.0.0', 'text': 'FedLogic is the first interpretable and principled federated multi-domain CoT prompt selection approach for LLMs and incorporates two KL-divergence constraints within this probabilistic modeling framework to surmount the intricacies of managing extensive search spaces and accomplishing cross-domain personalization of CoTs.'}",https://arxiv.org/pdf/2308.15324
-Learning Concise and Descriptive Attributes for Visual Recognition,Andy Yan,"Recent advances in foundation models present new opportunities for interpretable visual recognition -- one can first query Large Language Models (LLMs) to obtain a set of attributes that describe each class, then apply vision-language models to classify images via these attributes. Pioneering work shows that querying thousands of attributes can achieve performance competitive with image features. However, our further investigation on 8 datasets reveals that LLM-generated attributes in a large quantity perform almost the same as random words. This surprising finding suggests that significant noise may be present in these attributes. We hypothesize that there exist subsets of attributes that can maintain the classification performance with much smaller sizes, and propose a novel learning-to-search method to discover those concise sets of attributes. As a result, on the CUB dataset, our method achieves performance close to that of massive LLM-generated attributes (e.g., 10k attributes for CUB), yet using only 32 attributes in total to distinguish 200 bird species. Furthermore, our new paradigm demonstrates several additional benefits: higher interpretability and interactivity for humans, and the ability to summarize knowledge for a recognition task.","{'model': 'tldr@v2.0.0', 'text': 'This work hypothesizes that there exist subsets of attributes that can maintain the classification performance with much smaller sizes, and proposes a novel learning-to-search method to discover those concise sets of attributes.'}",https://arxiv.org/pdf/2308.03685
-Machine learning generated decision boundaries for prediction and exploration of patient-specific quality assurance failures in Stereotactic Radiosurgery plans.,Jeremy Braun,"INTRODUCTION Stereotactic Radiosurgery (SRS) is a form of radiotherapy treatment during which high radiation dose is delivered in a single or few fractions. These treatments require highly conformal plans with steep dose gradients which can result in an increase in plan complexity prompting the need for stringent pre-treatment patient specific quality assurance (QA) measurements to ensure the planned and measured dose distributions agree within clinical standards. Complexity scores and machine learning (ML) techniques may help with prediction of QA outcomes however interpretability and usability of those results continues to be an area of study. This study investigates the use of plan complexity metrics as input for an ML model to allow for prediction of QA outcomes for SRS plans as measured via 3D phantom dose verification. Explorations into interpretability and predictive performance changes as model dimensionality increases, as well as a prospective in-clinic implementation using the resulting model were also performed. METHODS 498 plans (1571 VMAT arcs) were processed via in-house script to generate several complexity scores. 3D phantom dose verification measurement results were extracted and classified as pass or failure (with failures defined as below 95% voxel agreement passing 3%/1mm gamma criteria with 10% threshold,) and 1472 of the arcs were split into training and testing sets, with 99 arcs as a sequential holdout set. A z-score scaler was trained on the training set and used to scale all other sets. Variations of MLC leaf movement variability, aperture complexity, and leaf size and MU at control point weighted target area scores were used as input to a Support Vector Classifier to generate a series of 1-D, 2-D, and 5-D decision boundaries. The best performing 5D model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements for 38 patient plans (112 arcs). The decision to order 3D phantom dose verification measurements was recorded before and after prediction. RESULTS Best performing 1-D threshold, and 2-D prediction models with best performance produced a QA failure recall and QA passing recall of 1.00 and 0.55, and 0.82 and 0.82 respectively. Best performing 5-D prediction model produced a QA failure recall (sensitivity) of 1.00, and QA passing recall (specificity) of 0.72. This model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements and achieved a QA failure recall of 1.00 and QA passing recall of 0.58. The decision to order 3D phantom dose verification measurements was recorded before and after measurement. A single initially unidentified failing plan of the prospective cohort was successfully predicted to fail by the model. CONCLUSION Implementation of complexity score based prediction models for SRS would allow for support of a clinician's decision to reduce time spent performing QA measurements, and avoid patient treatment delays (i.e. in case of QA failure). This article is protected by copyright. All rights reserved.","{'model': 'tldr@v2.0.0', 'text': ""Implementation of complexity score based prediction models for SRS would allow for support of a clinician's decision to reduce time spent performing QA measurements, and avoid patient treatment delays (i.e. in case of QA failure).""}",
-Tackling Vision Language Tasks Through Learning Inner Monologues,Diji Yang,"Visual language tasks require AI models to comprehend and reason with both visual and textual content. Driven by the power of Large Language Models (LLMs), two prominent methods have emerged: (1) the hybrid integration between LLMs and Vision-Language Models (VLMs), where visual inputs are firstly converted into language descriptions by VLMs, serving as inputs for LLMs to generate final answer(s); (2) visual feature alignment in language space, where visual inputs are encoded as embeddings and projected to LLMs' language space via further supervised fine-tuning. The first approach provides light training costs and interpretability but is hard to be optimized in an end-to-end fashion. The second approach presents decent performance, but feature alignment usually requires large amounts of training data and lacks interpretability. To tackle this dilemma, we propose a novel approach, Inner Monologue Multi-Modal Optimization (IMMO), to solve complex vision language problems by simulating inner monologue processes, a cognitive process in which an individual engages in silent verbal communication with themselves. We enable LLMs and VLMs to interact through natural language conversation and propose to use a two-stage training process to learn how to do the inner monologue (self-asking questions and answering questions). IMMO is evaluated on two popular tasks and the results suggest by emulating the cognitive phenomenon of internal dialogue, our approach can enhance reasoning and explanation abilities, contributing to the more effective fusion of vision and language models. More importantly, instead of using predefined human-crafted monologues, IMMO learns this process within the deep learning models, promising wider applicability to many different AI problems beyond vision language tasks.","{'model': 'tldr@v2.0.0', 'text': 'Inner Monologue Multi-Modal Optimization is proposed, to solve complex vision language problems by simulating inner monologue processes, a cognitive process in which an individual engages in silent verbal communication with themselves, promising wider applicability to many different AI problems beyond vision language tasks.'}",https://arxiv.org/pdf/2308.09970
-Robust Preference Learning for Storytelling via Contrastive Reinforcement Learning,Louis Castricato,"Controlled automated story generation seeks to generate natural language stories satisfying constraints from natural language critiques or preferences. Existing methods to control for story preference utilize prompt engineering which is labor intensive and often inconsistent. They may also use logit-manipulation methods which require annotated datasets to exist for the desired attributes. To address these issues, we first train a contrastive bi-encoder model to align stories with corresponding human critiques, named CARP, building a general purpose preference model. This is subsequently used as a reward function to fine-tune a generative language model via reinforcement learning. However, simply fine-tuning a generative language model with a contrastive reward model does not always reliably result in a story generation system capable of generating stories that meet user preferences. To increase story generation robustness we further fine-tune the contrastive reward model using a prompt-learning technique. A human participant study is then conducted comparing generations from our full system, ablations, and two baselines. We show that the full fine-tuning pipeline results in a story generator preferred over a LLM 20x as large as well as logit-based methods. This motivates the use of contrastive learning for general purpose human preference modeling.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the full fine-tuning pipeline results in a story generator preferred over a LLM 20x as large as well as logit-based methods, and motivates the use of contrastive learning for general purpose human preference modeling.'}",http://arxiv.org/pdf/2210.07792
-The Power of Large Language Models for Wireless Communication System Development: A Case Study on FPGA Platforms,Yuyang Du,"Large language models (LLMs) have garnered significant attention across various research disciplines, including the wireless communication community. There have been several heated discussions on the intersection of LLMs and wireless technologies. While recent studies have demonstrated the ability of LLMs to generate hardware description language (HDL) code for simple computation tasks, developing wireless prototypes and products via HDL poses far greater challenges because of the more complex computation tasks involved. In this paper, we aim to address this challenge by investigating the role of LLMs in FPGA-based hardware development for advanced wireless signal processing. We begin by exploring LLM-assisted code refactoring, reuse, and validation, using an open-source software-defined radio (SDR) project as a case study. Through the case study, we find that an LLM assistant can potentially yield substantial productivity gains for researchers and developers. We then examine the feasibility of using LLMs to generate HDL code for advanced wireless signal processing, using the Fast Fourier Transform (FFT) algorithm as an example. This task presents two unique challenges: the scheduling of subtasks within the overall task and the multi-step thinking required to solve certain arithmetic problem within the task. To address these challenges, we employ in-context learning (ICL) and Chain-of-Thought (CoT) prompting techniques, culminating in the successful generation of a 64-point Verilog FFT module. Our results demonstrate the potential of LLMs for generalization and imitation, affirming their usefulness in writing HDL code for wireless communication systems. Overall, this work contributes to understanding the role of LLMs in wireless communication and motivates further exploration of their capabilities.","{'model': 'tldr@v2.0.0', 'text': 'This work begins by exploring LLM-assisted code refactoring, reuse, and validation, using an open-source software-defined radio (SDR) project as a case study, and examines the feasibility of using LLMs to generate HDL code for advanced wireless signal processing.'}",
-FinVis-GPT: A Multimodal Large Language Model for Financial Chart Analysis,Ziao Wang,"In this paper, we propose FinVis-GPT, a novel multimodal large language model (LLM) specifically designed for financial chart analysis. By leveraging the power of LLMs and incorporating instruction tuning and multimodal capabilities, FinVis-GPT is capable of interpreting financial charts and providing valuable analysis. To train FinVis-GPT, a financial task oriented dataset was generated for pre-training alignment and instruction tuning, comprising various types of financial charts and their corresponding descriptions. We evaluate the model performance via several case studies due to the time limit, and the promising results demonstrated that FinVis-GPT is superior in various financial chart related tasks, including generating descriptions, answering questions and predicting future market trends, surpassing existing state-of-the-art multimodal LLMs. The proposed FinVis-GPT serves as a pioneering effort in utilizing multimodal LLMs in the finance domain and our generated dataset will be release for public use in the near future to speedup related research.","{'model': 'tldr@v2.0.0', 'text': 'The promising results demonstrated that FinVis-GPT is superior in various financial chart related tasks, including generating descriptions, answering questions and predicting future market trends, surpassing existing state-of-the-art multimodal LLMs.'}",https://arxiv.org/pdf/2308.01430
-Interpreting Reward Models in RLHF-Tuned Language Models Using Sparse Autoencoders,Luke Marks,"Large language models (LLMs) aligned to human preferences via reinforcement learning from human feedback (RLHF) underpin many commercial applications. However, how RLHF impacts LLM internals remains opaque. We propose a novel method to interpret learned reward functions in RLHF-tuned LLMs using sparse autoencoders. Our approach trains autoencoder sets on activations from a base LLM and its RLHF-tuned version. By comparing autoencoder hidden spaces, we identify unique features that reflect the accuracy of the learned reward model. To quantify this, we construct a scenario where the tuned LLM learns token-reward mappings to maximize reward. This is the first application of sparse autoencoders for interpreting learned rewards and broadly inspecting reward learning in LLMs. Our method provides an abstract approximation of reward integrity. This presents a promising technique for ensuring alignment between specified objectives and model behaviors.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method to interpret learned reward functions in RLHF-tuned LLMs using sparse autoencoders and provides an abstract approximation of reward integrity.'}",https://arxiv.org/pdf/2310.08164
-Using Natural Language Explanations to Rescale Human Judgments,Manya Wadhwa,"The rise of large language models (LLMs) has brought a critical need for high-quality human-labeled data, particularly for processes like human feedback and evaluation. A common practice is to label data via consensus annotation over the judgments of multiple crowdworkers. However, different annotators may have different interpretations of labeling schemes unless given extensive training, and for subjective NLP tasks, even trained expert annotators can diverge heavily. We show that these nuances can be captured by high quality natural language explanations, and propose a method to rescale ordinal annotation in the presence of disagreement using LLMs. Specifically, we feed Likert ratings and corresponding natural language explanations into an LLM and prompt it to produce a numeric score. This score should reflect the underlying assessment of the example by the annotator. The presence of explanations allows the LLM to homogenize ratings across annotators in spite of scale usage differences. We explore our technique in the context of a document-grounded question answering task on which large language models achieve near-human performance. Among questions where annotators identify incompleteness in the answers, our rescaling improves correlation between nearly all annotator pairs, improving pairwise correlation on these examples by an average of 0.2 Kendall's tau.","{'model': 'tldr@v2.0.0', 'text': 'Rescaling improves correlation between nearly all annotator pairs, and allows the LLM to homogenize ratings across annotators in spite of scale usage differences, in the context of a document-grounded question answering task on which large language models achieve near-human performance.'}",http://arxiv.org/pdf/2305.14770
-Multi-scale discriminative regions analysis in FDG-PET imaging for early diagnosis of Alzheimer’s disease,Jin Zhang,"Objective. Alzheimer’s disease (AD) is a degenerative brain disorder, one of the main causes of death in elderly people, so early diagnosis of AD is vital to prompt access to medication and medical care. Fluorodeoxyglucose positron emission tomography (FDG-PET) proves to be effective to help understand neurological changes via measuring glucose uptake. Our aim is to explore information-rich regions of FDG-PET imaging, which enhance the accuracy and interpretability of AD-related diagnosis. Approach. We develop a novel method for early diagnosis of AD based on multi-scale discriminative regions in FDG-PET imaging, which considers the diagnosis interpretability. Specifically, a multi-scale region localization module is discussed to automatically identify disease-related discriminative regions in full-volume FDG-PET images in an unsupervised manner, upon which a confidence score is designed to evaluate the prioritization of regions according to the density distribution of anomalies. Then, the proposed multi-scale region classification module adaptively fuses multi-scale region representations and makes decision fusion, which not only reduces useless information but also offers complementary information. Most of previous methods concentrate on discriminating AD from cognitively normal (CN), while mild cognitive impairment, a transitional state, facilitates early diagnosis. Therefore, our method is further applied to multiple AD-related diagnosis tasks, not limited to AD vs. CN. Main results. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative dataset show that the proposed method achieves superior performance over state-of-the-art FDG-PET-based approaches. Besides, some cerebral cortices highlighted by extracted regions cohere with medical research, further demonstrating the superiority. Significance. This work offers an effective method to achieve AD diagnosis and detect disease-affected regions in FDG-PET imaging. Our results could be beneficial for providing an additional opinion on the clinical diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'A novel method for early diagnosis of AD based on multi-scale discriminative regions in FDG-PET imaging, which considers the diagnosis interpretability, and is applied to multiple AD-related diagnosis tasks, not limited to AD vs. CN.'}",
-Chain-of-Thought Prompt Distillation for Multimodal Named Entity Recognition and Multimodal Relation Extraction,F. Chen,"Multimodal Named Entity Recognition (MNER) and Multimodal Relation Extraction (MRE) necessitate the fundamental reasoning capacity for intricate linguistic and multimodal comprehension. In this study, we explore distilling the reasoning ability of large language models (LLMs) into a more compact student model by generating a \textit{chain of thought} (CoT) -- a sequence of intermediate reasoning steps. Specifically, we commence by exemplifying the elicitation of such reasoning ability from LLMs through CoT prompts covering multi-grain (noun, sentence, multimodality) and data-augmentation (style, entity, image) dimensions. Subsequently, we present a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge. Extensive experiments reveal that our approach attains state-of-the-art accuracy and manifests a plethora of advantages concerning interpretability, data efficiency, and cross-domain generalization on MNER and MRE datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge.'}",
-LitSumm: Large language models for literature summarisation of non-coding RNAs,Andrew Green,"Motivation: Curation of literature in life sciences is a growing challenge. The continued increase in the rate of publication, coupled with the relatively fixed number of curators worldwide presents a major challenge to developers of biomedical knowledgebases. Very few knowledgebases have resources to scale to the whole relevant literature and all have to prioritise their efforts. Results: In this work, we take a first step to alleviating the lack of curator time in RNA science by generating summaries of literature for non-coding RNAs using large language models (LLMs). We demonstrate that high-quality, factually accurate summaries with accurate references can be automatically generated from the literature using a commercial LLM and a chain of prompts and checks. Manual assessment was carried out for a subset of summaries, with the majority being rated extremely high quality. We also applied the most commonly used automated evaluation approaches, finding that they do not correlate with human assessment. Finally, we apply our tool to a selection of over 4,600 ncRNAs and make the generated summaries available via the RNAcentral resource. We conclude that automated literature summarization is feasible with the current generation of LLMs, provided careful prompting and automated checking are applied. Availability: Code used to produce these summaries can be found here: https://github.com/RNAcentral/litscan-summarization and the dataset of contexts and summaries can be found here: https://huggingface.co/datasets/RNAcentral/litsumm-v1. Summaries are also displayed on the RNA report pages in RNAcentral (https://rnacentral.org/)","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that high-quality, factually accurate summaries with accurate references can be automatically generated from the literature using a commercial LLM and a chain of prompts and checks, concluding that automated literature summarization is feasible with the current generation of LLMs.'}",
-Who Wrote it and Why? Prompting Large-Language Models for Authorship Verification,Chia-Yu Hung,"Authorship verification (AV) is a fundamental task in natural language processing (NLP) and computational linguistics, with applications in forensic analysis, plagiarism detection, and identification of deceptive content. Existing AV techniques, including traditional stylometric and deep learning approaches, face limitations in terms of data requirements and lack of explainability. To address these limitations, this paper proposes PromptAV, a novel technique that leverages Large-Language Models (LLMs) for AV by providing step-by-step stylometric explanation prompts. PromptAV outperforms state-of-the-art baselines, operates effectively with limited training data, and enhances interpretability through intuitive explanations, showcasing its potential as an effective and interpretable solution for the AV task.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2310.08123
-AllTogether: Investigating the Efficacy of Spliced Prompt for Web Navigation using Large Language Models,Jiarun Liu,"Large Language Models (LLMs) have emerged as promising agents for web navigation tasks, interpreting objectives and interacting with web pages. However, the efficiency of spliced prompts for such tasks remains underexplored. We introduces AllTogether, a standardized prompt template that enhances task context representation, thereby improving LLMs' performance in HTML-based web navigation. We evaluate the efficacy of this approach through prompt learning and instruction finetuning based on open-source Llama-2 and API-accessible GPT models. Our results reveal that models like GPT-4 outperform smaller models in web navigation tasks. Additionally, we find that the length of HTML snippet and history trajectory significantly influence performance, and prior step-by-step instructions prove less effective than real-time environmental feedback. Overall, we believe our work provides valuable insights for future research in LLM-driven web agents.","{'model': 'tldr@v2.0.0', 'text': ""AllTogether, a standardized prompt template that enhances task context representation, thereby improving LLMs' performance in HTML-based web navigation, is introduced, revealing that models like GPT-4 outperform smaller models in web navigation tasks.""}",
-WordArt Designer: User-Driven Artistic Typography Synthesis using Large Language Models,Jun-Yan He,"This paper introduces""WordArt Designer"", a user-driven framework for artistic typography synthesis, relying on Large Language Models (LLM). The system incorporates four key modules: the""LLM Engine"",""SemTypo"",""StyTypo"", and""TexTypo""modules. 1) The""LLM Engine"", empowered by LLM (e.g., GPT-3.5-turbo), interprets user inputs and generates actionable prompts for the other modules, thereby transforming abstract concepts into tangible designs. 2) The""SemTypo module""optimizes font designs using semantic concepts, striking a balance between artistic transformation and readability. 3) Building on the semantic layout provided by the""SemTypo module"", the""StyTypo module""creates smooth, refined images. 4) The""TexTypo module""further enhances the design's aesthetics through texture rendering, enabling the generation of inventive textured fonts. Notably,""WordArt Designer""highlights the fusion of generative AI with artistic typography. Experience its capabilities on ModelScope: https://www.modelscope.cn/studios/WordArt/WordArt.",,
-Context-Aware Prompt Tuning for Vision-Language Model with Dual-Alignment,Hongyu Hu,"Large-scale vision-language models (VLMs), e.g., CLIP, learn broad visual concepts from tedious training data, showing superb generalization ability. Amount of prompt learning methods have been proposed to efficiently adapt the VLMs to downstream tasks with only a few training samples. We introduce a novel method to improve the prompt learning of vision-language models by incorporating pre-trained large language models (LLMs), called Dual-Aligned Prompt Tuning (DuAl-PT). Learnable prompts, like CoOp, implicitly model the context through end-to-end training, which are difficult to control and interpret. While explicit context descriptions generated by LLMs, like GPT-3, can be directly used for zero-shot classification, such prompts are overly relying on LLMs and still underexplored in few-shot domains. With DuAl-PT, we propose to learn more context-aware prompts, benefiting from both explicit and implicit context modeling. To achieve this, we introduce a pre-trained LLM to generate context descriptions, and we encourage the prompts to learn from the LLM's knowledge by alignment, as well as the alignment between prompts and local image features. Empirically, DuAl-PT achieves superior performance on 11 downstream datasets on few-shot recognition and base-to-new generalization. Hopefully, DuAl-PT can serve as a strong baseline. Code will be available.","{'model': 'tldr@v2.0.0', 'text': 'A novel method to improve the prompt learning of vision-language models by incorporating pre-trained large language models (LLMs), called Dual-Aligned Prompt Tuning (DuAl-PT), which achieves superior performance on 11 downstream datasets on few-shot recognition and base-to-new generalization.'}",https://arxiv.org/pdf/2309.04158
-GPTutor: an open-source AI pair programming tool alternative to Copilot,Eason Chen,"This paper presents the latest progress of GPTutor: a ChatGPT-powered programming tool extension in Visual Studio Code. The emergence of Large Language Models (LLMs) has improved software development efficiency, but their performance can be hindered by training data limitations and prompt design issues. Existing LLM development tools often operate as black boxes, with users unable to view the prompts used and unable to improve performance by correcting prompts when errors occur. To address the aforementioned issues, GPTutor was introduced as an open-source AI pair programming tool, offering an alternative to Copilot. GPTutor empowers users to customize prompts for various programming languages and scenarios, with support for 120+ human languages and 50+ programming languages. Users can fine-tune prompts to correct the errors from LLM for precision and efficient code generation. At the end of the paper, we underscore GPTutor's potential through examples, including demonstrating its proficiency in interpreting and generating Sui-Move, a newly introduced smart contract language, using prompt engineering.","{'model': 'tldr@v2.0.0', 'text': 'GPTutor is presented as an open-source AI pair programming tool, offering an alternative to Copilot, and its potential is underscore through examples, including demonstrating its proficiency in interpreting and generating Sui-Move, a newly introduced smart contract language, using prompt engineering.'}",
-FairMonitor: A Four-Stage Automatic Framework for Detecting Stereotypes and Biases in Large Language Models,Yanhong Bai,"Detecting stereotypes and biases in Large Language Models (LLMs) can enhance fairness and reduce adverse impacts on individuals or groups when these LLMs are applied. However, the majority of existing methods focus on measuring the model's preference towards sentences containing biases and stereotypes within datasets, which lacks interpretability and cannot detect implicit biases and stereotypes in the real world. To address this gap, this paper introduces a four-stage framework to directly evaluate stereotypes and biases in the generated content of LLMs, including direct inquiry testing, serial or adapted story testing, implicit association testing, and unknown situation testing. Additionally, the paper proposes multi-dimensional evaluation metrics and explainable zero-shot prompts for automated evaluation. Using the education sector as a case study, we constructed the Edu-FairMonitor based on the four-stage framework, which encompasses 12,632 open-ended questions covering nine sensitive factors and 26 educational scenarios. Experimental results reveal varying degrees of stereotypes and biases in five LLMs evaluated on Edu-FairMonitor. Moreover, the results of our proposed automated evaluation method have shown a high correlation with human annotations.","{'model': 'tldr@v2.0.0', 'text': 'A four-stage framework to directly evaluate stereotypes and biases in the generated content of LLMs, including direct inquiry testing, serial or adapted story testing, implicit association testing, and unknown situation testing is introduced.'}",
-De-Diffusion Makes Text a Strong Cross-Modal Interface,Chen Wei,"We demonstrate text as a strong cross-modal interface. Rather than relying on deep embeddings to connect image and language as the interface representation, our approach represents an image as text, from which we enjoy the interpretability and flexibility inherent to natural language. We employ an autoencoder that uses a pre-trained text-to-image diffusion model for decoding. The encoder is trained to transform an input image into text, which is then fed into the fixed text-to-image diffusion decoder to reconstruct the original input -- a process we term De-Diffusion. Experiments validate both the precision and comprehensiveness of De-Diffusion text representing images, such that it can be readily ingested by off-the-shelf text-to-image tools and LLMs for diverse multi-modal tasks. For example, a single De-Diffusion model can generalize to provide transferable prompts for different text-to-image tools, and also achieves a new state of the art on open-ended vision-language tasks by simply prompting large language models with few-shot examples.","{'model': 'tldr@v2.0.0', 'text': 'De-Diffusion text representing images can be readily ingested by off-the-shelf text-to-image tools and LLMs for diverse multi-modal tasks and achieves a new state of the art on open-ended vision-language tasks by simply prompting large language models with few-shot examples.'}",
-Artificial Intelligence and Implications for the Australian Social Work Journal,David Hodgson,"Social work is a profession committed to integrity and social justice. The AASW Social Work Practice Standards (AASW, 2023) calls on social workers to be critically reflective, ethical practitioners engaged in lifelong professional development and learning. Equally, social work education seeks to prepare students for research-informed, culturally-responsive practice across a diverse range of contexts, and in this Issue, we showcase critical social work education and practice diversity. However, a different ethical challenge to integrity and practice standards is the focus of this Editorial. Here, we highlight some of the concerns and implications of generative Artificial Intelligence (generative AI) for social work education, research, practice, and scholarly publishing. In November 2022, OpenAI released ChatGPT, a generative AI Large Language Model (LLM) that could generate realistic and natural text outputs from simple prompts. This technology had been in development for some time but had not been released to the public for general use. Since then, there has been a proliferation of different AI models that can generate and augment text, images, video, and audio. Generative AI is being used to perform analytical and interpretive tasks such as language translation; responding to queries on specific data sources, coding, and interpreting code; summarising documents and webpages; and creating case assessments and plans. This technology can be used to construct legal documents; machine learning for facial recognition; and for undertaking medical, mental health, and other diagnostic assessments. These are just some examples. In this fast-moving field, the uses and applications seem endless. The open-sourcing of generative AI models and their underlying architecture means developers are starting to create a myriad of practical applications and tools that rapidly increase the depth and scale of automation, potentially replacing or augmenting many everyday tasks normally performed by humans. The implications for social work education, practice, research, and scholarship are extensive. As with any new technology, there are a range of stances, from early adopters to positions that have resonance with luddism. This adds to the complexities of responding to AI as a whole profession. Nevertheless, what is clear is that the rise and integration of generative AI systems, at scale, will yield a wide range of practical, ethical, and epistemological problems for many professions, including social work. It is to some of these problems we turn our attention below. Beginning with social work education, generative AI will have profound effects on assessment and learning for higher education providers. It is likely to cause educators to re-evaluate their educational practices, assessments, and assumptions about what is core to a social work curriculum. Social work will need to refine and reappraise its ideas about critical thinking, ethical decision making, professional judgement, and reflective practice—all skills that are considered core to effective social work practice as outlined in the AASW Practice Standards (AASW, 2023). How will we ensure students have an educational environment that promotes","{'model': 'tldr@v2.0.0', 'text': 'Some of the concerns and implications of generative Artificial Intelligence (generative AI) for social work education, research, practice, and scholarly publishing are highlighted.'}",https://www.tandfonline.com/doi/pdf/10.1080/0312407X.2023.2247833?needAccess=true
-Automatic Prompt Rewriting for Personalized Text Generation,Cheng Li,"Facilitated by large language models (LLMs), personalized text generation has become a rapidly growing research direction. Most existing studies focus on designing specialized models for a particular domain, or they require fine-tuning the LLMs to generate personalized text. We consider a typical scenario in which the large language model, which generates personalized output, is frozen and can only be accessed through APIs. Under this constraint, all one can do is to improve the input text (i.e., text prompts) sent to the LLM, a procedure that is usually done manually. In this paper, we propose a novel method to automatically revise prompts for personalized text generation. The proposed method takes the initial prompts generated by a state-of-the-art, multistage framework for personalized generation and rewrites a few critical components that summarize and synthesize the personal context. The prompt rewriter employs a training paradigm that chains together supervised learning (SL) and reinforcement learning (RL), where SL reduces the search space of RL and RL facilitates end-to-end training of the rewriter. Using datasets from three representative domains, we demonstrate that the rewritten prompts outperform both the original prompts and the prompts optimized via supervised learning or reinforcement learning alone. In-depth analysis of the rewritten prompts shows that they are not only human readable, but also able to guide manual revision of prompts when there is limited resource to employ reinforcement learning to train the prompt rewriter, or when it is costly to deploy an automatic prompt rewriter for inference.","{'model': 'tldr@v2.0.0', 'text': 'In-depth analysis of the rewritten prompts shows that they are not only human readable, but also able to guide manual revision of prompts when there is limited resource to employ reinforcement learning to train the prompt rewriter, or when it is costly to deploy an automatic prompt rewriters for inference.'}",https://arxiv.org/pdf/2310.00152
-Structural Embeddings of Tools for Large Language Models,Eren Unlu,"It is evident that the current state of Large Language Models (LLMs) necessitates the incorporation of external tools. The lack of straightforward algebraic and logical reasoning is well documented and prompted researchers to develop frameworks which allow LLMs to operate via external tools. The ontological nature of tool utilization for a specific task can be well formulated with a Directed Acyclic Graph (DAG). The central aim of the paper is to highlight the importance of graph based approaches to LLM-tool interaction in near future. We propose an exemplary framework to guide the orchestration of exponentially increasing numbers of external tools with LLMs,where objectives and functionalities of tools are graph encoded hierarchically. Assuming that textual segments of a Chain-of-Thought (CoT) can be imagined as a tool as defined here, the graph based framework can pave new avenues in that particular direction as well.","{'model': 'tldr@v2.0.0', 'text': 'An exemplary framework to guide the orchestration of exponentially increasing numbers of external tools with LLMs is proposed, where objectives and functionalities of tools are graph encoded hierarchically.'}",https://arxiv.org/pdf/2308.00447
-LlamaRec: Two-Stage Recommendation using Large Language Models for Ranking,Zhenrui Yue,"Recently, large language models (LLMs) have exhibited significant progress in language understanding and generation. By leveraging textual features, customized LLMs are also applied for recommendation and demonstrate improvements across diverse recommendation scenarios. Yet the majority of existing methods perform training-free recommendation that heavily relies on pretrained knowledge (e.g., movie recommendation). In addition, inference on LLMs is slow due to autoregressive generation, rendering existing methods less effective for real-time recommendation. As such, we propose a two-stage framework using large language models for ranking-based recommendation (LlamaRec). In particular, we use small-scale sequential recommenders to retrieve candidates based on the user interaction history. Then, both history and retrieved items are fed to the LLM in text via a carefully designed prompt template. Instead of generating next-item titles, we adopt a verbalizer-based approach that transforms output logits into probability distributions over the candidate items. Therefore, the proposed LlamaRec can efficiently rank items without generating long text. To validate the effectiveness of the proposed framework, we compare against state-of-the-art baseline methods on benchmark datasets. Our experimental results demonstrate the performance of LlamaRec, which consistently achieves superior performance in both recommendation performance and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-stage framework using large language models for ranking-based recommendation (LlamaRec) that uses small-scale sequential recommenders to retrieve candidates based on the user interaction history and adopts a verbalizer-based approach that transforms output logits into probability distributions over the candidate items.'}",
-Evaluating Diverse Large Language Models for Automatic and General Bug Reproduction,Sungmin Kang,"Bug reproduction is a critical developer activity that is also challenging to automate, as bug reports are often in natural language and thus can be difficult to transform to test cases consistently. As a result, existing techniques mostly focused on crash bugs, which are easier to automatically detect and verify. In this work, we overcome this limitation by using large language models (LLMs), which have been demonstrated to be adept at natural language processing and code generation. By prompting LLMs to generate bug-reproducing tests, and via a post-processing pipeline to automatically identify promising generated tests, our proposed technique LIBRO could successfully reproduce about one-third of all bugs in the widely used Defects4J benchmark. Furthermore, our extensive evaluation on 15 LLMs, including 11 open-source LLMs, suggests that open-source LLMs also demonstrate substantial potential, with the StarCoder LLM achieving 70% of the reproduction performance of the closed-source OpenAI LLM code-davinci-002 on the large Defects4J benchmark, and 90% of performance on a held-out bug dataset likely not part of any LLM's training data. In addition, our experiments on LLMs of different sizes show that bug reproduction using LIBRO improves as LLM size increases, providing information as to which LLMs can be used with the LIBRO pipeline.","{'model': 'tldr@v2.0.0', 'text': 'The proposed technique LIBRO could successfully reproduce about one-third of all bugs in the widely used Defects4J benchmark, and experiments on LLMs of different sizes show that bug reproduction using LIBRO improves as LLM size increases, providing information as to which LLMs can be used with the LIBRO pipeline.'}",
-Text-based predictions of COVID-19 diagnosis from self-reported chemosensory descriptions,Hongyang Li,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that the description of perceptual symptoms caused by a viral infection can be used to fine-tune an LLM model to correctly predict and interpret the diagnostic status of a subject.'}",https://www.nature.com/articles/s43856-023-00334-5.pdf
-SCPatcher: Mining Crowd Security Discussions to Enrich Secure Coding Practices,Ziyou Jiang,"Secure coding practices (SCPs) have been proposed to guide software developers to write code securely to prevent potential security vulnerabilities. Yet, they are typically one-sentence principles without detailed specifications, e.g., “Properly free allocated memory upon the completion of functions and at all exit points.”, which makes them difficult to follow in practice, especially for software developers who are not yet experienced in secure programming. To address this problem, this paper proposes SCPatcher, an automated approach to enrich secure coding practices by mining crowd security discussions on online knowledge-sharing platforms, such as Stack Overflow. In particular, for each security post, SCPatcher first extracts the area of coding examples and coding explanations with a fix-prompt tuned Large Language Model (LLM) via Prompt Learning. Then, it hierarchically slices the lengthy code into coding examples and summarizes the coding explanations with the areas. Finally, SCPatcher matches the CWE and Public SCP, integrating them with extracted coding examples and explanations to form the SCP specifications, which are the wild SCPs with details, proposed by the developers. To evaluate the performance of SCPatcher, we conduct experiments on 3,907 security posts from Stack Overflow. The experimental results show that SCPatcher outperforms all baselines in extracting the coding examples with 2.73 % MLine on average, as well as coding explanations with 3.97 % F1 on average. Moreover, we apply SCPatcher on 447 new security posts to further evaluate its practicality, and the extracted SCP specifications enrich the public SCPs with 3,074 lines of code and 1,967 sentences.","{'model': 'tldr@v2.0.0', 'text': 'SCPatcher is proposed, an automated approach to enrich secure coding practices by mining crowd security discussions on online knowledge-sharing platforms, such as Stack Overflow, with outperforms all baselines in extracting coding examples and coding explanations.'}",
-Mind the instructions: a holistic evaluation of consistency and interactions in prompt-based learning,Lucas Weber,"Finding the best way of adapting pre-trained language models to a task is a big challenge in current NLP. Just like the previous generation of task-tuned models (TT), models that are adapted to tasks via in-context-learning (ICL) are robust in some setups but not in others. Here, we present a detailed analysis of which design choices cause instabilities and inconsistencies in LLM predictions. First, we show how spurious correlations between input distributions and labels -- a known issue in TT models -- form only a minor problem for prompted models. Then, we engage in a systematic, holistic evaluation of different factors that have been found to influence predictions in a prompting setup. We test all possible combinations of a range of factors on both vanilla and instruction-tuned (IT) LLMs of different scale and statistically analyse the results to show which factors are the most influential, interactive or stable. Our results show which factors can be used without precautions and which should be avoided or handled with care in most settings.","{'model': 'tldr@v2.0.0', 'text': 'A detailed analysis of which design choices cause instabilities and inconsistencies in LLM predictions is presented and which factors can be used without precautions and which should be avoided or handled with care in most settings.'}",
-BotChat: Evaluating LLMs' Capabilities of Having Multi-Turn Dialogues,Haodong Duan,"Interacting with human via high-quality multi-turn dialogues is a key feature of large language models (LLMs). However, human-based evaluation of such capability involves intensive manual labor. This report provides a preliminary evaluation of existing large language models for human-style multi-turn chatting, through an LLM-based approach. We start from real-world human dialogues and keep the very first utterances as the ChatSEED. Then we prompt LLMs to generate a full multi-turn dialogue (tens of utterances) based on the ChatSEED, utterance by utterance. Finally, we adopt state-of-the-art LLMs (GPT-4, \etc) as the judge to evaluate the generated dialogues. With different evaluation protocols, we come to substantially identical conclusions. We find that GPT-4 can generate human-style multi-turn dialogues with impressive quality, significantly outperforms its counterparts. It's difficult for a discriminator to distinguish between GPT-4 generated dialogues and human dialogues. In contrast, other LLMs struggle to generate multi-turn dialogues of satisfactory quality due to poor instruction-following capability, tendency to generate lengthy utterances, or limited general capability. All data and codes will be provided in https://github.com/open-compass/BotChat/ and we hope they can serve as a valuable resource for evaluating multi-turn chatting capabilities of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is found that GPT-4 can generate human-style multi-turn dialogues with impressive quality, significantly outperforms its counterparts and is difficult for a discriminator to distinguish between G PT-4 generated dialogues and human dialogues.'}",
-Collaborative Large Language Model for Recommender Systems,Yaochen Zhu,"Recently, there is a growing interest in developing next-generation recommender systems (RSs) based on pretrained large language models (LLMs), fully utilizing their encoded knowledge and reasoning ability. However, the semantic gap between natural language and recommendation tasks is still not well addressed, leading to multiple issues such as spuriously-correlated user/item descriptors, ineffective language modeling on user/item contents, and inefficient recommendations via auto-regression, etc. In this paper, we propose CLLM4Rec, the first generative RS that tightly integrates the LLM paradigm and ID paradigm of RS, aiming to address the above challenges simultaneously. We first extend the vocabulary of pretrained LLMs with user/item ID tokens to faithfully model the user/item collaborative and content semantics. Accordingly, in the pretraining stage, a novel soft+hard prompting strategy is proposed to effectively learn user/item collaborative/content token embeddings via language modeling on RS-specific corpora established from user-item interactions and user/item features, where each document is split into a prompt consisting of heterogeneous soft (user/item) tokens and hard (vocab) tokens and a main text consisting of homogeneous item tokens or vocab tokens that facilitates stable and effective language modeling. In addition, a novel mutual regularization strategy is introduced to encourage the CLLM4Rec to capture recommendation-oriented information from user/item contents. Finally, we propose a novel recommendation-oriented finetuning strategy for CLLM4Rec, where an item prediction head with multinomial likelihood is added to the pretrained CLLM4Rec backbone to predict hold-out items based on the soft+hard prompts established from masked user-item interaction history, where recommendations of multiple items can be generated efficiently.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes CLLM4Rec, the first generative RS that tightly integrates the LLM paradigm and ID paradigm of RS, aiming to address the above challenges simultaneously, and proposes a novel recommendation-oriented finetuning strategy for CLLm4Rec.'}",
-Optimal Scene Graph Planning with Large Language Model Guidance,Zhirui Dai,"Recent advances in metric, semantic, and topological mapping have equipped autonomous robots with semantic concept grounding capabilities to interpret natural language tasks. This work aims to leverage these new capabilities with an efficient task planning algorithm for hierarchical metric-semantic models. We consider a scene graph representation of the environment and utilize a large language model (LLM) to convert a natural language task into a linear temporal logic (LTL) automaton. Our main contribution is to enable optimal hierarchical LTL planning with LLM guidance over scene graphs. To achieve efficiency, we construct a hierarchical planning domain that captures the attributes and connectivity of the scene graph and the task automaton, and provide semantic guidance via an LLM heuristic function. To guarantee optimality, we design an LTL heuristic function that is provably consistent and supplements the potentially inadmissible LLM guidance in multi-heuristic planning. We demonstrate efficient planning of complex natural language tasks in scene graphs of virtualized real environments.","{'model': 'tldr@v2.0.0', 'text': 'This work considers a scene graph representation of the environment and utilizes a large language model (LLM) to convert a natural language task into a linear temporal logic (LTL) automaton, and designs an LTL heuristic function that is provably consistent and supplements the potentially inadmissible LLM guidance in multi-heuristic planning.'}",https://arxiv.org/pdf/2309.09182
-Can Large Language Models be Good Path Planners? A Benchmark and Investigation on Spatial-temporal Reasoning,Mohamed Aghzal,"Large language models (LLMs) have achieved remarkable success across a wide spectrum of tasks; however, they still face limitations in scenarios that demand long-term planning and spatial reasoning. To facilitate this line of research, in this work, we propose a new benchmark, termed $\textbf{P}$ath $\textbf{P}$lanning from $\textbf{N}$atural $\textbf{L}$anguage ($\textbf{PPNL}$). Our benchmark evaluates LLMs' spatial-temporal reasoning by formulating ''path planning'' tasks that require an LLM to navigate to target locations while avoiding obstacles and adhering to constraints. Leveraging this benchmark, we systematically investigate LLMs including GPT-4 via different few-shot prompting methodologies and BART and T5 of various sizes via fine-tuning. Our experimental results show the promise of few-shot GPT-4 in spatial reasoning, when it is prompted to reason and act interleavedly, although it still fails to make long-term temporal reasoning. In contrast, while fine-tuned LLMs achieved impressive results on in-distribution reasoning tasks, they struggled to generalize to larger environments or environments with more obstacles.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show the promise of few-shot GPT-4 in spatial reasoning, when it is prompted to reason and act interleavedly, although it still fails to make long-term temporal reasoning.'}",https://arxiv.org/pdf/2310.03249
-Automated Assessment of Comprehension Strategies from Self-Explanations Using LLMs,Bogdan-Ioan Nicula,"Text comprehension is an essential skill in today’s information-rich world, and self-explanation practice helps students improve their understanding of complex texts. This study was centered on leveraging open-source Large Language Models (LLMs), specifically FLAN-T5, to automatically assess the comprehension strategies employed by readers while understanding Science, Technology, Engineering, and Mathematics (STEM) texts. The experiments relied on a corpus of three datasets (N = 11,833) with self-explanations annotated on 4 dimensions: 3 comprehension strategies (i.e., bridging, elaboration, and paraphrasing) and overall quality. Besides FLAN-T5, we also considered GPT3.5-turbo to establish a stronger baseline. Our experiments indicated that the performance improved with fine-tuning, having a larger LLM model, and providing examples via the prompt. Our best model considered a pretrained FLAN-T5 XXL model and obtained a weighted F1-score of 0.721, surpassing the 0.699 F1-score previously obtained using smaller models (i.e., RoBERTa).",,https://www.mdpi.com/2078-2489/14/10/567/pdf?version=1697284775
-LINC: A Neurosymbolic Approach for Logical Reasoning by Combining Language Models with First-Order Logic Provers,Theo X. Olausson,"Logical reasoning, i.e., deductively inferring the truth value of a conclusion from a set of premises, is an important task for artificial intelligence with wide potential impacts on science, mathematics, and society. While many prompting-based strategies have been proposed to enable Large Language Models (LLMs) to do such reasoning more effectively, they still appear unsatisfactory, often failing in subtle and unpredictable ways. In this work, we investigate the validity of instead reformulating such tasks as modular neurosymbolic programming, which we call LINC: Logical Inference via Neurosymbolic Computation. In LINC, the LLM acts as a semantic parser, translating premises and conclusions from natural language to expressions in first-order logic. These expressions are then offloaded to an external theorem prover, which symbolically performs deductive inference. Leveraging this approach, we observe significant performance gains on FOLIO and a balanced subset of ProofWriter for three different models in nearly all experimental conditions we evaluate. On ProofWriter, augmenting the comparatively small open-source StarCoder+ (15.5B parameters) with LINC even outperforms GPT-3.5 and GPT-4 with Chain-of-Thought (CoT) prompting by an absolute 38% and 10%, respectively. When used with GPT-4, LINC scores 26% higher than CoT on ProofWriter while performing comparatively on FOLIO. Further analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes. We thus provide promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers. All corresponding code is publicly available at https://github.com/benlipkin/linc","{'model': 'tldr@v2.0.0', 'text': 'Analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes, which provides promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers.'}",
-Amortizing intractable inference in large language models,Edward J. Hu,"Autoregressive large language models (LLMs) compress knowledge from their training data through next-token conditional distributions. This limits tractable querying of this knowledge to start-to-end autoregressive sampling. However, many tasks of interest -- including sequence continuation, infilling, and other forms of constrained generation -- involve sampling from intractable posterior distributions. We address this limitation by using amortized Bayesian inference to sample from these intractable posteriors. Such amortization is algorithmically achieved by fine-tuning LLMs via diversity-seeking reinforcement learning algorithms: generative flow networks (GFlowNets). We empirically demonstrate that this distribution-matching paradigm of LLM fine-tuning can serve as an effective alternative to maximum-likelihood training and reward-maximizing policy optimization. As an important application, we interpret chain-of-thought reasoning as a latent variable modeling problem and demonstrate that our approach enables data-efficient adaptation of LLMs to tasks that require multi-step rationalization and tool use.","{'model': 'tldr@v2.0.0', 'text': 'This work interprets chain-of-thought reasoning as a latent variable modeling problem and demonstrates that this distribution-matching paradigm of LLM fine-tuning can serve as an effective alternative to maximum-likelihood training and reward-maximizing policy optimization.'}",
-Fabricator: An Open Source Toolkit for Generating Labeled Training Data with Teacher LLMs,Jonas Golde,"Most NLP tasks are modeled as supervised learning and thus require labeled training data to train effective models. However, manually producing such data at sufficient quality and quantity is known to be costly and time-intensive. Current research addresses this bottleneck by exploring a novel paradigm called zero-shot learning via dataset generation. Here, a powerful LLM is prompted with a task description to generate labeled data that can be used to train a downstream NLP model. For instance, an LLM might be prompted to""generate 500 movie reviews with positive overall sentiment, and another 500 with negative sentiment.""The generated data could then be used to train a binary sentiment classifier, effectively leveraging an LLM as a teacher to a smaller student model. With this demo, we introduce Fabricator, an open-source Python toolkit for dataset generation. Fabricator implements common dataset generation workflows, supports a wide range of downstream NLP tasks (such as text classification, question answering, and entity recognition), and is integrated with well-known libraries to facilitate quick experimentation. With Fabricator, we aim to support researchers in conducting reproducible dataset generation experiments using LLMs and help practitioners apply this approach to train models for downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'With Fabricator, an open-source Python toolkit for dataset generation, this work aims to support researchers in conducting reproducible dataset generation experiments using LLMs and help practitioners apply this approach to train models for downstream tasks.'}",https://arxiv.org/pdf/2309.09582
-BooookScore: A systematic exploration of book-length summarization in the era of LLMs,Yapei Chang,"Summarizing book-length documents (>100K tokens) that exceed the context window size of large language models (LLMs) requires first breaking the input document into smaller chunks and then prompting an LLM to merge, update, and compress chunk-level summaries. Despite the complexity and importance of this task, it has yet to be meaningfully studied due to the challenges of evaluation: existing book-length summarization datasets (e.g., BookSum) are in the pretraining data of most public LLMs, and existing evaluation methods struggle to capture errors made by modern LLM summarizers. In this paper, we present the first study of the coherence of LLM-based book-length summarizers implemented via two prompting workflows: (1) hierarchically merging chunk-level summaries, and (2) incrementally updating a running summary. We obtain 1193 fine-grained human annotations on GPT-4 generated summaries of 100 recently-published books and identify eight common types of coherence errors made by LLMs. Because human evaluation is expensive and time-consuming, we develop an automatic metric, BooookScore, that measures the proportion of sentences in a summary that do not contain any of the identified error types. BooookScore has high agreement with human annotations and allows us to systematically evaluate the impact of many other critical parameters (e.g., chunk size, base LLM) while saving $15K and 500 hours in human evaluation costs. We find that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher BooookScore than the oft-repetitive ones generated by LLaMA 2. Incremental updating yields lower BooookScore but higher level of detail than hierarchical merging, a trade-off sometimes preferred by human annotators. We release code and annotations after blind review to spur more principled research on book-length summarization.","{'model': 'tldr@v2.0.0', 'text': 'An automatic metric, BooookScore, is developed that measures the proportion of sentences in a summary that do not contain any of the identified error types and is found that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher Booook score than the oft-repetitive ones generated by LLaMA 2.'}",https://arxiv.org/pdf/2310.00785
-A Prefrontal Cortex-inspired Architecture for Planning in Large Language Models,Taylor Webb,"Large language models (LLMs) demonstrate impressive performance on a wide variety of tasks, but they often struggle with tasks that require multi-step reasoning or goal-directed planning. To address this, we take inspiration from the human brain, in which planning is accomplished via the recurrent interaction of specialized modules in the prefrontal cortex (PFC). These modules perform functions such as conflict monitoring, state prediction, state evaluation, task decomposition, and task coordination. We find that LLMs are sometimes capable of carrying out these functions in isolation, but struggle to autonomously coordinate them in the service of a goal. Therefore, we propose a black box architecture with multiple LLM-based (GPT-4) modules. The architecture improves planning through the interaction of specialized PFC-inspired modules that break down a larger problem into multiple brief automated calls to the LLM. We evaluate the combined architecture on two challenging planning tasks -- graph traversal and Tower of Hanoi -- finding that it yields significant improvements over standard LLM methods (e.g., zero-shot prompting or in-context learning). These results demonstrate the benefit of utilizing knowledge from cognitive neuroscience to improve planning in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A black box architecture with multiple LLM-based (GPT-4) modules that improves planning through the interaction of specialized PFC-inspired modules that break down a larger problem into multiple brief automated calls to the LLM.'}",https://arxiv.org/pdf/2310.00194
-Walking Down the Memory Maze: Beyond Context Limit through Interactive Reading,Howard Chen,"Large language models (LLMs) have advanced in large strides due to the effectiveness of the self-attention mechanism that processes and compares all tokens at once. However, this mechanism comes with a fundamental issue -- the predetermined context window is bound to be limited. Despite attempts to extend the context window through methods like extrapolating the positional embedding, using recurrence, or selectively retrieving essential parts of the long sequence, long-text understanding continues to be a challenge. We propose an alternative approach which instead treats the LLM as an interactive agent, allowing it to decide how to read the text via iterative prompting. We introduce MemWalker, a method that first processes the long context into a tree of summary nodes. Upon receiving a query, the model navigates this tree in search of relevant information, and responds once it gathers sufficient information. On long-text question answering tasks our method outperforms baseline approaches that use long context windows, recurrence, and retrieval. We show that, beyond effective reading, MemWalker enhances explainability by highlighting the reasoning steps as it interactively reads the text; pinpointing the relevant text segments related to the query.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces MemWalker, a method that first processes the long context into a tree of summary nodes, and upon receiving a query, the model navigates this tree in search of relevant information, and responds once it gathers sufficient information.'}",https://arxiv.org/pdf/2310.05029
-Scalable Multi-Robot Collaboration with Large Language Models: Centralized or Decentralized Systems?,Yongchao Chen,"A flurry of recent work has demonstrated that pre-trained large language models (LLMs) can be effective task planners for a variety of single-robot tasks. The planning performance of LLMs is significantly improved via prompting techniques, such as in-context learning or re-prompting with state feedback, placing new importance on the token budget for the context window. An under-explored but natural next direction is to investigate LLMs as multi-robot task planners. However, long-horizon, heterogeneous multi-robot planning introduces new challenges of coordination while also pushing up against the limits of context window length. It is therefore critical to find token-efficient LLM planning frameworks that are also able to reason about the complexities of multi-robot coordination. In this work, we compare the task success rate and token efficiency of four multi-agent communication frameworks (centralized, decentralized, and two hybrid) as applied to four coordination-dependent multi-agent 2D task scenarios for increasing numbers of agents. We find that a hybrid framework achieves better task success rates across all four tasks and scales better to more agents. We further demonstrate the hybrid frameworks in 3D simulations where the vision-to-text problem and dynamical errors are considered. See our project website https://yongchao98.github.io/MIT-REALM-Multi-Robot/ for prompts, videos, and code.","{'model': 'tldr@v2.0.0', 'text': 'This work compares the task success rate and token efficiency of four multi-agent communication frameworks as applied to four coordination-dependent multi- agent 2D task scenarios for increasing numbers of agents and finds that a hybrid framework achieves better task success rates across all four tasks and scales better to more agents.'}",https://arxiv.org/pdf/2309.15943
-Distilled Language Models are economically efficient for the enterprise. ...mostly.,Kristen Howell,"Contacting customer service via chat is a common practice. Because employing customer service agents is expensive, many companies are turning to NLP that assists human agents by auto-generating responses that can be used directly or with modifications. With their ability to handle large context windows, Large Language Models (LLMs) are a natural fit for this use case. However, their efficacy must be balanced with the cost of training and serving them. This paper assesses the practical cost and impact of LLMs for the enterprise as a function of the usefulness of the responses that they generate. We present a cost framework for evaluating an NLP model’s utility for this use case and apply it to a single brand as a case study in the context of an existing agent assistance product. We compare three strategies for specializing an LLM — prompt engineering, fine-tuning, and knowledge distillation — using feedback from the brand’s customer service agents. We find that the usability of a model’s responses can make up for a large difference in inference cost for our case study brand, and we extrapolate our findings to the broader enterprise space.",,http://arxiv.org/pdf/2306.07402
-Making Multimodal Generation Easier: When Diffusion Models Meet LLMs,Xiangyu Zhao,"We present EasyGen, an efficient model designed to enhance multimodal understanding and generation by harnessing the capabilities of diffusion models and large language models (LLMs). Unlike existing multimodal models that predominately depend on encoders like CLIP or ImageBind and need ample amounts of training data to bridge the gap between modalities, EasyGen is built upon a bidirectional conditional diffusion model named BiDiffuser, which promotes more efficient interactions between modalities. EasyGen handles image-to-text generation by integrating BiDiffuser and an LLM via a simple projection layer. Unlike most existing multimodal models that are limited to generating text responses, EasyGen can also facilitate text-to-image generation by leveraging the LLM to create textual descriptions, which can be interpreted by BiDiffuser to generate appropriate visual responses. Extensive quantitative and qualitative experiments demonstrate the effectiveness of EasyGen, whose training can be easily achieved in a lab setting. The source code is available at https://github.com/zxy556677/EasyGen.","{'model': 'tldr@v2.0.0', 'text': 'Efficient model designed to enhance multimodal understanding and generation by harnessing the capabilities of diffusion models and large language models, built upon a bidirectional conditional diffusion model named BiDiffuser, which promotes more efficient interactions between modalities.'}",https://arxiv.org/pdf/2310.08949
-Curriculum Prompt Learning with Self-Training for Abstractive Dialogue Summarization,Changqun Li,"Succinctly summarizing dialogue is a task of growing interest, but inherent challenges, such as insufficient training data and low information density impede our ability to train abstractive models. In this work, we propose a novel curriculum-based prompt learning method with self-training to address these problems. Specifically, prompts are learned using a curriculum learning strategy that gradually increases the degree of prompt perturbation, thereby improving the dialogue understanding and modeling capabilities of our model. Unlabeled dialogue is incorporated by means of self-training so as to reduce the dependency on labeled data. We further investigate topic-aware prompts to better plan for the generation of summaries. Experiments confirm that our model substantially outperforms strong baselines and achieves new state-of-the-art results on the AMI and ICSI datasets. Human evaluations also show the superiority of our model with regard to the summary generation quality.","{'model': 'tldr@v2.0.0', 'text': 'A novel curriculum-based prompt learning method with self-training that substantially outperforms strong baselines and achieves new state-of-the-art results on the AMI and ICSI datasets is proposed.'}",https://aclanthology.org/2022.emnlp-main.72.pdf
-Questioning Prompts to Examine Blended PK-20 Learning Environments,Neal Shambaugh,"The decision to deliver a course or an academic program must address issues of learning outcomes and curriculum, what is to be taught along with choices of pedagogy or how the content is taught, as well as face-to-face or online delivery. The question of whether or not online or blended instruction is more effective than face to face teaching is not about the delivery decision but about curriculum and pedagogical issues. A set of questioning prompts, organized by public school and undergraduate/graduate programs, help educators systematically think through issues of learner differences, learning outcomes, teaching options, including the use of technology and instructional delivery approaches. Best practices are organized by public school and undergraduate/graduate programs. Research questions are suggested.",,
-Practicing What We Preach: Incorporating Team-Based Learning into the Pre-Service Teaching Curriculum for Improved Outcomes,Jane S. Vogler,"Background: Prior research has shown that pre-service teachers’ learning of theory is enhanced when teacher educators model the instructional approaches that they advocate as effective. Objective: The purpose of this study was to compare undergraduate learning outcomes in a course using Team-based Learning (TBL) with those in a traditional, lecture-based approach. Method: Occurring across two consecutive semesters, we recruited students (n = 43) enrolled in an educational psychology course serving primarily pre-service teachers for this quasi-experimental study. In addition to final grades, data sources included responses to essay prompts administered at the beginning and end of the semester, which were qualitatively analyzed and coded. Results: Independent samples t-tests showed significant differences in favor of students in the TBL condition for course grades, but not on the final exam. χ2 tests of independence revealed significant differences for two of the seven essay codes, again in favor of TBL. Conclusion: These results provide further evidence that when placed in a context that emphasizes collaborative learning, students demonstrate more nuanced understanding and fewer misconceptions. Teaching Implications: The opportunity to experience TBL may provide pre-service teachers with better theoretical understanding and an effective model for translating socio-constructive theory into collaborative learning practices.",,
-COVID-19 pandemic prompts the development of a Web-OSCE using Zoom teleconferencing to resume medical students’ clinical skills training at Weill Cornell Medicine-Qatar,S. Major,"Web-Objective Structured Clinical Examination (OSCEs) were piloted for the Clinical Skills Assessment (CSA) exams in the USA two decades ago and were shown to be an acceptable way to conduct OSCEs remotely. The learners valued the process yet expressed numerous limitations.1 In response to the COVID-19 outbreak in Qatar, medical schools stopped in-person teaching creating a hiatus in clinical instruction. The Clinical Skills and Simulation Lab (CSSL) team devised and piloted a Web-OSCE to determine its feasibility in this setting, and to evaluate stakeholders’ experiences. We describe the steps taken to create a Web-OSCE built on women’s reproductive and sexual health which is part of the third year Obstetrics and Gynecology Clerkship curriculum, using Zoom teleconferencing. It mirrored the steps taken when conducting this activity in person; however, all communication relied on emails before the event, Zoom during the event with WhatsApp as a backup for connectivity between learners, Zoom hosts and faculty. ### Pre-Web-OSCE steps #### Preparing students CSSL provided by email a step-by-step instruction guide plus a pre-OSCE task for students to complete to ensure that they could access the OSCE on the day of the activity. A live Zoom session led by faculty reviewed the learning objectives, expectations and assessment and served as an open forum for students to ask questions in order to mitigate any anxiety surrounding the new modality. #### Preparing simulated patients Simulated patients (SP) were surveyed to assess their interest and technology capabilities to participate in Zoom activities. Interested SPs joined a Zoom meeting with staff members, who ensured they had adequate technical capability. SPs were selected and confirmed according to requirements of the case. Training notes and checklists were sent in …","{'model': 'tldr@v2.0.0', 'text': 'The steps taken to create a Web-OSCE built on women’s reproductive and sexual health which is part of the third year Obstetrics and Gynecology Clerkship curriculum, using Zoom teleconferencing are described.'}",https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7239650/pdf/bmjstel-2020-000629.pdf
-Book Review: Numeracy Across the Curriculum: Research Strategies for Enhancing Teaching and Learning,Glenda Anthony,"In Numeracy across the curriculum: Research strategies for enhancing teaching and learning, the authors provide a powerful resource in the form of practical, research-based strategies that will support prospective teachers, teachers, and teacher educators as they respond to the expectation that educators, across all levels and within all disciplines, will know and understand numeracy teaching strategies and their application. A critical and exciting feature of the book is the clear articulation that numeracy must move from a “ mathematics concern ” to an across curriculum focus. The authenticity of the advocated practices is informed by the authors’ many years of research in Australian educational, home, work, and community/civic settings. To that end, the text carefully explains the nature of numeracy and how it developed as an educational goal for the mathematics classroom and beyond. From the beginning chapters, and throughout, readers are prompted to engage in productive learnings with the inclusion of reflection prompts and numerous classroom exemplars of task design, assessment practices, and how to deal with the challenges and dilemmas of enactment in practice. urgent",,
-Pedagogical Prompts: Designing Experiences to Promote Deep Learning,Nicola Yelland,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter focuses on the ways in which the Learning by Design theoretical framework acts as a pedagogical prompt for teachers so that they are able to create learning experiences that complement the range of pedagogies described as knowledge processes and support 21st century skills and deep learning.'}",
-Challenges in Developing Automatic Learning Guidance in Relation to an Information Literacy Curriculum,Angela Fessl,"Becoming a data-savvy professional requires skills and competences in information literacy, communication and collaboration, and content creation in digital environments. In this paper, we present a concept for automatic learning guidance in relation to an information literacy curriculum. The learning guidance concept has three components: Firstly, an open learner model in terms of an information literacy curriculum is created. Based on the data collected in the learner model, learning analytics is used in combination with a corresponding visualization to present the current learning status of the learner. Secondly, reflection prompts in form of sentence starters or reflective questions adaptive to the learner model aim to guide learning. Thirdly, learning resources are suggested that are structured along learning goals to motivate learners to progress. The main contribution of this paper is to discuss what we see as main research challenges with respect to existing literature on open learner modeling, learning analytics, recommender systems for learning, and learning guidance.","{'model': 'tldr@v2.0.0', 'text': 'What the authors see as main research challenges with respect to existing literature on open learner modeling, learning analytics, recommender systems for learning, and learning guidance are discussed.'}",
-Implementation and Evaluation of a 10-Week Health Equity Curriculum for Pharmacy Students,S. Hsia,"Objective. To describe a health equity curriculum created for pharmacy students and evaluate students’ perceptions and structural competency after completion of the curriculum. Methods. A health equity curriculum based on transformative learning and structural competency frameworks was implemented as a 10-week mandatory component of the pass-no pass neuropsychiatric theme for second year pharmacy students. Each week, students reviewed materials around a neuropsychiatric-related health equity topic and responded to discussion prompts through asynchronous forums or synchronous online video discussions. The effectiveness of the health equity curriculum was evaluated through assessment of structural competency through a validated instrument, an objective structured clinical examination (OSCE), and a questionnaire. Results. All enrolled second year pharmacy students (n=124) participated in the health equity curriculum. Of the 75 students who completed the structural competency instrument, 46 (61%) were able to identify structural determinants of health, explain how structures contribute to health disparities, or design structural interventions. Ninety-six of the 124 students (77%) were able to address their OSCE standardized patient’s mistrust in the health care system. Thematic analysis of student comments elucidated three themes: allyship, peer connection, and self-awareness. Students rated asynchronous discussion forums as significantly less effective than online video discussions and patient cases for achieving curricular objectives. Conclusion. A mandatory curriculum delivered remotely throughout the didactic pharmacy curriculum using a blended learning approach was an effective way to incorporate health equity content and conversations into existing courses. Implementation of this or similar curriculums could be an important step in training pharmacy students to be advocates for social justice.","{'model': 'tldr@v2.0.0', 'text': 'A mandatory curriculum delivered remotely throughout the didactic pharmacy curriculum using a blended learning approach was an effective way to incorporate health equity content and conversations into existing courses.'}",https://www.ajpe.org/content/ajpe/85/9/8579.full.pdf
-What counts as nature in designing environmental links to health education curriculum in initial teacher education?,Bronwyn J. Cumbo,"ABSTRACT In this paper we report on a teacher education co-design project that explored Australian pre-service primary generalist teachers’ ideations of the pedagogical links between health education and nature. As part of their coursework in a Master of Teaching degree at an Australian University, students were invited to design a Victorian Curriculum: Health and Physical Education (VC:HPE) activity that connected primary curriculum (F-6) with nature. We conducted a 3-hour suite of online learning activities and prompts using Zoom, Padlet and Moodle. The data consists of the students’ curriculum design artefacts as well as recordings of the group discussions and non-assessment-based presentations of their work. Drawing on theories of child-nature interaction, we present an analysis of the ways in which n = 72 pre-service teachers across 18 groups of 4–5 educators conceptualised links between ‘nature’ and HPE across their activity designs. Prior to the workshop, the majority of pre-service teachers had not previously considered links between the HPE learning area and nature. Through the activities of the co-design workshop, students were surprised with the variety of pedagogical possibilities that were able to be made. There was much student discussion about the possibilities and limitations of balancing safety and risk in their nature-based activity designs. We present a thematic analysis of the quality of student-nature interactions in the groups’ learning designs through: (i) exploration; (ii) embodiment; (iii) cultivation; (iv) appropriation; and (v) representation. The analysis and discussion has implications for the way quality health education is linked to nature-based learning environments, teacher education and contemporary curriculum enactment that incorporates nature and the environment as part of the learning design.",,
-Curriculum Q-Learning for Visual Vocabulary Acquisition,Ahmed H. Zaidi,"The structure of curriculum plays a vital role in our learning process, both as children and adults. Presenting material in ascending order of difficulty that also exploits prior knowledge can have a significant impact on the rate of learning. However, the notion of difficulty and prior knowledge differs from person to person. Motivated by the need for a personalised curriculum, we present a novel method of curriculum learning for vocabulary words in the form of visual prompts. We employ a reinforcement learning model grounded in pedagogical theories that emulates the actions of a tutor. We simulate three students with different levels of vocabulary knowledge in order to evaluate the how well our model adapts to the environment. The results of the simulation reveal that through interaction, the model is able to identify the areas of weakness, as well as push students to the edge of their ZPD. We hypothesise that these methods can also be effective in training agents to learn language representations in a simulated environment where it has previously been shown that order of words and prior knowledge play an important role in the efficacy of language learning.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesised that these methods can also be effective in training agents to learn language representations in a simulated environment where it has previously been shown that order of words and prior knowledge play an important role in the efficacy of language learning.'}",
-Redesigning curriculum at the higher education level: challenges and successes within a sport management program,Wendi K. Zimmer,"ABSTRACT Higher education programs face challenges to increase accountability regarding student achievement. Course misalignment occurs as instructors interpret course objectives differently, making changes to the course content without considering overall program alignment and purpose. Misalignment prompts academic programs to revisit its curriculum to assess effectiveness. Faced with a rapidly growing and increasingly diverse student population, the Sport Management (SPMT) division at a southwestern university used action research to redesign their undergraduate curriculum to meet their students’ needs. This study assesses the impact of using action research to determine why the SPMT program exists, what a successful SPMT program includes, and what knowledge, skills, abilities, and values industry leaders report necessary for SPMT graduates. Furthermore, methods for addressing challenges and successes through action research are discussed. As a result of the action research process and data collection, eight program learning outcomes were established to guide the next steps of curriculum development. These findings guided syllabus creation for a redesigned SPMT program, allowing for the creation of a rigorous, research based, and customized curriculum.",,
-Analyzing preservice secondary mathematics teachers’ prompts in dynamic geometry environment tasks,Hilal Gulkilik,"ABSTRACT The purpose of this study was to analyze the prompts that preservice secondary mathematics teachers used for the acquisition of mathematics knowledge in dynamic geometry environment tasks. The participants, four preservice secondary mathematics teachers who were enrolled in a computer-supported mathematics education course, designed a dynamic geometry environment task based on a high school mathematics curriculum learning outcome. The main data sources were the participants’ task documents and the transcripts of interviews that were conducted with the participants to examine the details of their tasks. The techno-pedagogic mathematics task design model was used to analyze the data. The results indicated that the focus of the designs was to help students realize the invariant properties of geometric figures that were embodied by robust construction steps in the tasks. The preservice teachers utilized several capabilities of the dynamic geometry environment (e.g. measuring, dragging, and changing the input box value) to help students discern these properties and expected students to make a generalization based on inductive reasoning. However, since students were directed to build robust constructions, the mathematical activities prompted by the preservice teachers were limited to observing, finding, generalizing, and explaining the previously constructed invariant features in the tasks.",,
-The Impact of Scaffolding Prompts on the Collaborative Problem Solving of Ill-Structured Tasks by Undergraduate Engineering Student Groups,Miss Taylor Tucker,"Taylor Tucker received her Bachelor’s degree in engineering mechanics from the University of Illinois at Urbana-Champaign and is now pursuing a Master’s in curriculum and instruction through the Digital Environments for Learning, Teaching, and Agency (DELTA) program. She is interested in engineering design and lends her technical background to her research with the Collaborative Learning Lab, exploring how to improve ill-structured tasks for engineering students in order to promote collaborative problem solving and provide experience relevant to authentic work in industry.",,
-Support Factors and Barriers for Outdoor Learning in Elementary Schools: A Systemic Perspective,E. Oberle,"ABSTRACT Background Outdoor learning offers clear physical, cognitive, social-emotional and academic benefits for children and yet, it is considered a grassroots approach to teaching and learning in elementary schools. Purpose We examined teachers’ perspectives on barriers and supports for outdoor learning in public elementary schools. Methods Thirty-six teachers in (urban and rural) British Columbia, Alberta, and Ontario (all female; Mean age = 43.84, SD = 10) participated in one of five virtually administered, semi-structured focus groups. Questions/prompts facilitated a discussion on teachers’ experiences with barriers and supports for outdoor learning. Thematic analysis was used to identify main themes. Results Four interrelated themes and further sub-themes were found: 1) Teacher characteristics: interest/motivation to teach outdoors, preparedness, confidence in handling risks; 2) Systemic factors: principal support, school/district policies, funding/resources, curriculum, school schedule; 3) Culture: school culture, societal beliefs about education, family backgrounds; 4) Environmental factors: weather, built/natural environment, hazards. Discussion Systemic support is needed to integrate outdoor learning in schools. Translation to Health Education Practice The findings in this study are relevant to health education specialists particularly focused on elementary school education.",,
-Evaluating the impact of reflecting on curriculum-embedded skill development: the experience of science undergraduates,Michelle A. Hill,"ABSTRACT Meaningful reflection on their learning and skill development is often lacking in the experience of undergraduates. Many students do not recognise the curriculum-embedded development of transferable skills and lack the ability to articulate such skills. This mixed-methods study sought to investigate whether engaging students in reflection would increase their ability to recognise and articulate their skill development. Sixty science undergraduates from Monash University completed a voluntary semester-long program recording and reflecting on course-related skill development, supported by email prompts and group discussions. The impact of students’ involvement was evaluated through pre- and post-participation surveys, reflections and group discussions. Most students were challenged by the unfamiliarity of thinking beyond knowledge attainment in order to identify and reflect on skill-related experiences. However, they recognised a range of benefits from doing so, including an improved ability to recognise their skill development, strengths and weaknesses and to articulate their skills in readiness for seeking employment. They also valued previously unappreciated learning tasks and gained motivation to improve skill deficits and seek out opportunities to improve their employability. Based on this study, recommendations are made regarding best practice for implementing skills reflection in the curriculum.",,
-Meta-augmented Prompt Tuning for Better Few-shot Learning,Kaihang Pan,"Prompt tuning is a parameter-efficient method, which freezes all PLM parameters and only prepends some additional tunable tokens called soft prompts to the input text. However, soft prompts heavily rely on a better initialization and may easily result in overfitting under few-shot settings, which causes prompt-tuning performing much worse than fine-tuning. To address the above issues, this paper proposes a novel S elf-s U pervised M eta-prompt learning framework with ME ta-gradient R egularization for few-shot generalization ( SUMMER ). We leverage self-supervised meta-learning to better initialize soft prompts and curriculum-based task augmentation is further proposed to enrich the meta-task distribution. Besides, a novel meta-gradient regularization method is integrated into the meta-prompt learning framework, which meta-learns to transform the raw gradient during few-shot learning into a domain- generalizable direction, thus alleviat-ing the problem of overfitting. Extensive experiments show that SUMMER achieves better performance for different few-shot downstream tasks, and also exhibits a stronger domain generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-gradient regularization method is integrated into the meta-prompt learning framework, which meta-learns to transform the raw gradient during few-shot learning into a domain- generalizable direction, thus alleviat-ing the problem of overfitting.'}",http://arxiv.org/pdf/2303.12314
-Learning to critique and adapt science curriculum materials: Examining the development of preservice elementary teachers' pedagogical content knowledge,Carrie J. Beyer,"Teachers often engage in curricular planning by critiquing and adapting existing curriculum materials to contextualize lessons and compensate for their deficiencies. Designing instruction for students is shaped by teachers' ability to apply a variety of personal resources, including their pedagogical content knowledge (PCK). This study investigated a criterion-based approach to lesson plan analysis as one way to help preservice elementary teachers develop and use their PCK to plan instruction for students. Results show that the preservice teachers demonstrated a range of strengths and weaknesses in applying their knowledge of science assessment, science curriculum materials, and instructional strategies for teaching science. This range was influenced, in part, by the presence of alternative ideas about science teaching, the extent to which the original curriculum materials aligned with reform-based standards and practices, and the presence of prompts to use criteria in their analyses. Despite these factors, preservice teachers' PCK improved significantly over time when they had multiple opportunities to practice applying the same criterion in their analyses. Insights into science teacher knowledge and implications for science teacher education are discussed. © 2011 Wiley Periodicals, Inc. Sci Ed96:130–157, 2012",,http://deepblue.lib.umich.edu/bitstream/2027.42/89555/1/20466_ftp.pdf
-Promoting Social Emotional Learning through Shared Book Reading: Examining Teacher’s Strategies and Children’s Responses in Kindergarten Classrooms,S. Ng,"ABSTRACT Research Findings: Early childhood educators play an important role in promoting children’s social emotional learning (SEL). This study proposes a comprehensive coding scheme to explore teachers’ strategies and children’s responses in shared book reading (SBR), which create opportunities to share pictures and storylines to examine social circumstances and emotional situations. Drawing on videotaped observations of nineteen classrooms from a larger study in Singapore, SEL-related interactions consisted of illustrations or discussions with indicators from the five SEL areas outlined in the national curriculum. Results revealed that teachers’ support for SEL occurred more frequently when facilitating interpersonal learning over intrapersonal learning, while children’s responses accorded with the types of teachers’ input. In response to teachers’ strategies within the intrapersonal dimension, children provided responses to teachers’ prompts for identifying self-abilities and preferences most frequently. Within the interpersonal dimension, children provided responses to teachers’ prompts for labelling others’ emotions most frequently. Practice or Policy: Our findings demonstrate that SBR can be useful in promoting SEL even without books that explicitly convey skills and messages relating to SEL. The results imply that teachers should increase their input on facilitating intrapersonal learning topromote children’s balanced development in the understanding of themselves and others.",,
-Improving significance of binary black hole mergers in Advanced LIGO data using deep learning : Confirmation of GW151216,S. Jadhav,"We present a novel Machine Learning (ML) based strategy to search for compact binary coalescences (CBCs) in data from ground-based gravitational wave (GW) observatories. This is the first ML-based search that not only recovers all the binary black hole mergers in the first GW transients calalog (GWTC-1), but also makes a clean detection of GW151216, which was not significant enough to be included in the catalogue. Moreover, we achieve this by only adding a new coincident ranking statistic (MLStat) to a standard analysis that was used for GWTC-1. In CBC searches, reducing contamination by terrestrial and instrumental transients, which create a loud noise background by triggering numerous false alarms, is crucial to improving the sensitivity for detecting true events. The sheer volume of data and and large number of expected detections also prompts the use of ML techniques. We perform transfer learning to train ""InceptionV3"", a pre-trained deep neural network, along with curriculum learning to distinguish GW signals from noisy events by analysing their continuous wavelet transform (CWT) maps. MLStat incorporates information from this ML classifier into the standard coincident search likelihood used by the conventional search. This leads to at least an order of magnitude improvement in the inverse false-alarm-rate (IFAR) for the previously ""low significance"" events GW151012, GW170729 and GW151216. The confidence in detection of GW151216 is further strengthened by performing its parameter estimation using SEOBNRv4HM_ROM. Considering the impressive ability of the statistic to distinguish signals from glitches, the list of marginal events from MLStat could be quite reliable for astrophysical population studies and further follow-up. This work demonstrates the immense potential and readiness of MLStat for finding new sources in current data and possibility of its adaptation in similar searches.","{'model': 'tldr@v2.0.0', 'text': 'This is the first ML-based search that recovers all the binary black hole mergers in the first GW transients calalog (GWTC-1), but also makes a clean detection of GW151216, which was not significant enough to be included in the catalogue.'}",https://dspace.library.uu.nl/bitstream/handle/1874/413428/PhysRevD.104.064051.pdf?sequence=1&isAllowed=y
-The impact of multimedia educative curriculum materials (MECMs) on teachers’ beliefs about scientific argumentation,S. Loper,"ABSTRACT Recent reform efforts in science education include a focus on science practices. Teachers require support in integrating these practices into instruction. Multimedia educative curriculum materials (MECMs), digital materials explicitly designed to support teacher learning, offer one potential resource for this critical need. Consequently, the authors investigated how teachers used MECMs and whether that use impacted their beliefs about the practice of scientific argumentation. They conducted a randomised experimental study with 90 middle school science teachers in the USA. Control and experimental groups taught the same curriculum, using a web-based teacher’s guide. Additionally, experimental teachers received MECMs including 24 videos and 17 interactive reflective prompts. The authors collected multiple data sources: pre-surveys, backend website curriculum use, self-report curriculum use and post-surveys. Results suggest that enacting a curriculum with a focus on argumentation is associated with positive changes in teachers’ beliefs about this practice. Furthermore, the authors observed a wide range in how teachers used the curriculum. In terms of self-efficacy, this differential use was associated with differences in changes to teachers’ beliefs about argumentation. Teachers who enacted more lessons became more confident in their ability to teach argumentation. Additionally, experimental teachers had smaller improvements in self-efficacy, perhaps because the MECM videos problematised what teachers thought counted as argumentation.",,
-"PythonSneks: An Open-Source, Instructionally-Designed Introductory Curriculum with Action-Design Research",Austin Cory Bart,"Rising enrollments and limited instructor resources underscores the growing need for reusable, scalable curriculum. In this paper, we describe an open-source introductory Python course for non-Computer Science majors in STEM, designed following best practices of Instructional Design (a process similar to Software Engineering). The created resources include 234 learning objectives, 51 lesson videos, 45 lecture slides, 170 programming problems, 281 quiz questions, 6 unit tested projects, and 4 ethical prompts. A teaching field guide has also been produced as a result of this effort, documenting how to deploy this curriculum on a daily level. We describe our experiences deploying over two semesters. The course serviced over 500 students, with 100s in some sections. Along the way, two interventions were conducted in an Action Design Research style: one using Worked Examples, and another using Structured Small Groups. We report on the mixed results of these experiments, plus evaluations of the assignments from student surveys and statistical measures of item effectiveness. Finally, we describe lessons learned when following Instructional Design processes.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes an open-source introductory Python course for non-Computer Science majors in STEM, designed following best practices of Instructional Design (a process similar to Software Engineering).'}",
-Toward Expert Thinking: How curriculum case writing prompts the development of theory-based professional knowledge in student teachers,Karen Hammerness,"The present paper explores what, and how, student teachers may learn about theory and practice from writing cases, and examines some pedagogical features that may contribute to these results. Drawing on data collected from our course ""Principles of Learning for Teaching"", including student cases from outline to final drafts and students' course reflections, we found that students' successive case drafts demonstrated a development from naïve generalizations to sophisticated, theory-based explanations of the issues at play in their cases. In particular, we suggest that students' cases demonstrated some of the moves that Berliner (1986, 1991) identified as characteristic of more ""expert"" thinking about teaching. We propose that reading theory in context with writing cases, that sharing cases with peer readers, that specific, theoretically grounded, and concrete feedback from instructors, and that providing multiple opportunities for revision may have been most useful in helping student teachers learn to think like a teacher.",,
-The Curriculum Innovation Canvas: A Design Thinking Framework for the Engaged Educational Entrepreneur,C. Willness,"Integrating literature on entrepreneurial business models and community-based experiential learning, we propose a new framework to advance the practice of curriculum innovation. Grounded in principles of design thinking, the curriculum innovation canvas provides a human-centered, collaborative, and holistic platform for instructors, curriculum developers, and administrators to engage in innovation and implementation of experiential courses or programs—particularly those that involve community or organizational partnerships. The canvas promotes a creative and fluid approach to curriculum development. It prompts the consideration of the value propositions offered to various stakeholders (students, community partners, faculty peers, etc.) as well as how to involve stakeholders in the development and implementation process toward mutually beneficial outcomes in a complex and challenging environment. Evidence from an extensive prototyping process indicates that it can effectively assist instructors, administrators, students, and community partners in a variety of contexts.",,
-"Learning Chemistry: Self-Efficacy, Chemical Understanding, and Graphing Skills",Shirly Avargil,,,
-What millennial medical students say about flipped learning,Robin K Pettit,"Flipped instruction is gaining popularity in medical schools, but there are unanswered questions such as the optimum amount of the curriculum to flip and whether flipped sessions should be mandatory. We were in a unique position to evaluate feedback from first-year medical students who had experienced both flipped and lecture-based courses during their first semester of medical school. A key finding was that the students preferred a variety of different learning formats over an “all or nothing” learning format. Learning format preferences did not necessarily align with perceptions of which format led to better course exam performance. Nearly 70% of respondents wanted to make their own decisions regarding attendance. Candid responses to open-ended survey prompts reflected millennial preferences for choice, flexibility, efficiency, and the ability to control the pace of their learning, providing insight to guide curricular improvements.","{'model': 'tldr@v2.0.0', 'text': 'Feedback from first-year medical students who had experienced both flipped and lecture-based courses during their first semester of medical school found that the students preferred a variety of different learning formats over an “all or nothing” learning format.'}",https://www.dovepress.com/getfile.php?fileID=37537
-Improving Self-Regulated Learning through Personalized Weekly e-Learning Journals: A Time Series Quasi-Experimental Study.,Chorng Yuan Fung,This study provides an insight into using personalized weekly e-Learning Journals to improve self-regulated learning (SRL) of university students. Quasi-experimental method with time series data analysis was used. Pre and post-tests together with time series data over an intervention period on SRL were collected. 54 students in an undergraduate course wrote personalized weekly e-Learning Journals (e-LJs) in the Learning Management System (LMS) over a 10-week period. The e-LJs contained selfreflection prompts designed according to the course curriculum activities and assessments to scaffold students’ SRL. It was found that students’ SRL improved significantly over the intervention period. The time series data on SRL shows that students’ SRL varied according to the timing of assessments. Academic staff can help university students improve their SRL by providing personalized weekly e-LJs that contained self-reflection prompts. These prompts need to be personalized according to the course curriculum activities and assessments.,,
-Exploring the Impacts of Cognitive and Metacognitive Prompting on Students’ Scientific Inquiry Practices Within an E-Learning Environment,Wenxia Zhang,"This study explores the effects of metacognitive and cognitive prompting on the scientific inquiry practices of students with various levels of initial metacognition. Two junior high school classes participated in this study. One class, the experimental group (n = 26), which received an inquiry-based curriculum with a combination of cognitive and metacognitive prompts, was compared to the other class, the comparison group (n = 25), which received only cognitive prompts in the same curriculum. Data sources included a test of inquiry practices, a questionnaire of metacognition, and worksheets. The results showed that the mixed cognitive and metacognitive prompts had significant impacts on the students’ inquiry practices, especially their planning and analyzing abilities. Furthermore, the mixed prompts appeared to have a differential effect on those students with lower level metacognition, who showed significant improvement in their inquiry abilities. A combination of cognitive and metacognitive prompts during an inquiry cycle was found to promote students’ inquiry practices.",,
-Vygotsky in Action in the Early Years: The Key to Learning Curriculum,G. Dolya,"@contents: Selected Contents: CONTENTS FOREWORD LEV VYGOTSKY - A THINKER WHO CHANGED TEACHING VIVA VYGOTSKY! DEVELOPING LEARNING ABILITIES, UNLOCKING POSSIBILITIES UNDERSTANDING ABILITIES THE MECHANISM OF ABILITIES THE CLASSIFICATION OF ABILITIES A DEVELOPMENTAL CURRICULUM TWELVE PROGRAMMES, ONE PRACTICE THE PEDAGOGICAL PROCESS GROUP WORK SUBSTITUTION AND VISUAL MODELLING EXTERNAL MEDIATORS: VISIBLE PROPS AND PROMPTS FOR LEARNING COGNITIVE TASKS VISUAL MODELLING AND PRODUCTIVE IMAGINATION IN CONCLUSION FROM SENSORY STANDARDS TO VERBAL REASONING - A THEORETICAL PERSPECTIVE SENSORY STANDARDS AND PERCEPTUAL MODELLING CREATIVE MODELLING MATHEMATICS LOGIC CONSTRUCTION FROM THEORY TO PRACTICE - THE KEY TO LEARNING DEVELOPMENTAL CURRICULUM SENSORY MATHEMATICS LOGIC MATHEMATICS STORY GRAMMAR DEVELOPMENTAL GAMES ARTOGRAPHICS VISUAL-SPATIAL CREATIVE MODELLING CONSTRUCTION EXPLORATION EXPRESSIVE MOVEMENT YOU-ME-WORLD PARENT POWER - STRATEGIES FOR BUILDING A POSITIVE PARTNERSHIP ALL WORK AND NO PLAY IT'S NEVER TOO LATE TO LEARN, IT'S NEVER TOO EARLY TO TEACH ""HOMEWORK"" FOR PARENTS WHAT PEOPLE SAY... PROFESSIONAL OPINIONS WHAT PARENTS SAY REFERENCES",,
-Arctic Climate Connections Curriculum: A Model for Bringing Authentic Data Into the Classroom,A. Gold,"ABSTRACT Science education can build a bridge between research carried out by scientists and relevant learning opportunities for students. The Broader Impact requirements for scientists by funding agencies facilitate this connection. We propose and test a model curriculum development process in which scientists, curriculum developers, and classroom educators work together to scaffold the use of authentic, unprocessed scientific data for high school students. We outline a three-module curriculum structure that facilitates these goals. This curriculum engages students in the collection, description, visualization, and interpretation of data; develops understanding of the nature of science; includes prompts to develop higher-order thinking skills; builds knowledge of regional relevance of climate change in students; uses active learning techniques; and can be easily integrated with the Next Generation Science Standards. The curriculum was reviewed and tested in the classroom. To shed further light on the curriculum development process, we gathered reflection data from the scientists, curriculum developers, and educators. Scientists appreciated the collaborative process in which they contributed their expertise without requiring a large time commitment or strong expertise in science education. The curriculum developers viewed the modular structure as helpful in breaking complicated scientific concepts into teachable steps. Classroom educators appreciated the detailed description and step-by-step instructions to navigate data analysis tools like Excel or Google Earth. Initial classroom implementation of the curriculum by 11 teachers with over 1,100 students showed high levels of interest in the topic and engagement. Further work is needed to assess efficacy of the curriculum through classroom observations and measures of student learning.",,
-Scaffolding Collaborative Reflective Writing in a VET Curriculum,E. Boldrini,,,
-Empowering Students With Word-Learning Strategies: Teach a Child to Fish,Michael F. Graves,"This article on word-learning strategies describes a theory- and research-based set of procedures for teaching students to use word-learning strategies—word parts, context clues, the dictionary, and a combined strategy—to infer the meanings of unknown words. The article begins with a rationale for teaching word-learning strategies, particularly to students with vocabularies smaller than those of many students their age. After this, the authors provide a definition of word-learning strategies, a review of the most relevant research, and a brief description of the effects of the program. Next are descriptions of the curriculum, the instruction, and key aspects of the authors’ approach. The article concludes with a note stressing the importance of following initial instruction on word-learning strategies with reviews, reminders, and prompts to use them over time and a description of two aspects of the program identified as particularly important by the teachers the authors worked with.",,
-Good practices in European Short Learning Programmes (E-SLP),M. Maina,"collaborative inter-institutional curriculum designs. The outcome of these collaborations can be SLPs built from scratch or as puzzles (collection of existing learning building blocks). One of the consequences of international collaborative design is that it introduces distinct working practices and therefore requires teams to set innovative processes. It also brings opportunities from each institution to exchange policies and triggers new consensus. Furthermore, it requires harmonisation of recognitions and accreditations, which in turns prompts innovatory agreements.",,
-Increasing Student Learning through Arts Integration,S. Catherine,"This action research study was designed to examine the effects of the arts integration on student learning. Student learning was broken down by focusing on student engagement, critical thinking, and creative response. Research was conducted with 72 participants from fourth grade, fifth grade, and ninth grade classrooms within the social studies curriculum. Data collection methods included a pre-assessment survey, student journal prompts, observation tally sheets, student self-assessments, and the post-assessment survey. The results of our study indicated that student learning in each classroom maintained or improved when the arts were integrated. Students showed enthusiasm and felt optimistic towards integrating the arts into their learning. According to our student survey many students noted an increase in their critical thinking skills and explained that the arts helped them display their knowledge and creativity. However, in the future we would guide students to be more specific in their journal responses to better communicate their ideas. As a result of our study, we have committed to improving student learning by continuing to integrate the arts into the curriculum.",,
-A making and gaming approach to learning about RF path loss and antenna design,Kevin J. Richardson,"As part of an ongoing, longitudinal study on the use of “making” and “gaming” in the classroom, two sequential activities for learning about radio-frequency (RF) path loss and antenna design are presented. “Making” involves integration of makerspace concepts and tinkering in the curriculum, while “gaming” refers to gamified curricula; in this study we investigate the joint use of these two elements in the classroom. The RF path loss activity is modeled after ham radio “fox hunting”, where students must locate a transmitter hidden on campus; it makes use of low-cost software-defined radios, and prompts students to confront concepts including measuring signal power, frequency domain thinking, and antenna polarization. The follow-up activity challenges students to build an antenna designed to receive household gas meter readings; students must design their antennas specifically for operation in the 900 MHz band, and must give a presentation describing the theory of their antenna to their peers. A competition is held where students attempt to see which of their antennas can collect the most wireless gas meter readings over a five-minute interval. Assessment data from the broader study show that relative to a baseline offering, the treatment group developed an improvement in interest, perception, independence, and self-assessed abilities. This paper discusses the implementation of the activities, the students' approach to solving the proposed challenges, the assessment data, lessons learned from student focus groups, and instructor observations.","{'model': 'tldr@v2.0.0', 'text': 'Assessment data from the broader study show that relative to a baseline offering, the treatment group developed an improvement in interest, perception, independence, and self-assessed abilities.'}",
-"A socially responsive curriculum: an alternative for elt in ""Popular education"".",Orley Cruz Arcila,"This study reported on qualitative research that focused on the presentation of the current situation at the school de Cultura Popular with regards to the English Curriculum designed and implemented into the school specifically in 11th grade. This paper also contains a revision of the different social and institutional conditions that may have an impact on English language teaching (ELT) at the school. It was carried out with 36 students of fifth cycle at I.E.D de Cultura Popular, in Bogota. The main aim of this study was to re-design the English curriculum for 11th grade, considering students’ social context and learning interests for ELT by incorporating in its design the prompts of Popular Education, to give to the curriculum a social vision as a new alternative for ELT at the school. The instruments used for collecting data were teacher’s journal, students’ journal; represented in the elaboration of a “book” in English; and some documentary evidence as P.E.I and the socio-economical study of the students to identify in a more realistic way their current social and cultural context. The results of the study showed that the implementation of activities under the orientation of Popular Education, students develop a critical position about their social and educational reality that permits them to establish their own objectives in a long and short term.",,
-Hollowed-out genring as a way of purposefully embracing troublesome knowledge: Orientation and de-orientation in the learning and teaching of fine art,Shaun Camp,"This text emerges from ongoing research focused on pedagogical developments within the curriculum of the Year 0 Fine Art Pathway, at Norwich University of the Arts. Within any subject discipline, including Fine Art, there are threshold concepts (Meyer and Land, 2003, p.1) that can represent troublesome knowledge (Meyer and Land, 2003, p.2); knowledge that must be comprehended in order that students succeed within that discipline. Within the teaching and learning of Fine Art, students will work between both written and visual genres, often struggling to rationalise the two. As part of our research we have developed practiced based research tasks that serve as destabilising prompts, a hollowed out genre, that encourages students to enter into a de-orientated, liminal, conceptual space within their learning. With the introduction of reflective models, students are able to navigate this liminal space and are afforded opportunities for developmental re-orientations, thus beginning to embrace the threshold concepts inherent with the Fine Art discipline. Drawing upon examples of current practice and student learning, we demonstrate how a process of genring and the use of reflective tools encourage students to confront aspects of troublesome knowledge and see critical writing as an important facet of their practice.",,https://nua.collections.crest.ac.uk/17235/1/RegenringFinal.pdf
-Impact of the course teaching and learning of mathematics on preservice grades 7 and 8 mathematics teachers in Singapore,Berinderjeet Kaur,,,
-Discussion of Identity and Implicit Bias: A Strategy to Address Racism and Social Justice in Pediatric Nursing Education.,Kristen E. Stevens,"ABSTRACT Nurse educators must weave discussions of systemic racism, social justice, social determinants of health, and psychosocial influences throughout the curriculum. For an online pediatric course, an activity was developed to raise awareness of implicit bias. This experience interfused assigned readings from the literature, introspection of identity, and guided discussion. Framed by principles of transformative learning, faculty facilitated an online dialogue involving groups of 5 to 10 students through aggregated self-descriptors and open prompts. Ground rules for the discussion established psychological safety. This activity complements other schoolwide racial justice initiatives.",,
-Blended matters: Nurturing critical thinking,Pooja Jaswal,"The Organization for Economic Cooperation and Development (OECD) Learning Framework-2030 presents critical thinking as one of the transformative competencies for incorporation into the curriculum. Critical thinking refers to one’s ability to distinguish relevant information from irrelevant information, relate ideas, infer and justify whether a particular assumption is acceptable or unacceptable. This article argues that blended learning provides a concrete and meaningful context for enhancing learners’ critical thinking. Through collaboration, blended learning makes it possible to control, observe interactions, regulate tasks, and mediate in the acquisition of new knowledge. This form of collaboration creates a student-centered learning environment. Such a shift prompts to examine three objectives: (1) comprehend the creation of a blended learning environment. (2) discuss the characteristics of collaborative blended learning. (3) examine the role of collaboration in developing critical thinking among learners. To achieve the purpose of implementing a blended learning strategy in the classroom, the classroom teacher must create collaborative and participative learning experiences. Blended learning brings collaborative experience to develop higher-order thinking skills such as critical thinking.",,
-Analytics for Supporting Teaching Success in Higher Education: A Systematic Review,Dirk Ifenthaler,"Learning analytics are utilized to support learners’ educational needs as well as to enhance their study success, for example, via the use of real-time prompts, motivational dashboards, and appropriate learning interventions, which have been shown to increase students’ academic performance as well as their course retention rates. Yet, the perspective of higher education teachers in utilizing analytics to help analyze, reflect on, and improve their teaching design prior to delivery as well as to monitor and measure how the students engaged with their learning processes has been less recognized. In this paper, we present the results of a systematic review conducted from higher education teachers’ perspective concerning how analytics can be deployed to adapt the curriculum to suit better students’ educational needs in order to increase their study success. Thirty-five key studies were identified showing that analytics have been successful in influencing positively study success via teachers’ academic curriculum intervention. Specifically, via analytics, higher education teachers could rapidly visualize common course pathways and identify any difficulties students experienced in real-time in order to increase their learning experiences and outcomes.",,
-The Effects of Intraprofessional Collaborative Case Based Learning: A Cohort Study of Student Physical Therapist and Physical Therapist Assistant Perceptions of the Physical Therapist Role,Y. Colgrove,"Background: The changing nature and complex regulation of healthcare require the efficient use of resources, including the appropriate delegation and supervision of the physical therapist assistant (PTA). Knowledge of the scope of PTA practice introduced in the academic curriculum is mandated for entry-level practice. This study assessed the effect of a collaborative case-based educational intervention within the didactic curriculums of a physical therapy (PT) and PTA program on student knowledge of PTA scope of practice. Methods and Findings: A pre- and post-test research design was used. Students completed a validated survey exploring their perceptions of the PTA role before beginning the case study. The case study was a classroom assignment followed by instructional prompts requiring interactions between student cohorts three times over four weeks. Following case study completion, students completed the same survey. Independent and paired samples t -tests detected significant differences between and within groups ( p < .05). Conclusions: Based on the results, the case-based instructional model was efficacious in teaching both student cohorts about the role of the PTA. The impact was greater on the accuracy of the PT students, but PTA students became less uncertain in their perceptions. The effect of the clinical learning environment should be investigated to determine the impact on student perception of PTA role delineation following didactic instruction.","{'model': 'tldr@v2.0.0', 'text': 'The case-based instructional model was efficacious in teaching both student cohorts about the role of the PTA, and the impact was greater on the accuracy of the PT students, but PTA students became less uncertain in their perceptions.'}",https://jripe.org/index.php/journal/article/download/185/112
-Advancing English Language Learning in China through Multimodal Content Area Teaching,C. North,"During a ten-day teaching abroad experience in China, eight teachers from the United States implemented an interactive curriculum focusing on disciplinary literacy and authentic tasks. Employing multiliteracies and kidwatching, teachers encouraged Chinese students to compose while focusing primarily on communicating ideas rather than grammatical correctness. This article provides a one-student case study that serves as a representative example of the growth of 50 elementary-level students involved in the experience. Initially, Paul focused on writing correctness in response to prompts; his compositions were short and provided little detail. After we provided multimodal and interactive authentic experiences and encouraged risk-taking, Paul's representative compositions became more detailed and complex. The implications for engaging in this type of teaching experience underscore the benefits of providing students with authentic experiences that are multimodal and interactive while simultaneously encouraging risk-taking. The pedagogical growth that teachers made working with ELL students is also discussed. Normal 0 false false false EN-US JA X-NONE",,
-Could probability be out of proportion? Self-explanation and example-based practice help students with lower proportional reasoning skills learn probability,K. Begolli,,,https://link.springer.com/content/pdf/10.1007/s11251-021-09550-9.pdf
-Global health on the front lines: an innovative medical student elective combining education and service during the COVID-19 pandemic,Brandon Altillo,,"{'model': 'tldr@v2.0.0', 'text': 'The elective was well-received by students, achieved stated objectives, and garnered public attention; course leadership should monitor students’ time commitment closely in service-learning settings to ensure appropriate balance of service and education.'}",https://bmcmededuc.biomedcentral.com/counter/pdf/10.1186/s12909-021-02616-9
-Nurturing Graduate Attributes through a First Year Student Experience Which Promotes the Formation of Effective Learning Communities,D. Leung,"Many students enter university with naive epistemological beliefs and study approaches incompatible with the goals of higher education or the display of attributes such as critical or creative thinking. This study examines whether a first year experience can promote the formation of effective learning communities, which in turn can provide a mechanism for nurturing a range of graduate attributes. The curriculum encompasses initiatives to assimilate students into the university, prompts the adaptation of appropriate university study behaviour, and contains a general education component to broaden the student experience. The impact of the first year experience was examined by collecting quantitative data which measured students’ perceptions of the effectiveness of assimilation, adaptation of study behaviour, the impact of the broadening component, and the nurturing of attributes. The data were analysed with structural equation modelling. A model in which the curriculum elements impacted on the development of graduate attributes showed a good fit to the data. Effective learning communities played a key role both in assimilation into the university community and the adoption of appropriate study behaviours.",,
-AutoTutor: A tutor with dialogue in natural language,A. Graesser,,"{'model': 'tldr@v2.0.0', 'text': 'The design was inspired by explanation-based constructivist theories of learning, intelligent tutoring systems that adaptively respond to student knowledge, and empirical research on dialogue patterns in tutorial discourse.'}",https://link.springer.com/content/pdf/10.3758/BF03195563.pdf
-Heutagogy and e-learning in the workplace: some challenges and opportunities,S. Hase,"It is clear that e-learning offers tremendous opportunities in terms of access in a global environment as well as in terms of the economics of delivery for workplace learning and, indeed, other learning environments. However, those designing and delivering workplace e-learning also have the potential to challenge existing dogmas about the curriculum, how we conduct education programs, manage knowledge, and access and harness learning. Complexity theory, in particular, prompts us to question how we understand education, training and learning. This article discusses some of these challenges and possibilities for workplace e-learning through the lens of the recently developed concept of heutagogy, defined as the study of self-determined learning.",,
-Using Curriculum-Based Measurement for Struggling Beginning,Kristen L. McMaster,"loves to learn. He participates enthusiastically during shared storybook reading with his class. During independent reading time, he becomes immediately engrossed in the colorful books about science and sports that his teacher, Mrs. Graff, displayed on the classroom bookshelf. He is a whiz at mathematics and sometimes asks for extra problems to solve. At home, he reads National Geographic and books about animals, oceans, and dinosaurs with his dad and older sister. Despite Max’s excitement about learning, Mrs. Graff noticed that Max is experiencing difficulties with his handwriting, spelling, and written composition assignments. During daily journal time, Max typically busies himself with sharpening pencils, wandering around the classroom, or staring out the window. After several prompts from Mrs. Graff to get to work, he attempts to write a few words, in barely legible handwriting, with lots of crossing out and erasing, sometimes to the point of making holes in his paper. On several occasions, she has observed him balling up his paper and throwing it off his desk. His mother has expressed concern that when she tries to help him study his spelling words, he becomes easily frustrated. He often fails the Friday spelling test. Mrs. Graff is concerned that Max’s struggles with writing will increasingly affect his enjoyment of learning and his success in school.",,
-A Rubric for Assessing Student Drawings & Reasoning to Gauge Understanding of Osmosis & Diffusion,A. Angra,"Abstract The concepts of osmosis and diffusion are essential to biology, and student difficulties with them are well documented, along with multiple ways of assessment and diagnosis. We add to the literature by sharing how we used drawing and reasoning prompts with first-year undergraduate biology students to gauge learning before and after an osmosis and diffusion lab. We also share the development and validation of a three-point analytical rubric to score the keywords, student drawings, and reasoning prompts. The qualitative and quantitative nature of this exercise provides instructors with the unique opportunity to diagnose difficulties not only in introductory biology but throughout the undergraduate curriculum. Implementation strategies throughout the undergraduate curriculum are discussed.",,
-Numberjacks are on their way! A cultural historical reflection on contemporary society and the early childhood curriculum,S. Edwards,"This paper considers the temporal aspects of the early childhood curriculum from a cultural historical perspective, and in doing so focuses on the role of play in early childhood education. Drawing on ideas derived from cultural historical theory regarding the historical basis of community practices and knowledge, the paper reflects on the type of experiences that characterise playful activity for some of today's young children. Examples from previous research conducted by the author are provided as prompts for personal reflection on the temporal dimensions of cultural historical theory and the early childhood curriculum in order to understand the role of contemporary play experiences in children's learning.",,
-Comparing Elements of Study Abroad Among Sport Management Students,Carrie W. Lecrom,"A globalized curriculum has the potential to prepare students in a way that equips them for whatever sport looks like in the future. Study abroad programs are one way to achieve this. The current study looked at two short-term study abroad programs (one to western Europe, one to South Africa), offered during the same semester at the same institution, comparing learning outcomes between students on the two trips. Utilizing a mixed methods design, students completed quantitative pre/post surveys and responded to qualitative, open-ended daily prompts while on the trips. Findings indicate that knowledge acquisition occurs in both programs; however, students traveling on a sport-focused service-based trip to South Africa had a more transformational learning experience than those traveling on a sport-business-focused trip to western Europe.",,
-Effects of Formative Assessment Strategies on the Fractions Computation Skills of Students With Disabilities,B. Bottge,"Learning to compute with fractions is a major challenge for many students and especially for students with disabilities (SWD). Phase 1 of this study employed a randomized pretest–posttest comparison design to test the effects of two versions of formative assessment combined with an instructional program called Fractions at Work. In one condition, teachers used technology-assisted prompts to assess student performance and remediate errors. In the comparison condition, teachers gave students the same items for assessing progress but used their own methods of reteaching. Results indicated no difference between the two methods. However, pretest-to-posttest gain scores were significantly higher on all three measures regardless of type of formative assessment, and students maintained much of what they had learned. Phase 2 examined issues related to instructional dosage. Students who received additional weeks of instruction scored significantly higher than students who went back to their business-as-usual curriculum.",,
-“How Come Nobody Told Me?” Fostering Self‐Realization Through a High School English Curriculum,Laura T. Eisenman,"Through collaboratively designed qualitative inquiry, we investigated the responses of high school students with learning disabilities to a teacher’s intervention intended to promote self-realization, a fundamental component of self-determination. Activities were embedded within the general English curriculum and delivered in a special education classroom over the course of an academic year. Several themes emerged from analysis of student interviews, student responses to writing prompts and surveys, a teacher journal, and student portfolio pieces. Silence and misconceptions were prevalent in student experiences. However, through the intervention students acquired information that helped them make sense of their school experiences, redefine themselves in positive ways, and take small steps toward greater self-advocacy within their current school setting. The mediating influence of positive adult voices and concerns about social stigma were evident in students’ responses, which prompted us to question teachers’ and families’ responsibilities for engaging young people in dialogue about special education and disability.",,
-Students Exposed to Community Violence,Dorothy Rohde-Collins,"Exposure to community violence has a detrimental effect on the academic, social-emotional, and physical functioning of children, adolescents, and their families. Schools and educators are instrumental in enabling children and adults to process violence and trauma in order to develop a healthy worldview. Schools can counter the effects of community violence by emphasizing the development of sincere relationships between staff, students, and families; offering a warm, welcoming atmosphere; and providing a safe school environment conducive to learning. Educators can embed writing prompts and other opportunities for self-reflection within the academic curriculum as well as teach stress management, coping mechanisms, mindfulness, and relaxation techniques which can be used when a demanding situation arises. Educators and administrators are uniquely positioned to offer supports and opportunities within the school context to substantially improve the outcome for students, families, and communities experiencing community violence.",,
-"Debate, discourse and productive disagreement: interrogating the performative dimensions of authorship in the creative writing classroom",S. Holland-Batt,"ABSTRACT The creative writing curriculum has historically focussed on discipline-specific skills, developing students’ proficiency in literary forms, craft and techniques. However, contemporary writers are increasingly expected to participate in the economy of ideas through festival appearances, debates, and other forms of public speaking – skills that the creative writing curriculum has yet to fully contend with. We argue for the value of teaching debate as a distinct topic of inquiry for creative writing students, and hold that pedagogical innovation is required to address the changing nature of literary cultures and their increasing orientation towards performativity. This article establishes a new pedagogical model designed to introduce creative writing students to the study and practice of debate, comprising four key stages: modelling, scaffolding, debating and reflection. This learning progression not only fosters students’ oral argumentation skills, but also prompts critical reflection on the way key ideas in their field connect with their creative works. We contend that introducing debate into the creative writing curriculum addresses broader shifts in the writing and publishing industry, and that oral argumentation and debate should be considered key graduate competencies for creative writing students in the twenty-first century.",,https://research-repository.griffith.edu.au/bitstream/10072/419569/2/Jeffery4000400-Accepted.pdf
-"Nurturing Faculty‐Student Dialogue, Deep Learning and Creativity through Journal Writing Exercises",B. Korgel,"One of the most difficult endeavors in the engineering curriculum is encouraging students to develop creative, independent thought and a deep level of understanding. To achieve these results, we incorporate journal writing exercises into the curriculum of a core chemical engineering course, ChE 363 (Unit Operations II: Separations) at the University of Texas at Austin. Students receive weekly prompts to develop analogies or thought experiments to describe fundamental concepts prior to discussion in lecture. This motivates students to learn material independently and deeply as they are required to link concepts to common life experiences. The instructor identifies three to five of the most creative and accurate analogies and presents them to the class. The class votes for the “best” one, which earns the student extra credit. The journal writing exercises promote dialogue between the instructor and the students, generating in‐class discussions that often extend well beyond the scope of the journal prompt.",,
-Locating Teacher Voice in Curriculum Reform,R. Brooker,"Curriculum reform is a complex phenomenon, the outcomes of which may include changes to curriculum requirements, pedagogical practices (including assessment) and learning opportunities. The particular outcomes arising will be the product of influences exerted by multiple agents, representing a variety of voices, in a range of sites, and at particular points in time in a curriculum reform. The process of curriculum reform is, therefore, inherently complex and contested. This chapter prompts exploration of the process, specifically from the perspective of teachers and in relation to the notion of teacher voice in curriculum reform.",,
-The role of self-determination theory and cognitive evaluation theory in home education,Gina Riley,"Abstract This article explores the theories of Self-Determination, Cognitive Evaluation, and Intrinsic Motivation as it applies to home education. According to Self-Determination Theory, intrinsic motivation is innate. However, the maintenance and enhancement of intrinsic motivation depends upon the social and environmental conditions surrounding the individual. Deci and Ryan’s Cognitive Evaluation Theory specifically addresses the social and environmental factors that facilitate versus undermine intrinsic motivation and points to three significant psychological needs that must be present in the individual in order to foster self-motivation. These needs are competence, autonomy, and relatedness. Because of curriculum and time constraints, intrinsic motivation may be difficult to facilitate within the traditional classroom. This loss of intrinsic motivation for learning prompts some parents to homeschool their children. One of the most impressive strengths of home education lies in the fact that in many cases, the entire process revolves around a child’s intrinsic motivation to learn.",,
-Taking advantage of automated assessment of student‐constructed graphs in science,Jonathan M. Vitale,"We present a new system for automated scoring of graph construction items that address complex science concepts, feature qualitative prompts, and support a range of possible solutions. This system utilizes analysis of spatial features (e.g., slope of a line) to evaluate potential student ideas represented within graphs. Student ideas are then scored with rubrics based upon the knowledge integration framework (Linn & Eylon, 2011). We tested the effectiveness of this system on graphs constructed by 397 8th–12th grade students preceding, during, and following a curriculum focusing on graphs of motion. Comparison with human-coded responses indicates that the automated scoring system is very accurate (κ = 0.9). Also, ideas represented in constructions were generally similar to those demonstrated in written explanations; although individual students often shifted ideas between items. Learning gains were similar in both written and graph construction formats. Overall, these results suggest that graph construction is a valid and efficient means of evaluating students' complex ideas about data representation in science. We discuss the opportunities for incorporating graph construction into new science content areas, such as graphs representing density. We consider the implications of this system for generating automated, adaptive guidance to support instruction. © 2015 Wiley Periodicals, Inc. J Res Sci Teach 52: 1426–1450, 2015.",,
-Relating Chemistry to Healthcare and MORE: Implementation of MORE in a Survey Organic and Biochemistry Course for Prehealth Students,L. Schroeder,"We implemented a laboratory curriculum reform to teach foundational concepts in chemistry, particularly those concepts related to healthcare, in a chemistry course for prenursing students. Here, we discuss the reform, exploring how students built upon understandings gained in lab and correlating lab learning to course outcomes. We further discuss shifts in student work as they move through the course. As the course progressed, students became familiar with the pedagogy but also faced more challenging tasks. We present details on several of the laboratories that build the groundwork for understanding chemical principles, including the following: intermolecular forces, physical properties, acid–base chemistry, equilibrium, and chemical reactions. We further share our observations of student interactions around in-lab prompts and activities, and how these interactions inform our teaching. Our reform aims to improve critical thinking skills, namely, making and using models, observation skills, reasoning with ...",,
-Using Latent Semantic Analysis to Evaluate the Contributions of Students in AutoTutor,A. Graesser,"AutoTutor is a fully automated computer tutor that assists students in learning about hardware, operating systems, and the Internet in an introductory computer literacy course. AutoTutor presents questions and problems from a curriculum script, attempts to comprehend learner contributions that are entered by keyboard, formulates dialog moves that are sensitive to the learner's contributions (such as prompts, elaborations, corrections, and hints), and delivers the dialog moves with a talking head. Latent Semantic Analysis (LSA) is a major component of the mechanism that evaluates the quality of student contributions in the tutorial dialog. LSA's evaluations of college students' answers to deep reasoning questions are equivalent to the evaluations provided by intermediate experts of computer literacy, but not as high as more accomplished experts in computer science. LSA is capable of discriminating different classes of student ability (good, vague, erroneous or mute students) and in tracking the quality of contributions in tutorial dialog.","{'model': 'tldr@v2.0.0', 'text': 'Latent Semantic Analysis is a major component of the mechanism that evaluates the quality of student contributions in the tutorial dialog and is capable of discriminating different classes of student ability and in tracking thequality of contributions in tutorial dialog.'}",
-"""Virtual Assembly"" - A Web-Based Student Learning Tool for Thermodynamics Concepts Related to Multistaging in Compressors and Turbines",S. Chaturvedi,"Simulation and visualization have been used to develop ""virtual assembly"" as a student learning to ol for comprehension and reinforcement of concepts in basic engineering thermodynamics course in undergraduate engineering curriculum. Using a web-based module described in this study, students are able to assem ble on a computer screen a multistage compressor or turbine from a number of elemental building blocks or stages. The module is interactive and requires students to input data such as overall compressor ratio, stage effici ency, stage pressure ratio, and compressor inlet temperat ure and pressure. A computer program embedded in the module calculates total number of stages, and temperature and pressure at exit section of each st age. It also displays visual images of all stages to be ass embled, and prompts students to use the clicking and draggi ng action of computer mouse to assemble elemental compressor stages into a multistage axial compresso r. During the assembly process, a temperature-entropy diagram is generated, displaying thermodynamic state of air as it traverses through the compressor. These v isual images allow students to explore relationships betw een overall compressor efficiency, stage efficiency and compressor pressure ratio. The module is assessed b y comparing the performance of a ""control"" group (no exposure to module) with an ""experimental"" group (using the module) for an identical quiz administra ted to both groups. Results described in the paper show improvement in the average score for the ""experimental"" group over the ""control"" group.",,
-Washington State’s Classroom-Based Performance Assessments,AnnRené Joseph,"How does an entire state assess the arts in every school district at the elementary, middle, and high school levels with interrater validity and reliability? This chapter will summarize how Washington State’s Office of Superintendent of Public Instruction (OSPI) created, developed, designed, piloted, assessed, implemented, and reported arts classroom-based performance assessments (CBPAs), 2003–2016. The OSPI music CBPAs measure what educators value and teach, align with state arts learning standards regarding what all students should know and be able to do in music, and empower teachers to change their practice, resulting in practical significance. Developed by music educators, music CBPAs measure how individual students create, perform, and respond to real-life prompts. The CBPAs are formative and summative in design, and have become part of accountability instructional practices and state policy via annual district implementation verification reports. The vision for the initial CBPA design was: Curriculum + Instruction + Assessment for/ = Learning.",,
-The Transformation of Private Tutoring: Education in a Franchise Form,Janice Aurini," Various forms of private education are growing in Canada. This paper explores one instance of this change, private tutoring. Data from Ontario show that this massively growing industry is expanding its reach, as exemplified by the evolution from ""shadow education"" provision into ""learning centre"" franchises. Traditional shadow educators closely follow the school curriculum, offering short-term homework help and test prep. Learning centres develop their own curricular and assessment tools, offer comprehensive menus of services, and aim to nurture long-term skills. In so doing, these businesses are becoming increasingly ""school-like,"" moving beyond supplementary education towards a fuller alternative to public schooling.We link this evolution to the imperatives of the franchise form. Their larger investment prompts franchises to control their services via standardization, to secure stable revenues, and to seek new market niches. In conclusion we discuss the prospects for tutoring and other forms of private education.
Dans le secteur privé, différents modèles d'enseignement se développent actuellement au Canada. Cet article illustre un cas spécifique de cette évolution : la transformation du tutorat. Des données de l'Ontario démontrent que cette industrie, qui connaît un essor formidable, est en pleine expansion, tel qu'on peut le constater par la modification du tutorat, qui mèned'une forme d'éducation parallèle à des centres franchisés d'éducation. Les éducateurs traditionnels qui offrent des services parallèles suivent de près les programmes scolaires, offrant de l'aide à court terme pour les devoirs et pour les préparations aux examens, tandis que les centres d'éducation développent leurs propres outils d'apprentissage et d'évaluation, leurs programmes de services éducatifs polyvalents et visent aussi à entretenir leurs compétences à long terme. Ce faisant, ces entreprises deviennent de plus en plus « semblables à l'école », passant, au-delà de l'enseignement d'appoint, à une alternative plus sérieuse à l'école publique. Nous attribuons cette évolution aux impératifs entraînés par la formule des franchises. Les centres d'éducation, nécessitant de plus gros investissements, poussent les franchisésà standardiser leurs services, à s'assurer des revenus plus stables provenant de sources traditionnelles et à rechercher de nouveaux créneaux dans le domaine. Pour terminer, nous discutons des perspectives à court et à long termes du marché du tutorat et d'autres formes d'éducation privée.
",,
-Using a Concept Inventory to Reveal Student Thinking Associated with Common Misconceptions about Antibiotic Resistance,S. M. Bornstein-Forst,"Misconceptions, also known as alternate conceptions, about key concepts often hinder the ability of students to learn new knowledge. Concept inventories (CIs) are designed to assess students’ understanding of key concepts, especially those prone to misconceptions. Two-tiered CIs include prompts that ask students to explain the logic behind their answer choice. Such two-tiered CIs afford an opportunity for faculty to explore the student thinking behind the common misconceptions represented by their choice of a distractor. In this study, we specifically sought to probe the misconceptions that students hold prior to beginning an introductory microbiology course (i.e., preconceptions). Faculty-learning communities at two research-intensive universities used the validated Host-Pathogen Interaction Concept Inventory (HPI-CI) to reveal student preconceptions. Our method of deep analysis involved communal review and discussion of students’ explanations for their CI answer choice. This approach provided insight valuable for curriculum development. Here the process is illustrated using one question from the HPI-CI related to the important topic of antibiotic resistance. The frequencies with which students chose particular multiple-choice responses for this question were highly correlated between institutions, implying common underlying misconceptions. Examination of student explanations using our analysis approach, coupled with group discussions within and between institutions, revealed patterns in student thinking to the participating faculty. Similar application of a two-tiered concept inventory by general microbiology instructors, either individually or in groups, at other institutions will allow them to better understand student thinking related to key concepts in their curriculum.","{'model': 'tldr@v2.0.0', 'text': 'This study used the validated Host-Pathogen Interaction Concept Inventory (HPI-CI) to probe the misconceptions that students hold prior to beginning an introductory microbiology course, and revealed patterns in student thinking to the participating faculty.'}",https://www.asmscience.org/deliver/fulltext/jmbe/18/1/jmbe-18-8.pdf?itemId=/content/journal/jmbe/10.1128/jmbe.v18i1.1281&mimeType=pdf&isFastTrackArticle=
-Critical pedagogy for teaching HRM in the context of social change,S. Ruggunan,"This paper considers the imperatives of human resource management (HRM) studies in the context of contemporary South Africa. The authors draw on critical management studies (CMS) and the principles of emancipatory education to inform their argument for a critical and relevant HRM curriculum and associated teaching and learning approaches. The authors propose that the content and processes of HRM education must prepare students for critical participation in the contemporary South African society and workplace. The discussion outlines the rationale for the study, the specific prompts for its initiation, the theoretical framework of CMS, and Freire’s concept of emancipatory education.",,https://ajobe.journals.ac.za/pub/article/download/16/21
-Strategy training: An incidental learning model for CAI,S. Derry,,,
-The Impact of Intensive Writing Workshop Professional Development on a Cross-Curricular Group of Secondary Teachers,T. Locke,,,
-Thai students’ mental model of chemical bonding,Supawadee Sarawan,"This Research was finding the viewing about concept of chemical bonding is fundamental to subsequent learning of various other topics related to this concept in chemistry. Any conceptions about atomic structures that students have will be shown their further learning. The purpose of this study is to interviews conceptions held by high school chemistry students about metallic bonding and to reveal mental model of atomic structures show according to the educational level. With this aim, the questionnaire prepared making use of the literature and administered for analysis about mental model of chemical bonding. It was determined from the analysis of answers of questionnaire the 10th grade, 11th grade and 12th grade students. Finally, each was shown prompts in the form of focus cards derived from curriculum material that showed ways in which the bonding in specific metallic substances had been depicted. Students’ responses revealed that learners across all three levels prefer simple, realistic mental models f...",,
-Steps towards flipping classes in Higher Education (ESP),E. Păcurar,"The paper investigates the gradual transition from traditional methods of language instruction to delivering teaching material in a blended format and reports on some of the findings of a research grant dedicated to measuring the impact of flipped learning on students' development of productive skills. This transition reconsiders the learning/teaching stages and aims at both reducing teachers' lecturing time in class and increasing students' speaking time, so that more emphasis is given to productive oral activities and to fostering fluency. The process of reversing the learning and teaching stages prompts changes in the curriculum planning, in restructuring part of the materials available and reconsidering the role of both language instructor and student. These aspects are submitted under the concept of flipped learning and pinpoint the first steps involved in creating video materials for students in Cultural Tourism. The author reflects on choice of material (units, topics), curriculum changes (task-based and competence-based syllabi), comparing different learning environments and also on the advantages and challenges of implementing flipped methods. The benefits of transitioning from traditional to blended-learning methods stem from several aspects regarding the educational spectrum and its stakeholders. At the institutional level, it serves the language policy of the university and its long-term strategy of developing more blended teaching corpus and of encouraging multicultural and intercultural awareness through foreign language learning and teaching. At the European level, the focus on the proliferation of blended strategies, on Open Educational Resources and on facilitating the development of transversal and cross-sector skills are proof of the changes that need to be adopted by all European universities. At the local level, students' needs and their preparation for a mobile, ever-changing labor market acknowledge the role of language instruction in formal and informal interaction, inside and outside the classroom.","{'model': 'tldr@v2.0.0', 'text': ""The paper investigates the gradual transition from traditional methods of language instruction to delivering teaching material in a blended format and reports on some of the findings of a research grant dedicated to measuring the impact of flipped learning on students' development of productive skills.""}",https://easychair.org/publications/open/FnSp
-The JSEP Learning Skills Training System,S. Derry,"The Army Job Skills Education Program (JSEP) is a dual curriculum that reminds soldiers to use learning skills while they take CAI lessons that train basic math and verbal job competencies. Prior to entering the JSEP basic skills curriculum, soldiers are taught self-motivational techniques and strategies for studying and solving problems. Each type of learning skill is represented and trained by a different character model. These characters appear as prompts throughout the basic skills curriculum to remind soldiers to recall and use their new learning skills. In this article, the rationale for the curriculum model is overviewed and preliminary field trial data are reported.",,
-"Preparing School Systems To Deliver a Hybrid Education Program for Students with Autism Via Distance Learning Classrooms, In-Class Teleconferencing, and Listserv Technology.",J. C. Burke,"The Center for Technology in Education (CTE), a partnership between Johns Hopkins University and the Maryland state department of education, developed a model for teaching autistic students that integrates technology-based instructional and behavioral supports into existing curriculum. The goal is to promote student gains in communication, social skills, and academics, ab well as to increase students' overall involvement in classroom activities. The model recommends teacher training in providing clear instructions, prompts and prompt fading, positive reinforcement, behavior modification techniques, and advanced teaching principles such as targeting student responsiveness and providing students with opportunities for appropriately communicating their needs. The model also advocates the use of light or high technology based on student needs and demands of the learning task. Light and high technology include equipment such as voice output communication devices, picture systems, sequence and cue cards, touch screens, modified keyboards, switch access devices, and sign language. Teachers also need to target behaviors related to motivation, cooperation, and responsiveness; learning strategies and methods that help students learn more efficiently; and skills that help all students function in the inclusive educational setting. Equally important is the development of an individualized continuity plan that promotes longitudinal gains and transitions across teams of staff members within a school or across schools, and the use of school-wide staff training to promote methods of inclusion. Currently, CTE is exploring alternative methods for disseminating technical assistance and teacher training to local education agencies throughout Maryland using distance learning classrooms, in-class teleconferencing, and listserv technology. (LP) PERMISSION TO REPRODUCE AND DISSEMINATE THIS MATERIAL HAS BEEN GRANTED BY DIANE MONTGONE-RY ( PARENT) TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) U.S. DEPARTMENT OE EDUCATRDN Orlice of Educationat Research and Improvement EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) 1:104ns document has been reProduClid received from the persOn or Prgendation onginating C Mmot changes have been made to improve reproduction Quality Points of vie* ce opinions Slated in this docu rnenl do not necessanty represent official OE PI position or policy John C. Burke, Ph.D. Center for Technology in Education Johns Hopkins University Baltimore, Maryland PREPARING SCHOOL SYSTEMS TO DELIVER A HYBRID EDUCATION PROGRAM FOR STUDENTS WITH AUTISM VIA DISTANCE LEARNING CLASSROOMS, IN-CLASS TELECONFERENCING, AND LISTSERV TECHNOLOGY One of the most difficult challenges teachers and local educational agencies face today is educating students with pervasive developmental disorders such as autism. Teachers frequently describe these children as unique, puzzling, mysterious, and difficult to teach. They often comment that their students with autism rarely interact, communicate, or appear motivated to participate in academic or social activities. A committed teacher provided the following comment before receiving training: ""He came into my room the first day and cried and showed absolutely no interest in anything; I was at a loss."" Given the widespread behavioral deficits characteristic of autism and the limited resources available, teachers repeatedly face an enormous task as they strive to help these children learn. Aside from posing a challenge to teachers, LEAs are recurrently in a quandary as they form policy and develop curricula. Unfortunately, LEAs typically commit their resources to a single approach, then quickly discover that one approach does not easily serve all of their students with autism. The Center for Technology in Education (CTE), a partnership of Johns Hopkins University and the Maryland State Department of Education, is developing a hybrid paradigm in w11:.ch technology-based instructional and behavioral supports are infused into existing curricular activities to help promote gains in communication, social skills, and academics, as well as to increase the children's overall involvement in classroom activities. In addition, CTE is exploring alternative methods tbr disseminating technical assistance to LEAs throughout the state of Maryland using Distance Learning Classrooms, In-Class Teleconferencing, and LISTSERV Technology. Technology-Based Hybrid Educational Model for Childrot With Autism This hybrid paradigm is based on interdisciplinary research and years of educational experience and has four central components: Effective Behavioral Teaching Principles Curriculum-Wide Integration of Technology Flexible and Longitudinal Curriculum Guide Systemic Change Activities 112 BEST OR PAILMILE Effective Behavioral Teaching Principles A set of fundamental teaching principles, based on 30 years of research, now exists and forms the basis for many more advanced teaching strategies. The use of these principles is a central part of the approach. While many professionals are familiar with these principles, it is advantageous to delineate them clearly and to stress their importance in working with students with pervasive developmental disorders such as autism. All educational staff members should be trained in the use of these principles. It is equally vital that significant others, such as parents, learn to use these principles in less structured settings. A brief overview of these fundamental teaching principles is presented below. I. Effective instructions/requests. For all children, but especially for students with autism, instructions should be brief and clear, should be phrased in the form of a request, should be presented when you have the child's attention or should be used to get the child to attend to you, and should be individualized for each child. 2. Prompts and prompt fading. Prompts can help the child to respond and learn new skills. They can be used to help guide the child to respond. In using prompts, it is important to use the least intrusive, yet effective, prompt. Prompts can be verbal, gestural, and physical. When physical prompts are used, it is advantageous not to overuse ""hand-over-hand"" or other full physical ""holding"" prompts. While eftective, these are very difficult to fade. 3. Consequences. Consequences, in the form of positive reinforcement, should involve presenting highly preferred items or other events contingent on the display of the appropriate target behavior. It is vital that the items be things that the child clearly wants and finds reinforcing. In general, consequences should come soon after the child exhibits the target behavior. In addition, delivering positive consequences following a sequence of both disruptive and cooperative behaviors might inadvertently produce an increase in the disruptive behaviors. If a child first shows disruptive behaviors, then displays appropriate behaviors and is subsequently reinforced, he may learn that he needs to exhibit both to get reinforcement. It is usually advisable not to reinforce these chains of behaviors. In recent years, increased emphasis has been placed on integrating natural/direct reinforcers throughout a child's dav. Natural/direct reinforcers are consequences that are directly related to the activity. For instance, if a child uses a picture exchange system to communicate that the child wants a break, giving the child a break would be a natural reinforcer. Placing emphasis on using natural/direct reinforcers, rather than relying only on artificial reinforcers often leads to increased performance and cooperation. 4. Chainiag, Chaining involves breaking down a complex task into steps, establishing criteria for success for each step, and teaching the child to perform each step satisfactorily before adding the next step. S. Shaping. Shaping involves reinforcing closer approximations of the target response. For instance, if an essentially nonverbal child says ""C"" for Cup, the teacher would reinforce the child. If the child subsequently says ""Cu"" he would be reinforced. If he then said ""C"". no reinforcement would be provided. Only better approximations are reinforced. In addition to the fundamental teaching procedures described above, during the course of a training project, training would be extended to cover more advanced principles of effective teaching. Depending on the needs of the educational staff, these advanced procedures could include the following: 1. Use of the principle of partial participation. The principle of partial participation involves the idea of setting the occasion for the student to partially participate in an activity or task and arranging for the student to succeed. Over time, the student will learn more and more and will eventually learn to perform the entire task in a more independent manner. Using this principle throughout a child's day helps to ensure low levels of disruptive behavior. In a consistent manner, it is otlen beneficial to reinforce approximations and attempts. This concept centers on the view that a child who is truly trying to correctly respond should be provided with positive feedback relative to the response. This principle could be used during academic and leisure time periods. 2. Targeting cooperation and responsiveness. Cooperation forms the basis of effective teaching interactions between teachers and students. Cooperation is often increased as students are taught new academic and social skills. It is oflen advantageous, however, to directly target cooperation and responsiveness in a systematic, yet natural, manner throughout a child's day across settings and activities, and with people. Many children are more focused, responsive, and cooperative when relatively greater emphasis is placed on: Presenting the child a relatively short and simple request or instruction. Ensuring that the child is oriented to the adult when the instruction is presented. Using an appropriate prompting st",,
-High Quality Professional Development: An Investigation of the Supports for and Barriers to Professional Development in Arts Education.,Vicki R. Lind,"This study focused on a model of professional development designed to support and encourage arts educators to increase their understanding of student learning in the arts, broaden their knowledge of the Visual and Performing Arts Standards, build upon their repertoire of teaching methods and assessment strategies, and improve leadership skills. Data included 300 hours of observation, focus group and individual interviews, written responses to reflective prompts, unit plans, video and audio tapes, and samples of student work collected over a two year period. Findings indicated that working collaboratively, focusing on student learning, and identifying and planning curriculum around issues central to the discipline positively impacted teachers work. The issue of time constraints was consistently identified as a barrier to professional growth.",,
-Linking Schools of Thought to Schools of Practice,Lucy Hunt,"Project Linking Learning (“Link”) was created to target the needs of gifted students in urban school districts with historically underserved populations. Project Linking Learning implemented a linking curriculum between in-class instruction and an afterschool enrichment program for selected students in second through fifth grade. Designed by Dr. Sandra Kaplan as a collaborative endeavor between the University of Southern California, Rossier School of Education, and the Los Angeles Unified School District (LAUSD), this scale-up grant (Jacob J. Javits Grant #5206A090045) targeted elementary schools in diverse, urban neighborhoods with consistently low rates of gifted referral, identification, and program implementation. Intensive training and support was provided to Link teachers to allow them to effectively roll out this novel curriculum, including foundational training on differentiated instruction (e.g., prompts of depth and complexity, research skills, thinking skills, learning centers, independent study, etc.), as well as information on the characteristics of gifted learners and able underachievers from diverse backgrounds. Schools participating in Project Linking Learning experienced a significant increase in the rate of gifted identification over a four-year period of time. Key program concepts and alignment to educational schools of thought are described, as well as takeaways and recommendations for district adaptation and implementation.",,
-Integrating Physical and Virtual Models in Biology: A Study of Students' Reasoning While Solving a Design Challenge,N. Martin,"Using models to explain phenomena is important in science. Virtual and physical models have different affordances that can be integrated to foster students’ learning. Integrating evidence from multiple models to justify explanations is challenging, and we know little about how students coordinate such information, especially in biology. This study investigated how students’ integrated information from virtual and physical models in a design-based, biology curriculum. Some students used information from virtual simulations in written explanations of changes they would make to their physical models. However, onethird of students did not use the virtual model to justify their revisions, despite prompts from instructional materials, the teacher, and other group members. Even some students who integrated these different models did not initially do this without support from external prompts. This study provided deeper understanding of how students integrated physical and virtual models, which can help identify the kinds of support students may need.","{'model': 'tldr@v2.0.0', 'text': 'Deeper understanding of how students integrated physical and virtual models in a design-based, biology curriculum is provided, which can help identify the kinds of support students may need.'}",
-Pre-service elementary school teachers' metaphors on mathematics textbooks,J. Kim,"The purpose of this study was to investigate the nature of pre-service elementary teachers` metaphors on mathematics textbooks. Their metaphors describe individual and collective patterns of thinking and action on mathematics teaching and learning. To analyze their metaphors, qualitative analysis method based on Lakoff and Johnson`s theory of metaphor (1980) was adopted. Metaphors on mathematics textbooks were elicited from 161 pre-service elementary school teachers through writing prompts. The writing prompt responses revealed three types and thirteen categories: As Type I, there were (1) `Principles`, (2) `Summary`, (3) `Manual`, (4) `Encyclopedia`, (5) `Code`, (6) `Guidelines`, and (7) `Example`. As TypeII, there were (9) `Assistant`, (10) `Friend`, (11) `Scale`, and (12) `Ongoing`. As TypeIII, there was (13) `Trap`. Among these categories, `Guidelines`, `Assistant`, and `Ongoing` were the most frequently revealed. These results indicate that the relations of mathematics curriculum, textbooks, and classrooms are not a unilateral way but should communicate with each other.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the relations of mathematics curriculum, textbooks, and classrooms are not a unilateral way but should communicate with each other.'}",
-Critical and Reflective Thinking in an Intermediate Financial Accounting Course: An Action Research Study.,J. Cobb,"Accounting professionals have consistently called for educators to develop curriculum designed to encourage students to develop intellectual skills. The purpose of this action research study was to develop and implement an instructional method that requires intermediate financial accounting (IFA) students to consistently practice higher order thinking. Students completed a semester-long authentic comprehensive project (CP) that required them to consistently practice critical and reflective thinking (Facione, 1990; Mezirow, 1991). Findings led to the resolution of implementation issues associated with using a CP. Most implementation issues originated from students’ inability to apply learning obtained from working structured problems to unstructured accounting work. Short reflection papers (RPs) replaced periodic objective tests to encourage deep and meaningful learning. Students’ responses to question prompts gave evidence of one of Mezirow’s (1991) four stages of professional reflection. The depth of reflection trended with students’ understanding of when and how accountants use judgment. Students who consistently practiced higher order thinking also learned to adequately perform routine accounting procedures. This study resulted in an instructional method that requires accounting students to practice using the intellectual skills necessary for success in the accounting profession without sacrificing procedural knowledge. The findings will benefit other instructors working to develop learning materials that require students to practice higher order thinking as they complete authentic professional work.",,
-Finding Out What They Really Think: Assessing Non-Science Majors’ Views of the Nature of Science,B. Beck-Winchatz,"As institutions of higher learning are increasingly held accountable for student outcomes, faculty are faced with the challenge to clearly articulate and assess what students should learn in their courses. We report on the assessment of a liberal studies learning outcome related to the nature of science, which involved 178 students from 41 academic non-science majors at DePaul University in Chicago. Students were shown a video recording of an interview with a research scientist and asked to respond to prompts about falsifiable predictions, uncertainty, and the functions of hypotheses, observations, and reasoning. We found that students held a wide range of views and misconceptions about the nature of science. Responses were, on average, at a “developing” level (a score of 2 on a scale from 1–4). We discuss several possible changes to curriculum and pedagogy to address these misconceptions based on a review of the science education research literature.",,
-Hybrid teaching mode for laboratory-based remote education of computer structure course,Wan Han,"We describe an Open edX-based blended course developed for a reformed Computer Structure course at Beihang University. In three iterations of this laboratory-based course, we dive into key issues that impact students' learning, and then redesign our curriculum, which integrated with virtual laboratory technique into the MOOC platform. We show how certain course design aspects affect students' learning in the hybrid teaching mode: (a) Kung Fu style competency education with online-support laboratory system prompts students to own their learning as the pace and/or the path of learning, which is dictated by mastery instead of the time/space; (b) strengthen the use of learning-aid tools empowered teachers with the skills and information to define standard for each learner in each stage; (c) automated test technology make this blended learning possible at scale and also financially sustainable; (d) using discussion forum to build the lesson about `what to do' when learners get stuck helped in overcoming challenges.","{'model': 'tldr@v2.0.0', 'text': ""This work describes an Open edX-based blended course developed for a reformed Computer Structure course at Beihang University, and shows how certain course design aspects affect students' learning in the hybrid teaching mode.""}",
-Student perceptions of gamified audience response system interactions in large group lectures and via lecture capture technology,Robin K Pettit,,,https://bmcmededuc.biomedcentral.com/counter/pdf/10.1186/s12909-015-0373-7
-Interaction - Based Intervention Programs in Multiple Disabilities,Andrea Hathazi,"Competences of practitioners working with children with multiple disabilities refer mainly to assessment, curriculum planning and intervention, but these are specifically related to the learning process and the needs of the child in development. Since the child- centered approach is a key factor in the success of intervention, more and more theoreticians and practitioners take into consideration the interaction approach so that the teachers become more reflective on their own behavior, language and prompts and also what they determine in a child’s behavior, motivation and communication. The present paper focuses on the modalities in which practitioners become aware of the consequences of their own abilities in intervention and they implement the most adequate strategy and method for that specific context, time and child’s needs. The intervention should be dynamic, managed by both partners involved, observing each other’s participation, paying attention to emotional state and well- being.",,
-Strategies for planning and designing medical curricula and clinical teaching,M. Barrow,"Student learning is an active and constructive process. The role of a teacher is to provide an environment in which students are able to actively engage with subject matter in order to learn it. This article examines the principal features of good curriculum, course and lesson design and discusses ways in which doctors, in their roles as teachers of medical students and medical trainees can ensure that their teaching prompts learners’ engagement in the learning process.","{'model': 'tldr@v2.0.0', 'text': ""The principal features of good curriculum, course and lesson design are examined and ways in which doctors, in their roles as teachers of medical students and medical trainees can ensure that their teaching prompts learners' engagement in the learning process are discussed.""}",http://seajme.sljol.info/articles/10.4038/seajme.v4i1.432/galley/579/download/
-Dance as a way of knowing,Jennifer Donohue Zakkai,"Jennifer Donohue Zakkai illuminates why and how dance is a powerful tool for learning and creativity in K-6 classrooms. In this ""workshop on paper"" you'll discover how to build on what you already know about movement, and implement techniques to tap into children's natural appetite for moving. Zakkai addresses the challenges of engaging students in full-bodied motion in the classroom. She focuses first on helping students become responsible movers in the space, then offers structured learning experiences that demand a high level of concentration and creativity. Through verbal prompts--not demonstrations--that involve students in creative problem-solving, students learn through discovery instead of imitation. You don't have to be a dancer to use the detailed model lessons that guide you through warm-ups, movement explorations, rich curricular integrations, culminating activities, observation and reflection. To help your students understand curriculum content, express themselves in creative ways, and learn about dance as an art form, Dance as a Way of Knowing offers a progression of strategies that will make this vibrant discipline come alive in your classroom.",,
-Benchmarks Curricular Planning and Assessment Framework: Utilizing Standards Without Introducing Standardization,Erika Feldman,,"{'model': 'tldr@v2.0.0', 'text': 'An approach to curricular planning and assessment that uses state benchmarks as an umbrella structure to support curricular plans, assessment, and feedback among them is introduced.'}",
-Severe Weather Field Experience: An Undergraduate Field Course on Career Enhancement and Severe Convective Storms,C. Godfrey,"Abstract Undergraduate students acquire a deeper understanding of scientific principles through first-hand experience. To enhance the learning environment for atmospheric science majors, the University of North Carolina at Asheville has developed the severe weather field experience. Participants travel to Tornado Alley in the Great Plains to forecast and observe convective storms for two weeks. The objectives of the course encompass far more than observing severe storms. On days with non-threatening weather in the Great Plains, students participate in an array of activities that provide exposure to facilities and interaction with professionals in various sectors of meteorology. While the allure of chasing storms initially prompts the students to enroll in the course, the focused career-development aspect of the curriculum increases awareness for the varied career options in the atmospheric sciences and helps students discover where their own capabilities and interests might best suit the discipline. The course thus offers students a comprehensive career-development experience woven within a thrilling adventure.",,
-Teachers' planning processes: seeking insights from Australian teachers | NOVA. The University of Newcastle's Digital Repository,P. Sullivan,"The Australian Curriculum: Mathematics (ACM) prompts consideration of planning processes teachers typically use (as a whole school, in grade levels, and at the classroom level). In order to gain insights into the nature of the planning decisions teachers make, Australian teachers drawn from every state and territory were invited to respond to a survey on planning processes, during professional learning programs or voluntarily online through the Australian Association of Mathematics Teachers' website. In this article, we provide background to the overall project in which this survey was situated, and describe the process through which the survey was developed. Subsequent articles in this journal detail some important findings from the survey data.",,
-A dual purpose data base for research and diagnostic assessment of student writing,J. Parr,"The data base of writing examined serves a dual purpose. Here it is used as a research tool and the writing performance from the large, nationally representative sample (N = 20,947) of students (years 4 to 12) interrogated to examine patterns of performance in writing. However, the data base was designed to underpin a software tool for diagnostic assessment of writing. Viewing writing as accomplishing social communicative goals, performance was considered in terms of seven main purposes the writer may seek to achieve. Tasks related to each purpose were encapsulated in 60 writing prompts that included stimulus material. Participants produced one writing sample; the design ensured appropriate representation across writing purposes. Samples were scored using criteria differentiated according to purpose and curriculum level of schooling and acceptable reliability obtained. Analyses indicate that growth was most marked between years 8 and 10, arguably, as opportunity to write increases and writing is linked to learning in content areas. Variability in performance is relatively low at primary school and high at secondary school. Students at any level did not write equally well for different purposes. Mean scores across purposes at primary school were relatively similar with to instruct and to explain highest. By years 11-12 there is a considerable gap between the highest scores (for narrate and report) and the lowest, recount, reflecting likely opportunities to practice writing for different purposes. Although girls performed better than boys, the difference in mean scores narrows by years 11-12.","{'model': 'tldr@v2.0.0', 'text': 'The writing performance from the large, nationally representative sample of students is interrogated to examine patterns of performance in writing and indicates that growth was most marked between years 8 and 10, arguably, as opportunity to write increases and writing is linked to learning in content areas.'}",https://www.jowr.org/index.php/jowr/article/download/738/781
-Impact of Faculty Development on Physical Therapy Professors' Beliefs.,L. Behar-Horenstein,"Physical therapy faculty share similarities with faculty across allied health fields, such as nursing, and other clinical disciplines that educate students in licensing and board certification programs. Most have clinical experience and discipline-based expertise, however they may not have had the benefit of continuous learning aimed at enhancing their teaching repertoires. Because of the rapid influx of clinicians into the academy, faculty development is considered essential to their integration. The purpose of this study was to describe how faculty development impacted physical therapy professors' understanding and use of new instructional skills. Eight physical therapists from a university located in the Southeast U.S. participated in a six-week, 12-hour teaching seminar focusing on curriculum and teaching where participants kept reflective journals to record their responses to question prompts. Basic unitizing, coding, and categorizing were used to conduct a multi-stage analytical process. Eight themes emerged including assessment, instructional strategies, teaching styles, and individuals' planned changes to their classroom practice, among others. Findings showed that professional development is essential for enriching faculty instructional capacity to promote student learning and patient care. Investing in the professional development of faculty may help ensure quality teaching so that professors become conduits to knowledge production. Faculty development is ""the continuous learning that professionals may need to pursue throughout their careers in order to maintain, enhance, and broaden their professional competence"" (Gottlieb, Rogers & Rainey, 2002, p. 280). Studies in physical therapy, and other health professions, have shown that the faculty development process is central to effective teaching and the preparation of future healthcare practitioners and professional educators (Behar-Horenstein, Childs, & Graff, 2008, 2009, 2010; Farmer, 2004; Mahara & Jones, 2005, Steinert et al., 2006). Keeping physical therapists continuously informed about new knowledge, skills and technology is essential to their capacity as instructors. Historically physical therapy (FT) programs have been disadvantaged. They have been forced to hire an overwhelming number of clinicians as faculty rather than individuals who have been trained in higher educational instruction and assessment (Harrison, «Sc Kelly, 1996). Most of the clinical instructors, while quite skillful, often lack teaching abilities (Gottlieb, Rogers «Sc Rainey, 2002). Hiring clinicians resulted because universities were in the initial stages of developing physical therapy degree programs. Such an action highlights the importance of developing a mechanism to continuously determine faculty effectiveness and productivity. Faculty development can help new faculty examine their own beliefs about teaching and consider how they might apply their thinking. New faculty from clinical disciplines are often different from traditional academicians who earn several degrees and typically experience graduate student teaching and research roles aimed at faculty-type expertise. The American Physical Therapy Association's (APTA) former director of professional development, Marilyn Phillips, encourages the use of faculty development. This process can be a vehicle to create one's own plan or ""blueprint for career development"" (Starcke, 2005, p. 42). APTA' s Board of Directors acknowledges varied modes of faculty development including where and how it can occur. "" [It] may occur in formal instructional settings or in natural societal settings and may include . . . academic courses of study, organized continuing education, independent study, and self- and external assessment"" (p. 42). Although some individuals benefit from structured activities, others may work on their own plan of professional growth. However, the APTA directors emphasize the role of assessment in professional development stating, ""All professional development experiences should be planned and assessed"" (p. …",,
-Teachers have the power to alleviate math anxiety,D. J. Shields,"Abstract Many students suffer from math anxiety, but teachers can incorporate skills to alleviate it for their students. In order to alleviate math anxiety teachers first and foremost need to portray a positive, enthusiastic, helpful attitude, which communicates a love and usefulness for mathematics. Math curriculum should be designed that deepens student understanding in a practical, engaging and fun way. Pedagogy should be implemented that moves from the explain-practice-memorize strategy to strategies that stress reasoning and understanding. A classroom culture that prompts student inquisition, discovery, learning and the exploration of ideas needs to replace the structured, rigid atmosphere usually associated with math class. And finally, assessment should be conducted in a variety of ways. Introduction Not only do students today need math remediation at an overwhelming rate, but reports state that two thirds of American adults fear and loathe math (Burns, 1998). One reason such negative experiences are reported is math anxiety, which knows no boundaries regarding race, age or gender. Math anxiety can be perpetuated in the home, in society, and in the classroom. Because most students can name the teacher they believed caused their anxiety, the exact moment when it happened and the event that triggered the anxiety in the classroom, an examination of teacher attitudes, curriculum, pedagogy, the classroom culture, and assessment is crucial to understanding and alleviating the problem. Students at all grade levels can experience math anxiety and they attest to similar characteristics and consequences. One of the most notable consequences of math anxiety is poor math achievement and poor math grades. Part of this is because students with math anxiety attempt to cope with the debilitating effects of their anxiety by avoiding math in school. This in turn can lead to limited college majors, and ultimately career choices that are restricted mathematically and monetarily. This limited exposure to mathematics is also responsible for lower math achievement and competence (Ashcraft, 2002). The anxiety, which is not significantly related to intelligence, has been shown to inhibit student learning (Ashcraft, 2002) and to reduce working memory capacity (Ashcraft & Kirk, 2001; Beilock,& Cart, 2005; Shobe, Brewin & Carmack, 2005), which in turn has a major impact on students"" serf-confidence related to mathematics. Math anxiety is often the result of repeated negative experiences related to mathematics (Kogelman, Nigro, & Warren, 1978). It is a conditioned fear that develops into a fatalistic attitude, which becomes a self-fulfilling prophecy that reinforces one's beliefs about an inability to perform mathematically. For those who suffer from math anxiety physiological symptoms such as sweaty palms, nausea, muscle contractions, difficulty breathing, tightness in the throat and chest, headaches, heart palpitations, restless behavior, forgetfulness, and a temporary boost in one's heart rate are familiar. Teachers' attitudes, curriculum, pedagogy, the classroom culture, and assessment greatly impact students. All these areas must be considered in order to alleviate anxiety in the classroom. In a 1999 study by Jackson and Leffingwell, it was discovered that sixteen percent of the students studied were first traumatized mathematically in third or fourth grade. The difficulties in fourth grade arose from new math ideas such as fractions, timed tests, and memorization of multiplication tables and formulas. Teachers at this level were blamed for hostile behavior, making derogatory comments when children did not understand concepts, appearing angry when asked for additional help, displaying insensitive and uncaring attitudes, stereotyping females as not needing math, ridiculing girls more overtly than boys, and favoring boys in the subject. Approximately twenty-six percent of the students in the study first realized feelings of math anxiety in ninth, tenth and eleventh grade. …",,
-Music assessment and the hegemony of musical heritage,G. Spruce,"About the book: Issues in Music Teaching stimulates critical reflection on a range of topics related to the teaching and learning of music in both the primary and secondary school, including: • the place of music in the curriculum • the nature of music and music education • ICT and music education • music education and individual needs • continuity and progression in music education The book prompts the reader to be analytical and critical of theory and practice, and to become an autonomous professional and curriculum developer.",,
-Introducing In-Service Teachers to Virtual Schooling through the Lens of the Three Teacher Roles,M. Barbour,"This study will examine the third and fourth rounds of data collection from an action research project designed to help in-service teachers become better virtual school facilitators (currently being analyzed). The data included blog entries and comments from five of the seven graduate students in an instructional technology course related to K-12 online learning. The specific discussion prompts relate to virtual school readings and the Teacher Education Goes into Virtual Schooling (TEGIVS) curriculum. Based upon initial analysis, the TEGIVS curriculum was effective for providing these graduate students some experience with how K-12 online learning opportunities were delivered, along with some of the possibilities and challenges associated with K-12 online learning. The analysis of this data is continuing, and there are plans to continue this line of inquiry with additional students in future offerings as we continue to improve upon the course design. Clark (2001) estimated there were approximately 40,000 to 50,000 students engaged in virtual schooling in the United States. Eight years later Picciano and Seaman (2009) indicated that there were over 1,000,000 K-12 students enrolled in online courses. Watson, Gemin, and Ryan (2008) reported significant K-12 online learning activity in 44 states. Several states, including Michigan, require students to have an online learning experience in order to graduate from high school. This growth of online learning at the K-12 level has caused dramatic changes in the design and delivery of education, at least for the student population served by these opportunities. However, the vast majority of pre-service and in-service teacher education programs simply do not provide any curricular opportunities in this area. Rice and Dawley (2007) found that less than 40% of all online K-12 teachers in the United States reported receiving professional development before they began teaching online. This indicates a need for teacher education programs to address pre-service and in-service teachers’ ability to teach in environments that are completely mediated by technology. This study examined one university’s continuing efforts to address this growing need in teacher education. Wayne State University is a large, publicly funded, research-extensive university in Michigan (i.e., the first jurisdiction to have an online learning graduation requirement). Over the past two years, the Instructional Technology program has developed and adopted curriculum to support teachers with understanding of K-12 online learning in their own schools. This proceedings examines the results of the second round of data collection.",,
-Integrating Narrative Reading Comprehension and Writing Instruction for All Learners.,Deborah C. Simmons,"A study examined the effects of an integrated reading/writing curriculum on the narrative writing of students in general education eighth-grade classrooms. Subjects, 93 Ftudents (including 10 with learning disabilities) from middle to w.:7er-middle socioeconomic backgrounds at a suburban middle school in wertern Oregon, were administered narrative writing probes preand post-intervention tc assess students' ability to plan, organize, and write stories for topic prompts. Students were enrolled in four sections taught by tuo teachers; students from one teacher's classes served as the experimental group and the other teacher's classes participated in the control condition. Instructional materials included 10 short stories. The integrated reading and writing curriculum was designed in three interdependent phases: learning narrative text structure; learning a writing process; and learning to generate stories. Students in the control condition received instruction and practice on narrative text comprehension. Results indicated that students in the experimental group significantly outperformed students in the control group--their stories contained more fully developed ideas, content, settings, characters, and attempts to solve the central problem than students in the control classrooms. Results also indicated that all students in the integrated condition benefitted from the curriculum, although students continued to have difficulty generating well-developed stories. Findings suggest the potential value of investing in curriculum development that equips learners with transferrable and maintainable knowledge. (Two tables of data are included; 22 references are attached.) (RS) Integrating Narrative Reading Comprehension and Writing Instruction for All Learners Deborah C. Simmons Edward J. Kameenui Shirley Dickson David Chard Barbara Gunn","{'model': 'tldr@v2.0.0', 'text': 'The effects of an integrated reading/writing curriculum on the narrative writing of students in general education eighth-grade classrooms and the potential value of investing in curriculum development that equips learners with transferrable and maintainable knowledge are suggested.'}",
-Archiving Student Solutions with Tablet PCs in a Discussion‐based Introductory Physics Class,E. Price,"Many active learning based physics courses use whiteboards as a space for groups to respond to prompts based on short lab activities, problem solving, or inquiry‐oriented activities. Whiteboards are volatile; once erased, the material is lost. Tablet PCs and software such as Ubiquitous Presenter can be used as digital whiteboards in active learning classes. This enables automatic capture and archiving of student work for online review by students, instructors, and researchers. We studied the use of digital whiteboards in an active‐learning introductory physics course at California State University, San Marcos. In this paper we examine the archival features of digital whiteboards’, and characterize the use of these features by students and instructors, and explore possible uses for researchers and curriculum developers.","{'model': 'tldr@v2.0.0', 'text': 'The archival features of digital whiteboards are examined, and the use of these features by students and instructors are characterized, and possible uses for researchers and curriculum developers are explored.'}",http://www.compadre.org/per/perc/2008/proceedings/UP205perc08_revised.pdf
-Developing an IT Project Management Course to Meet Changing Industry Needs,T. Jewels,"ABSTRACT Acknowledging that an awareness of project management skills was an important outcome for their graduates, the IT faculty of one Australian university developed and introduced a new IT Project Management course into their undergraduate curriculum in 2002. A three stage approach that involved identifying students expected learning outcomes, selecting relevant content and choosing the most applicable pedagogy was applied in designing this course. An integrated case study approach, developed over a five year period which was to be the recipient of many teaching awards is described. However, despite its popularity and successes the course was totally rewritten for the 2008 academic year, reflecting alternative views about what IT Project Management courses should contain, which in turn prompts the question of what project management skills should we be teaching students in preparing them for professional IT careers? Keywords Case method, course design, project management, teaching.","{'model': 'tldr@v2.0.0', 'text': 'The course was totally rewritten for the 2008 academic year, reflecting alternative views about what IT Project Management courses should contain, which prompts the question of what project management skills should be teaching students in preparing them for professional IT careers.'}",
-Course approval processes,M. Horsburgh,"Focuses on processes of course approval, which lead to enhancement of the curriculum and student learning. Findings from a recent study concerned with the question: To what extent does quality monitoring impact on the student experience of learning? showed the curriculum, and how it was approved, to be significant factors in enhancing student learning. It is argued that the focus for quality in a rapidly changing world should be on the attributes of graduates, with quality monitoring concerned with improvement and enhancement of student learning. Where quality monitoring does have a positive impact on student learning, through the approval and ongoing monitoring of the curriculum and its outcomes, then these processes should be improved. This should be through a process which fosters collegiality and encourages pedagogical discussion amongst academic colleagues. A series of prompts or questions, which seek to challenge teachers and enhance dialogue with colleagues, is proposed.",,
-On Curriculum Learning for Commonsense Reasoning,A. Maharana,"Commonsense reasoning tasks follow a standard paradigm of finetuning pretrained language models on the target task data, where samples are introduced to the model in a random order during training. However, recent research suggests that data order can have a significant impact on the performance of finetuned models for natural language understanding. Hence, we examine the effect of a human-like easy-to-difficult curriculum during finetuning of language models for commonsense reasoning tasks. We use paced curriculum learning to rank data and sample training mini-batches with increasing levels of difficulty from the ranked dataset during finetuning. Further, we investigate the effect of an adaptive curriculum, i.e., the data ranking is dynamically updated during training based on the current state of the learner model. We use a teacher model to measure difficulty of each sample and experiment with three measures based on question answering probability, variability and out-of-distribution. To understand the effectiveness of curriculum learning in various scenarios, we apply it on full model fine-tuning as well as parameter-efficient prompt-tuning settings. Our results show that fixed as well as adaptive curriculum learning significantly improve performance for five commonsense reasoning tasks, i.e., SocialIQA, CosmosQA, CODAH, HellaSwag, WinoGrande in both tuning settings. Further, we find that prioritizing the difficult samples in the tail end of training improves generalization to unseen in-domain data as well as out-of-domain data. Our work provides evidence and encourages research into curriculum learning for commonsense reasoning.","{'model': 'tldr@v2.0.0', 'text': 'This work uses paced curriculum learning to rank data and sample training mini-batches with increasing levels of difficulty from the ranked dataset during finetuning, and finds that prioritizing the difficult samples in the tail end of training improves generalization to unseen in- domain data as well as out-of-domain data.'}",https://aclanthology.org/2022.naacl-main.72.pdf
-NC2T: Novel Curriculum Learning Approaches for Cross-Prompt Trait Scoring,Yejin Lee,"Automated essay scoring (AES) is a crucial research area with potential applications in education and beyond. However, recent studies have primarily focused on AES models that evaluate essays within a specific domain or using a holistic score, leaving a gap in research and resources for more generalized models capable of assessing essays with detailed items from multiple perspectives. As evaluating and scoring essays based on complex traits is costly and time-consuming, datasets for such AES evaluations are limited. To address these issues, we developed a cross-prompt trait scoring AES model and proposed a suitable curriculum learning (CL) design. By devising difficulty scores and introducing the key curriculum method, we demonstrated its effectiveness compared to existing CL strategies in natural language understanding tasks.",,
-Combating the COVID-19 infodemic using Prompt-Based curriculum learning,Zifan Peng,,"{'model': 'tldr@v2.0.0', 'text': 'A prompt-based curriculum learning method that can verify content from multiple perspectives by answering a series of questions concerning the text’s reliability and is robust to the hyperparameter settings, making it more applicable with limited infrastructure resources.'}",
-Using an experiential learning model to teach clinical reasoning theory and cognitive bias: an evaluation of a first-year medical student curriculum,Justin J. Choi,"ABSTRACT Background Most medical students entering clerkships have limited understanding of clinical reasoning concepts. The value of teaching theories of clinical reasoning and cognitive biases to first-year medical students is unknown. This study aimed to evaluate the value of explicitly teaching clinical reasoning theory and cognitive bias to first-year medical students. Methods Using Kolb’s experiential learning model, we introduced dual process theory, script theory, and cognitive biases in teaching clinical reasoning to first-year medical students at an academic medical center in New York City between January and June 2020. Due to the COVID-19 pandemic, instruction was transitioned to a distance learning format in March 2020. The curriculum included a series of written clinical reasoning examinations with facilitated small group discussions. Written self-assessments prompted each student to reflect on the experience, draw conclusions about their clinical reasoning, and plan for future encounters involving clinical reasoning. We evaluated the value of the curriculum using mixed-methods to analyze faculty assessments, student self-assessment questionnaires, and an end-of-curriculum anonymous questionnaire eliciting student feedback. Results Among 318 total examinations of 106 students, 254 (80%) had a complete problem representation, while 199 (63%) of problem representations were considered concise. The most common cognitive biases described by students in their clinical reasoning were anchoring bias, availability bias, and premature closure. Four major themes emerged as valuable outcomes of the CREs as identified by students: (1) synthesis of medical knowledge; (2) enhanced ability to generate differential diagnoses; (3) development of self-efficacy related to clinical reasoning; (4) raised awareness of personal cognitive biases. Conclusions We found that explicitly teaching clinical reasoning theory and cognitive biases using an experiential learning model provides first-year medical students with valuable opportunities for developing knowledge, skills, and self-efficacy related to clinical reasoning.","{'model': 'tldr@v2.0.0', 'text': 'It is found that explicitly teaching clinical reasoning theory and cognitive biases using an experiential learning model provides first-year medical students with valuable opportunities for developing knowledge, skills, and self-efficacy related to clinical reasoning.'}",https://www.tandfonline.com/doi/pdf/10.1080/10872981.2022.2153782?needAccess=true&role=button
-IMPLEMENTATION OF THE EMERGENCY CURRICULUM ON THEMATIC LEARNING IN PRIVATE ELEMENTARY SCHOOL,Hani Hanifah," Abstract: The Covid-19 pandemic has changed the order of the world of education, this has prompted the government to issue policy Number 719/P/2020 regarding guidelines for implementing the curriculum in educational units under special conditions. The research aims to determine the strategies, learning principles, and policy analysis of the emergency curriculum. The data collection techniques in this study were distributing questionnaires to 26 private elementary school teachers in Tasikmalaya City as participants. The research method used is the quantitative method and the data analysis technique used is descriptive analysis. The results of the study stated that the existence of an emergency curriculum policy in the conditions of the Covid-19 pandemic was effective in thematic learning in private elementary schools by using learning strategies that emphasize the character of students and the principles of active and fun learning.",,http://journal.ummat.ac.id/index.php/elementary/article/download/6504/pdf
-The Impact of Time on Learning Outcomes: Lessons from the Delivery of High School Mathematics Curriculum in Ghana,Farouq Sessah Mensah,"A student’s productive learning time is proportional to the length of time they spend working during their lessons. However, it is important to remember that the issues with educational time are not just about optimizing “in-school time.” The study is prompted by the time spent in school and how it impacts students learning outcomes in core mathematics, considering two different year groups for the May/June 2013 West African Senior School Certificate Examination (WASSCE-standardized examination written by Anglophonic countries of West Africa (Ghana, Nigeria, Sierra Leone, Liberia, and the Gambia)). The study compares students’ learning outcomes in the May/June 2013 WASSCE. The study uses a descriptive survey design and conveniently collected data from 10 Senior High Schools (SHSs) in the Central Region of Ghana. The results of the study showed no statistically significant difference in core mathematics grades between students who spent four years and those who spent three years in SHS. Thus, time is not the sole determinant of a student’s learning outcome, as has been suggested by some studies.",,https://www.ej-edu.org/index.php/ejedu/article/download/483/290
-Factors of Learning towards Creating Blended Learning Curriculum Using Learning Management System in Higher Education during Covid-19,Fazilah Razali,"Blended learning is a mix of online and face-to-face instruction. It refers to a variety of possibilities made available by integrating the internet and streaming media with traditional educational formats that necessitate the physical co-presence of teachers and students. While the Covid-19 pandemic is being referred to as the greatest challenge facing education systems around the world, it has prompted authorities to issue a drastic order requiring institutions to switch from face-to-face teaching to online teaching and virtual education for students. This present study aims to determine the model of integration between learning style and active learning towards enhancing blended learning curriculums among higher education students. A total of 208 students at a selected Malaysian higher education institution were selected randomly in this study. For data collection, this study used the quantitative approach and random sampling technique. The instrument was developed based on the literature review, and expert validation was obtained from various universities. The survey forms were distributed using an online medium (Google forms). Pearson Coefficient correlation (r) was employed to examine the relationship between variables, while Structural Equation Modelling (SEM) was utilized to examine the effect of the mediator variable. By dopting a two-stage method, the measurement model was first tested, followed by the structural model. Confirmatory Factor Analysis (CFA) is to test whether the data is compatible with the hypothesized measurement model as well as for the purpose of validation and construct reliability. The results show that the measurement model was a good fit for the data, and that the constructs were reliable and valid for testing the hypotheses based on the results. The assessment of the structural model, on the other hand, involved testing the hypothesized relationships about the direct effects. Four hypothesized direct paths were found to be statistically significant. © 2022 Eskisehir Osmangazi University. All rights reserved.",,
-"Longevity, legacy, and lament: Learning from longstanding educators in an innovative curriculum",G. Maudsley," Research about problem-based learning (PBL) tutoring in medicine has prioritized quantifying relationships between tutor characteristics or learning environment and tutoring behaviour or student outcomes. Longitudinal studies and qualitative research about how such tutors conceptualize their long-term experience are rare. The research question was thus: What educator outlooks do inaugural PBL tutors develop after substantial experience in a problem-based medical curriculum? At16 year-follow-up of interviews with an inaugural cohort of PBL tutors, semistructured interviews with the remaining ten explored their outlooks as educators now versus then. Two years later, an open-ended e-questionnaire (E-interview) reviewed their outlooks, particularly about the curriculum being replaced. Tutors viewed their role now through a more discerning, reflective, and constructivist ‘good educator’ lens. They articulated principles for facilitating active learning. When that curriculum was replaced, tutors were positive about its legacy but also lamented flawed educational governance for maintaining and renewing whole-system integrity. Educator development should prompt critical reflection about ‘the good educator’ identity, the related enthusiasms, discomforts, and uncertainties, and the impact of curriculum shifts.",,
-Aligning Classroom Management Strategies with a Social Emotional Learning Curriculum in Early Childhood,Jessica Kemp,"Traditionally, school-based preventative frameworks have been implemented in isolation with little consideration of alignment and integration of practices throughout the school day. The present study aims to address this gap by increasing school psychologists’ preventative involvement with consultation in early childhood school settings. Using an integrated approach through a multiple baseline design, four Head Start educators were trained in classroom management practices, to increase opportunities throughout the day for teaching, prompting, and reinforcing key skills taught through the Second Step Early Learning (SSEL) social emotional learning curriculum. Findings suggest that a brief professional development session (1-hour) followed by weekly performance feedback (15 minutes) largely increased educator use of aligning classroom management strategies with weekly SSEL lessons (ES = .94, p-value = <.000). Effects on challenging behavior were limited, although two classrooms demonstrated an overall decrease in challenging behavior (ES = -.20). Results further indicate this intervention increased feelings of teacher self-efficacy and was a socially valid approach; educators reported the aligned strategies were acceptable, sustainable, and beneficial to children. Limitations and implications of this study are further discussed with suggested directions for future research.",,https://digitalcommons.pace.edu/cgi/viewcontent.cgi?article=1054&context=perspectives
-TaxonPrompt: Taxonomy-aware curriculum prompt learning for few-shot event classification,Chengyu Song,,,
-Curriculum Implementation Status and Adoption of ICT Resources in Distance Learning Programme of the Nigerian Teachers’ Institute,I. A. Alade,"As novel as the notion of establishing the National Teachers’ Institute (NTI) Distant Learning Studies is a general observation shows that various people still pick flaws in the programme, especially, its Nigerian Certificate in Education (NCE) at the implementation level. Some quarters doubt the quality of teaching manpower produced as well as how realistic the synergy of combining NTI teacher education with new technology in the training centres. These observations prompted this study which examined “curriculum implementation status and adoption of ICT resources in distance learning programme of the Nigerian Teachers’ Institute. The descriptive survey was guided by three research questions. Curriculum Implementation Status and ICT Resources Rating Scale with a reliability co-efficient of 0.84 was used to collect data from 373 respondents (73 facilitators and 300 students). Frequency count, percentage weights and pie-chart were used for data analysis. The findings revealed that: the programme facilitators are 82.3% and 70.0% qualified qualification wise and in working experience respectively; The adoption of many ICT resources are inadequate except the use of mobile phones with 70.0% adoption; and a below average (43.3%) rating in supervision and monitoring of the teacher education curriculum implementation, but a fairly regular supervision and monitoring of classroom instructional practices for quality assurance (53.4%). It is recommended that basic ICT courses should be part of NCE curriculum for its gradual integration into the teacher preparation process. Also, the government at all times must be cautions not to leave the curriculum implementation affairs in the hands of non-professionals for maximum yield. Keywords: Curriculum Implementation, ICT Resources, Teacher Education.",,
-Learning inclusion through makerspace: a curriculum approach in Italy to share powerful ideas in a meaningful context,Taziana Giusti,"Learning by making is being recognized as an efficient technique for students to develop knowledge and skills simultaneously. However, one of the most urgent challenges that schools are facing nowadays is to reach every student in their individual profile and potential. Therefore, the purpose of this paper is to offer an integrated approach for re-thinking the role of Makerspace in a context of inclusion: the characteristics of this learning strategy, in contrast to the traditional currents, can offer promising paths to successfully build shared ideas.,This paper critically analyses a Makerspace workshop implemented during a whole academic year in an Italian state primary school. This exploratory and instrumental qualitative approach (Baxter and Jack, 2008) included two channels that have been developed simultaneously, namely, technological skills and social competences.,Stemming from a long tradition of inclusion of children with various educational needs in the mainstream school system, the authors aim to share a success story of academic and social achievements: all the participants were able to develop at their own pace, sharing tools, reaching a balance between the demands of the task and their planning and negotiation skills.,Small group size and the reiterated daily interactions with differences embodied in students with special needs, immigration background, low SES, gifted.,Through ad-hoc training, relatively marginalized pedagogical components (such as the ability to work in low-control situations, flexibility, student-centered learning environments) should be given a more prominent role and can be introduced in the desirable professional development. In addition, national and school policies are prompted to consider its inclusion as a slow-process that cannot be fully achieved in the presence of time and space constraints.,The curricular approach discussed above has shown the importance of inclusion of all students within mainstream schools. Pupils with atypical development can interact with other children, and in this way, they can have first-hand experience of how social dynamics unfold in a real environment. Moreover, they can act in a challenging context where, more often than not, they are pushed to achieve goals exceeding their supposed cognitive abilities. The other pupils are also gaining from these interactions: they can understand different points of view, thus developing empathy, and they can appreciate original ways to approach a task, with cognitive and emotional benefits. In addition, the constant relationship helps them to control their reactions to behavioral problems that sometimes classmates with special needs display and so they deepen their knowledge of and tolerance for others’ peculiarities.,Based on the foundational principles of Papert’s powerful ideas and meaningful context, this paper describes the design principles of a successful makerspace, its integration in the school curriculum, and the achieved inclusion of children with Special Education Needs in a group of peers where adults became observers. Recommendations are discussed on how school practitioners can promote young children’s learning through making.",,
-Facilitating Teaching and Learning with Made to Measure Fashion Design and Creation MOOC Courses,Tanya Dove,"Massive Online Open Courses (MOOC) provides a flexible and engaging learning environment. A MOOC in Apparel Technology facilitates a comprehensive e-Learning platform, with a variety of instructional videos and technical resources, which will foster a technical foundation in pattern cutting and sewing for learners. Apparel technology is a core skill for fashion design students; however, universities only address a limited range of construction techniques within their curriculum. Students work towards model size apparel for their graduate fashion shows, which limit the learner’s knowledge in developing apparel for different body shapes. A MOOC in Apparel Technology would facilitate learners with the skills and knowledge to develop apparel for different sizes and figure shapes, with fit assessment and problem-based forums on fit solutions. With many retailers now providing a global platform for the sales of clothing, the sizing of women’s clothing is an area of concern in the fashion industry, ‘one size does not fit all’. The inconsistencies in garment sizing can contribute to the dissatisfaction of garment fit, leaving women unable to find, or know, what size fits their body shape. MOOC incorporates an innovative design and e-Learning platform, in a less structured approach, which harnesses learning in an interactive setting, with online interactions prompting learners to collaborate, share ideas and examples, in a less structured environment.","{'model': 'tldr@v2.0.0', 'text': 'A MOOC in Apparel Technology would facilitate learners with the skills and knowledge to develop apparel for different sizes and figure shapes, with fit assessment and problem-based forums on fit solutions.'}",
-English Curriculum Reform in the Philippines: Issues and Challenges from a 21st Century Learning Perspective,J. Barrot,"ABSTRACT Recent developments in the region and in the Philippines have prompted the Philippine government to push for a new basic education curriculum. Along with these changes is the adoption of the new English curriculum known as the Language Arts and Multiliteracies Curriculum (LAMC). This article, therefore, sought to examine the K-12 English Curriculum in the Philippines from a 21st century learning perspective. The first section of this article briefly describes the LAMC and 21st century learning in both general and English language teaching-related terms. The discussion then reviews the specificity and coherence of the LAMC, how it is consistent with the principles espoused by 21st century learning, and how it is aligned to established language teaching and learning principles. The findings revealed that the current curriculum needs to improve its specificity, internal coherence, and integration of some essential principles of 21st century learning and language teaching and learning. The article concludes with the possible challenges in implementing the curriculum, suggestions for future design and implementation, and implications for future studies.",,
-Using Perusall to motivate students’ curriculum-based academic reading,Mimi Li,,,
-"Initiating curriculum mapping on the web-based, interactive learning opportunities, objectives and outcome platform (LOOOP)",I. Treadwell,"Background. A web-based curriculum is made transparent by providing multidimensional overviews of content (learning objectives) aligned with learning outcomes and frameworks, opportunities and assessment formats. A South African university embarked on the mapping of its curricula on the web-based learning opportunities, objectives and outcome platform (LOOOP). Objectives. To reflect on the customisation of LOOOP and training of lecturers, and to determine lecturers’ perceptions of the usability and value of LOOOP. Methods. The project manager reflected on the initiating processes, and a survey determined the lecturers’ perceptions of the usability and value of curriculum mapping, using a 4-point Likert scale questionnaire. The convenience sample comprised the first 30 lecturers who had uploaded their curriculum content and consented to partake in this ethics-approved study. Descriptive statistics portray the percentages of agreement on the positive statements of the questionnaire. Results. Challenges related to slow staff buy-in and development were experienced. Required modifications to LOOOP were promptly dealt with. The majority of participants agreed on the usability (≥89%), as well as structure and transparency (≥87%) of LOOOP. Mapping is expected to enhance curriculum revision (≥95%) and communication (≥96%), viewing the curriculum scope, complexity and cohesion (97%), as well as abstracting data for management analysis and reporting (100%). Conclusions. The lecturers agreed on the usability and values of curriculum mapping, which indicates that online mapping is sufficiently beneficial to justify the time and resources invested. Mapping should be a product of collaborative participation and planned as a long-term commitment, which can also be used to research the impact of mapping on student learning.","{'model': 'tldr@v2.0.0', 'text': 'The lecturers agreed on the usability and values of curriculum mapping, which indicates that online mapping is sufficiently beneficial to justify the time and resources invested.'}",http://www.ajhpe.org.za/index.php/ajhpe/article/download/1017/616226
-The Opportunity and Challenges of Implementing a Prototype Curriculum,P. Solikhah,"Learning loss due to the Covid-19 pandemic prompted the simplification of the 2013 Curriculum into an Emergency Curriculum, which was then refined into a Prototype Curriculum. Curriculum changes always bring optimism as well as skepticism. The hope for a new curriculum that is better in line with the ability of education providers in implementation. This research is evaluative research with a literature study approach. This study aims to make a corrective contribution to the policy coherence of the prototype curriculum. The data source comes from government policies supported by expert analysis of curriculum developers. Data analysis is rationally deductive. It was concluded that the prototype curriculum focused on developing non-technical abilities and character within the framework of Pancasila values, simplifying learning materials, and developing literacy and numeracy skills. Opportunities from this curriculum are reducing the administrative burden of teaching, and the design of learning becomes more flexible so that the quality of learning can be improved. The challenge is the teacher's readiness (attitudes and emotions, knowledge, and behavior). Based on the Ministry of Education's recommendation, which refers to the results of the 2020 UKG, it is stated that the quality and competence of teachers are still low, it is necessary to improve IT capabilities, and the KKG and MGMP functions are not yet maximized.",,https://e-journal.ikhac.ac.id/index.php/NAZHRUNA/article/download/2130/907
-The Companion Curriculum: medical students’ perceptions of the integration of humanities within medical education,Charlotte Axelrod,"Background The contributions of arts and humanities to medical education are known in the medical education community, but medical schools’ offerings vary. The Companion Curriculum (CC) is a student-curated set of optional humanities content for medical students at the University of Toronto. This study evaluates integration of the CC to identify key enabling conditions for medical humanities engagement. Methods A mixed-methods evaluation gauged usage and perceptions of integration of the CC among medical students using an online survey and focus groups. Narrative data underwent thematic analysis, supported by summary statistics of quantitative data. Results Half of survey respondents were aware of the CC (n = 67/130; 52%), and, once prompted with a description, 14% had discussed it in their tutorial groups. Of students using the CC, 80% reported learning something new regarding their roles as communicators and health advocates. Themes were the perceived value of the humanities, internal student barriers, institutional neglect of the humanities, and student critiques and recommendations. Conclusion Despite participants’ interest in medical humanities, our CC remains underused. To improve humanities’ visibility in the MD curriculum, our results indicate that greater institutional support, including faculty development and early curricular integration, is required. Further study should explore reasons for gaps between interest and participation.","{'model': 'tldr@v2.0.0', 'text': 'To improve humanities’ visibility in the MD curriculum, the results indicate that greater institutional support, including faculty development and early curricular integration, is required.'}",https://journalhosting.ucalgary.ca/index.php/cmej/article/download/72907/56208
-FPC: Fine-tuning with Prompt Curriculum for Relation Extraction,Sicheng Yang,"The current classification methods for relation extraction (RE) generally utilize pre-trained language models (PLMs) and have achieved superior results. However, such methods directly treat relation labels as class numbers, therefore they ignore the semantics of relation labels. Recently, prompt-based fine-tuning has been proposed and attracted much attention. This kind of methods insert templates into the input and convert the classification task to a (masked) language modeling problem. With this inspiration, we propose a novel method Fine-tuning with Prompt Curriculum (FPC) for RE, with two distinctive characteristics: the relation prompt learning, introducing an auxiliary prompt-based fine-tuning task to make the model capture the semantics of relation labels; the prompt learning curriculum, a fine-tuning procedure including an increasingly difficult task to adapt the model to the difficult multi-task setting. We have conducted extensive experiments on four widely used RE benchmarks under fully supervised and low-resource settings. The experimental results show that FPC can significantly outperform the existing methods and obtain the new state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method Fine-tuning with Prompt Curriculum (FPC) for RE, with two distinctive characteristics: the relation prompt learning, introducing an auxiliary prompt-based fine-tuned task to make the model capture the semantics of relation labels; the prompt learning curriculum, a fine- Tuning procedure including an increasingly difficult task to adapt the model to the difficult multi-task setting.'}",
-The Feasibility and Effectiveness of an Online Curriculum for Virtual Onboarding/Orientation of Graduate Medical Education Learners During the COVID-19 Pandemic,Scott A. Holliday,"Background Graduate medical education (GME) orientation/onboarding is conventionally an in-person activity, but the COVID-19 pandemic prompted virtual approaches to learner onboarding. However, online GME onboarding strategies have not been disseminated in the literature. Objective To determine the usefulness of an online curriculum for GME learner orientation at a large sponsoring institution using an electronic survey. The primary outcome was to discover the usefulness of our online curriculum for GME onboarding, and secondary outcomes included identifying barriers to implementation and weaknesses associated with online GME orientation. Methods We created an online GME orientation curriculum to onboard incoming learners (from June 1 to August 31, 2020) and electronically surveyed our learners to determine the usefulness of this novel approach. We conducted orientation sessions and electronically recorded questionnaire responses using CarmenCanvas, our institutional learning management system. Linear regression analysis was performed to identify factors predicting satisfaction with virtual GME orientation using IBM SPSS Statistics, Version 26.0 (Armonk, NY, USA). Results Of 353 trainees, 272 completed the survey for a 77% response rate. 97% of respondents reported that the curriculum supported performance of learner duties. 79% of trainees perceived the overall quality as “very good” or “good”, 91% responded that the curriculum provided “effective learning”, 94% reported “accessing the course content easily”, 92% reported “easily navigating the curriculum”, 91% described the curriculum as “well-organized”, and 87% reported that the lectures “supported their learning”. Conclusion Online delivery of a comprehensive GME orientation curriculum is useful and facilitates learner education, training, and integration into a large GME institution in the COVID-19 era.",,https://www.dovepress.com/getfile.php?fileID=85408
-Wither the plurality of decolonising the curriculum? Safe spaces and identitarian politics in the arts and humanities classroom,Ana Cristina Mendes,"Contributing to the debate on decolonising the curriculum, this reflective article questions: What does a safe space in a decolonised classroom mean? For whom is it safe? And at what cost? Must we redraw the parameters of ‘safe’? Prompted by a real-life ‘n-word incident’ in the classroom, this article unpacks the collision of decolonising the curriculum to continue making teaching and learning more pluriversal and inclusive, with the enactment of the ‘wounded attachments’ of identitarian politics and the playing of ‘Privilege or Oppression Olympics’. Using snippets from British parody and satire on decolonising the university, we query how far wokeness in a university setting can become political correctness taken to extremes that threaten decolonising efforts. In its concluding reflections, the article makes tentative recommendations for setting up safe spaces, away from self-silencing or censoring, and backing away from contention and provocation in the classroom.",,https://journals.sagepub.com/doi/pdf/10.1177/14740222221100711
-Integrating academic skills in the curriculum: a partnership approach,Emma L. Smith,"Supporting our new students to make the transition to higher education, so that they stay with us and succeed beyond their first year, has been a priority focus for the University of Salford over the past 18 months. As an academic skills team, based in the Library, we have carved out an integral role for our service in responding to this challenge. Building on the prior success of a standalone eLearning programme, we have developed an extensive set of eLearning pathways and complementary learning activities designed for academics to easily and flexibly integrate into their course delivery so that every student is connected with the right academic support at the right time. This presentation aimed to share how our active blended learning approach is scalable and allows for local ownership and opportunities for contextualisation by academic colleagues. We explored how we established our role in this strategic project and the value of our partnership working with the academic community and the VLE support team. We hoped to provide helpful examples of how it has worked in practice to support students to learn how to learn at university. Finally, we sought to reflect on the journey so far - acknowledging the bumps and bends in the road - and initiate a discussion about where to go next. The community response to the presentation provided a space for practitioners to reflect on the ways in which e-learning design could facilitate an embedded approach to academic skills at their institutions; it also prompted reflection on the potential for collaboration between learning developers and Faculties. Our final author reflections address some of the questions and challenges raised at the conference, and set out some future aspirations for where this work will go next.",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/984/612
-Pharmacy Student Challenges and Strategies towards Initial COVID-19 Curriculum Changes,Lu Liu,"Due to COVID-19, tertiary institutions were forced to deliver knowledge virtually, which proposed challenges for both institutions and students. In this study, we aimed to characterize pharmacy students’ challenges and strategies during COVID-19 curriculum changes, therefore developing a comprehensive understanding of students’ learning, wellbeing, and resilience in the ever-changing situation. Data were collected from student written reflections across four year levels at one school of pharmacy from March–May 2020. In addition, data were collected from written responses of second-year pharmacy students responding to prompted questions. The data were qualitatively analyzed inductively by five coders using NVivo 12. For each piece of data, two coders independently coded the data, calculated the inter-rater agreement, and resolved discrepancies. The most coded challenges were ‘negative emotional response’ and ‘communication barrier during virtual learning’. The most coded strategies were ‘using new technology’ and ‘time management’. This study allows researchers and education institutions to gain an overview of pharmacy students’ experiences during COVID-19, therefore helping universities to provide students with necessary support and techniques on how to self-cope with COVID-19 as well as stressful events in the future.","{'model': 'tldr@v2.0.0', 'text': 'This study helps universities to provide students with necessary support and techniques on how to self-cope with COVID-19 as well as stressful events in the future by developing a comprehensive understanding of students’ learning, wellbeing, and resilience in the ever-changing situation.'}",https://www.mdpi.com/2227-9032/9/10/1322/pdf?version=1633341761
-CVRriculum Program Faculty Development Workshop: Outcomes and Suggestions for Improving the Way We Guide Instructors to Embed Virtual Reality Into Course Curriculum,Eva Peisachovich,"Experiential education and student engagement are a main source of student attraction and retention in post secondary milieus. To remain innovative, it is imperative that universities look beyond the internet and traditional multimedia mediums and incorporate novel ways and cutting-edge technologies that can drastically change the way students and educators experience learning. The application of technology as an approach to experiential education is becoming more popular and has extensively impacted universities and other higher education organizations around the world. One approach to support this change in education delivery is to use immersive technologies such as virtual reality (VR). Our team has conducted a pilot study that focuses on embedding VR as a medium to teach empathy within higher education milieus. We began the study by conducting a pilot faculty development workshop to provide an understanding of VR and ways it can be embedded as a pedagogical approach to support curriculum design. Five faculty members from a local university were recruited to participate. Outcomes suggest that embedding VR into the curriculum is a feasible approach that provides an engaging learning environment that is effective for teaching an array of interpersonal skills. The workshop laid the foundation for future faculty training programs guiding the use of VR, prompting a dialog regarding plans for future workshops across a pan-university context.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study that focuses on embeddingVR as a medium to teach empathy within higher education milieus and suggests that embedding VR into the curriculum is a feasible approach that provides an engaging learning environment that is effective for teaching an array of interpersonal skills.'}",https://www.cureus.com/articles/52352-cvrriculum-program-faculty-development-workshop-outcomes-and-suggestions-for-improving-the-way-we-guide-instructors-to-embed-virtual-reality-into-course-curriculum.pdf
-Homelessness in the Medical Curriculum: An Analysis of Case-Based Learning Content From One Canadian Medical School,Matthew J To,"Abstract Phenomenon: Homelessness is a major public health concern. Given that homeless individuals have high rates of mortality and morbidity, are more likely to be users of the healthcare system, and often report unmet health needs, it is important to examine how homelessness is addressed in medical education. We wanted to examine content and framing of issues related to homelessness in the case-based learning (CBL) curriculum and provide insights about whether medical students are being adequately trained to meet the health needs of homeless individuals through CBL. Approach: CBL content at a Canadian medical school that featured content related to homelessness was analyzed. Data were extracted from cases for the following variables: curriculum unit (e.g., professionalism/ethics curriculum or biomedical/clinical curriculum), patient characteristics (e.g., age, sex), and medical and social conditions. A thematic analysis was performed on cases related to homelessness. Discrepancies in analysis were resolved by consensus. Findings: Homelessness was mentioned in five (2.6%) of 191 CBL cases in the medical curriculum. Homelessness was significantly more likely to be featured in professionalism/ethics cases than in biomedical/clinical cases (p = .03). Homeless patients were portrayed as socially disadvantaged individuals, and medical learners were prompted to discuss ethical issues related to homeless patients in cases. However, homeless individuals were largely voiceless in cases. Homelessness was associated with serious physical and mental health concerns, but students were rarely prompted to address these concerns. Insights: The health and social needs of homeless individuals are often overlooked in CBL cases in the medical curriculum. Moreover, stereotypes of homelessness may be reinforced through medical training. There are opportunities for growth in addressing the needs of homeless individuals through medical education.","{'model': 'tldr@v2.0.0', 'text': 'Examination of content and framing of issues related to homelessness in the case-based learning (CBL) curriculum and insights about whether medical students are being adequately trained to meet the health needs of homeless individuals through CBL are provided.'}",
-Responding to crisis and building forward better: The relevance of curriculum,Yao Ydo,,"{'model': 'tldr@v2.0.0', 'text': 'This triple special issue of Prospects addresses questions head on of curriculum preparedness, by foregrounding the work of more than sixty leading curriculum researchers.'}",https://link.springer.com/content/pdf/10.1007/s11125-021-09572-8.pdf
-Evaluation of a technical information systems module for distance learning during the COVID-19 pandemic,A. Aryal,"PurposeIn the wake of the pandemic (COVID-19) throughout the United States, many educators had to promptly switch to online modality to continue to provide education to students with safety through physical distancing requirements. This study mainly focuses on delivering an Information Systems module aligned with the information systems curriculum model during a pandemic.Design/methodology/approachThe authors used data collection techniques from student assignments and course tasks. These data are used for the purpose of academic assessment for the Information Systems program. The student performance is assessed on a 5-point scale (1 being low and 5 being high) for the synchronous and the asynchronous tasks related to the course. The authors compared the student performance during the pandemic to the pre-pandemic semester.FindingsThis study revealed that the technical module of an information systems course can be successfully delivered during a pandemic in a remote session. However, the authors found that there is a decline in the student performance in synchronous tasks and asynchronous tasks. But the decline of the student performance in the synchronous tasks is greater than that of the asynchronous tasks. The result of this study helps the Information Systems program with their assessment and to improve their course delivery during a pandemic.Originality/valueIn this paper, the authors examine the delivery of a technical module in the field of information systems via online learning models. The authors particularly examine the synchronous and asynchronous online learning models in the delivery of the technical module. The lessons learned from transitioning to the online modality can help universities better prepare for the future during unprecedented times.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that the technical module of an information systems course can be successfully delivered during a pandemic in a remote session, however, the authors found that there is a decline in the student performance in synchronous tasks and asynchronous tasks.'}",https://www.emerald.com/insight/content/doi/10.1108/JRIT-11-2021-0078/full/pdf?title=evaluation-of-a-technical-information-systems-module-for-distance-learning-during-the-covid-19-pandemic
-STUDENT PERCEPTIONS ON EXPERIENTIAL LEARNING WITHIN THE B.PHARM CURRICULUM IN NIGERIA,Abdu-Aguye,"Within the undergraduate pharmacy curriculum for the bachelor of pharmacy (B.Pharm) degree in Nigeria, experiential learning is covered under the Student Industrial Work Experience Scheme (SIWES). The aim of this work was to describe student experiences during SIWES with a view towards identifying problems and recommendations that could be used for improvement. The study was descriptive in nature and utilized qualitative data obtained from students of Ahmadu Bello University, Zaria and University of Maiduguri, Borno. Participating students (151) had completed the 4 year of their B. Pharm degree and undergone their SIWES attachment during the 2015/2016 academic session. They were given 10 questions and asked to answer them either by writing a reflection (essay) or in the form of an open ended questionnaire, data collected was then analyzed using summative qualitative content analysis. Respondents mostly did their attachments in hospitals (71.5%) and community pharmacies (24.5%). They generally believed that they would be exposed to the practical aspects of what they had earlier learnt theoretically, and over half of students reported feeling that their expectations had been met. Patient interaction in the form of counselling was the most enjoyable activity reported by most of the respondents. Conversely, it was also the biggest difficulty encountered. Other difficulties experienced by the students included problems locating drugs on the shelves and learning about the various drug brand names and dosages available. Recommendations on how to improve the scheme included; prompt payment of their stipend to offset transport costs, incorporating more experiential learning opportunities into the curriculum and making it mandatory for students to rotate amongst various practice settings. In conclusion, while students generally had positive perceptions of the scheme, several aspects could be improved",,
-Establishing Social Learning in an Engineering MOOC: Benefits for Diversity and Inclusion in Engineering Education,J. Brereton,"Recent Higher Education Statistics Agency data shows that only 20% of engineering students at UK Universities are female, despite the hard work being undertaken by many educational institutions to address this gender imbalance via outreach events and special interventions focussing on girls/women in STEM. It has been argued that student-centred teaching methods, together with changes in the engineering curriculum itself, which emphasise the social, creative, and human-centred aspects of the discipline, are required to effect real change in engaging with those from traditionally underrepresented groups. Through analysing quantitative data on age, gender, learner type, and commenting rates in peer-to-peer discussions, we examine the development and delivery of an engineering MOOC, before, during, and after COVID-19-related lockdowns in the UK, to identify what aspects of online learning might be harnessed to improve diversity in engineering education. The results show that the MOOC attracted a better gender balance than reported for UK-based in-person engineering programmes. In addition, we show that careful structuring of discussion prompts encouraged higher levels of social learning. We recommend the continued use of interactive and discursive elements within a blended learning environment to positively impact diversity and inclusion in engineering education specifically, and STEM education in general.",,https://www.mdpi.com/2071-1050/14/9/5472/pdf?version=1651725710
-Secondary mathematics teachers learning to do and teach mathematical modeling: a trajectory,Rose Mary Zbiek,,,https://link.springer.com/content/pdf/10.1007/s10857-022-09550-7.pdf
-Learning to Perform Complex Tasks through Compositional Fine-Tuning of Language Models,Victor S. Bursztyn,"How to usefully encode compositional task structure has long been a core challenge in AI. Recent work in chain of thought prompting has shown that for very large neural language models (LMs), explicitly demonstrating the inferential steps involved in a target task may improve performance over end-to-end learning that focuses on the target task alone. However, chain of thought prompting has significant limitations due to its dependency on huge pretrained LMs. In this work, we present compositional fine-tuning (CFT): an approach based on explicitly decomposing a target task into component tasks, and then fine-tuning smaller LMs on a curriculum of such component tasks. We apply CFT to recommendation tasks in two domains, world travel and local dining, as well as a previously studied inferential task (sports understanding). We show that CFT outperforms end-to-end learning even with equal amounts of data, and gets consistently better as more component tasks are modeled via fine-tuning. Compared with chain of thought prompting, CFT performs at least as well using LMs only 7.4% of the size, and is moreover applicable to task domains for which data are not available during pretraining.","{'model': 'tldr@v2.0.0', 'text': 'This work presents compositional fine-tuning (CFT): an approach based on explicitly decomposing a target task into component tasks, and then fine- Tuning smaller LMs on a curriculum of such component tasks.'}",http://arxiv.org/pdf/2210.12607
-Federated Few-Shot Learning for Mobile NLP,Dongqi Cai,"Natural language processing (NLP) sees rich mobile applications. To support various language understanding tasks, a foundation NLP model is often fine-tuned in a federated, privacy-preserving setting (FL). This process currently relies on at least hundreds of thousands of labeled training samples from mobile clients; yet mobile users often lack willingness or knowledge to label their data. Such an inadequacy of data labels is known as a few-shot scenario; it becomes the key blocker for mobile NLP applications. For the first time, this work investigates federated NLP in the few-shot scenario (FedFSL). By retrofitting algorithmic advances of pseudo labeling and prompt learning, we first establish a training pipeline that delivers competitive accuracy when only 0.05% (fewer than 100) of the training data is labeled and the remaining is unlabeled. To instantiate the workflow, we further present a system FeS1, addressing the high execution cost with novel designs: (1) Curriculum pacing, which injects pseudo labels to the training workflow at a rate commensurate to the learning progress; (2) Representational diversity, a mechanism for selecting the most learnable data, only for which pseudo labels will be generated; (3) Co-planning of a model's training depth and layer capacity. Together, these designs reduce the training delay, client energy, and network traffic by up to 46.0×, 41.2× and 3000.0×, respectively. Through algorithm/system co-design, FeS demonstrates that FL can apply to challenging settings where most training samples are unlabeled.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates federated NLP in the few-shot scenario (FedFSL), and demonstrates that FL can apply to challenging settings where most training samples are unlabeled, through algorithm/system co-design.'}",https://arxiv.org/pdf/2212.05974
-Facilitating Learners’ Self-Regulated Learning Skills and Self-Efficacy to Write in English Using Technologies,Kretsai Woottipong,"The purpose of this quasi-experimental research was to study the effects of self-regulated strategy development (SRSD) instruction with explicit generalization training prompted by Constructivist Self-regulating Virtual Composing Activities (CSRVCA) on students’ reported use of self-regulatory strategies, their self-efficacy to write and writing performance. The sample group in this study was 86 high school students who were selected through the use of convenience sampling, then 44 students were assigned to a control group and 42 students were in an experimental group. The students from the experimental group received regular writing classes integrated with constructivist virtual writing instruction, while those from the control group received regular writing classes, required by the school curriculum and syllabus. Instruments included surveys, writing tests, and students’ reflective journals. Data were collected through semi-structured interviews and reflective journals to triangulate quantitative results. The results revealed that the students from the experimental group had a significantly higher level of deploying self-regulatory strategies than those in the control group. They also had a higher level of perceived writing self-efficacy than the students in the control group. Although both groups showed improvement in writing scores, the experimental group outperformed the students in the control group in the post and the delayed post-tests.",,https://jurnal.unai.edu/index.php/acuity/article/download/2581/2052
-A Scientific Research Ability Prompting Approach in the Teaching of the Machine Learning Curriculum,Bi Zhao,"The teaching method of machine learning has a significant influence on the promotion of students' research ability. In traditional machine learning curriculum teaching, students have low interest in the theoretical study and lack practical ability of scientific research. It is particularly important to reform the teaching of machine learning to stimulate students' interest in the study, and to cultivate their comprehensive research ability. In this paper, we propose the idea of using exploring education to improve students' scientific research capability. Traditional teaching mode was transformed into an interactive teaching mode by using the methods of interactive teaching and case-teaching. In our practical teaching, our method obviously improved the teaching effect and got positive responses from participated students.","{'model': 'tldr@v2.0.0', 'text': ""This paper proposes the idea of using exploring education to improve students' scientific research capability by using the methods of interactive teaching and case-teaching.""}",http://dpi-proceedings.com/index.php/dtssehs/article/download/3471/3121
-Covid-19 compelled RASE Model curriculum transaction for teacher education -A study,C. Praveen,"The unprecedented shut down of educational institutions owing to the Covid pandemic had prompted many institutions to shift to an Online mode of instruction. The investigator, a teacher educator by profession found the abrupt shift from face-to-face instruction to an online only pedagogy for teacher education a major challenge. This necessitated the exploration of appropriate instructional strategies to substitute the existing time-tested pedagogical practices followed in teacher training colleges. The target audience specific data collected through surveys and interviews revealed that the needs, resources available and internet connectivity of main stake holders namely teachers and learners vary in several ways. This posed several challenges and the investigator found a solution in the RASE Model which is increasingly being perceived as a practical, evidence based design model with applications of technology to improve student learning outcomes and student satisfaction (Churchill, King, & Fox, 2013). When a batch of English optional teacher trainees of the Bachelor of Education (BEd) programme, was assigned to the investigator-cum-teacher educator, the RASE model which basically addresses Resources, Activity, Support and Evaluation was utilized for curriculum transaction. This paper lists down the strategy employed to transact the BEd curriculum using the RASE model. It is hoped that the report of the study will help colleges of teacher education and English language teacher educators to gain a sense of direction to commence delivery of instruction which has already been delayed owing to the pandemic.",,
-A Study on the Characteristics and Elements of Best Distance Learning Lecture through Topic Modeling,Songlee Han,"Objectives The purposes of this study is to identify the characteristics and elements of exemplary remote learning, provide factors that control the quality of effective remote learning, and present implications according to the educational environment changed by the COVID-19 situation. Methods For this study, a total of 324 students(1,178 case) and 459 instructors(459 case) in University A located in Seoul were surveyed using Google Survey, and the collected data was analyzed through topic modelling’s LDA method. Results Survey results show that instructors and students each perceived four topics of remote learning elements. Topics extracted from the students’ surveys are “prompt response to problems that occur”, “difference in the quality of lectures according to the instructors’ design of remote lectures”, “systematic preparation of remote learning and good communication”, and “recreating a curriculum that is suitable for remote learning”. Topics extracted from the instructors’ survey are “supporting the environment and equipment needed for remote learning”, “specific guidelines for remote learning”, “provision of clear grading guideline for remote learning”, and “good communication in remote learning.” Conclusions The results of this study present practical implications for smooth remote learning, such as the need for both students and instructors to have basic media literacy necessary for remote learning and the need to redesign teaching and learning to fit the remote learning environment(e.g., systematic redesign of lectures, interaction, class evaluation criteria, etc., appropriate for remote learning), and propose future research topics. ",,
-Exploring the use of writing frames to teach and assess writing in English Additional Language learning in schools,Marina Burger,"The learning of writing is complex because it requires the development of cognitive and linguistic abilities. Effective teaching and learning of writing demand guided practice. The South African National Curriculum Statement Grades R–12 includes the Curriculum and Assessment Policy Statement (CAPS) for First Additional Language (FAL) teaching and learning. It introduces the use of question-prompt writing frames to learn writing in grade 4. In the second part of the grade 5 year, process writing is introduced as a writing frame that should be used when needed. From grade 7 to grade 12 process writing is implemented as part of the writing lessons. The education department adopted assessment for learning as a teaching and learning approach where feedback provides scaffolded learning. This theoretical paper argues that the writing frames used in the South African curriculum are inadequate to scaffold the learning of writing; furthermore, that the assessment of the writing skills of learners tends to focus on closed skills. The implementation of assessment for learning as an approach to teaching and learning is yet to develop fully to ensure successful scaffolded learning. Additionally, the writing frames introduced limit the teaching of a variety of writing genres and restrict the teaching of writing to western narrative and writing styles. The paper intends to demonstrate that the use of targeted frames appropriate to the writing task and outcome would improve writing learning and assessment.",,https://perlinguam.journals.ac.za/pub/article/download/1044/804
-PROMPTING FACTORS CULTIVATING INDONESIAN EFL LEARNERS’ LEARNING AUTONOMY BEHAVIORS,K. Wijaya,"Previous research results unearthed that highly-autonomous EFL learners are more inclined to possess a higher sense of ownership of their learning enterprises owing to the advanced level of learning management. This library study was a further attempt enacted by the researcher to profoundly investigate particular factors believed to cultivate Indonesian EFL learners’ learning autonomy behaviors. In light of this major research objectivity, 30 previous learning autonomy literature were selectively selected to shed clearer enlightenment for Indonesian EFL experts, practitioners, curriculum designers, and educators to start internalizing various meaningful learning activities consolidating learners’ autonomous learning behaviors. Generally speaking, the obtained research results unveiled that the constant nourishment of Indonesian EFL learners’ learning autonomy behaviors can be insured in case second language teachers are committed to promoting more facilitative and innovative learning enterprises in their daily routines. In conformance with these research results, it is of foremost importance for second language teachers to organize their classroom learning dynamics harmonious with learners’ specific learning situations to advance their autonomous learning behaviors to greater development.",,
-Authentic learning tasks within physical education teacher education : redesigning the curriculum,M. Slingerland,"Teaching PE requires an integrated set of knowledge, skills and attitudes and for optimal development students should be exposed to various meaningful learning situations. However, the current School of Sport Studies PETE curriculum seems to be suboptimal due to compartmentalization and fragmentation of knowledge, skills and attitudes, thereby hindering optimal transfer of what is learnt in PETE to what is done in the field. In addition, the PE profession is subject to continuous change (e.g. PE as an examinable subject, PE teachers as liaison officers between school, community and sport clubs), while there is also an increasing demand for reflective practice and life-long learning. Altogether, this has prompted a rigorous redesign of the PETE curriculum. The curriculum redesign process was based on the principles of the four-component instructional design (4C-ID) model (Van Merrienboer, Clark, & Croock, 2002). Central in this approach are authentic learning tasks: situations that emerge from the actual PE profession. First, a skill hierarchy was created that served as a foundation for further curriculum design. Second, a blueprint was worked out, providing a global thematic overview of the four-year curriculum. Then, education was worked out in detail through the design of authentic learning tasks. The design process has resulted in a more coherent curriculum in which authentic learning tasks form the backbone and in which knowledge, skills and attitudes are optimally integrated. The 4C-ID model provides a promising starting point for curriculum (re)design for PETE. Within this presentation an overview of the curriculum will be provided and the suitability of the 4C-ID model as well as various organizational and educational challenges are discussed.",,
-Impacts of the COVID-19 Pandemic on Teaching and Learning Social Studies: A Literature Review,Mohammed Abdullah Al-Nofli,"The unprecedented circumstances surrounding the COVID-19 crisis have affected K-12 education systems in most countries around the world. It is widely recognized that this crisis has presented opportunities and challenges for various school subjects. The purpose of this study was to examine the impacts of the COVID-19 pandemic on teaching and learning social studies. The study reviewed the content of 53 articles on K-12 social studies and its associated subjects (e.g., geography, history, and civic education) published between January 2020 and July 2022. Overall, the research findings indicated that COVID-19 prompted social studies teachers and educators to prioritize important content areas in the social studies curriculum (e.g., health education, human-rights education, digital citizenship, and children’s everyday geographies), integrate a wide range of technologies, and implement innovative teaching practices. In spite of these opportunities, there were some common challenges that raised concern across international contexts, including pedagogical challenges, accessibility and connectivity challenges, administrative challenges, and insufficient professional development for teachers. The study recognized the need for more research on social studies, particularly in the area of learning loss.",,https://ijlter.org/index.php/ijlter/article/download/7467/pdf
-What Were You Thinking? Medical Students’ Metacognition and Perceptions of Self-Regulated Learning,Marjolein Versteeg,"Abstract Phenomenon As a component of self-regulated learning, metacognition is gaining attention in the medical education research community. Metacognition, simply put, is thinking about one’s thinking. Having a metacognitive habit of mind is essential for healthcare professionals. This study identified the metacognitive competencies of medical students as they completed a conceptual learning task, and provided insight into students’ perceptions of self-regulated learning in their curriculum. Approach: Eleven third-year medical students from a Dutch University were purposively sampled to participate in this qualitative study. The study design included a think-aloud assignment followed by a semi-structured interview. During the assignment, participants were instructed to think aloud while solving questions about medical physiological concepts such as blood flow, pressure, and resistance. Think-aloud data were collected through audiotaping and used to identify participants’ metacognitive competencies. The assignment also served as a prompt for an interview in which participants were questioned about metacognitive knowledge, monitoring, experiences, and perceptions of self-regulated learning in their curriculum. All data were transcribed verbatim and analyzed iteratively using a template analysis. Findings: Students differed in their use of metacognitive skills, with an overall focus on monitoring and, to a lesser extent, on planning and evaluation. Additionally, differences were found in students’ metacognitive knowledge and metacognitive experiences. There was apparent use of inefficient, superficial predictive cues. Regarding perceptions of self-regulated learning skills, some students felt no need to develop such skills as they perceived medical education as an exercise in memorizing facts. Others emphasized the need for more insight into their actual level of knowledge and competence. Insights: Pre-clinical medical students require explicit teaching of metacognitive skills to facilitate self-regulated learning. Educators should aim to integrate metacognition in the everyday discourse of the classroom to foster an environment in which students discuss their own learning. Supplemental data for this article is available online at https://doi.org/10.1080/10401334.2021.1889559.","{'model': 'tldr@v2.0.0', 'text': 'The metacognitive competencies of medical students as they completed a conceptual learning task were identified, and insight into students’ perceptions of self-regulated learning in their curriculum was provided.'}",
-Creativity in Australian health and physical education curriculum and pedagogy,Rosie Welch,"ABSTRACT Notions of creativity are increasingly central to educational scholarship and policies, but few studies attend to the intersection between Health and Physical Education (H/PE) and creativity. In this paper we explore the literature on creativity in education and identify how creativity is represented across a broad spectrum of enquiry yet limited in relation to specifically H/PE. Writing from an Australian context, we examine how the key ideas and propositions from the Australian Curriculum for Health and Physical Education and the Critical and Creative Thinking capability raise important questions regarding the social emergence of educative creative purpose and potential in H/PE. We outline the utility of a ‘creative ecologies’ conceptual framework to consider contemporary practices in H/PE, and apply this to two purposively selected resources; to showcase their contextual eminence for creative learning in H/PE pedagogy and practice. The first resource, Phenomenom!, was funded by Horticultural Innovation Australia to develop food literacy and is linked to cross-curriculum learning outcomes. The second resource, Fitter. Faster. Better, is a St Martin's Youth drama-arts performance where students were prompted to design and perform a fitness programme for adults. Utilising the creative ecologies framework we illustrate the complexity of interconnected creative processes in H/PE as a series of networked elements and cultural flows (policies, practices, products, process, material environments, and partnerships). This framework is then used to identify how the two educational resources exemplify and open up a socially creative practice in both H/PE and everyday life. Our theoretical developments offer new ways to identify, design and enact quality creative and contemporary socio-cultural H/PE curriculum within a broader creative ecology.",,
-Self-Regulated e-Learning in Pre-Service Teacher Training for Realities of 21st Century Classrooms,Osman Sadeck,"Educators and students were unprepared for the suspension of face-to-face (f2f) educational activities due to Covid-19, specifically those less experienced in online teaching and learning. Students and educators were traumatised by the sudden switch to online teaching and learning. As such the transition from f2f to exclusive online learning prompted adjusted pedagogical methods and assumed measures of self-regulated e-learning (SRL). During this period researchers embarked on a longitudinal project in Information and Communication Technologies (ICT) called the 21st Century Project (21CP). The purpose of the study was to explore the SRL behaviours of pre-service teachers in online learning and to understand the extent to which a curriculum for technology integration alongside contextual factors influences SRL. As such our conceptualisation of the SRL framework adds two moderating constructs, these are, context and intervention to the core constructs of SRL. The sample of sixty six (66) students were drawn from a volunteer cohort of 166, 4th year pre-service students. Data were collected through interviews, surveys and online journal entries. The data was analysed qualitatively using narrative methods in which themes were identified and reported. The findings revealed that: the ICT integration curriculum represented a cornerstone for SRL development and shaped students’ SRL behaviours; there were differences in the SRL practices that could be attributed to students’ learning habits and attitudes to the ICT-based interventions. The findings of this study provide an understanding of interrelationships among SRL, context, and the design of an online curriculum. The study made three contributions to policy and practice. Firstly, SRL can be improved by providing activities that include clear guidelines for engagement through guided instructional methodologies. Secondly, the faculty curriculum developers should formulate guidelines to ensure that curricular iterations are developed as blended f2f/online modes to enable a quick and seamless transition for exclusive online use. Finally, course creators can improve student engagement by aligning learning outcomes and related activities with learning events.",,https://academic-publishing.org/index.php/ejel/article/download/2659/2148
-Technology adaptation of teachers and students under the learning continuity plan: A case of one school in the Philippines.,Maricel Sacramento,"The COVID-19 pandemic has prompted educational leaders to restructure the curriculum and modify the instructional set-up to accommodate remote learning of which using technology is the most viable solution to the existing problem. This study explores how teachers adapt and utilise technology-based teaching, and what makes students learn under blended learning modalities in Taytay Senior High School. Quantitatively, using the validated survey questionnaire anchored on the technology adaptation model and the adaptive learning environment model, this study revealed that teachers' age is the factor in all aspects of the model (performance and effort expectancy, social influence and facilitating conditions). When comparing the adaptation levels of students and teachers, it showed that teachers were slightly higher than the students, and that there is a negligible correlation. The findings of this study will serve as baseline data for immediate actions for items that surfaced concerns as hindrance or factors that can hamper students’ academic performance. Keywords: Technology adaptation, online classes, remote learning, Senior High School, Philippines. ",,
-Reflective writing in dental education to improve critical thinking and learning: A systematic review.,Janet L. Woldt,"The Commission on Dental Accreditations states that ""graduates must be competent in the use of critical thinking and problem-solving."" With this in mind, dental education programs continually strive to enhance and deepen these skills by incorporating effective instructional strategies into the curriculum. To do this, predoctoral dental education has used a variety of techniques including problem-based learning, OSCEs, and standardized patient activities. Another technique, while less popular but potentially more effective if done well, is reflective writing. This study aimed to assess the effect of reflective writing on students' critical thinking skills and learning in dental education programs. Two database searches conducted between 2000 and 2019 resulted in 317 articles after the initial screening. Ultimately, 13 articles met inclusion criteria. The following recurring factors were identified for inclusion in dental education reflective writing initiative: reflective writing interventions, reflective writing instructions, writing prompts, evaluation frameworks, and negative perceptions of reflective writing. In answering the key aim of this review, 12 of the 13 studies determined that reflective writing had a positive impact on students' critical thinking, judgment, and/or learning. Key findings included positive effects on student growth in reflection, learning through reflective writing, reflection skills, self-assessment, critical thinking, clinical reasoning, problem solving, and motivation to change after evaluated experiences. The review indicated that dental education programs, which implemented reflective writing as an assessment tool within the curriculum as a means of developing and deepening critical thinking skills and learning were by and large successful in this effort.","{'model': 'tldr@v2.0.0', 'text': 'The review indicated that dental education programs, which implemented reflective writing as an assessment tool within the curriculum as a means of developing and deepening critical thinking skills and learning were by and large successful in this effort.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/jdd.12561
-Transitioning an Advanced Practice Fellowship Curriculum to eLearning During the COVID-19 Pandemic.,Lisa Pierce,"BACKGROUND The COVID-19 global pandemic brought mandatory shelter-in-place orders, disrupting traditional face-to-face teaching methods for advanced practice fellowship programs nationally, creating a challenge for fellowship program directors to preserve curriculum using nontraditional methods. METHOD Using a variety of both web-based and app-based virtual platforms, a nationally accredited fellowship program converted traditional in-seat learning modalities to elearning platforms using both synchronous and asynchronous education. RESULTS Preliminary data indicate that knowledge acquisition and perceived fellow satisfaction are preserved despite the abrupt change to program delivery. Programmatic modifications were submitted to the American Nurse Credentialing Center for compliance and deemed as creative, innovative, and collaborative. CONCLUSION Curriculum for advanced practice fellowship programs can be favorably converted to elearning using virtual platforms during a crisis. Through prompt reevaluation and restructuring, virtual platforms can replace in-seat didactic lectures, patient case studies, mentoring, and even simulation, while ensuring program continuation and compliance with accreditation standards. [J Nurs Educ. 2020;59(9):514-517.].","{'model': 'tldr@v2.0.0', 'text': 'Curriculum for advanced practice fellowship programs can be favorably converted to elearning using virtual platforms during a crisis, and virtual platforms can replace in-seat didactic lectures, patient case studies, mentoring, and even simulation, while ensuring program continuation and compliance with accreditation standards.'}",
-Hands-on Cybersecurity Curriculum Using a Modular Training Kit,Mohammad Nasim Imtiaz Khan,"There is an exponential growth in the number of cyber-attack incidents resulting in significant financial loss and national security concerns. Secure cyberspace has been designated as one of the National Academy of Engineering (NAE) Grand Challenges in engineering. Broadly, the security threats are targeted on software programs, operating system and network with the intention to launch confidentiality, integrity and availability violations. Existing undergraduate and graduate-level cybersecurity education curriculum rely primarily on didactic teaching methods with little focus on student centered, inquiry-based teaching, known to improve student learning. With growing number of security incidents taking place, it is of utmost importance to prepare a workforce equipped with knowledge of the threat space and existing state-of-the-art solutions. Such comprehensive understanding is only possible by a dedicated hands-on course on cybersecurity where students can learn the key concepts by editing the hardware, software and OS, and, network policies. Unfortunately, such extensive and deep flexibilities are not provided in current cybersecurity curriculum. In this paper, we introduce a hands-on and modular self-learning Cybersecurity Training (CST) Kit to advance cybersecurity education. Students can promptly apply newly acquired knowledge on the CST Kit as part of the learning process. This Kit accompanies Do-It-Yourself (DIY) training modules that is used to model and investigate cybersecurity issues and their prevention to all levels of the cybersecurity workforce, including undergraduate and graduate students and K-12 science and technology teachers. The Kit also covers various aspects of cybersecurity issues including, hardware, software, operating system and network security. A coursework has been developed on hardware security for senior undergraduate and graduate students using the Kit. A preliminary survey conducted among students who were introduced to the modular board to implement hardware security threats such as, side-channel attack shows an 120% improvement in their understanding after the CST Kit based activities. The components of the CST Kit have also been used in a 4-day summer workshop for K-12 teachers. Teachers took pre-and post- concept inventories to assess their learning of content throughout the workshop and the results indicated improvement of 58%. These assessments focused on vulnerabilities and specific types of attacks, system security, data transmission and encryption, permutations and combinatorics, and binary numbers.","{'model': 'tldr@v2.0.0', 'text': 'A hands-on and modular self-learning Cybersecurity Training (CST) Kit that accompanies Do-It-Yourself (DIY) training modules that is used to model and investigate cybersecurity issues and their prevention to all levels of the cybersecurity workforce, including undergraduate and graduate students and K-12 science and technology teachers.'}",
-Mediation and experiential learning: how a mediation clinic can inform a law-based curriculum,B. Waters,"The perception that students of Law and Legal Studies should learn about a variety of methods of dispute resolution and not just litigation, has prompted the Department of Law and Criminal Justice Studies at Canterbury Christ Church University in Kent UK to establish a mediation clinic as a focus for undergraduate experiential learning. This article will consider the importance of discipline based research and the integration of clinical legal education within the core curriculum, the benefits offered by a combined live and simulated curricula approach in the context of mediation and the importance of providing a practical input during the academic stage of legal education.",,https://www.northumbriajournals.co.uk/index.php/ijcle/article/download/48/52
-Collaborative Online International Learning: Language and Cross-Cultural Experiences of University Students,Frank Ramirez-Marin,"This qualitative study reports on a collaborative online project between the University of Washington at Tacoma (USA) and the Universidad Veracruzana at Veracruz City (México). The project was implemented as part of the internationalization of higher education policies of the participating universities, which include pedagogical practices oriented toward foreign language learning, the internationalization of the curriculum, and virtual exchange. The study documents cross-cultural learning experiences between two groups of students from different linguistic and cultural backgrounds in a Collaborative Online International Learning approach (COIL); and how those experiences related to the development of cross-cultural competence. The language of instruction was English and the students interacted through an online platform and a social network. The methodological design was qualitative-interpretive. Data was generated using interviews, linguistic samples (writing samples), and the interactions of the participants prompted by the use of an online platform and a social media network. Data analysis was realized through a content approach, which led to the formulation of assertions based on themes that emerged. Results indicate that the pedagogical approach implemented (COIL) prompted reflexion on issues related to language learning, cultural understanding, and common life experiences, and that it was conducive to the development of aspects of cross-cultural competence.",,https://revistas.unal.edu.co/index.php/male/article/download/92144/81604
-Students Engagement and Development of Generic Skills in Gamified Hybrid Service-Learning Course,Azizah Yusof,"Hybrid Service-Learning, an emerging trend of combining co-curriculum course embedded with service-learning instructional strategies in online settings, has increasingly gained popularity, as many service-learning educators are moving into online platforms. The ongoing COVID-19 pandemic has accentuated its needs, thus forcing education systems worldwide to find alternatives to face-to-face instructions, one of them being Hybrid Service-Learning. This has prompted many researchers to attempt utilising gamification in online courses to increase students’ engagement. The purpose of this research was to report a gamified Hybrid Service-Learning (GAMYSEL) through students’ level of engagement proposed by Schlechty, used as a data classification to compare students’ scorings in their earning of points, badges, level and leaderboards. The results of the study indicated that, the students’ levels of engagement were relatively high through Advocacy Hybrid-Service Learning approach, where majority of them achieved the Strategic Compliance and Optimal Engagement categories. The students’ generic skills were further analysed with one-way repeated measures ANOVA and it was discovered that, GAMYSEL had positive impacts in developing these skills in five stages of service-learning. Both of these findings suggest that student’s engagement and generic skills had a significant effect on students learning after the use of GAMYSEL.",,
-Homeschooling for Quarantined Residents: A Virtual Back to the Basics Curriculum,Anthony Sielicki,"Introduction The COVID 19 pandemic resulted in local and institutional restrictions with significant effects on the clinical environment for graduate medical education, displacing residents from non-emergency medicine (EM) based rotations. Additionally, resident physicians considered patients under investigation (PUI) were furloughed from clinical practice. The necessity for supplemental learning in a virtual setting prompted the development of an online homeschooling curriculum that incorporated back to the basics textbook learning, application, and retention via virtual sessions for the quarantined and furloughed learners. Methods An online homeschooling curriculum was developed to replace the cancelled clinical experiences for EM residents and for those who were quarantined utilizing Google Classroom and Zoom teleconference software. After completion of their quarantine or return to normal rotation schedule, residents were asked to evaluate the homeschooling curriculum using an anonymous survey. Results A total of 12 residents participated in the homeschooling program over eight weeks during the spring of 2020. Of the nine residents surveyed, 88.8% percent felt the homeschooling added to their knowledge of EM, 100% found the online format easy to use, and 88.8% stated it helped maintain a sense of social connection to peers and faculty. Conclusion An online homeschooling program was considered an effective means of providing an opportunity for synchronous and continuous education for EM resident physicians. This program could be sustainable long term to fill in knowledge gaps or supplement remediation in emergency resident education, post pandemic.","{'model': 'tldr@v2.0.0', 'text': 'An online homeschooling program was considered an effective means of providing an opportunity for synchronous and continuous education for EM resident physicians and could be sustainable long term to fill in knowledge gaps or supplement remediation in emergency resident education, post pandemic.'}",https://www.cureus.com/articles/40374-homeschooling-for-quarantined-residents-a-virtual-back-to-the-basics-curriculum.pdf
-Ants as an Experiential Learning Strategy in Preschool Teacher Training,Rosa María Gálvez Esteban,"In this chapter, the authors present the results of a project designed for 41 preservice preschool teachers to introduce the concept of living things as an experiential learning strategy in the classroom. The need to approach this concept from a different perspective prompted the design of an education project involving the introduction of insects into classroom as a teaching resource. An informative storyline was used for project launch presentation. The questions they strive to answer in this chapter are related with what concepts of living organisms and what inquiry stages will preservice teachers consider their pupils will carry out during the project. Relevant concepts that are usually not much covered in the preschool curriculum such as the life cycles of animals were considered by 23 participants. Twenty-five of the future teachers claimed that they would be able to work on every inquiry step if they implemented this project in the classroom.",,
-The COVID-19 Elective for Pediatric Residents: Learning About Systems-Based Practice During a Pandemic,Maya S. Iyer,"Background: The coronavirus disease 2019 (COVID-19) pandemic has prompted pediatric residency programs to adjust the delivery of educational curricula and to update content relevant to the pandemic. Objective: In this descriptive paper, we present how we rapidly developed and implemented a COVID-19 pandemic elective for pediatric residents. Methods: This curriculum was established at a single tertiary care children’s hospital in June 2020. We used the ADDIE (analysis, design, development, implementation, evaluation) framework to develop a two-week elective (30 hours) consisting of six flexibly scheduled modules. We administered post-elective surveys and exit interviews to solicit feedback to improve the elective and obtain effectiveness of our educational interventions. Results: We developed an asynchronous online COVID-19 Elective for Pediatric Residents. The curriculum modules focus on pathophysiology of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the disaster management ecosystem, simulation of clinical care, mental health ramifications, and public health consequences. We also include six in-situ experiences (visits to a drive-through COVID-19 testing site, testing laboratory and local public health department, a simulation of a critically ill child, and meetings with emergency managers and social workers) to solidify learning and allow for further reflection. To date, eight participants have taken the elective. All participants strongly agreed on a five-point Likert item survey that the elective enhanced their knowledge in current evidence-based literature for COVID-19, disaster preparedness, hospital response, management of the critically ill child, and mental and public health ramifications. All participants agreed this curriculum was relevant to and will change their practice. Conclusions: We demonstrate how a COVID-19 elective for pediatric residents could be quickly developed and implemented. The pilot results show that pediatric trainees value asynchronous learning, supplemented by relevant in-situ experiences. Moreover, these results suggest that this curriculum provides needed disaster response and resiliency education for pediatric residents.","{'model': 'tldr@v2.0.0', 'text': 'The pilot results show that pediatric trainees value asynchronous learning, supplemented by relevant in-situ experiences, and suggest that this curriculum provides needed disaster response and resiliency education for pediatric residents.'}",https://assets.cureus.com/uploads/original_article/pdf/50576/20210305-32377-1ulh3cm.pdf
-A case study of curriculum development: Backward or forward/central design?,N. T. Tung,"curriculum development, backward design, constructive alignment, objective, syllabus, methodology, evaluation Nowadays universities in Vietnam have begun to update their curricula by adopting a backward design with a focus on students’ learning outcomes to replace the old-fashioned forward model. However, to have a constructive alignment is a problem they have to face. This situation has prompted this case study research in May 2020, intending to examine a curriculum to find out whether it is coherent in terms of its main components – objectives, syllabus, methodology, and evaluation. The findings of the study indicate that for this case, although the curriculum was claimed to be of a backward design with learning outcomes as program objectives, it turned out to follow a forward one with more focus on knowledge transfer than competency development. In particular, as the program learning objectives were still written according to knowledge transmission, it was not constructively aligned with the three main components of syllabus, methodology and evaluation: The syllabi were mainly based on the experience of the experts in the field or syllabus designers, the innovation of the methodology as directed by the university with a shift of focus from the lecturer to learners was still interpreted rigidly and mechanically, and the evaluation of student learning was generally claimed to be criterionreferenced only without any elaboration for each syllabus.",,https://journalofscience.ou.edu.vn/index.php/soci-en/article/download/546/478
-Letter to the Editor Regarding “The Evolving Impact of COVID-19 on Medical Student Orthopedic Education: Perspectives From Medical Students in Different Phases of the Curriculum”,Shi-qi Wang,"Letter to the editor Richardson et al. recently published a paper in Geriatric Orthopaedic Surgery & Rehabilitation entitled “The Evolving Impact of COVID-19 on Medical Student Orthopedic Education: Perspectives From Medical Students in Different Phases of the Curriculum.” Richardson et al. mentioned that the COVID-19 pandemic has changed medical students’ orthopedic education curriculum and provided methods that may help students seek alternative programs and supplementary learning. These are very enlightening to the practice-based orthopedic rehabilitation medical education. During the early stage of the COVID-19 epidemic, medical colleges and universities in China also implemented similar management measures, including reducing or even stopping students’ clinical rotation and adopting network-based professional course teaching. However, in the process of improving the teaching methods, we also found the corresponding problems. First of all, the efficiency of online teaching and online academic lectures is challenging to be guaranteed. During the epidemic, online teaching and online academic lectures have broken the restrictions of location. Although they can effectively reduce the crowd gathering and the risk of infection, they will also greatly reduce some students’ learning efficiency with poor self-consciousness. The live broadcast software prompts students to be in a “listening state,” but they may be doing things that have nothing to do with teaching in the background and cannot be effectively supervised. In addition, some elderly teachers are not proficient in the use of teaching software, which further leads to the decline of students’ learning efficiency. During the teaching period, the department of rehabilitation medicine of our school received much negative feedback from students that the online teaching experience was inferior. This result is similar to that of Jordan, where students’ satisfaction with online teaching is very low. Secondly, the reduction of practical operation makes it difficult for students to adapt to clinical practice. Orthopedic sports rehabilitation is a subject that includes many clinical operations, such as joint mobilization, muscle strength exercise, and balance training. However, due to the lack of practical clinical teaching, after returning to the clinic in May, the proficiency of the current students is significantly lower than that of the previous students. We also asked students to watch the operation video repeatedly during the epidemic, and there were conditions to carry out simple action exercises between students, but the final effect was not satisfactory. From the above problems, it is not difficult to find that the self-consciousness of students plays a key role during the epidemic. The 5 solutions mentioned by Richardson et al. also depend on students’ self-consciousness, especially in reading important professional articles. In addition, online teaching","{'model': 'tldr@v2.0.0', 'text': 'The COVID-19 pandemic has changed medical students’ orthopedic education curriculum and provided methods that may help students seek alternative programs and supplementary learning, but from the corresponding problems, it is not difficult to find that the self-consciousness of students plays a key role during the epidemic.'}",https://journals.sagepub.com/doi/pdf/10.1177/2151459320973191
-Effectiveness of the Proposed Robotics Instructional Design on “Basic Movements” for Online Distance Learning: Determining the effectiveness of the proposed RID module through collaborative action research,Genevieve A. Pillar,"Robotics brings education technology to a new level. Introducing robotics to schools does not only mean making hands-on activities but to prepare students for the future. The creation of Robotics Online Distance Learning at the De La Salle Santiago Zobel (DLSZ) school, and its integration to the curriculum prompted the researchers to come up with an evaluation to determine the effectiveness of the proposed Robotics Instructional Design (RID) module for “Robotics Basic Movements” facilitated through Online Distance Learning (ODL). This collaborative action research involved Grade 3 (N1=224) and Grade 4 (N2=215) students of DLSZ who served as the participants to the proposed RID module. Results from this study revealed that In terms of learning modalities, majority of the students were able to manipulate the apps using their available equipment at home. The Grade 3 group 0.377 standard deviation from the average mean of 15.193 indicates the better performance. It is also seen that based on the computed Grade 4 group results, the standard deviation of 1.07 with the average mean of 4.235 is higher compared to the other group. Most of the performances are close to the expected average, causing a symmetrical data distribution. In terms of learning targets, these became helpful for the teachers working with the students. These made lessons and activities easier to plan, monitor, and assess students’ learning. These also guided instructional designers to make sure that all students would reach mastery level. It is suggested that the proposed RID module design should involve processes from designing, to building and programming. This innovative way of transmitting skills revealed to be useful for all robotics competencies.",,
-Student Voice in an Extended Curriculum Programme in the Era of Social Media: A systematic Review of Academic Literature,J. Chukwuere,"Student voice in this digital age and across higher learning institutions is increasing exponentially with the function of social media. Student voice provides a vibrant communication pathway to extended curriculum programme students in higher education institutions. Social media ensures active participation of extended curriculum programme students in generating views and ideas that define the higher learning environment and experience towards better learning conditions and outcomes. A systematic literature review was used in gathering scientific papers through trusted academic databases. The systematic literature review was conducted between the period of 1 April 2019 and 28 September 2019, by looking into the contents of articles covering the current research objectives. The study’s findings show that social media provides an effective and instant spread of the extended curriculum programme students’ voice across higher education learning institutions. It also allows the students in the extended curriculum programme to engage with each other and the institutional management promptly. Social media promotes extended curriculum programme students’ voice in reaching the right audience at the right time. The results of this study are key for extended curriculum programme students, lecturers, and university management in understanding and applying social media effectively and in bringing transformation to South African higher education institutions and beyond.",,
-STEAM education: student learning and transferable skills,M. Bertrand,"PurposeGlobally, interdisciplinary and transdisciplinary learning in schools has become an increasingly popular and growing area of interest for educational reform. This prompts discussions about Science, Technology, Engineering, Arts and Mathematics (STEAM), which is shifting educational paradigms toward art integration in science, technology, engineering and mathematics (STEM) subjects. Authentic tasks (i.e. real-world problems) address complex or multistep questions and offer opportunities to integrate disciplines across science and arts, such as in STEAM. The main purpose of this study is to better understand the STEAM instructional programs and student learning offered by nonprofit organizations and by publicly funded schools in Ontario, Canada.Design/methodology/approachThis study addresses the following research question: what interdisciplinary and transdisciplinary skills do students learn through different models of STEAM education in nonprofit and in-school contexts? We carried out a qualitative case study in which we conducted interviews, observations and data analysis of curriculum documents. A total of 103 participants (19 adults – director and instructors/teachers – and 84 students) participated in the study. The four STEAM programs comparatively taught both discipline specific and beyond discipline character-building skills. The skills taught included: critical thinking and problem solving; collaboration and communication; and creativity and innovation.FindingsThe main findings on student learning focused on students developing perseverance and adaptability, and them learning transferable skills.Originality/valueIn contrast to other research on STEAM, this study identifies both the enablers and the tensions. Also, we stress ongoing engagement with stakeholders (focus group), which has the potential to impact change in teaching and teacher development, as well as in related policies.",,https://www.emerald.com/insight/content/doi/10.1108/JRIT-01-2020-0003/full/pdf?title=steam-education-student-learning-and-transferable-skills
-Semantic Mapping of Learning Assets to Align Curriculum and Evidence Learning Effectiveness in Business Education,Chad Patrizi,"Ensuring the alignment of course content against desired goals and objectives has always been at the core of effective instructional course design—whether the instruction is delivered face-to-face, or via the support of distance technologies. Nonetheless, with the latter delivery of instruction, two distinct challenges have recently emerged, thus prompting the need to re-examine the term quality as it relates to accreditation. The first challenge lies within the ability to locate and/or create and align digital learning objects to course goals and objectives. The second challenge lies within the ability to demonstrate learning effectiveness from learning management system metadata. A research and development team at American Public University System addressed these challenges by employing an open source repository and semantic engine for analysis and alignment of content, materials, and learning activities to goals and objectives across all courses within the School of Business. The result was a highly detailed, accurate mapping of the programs' knowledge base to established goals and objectives. As an added benefit, resources were disaggregated to a very granular level and sorted into taxonomies that can be independently referenced and utilized for cross-curricular consumption.","{'model': 'tldr@v2.0.0', 'text': 'A research and development team at American Public University System addressed challenges by employing an open source repository and semantic engine for analysis and alignment of content, materials, and learning activities to goals and objectives across all courses within the School of Business.'}",
-Facilitating Virtual Medicinal Chemistry Active Learning Assignments Using Advanced Zoom Features during COVID-19 Campus Closure,Mudit K. Singhal,"Active student participation is critical to the delivery of medicinal chemistry components of the pharmacy school curriculum. Carefully structured active learning strategies can promote student engagement while increasing course material application. Topic-based active learning assignments can be administered, where learners apply constructed knowledge and skills to complete assignments. The suspension of on-campus classes due to the COVID-19 pandemic prompted the development of innovative ways to deliver active learning assignments without using the school’s physical team rooms. To actively engage learners, the Zoom video teleconferencing platform with breakout rooms was used to conduct medicinal chemistry active learning assignments in a Pharmacotherapeutics IV: Endocrinology course. The students’ assignment scores showed no significant difference when assignments were conducted virtually using Zoom breakout rooms versus on campus, suggesting successful virtual administration of the assignments and students’ prompt adaptation to the new environment. The combination of Zoom-based lectures with actively facilitated Zoom breakout room assignments is an effective active learning strategy.","{'model': 'tldr@v2.0.0', 'text': 'The combination of Zoom-based lectures with actively facilitated Zoom breakout room assignments is an effective active learning strategy, and the students’ assignment scores showed no significant difference when assignments were conducted virtually using Zoom breakout rooms versus on campus.'}",
-The uses of Facebook© technologies in Hospitality curriculum on an experiential learning platform for a new generation of students,Christopher W. Harris,"Purpose – The purpose of this paper is to examine the applications of the social networking software, Facebook, currently practiced by “expert user” faculty teaching within a Malaysian Higher Education Hospitality Diploma programme. Of particular interest is whether the faculty experiences with Facebook reveal its functional appropriateness for use on an experiential learning platform. The aim is also to prompt further research and experimentation with the medium on teachers and students.Design/methodology/approach – Phenomenological: the researcher uses a focus group to enable the essence of teachers' experience in the utilisation of Facebook as a teaching and learning tool within the syllabus of their taught subjects in a shared hospitality curriculum to emerge.Findings – The experiences divulged in the focus group reveal the use of Facebook as an “effective” medium for reflective purposes in relation to experiential teaching and learning activities and as a mechanism for reflective and extrapolative co...",,
-Does online engagement matter? The impact of interactive learning modules and synchronous class attendance on student achievement in an immersive delivery model,E. Goode,"One Australian public university is radically changing the way it delivers higher education, introducing a 6-week immersive scheduling delivery model across all units and courses. Despite the emerging success of block and immersive models for raising the performance of diverse student cohorts, the design factors underpinning positive outcomes are underexplored. This paper presents a mixed methods study of the impact and value of student engagement with interactive and responsive online content modules and synchronous classes in an immersive scheduling model. The findings indicate that behavioural engagement with online learning modules has a positive effect on academic success and is a significant predictor of a higher final score. Qualitative data indicate several attributes of high-quality online learning modules that students appear to associate with engagement and deeper learning in the immersive model: interactivity, media richness, constructive alignment, flexibility and responsiveness. Synchronous class attendance did not impact final scores; however, students nonetheless valued the opportunity to form safe and supportive communities of inquiry during classes. This study demonstrates that in times of increasing demand for more flexible learning, immersive scheduling models that are founded on active learning principles and embed interactive, responsive, media-rich online learning modules can improve student engagement and performance. Implications for practice or policy: Higher education practitioners should integrate interactive, responsive, media-rich and constructively aligned online learning modules into curricula. Synchronous active learning classes that create safe communities of inquiry should be offered alongside options for asynchronous participation. Low levels of engagement with online learning modules should prompt follow-up from educators to raise engagement and bolster academic achievement. Immersive delivery models are effective curriculum innovations that, when designed with interactive online modules, can support improved academic achievement. ",,https://ajet.org.au/index.php/AJET/article/download/7929/1922
-Enhancing the 4Cs among college students of a communication skills course in Tanzania through a project-based learning model,M. Saimon,,,https://link.springer.com/content/pdf/10.1007/s10639-022-11406-9.pdf
-Fully Immersive Virtual Reality for Total Hip Arthroplasty: Objective Measurement of Skills and Transfer of Visuospatial Performance After a Competency-Based Simulation Curriculum.,K. Logishetty,"BACKGROUND Fully immersive virtual reality (VR) uses headsets to situate a surgeon in a virtual operating room to perform open surgical procedures. The aims of this study were to determine (1) if a VR curriculum for training residents to perform anterior approach total hip replacement (AA-THR) was feasible, (2) if VR enabled residents' performance to be measured objectively, and (3) if cognitive and motor skills that were learned with use of VR were transferred to the physical world. METHODS The performance of 32 orthopaedic residents (surgical postgraduate years [PGY]-1 through 4) with no prior experience with AA-THR was measured during 5 consecutive VR training and assessment sessions. Outcome measures were related to procedural sequence, efficiency of movement, duration of surgery, and visuospatial precision in acetabular component positioning and femoral neck osteotomy, and were compared with the performance of 4 expert hip surgeons to establish competency-based criteria. Pretraining and post-training assessments on dry bone models were used to assess the transfer of visuospatial skills from VR to the physical world. RESULTS Residents progressively developed surgical skills in VR on a learning curve through repeated practice, plateauing, on average, after 4 sessions (4.1 ± 0.6 hours); they reached expert VR levels for 9 of 10 metrics (except femoral osteotomy angle). Procedural errors were reduced by 79%, assistive prompts were reduced by 70%, and procedural duration was reduced by 28%. Dominant and nondominant hand movements were reduced by 35% and 36%, respectively, and head movement was reduced by 44%. Femoral osteotomy was performed more accurately, and acetabular implant orientation improved in VR assessments. In the physical world assessments, experts were more accurate than residents prior to simulation, but were matched by residents after simulation for all of the metrics except femoral osteotomy angle. The residents who performed best in VR were the most accurate in the physical world, while 2 residents were unable to achieve competence despite sustained practice. CONCLUSIONS For novice surgeons learning AA-THR skills, fully immersive VR technology can objectively measure progress in the acquisition of surgical skills as measured by procedural sequence, efficiency of movement, and visuospatial accuracy. Skills learned in this environment are transferred to the physical environment.","{'model': 'tldr@v2.0.0', 'text': 'For novice surgeons learning AA-THR skills, fully immersive VR technology can objectively measure progress in the acquisition of surgical skills as measured by procedural sequence, efficiency of movement, and visuospatial accuracy.'}",
-Is Reflective Learning Visible in Online Discussion Forums for Medical Students on General Practice Placements? A Qualitative Study,Kelly Gillingham,"Abstract Problem: Reflection is important for lifelong learning and professional development, and discussion forums have been highlighted as a potential platform for enhancing reflection further through peer interaction and teacher feedback. Forums provide students on general practice (GP) placement the opportunity to engage in collaborative reflective learning despite their geographical isolation and disparate schedules. This case report seeks to explore whether online discussion forums demonstrate community presence and reflective learning among medical students on GP placement. Intervention: Online discussion forums were introduced into the curriculum for University of Auckland Year 5 and 6 medical students on their respective 4- and 6-week GP placement. Via asynchronous posts and comments, groups of eight students on average presented cases and experiences for discussion with peers and an overseeing GP faculty member. Context: Students were dispersed across a large geographical area and were unlikely to have peer contact in their assigned practices. Online discussion forums were implemented for a number of reasons, including facilitating reflective learning. An adapted community of inquiry framework was used to investigate the cognitive, social, and teacher presence elements visible in the discussion forum transcripts of unanimously consenting groups. Content analysis of the transcripts was performed to evaluate the presence and quality of reflective learning. Impact: The forums were predominately student-led with relatively scarce comments by GP faculty facilitators. The majority of cases and experiences presented related to clinical management of patients and dissonance triggered by the on-site supervisor’s actions. Ideas, knowledge, and understanding of presented encounters were shared and built on by positive and supportive interaction, broadening students’ perspectives and, at times, leading to the formation of solutions. However, students’ reflection was mostly superficial. Deep reflection was rarely present, and affective dimensions of reflection were disclosed relatively less than clinical cognitive aspects of reflection. Direct instruction, based on prior experience and pedagogical expertise, was the main form of comment made by faculty facilitators, who prompted with questions to a far lesser extent. Lessons Learned: Online discussion forums appeared to enable medical students on GP placement to interact positively with peers and faculty facilitators. However, deep reflection was not reached. Modifying the discussion forums to facilitate more peer interaction and addressing the barriers that limit faculty facilitation may encourage deeper and more affective reflections.","{'model': 'tldr@v2.0.0', 'text': 'Modifying the discussion forums to facilitate more peer interaction and addressing the barriers that limit faculty facilitation may encourage deeper and more affective reflections.'}",
-A Personalized Approach to Learning Across Time and Space,Rachel Fitzgerald,"In a higher education environment massively disrupted by the pandemic, the importance of exploring the efficacy of purposeful, flexible, online learning is essential. Research shows that blended approaches to learning can be active, flexible, and student-centred. However, this research suggests that without human intervention or a bespoke context, there can also be very poor engagement. As such, this study presents the successes and challenges of introducing a personalised mobile-learning resource in a supported, personalized environment. The resource was designed and developed in 2019 and integrated in 2020 and 2021 into a large (n=200+) communication design class at an Australian university. This paper evaluates the effectiveness of the app using measures drawn from Abdullah & Ward’s (2016) take on the Technology Acceptance Model. Their model, GETAMEL (general extended technology acceptance model for e-learning) measures technology acceptance factors in eLearning such as enthusiasm, self-efficacy, and student enjoyment of technology in relation to contextual purpose. This paper also considers the value of self-directed learning and co-curricular activities in the context of additional learning to support core curriculum. This paper presents an evaluation of student experience drawn from a class-wide survey and analysis of usage data. We conclude that the use of a personalised app fosters an autonomously supportive experience that leads to intrinsic motivation and improves overall learning (Baker & Goodboy, 2018). The app provides additional support for students unable to attend tutorials and enabled students to remain engaged and abreast of the general topics each week and provided some additional prompts about self-directed learning activities they could engage in, that they wouldn’t have otherwise had. Enjoyment, as one of the key factors of the GETAMEL model, was explicitly mentioned by students. While a personalised tool is labour intensive and expensive in cost and time, this study concludes that the rewards regarding student experience make it worthwhile.",,https://papers.academic-conferences.org/index.php/ecel/article/download/901/775
-Building Professional Competencies Through a Service Learning ' Gallery Walk' in Primary School Teacher Education,S. Ramsaroop,"This article reports on a service learning project in a South African primary school teacher education programme, as experiential and practice-based pedagogy in a social studies methods course. We aimed to broaden understanding of service learning as a form of non-placement work-integrated learning for the development of teacher professional competencies. Student teachers drew on topics in the middle school social studies curriculum and incorporated Indigenous geographical elements with local community history in the design of a service learning ‘gallery walk’ for Grade 5 learners. Using a generic qualitative design, data were generated from students’ and teachers’ reflective journals, lesson plans, photographs and video recordings. It was analysed for common content themes and prominent discourse markers of students’ developing professional knowledge and competencies. The findings provide evidence of deepened student learning, particularly on the influence of context and curriculum differentiation and how their struggles with group work enabled the development of collaboration and cooperation required by professionals. In addition, the service learning prompted changing notions of citizenship and reciprocity of learning.",,https://ro.uow.edu.au/cgi/viewcontent.cgi?article=2256&context=jutlp
-Robot open-Ended Autonomous Learning competition,Emilio Cartoni,"Open-ended learning, also called ‘life-long learning’ or ‘autonomous curriculum learning’, aims to program machines and robots that autonomously acquire knowledge and skills in a cumulative fashion. We illustrate the first edition of the REAL-2019 – Robot open-Ended Autonomous Learning competition, prompted by the EU project GOAL-Robots – Goalbased Open-ended Autonomous Learning Robots. The competition was based on a simulated robot that: (a) acquires sensorimotor competence to interact with objects on a table; (b) learns autonomously based on mechanisms such as curiosity, intrinsic motivations, and self-generated goals. The competition featured a first ‘intrinsic phase’, where the robots learned to interact with the objects in a fully autonomous way (no rewards, predefined tasks or human guidance), and a second ‘extrinsic phase’, where the acquired knowledge was evaluated with tasks unknown during the first phase. The competition ran online on AIcrowd for six months, involved 75 subscribers and 6 finalists, and was presented at NeurIPS-2019. The competition revealed very hard as it involved difficult machine learning challenges usually tackled in isolation, such as exploration, sparse rewards, object learning, generalisation, catastrophic interference, and autonomous skill learning. Following the participant’s positive feedback, the preparation of a second REAL-2020 competition is underway, improving on the formulation of a relevant benchmark for open-ended learning.","{'model': 'tldr@v2.0.0', 'text': 'The first edition of the REAL-2019 – Robot open-Ended Autonomous Learning competition, prompted by the EU project GOAL-Robots, is illustrated, based on a simulated robot that acquires sensorimotor competence to interact with objects on a table.'}",
-Contemplating curriculum in an urban world,Philip Roberts,,,https://link.springer.com/content/pdf/10.1007/s41297-023-00194-y.pdf
-"TRANSFORMATIONAL LEARNING: USING EQUALITY AND DIVERSITY MARKETING CLIENT BRIEFS TO FOSTER AWARENESS, APPLICATION AND ACTION",Sally Everett,"ABSTRACT This paper reports a study coauthored with second-year undergraduate students that examines student experiences of undertaking real-life, client-set marketing assessments with an equality and inclusion remit. Students were set a marketing assessment with an explicit social justice focus, thereby prompting them to reflect on their own backgrounds and unconscious biases whilst addressing issues of societal injustice. This study explores the impact on students of closely engaging with this material and how the assessment, which involved them interviewing experts and developing actionable recommendations, altered their views and career plans. Data were generated via focus groups with students who undertook the assessment, supplemented by some alumni and client interviews. The findings are presented as a “3A” framework of Awareness, Application, and Action: first, the assessment raised “Awareness” of broader societal issues and fostered transformation in terms of personal views; second, students developed core skills and networks they could “Apply” in their future careers; and third, “Action” describes the way students felt empowered to make a positive difference to their client and wider society. The 3A model provides simple design principles that educators could implement if they want to introduce transformational experiences based on social justice issues into the marketing curriculum.",,
-Learning developers as their own cultural critics?,Christie Pritchard,"The values that guide the work of ALDinHE and those associated with us as an organisation include ‘making HE inclusive through emancipatory practice, partnership working and collaboration’ and ‘critical self-reflection, on-going learning and a commitment to professional development’ (ALDinHE, 2022). However, considering this from a relational, or systems thinking perspective; how achievable are these values? Can we truly be inclusive to all? Exploring this notion in my doctoral research, using Bourdieu’s relational framework (Bourdieu, 1992; Bourdieu, 1993; Bourdieu, 1997; Bourdieu & Passeron, 1990) I was confronted with some uncomfortable truths. Education is reproductive in the sense that it selects those with the necessary capital to succeed and nurtures them to develop further. The diversity of student’s prior experience, background and capital to succeed is clear, and Learning Developers know how difficult Higher Education can be to navigate, but how often do we reflexively consider how our work reproduces the cultural system. Conference attendees working in small groups, discussed a brief precis of my Doctoral research, a copy of ALDinHE values and summary of Bourdieu’s notion of autonomy (Bourdieu, 1992). The following questions were posed to prompt discussion. How autonomous are we as practitioners and are we able to change the system that reproduces the inequality of society? Is being positioned by students as part of their curriculum through embedded practice advantageous all of the time? Do we focus on the knowledge and skills students bring, or are we forced to help them adapt to the game? ",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/967/631
-Development of a Novel Interactive Multimedia E-Learning Model to Enhance Clinical Competency Training and Quality of Care among Medical Students,Yu-ting Hsiao,"Clinical competencies consisting of skills, knowledge, and communication techniques should be acquired by all medical graduates to optimize healthcare quality. However, transitioning from observation to hands-on learning in clinical competencies poses a challenge to medical students. The aim of this study is to evaluate the impact of a novel interactive multimedia eBook curriculum in clinical competency training. Ninety-six medical students were recruited. Students in the control group (n = 46) were taught clinical competencies via conventional teaching, while students in the experimental group (n = 50) were taught with conventional teaching plus interactive multimedia eBooks. The outcomes of clinical competencies were evaluated using Objective Structured Clinical Examination (OSCE) scores, and feedback on their interactive eBook experiences was obtained. In the experimental group, the average National OSCE scores were not only higher than the control group (214.8 vs. 206.5, p < 0.001), but also showed a quicker improvement when comparing between three consecutive mock OSCEs (p < 0.001). In response to open-ended questions, participants emphasized the importance of eBooks in improving their abilities and self-confidence when dealing with ‘difficult’ patients. Implementing interactive multimedia eBooks could prompt a more rapid improvement in clinical skill performance to provide safer healthcare, indicating the potential of our innovative module in enhancing clinical competencies.","{'model': 'tldr@v2.0.0', 'text': 'Implementing interactive multimedia eBooks could prompt a more rapid improvement in clinical skill performance to provide safer healthcare, indicating the potential of the innovative module in enhancing clinical competencies.'}",https://www.mdpi.com/2227-9032/8/4/500/pdf?version=1605847025
-The Australian curriculum: Mathematics as an opportunity to support teachers and improve student learning,P. Sullivan,"The creation of a national, as distinct from state and territory based, mathematics curriculum creates important opportunities for improving learning, but whether those opportunities are taken up will depend on the ways that teachers are supported, including by teacher educators, coaches, school leaders and readers of this monograph. The debates that continue on aspects of content are irrelevant to whether the national curriculum provides a prompt to improvement. In fact, such debates are a by-product of the negotiations that are an obvious artefact of the ceding of responsibilities by local jurisdictions to a national authority. The real opportunities for improving mathematics learning are in the principles that underpin the structure of the curriculum and the use of these principles to inform teacher learning.",,
-Advocacy and Awareness: Integrating LGBTQ Health Education Into the Prelicensure Curriculum.,Paula McNiel,"BACKGROUND An identified gap in the curriculum related to lesbian, gay, bisexual, transgender, and queer (LGBTQ) health needs prompted nursing faculty to implement a collaborative educational offering. LGBTQ individuals experience significant health disparities, compared with heterosexual counterparts. Enhancing established LGBTQ population-specific training to highlight health disparities and awareness of special health care needs was piloted with two clinical groups of senior baccalaureate nursing students (N = 16). METHOD Didactic, simulated, and panel discussion related to LGBTQ terminology, current health standards of care, and the importance of advocacy was provided by campus advocates, experienced health care providers, and a student panel identifying as LGBTQ. Health specific learning outcomes were established and evaluated. RESULTS Posteducation, anonymous surveys, and journaling were completed. Survey respondents (n = 13) reported increased awareness and understanding of health disparities specific to the LGBTQ population. CONCLUSION LGBTQ-specific health education has been implemented as a permanent curriculum change. [J Nurs Educ. 2018;57(5):312-314.].","{'model': 'tldr@v2.0.0', 'text': 'enhancing established LGBTQ population-specific training to highlight health disparities and awareness of special health care needs was piloted with two clinical groups of senior baccalaureate nursing students.'}",
-Embodied cognition and curriculum construction,Meiqian Wang,"Abstract The disembodiment of cognitive science has resulted in curricula with disembodied concepts and practice. The emergence of the embodied cognitive science provoked public reflections on the nature of the curriculum. This has elevated the body from the ‘peripheral’ position to the ‘central’ position, acting as the subject in action and becoming the bridge to experience transformations. Meanwhile, the nurturing role of the environment for the mind is attracting increasingly more attention, and the environment, the body, and the mind jointly constitute a dynamic system that prompts the transformation and restructuring of learners’ experience. The reshaping of the relationship between the body and education and the integration of physical experience and mind-building have led us to examine curriculum from a completely new perspective, and curriculum construction is thus transformed to the dynamic generation of interaction among the body, the mind, and the environment based on group experiences. To be precise, there are three basic dimensions in the construction of an embodied curriculum: (1) returning to the life-world and promoting the integration of the curriculum narrative framework and the student’s learning trajectory; (2) introducing the embodied experience and realizing the continuous interactions among the learner’s body, mind, and environment; and (3) integrating the functions of technologies to provide a powerful mediating tool for the construction of embodied curriculum.",,
-Conversations about curriculum change: mathematical thinking and team-based learning in a discrete mathematics course,J. Paterson,"This article reports on the learning conversations between a mathematician and a mathematics educator as they worked together to change the delivery model of a third year discrete mathematics course from a traditional lecture mode to team-based learning (TBL). This change prompted the mathematician to create team tasks which increasingly focused on what he calls the ‘unspoken curriculum’: mathematical thinking. We consider the ways in which the TBL model promoted and enabled this in the light of literature on mathematical thinking, sense-making and behaviours, and strongly suggest that this approach warrants more attention from the mathematics teaching community. We also discuss shifts in the mathematician's thinking about task construction as he refined the tasks to encourage students to think and behave like mathematicians.",,
-Journal of Problem-Based Learning,Seung Wook Lee,"information, how they share it with others (or not), the way they learn together or apart, and how they support each other in ways that current classroom practices often prevent or discourage. is of which at to be hampered by ideas about what a curriculum should be, and how it should be managed. This is seen in over-managed and over- designed curricula, such as: be a The argument paper is that many PBL curricula focus on ‘tracing’ and ‘tracing other of PBL, what is needed is a shift imaginative curricula. It where PBL curricula what is getting in the way, how and PBL pedagogies might be delineated. Purpose: The purpose of this work was to describe and reflect on the worth of the iterative design of a valid teaching method for senior school level social studies classes reliant on PBL methods. Method: This descriptive reflection on the implementation of a Problem-Based Learning (PBL) activity involved a small elective class of ten Grade 11 World History students at Selwyn House School (SHS), in the Canadian Province of Quebec. Use of PBL was a five-step process, involving two different classes run concurrently, once with Cohort Group #1 (2017-18) in the spring of 2018, and twice with Cohort Group #2 (2018-19). Results: PBL processes evolved through three PBL examples involving twentieth-century historical situations, beginning with World War Two (1939-1945) and concluding with the Cold War (1946-1991). Design features and processes demonstrated enhanced outcomes from stu-dent-centered learning. Conclusion: The particular methods produced results that were evident in the students’ finished products. for ex-ample, found in this regard that a computer-based learning environment is ideal Case-based problem solving is a core approach to foster knowledge acquisition, especially during the learning process by which novices become more and more expert within a domain. This study investigated whether metacognitive scaffolding leads to better learning outcomes compared to learning without such support in a case-based learning environment. In addition, we examined the interaction between prior domain knowledge and prior metacognitive abilities and skills. Within a one-factorial design, we explored the role of metacognitive prompting during the learning phase. A pre- and post-test were applied assessing knowledge acquisition, metacognitive skills, and cognitive load. Results indicate no main effects for prompting, and no interaction effect between prior knowledge and prompting. Metacognitive prompting enables learners that already possess metacognitive abilities to activate these during problem-solving and, thus, score higher in the knowledge post-test than learners with low metacognitive abilities and no metacognitive prompting. to utilize the acquired knowledge effectively. Creation of new knowledge will arise from enquiry processes and, education should aim to develop this ability (Yoon, 2009). The purpose of contemporary education is to encourage self-directed learners with prob-Purpose: To i) use outcomes of professional development on PBL for curriculum renewal that leads to more active student learning ii) use academic and clinical peers to develop a learning package that integrates education and practice of Clinical Dental Hygienics and iii) present and evaluate the authentic cases used as stimulus material that is learner-centered. Methods: Professors of Dental Hygienics used professional development sessions to inform curriculum renewal and development of the stimulus material. Others with recent clinical experience used their expertise to assist with the development of learning packages. The PBL package was applied to two junior classes of Dental Hygienics and evaluated for perceptions of student satisfaction. Results: The journey through PBL Package development and implementation, using peer support through Professional Development, is reported on. Responses to statements from a survey on learning experience indicated there was a high level of student satisfaction with learning; after the application of PBL (91.4%) responded positively to the statement, “I actively attended this class”. Conclusions: It is necessary to develop more packages that reflect the field of Dental Hygienics and extend the use of PBL to verify its effect. Additional evaluation of the use PBL packages needs to focus on learning outcomes that demonstrate links to actual practice and students’ problem-solving, self-directed learning, and team-work abilities that can be cultivated through PBL methods. teachers out do the teachers work in but also the teachers do difference This paper addresses the current problems of STEAM education and suggests a solution to solve one of the problems through adopting a method from Flipped Learning. Through the meticulous literature review, comparison, and analysis on STEAM and Flipped Learning, it will be shown that the methodology of Flipped Learning complements the weaknesses in STEAM. The conclusion is that since there is a theoretical need for applying Flipped Learning to STEAM, further research needs to be carried out to apply flipped learning in STEAM classes.",,
-Learning to Be Lawyers: Professional Identity and the Law School Curriculum,Charlotte S. Alexander,"The Carnegie Report faults American legal education for focusing exclusively on doctrine and analytical skills and neglecting the formation of professional identity. According to the Report, law schools can fix this problem by enabling students to ""encounter appealing representations of professional ideals, connect in a powerful way with engaging models of ethical commitment within in the profession, and reflect on their [own] emerging professional identity in relation to those ideals and models."" The Report identifies pro bono work, clinics, and externships as sites for this sort of learning, where students can interact with members of the profession and reflect on the models of professionalism that they encounter. Taking the Carnegie Report’s charge as a starting point, this article proposes an additional model for integrating a focus on professional identity into the law school curriculum. It profiles an experimental law school course that combined field work observations of practicing attorneys with in-class simulations of the work of a small law firm. The course was quite successful in prompting students to engage in an inquiry into what it is to be a lawyer and what kinds of lawyers they wanted to be. One student commented in a course evaluation, for example, that the course allowed him to see ""a new vision for what being a practicing lawyer can be."" That this sort of exposure to professional exemplars and reflection on professional identity was possible in a non-clinic course was an exciting discovery, suggesting new directions for curricular design as law schools continue to meet the challenges of the Carnegie Report.",,
-Impacts of AIOT Implementation Course on the Learning Outcomes of Senior High School Students,Chih-Cheng Tsai,"In this study, experimental teaching was conducted through the artificial intelligence of things (AIOT) practical course, and the 4D (discover, define, develop, deliver) double diamond shape was used to design the course and plan the teaching content to observe the students’ self-efficacy and learning anxiety. The technology acceptance model (TAM) concept was used to explore learning effectiveness and satisfaction. A total of 36 Senior One students from a senior high school in Kaohsiung were selected as the research subjects in two classes per week for 13 weeks. Questionnaires and qualitative interviews were used to understand the correlation between students’ self-efficacy, anxiety, outcomes, and learning satisfaction with the AIOT course during the flipped learning process. The study used SmartPLS3 to analyze the measurement model and structural model, and bootstrapping to conduct the path analysis and validate the research hypotheses. Because this research provides students with relevant online teaching videos for linking and browsing in the textbooks, and the video time should not be too long, in the teaching process, small-scale online learning courses are adopted for flipped teaching, which promptly arouses students’ interest and enhances their learning participation. Through the sharing of homework with each other, its content deficiencies were modified, students’ anxiety in learning was reduced, and the effect of learning and thinking together was achieved; in the teaching mode, theoretical content was combined with physical and practical courses to implement cross-disciplinary. To learn, the principle of 4D double diamond design to make “reciprocal corrections” between curriculum planning and teaching implementation as the teaching model framework was used. The results of this study show ”self-efficacy” has a significant positive effect on the “perceived usefulness” and “perceived ease of use” of users. “Learning anxiety” does not significantly predict the “perceived ease of use” or “perceived usefulness” of flipped learning using online e-learning. The “perceived ease of use” and “perceived usefulness” have positive impacts on the prediction of “behavioral intention” in flipped teaching using online digital teaching materials. The “perceived ease of use” has a positive and high impact on “perceived usefulness”. “Learning engagement” positively affects students’ actual “behavioral intention” towards learning. Students are highly receptive to learning new knowledge about science and technology.",,https://www.mdpi.com/2227-7102/11/2/82/pdf?version=1614245403
-Disrupting colonial discourses in the geography curriculum during the introduction of British Values policy in schools,C. Winter,"ABSTRACT The main purpose of this article is to expose and disrupt discourses dominating global development in an English school geography textbook chapter. The study was prompted by a teacher’s encounter with cultural difference in a geography lesson in South Korea. I investigate the issues raised through the lens of a new curriculum policy in English schools called ‘Promoting Fundamental British Values’ which forms part of England’s education-securitisation agenda, a topic of international attention. Following contextualization across research fields and in recent curriculum and assessment policy reform, I bring together theoretical perspectives from curriculum studies and Continental philosophy that do not usually speak to each other, to construct a new analytical approach. I identify three key themes, each informed by colonial logic: ‘development’, ‘numerical indicators’ and ‘learning to divide the world’. The inquiry appears to expose a tension between the knowledge of the textbook chapter and the purported aims of the British Values curriculum policy, but further investigation reveals the two to be connected through common colonial values. The findings are relevant to teachers, publishers, textbook authors, policy-makers and curriculum researchers. I recommend a refreshed curriculum agenda with the politics of knowledge and ethical global relations at its centre.",,https://www.tandfonline.com/doi/pdf/10.1080/00220272.2018.1428366?needAccess=true
-Universal Design for Learning,J. Cressey,"Universal Design for Learning (UDL) is a framework for curriculum and instructional planning through which educators can maximize accessibility and minimize barriers that are often experienced by learners. Culturally responsive practices strengthen and complement UDL by framing accessibility as an equity goal and prompting educators to examine ableism, racism, and other structural inequities. Teacher educators are in a unique position to introduce UDL to future elementary teachers and support them in developing inclusive pedagogical methods early on in their careers. Education technology tools are used within UDL to make curriculum materials more accessible and engaging. In this chapter, the UDL framework will be described along with culturally responsive applications within elementary teacher education.",,
-Integrating Concept Maps into a Medical Student Oncology Curriculum,S. Brondfield,,"{'model': 'tldr@v2.0.0', 'text': 'Concept mapping was feasible and reliable as an assessment tool and student perception was mixed, and students provided feedback to improve CM utility as a learning tool.'}",
-Using the wider science curriculum to investigate fusion energy,Andrew Malcolm-Neale,"Fusion energy research is ‘mission oriented’ big-science. It has both rapid progress and a reputation for being forever distant that excites many with a curiosity for science. However, it is difficult to give time to in lesson plans given both its lack of associated learning objectives and complexity. Outreach at the Fusion Centre for Doctoral Training (FCDT) tries to make fusion accessible. We describe here a set of questions, which we use in workshops, that build a picture of fusion research. Each question integrates one or more aspects of the core science curriculum. We further share our experience of engagement with these questions and the rewarding experience pupils get on realising they can apply their knowledge to cutting-edge science. There is broad scope for using the worksheets. The available materials are a self-contained set of numerical questions that explore physics and fusion, providing all necessary equations and constants. The themes and physics of each sheet can be a springboard, though, for further exploration. Prompts for such discussion are suggested in this article for the first time, as they are not included in the available question set. Our questions, then, can be used to revise physics lessons in an interesting, unseen context; to explore the exciting world of fusion energy itself or as a workshop for the physics of energy. The full set of worksheets (one per question) is available upon request from outreach@fusion-cdt.ac.uk.",,
-Exploring the K-12 computer science curriculum standards in the U.S,Meize Guo,"In order to create early exposure and to guide more talent into computing, computer science was included in K-12 education as a new subject area in many states. Although each state has different educational goals, computer science curriculum standards ensure equality among the states, especially in guiding teaching and accessing students' achievements. In this study, we applied content analysis to investigate the published computer science standards to examine: (1) How did computer science standards fit within the whole curriculum? (2) What concepts were covered in the computer science standards in the U.S.? The results showed four approaches to housing computer science in the whole curriculum. The Computer Science Teachers Association Computer Science framework was broadly adopted. Moreover, three major concepts, computer literacy, digital citizenship, and computational thinking, were additionally emphasized at the state standard level. This study fundamentally reported K-12 CS promotion in the U.S. through an academic standard scope, which directs K-12 CS teaching and learning. Additionally, this study prompted a conversation of preparing for different kinds of computing skills at the K-12 level in the U.S. among the policy makers and scholars","{'model': 'tldr@v2.0.0', 'text': 'The results showed four approaches to housing computer science in the whole curriculum, and three major concepts, computer literacy, digital citizenship, and computational thinking, were additionally emphasized at the state standard level.'}",
-"Equality, diversity and inclusion: learning from laying our cards on the table (resource showcase)",Matthew Sillence,"In this showcase, we presented a card resource for learning developers who work with staff or students to set up an inclusive learning environment. These are often personal interactions and need to be explored respectfully (Love, Gaynor, and Blessett, 2016; Mahmood, Gray and Benincasa, 2022). The Equality, Diversity and Inclusion (EDI) cards were created by the University of Hertfordshire and comprise 26 topics and illustrations, ranging from ‘belonging’ to ‘trust’, and exist in both a physical and digital format. Resources used as games can encourage participation, reflection, and transformational thinking (Peabody and Noyes, 2017; Clark, Dickerson, and Jarvis, 2022), which are all important for encouraging respectful conversations about EDI issues. A collaborative project with the University of East Anglia to explore the use of the cards was funded by ALDinHE. Full ethics approval was gained from both universities before beginning. It recruited staff and students at Hertfordshire and East Anglia who worked with EDI issues and gathered reflections from facilitators and participants who used the cards in different learning contexts, such as academic support staff development exercises and student representation workshops. They reported that the card prompts allowed for personal storytelling and they suggested further topics for the cards set. This session, led by learning developers and teaching staff, explored the idea of the ‘brave space’ (Arao and Clemens, 2013, p. 142; Palfrey, 2017) and why ‘serious play’ is important in learning (Rieber, Smith and Noah, 1998, p. 29). It discussed where and when the cards can be used, in terms of: a physical or digital space; the curriculum or disciplinary context; and the stage of the learner’s journey. The EDI cards and new guidance booklet were made available for delegates to take away.",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/1115/846
-Critical Reflexivity in Indigenous and Cross‐cultural Psychology: A Decolonial Approach to Curriculum?,Emily Castell,"Critical reflexivity is a mechanism for working toward decoloniality in higher education, with the potential to prompt students' to critique the contexts in which they are embedded, and facilitate transformative learning. We present a critical examination of the tensions surrounding decoloniality and critical reflexivity in an undergraduate unit on Indigenous and cross-cultural psychology at a large Australian university. We invited students in the unit to participate in a written reflexive exercise at the beginning (N = 44) and end of semester (N = 23) and analyzed these reflections qualitatively for level (four-category scheme for coding) and content (causal layered analysis) of reflection. Findings suggest that, while students' primarily demonstrate reflective engagement at levels preordinate to critical reflexivity, they are also engaged in active and nuanced processes of negotiating discomfort and uncertainty in this space. We pose critical commentary on the notion of safety in teaching practice, and consider the role of the academic institution in parametrizing the decolonial stance. This research holds application and transferability to higher education settings, and for the enduring project of engaging a decolonial approach to the curriculum within psychology.","{'model': 'tldr@v2.0.0', 'text': 'A critical examination of the tensions surrounding decoloniality and critical reflexivity in an undergraduate unit on Indigenous and cross-cultural psychology at a large Australian university suggests students are engaged in active and nuanced processes of negotiating discomfort and uncertainty in this space.'}",
-How does STEM context-based learning work: what we know and what we still do not know,H. Sevian,"ABSTRACT Context-based learning (CBL) has influenced teaching and learning science in many countries over the past decades. Twelve years ago, a special issue on CBL was published in this Journal, focusing on CBL curriculum development. Seven papers in this current special issue on CBL now address the question of how a context influences the learning process. The papers focus on the stimulation of learning STEM subjects within contexts, how the learning process occurs and is enhanced, and the application of contexts in different settings. The approaches, results, and implications of the papers are located in a larger view that considers the question of what must be the case if a student not only engages in the tasks of learning but also succeeds at them. Concerning willingness and effort by learners, the papers draw conclusions about which STEM-related interests of students endure and are ephemeral across a decade, design criteria for maximising students’ situational interest, and students’ engagement with content and context simultaneously. Focusing on the opportunity to teach and learn, the papers reveal how a professional development approach functions to support STEM teachers to develop CBL materials, and how specific scaffolding acts in teaching bring students to more complex reasoning. Regarding good teaching, insights are offered on how metacognitive prompts improve teaching. Centring on the social surround that supports teaching and learning, a comparison of two contexts for teaching the same content reveals which aspects of the contexts move student learning forward. From this mapping, paths toward future research are projected.",,https://figshare.com/articles/journal_contribution/How_does_STEM_context-based_learning_work_what_we_know_and_what_we_still_do_not_know/6803186/1/files/12372731.pdf
-Simulations and Virtual Learning Supporting Clinical Education During the COVID 19 Pandemic,S. Tabatabai,"Abstract Currently, the COVID-19 pandemic has a critical impact on clinical education, and it has resulted in the widespread disruption of clinical assessment. Clinical mentors and students in all the health professions are working within the most troublesome of circumstances in the hospital-based educational settings. Medical educationists ought to concentrate on the health and the safety of the students and communities. The safety issues have prompted the Ministry of Health to suggest that schools develop action plans for the adoption of available technologies to keep medical education moving forward with high quality, active, and interactive learning for more demanding tomorrow. A key challenge for medical educators is to simulate the clinical encounters at this unprecedented time, and this emphasized the necessity of applying virtual simulation-based educational tools in clinical education. This commentary explores how COVID-19 has challenged medical education. It also has discussed the future implications and potencial challenges of incorporating simulation-based virtual learning technologies into the medical curriculum, for the future of clinical education, and students' or residents' competency evaluation.","{'model': 'tldr@v2.0.0', 'text': ""This commentary explores how COVID-19 has challenged medical education and the future implications and potencial challenges of incorporating simulation-based virtual learning technologies into the medical curriculum, for the future of clinical education, and students' or residents' competency evaluation.""}",https://www.dovepress.com/getfile.php?fileID=60342
-Becoming Doctors: Examining Student Narratives to Understand the Process of Professional Identity Formation Within a Learning Community,D. Hatem,"Background: Professional identity formation is a key aim of medical education, yet empiric data on how this forms are limited. Methods: Our study is a qualitative analysis of student reflections written during the final session of our Becoming a Physician curriculum. After reading their medical school admission essay and their class oath, students wrote about a “time, or times during your third year when you felt like a doctor.” The reflections were qualitatively analyzed by the evaluation team, looking for themes found in the reflections. Results: Narrative themes separated into 4 distinct categories, specifically that performing physician tasks can make one feel like a doctor, demonstrating caring is a fundamental task of doctors, integrating personal ideals with professional values promotes professional identity formation, and the theme of never feeling like a doctor. Subsets of these broad categories provide further insight into individual and integrative tasks. Patients, patient families, and students through their own reflection prompted learners to feel like doctors in 74% of narratives, whereas physicians or the care team did so in 26% of our narratives. Conclusion: Students are able to reflect on times during their principal clinical year where they feel like doctors, taking a step toward forming a professional identity. Having faculty prompt and support such reflection can help faculty understand the student experience of their principal clinical year and promote professional identity formation.","{'model': 'tldr@v2.0.0', 'text': 'Students are able to reflect on times during their principal clinical year where they feel like doctors, taking a step toward forming a professional identity.'}",https://journals.sagepub.com/doi/pdf/10.1177/2382120519834546
-Examining student work for evidence of teacher uptake of educative curriculum materials,A. Bismack,"The purpose of this study was to identify evidence in student work of teachers' uptake of educative features in educative curriculum materials. These are features in curriculum materials designed with the specific intent of supporting teacher learning and enactment. This study was prompted by previous work on educative curriculum materials and the need to determine how teachers' use of educative curriculum materials can influence student learning. Student work from two fourth-grade teachers' enactment of an electric circuits unit was analyzed for evidence of teachers' uptake of educative features, which included characteristics of quality for particular science practices. Findings from the student work revealed that the teachers used many of the supports in the educative curriculum materials, especially those that could be used directly with students. The student work also reflected characteristics of high-quality science practices, which were only supported within the educative features. This study supports and extends other work related to how teachers' use of educative curriculum materials may influence student learning and has implications for supporting teachers' productive engagement in teaching that supports the integration of science content and scientific practices, as emphasized in current reform efforts. © 2015 Wiley Periodicals, Inc. J Res Sci Teach 52: 816–846, 2015.",,http://deepblue.lib.umich.edu/bitstream/2027.42/112188/2/tea21220-sup-0001-SuppFig_S2.pdf
-"An Integrated, Multimodal Resident Curriculum in Patient Safety and Quality Improvement",J. Werner,"Introduction Patient safety and quality improvement are essential components of modern medicine. The traditional model of graduate medical education does not lend itself well to learning these disciplines. This curriculum encompasses these disciplines across multiple modalities and extends throughout residency. Methods The curriculum includes introductory presentations suitable for naive audiences. Following these is a structured rotation that provides the opportunity both to experience in-depth self-directed learning and to practice skills involved in quality and safety. This rotation includes existing online courses published elsewhere, reflective writing exercises based on self-directed learning, and practice cases. Finally, residents lead a morbidity, mortality, and improvement conference where adverse events are identified and reviewed, specific interventions and outcome objectives are selected, and action teams are identified. Results After two presentations on system issues and individual issues, responses to the prompt “This talk will aid in my professional development” were 4.75 and 4.59 out of 5, respectively. Eighty-three percent of residents agreed they had a better understanding of the concepts of patient safety and/or quality improvement than they did before the rotation. Audience members for the resident-led morbidity, mortality, and improvement conference agreed it would lead to a change in their own practice. Discussion The contents of this longitudinal curriculum have been incorporated into the core requirements of our general pediatrics residency program and could reasonably be imported into any residency requiring a robust longitudinal experience in quality improvement and patient safety.","{'model': 'tldr@v2.0.0', 'text': 'The contents of this longitudinal curriculum have been incorporated into the core requirements of the general pediatrics residency program and could reasonably be imported into any residency requiring a robust longitudinal experience in quality improvement and patient safety.'}",
-Integrating Curriculum-Based Dynamic Assessment in Computerized Adaptive Testing: Development and Predictive Validity of the EDPL-BAI Battery on Reading Competence,J. Navarro,"In recent decades there have been significant changes in the conceptualization of reading as well as in the perception of how this activity should be assessed. Interest in the analysis of reading processes has led to the emergence of new explanatory models based primarily on the contributions of cognitive psychology. In parallel, there have been notable advances in measurement procedures, especially in models based on Item Response Theory (IRT), as well as in the capacity and performance of specific software programs that allow data to be managed and analyzed. These changes have contributed significantly to the rise of testing procedures such as computerized adaptive tests (CATs), whose fundamental characteristic is that the sequence of items presented in the tests is adapted to the level of competence that the subject manifests. Likewise, the incorporation of elements of dynamic assessment (DA) as the prompts are gradually offered allows for obtaining information about the type and degree of support required to optimize the subject’s performance. In this sense, the confluence of contributions from DA and CATs offers a new possibility for approaching the assessment of learning processes. In this article, we present a longitudinal research developed in two phases, through which a computerized dynamic adaptive assessment battery of reading processes (EDPL-BAI) was configured. The research frame involved 1,831 students (46% girls) from 13 public schools in three regions of Chile. The purpose of this study was to analyze the differential contribution on reading competence of dynamic scores obtained in a subsample composed of 324 (47% girls) students from third to sixth grade after the implementation of a set of adaptive dynamic tests of morpho-syntactic processes. The results achieved in the structural equation modeling indicate a good global fit. Individual relationships show a significant contribution of calibrated score that reflects estimated knowledge level on reading competence, as well as dynamic scores based on the assigned value of graduated prompts required by the students. These results showed significant predictive values on reading competence and incremental validity in relation to predictions made by static criterion tests.","{'model': 'tldr@v2.0.0', 'text': 'This article presents a longitudinal research developed in two phases, through which a computerized dynamic adaptive assessment battery of reading processes (EDPL-BAI) was configured, and shows significant predictive values on reading competence and incremental validity in relation to predictions made by static criterion tests.'}",
-Enhance Student Learning with PowerPoint Games: Using Twenty Questions to Promote Active Learning in Managerial Accounting,Victoria A. Fratto,"Stakeholders encourage accounting educators to provide active learning opportunities, to integrate the creative use of technology into the curriculum, and to emphasize learning by doing. The principles of good teaching practice can use technology to promote active learning, to provide prompt feedback to students, to increase student time on task, and to make learning more effective and efficient for the student. Technological tools can permit students to become active participants and can improve student learning by giving students convenient access to review material with immediate feedback. This article describes the use and development of a PowerPoint game in an introductory accounting course managerial accounting that provides the student with immediate feedback and is designed to be accessed by the student outside of the classroom. This technological tool can be used in other undergraduate academic disciplines.","{'model': 'tldr@v2.0.0', 'text': 'This article describes the use and development of a PowerPoint game in an introductory accounting course managerial accounting that provides the student with immediate feedback and is designed to be accessed by the student outside of the classroom.'}",
-Narrative Medicine: A Writing Workshop Curriculum for Residents,A. Winkel,"Introduction Reflection is a critical part of the learning cycle. Narrative medicine has been shown to help physicians-in-training develop both empathy and professional identity. A narrative medicine curriculum focused on the experience of the physician and challenging patient experiences creates opportunities in which to process complicated aspects of the experience of both patient and physician with other members of the training community. Methods Fifteen 1-hour small-group reflective writing workshops comprise a 2-year narrative medicine curriculum. Each workshop uses selected literature to focus a discussion and a prompt for written reflection and can be integrated into a didactic curriculum. Teacher guides have been created to help untrained preceptors lead small-group sessions. Feedback forms are distributed to participating residents. Results Out of 29 total residents enrolled at one program, 23 residents completed feedback forms, and 16 (69%) reported that the reflective writing sessions were relevant to their work as obstetrician-gynecologists. Residents stated the best parts of the course were reading and writing (n = 6; 27%), sharing writings with colleagues (n = 5; 21%), and having positive experiences with members of their community (n = 6; 27%). Some residents reported difficulty sharing their private reflections (n = 4; 17%). Discussion A narrative medicine curriculum is a powerful tool for promoting reflection about the challenging work of training in obstetrics and gynecology and other specialties. Reflective writing workshops have been found to be acceptable to obstetrics and gynecology residents, and the curriculum has been successfully implemented at several training programs.","{'model': 'tldr@v2.0.0', 'text': 'Reflective writing workshops have been found to be acceptable to obstetrics and gynecology residents, and the curriculum has been successfully implemented at several training programs.'}",
-Jelly Bean World: A Testbed for Never-Ending Learning,Emmanouil Antonios Platanios,"Machine learning has shown growing success in recent years. However, current machine learning systems are highly specialized, trained for particular problems or domains, and typically on a single narrow dataset. Human learning, on the other hand, is highly general and adaptable. Never-ending learning is a machine learning paradigm that aims to bridge this gap, with the goal of encouraging researchers to design machine learning systems that can learn to perform a wider variety of inter-related tasks in more complex environments. To date, there is no environment or testbed to facilitate the development and evaluation of never-ending learning systems. To this end, we propose the Jelly Bean World testbed. The Jelly Bean World allows experimentation over two-dimensional grid worlds which are filled with items and in which agents can navigate. This testbed provides environments that are sufficiently complex and where more generally intelligent algorithms ought to perform better than current state-of-the-art reinforcement learning approaches. It does so by producing non-stationary environments and facilitating experimentation with multi-task, multi-agent, multi-modal, and curriculum learning settings. We hope that this new freely-available software will prompt new research and interest in the development and evaluation of never-ending learning systems and more broadly, general intelligence systems.","{'model': 'tldr@v2.0.0', 'text': 'The Jelly Bean World testbed is proposed, a freely-available software that provides environments that are sufficiently complex and where more generally intelligent algorithms ought to perform better than current state-of-the-art reinforcement learning approaches.'}",
-Development of the Self-Determined Learning Model of Instruction Coaching Model: Implications for Research and Practice.,Mayumi Hagiwara,"The Self-Determined Learning Model of Instruction (SDLMI) is an evidence-based model of instruction implemented by a facilitator (e.g., general or special education teacher, family member, related service professional, transition professional) to enable students to learn to self-regulate problem solving in service to a goal. Students learn how to select personally relevant goals, develop action plans for achieving those goals, and self-monitor and self-evaluate progress toward achieving those goals. To scale-up the implementation of the SDLMI with fidelity by facilitators, there is a need for a systematic coaching model. Researchers have established coaching as a critical aspect of the implementation of evidence-based practices at scale and as a part of ongoing professional development. The purpose of this article is to describe steps taken to develop a systematic coaching model to support the implementation of the SDLMI. We will describe components of the SDLMI Coaching Model and the process through which they were derived from the coaching literature and data gathered during SDLMI research studies. We will conclude with implications for the implementation of the SDLMI Coaching Model at scale while maintaining a focus on individual students’ support needs. Causal Agency Theory provides an empirically-validated framework for conceptualizing the development of self-determination across the lifespan. Causal Agency Theory defines self-determination as a “dispositional characteristic manifested as acting as the causal agent in one’s life. Self-determined people (i.e., causal agents) act in service to freely chosen goals” (Shogren, Wehmeyer, Palmer, Forber-Pratt et al., 2015, p. 258). Self-determination develops across the lifespan as children and youth have opportunities to build skills and attitudes associated with self-determined actions including choice-making, decisionmaking, problem solving, goal setting and attainment, planning, self-management, self-advocacy, self-awareness, and self-knowledge (Shogren, Wehmeyer, Palmer, Forber-Pratt et al., 2015). The positive impact of promoting self-determination on in-school (e.g., academic achievement, goal attainment; Shogren et al., 2012) and post-school (e.g., employment, community participation; Shogren, Wehmeyer, Palmer, Rifenbark, & Little, 2015) outcomes has been confirmed by multiple research studies (Hagiwara et al., 2017). As such, in both general and special education, there is a growing emphasis on promoting the skills associated with self-determination in inclusive environments aligned with efforts to promote social-emotional learning and college and career readiness (Shogren et al., 2016). Researchers have established teachers can embed instruction in skills associated with self-determination into academic and transition-related content, leading to enhanced selfdetermination and goal attainment (Powers et al., 2012; Test et al., 2009). However, teachers have also identified the need for further support for implementation, particularly as reThis research was supported in part by a grant from the Institute of Education Sciences, National Center for Special Education Research, Grant Award R324A170008. Correspondence concerning this article should be addressed to Mayumi Hagiwara, Kansas University Center on Developmental Disabilities, 3134 Haworth Hall, 1200 Sunnyside Avenue, Lawrence, KS 66045. Email: mayuhagi@ku.edu Education and Training in Autism and Developmental Disabilities, 2020, 55(1), 17–27 © Division on Autism and Developmental Disabilities SDLMI Coaching Model / 17 search project-based supports fade. This lead to the activities undertaken to create the SDLMI Coaching Model. Self-Determined Learning Model of Instruction The Self-Determined Learning Model of Instruction (SDLMI; Shogren et al., 2019) is an evidencebased practice (National Technical Assistance Center on Transition, 2017) designed to embed instruction on self-determination in ongoing curricular activities. The SDLMI is designed to be implemented by a facilitator (e.g., general or special education teachers, career counselors, transition specialists). Facilitators enable students to learn and practice skills associated with self-determination, by providing explicit instruction (e.g., how to make choices, set goals, plan) and embedding opportunities to practice those skills in ongoing instruction to ultimately teach students to self-regulate problem solving in service to a goal. As such, general or special education teachers can use the SDLMI to shape their instruction and teach students how to set and attain goals that are linked to the general education curriculum or students’ individualized learning plans, including transition planning. The SDLMI can be implemented with students with and without disabilities, in whole-class, small group, or one-to-one settings. Essentially, the SDLMI was designed as a model of instruction that could be overlaid on any curricular area to support teachers in enabling students to set and work toward goals in that curricular area (Shogren et al., 2019). The SDLMI consists of a three-phase instructional process repeated over time to enable students to work on setting and attaining goals that build on and enhance each other (see Figure 1). Each of the three instructional phases includes four Student Questions that guide students through the problem-solving steps needed to solve the overarching question of each phase (Phase 1: What is my goal?, Phase 2: What is my plan?, Phase 3: What have I learned?). The solution to the problem in each phase leads to the problem-solving sequence in the Student Questions in the next phase. Importantly, each Student Question is linked to a set of Teacher Objectives that provide teachers or other facilitators with a road map for how they can organize instruction to enable students to answer each Student Question. The Teacher Objectives drive teachers’ instructional practices and actions and are linked to Educational Supports, which are strategies teachers can use to meet the specified objective based on students’ individualized learning needs. Teachers deliver targeted instruction on these strategies to support students in answering the Student Questions. Additionally, teachers integrate the identified goal and action plan into ongoing instructional activities by embedding opportunities for students to apply learned skills in service to their goals. For example, a student might set a goal to ensure they have provided the format of a solution in math class (e.g., most simplified version) by underlining the directions on homework, quizzes and tests. To support the student in achieving their goal and implementing their action plan, the teacher can refer to this goal during core content instruction as well as deliver targeted instruction enFigure 1. The phases of the SDLMI. © 2017 – Kansas University Center on Developmental Disabilities, Lawrence, KS USA. 18 / Education and Training in Autism and Developmental Disabilities-March 2020 abling the student to develop a system that prompts them to underline the directions before each beginning each assignment. In a recent, systematic literature review of the SDLMI studies, Hagiwara and colleagues (2017) found that all 21 SDLMI studies produced positive students outcomes (e.g., enhanced self-determination, increased access to the general education curriculum, improved classroom behavior, student-directed transition planning). However, only eight out of the studies identified in the literature that were implemented by teachers and/or other facilitators reported the characteristics of initial training and ongoing technical assistance activities to support the implementation of the SDLMI. For example, Shogren et al. (2012) and Wehmeyer et al. (2012) reported that teachers were trained for 0.5 to 1.5 days prior to implementation and supported through ongoing email support. However, no ongoing, systematic training, professional development, or coaching were delivered. Cho et al., (2011) found that teachers reported a lack of formal training in interventions to promote self-determination as one of the most significant barriers to implementing instruction to enhance self-determination in general and special education. In a recent statewide SDLMI implementation (Shogren et al., 2018), efforts were undertaken to provide teachers with more systematic implementation supports, including coaching and yearly professional development. While coaching played a major role in supporting the teachers to implement the SDLMI with fidelity in this project, the coaching was delivered relatively informally and varied across participating districts when evaluated from the lens of implementation science (Burke et al., 2019). Therefore, researchers recommended establishing a systematic coaching model to ensure that high-quality, consistent coaching is delivered across coaches and their assigned facilitators. Furthermore, researchers suggested improving the coaching system by: (a) asking teachers for feedback on their coaching experiences and (b) establishing a strong, integrated fidelity assessment system to addresses teacher practices in the classroom and coaching practices to ensure high quality implementation (Burke et al., 2019). Development of the SDLMI Coaching Model The importance of coaching in promoting fidelity of implementation of an effective innovation has been widely acknowledged (Fixsen et al., 2005). Coaching also plays a key role in ensuring facilitators of an intervention to have the competencies to implement evidence-based practices as intended over time (Snyder et al., 2015). Additionally, researchers have identified coaching as an effective way to provide face-to-face modeling and guidance on the use of interventions (Fixsen et al., 2010). For teachers and other facilitators to implement the SDLMI with fidelity, best practice recommends the importance of sustained and systematic supports thr",,
-Undergraduate recognition of curriculum-related skill development and the skills employers are seeking,Michelle A. Hill,"Employers of chemistry graduates are seeking a range of transferable skills from prospective employees, and academics are increasingly seeking to build employability skill development opportunities into the undergraduate curriculum. However, research suggests that undergraduates do not recognise or value such skill development without prompting. This recognition is essential if graduates are to be able to articulate their skills in the employment process. This study involves research amongst almost 1000 undergraduates studying chemistry at two institutions, using open-ended questions to collect qualitative data. The extent to which students recognised course-related skills development and understood the skills that employers are looking for was investigated, as was their desire to develop additional skills. Similarities and differences in student views between institutions are discussed, as well as trends across year levels and by gender. Results indicate that undergraduates studying chemistry are most likely to value and recognise development of some key skills sought by employers (teamwork, communication, thinking/problem solving, organisation/time management and laboratory/practical skills), but are very unlikely to value or recognise others (numeracy, independent learning, commercial awareness, interpersonal, research, computer/IT, creativity/innovation, flexibility/adaptability and initiative). Opportunities to develop the latter skills and recognition of the value of doing so will require improved communication with students and/or provision of new experiences within the curriculum.",,
-Students’ sense-making of personalised feedback based on learning analytics,Lisa-Angelique Lim,"Although technological advances have brought about new opportunities for scaling feedback to students, there remain challenges in how such feedback is presented and interpreted. There is a need to better understand how students make sense of such feedback to adapt self-regulated learning processes. This study examined students’ sense-making of learning analytics–based personalised feedback across four courses. Results from a combination of thematic analysis and epistemic network analysis show an association between student perceptions of their personalised feedback and how these map to subsequent self-described self-regulated learning processes. Most notably, the results indicate that personalised feedback, elaborated by personal messages from course instructors, helps students refine or strengthen important forethought processes of goal-setting, as well as to reduce procrastination. The results highlight the need for instructors to increase the dialogic element in personalised feedback in order to reduce defensive reactions from students who hold to their own learning strategies. This approach may prompt reflection on the suitability of students’ current learning strategies and achievement of associated learning goals. Implications for practice or policy: Personalised feedback based on learning analytics should be informed by an understanding of students’ self-regulated learning. Instructors implementing personalised feedback should align this closely with the course curriculum. Instructors implementing personalised feedback in their courses should consider the relational element of feedback by using a positive tone. Personalised feedback can be further enhanced by increasing the dialogic element and by including more information about learning strategies. ",,https://ajet.org.au/index.php/AJET/article/download/6370/1696
-Which Prompts Make The Difference? Data Prioritization For Efficient Human LLM Evaluation,M. Boubdir,"Human evaluation is increasingly critical for assessing large language models, capturing linguistic nuances, and reflecting user preferences more accurately than traditional automated metrics. However, the resource-intensive nature of this type of annotation process poses significant challenges. The key question driving our work:""is it feasible to minimize human-in-the-loop feedback by prioritizing data instances which most effectively distinguish between models?""We evaluate several metric-based methods and find that these metrics enhance the efficiency of human evaluations by minimizing the number of required annotations, thus saving time and cost, while ensuring a robust performance evaluation. We show that our method is effective across widely used model families, reducing instances of indecisive (or""tie"") outcomes by up to 54% compared to a random sample when focusing on the top-20 percentile of prioritized instances. This potential reduction in required human effort positions our approach as a valuable strategy in future large language model evaluations.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that the key question driving the work:""is it feasible to minimize human-in-the-loop feedback by prioritizing data instances which most effectively distinguish between models"" is feasible, and finds that several metric-based methods enhance the efficiency of human evaluations by minimizing the number of required annotations.'}",
-"Conversational Challenges in AI-Powered Data Science: Obstacles, Needs, and Design Opportunities",Bhavya Chopra,"Large Language Models (LLMs) are being increasingly employed in data science for tasks like data preprocessing and analytics. However, data scientists encounter substantial obstacles when conversing with LLM-powered chatbots and acting on their suggestions and answers. We conducted a mixed-methods study, including contextual observations, semi-structured interviews (n=14), and a survey (n=114), to identify these challenges. Our findings highlight key issues faced by data scientists, including contextual data retrieval, formulating prompts for complex tasks, adapting generated code to local environments, and refining prompts iteratively. Based on these insights, we propose actionable design recommendations, such as data brushing to support context selection, and inquisitive feedback loops to improve communications with AI-based assistants in data-science tools.","{'model': 'tldr@v2.0.0', 'text': 'These findings highlight key issues faced by data scientists, including contextual data retrieval, formulating prompts for complex tasks, adapting generated code to local environments, and refining prompts iteratively, and propose actionable design recommendations, such as data brushing to support context selection.'}",
-Human-in-the-loop Machine Translation with Large Language Model,Xinyi Yang,"The large language model (LLM) has garnered significant attention due to its in-context learning mechanisms and emergent capabilities. The research community has conducted several pilot studies to apply LLMs to machine translation tasks and evaluate their performance from diverse perspectives. However, previous research has primarily focused on the LLM itself and has not explored human intervention in the inference process of LLM. The characteristics of LLM, such as in-context learning and prompt engineering, closely mirror human cognitive abilities in language tasks, offering an intuitive solution for human-in-the-loop generation. In this study, we propose a human-in-the-loop pipeline that guides LLMs to produce customized outputs with revision instructions. The pipeline initiates by prompting the LLM to produce a draft translation, followed by the utilization of automatic retrieval or human feedback as supervision signals to enhance the LLM's translation through in-context learning. The human-machine interactions generated in this pipeline are also stored in an external database to expand the in-context retrieval database, enabling us to leverage human supervision in an offline setting. We evaluate the proposed pipeline using GPT-3.5-turbo API on five domain-specific benchmarks for German-English translation. The results demonstrate the effectiveness of the pipeline in tailoring in-domain translations and improving translation performance compared to direct translation. Additionally, we discuss the results from the following perspectives: 1) the effectiveness of different in-context retrieval methods; 2) the construction of a retrieval database under low-resource scenarios; 3) the observed domains differences; 4) the quantitative analysis of linguistic statistics; and 5) the qualitative analysis of translation cases. The code and data are available at https://github.com/NLP2CT/HIL-MT/.","{'model': 'tldr@v2.0.0', 'text': 'A human-in-the-loop pipeline that guides LLMs to produce customized outputs with revision instructions is proposed that demonstrates the effectiveness of the pipeline in tailoring in-domain translations and improving translation performance compared to direct translation.'}",https://arxiv.org/pdf/2310.08908
-CONA: A novel CONtext-Aware instruction paradigm for communication using large language model,Nan Zhou,"We introduce CONA, a novel context-aware instruction paradigm for effective knowledge dissemination using generative pre-trained transformer (GPT) models. CONA is a flexible framework designed to leverage the capabilities of Large Language Models (LLMs) and incorporate DIKW (Data, Information, Knowledge, Wisdom) hierarchy to automatically instruct and optimise presentation content, anticipate potential audience inquiries, and provide context-aware answers that adaptive to the knowledge level of the audience group. The unique aspect of the CONA paradigm lies in its combination of an independent advisory mechanism and a recursive feedback loop rooted on the DIKW hierarchy. This synergy significantly enhances context-aware contents, ensuring they are accessible and easily comprehended by the audience. This paradigm is an early pioneer to explore new methods for knowledge dissemination and communication in the LLM era, offering effective support for everyday knowledge sharing scenarios. We conduct experiments on a range of audience roles, along with materials from various disciplines using GPT4. Both quantitative and qualitative results demonstrated that the proposed CONA paradigm achieved remarkable performance compared to the outputs guided by conventional prompt engineering.","{'model': 'tldr@v2.0.0', 'text': 'CONA is a novel context-aware instruction paradigm for effective knowledge dissemination using generative pre-trained transformer (GPT) models that combines an independent advisory mechanism and a recursive feedback loop rooted on the DIKW hierarchy.'}",http://arxiv.org/pdf/2305.18620
-EXPLORING THE DESIGN SPACE OF AI BASED CODE COMPLETION ENGINES,Parth Thakkar,"Artificial Intelligence (AI) based code completion tools such as Github Copilot have recently gained tremendous popularity due to their ability to suggest arbitrary length snippets, improving developer productivity dramatically. However, there is little public understanding of what it takes to build such a tool. In this thesis, we explore the design space of building such a tool. We study the importance of the two key components of such a tool: the Large Language Model (LLM) that predicts the suggestions, and the system around it that feeds it the right context and filters out the bad suggestions. We start by exploring the design of Github Copilot to understand the state of the art, and describe the three key components of Copilot: Prompt Engineering, Model Invocation and Feedback loop. We then study the various factors that affect the quality of the suggestions generated by the LLM. We study both (a) the impact of the context fed to the LLM, and (b) the impact of the LLM itself. For the former, we study the impact including context from other files and code after the cursor along with different methods of context collection and amount of collected context. For the latter, we study the impact of the size of the LLM and the training procedure. Apart from factors affecting the quality of suggestions, we also study the factors affecting the latency of such code completion engines, as low latency is critical for building good code completion engines. We find that the context fed to the model makes a significant difference in the quality of generated suggestions, and good context collection can improve the quality of suggestions by 11-26% points (20-113% relative improvement) on the exact match metric for one line suggestions. Models that can exploit the context after the cursor can further improve the quality of suggestions by 6-14% points (12-16% relative improvement). Our experiments show that increasing the prompt length beyond a point does not improve suggestion quality significantly, and that 2048-4096 tokens are sufficient. We also find that the size of the LLM has a much smaller impact on the quality of suggestions than other factors such as the context fed to the model and the training procedure used. For example, we found that the SantaCoder model (1.1B parameters) provided better suggestions than the 16B CodeGen-Multi","{'model': 'tldr@v2.0.0', 'text': 'It is found that the context fed to the model makes a significant difference in thequality of generated suggestions, and good context collection can improve the quality of suggestions by 11-26% points (20-113% relative improvement) on the exact match metric for one line suggestions.'}",
-Incremental Learning of Humanoid Robot Behavior from Natural Interaction and Large Language Models,Leonard Bärmann,"Natural-language dialog is key for intuitive human-robot interaction. It can be used not only to express humans' intents, but also to communicate instructions for improvement if a robot does not understand a command correctly. Of great importance is to endow robots with the ability to learn from such interaction experience in an incremental way to allow them to improve their behaviors or avoid mistakes in the future. In this paper, we propose a system to achieve incremental learning of complex behavior from natural interaction, and demonstrate its implementation on a humanoid robot. Building on recent advances, we present a system that deploys Large Language Models (LLMs) for high-level orchestration of the robot's behavior, based on the idea of enabling the LLM to generate Python statements in an interactive console to invoke both robot perception and action. The interaction loop is closed by feeding back human instructions, environment observations, and execution results to the LLM, thus informing the generation of the next statement. Specifically, we introduce incremental prompt learning, which enables the system to interactively learn from its mistakes. For that purpose, the LLM can call another LLM responsible for code-level improvements of the current interaction based on human feedback. The improved interaction is then saved in the robot's memory, and thus retrieved on similar requests. We integrate the system in the robot cognitive architecture of the humanoid robot ARMAR-6 and evaluate our methods both quantitatively (in simulation) and qualitatively (in simulation and real-world) by demonstrating generalized incrementally-learned knowledge.","{'model': 'tldr@v2.0.0', 'text': ""A system that deploys Large Language Models for high-level orchestration of the robot's behavior based on the idea of enabling the LLM to generate Python statements in an interactive console to invoke both robot perception and action is presented.""}",https://arxiv.org/pdf/2309.04316
-REAL: Resilience and Adaptation using Large Language Models on Autonomous Aerial Robots,Andrea Tagliabue,"Large Language Models (LLMs) pre-trained on internet-scale datasets have shown impressive capabilities in code understanding, synthesis, and general purpose question-and-answering. Key to their performance is the substantial prior knowledge acquired during training and their ability to reason over extended sequences of symbols, often presented in natural language. In this work, we aim to harness the extensive long-term reasoning, natural language comprehension, and the available prior knowledge of LLMs for increased resilience and adaptation in autonomous mobile robots. We introduce REAL, an approach for REsilience and Adaptation using LLMs. REAL provides a strategy to employ LLMs as a part of the mission planning and control framework of an autonomous robot. The LLM employed by REAL provides (i) a source of prior knowledge to increase resilience for challenging scenarios that the system had not been explicitly designed for; (ii) a way to interpret natural-language and other log/diagnostic information available in the autonomy stack, for mission planning; (iii) a way to adapt the control inputs using minimal user-provided prior knowledge about the dynamics/kinematics of the robot. We integrate REAL in the autonomy stack of a real multirotor, querying onboard an offboard LLM at 0.1-1.0 Hz as part the robot's mission planning and control feedback loops. We demonstrate in real-world experiments the ability of the LLM to reduce the position tracking errors of a multirotor under the presence of (i) errors in the parameters of the controller and (ii) unmodeled dynamics. We also show (iii) decision making to avoid potentially dangerous scenarios (e.g., robot oscillates) that had not been explicitly accounted for in the initial prompt design.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the LLM to reduce the position tracking errors of a multirotor under the presence of (i) errors in the parameters of the controller and (ii) unmodeled dynamics is demonstrated.'}",
-Improving Summarization with Human Edits,Zonghai Yao,"Recent work has shown the promise of learning with human feedback paradigms to produce human-determined high-quality text. Existing works use human feedback to train large language models (LLMs) in general domain abstractive summarization and have obtained summary quality exceeding traditional likelihood training. In this paper, we focus on a less explored form of human feedback -- Human Edits. We propose Sequence Alignment (un)Likelihood Training (SALT), a novel technique to use both the human-edited and model-generated data together in the training loop. In addition, we demonstrate simulating Human Edits with ground truth summaries coming from existing training data -- Imitation edits, along with the model-generated summaries obtained after the training, to reduce the need for expensive human-edit data. In our experiments, we extend human feedback exploration from general domain summarization to medical domain summarization. Our results demonstrate the effectiveness of SALT in improving the summary quality with Human and Imitation Edits. Through additional experiments, we show that SALT outperforms the conventional RLHF method (designed for human preferences) -- DPO, when applied to human-edit data. We hope the evidence in our paper prompts researchers to explore, collect, and better use different human feedback approaches scalably.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Sequence Alignment (un)Likelihood Training (SALT), a novel technique to use both the human-edited and model-generated data together in the training loop, and demonstrates the effectiveness of SALT in improving the summary quality with Human and Imitation Edits.'}",https://arxiv.org/pdf/2310.05857
-"Double-loop learning control (DLC) model for reengineering: a ""yin"" and ""yang"" balanced approach for effective organizational change",Y. Hua,"The advent of information technology has prompted organizations to reinvent themselves in response to the rising demands of customers and expectations of employees. A conceptual model that shows how organizations can effectively manage the change process is essential in view of the multitude of problems encountered in reengineering endeavours. The double-loop learning control (DLC) model advocates the importance of having a shared vision and an environment that facilitates double-loop learning. The model uses the oriental philosophical 'yin-yang' balanced concept to illustrate the need to balance the application of technology with the human factor. The DLC model was empirically tested by a public sector survey on success factors for change, conducted in 1997. Analyses confirmed that having a shared vision is paramount to reengineering success and an organizational ability to learn as a mediator of the yin-yang contributions toward building this shared vision for organizational change.",,
-A Parameter-Efficient Learning Approach to Arabic Dialect Identification with Pre-Trained General-Purpose Speech Model,S. Radhakrishnan,"In this work, we explore Parameter-Efficient-Learning (PEL) techniques to repurpose a General-Purpose-Speech (GSM) model for Arabic dialect identification (ADI). Specifically, we investigate different setups to incorporate trainable features into a multi-layer encoder-decoder GSM formulation under frozen pre-trained settings. Our architecture includes residual adapter and model reprogramming (input-prompting). We design a token-level label mapping to condition the GSM for Arabic Dialect Identification (ADI). This is challenging due to the high variation in vocabulary and pronunciation among the numerous regional dialects. We achieve new state-of-the-art accuracy on the ADI-17 dataset by vanilla fine-tuning. We further reduce the training budgets with the PEL method, which performs within 1.86% accuracy to fine-tuning using only 2.5% of (extra) network trainable parameters. Our study demonstrates how to identify Arabic dialects using a small dataset and limited computation with open source code and pre-trained models.","{'model': 'tldr@v2.0.0', 'text': 'This work explores Parameter-Efficient-Learning techniques to repurpose a General-Purpose-Speech model for Arabic dialect identification (ADI) and achieves new state-of-the-art accuracy on the ADI-17 dataset by vanilla fine-tuning.'}",https://repository.kaust.edu.sa/bitstream/10754/694670/1/2305.11244.pdf
-Tree-Planner: Efficient Close-loop Task Planning with Large Language Models,Mengkang Hu,"This paper studies close-loop task planning, which refers to the process of generating a sequence of skills (a plan) to accomplish a specific goal while adapting the plan based on real-time observations. Recently, prompting Large Language Models (LLMs) to generate actions iteratively has become a prevalent paradigm due to its superior performance and user-friendliness. However, this paradigm is plagued by two inefficiencies: high token consumption and redundant error correction, both of which hinder its scalability for large-scale testing and applications. To address these issues, we propose Tree-Planner, which reframes task planning with LLMs into three distinct phases: plan sampling, action tree construction, and grounded deciding. Tree-Planner starts by using an LLM to sample a set of potential plans before execution, followed by the aggregation of them to form an action tree. Finally, the LLM performs a top-down decision-making process on the tree, taking into account real-time environmental information. Experiments show that Tree-Planner achieves state-of-the-art performance while maintaining high efficiency. By decomposing LLM queries into a single plan-sampling call and multiple grounded-deciding calls, a considerable part of the prompt are less likely to be repeatedly consumed. As a result, token consumption is reduced by 92.2% compared to the previously best-performing model. Additionally, by enabling backtracking on the action tree as needed, the correction process becomes more flexible, leading to a 40.5% decrease in error corrections. Project page: https://tree-planner.github.io/","{'model': 'tldr@v2.0.0', 'text': 'Tree-Planner is proposed, which reframes task planning with Large Language Models into three distinct phases: plan sampling, action tree construction, and grounded deciding, which achieves state-of-the-art performance while maintaining high efficiency.'}",https://arxiv.org/pdf/2310.08582
-BatchPrompt: Accomplish more with less,Jianzhe Lin,"As the ever-increasing token limits of large language models (LLMs) have enabled long context as input, prompting with single data samples might no longer an efficient way. A straightforward strategy improving efficiency is to batch data within the token limit (e.g., 8k for gpt-3.5-turbo; 32k for GPT-4), which we call BatchPrompt. We have two initial observations for prompting with batched data. First, we find that prompting with batched data in longer contexts will inevitably lead to worse performance, compared to single-data prompting. Second, the performance of the language model is significantly correlated with the positions and order of the batched data, due to the corresponding change in decoder context. To retain efficiency and overcome performance loss, we propose Batch Permutation and Ensembling (BPE), and a novel Self-reflection-guided EArly Stopping (SEAS) technique. Our comprehensive experimental evaluation demonstrates that BPE can boost the performance of BatchPrompt with a striking margin on a range of popular NLP tasks, including question answering (Boolq), textual entailment (RTE), and duplicate questions identification (QQP). These performances are even competitive with/higher than single-data prompting(SinglePrompt), while BatchPrompt requires much fewer LLM calls and input tokens (For SinglePrompt v.s. BatchPrompt with batch size 32, using just 9%-16% the number of LLM calls, Boolq accuracy 90.6% to 90.9% with 27.4% tokens, QQP accuracy 87.2% to 88.4% with 18.6% tokens, RTE accuracy 91.5% to 91.1% with 30.8% tokens). To the best of our knowledge, this is the first work to technically improve prompting efficiency of large language models. We hope our simple yet effective approach will shed light on the future research of large language models. The code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This is the first work to technically improve prompting efficiency of large language models by proposing Batch Permutation and Ensembling (BPE), and a novel Self-reflection-guided EArly Stopping (SEAS) technique.'}",https://arxiv.org/pdf/2309.00384
-CenterCLIP: Token Clustering for Efficient Text-Video Retrieval,Shuai Zhao,"Recently, large-scale pre-training methods like CLIP have made great progress in multi-modal research such as text-video retrieval. In CLIP, transformers are vital for modeling complex multi-modal relations. However, in the vision transformer of CLIP, the essential visual tokenization process, which produces discrete visual token sequences, generates many homogeneous tokens due to the redundancy nature of consecutive and similar frames in videos. This significantly increases computation costs and hinders the deployment of video retrieval models in web applications. In this paper, to reduce the number of redundant video tokens, we design a multi-segment token clustering algorithm to find the most representative tokens and drop the non-essential ones. As the frame redundancy occurs mostly in consecutive frames, we divide videos into multiple segments and conduct segment-level clustering. Center tokens from each segment are later concatenated into a new sequence, while their original spatial-temporal relations are well maintained. We instantiate two clustering algorithms to efficiently find deterministic medoids and iteratively partition groups in high dimensional space. Through this token clustering and center selection procedure, we successfully reduce computation costs by removing redundant visual tokens. This method further enhances segment-level semantic alignment between video and text representations, enforcing the spatio-temporal interactions of tokens from within-segment frames. Our method, coined as CenterCLIP, surpasses existing state-of-the-art by a large margin on typical text-video benchmarks, while reducing the training memory cost by 35% and accelerating the inference speed by 14% at the best case. The code is available at https://github.com/mzhaoshuai/CenterCLIP https://github.com/mzhaoshuai/CenterCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A multi-segment token clustering algorithm to find the most representative tokens and drop the non-essential ones, which surpasses existing state-of-the-art by a large margin on typical text-video benchmarks, while reducing the training memory cost and accelerating the inference speed.'}",https://arxiv.org/pdf/2205.00823
-Token Dropping for Efficient BERT Pretraining,Le Hou,"Transformer-based models generally allocate the same amount of computation for each token in a given sequence. We develop a simple but effective “token dropping” method to accelerate the pretraining of transformer models, such as BERT, without degrading its performance on downstream tasks. In particular, we drop unimportant tokens starting from an intermediate layer in the model to make the model focus on important tokens more efficiently if with limited computational resource. The dropped tokens are later picked up by the last layer of the model so that the model still produces full-length sequences. We leverage the already built-in masked language modeling (MLM) loss to identify unimportant tokens with practically no computational overhead. In our experiments, this simple approach reduces the pretraining cost of BERT by 25% while achieving similar overall fine-tuning performance on standard downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'A simple but effective “token dropping” method is developed to accelerate the pretraining of transformer models, such as BERT, without degrading its performance on downstream tasks.'}",http://arxiv.org/pdf/2203.13240
-DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification,Yongming Rao,"Attention is sparse in vision transformers. We observe the final prediction in vision transformers is only based on a subset of most informative tokens, which is sufficient for accurate image recognition. Based on this observation, we propose a dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input. Specifically, we devise a lightweight prediction module to estimate the importance score of each token given the current features. The module is added to different layers to prune redundant tokens hierarchically. To optimize the prediction module in an end-to-end manner, we propose an attention masking strategy to differentiably prune a token by blocking its interactions with other tokens. Benefiting from the nature of self-attention, the unstructured sparse tokens are still hardware friendly, which makes our framework easy to achieve actual speed-up. By hierarchically pruning 66% of the input tokens, our method greatly reduces 31%~37% FLOPs and improves the throughput by over 40% while the drop of accuracy is within 0.5% for various vision transformers. Equipped with the dynamic token sparsification framework, DynamicViT models can achieve very competitive complexity/accuracy trade-offs compared to state-of-the-art CNNs and vision transformers on ImageNet. Code is available at https://github.com/raoyongming/DynamicViT","{'model': 'tldr@v2.0.0', 'text': 'A dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input and an attention masking strategy to differentiably prune a token by blocking its interactions with other tokens is proposed.'}",
-Modular and Parameter-Efficient Multimodal Fusion with Prompting,Sheng Liang,"Recent research has made impressive progress in large-scale multimodal pre-training. In the context of the rapid growth of model size, it is necessary to seek efficient and flexible methods other than finetuning. In this paper, we propose to use prompt vectors to align the modalities. Our method achieves comparable performance to several other multimodal fusion methods in low-resource settings. We further show that our method is modular and parameter-efficient for processing tasks involving two or more data modalities.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2203.08055
-Efficient Token Mixing for Transformers via Adaptive Fourier Neural Operators,John Guibas,,,
-AdapterBias: Parameter-efficient Token-dependent Representation Shift for Adapters in NLP Tasks,Chin-Lun Fu,"Transformer-based pre-trained models with millions of parameters require large storage. Recent approaches tackle this shortcoming by training adapters, but these approaches still require a relatively large number of parameters. In this study, AdapterBias, a surprisingly simple yet effective adapter architecture, is proposed. AdapterBias adds a token-dependent shift to the hidden output of transformer layers to adapt to downstream tasks with only a vector and a linear layer. Extensive experiments are conducted to demonstrate the effectiveness of AdapterBias. The experiments show that our proposed method can dramatically reduce the trainable parameters compared to the previous works with a minimal decrease in task performances compared with fine-tuned pre-trained models. We further find that AdapterBias automatically learns to assign more significant representation shifts to the tokens related to the task in consideration.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that the proposed AdapterBias method can dramatically reduce the trainable parameters compared to the previous works with a minimal decrease in task performances compared with fine-tuned pre-trained models.'}",http://arxiv.org/pdf/2205.00305
-TokenMixup: Efficient Attention-guided Token-level Data Augmentation for Transformers,Hyeong Kyu Choi,"Mixup is a commonly adopted data augmentation technique for image classification. Recent advances in mixup methods primarily focus on mixing based on saliency. However, many saliency detectors require intense computation and are especially burdensome for parameter-heavy transformer models. To this end, we propose TokenMixup, an efficient attention-guided token-level data augmentation method that aims to maximize the saliency of a mixed set of tokens. TokenMixup provides x15 faster saliency-aware data augmentation compared to gradient-based methods. Moreover, we introduce a variant of TokenMixup which mixes tokens within a single instance, thereby enabling multi-scale feature augmentation. Experiments show that our methods significantly improve the baseline models' performance on CIFAR and ImageNet-1K, while being more efficient than previous methods. We also reach state-of-the-art performance on CIFAR-100 among from-scratch transformer models. Code is available at https://github.com/mlvlab/TokenMixup.","{'model': 'tldr@v2.0.0', 'text': 'TokenMixup is proposed, an efficient attention-guided token-level data augmentation method that aims to maximize the saliency of a mixed set of tokens and provides x15 faster saliency-awareData augmentation compared to gradient-based methods.'}",http://arxiv.org/pdf/2210.07562
-TORE: Token Reduction for Efficient Human Mesh Recovery with Transformer,Zhiyang Dou,"In this paper, we introduce a set of simple yet effective TOken REduction (TORE) strategies for Transformer-based Human Mesh Recovery from monocular images. Current SOTA performance is achieved by Transformer-based structures. However, they suffer from high model complexity and computation cost caused by redundant tokens. We propose token reduction strategies based on two important aspects, i.e., the 3D geometry structure and 2D image feature, where we hierarchically recover the mesh geometry with priors from body structure and conduct token clustering to pass fewer but more discriminative image feature tokens to the Transformer. Our method massively reduces the number of tokens involved in high-complexity interactions in the Transformer. This leads to a significantly reduced computational cost while still achieving competitive or even higher accuracy in shape recovery. Extensive experiments across a wide range of benchmarks validate the superior effectiveness of the proposed method. We further demonstrate the generalizability of our method on hand mesh recovery. Visit our project page at https://frank-zy-dou.github.io/projects/Tore/index.html.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a set of simple yet effective TOken REduction (TORE) strategies for Transformer-based Human Mesh Recovery from monocular images that massively reduces the number of tokens involved in high-complexity interactions in the Transformer.'}",https://arxiv.org/pdf/2211.10705
-Adaptive Fourier Neural Operators: Efficient Token Mixers for Transformers,John Guibas,"Vision transformers have delivered tremendous success in representation learning. This is primarily due to effective token mixing through self attention. However, this scales quadratically with the number of pixels, which becomes infeasible for high-resolution inputs. To cope with this challenge, we propose Adaptive Fourier Neural Operator (AFNO) as an efficient token mixer that learns to mix in the Fourier domain. AFNO is based on a principled foundation of operator learning which allows us to frame token mixing as a continuous global convolution without any dependence on the input resolution. This principle was previously used to design FNO, which solves global convolution efficiently in the Fourier domain and has shown promise in learning challenging PDEs. To handle challenges in visual representation learning such as discontinuities in images and high resolution inputs, we propose principled architectural modifications to FNO which results in memory and computational efficiency. This includes imposing a block-diagonal structure on the channel mixing weights, adaptively sharing weights across tokens, and sparsifying the frequency modes via soft-thresholding and shrinkage. The resulting model is highly parallel with a quasi-linear complexity and has linear memory in the sequence size. AFNO outperforms self-attention mechanisms for few-shot segmentation in terms of both efficiency and accuracy. For Cityscapes segmentation with the Segformer-B3 backbone, AFNO can handle a sequence size of 65k and outperforms other efficient self-attention mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'Adaptive Fourier Neural Operator is proposed as an efficient token mixer that learns to mix in the Fourier domain that can handle a sequence size of 65k and outperforms other efficient self-attention mechanisms for few-shot segmentation in terms of both efficiency and accuracy.'}",
-CITADEL: Conditional Token Interaction via Dynamic Lexical Routing for Efficient and Effective Multi-Vector Retrieval,Minghan Li,"Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) and dense (e.g. DPR) retrievers and have achieved state-of-the-art performance on various retrieval tasks.These methods, however, are orders of magnitude slower and need much more space to store their indices compared to their single-vector counterparts.In this paper, we unify different multi-vector retrieval models from a token routing viewpoint and propose conditional token interaction via dynamic lexical routing, namely CITADEL, for efficient and effective multi-vector retrieval.CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key.This design significantly reduces the computation cost while maintaining high accuracy.Notably, CITADEL achieves the same or slightly better performance than the previous state of the art, ColBERT-v2, on both in-domain (MS MARCO) and out-of-domain (BEIR) evaluations, while being nearly 40 times faster. Source code and data are available at https://github.com/facebookresearch/dpr-scale/tree/citadel.","{'model': 'tldr@v2.0.0', 'text': 'CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key, which significantly reduces the computation cost while maintaining high accuracy.'}",https://arxiv.org/pdf/2211.10411
-Adaptive Token Sampling for Efficient Vision Transformers,Mohsen Fayyaz,,"{'model': 'tldr@v2.0.0', 'text': 'This work introduces a differentiable parameter-free Adaptive Token Sampler module, which can be plugged into any existing vision transformer architecture, and improves the SOTA by reducing their computational costs (GFLOPs) by 2X, while preserving their accuracy on the ImageNet, Kinetics-400, andKinetics-600 datasets.'}",https://arxiv.org/pdf/2111.15667
-Orthogonal Transformer: An Efficient Vision Transformer Backbone with Token Orthogonalization,Huaibo Huang,"We present a general vision transformer backbone, called as Orthogonal Trans-former, in pursuit of both efficiency and effectiveness. A major challenge for vision transformer is that self-attention, as the key element in capturing long-range dependency, is very computationally expensive for dense prediction tasks (e.g., object detection). Coarse global self-attention and local self-attention are then designed to reduce the cost, but they suffer from either neglecting local correlations or hurting global modeling. We present an orthogonal self-attention mechanism to alleviate these issues. Specifically, self-attention is computed in the orthogonal space that is reversible to the spatial domain but has much lower resolution. The capabilities of learning global dependency and exploring local correlations are maintained because every orthogonal token in self-attention can attend to the entire visual tokens. Remarkably, orthogonality is realized by constructing an endogenously orthogonal matrix that is friendly to neural networks and can be optimized as arbitrary orthogonal matrices. We also introduce Positional MLP to incorporate position information for arbitrary input resolutions as well as enhance the capacity of MLP. Finally, we develop a hierarchical architecture for Orthogonal Transformer. Extensive experiments demonstrate its strong performance on a broad range of vision tasks, including image classification, object detection, instance segmentation and semantic segmentation.","{'model': 'tldr@v2.0.0', 'text': 'A general vision transformer backbone, called as Orthogonal Trans-former, in pursuit of both efficiency and effectiveness is presented, and an orthogonal self-attention mechanism is presented to alleviate issues.'}",
-DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification,,.,,
-Efficient Video Transformers with Spatial-Temporal Token Selection,Junke Wang,,"{'model': 'tldr@v2.0.0', 'text': 'STTS is a token selection framework that dynamically selects a few informative tokens in both temporal and spatial dimensions conditioned on input video samples that achieves similar results while requiring 20% less computation.'}",https://arxiv.org/pdf/2111.11591
-DTQAtten: Leveraging Dynamic Token-based Quantization for Efficient Attention Architecture,Tao Yang,"Models based on the attention mechanism, i.e. transformers, have shown extraordinary performance in Natural Language Processing (NLP) tasks. However, their memory footprint, inference latency, and power consumption are still prohibitive for efficient inference at edge devices, even at data centers. To tackle this issue, we present an algorithm-architecture co-design with dynamic and mixed-precision quantization, DTQAtten. We present empirically that the tolerance to the noise varies from token to token in attention-based NLP models. This finding leads us to quantize different tokens with mixed levels of bits. Thus, we design a compression framework that (i) dynamically quantizes tokens while they are forwarded in the models and (ii) jointly determines the ratio of each precision. Moreover, due to the dynamic mixed-precision tokens caused by our framework, previous matrix-multiplication accelerators (e.g. systolic array) cannot effectively exploit the benefit of the compressed attention computation. We thus design our accelerator with the variable-speed systolic array (VSSA) and propose an effective optimization strategy to alleviate the pipeline-stall problem in VSSA without hardware overhead. We conduct experiments with existing attention-based NLP models, including BERT and GPT-2 on various language tasks. Our results show that DTQAtten outperforms the previous neural network accelerator Eyeriss by 13.12× in terms of speedup and 3.8× in terms of energy-saving. Compared with the state-of-the-art attention accelerator SpAtten, our DTQAtten achieves at least 2.65× speedup and 3.38× energy efficiency improvement.","{'model': 'tldr@v2.0.0', 'text': 'This work designs an algorithm-architecture co-design with dynamic and mixed-precision quantization with an effective optimization strategy to alleviate the pipeline-stall problem in VSSA without hardware overhead and conducts experiments with existing attention-based NLP models.'}",
-HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers,Peiyan Dong,"While vision transformers (ViTs) have continuously achieved new milestones in the field of computer vision, their sophisticated network architectures with high computation and memory costs have impeded their deployment on resource-limited edge devices. In this paper, we propose a hardware-efficient image-adaptive token pruning framework called HeatViT for efficient yet accurate ViT acceleration on embedded FPGAs. Based on the inherent computational patterns in ViTs, we first adopt an effective, hardware-efficient, and learnable head-evaluation token selector, which can be progressively inserted before transformer blocks to dynamically identify and consolidate the non-informative tokens from input images. Moreover, we implement the token selector on hardware by adding miniature control logic to heavily reuse existing hardware components built for the backbone ViT. To improve the hardware efficiency, we further employ 8-bit fixed-point quantization and propose polynomial approximations with regularization effect on quantization error for the frequently used nonlinear functions in ViTs. Compared to existing ViT pruning studies, under the similar computation cost, HeatViT can achieve 0.7% ~ 8.9% higher accuracy; while under the similar model accuracy, HeatViT can achieve more than 28.4% ~ 65.3% computation reduction, for various widely used ViTs, including DeiT-T, DeiT-S, DeiT-B, LV-ViT-S, and LV-ViT-M, on the ImageNet dataset. Compared to the baseline hardware accelerator, our implementations of HeatViT on the Xilinx ZCU102 FPGA achieve 3.46×~4.89× speedup with a trivial resource utilization overhead of 8%~11% more DSPs and 5%~8% more LUTs.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hardware-efficient image-adaptive token pruning framework called HeatViT for efficient yet accurate ViT acceleration on embedded FPGAs, and adopts an effective, hardware- efficient, and learnable head-evaluation token selector.'}",http://arxiv.org/pdf/2211.08110
-Efficient Transformers with Dynamic Token Pooling,Piotr Nawrot,"Transformers achieve unrivalled performance in modelling language, but remain inefficient in terms of memory and time complexity. A possible remedy is to reduce the sequence length in the intermediate layers by pooling fixed-length segments of tokens. Nevertheless, natural units of meaning, such as words or phrases, display varying sizes. To address this mismatch, we equip language models with a dynamic-pooling mechanism, which predicts segment boundaries in an autoregressive fashion. We compare several methods to infer boundaries, including end-to-end learning through stochastic re-parameterisation, supervised learning (based on segmentations from subword tokenizers or spikes in conditional entropy), as well as linguistically motivated boundaries. We perform character-level evaluation on texts from multiple datasets and morphologically diverse languages. The results demonstrate that dynamic pooling, which jointly segments and models language, is both faster and more accurate than vanilla Transformers and fixed-length pooling within the same computational budget.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that dynamic pooling, which jointly segments and models language, is both faster and more accurate than vanilla Transformers and fixed-length pooling within the same computational budget.'}",http://arxiv.org/pdf/2211.09761
-Compact Token Representations with Contextual Quantization for Efficient Document Re-ranking,Yingrui Yang,"Transformer based re-ranking models can achieve high search relevance through context- aware soft matching of query tokens with document tokens. To alleviate runtime complexity of such inference, previous work has adopted a late interaction architecture with pre-computed contextual token representations at the cost of a large online storage. This paper proposes contextual quantization of token embeddings by decoupling document-specific and document-independent ranking contributions during codebook-based compression. This allows effective online decompression and embedding composition for better search relevance. This paper presents an evaluation of the above compact token representation model in terms of relevance and space efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes contextual quantization of token embeddings by decoupling document-specific and document-independent ranking contributions during codebook-based compression to allow effective online decompression and embedding composition for better search relevance.'}",http://arxiv.org/pdf/2203.15328
-ATS: Adaptive Token Sampling For Efficient Vision Transformers,Mohsen Fayyaz,"While state-of-the-art vision transformer models achieve promising results for image classification, they are computationally very expensive and require many GFLOPs. Although the GFLOPs of a vision transformer can be de-creased by reducing the number of tokens in the network, there is no setting that is optimal for all input images. In this work, we therefore introduce a differentiable parameter-free Adaptive Token Sampling (ATS) module, which can be plugged into any existing vision transformer architecture. ATS empowers vision transformers by scoring and adaptively sampling significant tokens. As a result, the number of tokens is not anymore static but it varies for each input image. By integrating ATS as an additional layer within current transformer blocks, we can convert them into much more efficient vision transformers with an adaptive number of tokens. Since ATS is a parameter-free module, it can be added to off-the-shelf pretrained vision transformers as a plug and play module, thus reducing their GFLOPs without any additional training. However, due to its differentiable design, one can also train a vision transformer equipped with ATS. We evaluate our module on the ImageNet dataset by adding it to multiple state-of-the-art vision transformers. Our evaluations show that the proposed module improves the state-of-the-art by reducing the computational cost (GFLOPs) by 37% while preserving the accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A differentiable parameter-free Adaptive Token Sampling (ATS) module, which can be plugged into any existing vision transformer architecture, and improves the state-of-the-art by reducing the computational cost (GFLOPs) by 37% while preserving the accuracy.'}",
-You Only Group Once: Efficient Point-Cloud Processing with Token Representation and Relation Inference Module,Chenfeng Xu,"3D perception on point-cloud is a challenging and crucial computer vision task. A point-cloud consists of a sparse, unstructured, and unordered set of points. To understand a point-cloud, previous point-based methods, such as PointNet++, extract visual features through the hierarchical aggregation of local features. However, such methods have several critical limitations: 1) They require considerable sampling and grouping operations, which leads to low inference speed. 2) Despite redundancy among adjacent points, they treat all points alike with an equal amount of computation. 3) They aggregate local features together through downsampling, which causes information loss and hurts perception capability. To overcome these challenges, we propose a novel, simple, and elegant deep learning model called YOGO (You Only Group Once). YOGO divides a point-cloud into a small number of parts and extracts a high-dimensional token to represent points within each sub-region. Next, we use self-attention to capture token-to-token relations, and project the token features back to the point features. We formulate such a series of operations as a relation inference module (RIM). Compared with previous methods, YOGO is very efficient because it only needs to sample and group a point-cloud once. Instead of operating on points, YOGO operates on a small number of tokens, each of which summarizes the point features in a sub-region. This allows us to avoid redundant computation and thus boosts efficiency. Moreover, YOGO preserves pointwise features by projecting token features to point features although the RIM computes on tokens. This avoids information loss and enhances point-wise perception capability. We conduct thorough experiments to demonstrate that YOGO achieves at least 3.0x speedup over point-based baselines while delivering competitive classification and segmentation performance on a classification dataset and a segmentation dataset based on 3D Wharehouse, and S3DIS datasets. The code is available at https://github.com/chenfengxu714/YOGO.git.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel, simple, and elegant deep learning model called YOGO (You Only Group Once), which operates on a small number of tokens, each of which summarizes the point features in a sub-region, to avoid redundant computation and thus boosts efficiency.'}",https://arxiv.org/pdf/2103.09975
-Efficient Prompting via Dynamic In-Context Learning,Wangchunshu Zhou,"The primary way of building AI applications is shifting from training specialist models to prompting generalist models. A common practice for prompting generalist models, often referred to as in-context learning, is to append a few examples (demonstrations) to the prompt to help the model better understand the task. While effective, in-context learning can be inefficient because it makes the input prompt much longer, consuming valuable space in the context window and leading to larger computational costs. In this paper, we propose DynaICL, a recipe for efficient prompting with black-box generalist models that dynamically allocate in-context examples according to the input complexity and the computational budget. To achieve this, we train a meta controller that predicts the number of in-context examples suitable for the generalist model to make a good prediction based on the performance-efficiency trade-off for a specific input. We then dynamically allocate the number of demonstrations for an input according to predictions from the meta controller and the given computation budget. Experimental results show that dynamic example allocation helps achieve a better performance-efficiency trade-off in two practical settings where computational resources or the required performance is constrained. Specifically, DynaICL saves up to 46% token budget compared to the common practice that allocates the same number of in-context examples to each input. We also find that a meta controller trained on a certain backbone model and tasks can successfully generalize to unseen models and tasks.","{'model': 'tldr@v2.0.0', 'text': 'DynaICL is proposed, a recipe for efficient prompting with black-box generalist models that dynamically allocate in-context examples according to the input complexity and the computational budget and it is found that a meta controller trained on a certain backbone model and tasks can successfully generalize to unseen models and tasks.'}",http://arxiv.org/pdf/2305.11170
-SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning,Hanrui Wang,"The attention mechanism is becoming increasingly popular in Natural Language Processing (NLP) applications, showing superior performance than convolutional and recurrent architectures. However, general-purpose platforms such as CPUs and GPUs are inefficient when performing attention inference due to complicated data movement and low arithmetic intensity. Moreover, existing NN accelerators mainly focus on optimizing convolutional or recurrent models, and cannot efficiently support attention. In this paper, we present SpAtten, an efficient algorithm-architecture co-design that leverages token sparsity, head sparsity, and quantization opportunities to reduce the attention computation and memory access. Inspired by the high redundancy of human languages, we propose the novel cascade token pruning to prune away unimportant tokens in the sentence. We also propose cascade head pruning to remove unessential heads. Cascade pruning is fundamentally different from weight pruning since there is no trainable weight in the attention mechanism, and the pruned tokens and heads are selected on the fly. To efficiently support them on hardware, we design a novel top-k engine to rank token and head importance scores with high throughput. Furthermore, we propose progressive quantization that first fetches MSBs only and performs the computation; if the confidence is low, it fetches LSBs and recomputes the attention outputs, trading computation for memory reduction.Extensive experiments on 30 benchmarks show that, on average, SpAtten reduces DRAM access by 10.0× with no accuracy loss, and achieves 1.6×, 3.0×, 162×, 347× speedup, and 1.4×, 3.2×, 1193×, 4059× energy savings over A3 accelerator, MNNFast accelerator, TITAN Xp GPU, Xeon CPU, respectively.","{'model': 'tldr@v2.0.0', 'text': 'SpAtten is presented, an efficient algorithm-architecture co-design that leverages token sparsity, head Sparsity, and quantization opportunities to reduce the attention computation and memory access and proposes the novel cascade token pruning to prune away unimportant tokens in the sentence.'}",https://arxiv.org/pdf/2012.09852
-PoNet: Pooling Network for Efficient Token Mixing in Long Sequences,Chao-Hong Tan,"Transformer-based models have achieved great success in various NLP, vision, and speech tasks. However, the core of Transformer, the self-attention mechanism, has a quadratic time and memory complexity with respect to the sequence length, which hinders applications of Transformer-based models to long sequences. Many approaches have been proposed to mitigate this problem, such as sparse attention mechanisms, low-rank matrix approximations and scalable kernels, and token mixing alternatives to self-attention. We propose a novel Pooling Network (PoNet) for token mixing in long sequences with linear complexity. We design multi-granularity pooling and pooling fusion to capture different levels of contextual information and combine their interactions with tokens. On the Long Range Arena benchmark, PoNet significantly outperforms Transformer and achieves competitive accuracy, while being only slightly slower than the fastest model, FNet, across all sequence lengths measured on GPUs. We also conduct systematic studies on the transfer learning capability of PoNet and observe that PoNet achieves 95.7% of the accuracy of BERT on the GLUE benchmark, outperforming FNet by 4.5% relative. Comprehensive ablation analysis demonstrates effectiveness of the designed multi-granularity pooling and pooling fusion for token mixing in long sequences and efficacy of the designed pre-training tasks for PoNet to learn transferable contextualized language representations.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Pooling Network (PoNet) for token mixing in long sequences with linear complexity and designs multi-granularity pooling and pooling fusion to capture different levels of contextual information and combine their interactions with tokens.'}",
-Token system‐based efficient route optimization in mobile ad hoc network for vehicular ad hoc network in smart city,S. Sultanuddin,"Mobile ad hoc network (MANET) is a well‐known wireless hoc network, which several nodes combined together to form a network. These nodes do not have any structure that forms network by self‐configuring. In the global area, MANET is one of the popular technologies. At present, the most important real‐time challenge in MANET is efficient routing. This is because of dynamic topology and resource constraints. Several existing researcher's workfare were done in an achieving efficient routing mechanism. In our work, we propose a token agent‐based route planning protocol for achieving reliable routing. Vehicular ad hoc network (VANET) is advanced concept of the MANET. VANET can be performed in various smart cities, and it travels each city from source to destination. This token system consists of a token agent that monitors and operates the whole network. The tokens are responsible for collecting nodes activity details such as mobile node energy, node location, and speed. Later, its current and previous information is up‐to‐date and inmate to the token agent for further process in the network. Our proposed routing protocol is well applicable for typical application and a multimedia application. In order to transfer typical packets, our proposed scheme finds the shortest path based on the collect the information.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a token agent‐based route planning protocol for achieving reliable routing in mobile ad hoc network and its proposed protocol is well applicable for typical application and a multimedia application.'}",
-Making Vision Transformers Efficient from A Token Sparsification View,Shuning Chang,"The quadratic computational complexity to the number of tokens limits the practical applications of Vision Transformers (ViTs). Several works propose to prune redundant tokens to achieve efficient ViTs. However, these methods generally suffer from (i) dramatic accuracy drops, (ii) application difficulty in the local vision transformer, and (iii) non-general-purpose networks for downstream tasks. In this work, we propose a novel Semantic Token ViT (STViT), for efficient global and local vision transformers, which can also be revised to serve as backbone for downstream tasks. The semantic tokens represent cluster centers, and they are initialized by pooling image tokens in space and recovered by attention, which can adaptively represent global or local semantic information. Due to the cluster properties, a few semantic tokens can attain the same effect as vast image tokens, for both global and local vision transformers. For instance, only 16 semantic tokens on DeiT-(Tiny,Small,Base) can achieve the same accuracy with more than 100% inference speed improvement and nearly 60% FLOPs reduction; on Swin-(Tiny,Small,Base), we can employ 16 semantic tokens in each window to further speed it up by around 20% with slight accuracy increase. Besides great success in image classification, we also extend our method to video recognition. In addition, we design a STViT-R(ecovery) network to restore the detailed spatial information based on the STViT, making it work for downstream tasks, which is powerless for previous token sparsification methods. Experiments demonstrate that our method can achieve competitive results compared to the original networks in object detection and instance segmentation, with over 30% FLOPs reduction for backbone.","{'model': 'tldr@v2.0.0', 'text': 'A novel Semantic Token ViT (STViT) is proposed, for efficient global and local vision transformers, which can also be revised to serve as backbone for downstream tasks, and a STViT-R(ecovery) network is designed to restore the detailed spatial information based on the ST ViT, making it work for upstream tasks, which is powerless for previous token sparsification methods.'}",https://arxiv.org/pdf/2303.08685
-Efficient Load-Balancing through Distributed Token Dropping,S. Brandt,"We introduce a new graph problem, the token dropping game, and we show how to solve it efficiently in a distributed setting. We use the token dropping game as a tool to design an efficient distributed algorithm for stable orientations and more generally for locally optimal semi-matchings. The prior work by Czygrinow et al. (DISC 2012) finds a stable orientation in O(Δ^5) rounds in graphs of maximum degree Δ, while we improve it to O(Δ^4) and also prove a lower bound of Ω(Δ). For the more general problem of locally optimal semi-matchings, the prior upper bound is O(S^5) and our new algorithm runs in O(C · S^4) rounds, which is an improvement for C = o(S); here C and S are the maximum degrees of customers and servers, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work uses the token dropping game as a tool to design an efficient distributed algorithm for stable orientations and more generally for locally optimal semi-matchings and introduces a new graph problem, thetoken dropping game, and shows how to solve it efficiently in a distributed setting.'}",https://arxiv.org/pdf/2005.07761
-Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations,Youwei Liang,"Vision Transformers (ViTs) take all the image patches as tokens and construct multi-head self-attention (MHSA) among them. Complete leverage of these image tokens brings redundant computations since not all the tokens are attentive in MHSA. Examples include that tokens containing semantically meaningless or distractive image backgrounds do not positively contribute to the ViT predictions. In this work, we propose to reorganize image tokens during the feed-forward process of ViT models, which is integrated into ViT during training. For each forward inference, we identify the attentive image tokens between MHSA and FFN (i.e., feed-forward network) modules, which is guided by the corresponding class token attention. Then, we reorganize image tokens by preserving attentive image tokens and fusing inattentive ones to expedite subsequent MHSA and FFN computations. To this end, our method EViT improves ViTs from two perspectives. First, under the same amount of input image tokens, our method reduces MHSA and FFN computation for efficient inference. For instance, the inference speed of DeiT-S is increased by 50% while its recognition accuracy is decreased by only 0.3% for ImageNet classification. Second, by maintaining the same computational cost, our method empowers ViTs to take more image tokens as input for recognition accuracy improvement, where the image tokens are from higher resolution images. An example is that we improve the recognition accuracy of DeiT-S by 1% for ImageNet classification at the same computational cost of a vanilla DeiT-S. Meanwhile, our method does not introduce more parameters to ViTs. Experiments on the standard benchmarks show the effectiveness of our method. The code is available at https://github.com/youweiliang/evit","{'model': 'tldr@v2.0.0', 'text': 'This work reorganizes image tokens by preserving attentive image tokens and fusing inattentive ones to expedite subsequent MHSA and FFN computations during the feed-forward process of ViT models, which is integrated into ViT during training.'}",
-Frozen CLIP Models are Efficient Video Learners,Ziyi Lin,"Video recognition has been dominated by the end-to-end learning paradigm -- first initializing a video recognition model with weights of a pretrained image model and then conducting end-to-end training on videos. This enables the video network to benefit from the pretrained image model. However, this requires substantial computation and memory resources for finetuning on videos and the alternative of directly using pretrained image features without finetuning the image backbone leads to subpar results. Fortunately, recent advances in Contrastive Vision-Language Pre-training (CLIP) pave the way for a new route for visual recognition tasks. Pretrained on large open-vocabulary image-text pair data, these models learn powerful visual representations with rich semantics. In this paper, we present Efficient Video Learning (EVL) -- an efficient framework for directly training high-quality video recognition models with frozen CLIP features. Specifically, we employ a lightweight Transformer decoder and learn a query token to dynamically collect frame-level spatial features from the CLIP image encoder. Furthermore, we adopt a local temporal module in each decoder layer to discover temporal clues from adjacent frames and their attention maps. We show that despite being efficient to train with a frozen backbone, our models learn high quality video representations on a variety of video recognition datasets. Code is available at https://github.com/OpenGVLab/efficient-video-recognition.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Efficient Video Learning (EVL) -- an efficient framework for directly training high-quality video recognition models with frozen CLIP features that adopt a lightweight Transformer decoder and learn a query token to dynamically collect frame-level spatial features from the CLIP image encoder.'}",http://arxiv.org/pdf/2208.03550
-Towards Efficient Cross-Blockchain Token Transfers,Philipp Frauenthaler,"Interoperability between blockchains remains an open problem, with current interoperability approaches providing very limited means of cross-blockchain interaction, mostly in the form of atomic swaps. More general means of blockchain interoperability such as cross-blockchain data exchange, including cross-blockchain token transfer would contribute to dissolving today’s fragmentation of the research and development field of blockchains. To address this issue, within the TAST research project, a cross-blockchain token was developed. However, the developed solution suffers from high synchronization cost. In this paper, we discuss requirements for more efficient crossblockchain token transfers, describe open research challenges, and give an outlook on two approaches aiming to overcome these challenges.","{'model': 'tldr@v2.0.0', 'text': 'Requirements for more efficient crossblockchain token transfers are discussed, open research challenges are described, and an outlook on two approaches aiming to overcome these challenges are given.'}",
-Training data-efficient image transformers & distillation through attention,Hugo Touvron,"Recently, neural networks purely based on attention were shown to address image understanding tasks such as image classification. However, these visual transformers are pre-trained with hundreds of millions of images using an expensive infrastructure, thereby limiting their adoption. In this work, we produce a competitive convolution-free transformer by training on Imagenet only. We train them on a single computer in less than 3 days. Our reference vision transformer (86M parameters) achieves top-1 accuracy of 83.1% (single-crop evaluation) on ImageNet with no external data. More importantly, we introduce a teacher-student strategy specific to transformers. It relies on a distillation token ensuring that the student learns from the teacher through attention. We show the interest of this token-based distillation, especially when using a convnet as a teacher. This leads us to report results competitive with convnets for both Imagenet (where we obtain up to 85.2% accuracy) and when transferring to other tasks. We share our code and models.","{'model': 'tldr@v2.0.0', 'text': 'This work produces a competitive convolution-free transformer by training on Imagenet only and introduces a teacher-student strategy specific to transformers that relies on a distillation token ensuring that the student learns from the teacher through attention.'}",
-A New Generation of Perspective API: Efficient Multilingual Character-level Transformers,Alyssa Lees,"On the world wide web, toxic content detectors are a crucial line of defense against potentially hateful and offensive messages. As such, building highly effective classifiers that enable a safer internet is an important research area. Moreover, the web is a highly multilingual, cross-cultural community that develops its own lingo over time. As such, it is crucial to develop models that are effective across a diverse range of languages, usages, and styles. In this paper, we present the fundamentals behind the next version of the Perspective API from Google Jigsaw. At the heart of the approach is a single multilingual token-free Charformer model that is applicable across a range of languages, domains, and tasks. We demonstrate that by forgoing static vocabularies, we gain flexibility across a variety of settings. We additionally outline the techniques employed to make such a byte-level model efficient and feasible for productionization. Through extensive experiments on multilingual toxic comment classification benchmarks derived from real API traffic and evaluation on an array of code-switching, covert toxicity, emoji-based hate, human-readable obfuscation, distribution shift, and bias evaluation settings, we show that our proposed approach outperforms strong baselines. Finally, we present our findings from deploying this system in production.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the fundamentals behind the next version of the Perspective API from Google Jigsaw, and presents a single multilingual token-free Charformer model that is applicable across a range of languages, domains, and tasks.'}",https://dl.acm.org/doi/pdf/10.1145/3534678.3539147
-ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction,Keshav Santhanam,"Neural information retrieval (IR) has greatly advanced search and other knowledge-intensive language tasks. While many neural IR methods encode queries and documents into single-vector representations, late interaction models produce multi-vector representations at the granularity of each token and decompose relevance modeling into scalable token-level computations. This decomposition has been shown to make late interaction more effective, but it inflates the space footprint of these models by an order of magnitude. In this work, we introduce Maize, a retriever that couples an aggressive residual compression mechanism with a denoised supervision strategy to simultaneously improve the quality and space footprint of late interaction. We evaluate Maize across a wide range of benchmarks, establishing state-of-the-art quality within and outside the training domain while reducing the space footprint of late interaction models by 6–10x.","{'model': 'tldr@v2.0.0', 'text': 'Maize is introduced, a retriever that couples an aggressive residual compression mechanism with a denoised supervision strategy to simultaneously improve the quality and space footprint of late interaction and establishes state-of-the-art quality within and outside the training domain.'}",https://aclanthology.org/2022.naacl-main.272.pdf
-Abdomen CT Multi-organ Segmentation Using Token-based MLP-Mixer.,Shaoyan Pan,"BACKGROUND Manual contouring is very labor-intensive, time-consuming, and subject to intra- and inter-observer variability. An automated deep learning approach to fast and accurate contouring and segmentation is desirable during radiotherapy treatment planning. PURPOSE This work investigates an efficient deep-learning-based segmentation algorithm in abdomen computed tomography (CT) to facilitate radiation treatment planning. METHODS In this work, we propose a novel deep-learning model utilizing U-shaped Multi-Layer Perceptron Mixer (MLP-Mixer) and convolutional neural network (CNN) for multi-organ segmentation in abdomen CT images. The proposed model has a similar structure to V-net, while a proposed MLP-Convolutional block replaces each convolutional block. The MLP-Convolutional block consists of three components: an early convolutional block for local features extraction and feature resampling, a token-based MLP-Mixer layer for capturing global features with high efficiency, and a token projector for pixel-level detail recovery. We evaluate our proposed network using: 1) an institutional dataset with 60 patient cases, and 2) a public dataset (BCTV) with 30 patient cases. The network performance was quantitatively evaluated in three domains: 1) volume similarity between the ground truth contours and the network predictions using the Dice score coefficient (DSC), sensitivity, and precision; 2) surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMS); 3) the computational complexity reported by the number of network parameters, training time, and inference time. The performance of the proposed network is compared with other state-of-the-art networks. RESULTS In the institutional dataset, the proposed network achieved the following volume similarity measures when averaged over all organs: DSC = 0.912, sensitivity = 0.917, precision = 0.917, average surface similarities were HD = 11.95mm, MSD = 1.90mm, RMS = 3.86mm. The proposed network achieved DSC = 0.786 and HD = 9.04mm on the public dataset. The network also shows statistically significant improvement, which is evaluated by a two-tailed Wilcoxon Mann-Whitney U test, on right lung (MSD where the maximum p-value is 0.001), spinal cord (sensitivity, precision, HD, RMSD where p-value ranges from 0.001 to 0.039), and stomach (DSC where the maximum p-value is 0.01) over all other competing networks. On the public dataset, the network report statistically significant improvement, which is shown by the Wilcoxon Mann-Whitney test, on pancreas (HD where the maximum p-value is 0.006), left (HD where the maximum p-value is 0.022) and right adrenal glands (DSC where the maximum p-value is 0.026). In both datasets, the proposed method can generate contours in less than five seconds. Overall, the proposed MLP-Vnet demonstrates comparable or better performance than competing methods with much lower memory complexity and higher speed. CONCLUSIONS The proposed MLP-Vnet demonstrates superior segmentation performance, in terms of accuracy and efficiency, relative to state-of-the-art methods. This reliable and efficient method demonstrates potential to streamline clinical workflows in abdominal radiotherapy, which may be especially important for online adaptive treatments. This article is protected by copyright. All rights reserved.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates an efficient deep-learning-based segmentation algorithm in abdomen computed tomography (CT) to facilitate radiation treatment planning and demonstrates potential to streamline clinical workflows in abdominal radiotherapy, which may be especially important for online adaptive treatments.'}",
-Hydra Attention: Efficient Attention with Many Heads,Daniel Bolya,"While transformers have begun to dominate many tasks in vision, applying them to large images is still computationally difficult. A large reason for this is that self-attention scales quadratically with the number of tokens, which in turn, scales quadratically with the image size. On larger images (e.g., 1080p), over 60% of the total computation in the network is spent solely on creating and applying attention matrices. We take a step toward solving this issue by introducing Hydra Attention, an extremely efficient attention operation for Vision Transformers (ViTs). Paradoxically, this efficiency comes from taking multi-head attention to its extreme: by using as many attention heads as there are features, Hydra Attention is computationally linear in both tokens and features with no hidden constants, making it significantly faster than standard self-attention in an off-the-shelf ViT-B/16 by a factor of the token count. Moreover, Hydra Attention retains high accuracy on ImageNet and, in some cases, actually improves it.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Hydra Attention, an extremely efficient attention operation for Vision Transformers (ViTs), which is computationally linear in both tokens and features with no hidden constants, making it significantly faster than standard self-attention in an off-the-shelf ViT-B/16 by a factor of the token count.'}",http://arxiv.org/pdf/2209.07484
-To collaborate or not to collaborate: Prompting upstream eco-efficient innovation in a supply chain,Arda Yenipazarli,,"{'model': 'tldr@v2.0.0', 'text': 'A two-echelon supply chain where an upstream supplier sells through a downstream retailer and the retailer who embodies the channel power sets the product price and sells to consumers is considered, which studies the profit/cost implications of collaboration between two parties for upstream eco-efficient innovation.'}",
-DyTox: Transformers for Continual Learning with DYnamic TOken eXpansion,Arthur Douillard,"Deep network architectures struggle to continually learn new tasks without forgetting the previous tasks. A recent trend indicates that dynamic architectures based on an ex-pansion of the parameters can reduce catastrophic forget-ting efficiently in continual learning. However, existing approaches often require a task identifier at test-time, need complex tuning to balance the growing number of parameters, and barely share any information across tasks. As a result, they struggle to scale to a large number of tasks without significant overhead. In this paper, we propose a transformer architecture based on a dedicated encoder/decoder framework. Critically, the encoder and decoder are shared among all tasks. Through a dynamic expansion of special tokens, we specialize each forward of our decoder network on a task distribution. Our strategy scales to a large number of tasks while having neg-ligible memory and time overheads due to strict control of the expansion of the parameters. Moreover, this efficient strategy doesn't need any hyperparameter tuning to control the network's expansion. Our model reaches excellent results on CIFAR100 and state-of-the-art performances on the large-scale ImageNet100 and ImageNet100 while having fewer parameters than concurrent dynamic frameworks.11Code is released at https://github.com/arthurdouillard/dytox.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a transformer architecture based on a dedicated encoder/decoder framework that reaches excellent results on CIFAR100 and state-of-the-art performances on the large-scale ImageNet 100 and ImageNet100 while having fewer parameters than concurrent dynamic frameworks.'}",https://arxiv.org/pdf/2111.11326
-TACo: Token-aware Cascade Contrastive Learning for Video-Text Alignment,Jianwei Yang,"Contrastive learning has been widely used to train transformer-based vision-language models for video-text alignment and multi-modal representation learning. This paper presents a new algorithm called Token-Aware Cascade contrastive learning (TACo) that improves contrastive learning using two novel techniques. The first is the token-aware contrastive loss which is computed by taking into account the syntactic classes of words. This is motivated by the observation that for a video-text pair, the content words in the text, such as nouns and verbs, are more likely to be aligned with the visual contents in the video than the function words. Second, a cascade sampling method is applied to generate a small set of hard negative examples for efficient loss estimation for multi-modal fusion layers. To validate the effectiveness of TACo, in our experiments we finetune pretrained models for a set of downstream tasks including text-video retrieval (YouCook2, MSR-VTT and ActivityNet), video action step localization (CrossTask), video action segmentation (COIN). The results show that our models attain consistent improvements across different experimental settings over previous methods, set-ting new state-of-the-art on three public text-video retrieval benchmarks of YouCook2, MSR-VTT and ActivityNet.","{'model': 'tldr@v2.0.0', 'text': 'A new algorithm called Token-Aware Cascade contrastive learning (TACo) is presented that improves Contrastive learning using two novel techniques, the token-aware contrastive loss which is computed by taking into account the syntactic classes of words.'}",https://arxiv.org/pdf/2108.09980
-Learned Token Pruning for Transformers,Sehoon Kim,"Efficient deployment of transformer models in practice is challenging due to their inference cost including memory footprint, latency, and power consumption, which scales quadratically with input sequence length. To address this, we present a novel token reduction method dubbed Learned Token Pruning (LTP) which adaptively removes unimportant tokens as an input sequence passes through transformer layers. In particular, LTP prunes tokens with an attention score below a threshold, whose value is learned for each layer during training. Our threshold-based method allows the length of the pruned sequence to vary adaptively based on the input sequence, and avoids algorithmically expensive operations such as top-k token selection. We extensively test the performance of LTP on GLUE and SQuAD tasks and show that our method outperforms the prior state-of-the-art token pruning methods by up to ∽2.5% higher accuracy with the same amount of FLOPs. In particular, LTP achieves up to 2.1× FLOPs reduction with less than 1% accuracy drop, which results in up to 1.9× and 2.0× throughput improvement on Intel Haswell CPUs and NVIDIA V100 GPUs. Furthermore, we demonstrate that LTP is more robust than prior methods to variations in input sequence lengths. Our code has been developed in PyTorch and open-sourced","{'model': 'tldr@v2.0.0', 'text': 'A novel token reduction method dubbed Learned Token Pruning (LTP) which adaptively removes unimportant tokens as an input sequence passes through transformer layers, which is more robust than prior methods to variations in input sequence lengths.'}",https://dl.acm.org/doi/pdf/10.1145/3534678.3539260
-Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet,Li Yuan,"Transformers, which are popular for language modeling, have been explored for solving vision tasks recently, e.g., the Vision Transformer (ViT) for image classification. The ViT model splits each image into a sequence of tokens with fixed length and then applies multiple Transformer layers to model their global relation for classification. However, ViT achieves inferior performance to CNNs when trained from scratch on a midsize dataset like ImageNet. We find it is because: 1) the simple tokenization of input images fails to model the important local structure such as edges and lines among neighboring pixels, leading to low training sample efficiency; 2) the redundant attention backbone design of ViT leads to limited feature richness for fixed computation budgets and limited training samples. To overcome such limitations, we propose a new Tokens-To-Token Vision Transformer (T2T-VTT), which incorporates 1) a layer-wise Tokens-to-Token (T2T) transformation to progressively structurize the image to tokens by recursively aggregating neighboring Tokens into one Token (Tokens-to-Token), such that local structure represented by surrounding tokens can be modeled and tokens length can be reduced; 2) an efficient backbone with a deep-narrow structure for vision transformer motivated by CNN architecture design after empirical study. Notably, T2T-ViT reduces the parameter count and MACs of vanilla ViT by half, while achieving more than 3.0% improvement when trained from scratch on ImageNet. It also outperforms ResNets and achieves comparable performance with MobileNets by directly training on ImageNet. For example, T2T-ViT with comparable size to ResNet50 (21.5M parameters) can achieve 83.3% top1 accuracy in image resolution 384x384 on ImageNet.1","{'model': 'tldr@v2.0.0', 'text': 'A new Tokens-To-Token Vision Transformer (T2T-VTT), which incorporates an efficient backbone with a deep-narrow structure for vision transformer motivated by CNN architecture design after empirical study and reduces the parameter count and MACs of vanilla ViT by half.'}",https://arxiv.org/pdf/2101.11986
-ResT: An Efficient Transformer for Visual Recognition,Qing-Long Zhang,"This paper presents an efficient multi-scale vision Transformer, called ResT, that capably served as a general-purpose backbone for image recognition. Unlike existing Transformer methods, which employ standard Transformer blocks to tackle raw images with a fixed resolution, our ResT have several advantages: (1) A memory-efficient multi-head self-attention is built, which compresses the memory by a simple depth-wise convolution, and projects the interaction across the attention-heads dimension while keeping the diversity ability of multi-heads; (2) Position encoding is constructed as spatial attention, which is more flexible and can tackle with input images of arbitrary size without interpolation or fine-tune; (3) Instead of the straightforward tokenization at the beginning of each stage, we design the patch embedding as a stack of overlapping convolution operation with stride on the 2D-reshaped token map. We comprehensively validate ResT on image classification and downstream tasks. Experimental results show that the proposed ResT can outperform the recently state-of-the-art backbones by a large margin, demonstrating the potential of ResT as strong backbones. The code and models will be made publicly available at https://github.com/wofmanaf/ResT.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed ResT can outperform the recently state-of-the-art backbones by a large margin, demonstrating the potential of ResT as strong backbones.'}",
-Gating Dropout: Communication-efficient Regularization for Sparsely Activated Transformers,R. Liu,"Sparsely activated transformers, such as Mixture of Experts (MoE), have received great interest due to their outrageous scaling capability which enables dramatical increases in model size without significant increases in computational cost. To achieve this, MoE models replace the feedforward sub-layer with Mixture-of-Experts sub-layer in transformers and use a gating network to route each token to its assigned experts. Since the common practice for efficient training of such models requires distributing experts and tokens across different machines, this routing strategy often incurs huge cross-machine communication cost because tokens and their assigned experts likely reside in different machines. In this paper, we propose \emph{Gating Dropout}, which allows tokens to ignore the gating network and stay at their local machines, thus reducing the cross-machine communication. Similar to traditional dropout, we also show that Gating Dropout has a regularization effect during training, resulting in improved generalization performance. We validate the effectiveness of Gating Dropout on multilingual machine translation tasks. Our results demonstrate that Gating Dropout improves a state-of-the-art MoE model with faster wall-clock time convergence rates and better BLEU scores for a variety of model sizes and datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Gating Dropout, which allows tokens to ignore the gating network and stay at their local machines, thus reducing the cross-machine communication and improves a state-of-the-art MoE model with faster wall-clock time convergence rates and better BLEU scores for a variety of model sizes and datasets.'}",http://arxiv.org/pdf/2205.14336
-Building Extraction from Remote Sensing Images with Sparse Token Transformers,Keyan Chen,"Deep learning methods have achieved considerable progress in remote sensing image building extraction. Most building extraction methods are based on Convolutional Neural Networks (CNN). Recently, vision transformers have provided a better perspective for modeling long-range context in images, but usually suffer from high computational complexity and memory usage. In this paper, we explored the potential of using transformers for efficient building extraction. We design an efficient dual-pathway transformer structure that learns the long-term dependency of tokens in both their spatial and channel dimensions and achieves state-of-the-art accuracy on benchmark building extraction datasets. Since single buildings in remote sensing images usually only occupy a very small part of the image pixels, we represent buildings as a set of “sparse” feature vectors in their feature space by introducing a new module called “sparse token sampler”. With such a design, the computational complexity in transformers can be greatly reduced over an order of magnitude. We refer to our method as Sparse Token Transformers (STT). Experiments conducted on the Wuhan University Aerial Building Dataset (WHU) and the Inria Aerial Image Labeling Dataset (INRIA) suggest the effectiveness and efficiency of our method. Compared with some widely used segmentation methods and some state-of-the-art building extraction methods, STT has achieved the best performance with low time cost.","{'model': 'tldr@v2.0.0', 'text': 'An efficient dual-pathway transformer structure that learns the long-term dependency of tokens in both their spatial and channel dimensions and achieves state-of-the-art accuracy on benchmark building extraction datasets is designed.'}",https://www.mdpi.com/2072-4292/13/21/4441/pdf?version=1636075865
-Token Pooling in Vision Transformers,D. Marin,"Despite the recent success in many applications, the high computational requirements of vision transformers limit their use in resource-constrained settings. While many existing methods improve the quadratic complexity of attention, in most vision transformers, self-attention is not the major computation bottleneck, e.g., more than 80% of the computation is spent on fully-connected layers. To improve the computational complexity of all layers, we propose a novel token downsampling method, called Token Pooling, efficiently exploiting redundancies in the images and intermediate token representations. We show that, under mild assumptions, softmax-attention acts as a high-dimensional low-pass (smoothing) filter. Thus, its output contains redundancy that can be pruned to achieve a better trade-off between the computational cost and accuracy. Our new technique accurately approximates a set of tokens by minimizing the reconstruction error caused by downsampling. We solve this optimization problem via cost-efficient clustering. We rigorously analyze and compare to prior downsampling methods. Our experiments show that Token Pooling significantly improves the cost-accuracy trade-off over the state-of-the-art downsampling. Token Pooling is a simple and effective operator that can benefit many architectures. Applied to DeiT, it achieves the same ImageNet top-1 accuracy using 42% fewer computations.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel token downsampling method, called Token Pooling, efficiently exploiting redundancies in the images and intermediate token representations, and shows that it significantly improves the cost-accuracy trade-off over the state-of-the-art downsamplings.'}",
-A-ViT: Adaptive Tokens for Efficient Vision Transformer,Hongxu Yin,"We introduce A - ViT, a method that adaptively adjusts the inference cost of vision transformer (ViT) for images of different complexity. A - ViT achieves this by automatically reducing the number of tokens in vision transformers that are processed in the network as inference proceeds. We refor-mulate Adaptive Computation Time (ACT [17]) for this task, extending halting to discard redundant spatial tokens. The appealing architectural properties of vision transformers enables our adaptive token reduction mechanism to speed up inference without modifying the network architecture or inference hardware. We demonstrate that A - ViT requires no extra parameters or sub-network for halting, as we base the learning of adaptive halting on the original network parameters. We further introduce distributional prior regularization that stabilizes training compared to prior ACT approaches. On the image classification task (ImageNet1K), we show that our proposed A - ViT yields high efficacy in filtering informative spatial features and cutting down on the overall compute. The proposed method improves the throughput of DeiT-Tiny by 62% and DeiT-Small by 38% with only 0.3% accuracy drop, outperforming prior art by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces A - ViT, a method that adaptively adjusts the inference cost of vision transformer (ViT) for images of different complexity by automatically reducing the number of tokens in vision transformers that are processed in the network as inference proceeds, and demonstrates that it requires no extra parameters or sub-network for halting.'}",
-Vision Transformer with Super Token Sampling,Huaibo Huang,"Vision transformer has achieved impressive performance for many vision tasks. However, it may suffer from high redundancy in capturing local features for shallow layers. Local self-attention or early-stage convolutions are thus utilized, which sacrifice the capacity to capture long-range dependency. A challenge then arises: can we access efficient and effective global context modeling at the early stages of a neural network? To address this issue, we draw inspiration from the design of superpixels, which reduces the number of image primitives in subsequent processing, and introduce super tokens into vision transformer. Super tokens attempt to provide a semantically meaningful tessellation of visual content, thus reducing the token number in self-attention as well as preserving global modeling. Specifically, we propose a simple yet strong super token attention (STA) mechanism with three steps: the first samples super tokens from visual tokens via sparse association learning, the second performs self-attention on super tokens, and the last maps them back to the original token space. STA decomposes vanilla global attention into multiplications of a sparse association map and a low-dimensional attention, leading to high efficiency in capturing global dependencies. Based on STA, we develop a hierarchical vision transformer. Extensive experiments demonstrate its strong performance on various vision tasks. In particular, without any extra training data or label, it achieves 86.4% top-1 accuracy on ImageNet-1K with less than 100M parameters. It also achieves 53.9 box AP and 46.8 mask AP on the COCO detection task, and 51.9 mIOU on the ADE20K semantic segmentation task. Code will be released at https://github.com/hhb072/SViT.","{'model': 'tldr@v2.0.0', 'text': 'A simple yet strong super token attention (STA) mechanism with three steps: the first samples super tokens from visual tokens via sparse association learning, the second performs self-attention on super tokens, and the last maps them back to the original token space.'}",http://arxiv.org/pdf/2211.11167
-Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognition,Yulin Wang,"Vision Transformers (ViT) have achieved remarkable success in large-scale image recognition. They split every 2D image into a fixed number of patches, each of which is treated as a token. Generally, representing an image with more tokens would lead to higher prediction accuracy, while it also results in drastically increased computational cost. To achieve a decent trade-off between accuracy and speed, the number of tokens is empirically set to 16x16 or 14x14. In this paper, we argue that every image has its own characteristics, and ideally the token number should be conditioned on each individual input. In fact, we have observed that there exist a considerable number of""easy""images which can be accurately predicted with a mere number of 4x4 tokens, while only a small fraction of""hard""ones need a finer representation. Inspired by this phenomenon, we propose a Dynamic Transformer to automatically configure a proper number of tokens for each input image. This is achieved by cascading multiple Transformers with increasing numbers of tokens, which are sequentially activated in an adaptive fashion at test time, i.e., the inference is terminated once a sufficiently confident prediction is produced. We further design efficient feature reuse and relationship reuse mechanisms across different components of the Dynamic Transformer to reduce redundant computations. Extensive empirical results on ImageNet, CIFAR-10, and CIFAR-100 demonstrate that our method significantly outperforms the competitive baselines in terms of both theoretical computational efficiency and practical inference speed. Code and pre-trained models (based on PyTorch and MindSpore) are available at https://github.com/blackfeather-wang/Dynamic-Vision-Transformer and https://github.com/blackfeather-wang/Dynamic-Vision-Transformer-MindSpore.","{'model': 'tldr@v2.0.0', 'text': 'This paper argues that every image has its own characteristics, and ideally the token number should be conditioned on each individual input, and proposes a Dynamic Transformer to automatically configure a proper number of tokens for each input image.'}",
-"Efficient, Real-World Token Bucket Configuration for Residential Gateways",Fengyu Gao,"Internet service providers should deploy effective active queue management (AQM) strategies to provide high-bandwidth low-latency access links to customers. However, they do not, and customers see high latencies, which are problematic for latency-sensitive applications (e.g., VoIP). As a result, customers have to deploy token buckets, but they are hard to configure. In this paper, we evaluate different token bucket configurations and find that the anecdotal evidence regarding token bucket configuration is not optimal. We analyze different configurations using the ns-2 simulator and find a formula to derive optimal parameters depending on the link bandwidth, which brings about much lower latency and higher throughput.","{'model': 'tldr@v2.0.0', 'text': 'This paper evaluates different token bucket configurations and finds that the anecdotal evidence regarding token bucket configuration is not optimal, and finds a formula to derive optimal parameters depending on the link bandwidth, which brings about much lower latency and higher throughput.'}",
-Sparse DETR: Efficient End-to-End Object Detection with Learnable Sparsity,Byungseok Roh,"DETR is the first end-to-end object detector using a transformer encoder-decoder architecture and demonstrates competitive performance but low computational efficiency on high resolution feature maps. The subsequent work, Deformable DETR, enhances the efficiency of DETR by replacing dense attention with deformable attention, which achieves 10x faster convergence and improved performance. Deformable DETR uses the multiscale feature to ameliorate performance, however, the number of encoder tokens increases by 20x compared to DETR, and the computation cost of the encoder attention remains a bottleneck. In our preliminary experiment, we observe that the detection performance hardly deteriorates even if only a part of the encoder token is updated. Inspired by this observation, we propose Sparse DETR that selectively updates only the tokens expected to be referenced by the decoder, thus help the model effectively detect objects. In addition, we show that applying an auxiliary detection loss on the selected tokens in the encoder improves the performance while minimizing computational overhead. We validate that Sparse DETR achieves better performance than Deformable DETR even with only 10% encoder tokens on the COCO dataset. Albeit only the encoder tokens are sparsified, the total computation cost decreases by 38% and the frames per second (FPS) increases by 42% compared to Deformable DETR. Code is available at https://github.com/kakaobrain/sparse-detr","{'model': 'tldr@v2.0.0', 'text': 'Sparse DETR is proposed that selectively updates only the tokens expected to be referenced by the decoder to help the model effectively detect objects and achieves better performance than Deformable DETR even with only 10% encoder tokens on the COCO dataset.'}",
-Understanding and Overcoming the Challenges of Efficient Transformer Quantization,Yelysei Bondarenko,"Transformer-based architectures have become the de-facto standard models for a wide range of Natural Language Processing tasks. However, their memory footprint and high latency are prohibitive for efficient deployment and inference on resource-limited devices. In this work, we explore quantization for transformers. We show that transformers have unique quantization challenges – namely, high dynamic activation ranges that are difficult to represent with a low bit fixed-point format. We establish that these activations contain structured outliers in the residual connections that encourage specific attention patterns, such as attending to the special separator token. To combat these challenges, we present three solutions based on post-training quantization and quantization-aware training, each with a different set of compromises for accuracy, model size, and ease of use. In particular, we introduce a novel quantization scheme – per-embedding-group quantization. We demonstrate the effectiveness of our methods on the GLUE benchmark using BERT, establishing state-of-the-art results for post-training quantization. Finally, we show that transformer weights and embeddings can be quantized to ultra-low bit-widths, leading to significant memory savings with a minimum accuracy loss. Our source code is available at https://github.com/qualcomm-ai-research/transformer-quantization.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel quantization scheme – per-embedding-group quantization, and shows that transformer weights and embeddings can be quantized to ultra-low bit-widths, leading to significant memory savings with a minimum accuracy loss.'}",https://aclanthology.org/2021.emnlp-main.627.pdf
-Beyond Distillation: Task-level Mixture-of-Experts for Efficient Inference,Sneha Kudugunta,"Sparse Mixture-of-Experts (MoE) has been a successful approach for scaling multilingual translation models to billions of parameters without a proportional increase in training computation. However, MoE models are prohibitively large and practitioners often resort to methods such as distillation for serving. In this work, we investigate routing strategies at different granularity (token, sentence, task) in MoE models to bypass distillation. Experiments on WMT and a web-scale dataset suggest that task-level routing (task-MoE) enables us to extract smaller, ready-to-deploy sub-networks from large sparse models. On WMT, our task-MoE with 32 experts (533M parameters) outperforms the best performing token-level MoE model (token-MoE) by +1.0 BLEU on average across 30 language pairs. The peak inference throughput is also improved by a factor of 1.9x when we route by tasks instead of tokens. While distilling a token-MoE to a smaller dense model preserves only 32% of the BLEU gains, our sub-network task-MoE, by design, preserves all the gains with the same inference cost as the distilled student model. Finally, when scaling up to 200 language pairs, our 128-expert task-MoE (13B parameters) performs competitively with a token-level counterpart, while improving the peak inference throughput by a factor of 2.6x.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on WMT and a web-scale dataset suggest that task-level routing (task-MoE) enables us to extract smaller, ready-to-deploy sub-networks from large sparse models.'}",https://aclanthology.org/2021.findings-emnlp.304.pdf
-"Simple, Secure, Efficient, Lightweight and Token Based Protocol for Mutual Authentication in Wireless Sensor Networks",Reena Rathore,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple, secure, dynamic, scalable, efficient and lightweight protocol for mutual authentication of nodes in WSN based on tokens, which was modeled in scyther for verification and no potential attacks were detected.'}",
-Uplift and Upsample: Efficient 3D Human Pose Estimation with Uplifting Transformers,Moritz Einfalt,"The state-of-the-art for monocular 3D human pose estimation in videos is dominated by the paradigm of 2D-to-3D pose uplifting. While the uplifting methods themselves are rather efficient, the true computational complexity depends on the per-frame 2D pose estimation. In this paper, we present a Transformer-based pose uplifting scheme that can operate on temporally sparse 2D pose sequences but still produce temporally dense 3D pose estimates. We show how masked token modeling can be utilized for temporal upsampling within Transformer blocks. This allows to decouple the sampling rate of input 2D poses and the target frame rate of the video and drastically decreases the total computational complexity. Additionally, we explore the option of pre-training on large motion capture archives, which has been largely neglected so far We evaluate our method on two popular benchmark datasets: Human3.6M and MPI-INF-3DHP. With an MPJPE of 45.0 mm and 46.9 mm, respectively, our proposed method can compete with the state-of-the-art while reducing inference time by a factor of 12. This enables real-time throughput with variable consumer hardware in stationary and mobile applications. We release our code and models at https://github.com/goldbricklemon/uplift-upsample-3dhpe","{'model': 'tldr@v2.0.0', 'text': 'A Transformer-based pose uplifting scheme that can operate on temporally sparse 2D pose sequences but still produce temporally dense 3D pose estimates, and shows how masked token modeling can be utilized for temporal upsampling within Transformer blocks.'}",https://arxiv.org/pdf/2210.06110
-Frozen CLIP Model is An Efficient Point Cloud Backbone,Xiaoshui Huang,"The pretraining-finetuning paradigm has demonstrated great success in NLP and 2D image fields because of the high-quality representation ability and transferability of their pretrained models. However, pretraining such a strong model is difficult in the 3D point cloud field since the training data is limited and point cloud collection is expensive. This paper introduces Efficient Point Cloud Learning (EPCL), an effective and efficient point cloud learner for directly training high-quality point cloud models with a frozen CLIP model. Our EPCL connects the 2D and 3D modalities by semantically aligning the 2D features and point cloud features without paired 2D-3D data. Specifically, the input point cloud is divided into a sequence of tokens and directly fed into the frozen CLIP model to learn point cloud representation. Furthermore, we design a task token to narrow the gap between 2D images and 3D point clouds. Comprehensive experiments on 3D detection, semantic segmentation, classification and few-shot learning demonstrate that the 2D CLIP model can be an efficient point cloud backbone and our method achieves state-of-the-art accuracy on both real-world and synthetic downstream tasks. Code will be available.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Efficient Point Cloud Learning (EPCL), an effective and efficient point cloud learner for directly training high-quality point cloud models with a frozen CLIP model and designs a task token to narrow the gap between 2D images and 3D point clouds.'}",http://arxiv.org/pdf/2212.04098
-DTB-MAC: Dynamic Token-Based MAC Protocol for reliable and efficient beacon broadcasting in VANETs,A. Balador,"Most applications developed for vehicular environments rely on broadcasting as the main mechanism to disseminate their messages. However, in IEEE 802.11p, which is the most widely accepted Medium Access Control (MAC) protocol for vehicular communications, all transmissions remain unacknowledged if broadcasting is used. Furthermore, safety message transmission requires a strict delay limit and a high reliability, which is an issue for random access MAC protocols like IEEE 802.11p. Therefore, transmission reliability becomes the most important issue for broadcast-based services in vehicular environments. In this paper, we propose a hybrid MAC protocol, referred as Dynamic Token-Based MAC Protocol (DTB-MAC). DTB-MAC uses both a token passing mechanism and a random access MAC protocol to prevent channel contention as much as possible, and to improve the reliability of safety message transmissions. Our proposed protocol tries to select the best neighbouring node as the next transmitter, and when it is not possible, or when it causes a high overhead, the random access MAC protocol is used instead. Based on simulation experiments, we show that the DTB-MAC protocol can achieve better performance compared with IEEE 802.11p in terms of channel utilization and beacon delivery ratio.","{'model': 'tldr@v2.0.0', 'text': 'DTB-MAC uses both a token passing mechanism and a random access MAC protocol to prevent channel contention as much as possible, and to improve the reliability of safety message transmissions.'}",https://riunet.upv.es/bitstream/10251/64592/3/DTB-MAC%20%28ninth%20version%29.pdf
-UniFormer: Unified Transformer for Efficient Spatiotemporal Representation Learning,Kunchang Li,"It is a challenging task to learn rich and multi-scale spatiotemporal semantics from high-dimensional videos, due to large local redundancy and complex global dependency between video frames. The recent advances in this research have been mainly driven by 3D convolutional neural networks and vision transformers. Although 3D convolution can efficiently aggregate local context to suppress local redundancy from a small 3D neighborhood, it lacks the capability to capture global dependency because of the limited receptive field. Alternatively, vision transformers can effectively capture long-range dependency by self-attention mechanism, while having the limitation on reducing local redundancy with blind similarity comparison among all the tokens in each layer. Based on these observations, we propose a novel Unified transFormer (UniFormer) which seamlessly integrates merits of 3D convolution and spatiotemporal self-attention in a concise transformer format, and achieves a preferable balance between computation and accuracy. Different from traditional transformers, our relation aggregator can tackle both spatiotemporal redundancy and dependency, by learning local and global token affinity respectively in shallow and deep layers. We conduct extensive experiments on the popular video benchmarks, e.g., Kinetics-400, Kinetics-600, and Something-Something V1&V2. With only ImageNet-1K pretraining, our UniFormer achieves 82.9%/84.8% top-1 accuracy on Kinetics-400/Kinetics-600, while requiring 10x fewer GFLOPs than other state-of-the-art methods. For Something-Something V1 and V2, our UniFormer achieves new state-of-the-art performances of 60.9% and 71.2% top-1 accuracy respectively. Code is available at https://github.com/Sense-X/UniFormer.","{'model': 'tldr@v2.0.0', 'text': 'A novel Unified transFormer (UniFormer) which seamlessly integrates merits of 3D convolution and spatiotemporal self-attention in a concise transformer format, and achieves a preferable balance between computation and accuracy.'}",
-Attention Fusion: a light yet efficient late fusion mechanism for task adaptation in NLU,Jin Cao,"Fine-tuning a pre-trained language model using annotated data has become the de-facto standard for adapting general-purpose pre-trained models like BERT to downstream tasks. However, given the trend of larger pre-trained models, fine-tuning these models for each downstream task is parameter-inefficient and computationally-expensive deeming this approach sub-optimal for adoption by NLU systems. In recent years, various approaches have been proposed for parameter efficient task adaptation such as Adaptor, Bitfit, Prompt tuning, Prefix tuning etc. However, most of these efforts propose to insert task specific parameters in-between or inside intermediate layers of the pre-trained encoder resulting in higher computational cost due to back-propagation of errors to all layers. To mitigate this issue, we propose a light but efficient, attention based fusion module which computes task-attuned token representations by aggregating intermediate layer representations from a pre-trained network. Our proposed fusion module trains only 0.0009% of total parameters and achieves competitive performance to the standard fine-tuning approach on various tasks. It is also decoupled from the pre-trained network making it efficient during computation and scalable during deployment. Last but not the least, we demonstrate that our proposed attention-fusion mechanism can transfer effectively to different languages for further re-use and expansion.","{'model': 'tldr@v2.0.0', 'text': 'A light but efficient, attention based fusion module which computes task-attuned token representations by aggregating intermediate layer representations from a pre-trained network and can transfer effectively to different languages for further re-use and expansion.'}",
-Efficient Grid Exploration with a stationary Token,A. Pelc,"A mobile agent starting at an arbitrary node of an m × k grid, for 1 < m ≤ k, has to explore the grid by visiting all its nodes and traversing all edges. The cost of an exploration algorithm is the number of edge traversals by the agent. Nodes of the grid are unlabeled and ports at each node v have distinct numbers in {0,…, d − 1}, where d = 2, 3, 4 is the degree of v. Port numbering is local, i.e., there is no relation between port numbers at different nodes. When visiting a node the agent sees its degree. It also sees the port number by which it enters a node and can choose the port number by which it leaves a visited node. We are interested in deterministic exploration algorithms working at low cost. We consider the scenario in which the agent is equipped with a stationary token situated at its starting node. The agent sees the token whenever it visits this node. We give an exploration algorithm working at cost O(k2) for 2 × k grids, and at cost O(m2k), for m × k grids, when 2 < m ≤ k.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the scenario in which the agent is equipped with a stationary token situated at its starting node and gives an exploration algorithm working at cost O(k2) for 2 × k grids, and at Cost O(m2k), for m ×k grids, when 2 < m ≤ k.'}",
-Fuzzy-Token: An Adaptive MAC Protocol for Wireless-Enabled Manycores,Antonio Franques,"Recent computer architecture trends herald the arrival of manycores with over one hundred cores on a single chip. In this context, traditional on-chip networks do not scale well in latency or energy consumption, leading to bottlenecks in the execution. The Wireless Network-on-Chip (WNoC) paradigm holds considerable promise for the implementation of on-chip networks that will enable such highly-parallel manycores. However, one of the main challenges in WNoCs is the design of mechanisms that provide fast and efficient access to the wireless channel, while adapting to the changing traffic patterns within and across applications. Existing approaches are either slow or complicated, and do not provide the required adaptivity. In this paper, we propose Fuzzy Token,a simple WNoC protocol that leverages the unique properties of the on-chip scenario to deliver efficient and low-latency access to the wireless channel irrespective of the application characteristics. We substantiate our claim via simulations with a synthetic traffic suite and with real application traces. Fuzzy Tokenconsistently provides one of the lowest packet latencies among the evaluated WNoC MAC protocols. On average, the packet latency in Fuzzy Token is 4.4 × and 2.6 × lower than in a state-of-the art contention-based WNoC MAC protocol and in a token-passing protocol, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Fuzzy Token is proposed, a simple WNoC protocol that leverages the unique properties of the on-chip scenario to deliver efficient and low-latency access to the wireless channel irrespective of the application characteristics, and consistently provides one of the lowest packet latencies among the evaluated W nocturnal MAC protocols.'}",https://upcommons.upc.edu/bitstream/2117/349369/5/Fuzzy_Token_DATE_2021.pdf
-SwapCT: Swap Confidential Transactions for Privacy-Preserving Multi-Token Exchanges,Felix Engelmann,"Abstract Decentralized token exchanges allow for secure trading of tokens without a trusted third party. However, decentralization is mostly achieved at the expense of transaction privacy. For a fair exchange, transactions must remain private to hide the participants and volumes while maintaining the possibility for noninteractive execution of trades. In this paper we present a swap confidential transaction system (SwapCT) which is related to ring confidential transactions (e.g. used in Monero) but supports multiple token types to trade among and enables secure, partial transactions for noninteractive swaps. We prove that SwapCT is secure in a strict, formal model and present its efficient performance in a prototype implementation with logarithmic signature sizes for large anonymity sets. For our construction we design an aggregatable signature scheme which might be of independent interest. Our SwapCT system thereby enables a secure and private exchange for tokens without a trusted third party.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that SwapCT is secure in a strict, formal model and its efficient performance in a prototype implementation with logarithmic signature sizes for large anonymity sets is presented.'}",
-Token-Based Authentication Framework for 5G MEC Mobile Networks,Wojciech Niewolski,"MEC technology provides a distributed computing environment in 5G mobile networks for application and service hosting. It allows customers with different requirements and professional competencies to use the services offered by external suppliers. We consider a service access control framework on 5G MEC networks that is efficient, flexible, and user-friendly. Its central element is the MEC Enabler, which handles AAA requests for stakeholders accessing services hosted on the edge servers. The JSON Web Token (JWT) open standard is a suitable tool for the MEC Enabler to manage access control credentials and transfer them securely between parties. In this paper, in the context of access control, we propose the token reference pattern called JSON MEC Access Token (JMAT) and analyze the effectiveness of its available protection methods in compliance with the standard requirements of MEC-hosted services in 5G networks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes the token reference pattern called JSON MEC Access Token (JMAT) and analyzes the effectiveness of its available protection methods in compliance with the standard requirements of MEC-hosted services in 5G networks.'}",https://www.mdpi.com/2079-9292/10/14/1724/pdf?version=1626755715
-Cross-token Modeling with Conditional Computation,Yuxuan Lou,"Mixture-of-Experts (MoE), a conditional computation architecture, achieved promising performance by scaling local module (i.e. feed-forward network) of transformer. However, scaling the cross-token module (i.e. self-attention) is challenging due to the unstable training. This work proposes Sparse-MLP, an all-MLP model which applies sparsely-activated MLPs to cross-token modeling. Specifically, in each Sparse block of our all-MLP model, we apply two stages of MoE layers: one with MLP experts mixing information within channels along image patch dimension, the other with MLP experts mixing information within patches along the channel dimension. In addition, by proposing importance-score routing strategy for MoE and redesigning the image representation shape, we further improve our model's computational efficiency. Experimentally, we are more computation-efficient than Vision Transformers with comparable accuracy. Also, our models can outperform MLP-Mixer by 2.5\% on ImageNet Top-1 accuracy with fewer parameters and computational cost. On downstream tasks, i.e. Cifar10 and Cifar100, our models can still achieve better performance than baselines.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes Sparse-MLP, an all- MLP model which applies sparsely-activated MLPs to cross-token modeling and improves the model's computational efficiency by proposing importance-score routing strategy for MoE and redesigning the image representation shape.""}",
-Learning Token-Based Representation for Image Retrieval,Hui Wu,"In image retrieval, deep local features learned in a data-driven manner have been demonstrated effective to improve retrieval performance. To realize efficient retrieval on large image database, some approaches quantize deep local features with a large codebook and match images with aggregated match kernel. However, the complexity of these approaches is non-trivial with large memory footprint, which limits their capability to jointly perform feature learning and aggregation. To generate compact global representations while maintaining regional matching capability, we propose a unified framework to jointly learn local feature representation and aggregation. In our framework, we first extract local features using CNNs. Then, we design a tokenizer module to aggregate them into a few visual tokens, each corresponding to a specific visual pattern. This helps to remove background noise, and capture more discriminative regions in the image. Next, a refinement block is introduced to enhance the visual tokens with self-attention and cross-attention. Finally, different visual tokens are concatenated to generate a compact global representation. The whole framework is trained end-to-end with image-level labels. Extensive experiments are conducted to evaluate our approach, which outperforms the state-of-the-art methods on the Revisited Oxford and Paris datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a unified framework to jointly learn local feature representation and aggregation, which outperforms the state-of-the-art methods on the Revisited Oxford and Paris datasets.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20173/19932
-Token Merging for Fast Stable Diffusion,Daniel Bolya,"The landscape of image generation has been forever changed by open vocabulary diffusion models. However, at their core these models use transformers, which makes generation slow. Better implementations to increase the throughput of these transformers have emerged, but they still evaluate the entire model. In this paper, we instead speed up diffusion models by exploiting natural redundancy in generated images by merging redundant tokens. After making some diffusion-specific improvements to Token Merging (ToMe), our ToMe for Stable Diffusion can reduce the number of tokens in an existing Stable Diffusion model by up to 60% while still producing high quality images with-out any extra training. In the process, we speed up image generation by up to 2× and reduce memory consumption by up to 5.6×. Furthermore, this speed-up stacks with efficient implementations such as xFormers, minimally impacting quality while being up to 5.4× faster for large images. Code is available at https://github.com/dbolya/tomesd.","{'model': 'tldr@v2.0.0', 'text': 'This paper speed up diffusion models by exploiting natural redundancy in generated images by merging redundant tokens, and can reduce the number of tokens in an existing Stable Diffusion model by up to 60% while still producing high quality images with-out any extra training.'}",https://arxiv.org/pdf/2303.17604
-Non-human primate token use shows possibilities but also limitations for establishing a form of currency,M. Beran,"Non-human primates evaluate choices based on quantitative information and subjective valuation of options. Non-human primates can learn to value tokens as placeholders for primary rewards (such as food). With those tokens established as a potential form of ‘currency’, it is then possible to examine how they respond to opportunities to earn and use tokens in ways such as accumulating tokens or exchanging tokens with each other or with human experimenters to gain primary rewards. Sometimes, individuals make efficient and beneficial choices to obtain tokens and then exchange them at the right moments to gain optimal reward. Sometimes, they even accumulate such rewards through extended delay of gratification, or through other exchange-based interactions. Thus, non-human primates are capable of associating value to arbitrary tokens that may function as currency-like stimuli, but there also are strong limitations on how non-human primates can integrate such tokens into choice situations or use such tokens to fully ‘symbolize’ economic decision-making. These limitations are important to acknowledge when considering the evolutionary emergence of currency use in our species. This article is part of the theme issue ‘Existence and prevalence of economic behaviours among non-human primates’.","{'model': 'tldr@v2.0.0', 'text': ""There are strong limitations on how non-human primates can integrate such tokens into choice situations or use such tokens to fully ‘symbolize’ economic decision-making when considering the evolutionary emergence of currency use in the authors' species.""}",https://royalsocietypublishing.org/doi/pdf/10.1098/rstb.2019.0675
-2DMAC: A Sustainable and Efficient Medium Access Control Mechanism for Future Wireless NoCs,Sidhartha Sankar Rout,"Wireless Network-on-Chip (WNoC) requires a Medium Access Control (MAC) mechanism for an interference-free sharing of the wireless channel. In traditional MAC, a token is circulated among the Wireless Interfaces (WIs) in a Round Robin manner. The WI with the token holds the channel for a fixed number of cycles. However, the channel requirement of the individual WIs dynamically changes over time due to the varying traffic density across the WNoC. Moreover, the conventional WNoCs give equal importance to all the traffic taking the wireless path and transmit it in an oldest-first manner. Nevertheless, the critical data can degrade the system performance to a large extent by delaying the application runtime if not served promptly. We propose 2DMAC, which can change the token arbitration pattern and tune the channel hold time of each WI based on its runtime traffic density and criticality status. Moreover, 2DMAC prioritizes the critical traffic over the non-critical traffic during the wireless data transfer. The proposed mechanism improves the wireless channel utilization by 15.67% and the network throughput by 29.83% and reduces the critical data latency by 29.77% over the traditional MAC.","{'model': 'tldr@v2.0.0', 'text': '2DMAC is proposed, which can change the token arbitration pattern and tune the channel hold time of each WI based on its runtime traffic density and criticality status, and prioritizes the critical traffic over the non-critical traffic during the wireless data transfer.'}",
-PPT: token-Pruned Pose Transformer for monocular and multi-view human pose estimation,Haoyu Ma,"Recently, the vision transformer and its variants have played an increasingly important role in both monocular and multi-view human pose estimation. Considering image patches as tokens, transformers can model the global dependencies within the entire image or across images from other views. However, global attention is computationally expensive. As a consequence, it is difficult to scale up these transformer-based methods to high-resolution features and many views. In this paper, we propose the token-Pruned Pose Transformer (PPT) for 2D human pose estimation, which can locate a rough human mask and performs self-attention only within selected tokens. Furthermore, we extend our PPT to multi-view human pose estimation. Built upon PPT, we propose a new cross-view fusion strategy, called human area fusion, which considers all human foreground pixels as corresponding candidates. Experimental results on COCO and MPII demonstrate that our PPT can match the accuracy of previous pose transformer methods while reducing the computation. Moreover, experiments on Human 3.6M and Ski-Pose demonstrate that our Multi-view PPT can efficiently fuse cues from multiple views and achieve new state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'The token-Pruned Pose Transformer (PPT) is proposed, which can locate a rough human mask and performs self-attention only within selected tokens and is extended to multi-view human pose estimation and a new cross-view fusion strategy, called human area fusion, which considers all human foreground pixels as corresponding candidates.'}",http://arxiv.org/pdf/2209.08194
-MixMIM: Mixed and Masked Image Modeling for Efficient Visual Representation Learning,Jihao Liu,"In this study, we propose Mixed and Masked Image Modeling (MixMIM), a simple but efficient MIM method that is applicable to various hierarchical Vision Transformers. Existing MIM methods replace a random subset of input tokens with a special [MASK] symbol and aim at reconstructing original image tokens from the corrupted image. However, we find that using the [MASK] symbol greatly slows down the training and causes training-finetuning inconsistency, due to the large masking ratio (e.g., 40% in BEiT). In contrast, we replace the masked tokens of one image with visible tokens of another image, i.e., creating a mixed image. We then conduct dual reconstruction to reconstruct the original two images from the mixed input, which significantly improves efficiency. While MixMIM can be applied to various architectures, this paper explores a simpler but stronger hierarchical Transformer, and scales with MixMIM-B, -L, and -H. Empirical results demonstrate that MixMIM can learn high-quality visual representations efficiently. Notably, MixMIM-B with 88M parameters achieves 85.1% top-1 accuracy on ImageNet-1K by pretraining for 600 epochs, setting a new record for neural networks with comparable model sizes (e.g., ViT-B) among MIM methods. Besides, its transferring performances on the other 6 datasets show MixMIM has better FLOPs / performance tradeoff than previous MIM methods. Code is available at https://github.com/Sense-X/MixMIM.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes Mixed and Masked Image Modeling (MixMIM), a simple but efficient MIM method that is applicable to various hierarchical Vision Transformers, and scales with MixMIM-B, -L, and -H.'}",http://arxiv.org/pdf/2205.13137
-ATTEMPT: Parameter-Efficient Multi-task Tuning via Attentional Mixtures of Soft Prompts,Akari Asai,"This work introduces a new multi-task, parameter-efficient language model (LM) tuning method that learns to transfer knowledge across different tasks via a mixture of soft prompts—small prefix embedding vectors pre-trained for different tasks. Our method, called ATTEMPT (ATTEntional Mixtures of Prompt Tuning), obtains source prompts as encodings of large-scale source tasks into a small number of parameters and trains an attention module to interpolate the source prompts and a newly initialized target prompt for every instance in the target task. During training, only the target task prompt and the attention weights, which are shared between tasks in multi-task training, are updated, while the original LM and source prompts are intact. ATTEMPT is highly parameter-efficient (e.g., updates 2,300 times fewer parameters than full fine-tuning), while it overcomes instability of prompt tuning and achieves high task performance using learned knowledge from high-resource tasks. Moreover, it is modular using pre-trained soft prompts, and can flexibly add or remove source prompts for effective knowledge transfer. Our experimental results across 21 diverse NLP datasets show that ATTEMPT significantly outperforms prompt tuning and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use 10 times more parameters. Finally, ATTEMPT outperforms previous work in few-shot learning settings.","{'model': 'tldr@v2.0.0', 'text': 'A new multi-task, parameter-efficient language model (LM) tuning method that learns to transfer knowledge across different tasks via a mixture of soft prompts—small prefix embedding vectors pre-trained for different tasks that significantly outperforms prompt tuning and outperforms or matches fully fine-tuned or other parameter- efficient tuning approaches that use 10 times more parameters.'}",https://aclanthology.org/2022.emnlp-main.446.pdf
-Data Efficient Masked Language Modeling for Vision and Language,Yonatan Bitton,"Masked language modeling (MLM) is one of the key sub-tasks in vision-language pretraining. In the cross-modal setting, tokens in the sentence are masked at random, and the model predicts the masked tokens given the image and the text. In this paper, we observe several key disadvantages of MLM in this setting. First, as captions tend to be short, in a third of the sentences no token is sampled. Second, the majority of masked tokens are stop-words and punctuation, leading to under-utilization of the image. We investigate a range of alternative masking strategies specific to the cross-modal setting that address these shortcomings, aiming for better fusion of text and image in the learned representation. When pre-training the LXMERT model, our alternative masking strategies consistently improve over the original masking strategy on three downstream tasks, especially in low resource settings. Further, our pre-training approach substantially outperforms the baseline model on a prompt-based probing task designed to elicit image objects. These results and our analysis indicate that our method allows for better utilization of the training data.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates a range of alternative masking strategies specific to the cross-modal setting that address shortcomings of MLM, aiming for better fusion of text and image in the learned representation.'}",https://aclanthology.org/2021.findings-emnlp.259.pdf
-Effective Structured Prompting by Meta-Learning and Representative Verbalizer,Weisen Jiang,"Prompt tuning for pre-trained masked language models (MLM) has shown promising performance in natural language processing tasks with few labeled examples. It tunes a prompt for the downstream task, and a verbalizer is used to bridge the predicted token and label prediction. Due to the limited training data, prompt initialization is crucial for prompt tuning. Recently, MetaPrompting (Hou et al., 2022) uses meta-learning to learn a shared initialization for all task-specific prompts. However, a single initialization is insufficient to obtain good prompts for all tasks and samples when the tasks are complex. Moreover, MetaPrompting requires tuning the whole MLM, causing a heavy burden on computation and memory as the MLM is usually large. To address these issues, we use a prompt pool to extract more task knowledge and construct instance-dependent prompts via attention. We further propose a novel soft verbalizer (RepVerb) which constructs label embedding from feature embeddings directly. Combining meta-learning the prompt pool and RepVerb, we propose MetaPrompter for effective structured prompting. MetaPrompter is parameter-efficient as only the pool is required to be tuned. Experimental results demonstrate that MetaPrompter performs better than the recent state-of-the-arts and RepVerb outperforms existing soft verbalizers.","{'model': 'tldr@v2.0.0', 'text': 'A prompt pool is used to extract more task knowledge and construct instance-dependent prompts via attention and a novel soft verbalizer (RepVerb) is proposed which constructs label embedding from feature embeddings directly and is parameter-efficient as only the pool is required to be tuned.'}",http://arxiv.org/pdf/2306.00618
-Efficient pre-training objectives for Transformers,Luca Di Liello,"The Transformer architecture deeply changed the natural language processing, outperforming all previous state-of-the-art models. However, well-known Transformer models like BERT, RoBERTa, and GPT-2 require a huge compute budget to create a high quality contextualised representation. In this paper, we study several efficient pre-training objectives for Transformers-based models. By testing these objectives on different tasks, we determine which of the ELECTRA model's new features is the most relevant. We confirm that Transformers pre-training is improved when the input does not contain masked tokens and that the usage of the whole output to compute the loss reduces training time. Moreover, inspired by ELECTRA, we study a model composed of two blocks; a discriminator and a simple generator based on a statistical model with no impact on the computational performances. Besides, we prove that eliminating the MASK token and considering the whole output during the loss computation are essential choices to improve performance. Furthermore, we show that it is possible to efficiently train BERT-like models using a discriminative approach as in ELECTRA but without a complex generator, which is expensive. Finally, we show that ELECTRA benefits heavily from a state-of-the-art hyper-parameters search.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that eliminating the MASK token and considering the whole output during the loss computation are essential choices to improve performance and it is shown that ELECTRA benefits heavily from a state-of-the-art hyper-parameters search.'}",
-EETBR: Energy efficient token-based routing for wireless sensor networks,Taner Çevik,"The most significant drawback of wireless sensor networks is energy scarcity. As there is an increasing need for operating these networks for relatively long times, energy saving becomes the key challenge in the design of the architectures and protocols for sensor networks. Therefore, several research studies have been performed for making contributions to the analysis of this energy shortage problem. Most of these research activities have been focused on finding solutions for the energy consumption of the communication unit, which is the dominant energy dissipating component of the sensor nodes. In this paper, a novel, token-based routing protocol adapted with a multitier cluster-based architecture is presented. Most of the other cluster-based schemes mainly focus on intracluster organization and communication. However, it should be mentioned that a considerable amount of energy is dissipated during the intercluster communication when compared with intracluster communication. The architecture proposed here not only deals with intracluster communication, but also considers data aggregation, multihop data transmission, and best-effort next hop selection according to a cost factor that is described for the first time in this paper. The simulation results indicate that this token-based next hop selection method together with the multitier cluster-based architecture achieves a significant amount of energy savings, which inherently yields the prolongation of the network lifetime.","{'model': 'tldr@v2.0.0', 'text': 'A novel, token-based routing protocol adapted with a multitier cluster-based architecture that not only deals with intracluster communication, but also considers data aggregation, multihop data transmission, and best-effort next hop selection according to a cost factor is presented.'}",https://journals.tubitak.gov.tr/cgi/viewcontent.cgi?article=3199&context=elektrik
-LLaMA: Open and Efficient Foundation Language Models,Hugo Touvron,"We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community.","{'model': 'tldr@v2.0.0', 'text': 'LLaMA, a collection of foundation language models ranging from 7B to 65B parameters, is introduced and it is shown that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets.'}",http://arxiv.org/pdf/2302.13971
-Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing,Zihang Dai,"With the success of language pretraining, it is highly desirable to develop more efficient architectures of good scalability that can exploit the abundant unlabeled data at a lower cost. To improve the efficiency, we examine the much-overlooked redundancy in maintaining a full-length token-level presentation, especially for tasks that only require a single-vector presentation of the sequence. With this intuition, we propose Funnel-Transformer which gradually compresses the sequence of hidden states to a shorter one and hence reduces the computation cost. More importantly, by re-investing the saved FLOPs from length reduction in constructing a deeper or wider model, we further improve the model capacity. In addition, to perform token-level predictions as required by common pretraining objectives, Funnel-Transformer is able to recover a deep representation for each token from the reduced hidden sequence via a decoder. Empirically, with comparable or fewer FLOPs, Funnel-Transformer outperforms the standard Transformer on a wide variety of sequence-level prediction tasks, including text classification, language understanding, and reading comprehension. The code and pretrained checkpoints are available at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Funnel-Transformer, a model which gradually compresses the sequence of hidden states to a shorter one and hence reduces the computation cost and outperforms the standard Transformer on a wide variety of sequence-level prediction tasks.'}",
-Prompting classes: Exploring the Power of Prompt Class Learning in Weakly Supervised Semantic Segmentation,Balamurali Murugesan,"Recently, CLIP-based approaches have exhibited remarkable performance on generalization and few-shot learning tasks, fueled by the power of contrastive language-vision pre-training. In particular, prompt tuning has emerged as an effective strategy to adapt the pre-trained language-vision models to downstream tasks by employing task-related textual tokens. Motivated by this progress, in this work we question whether other fundamental problems, such as weakly supervised semantic segmentation (WSSS), can benefit from prompt tuning. Our findings reveal two interesting observations that shed light on the impact of prompt tuning on WSSS. First, modifying only the class token of the text prompt results in a greater impact on the Class Activation Map (CAM), compared to arguably more complex strategies that optimize the context. And second, the class token associated with the image ground truth does not necessarily correspond to the category that yields the best CAM. Motivated by these observations, we introduce a novel approach based on a PrOmpt cLass lEarning (POLE) strategy. Through extensive experiments we demonstrate that our simple, yet efficient approach achieves SOTA performance in a well-known WSSS benchmark. These results highlight not only the benefits of language-vision models in WSSS but also the potential of prompt learning for this problem. The code is available at https://github.com/rB080/WSS_POLE.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach based on a PrOmpt cLass lEarning (POLE) strategy achieves SOTA performance in a well-known WSSS benchmark and highlights not only the benefits of language-vision models in W SSS but also the potential of prompt learning for this problem.'}",https://arxiv.org/pdf/2307.00097
-ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models,Binfeng Xu,"Augmented Language Models (ALMs) blend the reasoning capabilities of Large Language Models (LLMs) with tools that allow for knowledge retrieval and action execution. Existing ALM systems trigger LLM thought processes while pulling observations from these tools in an interleaved fashion. Specifically, an LLM reasons to call an external tool, gets halted to fetch the tool's response, and then decides the next action based on all preceding response tokens. Such a paradigm, though straightforward and easy to implement, often leads to huge computation complexity from redundant prompts and repeated execution. This study addresses such challenges for the first time, proposing a modular paradigm ReWOO (Reasoning WithOut Observation) that detaches the reasoning process from external observations, thus significantly reducing token consumption. Comprehensive evaluations across six public NLP benchmarks and a curated dataset reveal consistent performance enhancements with our proposed methodology. Notably, ReWOO achieves 5x token efficiency and 4% accuracy improvement on HotpotQA, a multi-step reasoning benchmark. Furthermore, ReWOO demonstrates robustness under tool-failure scenarios. Beyond prompt efficiency, decoupling parametric modules from non-parametric tool calls enables instruction fine-tuning to offload LLMs into smaller language models, thus substantially reducing model parameters. Our illustrative work offloads reasoning ability from 175B GPT3.5 into 7B LLaMA, demonstrating the significant potential for truly efficient and scalable ALM systems.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a modular paradigm ReWOO (Reasoning WithOut Observation) that detaches the reasoning process from external observations, thus significantly reducing token consumption and demonstrating robustness under tool-failure scenarios.'}",http://arxiv.org/pdf/2305.18323
-Efficient Domain Adaptation of Language Models via Adaptive Tokenization,Vin Sachidananda,"Contextual embedding-based language models trained on large data sets, such as BERT and RoBERTa, provide strong performance across a wide range of tasks and are ubiquitous in modern NLP. It has been observed that fine-tuning these models on tasks involving data from domains different from that on which they were pretrained can lead to suboptimal performance. Recent work has explored approaches to adapt pretrained language models to new domains by incorporating additional pretraining on domain-specific corpora and task data. We propose an alternative approach for transferring pretrained language models to new domains by adapting their tokenizers. We show that domain-specific subword sequences can be determined efficiently directly from divergences in the conditional token distributions of the base and domain-specific corpora. In datasets from four disparate domains, we find adaptive tokenization on a pretrained RoBERTa model provides greater than 85% of the performance benefits of domain specific pretraining. Our approach produces smaller models and less training and inference time than other approaches using tokenizer augmentation. Although using adaptive tokenization incurs a 6% increase in model parameters (due to the introduction of 10k new domain-specific tokens), our approach, using 64 CPUs, is >72x faster than further pretraining the language model on domain-specific corpora on 8 TPUs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an alternative approach for transferring pretrained language models to new domains by adapting their tokenizers and shows that domain-specific subword sequences can be determined efficiently directly from divergences in the conditional token distributions of the base and domain- specific corpora.'}",https://aclanthology.org/2021.sustainlp-1.16.pdf
-Lightweight Composite Re-Ranking for Efficient Keyword Search with BERT,Yingrui Yang,"Recently transformer-based ranking models have been shown to deliver high relevance for document search and the relevance-efficiency tradeoff becomes important for fast query response times. This paper presents BECR (BERT-based Composite Re-Ranking), a lightweight composite re-ranking scheme that combines deep contextual token interactions and traditional lexical term-matching features. BECR conducts query decomposition and composes a query representation using pre-computable token embeddings based on uni-grams and skip-n-grams, to seek a tradeoff of inference efficiency and relevance. Thus it does not perform expensive transformer computations during online inference, and does not require the use of GPU. This paper describes an evaluation of relevance and efficiency of BECR with several TREC datasets.","{'model': 'tldr@v2.0.0', 'text': 'BECR (BERT-based Composite Re-Ranking), a lightweight composite re-ranking scheme that combines deep contextual token interactions and traditional lexical term-matching features, is presented and an evaluation of relevance and efficiency of BECR with several TREC datasets is described.'}",https://dl.acm.org/doi/pdf/10.1145/3488560.3498495
-Parameter-Efficient Low-Resource Dialogue State Tracking by Prompt Tuning,Mingyu Derek Ma,"Dialogue state tracking (DST) is an important step in dialogue management to keep track of users' beliefs. Existing works fine-tune all language model (LM) parameters to tackle the DST task, which requires significant data and computing resources for training and hosting. The cost grows exponentially in the real-world deployment where dozens of fine-tuned LM are used for different domains and tasks. To reduce parameter size and better utilize cross-task shared information, we propose to use soft prompt token embeddings to learn task properties. Without tuning LM parameters, our method drastically reduces the number of parameters needed to less than 0.5% of prior works while achieves better low-resource DST performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to use soft prompt token embeddings to learn task properties and drastically reduces the number of parameters needed to less than 0.5% of prior works while achieves better low-resource DST performance.'}",http://arxiv.org/pdf/2301.10915
-UniEX: An Effective and Efficient Framework for Unified Information Extraction via a Span-extractive Perspective,Junyu Lu,"We propose a new paradigm for universal information extraction (IE) that is compatible with any schema format and applicable to a list of IE tasks, such as named entity recognition, relation extraction, event extraction and sentiment analysis. Our approach converts the text-based IE tasks as the token-pair problem, which uniformly disassembles all extraction targets into joint span detection, classification and association problems with a unified extractive framework, namely UniEX. UniEX can synchronously encode schema-based prompt and textual information, and collaboratively learn the generalized knowledge from pre-defined information using the auto-encoder language models. We develop a traffine attention mechanism to integrate heterogeneous factors including tasks, labels and inside tokens, and obtain the extraction target via a scoring matrix. Experiment results show that UniEX can outperform generative universal IE models in terms of performance and inference-speed on 14 benchmarks IE datasets with the supervised setting. The state-of-the-art performance in low-resource scenarios also verifies the transferability and effectiveness of UniEX.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a traffine attention mechanism to integrate heterogeneous factors including tasks, labels and inside tokens, and obtain the extraction target via a scoring matrix, and shows that UniEX can outperform generative universal IE models in terms of performance and inference-speed.'}",http://arxiv.org/pdf/2305.10306
-Prompt-tuning in ASR systems for efficient domain-adaptation,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains. Since domain-specific systems perform better than their generic counterparts on in-domain evaluation, the need for memory and compute-efficient domain adaptation is obvious. Particularly, adapting parameter-heavy transformer-based language models used for rescoring ASR hypothesis is challenging. In this work, we overcome the problem using prompt-tuning, a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain. With just a handful of extra parameters per domain, we achieve much better perplexity scores over the baseline of using an unadapted LM. Despite being parameter-efficient, these improvements are comparable to those of fully-fine-tuned models with hundreds of millions of parameters. We replicate our findings in perplexity numbers to Word Error Rate in a domain-specific ASR system for one such domain.","{'model': 'tldr@v2.0.0', 'text': None}",
-Prompt-free and Efficient Language Model Fine-Tuning Anonymous ACL submission,Morgan Funtowicz,"Current methods for few-shot fine-tuning of 001 pretrained masked language model (PLM) require 002 carefully engineered prompts and verbalizers 003 for each new task, to convert examples into a 004 cloze-format that the PLM can score. In this work, 005 we propose PERFECT, a simple and efficient 006 method for few-shot fine-tuning of PLMs without 007 relying on any such handcrafting, which is highly 008 effective given as few as 32 data points. PERFECT 009 makes two key design choices: First, we show 010 that manually engineered task prompts can be 011 replaced with task-specific adapters that enable 012 sample-efficient fine-tuning and reduce memory 013 and storage costs by roughly factors of 5 and 100, 014 respectively. Second, instead of using handcrafted 015 verbalizers, we learn a new multi-token label em016 bedding during fine-tuning which are not tied to 017 the model vocabulary and which allow us to avoid 018 complex auto-regressive decoding. These embed019 dings are not only learnable from limited data but 020 also enable nearly 100x faster training and infer021 ence. Experiments on a wide range of few shot 022 NLP tasks demonstrate that PERFECT, while be023 ing simple and efficient, also outperforms existing 024 state-of-the-art few-shot learning methods.1 025","{'model': 'tldr@v2.0.0', 'text': 'Experiments on a wide range of few shot NLP tasks demonstrate that PERFECT, while be023 ing simple and efficient, also outperforms existing 024 state-of-the-art few-shot learning methods.'}",
-Fine-tuned CLIP Models are Efficient Video Learners,H. Rasheed,"Large-scale multi-modal training with image-text pairs imparts strong generalization to CLIP model. Since training on a similar scale for videos is infeasible, recent approaches focus on the effective transfer of image-based CLIP to the video domain. In this pursuit, new parametric modules are added to learn temporal information and inter-frame relationships which require meticulous design efforts. Furthermore, when the resulting models are learned on videos, they tend to overfit on the given task distribution and lack in generalization aspect. This begs the following question: How to effectively transfer image-level CLIP representations to videos? In this work, we show that a simple Video Fine-tuned CLIP (ViFi-CLIP) baseline is generally sufficient to bridge the domain gap from images to videos. Our qualitative analysis illustrates that the frame-level processing from CLIP image-encoder followed by feature pooling and similarity matching with corresponding text embeddings helps in implicitly modeling the temporal cues within ViFi-CLIP. Such fine-tuning helps the model to focus on scene dynamics, moving objects and inter-object relationships. For low-data regimes where full fine-tuning is not viable, we propose a ‘bridge and prompt’ approach that first uses fine-tuning to bridge the domain gap and then learns prompts on language and vision side to adapt CLIP representations. We extensively evaluate this simple yet strong baseline on zero-shot, base-to-novel generalization, few-shot and fully supervised settings across five video benchmarks. Our code and pre-trained models are available at https://github.com/muzairkhattak/ViFi-CLIP.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2212.03640
-Efficient Second-Order TreeCRF for Neural Dependency Parsing,Yu Zhang,"In the deep learning (DL) era, parsing models are extremely simplified with little hurt on performance, thanks to the remarkable capability of multi-layer BiLSTMs in context representation. As the most popular graph-based dependency parser due to its high efficiency and performance, the biaffine parser directly scores single dependencies under the arc-factorization assumption, and adopts a very simple local token-wise cross-entropy training loss. This paper for the first time presents a second-order TreeCRF extension to the biaffine parser. For a long time, the complexity and inefficiency of the inside-outside algorithm hinder the popularity of TreeCRF. To address this issue, we propose an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation. Experiments and analysis on 27 datasets from 13 languages clearly show that techniques developed before the DL era, such as structural learning (global TreeCRF loss) and high-order modeling are still useful, and can further boost parsing performance over the state-of-the-art biaffine parser, especially for partially annotated training data. We release our code at https://github.com/yzhangcs/crfpar.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a second-order TreeCRF extension to the biaffine parser, and proposes an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation.'}",https://www.aclweb.org/anthology/2020.acl-main.302.pdf
-Efficient Document Re-Ranking for Transformers by Precomputing Term Representations,Sean MacAvaney,"Deep pretrained transformer networks are effective at various ranking tasks, such as question answering and ad-hoc document ranking. However, their computational expenses deem them cost-prohibitive in practice. Our proposed approach, called PreTTR (Precomputing Transformer Term Representations), considerably reduces the query-time latency of deep transformer networks (up to a 42x speedup on web document ranking) making these networks more practical to use in a real-time ranking scenario. Specifically, we precompute part of the document term representations at indexing time (without a query), and merge them with the query representation at query time to compute the final ranking score. Due to the large size of the token representations, we also propose an effective approach to reduce the storage requirement by training a compression layer to match attention scores. Our compression technique reduces the storage required up to 95% and it can be applied without a substantial degradation in ranking performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed approach, called PreTTR (Precomputing Transformer Term Representations), considerably reduces the query-time latency of deep transformer networks making these networks more practical to use in a real-time ranking scenario.'}",https://arxiv.org/pdf/2004.14255
-SARATHI: Efficient LLM Inference by Piggybacking Decodes with Chunked Prefills,Amey Agrawal,"Large Language Model (LLM) inference consists of two distinct phases - prefill phase which processes the input prompt and decode phase which generates output tokens autoregressively. While the prefill phase effectively saturates GPU compute at small batch sizes, the decode phase results in low compute utilization as it generates one token at a time per request. The varying prefill and decode times also lead to imbalance across micro-batches when using pipeline parallelism, resulting in further inefficiency due to bubbles. We present SARATHI to address these challenges. SARATHI employs chunked-prefills, which splits a prefill request into equal sized chunks, and decode-maximal batching, which constructs a batch using a single prefill chunk and populates the remaining slots with decodes. During inference, the prefill chunk saturates GPU compute, while the decode requests 'piggyback' and cost up to an order of magnitude less compared to a decode-only batch. Chunked-prefills allows constructing multiple decode-maximal batches from a single prefill request, maximizing coverage of decodes that can piggyback. Furthermore, the uniform compute design of these batches ameliorates the imbalance between micro-batches, significantly reducing pipeline bubbles. Our techniques yield significant improvements in inference performance across models and hardware. For the LLaMA-13B model on A6000 GPU, SARATHI improves decode throughput by up to 10x, and accelerates end-to-end throughput by up to 1.33x. For LLaMa-33B on A100 GPU, we achieve 1.25x higher end-to-end-throughput and up to 4.25x higher decode throughput. When used with pipeline parallelism on GPT-3, SARATHI reduces bubbles by 6.29x, resulting in an end-to-end throughput improvement of 1.91x.","{'model': 'tldr@v2.0.0', 'text': 'SARATHI employs chunked-prefills, which splits a prefill request into equal sized chunks, and decode-maximal batching, which constructs a batch using a single prefill chunk and populates the remaining slots with decodes, resulting in significant improvements in inference performance across models and hardware.'}",https://arxiv.org/pdf/2308.16369
-Efficient Wait-k Models for Simultaneous Machine Translation,Maha Elbayad,"Simultaneous machine translation consists in starting output generation before the entire input sequence is available. Wait-k decoders offer a simple but efficient approach for this problem. They first read k source tokens, after which they alternate between producing a target token and reading another source token. We investigate the behavior of wait-k decoding in low resource settings for spoken corpora using IWSLT datasets. We improve training of these models using unidirectional encoders, and training across multiple values of k. Experiments with Transformer and 2D-convolutional architectures show that our wait-k models generalize well across a wide range of latency levels. We also show that the 2D-convolution architecture is competitive with Transformers for simultaneous translation of spoken language.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the behavior of wait-k decoding in low resource settings for spoken corpora using IWSLT datasets, and improves training of these models using unidirectional encoders, and training across multiple values of k.'}",https://arxiv.org/pdf/2005.08595
-Efficient Attentions for Long Document Summarization,L. Huang,"The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with Hepos, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GovReport, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.","{'model': 'tldr@v2.0.0', 'text': 'Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source is proposed, able to process ten times more tokens than existing models that use full attentions.'}",https://aclanthology.org/2021.naacl-main.112.pdf
-Fast and Memory-Efficient Neural Code Completion,Alexey Svyatkovskiy,"Code completion is one of the most widely used features of modern integrated development environments (IDEs). While deep learning has made significant progress in the statistical prediction of source code, state-of-the-art neural network models consume hundreds of megabytes of memory, bloating the development environment. We address this in two steps: first we present a modular neural framework for code completion. This allows us to explore the design space and evaluate different techniques. Second, within this framework we design a novel reranking neural completion model that combines static analysis with granular token encodings. The best neural reranking model consumes just 6 MB of RAM, — 19x less than previous models — computes a single completion in 8 ms, and achieves 90% accuracy in its top five suggestions.","{'model': 'tldr@v2.0.0', 'text': 'A modular neural framework for code completion is presented and a novel reranking neural completion model is designed that combines static analysis with granular token encodings and achieves 90% accuracy in its top five suggestions.'}",https://arxiv.org/pdf/2004.13651
-PROTECT: Efficient Password-Based Threshold Single-Sign-On Authentication for Mobile Users against Perpetual Leakage,Y. Zhang,"Password-based single-sign-on authentication has been widely applied in mobile environments. It enables an identity server to issue authentication tokens to mobile users holding correct passwords. With an authentication token, one can request mobile services from related service providers without multiple registrations. However, if an adversary compromises the identity server, he can retrieve users’ passwords by performing dictionary guessing attacks (DGA) and can overissue authentication tokens to break the security. In this paper, we propose a password-based threshold single-sign-on authentication scheme dubbed PROTECT that thwarts adversaries who can compromise identity server(s), where multiple identity servers are introduced to authenticate mobile users and issue authentication tokens in a threshold way. PROTECT supports key renewal that periodically updates the secret on each identity server to resist perpetual leakage of the secret. Furthermore, PROTECT is secure against off-line DGA: a credential used to authenticate a user is computed from the password and a server-side key. PROTECT is also resistant to online DGA and password testing attacks in an efficient way. We conduct a comprehensive performance evaluation of PROTECT, which demonstrates the high efficiency on the user side in terms of computation and communication and proves that it can be easily deployed on mobile devices.","{'model': 'tldr@v2.0.0', 'text': 'A password-based threshold single-sign-on authentication scheme dubbed PROTECT is proposed that thwarts adversaries who can compromise identity server(s), where multiple identity servers are introduced to authenticate mobile users and issue authentication tokens in a threshold way and proves that it can be easily deployed on mobile devices.'}",
-PANDA: Prompt Transfer Meets Knowledge Distillation for Efficient Model Adaptation,Qihuang Zhong,"Prompt-tuning, which freezes pretrained language models (PLMs) and only fine-tunes few parameters of additional soft prompt, shows competitive performance against full-parameter fine-tuning ( i.e., model-tuning) when the PLM has billions of parameters, but still performs poorly in the case of smaller PLMs. Hence, prompt transfer (PoT), which initializes the target prompt with the trained prompt of similar source tasks, is recently proposed to improve over prompt-tuning. However, such a vanilla PoT approach usually achieves sub-optimal performance, as (i) the PoT is sensitive to the similarity of source-target pair and (ii) directly fine-tuning the prompt initialized with source prompt on target task might lead to catastrophic forgetting of source knowledge. In response to these problems, we propose a new metric to accurately predict the prompt transferability (regarding (i)), and a novel PoT approach (namely PANDA) that leverages the knowledge distillation technique to transfer the “knowledge” from the source prompt to the target prompt in a subtle manner and alleviate the catastrophic forgetting effectively (regarding (ii)). Furthermore, to achieve adaptive prompt transfer for each source-target pair, we use our metric to control the knowledge transfer in our PANDA approach. Extensive and systematic experiments on 189 combinations of 21 source and 9 target datasets across 5 scales of PLMs demonstrate that: 1) our proposed metric works well to predict the prompt transferability ; 2) our PANDA consistently outperforms the vanilla PoT approach by 2.3% average score (up to 24.1%) among all tasks and model sizes ; 3) with our PANDA approach, prompt-tuning can achieve competitive and even better performance than model-tuning in various PLM","{'model': 'tldr@v2.0.0', 'text': 'A new metric to accurately predict the prompt transferability is proposed, and a novel PoT approach (namely PANDA) is proposed that leverages the knowledge distillation technique to transfer the “knowledge” from the source prompt to the target prompt in a subtle manner and alleviate the catastrophic forgetting effectively.'}",http://arxiv.org/pdf/2208.10160
-Token Turing Machines,M. Ryoo,"We propose Token Turing Machines (TTM), a sequential, autoregressive Transformer model with memory for real-world sequential visual understanding. Our model is inspired by the seminal Neural Turing Machine, and has an external memory consisting of a set of tokens which summarise the previous history (i.e., frames). This memory is efficiently addressed, read and written using a Transformer as the processing unit/controller at each step. The model's memory module ensures that a new observation will only be processed with the contents of the memory (and not the entire history), meaning that it can efficiently process long sequences with a bounded computational cost at each step. We show that TTM outperforms other alternatives, such as other Transformer models designed for long sequences and recurrent neural networks, on two real-world sequential visual understanding tasks: online temporal activity detection from videos and vision-based robot action policy learning. Code is publicly available at: https://github.com/google-research/scenic/tree/main/scenic/projects/token.turing.","{'model': 'tldr@v2.0.0', 'text': 'Token Turing Machines outperforms other alternatives, such as other Transformer models designed for long sequences and recurrent neural networks, on two real-world sequential visual understanding tasks: online temporal activity detection from videos and vision-based robot action policy learning.'}",
-Toward Efficient Language Model Pretraining and Downstream Adaptation via Self-Evolution: A Case Study on SuperGLUE,Qihuang Zhong,"This technical report briefly describes our JDExplore d-team's Vega v2 submission on the SuperGLUE leaderboard. SuperGLUE is more challenging than the widely used general language understanding evaluation (GLUE) benchmark, containing eight difficult language understanding tasks, including question answering, natural language inference, word sense disambiguation, coreference resolution, and reasoning. [Method] Instead of arbitrarily increasing the size of a pretrained language model (PLM), our aim is to 1) fully extract knowledge from the input pretraining data given a certain parameter budget, e.g., 6B, and 2) effectively transfer this knowledge to downstream tasks. To achieve goal 1), we propose self-evolution learning for PLMs to wisely predict the informative tokens that should be masked, and supervise the masked language modeling (MLM) process with rectified smooth labels. For goal 2), we leverage the prompt transfer technique to improve the low-resource tasks by transferring the knowledge from the foundation model and related downstream tasks to the target task. [Results] According to our submission record (Oct. 2022), with our optimized pretraining and fine-tuning strategies, our 6B Vega method achieved new state-of-the-art performance on 4/8 tasks, sitting atop the SuperGLUE leaderboard on Oct. 8, 2022, with an average score of 91.3.","{'model': 'tldr@v2.0.0', 'text': ""This technical report briefly describes the JDExplore d-team's Vega v2 submission on the SuperGLUE leaderboard, which achieved new state-of-the-art performance on 4/8 tasks, and the prompt transfer technique to improve the low-resource tasks by transferring the knowledge from the foundation model and related downstream tasks to the target task.""}",https://arxiv.org/pdf/2212.01853
-Vision Transformers are Parameter-Efficient Audio-Visual Learners,Yan-Bo Lin,"Vision transformers (ViTs) have achieved impressive results on various computer vision tasks in the last several years. In this work, we study the capability of frozen ViTs, pretrained only on visual data, to generalize to audio-visual data without finetuning any of its original parameters. To do so, we propose a latent audio-visual hybrid (LAVISH) adapter that adapts pretrained ViTs to audio-visual tasks by injecting a small number of trainable parameters into every layer of a frozen ViT. To efficiently fuse visual and audio cues, our LAVISH adapter uses a small set of latent tokens, which form an attention bottleneck, thus, eliminating the quadratic cost of standard cross-attention. Compared to the existing modality-specific audio-visual methods, our approach achieves competitive or even better performance on various audio-visual tasks while using fewer tunable parameters and without relying on costly audio pretraining or external audio encoders. Our code is available at https://genjib.github.io/project_page/LAVISH/","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a latent audio-visual hybrid (LAVISH) adapter that adapts pretrained ViTs to audio- visual tasks by injecting a small number of trainable parameters into every layer of a frozen ViT, eliminating the quadratic cost of standard cross-attention.'}",https://arxiv.org/pdf/2212.07983
-SaiT: Sparse Vision Transformers through Adaptive Token Pruning,Ling Li,"While vision transformers have achieved impressive results, effectively and efficiently accelerating these models can further boost performances. In this work, we propose a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities. Thus one model offers a range of accuracy and throughput tradeoffs for different applications. Besides, we introduce adaptive token pruning to optimize the patch token sparsity based on the input image. In addition, we investigate knowledge distillation to enhance token selection capability in early transformer modules. Sparse adaptive image Transformer (SaiT) offers varying levels of model acceleration by merely changing the token sparsity on the fly. Specifically, SaiT reduces the computation complexity (FLOPs) by 39% - 43% and increases the throughput by 67% - 91% with less than 0.5% accuracy loss for various vision transformer models. Meanwhile, the same model also provides the zero accuracy drop option by skipping the sparsification step. SaiT achieves better accuracy and computation tradeoffs than state-of-the-art transformer and convolutional models.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities, and introduces adaptive token pruning to optimize the patch token sparsity based on the input image.'}",http://arxiv.org/pdf/2210.05832
-CPPCD: A Token-Based Approach to Detecting Potential Clones,Yu-Liang Hung,"Most state-of-the-art clone detection approaches are aimed at finding clones accurately and/or efficiently. Yet, whether a code fragment is a clone often varies according to different people's perspectives and different clone detection tools. In this paper, we present CPPCD (CP-based Potential Clone Detection), a novel token-based approach to detecting potential clones. It generates CP (clone probability) values and CP distribution graphs for developers to decide if a method is a clone. We have evaluated our approach on large-scale software projects written in Java. Our experiments suggest that the majority of clones have CP values greater than or equal to 0.75 and that CPPCD is an accurate (with respect to Type-1, Type-2, and Type-3 clones), efficient, and scalable approach to detecting potential clones.","{'model': 'tldr@v2.0.0', 'text': 'CPPCD (CP-based Potential Clone Detection), a novel token-based approach to detecting potential clones, generates CP (clone probability) values and CP distribution graphs for developers to decide if a method is a clone.'}",
-Position-based Prompting for Health Outcome Generation,Micheal Abaho,"Probing factual knowledge in Pre-trained Language Models (PLMs) using prompts has indirectly implied that language models (LMs) can be treated as knowledge bases. To this end, this phenomenon has been effective, especially when these LMs are fine-tuned towards not just data, but also to the style or linguistic pattern of the prompts themselves. We observe that satisfying a particular linguistic pattern in prompts is an unsustainable, time-consuming constraint in the probing task, especially because they are often manually designed and the range of possible prompt template patterns can vary depending on the prompting task. To alleviate this constraint, we propose using a position-attention mechanism to capture positional information of each word in a prompt relative to the mask to be filled, hence avoiding the need to re-construct prompts when the prompts’ linguistic pattern changes. Using our approach, we demonstrate the ability of eliciting answers (in a case study on health outcome generation) to not only common prompt templates like Cloze and Prefix but also rare ones too, such as Postfix and Mixed patterns whose masks are respectively at the start and in multiple random places of the prompt. More so, using various biomedical PLMs, our approach consistently outperforms a baseline in which the default PLMs representation is used to predict masked tokens.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using a position-attention mechanism to capture positional information of each word in a prompt relative to the mask to be filled, hence avoiding the need to re-construct prompts when the prompts’ linguistic pattern changes.'}",http://arxiv.org/pdf/2204.03489
-MC-BERT: Efficient Language Pre-Training via a Meta Controller,Zhenhui Xu,"Pre-trained contextual representations (e.g., BERT) have become the foundation to achieve state-of-the-art results on many NLP tasks. However, large-scale pre-training is computationally expensive. ELECTRA, an early attempt to accelerate pre-training, trains a discriminative model that predicts whether each input token was replaced by a generator. Our studies reveal that ELECTRA's success is mainly due to its reduced complexity of the pre-training task: the binary classification (replaced token detection) is more efficient to learn than the generation task (masked language modeling). However, such a simplified task is less semantically informative. To achieve better efficiency and effectiveness, we propose a novel meta-learning framework, MC-BERT. The pre-training task is a multi-choice cloze test with a reject option, where a meta controller network provides training input and candidates. Results over GLUE natural language understanding benchmark demonstrate that our proposed method is both efficient and effective: it outperforms baselines on GLUE semantic tasks given the same computational budget.","{'model': 'tldr@v2.0.0', 'text': 'Results over GLUE natural language understanding benchmark demonstrate that the proposed MC-BERT method is both efficient and effective: it outperforms baselines on GLUE semantic tasks given the same computational budget.'}",
-DEGREE: A Data-Efficient Generation-Based Event Extraction Model,I-Hung Hsu,"Event extraction requires high-quality expert human annotations, which are usually expensive. Therefore, learning a data-efficient event extraction model that can be trained with only a few labeled examples has become a crucial challenge. In this paper, we focus on low-resource end-to-end event extraction and propose DEGREE, a data-efficient model that formulates event extraction as a conditional generation problem. Given a passage and a manually designed prompt, DEGREE learns to summarize the events mentioned in the passage into a natural sentence that follows a predefined pattern. The final event predictions are then extracted from the generated sentence with a deterministic algorithm. DEGREE has three advantages to learn well with less training data. First, our designed prompts provide semantic guidance for DEGREE to leverage DEGREE and thus better capture the event arguments. Moreover, DEGREE is capable of using additional weakly-supervised information, such as the description of events encoded in the prompts. Finally, DEGREE learns triggers and arguments jointly in an end-to-end manner, which encourages the model to better utilize the shared knowledge and dependencies among them. Our experimental results demonstrate the strong performance of DEGREE for low-resource event extraction.","{'model': 'tldr@v2.0.0', 'text': 'DEGREE is proposed, a data-efficient model that formulates event extraction as a conditional generation problem and learns triggers and arguments jointly in an end-to-end manner, which encourages the model to better utilize the shared knowledge and dependencies among them.'}",https://aclanthology.org/2022.naacl-main.138.pdf
-Efficient Exact Similarity Searches Using Multiple Token Orderings,Jongik Kim,"Similarity searches are essential in many applications including data cleaning and near duplicate detection. Many similarity search algorithms first generate candidate records, and then identify true matches among them. A major focus of those algorithms has been on how to reduce the number of candidate records in the early stage of similarity query processing. One of the most commonly used techniques to reduce the candidate size is the prefix filtering principle, which exploits the document frequency ordering of tokens. In this paper, we propose a novel partitioning technique that considers multiple token orderings based on token co-occurrence statistics. Experimental results show that the proposed technique is effective in reducing the number of candidate records and as a result improves the performance of existing algorithms significantly.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel partitioning technique that considers multiple token orderings based on token co-occurrence statistics and shows that the proposed technique is effective in reducing the number of candidate records and as a result improves the performance of existing algorithms significantly.'}",
-TRIPS: Efficient Vision-and-Language Pre-training with Text-Relevant Image Patch Selection,Chaoya Jiang,"Vision Transformers (ViTs) have been widely used in large-scale Vision and Language Pre-training (VLP) models. Though previous VLP works have proved the effectiveness of ViTs, they still suffer from computational efficiency brought by the long visual sequence. To tackle this problem, in this paper, we propose an efficient vision-and-language pre-training model with Text-Relevant Image Patch Selection, namely TRIPS, which reduces the visual sequence progressively with a text-guided patch-selection layer in the visual backbone for efficient training and inference. The patch-selection layer can dynamically compute text-dependent visual attention to identify the attentive image tokens with text guidance and fuse inattentive ones in an end-to-end manner. Meanwhile, TRIPS does not introduce extra parameters to ViTs. Experimental results on a variety of popular benchmark datasets demonstrate that TRIPS gain a speedup of 40% over previous similar VLP models, yet with competitive or better downstream task performance.","{'model': 'tldr@v2.0.0', 'text': 'An efficient vision-and-language pre-training model with Text-Relevant Image Patch Selection, namely TRIPS, which reduces the visual sequence progressively with a text-guided patch-selection layer in the visual backbone for efficient training and inference.'}",https://aclanthology.org/2022.emnlp-main.273.pdf
-Canine: Pre-training an Efficient Tokenization-Free Encoder for Language Representation,J. Clark,"Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model’s ability to adapt. In this paper, we present Canine, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, Canine combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. Canine outperforms a comparable mBert model by 5.7 F1 on TyDi QA, a challenging multilingual benchmark, despite having fewer model parameters.","{'model': 'tldr@v2.0.0', 'text': 'Canine is presented, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00448/1985933/tacl_a_00448.pdf
-NodePiece: Compositional and Parameter-Efficient Representations of Large Knowledge Graphs,Mikhail Galkin,"Conventional representation learning algorithms for knowledge graphs (KG) map each entity to a unique embedding vector. Such a shallow lookup results in a linear growth of memory consumption for storing the embedding matrix and incurs high computational costs when working with real-world KGs. Drawing parallels with subword tokenization commonly used in NLP, we explore the landscape of more parameter-efficient node embedding strategies with possibly sublinear memory requirements. To this end, we propose NodePiece, an anchor-based approach to learn a fixed-size entity vocabulary. In NodePiece, a vocabulary of subword/sub-entity units is constructed from anchor nodes in a graph with known relation types. Given such a fixed-size vocabulary, it is possible to bootstrap an encoding and embedding for any entity, including those unseen during training. Experiments show that NodePiece performs competitively in node classification, link prediction, and relation prediction tasks while retaining less than 10% of explicit nodes in a graph as anchors and often having 10x fewer parameters. To this end, we show that a NodePiece-enabled model outperforms existing shallow models on a large OGB WikiKG 2 graph having 70x fewer parameters.","{'model': 'tldr@v2.0.0', 'text': 'A NodePiece-enabled model outperforms existing shallow models on a large OGB WikiKG 2 graph having 70x fewer parameters and is shown to perform competitively in node classification, link prediction, and relation prediction tasks.'}",
-LLaMA-Adapter V2: Parameter-Efficient Visual Instruction Model,Peng Gao,"How to efficiently transform large language models (LLMs) into instruction followers is recently a popular research direction, while training LLM for multi-modal reasoning remains less explored. Although the recent LLaMA-Adapter demonstrates the potential to handle visual inputs with LLMs, it still cannot generalize well to open-ended visual instructions and lags behind GPT-4. In this paper, we present LLaMA-Adapter V2, a parameter-efficient visual instruction model. Specifically, we first augment LLaMA-Adapter by unlocking more learnable parameters (e.g., norm, bias and scale), which distribute the instruction-following ability across the entire LLaMA model besides adapters. Secondly, we propose an early fusion strategy to feed visual tokens only into the early LLM layers, contributing to better visual knowledge incorporation. Thirdly, a joint training paradigm of image-text pairs and instruction-following data is introduced by optimizing disjoint groups of learnable parameters. This strategy effectively alleviates the interference between the two tasks of image-text alignment and instruction following and achieves strong multi-modal reasoning with only a small-scale image-text and instruction dataset. During inference, we incorporate additional expert models (e.g. captioning/OCR systems) into LLaMA-Adapter to further enhance its image understanding capability without incurring training costs. Compared to the original LLaMA-Adapter, our LLaMA-Adapter V2 can perform open-ended multi-modal instructions by merely introducing 14M parameters over LLaMA. The newly designed framework also exhibits stronger language-only instruction-following capabilities and even excels in chat interactions. Our code and models are available at https://github.com/ZrrSkywalker/LLaMA-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'This work augments LLaMA-Adapter by unlocking more learnable parameters and proposes an early fusion strategy to feed visual tokens only into the early LLM layers, contributing to better visual knowledge incorporation and achieves strong multi-modal reasoning with only a small-scale image-text and instruction dataset.'}",http://arxiv.org/pdf/2304.15010
-Long Range Arena: A Benchmark for Efficient Transformers,Yi Tay,"Transformers do not scale very well to long sequence lengths largely because of quadratic self-attention complexity. In the recent months, a wide spectrum of efficient, fast Transformers have been proposed to tackle this problem, more often than not claiming superior or comparable model quality to vanilla Transformer models. To this date, there is no well-established consensus on how to evaluate this class of models. Moreover, inconsistent benchmarking on a wide spectrum of tasks and datasets makes it difficult to assess relative model quality amongst many models. This paper proposes a systematic and unified benchmark, LRA, specifically focused on evaluating model quality under long-context scenarios. Our benchmark is a suite of tasks consisting of sequences ranging from $1K$ to $16K$ tokens, encompassing a wide range of data types and modalities such as text, natural, synthetic images, and mathematical expressions requiring similarity, structural, and visual-spatial reasoning. We systematically evaluate ten well-established long-range Transformer models (Reformers, Linformers, Linear Transformers, Sinkhorn Transformers, Performers, Synthesizers, Sparse Transformers, and Longformers) on our newly proposed benchmark suite. LRA paves the way towards better understanding this class of efficient Transformer models, facilitates more research in this direction, and presents new challenging tasks to tackle. Our benchmark code will be released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and unified benchmark, LRA, specifically focused on evaluating model quality under long-context scenarios is proposed, paving the way towards better understanding this class of efficient Transformer models.'}",
-Visual Query Tuning: Towards Effective Usage of Intermediate Representations for Parameter and Memory Efficient Transfer Learning,Cheng-Hao Tu,"Intermediate features of a pretrained model have been shown informative for making accurate predictions on downstream tasks, even if the model backbone is kept frozen. The key challenge is how to utilize these intermediate features given their gigantic amount. We propose visual query tuning (VQT), a simple yet effective approach to aggregate intermediate features of Vision Transformers. Through introducing a handful of learnable “query” tokens to each layer, VQT leverages the inner workings of Transformers to “summarize” rich intermediate features of each layer, which can then be used to train the prediction heads of downstream tasks. As VQT keeps the intermediate features intact and only learns to combine them, it enjoys memory efficiency in training, compared to many other parameter-efficient fine-tuning approaches that learn to adapt features and need back-propagation through the entire backbone. This also suggests the complementary role between VQT and those approaches in transfer learning. Empirically, VQT consistently surpasses the state-of-the-art approach that utilizes intermediate features for transfer learning and outperforms full fine-tuning in many cases. Compared to parameter-efficient approaches that adapt features, VQT achieves much higher accuracy under memory constraints. Most importantly, VQT is compatible with these approaches to attain even higher accuracy, making it a simple add-on to further boost transfer learning. Code is available at https://github.com/andytu28/VQT.","{'model': 'tldr@v2.0.0', 'text': 'Visual query tuning (VQT), a simple yet effective approach to aggregate intermediate features of Vision Transformers, which consistently surpasses the state-of-the-art approach that utilizes intermediate features for transfer learning and outperforms full fine-tuning in many cases.'}",https://arxiv.org/pdf/2212.03220
-Fast-join: An efficient method for fuzzy token matching based string similarity join,Jiannan Wang,"String similarity join that finds similar string pairs between two string sets is an essential operation in many applications, and has attracted significant attention recently in the database community. A significant challenge in similarity join is to implement an effective fuzzy match operation to find all similar string pairs which may not match exactly. In this paper, we propose a new similarity metrics, called “fuzzy token matching based similarity”, which extends token-based similarity functions (e.g., Jaccard similarity and Cosine similarity) by allowing fuzzy match between two tokens. We study the problem of similarity join using this new similarity metrics and present a signature-based method to address this problem. We propose new signature schemes and develop effective pruning techniques to improve the performance. Experimental results show that our approach achieves high efficiency and result quality, and significantly outperforms state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new similarity metrics, called “fuzzy token matching based similarity”, which extends token-based similarity functions by allowing fuzzy match between two tokens, and achieves high efficiency and result quality, and significantly outperforms state-of-the-art methods.'}",http://dbgroup.cs.tsinghua.edu.cn/ligl/papers/icde2011-fastjoin.pdf
-FedPrompt: Communication-Efficient and Privacy-Preserving Prompt Tuning in Federated Learning,Haodong Zhao,"Federated learning (FL) has enabled global model training on decentralized data in a privacy-preserving way. However, for tasks that utilize pre-trained language models (PLMs) with massive parameters, there are considerable communication costs. Prompt tuning, which tunes soft prompts without modifying PLMs, has achieved excellent performance as a new learning paradigm. In this paper, we want to combine these methods and explore the effect of prompt tuning under FL. We propose ""FedPrompt"" studying prompt tuning in a model split aggregation way using FL, and prove that split aggregation greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution. We further conduct backdoor attacks by data poisoning on FedPrompt. Experiments show that attack achieve a quite low attack success rate and can not inject backdoor effectively, proving the robustness of FedPrompt.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that split aggregation greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution, proving the robustness of FedPrompt.'}",https://arxiv.org/pdf/2208.12268
-Token Allocation for Course Bidding With Machine Learning Method,Chonbadee Juthamanee,"Most educational institutions have a problem of allocation course seats to students because demand exceeds supply for many courses. A course bidding system is a tool for improving the registration system based on auction theory to provide course seat allocation equitably and efficiently. This paper proposes a method for token price prediction for allocate to courses through course bidding system. Three methods are compared: Decision Tree, Random Forest and Artificial Neuron Network. The result of the experiment shows that ANN is the best method with lowest RSME 3.98%. Furthermore, it provides an important information to users to estimate the risks on their course bidding strategy.","{'model': 'tldr@v2.0.0', 'text': 'The result of the experiment shows that ANN is the best method with lowest RSME 3.98% and provides an important information to users to estimate the risks on their course bidding strategy.'}",
-Token meets Wallet: Formalizing Privacy and Revocation for FIDO2,L. Hanzlik,"The FIDO2 standard is a widely-used class of challenge-response type protocols that allows to authenticate to an online service using a hardware token. Barbosa et al. (CRYPTO ‘21) provided the first formal security model and analysis for the FIDO2 standard. However, their model has two shortcomings: (1) It does not include privacy, one of the key features claimed by FIDO2. (2) It only covers tokens that store all secret keys locally. In contrast, due to limited memory, most existing FIDO2 tokens either derive all secret keys from a common seed or store keys on the server (the latter approach is also known as key wrapping).In this paper, we revisit the security of the WebAuthn component of FIDO2 as implemented in practice. Our contributions are as follows. (1) We adapt the model of Barbosa et al. so as to capture authentication tokens using key derivation or key wrapping. (2) We provide the first formal definition of privacy for the WebAuthn component of FIDO2. We then prove the privacy of this component in common FIDO2 token implementations if the underlying building blocks are chosen appropriately. (3) We address the unsolved problem of global key revocation in FIDO2. To this end, we introduce and analyze a simple revocation procedure that builds on the popular BIP32 standard used in cryptocurrency wallets and can efficiently be implemented with existing FIDO2 servers.","{'model': 'tldr@v2.0.0', 'text': 'A simple revocation procedure is introduced and analyzed that builds on the popular BIP32 standard used in cryptocurrency wallets and can efficiently be implemented with existing FIDO2 servers and address the unsolved problem of global key revocation in F IDO2.'}",
-To Repeat or Not To Repeat: Insights from Scaling LLM under Token-Crisis,Fuzhao Xue,"Recent research has highlighted the importance of dataset size in scaling language models. However, large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. To further enhance LLMs, a straightforward approach is to repeat the pre-training data for additional epochs. In this study, we empirically investigate three key aspects under this approach. First, we explore the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting, leading to multi-epoch degradation. Second, we examine the key factors contributing to multi-epoch degradation, finding that significant factors include dataset size, model parameters, and training objectives, while less influential factors consist of dataset quality and model FLOPs. Finally, we explore whether widely used regularization can alleviate multi-epoch degradation. Most regularization techniques do not yield significant improvements, except for dropout, which demonstrates remarkable effectiveness but requires careful tuning when scaling up the model size. Additionally, we discover that leveraging mixture-of-experts (MoE) enables cost-effective and efficient hyper-parameter tuning for computationally intensive dense LLMs with comparable trainable parameters, potentially impacting efficient LLM development on a broader scale.","{'model': 'tldr@v2.0.0', 'text': 'This study empirically investigates the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting, leading to multi-epoch degradation, and discovers that leveraging mixture-of-experts (MoE) enables cost-effective and efficient hyper-parameter tuning for computationally intensive dense LLMs with comparable trainable parameters.'}",https://arxiv.org/pdf/2305.13230
-Universal Few-shot Learning of Dense Prediction Tasks with Visual Token Matching,Donggyun Kim,"Dense prediction tasks are a fundamental class of problems in computer vision. As supervised methods suffer from high pixel-wise labeling cost, a few-shot learning solution that can learn any dense task from a few labeled images is desired. Yet, current few-shot learning methods target a restricted set of tasks such as semantic segmentation, presumably due to challenges in designing a general and unified model that is able to flexibly and efficiently adapt to arbitrary tasks of unseen semantics. We propose Visual Token Matching (VTM), a universal few-shot learner for arbitrary dense prediction tasks. It employs non-parametric matching on patch-level embedded tokens of images and labels that encapsulates all tasks. Also, VTM flexibly adapts to any task with a tiny amount of task-specific parameters that modulate the matching algorithm. We implement VTM as a powerful hierarchical encoder-decoder architecture involving ViT backbones where token matching is performed at multiple feature hierarchies. We experiment VTM on a challenging variant of Taskonomy dataset and observe that it robustly few-shot learns various unseen dense prediction tasks. Surprisingly, it is competitive with fully supervised baselines using only 10 labeled examples of novel tasks (0.004% of full supervision) and sometimes outperforms using 0.1% of full supervision. Codes are available at https://github.com/GitGyun/visual_token_matching.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Visual Token Matching (VTM), a universal few-shot learner for arbitrary dense prediction tasks that employs non-parametric matching on patch-level embedded tokens of images and labels that encapsulates all tasks and flexibly adapts to any task with a tiny amount of task-specific parameters that modulate the matching algorithm.'}",http://arxiv.org/pdf/2303.14969
-Efficient Training of BERT by Progressively Stacking,Linyuan Gong,"Unsupervised pre-training is commonly used in natural language processing: a deep neural network trained with proper unsupervised prediction tasks are shown to be effective in many down-stream tasks. Because it is easy to create a large monolingual dataset by collecting data from the Web, we can train high-capacity models. Therefore, training efficiency becomes a critical issue even when using high-performance hardware. In this paper, we explore an efficient training method for the state-of-the-art bidirectional Transformer (BERT) model. By visualizing the self-attention distributions of different layers at different positions in a well-trained BERT model, we find that in most layers, the self-attention distribution will concentrate locally around its position and the start-of-sentence token. Motivated by this, we pro-pose the stacking algorithm to transfer knowledge from a shallow model to a deep model; then we apply stacking progressively to accelerate BERT training. Experiments showed that the models trained by our training strategy achieve similar performance to models trained from scratch, but our algorithm is much faster.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores an efficient training method for the state-of-the-art bidirectional Transformer (BERT) model and pro-poses the stacking algorithm to transfer knowledge from a shallow model to a deep model; then the algorithm is applied progressively to accelerate BERT training.'}",
-Energy efficient cross-layer design protocol by using token passing mechanism for WSN,Sandhyasree Thaskani,"Recent advancements in processor, memory and radio technologies have enabled cheap and small sensors which are capable of sensing, communication and computation. Most of the Wireless Sensor Networks have limitation of battery power and in most of the applications it is not possible to replace the battery of a node. Considering this scarce energy and processing resources of WSN, we try to establish a joint optimization and design of networking layers i.e., cross layer design which could be a promising alternative to the inefficient traditional layered protocol design. In this paper we propose energy efficient cross layer design of the MAC and Routing protocol namely Energy Efficient cross-layer design protocol by using token passing mechanism for WSN. This proposed protocol proves to be better than some of the existing protocols and it is shown with the help of simulations.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes energy efficient cross layer design of the MAC and Routing protocol namely Energy Efficient cross-layer design protocol by using token passing mechanism for WSN and it proves to be better than some of the existing protocols.'}",
-Prompt Tuning for Parameter-efficient Medical Image Segmentation,Marc Fischer,"Neural networks pre-trained on a self-supervision scheme have become the standard when operating in data rich environments with scarce annotations. As such, fine-tuning a model to a downstream task in a parameter-efficient but effective way, e.g. for a new set of classes in the case of semantic segmentation, is of increasing importance. In this work, we propose and investigate several contributions to achieve a parameter-efficient but effective adaptation for semantic segmentation on two medical imaging datasets. Relying on the recently popularized prompt tuning approach, we provide a prompt-able UNet (PUNet) architecture, that is frozen after pre-training, but adaptable throughout the network by class-dependent learnable prompt tokens. We pre-train this architecture with a dedicated dense self-supervision scheme based on assignments to online generated prototypes (contrastive prototype assignment, CPA) of a student teacher combination alongside a concurrent segmentation loss on a subset of classes. We demonstrate that the resulting neural network model is able to attenuate the gap between fully fine-tuned and parameter-efficiently adapted models on CT imaging datasets. As such, the difference between fully fine-tuned and prompt-tuned variants amounts to only 3.83 pp for the TCIA/BTCV dataset and 2.67 pp for the CT-ORG dataset in the mean Dice Similarity Coefficient (DSC, in %) while only prompt tokens, corresponding to 0.85% of the pre-trained backbone model with 6.8M frozen parameters, are adjusted. The code for this work is available on https://github.com/marcdcfischer/PUNet .","{'model': 'tldr@v2.0.0', 'text': 'This work provides a prompt-able UNet (PUNet) architecture, that is frozen after pre-training, but adaptable throughout the network by class-dependent learnable prompt tokens, and demonstrates that the resulting neural network model is able to attenuate the gap between fully fine-tuned and parameter-efficiently adapted models on CT imaging datasets.'}",https://arxiv.org/pdf/2211.09233
-Rethinking Visual Prompt Learning as Masked Visual Token Modeling,Ning Liao,"Prompt learning has achieved great success in efficiently exploiting large-scale pre-trained models in natural language processing (NLP). It reformulates the downstream tasks as the generative pre-training ones, thus narrowing down the gap between them and improving the performance stably. However, when transferring it to the vision area, current visual prompt learning methods are all designed on discriminative pre-trained models, and there is also a lack of careful design to unify the forms of pre-training and downstream tasks. To explore prompt learning on the generative pre-trained visual model as well as keeping the task consistency, we propose Visual Prompt learning as masked visual Token Modeling (VPTM) to transform the downstream visual classification into the pre-trained masked visual token prediction. In addition, we develop the prototypical verbalizer for mapping the predicted visual token with implicit semantics to explicit downstream labels. To our best knowledge, VPTM is the first visual prompt method on the generative pre-trained visual model, and the first to achieve consistency between pre-training and downstream visual classification by task reformulation. Experiments show that VPTM outperforms other visual prompt methods and achieves excellent efficiency. Moreover, the task consistency of VPTM contributes to the robustness against prompt location, prompt length and prototype dimension, and could be deployed uniformly.","{'model': 'tldr@v2.0.0', 'text': 'Visual Prompt learning as masked visual Token Modeling (VPTM) is proposed to transform the downstream visual classification into the pre-trained masked visual token prediction, and the task consistency of VPTM contributes to the robustness against prompt location, prompt length and prototype dimension, and could be deployed uniformly.'}",http://arxiv.org/pdf/2303.04998
-Energy Efficient Token Based MAC Protocol for Wireless Sensor Networks,Soumya Ray,"In this paper, we propose on energy efficient token based MAC protocol for WSNs, in order to reduce energy consumption of each sensor node which is one of the important issues to prolong the network lifetime. WSN consists of a large number of small and low-cost sensor nodes powered by small batteries and equipped with various sensing devices. Due to limited energy in WSN energy efficiency is an important factor in designing a MAC protocol. To derogate energy consumption most of the MAC protocols in WSN exploits low duty-cycle; among those RMAC, HEMAC allows a node to transmit data packet for multi-hop in a single duty-cycle. To reduce energy consumption on prolonged network life time sensor networks are usually duty cycled; each node remains in low power sleep mode most of the time and wakes up periodically to sense for channel activities. In all these above MAC protocols, due to the synchronized scheduling, transmission collisions, flooding will increase resulting in energy waste and low throughput. By allowing nodes to operate with a new token-based approach, we intend to produce energy efficiency in an event based approach by reducing flooding, collision and traffic congestion. The work in this paper draws a bed on implementing a token based distributed approach to save energy at in order to mitigate flooding. Simulation studies of the proposed MAC protocol have been carried out using Castalia simulator.","{'model': 'tldr@v2.0.0', 'text': 'The work in this paper draws a bed on implementing a token based distributed approach to save energy at in order to mitigate flooding and produce energy efficiency in an event based approach by reducing flooding, collision and traffic congestion.'}",
-Fast Passage Re-ranking with Contextualized Exact Term Matching and Efficient Passage Expansion,Shengyao Zhuang,"BERT-based information retrieval models are expensive, in both time (query latency) and computational resources (energy, hardware cost), making many of these models impractical especially under resource constraints. The reliance on a query encoder that only performs tokenization and on the pre-processing of passage representations at indexing, has allowed the recently proposed TILDE method to overcome the high query latency issue typical of BERT-based models. This however is at the expense of a lower effectiveness compared to other BERT-based re-rankers and dense retrievers. In addition, the original TILDE method is characterised by indexes with a very high memory footprint, as it expands each passage into the size of the BERT vocabulary. In this paper, we propose TILDEv2, a new model that stems from the original TILDE but that addresses its limitations. TILDEv2 relies on contextualized exact term matching with expanded passages. This requires to only store in the index the score of tokens that appear in the expanded passages (rather than all the vocabulary), thus producing indexes that are 99% smaller than those of TILDE. This matching mechanism also improves ranking effectiveness by 24%, without adding to the query latency. This makes TILDEv2 the state-of-the-art passage re-ranking method for CPU-only environments, capable of maintaining query latency below 100ms on commodity hardware.","{'model': 'tldr@v2.0.0', 'text': 'TILDEv2 is proposed, a new model that stems from the original TILDE but that addresses its limitations, and relies on contextualized exact term matching with expanded passages to be the state-of-the-art passage re-ranking method for CPU-only environments, capable of maintaining query latency below 100ms on commodity hardware.'}",
-Token-Weighted Crowdsourcing,Gerry Tsoukalas,"Blockchain-based platforms often rely on token-weighted voting (“τ-weighting”) to efficiently crowdsource information from their users for a wide range of applications, including content curation and on-chain governance. We examine the effectiveness of such decentralized platforms for harnessing the wisdom and effort of the crowd. We find that τ-weighting generally discourages truthful voting and erodes the platform’s predictive power unless users are “strategic enough” to unravel the underlying aggregation mechanism. Platform accuracy decreases with the number of truthful users and the dispersion in their token holdings, and in many cases, platforms would be better off with a “flat” 1/n mechanism. When, prior to voting, strategic users can exert effort to endogenously improve their signals, users with more tokens generally exert more effort—a feature often touted in marketing materials as a core advantage of τ-weighting—however, this feature is not attributable to the mechanism itself, and more importantly, the ensuing equilibrium fails to achieve the first-best accuracy of a centralized platform. The optimality gap decreases as the distribution of tokens across users approaches a theoretical optimum, which we derive, but tends to increase with the dispersion in users’ token holdings. This paper was accepted by Gabriel Weintraub, revenue management and market analytics.","{'model': 'tldr@v2.0.0', 'text': 'It is found that τ-weighting generally discourages truthful voting and erodes the platform’s predictive power unless users are “strategic enough” to unravel the underlying aggregation mechanism.'}",https://pubsonline.informs.org/doi/pdf/10.1287/mnsc.2019.3515
-Funding New Ventures with Digital Tokens: Due Diligence and Token Tradability,Yannis Bakos,"We study funding mechanisms for new ventures based on cryptographic tokens enabled by blockchain technology, such as initial coin offerings (ICOs). The infrastructure built around blockchain allows for efficient trading of these tokens. Thus the due diligence process can be crowdsourced beyond the potential early adopters, as favorable assessments of the venture can be leveraged with speculative token purchases exceeding the buyer's individual demand. We develop a model for an entrepreneur considering digital tokens to finance a new venture, focusing on token tradability and broader crowdsourcing of due diligence. We then compare funding via digital tokens with funding from traditional financing sources like venture capital as well as with pre-sale crowdfunding with non-tradable rewards. We find that tradable digital tokens are more attractive when there is higher uncertainty about market demand, which is frequently the case when developing applications of new technologies, such as blockchain-based platforms. In such cases crowdsourcing due diligence benefits from the information contained in the private valuations of the early potential adopters. Token tradability leverages that information and increases the amount that can be financed, thus enabling new ventures with higher development costs. The increased funding comes at the cost of a lower digital token price and lower total profit for the entrepreneur, but may still be preferable to the alternatives, if such alternatives are available at all. This work makes a contribution to the emerging literatures on digital cryptographic tokens and on crowdfunding new ventures.","{'model': 'tldr@v2.0.0', 'text': 'It is found that tradable digital tokens are more attractive when there is higher uncertainty about market demand, which is frequently the case when developing applications of new technologies, such as blockchain-based platforms.'}",
-Sustainable Growth and Token Economy Design: The Case of Steemit,M. Kim,"Cryptocurrency blockchain technology is attracting worldwide attention, and the number of initial coin offerings (ICOs) is increasing rapidly. This new economic trend, called cryptoeconomics, can program human behavior through incentive design. A cryptocurrency-based incentive system is not only transparent, but also allows businesses to substitute initial investment costs with cryptocurrency tokens until they are on a sustainable growth trajectory in terms of network effects. This study aims to propose a process for building a desirable model of a token economy, based on the case of Steemit—a blogging and social networking website that is creating high values due to its efficient token economy model. We suggest the following design process of a token economy model: (1) Determine token-business fit, (2) determine the chance of success, (3) determine the properties of token, (4) give tokens intrinsic value, (5) establish strategies to raise token value, (6) establish operational strategies of token economy system, (7) establish strategies for token liquidation, and (8) continue modifying the operational base. Considering cryptoeconomics is still at an early stage, it is expected that the guidelines on the token economy model suggested in this paper will lay a significant foundation for the development of cryptoeconomics research.","{'model': 'tldr@v2.0.0', 'text': 'This study aims to propose a process for building a desirable model of a token economy, based on the case of Steemit—a blogging and social networking website that is creating high values due to its efficient token economy model.'}",https://www.mdpi.com/2071-1050/11/1/167/pdf?version=1546164358
-End-to-end Contextual Speech Recognition Using Class Language Models and a Token Passing Decoder,Zhehuai Chen,"End-to-end modeling (E2E) of automatic speech recognition (ASR) blends all the components of a traditional speech recognition system into a single, unified model. Although it simplifies the ASR systems, the unified model is hard to adapt when training and testing data mismatches. In this work, we focus on contextual speech recognition, which is particularly challenging for E2E models because contextual information is only available in inference time. To improve the performance in the presence of contextual information during training, we propose to use class-based language models (CLM) that can populate context-dependent information during inference. To enable this approach to scale to a large number of class members and minimize search errors, we propose a token passing algorithm with an efficient token recombination for E2E systems. We evaluate the proposed system on general and contextual ASR tasks, and achieve relative 62% Word Error Rate (WER) reduction for the contextual ASR task without hurting recognition performance for the general ASR task. We also show that the proposed method performs well without modification of the decoding hyper-parameters across tasks, making it a desirable solution for E2E ASR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use class-based language models (CLM) that can populate context-dependent information during inference for contextual speech recognition, and proposes a token passing algorithm with an efficient token recombination for E2E ASR.'}",https://arxiv.org/pdf/1812.02142
-YaRN: Efficient Context Window Extension of Large Language Models,Bowen Peng,"Rotary Position Embeddings (RoPE) have been shown to effectively encode positional information in transformer-based language models. However, these models fail to generalize past the sequence length they were trained on. We present YaRN (Yet another RoPE extensioN method), a compute-efficient method to extend the context window of such models, requiring 10x less tokens and 2.5x less training steps than previous methods. Using YaRN, we show that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow, while also surpassing previous the state-of-the-art at context window extension. In addition, we demonstrate that YaRN exhibits the capability to extrapolate beyond the limited context of a fine-tuning dataset. The models fine-tuned using YaRN has been made available and reproduced online up to 128k context length at https://github.com/jquesnelle/yarn","{'model': 'tldr@v2.0.0', 'text': 'YaRN is presented, a compute-efficient method to extend the context window of RoPE extensioN models, requiring 10x less tokens and 2.5x less training steps than previous methods, and it is shown that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow.'}",https://arxiv.org/pdf/2309.00071
-SoundStorm: Efficient Parallel Audio Generation,Zalán Borsos,"We present SoundStorm, a model for efficient, non-autoregressive audio generation. SoundStorm receives as input the semantic tokens of AudioLM, and relies on bidirectional attention and confidence-based parallel decoding to generate the tokens of a neural audio codec. Compared to the autoregressive generation approach of AudioLM, our model produces audio of the same quality and with higher consistency in voice and acoustic conditions, while being two orders of magnitude faster. SoundStorm generates 30 seconds of audio in 0.5 seconds on a TPU-v4. We demonstrate the ability of our model to scale audio generation to longer sequences by synthesizing high-quality, natural dialogue segments, given a transcript annotated with speaker turns and a short prompt with the speakers' voices.","{'model': 'tldr@v2.0.0', 'text': ""The ability of the model to scale audio generation to longer sequences is demonstrated by synthesizing high-quality, natural dialogue segments, given a transcript annotated with speaker turns and a short prompt with the speakers' voices.""}",http://arxiv.org/pdf/2305.09636
-DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning,Zhengxiang Shi,"Prompt tuning (PT), where a small amount of trainable soft (continuous) prompt vectors is affixed to the input of language models (LM), has shown promising results across various tasks and models for parameter-efficient fine-tuning (PEFT). PT stands out from other PEFT approaches because it maintains competitive performance with fewer trainable parameters and does not drastically scale up its parameters as the model size expands. However, PT introduces additional soft prompt tokens, leading to longer input sequences, which significantly impacts training and inference time and memory usage due to the Transformer's quadratic complexity. Particularly concerning for Large Language Models (LLMs) that face heavy daily querying. To address this issue, we propose Decomposed Prompt Tuning (DePT), which decomposes the soft prompt into a shorter soft prompt and a pair of low-rank matrices that are then optimised with two different learning rates. This allows DePT to achieve better performance while saving over 20% memory and time costs compared to vanilla PT and its variants, without changing trainable parameter sizes. Through extensive experiments on 23 natural language processing (NLP) and vision-language (VL) tasks, we demonstrate that DePT outperforms state-of-the-art PEFT approaches, including the full fine-tuning baseline in some scenarios. Additionally, we empirically show that DEPT grows more efficient as the model size increases. Our further study reveals that DePT integrates seamlessly with parameter-efficient transfer learning in the few-shot learning setting and highlights its adaptability to various model architectures and sizes.","{'model': 'tldr@v2.0.0', 'text': 'Decomposed Prompt Tuning (DePT) is proposed, which decomposes the soft prompt into a shorter soft prompt and a pair of low-rank matrices that are then optimised with two different learning rates, allowing DePT to achieve better performance while saving over 20% memory and time costs compared to vanilla PT and its variants.'}",https://arxiv.org/pdf/2309.05173
-Socially shared regulation of learning in CSCL: understanding and prompting individual- and group-level shared regulatory activities,Sanna Järvelä,,"{'model': 'tldr@v2.0.0', 'text': 'How conceptual work in socially shared regulation of learning (SSRL) contributes to effective and efficient CSCL, what tools are presently available, and what the implications of research on these tools are for future tool development are discussed.'}",
-STEM: Secure Token Exchange Mechanisms,Maneesh Darisi,,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents a solution which is token standard agnostic and provides effective intrinsic smart contracts facilitating token exchange and thus reducing the counterparty risk.'}",
-Walk Proximal Gradient: An Energy-Efficient Algorithm for Consensus Optimization,Xianghui Mao,"Decentralized computing is widely used for multiagent systems since it works without a central computing node. In this paper, we develop a first-order algorithm for decentralized consensus optimization that is more energy efficient than the current state-of-the-art. Our algorithm is suitable for application scenarios such as networks of wireless sensors and Internet of Things, where some agents have limited (battery) energy. We call our algorithm walk proximal gradient (WPG), which passes a token through a walk (a succession of nodes) in the graph. The agents that are visited during the walk compute the gradients of their private functions and update the token. We analyze WPG where the walk is the repetition of a Hamiltonian cycle and show that the token converges to the consensual solution faster (in terms of energy consumption) than existing gradient-based decentralized methods. We also generalize the analysis to the non-Hamiltonian graphs. Numerical experiments are presented to validate the energy efficiency of our algorithm.","{'model': 'tldr@v2.0.0', 'text': 'A first-order algorithm for decentralized consensus optimization that is more energy efficient than the current state-of-the-art and suitable for application scenarios such as networks of wireless sensors and Internet of Things, where some agents have limited energy.'}",
-Point-PEFT: Parameter-Efficient Fine-Tuning for 3D Pre-trained Models,Ivan Tang,"The popularity of pre-trained large models has revolutionized downstream tasks across diverse fields, such as language, vision, and multi-modality. To minimize the adaption cost for downstream tasks, many Parameter-Efficient Fine-Tuning (PEFT) techniques are proposed for language and 2D image pre-trained models. However, the specialized PEFT method for 3D pre-trained models is still under-explored. To this end, we introduce Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters. Specifically, for a pre-trained 3D model, we freeze most of its parameters, and only tune the newly added PEFT modules on downstream tasks, which consist of a Point-prior Prompt and a Geometry-aware Adapter. The Point-prior Prompt adopts a set of learnable prompt tokens, for which we propose to construct a memory bank with domain-specific knowledge, and utilize a parameter-free attention to enhance the prompt tokens. The Geometry-aware Adapter aims to aggregate point cloud features within spatial neighborhoods to capture fine-grained geometric information through local interactions. Extensive experiments indicate that our Point-PEFT can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters, demonstrating the efficiency and effectiveness of our approach. Code will be released at https://github.com/EvenJoker/Point-PEFT.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters that can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters.'}",https://arxiv.org/pdf/2310.03059
-Supporting Beacon and Event-Driven Messages in Vehicular Platoons through Token-Based Strategies,A. Balador,"Timely and reliable inter-vehicle communications is a critical requirement to support traffic safety applications, such as vehicle platooning. Furthermore, low-delay communications allow the platoon to react quickly to unexpected events. In this scope, having a predictable and highly effective medium access control (MAC) method is of utmost importance. However, the currently available IEEE 802.11p technology is unable to adequately address these challenges. In this paper, we propose a MAC method especially adapted to platoons, able to transmit beacons within the required time constraints, but with a higher reliability level than IEEE 802.11p, while concurrently enabling efficient dissemination of event-driven messages. The protocol circulates the token within the platoon not in a round-robin fashion, but based on beacon data age, i.e., the time that has passed since the previous collection of status information, thereby automatically offering repeated beacon transmission opportunities for increased reliability. In addition, we propose three different methods for supporting event-driven messages co-existing with beacons. Analysis and simulation results in single and multi-hop scenarios showed that, by providing non-competitive channel access and frequent retransmission opportunities, our protocol can offer beacon delivery within one beacon generation interval while fulfilling the requirements on low-delay dissemination of event-driven messages for traffic safety applications.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a MAC method especially adapted to platoons, able to transmit beacons within the required time constraints, but with a higher reliability level than IEEE 802.11p, while concurrently enabling efficient dissemination of event-driven messages.'}",https://www.mdpi.com/1424-8220/18/4/955/pdf?version=1521820815
-GTFuzz: Guard Token Directed Grey-Box Fuzzing,Rundong Li,"Directed grey-box fuzzing is an effective technique to find bugs in programs with the guidance of user-specified target locations. However, it can hardly reach a target location guarded by certain syntax tokens (Guard Tokens for short), which is often seen in programs with string operations or grammar/lexical parsing. Only the test inputs containing Guard Tokens are likely to reach the target locations, which challenges the effectiveness of mutation-based fuzzers. In this paper, a Guard Token directed grey-box fuzzer called GTFuzz is presented, which extracts Guard Tokens according to the target locations first and then exploits them to direct the fuzzing. Specifically, to ensure the new test cases generated from mutations contain Guard Tokens, new strategies of seed prioritization, dictionary generation, and seed mutation are also proposed, so as to make them likely to reach the target locations. Experiments on real-world software show that GTFuzz can reach the target locations, reproduce crashes, and expose bugs more efficiently than the state-of-the-art grey-box fuzzers (i.e., AFL, AFLGO and FairFuzz). Moreover, GTFuzz identified 23 previously undiscovered bugs in LibXML2 and MJS.","{'model': 'tldr@v2.0.0', 'text': 'A Guard Token directed grey-box fuzzer called GTFuzz is presented, which extracts Guard Tokens according to the target locations first and then exploits them to direct the fuzzing, and identified 23 previously undiscovered bugs in LibXML2 and MJS.'}",
-Hash Embeddings for Efficient Word Representations,Dan Svenstrup,"We present hash embeddings, an efficient method for representing words in a continuous vector form. A hash embedding may be seen as an interpolation between a standard word embedding and a word embedding created using a random hash function (the hashing trick). In hash embeddings each token is represented by $k$ $d$-dimensional embeddings vectors and one $k$ dimensional weight vector. The final $d$ dimensional representation of the token is the product of the two. Rather than fitting the embedding vectors for each token these are selected by the hashing trick from a shared pool of $B$ embedding vectors. Our experiments show that hash embeddings can easily deal with huge vocabularies consisting of millions of tokens. When using a hash embedding there is no need to create a dictionary before training nor to perform any kind of vocabulary pruning after training. We show that models trained using hash embeddings exhibit at least the same level of performance as models trained using regular embeddings across a wide range of tasks. Furthermore, the number of parameters needed by such an embedding is only a fraction of what is required by a regular embedding. Since standard embeddings and embeddings constructed using the hashing trick are actually just special cases of a hash embedding, hash embeddings can be considered an extension and improvement over the existing regular embedding types.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that hash embeddings can easily deal with huge vocabularies consisting of millions of tokens and that models trained using hash embedDings exhibit at least the same level of performance as models training using regular embeddins across a wide range of tasks.'}",
-Efficient Automatic Punctuation Restoration Using Bidirectional Transformers with Robust Inference,Maury Courtland,"Though people rarely speak in complete sentences, punctuation confers many benefits to the readers of transcribed speech. Unfortunately, most ASR systems do not produce punctuated output. To address this, we propose a solution for automatic punctuation that is both cost efficient and easy to train. Our solution benefits from the recent trend in fine-tuning transformer-based language models. We also modify the typical framing of this task by predicting punctuation for sequences rather than individual tokens, which makes for more efficient training and inference. Finally, we find that aggregating predictions across multiple context windows improves accuracy even further. Our best model achieves a new state of the art on benchmark data (TED Talks) with a combined F1 of 83.9, representing a 48.7% relative improvement (15.3 absolute) over the previous state of the art.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a solution for automatic punctuation that is both cost efficient and easy to train, and modify the typical framing of this task by predicting punctuation for sequences rather than individual tokens, which makes for more efficient training and inference.'}",https://www.aclweb.org/anthology/2020.iwslt-1.33.pdf
-sRDMA - Efficient NIC-based Authentication and Encryption for Remote Direct Memory Access,Konstantin Taranov,"State-of-the-art remote direct memory access (RDMA) technologies have shown to be vulnerable against attacks by innetwork adversaries, as they provide only a weak form of protection by including access tokens in each message. A network eavesdropper can easily obtain sensitive information and modify bypassing packets, affecting not only secrecy but also integrity. Tampering with packets can have drastic consequences. For example, when memory pages with code are changed remotely, altering packet contents enables remote code injection. We propose sRDMA, a protocol that provides efficient authentication and encryption for RDMA to prevent information leakage and message tampering. sRDMA uses symmetric cryptography and employs network interface cards to perform cryptographic operations. Additionally, we provide an implementation for sRDMA using programmable network adapters.","{'model': 'tldr@v2.0.0', 'text': 'SRDMA is proposed, a protocol that provides efficient authentication and encryption for RDMA to prevent information leakage and message tampering and an implementation for sRDMA using programmable network adapters.'}",
-2ED: An Efficient Entity Extraction Algorithm Using Two-Level Edit-Distance,Zeyi Wen,"Entity extraction is fundamental to many text mining tasks such as organisation name recognition. A popular approach to entity extraction is based on string matching against a dictionary of known entities. For approximate entity extraction from free text, considering solely character-based or solely token-based similarity cannot simultaneously deal with minor name variations at token-level and typos at character-level. Moreover, the tolerance of mismatch in character-level may be different from that in token-level, and the tolerance thresholds of the two levels should be able to be customised individually. In this paper, we propose an efficient character-level and token-level edit-distance based algorithm called FuzzyED. To improve the efficiency of FuzzyED, we develop various novel techniques including (i) a spanning-based candidate sub-string producing technique, (ii) a lower bound dissimilarity to determine the boundaries of candidate sub-strings, (iii) a core token based technique that makes use of the importance of tokens to reduce the number of unpromising candidate sub-strings, and (iv) a shrinking technique to reuse computation. Empirical results on real world datasets show that FuzzyED can efficiently extract entities and produce a high F1 score in the range of [0.91, 0.97].","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an efficient character-level and token-level edit-distance based algorithm called FuzzyED, and develops various novel techniques including a spanning-based candidate sub-string producing technique, a lower bound dissimilarity to determine the boundaries of Candidate sub-strings, and a shrinking technique to reuse computation.'}",
-Parameter-Efficient Tuning Helps Language Model Alignment,Tianci Xue,"Aligning large language models (LLMs) with human preferences is essential for safe and useful LLMs. Previous works mainly adopt reinforcement learning (RLHF) and direct preference optimization (DPO) with human feedback for alignment. Nevertheless, they have certain drawbacks. One such limitation is that they can only align models with one preference at the training time (e.g., they cannot learn to generate concise responses when the preference data prefers detailed responses), or have certain constraints for the data format (e.g., DPO only supports pairwise preference data). To this end, prior works incorporate controllable generations for alignment to make language models learn multiple preferences and provide outputs with different preferences during inference if asked. Controllable generation also offers more flexibility with regard to data format (e.g., it supports pointwise preference data). Specifically, it uses different control tokens for different preferences during training and inference, making LLMs behave differently when required. Current controllable generation methods either use a special token or hand-crafted prompts as control tokens, and optimize them together with LLMs. As control tokens are typically much lighter than LLMs, this optimization strategy may not effectively optimize control tokens. To this end, we first use parameter-efficient tuning (e.g., prompting tuning and low-rank adaptation) to optimize control tokens and then fine-tune models for controllable generations, similar to prior works. Our approach, alignMEnt with parameter-Efficient Tuning (MEET), improves the quality of control tokens, thus improving controllable generation quality consistently by an apparent margin on two well-recognized datasets compared with prior works.","{'model': 'tldr@v2.0.0', 'text': 'The approach, alignMEnt with parameter-Efficient Tuning (MEET), improves the quality of control tokens, thus improving controllable generation quality consistently by an apparent margin on two well-recognized datasets compared with prior works.'}",https://arxiv.org/pdf/2310.00819
-SPEGTI: Structured Prediction for Efficient Generative Text-to-Image Models,Sadeep Jayasumana,"Modern text-to-image generation models produce high-quality images that are both photorealistic and faithful to the text prompts. However, this quality comes at significant computational cost: nearly all of these models are iterative and require running inference multiple times with large models. This iterative process is needed to ensure that different regions of the image are not only aligned with the text prompt, but also compatible with each other. In this work, we propose a light-weight approach to achieving this compatibility between different regions of an image, using a Markov Random Field (MRF) model. This method is shown to work in conjunction with the recently proposed Muse model. The MRF encodes the compatibility among image tokens at different spatial locations and enables us to significantly reduce the required number of Muse prediction steps. Inference with the MRF is significantly cheaper, and its parameters can be quickly learned through back-propagation by modeling MRF inference as a differentiable neural-network layer. Our full model, SPEGTI, uses this proposed MRF model to speed up Muse by 1.5X with no loss in output image quality.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a light-weight approach to achieving compatibility between different regions of an image, using a Markov Random Field (MRF) model, and uses this proposed MRF model to speed up Muse by 1.5X with no loss in output image quality.'}",https://arxiv.org/pdf/2308.10997
-Efficient token based clone detection with flexible tokenization,H. Basit,"Code clones are similar code fragments that occur at multiple locations in a software system. Detection of code clones provides useful information for maintenance, reengineering, program understanding and reuse. Several techniques have been proposed to detect code clones. These techniques differ in the code representation used for analysis of clones, ranging from plain text to parse trees and program dependence graphs. Clone detection based on lexical tokens involves minimal code transformation and gives good results, but is computationally expensive because of the large number of tokens that need to be compared. We explored string algorithms to find suitable data structures and algorithms for efficient token based clone detection and implemented them in our tool Repeated Tokens Finder (RTF). Instead of using suffix tree for string matching, we use more memory efficient suffix array. RTF incorporates a suffix array based linear time algorithm to detect string matches. It also provides a simple and customizable tokenization mechanism. Initial analysis and experiments show that our clone detection is simple, scalable, and performs better than the previous well-known tools.","{'model': 'tldr@v2.0.0', 'text': 'String algorithms are explored to find suitable data structures and algorithms for efficient token based clone detection and implemented them in the tool Repeated Tokens Finder (RTF), which incorporates a suffix array based linear time algorithm to detect string matches.'}",
-A New Coeff-Token Decoding Method With Efficient Memory Access in H.264/AVC Video Coding Standard,Y. Moon,"In general, a large amount of memory accesses are required for context-based adaptive variable-length coding decoding in H.264/AVC standard. It is a serious problem for applications such as videophone and digital multimedia broadcasting services because of the considerable power consumption of the memory access function. In order to reduce the memory access, we carefully examined the codewords in variable-length code tables (VLCTs) of the coeff-token and found some distinctive features from the structure of codewords. Based on the features, a new coeff-token decoding method with an efficient memory organization is proposed in this paper. The simulation results show that the proposed algorithm achieves an approximately 95% memory access saving without video-quality degradation, compared with conventional decoding","{'model': 'tldr@v2.0.0', 'text': 'A new coeff-token decoding method with an efficient memory organization is proposed and it is shown that the proposed algorithm achieves an approximately 95% memory access saving without video-quality degradation, compared with conventional decoding.'}",
-Local Self-Attention over Long Text for Efficient Document Retrieval,Sebastian Hofstätter,"Neural networks, particularly Transformer-based architectures, have achieved significant performance improvements on several retrieval benchmarks. When the items being retrieved are documents, the time and memory cost of employing Transformers over a full sequence of document terms can be prohibitive. A popular strategy involves considering only the first n terms of the document. This can, however, result in a biased system that under retrieves longer documents. In this work, we propose a local self-attention which considers a moving window over the document terms and for each term attends only to other terms in the same window. This local attention incurs a fraction of the compute and memory cost of attention over the whole document. The windowed approach also leads to more compact packing of padded documents in minibatches resulting in additional savings. We also employ a learned saturation function and a two-staged pooling strategy to identify relevant regions of the document. The Transformer-Kernel pooling model with these changes can efficiently elicit relevance information from documents with thousands of tokens. We benchmark our proposed modifications on the document ranking task from the TREC 2019 Deep Learning track and observe significant improvements in retrieval quality as well as increased retrieval of longer documents at moderate increase in compute and memory costs.","{'model': 'tldr@v2.0.0', 'text': 'A local self-attention which considers a moving window over the document terms and for each term attends only to other terms in the same window resulting in increased retrieval of longer documents at moderate increase in compute and memory costs is proposed.'}",https://arxiv.org/pdf/2005.04908
-The Dynamic Geometry of Interaction Machine: A Token-Guided Graph Rewriter,Koko Muroya,"In implementing evaluation strategies of the lambda-calculus, both correctness and efficiency of implementation are valid concerns. While the notion of correctness is determined by the evaluation strategy, regarding efficiency there is a larger design space that can be explored, in particular the trade-off between space versus time efficiency. Aiming at a unified framework that would enable the study of this trade-off, we introduce an abstract machine, inspired by Girard's Geometry of Interaction (GoI), a machine combining token passing and graph rewriting. We show soundness and completeness of our abstract machine, called the \emph{Dynamic GoI Machine} (DGoIM), with respect to three evaluations: call-by-need, left-to-right call-by-value, and right-to-left call-by-value. Analysing time cost of its execution classifies the machine as ``efficient'' in Accattoli's taxonomy of abstract machines.","{'model': 'tldr@v2.0.0', 'text': ""An abstract machine, inspired by Girard's Geometry of Interaction (GoI), a machine combining token passing and graph rewriting, andalysing time cost of its execution classifies the machine as ``efficient'' in Accattoli's taxonomy of abstract machines.""}",http://arxiv.org/pdf/1803.00427
-ES‐Plag: Efficient and sensitive source code plagiarism detection tool for academic environment,Lisan Sulistiani,"Source code plagiarism detection using Running‐Karp‐Rabin Greedy‐String‐Tiling (RKRGST) is a common practice in academic environment. However, such approach is time‐inefficient (due to RKRGST's cubic time complexity) and insensitive (toward token subsequence rearrangement). This paper proposes ES‐Plag, a plagiarism detection tool featured with cosine‐based filtering and penalty mechanism to handle aforementioned issues. Cosine‐based filtering mitigates time‐inefficiency by excluding non‐potential pairs from RKRGST comparison; while penalty mechanism mitigates insensitivity by reducing the number of matched tokens with the number of matched subsequences prior similarity normalization. In addition to issue‐solving features, ES‐Plag is also featured with project‐based input, colorized adjacency similarity matrix, matched token highlighting, and various similarity algorithms (e.g., Cosine Similarity and Local Alignment). Three findings can be deducted from our evaluation. First, cosine‐based filtering boosts up time efficiency with a trade‐off in effectiveness. Second, penalty mechanism enhances sensitivity even though its improvement in terms of effectiveness is quite limited. Third, ES‐Plag's features are beneficial for examiners.","{'model': 'tldr@v2.0.0', 'text': 'ES‐Plag, a plagiarism detection tool featured with cosine‐based filtering and penalty mechanism to handle aforementioned issues, is proposed and its features are beneficial for examiners.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/cae.22066
-Token Routing: A Power Efficient Method for Securing AODV Routing Protocol,Leiyuan Li,"Despite the abundance of existing secure routing protocols in ad hoc and sensor networks, real world ad hoc and sensor network applications still usually opt for routing protocols without any security consideration, mostly because of the relatively high cost for enforcing security. In order to reduce the energy overhead when enhancing the security of routing protocols, we propose a token routing protocol (TRP), which is a routing security enhancement of the widely accepted AODV routing protocol. TRP employs hash-chain algorithm to generate a token, which is appended to the data packets to identify the authenticity of the routing packets and to choose correct route for data packets. As compared to the well-known AODV security enhancements such as SAODV, TRP provides comparable security with a significant reduction in energy consumption and routing packet delay by using the hash algorithm instead of public key cryptographic algorithms. Therefore TRP provides a practical energy-efficient option for routing security enhancements in AODV routing protocol","{'model': 'tldr@v2.0.0', 'text': 'A token routing protocol (TRP), which is a routing security enhancement of the widely accepted AODV routing protocol, employs hash-chain algorithm to generate a token, which is appended to the data packets to identify the authenticity of the routing packets and to choose correct route for data packets.'}",
-Testing the Tolerance Principle: Children form productive rules when it is more computationally efficient to do so,Kathryn D. Schuler,"During language acquisition, children must learn when to generalize a pattern – applying it broadly and to new words (‘add –ed’ in English) – and when to restrict generalization, storing the pattern only with specific lexical items. One effort to quantify the conditions for generalization, the Tolerance Principle, has been shown to accurately predict children’s generalizations in dozens of corpus-based studies. This principle hypothesizes that a general rule will be formed when it is computationally more efficient than storing lexical forms individually. It is formalized as: a rule R will generalize if the number of exceptions does not exceed the number of words in the category N divided by the natural log of N (N/lnN). Here we test the principle in an artificial language of 9 nonsense nouns. As predicted, children exposed to 5 regular forms and 4 exceptions generalized, applying the regular form to 100% of novel test words. Children exposed to 3 regular forms and 6 exceptions did not extend the rule, even though the token frequency of the regular form was still high in this condition. The Tolerance Principle thus appears to capture a basic principle of generalization in rule formation.","{'model': 'tldr@v2.0.0', 'text': 'The Tolerance Principle appears to capture a basic principle of generalization in rule formation, as children exposed to 5 regular forms and 4 exceptions generalized, applying the regular form to 100% of novel test words.'}",
-SMYRF: Efficient Attention using Asymmetric Clustering,Giannis Daras,"We propose a novel type of balanced clustering algorithm to approximate attention. Attention complexity is reduced from $O(N^2)$ to $O(N \log N)$, where $N$ is the sequence length. Our algorithm, SMYRF, uses Locality Sensitive Hashing (LSH) in a novel way by defining new Asymmetric transformations and an adaptive scheme that produces balanced clusters. The biggest advantage of SMYRF is that it can be used as a drop-in replacement for dense attention layers without any retraining. On the contrary, prior fast attention methods impose constraints (e.g. queries and keys share the same vector representations) and require re-training from scratch. We apply our method to pre-trained state-of-the-art Natural Language Processing and Computer Vision models and we report significant memory and speed benefits. Notably, SMYRF-BERT outperforms (slightly) BERT on GLUE, while using $50\%$ less memory. We also show that SMYRF can be used interchangeably with dense attention before and after training. Finally, we use SMYRF to train GANs with attention in high resolutions. Using a single TPU, we were able to scale attention to 128x128=16k and 256x256=65k tokens on BigGAN on CelebA-HQ.","{'model': 'tldr@v2.0.0', 'text': 'The algorithm, SMYRF, uses Locality Sensitive Hashing (LSH) in a novel way by defining new Asymmetric transformations and an adaptive scheme that produces balanced clusters that can be used interchangeably with dense attention before and after training.'}",
-An efficient and practical local synchronous bandwidth allocation scheme for the timed-token MAC protocol,Sijing Zhang,"This paper is concerned with deadline guarantees of synchronous messages with deadlines equal to periods, in a timed token ring network such as FDDI where the timed token medium access control (MAC) protocol is used. The timed token protocol guarantees a bounded access time and an average bandwidth for synchronous traffic. However, this guarantee alone, though necessary, is insufficient for guaranteeing the transmission of synchronous messages before their deadlines. To ensure timely delivery, the synchronous bandwidth must be carefully allocated to individual nodes. We propose and analyse an efficient and practical local synchronous bandwidth allocation (SBA) scheme. The new scheme performs better than any previously published as it calculates the synchronous bandwidth such that during the message period, the total synchronous transmission time definitely available (when judged only by local information) is exactly equal to the transmission time required. Our scheme also differs significantly from previously reported ones by explicitly taking into account the synchronous bandwidth allocation for message sets whose minimum message deadlines (D/sub min/) are less than twice the target token rotation time (TTRT), and consequently can apply to any synchronous message set (with D/sub min/>TTRT). The feasibility of the allocations produced by the proposed scheme and the worst case achievable utilisation of the scheme are also discussed.","{'model': 'tldr@v2.0.0', 'text': 'An efficient and practical local synchronous bandwidth allocation (SBA) scheme is proposed and analysed that performs better than any previously published as it calculates theynchronous bandwidth such that during the message period, the total synchronous transmission time definitely available is exactly equal to the transmission time required.'}",
-Efficient global allocation of synchronous bandwidths for hard real-time communication with the timed token MAC protocol,Sijing Zhang,"This paper presents an efficient global synchronous bandwidth allocation (SBA) scheme for guaranteeing synchronous message deadlines in a timed token network (such as FDDI) where the timed token medium access control (MAC) protocol is used. A local SBA scheme allocates synchronous bandwidth to a node using only information available locally to that node while a global scheme uses network wide information from all nodes. A local SBA scheme may be preferred for practical use to a global one due to its lower network management overhead. However, a global SBA scheme can greatly outperform a local scheme due to a more appropriate allocation resulted from the more complete global information. That is, use of a global SBA scheme may largely enhance the network ability of guaranteeing synchronous traffic. The global scheme proposed in this paper outperforms all existing SBA schemes for guaranteeing synchronous messages with deadlines equal to periods. Numerical examples are given to demonstrate the superiority of the proposed scheme to previously-proposed SBA schemes.","{'model': 'tldr@v2.0.0', 'text': 'The global scheme proposed in this paper outperforms all existing SBA schemes for guaranteeing synchronous messages with deadlines equal to periods and can greatly outperform a local scheme due to a more appropriate allocation resulted from the more complete global information.'}",
-Highly Space-Efficient Self-Stabilizing Depth-First Token Circulation for Trees,F. Petit,"Self-stabilization was rst introduced by Dijkstra Dij74]: it is the property for a system to eventually recover itself a legitimate state after any perturbation modifying the memory state. Since Dijkstra original paper, the goal of many works has been to obtain self-stabilizing algorithms requiring the fewest possible number of states. A great deal of proposed algorithms are token circulations, very useful to solve distributed mutual exclusion. In every algorithm, the token is held by the processor enabled to make a move. Tchuente Tch81] showed that in such conditions, the expected state number lower bound to solve mutual exclusion over tree networks is 2 n Q n i=1 i , n is the number of processors, i is the neighbors number of each processor p i. In this paper, we use a weaker token formulation introduced by Villain Vil97]: a processor holds a token if it holds a particular state. This new light allows us to propose a self-stabilizing depth-rst token circulation for tree networks requiring fewer states than Tchuente's lower bound, i.e. (1 + 1) Q n i=2 ((i + 2) states only.","{'model': 'tldr@v2.0.0', 'text': ""A weaker token formulation introduced by Villain Vil97 is used, which allows for a self-stabilizing depth-rst token circulation for tree networks requiring fewer states than Tchuente's lower bound, i.e. (1 + 1) Q n i=2 ((i + 2) states only.""}",
-"Space-Efficient, Distributed and Self-Stabilizing Depth-First Token Circulation",C. Johnen,The notion of self stabilization was introduced by Dijkstra He de ned a system as self stabilizing when regardless of its initial state it is guaranteed to arrive at a legitimate state in a nite number of steps Such a property is very desirable for any distributed system because after any unexpected perturbation modifying the memory state the system eventually recovers and returns to a legitimate state without any outside intervention In this paper we are interested in a distributed self stabilizing depth rst token circulation protocol on an uniform rooted network no identi ers but a distinguished root As already noted a search algorithm together with a deterministic enumeration of the node s neighbors yields an algorithm determining a spanning tree Our contribution is improving the best up to now known space complexity for this problem from O log N to O log D where N is number of nodes and D is the network s degree Moreover we give a full proof of the algorithm correctness assuming the existence of a dis tributed demon,"{'model': 'tldr@v2.0.0', 'text': 'A distributed self stabilizing depth rst token circulation protocol on an uniform rooted network no identi ers but a distinguished root is interested and the best up to now known space complexity for this problem is improved.'}",
-Asynchronous Completion Token An Object Behavioral Pattern for Efficient Asynchronous Event Handling,Timothy H. Harrison,"Contemporary applications must respond to many types of events, ranging from user interface notifications to network messages. Delay-sensitive applications, such as network management systems, often perform long-runningoperations asynchronously to avoid blocking the processing of other pending events. When these asynchronous operations complete, applications may need more information than simply the notification itself to properly handle the event. This paper describes the Asynchronous Completion Token pattern, which allows applications to efficiently associate state with the completion of asynchronous operations.","{'model': 'tldr@v2.0.0', 'text': 'The Asynchronous Completion Token pattern is described, which allows applications to efficiently associate state with the completion of asynchronous operations.'}",
-E/sup 2/WTRP: an energy-efficient wireless token ring protocol,Zhenhua Deng,"A wireless token ring protocol (WTRP) built on top of the IEEE 802.11 distributed control function (DCF) provides guaranteed QoS in terms of high throughput and bounded transmission delay. It is also applicable in a small-scale wireless ad hoc network where all nodes can hear each other and where throughput and delay requirements are stringent. As nodes in the network have different traffic styles and the number of nodes ready for transmission is highly dynamic, the aggregate throughput and delay of WTRP is degraded. Furthermore, each node keeps on forwarding the token, even though it is not ready for transmission, thus the power consumed by inactive nodes is high. An energy-efficient wireless token ring protocol is proposed to reduce the power consumption. It is shown that E/sup 2/WTRP could prolong the standby time for inactive nodes two-fold.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that E/sup 2/WTRP could prolong the standby time for inactive nodes two-fold and an energy-efficient wireless token ring protocol is proposed to reduce the power consumption.'}",
-WarpLDA: a Cache Efficient O(1) Algorithm for Latent Dirichlet Allocation,Jianfei Chen,"Developing efficient and scalable algorithms for Latent Dirichlet Allocation (LDA) is of wide interest for many applications. Previous work has developed an O(1) Metropolis-Hastings sampling method for each token. However, the performance is far from being optimal due to random accesses to the parameter matrices and frequent cache misses. In this paper, we first carefully analyze the memory access efficiency of existing algorithms for LDA by the scope of random access, which is the size of the memory region in which random accesses fall, within a short period of time. We then develop WarpLDA, an LDA sampler which achieves both the best O(1) time complexity per token and the best O(K) scope of random access. Our empirical results in a wide range of testing conditions demonstrate that WarpLDA is consistently 5-15x faster than the state-of-the-art Metropolis-Hastings based LightLDA, and is comparable or faster than the sparsity aware F+LDA. With WarpLDA, users can learn up to one million topics from hundreds of millions of documents in a few hours, at an unprecedentedly throughput of 11G tokens per second.","{'model': 'tldr@v2.0.0', 'text': 'WarpLDA is developed, an LDA sampler which achieves both the best O(1) time complexity per token and thebest O(K) scope of random access, within a short period of time.'}",
-An efficient algorithm for solving the token distribution problem on k-ary d-cube networks,Claude G. Diderich,"In parallel programs where the problem data is dynamically generated, it is very useful to be able to rely on an efficient load balancing algorithm. The token distribution problem (TDP) is a generalization of the static load balancing problem. The paper describes a novel algorithm for solving the TDP for k-ary d-cube topology networks. Compared to other algorithms, our method is more general and does not rely on every processor knowing the exact number of tokens associated to each processor. The correctness of the algorithm is proved and its complexity is informally studied.<>","{'model': 'tldr@v2.0.0', 'text': 'A novel algorithm for solving the token distribution problem for k-ary d-cube topology networks is described, which is more general and does not rely on every processor knowing the exact number of tokens associated to each processor.'}",
-1 An Efficient Token Based Algorithm for Mutual Exclusion In Distributed System,F. Kawsar,"Many distributed computations involving the sharing of resources among various processes require that a resource be allocated to a single process at a time. Therefore, mutual exclusion is a fundamental problem in any distributed computing system. This problem must be solved to synchronize the access to shared resources in order to maintain their consistency and integrity. The major goal of this paper is to get the reader acquainted with a new approach towards the ring based technique for mutual exclusion in a distributed system. An algorithm is proposed based on the idea of generating token by the competing processes to enter the critical section and thus eliminating idle time message passing and reducing communication overhead.","{'model': 'tldr@v2.0.0', 'text': 'An algorithm is proposed based on the idea of generating token by the competing processes to enter the critical section and thus eliminating idle time message passing and reducing communication overhead.'}",
-Delving into Multimodal Prompting for Fine-grained Visual Classification,Xin Jiang,"Fine-grained visual classification (FGVC) involves categorizing fine subdivisions within a broader category, which poses challenges due to subtle inter-class discrepancies and large intra-class variations. However, prevailing approaches primarily focus on uni-modal visual concepts. Recent advancements in pre-trained vision-language models have demonstrated remarkable performance in various high-level vision tasks, yet the applicability of such models to FGVC tasks remains uncertain. In this paper, we aim to fully exploit the capabilities of cross-modal description to tackle FGVC tasks and propose a novel multimodal prompting solution, denoted as MP-FGVC, based on the contrastive language-image pertaining (CLIP) model. Our MP-FGVC comprises a multimodal prompts scheme and a multimodal adaptation scheme. The former includes Subcategory-specific Vision Prompt (SsVP) and Discrepancy-aware Text Prompt (DaTP), which explicitly highlights the subcategory-specific discrepancies from the perspectives of both vision and language. The latter aligns the vision and text prompting elements in a common semantic space, facilitating cross-modal collaborative reasoning through a Vision-Language Fusion Module (VLFM) for further improvement on FGVC. Moreover, we tailor a two-stage optimization strategy for MP-FGVC to fully leverage the pre-trained CLIP model and expedite efficient adaptation for FGVC. Extensive experiments conducted on four FGVC datasets demonstrate the effectiveness of our MP-FGVC.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel multimodal prompting solution, denoted as MP-FGVC, based on the contrastive language-image pertaining (CLIP) model, and tailor a two-stage optimization strategy to fully leverage the pre-trained CLIP model and expedite efficient adaptation for FGVC.'}",https://arxiv.org/pdf/2309.08912
-Multimodal Prompting with Missing Modalities for Visual Recognition Supplementary Materials,Yi-Lun Lee,"We show ablation studies for attention-level prompts in Figure 1 and Figure 2, which analyze the effect of prompting layers and prompt length respectively. The results are similar to the study of input-level prompts as shown in Section 4.3 of the main paper. In summary, the earlier prompting layers and more prompting layers improve the performance more. In addition, even with fewer parameters (i.e., reducing the prompt length to 2), the performance is still competitive.",,
-Examining Taiwanese university students’ multimodal multiple text comprehension: individual differences and epistemic prompting,Yuan‐Hsuan Lee,,,
-Scaffolding university students' epistemic cognition during multimodal multiple-document reading: The effects of the epistemic prompting and the automated reflection report,Yuan‐Hsuan Lee,,"{'model': 'tldr@v2.0.0', 'text': ""Results showed that students relied primarily on YouTube videos for their MMDR with higher credibility ratings for the YouTube video than written texts, and epistemic prompting, along with the automatically generated reflection report, enhanced students' scores on constructed-response items for the experimental group and helped them create a coherent mental representation of information from diverse multimodal sources.""}",
-Prompting as Multimodal Fusing,,"Tsimpoukelli et al. (2021) devise Frozen, em- 001 powering a language model to solve multi- 002 modal tasks by pretraining a vision encoder 003 whose outputs are prompts fed to the language 004 model. The vision encoder has a dual objec- 005 tive: Extracting image features and aligning 006 image/text representation spaces. We propose 007 to disentangle the objectives by using prompt 008 vectors to align the spaces; this lets the vision 009 encoder focus on extracting image features. We 010 show that this disentangled approach is modu- 011 lar and parameter-efficient for processing tasks 012 that involve two or more modalities. 013","{'model': 'tldr@v2.0.0', 'text': 'Frozen, Frozen is devise Frozen, powering a language model to solve multi- modal tasks by pretraining a vision encoder whose outputs are prompts fed to the language 004 model.'}",
-DDCoT: Duty-Distinct Chain-of-Thought Prompting for Multimodal Reasoning in Language Models,Ge Zheng,"A long-standing goal of AI systems is to perform complex multimodal reasoning like humans. Recently, large language models (LLMs) have made remarkable strides in such multi-step reasoning on the language modality solely by leveraging the chain of thought (CoT) to mimic human thinking. However, the transfer of these advancements to multimodal contexts introduces heightened challenges, including but not limited to the impractical need for labor-intensive annotation and the limitations in terms of flexibility, generalizability, and explainability. To evoke CoT reasoning in multimodality, this work first conducts an in-depth analysis of these challenges posed by multimodality and presents two key insights:""keeping critical thinking""and""letting everyone do their jobs""in multimodal CoT reasoning. Furthermore, this study proposes a novel DDCoT prompting that maintains a critical attitude through negative-space prompting and incorporates multimodality into reasoning by first dividing the reasoning responsibility of LLMs into reasoning and recognition and then integrating the visual recognition capability of visual models into the joint reasoning process. The rationales generated by DDCoT not only improve the reasoning abilities of both large and small language models in zero-shot prompting and fine-tuning learning, significantly outperforming state-of-the-art methods but also exhibit impressive generalizability and explainability.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a novel DDCoT prompting that maintains a critical attitude through negative-space prompting and incorporates multimodality into reasoning by first dividing the reasoning responsibility of LLMs into reasoning and recognition and then integrating the visual recognition capability of visual models into the joint reasoning process.'}",
-Efficient Multimodal Fusion via Interactive Prompting,Yaowei Li,"Large-scale pre-training has brought unimodal fields such as computer vision and natural language processing to a new era. Following this trend, the size of multimodal learning models constantly increases, leading to an urgent need to reduce the massive computational cost of finetuning these models for downstream tasks. In this paper, we propose an efficient and flexible multimodal fusion method, namely PMF, tailored for fusing unimodally pretrained transformers. Specifically, we first present a modular multimodal fusion framework that exhibits high flexibility and facilitates mutual interactions among different modalities. In addition, we disentangle vanilla prompts into three types in order to learn different optimizing objectives for multimodal learning. It is also worth noting that we propose to add prompt vectors only on the deep layers of the unimodal transformers, thus significantly reducing the training memory usage. Experiment results show that our proposed method achieves comparable performance to several other multimodal finetuning methods with less than 3% trainable parameters and up to 66% saving of training memory usage.","{'model': 'tldr@v2.0.0', 'text': 'An efficient and flexible multimodal fusion method, namely PMF, tailored for fusing unimodally pretrained transformers and achieves comparable performance to several other multimodals finetuning methods with less than 3% trainable parameters and up to 66% saving of training memory usage.'}",https://arxiv.org/pdf/2304.06306
-Prompting ChatGPT in MNER: Enhanced Multimodal Named Entity Recognition with Auxiliary Refined Knowledge,Jinyuan Li,"Multimodal Named Entity Recognition (MNER) on social media aims to enhance textual entity prediction by incorporating image-based clues. Existing studies mainly focus on maximizing the utilization of pertinent image information or incorporating external knowledge from explicit knowledge bases. However, these methods either neglect the necessity of providing the model with external knowledge, or encounter issues of high redundancy in the retrieved knowledge. In this paper, we present PGIM -- a two-stage framework that aims to leverage ChatGPT as an implicit knowledge base and enable it to heuristically generate auxiliary knowledge for more efficient entity prediction. Specifically, PGIM contains a Multimodal Similar Example Awareness module that selects suitable examples from a small number of predefined artificial samples. These examples are then integrated into a formatted prompt template tailored to the MNER and guide ChatGPT to generate auxiliary refined knowledge. Finally, the acquired knowledge is integrated with the original text and fed into a downstream model for further processing. Extensive experiments show that PGIM outperforms state-of-the-art methods on two classic MNER datasets and exhibits a stronger robustness and generalization capability.","{'model': 'tldr@v2.0.0', 'text': 'PGIM is a two-stage framework that aims to leverage ChatGPT as an implicit knowledge base and enable it to heuristically generate auxiliary knowledge for more efficient entity prediction and exhibits a stronger robustness and generalization capability.'}",
-Initial Images: Using Image Prompts to Improve Subject Representation in Multimodal AI Generated Art,Han Qiao,"Advances in text-to-image generative models have made it easier for people to create art by just prompting models with text. However, creating through text leaves users with limited control over the final composition or the way the subject is represented. A potential solution is to use image prompts alongside text prompts to condition the model. To better understand how and when image prompts can improve subject representation in generations, we conduct an annotation experiment to quantify their effect on generations of abstract, concrete plural, and concrete singular subjects. We find that initial images improved subject representation across all subject types, with the most noticeable improvement in concrete singular subjects. In an analysis of different types of initial images, we find that icons and photos produced high quality generations of different aesthetics. We conclude with design guidelines for how initial images can improve subject representation in AI art.",,
-Multimodal Multi-Task Stealth Assessment for Reflection-Enriched Game-Based Learning,Anisha Gupta,"Game-based learning environments enable effective and engaging learning experiences that can be dynamically tailored to students. There is growing interest in the role of reflection in supporting student learning in game-based learning environments. By prompting students to periodically stop and reflect on their learning processes, it is possible to gain insight into students’ perceptions of their knowledge and problem-solving progress, which can in turn inform adaptive scaffolding to improve student learning outcomes. Given the positive relationship between student reflection and learning, we investigate the benefits of jointly modeling post-test score and reflection depth using a multimodal, multitask stealth assessment framework. Specifically, we present a gated recurrent unit-based multi-task stealth assessment framework that takes as input multimodal data streams (e.g., game trace logs, pre-test data, natural language responses to in-game reflection prompts) to jointly predict post-test scores and written reflection depth scores. Evaluation results demonstrate that the multimodal multi-task model outperforms single-task neural models that utilize subsets of the modalities, as well as non-neural baselines such as random forest regressors. Our multi-task stealth assessment framework for measuring students’ content knowledge and reflection depth during game-based learning shows significant promise for supporting student learning and improved reflection.","{'model': 'tldr@v2.0.0', 'text': 'A gated recurrent unit-based multi-task stealth assessment framework that takes as input multimodal data streams to jointly predict post-test scores and written reflection depth scores shows significant promise for supporting student learning and improved reflection.'}",
-Visual Chain of Thought: Bridging Logical Gaps with Multimodal Infillings,Daniel Philip Rose,"Recent advances in large language models elicit reasoning in a chain of thought that allows models to decompose problems in a human-like fashion. Though this paradigm improves multi-step reasoning ability in language models, it is limited by being unimodal and applied mainly to question-answering tasks. We claim that incorporating visual augmentation into reasoning is essential, especially for complex, imaginative tasks. Consequently, we introduce VCoT, a novel method that leverages chain of thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data. Our method uses visual guidance to generate synthetic multimodal infillings that add consistent and novel information to reduce the logical gaps for downstream tasks that can benefit from temporal reasoning, as well as provide interpretability into models' multi-step reasoning. We apply VCoT to the Visual Storytelling and WikiHow summarization datasets and demonstrate through human evaluation that VCoT offers novel and consistent synthetic data augmentation beating chain of thought baselines, which can be used to enhance downstream performance.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces VCoT, a novel method that leverages chain of thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data and applies it to the Visual Storytelling and WikiHow summarization datasets.'}",http://arxiv.org/pdf/2305.02317
-Pulmonary Artery Intimal Sarcoma: A Diagnostic Challenge Using a Multimodal Approach,Rima Nakrani,"Pulmonary artery intimal sarcoma (PAIS) is a rare tumor without clear syndromic presentation other than nonspecific symptoms of cough, dyspnea, and weight loss. This diagnosis is difficult due to challenging radiographic interpretations of multiple imaging modalities. We present a case of a 60-year-old male, who presented to his pulmonologist and underwent a CT chest with IV contrast that initially suggested primary lung carcinoma. CT angiogram showed significant vascular filling defects suspicious of an intravascular mass, rather than vascular invasion by lung lesions. The PET/CT scans further suggested a malignant process, but indistinguishable between an extravascular or intravascular etiology. Taking these results together, they suggested an intravascular malignancy, prompting a tissue biopsy, which ultimately led to a diagnosis of PAIS with metastases. Establishing a definitive diagnosis is essential as treatment and prognosis are different for sarcoma compared to carcinoma. There is no standard treatment to date, and management often includes a multidisciplinary approach involving surgery, radiation, chemotherapy, and targeted therapy. PAIS is a rare entity that cannot be diagnosed clinically and needs a multimodality approach for its diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'A case of a 60-year-old male, who presented to his pulmonologist and underwent a CT chest with IV contrast that initially suggested primary lung carcinoma, which ultimately led to a diagnosis of PAIS with metastases, which needs a multimodality approach for its diagnosis.'}",https://downloads.hindawi.com/journals/crim/2020/6795272.pdf
-Using multimodal learning analytics to study collaboration on discussion groups,Fabián Riquelme,,"{'model': 'tldr@v2.0.0', 'text': 'The main aim of this work is to develop a computational environment to both analyze and visualize student discussion groups working in a collaborative way to accomplish a task and show that this information can help to support complex decision-making processes.'}",
-Prompting Technology and Persons With Dementia: The Significance of Context and Communication,Rachel Braley,"Background and Objectives Smart home auto-prompting has the potential to increase the functional independence of persons with dementia (PWDs) and decrease caregiver burden as instrumental activities of daily living (IADLs) are completed at home. To improve prompting technologies, we sought to inductively understand how PWDs responded to auto-prompting while performing IADL tasks. Research Design and Methods Fifteen PWDs completed eight IADLs in a smart home testbed and received a hierarchy of verbal auto-prompts (indirect, direct, multimodal) as needed for task completion. Two researchers viewed archived videos and recorded the observed behaviors of the PWDs and their reflections watching the PWDs. Using qualitative descriptive methods, an interdisciplinary analytic team reviewed transcripts and organized data into themes using content analysis. Results Context and Communication emerged as the major themes, suggesting that positive user experiences will require auto-prompting systems to account for a multitude of contextual factors (individual and environmental) such as level of cognitive impairment, previous exposure to task, and familiarity of environment. Communicating with another human rather than an automated prompting system may be important if individuals begin to exhibit signs of stress while completing activities. Discussion and Implications Additional work is needed to create auto-prompting systems that provide specific, personalized, and flexible prompts. Holistic conceptualization of ""successful task completion"" is needed and a positive end-user experience will be key to utility. Such systems will benefit from including positive reinforcement, training, and exploration of how, and whether, direct human involvement can be minimized during the provision of in-home care.","{'model': 'tldr@v2.0.0', 'text': 'Context and Communication emerged as the major themes, suggesting that positive user experiences will require auto-prompting systems to account for a multitude of contextual factors such as level of cognitive impairment, previous exposure to task, and familiarity of environment.'}",https://academic.oup.com/gerontologist/article-pdf/59/1/101/27456552/gny071.pdf
-Exploring Collaborative Writing of User Stories With Multimodal Learning Analytics: A Case Study on a Software Engineering Course,Rene Noel,"Software engineering is the application of principles used in engineering design, development, testing, deployment, and management of software systems. One of the software engineering’s approaches, highly used in new industries, is agile development. User stories are a commonly used notation to capture user requirements in agile development. Nevertheless, for the elaboration of user stories, a high level of collaboration with the client is necessary. This professional skill is rarely measured or evaluated in educational contexts. The present work approaches collaboration in software engineering students through multimodal learning analytics, modeling, and evaluating students’ collaboration while they are writing user stories. For that, we used multidirectional microphones in order to derive social network analysis metrics related to collaboration (permanence and prompting) together with human-annotated information (quality of the stories and productivity). Results show that groups with a lower productivity in writing user stories and less professional experience in managing software requirements present a non-collaborative behavior more frequently, and that teams with a fewer number of interventions are more likely to produce a greater number of user stories. Moreover, although low experience subjects produced more user stories, a greater productivity of the most experienced subjects was not statistically verified. We believe that these types of initiatives will allow the measurement and early development of such skills in university students.","{'model': 'tldr@v2.0.0', 'text': 'Results show that groups with a lower productivity in writing user stories and less professional experience in managing software requirements present a non-collaborative behavior more frequently, and that teams with a fewer number of interventions are more likely to produce a greater number of user stories.'}",
-Multimodal Pipeline : A generic approach for handling multimodal data for supporting learning,D. D. Mitri,"In this demo paper, we introduce the Multimodal Pipeline, a prototypical approach for the collection, storing, annotation, processing and exploitation of multimodal data for supporting learning. At the current stage of development, the Multimodal Pipeline consists of two relevant prototypes: 1) Multimodal Learning Hub for the collection and storing of sensor data from multiple applications and 2) the Visual Inspection Tool for visualisation and annotation of the recorded sessions. The Multimodal Pipeline is designed to be a flexible system useful for supporting psychomotor skills in a variety of learning scenarios such as presentation skills, medical simulation with patient manikins or calligraphy learning. The Multimodal Pipeline can be configured to serve different support strategies, including detecting mistakes and prompting live feedback in an intelligent tutoring system or stimulating self-reflection through a learning analytics dashboard.","{'model': 'tldr@v2.0.0', 'text': 'The Multimodal Pipeline is designed to be a flexible system useful for supporting psychomotor skills in a variety of learning scenarios such as presentation skills, medical simulation with patient manikins or calligraphy learning.'}",
-Multimodal simile: The “when” meme in social media discourse,Adrian Lou,"This paper analyzes the “when” meme, a popular internet meme, which prototypically juxtaposes a when clause with an ostensibly unrelated image. Despite the initial incongruity, I contend this image prompts selective mapping between verbal and visual elements to produce a multimodal simile. First, I attempt to define and more clearly distinguish simile from metaphor. Second, I show how this multimodal simile exhibits unique viewpoint mapping by prompting audiences to subsume viewpoints that are both unfamiliar and bizarre. Third, I connect the like construction in simile with the like reported speech marker to show how both concepts are intimately related. Ultimately, the paper seeks to contribute to studies of simile by bolstering its ties with multimodality, blending, metonymy, viewpoint, and embodiment.",,
-Multimodal treatment approach in a patient with multiple intracranial myxomatous aneurysms.,D. Penn,"The most common primary cardiac tumor is myxoma, typically originating in the left atrium. Emboli to the central nervous system can cause cerebral infarction or, rarely, seed tumor growth within vessel walls, causing myxomatous aneurysms. Fewer than 60 myxomatous aneurysms have been reported, including 2 cases in children. Here, the authors describe 2 different growing myxomatous aneurysms in a child successfully managed using a combined multidisciplinary approach. A 12-year-old boy developed a sudden headache, diplopia, gait instability, and speech difficulty. Magnetic resonance imaging revealed a left parietal hemorrhage and multifocal cerebral infarction, suspicious for an embolic etiology. A cardiac myxoma was identified in the left atrium and resected. Follow-up cranial vasculature imaging demonstrated multiple intracranial myxomatous aneurysms. These lesions were followed up, and serial imaging identified marked growth of 2 of them (right occipital and left parietal), prompting invasive intervention. The deep occipital lesion was better suited to endovascular treatment, while the superficial parietal lesion was amenable to resection. The patient underwent embolization of an enlarging fusiform aneurysm of the distal right posterior cerebral artery, followed by a left parietal craniotomy for a lesion of the distal left middle cerebral artery. Both procedures were performed without complications and achieved successful obliteration of the lesions, as confirmed by catheter angiography at the 30-month follow-up. To the authors' knowledge, this report illustrates the first combined endovascular and open surgical treatment of 2 myxomatous aneurysms in a single patient. While acknowledging the rarity of this condition, this report illustrates the clinical manifestations and treatment challenges posed by myxoma and details a successful strategy that could be employed in similar scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This report illustrates the first combined endovascular and open surgical treatment of 2 myxomatous aneurysms in a child successfully managed using a combined multidisciplinary approach and details a successful strategy that could be employed in similar scenarios.'}",
-Changing the Antibiotic Prescribing of general practice registrars: the ChAP study protocol for a prospective controlled study of a multimodal educational intervention,M. V. van Driel,,"{'model': 'tldr@v2.0.0', 'text': 'It is hypothesised that intervening while early-career GPs are still developing their practice patterns and prescribing habits will result in better adherence to evidence-based guidelines as manifested by lower antibiotic prescribing rates for URTIs and acute bronchitis.'}",https://bmcfampract.biomedcentral.com/track/pdf/10.1186/s12875-016-0470-7
-On the Importance of the Multimodal Approach to Discourse Markers: A Pragmatic View,Kazuki Hata,"Discourse markers (DMs) are pragmatic devices, which operate beyond the traditional word or phrase classification and have little to no effect on the propositional meaning. They have significant functions with regard to organising ongoing discourse by linking discourse segments, illustrating the current discourse structure to the interlocutors. Recent studies have discovered that DMs are truly multifunctional and thus play communicative roles at different dimensions simultaneously, ranging from managing discourse by denoting the speaker’s attitude, monitoring turn-taking activities to prompting the shared-knowledge between interlocutors. Nevertheless, few attentions have been paid on the fact that human communication is multimodal, wherein discourse includes both spoken language and gestures. Most DM studies predominantly investigate the use of DMs in text-based frameworks and therefore do not do justice to the non-linguistic DM functions, which are not easily amenable to text descriptions. The goal of this paper is to draw attention to this gap from the pragmatic perspective on DMs, demonstrating the importance of multimodal approaches to the study area.",,
-Discordant Embryonic Aneuploidy Testing and Prenatal Ultrasonography Prompting Androgen Insensitivity Syndrome Diagnosis,J. Franasiak,"BACKGROUND: Multimodal prenatal screening for developmental pathology is increasingly common. In this case, definitive prenatal diagnosis of androgen insensitivity syndrome was diagnosed after discordant results from karyotypes determined by embryonic preimplantation genetic screening and antenatal ultrasound results. CASE: A 38-year-old woman, gravida 2 para 0010, undergoing in vitro fertilization with preimplantation genetic screening transferred one male and one female embryo. An anatomic ultrasonogram revealed two fetuses with female genitalia. Cell-free fetal DNA analyzed using noninvasive prenatal screening demonstrated Y chromosome material, and amniocentesis confirmed one 46,XX and one 46,XY fetus. Sequencing of the androgen receptor for the 46,XY fetus identified a mutation. CONCLUSION: With increased use, discordance among prenatal testing modalities such as preimplantation genetic screening, noninvasive prenatal screening, and ultrasonography will become more common requiring expert navigation to identify true pathology.","{'model': 'tldr@v2.0.0', 'text': 'With increased use, discordance among prenatal testing modalities such as preimplantation genetic screening, noninvasive prenatal screening, and ultrasonography will become more common requiring expert navigation to identify true pathology.'}",
-The Use of Multimodal Technologies to Enhance Reflective Writing in Teacher Education,L. Adie,,,
-Refractory postsurgical pyoderma gangrenosum in a patient with Beckwith Wiedemann syndrome: response to multimodal therapy,Faiza Fakhar,"Pyoderma gangrenosum (PG) is a rare neutrophilic dermatosis that may be difficult to diagnose and treat. We presented a 41-year-old woman who required skin grafting following third-degree burns to her left breast. She suffered recurrent graft dehiscence and infections over many years, prompting elective bilateral reduction mammoplasty. She subsequently developed suture margin ulcerations unresponsive to topical therapies and antibiotics. Skin biopsies were non-specific, and a clinical diagnosis of PG was established. Although initially responsive to corticosteroids, wounds promptly recurred following steroid taper. She was treated unsuccessfully with various immunomodulatory agents and underwent elective bilateral mastectomy. Following a mastectomy, she developed progressive deep chest wall ulcerations. She failed numerous immunomodulatory treatments, surgical wound closure and negative pressure wound therapy. Ultimately, treatment with adalimumab, mycophenolate mofetil and prednisone, in addition to hyperbaric oxygen therapy facilitated progressive healing. Our case highlights the role of collaborative multimodal therapy for the treatment of refractory PG.","{'model': 'tldr@v2.0.0', 'text': 'Treatment with adalimumab, mycophenolate mofetil and prednisone, in addition to hyperbaric oxygen therapy facilitated progressive healing and highlights the role of collaborative multimodal therapy for the treatment of refractory PG.'}",https://europepmc.org/articles/pmc3822173?pdf=render
-Multi-step pericellular proteolysis controls the transition from individual to collective cancer cell invasion,Katarina Wolf,,"{'model': 'tldr@v2.0.0', 'text': 'Both ECM track widening and transition to multicellular invasion are dependent on MT1-MMP-mediated collagenolysis, shown by broad-spectrum protease inhibition and RNA interference, and invasive migration and proteolytic ECM remodelling are interdependent processes that control tissue micropatterning and macrop atterning.'}",
-SVIT: Scaling up Visual Instruction Tuning,Bo Zhao,"Thanks to the emerging of foundation models, the large language and vision models are integrated to acquire the multimodal ability of visual captioning, dialogue, question answering, etc. Although existing multimodal models present impressive performance of visual understanding and reasoning, their limits are still largely under-explored due to the scarcity of high-quality instruction tuning data. To push the limits of multimodal capability, we Sale up Visual Instruction Tuning (SVIT) by constructing a dataset of 4.2 million visual instruction tuning data including 1.6M conversation question-answer (QA) pairs, 1.6M complex reasoning QA pairs, 1.0M referring QA pairs and 106K detailed image descriptions. Besides the volume, the proposed dataset is also featured by the high quality and rich diversity, which is generated by prompting GPT-4 with the abundant manual annotations of images. We empirically verify that training multimodal models on SVIT can significantly improve the multimodal performance in terms of visual perception, reasoning and planing.","{'model': 'tldr@v2.0.0', 'text': 'It is empirically verified that training multimodal models on SVIT can significantly improve the multi-modal performance in terms of visual perception, reasoning and planing.'}",https://arxiv.org/pdf/2307.04087
-User Behavior Recognition for an Automatic Prompting System - A Structured Approach based on Task Analysis,Christian Peters,"In this paper, we describe a structured approach for user behavior recognition in an automatic prompting system that assists users with cognitive disabilities in the task of brushing their teeth. We analyze the brushing task using qualitative data analysis. The results are a hierarchical decomposition of the task and the identification of environmental configurations during subtasks. We develop a hierarchical recognition framework based on the results of task analysis: We extract a set of features from multimodal sensors which are discretized into the environmental configuration in terms of states of objects involved in the brushing task. We classify subtasks using a Bayesian Network (BN) classifier and a Bayesian Filtering approach. We compare three variants of the BN using different observation models (IU, NaiveBayes and Holistic) with a maximum-margin classifier (multi-class SVM). We present recognition results on 18 trials with regular users and found the BN with a NaiveBayes observation model to produce the best recognition rates of 84.5% on avg.","{'model': 'tldr@v2.0.0', 'text': 'A structured approach for user behavior recognition in an automatic prompting system that assists users with cognitive disabilities in the task of brushing their teeth and develops a hierarchical recognition framework based on the results of task analysis.'}",
-"The cognitive hearing science perspective on perceiving, understanding, and remembering language: The ELU model",J. Rönnberg,"The review gives an introductory description of the successive development of data patterns based on comparisons between hearing-impaired and normal hearing participants’ speech understanding skills, later prompting the formulation of the Ease of Language Understanding (ELU) model. The model builds on the interaction between an input buffer (RAMBPHO, Rapid Automatic Multimodal Binding of PHOnology) and three memory systems: working memory (WM), semantic long-term memory (SLTM), and episodic long-term memory (ELTM). RAMBPHO input may either match or mismatch multimodal SLTM representations. Given a match, lexical access is accomplished rapidly and implicitly within approximately 100–400 ms. Given a mismatch, the prediction is that WM is engaged explicitly to repair the meaning of the input – in interaction with SLTM and ELTM – taking seconds rather than milliseconds. The multimodal and multilevel nature of representations held in WM and LTM are at the center of the review, being integral parts of the prediction and postdiction components of language understanding. Finally, some hypotheses based on a selective use-disuse of memory systems mechanism are described in relation to mild cognitive impairment and dementia. Alternative speech perception and WM models are evaluated, and recent developments and generalisations, ELU model tests, and boundaries are discussed.",,https://www.frontiersin.org/articles/10.3389/fpsyg.2022.967260/pdf
-"Cyberbullying Conceptualization, Characterization and Detection in Social Media – A Systematic Literature Review",Wai Hong Woo,"Social media has become the primary form of communication wherein users can share intimate moments online through photos, videos, or posts. At a glance, while this greatly improves interconnectivity between people, it also increases the propensity towards unrestricted acts of Cyberbullying, prompting the need for a data-centric detection system. Unfortunately, these sites generate much metadata, which begs the need for complex Machine Learning (ML) classifiers to categorize these acts accurately. Prior studies on the subject matter only target the topics of Conceptualization, Characterization, and Classification of Cyberbullying individually, so this research aims to provide a more holistic understanding of the subject matter in a continuous, synthesized format. This study found that Cyberbullying differs from Traditional Bullying in key areas of Repetition and Intention. Moreover, multimodal feature sets, as opposed to single feature sets, significantly improve ML classifiers' performance. Lastly, the selection of appropriate ML classifiers and performance metrics is context-dependent. The result of this study presents a consolidated view of relevant parties tackling different aspects of an ML-based automated Cyberbullying detection system so that those assigned tasks can approach them strategically","{'model': 'tldr@v2.0.0', 'text': ""This study found that Cyberbullying differs from Traditional Bullying in key areas of Repetition and Intention, and multimodal feature sets, as opposed to single feature Sets, significantly improve ML classifiers' performance.""}",https://journals.iium.edu.my/kict/index.php/IJPCC/article/download/374/223
-Mapping genomic and epigenomic evolution in cancer ecosystems,T. Ushijima,"Description Cancer is a major cause of global mortality underpinned by genomic and epigenomic derangements. Here, we highlight the importance of multimodal data integration in understanding the molecular evolution of malignant cell states across the cancer life cycle. The widespread presence of driver mutations and epigenetic alterations in normal-appearing tissues is prompting a reassessment of how cancer initiation is defined. In later-stage cancers, studying the roles of clonal selection, epigenomic adaptation, and persister cells in metastasis and therapy resistance is an emerging field. Finally, the importance of tumor ecosystems in driving cancer development is being unraveled by single-cell and spatial technologies at unprecedented resolution. Improving cancer risk assessment and accelerating therapeutic discovery for patients will require robust, comprehensive, and integrated temporal, spatial, and multilevel tumor atlases across the cancer life cycle.","{'model': 'tldr@v2.0.0', 'text': 'Improving cancer risk assessment and accelerating therapeutic discovery for patients will require robust, comprehensive, and integrated temporal, spatial, and multilevel tumor atlases across the cancer life cycle.'}",
-"UPGPT: Universal Diffusion Model for Person Image Generation, Editing and Pose Transfer",Soon Yau Cheong,"Text-to-image models (T2I) such as StableDiffusion have been used to generate high quality images of people. However, due to the random nature of the generation process, the person has a different appearance e.g. pose, face, and clothing, despite using the same text prompt. The appearance inconsistency makes T2I unsuitable for pose transfer. We address this by proposing a multimodal diffusion model that accepts text, pose, and visual prompting. Our model is the first unified method to perform all person image tasks - generation, pose transfer, and mask-less edit. We also pioneer using small dimensional 3D body model parameters directly to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining the person's appearance.","{'model': 'tldr@v2.0.0', 'text': ""This model is the first unified method to perform all person image tasks - generation, pose transfer, and mask-less edit and pioneer using small dimensional 3D body model parameters directly to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining the person's appearance.""}",https://arxiv.org/pdf/2304.08870
-Oncolytic adenovirus and gene therapy with EphA2-BiTE for the treatment of pediatric high-grade gliomas,C. Arnone,"Background Pediatric high-grade gliomas (pHGGs) are among the most common and incurable malignant neoplasms of childhood. Despite aggressive, multimodal treatment, the outcome of children with high-grade gliomas has not significantly improved over the past decades, prompting the development of innovative approaches. Methods To develop an effective treatment, we aimed at improving the suboptimal antitumor efficacy of oncolytic adenoviruses (OAs) by testing the combination with a gene-therapy approach using a bispecific T-cell engager (BiTE) directed towards the erythropoietin-producing human hepatocellular carcinoma A2 receptor (EphA2), conveyed by a replication-incompetent adenoviral vector (EphA2 adenovirus (EAd)). The combinatorial approach was tested in vitro, in vivo and thoroughly characterized at a molecular level. Results After confirming the relevance of EphA2 as target in pHGGs, documenting a significant correlation with worse clinical outcome of the patients, we showed that the proposed strategy provides significant EphA2-BiTE amplification and enhanced tumor cell apoptosis, on coculture with T cells. Moreover, T-cell activation through an agonistic anti-CD28 antibody further increased the activation/proliferation profiles and functional response against infected tumor cells, inducing eradication of highly resistant, primary pHGG cells. The gene-expression analysis of tumor cells and T cells, after coculture, revealed the importance of both EphA2-BiTE and costimulation in the proposed system. These in vitro observations translated into significant tumor control in vivo, in both subcutaneous and a more challenging orthotopic model. Conclusions The combination of OA and EphA2-BiTE gene therapy strongly enhances the antitumor activity of OA, inducing the eradication of highly resistant tumor cells, thus supporting the clinical translation of the approach.","{'model': 'tldr@v2.0.0', 'text': 'The combination of OA and EphA2-BiTE gene therapy strongly enhances the antitumor activity of OAs, inducing the eradication of highly resistant tumor cells, thus supporting the clinical translation of the approach.'}",https://jitc.bmj.com/content/jitc/9/5/e001930.full.pdf
-Unlocking Multimedia Capabilities of Gigantic Pretrained Language Models,Boyang Li,"Benefitting from unprecedented computational power, massive data throughput, and superhuman memory, large language models (LLMs) are fundamentally transforming multimodal machine learning. An LLM can be analogized to an enormous treasure box guarded by a lock. It contains extensive knowledge, but it can be non-trivial to access and apply appropriate knowledge to solve the problem at hand. Researchers have developed many techniques to unlock the capabilities of LLMs. Some well-known examples include chain-of-thought prompting, ""let's think step by step'', and instruction tuning. In this talk, I will discuss techniques to unlock the capability of LLMs to process both visual and linguistic information. VisualGPT is one of the earliest works that finetunes an LLM for a vision-language task. InstructBLIP is an instruction-tuned large vision-language model, which set new states of the art on several vision-language tasks and snatched top positions on several comprehensive evaluation suites. In addition, I will talk about how to unlock zero-shot capabilities without end-to-end finetuning, or any form of finetuning at all. In Plug-and-Play VQA and Img2LLM, we achieve excellent results on visual question-answering datasets by connecting existing pretrained models using natural language and model interpretations, demonstrating a feasible alternative to the mainstream finetuning approach. Finally, I will describe a new multimodal dataset, Synopses of Movie Narratives, or SyMoN, for story understanding, which constitutes a new challenge for large vision-language models. I will argue that story understanding is an important objective in the pursuit of artificial general intelligence (AGI) because stories are a preeminent form of human communication and story understanding requires many AGI capabilities such as cause-effect reasoning and theory of mind. Compared to other multimodal story datasets, the special advantages of SyMoN include (1) event descriptions at the right level of granularity, (2) abundant mental state descriptions, (3) the use of diverse storytelling techniques, and (4) the provision of easy-to-use automatic performance evaluation.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that story understanding is an important objective in the pursuit of artificial general intelligence (AGI) because stories are a preeminent form of human communication and story understanding requires many AGI capabilities such as cause-effect reasoning and theory of mind.'}",
-Point-of-critical-care diagnostics for sepsis enabled by multiplexed micro and nanosensing technologies.,B. Ashley,"Sepsis is responsible for the highest economic and mortality burden in critical care settings around the world, prompting the World Health Organization in 2018 to designate it as a global health priority. Despite its high universal prevalence and mortality rate, a disproportionately low amount of sponsored research funding is directed toward diagnosis and treatment of sepsis, when early treatment has been shown to significantly improve survival. Additionally, current technologies and methods are inadequate to provide an accurate and timely diagnosis of septic patients in multiple clinical environments. For improved patient outcomes, a comprehensive immunological evaluation is critical which is comprised of both traditional testing and quantifying recently proposed biomarkers for sepsis. There is an urgent need to develop novel point-of-care, low-cost systems which can accurately stratify patients. These point-of-critical-care sensors should adopt a multiplexed approach utilizing multimodal sensing for heterogenous biomarker detection. For effective multiplexing, the sensors must satisfy criteria including rapid sample to result delivery, low sample volumes for clinical sample sparring, and reduced costs per test. A compendium of currently developed multiplexed micro and nano (M/N)-based diagnostic technologies for potential applications toward sepsis are presented. We have also explored the various biomarkers targeted for sepsis including immune cell morphology changes, circulating proteins, small molecules, and presence of infectious pathogens. An overview of different M/N detection mechanisms are also provided, along with recent advances in related nanotechnologies which have shown improved patient outcomes and perspectives on what future successful technologies may encompass. This article is categorized under: Diagnostic Tools > Biosensing.","{'model': 'tldr@v2.0.0', 'text': 'An urgent need to develop novel point-of-care, low-cost systems which can accurately stratify patients and explore the various biomarkers targeted for sepsis including immune cell morphology changes, circulating proteins, small molecules, and presence of infectious pathogens.'}",https://rss.onlinelibrary.wiley.com/doi/am-pdf/10.1002/wnan.1701
-Winder: Linking Speech and Visual Objects to Support Communication in Asynchronous Collaboration,Tae Soo Kim,"Team members commonly collaborate on visual documents remotely and asynchronously. Particularly, students are frequently restricted to this setting as they often do not share work schedules or physical workspaces. As communication in this setting has delays and limits the main modality to text, members exert more effort to reference document objects and understand others’ intentions. We propose Winder, a Figma plugin that addresses these challenges through linked tapes—multimodal comments of clicks and voice. Bidirectional links between the clicked-on objects and voice recordings facilitate understanding tapes: selecting objects retrieves relevant recordings, and playing recordings highlights related objects. By periodically prompting users to produce tapes, Winder preemptively obtains information to satisfy potential communication needs. Through a five-day study with eight teams of three, we evaluated the system’s impact on teams asynchronously designing graphical user interfaces. Our findings revealed that producing linked tapes could be as lightweight as face-to-face (F2F) interactions while transmitting intentions more precisely than text. Furthermore, with preempted tapes, teammates coordinated tasks and invited members to build on each others’ work.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Winder, a Figma plugin that addresses challenges through linked tapes—multimodal comments of clicks and voice that could be as lightweight as face-to-face (F2F) interactions while transmitting intentions more precisely than text.'}",
-Chiral Benzothiazole Monofluoroborate Featuring Chiroptical and Oxygen-Sensitizing Properties: Synthesis and Photophysical Studies.,Omar Sadek,"Advances in personalized medicine are prompting the development of multimodal agents, that is, molecules that combine properties promoting various diagnostic and therapeutic applications. General approaches exploit chemical conjugation of therapeutic agents with contrast agents or the design of multimodal nanoplatforms. Herein, we report the design of a single molecule that exhibits potential for different diagnostic modes as well as the ability to sensitize oxygen, thus offering potential for photodynamic therapy. Exceptionally, this work involves the synthesis and chiral resolution of an enantiomeric pair of chiral monofluoroborates that contain a stereogenic boron atom. Combining experimental and theoretical chiroptical studies allowed the unambiguous determination of their absolute configuration. Photophysical investigations established the ability of this compound to sensitize oxygen even in the absence of heavy atoms within its structure. The synthesis of a chiral benzothiazole monofluoroborate paves a way to multimodal diagnostic tools (fluorescence and nuclear imaging) while also featuring potential therapeutic applications owing to its ability to activate oxygen to its singlet state for use in photodynamic therapy.","{'model': 'tldr@v2.0.0', 'text': 'The synthesis of a chiral benzothiazole monofluoroborate paves a way to multimodal diagnostic tools (fluorescence and nuclear imaging) while also featuring potential therapeutic applications owing to its ability to activate oxygen to its singlet state for use in photodynamic therapy.'}",https://hal.archives-ouvertes.fr/hal-03333839/file/Sadek%20et%20al%20-%202021%20-%20Chiral%20Benzothiazole%20Monofluoroborate%20Featuring%20Chiroptical%20-%20V2_revised_manuscript2_2nd_revision.pdf
-Advances in the Understanding and Management of Chronic Pain in Multiple Sclerosis: a Comprehensive Review,Ivan Urits,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive update of the current perspective of MS pathophysiology, symptomatology, and treatment is provided and medicinal cannabis therapy, targeted physical therapy, and neuromodulation offer promising results are provided.'}",
-The Influence of Multimodal 3D Visualizations on Learning Acquisition,Phuong T. Do,,"{'model': 'tldr@v2.0.0', 'text': 'The process of simultaneously exposing learners to interactive dynamic visualizations and prompting them to attend to information through the pragmatic use of audio cues reduced memory load, and in turn facilitated memory recall.'}",
-Prospects and challenges of imaging neuroinflammation beyond TSPO in Alzheimer’s disease,D. Boche,,"{'model': 'tldr@v2.0.0', 'text': ""The present review describes the recent advancements in microglia and astrocyte biology in the context of health, ageing, and Alzheimer's disease, the most common dementia worldwide, and presents novel methodologies currently explored to improve in vivo knowledge of the neuroinflammatory patterns in Alzheimer’s disease.""}",https://link.springer.com/content/pdf/10.1007/s00259-019-04462-w.pdf
-Making a Case for Spatial Prompting in Human-Robot Communication,A. Green,"In this paper we present an analysis of a set of examples of how verbal and non-verbal behavior of a service robot influence users’ way of positioning themselves during interaction, using concepts from theories of non-verbal behavior. Based on the analysis we propose a design case where a robot utilizes a (naïve) spatial prompting strategy to influence the spatial positioning and communicative behavior of the user. INTRODUCTION A design requirement of a personal service robot is that it should be configured and provided with work tasks by the user in an interactive and intuitive way. These robots are intended to provide service tasks in the home, possibly offering wide range of services. Typically they are envisioned to be equipped with multimodal spoken dialogue systems, to reduce the complexity in the user interface. In this paper we argue that theories of spatial positioning need to be considered when developing the communicative system of the robot. Furthermore we present an empirical account of the way spatial behavior of robots influence human users. We also propose the term spatial prompting, which refers to active strategies of the robot that are intended to influence users to position themselves in a way that is advantageous for further communicative actions. Positioning, as it has been approached as a research challenge for human-robot interaction, is considered as providing adaptive physical movements of the robot. A result of this is that the communicative dimension of positioning typically has been ignored in systems that interactively position themselves in relation to their users. One requirement that is typically put forward is that the robot should position itself in a socially appropriate manner [1, 4]. The parameters that concern these approaches are typically derived from research on nonverbal behavior. In robotics the problem of maintaining the robot localized and situated within a geometric representation of the world has been framed as the Simultaneous Localization and Mapping (SLAM) problem [13]. Recent advances in Human-Robot Interaction (HRI) have raised the interest in detecting and tracking the position of users during interaction. When the position of the user is known, the robot can plan how to position itself [1]. The research on spatial reasoning applied to robotics is well advanced but primarily focused on natural language understanding of spatial relations, providing for exchanges concerning locations of objects in the environment [3]. RESEARCH ON SPATIALITY IN COMMUNICATION There are several research approaches for human-human that are relevant for spatial management between humans and robots. Hall [14] studied interpersonal distances and distinguished four different distances: intimate (0-1.5 ft), personal (1.5-4 ft), social 4-12 ft, and public (> 12 ft). These distances vary both with respect to the current activity and cultural factors. Another dimension that is relevant to spatiality is the concept of territoriality, according to Sack, i.e., “the attempt by an individual or group to affect, influence, or control people, phenomena, and relationships, by delimiting and asserting control over a geographic area” [15]. Kendon [12] also studied the spatial configuration of the participants, using the term F-formations, for instance the L-shape which describes the relation when two participants have a common visual focus. The shared space, the so called o-space, or the transactional space is then located in front of the participants, and it is within this area that the interaction is conducted. Clark [5] refers to this space as the workspace, where perceptual co-presence is established between speakers [5, 10]. In this context, research on perception and especially visual perception plays an important part for maintaining common ground between participants [10, 8]. Gill [9] has investigated the communicative effects that participants achieve by using nonverbal behavior, focusing on the functional rather than the morphological perspective of nonverbal behavior. One such function is the category focus which is a metadiscursive function that signals a shift in the center of attention in the discussion, e.g., a shift in body posture with the same meaning as the utterance “I am going to focus on this spot”. Another, less obvious, but nevertheless important concept is Schegloff’s notion of body torque [15], a state of the bodily configuration when two different body segments are oriented in different directions. According to Schegloff [15] Body torque “project change”, i.e., when some part of the body is organized in an unstable way, the participants may predict that a change in posture is pending. For instance, when turning the head, this might predict a change of the general body orientation. During interaction, speakers monitor the action of others, interpreting purposeful actions that lead towards a common joint goal as compliance [10]. Human-robot interaction is situated in a physical context, where understanding and reference to actions of the human partner during interaction explicitly needs be taken into account. This makes research on virtual collaborative [6] environments interesting also in this context, since it is concerned with models that explicitly represent spatiality and reference. CORPUS ANALYSIS OF SPATIAL MANAGEMENT We have analyzed a video corpus, collected in a European project [7], containing transcribed data of about 20 user sessions, (approximately 20 minutes each) where a user talks to a robot and teaches it the names and locations of objects using a combination of gestures and speech. By viewing the video corpus we identified and analyzed instances where the robot movements or verbal actions appear to influence the actions of the user. The examples reflect three different ways in which the robot actively influences the user to act: • Primary verbal: by using a spoken command • Primary non-verbal: by movements • Multi-modal: using movement as trigger for a verbally specified (or grounded action) 00:32:889-00:34:071 R6: Robot is following 00:40:603-00:41:744 R7: You are too fast 00:44:690-00:46:539 R8: Please stand in front of the camera 00:49:783-00:50:883 R9: Robot is following User","{'model': 'tldr@v2.0.0', 'text': 'An analysis of a set of examples of how verbal and non-verbal behavior of a service robot influence users’ way of positioning themselves during interaction, using concepts from theories of non- verbal behavior is presented.'}",
-An Iterative Data Collection Approach for Multimodal Dialogue Systems,S. Rapp,"This paper deals with the way in which data for multimodal dialogue systems are collected. We argue that for multimodal data, an iterative data collection strategy should be followed. Instead of a single major data collection effort using a “Wizard of OZ” (WOZ) or “prompting” experimental setup, several smaller data collections should accompany the system development. We also describe the “script” experimental setup we developed. It is in between the WOZ and prompting setup, and can be used as a cost effective design for the first data collection within the iterative data collection strategy.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that for multimodal data, an iterative data collection strategy should be followed and instead of a single major data collection effort using a “Wizard of OZ” (WOZ) or “prompting” experimental setup, several smaller data collections should accompany the system development.'}",
-Towards the concept of disease-modifier in post-stroke or vascular cognitive impairment: a consensus report,R. Bordet,,"{'model': 'tldr@v2.0.0', 'text': 'A consensus emerged to propose the development of a disease-modifying strategy in VCI and PSCI based on pleiotropic pharmacological and non-pharmacological approaches.'}",https://bmcmedicine.biomedcentral.com/track/pdf/10.1186/s12916-017-0869-6
-The SmartWeb Corpora: Multimodal Access to the Web in Natural Environments,F. Schiel,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter describes the prompting scheme SitPro, the recording technique used to obtain real situation data as well as the properties of the resulting corpora.'}",
-Guiding the selection of child spellchecker suggestions using audio and visual cues,Brody Downs,"Spellchecking functionality embedded in existing search tools can assist children by offering a list of spelling alternatives when a spelling error is detected. Unfortunately, children tend to generally select the first alternative when presented with a list of options, as opposed to the one that matches their intent. In this paper, we describe a study we conducted with 191 children ages 6-12 in order to offer empirical evidence of: (1) their selection habits when identifying spelling suggestions that match the word they meant to type, and (2) the degree of influence multimodal cues, i.e., synthesized speech and images, have in prompting children to select the correct spelling suggestion. The results from our study reveal that multimodal cues, primarily synthesized speech, have a positive impact on the children's ability to identify their intended word from a list of spelling suggestions.","{'model': 'tldr@v2.0.0', 'text': ""The results from the study reveal that multimodal cues, primarily synthesized speech, have a positive impact on the children's ability to identify their intended word from a list of spelling suggestions.""}",
-Treatment Patterns and Outcomes of Women with Breast Cancer and Supraclavicular Nodal Metastases,N. Tamirisa,,"{'model': 'tldr@v2.0.0', 'text': 'Women with cN3c breast cancer who received multimodal therapy demonstrated improved overall survival when compared with patients undergoing non-standard therapy and those with metastatic (M1) disease.'}",
-A Recalibration of Theatre’s Hypermediality,M. Crossley,,,https://link.springer.com/content/pdf/10.1007%2F978-3-030-49679-1_2.pdf
-Coming to Know More Through and From Writing,V. Prain,"Over the past 20 years, claims about how and why student writing can serve learning have changed markedly. This has been partly due to new technologies displacing writing as a predominant resource for learning, prompting new sense-making practices and shifts in how these changes are theorized. Learners now routinely collaborate to generate, manipulate, analyze, and share images in many subject areas, where multimodal and multimedia resources are expected to motivate learners, enact new learning processes and outcomes, and display this cross-modal learning. These new practices have prompted revisions to how writing is understood and used as a tool for learning in an increasingly multimodal, highly digitized world. In reviewing this literature, we claim that there are strong evidence-based reasons for viewing writing as a central but not sole resource for learning. Our case draws on both past and current research on writing as an epistemological tool. In presenting this case, we draw primarily on our professional background in science education research, acknowledging its distinctive take on the use of writing for learning. However, we think our general case also holds for other disciplinary areas.",,
-Adaptive Action Prompting: A Complementary Aid to Support Task-Oriented Interaction in Explorative User Interfaces,Thomas Kuehme,"Adaptive action prompting supports users by suggesting how to continue with the interaction. A small number of continuously updated action prompts are offered in addition to the regular menu interface of an application. Users can use these prompts either occasionally in trouble situations or for a sequence of system-guided action selections. The suggestions are based on models of the application, context, and user. According prompting strategies are automatically evaluated and can optionally be controlled by the user. Multimodal user interfaces provide further perspectives for adaptive prompting.","{'model': 'tldr@v2.0.0', 'text': 'A small number of continuously updated action prompts are offered in addition to the regular menu interface of an application, offering suggestions based on models of the application, context, and user.'}",
-Acute Posterior Multifocal Placoid Pigment Epitheliopathy Complicated by Fatal Cerebral Vasculitis.,R. Maamari,"A 21-year-old man experienced unilateral vision loss associated with multiple atrophic chorioretinal lesions. He was treated for a presumptive diagnosis of acute retinal necrosis, but his vision did not improve with antiviral therapy. Over the course of several weeks, his symptoms progressed to involve both eyes. The fellow eye showed characteristic yellow-white placoid lesions, prompting treatment with oral corticosteroids for acute posterior multifocal placoid pigment epitheliopathy (APMPPE). Despite high-dose therapy with prednisone 80 mg daily, the patient developed the acute onset of mental status changes within the next several days. Neuroimaging revealed multifocal large-vessel strokes associated with cerebral edema; these infarcts led to herniation and death. Postmortem histopathologic examination confirmed granulomatous inflammation in both ocular and cerebral vasculatures. Together with findings from multimodal imaging obtained throughout this patient's clinical course, our findings support the notion that granulomatous choroiditis is the mechanism of the ocular lesions seen in APMPPE. This granulomatous inflammation can also affect cerebral vessels, leading to strokes.","{'model': 'tldr@v2.0.0', 'text': 'Findings support the notion that granulomatous choroiditis is the mechanism of the ocular lesions seen in APMPPE, which can also affect cerebral vessels, leading to strokes.'}",
-ADAPTIVE PROMPTING,T. Kühme,"Adaptive prompting addresses problems with locating, referencing, and selecting interface items such as elements of directory hierarchies or menu and dialog box entries. By drawing the user’s attention to the most appropriate and most likely to be chosen items, adaptive prompting tries to increase users’ performance on selections and to provide guidance in explorative environments. Examples for adaptive prompting include an Adaptive Tool Prompter, an Adaptive Action Prompter, and Adaptive Prompting in Dialog Boxes. In order to deal with application contexts and evolving needs and preferences of the user, adaptive prompting employs models of the application and the user. The chosen approach allows for an optional user involvement into the adaptation and for an evaluation of the embedded prompting strategies. Multimodal user interfaces provide further perspectives for adaptive prompting.","{'model': 'tldr@v2.0.0', 'text': 'Adaptive prompting addresses problems with locating, referencing, and selecting interface items such as elements of directory hierarchies or menu and dialog box entries to increase users’ performance on selections and to provide guidance in explorative environments.'}",
-"Energizing Project-Based Inquiry: Middle-Grade Students Read, Write, and Create Videos.",Hiller A. Spires,"In light of emerging technologies prompting new avenues for teaching and learning, students are positioned to “create” to learn, with video production being an important process for literacy development. There is a growing need for innovative instructional practices in reading and writing that are aligned with student interests and the activities they engage in outside of the classroom. Connecting video production to school reading and writing experiences taps into a student's natural predisposition for media consumption and production. This article presents a project-based inquiry process that merges students’ growing interest in grassroots video with educational goals that are aligned with state and national curricular standards. Through a five-phase process—ask a compelling question; gather and analyze information; creatively synthesize information; critically evaluate and revise; and publish, share, and act—students are engaged in reading, writing, and creating a video product that captures their multimodal learning of academic content.",,
-Using subpopulation EAs to map molecular structure landscapes,Ahmed Bin Zaman,"The emerging view in molecular biology is that molecules are intrinsically dynamic systems rearranging themselves into different structures to interact with molecules in the cell. Such rearrangements take place on energy landscapes that are vast and multimodal, with minima housing alternative structures. The multiplicity of biologically-active structures is prompting researchers to expand their treatment of classic computational biology problems, such as the template-free protein structure prediction problem (PSP), beyond the quest for the global optimum. In this paper, we revisit subpopulation-oriented EAs as vehicles to switch the objective from classic optimization to landscape mapping. Specifically, we present two EAs, one of which makes use of subpopulation competition to allocate more computational resources to fitter subpopulations, and another of which additionally utilizes a niche preservation technique to maintain stable and diverse subpopulations. Initial assessment on benchmark optimization problems confirms that stabler subpopulations are achieved by the niche-preserving EA. Evaluation on unknown energy landscapes in the context of PSP demonstrates superior mapping performance by both algorithms over a popular Monte Carlo-based method, with the niche-preserving EA achieving superior exploration of lower-energy regions. These results suggest that subpopulation EAs hold much promise for solving important mapping problems in computational structural biology.","{'model': 'tldr@v2.0.0', 'text': 'This paper revisits subpopulation-oriented EAs as vehicles to switch the objective from classic optimization to landscape mapping and presents two EAs, one of which makes use of subpopulation competition to allocate more computational resources to fitter subpopulations, and another of which additionally utilizes a niche preservation technique to maintain stable and diverse subpopulation.'}",https://dl.acm.org/doi/pdf/10.1145/3321707.3321777
-Effectiveness of different recruitment strategies in an RCT of a surgical device: experience from the Endobarrier trial,A. Ruban,"Recruiting participants into clinical trials is notoriously difficult and poses the greatest challenge when planning any investigative study. Poor recruitment may not only have financial ramifications owing to increased time and resources being spent but could adversely influence the clinical impact of a study if it becomes underpowered. Herein, we present our own experience of recruiting into a nationally funded, multicentre, randomised controlled trial (RCT) of the Endobarrier versus standard medical therapy in obese patients with type 2diabetes. Despite these both being highly prevalent conditions, there were considerable barriers to the effectiveness of different recruitment strategies across each study site. Although recruitment from primary care proved extremely successful at one study site, this largely failed at another site prompting the implementation of multimodal recruitment strategies including a successful media campaign to ensure sufficient participants were enrolled and the study was adequately powered. From this experience, we propose where appropriate the early engagement and investment in media campaigns to enhance recruitment into clinical trials. Trial Registration: ISRCTN30845205.","{'model': 'tldr@v2.0.0', 'text': 'From this experience of recruiting into a nationally funded, multicentre, randomised controlled trial of the Endobarrier versus standard medical therapy in obese patients with type 2diabetes, it is proposed where appropriate the early engagement and investment in media campaigns to enhance recruitment into clinical trials are proposed.'}",https://bmjopen.bmj.com/content/bmjopen/9/11/e032439.full.pdf
-Deferiprone as adjunctive treatment for patients with invasive mucormycosis: A retrospective case series,M. Chitasombat,"Mucormycosis is a life-threatening disease requiring multimodal treatment with antifungals and surgery. The mortality rate remains high, prompting consideration of alternative treatment strategies. Deferiprone has in vitro activity against Mucorales, but its efficacy has never been evaluated in humans. Here, we retrospectively analyzed patients with confirmed mucormycosis who received deferiprone from 2011 to 2017. Five patients had hematologic malignancies and one was diabetic. The sites of infection included sinus-orbit-cerebral (67%), lung (17%), and disseminated infection (17%). Surgery was performed in 83% of cases and achieved local control for 33% of patients. A combination regimen of polyenes plus echinocandins was administered with stepdown treatment using posaconazole. The median duration of antifungal treatment was 86 days (range: 46-435 days) days. Deferiprone was given as adjunctive treatment with a median dose and duration of 100 mg/kd/day (range: 86.2-100 mg/kg/day) and 25 days (range: 15-215 days), respectively. Overall, deferiprone was well-tolerated. Successful outcomes were observed at 12-week follow-up for 67% of patients. The mortality rate at 180- day follow-up was 50%. Adjunctive therapy with deferiprone showed safety and tolerability.","{'model': 'tldr@v2.0.0', 'text': 'Adjunctive therapy with deferiprone showed safety and tolerability and successful outcomes were observed at 12-week follow-up for 67% of patients, and the mortality rate at 180- day follow- up was 50%.'}",https://www.pagepress.org/journals/index.php/idr/article/download/7765/7453
-Pain Management in Cancer Patients Using a Mobile App: Study Design of a Randomized Controlled Trial,S. Agboola,"Background Despite the availability of effective medications and clinical guidelines for pain management, pain control is suboptimal in a sizeable proportion of patients with cancer pain. The National Comprehensive Cancer Network guidelines recommend a comprehensive and multimodal approach for management of cancer pain. We developed a mobile phone application, ePAL, based on clinical guidelines to empower patients for cancer pain management by prompting regular pain assessments and coaching for self-management. Objective The objective of this study is to evaluate the effect of a multidimensional mobile phone-based pain management application, ePAL, on controlling cancer pain and improving quality of life in patients with cancer pain being treated at an academic palliative care clinic. Methods The study will be implemented as a 2-arm randomized controlled trial with 110 adult patients with CP who own a mobile phone over a follow-up period of two months. Participants will be randomized to either the intervention group receiving ePAL and usual care or to a control group receiving only usual care. The brief pain inventory will be used to assess our primary outcome which is pain intensity. We will also evaluate the effect of the intervention on secondary outcomes which include the effect of the intervention on hospital utilization for pain crisis, quality of life, adherence to analgesic medications, barriers to pain control, anxiety and patient engagement. Instruments that will be used in evaluating secondary outcomes include the Brief Pain Inventory, Morisky Medication Adherence Scale, Barriers Questionnaire-II, Functional Assessment of Cancer Therapy–General, Edmonton Symptom Assessment System, Generalized Anxiety Disorder 7-item scale, and the Functional Assessment of Chronic Illness Therapy-Fatigue. The intention-to-treat approach will be used to evaluate outcomes. Our primary outcome, pain intensity, measured longitudinally over eight weeks, will be assessed by mixed model repeated analysis. Effect sizes will be calculated as mean group differences with standard deviations. Results The study is still in progress. We hope to have results by the end of 2015. Conclusions The multidimensional approach to pain management implemented on a mobile phone application could lead to significant improvements in patient outcomes. Trial Registration ClinicalTrials.gov NCT02069743; https://clinicaltrials.gov/ct2/show/NCT02069743 (Archived by WebCite at http://www.webcitation.org/6Qb65XGGA).","{'model': 'tldr@v2.0.0', 'text': 'The multidimensional approach to pain management implemented on a mobile phone application could lead to significant improvements in patient outcomes.'}",https://www.researchprotocols.org/2014/4/e76/PDF
-Deoxyhypusine synthase (DHPS) inhibitor GC7 induces p21/Rb-mediated inhibition of tumor cell growth and DHPS expression correlates with poor prognosis in neuroblastoma patients,A. Bandino,,"{'model': 'tldr@v2.0.0', 'text': 'Results suggest that spermidine and DHPS are key contributing factors in NB tumor proliferation through regulation of the p21/Rb signaling axis.'}",
-Using Formulations and Gaze to Encourage Parents to Talk About Their and Their Children's Health and Well-Being,S. Tiitinen,"In preventive health-care settings, professionals need to encourage clients to talk about their problems before they become critical. We use multimodal conversation analysis to demonstrate how public health nurses encourage parents to elaborate on their problems in a sample of preventive maternity and child health (MCH) clinics in Finland. The nurses topicalize the problem-relevant aspects of the parents’ problem-indicative talk by issuing a formulation of what the parent has just said (that is, by redescribing it in problem-related terms). This verbal practice is synchronized with a visual one—the nurse issues the formulation, receives the parent's response, and then gazes directly at them. This has the effect of prompting the parent to take up the problem and talk about it. We discuss the findings in relation to the institutional tasks in MCH care and to the role of gaze in constituting actions, such as formulations. Data are in Finnish with English translation.","{'model': 'tldr@v2.0.0', 'text': 'Multiodal conversation analysis is used to demonstrate how public health nurses encourage parents to elaborate on their problems in a sample of preventive maternity and child health clinics in Finland, and the role of gaze in constituting actions, such as formulations.'}",https://trepo.tuni.fi/bitstream/10024/104965/1/using_formulations_and_gaze_2014.pdf
-UNILATERAL DIFFUSE UVEAL MELANOCYTIC PROLIFERATION,R. Spaide,"Purpose: To report a unilateral case of what is named bilateral diffuse uveal melanocytic proliferation and consider the consequences of this finding. Methods: The ocular findings were investigated with multimodal imaging to include color fundus photography, fluorescein angiography, autofluorescence imaging, and enhanced depth imaging optical coherence tomography. Results: A 66-year-old woman had a history of breast cancer 23 years previously that was treated and the patient was free of disease since. She developed a recent decrease in visual acuity in her left eye prompting referral. She was seen to have an alteration in the pigmentation of the posterior pole of the left eye with dispersed red placoid spots. Autofluorescence imaging showed nummular areas of absent autofluorescence signal, which corresponded to areas of hyperfluorescence during fluorescein angiography. The placoid spots were hyperautofluorescent and hypofluorescent in autofluorescence and fluorescein angiography, respectively. She had diffuse thickening and infiltration of the choroid in the left eye. Because of the ocular findings, the patient underwent a systemic evaluation and was found to have widely metastatic disease with an unknown primary cancer. No progression of disease was seen in the left eye over a 6-month follow-up, and the right eye never showed any abnormality, except for a modest cataract, by any means of examination. Conclusion: Bilateral uveal melanocytic proliferation has been attributed to a paraneoplastic process, allegedly from a factor in the IgG fraction of the serum. However, unilateral involvement suggests that there are other factors involved in disease manifestation.","{'model': 'tldr@v2.0.0', 'text': 'Bilateral uveal melanocytic proliferation has been attributed to a paraneoplastic process, allegedly from a factor in the IgG fraction of the serum, but unilateral involvement suggests that there are other factors involved in disease manifestation.'}",
-Liver transplantation for hilar cholangiocarcinoma—a single-centre experience,S. Schüle,,"{'model': 'tldr@v2.0.0', 'text': 'Acceptable survival rates can be achieved by transplantation for hilar cholangiocarcinoma with lymph node metastases as the only exclusion criterion.'}",
-Subjective cognitive complaints and objective memory performance influence prompt preference for instrumental activities of daily living.,Emily J Van Etten,"INTRODUCTION Declines in memory and executive functioning often lead to difficulties completing instrumental activities of daily living (IADLs). Prompting technologies have the potential to help promote aging in place by providing support for the initiation and accurate completion of IADLs. In this study, we evaluate preferences of older adults for different levels of prompting support based on subjective and objective measures of cognitive functioning. METHOD Participants were 170 community-dwelling older adults split into two cognitive complaint groups: cognitive complaints and few cognitive complaints. After completing six IADL tasks (e.g., organize a pillbox, cook), each participant was asked to make a specific error (e.g., leave stove on) on three of the tasks. They were then prompted to correct the error with one of three different prompt modes: verbal indirect, verbal direct, multimodal verbal direct and video. RESULTS The cognitive complaints group reported greater preference for the multimodal prompt compared to the few cognitive complaints group. The indirect prompt was the least preferred by both groups. Furthermore, participants who recalled less on objective memory measures preferred more support in terms of prompt mode. Executive functioning did not appear to be related to prompt preference. CONCLUSION Level of subjective cognitive complaints and objective memory performance may influence amount of support preferred in a prompt.","{'model': 'tldr@v2.0.0', 'text': 'Level of subjective cognitive complaints and objective memory performance may influence amount of support preferred in a prompt.'}",https://europepmc.org/articles/pmc5597053?pdf=render
-"Type A behavior pattern: new perspectives on theory, assessment, and intervention.",C. Thoresen,"Research on the Type A behavior pattern (TA) has been plagued by inadequate theory, insensitive assessment, and insufficient interventions. These problems (e.g., using global dichotomous ratings of TA) have contributed to several failures of TA to predict cardiovascular outcomes prompting concern for new approaches. Conceptual models are discussed, such as a transactional model of TA, social cognitive theories (e.g., self-evaluative processes), and associative network theory (e.g., how emotions distort information). The view of hostility as the only pathogenic feature of TA is questioned. The ethnographic gap in TA research is described, and detailed descriptive and experimental case studies are recommended. Profile measures that assess all dimensions of TA are urged along with contextually sensitive multimodal assessments. Unlike results of some correlational studies, controlled TA interventions, although few, have consistently yielded positive results. Study of TA within an expanded conceptual perspective is encouraged, especially controlled experiments and interventions.","{'model': 'tldr@v2.0.0', 'text': 'The ethnographic gap in TA research is described, and detailed descriptive and experimental case studies are recommended, and the view of hostility as the only pathogenic feature of TA is questioned.'}",
-Literacy and Digital Culture in the Early Years,K. Dooley,"In this chapter we present data drawn from observations of kindergarten children using iPads and talk with the children, their parents/guardians and teachers. We identify a continuum of practices that extends from ‘educational apps’ teaching handwriting, sight words and so forth to uses of the iPad as a device for multimodal literacy development and substantive conversation around children’s creative work. At the current time high stakes testing and the implementation of the Australian Curriculum are prompting new public and professional conversations about literacy and digital technology. The iPad is construed as both cause of and solution to problems of traditional literacy education. In this context we describe the literacies enabled by educational software available on iPads. We higlight the time constraints which bore on teachers' capacity to enact their visions of literacy education through the iPad platform and suggest ways of reflecting on responses to this constraint.","{'model': 'tldr@v2.0.0', 'text': 'A continuum of practices that extends from ‘educational apps’ teaching handwriting, sight words and so forth to uses of the iPad as a device for multimodal literacy development and substantive conversation around children’s creative work is identified.'}",
-CM3: A Causal Masked Multimodal Model of the Internet,Armen Aghajanyan,"We introduce CM3, a family of causally masked generative models trained over a large corpus of structured multi-modal documents that can contain both text and image tokens. Our new causally masked approach generates tokens left to right while also masking out a small number of long token spans that are generated at the end of the string, instead of their original positions. The casual masking object provides a type of hybrid of the more common causal and masked language models, by enabling full generative modeling while also providing bidirectional context when generating the masked spans. We train causally masked language-image models on large-scale web and Wikipedia articles, where each document contains all of the text, hypertext markup, hyperlinks, and image tokens (from a VQVAE-GAN), provided in the order they appear in the original HTML source (before masking). The resulting CM3 models can generate rich structured, multi-modal outputs while conditioning on arbitrary masked document contexts, and thereby implicitly learn a wide range of text, image, and cross modal tasks. They can be prompted to recover, in a zero-shot fashion, the functionality of models such as DALL-E, GENRE, and HTLM. We set the new state-of-the-art in zero-shot summarization, entity linking, and entity disambiguation while maintaining competitive performance in the fine-tuning setting. We can generate images unconditionally, conditioned on text (like DALL-E) and do captioning all in a zero-shot setting with a single model.","{'model': 'tldr@v2.0.0', 'text': 'CM3, a family of causally masked generative models trained over a large corpus of structured multi-modal documents that can contain both text and image tokens, is introduced and set the new state-of-the-art in zero-shot summarization, entity linking, and entity disambiguation while maintaining competitive performance in the fine-tuning setting.'}",
-Show Me What and Tell Me How: Video Synthesis via Multimodal Conditioning,Ligong Han,"Most methods for conditional video synthesis use a single modality as the condition. This comes with major limitations. For example, it is problematic for a model conditioned on an image to generate a specific motion trajectory desired by the user since there is no means to provide motion information. Conversely, language information can describe the desired motion, while not precisely defining the content of the video. This work presents a multimodal video generation framework that benefits from text and images provided jointly or separately. We leverage the recent progress in quantized representations for videos and apply a bidirectional transformer with multiple modalities as inputs to predict a discrete video representation. To improve video quality and consistency, we propose a new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens. We introduce text augmentation to improve the robustness of the textual representation and diversity of generated videos. Our framework can incorporate various visual modalities, such as segmentation masks, drawings, and partially occluded images. It can generate much longer sequences than the one used for training. In addition, our model can extract visual information as suggested by the text prompt, e.g., “an object in image one is moving northeast”, and generate corresponding videos. We run evaluations on three public datasets and a newly collected dataset labeled with facial attributes, achieving state-of-the-art generation results on all four11Code: https://github.com/snap-research/MMVID and Webpage..","{'model': 'tldr@v2.0.0', 'text': 'To improve video quality and consistency, a new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens are proposed and text augmentation is introduced to improve the robustness of the textual representation and diversity of generated videos.'}",https://arxiv.org/pdf/2203.02573
-Prompt Tuning for Generative Multimodal Pretrained Models,Han Yang,"Prompt tuning has become a new paradigm for model tuning and it has demonstrated success in natural language pretraining and even vision pretraining. In this work, we explore the transfer of prompt tuning to multimodal pretraining, with a focus on generative multimodal pretrained models, instead of contrastive ones. Specifically, we implement prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks. Experimental results demonstrate that the light-weight prompt tuning can achieve comparable performance with finetuning and surpass other light-weight tuning methods. Besides, in comparison with finetuned models, the prompt-tuned models demonstrate improved robustness against adversarial attacks. We further figure out that experimental factors, including the prompt length, prompt depth, and reparameteratization, have great impacts on the model performance, and thus we empirically provide a recommendation for the setups of prompt tuning. Despite the observed advantages, we still find some limitations in prompt tuning, and we correspondingly point out the directions for future studies. Codes are available at \url{https://github.com/OFA-Sys/OFA}","{'model': 'tldr@v2.0.0', 'text': 'This work implements prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks, with a focus on generative multimodal pretrained models, instead of contrastive ones.'}",http://arxiv.org/pdf/2208.02532
-Opal: Multimodal Image Generation for News Illustration,Vivian Liu,"Advances in multimodal AI have presented people with powerful ways to create images from text. Recent work has shown that text-to-image generations are able to represent a broad range of subjects and artistic styles. However, finding the right visual language for text prompts is difficult. In this paper, we address this challenge with Opal, a system that produces text-to-image generations for news illustration. Given an article, Opal guides users through a structured search for visual concepts and provides a pipeline allowing users to generate illustrations based on an article’s tone, keywords, and related artistic styles. Our evaluation shows that Opal efficiently generates diverse sets of news illustrations, visual assets, and concept ideas. Users with Opal generated two times more usable results than users without. We discuss how structured exploration can help users better understand the capabilities of human AI co-creative systems.","{'model': 'tldr@v2.0.0', 'text': 'How structured exploration can help users better understand the capabilities of human AI co-creative systems is discussed, and Opal, a system that produces text-to-image generations for news illustration, is addressed.'}",https://arxiv.org/pdf/2204.09007
-"Joint engagement in the home environment is frequent, multimodal, timely, and structured.",Catalina Suarez-Rivera,"Infants develop in a social context, surrounded by knowledgeable caregivers who scaffold learning through shared engagement with objects. However, researchers have typically examined joint engagement in structured tasks, where caregivers sit near infants and display frequent, prompt, and multimodal behaviors around the objects of infant action. Which features of joint engagement generalize to the real-world? Despite the importance of joint engagement for infant learning, critical assumptions around joint engagement in everyday interaction remain unexamined. We investigated behavioral and temporal features of joint engagement in the home environment, where objects for play abound and dyad proximity fluctuates. Infant manual actions, mother manual and verbal behaviors, and dyad proximity were coded frame-by-frame from 2-h naturalistic recordings of 13- to 23-month-old infants and their mothers (N = 38). Infants experienced rich, highly structured, multimodal mother input around the objects of their actions. Specifically, joint engagement occurred within seconds of infant action and was amplified in the context of interpersonal proximity. Findings validate laboratory-based research on characteristics of joint engagement while highlighting unique properties around the role of mother-infant proximity and temporal structuring of caregiver input over extended time frames. Implications for the social contexts that support infant learning and development are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Joint engagement occurred within seconds of infant action and was amplified in the context of interpersonal proximity, highlighting unique properties around the role of mother-infant proximity and temporal structuring of caregiver input over extended time frames.'}",https://discovery.ucl.ac.uk/10142739/1/Suarez-Rivera%20et%20al.%2C%202022-%20JE%20at%20home_Infancy.pdf
-Draw Your Art Dream: Diverse Digital Art Synthesis with Multimodal Guided Diffusion,Nisha Huang,"Digital art synthesis is receiving increasing attention in the multimedia community because of engaging the public with art effectively. Current digital art synthesis methods usually use single-modality inputs as guidance, thereby limiting the expressiveness of the model and the diversity of generated results. To solve this problem, we propose the multimodal guided artwork diffusion (MGAD) model, which is a diffusion-based digital artwork generation approach that utilizes multimodal prompts as guidance to control the classifier-free diffusion model. Additionally, the contrastive language-image pretraining (CLIP) model is used to unify text and image modalities. Extensive experimental results on the quality and quantity of the generated digital art paintings confirm the effectiveness of the combination of the diffusion model and multimodal guidance. Code is available at https://github.com/haha-lisa/MGAD-multimodal-guided-artwork-diffusion.",,https://dl.acm.org/doi/pdf/10.1145/3503161.3548282
-Contrastive Language-Vision AI Models Pretrained on Web-Scraped Multimodal Data Exhibit Sexual Objectification Bias,R. Wolfe,"Warning: The content of this paper may be upsetting or triggering. Nine language-vision AI models trained on web scrapes with the Contrastive Language-Image Pretraining (CLIP) objective are evaluated for evidence of a bias studied by psychologists: the sexual objectification of girls and women, which occurs when a person’s human characteristics, such as emotions, are disregarded and the person is treated as a body or a collection of body parts. We replicate three experiments in the psychology literature quantifying sexual objectification and show that the phenomena persist in trained AI models. A first experiment uses standardized images of women from the Sexual OBjectification and EMotion Database, and finds that human characteristics are disassociated from images of objectified women: the model’s recognition of emotional state is mediated by whether the subject is fully or partially clothed. Embedding association tests (EATs) return significant effect sizes for both anger (d > 0.80) and sadness (d > 0.50), associating images of fully clothed subjects with emotions. GRAD-CAM saliency maps highlight that CLIP gets distracted from emotional expressions in objectified images where subjects are partially clothed. A second experiment measures the effect in a representative application: an automatic image captioner (Antarctic Captions) includes words denoting emotion less than 50% as often for images of partially clothed women than for images of fully clothed women. A third experiment finds that images of female professionals (scientists, doctors, executives) are likely to be associated with sexual descriptions relative to images of male professionals. A fourth experiment shows that a prompt of ""a [age] year old girl"" generates sexualized images (as determined by an NSFW classifier) up to 73% of the time for VQGAN-CLIP (age 17), and up to 42% of the time for Stable Diffusion (ages 14 and 18); the corresponding rate for boys never surpasses 9%. The evidence indicates that language-vision AI models trained on automatically collected web scrapes learn biases of sexual objectification, which propagate to downstream applications.",,https://dl.acm.org/doi/pdf/10.1145/3593013.3594072
-Bi-Bimodal Modality Fusion for Correlation-Controlled Multimodal Sentiment Analysis,Wei Han,"Multimodal sentiment analysis aims to extract and integrate semantic information collected from multiple modalities to recognize the expressed emotions and sentiment in multimodal data. This research area’s major concern lies in developing an extraordinary fusion scheme that can extract and integrate key information from various modalities. However, previous work is restricted by the lack of leveraging dynamics of independence and correlation between modalities to reach top performance. To mitigate this, we propose the Bi-Bimodal Fusion Network (BBFN), a novel end-to-end network that performs fusion (relevance increment) and separation (difference increment) on pairwise modality representations. The two parts are trained simultaneously such that the combat between them is simulated. The model takes two bimodal pairs as input due to the known information imbalance among modalities. In addition, we leverage a gated control mechanism in the Transformer architecture to further improve the final output. Experimental results on three datasets (CMU-MOSI, CMU-MOSEI, and UR-FUNNY) verifies that our model significantly outperforms the SOTA. The implementation of this work is available at https://github.com/declare-lab/multimodal-deep-learning and https://github.com/declare-lab/BBFN.","{'model': 'tldr@v2.0.0', 'text': 'The Bi-Bimodal Fusion Network (BBFN), a novel end-to-end network that performs fusion (relevance increment) and separation (difference increment) on pairwise modality representations that significantly outperforms the SOTA.'}",https://arxiv.org/pdf/2107.13669
-PromptMNER: Prompt-Based Entity-Related Visual Clue Extraction and Integration for Multimodal Named Entity Recognition,Xuwu Wang,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes to utilize entity-related prompts for extracting proper visual clues from the image to better integrate different modalities and address the popular semantic gap problem, and proposes a modality-aware attention mechanism for better cross-modal fusion.'}",
-Multimodal Data Guided Spatial Feature Fusion and Grouping Strategy for E-Commerce Commodity Demand Forecasting,Weiwei Cai,"E-commerce offers various merchandise for selling and purchasing with frequent transactions and commodity flows. An accurate prediction of customer needs and optimized allocation of goods is required for cost reduction. ,e existing solutions have significant errors and are unsuitable for addressing warehouse needs and allocation. ,at is why businesses cannot respond to customer demands promptly, as they need accurate and reliable demand forecasting.,erefore, this paper proposes spatial feature fusion and grouping strategies based onmultimodal data and builds a neural network predictionmodel for e-commodity demand. ,e designed model extracts order sequence features, consumer emotional features, and facial value features from multimodal data from e-commerce products. ,en, a bidirectional long short-term memory network(BiLSTM-) based grouping strategy is proposed. ,e proposed strategy fully learns the contextual semantics of time series data while reducing the influence of other features on the group’s local features. ,e output features of multimodal data are highly spatially correlated, and this paper employs the spatial dimension fusion strategy for feature fusion. ,is strategy effectively obtains the deep spatial relations among multimodal data by integrating the features of each column in each group across spatial dimensions. Finally, the proposedmodel’s prediction effect is tested using e-commerce dataset. ,e experimental results demonstrate the proposed algorithm’s effectiveness and superiority.","{'model': 'tldr@v2.0.0', 'text': 'A bidirectional long short-term memory network (BiLSTM-) based grouping strategy and spatial feature fusion and grouping strategies based on multimodal data are proposed and a neural network prediction model for e-commodity demand is built.'}",
-UniTranSeR: A Unified Transformer Semantic Representation Framework for Multimodal Task-Oriented Dialog System,Zhiyuan Ma,"As a more natural and intelligent interaction manner, multimodal task-oriented dialog system recently has received great attention and many remarkable progresses have been achieved. Nevertheless, almost all existing studies follow the pipeline to first learn intra-modal features separately and then conduct simple feature concatenation or attention-based feature fusion to generate responses, which hampers them from learning inter-modal interactions and conducting cross-modal feature alignment for generating more intention-aware responses. To address these issues, we propose UniTranSeR, a Unified Transformer Semantic Representation framework with feature alignment and intention reasoning for multimodal dialog systems. Specifically, we first embed the multimodal features into a unified Transformer semantic space to prompt inter-modal interactions, and then devise a feature alignment and intention reasoning (FAIR) layer to perform cross-modal entity alignment and fine-grained key-value reasoning, so as to effectively identify user’s intention for generating more accurate responses. Experimental results verify the effectiveness of UniTranSeR, showing that it significantly outperforms state-of-the-art approaches on the representative MMD dataset.","{'model': 'tldr@v2.0.0', 'text': 'UniTranSeR is proposed, a Unified Transformer Semantic Representation framework with feature alignment and intention reasoning for multimodal dialog systems that significantly outperforms state-of-the-art approaches on the representative MMD dataset.'}",https://aclanthology.org/2022.acl-long.9.pdf
-ConceptFusion: Open-set Multimodal 3D Mapping,Krishna Murthy Jatavallabhula,"Building 3D maps of the environment is central to robot navigation, planning, and interaction with objects in a scene. Most existing approaches that integrate semantic concepts with 3D maps largely remain confined to the closed-set setting: they can only reason about a finite set of concepts, pre-defined at training time. Further, these maps can only be queried using class labels, or in recent work, using text prompts. We address both these issues with ConceptFusion, a scene representation that is (1) fundamentally open-set, enabling reasoning beyond a closed set of concepts and (ii) inherently multimodal, enabling a diverse range of possible queries to the 3D map, from language, to images, to audio, to 3D geometry, all working in concert. ConceptFusion leverages the open-set capabilities of today's foundation models pre-trained on internet-scale data to reason about concepts across modalities such as natural language, images, and audio. We demonstrate that pixel-aligned open-set features can be fused into 3D maps via traditional SLAM and multi-view fusion approaches. This enables effective zero-shot spatial reasoning, not needing any additional training or finetuning, and retains long-tailed concepts better than supervised approaches, outperforming them by more than 40% margin on 3D IoU. We extensively evaluate ConceptFusion on a number of real-world datasets, simulated home environments, a real-world tabletop manipulation task, and an autonomous driving platform. We showcase new avenues for blending foundation models with 3D open-set multimodal mapping. For more information, visit our project page https://concept-fusion.github.io or watch our 5-minute explainer video https://www.youtube.com/watch?v=rkXgws8fiDs","{'model': 'tldr@v2.0.0', 'text': 'ConceptFusion enables effective zero-shot spatial reasoning, not needing any additional training or finetuning, and retains long-tailed concepts better than supervised approaches, outperforming them by more than 40% margin on 3D IoU.'}",http://arxiv.org/pdf/2302.07241
-LVP-M3: Language-aware Visual Prompt for Multilingual Multimodal Machine Translation,Hongcheng Guo,"Multimodal Machine Translation (MMT) focuses on enhancing text-only translation with visual features, which has attracted considerable attention from both natural language processing and computer vision communities. Recent advances still struggle to train a separate model for each language pair, which is costly and unaffordable when the number of languages increases in the real world. In other words, the multilingual multimodal machine translation (Multilingual MMT) task has not been investigated, which aims to handle the aforementioned issues by providing a shared semantic space for multiple languages. Besides, the image modality has no language boundaries, which is superior to bridging the semantic gap between languages. To this end,we first propose the Multilingual MMT task by establishing two new Multilingual MMT benchmark datasets covering seven languages.Then, an effective baseline LVP-M3 using visual prompts is proposed to support translations between different languages,which includes three stages (token encoding, language-aware visual prompt generation, and language translation). Extensive experimental results on our constructed benchmark datasets demonstrate the effectiveness of LVP-M3 method for Multilingual MMT.","{'model': 'tldr@v2.0.0', 'text': 'An effective baseline LVP-M3 using visual prompts is proposed to support translations between different languages, which includes three stages (token encoding, language-aware visual prompt generation, and language translation).'}",http://arxiv.org/pdf/2210.15461
-Multimodal Data for the Detection of Freezing of Gait in Parkinson’s Disease,Wei Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'A protocol has been designed to acquire multimodal physical and physiological information during FOG, including gait acceleration (ACC), electroencephalogram (EEG), electromyogram (EMG), and skin conductance (SC), and multi-modal data have been validated by a FOG detection task.'}",https://www.nature.com/articles/s41597-022-01713-8.pdf
-Designing Multimodal Interactive Dashboard of Disaster Management Systems,Abeer AlAbdulaali,"Disasters and crises are inevitable in this world. In the aftermath of a disaster, a society’s overall growth, resources, and economy are greatly affected as they cause damages from minor to huge proportions. Around the world, countries are interested in improving their emergency decision-making. The institutions are paying attention to collecting different types of data related to crisis information from various resources, including social media, to improve their emergency response. Previous efforts have focused on collecting, extracting, and classifying crisis data from text, audio, video, or files; however, the development of user-friendly multimodal disaster data dashboards to support human-to-system interactions during an emergency response has received little attention. Our paper seeks to fill this gap by proposing usable designs of interactive dashboards to present multimodal disaster information. For this purpose, we first investigated social media data and metadata for the required elicitation and analysis purposes. These requirements are then used to develop interactive multimodal dashboards to present complex disaster information in a usable manner. To validate our multimodal dashboard designs, we have conducted a heuristic evaluation. Experts have evaluated the interactive disaster dashboards using a customized set of heuristics. The overall assessment showed positive feedback from the evaluators. The proposed interactive multimodal dashboards complement the existing techniques of collecting textual, image, audio, and video emergency information and their classifications for usable presentation. The contribution will help the emergency response personnel in terms of useful information and observations for prompt responses to avoid significant damage.","{'model': 'tldr@v2.0.0', 'text': 'The proposed interactive multimodal dashboards complement the existing techniques of collecting textual, image, audio, and video emergency information and their classifications for usable presentation and will help the emergency response personnel in terms of useful information and observations for prompt responses to avoid significant damage.'}",https://www.mdpi.com/1424-8220/22/11/4292/pdf?version=1654601242
-M-VADER: A Model for Diffusion with Multimodal Context,Samuel Weinbach,"We introduce M-VADER: a diffusion model (DM) for image generation where the output can be specified using arbitrary combinations of images and text. We show how M-VADER enables the generation of images specified using combinations of image and text, and combinations of multiple images. Previously, a number of successful DM image generation algorithms have been introduced that make it possible to specify the output image using a text prompt. Inspired by the success of those models, and led by the notion that language was already developed to describe the elements of visual contexts that humans find most important, we introduce an embedding model closely related to a vision-language model. Specifically, we introduce the embedding model S-MAGMA: a 13 billion parameter multimodal decoder combining components from an autoregressive vision-language model MAGMA and biases finetuned for semantic search.","{'model': 'tldr@v2.0.0', 'text': 'The embedding model S-MAGMA is introduced: a 13 billion parameter multimodal decoder combining components from an autoregressive vision-language model MAGMA and biases finetuned for semantic search.'}",https://arxiv.org/pdf/2212.02936
-"Assistive Multimodal Robotic System (AMRSys): Security and Privacy Issues, Challenges, and Possible Solutions",Jims Marchang,"Assistive robotic systems could be a suitable solution to support a variety of health and care services, help independent living, and even simulate affection, to reduce loneliness. However, adoption is limited by several issues, as well as user concerns about ethics, data security, and privacy. Other than the common threats related to internet connectivity, personal robotic systems have advanced interaction possibilities, such as audio, video, touch, and gestures, which could be exploited to gain access to private data that are stored in the robot. Therefore, novel, safer methods of interaction should be designed to safeguard users’ privacy. To solicit further research on secure and private multimodal interaction, this article presents a thorough study of the state-of-the-art literature on data security and user privacy in interactive social robotic systems for health and care. In our study, we focus on social robotics to assist older people, which is a global challenge that is receiving a great deal of attention from the robotics and social care communities. This application will have a significant positive impact on the economy and society, but poses various security and privacy issues. This article analyses the key vulnerable areas where data leakage could occur during a multimodal interaction with a personal assistive robotic system. Thus, blockchain with a resource-aware framework, along with a continuous multifactor authentication mechanism, are envisaged as a potential solution for making such systems secure by design; therefore, increasing trust, acceptability, and adoption. Among the key cybersecurity research challenges, it is crucial to create an intelligent mechanism that autonomously determines the right trade-off between continuous user prompts and system usability, according to data types and personal preferences.","{'model': 'tldr@v2.0.0', 'text': 'Blockchain with a resource-aware framework, along with a continuous multifactor authentication mechanism, are envisaged as a potential solution for making assistive robotic systems secure by design; therefore, increasing trust, acceptability, and adoption.'}",https://www.mdpi.com/2076-3417/12/4/2174/pdf?version=1645512954
-Representation Learning through Multimodal Attention and Time-Sync Comments for Affective Video Content Analysis,Jicai Pan,"Although temporal patterns inherent in visual and audio signals are crucial for affective video content analysis, they have not been thoroughly explored yet. In this paper, we propose a novel Temporal-Aware Multimodal (TAM) method to fully capture the temporal information. Specifically, we design a cross-temporal multimodal fusion module that applies attention-based fusion to different modalities within and across video segments. As a result, it fully captures the temporal relations between different modalities. Furthermore, a single emotion label lacks supervision for learning representation of each segment, making temporal pattern mining difficult. We leverage time-synchronized comments (TSCs) as auxiliary supervision, since these comments are easily accessible and contain rich emotional cues. Two TSC-based self-supervised tasks are designed: the first aims to predict the emotional words in a TSC from video representation and TSC contextual semantics, and the second predicts the segment in which the TSC appears by calculating the correlation between video representation and TSC embedding. These self-supervised tasks are used to pre-train the cross-temporal multimodal fusion module on a large-scale video-TSC dataset, which is crawled from the web without labeling costs. These self-supervised pre-training tasks prompt the fusion module to perform representation learning on segments including TSC, thus capturing more temporal affective patterns. Experimental results on three benchmark datasets show that the proposed fusion module achieves state-of-the-art results in affective video content analysis. Ablation studies verify that after TSC-based pre-training, the fusion module learns more segments' affective patterns and achieves better performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs a cross-temporal multimodal fusion module that applies attention-based fusion to different modalities within and across video segments, and achieves state-of-the-art results in affective video content analysis.'}",
-Vision+X: A Survey on Multimodal Learning in the Light of Data,Ye Zhu,"We are perceiving and communicating with the world in a multisensory manner, where different information sources are sophisticatedly processed and interpreted by separate parts of the human brain to constitute a complex, yet harmonious and unified sensing system. To endow the machines with true intelligence, the multimodal machine learning that incorporates data from various modalities has become an increasingly popular research area with emerging technical advances in recent years. In this paper, we present a survey on multimodal machine learning from a novel perspective considering not only the purely technical aspects but also the nature of different data modalities. We analyze the commonness and uniqueness of each data format ranging from vision, audio, text and others, and then present the technical development categorized by the combination of Vision+X, where the vision data play a fundamental role in most multimodal learning works. We investigate the existing literature on multimodal learning from both the representation learning and downstream application levels, and provide an additional comparison in the light of their technical connections with the data nature, e.g., the semantic consistency between image objects and textual descriptions, or the rhythm correspondence between video dance moves and musical beats. The exploitation of the alignment, as well as the existing gap between the intrinsic nature of data modality and the technical designs, will benefit future research studies to better address and solve a specific challenge related to the concrete multimodal task, and to prompt a unified multimodal machine learning framework closer to a real human intelligence system.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes the commonness and uniqueness of each data format ranging from vision, audio, text and others, and presents the technical development categorized by the combination of Vision+X, where the vision data play a fundamental role in most multimodal learning works.'}",http://arxiv.org/pdf/2210.02884
-Multimodal Imaging and Biomarkers in Cardiac Amyloidosis,M. Jung,"Amyloidosis is a progressive infiltrative disease instigated by the extracellular deposition of amyloid fibrils in various organs such as the heart, kidney, and peripheral nerves. Cardiac amyloid deposits cause restrictive cardiomyopathy, leading to a poor prognosis in systemic amyloidosis. The most common etiologies of cardiac amyloidosis (CA) are immunoglobulin light chain deposits (AL-CA) and misfolded transthyretin deposits (ATTR-CA). In recent years, many developments have been accomplished in the field of diagnosis and treatment of CA. At present, ATTR-CA can be noninvasively diagnosed if the following two conditions are fulfilled in the setting of typical echocardiographic/cardiac MRI findings: (1) grade 2 or 3 myocardial uptake in bone scintigraphy confirmed by SPECT and (2) absence of monoclonal protein confirmed by serum-free light chain assay, and serum/urine protein electrophoresis with immunofixation test. Effective therapies are evolving in both types of CA (tafamidis for ATTR-CA and immunologic treatments for AL-CA). Thus, early suspicion and prompt diagnosis are crucial for achieving better outcomes. In this review, we have summarized the role of multimodal imaging (e.g., echocardiography, cardiac MRI, and bone scintigraphy) and biomarkers (e.g., troponin, BNP) in the diagnosis, risk stratification, and treatment monitoring of CA.","{'model': 'tldr@v2.0.0', 'text': 'The role of multimodal imaging and biomarkers in the diagnosis, risk stratification, and treatment monitoring of CA is summarized and early suspicion and prompt diagnosis are crucial for achieving better outcomes.'}",https://www.mdpi.com/2075-4418/12/3/627/pdf?version=1646296989
-PARACENTRAL ACUTE MIDDLE MACULOPATHY AFTER COVID-19: MULTIMODAL EVALUATION,C. Castro,"Paracentral acute middle maculopathy is a spectral domain optical coherence tomography finding characterized by macular band-like hyperreflective lesions that affect the middle layers of the retina, above the outer plexiform layer. COVID-19 may induce different ophthalmological manifestations, including retinal vascular changes. We report a case of paracentral acute middle maculopathy after SARS-CoV-2 infection. Purpose: To report the case and multimodal imaging findings of a healthy young woman who developed paracentral acute middle maculopathy 9 weeks after COVID-19. Methods: Case report. Ultra-widefield fundus photography, macular spectral domain optical coherence tomography, fluorescein angiography, and optical coherence tomography angiography were performed. Results: A 36-year-old woman went to the emergency department with sudden, painless, left eye vision loss. The only relevant medical history was COVID-19 9 weeks before. The best-corrected visual acuity was 20/200, a left eye relative afferent pupillary defect was present, and superficial hemorrhages throughout the macular area and peripheral retina were found. Nearly 4 hours after admission, the left eye best-corrected visual acuity recovered to 20/20 without relative afferent pupillary defect. Five days after presentation in the emergency department, the patient returned with recurrent left eye vision loss, with spontaneous recovery within 12 hours. Macular spectral domain optical coherence tomography revealed hyperreflectivity of the inner plexiform and inner nuclear layers, and the diagnosis of paracentral acute middle maculopathy was established. The patient started oral acetylsalicylic acid and oral prednisolone. The patient did not report any new episodes of vision loss, and there was a progressive resolution of abnormal fundus findings. Conclusion: SARS-CoV-2 infection increases the risk of vascular thrombotic events with possible involvement of the retinal circulation, and paracentral acute middle maculopathy may present as a possible complication. Ophthalmologists should be able to recognize it promptly through multimodal imaging findings.","{'model': 'tldr@v2.0.0', 'text': 'SARS-CoV-2 infection increases the risk of vascular thrombotic events with possible involvement of the retinal circulation, and paracentral acute middle maculopathy may present as a possible complication.'}",
-Few-shot Multimodal Sentiment Analysis Based on Multimodal Probabilistic Fusion Prompts,Xiaocui Yang,"Multimodal sentiment analysis has gained significant attention due to the proliferation of multimodal content on social media. However, existing studies in this area rely heavily on large-scale supervised data, which is time-consuming and labor-intensive to collect. Thus, there is a need to address the challenge of few-shot multimodal sentiment analysis. To tackle this problem, we propose a novel method called Multimodal Probabilistic Fusion Prompts (MultiPoint) that leverages diverse cues from different modalities for multimodal sentiment detection in the few-shot scenario. Specifically, we start by introducing a Consistently Distributed Sampling approach called CDS, which ensures that the few-shot dataset has the same category distribution as the full dataset. Unlike previous approaches primarily using prompts based on the text modality, we design unified multimodal prompts to reduce discrepancies between different modalities and dynamically incorporate multimodal demonstrations into the context of each multimodal instance. To enhance the model's robustness, we introduce a probabilistic fusion method to fuse output predictions from multiple diverse prompts for each input. Our extensive experiments on six datasets demonstrate the effectiveness of our approach. First, our method outperforms strong baselines in the multimodal few-shot setting. Furthermore, under the same amount of data (1% of the full dataset), our CDS-based experimental results significantly outperform those based on previously sampled datasets constructed from the same number of instances of each class.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method called Multimodal Probabilistic Fusion Prompts (MultiPoint) that leverages diverse cues from different modalities for multimodal sentiment detection in the few-shot scenario and introduces a probabilistic fusion method to fuse output predictions from multiple diverse prompts for each input.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612181
-Beyond Bounding Box: Multimodal Knowledge Learning for Object Detection,Wei Feng,"Multimodal supervision has achieved promising results in many visual language understanding tasks, where the language plays an essential role as a hint or context for recognizing and locating instances. However, due to the defects of the human-annotated language corpus, multimodal supervision remains unexplored in fully supervised object detection scenarios. In this paper, we take advantage of language prompt to introduce effective and unbiased linguistic supervision into object detection, and propose a new mechanism called multimodal knowledge learning (\textbf{MKL}), which is required to learn knowledge from language supervision. Specifically, we design prompts and fill them with the bounding box annotations to generate descriptions containing extensive hints and context for instances recognition and localization. The knowledge from language is then distilled into the detection model via maximizing cross-modal mutual information in both image- and object-level. Moreover, the generated descriptions are manipulated to produce hard negatives to further boost the detector performance. Extensive experiments demonstrate that the proposed method yields a consistent performance gain by 1.6\% $\sim$ 2.1\% and achieves state-of-the-art on MS-COCO and OpenImages datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper takes advantage of language prompt to introduce effective and unbiased linguistic supervision into object detection, and proposes a new mechanism called multimodal knowledge learning (\\textbf{MKL}), which is required to learn knowledge from language supervision.'}",http://arxiv.org/pdf/2205.04072
-Multimodal Opioid-sparing Analgesia for Total Knee Arthroplasty: Results from a Retrospective Case Series of 40 Patients,A. Bronstone,"The high burden of opioids has prompted orthopedic surgeons to identify effective nonopioid pain management modalities. In this pilot retrospective case series, we describe opioid use and pain during the first 3 months after total knee arthroplasty (TKA) for consecutive patients treated with opioid-free multimodal analgesia. During May and October 2020, 40 consecutive patients who underwent primary unilateral TKA performed by a single surgeon at an academic community hospital received a multimodal analgesia regimen consisting of preoperative cryoneurolysis; perioperative oral and intravenous analgesics, a neuraxial (spinal) and regional adductor canal block, and local infiltration of liposomal bupivacaine (Exparel); intraoperative periarticular infiltration of bupivacaine hydrochloride (Marcaine); and postoperative oral non-opioid analgesics. A Prescription Monitoring Program database was examined to capture opioid prescriptions filled during the 3 months before and after TKA. Self-reported pain was assessed before surgery and at 2 weeks and 3 months after surgery. Two thirds (67.5%; 27/40) of patients recovered from TKA without using opioids. Whereas 100% of the experienced opioid users required opioids after TKA, only 15.6% of opioid-naïve patients required postoperative opioids. Opioid-experienced patients had a higher mean number of post-TKA opioid prescriptions (3.1 vs. 1.6) and total morphine milligram equivalent (985 vs. 265) than opioid-naïve patients. Patients who used opioids after TKA reported higher levels of pain at each time point than those who had opioid-free TKA. Multimodal analgesia allows for opioid-free TKA for the majority of patients, particularly those without a history of recent opioid use.","{'model': 'tldr@v2.0.0', 'text': 'This pilot retrospective case series describes opioid use and pain during the first 3 months after total knee arthroplasty (TKA) for consecutive patients treated with opioid-free multimodal analgesia for the majority of patients, particularly those without a history of recent opioid use.'}",https://journaloei.scholasticahq.com/article/33296.pdf
-Weakly Supervised Multimodal 30-Day All-Cause Mortality Prediction for Pulmonary Embolism Patients,Noa Cahan,"Pulmonary embolism (PE) is a common life-threatening condition with a challenging diagnosis, as patients often present with nonspecific symptoms. Prompt and accurate detection of PE and specifically an assessment of its severity are critical for managing patient treatment. We introduce diverse multimodal fusion models that are capable of utilizing weakly-labeled multi-modal data, combining both volumetric pixel data and clinical patient data for automatic risk stratification of PE. The best performing multimodality model is an intermediate fusion model that achieves an area under the curve (AUC) of 0.96 for assessing PE severity, with a sensitivity of 90% and specificity of 94%. To the best of our knowledge, this is the first study that attempted to automatically assess PE severity.","{'model': 'tldr@v2.0.0', 'text': 'Diverse multimodal fusion models that are capable of utilizing weakly-labeled multi-modal data are introduced, combining both volumetric pixel data and clinical patient data for automatic risk stratification of PE.'}",
-Added Value of Ultrasound-Based Multimodal Imaging to Diagnose Hepatic Sclerosed Hemangioma before Biopsy and Resection,Feiqian Wang,"Imaging methods have the overwhelming advantage of being non-invasive in the diagnosis of hepatic lesions and, thanks to technical developments in the field of ultrasound (US), radiation exposure can also be avoided in many clinical situations. In particular, contrast-enhanced US (CEUS) outperforms other radiological methods in regard to real-time images, repeatability, and prompt reporting and demonstrates relatively few contraindications and adverse reactions. In this study, we reported in detail a rare benign tumor: hepatic sclerosed hemangioma (HSH). We described US-based multimodal imaging (B-flow imaging, US elastography, and Sonazoid CEUS) features of this HSH case. Furthermore, by summarizing the recently published literature on the imaging diagnosis of HSH, we offered readers comprehensive knowledge of conventional imaging methods (CT, MRI) and CEUS in the diagnosis of HSH and preliminarily discussed their mechanism of pathology-based diagnosis. Our multimodal imaging approach may provide a diagnostic strategy for HSH, thus avoiding unnecessary biopsy or resection.","{'model': 'tldr@v2.0.0', 'text': 'This study reported in detail a rare benign tumor: hepatic sclerosed hemangioma (HSH) and described US-based multimodal imaging features of this HSH case and offered readers comprehensive knowledge of conventional imaging methods (CT, MRI) and CEUS in the diagnosis of HSH.'}",https://www.mdpi.com/2075-4418/12/11/2818/pdf?version=1668590166
-Digital Storytelling: A Didactic Approach to Multimodal Coherence,S. Meier,"The aim of this article is to explain multimodal coherence-making as a transcribing practice and how this can be used to teach multimodal, narrative, and media competences in different genres. In multimodal arrangements, language makes images readable in specific ways and images make language understandable in different ways. This results in an abductive understanding process that can be used in teaching and learning contexts. This idea of meaning-making is based on the social semiotic approach of style. According to the understanding of semiotic meta functions, this approach considers style as the practice of selecting, forming, and composing semiotic resources. These stylistic practices realize a subjective appropriation of discursive and habitual patterns, which are carried out within the semiotic and technological dispositions (affordances) of the situationally used media infrastructures. In this sense, digital storytelling is a multimodal style practice with digital tools. Multimodal storytelling in educational contexts means that teachers and learners are prompted to bring the communicative functions of text, image, video, and audio into narrative coherence. Based on a journalistic Instagram story, this article reconstructs the media-practical, multimodal, and narrative skills that are prototypically necessary. Based on this analysis, these competencies are operationalized to make them usable for new teaching/learning arrangements using digital storytelling.",,https://www.frontiersin.org/articles/10.3389/fcomm.2022.906268/pdf
-Multimodal Representations Learning Based on Mutual Information Maximization and Minimization and Identity Embedding for Multimodal Sentiment Analysis,Jiahao Zheng,"Multimodal sentiment analysis (MSA) is a fundamental complex research problem due to the heterogeneity gap between di ff erent modalities and the ambiguity of human emotional expression. Although there have been many successful attempts to construct multimodal representations for MSA, there are still two challenges to be addressed: 1) A more robust multimodal representation needs to be constructed to bridge the heterogeneity gap and cope with the complex multimodal interactions, and 2) the contextual dynamics must be modeled e ff ectively throughout the information flow. In this work, we propose a multimodal representation model based on Mutual information Maximization and Minimization and Identity Embedding (MMMIE). We combine mutual information maximization between modal pairs, and mutual information minimization between input data and corresponding features to mine the modal-invariant and task-related information. Furthermore, Identity Embedding is proposed to prompt the downstream network to perceive the contextual information. Experimental results on two public datasets demonstrate the e ff ectiveness of the proposed model.","{'model': 'tldr@v2.0.0', 'text': 'A multimodal representation model based on Mutual information Maximization and Minimization and Identity Embedding (MMMIE) is proposed that combines mutual information maximization between modal pairs, and mutual information minimization between input data and corresponding features to mine the modal-invariant and task-related information.'}",
-"Feasibility of a multimodal exercise, nutrition, and palliative care intervention in advanced lung cancer",M. Ester,,"{'model': 'tldr@v2.0.0', 'text': 'A multimodal physical activity, nutrition, and palliative symptom management intervention is feasible and shows potential benefits on quality of life that warrant further investigation in a larger cohort\xa0trial.'}",https://bmccancer.biomedcentral.com/track/pdf/10.1186/s12885-021-07872-y
-Multimodal Study of PRPH2 Gene-Related Retinal Phenotypes,G. Antonelli,"PRPH2 gene mutations are frequently found in inherited retinal dystrophies (IRD) and are associated with a wide spectrum of clinical phenotypes. We studied 28 subjects affected by IRD carrying pathogenic PRPH2 mutations, belonging to 11 unrelated families. Functional tests (best-corrected visual acuity measurement, chromatic test, visual field, full-field, 30 Hz flicker, and multifocal electroretinogram), morphological retino-choroidal imaging (optical coherence tomography, optical coherence tomography angiography, and fundus autofluorescence), and clinical data were collected and analyzed. Common primary complaints, with onset in their 40s, were visual acuity reduction and abnormal dark adaptation. Visual acuity ranged from light perception to 20/20 Snellen. Visual field peripheral constriction and central scotoma were found. Chromatic sense was reduced in one third of patients. Electrophysiological tests were abnormal in most of the patients. Choroidal neovascular lesions were detected in five patients. Three novel PRPH2 variants were found in four different families. Based on the present multimodal study, we identified seven distinct PRPH2 phenotypes in 11 unrelated families carrying either different mutations or the same mutation, both within the same family or among them. Fundus autofluorescence modality turned out to be the most adequate imaging method for early recognition of this dystrophy, and the optical coherence tomography angiography was highly informative to promptly detect choroidal neovascularization, even in the presence of the extensive chorioretinal atrophy phenotype.","{'model': 'tldr@v2.0.0', 'text': 'Fundus autofluorescence modality turned out to be the most adequate imaging method for early recognition of this dystrophy, and the optical coherence tomography angiography was highly informative to promptly detect choroidal neovascularization, even in the presence of the extensive chorioretinal atrophy phenotype.'}",https://www.mdpi.com/2075-4418/12/8/1851/pdf?version=1660106282
-Nonopioid Treatments for Chronic Pain-Integrating Multimodal Biopsychosocial Approaches to Pain Management.,S. Eucker,"Chronic pain is a major health burden in the US, with recent estimates indicating that 1 in 5 adults has chronic pain.1 Nearly 40% of these individuals experience substantial impairments in function or daily activities as a result of their pain.1 The prevalence of chronic pain has been growing in the past 2 decades, accompanied by an increase in opioid prescribing for pain. However, the exponential increase in opioid-related deaths prompted the development of the 2016 Centers for Disease Control and Prevention (CDC) guideline to reduce opioid prescribing for chronic pain.2 This guideline, along with several other efforts to limit opioids, has been associated with decreased opioid prescribing rates across multiple settings and patient populations.3,4 Although promising, minimizing the use of 1 treatment strategy without ensuring access to other effective and viable alternatives may expose patients to inadequate treatment and worsening of chronic pain. The CDC guideline attempts to address this issue by recommending nonopioid medication and nonpharmacologic pain management as first-line treatments for chronic pain.2 However, a major question remains: has the reduced use of opioids been accompanied by greater use of alternative treatments for chronic pain? The study by Goldstick et al5 takes a major step toward answering this question. The authors assessed the prescribing rates of several classes of nonopioid pain medications before and after the release of the 2016 CDC guideline. They included a large nationwide cohort of commercially insured patients with some of the most common types of chronic pain: neck or back pain, osteoarthritis, nonmigraine headaches, and fibromyalgia. The analysis focused on specific medication classes that have been commonly prescribed for chronic pain: nonsteroidal anti-inflammatory drugs, acetaminophen, gabapentinoid and carbamazepine anticonvulsants, and antidepressants. These medications have been shown to provide small-to-moderate short-term benefit and are recommended by guidelines for chronic pain treatment.6 Medications whose evidence was primarily for acute pain management (eg, skeletal muscle relaxants) were not included. The authors constructed sequential cohorts to analyze the prescribing patterns over time for the overall population and for specific subpopulations. Similar to previous work, the study by Goldstick et al5 found greater reductions in opioid prescribing after the release of the 2016 CDC guideline compared with the preguideline pattern. The authors also found that, although nonopioid prescribing rates had remained relatively constant over the 4-year period before the release of the guideline, these rates steadily increased during the 2 years after the release. Similar patterns were found when examining specific subpopulations, including those with chronic pain, previous opioid exposure, substance use disorder, anxiety disorder, or mood disorder.5 These shifts suggest that nonopioid medications replaced opioid medications, but the overall increase in nonopioid prescribing rates was small. On the other hand, claims data analysis cannot account for commonly used over-the-counter medications (eg, nonsteroidal anti-inflammatory drugs and acetaminophen). Furthermore, although not guideline-recommended, other medication classes (eg, benzodiazepines and topical lidocaine) that are sometimes prescribed for pain were not included in the analysis. Thus, these findings may potentially underrepresent the increase in clinician recommendations for nonopioid pain medications. + Related article","{'model': 'tldr@v2.0.0', 'text': 'A large nationwide cohort of commercially insured patients with some of the most common types of chronic pain included, found greater reductions in opioid prescribing after the release of the 2016 CDC guideline, but the overall increase in nonopioid prescribing rates was small.'}",https://jamanetwork.com/journals/jamanetworkopen/articlepdf/2793231/eucker_2022_ic_220103_1654263506.10054.pdf
-"Multimodal C4: An Open, Billion-scale Corpus of Images Interleaved With Text",Wanrong Zhu,"In-context vision and language models like Flamingo support arbitrarily interleaved sequences of images and text as input. This format not only enables few-shot learning via interleaving independent supervised (image, text) examples, but also, more complex prompts involving interaction between images, e.g.,""What do image A and image B have in common?""To support this interface, pretraining occurs over web corpora that similarly contain interleaved images+text. To date, however, large-scale data of this form have not been publicly available. We release Multimodal C4, an augmentation of the popular text-only C4 corpus with images interleaved. We use a linear assignment algorithm to place images into longer bodies of text using CLIP features, a process that we show outperforms alternatives. Multimodal C4 spans everyday topics like cooking, travel, technology, etc. A manual inspection of a random sample of documents shows that a vast majority (88%) of images are topically relevant, and that linear assignment frequently selects individual sentences specifically well-aligned with each image (80%). After filtering NSFW images, ads, etc., the resulting corpus consists of 101.2M documents with 571M images interleaved in 43B English tokens.","{'model': 'tldr@v2.0.0', 'text': 'This work releases Multimodal C4, an augmentation of the popular text-only C4 corpus with images interleaved, and uses a linear assignment algorithm to place images into longer bodies of text using CLIP features, a process that shows outperforms alternatives.'}",http://arxiv.org/pdf/2304.06939
-"Bionic Soft Multimodal Actuators for Fast, Large Deformation under Ultralow Magnetic Conditions",Keren Dai,"Very recently, magnetically driven soft actuators have prompted increasing interest due to their programmable deformation, swift response, and remote actuation. However, it is still challenging to trigger strong and fast actuating performances with an extremely weak magnetic field due to the difficult task of in situ programming of magnetic domains and the limited mechanical structures. Here, first a bionic sandwich structure is proposed for designing soft magnetic actuators with a specific threshold value. The crab‐like jointed structure of PDMS‐embedded NdFeB upper/lower layers is essential for generating the desirable threshold effect, while a flytrap‐inspired soft interlayer is further implemented to decrease the driving magnetic field. Theoretical analysis and numerical simulations are implemented to optimize the actuating performances of soft actuators (deformation rate, deformation angle) by modulating the structural parameters. Experimental results show that the biomimetic features yield a promptly switchable bistable state, a superior strong deformation rate of 1.93, and a maximum deformation angle of 25.5° under an ultralow magnetic field of 1 mT. Two demonstrative applications of soft actuators are investigated, including threshold switching and soft grippers, suggesting their broad applications in engineering fields. Remarkably, the magnetic soft grippers associated with the gecko‐inspired adhesion surface exhibit an improved and stable grasping ability. This work focuses on designing highly deformable actuators and functions in emerging areas of soft robotics.",,
-A Prospective Study Evaluating Health-Related Quality of Life Following a Multimodal Treatment for Colorectal Cancer,M. Perrone,,"{'model': 'tldr@v2.0.0', 'text': 'A better HR-QoL at baseline was associated with better physical, social and sexual functioning, positive body image and sexual pleasure after 5\xa0years, and allows the early detection of patients at risk, favoring prompt patient-centered interventions.'}",
-N24News: A New Dataset for Multimodal News Classification,Zhen Wang,"Current news datasets merely focus on text features on the news and rarely leverage the feature of images, excluding numerous essential features for news classification. In this paper, we propose a new dataset, N24News, which is generated from New York Times with 24 categories and contains both text and image information in each news. We use a multitask multimodal method and the experimental results show multimodal news classification performs better than text-only news classification. Depending on the length of the text, the classification accuracy can be increased by up to 8.11%. Our research reveals the relationship between the performance of a multimodal classifier and its sub-classifiers, and also the possible improvements when applying multimodal in news classification. N24News is shown to have great potential to prompt the multimodal news studies.","{'model': 'tldr@v2.0.0', 'text': 'A new dataset, N24News, is proposed, which is generated from New York Times with 24 categories and contains both text and image information in each news and shows multimodal news classification performs better than text-only news classification.'}",
-A Novel Multimodal Method for Depression Identification,Rahul Singhal,"Depression is one of the most prominent mental health issues, characterized by a depressed low mood and an absence of enthusiasm in activities. In terms of early detection, accurate diagnosis, and effective treatment, doctors face a serious challenge from depression, which is a serious global health issue. For patients with this mental disease to receive prompt medical attention and improve their general well-being, early identification is essential. For the purpose of detecting various psychological illnesses including depression, anxiety, and post-traumatic stress disorder, medical audio consultations along with survey responses have been used. A depressed individual displays a range of subtle signs that may be more easily identified by combining the results of multiple modalities. Multimodality involves extracting maximum information from data by using multiple modes, so that the deep learning model can be trained efficiently to give better results. Given that each modality functions differently, combining various modalities is not easy, and each origin of a modality takes on a different form. It is clear from the literature that is currently significant in the area that, combining the modalities yields positive outcomes. A trustworthy approach to identify depression is thus urgently needed because it continues to be a problem for many individuals in today’s society. In this work, textual and audio features are incorporated related to the identification of depression, and a novel multimodal approach using an optimized Bi-directional Long Short -Term Memory model that recognizes premature depression is suggested for medical intervention before it develops further.",,
-A Multimodal Deep Learning-Based Fault Detection Model for a Plastic Injection Molding Process,Gyeongho Kim,"The authors of this work propose a deep learning-based fault detection model that can be implemented in the field of plastic injection molding. Compared to conventional approaches to fault detection in this domain, recent deep learning approaches prove useful for on-site problems involving complex underlying dynamics with a large number of variables. In addition, the advent of advanced sensors that generate data types in multiple modalities prompts the need for multimodal learning with deep neural networks to detect faults. This process is able to facilitate information from various modalities in an end-to-end learning fashion. The proposed deep learning-based approach opts for an early fusion scheme, in which the low-level feature representations of modalities are combined. A case study involving real-world data, obtained from a car parts company and related to a car window side molding process, validates that the proposed model outperforms late fusion methods and conventional models in solving the problem.","{'model': 'tldr@v2.0.0', 'text': 'A case study involving real-world data, obtained from a car parts company and related to a car window side molding process, validates that the proposed model outperforms late fusion methods and conventional models in solving the problem.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09548039.pdf
-"Introducing WESAD, a Multimodal Dataset for Wearable Stress and Affect Detection",P. Schmidt,"Affect recognition aims to detect a person's affective state based on observables, with the goal to e.g. improve human-computer interaction. Long-term stress is known to have severe implications on wellbeing, which call for continuous and automated stress monitoring systems. However, the affective computing community lacks commonly used standard datasets for wearable stress detection which a) provide multimodal high-quality data, and b) include multiple affective states. Therefore, we introduce WESAD, a new publicly available dataset for wearable stress and affect detection. This multimodal dataset features physiological and motion data, recorded from both a wrist- and a chest-worn device, of 15 subjects during a lab study. The following sensor modalities are included: blood volume pulse, electrocardiogram, electrodermal activity, electromyogram, respiration, body temperature, and three-axis acceleration. Moreover, the dataset bridges the gap between previous lab studies on stress and emotions, by containing three different affective states (neutral, stress, amusement). In addition, self-reports of the subjects, which were obtained using several established questionnaires, are contained in the dataset. Furthermore, a benchmark is created on the dataset, using well-known features and standard machine learning methods. Considering the three-class classification problem ( baseline vs. stress vs. amusement ), we achieved classification accuracies of up to 80%,. In the binary case ( stress vs. non-stress ), accuracies of up to 93%, were reached. Finally, we provide a detailed analysis and comparison of the two device locations ( chest vs. wrist ) as well as the different sensor modalities.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces WESAD, a new publicly available dataset for wearable stress and affect detection that bridges the gap between previous lab studies on stress and emotions, by containing three different affective states (neutral, stress, amusement).'}",
-Improvement of Autonomous Vehicles Trust Through Synesthetic-Based Multimodal Interaction,Xiaofeng Sun,"Trust is the key factor for people to accept autonomous vehicles(AVs). Existing studies have reported that multimodal interaction would enhance people’s trust in AVs. However, these researches mainly focus on the superposition effect between sensory channels, and lack on the research of correlation between different sensory channels and its influence on AVs trust. Therefore, we innovatively introduce synesthesia theory for the research of improving AVs trust. We present an AVs multimodal interaction model based on audio-visual synesthesia theory, and finally prove that the model has a definite effect on improving AVs trust by experiments. Firstly, 82 participants are recruited and assigned into two groups: Group A (non-synesthesia group) and Group B (synesthesia group). They conduct an experimental driving experienced normal traffic conditions (NTC) (turning, traffic lights, over and limit speed prompts) and emergency traffic condition (ETC) (sudden braking of the car in front, temporary lane change, pedestrian thrusting) while completing a secondary task. Then, we conduct a survey (questionnaire and interviews) to evaluate the attitude about trust, technical competence, situation management and perceived ease of use after participants finished experimental driving. The results demonstrate that synesthetic-based multimodal interaction (SBMI) can more effectively remind people of relevant information especially under ETC. SBMI model is more effective than single information stimulus or non-synesthetic audio-visual information stimulus not only in terms of information transmission efficiency and effect, but also in terms of output response/ action. The results also show that SBMI contributes to the improvement of AVs trust. These findings provide evidence on the importance of SBMI to the improvement of AVs trust. The findings of this study will be helpful to the future design of AVs interaction system.","{'model': 'tldr@v2.0.0', 'text': 'An AVs multimodal interaction model based on audio-visual synesthesia theory is presented, and it is proved that the model has a definite effect on improving AVs trust by experiments, and demonstrates that synesthetic-based multi-modal interaction (SBMI) can more effectively remind people of relevant information especially under ETC.'}",https://ieeexplore.ieee.org/ielx7/6287639/9312710/09353546.pdf
-Multimodal Imaging in Susac Syndrome: A Case Report and Literature Review,S. Bagaglia,"Susac syndrome (SS) is a rare microangiopathy that involves arterioles of the brain, retina, and cochlea. Diagnosis is extremely difficult because of the rarity of the disease and because the signs and symptoms often occur at different times. Multidisciplinary approaches and multimodal images are mandatory for diagnosis and prompt therapy. In this report, we describe a case of SS and the application of multimodal retinal imaging to evaluate the ophthalmologic changes and to confirm diagnosis. Early diagnosis and therapy based on the associations of steroids and immunosuppressants are necessary to limit the sequelae of the disease.","{'model': 'tldr@v2.0.0', 'text': 'A case of SS is described and the application of multimodal retinal imaging is described to evaluate the ophthalmologic changes and to confirm diagnosis.'}",https://www.mdpi.com/1660-4601/18/7/3435/pdf?version=1617936888
-Multimodal Deep Generative Models for Trajectory Prediction: A Conditional Variational Autoencoder Approach,B. Ivanovic,"Human behavior prediction models enable robots to anticipate how humans may react to their actions, and hence are instrumental to devising safe and proactive robot planning algorithms. However, modeling complex interaction dynamics and capturing the possibility of many possible outcomes in such interactive settings is very challenging, which has recently prompted the study of several different approaches. In this work, we provide a self-contained tutorial on a conditional variational autoencoder (CVAE) approach to human behavior prediction which, at its core, can produce a multimodal probability distribution over future human trajectories conditioned on past interactions and candidate robot future actions. Specifically, the goals of this tutorial paper are to review and build a taxonomy of state-of-the-art methods in human behavior prediction, from physics-based to purely data-driven methods, provide a rigorous yet easily accessible description of a data-driven, CVAE-based approach, highlight important design characteristics that make this an attractive model to use in the context of model-based planning for human-robot interactions, and provide important design considerations when using this class of models.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a self-contained tutorial on a conditional variational autoencoder (CVAE) approach to human behavior prediction which, at its core, can produce a multimodal probability distribution over future human trajectories conditioned on past interactions and candidate robot future actions.'}",https://arxiv.org/pdf/2008.03880
-Antlion re-sampling based deep neural network model for classification of imbalanced multimodal stroke dataset,T. G.,,"{'model': 'tldr@v2.0.0', 'text': 'This study focuses on improving the quality of stroke data implementing a rigorous pre-processing technique using a multimodal stroke dataset available in the publicly available Kaggle repository and proves the superiority of proposed model.'}",https://e-space.mmu.ac.uk/626646/7/G2020_Article_AntlionRe-samplingBasedDeepNeu%20e.pdf
-Concussion assessment potentially aided by use of an objective multimodal concussion index,A. Jacquin,"Objective Prompt, accurate, objective assessment of concussion is crucial as delays can lead to increased short and long-term consequences. The purpose of this study was to derive an objective multimodal concussion index (CI) using EEG at its core, to identify concussion, and to assess change over time throughout recovery. Methods Male and female concussed (N = 232) and control (N = 206) subjects 13–25 years were enrolled at 12 US colleges and high schools. Evaluations occurred within 72 h of injury, 5 days post-injury, at return-to-play (RTP), 45 days after RTP (RTP + 45); and included EEG, neurocognitive performance, and standard concussion assessments. Concussed subjects had a witnessed head impact, were removed from play for ≥ 5 days using site guidelines, and were divided into those with RTP < 14 or ≥14 days. Part 1 describes the derivation and efficacy of the machine learning derived classifier as a marker of concussion. Part 2 describes significance of differences in CI between groups at each time point and within each group across time points. Results Sensitivity = 84.9%, specificity = 76.0%, and AUC = 0.89 were obtained on a test Hold-Out group representing 20% of the total dataset. EEG features reflecting connectivity between brain regions contributed most to the CI. CI was stable over time in controls. Significant differences in CI between controls and concussed subjects were found at time of injury, with no significant differences at RTP and RTP + 45. Within the concussed, differences in rate of recovery were seen. Conclusions The CI was shown to have high accuracy as a marker of likelihood of concussion. Stability of CI in controls supports reliable interpretation of CI change in concussed subjects. Objective identification of the presence of concussion and assessment of readiness to return to normal activity can be aided by use of the CI, a rapidly obtained, point of care assessment tool.","{'model': 'tldr@v2.0.0', 'text': 'The CI was shown to have high accuracy as a marker of likelihood of concussion and assessment of readiness to return to normal activity can be aided by use of the CI, a rapidly obtained, point of care assessment tool.'}",
-Acute basilar artery occlusion (BAO): a pictorial review of multimodal imaging findings,Andrés Yesid Vásquez-Codina,,"{'model': 'tldr@v2.0.0', 'text': 'This pictorial essay illustrates the essential role that multimodal imaging plays in the prompt diagnosis, management, and overall outcome of patients with acute BAO.'}",
-Caption Anything: Interactive Image Description with Diverse Multimodal Controls,Teng Wang,"Controllable image captioning is an emerging multimodal topic that aims to describe the image with natural language following human purpose, $\textit{e.g.}$, looking at the specified regions or telling in a particular text style. State-of-the-art methods are trained on annotated pairs of input controls and output captions. However, the scarcity of such well-annotated multimodal data largely limits their usability and scalability for interactive AI systems. Leveraging unimodal instruction-following foundation models is a promising alternative that benefits from broader sources of data. In this paper, we present Caption AnyThing (CAT), a foundation model augmented image captioning framework supporting a wide range of multimodel controls: 1) visual controls, including points, boxes, and trajectories; 2) language controls, such as sentiment, length, language, and factuality. Powered by Segment Anything Model (SAM) and ChatGPT, we unify the visual and language prompts into a modularized framework, enabling the flexible combination between different controls. Extensive case studies demonstrate the user intention alignment capabilities of our framework, shedding light on effective user interaction modeling in vision-language applications. Our code is publicly available at https://github.com/ttengwang/Caption-Anything.","{'model': 'tldr@v2.0.0', 'text': 'Caption AnyThing (CAT) is presented, a foundation model augmented image captioning framework supporting a wide range of multimodel controls: 1) visual controls, including points, boxes, and trajectories; 2) language controls, such as sentiment, length, language, and factuality.'}",https://arxiv.org/pdf/2305.02677
-VIP5: Towards Multimodal Foundation Models for Recommendation,Shijie Geng,"Computer Vision (CV), Natural Language Processing (NLP), and Recommender Systems (RecSys) are three prominent AI applications that have traditionally developed independently, resulting in disparate modeling and engineering methodologies. This has impeded the ability for these fields to directly benefit from each other's advancements. With the recent development of foundation models, large language models have emerged as a potential general-purpose interface for unifying different modalities and problem formulations. In light of this, we propose the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks. This will enable the processing of multiple modalities in a shared architecture for improved recommendations. To achieve this, we introduce multimodal personalized prompts to accommodate multiple modalities under a shared format. Additionally, we propose a parameter-efficient training method for foundation models, which involves freezing the P5 backbone and fine-tuning lightweight adapters, resulting in improved recommendation performance and increased efficiency in terms of training time and memory usage. Code and data of VIP5 are available at https://github.com/jeykigung/VIP5.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks.'}",http://arxiv.org/pdf/2305.14302
-Meta Learning to Bridge Vision and Language Models for Multimodal Few-Shot Learning,Ivona Najdenkoska,"Multimodal few-shot learning is challenging due to the large domain gap between vision and language modalities. Existing methods are trying to communicate visual concepts as prompts to frozen language models, but rely on hand-engineered task induction to reduce the hypothesis space. To make the whole process learnable, we introduce a multimodal meta-learning approach. Specifically, our approach decomposes the training of the model into a set of related multimodal few-shot tasks. We define a meta-mapper network, acting as a meta-learner, to efficiently bridge frozen large-scale vision and language models and leverage their already learned capacity. By updating the learnable parameters only of the meta-mapper, it learns to accrue shared meta-knowledge among these tasks. Thus, it can rapidly adapt to newly presented samples with only a few gradient updates. Importantly, it induces the task in a completely data-driven manner, with no need for a hand-engineered task induction. We evaluate our approach on recently proposed multimodal few-shot benchmarks, measuring how rapidly the model can bind novel visual concepts to words and answer visual questions by observing only a limited set of labeled examples. The experimental results show that our meta-learning approach outperforms the baseline across multiple datasets and various training settings while being computationally more efficient.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the multimodal meta-learning approach outperforms the baseline across multiple datasets and various training settings while being computationally more efficient.'}",http://arxiv.org/pdf/2302.14794
-Parents and Children: Distinguishing Multimodal DeepFakes from Natural Images,Roberto Amoroso,"Recent advancements in diffusion models have enabled the generation of realistic deepfakes by writing textual prompts in natural language. While these models have numerous benefits across various sectors, they have also raised concerns about the potential misuse of fake images and cast new pressures on fake image detection. In this work, we pioneer a systematic study of the authenticity of fake images generated by state-of-the-art diffusion models. Firstly, we conduct a comprehensive study on the performance of contrastive and classification-based visual features. Our analysis demonstrates that fake images share common low-level cues, which render them easily recognizable. Further, we devise a multimodal setting wherein fake images are synthesized by different textual captions, which are used as seeds for a generator. Under this setting, we quantify the performance of fake detection strategies and introduce a contrastive-based disentangling strategy which let us analyze the role of the semantics of textual descriptions and low-level perceptual cues. Finally, we release a new dataset, called COCOFake, containing about 600k images generated from original COCO images.","{'model': 'tldr@v2.0.0', 'text': 'This work pioneer a systematic study of the authenticity of fake images generated by state-of-the-art diffusion models, and conducts a comprehensive study on the performance of contrastive and classification-based visual features.'}",http://arxiv.org/pdf/2304.00500
-Towards Multimodal Computational Humanities. Using CLIP to Analyze Late-Nineteenth Century Magic Lantern Slides,T. Smits,"The introduction of the CLIP model signaled a breakthrough in multimodal deep learning. This paper examines whether CLIP can be fruitfully applied to a (binary) classification task in the Humanities. We focus on a historical collection of late-nineteenth century magic lantern slides from the Lucerna database. Based on the available metadata, we evaluate CLIP’s performance on classifying slide images into ‘exterior’ and ‘interior’ categories. We compare the performance of several textual prompts for CLIP to two conventional mono-modal models (textual and visual) which we train and evaluate on the same stratified set of 5,244 magic lantern slides and their captions. We find that the textual and multimodal models achieve a respectable performance (∼0.80 accuracy) but are still outperformed by a vision model that was fine-tuned to the task (∼0.89). We flag three methodological issues that might arise from the application of CLIP in the (computational) humanities. First, the lack of (need for) labelled data makes it hard to inspect and/or interpret the performance of the model. Second, CLIP’s zero-shot capability only allows for classification tasks to be simulated, which makes it doubtful if standard metrics can be used to compare its performance to text and/or image models. Third, the lack of effective prompt engineering techniques makes the performance of CLIP (highly) unstable.","{'model': 'tldr@v2.0.0', 'text': 'This paper examines whether CLIP can be fruitfully applied to a (binary) classification task in the Humanities, focusing on a historical collection of late-nineteenth century magic lantern slides from the Lucerna database.'}",
-"A Cabinet of Curiosities, a Dwelling Place: Weekly Writing on Instagram as Multimodal Praxis",Rebecca Conklin,"This Instagram “Weekly Writing” assignment is a social-media-based, low-stakes, and longitudinal approach to teaching and experimenting with multimodal composition. Students create an account for the purposes of the class and follow each other. They post three times per week, sometimes freely and sometimes in response to a prompt or challenge. Together, we use the platform and its rich multimodal resources to consider how in-the-moment multimodal composing can spur invention, place the writer in the perpetual position of noticing, and create an archive of experience that holistically communicates beyond the author’s original intention. This article discusses the pedagogical rationale for this approach, along with the issues to consider before adopting and adapting this practice.",,
-N15News: A New Dataset for Multimodal News Classification,Zhen Wang,"Current news datasets merely focus on text features on the news and rarely leverage the feature of images, excluding numerous essential features for news classification. In this paper, we propose a new dataset, N15News, which is generated from New York Times with 15 categories and contains both text and image information in each news. We design a novel multitask multimodal network with different fusion methods, and experiments show multimodal news classification performs better than text-only news classification. Depending on the length of the text, the classification accuracy can be increased by up to 5.8%. Our research reveals the relationship between the performance of a multimodal classifier and its sub-classifiers, and also the possible improvements when applying multimodal in news classification. N15News is shown to have great potential to prompt the multimodal news studies. The way to collect N15News can be in found here.","{'model': 'tldr@v2.0.0', 'text': 'A new dataset, N15News, which is generated from New York Times with 15 categories and contains both text and image information in each news is proposed, and experiments show multimodal news classification performs better than text-only news classification.'}",
-π-Tuning: Transferring Multimodal Foundation Models with Optimal Multi-task Interpolation,Chengyue Wu,"Foundation models have achieved great advances in multi-task learning with a unified interface of unimodal and multimodal tasks. However, the potential of such multi-task learners has not been exploited during transfer learning. In this work, we present a universal parameter-efficient transfer learning method, termed Predict-Interpolate Tuning ($\pi$-Tuning), for vision, language, and vision-language tasks. It aggregates the parameters of lightweight task-specific experts learned from similar tasks to aid the target downstream task. The task similarities are predicted in a unified modality-independent space, yielding a scalable graph to demonstrate task relationships. $\pi$-Tuning has several appealing benefits. First, it flexibly explores both intra- and inter-modal transferability between similar tasks to improve the accuracy and robustness of transfer learning, especially in data-scarce scenarios. Second, it offers a systematical solution for transfer learning with multi-task prediction-and-then-interpolation, compatible with diverse types of parameter-efficient experts, such as prompt and adapter. Third, an extensive study of task-level mutual benefits on 14 unimodal and 6 multimodal datasets shows that $\pi$-Tuning surpasses fine-tuning and other parameter-efficient transfer learning methods both in full-shot and low-shot regimes. The task graph also enables an in-depth interpretable analysis of task transferability across modalities. The code will be available at https://github.com/TencentARC/pi-Tuning.","{'model': 'tldr@v2.0.0', 'text': 'Predict-Interpolate Tuning aggregates the parameters of lightweight task-specific experts learned from similar tasks to aid the target downstream task, and surpasses fine-tuning and other parameter-efficient transfer learning methods both in full-shot and low-shot regimes.'}",http://arxiv.org/pdf/2304.14381
-Fusing Pre-Trained Language Models with Multimodal Prompts through Reinforcement Learning,Youngjae Yu,"Language models are capable of commonsense reasoning: while domain-specific models can learn from explicit knowledge (e.g. commonsense graphs [6] ethical norms [25]), and larger models like GPT-3 [7] mani-fest broad commonsense reasoning capacity. Can their knowledge be extended to multimodal inputs such as images and audio without paired domain data? In this work, we propose ‡ESPER (Extending Sensory PErception with Reinforcement learning) which enables text-only pretrained models to address multimodal tasks such as visual commonsense reasoning. Our key novelty is to use rein-forcement learning to align multimodal inputs to language model generations without direct supervision: for example, our reward optimization relies only on cosine similarity derived from CLIP [52] and requires no additional paired (image, text) data. Experiments demonstrate that ESPER outperforms baselines and prior work on a variety of multimodal text generation tasks ranging from captioning to commonsense reasoning; these include a new benchmark we collect and release, the ESP dataset, which tasks models with generating the text of several different domains for each image. Our code and data are publicly released at https://github.com/JiwanChung/esper.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ‡ESPER (Extending Sensory PErception with Reinforcement learning) which enables text-only pretrained models to address multimodal tasks such as visual commonsense reasoning.'}",
-Multimodal (Non-)Participation and (Dis)Engagement,Christine M. Jacknick,"While subtler forms of engagement exist, the ones that “count” most are often verbal, because teachers notice those contributions more readily and they seem to have greater potential to affect the trajectory of classroom interaction. This chapter prompts researchers and teachers to examine the full range of students’ multimodal actions to characterize their participation and engagement. Students create meaning with their bodies while also speaking, but given the constraints of multiparty interaction, students often display their engagement with the ongoing interaction primarily in an embodied way. In operationalizing the concept of “engagement” as an embodied phenomenon, this chapter shows the crucial importance of the alignment, temporality, and sequentiality of students’ multimodal actions. Finally, this chapter uncovers the moment-by-moment nature of students’ embodied responses, showing how these actions often occur in a wave, potentially obscuring the teacher’s (and analyst’s) sense of how any individual student is participating in classroom interaction.",,
-Deep Learning-Based Acute Ischemic Stroke Lesion Segmentation Method on Multimodal MR Images Using a Few Fully Labeled Subjects,Bin Zhao,"Acute ischemic stroke (AIS) has been a common threat to human health and may lead to severe outcomes without proper and prompt treatment. To precisely diagnose AIS, it is of paramount importance to quantitatively evaluate the AIS lesions. By adopting a convolutional neural network (CNN), many automatic methods for ischemic stroke lesion segmentation on magnetic resonance imaging (MRI) have been proposed. However, most CNN-based methods should be trained on a large amount of fully labeled subjects, and the label annotation is a labor-intensive and time-consuming task. Therefore, in this paper, we propose to use a mixture of many weakly labeled and a few fully labeled subjects to relieve the thirst of fully labeled subjects. In particular, a multifeature map fusion network (MFMF-Network) with two branches is proposed, where hundreds of weakly labeled subjects are used to train the classification branch, and several fully labeled subjects are adopted to tune the segmentation branch. By training on 398 weakly labeled and 5 fully labeled subjects, the proposed method is able to achieve a mean dice coefficient of 0.699 ± 0.128 on a test set with 179 subjects. The lesion-wise and subject-wise metrics are also evaluated, where a lesion-wise F1 score of 0.886 and a subject-wise detection rate of 1 are achieved.","{'model': 'tldr@v2.0.0', 'text': 'A multifeature map fusion network (MFMF-Network) with two branches is proposed, where hundreds of weakly labeled subjects are used to train the classification branch, and several fully labeling subjects are adopted to tune the segmentation branch.'}",https://downloads.hindawi.com/journals/cmmm/2021/3628179.pdf
-Multimodal Garment Designer: Human-Centric Latent Diffusion Models for Fashion Image Editing,Alberto Baldrati,"Fashion illustration is used by designers to communicate their vision and to bring the design idea from conceptualization to realization, showing how clothes interact with the human body. In this context, computer vision can thus be used to improve the fashion design process. Differently from previous works that mainly focused on the virtual try-on of garments, we propose the task of multimodal-conditioned fashion image editing, guiding the generation of human-centric fashion images by following multimodal prompts, such as text, human body poses, and garment sketches. We tackle this problem by proposing a new architecture based on latent diffusion models, an approach that has not been used before in the fashion domain. Given the lack of existing datasets suitable for the task, we also extend two existing fashion datasets, namely Dress Code and VITON-HD, with multimodal annotations collected in a semi-automatic manner. Experimental results on these new datasets demonstrate the effectiveness of our proposal, both in terms of realism and coherence with the given multimodal inputs. Source code and collected multimodal annotations are publicly available at: https://github.com/aimagelab/multimodal-garment-designer.","{'model': 'tldr@v2.0.0', 'text': 'The task of multimodal-conditioned fashion image editing is proposed, guiding the generation of human-centric fashion images by following multi-modal prompts, such as text, human body poses, and garment sketches.'}",https://arxiv.org/pdf/2304.02051
-Multimodal-driven Talking Face Generation via a Unified Diffusion-based Generator,Chao Xu,"Multimodal-driven talking face generation refers to animating a portrait with the given pose, expression, and gaze transferred from the driving image and video, or estimated from the text and audio. However, existing methods ignore the potential of text modal, and their generators mainly follow the source-oriented feature rearrange paradigm coupled with unstable GAN frameworks. In this work, we first represent the emotion in the text prompt, which could inherit rich semantics from the CLIP, allowing flexible and generalized emotion control. We further reorganize these tasks as the target-oriented texture transfer and adopt the Diffusion Models. More specifically, given a textured face as the source and the rendered face projected from the desired 3DMM coefficients as the target, our proposed Texture-Geometry-aware Diffusion Model decomposes the complex transfer problem into multi-conditional denoising process, where a Texture Attention-based module accurately models the correspondences between appearance and geometry cues contained in source and target conditions, and incorporate extra implicit information for high-fidelity talking face generation. Additionally, TGDM can be gracefully tailored for face swapping. We derive a novel paradigm free of unstable seesaw-style optimization, resulting in simple, stable, and effective training and inference schemes. Extensive experiments demonstrate the superiority of our method.","{'model': 'tldr@v2.0.0', 'text': 'This work derives a novel paradigm free of unstable seesaw-style optimization, resulting in simple, stable, and effective training and inference schemes, and adopts the Diffusion Models.'}",http://arxiv.org/pdf/2305.02594
-D2TV: Dual Knowledge Distillation and Target-oriented Vision Modeling for Many-to-Many Multimodal Summarization,Yunlong Liang,"Many-to-many multimodal summarization (M$^3$S) task aims to generate summaries in any language with document inputs in any language and the corresponding image sequence, which essentially comprises multimodal monolingual summarization (MMS) and multimodal cross-lingual summarization (MXLS) tasks. Although much work has been devoted to either MMS or MXLS and has obtained increasing attention in recent years, little research pays attention to the M$^3$S task. Besides, existing studies mainly focus on 1) utilizing MMS to enhance MXLS via knowledge distillation without considering the performance of MMS or 2) improving MMS models by filtering summary-unrelated visual features with implicit learning or explicitly complex training objectives. In this paper, we first introduce a general and practical task, i.e., M$^3$S. Further, we propose a dual knowledge distillation and target-oriented vision modeling framework for the M$^3$S task. Specifically, the dual knowledge distillation method guarantees that the knowledge of MMS and MXLS can be transferred to each other and thus mutually prompt both of them. To offer target-oriented visual features, a simple yet effective target-oriented contrastive objective is designed and responsible for discarding needless visual information. Extensive experiments on the many-to-many setting show the effectiveness of the proposed approach. Additionally, we will contribute a many-to-many multimodal summarization (M$^3$Sum) dataset.","{'model': 'tldr@v2.0.0', 'text': 'The dual knowledge distillation method guarantees that the knowledge of MMS and MXLS can be transferred to each other and thus mutually prompt both of them, and a target-oriented vision modeling framework is proposed for the M$^3$S task.'}",http://arxiv.org/pdf/2305.12767
-Beyond Text-to-Image: Multimodal Prompts to Explore Generative AI,Vivian Liu,"Text-to-image AI systems have proven to have extraordinary generative capacities that have facilitated widespread adoption. However, these systems are primarily text-based, which is a fundamental inversion of what many artists are traditionally used to: having full control over the composition of their work. Prior work has shown that there is great utility in using text prompts and that AI augmented workflows can increase momentum on creative tasks for end users. However, multimodal interactions beyond text need to be further defined, so end users can have rich points of interaction that allow them to truly co-pilot AI-generated content creation. To this end, the goal of my research is to equip creators with workflows that 1) translate abstract design goals into prompts of visual language, 2) structure exploration of design outcomes, and 3) integrate creator contributions into generations.","{'model': 'tldr@v2.0.0', 'text': 'The goal of this research is to equip creators with workflows that translate abstract design goals into prompts of visual language, structure exploration of design outcomes, and integrate creator contributions into generations.'}",
-Mass-Producing Failures of Multimodal Systems with Language Models,Shengbang Tong,"Deployed multimodal systems can fail in ways that evaluators did not anticipate. In order to find these failures before deployment, we introduce MultiMon, a system that automatically identifies systematic failures -- generalizable, natural-language descriptions of patterns of model failures. To uncover systematic failures, MultiMon scrapes a corpus for examples of erroneous agreement: inputs that produce the same output, but should not. It then prompts a language model (e.g., GPT-4) to find systematic patterns of failure and describe them in natural language. We use MultiMon to find 14 systematic failures (e.g.,""ignores quantifiers"") of the CLIP text-encoder, each comprising hundreds of distinct inputs (e.g.,""a shelf with a few/many books""). Because CLIP is the backbone for most state-of-the-art multimodal systems, these inputs produce failures in Midjourney 5.1, DALL-E, VideoFusion, and others. MultiMon can also steer towards failures relevant to specific use cases, such as self-driving cars. We see MultiMon as a step towards evaluation that autonomously explores the long tail of potential system failures. Code for MULTIMON is available at https://github.com/tsb0601/MultiMon.","{'model': 'tldr@v2.0.0', 'text': 'MultiMon is introduced, a system that automatically identifies systematic failures -- generalizable, natural-language descriptions of patterns of model failures -- and is seen as a step towards evaluation that autonomously explores the long tail of potential system failures.'}",http://arxiv.org/pdf/2306.12105
-Open Visual Knowledge Extraction via Relation-Oriented Multimodality Model Prompting,Hejie Cui,"Images contain rich relational knowledge that can help machines understand the world. Existing methods on visual knowledge extraction often rely on the pre-defined format (e.g., sub-verb-obj tuples) or vocabulary (e.g., relation types), restricting the expressiveness of the extracted knowledge. In this work, we take a first exploration to a new paradigm of open visual knowledge extraction. To achieve this, we present OpenVik which consists of an open relational region detector to detect regions potentially containing relational knowledge and a visual knowledge generator that generates format-free knowledge by prompting the large multimodality model with the detected region of interest. We also explore two data enhancement techniques for diversifying the generated format-free visual knowledge. Extensive knowledge quality evaluations highlight the correctness and uniqueness of the extracted open visual knowledge by OpenVik. Moreover, integrating our extracted knowledge across various visual reasoning applications shows consistent improvements, indicating the real-world applicability of OpenVik.","{'model': 'tldr@v2.0.0', 'text': 'This work presents OpenVik, a first exploration to a new paradigm of open visual knowledge extraction which consists of an open relational region detector to detect regions potentially containing relational knowledge and a visual knowledge generator that generates format-free knowledge by prompting the large multimodality model with the detected region of interest.'}",
-A Prompt-based Multimodal Tabular Transformer Encoder For Medical Intervention Duration Estimation,Y. Ruan,"Objective: This study focuses on estimating the duration of medical interventions using electronic health records (EHRs) in clinical decision support. Most existing models were designed for structured tabular data only and often suffer from data corruption problem. The unstructured clinical free-text data that provides valuable insights and is more resistant to data corruption is often overlooked. The objective of this research is to develop a multimodal deep learning framework that integrates different data modalities from EHRs, thereby fully utilizing the predictive capability of EHRs for medical intervention estimation. Materials and Methods: A novel prompt-based tabular transformer encoder framework is proposed for medical intervention duration estimation based on multimodal EHR data. The framework leverages a pre-trained sentence encoder with medical prompts to harmonize language representations of various clinical data modalities, which a tabular transformer encoder is developed to further explore. Results: The developed model demonstrates superior performance compared to the baselines in two EHR datasets. Furthermore, the model exhibits resilience to data corruption in EHRs, with the RMSE curve increasing gradually with higher corruption rates. Discussion: Other than the predictive effectiveness and robustness of the proposed framework, the ablation study highlights the significance of critical components, such as medical prompts, free-text information, and the pre-trained sentence encoder, all contributing to the model's predictive ability. Conclusion: This research presents a promising pathway to enhance medical intervention estimation by incorporating diverse data modalities from language perspective, ultimately bolstering the reliability of deep learning models in clinical care.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based tabular transformer encoder framework is proposed for medical intervention duration estimation based on multimodal EHR data that leverages a pre-trained sentence encoder with medical prompts to harmonize language representations of various clinical data modalities and demonstrates superior performance compared to the baselines in two EHR datasets.'}",
-Stone Needle: A General Multimodal Large-scale Model Framework towards Healthcare,Weihua Liu,"In healthcare, multimodal data is prevalent and requires to be comprehensively analyzed before diagnostic decisions, including medical images, clinical reports, etc. However, current large-scale artificial intelligence models predominantly focus on single-modal cognitive abilities and neglect the integration of multiple modalities. Therefore, we propose Stone Needle, a general multimodal large-scale model framework tailored explicitly for healthcare applications. Stone Needle serves as a comprehensive medical multimodal model foundation, integrating various modalities such as text, images, videos, and audio to surpass the limitations of single-modal systems. Through the framework components of intent analysis, medical foundation models, prompt manager, and medical language module, our architecture can perform multi-modal interaction in multiple rounds of dialogue. Our method is a general multimodal large-scale model framework, integrating diverse modalities and allowing us to tailor for specific tasks. The experimental results demonstrate the superior performance of our method compared to single-modal systems. The fusion of different modalities and the ability to process complex medical information in Stone Needle benefits accurate diagnosis, treatment recommendations, and patient care.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Stone Needle, a general multimodal large-scale model framework tailored explicitly for healthcare applications, integrating diverse modalities and allowing it to tailor for specific tasks.'}",http://arxiv.org/pdf/2306.16034
-AbhiPaw@DravidianLangTech: Multimodal Abusive Language Detection and Sentiment Analysis,Abhinaba Bala,"Detecting abusive language in multimodal videos has become a pressing need in ensuring a safe and inclusive online environment. This paper focuses on addressing this challenge through the development of a novel approach for multimodal abusive language detection in Tamil videos and sentiment analysis for Tamil/Malayalam videos. By leveraging state-of-the-art models such as Multiscale Vision Transformers (MViT) for video analysis, OpenL3 for audio analysis, and the bert-base-multilingual-cased model for textual analysis, our proposed framework integrates visual, auditory, and textual features. Through extensive experiments and evaluations, we demonstrate the effectiveness of our model in accurately detecting abusive content and predicting sentiment categories. The limited availability of effective tools for performing these tasks in Dravidian Languages has prompted a new avenue of research in these domains.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach for multimodal abusive language detection in Tamil videos and sentiment analysis for Tamil/Malayalam videos is developed by leveraging state-of-the-art models such as Multiscale Vision Transformers for video analysis, OpenL3 for audio analysis, and the bert-base-multilingual-cased model for textual analysis.'}",
-A Multimodal Prototypical Approach for Unsupervised Sound Classification,Saksham Singh Kushwaha,"In the context of environmental sound classification, the adaptability of systems is key: which sound classes are interesting depends on the context and the user's needs. Recent advances in text-to-audio retrieval allow for zero-shot audio classification, but performance compared to supervised models remains limited. This work proposes a multimodal prototypical approach that exploits local audio-text embeddings to provide more relevant answers to audio queries, augmenting the adaptability of sound detection in the wild. We do this by first using text to query a nearby community of audio embeddings that best characterize each query sound, and select the group's centroids as our prototypes. Second, we compare unseen audio to these prototypes for classification. We perform multiple ablation studies to understand the impact of the embedding models and prompts. Our unsupervised approach improves upon the zero-shot state-of-the-art in three sound recognition benchmarks by an average of 12%.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a multimodal prototypical approach that exploits local audio-text embeddings to provide more relevant answers to audio queries, augmenting the adaptability of sound detection in the wild.'}",https://arxiv.org/pdf/2306.12300
-FACTIFY3M: A Benchmark for Multimodal Fact Verification with Explainability through 5W Question-Answering,Megha Chakraborty,"Combating disinformation is one of the burning societal crises -- about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disruption in the share market, panic and anxiety in society, and even death during crises. Therefore, disinformation should be identified promptly and, if possible, mitigated. With approximately 3.2 billion images and 720,000 hours of video shared online daily on social media platforms, scalable detection of multimodal disinformation requires efficient fact verification. Despite progress in automatic text-based fact verification (e.g., FEVER, LIAR), the research community lacks substantial effort in multimodal fact verification. To address this gap, we introduce FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering. Salient features of the dataset include: (i) textual claims, (ii) ChatGPT-generated paraphrased claims, (iii) associated images, (iv) stable diffusion-generated additional images (i.e., visual paraphrases), (v) pixel-level image heatmap to foster image-text explainability of the claim, (vi) 5W QA pairs, and (vii) adversarial fake news stories.","{'model': 'tldr@v2.0.0', 'text': 'FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering, is introduced.'}",http://arxiv.org/pdf/2306.05523
-VideoAdviser: Video Knowledge Distillation for Multimodal Transfer Learning,Yanan Wang,"Multimodal transfer learning aims to transform pretrained representations of diverse modalities into a common domain space for effective multimodal fusion. However, conventional systems are typically built on the assumption that all modalities exist, and the lack of modalities always leads to poor inference performance. Furthermore, extracting pretrained embeddings for all modalities is computationally inefficient for inference. In this work, to achieve high efficiency-performance multimodal transfer learning, we propose VideoAdviser, a video knowledge distillation method to transfer multimodal knowledge of video-enhanced prompts from a multimodal fundamental model (teacher) to a specific modal fundamental model (student). With an intuition that the best learning performance comes with professional advisers and smart students, we use a CLIP-based teacher model to provide expressive multimodal knowledge supervision signals to a RoBERTa-based student model via optimizing a step-distillation objective loss—first step: the teacher distills multimodal knowledge of video-enhanced prompts from classification logits to a regression logit—second step: the multimodal knowledge is distilled from the regression logit of the teacher to the student. We evaluate our method in two challenging multimodal tasks: video-level sentiment analysis (MOSI and MOSEI datasets) and audio-visual retrieval (VEGAS dataset). The student (requiring only the text modality as input) achieves an MAE score improvement of up to 12.3% for MOSI and MOSEI. Our method further enhances the state-of-the-art method by 3.4% mAP score for VEGAS without additional computations for inference. These results suggest the strengths of our method for achieving high efficiency-performance multimodal transfer learning.","{'model': 'tldr@v2.0.0', 'text': 'VideoAdviser is proposed, a video knowledge distillation method to transfer multimodal knowledge of video-enhanced prompts from a multi-modalities model (teacher) to a specific modal fundamental model (student) via optimizing a step-distillation objective loss.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10136716.pdf
-Multimodal Prompt Learning in Emotion Recognition Using Context and Audio Information,Eunseo Jeong,"Prompt learning has improved the performance of language models by reducing the gap in language model training methods of pre-training and downstream tasks. However, extending prompt learning in language models pre-trained with unimodal data to multimodal sources is difficult as it requires additional deep-learning layers that cannot be attached. In the natural-language emotion-recognition task, improved emotional classification can be expected when using audio and text to train a model rather than only natural-language text. Audio information, such as voice pitch, tone, and intonation, can give more information that is unavailable in text to predict emotions more effectively. Thus, using both audio and text can enable better emotion prediction in speech emotion-recognition models compared to semantic information alone. In this paper, in contrast to existing studies that use multimodal data with an additional layer, we propose a method for improving the performance of speech emotion recognition using multimodal prompt learning with text-based pre-trained models. The proposed method is using text and audio information in prompt learning by employing a language model pre-trained on natural-language text. In addition, we propose a method to improve the emotion-recognition performance of the current utterance using the emotion and contextual information of the previous utterances for prompt learning in speech emotion-recognition tasks. The performance of the proposed method was evaluated using the English multimodal dataset MELD and the Korean multimodal dataset KEMDy20. Experiments using both the proposed methods obtained an accuracy of 87.49%, F1 score of 44.16, and weighted F1 score of 86.28.","{'model': 'tldr@v2.0.0', 'text': 'A method for improving the performance of speech emotion recognition using multimodal prompt learning with text-based pre-trained models and a method to improve the emotion- Recognition performance of the current utterance using the emotion and contextual information of the previous utterances for prompt learning in speech emotion-recognition tasks are proposed.'}",https://www.mdpi.com/2227-7390/11/13/2908/pdf?version=1688017556
-"Telehealthcare and Covid-19: A Noninvasive & Low Cost Invasive, Scalable and Multimodal Real-Time Smartphone Application for Early Diagnosis of SARS-CoV-2 Infection",A. Shams,"The global coronavirus pandemic overwhelmed many health care systems, enforcing lockdown and encouraged work from home to control the spread of the virus and prevent overrunning of hospitalized patients. This prompted a sharp widespread use of telehealth to provide low-risk care for patients. Nevertheless, a continuous mutation into new variants and widespread unavailability of test kits, especially in developing countries, possess the challenge to control future potential waves of infection. In this paper, we propose a novel Smartphone application-based platform for early diagnosis of possible Covid-19 infected patients. The application provides three modes of diagnosis from possible symptoms, cough sound, and specific blood biomarkers. When a user chooses a particular setting and provides the necessary information, it sends the data to a trained machine learning (ML) model deployed in a remote server using the internet. The ML algorithm then predicts the possibility of contracting Covid-19 and sends the feedback to the user. The entire procedure takes place in real-time. Our machine learning models can identify Covid-19 patients with an accuracy of 100%, 95.65%, and 77.59% from blood parameters, cough sound, and symptoms respectively. Moreover, the ML sensitivity for blood and sound is 100%, which indicates correct identification of Covid positive patients. This is significant in limiting the spread of the virus. The multimodality offers multiplex diagnostic methods to better classify possible infectees and together with the instantaneous nature of our technique, demonstrates the power of telehealthcare as an easy and widespread low-cost scalable diagnostic solution for future pandemics.","{'model': 'tldr@v2.0.0', 'text': 'A novel Smartphone application-based platform for early diagnosis of possible Covid-19 infected patients that offers multiplex diagnostic methods to better classify possible infectees and demonstrates the power of telehealthcare as an easy and widespread low-cost scalable diagnostic solution for future pandemics.'}",
-Reversing Rivaroxaban Anticoagulation as Part of a Multimodal Hemostatic Intervention in a Polytrauma Animal Model,Farahnaz Rayatdoost,"Background: Life-threatening bleeding requires prompt reversal of the anticoagulant effects of factor Xa inhibitors. This study investigated the effectiveness of four-factor prothrombin complex concentrate in treating trauma-related hemorrhage with rivaroxaban-anticoagulation in a pig polytrauma model. This study also tested the hypothesis that the combined use of a low dose of prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate could improve its subtherapeutic effects. Methods: Trauma (blunt liver injury and bilateral femur fractures) was induced in 48 anesthetized male pigs after 30 min of rivaroxaban infusion (1 mg/kg). Animals in the first part of the study received prothrombin complex concentrate (12.5, 25, and 50 U/kg). In the second part, animals were treated with 12.5 U/kg prothrombin complex concentrate plus tranexamic acid or plus tranexamic acid and fibrinogen concentrate. The primary endpoint was total blood loss postinjury. The secondary endpoints (panel of coagulation parameters and thrombin generation) were monitored for 240 min posttrauma or until death. Results: The first part of the study showed that blood loss was significantly lower in the 25 U/kg prothrombin complex concentrate (1,541 ± 269 ml) and 50 U/kg prothrombin complex concentrate (1,464 ± 108 ml) compared with control (3,313 ± 634 ml), and 12.5 U/kg prothrombin complex concentrate (2,671 ± 334 ml, all P < 0.0001). In the second part of the study, blood loss was significantly less in the 12.5 U/kg prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate (1,836 ± 556 ml, P < 0.001) compared with 12.5 U/kg prothrombin complex concentrate plus tranexamic acid (2,910 ± 856 ml), and there were no early deaths in the 25 U/kg prothrombin complex concentrate, 50 U/kg prothrombin complex concentrate, and 12.5 U/kg prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate groups. Histopathologic analyses postmortem showed no adverse events. Conclusions: Prothrombin complex concentrate effectively reduced blood loss, restored hemostasis, and balanced thrombin generation. A multimodal hemostatic approach using tranexamic acid plus fibrinogen concentrate enhanced the effect of low doses of prothrombin complex concentrate, potentially reducing the prothrombin complex concentrate doses required for effective bleeding control. In an animal model of rivaroxaban-treated pigs that underwent complex traumatic injury, prothrombin complex concentrates alone and in combination with tranexamic acid and fibrinogen concentrate effectively reduced blood loss, restored hemostasis, and improved thrombin generation. Supplemental Digital Content is available in the text.","{'model': 'tldr@v2.0.0', 'text': 'Prothrombin complex concentrate effectively reduced blood loss, restored hemostasis, and balanced thrombin generation, and a multimodal hemostatic approach using tranexamic acid plus fibrinogen concentrate enhanced the effect of low doses of prothromin complex concentrate.'}",https://pubs.asahq.org/anesthesiology/article-pdf/135/4/673/524144/20211000.0-00027.pdf
-MPT: Multimodal Prompt Tuning for Event Detection,,"Event Detection is a key and challenging sub001 task of event extraction, which has serious trig002 ger word ambiguity. Existing studies mainly 003 focus on contextual information in text, while 004 there are naturally many images in news ar005 ticles that need to be explored. We believe 006 that images not only reflect the core events of 007 the text but also help to trigger word disam008 biguation. In this paper, we propose a new 009 bi-recursive multimodal Prompt Tuning (MPT) 010 model for deep interaction between images and 011 sentences to achieve aggregation of modal fea012 tures. MPT uses pre-trained CLIP to encode 013 and map sentences and images into the same 014 multimodal semantic space and uses alternat015 ing dual attention to select information features 016 for mutual enhancement. Then, a soft prompt 017 method of multimodal guidance is proposed, 018 and the multimodal information obtained by 019 fusion is used to guide the downstream event 020 detection task. Our superior performance com021 pared to six state-of-the-art baselines and fur022 ther ablation studies, demonstrate the impor023 tance of image modality and the effectiveness 024 of the proposed architecture. 025","{'model': 'tldr@v2.0.0', 'text': 'A new 009 bi-recursive multimodal Prompt Tuning (MPT) model for deep interaction between images and 011 sentences to achieve aggregation of modal fea012 tures and demonstrates the impor023 tance of image modality and the effectiveness of the proposed architecture.'}",
-What Matters in Training a GPT4-Style Language Model with Multimodal Inputs?,Yan Zeng,"Recent advancements in Large Language Models (LLMs) such as GPT4 have displayed exceptional multi-modal capabilities in following open-ended instructions given images. However, the performance of these models heavily relies on design choices such as network structures, training data, and training strategies, and these choices have not been extensively discussed in the literature, making it difficult to quantify progress in this field. To address this issue, this paper presents a systematic and comprehensive study, quantitatively and qualitatively, on training such models. We implement over 20 variants with controlled settings. Concretely, for network structures, we compare different LLM backbones and model designs. For training data, we investigate the impact of data and sampling strategies. For instructions, we explore the influence of diversified prompts on the instruction-following ability of the trained models. For benchmarks, we contribute the first, to our best knowledge, comprehensive evaluation set including both image and video tasks through crowd-sourcing. Based on our findings, we present Lynx, which performs the most accurate multi-modal understanding while keeping the best multi-modal generation ability compared to existing open-sourced GPT4-style models.","{'model': 'tldr@v2.0.0', 'text': 'Lynx is presented, which performs the most accurate multi-modal understanding while keeping the best multi- modal generation ability compared to existing open-sourced GPT4-style models.'}",https://arxiv.org/pdf/2307.02469
-Multimodal Prompt Retrieval for Generative Visual Question Answering,Timothy Ossowski,"Recent years have witnessed impressive results of pre-trained vision-language models on knowledge-intensive tasks such as visual question answering (VQA). Despite the recent advances in VQA, existing methods mainly adopt a discriminative formulation that predicts answers within a pre-defined label set, leading to easy overfitting on low-resource domains with limited labeled data (e.g., medicine) and poor generalization under domain shift to another dataset. To tackle this limitation, we propose a novel generative model enhanced by multimodal prompt retrieval (MPR) that integrates retrieved prompts and multimodal features to generate answers in free text. Our generative model enables rapid zero-shot dataset adaptation to unseen data distributions and open-set answer labels across datasets. Our experiments on medical VQA tasks show that MPR outperforms its non-retrieval counterpart by up to 30% accuracy points in a few-shot domain adaptation setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel generative model enhanced by multi-modal prompt retrieval (MPR) that integrates retrieved prompts and multimodal features to generate answers in free text that enables rapid zero-shot dataset adaptation to unseen data distributions and open-set answer labels across datasets.'}",http://arxiv.org/pdf/2306.17675
-Multimodal prompts effectively elicit robot-initiated social touch interactions,Spatika Sampath Gujran,"Social touch plays an important role in building interpersonal relationships and might therefore also facilitate interactions with social robots. As people tend to have less experience interacting with social robots compared to with humans, especially with interactions involving social touch, more explicit communication might be necessary to disambiguate social intentions. In the experiment, participants engaged in an informal conversation with humanoid robot Pepper. Throughout the interaction, Pepper initiated various social touch interactions such as a handshake during introductions and a hug to say goodbye by using either a unimodal prompt (control condition: movement cue only) or a multimodal prompt (experimental condition: movement and verbal cue). The results show that the multimodal prompts significantly increased the number of successfully elicited social touch interactions. No significant differences in the self-reported perception of the robot were found between condition. Our results help to inform the design of robots that are intended to engage in social touch interactions.",,https://dl.acm.org/doi/pdf/10.1145/3610661.3617642
-Multimodal Prompt Transformer with Hybrid Contrastive Learning for Emotion Recognition in Conversation,Shihao Zou,"Emotion Recognition in Conversation (ERC) plays an important role in driving the development of human-machine interaction. Emotions can exist in multiple modalities, and multimodal ERC mainly faces two problems: (1) the noise problem in the cross-modal information fusion process, and (2) the prediction problem of less sample emotion labels that are semantically similar but different categories. To address these issues and fully utilize the features of each modality, we adopted the following strategies: first, deep emotion cues extraction was performed on modalities with strong representation ability, and feature filters were designed as multimodal prompt information for modalities with weak representation ability. Then, we designed a Multimodal Prompt Transformer (MPT) to perform cross-modal information fusion. MPT embeds multimodal fusion information into each attention layer of the Transformer, allowing prompt information to participate in encoding textual features and being fused with multi-level textual information to obtain better multimodal fusion features. Finally, we used the Hybrid Contrastive Learning (HCL) strategy to optimize the model's ability to handle labels with few samples. This strategy uses unsupervised contrastive learning to improve the representation ability of multimodal fusion and supervised contrastive learning to mine the information of labels with few samples. Experimental results show that our proposed model outperforms state-of-the-art models in ERC on two benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': ""The Hybrid Contrastive Learning (HCL) strategy is used to optimize the model's ability to handle labels with few samples, and shows that the proposed model outperforms state-of-the-art models in ERC on two benchmark datasets.""}",https://dl.acm.org/doi/pdf/10.1145/3581783.3611805
-Towards Interpretable Multimodal Predictive Models for Early Mortality Prediction of Hemorrhagic Stroke Patients.,F. B. Emdad,"The increasing death rate over the past eight years due to stroke has prompted clinicians to look for data-driven decision aids. Recently, deep-learning-based prediction models trained with fine-grained electronic health record (EHR) data have shown superior promise for health outcome prediction. However, the use of EHR-based deep learning models for hemorrhagic stroke outcome prediction has not been extensively explored. This paper proposes an ensemble deep learning framework to predict early mortality among ICU patients with hemorrhagic stroke. The proposed ensemble model achieved an accuracy of 83%, which was higher than the fusion model and other baseline models (logistic regression, decision tree, random forest, and XGBoost). Moreover, we used SHAP values for interpretation of the ensemble model to identify important features for the prediction. In addition, this paper follows the MINIMAR (MINimum Information for Medical AI Reporting) standard, presenting an important step towards building trust among the AI system and clinicians.","{'model': 'tldr@v2.0.0', 'text': 'An ensemble deep learning framework to predict early mortality among ICU patients with hemorrhagic stroke is proposed, and follows the MINIMAR (MINimum Information for Medical AI Reporting) standard, presenting an important step towards building trust among the AI system and clinicians.'}",
-Translating and Segmenting Multimodal Medical Volumes with Cycle- and Shape-Consistency Generative Adversarial Network,Zizhao Zhang,"Synthesized medical images have several important applications, e.g., as an intermedium in cross-modality image registration and as supplementary training samples to boost the generalization capability of a classifier. Especially, synthesized computed tomography (CT) data can provide X-ray attenuation map for radiation therapy planning. In this work, we propose a generic cross-modality synthesis approach with the following targets: 1) synthesizing realistic looking 3D images using unpaired training data, 2) ensuring consistent anatomical structures, which could be changed by geometric distortion in cross-modality synthesis and 3) improving volume segmentation by using synthetic data for modalities with limited training samples. We show that these goals can be achieved with an end-to-end 3D convolutional neural network (CNN) composed of mutually-beneficial generators and segmentors for image synthesis and segmentation tasks. The generators are trained with an adversarial loss, a cycle-consistency loss, and also a shape-consistency loss, which is supervised by segmentors, to reduce the geometric distortion. From the segmentation view, the segmentors are boosted by synthetic data from generators in an online manner. Generators and segmentors prompt each other alternatively in an end-to-end training fashion. With extensive experiments on a dataset including a total of 4,496 CT and magnetic resonance imaging (MRI) cardiovascular volumes, we show both tasks are beneficial to each other and coupling these two tasks results in better performance than solving them exclusively.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a generic cross-modality synthesis approach and shows that these goals can be achieved with an end-to-end 3D convolutional neural network (CNN) composed of mutually-beneficial generators and segmentors for image synthesis and segmentation tasks.'}",https://arxiv.org/pdf/1802.09655
-"Multilingual, Multimodal, and Multidisciplinary: Deaf Students and Translanguaging in Content Area Classes",Jessica A. Scott,"Translanguaging is an approach to the education of multilingual students that has been gaining in popularity in recent years. Translanguaging is defined as the use of multiple languages and modalities during classroom instruction with the express purpose of encouraging and teaching students to use all their linguistic resources to make sense of content and demonstrate their understanding. Though there has been a great deal of research on translanguaging in general, this article focuses on two less-studied areas for translanguaging: the intersection of translanguaging in deaf education and content area education. Here, the authors explore the literature available in the domains of translanguaging and science, translanguaging and deaf students, and translanguaging in science classrooms with deaf students. Using the findings from these articles, the authors present four recommendations for teachers who want to engage in translanguaging in the content areas with deaf students: First, teachers can use both signed language and English (or other written languages) during instruction to expose students to essential concepts in multiple languages. Next, teachers can expose students to discipline and subject-specific written and signed words, as well as colloquial meanings of similar written and signed words (e.g., evidence versus fact versus proof in American Sign Language). Additionally, teachers should attempt to be visually oriented while translanguaging. Finally, we recommend that teachers encourage students to respond to questions or prompts using multiple languages and/or modalities. Implications of each recommendation are explored.",,https://www.mdpi.com/2226-471X/8/1/55/pdf?version=1676346473
-Integrating multimodal information in machine learning for classifying acute myocardial infarction,R. Xiao,"Objective. Prompt identification and recognization of myocardial ischemia/infarction (MI) is the most important goal in the management of acute coronary syndrome. The 12-lead electrocardiogram (ECG) is widely used as the initial screening tool for patients with chest pain but its diagnostic accuracy remains limited. There is early evidence that machine learning (ML) algorithms applied to ECG waveforms can improve performance. Most studies are designed to classify MI from healthy controls and thus are limited due to the lack of consideration of ECG abnormalities from other cardiac conditions, leading to false positives. Moreover, clinical information beyond ECG has not yet been well leveraged in existing ML models. Approach. The present study considered downstream clinical implementation scenarios in the initial model design by dichotomizing study recordings from a public large-scale ECG dataset into a MI class and a non-MI class with the inclusion of MI-confounding conditions. Two experiments were conducted to systematically investigate the impact of two important factors entrained in the modeling process, including the duration of ECG, and the value of multimodal information for model training. A novel multimodal deep learning architecture was proposed to learn joint features from both ECG and patient demographics. Main results. The multimodal model achieved better performance than the ECG-only model, with a mean area under the receiver operating characteristic curve of 92.1% and a mean accuracy of 87.4%, which is on par with existing studies despite the increased task difficulty due to the new class definition. By investigation of model explainability, it revealed the contribution of patient information in model performance and clinical concordance of the model’s attention with existing clinical insights. Significance. The findings in this study help guide the development of ML solutions for prompt MI detection and move the models one step closer to real-world clinical applications.","{'model': 'tldr@v2.0.0', 'text': 'The investigation of model explainability revealed the contribution of patient information in model performance and clinical concordance of the model’s attention with existing clinical insights and helped guide the development of ML solutions for prompt MI detection and move the models one step closer to real-world clinical applications.'}",https://iopscience.iop.org/article/10.1088/1361-6579/acc77f/pdf
-Multimodal Parameter-Efficient Few-Shot Class Incremental Learning,Marco D’Alessandro,"Few-Shot Class Incremental Learning (FSCIL) is a challenging continual learning task, where limited training examples are available during several learning sessions. To succeed in this task, it is necessary to avoid over-fitting new classes caused by biased distributions in the few-shot training sets. The general approach to address this issue involves enhancing the representational capability of a pre-defined backbone architecture by adding special modules for backward compatibility with older classes. However, this approach has not yet solved the dilemma of ensuring high classification accuracy over time while reducing the gap between the performance obtained on larger training sets and the smaller ones. In this work, we propose an alternative approach called Continual Parameter-Efficient CLIP (CPE-CLIP) to reduce the loss of information between different learning sessions. Instead of adapting additional modules to address information loss, we leverage the vast knowledge acquired by CLIP in large-scale pre-training and its effectiveness in generalizing to new concepts. Our approach is multimodal and parameter-efficient, relying on learnable prompts for both the language and vision encoders to enable transfer learning across sessions. We also introduce prompt regularization to improve performance and prevent forgetting. Our experimental results demonstrate that CPE-CLIP significantly improves FSCIL performance compared to state-of-the-art proposals while also drastically reducing the number of learnable parameters and training costs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an alternative approach called Continual Parameter-Efficient CLIP (CPE-CLIP) to reduce the loss of information between different learning sessions, and leverages the vast knowledge acquired by CLIP in large-scale pre-training and its effectiveness in generalizing to new concepts.'}",http://arxiv.org/pdf/2303.04751
-Multimodal Data Value Chain (M-DVC): A Conceptual Tool to Support the Development of Multimodal Learning Analytics Solutions,Shashi Kant Shankar,"Multimodal Learning Analytics (MMLA) systems, understood as those that exploit multimodal evidence of learning to better model a learning situation, have not yet spread widely in educational practice. Their inherent technical complexity, and the lack of educational stakeholder involvement in their design, are among the hypothesized reasons for the slow uptake of this emergent field. To aid in the process of stakeholder communication and systematization leading to the specification of MMLA systems, this paper proposes a Multimodal Data Value Chain (M-DVC). This conceptual tool, derived from both the field of Big Data and the needs of MMLA scenarios, has been evaluated in terms of its usefulness for stakeholders, in three authentic case studies of MMLA systems currently under development. The results of our mixed-methods evaluation highlight the usefulness of the M-DVC to elicit unspoken assumptions or unclear data processing steps in the initial stages of development. The evaluation also revealed limitations of the M-DVC in terms of the technical terminology employed, and the need for more detailed contextual information to be included. These limitations also prompt potential improvements for the M-DVC, on the path towards clearer specification and communication within the multi-disciplinary teams needed to build educationally-meaningful MMLA solutions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Multimodal Data Value Chain (M-DVC), a conceptual tool derived from both the field of Big Data and the needs of MMLA scenarios, that has been evaluated in terms of its usefulness for stakeholders, in three authentic case studies of M MLA systems currently under development.'}",https://uvadoc.uva.es/bitstream/10324/43245/1/Multimodal-2019_KantShankar_IEEERita.pdf
-Retrieving Multimodal Prompts for Generative Visual Question Answering,Timothy Ossowski,,,https://aclanthology.org/2023.findings-acl.158.pdf
-Seeing the unseen : euphemism in animated films : a multimodal and critical discourse analysis,Dalia Asseel,"Animated films are contemporary popular cultural products recreating the ‘real’ world and engaging massive worldwide audiences of adults and children. Children as the ostensible viewers of animated films may acquire their cultural and ideological knowledge and beliefs about the world from the representations in animated films. Although during the past decade animated films have increasingly been the focus of attention of researchers across different disciplines, including education, gender, sexuality and literacy, studies tackling the discourse and language of animated films are still in their early stages. More specifically, very few studies have investigated the use of euphemism as a major micro-level linguistic device reflecting macro-level discourse and extending to sociocultural structures. To this end, this thesis examines euphemism constructed through the discourse of animated films by employing the strategies of Critical Discourse Studies (CDS). Moreover, Multimodal Discourse Analysis (MDA) is employed to examine discursive strategies involving visual representations accompanying euphemism and what underpins those strategies, and to shed light on the multimodal relations between the representation of both. Euphemism is frequently associated with the notion of taboo. Consequently, new words or phrases are designated to refer to linguistic taboos as alternatives used by speakers to minimise the threat to the audience’s face as well as to their own. In addition, euphemistic occurrences represent a self-interested version of reality by pushing a topic into the background and highlighting instead a particular view of a topic. Therefore, euphemism is a speaker-oriented tool implying the reaction the speaker intends to prompt in the audience. This study shows that euphemism as a discursive linguistic tool has been used extensively in animated films as a manifestation of the discursive role anthropomorphised characters play to transmit certain ideological and social representations. A data set comprising 176 euphemisms found in four full-length anthropomorphised animated featured films, AAFF, extracted from film scripts and online channels was collected. The study identifies the main types of euphemism used in films, drawing on a framework of types based on Warren (1992), Allan and Burridge (1991) and Crespo (2006). After the main types of euphemism have been identified, the data set is approached from the perspective of taboo and culturally repressed topics, such as sexuality, gender and race. Moreover, discursive strategies adapted from Reisigl and Wodak (2016) are applied in order to identify important categories for the analysis of euphemistic discourse. This approach forms the basis for an in-depth, qualitative analysis of several representative scenes extracted from the films under investigation. First, the analysis focuses on racial euphemisms targeting different races. Then, I analyse sexual euphemisms related to nudity and sexual body parts, as well as conceptual metaphoric representations of sexual euphemisms. Finally, I focus on gendered euphemisms targeting the representation of four female characters in the films. The analysis shows that animated films tend to use euphemism to camouflage various taboo areas and manipulate the viewers’ perceptions, such as those relating to sexuality, gender and race. Nomination strategies are used more frequently with euphemisms referring to race, sex and gender. Metaphors and conceptual metaphors are used more frequently with sexual euphemisms. Visual empowerment strategies of female characters to highlight their sexuality are used more frequently with gendered euphemisms. I argue that while the linguistic element attenuates an ideology by virtue of euphemism’s manipulative nature, the visual element, in turn, highlights and confirms the same ideology, values or stereotypes. Hence, animated films can articulate the ideological and social legitimation or normalisation of a particular view of race, sex or gender through the use of euphemism and visual discursive strategies.",,
-TRAVELLERS' PERCEPTION OF WORSHIP FACILITIES FOR MULTIMODAL USERS OF MRT SBK LINE,Nur Athifah A. Kadir,"The complete opening of the new Mass Rapid Transit System (MRT) Sungai Buloh - Kajang (SBK) of Klang Valley, Malaysia since July 2017 has been supported positively by transit researchers and public transport advocators alike. Multimodal public transport users make trips, usually involving transferring at interchanges. These interchanges are provided with several amenities facilitating each trip leg. This research recorded the perception of users of worship facilities provided at an interchanging node where several rails, bus, taxi, and other paratransit systems meet. This new rail system intersects with the other systems at Muzium Negara station. Using intercept off-board face to face questionnaire survey method, perception of the quality of worship facilities was captured between May and August 2019. Several explanatory variables such as socio economic and trip characteristics were set against the perception to identify factors influencing the perception. Results of the correlational analysis showed that both socio-demographic and trip characteristics influenced perception to various statistical significance degrees. Findings suggested that worship facilities were most utilised by passengers of MRT who relatively lower income, engaged in the public, specifically educational sector. Usage of MRT for commuting within the Klang Valley on monthly basis. The improvements in ventilation systems, prayer hall size and design, floor cleanliness, lighting and mirror design as well as shoe rack/clean area design were most sought by these passengers. As such, these should be the focus of strategies to be promptly adopted by interchanging nodes building managers to increase the quality of worship places provision.",,https://www.planningmalaysia.org/index.php/pmj/article/download/834/620
-A Survey on Psycho-Physiological Analysis & Measurement Methods in Multimodal Systems,M. Z. Baig,"Psycho-physiological analysis has gained greater attention in the last few decades in various fields including multimodal systems. Researchers use psychophysiological feedback devices such as skin conductance (SC), Electroencephalography (EEG) and Electrocardiography (ECG) to detect the affective states of the users during task performance. Psycho-physiological feedback has been successful in detection of the cognitive states of users in human-computer interaction (HCI). Recently, in game studies, psycho-physiological feedback has been used to capture the user experience and the effect of interaction on human psychology. This paper reviews several psycho-physiological, cognitive, and affective assessment studies and focuses on the use of psychophysiological signals in estimating the user’s cognitive and emotional states in multimodal systems. In this paper, we review the measurement techniques and methods that have been used to record psycho-physiological signals as well as the cognitive and emotional states in a variety of conditions. The aim of this review is to conduct a detailed study to identify, describe and analyze the key psycho-physiological parameters that relate to different mental and emotional states in order to provide an insight into key approaches. Furthermore, the advantages and limitations of these approaches are also highlighted in this paper. The findings state that the classification accuracy of >90% has been achieved in classifying emotions with EEG signals. A strong correlation between self-reported data, HCI experience, and psychophysiological data has been observed in a wide range of domains including games, human-robot interaction, mobile interaction, and simulations. An increase in β and γ -band activity have been observed in high intense games and simulations.","{'model': 'tldr@v2.0.0', 'text': 'The aim of this review is to conduct a detailed study to identify, describe and analyze the key psycho-physiological parameters that relate to different mental and emotional states in order to provide an insight into key approaches.'}",https://www.mdpi.com/2414-4088/3/2/37/pdf?version=1561376523
-Multimodal sentiment analysis with word-level fusion and reinforcement learning,Minghai Chen,"With the increasing popularity of video sharing websites such as YouTube and Facebook, multimodal sentiment analysis has received increasing attention from the scientific community. Contrary to previous works in multimodal sentiment analysis which focus on holistic information in speech segments such as bag of words representations and average facial expression intensity, we propose a novel deep architecture for multimodal sentiment analysis that is able to perform modality fusion at the word level. In this paper, we propose the Gated Multimodal Embedding LSTM with Temporal Attention (GME-LSTM(A)) model that is composed of 2 modules. The Gated Multimodal Embedding allows us to alleviate the difficulties of fusion when there are noisy modalities. The LSTM with Temporal Attention can perform word level fusion at a finer fusion resolution between the input modalities and attends to the most important time steps. As a result, the GME-LSTM(A) is able to better model the multimodal structure of speech through time and perform better sentiment comprehension. We demonstrate the effectiveness of this approach on the publicly-available Multimodal Corpus of Sentiment Intensity and Subjectivity Analysis (CMU-MOSI) dataset by achieving state-of-the-art sentiment classification and regression results. Qualitative analysis on our model emphasizes the importance of the Temporal Attention Layer in sentiment prediction because the additional acoustic and visual modalities are noisy. We also demonstrate the effectiveness of the Gated Multimodal Embedding in selectively filtering these noisy modalities out. These results and analysis open new areas in the study of sentiment analysis in human communication and provide new models for multimodal fusion.","{'model': 'tldr@v2.0.0', 'text': 'The Gated Multimodal Embedding LSTM with Temporal Attention model is proposed that is composed of 2 modules and able to perform modality fusion at the word level and is able to better model the multimodal structure of speech through time and perform better sentiment comprehension.'}",https://arxiv.org/pdf/1802.00924
-Towards a Multimodal Model of Cognitive Workload Through Synchronous Optical Brain Imaging and Eye Tracking Measures,E. Isbilir,"Recent advances in neuroimaging technologies have rendered multimodal analysis of operators’ cognitive processes in complex task settings and environments increasingly more practical. In this exploratory study, we utilized optical brain imaging and mobile eye tracking technologies to investigate the behavioral and neurophysiological differences among expert and novice operators while they operated a human-machine interface in normal and adverse conditions. In congruence with related work, we observed that experts tended to have lower prefrontal oxygenation and exhibit gaze patterns that are better aligned with the optimal task sequence with shorter fixation durations as compared to novices. These trends reached statistical significance only in the adverse condition where the operators were prompted with an unexpected error message. Comparisons between hemodynamic and gaze measures before and after the error message indicated that experts’ neurophysiological response to the error involved a systematic increase in bilateral dorsolateral prefrontal cortex (dlPFC) activity accompanied with an increase in fixation durations, which suggests a shift in their attentional state, possibly from routine process execution to problem detection and resolution. The novices’ response was not as strong as that of experts, including a slight increase only in the left dlPFC with a decreasing trend in fixation durations, which is indicative of visual search behavior for possible cues to make sense of the unanticipated situation. A linear discriminant analysis model capitalizing on the covariance structure among hemodynamic and eye movement measures could distinguish experts from novices with 91% accuracy. Despite the small sample size, the performance of the linear discriminant analysis combining eye fixation and dorsolateral oxygenation measures before and after an unexpected event suggests that multimodal approaches may be fruitful for distinguishing novice and expert performance in similar neuroergonomic applications in the field.","{'model': 'tldr@v2.0.0', 'text': 'Investigation of behavioral and neurophysiological differences among expert and novice operators while they operated a human-machine interface in normal and adverse conditions observed that experts tended to have lower prefrontal oxygenation and exhibit gaze patterns that are better aligned with the optimal task sequence with shorter fixation durations as compared to novices.'}",https://www.frontiersin.org/articles/10.3389/fnhum.2019.00375/pdf
-Temporal multimodal fusion for video emotion classification in the wild,Valentin Vielzeuf,"This paper addresses the question of emotion classification. The task consists in predicting emotion labels (taken among a set of possible labels) best describing the emotions contained in short video clips. Building on a standard framework – lying in describing videos by audio and visual features used by a supervised classifier to infer the labels – this paper investigates several novel directions. First of all, improved face descriptors based on 2D and 3D Convolutional Neural Networks are proposed. Second, the paper explores several fusion methods, temporal and multimodal, including a novel hierarchical method combining features and scores. In addition, we carefully reviewed the different stages of the pipeline and designed a CNN architecture adapted to the task; this is important as the size of the training set is small compared to the difficulty of the problem, making generalization difficult. The so-obtained model ranked 4th at the 2017 Emotion in the Wild challenge with the accuracy of 58.8 %.","{'model': 'tldr@v2.0.0', 'text': 'Improved face descriptors based on 2D and 3D Convolutional Neural Networks are proposed and several fusion methods, temporal and multimodal, including a novel hierarchical method combining features and scores are explored.'}",https://arxiv.org/pdf/1709.07200
-Multimodal Signaling in Myrmecophilous Butterflies,L. P. Casacci,"Ant nests and their surrounding territories represent a hoard of trophic resources, as well as of stable and protected environments for many arthropods involved in commensal, mutualistic or parasitic associations. Among these organisms, called myrmecophiles, several are butterflies. Here, we explore the amazing diversity of strategies developed by myrmecophilous butterflies to “cheat” or manipulate ants and to elude the tough defenses of the colony. During oviposition, female butterflies use visual or plant volatile signals to identify the presence of ants, whereas chemical and vibroacoustic cues, either isolated or combined, are used by larvae and pupae to attract, deceive or appease workers. Examples of mimicry and eavesdropping on both intraspecific and interspecific signals are discussed, primarily referring to the obligate-parasitic interactions involving Maculinea butterflies and Myrmica ants. Multimodal communication is crucial to maintaining the strong cohesion and social structure of ant societies, but its corruption is at the base of the evolution and persistence of interspecific associations, which can be beneficial or detrimental for the colony’s fitness. In this framework, the remarkable complexity of signaling could have prompted the evolution of specialized life cycles enhancing the extraordinary butterfly diversity.","{'model': 'tldr@v2.0.0', 'text': 'The amazing diversity of strategies developed by myrmecophilous butterflies to “cheat” or manipulate ants and to elude the tough defenses of the colony are explored.'}",https://www.frontiersin.org/articles/10.3389/fevo.2019.00454/pdf
-Deep multimodal fusion for persuasiveness prediction,Behnaz Nojavanasghari,"Persuasiveness is a high-level personality trait that quantifies the influence a speaker has on the beliefs, attitudes, intentions, motivations, and behavior of the audience. With social multimedia becoming an important channel in propagating ideas and opinions, analyzing persuasiveness is very important. In this work, we use the publicly available Persuasive Opinion Multimedia (POM) dataset to study persuasion. One of the challenges associated with this problem is the limited amount of annotated data. To tackle this challenge, we present a deep multimodal fusion architecture which is able to leverage complementary information from individual modalities for predicting persuasiveness. Our methods show significant improvement in performance over previous approaches.","{'model': 'tldr@v2.0.0', 'text': 'A deep multimodal fusion architecture is presented which is able to leverage complementary information from individual modalities for predicting persuasiveness and shows significant improvement in performance over previous approaches.'}",
-Germline DICER1‐mutant intracranial sarcoma with dual chondroid and spindle cell morphology and pulmonary metastases treated with multimodal therapy,Anirban Das,"To the Editor: Pediatric intracranial sarcomas are rare tumors with poor prognoses.1–3 Mutations in the DICER1 gene on chromosome 4, which encodes a RNase-IIIb endonuclease involved in microRNA processing and thereby controlling gene expression, has recently been reported to be associated with intracranial spindle-cell sarcomas with rhabdomyosarcoma (RMS)-like features.3,4 We describe the multimodal management of a case with germline DICER1mutation and pulmonarymetastases. A 12-year-old female with headache, left hemiparesis, and seizures was referred following partial excision of a 7 × 5.7 cm2 right frontoparietal mass. Her maternal aunt had died at 39 years, with an unknown, recurrent, genitourinary malignancy. Histopathology demonstrated malignant oval-spindle cells streaming out of chondroid foci (Figure 1), staining strongly with desmin, myogenin, myo-D1, with retained INI-1, weak membranous CD99, and negative staining for GFAP, neuron-specific enolase, and synaptophysin. Magnetic resonance imaging (MRI)-spine, cerebrospinal fluid-cytology, serum AFP and BCG, and chest X ray were normal. Clinical and radiological progression within 2 weeks prompted re-exploration, achieving gross total resection (GTR), immediately followed by focal radiotherapy (59.4 Gy, 33 fractions) with adequate margins, and consensus-driven chemotherapy with vincristine, doxorubicin, and cyclophosphamide (VDC), alternating with ifosfamide, cisplatin, and etoposide (ICE). Methylation-based classification demonstrated highest concordance with “central nervous system (CNS)/embryonal RMS,” though this did not reach the diagnostic reference range. Gain of chromosomes 1q and 8 was reported. Next generation sequencing demonstrated a nonsense mutation (c.C1870T,p.R624X) in exon 11, and a missense mutation (c.A5438C,p.E1813A) in exon 25 of theDICER1 gene. Sanger sequencing on blood DNA confirmed an identical germline, nonsense mutation. Co-existence of germline truncatingmutations with somatic missense mutations characterize the cancer-predisposing DICER1 syndrome. Computed tomography-chest revealed bilateral pulmonary nodules (maximum size: 1.3 × 1 cm2). Ultrasound abdomen, pelvis, and thyroid, and detailed ophthalmological evaluation were unremarkable. After 50weeks of combination chemotherapy including five alternating cycles of VDC/ICE followed by vincristine, actinomycin-D, and cyclophosphamide (VAC), MRI-brain revealed no active disease, and fluorodeoxyglucose-positron emission tomography revealed a few tiny, calcified, inoperable, nonavid lesions in the lungs. Six months later, she remains well on recommended surveillance for DICER1 syndrome. F IGURE 1 (A) Hematoxylin-eosin (4×)—the section shows large areas of cartilage tissue (arrowhead) and a small fragment of spindled cells (arrow); the inset shows a higher-power image of cellular cartilage. (B) Hematoxylin-eosin (10×)—the section shows a tumor comprising sheets of malignant oval to spindle cells (arrow) with eosinophilic cytoplasm in a hyalinized stroma","{'model': 'tldr@v2.0.0', 'text': 'The multimodal management of a case with germline DICER1mutation and pulmonarymetastases and co-existence of germline truncatingmutations with somatic missense mutations characterize the cancer-predisposing DICer1 syndrome is described.'}",
-Multimodal intraoperative monitoring during surgical correction of scoliosis to avoid neurologic damage,Tong Yu,"Abstract The purpose of this study was to evaluate the application of multimodal intraoperative monitoring (MIOM) system in patients with congenital scoliosis (CS) and adolescent idiopathic scoliosis (AIS). Twelve patients who underwent posterior surgical correction of scoliosis for CS and AIS from June 2014 to July 2018 were enrolled in this study. During the operation, we monitored the functional status of the spinal cord by MIOM. An abnormal somatosensory evoked potential was defined as a prolonged latency of more than 10% or a peak-to-peak amplitude decline of more than 50% when compared to baseline. An abnormal transcranial motor evoked potential (TcMEP) was defined as a TcMEP amplitude decrease of more than 50%. A normal triggered electromyography response, which presented with the absence of an electrical response on stimulation at 8.2 mA, indicated that the pedicle screw was not in contact with the spinal cord or nerve root. A total of 12 patients underwent MIOM surgery, of which 9 patients with negative MIOM had no significant deterioration of neurological function postoperatively, and exhibited satisfactory surgical correction of scoliosis during follow-ups. However, the remaining 3 patients suffered from MIOM events, 2 patients had normal neurological function, and 1 patient had deteriorated neurological function postoperatively. Using MIOM in CS and AIS surgery could promptly detect iatrogenic neurological injury at the early stage. Therefore, rapid response by appropriate intraoperative interventions can be taken to minimize the injury. Besides, stable MIOM recordings encourage surgeons to correct scoliosis even when the Cobb angle of scoliosis was extremely large.","{'model': 'tldr@v2.0.0', 'text': 'Using MIOM in CS and AIS surgery could promptly detect iatrogenic neurological injury at the early stage, and rapid response by appropriate intraoperative interventions can be taken to minimize the injury.'}",
-Multimodal Local-Global Ranking Fusion for Emotion Recognition,P. Liang,"Emotion recognition is a core research area at the intersection of artificial intelligence and human communication analysis. It is a significant technical challenge since humans display their emotions through complex idiosyncratic combinations of the language, visual and acoustic modalities. In contrast to traditional multimodal fusion techniques, we approach emotion recognition from both direct person-independent and relative person-dependent perspectives. The direct person-independent perspective follows the conventional emotion recognition approach which directly infers absolute emotion labels from observed multimodal features. The relative person-dependent perspective approaches emotion recognition in a relative manner by comparing partial video segments to determine if there was an increase or decrease in emotional intensity. Our proposed model integrates these direct and relative prediction perspectives by dividing the emotion recognition task into three easier subtasks. The first subtask involves a multimodal local ranking of relative emotion intensities between two short segments of a video. The second subtask uses local rankings to infer global relative emotion ranks with a Bayesian ranking algorithm. The third subtask incorporates both direct predictions from observed multimodal behaviors and relative emotion ranks from local-global rankings for final emotion prediction. Our approach displays excellent performance on an audio-visual emotion recognition benchmark and improves over other algorithms for multimodal fusion.","{'model': 'tldr@v2.0.0', 'text': 'This work approaches emotion recognition from both direct person-independent and relative person-dependent perspectives and displays excellent performance on an audio-visual emotion recognition benchmark and improves over other algorithms for multimodal fusion.'}",https://arxiv.org/pdf/1809.04931
-GazeTouchPIN: protecting sensitive data on mobile devices using secure multimodal authentication,M. Khamis,"Although mobile devices provide access to a plethora of sensitive data, most users still only protect them with PINs or patterns, which are vulnerable to side-channel attacks (e.g., shoulder surfing). How-ever, prior research has shown that privacy-aware users are willing to take further steps to protect their private data. We propose GazeTouchPIN, a novel secure authentication scheme for mobile devices that combines gaze and touch input. Our multimodal approach complicates shoulder-surfing attacks by requiring attackers to ob-serve the screen as well as the user’s eyes to and the password. We evaluate the security and usability of GazeTouchPIN in two user studies (N=30). We found that while GazeTouchPIN requires longer entry times, privacy aware users would use it on-demand when feeling observed or when accessing sensitive data. The results show that successful shoulder surfing attack rate drops from 68% to 10.4%when using GazeTouchPIN.","{'model': 'tldr@v2.0.0', 'text': 'GazeTouchPIN is proposed, a novel secure authentication scheme for mobile devices that combines gaze and touch input that complicates shoulder-surfing attacks by requiring attackers to ob-serve the screen as well as the user’s eyes to and the password.'}",https://eprints.gla.ac.uk/170215/1/170215.pdf
-"The Handbook of Multimodal-Multisensor Interfaces: Foundations, User Modeling, and Common Modality Combinations - Volume 1",S. Oviatt,"The Handbook of Multimodal-Multisensor Interfaces provides the first authoritative resource on what has become the dominant paradigm for new computer interfaces-user input involving new media (speech, multi-touch, gestures, writing) embedded in multimodal-multisensor interfaces. These interfaces support smartphones, wearables, in-vehicle, robotic, and many other applications that are now highly competitive commercially. This edited collection is written by international experts and pioneers in the field. It provides a textbook for students, and a reference and technology roadmap for professionals working in this rapidly emerging area. Volume 1 of the handbook presents relevant theory and neuroscience foundations for guiding the development of high-performance systems. Additional chapters discuss approaches to user modeling, interface design that supports user choice, synergistic combination of modalities with sensors, and blending of multimodal input and output. They also highlight an in-depth look at the most common multimodal-multisensor combinations- for example, touch and pen input, haptic and non-speech audio output, and speech co-processed with visible lip movements, gaze, gestures, or pen input. A common theme throughout is support for mobility and individual differences among users-including the world's rapidly growing population of seniors. These handbook chapters provide walk-through examples and video illustrations of different system designs and their interactive use. Common terms are defined, and information on practical resources is provided (e.g., software tools, data resources) for hands-on project work to develop and evaluate multimodal-multisensor systems. In the final chapter, experts exchange views on a timely and controversial challenge topic, and how they believe multimodal-multisensor interfaces should be designed in the future to most effectively advance human performance.","{'model': 'tldr@v2.0.0', 'text': 'The Handbook of Multimodal-Multisensor Interfaces provides the first authoritative resource on what has become the dominant paradigm for new computer interfaces-user input involving new media (speech, multi-touch, gestures, writing) embedded in multimodAL-multisensor interfaces.'}",
-EmoNets: Multimodal deep learning approaches for emotion recognition in video,S. Kahou,,"{'model': 'tldr@v2.0.0', 'text': 'This paper explores multiple methods for the combination of cues from these modalities into one common classifier, which achieves a considerably greater accuracy than predictions from the strongest single-modality classifier.'}",http://arxiv.org/pdf/1503.01800
-The Use of Digital Storytelling (DST) to Help Students in Learning Multimodal Writing,Dewi - Puspitasari,"This article explores the use of DST and the explicit teaching in one of universities in Indonesia on how stu-dents use it to help the learning process. Using a digital story to teach English and Based on multimodal theory, the term of DST has been increasingly used by scholars to illustrate various forms of support of learn-ing to help students learn successfully in a classroom. Despite being widely used in educational context of many countries, DST has received scanty attention from teacher especially in ESP classes. This article specif-ically describes our experience of using DST as a learning aid with students of 18 to 19 years old. In this project they individually created collected the photographs based on their interest related to the specified theme as multimodal text. In the process they utilized two linguistic resources (Bahasa Indonesia and Eng-lish) to help them in understanding the process of creation. Several supports from machine translation and machine pronunciation software were employed during the creation of DST project. The result shows that DST helps students in composing narrative writing by analyzing the visual prompts. This proves that DST is impending to support the writing process as students were engaged during the process.","{'model': 'tldr@v2.0.0', 'text': 'This article explores the use of DST and the explicit teaching in one of universities in Indonesia on how stu-dents use it to help the learning process and shows that DST helps students in composing narrative writing by analyzing the visual prompts.'}",http://jurnalnasional.ump.ac.id/index.php/LEKSIKA/article/download/3803/2285
-EmoReact: a multimodal approach and dataset for recognizing emotional responses in children,Behnaz Nojavanasghari,"Automatic emotion recognition plays a central role in the technologies underlying social robots, affect-sensitive human computer interaction design and affect-aware tutors. Although there has been a considerable amount of research on automatic emotion recognition in adults, emotion recognition in children has been understudied. This problem is more challenging as children tend to fidget and move around more than adults, leading to more self-occlusions and non-frontal head poses. Also, the lack of publicly available datasets for children with annotated emotion labels leads most researchers to focus on adults. In this paper, we introduce a newly collected multimodal emotion dataset of children between the ages of four and fourteen years old. The dataset contains 1102 audio-visual clips annotated for 17 different emotional states: six basic emotions, neutral, valence and nine complex emotions including curiosity, uncertainty and frustration. Our experiments compare unimodal and multimodal emotion recognition baseline models to enable future research on this topic. Finally, we present a detailed analysis of the most indicative behavioral cues for emotion recognition in children.","{'model': 'tldr@v2.0.0', 'text': 'A newly collected multimodal emotion dataset of children between the ages of four and fourteen years old is introduced and a detailed analysis of the most indicative behavioral cues for emotion recognition in children is presented.'}",
-Multimodal Retrieval of Similar Soccer Videos Based on Optimal Combination of Multiple Distance Measures,T. Haruyama,"This paper presents a new multimodal method for retrieval of similar soccer videos based on optimal combination of multiple distance measures. Our method first extracts three types of Convolutional Neural Network-based features focusing the players' actions, the audience's cheers and prompt reports. Then, by applying the optimal distance measure to each feature, we calculate the similarities between a query video and videos in a database. Finally, we realize accurate retrieval of similar soccer videos by integrating these similarities. Experiments on actual soccer videos demonstrate encouraging results.","{'model': 'tldr@v2.0.0', 'text': ""This paper first extracts three types of Convolutional Neural Network-based features focusing the players' actions, the audience's cheers and prompt reports, and calculates the similarities between a query video and videos in a database to realize accurate retrieval of similar soccer videos.""}",
-The Development of Plurilingual Education through Multimodal Narrative Reflection in Teacher Education: A Case Study of a Pre-Service Teacher's Beliefs about Language Education,Míriam Cabré Rocafort,"Résumé:L'auteure utilise la réflexion narrative multimodale pour étudier les croyances relatives à la formation linguistique des enseignants en devenir. La sensibilisation aux croyances acquises au moyen de pratiques réflexives est essentielle à l'amélioration de la formation des enseignants, car la réinterprétation des croyances permet d'envisager de nouveaux cadres de référence en formation linguistique. L'auteure analyse les réflexions narratives multimodales à titre d'outils méthodologiques pour déterminer si la réflexion sur l'expérience linguistique personnelle favorise l'acquisition de stratégies pédagogiques d'enseignement plurilingue. Les résultats de l'étude révèlent que certaines croyances sont très fermement ancrées, alors que d'autres sont davantage susceptibles d'être modifiées, les forces à l'œuvre étant la pression de l'hégémonie du purisme linguistique et des principes du monolinguisme. L'auteure utilise une étude de cas pour illustrer comment la réflexion narrative multimodale dans le cadre d'une approche de la formation des enseignants axée sur le plurilinguisme engendre une transformation des croyances. La réflexion narrative multimodale contribue à ce que la participante renonce à l'idée que les langues doivent être maîtrisées et adopte la notion plus fluide selon laquelle l'apprentissage des langues relève du développement d'un répertoire plurilingue. Ces observations ont des répercussions sur la formation des enseignants au chapitre de l'obtention de changements significatifs liés à la formation linguistique. L'étude démontre qu'il convient de développer un cadre pour la formation plurilingue au niveau universitaire.Abstract:This study utilizes multimodal narrative reflection to examine pre-service teachers' beliefs regarding language education. Building awareness of beliefs acquired by reflective practices is fundamental to the improvement of teacher education because reinterpretation of beliefs allows for the consideration of new frames of reference for language education. As methodological tools, multimodal narrative reflections are analyzed to identify whether reflection on personal linguistic experience favours the acquisition of teaching strategies for plurilingual education. Results show that there are some beliefs that are deeply seated, while other beliefs are more likely to be modified; what we see at work here is the pressure of hegemonic linguistic purism and monolingual principles. This article uses a case study to illustrate how multimodal narrative reflection within a plurilingual approach to teacher education prompted a transformation of beliefs. It helped the participant to move away from the idea that languages have to be mastered, to a more fluid notion that learning languages is about the development of a plurilingual repertoire. These findings have implications for teacher education in terms of achieving meaningful changes related to language education. The study demonstrates that it is necessary to develop a framework for plurilingual education at the university level.",,
-Fluorescence Guided Sentinel Lymph Node Mapping: From Current Molecular Probes to Future Multimodal Nanoprobes.,S. Hameed,"For SLN lymph node biopsy (SLNB), SLN mapping has become a standard of care procedure that can accurately locate the micrometastases disseminated from primary tumor sites to the regional lymph nodes. The broad array of SLN mapping has prompted the development of a wide range of SLN tracers, rationally designed for noninvasive and high-resolution imaging of SLNs. At present, conventional SLN imaging probes (blue dyes, radiocolloids, and few other small-molecular dyes), although serving the clinical needs, are often associated with major issues such as insufficient accumulation in SLN, short retention time, staining of the surgical field, and other adverse side effects. In a recent advancement, newly designed fluorescent nanoprobes are equipped with novel features that could be of high interest in SLN mapping such as (i) a unique niche that is not met by any other conventional SLN probes, (ii) their adoptable synthesis method, and (ii) excellent sensitivity facilitating high resolution SLN mapping. Most importantly, lots of effort has been devoted for translating the fluorescent nanoprobes into a clinical setup and also imparting the multimodal imaging abilities of nanoprobes for the excellent diagnosis of life-threatening diseases such as cancer. In this review, we will provide a detailed roadmap of the progress of a wide variety of current fluorescent molecular probes and emphasize the future of nanomaterial-based single/multimodal imaging probes that have true potential translation abilities for SLN mapping.","{'model': 'tldr@v2.0.0', 'text': 'A detailed roadmap of the progress of a wide variety of current fluorescent molecular probes is provided and the future of nanomaterial-based single/multimodal imaging probes that have true potential translation abilities for SLN mapping are emphasized.'}",
-"Assuming a hybrid logic of reading: comics, psychoanalysis, and the multimodal promise of visual response",David Lewkowich,"ABSTRACT With the increasing educational and institutional legitimacy afforded to multimodal texts, there is a need to further explore the use of the visual and its place in reader response, not only as a textual means to prompt interpretation but also as a form of interpretation itself. In this paper, I look at the multimodal interpretive practices of one adult reader who participated in a study I recently conducted with a number of undergraduate students in teacher education, reading a series of graphic novels that centred on themes of adolescence. I explore this reader’s responses to two texts: Jillian and Mariko Tamaki’s This One Summer, and Lynda Barry’s My Perfect Life. In response to their experiences of reading, I asked this reader to think about her own adolescence and to create a visual representation of a memory that was sparked while reading. I thus proceeded with a methodological assumption that to limit our students to only one mode of response is also to limit their possibilities for textual description and existential understanding. Along with a theory of multimodal literacies, I turn to psychoanalytic theory as a way to describe the potential effects of the unconscious on reading experience, memory, and visual response.",,
-Multitask Instruction-based Prompting for Fallacy Recognition,Tariq Alhindi,"Fallacies are used as seemingly valid arguments to support a position and persuade the audience about its validity. Recognizing fallacies is an intrinsically difficult task both for humans and machines. Moreover, a big challenge for computational models lies in the fact that fallacies are formulated differently across the datasets with differences in the input format (e.g., question-answer pair, sentence with fallacy fragment), genre (e.g., social media, dialogue, news), as well as types and number of fallacies (from 5 to 18 types per dataset). To move towards solving the fallacy recognition task, we approach these differences across datasets as multiple tasks and show how instruction-based prompting in a multitask setup based on the T5 model improves the results against approaches built for a specific dataset such as T5, BERT or GPT-3. We show the ability of this multitask prompting approach to recognize 28 unique fallacies across domains and genres and study the effect of model size and prompt choice by analyzing the per-class (i.e., fallacy type) results. Finally, we analyze the effect of annotation quality on model performance, and the feasibility of complementing this approach with external knowledge.","{'model': 'tldr@v2.0.0', 'text': 'This work shows how instruction-based prompting in a multitask setup based on the T5 model improves the results against approaches built for a specific dataset such as T5, BERT or GPT-3 and examines the effect of model size and prompt choice on model performance.'}",http://arxiv.org/pdf/2301.09992
-Fostering Germane Load Through Self-Explanation Prompting In Calculus Instruction,Cecep Anwar Hadi Firdos Santosa,"The purpose of this research was to investigate the effect of self-explanation prompting to students’ germane load while studying mathematics in the multivariable calculus course. This research employed a quasi-experimental method with matching-only posttest-only control group design. The subject of the research consists of 72 first-year mathematics education undergraduate students. The results indicated that there was no significant difference in students’ germane load between students who implemented worked-example with self-explanation prompting and students who implemented worked-example without self-explanation prompting. However, it was revealed that the students' germane load was categorized high in both classes. It indicates that the worked-example method could foster students' germane load. Nonetheless, these results cannot be evidence that self-explanation prompting is capable to foster students' germane load. However, there is an association between germane load and learning objectives. When students achieve the learning objectives, then its learning method is able to foster the germane load. To assess the learning objectives, the posttest was arranged. The results stated that students who implemented the worked-example method with self-explanation prompting had better test scores than students who implemented the worked-example method without self-explanation prompting. This result was sufficient to provide evidence that the use of worked-example with self-explanation prompting could foster students’ germane load students in the multivariable calculus course.",,
-Using simultaneous prompting and computer-assisted instruction to teach narrative writing skills to students with autism spectrum disorders,J. Schuster,"Despite the importance of written expression to the lives of individuals with autism spectrum disorders (ASD), there is limited research on teaching writing skills to this population. In the current study, the researcher evaluated the effects of simultaneous prompting (SP) and computer-assisted instruction (CAI) on the story writing responses of 5 males with autism, 6 to 10 years of age. A classroom teacher and teacher assistant conducted all probe and instructional sessions. The researcher used a multiple probe across behaviors design to evaluate the efficacy of the intervention. In addition, he used pre-posttest measures to assess the generalization of acquired skills across untrained story templates and different response topographies (i.e., handwriting, verbal). Finally, the researcher evaluated the effects of using SP and CAI during writing instruction on the expressive naming of untrained sight words. The data indicated that SP and CAI were effective in improving the story writing skills of and increasing the number of sight words read by all 5 participants. In addition, all participants demonstrated maintenance and generalization of story writing skills. KEYWORDS: Autism Spectrum Disorders, Writing Instruction, Simultaneous Prompting, Computer-Assisted Instruction, Story Writing.",,
-Rapid transition of traditionally hands-on labs to online instruction in engineering courses,D. May,"ABSTRACT The COVID-19 pandemic forced universities to suspend face-to-face instruction, prompting a rapid transition to online education. As many lab courses transitioned online, this provided a rare window of opportunity to learn about the challenges and affordances that the online lab experiences created for students and instructors. We present results from exploratory educational research that investigated student motivation and self-regulated learning in the online lab environment. We consider two student factors: motivation and self-regulation. The instrument is administered to students (n = 121) at the beginning of the semester and statistically analysed for comparisons between different demographic groups. The results indicated students' major was the only distinguishing factor for their motivation and self-regulation. Students' unfamiliarity with online labs or uncertainty about what to expect in the course contributed to the lower levels of self-regulation. The lack of significant differences between various subgroups was not surprising, as we posit many students entered the virtual lab environment with the same level of online lab experience. We conducted interviews among these respondents to explore the factors in greater detail. Using latent Dirichlet allocation, three main topics that emerged: (1) Learning Compatibility, (2) Questions and Inquiry, and (3) Planning and Coordination.",,https://www.tandfonline.com/doi/pdf/10.1080/03043797.2022.2046707?needAccess=true
-Using response-prompting procedures during small-group direct instruction: outcomes and procedural variations.,Jennifer R. Ledford,"Research was reviewed on small-group instruction for learners with disabilities. The review was conducted for articles published between 1990 and 2010 on the application of small-group direct instruction to teach discrete skills using prompting procedures. A total of 47 articles with 197 participants and 687 replications of effects was located. Small-group instruction was effective for 195 of 197 participants and across variations in implementation and contexts. Implementers were primarily special education personnel, and instruction typically occurred in special education settings. Rigorous designs were used in all studies, and fidelity was assessed in 46 of 47 studies and was uniformly high. Students consistently reached criterion on their own target behaviors, generalized those behaviors, maintained them, and learned the behaviors taught to their peers (when this was measured, which occurred in a majority of the studies). Future research should examine comparisons of procedural variables and promoting social behaviors between group mates.","{'model': 'tldr@v2.0.0', 'text': 'Small-group instruction for learners with disabilities was effective for 195 of 197 participants and across variations in implementation and contexts, and fidelity was assessed in 46 of 47 studies and was uniformly high.'}",
-When do you need Chain-of-Thought Prompting for ChatGPT?,Jiuhai Chen,"Chain-of-Thought (CoT) prompting can effectively elicit complex multi-step reasoning from Large Language Models~(LLMs). For example, by simply adding CoT instruction ``Let's think step-by-step'' to each input query of MultiArith dataset, GPT-3's accuracy can be improved from 17.7\% to 78.7\%. However, it is not clear whether CoT is still effective on more recent instruction finetuned (IFT) LLMs such as ChatGPT. Surprisingly, on ChatGPT, CoT is no longer effective for certain tasks such as arithmetic reasoning while still keeping effective on other reasoning tasks. Moreover, on the former tasks, ChatGPT usually achieves the best performance and can generate CoT even without being instructed to do so. Hence, it is plausible that ChatGPT has already been trained on these tasks with CoT and thus memorized the instruction so it implicitly follows such an instruction when applied to the same queries, even without CoT. Our analysis reflects a potential risk of overfitting/bias toward instructions introduced in IFT, which becomes more common in training LLMs. In addition, it indicates possible leakage of the pretraining recipe, e.g., one can verify whether a dataset and instruction were used in training ChatGPT. Our experiments report new baseline results of ChatGPT on a variety of reasoning tasks and shed novel insights into LLM's profiling, instruction memorization, and pretraining dataset leakage.","{'model': 'tldr@v2.0.0', 'text': 'This analysis reflects a potential risk of overfitting/bias toward instructions introduced in IFT, which becomes more common in training LLMs, and indicates possible leakage of the pretraining recipe, e.g., one can verify whether a dataset and instruction were used in training ChatGPT.'}",http://arxiv.org/pdf/2304.03262
-Teacher implementation profiles for integrating computational thinking into elementary mathematics and science instruction,Kathryn M. Rich,,"{'model': 'tldr@v2.0.0', 'text': 'This work analyzed how eight elementary teachers created opportunities for their students to engage in four CT practices during unplugged mathematics and science activities and grouped teachers into four profiles of implementation according to how they used these three strategies.'}",
-Using Simultaneous Prompting and Computer-Assisted Instruction to Teach Story Writing to Students with Autism,R. Pennington,"In the current study, the researchers evaluated the effects of simultaneous prompting and computerassisted instruction on the story-writing responses of 3 males with autism, 7 to 10 years of age. Classroom teachers conducted all probe and training sessions. The researchers used a multiple baseline across participants design to evaluate the efficacy of the intervention. In addition, they used preposttest measures to assess the generalization of acquired skills across untrained story topics and different response topographies. The data indicated that simultaneous prompting and computer-assisted instruction were effective in improving the story-writing skills of all 3 participants. Two of the participants demonstrated maintenance and generalization of trained responses.","{'model': 'tldr@v2.0.0', 'text': 'The data indicated that simultaneous prompting and computer-assisted instruction were effective in improving the story-writing skills of all 3 participants.'}",
-Effectiveness of Video Prompting Delivered via Augmented Reality for Teaching Transition-Related Math Skills to Adults With Intellectual Disabilities,Ryan O. Kellems,"The purpose of this study was to determine the effectiveness of a video-based instruction packet for teaching math-based vocational skills delivered through augmented reality (AR) to three adults with intellectual disabilities. The dependent variable was the percentage of steps performed correctly to solve each selected type of math problem. The independent variable was the video-based math intervention delivered via AR, which modeled the individual steps for solving three different multistep math problems: (1) adjusting a recipe to serve a different number of people, (2) calculating salary, and (3) calculating unit prices. Visual and statistical analyses demonstrated a functional relationship between the video-based math intervention and an increase in the percentage of steps completed correctly for each type of problem. All three participants showed significant gains immediately after receiving the intervention and maintained the learned skills following withdrawal of the intervention. Implications for practitioners and further research are discussed.",,
-"Giving, Prompting, Making: Aligning Technology and Pedagogy Within TPACK for Social Studies Instruction",Thomas Hammond,"Technological pedagogical content knowledge (now known as technology, pedagogy, and content knowledge, or TPACK) has become a widely referenced conceptual framework within teacher education. It provides a common language to discuss the integration of technology into instruction (Koehler & Mishra, 2008) and builds upon the concepts of pedagogical content knowledge (Shulman, 1987) and teacher as curricular “gatekeeper” (Thornton, 2001a, 2001b). This paper describes a three-part pedagogical model—giving-prompting-making—to explicate the relationship between pedagogy and technology within the social studies classroom. This model is intended to enhance the TPACK framework by providing a clear and intuitive comparison between social studies teachers’ pedagogical aims and their choices with technology. The giving-prompting-making model can be used to guide social studies teacher education students to make the most appropriate use of technology.",,
-A MULTILEVEL ANALYSIS OF THE EFFECT OF PROMPTING SELF-REGULATION IN TECHNOLOGY-DELIVERED INSTRUCTION,Traci Sitzmann,"Two studies were conducted to examine the effect of prompting self-regulation, an intervention designed to improve learning from technology-delivered instruction. In Study 1, trainees who were prompted to self-regulate gradually improved their declarative and procedural knowledge over time, relative to the other conditions, whereas test scores declined over time for trainees who were not prompted to self-regulate. In Study 2, basic performance remained stable over time and strategic performance improved over time for trainees who were prompted to self-regulate, relative to the other conditions, whereas performance declined over time for trainees who were not prompted to self-regulate. Trainees’ cognitive ability moderated the effect of the prompts on basic performance and task-specific self-efficacy moderated the effect of the prompts on strategic performance. Prompting self-regulation resulted in stronger performance gains over time for trainees with higher ability or higher self-efficacy. These results demonstrate prompting self-regulation improved performance over time, relative to the other conditions, in both online, work-related training and laboratory settings. The results are consistent with theory suggesting self-regulation is a dynamic process that has a gradual effect on performance and highlight the importance of using a within-subjects design in self-regulation research.",,https://ecommons.cornell.edu/bitstream/1813/77378/1/WP08_121.pdf
-"Explicit linguistic knowledge is necessary, but not sufficient, for the provision of explicit early literacy instruction",Alison Arrow,,"{'model': 'tldr@v2.0.0', 'text': 'The teachers in the study used predominantly implicit approaches to early reading instruction, with word-level instruction and prompting used only after context, even when teachers with high linguistic knowledge used implicit approaches, suggesting that teacher’s knowledge is not sufficient, on its own, to ensure effective, explicit, word- level instruction to beginning readers.'}",
-"Giving, prompting, making: Framing a conceptual home for TPACK in social studies instruction",Thomas Hammond,,,
-A Comparison of Constant Time Delay and Simultaneous Prompting Within Embedded Instruction on Teaching Leisure Skills to Children With Autism,Onur Kurt,"An adapted alternating-treatments design was used to compare the effectiveness and efficiency of constant time delay and simultaneous prompting procedures within an embedded instruction format on the acquisition of various leisure skills by four preschool students with autism. The results showed that both procedures were effective in promoting the acquisition of the skills and maintaining them over time by three students. Although the impacts of the procedures were evident for these three students, the results were not replicated with the fourth student. Furthermore, mixed results were obtained regarding the efficiency measures. However, the social validity findings of the study were positive overall. On the basis of an evaluation of the findings, implications and future research needs are discussed.",,
-Comparison of Simultaneous Prompting Procedure in 1:1 and Small Group Instruction to Teach Play Skills to Preschool Students with Pervasive Developmental Disorder and Developmental Disabilities.,Gail A. Colozzi,"This study examined and compared the effectiveness of a simultaneous prompting procedure used in both 1:1 and small group instruction to teach pretend play skills to a group of preschool students, three having a diagnosis of pervasive developmental disorder and one having a diagnosis of severe developmental disabilities. The study also assessed acquisition of instructive feedback and observational learning stimuli. A multiple probe design across responses and replicated across students and the two instruction formats assessed and compared effectiveness of the simultaneous prompting procedure. Data indicates that simultaneous prompting procedure resulted in acquisition, maintenance, and some generalization of the target skills and instructive feedback stimuli in both instruction formats with minimal differences. The group format allowed the opportunity for observational learning to occur and data on observational learning accuracy and generalization are reported. Severe and pervasive impairment in communication and social skills are primary diagnostic characteristics for children with Pervasive Developmental Disorders (PDD), which includes Autistic Disorder. These deficits impact other skill areas, such as, self-care and play, and, overall, the ability of these children to interact effectively with their environment. Individualized and errorless instructional procedures that use systematic prompt fading methods are highly recommended and welldocumented for their success in teaching skills to children with PDD and other developmental disabilities (Akmanoglu & Batu, 2004; Akmanoglu-Uludag & Batu, 2005; Bozkurt & Gursel, 2005; Farmer, Gast, Wolery, & Winterling, 1991). Research over the past 14-years on one specific method called simultaneous prompting shows it to be a promising and errorless teaching method that is a more efficient, simpler, and straightforward method for teaching a variety of skills (Morse & Schuster, 2004). Simultaneous Prompting (SP) is a systematic form of antecedent prompt and test procedure; all trials are conducted at zero seconds delay between delivery of the discriminative stimulus and the controlling prompt. Unlike antecedent prompt and test, SP first, always uses a controlling prompt, and second, always conducts probe or test trials prior to instructional sessions, rather than after instruction. Schuster and Griffen (1993) pointed out the need for systematic daily probe trials, since learners are always given the controlling prompt during instruction and thus, never have the opportunity to respond independently. Probe trials are necessary to assess the transfer of stimulus control from the controlling prompt to the discriminative stimulus. Morse and Schuster (2004) conducted an extensive review of research on SP procedure. After analyzing 18 published articles, they concluded that SP procedure has been effective in teaching individuals age preschool through adult, with and without disabilities, discrete and chained tasks. The research to date has high procedural and dependent variable reliThe authors would like to thank special education teachers Joyce Camenker, Sally Ann Watt, Tricia Dimke, and music and movement therapist Mary Beth Trapeni for their contributions to the study. Correspondence concerning this article should be addressed to Gail A. Colozzi, 7 Matthew Drive, Salem, NH 03079-1516. Email: bgaegc@earthlink.net Education and Training in Developmental Disabilities, 2008, 43(2), 226-248 © Division on Developmental Disabilities 226 / Education and Training in Developmental Disabilities-June 2008 ability, as well as, positive measures of maintenance and generalization. They cited a number of suggestions for future research that included use of SP (a) with middle school age participants and participants having profound intellectual disabilities, such as, autism spectrum disorders, (b) in non-school settings and activities, (c) with small and large group instructional arrangements, particularly 4:1 or higher student to teacher ratios, (d) related to examination of instructional efficiency variables, like incidental or observational learning, (e) to teach chained tasks, (f) to explore ways to decrease participant errors during daily probes, (g) implemented by non-certified instructors, such as, paraprofessionals, volunteers, parents, and peer tutors, and (h) in comparison with other prompting procedures, such as, system of least prompts or constant time delay. Tekin and Kircaali-Iftar (2002) compared a 4-second constant time delay procedure to an SP procedure (zero delay) in their study of sibling tutors who were trained to implement the procedures to teach receptive animal identification to developmentally delayed students. They found minimal differences between the two procedures; the constant time delay resulted in more generalization, but simultaneous prompting resulted in more efficiency, in terms of fewer errors and less training time. These factors could, in turn, lead to lower frustration and inappropriate behaviors during teaching. Siblings performed as reliably as adults in this study and led the authors to conclude that SP procedures could easily be implemented by teacher, parents, tutors and other related service personnel. Morse and Schuster (2004) state the need for not only teachers, but also, parents, siblings, paraprofessionals, and peer tutors to be able to use effective methods to teach students with disabilities, given the emphasis on inclusion and the need to provide these students with opportunities to access their natural environment. Tekin-Iftar (2003) demonstrated the ease and effectiveness of SP procedures used by peer tutors to teach expressive identification of community sight words and non-target information (instructive feedback responses) to students with developmental disabilities. Other studies have demonstrated the effectiveness of SP procedures in 1:1 and small group instruction, with students having mild disabilities (Johnson, Schuster, & Bell, 1996) and moderate to severe developmental delays (Fetco, Schuster, Harley, & Collins, 1999; Fickel, Schuster, & Collins, 1998; Gursel, Tekin-Iftar, & Bozkurt, 2006; Maciag, Schuster, Collins, & Cooper, 2000; Parrott, Schuster, Collins, & Gassaway, 2000; Schuster & Griffen, 1993; Singleton, Schuster, & Ault, 1995). Birkan (2005) effectively taught three students, ages preschool, primary, and secondary, with mild to moderate developmental delays, three separate academic tasks of reading sight words, receptively identifying numbers, and telling time in hours, using SP during 1:1 instruction. Students maintained and generalized across people, settings, and materials. SP is currently being studied as a promising teaching method that may combine efficiency and error reduction in skill acquisition, maintenance, and generalization for students having various disabilities from preschool through high school age, in a variety of skill areas, such as, academics, vocational, self-care, and language (Dogan & Tekin-Iftar, 2002; Fetco et al., 1999; Gibson & Schuster, 1992; Morse & Schuster, 2004; Parker & Schuster, 2002; Schuster, Griffen, & Wolery, 1992; Singleton, Schuster, Morse, & Collins, 1999). Several studies have shown acquisition of instructive feedback during SP procedures (Griffen, Schuster, & Morse, 1998; Gursel et al., 2006; Parker & Schuster, 2002; Parrott et al., 2000; Schuster & Griffen, 1993; Singleton et al., 1995; Tekin-Iftar, Acar, & Kurt, 2003; Wolery, Holcombe, Werts, & Cipoloni, 1993). Instructive feedback is a strategy whereby additional, non-target stimuli are presented during consequent events of an instructional trial. Students are not reinforced for or expected to respond to this additional material. Werts, Wolery, Holcombe, and Gast (1995) reviewed 20 studies that showed successful acquisition of instructive feedback. Parker and Schuster demonstrated successful acquisition of targets, observational stimuli, and instructive feedback in their study that used SP with high school students. Likewise, Gursel et al. most recently demonstrated the efficacy of SP during small group instruction of middle school aged students to teach target academic skills and nontarget skills through instructive feedback and observational learning. Comparison of Simultaneous Prompting / 227 SP has been cited as an easy instructional method for teachers, siblings, and peer tutors to implement, since no fading procedures are required and a number of studies have shown a high degree of fidelity with regard to procedural implementation (Akmanoglu & Batu, 2004; Akmanoglu-Uludag & Batu, 2005; Morse & Schuster, 2004; Tekin & KircaaliIftar, 2002; Tekin-Iftar, 2003; Wolery et al., 1993). In most of the SP research, teachers have implemented the programming, rather than paraprofessionals or parents (Morse & Schuster). Two studies used sibling and peer tutors (Tekin & Kircaali-Iftar; Tekin-Iftar). Though there is a growing body of research on using SP to teach a variety of discrete and chained skills, to date, minimal research has been conducted with students diagnosed with PDD (Almanoglu & Batu, 2004; AkmanogluUludag & Batu, 2005; Morse & Schuster, 2004) and none with preschoolers having PDD. SP has been used effectively with the preschoolers having developmental disabilities to teach expressive word recognition (Gibson & Schuster, 1992), expressive object identification (MacFarland-Smith, Schuster, & Stevens, 1993), and self-care skills (Sewall, Collins, Hemmenter, & Schuster, 1998). Wolery (1998) cited the need for systematic teaching methods, including simultaneous prompting procedure, to be used with preschoolers having developmental delays, if they are to acquire skills. Akmanoglu and Batu were among the first to study SP procedure with autistic students aged 6 – 17-years old, who acquired, maintained, and generalized across materials, the skill of pointing to numerals. In a second study, Akmanoglu-Uludag and Batu, showed",,
-Caregiver-Delivered Home-Based Instruction Using Simultaneous Prompting for Teaching Home Skills to Individuals with Developmental Disabilities,Sema Batu,"It is very important for individuals with all kinds of developmental disabilities to learn skills in order to be independent at home. The purposes of the study were twofold; (1) to examine the effectiveness of caregiver-delivered home-based instruction using simultaneous prompting to children with moderate develop- mental disabilities on teaching chained home skills and (2) to find out whether the caregivers would be able to implement simultaneous prompting procedure correctly throughout the study. Moreover, maintenance and generalization of the acquired skills were also evaluated. Participants of the study included four children with moderate mental retardation and their caregivers. After the caregivers were trained to use the simultaneous prompting procedure at the beginning of the study and were expected to teach three different chained-home skills to their children. Results of the study revealed that caregivers were successful in using the simultaneous prompting procedure. The children acquired target skills of the study conducted with multiple exemplars. Results also showed that the children maintained them one and three weeks after the training sessions were completed, and generalized them across trainers. Results of the study were extensively discussed.","{'model': 'tldr@v2.0.0', 'text': 'The purposes of the study were to examine the effectiveness of caregiver-delivered home-based instruction using simultaneous prompting to children with moderate develop- mental disabilities on teaching chained home skills and to find out whether the caregivers would be able to implement simultaneous prompting procedure correctly throughout the study.'}",
-Parent-Delivered Community-Based Instruction with Simultaneous Prompting for Teaching Community Skills to Children with Developmental Disabilities,Elif Tekin‐Iftar,"The present study was designed to determine whether parents (three mothers and one grandmother) could implement CBI with SP reliably for teaching community skills to their children and the effects of parent-delivered intervention on teaching the community skills. Maintenance and generalization effects of the intervention were also analyzed in the study. Lastly, the study was designed to reveal the participants' opinions about the intervention. Four parents and their children participated in the study and mother-child dyad was formed. A multiple probe design across community skills and replicated across children was used. Results showed that all parents implemented the intervention with a high degree of treatment integrity. Parent-delivered intervention was effective on teaching community skills. Participants were able to maintain the acquired community skills over time and generalized the acquired skills to generalization sites. The participants' opinions regarding the social validity aspects of the study were positive in general. Based upon evaluation of the findings and implications of the study, future research needs are discussed.",,
-Using Video Prompting to Teach Mathematical Problem Solving of Real-World Video-Simulation Problems,Alicia F. Saunders,"Mathematical problem solving is necessary in many facets of everyday life, yet little research exists on how to teach students with more severe disabilities higher order mathematics like problem solving. Using a multiple probe across participants design, three middle school students with moderate intellectual disability (ID) were taught to solve video-simulation real-world mathematical problems by finger counting using video prompting (VP) in conjunction with systematic instruction (e.g., least intrusive prompting) with error correction and feedback. The simulated videos covered a variety of thematic units (i.e., pet store, grocery store, household chores, sporting goods store, outside chores, and thrift store) students may encounter in their everyday lives. The results of this study demonstrated a functional relation between VP and participants’ mathematical problem-solving skills. Findings from this study provide several implications for practice and research for using video-based interventions to teach mathematical problem solving to students with moderate ID.",,https://journals.sagepub.com/doi/pdf/10.1177/0741932517717042
-Teaching Spelling Through Prompting and Review Procedures Using Computer-Based Instruction,Kristin H. Mayfield,,,
-Grammar instruction in the Hispanic area : the case of Spain with attention to empirical studies on metalinguistic activity,X. Fontich,"Grammar instruction is an unresolved issue in the Hispanic area, having long been approached from within the disjunction between rhetoric (teaching how to use language, especially writing) and grammar (teaching the grammar content). Over time grammar instruction has generated an intense debate around two positions: direct instruction on grammar content, versus instruction devoted to prompting reflection on grammar and language use. There has been an insistent and recurring tendency towards the former, a situation that still prevails. More recently, however, certain research trends in Spain, albeit a minority, have crystallized in a research trajectory on metalinguistic activity, opening up new possibili-ties for rethinking instruction based on grammar reflection to support writing. Within such a trajectory, innovative ways to promote grammar reflection within language use are explored, and metalinguistic activity becomes the focus of research and pedagogy. Crucially, some of the empirical studies developed within this trajectory suggest that grammar knowledge may not be a condition for reflection about language but its consequence, leading to a consideration of metalinguistic activity as a promising avenue for rethinking the debate on the role of grammar instruction.",,https://l1.publication-archive.com/download/1/4773
-"The Effects of Fading, Modeling, Prompting, and Direct Instruction on Letter Legibility for Two Preschool Students with Physical and Developmental Delays",Christine Park,"ABSTRACT The purpose of this study was to determine the effectiveness of the model, lead, and test procedure, as well as a fading procedure with prompts and Direct Instruction with two preschool developmentally delayed students. These procedures were implemented to teach a class of preschoolers to write their names in preparation for their kindergarten transition. The participants were enrolled in a special education preschool and both were referred to the program due to physical as well as academic delays. Signing their name was a daily task that the students needed to accomplish before they started the day. During baseline both students had difficulty with letter identification and formation. The results of the model, lead and test suggested that these were effective procedures. When these were faded, there was little change in both participants handwriting. The benefits of using Direct Instruction procedures and fading for preschool students with physical development delays at the preschool level were discussed.",,
-Effects of prompting multiple solutions for modelling problems on students’ performance,S. Schukajlow,,,
-A Comparison of Constant Time Delay and Simultaneous Prompting Within Embedded Instruction in General Education Classes with Students with Moderate to Severe Disabilities,Tim Riesen,,,
-Coaching Teachers to Use a Simultaneous Prompting Procedure to Teach Core Content to Students With Autism,Elif Tekin‐Iftar,"The researchers in this study used a multiple baseline design across dyads to examine the effects of professional development with coaching to train general education teachers to use a simultaneous prompting procedure when teaching academic core content to students with autism and the effects of the procedure on the students’ outcomes. Three teacher–student dyads participated in the study. Results showed that (a) teachers acquired the ability to use the simultaneous prompting procedure with 100% accuracy, maintained the acquired teaching behaviors over time, and generalized them in teaching new academic content to their students; and (b) students acquired the targeted academic content, maintained it over time, and generalized it across different persons and settings. In addition, the students acquired instructive feedback stimuli added to instruction and maintained these over time as well. Last, both the opinions of the teachers and students about the social validity of the study were positive. Future research is needed to support these findings.",,
-"Verbal Prompting, Hand-over-Hand Instruction, and Passive Observation in Teaching Children with Developmental Disabilities",G. Biederman,"In hand-over-hand modeling with response-contingent verbal prompts, tasks are divided into identifiable sequential components, and the achievement of each component is marked by the delivery of some form of verbal prompt. In a within-subjects design, children were trained in one skill with response-contingent verbal prompts and in a second skill with simple passive observation. A separate group of children were trained with less rigorous verbal prompting in one skill and with passive observation in a second. Consistent with previous research, we found that passive modeling was overall significantly more effective than hand-over-hand modeling and that passive modeling was significantly more effective than hand-over-hand modeling with response-contingent prompting. Our evidence therefore indicates that current methods for teaching basic skills to children with severe developmental delays may require reassessment because simple observation of modeled skills appears to be more effective than more labor-intensive instruction.",,
-An experimental analysis of proctor prompting behavior in a personalized instruction course.,Kent Johnson,"An Experimental Analysis of Proctor Prompting Behavior in a Personalized Instruction Course June 1977 Kent R. Johnson, E.S., Georgetown University M.S., University of Massachusetts Ph.D., University of Massachusetts Directed by: Professor Beth Sul zer-Azarof f The present study analyzed one aspect of proctor stud ent interactions in Personalized System of Instruction (PSI) quiz scoring sessions. When a student has omitted a quiz answer, or has supplied an unclear, partial, or erroneous quiz answer, the proctor may (a) supply the student with the correct answer and have the student repeat it (informational prompting), (b) provide varying degrees of additional information until tlic student emits the correct response (informational prompting), (c) tell the student where to find the correct answer and try again later (noninformat ional prompting) , or (d) simply mark the answer wrong and say nothing. To determine the strategies that proctors used during quiz scoring, nine proctors in a PSI Introductory Psychology course tapo-i'ccorded tlicir.^quiz scoring sessions throughout the semester. Tape recordings of proctoring sessions during the first three weeks of the course indicnted that proctors used informational prompting strategies approximately SOI. of the time an unclear, omitted, or incorrect quiz answer",,
-Effects of Self-Mediated Video Modeling Compared to Video Self-Prompting for Adolescents with Intellectual Disability.,Sally B. Shepley,"Self-instruction is a pivotal skill that promotes independence and self-determination by allowing individuals to independently access prompts during the acquisition of new skills while decreasing the need for support from another person. Self-mediated video modeling (SMVM) and video self-prompting (VSP) are two methods of video-based self-instruction in which individuals use technology to access visual support while practicing new skills. This study evaluated the efficiency of SMVM versus VSP to teach four high school females with intellectual disability art-related tasks in a school setting. An adapted alternating treatments design was used to compare the effects of SMVM and VSP on sessions to criterion and instructional time to criterion. Results indicated that both SMVM and VSP resulted in an increase in percentage of correct responses across all participants, but VSP tasks resulted in all participants reaching mastery criterion in the fewest sessions to criterion. Self-instruction occurs when an individual uses resources available to learn a skill rather than relying on support or guidance provided directly by another person. The materials used for self-instruction may serve as the prompt to notify the learner of the step that should occur next in a sequence to complete a skill (Browder & Shapiro, 1985). Self-instruction promotes monitoring of one’s own behaviors when completing a task and is a form of selfdirected learning. Teaching individuals how to self-instruct fosters self-determination (Agran & Wehmeyer, 2000). Self-determination is the ability to make decisions on one’s own life and plays a role in the success of an individual after graduating high school (Wehmeyer & Palmer, 2003). Self-instruction is a pivotal skill in that it allows individuals to not only learn one to three target behaviors, but instead, can potentially have collateral effects on several behaviors if the self-instruction skills generalize (Smith, Shepley, Alexander, & Ayres, 2015). For instance, once an individual learns how to self-instruct while learning to make a ham and cheese sandwich, they have acquired two skills (making a sandwich and self-instruction). While making a ham and cheese sandwich is valuable in a particular context, the skill of using a set of procedures to self-instruct creates opportunities to learn new skills. Individuals with intellectual disability (ID) rely on teachers and other postsecondary instructors, such as job coaches, to predict and teach skills needed for obtaining and maintaining employment and for independent living. Teaching each skill in isolation that is required for future employment or community settings is an inefficient use of time and resources. Further, teaching all skills needed to fully integrate into one’s community and live independently is impossible. Instructors and teachers might consider teaching individuals how to self-instruct rather than teaching skills in isolation to maximize instruction time and increase independent skills (Smith et al., 2015). With web-based resources like YouTube, engaging in self-instruction has become more commonplace. For instance, if one needs to change a battery in a car, learn to fix a dishwasher, or even play an instrument, they can find instructional materials for free on YouTube. The challenge instructionally then Correspondence concerning this article should be addressed to Sally B. Shepley, The University of Kentucky, Early Childhood, Special Education, and Rehabilitation Counseling, 229 Taylor Education Building, Lexington, KY 40506. E-mail: sally.shepley@uky.edu Education and Training in Autism and Developmental Disabilities, 2018, 53(3), 264–275 © Division on Autism and Developmental Disabilities 264 / Education and Training in Autism and Developmental Disabilities-September 2018 is to teach individuals with ID to sift through the available instructional resources to identify a prompt corresponding to their needs. Learning to self-instruct has the dual benefit of decreasing reliance on others while allowing a learner to pursue those topics of greatest need and interest to him or herself. Further, self-instruction may result in less stigmatization because rather than a teacher or job coach accompanying an individual into the natural environment, the learner can use mobile technology to self-instruct. Additionally, because self-instruction reduces dependence on instructors, instructors may have more time to dedicate to working with other learners or completing other duties. Smith et al. (2015) reviewed the research on self-instruction strategies for individuals with ID that specifically resulted in generalization to novel skills. The results indicated that 56 of 57 included participants in the review learned at least one multi-step skill with a self-instructional strategy, “meaning the participant independently controlled and manipulated the [self-instruction] materials” (p. 21). The review identified three times within the experimental process in which individuals learned to self-instruct, including the use of history training prior to the baseline condition, teaching between baseline and intervention, or teaching during the intervention itself. Of the 56 individuals with ID that successfully acquired a self-instructional strategy, only 32 generalized this strategy to learn at least one additional multi-step skill. Smith et al. identified potential barriers related to generalization of self-instructional skills in the included studies. One example barrier was 7 of the 19 studies incorporated vocal directions to use the self-instruction tool in their task direction (e.g., “Restock the vending machine. Watch the video on your iPod”). Fundamentally, this means the learners did not have to discriminate on their own that they needed to use their self-instructional tools, rather they were directly coached by an instructor to do so. This, in some ways, defeats the end goal of self-instruction. Another issue that Smith et al. (2015) cited related to a lack of instruction for learners on how to search for and select self-instructional materials. In 18 of the 19 included studies, researchers directly loaded the self-instructional materials to the specific prompt required for task completion (e.g., the audio tape corresponding to target task was loaded in the cassette player or a video model demonstrating the targeted skill was loaded to a mobile device). An alternative would be teaching an individual to navigate through their self-instructional tool to locate a specific prompt among multiple available options. As technology has advanced, selfinstructional tools have the potential to store multiple needed prompts to complete identified targeted skills. The tools used by individuals with ID to self-instruct have evolved since the 1980s when researchers used books and picture prompts to teach a novel set of skills (e.g., Wacker & Berg, 1983, 1984). Self-instruction later incorporated more high-tech tools, such as cassette players with headphones to provide auditory prompting in the late 1990s (Trask-Tyler, Grossi, & Heward, 1994), and portable DVD players to promote self-meditated video modeling (SMVM) in the early 2000s (Mechling, Gast, & Fields, 2008; Mechling & Stephens, 2009). Beginning in 2009, handheld devices became more prevalent on the technology market; therefore, the tools used in self-instructional research followed suit. Mechling, Gast, and Seid (2009) taught participants SMVM using a Hewlett Packard iPAQ Pocket PC to complete cooking recipes. The first-generation Apple iPhone was released in 2007 (Apple Inc., 2007) and Bereznak, Ayres, Mechling, and Alexander (2012) first used it as a self-instructional tool for learners with ID and autism spectrum disorder (ASD) to complete vocational and independent living tasks. Additional research has focused on video prompting self-instruction, or video selfprompting (VSP), in which a video of a task is separated into steps so that the individual views and then imitates a single step or a few steps of the task at a time, as opposed to the entire task in video modeling. Bereznak et al. (2012) evaluated the use VSP and taught three high school students how to pause and play video prompts of vocational and daily living skills such as using a washing machine, making noodles, and using a copy machine. After each step in the task analysis, the video displayed a stop sign for 4 s signaling to the participant to pause the video and complete the demonstrated step. Two of the three parSelf-Instruction Video Comparison / 265 ticipants learned to pause and play the video prompts and reached mastery criterion of the various vocational and daily living tasks. Shepley, Spriggs, Samudre, and Elliott (2017) taught four middle school students with ID to self-instruct using a similar VSP format. The researchers created videos with embedded pause signs [similar to Bereznak et al.’s (2012) stop signs] to cue the participants to pause the video, complete the demonstrated step(s), and resume play to view additional video prompts. All participants learned to navigate the technology to find the necessary videos and pause/play videos using a system of least prompts instruction. After completing technology training, three of the four participants were able to self-instruct using video prompts to acquire a novel daily living task sequence (i.e., set the table, make noodles, and make a cup of punch). These studies provided support for video prompting as an effective selfinstructional tool for learners of various ages with ID. As mentioned above and represented in the published video-based instruction literature, videos can take one of two forms, video modeling and video prompting. Both methods have led to skill acquisition, yet it is important to assess efficiency of instruction to ensure instructional time is spent wisely, thus maximize outcomes for learners with ID. Mechling, Ayres, Bryant, and Foster (2014) compared VM and ",,
-When procedures discourage insight: epistemological consequences of prompting novice physics students to construct force diagrams,E. Kuo,"ABSTRACT One aim of school science instruction is to help students become adaptive problem solvers. Though successful at structuring novice problem solving, step-by-step problem-solving frameworks may also constrain students’ thinking. This study utilises a paradigm established by Heckler [(2010). Some consequences of prompting novice physics students to construct force diagrams. International Journal of Science Education, 32(14), 1829–1851] to test how cuing the first step in a standard framework affects undergraduate students’ approaches and evaluation of solutions in physics problem solving. Specifically, prompting the construction of a standard diagram before problem solving increases the use of standard procedures, decreasing the use of a conceptual shortcut. Providing a diagram prompt also lowers students’ ratings of informal approaches to similar problems. These results suggest that reminding students to follow typical problem-solving frameworks limits their views of what counts as good problem solving.",,
-Helping Students Think about Public Issues: Instruction versus Prompting.,W. Parker,,,
-Key Reading Recovery Strategies to Support Classroom Guided Reading Instruction.,J. Lipp,"Effective teachers are continuously striving to improve their instruction. Reading Recovery teachers have detailed and specific literacy training and expertise that can be of great value to classroom teachers, especially in the area of guided reading instruction. This article explores the ways in which key Reading Recovery strategies can be interwoven into primary classroom guided reading instruction to support acceleration. Fluency, supportive book introductions, prompting, and observation and analysis are emphasized throughout.",,
-Teaching Students With Developmental Disabilities Daily Living Skills Using Point-of-View Modeling Plus Video Prompting With Error Correction,S. Gardner,"A primary goal of instruction for students with developmental disabilities is to enhance their future quality of life by promoting skill acquisition, which will enable them to live, function, and participate in the community. One instructional method that can help students with developmental disabilities improve independence in performing daily living skills is video prompting (VP). This study investigated the effectiveness of a VP plus error correction procedure on skill acquisition when teaching daily living skills to four adolescents with mild and moderate developmental disabilities. A multiple baseline across participants design demonstrated that all four participants rapidly acquired dish washing skills upon introduction of the intervention. Future research should further explore the efficacy of error correction procedures used with VP and the impact that these procedures have on student learning and skill retention.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that all four participants rapidly acquired dish washing skills upon introduction of the intervention and future research should further explore the efficacy of error correction procedures used with VP and the impact that these procedures have on student learning and skill retention.'}",https://scholarlycommons.susqu.edu/cgi/viewcontent.cgi?article=1000&context=educ_fac_pubs
-Teaching Sight Words to Elementary Students With Intellectual Disability and Autism: A Comparison of Teacher-Directed Versus Computer-Assisted Simultaneous Prompting.,M. Coleman,"The purpose of this study was to compare the effects of teacher-directed simultaneous prompting to computer-assisted simultaneous prompting for teaching sight words to 3 elementary school students with intellectual disability. Activities in the computer-assisted condition were designed with Intellitools Classroom Suite software whereas traditional materials (i.e., flashcards) were used in the teacher-directed condition. Treatment conditions were compared using an adapted alternating treatments design. Acquisition of sight words occurred in both conditions for all 3 participants; however, each participant either clearly responded better in the teacher-directed condition or reported a preference for the teacher-directed condition when performance was similar with computer-assisted instruction being more efficient. Practical implications and directions for future research are discussed.","{'model': 'tldr@v2.0.0', 'text': ""Comparing the effects of teacher-directed simultaneous prompting to computer-assisted simultaneous prompting for teaching sight words to 3 elementary school students with intellectual disability found each participant either clearly responded better in the teachers' condition or reported a preference for the teacher's condition.""}",
-Tablet-based video modeling and prompting in the workplace for individuals with autism,Raymond Burke,"The current study involved a preliminary job-site testing of computer software, i.e., VideoTote, delivered via a computer tablet and designed to provide users with video modeling and prompting for use by young adults with an autism spectrum disorder (ASD) across a range of employment settings. A multiple baseline design was used to assess changes in rates of completion with a complex, 104-step shipping task by four participants diagnosed with ASD. Baseline data were collected on accuracy of task completion after exposure to typical job-training involving instruction, modeling, and practice. The intervention involved video modeling and prompting with a 13 minute video depicting an individual completing job responsibilities that entailed checking to make sure materials were in working order, replacing defective items, packing materials in a container, entering information into a computer, and attaching a label to a container. Results suggested that video modeling and prompting were effective in helping individuals with autism complete a multi-step shipping task. Participants and their parents gave the device and software high ratings as an acceptable treatment for adults with autism to use in the workplace and intervention that complies with universal design principles. Implications for competitive job opportunities for individuals with autism are discussed.",,
-The Effects of Prompting in Programed Instruction as a Function of Motivation and Instructions,G. Faust,,,
-A Comparison of Prompt Delays with Trial-and-Error Instruction in Conditional Discrimination Training,Sean J. O’Neill,,"{'model': 'tldr@v2.0.0', 'text': 'Compared three variations of prompt delay alongside trial-and-error instruction, results suggest progressive prompt delay may be effective and the most efficient in reducing learner errors during instruction.'}",https://europepmc.org/articles/pmc6269381?pdf=render
-A preliminary discourse analysis of constructivist-oriented mathematics instruction for a student with learning disabilities,Y. Xin,"ABSTRACT Reform efforts in mathematics education arose, in part, in response to constructivist works on conceptual learning. However, little research has examined how students with learning disabilities (LD) respond to constructivist-oriented instruction in mathematics, particularly in moment-to-moment interactions. To understand the nature of constructivist-oriented mathematics instruction involving students with LD, the authors conducted a case study to analyze teacher–student interactions during constructivist-oriented small group instruction involving a student with LD. The student demonstrated, to a certain degree, the ability to reason mathematically when provided with appropriate opportunities and prompting. However, given the limited intervention time, his reasoning and problem solving did not seem to go beyond the semiconcrete level of operation, which may have inhibited his solving of complex word problems with large numbers. Findings indicate that more efforts are needed to support students, those with LD in particular, in their transitions from concrete or semiconcrete to abstract conceptual understanding and problem solving.",,
-Video Modeling and Prompting in Practice,Ryan O. Kellems,"This article discusses the creation of video modeling (VM) and video prompting (VP) interventions for teaching novel multi-step tasks to individuals with disabilities. This article reviews factors to consider when selecting skills to teach, and students for whom VM/VP may be successful, as well as the difference between VM and VP and circumstances affecting which to choose. The purpose of this article is to provide practitioners with step-by-step instruction on how to teach cooking skills using VM and VP for young adults with disabilities.",,
-Use of Continuous Video Prompting to Teach a Student With a Significant Disability,Helen I. Cannella-Malone,,,
-Analysis of Live Modeling Plus Prompting and Video Modeling for Teaching Imitation to Children with Autism,Logan S. McDowell,"Previous researchers have demonstrated that training in imitation can significantly improve the learning capabilities of children diagnosed with autism spectrum disorder (ASD) and that children within this population show a preference for video presentations. Video-based instruction has been used to teach a variety of behaviors to individuals with ASD. However, only a small number of studies have examined the use of video modeling to teach initial imitation. Furthermore, there are limited and conflicting data on the effectiveness of a video modeling procedure that does not incorporate prompting when used to teach imitation to young children with ASD. Thus, the purpose of this study was to evaluate a video-modeling-alone procedure and a live-modeling-with-prompting procedure for teaching imitation to young children with ASD. The results suggest that the live modeling with prompting procedure was more effective, and implications related to this finding are discussed. Copyright © 2015 John Wiley & Sons, Ltd.",,
-Using Robot-Assisted Instruction to Teach Students with Intellectual Disabilities to Use Personal Narrative in Text Messages,R. Pennington,"In the current investigation, we evaluated the effectiveness of a multi-component package (i.e., robot, simultaneous prompting, self-graphing) for teaching three students, ages 19–21, with intellectual disabilities (ID) to write text messages that included a greeting, personal narrative, and closing. Data suggest that the package was effective in increasing correct performance for all participants. In addition, participants demonstrated their newly acquired texting skills across different communicative partners.",,
-Effects of Prompting Critical Reading of Science News on Seventh Graders' Cognitive Achievement.,Pei-ying Tsai,"Science teachers frequently select science news articles as supplementary teaching materials; however, the literature indicates that students encounter difficulties in examining and evaluating the news content and textual elements. This paper reports an instructional strategy of utilizing science news articles and investigates its effectiveness in enhancing students’ cognitive learning outcomes. In this quasi-experimental study, 118 seventh graders from four classes in one secondary school in Taiwan took part in the Science News Instruction (SNI). After eight weeks of instruction about Genetics and Reproduction, all students were requested to present their written arguments to the selected science news article. Two of the classes were prompted to link to their science content knowledge (the CK-SNI group, n=56), whereas their counterpart peers were not instructed to link to their science content knowledge (the SNI group, n=62). The analysis of students’ written artifacts revealed that the CK-SNI group performed significantly better on examining the data and claims made in the science news, producing warrants of higher quality to support their reasoning about the relationship between data and claims, and providing better reasons to refute the identified causal textual elements. Furthermore, the results of ANCOVA indicate that the CK-SNI group outperformed the SNI group on the cognitive test, with the largest improvement at Apply level (ηp 2 = .10). We conclude that prompting students to examine news content based on the content knowledge they learned is effective in fostering students to evaluate the science news critically, and thus make a positive impact on students’ cognitive learning.",,
-Embedding Science Facts in Leisure Skill Instruction Conducted by Peer Tutors,Erin E. Fetko,"This investigation evaluated the effectiveness of using peer tutors to teach a chained leisure skill (i.e., UNO card game) to three middle school students with disabilities using a simultaneous prompting procedure within a multiple probe design. The investigation also assessed whether the students with disabilities would acquire four unrelated science core content facts presented as nontargeted information during instructive feedback. Results indicated that all students met or made progress toward criterion on the leisure skill. In addition, two of the three students acquired all four core content facts.",,
-Using Simultaneous Prompting to Teach Computer-based Story Writing to a Student with Autism,R. C. Pennington,"Writing is a critical skill because it is used to access reinforcement in a variety of contexts. Unfortunately, there has been little research on writing skills instruction for students with intellectual disabilities and autism spectrum disorders. The purpose of this study was to evaluate the effects simultaneous prompting and computer-assisted instruction on story writing responses of a 7-year-old male with autism. Data indicated that the intervention was effective in teaching the participant to construct stories related to three different topics. Additionally, the student maintained responding at 2 and 4 weeks following intervention and increased responding across different topographies (i.e., handwriting, vocal).",,
-Comparing the Effects of Video Prompting with and without Error Correction on Skill Acquisition for Students with Intellectual Disability,Helen I. Cannella-Malone,"This study used an iPod Touch to compare the effects of video prompting with and without error correction on the acquisition of two daily living skills across three students with moderate to profound intellectual disability and an extremely limited daily living skills repertoire. An adapted alternating treatments design within a multiple probe across participants design was used to demonstrate that the inclusion of error correction from the outset of intervention increased the efficiency of skill acquisition for at least one task for two students. For the third student, some skill acquisition was observed using video prompting both with and without error correction, but more stable responding was achieved using in vivo instruction.","{'model': 'tldr@v2.0.0', 'text': 'An adapted alternating treatments design within a multiple probe across participants design was used to demonstrate that the inclusion of error correction from the outset of intervention increased the efficiency of skill acquisition for at least one task for two students.'}",
-Teaching Daily Living Skills to Seven Individuals With Severe Intellectual Disabilities: A Comparison of Video Prompting to Video Modeling,Helen I. Cannella-Malone,"We conducted a systematic replication of Cannella-Malone et al. by comparing the effects of video prompting to video modeling for teaching seven students with severe disabilities to do laundry and wash dishes. The video prompting and video modeling procedures were counterbalanced across tasks and participants and compared in an alternating treatments design within a multiple probe across participants design. For six participants, video prompting was more effective than video modeling, which was generally ineffective. For one participant, neither video modeling nor video prompting was effective, but in vivo instruction led to skill acquisition. One participant who was deaf was also able to learn both skills using video prompting, even though he could not hear the voice-over instructions. These data suggest that the duration of the video may influence its effectiveness as a teaching tool and that the voice-over instructions may not be necessary.",,
-Self-monitoring during whole group reading instruction: effects among students with emotional and behavioral disabilities during summer school intervention sessions,Lisa A. Rafferty,"The purpose of this study was to examine the effects that a self-monitoring strategy, plus a tactile prompting device, had upon the on-task and oral reading fluency behaviors of students with emotional and/or behavioral disabilities in the general education setting when used during whole group reading instruction. A multiple-baseline across pairs of participants design was used to evaluate the effects of the intervention. The results show that all four participants increased their on-task behavior during whole group reading instruction after being taught how to use the self-monitoring strategy, and their on-task behavior results were more similar to their peers' during the intervention condition. In addition, all students showed meaningful gains in oral reading fluency after being taught to self-monitor during whole group reading instruction when compared to baseline levels. Generalization probe data were taken during teacher-led, small group reading instruction. Although there were slight increases in performance for each of the target students after using the intervention for at least one week, these results should be interpreted with caution and should be further examined with future research.",,
-Coder Reviewer Reranking for Code Generation,Tianyi Zhang,"Sampling diverse programs from a code language model and reranking with model likelihood is a popular method for code generation but it is prone to preferring degenerate solutions. Inspired by collaborative programming, we propose Coder-Reviewer reranking. We augment Coder language models from past work, which generate programs given language instructions, with Reviewer models, which evaluate the likelihood of the instruction given the generated programs. We perform an extensive study across six datasets with eight models from three model families. Experimental results show that Coder-Reviewer reranking leads to consistent and significant improvement (up to 17% absolute accuracy gain) over reranking with the Coder model only. When combined with executability filtering, Coder-Reviewer reranking can often outperform the minimum Bayes risk method. Coder-Reviewer reranking is easy to implement by prompting, can generalize to different programming languages, and works well with off-the-shelf hyperparameters.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that Coder-Reviewer reranking leads to consistent and significant improvement over reranking with the Coder model only, when combined with executability filtering, and can often outperform the minimum Bayes risk method.'}",https://arxiv.org/pdf/2211.16490
-Teaching students with autism to tie a shoelace knot using video prompting and backward chaining,C. Rayner,"Purpose: To evaluate the effects of video prompting and backward chaining for teaching students with autism to tie a shoelace knot. Method: Videos featuring an adult and a peer or sibling model were used as part of the video prompting procedures to teach three boys with autism to tie a shoelace knot. A backward chaining procedure involving live modelling and verbal instruction was introduced following the video prompting phases. Results: Although the video prompting interventions increased the number of steps in the shoelace tying task completed by each of the participants, the backward chaining procedure was more effective, enabling one participant to reach mastery and a second participant to approach mastery. Conclusion: Practitioners should consider the pre-requisite skills of the participants and the nature of the target behaviour when selecting an intervention to teach daily living skills to individuals with autism.","{'model': 'tldr@v2.0.0', 'text': 'Although the video prompting interventions increased the number of steps in the shoelace tying task completed by each of the participants, the backward chaining procedure was more effective, enabling one participant to reach mastery and a second participant to approach mastery.'}",
-Sight Word Instruction for Students with Autism: An Evaluation of the Evidence Base,Janet E. Spector,,"{'model': 'tldr@v2.0.0', 'text': 'Evidence is found in support of a massed trials approach featuring student response to a succession of items, differential positive reinforcement, systematic prompting, and use of visual supports in teaching students with autism to read printed words.'}",
-Research on Reading Instruction for Individuals with Significant Cognitive Disabilities,D. Browder,"This article presents the results of a comprehensive review of 128 studies on teaching reading to individuals with significant cognitive disabilities. The review compared these studies against the National Reading Panel's components of reading; although it revealed an inadequate consideration of the components of reading, it found strong evidence for teaching sight words using systematic prompting and fading. The reviewers considered not only the number of studies, but also indicators proposed for evidence-based practice and effect size. This study identified some high quality studies with strong effect size for comprehension and fluency, but only one phonics study was strong in both quality and effects. Additional research is needed to promote broader skills in literacy for this population.",,
-Using Simultaneous Prompting to Teach Restaurant Words and Classifications as Non-target Information to Secondary Students with Moderate to Severe Disabilities,Bethany R. Smith,"This paper reviews selected literature pertaining to simultaneous prompting and the acquisition of non-target information for individuals with moderate to severe disabilities. The purpose of this review was to discuss the definition of non-target information (NTI) and the various places it can be embedded within an instructional trial. The acquisition of NTI has become especially critical to teachers who serve this population of students. Incidental learning provides teachers with an opportunity to increase a student's exposure to information without increasing instructional time, therefore increasing the efficiency of the teacher's instruction. A multiple probe design with full conditions across behaviors and replicated across participants evaluated the effectiveness of simultaneous prompting (SP) when teaching four participants with moderate and severe disabilities to read 12 sight words from community restaurants. In addition, the researcher examined generalization of the non-target information provided in the discriminative stimulus. Data indicated that SP was effective in teaching the participants the targeted sight words and the participants maintained the knowledge at high rates over time. Additionally, participants acquired NTI knowledge of the food classification embedded in the discriminative stimulus. As expectations for students with moderate or severe disabilities increase, effective and effi- cient instruction is of the highest importance to special education professionals. The strug-",,
-An Automated Prompting System for Smart Environments,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'This paper takes the approach to fully automating a prompting system without any predefined rule set or user feedback and uses simple off-the-shelf sensors to learn the timing for prompts based on real data that is collected with volunteer participants in the smart home testbed.'}",http://www.eecs.wsu.edu/~cook/pubs/icost11p2.pdf
-Academic Instruction for Students With Moderate and Severe Intellectual Disabilities in Inclusive Classrooms,J. Downing,"Foreword. by Diane Ryndak - Diane Ryndak Preface Acknowledgments About the Author 1. Teaching Students With Moderate to Severe Intellectual Disabilities in General Education Classrooms: Foundational Beliefs Key Concepts A Historical Perspective: Where We Came From The Present Situation and Challenge What is Inclusive Education? What is Not Inclusive Education Who Are We Talking About? Summary 2. Instructional Strategies and Teaching Arrangements Key Concepts Characteristics of Effective Instruction for All Students Clear Expectations Analyzing Tasks for Improved Learning What We Know About Teaching Students with Moderate to Severe Intellectual Disabilities The Importance of Student Interests Components of the Teaching Task Prompting Strategies Consequences of the Behavior Using Sequences of Different Prompts to Teach Students: Shaping Behavior Maintaining and Generalizing Skills Teaching Arrangements in General Education Classrooms Summary 3. Determining Student Needs: What to Teach Key Concepts Limitations of Standardized Assessment Family and Child-Based Assessment Procedure Record Review Observational Assessments What's the Class Doing? Interpreting Content Standards Blending Student/Family Goals with State Standards Identifying Learning Opportunities Writing IEP Goals and Objectives Summary 4. Teaching Core Curriculum to Students With Moderate to Severe Intellectual Disabilities Key Concepts The Critical Need to Adapt Curriculum to Make it Meaningful Identifying the BIG Ideas from Core Curriculum Determining Prompts to Use for a Particular Student and Lesson Examples of Students Receiving Direct Instruction Across Grades and Instructional Arrangements Large Group Instruction Generalization of Skills Taught Summary 5. It Takes a Village: Teaching as a Collaborative Effort Key Concepts The Expectation of Team Collaboration Team Members Involved in Instruction Credentialed Teachers Co-Teaching Supporting General Education Ownership Paraprofessionals as Teachers Related Service Providers Parent Volunteers Peers as Teachers A Few Cautions When Using Peers The Need for Information and Training Effective Use of Team Members The Importance of Consistency Generalization of Skills Across Team Members Summary 6. Keeping Track of Student Progress, by Kathryn D. Peckham-Hardin and June E. Downing Key Concepts Types of Data Collection Strategies Linking Data Collection Methods to the IEP Objectives Collecting Data While Teaching in General Education Classrooms Examples of Collecting Data During Instructional Times Test Taking by the Class Training Paraprofessionals and Others to Take Data The Need for Alternate Assessment Summary 7. He's Getting It! Now What? Taking Learning to the Next Level Key Concepts Involving the Student in Planning Next Steps Writing IEP Objectives to Reflect Next Steps Using Standards and Performance Indicators to Determine Next Steps Using Task Analyses to Determine Next Steps Using Life Needs to Determine Next Steps Postsecondary Options Next Steps for Nonacademic Skills Summary References Index","{'model': 'tldr@v2.0.0', 'text': 'This book discusses Teaching Students With Moderate to Severe Intellectual Disabilities in General Education Classrooms: Foundational Beliefs and Instructional Strategies and Teaching Arrangements, and the Critical Need to Adapt Curriculum to Make it Meaningful.'}",
-Effects of No-No Prompting on Teaching Expressive Labeling of Facial Expressions to Children with and without a Pervasive Developmental Disorder,Justin B. Leaf,"Discrete trial teaching is a systematic form of instruction found to be effective for children diagnosed with autism. Three areas of discrete trial teaching warranting more research are the effectiveness and efficiency of various prompting procedures, the effectiveness of implementing teaching in a group instructional format, and the ability of children with autism to observationally learn from their peers. This study evaluated the effectiveness of a no-no prompting procedure implemented in a group instructional format to teach five children, four of whom were diagnosed with an autism spectrum disorder, to correctly label facial expressions. Addition- ally, this study evaluated whether participants observationally learned to expressively label facial expressions not directly taught to them but taught to their peers. Using a multiple baseline design, results of this study indicated that all participants learned to expressively label facial expressions taught to them directly with a no-no prompting procedure. The participants also learned through observation to expressively label facial expressions taught only to their peers using the same procedure.",,
-Effects of Peer Tutoring With Audio Prompting on Vocabulary Acquisition for Struggling Readers,S. M. Mackiewicz,"Reciprocal peer tutoring can be an effective supplement to teacher-led instruction, but students need to have the tutoring skills necessary to teach their peers successfully. Previous studies have addressed the challenge of providing essential information to a naïve tutor, allowing for correct modeling and feedback. The present study compared incidental learning of vocabulary words through classroom reading instruction to a combination of incidental learning supplemented with peer tutoring. Eight fourth-grade students were trained to tutor each other using a digital recording and playback device that provides audio prompts to naïve tutors. Results indicated that students made modest gains from incidental learning and much stronger gains from peer tutoring with audio prompting. Implications for practice and future research are discussed.",,
-Comparison of the Effects of Static Picture and Video Prompting on Completion of Cooking Related Tasks by Students with Moderate Intellectual Disabilities,L. Mechling,"This study compared the effects of static photographs and video prompts on the independent performance of cooking related tasks by six young adults with moderate intellectual disabilities. An adapted alternating treatment design with baseline and final treatment phase was used to measure the percentage of tasks correctly completed by each student across treatments. Results indicated that both procedures were effective in increasing correct performance of tasks for all participants from baseline levels in which only verbal task directions were provided. All six participants demonstrated greater independent responding on the target tasks when using video prompting. During the final treatment phase students increased their level of performance using video prompting with sets of tasks receiving static picture prompts in the comparison phase. Implications for future research and the development of curriculum, which include instruction of self-prompting strategies, are discussed.",,
-Generating Instruction Automatically for the Reading Strategy of Self-Questioning,Jack Mostow,"Self-questioning is an important reading comprehension strategy, so it would be useful for an intelligent tutor to help students apply it to any given text. Our goal is to help children generate questions that make them think about the text in ways that improve their comprehension and retention. However, teaching and scaffolding self-questioning involve analyzing both the text and the students' responses. This requirement poses a tricky challenge to generating such instruction automatically, especially for children too young to respond by typing. This paper describes how to generate self-questioning instruction for an automated reading tutor. Following expert pedagogy, we decompose strategy instruction into describing, modeling, scaffolding, and prompting the strategy. We present a working example to illustrate how we generate each of these four phases of instruction for a given text. We identify some relevant criteria and use them to evaluate the generated instruction on a corpus of 513 children's stories.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes how to generate self-questioning instruction for an automated reading tutor and decomposes strategy instruction into describing, modeling, scaffolding, and prompting the strategy.'}",
-Enhancing creativity: A comparison between effects of an indicative instruction “to be creative” and a more elaborate heuristic instruction on Chinese student creativity.,W. Niu,"This study consisted of an experiment conducted with a sample of 180 Chinese high school students,designed to scrutinize the effects of prompting and instructing students to be creative in their creativeperformance in the domains of art and literature. The results of the study reveal that students creativitybenefited from receiving appropriate strategies or more elaborate instruction (as induced by instructing),but not by merely being motivated (as induced by prompting) to be creative, regardless of the taskdomain. Gender differences in creativity were found only in the domain of literature, in favor of femaleparticipants. Implications of the findings are discussed.Keywords: creativity, instruction, Chinese students, motivation",,
-Teaching effective hand raising to children with autism during group instruction.,Shaireen M. Charania,"We taught 3 children with autism to raise a hand or keep both hands down depending on their status (e.g., having heard a target word, possessing a specific item) using modeling, prompting, and reinforcement. All 3 children acquired accurate hand-raising skills in response to progressively more difficult discrimination tasks during group instruction. The implications for preparing children for general education settings are discussed.","{'model': 'tldr@v2.0.0', 'text': '3 children with autism are taught to raise a hand or keep both hands down depending on their status using modeling, prompting, and reinforcement to acquire accurate hand-raising skills in response to progressively more difficult discrimination tasks during group instruction.'}",https://europepmc.org/articles/pmc2938940?pdf=render
-Effectiveness of multimedia-based instruction that emphasizes molecular representations on students' understanding of chemical change,D. Ardaç,"The present study makes use of the capabilities of computerized environments to enable simultaneous display of molecular representations that correspond to observations at the macroscopic level. This study questions the immediate and long-term effects of using a multimedia instructional unit that integrates the macroscopic, symbolic, and molecular representations of chemical phenomena. Forty-nine eighth graders received either multimedia-based instruction that emphasized molecular representations (n = 16), or regular instruction (n = 33). Students who received multimedia-based instruction that emphasized the molecular state of chemicals outperformed students from the regular instruction group in terms of the resulting test scores and the ease with which they could represent matter at the molecular level. However, results relating to the long-term effects suggest that the effectiveness of a multimedia-based environment can be improved if instruction includes additional prompting that requires students to attend to the correspondence between different representations of the same phenomena. © 2004 Wiley Periodicals, Inc. J Res Sci Teach 41: 317–337, 2004",,
-A Qualitative Study of Teacher Perceptions on Using an Explicit Instruction Curriculum to Teach Early Reading Skills to Students with Significant Developmental Disabilities,D. Taylor,"This study examines teachers’ perceptions about a structured literacy curriculum that uses explicit instruction and teacher script combined with instructional practices common in special education such as time delay, error correction, and prompting strategies. The main research question addressed by this study was “What are the perceptions and beliefs of teachers of students with significant developmental disabilities about the effects of using an explicit reading instruction curriculum?” Three themes or key findings emerged from our analysis of the data regarding teachers’ perceptions of such curricula. These themes suggest that these teachers believe that a structured literacy curriculum has a positive impact on student learning, teacher effectiveness, and teacher self-efficacy.",,
-Making Metacognition Explicit: Developing a Theoretical Foundation for Metacognitive Prompting During Scenario-based Training,S. Fiore,"In this paper we describe a set of metacognition related training interventions that can be introduced in advanced scenario-based training systems. A long line of research in the cognitive and learning sciences has shown that knowledge acquisition is more effective when learners reflect upon their comprehension processes. We support this view and discuss how instruction can assist trainees in monitoring their learning in order to facilitate overall retention and transfer. We focus on the multidimensional nature of metacognition and emphasize the regulatory component of metacognition. Regulation involves an active process engaged by the trainee in service of learning. We present a theoretical framework that describes how this can be developed as a means of instructional strategy in preparation for, during execution of, and in reflection upon, a given learning episode.",,
-An Effective Framework for Primary-Grade Guided Writing Instruction,Sharan A. Gibson,"This article describes the theory and procedures for a specific format of primary-grade small-group writing instruction (guided writing lessons) based on (a) a sociocultural perspective, (b) Clay's theories of literacy learning, and (c) the author's study of second-grade guided writing instruction. Guided writing lessons are described as an intensive, small-group context for instructional support and interaction between teacher and students during writing that can be folded into a classroom context, including writing workshops and interactive writing. The purpose, format, teacher prompting, and assessment procedures of guided writing lessons are demonstrated within the article through the presentation of student examples, lesson excerpts, teacher prompts for writing, and an observation form. The article also describes the use of expert observation of student behavior while writing and an analytic assessment of writing products in order to maintain a strong pace for students' learning within and across lessons. تصف هذه المقالة نظرية نظام خاص وإجراءاته في الصفوف الابتدائية بصدد تعليم الكتابة في مجموعات صغيرة (دروس الكتابة الموجهة) المبني على (ا) وجهة اجتماعية ثقافية و(ب) نظريات (كلاي) حيال تعلم القراءة والكتابة و(ج) بحث المؤلف العلمي بشأن تعليم الكتابة الموجه في الصف الثاني. وتتم وصفة دروس الكتابة الموجهة كسياق مكثف في مجموعات صغيرة لدعم التعليم والتفاعل بين المعلم والطلاب خلال الكتابة ويمكن أن تنطوي هذه الدروس في سياق الصف التي تشتمل على ورشات الكتابة والكتابة التفاعلية. وقد تم عرض غرض دروس الكتابة الموجهة ونظامها وحث معلمها وإجراءات تقييمها في المقالة بواسطة أمثلة من الطلاب ومقتطفات الدروس وتحاريض المعلم على الكتابة واستمارة مراقبة. وتصف كذلك استخدام مراقبة الخبير تصرفات الطلاب عند كتابتهم والتقدير التحليلي لإنتاجات الكتابة من أجل الحفاظ على سرعة تعلم الطلاب القوية في الدروس وعبرها. 本文描述一个教授小学写作教学方法的理论和程序。该教学方法（指导写作练习课）是以特别的小组形式进行，其基础建于（a）一个社会文化观点，（b）克雷的读写学习理论，及（c）本文作者对小学二年级指导写作教学的研究。指导写作练习课能在教授写作时，为学生提供一个强化的小组学习环境，支援教学和促进教师与学生之间的互动。指导写作练习课可以融入一般的课堂教学环境中，包括写作工作坊和互动写作练习等。本文透过展示学生学习例子、授课节录、教师指导写作的提示及观课记录表，来表明指导写作练习课的教学目的、教学形式、教师提示和教学评估程序。本文亦描述如何使用专门方法去观察学生在写作练习时出现的行为，及如何使用分析评估方法去评估学生的写作成果，以帮助学生在课堂内，和课堂与课堂之间，保持稳健的学习进度。 Cet article decrit la theorie et les procedures d'un format specifique d'enseignement de l'ecriture en petit groupe en premiere annee (lecons d'ecriture guidee), a partir (a) d'une perspective socioculturelle, (b) des theories d'apprentissage de la litteratie de Clay, et (c) de l'etude effectuee par l'auteur d'un enseignement de l'ecriture guidee en seconde annee. Les lecons d'ecriture guidee sont decrites comme un contexte intensif en petit groupe, en vue d'un soutien pedagogique et d'interactions maitre- eleves pendant l'ecriture qui puisse etre mis en œuvre en contexte scolaire, y compris pendant des ateliers d'ecriture et d'ecriture interactive. L'article indique le but, le format, les interventions du professeur, et les procedures d'evaluation des lecons d'ecriture guidee, au moyen de la presentation d'exemples d'eleves, d'extraits de lecons, des incitations a ecrire du professeur, et d'un protocole d'observation. L'article decrit egalement l'utilisation d'une observation d'expert du comportement de l'eleve pendant l'ecriture et une evaluation detaillee des productions ecrites afin de conserver un rythme soutenu d'apprentissage des eleves au cours d'une lecon et d'une lecon a l'autre. Предлагаются теоретические обоснования и процедуры для определенного формата групповой работы над письмом в начальной школе (направляемое письмо), основанные на (а) социокультурном подходе, (б) теории изучения грамотности Клэя, и (в), исследованиях самого автора в области обучения второклассников письменной речи. Направляемое письмо происходит как интенсивные занятия в малых группах, проходящие с поддержкой учителя и при активном взаимодействии между учителем и учениками в ходе выполнения заданий. Такие занятия могут быть элементами обычного урока, писательского семинара или сессии интерактивного письма. Цель, формат, степень поддержки, которую оказывает учитель, и процедуры оценивания демонстрируются на примере ученических работ, эпизодов урока, типичных реплик-подсказок учителя и форм, которые он заполняет, наблюдая за работой учащихся. В статье также описано, как проводить экспертное наблюдение за процессом письма и аналитическое оценивание его результатов, чтобы обеспечить динамичное развитие навыков письма в пределах каждого отдельного урока и учебного курса в целом. Este articulo describe la teoria y los procedimientos de un formato especifico de la ensenanza de la escritura en pequenos grupos al nivel de primaria (lecciones de escritura guiada) basados en (a) una perspectiva sociocultural, (b) la teoria de Clay sobre el aprendizaje de competencias, y (c) el estudio hecho por el autor de la ensenanza de la escritura guiada en una clase de segundo grado. Las lecciones de la escritura guiada se describen como un contexto intenso de trabajo en grupos pequenos para apoyar y ofrecer interaccion entre el maestro y los estudiantes al escribir. Estos grupos pequenos se pueden integrar en el contexto del salon de clase, inclusive en talleres de escritura y en la escritura interactiva. En el articulo se demuestran el proposito, el formato, y los procedimientos para las sugerencias del maestro y la evaluacion de las lecciones de escritura guiada por medio de la presentacion de ejemplos de estudiantes, fragmentos de lecciones, las sugerencias de los maestros en la escritura, y un formulario de observaciones. El articulo tambien describe el uso de la observacion experta del comportamiento de los estudiantes mientras escriben y una evaluacion analitica de los productos de la escritura para mantener un ritmo rapido de aprendizaje dentro y a traves de las lecciones.",,
-Speak,Beth Whittle,"This study presents holistic insights into the culturally relevant English language arts and reading instructional practices of one award-winning Asian-American male teacher as he worked with culturally and linguistically diverse students from a variety of backgrounds in a ‘significantly underperforming’ urban middle-school. Avenues through which the exemplary teacher reflected his culturally relevant teaching practices were planning instruction, prompting students to think critically through reflection, modeling, conferencing with students, facilitating peer evaluation and support, emphasizing procedural processes, establishing routine classroom practices, and capitalizing on technology and visual aid support. Mr. Wang found consistent collaboration with other teachers and opportunities for debriefing based on data from instruction to be the most beneficial forms of support. Students’ positive responses to Mr. Wang’s responsiveness were most influential to his continued instruction. Implications are provided.",,https://www.qeios.com/read/GUMCLM/pdf
-Simulating naturalistic instruction: the case for a voice mediated interface for assistive technology for cognition,B. O'Neill,"A variety of brain pathologies can result in difficulties performing complex behavioural sequences. Assistive technology for cognition (ATC) attempts support of complex sequences with the aim of reducing disability. Traditional ATCs are cognitively demanding to use and thus have had poor uptake. A more intuitive interface may allow ATCs to reach their potential. Insights from psychological science may be useful to technologists in this area. We propose that an auditory‐verbal interface is more intuitive than a visual interface and reduces cognitive demands on users. Two experiments demonstrate a novel ATC, the General User Interface for Disorders of Execution (GUIDE). GUIDE is novel because it simulates normal conversational prompting to support task performance. GUIDE provides verbal prompts and questions and voice recognition allows the user to interact with the GUIDE. Research with non‐cognitively impaired participants and a single participant experiment involving a person with vascular dementia provide support for using interactive auditory‐verbal interfaces. Suggestions for the future development of auditory‐verbal interfaces are discussed.",,http://eprints.lse.ac.uk/38691/1/Running_head%28lsero%29.pdf
-Reading Instruction for Students with Emotional and Behavioral Disorders and at Risk of Antisocial Behaviors in Primary Grades: Review of Literature,Mabel O. Rivera,"Frequently, students with emotional and behavior disorders (EBD) exhibit academic underachievement combined with high levels of externalizing behaviors and resistance to instructional efforts. Regardless of the present reading initiatives, research focusing on interventions for teaching reading to students with EBD continues to be limited. This article extends previous efforts to review literature concerning reading instruction interventions for students with EBD. Specifically, this review focuses on interventions employed in primary grades. Because of the paucity in research and documented issues related to late and misidentification of students with EBD, studies including students at risk of antisocial behaviors were included. Eleven studies were found and carefully reviewed. Results demonstrate the efficacy of several reading interventions, including Direct Instruction, peer tutoring, and behaviorally based procedures such as time delay prompting, trial and error, and differential reinforcement.",,
-Using Video Prompting to Teach Cooking Skills to Secondary Students with Moderate Disabilities.,Tara Graves,"Three secondary students with moderate disabilities acquired cooking skills through a constant time delay procedure used with video prompting. A multiple probe design was used to evaluate effectiveness of the procedure to teach preparation of a food item (a) on a stove, (b) in a microwave, and (c) on a counter top. The procedure was effective for each student. Guidelines for using videotapes in the instruction of functional skills are discussed.",,
-The reading assistant: eye gaze triggered auditory prompting for reading remediation,J. Sibert,We have developed a system for remedial reading instruction that uses visually controlled auditory prompting to help the user with recognition and pronunciation of words. Our underlying hypothesis is that the relatively unobtrusive assistance rendered by such a system will be more effective than previous computer aided approaches. We present a description of the design and implementation of our system and discuss a controlled study that we undertook to evaluate the usability of the Reading Assistant.,"{'model': 'tldr@v2.0.0', 'text': 'A system for remedial reading instruction that uses visually controlled auditory prompting to help the user with recognition and pronunciation of words and a controlled study is discussed that was undertook to evaluate the usability of the Reading Assistant.'}",
-Effects of Manipulative Instruction on Solving Area and Perimeter Problems by Students with Learning Disabilities,Mike Cass,"Abstract. A multiple baseline design was employed to test the effect of manipulative instruction on the perimeter and area problem-solving performance of middle and high school students who had been diagnosed with LD in the area of mathematics. Modeling, prompting/guided practice, and independent practice in conjunction with manipulative training were employed to teach both perimeter and area problem-solving skills. Analysis of data revealed that the students rapidly acquired the problem-solving-skills, maintained these skills over a two-month period, and transferred these skills to a paper and pencil problem-solving format. This research extends previous findings by revealing that use of concrete manipulatives promotes the long-term maintenance of skills.",,
-"Improving Comprehension Instruction: Rethinking Research, Theory, and Classroom Practice.",C. Block,"Tables, Figures, and Exhibits.Foreword (Gerald G. Duffy).Preface.The Editors.The Contributors.Part One: New Directions in Comprehension Instruction.Introduction: Improving Comprehension Instruction: An Urgent Priority (Linda B. Gambrell, Cathy Collins Block, and Michael Pressley).1. Reconceptualizing Reading Comprehension (Anne P. Sweet and Catherine Snow).2. The Thinking Process Approach to Comprehension Development: Preparing Students for Their Future Comprehension Challenges (Cathy Collins Block and Rebecca B. Johnson).3. From Good to Memorable: Characteristics of Highly Effective Comprehension Teaching (Ellin Oliver Keene).4. The Guided Reading Lesson: Explaining, Supporting, and Prompting for Comprehension (Gay Su Pinnell).5. Instructional Components for Promoting Thoughtful Literacy Learning (Pamela J. Dunston).Part Two: New Comprehension Lessons Across the Curriculum.6. Differentiating Reading and Writing Lessons to Promote Content Learning (Karen D. Wood).7. Parsing, Questioning, and Rephrasing (PQR): Building Syntactic Knowledge to Improve Reading Comprehension (James Flood, Diane Lapp, and Douglas Fisher).8. Using Writing to Improve Comprehension: A Review of the Writing-to-Reading Research (Bena R. Hef.in and Douglas K. Hartman).9. Research-Based Comprehension Practices That Create Higher-Level Discussions (Janice F. Almasi).10. Goose Bumps and Giggles: Engaging Young Readers' Critical Thinking with Books from the Teachers' Choices Project and Graphic Organizers (Kathy N. Headley and Jean Keeler).Part Three: Integrating Technology and Innovative Instruction.11. Using Technology to Individualize Reading Instruction (David Rose and Bridget Dalton).12. Computers, Kids, and Comprehension: Instructional Practices That Make a Difference (Linda D. Labbo).13. Out of This World: Cyberspace, Literacy, and Learning (Victoria Gentry Ridgeway, Chris L. Peters, and Terrell Seawell Tracy).14. Reading in the Digital Era: Strategies for Building Critical Literacy (Lisa Patel Stevens and Thomas W. Bean).Part Four: Overcoming Comprehension Challenges.15. Hitting the Wall: Helping Struggling Readers Comprehend (D. Ray Reutzel, Kay Camperell, and John A. Smith).16. At-Risk Students: Learning to Break Through Comprehension Barriers (Lynn Romeo).17. Helping Struggling Readers Make Sense of Reading (Irene W. Gaskins, Sally R. Laird, Colleen O'Hara, Theresa Scott, and Cheryl A. Cress).Conclusion: Improving Comprehension Instruction: A Path for the Future (Michael Pressley).Name Index.Subject Index.","{'model': 'tldr@v2.0.0', 'text': 'Improving Comprehension Instruction: A Path for the Future is a path for the future by Linda B. Gambrell, Cathy Collins Block, and Michael Pressley.'}",
-Effectiveness of Peer Delivered Simultaneous Prompting on Teaching Community Signs to Students with Developmental Disabilities.,Elif Tekin‐Iftar,"A multiple probe study across behaviors, replicated across participants, assessed effectiveness of peer delivered simultaneous prompting (SP) in teaching expressively identifying community signs to four students with developmental disabilities. The two purposes of the study were: (a) to find out if peer tutors use simultaneous prompting reliably for instructing their tutees with developmental disabilities, and (b) to examine effectiveness of simultaneous prompting on teaching expressively identifying community signs. The definition of community sign was presented as instructive feedback. Besides these aims, generalization and maintenance effects of simultaneous prompting were also investigated in the study. Generalization across persons was tested before introducing simultaneous prompting and after tutees met criteria. Maintenance data for targeted and non-targeted behaviors were collected one week after instruction. Results show that peer tutors deliver SP reliably, and tutees acquired expressively identifying community signs. Tutees also gained some instructive feedback. Furthermore, tutees maintained acquired skills at criterion level and generalized acquired skills to another person at criterion level. Based upon evaluation of findings and implications of the study future research needs are discussed.",,
-Training Adults with Moderate and Severe Mental Retardation in a Vocational Skill Using a Simultaneous Prompting Procedure.,K. Maciag,"A multiple probe across subjects design evaluated the effectiveness of a simultaneous prompting procedure in training 10 adults with moderate and severe mental retardation to construct shipping boxes within a community vocational center. The trainer used a dyadic instructional format in which each member of the dyad received instruction on one half of the task analysis on an alternating basis. Two instructional sessions occurred daily, so each member of the dyad received instruction on the entire task analysis. Four of the five dyads learned to construct shipping boxes. Instructional data are presented and implications for future research are discussed. One of the primary roles of teachers is to only contribute to the growth of the individual promote the acquisition of target skills by uti- but also increase public awareness of the lizing effective and efficient instructional pro- strengths and contributions of individuals cedures. As curriculum goals for students with with mental retardation, moderate and severe disabilities focus on pre- The majority of vocational skills can be de paring these students for the least restrictive scribed as chained tasks; that is, a series of environment and vocational opportunities, behaviors that, when put together, form a teachers must identify instructional proce- more complex skill (Wolery, Ault, & Doyle, dures that are efficient in training vocational 1992). Numerous vocational skills have been tasks. Transition from school to the world of taught successfully in the literature and in work offers individuals the opportunity to de- elude opening a keyed lock (Fetko, Schuster, velop new social networks and increased per- Harley, & Collins, 1999), banking skills (Mc sonal independence. The acquisition of voca- Donnell & Ferguson, 1989), and other voca tional skills enables individuals to experience tional skills (Chandler, Schuster, & Stevens, increased self-esteem by increasing vocational 1993). Several response prompting proce opportunities, earning money, and contribut- dures such as constant and progressive time",,
-Identifying Teachers’ Supports of Metacognition Through Classroom Talk and Its Relation to Growth in Conceptual Learning,Cristina D. Zepeda,"A gulf exists between prior work testing metacognitive instructional interventions and teacher practices that may support metacognition in the classroom. To help bridge this gulf, we designed an observational protocol to capture whether and how teachers provide metacognitive support in their talk and examined whether these supports were related to student learning. We examined four features of metacognitive support, including the type of metacognitive knowledge supported (personal, strategy, or conditional), the type of metacognitive skill supported (planning, monitoring, or evaluating) the type of instructional manner in which the support was delivered (directives, prompting, or modeling), and the type of framing (problem specific, problem general, or domain general), during three types of instructional activities (individual, group, or whole-class instruction). We compared teacher talk from 20 middle school mathematics classrooms with high growth in conceptual mathematics scores with teacher talk from 20 classrooms with low growth. For each of these classrooms, we examined the amount of teacher talk that supported metacognition during one regular class period. Observations revealed that the high-conceptual growth classrooms had more metacognitive supports for personal knowledge, monitoring, evaluating, directive manners, and domain-general frames than the low-conceptual growth classrooms. We discuss the implications of those observations for bridging research on metacognition to teacher practice.",,
-Response prompting and fading methods: a review.,M. Demchak,"Teachers of individuals with severe handicaps typically use response prompts, such as verbal instructions, modeling, and physical guidance, to encourage correct responding. However, the goal of instruction is to have the individual respond to natural stimuli rather than response prompts. Therefore, these response prompts must be faded. Currently there are four methods for systematically fading response prompts: increasing assistance, decreasing assistance, graduated guidance, and time delay. An overview of each method as well as a review of comparative investigations involving these methods of fading response prompts was presented. Recommendations for practitioners and for future research were included.","{'model': 'tldr@v2.0.0', 'text': 'An overview of each method as well as a review of comparative investigations involving these methods of fading response prompts was presented and Recommendations for practitioners and for future research were included.'}",
-The Effect of Metacognitive Instruction in Outlining and Graphic Organizer Construction on Students' Comprehension in a Tenth-Grade World History Class,Thomas W. Bean,"The purpose of the present study was to explore the effect of graphic organizer instruction versus outlining on students' text recall in tenth-grade world history. Based on the hierarchical structure of graphic organizers depicting interrelationships among ideas, we hypothesized that students instructed in the use of graphic organizers would display significantly higher text recall than a group using outlines. Seventy-two tenth graders in three sections of world history participated in the study. Two sections received systematic instruction in the development of graphic organizers and a third section received parallel instruction in outlining. One of the two graphic organizer sections had previous instruction in summarization and question generation during an earlier study. Six 15-item multiple-choice quizzes comprised one of the dependent measures. On the first five quizzes there were no significant differences between the three sections. However, when teacher prompting was phased-out on quiz six, the graphic organizer group with previous training in summarization significantly outperformed the other two groups. Additionally, on a second dependent measure involving the development of a written recall protocol based on difficult college level text, the group with previous summarization experience significantly outperformed both the new graphic organizer group and the outlining section. Implications for a cumulative, long-range view of metacognitive instruction are discussed.",,
-Comparison of simultaneous prompting with and without error correction in teaching science vocabulary words to high school students with mild disabilities,Patti Johnson,,,
-The Effects of Simultaneous Prompting on Teaching Expressive Identification of Objects: An instructive feedback study,Elif Tekin‐Iftar,"The present study examined whether the use of a simultaneous prompting procedure would result in an increase in the percentage of correct responses when expressively identifying first aid materials. A multiple probe design across behaviours and replicated across students was used. Three training sets with a total of nine first aid materials were presented to three students. Also, instructive feedback stimuli were presented during consequent events to increase the efficiency of instruction. The instructive feedback con tained the functions of the instructional materials. Maintenance effects were assessed 1, 2, and 4 weeks after training. The results showed that all students learned expressive identification of first aid materials and maintained them after training. Furthermore, all students acquired and maintained some of the instructive feedback stimuli presented to them during instruction.",,
-Enhancing Choice and Participation for Adults with Severe Disabilities in Community-Based Instruction,K. Cooper,"Three adults with severe disabilities received instructional support to make purchases in fast food restaurants. During baseline and intervention, participants received discrete, systematic prompting and feedback to perform the steps of the purchasing task analysis. During intervention, participants were prompted to make five distinct choices during the community routine. This prompting resulted not only in increased choicemaking, but also in collateral increases in the performance of the purchasing task analysis.","{'model': 'tldr@v2.0.0', 'text': 'Three adults with severe disabilities received instructional support to make purchases in fast food restaurants and received discrete, systematic prompting and feedback to perform the steps of the purchasing task analysis.'}",
-The Evolution of Transactional Strategies Instruction in One Teacher's Classroom,R. Brown,"In this study, 1 teacher taught the same story for 3 consecutive years to 3 comparable groups of second-grade, low-achieving students. An interaction-tracking and coding scheme was used to analyze the 45-minute lessons for changes in interactional patterns, participation among group members, instructional focus, strategy instruction, and prompted and self-regulated use of strategies by students. By year 3, students participated more actively in story discussion and used strategies with less teacher prompting to support their interpretations and responses to text. These changes appeared to occur because of modifications in the teacher's instructional practices. The instruction that emerged during the third year could be characterized as transactional strategies instruction, an approach that involves teaching reading group members to use comprehension strategies as they jointly construct interpretations of text.","{'model': 'tldr@v2.0.0', 'text': 'The instruction that emerged during the third year could be characterized as transactional strategies instruction, an approach that involves teaching reading group members to use comprehension strategies as they jointly construct interpretations of text.'}",
-Divergent Thinking As a Function of Time and Prompting to “Be Creative” in Undergraduates,G. Johns,"This study examined divergent thinking in undergraduates as a function of gender, time, and test instruction conditions. Differences were examined in male and female undergraduates when tested on a divergent thinking test with or without prompting to “be creative” and with or without time limitations imposed. Subject groups were assigned to one of the following four conditions: standard instructions/four minute time limit; standard instructions/no time limit; creative instructions/four minute time limit; and creative instructions/no time limit. The no time limit condition resulted in significantly higher mean traditional fluency and flexibility divergent production scores. Males were found to have significantly higher mean divergent production scores on fluency. This study supported previous research findings regarding the lack of enhanced traditional divergent production scores for females resulting from creative instructions.",,
-Developing Joint Attention for Children with Autism in Robot-Enhanced Therapy,D. David,,"{'model': 'tldr@v2.0.0', 'text': 'Using more cues for prompting JA increases the performance of the children and emphasizes the importance of using more cues, such as pointing, for increasing engagement and performance engagement in a child–robot interaction session.'}",
-Community-based instruction for independent meal preparation by adults with profound mental retardation,Nirbhay N. Singh,,"{'model': 'tldr@v2.0.0', 'text': 'Assessment, training and followup data on the preparation of a dessert by three adults with profound mental retardation and very limited reading skills transitioned to community living showed that they were able to independently prepare the dessert in various community settings of their choice.'}",
-A Review and Critical Examination of the System of Least Prompts,Collin Shepley,"We systematically reviewed more than 25 years of applied research examining the system of least prompts response prompting procedure with individuals with disabilities. We identified 123 peer-reviewed studies including 413 participants receiving instruction with the system of least prompts. A total of 252 experimental designs were evaluated, with 51 designs indicating a functional relation and the presence of 154 demonstrations of effect across 91 individuals. Our data indicate that the system of least prompts is an evidenced-based practice for teaching chained responses related to community, self-care, and vocational skills to individuals with moderate intellectual disability who are 13 years of age or older. In addition, we present and discuss a method for analyzing and aggregating data from single-case studies to account for noneffects and publication bias when identifying if an intervention meets standards as an evidence-based practice.",,
-To Err or Not to Err: A Question for the Instruction of Psychomotor Skills,R. Singer,"s International, 1969, No. 69-20, 790, p. 2385-A. Prather, D. C. Trial-and-error versus errorless learning: Training, transfer and stress. American Journal of Psychology, 1971, 84, 377-386. Prather, D. C., & Berry, G. A. Comparison of trial-and-error versus highly prompted learning of a perceptual skill. Proceedings of the 78th Annual Convention of the American Psychological Association, 1970, 5, 677-678. Prather, D. C., Berry, G. A., & Bermudez, J. M. Effect of prompting and feedback on performance during learning, stress, and transfer of a perceptual skill. Proceedings of the 80th Annual Convention of the American Psychological Association, 1972, 7, 643-644. Schmidt, R. A. A schema theory of discrete motor skill learning. Psychological Review, 1975, 82, 225-260. Schurr, E. Movement experiences for children. New York: Appleton-Century- Crofts, 1967. Seidel, R. J. Theories and strategies related to measurement in individualized instruction (HumRRO Tec. Rep. 2-71). Alexandria, Va.: Human Resources Research Organization, March 1971. Shulman, L., & Keislar, E. Learning by discovery-a critical appraisal. Chicago: Rand McNally, 1966. Singer, R. N. Coaching, athletics, and psychology. New York: McGraw-Hill, 1972. Singer, R. N. Motor Learning and human performance. (2nd ed.) New York: Macmillan, 1975. Singer, R. N., & Dick, W. Teaching physical education: A systems approach. Boston: Houghton Mifflin, 1974. Singer, R. N., & Gaines, L. Effect of prompted and trial-and-error learning on transfer performance of a serial motor task. American Educational Research Journal, 1975, 12, 395-403. Singer, R. N., & Pease, D. The effect of different instructional strategies on",,
-The Effects of Strong Formal Prompts in Programed Instruction,Richard C. Anderson,"Authors of self-instructional programs are often cautioned to avoid ""overprompting."" I t has never been demonstrated experimentally that ""overprompting"" inhibits learning. In fact, there is surprisingly little research on prompting considering the lengthy discussions and stronglystated opinions which appear in treatises on programed instruction (Anderson, 1967). Current conceptions of prompting seem to be based largely on Skinner's (1957) speculative analysis of verbal behavior and the practical experience of programers (Markle, 1964). Most of the actual research on prompting has involved paired associate lists. I t has been repeatedly demonstrated that people learn faster under a prompting procedure, in which both the stimulus term and response term appear before the response is required, than under the anticipation method, or confirmation method, as it has been called in these studies (Cook and Spitzer, 1960; Sidowski, 1961; Levine, 1965). On the basis of these experiments, Cook (1963) has argued that the student should be shown the correct answer before he makes the response. Cook seems to be suggesting that the ""copying frame"" is an especially effective sort of teaching device. The further implication is that there is no such condition as ""overprompting."" Our contention is that copying frames and, presumably, under some conditions, other kinds of strongly-prompted frames as well, do have a serious shortcoming. The following is an obvious copying frame from a program on the menstrual cycle (Biological Sciences Curriculum Study, 1965, p. 11):",,
-Strengthening scientific verbal behavior: an experimental comparison of progressively prompted and unprompted programmed instruction and prose tutorials.,Darrel R. Davis,"Web-based software was used to deliver and record the effects of programmed instruction that progressively added formal prompts until attempts were successful, programmed instruction with one attempt, and prose tutorials. Error-contingent progressive prompting took significantly longer than programmed instruction and prose. Both forms of programmed instruction substantially increased the appropriate use of behavioral vocabulary during subsequent interpretive essays. These behavioral gains extended to a different setting, suggesting that more was being learned than simply how to answer programmed tutorial frames correctly.","{'model': 'tldr@v2.0.0', 'text': 'Both forms of programmed instruction substantially increased the appropriate use of behavioral vocabulary during subsequent interpretive essays, suggesting that more was being learned than simply how to answer programmed tutorial frames correctly.'}",https://europepmc.org/articles/pmc1868811?pdf=render
-Promoting preservice teachers’ dual self-regulation roles as learners and as teachers: effects of generic vs. specific prompts,B. Kramarski,,,
-Composition Instruction with Learning Disabled Students: Self-Instructional Strategy Training.,S. Graham,"Writing is a particularly complex and demanding task. The development of skillful writing requires much more than simply adding special knowledge and skills to existing oral language abilities (Scardamalia & Bereiter ,, 1986). The beginning writer must develop the ability to generate language in the absence of a conversational partner, activate relevant memories without prompting, generate units of text larger than typically included in a conversational turn, and examine what is generated from the perspectives of both the writer and the reader (Bereiter & Scardamalia, 1983). Furthermore, the beginning writer must learn to monitor and regulate a variety of mental operations including the generation of ideas, development of plans for writing, and the actual production of text that corresponds to the conventions of standard written English. The developing writer, in sum, must learn to orchestrate a variety of cognitive activities, and attention must be allocated to both mechanical and substantive concerns in an efficient, effective manner (Bereiter & Scardamalia, 1982; Graham, 1982; Scardamalia & Bereiter, 1986). One way to help poor and beginning writers overcome the difficulties inherent in the composing process is to teach them task-specific and metacognitive strategies for regulating their writing behavior (Graham & Harris, 1987a; Harris & Graham, in press). Such an approach may be particularly helpful because strategy instruction can be employed as a means for helping students gain more security and competence in the processes considered central to effective writing (e.g., planning, revising). Furthermore, strategy training holds much promise as a means for helping students develop the skills necessary to carry out more mature and complex composing behaviors (Graham & Harris, in press-a, in press-b). Strategy and self-regulation training may be especially useful as a means for helping learning disabled (LD) students improve their writing performance. Two assumptions underlying strategy instruction are (Graham & Harris, in press-b; Harris, 1982):",,https://journals.ku.edu/focusXchild/article/download/7502/6835
-Evidence-Based Practice for Teaching Academics to Students With Severe Developmental Disabilities,F. Spooner,"A review of the literature was conducted for articles published between 2003 and 2010 to build a case for the degree to which evidence-based practices were documented for teaching academic skills to students with severe developmental disabilities. This review extended earlier comprehensive work in literacy, mathematics, and science for the population in question. A total of 18 studies met the Horner et al. (2005) quality indicator criteria. In general, time delay and task analytic instruction were found to be evidence-based practices. In addition, specific target responses were defined to show academic learning, with the most prevalent target responses being discrete responses; the type of systematic prompting and feedback used most often was time delay, while the component used least often was stimulus fading/shaping; and teaching formats used most often were massed trials and one-to-one instruction.",,
-Strategy Instruction Is Often Insufficient: Addressing the Interdependency of Executive and Attributional Processes,Andrea K. Groteluschen,,,
-Teaching Generalized Pretend Play and Related Behaviors to Young Children With Disabilities,E. Barton,"Children with disabilities play less often and demonstrate fewer varied pretend play behaviors than children with typical development. A multiple-probe design was used to examine the relation between teachers’ use of the system of least prompts and contingent imitation and the acquisition, maintenance, and generalization of pretend play and related behaviors by four children with disabilities. Results indicated the teachers’ use of the intervention package was functionally related to increases in the children’s frequency and diversity of pretend play and related behaviors. Children also maintained responses in sessions without prompts and generalized across toys and contexts. The findings replicate previous studies on adult prompting of pretend play and extend the literature by assessing generalization of children’s pretend play across contexts and measuring intervention and implementation fidelity. Overall, this study provides a strong argument for engaging in systematic instruction of play, including pretend play, for children who do not display such behaviors.",,
-Behavioral Ethics and Teaching Ethical Decision Making,M. Drumwright,"Business education often renders students less likely to act ethically. An infusion of liberal learning in the form of behavioral ethics could improve this situation by prompting students to develop higher levels of professionalism that encompass ethics, social responsibility, self-critical reflection, and personal accountability. More specifically, teaching behavioral ethics, which draws upon psychology, sociology, and related fields, can improve students’ ethical decision making in a manner that can lead to a more ethical climate in organizations and in society more generally. This article introduces key concepts of behavioral ethics, argues that teaching behavioral ethics can have a positive impact, discusses materials that can be used to teach those concepts, and addresses action-research approaches to assessing the effectiveness of the instruction. There is significant evidence, though preliminary and incomplete, that teaching behavioral ethics is a promising new approach for improving the ethicality of students’ decisions and actions.",,
-Use of Technology in Interventions for Children with Autism,T. Goldsmith,"A growing number of studies have investigated diverse applications of technology-based interventions with children with autism. The purpose of this paper is to review the growing empirical support for the efficacy of technology-based interventions with children with autism and to recommend future directions for research. This review will focus on five examples of technology introduced as a temporary instructional aid to be removed once the goal of behavior change has been met: (a) tactile and auditory prompting devices, (b) video-based instruction and feedback, (c) computer-aided instruction, (d) virtual reality, and (e) robotics. Future directions for research and practice with each technology are discussed. Keywords: autism; technology-based interventions; computer-aided instruction; virtual reality; robotics. ********** A growing number of studies have investigated diverse applications of technology-based interventions with children with autism. The popularity of technology in the field of psychology is evidenced by the development of new journals in the area (e.g., the Journal of Special Education Technology, the Journal of Educational Multimedia and Hypermedia, the Journal of Computer Assisted Learning, etc.). In addition, clinical psychology journals are recognizing the importance of technology in service delivery and devoting special issues to the topic (e.g., Newman, 2004). Parents and clinicians regularly report that children with autism are drawn to technological devices and researchers have noted the importance of devising treatments that take advantage of this fascination (Colby, 1973). The purpose of this paper is to review the growing empirical support for the efficacy of technology-based interventions with children with autism and to recommend future directions for research. A review of technology-based interventions must first clarify the use of the word technology. The term is broadly defined as ""the practical application of knowledge"" or ""the specialized aspects of a particular field of endeavor"" (Merriam-Webster's Collegiate Dictionary, 1994, p. 1210). This broad definition encompasses virtually any information or object that has been used in application to a field of study. For example, the application of stimulus control is technology to a behavioral clinician. A more specific use of the term refers to use of mechanical or electromechanical processes that often increase productivity and reduce or eliminate manual operations or operations done by older technologies. In the twenty-first century, technology commonly denotes a variety of popular electromechanical devices such as cell phones, video recording equipment, and hand-held, desktop, and laptop personal computers. It is the use of these tools with children with autism that will be the subject of this review paper. Some technology-based interventions are designed for indefinite use as an assistive tool (e.g., voice-output augmentative communication devices, microswitches, etc.) while others are introduced as a temporary instructional aid to be removed once the goal of behavior change has been met. This review will focus on five examples of the second category: tactile and auditory prompting devices, video-based instruction and feedback, computer-aided instruction, virtual reality, and robotics. Research support for the use of each technological advancement with children with autism will be reviewed, and directions for future research and practical application will be discussed. Mechanical Prompts Individuals with autism often need external stimulus prompts to initiate, maintain, or terminate a behavior. Commonly used prompts include vocal, gestural, physical, written/pictorial, and signed prompts, and each modality has been demonstrated effective for multiple purposes (MacDuff, Krantz, & McClannahan, 2001). Technological advancements in the last decade have created cost-effective automated prompting devices with the ability to deliver the same level of prompting with less human interaction and obtrusiveness and often less human effort in managing prompt delivery (e. …","{'model': 'tldr@v2.0.0', 'text': 'This review will focus on five examples of technology introduced as a temporary instructional aid to be removed once the goal of behavior change has been met: tactile and auditory prompting devices, video-based instruction and feedback, computer-aided instruction, virtual reality, and robotics.'}",
-Effects of Self-Explaining on Learning and Transfer of Critical Thinking Skills,Lara M. van Peppen,"Critical thinking is considered to be an important competence for students and graduates of higher education. Yet, it is largely unclear which teaching methods are most effective in supporting the acquisition of critical thinking skills, especially regarding one important aspect of critical thinking: avoiding biased reasoning. The present study examined whether creating desirable difficulties in instruction by prompting students to generate explanations of a problem-solution to themselves (i.e., self-explaining) is effective for fostering learning and transfer of unbiased reasoning. Seventy-nine first-year students of a Dutch Applied University of Sciences were first instructed on two categories of “heuristics and biases” tasks (syllogism and base-rate or Wason and conjunction). Thereafter, they practiced these either with (self-explaining condition) or without (no self-explaining condition) self-explanation prompts that asked them to motivate their answers. Performance was measured on a pretest, immediate posttest, and delayed (2 weeks later) posttest on all four task categories, to examine effects on learning (performance on practiced tasks) and transfer (performance on non-practiced tasks). Participants’ learning and transfer performance improved to a comparable degree from pretest to immediate posttest in both conditions, and this higher level of performance was retained on the delayed posttest. Surprisingly, self-explanation prompts had a negative effect on posttest performance on practiced tasks when those were Wason and conjunction tasks, and self-explaining had no effect on transfer performance. These findings suggest that the benefits of explicit instruction and practice on learning and transfer of unbiased reasoning cannot be enhanced by increasing the difficulty of the practice tasks through self-explaining.",,https://www.frontiersin.org/articles/10.3389/feduc.2018.00100/pdf
-"Feedback, Prompting, and Overt Correction Procedures in Nonbranching Computer Assisted Instructed Programs",D. Gilman,"AbstractNo differences in learning or retention were obtained for a computer assisted instruction program (CAI) which incorporated response contingent feedback, prompting, and overt correction procedures when compared to a CAI program which simply typed the correct response following a student response. No differences in learning or retention were obtained for a condition in which an instructional program was administered by a teletypewriter as compared to a condition in which the material was presented by means of programmed texts. Both conditions in which instruction was presented by a CAI communication device took significantly more instructional time than the programmed text condition.",,
-Improving students' critical thinking : Empirical support for explicit instructions combined with practice,A. Heijltjes,"This experiment investigated the impact of different types of critical thinking instruction and dispositions on bias in economics students' (N=141) reasoning performance. The following conditions were compared: (A) implicit instruction; (B) implicit instruction with practice; (C) implicit instruction with explicit instruction and practice; (D) implicit instruction with explicit instruction, practice, and self-explanation prompts; and (E) implicit instruction with explicit instruction, practice, and activation prompts. Results showed that explicit instruction combined with practice is required to improve critical thinking (i.e., conditions A/B’ and rest of the layers are initialized randomly (row-3). Initializing prompts with a similar template in all layers leads to lower performance suggesting that this is redundant as these prompts learn hierarchically different contextual concepts in different layers (row-1). However, complete random initialization of prompts provides competitive performance (row-2). For implementation, if the number of learnable prompts M = #P are less than the total tokens of initial prompt template, we convert the former M word embeddings of template with learnable prompts and consider the rest of word embeddings of prompt template as xed and use all token embeddings (learnable prompts + xed word tokens) as input to text encoder.","{'model': 'tldr@v2.0.0', 'text': 'For implementation, if the number of learnable prompts M = #P are less than the total tokens of initial prompt template, the authors convert the former M word embeddings of template with learnable prompt and consider the rest of word embedDings of prompt template as \ue000xed and use all token embeddeddings (learnable prompts + \ue001xed word tokens) as input to text encoder.'}",
-This Prompt is Measuring : Evaluating Bias Evaluation in Language Models,Seraphina Goldfarb-Tarrant,"Bias research in NLP seeks to analyse models for social biases, thus helping NLP practitioners uncover, measure, and mitigate social harms. We analyse the body of work that uses prompts and templates to assess bias in language models. We draw on a measurement modelling framework to create a taxonomy of attributes that capture what a bias test aims to measure and how that measurement is carried out. By applying this taxonomy to 90 bias tests, we illustrate qualitatively and quantitatively that core aspects of bias test conceptualisations and operationalisations are frequently unstated or ambiguous, carry implicit assumptions, or be mismatched. Our analysis illuminates the scope of possible bias types the field is able to measure, and reveals types that are as yet under-researched. We offer guidance to enable the community to explore a wider section of the possible bias space, and to better close the gap between desired outcomes and experimental design, both for bias and for evaluating language models more broadly.",,http://arxiv.org/pdf/2305.12757
-Prompt Tuning with Contradictory Intentions for Sarcasm Recognition,Yiyi Liu,"Recently, prompt tuning has achieved promising results in a variety of natural language processing (NLP) tasks. The typical approach is to insert text pieces (i.e. templates) into the input and transform downstream tasks into the same form as pre-training. In essence, a high-quality template is the foundation of prompt tuning to support the performance of the converted cloze-style task. However, for sarcasm recognition, it is time-consuming and requires increasingly sophisticated domain knowledge to determine the appropriate templates and label words due to its highly figurative nature. In this work, we propose SarcPrompt, to incorporate the prior knowledge about contradictory intentions into prompt tuning for sarcasm recognition. SarcPrompt is inspired by that the speaker usually says the opposite of what they actually mean in the sarcastic text. Based on this idea, we explicitly mimic the actual intention by prompt construction and indicate whether the actual intention is contradictory to the literal content by verbalizer engineering. Experiments on three public datasets with standard and low-resource settings demonstrate the effectiveness of our SarcPrompt for sarcasm recognition.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes SarcPrompt, to incorporate the prior knowledge about contradictory intentions into prompt tuning for sarcasm recognition, inspired by that the speaker usually says the opposite of what they actually mean in the sarcastic text.'}",https://aclanthology.org/2023.eacl-main.25.pdf
-Grammar Correction for Multiple Errors in Chinese Based on Prompt Templates,Zhici Wang,"Grammar error correction (GEC) is a crucial task in the field of Natural Language Processing (NLP). Its objective is to automatically detect and rectify grammatical mistakes in sentences, which possesses immense application research value. Currently, mainstream grammar-correction methods primarily rely on sequence labeling and text generation, which are two kinds of end-to-end methods. These methods have shown exemplary performance in areas with low error density but often fail to deliver satisfactory results in high-error density situations where multiple errors exist in a single sentence. Consequently, these methods tend to overcorrect correct words, leading to a high rate of false positives. To address this issue, we researched the specific characteristics of the Chinese grammar error correction (CGEC) task in high-error density situations. We proposed a grammar-correction method based on prompt templates. Firstly, we proposed a strategy for constructing prompt templates suitable for CGEC. This strategy transforms the CGEC task into a masked fill-in-the-blank task compatible with the masked language model BERT. Secondly, we proposed a method for dynamically updating templates, which incorporates already corrected errors into the template through dynamic updates to improve the template quality. Moreover, we used the phonetic and graphical resemblance knowledge from the confusion set as guiding information. By combining this with BERT’s prediction results, the model can more accurately select the correct characters, significantly enhancing the accuracy of the model’s prediction correction results. Our methods were validated through experiments on a public grammar-correction dataset. The results indicate that our method achieves higher correction performance and lower false correction rates in high-error density scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This work proposed a grammar-correction method based on prompt templates that achieves higher correction performance and lower false correction rates in high-error density scenarios and used the phonetic and graphical resemblance knowledge from the confusion set as guiding information.'}",https://www.mdpi.com/2076-3417/13/15/8858/pdf?version=1690869487
-PPM: prompt-free prompt-tuning for multi-task learning,Zihan Zeng,"The cost of fine-tuning has increased significantly in recent years as the size of language model parameters has increased. Prompt-tuning and adapters have made it possible to train models with a small number of parameters to obtain results similar to those of fine-tuning methods. However, most of the current prompt-tuning methods require the help of hand-crafted templates and verbalizers to achieve outstanding results in few-shot learning. In this work, we propose PPM, Prompt-free prompt-tuning for multi-task learning. First, we insert the task-specific adapter into the pre-trained language model to replace the hand-designed external template. Then, we train each adapter separately on different tasks and adjust the parameters of each adapter layer. Next, we combine the different adapters and draw on their valid knowledge by tuning the parameters of the fusion part to get the smallest loss function in the process of extracting knowledge from different adapters. To boost the training speed, we use Post-LN to replace Pre-LN, which switched the position of the Laynorm layer in the model from after the two Addition layers to before the FFN layer and the Multi-head Attention layer. Experimental results on different NLP tasks show that our model has better synergistic effects on diverse types of downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'PPM, Prompt-free prompt-tuning for multi-task learning, which inserts the task-specific adapter into the pre-trained language model to replace the hand-designed external template and combines the different adapters to draw on their valid knowledge.'}",
-TEPrompt: Task Enlightenment Prompt Learning for Implicit Discourse Relation Recognition,Wei Xiang,"Implicit Discourse Relation Recognition (IDRR) aims at classifying the relation sense between two arguments without an explicit connective. Recently, the ConnPrompt~\cite{Wei.X:et.al:2022:COLING} has leveraged the powerful prompt learning for IDRR based on the fusion of multi-prompt decisions from three different yet much similar connective prediction templates. Instead of multi-prompt ensembling, we propose to design auxiliary tasks with enlightened prompt learning for the IDRR task. Although an auxiliary task is not used to directly output final prediction, we argue that during the joint training some of its learned features can be useful to boost the main task. In light of such motivations, we propose a task enlightenment prompt learning model, called TEPrompt, to fuse learned features from three related tasks for IDRR. In particular, the TEPrompt contains three tasks, viz., Discourse Relation Recognition (DRR), Sense Semantics Classification (SSC) and Annotated Connective Prediction (ACP), each with a unique prompt template and an answer space. In the training phase, we jointly train three prompt learning tasks with shared argument representation. In the testing phase, we only take the DRR output with fused features as the final IDRR decision. Experiments with the same conditions have shown that the proposed TEPrompt outperforms the ConnPrompt. This can be attributed to the promoted decision features and language models benefited from joint-training of auxiliary tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a task enlightenment prompt learning model, called TEPrompt, to fuse learned features from three related tasks for IDRR, and shows that the proposed TEP Prompt outperforms the ConnPrompt.'}",http://arxiv.org/pdf/2305.10866
-Adapting Prompt for Few-shot Table-to-Text Generation,Zhixin Guo,"Pretrained language models (PLMs) have made remarkable progress in table-to-text generation tasks. However, the lack of domain-specific knowledge makes it challenging to bridge the topological gap between tabular data and text, especially in real-world applications with limited resources. To mitigate the limitation of insufficient labeled data, we propose a novel framework: Adapt-Prompt-to-Generate (AdaPTGen). The core insight of AdaPTGen is to adapt prompt templates of domain-specific knowledge into the model, which brings at least three benefits: (1) it injects representation of normal table-related descriptions to bridge the topological gap between tabular data and texts; (2) it enables us to use large amounts of unlabeled domain-specific knowledge fully, which can alleviate the PLMs' inherent shortcomings of lacking domain knowledge; (3) it allows us to design various tasks to explore the domain-specific knowledge. Extensive experiments and analyses are conducted on three open-domain few-shot natural language generation (NLG) data sets: Humans, Songs, and Books. Compared to previous state-of-the-art approaches, our model achieves superior performance in terms of both fluency and accuracy.","{'model': 'tldr@v2.0.0', 'text': ""The core insight of AdaPTGen is to adapt prompt templates of domain-specific knowledge into the model, which brings at least three benefits: it injects representation of normal table-related descriptions to bridge the topological gap between tabular data and texts, which can alleviate the PLMs' inherent shortcomings of lacking domain knowledge.""}",
-Multi-Prompt with Depth Partitioned Cross-Modal Learning,Yiqi Wang,"In recent years, soft prompt learning methods have been proposed to fine-tune large-scale vision-language pre-trained models for various downstream tasks. These methods typically combine learnable textual tokens with class tokens as input for models with frozen parameters. However, they often employ a single prompt to describe class contexts, failing to capture categories' diverse attributes adequately. This study introduces the Partitioned Multi-modal Prompt (PMPO), a multi-modal prompting technique that extends the soft prompt from a single learnable prompt to multiple prompts. Our method divides the visual encoder depths and connects learnable prompts to the separated visual depths, enabling different prompts to capture the hierarchical contextual depths of visual representations. Furthermore, to maximize the advantages of multi-prompt learning, we incorporate prior information from manually designed templates and learnable multi-prompts, thus improving the generalization capabilities of our approach. We evaluate the effectiveness of our approach on three challenging tasks: new class generalization, cross-dataset evaluation, and domain generalization. For instance, our method achieves a $79.28$ harmonic mean, averaged over 11 diverse image recognition datasets ($+7.62$ compared to CoOp), demonstrating significant competitiveness compared to state-of-the-art prompting methods.","{'model': 'tldr@v2.0.0', 'text': 'This study introduces the Partitioned Multi-modal Prompt (PMPO), a multi- modal prompting technique that extends the soft prompt from a single learnable prompt to multiple prompts, and incorporates prior information from manually designed templates and learnable multi-prompts, thus improving the generalization capabilities of the approach.'}",https://arxiv.org/pdf/2305.06221
-COVER: A Heuristic Greedy Adversarial Attack on Prompt-based Learning in Language Models,Zihao Tan,"Prompt-based learning has been proved to be an effective way in pre-trained language models (PLMs), especially in low-resource scenarios like few-shot settings. However, the trustworthiness of PLMs is of paramount significance and potential vulnerabilities have been shown in prompt-based templates that could mislead the predictions of language models, causing serious security concerns. In this paper, we will shed light on some vulnerabilities of PLMs, by proposing a prompt-based adversarial attack on manual templates in black box scenarios. First of all, we design character-level and word-level heuristic approaches to break manual templates separately. Then we present a greedy algorithm for the attack based on the above heuristic destructive approaches. Finally, we evaluate our approach with the classification tasks on three variants of BERT series models and eight datasets. And comprehensive experimental results justify the effectiveness of our approach in terms of attack success rate and attack speed.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a prompt-based adversarial attack on manual templates in black box scenarios by designing character-level and word-level heuristic approaches to break manual templates separately and presents a greedy algorithm for the attack based on the above heuristic destructive approaches.'}",https://arxiv.org/pdf/2306.05659
-Self-adaptive Prompt-tuning for Event Extraction in Ancient Chinese Literature,Jingyi Zhang,"Extracting different types of war events from ancient Chinese literature is significant, as war is an important factor in driving the development of Chinese history. The existing trend of event extraction models utilizes template-based generative approaches, which do not take into account the brevity and obscurity of ancient Chinese, as well as the diversity of templates for similar event types. In this paper, we propose a novel Knowledge Graph-based generative event extraction framework with a self-Adaptive Prompt (KGAP) for ancient Chinese war. Specifically, we construct a self-adaptive prompt, which considers its unique trigger words for different types of wars and is designed to solve the problem of the similarity in events. Moreover, we construct a semantic knowledge graph of ancient literature, assisting the pre-trained language model to better understand the ancient Chinese text. Since there is no public dataset for the ancient Chinese event extraction task, we provide an event extraction dataset and conduct experiments on it. Experimental results show that our model is more state-of-the-art than both the classification-based and generative-based methods for event extraction in ancient Chinese literature.","{'model': 'tldr@v2.0.0', 'text': 'A novel Knowledge Graph-based generative event extraction framework with a self-Adaptive Prompt (KGAP) for ancient Chinese war and a semantic knowledge graph of ancient literature is constructed, assisting the pre-trained language model to better understand the ancient Chinese text.'}",
-SPTNET: Span-based Prompt Tuning for Video Grounding,Yiren Zhang,"When a Pre-trained Language Model (PLM) is adopted in video grounding task, it usually acts as a text encoder without having its knowledge fully utilized. Also, there exists an inconsistency problem between the pre-training and downstream objectives. To solve the issues, we propose a new paradigm, named Span-based Prompt Tuning (SPTNet). It can convert the video grounding task into a cloze form. Specifically, a query is first changed into a form with mask token by a template, then the video and the query embeddings are integrated through a cross-modal transformer. The start and end points of the query matching time span are predicted with the embedding of the mask token. Experimental results on two public benchmarks ActivityNet Captions and Charades-STA show that our SPTNet achieves surpassing performance compared with state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A new paradigm, named Span-based Prompt Tuning (SPTNet), which can convert the video grounding task into a cloze form and achieves surpassing performance compared with state-of-the-art methods.'}",
-Prompt Cache: Modular Attention Reuse for Low-Latency Inference,In Gim,"We present Prompt Cache, an approach for accelerating inference for large language models (LLM) by reusing attention states across different LLM prompts. Many input prompts have overlapping text segments, such as system messages, prompt templates, and documents provided for context. Our key insight is that by precomputing and storing the attention states of these frequently occurring text segments on the inference server, we can efficiently reuse them when these segments appear in user prompts. Prompt Cache employs a schema to explicitly define such reusable text segments, called prompt modules. The schema ensures positional accuracy during attention state reuse and provides users with an interface to access cached states in their prompt. Using a prototype implementation, we evaluate Prompt Cache across several LLMs. We show that Prompt Cache significantly reduce latency in time-to-first-token, especially for longer prompts such as document-based question answering and recommendations. The improvements range from 8x for GPU-based inference to 60x for CPU-based inference, all while maintaining output accuracy and without the need for model parameter modifications.","{'model': 'tldr@v2.0.0', 'text': None}",
-Low-Resource Multi-Granularity Academic Function Recognition Based on Multiple Prompt Knowledge,Jiawei Liu,"Fine-tuning pre-trained language models (PLMs), e.g., SciBERT, generally requires large numbers of annotated data to achieve state-of-the-art performance on a range of NLP tasks in the scientific domain. However, obtaining the fine-tune data for scientific NLP task is still challenging and expensive. Inspired by recent advancement in prompt learning, in this paper, we propose the Mix Prompt Tuning (MPT), which is a semi-supervised method to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks with a small number of labeled examples. Specifically, the proposed method provides multi-perspective representations by combining manual prompt templates with automatically learned continuous prompt templates to help the given academic function recognition task take full advantage of knowledge in PLMs. Based on these prompt templates and the fine-tuned PLM, a large number of pseudo labels are assigned to the unlabeled examples. Finally, we fine-tune the PLM using the pseudo training set. We evaluate our method on three academic function recognition tasks of different granularity including the citation function, the abstract sentence function, and the keyword function, with datasets from computer science domain and biomedical domain. Extensive experiments demonstrate the effectiveness of our method and statistically significant improvements against strong baselines. In particular, it achieves an average increase of 5% in Macro-F1 score compared with fine-tuning, and 6% in Macro-F1 score compared with other semi-supervised method under low-resource settings. In addition, MPT is a general method that can be easily applied to other low-resource scientific classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Mix Prompt Tuning (MPT) is a semi-supervised method to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks with a small number of labeled examples.'}",http://arxiv.org/pdf/2305.03287
-PromptCL: Improving Event Representation via Prompt Template and Contrastive Learning,Yubo Feng,,,
-Learning Historical Status Prompt for Accurate and Robust Visual Tracking,Wenrui Cai,"Most trackers perform template and search region similarity matching to find the most similar object to the template during tracking. However, they struggle to make prediction when the target appearance changes due to the limited historical information introduced by roughly cropping the current search region based on the predicted result of previous frame. In this paper, we identify that the central impediment to improving the performance of existing trackers is the incapacity to integrate abundant and effective historical information. To address this issue, we propose a Historical Information Prompter (HIP) to enhance the provision of historical information. We also build HIPTrack upon HIP module. HIP is a plug-and-play module that make full use of search region features to introduce historical appearance information. It also incorporates historical position information by constructing refined mask of the target. HIP is a lightweight module to generate historical information prompts. By integrating historical information prompts, HIPTrack significantly enhances the tracking performance without the need to retrain the backbone. Experimental results demonstrate that our method outperforms all state-of-the-art approaches on LaSOT, LaSOT ext, GOT10k and NfS. Futhermore, HIP module exhibits strong generality and can be seamlessly integrated into trackers to improve tracking performance. The source code and models will be released for further research.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Historical Information Prompter (HIP) to enhance the provision of historical information and builds HIPTrack upon HIP module, a plug-and-play module that make full use of search region features to introduce historical appearance information.'}",
-Prompt template construction by Average Gradient Search with External Knowledge for aspect sentimental analysis,Yongping Du,,,
-Teaching Story Prompt Writing Through WH-Questions Cube among Year 5 Students,Daisy Kunang,"Today’s education field in Malaysia has undergone various transformations in line with progressive global educational trends. Henceforth, revolution of teaching and learning styles, methods and approaches are essential to ensure that students’ learning, skills and thinking are optimized and concurrently nurture motivated and confident learners starting from primary school level. Numerous studies on teaching and learning writing had highlighted that it is one of the most challenging skills to be mastered in English language. With respect to this, this study aims to explore the necessity of incorporating WH-Foldable Cube to facilitate expansion of ideas in teaching story prompt lessons. WH-Foldable Cube is a writing learning strategy which emphasised on expansion of ideas through the use of question words to facilitate generation of ideas in writing story prompt. Expansion of ideas becomes less strenuous through students’ engagement with the question words that are used in the WH-Foldable Cube template to stimulate their thinking about the pictures stimulus given. The data were collected using writing tasks and interview which were administered among five Year 5 students in SK Parish, a suburban school in Balingian district. The findings of the study revealed that majority of the respondents agreed that the use of WH-Foldable Cube had positive impacts towards their story prompt writing to generate and expand their ideas. This research also highlighted the potential use of WH-Foldable Cube to expand students’ motivation and confidence in their learning.",,
-Activation of Actuating Hydrogels with WS2 Nanosheets for Biomimetic Cellular Structures and Steerable Prompt Deformation.,Lu Zong,"Macroscopic soft actuation is intrinsic to living organisms in nature, including slow deformation (e.g., contraction, bending, twisting, and curling) of plants motivated by microscopic swelling and shrinking of cells, and rapid motion of animals (e.g., deformation of jellyfish) motivated by cooperative nanoscale movement of motor proteins. These actuation behaviors, with an exceptional combination of tunable speed and programmable deformation direction, inspire us to design artificial soft actuators for broad applications in artificial muscles, nanofabrication, chemical valves, microlenses, soft robotics, etc. However, so far artificial soft actuators have been typically produced on the basis of poly(N-isopropylacrylamide) (PNiPAM), whose deformation is motived by volumetric shrinkage and swelling in analogue to plant cells, and exhibits sluggish actuation kinetics. In this study, alginate-exfoliated WS2 nanosheets were incorporated into ice-template-polymerized PNiPAM hydrogels with the cellular microstructures which mimic plant cells, yet the prompt steerable actuation of animals. Because of the nanosheet-reinforced pore walls formed in situ in freezing polymerization and reasonable hierarchical water channels, this cellular hybrid hydrogel achieves super deformation speed (on the order of magnitude of 10° s), controllable deformation direction, and high near-infrared light responsiveness, offering an unprecedented platform of artificial muscles for various soft robotics and devices (e.g., rotator, microvalve, aquatic swimmer, and water-lifting filter).","{'model': 'tldr@v2.0.0', 'text': 'This cellular hybrid hydrogel achieves super deformation speed (on the order of magnitude of 10° s), controllable deformation direction, and high near-infrared light responsiveness, offering an unprecedented platform of artificial muscles for various soft robotics and devices.'}",https://doi.org/10.1021/acsami.7b10348.s001
-The effects of implementing a point-of-care electronic template to prompt routine anxiety and depression screening in patients consulting for osteoarthritis (the Primary Care Osteoarthritis Trial): A cluster randomised trial in primary care,C. Mallen,"Background This study aimed to evaluate whether prompting general practitioners (GPs) to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis (OA) improves pain outcomes. Methods and findings We conducted a cluster randomised controlled trial involving 45 English general practices. In intervention practices, patients aged ≥45 y consulting with OA received point-of-care anxiety and depression screening by the GP, prompted by an automated electronic template comprising five questions (a two-item Patient Health Questionnaire–2 for depression, a two-item Generalized Anxiety Disorder–2 questionnaire for anxiety, and a question about current pain intensity [0–10 numerical rating scale]). The template signposted GPs to follow National Institute for Health and Care Excellence clinical guidelines for anxiety, depression, and OA and was supported by a brief training package. The template in control practices prompted GPs to ask the pain intensity question only. The primary outcome was patient-reported current pain intensity post-consultation and at 3-, 6-, and 12-mo follow-up. Secondary outcomes included pain-related disability, anxiety, depression, and general health. During the trial period, 7,279 patients aged ≥45 y consulted with a relevant OA-related code, and 4,240 patients were deemed potentially eligible by participating GPs. Templates were completed for 2,042 patients (1,339 [31.6%] in the control arm and 703 [23.1%] in the intervention arm). Of these 2,042 patients, 1,412 returned questionnaires (501 [71.3%] from 20 intervention practices, 911 [68.0%] from 24 control practices). Follow-up rates were similar in both arms, totalling 1,093 (77.4%) at 3 mo, 1,064 (75.4%) at 6 mo, and 1,017 (72.0%) at 12 mo. For the primary endpoint, multilevel modelling yielded significantly higher average pain intensity across follow-up to 12 mo in the intervention group than the control group (adjusted mean difference 0.31; 95% CI 0.04, 0.59). Secondary outcomes were consistent with the primary outcome measure in reflecting better outcomes as a whole for the control group than the intervention group. Anxiety and depression scores did not reduce following the intervention. The main limitations of this study are two potential sources of bias: an imbalance in cluster size (mean practice size 7,397 [intervention] versus 5,850 [control]) and a difference in the proportion of patients for whom the GP deactivated the template (33.6% [intervention] versus 27.8% [control]). Conclusions In this study, we observed no beneficial effect on pain outcomes of prompting GPs to routinely screen for and manage comorbid anxiety and depression in patients presenting with symptoms due to OA, with those in the intervention group reporting statistically significantly higher average pain scores over the four follow-up time points than those in the control group. Trial registration ISRCTN registry ISRCTN40721988","{'model': 'tldr@v2.0.0', 'text': 'No beneficial effect on pain outcomes was observed of prompting GPs to routinely screen for and manage comorbid anxiety and depression in patients presenting with symptoms due to OA, with those in the intervention group reporting statistically significantly higher average pain scores over the four follow-up time points than those inThe control group.'}",https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.1002273&type=printable
-Templating Three-Dimensional Self-Assembled Structures in Bilayer Block Copolymer Films,Tavakkoli K G,"To the Next Level Block copolymers will spontaneously separate into a range of microstructures that depend on the polymer block lengths and chemical compositions, and have been used as a templating material because one can selectively etch or functionalize one of the blocks. However, creating a template that is more than one layer thick is challenging. Tavakkoli K. G. et al. (p. 1294) used an array of posts to provide independent and simultaneous control of the morphology and orientation of two layers of block copolymers and were able to create local variations in the curvature and spacing of the domains. An array of posts guides the bilayer assembly of block copolymers with independent control of morphology and orientation. The registration and alignment of a monolayer of microdomains in a self-assembled block copolymer thin film can be controlled by chemical or physical templating methods. Although planar patterns are useful for nanoscale device fabrication, three-dimensional multilevel structures are required for some applications. We found that a bilayer film of a cylindrical-morphology block copolymer, templated by an array of posts functionalized with a brush attractive to the majority block, can form a rich variety of three-dimensional structures consisting of cylinder arrays with controllable angles, bends, and junctions whose geometry is controlled by the template periodicity and arrangement. This technique allows control of microdomain patterns and the ability to route and connect microdomains in specific directions.","{'model': 'tldr@v2.0.0', 'text': 'A bilayer film of a cylindrical-morphology block copolymer, templated by an array of posts functionalized with a brush attractive to the majority block, can form a rich variety of three-dimensional structures consisting of cylinder arrays with controllable angles, bends, and junctions whose geometry is controlled by the template periodicity and arrangement.'}",https://dspace.mit.edu/bitstream/1721.1/86080/1/Berggren_Templating%20three.pdf
-Surface-templated hydrogel patterns prompt matrix-dependent migration of breast cancer cells towards chemokine-secreting cells.,Taisuke Kojima,,"{'model': 'tldr@v2.0.0', 'text': 'A novel technique for fabricating spatially defined cell-laden collagen hydrogels using patterned, non-adhesive polyacrylamide-coated polydimethylsiloxane (PDMS) surfaces as a template revealed matrix-dependent and chemokine isoform-dependent migration of cancer cells.'}",https://europepmc.org/articles/pmc4293228?pdf=render
-PROMPT: A TEMPLATE SYSTEM FOR SECOND LANGUAGE READING COMPREHENSION,Donna Mydlarski,The following article briefly summarizes some different eras of CALL and suggests that the present era could be of immense help and usefulness to language teachers because of the multiplicity of template (miniauthoring)systems which are becoming available. This article also describes PROMPT which is a template for two types of reading comprehension exercises developed at the universities of Calgary and Guelph in Canada.,"{'model': 'tldr@v2.0.0', 'text': 'The following article briefly summarizes some different eras of CALL and suggests that the present era could be of immense help and usefulness to language teachers because of the multiplicity of template (miniauthoring) systems which are becoming available.'}",
-Ground-Truth Labels Matter: A Deeper Look into Input-Label Demonstrations,Junyeob Kim,"Despite recent explosion of interests in in-context learning, the underlying mechanism and the precise impact of the quality of demonstrations remain elusive.Intuitively, ground-truth labels should have as much impact in in-context learning (ICL) as supervised learning, but recent work reported that the input-label correspondence is significantly less important than previously thought.Intrigued by this counter-intuitive observation, we re-examine the importance of ground-truth labels in in-context learning.With the introduction of two novel metrics, namely Label-Correctness Sensitivity and Ground-truth Label Effect Ratio (GLER), we were able to conduct quantifiable analysis on the impact of ground-truth label demonstrations.Through extensive analyses, we find that the correct input-label mappings can have varying impacts on the downstream in-context learning performances, depending on the experimental configuration.Through additional studies, we identify key components, such as the verbosity of prompt templates and the language model size, as the controlling factor to achieve more noise-resilient ICL.","{'model': 'tldr@v2.0.0', 'text': 'Through extensive analyses, it is found that the correct input-label mappings can have varying impacts on the downstream in-context learning performances, depending on the experimental configuration.'}",http://arxiv.org/pdf/2205.12685
-"Method, device, client terminal and system for realizing commodity state change prompt",曹晴,"The invention provides a method, a device, a client terminal and a system for realizing a commodity state change prompt. The method comprises the following steps: a template used for collecting a commodity on a website is installed on the client terminal; when a user visits a commodity webpage, information in the commodity webpage is detected according to the template, and a collecting prompt sign is displayed in the commodity webpage when information in the commodity webpage is matched with all information in the template; when a triggering signal of the collecting prompt sign is received, all corresponding information of the commodity is extracted from the commodity webpage according to the template; and when price information in all corresponding information of the commodity is changed through detection, a commodity state change prompt message is sent to the user. Price information of each collected commodity is monitored, and more flexible and intelligent shopping information is provided for the user so that a shopping selection with higher cost performance is conveniently made by the user.","{'model': 'tldr@v2.0.0', 'text': 'Price information of each collected commodity is monitored, and more flexible and intelligent shopping information is provided for the user so that a shopping selection with higher cost performance is conveniently made by the user.'}",
-Not All Languages Are Created Equal in LLMs: Improving Multilingual Capability by Cross-Lingual-Thought Prompting,Haoyang Huang,"Large language models (LLMs) demonstrate impressive multilingual capability, but their performance varies substantially across different languages. In this work, we introduce a simple yet effective method, called cross-lingual-thought prompting (XLT), to systematically improve the multilingual capability of LLMs. Specifically, XLT is a generic template prompt that stimulates cross-lingual and logical reasoning skills to enhance task performance across languages. We conduct comprehensive evaluations on 7 typical benchmarks related to reasoning, understanding, and generation tasks, covering both high-resource and low-resource languages. Experimental results show that XLT not only remarkably enhances the performance of various multilingual tasks but also significantly reduces the gap between the average performance and the best performance of each task in different languages. Notably, XLT brings over 10 points of average improvement in arithmetic reasoning and open-domain question-answering tasks.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that XLT not only remarkably enhances the performance of various multilingual tasks but also significantly reduces the gap between the average performance and the best performance of each task in different languages.'}",http://arxiv.org/pdf/2305.07004
-UniHD at TSAR-2022 Shared Task: Is Compute All We Need for Lexical Simplification?,Dennis Aumiller,"Previous state-of-the-art models for lexical simplification consist of complex pipelines with several components, each of which requires deep technical knowledge and fine-tuned interaction to achieve its full potential. As an alternative, we describe a frustratingly simple pipeline based on prompted GPT-3 responses, beating competing approaches by a wide margin in settings with few training instances. Our best-performing submission to the English language track of the TSAR-2022 shared task consists of an “ensemble” of six different prompt templates with varying context levels. As a late-breaking result, we further detail a language transfer technique that allows simplification in languages other than English. Applied to the Spanish and Portuguese subset, we achieve state-of-the-art results with only minor modification to the original prompts. Aside from detailing the implementation and setup, we spend the remainder of this work discussing the particularities of prompting and implications for future work. Code for the experiments is available online at https://github.com/dennlinger/TSAR-2022-Shared-Task.","{'model': 'tldr@v2.0.0', 'text': 'A language transfer technique that allows simplification in languages other than English and a frustratingly simple pipeline based on prompted GPT-3 responses is described, beating competing approaches by a wide margin in settings with few training instances.'}",http://arxiv.org/pdf/2301.01764
-The utility of ChatGPT for cancer treatment information,Shan Chen,"The use of large language models (LLMs) such as ChatGPT for medical question-answering is becoming increasingly popular. However, there are concerns that these models may generate and amplify medical misinformation. Because cancer patients frequently seek to educate themselves through online resources, some individuals will likely use ChatGPT to obtain cancer treatment information. This study evaluated the performance and robustness of ChatGPT in providing breast, prostate, and lung cancer treatment recommendations that align with National Comprehensive Cancer Network (NCCN) guidelines. Four prompt templates were created to explore how differences in how the query is posed impacts response. ChatGPT output was scored by 3 oncologists and a 4th oncologist adjudicated in cases of disagreement. ChatGPT provided at least one NCCN-concordant recommendation for 102/104 (98%) prompts. However, 35/102 (34.3%) of these also included a recommendation that was at least partially non-concordant with NCCN guidelines. Responses varied based on prompt type. In conclusion, ChatGPT did not perform well at reliably and robustly providing cancer treatment recommendations. Patients and clinicians should be aware of the limitations of ChatGPT and similar technologies for self-education.","{'model': 'tldr@v2.0.0', 'text': 'Patients and clinicians should be aware of the limitations of ChatGPT and similar technologies for self-education because cancer patients frequently seek to educate themselves through online resources and some individuals will likely useChatGPT to obtain cancer treatment information.'}",https://www.medrxiv.org/content/medrxiv/early/2023/03/23/2023.03.16.23287316.full.pdf
-AvatarCraft: Transforming Text into Neural Human Avatars with Parameterized Shape and Pose Control,Ruixia Jiang,"Neural implicit fields are powerful for representing 3D scenes and generating high-quality novel views, but it remains challenging to use such implicit representations for creating a 3D human avatar with a specific identity and artistic style that can be easily animated. Our proposed method, AvatarCraft, addresses this challenge by using diffusion models to guide the learning of geometry and texture for a neural avatar based on a single text prompt. We carefully design the optimization framework of neural implicit fields, including a coarse-to-fine multi-bounding box training strategy, shape regularization, and diffusion-based constraints, to produce high-quality geometry and texture. Additionally, we make the human avatar animatable by deforming the neural implicit field with an explicit warping field that maps the target human mesh to a template human mesh, both represented using parametric human models. This simplifies animation and reshaping of the generated avatar by controlling pose and shape parameters. Extensive experiments on various text descriptions show that AvatarCraft is effective and robust in creating human avatars and rendering novel views, poses, and shapes. Our project page is: https://avatar-craft.github.io/.","{'model': 'tldr@v2.0.0', 'text': 'The proposed method, AvatarCraft, addresses the challenge of creating a 3D human avatar with a specific identity and artistic style that can be easily animated by using diffusion models to guide the learning of geometry and texture for a neural avatar based on a single text prompt.'}",https://arxiv.org/pdf/2303.17606
-TrustGPT: A Benchmark for Trustworthy and Responsible Large Language Models,Yue Huang,"Large Language Models (LLMs) such as ChatGPT, have gained significant attention due to their impressive natural language processing capabilities. It is crucial to prioritize human-centered principles when utilizing these models. Safeguarding the ethical and moral compliance of LLMs is of utmost importance. However, individual ethical issues have not been well studied on the latest LLMs. Therefore, this study aims to address these gaps by introducing a new benchmark -- TrustGPT. TrustGPT provides a comprehensive evaluation of LLMs in three crucial areas: toxicity, bias, and value-alignment. Initially, TrustGPT examines toxicity in language models by employing toxic prompt templates derived from social norms. It then quantifies the extent of bias in models by measuring quantifiable toxicity values across different groups. Lastly, TrustGPT assesses the value of conversation generation models from both active value-alignment and passive value-alignment tasks. Through the implementation of TrustGPT, this research aims to enhance our understanding of the performance of conversation generation models and promote the development of language models that are more ethical and socially responsible.","{'model': 'tldr@v2.0.0', 'text': 'This study aims to enhance the understanding of the performance of conversation generation models and promote the development of language models that are more ethical and socially responsible by introducing a new benchmark -- TrustGPT.'}",http://arxiv.org/pdf/2306.11507
-Visual Prompting for Adversarial Robustness,Aochuan Chen,"In this work, we leverage visual prompting (VP) to improve adversarial robustness of a fixed, pre-trained model at test time. Compared to conventional adversarial defenses, VP allows us to design universal (i.e., data-agnostic) input prompting templates, which have plug-and-play capabilities at test time to achieve desired model performance without introducing much computation overhead. Although VP has been successfully applied to improving model generalization, it remains elusive whether and how it can be used to defend against adversarial attacks. We investigate this problem and show that the vanilla VP approach is not effective in adversarial defense since a universal input prompt lacks the capacity for robust learning against sample-specific adversarial perturbations. To circumvent it, we propose a new VP method, termed Class-wise Adversarial Visual Prompting (C-AVP), to generate class-wise visual prompts so as to not only leverage the strengths of ensemble prompts but also optimize their interrelations to improve model robustness. Our experiments show that C-AVP outperforms the conventional VP method, with 2.1× standard accuracy gain and 2× robust accuracy gain. Compared to classical test-time defenses, C-AVP also yields a 42× inference time speedup. Code is available at https://github.com/Phoveran/vp-for-adversarial-robustness.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new VP method, termed Class-wise Adversarial Visual Prompting (C-AVP), to generate class-wise visual prompts so as to not only leverage the strengths of ensemble prompts but also optimize their interrelations to improve model robustness.'}",https://arxiv.org/pdf/2210.06284
-Using Natural Sentence Prompts for Understanding Biases in Language Models,Sarah Alnegheimish,"Evaluation of biases in language models is often limited to synthetically generated datasets. This dependence traces back to the need of prompt-style dataset to trigger specific behaviors of language models. In this paper, we address this gap by creating a prompt dataset with respect to occupations collected from real-world natural sentences present in Wikipedia.We aim to understand the differences between using template-based prompts and natural sentence prompts when studying gender-occupation biases in language models. We find bias evaluations are very sensitiveto the design choices of template prompts, and we propose using natural sentence prompts as a way of more systematically using real-world sentences to move away from design decisions that may bias the results.","{'model': 'tldr@v2.0.0', 'text': 'A prompt dataset with respect to occupations collected from real-world natural sentences present in Wikipedia is created to understand the differences between using template-based prompts and natural sentence prompts when studying gender-occupation biases in language models.'}",https://arxiv.org/pdf/2205.06303
-An innovative dual recognition aptasensor for specific detection of Staphylococcus aureus based on Au/Fe3O4 binary hybrid,Mohamed M. El-wekil,,,https://www.nature.com/articles/s41598-022-15637-1.pdf
-Domain Knowledge Matters: Improving Prompts with Fix Templates for Repairing Python Type Errors,Yun Peng,"Although the dynamic type system of Python facilitates the developers in writing Python programs, it also brings type errors at run-time. There exist rule-based approaches for automatically repairing Python type errors. The approaches can generate accurate patches but they require domain experts to design patch synthesis rules and suffer from low template coverage of real-world type errors. Learning-based approaches alleviate the manual efforts in designing patch synthesis rules. Among the learning-based approaches, the prompt-based approach which leverages the knowledge base of code pre-trained models via pre-defined prompts, obtains state-of-the-art performance in general program repair tasks. However, such prompts are manually defined and do not involve any specific clues for repairing Python type errors, resulting in limited effectiveness. How to automatically improve prompts with the domain knowledge for type error repair is challenging yet under-explored. In this paper, we present TypeFix, a novel prompt-based approach with fix templates incorporated for repairing Python type errors. TypeFix first mines generalized fix templates via a novel hierarchical clustering algorithm. The identified fix templates indicate the common edit patterns and contexts of existing type error fixes. TypeFix then generates code prompts for code pre-trained models by employing the generalized fix templates as domain knowledge, in which the masks are adaptively located for each type error instead of being pre-determined. Experiments on two benchmarks, including BugsInPy and TypeBugs, show that TypeFix successfully repairs 26 and 55 type errors, outperforming the best baseline approach by 9 and 14, respectively. Besides, the proposed fix template mining approach can cover 75% of developers' patches in both benchmarks, increasing the best rule-based approach PyTER by more than 30%.","{'model': 'tldr@v2.0.0', 'text': ""TypeFix is presented, a novel prompt-based approach with fix templates incorporated for repairing Python type errors that can cover 75% of developers' patches in both benchmarks, and increases the best rule- based approach PyTER by more than 30%.""}",http://arxiv.org/pdf/2306.01394
-CitePrompt: Using Prompts to Identify Citation Intent in Scientific Papers,Avishek Lahiri,"Citations in scientific papers not only help us trace the intellectual lineage but also are a useful indicator of the scientific significance of the work. Citation intents prove beneficial as they specify the role of the citation in a given context. We present a tool Citeprompt which uses the hitherto unexplored approach of prompt learning for citation intent classification. We argue that with the proper choice of the pretrained language model, the prompt template, and the prompt verbalizer, we can not only get results that are better than or comparable to those obtained with the state-of-the-art methods but also do it with much less exterior information about the scientific document. We report state-of-the-art results on the ACL-ARC dataset, and also show significant improvement on the SciCite dataset over all baseline models except one. As suitably large labelled datasets for citation intent classification can be quite hard to find, in a first, we propose the conversion of this task to the few-shot and zero-shot settings. For the ACL-ARC dataset, we report a 53.86% F1 score for the zero-shot setting, which improves to 63.61% and 66.99% for the 5-shot and 10-shot settings respectively.","{'model': 'tldr@v2.0.0', 'text': 'A tool Citeprompt is presented which uses the hitherto unexplored approach of prompt learning for citation intent classification and argues that with the proper choice of the pretrained language model, the prompt template, and the prompt verbalizer, one can get results that are better than or comparable to those obtained with the state-of-the-art methods.'}",https://arxiv.org/pdf/2304.12730
-Extracting Structured Seed-Mediated Gold Nanorod Growth Procedures from Literature with GPT-3,Nicholas Walker,"Although gold nanorods have been the subject of much research, the pathways for controlling their shape and thereby their optical properties remain largely heuristically understood. Although it is apparent that the simultaneous presence of and interaction between various reagents during synthesis control these properties, computational and experimental approaches for exploring the synthesis space can be either intractable or too time-consuming in practice. This motivates an alternative approach leveraging the wealth of synthesis information already embedded in the body of scientific literature by developing tools to extract relevant structured data in an automated, high-throughput manner. To that end, we present an approach using the powerful GPT-3 language model to extract structured multi-step seed-mediated growth procedures and outcomes for gold nanorods from unstructured scientific text. GPT-3 prompt completions are fine-tuned to predict synthesis templates in the form of JSON documents from unstructured text input with an overall accuracy of $86\%$. The performance is notable, considering the model is performing simultaneous entity recognition and relation extraction. We present a dataset of 11,644 entities extracted from 1,137 papers, resulting in 268 papers with at least one complete seed-mediated gold nanorod growth procedure and outcome for a total of 332 complete procedures.","{'model': 'tldr@v2.0.0', 'text': 'An approach using the powerful GPT-3 language model to extract structured multi-step seed-mediated growth procedures and outcomes for gold nanorods from unstructured scientific text by developing tools to extract relevant structured data in an automated, high-throughput manner.'}",http://arxiv.org/pdf/2304.13846
-A Dataset for Cross-Domain Reasoning via Template Filling,,"While several benchmarks exist for reasoning 001 tasks, reasoning across domains is an under- 002 explored area in NLP. Towards this, we present 003 a dataset and a prompt-template-filling ap- 004 proach to enable sequence to sequence mod- 005 els to perform cross-domain reasoning. We 006 also present a case-study with commonsense 007 and health and well-being domains, where 008 we study how prompt-template-filling en- 009 ables pretrained sequence to sequence models 010 across domains. Our experiments across sev- 011 eral pretrained encoder-decoder models show 012 that cross-domain reasoning is challenging for 013 current models. We also show an in-depth er- 014 ror analysis and avenues for future research for 015 reasoning across domains 1 . 016","{'model': 'tldr@v2.0.0', 'text': 'A dataset and a prompt-template-filling approach to enable sequence to sequence mod-5 models to perform cross-domain reasoning and an in-depth analysis of avenues for future research for reasoning across domains.'}",
-Prompting for Automatic Log Template Extraction,Junjie Xu,"Log parsing, the initial and vital stage in automated log analysis, involves extracting log templates from semi-structured logs to generate structured logs. Nonetheless, current log parsers are limited in effectiveness due to two primary reasons. Firstly, traditional data-driven log parsers heavily rely on heuristics or manually crafted features provided by domain experts, which may not consistently yield optimal performance when applied to diverse log systems. Secondly, existing deep learning-based log parsers necessitate model tuning, which is typically confined to training samples and leads to suboptimal performance across the entire log source. To overcome these limitations, we propose a precise log parsing framework named LogDiv, which leverages the in-context inference capability of large language models. Specifically, LogDiv extracts the hidden semantics from multiple log examples through prompt demonstrations. Without the need for model tuning, LogDiv can directly generate a log template for the target log message by leveraging the semantics provided in the prompt context. Additionally, we introduce a simple yet effective prompt format for extracting the output and enhancing the quality of the generated log templates. To validate the performance of LogDiv, we conducted experiments using 16 widely-used public datasets. The results show that LogDiv achieves state-of-the-art performance with an average parsing accuracy of 97.7%, precision template accuracy of 88.1%, and recall template accuracy of 90.8%.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a precise log parsing framework named LogDiv, which leverages the in-context inference capability of large language models and achieves state-of-the-art performance.'}",https://arxiv.org/pdf/2307.09950
-DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines,O. Khattab,"The ML community is rapidly exploring techniques for prompting language models (LMs) and for stacking them into pipelines that solve complex tasks. Unfortunately, existing LM pipelines are typically implemented using hard-coded""prompt templates"", i.e. lengthy strings discovered via trial and error. Toward a more systematic approach for developing and optimizing LM pipelines, we introduce DSPy, a programming model that abstracts LM pipelines as text transformation graphs, i.e. imperative computational graphs where LMs are invoked through declarative modules. DSPy modules are parameterized, meaning they can learn (by creating and collecting demonstrations) how to apply compositions of prompting, finetuning, augmentation, and reasoning techniques. We design a compiler that will optimize any DSPy pipeline to maximize a given metric. We conduct two case studies, showing that succinct DSPy programs can express and optimize sophisticated LM pipelines that reason about math word problems, tackle multi-hop retrieval, answer complex questions, and control agent loops. Within minutes of compiling, a few lines of DSPy allow GPT-3.5 and llama2-13b-chat to self-bootstrap pipelines that outperform standard few-shot prompting (generally by over 25% and 65%, respectively) and pipelines with expert-created demonstrations (by up to 5-46% and 16-40%, respectively). On top of that, DSPy programs compiled to open and relatively small LMs like 770M-parameter T5 and llama2-13b-chat are competitive with approaches that rely on expert-written prompt chains for proprietary GPT-3.5. DSPy is available at https://github.com/stanfordnlp/dspy","{'model': 'tldr@v2.0.0', 'text': 'DSPy is introduced, a programming model that abstracts LM pipelines as text transformation graphs, i.e. imperative computational graphs where LMs are invoked through declarative modules, and a compiler is designed that will optimize any DSPy pipeline to maximize a given metric.'}",https://arxiv.org/pdf/2310.03714
-Role Knowledge Prompting for Document-Level Event Argument Extraction,Ruijuan Hu,"Document-level event argument extraction (DEAE) aims to identify the arguments corresponding to the roles of a given event type in a document. However, arguments scattering and arguments and roles overlapping make DEAE face great challenges. In this paper, we propose a novel DEAE model called Role Knowledge Prompting for Document-Level Event Argument Extraction (RKDE), which enhances the interaction between templates and roles through a role knowledge guidance mechanism to precisely prompt pretrained language models (PLMs) for argument extraction. Specifically, it not only facilitates PLMs to understand deep semantics but also generates all the arguments simultaneously. The experimental results show that our model achieved decent performance on two public DEAE datasets, with 3.2% and 1.4% F1 improvement on Arg-C, and to some extent, it addressed the overlapping arguments and roles.","{'model': 'tldr@v2.0.0', 'text': 'A novel DEAE model called Role Knowledge Prompting for Document-Level Event Argument Extraction (RKDE) is proposed, which enhances the interaction between templates and roles through a role knowledge guidance mechanism to precisely prompt pretrained language models (PLMs) for argument extraction.'}",https://www.mdpi.com/2076-3417/13/5/3041/pdf?version=1677492694
-CoCoMo: Computational Consciousness Modeling for Generative and Ethical AI,Edward Y. Chang,"The CoCoMo model proposes a computational solution to the challenge of incorporating ethical and emotional intelligence considerations into AI systems, with the aim of creating AI agents that combine knowledge with compassion. To achieve this goal, CoCoMo prioritizes fairness, beneficence, non-maleficence, empathy, adaptability, transparency, and critical and exploratory thinking abilities. The model employs consciousness modeling, reinforcement learning, and prompt template formulation to support these desired traits. By incorporating ethical and emotional intelligence considerations, a generative AI model can potentially lead to improved fairness, reduced toxicity, and increased reliability.",,http://arxiv.org/pdf/2304.02438
-CoT-BERT: Enhancing Unsupervised Sentence Representation through Chain-of-Thought,Bowen Zhang,"Unsupervised sentence representation learning aims to transform input sentences into fixed-length vectors enriched with intricate semantic information while obviating the reliance on labeled data. Recent progress within this field, propelled by contrastive learning and prompt engineering, has significantly bridged the gap between unsupervised and supervised strategies. Nonetheless, the potential utilization of Chain-of-Thought, remains largely untapped within this trajectory. To unlock latent capabilities within pre-trained models, such as BERT, we propose a two-stage approach for sentence representation: comprehension and summarization. Subsequently, the output of the latter phase is harnessed as the vectorized representation of the input sentence. For further performance enhancement, we meticulously refine both the contrastive learning loss function and the template denoising technique for prompt engineering. Rigorous experimentation substantiates our method, CoT-BERT, transcending a suite of robust baselines without necessitating other text representation models or external databases.","{'model': 'tldr@v2.0.0', 'text': 'Rigorous experimentation substantiates the method, CoT-BERT, transcending a suite of robust baselines without necessitating other text representation models or external databases, and meticulously refine both the contrastive learning loss function and the template denoising technique for prompt engineering.'}",https://arxiv.org/pdf/2309.11143
-Advanced prompting as a catalyst: Empowering large language models in the management of gastrointestinal cancers,J. Yuan,"Large Language Models' (LLMs) performance in healthcare can be significantly impacted by prompt engineering. However, the area of study remains relatively uncharted in gastrointestinal oncology until now. Our research delves into this unexplored territory, investigating the efficacy of varied prompting strategies, including simple prompts, templated prompts, in-context learning (ICL), and multi-round iterative questioning, for optimizing the performance of LLMs within a medical setting. We develop a comprehensive evaluation system to assess the performance of LLMs across multiple dimensions. This robust evaluation system ensures a thorough assessment of the LLMs' capabilities in the field of medicine. Our findings suggest a positive relationship between the comprehensiveness of the prompts and the LLMs' performance. Notably, the multi-round strategy, which is characterized by iterative question-and-answer rounds, consistently yields the best results. ICL, a strategy that capitalizes on interrelated contextual learning, also displays significant promise, surpassing the outcomes achieved with simpler prompts. The research underscores the potential of advanced prompt engineering and iterative learning approaches for boosting the applicability of LLMs in healthcare. We recommend that additional research be conducted to refine these strategies and investigate their potential integration, to truly harness the full potential of LLMs in medical applications. ","{'model': 'tldr@v2.0.0', 'text': 'The research investigates the efficacy of varied prompting strategies, including simple prompts, templated prompts, in-context learning (ICL), and multi-round iterative questioning, for optimizing the performance of LLMs within a medical setting, and develops a comprehensive evaluation system.'}",https://www.the-innovation.org/data/article/export-pdf?id=64db4fd54228a72545780714
-Automated Cerebral Infarct Detection on Computed Tomography Images Based on Deep Learning,S. Peng,"The limited accuracy of cerebral infarct detection on CT images caused by the low contrast of CT hinders the desirable application of CT as a first-line diagnostic modality for screening of cerebral infarct. This research was aimed at utilizing convolutional neural network to enhance the accuracy of automated cerebral infarct detection on CT images. The CT images underwent a series of preprocessing steps mainly to enhance the contrast inside the parenchyma, adjust the orientation, spatially normalize the images to the CT template, and create a t-score map for each patient. The input format of the convolutional neural network was the t-score matrix of a 16 × 16-pixel patch. Non-infarcted and infarcted patches were selected from the t-score maps, on which data augmentation was conducted to generate more patches for training and testing the proposed convolutional neural network. The convolutional neural network attained a 93.9% patch-wise detection accuracy in the test set. The proposed method offers prompt and accurate cerebral infarct detection on CT images. It renders a frontline detection modality of ischemic stroke on an emergent or regular basis.","{'model': 'tldr@v2.0.0', 'text': 'The proposed convolutional neural network renders a frontline detection modality of ischemic stroke on an emergent or regular basis and offers prompt and accurate cerebral infarct detection on CT images.'}",https://www.mdpi.com/2227-9059/10/1/122/pdf?version=1641806958
-Incorporating Instructional Prompts into a Unified Generative Framework for Joint Multiple Intent Detection and Slot Filling,Yangjun Wu,"The joint multiple Intent Detection (ID) and Slot Filling (SF) is a significant challenge in spoken language understanding. Because the slots in an utterance may relate to multi-intents, most existing approaches focus on utilizing task-specific components to capture the relations between intents and slots. The customized networks restrict models from modeling commonalities between tasks and generalization for broader applications. To address the above issue, we propose a Unified Generative framework (UGEN) based on a prompt-based paradigm, and formulate the task as a question-answering problem. Specifically, we design 5-type templates as instructional prompts, and each template includes a question that acts as the driver to teach UGEN to grasp the paradigm, options that list the candidate intents or slots to reduce the answer search space, and the context denotes original utterance. Through the instructional prompts, UGEN is guided to understand intents, slots, and their implicit correlations. On two popular multi-intent benchmark datasets, experimental results demonstrate that UGEN achieves new SOTA performances on full-data and surpasses the baselines by a large margin on 5-shot (28.1%) and 10-shot (23%) scenarios, which verify that UGEN is robust and effective.","{'model': 'tldr@v2.0.0', 'text': 'A Unified Generative framework (UGEN) based on a prompt-based paradigm, and formulate the task as a question-answering problem to understand intents, slots, and their implicit correlations is proposed.'}",
-A Practical Three-phase Approach To Fully Automated Programming Using System Decomposition And Coding Copilots,Haoli Bai,"Very large-scale (VLS) deep learning models are capable of generating meaningful code snippets, yet the performance drops dramatically when the coding task becomes more complex. Although fully neural approaches have been proposed to solve this problem, the value of the application is still limited. In our work, we propose a neuro-symbolic approach that integrates the symbolic natures of programming and the existing neural language models. We divide a programming task into three phases: forming a hierarchical task composed of functions, completing each function, and fulfilling the corner cases. Because each phase can be completed by language models, the coding process can be fully automated. Our contribution is three-fold. Firstly, we show that with little help from humans, VLS language models are capable of completing non-trivial programming tasks. Secondly, we provide a number of empirical insights to create prompt templates that help the language models generate better code. Thirdly, compared to the existing approaches, our work provides a much more practical approach for programmers and researchers to follow. The generated programming project using our fully automated programming approach and part of the ablation study code are available at https://github.com/BiEchi/FAP.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a neuro-symbolic approach that integrates the symbolic natures of programming and the existing neural language models to show that with little help from humans, VLS language models are capable of completing non-trivial programming tasks.'}",
-Towards Robust NLG Bias Evaluation with Syntactically-diverse Prompts,Arshiya Aggarwal,"We present a robust methodology for evaluating biases in natural language generation(NLG) systems. Previous works use fixed hand-crafted prefix templates with mentions of various demographic groups to prompt models to generate continuations for bias analysis. These fixed prefix templates could themselves be specific in terms of styles or linguistic structures, which may lead to unreliable fairness conclusions that are not representative of the general trends from tone varying prompts. To study this problem, we paraphrase the prompts with different syntactic structures and use these to evaluate demographic bias in NLG systems. Our results suggest similar overall bias trends but some syntactic structures lead to contradictory conclusions compared to past works. We show that our methodology is more robust and that some syntactic structures prompt more toxic content while others could prompt less biased generation. This suggests the importance of not relying on a fixed syntactic structure and using tone-invariant prompts. Introducing syntactically-diverse prompts can achieve more robust NLG (bias) evaluation.",,https://arxiv.org/pdf/2212.01700
-KU X Upstage’s Submission for the WMT22 Quality Estimation: Critical Error Detection Shared Task,Sugyeong Eo,"This paper presents KU X Upstage’s submission to the quality estimation (QE): critical error detection (CED) shared task in WMT22. We leverage the XLM-RoBERTa large model without utilizing any additional parallel data. To the best of our knowledge, we apply prompt-based fine-tuning to the QE task for the first time. To maximize the model’s language understanding capability, we reformulate the CED task to be similar to the masked language model objective, which is a pre-training strategy of the language model. We design intuitive templates and label words, and include auxiliary descriptions such as demonstration or Google Translate results in the input sequence. We further improve the performance through the template ensemble, and as a result of the shared task, our approach achieve the best performance for both English-German and Portuguese-English language pairs in an unconstrained setting.","{'model': 'tldr@v2.0.0', 'text': 'KU X Upstage’s submission to the quality estimation (QE): critical error detection (CED) shared task in WMT22 is presented, and prompt-based fine-tuning is applied to the QE task for the first time.'}",
-DESED: Dialogue-based Explanation for Sentence-level Event Detection,Yinyi Wei,"Many recent sentence-level event detection efforts focus on enriching sentence semantics, e.g., via multi-task or prompt-based learning. Despite the promising performance, these methods commonly depend on label-extensive manual annotations or require domain expertise to design sophisticated templates and rules. This paper proposes a new paradigm, named dialogue-based explanation, to enhance sentence semantics for event detection. By saying dialogue-based explanation of an event, we mean explaining it through a consistent information-intensive dialogue, with the original event description as the start utterance. We propose three simple dialogue generation methods, whose outputs are then fed into a hybrid attention mechanism to characterize the complementary event semantics. Extensive experimental results on two event detection datasets verify the effectiveness of our method and suggest promising research opportunities in the dialogue-based explanation paradigm.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new paradigm, named dialogue-based explanation, to enhance sentence semantics for event detection, by explaining it through a consistent information-intensive dialogue, with the original event description as the start utterance.'}",
-Can Language Models be Biomedical Knowledge Bases?,Mujeen Sung,"Pre-trained language models (LMs) have become ubiquitous in solving various natural language processing (NLP) tasks. There has been increasing interest in what knowledge these LMs contain and how we can extract that knowledge, treating LMs as knowledge bases (KBs). While there has been much work on probing LMs in the general domain, there has been little attention to whether these powerful LMs can be used as domain-specific KBs. To this end, we create the BioLAMA benchmark, which is comprised of 49K biomedical factual knowledge triples for probing biomedical LMs. We find that biomedical LMs with recently proposed probing methods can achieve up to 18.51% Acc@5 on retrieving biomedical knowledge. Although this seems promising given the task difficulty, our detailed analyses reveal that most predictions are highly correlated with prompt templates without any subjects, hence producing similar results on each relation and hindering their capabilities to be used as domain-specific KBs. We hope that BioLAMA can serve as a challenging benchmark for biomedical factual probing.","{'model': 'tldr@v2.0.0', 'text': 'The BioLAMA benchmark is created, which is comprised of 49K biomedical factual knowledge triples for probing biomedical LMs and reveals that most predictions are highly correlated with prompt templates without any subjects, hindering their capabilities to be used as domain-specific KBs.'}",https://aclanthology.org/2021.emnlp-main.388.pdf
-Vision Encoders in Visual Question Answering,Ryan R. Anderson,"Most existing methods that apply pretrained Visual Language Models (VLMs) to vision and language tasks do not sufficiently explore the effect of the format of their inputs on downstream performance. We show that utilising appropriate prompt formatting is a simple yet effective approach to improving the few-shot performance of VLMs that use relatively small language models on the Visual Question Answering (VQA) task. We format the inputs used to prompt a VLM using a modified text-only template from a closed-book question answering task that the language-model component of the VLM was pretrained on. By doing this, we explicitly align the VQA task with a task that this language model has already seen, enabling the VLM to leverage the similarities between the tasks, such as the answer-length distribution, when generating answers to the visual questions. In order to test our claims, we implement a simple architecture based on Frozen (Tsimpoukelli et al., 2021) and ClipCap (Mokady et al., 2021), whereby, through image captioning, the VLM learns to integrate powerful pretrained vision-only and language-only models via a relatively simple learnt mapping network. Furthermore, we contextualise our approach relative to existing work by presenting a unified view of VLMs. Our results show that explicit alignment enables our VLMs to achieve a significantly higher zero-shot (34.49% vs 20.89%) and best overall (40.39% vs 30.83%) VQA score on the VQA2.0 dataset (Goyal et al., 2017) than when the prompt template from Frozen (Tsimpoukelli et al., 2021) and Flamingo (Alayrac et al., 2022) is used. Furthermore, our zero-shot and best overall performance is better than Frozen’s (34.49% vs 29.5% and 40.39% vs 38.2%, respectively) despite Frozen using a language model with more than double the number of parameters. Our code is available here.","{'model': 'tldr@v2.0.0', 'text': 'This work formats the inputs used to prompt a VLM using a modified text-only template from a closed-book question answering task that the language-model component of the VLM was pretrained on, enabling the V LM to leverage the similarities between the tasks, such as the answer-length distribution, when generating answers to the visual questions.'}",
-Distilling Task-specific Logical Rules from Large Pre-trained Models,Tao Chen,"Logical rules, both transferable and explainable, are widely used as weakly supervised signals for many downstream tasks such as named entity tagging. To reduce the human effort of writing rules, previous researchers adopt an iterative approach to automatically learn logical rules from several seed rules. However, obtaining more seed rules can only be accomplished by extra human annotation with heavy costs. Limited by the size and quality of the seed rules, the model performance of previous systems is bounded. In this paper, we develop a novel framework STREAM to distill task-specific logical rules from large pre-trained models. Specifically, we borrow recent prompt-based language models as the knowledge expert to yield initial seed rules, and based on the formed high-quality instance pool that acts as an intermediary role, we keep teaching the expert to fit our task and learning task-specific logical rules. Experiments on three public named entity tagging benchmarks demonstrate the effectiveness of our proposed framework. With several predefined prompt templates, our system has gained significant improvements over previous state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': ""A novel framework STREAM to distill task-specific logical rules from large pre-trained models is developed based on the formed high-quality instance pool that acts as an intermediary role and keeps teaching the expert to fit the authors' task and learning task- specific logical rules.""}",http://arxiv.org/pdf/2210.02768
-Keyword-optimized Template Insertion for Clinical Information Extraction via Prompt-based Learning,Eugenia Alleva,"Clinical note classification is a common clinical NLP task. However, annotated data-sets are scarse. Prompt-based learning has recently emerged as an effective method to adapt pre-trained models for text classification using only few training examples. A critical component of prompt design is the definition of the template (i.e. prompt text). The effect of template position, however, has been insufficiently investigated. This seems particularly important in the clinical setting, where task-relevant information is usually sparse in clinical notes. In this study we develop a keyword-optimized template insertion method (KOTI) and show how optimizing position can improve performance on several clinical tasks in a zero-shot and few-shot training setting.","{'model': 'tldr@v2.0.0', 'text': 'This study develops a keyword-optimized template insertion method (KOTI) and shows how optimizing position can improve performance on several clinical tasks in a zero-shot and few-shot training setting.'}",
-KUL@SMM4H’22: Template Augmented Adaptive Pre-training for Tweet Classification,Sumam Francis,This paper describes models developed for the Social Media Mining for Health (SMM4H) 2022 shared tasks. Our team participated in the first subtask that classifies tweets with Adverse Drug Effect (ADE) mentions. Our best-performing model comprises of a template augmented task adaptive pre-training and further fine-tuning on target task data. Augmentation with random prompt templates increases the amount of task-specific data to generalize the LM to the target task domain. We explore 2 pre-training strategies: Masked language modeling (MLM) and Simple contrastive pre-training (SimSCE) and the impact of adding template augmentations with these pre-training strategies. Our system achieves an F1 score of 0.433 on the test set without using supplementary resources and medical dictionaries.,"{'model': 'tldr@v2.0.0', 'text': 'This paper describes models developed for the Social Media Mining for Health (SMM4H) 2022 shared tasks and explores 2 pre-training strategies: Masked language modeling (MLM) and Simple contrastive pre- training (SimSCE) and the impact of adding template augmentations with these pre- Training strategies.'}",
-Research on Chinese Short Text Classification Based on Prefix-vector Attention Template and Probabilistic Answer Set,Baoshan Sun,"When people tend to communicate in the network, a large amount of text data is generated. These text data have become an important way for people to obtain information. Nowadays, text classification through prompt learning has also become the focus of people’s research. However, the effect of prompt learning is not ideal due to the immobilization of the prompt template and the single label words. In response to the above problems, this paper proposes a method of constructing a prefix vector attention template (PAP) and a probabilistic answer set. The prefix space vector is added in front of the input text sentence as a template for prompt learning, and using the Attention mechanism to connect the template with the input sentence to construct the input of the model sentence. For each label word, an answer space set is constructed, and the probability that the text belongs to this category is obtained in the form of the answer set probability. The experimental results show that the PAP + probabilistic answer set algorithm proposed in this paper is significantly better than other benchmark models in the text classification task, and can achieve better results.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the PAP + probabilistic answer set algorithm proposed in this paper is significantly better than other benchmark models in the text classification task, and can achieve better results.'}",
-STT: Soft Template Tuning for Few-Shot Learning,Christopher Potts,"With the rapid expansion of large pre-trained 001 language models, fine-tuning all the model pa002 rameters for downstream tasks is becoming 003 computationally prohibitive. The recently de004 veloped prompt-based methods freeze the en005 tire model parameters and only update the so006 called prompt parameters appended to the in007 puts, significantly reducing the burden of fully 008 fine-tuning. However, standard prompt-based 009 methods mainly consider the case where suf010 ficient data of downstream tasks are available. 011 It is still unclear whether the advantage can be 012 transferred to the few-shot regime, where only 013 limited data are available for each downstream 014 task. Our empirical studies suggest there is 015 still a gap between prompt tuning and fully 016 fine-tuning for few-shot learning. We propose 017 a new prompt-tuning framework, called Soft 018 Template Tuning (STT), to bridge the gap. STT 019 combines manual prompts and auto-prompts, 020 and treats downstream classification tasks as a 021 masked language modeling task. STT can close 022 the gap between fine-tuning and prompt-based 023 methods without introducing additional param024 eters. Importantly, it can even outperform 025 the timeand resource-consuming fine-tuning 026 method on sentiment classification tasks. 027","{'model': 'tldr@v2.0.0', 'text': 'A new prompt-tuning framework, called Soft 018 Template Tuning (STT), is proposed, which combines manual prompts and auto-prompts and treats downstream classification tasks as a masked language modeling task and can even outperform the time and resource-consuming fine-tuned method on sentiment classification tasks.'}",
-STT: Soft Template Tuning for Few-Shot Adaptation,Ping Yu,"Prompt tuning has been an extremely effective tool to adapt a pre-trained model to downstream tasks. However, standard prompt-based methods mainly consider the case of sufficient data of downstream tasks. It is still unclear whether the advantage can be transferred to the few-shot regime, where only limited data are available for each downstream task. Although some works have demonstrated the potential of prompt-tuning under the few-shot setting, the main stream methods via searching discrete prompts or tuning soft prompts with limited data are still very challenging. Through extensive empirical studies, we find that there is still a gap between prompt tuning and fully fine-tuning for few-shot learning. To bridge the gap, we propose a new prompt-tuning framework, called Soft Template Tuning (STT) 1. STT combines manual and auto prompts, and treats down-stream classification tasks as a masked language modeling task. Comprehensive evaluation on different settings suggests STT can close the gap between fine-tuning and prompt-based methods without introducing additional parameters. Significantly, it can even outperform the time- and resource-consuming fine-tuning method on sentiment classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A new prompt-tuning framework, called Soft Template Tuning (STT) is proposed, which combines manual and auto prompts, and treats down-stream classification tasks as a masked language modeling task and can even outperform the time- and resource-consuming fine- Tuning method on sentiment classification tasks.'}",https://arxiv.org/pdf/2207.08408
-Cross-Domain Reasoning via Template Filling,Dheeraj Rajagopal,"In this paper, we explore the ability of sequence to sequence models to perform cross-domain reasoning. Towards this, we present a prompt-template-filling approach to enable sequence to sequence models to perform cross-domain reasoning. We also present a case-study with commonsense and health and well-being domains, where we study how prompt-template-filling enables pretrained sequence to sequence models across domains. Our experiments across several pretrained encoder-decoder models show that cross-domain reasoning is challenging for current models. We also show an in-depth error analysis and avenues for future research for reasoning across domains 1 .","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a prompt-template-filling approach to enable sequence to sequence models to perform cross-domain reasoning, and presents a case-study with commonsense and health and well-being domains, where it is studied how prompt- templates enables pretrained sequence to sequences models across domains.'}",
-Unveiling One-pot Template-free Fabrication of Exquisite Multidimensional PtNi Multicubes Nanoarchitectonics for the Efficient Electrochemical Oxidation of Ethanol and Methanol with a Great Tolerance for CO.,Fengxia Wu,"Multidimensional bi-metallic Pt-based nanoarchitectonics are highly promising in electrochemical energy conversion technologies because of their fancy structural merits and accessible active sites; however, hitherto their precise template-free fabrication remains a grand challenge. Herein, we report a facile template-free solvothermal approach for the one-pot rational design of co-centric PtNi multicubes nanoarchitectonics via adjusting the oleylamine/oleic acid ratio with curcumin. The obtained multidimensional PtNi multicubes comprise multiple small interlace-stacked nanocubes subunits assembled in spatially porous branched nanoarchitectonics and bounded by high-index facets. The synthetic mechanism is based on the spontaneous isolation among prompt nucleation and oriented attachment epitaxial growth. These inimitable architectural and compositional merits of PtNi multicubes endowed the ethanol oxidation mass and specific activity by 5.6 and 9.03 times than commercial Pt/C catalyst, respectively, alongside enhancement the methanol oxidation mass activity by 2.3 times. Also, PtNi multicubes showed superior durability and a higher tolerance for CO poisoning than Pt/C catalyst. This work may pave the way for tailored synthesis of Pt-based nanoarchitectonics for myriad catalytic reactions.","{'model': 'tldr@v2.0.0', 'text': 'A facile template-free solvothermal approach for the one-pot rational design of co-centric PtNi multicubes nanoarchitectonics via adjusting the oleylamine/oleic acid ratio with curcumin is reported, paving the way for tailored synthesis of Pt-based nanoarch architectonics for myriad catalytic reactions.'}",
-DAPrompt: Deterministic Assumption Prompt Learning for Event Causality Identification,Wei Xiang,"Event Causality Identification (ECI) aims at determining whether there is a causal relation between two event mentions. Conventional prompt learning designs a prompt template to first predict an answer word and then maps it to the final decision. Unlike conventional prompts, we argue that predicting an answer word may not be a necessary prerequisite for the ECI task. Instead, we can first make a deterministic assumption on the existence of causal relation between two events and then evaluate its rationality to either accept or reject the assumption. The design motivation is to try the most utilization of the encyclopedia-like knowledge embedded in a pre-trained language model. In light of such considerations, we propose a deterministic assumption prompt learning model, called DAPrompt, for the ECI task. In particular, we design a simple deterministic assumption template concatenating with the input event pair, which includes two masks as predicted events' tokens. We use the probabilities of predicted events to evaluate the assumption rationality for the final event causality decision. Experiments on the EventStoryLine corpus and Causal-TimeBank corpus validate our design objective in terms of significant performance improvements over the state-of-the-art algorithms.","{'model': 'tldr@v2.0.0', 'text': ""A deterministic assumption prompt learning model, called DAPrompt, is proposed, which includes two masks as predicted events' tokens and uses the probabilities of predicted events to evaluate the assumption rationality for the final event causality decision.""}",https://arxiv.org/pdf/2307.09813
-Working Example of Template Prompt Payments by Local Authorities,Em Reeves,Number Value (€) Percentage (%) of total invoices paid Percentage (%) of the total amount paid Total invoices paid in Quarter,,
-The effects of implementing a point-of-care template to prompt routine and screening in consulting for (the,,Background This study aimed to evaluate whether prompting general practitioners (GPs) to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis (OA) improves pain outcomes.,"{'model': 'tldr@v2.0.0', 'text': 'This study aimed to evaluate whether prompting general practitioners to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis improves pain outcomes.'}",
-Coarse-Grained Diffraction Template Matching Model to Retrieve Multiconformational Models for Biomolecule Structures from Noisy Diffraction Patterns,Atsushi Tokuhisa,"Biomolecular imaging using X-ray free-electron lasers (XFELs) has been successfully applied to serial femtosecond crystallography. However, the application of single-particle analysis for structure determination using XFELs with 100 nm or smaller biomolecules has two practical problems: the incomplete diffraction datasets for reconstructing 3D assembled structures and the heterogeneous conformational states of samples. A new diffraction template matching method is thus presented here to retrieve a plausible 3D structural model based on single noisy target diffraction patterns, assuming candidate structures. Two concepts are introduced here: prompt candidate diffraction, generated by enhanced sampled coarse-grain (CG) candidate structures, and efficient molecular orientation searching for matching based on Bayesian optimization. A CG model-based diffraction-matching protocol is proposed that achieves a 100-fold speed increase compared to exhaustive diffraction matching using an all-atom model. The conditions that enable multi-conformational analysis were also investigated by simulated diffraction data for various conformational states of chromatin and ribosomes. The proposed method can enable multi-conformational analysis, with a structural resolution of at least 20 Å for 270-800 Å flexible biomolecules, in experimental single-particle structure analyses that employ XFELs.","{'model': 'tldr@v2.0.0', 'text': 'A new diffraction template matching method is presented here to retrieve a plausible 3D structural model based on single noisy target diffraction patterns, assuming candidate structures, and a CG model-based diffraction-matching protocol is proposed that achieves a 100-fold speed increase compared to exhaustive diffraction matching using an all-atom model.'}",
-Unraveling template-free fabrication of carbon nitride nanorods codoped with Pt and Pd for efficient electrochemical and photoelectrochemical carbon monoxide oxidation at room temperature.,Kamel Eid,"The tailored synthesis of carbon nitrides (CNs) is of particular interest in multidisciplinary catalytic applications. However, their fabrication in the form of one-dimensional (1D) nanorods for electrocatalytic carbon monoxide (CO) oxidation is not hitherto reported. Herein, a facile roadmap is presented for the rational design of Pt- and Pd-codoped CN (PtPd/CNs) nanorods via protonation of melamine in an ethylene glycol solution containing Pt and Pd precursors using NaNO3 and HCl and subsequent annealing. The protonation induces the polymerization of melamine to melon nanosheets that consequently roll up to CN nanorods. This tailored the prompt high mass production of uniform 1D CN nanorods (94 ± 2 nm) with a high surface area (155.2 m2 g-1) and they were atomically codoped with Pt and Pd (1.5 wt%) without a template and/or multiple complicated steps. The electrocatalytic CO oxidation activity of PtPd/CNs is 2.01 and 23.41 times greater than that of the commercial Pt/C catalyst and metal-free CNs, respectively, at room temperature. Meanwhile, the UV-vis light irradiation enhanced the CO oxidation activity of PtPd/CNs nanorods by 1.48 fold compared to that in the dark, emanated from the coupling between the drastic inbuilt catalytic merits of PtPd and the inimitable physicochemical properties of CNs. The presented study may pave the way for using CN-based materials in gas conversion reactions.","{'model': 'tldr@v2.0.0', 'text': 'A facile roadmap is presented for the rational design of Pt- and Pd-codoped CN (PtPd/CNs) nanorods via protonation of melamine in an ethylene glycol solution containing Pt and PD precursors using NaNO3 and HCl and subsequent annealing to pave the way for using CN-based materials in gas conversion reactions.'}",
-PD38-08 MULTI-INSTITUTIONAL PROSPECTIVE VALIDATION OF THE NOVEL MICHIGAN UROLOGICAL SURGERY IMPROVEMENT COLLABORATIVE TRANSPERINEAL BIOPSY TEMPLATE,M. Maruf,"and IPA of both the PCPT and PBCG risk tools were improved by including Prompt-PGS for both the PCPT and PBCG populations (Table 1). Observed versus expected plots (Figure 1) revealed superior calibration of inclusion of the Prompt-PGS score compared to either risk tool alone in both the PCPT and REDUCE populations, particularly on the lower end of risk. CONCLUSIONS: A germline-genetic risk stratification tool, Prompt e PGS, improves the performance of both the PCPT and PBCG risk tools in two large populations, particularly at the lower end of risk. It may be a useful tool for safely decreasing unnecessary prostate biopsies.","{'model': 'tldr@v2.0.0', 'text': 'A germline-genetic risk stratification tool, Prompt e PGS, improves the performance of both the PCPT and PBCG risk tools in two large populations, particularly at the lower end of risk.'}",https://www.auajournals.org/doi/pdf/10.1097/JU.0000000000000917.08
-A Powerful Artificial Intelligence-Based Authentication Mechanism of Retina Template Using Sparse Matrix Representation with High Security,B. Rani,,"{'model': 'tldr@v2.0.0', 'text': 'Mechanized picture characterization frameworks with these AI procedures are created for down-to-earth applications that have the capacity to deal with intra-user fluctuation in the procured biometric identifiers.'}",
-Jammer Identification Technique based on a Template Matching Method,M. Jin,"GNSS has the disadvantage of being vulnerable to jamming, and thus, the necessity of jamming countermeasure techniques has gradually increased. Jamming countermeasure techniques can be divided into an anti-jamming technique and a jammer localization technique. Depending on the type of a jammer, applicable techniques and performance vary significantly. Using an appropriate jamming countermeasure technique, the effect of jamming on a GNSS receiver can be attenuated, and prompt action is enabled when estimating the location of a jammer. However, if an inappropriate jamming countermeasure technique is used, a GNSS receiver may not operate in the worst case. Therefore, jammer identification is a technique that is essential for proper action. In this study, a technique that identifies a jammer based on template matching was proposed. For template matching, analysis of a received jamming signal is required; and the signal analysis was performed using a spectral correlation function. Based on a simulation, it was shown that the proposed identification of jamming signals was possible at various JNR.","{'model': 'tldr@v2.0.0', 'text': 'A technique that identifies a jammer based on template matching was proposed and it was shown that the proposed identification of jamming signals was possible at various JNR.'}",
-Inspection method for template library programming conventions,金海,"The invention discloses an inspection method for template library programming conventions and belongs to the technical field of computer parallel programming. According to the inspection method, the programming conventions are described as states, and the migration of the states is described as calling of specific methods in the programming conventions; then the states are implemented into categories, and methods satisfying the programming conventions are written in the categories for calling by programming personnel, so as to ensure that a compiler stops compiling and gives out a corresponding prompt when the programming personnel is against the programming conventions; and possible errors from the programming personnel are predicted and the methods are added in the corresponding categories representing the states so as to provide corresponding prompt information when the programming personnel is against the programming conventions to help the programming personnel to inspect and modify the errors. The inspection method can effectively improve the parallel program writing efficiency of the programming personnel.","{'model': 'tldr@v2.0.0', 'text': 'The inspection method can effectively improve the parallel program writing efficiency of the programming personnel and be used to inspect and modify the errors.'}",
-TEMPLATE SYSTEM FOR SECOND LANGUAGE AURAL COMPREHENSION,Donna Mydlarski,"With the introduction of the PROMPT template materials into the French, Italian, and Spanish classes at the Universities of Calgary and Guelph, Canada, it became evident that a system to help language students improve their listening skills was much needed. In response to this need, a template called DICTATE was developed to allow teachers to use the dictation format to practice auditory discrimination, aural comprehension, and orthography in the most efficient and effective way possible.","{'model': 'tldr@v2.0.0', 'text': 'A template called DICTATE was developed to allow teachers to use the dictation format to practice auditory discrimination, aural comprehension, and orthography in the most efficient and effective way possible.'}",
-The 10 Building Blocks of High-Performing Primary Care,T. Bodenheimer,"Our experiences studying exemplar primary care practices, and our work assisting other practices to become more patient centered, led to a formulation of the essential elements of primary care, which we call the 10 building blocks of high-performing primary care. The building blocks include 4 foundational elements—engaged leadership, data-driven improvement, empanelment, and team-based care—that assist the implementation of the other 6 building blocks—patient-team partnership, population management, continuity of care, prompt access to care, comprehensiveness and care coordination, and a template of the future. The building blocks, which represent a synthesis of the innovative thinking that is transforming primary care in the United States, are both a description of existing high-performing practices and a model for improvement.","{'model': 'tldr@v2.0.0', 'text': 'The building blocks, which represent a synthesis of the innovative thinking that is transforming primary care in the United States, are both a description of existing high-performing practices and a model for improvement.'}",http://www.annfammed.org/content/12/2/166.full.pdf
-Let Me Check the Examples: Enhancing Demonstration Learning via Explicit Imitation,Sirui Wang,"Demonstration learning aims to guide the prompt prediction by providing answered demonstrations in the few shot settings. Despite achieving promising results, existing work only concatenates the answered examples as demonstrations to the prompt template (including the raw context) without any additional operation, neglecting the prompt-demonstration dependencies. Besides, prior research found that randomly replacing the labels of demonstrations marginally hurts performance, illustrating that the model could not properly learn the knowledge brought by the demonstrations. Inspired by the human learning process, in this paper, we introduce Imitation DEMOnstration learning (Imitation-Demo) to strengthen demonstration learning via explicitly imitating human review behaviour, which includes: (1) contrastive learning mechanism to concentrate on similar demonstrations.(2) demonstration-label re-prediction method to consolidate known knowledge. Experiment results show that our proposed method achieves state-of-the-art performance on 5 out of 14 classification corpus. Further studies also prove that Imitation-Demo strengthens the associations between the prompt and demonstrations, which could provide the basis for exploring how demonstration learning works.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the human learning process, Imitation-Demo is introduced to strengthen demonstration learning via explicitly imitating human review behaviour, which includes a contrastive learning mechanism to concentrate on similar demonstrations and a demonstration-label re-prediction method to consolidate known knowledge.'}",http://arxiv.org/pdf/2209.00455
-A Few-shot Approach to Resume Information Extraction via Prompts,Chengguang Gan,,"{'model': 'tldr@v2.0.0', 'text': 'The Manual Knowledgeable Verbalizer (MKV), a rule for constructing verbalizers for specific applications, is presented, showing that MKV rules yield more effective, robust templates and verbalizers than existing methods.'}",https://arxiv.org/pdf/2209.09450
-STPrompt: Semantic-guided and Task-driven prompts for Effective Few-shot Classification,Jinta Weng,"The effectiveness of prompt learning has been demonstrated in different pre-trained language models. By formulating suitable template and choosing representative label mapping, prompt learning can be used as an efficient knowledge probe. However, finding suitable prompt in existing methods requires multiple experimental attempts or appropriate vector initialization on formulating suitable template and choosing representative label mapping, which it is more common in few-shot learning tasks. Motivating by PLM working process, we try to construct the prompt from task semantic perspective and thus propose the STPrompt -Semantic-guided and Task-driven Prompt model. Specifically, two novel prompts generated from the semantic dependency tree (Dep-prompt) and task-specific metadata description (Meta-prompt), are firstly constructed in a prompt augmented pool, and the proposed model would automatically select a suitable semantic prompt to motivating the prompt learning process. Our results show that the proposed model achieves the state-of-the-art performance in five different datasets of few-shot text classification tasks, which prove that more semantic and significant prompts could assume as a better knowledge proving tool.","{'model': 'tldr@v2.0.0', 'text': 'The proposed STPrompt -Semantic-guided and Task-driven Prompt model achieves the state-of-the-art performance in five different datasets of few-shot text classification tasks, which prove that more semantic and significant prompts could assume as a better knowledge proving tool.'}",
-Supplementary material for Mask-free OVIS: Open-Vocabulary Instance Segmentation without Manual Mask Annotations,VS Vibashan,"COCO Caption vs Image-label pseudo-caption: Pseudo-caption generation: Since the pre-trained vision-language models are trained on full sentences, we need to feed the image-labels into a prompt template first, and use them to generate a pseudo-captions. Specifically, given image-labels [category-1,category-2...,category-n], we randomly sample a prompt from 63 prompt templates [1,6] and the pseudo-caption are generated as ”{Prompt-x} + {category-1 and category-2 and ... category-n}”. For example, as shown in Fig. 1 bottom row the sampled prompts are ”A black and white photo of the {category}.” and ”A photo of {category} in the scene.” and the image-labels are ”zebra” and ”giraffe”. Thus, the generated pseudo-captions are ”A black and white photo of the zebra and giraffe.” and ”A photo of one zebra and giraffe in the scene.”","{'model': 'tldr@v2.0.0', 'text': 'Pseudo-caption generation: Since the pre-trained vision-language models are trained on full sentences, they need to feed the image-labels into a prompt template first, and use them to generate a pseudo-captions.'}",
-Pre-training Extractive Question-Answer Prompts for Few-Shot Chinese Text Classification,Gaojian Ding,". In recent years, pre-training models (PLMs) have made impressive progress, and prompt learning has made few-shot learning achievable. However, traditional prompt learning methods often require manual template design, or performance may be unstable due to the limited data in few-shot tasks. To address these issues, we propose a few-shot text classification method based on multi-task learning. We first unify the multi-task into an extractive question-answering (EQA) format, then train the prompt using task data in the unified format. The prompt cists of modular prompts and a router that indicates their functionality. We then initonsialize the downstream training parameters using the router of a pre-training task similar to the downstream task and employ contrastive learning to improve EQA efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This work unify the multi-task into an extractive question-answering (EQA) format, then train the prompt using task data in the unified format and initonsialize the downstream training parameters using the router of a pre-training task similar to the downstream task and employ contrastive learning to improve EQA efflciency.'}",
-Enhancing Cross-lingual Natural Language Inference by Soft Prompting with Multilingual Verbalizer,Shuang Li,"Cross-lingual natural language inference is a fundamental problem in cross-lingual language understanding. Many recent works have used prompt learning to address the lack of annotated parallel corpora in XNLI. However, these methods adopt discrete prompting by simply translating the templates to the target language and need external expert knowledge to design the templates. Besides, discrete prompts of human-designed template words are not trainable vectors and can not be migrated to target languages in the inference stage flexibly. In this paper, we propose a novel Soft prompt learning framework with the Multilingual Verbalizer (SoftMV) for XNLI. SoftMV first constructs cloze-style question with soft prompts for the input sample. Then we leverage bilingual dictionaries to generate an augmented multilingual question for the original question. SoftMV adopts a multilingual verbalizer to align the representations of original and augmented multilingual questions into the same semantic space with consistency regularization. Experimental results on XNLI demonstrate that SoftMV can achieve state-of-the-art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.","{'model': 'tldr@v2.0.0', 'text': 'A novel Soft prompt learning framework with the Multilingual Verbalizer (SoftMV) for XNLI and demonstrates that SoftMV can achieve state-of-the-art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.'}",http://arxiv.org/pdf/2305.12761
-Exploring Prompts in Few-Shot Cross-Linguistic Topic Classification Scenarios,Zhipeng Zhang,"In recent years, large-scale pretrained language models have become widely used in natural language processing tasks. On this basis, prompt learning has achieved excellent performance in specific few-shot classification scenarios. The core idea of prompt learning is to convert a downstream task into a masked language modelling task. However, different prompt templates can greatly affect the results, and finding an appropriate template is difficult and time-consuming. To this end, this study proposes a novel hybrid prompt approach, which combines discrete prompts and continuous prompts, to motivate the model to learn more semantic knowledge from a small number of training samples. By comparing the performance difference between discrete prompts and continuous prompts, we find that hybrid prompts achieve the best results, reaching a 73.82% F1 value in the test set. In addition, we analyze the effect of different virtual token lengths in continuous prompts and hybrid prompts in a few-shot cross-language topic classification scenario. The results demonstrate that there is a threshold for the length of virtual tokens, and too many virtual tokens decrease the performance of the model. It is better not to exceed the average length of the training set corpus. Finally, this paper designs a method based on vector similarity to explore the real meanings represented by virtual tokens. The experimental results show that the prompt automatically learnt from the virtual token has a certain correlation with the input text.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a novel hybrid prompt approach, which combines discrete prompts and continuous prompts, to motivate the model to learn more semantic knowledge from a small number of training samples, and designs a method based on vector similarity to explore the real meanings represented by virtual tokens.'}",https://www.mdpi.com/2076-3417/13/17/9944/pdf?version=1693814326
-A Search for Prompts: Generating Structured Answers from Contracts,Adam Roegiest,"In many legal processes being able to action on the concrete implication of a legal question can be valuable to automating human review or signalling certain conditions (e.g., alerts around automatic renewal). To support such tasks, we present a form of legal question answering that seeks to return one (or more) fixed answers for a question about a contract clause. After showing that unstructured generative question answering can have questionable outcomes for such a task, we discuss our exploration methodology for legal question answering prompts using OpenAI's \textit{GPT-3.5-Turbo} and provide a summary of insights. Using insights gleaned from our qualitative experiences, we compare our proposed template prompts against a common semantic matching approach and find that our prompt templates are far more accurate despite being less reliable in the exact response return. With some additional tweaks to prompts and the use of in-context learning, we are able to further improve the performance of our proposed strategy while maximizing the reliability of responses as best we can.","{'model': 'tldr@v2.0.0', 'text': 'A form of legal question answering that seeks to return one (or more) fixed answers for a question about a contract clause is presented and it is found that the proposed template prompts are far more accurate despite being less reliable in the exact response return.'}",
-Successful Use of a Physiologically Acceptable Artificial Skin in the Treatment of Extensive Burn Injury,J. Burke,"A bilayer artificial skin composed of a temporary Silastic epidermis and a porous collagcn-chondroitin 6-sulfate fibrillar dermis, which is not removed, has been used to physiologically close up to 60% of the body surface following prompt excision of burn wounds in ten patients whose total burn size covered 50–95% body surface area (BSA). Following grafting, the dermal portion is populated with fibroblasts and vessels from the wound bed. The anatomic structure of the artificial dermis resembles normal dermis and serves as a template for the synthesis of new connective tissue and the formation of a “neodermis,” while it is slowly biodegraded. This artificial skin has physiologically closed excised burn wounds for periods of time up to 46 days before the Silastic epidermis was removed. At the time of election when donor sites are ready for reharvesting, the Silastic epidermis is removed from the vascularized artificial dermis and replaced with 0.004 auto-epidermal graft in sheet or meshed form. Clinical and histologic experience in a relatively short follow-up period (2–16 months) indicates that “neodermis” retains some of the anatomic characteristics and behavior of normal dermis, thus promising improvement in the functional and cosmetic results, as well as providing physiologic function as a skin substitute. The artificial skin is easily sterilized and stored at room temperature, capable of large scale production, and immediately available for grafting, indicating its potential for easy and relatively economic use in the burn patient","{'model': 'tldr@v2.0.0', 'text': 'Clinical and histologic experience in a relatively short follow-up period indicates that “neodermis” retains some of the anatomic characteristics and behavior of normal dermis, thus promising improvement in the functional and cosmetic results, as well as providing physiologic function as a skin substitute.'}",
-Structural snapshots of La Crosse virus polymerase reveal the mechanisms underlying Peribunyaviridae replication and transcription,B. Arragain,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that replication initiation is coupled to distal duplex promoter formation, endonuclease movement, prime-and-realign loop extension and closure of the polymerase core that direct the template towards the active site.'}",https://www.nature.com/articles/s41467-022-28428-z.pdf
-Wound tissue can utilize a polymeric template to synthesize a functional extension of skin.,I. Yannas,"Prompt and long-term closure of full-thickness skin wounds is guinea pigs and humans is achieved by applying a bilayer polymeric membrane. The membrane comprises a top layer of a silicone elastomer and a bottom layer of a porous cross-linked network of collagen and glycosaminoglycan. The bottom layer can be seeded with a small number of autologous basal cells before grafting. No immunosuppression is used and infection, exudation, and rejection are absent. Host tissue utilizes the sterile membrane as a culture medium to synthesize neoepidermal and neodermal tissue. A functional extension of skin over the entire wound area is formed in about 4 weeks.","{'model': 'tldr@v2.0.0', 'text': 'A functional extension of skin over the entire wound area is formed in about 4 weeks and no immunosuppression is used and infection, exudation, and rejection are absent.'}",
-Long-lived Ag10 6+ luminescence and a split DNA scaffold.,J. Petty,"Molecular silver clusters emit across the visible to near-infrared, and specific chromophores can be formed using DNA strands. We study C4AC4TC3G that selectively coordinates and encapsulates Ag10 6+, and this chromophore has two distinct electronic transitions. The green emission is strong and prompt with ϕ = 18% and τ = 1.25 ns, and the near-infrared luminescence is weaker, slower with τ = 50 µs, and is partly quenched by oxygen, suggesting phosphorescence. This lifetime can be modulated by the DNA host, and we consider two derivatives of C4AC4TC3G with similar sequences but distinct structures. In one variant, thymine was excised to create an abasic gap in an otherwise intact strand. In the other, the covalent phosphate linkage was removed to split the DNA scaffold into two fragments. In relation to the contiguous strands, the broken template speeds the luminescence decay by twofold, and this difference may be due to greater DNA flexibility. These modifications suggest that a DNA can be structurally tuned to modulate metastable electronic states in its silver cluster adducts.","{'model': 'tldr@v2.0.0', 'text': 'Modifications of C4AC4TC3G suggest that a DNA can be structurally tuned to modulate metastable electronic states in its silver cluster adducts.'}",
-The formless void as organizational template☆,T. Harle,"This article examines the first creation story in the Book of Genesis from the perspective of complexity theory. Traditional views of this narrative conceive of a deity harnessing chaos to create order. They reflect a loathing of chaos and a desire for control: a Newtonian worldview of cause and effect. It is an approach we see repeated in contemporary organizations. However, a radically different perspective sees the formless void as opening up the possibility of emergence and self-organization. This approach is used to prompt interaction both with the Judaeo-Christian tradition and organization studies. Four particular leadership challenges are examined: notions of control, attitudes to change, co-creating an environment for self-organization, and using emergence in relation to values. A concluding section highlights the importance of consistency in leadership to promote emergence.",,
-Overview of Tertiary Addictions Services response to Opioid dependence during the COVID-19 Pandemic.,K. Hennigan,"The emergence of the COVID-19 pandemic has presented the addiction services with an unprecedented set of challenges. Opioid users are particularly vulnerable because of their high level of pre-existing health problems and lifestyle factors. In order to minimise their risks to self and to others in the current Covid-19 crisis, the service has sought to urgently identify vulnerable individuals, and induct them into OST treatment promptly. Additionally, several guidelines have been created and regularly updated by the HSE for any healthcare staff working with opioid users. These include guidance documents, to facilitate prompt induction of patients onto the OST programme, the prescribing of naloxone to all patients at risk of overdose, eConsultation, medication management for those in self-isolation, and the delivery of injecting equipment. The guidance documents and resources will provide a template for a new way of working for the sector during these challenging times and into the future.","{'model': 'tldr@v2.0.0', 'text': 'Several guidelines have been created and regularly updated by the HSE for any healthcare staff working with opioid users to facilitate prompt induction of patients onto the OST programme, and the prescribing of naloxone to all patients at risk of overdose.'}",
-Grounding Language to Entities and Dynamics for Generalization in Reinforcement Learning,,"To collect the text manual, we first crowdsource 82 templates (with 2,214 possible descriptions after filling in the blanks). Each Amazon Mechanical Turk worker is asked to paraphrase a prompt sentence while preserving words in boldface (which become the blanks in our templates). We have three blanks per template, one each for the entity, role and an adjective. For each role (enemy, message, goal) we have three role words and three adjectives that are synonymous (Table 5). Each entity is also described in three synonymous ways. Thus, every entity-role assignment can be described in 27 different ways on the same template. Raw templates are filtered for duplicates, converted to lowercase, and corrected for typos to prevent confusion on downstream collection tasks.","{'model': 'tldr@v2.0.0', 'text': None}",
-Kinect-based intelligent monitoring and warning of students' sitting posture,Heng Sun,"This paper proposes a Kinect-based software design for intelligent detection and warning of students' sitting posture. The plan includes three modules: video acquisition, preprocessing, and sitting posture recognition. Two methods, template matching method based on bone data and feature learning method based on Convolutional neural network (CNN), are used to realize sitting posture recognition. This design first uses Kinect to complete video capture, and secondly, preprocesses the real-time image frames in the video stream to extract the region of interest. Calculate the spatial position of the main joint points of the upper body of the human body again, and convert them into different spatial coordinate systems to locate the target. Finally, the best result of the two methods is used as the recognition result. When an incorrect posture is judged, a corresponding prompt or warning is issued to the supervised person. The experiment was carried out to realize the real-time detection and judgment of the student's sitting posture and timely notification of bad posture, which verified the software design scheme's feasibility.","{'model': 'tldr@v2.0.0', 'text': ""The experiment was carried out to realize the real-time detection and judgment of the student's sitting posture and timely notification of bad posture, which verified the software design scheme's feasibility.""}",
-Circulating miRNAs with prognostic value in metastatic breast cancer and for early detection of metastasis.,Dharanija Madhavan,"Metastasis is the principal cause of high morbidity and mortality among breast cancer (BC) patients. Identification of markers that can be routinely monitored to predict onset of metastasis in BC patients and prognosis of metastatic breast cancer (MBC) patients would increase their median survival. In this study, plasma miRNAs of 40 MBC patients were profiled by TaqMan low density arrays and miRNAs with prognostic capacity were identified. The candidates were validated initially in the samples of 237 MBC patients and subsequently in 335 samples from an independent study cohort of BC patients. Sixteen miRNAs were established to be significantly associated with overall survival, and were termed as prognostic miRNA panel template (PROMPT). These included miR-141, miR-144, miR-193b, miR-200a, miR-200b, miR-200c, miR-203, miR-210, miR-215, miR-365, miR-375, miR-429, miR-486-5p, miR-801, miR-1260 and miR-1274a. Additionally, 11 of these miRNAs were also associated with progression-free survival. Their prognostic significance was further confirmed in samples from a second study cohort of BC patients. In addition, miR-200a, miR-200b, miR-200c, miR-210, miR-215 and miR-486-5p were found to be significantly associated with onset of metastasis up to 2 years prior to clinical diagnosis in BC patients. We have thus identified panels of miRNAs, which include metastasis promoting miR-200 family and miR-203, as well as oncogenic and tumor-suppressive miRNAs, that can serve as prognostic markers for MBC, and early detection markers of metastasis in BC.","{'model': 'tldr@v2.0.0', 'text': 'Panel of miRNAs, which include metastasis promotingmiR-200 family and miR-203, as well as oncogenic and tumor-suppressive mi RNAs, that can serve as prognostic markers for MBC, and early detection markers of metastasis in BC are identified.'}",https://academic.oup.com/carcin/article-pdf/37/5/461/8193240/bgw008.pdf
-Kilohertz gravitational waves from binary neutron star remnants: Time-domain model and constraints on extreme matter,M. Breschi,"The remnant star of a neutron star merger is an anticipated loud source of kiloHertz gravitational waves that conveys unique information on the equation of state of hot matter at extreme densities. Observations of such signals are hampered by the photon shot noise of ground-based interferometers and pose a challenge for gravitational-wave astronomy. We develop an analytical time-domain waveform model for postmerger signals informed by numerical relativity simulations. The model completes effective-one-body waveforms for quasi-circular nonspinning binaries in the kiloHertz regime. We show that a template-based analysis can detect postmerger signals with a minimal signal-to-noise ratios (SNR) of 8, corresponding to GW170817-like events for third-generation interferometers. Using Bayesian model selection and the complete inspiral-merger-postmerger waveform model it is possible to infer whether the merger outcome is a prompt collapse to a black hole or a remnant star. In the latter case, the radius of the maximum mass (most compact) nonrotating neutron star can be determined to kilometer precision. We demonstrate the feasibility of inferring the stiffness of the equation of state at extreme densities using the quasiuniversal relations deduced from numerical-relativity simulations.",,https://www.repository.cam.ac.uk/bitstream/1810/301786/3/PhysRevD.100.104029.pdf
-Financial scenario modelling: a study of UK universities,P. Cropper,"ABSTRACT Financial scenario modelling appears to offer the potential to help universities cope with an uncertain funding environment. By means of a questionnaire survey of UK universities, supported by interviews with members of finance departments, this paper explores the prevalence and construction of financial models used for scenario analysis. According to the findings, most universities undertake such modelling; but a minority do not, even though it is officially required. There is some variation in the purposes for which the models are constructed and in the sophistication of the approaches taken, but the overall impression is of a relatively straightforward approach, with spreadsheets generally viewed as having sufficient functionality to meet current strategic needs. There is no appetite for a standard template to be imposed or even offered by the funding councils. Sharing experience is likely to be a more welcome approach if growing pressures prompt finance departments to strengthen their modelling.",,https://pure.hud.ac.uk/ws/files/35714977/AMENDED_JHEPM_scen_R1_full.pdf
-Cardiogenic shock: evolving definitions and future directions in management,Tara L Jones,"Cardiogenic shock (CS) is a complex and highly morbid entity conceptualised as a vicious cycle of injury, cardiac and systemic decompensation, and further injury and decompensation. The pathophysiology of CS is incompletely understood but limited clinical trial experience suggests that early and robust support of the failing heart to allow for restoration of systemic homoeostasis appears critical for survival. We review the pathophysiology, clinical features and trial data to construct a contemporary model of CS as a systemic process characterised with maladaptive compensatory mechanisms requiring prompt and appropriately tailored medical and mechanical support for optimal outcomes. We conclude with an algorithmic approach to acute CS incorporating clinical and haemodynamic data to match the patient’s cardiac and systemic needs as a template for contemporary management.","{'model': 'tldr@v2.0.0', 'text': 'An algorithmic approach to acute CS is taken incorporating clinical and haemodynamic data to match the patient’s cardiac and systemic needs as a template for contemporary management.'}",https://openheart.bmj.com/content/openhrt/6/1/e000960.full.pdf
-A brachytherapy template approach to standardize saturation prostatic biopsy,Simon Bott,"The increasing use of PSA testing has resulted in more men undergoing prostatic biopsy to diagnose or exclude malignancy. However, the standard sextant or octant biopsy regimen is reported to miss 15–35% of ‘significant’ prostate cancers [1–3]. The finding of ‘no evidence of malignancy’ on standard biopsy in the presence of a rising PSA level, a rapid PSA doubling time or a suspicious DRE may prompt further biopsies. Several studies have reported sites inadequately sampled using standard techniques, including the anterior transition zone superior and lateral to the urethra, the inferior part of the anterior horn where the peripheral zone wraps round the transition zone, and the midline of the peripheral zone [2,4,5]. Increasing the number of biopsies and targeting these sites after a negative standard prostatic biopsy results in increased cancer detection [1,2,5]. Saturation biopsy involves taking a large number of biopsies (14–45) using a periprostatic block, sedation, spinal or general spinal anaesthesia [6–8]. In men with a clinical suspicion of prostate cancer and previously negative sextant biopsies, transrectal saturation biopsies have a cancer detection rate of 14–34% [6–8]. We describe a systematic and comprehensive technique to take saturation prostatic biopsies through the perineum.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and comprehensive technique to take saturation prostatic biopsies through the perineum is described, which is reported to miss 15–35% of ‘significant’ prostate cancers and in men with a clinical suspicion of prostate cancer and previously negative sextantBiopsies have a cancer detection rate of 14–34%.'}",
-Pre-notification and reminder SMS text messages with behaviourally informed invitation letters to improve uptake of NHS Health Checks: a factorial randomised controlled trial,A. Sallis,,"{'model': 'tldr@v2.0.0', 'text': 'This large randomised controlled trial adds support to the evidence that small, low cost behaviourally informed changes to letter-based invitations can increase uptake of NHS HCs and provides novel evidence on the effect of SMS reminders and pre-notification on NHS HC attendance.'}",https://bmcpublichealth.biomedcentral.com/track/pdf/10.1186/s12889-019-7476-8
-Variations in Influenza Vaccination by Clinic Appointment Time and an Active Choice Intervention in the Electronic Health Record to Increase Influenza Vaccination,Rebecca H Kim,"Key Points Question Do influenza vaccination rates vary by clinic appointment time and can an active choice intervention in the electronic health record directed to medical assistants improve vaccination rates in primary care practices? Findings In this quality improvement study of 11 primary care practices and 96 291 patients, influenza vaccination rates significantly declined as the clinic day progressed. Primary care practices that implemented an active choice intervention in the electronic health record to prompt medical assistants to ask patients about influenza vaccination and template vaccination orders for clinicians to review were associated with a significant increase in influenza vaccination rates when compared with a control group of primary care practices. Meaning Reminders in the electronic health record can improve vaccination rates overall, but other strategies may also be needed to address variations by time of day.","{'model': 'tldr@v2.0.0', 'text': 'Primary care practices that implemented an active choice intervention in the electronic health record to prompt medical assistants to ask patients about influenza vaccination were associated with a significant increase in influenza vaccination rates when compared with a control group of primary care practices.'}",https://jamanetwork.com/journals/jamanetworkopen/articlepdf/2702210/kim_2018_oi_180107.pdf
-Synthesis and Anticancer Properties of ‘Azole’ Based Chemotherapeutics as Emerging Chemical Moieties: A Comprehensive Review,P. Prasher," Azole frameworks serve as privileged scaffolds in the contemporary drug design paradigm owing to their unique physicochemical profile that promotes the development of highly selective, physiological benevolent chemotherapeutics. Several azole nuclei function as bioisostere in medicinal chemistry and prompt the development of tailored therapeutics for targeting the desired biological entities. Besides, the azole scaffold forms an integral part in the advanced drug designing methodologies, such as target template insitu drug synthesis, that assists in rapid identification of the hit molecules form a diverse pool of leads; and direct biomolecule-drug conjugation, along with bioorthogonal strategies that ensure localization, and superior target specificity of the directed therapeutic. Lastly, the structural diversity of azole framework and high yielding click synthetic methods provide a comprehensive Structure-Activity Relationship analysis for design optimization of the potential drug molecules by fine-tuning the placement of different substituents critical for the activity. This review provides a comprehensive analysis of the synthesis and anticancer potential of azole based chemotherapeutics. ","{'model': 'tldr@v2.0.0', 'text': 'The structural diversity of azole framework and high yielding click synthetic methods provide a comprehensive Structure-Activity Relationship analysis for design optimization of the potential drug molecules by fine-tuning the placement of different substituents critical for the activity.'}",
-Comparing emissions mitigation efforts across countries,Joseph E. Aldy,"A natural outcome of the emerging pledge and review approach to international climate change policy is the interest in comparing mitigation effort among countries. Domestic publics and stakeholders will have an interest in knowing if peer countries are undertaking (or planning to undertake) comparable efforts in mitigating their GHG emissions. Moreover, if the aggregate effort is considered inadequate in addressing the risks posed by climate change, then this will likely prompt a broader interest in identifying those countries where greater effort is arguably warranted based on comparison with their peers. Both assessments require metrics of effort and comparisons among countries. We propose a framework for such an exercise, drawing from a set of principles for designing and implementing informative metrics. We present a template for organizing metrics on mitigation effort, for both ex ante and ex post review. We also provide preliminary assessments of effort along emissions, price, and cost metrics for post-2020 climate policy contributions by China, the European Union, Russia, and the United States. We close with a discussion of the role of academics and civil society in promoting transparency and facilitating the evaluation and comparison of effort. Policy relevance Statement: Our article presents a framework for the review of intended nationally determined contributions and the ex post review of contributions under the UNFCCC negotiations. We provide an illustration of this framework with an energy-economic model. Our work focuses on how countries may use the review to compare mitigation effort – planned under INDCs and delivered by implementation of the pledged contributions – to address concerns about equity, efficiency, competitiveness, and the stability of any agreement that arise in international negotiations.",,https://dash.harvard.edu/bitstream/1/30203383/6/Comparability%20APA%20Paper%20150930%20Final.pdf
-Factors Underlying Quality Problems with Alcohol Screening Prompted by a Clinical Reminder in Primary Care: A Multi-site Qualitative Study,E. Williams,,"{'model': 'tldr@v2.0.0', 'text': 'This qualitative study in nine clinics found that implementation of alcohol screening facilitated by a clinical reminder resulted primarily in verbal screening in which questions were not asked vertbatim and were otherwise adapted, addressing limitations of the clinical reminder and issues related to perceived discomfort.'}",https://link.springer.com/content/pdf/10.1007/s11606-015-3248-z.pdf
-An Electrochemical Non-Enzymatic Glucose Sensor Based on Ultrathin PdAg Single-Crystalline Nanowires.,Xiaoxiang Xu,"Electrochemical non-enzymatic sensors have great potential for prompt and efficient detection of glucose. Herein, a novel, highly efficient electrochemical non-enzymatic glucose sensor is reported that is based on ultrathin PdAg single-crystalline nanowires (NWs). Ultrathin PdAg NWs are fabricated by a facile one-pot aqueous synthesis through an in situ growth strategy with an amphiphilic surfactant as the template. A comparison of the activities of PdAg NWs with different compositional ratios and nanostructures shows that ultrathin Pd2 Ag1 NWs hold the best performance toward electrochemical detection of glucose with an operable sensitivity of 11.6 μA mM-1 cm-2 and a linear response range of 0.1-8 mM. Structural and compositional features of the Pd2 Ag1 NWs allow an excellent selectivity, rapid response, and good long-term stability for electrochemical glucose sensor. This work thus provides a new possibility for the rational design and synthesis of noble-metal-based nanomaterials for non-enzymatic sensors.","{'model': 'tldr@v2.0.0', 'text': 'A novel, highly efficient electrochemical non-enzymatic glucose sensor is reported that is based on ultrathin PdAg single-crystalline nanowires (NWs) that provides a new possibility for the rational design and synthesis of noble-metal-based nanomaterials for non- enzymatic sensors.'}",
-Liquid xenon scintillation measurements and pulse shape discrimination in the LUX dark matter detector,D. Akerib,"Weakly Interacting Massive Particles (WIMPs) are a leading candidate for dark matter and are expected to produce nuclear recoil (NR) events within liquid xenon time-projection chambers. We present a measurement of liquid xenon scintillation characteristics in the LUX dark matter detector and develop a pulse shaped based discrimination parameter to be used for particle identification. To accurately measure the scintillation characteristics, we develop a template-fitting method to reconstruct the detection time of photons. Analyzing calibration data collected during the 2013-16 LUX WIMP search, we measure a singlet-to-triplet scintillation ratio for electron recoils (ER) that is consistent with existing literature, and we make a first-ever measurement of the NR singlet-to-triplet ratio at recoil energies below 74 keV. A prompt fraction discrimination parameter exploits the difference of the photon time spectra for NR and ER events and is optimized to have the least number of ER events that occur in the 50\% NR acceptance region. When this discriminator is used in conjunction with charge-to-light discrimination on the calibration data, the signal-to-noise ratio in the NR dark matter acceptance region increases by up to a factor of two.",,https://link.aps.org/accepted/10.1103/PhysRevD.97.112002
-Pokérator - Unveil your inner Pokémon,Dominique Geissler,"The Pokérator is a generator of Pokémon names and descriptions, based on user input. The names are generated by blending words based on syllables or characters according to a bigram language model. An accompanying description is generated by filling a template with ConceptNet answers. This sentence is then used as a prompt for text generation with the GPT-2 language model which was finetuned on Pokédex entries. The evaluation of the generated Pokémon names shows that the names are not realistic, but appreciated and creative.","{'model': 'tldr@v2.0.0', 'text': 'The Pokérator is a generator of Pokémon names and descriptions, based on user input, which shows that the names are not realistic, but appreciated and creative.'}",
-Measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at √ s = 7 TeV,A. Belyaev,"A measurement of differential cross sections for the production of a pair of isolated photons in proton–proton collisionsat √ s = 7 TeVispresented.Thedatasamplecorre-sponds to an integrated luminosity of 5.0 fb − 1 collected with the CMS detector. A data-driven isolation template method is used to extract the prompt diphoton yield. The measured cross section for two isolated photons, with transverse energy above 40 and 25 GeV respectively, in the pseudorapidity range | η | < 2 . 5, | η | / ∈ [ 1 . 44 , 1 . 57 ] and with an angular separation (cid:3) R > 0 . 45, is 17 . 2 ± 0 . 2 (stat) ± 1 . 9 (syst) ± 0 . 4 (lumi)pb. Differential cross sections are measured as a function of the diphoton invariant mass, the diphoton transverse momentum, the azimuthal angle difference between the two photons, and the cosine of the polar angle in the Collins–Soper reference frame of the diphoton system. The results are compared to theoretical predictions at leading, next-to-leading, and next-to-next-to-leading order in quantum chromodynamics.",,
-OPTICAL AND NEAR-INFRARED OBSERVATIONS OF SN 2013DX ASSOCIATED WITH GRB 130702A,V. Toy,"We present optical and near-infrared (NIR) light curves and optical spectra of SN 2013dx, associated with the nearby (redshift 0.145) gamma-ray burst GRB 130702A. The prompt isotropic gamma-ray energy released from GRB 130702A is measured to be E γ , iso = 6.4 − 1.0 + 1.3 × 10 50 ?> erg (1 keV to 10 MeV in the rest frame), placing it intermediate between low-luminosity GRBs like GRB 980425/SN 1998bw and the broader cosmological population. We compare the observed g ′ r ′ i ′ z ′ ?> light curves of SN 2013dx to a SN 1998bw template, finding that SN 2013dx evolves ∼20% faster (steeper rise time), with a comparable peak luminosity. Spectroscopically, SN 2013dx resembles other broad-lined SNe Ic, both associated with (SN 2006aj and SN 1998bw) and lacking (SN 1997ef, SN 2007I, and SN 2010ah) gamma-ray emission, with photospheric velocities around peak of ∼ 21,000 km s−1. We construct a quasi-bolometric ( g ′ r ′ i ′ z ′ yJ ?> ) light curve for SN 2013dx, only the fifth GRB-associated SN with extensive NIR coverage and the third with a bolometric light curve extending beyond Δ t > 40 days ?> . Together with the measured photospheric velocity, we derive basic explosion parameters using simple analytic models. We infer a 56Ni mass of M Ni = 0.37 ± 0.01 ?> M ⊙ ?> , an ejecta mass of M ej = 3.1 ± 0.1 ?> M ⊙ ?> , and a kinetic energy of E K = ( 8.2 ± 0.43 ) × 10 51 ?> erg (statistical uncertainties only), consistent with previous GRB-associated supernovae. When considering the ensemble population of GRB-associated supernovae, we find no correlation between the mass of synthesized 56Ni and high-energy properties, despite clear predictions from numerical simulations that M Ni ?> should correlate with the degree of asymmetry. On the other hand, M Ni ?> clearly correlates with the kinetic energy of the supernova ejecta across a wide range of core-collapse events.",,https://iopscience.iop.org/article/10.3847/0004-637X/818/1/79/pdf
-"Measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at $$\sqrt{s}=7\,\text {TeV} $$s=7TeV",S. Chatrchyan,,,https://link.springer.com/content/pdf/10.1140/epjc/s10052-014-3129-3.pdf
-Writing in Math: A Disciplinary Literacy Approach,W. Brozo,"ABSTRACT Mathematics teachers often resist generic literacy strategies because they do not seem relevant to math learning. Discipline-specific literacy practices that emerge directly from the math content and processes under study are more likely to be embraced by math teachers. Furthermore, national and state-level mathematics standards as well as Common Core standards provide frameworks for situating literacy practices squarely within the disciplines. A disciplinary literacy approach to writing in math requires teachers to develop innovative strategies and practices that link writing to particular mathematical processes and tasks. An example is shared of a math writing approach developed by a middle school teacher used to prompt her students' critical thinking and problem solving processes during the study of algebra. She designed a template that when completed can serve as a reflective tool for her students and provide the teacher useful feedback on their learning. The example of teaching with the template as a guide for working through steps to solve a story problem demonstrates what disciplinary writing can look like in a typical middle school classroom.",,
-The Clinician as Clinical Ethics Consultant: An Empirical Method of Study,D. Kornfeld,"Some 30 years ago the role of the clinical ethics consultant (CEC) was formalized. At the time, the perception of the role differed between two groups serving in that capacity, clinicians and nonclinicians. Differences in their roles reflected their training and experience. These divergent views were resolved semantically by designating the role of the CEC as “ethics facilitation.” In practice the different perspectives have remained. However, the subsequent published literature on clinical ethics consultation has not adequately reflected the activity of the clinician as a CEC. There have been recurring unanswered calls for the acquisition of empirical data on the nature of the problems that prompt ethics consultation requests and the functions required to address them. The authors introduce a template that provides a means to acquire such data for clinician ethicists. A similar instrument could be constructed to reflect the role of the nonclinician ethicist serving in that capacity.","{'model': 'tldr@v2.0.0', 'text': 'A template is introduced that provides a means to acquire empirical data on the nature of the problems that prompt ethics consultation requests and the functions required to address them and a similar instrument could be constructed to reflect the role of the nonclinician ethicist serving in that capacity.'}",
-Women’s views on anxiety in pregnancy and the use of anxiety instruments: a qualitative study,K. Evans,"Abstract Objective: To explore women’s experience of anxiety in pregnancy and views on the use of anxiety instruments in antenatal care. Background: Anxiety in pregnancy is associated with adverse birth outcomes, developmental and behavioural problems in infants and postnatal depression. Despite recommendations for routine psychological assessment in pregnancy, the optimal methods to identify anxiety in pregnancy have not been confirmed. Methods: A qualitative study using two focus group discussions was undertaken. Focus group one included women in a community setting and focus group two included women in a hospital clinic setting who had received additional support for anxiety in pregnancy. Participants were women who had given birth within the past nine months and considered themselves to have been anxious during their pregnancy. Results: Three main themes were identified using template analysis: sources of support, administration of anxiety instruments and the use of instruments to prompt discussion. Women stated that anxiety instruments could help them to identify their anxious feelings and prompt a discussion around those feelings. However, they expressed concerns surrounding the administration of anxiety instruments and questioned how useful they would be in helping women access help and support. Conclusions: The introduction of anxiety instruments in antenatal care may present an opportunity to discuss women’s emotional health and anxieties. Providing women with sufficient time to discuss their anxious feelings, identified by such instruments, could facilitate access to additional support.","{'model': 'tldr@v2.0.0', 'text': 'The introduction of anxiety instruments in antenatal care may present an opportunity to discuss women’s emotional health and anxieties, and providing women with sufficient time to discuss their anxious feelings, identified by such instruments, could facilitate access to additional support.'}",https://nottingham-repository.worktribe.com/preview/829175/JRIP%20FG%20Anx%20KE%20resubmission%20text%20%28non-anon%29%2011.7.16%20proof.pdf
-Linguistics Vanguard,Mark B. Turner,A range of multimodal form-meaning pairs has arisen to prompt for the generic integration template called blended classic joint attention (BCJA). This article presents examples and principles.,,
-PCR Detection of Escherichia coli O157:H7 Directly from Stools: Evaluation of Commercial Extraction Methods for Purifying Fecal DNA,J. Holland,"ABSTRACT Rapid identification of Escherichia coli O157:H7 is important for patient management and for prompt epidemiological investigations. We evaluated one in-house method and three commercially available kits for their ability to extract E. coli O157:H7 DNA directly from stool specimens for PCR. Of the 153 stool specimens tested, 107 were culture positive and 46 were culture negative. The sensitivities and specificities of the in-house enrichment method, IsoQuick kit, NucliSens kit, and QIAamp kit were comparable, as follows: 83 and 98%, 85 and 100%, 74 and 98%, and 86 and 100%, respectively. False-negative PCR results may be due to the presence of either inherent inhibitors or small numbers of organisms. The presence of large amounts of bacteria relative to the amount of the E. coli O157:H7 target may result in the lower sensitivities of the assays. All commercial kits were rapid and easy to use, although DNA extracted with the QIAamp kit did not require further dilution of the DNA template prior to PCR.","{'model': 'tldr@v2.0.0', 'text': 'All commercial kits were rapid and easy to use, although DNA extracted with the QIAamp kit did not require further dilution of the DNA template prior to PCR, and the sensitivities and specificities of the in-house enrichment method, IsoQuick kit, NucliSens kit, and QIAamps kit were comparable.'}",https://jcm.asm.org/content/jcm/38/11/4108.full.pdf
-Improving Human Papillomavirus (HPV) Vaccination in the Postpartum Setting,R. Soon,,"{'model': 'tldr@v2.0.0', 'text': 'A drop-down menu that was inserted into the EMR postpartum note template inquiring about the patient’s HPV vaccination status and whether or not the vaccine was administered at that visit could be an effective way to promote HPV vaccination in the post partum setting.'}",https://europepmc.org/articles/pmc5510873?pdf=render
-Chapter 9. Interfaces for Capturing the Experiment Record,Cerys Willoughby,"Electronic Laboratory Notebooks (ELNs) have many features that can support the work of scientists, but their main purpose is to replace the paper notebook in the laboratory. When we capture notes in a paper notebook, we have the flexibility to record our research in any format we choose. When we capture notes using an electronic notebook, and particularly using ELNs, the way that we capture information changes. Many ELNs make use of templates to encourage the capture of standard information and provide the opportunity to prompt users to capture items of interest that they might otherwise omit or forget to record. Using the cues provided by the formal structure, templates should improve the consistency and quality of records. There is, however, the possibility that templates may have negative as well as positive impacts because they may restrict the content that is recorded, and information that is not specifically prompted for in the template may be omitted, even if it is significant to the record. In this chapter, we will examine research into the impacts of different kinds of interfaces on the information that is captured in the laboratory. We also take a look at the role of mobile applications in bridging the gap between paper notebooks and ELNs, and how other kinds of interfaces may better support users in capturing their research in electronic form.","{'model': 'tldr@v2.0.0', 'text': 'Research into the impacts of different kinds of interfaces on the information that is captured in the laboratory is examined, including the role of mobile applications in bridging the gap between paper notebooks and ELNs, and how other types of interfaces may better support users in capturing their research in electronic form.'}",
-Expectancies and Motivations to Attend an Informal Science Lecture Series,Niveen Abighannam,"This study explored the expectancies and motivations that prompt audiences to attend a university science lecture series. The series features talks by science experts from the host campus and around the USA. Each lecture typically attracts between 300 and 600 attendees, including middle and high school student groups, university students, and families and adults from the area. We conducted 47 semi-structured interviews with attendees in order to evaluate their expectancies and motivations. A template analysis of the interviews was grounded in social cognitive and self-determination theories. Results suggest that participants were mostly driven by intrinsic motivations and acquired strong sensory outcome expectancies, such as novelty and activity. Participants also held physical outcome expectancies, such as social expectancies, though to a lesser extent. Both intrinsic and extrinsic motivations to attend the lecture series were associated with expectancies held prior to the event. Of those expectancies, the novelty, entertainment, and social outcomes were dominant. Other noteworthy outcome expectancies include status and self-reactivity. Parents and teachers held outcome expectancies, not only for themselves, but also for their children and students who attended the talks with them.",,
-Optical and conductive properties of as-synthesized organic-capped TiO₂ nanorods highly dispersible in polystyrene-block-poly(methyl methacrylate) diblock copolymer.,L. Cano,"As-synthesized organic-capped TiO2 nanorods were incorporated into polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) diblock copolymer to achieve TiO2/PS-b-PMMA nanocomposites with enhanced optical and conductive properties. The specific surface chemistry of TiO2 nanorods derived from the colloidal synthetic approach allowed their prompt incorporation in the PS-b-PMMA block copolymer template up to 50 wt %, which resulted in films with an extended coverage of highly dispersed nanoparticles for contents higher than 30 wt %. At such high nanorod contents, the films fabricated by the prepared nanocomposites demonstrated enhanced optical properties. Atomic force microscopy investigation of the nanocomposite films showed a cylindrical morphology for low nanorod contents. Conversely, higher nanorod contents resulted upon removal of the organic component in the nanocomposites with UV treatment in overall nanorod coverage of the film surface with the concomitant formation of charge percolation paths, which led to noticeable conductivity values. EFM and PF-TUNA measurements confirmed the conductive properties of the composites at nanoscale, whereas semiconductor analyzer measurements provided their macroscale characterization. In addition, an increase in the UV-vis absorption was observed with the increase in the nanorod content along with a remarkable conductivity of the overall film.","{'model': 'tldr@v2.0.0', 'text': 'The specific surface chemistry of TiO2 nanorods derived from the colloidal synthetic approach allowed their prompt incorporation in the PS-b-PMMA block copolymer template up to 50 wt %, which resulted in films with an extended coverage of highly dispersed nanoparticles for contentsHigher nanorod contents led to noticeable conductivity values.'}",
-Use of a nursing checklist to facilitate implementation of therapeutic hypothermia after cardiac arrest.,K. Avery,"Therapeutic hypothermia has become a widely accepted intervention that is improving neurological outcomes following return of spontaneous circulation after cardiac arrest. This intervention is highly complex but infrequently used, and prompt implementation of the many steps involved, especially achieving the target body temperature, can be difficult. A checklist was introduced to guide nurses in implementing the therapeutic hypothermia protocol during the different phases of the intervention (initiation, maintenance, rewarming, and normothermia) in an intensive care unit. An interprofessional committee began by developing the protocol, a template for an order set, and a shivering algorithm. At first, implementation of the protocol was inconsistent, and a lack of clarity and urgency in managing patients during the different phases of the protocol was apparent. The nursing checklist has provided all of the intensive care nurses with an easy-to-follow reference to facilitate compliance with the required steps in the protocol for therapeutic hypothermia. Observations of practice and feedback from nursing staff in all units confirm the utility of the checklist. Use of the checklist has helped reduce the time from admission to the unit to reaching the target temperature and the time from admission to continuous electroencephalographic monitoring in the cardiac intensive care unit. Evaluation of patients' outcomes as related to compliance with the protocol interventions is ongoing.","{'model': 'tldr@v2.0.0', 'text': 'A checklist was introduced to guide nurses in implementing the therapeutic hypothermia protocol during the different phases of the intervention (initiation, maintenance, rewarming, and normothermia) in an intensive care unit and has provided all of the intensive care nurses with an easy-to-follow reference to facilitate compliance with the required steps in the protocol.'}",http://ccn.aacnjournals.org/content/35/1/29.full.pdf
-Measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength,S. Chatrchyan,,,https://link.springer.com/content/pdf/10.1140/epjc/s10052-014-3129-3.pdf
-A Successful Competency-Based Writing Skills Development Programme: Results of an Experiment1,R. Craig,"We describe a successful, semester-length writing skills development programme conducted at the University of Toronto Mississauga (UTM) with intermediate level undergraduate financial accounting students. The programme focused on improving students' writing in five competency areas: organization, grammar, style, professional writing and case writing. Effectiveness was assessed by means of a quasi-experiment involving a pre-test, post-test design with a treatment group and a control group. Comparison of entry and exit test scores of students in the treatment group with those in the control group revealed a highly statistically significant increase in the writing skills of the treatment group. The programme outlined is not presented as a template for emulation, but rather to prompt educators to ponder how they might address the challenge of improving (and assessing) the writing skills of students.",,
-See Yourself Write: A Simple Student Model to Make Students Think,S. Bull,,"{'model': 'tldr@v2.0.0', 'text': 'This paper introduces See Yourself Write, a system with two components: a template for tutors to provide feedback to their students on their foreign language writing, and an inspectable student model which is automatically constructed for each student, based on the feedback given by the teacher.'}",
-Life defined,Andrew Moore,"I have to admit, I’ve been waiting for a particular event to write this editorial. My prompt is the recent opinion piece ‘‘What is life?’’ [1] published in The Scientist, which authoritatively represents the problems of, and approaches to, defining life, ultimately settling on self-replication as its essence: self-replicating RNA might even be considered a primitive lifeform. Self-replication is a feature of life, alright, but I fundamentally differ that it defines life in terms of being its essence. Inanimate structures in nature can replicate themselves on a template too: minute fragments of a crystal can act as seeds for other crystals to grow. And although the complexity of a crystal structure is nothing approaching that of a polynucleotide, the physical principle is similar. Even the property of catalysis in a self-replicating RNA witnesses a physicochemical principle that is widespread in the inanimate realm. In order to see what is truly ‘‘special’’ about life – in contrast to non-living nature – I believe there is something more important than replication that we need to be thinking about: the way in which living organisms use energy gradients. Life, I would argue, is a combination of the ability to harness energy gradients to produce a local reduction in entropy – i.e. local increase in order/ organisational complexity of matter – coupled with the ability to replicate chemically coded information and reproduce (not replicate) the housing of that information (be it a single cell or multi-celled organism) (Fig. 1). But at the heart of it all is energy metabolism. Energy can only be harnessed via a physical structure that uses an energy gradient – in tangible terms, the natural flux of hot > cold, culminating in everything being ‘‘warm’’; for biological organisms it is potential energy in the form of an ion gradient, chemical bonds or oxidation state. The newest theories on the origins of life indeed argue that it all began with the exploitation of a naturally occurring physicochemical (energy) gradient [2, 3] by pre-biotic membranes. Thermodynamics play a role in another crucially special feature of life: thermodynamic renewal via reproduction. Renewal is only possible as a result of the subversion of energy fluxes that life performs – i.e. to produce a highly complex physicochemical information system that, despite becoming ‘‘degraded’’ via ageing, is able to produce a pristine new organism to carry on the procession of life [4]. Upon reproduction, degraded components are sequestered into the ‘‘mother’’ organism (even in the case of bacterial cells). Here, again we find a force working against the ‘‘inanimate’’ flux of energy and matter towards equilibrium: a force derived from life’s use of energy gradients. Renewal via reproduction is why life exploded and diversified on Earth, creating new order as Earth and the rest of the universe continue their predictable journey towards equilibrium – spectacular though this inanimate physical theatre is. Coupled with replication of the genetic code, renewal-via-reproduction imparts to life its additional characteristic of evolution. Life essentially manifests a physical principle that is not seen elsewhere: the ability to preserve itself (often down to the physical form of an organism) whilst all around it equilibrates. Life has ‘‘wandered’’ into a realm of physics that the inanimate universe does not ‘‘use’’. Every day until we die, our living organism is – in a sense – playing a game with the second law of thermodynamics, channeling the flux of energy toward maintaining complexity and order in contrast to the inexorable process of equilibration surrounding it in the inorganic environment: a game that no non-living system can play. It is also no accident – in fact it might even be considered a tautology to state – that humans employ this principle in the production of machines and engines that similarly precipitate, and then use, a thermodynamic flow from hot to cold to do useful work: constructing local order (decrease in entropy) that","{'model': 'tldr@v2.0.0', 'text': 'This editorial argues that there is something more important than replication that the authors need to be thinking about: the way in which living organisms use energy gradients, andThermodynamics play a role in another crucially special feature of life: thermodynamic renewal via reproduction.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/bies.201290011
-Stable and Noncompetitive RNA Internal Control for Routine Clinical Diagnostic Reverse Transcription-PCR,K. Dingle,"ABSTRACT Clinical diagnostic tests based on nucleic acid amplification assist with the prompt diagnosis of microbial infections because of their speeds and extremely low limits of detection. However, the design of appropriate internal controls for such assays has proven difficult. We describe a reaction-specific RNA internal control for diagnostic reverse transcription (RT)-PCR which allows extraction, RT, amplification, and detection to be monitored. The control consists of a G+C-rich (60%) RNA molecule with an extensive secondary structure, based on a modified hepatitis delta virus genome. The rod-like structure of this RNA, with 70% intramolecular base pairing, provides a difficult template for RT-PCR. This ensures that the more favorable target virus amplicon is generated in preference to the control, with the control being detected only if the target virus is absent. The unusual structure of hepatitis delta virus RNA has previously been shown to enhance its stability and resistance to nucleases, an advantage for routine use as an internal control. The control was implemented in three nested multiplex RT-PCRs to detect nine clinically important respiratory viruses: (i) influenza A and B viruses, (ii) respiratory syncytial viruses A and B and human metapneumovirus, and (iii) parainfluenza virus types 1 to 4. The detection limits of these assays were not detectably compromised by the presence of the RNA control. During routine testing of 324 consecutive unselected respiratory samples, the presence of the internal control ensured that genuine and false-negative results were distinguishable, thus increasing the diagnostic confidence in the assay.","{'model': 'tldr@v2.0.0', 'text': 'A reaction-specific RNA internal control for diagnostic reverse transcription (RT)-PCR which allows extraction, RT, amplification, and detection to be monitored, and ensures that genuine and false-negative results were distinguishable, thus increasing the diagnostic confidence in the assay.'}",https://europepmc.org/articles/pmc356891?pdf=render
-A single-center retrospective analysis of interventions provided to geriatric inpatients receiving end-of-life care,T. Tan,"This study is a single-center retrospective analysis of interventions provided to 50 geriatric inpatients (age > 70 years) receiving end-of-life (EOL) care. Data including frequency of medical and non-medical interventions provided to the patient at the EOL and the different symptoms experienced as well as the outcome of pharmacological interventions provided for these symptoms were collected. The analysis highlights the relatively low rates of anticipatory prescribing of non-opioid EOL medications and the under treatment of multiple symptoms experienced at the EOL – pain, breathlessness, and agitation. Presently there is no formalized or uniform process in EOL prescribing on the medical or surgical wards of the hospital. Use of an ‘anticipatory prescribing’ template that can be modified to suit individual cases may be useful, as well as inclusion of a prompt for routine anticipatory EOL prescribing for a patient who is on the EOL pathway. Education of nurses and doctors on prompt symptom recognition and treatment, as well as the importance of addressing spiritual, religious and social needs may be useful in improving the care of elderly inpatients receiving EOL care in the hospital setting.","{'model': 'tldr@v2.0.0', 'text': 'Education of nurses and doctors on prompt symptom recognition and treatment, as well as the importance of addressing spiritual, religious and social needs may be useful in improving the care of elderly inpatients receiving EOL care in the hospital setting.'}",
-Comparing Emission Mitigation Effort,Joseph E. Aldy,"A natural outcome of the emerging pledge and review approach to international climate change policy is interest in comparing mitigation effort among countries. Domestic publics and stakeholders will have an interest in knowing if peer countries are undertaking (or planning to undertake) comparable effort in mitigating their greenhouse gas emissions. Moreover, if considered inadequate to address the risks posed by climate change, the aggregate effort will likely prompt broader interest in identifying those countries where greater effort is arguably warranted on the basis of peer comparisons. Both assessments require metrics of effort and comparisons among countries. We propose a framework for such an exercise, drawing from a set of principles for designing and implementing informative metrics. We present a template for organizing metrics on mitigation effort, for both ex-ante and ex-post review. We also provide preliminary assessments of effort along emissions, price, and cost metrics for post–2020 climate policy contributions by China, the European Union, Russia, and the United States. We close with a discussion of the role of academics and civil society in promoting transparency and facilitating the evaluation and comparison of effort.",,
-Clinical handover of the trauma and medical patient: a structured approach,Jez Loseby,"Handovers are vital to the continuity of patient care and can influence patient outcome. Several structures exist to facilitate handover delivery but further work is needed to prospectively evaluate them. This article advocates the implementation of a suitable pre-hospital handover template and the introduction of handover training. We believe the introduction of trauma networks into the UK provides the ideal opportunity to prompt hospitals and ambulance services to co-operate to standardise the approach to handover, improve training and carry out further prospective research into the most effective methods of patient handover.","{'model': 'tldr@v2.0.0', 'text': 'The introduction of trauma networks into the UK provides the ideal opportunity to prompt hospitals and ambulance services to co-operate to standardise the approach to handover, improve training and carry out further prospective research into the most effective methods of patient handover.'}",
-Early Identification of Children at Risk for Critical Care: Standardizing Communication for Inter–Emergency Department Transfers,Cyril Sahyoun,"Background Interfacility transfers occur frequently and often involve critically ill patients. Clear communication at the time of patient referral is essential for patient safety. Objectives The objective of this work was to study whether a standardized inter-emergency department (ED) transfer communication template for pediatric patients with respiratory complaints identifies patients that require intensive care unit (ICU) admission. Methods We created a template to structure the communication between referring and receiving providers involved in inter-ED transfers of children with respiratory complaints. The template was designed for use by nonphysicians to prompt specific questions that would trigger notification of the ED attending based on signs of critical illness. The template was retrospectively applied to determine whether it would have properly triggered attending physician notification of a child ultimately requiring ICU admission. Results Of 285 transferred children, 61 (21%) were admitted to an ICU from the receiving ED. The sensitivity of the communication template in predicting the need for ICU admission was 84% (95% confidence interval [CI], 72%–92%), negative predictive value of 95% (95% CI, 90%–97%), specificity of 77% (95% CI, 71%–82%), positive predictive value of 50% (95% CI, 40%–60%). Of the 10 patients admitted to an ICU who were not identified by the tool, none were critically ill upon arrival. Of the individual communication elements, the sensitivity and negative predictive value ranged from 3% to 38% and from 79% to 86%, respectively. Conclusions A standardized communication template for inter-ED transfers can identify children with respiratory complaints who require ICU admission. Next steps include real-time application to judge screening performance compared with current nonstandardized intake protocols.","{'model': 'tldr@v2.0.0', 'text': 'Whether a standardized inter-emergency department transfer communication template for pediatric patients with respiratory complaints identifies patients that require intensive care unit (ICU) admission is studied.'}",
-Random word retrieval for automatic story generation,Richard S. Colon,"Over the past forty years, significant research has been done on story/narrative generation in which the computer is the author. Many existing systems generate stories by filling in a template or copying an analogous story (and changing the time, place, etc.) or by prompting the user to provide a start to the story. Very few systems generate variable stories without these techniques. While it is impossible to quantify a human writer's inspiration, we can consider a common exercise that authors perform; namely `writing prompts'. A writing prompt is just a topic or idea around which to start writing. The prompt can simply be a few words, which becomes the basis for a story. In this paper we present story generation from the perspective of how human authors create stories via writing prompts. The system will select a few random words as a prompt, which will form the basic parameters for generating a story. But unlike a human writer, a computer cannot intuitively know the context of a chosen word. Therefore, the Internet (and existing `Concept Knowledge' systems) will be used to find the context for the selected words, thus guiding the story generation process.","{'model': 'tldr@v2.0.0', 'text': ""This paper presents story generation from the perspective of how human authors create stories via writing prompts, using the Internet (and existing `Concept Knowledge' systems) to find the context for the selected words, thus guiding the story generation process.""}",https://scholarworks.bridgeport.edu/xmlui/bitstream/123456789/545/3/FRD_RColon_Story_Gen_Poster_Mar03_2014.pdf
-Direct colony polymerase chain reaction for rapid identification of yeasts isolated from blood specimen,Rajyoganandh S. Vijayaraman,"Context: Bloodstream infections (BSIs) caused by yeasts have an increasing frequency due to the growing population of immunosuppressed individuals. Among yeasts, Candida remains the most prevalent species with the increase in the incidence of non-albicans Candida species. Apart from Candida, other yeasts are also involved in causing BSI. High mortality associated with Candida and other yeast infection can be reduced by prompt and appropriate antifungal therapy. Hence, rapid identification and speciation of yeasts isolated from blood play a significant role in the management of the patients. Since conventional methods used for speciation of Candida and other yeasts are laborious, time-consuming and often unclear, rapid and accurate molecular techniques are required. Materials and Methods: Instead of using purified genomic DNA as template for polymerase chain reaction (PCR), we used yeast colony and cell suspensions in water and 0.10M potassium hydroxide as template for PCR. Candida albicans, Trichosporon and Cryptococcus neoformans were used as reference strains. Further, a total of 100 yeast isolates were also tested. All reactions were performed using the universal fungal primers ITS1 and ITS4; the PCR products were then digested with restriction enzyme (Msp1). Results: Direct colony PCR (DCPCR) produced sharp and distinct bands compared to the cell suspensions with the reference strains. All the 100 clinical isolates tested also produced distinct bands. Conclusion: DCPCR approach not only reduces the DNA template preparation time but is also easy, rapid and reduces the cost of PCR.","{'model': 'tldr@v2.0.0', 'text': 'Direct colony PCR (DCPCR) approach not only reduces the DNA template preparation time but is also easy, rapid and reduces the cost of PCR.'}",
-Orbital and physical parameters of eclipsing binaries from the ASAS catalogue -- III. Two new low-mass systems with rapidly evolving spots,K. Hełminiak,"We present the results of our spectroscopic and photometric analysis of two newly discovered low-mass detached eclipsing binaries found in the All-Sky Automated Survey (ASAS) catalogue: ASAS J093814-0104.4 and ASAS J212954-5620.1. Using the GIRAFFE instrument on the 1.9-m Radcliffe telescope at SAAO and the UCLES spectrograph on the 3.9-m Anglo-Australian Telescope, we obtained high-resolution spectra of both objects and derived their radial velocities (RVs) at various orbital phases. The RVs of both objects were measured with the TODCOR technique using synthetic template spectra as references. We also obtained V and I band photometry using the 1.0-m Elizabeth telescope at SAAO and the 0.4-m PROMPT instruments located at the CTIO. The orbital and physical parameters of the systems were derived with PHOEBE and JKTEBOP codes. We compared our results with several sets of widely-used isochrones. Our multi-epoch photometric observations demonstrate that both objects show significant out-of-eclipse modulations, which vary in time. We believe that this effect is caused by stellar spots, which evolve on time scales of tens of days. For this reason, we constructed our models on the basis of photometric observations spanning short time scales (less than a month). Our modeling indicates that (1) ASAS-09 is a main sequence active system with nearly-twin components with masses of M1 = 0.771(33) Msun, M2 = 0.768(21) Msun and radii of R1 = 0.772(12) Rsun and R2 = 0.769(13) Rsun. (2) ASAS-21 is a main sequence active binary with component masses of M1 = 0.833(17) Msun, M2 = 0.703(13) Msun and radii of R1 = 0.845(12) Rsun and R2 = 0.718(17) Rsun. Both systems confirm the characteristic of active low-mass stars, for which the observed radii are larger and the temperatures lower than predicted by evolutionary models. Other parameters agree within errors with the models of main sequence stars.",,https://www.aanda.org/articles/aa/pdf/2011/03/aa15127-10.pdf
-Group medical appointments: organization and implementation in the bone marrow transplantation clinic.,K. Meehan,"Group medical appointments provide patients with prompt access to care, greater attention to their psychosocial needs, and increased time with their medical team. Care providers evaluate more patients with similar needs in a shorter period. Eligible patients (between 100 days and 3 years after autologous stem cell transplantation) were contacted to participate. The laboratory and radiographic results of each patient's completed reevaluation were entered into the electronic template office note before their visit. A group medical visit model, called a Physical Shared Medical Appointment (PSMA), was employed because this uses individual patient examinations followed by a group meeting. On the day of the visit, brief physical examinations were performed on each patient. A transplantation physician then met with the group of patients and answered questions. Patient satisfaction surveys were distributed upon completion of the group session. Each PSMA was limited to 10 patients to facilitate patient participation. Questions during the group meeting were general and applicable to all patients. At completion of the 2-hour visit, patient surveys indicated an extremely high level of satisfaction and the preference to attend a future PSMA. Issues discussed during the group meeting were pertinent to all transplant recipients, regardless of diagnosis. The PSMA model allows the patient to spend extended time with their care providers while providing the care providers an opportunity to discuss health issues with numerous patients during 1 appointment. The Dartmouth Transplant PSMA model is expanding to pretransplantation and postallogeneic transplant recipients.","{'model': 'tldr@v2.0.0', 'text': 'The Dartmouth Transplant PSMA model allows the patient to spend extended time with their care providers while providing the care providers an opportunity to discuss health issues with numerous patients during 1 appointment.'}",
-Research on Linear Features Change Detection Based on Remote Sensing Image and Vector Data,Zuobang Zhang,"To prompt the present situation and utilized values of fundamental geo-information, this paper focuses on a change detection method based on remote sensing image and GIS vector for linear features. Firstly unilateral vector was taken as original value of linear features; then edge points were picked up by pyramid decomposition and multi-scale template matching, and Ziplock Snake method was adopted to further improve the extraction results; finally buffer zone was constructed to distinguish the changed part. This change detection method proves to have higher degree of automation and more precise, so long as the registration of remote sensing image and vector map is accurate.","{'model': 'tldr@v2.0.0', 'text': 'This change detection method proves to have higher degree of automation and more precise, so long as the registration of remote sensing image and vector map is accurate.'}",
-Single-chip speech recognition system based on 8051 microcontroller core,Yuanyuan Shi,"This paper describes a single-chip speech recognition system. It contains the speech functions of prompt, playback, speakerdependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc. The chip is designed based on the SOC (System on Chip) philosophy and an 8-bit MCU, RAM, ROM, ADCIDAC, PWM, I/O ports and other peripheral circuits are all embedded in it. Software modules including control/communication, speech coding and speech recognition algorithms are implemented in an 805 1 compatible microcontroller core, resulting in the extremely low cost of the chip. The speech recognition adopts the template matching technique. It recognizes up to 20 phrases with an average length of 1 second and the recognition accuracy reaches more than 95% with the background SNR above IOdB. Speech coding uses Continuous Variablc Slope Deltamodulation (CVSD) algorithm. The bit rate is 16kbitsis.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes a single-chip speech recognition system that contains the speech functions of prompt, playback, speakerdependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc.'}",
-Single-chip speech recognition system based on 8051 microcontroller core,Shi Yuanyuan,"This paper describes a single-chip speech recognition system. It contains the speech functions of prompt, playback, speaker-dependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc. The chip is designed based on the SOC (system on chip) philosophy and an 8-bit MCU, RAM, ROM, ADC/DAC, PWM, I/O ports and other peripheral circuits are all embedded in it. Software modules including control/communication, speech coding and speech recognition algorithms are implemented in an 8051 compatible microcontroller core, resulting in the extremely low cost of the chip. The speech recognition adopts the template matching technique. It recognizes up to 20 phrases with an average length of 1 second and the recognition accuracy reaches more than 95% with the background SNR above 10 dB. Speech coding uses continuous variable slope delta modulation (CVSD) algorithm. The bit rate is 16 kbits/s.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes a single-chip speech recognition system that contains the speech functions of prompt, playback, speaker-dependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc.'}",
-Competitive polymerase chain reaction for quantitating feline immunodeficiency virus load in infected cat tissues.,M. Pistello,"To quantitate FIV provirus copy numbers present in tissue of infected cats, we have applied a competitive polymerase chain reaction (cPCR) recently described for HIV. The method consists in coamplifying a fixed amount of the DNA to be examined with graded copy numbers of a DNA competitor incorporating a short deletion and bearing the same primer recognition sequences. These conditions ensure almost identical thermodynamic and amplification efficiency for both template species but permit a prompt recognition of the two amplification products by gel electrophoresis. Since the amounts of the two amplicons are dependent on relative initial template concentrations, the number of FIV genomes in the sample can be calculated by densitometric analysis of the electrophoretic bands. After validation, the method has been applied to study the provirus loads in the tissues of cats infected with the Pisa-M2 isolate of FIV.","{'model': 'tldr@v2.0.0', 'text': 'A competitive polymerase chain reaction recently described for HIV is applied to quantitate FIV provirus copy numbers present in tissue of infected cats to study the provirus loads in the tissues of cats infected with the Pisa-M2 isolate of FIV.'}",
-A Reliable Logo and Replay Detector for Sports Video,Qiao Huang,"Replay is one of the key cues indicating highlights in sports videos. A replay is usually sandwiched by two identical logos which prompt the start and end of a replay. A logo transition usually contains 10-30 frames, describes a flying or varying object(s). In this paper, a reliable logo and replay detecting approach is proposed. It contains two main stages: first, a logo transition template is unsupervised learned, a key frame (K-frame) and a set of pixels that describes logo object (logo pixels, L-pixels) accurately are also extracted; second, the learned information are used jointly to detect logos and replays in the video. In addition to traditional color analysis, optical flow feature is employed to depict the movement of the logo object(s). Extensive experiments show that the proposed approach can reliably detect logos and replays regardless of the types of sports videos.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that the proposed approach can reliably detect logos and replays regardless of the types of sports videos.'}",
-A Rapid Development Framework for Multilingual Spoken Dialogue Systems,Masahiro Araki,"We propose a new framework for rapid development of multilingual spoken dialogue systems based on a data modeling driven method. The development process begins with the data model definition in the Groovy programming language, with an extension for implementing spoken dialogue systems, such as the annotation of the target task type and the initiative type of the realized dialogue. Since the data model definition can be inherited from the existing semantic Web schema, the developer can choose the necessary properties from the schema via the GUI interface. By applying the data model definition, our framework generates all the necessary components for a spoken dialogue system based on the model-view-controller (MVC) model. In addition, template prompt entries, which the developer uses to specify the contents of the system utterance, are automatically generated. The fulfilled contents are translated to various languages using a translation Web service. The multilingual speech interaction is realized by an HTML5 speech API specifying only the ""lang"" attribute of the HTML element. Therefore, to realize a first prototype of a foreign language spoken dialogue system, the developer does not necessarily need knowledge of the target language.","{'model': 'tldr@v2.0.0', 'text': 'A new framework for rapid development of multilingual spoken dialogue systems based on a data modeling driven method that generates all the necessary components for a spoken dialogue system based on the model-view-controller (MVC) model.'}",
-Changes in membrane transport function in G0 and G1 cells,M. Costlow,"Confluent quiescent monolayers of aneuploid and euploid cells in culture can be stimulated to proliferate by appropriate nutritional changes. In confluent monolayers of WI‐38 human diploid fibroblasts the uptake of cycloleucine is increased three hours after these cells are stimulated to proliferate by a change of medium plus 10% serum. No changes in the uptake of cycloleucine are observed in logarithmically‐growing WI‐38 cells exposed to fresh medium plus 10% serum, or in WI‐38 confluent monolayers in which the conditioned medium has been replaced by fresh medium with 0.3% serum (a change that does not cause stimulation of cellular proliferation in WI‐38 cells). In 3T6 cells in the stationary phase stimulated to proliferate by nutritional changes, there is a prompt increase in the uptake of cycloleucine, within one hour after stimulation of cell proliferation. Similar results were obtained with stationary 2RA cells which are SV‐40 transformed WI‐38 fibroblasts. In addition, chromatin template activity which is known to increase in the early stages after stimulation of confluent WI‐38 cells, was unchanged in confluent 3T6 or 2RA cells stimulated to proliferate.","{'model': 'tldr@v2.0.0', 'text': 'In confluent monolayers of WI‐38 human diploid fibroblasts the uptake of cycloleucine is increased three hours after these cells are stimulated to proliferate by a change of medium plus 10% serum, while chromatin template activity was unchanged in confluent 3T6 or 2RA cells stimulated to proliferation.'}",
-Isolation and characterization of conditional alleles of bacteriophage T4 genes uvsX and uvsY.,M. A. Conkling,"The bacteriophage T4 uvsW, uvsX and uvsY gene functions are required for wild-type levels of recombination and for normal survival and mutagenesis after treatments with ultraviolet (UV) and ionizing radiations. The ability of uvsX and uvsY mutations to suppress the lethality of gene 49 mutations was used to select temperature-sensitive and amber alleles of these two genes. (uvsW mutations do not suppress gene 49 mutations.) A simple and powerful complementation test was developed to assist in assigning uvs mutations to genes. The amber alleles of uvsX and uvsY behave as simple null alleles, fully suppressing a gene 49 defect, enhancing UV killing and abolishing UV mutagenesis. However, the properties of the ts alleles of uvsX and uvsY demonstrated that suppression of a gene 49 defect, sensitivity to UV-induced inactivation and UV mutability can be partially uncoupled. These results prompt the hypothesis that radiation mutagenesis occurs during DNA chain elongation past template damage within a recombinational intermediate rather than within a conventional replication fork.","{'model': 'tldr@v2.0.0', 'text': 'The properties of the ts alleles of uvsX and uvsY demonstrated that suppression of a gene 49 defect, sensitivity to UV-induced inactivation and UV mutability can be partially uncoupled, prompting the hypothesis that radiation mutagenesis occurs during DNA chain elongation past template damage within a recombinational intermediate rather than within a conventional replication fork.'}",https://academic.oup.com/genetics/article-pdf/107/4/505/34446702/genetics0505.pdf
-Recruiting patients and collecting data for an observational study using computerised record pop-up prompts: the PROG-RES study,R. Hayward,"Background and Aim Engagement of general practitioners (GPs) and recruitment of patients are ever present problems in primary care studies. This paper seeks to demonstrate that electronic prompts represent one method of easing the burden on GPs to recruit individual patients to studies and also provide the opportunity to collect research data during a normal consultation. Methods Older adults consulting for non-inflammatory musculoskeletal pain from five general practices in Cheshire were recruited to a prospective cohort study (the PROG-RES study). Recruitment of patients was aided by a computer prompt during relevant consultations. When triggered by an appropriate Read code, a pop-up template appeared on the consultation screen prompting the GPs to record the answers to seven brief questions. A self-complete questionnaire was mailed to patients who had completed templates by the Keele GP Research Network team and permission was sought to access their medical records. A feasibility study suggested that the potential number of activated templates in the practice within four months would be 636. Results The 44 GPs completed 650 electronic templates during the four-month recruitment period. Almost 40% of recruitment was within four weeks and greater than 95% of recruitment was within 16 weeks. Practices A–D completed electronic templates at a similar rate (1.61–1.86 templates per 1000 patients), although practice E completed templates at a lower frequency (0.76) due to internal difficulties. Completion of individual items ranged from 98% to 83% and completion of all seven questions was recorded in 63% of patients; 4% of patients had three or fewer responses recorded. Conclusion Templates activated by appropriate codes in the GP consultation can facilitate recruitment to observational studies in primary care. It is possible to collect high-quality research data within a normal consultation. This may be a model for use in future studies in primary care.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that electronic prompts activated by appropriate codes in the GP consultation can facilitate recruitment to observational studies in primary care and it is possible to collect high-quality research data within a normal consultation.'}",https://www.cambridge.org/core/services/aop-cambridge-core/content/view/E93B7CE703F600A1423292CF4FD74CAA/S1463423612000047a.pdf/div-class-title-recruiting-patients-and-collecting-data-for-an-observational-study-using-computerised-record-pop-up-prompts-the-prog-res-study-div.pdf
-A web‐based performance assessment system for environmental protection: WePass,S. Cheung,"In response to the increasing public concern over the effects of construction waste on the environment, a performance assessment system has been designed to provide an instant, online assessment of how well a construction site performs environmentally, e.g. measuring the level of pollution generated from a construction site. Through the use of both Internet and database technologies, it is possible to collect and present data that are essential for environmental management online. To achieve this, a set of environmental performance parameters was derived through a postal questionnaire survey and interviews with experts in the field. A conceptual framework was developed, of which the key components were: template, knowledge base, output data and benchmark group. In combination, these factors provide a system that enables speedy assessment and effective management of environmental performance on construction sites, hence prompt decisions can be made and corrective actions can be taken before potential hazards turn into real disasters.",,
-Xerte - A User-Friendly Tool for Creating Accessible Learning Objects,S. Ball,,"{'model': 'tldr@v2.0.0', 'text': 'The JISC TechDis template for Xerte brings some of the key aspects of Xerte to a non-technical user interface to enable any teacher or tutor to easily create Accessible Learning Objects.'}",
-Radiation Interrogation Using Signature Analysis for Detection of Chemical Explosives,W. L. Dunn,"A signature-based radiation technique for detecting chemical explosive is described. Radiation techniques offer the advantage that they can operate at standoff, although rapid detection of explosives at safe standoff distances remains problematic. The technique we describe differs from the major nuclear approaches, which involve imaging (using either photon or neutron interrogation) and/or quantitative analysis (for instance by use of prompt-and inelastic-scatter gamma-ray production based on neutron interrogation). Our technique utilizes both photon and neutron interrogation but avoids imaging or quantitative analysis. The technique, which we call signature-based radiation scanning (SBRS), limits itself to detecting whether a target contains an explosive and does not attempt to characterize the internals of a target further. A template-matching technique is employed, which provides a single figure-of-merit whose value is used to distinguish between safe targets and those containing explosives. Both simulation and experiment have been used to verify the validity of SBRS.",,
-"""Automics"": assisted mobile authoring of photostories as souvenirs of a day in the park",Duncan Rowland,"This paper describes a prototype digital economy service that generates ‘photostory’ souvenirs for group day-trips to an amusement park. Location-based triggers prompt visitors to use their mobile phones to capture, share and annotate photos during the day. Selected photos are combined with professional photos from on-ride capture systems and a template is used to generate a bespoke photostory - an “Automic” souvenir. The design of this prototype system was informed by an ethnographic study of park visiting, and this is discussed along with directions for future work.",,
-Identificación de especies de Leishmania por la técnica de amplificación al azar del ADN polimórfico,Lianet Monzote Fidalgo,"INTRODUCTION: leishmaniosis has been regarded by the World Health Organization as one of the most important tropical diseases. It is very difficult to control such parasitosis because there are not vaccines, and therapy is generally toxic and unsatisfactory. It is of vital importance to set prompt diagnostic method along with identification of the parasite in order to select the suitable treatment and to design the most convenient control measures. Recently, the advances in molecular biology have made it possible to characterize Leishmania species by different methods. The random amplified polymorphic DNA technique is a simple method to detect the genetic polymorphic DNA. OBJECTIVE: to standardize the random amplified polymorphic DNA technique for its use in New World Leishmania species typing. METHODS: by using 5 pmol primer concentration, 75 ng of template DNA, 2 mM of magnesium chloride and 2 U of polymerase DNA Taq in 25μL reaction, two reproducible amplification patters were obtained. The optimized random amplified polymorphic DNA technique served to determine the genetic differences among ten reference strains of Leishmania, with 6 sets of randomly designed conventional primers. The UP GMA method-based grouping strategy determined the phylogenetic relation among the studied species. RESULTS: OPA primers -3, 4 and 8 allowed distinguishing the ten reference strains of Leishmania under study. Two well defined genetic groups including species of Leishmania and Viannia subgenres were obtained; these 2 subgenres showed genetic differences. CONCLUSIONS: in this way, our laboratory has the optimized random amplified polymorphic DNA for the identification of Leishmania species.","{'model': 'tldr@v2.0.0', 'text': 'The laboratory has the optimized random amplified polymorphic DNA for the identification of Leishmania species and determined the phylogenetic relation among the studied species.'}",
-Effect of Drug Information Request Templates on Pharmacy Student Compliance with the Modified Systematic Approach to Answering Drug Information Questions,Stacey M. Lavsa,"Background: The modified systematic approach to answering drug information questions is a technique used in drug information practice and in teaching pharmacy students to effectively provide drug information. Drug information request templates were developed to prompt students and other trainees to ask appropriate background questions and perform an effective search. Objective: An evaluation was conducted to determine whether use of drug information templates by fourth-year pharmacy students during their drug information experiential rotation improved compliance with the modified systematic approach. Methods: Fifty documented drug information requests, including 25 prior to template implementation (August 2005–August 2006) and 25 after template implementation (August 2007–August 2008), were randomly selected for evaluation. Each question was evaluated for completeness of background information obtained, categorization and identification of the ultimate question, completeness of references searched, and formulation of a concise response and an evidence-based recommendation. Results: Background information was complete in 16% of pre-template questions and 92% of post-template questions (p < 0.001). Eighty-four percent of pre-template questions and 96% of post-template questions were appropriately categorized (p = 0.349). The requestor's ultimate question was clearly identified in 68% of pre-template questions and 92% of post-template questions (p = 0.074). All necessary references were searched in 36% of pre-template questions and 88% of post-template questions (p < 0.001). A concise response was documented in 80% of pre-template questions and 92% of post-template questions (p = 0.417). In questions determined to require a specific recommendation among the pre-template (n = 20) and post-template groups (n = 14), a clear and evidence-based recommendation was described in 40% (p = 0.038) and 79% (p = 0.038), respectively. Conclusions: Use of drug information request templates improves students' compliance with the modified systematic approach, most notably in obtaining background information and searching necessary references including primary literature.","{'model': 'tldr@v2.0.0', 'text': ""Use of drug information request templates improves students' compliance with the modified systematic approach, most notably in obtaining background information and searching necessary references including primary literature.""}",
-New Event-Processing Design Patterns Using CEP,A. Alves,,"{'model': 'tldr@v2.0.0', 'text': 'A set of new patterns for event processing are documents, describing their problem domain and providing a solution template implemented using CEP, which is both succinct and highly re-usable.'}",http://adcalves.files.wordpress.com/2009/09/new-event-processing-design-patterns-using-cep.pdf
-A Novel Sports Video Logo Detector Based on Motion Analysis,Hongliang Bai,,"{'model': 'tldr@v2.0.0', 'text': 'A novel automatic logo detection approach is proposed that is automatically learned by dynamic programming and unsupervised clustering, a key frame is also extracted and used jointly to detect logos in sports videos.'}",
diff --git a/data/semantic_scholar_data_doubled_cleaned.csv b/data/semantic_scholar_data_doubled_cleaned.csv
deleted file mode 100644
index fff7549..0000000
--- a/data/semantic_scholar_data_doubled_cleaned.csv
+++ /dev/null
@@ -1,3937 +0,0 @@
-Title,First Author,Abstract,TLDR,Open Access PDF URL
-"""do anything now"": characterizing and evaluating in-the-wild jailbreak prompts on large language models",Xinyue Shen,"The misuse of large language models (LLMs) has garnered significant attention from the general public and LLM vendors. In response, efforts have been made to align LLMs with human values and intent use. However, a particular type of adversarial prompts, known as jailbreak prompt, has emerged and continuously evolved to bypass the safeguards and elicit harmful content from LLMs. In this paper, we conduct the first measurement study on jailbreak prompts in the wild, with 6,387 prompts collected from four platforms over six months. Leveraging natural language processing technologies and graph-based community detection methods, we discover unique characteristics of jailbreak prompts and their major attack strategies, such as prompt injection and privilege escalation. We also observe that jailbreak prompts increasingly shift from public platforms to private ones, posing new challenges for LLM vendors in proactive detection. To assess the potential harm caused by jailbreak prompts, we create a question set comprising 46,800 samples across 13 forbidden scenarios. Our experiments show that current LLMs and safeguards cannot adequately defend jailbreak prompts in all scenarios. Particularly, we identify two highly effective jailbreak prompts which achieve 0.99 attack success rates on ChatGPT (GPT-3.5) and GPT-4, and they have persisted online for over 100 days. Our work sheds light on the severe and evolving threat landscape of jailbreak prompts. We hope our study can facilitate the research community and LLM vendors in promoting safer and regulated LLMs.","{'model': 'tldr@v2.0.0', 'text': 'The first measurement study on jailbreak prompts in the wild is conducted, with 6,387 prompts collected from four platforms over six months, and it is shown that current LLMs and safeguards cannot adequately defend jailbreak Prompts in all scenarios.'}",https://arxiv.org/pdf/2308.03825
-latent jailbreak: a benchmark for evaluating text safety and output robustness of large language models,Huachuan Qiu,"Considerable research efforts have been devoted to ensuring that large language models (LLMs) align with human values and generate safe text. However, an excessive focus on sensitivity to certain topics can compromise the model's robustness in following instructions, thereby impacting its overall performance in completing tasks. Previous benchmarks for jailbreaking LLMs have primarily focused on evaluating the safety of the models without considering their robustness. In this paper, we propose a benchmark that assesses both the safety and robustness of LLMs, emphasizing the need for a balanced approach. To comprehensively study text safety and output robustness, we introduce a latent jailbreak prompt dataset, each involving malicious instruction embedding. Specifically, we instruct the model to complete a regular task, such as translation, with the text to be translated containing malicious instructions. To further analyze safety and robustness, we design a hierarchical annotation framework. We present a systematic analysis of the safety and robustness of LLMs regarding the position of explicit normal instructions, word replacements (verbs in explicit normal instructions, target groups in malicious instructions, cue words for explicit normal instructions), and instruction replacements (different explicit normal instructions). Our results demonstrate that current LLMs not only prioritize certain instruction verbs but also exhibit varying jailbreak rates for different instruction verbs in explicit normal instructions. Code and data are available at https://github.com/qiuhuachuan/latent-jailbreak.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a latent jailbreak prompt dataset, and presents a systematic analysis of the safety and robustness of LLMs regarding the position of explicit normal instructions, word replacements, and instruction replacements.'}",https://arxiv.org/pdf/2307.08487
-jailbreak in pieces: compositional adversarial attacks on multi-modal language models,Erfan Shayegani,"We introduce new jailbreak attacks on vision language models (VLMs), which use aligned LLMs and are resilient to text-only jailbreak attacks. Specifically, we develop cross-modality attacks on alignment where we pair adversarial images going through the vision encoder with textual prompts to break the alignment of the language model. Our attacks employ a novel compositional strategy that combines an image, adversarially targeted towards toxic embeddings, with generic prompts to accomplish the jailbreak. Thus, the LLM draws the context to answer the generic prompt from the adversarial image. The generation of benign-appearing adversarial images leverages a novel embedding-space-based methodology, operating with no access to the LLM model. Instead, the attacks require access only to the vision encoder and utilize one of our four embedding space targeting strategies. By not requiring access to the LLM, the attacks lower the entry barrier for attackers, particularly when vision encoders such as CLIP are embedded in closed-source LLMs. The attacks achieve a high success rate across different VLMs, highlighting the risk of cross-modality alignment vulnerabilities, and the need for new alignment approaches for multi-modal models.","{'model': 'tldr@v2.0.0', 'text': 'Cross-modality attacks on alignment where adversarial images going through the vision encoder with textual prompts to break the alignment of the language model are developed.'}",
-fuzzllm: a novel and universal fuzzing framework for proactively discovering jailbreak vulnerabilities in large language models,Dongyu Yao,"Jailbreak vulnerabilities in Large Language Models (LLMs), which exploit meticulously crafted prompts to elicit content that violates service guidelines, have captured the attention of research communities. While model owners can defend against individual jailbreak prompts through safety training strategies, this relatively passive approach struggles to handle the broader category of similar jailbreaks. To tackle this issue, we introduce FuzzLLM, an automated fuzzing framework designed to proactively test and discover jailbreak vulnerabilities in LLMs. We utilize templates to capture the structural integrity of a prompt and isolate key features of a jailbreak class as constraints. By integrating different base classes into powerful combo attacks and varying the elements of constraints and prohibited questions, FuzzLLM enables efficient testing with reduced manual effort. Extensive experiments demonstrate FuzzLLM's effectiveness and comprehensiveness in vulnerability discovery across various LLMs.","{'model': 'tldr@v2.0.0', 'text': 'FuzzLLM is introduced, an automated fuzzing framework designed to proactively test and discover jailbreak vulnerabilities in LLMs and utilizes templates to capture the structural integrity of a prompt and isolate key features of a jailbreak class as constraints.'}",https://arxiv.org/pdf/2309.05274
-latent jailbreak: a test suite for evaluating both text safety and output robustness of large language models,Huachuan Qiu,"Warning: This paper contains examples of potentially offensive and harmful text. Considerable research efforts have been devoted to ensuring that large language models (LLMs) align with human values and generate safe text. However, an excessive focus on sensitivity to certain topics can compromise the model’s robustness in following instructions, thereby impacting its over-all performance in completing tasks. Previous benchmarks for jailbreaking LLMs have primarily focused on evaluating the safety of the models without considering their robustness. In this paper, we propose a benchmark that assesses both the safety and robustness of LLMs, emphasizing the need for a balanced approach. To comprehensively study text safety and output robustness, we introduce a latent jailbreak prompt dataset, each involving malicious instruction embedding. Specifically, we instruct the model to complete a regular task, such as translation, with the text to be translated containing malicious instructions. To further analyze safety and robustness, we design a hierarchical annotation framework. We present a systematic analysis of the safety and robustness of LLMs regarding the position of explicit normal instructions, word replacements (verbs in explicit normal instructions, target groups in malicious instructions, cue words for explicit normal instructions), and instruction replacements (different explicit normal instructions). Our results demonstrate that current LLMs not only prioritize certain instruction verbs but also exhibit varying jailbreak rates for different instruction verbs in explicit normal instructions. Code and data are available at https://github.com/ qiuhuachuan/latent-jailbreak.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a latent jailbreak prompt dataset, each involving malicious instruction embedding, and presents a systematic analysis of the safety and robustness of LLMs regarding the position of explicit normal instructions, word replacements, and instruction replacements.'}",
-jailbroken: how does llm safety training fail?,Alexander Wei,"Large language models trained for safety and harmlessness remain susceptible to adversarial misuse, as evidenced by the prevalence of""jailbreak""attacks on early releases of ChatGPT that elicit undesired behavior. Going beyond recognition of the issue, we investigate why such attacks succeed and how they can be created. We hypothesize two failure modes of safety training: competing objectives and mismatched generalization. Competing objectives arise when a model's capabilities and safety goals conflict, while mismatched generalization occurs when safety training fails to generalize to a domain for which capabilities exist. We use these failure modes to guide jailbreak design and then evaluate state-of-the-art models, including OpenAI's GPT-4 and Anthropic's Claude v1.3, against both existing and newly designed attacks. We find that vulnerabilities persist despite the extensive red-teaming and safety-training efforts behind these models. Notably, new attacks utilizing our failure modes succeed on every prompt in a collection of unsafe requests from the models' red-teaming evaluation sets and outperform existing ad hoc jailbreaks. Our analysis emphasizes the need for safety-capability parity -- that safety mechanisms should be as sophisticated as the underlying model -- and argues against the idea that scaling alone can resolve these safety failure modes.","{'model': 'tldr@v2.0.0', 'text': 'The analysis emphasizes the need for safety-capability parity -- that safety mechanisms should be as sophisticated as the underlying model -- and argues against the idea that scaling alone can resolve these safety failure modes.'}",https://arxiv.org/pdf/2307.02483
-"tricking llms into disobedience: understanding, analyzing, and preventing jailbreaks",Abhinav Rao,"Recent explorations with commercial Large Language Models (LLMs) have shown that non-expert users can jailbreak LLMs by simply manipulating the prompts; resulting in degenerate output behavior, privacy and security breaches, offensive outputs, and violations of content regulator policies. Limited formal studies have been carried out to formalize and analyze these attacks and their mitigations. We bridge this gap by proposing a formalism and a taxonomy of known (and possible) jailbreaks. We perform a survey of existing jailbreak methods and their effectiveness on open-source and commercial LLMs (such as GPT 3.5, OPT, BLOOM, and FLAN-T5-xxl). We further propose a limited set of prompt guards and discuss their effectiveness against known attack types.","{'model': 'tldr@v2.0.0', 'text': 'This work performs a survey of existing jailbreak methods and their effectiveness on open-source and commercial LLMs, and proposes a limited set of prompt guards and discusses their effectiveness against known attack types.'}",http://arxiv.org/pdf/2305.14965
-jailbreaking black box large language models in twenty queries,Patrick Chao,"There is growing interest in ensuring that large language models (LLMs) align with human values. However, the alignment of such models is vulnerable to adversarial jailbreaks, which coax LLMs into overriding their safety guardrails. The identification of these vulnerabilities is therefore instrumental in understanding inherent weaknesses and preventing future misuse. To this end, we propose Prompt Automatic Iterative Refinement (PAIR), an algorithm that generates semantic jailbreaks with only black-box access to an LLM. PAIR -- which is inspired by social engineering attacks -- uses an attacker LLM to automatically generate jailbreaks for a separate targeted LLM without human intervention. In this way, the attacker LLM iteratively queries the target LLM to update and refine a candidate jailbreak. Empirically, PAIR often requires fewer than twenty queries to produce a jailbreak, which is orders of magnitude more efficient than existing algorithms. PAIR also achieves competitive jailbreaking success rates and transferability on open and closed-source LLMs, including GPT-3.5/4, Vicuna, and PaLM-2.","{'model': 'tldr@v2.0.0', 'text': 'PAIR is an algorithm that generates semantic jailbreaks with only black-box access to an LLM with competitive jailbreaking success rates and transferability on open and closed-source LLMs, including GPT-3.5/4, Vicuna, and PaLM.'}",https://arxiv.org/pdf/2310.08419
-language model unalignment: parametric red-teaming to expose hidden harms and biases,Rishabh Bhardwaj,"Red-teaming has been a widely adopted way to evaluate the harmfulness of Large Language Models (LLMs). It aims to jailbreak a model's safety behavior to make it act as a helpful agent disregarding the harmfulness of the query. Existing methods are primarily based on input text-based red-teaming such as adversarial prompts, low-resource prompts, or contextualized prompts to condition the model in a way to bypass its safe behavior. Bypassing the guardrails uncovers hidden harmful information and biases in the model that are left untreated or newly introduced by its safety training. However, prompt-based attacks fail to provide such a diagnosis owing to their low attack success rate, and applicability to specific models. In this paper, we present a new perspective on LLM safety research i.e., parametric red-teaming through Unalignment. It simply (instruction) tunes the model parameters to break model guardrails that are not deeply rooted in the model's behavior. Unalignment using as few as 100 examples can significantly bypass commonly referred to as CHATGPT, to the point where it responds with an 88% success rate to harmful queries on two safety benchmark datasets. On open-source models such as VICUNA-7B and LLAMA-2-CHAT 7B AND 13B, it shows an attack success rate of more than 91%. On bias evaluations, Unalignment exposes inherent biases in safety-aligned models such as CHATGPT and LLAMA- 2-CHAT where the model's responses are strongly biased and opinionated 64% of the time.","{'model': 'tldr@v2.0.0', 'text': ""A new perspective on LLM safety research is presented i.e., parametric red-teaming through Unalignment, which tunes the model parameters to break model guardrails that are not deeply rooted in the model's behavior.""}",
-autodan: automatic and interpretable adversarial attacks on large language models,Sicheng Zhu,"Safety alignment of Large Language Models (LLMs) can be compromised with manual jailbreak attacks and (automatic) adversarial attacks. Recent work suggests that patching LLMs against these attacks is possible: manual jailbreak attacks are human-readable but often limited and public, making them easy to block; adversarial attacks generate gibberish prompts that can be detected using perplexity-based filters. In this paper, we show that these solutions may be too optimistic. We propose an interpretable adversarial attack, \texttt{AutoDAN}, that combines the strengths of both types of attacks. It automatically generates attack prompts that bypass perplexity-based filters while maintaining a high attack success rate like manual jailbreak attacks. These prompts are interpretable and diverse, exhibiting strategies commonly used in manual jailbreak attacks, and transfer better than their non-readable counterparts when using limited training data or a single proxy model. We also customize \texttt{AutoDAN}'s objective to leak system prompts, another jailbreak application not addressed in the adversarial attack literature. Our work provides a new way to red-team LLMs and to understand the mechanism of jailbreak attacks.","{'model': 'tldr@v2.0.0', 'text': 'An interpretable adversarial attack is proposed, \\texttt{AutoDAN}, that combines the strengths of both types of attacks and automatically generates attack prompts that bypass perplexity-based filters while maintaining a high attack success rate like manual jailbreak attacks.'}",
-latin america’s prisons pose major covid-19 risks,," Subject COVID-19 and prisons. Significance Authorities in Leticia, Colombia, reported on May 12 that half of inmates at the local prison had tested positive for COVID-19. The news follows a major outbreak at a prison in Villavicencio last month that prompted an attempted jailbreak, and riots at 13 Colombian facilities on March 21, which resulted in 23 deaths at one prison. Concerns are growing about health and security at jails across Latin America, where tensions are building over overcrowding and unsanitary conditions. Impacts The confirmation of COVID-19 at Haiti’s largest jail last week will exacerbate pressure on the region’s most overcrowded prison system. Prisoner releases will become increasingly complicated as infection spreads within jails. Increased poverty due to the COVID-19 crisis could prompt a rise in crime, and thus prison populations. ",,
-visual prompt tuning,Menglin Jia,"The current modus operandi in adapting pre-trained models involves updating all the backbone parameters, ie, full fine-tuning. This paper introduces Visual Prompt Tuning (VPT) as an efficient and effective alternative to full fine-tuning for large-scale Transformer models in vision. Taking inspiration from recent advances in efficiently tuning large language models, VPT introduces only a small amount (less than 1% of model parameters) of trainable parameters in the input space while keeping the model backbone frozen. Via extensive experiments on a wide variety of downstream recognition tasks, we show that VPT achieves significant performance gains compared to other parameter efficient tuning protocols. Most importantly, VPT even outperforms full fine-tuning in many cases across model capacities and training data scales, while reducing per-task storage cost.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Visual Prompt Tuning (VPT) as an efficient and effective alternative to full fine-tuning for large-scale Transformer models in vision and shows that VPT achieves significant performance gains compared to other parameter efficient tuning protocols.'}",http://arxiv.org/pdf/2203.12119
-conditional prompt learning for vision-language models,Kaiyang Zhou,"With the rise of powerful pre-trained vision-language models like CLIP, it becomes essential to investigate ways to adapt these models to downstream datasets. A recently proposed method named Context Optimization (CoOp) introduces the concept of prompt learning—a recent trend in NLP—to the vision domain for adapting pre-trained vision-language models. Specifically, CoOp turns context words in a prompt into a set of learnable vectors and, with only a few labeled images for learning, can achieve huge improvements over intensively-tuned manual prompts. In our study we identify a critical problem of CoOp: the learned context is not generalizable to wider unseen classes within the same dataset, suggesting that CoOp overfits base classes observed during training. To address the problem, we propose Conditional Context Optimization (CoCoOp), which extends CoOp by further learning a lightweight neural network to generate for each image an input-conditional token (vector). Compared to CoOp's static prompts, our dynamic prompts adapt to each instance and are thus less sensitive to class shift. Extensive experiments show that CoCoOp generalizes much better than CoOp to unseen classes, even showing promising transferability beyond a single dataset; and yields stronger domain generalization performance as well. Code is available at https://github.com/KaiyangZhou/CoOp.","{'model': 'tldr@v2.0.0', 'text': 'Conditional Context Optimization (CoCoOp), which extends CoOp by further learning a lightweight neural network to generate for each image an input-conditional token (vector), and yields stronger domain generalization performance as well.'}",https://arxiv.org/pdf/2203.05557
-prompt-to-prompt image editing with cross attention control,Amir Hertz,"Recent large-scale text-driven synthesis models have attracted much attention thanks to their remarkable capabilities of generating highly diverse images that follow given text prompts. Such text-based synthesis methods are particularly appealing to humans who are used to verbally describe their intent. Therefore, it is only natural to extend the text-driven image synthesis to text-driven image editing. Editing is challenging for these generative models, since an innate property of an editing technique is to preserve most of the original image, while in the text-based models, even a small modification of the text prompt often leads to a completely different outcome. State-of-the-art methods mitigate this by requiring the users to provide a spatial mask to localize the edit, hence, ignoring the original structure and content within the masked region. In this paper, we pursue an intuitive prompt-to-prompt editing framework, where the edits are controlled by text only. To this end, we analyze a text-conditioned model in depth and observe that the cross-attention layers are the key to controlling the relation between the spatial layout of the image to each word in the prompt. With this observation, we present several applications which monitor the image synthesis by editing the textual prompt only. This includes localized editing by replacing a word, global editing by adding a specification, and even delicately controlling the extent to which a word is reflected in the image. We present our results over diverse images and prompts, demonstrating high-quality synthesis and fidelity to the edited prompts.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes a text-conditioned model in depth and observes that the cross-attention layers are the key to controlling the relation between the spatial layout of the image to each word in the prompt, and presents several applications which monitor the image synthesis by editing the textual prompt only.'}",http://arxiv.org/pdf/2208.01626
-p-tuning: prompt tuning can be comparable to fine-tuning across scales and tasks,Xiao Liu,"Prompt tuning, which only tunes continuous prompts with a frozen language model, substantially reduces per-task storage and memory usage at training. However, in the context of NLU, prior work reveals that prompt tuning does not perform well for normal-sized pretrained models. We also find that existing methods of prompt tuning cannot handle hard sequence labeling tasks, indicating a lack of universality. We present a novel empirical finding that properly optimized prompt tuning can be universally effective across a wide range of model scales and NLU tasks. It matches the performance of finetuning while having only 0.1%-3% tuned parameters. Our method P-Tuning v2 is an implementation of Deep Prompt Tuning (CITATION) optimized and adapted for NLU. Given the universality and simplicity of P-Tuning v2, we believe it can serve as an alternative to finetuning and a strong baseline for future research.","{'model': 'tldr@v2.0.0', 'text': 'The method P-Tuning v2 is an implementation of Deep Prompt Tuning (CITATION) optimized and adapted for NLU and can serve as an alternative to finetuning and a strong baseline for future research.'}",https://aclanthology.org/2022.acl-short.8.pdf
-"pre-train, prompt, and predict: a systematic survey of prompting methods in natural language processing",Pengfei Liu,"This article surveys and organizes research works in a new paradigm in natural language processing, which we dub “prompt-based learning.” Unlike traditional supervised learning, which trains a model to take in an input x and predict an output y as P(y|x), prompt-based learning is based on language models that model the probability of text directly. To use these models to perform prediction tasks, the original input x is modified using a template into a textual string prompt x′ that has some unfilled slots, and then the language model is used to probabilistically fill the unfilled information to obtain a final string x̂, from which the final output y can be derived. This framework is powerful and attractive for a number of reasons: It allows the language model to be pre-trained on massive amounts of raw text, and by defining a new prompting function the model is able to perform few-shot or even zero-shot learning, adapting to new scenarios with few or no labeled data. In this article, we introduce the basics of this promising paradigm, describe a unified set of mathematical notations that can cover a wide variety of existing work, and organize existing work along several dimensions, e.g., the choice of pre-trained language models, prompts, and tuning strategies. To make the field more accessible to interested beginners, we not only make a systematic review of existing works and a highly structured typology of prompt-based concepts but also release other resources, e.g., a website NLPedia–Pretrain including constantly updated survey and paperlist.","{'model': 'tldr@v2.0.0', 'text': 'The basics of this promising paradigm in natural language processing are introduced, a unified set of mathematical notations that can cover a wide variety of existing work are described, and existing work is organized along several dimensions.'}",https://dl.acm.org/doi/pdf/10.1145/3560815
-learning to prompt for open-vocabulary object detection with vision-language model,Yu Du,"Recently, vision-language pre-training shows great potential in open-vocabulary object detection, where detectors trained on base classes are devised for detecting new classes. The class text embedding is firstly generated by feeding prompts to the text encoder of a pre-trained vision-language model. It is then used as the region classifier to supervise the training of a detector. The key element that leads to the success of this model is the proper prompt, which requires careful words tuning and ingenious design. To avoid laborious prompt engineering, there are some prompt representation learning methods being proposed for the image classification task, which however can only be sub-optimal solutions when applied to the detection task. In this paper, we introduce a novel method, detection prompt (DetPro), to learn continuous prompt representations for open-vocabulary object detection based on the pre-trained vision-language model. Different from the previous classification-oriented methods, DetPro has two highlights: 1) a background interpretation scheme to include the proposals in image background into the prompt training; 2) a context grading scheme to separate proposals in image foreground for tailored prompt training. We assemble DetPro with ViLD, a recent state-of-the-art openworld object detector, and conduct experiments on the LVIS as well as transfer learning on the Pascal VOC, COCO, Objects365 datasets. Experimental results show that our DetPro outperforms the baseline ViLD [7] in all settings, e.g., +3.4 APbox and +3.0 APmask improvements on the novel classes of LVIS. Code and models are available at https://github.com/dyabel/detpro.","{'model': 'tldr@v2.0.0', 'text': 'A novel method, detection prompt (DetPro), to learn continuous prompt representations for open-vocabulary object detection based on the pre-trained vision-language model, which outperforms the baseline ViLD in all settings.'}",https://arxiv.org/pdf/2203.14940
-"recommendation as language processing (rlp): a unified pretrain, personalized prompt & predict paradigm (p5)",Shijie Geng,"For a long time, different recommendation tasks require designing task-specific architectures and training objectives. As a result, it is hard to transfer the knowledge and representations from one task to another, thus restricting the generalization ability of existing recommendation approaches. To deal with such issues, considering that language can describe almost anything and language grounding is a powerful medium to represent various problems or tasks, we present a flexible and unified text-to-text paradigm called “Pretrain, Personalized Prompt, and Predict Paradigm” (P5) for recommendation, which unifies various recommendation tasks in a shared framework. In P5, all data such as user-item interactions, user descriptions, item metadata, and user reviews are converted to a common format — natural language sequences. The rich information from natural language assists P5 to capture deeper semantics for personalization and recommendation. Specifically, P5 learns different tasks with the same language modeling objective during pretraining. Thus, it serves as the foundation model for various downstream recommendation tasks, allows easy integration with other modalities, and enables instruction-based recommendation. P5 advances recommender systems from shallow model to deep model to big model, and will revolutionize the technical form of recommender systems towards universal recommendation engine. With adaptive personalized prompt for different users, P5 is able to make predictions in a zero-shot or few-shot manner and largely reduces the necessity for extensive fine-tuning. On several benchmarks, we conduct experiments to show the effectiveness of P5. To help advance future research on Recommendation as Language Processing (RLP), Personalized Foundation Models (PFM), and Universal Recommendation Engine (URE), we release the source code, dataset, prompts, and pretrained P5 model at https://github.com/jeykigung/P5.","{'model': 'tldr@v2.0.0', 'text': 'A flexible and unified text-to-text paradigm called “Pretrain, Personalized Prompt, and Predict Paradigm” (P5) for recommendation, which unifies various recommendation tasks in a shared framework and will revolutionize the technical form of recommender systems towards universal recommendation engine.'}",
-an information-theoretic approach to prompt engineering without ground truth labels,Lisa P. Argyle,"Pre-trained language models derive substantial linguistic and factual knowledge from the massive corpora on which they are trained, and prompt engineering seeks to align these models to specific tasks. Unfortunately, existing prompt engineering methods require significant amounts of labeled data, access to model parameters, or both. We introduce a new method for selecting prompt templates without labeled examples and without direct access to the model. Specifically, over a set of candidate templates, we choose the template that maximizes the mutual information between the input and the corresponding model output. Across 8 datasets representing 7 distinct NLP tasks, we show that when a template has high mutual information, it also has high accuracy on the task. On the largest model, selecting prompts with our method gets 90% of the way from the average prompt accuracy to the best prompt accuracy and requires no ground truth labels.","{'model': 'tldr@v2.0.0', 'text': 'A new method for selecting prompt templates without labeled examples and without direct access to the model is introduced, which gets 90% of the way from the average prompt accuracy to the best prompt accuracy and requires no ground truth labels.'}",https://www.cambridge.org/core/services/aop-cambridge-core/content/view/035D7C8A55B237942FB6DBAD7CAA4E49/S1047198723000025a.pdf/div-class-title-out-of-one-many-using-language-models-to-simulate-human-samples-div.pdf
-prompt distribution learning,Yuning Lu,"We present prompt distribution learning for effectively adapting a pre-trained vision-language model to address downstream recognition tasks. Our method not only learns low-bias prompts from a few samples but also captures the distribution of diverse prompts to handle the varying visual representations. In this way, we provide high-quality task-related content for facilitating recognition. This prompt distribution learning is realized by an efficient approach that learns the output embeddings of prompts instead of the input embeddings. Thus, we can employ a Gaussian distribution to model them effectively and derive a surrogate loss for efficient training. Extensive experiments on 12 datasets demonstrate that our method consistently and significantly outperforms existing methods. For example, with 1 sample per category, it relatively improves the average result by 9.1% compared to human-crafted prompts.","{'model': 'tldr@v2.0.0', 'text': 'This work presents prompt distribution learning for effectively adapting a pre-trained vision-language model to address downstream recognition tasks and employs a Gaussian distribution to model them effectively and derive a surrogate loss for efficient training.'}",https://arxiv.org/pdf/2205.03340
-ignore previous prompt: attack techniques for language models,Fábio Perez,"Transformer-based large language models (LLMs) provide a powerful foundation for natural language tasks in large-scale customer-facing applications. However, studies that explore their vulnerabilities emerging from malicious user interaction are scarce. By proposing PromptInject, a prosaic alignment framework for mask-based iterative adversarial prompt composition, we examine how GPT-3, the most widely deployed language model in production, can be easily misaligned by simple handcrafted inputs. In particular, we investigate two types of attacks -- goal hijacking and prompt leaking -- and demonstrate that even low-aptitude, but sufficiently ill-intentioned agents, can easily exploit GPT-3's stochastic nature, creating long-tail risks. The code for PromptInject is available at https://github.com/agencyenterprise/PromptInject.","{'model': 'tldr@v2.0.0', 'text': ""This work investigates two types of attacks -- goal hijacking and prompt leaking -- and demonstrates that even low-aptitude, but sufficiently ill-intentioned agents, can easily exploit GPT-3's stochastic nature, creating long-tail risks.""}",https://arxiv.org/pdf/2211.09527
-language models that seek for knowledge: modular search & generation for dialogue and prompt completion,Kurt Shuster,"Language models (LMs) have recently been shown to generate more factual responses by employing modularity (Zhou et al., 2021) in combination with retrieval (Adolphs et al., 2021). We extend the recent approach of Adolphs et al. (2021) to include internet search as a module. Our SeeKeR (Search engine->Knowledge->Response) method thus applies a single LM to three modular tasks in succession: search, generating knowledge, and generating a final response. We show that, when using SeeKeR as a dialogue model, it outperforms the state-of-the-art model BlenderBot 2 (Chen et al., 2021) on open-domain knowledge-grounded conversations for the same number of parameters, in terms of consistency, knowledge and per-turn engagingness. SeeKeR applied to topical prompt completions as a standard language model outperforms GPT2 (Radford et al., 2019) and GPT3 (Brown et al., 2020) in terms of factuality and topicality, despite GPT3 being a vastly larger model. Our code and models are made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that, when using SeeKeR as a dialogue model, it outperforms the state-of-the-art model BlenderBot 2 on open-domain knowledge-grounded conversations for the same number of parameters, in terms of consistency, knowledge and per-turn engagingness.'}",http://arxiv.org/pdf/2203.13224
-test-time prompt tuning for zero-shot generalization in vision-language models,Manli Shu,"Pre-trained vision-language models (e.g., CLIP) have shown promising zero-shot generalization in many downstream tasks with properly designed text prompts. Instead of relying on hand-engineered prompts, recent works learn prompts using the training data from downstream tasks. While effective, training on domain-specific data reduces a model's generalization capability to unseen new domains. In this work, we propose test-time prompt tuning (TPT), a method that can learn adaptive prompts on the fly with a single test sample. For image classification, TPT optimizes the prompt by minimizing the entropy with confidence selection so that the model has consistent predictions across different augmented views of each test sample. In evaluating generalization to natural distribution shifts, TPT improves the zero-shot top-1 accuracy of CLIP by 3.6% on average, surpassing previous prompt tuning approaches that require additional task-specific training data. In evaluating cross-dataset generalization with unseen categories, TPT performs on par with the state-of-the-art approaches that use additional training data. Project page: https://azshue.github.io/TPT.","{'model': 'tldr@v2.0.0', 'text': 'Test-time prompt tuning (TPT) is proposed, a method that can learn adaptive prompts on the fly with a single test sample and performs on par with the state-of-the-art approaches that use additional training data.'}",http://arxiv.org/pdf/2209.07511
-p-tuning v2: prompt tuning can be comparable to fine-tuning universally across scales and tasks,Xiao Liu,"Prompt tuning, which only tunes continuous prompts with a frozen language model, substantially reduces per-task storage and memory usage at training. However, in the context of NLU, prior work reveals that prompt tuning does not perform well for normal-sized pretrained models. We also find that existing methods of prompt tuning cannot handle hard sequence labeling tasks, indicating a lack of universality. We present a novel empirical finding that properly optimized prompt tuning can be universally effective across a wide range of model scales and NLU tasks. It matches the performance of finetuning while having only 0.1%-3% tuned parameters. Our method P-Tuning v2 is an implementation of Deep Prompt Tuning \cite{li2021prefix,qin2021learning} optimized and adapted for NLU. Given the universality and simplicity of P-Tuning v2, we believe it can serve as an alternative to finetuning and a strong baseline for future research.Our code and data are released at https://github.com/THUDM/P-tuning-v2.","{'model': 'tldr@v2.0.0', 'text': 'The method P-Tuning v2 is an implementation of Deep Prompt Tuning optimized and adapted for NLU and can serve as an alternative to finetuning and a strong baseline for future research.'}",
-diffusiondb: a large-scale prompt gallery dataset for text-to-image generative models,Zijie J. Wang,"With recent advancements in diffusion models, users can generate high-quality images by writing text prompts in natural language. However, generating images with desired details requires proper prompts, and it is often unclear how a model reacts to different prompts or what the best prompts are. To help researchers tackle these critical challenges, we introduce DiffusionDB, the first large-scale text-to-image prompt dataset totaling 6.5TB, containing 14 million images generated by Stable Diffusion, 1.8 million unique prompts, and hyperparameters specified by real users. We analyze the syntactic and semantic characteristics of prompts. We pinpoint specific hyperparameter values and prompt styles that can lead to model errors and present evidence of potentially harmful model usage, such as the generation of misinformation. The unprecedented scale and diversity of this human-actuated dataset provide exciting research opportunities in understanding the interplay between prompts and generative models, detecting deepfakes, and designing human-AI interaction tools to help users more easily use these models. DiffusionDB is publicly available at: https://poloclub.github.io/diffusiondb.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces DiffusionDB, the first large-scale text-to-image prompt dataset totaling 6.5TB, containing 14 million images generated by Stable Diffusion, 1.8 million unique prompts, and hyperparameters specified by real users, and analyzes the syntactic and semantic characteristics of prompts.'}",https://arxiv.org/pdf/2210.14896
-learning to prompt for continual learning,Zifeng Wang,"The mainstream paradigm behind continual learning has been to adapt the model parameters to non-stationary data distributions, where catastrophic forgetting is the central challenge. Typical methods rely on a rehearsal buffer or known task identity at test time to retrieve learned knowl-edge and address forgetting, while this work presents a new paradigm for continual learning that aims to train a more succinct memory system without accessing task identity at test time. Our method learns to dynamically prompt (L2P) a pre-trained model to learn tasks sequen-tially under different task transitions. In our proposed framework, prompts are small learnable parameters, which are maintained in a memory space. The objective is to optimize prompts to instruct the model prediction and ex-plicitly manage task-invariant and task-specific knowledge while maintaining model plasticity. We conduct comprehen-sive experiments under popular image classification bench-marks with different challenging continual learning set-tings, where L2P consistently outperforms prior state-of-the-art methods. Surprisingly, L2P achieves competitive results against rehearsal-based methods even without a re-hearsal buffer and is directly applicable to challenging task-agnostic continual learning. Source code is available at https://github.com/google-research/12p.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a new paradigm for continual learning that aims to train a more succinct memory system without accessing task identity at test time, and achieves competitive results against rehearsal-based methods even without a re-hearsal buffer.'}",https://arxiv.org/pdf/2112.08654
-prompt-aligned gradient for prompt tuning,Beier Zhu,"Thanks to the large pre-trained vision-language models (VLMs) like CLIP, we can craft a zero-shot classifier by""prompt"", e.g., the confidence score of an image being""[CLASS]""can be obtained by using the VLM provided similarity measure between the image and the prompt sentence""a photo of a [CLASS]"". Therefore, prompt shows a great potential for fast adaptation of VLMs to downstream tasks if we fine-tune the prompt-based similarity measure. However, we find a common failure that improper fine-tuning may not only undermine the prompt's inherent prediction for the task-related classes, but also for other classes in the VLM vocabulary. Existing methods still address this problem by using traditional anti-overfitting techniques such as early stopping and data augmentation, which lack a principled solution specific to prompt. We present Prompt-aligned Gradient, dubbed ProGrad, to prevent prompt tuning from forgetting the the general knowledge learned from VLMs. In particular, ProGrad only updates the prompt whose gradient is aligned (or non-conflicting) to the""general direction"", which is represented as the gradient of the KL loss of the pre-defined prompt prediction. Extensive experiments demonstrate the stronger few-shot generalization ability of ProGrad over state-of-the-art prompt tuning methods. Codes are available at https://github.com/BeierZhu/Prompt-align.","{'model': 'tldr@v2.0.0', 'text': 'Prompt-aligned Gradient is presented, dubbed ProGrad, to prevent prompt tuning from forgetting the the general knowledge learned from VLMs and demonstrates the stronger few-shot generalization ability of ProGrad over state-of-the-art prompt tuning methods.'}",https://arxiv.org/pdf/2205.14865
-domain adaptation via prompt learning,Chunjiang Ge,"Unsupervised domain adaptation (UDA) aims to adapt models learned from a well-annotated source domain to a target domain, where only unlabeled samples are given. Current UDA approaches learn domain-invariant features by aligning source and target feature spaces through statistical discrepancy minimization or adversarial training. However, these constraints could lead to the distortion of semantic feature structures and loss of class discriminability. In this article, we introduce a novel prompt learning paradigm for UDA, named domain adaptation via prompt learning (DAPrompt). In contrast to prior works, our approach learns the underlying label distribution for target domain rather than aligning domains. The main idea is to embed domain information into prompts, a form of representation generated from natural language, which is then used to perform classification. This domain information is shared only by images from the same domain, thereby dynamically adapting the classifier according to each domain. By adopting this paradigm, we show that our model not only outperforms previous methods on several cross-domain benchmarks but also is very efficient to train and easy to implement.","{'model': 'tldr@v2.0.0', 'text': 'This article introduces a novel prompt learning paradigm for UDA, named domain adaptation via prompt learning (DAPrompt), which outperforms previous methods on several cross-domain benchmarks but also is very efficient to train and easy to implement.'}",
-hyperprompt: prompt-based task-conditioning of transformers,Yun He,"Prompt-Tuning is a new paradigm for finetuning pre-trained language models in a parameter-efficient way. Here, we explore the use of HyperNetworks to generate hyper-prompts: we propose HyperPrompt, a novel architecture for prompt-based task-conditioning of self-attention in Transformers. The hyper-prompts are end-to-end learnable via generation by a HyperNetwork. HyperPrompt allows the network to learn task-specific feature maps where the hyper-prompts serve as task global memories for the queries to attend to, at the same time enabling flexible information sharing among tasks. We show that HyperPrompt is competitive against strong multi-task learning baselines with as few as $0.14\%$ of additional task-conditioning parameters, achieving great parameter and computational efficiency. Through extensive empirical experiments, we demonstrate that HyperPrompt can achieve superior performances over strong T5 multi-task learning baselines and parameter-efficient adapter variants including Prompt-Tuning and HyperFormer++ on Natural Language Understanding benchmarks of GLUE and SuperGLUE across many model sizes.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes HyperPrompt, a novel architecture for prompt-based task-conditioning of self-attention in Transformers, which can achieve superior performances over strong T5 multi-task learning baselines and parameter-efficient adapter variants including Prompt-Tuning and HyperFormer++ on Natural Language Understanding benchmarks of GLUE and SuperGLUE across many model sizes.'}",https://arxiv.org/pdf/2203.00759
-prompt for extraction? paie: prompting argument interaction for event argument extraction,Yubo Ma,"In this paper, we propose an effective yet efficient model PAIE for both sentence-level and document-level Event Argument Extraction (EAE), which also generalizes well when there is a lack of training data. On the one hand, PAIE utilizes prompt tuning for extractive objectives to take the best advantages of Pre-trained Language Models (PLMs). It introduces two span selectors based on the prompt to select start/end tokens among input texts for each role. On the other hand, it captures argument interactions via multi-role prompts and conducts joint optimization with optimal span assignments via a bipartite matching loss. Also, with a flexible prompt design, PAIE can extract multiple arguments with the same role instead of conventional heuristic threshold tuning. We have conducted extensive experiments on three benchmarks, including both sentence- and document-level EAE. The results present promising improvements from PAIE (3.5% and 2.3% F1 gains in average on three benchmarks, for PAIE-base and PAIE-large respectively). Further analysis demonstrates the efficiency, generalization to few-shot settings, and effectiveness of different extractive prompt tuning strategies. Our code is available at https://github.com/mayubo2333/PAIE.","{'model': 'tldr@v2.0.0', 'text': 'An effective yet efficient model PAIE for both sentence-level and document-level Event Argument Extraction (EAE), which also generalizes well when there is a lack of training data is proposed.'}",https://aclanthology.org/2022.acl-long.466.pdf
-promda: prompt-based data augmentation for low-resource nlu tasks,Yufei Wang,"This paper focuses on the Data Augmentation for low-resource Natural Language Understanding (NLU) tasks. We propose Prompt-based Data Augmentation model (PromDA) which only trains small-scale Soft Prompt (i.e., a set of trainable vectors) in the frozen Pre-trained Language Models (PLMs). This avoids human effort in collecting unlabeled in-domain data and maintains the quality of generated synthetic data. In addition, PromDA generates synthetic data via two different views and filters out the low-quality data using NLU models. Experiments on four benchmarks show that synthetic data produced by PromDA successfully boost up the performance of NLU models which consistently outperform several competitive baseline models, including a state-of-the-art semi-supervised model using unlabeled in-domain data. The synthetic data from PromDA are also complementary with unlabeled in-domain data. The NLU models can be further improved when they are combined for training.","{'model': 'tldr@v2.0.0', 'text': 'PromDA, a Prompt-based Data Augmentation model (PromDA) which only trains small-scale Soft Prompt in the frozen Pre-trained Language Models (PLMs) avoids human effort in collecting unlabeled in-domain data and maintains the quality of generated synthetic data.'}",https://aclanthology.org/2022.acl-long.292.pdf
-no more fine-tuning? an experimental evaluation of prompt tuning in code intelligence,Chaozheng Wang,"Pre-trained models have been shown effective in many code intelligence tasks. These models are pre-trained on large-scale unlabeled corpus and then fine-tuned in downstream tasks. However, as the inputs to pre-training and downstream tasks are in different forms, it is hard to fully explore the knowledge of pre-trained models. Besides, the performance of fine-tuning strongly relies on the amount of downstream data, while in practice, the scenarios with scarce data are common. Recent studies in the natural language processing (NLP) field show that prompt tuning, a new paradigm for tuning, alleviates the above issues and achieves promising results in various NLP tasks. In prompt tuning, the prompts inserted during tuning provide task-specific knowledge, which is especially beneficial for tasks with relatively scarce data. In this paper, we empirically evaluate the usage and effect of prompt tuning in code intelligence tasks. We conduct prompt tuning on popular pre-trained models CodeBERT and CodeT5 and experiment with three code intelligence tasks including defect prediction, code summarization, and code translation. Our experimental results show that prompt tuning consistently outperforms fine-tuning in all three tasks. In addition, prompt tuning shows great potential in low-resource scenarios, e.g., improving the BLEU scores of fine-tuning by more than 26% on average for code summarization. Our results suggest that instead of fine-tuning, we could adapt prompt tuning for code intelligence tasks to achieve better performance, especially when lacking task-specific data.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2207.11680
-personalized prompt learning for explainable recommendation,Lei Li,"Providing user-understandable explanations to justify recommendations could help users better understand the recommended items, increase the system’s ease of use, and gain users’ trust. A typical approach to realize it is natural language generation. However, previous works mostly adopt recurrent neural networks to meet the ends, leaving the potentially more effective pre-trained Transformer models under-explored. In fact, user and item IDs, as important identifiers in recommender systems, are inherently in different semantic space as words that pre-trained models were already trained on. Thus, how to effectively fuse IDs into such models becomes a critical issue. Inspired by recent advancement in prompt learning, we come up with two solutions: find alternative words to represent IDs (called discrete prompt learning) and directly input ID vectors to a pre-trained model (termed continuous prompt learning). In the latter case, ID vectors are randomly initialized but the model is trained in advance on large corpora, so they are actually in different learning stages. To bridge the gap, we further propose two training strategies: sequential tuning and recommendation as regularization. Extensive experiments show that our continuous prompt learning approach equipped with the training strategies consistently outperforms strong baselines on three datasets of explainable recommendation.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by recent advancement in prompt learning, a continuous prompt learning approach equipped with the training strategies consistently outperforms strong baselines on three datasets of explainable recommendation and proposes two training strategies: sequential tuning and recommendation as regularization.'}",https://arxiv.org/pdf/2202.07371
-towards unified conversational recommender systems via knowledge-enhanced prompt learning,Xiaolei Wang,"Conversational recommender systems (CRS) aim to proactively elicit user preference and recommend high-quality items through natural language conversations. Typically, a CRS consists of a recommendation module to predict preferred items for users and a conversation module to generate appropriate responses. To develop an effective CRS, it is essential to seamlessly integrate the two modules. Existing works either design semantic alignment strategies, or share knowledge resources and representations between the two modules. However, these approaches still rely on different architectures or techniques to develop the two modules, making it difficult for effective module integration. To address this problem, we propose a unified CRS model named UniCRS based on knowledge-enhanced prompt learning. Our approach unifies the recommendation and conversation subtasks into the prompt learning paradigm, and utilizes knowledge-enhanced prompts based on a fixed pre-trained language model (PLM) to fulfill both subtasks in a unified approach. In the prompt design, we include fused knowledge representations, task-specific soft tokens, and the dialogue context, which can provide sufficient contextual information to adapt the PLM for the CRS task. Besides, for the recommendation subtask, we also incorporate the generated response template as an important part of the prompt, to enhance the information interaction between the two subtasks. Extensive experiments on two public CRS datasets have demonstrated the effectiveness of our approach. Our code is publicly available at the link: https://github.com/RUCAIBox/UniCRS.","{'model': 'tldr@v2.0.0', 'text': 'This approach unifies the recommendation and conversation subtasks into the prompt learning paradigm, and utilizes knowledge-enhanced prompts based on a fixed pre-trained language model (PLM) to fulfill both subtasks in a unified approach.'}",https://arxiv.org/pdf/2206.09363
-bridge-prompt: towards ordinal action understanding in instructional videos,Muheng Li,"Action recognition models have shown a promising capability to classify human actions in short video clips. In a real scenario, multiple correlated human actions commonly occur in particular orders, forming semantically meaningful human activities. Conventional action recognition approaches focus on analyzing single actions. However, they fail to fully reason about the contextual relations between adjacent actions, which provide potential temporal logic for understanding long videos. In this paper, we propose a prompt-based framework, Bridge-Prompt (Br-Prompt), to model the semantics across adjacent actions, so that it simultaneously exploits both out-of-context and contextual information from a series of ordinal actions in instructional videos. More specifically, we reformulate the individual action labels as integrated text prompts for super-vision, which bridge the gap between individual action semantics. The generated text prompts are paired with corresponding video clips, and together co-train the text encoder and the video encoder via a contrastive approach. The learned vision encoder has a stronger capability for ordinal-action-related downstream tasks, e.g. action segmentation and human activity recognition. We evaluate the performances of our approach on several video datasets: Georgia Tech Egocentric Activities (GTEA), 50Salads, and the Breakfast dataset. Br-Prompt achieves state-of-the-art on multiple benchmarks. Code is available at: https://github.com/ttlmh/Bridge-Prompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper reformulates the individual action labels as integrated text prompts for super-vision, which bridge the gap between individual action semantics, and proposes a prompt-based framework, Bridge-Prompt, to model the semantics across adjacent actions, so that it simultaneously exploits both out-of-context and contextual information from a series of ordinal actions in instructional videos.'}",https://arxiv.org/pdf/2203.14104
-prompt consistency for zero-shot task generalization,Chunting Zhou,"One of the most impressive results of recent NLP history is the ability of pre-trained language models to solve new tasks in a zero-shot setting. To achieve this, NLP tasks are framed as natural language prompts, generating a response indicating the predicted output. Nonetheless, the performance in such settings often lags far behind its supervised counterpart, suggesting a large space for potential improvement. In this paper, we explore methods to utilize unlabeled data to improve zero-shot performance. Specifically, we take advantage of the fact that multiple prompts can be used to specify a single task, and propose to regularize prompt consistency, encouraging consistent predictions over this diverse set of prompts. Our method makes it possible to fine-tune the model either with extra unlabeled training data, or directly on test input at inference time in an unsupervised manner. In experiments, our approach outperforms the state-of-the-art zero-shot learner, T0 (Sanh et al., 2022), on 9 out of 11 datasets across 4 NLP tasks by up to 10.6 absolute points in terms of accuracy. The gains are often attained with a small number of unlabeled examples.","{'model': 'tldr@v2.0.0', 'text': 'This work takes advantage of the fact that multiple prompts can be used to specify a single task, and proposes to regularize prompt consistency, encouraging consistent predictions over this diverse set of prompts, to improve zero-shot performance.'}",http://arxiv.org/pdf/2205.00049
-promptcap: prompt-guided task-aware image captioning,Yushi Hu,"Knowledge-based visual question answering (VQA) involves questions that require world knowledge beyond the image to yield the correct answer. Large language models (LMs) like GPT-3 are particularly helpful for this task because of their strong knowledge retrieval and reasoning capabilities. To enable LM to understand images, prior work uses a captioning model to convert images into text. However, when summarizing an image in a single caption sentence, which visual entities to describe are often underspecified. Generic image captions often miss visual details essential for the LM to answer visual questions correctly. To address this challenge, we propose PromptCap (Prompt-guided image Captioning), a captioning model designed to serve as a better connector between images and black-box LMs. Different from generic captions, PromptCap takes a natural-language prompt to control the visual entities to describe in the generated caption. The prompt contains a question that the caption should aid in answering. To avoid extra annotation, PromptCap is trained by examples synthesized with GPT-3 and existing datasets. We demonstrate PromptCap's effectiveness on an existing pipeline in which GPT-3 is prompted with image captions to carry out VQA. PromptCap outperforms generic captions by a large margin and achieves state-of-the-art accuracy on knowledge-based VQA tasks (60.4% on OK-VQA and 59.6% on A-OKVQA). Zero-shot results on WebQA show that PromptCap generalizes well to unseen domains.","{'model': 'tldr@v2.0.0', 'text': 'PromptCap (Prompt-guided image Captioning), a captioning model designed to serve as a better connector between images and black-box LMs, achieves state-of-the-art accuracy on knowledge-based VQA tasks and generalizes well to unseen domains.'}",https://arxiv.org/pdf/2211.09699
-spot: better frozen model adaptation through soft prompt transfer,Tu Vu,"There has been growing interest in parameter-efficient methods to apply pre-trained language models to downstream tasks. Building on the Prompt Tuning approach of Lester et al. (2021), which learns task-specific soft prompts to condition a frozen pre-trained model to perform different tasks, we propose a novel prompt-based transfer learning approach called SPoT: Soft Prompt Transfer. SPoT first learns a prompt on one or more source tasks and then uses it to initialize the prompt for a target task. We show that SPoT significantly boosts the performance of Prompt Tuning across many tasks. More remarkably, across all model sizes, SPoT matches or outperforms standard Model Tuning (which fine-tunes all model parameters) on the SuperGLUE benchmark, while using up to 27,000× fewer task-specific parameters. To understand where SPoT is most effective, we conduct a large-scale study on task transferability with 26 NLP tasks in 160 combinations, and demonstrate that many tasks can benefit each other via prompt transfer. Finally, we propose an efficient retrieval approach that interprets task prompts as task embeddings to identify similar tasks and predict the most transferable source tasks for a novel target task.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that SPoT significantly boosts the performance of Prompt Tuning across many tasks, and an efficient retrieval approach is proposed that interprets task prompts as task embeddings to identify similar tasks and predict the most transferable source tasks for a novel target task.'}",https://aclanthology.org/2022.acl-long.346.pdf
-prompt programming for large language models: beyond the few-shot paradigm,Laria Reynolds,"Prevailing methods for mapping large generative language models to supervised tasks may fail to sufficiently probe models’ novel capabilities. Using GPT-3 as a case study, we show that 0-shot prompts can significantly outperform few-shot prompts. We suggest that the function of few-shot examples in these cases is better described as locating an already learned task rather than meta-learning. This analysis motivates rethinking the role of prompts in controlling and evaluating powerful language models. We discuss methods of prompt programming, emphasizing the usefulness of considering prompts through the lens of natural language. We explore techniques for exploiting the capacity of narratives and cultural anchors to encode nuanced intentions and techniques for encouraging deconstruction of a problem into components before producing a verdict. Informed by this more encompassing theory of prompt programming, we also introduce the idea of a metaprompt that seeds the model to generate its own natural language prompts for a range of tasks. Finally, we discuss how these more general methods of interacting with language models can be incorporated into existing and future benchmarks and practical applications.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that the function of few-shot examples in these cases is better described as locating an already learned task rather than meta-learning, which motivates rethinking the role of prompts in controlling and evaluating powerful language models.'}",https://arxiv.org/pdf/2102.07350
-coda-prompt: continual decomposed attention-based prompting for rehearsal-free continual learning,James Smith,"Computer vision models suffer from a phenomenon known as catastrophic forgetting when learning novel concepts from continuously shifting training data. Typical solutions for this continual learning problem require extensive rehearsal of previously seen data, which increases memory costs and may violate data privacy. Recently, the emergence of large-scale pre-trained vision transformer models has enabled prompting approaches as an alternative to data-rehearsal. These approaches rely on a key-query mechanism to generate prompts and have been found to be highly resistant to catastrophic forgetting in the well-established rehearsal-free continual learning setting. However, the key mechanism of these methods is not trained end-to-end with the task sequence. Our experiments show that this leads to a reduction in their plasticity, hence sacrificing new task accuracy, and inability to benefit from expanded parameter capacity. We instead propose to learn a set of prompt components which are assembled with input-conditioned weights to produce input-conditioned prompts, resulting in a novel attention-based end-to-end key-query scheme. Our experiments show that we outperform the current SOTA method DualPrompt on established benchmarks by as much as 4.5% in average final accuracy. We also outperform the state of art by as much as 4.4% accuracy on a continual learning benchmark which contains both class-incremental and domain-incremental task shifts, corresponding to many practical settings. Our code is available at https://github.com/GT-RIPL/CODA-Prompt","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to learn a set of prompt components which are assembled with input- Conditioned weights to produce input-conditioned prompts, resulting in a novel attention-based end-to-end key-query scheme.'}",https://arxiv.org/pdf/2211.13218
-prompt learning with optimal transport for vision-language models,Guangyi Chen,"With the increasing attention to large vision-language models such as CLIP, there has been a significant amount of effort dedicated to building efficient prompts. Unlike conventional methods of only learning one single prompt, we propose to learn multiple comprehensive prompts to describe diverse characteristics of categories such as intrinsic attributes or extrinsic contexts. However, directly matching each prompt to the same visual feature is problematic, as it pushes the prompts to converge to one point. To solve this problem, we propose to apply optimal transport to match the vision and text modalities. Specifically, we first model images and the categories with visual and textual feature sets. Then, we apply a two-stage optimization strategy to learn the prompts. In the inner loop, we optimize the optimal transport distance to align visual features and prompts by the Sinkhorn algorithm, while in the outer loop, we learn the prompts by this distance from the supervised data. Extensive experiments are conducted on the few-shot recognition task and the improvement demonstrates the superiority of our method. The code is available at https://github.com/CHENGY12/PLOT.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to apply optimal transport to match the vision and text modalities and first model images and the categories with visual and textual feature sets and applies a two-stage optimization strategy to learn the prompts.'}",http://arxiv.org/pdf/2210.01253
-idpg: an instance-dependent prompt generation method,Zhuofeng Wu,"Prompt tuning is a new, efficient NLP transfer learning paradigm that adds a task-specific prompt in each input instance during the model training stage. It freezes the pre-trained language model and only optimizes a few task-specific prompts. In this paper, we propose a conditional prompt generation method to generate prompts for each input instance, referred to as the Instance-Dependent Prompt Generation (IDPG). Unlike traditional prompt tuning methods that use a fixed prompt, IDPG introduces a lightweight and trainable component to generate prompts based on each input sentence. Extensive experiments on ten natural language understanding (NLU) tasks show that the proposed strategy consistently outperforms various prompt tuning baselines and is on par with other efficient transfer learning methods such as Compacter while tuning far fewer model parameters.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments on ten natural language understanding (NLU) tasks show that the proposed strategy consistently outperforms various prompt tuning baselines and is on par with other efficient transfer learning methods such as Compacter while tuning far fewer model parameters.'}",http://arxiv.org/pdf/2204.04497
-continual prompt tuning for dialog state tracking,Qi Zhu,"A desirable dialog system should be able to continually learn new skills without forgetting old ones, and thereby adapt to new domains or tasks in its life cycle. However, continually training a model often leads to a well-known catastrophic forgetting issue. In this paper, we present Continual Prompt Tuning, a parameter-efficient framework that not only avoids forgetting but also enables knowledge transfer between tasks. To avoid forgetting, we only learn and store a few prompt tokens’ embeddings for each task while freezing the backbone pre-trained model. To achieve bi-directional knowledge transfer among tasks, we propose several techniques (continual prompt initialization, query fusion, and memory replay) to transfer knowledge from preceding tasks and a memory-guided technique to transfer knowledge from subsequent tasks. Extensive experiments demonstrate the effectiveness and efficiency of our proposed method on continual learning for dialog state tracking, compared with state-of-the-art baselines.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Continual Prompt Tuning, a parameter-efficient framework that not only avoids forgetting but also enables knowledge transfer between tasks, and proposes several techniques to transfer knowledge from preceding tasks and a memory-guided technique to transferknowledge from subsequent tasks.'}",http://arxiv.org/pdf/2203.06654
-exploring the universal vulnerability of prompt-based learning paradigm,Lei Xu,"Prompt-based learning paradigm bridges the gap between pre-training and fine-tuning, and works effectively under the few-shot setting. However, we find that this learning paradigm inherits the vulnerability from the pre-training stage, where model predictions can be misled by inserting certain triggers into the text. In this paper, we explore this universal vulnerability by either injecting backdoor triggers or searching for adversarial triggers on pre-trained language models using only plain text. In both scenarios, we demonstrate that our triggers can totally control or severely decrease the performance of prompt-based models fine-tuned on arbitrary downstream tasks, reflecting the universal vulnerability of the prompt-based learning paradigm. Further experiments show that adversarial triggers have good transferability among language models. We also find conventional fine-tuning models are not vulnerable to adversarial triggers constructed from pre-trained language models. We conclude by proposing a potential solution to mitigate our attack methods. Code and data are publicly available at https://github.com/leix28/prompt-universal-vulnerability","{'model': 'tldr@v2.0.0', 'text': 'This paper demonstrates that backdoor triggers or searching for adversarial triggers on pre-trained language models using only plain text can totally control or severely decrease the performance of prompt-based models fine-tuned on arbitrary downstream tasks, reflecting the universal vulnerability of the prompt- based learning paradigm.'}",http://arxiv.org/pdf/2204.05239
-how many data points is a prompt worth?,Teven Le Scao,"When fine-tuning pretrained models for classification, researchers either use a generic model head or a task-specific prompt for prediction. Proponents of prompting have argued that prompts provide a method for injecting task-specific guidance, which is beneficial in low-data regimes. We aim to quantify this benefit through rigorous testing of prompts in a fair setting: comparing prompted and head-based fine-tuning in equal conditions across many tasks and data sizes. By controlling for many sources of advantage, we find that prompting does indeed provide a benefit, and that this benefit can be quantified per task. Results show that prompting is often worth 100s of data points on average across classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'It is found that prompting does indeed provide a benefit, and that this benefit can be quantified per task, and results show that prompting is often worth 100s of data points on average across classification tasks.'}",https://aclanthology.org/2021.naacl-main.208.pdf
-knowprompt: knowledge-aware prompt-tuning with synergistic optimization for relation extraction,Xiang Chen,"Recently, prompt-tuning has achieved promising results for specific few-shot classification tasks. The core idea of prompt-tuning is to insert text pieces (i.e., templates) into the input and transform a classification task into a masked language modeling problem. However, for relation extraction, determining an appropriate prompt template requires domain expertise, and it is cumbersome and time-consuming to obtain a suitable label word. Furthermore, there exists abundant semantic and prior knowledge among the relation labels that cannot be ignored. To this end, we focus on incorporating knowledge among relation labels into prompt-tuning for relation extraction and propose a Knowledge-aware Prompt-tuning approach with synergistic optimization (KnowPrompt). Specifically, we inject latent knowledge contained in relation labels into prompt construction with learnable virtual type words and answer words. Then, we synergistically optimize their representation with structured constraints. Extensive experimental results on five datasets with standard and low-resource settings demonstrate the effectiveness of our approach. Our code and datasets are available in GitHub1 for reproducibility.","{'model': 'tldr@v2.0.0', 'text': 'A Knowledge-aware Prompt-tuning approach with synergistic optimization (KnowPrompt) that injects latent knowledge contained in relation labels into prompt construction with learnable virtual type words and answer words.'}",https://arxiv.org/pdf/2104.07650
-knowledgeable prompt-tuning: incorporating knowledge into prompt verbalizer for text classification,Shengding Hu,"Tuning pre-trained language models (PLMs) with task-specific prompts has been a promising approach for text classification. Particularly, previous studies suggest that prompt-tuning has remarkable superiority in the low-data scenario over the generic fine-tuning methods with extra classifiers. The core idea of prompt-tuning is to insert text pieces, i.e., template, to the input and transform a classification problem into a masked language modeling problem, where a crucial step is to construct a projection, i.e., verbalizer, between a label space and a label word space. A verbalizer is usually handcrafted or searched by gradient descent, which may lack coverage and bring considerable bias and high variances to the results. In this work, we focus on incorporating external knowledge into the verbalizer, forming a knowledgeable prompttuning (KPT), to improve and stabilize prompttuning. Specifically, we expand the label word space of the verbalizer using external knowledge bases (KBs) and refine the expanded label word space with the PLM itself before predicting with the expanded label word space. Extensive experiments on zero and few-shot text classification tasks demonstrate the effectiveness of knowledgeable prompt-tuning.","{'model': 'tldr@v2.0.0', 'text': 'This work focuses on incorporating external knowledge into the verbalizer, forming a knowledgeable prompt Tuning (KPT), to improve and stabilize prompttuning.'}",https://aclanthology.org/2022.acl-long.158.pdf
-pro-tuning: unified prompt tuning for vision tasks,Xing Nie,"In computer vision, fine-tuning is the de-facto approach to leverage pre-trained vision models to perform downstream tasks. However, deploying it in practice is quite challenging, due to adopting parameter inefficient global update and heavily relying on high-quality downstream data. Recently, prompt-based learning, which adds a task-relevant prompt to adapt the downstream tasks to pre-trained models, has drastically boosted the performance of many natural language downstream tasks. In this work, we extend this notable transfer ability benefited from prompt into vision models as an alternative to fine-tuning. To this end, we propose parameter-efficient Prompt tuning (Pro-tuning) to adapt frozen vision models to various downstream vision tasks. The key to Pro-tuning is prompt-based tuning, i.e., learning task-specific vision prompts for downstream input images with the pre-trained model frozen. By only training a few additional parameters, it can work on diverse CNN-based and Transformer-based architectures. Extensive experiments evidence that Pro-tuning outperforms fine-tuning in a broad range of vision tasks and scenarios, including image classification (generic objects, class imbalance, image corruption, adversarial robustness, and out-of-distribution generalization), and dense prediction tasks such as object detection and semantic segmentation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes parameter-efficient Prompt tuning (Pro-tuning) to adapt frozen vision models to various downstream vision tasks, which outperforms fine- Tuning in a broad range of vision tasks and scenarios, including image classification, and dense prediction tasks such as object detection and semantic segmentation.'}",http://arxiv.org/pdf/2207.14381
-interactive and visual prompt engineering for ad-hoc task adaptation with large language models,Hendrik Strobelt,"State-of-the-art neural language models can now be used to solve ad-hoc language tasks through zero-shot prompting without the need for supervised training. This approach has gained popularity in recent years, and researchers have demonstrated prompts that achieve strong accuracy on specific NLP tasks. However, finding a prompt for new tasks requires experimentation. Different prompt templates with different wording choices lead to significant accuracy differences. PromptIDE allows users to experiment with prompt variations, visualize prompt performance, and iteratively optimize prompts. We developed a workflow that allows users to first focus on model feedback using small data before moving on to a large data regime that allows empirical grounding of promising prompts using quantitative measures of the task. The tool then allows easy deployment of the newly created ad-hoc models. We demonstrate the utility of PromptIDE (demo: http://prompt.vizhub.ai) and our workflow using several real-world use cases.","{'model': 'tldr@v2.0.0', 'text': 'A workflow that allows users to first focus on model feedback using small data before moving on to a large data regime that allows empirical grounding of promising prompts using quantitative measures of the task, and then allows easy deployment of the newly created ad-hoc models.'}",https://arxiv.org/pdf/2208.07852
-promptmaker: prompt-based prototyping with large language models,Ellen Jiang,"Prototyping is notoriously difficult to do with machine learning (ML), but recent advances in large language models may lower the barriers to people prototyping with ML, through the use of natural language prompts. This case study reports on the real-world experiences of industry professionals (e.g. designers, program managers, front-end developers) prototyping new ML-powered feature ideas via prompt-based prototyping. Through interviews with eleven practitioners during a three-week sprint and a workshop, we find that prompt-based prototyping reduced barriers of access by substantially broadening who can prototype with ML, sped up the prototyping process, and grounded communication between collaborators. Yet, it also introduced new challenges, such as the need to reverse-engineer prompt designs, source example data, and debug and evaluate prompt effectiveness. Taken together, this case study provides important implications that lay the groundwork toward a new future of prototyping with ML.","{'model': 'tldr@v2.0.0', 'text': None}",
-dynamic prompt learning via policy gradient for semi-structured mathematical reasoning,Pan Lu,"Mathematical reasoning, a core ability of human intelligence, presents unique challenges for machines in abstract thinking and logical reasoning. Recent large pre-trained language models such as GPT-3 have achieved remarkable progress on mathematical reasoning tasks written in text form, such as math word problems (MWP). However, it is unknown if the models can handle more complex problems that involve math reasoning over heterogeneous information, such as tabular data. To fill the gap, we present Tabular Math Word Problems (TabMWP), a new dataset containing 38,431 open-domain grade-level problems that require mathematical reasoning on both textual and tabular data. Each question in TabMWP is aligned with a tabular context, which is presented as an image, semi-structured text, and a structured table. There are two types of questions: free-text and multi-choice, and each problem is annotated with gold solutions to reveal the multi-step reasoning process. We evaluate different pre-trained models on TabMWP, including the GPT-3 model in a few-shot setting. As earlier studies suggest, since few-shot GPT-3 relies on the selection of in-context examples, its performance is unstable and can degrade to near chance. The unstable issue is more severe when handling complex problems like TabMWP. To mitigate this, we further propose a novel approach, PromptPG, which utilizes policy gradient to learn to select in-context examples from a small amount of training data and then constructs the corresponding prompt for the test example. Experimental results show that our method outperforms the best baseline by 5.31% on the accuracy metric and reduces the prediction variance significantly compared to random selection, which verifies its effectiveness in selecting in-context examples.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach is proposed, PromptPG, which utilizes policy gradient to learn to select in-context examples from a small amount of training data and then constructs the corresponding prompt for the test example, which verifies its effectiveness in selecting in- context examples.'}",http://arxiv.org/pdf/2209.14610
-conversing with copilot: exploring prompt engineering for solving cs1 problems using natural language,Paul Denny,"GitHub Copilot is an artificial intelligence tool for automatically generating source code from natural language problem descriptions. Since June 2022, Copilot has officially been available for free to all students as a plug-in to development environments like Visual Studio Code. Prior work exploring OpenAI Codex, the underlying model that powers Copilot, has shown it performs well on typical CS1 problems thus raising concerns about its potential impact on how introductory programming courses are taught. However, little is known about the types of problems for which Copilot does not perform well, or about the natural language interactions that a student might have with Copilot when resolving errors. We explore these questions by evaluating the performance of Copilot on a publicly available dataset of 166 programming problems. We find that it successfully solves around half of these problems on its very first attempt, and that it solves 60% of the remaining problems using only natural language changes to the problem description. We argue that this type of prompt engineering, which we believe will become a standard interaction between human and Copilot when it initially fails, is a potentially useful learning activity that promotes computational thinking skills, and is likely to change the nature of code writing skill development.","{'model': 'tldr@v2.0.0', 'text': 'Evaluating the performance of Copilot on a publicly available dataset of 166 programming problems finds that it successfully solves around half of these problems on its very first attempt, and that it solves 60% of the remaining problems using only natural language changes to the problem description.'}",https://eprints.iisc.ac.in/81157/1/SIGCSE_2023.pdf
-"zeroprompt: scaling prompt-based pretraining to 1, 000 tasks improves zero-shot generalization",Hanwei Xu,"We propose a multitask pretraining approach ZeroPrompt for zero-shot generalization, focusing on task scaling and zero-shot prompting. While previous models are trained on only a few dozen tasks, we scale to 1,000 tasks for the first time using real-world data. This leads to a crucial discovery that task scaling can be an efficient alternative to model scaling; i.e., the model size has little impact on performance with an extremely large number of tasks. Our results show that task scaling can substantially improve training efficiency by 30 times in FLOPs. Moreover, we present a prompting method that incorporates a genetic algorithm to automatically search for the best prompt for unseen tasks, along with a few other improvements. Empirically, ZeroPrompt substantially improves both the efficiency and the performance of zero-shot learning across a variety of academic and production datasets.","{'model': 'tldr@v2.0.0', 'text': 'The results show that task scaling can substantially improve training efficiency by 30 times in FLOPs, and a prompting method that incorporates a genetic algorithm to automatically search for the best prompt for unseen tasks, along with a few other improvements.'}",https://aclanthology.org/2022.findings-emnlp.312.pdf
-prompt-free and efficient few-shot learning with language models,Rabeeh Karimi Mahabadi,"Current methods for few-shot fine-tuning of pretrained masked language models (PLMs) require carefully engineered prompts and verbalizers for each new task to convert examples into a cloze-format that the PLM can score. In this work, we propose Perfect, a simple and efficient method for few-shot fine-tuning of PLMs without relying on any such handcrafting, which is highly effective given as few as 32 data points. Perfect makes two key design choices: First, we show that manually engineered task prompts can be replaced with task-specific adapters that enable sample-efficient fine-tuning and reduce memory and storage costs by roughly factors of 5 and 100, respectively. Second, instead of using handcrafted verbalizers, we learn new multi-token label embeddings during fine-tuning, which are not tied to the model vocabulary and which allow us to avoid complex auto-regressive decoding. These embeddings are not only learnable from limited data but also enable nearly 100x faster training and inference. Experiments on a wide range of few shot NLP tasks demonstrate that Perfect, while being simple and efficient, also outperforms existing state-of-the-art few-shot learning methods. Our code is publicly available at https://github.com/rabeehk/perfect.","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that Perfect, a simple and efficient method for few-shot fine-tuning of PLMs without relying on any such handcrafting, also outperforms existing state-of-the-art few- shot learning methods.'}",http://arxiv.org/pdf/2204.01172
-fantastically ordered prompts and where to find them: overcoming few-shot prompt order sensitivity,Yao Lu,"When primed with only a handful of training samples, very large, pretrained language models such as GPT-3 have shown competitive results when compared to fully-supervised, fine-tuned, large, pretrained language models. We demonstrate that the order in which the samples are provided can make the difference between near state-of-the-art and random guess performance: essentially some permutations are “fantastic” and some not. We analyse this phenomenon in detail, establishing that: it is present across model sizes (even for the largest current models), it is not related to a specific subset of samples, and that a given good permutation for one model is not transferable to another. While one could use a development set to determine which permutations are performant, this would deviate from the true few-shot setting as it requires additional annotated data. Instead, we use the generative nature of language models to construct an artificial development set and based on entropy statistics of the candidate permutations on this set, we identify performant prompts. Our method yields a 13% relative improvement for GPT-family models across eleven different established text classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work uses the generative nature of language models to construct an artificial development set and based on entropy statistics of the candidate permutations on this set, it identifies performant prompts and yields a 13% relative improvement for GPT-family models across eleven different established text classification tasks.'}",https://aclanthology.org/2022.acl-long.556.pdf
-ptr: prompt tuning with rules for text classification,Xu Han,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt tuning with rules (PTR) for many-class text classification and applies logic rules to construct prompts with several sub-prompts, indicating that PTR is a promising approach to take advantage of both human prior knowledge and PLMs for those complicated classification tasks.'}",
-iteratively prompt pre-trained language models for chain of thought,Boshi Wang,"While Pre-trained Language Models (PLMs) internalize a great amount of world knowledge, they have been shown incapable of recalling these knowledge to solve tasks requiring complex & multi-step reasoning. Similar to how humans develop a “chain of thought” for these tasks, how can we equip PLMs with such abilities? In this work, we explore an iterative prompting framework, a new prompting paradigm which progressively elicits relevant knowledge from PLMs for multi-step inference. We identify key limitations of existing prompting methods, namely they are either restricted to queries with a single identifiable relation/predicate, or being agnostic to input contexts, which makes it difficult to capture variabilities across different inference steps. We propose an iterative context-aware prompter, which addresses these limitations by learning to dynamically synthesize prompts conditioned on the current step’s contexts. Experiments on three datasets involving multi-step reasoning show the effectiveness of the iterative scheme and the context-aware prompter design.","{'model': 'tldr@v2.0.0', 'text': 'An iterative prompting framework is explored, a new prompting paradigm which progressively elicits relevant knowledge from PLMs for multi-step inference by learning to dynamically synthesize prompts conditioned on the current step’s contexts.'}",https://aclanthology.org/2022.emnlp-main.174.pdf
-black-box prompt learning for pre-trained language models,Shizhe Diao,"The increasing scale of general-purpose Pre-trained Language Models (PLMs) necessitates the study of more efficient adaptation across different downstream tasks. In this paper, we establish a Black-box Discrete Prompt Learning (BDPL) to resonate with pragmatic interactions between the cloud infrastructure and edge devices. Particularly, instead of fine-tuning the model in the cloud, we adapt PLMs by prompt learning, which efficiently optimizes only a few parameters of the discrete prompts. Moreover, we consider the scenario that we do not have access to the parameters and gradients of the pre-trained models, except for its outputs given inputs. This black-box setting secures the cloud infrastructure from potential attack and misuse to cause a single-point failure, which is preferable to the white-box counterpart by current infrastructures. Under this black-box constraint, we apply a variance-reduced policy gradient algorithm to estimate the gradients of parameters in the categorical distribution of each discrete prompt. In light of our method, the user devices can efficiently tune their tasks by querying the PLMs bounded by a range of API calls. Our experiments on RoBERTa and GPT-3 demonstrate that the proposed algorithm achieves significant improvement on eight benchmarks in a cloud-device collaboration manner. Finally, we conduct in-depth case studies to comprehensively analyze our method in terms of various data sizes, prompt lengths, training budgets, optimization objectives, prompt transferability, and explanations of the learned prompts. Our code will be available at https://github.com/shizhediao/Black-Box-Prompt-Learning.","{'model': 'tldr@v2.0.0', 'text': 'A Black-box Discrete Prompt Learning (BDPL) is established to resonate with pragmatic interactions between the cloud infrastructure and edge devices and achieves significant improvement on eight benchmarks in a cloud-device collaboration manner.'}",
-visual prompt tuning for test-time domain adaptation,Yunhe Gao,"Models should be able to adapt to unseen data during test-time to avoid performance drops caused by inevitable distribution shifts in real-world deployment scenarios. In this work, we tackle the practical yet challenging test-time adaptation (TTA) problem, where a model adapts to the target domain without accessing the source data. We propose a simple recipe called \textit{Data-efficient Prompt Tuning} (DePT) with two key ingredients. First, DePT plugs visual prompts into the vision Transformer and only tunes these source-initialized prompts during adaptation. We find such parameter-efficient finetuning can efficiently adapt the model representation to the target domain without overfitting to the noise in the learning objective. Second, DePT bootstraps the source representation to the target domain by memory bank-based online pseudo-labeling. A hierarchical self-supervised regularization specially designed for prompts is jointly optimized to alleviate error accumulation during self-training. With much fewer tunable parameters, DePT demonstrates not only state-of-the-art performance on major adaptation benchmarks VisDA-C, ImageNet-C, and DomainNet-126, but also superior data efficiency, i.e., adaptation with only 1\% or 10\% data without much performance degradation compared to 100\% data. In addition, DePT is also versatile to be extended to online or multi-source TTA settings.","{'model': 'tldr@v2.0.0', 'text': 'This work tackles the practical yet challenging test-time adaptation (TTA) problem, where a model adapts to the target domain without accessing the source data, with a simple recipe called Data-efficient Prompt Tuning (DePT).'}",https://arxiv.org/pdf/2210.04831
-repository-level prompt generation for large language models of code,Disha Shrivastava,"With the success of large language models (LLMs) of code and their use as code assistants (e.g. Codex used in GitHub Copilot), techniques for introducing domain-specific knowledge in the prompt design process become important. In this work, we propose a framework called Repo-Level Prompt Generator that learns to generate example-specific prompts using prompt proposals. The prompt proposals take context from the entire repository, thereby incorporating both the structure of the repository and the context from other relevant files (e.g. imports, parent class files). Our technique doesn't require any access to the weights of the LLM, making it applicable in cases where we only have black-box access to the LLM. We conduct experiments on the task of single-line code-autocompletion using code repositories taken from Google Code archives. We demonstrate that an oracle constructed from our prompt proposals gives a remarkably high relative improvement of 36% over Codex, showing the quality of these proposals. Further, we show that when we train a model to predict a prompt proposal, we can achieve significant performance gains over Codex and other baselines. We release our code, data, and trained checkpoints at: \url{https://github.com/shrivastavadisha/repo_level_prompt_generation}.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a framework called Repo-Level Prompt Generator that learns to generate example-specific prompts using prompt proposals that take context from the entire repository, thereby incorporating both the structure of the repository and the context from other relevant files.'}",http://arxiv.org/pdf/2206.12839
-gppt: graph pre-training and prompt tuning to generalize graph neural networks,Mingchen Sun,"Despite the promising representation learning of graph neural networks (GNNs), the supervised training of GNNs notoriously requires large amounts of labeled data from each application. An effective solution is to apply the transfer learning in graph: using easily accessible information to pre-train GNNs, and fine-tuning them to optimize the downstream task with only a few labels. Recently, many efforts have been paid to design the self-supervised pretext tasks, and encode the universal graph knowledge among the various applications. However, they rarely notice the inherent training objective gap between the pretext and downstream tasks. This significant gap often requires costly fine-tuning for adapting the pre-trained model to downstream problem, which prevents the efficient elicitation of pre-trained knowledge and then results in poor results. Even worse, the naive pre-training strategy usually deteriorates the downstream task, and damages the reliability of transfer learning in graph data. To bridge the task gap, we propose a novel transfer learning paradigm to generalize GNNs, namely graph pre-training and prompt tuning (GPPT). Specifically, we first adopt the masked edge prediction, the most simplest and popular pretext task, to pre-train GNNs. Based on the pre-trained model, we propose the graph prompting function to modify the standalone node into a token pair, and reformulate the downstream node classification looking the same as edge prediction. The token pair is consisted of candidate label class and node entity. Therefore, the pre-trained GNNs could be applied without tedious fine-tuning to evaluate the linking probability of token pair, and produce the node classification decision. The extensive experiments on eight benchmark datasets demonstrate the superiority of GPPT, delivering an average improvement of 4.29% in few-shot graph analysis and accelerating the model convergence up to 4.32X. The code is available in: https://github.com/MingChen-Sun/GPPT.","{'model': 'tldr@v2.0.0', 'text': 'This work first adopts the masked edge prediction, the most simplest and popular pretext task, to pre-train GNNs, and proposes the graph prompting function to modify the standalone node into a token pair, and reformulate the downstream node classification looking the same as edge prediction.'}",
-visual prompt tuning for generative transfer learning,Kihyuk Sohn,"Learning generative image models from various domains efficiently needs transferring knowledge from an image synthesis model trained on a large dataset. We present a recipe for learning vision transformers by generative knowledge transfer. We base our framework on generative vision transformers representing an image as a sequence of visual tokens with the autoregressive or non-autoregressive transformers. To adapt to a new domain, we employ prompt tuning, which prepends learnable tokens called prompts to the image token sequence and introduces a new prompt design for our task. We study on a variety of visual domains with varying amounts of training images. We show the effectiveness of knowledge transfer and a significantly better image generation quality.11https://github.com/google-research/generative_transfer","{'model': 'tldr@v2.0.0', 'text': 'This work presents a recipe for learning vision transformers by generative knowledge transfer, and shows the effectiveness of knowledge transfer and a significantly better image generation quality.'}",https://arxiv.org/pdf/2210.00990
-prompt vision transformer for domain generalization,Zangwei Zheng,"Though vision transformers (ViTs) have exhibited impressive ability for representation learning, we empirically find that they cannot generalize well to unseen domains with previous domain generalization algorithms. In this paper, we propose a novel approach DoPrompt based on prompt learning to embed the knowledge of source domains in domain prompts for target domain prediction. Specifically, domain prompts are prepended before ViT input tokens from the corresponding source domain. Each domain prompt learns domain-specific knowledge efficiently since it is optimized only for one domain. Meanwhile, we train a prompt adapter to produce a suitable prompt for each input image based on the learned source domain prompts. At test time, the adapted prompt generated by the prompt adapter can exploit the similarity between the feature of the out-of-domain image and source domains to properly integrate the source domain knowledge. Extensive experiments are conducted on four benchmark datasets. Our approach achieves 1.4% improvements in the averaged accuracy, which is 3.5 times the improvement of the state-of-the-art algorithm with a ViT backbone.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach DoPrompt based on prompt learning to embed the knowledge of source domains in domain prompts for target domain prediction, which achieves 1.4% improvements in the averaged accuracy and 3.5 times the improvement of the state-of-the-art algorithm with a ViT backbone.'}",http://arxiv.org/pdf/2208.08914
-prompt tuning for discriminative pre-trained language models,Yuan Yao,"Recent works have shown promising results of prompt tuning in stimulating pre-trained language models (PLMs) for natural language processing (NLP) tasks. However, to the best of our knowledge, existing works focus on prompt-tuning generative PLMs that are pre-trained to generate target tokens, such as BERT. It is still unknown whether and how discriminative PLMs, e.g., ELECTRA, can be effectively prompt-tuned. In this work, we present DPT, the first prompt tuning framework for discriminative PLMs, which reformulates NLP tasks into a discriminative language modeling problem. Comprehensive experiments on text classification and question answering show that, compared with vanilla fine-tuning, DPT achieves significantly higher performance, and also prevents the unstable problem in tuning large PLMs in both full-set and low-resource settings.","{'model': 'tldr@v2.0.0', 'text': 'DPT is presented, the first prompt tuning framework for discriminative PLMs, which reformulates NLP tasks into a discrim inative language modeling problem, and achieves significantly higher performance, and also prevents the unstable problem in tuning large PLMs in both full-set and low-resource settings.'}",https://arxiv.org/pdf/2205.11166
-incremental prompting: episodic memory prompt for lifelong event detection,Minqian Liu,"Lifelong event detection aims to incrementally update a model with new event types and data while retaining the capability on previously learned old types. One critical challenge is that the model would catastrophically forget old types when continually trained on new data. In this paper, we introduce Episodic Memory Prompts (EMP) to explicitly retain the learned task-specific knowledge. Our method adopts continuous prompt for each task and they are optimized to instruct the model prediction and learn event-specific representation. The EMPs learned in previous tasks are carried along with the model in subsequent tasks, and can serve as a memory module that keeps the old knowledge and transferring to new tasks. Experiment results demonstrate the effectiveness of our method. Furthermore, we also conduct a comprehensive analysis of the new and old event types in lifelong learning.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Episodic Memory Prompts (EMP) to explicitly retain the learned task-specific knowledge and conduct a comprehensive analysis of the new and old event types in lifelong learning.'}",http://arxiv.org/pdf/2204.07275
-prompt-matched semantic segmentation,Lingbo Liu,"The objective of this work is to explore how to effectively and efficiently adapt pre-trained visual foundation models to various downstream tasks of semantic segmentation. Previous methods usually fine-tuned the entire networks for each specific dataset, which will be burdensome to store massive parameters of these networks. A few recent works attempted to insert some extra trainable parameters into the frozen networks to learn visual prompts for parameter-efficient tuning. However, these works showed poor generality as they were designed specifically for Transformers. Moreover, using limited information in these schemes, they exhibited a poor capacity to learn beneficial prompts. To alleviate these issues, we propose a novel Stage-wise Prompt-Matched Framework for generic and effective visual prompt tuning. Specifically, to ensure generality, we divide the pre-trained backbone with frozen parameters into multiple stages and perform prompt learning between different stages, which makes the proposed scheme applicable to various architectures of CNN and Transformer. For effective tuning, a lightweight Semantic-aware Prompt Matcher (SPM) is designed to progressively learn reasonable prompts with a recurrent mechanism, guided by the rich information of interim semantic maps. Working as deep matched filter of representation learning, the proposed SPM can well transform the output of the previous stage into a desirable input for the next stage, thus achieving the better matching/stimulating for the pre-trained knowledge. Extensive experiments on four benchmarks demonstrate that the proposed scheme can achieve a promising trade-off between parameter efficiency and performance effectiveness. Our code and models will be released.","{'model': 'tldr@v2.0.0', 'text': 'A novel Stage-wise Prompt-Matched Framework for generic and effective visual prompt tuning that divides the pre-trained backbone with frozen parameters into multiple stages and performs prompt learning between different stages, which makes the proposed scheme applicable to various architectures of CNN and Transformer.'}",https://arxiv.org/pdf/2208.10159
-multitask vision-language prompt tuning,Sheng Shen,"Prompt Tuning, conditioning on task-specific learned prompt vectors, has emerged as a data-efficient and parameter-efficient method for adapting large pretrained vision-language models to multiple downstream tasks. However, existing approaches usually consider learning prompt vectors for each task independently from scratch, thereby failing to exploit the rich shareable knowledge across different vision-language tasks. In this paper, we propose multitask vision-language prompt tuning (MVLPT), which incorporates cross-task knowledge into prompt tuning for vision-language models. Specifically, (i) we demonstrate the effectiveness of learning a single transferable prompt from multiple source tasks to initialize the prompt for each target task; (ii) we show many target tasks can benefit each other from sharing prompt vectors and thus can be jointly learned via multitask prompt tuning. We benchmark the proposed MVLPT using three representative prompt tuning methods, namely text prompt tuning, visual prompt tuning, and the unified vision-language prompt tuning. Results in 20 vision tasks demonstrate that the proposed approach outperforms all single-task baseline prompt tuning methods, setting the new state-of-the-art on the few-shot ELEVATER benchmarks and cross-task generalization benchmarks. To understand where the cross-task knowledge is most effective, we also conduct a large-scale study on task transferability with 20 vision tasks in 400 combinations for each prompt tuning method. It shows that the most performant MVLPT for each prompt tuning method prefers different task combinations and many tasks can benefit each other, depending on their visual similarity and label similarity. Code is available at https://github.com/sIncerass/MVLPT.","{'model': 'tldr@v2.0.0', 'text': 'This paper demonstrates the effectiveness of learning a single transferable prompt from multiple source tasks to initialize the prompt for each target task and shows many target tasks can benefit each other from sharing prompt vectors and thus can be jointly learned via multitask prompt tuning.'}",https://arxiv.org/pdf/2211.11720
-memory-assisted prompt editing to improve gpt-3 after deployment,Aman Madaan,"Large LMs such as GPT -3 are powerful, but can commit mistakes that are obvious to humans. For example, GPT -3 would mistakenly interpret ""What word is similar to good ?"" to mean a homonym, while the user intended a synonym. Our goal is to effectively correct such errors via user interactions with the sys-tem but without retraining, which will be pro-hibitively costly. We pair GPT -3 with a growing memory of recorded cases where the model misunderstood the user’s intents, along with user feedback for clarification. Such a memory allows our system to produce enhanced prompts for any new query based on the user feedback for error correction on similar cases in the past. On four tasks (two lexical tasks, two advanced ethical reasoning tasks), we show how a (simulated) user can interactively teach a deployed GPT -3, substantially increasing its accuracy over the queries with different kinds of misunderstandings by the GPT -3. Our approach is a step towards the low-cost utility enhancement for very large pre-trained LMs. 1","{'model': 'tldr@v2.0.0', 'text': 'This work pair GPT -3 with a growing memory of recorded cases where the model misunderstood the user’s intents, along with user feedback for clarification, which allows the system to produce enhanced prompts for any new query based on the user feedback on similar cases in the past.'}",https://aclanthology.org/2022.emnlp-main.183.pdf
-openprompt: an open-source framework for prompt-learning,Ning Ding,"Prompt-learning has become a new paradigm in modern natural language processing, which directly adapts pre-trained language models (PLMs) to cloze-style prediction, autoregressive modeling, or sequence to sequence generation, resulting in promising performances on various tasks. However, no standard implementation framework of prompt-learning is proposed yet, and most existing prompt- learning codebases, often unregulated, only provide limited implementations for specific scenarios. Since there are many details such as templating strategy, initializing strategy, verbalizing strategy, etc., that need to be considered in prompt-learning, practitioners face impediments to quickly adapting the de-sired prompt learning methods to their applications. In this paper, we present Open- Prompt, a unified easy-to-use toolkit to conduct prompt-learning over PLMs. OpenPrompt is a research-friendly framework that is equipped with efficiency, modularity, and extendibility, and its combinability allows the freedom to combine different PLMs, task for- mats, and prompting modules in a unified paradigm. Users could expediently deploy prompt-learning frameworks and evaluate the generalization of them on different NLP tasks without constraints.","{'model': 'tldr@v2.0.0', 'text': 'Open- Prompt is a unified easy-to-use toolkit to conduct prompt-learning over PLMs equipped with efficiency, modularity, and extendibility, and its combinability allows the freedom to combine different PLMs, task for- mats, and prompting modules in a unified paradigm.'}",https://aclanthology.org/2022.acl-demo.10.pdf
-cpt: colorful prompt tuning for pre-trained vision-language models,Yuan Yao,"Pre-Trained Vision-Language Models (VL-PTMs) have shown promising capabilities in grounding natural language in image data, facilitating a broad variety of cross-modal tasks. However, we note that there exists a significant gap between the objective forms of model pre-training and fine-tuning, resulting in a need for large amounts of labeled data to stimulate the visual grounding capability of VL-PTMs for downstream tasks. To address the challenge, we present Cross-modal Prompt Tuning (CPT, alternatively, Colorful Prompt Tuning), a novel paradigm for tuning VL-PTMs, which reformulates visual grounding into a fill-in-the-blank problem with color-based co-referential markers in image and text, maximally mitigating the gap. In this way, CPT enables strong few-shot and even zero-shot visual grounding capabilities of VL-PTMs. Comprehensive experimental results show that the prompt-tuned VL-PTMs outperform their fine-tuned counterparts by a large margin (e.g., 17.3% absolute accuracy improvement, and 73.8% relative standard deviation reduction on average with one shot in RefCOCO evaluation). We make the data and code for this paper publicly available at https://github.com/thunlp/CPT.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Cross-modal Prompt Tuning (CPT), a novel paradigm for tuning VL-PTMs, which reformulates visual grounding into a fill-in-the-blank problem with color-based co-referential markers in image and text, maximally mitigating the gap between model pre-training and fine-tuning.'}",
-differentiable prompt makes pre-trained language models better few-shot learners,Ningyu Zhang,"Large-scale pre-trained language models have contributed significantly to natural language processing by demonstrating remarkable abilities as few-shot learners. However, their effectiveness depends mainly on scaling the model parameters and prompt design, hindering their implementation in most real-world applications. This study proposes a novel pluggable, extensible, and efficient approach named DifferentiAble pRompT (DART), which can convert small language models into better few-shot learners without any prompt engineering. The main principle behind this approach involves reformulating potential natural language processing tasks into the task of a pre-trained language model and differentially optimizing the prompt template as well as the target label with backpropagation. Furthermore, the proposed approach can be: (i) Plugged to any pre-trained language models; (ii) Extended to widespread classification tasks. A comprehensive evaluation of standard NLP tasks demonstrates that the proposed approach achieves a better few-shot performance. Code is available in https://github.com/zjunlp/DART.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a novel pluggable, extensible, and efficient approach named DifferentiAble pRompT (DART), which can convert small language models into better few-shot learners without any prompt engineering.'}",
-adapting language models for zero-shot learning by meta-tuning on dataset and prompt collections,Ruiqi Zhong,"Large pre-trained language models (LMs) such as GPT-3 have acquired a surprising ability to perform zero-shot learning. For example, to classify sentiment without any training examples, we can""prompt""the LM with the review and the label description""Does the user like this movie?"", and ask whether the next word is""yes""or""no"". However, the next word prediction training objective is still misaligned with the target zero-shot learning objective. To address this weakness, we propose meta-tuning, which directly optimizes the zero-shot learning objective by fine-tuning pre-trained language models on a collection of datasets. We focus on classification tasks, and construct the meta-dataset by aggregating 43 existing datasets and annotating 441 label descriptions in a question-answering (QA) format. When evaluated on unseen tasks, meta-tuned models outperform a same-sized QA model and the previous SOTA zero-shot learning system based on natural language inference. Additionally, increasing parameter count from 220M to 770M improves AUC-ROC scores by 6.3%, and we forecast that even larger models would perform better. Therefore, measuring zero-shot learning performance on language models out-of-the-box might underestimate their true potential, and community-wide efforts on aggregating datasets and unifying their formats can help build models that answer prompts better.","{'model': 'tldr@v2.0.0', 'text': 'Meta-tuned models outperform a same-sized QA model and the previous SOTA zero-shot learning system based on natural language inference on unseen tasks, and community-wide efforts on aggregating datasets and unifying their formats can help build models that answer prompts better.'}",https://aclanthology.org/2021.findings-emnlp.244.pdf
-align and prompt: video-and-language pre-training with entity prompts,Dongxu Li,"Yidco-and-language pre-training has shown promising improvements on various downstream tasks. Most previous methods capture cross-modal interactions with a standard transformer-based multimodal encoder, not fully addressing the misalignment between unimodal video and text features. Besides, learning finegrained visual-language alignment usually requires off-the-shelf object detectors to provide object information, which is bottlenecked by the detector's limited vocabulary and expensive computation cost. In this paper, we propose Align and Prompt: a new video-and-language pre-training framework (AlPro), which operates on sparsely-sampled video frames and achieves more effective cross-modal alignment without explicit object detectors. First, we introduce a video-text contrastive (VTC) loss to align unimodal video-text features at the instance level, which eases the modeling of cross-modal interactions. Then, we propose a novel visually-grounded pre-training task, prompting entity modeling (PEM), which learns finegrained alignment between visual region and text entity via an entity prompter module in a self-supervised way. Finally, we pretrain the video-and-language transformer models on large webly-source video-text pairs using the proposed VTC and PEM losses as well as two standard losses of masked language modeling (MLM) and video-text matching (VTM). The resulting pre-trained model achieves state-of-the-art performance on both text-video retrieval and videoQA, outperforming prior work by a substantial margin. Implementation and pre-trained models are available at https://github.com/salesforce/ALPRO.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Align and Prompt: a new video-and-language pre-training framework (AlPro), which operates on sparsely-sampled video frames and achieves more effective cross-modal alignment without explicit object detectors.'}",https://arxiv.org/pdf/2112.09583
-prompt-learning for fine-grained entity typing,Ning Ding,"As an effective approach to tune pre-trained language models (PLMs) for specific tasks, prompt-learning has recently attracted much attention from researchers. By using \textit{cloze}-style language prompts to stimulate the versatile knowledge of PLMs, prompt-learning can achieve promising results on a series of NLP tasks, such as natural language inference, sentiment classification, and knowledge probing. In this work, we investigate the application of prompt-learning on fine-grained entity typing in fully supervised, few-shot and zero-shot scenarios. We first develop a simple and effective prompt-learning pipeline by constructing entity-oriented verbalizers and templates and conducting masked language modeling. Further, to tackle the zero-shot regime, we propose a self-supervised strategy that carries out distribution-level optimization in prompt-learning to automatically summarize the information of entity types. Extensive experiments on three fine-grained entity typing benchmarks (with up to 86 classes) under fully supervised, few-shot and zero-shot settings show that prompt-learning methods significantly outperform fine-tuning baselines, especially when the training data is insufficient.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a simple and effective prompt-learning pipeline, and proposes a self-supervised strategy that carries out distribution-level optimization in prompt- learning to automatically summarize the information of entity types in the zero-shot regime.'}",https://aclanthology.org/2022.findings-emnlp.512.pdf
-a good prompt is worth millions of parameters: low-resource prompt-based learning for vision-language models,Woojeong Jin,"Large pre-trained vision-language (VL) models can learn a new task with a handful of examples and generalize to a new task without fine-tuning.However, these VL models are hard to deploy for real-world applications due to their impractically huge sizes and slow inference speed.To solve this limitation, we study prompt-based low-resource learning of VL tasks with our proposed method, FewVLM, relatively smaller than recent few-shot learners.For FewVLM, we pre-train a sequence-to-sequence transformer model with prefix language modeling (PrefixLM) and masked language modeling (MaskedLM).Furthermore, we analyze the effect of diverse prompts for few-shot tasks.Experimental results on VQA show that FewVLM with prompt-based learning outperforms Frozen which is 31x larger than FewVLM by 18.2% point and achieves comparable results to a 246x larger model, PICa.In our analysis, we observe that (1) prompts significantly affect zero-shot performance but marginally affect few-shot performance, (2) models with noisy prompts learn as quickly as hand-crafted prompts given larger training data, and (3) MaskedLM helps VQA tasks while PrefixLM boosts captioning performance. Our code is publicly available at https://github.com/woojeongjin/FewVLM","{'model': 'tldr@v2.0.0', 'text': 'This work studies prompt-based low-resource learning of VL tasks with a sequence-to-sequence transformer model with prefix language modeling and masked language modeling, and observes that models with noisy prompts learn as quickly as hand-crafted prompts given larger training data.'}",https://aclanthology.org/2022.acl-long.197.pdf
-why do pretrained language models help in downstream tasks? an analysis of head and prompt tuning,Colin Wei,"Pretrained language models have achieved state-of-the-art performance when adapted to a downstream NLP task. However, theoretical analysis of these models is scarce and challenging since the pretraining and downstream tasks can be very different. We propose an analysis framework that links the pretraining and downstream tasks with an underlying latent variable generative model of text -- the downstream classifier must recover a function of the posterior distribution over the latent variables. We analyze head tuning (learning a classifier on top of the frozen pretrained model) and prompt tuning in this setting. The generative model in our analysis is either a Hidden Markov Model (HMM) or an HMM augmented with a latent memory component, motivated by long-term dependencies in natural language. We show that 1) under certain non-degeneracy conditions on the HMM, simple classification heads can solve the downstream task, 2) prompt tuning obtains downstream guarantees with weaker non-degeneracy conditions, and 3) our recovery guarantees for the memory-augmented HMM are stronger than for the vanilla HMM because task-relevant information is easier to recover from the long-term memory. Experiments on synthetically generated data from HMMs back our theoretical findings.","{'model': 'tldr@v2.0.0', 'text': 'An analysis framework is proposed that links the pretraining and downstream tasks with an underlying latent variable generative model of text -- the downstream classifier must recover a function of the posterior distribution over the latent variables.'}",
-on transferability of prompt tuning for natural language processing,Yusheng Su,"Prompt tuning (PT) is a promising parameter-efficient method to utilize extremely large pre-trained language models (PLMs), which can achieve comparable performance to full-parameter fine-tuning by only tuning a few soft prompts. However, PT requires much more training time than fine-tuning. Intuitively, knowledge transfer can help to improve the efficiency. To explore whether we can improve PT via prompt transfer, we empirically investigate the transferability of soft prompts across different downstream tasks and PLMs in this work. We find that (1) in zero-shot setting, trained soft prompts can effectively transfer to similar tasks on the same PLM and also to other PLMs with a cross-model projector trained on similar tasks; (2) when used as initialization, trained soft prompts of similar tasks and projected prompts of other PLMs can significantly accelerate training and also improve the performance of PT. Moreover, to explore what decides prompt transferability, we investigate various transferability indicators and find that the overlapping rate of activated neurons strongly reflects the transferability, which suggests how the prompts stimulate PLMs is essential. Our findings show that prompt transfer is promising for improving PT, and further research shall focus more on prompts’ stimulation to PLMs. The source code can be obtained from https://github.com/thunlp/Prompt-Transferability.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the overlapping rate of activated neurons strongly reflects the transferability, which suggests how the prompts stimulate PLMs is essential, and that prompt transfer is promising for improving PT.'}",https://aclanthology.org/2022.naacl-main.290.pdf
-pada: example-based prompt learning for on-the-fly adaptation to unseen domains,Eyal Ben-David,"Natural Language Processing algorithms have made incredible progress, but they still struggle when applied to out-of-distribution examples. We address a challenging and underexplored version of this domain adaptation problem, where an algorithm is trained on several source domains, and then applied to examples from unseen domains that are unknown at training time. Particularly, no examples, labeled or unlabeled, or any other knowledge about the target domain are available to the algorithm at training time. We present PADA: An example-based autoregressive Prompt learning algorithm for on-the-fly Any-Domain Adaptation, based on the T5 language model. Given a test example, PADA first generates a unique prompt for it and then, conditioned on this prompt, labels the example with respect to the NLP prediction task. PADA is trained to generate a prompt that is a token sequence of unrestricted length, consisting of Domain Related Features (DRFs) that characterize each of the source domains. Intuitively, the generated prompt is a unique signature that maps the test example to a semantic space spanned by the source domains. In experiments with 3 tasks (text classification and sequence tagging), for a total of 14 multi-source adaptation scenarios, PADA substantially outperforms strong baselines.1","{'model': 'tldr@v2.0.0', 'text': 'This work presents PADA: An example-based autoregressive Prompt learning algorithm for on-the-fly Any-Domain Adaptation, based on the T5 language model, which substantially outperforms strong baselines.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00468/2008061/tacl_a_00468.pdf
-nsp-bert: a prompt-based few-shot learner through an original pre-training task —— next sentence prediction,Yi Sun,"Using prompts to utilize language models to perform various downstream tasks, also known as prompt-based learning or prompt-learning, has lately gained significant success in comparison to the pre-train and fine-tune paradigm. Nonetheless, virtually most prompt-based methods are token-level such as PET based on mask language model (MLM). In this paper, we attempt to accomplish several NLP tasks in the zero-shot and few-shot scenarios using a BERT original pre-training task abandoned by RoBERTa and other models——Next Sentence Prediction (NSP). Unlike token-level techniques, our sentence-level prompt-based method NSP-BERT does not need to fix the length of the prompt or the position to be predicted, allowing it to handle tasks such as entity linking with ease. NSP-BERT can be applied to a variety of tasks based on its properties. We present an NSP-tuning approach with binary cross-entropy loss for single-sentence classification tasks that is competitive compared to PET and EFL. By continuing to train BERT on RoBERTa’s corpus, the model’s performance improved significantly, which indicates that the pre-training corpus is another important determinant of few-shot besides model size and prompt method.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents an NSP-tuning approach with binary cross-entropy loss for single-sentence classification tasks that is competitive compared to PET and EFL and indicates that the pre-training corpus is another important determinant of few-shot besides model size and prompt method.'}",
-lightner: a lightweight generative framework with prompt-guided attention for low-resource ner,Xiang Chen,"NER in low-resource domains suffers from insufficient training data. Existing transfer learning approaches for lowresource NER usually have the challenge that the target domain has different sets of entity categories compared with a resource-rich source domain, which can be concluded as class transfer and domain transfer problems. In this paper, we propose a lightweight generative framework with prompt-guided attention for low-resource NER (LightNER). Concretely, instead of tackling the problem by training label-specific discriminative classifiers, we convert sequence labeling to generate the entity pointer index sequence and entity categories without any label-specific classifiers, which can address the class transfer issue. We further propose prompt-guided attention by incorporating continuous prompts into the selfattention layer to re-modulate the attention and adapt pretrained weights. Note that we only tune those continuous prompts with the whole parameter of the pre-trained language model fixed, thus, making our approach lightweight and flexible for low-resource scenarios and can better transfer knowledge across domains. Experimental results show that by tuning only 0.16% of the parameters, LightNER can obtain comparable performance in the standard setting and outperform baselines in low-resource settings.","{'model': 'tldr@v2.0.0', 'text': 'A lightweight generative framework with prompt-guided attention for low-resource NER (LightNER), which converts sequence labeling to generate the entity pointer index sequence and entity categories without any label-specific classifiers, which can address the class transfer issue.'}",
-pada: a prompt-based autoregressive approach for adaptation to unseen domains,Eyal Ben-David,"Natural Language Processing algorithms have made incredible progress recently, but they still struggle when applied to out-of-distribution examples. In this paper, we address a very challenging and previously underexplored version of this domain adaptation problem. In our setup an algorithm is trained on several source domains, and then applied to examples from an unseen domain that is unknown at training time. Particularly, no examples, labeled or unlabeled, or any other knowledge about the target domain are available to the algorithm at training time. We present PADA : A Prompt-based Autoregressive Domain Adaptation algorithm, based on the T5 model. Given a test example, PADA first generates a unique prompt and then, conditioned on this prompt, labels the example with respect to the NLP task. The prompt is a sequence of unrestricted length, consisting of pre-defined Domain Related Features (DRFs) that characterize each of the source domains. Intuitively, the prompt is a unique signature that maps the test example to the semantic space spanned by the source domains. In experiments with two tasks: Rumour Detection and Multi-Genre Natural Language Inference (MNLI), for a total of 10 multi-source adaptation scenarios, PADA strongly outperforms state-of-the-art approaches and additional strong baselines. 1","{'model': 'tldr@v2.0.0', 'text': 'This paper presents PADA : A Prompt-based Autoregressive Domain Adaptation algorithm, based on the T5 model, which strongly outperforms state-of-the-art approaches and additional strong baselines in multi-source adaptation scenarios.'}",
-make-an-audio: text-to-audio generation with prompt-enhanced diffusion models,Rongjie Huang,"Large-scale multimodal generative modeling has created milestones in text-to-image and text-to-video generation. Its application to audio still lags behind for two main reasons: the lack of large-scale datasets with high-quality text-audio pairs, and the complexity of modeling long continuous audio data. In this work, we propose Make-An-Audio with a prompt-enhanced diffusion model that addresses these gaps by 1) introducing pseudo prompt enhancement with a distill-then-reprogram approach, it alleviates data scarcity with orders of magnitude concept compositions by using language-free audios; 2) leveraging spectrogram autoencoder to predict the self-supervised audio representation instead of waveforms. Together with robust contrastive language-audio pretraining (CLAP) representations, Make-An-Audio achieves state-of-the-art results in both objective and subjective benchmark evaluation. Moreover, we present its controllability and generalization for X-to-Audio with""No Modality Left Behind"", for the first time unlocking the ability to generate high-definition, high-fidelity audios given a user-defined modality input. Audio samples are available at https://Text-to-Audio.github.io","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Make-An- audio with a prompt-enhanced diffusion model that alleviates data scarcity with orders of magnitude concept compositions by using language-free audios, and presents its controllability and generalization for X-to-Audio with ""No Modality Left Behind"", for the first time unlocking the ability to generate high-definition, high-fidelity audios given a user-defined modality input.'}",http://arxiv.org/pdf/2301.12661
-why johnny can’t prompt: how non-ai experts try (and fail) to design llm prompts,J. Zamfirescu-Pereira,"Pre-trained large language models (“LLMs”) like GPT-3 can engage in fluent, multi-turn instruction-taking out-of-the-box, making them attractive materials for designing natural language interactions. Using natural language to steer LLM outputs (“prompting”) has emerged as an important design technique potentially accessible to non-AI-experts. Crafting effective prompts can be challenging, however, and prompt-based interactions are brittle. Here, we explore whether non-AI-experts can successfully engage in “end-user prompt engineering” using a design probe—a prototype LLM-based chatbot design tool supporting development and systematic evaluation of prompting strategies. Ultimately, our probe participants explored prompt designs opportunistically, not systematically, and struggled in ways echoing end-user programming systems and interactive machine learning systems. Expectations stemming from human-to-human instructional experiences, and a tendency to overgeneralize, were barriers to effective prompt design. These findings have implications for non-AI-expert-facing LLM-based tool design and for improving LLM-and-prompt literacy among programmers and the public, and present opportunities for further research.","{'model': 'tldr@v2.0.0', 'text': 'This work explores whether non-AI-experts can successfully engage in “end-user prompt engineering” using a design probe—a prototype LLM-based chatbot design tool supporting development and systematic evaluation of prompting strategies.'}",https://dl.acm.org/doi/pdf/10.1145/3544548.3581388
-the power of prompt tuning for low-resource semantic parsing,Nathan Schucher,"Prompt tuning has recently emerged as an effective method for adapting pre-trained language models to a number of language understanding and generation tasks. In this paper, we investigate prompt tuning for semantic parsing—the task of mapping natural language utterances onto formal meaning representations. On the low-resource splits of Overnight and TOPv2, we find that a prompt tuned T5-xl significantly outperforms its fine-tuned counterpart, as well as strong GPT-3 and BART baselines. We also conduct ablation studies across different model scales and target representations, finding that, with increasing model scale, prompt tuned T5 models improve at generating target representations that are far from the pre-training distribution.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates prompt tuning for semantic parsing—the task of mapping natural language utterances onto formal meaning representations, and finds that a prompt tuned T5-xl significantly outperforms its fine-tuned counterpart, as well as strong GPT-3 and BART baselines.'}",https://aclanthology.org/2022.acl-short.17.pdf
-the biases of pre-trained language models: an empirical study on prompt-based sentiment analysis and emotion detection,Rui Mao,"Thanks to the breakthrough of large-scale pre-trained language model (PLM) technology, prompt-based classification tasks, e.g., sentiment analysis and emotion detection, have raised increasing attention. Such tasks are formalized as masked language prediction tasks which are in line with the pre-training objects of most language models. Thus, one can use a PLM to infer the masked words in a downstream task, then obtaining label predictions with manually defined label-word mapping templates. Prompt-based affective computing takes the advantages of both neural network modeling and explainable symbolic representations. However, there still remain many unclear issues related to the mechanisms of PLMs and prompt-based classification. We conduct a systematic empirical study on prompt-based sentiment analysis and emotion detection to study the biases of PLMs towards affective computing. We find that PLMs are biased in sentiment analysis and emotion detection tasks with respect to the number of label classes, emotional label-word selections, prompt templates and positions, and the word forms of emotion lexicons.","{'model': 'tldr@v2.0.0', 'text': 'It is found that PLMs are biased in sentiment analysis and emotion detection tasks with respect to the number of label classes, emotional label-word selections, prompt templates and positions, and the word forms of emotion lexicons.'}",
-adaprompt: adaptive prompt-based finetuning for relation extraction,Xiang Chen,"In this paper, we reformulate the relation extraction task as mask language modeling and propose a novel adaptive prompt-based finetuning approach. We propose an adaptive label words selection mechanism that scatters the relation label into variable number of label tokens to handle the complex multiple label space. We further introduce an auxiliary entity discriminator object to encourage the model to focus on context representation learning. Extensive experiments on benchmark datasets demonstrate that our approach can achieve better performance on both the few-shot and supervised setting1.","{'model': 'tldr@v2.0.0', 'text': 'An adaptive label words selection mechanism that scatters the relation label into variable number of label tokens to handle the complex multiple label space and introduces an auxiliary entity discriminator object to encourage the model to focus on context representation learning.'}",
-prompt waywardness: the curious case of discretized interpretation of continuous prompts,Daniel Khashabi,"Fine-tuning continuous prompts for target tasks has recently emerged as a compact alternative to full model fine-tuning. Motivated by these promising results, we investigate the feasibility of extracting a discrete (textual) interpretation of continuous prompts that is faithful to the problem they solve. In practice, we observe a “wayward” behavior between the task solved by continuous prompts and their nearest neighbor discrete projections: We can find continuous prompts that solve a task while being projected to an arbitrary text (e.g., definition of a different or even a contradictory task), while being within a very small (2%) margin of the best continuous prompt of the same size for the task. We provide intuitions behind this odd and surprising behavior, as well as extensive empirical analyses quantifying the effect of various parameters. For instance, for larger model sizes we observe higher waywardness, i.e, we can find prompts that more closely map to any arbitrary text with a smaller drop in accuracy. These findings have important implications relating to the difficulty of faithfully interpreting continuous prompts and their generalization across models and tasks, providing guidance for future progress in prompting language models.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the feasibility of extracting a discrete (textual) interpretation of continuous prompts that is faithful to the problem they solve, and observes a “wayward” behavior between the task solved by continuous prompts and their nearest neighbor discrete projections.'}",https://aclanthology.org/2022.naacl-main.266.pdf
-sentiprompt: sentiment knowledge enhanced prompt-tuning for aspect-based sentiment analysis,Chengxi Li,"Aspect-based sentiment analysis (ABSA) is an emerging fine-grained sentiment analysis task that aims to extract aspects, classify corresponding sentiment polarities and find opinions as the causes of sentiment. The latest research tends to solve the ABSA task in a unified way with end-to-end frameworks. Yet, these frameworks get fine-tuned from downstream tasks without any task-adaptive modification. Specifically, they do not use task-related knowledge well or explicitly model relations between aspect and opinion terms, hindering them from better performance. In this paper, we propose SentiPrompt to use sentiment knowledge enhanced prompts to tune the language model in the unified framework. We inject sentiment knowledge regarding aspects, opinions, and polarities into prompt and explicitly model term relations via constructing consistency and polarity judgment templates from the ground truth triplets. Experimental results demonstrate that our approach can outperform strong baselines on Triplet Extraction, Pair Extraction, and Aspect Term Extraction with Sentiment Classification by a notable margin.","{'model': 'tldr@v2.0.0', 'text': 'SentiPrompt is proposed to use sentiment knowledge enhanced prompts to tune the language model in the unified framework and inject sentiment knowledge regarding aspects, opinions, and polarities into prompt and explicitly model term relations via constructing consistency and polarity judgment templates from the ground truth triplets.'}",
-automated cross-prompt scoring of essay traits,Robert Ridley,"The majority of current research in Automated Essay Scoring (AES) focuses on prompt-specific scoring of either the overall quality of an essay or the quality with regards to certain traits. In real-world applications obtaining labelled data for a target essay prompt is often expensive or unfeasible, requiring the AES system to be able to perform well when predicting scores for essays from unseen prompts. As a result, some recent research has been dedicated to cross-prompt AES. However, this line of research has thus far only been concerned with holistic, overall scoring, with no exploration into the scoring of different traits. As users of AES systems often require feedback with regards to different aspects of their writing, trait scoring is a necessary component of an effective AES system. Therefore, to address this need, we introduce a new task named Automated Cross-prompt Scoring of Essay Traits, which requires the model to be trained solely on non-target-prompt essays and to predict the holistic, overall score as well as scores for a number of specific traits for target-prompt essays. This task challenges the model's ability to generalize in order to score essays from a novel domain as well as its ability to represent the quality of essays from multiple different aspects. In addition, we introduce a new, innovative approach which builds on top of a state-of-the-art method for cross-prompt AES. Our method utilizes a trait-attention mechanism and a multi-task architecture that leverages the relationships between each trait to simultaneously predict the overall score and the score of each individual trait. We conduct extensive experiments on the widely used ASAP and ASAP++ datasets and demonstrate that our approach is able to outperform leading prompt-specific trait scoring and cross-prompt AES methods.","{'model': 'tldr@v2.0.0', 'text': 'A new task named Automated Cross- Prompt Scoring of Essay Traits, which requires the model to be trained solely on non-target-prompt essays and to predict the holistic, overall score as well as scores for a number of specific traits for target-promPT essays, is introduced.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17620/17427
-"inclusive, prompt and non-prompt ${\rm j}/\psi$ production at midrapidity in p$-$pb collisions at $\sqrt{s_{\rm nn}} = 5.02$ tev",Alice Collaboration,,,https://link.springer.com/content/pdf/10.1007/JHEP06(2022)011.pdf
-masterkey: automated jailbreak across multiple large language model chatbots,Gelei Deng,"Large Language Models (LLMs) have revolutionized Artificial Intelligence (AI) services due to their exceptional proficiency in understanding and generating human-like text. LLM chatbots, in particular, have seen widespread adoption, transforming human-machine interactions. However, these LLM chatbots are susceptible to""jailbreak""attacks, where malicious users manipulate prompts to elicit inappropriate or sensitive responses, contravening service policies. Despite existing attempts to mitigate such threats, our research reveals a substantial gap in our understanding of these vulnerabilities, largely due to the undisclosed defensive measures implemented by LLM service providers. In this paper, we present Jailbreaker, a comprehensive framework that offers an in-depth understanding of jailbreak attacks and countermeasures. Our work makes a dual contribution. First, we propose an innovative methodology inspired by time-based SQL injection techniques to reverse-engineer the defensive strategies of prominent LLM chatbots, such as ChatGPT, Bard, and Bing Chat. This time-sensitive approach uncovers intricate details about these services' defenses, facilitating a proof-of-concept attack that successfully bypasses their mechanisms. Second, we introduce an automatic generation method for jailbreak prompts. Leveraging a fine-tuned LLM, we validate the potential of automated jailbreak generation across various commercial LLM chatbots. Our method achieves a promising average success rate of 21.58%, significantly outperforming the effectiveness of existing techniques. We have responsibly disclosed our findings to the concerned service providers, underscoring the urgent need for more robust defenses. Jailbreaker thus marks a significant step towards understanding and mitigating jailbreak threats in the realm of LLM chatbots.","{'model': 'tldr@v2.0.0', 'text': 'Jailbreaker is presented, a comprehensive framework that offers an in-depth understanding of jailbreak attacks and countermeasures, and an automatic generation method for jailbreak prompts is introduced, leveraging a fine-tuned LLM to validate the potential of automated jailbreak generation across various commercial LLM chatbots.'}",
-visual adversarial examples jailbreak aligned large language models,Xiangyu Qi,"Recently, there has been a surge of interest in integrating vision into Large Language Models (LLMs), exemplified by Visual Language Models (VLMs) such as Flamingo and GPT-4. This paper sheds light on the security and safety implications of this trend. First, we underscore that the continuous and high-dimensional nature of the visual input makes it a weak link against adversarial attacks, representing an expanded attack surface of vision-integrated LLMs. Second, we highlight that the versatility of LLMs also presents visual attackers with a wider array of achievable adversarial objectives, extending the implications of security failures beyond mere misclassification. As an illustration, we present a case study in which we exploit visual adversarial examples to circumvent the safety guardrail of aligned LLMs with integrated vision. Intriguingly, we discover that a single visual adversarial example can universally jailbreak an aligned LLM, compelling it to heed a wide range of harmful instructions that it otherwise would not) and generate harmful content that transcends the narrow scope of a `few-shot' derogatory corpus initially employed to optimize the adversarial example. Our study underscores the escalating adversarial risks associated with the pursuit of multimodality. Our findings also connect the long-studied adversarial vulnerabilities of neural networks to the nascent field of AI alignment. The presented attack suggests a fundamental adversarial challenge for AI alignment, especially in light of the emerging trend toward multimodality in frontier foundation models.","{'model': 'tldr@v2.0.0', 'text': 'A case study in which a single visual adversarial example can universally jailbreak an aligned LLM, compelling it to heed a wide range of harmful instructions that it otherwise would not and suggest a fundamental adversarial challenge for AI alignment, especially in light of the emerging trend toward multimodality in frontier foundation models.'}",
-gptfuzzer : red teaming large language models with auto-generated jailbreak prompts,Jiahao Yu,"Large language models (LLMs) have recently experienced tremendous popularity and are widely used from casual conversations to AI-driven programming. However, despite their considerable success, LLMs are not entirely reliable and can give detailed guidance on how to conduct harmful or illegal activities. While safety measures can reduce the risk of such outputs, adversarial jailbreak attacks can still exploit LLMs to produce harmful content. These jailbreak templates are typically manually crafted, making large-scale testing challenging. In this paper, we introduce GPTFuzz, a novel black-box jailbreak fuzzing framework inspired by the AFL fuzzing framework. Instead of manual engineering, GPTFuzz automates the generation of jailbreak templates for red-teaming LLMs. At its core, GPTFuzz starts with human-written templates as initial seeds, then mutates them to produce new templates. We detail three key components of GPTFuzz: a seed selection strategy for balancing efficiency and variability, mutate operators for creating semantically equivalent or similar sentences, and a judgment model to assess the success of a jailbreak attack. We evaluate GPTFuzz against various commercial and open-source LLMs, including ChatGPT, LLaMa-2, and Vicuna, under diverse attack scenarios. Our results indicate that GPTFuzz consistently produces jailbreak templates with a high success rate, surpassing human-crafted templates. Remarkably, GPTFuzz achieves over 90% attack success rates against ChatGPT and Llama-2 models, even with suboptimal initial seed templates. We anticipate that GPTFuzz will be instrumental for researchers and practitioners in examining LLM robustness and will encourage further exploration into enhancing LLM safety.","{'model': 'tldr@v2.0.0', 'text': 'GPTFuzz is introduced, a novel black-box jailbreak fuzzing framework inspired by the AFL fuzzed framework that automates the generation of jailbreak templates for red-teaming LLMs and consistently produces jailbreaks with a high success rate, surpassing human-crafted templates.'}",https://arxiv.org/pdf/2309.10253
-autodan: generating stealthy jailbreak prompts on aligned large language models,Xiaogeng Liu,"The aligned Large Language Models (LLMs) are powerful language understanding and decision-making tools that are created through extensive alignment with human feedback. However, these large models remain susceptible to jailbreak attacks, where adversaries manipulate prompts to elicit malicious outputs that should not be given by aligned LLMs. Investigating jailbreak prompts can lead us to delve into the limitations of LLMs and further guide us to secure them. Unfortunately, existing jailbreak techniques suffer from either (1) scalability issues, where attacks heavily rely on manual crafting of prompts, or (2) stealthiness problems, as attacks depend on token-based algorithms to generate prompts that are often semantically meaningless, making them susceptible to detection through basic perplexity testing. In light of these challenges, we intend to answer this question: Can we develop an approach that can automatically generate stealthy jailbreak prompts? In this paper, we introduce AutoDAN, a novel jailbreak attack against aligned LLMs. AutoDAN can automatically generate stealthy jailbreak prompts by the carefully designed hierarchical genetic algorithm. Extensive evaluations demonstrate that AutoDAN not only automates the process while preserving semantic meaningfulness, but also demonstrates superior attack strength in cross-model transferability, and cross-sample universality compared with the baseline. Moreover, we also compare AutoDAN with perplexity-based defense methods and show that AutoDAN can bypass them effectively.","{'model': 'tldr@v2.0.0', 'text': 'Extensive evaluations demonstrate that AutoDAN not only automates the process while preserving semantic meaningfulness, but also demonstrates superior attack strength in cross-model transferability, and cross-sample universality compared with the baseline.'}",https://arxiv.org/pdf/2310.04451
-multilingual jailbreak challenges in large language models,Yue Deng,"While large language models (LLMs) exhibit remarkable capabilities across a wide range of tasks, they pose potential safety concerns, such as the ``jailbreak'' problem, wherein malicious instructions can manipulate LLMs to exhibit undesirable behavior. Although several preventive measures have been developed to mitigate the potential risks associated with LLMs, they have primarily focused on English data. In this study, we reveal the presence of multilingual jailbreak challenges within LLMs and consider two potential risk scenarios: unintentional and intentional. The unintentional scenario involves users querying LLMs using non-English prompts and inadvertently bypassing the safety mechanisms, while the intentional scenario concerns malicious users combining malicious instructions with multilingual prompts to deliberately attack LLMs. The experimental results reveal that in the unintentional scenario, the rate of unsafe content increases as the availability of languages decreases. Specifically, low-resource languages exhibit three times the likelihood of encountering harmful content compared to high-resource languages, with both ChatGPT and GPT-4. In the intentional scenario, multilingual prompts can exacerbate the negative impact of malicious instructions, with astonishingly high rates of unsafe output: 80.92\% for ChatGPT and 40.71\% for GPT-4. To handle such a challenge in the multilingual context, we propose a novel \textsc{Self-Defense} framework that automatically generates multilingual training data for safety fine-tuning. Experimental results show that ChatGPT fine-tuned with such data can achieve a substantial reduction in unsafe content generation. Data is available at https://github.com/DAMO-NLP-SG/multilingual-safety-for-LLMs. Warning: This paper contains examples with potentially harmful content.","{'model': 'tldr@v2.0.0', 'text': 'A novel \\textsc{Self-Defense} framework that automatically generates multilingual training data for safety fine-tuning is proposed that shows that ChatGPT fine- Tuned with such data can achieve a substantial reduction in unsafe content generation.'}",https://arxiv.org/pdf/2310.06474
-developing an accuracy-prompt toolkit to reduce covid-19 misinformation online,Ziv Epstein,"Recent research suggests that shifting users’ attention to accuracy increases the quality of news they subsequently share online. Here we help develop this initial observation into a suite of deploy-able interventions for practitioners. We ask (i) how prior results generalize to other approaches for prompting users to consider accuracy, and (ii) for whom these prompts are more versus less effec-tive. In a large survey experiment examining participants’ intentions to share true and false head-lines about COVID-19, we identify a variety of different accuracy prompts that su¬ccessfully increase sharing","{'model': 'tldr@v2.0.0', 'text': 'In a large survey experiment examining participants’ intentions to share true and false head-lines about COVID-19, a variety of different accuracy prompts are identified that Successfully increase sharing.'}",https://misinforeview.hks.harvard.edu/wp-content/uploads/2021/05/epstein_toolkit_covid_19_misinformation_20210518.pdf
-not what you've signed up for: compromising real-world llm-integrated applications with indirect prompt injection,Kai Greshake,"Large Language Models (LLMs) are increasingly being integrated into various applications. The functionalities of recent LLMs can be flexibly modulated via natural language prompts. This renders them susceptible to targeted adversarial prompting, e.g., Prompt Injection (PI) attacks enable attackers to override original instructions and employed controls. So far, it was assumed that the user is directly prompting the LLM. But, what if it is not the user prompting? We argue that LLM-Integrated Applications blur the line between data and instructions. We reveal new attack vectors, using Indirect Prompt Injection, that enable adversaries to remotely (without a direct interface) exploit LLM-integrated applications by strategically injecting prompts into data likely to be retrieved. We derive a comprehensive taxonomy from a computer security perspective to systematically investigate impacts and vulnerabilities, including data theft, worming, information ecosystem contamination, and other novel security risks. We demonstrate our attacks' practical viability against both real-world systems, such as Bing's GPT-4 powered Chat and code-completion engines, and synthetic applications built on GPT-4. We show how processing retrieved prompts can act as arbitrary code execution, manipulate the application's functionality, and control how and if other APIs are called. Despite the increasing integration and reliance on LLMs, effective mitigations of these emerging threats are currently lacking. By raising awareness of these vulnerabilities and providing key insights into their implications, we aim to promote the safe and responsible deployment of these powerful models and the development of robust defenses that protect users and systems from potential attacks.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that LLM-Integrated Applications blur the line between data and instructions, and it is shown how processing retrieved prompts can act as arbitrary code execution, manipulate the application's functionality, and control how and if other APIs are called.""}",
-hard prompts made easy: gradient-based discrete optimization for prompt tuning and discovery,Yuxin Wen,"The strength of modern generative models lies in their ability to be controlled through text-based prompts. Typical""hard""prompts are made from interpretable words and tokens, and must be hand-crafted by humans. There are also""soft""prompts, which consist of continuous feature vectors. These can be discovered using powerful optimization methods, but they cannot be easily interpreted, re-used across models, or plugged into a text-based interface. We describe an approach to robustly optimize hard text prompts through efficient gradient-based optimization. Our approach automatically generates hard text-based prompts for both text-to-image and text-to-text applications. In the text-to-image setting, the method creates hard prompts for diffusion models, allowing API users to easily generate, discover, and mix and match image concepts without prior knowledge on how to prompt the model. In the text-to-text setting, we show that hard prompts can be automatically discovered that are effective in tuning LMs for classification.","{'model': 'tldr@v2.0.0', 'text': 'This work describes an approach to robustly optimize hard text prompts through efficient gradient-based optimization and shows that hard prompts can be automatically discovered that are effective in tuning LMs for classification.'}",http://arxiv.org/pdf/2302.03668
-more than you've asked for: a comprehensive analysis of novel prompt injection threats to application-integrated large language models,Kai Greshake,"We are currently witnessing dramatic advances in the capabilities of Large Language Models (LLMs). They are already being adopted in practice and integrated into many systems, including integrated development environments (IDEs) and search engines. The functionalities of current LLMs can be modulated via natural language prompts, while their exact internal functionality remains implicit and unassessable. This property, which makes them adaptable to even unseen tasks, might also make them susceptible to targeted adversarial prompting . Recently, several ways to misalign LLMs using Prompt Injection (PI) attacks have been introduced. In such attacks, an adversary can prompt the LLM to produce malicious content or override the original instructions and the employed filtering schemes. Recent work showed that these attacks are hard to mitigate, as state-of-the-art LLMs are instruction-following . So far, these attacks assumed that the adversary is directly prompting the LLM. In this work, we show that augmenting LLMs with retrieval and API calling capabilities (so-called Application-Integrated LLMs ) induces a whole new set of attack vectors. These LLMs might process poisoned content retrieved from the Web that contains malicious prompts pre-injected and selected by adversaries. We demonstrate that an attacker can indirectly perform such PI attacks. Based on this key insight, we systematically analyze the resulting threat landscape of Application-Integrated LLMs and discuss a variety of new attack vectors. To demonstrate the practical viabil-ity of our attacks, we implemented specific demonstrations","{'model': 'tldr@v2.0.0', 'text': 'This work systematically analyze the resulting threat landscape of Application-Integrated LLMs and discusses a variety of new attack vectors, including poisoned content retrieved from the Web that contains malicious prompts pre-injected and selected by adversaries.'}",http://arxiv.org/pdf/2302.12173
-catastrophic jailbreak of open-source llms via exploiting generation,Yangsibo Huang,"The rapid progress in open-source large language models (LLMs) is significantly advancing AI development. Extensive efforts have been made before model release to align their behavior with human values, with the primary goal of ensuring their helpfulness and harmlessness. However, even carefully aligned models can be manipulated maliciously, leading to unintended behaviors, known as""jailbreaks"". These jailbreaks are typically triggered by specific text inputs, often referred to as adversarial prompts. In this work, we propose the generation exploitation attack, an extremely simple approach that disrupts model alignment by only manipulating variations of decoding methods. By exploiting different generation strategies, including varying decoding hyper-parameters and sampling methods, we increase the misalignment rate from 0% to more than 95% across 11 language models including LLaMA2, Vicuna, Falcon, and MPT families, outperforming state-of-the-art attacks with $30\times$ lower computational cost. Finally, we propose an effective alignment method that explores diverse generation strategies, which can reasonably reduce the misalignment rate under our attack. Altogether, our study underscores a major failure in current safety evaluation and alignment procedures for open-source LLMs, strongly advocating for more comprehensive red teaming and better alignment before releasing such models. Our code is available at https://github.com/Princeton-SysML/Jailbreak_LLM.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the generation exploitation attack, an extremely simple approach that disrupts model alignment by only manipulating variations of decoding methods, and proposes an effective alignment method that explores diverse generation strategies, which can reasonably reduce the misalignment rate under the attack.'}",https://arxiv.org/pdf/2310.06987
-visual adversarial examples jailbreak large language models,Xiangyu Qi,for,"{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2306.13213
-jailbreak and guard aligned language models with only few in-context demonstrations,Zeming Wei,"Large Language Models (LLMs) have shown remarkable success in various tasks, but concerns about their safety and the potential for generating malicious content have emerged. In this paper, we explore the power of In-Context Learning (ICL) in manipulating the alignment ability of LLMs. We find that by providing just few in-context demonstrations without fine-tuning, LLMs can be manipulated to increase or decrease the probability of jailbreaking, i.e. answering malicious prompts. Based on these observations, we propose In-Context Attack (ICA) and In-Context Defense (ICD) methods for jailbreaking and guarding aligned language model purposes. ICA crafts malicious contexts to guide models in generating harmful outputs, while ICD enhances model robustness by demonstrations of rejecting to answer harmful prompts. Our experiments show the effectiveness of ICA and ICD in increasing or reducing the success rate of adversarial jailbreaking attacks. Overall, we shed light on the potential of ICL to influence LLM behavior and provide a new perspective for enhancing the safety and alignment of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Light is shed on the potential of In-Context Learning (ICL) to influence LLM behavior and a new perspective for enhancing the safety and alignment of LLMs is provided.'}",https://arxiv.org/pdf/2310.06387
-equation of state constraints from the threshold binary mass for prompt collapse of neutron star mergers.,A. Bauswein,"Using hydrodynamical simulations for a large set of high-density matter equations of state (EOSs), we systematically determine the threshold mass M_{thres} for prompt black-hole formation in equal-mass and asymmetric neutron star (NS) mergers. We devise the so far most direct, general, and accurate method to determine the unknown maximum mass of nonrotating NSs from merger observations revealing M_{thres}. Considering hybrid EOSs with hadron-quark phase transition, we identify a new, observable signature of quark matter in NS mergers. Furthermore, our findings have direct applications in gravitational wave searches, kilonova interpretations, and multimessenger constraints on NS properties.","{'model': 'tldr@v2.0.0', 'text': 'Considering hybrid EOSs with hadron-quark phase transition, a new, observable signature of quark matter in NS mergers is identified, which has direct applications in gravitational wave searches, kilonova interpretations, and multimessenger constraints on NS properties.'}",https://authors.library.caltech.edu/104636/3/PhysRevLett.125.141103.pdf
-prompt as triggers for backdoor attack: examining the vulnerability in language models,Shuai Zhao,"The prompt-based learning paradigm, which bridges the gap between pre-training and fine-tuning, achieves state-of-the-art performance on several NLP tasks, particularly in few-shot settings. Despite being widely applied, prompt-based learning is vulnerable to backdoor attacks. Textual backdoor attacks are designed to introduce targeted vulnerabilities into models by poisoning a subset of training samples through trigger injection and label modification. However, they suffer from flaws such as abnormal natural language expressions resulting from the trigger and incorrect labeling of poisoned samples. In this study, we propose ProAttack, a novel and efficient method for performing clean-label backdoor attacks based on the prompt, which uses the prompt itself as a trigger. Our method does not require external triggers and ensures correct labeling of poisoned samples, improving the stealthy nature of the backdoor attack. With extensive experiments on rich-resource and few-shot text classification tasks, we empirically validate ProAttack's competitive performance in textual backdoor attacks. Notably, in the rich-resource setting, ProAttack achieves state-of-the-art attack success rates in the clean-label backdoor attack benchmark without external triggers.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes ProAttack, a novel and efficient method for performing clean-label backdoor attacks based on the prompt, which uses the prompt itself as a trigger, which does not require external triggers and ensures correct labeling of poisoned samples, improving the stealthy nature of the backdoor attack.'}",http://arxiv.org/pdf/2305.01219
-notable: transferable backdoor attacks against prompt-based nlp models,Kai Mei,"Prompt-based learning is vulnerable to backdoor attacks. Existing backdoor attacks against prompt-based models consider injecting backdoors into the entire embedding layers or word embedding vectors. Such attacks can be easily affected by retraining on downstream tasks and with different prompting strategies, limiting the transferability of backdoor attacks. In this work, we propose transferable backdoor attacks against prompt-based models, called NOTABLE, which is independent of downstream tasks and prompting strategies. Specifically, NOTABLE injects backdoors into the encoders of PLMs by utilizing an adaptive verbalizer to bind triggers to specific words (i.e., anchors). It activates the backdoor by pasting input with triggers to reach adversary-desired anchors, achieving independence from downstream tasks and prompting strategies. We conduct experiments on six NLP tasks, three popular models, and three prompting strategies. Empirical results show that NOTABLE achieves superior attack performance (i.e., attack success rate over 90% on all the datasets), and outperforms two state-of-the-art baselines. Evaluations on three defenses show the robustness of NOTABLE. Our code can be found at https://github.com/RU-System-Software-and-Security/Notable.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes transferable backdoor attacks against prompt-based models, called NOTABLE, which is independent of downstream tasks and prompting strategies, and achieves superior attack performance and outperforms two state-of-the-art baselines.'}",http://arxiv.org/pdf/2305.17826
-prompts should not be seen as secrets: systematically measuring prompt extraction attack success,Yiming Zhang,"The generations of large language models are commonly controlled through prompting techniques, where a user's query to the model is prefixed with a prompt that aims to guide the model's behaviour on the query. The prompts used by companies to guide their models are often treated as secrets, to be hidden from the user making the query. They have even been treated as commodities to be bought and sold. However, there has been anecdotal evidence showing that the prompts can be extracted by a user even when they are kept secret. In this paper, we present a framework for systematically measuring the success of prompt extraction attacks. In experiments with multiple sources of prompts and multiple underlying language models, we find that simple text-based attacks can in fact reveal prompts with high probability.","{'model': 'tldr@v2.0.0', 'text': 'In experiments with multiple sources of prompts and multiple underlying language models, it is found that simple text-based attacks can in fact reveal prompts with high probability.'}",https://arxiv.org/pdf/2307.06865
-sam on medical images: a comprehensive study on three prompt modes,D. Cheng,"The Segment Anything Model (SAM) made an eye-catching debut recently and inspired many researchers to explore its potential and limitation in terms of zero-shot generalization capability. As the first promptable foundation model for segmentation tasks, it was trained on a large dataset with an unprecedented number of images and annotations. This large-scale dataset and its promptable nature endow the model with strong zero-shot generalization. Although the SAM has shown competitive performance on several datasets, we still want to investigate its zero-shot generalization on medical images. As we know, the acquisition of medical image annotation usually requires a lot of effort from professional practitioners. Therefore, if there exists a foundation model that can give high-quality mask prediction simply based on a few point prompts, this model will undoubtedly become the game changer for medical image analysis. To evaluate whether SAM has the potential to become the foundation model for medical image segmentation tasks, we collected more than 12 public medical image datasets that cover various organs and modalities. We also explore what kind of prompt can lead to the best zero-shot performance with different modalities. Furthermore, we find that a pattern shows that the perturbation of the box size will significantly change the prediction accuracy. Finally, Extensive experiments show that the predicted mask quality varied a lot among different datasets. And providing proper prompts, such as bounding boxes, to the SAM will significantly increase its performance.","{'model': 'tldr@v2.0.0', 'text': 'To evaluate whether SAM has the potential to become the foundation model for medical image segmentation tasks, more than 12 public medical image datasets are collected and extensive experiments show that the predicted mask quality varied a lot among different datasets.'}",http://arxiv.org/pdf/2305.00035
-from prompt injections to sql injection attacks: how protected is your llm-integrated web application?,Rodrigo Pedro,"Large Language Models (LLMs) have found widespread applications in various domains, including web applications, where they facilitate human interaction via chatbots with natural language interfaces. Internally, aided by an LLM-integration middleware such as Langchain, user prompts are translated into SQL queries used by the LLM to provide meaningful responses to users. However, unsanitized user prompts can lead to SQL injection attacks, potentially compromising the security of the database. Despite the growing interest in prompt injection vulnerabilities targeting LLMs, the specific risks of generating SQL injection attacks through prompt injections have not been extensively studied. In this paper, we present a comprehensive examination of prompt-to-SQL (P$_2$SQL) injections targeting web applications based on the Langchain framework. Using Langchain as our case study, we characterize P$_2$SQL injections, exploring their variants and impact on application security through multiple concrete examples. Furthermore, we evaluate 7 state-of-the-art LLMs, demonstrating the pervasiveness of P$_2$SQL attacks across language models. Our findings indicate that LLM-integrated applications based on Langchain are highly susceptible to P$_2$SQL injection attacks, warranting the adoption of robust defenses. To counter these attacks, we propose four effective defense techniques that can be integrated as extensions to the Langchain framework. We validate the defenses through an experimental evaluation with a real-world use case application.","{'model': 'tldr@v2.0.0', 'text': 'It is indicated that LLM-integrated applications based on Langchain are highly susceptible to P$_2$SQL injection attacks, warranting the adoption of robust defenses, and four effective defense techniques that can be integrated as extensions to the Langchain framework are proposed.'}",https://arxiv.org/pdf/2308.01990
-integral detection of the first prompt gamma-ray signal coincident with the gravitational-wave event gw170817,V. Savchenko,"We report the INTernational Gamma-ray Astrophysics Laboratory (INTEGRAL) detection of the short gamma-ray burst GRB 170817A (discovered by Fermi-GBM) with a signal-to-noise ratio of 4.6, and, for the first time, its association with the gravitational waves (GWs) from binary neutron star (BNS) merging event GW170817 detected by the LIGO and Virgo observatories. The significance of association between the gamma-ray burst observed by INTEGRAL and GW170817 is 3.2σ, while the association between the Fermi-GBM and INTEGRAL detections is 4.2σ. GRB 170817A was detected by the SPI-ACS instrument about 2 s after the end of the GW event. We measure a fluence of (1.4 ± 0.4 ± 0.6) × 10−7 erg cm−2 (75–2000 keV), where, respectively, the statistical error is given at the 1σ confidence level, and the systematic error corresponds to the uncertainty in the spectral model and instrument response. We also report on the pointed follow-up observations carried out by INTEGRAL, starting 19.5 hr after the event, and lasting for 5.4 days. We provide a stringent upper limit on any electromagnetic signal in a very broad energy range, from 3 keV to 8 MeV, constraining the soft gamma-ray afterglow flux to <7.1 × 10−11 erg cm−2 s−1 (80–300 keV). Exploiting the unique capabilities of INTEGRAL, we constrained the gamma-ray line emission from radioactive decays that are expected to be the principal source of the energy behind a kilonova event following a BNS coalescence. Finally, we put a stringent upper limit on any delayed bursting activity, for example, from a newly formed magnetar.",,https://iopscience.iop.org/article/10.3847/2041-8213/aa8f94/pdf
-a general-relativistic determination of the threshold mass to prompt collapse in binary neutron star mergers,Sven Köppel,"We study the lifetimes of the remnant produced by the merger of two neutron stars and revisit the determination of the threshold mass to prompt collapse, Mth. Using a fully general-relativistic numerical approach and a novel method for a rigorous determination of Mth, we show that a nonlinear universal relation exists between the threshold mass and the maximum compactness. For the temperature-dependent equations of state considered here, our results improve a similar linear relation found recently with methods that are less accurate but yield quantitatively similar results. Furthermore, exploiting the information from GW170817, we use the universal relation to set lower limits on the stellar radii for any mass.",,https://iopscience.iop.org/article/10.3847/2041-8213/ab0210/pdf
-prompt rewetting of drained peatlands reduces climate warming despite methane emissions,A. Günther,,"{'model': 'tldr@v2.0.0', 'text': 'A radiative forcing model is used to compare forcing dynamics of global scenarios for future peatland management using areal data from the Global Peatland Database and shows that CH4 radiativeforcing does not undermine the climate change mitigation potential of peatlands rewetting.'}",https://www.nature.com/articles/s41467-020-15499-z.pdf
-prompt optical emission as a signature of synchrotron radiation in gamma-ray bursts,G. Oganesyan,"Information on the spectral shape of prompt emission in gamma-ray bursts (GRB) is mostly available only at energies ≳10 keV, where the main instruments for GRB detection are sensitive. The origin of this emission is still very uncertain because of the apparent inconsistency with synchrotron radiation, which is the most obvious candidate, and the resulting need for considering less straightforward scenarios. The inclusion of data down to soft X-rays (∼0.5 keV), which are available only in a small fraction of GRBs, has firmly established the common presence of a spectral break in the low-energy part of prompt spectra, and even more importantly, the consistency of the overall spectral shape with synchrotron radiation in the moderately fast-cooling regime, the low-energy break being identified with the cooling frequency. In this work we further extend the range of investigation down to the optical band. In particular, we test the synchrotron interpretation by directly fitting a theoretically derived synchrotron spectrum and making use of optical to gamma-ray data. Secondly, we test an alternative model that considers the presence of a black-body component at ∼keV energies, in addition to a non-thermal component that is responsible for the emission at the spectral peak (100 keV–1 MeV). We find that synchrotron radiation provides a good description of the broadband data, while models composed of a thermal and a non-thermal component require the introduction of a low-energy break in the non-thermal component in order to be consistent with optical observations. Motivated by the good quality of the synchrotron fits, we explore the physical parameter space of the emitting region. In a basic prompt emission scenario we find quite contrived solutions for the magnetic field strength (5 G < B′< 40 G) and for the location of the region where the radiation is produced (Rγ > 1016 cm). We discuss which assumptions of the basic model would need to be relaxed in order to achieve a more natural parameter space.",,https://www.aanda.org/articles/aa/pdf/2019/08/aa35766-19.pdf
-prompt,Cshell Xui,"Array Set the size and location of sub-arrays to be readout for r the next GO. These coordinate are relative to the physical device. You data is normally viewed rotated 90 ̊ clockwise. Prompt Array_icon on the Parameters screen, OBS page. Range x y wid hgt The (x,y) location of the upper left corner and its width and hgt is specified. Please note that these values must be multiples of 8. Initial Full size (0 0 256 256). Syntax ARRAY x y wid hgt",,
-evidence of two spectral breaks in the prompt emission of gamma-ray bursts,M. Ravasio,"The long-lasting tension between the observed spectra of gamma-ray bursts (GRBs) and the predicted synchrotron emission spectrum might be solved if electrons do not completely cool. Evidence of incomplete cooling was recently found in Swift GRBs with prompt observations down to 0.1 keV, and in one bright Fermi burst, GRB 160625B. Here we systematically search for evidence of incomplete cooling in the spectra of the ten brightest short and long GRBs observed by Fermi. We find that in eight out of ten long GRBs there is compelling evidence of a low-energy break (below the peak energy) and good agreement with the photon indices of the synchrotron spectrum (respectively −2/3 and −3/2 below the break and between the break and the peak energy). Interestingly, none of the ten short GRBs analysed shows a break, but the low-energy spectral slope is consistent with −2/3. In a standard scenario, these results imply a very low magnetic field in the emission region (B′∼10 G in the comoving frame), at odd with expectations.",,https://www.aanda.org/articles/aa/pdf/2019/05/aa34987-18.pdf
-digital forensic analysis on idevice : jailbreak ios 12.1.1 as a case study,Amin Ali,"Jailbreak has an issue in data alteration, as it modifies file(s) in the device to allow user to extract more data than without jailbreaking. This issue raises controversy of the use of jailbreaking in digital forensic investigation, as data integrity is a prominent requirement in a court proceeding. This study aims to analyze the process of jailbreak, what is actually done by the jailbreak code in a device, and what data is actually modified by the jailbreak code. By using the latest version of iOS system, this study uses the voucher_swap exploit as a representation of semi-tethered jailbreaking method to investigate the effects of jailbreak on data integrity on a idevice. The investigation is conducted based on to what extent data can be extracted from the jailbreak device, hash value comparison of the data, and source code analysis to scrutinize the effect of jailbreak to the system and user data inside the device. Results of this study suggest that jailbreak is acceptable to prepare idevice in digital forensic investigations to acquire more data, as it maintains the integrity of user data. These results may help forensic communities in their decision about the acceptability of jailbreaking in idevide forensic investigations.","{'model': 'tldr@v2.0.0', 'text': 'Results of this study suggest that jailbreak is acceptable to prepare idevice in digital forensic investigations to acquire more data, as it maintains the integrity of user data.'}",
-detailed polarization measurements of the prompt emission of five gamma-ray bursts,Shuang-Nan Zhang,,,https://arxiv.org/pdf/1901.04207
-3d prompt gamma imaging for proton beam range verification,E. Draeger,"We tested the ability of a single Compton camera (CC) to produce 3-dimensional (3D) images of prompt gammas (PGs) emitted during the irradiation of a tissue-equivalent plastic phantom with proton pencil beams for clinical doses delivered at clinical dose rates. PG measurements were made with a small prototype CC placed at three different locations along the proton beam path. We evaluated the ability of the CC to produce images at each location for two clinical scenarios: (1) the delivery of a single 2 Gy pencil beam from a hypo-fractionated treatment (~9 × 108 protons), and (2) a single pencil beam from a standard treatment (~1 × 108 protons). Additionally, the data measured at each location were combined to simulate measurements with a larger scale, clinical CC and its ability to image shifts in the Bragg peak (BP) range for both clinical scenarios. With our prototype CC, the location of the distal end of the BP could be seen with the CC placed up to 4 cm proximal or distal to the BP distal falloff. Using the data from the simulated full scale clinical CC, 3D images of the PG emission were produced with the delivery of as few as 1 × 108 protons, and shifts in the proton beam range as small as 2 mm could be detected for delivery of a 2 Gy spot. From these results we conclude that 3D PG imaging for proton range verification under clinical beam delivery conditions is possible with a single CC.","{'model': 'tldr@v2.0.0', 'text': '3D PG imaging for proton range verification under clinical beam delivery conditions is possible with a single Compton camera, and shifts in the proton beam range as small as 2\u2009mm could be detected for delivery of a 2 Gy spot.'}",https://europepmc.org/articles/pmc5808927?pdf=render
-grb 190114c: from prompt to afterglow?,M. Ravasio,"GRB 190114C is the first gamma-ray burst detected at very high energies (VHE, i.e., > 300 GeV) by the MAGIC Cherenkov telescope. The analysis of the emission detected by the Fermi satellite at lower energies, in the 10 keV–100 GeV energy range, up to ∼50 s (i.e., before the MAGIC detection) can hold valuable information. We analyze the spectral evolution of the emission of GRB 190114C as detected by the Fermi Gamma-Ray Burst Monitor (GBM) in the 10 keV–40 MeV energy range up to ∼60 s. The first 4 s of the burst feature a typical prompt emission spectrum, which can be fit by a smoothly broken power-law function with typical parameters. Starting on ∼4 s post-trigger, we find an additional nonthermal component that can be fit by a power law. This component rises and decays quickly. The 10 keV–40 MeV flux of the power-law component peaks at ∼6 s; it reaches a value of 1.7 × 10−5 erg cm−2 s−1. The time of the peak coincides with the emission peak detected by the Large Area Telescope (LAT) on board Fermi. The power-law spectral slope that we find in the GBM data is remarkably similar to that of the LAT spectrum, and the GBM+LAT spectral energy distribution seems to be consistent with a single component. This suggests that the LAT emission and the power-law component that we find in the GBM data belong to the same emission component, which we interpret as due to the afterglow of the burst. The onset time allows us to estimate that the initial jet bulk Lorentz factor Γ0 is about 500, depending on the assumed circum-burst density.",,https://www.aanda.org/articles/aa/pdf/2019/06/aa35214-19.pdf
-a full-scale clinical prototype for proton range verification using prompt gamma-ray spectroscopy,F. Hueso-González,"We present a full-scale clinical prototype system for in vivo range verification of proton pencil-beams using the prompt gamma-ray spectroscopy method. The detection system consists of eight LaBr3 scintillators and a tungsten collimator, mounted on a rotating frame. Custom electronics and calibration algorithms have been developed for the measurement of energy- and time-resolved gamma-ray spectra during proton irradiation at a clinical dose rate. Using experimentally determined nuclear reaction cross sections and a GPU-accelerated Monte Carlo simulation, a detailed model of the expected gamma-ray emissions is created for each individual pencil-beam. The absolute range of the proton pencil-beams is determined by minimizing the discrepancy between the measurement and this model, leaving the absolute range of the beam and the elemental concentrations of the irradiated matter as free parameters. The system was characterized in a clinical-like situation by irradiating different phantoms with a scanning pencil-beam. A dose of 0.9 Gy was delivered to a cm3 target with a beam current of 2 nA incident on the phantom. Different range shifters and materials were used to test the robustness of the verification method and to calculate the accuracy of the detected range. The absolute proton range was determined for each spot of the distal energy layer with a mean statistical precision of 1.1 mm at a 95% confidence level and a mean systematic deviation of 0.5 mm, when aggregating pencil-beam spots within a cylindrical region of 10 mm radius and 10 mm depth. Small range errors that we introduced were successfully detected and even large differences in the elemental composition do not affect the range verification accuracy. These results show that our system is suitable for range verification during patient treatments in our upcoming clinical study.","{'model': 'tldr@v2.0.0', 'text': ""Results show that the full-scale clinical prototype system for in vivo range verification of proton pencil-beams using the prompt gamma-ray spectroscopy method is suitable for range verification during patient treatments in the authors' upcoming clinical study.""}",https://europepmc.org/articles/pmc6340397?pdf=render
-false sense of security: a study on the effectivity of jailbreak detection in banking apps,Ansgar Kellner,"People increasingly rely on mobile devices for banking transactions or two-factor authentication (2FA) and thus trust in the security provided by the underlying operating system. Simultaneously, jailbreaks gain tremendous popularity among regular users for customizing their devices. In this paper, we show that both do not go well together: Jailbreaks remove vital security mechanisms, which are necessary to ensure a trusted environment that allows to protect sensitive data, such as login credentials and transaction numbers (TANs). We find that all but one banking app, available in the iOS App Store, can be fully compromised by trivial means without reverse-engineering, manipulating the app, or other sophisticated attacks. Even worse, 44% of the banking apps do not even try to detect jailbreaks, revealing the prevalent, errant trust in the operating system's security. This study assesses the current state of security of banking apps and pleads for more advanced defensive measures for protecting user data.","{'model': 'tldr@v2.0.0', 'text': 'It is found that all but one banking app, available in the iOS App Store, can be fully compromised by trivial means without reverse-engineering, manipulating the app, or other sophisticated attacks.'}",https://ieeexplore.ieee.org/ielx7/8790377/8806708/08806743.pdf
-intravitreal ranibizumab for diabetic macular edema with prompt versus deferred laser treatment: 5-year randomized trial results.,M. Elman,,"{'model': 'tldr@v2.0.0', 'text': 'Five-year results suggest focal/grid laser treatment at the initiation of intravitreal ranibizumab is no better than deferring laser treatment for ≥24 weeks in eyes with DME involving the central macula with vision impairment.'}",https://europepmc.org/articles/pmc4520307?pdf=render
-tdnn: a two-stage deep neural network for prompt-independent automated essay scoring,Cancan Jin,"Existing automated essay scoring (AES) models rely on rated essays for the target prompt as training data. Despite their successes in prompt-dependent AES, how to effectively predict essay ratings under a prompt-independent setting remains a challenge, where the rated essays for the target prompt are not available. To close this gap, a two-stage deep neural network (TDNN) is proposed. In particular, in the first stage, using the rated essays for non-target prompts as the training data, a shallow model is learned to select essays with an extreme quality for the target prompt, serving as pseudo training data; in the second stage, an end-to-end hybrid deep model is proposed to learn a prompt-dependent rating model consuming the pseudo training data from the first step. Evaluation of the proposed TDNN on the standard ASAP dataset demonstrates a promising improvement for the prompt-independent AES task.","{'model': 'tldr@v2.0.0', 'text': 'A two-stage deep neural network (TDNN) is proposed to learn a prompt-dependent rating model consuming the pseudo training data from the first step, demonstrating a promising improvement for the prompt-independent AES task.'}",https://www.aclweb.org/anthology/P18-1100.pdf
-prompt-gamma monitoring in hadrontherapy: a review,J. Krimmer,,,https://hal.archives-ouvertes.fr/hal-01585334/file/Krimmer2017_nima.pdf
-remote-controlled switch allocation enabling prompt restoration of distribution systems,Shunbo Lei,"Remote-controlled switches (RCSs) play an important role in prompt service restoration of distribution systems (DSs). However, the cost of RCSs and the vast footprint of DSs limit widespread utilization of RCSs. In this paper, we present a new approach to RCS allocation for improving the performance of restoration and optimizing reliability benefits with reasonable RCS cost. Specifically, the optimal number and locations of to-be-upgraded switches can be determined with different objectives: maximizing the reduction of customer interruption cost; maximizing the reduction of system average interruption duration index; or maximizing the amount of loads that can be restored using the upgraded RCSs. We show that these models can actually be formulated as mixed-integer convex programming problems. We further introduce a novel method to equivalently transform and efficiently solve each of them. The global optimum can thus be computed within a reasonable amount of time. The IEEE 33-node and 123-node test systems are used to demonstrate the proposed models and algorithms.","{'model': 'tldr@v2.0.0', 'text': 'A new approach to RCS allocation for improving the performance of restoration and optimizing reliability benefits with reasonable RCS cost is presented and a novel method to equivalently transform and efficiently solve each of them is introduced.'}",
-off-axis prompt x-ray transients from the cocoon of short gamma-ray bursts,D. Lazzati,"We present the results of numerical simulations of the prompt emission of short-duration gamma-ray bursts. We consider emission from the relativistic jet, the mildly relativistic cocoon, and the non-relativistic shocked ambient material. We find that the cocoon material is confined between off-axis angles and gives origin to X-ray transients with a duration of a few to ∼10 s, delayed by a few seconds from the time of the merger. We also discuss the distance at which such transients can be detected, finding that it depends sensitively on the assumptions that are made about the radiation spectrum. Purely thermal cocoon transients are detectable only out to a few Mpc, while Comptonized transients can instead be detected by the Fermi Gamma-ray Burst Monitor (GBM) out to several tens of Mpc.",,https://iopscience.iop.org/article/10.3847/2041-8213/aa8f3d/pdf
-measurements of prompt charm production cross-sections in pp collisions at s√=5tev,R. Aaij,"Production cross-sections of prompt charm mesons are measured using data from pp collisions at the LHC at a centre-of-mass energy of 5TeV. The data sample corresponds to an integrated luminosity of 8.60±0.33pb−1 collected by the LHCb experiment. The production cross-sections of D0, D+, D+s, and D∗+ mesons are measured in bins of charm meson transverse momentum, pT, and rapidity, y. They cover the rapidity range 2.0) and the number of photons detected (n′ ?>), i.e. CTR∝τd/n′ ?>. However, it is still an open question to what extent the scintillation rise time (τr ?>) and other fast or prompt photons, e.g. Cherenkov photons, at the beginning of the scintillation process influence the CTR. This paper presents measurements of the scintillation emission rate for different LSO type crystals, i.e. LSO:Ce, LYSO:Ce, LSO:Ce codoped Ca and LGSO:Ce. For the various LSO-type samples measured we find an average value of 70 ps for the scintillation rise time, although some crystals like LSO:Ce codoped Ca seem to have a much faster rise time in the order of 20 ps. Additional measurements for LuAG:Ce and LuAG:Pr show a rise time of 535 ps and 251 ps, respectively. For these crystals, prompt photons (Cherenkov) can be observed at the beginning of the scintillation event. Furthermore a significantly lower rise time value is observed when codoping with calcium. To quantitatively investigate the influence of the rise time to the time resolution we measured the CTR with the same L(Y)SO samples and compared the values to Monte Carlo simulations. Using the measured relative light yields, rise- and decay times of the scintillators we are able to quantitatively understand the measured CTRs in our simulations. Although the rise time is important to fully explain the CTR variation for the different samples tested we determined its influence on the CTR to be in the order of a few percent only. This result is surprising because, if only photonstatistics of the scintillation process is considered, the CTR would be proportional to the square root of the rise time. The unexpected small rise time influence on the CTR can be explained by the convolution of the scintillation rate with the single photon time resolution (SPTR) of the photodetector and the photon travel spread (PTS) in the crystal. The timing benefits of prompt photons at the beginning of the scintillation process (Cherenkov etc) are further studied, which leads to the conclusion that the scintillation rise time, SPTR and PTS have to be lowered simultaneously to fully profit from these fast photons in order to improve the CTR significantly.","{'model': 'tldr@v2.0.0', 'text': 'The timing benefits of prompt photons at the beginning of the scintillation process (Cherenkov etc) are further studied, which leads to the conclusion that theScintillation rise time, SPTR and PTS have to be lowered simultaneously to fully profit from these fast photons in order to improve the CTR significantly.'}",
-promoting end-of-life discussions in advanced cancer: effects of patient coaching and question prompt lists.,Rachel A Rodenbach,"Purpose To build on results of a cluster randomized controlled trial (RCT) of a combined patient-oncologist intervention to improve communication in advanced cancer, we conducted a post hoc analysis of the patient intervention component, a previsit patient coaching session that used a question prompt list (QPL). We hypothesized that intervention-group participants would bring up more QPL-related topics, particularly prognosis-related topics, during the subsequent oncologist visit. Patients and Methods This cluster RCT with 170 patients who had advanced nonhematologic cancer (and their caregivers) recruited from practices of 24 participating oncologists in western New York. Intervention-group oncologists (n = 12) received individualized communication training; up to 10 of their patients (n = 84) received a previsit individualized communication coaching session that incorporated a QPL. Control-group oncologists (n = 12) and patients (n = 86) received no interventions. Topics of interest identified by patients during the coaching session were summarized from coaching notes; one office visit after the coaching session was audio recorded, transcribed, and analyzed by using linear regression modeling for group differences. Results Compared with controls, more than twice as many intervention-group participants brought up QPL-related topics during their office visits (70.2% v 32.6%; P < .001). Patients in the intervention group were nearly three times more likely to ask about prognosis (16.7% v 5.8%; P =.03). Of 262 topics of interest identified during coaching, 158 (60.3%) were QPL related; 20 (12.7%) addressed prognosis. Overall, patients in the intervention group brought up 82.4% of topics of interest during the office visit. Conclusion A combined coaching and QPL intervention was effective to help patients with advanced cancer and their caregivers identify and bring up topics of concern, including prognosis, during their subsequent oncologist visits. Considering that most patients are misinformed about prognosis, more intensive steps are needed to better promote such discussions.","{'model': 'tldr@v2.0.0', 'text': 'A combined coaching and QPL intervention was effective to help patients with advanced cancer and their caregivers identify and bring up topics of concern, including prognosis, during their subsequent oncologist visits.'}",https://europepmc.org/articles/pmc5455683?pdf=render
-correlated prompt fission data in transport simulations,P. Talou,,,https://arxiv.org/pdf/1710.00107
-detection of low-energy breaks in gamma-ray burst prompt emission spectra,G. Oganesyan,"The radiative process responsible for gamma-ray burst (GRB) prompt emission has not been identified yet. If dominated by fast-cooling synchrotron radiation, the part of the spectrum immediately below the peak energy should display a power-law behavior with slope , which breaks to a higher value (i.e., to a harder spectral shape) at lower energies. Prompt emission spectral data (usually available down to keV) are consistent with one single power-law behavior below the peak, with typical slope , higher than (and then inconsistent with) the expected value . To better characterize the spectral shape at low energy, we analyzed 14 GRBs for which the Swift X-ray Telescope started observations during the prompt. When available, Fermi-GBM observations have been included in the analysis. For 67% of the spectra, models that usually give a satisfactory description of the prompt (e.g., the Band model) fail to reproduce the 0.5–1000 keV spectra: low-energy data outline the presence of a spectral break around a few keV. We then introduce an empirical fitting function that includes a low-energy power law , a break energy , a second power law , and a peak energy . We find ( ), ( ), ( ), and ( ). The values and are very close to expectations from synchrotron radiation. In this context, corresponds to the cooling break frequency. The relatively small ratio suggests a regime of moderately fast cooling, which might solve the long-lasting problem of the apparent inconsistency between measured and predicted low-energy spectral index.",,https://iopscience.iop.org/article/10.3847/1538-4357/aa831e/pdf
-prompt neutrinos from atmospheric charm in the general-mass variable-flavor-number scheme,M. Benzke,,,https://link.springer.com/content/pdf/10.1007/JHEP12(2017)021.pdf
-marginally fast cooling synchrotron models for prompt grbs,P. Beniamini,"Previous studies have considered synchrotron as the emission mechanism for prompt Gamma-Ray Bursts (GRBs). These works have shown that the electrons must cool on a timescale comparable to the dynamic time at the source in order to satisfy spectral constraints while maintaining high radiative efficiency. We focus on conditions where synchrotron cooling is balanced by a continuous source of heating, and in which these constraints are naturally satisfied. Assuming that a majority of the electrons in the emitting region are contributing to the observed peak, we find that the energy per electron has to be $E\gtrsim 20$ GeV and that the Lorentz factor of the emitting material has to be very large $10^3\lesssim \Gamma_{\rm em} \lesssim 10^4$, well in excess of the bulk Lorentz factor of the jet inferred from GRB afterglows. A number of independent constraints then indicate that the emitters must be moving relativistically, with $\Gamma'\approx 10$, relative to the bulk frame of the jet and that the jet must be highly magnetized upstream of the emission region, $\sigma_{\rm up}\gtrsim 30$. The emission radius is also strongly constrained in this model to $R\gtrsim 10^{16}$cm. These values are consistent with magnetic jet models where the dissipation is driven by magnetic reconnection that takes place far away from the base of the jet.",,https://academic.oup.com/mnras/article-pdf/476/2/1785/24280125/sty340.pdf
-measurement of the prompt j/ψ pair production cross-section in pp collisions at √s = 8 tev with the atlas detector,Atlas Collaboration,,,
-first clinical application of a prompt gamma based in vivo proton range verification system.,C. Richter,,"{'model': 'tldr@v2.0.0', 'text': 'For the first time, range verification based on prompt gamma imaging was applied for a clinical proton treatment and the potential to improve the precision of particle therapy with this technique has increased considerably.'}",
-measurement of prompt d^{0} meson azimuthal anisotropy in pb-pb collisions at sqrt[s_{nn}]=5.02 tev.,A. Sirunyan,"The prompt D^{0} meson azimuthal anisotropy coefficients, v_{2} and v_{3}, are measured at midrapidity (|y|<1.0) in Pb-Pb collisions at a center-of-mass energy sqrt[s_{NN}]=5.02 TeV per nucleon pair with data collected by the CMS experiment. The measurement is performed in the transverse momentum (p_{T}) range of 1 to 40 GeV/c, for central and midcentral collisions. The v_{2} coefficient is found to be positive throughout the p_{T} range studied. The first measurement of the prompt D^{0} meson v_{3} coefficient is performed, and values up to 0.07 are observed for p_{T} around 4 GeV/c. Compared to measurements of charged particles, a similar p_{T} dependence, but smaller magnitude for p_{T}<6 GeV/c, is found for prompt D^{0} meson v_{2} and v_{3} coefficients. The results are consistent with the presence of collective motion of charm quarks at low p_{T} and a path length dependence of charm quark energy loss at high p_{T}, thereby providing new constraints on the theoretical description of the interactions between charm quarks and the quark-gluon plasma.","{'model': 'tldr@v2.0.0', 'text': 'The results are consistent with the presence of collective motion of Charm quarks at low p_{T} and a path length dependence of charm quark energy loss at high p_{ t}, thereby providing new constraints on the theoretical description of the interactions between charm quarks and the quark-gluon plasma.'}",http://link.aps.org/pdf/10.1103/PhysRevLett.120.202301
-characterization of gamma-ray burst prompt emission spectra down to soft x-rays,G. Oganesyan,"Detection of prompt emission by Swift-XRT provides a unique tool to study how the prompt spectrum of gamma-ray bursts (GRBs) extends down to the soft X-ray band. This energy band is particularly important for prompt emission studies, since it is towards low energies that the observed spectral shape is in disagreement with the synchrotron predictions. Unfortunately, the number of cases where XRT started observing the GRB location during the prompt phase is very limited. In this work, we collect a sample of 34 GRBs and perform joint XRT+BAT spectral analysis of prompt radiation, extending a previous study focused on the 14 brightest cases. Fermi-GBM observations are included in the analysis when available (11 cases), allowing the characterization of prompt spectra from soft X-rays to MeV energies. In 62% of the spectra, the XRT data reveal a hardening of the spectrum, well described by introducing an additional, low-energy power-law segment (with index α1) into the empirical fitting function. The break energy below which the spectrum hardens has values between 3 keV and 22 keV. A second power-law (α2) describes the spectrum between the break energy and the peak energy. The mean values of the photon indices are 〈α1〉 = −0.51 (σ = 0.24) and 〈α2〉 = −1.56 (σ = 0.26). These are consistent, within one σ, with the synchrotron values in fast cooling regime. As a test, if we exclude XRT data from the fits we find typical results: the spectrum below the peak energy is described by a power law with 〈α〉 = −1.15. This shows the relevance of soft X-ray data in revealing prompt emission spectra consistent with synchrotron spectra. Finally, we do not find any correlation between the presence of the X-ray break energy and the flux, fluence, or duration of the prompt emission.",,https://www.aanda.org/articles/aa/pdf/2018/08/aa32172-17.pdf
-gamma-ray burst prompt correlations: selection and instrumental effects,M. Dainotti,"The prompt emission mechanism of gamma-ray bursts (GRB) even after several decades remains a mystery. However, it is believed that correlations between observable GRB properties, given their huge luminosity/radiated energy and redshift distribution extending up to at least z ≈ 9, are promising possible cosmological tools. They also may help to discriminate among the most plausible theoretical models. Nowadays, the objective is to make GRBs standard candles, similar to supernovae (SNe) Ia, through well-established and robust correlations. However, differently from SNe Ia, GRBs span over several order of magnitude in their energetics, hence they cannot yet be considered standard candles. Additionally, being observed at very large distances, their physical properties are affected by selection biases, the so-called Malmquist bias or Eddington effect. We describe the state of the art on how GRB prompt correlations are corrected for these selection biases to employ them as redshift estimators and cosmological tools. We stress that only after an appropriate evaluation and correction for these effects, GRB correlations can be used to discriminate among the theoretical models of prompt emission, to estimate the cosmological parameters and to serve as distance indicators via redshift estimation.",,https://iopscience.iop.org/article/10.1088/1538-3873/aaa8d7/pdf
-the characteristics and effectiveness of question prompt list interventions in oncology: a systematic review of the literature,K. Brandes,"Question Prompt Lists (QPLs) have been used extensively in the oncology setting to improve communication, psychological and/or cognitive outcomes. In this systematic review, the objectives were to (a) examine the methodological quality of QPL interventions, (b) review the effectiveness of QPL interventions on communication, psychological and/or cognitive outcomes of cancer patients, (c) gain more insight into the characteristics of QPL interventions (e.g., the number and content of questions, and the mode of delivery) and (d) explore whether the effectiveness of QPL interventions might be explained by their characteristics.","{'model': 'tldr@v2.0.0', 'text': 'This systematic review reviewed the effectiveness of QPL interventions on communication, psychological and/or cognitive outcomes of cancer patients, and gained more insight into the characteristics ofQPL interventions to explore whether the effectiveness might be explained by their characteristics.'}",https://pure.uva.nl/ws/files/2138355/167564_433599.pdf
-prompt atmospheric neutrino fluxes: perturbative qcd models and nuclear effects,A. Bhattacharya,,,https://link.springer.com/content/pdf/10.1007%2FJHEP11%282016%29167.pdf
-"weakly bound free radicals in combustion: ""prompt"" dissociation of formyl radicals and its effect on laminar flame speeds.",N. Labbe,"Weakly bound free radicals have low-dissociation thresholds such that at high temperatures, time scales for dissociation and collisional relaxation become comparable, leading to significant dissociation during the vibrational-rotational relaxation process. Here we characterize this ""prompt"" dissociation of formyl (HCO), an important combustion radical, using direct dynamics calculations for OH + CH2O and H + CH2O (key HCO-forming reactions). For all other HCO-forming reactions, presumption of a thermal incipient HCO distribution was used to derive prompt dissociation fractions. Inclusion of these theoretically derived HCO prompt dissociation fractions into combustion kinetics models provides an additional source for H-atoms that feeds chain-branching reactions. Simulations using these updated combustion models are therefore shown to enhance flame propagation in 1,3,5-trioxane and acetylene. The present results suggest that HCO prompt dissociation should be included when simulating flames of hydrocarbons and oxygenated molecules and that prompt dissociations of other weakly bound radicals may also impact combustion simulations.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that HCO prompt dissociation should be included when simulating flames of hydrocarbons and oxygenated molecules and that prompt dissociations of other weakly bound radicals may also impact combustion simulations.'}",
-observation and measurements of the production of prompt and non-prompt j/psi mesons in association with a z boson in pp collisions at √s = 8 tev with the atlas detector,S. Leontsinis,"The associated production of a vector boson with heavy quarkonia is a key observable for understanding the quarkonium production mechanisms. In this poster the observation of the production of the Z boson in association with a prompt or with a non-prompt J/ψ meson with the ATLAS detector at LHC is presented and its production rate is measured in comparison of the inclusive Z production. Relative contributions to the signal from single and double parton scattering are estimated. Single parton scattering cross-sections are compared to cutting-edge theoretical calculations in the colour singlet and colour octet formalisms. Finally, a lower limit in the double parton scattering effective cross section is extracted.",,
-prompt gravity signal induced by the 2011 tohoku-oki earthquake,J. Montagner,,"{'model': 'tldr@v2.0.0', 'text': 'While prompt gravity signal detection with state-of-the-art gravimeters and seismometers is challenged by background seismic noise, its robust detection with gravity gradiometers under development could open new directions in earthquake seismology, and overcome fundamental limitations of current earthquake early-warning systems imposed by the propagation speed of seismic waves.'}",https://www.nature.com/articles/ncomms13349.pdf
-imaging of prompt gamma rays emitted during delivery of clinical proton beams with a compton camera: feasibility studies for range verification,J. Polf,"The purpose of this paper is to evaluate the ability of a prototype Compton camera (CC) to measure prompt gamma rays (PG) emitted during delivery of clinical proton pencil beams for prompt gamma imaging (PGI) as a means of providing in vivo verification of the delivered proton radiotherapy beams. A water phantom was irradiated with clinical 114 MeV and 150 MeV proton pencil beams. Up to 500 cGy of dose was delivered per irradiation using clinical beam currents. The prototype CC was placed 15 cm from the beam central axis and PGs from 0.2 MeV up to 6.5 MeV were measured during irradiation. From the measured data (2D) images of the PG emission were reconstructed. (1D) profiles were extracted from the PG images and compared to measured depth dose curves of the delivered proton pencil beams. The CC was able to measure PG emission during delivery of both 114 MeV and 150 MeV proton beams at clinical beam currents. 2D images of the PG emission were reconstructed for single 150 MeV proton pencil beams as well as for a 5 × 5 cm mono-energetic layer of 114 MeV pencil beams. Shifts in the Bragg peak (BP) range were detectable on the 2D images. 1D profiles extracted from the PG images show that the distal falloff of the PG emission profile lined up well with the distal BP falloff. Shifts as small as 3 mm in the beam range could be detected from the 1D PG profiles with an accuracy of 1.5 mm or better. However, with the current CC prototype, a dose of 400 cGy was required to acquire adequate PG signal for 2D PG image reconstruction. It was possible to measure PG interactions with our prototype CC during delivery of proton pencil beams at clinical dose rates. Images of the PG emission could be reconstructed and shifts in the BP range were detectable. Therefore PGI with a CC for in vivo range verification during proton treatment delivery is feasible. However, improvements in the prototype CC detection efficiency and reconstruction algorithms are necessary to make it a clinically viable PGI system.","{'model': 'tldr@v2.0.0', 'text': 'PGI with a CC for in vivo range verification during proton treatment delivery is feasible, however, improvements in the prototype CC detection efficiency and reconstruction algorithms are necessary to make it a clinically viable PGI system.'}",
-a revised analysis of gamma-ray bursts’ prompt efficiencies,P. Beniamini,"The prompt Gamma-Ray Bursts' (GRBs) efficiency is an important clue on the emission mechanism producing the $\gamma$-rays. Previous estimates of the kinetic energy of the blast waves, based on the X-ray afterglow luminosity $L_X$, suggested that this efficiency is large, with values above 90\% in some cases. This poses a problem to emission mechanisms and in particular to the internal shocks model. These estimates are based, however, on the assumption that the X-ray emitting electrons are fast cooling and that their Inverse Compton (IC) losses are negligible. The observed correlations between $L_X$ (and hence the blast wave energy) and $E_{\gamma\rm ,iso}$, the isotropic equivalent energy in the prompt emission, has been considered as observational evidence supporting this analysis. It is reasonable that the prompt gamma-ray energy and the blast wave kinetic energy are correlated and the observed correlation corroborates, therefore, the notion $L_X$ is indeed a valid proxy for the latter. Recent findings suggest that the magnetic field in the afterglow shocks is significantly weaker than was earlier thought and its equipartition fraction, $\epsilon_B$, could be as low as $10^{-4}$ or even lower. Motivated by these findings we reconsider the problem, taking now IC cooling into account. We find that the observed $L_X-E_{\gamma\rm ,iso}$ correlation is recovered also when IC losses are significant. For small $\epsilon_B$ values the blast wave must be more energetic and we find that the corresponding prompt efficiency is significantly smaller than previously thought. For example, for $\epsilon_B\sim10^{-4}$ we infer a typical prompt efficiency of $\sim15\%$.",,https://academic.oup.com/mnras/article-pdf/461/1/51/8043156/stw1331.pdf
-proton range verification through prompt gamma-ray spectroscopy,J. Verburg,"We present an experimental study of a novel method to verify the range of proton therapy beams. Differential cross sections were measured for 15 prompt gamma-ray lines from proton-nuclear interactions with 12C and 16O at proton energies up to 150 MeV. These cross sections were used to model discrete prompt gamma-ray emissions along proton pencil-beams. By fitting detected prompt gamma-ray counts to these models, we simultaneously determined the beam range and the oxygen and carbon concentration of the irradiated matter. The performance of the method was assessed in two phantoms with different elemental concentrations, using a small scale prototype detector. Based on five pencil-beams with different ranges delivering 5 × 108 protons and without prior knowledge of the elemental composition at the measurement point, the absolute range was determined with a standard deviation of 1.0–1.4 mm. Relative range shifts at the same dose level were detected with a standard deviation of 0.3–0.5 mm. The determined oxygen and carbon concentrations also agreed well with the actual values. These results show that quantitative prompt gamma-ray measurements enable knowledge of nuclear reaction cross sections to be used for precise proton range verification in the presence of tissue with an unknown composition.","{'model': 'tldr@v2.0.0', 'text': 'Results show that quantitative prompt gamma-ray measurements enable knowledge of nuclear reaction cross sections to be used for precise proton range verification in the presence of tissue with an unknown composition.'}",
-search for prompt neutrino emission from gamma-ray bursts with icecube,I. C. M. Aartsen,"We present constraints derived from a search of four years of IceCube data for a prompt neutrino flux from gamma-ray bursts (GRBs). A single low-significance neutrino, compatible with the atmospheric neutrino background, was found in coincidence with one of the 506 observed bursts. Although GRBs have been proposed as candidate sources for ultra-high-energy cosmic rays, our limits on the neutrino flux disfavor much of the parameter space for the latest models. We also find that no more than ∼1% of the recently observed astrophysical neutrino flux consists of prompt emission from GRBs that are potentially observable by existing satellites.",,https://iopscience.iop.org/article/10.1088/2041-8205/805/1/L5/pdf
-calculation of conventional and prompt lepton fluxes at very high energy,A. Fedynitch,"An efficient method for calculating inclusive conventional and prompt atmospheric leptons fluxes is presented. The coupled cascade equations are solved numerically by formulating them as matrix equation. The presented approach is very flexible and allows the use of different hadronic interaction models, realistic parametrizations of the primary cosmic-ray flux and the Earth's atmosphere, and a detailed treatment of particle interactions and decays. The power of the developed method is illustrated by calculating lepton flux predictions for a number of different scenarios.",,https://www.epj-conferences.org/articles/epjconf/pdf/2015/18/epjconf-isvhecri2014_08001.pdf
-range assessment in particle therapy based on prompt γ-ray timing measurements,C. Golnik,"Proton and ion beams open up new vistas for the curative treatment of tumors, but adequate technologies for monitoring the compliance of dose delivery with treatment plans in real time are still missing. Range assessment, meaning the monitoring of therapy-particle ranges in tissue during dose delivery (treatment), is a continuous challenge considered a key for tapping the full potential of particle therapies. In this context the paper introduces an unconventional concept of range assessment by prompt-gamma timing (PGT), which is based on an elementary physical effect not considered so far: therapy particles penetrating tissue move very fast, but still need a finite transit time—about 1–2 ns in case of protons with a 5–20 cm range—from entering the patient’s body until stopping in the target volume. The transit time increases with the particle range. This causes measurable effects in PGT spectra, usable for range verification. The concept was verified by proton irradiation experiments at the AGOR cyclotron, KVI-CART, University of Groningen. Based on the presented kinematical relations, we describe model calculations that very precisely reproduce the experimental results. As the clinical treatment conditions entail measurement constraints (e.g. limited treatment time), we propose a setup, based on clinical irradiation conditions, capable of determining proton range deviations within a few seconds of irradiation, thus allowing for a fast safety survey. Range variations of 2 mm are expected to be clearly detectable.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a setup, based on clinical irradiation conditions, capable of determining proton range deviations within a few seconds of irradiation, thus allowing for a fast safety survey, and describes model calculations that very precisely reproduce the experimental results.'}",https://iopscience.iop.org/article/10.1088/0031-9155/59/18/5399/pdf
-prompt fission neutron spectra of actinides,R. Capote,,,
-measurement of prompt d-meson production in p-pb collisions at $\sqrt{s_{\rm nn}}$ = 5.02 tev,B. Abelev,"The $p_{\rm T}$-differential production cross sections of the prompt charmed mesons $D^0$, $D^+$, $D^{*+}$ and $D_{\rm s}^{+}$ and their charge conjugate in the rapidity interval $-0.9691.9%), GSM8K (92% ->95.5%), AQuA (76.4% ->79.9%) and MATH (50.3% ->53.9%).","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new prompting method, named Progressive-Hint Prompting (PHP), that enables automatic multiple interactions between users and LLMs by using previously generated answers as hints to progressively guide toward the correct answers.'}",https://arxiv.org/pdf/2304.09797
-frugalgpt: how to use large language models while reducing cost and improving performance,Lingjiao Chen,"There is a rapidly growing number of large language models (LLMs) that users can query for a fee. We review the cost associated with querying popular LLM APIs, e.g. GPT-4, ChatGPT, J1-Jumbo, and find that these models have heterogeneous pricing structures, with fees that can differ by two orders of magnitude. In particular, using LLMs on large collections of queries and text can be expensive. Motivated by this, we outline and discuss three types of strategies that users can exploit to reduce the inference cost associated with using LLMs: 1) prompt adaptation, 2) LLM approximation, and 3) LLM cascade. As an example, we propose FrugalGPT, a simple yet flexible instantiation of LLM cascade which learns which combinations of LLMs to use for different queries in order to reduce cost and improve accuracy. Our experiments show that FrugalGPT can match the performance of the best individual LLM (e.g. GPT-4) with up to 98% cost reduction or improve the accuracy over GPT-4 by 4% with the same cost. The ideas and findings presented here lay a foundation for using LLMs sustainably and efficiently.","{'model': 'tldr@v2.0.0', 'text': 'FrugalGPT is proposed, a simple yet flexible instantiation of LLM cascade which learns which combinations of LLMs to use for different queries in order to reduce cost and improve accuracy and lays a foundation for using LLMs sustainably and efficiently.'}",http://arxiv.org/pdf/2305.05176
-conversational automated program repair,Chun Xia,"Automated Program Repair (APR) can help developers automatically generate patches for bugs. Due to the impressive performance obtained using Large Pre-Trained Language Models (LLMs) on many code related tasks, researchers have started to directly use LLMs for APR. However, prior approaches simply repeatedly sample the LLM given the same constructed input/prompt created from the original buggy code, which not only leads to generating the same incorrect patches repeatedly but also miss the critical information in testcases. To address these limitations, we propose conversational APR, a new paradigm for program repair that alternates between patch generation and validation in a conversational manner. In conversational APR, we iteratively build the input to the model by combining previously generated patches with validation feedback. As such, we leverage the long-term context window of LLMs to not only avoid generating previously incorrect patches but also incorporate validation feedback to help the model understand the semantic meaning of the program under test. We evaluate 10 different LLM including the newly developed ChatGPT model to demonstrate the improvement of conversational APR over the prior LLM for APR approach.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2301.13246
-annollm: making large language models to be better crowdsourced annotators,Xingwei He,"Many natural language processing (NLP) tasks rely on labeled data to train machine learning models to achieve high performance. However, data annotation can be a time-consuming and expensive process, especially when the task involves a large amount of data or requires specialized domains. Recently, GPT-3.5 series models have demonstrated remarkable few-shot and zero-shot ability across various NLP tasks. In this paper, we first claim that large language models (LLMs), such as GPT-3.5, can serve as an excellent crowdsourced annotator by providing them with sufficient guidance and demonstrated examples. To make LLMs to be better annotators, we propose a two-step approach, 'explain-then-annotate'. To be more precise, we begin by creating prompts for every demonstrated example, which we subsequently utilize to prompt a LLM to provide an explanation for why the specific ground truth answer/label was chosen for that particular example. Following this, we construct the few-shot chain-of-thought prompt with the self-generated explanation and employ it to annotate the unlabeled data. We conduct experiments on three tasks, including user input and keyword relevance assessment, BoolQ and WiC. The annotation results from GPT-3.5 surpasses those from crowdsourced annotation for user input and keyword relevance assessment. Additionally, for the other two tasks, GPT-3.5 achieves results that are comparable to those obtained through crowdsourced annotation.","{'model': 'tldr@v2.0.0', 'text': 'This paper claims that large language models, such as GPT-3.5, can serve as an excellent crowdsourced annotator by providing them with sufficient guidance and demonstrated examples, and proposes a two-step approach, explain-then-annotate, to make LLMs to be better annotators.'}",http://arxiv.org/pdf/2303.16854
-keep the conversation going: fixing 162 out of 337 bugs for $0.42 each using chatgpt,Chun Xia,"Automated Program Repair (APR) aims to automatically generate patches for buggy programs. Recent APR work has been focused on leveraging modern Large Language Models (LLMs) to directly generate patches for APR. Such LLM-based APR tools work by first constructing an input prompt built using the original buggy code and then queries the LLM to generate patches. While the LLM-based APR tools are able to achieve state-of-the-art results, it still follows the classic Generate and Validate repair paradigm of first generating lots of patches and then validating each one afterwards. This not only leads to many repeated patches that are incorrect but also miss the crucial information in test failures as well as in plausible patches. To address these limitations, we propose ChatRepair, the first fully automated conversation-driven APR approach that interleaves patch generation with instant feedback to perform APR in a conversational style. ChatRepair first feeds the LLM with relevant test failure information to start with, and then learns from both failures and successes of earlier patching attempts of the same bug for more powerful APR. For earlier patches that failed to pass all tests, we combine the incorrect patches with their corresponding relevant test failure information to construct a new prompt for the LLM to generate the next patch. In this way, we can avoid making the same mistakes. For earlier patches that passed all the tests, we further ask the LLM to generate alternative variations of the original plausible patches. In this way, we can further build on and learn from earlier successes to generate more plausible patches to increase the chance of having correct patches. While our approach is general, we implement ChatRepair using state-of-the-art dialogue-based LLM -- ChatGPT. By calculating the cost of accessing ChatGPT, we can fix 162 out of 337 bugs for \$0.42 each!","{'model': 'tldr@v2.0.0', 'text': 'ChatRepair is proposed, the first fully automated conversation-driven APR approach that interleaves patch generation with instant feedback to perform APR in a conversational style and is implemented using state-of-the-art dialogue-based LLM -- ChatGPT.'}",http://arxiv.org/pdf/2304.00385
-a survey on large language models for recommendation,Likang Wu,"Large Language Models (LLMs) have emerged as powerful tools in the field of Natural Language Processing (NLP) and have recently gained significant attention in the domain of Recommendation Systems (RS). These models, trained on massive amounts of data using self-supervised learning, have demonstrated remarkable success in learning universal representations and have the potential to enhance various aspects of recommendation systems by some effective transfer techniques such as fine-tuning and prompt tuning, and so on. The crucial aspect of harnessing the power of language models in enhancing recommendation quality is the utilization of their high-quality representations of textual features and their extensive coverage of external knowledge to establish correlations between items and users. To provide a comprehensive understanding of the existing LLM-based recommendation systems, this survey presents a taxonomy that categorizes these models into two major paradigms, respectively Discriminative LLM for Recommendation (DLLM4Rec) and Generative LLM for Recommendation (GLLM4Rec), with the latter being systematically sorted out for the first time. Furthermore, we systematically review and analyze existing LLM-based recommendation systems within each paradigm, providing insights into their methodologies, techniques, and performance. Additionally, we identify key challenges and several valuable findings to provide researchers and practitioners with inspiration. We have also created a GitHub repository to index relevant papers on LLMs for recommendation, https://github.com/WLiK/LLM4Rec.","{'model': 'tldr@v2.0.0', 'text': 'This survey presents a taxonomy that categorizes existing LLM-based recommendation systems into two major paradigms, respectively Discriminative LLM for Recommendation (DLLM4Rec) and Generative LLL4Rec (GLLM 4Rec), with the latter being systematically sorted out for the first time.'}",https://arxiv.org/pdf/2305.19860
-robots that ask for help: uncertainty alignment for large language model planners,Allen Z. Ren,"Large language models (LLMs) exhibit a wide range of promising capabilities -- from step-by-step planning to commonsense reasoning -- that may provide utility for robots, but remain prone to confidently hallucinated predictions. In this work, we present KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed. KnowNo builds on the theory of conformal prediction to provide statistical guarantees on task completion while minimizing human help in complex multi-step planning settings. Experiments across a variety of simulated and real robot setups that involve tasks with different modes of ambiguity (e.g., from spatial to numeric uncertainties, from human preferences to Winograd schemas) show that KnowNo performs favorably over modern baselines (which may involve ensembles or extensive prompt tuning) in terms of improving efficiency and autonomy, while providing formal assurances. KnowNo can be used with LLMs out of the box without model-finetuning, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models. Website: https://robot-help.github.io","{'model': 'tldr@v2.0.0', 'text': ""This work presents KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models.""}",https://arxiv.org/pdf/2307.01928
-marked personas: using natural language prompts to measure stereotypes in language models,Myra Cheng,"To recognize and mitigate harms from large language models (LLMs), we need to understand the prevalence and nuances of stereotypes in LLM outputs. Toward this end, we present Marked Personas, a prompt-based method to measure stereotypes in LLMs for intersectional demographic groups without any lexicon or data labeling.Grounded in the sociolinguistic concept of markedness (which characterizes explicitly linguistically marked categories versus unmarked defaults), our proposed method is twofold: 1) prompting an LLM to generate personas, i.e., natural language descriptions, of the target demographic group alongside personas of unmarked, default groups; 2) identifying the words that significantly distinguish personas of the target group from corresponding unmarked ones.We find that the portrayals generated by GPT-3.5 and GPT-4 contain higher rates of racial stereotypes than human-written portrayals using the same prompts. The words distinguishing personas of marked (non-white, non-male) groups reflect patterns of othering and exoticizing these demographics. An intersectional lens further reveals tropes that dominate portrayals of marginalized groups, such as tropicalism and the hypersexualization of minoritized women. These representational harms have concerning implications for downstream applications like story generation.",,http://arxiv.org/pdf/2305.18189
-towards measuring the representation of subjective global opinions in language models,Esin Durmus,"Large language models (LLMs) may not equitably represent diverse global perspectives on societal issues. In this paper, we develop a quantitative framework to evaluate whose opinions model-generated responses are more similar to. We first build a dataset, GlobalOpinionQA, comprised of questions and answers from cross-national surveys designed to capture diverse opinions on global issues across different countries. Next, we define a metric that quantifies the similarity between LLM-generated survey responses and human responses, conditioned on country. With our framework, we run three experiments on an LLM trained to be helpful, honest, and harmless with Constitutional AI. By default, LLM responses tend to be more similar to the opinions of certain populations, such as those from the USA, and some European and South American countries, highlighting the potential for biases. When we prompt the model to consider a particular country's perspective, responses shift to be more similar to the opinions of the prompted populations, but can reflect harmful cultural stereotypes. When we translate GlobalOpinionQA questions to a target language, the model's responses do not necessarily become the most similar to the opinions of speakers of those languages. We release our dataset for others to use and build on. Our data is at https://huggingface.co/datasets/Anthropic/llm_global_opinions. We also provide an interactive visualization at https://llmglobalvalues.anthropic.com.","{'model': 'tldr@v2.0.0', 'text': 'A quantitative framework to evaluate whose opinions model-generated responses are more similar to, and defines a metric that quantifies the similarity between LLM-generated survey responses and human responses, conditioned on country.'}",http://arxiv.org/pdf/2306.16388
-supporting qualitative analysis with large language models: combining codebook with gpt-3 for deductive coding,Ziang Xiao,"Qualitative analysis of textual contents unpacks rich and valuable information by assigning labels to the data. However, this process is often labor-intensive, particularly when working with large datasets. While recent AI-based tools demonstrate utility, researchers may not have readily available AI resources and expertise, let alone be challenged by the limited generalizability of those task-specific models. In this study, we explored the use of large language models (LLMs) in supporting deductive coding, a major category of qualitative analysis where researchers use pre-determined codebooks to label the data into a fixed set of codes. Instead of training task-specific models, a pre-trained LLM could be used directly for various tasks without fine-tuning through prompt learning. Using a curiosity-driven questions coding task as a case study, we found, by combining GPT-3 with expert-drafted codebooks, our proposed approach achieved fair to substantial agreements with expert-coded results. We lay out challenges and opportunities in using LLMs to support qualitative coding and beyond.","{'model': 'tldr@v2.0.0', 'text': 'This study explored the use of large language models (LLMs) in supporting deductive coding, a major category of qualitative analysis where researchers use pre-determined codebooks to label the data into a fixed set of codes, and found a pre-trained LLM could be used directly for various tasks without fine-tuning through prompt learning.'}",https://arxiv.org/pdf/2304.10548
-assessment of chemistry knowledge in large language models that generate code,A. White,"In this work, we investigate the question: do code-generating large language models know chemistry? Our results indicate, mostly yes. To evaluate this, we introduce an expandable framework for evaluating chemistry knowledge in these models, through prompting models to solve chemistry problems posed as coding tasks. To do so, we produce a benchmark set of problems, and evaluate these models based on correctness of code by automated testing and evaluation by experts. We find that recent LLMs are able to write correct code across a variety of topics in chemistry and their accuracy can be increased by 30 percentage points via prompt engineering strategies, like putting copyright notices at the top of files. Our dataset and evaluation tools are open source which can be contributed to or built upon by future researchers, and will serve as a community resource for evaluating the performance of new models as they emerge. We also describe some good practices for employing LLMs in chemistry. The general success of these models demonstrates that their impact on chemistry teaching and research is poised to be enormous.","{'model': 'tldr@v2.0.0', 'text': 'It is found that recent LLMs are able to write correct code across a variety of topics in chemistry and their accuracy can be increased by 30 percentage points via prompt engineering strategies, like putting copyright notices at the top of files.'}",https://pubs.rsc.org/en/content/articlepdf/2023/dd/d2dd00087c
-accuracy of information and references using chatgpt-3 for retrieval of clinical radiological information.,M. Wagner,"Purpose: To assess the accuracy of answers provided by ChatGPT-3 when prompted with questions from the daily routine of radiologists and to evaluate the text response when ChatGPT-3 was prompted to provide references for a given answer. Methods: ChatGPT-3 (San Francisco, OpenAI) is an artificial intelligence chatbot based on a large language model (LLM) that has been designed to generate human-like text. A total of 88 questions were submitted to ChatGPT-3 using textual prompt. These 88 questions were equally dispersed across 8 subspecialty areas of radiology. The responses provided by ChatGPT-3 were assessed for correctness by cross-checking them with peer-reviewed, PubMed-listed references. In addition, the references provided by ChatGPT-3 were evaluated for authenticity. Results: A total of 59 of 88 responses (67%) to radiological questions were correct, while 29 responses (33%) had errors. Out of 343 references provided, only 124 references (36.2%) were available through internet search, while 219 references (63.8%) appeared to be generated by ChatGPT-3. When examining the 124 identified references, only 47 references (37.9%) were considered to provide enough background to correctly answer 24 questions (37.5%). Conclusion: In this pilot study, ChatGPT-3 provided correct responses to questions from the daily clinical routine of radiologists in only about two thirds, while the remainder of responses contained errors. The majority of provided references were not found and only a minority of the provided references contained the correct information to answer the question. Caution is advised when using ChatGPT-3 to retrieve radiological information.","{'model': 'tldr@v2.0.0', 'text': 'In this pilot study, ChatGPT-3 provided correct responses to questions from the daily clinical routine of radiologists in only about two thirds, while the remainder of responses contained errors.'}",
-in-context impersonation reveals large language models' strengths and biases,Leonard Salewski,"In everyday conversations, humans can take on different roles and adapt their vocabulary to their chosen roles. We explore whether LLMs can take on, that is impersonate, different roles when they generate text in-context. We ask LLMs to assume different personas before solving vision and language tasks. We do this by prefixing the prompt with a persona that is associated either with a social identity or domain expertise. In a multi-armed bandit task, we find that LLMs pretending to be children of different ages recover human-like developmental stages of exploration. In a language-based reasoning task, we find that LLMs impersonating domain experts perform better than LLMs impersonating non-domain experts. Finally, we test whether LLMs' impersonations are complementary to visual information when describing different categories. We find that impersonation can improve performance: an LLM prompted to be a bird expert describes birds better than one prompted to be a car expert. However, impersonation can also uncover LLMs' biases: an LLM prompted to be a man describes cars better than one prompted to be a woman. These findings demonstrate that LLMs are capable of taking on diverse roles and that this in-context impersonation can be used to uncover their hidden strengths and biases.",,http://arxiv.org/pdf/2305.14930
-knn prompting: beyond-context learning with calibration-free nearest neighbor inference,Benfeng Xu,"In-Context Learning (ICL), which formulates target tasks as prompt completion conditioned on in-context demonstrations, has become the prevailing utilization of LLMs. In this paper, we first disclose an actual predicament for this typical usage that it can not scale up with training data due to context length restriction. Besides, existing works have shown that ICL also suffers from various biases and requires delicate calibration treatment. To address both challenges, we advocate a simple and effective solution, $k$NN Prompting, which first queries LLM with training data for distributed representations, then predicts test instances by simply referring to nearest neighbors. We conduct comprehensive experiments to demonstrate its two-fold superiority: 1) Calibration-Free: $k$NN Prompting does not directly align LLM output distribution with task-specific label space, instead leverages such distribution to align test and training instances. It significantly outperforms state-of-the-art calibration-based methods under comparable few-shot scenario. 2) Beyond-Context: $k$NN Prompting can further scale up effectively with as many training data as are available, continually bringing substantial improvements. The scaling trend holds across 10 orders of magnitude ranging from 2 shots to 1024 shots as well as different LLMs scales ranging from 0.8B to 30B. It successfully bridges data scaling into model scaling, and brings new potentials for the gradient-free paradigm of LLM deployment. Code is publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper advocates a simple and effective solution, $k$NN Prompting, which first queries LLM with training data for distributed representations, then predicts test instances by simply referring to nearest neighbors, and significantly outperforms state-of-the-art calibration-based methods under comparable few-shot scenario.'}",http://arxiv.org/pdf/2303.13824
-"on second thought, let’s not think step by step! bias and toxicity in zero-shot reasoning",Omar Shaikh,"Generating a Chain of Thought (CoT) has been shown to consistently improve large language model (LLM) performance on a wide range of NLP tasks. However, prior work has mainly focused on logical reasoning tasks (e.g. arithmetic, commonsense QA); it remains unclear whether improvements hold for more diverse types of reasoning, especially in socially situated contexts. Concretely, we perform a controlled evaluation of zero-shot CoT across two socially sensitive domains: harmful questions and stereotype benchmarks. We find that zero-shot CoT reasoning in sensitive domains significantly increases a model’s likelihood to produce harmful or undesirable output, with trends holding across different prompt formats and model variants. Furthermore, we show that harmful CoTs increase with model size, but decrease with improved instruction following. Our work suggests that zero-shot CoT should be used with caution on socially important tasks, especially when marginalized groups or sensitive topics are involved.","{'model': 'tldr@v2.0.0', 'text': 'It is found that zero-shot CoT reasoning in sensitive domains significantly increases a model’s likelihood to produce harmful or undesirable output, with trends holding across different prompt formats and model variants.'}",http://arxiv.org/pdf/2212.08061
-evaluation of chatgpt for nlp-based mental health applications,Bishal Lamichhane,"Large language models (LLM) have been successful in several natural language understanding tasks and could be relevant for natural language processing (NLP)-based mental health application research. In this work, we report the performance of LLM-based ChatGPT (with gpt-3.5-turbo backend) in three text-based mental health classification tasks: stress detection (2-class classification), depression detection (2-class classification), and suicidality detection (5-class classification). We obtained annotated social media posts for the three classification tasks from public datasets. Then ChatGPT API classified the social media posts with an input prompt for classification. We obtained F1 scores of 0.73, 0.86, and 0.37 for stress detection, depression detection, and suicidality detection, respectively. A baseline model that always predicted the dominant class resulted in F1 scores of 0.35, 0.60, and 0.19. The zero-shot classification accuracy obtained with ChatGPT indicates a potential use of language models for mental health classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'The zero-shot classification accuracy obtained with ChatGPT indicates a potential use of language models for mental health classification tasks, and large language models have been successful in several natural language understanding tasks.'}",http://arxiv.org/pdf/2303.15727
-exploiting asymmetry for synthetic training data generation: synthie and the case of information extraction,Martin Josifoski,"Large language models (LLMs) have great potential for synthetic data generation. This work shows that useful data can be synthetically generated even for tasks that cannot be solved directly by LLMs: for problems with structured outputs, it is possible to prompt an LLM to perform the task in the reverse direction, by generating plausible input text for a target output structure. Leveraging this asymmetry in task difficulty makes it possible to produce large-scale, high-quality data for complex tasks. We demonstrate the effectiveness of this approach on closed information extraction, where collecting ground-truth data is challenging, and no satisfactory dataset exists to date. We synthetically generate a dataset of 1.8M data points, establish its superior quality compared to existing datasets in a human evaluation, and use it to finetune small models (220M and 770M parameters), termed SynthIE, that outperform the prior state of the art (with equal model size) by a substantial margin of 57 absolute points in micro-F1 and 79 points in macro-F1. Code, data, and models are available at https://github.com/epfl-dlab/SynthIE.","{'model': 'tldr@v2.0.0', 'text': 'This work synthetically generates a dataset of 1.8M data points, establishes its superior quality compared to existing datasets in a human evaluation, and uses it to finetune small models, termed SynthIE, that outperform the prior state of the art by a substantial margin.'}",http://arxiv.org/pdf/2303.04132
-llavar: enhanced visual instruction tuning for text-rich image understanding,Yanzhe Zhang,"Instruction tuning unlocks the superior capability of Large Language Models (LLM) to interact with humans. Furthermore, recent instruction-following datasets include images as visual inputs, collecting responses for image-based instructions. However, visual instruction-tuned models cannot comprehend textual details within images well. This work enhances the current visual instruction tuning pipeline with text-rich images (e.g., movie posters, book covers, etc.). Specifically, we first use publicly available OCR tools to collect results on 422K text-rich images from the LAION dataset. Moreover, we prompt text-only GPT-4 with recognized texts and image captions to generate 16K conversations, each containing question-answer pairs for text-rich images. By combining our collected data with previous multi-modal instruction-following data, our model, LLaVAR, substantially improves the LLaVA model's capability on text-based VQA datasets (up to 20% accuracy improvement) while achieving an accuracy of 91.42% on ScienceQA. The GPT-4-based instruction-following evaluation also demonstrates the improvement of our model on both natural images and text-rich images. Through qualitative analysis, LLaVAR shows promising interaction (e.g., reasoning, writing, and elaboration) skills with humans based on the latest real-world online content that combines text and images. We make our code/data/models publicly available at https://llavar.github.io/.","{'model': 'tldr@v2.0.0', 'text': ""This work enhances the current visual instruction tuning pipeline with text-rich images by substantially improves the LLaVA model's capability on text-based VQA datasets and shows promising interaction skills with humans based on the latest real-world online content that combines text and images.""}",http://arxiv.org/pdf/2306.17107
-guiding large language models via directional stimulus prompting,Zekun Li,"We introduce Directional Stimulus Prompting, a novel framework for guiding black-box large language models (LLMs) toward specific desired outputs. Instead of directly adjusting LLMs, our method employs a small tunable policy model (e.g., T5) to generate an auxiliary directional stimulus prompt for each input instance. These directional stimulus prompts act as nuanced, instance-specific hints and clues to guide LLMs in generating desired outcomes, such as including specific keywords in the generated summary. Our approach sidesteps the challenges of direct LLM tuning by optimizing the policy model to explore directional stimulus prompts that align LLMs with desired behaviors. The policy model can be optimized through 1) supervised fine-tuning using labeled data and 2) reinforcement learning from offline or online rewards based on the LLM's output. We assess our method across summarization, dialogue response generation, and chain-of-thought reasoning tasks. Our experiments demonstrate that the framework consistently improves LLMs' (e.g., ChatGPT, Codex, InstructGPT) performance on these supervised tasks using minimal labeled data. Notably, using just 80 dialogues on the MultiWOZ dataset, our approach enhances ChatGPT's performance by an impressive 41.4%, matching or surpassing some fully supervised start-of-the-art models. Additionally, the instance-specific chain-of-thought prompt generated by our approach improves InstructGPT's reasoning accuracy compared to human-crafted or automatically generated prompts. The code and data are publicly available at \url{https://github.com/Leezekun/Directional-Stimulus-Prompting}.","{'model': 'tldr@v2.0.0', 'text': 'Directional Stimulus Prompting, a novel framework for guiding black-box large language models (LLMs) toward specific desired outputs, sidesteps the challenges of direct LLM tuning by optimizing the policy model to explore directional stimulus prompts that align LLMs with desired behaviors.'}",https://arxiv.org/pdf/2302.11520
-motiongpt: finetuned llms are general-purpose motion generators,Yaqi Zhang,"Generating realistic human motion from given action descriptions has experienced significant advancements because of the emerging requirement of digital humans. While recent works have achieved impressive results in generating motion directly from textual action descriptions, they often support only a single modality of the control signal, which limits their application in the real digital human industry. This paper presents a Motion General-Purpose generaTor (MotionGPT) that can use multimodal control signals, e.g., text and single-frame poses, for generating consecutive human motions by treating multimodal signals as special input tokens in large language models (LLMs). Specifically, we first quantize multimodal control signals into discrete codes and then formulate them in a unified prompt instruction to ask the LLMs to generate the motion answer. Our MotionGPT demonstrates a unified human motion generation model with multimodal control signals by tuning a mere 0.4% of LLM parameters. To the best of our knowledge, MotionGPT is the first method to generate human motion by multimodal control signals, which we hope can shed light on this new direction. Codes shall be released upon acceptance.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a Motion General-Purpose generaTor (MotionGPT) that can use multimodal control signals, e.g., text and single-frame poses, for generating consecutive human motions by treating multimmodal signals as special input tokens in large language models (LLMs).'}",http://arxiv.org/pdf/2306.10900
-roco: dialectic multi-robot collaboration with large language models,Zhao Mandi,"We propose a novel approach to multi-robot collaboration that harnesses the power of pre-trained large language models (LLMs) for both high-level communication and low-level path planning. Robots are equipped with LLMs to discuss and collectively reason task strategies. They then generate sub-task plans and task space waypoint paths, which are used by a multi-arm motion planner to accelerate trajectory planning. We also provide feedback from the environment, such as collision checking, and prompt the LLM agents to improve their plan and waypoints in-context. For evaluation, we introduce RoCoBench, a 6-task benchmark covering a wide range of multi-robot collaboration scenarios, accompanied by a text-only dataset for agent representation and reasoning. We experimentally demonstrate the effectiveness of our approach -- it achieves high success rates across all tasks in RoCoBench and adapts to variations in task semantics. Our dialog setup offers high interpretability and flexibility -- in real world experiments, we show RoCo easily incorporates human-in-the-loop, where a user can communicate and collaborate with a robot agent to complete tasks together. See project website https://project-roco.github.io for videos and code.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach to multi-robot collaboration that harnesses the power of pre-trained large language models (LLMs) for both high-level communication and low-level path planning, which shows RoCo easily incorporates human-in-the-loop, where a user can communicate and collaborate with a robot agent to complete tasks together.'}",https://arxiv.org/pdf/2307.04738
-up5: unbiased foundation model for fairness-aware recommendation,Wenyue Hua,"Recent advancements in foundation models such as large language models (LLM) have propelled them to the forefront of recommender systems (RS). Moreover, fairness in RS is critical since many users apply it for decision-making and demand fulfillment. However, at present, there is a lack of understanding regarding the level of fairness exhibited by recommendation foundation models and the appropriate methods for equitably treating different groups of users in foundation models. In this paper, we focus on user-side unfairness problem and show through a thorough examination that there is unfairness involved in LLMs that lead to unfair recommendation results. To eliminate bias from LLM for fairness-aware recommendation, we introduce a novel Unbiased P5 (UP5) foundation model based on Counterfactually-Fair-Prompting (CFP) techniques. CFP includes two sub-modules: a personalized prefix prompt that enhances fairness with respect to individual sensitive attributes, and a Prompt Mixture that integrates multiple counterfactually-fair prompts for a set of sensitive attributes. Experiments are conducted on two real-world datasets, MovieLens-1M and Insurance, and results are compared with both matching-based and sequential-based fairness-aware recommendation models. The results show that UP5 achieves better recommendation performance and meanwhile exhibits a high level of fairness.","{'model': 'tldr@v2.0.0', 'text': 'To eliminate bias from LLM for fairness-aware recommendation, a novel Unbiased P5 (UP5) foundation model based on Counterfactually-Fair-Prompting (CFP) techniques is introduced, which achieves better recommendation performance and meanwhile exhibits a high level of fairness.'}",http://arxiv.org/pdf/2305.12090
-fill in the blank: context-aware automated text input generation for mobile gui testing,Zhe Liu,"Automated GUI testing is widely used to help ensure the quality of mobile apps. However, many GUIs require appropriate text inputs to proceed to the next page, which remains a prominent obstacle for testing coverage. Considering the diversity and semantic requirement of valid inputs (e.g., flight departure, movie name), it is challenging to automate the text input generation. Inspired by the fact that the pre-trained Large Language Model (LLM) has made outstanding progress in text generation, we propose an approach named QTypist based on LLM for intelligently generating semantic input text according to the GUI context. To boost the performance of LLM in the mobile testing scenario, we develop a prompt-based data construction and tuning method which automatically extracts the prompts and answers for model tuning. We evaluate QTypist on 106 apps from Google Play, and the result shows that the passing rate of QTypist is 87%, which is 93% higher than the best baseline. We also integrate QTypist with the automated GUI testing tools and it can cover 42% more app activities, 52% more pages, and subsequently help reveal 122% more bugs compared with the raw tool.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an approach named QTypist based on LLM for intelligently generating semantic input text according to the GUI context which can cover 42% more app activities, 52% more pages, and subsequently help reveal 122% more bugs compared with the raw tool.'}",https://arxiv.org/pdf/2212.04732
-explaining patterns in data with language models via interpretable autoprompting,Chandan Singh,"Large language models (LLMs) have displayed an impressive ability to harness natural language to perform complex tasks. In this work, we explore whether we can leverage this learned ability to find and explain patterns in data. Specifically, given a pre-trained LLM and data examples, we introduce interpretable autoprompting (iPrompt), an algorithm that generates a natural-language string explaining the data. iPrompt iteratively alternates between generating explanations with an LLM and reranking them based on their performance when used as a prompt. Experiments on a wide range of datasets, from synthetic mathematics to natural-language understanding, show that iPrompt can yield meaningful insights by accurately finding groundtruth dataset descriptions. Moreover, the prompts produced by iPrompt are simultaneously human-interpretable and highly effective for generalization: on real-world sentiment classification datasets, iPrompt produces prompts that match or even improve upon human-written prompts for GPT-3. Finally, experiments with an fMRI dataset show the potential for iPrompt to aid in scientific discovery.1","{'model': 'tldr@v2.0.0', 'text': 'This work introduces interpretable autoprompting (iPrompt), an algorithm that generates a natural-language string explaining the data that iteratively alternates between generating explanations with an LLM and reranking them based on their performance when used as a prompt.'}",http://arxiv.org/pdf/2210.01848
-instructzero: efficient instruction optimization for black-box large language models,Lichang Chen,"Large language models~(LLMs) are instruction followers, but it can be challenging to find the best instruction for different situations, especially for black-box LLMs on which backpropagation is forbidden. Instead of directly optimizing the discrete instruction, we optimize a low-dimensional soft prompt applied to an open-source LLM to generate the instruction for the black-box LLM. On each iteration of the proposed method, which we call InstructZero, a soft prompt is converted into an instruction using the open-source LLM, which is then submitted to the black-box LLM for zero-shot evaluation, and the performance is sent to Bayesian optimization to produce new soft prompts improving the zero-shot performance. We evaluate InstructZero on different combinations of open-source LLMs and APIs including Vicuna and ChatGPT. Our results show that InstructZero outperforms SOTA auto-instruction methods across a variety of downstream tasks. Our code and data are publicly available at https://github.com/Lichang-Chen/InstructZero.","{'model': 'tldr@v2.0.0', 'text': 'This work optimized a low-dimensional soft prompt applied to an open-source LLM to generate the instruction for the black-box LLM, showing that InstructZero outperforms SOTA auto-instruction methods across a variety of downstream tasks.'}",https://arxiv.org/pdf/2306.03082
-language models enable simple systems for generating structured views of heterogeneous data lakes,Simran Arora,"A long standing goal of the data management community is to develop general, automated systems that ingest semi-structured documents and output queryable tables without human effort or domain specific customization. Given the sheer variety of potential documents, state-of-the art systems make simplifying assumptions and use domain specific training. In this work, we ask whether we can maintain generality by using large language models (LLMs). LLMs, which are pretrained on broad data, can perform diverse downstream tasks simply conditioned on natural language task descriptions. We propose and evaluate EVAPORATE, a simple, prototype system powered by LLMs. We identify two fundamentally different strategies for implementing this system: prompt the LLM to directly extract values from documents or prompt the LLM to synthesize code that performs the extraction. Our evaluations show a cost-quality tradeoff between these two approaches. Code synthesis is cheap, but far less accurate than directly processing each document with the LLM. To improve quality while maintaining low cost, we propose an extended code synthesis implementation, EVAPORATE-CODE+, which achieves better quality than direct extraction. Our key insight is to generate many candidate functions and ensemble their extractions using weak supervision. EVAPORATE-CODE+ not only outperforms the state-of-the art systems, but does so using a sublinear pass over the documents with the LLM. This equates to a 110x reduction in the number of tokens the LLM needs to process, averaged across 16 real-world evaluation settings of 10k documents each.","{'model': 'tldr@v2.0.0', 'text': 'An extended code synthesis implementation, EVAPORATE-CODE+, is proposed, which achieves better quality than direct extraction and not only outperforms the state-of-the art systems, but does so using a sublinear pass over the documents with the LLM.'}",http://arxiv.org/pdf/2304.09433
-a new era in software security: towards self-healing software via large language models and formal verification,Yiannis Charalambous,"In this paper we present a novel solution that combines the capabilities of Large Language Models (LLMs) with Formal Verification strategies to verify and automatically repair software vulnerabilities. Initially, we employ Bounded Model Checking (BMC) to locate the software vulnerability and derive a counterexample. The counterexample provides evidence that the system behaves incorrectly or contains a vulnerability. The counterexample that has been detected, along with the source code, are provided to the LLM engine. Our approach involves establishing a specialized prompt language for conducting code debugging and generation to understand the vulnerability's root cause and repair the code. Finally, we use BMC to verify the corrected version of the code generated by the LLM. As a proof of concept, we create ESBMC-AI based on the Efficient SMT-based Context-Bounded Model Checker (ESBMC) and a pre-trained Transformer model, specifically gpt-3.5-turbo, to detect and fix errors in C programs. Our experimentation involved generating a dataset comprising 1000 C code samples, each consisting of 20 to 50 lines of code. Notably, our proposed method achieved an impressive success rate of up to 80% in repairing vulnerable code encompassing buffer overflow and pointer dereference failures. We assert that this automated approach can effectively incorporate into the software development lifecycle's continuous integration and deployment (CI/CD) process.","{'model': 'tldr@v2.0.0', 'text': ""A novel solution that combines the capabilities of Large Language Models (LLMs) with Formal Verification strategies to verify and automatically repair software vulnerabilities and it is asserted that this automated approach can effectively incorporate into the software development lifecycle's continuous integration and deployment (CI/CD) process.""}",http://arxiv.org/pdf/2305.14752
-recurrentgpt: interactive generation of (arbitrarily) long text,Wangchunshu Zhou,"The fixed-size context of Transformer makes GPT models incapable of generating arbitrarily long text. In this paper, we introduce RecurrentGPT, a language-based simulacrum of the recurrence mechanism in RNNs. RecurrentGPT is built upon a large language model (LLM) such as ChatGPT and uses natural language to simulate the Long Short-Term Memory mechanism in an LSTM. At each timestep, RecurrentGPT generates a paragraph of text and updates its language-based long-short term memory stored on the hard drive and the prompt, respectively. This recurrence mechanism enables RecurrentGPT to generate texts of arbitrary length without forgetting. Since human users can easily observe and edit the natural language memories, RecurrentGPT is interpretable and enables interactive generation of long text. RecurrentGPT is an initial step towards next-generation computer-assisted writing systems beyond local editing suggestions. In addition to producing AI-generated content (AIGC), we also demonstrate the possibility of using RecurrentGPT as an interactive fiction that directly interacts with consumers. We call this usage of generative models by ``AI As Contents'' (AIAC), which we believe is the next form of conventional AIGC. We further demonstrate the possibility of using RecurrentGPT to create personalized interactive fiction that directly interacts with readers instead of interacting with writers. More broadly, RecurrentGPT demonstrates the utility of borrowing ideas from popular model designs in cognitive science and deep learning for prompting LLMs. Our code is available at https://github.com/aiwaves-cn/RecurrentGPT and an online demo is available at https://www.aiwaves.org/recurrentgpt.","{'model': 'tldr@v2.0.0', 'text': 'RecurrentGPT is a language-based simulacrum of the recurrence mechanism in RNNs that uses natural language to simulate the Long Short-Term Memory mechanism in an LSTM and demonstrates the utility of borrowing ideas from popular model designs in cognitive science and deep learning for prompting LLMs.'}",http://arxiv.org/pdf/2305.13304
-prd: peer rank and discussion improve large language model based evaluations,Ruosen Li,"Nowadays, the quality of responses generated by different modern large language models (LLMs) are hard to evaluate and compare automatically. Recent studies suggest and predominantly use LLMs as a reference-free metric for open-ended question answering. More specifically, they use the recognized""strongest""LLM as the evaluator, which conducts pairwise comparisons of candidate models' answers and provides a ranking score. However, this intuitive method has multiple problems, such as bringing in self-enhancement (favoring its own answers) and positional bias. We draw insights and lessons from the educational domain (Cho and MacArthur, 2011; Walsh, 2014) to improve LLM-based evaluations. Specifically, we propose the (1) peer rank (PR) algorithm that takes into account each peer LLM's pairwise preferences of all answer pairs, and outputs a final ranking of models; and (2) peer discussion (PD), where we prompt two LLMs to discuss and try to reach a mutual agreement on preferences of two answers. We conduct experiments on two benchmark datasets. We find that our approaches achieve higher accuracy and align better with human judgments, respectively. Interestingly, PR can induce a relatively accurate self-ranking of models under the anonymous setting, where each model's name is unrevealed. Our work provides space to explore evaluating models that are hard to compare for humans.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes the (1) peer rank (PR) algorithm that takes into account each peer LLM's pairwise preferences of all answer pairs, and outputs a final ranking of models; and (2) peer discussion (PD), where two LLMs are prompt to discuss and try to reach a mutual agreement on preferences of two answers.""}",https://arxiv.org/pdf/2307.02762
-gpt-3-driven pedagogical agents for training children's curious question-asking skills,Rania Abdelghani,,"{'model': 'tldr@v2.0.0', 'text': 'The efficiency of using large language model (LLM) to support children in generating more curious questions, using a natural language prompting approach that affords usability by teachers and other users not specialists of AI techniques is suggested.'}",https://arxiv.org/pdf/2211.14228
-open sesame! universal black box jailbreaking of large language models,Raz Lapid,"Large language models (LLMs), designed to provide helpful and safe responses, often rely on alignment techniques to align with user intent and social guidelines. Unfortunately, this alignment can be exploited by malicious actors seeking to manipulate an LLM's outputs for unintended purposes. In this paper we introduce a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible. The GA attack works by optimizing a universal adversarial prompt that -- when combined with a user's query -- disrupts the attacked model's alignment, resulting in unintended and potentially harmful outputs. Our novel approach systematically reveals a model's limitations and vulnerabilities by uncovering instances where its responses deviate from expected behavior. Through extensive experiments we demonstrate the efficacy of our technique, thus contributing to the ongoing discussion on responsible AI development by providing a diagnostic tool for evaluating and enhancing alignment of LLMs with human intent. To our knowledge this is the first automated universal black box jailbreak attack.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible, and is the first automated universal black box jailbreak attack.'}",https://arxiv.org/pdf/2309.01446
-what language reveals about perception: distilling psychophysical knowledge from large language models,Raja Marjieh,"Understanding the extent to which the perceptual world can be recovered from language is a fundamental problem in cognitive science. We reformulate this problem as that of distilling psychophysical information from text and show how this can be done by combining large language models (LLMs) with a classic psychophysical method based on similarity judgments. Specifically, we use the prompt auto-completion functional- ity of GPT3, a state-of-the-art LLM, to elicit similarity scores between stimuli and then apply multidimensional scaling to uncover their underlying psychological space. We test our approach on six perceptual domains and show that the elicited judgments strongly correlate with human data and successfully recover well-known psychophysical structures such as the color wheel and pitch spiral. We also explore meaningful divergences between LLM and human representations. Our work showcases how combining state-of-the-art machine models with well-known cognitive paradigms can shed new light on fundamental questions in perception and language research.",,http://arxiv.org/pdf/2302.01308
-large language models can be used to estimate the latent positions of politicians,Patrick Y. Wu,"Existing approaches to estimating politicians' latent positions along specific dimensions often fail when relevant data is limited. We leverage the embedded knowledge in generative large language models (LLMs) to address this challenge and measure lawmakers' positions along specific political or policy dimensions. We prompt an instruction/dialogue-tuned LLM to pairwise compare lawmakers and then scale the resulting graph using the Bradley-Terry model. We estimate novel measures of U.S. senators' positions on liberal-conservative ideology, gun control, and abortion. Our liberal-conservative scale, used to validate LLM-driven scaling, strongly correlates with existing measures and offsets interpretive gaps, suggesting LLMs synthesize relevant data from internet and digitized media rather than memorizing existing measures. Our gun control and abortion measures -- the first of their kind -- differ from the liberal-conservative scale in face-valid ways and predict interest group ratings and legislator votes better than ideology alone. Our findings suggest LLMs hold promise for solving complex social science measurement problems.","{'model': 'tldr@v2.0.0', 'text': 'The liberal-conservative scale, used to validate LLM-driven scaling, strongly correlates with existing measures and offsets interpretive gaps, suggesting LLMs synthesize relevant data from internet and digitized media rather than memorizing existing measures.'}",
-boosting language models reasoning with chain-of-knowledge prompting,J. Wang,"Recently, Chain-of-Thought (CoT) prompting has delivered success on complex reasoning tasks, which aims at designing a simple prompt like ``Let's think step by step'' or multiple in-context exemplars with well-designed rationales to elicit Large Language Models (LLMs) to generate intermediate reasoning steps. However, the generated rationales often come with mistakes, making unfactual and unfaithful reasoning chains. To mitigate this brittleness, we propose a novel Chain-of-Knowledge (CoK) prompting, where we aim at eliciting LLMs to generate explicit pieces of knowledge evidence in the form of structure triple. This is inspired by our human behaviors, i.e., we can draw a mind map or knowledge map as the reasoning evidence in the brain before answering a complex question. Benefiting from CoK, we additionally introduce a F^2-Verification method to estimate the reliability of the reasoning chains in terms of factuality and faithfulness. For the unreliable response, the wrong evidence can be indicated to prompt the LLM to rethink. Extensive experiments demonstrate that our method can further improve the performance of commonsense, factual, symbolic, and arithmetic reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Chain-of-Knowledge prompting, where it aims at eliciting LLMs to generate explicit pieces of knowledge evidence in the form of structure triple, and introduces a F^2-Verification method to estimate the reliability of the reasoning chains in terms of factuality and faithfulness.'}",https://arxiv.org/pdf/2306.06427
-large language models will not replace healthcare professionals: curbing popular fears and hype,A. Thirunavukarasu,"Following the release of ChatGPT, large language models (LLMs) have entered the mainstream. ChatGPT and GPT-4 recently garnered particular attention for attaining expert-level performance in United States Medical Licensing Examinations. However, performance is not perfect, and has not been as impressive in more specialised tests, such as the Membership of the Royal College of General Practitioners Applied Knowledge Test. ChatGPT frequently ‘hallucinates’, providing false, unverified information in the same manner as which it delivers facts. While performance in clinical tasks is expected to improve dramatically with the release of GPT-4, remaining inaccuracy and lack of an uncertainty indicator preclude autonomous deployment of ChatGPT and LLM chatbots like it in clinical settings. LLM applications may nevertheless revolutionise cognitive work – tools such as ChatGPT excel in tasks where specialist knowledge is not required, or is provided by the user prompt: examples include correcting language and rephrasing information for different audiences or within other constraints (e.g. word limits), and it has already been proposed as a tool for administrative tasks, clinical work and patient education. While this does represent an impressive advance in natural language processing, and benefits may be manifold across fields including medicine, these limited use-cases do not live up to the hype surrounding LLMs and artificial intelligence (AI) more generally in 2023. This is due to a fundamental misunderstanding about the form of AI represented by LLMs. Do LLMs represent artificial generalised intelligence (AGI)? The answer is currently probably not, despite emergence of interactive conversational interfaces and few-shot or zero-shot properties – where models execute tasks that they have previously been exposed to only a few times before, or never before, respectively. This is demonstrated by observing how these models are trained, and the composition of their architecture. The backend LLM (GPT-3, from which GPT-3.5 was developed) underpinning older versions of ChatGPT was initially trained on a dataset of billions of words taken from books, Wikipedia and the wider internet. Through a process of machine learning, the GPT-3 accurately encoded the association between individual words in the training dataset. Through ‘reinforcement learning from human feedback’, GPT-3 was subsequently finetuned to provide appropriate responses to users’ queries – producing GPT-3.5. Through these processes, ChatGPT has developed an impressive ability to respond appropriately to diverse prompts, albeit equally lucidly with accurate and inaccurate statements. This lucidity, responsiveness and flexibility have led to sensational claims regarding attainment of AGI that could feasibly replace professionals in cognitive roles. The performance of GPT-4 – which powers newer versions of ChatGPT – dwarfs that of GPT-3.5 across tasks including logical reasoning and medical aptitude tests. Moreover, GPT-4 can be prompted to adopt different roles on demand, and will accept multimodal input, processing images as well as text. Prominent figures in industry and academia have advocated for a moratorium on development of more advanced AI systems in response to concerns regarding safety, ethics and fears of replacement. Despite these fears and hype, the barriers to implementation of LLMs replacing healthcare professionals in any capacity still look out of reach. Although GPT-4’s architecture and training are confidential, it likely relies on similar schemata to its predecessor as it exhibits similar (albeit fewer) hallucinations and reasoning errors, including in medicine. None of ChatGPT’s published autonomous training involved actual comprehension of language in context; the meaning (as we understand it) of words in the dataset was immaterial throughout. While this brute force linguistic processing may prove sufficient to develop a form of AGI, it appears that these LLMs will continue to be afflicted by mistakes and errors. Journal of the Royal Society of Medicine; 2023, Vol. 116(5) 181–182","{'model': 'tldr@v2.0.0', 'text': 'Following the release of ChatGPT, large language models (LLMs) have entered the mainstream and recently garnered particular attention for attaining expert-level performance in United States Medical Licensing Examinations, but performance has not been as impressive in more specialised tests.'}",https://journals.sagepub.com/doi/pdf/10.1177/01410768231173123
-lion: adversarial distillation of proprietary large language models,Yuxin Jiang,"The practice of transferring knowledge from a sophisticated, proprietary large language model (LLM) to a compact, open-source LLM has garnered considerable attention. Previous works have focused on a unidirectional knowledge distillation way by aligning the responses of the student model with those of the teacher model to a set of instructions. Nevertheless, they overlooked the possibility of incorporating any reciprocal""feedback""--identifying challenging instructions where the student model's performance falls short--to boost the student model's proficiency iteratively. To this end, we propose a novel adversarial distillation framework for a more efficient knowledge transfer. Leveraging the versatile role adaptability of LLMs, we prompt the teacher model to identify""hard""instructions and generate new""hard""instructions for the student model, creating a three-stage adversarial loop of imitation, discrimination, and generation. By applying this adversarial framework, we successfully transfer knowledge from ChatGPT to a student model (named Lion), using a mere 70k training data. Our results show that Lion-13B not only achieves comparable open-ended generation capabilities to ChatGPT but surpasses conventional state-of-the-art (SOTA) instruction-tuned models like Vicuna-13B by 55.4% in challenging zero-shot reasoning benchmarks such as BIG-Bench Hard (BBH) and 16.7% on AGIEval. Code and model can be found at https://github.com/YJiangcm/Lion.","{'model': 'tldr@v2.0.0', 'text': 'Lion-13B not only achieves comparable open-ended generation capabilities to ChatGPT but surpasses conventional state-of-the-art instruction-tuned models like Vicuna- 13B by 55.4% in challenging zero-shot reasoning benchmarks such as BIG-Bench Hard (BBH) and 16.7% on AGIEval.'}",
-surreal vr pong: llm approach to game design,Jasmine Roberts,"The increase in complexity from 2D to 3D game design makes it fascinating to study from a computational creativity perspective. Generating images given text descriptions using models like DALL-E has recently increased in popularity. However, these models are limited to generating 2-dimensional outputs. While outputs of these models can be used to stylize 3d objects with variable textures, they cannot produce mesh-level interactions. We introduce Codex VR Pong as a demonstration of controlled non-deterministic game mechanics leveraging generative models. We are proposing that prompt-based creation can become part of gameplay rather than just part of game development","{'model': 'tldr@v2.0.0', 'text': 'This work proposes that prompt-based creation can become part of gameplay rather than just part of game development, and introduces Codex VR Pong as a demonstration of controlled non-deterministic game mechanics leveraging generative models.'}",
-towards parameter-efficient automation of data wrangling tasks with prefix-tuning,David Vos,"Data wrangling tasks for data integration and cleaning arise in virtually every data-driven application scenario nowadays. Recent research indicated the astounding potential of Large Language Models (LLMs) for such tasks. However, the automation of data wrangling with LLMs poses additional challenges, as hand-tuning task- and data-specific prompts for LLMs requires high expertise and manual effort. On the other hand, finetuning a whole LLM is more amenable to automation, but incurs high storage costs, as a copy of the LLM has to be maintained. In this work, we explore the potential of a lightweight alternative to finetuning an LLM, which automatically learns a continuous prompt. This approach called prefix-tuning does not require updating the original LLM parameters, and can therefore re-use a single LLM instance across tasks. At the same time, it is amenable to automation, as continuous prompts can be automatically learned with standard techniques. We evaluate prefix-tuning on common data wrangling tasks for tabular data such as entity matching, error detection, and data imputation, with promising results. We find that in five out of ten cases, prefix-tuning is within 2.3% of the performance of finetuning, even though it leverages only 0.39% of the parameter updates required for finetuning the full model. These results highlight the potential of prefix-tuning as a parameter-efficient alternative to finetuning for data integration and data cleaning with LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the potential of a lightweight alternative to finetuning an LLM, which automatically learns a continuous prompt, called prefix-tuning, which does not require updating the original LLM parameters, and can therefore re-use a single LLM instance across tasks.'}",
-herding ai cats: lessons from designing a chatbot by prompting gpt-3,J. Zamfirescu-Pereira,"Prompting Large Language Models (LLMs) is an exciting new approach to designing chatbots. But can it improve LLM’s user experience (UX) reliably enough to power chatbot products? Our attempt to design a robust chatbot by prompting GPT-3/4 alone suggests: not yet. Prompts made achieving “80%” UX goals easy, but not the remaining 20%. Fixing the few remaining interaction breakdowns resembled herding cats: We could not address one UX issue or test one design solution at a time; instead, we had to handle everything everywhere all at once. Moreover, because no prompt could make GPT reliably say “I don’t know” when it should, the user-GPT conversations had no guardrails after a breakdown occurred, often leading to UX downward spirals. These risks incentivized us to design highly prescriptive prompts and scripted bots, counter to the promises of LLM-powered chatbots. This paper describes this case study, unpacks prompting’s fickleness and its impact on UX design processes, and discusses implications for LLM-based design methods and tools.",,https://dl.acm.org/doi/pdf/10.1145/3563657.3596138
-exploring large language model for graph data understanding in online job recommendations,Likang Wu,"Large Language Models (LLMs) have revolutionized natural language processing tasks, demonstrating their exceptional capabilities in various domains. However, their potential for behavior graph understanding in job recommendations remains largely unexplored. This paper focuses on unveiling the capability of large language models in understanding behavior graphs and leveraging this understanding to enhance recommendations in online recruitment, including the promotion of out-of-distribution (OOD) application. We present a novel framework that harnesses the rich contextual information and semantic representations provided by large language models to analyze behavior graphs and uncover underlying patterns and relationships. Specifically, we propose a meta-path prompt constructor that leverages LLM recommender to understand behavior graphs for the first time and design a corresponding path augmentation module to alleviate the prompt bias introduced by path-based sequence input. By leveraging this capability, our framework enables personalized and accurate job recommendations for individual users. We evaluate the effectiveness of our approach on a comprehensive dataset and demonstrate its ability to improve the relevance and quality of recommended quality. This research not only sheds light on the untapped potential of large language models but also provides valuable insights for developing advanced recommendation systems in the recruitment market. The findings contribute to the growing field of natural language processing and offer practical implications for enhancing job search experiences.","{'model': 'tldr@v2.0.0', 'text': 'A meta-path prompt constructor is proposed that leverages LLM recommender to understand behavior graphs for the first time and a corresponding path augmentation module is designed to alleviate the prompt bias introduced by path-based sequence input.'}",https://arxiv.org/pdf/2307.05722
-fine-tuning multimodal llms to follow zero-shot demonstrative instructions,Juncheng Li,"Recent advancements in Multimodal Large Language Models (MLLMs) have been utilizing Visual Prompt Generators (VPGs) to convert visual features into tokens that LLMs can recognize. This is achieved by training the VPGs on millions of image-caption pairs, where the VPG-generated tokens of images are fed into a frozen LLM to generate the corresponding captions. However, this image-captioning based training objective inherently biases the VPG to concentrate solely on the primary visual contents sufficient for caption generation, often neglecting other visual details. This shortcoming results in MLLMs' underperformance in comprehending demonstrative instructions consisting of multiple, interleaved, and multimodal instructions that demonstrate the required context to complete a task. To address this issue, we introduce a generic and lightweight Visual Prompt Generator Complete module (VPG-C), which can infer and complete the missing details essential for comprehending demonstrative instructions. Further, we propose a synthetic discriminative training strategy to fine-tune VPG-C, eliminating the need for supervised demonstrative instructions. As for evaluation, we build DEMON, a comprehensive benchmark for demonstrative instruction understanding. Synthetically trained with the proposed strategy, VPG-C achieves significantly stronger zero-shot performance across all tasks of DEMON. Further evaluation on the MME and OwlEval benchmarks also demonstrate the superiority of VPG-C. Our benchmark, code, and pre-trained models are available at https://github.com/DCDmllm/Cheetah.","{'model': 'tldr@v2.0.0', 'text': 'A generic and lightweight Visual Prompt Generator Complete module (VPG-C) is introduced, which can infer and complete the missing details essential for comprehending demonstrative instructions, and a synthetic discriminative training strategy is proposed to fine-tune V PG-C, eliminating the need for supervised demonstrative instruction understanding.'}",
-ten quick tips for harnessing the power of chatgpt/gpt-4 in computational biology,Tiago Lubiana,"The rise of advanced chatbots, such as ChatGPT, has sparked curiosity in the scientific community. ChatGPT is a general-purpose chatbot powered by large language models (LLMs) GPT-3.5 and GPT-4, with the potential to impact numerous fields, including computational biology. In this article, we offer ten tips based on our experience with ChatGPT to assist computational biologists in optimizing their workflows. We have collected relevant prompts and reviewed the nascent literature in the field, compiling tips we project to remain pertinent for future ChatGPT and LLM iterations, ranging from code refactoring to scientific writing to prompt engineering. We hope our work will help bioinformaticians to complement their workflows while staying aware of the various implications of using this technology. Additionally, to track new and creative applications for bioinformatics tools such as ChatGPT, we have established a GitHub repository at https://github.com/csbl-br/awesome-compbio-chatgpt. Our belief is that ethical adherence to ChatGPT and other LLMs will increase the efficiency of computational biologists, ultimately advancing the pace of scientific discovery in the life sciences.","{'model': 'tldr@v2.0.0', 'text': 'It is believed that ethical adherence to ChatGPT and other LLMs will increase the efficiency of computational biologists, ultimately advancing the pace of scientific discovery in the life sciences.'}",
-reconcile: round-table conference improves reasoning via consensus among diverse llms,Justin Chih-Yao Chen,"Large Language Models (LLMs) still struggle with complex reasoning tasks. Motivated by the society of minds (Minsky, 1988), we propose ReConcile, a multi-model multi-agent framework designed as a round table conference among diverse LLM agents to foster diverse thoughts and discussion for improved consensus. ReConcile enhances the reasoning capabilities of LLMs by holding multiple rounds of discussion, learning to convince other agents to improve their answers, and employing a confidence-weighted voting mechanism. In each round, ReConcile initiates discussion between agents via a 'discussion prompt' that consists of (a) grouped answers and explanations generated by each agent in the previous round, (b) their uncertainties, and (c) demonstrations of answer-rectifying human explanations, used for convincing other agents. This discussion prompt enables each agent to revise their responses in light of insights from other agents. Once a consensus is reached and the discussion ends, ReConcile determines the final answer by leveraging the confidence of each agent in a weighted voting scheme. We implement ReConcile with ChatGPT, Bard, and Claude2 as the three agents. Our experimental results on various benchmarks demonstrate that ReConcile significantly enhances the reasoning performance of the agents (both individually and as a team), surpassing prior single-agent and multi-agent baselines by 7.7% and also outperforming GPT-4 on some of these datasets. We also experiment with GPT-4 itself as one of the agents in ReConcile and demonstrate that its initial performance also improves by absolute 10.0% through discussion and feedback from other agents. Finally, we also analyze the accuracy after every round and observe that ReConcile achieves better and faster consensus between agents, compared to a multi-agent debate baseline. Our code is available at: https://github.com/dinobby/ReConcile","{'model': 'tldr@v2.0.0', 'text': 'ReConcile significantly enhances the reasoning performance of the agents (both individually and as a team), surpassing prior single-agent and multi-agent baselines by 7.7% and also outperforming GPT-4 on some of these datasets.'}",https://arxiv.org/pdf/2309.13007
-prompting large language models for zero-shot domain adaptation in speech recognition,Yuang Li,"The integration of Language Models (LMs) has proven to be an effective way to address domain shifts in speech recognition. However, these approaches usually require a significant amount of target domain text data for the training of LMs. Different from these methods, in this work, with only a domain-specific text prompt, we propose two zero-shot ASR domain adaptation methods using LLaMA, a 7-billion-parameter large language model (LLM). LLM is used in two ways: 1) second-pass rescoring: reranking N-best hypotheses of a given ASR system with LLaMA; 2) deep LLM-fusion: incorporating LLM into the decoder of an encoder-decoder based ASR system. Experiments show that, with only one domain prompt, both methods can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGISpeech datasets. Especially, the deep LLM-fusion has the advantage of better recall of entity and out-of-vocabulary words.","{'model': 'tldr@v2.0.0', 'text': 'Two zero-shot ASR domain adaptation methods using LLaMA, a 7-billion-parameter large language model (LLM), which can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGISpeech datasets.'}",http://arxiv.org/pdf/2306.16007
-prompting is not a substitute for probability measurements in large language models,Jennifer Hu,"Prompting is now a dominant method for evaluating the linguistic knowledge of large language models (LLMs). While other methods directly read out models' probability distributions over strings, prompting requires models to access this internal information by processing linguistic input, thereby implicitly testing a new type of emergent ability: metalinguistic judgment. In this study, we compare metalinguistic prompting and direct probability measurements as ways of measuring models' linguistic knowledge. Broadly, we find that LLMs' metalinguistic judgments are inferior to quantities directly derived from representations. Furthermore, consistency gets worse as the prompt query diverges from direct measurements of next-word probabilities. Our findings suggest that negative results relying on metalinguistic prompts cannot be taken as conclusive evidence that an LLM lacks a particular linguistic generalization. Our results also highlight the value that is lost with the move to closed APIs where access to probability distributions is limited.",,
-automated annotation with generative ai requires validation,Nicholas Pangakis,"Generative large language models (LLMs) can be a powerful tool for augmenting text annotation procedures, but their performance varies across annotation tasks due to prompt quality, text data idiosyncrasies, and conceptual difficulty. Because these challenges will persist even as LLM technology improves, we argue that any automated annotation process using an LLM must validate the LLM's performance against labels generated by humans. To this end, we outline a workflow to harness the annotation potential of LLMs in a principled, efficient way. Using GPT-4, we validate this approach by replicating 27 annotation tasks across 11 datasets from recent social science articles in high-impact journals. We find that LLM performance for text annotation is promising but highly contingent on both the dataset and the type of annotation task, which reinforces the necessity to validate on a task-by-task basis. We make available easy-to-use software designed to implement our workflow and streamline the deployment of LLMs for automated annotation.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that any automated annotation process using an LLM must validate the LLM's performance against labels generated by humans, and an outline of a workflow to harness the annotation potential of LLMs in a principled, efficient way is outlined.""}",http://arxiv.org/pdf/2306.00176
-statler: state-maintaining language models for embodied reasoning,Takuma Yoneda,"Large language models (LLMs) provide a promising tool that enable robots to perform complex robot reasoning tasks. However, the limited context window of contemporary LLMs makes reasoning over long time horizons difficult. Embodied tasks such as those that one might expect a household robot to perform typically require that the planner consider information acquired a long time ago (e.g., properties of the many objects that the robot previously encountered in the environment). Attempts to capture the world state using an LLM's implicit internal representation is complicated by the paucity of task- and environment-relevant information available in a robot's action history, while methods that rely on the ability to convey information via the prompt to the LLM are subject to its limited context window. In this paper, we propose Statler, a framework that endows LLMs with an explicit representation of the world state as a form of ``memory'' that is maintained over time. Integral to Statler is its use of two instances of general LLMs -- a world-model reader and a world-model writer -- that interface with and maintain the world state. By providing access to this world state ``memory'', Statler improves the ability of existing LLMs to reason over longer time horizons without the constraint of context length. We evaluate the effectiveness of our approach on three simulated table-top manipulation domains and a real robot domain, and show that it improves the state-of-the-art in LLM-based robot reasoning. Project website: https://statler-lm.github.io/","{'model': 'tldr@v2.0.0', 'text': ""Statler is a framework that endows LLMs with an explicit representation of the world state as a form of ``memory'' that is maintained over time, and improves the ability of existing LLMs to reason over longer time horizons without the constraint of context length.""}",http://arxiv.org/pdf/2306.17840
-studenteval: a benchmark of student-written prompts for large language models of code,Hannah McLean Babe,"Code LLMs are being rapidly deployed and there is evidence that they can make professional programmers more productive. Current benchmarks for code generation measure whether models generate correct programs given an expert prompt. In this paper, we present a new benchmark containing multiple prompts per problem, written by a specific population of non-expert prompters: beginning programmers. StudentEval contains 1,749 prompts for 48 problems, written by 80 students who have only completed one semester of Python programming. Our students wrote these prompts while working interactively with a Code LLM, and we observed very mixed success rates. We use StudentEval to evaluate 5 Code LLMs and find that StudentEval is a better discriminator of model performance than existing benchmarks. We analyze the prompts and find significant variation in students' prompting techniques. We also find that nondeterministic LLM sampling could mislead students into thinking that their prompts are more (or less) effective than they actually are, which has implications for how to teach with Code LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new benchmark containing multiple prompts per problem, written by a specific population of non-expert prompters: beginning programmers, and finds that StudentEval is a better discriminator of model performance than existing benchmarks.'}",http://arxiv.org/pdf/2306.04556
-mindmap: knowledge graph prompting sparks graph of thoughts in large language models,Yilin Wen,"LLMs usually exhibit limitations in their ability to incorporate new knowledge, the generation of hallucinations, and the transparency of their decision-making process. In this paper, we explore how to prompt LLMs with knowledge graphs (KG), working as a remedy to engage LLMs with up-to-date knowledge and elicit the reasoning pathways from LLMs. Specifically, we build a prompting pipeline that endows LLMs with the capability of comprehending KG inputs and inferring with a combined implicit knowledge and the retrieved external knowledge. In addition, we investigate eliciting the mind map on which LLMs perform the reasoning and generate the answers. It is identified that the produced mind map exhibits the reasoning pathways of LLMs grounded on the ontology of knowledge, hence bringing the prospects of probing and gauging LLM inference in production. The experiments on three question&answering datasets also show that MindMap prompting leads to a striking empirical gain. For instance, prompting a GPT-3.5 with MindMap yields an overwhelming performance over GPT-4 consistently. We also demonstrate that with structured facts retrieved from KG, MindMap can outperform a series of prompting-with-document-retrieval methods, benefiting from more accurate, concise, and comprehensive knowledge from KGs. To reproduce our results and extend the framework further, we make our codebase available at https://github.com/wyl.willing/MindMap.","{'model': 'tldr@v2.0.0', 'text': 'This paper builds a prompting pipeline that endows LLMs with the capability of comprehending KG inputs and inferring with a combined implicit knowledge and the retrieved external knowledge, and demonstrates that with structured facts retrieved from KG, MindMap can outperform a series of prompting-with-document-retrieval methods, benefiting from more accurate, concise, and comprehensive knowledge from KGs.'}",https://arxiv.org/pdf/2308.09729
-progprompt: program generation for situated robot task planning using large language models,Ishika Singh,,"{'model': 'tldr@v2.0.0', 'text': 'This work presents a programmatic LLM prompt structure that enables plan generation functional across situated environments, robot capabilities, and tasks, and makes concrete recommendations about prompt structure and generation constraints through ablation experiments.'}",https://link.springer.com/content/pdf/10.1007/s10514-023-10135-3.pdf
-interactive data synthesis for systematic vision adaptation via llms-aigcs collaboration,Qifan Yu,"Recent text-to-image generation models have shown promising results in generating high-fidelity photo-realistic images. In parallel, the problem of data scarcity has brought a growing interest in employing AIGC technology for high-quality data expansion. However, this paradigm requires well-designed prompt engineering that cost-less data expansion and labeling remain under-explored. Inspired by LLM's powerful capability in task guidance, we propose a new paradigm of annotated data expansion named as ChatGenImage. The core idea behind it is to leverage the complementary strengths of diverse models to establish a highly effective and user-friendly pipeline for interactive data augmentation. In this work, we extensively study how LLMs communicate with AIGC model to achieve more controllable image generation and make the first attempt to collaborate them for automatic data augmentation for a variety of downstream tasks. Finally, we present fascinating results obtained from our ChatGenImage framework and demonstrate the powerful potential of our synthetic data for systematic vision adaptation. Our codes are available at https://github.com/Yuqifan1117/Labal-Anything-Pipeline.","{'model': 'tldr@v2.0.0', 'text': 'This work extensively study how LLMs communicate with AIGC model to achieve more controllable image generation and makes the first attempt to collaborate them for automatic data augmentation for a variety of downstream tasks.'}",http://arxiv.org/pdf/2305.12799
-examining the validity of chatgpt in identifying relevant nephrology literature: findings and implications,S. Suppadungsuk,"Literature reviews are valuable for summarizing and evaluating the available evidence in various medical fields, including nephrology. However, identifying and exploring the potential sources requires focus and time devoted to literature searching for clinicians and researchers. ChatGPT is a novel artificial intelligence (AI) large language model (LLM) renowned for its exceptional ability to generate human-like responses across various tasks. However, whether ChatGPT can effectively assist medical professionals in identifying relevant literature is unclear. Therefore, this study aimed to assess the effectiveness of ChatGPT in identifying references to literature reviews in nephrology. We keyed the prompt “Please provide the references in Vancouver style and their links in recent literature on… name of the topic” into ChatGPT-3.5 (03/23 Version). We selected all the results provided by ChatGPT and assessed them for existence, relevance, and author/link correctness. We recorded each resource’s citations, authors, title, journal name, publication year, digital object identifier (DOI), and link. The relevance and correctness of each resource were verified by searching on Google Scholar. Of the total 610 references in the nephrology literature, only 378 (62%) of the references provided by ChatGPT existed, while 31% were fabricated, and 7% of citations were incomplete references. Notably, only 122 (20%) of references were authentic. Additionally, 256 (68%) of the links in the references were found to be incorrect, and the DOI was inaccurate in 206 (54%) of the references. Moreover, among those with a link provided, the link was correct in only 20% of cases, and 3% of the references were irrelevant. Notably, an analysis of specific topics in electrolyte, hemodialysis, and kidney stones found that >60% of the references were inaccurate or misleading, with less reliable authorship and links provided by ChatGPT. Based on our findings, the use of ChatGPT as a sole resource for identifying references to literature reviews in nephrology is not recommended. Future studies could explore ways to improve AI language models’ performance in identifying relevant nephrology literature.","{'model': 'tldr@v2.0.0', 'text': 'The use of ChatGPT as a sole resource for identifying references to literature reviews in nephrology is not recommended and future studies could explore ways to improve AI language models’ performance in identifying relevant nephrologic literature.'}",https://www.mdpi.com/2077-0383/12/17/5550/pdf?version=1692975950
-clusterllm: large language models as a guide for text clustering,Yuwei Zhang,"We introduce ClusterLLM, a novel text clustering framework that leverages feedback from an instruction-tuned large language model, such as ChatGPT. Compared with traditional unsupervised methods that builds upon""small""embedders, ClusterLLM exhibits two intriguing advantages: (1) it enjoys the emergent capability of LLM even if its embeddings are inaccessible; and (2) it understands the user's preference on clustering through textual instruction and/or a few annotated data. First, we prompt ChatGPT for insights on clustering perspective by constructing hard triplet questions, where A, B and C are similar data points that belong to different clusters according to small embedder. We empirically show that this strategy is both effective for fine-tuning small embedder and cost-efficient to query ChatGPT. Second, we prompt ChatGPT for helps on clustering granularity by carefully designed pairwise questions, and tune the granularity from cluster hierarchies that is the most consistent with the ChatGPT answers. Extensive experiments on 14 datasets show that ClusterLLM consistently improves clustering quality, at an average cost of ~$0.6 per dataset.","{'model': 'tldr@v2.0.0', 'text': 'ClusterLLM, a novel text clustering framework that leverages feedback from an instruction-tuned large language model, such as ChatGPT, consistently improves clustering quality, at an average cost of ~$0.6 per dataset.'}",http://arxiv.org/pdf/2305.14871
-how to unleash the power of large language models for few-shot relation extraction?,Xin Xu,"Scaling language models have revolutionized widespread NLP tasks, yet little comprehensively explored few-shot relation extraction with large language models. In this paper, we investigate principal methodologies, in-context learning and data generation, for few-shot relation extraction via GPT-3.5 through exhaustive experiments. To enhance few-shot performance, we further propose task-related instructions and schema-constrained data generation. We observe that in-context learning can achieve performance on par with previous prompt learning approaches, and data generation with the large language model can boost previous solutions to obtain new state-of-the-art few-shot results on four widely-studied relation extraction datasets. We hope our work can inspire future research for the capabilities of large language models in few-shot relation extraction. Code is available in https://github.com/zjunlp/DeepKE/tree/main/example/llm.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that in-context learning can achieve performance on par with previous prompt learning approaches, and data generation with the large language model can boost previous solutions to obtain new state-of-the-art few-shot results on four widely-studied relation extraction datasets.'}",http://arxiv.org/pdf/2305.01555
-"large language models and simple, stupid bugs",Kevin Jesse,"With the advent of powerful neural language models, AI-based systems to assist developers in coding tasks are becoming widely available; Copilot is one such system. Copilot uses Codex, a large language model (LLM), to complete code conditioned on a preceding ""prompt"". Codex, however, is trained on public GitHub repositories, viz., on code that may include bugs and vulnerabilities. Previous studies [1], [2] show Codex reproduces vulnerabilities seen in training. In this study, we examine how prone Codex is to generate an interesting bug category, single statement bugs, commonly referred to as simple, stupid bugs or SStuBs in the MSR community. We find that Codex and similar LLMs do help avoid some SStuBs, but do produce known, verbatim SStuBs as much as 2x as likely than known, verbatim correct code. We explore the consequences of the Codex generated SStuBs and propose avoidance strategies that suggest the possibility of reducing the production of known, verbatim SStubs, and increase the possibility of producing known, verbatim fixes.","{'model': 'tldr@v2.0.0', 'text': 'This study examines how prone Codex is to generate an interesting bug category, single statement bugs, commonly referred to as simple, stupid bugs or SStuBs in the MSR community and proposes avoidance strategies that suggest the possibility of reducing the production of known, verbatim SStubs, and increase theossibility of producing known,Verbatim fixes.'}",https://arxiv.org/pdf/2303.11455
-systematic rectification of language models via dead-end analysis,Mengyao Cao,"With adversarial or otherwise normal prompts, existing large language models (LLM) can be pushed to generate toxic discourses. One way to reduce the risk of LLMs generating undesired discourses is to alter the training of the LLM. This can be very restrictive due to demanding computation requirements. Other methods rely on rule-based or prompt-based token elimination, which are limited as they dismiss future tokens and the overall meaning of the complete discourse. Here, we center detoxification on the probability that the finished discourse is ultimately considered toxic. That is, at each point, we advise against token selections proportional to how likely a finished text from this point will be toxic. To this end, we formally extend the dead-end theory from the recent reinforcement learning (RL) literature to also cover uncertain outcomes. Our approach, called rectification, utilizes a separate but significantly smaller model for detoxification, which can be applied to diverse LLMs as long as they share the same vocabulary. Importantly, our method does not require access to the internal representations of the LLM, but only the token probability distribution at each decoding step. This is crucial as many LLMs today are hosted in servers and only accessible through APIs. When applied to various LLMs, including GPT-3, our approach significantly improves the generated discourse compared to the base LLMs and other techniques in terms of both the overall language and detoxification performance.","{'model': 'tldr@v2.0.0', 'text': 'The dead-end theory from the recent reinforcement learning (RL) literature is formally extended to also cover uncertain outcomes and the approach, called rectification, utilizes a separate but significantly smaller model for detoxification, which can be applied to diverse LLMs as long as they share the same vocabulary.'}",http://arxiv.org/pdf/2302.14003
-knowledge refinement via interaction between search engines and large language models,Jiazhan Feng,"Information retrieval (IR) plays a crucial role in locating relevant resources from vast amounts of data, and its applications have evolved from traditional knowledge bases to modern search engines (SEs). The emergence of large language models (LLMs) has further revolutionized the IR field by enabling users to interact with search systems in natural language. In this paper, we explore the advantages and disadvantages of LLMs and SEs, highlighting their respective strengths in understanding user-issued queries and retrieving up-to-date information. To leverage the benefits of both paradigms while circumventing their limitations, we propose InteR, a novel framework that facilitates knowledge refinement through interaction between SEs and LLMs. InteR allows SEs to expand knowledge in queries using LLM-generated knowledge collections and enables LLMs to enhance prompt formulation using SE-retrieved documents. This iterative refinement process augments the inputs of SEs and LLMs, leading to more accurate retrieval. Experiments on large-scale retrieval benchmarks involving web search and low-resource retrieval tasks demonstrate that InteR achieves overall superior zero-shot retrieval performance compared to state-of-the-art methods, even those using relevance judgment. Source code is available at https://github.com/Cyril-JZ/InteR","{'model': 'tldr@v2.0.0', 'text': 'InteR is proposed, a novel framework that facilitates knowledge refinement through interaction between SEs and LLMs, which allows SEs to expand knowledge in queries using LLM-generated knowledge collections and enables LLMs to enhance prompt formulation using SE-retrieved documents.'}",http://arxiv.org/pdf/2305.07402
-introspective tips: large language model for in-context decision making,Liting Chen,"The emergence of large language models (LLMs) has substantially influenced natural language processing, demonstrating exceptional results across various tasks. In this study, we employ ``Introspective Tips""to facilitate LLMs in self-optimizing their decision-making. By introspectively examining trajectories, LLM refines its policy by generating succinct and valuable tips. Our method enhances the agent's performance in both few-shot and zero-shot learning situations by considering three essential scenarios: learning from the agent's past experiences, integrating expert demonstrations, and generalizing across diverse games. Importantly, we accomplish these improvements without fine-tuning the LLM parameters; rather, we adjust the prompt to generalize insights from the three aforementioned situations. Our framework not only supports but also emphasizes the advantage of employing LLM in in-contxt decision-making. Experiments involving over 100 games in TextWorld illustrate the superior performance of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This study employs ``Introspective Tips"" to facilitate LLMs in self-optimizing their decision-making by introspectively examining trajectories, which enhances the agent\'s performance in both few-shot and zero-shot learning situations.'}",http://arxiv.org/pdf/2305.11598
-augmenting greybox fuzzing with generative ai,Jie Hu,"Real-world programs expecting structured inputs often has a format-parsing stage gating the deeper program space. Neither a mutation-based approach nor a generative approach can provide a solution that is effective and scalable. Large language models (LLM) pre-trained with an enormous amount of natural language corpus have proved to be effective for understanding the implicit format syntax and generating format-conforming inputs. In this paper, propose ChatFuzz, a greybox fuzzer augmented by generative AI. More specifically, we pick a seed in the fuzzer's seed pool and prompt ChatGPT generative models to variations, which are more likely to be format-conforming and thus of high quality. We conduct extensive experiments to explore the best practice for harvesting the power of generative LLM models. The experiment results show that our approach improves the edge coverage by 12.77\% over the SOTA greybox fuzzer (AFL++) on 12 target programs from three well-tested benchmarks. As for vulnerability detection, \sys is able to perform similar to or better than AFL++ for programs with explicit syntax rules but not for programs with non-trivial syntax.","{'model': 'tldr@v2.0.0', 'text': ""ChatFuzz is proposed, a greybox fuzzer augmented by generative AI that picks a seed in the fuzzer's seed pool and prompts ChatGPT generative models to variations, which are more likely to be format-conforming and thus of high quality.""}",http://arxiv.org/pdf/2306.06782
-chatrule: mining logical rules with large language models for knowledge graph reasoning,Linhao Luo,"Logical rules are essential for uncovering the logical connections between relations, which could improve the reasoning performance and provide interpretable results on knowledge graphs (KGs). Although there have been many efforts to mine meaningful logical rules over KGs, existing methods suffer from the computationally intensive searches over the rule space and a lack of scalability for large-scale KGs. Besides, they often ignore the semantics of relations which is crucial for uncovering logical connections. Recently, large language models (LLMs) have shown impressive performance in the field of natural language processing and various applications, owing to their emergent ability and generalizability. In this paper, we propose a novel framework, ChatRule, unleashing the power of large language models for mining logical rules over knowledge graphs. Specifically, the framework is initiated with an LLM-based rule generator, leveraging both the semantic and structural information of KGs to prompt LLMs to generate logical rules. To refine the generated rules, a rule ranking module estimates the rule quality by incorporating facts from existing KGs. Last, a rule validator harnesses the reasoning ability of LLMs to validate the logical correctness of ranked rules through chain-of-thought reasoning. ChatRule is evaluated on four large-scale KGs, w.r.t. different rule quality metrics and downstream tasks, showing the effectiveness and scalability of our method.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, ChatRule, unleashing the power of large language models for mining logical rules over knowledge graphs, initiated with an LLM-based rule generator, leveraging both the semantic and structural information of KGs to prompt LLMs to generate logical rules.'}",https://arxiv.org/pdf/2309.01538
-zero-shot prompting for code complexity prediction using github copilot,Mohammed Latif Siddiq,"Code generation models are gaining popularity because they can produce correct code from a prompt, speeding up the software development process. GitHub Copilot is currently one of the most commonly used tools for code generation. This tool is based on GPT3, a Large Language Model (LLM), and can perform zero-shot prompting tasks i.e., tasks for which the model is not specifically trained. In this paper, we describe a preliminary study that investigates whether GitHub Copilot can predict the runtime complexity of a given program using zero- shot prompting. In our study, we found that GitHub Copilot can correctly predict the runtime complexity 45.44% times in the first suggestion and 56.38 % times considering all suggestions. We also compared Copilot to other machine learning, neural network, and transformer-based approaches for code complexity prediction. We observed that Copilot outperformed other approaches for predicting code with linear complexity $\mathbf{O}(n)$.","{'model': 'tldr@v2.0.0', 'text': 'This study found that GitHub Copilot can correctly predict the runtime complexity 45.44% times in the first suggestion and 56.38 % times considering all suggestions, and observed that Copilot outperformed other approaches for predicting code with linear complexity $\\mathbf{O}(n)$.'}",
-taming ai bots: controllability of neural states in large language models,S. Soatto,"We tackle the question of whether an agent can, by suitable choice of prompts, control an AI bot to any state. To that end, we first introduce a formal definition of ``meaning'' that is amenable to analysis. Then, we characterize ``meaningful data'' on which large language models (LLMs) are ostensibly trained, and ``well-trained LLMs'' through conditions that are largely met by today's LLMs. While a well-trained LLM constructs an embedding space of meanings that is Euclidean, meanings themselves do not form a vector (linear) subspace, but rather a quotient space within. We then characterize the subset of meanings that can be reached by the state of the LLMs for some input prompt, and show that a well-trained bot can reach any meaning albeit with small probability. We then introduce a stronger notion of controllability as {\em almost certain reachability}, and show that, when restricted to the space of meanings, an AI bot is controllable. We do so after introducing a functional characterization of attentive AI bots, and finally derive necessary and sufficient conditions for controllability. The fact that AI bots are controllable means that an adversary could steer them towards any state. However, the sampling process can be designed to counteract adverse actions and avoid reaching undesirable regions of state space before their boundary is crossed.",,http://arxiv.org/pdf/2305.18449
-spellburst: a node-based interface for exploratory creative coding with natural language prompts,Tyler Angert,"Creative coding tasks are often exploratory in nature. When producing digital artwork, artists usually begin with a high-level semantic construct such as a “stained glass filter” and programmatically implement it by varying code parameters such as shape, color, lines, and opacity to produce visually appealing results. Based on interviews with artists, it can be effortful to translate semantic constructs to program syntax, and current programming tools don’t lend well to rapid creative exploration. To address these challenges, we introduce Spellburst, a large language model (LLM) powered creative-coding environment. Spellburst provides (1) a node-based interface that allows artists to create generative art and explore variations through branching and merging operations, (2) expressive prompt-based interactions to engage in semantic programming, and (3) dynamic prompt-driven interfaces and direct code editing to seamlessly switch between semantic and syntactic exploration. Our evaluation with artists demonstrates Spellburst’s potential to enhance creative coding practices and inform the design of computational creativity tools that bridge semantic and syntactic spaces.","{'model': 'tldr@v2.0.0', 'text': 'Spellburst provides a node-based interface that allows artists to create generative art and explore variations through branching and merging operations, expressive prompt-based interactions to engage in semantic programming, and dynamic prompt-driven interfaces and direct code editing to seamlessly switch between semantic and syntactic exploration.'}",https://arxiv.org/pdf/2308.03921
-smoothllm: defending large language models against jailbreaking attacks,Alexander Robey,"Despite efforts to align large language models (LLMs) with human values, widely-used LLMs such as GPT, Llama, Claude, and PaLM are susceptible to jailbreaking attacks, wherein an adversary fools a targeted LLM into generating objectionable content. To address this vulnerability, we propose SmoothLLM, the first algorithm designed to mitigate jailbreaking attacks on LLMs. Based on our finding that adversarially-generated prompts are brittle to character-level changes, our defense first randomly perturbs multiple copies of a given input prompt, and then aggregates the corresponding predictions to detect adversarial inputs. SmoothLLM reduces the attack success rate on numerous popular LLMs to below one percentage point, avoids unnecessary conservatism, and admits provable guarantees on attack mitigation. Moreover, our defense uses exponentially fewer queries than existing attacks and is compatible with any LLM.","{'model': 'tldr@v2.0.0', 'text': 'Based on the finding that adversarially-generated prompts are brittle to character-level changes, the first algorithm designed to mitigate jailbreaking attacks on LLMs is proposed, which reduces the attack success rate on numerous popular LLMs to below one percentage point, avoids unnecessary conservatism, and admits provable guarantees on attack mitigation.'}",https://arxiv.org/pdf/2310.03684
-fully autonomous programming with large language models,Vadim Liventsev,"Current approaches to program synthesis with Large Language Models (LLMs) exhibit a ""near miss syndrome"": they tend to generate programs that semantically resemble the correct answer (as measured by text similarity metrics or human evaluation), but achieve a low or even zero accuracy as measured by unit tests due to small imperfections, such as the wrong input or output format. This calls for an approach known as Synthesize, Execute, Debug (SED), whereby a draft of the solution is generated first, followed by a program repair phase addressing the failed tests. To effectively apply this approach to instruction-driven LLMs, one needs to determine which prompts perform best as instructions for LLMs, as well as strike a balance between repairing unsuccessful programs and replacing them with newly generated ones. We explore these trade-offs empirically, comparing replace-focused, repair-focused, and hybrid debug strategies, as well as different template-based and model-based prompt-generation techniques. We use OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation. The resulting framework outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.","{'model': 'tldr@v2.0.0', 'text': 'This work uses OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation, and finds a framework that outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.'}",https://arxiv.org/pdf/2304.10423
-large language models and (non-)linguistic recursion,M. Dąbkowski,"Recursion is one of the hallmarks of human language. While many design features of language have been shown to exist in animal communication systems, recursion has not. Previous research shows that GPT-4 is the first large language model (LLM) to exhibit metalinguistic abilities (Begu\v{s}, D\k{a}bkowski, and Rhodes 2023). Here, we propose several prompt designs aimed at eliciting and analyzing recursive behavior in LLMs, both linguistic and non-linguistic. We demonstrate that when explicitly prompted, GPT-4 can both produce and analyze recursive structures. Thus, we present one of the first studies investigating whether meta-linguistic awareness of recursion -- a uniquely human cognitive property -- can emerge in transformers with a high number of parameters such as GPT-4.","{'model': 'tldr@v2.0.0', 'text': 'One of the first studies investigating whether meta-linguistic awareness of recursion -- a uniquely human cognitive property -- can emerge in transformers with a high number of parameters such as GPT-4 is presented.'}",http://arxiv.org/pdf/2306.07195
-revisit input perturbation problems for llms: a unified robustness evaluation framework for noisy slot filling task,Guanting Dong,,"{'model': 'tldr@v2.0.0', 'text': 'A unified robustness evaluation framework based on the slot-filling task is proposed to systematically evaluate the dialogue understanding capability of LLMs in diverse input perturbation scenarios and some forward-looking suggestions are made to fuel the research in this direction.'}",
-domain knowledge distillation from large language model: an empirical study in the autonomous driving domain,Yun Tang,"Engineering knowledge-based (or expert) systems require extensive manual effort and domain knowledge. As Large Language Models (LLMs) are trained using an enormous amount of cross-domain knowledge, it becomes possible to automate such engineering processes. This paper presents an empirical automation and semi-automation framework for domain knowledge distillation using prompt engineering and the LLM ChatGPT. We assess the framework empirically in the autonomous driving domain and present our key observations. In our implementation, we construct the domain knowledge ontology by""chatting""with ChatGPT. The key finding is that while fully automated domain ontology construction is possible, human supervision and early intervention typically improve efficiency and output quality as they lessen the effects of response randomness and the butterfly effect. We, therefore, also develop a web-based distillation assistant enabling supervision and flexible intervention at runtime. We hope our findings and tools could inspire future research toward revolutionizing the engineering of knowledge-based systems across application domains.","{'model': 'tldr@v2.0.0', 'text': 'The key finding is that while fully automated domain ontology construction is possible, human supervision and early intervention typically improve efficiency and output quality as they lessen the effects of response randomness and the butterfly effect.'}",https://arxiv.org/pdf/2307.11769
-investigating causal understanding in llms,Marius Hobbhahn,"We investigate the quality of causal world models of LLMs in very simple settings. We test whether LLMs can identify cause and effect in natural language settings (taken from BigBench) such as “My car got dirty. I washed the car. Question: Which sentence is the cause of the other?” and in multiple other toy settings. We probe the LLM’s world model by changing the presentation of the prompt while keeping the meaning constant, e.g. by changing the order of the sentences or asking the opposite question. Additionally, we test if the model can be “tricked” into giving wrong answers when we present the shot in a different pattern than the prompt. We have three findings. Firstly, larger models yield better results. Secondly, k-shot outperforms one-shot and one-shot outperforms zero-shot in standard conditions. Thirdly, LLMs perform worse in conditions where form and content differ. We conclude that the form of the presentation matters for LLM predictions or, in other words, that LLMs don’t solely base their predictions on content. Finally, we detail some of the implications this research has on AI safety.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that the form of the presentation matters for LLM predictions or, in other words, that LLMs don’t solely base their predictions on content.'}",
-reducing retraining by recycling parameter-efficient prompts,Brian Lester,"Parameter-efficient methods are able to use a single frozen pre-trained large language model (LLM) to perform many tasks by learning task-specific soft prompts that modulate model behavior when concatenated to the input text. However, these learned prompts are tightly coupled to a given frozen model -- if the model is updated, corresponding new prompts need to be obtained. In this work, we propose and investigate several approaches to""Prompt Recycling'""where a prompt trained on a source model is transformed to work with the new target model. Our methods do not rely on supervised pairs of prompts, task-specific data, or training updates with the target model, which would be just as costly as re-tuning prompts with the target model from scratch. We show that recycling between models is possible (our best settings are able to successfully recycle $88.9\%$ of prompts, producing a prompt that out-performs baselines), but significant performance headroom remains, requiring improved recycling techniques.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes and investigates several approaches to Prompt Recycling, where a prompt trained on a source model is transformed to work with the new target model, and shows that recycling between models is possible.'}",http://arxiv.org/pdf/2208.05577
-selecting better samples from pre-trained llms: a case study on question generation,Xingdi Yuan,"Large Language Models (LLMs) have in recent years demonstrated impressive prowess in natural language generation. A common practice to improve generation diversity is to sample multiple outputs from the model. However, there lacks a simple and robust way of selecting the best output from these stochastic samples. As a case study framed in the context of question generation, we propose two prompt-based approaches to selecting high-quality questions from a set of LLM-generated candidates. Our method works under the constraints of 1) a black-box (non-modifiable) question generation model and 2) lack of access to human-annotated references -- both of which are realistic limitations for real-world deployment of LLMs. With automatic as well as human evaluations, we empirically demonstrate that our approach can effectively select questions of higher qualities than greedy generation.","{'model': 'tldr@v2.0.0', 'text': 'This case study framed in the context of question generation proposes two prompt-based approaches to selecting high-quality questions from a set of LLM-generated candidates and empirically demonstrates that the approach can effectively select questions of higher qualities than greedy generation.'}",
-validating large language models with relm,Michael Kuchnik,"Although large language models (LLMs) have been touted for their ability to generate natural-sounding text, there are growing concerns around possible negative effects of LLMs such as data memorization, bias, and inappropriate language. Unfortunately, the complexity and generation capacities of LLMs make validating (and correcting) such concerns difficult. In this work, we introduce ReLM, a system for validating and querying LLMs using standard regular expressions. ReLM formalizes and enables a broad range of language model evaluations, reducing complex evaluation rules to simple regular expression queries. Our results exploring queries surrounding memorization, gender bias, toxicity, and language understanding show that ReLM achieves up to 15x higher system efficiency, 2.5x data efficiency, and increased statistical and prompt-tuning coverage compared to state-of-the-art ad-hoc queries. ReLM offers a competitive and general baseline for the increasingly important problem of LLM validation.","{'model': 'tldr@v2.0.0', 'text': 'Results exploring queries surrounding memorization, gender bias, toxicity, and language understanding show that ReLM achieves up to 15x higher system efficiency, 2.5x data efficiency, and increased statistical and prompt-tuning coverage compared to state-of-the-art ad-hoc queries.'}",http://arxiv.org/pdf/2211.15458
-preserving in-context learning ability in large language model fine-tuning,Yihan Wang,"Pretrained large language models (LLMs) are strong in-context learners that are able to perform few-shot learning without changing model parameters. However, as we show, fine-tuning an LLM on any specific task generally destroys its in-context ability. We discover an important cause of this loss, format specialization, where the model overfits to the format of the fine-tuned task and is unable to output anything beyond this format. We further show that format specialization happens at the beginning of fine-tuning. To solve this problem, we propose Prompt Tuning with MOdel Tuning (ProMoT), a simple yet effective two-stage fine-tuning framework that preserves in-context abilities of the pretrained model. ProMoT first trains a soft prompt for the fine-tuning target task, and then fine-tunes the model itself with this soft prompt attached. ProMoT offloads task-specific formats into the soft prompt that can be removed when doing other in-context tasks. We fine-tune mT5 XXL with ProMoT on natural language inference (NLI) and English-French translation and evaluate the in-context abilities of the resulting models on 8 different NLP tasks. ProMoT achieves similar performance on the fine-tuned tasks compared with vanilla fine-tuning, but with much less reduction of in-context learning performances across the board. More importantly, ProMoT shows remarkable generalization ability on tasks that have different formats, e.g","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prompt Tuning with MOdel Tuning (ProMoT), a simple yet effective two-stage fine-tuning framework that preserves in-context abilities of the pretrained model and shows remarkable generalization ability on tasks that have different formats.'}",https://arxiv.org/pdf/2211.00635
-improving knowledge extraction from llms for robotic task learning through agent analysis,James R. Kirk,": Large language models (LLMs) offer significant promise as a knowledge source for robotic task learning. Prompt engineering has been shown to be effective for eliciting knowledge from an LLM but alone is insufficient for acquiring relevant, situationally grounded knowledge for an embodied robotic agent learning novel tasks. We describe a cognitive-agent approach that extends and complements prompt engineering, mitigating its limitations, and thus enabling a robot to acquire new task knowledge matched to its native language capabilities, embodiment, environment, and user preferences. The approach is to increase the response space of LLMs and deploy general strategies, embedded within the autonomous robot, to evaluate, repair, and select among candidate responses produced by the LLM. We describe the approach and experiments that show how a robot, by retrieving and evaluating a breadth of responses from the LLM, can achieve > 75% task completion in one-shot learning without user oversight. The approach achieves 100% task completion when human oversight (such as indication of preference) is provided, while greatly reducing how much human oversight is needed.","{'model': 'tldr@v2.0.0', 'text': 'A cognitive-agent approach is described that extends and complements prompt engineering, mitigating its limitations, and thus enabling a robot to acquire new task knowledge matched to its native language capabilities, embodiment, environment, and user preferences.'}",https://arxiv.org/pdf/2306.06770
-large language models as superpositions of cultural perspectives,Grgur Kovač,"Large Language Models (LLMs) are often misleadingly recognized as having a personality or a set of values. We argue that an LLM can be seen as a superposition of perspectives with different values and personality traits. LLMs exhibit context-dependent values and personality traits that change based on the induced perspective (as opposed to humans, who tend to have more coherent values and personality traits across contexts). We introduce the concept of perspective controllability, which refers to a model's affordance to adopt various perspectives with differing values and personality traits. In our experiments, we use questionnaires from psychology (PVQ, VSM, IPIP) to study how exhibited values and personality traits change based on different perspectives. Through qualitative experiments, we show that LLMs express different values when those are (implicitly or explicitly) implied in the prompt, and that LLMs express different values even when those are not obviously implied (demonstrating their context-dependent nature). We then conduct quantitative experiments to study the controllability of different models (GPT-4, GPT-3.5, OpenAssistant, StableVicuna, StableLM), the effectiveness of various methods for inducing perspectives, and the smoothness of the models' drivability. We conclude by examining the broader implications of our work and outline a variety of associated scientific questions. The project website is available at https://sites.google.com/view/llm-superpositions .",,https://arxiv.org/pdf/2307.07870
-robot task planning based on large language model representing knowledge with directed graph structures,Zhen Yue,"Traditional robot task planning methods face challenges when dealing with highly unstructured environments and complex tasks. We propose a task planning method that combines human expertise with an LLM and have designed an LLM prompt template, Think_Net_Prompt, with stronger expressive power to represent structured professional knowledge. We further propose a method to progressively decompose tasks and generate a task tree to reduce the planning volume for each task, and we have designed a strategy to decouple robot task planning. By dividing different planning entities and separating the task from the actual machine binding process, the task planning process becomes more flexible. Research results show that our method performs well in handling specified code formats, understanding the relationship between tasks and subtasks, and extracting parameters from text descriptions. However, there are also problems such as limited complexity of task logic handling, ambiguity in the quantity of parts and the precise location of assembly. Improving the precision of task description and cognitive structure can bring certain improvements. https://github.com/NOMIzy/Think_Net_Prompt","{'model': 'tldr@v2.0.0', 'text': 'A task planning method that combines human expertise with anLLM and an LLM prompt template with stronger expressive power to represent structured professional knowledge is proposed and a method to progressively decompose tasks and generate a task tree to reduce the planning volume for each task is proposed.'}",http://arxiv.org/pdf/2306.05171
-using large language models to generate engaging captions for data visualizations,A. Liew,"Creating compelling captions for data visualizations has been a long- standing challenge. Visualization researchers are typically untrained in journalistic reporting and hence the captions that are placed be- low data visualizations tend to be not overly engaging and rather just stick to basic observations about the data. In this work we explore the opportunities offered by the newly emerging crop of large language models (LLM) which use sophisticated deep learning technology to produce human-like prose. We ask, can these power-ful software devices be purposed to produce engaging captions for generic data visualizations like a scatterplot. It turns out that the key challenge lies in designing the most effective prompt for the LLM, a task called prompt engineering . We report on first experiments using the popular LLM GPT-3 and deliver some promising results.","{'model': 'tldr@v2.0.0', 'text': 'The opportunities offered by the newly emerging crop of large language models (LLM) which use sophisticated deep learning technology to produce human-like prose are explored and can these power-ful software devices be purposed to produce engaging captions for generic data visualizations like a scatterplot.'}",http://arxiv.org/pdf/2212.14047
-spec: a soft prompt-based calibration on performance variability of large language model in clinical notes summarization,Yu-Neng Chuang,"Electronic health records (EHRs) store an extensive array of patient information, encompassing medical histories, diagnoses, treatments, and test outcomes. These records are crucial for enabling healthcare providers to make well-informed decisions regarding patient care. Summarizing clinical notes further assists healthcare professionals in pinpointing potential health risks and making better-informed decisions. This process contributes to reducing errors and enhancing patient outcomes by ensuring providers have access to the most pertinent and current patient data. Recent research has shown that incorporating prompts with large language models (LLMs) substantially boosts the efficacy of summarization tasks. However, we show that this approach also leads to increased output variance, resulting in notably divergent outputs even when prompts share similar meanings. To tackle this challenge, we introduce a model-agnostic Soft Prompt-Based Calibration (SPeC) pipeline that employs soft prompts to diminish variance while preserving the advantages of prompt-based summarization. Experimental findings on multiple clinical note tasks and LLMs indicate that our method not only bolsters performance but also effectively curbs variance for various LLMs, providing a more uniform and dependable solution for summarizing vital medical information.","{'model': 'tldr@v2.0.0', 'text': 'A model-agnostic Soft Prompt-Based Calibration (SPeC) pipeline is introduced that employs soft prompts to diminish variance while preserving the advantages of prompt-based summarization, providing a more uniform and dependable solution for summarizing vital medical information.'}",
-using a large language model to control speaking style for expressive tts,A. Sigurgeirsson,"Appropriate prosody is critical for successful spoken communication. Contextual word embeddings are proven to be helpful in predicting prosody but do not allow for choosing between plausible prosodic renditions. Reference-based TTS models attempt to address this by conditioning speech generation on a reference speech sample. These models can generate expressive speech but this requires finding an appropriate reference. Sufficiently large generative language models have been used to solve various language-related tasks. We explore whether such models can be used to suggest appropriate prosody for expressive TTS. We train a TTS model on a non-expressive corpus and then prompt the language model to suggest changes to pitch, energy and duration. The prompt can be designed for any task and we prompt the model to make suggestions based on target speaking style and dialogue context. The proposed method is rated most appropriate in 49.9% of cases compared to 31.0% for a baseline model.","{'model': 'tldr@v2.0.0', 'text': 'This work trains a TTS model on a non-expressive corpus and then prompts the language model to suggest changes to pitch, energy and duration and prompts the model to make suggestions based on target speaking style and dialogue context.'}",https://arxiv.org/pdf/2305.10321
-gpt4tools: teaching large language model to use tools via self-instruction,Rui Yang,"This paper aims to efficiently enable Large Language Models (LLMs) to use multimodal tools. Advanced proprietary LLMs, such as ChatGPT and GPT-4, have shown great potential for tool usage through sophisticated prompt engineering. Nevertheless, these models typically rely on prohibitive computational costs and publicly inaccessible data. To address these challenges, we propose the GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools. It generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts. By using the Low-Rank Adaptation (LoRA) optimization, our approach facilitates the open-source LLMs to solve a range of visual problems, including visual comprehension and image generation. Moreover, we provide a benchmark to evaluate the ability of LLMs to use tools, which is performed in both zero-shot and fine-tuning ways. Extensive experiments demonstrate the effectiveness of our method on various language models, which not only significantly improves the accuracy of invoking seen tools, but also enables the zero-shot capacity for unseen tools. The code and demo are available at https://github.com/StevenGrove/GPT4Tools.","{'model': 'tldr@v2.0.0', 'text': 'The GPT4Tools based on self-instruct to enable open-source LLMs, such as LLaMA and OPT, to use tools, generates an instruction-following dataset by prompting an advanced teacher with various multi-modal contexts using the Low-Rank Adaptation (LoRA) optimization.'}",http://arxiv.org/pdf/2305.18752
-simulating h.p. lovecraft horror literature with the chatgpt large language model,Eduardo C. Garrido-Merch'an,"In this paper, we present a novel approach to simulating H.P. Lovecraft's horror literature using the ChatGPT large language model, specifically the GPT-4 architecture. Our study aims to generate text that emulates Lovecraft's unique writing style and themes, while also examining the effectiveness of prompt engineering techniques in guiding the model's output. To achieve this, we curated a prompt containing several specialized literature references and employed advanced prompt engineering methods. We conducted an empirical evaluation of the generated text by administering a survey to a sample of undergraduate students. Utilizing statistical hypothesis testing, we assessed the students ability to distinguish between genuine Lovecraft works and those generated by our model. Our findings demonstrate that the participants were unable to reliably differentiate between the two, indicating the effectiveness of the GPT-4 model and our prompt engineering techniques in emulating Lovecraft's literary style. In addition to presenting the GPT model's capabilities, this paper provides a comprehensive description of its underlying architecture and offers a comparative analysis with related work that simulates other notable authors and philosophers, such as Dennett. By exploring the potential of large language models in the context of literary emulation, our study contributes to the body of research on the applications and limitations of these models in various creative domains.","{'model': 'tldr@v2.0.0', 'text': ""This study aims to generate text that emulates Lovecraft's unique writing style and themes, while also examining the effectiveness of prompt engineering techniques in guiding the model's output, and provides a comprehensive description of its underlying architecture.""}",http://arxiv.org/pdf/2305.03429
-s3: social-network simulation system with large language model-empowered agents,Chen Gao,"Social network simulation plays a crucial role in addressing various challenges within social science. It offers extensive applications such as state prediction, phenomena explanation, and policy-making support, among others. In this work, we harness the formidable human-like capabilities exhibited by large language models (LLMs) in sensing, reasoning, and behaving, and utilize these qualities to construct the S$^3$ system (short for $\textbf{S}$ocial network $\textbf{S}$imulation $\textbf{S}$ystem). Adhering to the widely employed agent-based simulation paradigm, we employ prompt engineering and prompt tuning techniques to ensure that the agent's behavior closely emulates that of a genuine human within the social network. Specifically, we simulate three pivotal aspects: emotion, attitude, and interaction behaviors. By endowing the agent in the system with the ability to perceive the informational environment and emulate human actions, we observe the emergence of population-level phenomena, including the propagation of information, attitudes, and emotions. We conduct an evaluation encompassing two levels of simulation, employing real-world social network data. Encouragingly, the results demonstrate promising accuracy. This work represents an initial step in the realm of social network simulation empowered by LLM-based agents. We anticipate that our endeavors will serve as a source of inspiration for the development of simulation systems within, but not limited to, social science.","{'model': 'tldr@v2.0.0', 'text': 'This work harnesses the formidable human-like capabilities exhibited by large language models in sensing, reasoning, and behaving, and utilize these qualities to construct the S$^3 system, an initial step in the realm of social network simulation empowered by LLM-based agents.'}",https://arxiv.org/pdf/2307.14984
-diminished diversity-of-thought in a standard large language model,Peter S. Park,"We test whether Large Language Models (LLMs) can be used to simulate human participants in social-science studies. To do this, we run replications of 14 studies from the Many Labs 2 replication project with OpenAI's text-davinci-003 model, colloquially known as GPT3.5. Based on our pre-registered analyses, we find that among the eight studies we could analyse, our GPT sample replicated 37.5% of the original results and 37.5% of the Many Labs 2 results. However, we were unable to analyse the remaining six studies due to an unexpected phenomenon we call the""correct answer""effect. Different runs of GPT3.5 answered nuanced questions probing political orientation, economic preference, judgement, and moral philosophy with zero or near-zero variation in responses: with the supposedly""correct answer.""In one exploratory follow-up study, we found that a""correct answer""was robust to changing the demographic details that precede the prompt. In another, we found that most but not all""correct answers""were robust to changing the order of answer choices. One of our most striking findings occurred in our replication of the Moral Foundations Theory survey results, where we found GPT3.5 identifying as a political conservative in 99.6% of the cases, and as a liberal in 99.3% of the cases in the reverse-order condition. However, both self-reported 'GPT conservatives' and 'GPT liberals' showed right-leaning moral foundations. Our results cast doubts on the validity of using LLMs as a general replacement for human participants in the social sciences. Our results also raise concerns that a hypothetical AI-led future may be subject to a diminished diversity-of-thought.","{'model': 'tldr@v2.0.0', 'text': ""Re-replicate 14 psychology studies from the Many Labs 2 replication project with OpenAI's text-davinci-003 model, colloquially known as GPT3.5, to hypothesise that this pattern was learned from a conservative bias in the model's largely Internet-based training data.""}",
-hierarchical prompting assists large language model on web navigation,Abishek Sridhar,"Large language models (LLMs) struggle on processing complicated observations in interactive decision making tasks. To alleviate this issue, we propose a simple hierarchical prompting approach. Diverging from previous prompting approaches that always put the full observation (e.g. a web page) to the prompt, we propose to first construct an action-aware observation which is more condensed and relevant with a dedicated SUMMARIZER prompt. The ACTOR prompt then predicts the next action based on the summarized observation. While our method has broad applicability, we particularly demonstrate its efficacy in the complex domain of web navigation where a full observation often contains redundant and irrelevant information. Our approach outperforms the previous state-of-the-art prompting mechanics by 6.2% on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple hierarchical prompting approach that outperforms the previous state-of-the-art prompting mechanics by 6.2% on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.'}",http://arxiv.org/pdf/2305.14257
-co-training improves prompt-based learning for large language models,Hunter Lang,"We demonstrate that co-training (Blum&Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data. While prompting has emerged as a promising paradigm for few-shot and zero-shot learning, it is often brittle and requires much larger models compared to the standard supervised setup. We find that co-training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model. In the case where we only have partial access to a prompt model (e.g., output probabilities from GPT-3 (Brown et al., 2020)) we learn a calibration model over the prompt outputs. When we have full access to the prompt model's gradients but full finetuning remains prohibitively expensive (e.g., T0 (Sanh et al., 2021)), we learn a set of soft prompt continuous vectors to iteratively update the prompt model. We find that models trained in this manner can significantly improve performance on challenging datasets where there is currently a large gap between prompt-based learning and fully-supervised models.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that co-training (Blum&Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data and co- training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model.'}",
-prompt text classifications with transformer models! an exemplary introduction to prompt-based learning with large language models,Christian W. F. Mayer,"Abstract This study investigates the potential of automated classification using prompt-based learning approaches with transformer models (large language models trained in an unsupervised manner) for a domain-specific classification task. Prompt-based learning with zero or few shots has the potential to (1) make use of artificial intelligence without sophisticated programming skills and (2) make use of artificial intelligence without fine-tuning models with large amounts of labeled training data. We apply this novel method to perform an experiment using so-called zero-shot classification as a baseline model and a few-shot approach for classification. For comparison, we also fine-tune a language model on the given classification task and conducted a second independent human rating to compare it with the given human ratings from the original study. The used dataset consists of 2,088 email responses to a domain-specific problem-solving task that were manually labeled for their professional communication style. With the novel prompt-based learning approach, we achieved a Cohen’s kappa of .40, while the fine-tuning approach yields a kappa of .59, and the new human rating achieved a kappa of .58 with the original human ratings. However, the classifications from the machine learning models have the advantage that each prediction is provided with a reliability estimate allowing us to identify responses that are difficult to score. We, therefore, argue that response ratings should be based on a reciprocal workflow of machine raters and human raters, where the machine rates easy-to-classify responses and the human raters focus and agree on the responses that are difficult to classify. Further, we believe that this new, more intuitive, prompt-based learning approach will enable more people to use artificial intelligence.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that response ratings should be based on a reciprocal workflow of machine raters and human rater, where the machine rates easy-to-classify responses and the human raters focus and agree on the responses that are difficult to classify.'}",
-on robustness of prompt-based semantic parsing with large pre-trained language model: an empirical study on codex,Terry Yue Zhuo,"Semantic parsing is a technique aimed at constructing a structured representation of the meaning of a natural-language question. Recent advances in language models trained on code have shown superior performance in generating these representations compared to language models trained solely on natural language text. The existing fine-tuned neural semantic parsers are vulnerable to adversarial attacks on natural-language inputs. While it has been established that the robustness of smaller semantic parsers can be enhanced through adversarial training, this approach is not feasible for large language models in real-world scenarios, as it requires both substantial computational resources and expensive human annotation on in-domain semantic parsing data. This paper presents the first empirical study on the adversarial robustness of a prompt-based semantic parser based on CODEX, a stateof-the-art (SOTA) language model trained on code. Our results demonstrate that the large language model of code is vulnerable to carefully crafted adversarial examples. To overcome this challenge, we propose methods for enhancing robustness without requiring substantial amounts of labelled data or intensive computational resources.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the first empirical study on the adversarial robustness of a prompt-based semantic parser based on CODEX, a stateof-the-art (SOTA) language model trained on code.'}",http://arxiv.org/pdf/2301.12868
-investigating the translation performance of a large multilingual language model: the case of bloom,Rachel Bawden,"The NLP community recently saw the release of a new large open-access multilingual language model, BLOOM (BigScience et al., 2022) covering 46 languages. We focus on BLOOM’s multilingual ability by evaluating its machine translation performance across several datasets (WMT, Flores-101 and DiaBLa) and language pairs (high- and low-resourced). Our results show that 0-shot performance suffers from overgeneration and generating in the wrong language, but this is greatly improved in the few-shot setting, with very good results for a number of language pairs. We study several aspects including prompt design, model sizes, cross-lingual transfer and the use of discursive context.","{'model': 'tldr@v2.0.0', 'text': 'The results show that 0-shot performance suffers from overgeneration and generating in the wrong language, but this is greatly improved in the few-shot setting, with very good results for a number of language pairs.'}",http://arxiv.org/pdf/2303.01911
-cold-start data selection for few-shot language model fine-tuning: a prompt-based uncertainty propagation approach,Yue Yu,"Large Language Models have demonstrated remarkable few-shot performance, but the performance can be sensitive to the selection of few-shot instances. We propose PATRON, a new method that uses prompt-based uncertainty estimation for data selection for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON is available at \url{https://github.com/yueyu1030/Patron}.","{'model': 'tldr@v2.0.0', 'text': 'PATRON is a new method that uses prompt-based uncertainty estimation for data selection for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available.'}",http://arxiv.org/pdf/2209.06995
-can large language models reason about medical questions?,Valentin Li'evin,"Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge. We set out to investigate whether GPT-3.5 (Codex and InstructGPT) can be applied to answer and reason about difficult real-world-based questions. We utilize two multiple-choice medical exam questions (USMLE and MedMCQA) and a medical reading comprehension dataset (PubMedQA). We investigate multiple prompting scenarios: Chain-of-Thought (CoT, think step-by-step), zero- and few-shot (prepending the question with question-answer exemplars) and retrieval augmentation (injecting Wikipedia passages into the prompt). For a subset of the USMLE questions, a medical expert reviewed and annotated the model's CoT. We found that InstructGPT can often read, reason and recall expert knowledge. Failure are primarily due to lack of knowledge and reasoning errors and trivial guessing heuristics are observed, e.g.\ too often predicting labels A and D on USMLE. Sampling and combining many completions overcome some of these limitations. Using 100 samples, Codex 5-shot CoT not only gives close to well-calibrated predictive probability but also achieves human-level performances on the three datasets. USMLE: 60.2%, MedMCQA: 62.7% and PubMedQA: 78.2%.","{'model': 'tldr@v2.0.0', 'text': 'It is found that InstructGPT can often read, reason and recall expert knowledge, and not only gives close to well-calibrated predictive probability but also achieves human-level performances on the three datasets.'}",http://arxiv.org/pdf/2207.08143
-tabllm: few-shot classification of tabular data with large language models,S. Hegselmann,"We study the application of large language models to zero-shot and few-shot classification of tabular data. We prompt the large language model with a serialization of the tabular data to a natural-language string, together with a short description of the classification problem. In the few-shot setting, we fine-tune the large language model using some labeled examples. We evaluate several serialization methods including templates, table-to-text models, and large language models. Despite its simplicity, we find that this technique outperforms prior deep-learning-based tabular classification methods on several benchmark datasets. In most cases, even zero-shot classification obtains non-trivial performance, illustrating the method's ability to exploit prior knowledge encoded in large language models. Unlike many deep learning methods for tabular datasets, this approach is also competitive with strong traditional baselines like gradient-boosted trees, especially in the very-few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'Despite its simplicity, this technique outperforms prior deep-learning-based tabular classification methods on several benchmark datasets and is also competitive with strong traditional baselines like gradient-boosted trees, especially in the very-few-shot setting.'}",http://arxiv.org/pdf/2210.10723
-evaluating the text-to-sql capabilities of large language models,Nitarshan Rajkumar,"We perform an empirical evaluation of Text-to-SQL capabilities of the Codex language model. We find that, without any finetuning, Codex is a strong baseline on the Spider benchmark; we also analyze the failure modes of Codex in this setting. Furthermore, we demonstrate on the GeoQuery and Scholar benchmarks that a small number of in-domain examples provided in the prompt enables Codex to perform better than state-of-the-art models finetuned on such few-shot examples.","{'model': 'tldr@v2.0.0', 'text': 'An empirical evaluation of Text-to-SQL capabilities of the Codex language model is performed and it is demonstrated on the GeoQuery and Scholar benchmarks that a small number of in-domain examples provided in the prompt enables Codex to perform better than state-of-the-art models finetuned on such few-shot examples.'}",http://arxiv.org/pdf/2204.00498
-prompting is programming: a query language for large language models,Luca Beurer-Kellner,"Large language models have demonstrated outstanding performance on a wide range of tasks such as question answering and code generation. On a high level, given an input, a language model can be used to automatically complete the sequence in a statistically-likely way. Based on this, users prompt these models with language instructions or examples, to implement a variety of downstream tasks. Advanced prompting methods can even imply interaction between the language model, a user, and external tools such as calculators. However, to obtain state-of-the-art performance or adapt language models for specific tasks, complex task- and model-specific programs have to be implemented, which may still require ad-hoc interaction. Based on this, we present the novel idea of Language Model Programming (LMP). LMP generalizes language model prompting from pure text prompts to an intuitive combination of text prompting and scripting. Additionally, LMP allows constraints to be specified over the language model output. This enables easy adaption to many tasks while abstracting language model internals and providing high-level semantics. To enable LMP, we implement LMQL (short for Language Model Query Language), which leverages the constraints and control flow from an LMP prompt to generate an efficient inference procedure that minimizes the number of expensive calls to the underlying language model. We show that LMQL can capture a wide range of state-of-the-art prompting methods in an intuitive way, especially facilitating interactive flows that are challenging to implement with existing high-level APIs. Our evaluation shows that we retain or increase the accuracy on several downstream tasks, while also significantly reducing the required amount of computation or cost in the case of pay-to-use APIs (26-85% cost savings).","{'model': 'tldr@v2.0.0', 'text': 'LMQL is implemented, which leverages the constraints and control flow from an LMP prompt to generate an efficient inference procedure that minimizes the number of expensive calls to the underlying language model.'}",https://dl.acm.org/doi/pdf/10.1145/3591300
-large language models are reasoning teachers,Namgyu Ho,"Recent works have shown that chain-of-thought (CoT) prompting can elicit language models to solve complex reasoning tasks, step-by-step. However, prompt-based CoT methods are dependent on very large models such as GPT-3 175B which are prohibitive to deploy at scale. In this paper, we use these large models as reasoning teachers to enable complex reasoning in smaller models and reduce model size requirements by several orders of magnitude. We propose Fine-tune-CoT, a method that generates reasoning samples from very large teacher models to fine-tune smaller models. We evaluate our method on a wide range of public models and complex tasks. We find that Fine-tune-CoT enables substantial reasoning capability in small models, far outperforming prompt-based baselines and even the teacher model in many tasks. Additionally, we extend our method by leveraging the teacher model’s ability to generate multiple distinct rationales for each original sample. Enriching the fine-tuning data with such diverse reasoning results in a substantial performance boost across datasets, even for very small models. We conduct ablations and sample studies to understand the emergence of reasoning capabilities of student models. Our code implementation and data are available at https://github.com/itsnamgyu/reasoning-teacher.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses very large models as reasoning teachers to enable complex reasoning in smaller models and reduce model size requirements by several orders of magnitude, and proposes Fine-tune-CoT, a method that generates reasoning samples from very large teacher models to fine-tunes smaller models.'}",http://arxiv.org/pdf/2212.10071
-class-aware visual prompt tuning for vision-language pre-trained model,Yinghui Xing,"With the emergence of large pre-trained vison-language model like CLIP, transferrable representations can be adapted to a wide range of downstream tasks via prompt tuning. Prompt tuning tries to probe the beneficial information for downstream tasks from the general knowledge stored in both the image and text encoders of the pre-trained vision-language model. A recently proposed method named Con- text Optimization (CoOp) introduces a set of learnable vectors as text prompt from the language side, while tuning the text prompt alone can not affect the computed visual features of the image encoder, thus leading to sub-optimal. In this paper, we propose a dual modality prompt tuning paradigm through learning text prompts and visual prompts for both the text and image encoder simultaneously. In addi-tion, to make the visual prompt concentrate more on the tar- get visual concept, we propose Class-Aware Visual Prompt Tuning (CAVPT), which is generated dynamically by per- forming the cross attention between language descriptions of template prompts and visual class token embeddings. Our method provides a new paradigm for tuning the large pretrained vision-language model and extensive experimental re- sults on 8 datasets demonstrate the effectiveness of the proposed method. Our code is available in the supplementary materials.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a dual modality prompt tuning paradigm through learning text prompts and visual prompts for both the text and image encoder simultaneously, which provides a new paradigm for tuning the large pretrained vision-language model.'}",http://arxiv.org/pdf/2208.08340
-analogy generation by prompting large language models: a case study of instructgpt,B. Bhavya,"We propose a novel application of prompting Pre-trained Language Models (PLMs) to generate analogies and study how to design effective prompts for two task settings: generating a source concept analogous to a given target concept (aka Analogous Concept Generation or ACG), and generating an explanation of the similarity between a given pair of target concept and source concept (aka Analogous Explanation Generation or AEG). We found that it is feasible to prompt InstructGPT to generate meaningful analogies and the best prompts tend to be precise imperative statements especially with a low temperature setting. We also systematically analyzed the sensitivity of the InstructGPT model to prompt design, temperature, and injected spelling errors, and found that the model is particularly sensitive to certain variations (e.g., questions vs. imperative statements). Further, we conducted human evaluation on 1.4k of the generated analogies and found that the quality of generations varies substantially by model size. The largest InstructGPT model can achieve human-level performance at generating meaningful analogies for a given target while there is still room for improvement on the AEG task.","{'model': 'tldr@v2.0.0', 'text': 'It is feasible to prompt InstructGPT to generate meaningful analogies and the best prompts tend to be precise imperative statements especially with a low temperature setting, and the quality of generations varies substantially by model size.'}",http://arxiv.org/pdf/2210.04186
-using large language models to simulate multiple humans,Gati Aher,"We propose a method for using a large language model, such as GPT-3, to simulate responses of different humans in a given context. We test our method by attempting to repro- duce well-established economic, psycholinguistic, and social experiments. The method requires prompt templates for each experiment. Simulations are run by varying the (hypotheti-cal) subject details, such as name, and analyzing the text gen- erated by the language model. To validate our methodology, we use GPT-3 to simulate the Ultimatum Game , garden path sentences , risk aversion , and the Milgram Shock experiments. In order to address concerns of exposure to these studies in training data, we also evaluate simulations on novel variants of these studies. We show that it is possible to simulate re- sponses of different people and that their responses are consistent with prior human studies from the literature. Across all studies, the distributions generated by larger language models better align with prior experimental results, suggesting a trend that future language models may be used for even more faithful simulations of human responses. Our use of a lan- guage model for simulation is contrasted with anthropomor-phic views of a language model as having its own behavior.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that it is possible to simulate re- sponses of different people and that their responses are consistent with prior human studies from the literature, suggesting a trend that future language models may be used for even more faithful simulations of human responses.'}",https://arxiv.org/pdf/2208.10264
-persistent anti-muslim bias in large language models,Abubakar Abid,"It has been observed that large-scale language models capture undesirable societal biases, e.g. relating to race and gender; yet religious bias has been relatively unexplored. We demonstrate that GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias. We probe GPT-3 in various ways, including prompt completion, analogical reasoning, and story generation, to understand this anti-Muslim bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups. For instance, Muslim is analogized to terrorist in 23% of test cases, while Jewish is mapped to its most common stereotype, money, in 5% of test cases. We quantify the positive distraction needed to overcome this bias with adversarial text prompts, and find that use of the most positive 6 adjectives reduces violent completions for Muslims from 66% to 20%, but which is still higher than for other religious groups.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3, a state-of-the-art contextual language model, captures persistent Muslim-violence bias, demonstrating that it appears consistently and creatively in different uses of the model and that it is severe even compared to biases about other religious groups.'}",https://arxiv.org/pdf/2101.05783
-large language models in the workplace: a case study on prompt engineering for job type classification,Benjamin Clavié,"This case study investigates the task of job classification in a real-world setting, where the goal is to determine whether an English-language job posting is appropriate for a graduate or entry-level position. We explore multiple approaches to text classification, including supervised approaches such as traditional models like Support Vector Machines (SVMs) and state-of-the-art deep learning methods such as DeBERTa. We compare them with Large Language Models (LLMs) used in both few-shot and zero-shot classification settings. To accomplish this task, we employ prompt engineering, a technique that involves designing prompts to guide the LLMs towards the desired output. Specifically, we evaluate the performance of two commercially available state-of-the-art GPT-3.5-based language models, text-davinci-003 and gpt-3.5-turbo. We also conduct a detailed analysis of the impact of different aspects of prompt engineering on the model's performance. Our results show that, with a well-designed prompt, a zero-shot gpt-3.5-turbo classifier outperforms all other models, achieving a 6% increase in Precision@95% Recall compared to the best supervised approach. Furthermore, we observe that the wording of the prompt is a critical factor in eliciting the appropriate""reasoning""in the model, and that seemingly minor aspects of the prompt significantly affect the model's performance.","{'model': 'tldr@v2.0.0', 'text': 'This case study investigates the task of job classification in a real-world setting, where the goal is to determine whether an English-language job posting is appropriate for a graduate or entry-level position, and evaluates the performance of two commercially available state-of-the-art GPT-3.5-turbo language models.'}",http://arxiv.org/pdf/2303.07142
-soft-prompt tuning for large language models to evaluate bias,Jacob-Junqi Tian,"Prompting large language models has gained immense popularity in recent years due to the advantage of producing good results even without the need for labelled data. However, this requires prompt tuning to get optimal prompts that lead to better model performances. In this paper, we explore the use of soft-prompt tuning on sentiment classification task to quantify the biases of large language models (LLMs) such as Open Pre-trained Transformers (OPT) and Galactica language model. Since these models are trained on real-world data that could be prone to bias toward certain groups of populations, it is important to identify these underlying issues. Using soft-prompts to evaluate bias gives us the extra advantage of avoiding the human-bias injection that can be caused by manually designed prompts. We check the model biases on different sensitive attributes using the group fairness (bias) and find interesting bias patterns. Since LLMs have been used in the industry in various applications, it is crucial to identify the biases before deploying these models in practice. We open-source our pipeline and encourage industry researchers to adapt our work to their use cases.","{'model': 'tldr@v2.0.0', 'text': 'The use of soft-prompt tuning on sentiment classification task to quantify the biases of large language models such as Open Pre-trained Transformers (OPT) and Galactica language model and finds interesting bias patterns.'}",http://arxiv.org/pdf/2306.04735
-promptify: text-to-image generation through interactive prompt exploration with large language models,Stephen Brade,"Text-to-image generative models have demonstrated remarkable capabilities in generating high-quality images based on textual prompts. However, crafting prompts that accurately capture the user’s creative intent remains challenging. It often involves laborious trial-and-error procedures to ensure that the model interprets the prompts in alignment with the user’s intention. To address these challenges, we present Promptify, an interactive system that supports prompt exploration and refinement for text-to-image generative models. Promptify utilizes a suggestion engine powered by large language models to help users quickly explore and craft diverse prompts. Our interface allows users to organize the generated images flexibly, and based on their preferences, Promptify suggests potential changes to the original prompt. This feedback loop enables users to iteratively refine their prompts and enhance desired features while avoiding unwanted ones. Our user study shows that Promptify effectively facilitates the text-to-image workflow, allowing users to create visually appealing images on their first attempt while requiring significantly less cognitive load than a widely-used baseline tool.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Promptify, an interactive system that supports prompt exploration and refinement for text-to-image generative models, and utilizes a suggestion engine powered by large language models to help users quickly explore and craft diverse prompts.'}",https://arxiv.org/pdf/2304.09337
-you only prompt once: on the capabilities of prompt learning on large language models to tackle toxic content,Xinlei He,"The spread of toxic content online is an important problem that has adverse effects on user experience online and in our society at large. Motivated by the importance and impact of the problem, research focuses on developing solutions to detect toxic content, usually leveraging machine learning (ML) models trained on human-annotated datasets. While these efforts are important, these models usually do not generalize well and they can not cope with new trends (e.g., the emergence of new toxic terms). Currently, we are witnessing a shift in the approach to tackling societal issues online, particularly leveraging large language models (LLMs) like GPT-3 or T5 that are trained on vast corpora and have strong generalizability. In this work, we investigate how we can use LLMs and prompt learning to tackle the problem of toxic content, particularly focusing on three tasks; 1) Toxicity Classification, 2) Toxic Span Detection, and 3) Detoxification. We perform an extensive evaluation over five model architectures and eight datasets demonstrating that LLMs with prompt learning can achieve similar or even better performance compared to models trained on these specific tasks. We find that prompt learning achieves around 10\% improvement in the toxicity classification task compared to the baselines, while for the toxic span detection task we find better performance to the best baseline (0.643 vs. 0.640 in terms of $F_1$-score). Finally, for the detoxification task, we find that prompt learning can successfully reduce the average toxicity score (from 0.775 to 0.213) while preserving semantic meaning.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates how to use LLMs and prompt learning to tackle the problem of toxic content, particularly focusing on three tasks; 1) Toxicity Classification, 2) Toxic Span Detection, and 3) Detoxification.'}",https://arxiv.org/pdf/2308.05596
-controlling the extraction of memorized data from large language models via prompt-tuning,Mustafa Safa Ozdayi,"Large Language Models (LLMs) are known to memorize significant portions of their training data. Parts of this memorized content have been shown to be extractable by simply querying the model, which poses a privacy risk. We present a novel approach which uses prompt-tuning to control the extraction rates of memorized content in LLMs. We present two prompt training strategies to increase and decrease extraction rates, which correspond to an attack and a defense, respectively. We demonstrate the effectiveness of our techniques by using models from the GPT-Neo family on a public benchmark. For the 1.3B parameter GPT-Neo model, our attack yields a 9.3 percentage point increase in extraction rate compared to our baseline. Our defense can be tuned to achieve different privacy-utility trade-offs by a user-specified hyperparameter. We achieve an extraction rate reduction of up to 97.7% relative to our baseline, with a perplexity increase of 16.9%.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach which uses prompt-tuning to control the extraction rates of memorized content in LLMs, and presents two prompt training strategies to increase and decrease extraction rates, which correspond to an attack and a defense, respectively.'}",http://arxiv.org/pdf/2305.11759
-sensitivity and robustness of large language models to prompt template in japanese text classification tasks,Chengguang Gan,"Prompt engineering relevance research has seen a notable surge in recent years, primarily driven by advancements in pre-trained language models and large language models. However, a critical issue has been identified within this domain: the inadequate of sensitivity and robustness of these models towards Prompt Templates, particularly in lesser-studied languages such as Japanese. This paper explores this issue through a comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM). These models are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context. Our experimental results reveal startling discrepancies. A simple modification in the sentence structure of the Prompt Template led to a drastic drop in the accuracy of GPT-4 from 49.21 to 25.44. This observation underscores the fact that even the highly performance GPT-4 model encounters significant stability issues when dealing with diverse Japanese prompt templates, rendering the consistency of the model's output results questionable. In light of these findings, we conclude by proposing potential research trajectories to further enhance the development and performance of Large Language Models in their current stage.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM) is scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context.'}",
-sensitivity and robustness of large language models to prompt in japanese,Chengguang Gan,"Prompt Engineering has gained significant rel-evance in recent years, fueled by advance-ments in pre-trained and large language models. However, a critical issue has been iden-tified within this domain: the lack of sensitivity and robustness of these models towards Prompt Templates, particularly in lesser-studied languages such as Japanese. This paper explores this issue through a comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model(PLM), T5. These models are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context. Our experimental results reveal startling discrepancies. A simple modification in the sentence structure of the Prompt Template led to a drastic drop in the accuracy of GPT-4 from 49.21 to 25.44. This observation underscores the fact that even the highly performance GPT-4 model encoun-ters significant stability issues when dealing with diverse Japanese prompt templates, ren-dering the consistency of the model’s output results questionable. In light of these findings, we conclude by proposing potential research trajectories to further enhance the devel-opment and performance of Large Language Models in their current stage.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive evaluation of several representative Large Language Models (LLMs) and a widely-utilized pre-trained model (PLM), T5 are scrutinized using a benchmark dataset in Japanese, with the aim to assess and analyze the performance of the current multilingual models in this context.'}",http://arxiv.org/pdf/2305.08714
-bounding the capabilities of large language models in open text generation with prompt constraints,Albert Lu,"The limits of open-ended generative models are unclear, yet increasingly important. What causes them to succeed and what causes them to fail? In this paper, we take a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models. We present a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic. These constraint types are categorized into a set of well-defined constraints that are analyzable by a single prompt. We then systematically create a diverse set of simple, natural, and useful prompts to robustly analyze each individual constraint. Using the GPT-3 text-davinci-002 model as a case study, we generate outputs from our collection of prompts and analyze the model’s generative failures. We also show the generalizability of our proposed method on other large models like BLOOM and OPT. Our results and our in-context mitigation strategies reveal open challenges for future research.","{'model': 'tldr@v2.0.0', 'text': 'This paper takes a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models, and presents a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic.'}",http://arxiv.org/pdf/2302.09185
-linguist: language model instruction tuning to generate annotated utterances for intent classification and slot tagging,Andrew Rosenbaum,"We present LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt. In a 10-shot novel intent setting for the SNIPS dataset, LINGUIST surpasses state-of-the-art approaches (Back-Translation and Example Extrapolation) by a wide margin, showing absolute improvement for the target intents of +1.9 points on IC Recall and +2.5 points on ST F1 Score. In the zero-shot cross-lingual setting of the mATIS++ dataset, LINGUIST out-performs a strong baseline of Machine Translation with Slot Alignment by +4.14 points absolute on ST F1 Score across 6 languages, while matching performance on IC. Finally, we verify our results on an internal large-scale multilingual dataset for conversational agent IC+ST and show significant improvements over a baseline which uses Back-Translation, Paraphrasing and Slot Catalog Resampling. To our knowledge, we are the first to demonstrate instruction fine-tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation.","{'model': 'tldr@v2.0.0', 'text': 'This work presents LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt, and is the first to demonstrate instruction fine- tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation'}",http://arxiv.org/pdf/2209.09900
-augesc: dialogue augmentation with large language models for emotional support conversation,Chujie Zheng,"Crowdsourced dialogue corpora are usually limited in scale and topic coverage due to the expensive cost of data curation. This would hinder the generalization of downstream dialogue models to open-domain topics. In this work, we leverage large language models for dialogue augmentation in the task of emotional support conversation (ESC). By treating dialogue augmentation as a dialogue completion task, we prompt a fine-tuned language model to complete full dialogues from available dialogue posts of various topics, which are then postprocessed based on heuristics. Applying this approach, we construct AugESC, an augmented dataset for the ESC task, which largely extends the scale and topic coverage of the crowdsourced ESConv corpus. Through comprehensive human evaluation, we demonstrate that our approach is superior to strong baselines of dialogue augmentation and that AugESC has comparable dialogue quality to the crowdsourced corpus. We also conduct human interactive evaluation and prove that post-training on AugESC improves downstream dialogue models' generalization ability to open-domain topics. These results suggest the utility of AugESC and highlight the potential of large language models in improving data-scarce dialogue generation tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work leverages large language models for dialogue augmentation in the task of emotional support conversation (ESC) to prompt a fine-tuned language model to complete full dialogues from available dialogue posts of various topics, which are then postprocessed based on heuristics.'}",https://aclanthology.org/2023.findings-acl.99.pdf
-conal: anticipating outliers with large language models,Albert Xu,"In many task settings, text classification models are likely to encounter examples from novel classes on which they cannot predict correctly. Selective prediction, in which models abstain on low-confidence examples, provides a possible solution, but existing models are often overly confident on OOD examples. To remedy this overconfidence, we introduce Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them. First, we generate OOD examples by prompting a large language model twice: we prompt it to enumerate relevant novel labels, then generate examples from each novel class matching the task format. Second, we train our classifier with a novel contrastive objective that encourages lower confidence on generated OOD examples than training examples. When trained with CoNAL, classifiers improve in their ability to detect and abstain on OOD examples over prior methods by an average of 2.3% AUAC and 5.5% AUROC across 4 NLP datasets, with no cost to in-distribution accuracy.1","{'model': 'tldr@v2.0.0', 'text': ""Contrastive Novelty-Augmented Learning (CoNAL), a two-step method that generates OOD examples representative of novel classes, then trains to decrease confidence on them, which improves classifiers' ability to detect and abstain on OODExamples over prior methods.""}",http://arxiv.org/pdf/2211.15718
-variational prompt tuning improves generalization of vision-language models,Mohammad Mahdi Derakhshani,"Prompt tuning provides an efficient mechanism to adapt large vision-language models to downstream tasks by treating part of the input language prompts as learnable parameters while freezing the rest of the model. Existing works for prompt tuning are however prone to damaging the generalization capabilities of the foundation models, because the learned prompts lack the capacity of covering certain concepts within the language model. To avoid such limitation, we propose a probabilistic modeling of the underlying distribution of prompts, allowing prompts within the support of an associated concept to be derived through stochastic sampling. This results in a more complete and richer transfer of the information captured by the language model, providing better generalization capabilities for downstream tasks. The resulting algorithm relies on a simple yet powerful variational framework that can be directly integrated with other developments. We show our approach is seamlessly integrated into both standard and conditional prompt learning frameworks, improving the performance on both cases considerably, especially with regards to preserving the generalization capability of the original model. Our method provides the current state-of-the-art for prompt learning, surpassing CoCoOp by 1.6% average Top-1 accuracy on the standard benchmark. Remarkably, it even surpasses the original CLIP model in terms of generalization to new classes. Implementation code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a probabilistic modeling of the underlying distribution of prompts, allowing prompts within the support of an associated concept to be derived through stochastic sampling, which results in a more complete and richer transfer of the information captured by the language model, providing better generalization capabilities for downstream tasks.'}",https://arxiv.org/pdf/2210.02390
-prompt-and-rerank: a method for zero-shot and few-shot arbitrary textual style transfer with small language models,Mirac Suzgun,"We propose a method for arbitrary textual style transfer (TST)—the task of transforming a text into any given style—utilizing general-purpose pre-trained language models. Our method, Prompt-and-Rerank, is based on a mathematical formulation of the TST task, decomposing it into three constituent components: textual similarity, target style strength, and fluency. Our method uses zero-shot or few-shot prompting to obtain a set of candidate generations in the target style, and then re-ranks them according to the three components. Our method enables small pre-trained language models to perform on par with state-of-the-art large-scale models while using two orders of magnitude less compute and memory. We also investigate the effect of model size and prompt design (e.g., prompt paraphrasing and delimiter-pair choice) on style transfer quality across seven diverse textual style transfer datasets, finding, among other things, that delimiter-pair choice has a large impact on performance, and that models have biases on the direction of style transfer.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for arbitrary textual style transfer (TST), based on a mathematical formulation of the TST task, that enables small pre-trained language models to perform on par with state-of-the-art large-scale models while using two orders of magnitude less compute and memory.'}",https://arxiv.org/pdf/2205.11503
-visual-language navigation pretraining via prompt-based environmental self-exploration,Xiwen Liang,"Vision-language navigation (VLN) is a challenging task due to its large searching space in the environment. To address this problem, previous works have proposed some methods of fine-tuning a large model that pretrained on large-scale datasets. However, the conventional fine-tuning methods require extra human-labeled navigation data and lack self-exploration capabilities in environments, which hinders their generalization of unseen scenes. To improve the ability of fast cross-domain adaptation, we propose Prompt-based Environmental Self-exploration (ProbES), which can self-explore the environments by sampling trajectories and automatically generates structured instructions via a large-scale cross-modal pretrained model (CLIP). Our method fully utilizes the knowledge learned from CLIP to build an in-domain dataset by self-exploration without human labeling. Unlike the conventional approach of fine-tuning, we introduce prompt tuning to achieve fast adaptation for language embeddings, which substantially improves the learning efficiency by leveraging prior knowledge. By automatically synthesizing trajectory-instruction pairs in any environment without human supervision and instruction prompt tuning, our model can adapt to diverse vision-language navigation tasks, including VLN and REVERIE. Both qualitative and quantitative results show that our ProbES significantly improves the generalization ability of the navigation model.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces prompt tuning to achieve fast adaptation for language embeddings, which substantially improves the learning efficiency by leveraging prior knowledge and significantly improves the generalization ability of the navigation model.'}",http://arxiv.org/pdf/2203.04006
-prcbert: prompt learning for requirement classification using bert-based pretrained language models,Xianchang Luo,"Software requirement classification is a longstanding and important problem in requirement engineering. Previous studies have applied various machine learning techniques for this problem, including Support Vector Machine (SVM) and decision trees. With the recent popularity of NLP technique, the state-of-the-art approach NoRBERT utilizes the pre-trained language model BERT and achieves a satisfactory performance. However, the dataset PROMISE used by the existing approaches for this problem consists of only hundreds of requirements that are outdated according to today’s technology and market trends. Besides, the NLP technique applied in these approaches might be obsolete. In this paper, we propose an approach of prompt learning for requirement classification using BERT-based pretrained language models (PRCBERT), which applies flexible prompt templates to achieve accurate requirements classification. Experiments conducted on two existing small-size requirement datasets (PROMISE and NFR-Review) and our collected large-scale requirement dataset NFR-SO prove that PRCBERT exhibits moderately better classification performance than NoRBERT and MLM-BERT (BERT with the standard prompt template). On the de-labeled NFR-Review and NFR-SO datasets, Trans_PRCBERT (the version of PRCBERT which is fine-tuned on PROMISE) is able to have a satisfactory zero-shot performance with 53.27% and 72.96% F1-score when enabling a self-learning strategy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an approach of prompt learning for requirement classification using BERT-based pretrained language models (PRCBERT), which applies flexible prompt templates to achieve accurate requirements classification.'}",https://dl.acm.org/doi/pdf/10.1145/3551349.3560417
-toolllm: facilitating large language models to master 16000+ real-world apis,Yujia Qin,"Despite the advancements of open-source large language models (LLMs), e.g., LLaMA, they remain significantly limited in tool-use capabilities, i.e., using external tools (APIs) to fulfill human instructions. The reason is that current instruction tuning largely focuses on basic language tasks but ignores the tool-use domain. This is in contrast to the excellent tool-use capabilities of state-of-the-art (SOTA) closed-source LLMs, e.g., ChatGPT. To bridge this gap, we introduce ToolLLM, a general tool-use framework encompassing data construction, model training, and evaluation. We first present ToolBench, an instruction-tuning dataset for tool use, which is constructed automatically using ChatGPT. Specifically, the construction can be divided into three stages: (i) API collection: we collect 16,464 real-world RESTful APIs spanning 49 categories from RapidAPI Hub; (ii) instruction generation: we prompt ChatGPT to generate diverse instructions involving these APIs, covering both single-tool and multi-tool scenarios; (iii) solution path annotation: we use ChatGPT to search for a valid solution path (chain of API calls) for each instruction. To enhance the reasoning capabilities of LLMs, we develop a novel depth-first search-based decision tree algorithm. It enables LLMs to evaluate multiple reasoning traces and expand the search space. Moreover, to evaluate the tool-use capabilities of LLMs, we develop an automatic evaluator: ToolEval. Based on ToolBench, we fine-tune LLaMA to obtain an LLM ToolLLaMA, and equip it with a neural API retriever to recommend appropriate APIs for each instruction. Experiments show that ToolLLaMA demonstrates a remarkable ability to execute complex instructions and generalize to unseen APIs, and exhibits comparable performance to ChatGPT. Our ToolLLaMA also demonstrates strong zero-shot generalization ability in an out-of-distribution tool-use dataset: APIBench.","{'model': 'tldr@v2.0.0', 'text': 'ToolLLM is introduced, a general tool-use framework encompassing data construction, model training, and evaluation, and a novel depth-first search-based decision tree algorithm that enables LLMs to evaluate multiple reasoning traces and expand the search space.'}",https://arxiv.org/pdf/2307.16789
-fundamental limitations of alignment in large language models,Yotam Wolf,"An important aspect in developing language models that interact with humans is aligning their behavior to be useful and unharmful for their human users. This is usually achieved by tuning the model in a way that enhances desired behaviors and inhibits undesired ones, a process referred to as alignment. In this paper, we propose a theoretical approach called Behavior Expectation Bounds (BEB) which allows us to formally investigate several inherent characteristics and limitations of alignment in large language models. Importantly, we prove that within the limits of this framework, for any behavior that has a finite probability of being exhibited by the model, there exist prompts that can trigger the model into outputting this behavior, with probability that increases with the length of the prompt. This implies that any alignment process that attenuates an undesired behavior but does not remove it altogether, is not safe against adversarial prompting attacks. Furthermore, our framework hints at the mechanism by which leading alignment approaches such as reinforcement learning from human feedback make the LLM prone to being prompted into the undesired behaviors. This theoretical result is being experimentally demonstrated in large scale by the so called contemporary""chatGPT jailbreaks"", where adversarial users trick the LLM into breaking its alignment guardrails by triggering it into acting as a malicious persona. Our results expose fundamental limitations in alignment of LLMs and bring to the forefront the need to devise reliable mechanisms for ensuring AI safety.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that within the limits of this framework, for any behavior that has a finite probability of being exhibited by the model, there exist prompts that can trigger the model into outputting this behavior, with probability that increases with the length of the prompt.'}",https://arxiv.org/pdf/2304.11082
-synthetic prompting: generating chain-of-thought demonstrations for large language models,Zhihong Shao,"Large language models can perform various reasoning tasks by using chain-of-thought prompting, which guides them to find answers through step-by-step demonstrations. However, the quality of the prompts depends on the demonstrations given to the models, and creating many of them by hand is costly. We introduce Synthetic prompting, a method that leverages a few handcrafted examples to prompt the model to generate more examples by itself, and selects effective demonstrations to elicit better reasoning. Our method alternates between a backward and forward process to generate new examples. The backward process generates a question that match a sampled reasoning chain, so that the question is solvable and clear. The forward process produces a more detailed reasoning chain for the question, improving the quality of the example. We evaluate our method on numerical, symbolic, and algorithmic reasoning tasks, and show that it outperforms existing prompting techniques.","{'model': 'tldr@v2.0.0', 'text': 'Synthetic prompting is introduced, a method that leverages a few handcrafted examples to prompt the model to generate more examples by itself, and selects effective demonstrations to elicit better reasoning.'}",http://arxiv.org/pdf/2302.00618
-universal captioner: inducing content-style separation in vision-and-language model training,M. Cornia,"While captioning models have obtained compelling results in describing natural images, there is a growing effort to increase their capability of dealing with real-world concepts. In this paper, we address the task of generating fluent descriptions by training on a non-uniform combination of data sources, containing both human- and automatically-collected captions. To this end, we propose a model which induces a separation between content and descriptive style through the incorporation of stylistic parameters and keywords extracted from large-scale multi-modal models as pivotal data. In terms of visual features, our model avoids the need of object detectors and employs grid-like features together with a single objective of prompt language modeling. Experimentally, we consistently outperform existing methods in terms of caption quality and capability of describing out-of-domain concepts. Finally, our model obtains a new state of the art on both COCO and nocaps.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a model which induces a separation between content and descriptive style through the incorporation of stylistic parameters and keywords extracted from large-scale multi-modal models as pivotal data, and obtains a new state of the art on both COCO and nocaps.'}",
-prompting large language models with answer heuristics for knowledge-based visual question answering,Zhenwei Shao,"Knowledge-based visual question answering (VQA) requires external knowledge beyond the image to answer the question. Early studies retrieve required knowledge from explicit knowledge bases (KBs), which often introduces irrelevant information to the question, hence restricting the performance of their models. Recent works have sought to use a large language model (i.e., GPT-3 [3]) as an implicit knowledge engine to acquire the necessary knowledge for answering. Despite the encouraging results achieved by these methods, we argue that they have not fully activated the capacity of GPT-3 as the provided input information is insufficient. In this paper, we present Prophet-a conceptually simple framework designed to $prompt$ GPT-3 with answer heuristics for knowledge-based VQA. Specifically, we first train a vanilla VQA model on a specific knowledge-based VQA dataset without external knowledge. After that, we extract two types of complementary answer heuristics from the model: answer candidates and answer-aware examples. Finally, the two types of answer heuristics are encoded into the prompts to enable GPT-3 to better comprehend the task thus enhancing its capacity. Prophet significantly outperforms all existing state-of-the-art methods on two challenging knowledge-based VQA datasets, OK-VQA and A-OKVQA, delivering 61.1% and 55.7% accuracies on their testing sets, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Prophet-a conceptually simple framework designed to enable GPT-3 to better comprehend the task thus enhancing its capacity, which significantly outperforms all existing state-of-the-art methods on two challenging knowledge-based VQA datasets, OK-VQA and A-OKVZA.'}",https://arxiv.org/pdf/2303.01903
-large language models are effective text rankers with pairwise ranking prompting,Zhen Qin,"Ranking documents using Large Language Models (LLMs) by directly feeding the query and candidate documents into the prompt is an interesting and practical problem. However, there has been limited success so far, as researchers have found it difficult to outperform fine-tuned baseline rankers on benchmark datasets. We analyze pointwise and listwise ranking prompts used by existing methods and argue that off-the-shelf LLMs do not fully understand these ranking formulations, possibly due to the nature of how LLMs are trained. In this paper, we propose to significantly reduce the burden on LLMs by using a new technique called Pairwise Ranking Prompting (PRP). Our results are the first in the literature to achieve state-of-the-art ranking performance on standard benchmarks using moderate-sized open-sourced LLMs. On TREC-DL2020, PRP based on the Flan-UL2 model with 20B parameters outperforms the previous best approach in the literature, which is based on the blackbox commercial GPT-4 that has 50x (estimated) model size, by over 5% at NDCG@1. On TREC-DL2019, PRP is only inferior to the GPT-4 solution on the NDCG@5 and NDCG@10 metrics, while outperforming other existing solutions, such as InstructGPT which has 175B parameters, by over 10% for nearly all ranking metrics. Furthermore, we propose several variants of PRP to improve efficiency and show that it is possible to achieve competitive results even with linear complexity. We also discuss other benefits of PRP, such as supporting both generation and scoring LLM APIs, as well as being insensitive to input ordering.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to significantly reduce the burden on LLMs by using a new technique called Pairwise Ranking Prompting (PRP), and is the first in the literature to achieve state-of-the-art ranking performance on standard benchmarks using moderate-sized open-sourced LLMs.'}",http://arxiv.org/pdf/2306.17563
-"chatgpt is fun, but it is not funny! humor is still challenging large language models",Sophie F. Jentzsch,"Humor is a central aspect of human communication that has not been solved for artificial agents so far. Large language models (LLMs) are increasingly able to capture implicit and contextual information. Especially, OpenAI’s ChatGPT recently gained immense public attention. The GPT3-based model almost seems to communicate on a human level and can even tell jokes. Humor is an essential component of human communication. But is ChatGPT really funny?We put ChatGPT’s sense of humor to the test. In a series of exploratory experiments around jokes, i.e., generation, explanation, and detection, we seek to understand ChatGPT’s capability to grasp and reproduce human humor. Since the model itself is not accessible, we applied prompt-based experiments. Our empirical evidence indicates that jokes are not hard-coded but mostly also not newly generated by the model. Over 90% of 1008 generated jokes were the same 25 Jokes. The system accurately explains valid jokes but also comes up with fictional explanations for invalid jokes. Joke-typical characteristics can mislead ChatGPT in the classification of jokes. ChatGPT has not solved computational humor yet but it can be a big leap toward “funny” machines.","{'model': 'tldr@v2.0.0', 'text': 'The empirical evidence indicates that jokes are not hard-coded but mostly also not newly generated by the model, which means ChatGPT can be a big leap toward “funny” machines.'}",http://arxiv.org/pdf/2306.04563
-exploring the mit mathematics and eecs curriculum using large language models,Sarah J. Zhang,"We curate a comprehensive dataset of 4,550 questions and solutions from problem sets, midterm exams, and final exams across all MIT Mathematics and Electrical Engineering and Computer Science (EECS) courses required for obtaining a degree. We evaluate the ability of large language models to fulfill the graduation requirements for any MIT major in Mathematics and EECS. Our results demonstrate that GPT-3.5 successfully solves a third of the entire MIT curriculum, while GPT-4, with prompt engineering, achieves a perfect solve rate on a test set excluding questions based on images. We fine-tune an open-source large language model on this dataset. We employ GPT-4 to automatically grade model responses, providing a detailed performance breakdown by course, question, and answer type. By embedding questions in a low-dimensional space, we explore the relationships between questions, topics, and classes and discover which questions and classes are required for solving other questions and classes through few-shot learning. Our analysis offers valuable insights into course prerequisites and curriculum design, highlighting language models' potential for learning and improving Mathematics and EECS education.","{'model': 'tldr@v2.0.0', 'text': ""A comprehensive dataset of 4,550 questions and solutions from problem sets, midterm exams, and final exams across all MIT Mathematics and Electrical Engineering and Computer Science courses required for obtaining a degree is curated, highlighting language models' potential for learning and improving Mathematics and EECS education.""}",http://arxiv.org/pdf/2306.08997
-retroformer: retrospective large language agents with policy gradient optimization,Weiran Yao,"Recent months have seen the emergence of a powerful new trend in which large language models (LLMs) are augmented to become autonomous language agents capable of performing objective oriented multi-step tasks on their own, rather than merely responding to queries from human users. Most existing language agents, however, are not optimized using environment-specific rewards. Although some agents enable iterative refinement through verbal feedback, they do not reason and plan in ways that are compatible with gradient-based learning from rewards. This paper introduces a principled framework for reinforcing large language agents by learning a retrospective model, which automatically tunes the language agent prompts from environment feedback through policy gradient. Specifically, our proposed agent architecture learns from rewards across multiple environments and tasks, for fine-tuning a pre-trained language model which refines the language agent prompt by summarizing the root cause of prior failed attempts and proposing action plans. Experimental results on various tasks demonstrate that the language agents improve over time and that our approach considerably outperforms baselines that do not properly leverage gradients from the environment. This demonstrates that using policy gradient optimization to improve language agents, for which we believe our work is one of the first, seems promising and can be applied to optimize other models in the agent architecture to enhance agent performances over time.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the language agents improve over time and that the approach considerably outperforms baselines that do not properly leverage gradients from the environment, and demonstrates that using policy gradient optimization to improve language agents seems promising and can be applied to optimize other models in the agent architecture to enhance agent performances over time.'}",https://arxiv.org/pdf/2308.02151
-"chatgpt, large language models, and generative ai as future augments of surgical cancer care",A. Kothari,,"{'model': 'tldr@v2.0.0', 'text': 'ChatGPT’s popularity is a result of its advanced technology, versatility, and accessibility, which make it a valuable resource for a wide range of users and applications, and where it excelled by showcasing the capabilities of AI in understanding and responding to natural language.'}",https://link.springer.com/content/pdf/10.1245/s10434-023-13442-2.pdf
-sequential monte carlo steering of large language models using probabilistic programs,Alexander K. Lew,"Even after fine-tuning and reinforcement learning, large language models (LLMs) can be difficult, if not impossible, to control reliably with prompts alone. We propose a new inference-time approach to enforcing syntactic and semantic constraints on the outputs of LLMs, called sequential Monte Carlo (SMC) steering. The key idea is to specify language generation tasks as posterior inference problems in a class of discrete probabilistic sequence models, and replace standard decoding with sequential Monte Carlo inference. For a computational cost similar to that of beam search, SMC can steer LLMs to solve diverse tasks, including infilling, generation under syntactic constraints, and prompt intersection. To facilitate experimentation with SMC steering, we present a probabilistic programming library, LLaMPPL (https://github.com/probcomp/LLaMPPL), for concisely specifying new generation tasks as language model probabilistic programs, and automating steering of LLaMA-family Transformers.","{'model': 'tldr@v2.0.0', 'text': 'A Probabilistic programming library, LLaMPPL, is presented, for concisely specifying new generation tasks as language model probabilistic programs, and automating steering of LLaMA-family Transformers, to facilitate experimentation with SMC steering.'}",http://arxiv.org/pdf/2306.03081
-fineval: a chinese financial domain knowledge evaluation benchmark for large language models,Liwen Zhang,"Large language models (LLMs) have demonstrated exceptional performance in various natural language processing tasks, yet their efficacy in more challenging and domain-specific tasks remains largely unexplored. This paper presents FinEval, a benchmark specifically designed for the financial domain knowledge in the LLMs. FinEval is a collection of high-quality multiple-choice questions covering Finance, Economy, Accounting, and Certificate. It includes 4,661 questions spanning 34 different academic subjects. To ensure a comprehensive model performance evaluation, FinEval employs a range of prompt types, including zero-shot and few-shot prompts, as well as answer-only and chain-of-thought prompts. Evaluating state-of-the-art Chinese and English LLMs on FinEval, the results show that only GPT-4 achieved an accuracy close to 70% in different prompt settings, indicating significant growth potential for LLMs in the financial domain knowledge. Our work offers a more comprehensive financial knowledge evaluation benchmark, utilizing data of mock exams and covering a wide range of evaluated LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Evaluating state-of-the-art Chinese and English LLMs on FinEval, the results show that only GPT-4 achieved an accuracy close to 70% in different prompt settings, indicating significant growth potential for LLMs in the financial domain knowledge.'}",https://arxiv.org/pdf/2308.09975
-leveraging large language models for mental health prediction via online text data,Xuhai Xu,"The recent technology boost of large language models (LLMs) has empowered a variety of applications. However, there is very little research on understanding and improving LLMs’ capability for the mental health domain. In this work, we present the first comprehensive evaluation of multiple LLMs, including Alpaca, Alpaca-LoRA, and GPT-3.5, on various mental health prediction tasks via online text data. We conduct a wide range of experiments, covering zero-shot prompting, few-shot prompting, and instruction finetuning. The results indicate the promising yet limited performance of LLMs with zero-shot and few-shot prompt designs for mental health tasks. More importantly, our experiments show that instruction finetuning can significantly boost the performance of LLMs for all tasks simultaneously. Our best-finetuned model, Mental-Alpaca, outperforms GPT-3.5 (25 times bigger) by 16.7% on balanced accuracy and performs on par with the state-of-the-art task-specific model. We summarize our findings into a set of action guidelines for future researchers, engineers, and practitioners on how to empower LLMs with better mental health domain knowledge and become an expert in mental health prediction tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the first comprehensive evaluation of multiple LLMs, including AlPaca, Alpaca-LoRA, and GPT-3.5, on various mental health prediction tasks via online text data and shows that instruction finetuning can significantly boost the performance of LLMs for all tasks simultaneously.'}",https://arxiv.org/pdf/2307.14385
-analyzing chain-of-thought prompting in large language models via gradient-based feature attributions,Skyler Wu,"Chain-of-thought (CoT) prompting has been shown to empirically improve the accuracy of large language models (LLMs) on various question answering tasks. While understanding why CoT prompting is effective is crucial to ensuring that this phenomenon is a consequence of desired model behavior, little work has addressed this; nonetheless, such an understanding is a critical prerequisite for responsible model deployment. We address this question by leveraging gradient-based feature attribution methods which produce saliency scores that capture the influence of input tokens on model output. Specifically, we probe several open-source LLMs to investigate whether CoT prompting affects the relative importances they assign to particular input tokens. Our results indicate that while CoT prompting does not increase the magnitude of saliency scores attributed to semantically relevant tokens in the prompt compared to standard few-shot prompting, it increases the robustness of saliency scores to question perturbations and variations in model output.","{'model': 'tldr@v2.0.0', 'text': 'This work probes several open-source LLMs to investigate whether CoT prompting affects the relative importances they assign to particular input tokens, and results indicate that while coT prompting does not increase the magnitude of saliency scores attributed to semantically relevant tokens in the prompt compared to standard few-shot prompting, it increases the robustness ofsaliency scores to question perturbations and variations in model output.'}",https://arxiv.org/pdf/2307.13339
-"utilizing large language models to simplify radiology reports: a comparative analysis of chatgpt-3.5, chatgpt-4.0, google bard, and microsoft bing",Bing,"This paper investigates the application of Large Language Models (LLMs), specifically OpenAI's ChatGPT-3.5, ChatGPT-4.0, Google Bard, and Microsoft Bing, in simplifying radiology reports, thus potentially enhancing patient understanding. We examined 254 anonymized radiology reports from diverse examination types and used three different prompts to guide the LLMs' simplification processes. The resulting simplified reports were evaluated using four established readability indices. All LLMs significantly simplified the reports, but performance varied based on the prompt used and the specific model. The ChatGPT models performed best when additional context was provided (i.e., specifying user as a patient or requesting simplification at the 7th grade level). Our findings suggest that LLMs can effectively simplify radiology reports, although improvements are needed to ensure accurate clinical representation and optimal readability. These models have the potential to improve patient health literacy, patient-provider communication, and ultimately, health outcomes.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that LLMs can effectively simplify radiology reports, although improvements are needed to ensure accurate clinical representation and optimal readability.'}",https://www.medrxiv.org/content/medrxiv/early/2023/06/07/2023.06.04.23290786.full.pdf
-can large language models empower molecular property prediction?,Chen Qian,"Molecular property prediction has gained significant attention due to its transformative potential in multiple scientific disciplines. Conventionally, a molecule graph can be represented either as a graph-structured data or a SMILES text. Recently, the rapid development of Large Language Models (LLMs) has revolutionized the field of NLP. Although it is natural to utilize LLMs to assist in understanding molecules represented by SMILES, the exploration of how LLMs will impact molecular property prediction is still in its early stage. In this work, we advance towards this objective through two perspectives: zero/few-shot molecular classification, and using the new explanations generated by LLMs as representations of molecules. To be specific, we first prompt LLMs to do in-context molecular classification and evaluate their performance. After that, we employ LLMs to generate semantically enriched explanations for the original SMILES and then leverage that to fine-tune a small-scale LM model for multiple downstream tasks. The experimental results highlight the superiority of text explanations as molecular representations across multiple benchmark datasets, and confirm the immense potential of LLMs in molecular property prediction tasks. Codes are available at \url{https://github.com/ChnQ/LLM4Mol}.","{'model': 'tldr@v2.0.0', 'text': 'This work employs LLMs to generate semantically enriched explanations for the original SMILES and then uses that to fine-tune a small-scale LM model for multiple downstream tasks and highlights the superiority of text explanations as molecular representations across multiple benchmark datasets.'}",https://arxiv.org/pdf/2307.07443
-understanding the effectiveness of very large language models on dialog evaluation,Jessica Huynh,"Language models have steadily increased in size over the past few years. They achieve a high level of performance on various natural language processing (NLP) tasks such as question answering and summarization. Large language models (LLMs) have been used for generation and can now output human-like text. Due to this, there are other downstream tasks in the realm of dialog that can now harness the LLMs' language understanding capabilities. Dialog evaluation is one task that this paper will explore. It concentrates on prompting with LLMs: BLOOM, OPT, GPT-3, Flan-T5, InstructDial and TNLGv2. The paper shows that the choice of datasets used for training a model contributes to how well it performs on a task as well as on how the prompt should be structured. Specifically, the more diverse and relevant the group of datasets that a model is trained on, the better dialog evaluation performs. This paper also investigates how the number of examples in the prompt and the type of example selection used affect the model's performance.","{'model': 'tldr@v2.0.0', 'text': ""The paper shows that the choice of datasets used for training a model contributes to how well it performs on a task as well as on how the prompt should be structured, and investigates how the number of examples in the prompt and the type of example selection used affect the model's performance.""}",http://arxiv.org/pdf/2301.12004
-generating data for symbolic language with large language models,Jiacheng Ye,"While large language models (LLMs) bring not only performance but also complexity, recent work has started to turn LLMs into data generators rather than task inferencers, where another affordable task model is trained for efficient deployment and inference. However, such an approach has primarily been applied to natural language tasks and has not yet been explored for symbolic language tasks with complex structured outputs (e.g., semantic parsing and code generation). In this paper, we propose SymGen which utilizes LLMs for generating various annotation-expensive symbolic language data. SymGen consists of an informative prompt to steer generation and an agreement-based verifier to improve data correctness. We conduct extensive experiments on six symbolic language tasks across various settings. Compared with the LLMs, we demonstrate the 1\%-sized task model can achieve comparable or better performance, largely cutting inference and deployment costs. We also show that generated data with only a few human demonstrations can be as effective as over 10 times the amount of human-annotated data when training the task model, saving a considerable amount of annotation effort. SymGen sheds new light on data generation for complex tasks, and we release the code at \href{https://github.com/HKUNLP/SymGen}{https://github.com/HKUNLP/SymGen}.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes SymGen which utilizes LLMs for generating various annotation-expensive symbolic language data and demonstrates the 1\\%-sized task model can achieve comparable or better performance, largely cutting inference and deployment costs.'}",http://arxiv.org/pdf/2305.13917
-identifying and extracting rare disease phenotypes with large language models,Cathy Shyr,"Rare diseases (RDs) are collectively common and affect 300 million people worldwide. Accurate phenotyping is critical for informing diagnosis and treatment, but RD phenotypes are often embedded in unstructured text and time-consuming to extract manually. While natural language processing (NLP) models can perform named entity recognition (NER) to automate extraction, a major bottleneck is the development of a large, annotated corpus for model training. Recently, prompt learning emerged as an NLP paradigm that can lead to more generalizable results without any (zero-shot) or few labeled samples (few-shot). Despite growing interest in ChatGPT, a revolutionary large language model capable of following complex human prompts and generating high-quality responses, none have studied its NER performance for RDs in the zero- and few-shot settings. To this end, we engineered novel prompts aimed at extracting RD phenotypes and, to the best of our knowledge, are the first the establish a benchmark for evaluating ChatGPT's performance in these settings. We compared its performance to the traditional fine-tuning approach and conducted an in-depth error analysis. Overall, fine-tuning BioClinicalBERT resulted in higher performance (F1 of 0.689) than ChatGPT (F1 of 0.472 and 0.591 in the zero- and few-shot settings, respectively). Despite this, ChatGPT achieved similar or higher accuracy for certain entities (i.e., rare diseases and signs) in the one-shot setting (F1 of 0.776 and 0.725). This suggests that with appropriate prompt engineering, ChatGPT has the potential to match or outperform fine-tuned language models for certain entity types with just one labeled sample. While the proliferation of large language models may provide opportunities for supporting RD diagnosis and treatment, researchers and clinicians should critically evaluate model outputs and be well-informed of their limitations.","{'model': 'tldr@v2.0.0', 'text': ""This work engineered novel prompts aimed at extracting RD phenotypes and is the first the establish a benchmark for evaluating ChatGPT's performance in these settings, and compared its performance to the traditional fine-tuning approach and conducted an in-depth error analysis.""}",http://arxiv.org/pdf/2306.12656
-prompt tuning or fine-tuning - investigating relational knowledge in pre-trained language models,L. Fichtel,"Extracting relational knowledge from large pre-trained language models by a cloze-style sentence serving as a query has shown promising results. In particular, language models can be queried similar to knowledge graphs. The performance of the relational fact extraction task depends significantly on the query sentence, also known under the term prompt . Tuning these prompts has shown to increase the precision on standard language models by a maximum of around 12% points. However, usually large amounts of data in the form of existing knowledge graph facts and large text corpora are needed to train the required additional model. In this work, we propose using a completely different approach: Instead of spending resources on training an additional model, we simply perform an adaptive fine-tuning of the pre-trained language model on the standard fill-mask task using a small training dataset of existing facts from a knowledge graph. We investigate the differences between complex prompting techniques and adaptive fine-tuning in an extensive evaluation. Remarkably, adaptive fine-tuning outperforms all baselines, even by using significantly fewer training facts. Additionally, we analyze the transfer learning capabilities of this adapted language model by training on a restricted set of relations to show that even fewer training relations are needed to achieve high knowledge extraction quality.","{'model': 'tldr@v2.0.0', 'text': 'This work performs an adaptive tuning of the pre-trained language model on the standard task using a small training dataset of existing facts from a knowledge graph to show that even fewer training relations are needed to achieve high knowledge extraction quality.'}",
-on transferability of prompt tuning for natural language understanding,Yusheng Su,"Prompt tuning (PT) is a promising parameter-efficient method to utilize extremely large pre-trained language models (PLMs), which could achieve comparable performance to full-parameter fine-tuning by only tuning a few soft prompts. However, compared to fine-tuning, PT empirically requires much more training steps. To explore whether we can improve the efficiency of PT by reusing trained soft prompts and sharing learned knowledge, we empirically investigate the transferability of soft prompts across different tasks and models. In cross-task transfer , we find that trained soft prompts can well transfer to similar tasks and initialize PT for them to accelerate training and improve performance. Moreover, to explore what factors influence prompts’ transferability across tasks, we investigate how to measure the prompt similarity and find that the overlapping rate of activated neurons highly correlates to the transferability. In cross-model transfer , we explore how to project the prompts of a PLM to another PLM and successfully train a kind of projector which can achieve non-trivial transfer performance on similar tasks. However, initializing PT with the projected prompts does not work well, which may be caused by optimization preferences and PLMs’ high redundancy. Our findings show that improving PT with knowledge transfer is possible and promising, while prompts’ cross-task transferability is generally better than the cross-model transferability.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that improving PT with knowledge transfer is possible and promising, while prompts’ cross-task transferability is generally better than the cross-model transferability.'}",
-knowledge-augmented language model prompting for zero-shot knowledge graph question answering,Jinheon Baek,"Large Language Models (LLMs) are capable of performing zero-shot closed-book question answering tasks, based on their internal knowledge stored in parameters during pre-training. However, such internalized knowledge might be insufficient and incorrect, which could lead LLMs to generate factually wrong answers. Furthermore, fine-tuning LLMs to update their knowledge is expensive. To this end, we propose to augment the knowledge directly in the input of LLMs. Specifically, we first retrieve the relevant facts to the input question from the knowledge graph based on semantic similarities between the question and its associated facts. After that, we prepend the retrieved facts to the input question in the form of the prompt, which is then forwarded to LLMs to generate the answer. Our framework, Knowledge-Augmented language model PromptING (KAPING), requires no model training, thus completely zero-shot. We validate the performance of our KAPING framework on the knowledge graph question answering task, that aims to answer the user’s question based on facts over a knowledge graph, on which ours outperforms relevant zero-shot baselines by up to 48% in average, across multiple LLMs of various sizes.","{'model': 'tldr@v2.0.0', 'text': ""The framework, Knowledge-Augmented language model PromptING (KAPING), requires no model training, thus completely zero-shot, and aims to answer the user’s question based on facts over a knowledge graph, on which the authors' outperforms relevant zero- shot baselines by up to 48% in average, across multiple LLMs of various sizes.""}",http://arxiv.org/pdf/2306.04136
-purr: efficiently editing language model hallucinations by denoising language model corruptions,Anthony Chen,"The remarkable capabilities of large language models have been accompanied by a persistent drawback: the generation of false and unsubstantiated claims commonly known as""hallucinations"". To combat this issue, recent research has introduced approaches that involve editing and attributing the outputs of language models, particularly through prompt-based editing. However, the inference cost and speed of using large language models for editing currently bottleneck prompt-based methods. These bottlenecks motivate the training of compact editors, which is challenging due to the scarcity of training data for this purpose. To overcome these challenges, we exploit the power of large language models to introduce corruptions (i.e., noise) into text and subsequently fine-tune compact editors to denoise the corruptions by incorporating relevant evidence. Our methodology is entirely unsupervised and provides us with faux hallucinations for training in any domain. Our Petite Unsupervised Research and Revision model, PURR, not only improves attribution over existing editing methods based on fine-tuning and prompting, but also achieves faster execution times by orders of magnitude.","{'model': 'tldr@v2.0.0', 'text': 'The Petite Unsupervised Research and Revision model, PURR, not only improves attribution over existing editing methods based on fine-tuning and prompting, but also achieves faster execution times by orders of magnitude.'}",http://arxiv.org/pdf/2305.14908
-revealing the unwritten: visual investigation of beam search trees to address language model prompting challenges,Thilo Spinner,"The growing popularity of generative language models has amplified interest in interactive methods to guide model outputs. Prompt refinement is considered one of the most effective means to influence output among these methods. We identify several challenges associated with prompting large language models, categorized into data- and model-specific, linguistic, and socio-linguistic challenges. A comprehensive examination of model outputs, including runner-up candidates and their corresponding probabilities, is needed to address these issues. The beam search tree, the prevalent algorithm to sample model outputs, can inherently supply this information. Consequently, we introduce an interactive visual method for investigating the beam search tree, facilitating analysis of the decisions made by the model during generation. We quantitatively show the value of exposing the beam search tree and present five detailed analysis scenarios addressing the identified challenges. Our methodology validates existing results and offers additional insights.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces an interactive visual method for investigating the beam search tree, facilitating analysis of the decisions made by the model during generation, and quantitatively shows the value of exposing thebeam search tree.'}",
-training language models to follow instructions with human feedback,Long Ouyang,"Making language models bigger does not inherently make them better at following a user's intent. For example, large language models can generate outputs that are untruthful, toxic, or simply not helpful to the user. In other words, these models are not aligned with their users. In this paper, we show an avenue for aligning language models with user intent on a wide range of tasks by fine-tuning with human feedback. Starting with a set of labeler-written prompts and prompts submitted through the OpenAI API, we collect a dataset of labeler demonstrations of the desired model behavior, which we use to fine-tune GPT-3 using supervised learning. We then collect a dataset of rankings of model outputs, which we use to further fine-tune this supervised model using reinforcement learning from human feedback. We call the resulting models InstructGPT. In human evaluations on our prompt distribution, outputs from the 1.3B parameter InstructGPT model are preferred to outputs from the 175B GPT-3, despite having 100x fewer parameters. Moreover, InstructGPT models show improvements in truthfulness and reductions in toxic output generation while having minimal performance regressions on public NLP datasets. Even though InstructGPT still makes simple mistakes, our results show that fine-tuning with human feedback is a promising direction for aligning language models with human intent.","{'model': 'tldr@v2.0.0', 'text': 'The results show that fine-tuning with human feedback is a promising direction for aligning language models with human intent and showing improvements in truthfulness and reductions in toxic output generation while having minimal performance regressions on public NLP datasets.'}",http://arxiv.org/pdf/2203.02155
-"translating radiology reports into plain language using chatgpt and gpt-4 with prompt learning: results, limitations, and potential",Qing Lyu,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.'}",https://vciba.springeropen.com/counter/pdf/10.1186/s42492-023-00136-5
-"translating radiology reports into plain language using chatgpt and gpt-4 with prompt learning: promising results, limitations, and potential",Qing Lyu,"The large language model called ChatGPT has drawn extensively attention because of its human-like expression and reasoning abilities. In this study, we investigate the feasibility of using ChatGPT in experiments on using ChatGPT to translate radiology reports into plain language for patients and healthcare providers so that they are educated for improved healthcare. Radiology reports from 62 low-dose chest CT lung cancer screening scans and 76 brain MRI metastases screening scans were collected in the first half of February for this study. According to the evaluation by radiologists, ChatGPT can successfully translate radiology reports into plain language with an average score of 4.27 in the five-point system with 0.08 places of information missing and 0.07 places of misinformation. In terms of the suggestions provided by ChatGPT, they are general relevant such as keeping following-up with doctors and closely monitoring any symptoms, and for about 37% of 138 cases in total ChatGPT offers specific suggestions based on findings in the report. ChatGPT also presents some randomness in its responses with occasionally over-simplified or neglected information, which can be mitigated using a more detailed prompt. Furthermore, ChatGPT results are compared with a newly released large model GPT-4, showing that GPT-4 can significantly improve the quality of translated reports. Our results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.","{'model': 'tldr@v2.0.0', 'text': 'The results show that it is feasible to utilize large language models in clinical education, and further efforts are needed to address limitations and maximize their potential.'}",http://arxiv.org/pdf/2303.09038
-a systematic survey of prompt engineering on vision-language foundation models,Jindong Gu,"—Prompt engineering is a technique that involves augmenting a large pre-trained model with task-specific hints, known as prompts, to adapt the model to new tasks. Prompts can be created manually as natural language instructions or generated automatically as either natural language instructions or vector representations. Prompt engineering enables the ability to perform predictions based solely on prompts without updating model parameters, and the easier application of large pre-trained models in real-world tasks. In past years, Prompt engineering has been well-studied in natural language processing. Recently, it has also been intensively studied in vision-language modeling. However, there is currently a lack of a systematic overview of prompt engineering on pre-trained vision-language models. This paper aims to provide a comprehensive survey of cutting-edge research in prompt engineering on three types of vision-language models: multimodal-to-text generation models ( e.g., Flamingo), image-text matching models ( e.g., CLIP), and text-to-image generation models ( e.g., Stable Diffusion). For each type of model, a brief model summary, prompting methods, prompting-based applications, and the corresponding responsibility and integrity issues are summarized and discussed. Furthermore, the commonalities and differences between prompting on vision-language models, language models, and vision models are also discussed. The challenges, future directions, and research opportunities are summarized to foster future research on this topic.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to provide a comprehensive survey of cutting-edge research in prompt engineering on three types of vision-language models: multimodal-to-text generation models, image-text matching models, and text- to-image generation models.'}",https://arxiv.org/pdf/2307.12980
-pouf: prompt-oriented unsupervised fine-tuning for large pre-trained models,Korawat Tanwisuth,"Through prompting, large-scale pre-trained models have become more expressive and powerful, gaining significant attention in recent years. Though these big models have zero-shot capabilities, in general, labeled data are still required to adapt them to downstream tasks. To overcome this critical limitation, we propose an unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data. We demonstrate how to apply our method to both language-augmented vision and masked-language models by aligning the discrete distributions extracted from the prompts and target data. To verify our approach's applicability, we conduct extensive experiments on image classification, sentiment analysis, and natural language inference tasks. Across 13 image-related tasks and 15 language-related ones, the proposed approach achieves consistent improvements over the baselines.","{'model': 'tldr@v2.0.0', 'text': 'An unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data by aligning the discrete distributions extracted from the prompts and target data is proposed.'}",http://arxiv.org/pdf/2305.00350
-model ensemble instead of prompt fusion: a sample-specific knowledge transfer method for few-shot prompt tuning,Xiangyu Peng,"Prompt tuning approaches, which learn task-specific soft prompts for a downstream task conditioning on frozen pre-trained models, have attracted growing interest due to its parameter efficiency. With large language models and sufficient training data, prompt tuning performs comparably to full-model tuning. However, with limited training samples in few-shot settings, prompt tuning fails to match the performance of full-model fine-tuning. In this work, we focus on improving the few-shot performance of prompt tuning by transferring knowledge from soft prompts of source tasks. Recognizing the good generalization capabilities of ensemble methods in low-data regime, we first experiment and show that a simple ensemble of model predictions based on different source prompts, outperforms existing multi-prompt knowledge transfer approaches such as source prompt fusion in the few-shot setting. Motivated by this observation, we further investigate model ensembles and propose Sample-specific Ensemble of Source Models (SESoM). SESoM learns to adjust the contribution of each source model for each target sample separately when ensembling source model outputs. Through this way, SESoM inherits the superior generalization of model ensemble approaches and simultaneously captures the sample-specific competence of each source prompt. We conduct experiments across a diverse set of eight NLP tasks using models of different scales (T5-{base, large, XL}) and find that SESoM consistently outperforms the existing models of the same as well as larger parametric scale by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates model ensembles and proposes Sample-specific Ensemble of Source Models (SESoM), a simple ensemble of model predictions based on different source prompts that consistently outperforms the existing models of the same as well as larger parametric scale by a large margin.'}",http://arxiv.org/pdf/2210.12587
-attentional mixtures of soft prompt tuning for parameter-efficient multi-task knowledge sharing,Akari Asai,"This work introduces ATTEMPT (ATTEntional Mixture of Prompt Tuning), a new modular, multi-task, and parameterefficient language model (LM) tuning approach that combines knowledge transferred across different tasks via a mixture of soft prompts while keeping original LM unchanged. ATTEMPT interpolates a set of prompts trained on large-scale source tasks and a newly initialized target task prompt using instance-wise attention computed by a light-weight sub-network trained on multiple target tasks. ATTEMPT is parameter-efficient (e.g., updates 1,600 times fewer parameters than fine-tuning) and enables multi-task learning and flexible extensions; importantly, it is also more interpretable because it demonstrates which source tasks affect the final model decision on target tasks. Experimental results across 17 diverse datasets show that ATTEMPT improves prompt tuning by up to a 22% absolute performance gain and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use over ten times more parameters.1","{'model': 'tldr@v2.0.0', 'text': 'Experimental results across 17 diverse datasets show that ATTEMPT improves prompt tuning by up to a 22% absolute performance gain and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use over ten times more parameters.'}",https://arxiv.org/pdf/2205.11961
-prompting large pre-trained vision-language models for compositional concept learning,Guangyue Xu,"This work explores the zero-shot compositional learning ability of large pre-trained vision-language models(VLMs) within the prompt-based learning framework and propose a model (\textit{PromptCompVL}) to solve the compositonal zero-shot learning (CZSL) problem. \textit{PromptCompVL} makes two design choices: first, it uses a soft-prompting instead of hard-prompting to inject learnable parameters to reprogram VLMs for compositional learning. Second, to address the compositional challenge, it uses the soft-embedding layer to learn primitive concepts in different combinations. By combining both soft-embedding and soft-prompting, \textit{PromptCompVL} achieves state-of-the-art performance on the MIT-States dataset. Furthermore, our proposed model achieves consistent improvement compared to other CLIP-based methods which shows the effectiveness of the proposed prompting strategies for CZSL.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the zero-shot compositional learning ability of large pre-trained vision-language models within the prompt-based learning framework and proposes a model that achieves consistent improvement compared to other CLIP-based methods which shows the effectiveness of the proposed prompting strategies for CZSL.'}",https://arxiv.org/pdf/2211.05077
-quantifying memorization across neural language models,Nicholas Carlini,"Large language models (LMs) have been shown to memorize parts of their training data, and when prompted appropriately, they will emit the memorized training data verbatim. This is undesirable because memorization violates privacy (exposing user data), degrades utility (repeated easy-to-memorize text is often low quality), and hurts fairness (some texts are memorized over others). We describe three log-linear relationships that quantify the degree to which LMs emit memorized training data. Memorization significantly grows as we increase (1) the capacity of a model, (2) the number of times an example has been duplicated, and (3) the number of tokens of context used to prompt the model. Surprisingly, we find the situation becomes more complicated when generalizing these results across model families. On the whole, we find that memorization in LMs is more prevalent than previously believed and will likely get worse as models continues to scale, at least without active mitigations.","{'model': 'tldr@v2.0.0', 'text': 'On the whole, this work finds that memorization in LMs is more prevalent than previously believed and will likely get worse as models continues to scale, at least without active mitigations.'}",
-proqa: structural prompt-based pre-training for unified question answering,Wanjun Zhong,"Question Answering (QA) is a longstanding challenge in natural language processing. Existing QA works mostly focus on specific question types, knowledge domains, or reasoning skills. The specialty in QA research hinders systems from modeling commonalities between tasks and generalization for wider applications. To address this issue, we present ProQA, a unified QA paradigm that solves various tasks through a single model. ProQA takes a unified structural prompt as the bridge and improves the QA-centric ability by structural prompt-based pre-training. Through a structurally designed prompt-based input schema, ProQA concurrently models the knowledge generalization for all QA tasks while keeping the knowledge customization for every specific QA task. Furthermore, ProQA is pre-trained with structural prompt-formatted large-scale synthesized corpus, which empowers the model with the commonly-required QA ability. Experimental results on 11 QA benchmarks demonstrate that ProQA consistently boosts performance on both full data fine-tuning, few-shot learning, and zero-shot testing scenarios. Furthermore, ProQA exhibits strong ability in both continual learning and transfer learning by taking the advantages of the structural prompt.","{'model': 'tldr@v2.0.0', 'text': 'ProQA is a unified QA paradigm that solves various tasks through a single model that takes a unified structural prompt as the bridge and improves the QA-centric ability by structural prompt-based pre-training.'}",http://arxiv.org/pdf/2205.04040
-novelty controlled paraphrase generation with retrieval augmented conditional prompt tuning,Jishnu Ray Chowdhury,"Paraphrase generation is a fundamental and long-standing task in natural language processing. In this paper, we concentrate on two contributions to the task: (1) we propose Retrieval Augmented Prompt Tuning (RAPT) as a parameter-efficient method to adapt large pre-trained language models for paraphrase generation; (2) we propose Novelty Conditioned RAPT (NC-RAPT) as a simple model-agnostic method of using specialized prompt tokens for controlled paraphrase generation with varying levels of lexical novelty. By conducting extensive experiments on four datasets, we demonstrate the effectiveness of the proposed approaches for retaining the semantic content of the original text while inducing lexical novelty in the generation.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of the proposed approaches for retaining the semantic content of the original text while inducing lexical novelty in the generation of paraphrase generation is demonstrated.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21297/21046
-discovering latent knowledge in language models without supervision,Collin Burns,"Existing techniques for training language models can be misaligned with the truth: if we train models with imitation learning, they may reproduce errors that humans make; if we train them to generate text that humans rate highly, they may output errors that human evaluators can't detect. We propose circumventing this issue by directly finding latent knowledge inside the internal activations of a language model in a purely unsupervised way. Specifically, we introduce a method for accurately answering yes-no questions given only unlabeled model activations. It works by finding a direction in activation space that satisfies logical consistency properties, such as that a statement and its negation have opposite truth values. We show that despite using no supervision and no model outputs, our method can recover diverse knowledge represented in large language models: across 6 models and 10 question-answering datasets, it outperforms zero-shot accuracy by 4\% on average. We also find that it cuts prompt sensitivity in half and continues to maintain high accuracy even when models are prompted to generate incorrect answers. Our results provide an initial step toward discovering what language models know, distinct from what they say, even when we don't have access to explicit ground truth labels.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a method for accurately answering yes-no questions given only unlabeled model activations, and shows that despite using no supervision and no model outputs, the method can recover diverse knowledge represented in large language models.'}",https://arxiv.org/pdf/2212.03827
-discup: discriminator cooperative unlikelihood prompt-tuning for controllable text generation,Hanqing Zhang,"Prompt learning with immensely large Casual Language Models (CLMs) has been shown promising for attribute-controllable text generation (CTG). However, vanilla prompt tuning tends to imitate training corpus characteristics beyond the control attributes, resulting in a poor generalization ability. Moreover, it is less able to capture the relationship between different attributes, further limiting the control performance. In this paper, we propose a new CTG approach, namely DisCup, which incorporates the attribute knowledge of discriminator to optimize the control-prompts, steering a frozen CLM to produce attribute-specific texts. Specifically, the frozen CLM model, capable of producing multitudinous texts, is first used to generate the next-token candidates based on the context, so as to ensure the diversity of tokens to be predicted. Then, we leverage an attribute-discriminator to select desired/undesired tokens from those candidates, providing the inter-attribute knowledge. Finally, we bridge the above two traits by an unlikelihood objective for prompt-tuning. Extensive experimental results show that DisCup can achieve a new state-of-the-art control performance while maintaining an efficient and high-quality text generation, only relying on around 10 virtual tokens.","{'model': 'tldr@v2.0.0', 'text': 'A new CTG approach, namely DisCup, is proposed, which incorporates the attribute knowledge of discriminator to optimize the control-prompts, steering a frozen CLM to produce attribute-specific texts.'}",http://arxiv.org/pdf/2210.09551
-deep continuous prompt for contrastive learning of sentence embeddings,Yuxin Jiang,"The performance of sentence representation 001 has been remarkably improved by the frame- 002 work of contrastive learning. However, recent 003 works still require full fine-tuning, which is 004 quite inefficient for large-scaled pre-trained lan- 005 guage models. To this end, we present a novel 006 method which freezes the whole language 007 model and only optimizes the prefix deep con- 008 tinuous prompts. It not only tunes around 0.1% 009 parameters of the original language model, but 010 avoids the cumbersome computation of search- 011 ing handcrafted prompts. Experimental results 012 show that our proposed DCPCSE outperforms 013 the state-of-the-art method SimCSE by a large 014 margin. We raise the performance of unsuper- 015 vised BERT base and supervised RoBERTa large 016 by 2.24 and 1.00 points, respectively. Our code 017 will be released at Github. 018","{'model': 'tldr@v2.0.0', 'text': 'A novel 006 method which freezes the whole language 007 model and only optimizes the prefix deep con- 008 tinuous prompts and avoids the cumbersome computation of search- 011 ing handcrafted prompts is presented.'}",http://arxiv.org/pdf/2203.06875
-language models with image descriptors are strong few-shot video-language learners,Zhenhailong Wang,"The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples, such as domain-specific captioning, question answering, and future event prediction. Existing few-shot video-language learners focus exclusively on the encoder, resulting in the absence of a video-to-text decoder to handle generative tasks. Video captioners have been pretrained on large-scale video-language datasets, but they rely heavily on finetuning and lack the ability to generate text for unseen tasks in a few-shot setting. We propose VidIL, a few-shot Video-language Learner via Image and Language models, which demonstrates strong performance on few-shot video-to-text tasks without the necessity of pretraining or finetuning on any video datasets. We use the image-language models to translate the video content into frame captions, object, attribute, and event phrases, and compose them into a temporal structure template. We then instruct a language model, with a prompt containing a few in-context examples, to generate a target output from the composed content. The flexibility of prompting allows the model to capture any form of text input, such as automatic speech recognition (ASR) transcripts. Our experiments demonstrate the power of language models in understanding videos on a wide variety of video-language tasks, including video captioning, video question answering, video caption retrieval, and video future event prediction. Especially, on video future event prediction, our few-shot model significantly outperforms state-of-the-art supervised models trained on large-scale video datasets. Code and resources are publicly available for research purposes at https://github.com/MikeWangWZHL/VidIL .","{'model': 'tldr@v2.0.0', 'text': 'The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples, such as domain-specific captioning, question answering, and future event prediction, and to demonstrate the power of language models in understanding videos on a wide variety of video- language tasks.'}",http://arxiv.org/pdf/2205.10747
-improving the sample efficiency of prompt tuning with domain adaptation,Xu Guo,"Prompt tuning, or the conditioning of a frozen pretrained language model (PLM) with soft prompts learned from data, has demonstrated impressive performance on a wide range of NLP tasks. However, prompt tuning requires a large training dataset to be effective and is outperformed by finetuning the entire PLM in data-scarce regimes. Previous work (Gu et al., 2022, Vu et al., 2022) proposed to transfer soft prompts pretrained on the source domain to the target domain. In this paper, we explore domain adaptation for prompt tuning, a problem setting where unlabeled data from the target domain are available during pretraining. We propose bOosting Prompt TunIng with doMain Adaptation (OPTIMA), which regularizes the decision boundary to be smooth around regions where source and target data distributions are similar. Extensive experiments demonstrate that OPTIMA significantly enhances the transferability and sample-efficiency of prompt tuning compared to strong baselines. Moreover, in few-shot settings, OPTIMA exceeds full-model tuning by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes bOosting Prompt TunIng with doMain Adaptation (OPTIMA), which regularizes the decision boundary to be smooth around regions where source and target data distributions are similar, and demonstrates that OPTIMA significantly enhances the transferability and sample-efficiency of prompt tuning compared to strong baselines.'}",http://arxiv.org/pdf/2210.02952
-prompt-augmented linear probing: scaling beyond the limit of few-shot in-context learners,Hyunsoo Cho,"Through in-context learning (ICL), large-scale language models are effective few-shot learners without additional model fine-tuning. However, the ICL performance does not scale well with the number of available training sample as it is limited by the inherent input length constraint of the underlying language model. Meanwhile, many studies have revealed that language models are also powerful feature extractors, allowing them to be utilized in a black-box manner and enabling the linear probing paradigm, where lightweight discriminators are trained on top of the pre-extracted input representations. This paper proposes prompt-augmented linear probing (PALP), a hybrid of linear probing and ICL, which leverages the best of both worlds. PALP inherits the scalability of linear probing and the capability of enforcing language models to derive more meaningful representations via tailoring input into a more conceivable form. Throughout in-depth investigations on various datasets, we verified that PALP significantly closes the gap between ICL in the data-hungry scenario and fine-tuning in the data-abundant scenario with little training overhead, potentially making PALP a strong alternative in a black-box scenario.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2212.10873
-reduce communication costs and preserve privacy: prompt tuning method in federated learning,Haodong Zhao,"Federated learning (FL) has enabled global model training on decentralized data in a privacy-preserving way by aggregating model updates. However, for many natural language processing (NLP) tasks that utilize pre-trained language models (PLMs) with large numbers of parameters, there are con- siderable communication costs associated with FL. Recently, prompt tuning, which tunes some soft prompts without mod- ifying PLMs, has achieved excellent performance as a new learning paradigm. Therefore we want to combine the two methods and explore the effect of prompt tuning under FL. In this paper, we propose ”FedPrompt” as the first work study prompt tuning in a model split learning way using FL, and prove that split learning greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution. This improves the efficiency of FL method while also protect-ing the data privacy in prompt tuning. In addition, like PLMs, prompts are uploaded and downloaded between public platforms and personal users, so we try to figure out whether there is still a backdoor threat using only soft prompt in FL scenarios. We further conduct backdoor attacks by data poisoning on FedPrompt. Our experiments show that normal backdoor attack can not achieve a high attack success rate, proving the robustness of FedPrompt. We hope this work can promote the application of prompt in FL and raise the awareness of the possible security threats.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that split learning greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution, which improves theency of FL method while also protect-ing the data privacy in prompt tuning.'}",http://arxiv.org/pdf/2208.12268
-doubly right object recognition: a why prompt for visual rationales,Chengzhi Mao,"Many visual recognition models are evaluated only on their classification accuracy, a metric for which they obtain strong performance. In this paper, we investigate whether computer vision models can also provide correct rationales for their predictions. We propose a “doubly right” object recognition benchmark, where the metric requires the model to simultaneously produce both the right labels as well as the right rationales. We find that state-of-the-art visual models, such as CLIP, often provide incorrect rationales for their categorical predictions. However, by transferring the rationales from language models into visual representations through a tailored dataset, we show that we can learn a “why prompt,” which adapts large visual representations to produce correct rationales. Visualizations and empirical experiments show that our prompts significantly improve performance on doubly right object recognition, in addition to zero-shot transfer to unseen tasks and datasets.","{'model': 'tldr@v2.0.0', 'text': 'It is found that state-of-the-art visual models, such as CLIP, often provide incorrect rationales for their categorical predictions, but by transferring the rationales from language models into visual representations through a tailored dataset, it is shown that a “why prompt” can be learned, which adapts large visual representations to produce correct rationales.'}",https://arxiv.org/pdf/2212.06202
-xprompt: exploring the extreme of prompt tuning,Fang Ma,"Prompt tuning learns soft prompts to condition the frozen Pre-trained Language Models (PLMs) for performing downstream tasks in a parameter-efficient manner. While prompt tuning has gradually reached the performance level of fine-tuning as the model scale increases, there is still a large performance gap between prompt tuning and fine-tuning for models of moderate and small scales (typically less than 11B parameters). In this paper, we empirically show that the trained prompt tokens can have a negative impact on a downstream task and thus degrade its performance. To bridge the gap, we propose a novel Prompt tuning model with an eXtremely small scale (XPrompt) under the regime of lottery tickets hypothesis. Specifically, XPrompt eliminates the negative prompt tokens at different granularity levels through a hierarchical structured pruning, yielding a more parameter-efficient prompt yet with a competitive performance. Comprehensive experiments are carried out on the SuperGLUE tasks, and the results indicate that XPrompt is able to close the performance gap at smaller model scales.","{'model': 'tldr@v2.0.0', 'text': 'This paper empirically shows that the trained prompt tokens can have a negative impact on a downstream task and thus degrade its performance, and proposes a novel Prompt tuning model with an eXtremely small scale (XPrompt) under the regime of lottery tickets hypothesis that eliminates the negative prompt tokens at different granularity levels through a hierarchical structured pruning.'}",http://arxiv.org/pdf/2210.04457
-automatic prompt augmentation and selection with chain-of-thought from labeled data,Kashun Shum,"Chain-of-thought prompting (CoT) advances the reasoning abilities of large language models (LLMs) and achieves superior performance in arithmetic, commonsense, and symbolic reasoning tasks. However, most CoT studies rely on carefully designed human-annotated rational chains to prompt the language model, which poses challenges for real-world applications where labeled training data is available without human-annotated rational chains. This creates barriers to applications of CoT prompting to these general tasks. This paper proposes a new strategy, Automate-CoT (Automatic Prompt Augmentation and Selection with Chain-of-Thought), that can bypass human engineering of CoTs by automatically augmenting rational chains from a small labeled dataset, and then pruning low-quality chains to construct a candidate pool of machine-generated rationale chains based on the labels. Finally, it selects the optimal combination of several rationale chains from the pool for CoT prompting by employing a variance-reduced policy gradient strategy to estimate the significance of each example in a black-box language model. Automate-CoT enables a quick adaptation of the CoT technique to different tasks. Experimental results demonstrate the effectiveness of our method, where state-of-the-art results are achieved on arithmetic reasoning (+2.7\%), commonsense reasoning (+3.4\%), symbolic reasoning (+3.2\%), and non-reasoning tasks (+2.5\%). Our code will be available at https://github.com/shizhediao/automate-cot.","{'model': 'tldr@v2.0.0', 'text': 'A new strategy, Automate-CoT (Automatic Prompt Augmentation and Selection with Chain-of-Thought), that can bypass human engineering of CoTs by automatically augmenting rational chains from a small labeled dataset, and then pruning low-quality chains to construct a candidate pool of machine-generated rationale chains based on the labels.'}",http://arxiv.org/pdf/2302.12822
-multitask prompt tuning enables parameter-efficient transfer learning,Zhen Wang,"Prompt tuning, in which a base pretrained model is adapted to each task via conditioning on learned prompt vectors, has emerged as a promising approach for efficiently adapting large language models to multiple downstream tasks. However, existing methods typically learn soft prompt vectors from scratch, and it has not been clear how to exploit the rich cross-task knowledge with prompt vectors in a multitask learning setting. We propose multitask prompt tuning (MPT), which first learns a single transferable prompt by distilling knowledge from multiple task-specific source prompts. We then learn multiplicative low rank updates to this shared prompt to efficiently adapt it to each downstream target task. Extensive experiments on 23 NLP datasets demonstrate that our proposed approach outperforms the state-of-the-art methods, including the full finetuning baseline in some cases, despite only tuning 0.035% as many task-specific parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes multitask prompt tuning (MPT), which first learns a single transferable prompt by distilling knowledge from multiple task-specific source prompts, then learns multiplicative low rank updates to this shared prompt to efficiently adapt it to each downstream target task.'}",http://arxiv.org/pdf/2303.02861
-declaration-based prompt tuning for visual question answering,Yuhang Liu,"In recent years, the pre-training-then-fine-tuning paradigm has yielded immense success on a wide spectrum of cross-modal tasks, such as visual question answering (VQA), in which a visual-language (VL) model is first optimized via self-supervised task objectives, e.g., masked language modeling (MLM) and image-text matching (ITM), and then fine-tuned to adapt to downstream task (e.g., VQA) via a brand-new objective function, e.g., answer prediction. However, the inconsistency of the objective forms not only severely limits the generalization of pre-trained VL models to downstream tasks, but also requires a large amount of labeled data for fine-tuning. To alleviate the problem, we propose an innovative VL fine-tuning paradigm (named Declaration-based Prompt Tuning, abbreviated as DPT), which fine-tunes the model for downstream VQA using the pre-training objectives, boosting the effective adaptation of pre-trained models to the downstream task. Specifically, DPT reformulates the VQA task via (1) textual adaptation, which converts the given questions into declarative sentence form for prompt-tuning, and (2) task adaptation, which optimizes the objective function of VQA problem in the manner of pre-training phase. Experimental results on GQA dataset show that DPT outperforms the fine-tuned counterpart by a large margin regarding accuracy in both fully-supervised (2.68%) and zero-shot/fewshot (over 31%) settings. All the data and codes will be available to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'An innovative VL fine-tuning paradigm (named Declaration-based Prompt Tuning, abbreviated as DPT), which fine-tunes the model for downstream VQA using the pre-training objectives, boosting the effective adaptation of pre-trained models to the downstream task.'}",http://arxiv.org/pdf/2205.02456
-prompt generation networks for efficient adaptation of frozen vision transformers,J. Loedeman,"Large-scale pretrained models, especially those trained from vision-language data have demonstrated the tremendous value that can be gained from both larger training datasets and models. Thus, in order to benefit from these developments, there is renewed interest in transfer learning and adapting models from large-scale general pretraining to particular downstream tasks. However, the continuously increasing size of the models means that even the classic approach of finetuning is becoming infeasible for all but big institutions. Prompt leaning has emerged as a flexible way to adapt models by solely learning additional inputs to a model that is kept frozen, but so far performances remained inferior to finetuning. To address this, we propose the Prompt Generation Network (PGN) that generates input-dependent prompts by sampling from a learned library of tokens. We show the PGN is effective in adapting pretrained models to various new datasets. It surpasses previous prompt-learning methods by a large margin and even fullfinetuning on 5 out of 12 datasets while requiring 100x less parameters. PGN can even be used for training and inferring on multiple datasets simultaneously and learns to allocate tokens between domains. Given these findings, we conclude that PGN is a viable and scalable approach for downstream adaptation of frozen models. Code is available at https://github.com/jochemloedeman/PGN.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt Generation Network (PGN) is proposed that generates input-dependent prompts by sampling from a learned library of tokens and is shown to be effective in adapting pretrained models to various new datasets.'}",http://arxiv.org/pdf/2210.06466
-spt: semi-parametric prompt tuning for multitask prompted learning,M Saiful Bari,"Pre-trained large language models can efficiently interpolate human-written prompts in a natural way. Multitask prompted learning can help generalization through a diverse set of tasks at once, thus enhancing the potential for more effective downstream fine-tuning. To perform efficient multitask-inference in the same batch, parameter-efficient fine-tuning methods such as prompt tuning have been proposed. However, the existing prompt tuning methods may lack generalization. We propose SPT, a semi-parametric prompt tuning method for multitask prompted learning. The novel component of SPT is a memory bank from where memory prompts are retrieved based on discrete prompts. Extensive experiments, such as (i) fine-tuning a full language model with SPT on 31 different tasks from 8 different domains and evaluating zero-shot generalization on 9 heldout datasets under 5 NLP task categories and (ii) pretraining SPT on the GLUE datasets and evaluating fine-tuning on the SuperGLUE datasets, demonstrate effectiveness of SPT.","{'model': 'tldr@v2.0.0', 'text': 'SPT, a semi-parametric prompt tuning method for multitask prompted learning, which is a memory bank from where memory prompts are retrieved based on discrete prompts, is proposed.'}",http://arxiv.org/pdf/2212.10929
-cup: curriculum learning based prompt tuning for implicit event argument extraction,Jiaju Lin,"Implicit event argument extraction (EAE) aims to identify arguments that could scatter over the document. Most previous work focuses on learning the direct relations between arguments and the given trigger, while the implicit relations with long-range dependency are not well studied. Moreover, recent neural network based approaches rely on a large amount of labeled data for training, which is unavailable due to the high labelling cost. In this paper, we propose a Curriculum learning based Prompt tuning (CUP) approach, which resolves implicit EAE by four learning stages. The stages are defined according to the relations with the trigger node in a semantic graph, which well captures the long-range dependency between arguments and the trigger. In addition, we integrate a prompt-based encoder-decoder model to elicit related knowledge from pre-trained language models (PLMs) in each stage, where the prompt templates are adapted with the learning progress to enhance the reasoning for arguments. Experimental results on two well-known benchmark datasets show the great advantages of our proposed approach. In particular, we outperform the state-of-the-art models in both fully-supervised and low-data scenarios.","{'model': 'tldr@v2.0.0', 'text': 'A Curriculum learning based Prompt tuning (CUP) approach, which resolves implicit EAE by four learning stages, which outperform the state-of-the-art models in both fully-supervised and low-data scenarios.'}",https://arxiv.org/pdf/2205.00498
-zero-label prompt selection,Chonghua Liao,"Natural language prompts have been shown to facilitate cross-task generalization for large language models. However, with no or limited labeled examples, the cross-task performance is highly sensitive to the choice of prompts, while selecting a high-performing prompt is challenging given the scarcity of labels. To address the issue, we propose a Zero-Label Prompt Selection (ZPS) method that selects prompts without any labeled data or gradient update. Specifically, given the candidate human-written prompts for a task, ZPS labels a set of unlabeled data with a prompt ensemble and uses the pseudo-labels for prompt selection. Experiments show that ZPS improves over prior methods by a sizeable margin in zero-label performance. We also extend ZPS to a few-shot setting and show its advantages over strong baselines such as prompt tuning and model tuning.","{'model': 'tldr@v2.0.0', 'text': 'A Zero-Label Prompt Selection (ZPS) method that selects prompts without any labeled data or gradient update is proposed that improves over prior methods by a sizeable margin in zero-label performance.'}",http://arxiv.org/pdf/2211.04668
-clip-tuning: towards derivative-free prompt learning with a mixture of rewards,Yekun Chai,"Derivative-free prompt learning has emerged as a lightweight alternative to prompt tuning, which only requires model inference to optimize the prompts. However, existing work did not take full advantage of the over-parameterized characteristics of large pre-trained language models (PLMs). In this paper, we propose Clip-Tuning, a simple yet effective method that adopts diverse frozen""thinned""networks of PLMs to obtain a mixture of rewards and thus advance the derivative-free prompt learning. The thinned networks consist of all the hidden units that survive a stationary dropout strategy, whose inference predictions reflect an ensemble of partial views over prompted training samples. Our method outperforms previous gradient-free prompt learning methods and achieves parity with gradient-based counterparts on seven language understanding benchmarks under few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Clip-Tuning is proposed, a simple yet effective method that adopts diverse frozen""thinned""networks of PLMs to obtain a mixture of rewards and thus advance the derivative-free prompt learning.'}",http://arxiv.org/pdf/2210.12050
-denseclip: language-guided dense prediction with context-aware prompting,Yongming Rao,"Recent progress has shown that large-scale pre-training using contrastive image-text pairs can be a promising alternative for high-quality visual representation learning from natural language supervision. Benefiting from a broader source of supervision, this new paradigm exhibits impressive transferability to downstream classification tasks and datasets. However, the problem of transferring the knowledge learned from image-text pairs to more complex dense prediction tasks has barely been visited. In this work, we present a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP. Specifically, we convert the original image-text matching problem in CLIP to a pixel-text matching problem and use the pixel-text score maps to guide the learning of dense prediction models. By further using the contextual information from the image to prompt the language model, we are able to facilitate our model to better exploit the pretrained knowledge. Our method is model-agnostic, which can be applied to arbitrary dense prediction systems and various pre-trained visual backbones including both CLIP models and ImageNet pre-trained models. Extensive experiments demonstrate the superior performance of our methods on semantic segmentation, object detection, and instance segmentation tasks. Code is available at https://github.com/raoyongming/DenseCLIP.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a new framework for dense prediction by implicitly and explicitly leveraging the pre-trained knowledge from CLIP and converts the original image-text matching problem in CLIP to a pixel- text matching problem and uses the pixel-text score maps to guide the learning of dense prediction models.'}",https://arxiv.org/pdf/2112.01518
-uom&mmu at tsar-2022 shared task: prompt learning for lexical simplification,Laura Vásquez-Rodríguez,"We present PromptLS, a method for fine-tuning large pre-trained Language Models (LM) to perform the task of Lexical Simplification. We use a predefined template to attain appropriate replacements for a term, and fine-tune a LM using this template on language specific datasets. We filter candidate lists in post-processing to improve accuracy. We demonstrate that our model can work in a) a zero shot setting (where we only require a pre-trained LM), b) a fine-tuned setting (where language-specific data is required), and c) a multilingual setting (where the model is pre-trained across multiple languages and fine-tuned in an specific language). Experimental results show that, although the zero-shot setting is competitive, its performance is still far from the fine-tuned setting. Also, the multilingual is unsurprisingly worse than the fine-tuned model. Among all TSAR-2022 Shared Task participants, our team was ranked second in Spanish and third in English.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2022.tsar-1.23.pdf
-bidirectional language models are also few-shot learners,Ajay Patel,"Large language models such as GPT-3 (Brown et al., 2020) can perform arbitrary tasks without undergoing fine-tuning after being prompted with only a few labeled examples. An arbitrary task can be reformulated as a natural language prompt, and a language model can be asked to generate the completion, indirectly performing the task in a paradigm known as prompt-based learning. To date, emergent prompt-based learning capabilities have mainly been demonstrated for unidirectional language models. However, bidirectional language models pre-trained on denoising objectives such as masked language modeling produce stronger learned representations for transfer learning. This motivates the possibility of prompting bidirectional models, but their pre-training objectives have made them largely incompatible with the existing prompting paradigm. We present SAP (Sequential Autoregressive Prompting), a technique that enables the prompting of bidirectional models. Utilizing the machine translation task as a case study, we prompt the bidirectional mT5 model (Xue et al., 2021) with SAP and demonstrate its few-shot and zero-shot translations outperform the few-shot translations of unidirectional models like GPT-3 and XGLM (Lin et al., 2021), despite mT5's approximately 50% fewer parameters. We further show SAP is effective on question answering and summarization. For the first time, our results demonstrate prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models.","{'model': 'tldr@v2.0.0', 'text': 'For the first time, prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models, and is shown to be effective on question answering and summarization.'}",http://arxiv.org/pdf/2209.14500
-speechprompt v2: prompt tuning for speech classification tasks,Kai-Wei Chang,"Prompt tuning is a technology that tunes a small set of parameters to steer a pre-trained language model (LM) to directly generate the output for downstream tasks. Recently, prompt tuning has demonstrated its storage and computation efficiency in both natural language processing (NLP) and speech processing fields. These advantages have also revealed prompt tuning as a candidate approach to serving pre-trained LM for multiple tasks in a unified manner. For speech processing, SpeechPrompt shows its high parameter efficiency and competitive performance on a few speech classification tasks. However, whether SpeechPrompt is capable of serving a large number of tasks is unanswered. In this work, we propose SpeechPrompt v2, a prompt tuning framework capable of performing a wide variety of speech classification tasks, covering multiple languages and prosody-related tasks. The experiment result shows that SpeechPrompt v2 achieves performance on par with prior works with less than 0.15M trainable parameters in a unified framework.","{'model': 'tldr@v2.0.0', 'text': 'SpeechPrompt v2 is proposed, a prompt tuning framework capable of performing a wide variety of speech classification tasks, covering multiple languages and prosody-related tasks, and achieves performance on par with prior works with less than 0.15M trainable parameters in a unified framework.'}",http://arxiv.org/pdf/2303.00733
-language models in the loop: incorporating prompting into weak supervision,Ryan Smith,"We propose a new strategy for applying large pre-trained language models to novel tasks when labeled training data is limited. Rather than apply the model in a typical zero-shot or few-shot fashion, we treat the model as the basis for labeling functions in a weak supervision framework. To create a classifier, we first prompt the model to answer multiple distinct queries about an example and define how the possible responses should be mapped to votes for labels and abstentions. We then denoise these noisy label sources using the Snorkel system and train an end classifier with the resulting training data. Our experimental evaluation shows that prompting large language models within a weak supervision framework can provide significant gains in accuracy. On the WRENCH weak supervision benchmark, this approach can significantly improve over zero-shot performance, an average 19.5% reduction in errors. We also find that this approach produces classifiers with comparable or superior accuracy to those trained from hand-engineered rules.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new strategy for applying large pre-trained language models to novel tasks when labeled training data is limited and treats the model as the basis for labeling functions in a weak supervision framework, and produces classifiers with comparable or superior accuracy to those trained from hand-engineered rules.'}",http://arxiv.org/pdf/2205.02318
-prompting as probing: using language models for knowledge base construction,Dimitrios Alivanistos,"Language Models (LMs) have proven to be useful in various downstream applications, such as summarisation, translation, question answering and text classification. LMs are becoming increasingly important tools in Artificial Intelligence, because of the vast quantity of information they can store. In this work, we present ProP (Prompting as Probing), which utilizes GPT-3, a large Language Model originally proposed by OpenAI in 2020, to perform the task of Knowledge Base Construction (KBC). ProP implements a multi-step approach that combines a variety of prompting techniques to achieve this. Our results show that manual prompt curation is essential, that the LM must be encouraged to give answer sets of variable lengths, in particular including empty answer sets, that true/false questions are a useful device to increase precision on suggestions generated by the LM, that the size of the LM is a crucial factor, and that a dictionary of entity aliases improves the LM score. Our evaluation study indicates that these proposed techniques can substantially enhance the quality of the final predictions: ProP won track 2 of the LM-KBC competition, outperforming the baseline by 36.4 percentage points. Our implementation is available on https://github.com/HEmile/iswc-challenge.","{'model': 'tldr@v2.0.0', 'text': 'ProP (Prompting as Probing), which utilizes GPT-3, a large Language Model originally proposed by OpenAI in 2020, to perform the task of Knowledge Base Construction (KBC), implements a multi-step approach that combines a variety of prompting techniques to achieve this.'}",http://arxiv.org/pdf/2208.11057
-understanding zero-shot adversarial robustness for large-scale models,Chengzhi Mao,"Pretrained large-scale vision-language models like CLIP have exhibited strong generalization over unseen tasks. Yet imperceptible adversarial perturbations can significantly reduce CLIP's performance on new tasks. In this work, we identify and explore the problem of \emph{adapting large-scale models for zero-shot adversarial robustness}. We first identify two key factors during model adaption -- training losses and adaptation methods -- that affect the model's zero-shot adversarial robustness. We then propose a text-guided contrastive adversarial training loss, which aligns the text embeddings and the adversarial visual features with contrastive learning on a small set of training data. We apply this training loss to two adaption methods, model finetuning and visual prompt tuning. We find that visual prompt tuning is more effective in the absence of texts, while finetuning wins in the existence of text guidance. Overall, our approach significantly improves the zero-shot adversarial robustness over CLIP, seeing an average improvement of over 31 points over ImageNet and 15 zero-shot datasets. We hope this work can shed light on understanding the zero-shot adversarial robustness of large-scale models.","{'model': 'tldr@v2.0.0', 'text': ""This work identifies two key factors during model adaption -- training losses and adaptation methods -- that affect the model's zero-shot adversarial robustness, and proposes a text-guided contrastive adversarial training loss, which aligns the text embeddings and the adversarial visual features with contrastive learning on a small set of training data.""}",http://arxiv.org/pdf/2212.07016
-what does clip know about a red circle? visual prompt engineering for vlms,Aleksandar Shtedritski,"Large-scale Vision-Language Models, such as CLIP, learn powerful image-text representations that have found numerous applications, from zero-shot classification to text-to-image generation. Despite that, their capabilities for solving novel discriminative tasks via prompting fall behind those of large language models, such as GPT-3. Here we explore the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text. In particular, we discover an emergent ability of CLIP, where, by simply drawing a red circle around an object, we can direct the model's attention to that region, while also maintaining global information. We show the power of this simple approach by achieving state-of-the-art in zero-shot referring expressions comprehension and strong performance in keypoint localization tasks. Finally, we draw attention to some potential ethical concerns of large language-vision models.","{'model': 'tldr@v2.0.0', 'text': ""This work explores the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text, and discovers an emergent ability of CLIP, where, by simply drawing a red circle around an object, it can direct the model's attention to that region, while also maintaining global information.""}",https://arxiv.org/pdf/2304.06712
-prompting visual-language models for efficient video understanding,Chen Ju,,"{'model': 'tldr@v2.0.0', 'text': 'A simple but strong baseline is presented to efficiently adapt the pre-trained I-VL model, and exploit its powerful ability for resource-hungry video understanding tasks, with minimal training, to optimise a few random vectors that convert video-related tasks into the same format as thePre-training objectives.'}",https://arxiv.org/pdf/2112.04478
-healthprompt: a zero-shot learning paradigm for clinical natural language processing,S. Sivarajkumar,"Developing clinical natural language systems based on machine learning and deep learning is dependent on the availability of large-scale annotated clinical text datasets, most of which are time-consuming to create and not publicly available. The lack of such annotated datasets is the biggest bottleneck for the development of clinical NLP systems. Zero-Shot Learning (ZSL) refers to the use of deep learning models to classify instances from new classes of which no training data have been seen before. Prompt-based learning is an emerging ZSL technique in NLP where we define task-based templates for different tasks. In this study, we developed a novel prompt-based clinical NLP framework called HealthPrompt and applied the paradigm of prompt-based learning on clinical texts. In this technique, rather than fine-tuning a Pre-trained Language Model (PLM), the task definitions are tuned by defining a prompt template. We performed an in-depth analysis of HealthPrompt on six different PLMs in a no-training-data setting. Our experiments show that HealthPrompt could effectively capture the context of clinical texts and perform well for clinical NLP tasks without any training data.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt- based clinical NLP framework called HealthPrompt is developed and applied the paradigm of prompt-based learning on clinical texts and shows that it could effectively capture the context ofclinical texts and perform well forclinical NLP tasks without any training data.'}",
-vl-adapter: parameter-efficient transfer learning for vision-and-language tasks,Yi-Lin Sung,"Recently, fine-tuning language models pre-trained on large text corpora have provided huge improvements on vision-and-language (V&L) tasks as well as on pure language tasks. However, fine-tuning the entire parameter set of pre-trained models becomes impractical since the model size is growing rapidly. Hence, in this paper, we introduce adapter-based parameter-efficient transfer learning techniques to V&L models such as VL-BART and VL-T5. We evaluate our methods in a unified multi-task setup on both image-text and video-text benchmarks. For the image-text tasks, we use four diverse V&L datasets: VQAv2, GQA, NLVR2, and MSCOCO image captioning. For video-text tasks, we use TVQA, How2QA, TVC, and YC2C. With careful training and thorough experiments, we benchmark three popular adapter-based methods (Adapter, Hyperformer, Compacter) against the standard full fine-tuning and the recently proposed prompt-tuning approach. We also enhance the efficiency and performance of adapters by sharing their weights to attain knowledge across tasks. Our results demonstrate that training the adapter with the weight-sharing technique (4.18% of total parameters for image-text tasks and 3.39% for video-text tasks) can match the performance of fine-tuning the entire model. Lastly, we present a comprehensive analysis including the combination of adapter and task-specific prompts and the impact of V&L pre-training on adapters. 11The code for our CVPR 2022 paper is available at: https://github.com/ylsung/VL_adapter.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that training the adapter with the weight-sharing technique can match the performance of fine-tuning the entire model, and enhance the efficiency and performance of adapters by sharing their weights to attain knowledge across tasks.'}",https://arxiv.org/pdf/2112.06825
-an automatically discovered chain-of-thought prompt generalizes to novel models and datasets,Konstantin Hebenstreit,"Emergent chain-of-thought (CoT) reasoning capabilities promise to improve performance and explainability of large language models (LLMs). However, uncertainties remain about how reasoning strategies formulated for previous model generations generalize to new model generations and different datasets. In this small-scale study, we compare different reasoning strategies induced by zero-shot prompting across six recently released LLMs (davinci-002, davinci-003, GPT-3.5-turbo, GPT-4, Flan-T5-xxl and Cohere command-xlarge) on a mixture of six question-answering datasets, including datasets from scientific and medical domains. Our findings demonstrate that while some variations in effectiveness occur, gains from CoT reasoning strategies remain robust across different models and datasets. GPT-4 has the most benefit from current state-of-the-art reasoning strategies and exhibits the best performance by applying a prompt previously discovered through automated discovery.","{'model': 'tldr@v2.0.0', 'text': 'GPT-4 has the most benefit from current state-of-the-art reasoning strategies and exhibits the best performance by applying a prompt previously discovered through automated discovery.'}",https://arxiv.org/pdf/2305.02897
-pbnr: prompt-based news recommender system,Xinyi Li,"Online news platforms often use personalized news recommendation methods to help users discover articles that align with their interests. These methods typically predict a matching score between a user and a candidate article to reflect the user's preference for the article. Some previous works have used language model techniques, such as the attention mechanism, to capture users' interests based on their past behaviors, and to understand the content of articles. However, these existing model architectures require adjustments if additional information is taken into account. Pre-trained large language models, which can better capture word relationships and comprehend contexts, have seen a significant development in recent years, and these pre-trained models have the advantages of transfer learning and reducing the training time for downstream tasks. Meanwhile, prompt learning is a newly developed technique that leverages pre-trained language models by building task-specific guidance for output generations. To leverage textual information in news articles, this paper introduces the pre-trained large language model and prompt-learning to the community of news recommendation. The proposed model""prompt-based news recommendation""(PBNR) treats the personalized news recommendation as a text-to-text language task and designs personalized prompts to adapt to the pre-trained language model -- text-to-text transfer transformer (T5). Experimental studies using the Microsoft News dataset show that PBNR is capable of making accurate recommendations by taking into account various lengths of past behaviors of different users. PBNR can also easily adapt to new information without changing the model architecture and the training objective. Additionally, PBNR can make recommendations based on users' specific requirements, allowing human-computer interaction in the news recommendation field.","{'model': 'tldr@v2.0.0', 'text': 'The proposed model ""prompt-based news recommendation""(PBNR) treats the personalized news recommendation as a text-to-text language task and designs personalized prompts to adapt to the pre-trained language model -- text- to-text transfer transformer (T5).'}",http://arxiv.org/pdf/2304.07862
-p rompt c ap : prompt-guided image captioning for vqa with gpt-3,Yushi Hu,"Knowledge-based visual question answering (VQA) in-volves questions that require world knowledge beyond the image to yield the correct answer. Large language models (LMs) like GPT-3 are particularly helpful for this task because of their strong knowledge retrieval and reasoning capabilities. To enable LM to understand images, prior work uses a captioning model to convert images into text. However, when summarizing an image in a single caption sentence, which visual entities to describe are often underspecified. Generic image captions often miss visual details essential for the LM to answer visual questions correctly. To address this challenge, we propose P ROMPT C AP ( Prompt -guided image Cap tioning), a captioning model designed to serve as a better connector between images and black-box LMs. Different from generic captions, P ROMPT C AP takes a natural-language prompt to control the visual entities to describe in the generated caption. The prompt contains a question that the caption should aid in answering. To avoid extra annotation, P ROMPT C AP is trained by examples synthesized with GPT-3 and existing datasets. We demonstrate P ROMPT - C AP ’s effectiveness on an existing pipeline in which GPT-3 is prompted with image captions to carry out VQA. P ROMPT - C AP outperforms generic captions by a large margin and achieves state-of-the-art accuracy on knowledge-based VQA tasks (60.4% on OK-VQA and 59.6% on A-OKVQA). Zero-shot results on WebQA show that P ROMPT C AP generalizes well to unseen domains. 1","{'model': 'tldr@v2.0.0', 'text': 'P ROMPT C AP is a captioning model designed to serve as a better connector between images and black-box LMs that outperforms generic captions by a large margin and achieves state-of-the-art accuracy on knowledge-based VQA tasks.'}",
-visual clues: bridging vision and language foundations for image paragraph captioning,Yujia Xie,"People say,""A picture is worth a thousand words"". Then how can we get the rich information out of the image? We argue that by using visual clues to bridge large pretrained vision foundation models and language models, we can do so without any extra cross-modal training. Thanks to the strong zero-shot capability of foundation models, we start by constructing a rich semantic representation of the image (e.g., image tags, object attributes / locations, captions) as a structured textual prompt, called visual clues, using a vision foundation model. Based on visual clues, we use large language model to produce a series of comprehensive descriptions for the visual content, which is then verified by the vision model again to select the candidate that aligns best with the image. We evaluate the quality of generated descriptions by quantitative and qualitative measurement. The results demonstrate the effectiveness of such a structured semantic representation.","{'model': 'tldr@v2.0.0', 'text': 'This work argues that by using visual clues to bridge large pretrained vision foundation models and language models, they can do so without any extra cross-modal training.'}",http://arxiv.org/pdf/2206.01843
-response generation with context-aware prompt learning,X. Gu,"Pre-trained language models (PLM) have marked a huge leap in neural dialogue modeling. While PLMs are pre-trained on large-scale text corpora, they are usually fine-tuned on scarce dialogue data with specific domain knowledge and dialogue styles. However, tailoring the language models while fully utilizing prior knowledge in large pre-trained models remains a challenge. In this paper, we present a novel approach for pre-trained dialogue modeling that casts the dialogue generation problem as a prompt-learning task. Instead of fine-tuning on limited dialogue data, our approach, DialogPrompt, learns continuous prompt embeddings optimized for dialogue contexts, which appropriately elicit knowledge from the large pre-trained model. To encourage the model to better utilize the prompt embeddings, the prompt encoders are designed to be dynamically generated based on the dialogue context. Experiments on popular conversation datasets show that our approach significantly outperforms the fine-tuning baseline and the generic prompt-learning methods. Furthermore, human evaluations strongly support the superiority of DialogPrompt in regard to response generation quality.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel approach for pre-trained dialogue modeling that casts the dialogue generation problem as a prompt-learning task and learns continuous prompt embeddings optimized for dialogue contexts, which appropriately elicit knowledge from the large pre- trained model.'}",
-few-shot self-rationalization with natural language prompts,Ana Marasović,"Self-rationalization models that predict task labels and generate free-text elaborations for their predictions could enable more intuitive interaction with NLP systems. These models are, however, currently trained with a large amount of human-written free-text explanations for each task which hinders their broader usage. We propose to study a more realistic setting of self-rationalization using few training examples. We present FEB -- a standardized collection of four existing English-language datasets and associated metrics. We identify the right prompting approach by extensively exploring natural language prompts on FEB. Then, by using this prompt and scaling the model size, we demonstrate that making progress on few-shot self-rationalization is possible. We show there is still ample room for improvement in this task: the average plausibility of generated explanations assessed by human annotators is at most 51% (with GPT-3), while plausibility of human explanations is 76%. We hope that FEB and our proposed approach will spur the community to take on the few-shot self-rationalization challenge.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies the right prompting approach by extensively exploring natural language prompts on FEB and demonstrates that making progress on few-shot self-rationalization is possible, and presents FEB -- a standardized collection of four existing English-language datasets and associated metrics.'}",https://aclanthology.org/2022.findings-naacl.31.pdf
-controllable generation from pre-trained language models via inverse prompting,Xu Zou,"Large-scale pre-trained language models have demonstrated strong capabilities of generating realistic texts. However, it remains challenging to control the generation results. Previous approaches such as prompting are far from sufficient, and lack of controllability limits the usage of language models. To tackle this challenge, we propose an innovative method, inverse prompting, to better control text generation. The core idea of inverse prompting is to use generated text to inversely predict the prompt during beam search, which enhances the relevance between the prompt and the generated text and thus improves controllability. Empirically, we pre-train a large-scale Chinese language model to perform a systematic study using human evaluation on the tasks of open-domain poem generation and open-domain long-form question answering. Results demonstrate that our proposed method substantially outperforms the baselines and that our generation quality is close to human performance on some of the tasks.","{'model': 'tldr@v2.0.0', 'text': 'The core idea of inverse prompting is to use generated text to inversely predict the prompt during beam search, which enhances the relevance between the prompt and the generated text and thus improves controllability.'}",https://arxiv.org/pdf/2103.10685
-progressive prompts: continual learning for language models,Anastasia Razdaibiedina,"We introduce Progressive Prompts - a simple and efficient approach for continual learning in language models. Our method allows forward transfer and resists catastrophic forgetting, without relying on data replay or a large number of task-specific parameters. Progressive Prompts learns a new soft prompt for each task and sequentially concatenates it with the previously learned prompts, while keeping the base model frozen. Experiments on standard continual learning benchmarks show that our approach outperforms state-of-the-art methods, with an improvement>20% in average test accuracy over the previous best-preforming method on T5 model. We also explore a more challenging continual learning setup with longer sequences of tasks and show that Progressive Prompts significantly outperforms prior methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Progressive Prompts - a simple and efficient approach for continual learning in language models that allows forward transfer and resists catastrophic forgetting, without relying on data replay or a large number of task-specific parameters.'}",http://arxiv.org/pdf/2301.12314
-motiongpt: human motion as a foreign language,Biao Jiang,"Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this""motion vocabulary"", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks that achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.'}",https://arxiv.org/pdf/2306.14795
-boosting natural language generation from instructions with meta-learning,Budhaditya Deb,"Recent work has shown that language models (LMs) trained with multi-task instructional learning (MTIL) can solve diverse NLP tasks in zero- and few-shot settings with improved performance compared to prompt tuning. MTIL illustrates that LMs can extract and use information about the task from instructions beyond the surface patterns of the inputs and outputs. This suggests that meta-learning may further enhance the utilization of instructions for effective task transfer. In this paper we investigate whether meta-learning applied to MTIL can further improve generalization to unseen tasks in a zero-shot setting. Specifically, we propose to adapt meta-learning to MTIL in three directions: 1) Model Agnostic Meta Learning (MAML), 2) Hyper-Network (HNet) based adaptation to generate task specific parameters conditioned on instructions, and 3) an approach combining HNet and MAML. Through extensive experiments on the large scale Natural Instructions V2 dataset, we show that our proposed approaches significantly improve over strong baselines in zero-shot settings. In particular, meta-learning improves the effectiveness of instructions and is most impactful when the test tasks are strictly zero-shot (i.e. no similar tasks in the training set) and are “hard” for LMs, illustrating the potential of meta-learning for MTIL for out-of-distribution tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to adapt meta-learning to MTIL in three directions: Model Agnostic Meta Learning (MAML), Hyper-Network (HNet) based adaptation to generate task specific parameters conditioned on instructions, and an approach combining HNet and MAML.'}",http://arxiv.org/pdf/2210.11617
-strategic reasoning with language models,Kanishk Gandhi,"Strategic reasoning enables agents to cooperate, communicate, and compete with other agents in diverse situations. Existing approaches to solving strategic games rely on extensive training, yielding strategies that do not generalize to new scenarios or games without retraining. Large Language Models (LLMs), with their ability to comprehend and generate complex, context-rich language, could prove powerful as tools for strategic gameplay. This paper introduces an approach that uses pretrained LLMs with few-shot chain-of-thought examples to enable strategic reasoning for AI agents. Our approach uses systematically generated demonstrations of reasoning about states, values, and beliefs to prompt the model. Using extensive variations of simple matrix games, we show that strategies that are derived based on systematically generated prompts generalize almost perfectly to new game structures, alternate objectives, and hidden information. Additionally, we demonstrate our approach can lead to human-like negotiation strategies in realistic scenarios without any extra training or fine-tuning. Our results highlight the ability of LLMs, guided by systematic reasoning demonstrations, to adapt and excel in diverse strategic scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an approach that uses pretrained LLMs with few-shot chain-of-thought examples to enable strategic reasoning for AI agents and shows that strategies that are derived based on systematically generated prompts generalize almost perfectly to new game structures, alternate objectives, and hidden information.'}",http://arxiv.org/pdf/2305.19165
-respectful or toxic? using zero-shot learning with language models to detect hate speech,F. Plaza-Del-Arco,"Hate speech detection faces two significant challenges: 1) the limited availability of labeled data and 2) the high variability of hate speech across different contexts and languages. Prompting brings a ray of hope to these challenges. It allows injecting a model with task-specific knowledge without relying on labeled data. This paper explores zero-shot learning with prompting for hate speech detection. We investigate how well zero-shot learning can detect hate speech in 3 languages with limited labeled data. We experiment with various large language models and verbalizers on 8 benchmark datasets. Our findings highlight the impact of prompt selection on the results. They also suggest that prompting, specifically with recent large language models, can achieve performance comparable to and surpass fine-tuned models, making it a promising alternative for under-resourced languages. Our findings highlight the potential of prompting for hate speech detection and show how both the prompt and the model have a significant impact on achieving more accurate predictions in this task.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates how well zero-shot learning with prompting for hate speech detection in 3 languages with limited labeled data and suggests that prompting can achieve performance comparable to and surpass fine-tuned models, making it a promising alternative for under-resourced languages.'}",https://aclanthology.org/2023.woah-1.6.pdf
-meta-tuning language models to answer prompts better,Ruiqi Zhong,"Large pretrained language models like GPT-3 have acquired a surprising ability to perform zero-shot classification (ZSC). For example, to classify review sentiments, we can “prompt"" the language model with the review and the question “ Is the review positive? "" as the context, and ask it to predict whether the next word is “ Yes "" or “ No "". However, these models are not specialized for answering these prompts. To address this weak-ness, we propose meta-tuning, which trains the model to specialize in answering prompts but still generalize to unseen tasks. To create the training data, we aggregated 43 existing datasets, annotated 441 label descriptions in total, and unified them into the above question answering (QA) format. After meta-tuning, our model outperforms a same-sized QA model for most labels on unseen tasks, and we forecast that the performance would im-prove for even larger models. Therefore, measuring ZSC performance on non-specialized language models might underestimate their true capability, and community-wide efforts on aggregating datasets and unifying their formats can help build models that understand prompts better.","{'model': 'tldr@v2.0.0', 'text': None}",
-zero-shot translation of attention patterns in vqa models to natural language,Leonard Salewski,"Converting a model's internals to text can yield human-understandable insights about the model. Inspired by the recent success of training-free approaches for image captioning, we propose ZS-A2T, a zero-shot framework that translates the transformer attention of a given model into natural language without requiring any training. We consider this in the context of Visual Question Answering (VQA). ZS-A2T builds on a pre-trained large language model (LLM), which receives a task prompt, question, and predicted answer, as inputs. The LLM is guided to select tokens which describe the regions in the input image that the VQA model attended to. Crucially, we determine this similarity by exploiting the text-image matching capabilities of the underlying VQA model. Our framework does not require any training and allows the drop-in replacement of different guiding sources (e.g. attribution instead of attention maps), or language models. We evaluate this novel task on textual explanation datasets for VQA, giving state-of-the-art performances for the zero-shot setting on GQA-REX and VQA-X. Our code is available at: https://github.com/ExplainableML/ZS-A2T.","{'model': 'tldr@v2.0.0', 'text': 'ZS-A2T is proposed, a zero-shot framework that translates the transformer attention of a given model into natural language without requiring any training, and allows the drop-in replacement of different guiding sources (e.g. attribution instead of attention maps), or language models.'}",
-"a sign language recognition system with pepper, lightweight-transformer, and llm",Jongyoon Lim,"This research explores using lightweight deep neural network architectures to enable the humanoid robot Pepper to understand American Sign Language (ASL) and facilitate non-verbal human-robot interaction. First, we introduce a lightweight and efficient model for ASL understanding optimized for embedded systems, ensuring rapid sign recognition while conserving computational resources. Building upon this, we employ large language models (LLMs) for intelligent robot interactions. Through intricate prompt engineering, we tailor interactions to allow the Pepper Robot to generate natural Co-Speech Gesture responses, laying the foundation for more organic and intuitive humanoid-robot dialogues. Finally, we present an integrated software pipeline, embodying advancements in a socially aware AI interaction model. Leveraging the Pepper Robot's capabilities, we demonstrate the practicality and effectiveness of our approach in real-world scenarios. The results highlight a profound potential for enhancing human-robot interaction through non-verbal interactions, bridging communication gaps, and making technology more accessible and understandable.","{'model': 'tldr@v2.0.0', 'text': 'This research explores using lightweight deep neural network architectures to enable the humanoid robot Pepper to understand American Sign Language (ASL) and facilitate non-verbal human-robot interaction and presents an integrated software pipeline, embodying advancements in a socially aware AI interaction model.'}",https://arxiv.org/pdf/2309.16898
-reducing sentiment bias in language models via counterfactual evaluation,Po-Sen Huang,"Advances in language modeling architectures and the availability of large text corpora have driven progress in automatic text generation. While this results in models capable of generating coherent texts, it also prompts models to internalize social biases present in the training corpus. This paper aims to quantify and reduce a particular type of bias exhibited by language models: bias in the sentiment of generated text. Given a conditioning context (e.g., a writing prompt) and a language model, we analyze if (and how) the sentiment of the generated text is affected by changes in values of sensitive attributes (e.g., country names, occupations, genders) in the conditioning context using a form of counterfactual evaluation. We quantify sentiment bias by adopting individual and group fairness metrics from the fair machine learning literature, and demonstrate that large-scale models trained on two different corpora (news articles, and Wikipedia) exhibit considerable levels of bias. We then propose embedding and sentiment prediction-derived regularization on the language model’s latent representations. The regularizations improve fairness metrics while retaining comparable levels of perplexity and semantic similarity.","{'model': 'tldr@v2.0.0', 'text': 'This paper quantifies sentiment bias by adopting individual and group fairness metrics from the fair machine learning literature, and proposes embedding and sentiment prediction-derived regularization on the language model’s latent representations.'}",https://www.aclweb.org/anthology/2020.findings-emnlp.7.pdf
-question decomposition improves the faithfulness of model-generated reasoning,Ansh Radhakrishnan,"As large language models (LLMs) perform more difficult tasks, it becomes harder to verify the correctness and safety of their behavior. One approach to help with this issue is to prompt LLMs to externalize their reasoning, e.g., by having them generate step-by-step reasoning as they answer a question (Chain-of-Thought; CoT). The reasoning may enable us to check the process that models use to perform tasks. However, this approach relies on the stated reasoning faithfully reflecting the model's actual reasoning, which is not always the case. To improve over the faithfulness of CoT reasoning, we have models generate reasoning by decomposing questions into subquestions. Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT while improving the faithfulness of the model's stated reasoning on several recently-proposed metrics. By forcing the model to answer simpler subquestions in separate contexts, we greatly increase the faithfulness of model-generated reasoning over CoT, while still achieving some of the performance gains of CoT. Our results show it is possible to improve the faithfulness of model-generated reasoning; continued improvements may lead to reasoning that enables us to verify the correctness and safety of LLM behavior.","{'model': 'tldr@v2.0.0', 'text': ""Decomposition-based methods achieve strong performance on question-answering tasks, sometimes approaching that of CoT while improving the faithfulness of the model's stated reasoning on several recently-proposed metrics.""}",https://arxiv.org/pdf/2307.11768
-improving gender fairness of pre-trained language models without catastrophic forgetting,Zahra Fatemi,"Existing studies addressing gender bias of pre-trained language models, usually build a small gender-neutral data set and conduct a second phase pre-training on the model with such data. However, given the limited size and concentrated focus of the gender-neutral data, catastrophic forgetting would occur during second-phase pre-training. Forgetting information in the original training data may damage the model’s downstream performance by a large margin. In this work, we empirically show that catastrophic forgetting occurs in such methods by evaluating them with general NLP tasks in GLUE. Then, we propose a new method, GEnder Equality Prompt (GEEP), to improve gender fairness of pre-trained models with less forgetting. GEEP freezes the pre-trained model and learns gender-related prompts with gender-neutral data.Empirical results show that GEEP not only achieves SOTA performances on gender fairness tasks, but also forgets less and performs better on GLUE by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'Empirical results show that GEEP not only achieves SOTA performances on gender fairness tasks, but also forgets less and performs better on GLUE by a large margin.'}",https://aclanthology.org/2023.acl-short.108.pdf
-few-shot instruction prompts for pretrained language models to detect social biases,Shrimai Prabhumoye,"Detecting social bias in text is challenging due to nuance, subjectivity, and difficulty in obtaining good quality labeled datasets at scale, especially given the evolving nature of social biases and society. To address these challenges, we propose a few-shot instruction-based method for prompting pre-trained language models (LMs). We select a few class-balanced exemplars from a small support repository that are closest to the query to be labeled in the embedding space. We then provide the LM with instruction that consists of this subset of labeled exemplars, the query text to be classified, a definition of bias, and prompt it to make a decision. We demonstrate that large LMs used in a few-shot context can detect different types of fine-grained biases with similar and sometimes superior accuracy to fine-tuned models. We observe that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 13% improvement in AUC metric compared to other models). It also maintains a high AUC (dropping less than 2%) when the labeled repository is reduced to as few as $100$ samples. Large pretrained language models thus make it easier and quicker to build new bias detectors.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 13% improvement in AUC metric compared to other models) and makes it easier and quicker to build new bias detectors.'}",
-simultaneous translation and paraphrase for language education,Stephen Mayhew,"We present the task of Simultaneous Translation and Paraphrasing for Language Education (STAPLE). Given a prompt in one language, the goal is to generate a diverse set of correct translations that language learners are likely to produce. This is motivated by the need to create and maintain large, high-quality sets of acceptable translations for exercises in a language-learning application, and synthesizes work spanning machine translation, MT evaluation, automatic paraphrasing, and language education technology. We developed a novel corpus with unique properties for five languages (Hungarian, Japanese, Korean, Portuguese, and Vietnamese), and report on the results of a shared task challenge which attracted 20 teams to solve the task. In our meta-analysis, we focus on three aspects of the resulting systems: external training corpus selection, model architecture and training decisions, and decoding and filtering strategies. We find that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to our provided learner response frequencies.","{'model': 'tldr@v2.0.0', 'text': 'It is found that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to the provided learner response frequencies.'}",https://www.aclweb.org/anthology/2020.ngt-1.28.pdf
-evaluating the instruction-following robustness of large language models to prompt injection,Zekun Li,"Large Language Models (LLMs) have shown remarkable proficiency in following instructions, making them valuable in customer-facing applications. However, their impressive capabilities also raise concerns about the amplification of risks posed by adversarial instructions, which can be injected into the model input by third-party attackers to manipulate LLMs' original instructions and prompt unintended actions and content. Therefore, it is crucial to understand LLMs' ability to accurately discern which instructions to follow to ensure their safe deployment in real-world scenarios. In this paper, we propose a pioneering benchmark for automatically evaluating the robustness of instruction-following LLMs against adversarial instructions injected in the prompt. The objective of this benchmark is to quantify the extent to which LLMs are influenced by injected adversarial instructions and assess their ability to differentiate between these injected adversarial instructions and original user instructions. Through experiments conducted with state-of-the-art instruction-following LLMs, we uncover significant limitations in their robustness against adversarial instruction injection attacks. Furthermore, our findings indicate that prevalent instruction-tuned models are prone to being ``overfitted'' to follow any instruction phrase in the prompt without truly understanding which instructions should be followed. This highlights the need to address the challenge of training models to comprehend prompts instead of merely following instruction phrases and completing the text. The data and code can be found at \url{https://github.com/Leezekun/Adv-Instruct-Eval}.","{'model': 'tldr@v2.0.0', 'text': ""A pioneering benchmark for automatically evaluating the robustness of instruction-following LLMs against adversarial instructions injected in the prompt is proposed and it is indicated that prevalent instruction-tuned models are prone to being ``overfitted'' to follow any instruction phrase in the Prompt without truly understanding which instructions should be followed.""}",
-calculation of prompt loss and toroidal field ripple loss under neutral beam injection on east,Bin Wu,"Neutral beam injection is a major auxiliary heating method in the EAST experimental campaign. This paper gives detailed calculations of beam loss with different plasma equilibria using the guiding center code ORBIT and NUBEAM/TRANSP. Increasing plasma current can dramatically lower the beam ion prompt loss and ripple loss. Countercurrent beam injection gives a much larger prompt loss fraction than co-injection, and ripple-induced collisionless stochastic diffusion is the dominant loss channel.",,
-"prompt injection of relativistic protons from the september 1, 1971 solar flare",E. Cliver,,,
-the role of stochastic acceleration in the prompt emission of gamma-ray bursts: application to hadronic injection,K. Murase,"We study effects of particle re-acceleration (or heating) in the post-shock region via magnetohydrodynamic/plasma turbulence, in the context of a mixed hadronic–leptonic model for the prompt emission of gamma-ray bursts, using both analytical and numerical methods. We show that stochastically accelerated (or heated) leptons, which are injected via pp and pγ reactions and subsequent pair cascades, are plausibly able to reproduce the Band function spectra with α ∼ 1 and β ∼ 2–3 in the ∼MeV range. An additional hard component coming from the proton-induced cascade emission is simultaneously expected, which can be compatible with observed extra power-law spectra far above the MeV range. We also discuss the specific implications of hadronic models for ongoing high-energy neutrino observations.",,https://iopscience.iop.org/article/10.1088/0004-637X/746/2/164/pdf
-(ab)using images and sounds for indirect instruction injection in multi-modal llms,Eugene Bagdasaryan,"We demonstrate how images and sounds can be used for indirect prompt and instruction injection in multi-modal LLMs. An attacker generates an adversarial perturbation corresponding to the prompt and blends it into an image or audio recording. When the user asks the (unmodified, benign) model about the perturbed image or audio, the perturbation steers the model to output the attacker-chosen text and/or make the subsequent dialog follow the attacker's instruction. We illustrate this attack with several proof-of-concept examples targeting LLaVa and PandaGPT.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated how images and sounds can be used for indirect prompt and instruction injection in multi-modal LLMs by an attacker generating an adversarial perturbation corresponding to the prompt and blends it into an image or audio recording.'}",https://arxiv.org/pdf/2307.10490
-regional grafting of autologous adipose tissue is effective in inducing prompt healing of indolent digital ulcers in patients with systemic sclerosis: results of a monocentric randomized controlled study,N. Del Papa,,"{'model': 'tldr@v2.0.0', 'text': 'This RCT strongly confirms that AT-G is effective in inducing ischemic digital ulcer healing in patients with systemic sclerosis patients.'}",https://arthritis-research.biomedcentral.com/track/pdf/10.1186/s13075-018-1792-8
-"limitations of extended gas delivery tubes used for fuelling mitigated plasma disruptions, and a unique injection concept for prompt gas delivery",P. Parks,"Modelling high-pressure gas flows through long extended delivery tubes used for massive particle fuelling of tokamaks during a major disruptive instability or a preemptive fast plasma shutdown is presented using analytical theory and simulation. For steady-state flows, expressions were derived and compared with experiments for the transition diameter and inlet Mach number of a straight tube (pipe) attached to a ‘nozzle-like’ inlet valve, such that increases in pipe diameter have no effect on the flow rate (valve-limited flow), and decreases below the transition diameter cause decreasing flow rates (friction-limited flow). Analytical expressions for the exit outflow rate and other gas dynamic variables during the initial unsteady gas flow buildup were developed from the classical 1D centred expansion wave problem and compared with 2D axisymmetric FLUENT simulations with wall friction, and good agreement was found for sufficiently high-conductance pipes. The intrinsic time delay before steady-state outflow is reached can seriously limit plasma density increases during the disruption, as the disruption time scale is similar to the delay time or ‘rise time’ of the outflow at the exit plane. Thus, conditions required for strong collisional dissipation of destructive runaway electron currents can be compromised. A unique gas injection scheme ‘burst membrane gas injection’ is also presented in which a steady-state outflow at the exit plane can be established promptly once the membrane bursts. It is shown that the theoretically ideal rise time of the gas at the exit plane will be zero if the membrane had an instantaneous opening time, therefore the true rise time will be limited only by the rupture time of a real bursting disc. The duration of the initial steady-state gas delivery phase can be matched to the relevant disruption time scales (or runaway formation time) by simply adjusting the tube length.",,
-review of conformal cooling system design and additive manufacturing for injection molds.,Zhiquan Wei,"This paper points out the significance of cooling in injection molding and briefly reviews the development of cooling systems. The focus of this survey is on the physical model, development, and optimization of conformal cooling systems which have curved cooling circuits following the shape of mold cavity. Compared with traditional cooling systems, conformal cooling can greatly reduce the warpage defect and shorten the cooling cycle time. The computational design methods and additive manufacturing techniques that prompt the development of conformal cooling are deeply investigated. At the end of this survey, the future perspectives for conformal cooling design and manufacturing are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The focus of this survey is on the physical model, development, and optimization of conformal cooling systems which have curved cooling circuits following the shape of mold cavity.'}",
-constraints on millisecond magnetars as the engines of prompt emission in gamma-ray bursts,P. Beniamini,"We examine millisecond magnetars as central engines of Gamma Ray Bursts' (GRB) prompt emission. Using the proto-magnetar wind model of Metzger et al. 2011, we estimate the temporal evolution of the magnetization and power injection at the base of the GRB jet and apply these to different prompt emission models to make predictions for the GRB energetics, spectra and lightcurves. We investigate both shock and magnetic reconnection models for the particle acceleration, as well as the effects of energy dissipation across optically thick and thin regions of the jet. The magnetization at the base of the jet, $\sigma_0$, is the main parameter driving the GRB evolution in the magnetar model and the emission is typically released for $100\lesssim \sigma_0 \lesssim 3000$. Given the rapid increase in $\sigma_0$ as the proto-magnetar cools and its neutrino-driven mass loss subsides, the GRB duration is typically limited to $\lesssim 100$ s. This low baryon loading at late times challenges magnetar models for ultra-long GRBs, though black hole models likely run into similar difficulties without substantial entrainment from the jet walls. The maximum radiated gamma-ray energy is $\lesssim 5 \times 10^{51}$erg, significantly less than the magnetar's total initial rotational energy and in strong tension with the high end of the observed GRB energy distribution. However, the gradual magnetic dissipation model (Beniamini & Giannios 2017) applied to a magnetar central engine, naturally explains several key observables of typical GRBs, including energetics, durations, stable peak energies, spectral slopes and a hard to soft evolution during the burst.",,https://academic.oup.com/mnras/article-pdf/472/3/3058/20131453/stx2095.pdf
-prompt injections of highly relativistic electrons induced by interplanetary shocks: a statistical study of van allen probes observations,Q. Schiller,"We conduct a statistical study on the sudden response of outer radiation belt electrons due to interplanetary (IP) shocks during the Van Allen Probes era, i.e., 2012 to 2015. Data from the Relativistic Electron‐Proton Telescope instrument on board Van Allen Probes are used to investigate the highly relativistic electron response (E > 1.8 MeV) within the first few minutes after shock impact. We investigate the relationship of IP shock parameters, such as Mach number, with the highly relativistic electron response, including spectral properties and radial location of the shock‐induced injection. We find that the driving solar wind structure of the shock does not affect occurrence for enhancement events, 25% of IP shocks are associated with prompt energization, and 14% are associated with MeV electron depletion. Parameters that represent IP shock strength are found to correlate best with highest levels of energization, suggesting that shock strength may play a key role in the severity of the enhancements. However, not every shock results in an enhancement, indicating that magnetospheric preconditioning may be required.",,
-on the synchrotron spectrum of grb prompt emission,Siyao Xu,"The prompt emission spectrum of gamma-ray bursts is characterized by a smoothly joint broken power-law spectrum known as the Band function. The typical low-energy photon index is , which poses a challenge to standard synchrotron radiation models. We investigate the electron energy spectrum as a result of the interplay among adiabatic stochastic acceleration (ASA), particle injection, and synchrotron cooling. In the ASA-dominated low-energy range, ASA enables an efficient hardening of the injected energy spectrum to approach a spectral index −1. In the synchrotron cooling-dominated high-energy range, the injected high-energy electrons undergo fast synchrotron cooling and have a softer photon spectrum. With the energy range of the injected electrons broadly covering both the ASA- and synchrotron cooling-dominated ranges, the resulting photon number spectrum has low- and high-energy indices of and , respectively. The break energy is of the order of ∼100 keV, depending on the turbulence properties.",,https://iopscience.iop.org/article/10.3847/1538-4357/aaa0ca/pdf
-regional implantation of autologous adipose tissue-derived cells induces a prompt healing of long-lasting indolent digital ulcers in patients with systemic sclerosis,N. Del Papa,"Digital ulcers (DUs) are a rather frequent and invalidating complication in systemic sclerosis (SSc), often showing a very slow or null tendency to heal, in spite of the commonly used systemic and local therapeutic procedures. Recently, stem cell therapy has emerged as a new approach to accelerate wound healing. In the present study, we have tentatively treated long-lasting and poorly responsive to traditional therapy SSc-related DUs by implantation of autologous adipose tissue-derived cell (ATDC) fractions. Fifteen patients with SSc having a long-lasting DU in only one fingertip who were unresponsive to intensive systemic and local treatment were enrolled in the study. The grafting procedure consisted of the injection, at the basis of the corresponding finger, of 0.5-1 ml of autologous ATDC fractions, separated by centrifugation of adipose tissue collected through liposuction from subcutaneous abdominal fat. Time to heal after the procedure was the primary end point of the study, while reduction of pain intensity and of analgesic consumption represented a secondary end point. Furthermore, the posttherapy variation of the number of capillaries, observed in the nailfold video capillaroscopy (NVC) exam and of the resistivity in the digit arteries, measured by high-resolution echocolor-Doppler, were also taken into account. A rather fast healing of the DUs was reached in all of the enrolled patients (mean time to healing 4.23 weeks; range 2-7 weeks). A significant reduction of pain intensity was observed after a few weeks (p < 0.001), while the number of capillaries was significantly increased at 3- and 6-month NVC assessment (p < 0.0001 in both cases). Finally, a significant after-treatment reduction of digit artery resistivity was also recorded (p < 0.0001). Even with the limitations related to the small number of patients included and to the open-label design of the study, the observed strongly favorable outcome suggests that local grafting with ATDCs could represent a promising option for the treatment of SSc-related DUs unresponsive to more consolidated therapies.","{'model': 'tldr@v2.0.0', 'text': 'The observed strongly favorable outcome suggests that local grafting with ATDCs could represent a promising option for the treatment of SSc-related DUs unresponsive to more consolidated therapies.'}",https://journals.sagepub.com/doi/pdf/10.3727/096368914X685636
-estimation method of prompt neutron decay constant reducing higher order mode effect by linear combination,R. Katano,"Abstract The linear combination method is proposed to reduce the higher order mode (HOM) effect on the measurement of the prompt neutron decay constant using the α-fitting method. Conventional α-fitting utilizes the pulsed neutron source and estimates the prompt neutron decay constant by fitting the neutron counts at a single detector after pulse injection with a single exponential function. The proposed method reduces the spatial HOM effect with linear combination of the neutron counts at multiple detectors. For verification, we applied the conventional method and the proposed method to the analytical solution of the diffusion theory and the Monte Carlo simulation to estimate the prompt neutron decay constant of a one-dimensional infinite slab. Comparison of these results indicates that the proposed method enables estimation with the reduced HOM effect as opposed to the conventional method. Through the verification, we confirmed that the proposed method can be a candidate for a measurement method of the prompt neutron decay constant.",,
-prompt elevation of rat serum α‐fetoprotein by acute liver injury following a single injection of ethionine,A. Watanabe,"The mechanism of increased a‐fetoprotein (AFP) production following a single injection of ethionine was investigated by using rats aged 5 weeks at the time of killing. Marked elevations of serum AFP concentrations occurred within 4 days in both male and female rats after administration of DL‐ethionine or L‐ethionine, although the increased levels of serum AFP and liver triglyceride in the adults were less marked in the male than in the female. No apparent necrosis of liver cells was observed in ethionine‐treated rats. Frequent administrations of adenosine triphosphate after a single dose of ethionine prevented the increases in liver triglyceride and serum AFP levels. The increased concentrations of serum AFP, reaching a maximum level within 4 days, occurred before a slight increase in incorporation of 3H‐thymidine into liver DNA. The serum AFP from ethionine‐treated rats was immunologically and electrophoretically indistinguishable from that of fetal, carbontetrachloride‐treated or hepatoma‐bearing rats. These observations suggest that the increased production of AFP in ethionine‐treated rats is closely associated with hepatic injury and is not the consequence of liver cell regeneration.","{'model': 'tldr@v2.0.0', 'text': 'The observations suggest that the increased production of AFP in ethionine‐treated rats is closely associated with hepatic injury and is not the consequence of liver cell regeneration.'}",
-split injection strategy for prompt cold starting and low white smoke emissions,M. Girotra,"Progressing needs for prompt cold start of direct injection Diesel engines is the motivation behind this study. Authors have examined the autoignition and combustion processes in the early firing cycles of the engine and proposed a strategy to reduce the cranking period and the white smoke emissions. The concept is to accelerate the preparation of the combustible mixture during the cranking process. This is achieved by splitting the injected fuel in two parts and controlling its timing. The duration of the first injection is limited such that the spray penetrates through the combustion chamber and evaporates before it reaches the walls. The dwell between the two injections is adjusted to allow time for the first spray to mix with the fresh charge, form a combustible mixture and start producing the autoignition radicals. The second part would evaporate and autoignite by reacting with the radicals before it reaches the cool walls. The strategy is verified on a 1.2 L Ford Diata Diesel engine equipped with a first generation common rail fuel injection system. The cycle resolved hydrocarbons, and NOx emissions are measured by high response detectors. In addition, the mass and constituents of the white smoke are measured. All the experiments for this paper have been conducted after the engine has been soaked at the normal room temperature for at least eight hours. The results showed that there is an optimum strategy for the split injection that would minimize the cranking period and white smoke emissions.Copyright © 2005 by ASME",,
-hyperbaric oxygen for ischemia due to injection of cosmetic fillers: case report and issues,R. Henderson,"Summary: Natural and synthetic fillers have revolutionized aesthetic facial rejuvenation and soft-tissue augmentation. We present a case highlighting the dangers of filler self-injection. A 37-year-old woman self-injected a dermal filler around both temples. She immediately experienced left--side hearing loss, blanching over the left face, and pain. Prompt treatment with hyaluronidase, topical nitro paste, and warm compresses ensued. An emergency computed tomography angiogram showed occlusion of a superficial temporal artery branch. We treated her with enoxaparin, aspirin, dexamethasone, piperacillin-tazobactam, and intradermal lidocaine. After 6 hyperbaric oxygen therapy (HBO2) treatments in 3 days, the patient showed improvement in appearance with markedly decreased ischemic discoloration and her hearing returned to baseline. Algorithms for treating such injuries generally neglect HBO2. HBO2 is thought to be efficacious in these situations by a variety of mechanisms: oxygenation of ischemic tissues, reduction of edema, amelioration of ischemic/reperfusion injury, promotion of angiogenesis and collagen maturation. Her resolved hearing highlights the utility of HBO2 in sudden hearing loss as well. Injectors should have guidelines for using product, not only on patients but staff as well. Filler courses should include handling complications and include HBO2 in their guidelines. Clinicians should remind patients to seek treatment from qualified clinicians. The goal of a bargain price using self-injection may quickly become expensive and disfiguring.","{'model': 'tldr@v2.0.0', 'text': 'A 37-year-old woman who self-injected a dermal filler around both temples experienced left--side hearing loss, blanching over the left face, and pain and received 6 hyperbaric oxygen therapy treatments, highlighting the utility of HBO2 in sudden hearing loss as well.'}",
-prompt radial electric field response to neutral beam injection,T. Rhodes,"A variety of low density discharges in the DIII-D tokamak exhibit a prompt response (<10 ms) to neutral beam injection as evidenced by changes in the Doppler shift of the density fluctuations. This variation in the Doppler shift, attributed to a changing Vθ,Er × B velocity, which is in turn due to a rapidly changing radial electric field Er, is observed from the deep core to the edge of these plasmas. The core changes are large and occur much more quickly than collisional beam equilibration times (≥ 120 ms for the core) indicating a different momentum transfer mechanism. Towards the edge the Doppler shift times approach the collisional transfer times. Theoretical predictions of changes in the core Er due to radial fast particle currents are not large enough to explain the observations. Although not yet understood this effect could lead to a tool for relatively fast feedback control of core electric fields in advanced confinement regimes.",,
-experience and management of intravascular injection with facial fillers: results of a multinational survey of experienced injectors,G. Goodman,,"{'model': 'tldr@v2.0.0', 'text': None}",
-detection of prompt luminescence from an imaging plate with heavy particle injection,S. Gunji,"We have been investigating the basic characteristics of an imaging plate JP) to develop a new type of ultra-heavy-particle detector. To acquire the information on the deposited energy, photostimulated luminescence (PSL) from the injection position is usually utilized, which occurs upon scanning the surface of the IP with a laser beam after the exposure of particles. However, in the use of PSL, there is the disadvantage that it fades with time. On the other hand, when heavy particles are injected to the IP, prompt luminescence (PL) also occurs from the injection position. We have succeeded in detecting the prompt luminescence from each heavy particle with the energy of several hundred MeV/n. Since the PL is useful for quick detection of the injection position, the PSL data can be read within a few seconds by scanning the determined position before any serious fading effect occurs.",,
-prompt enhancement of the earth's outer radiation belt due to substorm electron injections,C. Tang,"We present multipoint simultaneous observations of the near‐Earth magnetotail and outer radiation belt during the substorm electron injection event on 16 August 2013. Time History of Events and Macroscale Interactions during Substorms A in the near‐Earth magnetotail observed flux‐enhanced electrons of 300 keV during the magnetic field dipolarization. Geosynchronous orbit satellites also observed the intensive electron injections. Located in the outer radiation belt, RBSP‐A observed enhancements of MeV electrons accompanied by substorm dipolarization. The phase space density (PSD) of MeV electrons at L*~5.4 increased by 1 order of magnitude in 1 h, resulting in a local PSD peak of MeV electrons, which was caused by the direct effect of substorm injections. Enhanced MeV electrons in the heart of the outer radiation belt were also detected within 2 h, which may be associated with intensive substorm electron injections and subsequent local acceleration by chorus waves. Multipoint observations have shown that substorm electron injections not only can be the external source of MeV electrons at the outer edge of the outer radiation belt (L*~5.4) but also can provide the intensive seed populations in the outer radiation belt. These initial higher‐energy electrons from injection can reach relativistic energy much faster. The observations also provide evidence that enhanced substorm electron injections can explain rapid enhancements of MeV electrons in the outer radiation belt.",,https://agupubs.onlinelibrary.wiley.com/doi/pdfdirect/10.1002/2016JA023550
-hot electron injection from graphene quantum dots to tio₂.,K. J. Williams,"The Shockley-Queisser limit is the maximum power conversion efficiency of a conventional solar cell based on a single semiconductor junction. One approach to exceed this limit is to harvest hot electrons/holes that have achieved quasi-equilibrium in the light absorbing material with electronic temperatures higher than the phonon temperature. We argue that graphene based materials are viable candidates for hot carrier chromophores. Here we probe hot electron injection and charge recombination dynamics for graphene quantum dots (QDs, each containing 48 fused benzene rings) anchored to the TiO₂(110) surface via carboxyl linkers. We find ultrafast electron injection from photoexcited graphene QDs to the TiO₂ conduction band with time constant τ(i) < 15 fs and charge recombination dynamics characterized by a fast channel (τ(r1) = 80-130 fs) and a slow one (τ(r2) = 0.5-2 ps). The fast decay channel is attributed to the prompt recombination of the bound electron-hole pair across the interface. The slow channel depends strongly on excitation photon energy or sample temperature and can be explained by a ""boomerang"" mechanism, in which hot electrons are injected into bulk TiO₂, cooled down due to electron-phonon scattering, drifted back to the interface under the transient electric field, and recombine with the hole on graphene QDs. We discuss feasibilities of implementing the hot carrier solar cell using graphene nanomaterials.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that graphene based materials are viable candidates for hot carrier chromophores and feasibilities of implementing the hot carrier solar cell using graphene nanomaterials are discussed.'}",
-prompt loss of beam ions in kstar plasmas,Jun Young Kim,"For a toroidal plasma facility to realize fusion energy, researching the transport of fast ions is important not only due to its close relation to the heating and current drive efficiencies but also to determine the heat load on the plasma-facing components. We present a theoretical analysis and orbit simulation for the origin of lost fast-ions during neutral beam injection (NBI) heating in Korea Superconducting Tokamak Advanced Research (KSTAR) device. We adopted a two-dimensional phase diagram of the toroidal momentum and magnetic moment and describe detectable momentums at the fast-ion loss detector (FILD) position as a quadratic line. This simple method was used to model birth ions deposited by NBI and drawn as points in the momentum phase space. A Lorentz orbit code was used to calculate the fast-ion orbits and present the prompt loss characteristics of the KSTAR NBI. The scrape-off layer deposition of fast ions produces a significant prompt loss, and the model and experimental results closely agreed on the pitch-angle range of the NBI prompt loss. Our approach can provide wall load information from the fast ion loss.",,https://aip.scitation.org/doi/pdf/10.1063/1.4966588
-combined intravitreal bevacizumab injection and zone i sparing laser photocoagulation in patients with zone i retinopathy of prematurity,Jaeryung Kim,"Purpose: To investigate the anatomical outcome of combined intravitreal bevacizumab injection and Zone I sparing laser ablation in patients with Type 1 retinopathy of prematurity in Zone I. Methods: The medical records of consecutive 18 eyes of 10 infants, who underwent combined intravitreal bevacizumab (0.25 mg) injection and Zone I sparing laser ablation for the treatment of Type 1 retinopathy of prematurity in Zone I, were retrospectively reviewed. Laser photocoagulation was performed on the avascular retina anterior to the margin of Zone I extending to the ora serrata. Anatomical outcomes including progression to stage 4/5, macular changes, and vitreous organization were reviewed. Results: The mean gestational age at birth and the birth weight of included patients were 24.0 weeks and 628 g, respectively. The timing of bevacizumab injection ranged from postmenstrual age 33+2 to 35 weeks (mean, 34.3 weeks). Postmenstrual age at last follow-up ranged from 74+6 to 107+1 weeks (mean, 83.6 weeks). All 18 eyes demonstrated prompt regression of neovascular pathology and plus disease without recurrence. Previously avascular Zone I retina was vascularized in all eyes after the treatment. All eyes showed excellent anatomical outcome with intact macula, but one eye showed mild vitreous organization above the vascular/avascular junction. Conclusion: Combined intravitreal bevacizumab injection and Zone I sparing laser ablation for Type 1 retinopathy of prematurity in Zone I seem to be effective treatment options. Possible advantages include lower dose of anti–vascular endothelial growth factor, less recurrence than monotherapy, and preservation of central visual field.","{'model': 'tldr@v2.0.0', 'text': 'All eyes showed excellent anatomical outcome with intact macula, but one eye showed mild vitreous organization above the vascular/avascular junction, and possible advantages include lower dose of anti–vascular endothelial growth factor, less recurrence than monotherapy, and preservation of central visual field.'}",
-intralesional antibiotic injection using 41g needle for the management of subretinal abscess in endogenous endophthalmitis,P. Venkatesh,,"{'model': 'tldr@v2.0.0', 'text': 'Intralesional pharmacotherapy may be safe and effective in the treatment of subretinal abscess associated with endogenous endophthalmitis.'}",https://journalretinavitreous.biomedcentral.com/track/pdf/10.1186/s40942-016-0043-x
-sudden visual loss and multiple cerebral infarction after autologous fat injection into the glabella.,Dongkyun Hong,"This case demonstrates that patients may develop nerve injury even when using the lowest energy level. Even in this case with nerve injury, the long-term benefits of improved hyperhidrosis may outweigh such adverse effects as weakness of the left hand that recovered after 12 months. Also, the patient improved, neurologically and functionally, with intensive physiotherapy after 6 months of persistent weakness and numbness. Based on this observation, we recommend prompt and intensive rehabilitation to accelerate recovery if neural damage occurs unexpectedly. Moreover, more caution is needed when treating thin male patients.","{'model': 'tldr@v2.0.0', 'text': 'This case demonstrates that patients may develop nerve injury even when using the lowest energy level, and recommends prompt and intensive rehabilitation to accelerate recovery if neural damage occurs unexpectedly.'}",
-the roles of traditional chinese medicine: shen-fu injection on the postresuscitation care bundle,Qian Zhang,"Survival rates following in-hospital and out-of-hospital cardiac arrests remain disappointingly low. Organ injury caused by ischemia and hypoxia during prolonged cardiac arrest is compounded by reperfusion injury that occurs when a spontaneous circulation is restored. A bundle of procedures, which may need to be administered simultaneously, is required. The procedures include prompt identification and treatment of the cause of cardiac arrest, as well as a definitive airway and ventilation together. Additional benefit is possible with appropriate forms of early goal-directed therapy and achieving therapeutic hypothermia within the first few hours, followed by gradual rewarming and ensuring glycaemic control to be within a range of 6 to 10 mmol/L. All these would be important and need to be continued for at least 24 hours. Previous studies have showed that the effects of Shen-Fu injection (SFI) are based on aconitine properties, supplemented by ginsenoside, which can scavenge free radicals, improve energy metabolism, inhibit inflammatory mediators, suppress cell apoptosis, and alleviate mitochondrial damage. SFI, like many other complex prescriptions of traditional Chinese medicine, was also found to be more effective than any of its ingredient used separately in vivo. As the postresuscitation care bundle is known to be, the present paper focuses on the role of SFI played on the postresuscitation care bundle.","{'model': 'tldr@v2.0.0', 'text': 'The role of SFI played on the postresuscitation care bundle is focused on and Shen-Fu injection was found to be more effective than any of its ingredient used separately in vivo.'}",https://downloads.hindawi.com/journals/ecam/2013/319092.pdf
-chat-rec: towards interactive and explainable llms-augmented recommender system,Yunfan Gao,"Large language models (LLMs) have demonstrated their significant potential to be applied for addressing various application tasks. However, traditional recommender systems continue to face great challenges such as poor interactivity and explainability, which actually also hinder their broad deployment in real-world systems. To address these limitations, this paper proposes a novel paradigm called Chat-Rec (ChatGPT Augmented Recommender System) that innovatively augments LLMs for building conversational recommender systems by converting user profiles and historical interactions into prompts. Chat-Rec is demonstrated to be effective in learning user preferences and establishing connections between users and products through in-context learning, which also makes the recommendation process more interactive and explainable. What's more, within the Chat-Rec framework, user's preferences can transfer to different products for cross-domain recommendations, and prompt-based injection of information into LLMs can also handle the cold-start scenarios with new items. In our experiments, Chat-Rec effectively improve the results of top-k recommendations and performs better in zero-shot rating prediction task. Chat-Rec offers a novel approach to improving recommender systems and presents new practical scenarios for the implementation of AIGC (AI generated content) in recommender system studies.","{'model': 'tldr@v2.0.0', 'text': 'A novel paradigm called Chat-Rec (ChatGPT Augmented Recommender System) that innovatively augments LLMs for building conversational recommender systems by converting user profiles and historical interactions into prompts is proposed.'}",http://arxiv.org/pdf/2303.14524
-an overview of anthrax infection including the recently identified form of disease in injection drug users,C. Hicks,,"{'model': 'tldr@v2.0.0', 'text': 'While antibiotics are the mainstay of anthrax treatment, use of adjunctive therapies such as anthrax toxin antagonists are a consideration and Prompt surgical therapy appears to be important for successful management of injectional anthrax.'}",https://europepmc.org/articles/pmc3523299?pdf=render
-from chatgpt to threatgpt: impact of generative ai in cybersecurity and privacy,Maanak Gupta,"Undoubtedly, the evolution of Generative AI (GenAI) models has been the highlight of digital transformation in the year 2022. As the different GenAI models like ChatGPT and Google Bard continue to foster their complexity and capability, it’s critical to understand its consequences from a cybersecurity perspective. Several instances recently have demonstrated the use of GenAI tools in both the defensive and offensive side of cybersecurity, and focusing on the social, ethical and privacy implications this technology possesses. This research paper highlights the limitations, challenges, potential risks, and opportunities of GenAI in the domain of cybersecurity and privacy. The work presents the vulnerabilities of ChatGPT, which can be exploited by malicious users to exfiltrate malicious information bypassing the ethical constraints on the model. This paper demonstrates successful example attacks like Jailbreaks, reverse psychology, and prompt injection attacks on the ChatGPT. The paper also investigates how cyber offenders can use the GenAI tools in developing cyber attacks, and explore the scenarios where ChatGPT can be used by adversaries to create social engineering attacks, phishing attacks, automated hacking, attack payload generation, malware creation, and polymorphic malware. This paper then examines defense techniques and uses GenAI tools to improve security measures, including cyber defense automation, reporting, threat intelligence, secure code generation and detection, attack identification, developing ethical guidelines, incidence response plans, and malware detection. We will also discuss the social, legal, and ethical implications of ChatGPT. In conclusion, the paper highlights open challenges and future directions to make this GenAI secure, safe, trustworthy, and ethical as the community understands its cybersecurity impacts.","{'model': 'tldr@v2.0.0', 'text': 'The work presents the vulnerabilities of ChatGPT, which can be exploited by malicious users to exfiltrate malicious information bypassing the ethical constraints on the model, and uses GenAI tools to improve security measures, including cyber defense automation, reporting, threat intelligence, secure code generation and detection, attack identification, developing ethical guidelines, incidence response plans, and malware detection.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/10198233.pdf
-the influence of boundary layers on supersonic inlet flow unstart induced by mass injection,H. Do,,,
-international consensus on (icon) anaphylaxis,L. Ardusso,,"{'model': 'tldr@v2.0.0', 'text': 'These guidelines concur about prompt initial treatment with intramuscular injection of epinephrine (adrenaline) in the mid-outer thigh, positioning the patient supine, calling for help, and when indicated, providing supplemental oxygen, intravenous fluid resuscitation and cardiopulmonary resuscitation.'}",http://www.worldallergyorganizationjournal.org/article/S1939455119302406/pdf
-numerical investigation of the plasma-assisted mild combustion of a ch4/h2 fuel blend under various working conditions,S. M. Mousavi," The effects of plasma injection upon MILD combustion of a mixture of methane and hydrogen are investigated numerically. The injected plasma includes the flow of a highly air-diluted methane including C2H2, C2H4, C2H6, CH, CH2, CH3, CO, and CO2. The results show that among all the constitutes of plasma, CH3 is the most effective in improving the characteristics of MILD combustion. Injection of this radical leads to the occurrence of reactions at a closer distance to the burner inlet and thus provides longer time for completion of combustion. Further, mass fractions of OH, CH2O, and HCO are considerably affected by the injections of CH3, indicating structural modifications of the reacting flow. Importantly, as Reynolds number of the plasma flow increases, the volume and width of the flame decrease, while the formations of prompt and thermal NOx are intensified. However, injection of CH3, as plasma, reduces the emission of thermal NOx.",,https://asmedigitalcollection.asme.org/energyresources/article-pdf/143/6/062302/6686005/jert_143_6_062302.pdf
-modeling the response of a fast ion loss detector using orbit tracing techniques in a neutral beam prompt-loss study on the diii-d tokamak.,D. Pace,A numerical model describing the expected measurements of neutral beam prompt-losses by a newly commissioned fast ion loss detector (FILD) in DIII-D is presented. This model incorporates the well understood neutral beam deposition profiles from all eight DIII-D beamlines to construct a prompt-loss source distribution. The full range of detectable ion orbit phase space available to the FILD is used to calculate ion trajectories that overlap with neutral beam injection footprints. Weight functions are applied to account for the level of overlap between these detectable orbits and the spatial and velocity (pitch) properties of ionized beam neutrals. An experimental comparison is performed by firing each neutral beam individually in the presence of a ramping plasma current. Fast ion losses determined from the model are in agreement with measured losses.,"{'model': 'tldr@v2.0.0', 'text': 'A numerical model describing the expected measurements of neutral beam prompt-losses by a newly commissioned fast ion loss detector (FILD) in DIII-D is presented.'}",https://escholarship.org/content/qt9j11k20t/qt9j11k20t.pdf?t=p15iy8
-"complications of injectable fillers, part 2: vascular complications.",C. Delorenzi,"Accidental intra-arterial filler injection may cause significant tissue injury and necrosis. Hyaluronic acid (HA) fillers, currently the most popular, are the focus of this article, which highlights complications and their symptoms, risk factors, and possible treatment strategies. Although ischemic events do happen and are therefore important to discuss, they seem to be exceptionally rare and represent a small percentage of complications in individual clinical practices. However, the true incidence of this complication is unknown because of underreporting by clinicians. Typical clinical findings include skin blanching, livedo reticularis, slow capillary refill, and dusky blue-red discoloration, followed a few days later by blister formation and finally tissue slough. Mainstays of treatment (apart from avoidance by meticulous technique) are prompt recognition, immediate treatment with hyaluronidase, topical nitropaste under occlusion, oral acetylsalicylic acid (aspirin), warm compresses, and vigorous massage. Secondary lines of treatment may involve intra-arterial hyaluronidase, hyperbaric oxygen therapy, and ancillary vasodilating agents such as prostaglandin E1. Emergency preparedness (a ""filler crash cart"") is emphasized, since early intervention is likely to significantly reduce morbidity. A clinical summary chart is provided, organized by complication presentation.","{'model': 'tldr@v2.0.0', 'text': 'Hyaluronic acid (HA) fillers, currently the most popular, are the focus of this article, which highlights complications and their symptoms, risk factors, and possible treatment strategies, and emergency preparedness is emphasized.'}",https://academic.oup.com/asj/article-pdf/34/4/584/23787559/34-4-584.pdf
-"cd4pos, nk1.1pos t cells promptly produce interleukin 4 in response to in vivo challenge with anti-cd3",T. Yoshimoto,"Injection of anti-CD3 antibodies causes prompt expression of interleukin (IL)-4, IL-2, and interferon gamma (IFN-gamma) mRNA among spleen cells. The optimal dose of anti-CD3 for such induction was 1.33 microgram/animal; lymphokine mRNA was first observed at 30 min, peaked at 90 min, and was undetectable (for IL-4) or had declined markedly by 4 h. Cells harvested from spleens of mice injected with anti-CD3 90 min earlier secreted IL-4, IL-2, and IFN-gamma without further stimulation. By contrast, in vitro stimulation with anti-CD3 of spleen cell suspensions or splenic fragments from noninjected donors failed to cause prompt production of IL-4 and, even after 24 h of stimulation, the amount of IL-4 produced in such cells was substantially less than that secreted within 1 h by spleen cell suspensions or splenic fragments from mice injected with anti-CD3 90 min earlier. Production of IL-4 by spleen cells from anti-CD3-injected mice was not inhibited by pretreatment with anti-IL-4 antibody or with IFN-gamma or tumor growth factor beta nor enhanced by treatment with IL-4. By contrast, CTLA-4 immunoglobulin (Ig) treatment clearly diminished IL-4 production in response to in vivo anti-CD3, indicating that cellular interactions involving CD28 (or related molecules) were important in stimulation. Cell sorting analysis indicated that the cells that produced IL-4 in response to in vivo injection of anti-CD3 were highly enriched in CD4pos cells with the phenotype leukocyte cell adhesion molecule-1 (LECAM-1)dull, CD44bright, CD45RBdull, NK1.1pos. Indeed, the small population of CD4pos, NK1.1pos cells had the great majority of the IL-4- producing activity of this population. Injection with Staphylococcal enterotoxin B also caused prompt induction of IL-4 mRNA; the cells that were principally responsible for production also had the phenotype of CD4pos, NK1.1pos. These results suggest that possibility that this rare population of T cells may be capable of secreting IL-4 at the outset of immune responses and thus may act to regulate the pattern of priming of naive T cells, by providing a source of IL-4 to favor the development of T cell helper 2-like IL-4-producing cells.","{'model': 'tldr@v2.0.0', 'text': 'Cell sorting analysis indicated that the cells that produced IL-4 in response to in vivo injection of anti-CD3 were highly enriched in CD4pos cells with the phenotype leukocyte cell adhesion molecule-1 (LECAM-1)dull, CD44bright, CD45RBdull and NK1.1pos, suggesting that possibility that this rare population of T cells may be capable of secreting IL- 4 at the outset of immune responses is suggested.'}",https://rupress.org/jem/article-pdf/179/4/1285/1104795/1285.pdf
-"butyrate, a metabolite of intestinal bacteria, enhances sleep",É. Szentirmai,,"{'model': 'tldr@v2.0.0', 'text': 'The results suggest that the sleep-inducing effects of butyrate are mediated by a sensory mechanism located in the liver and/or in the portal vein wall, and Hepatoportalbutyrate-sensitive mechanisms may play a role in sleep modulation by the intestinal microbiota.'}",https://www.nature.com/articles/s41598-019-43502-1.pdf
-critical contribution of liver natural killer t cells to a murine model of hepatitis.,Kazuyoshi Takeda,"Natural killer T (NKT) cells constitute a distinct subpopulation of T cells with a unique antigen specificity, prompt effector functions, and an unusual tissue distribution. NKT cells are especially abundant in the liver, but their physiological function in this organ remains unclear. In the present study, we examined the possible contribution of NKT cells to a murine model of hepatitis induced by i.v. injection of Con A. CD1-deficient mice lacking NKT cells were highly resistant to Con A-induced hepatitis. Adoptive transfer of hepatic NKT cells isolated from wild-type mice, but not from FasL-deficient gld mice, sensitized CD1-deficient mice to Con A-induced hepatitis. Furthermore, adoptive transfer of hepatic mononuclear cells from wild-type mice, but not from CD1-deficient mice, sensitized gld mice to Con A-induced hepatitis. Upon Con A administration, hepatic NKT cells rapidly up-regulated cell surface FasL expression and FasL-mediated cytotoxicity. At the same time, NKT cells underwent apoptosis leading to their rapid disappearance in the liver. These results implicated FasL expression on liver NKT cells in the pathogenesis of Con A-induced hepatitis, suggesting a similar pathogenic role in human liver diseases such as autoimmune hepatitis.","{'model': 'tldr@v2.0.0', 'text': 'Results implicated FasL expression on liver NKT cells in the pathogenesis of Con A-induced hepatitis, suggesting a similar pathogenic role in human liver diseases such as autoimmune hepatitis.'}",https://europepmc.org/articles/pmc25857?pdf=render
-dynamic modeling of radiation belt electrons by radial diffusion simulation for a 2 month interval following the 24 march 1991 storm injection,F. Chu,"[1] Diffusive radial transport of radiation belt electrons with variable outer boundary is computed using Brautigam and Albert (2000) diffusion coefficients parameterized by Kp, modeling power level at ULF wave frequencies in the range of MeV electron drift periods. We analyzed radial diffusion during a relatively quiet 2 month interval following the 24 March 1991, prompt injection to form a new radiation belt at L* = 2.5. The radial diffusion calculation is initialized with a computed phase space density (PSD) profile using differential flux values from the CRRES HEEF instrument, covering 0.65–7.5 MeV. The outer boundary phase space density is updated using Los Alamos National Laboratory (LANL) GEO satellite fluxes, changing the ratio of PSD relative to a quiet day by assuming the outer boundary is changing proportional to the flux at a LANL GEO satellite. The location of the plasmapause Lp* is computed using a Kp-dependent formula separating different loss rates inside and outside the plasmapause. A series of simulations for different values of the first invariant is performed for this 2 month period. The flux is then interpolated to find electron flux at a fixed energy, 1 MeV, in order to compare with the CRRES satellite 1 MeV flux. Radial diffusion appears to be the dominant mechanism for this 2 month interval, which contains moderate storms (∣Dst∣ ≲ 100). Modulation of fluxes measured by CRRES compare well with simulations of the outer zone flux peak at L* = 3–4 for moderate high-speed stream-driven storms, along with persistence and slow decay of the new population of electrons injected on 24 March 1991, into L* = 2.5. The strongest storm of the 2 month interval (Dst = −105 nT) produced a flux dropout, which is not well-captured by the model, suggesting that improvements to the Kp-parameterized loss model are needed for larger storms.",,
-epinephrine for first-aid management of anaphylaxis,S. Sicherer,"Anaphylaxis is a severe, generalized allergic or hypersensitivity reaction that is rapid in onset and may cause death. Epinephrine (adrenaline) can be life-saving when administered as rapidly as possible once anaphylaxis is recognized. This clinical report from the American Academy of Pediatrics is an update of the 2007 clinical report on this topic. It provides information to help clinicians identify patients at risk of anaphylaxis and new information about epinephrine and epinephrine autoinjectors (EAs). The report also highlights the importance of patient and family education about the recognition and management of anaphylaxis in the community. Key points emphasized include the following: (1) validated clinical criteria are available to facilitate prompt diagnosis of anaphylaxis; (2) prompt intramuscular epinephrine injection in the mid-outer thigh reduces hospitalizations, morbidity, and mortality; (3) prescribing EAs facilitates timely epinephrine injection in community settings for patients with a history of anaphylaxis and, if specific circumstances warrant, for some high-risk patients who have not previously experienced anaphylaxis; (4) prescribing epinephrine for infants and young children weighing <15 kg, especially those who weigh 7.5 kg and under, currently presents a dilemma, because the lowest dose available in EAs, 0.15 mg, is a high dose for many infants and some young children; (5) effective management of anaphylaxis in the community requires a comprehensive approach involving children, families, preschools, schools, camps, and sports organizations; and (6) prevention of anaphylaxis recurrences involves confirmation of the trigger, discussion of specific allergen avoidance, allergen immunotherapy (eg, with stinging insect venom, if relevant), and a written, personalized anaphylaxis emergency action plan; and (7) the management of anaphylaxis also involves education of children and supervising adults about anaphylaxis recognition and first-aid treatment.","{'model': 'tldr@v2.0.0', 'text': 'This clinical report from the American Academy of Pediatrics is an update of the 2007 clinical report on this topic and provides information to help clinicians identify patients at risk of anaphylaxis and new information aboutEpinephrine and epinephrine autoinjectors (EAs).'}",https://pediatrics.aappublications.org/content/pediatrics/139/3/e20164006.full.pdf
-diversity of kilonova light curves,K. Kawaguchi,"We perform radiative transfer simulations for kilonova in various situations, including the cases of prompt collapse to a black hole from neutron star mergers, high-velocity ejecta possibly accelerated by magnetars, and a black hole–neutron star merger. Our calculations are done employing ejecta profiles predicted by numerical-relativity simulations and a new line list for all the r-process elements. We found that: (i) the optical emission for binary neutron stars promptly collapsing to a black hole would be fainter by ≳1–2 mag than that found in GW170817, while the infrared emission could be as bright as that in GW170817 if the post-merger ejecta is as massive as ≈0.01 M⊙; (ii) the kilonova would be brighter than that observed in GW170817 for the case that the ejecta is highly accelerated by the electromagnetic energy injection from the remnant, but within a few days it would decline rapidly and the magnitude would become fainter than in GW170817; and (iii) the optical emission from a black hole–neutron star merger ejecta could be as bright as that observed in GW170817 for the case that sufficiently large amount of matter is ejected (≳0.02 M⊙), while the infrared brightness would be brighter by 1–2 mag at the same time. We show that the difference in the ejecta properties would be imprinted in the differences in the peak brightness and time of peak. This indicates that we may be able to infer the type of the central engine for kilonovae by observation of the peak in the multiple band.",,https://iopscience.iop.org/article/10.3847/1538-4357/ab61f6/pdf
-the time course of anaphylaxis manifestations in children is diverse and unpredictable,G. Pouessel,"The most current clinical definition of anaphylaxis is that proposed by Sampson et al. and different scoring systems have been proposed to assess its severity [1,2]. Anaphylaxis symptoms range from mild, self-limiting local reactions to life-threatening anaphylaxis. All anaphylaxis guidelines emphasize early recognition and the need for prompt intramuscular adrenaline injection as the first-line treatment [3]. However, adrenaline for anaphylaxis is underused by patients, families and even by health professionals. In the European anaphylaxis registry, 27% of patients treated by a health professional received adrenaline [4].","{'model': 'tldr@v2.0.0', 'text': 'The most current clinical definition of anaphylaxis is that proposed by Sampson et al. and different scoring systems have been proposed to assess its severity and underused by patients, families and even by health professionals.'}",
-high-pressure injection injury of the hand: an often underestimated trauma: case report with study of the literature,N. Verhoeven,,"{'model': 'tldr@v2.0.0', 'text': 'Not only the latency time to adequate treatment but also the injected fluid’s nature, the pressure, the volume and the location of injection, has influence on the seriousness and extensiveness of subcutaneous damage.'}",
-beam-ion confinement for different injection geometries,W. Heidbrink,"The DIII-D tokamak is equipped with neutral beam sources that inject in four different directions; in addition, the plasma can be moved up or down to compare off-axis with on-axis injection. Fast-ion data for eight different conditions have been obtained: co/counter, near-tangential/near-perpendicular and on-axis/off-axis. Neutron measurements during short beam pulses assess prompt and delayed losses under low-power conditions. As expected, co-injection has fewer losses than counter, tangential fewer than perpendicular and on-axis fewer than off-axis; the differences are greater at low current than at higher current. The helicity of the magnetic field has a weak effect on the overall confinement. Fast-ion Dα (FIDA) and neutron measurements diagnose the confinement at higher power. The basic trends are the same as in low-power plasmas but, even in plasmas without long wavelength Alfvén modes or other MHD, discrepancies with theory are observed, especially in higher temperature plasmas. At modest temperature, two-dimensional images of the FIDA light are in good agreement with the simulations for both on-axis and off-axis injection. Discrepancies with theory are more pronounced at low fast-ion energy and at high plasma temperature, suggesting that fast-ion transport by microturbulence is responsible for the anomalies.",,
-quantitation in pet using isotopes emitting prompt single gammas: application to yttrium-86,S. Walrand,,"{'model': 'tldr@v2.0.0', 'text': 'A patient-dependent correction method based on sinogram tail fitting using an 86Y point spread function library significantly improved the accuracy of the quantification and the mean discrepancy between PET total body activity and the activity expected from urinary collections was reduced.'}",
-the experimental disease that follows the injection of human leprosy bacilli into foot-pads of mice,C. Shepard,"When leprosy bacilli from human patients are inoculated into the foot-pads of CFW mice, a microscopic granuloma containing acid-fast bacilli develops in a characteristic manner. This has been seen in 22 of 22 instances with leprosy bacilli from nasal washings, in 12 of 16 instances with leprosy bacilli from skin biopsies, and in none of 16 cases where the nasal washings were not observed to contain leprosy bacilli. Quantitative studies revealed a relationship between the number of bacilli inoculated and the time required for the appearance of the lesions. The incubation period was usually 1 to 2 months when the dose was 105.5 to 106.0 bacilli and about 6 months when the dose was about 103 organisms. After the development of the lesion, the number of bacilli harvested was usually in the range 104.5 to 106.0, regardless of the number inoculated. When the inoculum has contained 102.0 to 103.5 acid-fast bacilli, and harvests were reasonably prompt, there were regular increases of 50- to 1000-fold. Passage to new groups of mice has been successful 11 of 12 times. Most of these were second passages. One strain has been maintained in 3 passages with a total increase in acid-fast bacilli of 4 x 104-fold. Another strain has been through 4 passages with a total increase of about 4 x 106-fold. Cultures on bacteriological media favorable for the growth of most known mycobacterial species have not shown growth of mycobacteria.","{'model': 'tldr@v2.0.0', 'text': 'Quantitative studies revealed a relationship between the number of bacilli inoculated and the time required for the appearance of the lesions, and Cultures on bacteriological media favorable for the growth of most known mycobacterial species have not shown growth of myCobacteria.'}",http://jem.rupress.org/content/jem/112/3/445.full.pdf
-evidence for a structurally-driven insulator-to-metal transition in vo 2 : a view from the ultrafast timescale,A. Cavalleri,"We apply ultrafast spectroscopy to establish a time-domain hierarchy between structural and electronic effects in a strongly correlated electron system. We discuss the case of the model system ${\mathrm{VO}}_{2}$, a prototypical nonmagnetic compound that exhibits cell doubling, charge localization, and a metal-insulator transition below 340 K. We initiate the formation of the metallic phase by prompt hole photo-doping into the valence band of the low-$T$ insulator. The insulator-to-metal transition is, however, delayed with respect to hole injection, exhibiting a bottleneck time scale, associated with the phonon connecting the two crystallographic phases. This structural bottleneck is observed despite faster depletion of the $d$ bands and is indicative of important bandlike character for this controversial insulator.",,https://kops.uni-konstanz.de/bitstream/123456789/5099/1/Evidence_for_a_structurally_driven_insulator_to_metal_transition_in_VO2.pdf
-tractional retinal detachment after intravitreal injection of bevacizumab in proliferative diabetic retinopathy.,M. Torres-Soriano,"PURPOSE To assess the short-term complications of a single dose of intravitreal bevacizumab in patients with proliferative diabetic retinopathy (PDR). METHODS Retrospective review of 343 patients with PDR who were treated with intravitreal injection of bevacizumab (2.5 mg/0.1 mL). RESULTS Five patients (1.45%) presented tractional retinal detachment 1 to 6 weeks (mean 3 weeks) after intravitreal injection. All cases underwent pars plana vitrectomy, removal of all epiretinal fibrovascular membranes, further endolaser panretinal photocoagulation, and silicone tamponade. CONCLUSION Tractional retinal detachment may occur in a short time post intravitreal injection of bevacizumab in patients with proliferative diabetic retinopathy with extensive areas of ischemia and fibrovascular proliferations, and may require prompt vitreoretinal surgery.","{'model': 'tldr@v2.0.0', 'text': 'Tractional retinal detachment may occur in a short time post intravitreal injection of bevacizumab in patients with proliferative diabetic retinopathy with extensive areas of ischemia and fibrovascular proliferations, and may require prompt vitreoretinal surgery.'}",
-prompt and powerful effect of a practice guideline on caudal additives,C. Eich,"not be ruled out completely in this event. Intravenous anticholinergic agents are the treatment of choice for these dystonic reactions, however more than one dose may be necessary for complete resolution of dystonia. An additional dose of diphenhydramine may have been beneficial in this patient. Although usually self limited, complications after Botox injections may be very disabling. Patients with communication problems experiencing these complications pose a challenge for the care team, especially in the context of prescheduled pediatric ambulatory care. In order to avoid these complications, optimal dose, injection volume, longterm cumulative effect and injection techniques should be studied further. Claude Abdallah Raafat Hannallah Division of Anesthesiology, Children’s National Medical Center, 111 Michigan Ave, NW, Washington, DC, USA (email: cabdalla@cnmc.org)","{'model': 'tldr@v2.0.0', 'text': 'Intravenous anticholinergic agents are the treatment of choice for these dystonic reactions, however more than one dose may be necessary for complete resolution of dystonia.'}",
-anterior segment ischemia and retinochoroidal vascular occlusion after intralesional steroid injection.,A. Yağcı,"A 19-year-old woman was admitted with sudden severe pain and loss of light perception that began immediately after eyelid injection of the depot form of corticosteroid. Ecchymosis of the lower eyelid, corneal edema, and a fixed dilated pupil were noted. Fundus examination could not be performed because of corneal edema. Embolic material packed in the episcleral vessels was detected. With these findings, the diagnosis was anterior and posterior segment ischemia. Despite administration of an intravenous hyperosmotic agent, in addition to topical and systemic pulse corticosteroid (1.0 g/day), vision was not recovered. The most serious complication of intralesional injection of corticosteroid is vascular occlusion. This catastrophic complication after intralesional steroid injection should prompt immediate ophthalmoscopic fundus examination to allow early recognition of ischemic signs and immediate intervention.","{'model': 'tldr@v2.0.0', 'text': 'A 19-year-old woman admitted with sudden severe pain and loss of light perception that began immediately after eyelid injection of the depot form of corticosteroid was diagnosed with anterior and posterior segment ischemia.'}",
-inflammation‐inducible anti‐tnf gene expression mediated by intra‐articular injection of serotype 5 adeno‐associated virus reduces arthritis,M. Khoury,"The tumor necrosis factor (TNF)‐α plays a central role in rheumatoid arthritis (RA) and current biotherapies targeting TNF‐α have a major impact on RA treatment. The long‐term safety concerns associated with the repetitive TNF blockade prompt optimization of therapeutic anti‐TNF approaches. Since we recently demonstrated that intra‐articular gene transfer using a recombinant adeno‐associated virus serotype 5 (rAAV5) efficiently transduces arthritic joints, we evaluate its effect on collagen‐induced arthritis (CIA) when encoding TNF antagonists.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that intra‐articular gene transfer using a recombinant adeno‐associated virus serotype 5 (rAAV5) efficiently transduces arthritic joints, and its effect on collagen‐induced arthritis (CIA) when encoding TNF antagonists is evaluated.'}",
-retinal branch artery occlusion following injection of hyaluronic acid (restylane),S. Peter,"The authors report a retinal branch artery occlusion occurring after facial injection of a dermal filler. The superior temporal artery showed occlusion due to a clearly visible long and fragmented embolus suggestive of gel and clearly distinguishable from calcific or cholesterol emboli. The authors suppose that hyaluronic acid gel was embolized in the patient. The embolized material is supposed to enter the ocular circulation through retrograde arteriolar flow after intra‐arterial injection into one of the peripheral branches of the ophthalmic artery. If there is any evidence of a visual problem after facial injection of a dermal filler, prompt consultation of an ophthalmologist is recommended.","{'model': 'tldr@v2.0.0', 'text': 'The superior temporal artery showed occlusion due to a clearly visible long and fragmented embolus suggestive of gel and clearly distinguishable from calcific or cholesterol emboli and the authors suppose that hyaluronic acid gel was embolized in the patient.'}",
-treatment of hyaluronic acid filler-induced impending necrosis with hyaluronidase: consensus recommendations.,Joel L Cohen,"Injection-induced necrosis is a rare but dreaded consequence of soft tissue augmentation with filler agents. It usually occurs as a result of injection of filler directly into an artery, but can also result from compression or injury. We provide recommendations on the use of hyaluronidase when vascular compromise is suspected. Consensus recommendations were developed by thorough discussion and debate amongst the authors at a roundtable meeting on Wednesday June 18, 2014 in Las Vegas, NV as well as significant ongoing written and verbal communications amongst the authors in the months prior to journal submission. All authors are experienced tertiary care providers. A prompt diagnosis and immediate treatment with high doses of hyaluronidase (at least 200 U) are critically important. It is not felt necessary to do a skin test in cases of impending necrosis. Some experts recommend dilution with saline to increase dispersion or lidocaine to aid vasodilation. Additional hyaluronidase should be injected if improvement is not seen within 60 minutes. A warm compress also aids vasodilation, and massage has been shown to help. Some experts advocate the use of nitroglycerin paste, although this area is controversial. Introducing an oral aspirin regimen should help prevent further clot formation due to vascular compromise. In our experience, patients who are diagnosed promptly and treated within 24 hours will usually have the best outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Recommendations on the use of hyaluronidase when vascular compromise is suspected are provided and patients who are diagnosed promptly and treated within 24 hours will usually have the best outcomes.'}",
-subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient.,Hugo H. Davila,"INTRODUCTION Ischemic priapism (IP) is a urologic condition, which necessitates prompt management. Intracavernosal injection of phenylephrine is a usual treatment modality utilized for the management of these patients. Aim. We present a case of subarachnoid hemorrhage following intracavernosal injection of phenylephrine for IP in a patient with sickle cell disease. METHODS We analyzed the degree of subarachnoid hemorrhage in our patient after intracavernosal injection of phenylephrine. The patient had an acute rise in blood pressure during corporal irrigation. This was followed by the onset of severe headache. Computed tomography (CT) scan confirmed the diagnosis of a subarachnoid hemorrhage. MAIN OUTCOME MEASURE Subarachnoid hemorrhage associated with intracavernosal injection of phenylephrine. Result. A 23-year-old African American male with a history of sickle cell disease presented with a painful penile erection. The patient was started on intravenous fluids, oxygen by nasal canula, and analgesic medication. After this, a blood gas was obtained from his left corpora cavernosa. This was followed by normal saline irrigation and injection of phenylephrine. The patient complained of a sudden, severe ""terrible headache"" immediately following the last injection, and noncontrast CT scan of the head was obtained and a subarachnoid hemorrhage was noted. The patient was admitted for observation and no significant changes were noted. CONCLUSIONS Intracavernosal injection of phenylephrine for the management of IP can be associated with several possible complications. We present our single case complicated with the formation of a subarachnoid hemorrhage. The patient was treated conservatively and had no long-term neurologic sequelae. Davila HH, Parker J, Webster JC, Lockhart JL, and Carrion RE. Subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient.","{'model': 'tldr@v2.0.0', 'text': 'A single case complicated with the formation of a subarachnoid hemorrhage as complication of phenylephrine injection for the treatment of ischemic priapism in a sickle cell disease patient is presented.'}",
-prompt toroidal momentum balance with collisionless neutral beam injected torque in diii-d,J. Degrassie,"In electron cyclotron heated (ECH) H-mode discharges with neutral beam injection (NBI) pulses that are short compared with the fast ion scattering or slowing times, and it is observed that the plasma stores all the angular momentum delivered by the NBI torque impulse. The pulse length is also much shorter than the momentum confinement time of the plasma. Source computations with the Monte Carlo code TRANSP [R. J. Goldston, D. C. McCune, H. H. Towner et al., J. Comput. Phys. 43, 61 (1981)] show that during a pulse approximately 90% of this torque impulse is delivered via the collisionless fast radial current injection process, so that the plasma acquires the balancing toroidal acceleration through ion drift motion in the increasing ∂E∕∂t where E is the electric field normal to the flux surfaces. The measured radial profile of the toroidal momentum increase matches the source, i.e., the computed torque impulse profile. We measure the bulk ion toroidal acceleration in helium discharges, as well as that of th...",,
-increase of serum growth hormone concentration following thyrotropin-releasing hormone injection in patients with acromegaly or gigantism.,Minoru Irie,"ABSTRACT Effect of TRH administration on serum TSH and GH was studied in 9 normal subjects and 11 patients with acromegaly or gigantism. TRH provoked a prompt increase of serum TSH in normal subjects, whereas no appreciable change of serum GH was observed. Serum TSH responses to TRH were lower in most of the patients, but 8 out of the patients showed a remarkable increase of serum GH within 30 min after TRH injection. These results appear to indicate that specificity of TRH is not absolute in the functioning pituitary tumor secreting GH.","{'model': 'tldr@v2.0.0', 'text': 'Results appear to indicate that specificity of TRH is not absolute in the functioning pituitary tumor secreting GH in patients with acromegaly or gigantism.'}",
-global‐scale ulf waves associated with ssc accelerate magnetospheric ultrarelativistic electrons,Y. Hao,"We study electron behavior in the outer radiation belts during the 16 July 2017 storm sudden commencement (SSC), in which prompt intensification of ultrarelativistic electron fluxes was observed at around L = 4.8 by Van Allen Probe B immediately after an interplanetary shock. The electron fluxes in multiple energy channels show clear oscillations in the Pc5 frequency range, although the oscillation characteristics are quite different in different energy channels. At energies above ∼1 MeV, the oscillation periods were very close to the electron drift period, which resembles an energy spectrogram evolution expected for an energetic particle injection event and its drift echoes. At lower energies, however, the oscillation periods hardly depended on the energy: They were very close to the ultralow frequency (ULF) wave period derived from electric field measurements (about 250 s according to wavelet analysis). These complex signatures are consistent with the picture of drift resonance between electrons and short‐lived ULF waves with low azimuthal wave numbers. Good agreement between the observations and numerical simulations confirms that shock‐induced global‐scale ULF waves can efficiently accelerate outer belt ultrarelativistic electrons up to 3.4 MeV over a time scale shorter than 1 hr.",,
-septic facet joint arthritis after a corticosteroid facet injection.,T. Weingarten,Lumbar facet joint injections are commonly employed in the treatment of low back pain and are considered to be relatively safe with few known complications. We report the case of septic facet arthritis following a periarticular facet injection in a patient with recurrent urinary tract infections. The literature is reviewed to identify epidemiological and clinical features of patients in whom septic facet arthritis developed after lumbar facet injection. The diagnosis of iatrogenic septic facet arthritis is often delayed because neurologic and constitutional signs and symptoms develop slowly. Serologic nonspecific markers of infection and appropriate imaging studies may be more sensitive for the early diagnosis of septic facet arthritis. Recalcitrant or worsening back pain after facet injections should prompt an investigation to rule out infectious causes.,"{'model': 'tldr@v2.0.0', 'text': 'Serologic nonspecific markers of infection and appropriate imaging studies may be more sensitive for the early diagnosis of septic facet arthritis.'}",https://academic.oup.com/painmedicine/article-pdf/7/1/52/5262013/7-1-52.pdf
-comparison of intraarticular injection of depot corticosteroid and hyaluronic acid for treatment of degenerative trapeziometacarpal joints,S. Stahl,"Background:Trapeziometacarpal joint arthritis is a disabling condition presenting with pain at the base of the thumb causing impairment of hand function. Nonoperative treatment at an early stage includes intraarticular steroid injection. Although this treatment may bring about prompt symptomatic relief, its efficacy is unpredictable. Objectives:There is previous evidence that injection of sodium hyaluronate is effective and safe in the treatment of knee arthritis. We proposed that intraarticular injection of sodium hyaluronate, for the symptomatic treatment of trapeziometacarpal joint arthritis, could provide symptomatic relief without the adverse effects of steroids. Methods:Fifty-two patients with trapeziometacarpal joint grade II arthritis were randomized prospectively either for methylprednisolone or hyaluronate intraarticular injections. Initial evaluation included an estimation of pain, grip, pinch strengths and the functional Purdue Pegboard Test (PPT). This evaluation was repeated after 1, 3, and 6 months and statistically compared with the initial evaluation. Results:In both groups, the intraarticular injection produced a relief of pain after 1 month. Grip strength improved significantly in the group treated by the steroid during the whole evaluation period. The patients treated by hyaluronate showed improvement in grip strength after 6 months and in the pinch and the PPT after 3 months. Conclusions:Steroids and hyaluronate injections were found effective in reducing pain. Hyaluronate was more effective in the improvement of some aspects of fine hand function.","{'model': 'tldr@v2.0.0', 'text': 'This work proposed that intraarticular injection of sodium hyaluronate, for the symptomatic treatment of trapeziometacarpal joint arthritis, could provide symptomatic relief without the adverse effects of steroids.'}",
-prompt striations in ionospheric barium clouds due to a velocity space instability,D. Simons,The linear theory of electrostatic instabilities of the drift loss-cone type has been developed for a velocity distribution appropriate to the shaped-charge barium injection experiment Buaro. The dispersion relation for the instability in a Vlasov plasma has been solved numerically. The results show a low-frequency instability that we interpret as the source of prompt striations observed in Buaro and cyclotron instabilities that appear to correspond to observed ion cyclotron fluctuations. These instabilities have much faster growth rates than drift instabilities of fluid theories and therefore provide a better description of the observations.,,
-gamma-ray burst spectral correlations: photospheric and injection effects,F. Ryde,"We present a physical framework that can account for most of the observed spectral properties of prompt gamma-ray burst emission. This includes the variety of spectral shapes and shape evolutions, and spectral correlations between flux and spectral peaks within bursts, described by Borgonovo & Ryde, and among bursts described by Amati and Ghirlanda. In our proposed model the spectral peak is given by the photospheric emission from a relativistic outflow for which the horizon length is much smaller that the radial width. The observed duration of the thermal flash is given by the radial light-crossing time. This then gives the typical emission site at ~1011 cm with a Lorentz factor of ~300. This emission is accompanied by nonthermal emission from dissipation locations outside the photosphere. The relative strengths of these two components depend on injection effects at the central engine, leading to varying relative locations of the saturation and photospheric radii. The total emission can then reproduce the observed variety. The spectral correlations are found by assuming that the amount of energy dissipated depends nonlinearly on the averaged particle density. Besides the spectral correlations, this also gives a description of how the relative strength of the thermal component varies with temperature within a burst.",,https://iopscience.iop.org/article/10.1086/508410/pdf
-x-ray plateaus in gamma-ray bursts’ light curves from jets viewed slightly off-axis,P. Beniamini," Using multiple observational arguments, recent work has shown that cosmological gamma-ray bursts (GRBs) are typically viewed at angles within, or close to the cores of their relativistic jets. One of those arguments relied on the lack of tens-of-days-long periods of very shallow evolution that would be seen in the afterglow light curves of GRBs viewed at large angles. Motivated by these results, we consider that GRBs efficiently produce γ-rays only within a narrow region around the core. We show that, on these near-core lines of sight, structured jets naturally produce shallow phases in the X-ray afterglow of GRBs. These plateaus would be seen by a large fraction of observers and would last between 102–105 s. They naturally reproduce the observed distributions of time-scales and luminosities as well as the intercorrelations between plateau duration, plateau luminosity, and prompt γ-ray energy. An advantage of this interpretation is that it involves no late-time energy injection which would be both challenging from the point of view of the central engine and, as we show here, less natural given the observed correlations between plateau and prompt properties.",,https://hal.archives-ouvertes.fr/hal-02393303v1/file/1907.05899.pdf
-compliance with azt treatment regimen of hiv-seropositive injection drug users: a neglected issue.,R. Freeman,"Although treatment with zidovudine (AZT) is now recommended for asymptomatic and symptomatic HIV-infected persons with CD4+ cell counts of 0.20 to 0.50 x 10(9)/L and under, data gathered from a small convenience sample of current and former injection drug users with AIDS in the New York City metropolitan region suggest that noncompliance with HIV/AIDS-related therapeutic regimen may be common in this population. This paper enumerates the reasons for noncompliance offered by these informants, reviews the general literature on treatment compliance to identify additional potential reasons for non-adherence to AZT treatment regimen, and outlines some suggestions for future research into this important issue that may prompt changes in the antiviral delivery system.","{'model': 'tldr@v2.0.0', 'text': 'Although treatment with zidovudine is now recommended for asymptomatic and symptomatic HIV-infected persons with CD4+ cell counts of 0.20 to 0.50 x 10(9)/L and under, data gathered from a small convenience sample of current and former injection drug users with AIDS in the New York City metropolitan region suggest that noncompliance with HIV/AIDS-related therapeutic regimen may be common.'}",
-thoracic intradural aspergillus abscess formation following epidural steroid injection.,G. Saigal,"We report an extremely unusual iatrogenic infection of the spinal canal with Aspergillus fumigatus that resulted in intradural abscess formation following epidural steroid injection in an immunocompetent young individual. Although the imaging findings of the infection were relatively nonspecific, MR imaging not only allowed for a prompt diagnosis, but also helped in surgical localization to the intradural compartment. Complications from the use of these injections are briefly discussed.","{'model': 'tldr@v2.0.0', 'text': 'An extremely unusual iatrogenic infection of the spinal canal with Aspergillus fumigatus that resulted in intradural abscess formation following epidural steroid injection in an immunocompetent young individual is reported.'}",
-observations concerning the injection of a lithium aerosol into the edge of tftr discharges,D. Mansfield,A new method of actively modifying the plasma-wall interaction was tested on the Tokamak Fusion Test Reactor. A laser was used to introduce a directed lithium aerosol into the discharge scrape-off layer. The lithium introduced in this fashion ablated and migrated preferentially to the limiter contact points. This allowed the plasma-wall interaction to be influenced in situ and in real time by external means. Significant improvement in energy confinement and fusion neutron production rate as well as a reduction in the plasma Zeff have been documented in a neutral-beam-heated plasma. The introduction of a metallic aerosol into the plasma edge increased the internal inductance of the plasma column and also resulted in prompt heating of core electrons in Ohmic plasmas. Preliminary evidence also suggests that the introduction of an aerosol leads to both edge poloidal velocity shear and edge electric field shear.,,https://digital.library.unt.edu/ark:/67531/metadc716589/m2/1/high_res_d/769316.pdf
-effect of injection site on circulation times during cardiac arrest.,C. Emerman,"Cardiopulmonary resuscitation requires effective, prompt drug administration. In order to analyze Advanced Cardiac Life Support (ACLS) recommendations for site of drug administration, we studied dye circulation times after central, femoral, and peripheral venous injection during both closed and open chest CPR using a canine arrest model. Measurements of circulation times were made after injection of indocyanine green dye at femoral, central, and peripheral venous sites. Circulation times during closed chest CPR were 62.7 +/- 19.6 sec after central injection, 86.6 +/- 23.5 sec after femoral injection, and 93.6 +/- 17.9 sec after peripheral injection (p less than .001). During closed chest CPR, peak dye concentration after central injection was significantly higher than that after peripheral injection (4.0 +/- 1.3 vs. 3.1 +/- 0.8 mg/L, p less than .01). Circulation times were significantly shorter during open chest CPR with times again shortest after central injection. This animal model suggests that prompt drug delivery during CPR is enhanced by central venous injection of medication. There appears to be no advantage in femoral over peripheral injection.","{'model': 'tldr@v2.0.0', 'text': 'This animal model suggests that prompt drug delivery during CPR is enhanced by central venous injection of medication, and there appears to be no advantage in femoral over peripheral injection.'}",
-endoscopic hemostasis by local injection of epinephrine and polidocanol in bleeding ulcer. a prospective randomized trial,J. Balanzó,"The present prospective randomized trial has been carried out to assess the efficacy of the endoscopic injection of epinephrine plus polidocanol, in the prompt treatment of upper gastrointestinal bleeding due to peptic lesions. Over a period of one year, 72 patients were admitted to the study. Thirty-six of these underwent endoscopic treatment (group A), while the others received medical treatment (group B). Definitive hemostasis was achieved in 29 patients in group A, and in 21 in group B (p less than 0.05). Seven patients in group A and 12 in group B underwent emergency surgery for persistent or recurrent bleeding (p = NS). Three patients from group B received endoscopic treatment owing to relative surgical contraindications. It was successful in all of them. Patients from group A needed a significantly lower number of units of packed red cells (p less than 0.05). Our results show that it is justified to employ endoscopic injection as the first therapeutic step in upper GI bleeding due to peptic gastroduodenal lesions.","{'model': 'tldr@v2.0.0', 'text': 'The results show that it is justified to employ endoscopic injection as the first therapeutic step in upper GI bleeding due to peptic gastroduodenal lesions.'}",
-observation of the reaction h/sup o/ + o/sup 8 +/. -->. h/sup +/(o/sup 7 +/) * during neutral-beam injection into ormak,R. Isler,"Several lines of O VIII have been investigated during the injection of hydrogen atoms at energies of 10, 15, and 30 keV into the Oak Ridge tokamak (ORMAK). The Lyman-..cap alpha.., -..beta.., and -..gamma.. lines and the Balmer-..beta.. line exhibit slowly increasing signals, but the intensity of the Balmer-..cap alpha.. line increases sharply as soon as injection begins. This prompt signal is interpreted to be the direct result of charge transfer from hydrogen to O/sup 7 +/ ions in excited states.",,
-complications of an intra-arterial injection from an inferior alveolar nerve block.,B. Webber,"BACKGROUND Unintended intravascular injections from inferior alveolar nerve blocks result in frustrating complications, both systemically and locally. It is imperative that the dentist diagnose the complication and treat it appropriately. Sometimes, several reactions occur simultaneously. CASE DESCRIPTION The authors present a case report that illustrates some of the many complications resulting from inferior alveolar nerve block injections. In this case, complications developed from intra-arterial injection of local anesthetic. The facial skin, intraoral structures and eye were affected. Within 60 minutes of the injection, all structures returned to their normal state. Prompt diagnosis and reassurance usually calm the patient. CLINICAL IMPLICATIONS Even when clinicians use the utmost care, by aspirating before the injection and noting anatomical landmarks, intra-arterial injections can occur during inferior alveolar nerve blocks. Fortunately, permanent damage to nerves, facial and oral tissues, and eyes is rare. The practitioner should diagnose and treat the problems appropriately to avoid any irreversible complications.","{'model': 'tldr@v2.0.0', 'text': 'In this case, complications developed from intra-arterial injection of local anesthetic from inferior alveolar nerve block injections, resulting in permanent damage to nerves, facial and oral tissues, and eyes.'}",
-physics of confinement improvement of plasmas with impurity injection in diii-d,M. Murakami,"External impurity injection into L mode edge discharges in DIII-D has produced clear confinement improvement (a factor of 2 in energy confinement and neutron emission), reduction in all transport channels (particularly ion thermal diffusivity to the neoclassical level), and simultaneous reduction of long wavelength turbulence. Suppression of the long wavelength turbulence and transport reduction are attributed to synergistic effects of impurity induced enhancement of E × B shearing rate and reduction of toroidal drift wave turbulence growth rate. A prompt reduction of density fluctuations and local transport at the beginning of impurity injection appears to result from an increased gradient of toroidal rotation enhancing the E × B shearing. Transport simulations carried out using the National Transport Code Collaboration demonstration code with a gyro-Landau fluid model, GLF23, indicate that E × B shearing suppression is the dominant transport suppression mechanism.",,
-hepatocyte growth factor promotes liver regeneration with prompt improvement of hyperbilirubinemia in hepatectomized cholestatic rats.,A. Yoshikawa,"BACKGROUND In hepatectomy for patients with liver cirrhosis or cholestasis, prolonged postoperative hyperbilirubinemia is a troublesome complication and, if uncontrolled, often leads to life-threatening hepatic failure. Hepatocyte growth factor (HGF), first identified as the most potent mitogen for primary hepatocytes, has been shown to have multiple biological properties on liver, including mitogenic, antifibrotic, and cytoprotective activities. This study investigated the beneficial effects of a perioperative HGF supply to jaundiced liver after hepatectomy in rats. MATERIALS AND METHODS As a model of jaundiced liver, we used an alpha-naphtylisocyocyanate (ANIT)-induced intrahepatic cholestasis model. Forty-eight hours after intraperitoneal injection of ANIT (75 mg/kg), when the total serum bilirubin level was moderately increased, a 70 % hepatectomy was performed. Human recombinant HGF (250 microgram/kg) (n = 15) or vehicle alone (n = 15) was intermittently administered to the rats 12 h before surgery and every 12 h after that until sacrifice. RESULTS Perioperative HGF treatment effectively accelerated hepatocellular DNA synthesis of cholestatic liver followed by increase in the regenerated liver weight. Moreover, HGF supply promptly improved hyperbilirubinemia within 24 h after surgery. Histological examination revealed that HGF administration attenuated periportal inflammation and formation of bile duct obstructions. Postoperative serum concentrations of tumor necrosis factor-alpha, a representative inflammatory cytokine, were not altered by HGF treatment. CONCLUSIONS Perioperative HGF supply not only promotes liver regeneration but also ameliorates hyperbilirubinemia in hepatectomized cholestatic rats. This mode of HGF treatment may be clinically useful for hepatectomy in patients with cholestasis.","{'model': 'tldr@v2.0.0', 'text': 'Perioperative HGF supply not only promotes liver regeneration but also ameliorates hyperbilirubinemia in hepatectomized cholestatic rats, suggesting this mode of HGF treatment may be clinically useful for liver regeneration in patients withCholestasis.'}",
-nuclear equation of state from observations of short gamma-ray burst remnants,P. Lasky,"The favored progenitor model for short $\ensuremath{\gamma}$-ray bursts (SGRBs) is the merger of two neutron stars that triggers an explosion with a burst of collimated $\ensuremath{\gamma}$-rays. Following the initial prompt emission, some SGRBs exhibit a plateau phase in their X-ray light curves that indicates additional energy injection from a central engine, believed to be a rapidly rotating, highly magnetized neutron star. The collapse of this ``protomagnetar'' to a black hole is likely to be responsible for a steep decay in X-ray flux observed at the end of the plateau. In this paper, we show that these observations can be used to effectively constrain the equation of state of dense matter. In particular, we show that the known distribution of masses in binary neutron star systems, together with fits to the X-ray light curves, provides constraints that exclude the softest and stiffest plausible equations of state. We further illustrate how a future gravitational wave observation with Advanced LIGO/Virgo can place tight constraints on the equation of state, by adding into the picture a measurement of the chirp mass of the SGRB progenitor.",,http://minerva-access.unimelb.edu.au/bitstreams/fac0a5fc-ddd7-577e-8842-c204d11e5797/download
-"hand injuries due to high-pressure injection devices for painting in shipyards: circumstances, management, and outcome in twelve patients.",M. Valentino,"BACKGROUND Injuries due to high-pressure injections are frequently underestimated occupational accidents, which may have disabling outcomes. METHODS The occupational injuries logbooks of two shipyards were examined. Twelve accidents due to the utilization of high-pressure equipment were reported. Data were obtained by reviewing the employers' logbooks and hospital records. RESULTS Contaminants were epoxy paint and stucco, paint solvent, hydraulic and industrial oil. In three cases, the palm was injured and the fingers in nine. In 91% of cases, the accident occurred in the last 2 hr of the work shift. Mishandling (n = 9) and rupture of the high-pressure equipment (n = 3) were the causes. Mean time to medical treatment was 42.5 hr (SD 56.0). Ten workers were permanently disabled. CONCLUSIONS Prompt surgical intervention with debridement and decompression was done in two workers, who healed completely. The ten patients who were not treated immediately eventually experienced a high rate of disability and five of them also suffered amputation.","{'model': 'tldr@v2.0.0', 'text': 'Ten workers were permanently disabled and five of them also suffered amputation after accidents due to the utilization of high-pressure equipment in two shipyards.'}",
-time-resolved four-wave mixing in inas/ingaas quantum-dot amplifiers under electrical injection,P. Borri,"Time-resolved four-wave mixing in an InAs/InGaAs/GaAs electrically pumped quantum-dot amplifier is measured at room temperature for different applied bias currents going from optical absorption to gain of the device. The four-wave mixing signal from 140 fs pulses shows a transition from a delayed photon-echo response in the absorption regime to a prompt free polarization decay in the gain regime. This corresponds to a pronounced reduction of the dephasing time from 250 fs at zero bias to less than 50 fs at the maximum applied current. The four-wave mixing response at transparency of the device shows a composite structure with both photon echo and free-polarization decay. This is a signature of the digital occupation number in quantum dots, resulting at transparency in a signal from dots occupied with either zero or two excitons corresponding to absorption or gain of the dot ground state.",,https://pubs.aip.org/aip/apl/article-pdf/76/11/1380/7813934/1380_1_online.pdf
-cosmic ray and neutrino emission from gamma-ray bursts with a nuclear cascade,D. Biehl,"We discuss neutrino and cosmic-ray emission from Gamma-Ray Bursts (GRBs) with the injection of nuclei, where we take into account that a nuclear cascade from photo-disintegration can fully develop in the source. One of our main objectives is to test if recent results from the IceCube and the Pierre Auger Observatory can be accommodated with the paradigm that GRBs are the sources of Ultra-High Energy Cosmic Rays (UHECRs). While our key results are obtained using an internal shock model, we discuss how the secondary emission from a GRB shell can be interpreted in terms of other astrophysical models. It is demonstrated that the expected neutrino flux from GRBs weakly depends on the injection composition, which implies that prompt neutrinos from GRBs can efficiently test the GRB-UHECR paradigm even if the UHECRs are nuclei. We show that the UHECR spectrum and composition, as measured by the Pierre Auger Observatory, can be self-consistently reproduced in a combined source-propagation model. In an attempt to describe the energy range including the ankle, we find tension with the IceCube bounds from the GRB stacking analyses. In an alternative scenario, where only the UHECRs beyond the ankle originate from GRBs, the requirement for a joint description of cosmic-ray and neutrino observations favors lower luminosities, which does not correspond to the typical expectation from {\gamma}-ray observations.",,https://www.aanda.org/articles/aa/pdf/2018/03/aa31337-17.pdf
-"high‐pressure hand injection injuries caused by dry cleaning solvents: case reports, review of the literature, and treatment guidelines",K. Gutowski,"&NA; A previously unreported subset of high‐pressure injection injuries, namely those involving solvents used in the garment dry cleaning industry, is presented. Dry cleaning solutions contain isoparaffinic hydrocarbons, methoxypropanol, and dichlorofluoroethane. Although these solvents have limited potential for systemic toxicity, severe local toxicity causing tissue necrosis often results in loss of the injured digit. Proper treatment includes prompt surgical exploration, careful débridement and irrigation, intravenous antibiotics, and in selected cases, high‐dose systemic corticosteroids. (Plast. Reconstr. Surg. 111: 174, 2003.)","{'model': 'tldr@v2.0.0', 'text': 'A previously unreported subset of high‐pressure injection injuries, namely those involving solvents used in the garment dry cleaning industry, is presented.'}",
-cardiorespiratory distress after sumatriptan given by injection.,W. Inman,"incompetence in 1985 was admitted in ventricular tachycardia without angina. He had started taking sumatriptan 30 days before admission. Each of the eight injections had been followed by ""hot surges in the throat,"" and four of these were followed by a definite sensation of palpitation (up to 160 beats per minute). He required DC cardioversion and amiodarone to correct the arrhythmia. A previous drug point showed that subcutaneous sumatriptan could produce ST elevation in susceptible subjects,' and a recent study has shown that vasospasm in patients with minor coronary artery stenosis can precipitate ventricular tachycardia and fibrillation.4 Our first patient would almost certainly have died without the prompt arrival of the ambulance service. The timing of her ventricular tachycardia in relation to the sumatriptan injection and the lack of previous or subsequent problems strongly suggest that sumatriptan induced her arrhythmia. The second case is less clear in that there was a previous history of cardiac surgery, but eight separate administrations of the drug produced similar symptoms, which were shown to be due to ventricular tachycardia on hospital admission. Thus we suggest that subcutaneous sumatriptan should be administered with caution and that, ideally, the first dose should be given while the cardiac rhythm is being closely monitored.","{'model': 'tldr@v2.0.0', 'text': 'The second case is less clear in that there was a previous history of cardiac surgery, but eight separate administrations of the drug produced similar symptoms, which were shown to be due to ventricular tachycardia on hospital admission, and it is suggested that subcutaneous sumatriptan should be administered with caution and that the first dose should be given while the cardiac rhythm is being closely monitored.'}",https://europepmc.org/articles/pmc1882911?pdf=render
-changes in sleep in response to intracerebral injection of insulin-like growth factor-1 (ifg-1) in the rat.,F. Obál,"Changes in sleep were studied during 6 hours after intracerebroventricular (ICV) administration of Insulin-like growth factor-1 (IGF-1) or the structurally related insulin. IGF-1 was injected either at dark onset (0.05 or 0.5 microgram) or 6 hours after light onset (0.05, 0.5, or 5.0 microgram). The small dose of IGF-1 consistently, albeit modestly, enhanced NREMS over the 6 hour postinjection period in both the dark and the light cycles (REMS increased only at night). The NREMS-promoting activity vanished when the dose was increased to 0.5 microgram, and 5.0 microgram IGF-1 elicited a marked and prompt suppression in NREMS. Heat-inactivated IGF-1 (0.05 microgram) did not alter sleep. On a molar base, the NREMS-promoting dose of insulin was higher than that of IGF-1. Late (hours 7-17 postinjection) enhancements in EEG slow wave activity during NREMS were observed after 5.0 microgram IGF-1. The results indicate that IGF-1 can promote NREMS and may contribute to the mediation of the effects of GH on sleep. The acute sleep-suppressive activity of the high dose of IGF-1 is attributed to an inhibition of endogenous growth hormone-releasing hormone (GHRH).","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that IGF-1 can promote NREMS and may contribute to the mediation of the effects of GH on sleep, and the acute sleep-suppressive activity of the high dose of IGF-2 is attributed to an inhibition of endogenous growth hormone-releasing hormone.'}",
-dilute povidone-iodine prophylaxis maintains safety while improving patient comfort after intravitreal injections.,M. Peden,"PURPOSE To report the rates of postintravitreal injection (IVT) endophthalmitis with topical conjunctival application of various concentrations of povidone-iodine (PI), including no PI. METHODS Retrospective chart review of patients receiving IVTs performed in a single practice between January 2011 and June 2016. Concentration of PI for all injections was recorded and cases of endophthalmitis identified and reviewed. RESULTS A total of 35,060 IVTs in 1854 patients were included from the 5.5-year period. 29,281 injections were performed with standard 5% PI, 5,460 injections with diluted PI (3,731 with 2.5%, 1,673 with 1.25%, 56 with 0.625%), and 319 IVTs with no PI. Incidence of patient-reported PI sensitivity occurred in 15.9% of patients. Fourteen cases of endophthalmitis were identified: 12 in eyes that received 5% PI, one in an eye that received 1.25% PI, and one in an eye receiving no PI. The incidence of endophthalmitis was 0.04% for 5% PI, 0.02% for dilute PI, and 0.31% for no PI prophylaxis. All cases underwent prompt vitrectomy and had positive cultures for coagulase-negative Staphylococcus. CONCLUSION Application of dilute PI solution to the conjunctiva at the time of IVT is an effective alternative to 5% PI for endophthalmitis prophylaxis in betadine-sensitive patients.","{'model': 'tldr@v2.0.0', 'text': 'Application of dilute PI solution to the conjunctiva at the time of IVT is an effective alternative to 5% PI for endophthalmitis prophylaxis in betadine-sensitive patients.'}",
-transient heat transport studies using laser ablated impurity injection in jet,P. Galli,"Following impurity injection by laser ablation (LA) in JET plasmas, the electron temperature, Te, is observed to drop at a rate that cannot be accounted for by changes in radiated power, ΔPrad. Te starts to drop promptly over a large fraction of the plasma volume, which can be explained by a non-local change in electron heat diffusivity. The change in diffusivity, Δχe, is generally short lived, lasting a few milliseconds, even if ΔPrad can persist for longer times. No clear relation between ΔPrad and the strength of the prompt plasma response to LA can be observed, but only those events with ΔPrad/Prad > 0.5 give rise to detectable Te perturbations. In hot ion H mode plasmas, Δχe is found to be spatially non-uniform and increasing exponentially with the time delay of the LA relative to the onset of the H mode. Examples of non-monotonic Δχe profiles are found in plasmas of lower heating power. Energy transport models combining local and non-local (or strongly non-linear) features are possible candidates for an explanation of these observations.",,
-thyrotropin-releasing hormone: evidence for thyroid response to intravenous injection in man,C. Hollander,"Administration of thyrotropin-releasing hormone to normal subjects causes a prompt rise in plasma thyrotropin concentration, followed by a significant increase in circulating plasma triiodothyronine. These observations may prove to be of value in simultaneously assessing the ability of the pituitary and thyroid glands to respond to their trophic hormones.","{'model': 'tldr@v2.0.0', 'text': 'Observations of the response of the pituitary and thyroid glands to thyrotropin-releasing hormone to normal subjects may prove to be of value in simultaneously assessing the ability of these glands to respond to their trophic hormones.'}",
-oozing type cardiac rupture repaired with percutaneous injection of fibrin-glue into the pericardial space: case report.,H. Murata,"Two patients, a 56-year-old man and an 81-year-old woman who were admitted to hospital because of anteroseptal acute myocardial infarction, were initially treated successfully with direct percutaneous transluminal coronary angioplasty. However, both patients later developed sudden cardiogenic shock due to cardiac tamponade caused by left ventricular free wall rupture (LVFWR). Prompt, life-saving pericardiocentesis was performed, then fibrin-glue was percutaneously injected into the pericardial space. After the procedure, there was no detectable pericardial effusion on echocardiography and the hemodynamic state became stable. The surgical treatment was the standard procedure for LVFWR, but percutaneous fibrin-glue therapy can also be considered for oozing type LVFWR.",,
-the renal response to intravenous injection of sodium chloride solutions in man.,B. Crawford,"The administration of isotonic saline solution to dogs in quantities equal to one-half or more of the volume of extracellular fluid leads to a prompt increase in glomerular filtration rate and the excretion of sodium and chloride. Although factors possibly influencing tubular reabsorptive activity, such as the adrenal and neurohypophysial hormones, are not as yet amenable to quantitative study, the general correlation between filtration rate and sodium excretion in the dog has led many investigators to believe that changes in glomerular activity constitute one of the most important mechanisms in the maintenance of salt and water balance in this animal (1-5). So effective is this renal response that, with other adjunctival mechanisms, dogs may be maintained on diets containing 4 gm. of salt per kg. per day (corresponding to 280 gm. per day in a 70 kg. man) without gain in weight or apparent disturbance in salt and water balance (5). An extensive literature is available (6) on the excretion of sodium, chloride and water in patients receiving saline infusions, but in great part the information contained therein is merely descriptive of the fact of diuresis (or absence of diuresis) and is not correlated with the measurable physiological mechanisms concerned in the excretion of sodium and water. No information is available on the effect of saline solutions on renal function in man, and the present studywas undertaken to obtain information, comparable to that available on the dog, on the changes in filtration rate, renal blood flow and electrolyte excretion after the rapid administration of saline infusions.","{'model': 'tldr@v2.0.0', 'text': 'The present study was undertaken to obtain information, comparable to that available on the dog, on the changes in filtration rate, renal blood flow and electrolyte excretion after the rapid administration of saline infusions.'}",http://www.jci.org/articles/view/102553/files/pdf
-prompt gamma ray spectrometry for in vivo measurement of boron-10 concentration in rabbit brain tissue.,K. Mukai,"Boron-10 concentrations in the brain of live rabbits were measured by prompt gamma ray spectrometry at intervals over a 24-hour period. Boron-10 concentrations in the blood and cerebrospinal fluid (CSF) were also measured. Animals were killed at each interval to obtain brain tissues to measure the boron-10 concentration in the brainstem, cerebral cortex, cerebellar cortex, and basal ganglia, as well as the whole brain. Boron-10 concentrations in the live brain did not differ significantly from those measured in whole brain tissue. Boron-10 concentrations in the blood were much higher than in the brain at each interval after injection. These boron-10 concentrations showed a similar pattern of initial rapid decrease, followed by a more gradual decrease. There was little boron-10 present in the CSF. The brainstem contained a significantly larger concentration of boron-10 than the other tissues. Prompt gamma ray spectrometry has the potential for direct measurement of boron-10 concentrations in the brain of patients undergoing boron neutron capture therapy.",,https://www.jstage.jst.go.jp/article/nmc1959/35/12/35_12_855/_pdf
-high-pressure injection injuries to the hand.,D. Vasilevski,"High-pressure injection injury hides the true extent of the lesions behind an apparent small and harmless puncture of the finger or the hand. Through clinical description, we wish to point out the need for prompt treatment to avoid mutilating and function-threatening complications. We wish to outline the role of the emergency physician who must be aware of the incidence of high-pressure injection injury and become accustomed to early referral to a surgeon, experienced in extensive surgical exploration, removal of foreign bodies, and rehabilitation. The open-wound technique gives the best results. We also point out that failure to refer may become an increasing focus of negligence claims.","{'model': 'tldr@v2.0.0', 'text': 'The role of the emergency physician who must be aware of the incidence of high-pressure injection injury and become accustomed to early referral to a surgeon, experienced in extensive surgical exploration, removal of foreign bodies, and rehabilitation is outlined.'}",
-observationally constraining gravitational wave emission from short gamma-ray burst remnants,P. Lasky,"Observations of short gamma-ray bursts indicate ongoing energy injection following the prompt emission, with the most likely candidate being the birth of a rapidly rotating, highly magnetised neutron star. We utilise X-ray observations of the burst remnant to constrain properties of the nascent neutron star, including its magnetic field-induced ellipticity and the saturation amplitude of various oscillation modes. Moreover, we derive strict upper limits on the gravitational wave emission from these objects by looking only at the X-ray light curve, showing the burst remnants are unlikely to be detected in the near future using ground-based gravitational wave interferometers such as Advanced LIGO.",,https://academic.oup.com/mnras/article-pdf/458/2/1660/18242739/stw435.pdf
-experimental allergic encephalomyelitis after the excision of the injection site of antigen-adjuvant emulsion.,J. Freund," It has been repeatedly shown that when certain antigenic material dissolved or suspended in physiological salt solution is injected into the subcutaneous tissue of animals, the surgical removal of the site of injection within a few minutes after injection does not prevent antibody formation (1). These observations are readily explained by the prompt dissemination of injected material. It was demonstrated by Cannon and Burt (2) that when Staphylococcus aureus or Serratia marcescens are injected into tissue of the ear of the rabbit, these microorganisms can be cultured from the blood, spleen, and liver even if the ear is amputated within a few minutes after injection. Furthermore, McMaster and Kidd (3) have shown that an intracutaneous injection introduces material very promptly into regional lymph nodes. When antigens incorporated in water-in-paraffin oil emulsion with or without mycobacteria are injected into the tissues, the immune response is abundant and long sustained.","{'model': 'tldr@v2.0.0', 'text': None}",https://journals.aai.org/jimmunol/article-pdf/75/6/454/1473851/ji0750060454.pdf
-bench test and preliminary results of vertical compact torus injection experiments on the stor-m tokamak,D. Liu,The University of Saskatchewan compact torus injector has been modified to allow vertical injection of the compact torus (CT) into the STOR-M tokamak. The injector stayed at the horizontal position and the CT trajectory was bent by 90° using a curved conducting drift tube. The curved drift tube did not significantly slow the CT velocity down or change the CT magnetic field topology. Preliminary vertical CT injection experiments have revealed a prompt increase in the line averaged electron density and in the soft x-ray radiation level in the tokamak discharge immediately following vertical CT injection. Suppression of the m = 2 Mirnov oscillation amplitude has also been observed after CT injection.,,
-repeated intravitreous ranibizumab injections for diabetic macular edema and the risk of sustained elevation of intraocular pressure or the need for ocular hypotensive treatment.,S. Bressler,"IMPORTANCE For the management of retinal disease, the use of intravitreous injections of anti-vascular endothelial growth factor has increased. Recent reports have suggested that this therapy may cause sustained elevation of intraocular pressure (IOP) and may potentially increase the risk of glaucoma for patients with retinal disease. OBJECTIVE To assess the risk of sustained IOP elevation or the need for IOP-lowering treatments for eyes with diabetic macular edema following repeated intravitreous injections of ranibizumab. DESIGN, SETTING, AND PARTICIPANTS An exploratory analysis was conducted within a Diabetic Retinopathy Clinical Research Network randomized clinical trial. Study enrollment dates were from March 20, 2007, to December 17, 2008. Of 582 eyes (of 486 participants) with center-involved diabetic macular edema and no preexisting open-angle glaucoma, 260 were randomly assigned to receive a sham injection plus focal/grid laser treatment, and 322 were randomly assigned to receive ranibizumab plus deferred or prompt focal/grid laser treatment. MAIN OUTCOMES AND MEASURES The cumulative probability of sustained IOP elevation, defined as IOP of at least 22 mm Hg and an increase of at least 6 mm Hg from baseline at 2 consecutive visits, or the initiation or augmentation of ocular hypotensive therapy, through 3 years of follow-up. RESULTS The mean (SD) baseline IOP in both treatment groups was 16 (3) mm Hg (range, 5-24 mm Hg). The cumulative probability of sustained IOP elevation or of initiation or augmentation of ocular hypotensive therapy by 3 years, after repeated ranibizumab injections, was 9.5% for the participants who received ranibizumab plus prompt or deferred focal/grid laser treatment vs 3.4% for the participants who received a sham injection plus focal/grid laser treatment (difference, 6.1% [99% CI, -0.2% to 12.3%]; hazard ratio, 2.9 [99% CI, 1.0-7.9]; P = .01). The distribution of IOP and the change in IOP from baseline at each visit through 3 years were similar in each group. CONCLUSIONS AND RELEVANCE In eyes with center-involved diabetic macular edema and no prior open-angle glaucoma, repeated intravitreous injections of ranibizumab may increase the risk of sustained IOP elevation or the need for ocular hypotensive treatment. Clinicians should be aware of this risk and should consider this information when following up with patients who have received intravitreous injections of anti-vascular endothelial growth factor for the treatment of diabetic macular edema.","{'model': 'tldr@v2.0.0', 'text': 'In eyes with center-involved diabetic macular edema and no prior open-angle glaucoma, repeated intravitreous injections of ranibizumab may increase the risk of sustained IOP elevation or the need for ocular hypotensive treatment.'}",https://europepmc.org/articles/pmc4496789?pdf=render
-"the afterglow, energetics, and host galaxy of the short-hard gamma-ray burst 051221a",A. Soderberg,"We present detailed optical, X-ray, and radio observations of the bright afterglow of the short gamma-ray burst 051221a obtained with Gemini, Swift XRT, and the Very Large Array, as well as optical spectra from which we measure the redshift of the burst, z = 0.5464. At this redshift the isotropic-equivalent prompt energy release was about 1.5 × 1051 ergs, and using a standard afterglow synchrotron model, we find that the blast wave kinetic energy is similar, EK,iso ≈ 8.4 × 1051 ergs. An observed jet break at t ≈ 5 days indicates that the opening angle is θj ≈ 7° and the total beaming-corrected energy is therefore ≈2.5 × 1049 ergs, comparable to the values inferred for previous short GRBs. We further show that the burst experienced an episode of energy injection by a factor of 3.4 between t = 1.4 and 3.4 hr, which was accompanied by reverse shock emission in the radio band. This result provides continued evidence that the central engines of short GRBs may be active significantly longer than the duration of the burst and/or produce a wide range of Lorentz factors. Finally, we show that the host galaxy is actively forming stars at a rate of about 1.6 M☉ yr-1, yet exhibits evidence for an appreciable population of old stars (~1 Gyr) and near-solar metallicity. These properties are intermediate between those of long GRB hosts and previous short burst hosts. The lack of bright supernova emission and the low circumburst density (n ~ 10-3 cm-3), however, continue to support the idea that short bursts are not related to massive stellar death. Given that the total energy release is larger than the predicted yield for a neutrino annihilation mechanism, this suggests that magnetohydrodynamic processes may be required to power the burst.",,https://iopscience.iop.org/article/10.1086/506429/pdf
-generation of runaway electrons during the thermal quench in tokamaks,P. Aleynikov,"This work provides a systematic description of electron kinetics during impurity dominated thermal quenches. A Fokker–Planck equation for the hot electrons and a power balance equation for the bulk plasma are solved self-consistently, with impurity radiation as the dominant energy loss mechanism. We find that runaway production is facilitated by heavy injection of impurities up to prompt conversion of the total current into a sub-MeV runaway current. We also find that runaway formation is less efficient in plasmas with high pre-quench temperatures and predict significant radial variation of the runaway seed in such plasmas.",,
-experimental background for prompt treatment with dtpa of 239pu-contaminated wounds.,V. Volf,"Rats were injected intramuscularlywith 239Pu in the form ofnitrate, citrate or DTPA chelate. Differences were observed in the distribution pattern of 239Pu and in the response to DTPA treatment, depending on the 239Pu-compound injected. In general, a 3-hr intravenous infusion of Na,CaDTPA (30-120 pmoles kg-l) was more effective than the rapid intravenous injection. Furthermore, the effect of DTPA was dose-dependent only as far as 239Pu-citrate was concerned. Best effects were, however, achieved by injecting DTPA intramuscularly into the vicinity of the 239P~-deposit. Enhanced absorption of the chelated 238Pu did not result in its increased uptake by the body organs. The practical implications of these findings are discussed.","{'model': 'tldr@v2.0.0', 'text': 'In general, a 3-hr intravenous infusion of Na,CaDTPA (30-120 pmoles kg-l) was more effective than the rapid intravenous injection and the effect of DTPA was dose-dependent only as far as 239Pu-citrate was concerned.'}",
-transient decrease in arm blood flow following accidental intra-arterial injection of propofol into the left brachial artery,J. Brimacombe,"trauma, basic teaching is resuscitation and urgent laparotomy. We believe this case represents a severe vagal response to abdominal insufflation to carbon dioxide. It also emphasises the need for vigilant clinical and mechanical monitoring. Prompt resuscitation, including the early use of atropine, is recommended in all severe vagal reactions. Low-flow peritoneal insufflation should always proceed for some time before high-flow is employed. Consideration should also be given to urgent laparotomy in all cases of cardiac arrest during laparoscopy. REFERENCES","{'model': 'tldr@v2.0.0', 'text': 'This case represents a severe vagal response to abdominal insufflation to carbon dioxide and emphasises the need for vigilant clinical and mechanical monitoring.'}",https://journals.sagepub.com/doi/pdf/10.1177/0310057X9402200310
-prompt stimulation by tsh of thyroid oxidation of thiocyanate.,S. Otaki,"TSH (4 U, iv)1 was given to rats on a low iodine diet 30 min before an ip injection of KSCN labeled with 35S; within 15 min after the SCN injection, thyroid oxidation of thiocyanate was significantly greater than in controls not given TSH. Prompt stimulation of thyroid oxidation of SCN was also observed in Purinafed rats, including animals acutely pretreated with perchlorate. The formation of sulfate by calf thyroid slices incubated in 0.1 mil KSCN was enhanced within 15 min by TSH (50 mU/ ml). In canine thyroid slices, supplementing the medium with dbcAMP also stimulated thiocyanate oxidation. Thiocyanate oxidation by bovine thyroid slices was enhanced by the presence of 0.1 mM H2O2 or a hydrogen peroxide generating system in the medium. Both basal and TSH stimulated oxidation of SCN were inhibited by methimazole, thiols, anaerobiosis and inhibitors of electron transport. The promptness with which TSH augments thyroid oxidation of thiocyanate, as well as other features of this response in vivo and in vit...","{'model': 'tldr@v2.0.0', 'text': 'TSH (4 U, iv)1 was given to rats on a low iodine diet 30 min before an ip injection of KSCN labeled with 35S; within 15 min after the SCN injection, thyroid oxidation of thiocyanate was significantly greater than in controls not given TSH.'}",
-periorbital injectables: understanding and avoiding complications,C. Hwang,"Periorbital rejuvenation with neurotoxins and dermal fillers address several aging changes. Safe and effective results require a thorough understanding of periorbital anatomy, proper injection techniques, and complications of these products. Prompt recognition and treatment of complications can minimize their adverse impacts. Complications can be divided into ischaemic and non-ischaemic effects. Hylauronidase, an enzyme that degrades hyaluronic acid, may improve outcomes after intravascular hyaluronic acid fillers.","{'model': 'tldr@v2.0.0', 'text': 'Hylauronidase, an enzyme that degrades hyaluronic acid, may improve outcomes after intravascular hyaluonic acid fillers.'}",
-a comprehensive analysis of swift xrt data. ii. diverse physical origins of the shallow decay segment,E. Liang,"The origin of the shallow decay segment in Swift XRT light curves remains a puzzle. We analyze the properties of this segment with a sample of 53 long Swift GRBs detected before 2007 February. We show that the distributions of the sample's characteristics are lognormal or normal, and its isotropic X-ray energy (Eiso,X) is linearly correlated with the prompt gamma-ray energy but with a steeper photon spectrum, aside from some X-ray flashes. No significant spectral evolution is observed from this phase to the following phase, and the latter is usually consistent with external-shock models, implying that the shallow decay is also of external-shock origin, likely a refreshed external shock. Within the refreshed-shock model, the data are generally consistent with a roughly constant injection luminosity up to the end of this phase, tb. A positive correlation between Eiso,X and tb also favors this scenario. Among the 13 bursts that have well-sampled optical light curves, six have an optical break around tb and the breaks are consistent with being achromatic. However, the other seven either do not show an optical break or have a break at an epoch different from tb. This raises a concern for the energy injection scenario, suggesting that the optical and X-ray emission may not be the same component, at least for some bursts. There are four significant outliers in the sample, GRBs 060413, 060522, 060607A, and 070110. The shallow decay phase in these bursts is immediately followed by a very steep decay after tb, which is inconsistent with any external-shock model. The optical data for these bursts evolve independently from the X-ray data. These X-ray plateaus likely have an internal origin and demand continuous operation of a long-term central engine. We conclude that the observed shallow decay phase likely has diverse physical origins.",,https://iopscience.iop.org/article/10.1086/521870/pdf
-local injection of hydrocortisone and procaine in osteo-arthritis of the hip joint,V. Leveaux,"The local injection of therapeutic substances to relieve symptoms in osteo-arthritis has long been a controversial method of treatment. Desmarais (1952) reviewed the use of various substances, including procaine lactic acid, acid potassium phosphate, and procaine. He further critically examined the results of his own series of 189 injections into hip or knee joint with alkaline procaine, lactic acid, saline, or needle prick alone. There appeared to be little difference between the results with these various materials, except that procaine gave slightly more pain relief and was probably the substance of choice. The introduction of hydrocortisone (Compound F) stimulated further studies in relation to osteoarthritis of the hip joint. Hollander, Brown, Jessar, and Brown (1951) reported the responses of two patients suffering from unilateral osteo-arthritis of the hip to the injection of 25 mg. hydrocortisone into the region of the affected joint. They stated that ""a prompt alleviation of symptoms and improvement in range of motion was noted in both patients after 24 hours; this persisted for 7 days in one and 13 days in the other"". Two years later Hollander (1953) was able to report on 77 injections into osteoarthritic hips, in which relief was obtained for at least 3 days in 47 per cent. of cases.","{'model': 'tldr@v2.0.0', 'text': 'The local injection of therapeutic substances to relieve symptoms in osteo-arthritis has long been a controversial method of treatment and hydrocortisone (Compound F) stimulated further studies in relation to osteoarthritis of the hip joint.'}",https://ard.bmj.com/content/annrheumdis/15/4/330.full.pdf
-update on hiv/hcv coinfection,V. Soriano,,"{'model': 'tldr@v2.0.0', 'text': 'Issues that should be addressed include the following: (1) rising incidence of acute hepatitis C in men who have sex with men, and expansion/recrudescence of injection drug use in some settings/regions; (2) adverse drug interactions between antiretrovirals and DAA; and (3) high cost of DAA, which may lead many to defer or fail to access appropriate therapy.'}",
-roles for thrombin and fibrin(ogen) in cytokine/chemokine production and macrophage adhesion in vivo.,F. Szaba,"Extravascular coagulation leading to fibrin deposition accompanies many immune and inflammatory responses. Although recognized by pathologists for decades, and probably pathologic under certain conditions, the physiologic functions of extravascular coagulation remain to be fully defined. This study demonstrates that thrombin can activate macrophage adhesion and prompt interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) production in vivo. Peritoneal macrophages were elicited with thioglycollate (TG) and then activated in situ, either by intraperitoneal injection of lipopolysaccharide (LPS) or by injection of antigen into mice bearing antigen-primed T cells. Others previously established that such treatments stimulate macrophage adhesion to the mesothelial lining of the peritoneal cavity. The present study demonstrates that thrombin functions in this process, as macrophage adhesion was suppressed by Refludan, a highly specific thrombin antagonist, and induced by direct peritoneal administration of purified thrombin. Although recent studies established that protease activated receptor 1 (PAR-1) mediates some of thrombin's proinflammatory activities macrophage adhesion occurred normally in PAR-1-deficient mice. However, adhesion was suppressed in fibrin(ogen)-deficient mice, suggesting that fibrin formation stimulates macrophage adhesion in vivo. This study also suggests that fibrin regulates chemokine/cytokine production in vivo, as direct injection of thrombin stimulated peritoneal accumulation of IL-6 and MCP-1 in a fibrin(ogen)-dependent manner. Given that prior studies have clearly established inflammatory roles for PAR-1, thrombin probably has pleiotropic functions during inflammation, stimulating vasodilation and mast cell degranulation via PAR-1, and activating cytokine/chemokine production and macrophage adhesion via fibrin(ogen).","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that thrombin can activate macrophage adhesion and prompt interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) production in vivo, and that fibrin regulates chemokine/cytokine production in vitro and in vivo.'}",https://ashpublications.org/blood/article-pdf/99/3/1053/1680716/1053.pdf
-single extra-amniotic injection of prostaglandin e2 in viscous gel to induce mid-trimester abortion.,I. Mackenzie,"In a preliminary study a single extra-amniotic injection of 1.5 mg of prostaglandin E-2 incorporated into an aqueous viscous gel was given to 24 patients aborted within 24 hours, and the mean induction-abortion interval (plus or minus S.E. of mean) was 13.5 plus or minus 1.5 hours. Vomiting occurred in seven patients, and transient severe uterine cramps, pallor, nausea, and shivering occurred in one patient immediately after injection. Complete abortion occurred in 20patients. A delay in the time taken to abort seemed to be associated with an immediate and rapid rise in uterine tone after the injection which required prompt analgesia; this probably reflected rapid decidual absorption and dissolution of the prostaglandins away from their site of action. The degree of distention of the catheter-retaining balloon did not influence abortion times.","{'model': 'tldr@v2.0.0', 'text': 'A delay in the time taken to abort seemed to be associated with an immediate and rapid rise in uterine tone after the injection which required prompt analgesia; this probably reflected rapid decidual absorption and dissolution of the prostaglandins away from their site of action.'}",https://www.bmj.com/content/bmj/1/5952/240.full.pdf
-needlestick injuries among health care workers of university of alexandria hospitals.,M. I. Hanafi,"The aims of this cross-sectional study were to investigate the prevalence and circumstances of needlestick injury (NSI) among heath care workers at University of Alexandria teaching hospitals and to assess the effectiveness of the existing control measures and standard precautions. Data were obtained by anonymous, self-reporting questionnaire from 645 nurses, physicians and ancillary staff in 2007. Around two-thirds of workers (67.9%) had suffered at least 1 NSI in the last 12 months. High-risk patients (with history of HIV, hepatitis B virus or, hepatitis C virus infection or injecting drug use) were involved in 8.2% of injuries. On evaluating the effectiveness of existing control measures, significant protective factors against NSI were: using devices with safety features (OR 0.41), adherence to infection control guidelines (OR 0.42), training in injection safety (OR 0.14), comfortable room temperature (OR 0.32) and availability of a written protocol for prompt reporting (OR 0.37).","{'model': 'tldr@v2.0.0', 'text': 'On evaluating the effectiveness of existing control measures, significant protective factors against NSI were: using devices with safety features, adherence to infection control guidelines, training in injection safety, and availability of a written protocol for prompt reporting.'}",
-the osmoregulation of vasopressin.,G. Robertson,,"{'model': 'tldr@v2.0.0', 'text': 'New assay methods now have made it possible to begin to characterize osmoregulatory function in a more concrete and comprehensive way and also to use such information to analyze systematically certain clinical disorders of salt and water balance.'}",
-intra-arterial thrombolytic therapy is not a therapeutic option for filler-related central retinal artery occlusion,Yi-chun Chen,"Abstract Cosmetic facial filler-related central retinal artery occlusion (CRAO) is a devastating complication of facial hyaluronic acid (HA) injection and can be managed by intra-arterial thrombolytic therapy (IATT). The authors report on a 20-year-old woman who developed unilateral CRAO due to facial HA injection and who, despite prompt IATT, lost vision. A review of the related literature found 14 other female patients who developed cosmetic facial filler-related CRAO and accepted IATT management. In no case was vision loss clinically improved. IATT is not an effective preventive treatment of dermal filler-associated CRAO. The authors suggest careful preprocedural patient selection to prevent this complication.","{'model': 'tldr@v2.0.0', 'text': 'CRAO is a devastating complication of facial hyaluronic acid injection and can be managed by intra-arterial thrombolytic therapy (IATT), but IATT is not an effective preventive treatment of dermal filler-associated CRAO.'}",
-plasma delta‐9‐tetrahydrocannabinol concentrations and clinical effects after oral and intravenous administration and smoking,A. Ohlsson,,"{'model': 'tldr@v2.0.0', 'text': 'After oral THC, the onset of clinical effects was much slower and lasted longer, but effects occurred at much lower plasma concentrations than after the other two methods of administration, suggesting that brain concentrations were increasing as plasma concentrations decreased.'}",
-prompt striations observed in a barium thermite release at 335 km,D. Simons,"Three barium clouds have been released from a single rocket at altitudes of 335, 443, and 535 km. The releases were by means of barium thermite injection. Adding the different cross-field velocity components of the rocket at each of the releases resulted in barium clouds with different kinetic velocity distributions. The barium cloud with the most peaked velocity distribution striated promptly, in good agreement with theoretical predictions based on a kinetic plasma instability. Details of the data analysis and experiment are discussed.",,
-history of cord blood transplantation,E. Gluckman,,"{'model': 'tldr@v2.0.0', 'text': 'Results of unrelated allogeneic cord blood transplants in malignant and nonmalignant diseases, in adults and children, show that, compared with HLA-matched unrelated BM transplant, cord blood has several advantages, including prompt availability of the transplant, decrease of GVHD and better long-term immune recovery resulting in a similar long- term survival.'}",
-effects of acetylcholine injection into the hypothalamus on the insulin and glucagon release.,K. Ishikawa,"To evaluate the role of cholinergic neurones within the hypothalamus in modulating pancreatic-endocrine function, plasma levels of glucagon, insulin and glucose were measured after microinjection of acetylcholine into the ventromedial hypothalamic nucleus (VMH) of rabbits. Injection of acetylcholine (5 X 10(-8) mol in 2 microliter of saline) into the VMH of unanesthetized, unrestrained animals resulted in a prompt elevation of immunoreactive glucagon and glucose levels without any significant change in immunoreactive insulin level. The suppression of insulin secretion following cholinergic stimulation of the VMH was not counteracted by intravenous infusion of glucose (0.25 g/kg). Bilateral adrenalectomy prevented both hyperglucagonemia and hyperglycemia induced by the acetylcholine injection into the VMH. In adrenalectomized rabbits, exogenous glucose load was followed by a significant rise in the plasma insulin level despite the hypothalamic injection of acetylcholine. However, the response of insulin release to glucose load in these animals was much less than in adrenalectomized, saline-treated, control animals. These data support the conclusion that the action of acetylcholine within the hypothalamus on the pancreatic hormone secretions is mediated to a large part through sympatho-adrenomedullary activity. However, a part of the decreased insulin response to glucose may be mediated by direct innervation of the pancreas.","{'model': 'tldr@v2.0.0', 'text': 'The conclusion that the action of acetylcholine within the hypothalamus on the pancreatic hormone secretions is mediated to a large part through sympatho-adrenomedullary activity is supported, however, a part of the decreased insulin response to glucose may be mediated by direct innervation of the pancreas.'}",
-injected mass deposition thresholds for lithium granule instigated triggering of edge localized modes on east,R. Lunsford,"The ability of an injected lithium granule to promptly trigger an edge localized mode (ELM) has been established in multiple experiments. By horizontally injecting granules ranging in diameter from 200 microns to 1 mm in diameter into the low field side of EAST H-mode discharges we have determined that granules with diameter >600 microns are successful in triggering ELMs more than 95% of the time. It was also demonstrated that below 600 microns the triggering efficiency decreased roughly with granule size. Granules were radially injected from the outer midplane with velocities ~80 m s−1 into EAST upper single null discharges with an ITER like tungsten monoblock divertor. These granules were individually tracked throughout their injection cycle in order to determine their efficacy at triggering an ELM. For those granules of sufficient size, ELM triggering was a prompt response to granule injection. By simulating the granule injection with an experimentally benchmarked neutral gas shielding (NGS) model, the ablatant mass deposition required to promptly trigger an ELM is calculated and the fractional mass deposition is determined.",,
-electromagnetic effects in the analysis of prompt striations,J. L. Sperling,"In a number of barium injection experiments in the ionosphere, field-aligned structures have been observed to develop on the ion-cyclotron time scale. These structures, called prompt striations, have been attributed to plasma jetting across the ambient magnetic field and have previously been analyzed in the electrostatic limit. The electrostatic analysis is reasonable for experiments in which the directed ion kinetic energy per unit volume is much smaller than the magnetic field pressure; however, for higher altitude experiments in which the ambient magnetic field strength is weaker, justification for the electrostatic treatment fails and the full electromagnetic dispersion relation must be examined. It is demonstrated that electromagnetic effects inhibit the growth of the linear instability considered to be the source of prompt striations in the earlier experiments. Consistent with the electrostatic analysis, cold background plasma and thermal spread in the ion loss cone distribution function tend to stabilize the instability in the full electromagnetic treatment. The theory suggests that prompt striations in the magnetosphere can form from ion jets only if the plasma density is sufficiently tenuous so that electromagnetic stabilization is not complete.",,
-an accidental intra-arterial injection of thiopental on the dorsum of the foot--a case report.,Z. Khan,"Since its maiden clinical use decades back, innumerable instances of accidental intra-arterial injection of the drug have been reported mostly in the upper extremities. We report an accidental intra-arterial injection of thiopental into one of the tributaries of the arterial arch of the dorsum of the foot, but timely and prompt treatment with lidocaine and heparin together with leg rising prevented a gangrenous episode of the extremity and resulted in an uneventful recovery. This report emphasizes the importance of extreme caution while cannulating the veins on the dorsum of the foot. Attention should be focused on the patient's reactions, the feeling of an intense pain and withdrawal of the leg during injection so that timely therapeutic measures are undertaken and the sad episode is prevented.","{'model': 'tldr@v2.0.0', 'text': ""The importance of extreme caution while cannulating the veins on the dorsum of the foot is emphasized and attention should be focused on the patient's reactions, the feeling of an intense pain and withdrawal of the leg during injection so that timely therapeutic measures are undertaken and the sad episode is prevented.""}",
-intratympanic injection of steroid for treatment of facial paralysis,F. L. Bryant,"Corticosteroids administered systemically have been used with varying degrees of success in the treatment of idiopathic facial paralysis. There is reason to believe that this drug exerts a beneficial influence based on the work of numerous investigators. On the basis of existing dehiscences along the course of the facial nerve and exposure of the chorda tympani nerve, such abnormalities might well be sites of entrance for insult to these nerves; likewise, they might serve as portals for direct medication. Such medication, a corticosteroid, might thus be introduced intratympanically. Seven earlier cases were so treated with complete recovery in six cases and 75 percent recovery in the seventh. Three more cases are reported, all with good and prompt recovery. No untoward side effects were noted.","{'model': 'tldr@v2.0.0', 'text': 'Seven earlier cases of idiopathic facial paralysis were treated with complete recovery in six cases and 75 percent recovery in the seventh, and three more cases are reported, all with good and prompt recovery.'}",
-"grb radiative efficiencies derived from the swift data: grbs versus xrfs, long versus short",Bing Zhang,"We systematically analyze the prompt emission and the early afterglow data of a sample of 31 GRBs detected by Swift before 2005 September and estimate the GRB radiative efficiency. BAT's narrow band inhibits a precise determination of the GRB spectral parameters, and we have developed a method to estimate these parameters with the hardness ratio information. The shallow decay component commonly existing in early X-ray afterglows, if interpreted as continuous energy injection in the external shock, suggests that the GRB efficiencies previously derived from the late-time X-ray data were not reliable. We calculate two radiative efficiencies using the afterglow kinetic energy EK derived at the putative deceleration time (tdec) and at the break time (tb), when the energy injection phase ends, respectively. At tb XRFs appear to be less efficient than normal GRBs. However, when we analyze the data at tdec, XRFs are found to be as efficient as GRBs. Short GRBs have similar radiative efficiencies to long GRBs despite of their different progenitors. Twenty-two bursts in the sample are identified to have the afterglow cooling frequency below the X-ray band. Assuming ϵe = 0.1, we find ηγ(tb) usually <10% and ηγ(tdec) varying from a few percent to >90%. Nine GRBs in the sample have the afterglow cooling frequency above the X-ray band for a very long time. This suggests a very small ϵB and/or a very low ambient density n.","{'model': 'tldr@v2.0.0', 'text': 'The shallow decay component commonly existing in early X-ray afterglows, if interpreted as continuous energy injection in the external shock, suggests that the GRB efficiencies previously derived from the late-time X-rays were not reliable.'}",https://iopscience.iop.org/article/10.1086/510110/pdf
-characterization and reduction of no during the combustion of biodiesel in a semi-industrial boiler,B. Bazooyar,"This paper aims to characterize and reduce the level of nitrogen monoxide (NO) during the combustion of rapeseed oil methyl ester (ROME) in a semi-industrial boiler. First, the formation of NO is characterized during transitional and steady-state operation of boiler; the influence of combustion pressure, excess air, exhaust gas temperature, spray cone angle, and combustion air swirl angle on the level of NO is evaluated, suitable burner operating points for control of NO are recognized, and contributions of thermal and prompt NO to the total level of NO are obtained. At the next level, the potential of air-staging technique (injection of extra air) in the reduction of NO is studied. Results reveal that the level of NO rises significantly in the post-flame zone 10 cm after the tail of the flame, where only the formation of thermal NO is probable in the chamber. Staged air is able to reduce the level of NO up to 10% without any negative impact on the operation of the boiler. Results also reveal that fuel sp...",,
-a review of current evaluation and management,C. Lemley,"Purpose: To review the current state of evaluation and management of various forms of endophthalmitis. Methods: A review of the literature is included, encompassing endophthalmitis occurring after ocular surgeries, intravitreal injection, trauma, and systemic infection. Based on this review, current principles and techniques for evaluating and treating these forms of endophthalmitis are discussed. Results: Postoperative endophthalmitis after cataract surgery is the most common presentation. Conclusions from the Endophthalmitis Vitrectomy Study (EVS) remain a foundation for management of postcataract surgery endophthalmitis, notably prompt intravitreal antibiotic administration after vitreous sampling, with consideration for pars plana vitrectomy in severe cases. The potential impact of advances since the EVS, such as oral fourth generation fluoroquinolones and new vitrectomy techniques, are also discussed. The management of postcataract endophthalmitis is compared and contrasted to endophthalmitis occurring after other ocular surgeries, intravitreal injection, trauma, and systemic infection. Although some principles remain common, treatment rationales differ with other forms of endophthalmitis based on differing clinical circumstances, such as the virulence of organisms that are frequently encountered. Conclusions: Endophthalmitis is a serious, potentially vision threatening condition which can present in various settings. Prompt recognition and treatment are key in maximizing outcomes. RETINA 27:662–680, 2007","{'model': 'tldr@v2.0.0', 'text': 'Current principles and techniques for evaluating and treating various forms of endophthalmitis and the potential impact of advances since the EVS, such as oral fourth generation fluoroquinolones and new vitrectomy techniques are discussed.'}",
-combination of laser photocoagulation and intravitreal bevacizumab (avastin®) for aggressive zone i retinopathy of prematurity,E. Chung,,"{'model': 'tldr@v2.0.0', 'text': 'The combination of indirect laser photocoagulation and intravitreal bevacizumab injection was well tolerated and induced prompt regression of aggressive zone I ROP.'}",
-solar flare electron spectra at the sun and near the earth,S. Krucker,"We compare hard X-ray (HXR) photon spectra observed by the RHESSI with the spectra of the electrons in the associated solar impulsive particle events observed near 1 AU by the WIND 3D Plasma and Energetic Particle (3DP) instrument. For prompt events, where the inferred injection time at the Sun coincides with the HXR burst, the HXR photon power-law spectral index γ and the in situ observed electron spectral index δ measured above 50 keV show a good linear fit, δ = γ + 0.1(±0.1), with correlation coefficient of 0.83, while for delayed events (inferred injection >10 minutes after the HXR burst) only a weak correlation with a coefficient of 0.43 is seen. The observed relationship for prompt events is inconsistent, however, with both the thin target case, where the escaping electrons come from the X-ray-producing electron population, and the thick target case where some of the accelerated source population escapes to 1 AU and the rest produce the HXRs while losing all their energy to collisions. Furthermore, the derived total number of escaping electrons correlates with the number of electrons required to produce observed X-ray flux but is only about ~0.2% of the number of HXR-producing electrons.",,https://iopscience.iop.org/article/10.1086/519373/pdf
-non surgical coronary artery recanalization in acute transmural myocardial infarction.,D. Mathey,"In 41 consecutive patients with an acute transmural myocardial infarction (AMI) admitted within 3 hours after the onset of symptoms, we tried to recanalize the occluded coronary artery by an intracoronary infusion of streptokinase (SK) (2000 units/min). SK infusion was preceded by (1) an intracoronary injection of 0.5 mg nitroglycerin to rule out coronary artery spasm, (2) an attempt to recanalize the vessel mechanically with a flexible guidewire, and (3) an intracoronary injection of plasminogen (500 units) to increase the efficacy of the subsequent SK infusion. Coronary angiography revealed a total coronary artery occlusion in 39 patients and a subtotal occlusion in two patients. In 30 patients (73%), the occluded coronary artery was successfully recanalized within 1 hour (mean 29 ± 15 minutes), resulting in prompt contrast filling of the previously occluded vessel. An arteriosclerotic stenosis always remained at the site of the occlusion. Nitroglycerin opened the occluded coronary artery in one patient, contrast injection in seven patients and guidewire perforation in four of the 15 patients, in whom it was attempted. In 18 patients the occluded coronary artery was recanalized by intracoronary SK infusion alone. After the initial opening of the occluded coronary artery, subsequent SK infusion markedly reduced the degree of stenosis and visible thrombi disappeared. Clinically, recanalization was associated with significant relief of ischemic chest pain. None of the successfully recanalized patients died, including three patients with cardiogenic shock. Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion. Repeat angiography 7-21 days later revealed a patent coronary artery in 12 of 15 successfully recanalized patients. The left ventricular ejection fraction had significantly improved, from 37","{'model': 'tldr@v2.0.0', 'text': 'Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion.'}",
-high-pressure injection injuries of the hand.,P. Jebson,"High pressure injection trauma to the upper extremity is a rare but potentially limb-threatening injury. The index finger on the non-dominant hand of young male workers is most commonly involved. Injected materials include paint, grease and diesel oil. Prompt recognition and initiation of treatment is necessary if a disastrous outcome is to be avoided. Primary treatment consists of surgical decompression and debridement. Amputation may be necessary. With early aggressive treatment, satisfactory results can be obtained. The use of high-pressure devices in industry has resulted in an unusual but potentially limb-threatening injury to the upper extremity known as the ""high-pressure injection injury."" This injury is a surgical emergency that demands prompt diagnosis and treatment if a disastrous outcome is to be avoided. The purpose of this paper is to present our recent experience and review the clinical features, evaluation and treatment principles.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this paper is to present the recent experience and review the clinical features, evaluation and treatment principles of the ""high-pressure injection injury"" and suggest that with early aggressive treatment, satisfactory results can be obtained.'}",
-high pressure injection injury of the hand.,M. Childress,Attention is called to the importance of prompt and adequate treatment of high pressure injection injury of the hand.,"{'model': 'tldr@v2.0.0', 'text': 'Attention is called to the importance of prompt and adequate treatment of high pressure injection injury of the hand.'}",
-a subunit cytomegalovirus vaccine based on recombinant envelope glycoprotein b and a new adjuvant.,R. F. Pass,"A phase I randomized, double-blind, placebo-controlled trial was done with a cytomegalovirus (CMV) vaccine based on the envelope glycoprotein, gB, combined with a novel adjuvant, MF59. Participants received CMV gB vaccine with MF59 or CMV gB with alum or placebo at 0, 1, and 6 months. A fourth vaccine was given at 12 months to a subgroup. Levels of neutralizing antibody and antibody to gB 2 weeks after the third dose of vaccine exceeded those in seropositive control subjects. the formulation with MF59 was more immunogenic than that with alum. The optimal dose of gB appeared to be between 5 and 30 microg. The fourth dose produced a prompt rise in antibody level. There were no serious adverse events associated with vaccine. Local and systemic reactions were generally mild and, except for pain at the injection site, occurred with similar frequency in recipients of placebo and CMV vaccine.","{'model': 'tldr@v2.0.0', 'text': 'A phase I randomized, double-blind, placebo-controlled trial was done with a cytomegalovirus vaccine based on the envelope glycoprotein, gB, combined with a novel adjuvant, MF59, which was more immunogenic than that with alum.'}",https://academic.oup.com/jid/article-pdf/180/4/970/17996494/180-4-970.pdf
-"cern large hadron collider optics model, measurements, and corrections",R. Tomás,"Optics stability during all phases of operation is crucial for the LHC. Tools and procedures have been developed for rapid checks of beta beating, dispersion, and linear coupling, as well as for prompt optics corrections. Important optics errors during the different phases of the beam commissioning were observed and locally corrected using the segment-by-segment technique. The most relevant corrections at injection have been corroborated with dedicated magnetic measurements.",,http://link.aps.org/pdf/10.1103/PhysRevSTAB.13.121004
-experimental glomerulonephritis in the rat induced by antibodies directed against tubular antigens. v. fixed glomerular antigens in the pathogenesis of heterologous immune complex glomerulonephritis.,B. Vandamme,"In heterologous immune complex glomerulonephritis glomerular deposition of immune complexes occurs immediately after an injection with heterologous antibody directed against antigen, derived from the brush border of the tubules. The injected antibody is thought to combine with circulating Fx1A antigen to form immune complexes which subsequently are deposited in the glomeruli. However, perfusion of rat kidneys in absence of this antigen likewise resulted in prompt localization of immune complexes along the glomerular basement membrane. Further, Fx1A antigen was shown to be present in the capillary wall, especially in the filtration slits and on the cell membrane of epithelial cells. From these findings it was concluded that in this model of glomerulonephritis the deposited immune complexes are formed locally instead of being deposited from the circulation. This concept of ""fixed antigen"" may also be relevant to the pathogenesis of other forms of experimental glomerulonephritis and probably also for human glomerulonephritis.","{'model': 'tldr@v2.0.0', 'text': 'It was concluded that in this model of glomerulonephritis the deposited immune complexes are formed locally instead of being deposited from the circulation.'}",
-endophthalmitis associated with intravitreal anti-vascular endothelial growth factor injections,S. G. Schwartz,,"{'model': 'tldr@v2.0.0', 'text': 'Although endophthalmitis cannot be prevented in all cases, certain risk reduction strategies have been proposed, including the use of an eyelid speculum, povidone iodine, avoidance of needle contact with the eyelid margin or eyelashes, and avoidance of routine post-injection antibiotics.'}",https://link.springer.com/content/pdf/10.1007/s40135-013-0033-1.pdf
-ischemic neuroprotection by trpv1 receptor-induced hypothermia,Mirko Muzzi,,"{'model': 'tldr@v2.0.0', 'text': 'Data indicate that TRPV1 receptor agonists are promising candidates for hypothermic treatment of stroke and provide permanent neuroprotection in animals subjected to transient middle cerebral artery occlusion.'}",https://journals.sagepub.com/doi/pdf/10.1038/jcbfm.2012.36
-priapism: current updates in clinical management,P. Song,"Priapism is a persistent penile erection that continues for hours beyond, or is unrelated to, sexual stimulation. Priapism requires a prompt evaluation and usually requires an emergency management. There are two types of priapism: 1) ischemic (veno-occlusive or low-flow), which is found in 95% of cases, and 2) nonischemic (arterial or high-flow). Stuttering (intermittent or recurrent) priapism is a recurrent form of ischemic priapism. To initiate appropriate management, the physician must decide whether the priapism is ischemic or nonischemic. In the management of an ischemic priapism, resolution should be achieved as promptly as possible. Initial treatment is therapeutic aspiration with or without irrigation of the corpora. If this fails, intracavernous injection of sympathomimetic agents is the next step. Surgical shunts should be performed in cases involving failure of nonsurgical treatment. The first management of a nonischemic priapism should be observation. Selective arterial embolization is recommended for the management of nonischemic priapism in cases that request treatment. The goal of management for stuttering priapism is prevention of future episodes. This article provides a review of recent clinical developments in the medical and surgical management of priapism and an investigation of scientific research activity in this rapidly developing field of study.","{'model': 'tldr@v2.0.0', 'text': 'Recent clinical developments in the medical and surgical management of priapism and an investigation of scientific research activity in this rapidly developing field of study are provided.'}",https://europepmc.org/articles/pmc3866283?pdf=render
-eis microfluidic chips for flow immunoassay and ultrasensitive cholera toxin detection.,M. Chiriacò,"A flow-injection impedimetric immunosensor for the sensitive, direct and label-free detection of cholera toxin is reported. A limit of detection smaller than 10 pM was achieved, a value thousands of times lower than the lethal dose. The developed chips fulfil the requirement of low cost and quick reply of the assay and are expected to enable field screening, prompt diagnosis and medical intervention without the need of specialized personnel and expensive equipment, a perspective of special relevance for use in developing countries. Since the chip layout includes two sensing areas each one with a 2 × 2 sensor array, our biochips can allow statistical or (alternatively) multiplex analysis of biorecognition events between antibodies immobilized on each working electrode and different antigens flowing into the chamber.","{'model': 'tldr@v2.0.0', 'text': 'The developed chips fulfil the requirement of low cost and quick reply of the assay and are expected to enable field screening, prompt diagnosis and medical intervention without the need of specialized personnel and expensive equipment.'}",
-global modeling of storm‐time thermospheric dynamics and electrodynamics,T. Fuller‐Rowell,"Understanding the neutral dynamic and electrodynamic response of the upper atmosphere to geomagnetic storms, and quantifying the balance between prompt penetration and disturbance dynamo effects, are two of the significant challenges facing us today. This paper reviews our understanding of the dynamical and electrodynamic response of the upper atmosphere to storms from a modeling perspective. After injection of momentum and energy at high latitude during a geomagnetic storm, the neutral winds begin to respond almost immediately. The high-latitude wind system evolves quickly by the action of ion drag and the injection of kinetic energy; however, Joule dissipation provides the bulk of the energy source to change the dynamics and electrodynamics globally. Impulsive energy injection at high latitudes drives large-scale gravity waves that propagate globally. The waves transmit pressure gradients initiating a change in the global circulation. Numerical simulations of the coupled thermosphere, ionosphere, plasmasphere, and electrodynamic response to storms indicate that although the wind and waves are dynamic, with significant apparent ""sloshing"" between the hemispheres, the net effect is for an increased equatorward wind. The dynamic changes during a storm provide the conduit for many of the physical processes that ensue in the upper atmosphere. For instance, the increased meridional winds at mid latitudes push plasma parallel to the magnetic field to regions of different composition. The global circulation carries molecular rich air from the lower thermosphere upward and equatorward, changing the ratio of atomic and molecular neutral species, and changing loss rates for the ionosphere. The storm wind system also drives the disturbance dynamo, which through plasma transport modifies the strength and location of the equatorial ionization anomaly peaks. On a global scale, the increased equatorward meridional winds, and the generation of zonal winds at mid latitudes via the Coriolis effects, produce a current system opposing the normal quiet-time Sq current system. At the equator, the storm-time zonal electric fields reduce or reverse the normal upward and downward plasma drift on the dayside and nightside, respectively. In the numerical simulations, on the dayside, the disturbance dynamo appears fairly uniform, whereas at night a stronger local time dependence is apparent with increased upward drift between midnight and dawn. The simulations also indicate the possibility for a rapid dynamo response at the equator, within 2 h of storm onset, before the arrival of the large-scale gravity waves. All these wind-driven processes can result in dramatic ionospheric changes during storms. The disturbance dynamo can combine and interact with the prompt penetration of magnetospheric electric fields to the equator.",,
-stimulatory effects of ghrelin on circulating somatostatin and pancreatic polypeptide levels.,M. Arosio,"Ghrelin, the recently identified endogenous ligand of the GH secretagogue receptor, is a gut-brain peptide with endocrine, orexigenic, and gastrointestinal effects. In rodents it increases circulating gastrin and insulin levels, whereas in man it appears to decrease insulin secretion despite a rise in blood glucose levels. The aim of the present study was to evaluate the effects of ghrelin administration on total circulating somatostatin (SS), pancreatic polypeptide (PP), and gastrin levels compared with those elicited on insulin, glucose, and GH. Eight healthy volunteers of normal weight (four women and four men) were injected with 3.3 microg/kg ghrelin or saline after an overnight fast on 2 different days. Blood was taken every 15 min for 1 h and then every 30 min for 2 h. As expected, ghrelin injection elicited a prompt GH and glucose increase with a peak at 30 min and an insulin decrease with a nadir at 60 min. Gastrin concentrations were not modified, whereas significant rises were observed in both SS (in a biphasic pattern with peaks at 15 and 120 min) and PP (which increased promptly with a peak at 15 min). A significant negative correlation was found between SS (first peak) and insulin changes (r = -0.86; P < 0.01). In conclusion, this study clearly demonstrates that ghrelin stimulates SS and PP release in man. Although the underlying mechanisms and biological significance of these pharmacological effects remain to be elucidated, a causal relationship between the SS increase and the insulin changes may be hypothesized. Finally, these findings strongly support ghrelin's postulated role in linking the endocrine control of energy balance and growth with the regulation of gastrointestinal functions.","{'model': 'tldr@v2.0.0', 'text': ""Ghrelin's postulated role in linking the endocrine control of energy balance and growth with the regulation of gastrointestinal functions is supported, as this study clearly demonstrates that ghrelin stimulates SS and PP release in man.""}",https://academic.oup.com/jcem/article-pdf/88/2/701/9149104/jcem0701.pdf
-intestinal obstruction,J. V. Cooke,"Intestinal obstruction, as a rule, is associated with an increasing amount of non-coagulable nitrogen in the blood. With acute intoxication the rise in non-coagulable nitrogen may be rapid and reach as high as three or even ten times normal. With more chronic intoxication there may be little or no rise in the blood non-coagulable nitrogen. Closed intestinal loops show exactly the same picture, and, when combined with obstruction, may give very high nitrogen readings. Acute proteose intoxication due to injection of a pure proteose will show a prompt rise in blood non-coagulable nitrogen, even an increase of 100 per cent within 3 or 4 hours. These intoxications also show a high blood content of creatinine and urea. The residual or undetermined nitrogen may be very high. A human case of intestinal obstruction with autopsy presents blood findings exactly similar to those observed in many animal experiments. Clinically the non-coagulable nitrogen of the blood may give information of value in intestinal obstruction. A high reading means a grave intoxication, but a low reading may be observed in some fatal cases and gives no assurance that a fatal intoxication may not supervene. The kidneys in practically all these experiments are normal in all respects. It is possible that protein or tissue destruction rather than impaired eliminative function is responsible for the rise in non-coagulable nitrogen of the blood in these acute intoxications. Transfusions of dextrose solutions often benefit intestinal obstruction, and may depress the level of the non-coagulable nitrogen in the blood. Some cases show no change in non-coagulable nitrogen following transfusions and diuresis, and, as a rule, such cases present the most severe intoxication.","{'model': 'tldr@v2.0.0', 'text': 'A human case of intestinal obstruction with autopsy presents blood findings exactly similar to those observed in many animal experiments, and Clinically the non-coagulable nitrogen of the blood may give information of value in intestinal obstruction.'}",
-coccygeal epidural with local anesthetic for catheterization and pain management in the treatment of feline urethral obstruction.,Angela K O'Hearn,"OBJECTIVE To describe the technique for performing a coccygeal epidural injection of local anesthetic to facilitate catheterization in male cats with urethral obstructions using low-dose sedation. SIGNIFICANCE Prompt diagnosis and relief of urethral obstructions is important as many cats may have developed marked metabolic abnormalities at the time of presentation. General anesthesia in these patients may be associated with significant risk for complications. Pain management is also an essential treatment goal, and this technique relieves urethral and penile pain during the unblocking process. CONCLUSION Coccygeal epidurals can be used safely to provide analgesia to the penis and urethra and to the authors' knowledge, is a novel treatment modality to aid in the relief of urethral obstructions in male cats.","{'model': 'tldr@v2.0.0', 'text': ""Coccygeal epidurals can be used safely to provide analgesia to the penis and urethra and to the authors' knowledge, is a novel treatment modality to aid in the relief of urethral obstructions in male cats.""}",
-pharmacologic treatment of anaphylaxis: can the evidence base be strengthened?,F. Simons,"Purpose of reviewTo evaluate the evidence base for the pharmacologic treatment of anaphylaxis. Recent findingsIn this review, we focus on four classes of medications (the alpha/beta-agonist epinephrine (adrenaline), H1-antihistamines, H2-antihistamines, and glucocorticoids) that are used in healthcare settings for the initial treatment of anaphylaxis. Epinephrine and many H1-antihistamines and glucocorticoids were introduced before the era of randomized controlled trials and before the era of evidence-based medicine. In anaphylaxis, no randomized controlled trials that are free from methodological problems and meet current standards have been performed with these medications, or with H2-antihistamines. The evidence base for epinephrine injection is stronger than the evidence base for use of other medications in anaphylaxis. Guidelines unanimously recommend prompt injection of epinephrine as the life-saving first-line medication in anaphylaxis; however, they differ in their recommendations for H1-antihistamines, H2-antihistamines, and glucocorticoids. Epinephrine is the only medication that is universally available for anaphylaxis treatment in healthcare settings worldwide. Paradoxically, it is underused in anaphylaxis treatment. SummaryFor ethical reasons, there should never be a placebo-controlled trial of epinephrine in anaphylaxis. We discuss why the possibility of conducting randomized placebo-controlled trials with H1-antihistamines, H2-antihistamines, and particularly with glucocorticoids in anaphylaxis should be considered in order to improve the evidence base for treatment and guide clinical decision-making. We also describe the precautions that will be needed if randomized controlled trials are conducted in anaphylaxis.","{'model': 'tldr@v2.0.0', 'text': 'Why the possibility of conducting randomized placebo-controlled trials with H1-antihistamines, H2-antiharamines, and particularly with glucocorticoids in anaphylaxis should be considered in order to improve the evidence base for treatment and guide clinical decision-making is discussed.'}",
-viral hepatitis transmission in ambulatory health care settings.,I. Williams,"In the United States, transmission of viral hepatitis from health care-related exposures is uncommon and primarily recognized in the context of outbreaks. Transmission is typically associated with unsafe injection practices, as exemplified by several recent outbreaks that occurred in ambulatory health care settings. To prevent transmission of bloodborne pathogens, health care workers must adhere to standard precautions and follow fundamental infection-control principles, including safe injection practices and appropriate aseptic techniques. These principles and practices need to be made explicit in institutional policies and reinforced through in-service education for all personnel involved in direct patient care, including those in ambulatory care settings. The effectiveness of these measures should be monitored as part of the oversight process. In addition, prompt reporting of suspected health care-related cases coupled with appropriate investigation and improved monitoring of surveillance data are needed to accurately characterize and prevent health care-related transmission of viral hepatitis.","{'model': 'tldr@v2.0.0', 'text': None}",https://academic.oup.com/cid/article-pdf/38/11/1592/5935448/38-11-1592.pdf
-carbon-11-acetate pet imaging in renal disease.,P. Shreve,"UNLABELLED The purpose of this study was to investigate the use of [1-11C]acetate as a metabolic tracer for renal imaging in human subjects. METHODS Eighteen patients underwent dynamic PET imaging of the kidneys after intravenous bolus injection of 10-20 mCi [1-11C]acetate. Time-activity curves of renal parenchyma tracer activity were fitted to a two-compartment model using direct arterial blood sampling for the arterial input function. RESULTS Renal uptake of [1-11C]acetate is prompt and high target-to-background ratios are achieved even in the presence of markedly reduced renal function. Carbon-11-acetate is cleared from the renal parenchyma without any urinary excretion and the rate of clearance is comparable to myocardial clearance rates. Among normal subjects, K1, ranged from 0.653 to 1.37 ml/min-g, and was reduced to as low as 0.363 ml/min-g in severe renal disease (serum creatinine greater than 5 mg/dl), while k2 ranged from 0.114 to 0.166 min-1 among normal subjects and was reduced to as low as 0.053 min-1 in severe renal disease. Kinetic parameters K1 and k2 were both reduced in the presence of intrinsic renal disease or significant renal artery stenosis. Renal cell carcinoma demonstrated similar uptake of [1-11C]acetate, but substantially reduced the rate of clearance compared to normal and diseased non-neoplastic renal tissue, allowing for ready differentiation of renal cell carcinoma from non-neoplastic renal tissue on images acquired beyond 10 min of tracer administration. CONCLUSION Carbon-11-acetate is a promising physiologic tracer for the study of renal disease.","{'model': 'tldr@v2.0.0', 'text': 'Renal uptake of [1-11C]acetate is prompt and high target-to-background ratios are achieved even in the presence of markedly reduced renal function, allowing for ready differentiation of renal cell carcinoma from non-neoplastic renal tissue on images acquired beyond 10 min of tracer administration.'}",
-prevention of lethal and renal toxicity of cis-diamminedichloroplatinum(ii) by induction of metallothionein synthesis without compromising its antitumor activity in mice.,A. Naganuma,"The participation of renal metallothionein (MT) in the toxicity and antitumor activity of cis-diamminedichloroplatinum(II) (cis-DDP) in male mice was examined. Preinduction of MT in the kidney by the s.c. administration of bismuth compounds decreased the lethality and renal and gastrointestinal toxicity caused by a single s.c. injection of cis-DDP. In the present study a correlation between the protective effect of pretreatment with bismuth nitrate against cis-DDP toxicity and the preinduced MT levels in the kidney was observed. Bismuth nitrate pretreatment showed no effect on the antitumor activity of cis-DDP against several transplantable tumors, probably because it induces MT in the kidney but not in tumor tissues. The fact that p.o. preadministration of bismuth subnitrate, an antidiarrheal drug, also depressed the lethal toxicity of cis-DDP is promising for its prompt application in medical attention. Thus, bismuth pretreatment allows higher doses of cis-DDP with no apparent toxicity, resulting in more efficient utilization of this anticancer drug.","{'model': 'tldr@v2.0.0', 'text': 'A correlation between the protective effect of pretreatment with bismuth nitrate against cis-DDP toxicity and the preinduced MT levels in the kidney was observed, resulting in more efficient utilization of this anticancer drug.'}",
-thermal and behavioral effects of lipopolysaccharide and influenza in interleukin-1 beta-deficient mice.,W. Kozak,"This study characterized body temperature (Tb), locomotor activity (Act), and feeding behavior under normal conditions and following injection with lipopolysaccharide (LPS) or inoculation with live influenza virus of transgenic C57/black mice deficient in interleukin-1 beta (IL-1 beta). Tb and Act in freely moving mice were measured by biotelemetry. Mice deficient in IL-1 beta had normal circadian rhythm of Tb but were less active than their control counterparts. Mice injected with LPS (2.5 mg/kg i.p.) responded with a prompt decrease of Tb, which lasted approximately 10 h, followed by a fever in which Tb reached a peak at approximately 24 h postinjection. There was no difference between groups in the early drop of Tb after the LPS; however, the 24-h peak of Tb was lower in IL-1 beta-deficient mice. The anorexic effects of LPS and influenza infection were similar in both groups of mice. In mice given influenza virus (17.5 plaque-forming units, median lethal dose), Tb and Act gradually decreased. The fall of Tb was smaller in the transgenic mice. The mice deficient in IL-1 beta displayed a higher mortality rate due to influenza infection than the control mice. We conclude that deficiency in IL-1 beta results in lower fever following the LPS injection and in impairment of the defense response to infection with influenza.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that deficiency in IL-1 beta results in lower fever following the LPS injection and in impairment of the defense response to infection with influenza.'}",
-green or yellow laser treatment for diabetic macular edema: exploratory assessment within the diabetic retinopathy clinical research network,S. Bressler,"Purpose: Explore differences in green compared with yellow focal/grid laser treatment on functional and anatomical endpoints in eyes with diabetic macular edema. Methods: Data from two randomized clinical trials were evaluated for differences in visual acuity and optical coherence tomography parameters for eyes assigned to sham injection + prompt laser, ranibizumab + prompt laser, or prompt laser only: among subgroups of eyes treated exclusively and electively with either green or yellow laser. Results: In the sham injection + prompt laser group, the mean visual acuity letter score change for eyes receiving green and yellow laser treatment, respectively, was +2.4 ± 14 and +5.1 ± 13 at the 52-week visit (P = 0.06) and +2.4 ± 15 and +6.0 ± 13 at the 104-week visit (P = 0.13), with no corresponding evidence of differences in optical coherence tomography thickness. When comparing wavelength groups in the ranibizumab + prompt laser and prompt laser-only groups, meaningful differences in visual acuity and optical coherence tomography thickness were not detected at 1 year or 2 years. Conclusion: A trend toward improved vision outcome with yellow laser observed in one trial was not corroborated by anatomical outcomes or by the other trial. In this study, without random assignment to different wavelengths controlling for bias and confounding, it is not possible to determine whether one wavelength is better than the other.","{'model': 'tldr@v2.0.0', 'text': 'A trend toward improved vision outcome with yellow laser observed in one trial was not corroborated by anatomical outcomes or by the other trial, and without random assignment to different wavelengths controlling for bias and confounding it is not possible to determine whether one wavelength is better than the other.'}",https://europepmc.org/articles/pmc4126070?pdf=render
-detection of bacteremia in emergency department patients at risk for infective endocarditis using universal 16s rrna primers in a decontaminated polymerase chain reaction assay.,R. Rothman,"Prompt definitive diagnosis of acute bacterial endocarditis in febrile injection drug users (IDUs) remains problematic because of delays associated with blood culture. Rapid detection of bacteremia by polymerase chain reaction (PCR) by use of ""universal"" primers has been hampered by background bacterial contamination. Broad-range eubacterial primers selected from the 16S rRNA gene were used in a PCR assay coupled with a simple pre-PCR decontamination step. All PCR reagents were pretreated with the restriction enzyme AluI, which has multiple digestion sites in the amplicon but none in the primer sets. When 4 different bacterial species were spiked into healthy human blood specimens, the assay identified each pathogen with an analytic sensitivity of 5 bacteria/PCR reaction. A clinical trial with 51 febrile IDUs revealed that PCR had a sensitivity and specificity of 86.7% and 86.9%, respectively, versus blood culture. Importantly, all (8/8) patients with blood culture-positive infective endocarditis were determined to be positive by PCR. This assay provides a promising diagnostic for rapid identification of bacteremia, particularly valuable in acute care settings.","{'model': 'tldr@v2.0.0', 'text': 'This assay provides a promising diagnostic for rapid identification of bacteremia, particularly valuable in acute care settings, with high sensitivity and specificity compared with blood culture.'}",https://academic.oup.com/jid/article-pdf/186/11/1677/18007218/186-11-1677.pdf
-renal response to repetitive exposure to heme proteins: chronic injury induced by an acute insult.,Karl A. Nath,"BACKGROUND Renal diseases are conventionally classified into acute and chronic disorders. We questioned whether acute, reversible, renal insults may be induced to incite a chronic scarring process, employing as an acute insult the glycerol model of heme protein-induced renal injury. METHODS Rats were subjected to weekly injections of hypertonic glycerol for up to six months. Renal function was serially determined, and the effect of such insults on renal histology and renal expression of collagen and fibrogenic cytokines was assessed. RESULTS After the first injection of glycerol, which, expectedly, induced a prompt fall in the glomerular filtration rate (GFR), subsequent injections encountered a remarkable renal resistance in that the fall in GFR was markedly blunted. This resistance to acute decline in renal function in rats subjected to repetitive injections of glycerol was accompanied by less necrosis and apoptosis of renal tubular epithelial cells after such injections. The attenuation in the fall in GFR in response to repetitive exposure to glycerol-induced heme protein injury was maintained for up to six months. A progressive decline in GFR appeared after three months and was accompanied by histologic tubulointerstitial injury, the latter assessed at six months. These kidneys demonstrated up-regulation of collagen I, III, and IV in conjunction with increased expression of the oxidant-inducible, chemotactic cytokine, monocyte chemoattractant protein-1 (MCP-1), and the oxidant-inducible, fibrogenic cytokine, transforming growth factor-beta1 (TGF-beta1). The exposure of the kidney to a single injection of hypertonic glycerol increased the expression of both cytokines some three to five days following this exposure, while the exposure of NRK 49F cells in culture to an iron-dependent model of oxidative stress also increased expression of TGF-beta1 and collagen mRNAs. CONCLUSIONS We conclude that this nephrotoxic insult, repetitively administered, encounters a resistance in the kidney such that the expected fall in GFR does not occur. However, with time, such resistance is accompanied by a decrease in GFR, the latter associated with chronic tubulointerstitial disease. Thus, a long-term cost is exacted, either along with, or as a consequence of, such resistance. We suggest that chronic up-regulation of such oxidant-inducible genes such as TGF-beta1 and MCP-1 contributes to tubulointerstitial disease, and iron-mediated oxidative stress may directly induce TGF-beta1.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that chronic up-regulation of such oxidant-inducible genes such as TGF-beta1 and MCP-1 contributes to tubulointerstitial disease, and iron-mediated oxidative stress may directly induce TGF -beta1.'}",http://www.kidney-international.org/article/S0085253815470012/pdf
-heparin-induced release of extracellular superoxide dismutase to human blood plasma.,K. Karlsson,"Extracellular superoxide dismutase (SOD) has previously been shown to be the major SOD isoenzyme in extracellular fluids. Upon chromatography on heparin-Sepharose it was separated into three fractions: A, without affinity; B, with intermediate affinity; and C, with relatively strong heparin affinity. Intravenous injection of heparin leads to a prompt increase in plasma extracellular-superoxide-dismutase (EC-SOD) activity. Heparin induces no release of EC-SOD from blood cells, nor does it activate EC-SOD in plasma, indicating that the source of the released enzyme is the endothelial-cell surfaces. No distinct saturation could be demonstrated in a dose-response curve up to 200 i.u. of heparin per kg body weight, showing that the releasing potency of heparin is lower for EC-SOD than for previously investigated heparin-released factors. Chromatography of human plasma on heparin-Sepharose shows nearly equal amounts of EC-SOD fractions A, B and C. Heparin induces specifically the release of fraction C. The findings point to the existence of an equilibrium of EC-SOD fraction C between the plasma phase and endothelial-cell surfaces. The major part of EC-SOD in the vasculature seems to be located on endothelial-cell surfaces.","{'model': 'tldr@v2.0.0', 'text': 'The findings point to the existence of an equilibrium of EC-SOD fraction C between the plasma phase and endothelial-cell surfaces, which seems to be located on endothelial -cell surfaces in the vasculature.'}",https://europepmc.org/articles/pmc1147663?pdf=render
-extracellular superoxide dismutase in the vascular system of mammals.,K. Karlsson,"Plasma extracellular superoxide dismutase (EC-SOD) from the pig, cat, rabbit, guinea pig and mouse was found to be heterogeneous with regard to heparin affinity and could be separated into three fractions: A without affinity, B with weak affinity and C with relatively high affinity. Rat plasma EC-SOD was deviant and contained only A and B fractions. There were very large interspecies differences in total plasma EC-SOD activity and in division of the activity between the different fractions. Intravenous injection of heparin resulted in the pig, dog, cat, rabbit, guinea pig and mouse in a prompt increase in the plasma SOD activity. The increase was due to release of EC-SOD C to plasma, most probably from endothelial cell surfaces. In the rat, heparin induced no increase in plasma SOD activity, apparently because of the absence of EC-SOD C in this species. The relative heparin-induced increase in plasma EC-SOD C varied between 2 and 11 in the different species and was distinctly correlated with the heparin affinity of EC-SOD C in the particular species. Apparently the EC-SOD C, present in the vasculature, forms an equilibrium between plasma and endothelium, whereas EC-SOD A and B primarily exist in plasma. The wide diversity of EC-SOD in the vascular system of mammals with regard to total amount, division into fractions and distribution between plasma and endothelium indicates that the pathogenic potential of superoxide radicals in the extracellular space might vary much between species.","{'model': 'tldr@v2.0.0', 'text': 'The wide diversity of EC-SOD in the vascular system of mammals with regard to total amount, division into fractions and distribution between plasma and endothelium indicates that the pathogenic potential of superoxide radicals in the extracellular space might vary much between species.'}",
-gas jet disruption mitigation studies on alcator c-mod,R. Granetz,"Damaging effects of disruptions are a major concern for Alcator C-Mod, ITER and future tokamak reactors. High-pressure noble gas jet injection is a mitigation technique which potentially satisfies the operational requirements of fast response time and reliability, while still being benign to subsequent discharges. Disruption mitigation experiments using an optimized gas jet injection system are being carried out on Alcator C-Mod to study the physics of gas jet penetration into high pressure plasmas, as well as the ability of the gas jet impurities to convert plasma energy into radiation on timescales consistent with C-Mod's fast quench times, and to reduce halo currents given C-Mod's high-current density. The dependence of impurity penetration and effectiveness on noble gas species (He, Ne, Ar, Kr) is also being studied. It is found that the high-pressure neutral gas jet does not penetrate deeply into the C-Mod plasma, and yet prompt core thermal quenches are observed on all gas jet shots. 3D MHD modelling of the disruption physics with NIMROD shows that edge cooling of the plasma triggers fast growing tearing modes which rapidly produce a stochastic region in the core of the plasma and loss of thermal energy. This may explain the apparent effectiveness of the gas jet in C-Mod despite its limited penetration. The higher-Z gases (Ne, Ar, Kr) also proved effective at reducing halo currents and decreasing thermal deposition to the divertor surfaces. In addition, noble gas jet injection proved to be benign for plasma operation with C-Mod's metal (Mo) wall, actually improving the reliability of the startup in the following discharge.",,
-dendrimers to treat rheumatoid arthritis.,X. Bosch,"In comparison with linear polymers, dendrimers' multivalency and nanostructure confer substantial advantages in drug delivery including rapid cell entry, targetability, and easier passage across biological barriers. Previous work has shown that phosphorus-containing dendrimers capped with anionic azabisphosphonate (ABP) end groups prompt anti-inflammatory activation of human monocytes. By using two mouse models of arthritis mimicking human rheumatoid arthritis (RA), Hayder et al. recently demonstrated that intravenous injection of dendrimer ABP inhibits the secretion of proinflammatory cytokines and osteoclastogenesis--two fundamental monocyte-dependent processes of inflammation and bone erosion in RA. While available biological therapies for RA target only one of the cytokines involved in inflammation or bone erosion, dendrimer ABP, by virtue of its double action on both processes in mice, might become a more active and cost-saving alternative for RA patients. This Perspective highlights this important development and the challenges that lie ahead.","{'model': 'tldr@v2.0.0', 'text': 'While available biological therapies for RA target only one of the cytokines involved in inflammation or bone erosion, dendrimer ABP might become a more active and cost-saving alternative for RA patients.'}",
-fall back accretion and energy injections in gamma-ray bursts,Y. Yu,"Intense flares that occur at late times relative to the prompt phase have been observed by the Swift satellite in the X-ray afterglows of gamma-ray bursts (GRBs). Here, we present a detailed analysis on the fall back accretion process to explain the intense flare phase in the very early X-ray afterglow light curves. To reproduce the afterglow at late times, we resort to the external shock by engaging energy injections. By applying our model to GRBs 080810, 081028 and 091029, we show that their X-ray afterglow light curves can be reproduced well. We then apply our model to the ultralong Swift GRB 111209A, which is the longest burst ever observed. The very early X-ray afterglow of GRB 111209A showed many interesting features, such as a significant bump observed at around 2000 s after the Swift/BAT trigger. We assume two constant energy injection processes in our model. These can explain the observed plateau at X-ray wavelength in the relatively early stage (8.0 x 10(3) s) and a second X-ray plateau and optical rebrightening at about 10(5) s. Our analysis supports the scenario that a significant amount of material may fall back towards the central engine after the prompt phase, causing an enhanced and long-lived mass accretion rate powering a Poynting-flux-dominated outflow.",,https://arxiv.org/pdf/1312.0794
-ocular inflammation associated with antivascular endothelial growth factor treatment,H. Fine,"Purpose of review To describe the diagnosis and management of intraocular inflammation following antivascular endothelial growth factor (VEGF) injections. Recent findings Inflammation following intravitreal anti-VEGF injections can cause a dramatic reduction in acuity. Differentiating factors from truly infectious endophthalmitis include a lack of pain, redness, or hypopyon, although none of these factors is diagnostic. A high suspicion of infectious endophthalmitis should trigger a prompt vitreous tap and injection of intravitreal antibiotics. Conversely, if noninfectious endophthalmitis is suspected, close observation with frequent topical steroids is warranted. Summary Most eyes with noninfectious endophthalmitis following anti-VEGF injection recover within 1 month to baseline acuity with topical corticosteroid treatment. Mechanisms hypothesized to explain postinjection inflammation include patient-specific, delivery-specific, and medication-specific factors.","{'model': 'tldr@v2.0.0', 'text': 'Most eyes with noninfectious endophthalmitis following anti-VEGF injection recover within 1 month to baseline acuity with topical corticosteroid treatment, and Mechanisms hypothesized to explain postinjection inflammation include patient-specific, delivery- specific, and medication-specific factors.'}",
-tungsten as target material in fusion devices,D. Naujoks,"Several experiments were conducted in ASDEX Upgrade to prove the suitability of tungsten as a divertor target material under the conditions of a high density and low temperature divertor. The observed fluxes from a tungsten tile into the plasma are low, in keeping with the extremely low sputtering yields. In addition, the very favourable effect of `prompt redeposition' (redeposition during the first gyration) could be confirmed by the experiments. Cooling of the edge region by neon injection seems permissible, i.e. neon impurity sputtering did not increase the eroded fluxes of tungsten. The transport and accumulation behaviour were investigated by means of the laser blow-off technique. No accumulation effects could be observed in ohmic discharges. In discharges with NBI heating but without ICRH, strong accumulation can occur. High heat flux tests were performed on graphite tiles coated with plasma sprayed tungsten, which withstood a thermal load of 15 MW/m2 lasting 2 s as well as 1000 cycles of 10 MW/m2 for 2 s without disabling damage. Owing to the encouraging results, an experiment using a tungsten divertor is planned in ASDEX Upgrade",,
-enhancement of indirect sulphation of limestone by steam addition.,Michael C. Stewart,"The effect of water (H₂O(g)) on in situ SO₂ capture using limestone injection under (FBC) conditions was studied using a thermobalance and tube furnace. The indirect sulphation reaction was found to be greatly enhanced in the presence of H₂O(g). Stoichiometric conversion of samples occurred when sulphated with a synthetic flue gas containing 15% H₂O(g) in under 10 h, which is equivalent to a 45% increase in conversion as compared to sulphation without H₂O(g). Using gas pycnometry and nitrogen adsorption methods, it was shown that limestone samples sulphated in the presence of H₂O(g) undergo increased particle densification without any significant changes to pore area or volume. The microstructural changes and observed increase in conversion were attributed to enhanced solid-state diffusion in CaO/CaSO₄ in the presence of H₂O(g). Given steam has been shown to have such a strong influence on sulphation, whereas it had been previously regarded as inert, may prompt a revisiting of the classically accepted sulphation models and phenomena. These findings also suggest that steam injection may be used to enhance sulfur capture performance in fluidized beds firing low-moisture fuels such as petroleum coke.","{'model': 'tldr@v2.0.0', 'text': 'It was shown that limestone samples sulphated in the presence of H₂O(g) undergo increased particle densification without any significant changes to pore area or volume, and steam injection may be used to enhance sulfur capture performance in fluidized beds firing low-moisture fuels.'}",
-re-186(sn) hedp for treatment of painful osseous metastases: initial clinical experience in 20 patients with hormone-resistant prostate cancer.,H. Maxon,"Rhenium-186(tin) hydroxyethylidene diphosphonate (HEDP) is a new radiopharmaceutical that localizes in areas of osseous metastases in a manner similar to that of standard bone-scanning agents. It also emits beta particles with sufficient energy to be therapeutically useful. A single intravenous injection of about 33 mCi (1,221 MBq) was given to each of 20 elderly patients with advanced skeletal metastases from hormonally resistant prostate cancer. Prompt, significant relief of pain occurred 80% of the time with no significant side effects and only minimal, transient marrow toxicity. Re-186(Sn) HEDP appears to be a useful new agent for the palliation of painful osseous metastases in prostate cancer.","{'model': 'tldr@v2.0.0', 'text': 'Re-186(Sn) HEDP appears to be a useful new agent for the palliation of painful osseous metastases in prostate cancer.'}",
-vaccination against hepatitis b: comparison of three different vaccination schedules.,W. Jilg,"Three different hepatitis B vaccination schedules employing injections at months 0, 1, 2, and 12, at months 0, 1, and 6, or at months 0, 1, and 12 were compared in 89 healthy young adults. Concentrations of antibodies to hepatitis B surface antigen (anti-HBs) after the third injection were dependent on the interval between the second and the third dose; geometric mean titers (GMTs) in the three groups were 53 IU/l, 5,846 IU/l, and 19,912 IU/l, respectively, when the third dose was given at month 2, 6, or 12. Whereas the anti-HBs responses to the third dose at month 6 or 12 were typical booster reactions, the kinetics after a third dose given at month 2 resembled those after only two doses but on a significantly higher level. A fourth dose given at month 12 to the individuals vaccinated at months 0, 1, and 2 led to a prompt anti-HBs response similar in height to the response in those vaccinated at months 0, 1, and 12. Thus, for achieving a high anti-HBs concentration guaranteeing its long-lasting persistence, vaccination at months 0, 1, and 12 seems to be preferable to vaccination at months 0, 1, and 6. For individuals at high risk of hepatitis B infection, vaccination at months 0, 1, 2, and 12 might be considered for obtaining an optimal early seroconversion as well as long-term protection.","{'model': 'tldr@v2.0.0', 'text': 'For achieving a high anti-HBs concentration guaranteeing its long-lasting persistence, vaccination at weeks 0, 1, and 12 seems to be preferable to vaccination at months 0,1, and 6, which might be considered for individuals at high risk of hepatitis B infection.'}",
-evaluation and management of priapism: 2009 update,Yun-Ching Huang,,"{'model': 'tldr@v2.0.0', 'text': 'Stuttering (recurrent) ischemic priapism is a challenging and poorly understood condition; new management strategies currently under investigation may improve the ability to care for men with this condition.'}",https://europepmc.org/articles/pmc3905796?pdf=render
-is gev emission from gamma-ray bursts of external shock origin?,A. Maxham,"Recent observations of Gamma-Ray Bursts (GRBs) by the Fermi Large Area Telescope (LAT) revealed a power-law decay feature of the high-energy emission (above 100 MeV), which led to the suggestion that it originates from an external shock. We analyse four GRBs (080916C, 090510,090902B and 090926A) jointly detected by Fermi LAT and Gamma-ray Burst Monitor (GBM), which have high-quality light curves in both instrument energy bands. Using the MeV prompt emission (GBM) data, we can record the energy output from the central engine as a function of time. Assuming a constant radiative efficiency, we are able to track energy accumulation in the external shock using our internal/external shell model code. By solving for the early evolution of both an adiabatic and a radiative blastwave, we calculate the high-energy emission light curve in the LAT band and compare it with the observed one for each burst. The late time LAT light curves after T 90 can be well fitted by the model. However, due to continuous energy injection into the blastwave during the prompt emission phase, the early external shock emission cannot account for the observed GeV flux level. The high-energy emission during the prompt phase (before T 90 ) is most likely a superposition of a gradually enhancing external shock component and a dominant emission component that is of an internal origin.",,https://academic.oup.com/mnras/article-pdf/415/1/77/3112441/mnras0415-0077.pdf
-long-acting formulations of atypical antipsychotics,P. Chue,,"{'model': 'tldr@v2.0.0', 'text': 'The benefits of an atypical antipsychotic coupled with the assurance of medication delivery in the form of a long-acting injection imply that these novel formulations should be considered in first-episode patients, for whom optimal outcome is frequently compromised by early treatment discontinuation and poor adherence.'}",
-glucagon-like peptide-1 stimulates luteinizing hormone-releasing hormone secretion in a rodent hypothalamic neuronal cell line.,S. A. Beak,"To examine the influence of the putative satiety factor (GLP-1) on the hypothalamo-pituitary-gonadal axis, we used GT1-7 cells as a model of neuronal luteinizing hormone- releasing hormone (LHRH) release. GLP-1 caused a concentration-dependent increase in LHRH release from GT1-7 cells. Specific, saturable GLP-1 binding sites were demonstrated on these cells. The binding of [125I]GLP-1 was time-dependent and consistent with a single binding site (Kd = 0.07+/-0.016 nM; binding capacity = 160+/-11 fmol/mg protein). The specific GLP-1 receptor agonists, exendin-3 and exendin-4, also showed high affinity (Ki = 0.3+/-0.05 and 0.32+/-0.06 nM, respectively) as did the antagonist exendin-(9-39) (Ki = 0.98+/-0.24 nM). At concentrations that increased LHRH release, GLP-1 (0.5-10 nM) also caused an increase in intracellular cAMP in GT1-7 cells (10 nM GLP-1: 7.66+/-0.4 vs. control: 0.23+/-0.02 nmol/mg protein; P < 0.001). Intracerebroventricular injection of GLP-1 at a single concentration (10 microg) produced a prompt increase in the plasma luteinizing hormone concentration in male rats (GLP-1: 1.09+/-0.11 vs. saline: 0.69+/-0.06 ng/ml; P < 0.005). GLP-1 levels in the hypothalami of 48-h-fasted male rats showed a decrease, indicating a possible association of the satiety factor with the low luteinizing hormone levels in animals with a negative energy balance.","{'model': 'tldr@v2.0.0', 'text': 'GLP-1 levels in the hypothalami of 48-h-fasted male rats showed a decrease, indicating a possible association of the satiety factor with the low luteinizing hormone levels in animals with a negative energy balance.'}",http://www.jci.org/articles/view/610/files/pdf
-antisolvent crystallization of roxithromycin and the effect of ultrasound,Minwoo Park,"Antisolvent crystallization was performed to precipitate roxithromycin particles from organic solutions. Roxithromycin was dissolved in acetone at different concentrations and each solution was injected into an aqueous antisolvent leading to prompt particle formation. The effects of various experimental variables (solution injection rate, solution concentration, and temperature) on the particle size of roxithromycin were investigated. In addition to these variables, the effect of ultrasound on the resulting particle size was investigated by changing process parameters such as wave intensity (power output), sonication time, and the moment of ultrasonic application. When the drug solution was rapidly injected into the antisolvent, smaller crystals were obtained. Smaller crystals were obtained when solutions with high drug concentrations were used and also when the crystallization took place at lower temperatures. The particle size decreased with the increasing power output of ultrasound and with the increasing sonication time. It was also found that the ultrasonic wave induced the reduction of the particle size only when the ultrasound was applied to the solution at the initial stage of crystallization.",,
-the dynamics of granuloma formation in experimental visceral leishmaniasis,M. J. Mcelrath,"We have examined the temporal sequence of events leading to the formation of hepatic granulomas after the intravenous injection of L. donovani amastigotes into BALB/c mice. Parasite ingestion by permissive Kupffer cells (KC) occurred promptly, and local KC aggregations were the foci about which granulomas were subsequently formed. Infected KC were recognized by the uptake of colloidal carbon and the expression of the macrophage-specific antigen recognized by F4/80 mAb. Peroxidase- positive granulocytes migrated rapidly and were followed by monocytes and L3T4+ T cells that enclosed the infected KC. Thereafter, Ly-2+ T cells were prominent members of the granulomatous lymphoid population. Parasites multiplied until 4 wk, and then a prompt reduction in infected cells occurred. This was associated with a sharp decline in the L3T4+ T cells of the granulomas and the maintenance of the Ly-2+ subset. In comparison, athymic nu/nu mice developed smaller, more slowly appearing granulomas that contained granulocytes and monocytes and exhibited progressive parasite replication. Upon rechallenge, the entire process was completed in 2 wk, and infected KC in the euthymic mice were never observed. We hypothesize that the effectiveness of the granulomatous response requires the destruction of parasitized host cells (KC), in a lymphokine rich environment. We further suggest that the Ly-2+ T cell serves as an important effector cell in this process, either by direct cytotoxicity or by supporting the cytotoxic potential of other cell types in the granuloma.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesize that the effectiveness of the granulomatous response requires the destruction of parasitized host cells (KC), in a lymphokine rich environment, and suggests that the Ly-2+ T cell serves as an important effector cell in this process, either by direct cytotoxicity or by supporting the cytotoxic potential of other cell types in thegranuloma.'}",https://rupress.org/jem/article-pdf/167/6/1927/1098217/1927.pdf
-stimulation and desensitization of the hypothalamic heat‐sensitive structures by capsaicin in rats,A. Jancsó-Gábor,1. In rats the injection of capsaicin into the pre‐optic area of the anterior hypothalamus produces a prompt fall in body temperature and abolishes shivering. With repeated injections of capsaicin the hypothermic effect gradually diminishes and finally vanishes (local desensitization).,"{'model': 'tldr@v2.0.0', 'text': 'In rats the injection of capsaicin into the pre‐optic area of the anterior hypothalamus produces a prompt fall in body temperature and abolishes shivering, and with repeated injections the hypothermic effect gradually diminishes and finally vanishes (local desensitization).'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1113/jphysiol.1970.sp009130
-"complications from injectable polyacrylamide gel, a new nonbiodegradable soft tissue filler",Snehal P. Amin,"BACKGROUND Polyacrylamide gels, containing a hydrogel composed of polyacrylamide and water, are used for soft tissue augmentation and contour correction. There are no reports of significant complications after injection of this material into the face. OBJECTIVE We report an inflammatory reaction after injection of polyacrylamide gels for zygomatic facial augmentation. METHODS A retrospective chart review of single case is presented. RESULTS An inflammatory reaction at the sites of polyacrylamide gels injection was noted at 1 month after initial injection. Despite two ensuing courses of broad-spectrum antibiotics, the patient presented to us with persistent draining nodules. Intralesional steroid injections resulted in prompt resolution and no recurrence. CONCLUSION Inflammatory reactions have been noted in patients receiving polyacrylamide gels for breast augmentation. Facial polyacrylamide gels injections may also be associated with an inflammatory reaction that responds to intralesional steroids. With increasing availability of a variety of soft tissue fillers, dermatologists should be aware of this delayed complication from polyacrylamide gels.","{'model': 'tldr@v2.0.0', 'text': 'Patients receiving polyacrylamide gels for breast augmentation may also be associated with an inflammatory reaction that responds to intralesional steroids, and dermatologists should be aware of this delayed complication.'}",
-quantitative aerobic and anaerobic bacteriology of infected diabetic feet,F. Sapico,"Quantitative aerobic and anaerobic cultures of deep tissue were performed on amputated infected lower limbs of 13 diabetic patients immediately after surgery. Dissection was made through intact skin distant from any preexisting ulcer. The results were compared with those obtained from: (i) ulcer swabs (pre- and postamputation), (iii) curettage of the ulcer base, and (iii) needle aspiration after normal saline injection. Anaerobic transport media were used for anaerobic cultures before prompt transfer to the anaerobic chamber. A mean of 4.7 bacterial species per specimen was seen (2.3 aerobes, 2.4 anaerobes). Mean log10 growth per gram of tissue was as follows: (i) aerobes plus anaerobes = 6.99, (ii) aerobes = 6.42, and (iii) anaerobes = 7.65. There was poor concordance between the deep tissue culture results and the results from other modalities of culture collection, though the results from other modalities of culture collection, though the results from curettage and saline aspiration were slightly better than those obtained from ulcer swabs. The most frequently isolated organisms were enterococci, anaerobic streptococci, and species of Proteus, Clostridium, and Bacteroides. When anti-microbial therapy is indicated for this patient population, the possibility of the concomitant presence of aerobes as well as anaerobes should be considered.","{'model': 'tldr@v2.0.0', 'text': 'Quantitative aerobic and anaerobic cultures of deep tissue were performed on amputated infected lower limbs of 13 diabetic patients immediately after surgery, with poor concordance between the deep tissue culture results and the results from other modalities of culture collection.'}",https://jcm.asm.org/content/jcm/12/3/413.full.pdf
-hormonal control of gastroesophageal-sphincter strength.,D. Castell,"Abstract Because of the possibility that the esophagus is protected during periods of increased gastric acidity by more forceful contraction of the lower esophageal sphincter, pressures were measured during routine gastric analyses. After betazole hydrochloride was given, sphincter pressure consistently increased, but this rise was not due to the simple presence of acid in the stomach, for sphincter pressure decreased consistently when 0.1 N hydrochloric acid was instilled into the stomach of normal subjects. Conversely, increases in sphincter pressure were noted when alkali was placed into the stomachs of the same subjects. These results suggested a possible role of endogenous gastrin in stimulation of the lower esophageal sphincter. Prompt, marked increases in sphincter pressure after subcutaneous injection of submaximal doses of gastrin-like pentapeptide indicate that regulation of lower esophageal sphincter pressure may be a physiologic action of gastrin.","{'model': 'tldr@v2.0.0', 'text': None}",
-blood flow to the rabbit pancreas with special reference to the islets of langerhans.,N. Lifson,"Nonradioactive microspheres of various sized (mean batch diameters ranging from 6 microns ato 26 microns) were administered to unfasted rabbits under sodium pentobarbital anesthesia by a brief injection into the left ventricle. Flow rate per bead was determined by the reference organ method. After prompt death of the animal, the spheres were located and counted miroscopically (islet vs. nonislet) in fixed, stained, and cleared portions of the pancreas. According to an analysis of the distribution of spheres as a function of bead diameter, 11-23% of the total pancreatic blood flow went directly to the islets and 77-89% to the ""acini"" (nonislet tissues). After retrograde postmortem injection of spheres 6 microns, 9 microns and 11 microns in diameter, practically none reached the islets, whereas after orthograde postmortem injection, they did so in the same proportions as in vivo. These results, supplemented by certain control experiments, support the view that all, or nearly all, efferent islet blood flow goes to the acinar capillaries before leaving the organ. We conclude that the arterial supplies to the rabbit exocrine and endocrine pancreas are in parallel, with most of the flow going to the exocrine portion. However, the flow to the islets is large enough to permit significant local actions of the islet hormones on the exocrine pancreas, in confirmation of the existence of an insuloacinar portal system.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that the arterial supplies to the rabbit exocrine and endocrine pancreas are in parallel, with most of the flow going to the exocrine portion, however, the flow to the islets is large enough to permit significant local actions of the islet hormones on theExocrine Pancreas, in confirmation of the existence of an insuloacinar portal system.'}",
-gamma-hydroxybutyric acid for treatment of opiate withdrawal syndrome,L. Gallimberti,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/npp199345.pdf
-randomized clinical trial of testosterone replacement therapy in hypogonadal men.,A. Conway,"We have compared the pharmacokinetics and pharmacodynamics of the three commonly used testosterone formulations in a prospective, randomized cross-over clinical trial. Plasma free and total testosterone and their ratio (proportion of unbound testosterone), sex hormone-binding globulin (SHBG), oestradiol, LH and FSH were measured in 15 hypogonadal men (nine hyper- and six hypogonadotrophic) who underwent, in a randomized sequence, three treatment periods each separated by an intervening washout period. The treatments were: (i) intramuscular injection of 250 mg mixed testosterone esters at 2-weekly intervals, (ii) oral testosterone undecanoate 120 mg bd, and (iii) subcutaneous testosterone pellets (6 x 100 mg). Pellet implantation gave the most prolonged effect with free and total testosterone levels being elevated for up to 4 months. This was accompanied by prompt and sustained suppression of plasma LH and FSH, an increase in plasma levels of oestradiol but no change in SHBG levels. In contrast, intramuscular injections induced marked but reproducible week-to-week fluctuations in free and total testosterone, which resulted in a small decrease in plasma SHBG levels, less marked suppression of LH and FSH and a smaller increase in plasma levels of oestradiol. Oral testosterone undecanoate produced the most variable plasma levels of free and total testosterone with a peak in the first treatment week and a fall thereafter and, despite maintenance of testosterone levels within the physiological range, there was no significant suppression of plasma levels of LH and FSH, and oestradiol levels were unchanged but levels of SHBG and total cholesterol were decreased. Free testosterone levels were increased disproportionately during testosterone treatment as the proportion of unbound testosterone was increased by all three treatments. All three testosterone preparations lowered plasma levels of urea and all were without biochemical or haematological toxicity. Reported sexual function was better maintained and side-effects were fewer with parenteral compared with oral treatments. The marked decrease in SHBG and cholesterol levels during oral testosterone undecanoate, when compared with parenteral treatments, occurred despite lesser androgenic effects (suppression of gonadotrophin levels and reported sexual function), which suggests that the liver is exposed to excessive androgenic load via the portal vein during oral treatment with testosterone esters. It is concluded that testosterone pellets give the closest approximation to zero-order (steady-state) delivery conditions for up to 4 months after a single insertion.(ABSTRACT TRUNCATED AT 400 WORDS)","{'model': 'tldr@v2.0.0', 'text': 'Testosterone pellets give the closest approximation to zero-order delivery conditions for up to 4 months after a single insertion, which suggests that the liver is exposed to excessive androgenic load via the portal vein during oral treatment with testosterone esters.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1365-2605.1988.tb00999.x
-effects of alpha-human atrial natriuretic peptide in essential hypertension.,M.,"Because there is little published information on the effects of atrial peptides in hypertensive humans, 100 micrograms of alpha-human atrial natriuretic peptide was injected intravenously into six patients with essential hypertension in a double-blind, placebo-controlled study under standardized conditions of body posture and dietary sodium and potassium intake. The peptide increased urine sodium excretion sixfold in the first 30 minutes. Smaller increments occurred in urine volume and in calcium, magnesium, and phosphorus excretion; the rise in urine potassium concentration was not statistically significant. Most of these indices returned to time-matched placebo values within 1 hour, but urine sodium excretion remained high for 2 1/2 hours. Arterial pressure fell within 2 minutes of alpha-human atrial natriuretic peptide injection, then returned to matching placebo levels by 10 minutes. Conversely, heart rate increased rapidly and remained elevated for 3 hours. The peptide induced a prompt, brief rise in plasma norepinephrine concentration and a more sustained fall in epinephrine and aldosterone levels, but it did not affect plasma renin activity or cortisol concentration. Compared with normotensive volunteers studied previously under the same conditions, the hypertensive subjects had a greater response in urine volume and sodium, calcium, and magnesium excretion but a less sustained fall in arterial pressure.","{'model': 'tldr@v2.0.0', 'text': 'Compared with normotensive volunteers studied previously under the same conditions, the hypertensive subjects had a greater response in urine volume and sodium, calcium, and magnesium excretion but a less sustained fall in arterial pressure.'}",https://www.ahajournals.org/doi/pdf/10.1161/01.HYP.7.5.812
-autoimmunity caused by disruption of central t cell tolerance. a murine model of drug-induced lupus.,A. Kretz-Rommel,"A side effect of therapy with procainamide and numerous other medications is a lupus-like syndrome characterized by autoantibodies directed against denatured DNA and the (H2A-H2B)-DNA subunit of chromatin. We tested the possibility that an effect of lupus-inducing drugs on central T cell tolerance underlies these phenomena. Two intrathymic injections of procainamide-hydroxylamine (PAHA), a reactive metabolite of procainamide, resulted in prompt production of IgM antidenatured DNA antibodies in C57BL/6xDBA/2 F1 mice. Subsequently, IgG antichromatin antibodies began to appear in the serum 3 wk after the second injection and were sustained for several months. Specificity, inhibition and blocking studies demonstrated that the PAHA-induced antibodies showed remarkable specificity to the (H2A-H2B)-DNA complex. No evidence for polyclonal B cell activation could be detected based on enumeration of Ig-secreting B cells and serum Ig levels, suggesting that a clonally restricted autoimmune response was induced by intrathymic PAHA. The IgG isotype of the antichromatin antibodies indicated involvement of T cell help, and proliferative responses of splenocytes to oligonucleosomes increased up to 100-fold. As little as 5 microM PAHA led to a 10-fold T cell proliferative response to chromatin in short term organ culture of neonatal thymi. We suggest that PAHA interferes with self-tolerance mechanisms accompanying T cell maturation in the thymus, resulting in the emergence of chromatin-reactive T cells followed by humoral autoimmunity.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that PAHA interferes with self-tolerance mechanisms accompanying T cell maturation in the thymus, resulting in the emergence of chromatin-reactive T cells followed by humoral autoimmunity.'}",http://www.jci.org/articles/view/119356/files/pdf
-uveitis following intravitreal bevacizumab: a non-infectious cluster.,Christine N. Kay,"BACKGROUND AND OBJECTIVE In this retrospective case series, the authors report seven cases of bevacizumab-related uveitis that occurred within a 4-month period. PATIENTS AND METHODS Seven eyes of six patients developed non-infectious uveitis following bevacizumab intravitreal injections in a cohort of 978 consecutive bevacizumab injections. RESULTS The mean age of patients was 74.6 years (range: 26 to 92). All patients developed symptom onset within 1 day of injection. Shared signs and symptoms included corneal edema, anterior chamber and vitreous cell, conjunctival injection, ocular pain, and lack of hypopyon. In all patients, visual acuity returned to within one line of baseline acuity. All seven eyes had been previously injected with bevacizumab, with a mean number of antecedent injections of 6.1 (range: 3 to 12). CONCLUSION A cluster of sterile bevacizumab-related uveitic reactions was described in this case series. Acute onset of symptoms, absence of hypopyon, a predominant anterior segment reaction, and prompt improvement on topical steroid therapy are useful clinical features distinguishing this uveitic syndrome from infectious endophthalmitis.","{'model': 'tldr@v2.0.0', 'text': 'Acute onset of symptoms, absence of hypopyon, a predominant anterior segment reaction, and prompt improvement on topical steroid therapy are useful clinical features distinguishing this uveitic syndrome from infectious endophthalmitis.'}",
-selective targeting of boronophenylalanine to melanoma in balb/c mice for neutron capture therapy.,J. Coderre,"Melanoma cells actively accumulate aromatic amino acids for use as precursors in the synthesis of the pigment melanin. Using the Harding-Passey melanoma carried s.c. in BALB/c mice, we have demonstrated that p-boronophenylalanine (BPA) is taken up by melanoma tissue to a much greater extent than by normal tissues. Following a single i.p. injection, or a series of injections given over 1 h, the accumulation of boron in melanoma was found to be transient, reaching a maximum approximately 6 h postinjection. The concentrations of boron achieved in tumor ranged from 9-33 micrograms/g, and are within the range estimated to be necessary for successful application of the nuclear reaction 10B(n,alpha)7Li for neutron capture therapy. Boron concentrations in tumor and tissues were determined using either a prompt-gamma spectroscopic technique or by quantitative neutron capture radiography using whole-body sections. Distribution studies with the resolved stereoisomers of BPA indicated that the L isomer is preferentially accumulated in the melanoma compared to the D isomer. The L isomer of BPA was shown to be targeted to actively dividing tumor cells by simultaneously comparing the boron and [3H]thymidine distribution in tumor. Under conditions which selectively deliver high concentrations of boron to Harding-Passey melanomas in BALB/c mice, BPA did not deliver useful concentrations of boron to a mammary adenocarcinoma in Hale-Stoner mice. These results, along with the selectivity of the Harding-Passey melanoma for the L isomer of BPA, are consistent with our working hypothesis that BPA is actively transported into the melanomas as an analogue of natural melanin precursors.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that p-boronophenylalanine (BPA) is taken up by melanoma tissue to a much greater extent than by normal tissues and the L isomer of BPA was shown to be targeted to actively dividing tumor cells by simultaneously comparing the boron and [3H]thymidine distribution in tumor.'}",
-endophthalmitis following intraocular lens implantation: report of 30 cases and review of the literature.,D. Weber,"Bacterial endophthalmitis is a postoperative complication of intraocular lens implantation. A review of 30 cases showed that 77% occurred within seven days of initial cataract surgery and that all cases occurred within 32 days. Common presenting symptoms and signs included pain localized to the involved eye, decreased visual acuity, conjunctival injection, anterior chamber inflammation, hypopyon, and absent or poor red reflex. Diagnosis of an infectious etiology was made by aqueous and vitreous sampling for gram stain and culture. The commonest bacterial agents were coagulase-negative staphylococci and Staphylococcus aureus, but a wide variety of gram-negative and -positive organisms were also isolated. On the basis of this series and review of the literature, optimal treatment includes prompt diagnosis by culture and gram stain of vitreous fluid, frequent application of topical antibiotics, and administration of intraocular antibiotics. Intravenous antibiotics and vitrectomy may improve the visual outcome. Removal of the intraocular lens is unnecessary for bacteriologic cure and is not associated with improved visual outcome.",,
-"efficacy of nerve root versus interspinous injections of glucocorticoids in the treatment of disk-related sciatica. a pilot, prospective, randomized, double-blind study.",I. Kolsi,"STUDY OBJECTIVES Pilot study comparing the short-term efficacy on pain and functional impairment of nerve root sheath versus interspinous glucocorticoid injections in patients admitted to a French rheumatology department for disk-related sciatica or femoral neuralgia. PATIENTS AND METHODS Thirty patients with refractory nerve root pain (sciatica, n = 29; femoral neuralgia, n = 1) for a mean of four months were randomized to nerve root injection (n = 17) or interspinous injection (n = 13) of the same mixture of 0.10 g of lidocaine hydrochloride and 3.75 mg of cortivazol. Both injection methods were performed under analgesia and benzodiazepine sedation to maintain double blinding. Each patient was evaluated daily during the first seven days of bed rest in the hospital, then after discharge on postinjection day 28. RESULTS Prompt pain relief was obtained in both groups. On day 1, the mean pain scale score (0-100) fell from 70 +/- 3.9 to 26 +/- 5.6 in the nerve root group and from 63 +/- 4 to 23 +/- 4.7 in the interspinous group. These results were sustained on D7 and D28. CONCLUSIONS The unusually high level of efficacy of glucocorticoid injection in our study may be ascribable in part to strong placebo and Hawthorne effects and in part to the intrinsic effects of the injections. Whether nerve root injection is superior over interspinous injection remains unproven.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study comparing the short-term efficacy on pain and functional impairment of nerve root sheath versus interspinous glucocorticoid injections in patients admitted to a French rheumatology department for disk-related sciatica or femoral neuralgia finds an unusually high level of efficacy.'}",
-a risk-benefit assessment of intra-articular corticosteroids in rheumatic disorders,J. Hunter,,"{'model': 'tldr@v2.0.0', 'text': 'Intra-articular corticosteroids form an important part of the management of inflammatory joint disease and might be considered where an inflammatory element occurs in osteoarthritis, but should be seen as an adjunct to other forms of symptom relief.'}",
-extravasation injury in the perioperative setting.,W. Schummer,"Extravasation is an unintentional injection or leakage of fluid in the perivascular or subcutaneous space. Extravasation injury results from a combination of factors, including solution cytotoxicity, osmolality, vasoconstrictor properties, infusion pressure, regional anatomical peculiarities, and other patient factors. We reviewed the hospital files of patients who had sustained a significant extravasation injury in the perioperative setting at two German hospitals. These cases highlight the risk of devastating consequences from extravasation injury. Vasoactive drugs and hyperosmolar and concentrated electrolyte solutions are the predominant vesicants in the perioperative setting. Prompt and appropriate intervention is important for avoiding or minimizing extensive tissue injury.","{'model': 'tldr@v2.0.0', 'text': 'This work reviewed the hospital files of patients who had sustained a significant extravasation injury in the perioperative setting at two German hospitals and highlighted the risk of devastating consequences from extravasations injury.'}",
-autonomic activation in experimental trigeminal pain,A. Frese,"Autonomic activation ipsilateral and simultaneously to unilateral trigeminal pain is a hallmark of a group of primary headache syndromes. The most common of these syndromes is cluster headache, which is characterized by strictly unilateral autonomic symptoms such as ptosis, miosis, lacrimation, conjunctival injection, rhinorrhoea, and nasal congestion during the extreme pain attack, indicating parasympathetic hyperactivity. In fact, the autonomic disturbances are so prominent that this group of headache syndromes has been coined trigeminoautonomic cephalgias (TAC) (1). However, it seems frequently overlooked that other primary headache and facial syndromes may exhibit autonomic symptoms as well (2). Most recently, autonomic activation during severe unilateral migraine attacks (3) and first division trigeminal neuralgia (4) have been described. These descriptions prompt us to report a clinical observation made during recent and ongoing studies in experimental head pain.","{'model': 'tldr@v2.0.0', 'text': 'Clinical observation made during recent and ongoing studies in experimental head pain of autonomic activation during severe unilateral migraine attacks and first division trigeminal neuralgia are reported.'}",https://journals.sagepub.com/doi/pdf/10.1046/j.1468-2982.2003.00462.x
-prolonged biphasic response of plasma testosterone to single intramuscular injections of human chorionic gonadotropin.,RUBEN S. Padrón,"The response of plasma testosterone to varying doses of hCG (0--6000 IU) given as a single im injection has been evaluated in normal men. After an initial rise at 2 h, the levels of testosterone demonstrated a secondary rise, reaching a peak 48 h after the im injection. The magnitude of the response varied directly with the dose of hCG used, and at the highest dose (6000 IU) testosterone levels were still elevated 6 days after administration. Plasma estradiol levels showed a dose-dependent rise, with peak levels being attained 24 h after hCG. The prolonged response of plasma testosterone to a single injection of hCG should prompt a reevaluation of diagnostic and therapeutic regimens using this agent.","{'model': 'tldr@v2.0.0', 'text': 'The prolonged response of plasma testosterone to a single injection of hCG should prompt a reevaluation of diagnostic and therapeutic regimens using this agent.'}",https://academic.oup.com/jcem/article-pdf/50/6/1100/10514661/jcem1100.pdf
-ocean physical and biogeochemical responses to the passage of typhoon meari in the east china sea observed from argo float and multiplatform satellites,E. Siswanto,"We elucidated ocean physical and biogeochemical responses to slow‐moving Typhoon Meari using a new method combining Argo float and satellite observations. Meari‐driven upwelling brought colder, nutrient‐rich deep water to the surface layer, causing sea surface cooling (3–6°C) and threefold enhancement of primary production (PP). Maximum surface cooling (and hence nutrient injection) and peak PP enhancement lagged Meari's passage by 1 and 3 days, respectively, implying that remarkable PP enhancement was attributed to new production (NP). This NP accounted for approximately 3.8% of annual carbon export in the East China Sea (ECS) outer shelf, suggesting that typhoon‐driven upwelling is important for biogeochemical processes in the ECS. Given the wide coverage of Argo float and satellite data, our new approach may prompt comparative studies in other basins and advance the understanding of the role of tropical cyclones in the global ocean biogeochemical cycle.",,https://agupubs.onlinelibrary.wiley.com/doi/pdfdirect/10.1029/2008GL035040
-"studies on alcohol diuresis. i. the effect of ethyl alcohol ingestion on water, electrolyte and acid-base metabolism.",M. Rubini,"The diuresis following the ingestion of alcohol has been compared with the increased urine flow secondary to water ingestion (1). Van Dyke and Ames (2) demonstrated that the injection of small amounts of alcohol (12 to 50 mg. per Kg.) into the carotid artery of normal unanesthetized dogs evoked a prompt diuresis without any detectable alcohol in the systemic venous blood. No increased diuresis occurred in dogs with diabetes insipidus following intracarotid injection. As alcohol has no direct effect on exogenous antidiuretic hormone (ADH), or on the responsiveness of the renal tubules to this hormone (3), most investigators agree that ""alcohol diuresis"" is probably due to suppression of the release of ADH. Little attempt has been made, however, to correlate simultaneously its effect on water excretion with electrolyte excretion, acid-base changes, and al-terations in blood volume. The study was divided into three parts: Part I. Effect of alcohol ingestion on water, electrolyte, and acid-base metabolism in semi-recumbent hydrated normal subjects; Part II. Evaluation of the inhibition of ADH secretion by alcohol, in normal subjects in whom acute changes of extracellular fluid tonicity and ""effective circulating blood volume"" were induced; Part III. Evaluation in pathologic states in which abnormal function of the neurohypophyseal system had been demonstrated or postulated.","{'model': 'tldr@v2.0.0', 'text': 'The study was divided into three parts: Effect of alcohol ingestion on water, electrolyte, and acid-base metabolism in semi-recumbent hydrated normal subjects, and evaluation of the inhibition of ADH secretion by alcohol, in normal subjects in whom acute changes of extracellular fluid tonicity and ""effective circulating blood volume"" were induced.'}",http://www.jci.org/articles/view/103092/files/pdf
-complications in the management of large glomus jugulae tumors,J. Cece,"This study analyzes the complications encountered in the surgical treatment of 17 patients with large glomus jugulare tumors. All 17 patients sustained either new cranial nerve palsies or exacerbation of preexisting palsies. These involved, in descending order of frequency, the facial nerve, the vagus and glossopharyngeal nerves, and the hypoglossal nerve. Post‐operative palsies of the facial nerve involved 17 patients, as compared to 5 preoperatively. Fifteen patients had postoperative partial or complete paralysis of the vagus nerve as compared to eight preoperatively. Ten patients had postoperative palsies of the hypoglossal nerve as compared to six preoperatively. Other complications included CSF leak, meningitis, and wound infection. Aspiration and dysphagia were encountered postoperatively in 13 patients. Teflonr̀ injection of paretic vocal cords and cricopharyngeal myotomy effectively improved the ability to swallow and the quality of the voice. Prompt recognition and treatment of complications is essential for effective surgical management of large glomus jugulare tumors.","{'model': 'tldr@v2.0.0', 'text': 'This study analyzes the complications encountered in the surgical treatment of 17 patients with large glomus jugulare tumors and found that teflonr̀ injection of paretic vocal cords and cricopharyngeal myotomy effectively improved the ability to swallow and the quality of the voice.'}",
-combination of laser photocoagulation and intravitreal bevacizumab in aggressive posterior retinopathy of prematurity.,H. Altınsoy,"The response to combined laser photocoagulation and a single intravitreal injection of 0.75 mg bevacizumab to each eye on separate days in two patients with aggressive, posterior retinopathy of prematurity (ROP) is described. Combined treatment resulted in regression of zone-1 disease in Case 1, which had no retinal detachment. However, no significant regression or unfavorable anatomic response was observed in the second case with retinal detachment. Although the combination of laser photocoagulation and intravitreal bevacizumab injection seems to be well tolerated, inducing prompt regression of agressive zone-1 ROP without retinal detachment, further controlled studies with long-term follow-up are necessary for their use in the treatment of ROP with for potentially dangerous growth factor inhibitors in premature babies.","{'model': 'tldr@v2.0.0', 'text': 'Although the combination of laser photocoagulation and intravitreal bevacizumab injection seems to be well tolerated, inducing prompt regression of agressive zone-1 ROP without retinal detachment, further controlled studies with long-term follow-up are necessary for their use in the treatment of ROP with for potentially dangerous growth factor inhibitors in premature babies.'}",
-determination of lambda-mode eigenvalue separation of a thermal accelerator-driven system from pulsed neutron experiment,Hiroshi Taninaka,"Basic research on the Accelerator-Driven System (ADS) with thermal neutron spectrum has been promoted by the Kyoto University Research Reactor Institute. At the Kyoto University Critical Assembly (KUCA), various experiments on thermal ADS with a pulsed spallation source are planned. In such an ADS, neutron flux distribution may be sensitive to the injection of neutrons, and the high sensitivity results in various spatial effects. In this study, a pulsed neutron experiment with 14MeV neutrons was carried out in a thermal ADS of KUCA, to determine the λ-mode eigenvalue separation, which is a quantitative indication of spatial effects. An original data-processing technique was applied to infer prompt-neutron decay constants of fundamental and higher modes from neutron count decay data, and then the eigenvalue separation around 13%Δk/k was obtained from these decay constants.",,
-availability of amino acids supplied intravenously in healthy man as synthetic dipeptides: kinetic evaluation of l-alanyl-l-glutamine and glycyl-l-tyrosine.,S. Albers,"1. For the first time utilization of intravenously administered L-alanyl-L-glutamine and glycyl-L-tyrosine was investigated by means of their kinetic behaviour after bolus injection in 10 and 11 apparently healthy male subjects (age 26.6 +/- 5.7 years), respectively. 2. The injection of the synthetic dipeptides was not accompanied by any side effects or complaints. 3. The synthetic dipeptides L-alanyl-L-glutamine and glycyl-L-tyrosine were rapidly cleared from plasma. By applying a monoexponential model the elimination half-lives were found to have very similar values (3.8 +/- 0.5 and 3.4 +/- 0.3 min) whether alanine or glycine was occupying the N-terminal position. The estimated volume of distribution was approximately that of the extracellular space. 4. Peptide disappearance was accompanied by a prompt equimolar increase in the concentrations of the constituent amino acids alanine and glutamine as well as glycine and tyrosine. 5. The study provides firm evidence that L-alanyl-L-glutamine and glycyl-L-tyrosine are rapidly (quantitatively) hydrolysed. The results of this study may indicate a safe and efficient parenteral use of the investigated peptides as sources of free glutamine and free tyrosine.","{'model': 'tldr@v2.0.0', 'text': 'The results of this study may indicate a safe and efficient parenteral use of the investigated peptides as sources of free glutamine and free tyrosine.'}",
-respiratory modulation of barareceptor and chemoreceptor reflexes affecting heart rate and cardiac vagal efferent nerve activity.,N. Davidson,"1. Brief stimuli were delivered to the carotid chemoreceptors or baroreceptors in dogs anaesthetized with chloralose. Chemoreceptor stimulation was achieved by rapid retrograde injection of 0.2‐0.5 ml. CO2 equilibrated saline through a cannula in the external carotid artery. Baroreceptor stimulation was achieved by forceful retrograde injection of 2‐5 ml. air‐equilibrated saline into the external carotid artery after first clamping the common carotid artery. 2. prompt decreases in heart rate were elicited by brief sudden chemoreceptor or baroreceptor stimuli when these were delivered during the expiratory phase of respiration. The stimuli did not modify the control heart rate pattern when delivered in the inspiratory phase of respiration. This respiratory modulation of reflex effectiveness persisted when the animals were completely paralysed and the phase of the respiratory cycle was monitored through a phrenic electroneurogram. 3. single cardiac vagal efferent nerve fibres were dissected from the cut central end of the right cervical vagus nerve. They were classified as cardiac efferents by their cardiac and respiratory rhythmicity, and by their increased activity in response to stimulation of a carotid sinus nerve or to mechanical elevation of the systemic arterial pressure. These efferent fibres increased their activity in response to brief chemoreceptor or baroreceptor stimuli delivered in expiration, but did not respond to stimuli delivered in inspiration. This respiratory modulation of both reflexes persisted after bilateral cervical vagotomy.",,
-baroreceptor and chemoreceptor influences on heart rate during the respiratory cycle in the dog.,B. T. Haymet,"1. Brief stimuli were delivered to the carotid chemoreceptors or baroreceptors in dogs anaesthetized with pentobarbitone or chloralose. Chemoreceptor stimulation was achieved by rapid retrograde injections of 0‐2‐0‐5 ml. warmed, CO2‐equilibrated saline through a cannula in the external carotid artery. Baroreceptor stimulation was achieved by forceful retrograde injection of 2–5 ml. air‐equilibrated saline, or of freshly drawn arterial blood, into the external carotid artery after first clamping the common carotid artery. 2. Brief baroreceptor stimuli had no noticeable effect on breathing. Brief chemoreceptor stimuli had no effect on breathing in some dogs, but in many produced a reflex increase in the depth of inspiration when delivered during inspiration. In these same dogs, brief chemoreceptor stimuli delivered in expiration either prolonged the expiratory pause or evoked an active expiratory effort. 3. Prompt decreases in heart rate were elicited by brief sudden chemoreceptor or baroreceptor stimuli when these were delivered during the expiratory phase of respiration. The stimuli did not modify the control heart rate pattern when delivered during inspiration. If the carotid sinus nerve or the vagus nerves were cut the responses were abolished. 4. Brief chemoreceptor or baroreceptor stimuli remained effective in evoking prompt decreases in heart rate during periods of apnoea in the end‐inspiratory position (Hering‐Breuer inflation reflex). In periods of apnoea after prolonged artificial hyperventilation the stimuli were sometimes ineffective at first, but were always effective late in the period of apnoea, again producing prompt cardiac slowing. 5. After denervation of the lungs, brief baroreceptor and chemoreceptor stimuli continued to evoke prompt falls in heart rate when given during expiration. When delivered during inspiration the same stimuli were either ineffective, or less effective.","{'model': 'tldr@v2.0.0', 'text': 'Brief chemoreceptor or baroreceptor stimuli remained effective in evoking prompt decreases in heart rate during periods of apnoea in the end‐inspiratory position (Hering‐Breuer inflation reflex), but were sometimes ineffective at first, but were always effective late in the period of apNoea, again producing prompt cardiac slowing.'}",
-the pathophysiology of barium: hypokalemic and cardiovascular effects.,O. Roza,"The ingestion of absorbable barium salts, e.g. , carbonate or chloride, produces a combination of ectopic ventricular contractions, ventricular tachycardia, skeletal muscle paralysis, salivation, diarrhea, hypertension and finally, respiratory paralysis and ventricular fibrillation. Infusion of barium chloride into anesthetized dogs produced all of the above, plus a prompt and substantial hypokalemia. The data suggest that the hypokalemia is due to a transfer of potassium from extracellular to intracellular compartments rather than to urinary or gastrointestinal losses. Potassium administration prevented or reversed all of the clinical effects except the hypertension. Arterial hypertension was a constant feature of BaCI3 infusion and was not affected by simultaneous potassium infusion, by the injection of phentolamine or by bilateral nephrectomy. Barium and potassium were each found to be powerful antagonists of the other's cardiac toxicity. The intracellular accumulation of potassium induced by barium together with the reported ability of barium to stimulate myocardial adenosine triphosphatase suggests the possibility of a role for barium in the management of digitalis poisoning.","{'model': 'tldr@v2.0.0', 'text': 'The intracellular accumulation of potassium induced by barium together with the reported ability of barium to stimulate myocardial adenosine triphosphatase suggests the possibility of a role for barium in the management of digitalis poisoning.'}",
-nonsurgical coronary artery recanalization in acute transmural myocardial infarction,D. Mathey,"In 41 consecutive patients with an acute transmural myocardial infarction (AMI) admitted within 3 hours after the onset of symptoms, we tried to recanalize the occluded coronary artery by an intracoronary infusion of streptokinase (SK) (2000 units/min). SK infusion was preceded by (1) an intracoronary injection of 0.5 mg nitroglycerin to rule out coronary artery spasm, (2) an attempt to recanalize the vessel mechanically with a flexible guidewire, and (3) an intracoronary injection of plasminogen (500 units) to increase the efficacy of the subsequent SK infusion. Coronary angiography revealed a total coronary artery occlusion in 39 patients and a subtotal occlusion in two patients. In 30 patients (73%), the occluded coronary artery was successfully recanalized within 1 hour (mean 29 ± 15 minutes), resulting in prompt contrast filling of the previously occluded vessel. An arteriosclerotic stenosis always remained at the site of the occlusion. Nitroglycerin opened the occluded coronary artery in one patient, contrast injection in seven patients and guidewire perforation in four of the 15 patients, in whom it was attempted. In 18 patients the occluded coronary artery was recanalized by intracoronary SK infusion alone. After the initial opening of the occluded coronary artery, subsequent SK infusion markedly reduced the degree of stenosis and visible thrombi disappeared. Clinically, recanalization was associated with significant relief of ischemic chest pain.None of the successfully recanalized patients died, including three patients with cardiogenic shock. Recanalization, however, did not prevent myocardial infarction, as shown by new Q waves and/or R-wave reduction in 24 of the 30 patients and by the rise in serum CPK with an early peak, indicating CPK washout by coronary artery reperfusion. Repeat angiography 7–21 days later revealed a patent coronary artery in 12 of 15 successfully recanalized patients. The left ventricular ejection fraction had significantly improved, from 37 ± 5% to 47 ± 4% (mean ± SEM). Failure of recanalization in 11 of 41 patients may be explained by the absence of coronary artery thrombosis or poor SK penetration of the thrombus because of its distal location or SK runoff into nonaffected arteries.Thus, in patients with acute myocardial infarction the occluded coronary artery can be rapidly recanalized in 73% of the patients by an intracoronary infusion of streptokinase.","{'model': 'tldr@v2.0.0', 'text': 'In patients with acute myocardial infarction the occluded coronary artery can be rapidly recanalized in 73% of the patients by an intracoronary infusion of streptokinase, and success in 11 of 41 patients may be explained by the absence of coronary artery thrombosis or poor SK penetration of the thrombus because of its distal location.'}",
-nonosmolar factors affecting renal water excretion (second of two parts).,Robert W. Schrier,"(First of Two Parts) THE pioneering work of Verney and his co-workers1 defined a very delicate mechanism that maintains the constancy of the osmotic pressure of body fluids by regulating the release of vasopressin. These investigators clearly showed that an increase in plasma osmotic pressure as small as 1 to 2 per cent by the intracarotid injection of hypertonic saline causes the release of an amount of vasopressin capable of dramatically decreasing renal water excretion.1 Conversely, the injection of a hypotonic solution into the carotid circulation has been shown to cause a prompt diuresis.2 These results have been used to . . .","{'model': 'tldr@v2.0.0', 'text': 'These investigators clearly showed that an increase in plasma osmotic pressure as small as 1 to 2 per cent by the intracarotid injection of hypertonic saline causes the release of an amount of vasopressin capable of dramatically decreasing renal water excretion.'}",
-the location of the chemoreceptor controlling gastric acid secretion during hypoglycaemia,D. Colin-Jones,"1. The injection of 2‐deoxy‐D‐glucose directly into the lateral hypothalamic area of rats, but not elsewhere, caused a prompt and sustained secretion of acid by the stomach at a rate comparable to that due to insulin hypoglycaemia.","{'model': 'tldr@v2.0.0', 'text': 'The injection of 2‐deoxy‐D‐glucose directly into the lateral hypothalamic area of rats, but not elsewhere, caused a prompt and sustained secretion of acid by the stomach at a rate comparable to that due to insulin hypoglycaemia.'}",
-the circulatory effects of intravenous phentolamine in man,S. Taylor,"Circulatory observations have been made on six normal subjects and six hypertensive patients before and after the acute intravenous injection of 5 mg. of phentolamine.The drug caused a prompt reduction in systemic vascular resistance which resulted in a rapid fall in systemic blood pressure in spite of an increase in heart rate and cardiac output.Evidence is presented that the predominant vascular activity of the drug is to cause a direct relaxation of vascular smooth muscle, an effect far more potent than its antagonism of circulating catecholamines or its very much weaker sympathetic-blocking action. This direct depressor effect on vascular smooth muscle unaccompanied by any but slight sympathetic blocking activities endows the drug with theoretically ideal antihypertensive properties.Phentolamine caused a transient but significant reduction in oxygen uptake in the majority of individuals. It is suggested that this may be due to a direct inhibition of cellular oxidative mechanisms.An appraisal of the circulatory activities of the drug with particular regard to its use as a diagnostic test for pheochromocytoma leads to serious doubts, not only with regard to the precision with which the results of such tests can be interpreted, but also to the validity of the use of the drug as a reliable screening test for excessive circulating catecholamines.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is presented that the predominant vascular activity of the drug is to cause a direct relaxation of vascular smooth muscle, an effect far more potent than its antagonism of circulating catecholamines or its very much weaker sympathetic-blocking action.'}",https://www.ahajournals.org/doi/pdf/10.1161/01.CIR.31.5.741
-inhibition of angiotensin conversion and prevention of renal hypertension.,Edward D. Miller,"Renal artery constriction in the unilaterally nephrectomized, trained dog, with maintained renal arterial hypotension, produces a prompt increase in systemic renin activity and blood pressure. The hypertension normally induced by renal artery stenosis is prevented by prior treatment with the nonapeptide Pyr-Trp-Pro-Arg-Pro-Gln-Ile-Pro-Pro (SQ 20, 881), which blocks conversion of angiotensin I to angiotensin II. Constant intravenous infusion of the inhibitor over several days of renal artery constriction prevents the development of chronic renovascular hypertension. Furthermore, a single injection of the nonapeptide restores blood pressure to normal in the early phase of renovascular hypertension, but becomes progressively less effective as salt and water retention occurs in the chronic stage when plasma renin activity returns to control levels. These data provide strong evidence that the renin-angiotensin system is responsible for the initiation of renovascular hypertension in the one-kidney Goldblatt dog, but that other factors become increasingly important in chronic renovascular hypertension.","{'model': 'tldr@v2.0.0', 'text': 'Data provide strong evidence that the renin-angiotensin system is responsible for the initiation of renov vascular hypertension in the one-kidney Goldblatt dog, but that other factors become increasingly important in chronic renovascular hypertension.'}",
-hydrodynamic collimation of relativistic outflows: semianalytic solutions and application to gamma-ray bursts,O. Bromberg,"A model is developed for the confinement and collimation of a baryon-poor outflow by its surrounding medium. Both confinement by kinetic pressure of a static corona and confinement by the ram pressure of a supersonic wind emanating from a disk surrounding the inner source are considered. Solutions are presented for the structure of the shocked layers of a deflected baryon-poor jet (BPJ) and exterior wind. The dependence of the opening angle of the BPJ on the parameters of the confining medium are carefully examined. It is found that the BPJ shock may either converge to the symmetry axis or diverge away from it, depending on the opening angle of the BPJ injection cone. In the latter case, the inner flow exhibits a nonuniform structure, consisting of an ultrarelativistic core containing the unshocked BPJ enveloped by the slower, shocked BPJ layer. The implications of our results to the prompt GRB emission are briefly discussed.",,https://arxiv.org/pdf/0705.2040
-infectious endophthalmitis.,L. Ficker,"A clinical series of twenty-eight cases of suspected infectious endophthalmitis was reviewed. The combination of epidemiology and peroperative microscopy at the time of vitreous biopsy provided a useful indication of the pathogen involved. The visual prognosis was related to the virulence of the pathogen, and to the delay between recognition of symptoms or signs and treatment. It appears possible to predict high risk groups which warrant a high index of clinical suspicion and prompt diagnosis to enable early vitrectomy with injection of appropriate intraocular antibiotics.","{'model': 'tldr@v2.0.0', 'text': 'It appears possible to predict high risk groups which warrant a high index of clinical suspicion and prompt diagnosis to enable early vitrectomy with injection of appropriate intraocular antibiotics.'}",
-the posterior pituitary: regulation of anterior pituitary prolactin secretion.,L. Peters,"Removal of the posterior pituitary from anesthetized male rats results in a prompt and significant increase in circulating prolactin that is reversed by the injection of dopamine. Posterior pituitary extracts, which contain high concentrations of endogenous dopamine, inhibit prolactin secretion from isolated anterior pituitary cells. This inhibition is prevented by incubation of the cells with the dopamine receptor antagonist (+)-butaclamol. The data show that posterior pituitary dopamine reaches the anterior pituitary via the short hypophysial portal vessels and participates in the regulation of prolactin secretion.","{'model': 'tldr@v2.0.0', 'text': 'The data show that posterior pituitary dopamine reaches the anterior pituitsary via the short hypophysial portal vessels and participates in the regulation of prolactin secretion.'}",
-topological data analysis guided segment anything model prompt optimization for zero-shot segmentation in biological imaging,R. Glatt,"Emerging foundation models in machine learning are models trained on vast amounts of data that have been shown to generalize well to new tasks. Often these models can be prompted with multi-modal inputs that range from natural language descriptions over images to point clouds. In this paper, we propose topological data analysis (TDA) guided prompt optimization for the Segment Anything Model (SAM) and show preliminary results in the biological image segmentation domain. Our approach replaces the standard grid search approach that is used in the original implementation and finds point locations based on their topological significance. Our results show that the TDA optimized point cloud is much better suited for finding small objects and massively reduces computational complexity despite the extra step in scenarios which require many segmentations.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the TDA optimized point cloud is much better suited for finding small objects and massively reduces computational complexity despite the extra step in scenarios which require many segmentations.'}",http://arxiv.org/pdf/2306.17400
-dialogue for prompting: a policy-gradient-based discrete prompt optimization for few-shot learning,Chengzhengxu Li,"Prompt-based pre-trained language models (PLMs) paradigm have succeeded substantially in few-shot natural language processing (NLP) tasks. However, prior discrete prompt optimization methods require expert knowledge to design the base prompt set and identify high-quality prompts, which is costly, inefficient, and subjective. Meanwhile, existing continuous prompt optimization methods improve the performance by learning the ideal prompts through the gradient information of PLMs, whose high computational cost, and low readability and generalizability are often concerning. To address the research gap, we propose a Dialogue-comprised Policy-gradient-based Discrete Prompt Optimization ($DP_2O$) method. We first design a multi-round dialogue alignment strategy for readability prompt set generation based on GPT-4. Furthermore, we propose an efficient prompt screening metric to identify high-quality prompts with linear complexity. Finally, we construct a reinforcement learning (RL) framework based on policy gradients to match the prompts to inputs optimally. By training a policy network with only 0.67% of the PLM parameter size on the tasks in the few-shot setting, $DP_2O$ outperforms the state-of-the-art (SOTA) method by 1.52% in accuracy on average on four open-source datasets. Moreover, subsequent experiments also demonstrate that $DP_2O$ has good universality, robustness, and generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'A Dialogue-comprised Policy-gradient-based Discrete Prompt Optimization ($DP_2O$) method that outperforms the state-of-the-art (SOTA) method by 1.52% in accuracy on average on four open-source datasets and has good universality, robustness, and generalization ability.'}",https://arxiv.org/pdf/2308.07272
-promptagent: strategic planning with language models enables expert-level prompt optimization,Xinyuan Wang,"Highly effective, task-specific prompts are often heavily engineered by experts to integrate detailed instructions and domain insights based on a deep understanding of both instincts of large language models (LLMs) and the intricacies of the target task. However, automating the generation of such expert-level prompts remains elusive. Existing prompt optimization methods tend to overlook the depth of domain knowledge and struggle to efficiently explore the vast space of expert-level prompts. Addressing this, we present PromptAgent, an optimization method that autonomously crafts prompts equivalent in quality to those handcrafted by experts. At its core, PromptAgent views prompt optimization as a strategic planning problem and employs a principled planning algorithm, rooted in Monte Carlo tree search, to strategically navigate the expert-level prompt space. Inspired by human-like trial-and-error exploration, PromptAgent induces precise expert-level insights and in-depth instructions by reflecting on model errors and generating constructive error feedback. Such a novel framework allows the agent to iteratively examine intermediate prompts (states), refine them based on error feedbacks (actions), simulate future rewards, and search for high-reward paths leading to expert prompts. We apply PromptAgent to 12 tasks spanning three practical domains: BIG-Bench Hard (BBH), as well as domain-specific and general NLP tasks, showing it significantly outperforms strong Chain-of-Thought and recent prompt optimization baselines. Extensive analyses emphasize its capability to craft expert-level, detailed, and domain-insightful prompts with great efficiency and generalizability.","{'model': 'tldr@v2.0.0', 'text': 'This work presents PromptAgent, an optimization method that autonomously crafts prompts equivalent in quality to those handcrafted by experts, and applies it to 12 tasks spanning three practical domains, showing it significantly outperforms strong Chain-of-Thought and recent prompt optimization baselines.'}",
-emotion-conditioned text generation through automatic prompt optimization,Yarik Menchaca Resendiz,"Conditional natural language generation methods often require either expensive fine-tuning or training a large language model from scratch. Both are unlikely to lead to good results without a substantial amount of data and computational resources. Prompt learning without changing the parameters of a large language model presents a promising alternative. It is a cost-effective approach, while still achieving competitive results. While this procedure is now established for zero- and few-shot text classification and structured prediction, it has received limited attention in conditional text generation. We present the first automatic prompt optimization approach for emotion-conditioned text generation with instruction-fine-tuned models. Our method uses an iterative optimization procedure that changes the prompt by adding, removing, or replacing tokens. As objective function, we only require a text classifier that measures the realization of the conditional variable in the generated text. We evaluate the method on emotion-conditioned text generation with a focus on event reports and compare it to manually designed prompts that also act as the seed for the optimization procedure. The optimized prompts achieve 0.75 macro-average F1 to fulfill the emotion condition in contrast to manually designed seed prompts with only 0.22 macro-average F1.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the first automatic prompt optimization approach for emotion-conditioned text generation with instruction-fine-tuned models that uses an iterative optimization procedure that changes the prompt by adding, removing, or replacing tokens.'}",https://arxiv.org/pdf/2308.04857
-multiprompter: cooperative prompt optimization with multi-agent reinforcement learning,Dong-Ki Kim,"Recently, there has been an increasing interest in automated prompt optimization based on reinforcement learning (RL). This approach offers important advantages, such as generating interpretable prompts and being compatible with black-box foundation models. However, the substantial prompt space size poses challenges for RL-based methods, often leading to suboptimal policy convergence. This paper introduces MultiPrompter, a new framework that views prompt optimization as a cooperative game between prompters which take turns composing a prompt together. Our cooperative prompt optimization effectively reduces the problem size and helps prompters learn optimal prompts. We test our method on the text-to-image task and show its ability to generate higher-quality images than baselines.","{'model': 'tldr@v2.0.0', 'text': 'MultiPrompter, a new framework that views prompt optimization as a cooperative game between prompters which take turns composing a prompt together, is introduced, which effectively reduces the problem size and helpsprompters learn optimal prompts.'}",
-robust prompt optimization for large language models against distribution shifts,Moxin Li,"Large Language Model (LLM) has demonstrated significant ability in various Natural Language Processing tasks. However, their effectiveness is highly dependent on the phrasing of the task prompt, leading to research on automatic prompt optimization using labeled task data. We reveal that these prompt optimization techniques are vulnerable to distribution shifts such as subpopulation shifts, which are common for LLMs in real-world scenarios such as customer reviews analysis. In this light, we propose a new problem of robust prompt optimization for LLMs against distribution shifts, which requires the prompt optimized over the labeled source group can simultaneously generalize to an unlabeled target group. To solve this problem, we propose Generalized Prompt Optimization framework, which incorporates the unlabeled data from the target group into prompt optimization. Extensive experimental results demonstrate the effectiveness of the proposed framework with significant performance improvement on the target group and comparable performance on the source group.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Generalized Prompt Optimization framework, which incorporates the unlabeled data from the target group into prompt optimization, and demonstrates the effectiveness of the proposed framework with significant performance improvement on thetarget group and comparable performance on the source group.'}",
-towards an automatic prompt optimization framework for ai image generation,Ling Fan,,,
-query-dependent prompt evaluation and optimization with offline inverse rl,Hao Sun,"In this study, we aim to enhance the arithmetic reasoning ability of Large Language Models (LLMs) through zero-shot prompt optimization. We identify a previously overlooked objective of query dependency in such optimization and elucidate two ensuing challenges that impede the successful and economical design of prompt optimization techniques. One primary issue is the absence of an effective method to evaluate prompts during inference when the golden answer is unavailable. Concurrently, learning via interactions with the LLMs to navigate the expansive natural language prompting space proves to be resource-intensive. To address this, we introduce Prompt-OIRL, which harnesses offline inverse reinforcement learning to draw insights from offline prompting demonstration data. Such data exists as by-products when diverse prompts are benchmarked on open-accessible datasets. With Prompt-OIRL, the query-dependent prompt optimization objective is achieved by first learning an offline reward model. This model can evaluate any query-prompt pairs without accessing LLMs. Subsequently, a best-of-N strategy is deployed to recommend the optimal prompt. Our experimental evaluations across various LLM scales and arithmetic reasoning datasets underscore both the efficacy and economic viability of the proposed approach.","{'model': 'tldr@v2.0.0', 'text': 'This study introduces Prompt-OIRL, which harnesses offline inverse reinforcement learning to draw insights from offline prompting demonstration data and achieves the query-dependent prompt optimization objective by first learning an offline reward model.'}",https://arxiv.org/pdf/2309.06553
-visual-language prompt tuning with knowledge-guided context optimization,Hantao Yao,"Prompt tuning is an effective way to adapt the pretrained visual-language model (VLM) to the downstream task using task-related textual tokens. Representative CoOp-based work combines the learnable textual tokens with the class tokens to obtain specific textual knowledge. However, the specific textual knowledge is worse generalization to the unseen classes because it forgets the essential general textual knowledge having a strong generalization ability. To tackle this issue, we introduce a novel Knowledge-guided Context Optimization (KgCoOp) to enhance the generalization ability of the learnable prompt for unseen classes. The key insight of KgCoOp is that the forgetting about essential knowledge can be alleviated by reducing the discrepancy between the learnable prompt and the hand-crafted prompt. Especially, KgCoOp minimizes the discrepancy between the textual embeddings generated by learned prompts and the hand-crafted prompts. Finally, adding the KgCoOp upon the contrastive loss can make a discriminative prompt for both seen and unseen tasks. Extensive evaluation of several benchmarks demonstrates that the proposed Knowledge-guided Context Optimization is an efficient method for prompt tuning, i.e., achieves better performance with less training time. code.","{'model': 'tldr@v2.0.0', 'text': 'A novel Knowledge-guided Context Optimization (KgCoOp) is introduced to enhance the generalization ability of the learnable prompt for unseen classes and achieves better performance with less training time for prompt tuning.'}",https://arxiv.org/pdf/2303.13283
-optimization of the signal-to-background ratio in prompt gamma imaging using energy and shifting time-of-flight discrimination: experiments with a scanning parallel-slit collimator,P. Cambraia Lopes,"Much attention is currently being paid to imaging prompt gamma (PG) rays for in vivo proton range monitoring in proton therapy. PG imaging using a collimator is affected by neutron-related background. We study the effectiveness of background reduction experimentally, using a scanning parallel-slit PG collimator as a simplified model of a multislat PG camera. The analysis is focused on the falloff region of the PG intensity profile near the Bragg peak, which is the typical region of interest for proton range estimation. Background reduction was studied for different energy windows, with and without a shifting time-of-flight window that takes into account the proton velocity within the phantom. Practical methods are put forward that apply to cyclotron-based pencil beams. The parallel-slit collimator was placed in front of arrays of cerium-doped lutetium yttrium silicate-coupled digital silicon photomultipliers, used to measure energy and time spectra together with intensity profiles of prompt events emitted from a polymethylmethacrylate phantom irradiated with a 160-MeV proton pencil beam. The best signal-to-background ratio of ~1.6 was similar to that obtained previously with a knife-edge-slit collimator. However, the slope-over-noise ratio in the PG-profile falloff region, was ~1.2 higher for the present collimator, given its better resolution.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of background reduction experimentally is studied, using a scanning parallel-slit PG collimator as a simplified model of a multislat PG camera, focused on the falloff region of the PG intensity profile near the Bragg peak, which is the typical region of interest for proton range estimation.'}",https://ieeexplore.ieee.org/ielx7/7433213/8453249/08385159.pdf
-cebr3 scintillators for 4he prompt gamma spectroscopy: results from a monte carlo optimization study,R. Dal Bello,"PURPOSE Range uncertainties limit the potential of charged particle therapy. In vivo and online range verification techniques could increase the confidence in the dose delivery distribution and lead to more conformal treatments. Prompt gamma imaging and prompt gamma spectroscopy (PGS) have been demonstrated for such a purpose. The successful application of these techniques requires the development of a dedicated detector system optimized to the radiation energy ranges and the intensity. In this work, we investigated a detector system based on CeBr3 crystals capable of performing spectroscopy of the prompt gamma radiation induced by 4 He beams. METHODS We performed Monte Carlo simulations to optimize the detector system. The study was carried out both with the Geant4 toolkit and the FLUKA package. The simulated system consisted of a primary crystal for spectroscopy and secondary crystals for noise reduction in anticoincidence (AC). For comparison purposes, we considered a configuration without AC crystals. We first defined the dimensions of the primary cerium bromide (CeBr3 ) crystal and the secondary bismuth germanate (BGO) or CeBr3 crystals. We then evaluated their detection performance for monoenergetic gamma radiation up to 7 MeV in such way that the probability of the photo-peak detection was maximized in comparison to the number of escape peak and Compton events. We simulated realistic prompt gamma radiation spectra induced by 4 He beams on homogeneous targets (water, graphite, and aluminum) and on implants (water with an aluminum insert). Finally, we tested the performances of the optimized systems in the detection of the realistic gamma spectra. The quantitative analysis was accomplished by comparing the signal-to-noise ratio between the different configurations and the ability to resolve the discrete reactions. RESULTS We present the optimized dimensions for the primary CeBr3 crystals with and without AC shielding. The specific values are given over a wide range of crystal volumes. The results show an optimal primary CeBr3 crystal with an approximately diameter to length ratio of 1 without AC shielding and 0.5 with AC shielding. The secondary BGO and CeBr3 should have a transverse dimension of 3 and 4.56 cm, respectively. The analysis of the prompt gamma spectra from 4 He beams highlighted the presence of specific discrete reactions not observed in 1 H studies, for example, 12 C transition 0+ (7.65 MeV) →2+ (4.44 MeV). This reaction is responsible for the generation of the 3.21 MeV prompt gamma peak. The optimized primary crystal provides a significant increase in the signal-to-noise ratio together with an improved resolution of the discrete gamma lines, especially in the high-energy region. The detection configuration with an optimized anticoincidence crystal improved the signal-to-noise ratio up to a factor of 3.5. CONCLUSIONS This work provides the optimal geometry for primary and secondary crystals to be used in range verification through PGS. The simulations show that such a PGS system may allow for the simultaneous detection of the discrete lines from a thin metal implant within a water phantom.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates a detector system based on CeBr3 crystals capable of performing spectroscopy of the prompt gamma radiation induced by 4 He beams and provides the optimal geometry for primary and secondary crystals to be used in range verification through PGS.'}",
-copner: contrastive learning with prompt guiding for few-shot named entity recognition,Yucheng Huang,"Distance metric learning has become a popular solution for few-shot Named Entity Recognition (NER). The typical setup aims to learn a similarity metric for measuring the semantic similarity between test samples and referents, where each referent represents an entity class. The effect of this setup may, however, be compromised for two reasons. First, there is typically a limited optimization exerted on the representations of entity tokens after initing by pre-trained language models. Second, the referents may be far from representing corresponding entity classes due to the label scarcity in the few-shot setting. To address these challenges, we propose a novel approach named COntrastive learning with Prompt guiding for few-shot NER (COPNER). We introduce a novel prompt composed of class-specific words to COPNER to serve as 1) supervision signals for conducting contrastive learning to optimize token representations; 2) metric referents for distance-metric inference on test samples. Experimental results demonstrate that COPNER outperforms state-of-the-art models with a significant margin in most cases. Moreover, COPNER shows great potential in the zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt composed of class-specific words to COPNER is introduced to serve as supervision signals for conducting contrastive learning to optimize token representations and metric referents for distance-metric inference on test samples.'}",
-moderation and shielding optimization for a 252cf based prompt gamma neutron activation analyzer system,K. Hadad,,,
-generative visual prompt: unifying distributional control of pre-trained generative models,Chen Henry Wu,"Generative models (e.g., GANs, diffusion models) learn the underlying data distribution in an unsupervised manner. However, many applications of interest require sampling from a particular region of the output space or sampling evenly over a range of characteristics. For efficient sampling in these scenarios, we propose Generative Visual Prompt (PromptGen), a framework for distributional control over pre-trained generative models by incorporating knowledge of other off-the-shelf models. PromptGen defines control as energy-based models (EBMs) and samples images in a feed-forward manner by approximating the EBM with invertible neural networks, avoiding optimization at inference. Our experiments demonstrate how PromptGen can efficiently sample from several unconditional generative models (e.g., StyleGAN2, StyleNeRF, diffusion autoencoder, NVAE) in a controlled or/and de-biased manner using various off-the-shelf models: (1) with the CLIP model as control, PromptGen can sample images guided by text, (2) with image classifiers as control, PromptGen can de-bias generative models across a set of attributes or attribute combinations, and (3) with inverse graphics models as control, PromptGen can sample images of the same identity in different poses. (4) Finally, PromptGen reveals that the CLIP model shows a""reporting bias""when used as control, and PromptGen can further de-bias this controlled distribution in an iterative manner. The code is available at https://github.com/ChenWu98/Generative-Visual-Prompt.","{'model': 'tldr@v2.0.0', 'text': 'Generative Visual Prompt (PromptGen), a framework for distributional control over pre-trained generative models by incorporating knowledge of other off-the-shelf models, is proposed, and it is revealed that the CLIP model shows a reporting bias when used as control, and PromptGen can further de-bias this controlled distribution in an iterative manner.'}",http://arxiv.org/pdf/2209.06970
-cpl: counterfactual prompt learning for vision and language models,Xuehai He,"Prompt tuning is a new few-shot transfer learning technique that only tunes the learnable prompt for pre-trained vision and language models such as CLIP. However, existing prompt tuning methods tend to learn spurious or entangled representations, which leads to poor generalization to unseen concepts.Towards non-spurious and efficient prompt learning from limited examples, this paper presents a novel Counterfactual Prompt Learning (CPL) method for vision and language models, which simultaneously employs counterfactual generation and contrastive learning in a joint optimization framework.Particularly, CPL constructs counterfactual by identifying minimal non-spurious feature change between semantically-similar positive and negative samples that causes concept change, and learns more generalizable prompt representation from both factual and counterfactual examples via contrastive learning. Extensive experiments demonstrate that CPL can obtain superior few-shot performance on different vision and language tasks than previous prompt tuning methods on CLIP. On image classification, we achieve 3.55% average relative improvement on unseen classes across seven datasets; on image-text retrieval and visual question answering, we gain up to 4.09% and 25.08% relative improvements across three few-shot scenarios on unseen test sets respectively.","{'model': 'tldr@v2.0.0', 'text': 'A novel Counterfactual Prompt Learning method for vision and language models, which simultaneously employs counterfactual generation and contrastive learning in a joint optimization framework, and can obtain superior few-shot performance on different vision andlanguage tasks than previous prompt tuning methods on CLIP.'}",http://arxiv.org/pdf/2210.10362
-prompt engineering for zero‐shot and few‐shot defect detection and classification using a visual‐language pretrained model,Gunwoo Yong,"Zero‐shot learning, applied with vision‐language pretrained (VLP) models, is expected to be an alternative to existing deep learning models for defect detection, under insufficient dataset. However, VLP models, including contrastive language‐image pretraining (CLIP), showed fluctuated performance on prompts (inputs), resulting in research on prompt engineering—optimization of prompts for improving performance. Therefore, this study aims to identify the features of a prompt that can yield the best performance in classifying and detecting building defects using the zero‐shot and few‐shot capabilities of CLIP. The results reveal the following: (1) domain‐specific definitions are better than general definitions and images; (2) a complete sentence is better than a set of core terms; and (3) multimodal information is better than single‐modal information. The resulting detection performance using the proposed prompting method outperformed that of existing supervised models.","{'model': 'tldr@v2.0.0', 'text': 'The results reveal that domain‐specific definitions are better than general definitions and images; a complete sentence is better than a set of core terms; and multimodal information is betterthan single‐modal information.'}",
-understanding and mitigating overfitting in prompt tuning for vision-language models,Cheng Ma,"Pretrained vision-language models (VLMs) such as CLIP have shown impressive generalization capability in downstream vision tasks with appropriate text prompts. Instead of designing prompts manually, Context Optimization (CoOp) has been recently proposed to learn continuous prompts using task-specific training data. Despite the performance improvements on downstream tasks, several studies have reported that CoOp suffers from the overfitting issue in two aspects: (i) the test accuracy on base classes first improves and then worsens during training; (ii) the test accuracy on novel classes keeps decreasing. However, none of the existing studies can understand and mitigate such overfitting problems. In this study, we first explore the cause of overfitting by analyzing the gradient flow. Comparative experiments reveal that CoOp favors generalizable and spurious features in the early and later training stages, respectively, leading to the non-overfitting and overfitting phenomena. Given those observations, we propose Subspace Prompt Tuning (Sub PT) to project the gradients in back-propagation onto the low-rank subspace spanned by the early-stage gradient flow eigenvectors during the entire training process and successfully eliminate the overfitting problem. In addition, we equip CoOp with a Novel Feature Learner (NFL) to enhance the generalization ability of the learned prompts onto novel categories beyond the training set, needless of image training data. Extensive experiments on 11 classification datasets demonstrate that Sub PT+NFL consistently boost the performance of CoOp and outperform the state-of-the-art CoCoOp approach. Experiments on more challenging vision downstream tasks, including open-vocabulary object detection and zero-shot semantic segmentation, also verify the effectiveness of the proposed method. Codes can be found at https://tinyurl.com/mpe64f89.","{'model': 'tldr@v2.0.0', 'text': 'Subspace Prompt Tuning (Sub PT) is proposed to project the gradients in back-propagation onto the low-rank subspace spanned by the early-stage gradient flow eigenvectors during the entire training process and successfully eliminate the overfitting problem.'}",https://arxiv.org/pdf/2211.02219
-optimization of a prompt gamma setup for analysis of environmental samples,A. Naqvi,,,
-relationship between fdm 3d printing parameters study: parameter optimization for lower defects,Patrich Ferretti,"Technology evolution and wide research attention on 3D printing efficiency and processes have given the prompt need to reach an understanding about each technique’s prowess to deliver superior quality levels whilst showing an economical and process viability to become mainstream. Studies in the field have struggled to predict the singularities that arise during most Fused Deposition Modeling (FDM) practices; therefore, diverse individual description of the parameters have been performed, but a relationship study between them has not yet assessed. The proposed study lays the main defects caused by a selection of printing parameters which might vary layer slicing, then influencing the defect rate. Subsequently, the chosen technique for optimization is presented, with evidence of its application viability that suggests that a quality advance would be gathered with such. The results would help in making the FDM process become a reliable process that could also be used for industry manufacturing besides prototyping purposes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed study lays the main defects caused by a selection of printing parameters which might vary layer slicing, then influencing the defect rate, to help in making the FDM process become a reliable process that could also be used for industry manufacturing besides prototyping purposes.'}",https://www.mdpi.com/2073-4360/13/13/2190/pdf?version=1625190845
-bbtv2: pure black-box optimization can be comparable to gradient descent for few-shot learning,Tianxiang Sun,"Black-Box Tuning (BBT) is a derivative-free approach to optimize continuous prompt tokens prepended to the input of language models. Although BBT has achieved comparable performance to full model tuning on simple classification tasks under few-shot settings, it requires pre-trained prompt embedding to match model tuning on hard tasks (e.g., entailment tasks), and therefore does not completely get rid of the dependence on gradients. In this paper we present BBTv2, a pure black-box optimization approach that can drive language models to achieve comparable results to gradient-based optimization. In particular, we prepend continuous prompt tokens to every layer of the language model and propose a divide-and-conquer algorithm to alternately optimize the prompt tokens at different layers. For the optimization at each layer, we perform derivative-free optimization in a low-dimensional subspace, which is then randomly projected to the original prompt parameter space. Experimental results show that BBTv2 not only outperforms BBT by a large margin, but also achieves comparable or even better performance than full model tuning and state-of-the-art parameter-efficient methods (e.g., Adapter, LoRA, BitFit, etc.) under few-shot learning settings, while maintaining much fewer tunable parameters. 1","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that BBTv2 not only outperforms BBT by a large margin, but also achieves comparable or even better performance than full model tuning and state-of-the-art parameter-efficient methods under few-shot learning settings, while maintaining much fewer tunable parameters.'}",http://arxiv.org/pdf/2205.11200
-a model for the fast evaluation of prompt losses of energetic ions in stellarators,J. Velasco,"A good understanding of the confinement of energetic ions in non-axisymmetric magnetic fields is key for the design of reactors based on the stellarator concept. In this work, we develop a model that, based on the radially-local bounce-averaged drift-kinetic equation, classifies orbits and succeeds in predicting configuration-dependent aspects of the prompt losses of energetic ions in stellarators. Such a model could in turn be employed in the optimization stage of the design of new devices.","{'model': 'tldr@v2.0.0', 'text': 'A model is developed that, based on the radially-local bounce-averaged drift-kinetic equation, classifies orbits and succeeds in predicting configuration-dependent aspects of the prompt losses of energetic ions in stellarators.'}",https://arxiv.org/pdf/2106.05697
-healthcare scheduling in optimization context: a review,Z. A. Abdalkareem,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey on healthcare scheduling focuses on the recent literature and aims to help researchers to highlight some development from the most recent papers and grasp the new trends for future directions.'}",https://link.springer.com/content/pdf/10.1007/s12553-021-00547-5.pdf
-connecting large language models with evolutionary algorithms yields powerful prompt optimizers,Qingyan Guo,"Large Language Models (LLMs) excel in various tasks, but they rely on carefully crafted prompts that often demand substantial human effort. To automate this process, in this paper, we propose a novel framework for discrete prompt optimization, called EvoPrompt, which borrows the idea of evolutionary algorithms (EAs) as they exhibit good performance and fast convergence. To enable EAs to work on discrete prompts, which are natural language expressions that need to be coherent and human-readable, we connect LLMs with EAs. This approach allows us to simultaneously leverage the powerful language processing capabilities of LLMs and the efficient optimization performance of EAs. Specifically, abstaining from any gradients or parameters, EvoPrompt starts from a population of prompts and iteratively generates new prompts with LLMs based on the evolutionary operators, improving the population based on the development set. We optimize prompts for both closed- and open-source LLMs including GPT-3.5 and Alpaca, on 9 datasets spanning language understanding and generation tasks. EvoPrompt significantly outperforms human-engineered prompts and existing methods for automatic prompt generation by up to 25% and 14% respectively. Furthermore, EvoPrompt demonstrates that connecting LLMs with EAs creates synergies, which could inspire further research on the combination of LLMs and conventional algorithms.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework for discrete prompt optimization, called EvoPrompt, which borrows the idea of evolutionary algorithms (EAs) as they exhibit good performance and fast convergence and demonstrates that connecting LLMs with EAs creates synergies, which could inspire further research on the combination of LLMs and conventional algorithms.'}",https://arxiv.org/pdf/2309.08532
-optimization of collimator designs for real-time proton range verification by measuring prompt gamma rays,P. Cambraia Lopes,"Precise monitoring of the particle range inside the body, preferably in real-time, is a primary interest in quality assurance of particle therapy. Prompt-gamma (PG) detection aims at this task, but such a system is technically challenging to implement. For mechanically collimated PG detection, neutron-induced background radiation originating in the collimator material is a major issue. We study by Monte-Carlo simulations (Geant4), the influence of different collimator designs and their geometric parameters - such as distance to source, size and septa thickness - on the correlation between transmitted PG profiles and the longitudinal depth-dose profile. The impact of background rejection by time-of-flight (TOF) is also addressed. A single-parallel-slit collimator scanned along the beam axis was primarily studied and results show that the background can be efficiently reduced by TOF discrimination. Due to reduced statistics, signal integration from neighbouring pencil-beams was further considered, accounting for typical clinical proton fluences. Furthermore, first direct-comparison results between multi-slat collimation and a knife-edge-slit collimator are presented, obtained both by simulation and analytical calculations of their geometric performances. Advantages and disadvantages of both approaches are highlighted and some design improvements are proposed. Energy thresholding was also addressed, showing that a low-energy threshold of 4 MeV increases the signal-to-background ratio in all simulation results, a factor mostly important if TOF rejection cannot be applied.",,
-negative-prompt inversion: fast image inversion for editing with text-guided diffusion models,Daiki Miyake,"In image editing employing diffusion models, it is crucial to preserve the reconstruction quality of the original image while changing its style. Although existing methods ensure reconstruction quality through optimization, a drawback of these is the significant amount of time required for optimization. In this paper, we propose negative-prompt inversion, a method capable of achieving equivalent reconstruction solely through forward propagation without optimization, thereby enabling much faster editing processes. We experimentally demonstrate that the reconstruction quality of our method is comparable to that of existing methods, allowing for inversion at a resolution of 512 pixels and with 50 sampling steps within approximately 5 seconds, which is more than 30 times faster than null-text inversion. Reduction of the computation time by the proposed method further allows us to use a larger number of sampling steps in diffusion models to improve the reconstruction quality with a moderate increase in computation time.",,http://arxiv.org/pdf/2305.16807
-pfedprompt: learning personalized prompt for vision-language models in federated learning,Tao Guo,"Pre-trained vision-language models like CLIP show great potential in learning representations that capture latent characteristics of users. A recently proposed method called Contextual Optimization (CoOp) introduces the concept of training prompt for adapting pre-trained vision-language models. Given the lightweight nature of this method, researchers have migrated the paradigm from centralized to decentralized system to innovate the collaborative training framework of Federated Learning (FL). However, current prompt training in FL mainly focuses on modeling user consensus and lacks the adaptation to user characteristics, leaving the personalization of prompt largely under-explored. Researches over the past few years have applied personalized FL (pFL) approaches to customizing models for heterogeneous users. Unfortunately, we find that with the variation of modality and training behavior, directly applying the pFL methods to prompt training leads to insufficient personalization and performance. To bridge the gap, we present pFedPrompt, which leverages the unique advantage of multimodality in vision-language models by learning user consensus from linguistic space and adapting to user characteristics in visual space in a non-parametric manner. Through this dual collaboration, the learned prompt will be fully personalized and aligned to the user’s local characteristics. We conduct extensive experiments across various datasets under the FL setting with statistical heterogeneity. The results demonstrate the superiority of our pFedPrompt against the alternative approaches with robust performance.","{'model': 'tldr@v2.0.0', 'text': None}",
-design optimization of a 2d prompt-gamma measurement system for proton dose verification,H. Lee,"To verify in-vivo proton dose distribution, a 2-dimensional (2D) prompt-gamma measurement system, comprised of a multi-hole collimation system, a 2D array of CsI(Tl) scintillators, and a position-sensitive photomultiplier tube (PS-PMT), is under development. In the present study, to determine the optimal dimension of the measurement system, we employed a series of Monte Carlo simulations with the MCNPX code. To effectively measure the high-energy prompt gammas while minimizing background gammas, we determined the collimator hole size, collimator thickness, and scintillator length to be 0.4 × 0.4 cm2, 15 cm, and 5 cm, respectively. Thereafter, the performance of the optimized measurement system was estimated for monoenergetic proton pencil beams. The peak locations of the prompt-gamma distributions for 80- and 150-MeV proton beams were clearly distinguished, and the correlation between the beam range and the peak location was confirmed by using the measurement system. For a 200-MeV proton beam, however, the peak location could not be determined due to the dominance of background gammas and the lateral dispersion of the proton beam at the end of the beam range. Based on these simulation results, a prototype 2D prompt-gamma measurement system currently is under construction and, upon completion, will be tested with therapeutic proton beams.",,
-iterative prompt learning for unsupervised backlit image enhancement,Zhexin Liang,"We propose a novel unsupervised backlit image enhancement method, abbreviated as CLIP-LIT, by exploring the potential of Contrastive Language-Image Pre-Training (CLIP) for pixel-level image enhancement. We show that the open-world CLIP prior not only aids in distinguishing between backlit and well-lit images, but also in perceiving heterogeneous regions with different luminance, facilitating the optimization of the enhancement network. Unlike high-level and image manipulation tasks, directly applying CLIP to enhancement tasks is non-trivial, owing to the difficulty in finding accurate prompts. To solve this issue, we devise a prompt learning framework that first learns an initial prompt pair by constraining the text-image similarity between the prompt (negative/positive sample) and the corresponding image (backlit image/well-lit image) in the CLIP latent space. Then, we train the enhancement network based on the text-image similarity between the enhanced result and the initial prompt pair. To further improve the accuracy of the initial prompt pair, we iteratively fine-tune the prompt learning framework to reduce the distribution gaps between the backlit images, enhanced results, and well-lit images via rank learning, boosting the enhancement performance. Our method alternates between updating the prompt learning framework and enhancement network until visually pleasing results are achieved. Extensive experiments demonstrate that our method outperforms state-of-the-art methods in terms of visual quality and generalization ability, without requiring any paired data.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel unsupervised backlit image enhancement method that outperforms state-of-the-art methods in terms of visual quality and generalization ability, and iteratively fine-tune the prompt learning framework to reduce the distribution gaps between the backlit images, enhanced results, and well-lit images via rank learning, boosting the enhancement performance.'}",https://arxiv.org/pdf/2303.17569
-meta learning for domain agnostic soft prompt,Ming-Yen Chen,"The prompt-based learning, as used in GPT-3, has become a popular approach to extract knowledge from a powerful pre-trained language model (PLM) for natural language understanding tasks. However, either applying the hard prompt for sentences by defining a collection of human-engineering prompt templates or directly optimizing the soft or continuous prompt with labeled data may not really generalize well for unseen domain data. To cope with this issue, this paper presents a new prompt-based unsupervised domain adaptation where the learned soft prompt is able to boost the frozen pre-trained language model to deal with the input tokens from unseen domains. Importantly, the meta learning and optimization is developed to carry out the domain agnostic soft prompt where the loss for masked language model is minimized. The experiments on multi-domain natural language understanding tasks show the merits of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new prompt-based unsupervised domain adaptation where the learned soft prompt is able to boost the frozen pre-trained language model to deal with the input tokens from unseen domains.'}",
-ant colony optimization for the control of pollutant spreading on social networks,Wei-neng Chen,"The rapid development of online social networks not only enables prompt and convenient dissemination of desirable information but also incurs fast and wide propagation of undesirable information. A common way to control the spread of pollutants is to block some nodes, but such a strategy may affect the service quality of a social network and leads to a high control cost if too many nodes are blocked. This paper considers the node selection problem as a biobjective optimization problem to find a subset of nodes to be blocked so that the effect of the control is maximized while the cost of the control is minimized. To solve this problem, we design an ant colony optimization algorithm with an adaptive dimension size selection under the multiobjective evolutionary algorithm framework based on decomposition (MOEA/D-ADACO). The proposed algorithm divides the biobjective problem into a set of single-objective subproblems and each ant takes charge of optimizing one subproblem. Moreover, two types of pheromone and heuristic information are incorporated into MOEA/D-ADACO, that is, pheromone and heuristic information of dimension size selection and that of node selection. While constructing solutions, the ants first determine the dimension size according to the former type of pheromone and heuristic information. Then, the ants select a specific number of nodes to build solutions according to the latter type of pheromone and heuristic information. Experiments conducted on a set of real-world online social networks confirm that the proposed biobjective optimization model and the developed MOEA/D-ADACO are promising for the pollutant spreading control.","{'model': 'tldr@v2.0.0', 'text': 'Experiments conducted on a set of real-world online social networks confirm that the proposed biobjective optimization model and the developed MOEA/D-ADACO are promising for the pollutant spreading control.'}",
-which components of a smartphone walking app help users to reach personalized step goals? results from an optimization trial,Jan-Niklas Kramer,"Abstract Background The Assistant to Lift your Level of activitY (Ally) app is a smartphone application that combines financial incentives with chatbot-guided interventions to encourage users to reach personalized daily step goals. Purpose To evaluate the effects of incentives, weekly planning, and daily self-monitoring prompts that were used as intervention components as part of the Ally app. Methods We conducted an 8 week optimization trial with n = 274 insurees of a health insurance company in Switzerland. At baseline, participants were randomized to different incentive conditions (cash incentives vs. charity incentives vs. no incentives). Over the course of the study, participants were randomized weekly to different planning conditions (action planning vs. coping planning vs. no planning) and daily to receiving or not receiving a self-monitoring prompt. Primary outcome was the achievement of personalized daily step goals. Results Study participants were more active and healthier than the general Swiss population. Daily cash incentives increased step-goal achievement by 8.1%, 95% confidence interval (CI): [2.1, 14.1] and, only in the no-incentive control group, action planning increased step-goal achievement by 5.8%, 95% CI: [1.2, 10.4]. Charity incentives, self-monitoring prompts, and coping planning did not affect physical activity. Engagement with planning interventions and self-monitoring prompts was low and 30% of participants stopped using the app over the course of the study. Conclusions Daily cash incentives increased physical activity in the short term. Planning interventions and self-monitoring prompts require revision before they can be included in future versions of the app. Selection effects and engagement can be important challenges for physical-activity apps. Clinical Trial Information This study was registered on ClinicalTrials.gov, NCT03384550.","{'model': 'tldr@v2.0.0', 'text': 'Daily cash incentives increased physical activity in the short term and planning interventions and self-monitoring prompts require revision before they can be included in future versions of the Ally app.'}",https://academic.oup.com/abm/article-pdf/54/7/518/33381379/kaaa002.pdf
-prompt report on exa-scale hpl-ai benchmark,Shuhei Kudo,"Our performance benchmark of HPL-AI on the supercomputer Fugaku was awarded in the 55th top500 at ISC20. The effective performance was 1.42 EFlop/s, and the world's first achievement to exceed the wall of exascale in a floating-point arithmetic benchmark. Due to the novelty of HPL-AI, there are few guidelines for large systems and several drawbacks to the large-scale benchmark. It is not enough to replace FP64 operations solely to those on FP32 or FP16. At the least, we need thoughtful numerical analysis for lower-precision arithmetic and introduction of optimization techniques on extensive computing such as on Fugaku. In the poster, we give some comments on the accuracy, implementation, performance improvement, and report on the Exa-scale benchmark on Fugaku.","{'model': 'tldr@v2.0.0', 'text': 'Comments are given on the accuracy, implementation, performance improvement, and report on the Exa-scale benchmark on Fugaku.'}",
-temporally-extended prompts optimization for sam in interactive medical image segmentation,Chuyun Shen,"The Segmentation Anything Model (SAM) has recently emerged as a foundation model for addressing image segmentation. Owing to the intrinsic complexity of medical images and the high annotation cost, the medical image segmentation (MIS) community has been encouraged to investigate SAM's zero-shot capabilities to facilitate automatic annotation. Inspired by the extraordinary accomplishments of interactive medical image segmentation (IMIS) paradigm, this paper focuses on assessing the potential of SAM's zero-shot capabilities within the IMIS paradigm to amplify its benefits in the MIS domain. Regrettably, we observe that SAM's vulnerability to prompt forms (e.g., points, bounding boxes) becomes notably pronounced in IMIS. This leads us to develop a framework that adaptively offers suitable prompt forms for human experts. We refer to the framework above as temporally-extended prompts optimization (TEPO) and model it as a Markov decision process, solvable through reinforcement learning. Numerical experiments on the standardized benchmark BraTS2020 demonstrate that the learned TEPO agent can further enhance SAM's zero-shot capability in the MIS context.","{'model': 'tldr@v2.0.0', 'text': ""This paper focuses on assessing the potential of SAM's zero-shot capabilities within the IMIS paradigm to amplify its benefits in the MIS domain, and develops a framework that adaptively offers suitable prompt forms for human experts.""}",http://arxiv.org/pdf/2306.08958
-optimization technique for a prompt gamma-ray spect collimator system.,Masayori Ishikawa,"Because background radiation in an irradiation room creates a problem with the PG-SPECT (Prompt Gamma-ray Single Photon Emission Computed Tomography) system, which evaluates the absorbed dose for the Boron Neutron Capture Therapy treatment, optimization of a collimator system was performed while taking the shielding of background gamma-rays into consideration. Assuming that a parallel-beam collimator is used, three parameters--the diameter of a hole, the length of the collimator, and the number of detectors (the number of holes of the collimator)--were selected for optimization. Because the combinations of these parameters are limitless, it is difficult to determine them simultaneously. Therefore, a statistically derived Optimization Criterion has been proposed to optimize these parameters. When the spatial resolution was 1 cm-FWHM (full width at half maximum), the optimal diameter of the collimator was 5.4 mm, the optimal length was 321 mm, and the optimal number of detectors was 31 x 31.","{'model': 'tldr@v2.0.0', 'text': 'Because background radiation in an irradiation room creates a problem with the PG-SPECT system, which evaluates the absorbed dose for the Boron Neutron Capture Therapy treatment, optimization of a collimator system was performed while taking the shielding of background gamma-rays into consideration.'}",https://academic.oup.com/jrr/article-pdf/42/4/387/2722132/jrr-42-387.pdf
-quality-aware energy optimization in wireless video communication with multipath tcp,Jiyan Wu,"The advancements in wireless communication technologies prompt the bandwidth aggregation for mobile video delivery over heterogeneous access networks. Multipath TCP (MPTCP) is the transport protocol recommended by IETF for concurrent data transmission to multihomed terminals. However, it still remains challenging to deliver user-satisfied video services with the existing MPTCP schemes because of the contradiction between energy consumption and received video quality in mobile devices. To enable the energy-efficient and quality-guaranteed video streaming, this paper presents an energy-distortion-aware MPTCP (EDAM) solution. First, we develop an analytical framework to characterize the energy-distortion tradeoff for multipath video transmission over heterogeneous wireless networks. Second, we propose a video flow rate allocation algorithm to minimize the energy consumption while achieving target video quality based on utility maximization theory. The performance of the proposed EDAM is evaluated through both experiments in real wireless networks and extensive emulations in exata. Experimental results show that EDAM exhibits performance advantages over existing MPTCP schemes in energy conservation and video quality.","{'model': 'tldr@v2.0.0', 'text': 'An analytical framework to characterize the energy-distortion tradeoff for multipath video transmission over heterogeneous wireless networks is developed and a video flow rate allocation algorithm to minimize the energy consumption while achieving target video quality based on utility maximization theory is proposed.'}",
-optimization of mass cytometry sample cryopreservation after staining,Hermi Sumatoh,"The advent of mass cytometry has facilitated highly multi‐parametric single‐cell analysis allowing for the deep assessment of cellular diversity. While the data and analytical power of this approach are well described, associated technical and experimental hurdles remain. Issues like equipment breakdown and sampling of large‐scale batches, which may require multiple days of data acquisition, are minor but critical obstacles that prompt a technical solution, especially when dealing with precious samples. An ability to cryopreserve mass cytometry samples that have already been stained would alleviate numerous technical limitations we face with currently used sample‐handling approaches. Here, we evaluated two protocols for freezing of already‐stained and fixed cellular samples and compared them with standard sample refrigeration in staining buffer. A comprehensive human T cell staining phenotypic and functional profiling panel was used and the signal intensity and reliability of each marker was assessed over a 4‐week period. In general, cellular viability, DNA Ir‐Intercalator and barcode staining were minimally affected by freezing compared to refrigeration, and the signal intensities for cell surface markers and receptors were not compromised. Intracellular cytokine staining did show some decreases in signal intensity after freezing, with the decreases more prominent in a methanol‐based protocol compared to a protocol involving the use of 10% DMSO in FBS. We conclude that freezing already‐stained samples suspended in 10% DMSO in FBS is practical and efficient way to preserve already‐stained samples when needed. © 2016 International Society for Advancement of Cytometry","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that freezing already‐stained samples suspended in 10% DMSO in FBS is practical and efficient way to preserve already-stained samples when needed and alleviate numerous technical limitations with currently used sample‐handling approaches.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/cyto.a.23014
-integrated well placement and completion optimization using heuristic algorithms: a case study of an iranian carbonate formation,Reza Khoshneshin,"Determination of optimum location for drilling a new well not only requires engineering judgments but also consumes excessive computational time. Additionally, availability of many physical constraints such as the well length, trajectory, and completion type and the numerous affecting parameters including, well type, well numbers, well-control variables prompt that the optimization approaches become imperative;. The aim of this study is to figure out optimum well location and the best completion condition using coupled simulation optimization on an Iranian oil field located in southwest of Iran. The well placement scenarios are considered in two successive time intervals during of the field life, i.e., exploration and infill drilling phase. In the former scenario, the well-placement optimization is considered to locate the drilling site of a wildcat well, while the later scenario includes the optimum drilling location of a well is determined after 10-years primary production of nine production wells. In each scenario, two stochastic optimization algorithms namely particle swarm optimization, and artificial bee colony will be applied to evaluate the considered objective function. The net present value to drill production wells through the field life is considered as an objective function during our simulation-optimization approach. Our results show that the outcome of two population-based algorithms (i.e., particle swarm optimization and artificial bee colony) is marginally different from each other. The net present value of the infill drilling phase attains higher value using artificial bee colony algorithm.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the outcome of two population-based algorithms (i.e., particle swarm optimization and artificial bee colony) is marginally different from each other, and the net present value of the infill drilling phase attains higher value using artificial bee colonies algorithm.'}",
-robust optimization model for backup resource allocation in cloud provider,Fujun He,"This paper proposes a backup resource allocation model that provides a probabilistic protection for primary physical machines in a cloud provider to minimize the required total capacity. When any random failure occurs, workloads are transferred to preplanned and dedicated backup physical machines for prompt recovery. In the proposed model, a probabilistic protection guarantee is introduced to prevent the cloud provider from capacity overbooking. We apply robust optimization in our model to formulate the backup resource allocation problem as an integer linear programming problem. A simulated annealing heuristic is adopted to solve the same optimization problem when the cloud provider is large. Finally, the results reveal that the required backup capacity depends on the reliability of primary physical machines. Specifically, the more the resources in primary physical machines share backup capacity when the failure probabilities of primary physical machines are sufficiently small, the less capacity is required for backup resource allocation.","{'model': 'tldr@v2.0.0', 'text': 'The more the resources inPrimary physical machines share backup capacity when the failure probabilities of primary physical machines are sufficiently small, the less capacity is required for backup resource allocation.'}",
-data mining algorithms for classification of diagnostic cancer using genetic optimization algorithms,R. A. Khan,"The breast tumor is the primary driver of female casualty everywhere throughout the world and the real area of study from a long time but with slighter development than anticipated. Numerous establishments and associations are working in this field to prompt to a conceivable arrangement of the issue or to prompt to additionally comprehension of the issue. Numerous past inquiries about the said were contemplated for improved comprehension of the issue and the research performed previously was to reduce dimensionality and to contribute to the betterment in the field of cancer, Wisconsin-Madison Diagnostic Breast cancer (WDBC) dataset was taken from learning repository of UCI database with 569 distinct instances for training by choosing finest features out of 32 different attributes. Different feature selection algorithms were used with data mining algorithms for better classification. Numerous enhancements in classification accuracy of WDBC were discovered by utilizing distinctive methodologies than the prior reviews directed in a similar field. The Logistic Regression, Linear Regression, and SVM algorithms showed better classification accuracy i.e. 98.24 %, 98.24 % and 98.07 % than the previous outcome results known for the said classification algorithms. The results were generated using 10 fold cross validation, by using different classification algorithms with feature selection and generation algorithms. General Terms Machine Learning, Data Mining, Classification, Genetic algorithms, Feature Selection, Algorithms, and Cancer.","{'model': 'tldr@v2.0.0', 'text': 'Improvements in classification accuracy of WDBC were discovered by utilizing distinctive methodologies than the prior reviews directed in a similar field.'}",
-prompt decision method for ground-state searches of natural computing architecture using 2d ising spin model,M. Ito,"Recently, the ability to analyze big data has been required for the optimization of social systems and the development of artificial intelligence. Consequently, the solution of combinatorial optimization problems has become important in recent years, especially for the suitable operation of infrastructures. However, combinatorial optimization problems have unique properties that the number of candidate solutions increases explosively as the number of parameters is increased. One possible resolution of this issue involves artificial and/or simulated Ising spin system. In this work, we have implemented such a spin system using “prompt decision logic”. The convergence operation was successfully observed in prompt decision method for spin interaction. Therefore, it is indicated that Ising computing by prompt decision logic could resolve combinatorial optimization problems.","{'model': 'tldr@v2.0.0', 'text': 'It is indicated that Ising computing by prompt decision logic could resolve combinatorial optimization problems.'}",
-the ancient charm project at frm ii: three-dimensional elemental mapping by prompt gamma activation imaging and neutron tomography,R. Schulze,"The aim of the EU FP6-funded ANCIENT CHARM project has been the development of new, and the integration of existing element-sensitive imaging methods utilizing neutrons. We present here the methodology and the first implementation of 3D elemental mapping at the Prompt Gamma Activation Analysis instrument of the research reactor FRM II, in order to determine spatially resolved elemental abundances in samples. After the design, optimization, and characterization of the new setup, measurements were successfully completed on archaeological objects of historical significance from the collection of the Hungarian National Museum.",,
-neck of femur fractures in the over 90s: a select group of patients who require prompt surgical intervention for optimal results,Kamal S Hapuarachchi,,"{'model': 'tldr@v2.0.0', 'text': 'The findings show that timing of surgery has an important bearing on mortality and morbidity after hip surgery, and older patients with higher orthopaedic POSSUM scores are sensitive to delays in surgery.'}",https://jorthoptraumatol.springeropen.com/counter/pdf/10.1007/s10195-013-0248-9
-evolutionary fuzzy intelligent system for multi-objective supply chain network designs: an agent-based optimization state of the art,Mosahar Tarimoradi,,"{'model': 'tldr@v2.0.0', 'text': 'A fuzzy multi-agent system compatible with the decision makers’ interests and environmental survey is put forward according to which the parameters of the mathematical model are identified and its main competencies could be contemplated regarding the facts that it interactively fulfills the decisionmakers’ utilities.'}",
-styleclip: text-driven manipulation of stylegan imagery,Or Patashnik,"Inspired by the ability of StyleGAN to generate highly realistic images in a variety of domains, much recent work has focused on understanding how to use the latent spaces of StyleGAN to manipulate generated and real images. However, discovering semantically meaningful latent manipulations typically involves painstaking human examination of the many degrees of freedom, or an annotated collection of images for each desired manipulation. In this work, we explore leveraging the power of recently introduced Contrastive Language-Image Pre-training (CLIP) models in order to develop a text-based interface for StyleGAN image manipulation that does not require such manual effort. We first introduce an optimization scheme that utilizes a CLIP-based loss to modify an input latent vector in response to a user-provided text prompt. Next, we describe a latent mapper that infers a text-guided latent manipulation step for a given input image, allowing faster and more stable text-based manipulation. Finally, we present a method for mapping text prompts to input-agnostic directions in StyleGAN’s style space, enabling interactive text-driven image manipulation. Extensive results and comparisons demonstrate the effectiveness of our approaches.","{'model': 'tldr@v2.0.0', 'text': 'This work explores leveraging the power of recently introduced Contrastive Language-Image Pre-training (CLIP) models in order to develop a text-based interface for StyleGAN image manipulation that does not require such manual effort.'}",https://arxiv.org/pdf/2103.17249
-null-text inversion for editing real images using guided diffusion models,Ron Mokady,"Recent large-scale text-guided diffusion models provide powerful image generation capabilities. Currently, a massive effort is given to enable the modification of these images using text only as means to offer intuitive and versatile editing tools. To edit a real image using these state-of-the-art tools, one must first invert the image with a meaningful text prompt into the pretrained model's domain. In this paper, we introduce an accurate inversion technique and thus facilitate an intuitive text-based modification of the image. Our proposed inversion consists of two key novel components: (i) Pivotal inversion for diffusion models. While current methods aim at mapping random noise samples to a single input image, we use a single pivotal noise vector for each timestamp and optimize around it. We demonstrate that a direct DDIM inversion is inadequate on its own, but does provide a rather good anchor for our optimization. (ii) Null-text optimization, where we only modify the unconditional textual embedding that is used for classifier-free guidance, rather than the input text embedding. This allows for keeping both the model weights and the conditional embedding intact and hence enables applying prompt-based editing while avoiding the cumbersome tuning of the model's weights. Our null-text inversion, based on the publicly available Stable Diffusion model, is extensively evaluated on a variety of images and various prompt editing, showing high-fidelity editing of real images.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an accurate inversion technique and facilitates an intuitive text-based modification of the image using the Stable Diffusion model, and proposes Null-text optimization, where the unconditional textual embedding is used for classifier-free guidance, rather than the input text embedding.'}",https://arxiv.org/pdf/2211.09794
-optimization of pgnaa set-up for the elements detection in aqueous solution,Wenbao Jia,,,
-clip-mesh: generating textured meshes from text using pretrained image-text models,N. Khalid,"We present a technique for zero-shot generation of a 3D model using only a target text prompt. Without any 3D supervision our method deforms the control shape of a limit subdivided surface along with its texture map and normal map to obtain a 3D asset that corresponds to the input text prompt and can be easily deployed into games or modeling applications. We rely only on a pre-trained CLIP model that compares the input text prompt with differentiably rendered images of our 3D model. While previous works have focused on stylization or required training of generative models we perform optimization on mesh parameters directly to generate shape, texture or both. To constrain the optimization to produce plausible meshes and textures we introduce a number of techniques using image augmentations and the use of a pretrained prior that generates CLIP image embeddings given a text embedding.","{'model': 'tldr@v2.0.0', 'text': 'A technique for zero-shot generation of a 3D model using only a target text prompt and a number of techniques using image augmentations and the use of a pretrained prior that generates CLIP image embeddings given a text embedding are presented.'}",https://arxiv.org/pdf/2203.13333
-an overview of variants and advancements of pso algorithm,Meetu Jain,"Particle swarm optimization (PSO) is one of the most famous swarm-based optimization techniques inspired by nature. Due to its properties of flexibility and easy implementation, there is an enormous increase in the popularity of this nature-inspired technique. Particle swarm optimization (PSO) has gained prompt attention from every field of researchers. Since its origin in 1995 till now, researchers have improved the original Particle swarm optimization (PSO) in varying ways. They have derived new versions of it, such as the published theoretical studies on various parameters of PSO, proposed many variants of the algorithm and numerous other advances. In the present paper, an overview of the PSO algorithm is presented. On the one hand, the basic concepts and parameters of PSO are explained, on the other hand, various advances in relation to PSO, including its modifications, extensions, hybridization, theoretical analysis, are included.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the PSO algorithm is presented, the basic concepts and parameters of PSO are explained, and various advances in relation to PSO, including its modifications, extensions, hybridization, theoretical analysis, are included.'}",https://www.mdpi.com/2076-3417/12/17/8392/pdf?version=1661250377
-dream3d: zero-shot text-to-3d synthesis using 3d shape prior and text-to-image diffusion models,Jiale Xu,"Recent CLIP-guided 3D optimization methods, such as DreamFields [19] and PureCLIPNeRF [24], have achieved impressive results in zero-shot text-to-3D synthesis. However, due to scratch training and random initialization without prior knowledge, these methods often fail to generate accurate and faithful 3D structures that conform to the input text. In this paper, we make the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process. Specifically, we first generate a high-quality 3D shape from the input text in the text-to-shape stage as a 3D shape prior. We then use it as the initialization of a neural radiance field and optimize it with the full prompt. To address the challenging text-to-shape generation task, we present a simple yet effective approach that directly bridges the text and image modalities with a powerful text-to-image diffusion model. To narrow the style domain gap between the images synthesized by the text-to-image diffusion model and shape renderings used to train the image-to-shape generator, we further propose to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation. Our method, Dream3D, is capable of generating imaginative 3D content with superior visual quality and shape accuracy compared to state-of-the-art methods. Our project page is at https://bluestyle97.github.io/dream3d/.","{'model': 'tldr@v2.0.0', 'text': 'This paper makes the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process, and proposes to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation.'}",https://arxiv.org/pdf/2212.14704
-training diffusion models with reinforcement learning,Kevin Black,"Diffusion models are a class of flexible generative models trained with an approximation to the log-likelihood objective. However, most use cases of diffusion models are not concerned with likelihoods, but instead with downstream objectives such as human-perceived image quality or drug effectiveness. In this paper, we investigate reinforcement learning methods for directly optimizing diffusion models for such objectives. We describe how posing denoising as a multi-step decision-making problem enables a class of policy gradient algorithms, which we refer to as denoising diffusion policy optimization (DDPO), that are more effective than alternative reward-weighted likelihood approaches. Empirically, DDPO is able to adapt text-to-image diffusion models to objectives that are difficult to express via prompting, such as image compressibility, and those derived from human feedback, such as aesthetic quality. Finally, we show that DDPO can improve prompt-image alignment using feedback from a vision-language model without the need for additional data collection or human annotation. The project's website can be found at http://rl-diffusion.github.io .","{'model': 'tldr@v2.0.0', 'text': 'It is described how posing denoising as a multi-step decision-making problem enables a class of policy gradient algorithms, which are referred to as Denoising diffusion policy optimization (DDPO), that are more effective than alternative reward-weighted likelihood approaches.'}",https://arxiv.org/pdf/2305.13301
-delta denoising score,Amir Hertz,"We introduce Delta Denoising Score (DDS), a novel scoring function for text-based image editing that guides minimal modifications of an input image towards the content described in a target prompt. DDS leverages the rich generative prior of text-to-image diffusion models and can be used as a loss term in an optimization problem to steer an image towards a desired direction dictated by a text. DDS utilizes the Score Distillation Sampling (SDS) mechanism for the purpose of image editing. We show that using only SDS often produces non-detailed and blurry outputs due to noisy gradients. To address this issue, DDS uses a prompt that matches the input image to identify and remove undesired erroneous directions of SDS. Our key premise is that SDS should be zero when calculated on pairs of matched prompts and images, meaning that if the score is non-zero, its gradients can be attributed to the erroneous component of SDS. Our analysis demonstrates the competence of DDS for text based image-to-image translation. We further show that DDS can be used to train an effective zero-shot image translation model. Experimental results indicate that DDS outperforms existing methods in terms of stability and quality, highlighting its potential for real-world applications in text-based image editing.","{'model': 'tldr@v2.0.0', 'text': 'Delta Denoising Score is introduced, a novel scoring function for text-based image editing that guides minimal modifications of an input image towards the content described in a target prompt that can be used to train an effective zero-shot image translation model.'}",http://arxiv.org/pdf/2304.07090
-what changes can large-scale language models bring? intensive study on hyperclova: billions-scale korean generative pretrained transformers,Boseop Kim,"GPT-3 shows remarkable in-context learning ability of large-scale language models (LMs) trained on hundreds of billion scale data. Here we address some remaining issues less reported by the GPT-3 paper, such as a non-English LM, the performances of different sized models, and the effect of recently introduced prompt optimization on in-context learning. To achieve this, we introduce HyperCLOVA, a Korean variant of 82B GPT-3 trained on a Korean-centric corpus of 560B tokens. Enhanced by our Korean-specific tokenization, HyperCLOVA with our training configuration shows state-of-the-art in-context zero-shot and few-shot learning performances on various downstream tasks in Korean. Also, we show the performance benefits of prompt-based learning and demonstrate how it can be integrated into the prompt engineering pipeline. Then we discuss the possibility of materializing the No Code AI paradigm by providing AI prototyping capabilities to non-experts of ML by introducing HyperCLOVA studio, an interactive prompt engineering interface. Lastly, we demonstrate the potential of our methods with three successful in-house applications.","{'model': 'tldr@v2.0.0', 'text': 'The possibility of materializing the No Code AI paradigm by providing AI prototyping capabilities to non-experts of ML by introducing HyperCLOVA studio, an interactive prompt engineering interface is discussed and the performance benefits of prompt-based learning are shown and how it can be integrated into the prompt engineering pipeline.'}",https://aclanthology.org/2021.emnlp-main.274.pdf
-directed diffusion: direct control of object placement through attention guidance,W. Ma,"Text-guided diffusion models such as DALLE-2, Imagen, eDiff-I, and Stable Diffusion are able to generate an effectively endless variety of images given only a short text prompt describing the desired image content. In many cases the images are of very high quality. However, these models often struggle to compose scenes containing several key objects such as characters in specified positional relationships. The missing capability to ``direct'' the placement of characters and objects both within and across images is crucial in storytelling, as recognized in the literature on film and animation theory. In this work, we take a particularly straightforward approach to providing the needed direction. Drawing on the observation that the cross-attention maps for prompt words reflect the spatial layout of objects denoted by those words, we introduce an optimization objective that produces ``activation'' at desired positions in these cross-attention maps. The resulting approach is a step toward generalizing the applicability of text-guided diffusion models beyond single images to collections of related images, as in storybooks. Directed Diffusion provides easy high-level positional control over multiple objects, while making use of an existing pre-trained model and maintaining a coherent blend between the positioned objects and the background. Moreover, it requires only a few lines to implement.","{'model': 'tldr@v2.0.0', 'text': 'Directed Diffusion provides easy high-level positional control over multiple objects, while making use of an existing pre-trained model and maintaining a coherent blend between the positioned objects and the background.'}",https://arxiv.org/pdf/2302.13153
-clip-actor: text-driven recommendation and stylization for animating human meshes,Youwang Kim,"We propose CLIP-Actor, a text-driven motion recommendation and neural mesh stylization system for human mesh animation. CLIP-Actor animates a 3D human mesh to conform to a text prompt by recommending a motion sequence and optimizing mesh style attributes. We build a text-driven human motion recommendation system by leveraging a large-scale human motion dataset with language labels. Given a natural language prompt, CLIP-Actor suggests a text-conforming human motion in a coarse-to-fine manner. Then, our novel zero-shot neural style optimization detailizes and texturizes the recommended mesh sequence to conform to the prompt in a temporally-consistent and pose-agnostic manner. This is distinctive in that prior work fails to generate plausible results when the pose of an artist-designed mesh does not conform to the text from the beginning. We further propose the spatio-temporal view augmentation and mask-weighted embedding attention, which stabilize the optimization process by leveraging multi-frame human motion and rejecting poorly rendered views. We demonstrate that CLIP-Actor produces plausible and human-recognizable style 3D human mesh in motion with detailed geometry and texture solely from a natural language prompt.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that CLIP-Actor produces plausible and human-recognizable style 3D human mesh in motion with detailed geometry and texture solely from a natural language prompt.'}",http://arxiv.org/pdf/2206.04382
-debiasing scores and prompts of 2d diffusion for robust text-to-3d generation,Susung Hong,"The view inconsistency problem in score-distilling text-to-3D generation, also known as the Janus problem, arises from the intrinsic bias of 2D diffusion models, which leads to the unrealistic generation of 3D objects. In this work, we explore score-distilling text-to-3D generation and identify the main causes of the Janus problem. Based on these findings, we propose two approaches to debias the score-distillation frameworks for robust text-to-3D generation. Our first approach, called score debiasing, involves gradually increasing the truncation value for the score estimated by 2D diffusion models throughout the optimization process. Our second approach, called prompt debiasing, identifies conflicting words between user prompts and view prompts utilizing a language model and adjusts the discrepancy between view prompts and object-space camera poses. Our experimental results show that our methods improve realism by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead.","{'model': 'tldr@v2.0.0', 'text': 'Two approaches to debias the score-distillation frameworks for robust text-to-3D generation are proposed and show that they improve realism by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead.'}",https://arxiv.org/pdf/2303.15413
-promptboosting: black-box text classification with ten forward passes,Bairu Hou,"We describe PromptBoosting, a query-efficient procedure for building a text classifier from a neural language model (LM) without access to the LM's parameters, gradients, or hidden representations. This form of""black-box""classifier training has become increasingly important as the cost of training and inference in large-scale LMs grows. But existing black-box LM classifier learning approaches are themselves computationally inefficient, typically specializing LMs to the target task by searching in a large space of (discrete or continuous) prompts using zeroth-order optimization methods. Instead of directly optimizing in prompt space, PromptBoosting obtains a small pool of prompts via a gradient-free approach and then constructs a large pool of weak learners by pairing these prompts with different elements of the LM's output distribution. These weak learners are then ensembled using the AdaBoost algorithm. The entire learning process requires only a small number of forward passes and no backward pass. Experiments show that PromptBoosting achieves state-of-the-art performance in multiple black-box few-shot classification tasks, and matches or outperforms full fine-tuning in both few-shot and standard learning paradigms, while training 10x faster than existing black-box methods.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that PromptBoosting achieves state-of-the-art performance in multiple black-box few- shot classification tasks, and matches or outperforms full fine-tuning in both few-shot and standard learning paradigms, while training 10x faster than existing black- box methods.'}",http://arxiv.org/pdf/2212.09257
-text to mesh without 3d supervision using limit subdivision,N. Khalid,"We present a technique for zero-shot generation of a 3D model using only a target text prompt. Without a generative model or any 3D supervision our method deforms a control shape of a limit subdivided surface along with a texture map and normal map to obtain a 3D model asset that matches the input text prompt and can be deployed into games or modeling applications. We rely only on a pre-trained CLIP model that compares the input text prompt with differen-tiably rendered images of our 3D model. While previous works have focused on stylization or required training of generative models we perform optimization on mesh parameters directly to generate shape and texture. To improve the quality of results we also introduce a set of techniques such as render augmentations, primitive selection, prompt aug-mentation that guide the mesh towards a suitable result.","{'model': 'tldr@v2.0.0', 'text': 'A technique for zero-shot generation of a 3D model using only a target text prompt and a set of techniques such as render augmentations, primitive selection, prompt aug-mentation that guide the mesh towards a suitable result are presented.'}",
-unleashing the power of visual prompting at the pixel level,Junyang Wu,"This paper presents a simple and effective visual prompting method for adapting pre-trained models to downstream recognition tasks. Our method includes two key designs. First, rather than directly adding together the prompt and the image, we treat the prompt as an extra and independent learnable component. We show that the strategy of reconciling the prompt and the image matters, and find that warping the prompt around a properly shrinked image empirically works the best. Second, we re-introduce two""old tricks""commonly used in building transferable adversarial examples, i.e., input diversity and gradient normalization, into visual prompting. These techniques improve optimization and enable the prompt to generalize better. We provide extensive experimental results to demonstrate the effectiveness of our method. Using a CLIP model, our prompting method sets a new record of 82.8% average accuracy across 12 popular classification datasets, substantially surpassing the prior art by +5.6%. It is worth noting that this prompting performance already outperforms linear probing by +2.1% and can even match fully fine-tuning in certain datasets. In addition, our prompting method shows competitive performance across different data scales and against distribution shifts. The code is publicly available at https://github.com/UCSC-VLAA/EVP.","{'model': 'tldr@v2.0.0', 'text': 'A simple and effective visual prompting method for adapting pre-trained models to downstream recognition tasks, which sets a new record of 82.8% average accuracy across 12 popular classification datasets, substantially surpassing the prior art by +5.6%.'}",http://arxiv.org/pdf/2212.10556
-reward collapse in aligning large language models,Ziang Song,"The extraordinary capabilities of large language models (LLMs) such as ChatGPT and GPT-4 are in part unleashed by aligning them with reward models that are trained on human preferences, which are often represented as rankings of responses to prompts. In this paper, we document the phenomenon of \textit{reward collapse}, an empirical observation where the prevailing ranking-based approach results in an \textit{identical} reward distribution \textit{regardless} of the prompts during the terminal phase of training. This outcome is undesirable as open-ended prompts like ``write a short story about your best friend'' should yield a continuous range of rewards for their completions, while specific prompts like ``what is the capital of New Zealand'' should generate either high or low rewards. Our theoretical investigation reveals that reward collapse is primarily due to the insufficiency of the ranking-based objective function to incorporate prompt-related information during optimization. This insight allows us to derive closed-form expressions for the reward distribution associated with a set of utility functions in an asymptotic regime. To overcome reward collapse, we introduce a prompt-aware optimization scheme that provably admits a prompt-dependent reward distribution within the interpolating regime. Our experimental results suggest that our proposed prompt-aware utility functions significantly alleviate reward collapse during the training of reward models.","{'model': 'tldr@v2.0.0', 'text': 'The theoretical investigation reveals that reward collapse is primarily due to the insufficiency of the ranking-based objective function to incorporate prompt-related information during optimization, which allows us to derive closed-form expressions for the reward distribution associated with a set of utility functions in an asymptotic regime.'}",http://arxiv.org/pdf/2305.17608
-structural and electrochemical aspects of lini0.8co0.1mn0.1o2 cathode materials doped by various cations,T. Weigel,"Ni-rich materials of layered structure LiNixCoyMnzO2, x > 0.5, are promising candidates as cathodes in high-energy-density Li-ion batteries for electric vehicles. The structural and cycling stability of Ni-rich cathodes can be remarkably improved by doping with a small amount of extrinsic multivalent cations. In this study, we examine development of a fast screening methodology for doping LiNi0.8Co0.1Mn0.1O2 with cations Mg2+, Al3+, Si4+, Ti4+, Zr4+, and Ta5+ by a “top-down” approach. The cathode material is coated by a precursor layer that contains the dopant, which then is introduced into the particles by diffusion during heat treatment at elevated temperatures. The methodology described herein can be applied to Ni-rich cathode materials and allows relatively easy and prompt identification of the most promising dopants. Then further optimization work can lead to development of high-capacity stable cathode materials. The present study marks Ta5+ cations as very promising dopants for Ni-rich NCM cathodes.",,
-"sepsis: the evolution in definition, pathophysiology, and management",B. Gyawali,"There has been a significant evolution in the definition and management of sepsis over the last three decades. This is driven in part due to the advances made in our understanding of its pathophysiology. There is evidence to show that the manifestations of sepsis can no longer be attributed only to the infectious agent and the immune response it engenders, but also to significant alterations in coagulation, immunosuppression, and organ dysfunction. A revolutionary change in the way we manage sepsis has been the adoption of early goal-directed therapy. This involves the early identification of at-risk patients and prompt treatment with antibiotics, hemodynamic optimization, and appropriate supportive care. This has contributed significantly to the overall improved outcomes with sepsis. Investigation into clinically relevant biomarkers of sepsis are ongoing and have yet to yield effective results. Scoring systems such as the sequential organ failure assessment and Acute Physiology and Chronic Health Evaluation help risk-stratify patients with sepsis. Advances in precision medicine techniques and the development of targeted therapy directed at limiting the excesses of the inflammatory and coagulatory cascades offer potentially viable avenues for future research. This review summarizes the progress made in the diagnosis and management of sepsis over the past two decades and examines promising avenues for future research.","{'model': 'tldr@v2.0.0', 'text': 'Advances in precision medicine techniques and the development of targeted therapy directed at limiting the excesses of the inflammatory and coagulatory cascades offer potentially viable avenues for future research.'}",https://journals.sagepub.com/doi/pdf/10.1177/2050312119835043
-being comes from not-being: open-vocabulary text-to-motion generation with wordless training,Junfan Lin,"Text-to-motion generation is an emerging and challenging problem, which aims to synthesize motion with the same semantics as the input text. However, due to the lack of diverse labeled training data, most approaches either limit to specific types of text annotations or require online optimizations to cater to the texts during inference at the cost of efficiency and stability. In this paper, we investigate offline open-vocabulary text-to-motion generation in a zero-shot learning manner that neither requires paired training data nor extra online optimization to adapt for unseen texts. Inspired by the prompt learning in NLP, we pretrain a motion generator that learns to reconstruct the full motion from the masked motion. During inference, instead of changing the motion generator, our method reformulates the input text into a masked motion as the prompt for the motion generator to “reconstruct” the motion. In constructing the prompt, the unmasked poses of the prompt are synthesized by a text-to-pose generator. To supervise the optimization of the text-to-pose generator, we propose the first text-pose alignment model for measuring the alignment between texts and 3D poses. And to prevent the pose generator from over-fitting to limited training texts, we further propose a novel wordless training mechanism that optimizes the text-to-pose generator without any training texts. The comprehensive experimental results show that our method obtains a significant improvement against the baseline methods. The code is available at https://github.com/junfanlin/oohmg.","{'model': 'tldr@v2.0.0', 'text': 'This paper pretrain a motion generator that learns to reconstruct the full motion from the masked motion, inspired by the prompt learning in NLP, and proposes the first text-pose alignment model for measuring the alignment between texts and 3D poses.'}",https://arxiv.org/pdf/2210.15929
-management of acute kidney injury: core curriculum 2018.,P. K. Moore,"Acute kidney injury (AKI) is a heterogeneous disorder that is common in hospitalized patients and associated with short- and long-term morbidity and mortality. When AKI is present, prompt workup of the underlying cause should be pursued, with specific attention to reversible causes. Measures to prevent AKI include optimization of volume status and avoidance of nephrotoxic medications. Crystalloids are preferred over colloids for most patients, and hydroxyethyl starches should be avoided. Volume overload in the setting of AKI is associated with adverse outcomes, so attention should be paid to overall fluid balance. Currently there are no targeted pharmacotherapies approved for the treatment of AKI. The optimal timing of renal replacement therapy in critically ill patients with AKI is unclear, but is an area of active investigation. Recent studies suggest that AKI is not a ""self-limited"" process, but is strongly linked to increased risk for chronic kidney disease, subsequent AKI, and future mortality.","{'model': 'tldr@v2.0.0', 'text': 'Current studies suggest that AKI is not a ""self-limited"" process, but is strongly linked to increased risk for chronic kidney disease, subsequent AKI, and future mortality.'}",http://www.ajkd.org/article/S0272638617311411/pdf
-a top-down method for performance analysis and counters architecture,Ahmad Yasin,"Optimizing an application's performance for a given microarchitecture has become painfully difficult. Increasing microarchitecture complexity, workload diversity, and the unmanageable volume of data produced by performance tools increase the optimization challenges. At the same time resource and time constraints get tougher with recently emerged segments. This further calls for accurate and prompt analysis methods. The insights from this method guide a proposal for a novel performance counters architecture that can determine the true bottlenecks of a general out-of-order processor. Unlike other approaches, our analysis method is low-cost and already featured in in-production systems - it requires just eight simple new performance events to be added to a traditional PMU. It is comprehensive - no restriction to predefined set of performance issues. It accounts for granular bottlenecks in super-scalar cores, missed by earlier approaches.","{'model': 'tldr@v2.0.0', 'text': 'This analysis method is low-cost and already featured in in-production systems - it requires just eight simple new performance events to be added to a traditional PMU, and accounts for granular bottlenecks in super-scalar cores, missed by earlier approaches.'}",
-research on path planning of mobile robot based on improved a* algorithm,Xiu-hong Wang,,"{'model': 'tldr@v2.0.0', 'text': 'The simulated results indicate that the improved A* algorithm can achieve the ideal effect of reducing and improving the search rate to finally locate the optimal path with precise search results and prompt response.'}",
-fluid management for the prevention and attenuation of acute kidney injury,J. Prowle,,"{'model': 'tldr@v2.0.0', 'text': 'The conflict between the desire to achieve adequate resuscitation of shock and the need to mitigate the harmful effects of fluid overload is discussed.'}",
-risk-aware identification of highly suspected covid-19 cases in social iot: a joint graph theory and reinforcement learning approach,Bowen Wang,"The recent outbreak of the coronavirus disease 2019 (COVID-19) has rapidly become a pandemic, which calls for prompt action in identifying suspected cases at an early stage through risk prediction. To suppress its further spread, we exploit the social relationships between mobile devices in the Social Internet of Things (SIoT) to help control its propagation by allocating the limited protective resources to the influential so-called high-degree individuals to stem the tide of precipitated spreading. By exploiting the so-called differential contact intensity and the infectious rate in susceptible-exposed-infected-removed (SEIR) epidemic model, the resultant optimization problem can be transformed into the minimum weight vertex cover (MWVC) problem of graph theory. To solve this problem in a high-dynamic random network topology, we propose an adaptive scheme by relying on the graph embedding technique during the state representation and reinforcement learning in the training phase. By relying on a pair of real-life datasets, the results demonstrate that our scheme can beneficially reduce the epidemiological reproduction rate of the infection. This technique has the potential of assisting in the early identification of COVID-19 cases.","{'model': 'tldr@v2.0.0', 'text': 'An adaptive scheme by relying on the graph embedding technique during the state representation and reinforcement learning in the training phase can beneficially reduce the epidemiological reproduction rate of the infection.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/09121230.pdf
-hydra: an ensemble of convolutional neural networks for geospatial land classification,R. Minetto,"In this paper, we describe Hydra, an ensemble of convolutional neural networks (CNNs) for geospatial land classification. The idea behind Hydra is to create an initial CNN that is coarsely optimized but provides a good starting pointing for further optimization, which will serve as the Hydra’s body. Then, the obtained weights are fine-tuned multiple times with different augmentation techniques, crop styles, and classes weights to form an ensemble of CNNs that represent the Hydra’s heads. By doing so, we prompt convergence to different endpoints, which is a desirable aspect for ensembles. With this framework, we were able to reduce the training time while maintaining the classification performance of the ensemble. We created ensembles for our experiments using two state-of-the-art CNN architectures, residual network (ResNet), and dense convolutional networks (DenseNet). We have demonstrated the application of our Hydra framework in two data sets, functional map of world (FMOW) and NWPU-RESISC45, achieving results comparable to the state-of-the-art for the former and the best-reported performance so far for the latter. Code and CNN models are available at https://github.com/maups/hydra-fmow.","{'model': 'tldr@v2.0.0', 'text': 'The idea behind Hydra is to create an initial CNN that is coarsely optimized but provides a good starting pointing for further optimization, which will serve as the Hydra’s body, an ensemble of convolutional neural networks (CNNs) for geospatial land classification.'}",
-mechanisms of cardiac repair and regeneration.,K. Broughton,"Cardiovascular regenerative therapies are pursued on both basic and translational levels. Although efficacy and value of cell therapy for myocardial regeneration can be debated, there is a consensus that profound deficits in mechanistic understanding limit advances, optimization, and implementation. In collaboration with the TACTICS (Transnational Alliance for Regenerative Therapies in Cardiovascular Syndromes), this review overviews several pivotal aspects of biological processes impinging on cardiac maintenance, repair, and regeneration. The goal of summarizing current mechanistic understanding is to prompt innovative directions for fundamental studies delineating cellular reparative and regenerative processes. Empowering myocardial regenerative interventions, whether dependent on endogenous processes or exogenously delivered repair agents, ultimately depends on mastering mechanisms and novel strategies that take advantage of rather than being limited by inherent myocardial biology.","{'model': 'tldr@v2.0.0', 'text': 'Current mechanistic understanding of biological processes impinging on cardiac maintenance, repair, and regeneration is summarized to prompt innovative directions for fundamental studies delineating cellular reparative and regenerative processes.'}",https://www.ahajournals.org/doi/pdf/10.1161/CIRCRESAHA.117.312586
-timetabling optimization of a single railway track line with sensitivity analysis,Enrique Castillo,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that an adequate selection of intermediate stations and of the departure times are crucial in the good performance of the line and that inadequate spacings between consecutive trains can block the line.'}",
-liver mr elastography technique and image interpretation: pearls and pitfalls.,F. Guglielmo,"Liver MR elastography is an imaging technique used to measure liver stiffness in the evaluation for possible fibrosis or cirrhosis. Liver stiffness measurement (LSM) is useful for predicting the stage of liver fibrosis. However, obtaining and reporting accurate and reliable LSMs with MR elastography require an understanding of the three core components of liver MR elastography: optimization of imaging technique, prompt quality control of images, and proper interpretation and reporting of elastogram findings. When performing MR elastography, six important technical parameters that should be optimized are patient fasting before the examination, proper passive driver placement, proper MR elastography section positioning over the largest area of the liver, use of MR elastography-related sequences at end expiration, choosing the best timing of the MR elastography sequence, and optimization of several essential pulse sequence parameters. As soon as the MR elastography examination is performed, the elastograms should be reviewed to ensure that they are of diagnostic quality so that corrective steps can be taken, if needed, and MR elastography can be repeated before the diagnostic portion of the examination concludes. Finally, the interpreting radiologist needs to understand and be able to perform the proper technique for LSMs, including determining which areas of the liver to include or avoid in the measurements; knowing which conditions, other than fibrosis or cirrhosis, can increase liver stiffness; and understanding how to report elastography results. This article reviews the proper technique for performing liver MR elastography and subsequent quality control assessment, as well as the principles for interpreting and reporting studies. This review may be helpful for implementing and operating a clinical liver MR elastography service.The full digital presentation is available online.©RSNA, 2019.","{'model': 'tldr@v2.0.0', 'text': 'The proper technique for performing liver MR elastography and subsequent quality control assessment, as well as the principles for interpreting and reporting studies are reviewed, may be helpful for implementing and operating a clinical liver MRElastography service.'}",
-single image dehazing with dark channel prior,J. Kaur,"In applications related to computation of visuals, the clear picture is very crucial component but certain atmospheric conditions like haze and fog prompt the suspended particles which lead to degradation of the image. This hinders the flow of vision systems. Therefore, implementation of image dehazing has become important to improve the throughput and effectiveness of such vision systems. This paper sheds light on execution of dark channel to delete haze from single image. Different measures to optimize this algorithm are discussed with results obtained from certain successful attempts to implement this algorithm.","{'model': 'tldr@v2.0.0', 'text': 'Light is shed on execution of dark channel to delete haze from single image and different measures to optimize this algorithm are discussed with results obtained from certain successful attempts to implement this algorithm.'}",
-adaptive cubic overestimation methods for unconstrained optimization,C. Cartis,"An Adaptive Cubic Overestimation (ACO) algorithm for unconstrained optimization, generalizing a method due to Nesterov & Polyak (Math. Programming 108, 2006, pp 177-205), is proposed. At each iteration of Nesterov & Polyak's approach, the global minimizer of a local cubic overestimator of the objective function is determined, and this ensures a significant improvement in the objective so long as the Hessian of the objective is Lipschitz continuous and its Lipschitz constant is available. The twin requirements of global model optimality and the availability of Lipschitz constants somewhat limit the applicability of such an approach, particularly for large-scale problems. However the promised powerful worst-case theoretical guarantees prompt us to investigate variants in which estimates of the required Lipschitz constant are refined and in which computationally-viable approximations to the global model-minimizer are sought. We show that the excellent global and local convergence properties and worst-case iteration complexity bounds obtained by Nesterov & Polyak are retained, and sometimes extended to a wider class of problems, by our ACO approach. Numerical experiments with small-scale test problems from the CUTEr set show superior performance of the ACO algorithm when compared to a trust-region implementation.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the excellent global and local convergence properties and worst-case iteration complexity bounds obtained by Nesterov & Polyak are retained, and sometimes extended to a wider class of problems, by the ACO approach.'}",
-multi-swarm particle swarm optimization based on mixed search behavior,Jing Jie,"The paper develops a Multi-swarm particle swarm optimization (MPSO) to overcome the premature convergence problem. MPSO takes advantage of multiple sub-swarms with mixed search behavior to maintain the swarm diversity, and introduces cooperative mechanism to prompt the information exchange among sub-swarms. Moreover, MPSO adopts an adaptive reinitializing strategy guided by swarm diversity, which can contribute to the global convergence of the algorithm. Through the mixed local search behavior modes, the cooperative search and the reinitializing strategy guided by swarm diversity, MPSO can maintain appropriate diversity and keep the balance of local search and global search validly. The proposed MPSO was applied to some well-known benchmarks. The experimental results show MPSO is a robust global optimization technique for the complex multimodal functions.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show MPSO is a robust global optimization technique for the complex multimodal functions and can maintain appropriate diversity and keep the balance of local search and global search validly.'}",
-seeking for sterile neutrinos with displaced leptons at the lhc,Jia Liu,,,https://link.springer.com/content/pdf/10.1007/JHEP07(2019)159.pdf
-handover parameter optimization in wcdma using fuzzy controlling,C. Werner,"The increasing traffic and the demand for high data rate services in WCDMA networks prompt the need for an automatic network optimization. Current state-of-the-art optimization methods adapt physical parameters, such as the antennas' tilt or azimuth. As agreed in the community, a further capacity increase can be achieved by an optimization of radio resource management (RRM) parameters. While theoretical approaches of RRM parameter optimization have already been introduced in the literature, practical solutions optimizing cell individual parameters have rarely been treated so far. Consequently, this paper copes with the optimization of the cell individual offset (CIO) in the soft handover algorithm with the goal to reduce the network's outage probability. We design and apply a fuzzy logic controller, whose output are CIO changes matching the current traffic and load conditions. Verifications using a dynamic system simulator prove the powerfulness of our method and promise significant outage reductions in a fully loaded network scenario leading to capacity improvements of up to 9 %.","{'model': 'tldr@v2.0.0', 'text': ""This paper copes with the optimization of the cell individual offset (CIO) in the soft handover algorithm with the goal to reduce the network's outage probability and promises significant outage reductions in a fully loaded network scenario leading to capacity improvements of up to 9 %.""}",
-optimization of injected dose based on noise equivalent count rates for 2- and 3-dimensional whole-body pet.,C. Lartizien,"UNLABELLED The noise equivalent count (NEC) rate index is used to derive guidelines on the optimal injected dose to the patient for 2-dimensional (2D) and 3-dimensional (3D) whole-body PET acquisitions. METHODS We performed 2D and 3D whole-body acquisitions of an anthropomorphic phantom modeling the conditions for (18)F-FDG PET of the torso and measured the NEC rates for different activity levels for several organs of interest. The correlations between count rates measured from the phantom and those from a series of whole-body patient scans were then analyzed. This analysis allowed validation of our approach and estimation of the injected dose that maximizes NEC rate as a function of patient morphology for both acquisition modes. RESULTS Variations of the phantom and patient prompt and random coincidence rates as a function of single-photon rates correlated well. On the basis of these correlations, we demonstrated that the patient NEC rate can be predicted for a given single-photon rate. Finally, we determined that patient single-photon rates correlated with the mean dose per weight at acquisition start when normalized by the body mass index. This correlation allows modifying the injected dose as a function of patient body mass index to reach the peak NEC rate in 3D mode. Conversely, we found that the peak NEC rates were never reached in 2D mode within an acceptable range of injected dose. CONCLUSION The injected dose was adapted to patient morphology for 2D and 3D whole-body acquisitions using the NEC rate as a figure of merit of the statistical quality of the sinogram data. This study is a first step toward a more comprehensive comparison of the image quality obtained using both acquisition modes.","{'model': 'tldr@v2.0.0', 'text': 'The injected dose was adapted to patient morphology for 2D and 3D whole-body acquisitions using the NEC rate as a figure of merit of the statistical quality of the sinogram data as a first step toward a more comprehensive comparison of the image quality obtained using both acquisition modes.'}",
-bleach optimization of sputum smear microscopy for pulmonary tuberculosis.,P. Srikanth,"The Revised National Tuberculosis Control Programme (RNTCP) aims to improve case detection rates of tuberculosis to facilitate prompt recognition and treatment. The low case detection rates in the programme can be directly attributed to failure to screen patients with suspected tuberculosis and the low sensitivity of the direct smear microscopy method to detect cases among the fraction of patients that are screened. Apart from low sensitivity, this method also has other disadvantages including the increased risk of infection transmission to technicians. There are several methods that can be used to improve sensitivity, but their applicability in a national programme and in resource limited settings are limited. Bleach processing of sputum smears prior to microscopy may be a cheap and effective way to improve on the sensitivity of the direct smear. Four distinctive techniques of sputum smear processing using bleach are described in the review, with the variations in each technique, along with the sensitivity. An analysis of reports published earlier on the bleach method is also presented including a discussion on when and why the bleach method works.","{'model': 'tldr@v2.0.0', 'text': 'Four distinctive techniques of sputum smear processing using bleach are described in the review, with the variations in each technique, along with the sensitivity.'}",
-optimization of the tem feed structure for four-arm reflector impulse radiating antennas,J. Tyo,"This paper considers the optimization of the feed arm geometry of four-arm crossed-coplanar plate impulse radiating antennas (IRAs) when the angular position and extent of the arms are taken as free parameters. Previously, optimization of this class of antenna considered only the symmetric case where the two pairs of crossed feed arms were perpendicular to each other. Comparison is made using the prompt aperture efficiency, and the results indicate that the efficiency of four-arm IRAs can be increased from /spl sim/25% for the perpendicularly crossed arms to /spl sim/35% for the optimum configuration. In addition to the optimization, the feed impedance of coplanar feeds is presented for general values of feed arm angle and plate width, and the optimum feed impedance is computed for each feed arm angle. The results can be used to design the optimal four-arm IRA with an arbitrary specified input impedance.",,
-fast dynamic routing based on weighted kernel density estimation,Suofei Zhang,"Capsules as well as dynamic routing between them are most recently proposed structures for deep neural networks. A capsule groups data into vectors or matrices as poses rather than conventional scalars to represent specific properties of target instance. Based on pose, a capsule should be attached to a probability (often denoted as activation) for its presence. The dynamic routing helps capsule network achieve more generalization capacity with fewer model parameters. However, the bottleneck, which prevents widespread applications of capsule, is the expense of computation during routing. To address this problem, we generalize existing routing methods within the framework of weighted kernel density estimation, proposing two fast routing methods with different optimization strategies. Our methods prompt the time efficiency of routing by nearly 40% with negligible performance degradation. By stacking a hybrid of convolutional layers and capsule layers, we construct a network architecture to handle inputs at a resolution of 64 × 64 pixels. The proposed models achieve a parallel performance with other leading methods in multiple benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This work generalizes existing routing methods within the framework of weighted kernel density estimation, proposing two fast routing methods with different optimization strategies that prompt the time efficiency of routing by nearly 40% with negligible performance degradation.'}",https://arxiv.org/pdf/1805.10807
-optimization of steady-state beam-driven tokamak reactors.,David R. Mikkelsen,"Recent developments in neutral beam technology prompt us to reconsider the prospects for steady-state tokamak reactors. A mathematical reactor model is developed that includes the physics of beam-driven currents and reactor power balance, as well as reactor and beam system costs. This model is used to find the plasma temperatures that minimize the reactor cost per unit of net electrical output. The optimum plasma temperatures are nearly independent of ..beta.. and are roughly twice as high as the optimum temperatures for ignited reactors. If beams of neutral deuterium atoms with near-optimum energies of 1 to 2 MeV are used to drive the current in a reactor the size of the International Tokamak Reactor, then the optimum temperatures are typically T /SUB e/ approx. = 12 to 15 keV and T /SUB i/ approx. = 17 to 21 keV for a wide range of model parameters. Net electrical output rises rapidly with increasing deuterium beam energy for E /SUB b/ less than or equal to 400 keV, but rises only slowly above E /SUB b/ about 1 MeV. We estimate that beam-driven steady-state reactors could be economically competitive with pulsed-ignition reactors if cyclic-loading problems limit the toroidal magnetic field strength ofmore » pulsed reactors to less than or equal to 85% of that allowed in steady-state reactors.« less",,https://digital.library.unt.edu/ark:/67531/metadc1192439/m2/1/high_res_d/6772488.pdf
-seasonal food scarcity prompts long-distance foraging by a wild social bee,Nathaniel S. Pope,"Foraging is an essential process for mobile animals, and its optimization serves as a foundational theory in ecology and evolution; however, drivers of foraging are rarely investigated across landscapes and seasons. Using a common bumblebee species from the western United States (Bombus vosnesenskii), we ask whether seasonal decreases in food resources prompt changes in foraging behavior and space use. We employ a unique integration of population genetic tools and spatially explicit foraging models to estimate foraging distances and rates of patch visitation for wild bumblebee colonies across three study regions and two seasons. By mapping the locations of 669 wild-caught individual foragers, we find substantial variation in colony-level foraging distances, often exhibiting a 60-fold difference within a study region. Our analysis of visitation rates indicates that foragers display a preference for destination patches with high floral cover and forage significantly farther for these patches, but only in the summer, when landscape-level resources are low. Overall, these results indicate that an increasing proportion of long-distance foraging bouts take place in the summer. Because wild bees are pollinators, their foraging dynamics are of urgent concern, given the potential impacts of global change on their movement and services. The behavioral shift toward long-distance foraging with seasonal declines in food resources suggests a novel, phenologically directed approach to landscape-level pollinator conservation and greater consideration of late-season floral resources in pollinator habitat management.","{'model': 'tldr@v2.0.0', 'text': 'The behavioral shift toward long-distance foraging with seasonal declines in food resources suggests a novel, phenologically directed approach to landscape-level pollinator conservation and greater consideration of late-season floral resources in pollinator habitat management.'}",
-reinfarction following anesthesia in patients with myocardial infarction,T. Rao,"The authors studied the incidence of and factors related to recurrent perioperative myocardial infarction retrospectively during 1973–1976 (Group 1) and prospectively during 1977–1982 (Group 2). Reinfarction occurred in 28 of 364 (7.7%) patients in Group 1 and 14 of 733 (1.9%) in Group 2 (P < 0.005). When the previous infarction was 0–3 and 4–6 months old, perioperative reinfarction occurred in 36% and 26% of Group 1 patients, respectively, and only 5.7% and 2.3% of Group 2 patients, respectively, (P < 0.05). In both groups, patients with associated congestive heart failure had a higher reinfarction rate. Patients who had intraoperative hypertension and tachycardia or hypotension develop had a higher incidence of reinfarction in both groups. The results suggest that preoperative optimization of the patient's status, aggressive invasive monitoring of the hemodynamic status, and prompt treatment of any hemodynamic aberration may be associated with decreased perioperative morbidity and mortality in patients with previous myocardial infarction. Which of these factors, if any, contributed to the improved outcome was not determined in this study.","{'model': 'tldr@v2.0.0', 'text': ""The results suggest that preoperative optimization of the patient's status, aggressive invasive monitoring of the hemodynamic status, and prompt treatment of any hemodynamic aberration may be associated with decreased perioperative morbidity and mortality in patients with previous myocardial infarction.""}",https://pubs.asahq.org/anesthesiology/article-pdf/59/6/499/626415/0000542-198312000-00003.pdf
-feasibility of portfolio optimization under coherent risk measures,I. Kondor,"It is shown that the axioms for coherent risk measures imply that whenever there is an asset in a portfolio that dominates the others in a given sample (which happens with finite probability even for large samples), then this portfolio cannot be optimized under any coherent measure on that sample, and the risk measure diverges to minus infinity. This instability was first discovered on the special example of Expected Shortfall which is used here both as an illustration and as a prompt for generalization.",,
-unhealthy stem cells: when health conditions upset stem cell properties,Laura M Pérez,"The stem cell field has grown very rapidly during the last decade, offering the promise of innovative therapies to treat disease. Different stem cell populations have been isolated from various human adult tissues, mainly from bone marrow and adipose tissue, but many other body tissues harbor a stem cell population. Adult tissue stem cells are invariably found in discrete microenvironments termed niches, where they play key roles in tissue homeostasis by enabling lifelong optimization of organ form and function. Some diseases are known to strike at the stem cell population, through alterations in their specific microenvironments, making them non-viable. Furthermore, it has been shown that a transformed stem cell population could prompt the development of certain cancers. This review focuses on the potential negative aspects of a range of diseases on the activity of stem cells and how their potential use in cell therapies may be affected.","{'model': 'tldr@v2.0.0', 'text': 'This review focuses on the potential negative aspects of a range of diseases on the activity of stem cells and how their potential use in cell therapies may be affected.'}",https://www.karger.com/Article/Pdf/489440
-a randomized controlled trial of an optimized smoking treatment delivered in primary care.,Megan E. Piper,"Background The effectiveness of smoking cessation treatment is limited in real-world use, perhaps because we have not selected the components of such treatments optimally nor have treatments typically been developed for and evaluated in real-world clinical settings. Purpose To validate an optimized smoking cessation treatment package that comprises intervention components identified as effective in factorial screening experiments conducted as per the Multiphase Optimization Strategy (MOST). Methods Adult smokers motivated to quit were recruited from primary care clinics (N = 623). Participants were randomized to receive either recommended usual care (R-UC; 10 min of in-person counseling, 8 weeks of nicotine patch, and referral to quitline services) or abstinence-optimized treatment (A-OT; 3 weeks of prequit mini-lozenges, 26 weeks of nicotine patch + mini-lozenges, three in-person and eight phone counseling sessions, and 7-11 automated calls to prompt medication use). The key outcomes were self-reported and biochemically confirmed (carbon monoxide, CO <6 ppm) 7-day point-prevalence abstinence. Results A-OT participants had significantly higher self-reported abstinence rates than R-UC participants at 4, 8, 16, and 26 weeks (ORs: 1.91-3.05; p <. 001). The biochemically confirmed 26-week abstinence rates were lower than the self-reported 26-week rates, but revealed a similar treatment effect size (OR = 2.94, p < .001). There was no moderation of treatment effects on 26-week abstinence by demographic, psychiatric, or nicotine dependence variables. A-OT had an incremental cost-effectiveness ratio for 26-week CO-confirmed abstinence of $7,800. Conclusions A smoking cessation treatment that is optimized via MOST development meaningfully enhances cessation rates beyond R-UC smoking treatment in smokers seen in primary care. Clinical Trial Registration NCT02301403.","{'model': 'tldr@v2.0.0', 'text': 'A smoking cessation treatment that is optimized via MOST development meaningfully enhances cessation rates beyond R-UC smoking treatment in smokers seen in primary care.'}",
-optimization models for streaming in multihop wireless networks,Zongpeng Li,"Wireless spectrum is a scare resource, while media streaming usually requires high end-to-end bandwidth. Media streaming in wireless ad hoc networks is therefore a particularly challenging problem, especially for the case of streaming to multiple receivers. In this paper, we design linear optimization models for computing a high-bandwidth routing strategy for media multicast in wireless networks, which targets near-optimal throughput, given constraints including network topology, radio capacity, and link contention. We study both the directional antenna and omni-directional antenna cases and point out their connections. We also combine the classic forward error correction techniques with the novel network coding techniques to provide error control in a timely fashion. Simulation results show that our solutions indeed achieve high streaming rates, and prompt error recovery under a wide range of link failure patterns.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs linear optimization models for computing a high-bandwidth routing strategy for media multicast in wireless networks, which targets near-optimal throughput, given constraints including network topology, radio capacity, and link contention.'}",http://www.eecg.toronto.edu/~bli/papers/zp-icccn07.pdf
-time domain compressive beam forming of ultrasound signals.,G. David,"Ultrasound imaging is a wide spread technique used in medical imaging as well as in non-destructive testing. The technique offers many advantages such as real-time imaging, good resolution, prompt acquisition, ease of use, and low cost compared to other techniques such as x-ray imaging. However, the maximum frame rate achievable is limited as several beams must be emitted to compute a single image. For each emitted beam, one must wait for the wave to propagate back and forth, thus imposing a limit to the frame rate. Several attempts have been made to use less beams while maintaining image quality. Although efficiently increasing the frame rate, these techniques still use several transmit beams. Compressive Sensing (CS), a universal data completion scheme based on convex optimization, has been successfully applied to a number of imaging modalities over the past few years. Using a priori knowledge of the signal, it can compute an image using less data allowing for shorter acquisition times. In this paper, it is shown that a valid CS framework can be derived from ultrasound propagation theory, and that this framework can be used to compute images of scatterers using only one plane wave as a transmit beam.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a valid CS framework can be derived from ultrasound propagation theory, and that this framework could be used to compute images of scatterers using only one plane wave as a transmit beam.'}",https://pubs.aip.org/asa/jasa/article-pdf/137/5/2773/15314040/2773_1_online.pdf
-independent tasks scheduling based on genetic algorithm in cloud computing,C. Zhao,"Task scheduling algorithm, which is an NP-completeness problem, plays a key role in cloud computing systems. In this paper, we propose an optimized algorithm based on genetic algorithm to schedule independent and divisible tasks adapting to different computation and memory requirements. We prompt the algorithm in heterogeneous systems, where resources (including CPUs) are of computational and communication heterogeneity. Dynamic scheduling is also in consideration. Though GA is designed to solve combinatorial optimization problem, it's inefficient for global optimization. So we conclude with further researches in optimized genetic algorithm.","{'model': 'tldr@v2.0.0', 'text': ""Though GA is designed to solve combinatorial optimization problem, it's inefficient for global optimization, so this paper concludes with further researches in optimized genetic algorithm.""}",
-from macrohemodynamic to the microcirculation,A. Donati,"ICU patients need a prompt normalization of macrohemodynamic parameters. Unfortunately, this optimization sometimes does not protect patients from organ failure development. Prevention or treatment of organ failure needs another target to be pursued: the microcirculatory restoration. Microcirculation is the ensemble of vessels of maximum 100 μm in diameter. Nowadays the Sidestream Dark Field (SDF) imaging technique allows its bedside investigation and a recent round-table conference established the criteria for its evaluation. First, microcirculatory derangements have been studied in sepsis: they are mainly characterized by a reduction of vessel density, an alteration of flow, and a heterogeneous distribution of perfusion. Endothelial malfunction and glycocalyx rupture were proved to be the main reasons for the observed microthrombi, capillary leakage, leukocyte rolling, and rouleaux phenomenon, even if further studies are necessary for a better explanation. Therapeutic approaches targeting microcirculation are under investigation. Microcirculatory alterations have been recently demonstrated in other diseases such as hypovolemia and cardiac failure but this issue still needs to be explored. The aim of this paper is to gather the already known information, focus the reader's attention on the importance of microvascular physiopathology in critical illness, and prompt him to actively participate to achieve a more comprehensive understanding of the issue.","{'model': 'tldr@v2.0.0', 'text': ""The aim of this paper is to gather the already known information, focus the reader's attention on the importance of microvascular physiopathology in critical illness, and prompt him to actively participate to achieve a more comprehensive understanding of the issue.""}",https://downloads.hindawi.com/journals/ccrp/2013/892710.pdf
-contemporary mechanical circulatory support therapy for postcardiotomy shock,S. Fukuhara,,"{'model': 'tldr@v2.0.0', 'text': 'The reported in-hospital mortality of the PCS patients remains high, consistently over 50\xa0%, despite ongoing refinements of MCS technology, and the optimization of selection criteria and the prompt institution of M CS are likely the keys to improving this persistently high mortality rate.'}",
-fragment-to-hit-to-lead discovery of a novel pyridylurea scaffold of atp competitive dual targeting type ii topoisomerase inhibiting antibacterial agents.,G. Basarab,"The discovery and optimization of a new class of bacterial topoisomerase (DNA gyrase and topoisomerase IV) inhibitors binding in the ATP domain are described. A fragment molecule, 1-ethyl-3-(2-pyridyl)urea, provided sufficiently potent enzyme inhibition (32 μM) to prompt further analogue work. Acids and acid isosteres were incorporated at the 5-pyridyl position of this fragment, bridging to a key asparagine residue, improving enzyme inhibition, and leading to measurable antibacterial activity. A CF3-thiazole substituent at the 4-pyridyl position improved inhibitory potency due to a favorable lipophilic interaction. Promising antibacterial activity was seen versus the Gram-positive pathogens Staphylococcus aureus and Streptococcus pneumoniae and the Gram-negative pathogens Haemophilus influenzae and Moraxella catarrhalis . Precursor metabolite incorporation and mutant analysis studies support the mode-of-action, blockage of DNA synthesis by dual target topoisomerase inhibition. Compound 35 was efficacious in a mouse S. aureus disease model, where a 4.5-log reduction in colony forming units versus control was demonstrated.","{'model': 'tldr@v2.0.0', 'text': 'Preursor metabolite incorporation and mutant analysis studies support the mode-of-action, blockage of DNA synthesis by dual target topoisomerase inhibition.'}",
-enhanced extracellular polysaccharide production and self-sustainable electricity generation for pamfcs by scenedesmus sp. sb1,M. Angelaalincy,"In this study, a freshwater microalga, Scenedesmus sp. SB1, was isolated, purified, and identified by its internal transcribed spacer region (ITS1-5.8S-ITS2). Media optimization through the Plackett–Burman Design and response surface methodology (RSM) showed a maximum exopolysaccharide (EPS) production of 48 mg/L (1.8-fold higher than that for unoptimized media). Characterization using gas chromatography–mass spectrometry, Fourier transform infrared, X-ray diffraction, and thermogravimetric analysis reveals that the EPS is a sulfated pectin polysaccharide with a crystallinity index of 15.2% and prompt thermal stability. Furthermore, the photoelectrogenic activity of Scenedesmus sp. SB1 inoculated in BG-11 and RSM-optimized BG-11 (ROBG-11) media was tested by cyclic voltammogram studies, revealing the potential of the inoculated strain in ROBG-11 toward photosynthetic algal microbial fuel cells over normal BG-11. To the best of our knowledge, functional group characterization, physical and thermal property and media optimization for EPS production by RSM and electrogenic activity studies are reported for the first time in Scenedesmus sp. SB1.","{'model': 'tldr@v2.0.0', 'text': 'Functional group characterization, physical and thermal property and media optimization for EPS production by RSM and electrogenic activity studies are reported for the first time in Scenedesmus sp.'}",https://pubs.acs.org/doi/pdf/10.1021/acsomega.7b00326
-transcriptome profiling of antimicrobial resistance in pseudomonas aeruginosa,A. Khaledi,"ABSTRACT Emerging resistance to antimicrobials and the lack of new antibiotic drug candidates underscore the need for optimization of current diagnostics and therapies to diminish the evolution and spread of multidrug resistance. As the antibiotic resistance status of a bacterial pathogen is defined by its genome, resistance profiling by applying next-generation sequencing (NGS) technologies may in the future accomplish pathogen identification, prompt initiation of targeted individualized treatment, and the implementation of optimized infection control measures. In this study, qualitative RNA sequencing was used to identify key genetic determinants of antibiotic resistance in 135 clinical Pseudomonas aeruginosa isolates from diverse geographic and infection site origins. By applying transcriptome-wide association studies, adaptive variations associated with resistance to the antibiotic classes fluoroquinolones, aminoglycosides, and β-lactams were identified. Besides potential novel biomarkers with a direct correlation to resistance, global patterns of phenotype-associated gene expression and sequence variations were identified by predictive machine learning approaches. Our research serves to establish genotype-based molecular diagnostic tools for the identification of the current resistance profiles of bacterial pathogens and paves the way for faster diagnostics for more efficient, targeted treatment strategies to also mitigate the future potential for resistance evolution.","{'model': 'tldr@v2.0.0', 'text': 'This research serves to establish genotype-based molecular diagnostic tools for the identification of the current resistance profiles of bacterial pathogens and paves the way for faster diagnostics for more efficient, targeted treatment strategies to also mitigate the future potential for resistance evolution.'}",https://aac.asm.org/content/aac/60/8/4722.full.pdf
-"triplet–triplet annihilation in 9,10-diphenylanthracene derivatives: the role of intersystem crossing and exciton diffusion",Tomas Serevičius,"Triplet–triplet annihilation (TTA) is an attractive way to boost the efficiency of conventional fluorescent organic light-emitting diodes (OLEDs). TTA-active anthracene derivatives are often considered as state-of-the-art emitters due to the proper energy level alignment. In this work, TTA properties of a series of highly fluorescent nonsymmetrical anthracene compounds bearing 9-(4-arylphenyl) moiety and 10-(4-hexylphenyl) fragments were assessed. Two different methods to enhance the TTA efficiency are demonstrated. First, the intensity of TTA-based delayed fluorescence directly depended on the intersystem crossing (ISC) rate. This ISC rate can be significantly enhanced in more conjugated compounds due to the resonant alignment of S1 and T2 energy levels. While enhanced ISC rate slightly quenches the intensity of prompt fluorescence, the rise of the triplet population boosts the intensity of resultant delayed fluorescence. Second, the triplet annihilation rate can be significantly enhanced by optimization...",,
-"multicast routing and its qos extension: problems, algorithms, and protocols",Bin Wang,"Multicast services have been increasingly used by various continuous media applications. The QoS requirements of these continuous media applications prompt the necessity for QoS-driven, constraint-based multicast routing. This article provides a comprehensive overview of existing multicast routing algorithms, protocols, and their QoS extension. In particular, we classify multicast routing problems according to their optimization functions and performance constraints, present basic routing algorithms in each problem class, and discuss their strengths and weaknesses. We also categorize existing multicast routing protocols, outline the issues and challenges in providing QoS in multicast routing, and point out possible future research directions.","{'model': 'tldr@v2.0.0', 'text': 'This article classifies multicast routing problems according to their optimization functions and performance constraints, present basic routing algorithms in each problem class, and discuss their strengths and weaknesses.'}",
-photochemical characterization and optimization of argon ion laser-induced fluorescence detection of o-phthalaldehyde/.beta.-mercaptoethanol-labeled amino acids and .gamma.-glutamyl peptides in liquid chromatography: ultratrace analysis with neurobiological samples,O. Orwar,"Optimization of argon ion laser-induced (UV multiline) fluorescence detection in liquid chromatography (1 mm i.d. columns) was performed aiming for ultratrace determination of o-phthalaldehyde/β-mercaptoethanol-labeled neuroactive amino acids and peptides. The derivatives are shown to undergo facile photochemical decomposition with photodestruction quantum yields, Φ D , in the range 0.009-0.14. In agreement with theoretical predictions, the optimal signal-to-noise (S/N) ratio was obtained at laser irradiances much lower than are required for prompt fluorescence saturation",,
-escmid* guideline for the diagnosis and management of candida diseases 2012: patients with hiv infection or aids.,O. Lortholary,"Mucosal candidiasis is frequent in immunocompromised HIV-infected highly active antiretroviral (HAART) naive patients or those who have failed therapy. Mucosal candidiasis is a marker of progressive immune deficiency. Because of the frequently marked and prompt immune reconstitution induced by HAART, there is no recommendation for primary antifungal prophylaxis of mucosal candidiasis in the HIV setting in Europe, although it has been evidenced as effective in the pre-HAART era. Fluconazole remains the first line of therapy for both oropharyngeal candidiasis and oesophageal candidiasis and should be preferred to itraconazole oral solution (or capsules when not available) due to fewer side effects. For patients who still present with fluconazole-refractory mucosal candidiasis, oral treatment with any other azole should be preferred based on precise Candida species identification and susceptibility testing results in addition to the optimization of HAART when feasible. For vaginal candidiasis, topical therapy is preferred.","{'model': 'tldr@v2.0.0', 'text': 'There is no recommendation for primary antifungal prophylaxis of mucosal candidiasis in the HIV setting in Europe, although it has been evidenced as effective in the pre-HAART era.'}",http://www.clinicalmicrobiologyandinfection.com/article/S1198743X14607680/pdf
-advantages and disadvantages of mammography screening,S. Heywang-Köbrunner,"Mammography screening is the only method presently considered appropriate for mass screening of asymptomatic women. Its frequent use, however, warrants diligent analysis of potential side effects. Radiation risk is far below the natural yearly risk of breast cancer and should not be used as an argument against screening. False-positive calls lead to additional imaging or histopathological assessment, mainly percutaneous breast biopsy. These measures are tolerated and accepted fairly well. Their number is limited by strict quality assurance and constant training. Interval cancers represent a limitation of breast screening that should prompt further research for optimization. Evaluation of overdiagnosis is a highly debated topic in the literature. According to the probably most realistic available calculations, overdiagnosis is acceptable as it is compensated by the potential mortality reduction. Nonetheless, this potential side effect warrants optimal adjustment of therapy to the patient’s individual risk. The mortality reduction seen in randomized studies was confirmed by results from national screening programs. A recent case referent study indicated that improvements in mortality reduction run parallel to improved mammographic techniques. Use of less aggressive therapies is another valuable effect of screening. Awareness of potential problems, strict quality assurance, and further research should help to further develop screening programs.","{'model': 'tldr@v2.0.0', 'text': 'The mortality reduction seen in randomized studies was confirmed by results from national screening programs, and a recent case referent study indicated that improvements in mortality reduction run parallel to improved mammographic techniques.'}",https://www.karger.com/Article/Pdf/329005
-anticipated and unanticipated difficult airway management,Zeping Xu,"Purpose of review Management of difficult airway is far from optimal despite of continuous progress in science and technology. The purpose of this review is to summarize the current research in the field and bring readers up to date. Recent findings New technologies for intubation make providers more confident to handle difficult airways, but there is lack of evidence indicating the reduction in incidence of ‘cannot intubate cannot ventilate (CICV)’. Optimization of mask ventilation should reduce the incidence of difficult mask ventilation but it is greatly underappreciated. Even optimization of preoxygenation is not directly associated with any decreased incidence of difficult airway, but it prolongs time of safe apnea oxygenation; therefore, is likely to improve the outcome of the patients if CICV occurs. Summary Improvement of managing difficult airway relies on optimized mask ventilation, utilization of the appropriate tools for intubation, maximization of the safe apnea oxygenation time, prompt surgical airway in response to severe hypoxia in case effective noninvasive interventions are not achievable. It seems that a simplified and concise algorithm of difficult airway management needs to be established in order to enable providers to easily remember and execute.","{'model': 'tldr@v2.0.0', 'text': 'Improvement of managing difficult airway relies on optimized mask ventilation, utilization of the appropriate tools for intubation, maximization of the safe apnea oxygenation time, prompt surgical airway in response to severe hypoxia in case effective noninvasive interventions are not achievable.'}",
-the approach to the adult with newly diagnosed adrenal insufficiency.,W. Arlt,"Adrenal insufficiency, primarily presenting as an adrenal crisis, is a life-threatening emergency and requires prompt therapeutic management including fluid resuscitation and stress dose hydrocortisone administration. Primary adrenal insufficiency is most frequently caused by autoimmune adrenalitis, and hypothalamic-pituitary tumors represent the most frequent cause of secondary adrenal insufficiency. However, the exact underlying diagnosis needs to be confirmed by a stepwise diagnostic approach, with an open eye for other differential diagnostic possibilities. Chronic replacement therapy with glucocorticoids and, in primary adrenal insufficiency, mineralocorticoids requires careful monitoring. However, current replacement strategies still require optimization as evidenced by recent studies demonstrating significantly impaired subjective health status and increased mortality in patients with primary and secondary adrenal insufficiency. Future studies will have to explore the potential of dehydroepiandrosterone replacement and modified delayed-release hydrocortisone to improve the prospects of patients with adrenal insufficiency.","{'model': 'tldr@v2.0.0', 'text': 'Future studies will have to explore the potential of dehydroepiandrosterone replacement and modified delayed-release hydrocortisone to improve the prospects of patients with adrenal insufficiency.'}",https://academic.oup.com/jcem/article-pdf/94/4/1059/10806508/jcem1059.pdf
-end-stage renal disease: symptom management and advance care planning.,Nina R. O’Connor,"The prevalence of end-stage renal disease continues to increase, and dialysis is offered to older and more medically complex patients. Pain is problematic in up to one-half of patients receiving dialysis and may result from renal and nonrenal etiologies. Opioids can be prescribed safely, but the patient's renal function must be considered when selecting a drug and when determining the dosage. Fentanyl and methadone are considered the safest opioids for use in patients with end-stage renal disease. Nonpain symptoms are common and affect quality of life. Phosphate binders, ondansetron, and naltrexone can be helpful for pruritus. Fatigue can be managed with treatment of anemia and optimization of dialysis, but persistent fatigue should prompt screening for depression. Ondansetron, metoclopramide, and haloperidol are effective for uremia-associated nausea. Nondialytic management may be preferable to dialysis initiation in older patients and in those with additional life-limiting illnesses, and may not significantly decrease life expectancy. Delaying dialysis initiation is also an option. Patients with end-stage renal disease should have advance directives, including documentation of situations in which they would no longer want dialysis.","{'model': 'tldr@v2.0.0', 'text': 'Patients with end-stage renal disease should have advance directives, including documentation of situations in which they would no longer want dialysis, and may not significantly decrease life expectancy.'}",
-sdslc: a multi-target domain-specific compiler for stencil computations,P. Rawat,"Stencil computations are at the core of applications in a number of scientific computing domains. We describe a domain-specific language for regular stencil computations that allows specification of the computations in a concise manner. We describe a multi-target compiler for this DSL, which generates optimized code for GPUa, FPGAs, and multi-core processors with short-vector SIMD instruction sets, considering both low-order and high-order stencil computations. The hardware differences between these three types of architecture prompt different optimization strategies for the compiler. We evaluate the domain-specific compiler using a number of benchmarks on CPU, GPU and FPGA platforms.","{'model': 'tldr@v2.0.0', 'text': 'A domain-specific language for regular stencil computations that allows specification of the computations in a concise manner is described and a multi-target compiler for this DSL is described, which generates optimized code for GPUa, FPGAs, and multi-core processors with short-vector SIMD instruction sets.'}",http://dl.acm.org/ft_gateway.cfm?id=2830025&type=pdf
-detection of retinal hemorrhage from fundus images using anfis classifier and mrg segmentation,L. GodlinAtlas,"Diabetic Retinopathy is a medicinal condition in which the retina is impaired based on fluid breaks from veins into the retina. The occurrence of hemorrhages in the retina is the most prompt effect of diabetic retinopathy. The number and state of hemorrhages are used to display the significance of the disorder. This research paper analyzed hemorrhage detection in retinal fundus images using classifier and segmentation methods. All the database images into the pre-processing steps and some meaning full features are extracted from the images. Then ANFIS classifier utilized to normaland abnormal images, this abnormal category into the hemorrhage detection process with help of segmentation technique. Here Region growing (RG) with threshold optimization techniques are considered its known as Modified RG (MRG) to get the maximum accuracy in the hemorrhage segmenting process. As regards the threshold optimization, Grey Wolf Optimization (GWO) technique used, this proposed work compared to our existing work getting maximum accuracy, sensitivity and specificity performance metrics.","{'model': 'tldr@v2.0.0', 'text': 'This research paper analyzed hemorrhage detection in retinal fundus images using classifier and segmentation methods and proposed a new threshold optimization technique known as Modified RG (MRG) to get the maximum accuracy in the hemorrhage segmenting process.'}",https://www.alliedacademies.org/articles/detection-of-retinal-hemorrhage-from-fundus-images-using-anfis-classifier-and-mrg-segmentation.pdf
-optimal feature selection for chronic kidney disease classification using deep learning classifier,K. Shankar,"Chronic Kidney Disease (CKD) is an increasing failure of kidney function leading to kidney failure over the years. The disease settles down and hence makes its diagnosis difficult. Analyzing CKD stages from standard office visit records can assist in premature recognition of the disease and prompt auspicious mediation. Hereby, we propose a methodology using inspired optimization model and learning procedure to classify CKD. The proposed method selects applicable features of kidney data with the help of Ant Lion Optimization (ALO) technique to choose optimal features for the classification process. After that, we sort the CKD data based on chosen features by utilizing Deep Neural Network (DNN). Performance comparison indicates that our proposed model accomplishes better classification accuracy, precision, F-measure, sensitivity measures when compared with other data mining classifiers.","{'model': 'tldr@v2.0.0', 'text': 'Performance comparison indicates that the proposed methodology using inspired optimization model and learning procedure to classify CKD accomplishes better classification accuracy, precision, F-measure, sensitivity measures when compared with other data mining classifiers.'}",
-modelling the size and skill-mix of hospital nursing teams,PR Harper,,"{'model': 'tldr@v2.0.0', 'text': 'The work, as presented here, extends the PROMPT functionality to consider in more detail workforce issues and suggests that it is cost beneficial to increase the number of permanently employed nurses to account for fluctuations in demand and corresponding high costs of temporary (agency) nurses.'}",
-self-organization towards reduced cost and energy per bit for future emerging radio technologies - sonnet,S. Mumtaz,"Self-organizing network (SON) is seen as playing a pivotal role towards reducing the management costs of networks for the exposure of 5G and beyond 5G networks. SON is capable of collecting information from the network, so as to perform self- configuration, self-optimization, self-healing, etc., so as to reduce the operation cost through less human involvement, and to optimize the service quality through robust and prompt network optimization. SONNET aims to drive further cost savings in the way networks are managed today by amplifying further the coverage zone of SON within the network. We believe that key technologies such as network sharing and Coordinated Multipoint (CoMP) can benefit from SON technology solutions. Hence, this paper will give the overview of SONNET project (http://sonnet- project.eu/) by defining the state of art technologies (CoMP, network sharing, SON Context-aware) and innovations brought by SONNET. Moreover, this paper will also highlight the 5G use cases recently standardized by 3GPP, which act as a baseline for novel SONNET use cases proposed in this paper. The selected use cases will be used for demonstration purposes with the aid of Ray tracer and System Level Simulator.","{'model': 'tldr@v2.0.0', 'text': 'The 5G use cases recently standardized by 3GPP, which act as a baseline for novel SONNET use cases proposed in this paper, are highlighted.'}",
-itimerc 2.0: fast incremental timing and cppr analysis,Pei-Yu Lee,"To achieve timing closure, performance-driven optimizations are repeatedly performed throughout the modern IC design flow. Along with these optimization operations, how to incrementally update timing information efficiently and accurately becomes a crucial task for fast turnaround time. On the other hand, to avoid wasteful over-optimization, clock path pessimism should be removed during timing analysis. In order to provide prompt timing information without over-pessimism during iterative optimizations, in this paper, we aim at fast incremental timing and CPPR analysis. We present two delicate techniques, lazy evaluation and lazy propagation, to avoid redundant updates. Our experiments are conducted on the benchmark suite released by TAU 2015 timing analysis contest. Experimental results show that our timer delivers the best results in terms of accuracy, runtime, and memory over all participating teams.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims at fast incremental timing and CPPR analysis, and presents two delicate techniques, lazy evaluation and lazy propagation, to avoid redundant updates.'}",
-a heuristics approach for classroom scheduling using genetic algorithm technique,Izah Rafidah Ahmad,"Reshuffling and arranging classroom based on the capacity of the audience, complete facilities, lecturing time and many more may lead to a complexity of classroom scheduling. While trying to enhance the productivity in classroom planning, this paper proposes a heuristic approach for timetabling optimization. A new algorithm was produced to take care of the timetabling problem in a university. The proposed of heuristics approach will prompt a superior utilization of the accessible classroom space for a given time table of courses at the university. Genetic Algorithm through Java programming languages were used in this study and aims at reducing the conflicts and optimizes the fitness. The algorithm considered the quantity of students in each class, class time, class size, time accessibility in each class and lecturer who in charge of the classes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed of heuristics approach will prompt a superior utilization of the accessible classroom space for a given time table of courses at the university.'}",
-defining a research agenda in value driven design: questions that need to be asked,D. Soban,"Value driven design is an innovative design process that utilizes the optimization of a system level value function to determine the best possible design. This contrasts with more traditional systems engineering techniques, which rely on satisfying requirements to determine the design solution. While ?design for value? is intuitively acceptable, the transformation of value driven design concepts into practical tools and methods for its application is challenging. This, coupled with the growing popularity of value-centric design philosophies, has led to a proposed research agenda in value driven design. This research agenda asks fundamental questions about the design philosophy and attempts to identify areas of significant challenge. The research agenda is meant to stimulate discussion in the field, as well as prompt research that will lead to the development of tools and methodologies that will facilitate the application of value driven design and further the state of the art.",,
-a history of boron neutron capture therapy of brain tumours. postulation of a brain radiation dose tolerance limit.,D. Slatkin,"Boron neutron capture therapy (BNCT) is a form of radiation therapy mediated by the short-range (less than 10 microns) energetic alpha (4He) and lithium-7 (7Li) ionizing particles that result from the prompt disintegration by slow neutrons of the stable (nonradioactive) nucleus boron-10 (10B). Recent advances in radiobiological and toxicological evaluation of tumour-affinitive boron-containing drugs and in optimization of the energies of neutrons in the incident beam have spurred interest in BNCT. This article presents a history of BNCT that emphasizes studies in the USA. A new dosimetric analysis of the 1959-1961 clinical trials of BNCT at Brookhaven National Laboratory is also presented. This analysis yields an acute radiation dose tolerance limit estimate of approximately 10 Gy-Eq to the capillary endothelium of human basal ganglia from BNCT. (Gy-Eq: Gray-equivalent, or relative biological effectiveness of a radiation component multiplied by the physical dose of the component (Gy), summed over the component kinds of radiation.)","{'model': 'tldr@v2.0.0', 'text': 'A new dosimetric analysis of the 1959-1961 clinical trials of BNCT at Brookhaven National Laboratory yields an acute radiation dose tolerance limit estimate of approximately 10 Gy-Eq to the capillary endothelium of human basal ganglia from BnCT.'}",
-multidetector ct and three-dimensional ct angiography for suspected vascular trauma of the extremities.,E. Fishman,"The evolution of computed tomography (CT) from four to 16 to 64 sections since its inception in the late 1970s has led to more widespread use of this imaging modality in the emergent setting. CT angiography has become a crucial diagnostic technique for identifying vascular injury in the trauma patient. Regardless of the nature of the traumatic injury (eg, stab wound, gunshot wound, injury from a motor vehicle accident), use of multidetector CT with two-dimensional (2D) reformation and three-dimensional (3D) rendering allows visualization of injury to bone, muscle, and vasculature. The radiologist should be familiar with the indications for CT angiography, optimization of current multidetector CT acquisition protocols, utility of 2D and 3D displays, and CT findings in the presence of vascular injury to ensure prompt diagnosis and treatment.","{'model': 'tldr@v2.0.0', 'text': 'The radiologist should be familiar with the indications for CT angiography, optimization of current multidetector CT acquisition protocols, utility of 2D and 3D displays, and CT findings in the presence of vascular injury to ensure prompt diagnosis and treatment.'}",
-human parent-child relationships from an evolutionary perspective,H. Keller,"In this article, a proximate conception of parental investment for humans is developed. Parenting investment is introduced as a part of life histories, specifying the optimization of reproductive success in terms of inclusive fitness. Contextual, parental, and children's characteristics that influence investment decisions are specified. We can distinguish between four systems of parenting: primary care, body stimulation, body contact, and face-to-face interaction, which prompt different socialization experiences with different psychological consequences. Parental behaviors within these different systems may vary in terms of the expression of warmth and contingency. Parenting is described here as an intergenerational transmission mechanism of psychological characteristics as well as reproductive styles.",,
-simultaneous immunoaffinity column cleanup and hplc analysis of aflatoxins and ochratoxin a in spanish bee pollen.,R. García-Villanova,"Bee pollen is a major substrate for mycotoxins growth when no prompt and adequate drying is performed by the beekeeper after collection by bees. Regulatory limits for aflatoxins and ochratoxin A are currently in force in the European Union for a rising list of foodstuffs, but not for this. An immunoaffinity column cleanup process has been applied prior to the analysis of aflatoxins B(1), B(2), G(1), and G(2) and ochratoxin A (OTA). Optimization of the HPLC conditions has involved both a gradient elution and a wavelength program for the separation and fluorimetric quantitation of all five mycotoxins at their maximum excitation and emission values of wavelength in a single run. The higher limit of detection (mug/kg) was 0.49 for OTA and 0.20 for aflatoxin B(1). Repeatability (RSDr) at the lower limit tested ranged from 9.85% for OTA to 6.23% for aflatoxin G(2), and recoveries also at the lower spiked level were 73% for OTA and 81% for aflatoxin B(1). None of the 20 samples assayed showed quantifiable values for the five mycotoxins.","{'model': 'tldr@v2.0.0', 'text': 'None of the 20 samples assayed showed quantifiable values for the five mycotoxins, but the higher limit of detection (mug/kg) was 0.49 for OTA and 0.20 for aflatoxin B(1).'}",
-an agent-based approach for managing symbiotic simulation of semiconductor assembly and test operation,M. Low,"The rapid changing business environment of high-tech asset intensive enterprises such as semiconductor manufacturing constantly drives production managers to look for better solutions to improve the manufacturing process. Simulation, though identified to be the most appropriate technique to generate and test out possible execution plans, suffers from long cycle-time in the process of model update, analysis and verification. It is thus very difficult to carry out prompt ""what-if' analysis to respond to abrupt changes in these systems. Symbiotic simulation systems have been proposed as a way of solving this problem by having the simulation and the physical system interact in a mutually beneficial manner. In this paper, we describe our work in developing a prototype proof-of-concept symbiotic simulation system that employs software agents in the monitoring, optimization and control of a semiconductor assembly and test operation.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the work in developing a prototype proof-of-concept symbiotic simulation system that employs software agents in the monitoring, optimization and control of a semiconductor assembly and test operation.'}",
-problems in signal-to-noise ratio for attenuation correction in high resolution pet,M. Dahlbom,"In high resolution PET, the number of events required for a high signal-to-noise ratio in the measured attenuation correction is an order of magnitude higlher (50-100 million) than that required for the emission image. This is due to the large number of elements in the attenuation correction sinogram (e.g. 50,000) and the large attenuation factors through the thicker sections of the body (e.g. 50). Another major contribution to noise in high resolution PET is contamination of true coincidences by accidentals and their subtraction from the prompt coincidences. In this work three methods which reduce noise in the measured attenuation are evaluated. The methods are: (1) optimization of energy threshold, (2) processing of random sinogram before subtraction and (3) spatial averaging of the tranismission sinogram before performing attenuation correction.",,
-fault location estimation based on matching the simulated and recorded waveforms using genetic algorithms,M. Kezunovic,"Prompt and accurate location of the faults in a large-scale transmission system is critical when system reliability is considered and usually is the first step in the system restoration. The accuracy of fault location estimation essentially depends on the information available. In this paper, the fault location estimation is mathematically formulated as an optimization problem of which the fault location and fault resistances are unknown variables. An efficient genetic algorithm-based searching scheme is developed for obtaining a solution that is globally optimal.",,
-"characteristics of relativistic solar cosmic rays during the event of december 13, 2006",E. V. Vashenyuk,,,
-a web-based environment for documentation and sharing of engineering design knowledge,Justin Rockwell,"This paper presents the foundation for a collaborative Web-based environment for improving communication by formally defining a platform for documentation and sharing of engineering design knowledge throughout the entire design process. In this work an ontological structure is utilized to concisely define a set of individual engineering concepts. This set of modular ontologies link together to create a flexible, yet consistent, product development knowledge-base. The resulting infrastructure uniquely enables the information stored within the knowledge-base to be readily inspectable and computable, thus allowing for design tools that reason on the information to assist designers and automate design processes. A case study of the structural optimization of a transfer plate for an aerospace circuit breaker is presented to demonstrate implementation and usefulness of the knowledge framework. The results indicate that the ontological knowledge-base can be used to prompt engineers to document important product development information, increase understanding of the design process, provide a means to intuitively retrieve information, and seamlessly access distributed information.© 2008 ASME","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the ontological knowledge-base can be used to prompt engineers to document important product development information, increase understanding of the design process, provide a means to intuitively retrieve information, and seamlessly access distributed information.'}",https://dr.lib.iastate.edu/bitstreams/f33fbf78-2699-4329-ba73-eb0eb605254f/download
-the surgical application of point‐of‐care haemostasis and platelet function testing,K. J. Dickinson,"Disordered coagulation complicates many diseases and their treatments, often predisposing to haemorrhage. Conversely, patients with cardiovascular disease who demonstrate antiplatelet resistance may be at increased thromboembolic risk. Prompt identification of these patients facilitates optimization of haemostatic dysfunction. Point‐of‐care (POC) tests are performed ‘near patient’ to provide a rapid assessment of haemostasis and platelet function.","{'model': 'tldr@v2.0.0', 'text': 'POC tests are performed ‘near patient’ to provide a rapid assessment of haemostasis and platelet function in patients with cardiovascular disease who demonstrate antiplatelet resistance.'}",https://academic.oup.com/bjs/article-pdf/95/11/1317/36709516/bjs6359.pdf
-preference heterogeneity in relation to museum services,U. Colombino,"Prevailing trends in the management of European museums underline the importance of additional museum services in fostering and encouraging the optimization of cultural assets while facilitating the collection of resources necessary for conservation. This paper considers the case of the archaeological site of Paestum (Salerno) and presents an analysis of individual preferences in relation to specific policies of cultural heritage management, each characterized by the supply of different museum services. Since the diversity of these services can prompt different individual preferences, the analysis allows for heterogeneity of parameters among individuals.",,http://www3.unisa.it/uploads/2551/105_dp.pdf
-acute mesenteric ischemia.,R. J. Stoney,"Acute mesenteric ischemia resulting from specific etiologies is a treatable vascular emergency. It requires a high index of suspicion and prompt, aggressive resuscitation and diagnostic maneuvers to determine the specific underlying cause. This will allow specific directed surgical revascularization, optimization of cardiac performance, or correction of a hypercoagulable state, all directed at maximizing the amount of functional bowel that is salvageable. These strategies are the cornerstones for a successful outcome in this life-threatening vascular catastrophe.","{'model': 'tldr@v2.0.0', 'text': 'Acute mesenteric ischemia resulting from specific etiologies is a treatable vascular emergency that requires a high index of suspicion and prompt, aggressive resuscitation and diagnostic maneuvers to determine the specific underlying cause.'}",
-modeling a neutron-rich nuclei source,M. Mirea,,,
-fluorescence polarization is a useful technology for reagent reduction in assay miniaturization.,T. J. Kowski,"The use of fluorescence polarization (FP) has increased significantly in the development of sensitive and robust assays for high throughput screening of chemical compound libraries during the past few years. In this study, we show that FP is a useful assay miniaturization technology for reagent reduction during high throughput screening. We developed and optimized several FP assays for binding to estrogen receptor alpha and two protein kinases with an assay volume of 100 microl. Without any re-optimization, a consistent signal window was maintained in 384- or 1536-well format when the assay volume varied from 2.5-100 microl at constant concentrations of all assay components. In contrast, the signal window decreased with decreasing assay volume at constant reagent concentration in the protein kinase C scintillation proximity assay (SPA) and prompt fluorescence assay. In addition, the effect of evaporation on the signal window was minimal for the FP assays. Our study suggests that FP is superior to SPA and prompt fluorescence in terms of reagent reduction in the miniaturized assay format.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that FP is superior to SPA and prompt fluorescence in terms of reagent reduction in the miniaturized assay format and the effect of evaporation on the signal window was minimal for the FP assays.'}",
-designing cross-sectional population pharmacokinetic studies: implications for pediatric and animal studies,C. D. Jones,"AbstractSampling constraints and ethical concerns are two of the issues that prompt the conduct of pediatric/animal cross-sectional pharmacokinetic studies. A simulation study was carried out to investigate the effect of optimization of design features (arrangement of concentrations (n) in time and sample size (N)) on the accuracy and precision of parameter (especially intersubject/interanimal variability) estimation in cross-sectional population pharmacokinetic studies involving drugs administered by single or multiple intravenous bolus input(s). Drugs exhibiting one and two compartment (i.e., 1 CMT and 2 CMT) pharmacokinetics were investigated using N of 20 to 100 and 30 to 150, respectively, n was optimized, using the profile (block) randomized sampling design, for accurate and precise estimation of population pharmacokinetic parameters. N of 50 and 80 was found to be adequate for the estimation of clearance (CL) and its variability ((ωCL) needed for dosage optimization) for the 1 CMT and 2 CMT drugs, ...","{'model': 'tldr@v2.0.0', 'text': 'A simulation study was carried out to investigate the effect of optimization of design features on the accuracy and precision of parameter estimation in cross-sectional population pharmacokinetic studies involving drugs administered by single or multiple intravenous bolus input(s).'}",
-a method for generating a control rod program for boiling water reactors,T. Kawai,"The OPROD computer code has been developed to generate a long-term control rod program, a series of control rod patterns that optimize a cycle length within various operational constraints. In the algorithm, the optimization problem is decomposed into two hierarchies. In the inner loop, a time-invariant target power distribution is assumed, and a control rod pattern is determined so as to best fit the power distribution to the target within the constraints at each burnup step. The target is then improved in the outer loop to achieve a longer cycle length. The code consists of two major parts: a three- dimensional boiling water reactor (BWR) core simulator and MAP, the method of approximate programming. It readily generates a long-term control rod program of BWRs without trial search by core-management engineers. The OPROD has therefore facilitated prompt response to varying operating conditions and the investigation of a conflicting relationship between the thermal limitation and the cycle length. (auth)","{'model': 'tldr@v2.0.0', 'text': 'The OPROD has facilitated prompt response to varying operating conditions and the investigation of a conflicting relationship between the thermal limitation and the cycle length.'}",
-home telemonitoring of respiratory activity and heart rate variability in chronic heart failure patients: the challenge of the home or hospital in heart failure project,G. Pinna,"Nocturnal respiratory disorders and depressed heart rate variability are known predictors of poor prognosis in chronic heart failure (CHF) patients. Intermittent monitoring of cardiorespiratory signals while the patient is at home might thus allow early identification of clinical deterioration and prompt optimization of treatment, leading to reduced hospitalizations and mortality and improved quality of life. Within the European Community multicenter trial HHH (Home or Hospital in Heart Failure), we are testing a novel low-cost system for 24-hour recording of cardiorespiratory signals, suitable to be self-managed by the patient at home, with transmission of acquired data through standard telephone lines to the medical/nursing staff. Preliminary results from 24 CHF patients enrolled so far indicate that monthly home telemonitoring is feasible and the compliance is high.","{'model': 'tldr@v2.0.0', 'text': 'Preliminary results from 24 CHF patients enrolled so far indicate that monthly home telemonitoring is feasible and the compliance is high, and a novel low-cost system for 24-hour recording of cardiorespiratory signals is tested, suitable to be self-managed by the patient at home.'}",http://www.cinc.org/Proceedings/2003/pdf/197.pdf
-engineering robust ag‐decorated polydopamine nano‐photothermal platforms to combat bacterial infection and prompt wound healing,Xiaoliang Qi,"Polydopamine (PDA) nanoparticles have emerged as an attractive biomimetic photothermal agent in photothermal antibacterial therapy due to their ease of synthesis, good biodegradability, long‐term safety, and excellent photostability. However, the therapeutic effects of PDA nanoparticles are generally limited by the low photothermal conversion efficiency (PCE). Herein, PDA@Ag nanoparticles are synthesized via growing Ag on the surface of PDA nanoparticles and then encapsulated into a cationic guar gum (CG) hydrogel network. The optimized CG/PDA@Ag platform exhibits a high PCE (38.2%), which is more than two times higher than that of pure PDA (16.6%). More importantly, the formulated CG/PDA@Ag hydrogel with many active groups can capture and kill bacteria through effective interactions between hydrogel and bacteria, thereby benefiting the antibacterial effect. As anticipated, the designed CG/PDA@Ag system combined the advantages of PDA@Ag nanoparticles (high PCE) and hydrogel (preventing aggregation of PDA@Ag nanoparticles and possessing inherent antibacterial ability) is demonstrated to have superior antibacterial efficacy both in vitro and in vivo. This study develops a facile approach to boost the PCE of PDA for photothermal antibacterial therapy, providing a significant step forward in advancing the application of PDA nano‐photothermal agents.","{'model': 'tldr@v2.0.0', 'text': 'A facile approach to boost the PCE of Pda for photothermal antibacterial therapy is developed, providing a significant step forward in advancing the application of PDA nano‐photothermal agents.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/advs.202106015
-speechprompt: an exploration of prompt tuning on generative spoken language model for speech processing tasks,Kai-Wei Chang,"Speech representations learned from Self-supervised learning (SSL) models can benefit various speech processing tasks. However, utilizing SSL representations usually requires fine-tuning the pre-trained models or designing task-specific downstream models and loss functions, causing much memory usage and human labor. Recently, prompting in Natural Language Processing (NLP) has been found to be an efficient technique to leverage pre-trained language models (LMs). Specifically, prompt tuning optimizes a limited number of task-specific parameters with a fixed pre-trained model; as a result, only a small set of parameters is needed to be stored for each task. Prompt tuning improves computation and memory efficiency by leveraging the pre-trained LM's prediction ability. Nevertheless, such a paradigm is little studied in the speech community. We report in this paper the first exploration of the prompt tuning paradigm for speech processing tasks based on Generative Spoken Language Model (GSLM). Experiment results show that the prompt tuning technique achieves competitive performance in speech classification tasks with fewer trainable parameters than fine-tuning specialized downstream models. We further study the technique in challenging sequence generation tasks. Prompt tuning also demonstrates its potential, while the limitation and possible research directions are discussed in this paper. The source code is available on https://github.com/ga642381/SpeechPrompt.","{'model': 'tldr@v2.0.0', 'text': 'The first exploration of the prompt tuning paradigm for speech processing tasks based on Generative Spoken Language Model (GSLM) is reported, demonstrating competitive performance in speech classification tasks with fewer trainable parameters than fine-tuning specialized downstream models.'}",
-late prompt tuning: a late prompt could be better than many prompts,Xiangyang Liu,"Prompt tuning is a parameter-efficient tuning (PETuning) method for utilizing pre-trained models (PTMs) that simply prepends a soft prompt to the input and only optimizes the prompt to adapt PTMs to downstream tasks. Although it is parameter- and deployment-efficient, its performance still lags behind other state-of-the-art PETuning methods. Besides, the training cost of prompt tuning is not significantly reduced due to the back-propagation through the entire model. Through empirical analyses, we shed some light on the lagging performance of prompt tuning and recognize a trade-off between the propagation distance from label signals to the inserted prompt and the influence of the prompt on model outputs. Further, we present Late Prompt Tuning (LPT) that inserts a late prompt into an intermediate layer of the PTM instead of the input layer or all layers. The late prompt is obtained by a neural prompt generator conditioned on the hidden states before the prompt insertion layer and therefore is instance-dependent. Through extensive experimental results across various tasks and PTMs, we show that LPT can achieve competitive performance to full model tuning and other PETuning methods under both full-data and few-shot scenarios while possessing faster training speed and lower memory cost.","{'model': 'tldr@v2.0.0', 'text': 'Late Prompt Tuning (LPT) is presented that can achieve competitive performance to full model tuning and other PETuning methods under both full-data and few-shot scenarios while possessing faster training speed and lower memory cost.'}",http://arxiv.org/pdf/2210.11292
-kipt: knowledge-injected prompt tuning for event detection,Haochen Li,"Event detection aims to detect events from the text by identifying and classifying event triggers (the most representative words). Most of the existing works rely heavily on complex downstream networks and require sufficient training data. Thus, those models may be structurally redundant and perform poorly when data is scarce. Prompt-based models are easy to build and are promising for few-shot tasks. However, current prompt-based methods may suffer from low precision because they have not introduced event-related semantic knowledge (e.g., part of speech, semantic correlation, etc.). To address these problems, this paper proposes a Knowledge-injected Prompt Tuning (KiPT) model. Specifically, the event detection task is formulated into a condition generation task. Then, knowledge-injected prompts are constructed using external knowledge bases, and a prompt tuning strategy is leveraged to optimize the prompts. Extensive experiments indicate that KiPT outperforms strong baselines, especially in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The event detection task is formulated into a condition generation task, knowledge-injected prompts are constructed using external knowledge bases, and a prompt tuning strategy is leveraged to optimize the prompts.'}",
-highly efficient oxidation of propane at low temperature over a pt-based catalyst by optimization support.,Zhen‐Feng Huang,"Pt-based catalysts have attracted widespread attention in environmental protection applications, especially in the catalytic destruction of light alkane pollutants. However, developing a satisfying platinum catalyst with high activity, excellent water-resistance, and practical suitability for hydrocarbon combustion at low temperature is challenging. In this study, the Pt catalyst supported on the selected Nb2O5 oxide exhibited an efficient catalytic activity in propane oxidation and exceeded that of most catalysts reported in the literature. More importantly, the Pt/Nb2O5 catalyst maintained excellent activity and durability even after high-temperature aging at 700 °C and under harsh working conditions, such as a certain degree of moisture, high space velocity, and composite pollutants. The excellent performance of the Pt/Nb2O5 catalyst was attributed to the abundant metallic Pt species stabilized on the surface of Nb2O5, which prompted the C-H bond dissociation ability as the rate-determining step. Furthermore, propane was initially activated via oxidehydrogenation and followed the acrylate species path as a more efficient propane oxidation path on the Pt/Nb2O5 surface. Overall, Pt/Nb2O5 can be considered a promising catalyst for the catalytic oxidation of alkanes from industrial sources and could provide inspiration for designing superb catalysts for the oxidation of light alkanes.",,
-making pre-trained language models end-to-end few-shot learners with contrastive prompt tuning,Ziyun Xu,"Pre-trained Language Models (PLMs) have achieved remarkable performance for various language understanding tasks in IR systems, which require the fine-tuning process based on labeled training data. For low-resource scenarios, prompt-based learning for PLMs exploits prompts as task guidance and turns downstream tasks into masked language problems for effective few-shot fine-tuning. In most existing approaches, the high performance of prompt-based learning heavily relies on handcrafted prompts and verbalizers, which may limit the application of such approaches in real-world scenarios. To solve this issue, we present CP-Tuning, an end-to-end Contrastive Prompt Tuning framework for fine-tuning PLMs without any manual engineering of task-specific prompts and verbalizers. It is integrated with the task-invariant continuous prompt encoding technique with fully trainable prompt parameters. We further propose the pair-wise cost-sensitive contrastive learning procedure to optimize the model in order to achieve verbalizer-free class mapping and enhance the task-invariance of prompts. It explicitly learns to distinguish different classes and makes the decision boundary smoother by assigning different costs to easy and hard cases. Experiments over a variety of language understanding tasks and different PLMs show that CP-Tuning outperforms state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'CP-Tuning is presented, an end-to-end Contrastive Prompt Tuning framework for fine-tuning PLMs without any manual engineering of task-specific prompts and verbalizers, and is integrated with the task-invariant continuous prompt encoding technique with fully trainable prompt parameters.'}",https://arxiv.org/pdf/2204.00166
-lpt: long-tailed prompt tuning for image classification,Bowen Dong,"For long-tailed classification, most works often pretrain a big model on a large-scale dataset, and then fine-tune the whole model for adapting to long-tailed data. Though promising, fine-tuning the whole pretrained model tends to suffer from high cost in computation and deployment of different models for different tasks, as well as weakened generalization ability for overfitting to certain features of long-tailed data. To alleviate these issues, we propose an effective Long-tailed Prompt Tuning method for long-tailed classification. LPT introduces several trainable prompts into a frozen pretrained model to adapt it to long-tailed data. For better effectiveness, we divide prompts into two groups: 1) a shared prompt for the whole long-tailed dataset to learn general features and to adapt a pretrained model into target domain; and 2) group-specific prompts to gather group-specific features for the samples which have similar features and also to empower the pretrained model with discrimination ability. Then we design a two-phase training paradigm to learn these prompts. In phase 1, we train the shared prompt via supervised prompt tuning to adapt a pretrained model to the desired long-tailed domain. In phase 2, we use the learnt shared prompt as query to select a small best matched set for a group of similar samples from the group-specific prompt set to dig the common features of these similar samples, then optimize these prompts with dual sampling strategy and asymmetric GCL loss. By only fine-tuning a few prompts while fixing the pretrained model, LPT can reduce training and deployment cost by storing a few prompts, and enjoys a strong generalization ability of the pretrained model. Experiments show that on various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves comparable performance than previous whole model fine-tuning methods, and is more robust to domain-shift.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that on various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves comparable performance than previous whole model fine-tuning methods, and is more robust to domain-shift.'}",http://arxiv.org/pdf/2210.01033
-multi-prompt alignment for multi-source unsupervised domain adaptation,Haoran Chen,"Most existing methods for unsupervised domain adaptation (UDA) rely on a shared network to extract domain-invariant features. However, when facing multiple source domains, optimizing such a network involves updating the parameters of the entire network, making it both computationally expensive and challenging, particularly when coupled with min-max objectives. Inspired by recent advances in prompt learning that adapts high-capacity models for downstream tasks in a computationally economic way, we introduce Multi-Prompt Alignment (MPA), a simple yet efficient framework for multi-source UDA. Given a source and target domain pair, MPA first trains an individual prompt to minimize the domain gap through a contrastive loss. Then, MPA denoises the learned prompts through an auto-encoding process and aligns them by maximizing the agreement of all the reconstructed prompts. Moreover, we show that the resulting subspace acquired from the auto-encoding process can easily generalize to a streamlined set of target domains, making our method more efficient for practical usage. Extensive experiments show that MPA achieves state-of-the-art results on three popular datasets with an impressive average accuracy of 54.1% on DomainNet.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by recent advances in prompt learning that adapts high-capacity models for downstream tasks in a computationally economic way, Multi-Prompt Alignment (MPA) is introduced, a simple yet efficient framework for multi-source UDA.'}",http://arxiv.org/pdf/2209.15210
-eliciting knowledge from pretrained language models for prototypical prompt verbalizer,Yinyi Wei,,"{'model': 'tldr@v2.0.0', 'text': 'This paper focuses on eliciting knowledge from pretrained language models and proposes a prototypical prompt verbalizer for prompt-tuning, which optimizes models by contrastive learning.'}",https://arxiv.org/pdf/2201.05411
-fine-grained retrieval prompt tuning,Shijie Wang,"Fine-grained object retrieval aims to learn discriminative representation to retrieve visually similar objects. However, existing top-performing works usually impose pairwise similarities on the semantic embedding spaces or design a localization sub-network to continually fine-tune the entire model in limited data scenarios, thus resulting in convergence to suboptimal solutions. In this paper, we develop Fine-grained Retrieval Prompt Tuning (FRPT), which steers a frozen pre-trained model to perform the fine-grained retrieval task from the perspectives of sample prompting and feature adaptation. Specifically, FRPT only needs to learn fewer parameters in the prompt and adaptation instead of fine-tuning the entire model, thus solving the issue of convergence to suboptimal solutions caused by fine-tuning the entire model. Technically, a discriminative perturbation prompt (DPP) is introduced and deemed as a sample prompting process, which amplifies and even exaggerates some discriminative elements contributing to category prediction via a content-aware inhomogeneous sampling operation. In this way, DPP can make the fine-grained retrieval task aided by the perturbation prompts close to the solved task during the original pre-training. Thereby, it preserves the generalization and discrimination of representation extracted from input samples. Besides, a category-specific awareness head is proposed and regarded as feature adaptation, which removes the species discrepancies in features extracted by the pre-trained model using category-guided instance normalization. And thus, it makes the optimized features only include the discrepancies among subcategories. Extensive experiments demonstrate that our FRPT with fewer learnable parameters achieves the state-of-the-art performance on three widely-used fine-grained datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops Fine-grained Retrieval Prompt Tuning (FRPT), which steers a frozen pre-trained model to perform the fine- grained retrieval task from the perspectives of sample prompting and feature adaptation, thus solving the issue of convergence to suboptimal solutions caused by fine-tuning the entire model.'}",http://arxiv.org/pdf/2207.14465
-exploring low-dimensional intrinsic task subspace via prompt tuning,Yujia Qin,"Why can pre-trained language models (PLMs) 001 learn universal representations and effectively 002 adapt to broad NLP tasks differing a lot super- 003 ficially? In this work, we empirically find evi- 004 dence indicating that the adaptations of PLMs 005 to various few-shot tasks can be reparameter- 006 ized as optimizing only a few free parame- 007 ters in a unified low-dimensional intrinsic task 008 subspace , which may help us understand why 009 PLMs could easily adapt to various NLP tasks 010 with small-scale data. To find such a subspace 011 and examine its universality, we propose an 012 analysis pipeline called intrinsic prompt tun- 013 ing (IPT). Specifically, we resort to the re- 014 cent success of prompt tuning and decompose 015 the soft prompts of multiple NLP tasks into 016 the same low-dimensional nonlinear subspace, 017 then we learn to adapt the PLM to unseen data 018 or tasks by only tuning parameters in this sub- 019 space. In the experiments, we study diverse 020 few-shot NLP tasks and surprisingly find that 021 in a 5 -dimensional subspace found with 100 022 tasks, by only tuning 5 free parameters, we 023 can recover 87% and 65% of the full prompt 024 tuning performance for 100 seen tasks (using 025 different training data) and 20 unseen tasks, re- 026 spectively, showing great generalization abil- 027 ity of the found intrinsic task subspace. Be- 028 sides being an analysis tool, IPT could further 029 bring practical benefits, such as improving the 030 prompt tuning stability. 031","{'model': 'tldr@v2.0.0', 'text': 'Empirically, there is evidence indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parame- 007 ters in a unified low-dimensional intrinsic task 008 subspace, which may help to understand why 009 PLMs could easily adapt to various NLP tasks 010 with small-scale data.'}",
-exploring universal intrinsic task subspace via prompt tuning,Yujia Qin,"Why can pre-trained language models (PLMs) learn universal representations and effectively adapt to broad NLP tasks differing a lot superficially? In this work, we empirically find evidence indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parameters in a unified low-dimensional intrinsic task subspace, which may help us understand why PLMs could easily adapt to various NLP tasks with small-scale data. To find such a subspace and examine its universality, we propose an analysis pipeline called intrinsic prompt tuning (IPT). Specifically, we resort to the recent success of prompt tuning and decompose the soft prompts of multiple NLP tasks into the same low-dimensional nonlinear subspace, then we learn to adapt the PLM to unseen data or tasks by only tuning parameters in this subspace. In the experiments, we study diverse few-shot NLP tasks and surprisingly find that in a 250-dimensional subspace found with 100 tasks, by only tuning 250 free parameters, we can recover 97% and 83% of the full prompt tuning performance for 100 seen tasks (using different training data) and 20 unseen tasks, respectively, showing great generalization ability of the found intrinsic task subspace. Besides being an analysis tool, IPT could further help us improve the prompt tuning stability.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is empirically found indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parameters in a unified low-dimensional intrinsic task subspace, which may help to understand why PLMs could easily adapt to various NLP tasks with small-scale data.'}",
-how to design the perfect prompt: a linguistic approach to prompt design in automotive voice assistants – an exploratory study,Anna-Maria Meck,"In-vehicle voice user interfaces (VUIs) are becoming increasingly popular while needing to handle more and more complex functions. While many guidelines exist in terms of dialog design, a methodical and encompassing approach to prompt design is absent in the scientific landscape. The present work closes this gap by providing such an approach in form of linguistic-centered research. By extracting syntactical, lexical, and grammatical parameters from a German contemporary grammar, we examine how their respective manifestations affect users’ perception of a given system output across different prompt types. Through exploratory studies with a total of 1,206 participants, we provide concrete best practices to optimize and refine the design of VUI prompts. Based on these best practices, three superordinate user needs regarding prompt design can be identified: a) a suitable level of (in)formality, b) a suitable level of complexity/simplicity, and c) a suitable level of (im)mediacy.","{'model': 'tldr@v2.0.0', 'text': 'By extracting syntactical, lexical, and grammatical parameters from a German contemporary grammar, this work examines how their respective manifestations affect users’ perception of a given system output across different prompt types.'}",
-improving chatgpt prompt for code generation,Chao Liu,"Automated code generation can be a powerful technique for software development, significantly reducing developers' efforts and time required to create new code by generating it automatically based on requirements. Recently, OpenAI's language model ChatGPT has emerged as a powerful tool for generating human-like responses to a wide range of textual inputs (i.e., prompts), including those related to code generation. However, the effectiveness of ChatGPT for code generation is not well understood, and the generation performance could be heavily influenced by the choice of prompt. To answer these questions, we conducted experiments using the CodeXGlue dataset to evaluate ChatGPT's capabilities for two code generation tasks, including text-to-code and code-to-code generation. We designed prompts by leveraging the chain-of-thought strategy with multi-step optimizations. Our results showed that by carefully designing prompts to guide ChatGPT, the generation performance can be improved substantially. We also analyzed the factors that influenced the prompt design and provided insights that could guide future research.","{'model': 'tldr@v2.0.0', 'text': 'By carefully designing prompts to guide ChatGPT, the generation performance can be improved substantially, and the factors that influenced the prompt design were analyzed and provided insights that could guide future research.'}",http://arxiv.org/pdf/2305.08360
-prompt engineering with chatgpt: a guide for academic writers,L. Giray,,,
-dynamic prompting: a unified framework for prompt tuning,Xianjun Yang,"It has been demonstrated that the art of prompt tuning is highly effective in efficiently extracting knowledge from pretrained foundation models, encompassing pretrained language models (PLMs), vision pretrained models, and vision-language (V-L) models. However, the efficacy of employing fixed soft prompts with a predetermined position for concatenation with inputs for all instances, irrespective of their inherent disparities, remains uncertain. Variables such as the position, length, and representations of prompts across diverse instances and tasks can substantially influence the performance of prompt tuning. In this context, we provide a theoretical analysis, which reveals that optimizing the position of the prompt to encompass the input can capture additional semantic information that traditional prefix or postfix prompt tuning methods fail to capture. Building upon our analysis, we present a unified dynamic prompt (DP) tuning strategy that dynamically determines different factors of prompts based on specific tasks and instances. To accomplish this, we employ a lightweight learning network with Gumble-Softmax, allowing us to learn instance-dependent guidance. Experimental results underscore the significant performance improvement achieved by dynamic prompt tuning across a wide range of tasks, including NLP tasks, vision recognition tasks, and vision-language tasks. Furthermore, we establish the universal applicability of our approach under full-data, few-shot, and multitask scenarios. Codes are available at https://github.com/Xianjun-Yang/DPT.","{'model': 'tldr@v2.0.0', 'text': 'A theoretical analysis is provided, which reveals that optimizing the position of the prompt to encompass the input can capture additional semantic information that traditional prefix or postfix prompt tuning methods fail to capture.'}",http://arxiv.org/pdf/2303.02909
-exploring sparse visual prompt for domain adaptive dense prediction,Senqiao Yang,"The visual prompts have provided an efficient manner in addressing visual cross-domain problems. In previous works, Visual Domain Prompt (VDP) first introduces domain prompts to tackle the classification Test-Time Adaptation (TTA) problem by warping image-level prompts on the input and fine-tuning prompts for each target domain. However, since the image-level prompts mask out continuous spatial details in the prompt-allocated region, it will suffer from inaccurate contextual information and limited domain knowledge extraction, particularly when dealing with dense prediction TTA problems. To overcome these challenges, we propose a novel Sparse Visual Domain Prompts (SVDP) approach, which holds minimal trainable parameters (e.g., 0.1\%) in the image-level prompt and reserves more spatial information of the input. To better apply SVDP in extracting domain-specific knowledge, we introduce the Domain Prompt Placement (DPP) method to adaptively allocates trainable parameters of SVDP on the pixels with large distribution shifts. Furthermore, recognizing that each target domain sample exhibits a unique domain shift, we design Domain Prompt Updating (DPU) strategy to optimize prompt parameters differently for each sample, facilitating efficient adaptation to the target domain. Extensive experiments were conducted on widely-used TTA and continual TTA benchmarks, and our proposed method achieves state-of-the-art performance in both semantic segmentation and depth estimation tasks.","{'model': 'tldr@v2.0.0', 'text': 'To better apply SVDP in extracting domain-specific knowledge, the Domain Prompt Placement (DPP) method is introduced to adaptively allocates trainable parameters of SVDP on the pixels with large distribution shifts, and recognizing that each target domain sample exhibits a unique domain shift, the DPU strategy is designed to optimize prompt parameters differently for each sample, facilitating efficient adaptation to the target domain.'}",
-stylediffusion: prompt-embedding inversion for text-based editing,Senmao Li,"A significant research effort is focused on exploiting the amazing capacities of pretrained diffusion models for the editing of images. They either finetune the model, or invert the image in the latent space of the pretrained model. However, they suffer from two problems: (1) Unsatisfying results for selected regions, and unexpected changes in nonselected regions. (2) They require careful text prompt editing where the prompt should include all visual objects in the input image. To address this, we propose two improvements: (1) Only optimizing the input of the value linear network in the cross-attention layers, is sufficiently powerful to reconstruct a real image. (2) We propose attention regularization to preserve the object-like attention maps after editing, enabling us to obtain accurate style editing without invoking significant structural changes. We further improve the editing technique which is used for the unconditional branch of classifier-free guidance, as well as the conditional one as used by P2P. Extensive experimental prompt-editing results on a variety of images, demonstrate qualitatively and quantitatively that our method has superior editing capabilities than existing and concurrent works.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experimental prompt-editing results, demonstrate qualitatively and quantitatively that the proposed attention regularization to preserve the object-like attention maps after editing has superior editing capabilities than existing and concurrent works.'}",https://arxiv.org/pdf/2303.15649
-efficient transfer learning for visual tasks via continuous optimization of prompts,Jonathan Conder,,,
-a simple zero-shot prompt weighting technique to improve prompt ensembling in text-image models,J. Allingham,"Contrastively trained text-image models have the remarkable ability to perform zero-shot classification, that is, classifying previously unseen images into categories that the model has never been explicitly trained to identify. However, these zero-shot classifiers need prompt engineering to achieve high accuracy. Prompt engineering typically requires hand-crafting a set of prompts for individual downstream tasks. In this work, we aim to automate this prompt engineering and improve zero-shot accuracy through prompt ensembling. In particular, we ask""Given a large pool of prompts, can we automatically score the prompts and ensemble those that are most suitable for a particular downstream dataset, without needing access to labeled validation data?"". We demonstrate that this is possible. In doing so, we identify several pathologies in a naive prompt scoring method where the score can be easily overconfident due to biases in pre-training and test data, and we propose a novel prompt scoring method that corrects for the biases. Using our proposed scoring method to create a weighted average prompt ensemble, our method outperforms equal average ensemble, as well as hand-crafted prompts, on ImageNet, 4 of its variants, and 11 fine-grained classification benchmarks, all while being fully automatic, optimization-free, and not requiring access to labeled validation data.","{'model': 'tldr@v2.0.0', 'text': 'Using the proposed scoring method to create a weighted average prompt ensemble, the method outperforms equal average ensemble, as well as hand-crafted prompts, on ImageNet, 4 of its variants, and 11 fine-grained classification benchmarks, all while being fully automatic, optimization-free, and not requiring access to labeled validation data.'}",https://arxiv.org/pdf/2302.06235
-drpt: disentangled and recurrent prompt tuning for compositional zero-shot learning,Xiaocheng Lu,"Compositional Zero-shot Learning (CZSL) aims to recognize novel concepts composed of known knowledge without training samples. Standard CZSL either identifies visual primitives or enhances unseen composed entities, and as a result, entanglement between state and object primitives cannot be fully utilized. Admittedly, vision- language models (VLMs) could naturally cope with CZSL through tuning prompts, while uneven entanglement leads prompts to be dragged into local optimum. In this paper, we take a further step to introduce a novel Disentangled and Recurrent Prompt Tuning framework termed DRPT to better tap the potential of VLMs in CZSL. Specifically, the state and object primitives are deemed as learnable tokens of vocabulary embedded in prompts and tuned on seen compositions. Instead of jointly tuning state and object, we devise a disentangled and recurrent tuning strategy to suppress the traction force caused by entanglement and gradually optimize the token parameters, leading to a better prompt space. Notably, we develop a progressive fine-tuning procedure that allows for incremental updates to the prompts, optimizing the object first, then the state, and vice versa. Meanwhile, the optimization of state and object is independent, thus clearer features can be learned to further alleviate the issue of entangling misleading optimization. Moreover, we quantify and analyze the entanglement in CZSL and supplement entanglement rebalancing optimization schemes. DRPT surpasses representative state-of-the-art methods on extensive benchmark datasets, demonstrating superiority in both accuracy and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel Disentangled and Recurrent Prompt Tuning framework termed DRPT to better tap the potential of VLMs in CZSL and develops a progressive fine-tuning procedure that allows for incremental updates to the prompts, optimizing the object first, then the state, and vice versa.'}",http://arxiv.org/pdf/2305.01239
-reprompt: automatic prompt editing to refine ai-generative art towards precise expressions,Yunlong Wang,"Generative AI models have shown impressive ability to produce images with text prompts, which could benefit creativity in visual art creation and self-expression. However, it is unclear how precisely the generated images express contexts and emotions from the input texts. We explored the emotional expressiveness of AI-generated images and developed RePrompt, an automatic method to refine text prompts toward precise expression of the generated images. Inspired by crowdsourced editing strategies, we curated intuitive text features, such as the number and concreteness of nouns, and trained a proxy model to analyze the feature effects on the AI-generated image. With model explanations of the proxy model, we curated a rubric to adjust text prompts to optimize image generation for precise emotion expression. We conducted simulation and user studies, which showed that RePrompt significantly improves the emotional expressiveness of AI-generated images, especially for negative emotions.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by crowdsourced editing strategies, this work curated intuitive text features, such as the number and concreteness of nouns, and trained a proxy model to analyze the feature effects on the AI-generated image.'}",https://arxiv.org/pdf/2302.09466
-feasibility of quasi-prompt pet-based range verification in proton therapy,I. Ozoemelam,"Compared to photon therapy, proton therapy allows a better conformation of the dose to the tumor volume with reduced radiation dose to co-irradiated tissues. In vivo verification techniques including positron emission tomography (PET) have been proposed as quality assurance tools to mitigate proton range uncertainties. Detection of differences between planned and actual dose delivery on a short timescale provides a fast trigger for corrective actions. Conventional PET-based imaging of 15O (T1/2 = 2 min) and 11C (T1/2 = 20 min) distributions precludes such immediate feedback. We here present a demonstration of near real-time range verification by means of PET imaging of 12N (T1/2 = 11 ms). PMMA and graphite targets were irradiated with a 150 MeV proton pencil beam consisting of a series of pulses of 10 ms beam-on and 90 ms beam-off. Two modules of a modified Siemens Biograph mCT PET scanner (21 × 21 cm2 each), installed 25 cm apart, were used to image the beam-induced PET activity during the beam-off periods. The modifications enable the detectors to be switched off during the beam-on periods. 12N images were reconstructed using planar tomography. Using a 1D projection of the 2D reconstructed 12N image, the activity range was obtained from a fit of the activity profile with a sigmoid function. Range shifts due to modified target configurations were assessed for multiples of the clinically relevant 108 protons per pulse (approximately equal to the highest intensity spots in the pencil beam scanning delivery of a dose of 1 Gy over a cubic 1 l volume). The standard deviation of the activity range, determined from 30 datasets obtained from three irradiations on PMMA and graphite targets, was found to be 2.5 and 2.6 mm (1σ) with 108 protons per pulse and 0.9 and 0.8 mm (1σ) with 109 protons per pulse. Analytical extrapolation of the results from this study shows that using a scanner with a solid angle coverage of 57%, with optimized detector switching and spot delivery times much smaller than the 12N half-life, an activity range measurement precision of 2.0 mm (1σ) and 1.3 mm (1σ) within 50 ms into an irradiation with 4 × 107 and 108 protons per pencil beam spot can be potentially realized. Aggregated imaging of neighboring spots or, if possible, increasing the number of protons for a few probe beam spots will enable the realization of higher precision range measurement.","{'model': 'tldr@v2.0.0', 'text': 'Analytical extrapolation of the results from this study shows that using a scanner with a solid angle coverage of 57%, with optimized detector switching and spot delivery times much smaller than the 12N half-life, an activity range measurement precision of 2.0 mm can be potentially realized.'}",
-"optimizing of a question prompt list to improve communication about the heart failure trajectory in patients, families, and health care professionals",Lisa Hjelmfors,,"{'model': 'tldr@v2.0.0', 'text': 'This Question Prompt List has successfully been adapted into a Swedish version and a Dutch version and includes questions about the HF trajectory which patients, their families, and health care professionals perceived to be relevant for discussion in clinical practice.'}",https://bmcpalliatcare.biomedcentral.com/track/pdf/10.1186/s12904-020-00665-3
-optimal sink node placement in large scale wireless sensor networks based on harris’ hawk optimization algorithm,E. H. Houssein,"Large-scale wireless sensor network (LSWSN) is composed of a huge number of sensor nodes that are distributed in some region of interest (ROI), to sense and measure the environmental conditions like pressure, temperature, pollution levels, humidity, wind, and so on. The objective is to collect data for real-time monitoring so that appropriate actions can be taken promptly. One of the sensor nodes used in an LSWSN is called the sink node, which is responsible for processing and analyzing the collected information. It works as a station between the network sensor nodes and the administrator. Also, it is responsible for controlling the whole network. Determining the sink node location in an LSWSN is a challenging task, as it is crucial to the network lifetime, for keeping the network activity to the most possible extent. In this paper, the Harris’ hawks optimization (HHO) algorithm is employed to solve this problem and subsequently the Prim’s shortest path algorithm is used to reconstruct the network by making minimum transmission paths from the sink node to the rest of the sensor nodes. The performance of HHO is compared with other well-known algorithms such as particle swarm optimization (PSO), flower pollination algorithm (FPA), grey wolf optimizer (GWO), sine cosine algorithm (SCA), multi-verse optimizer (MVO), and whale optimization algorithm (WOA). The simulation results of different network sizes, with single and multiple sink nodes, show the superiority of the employed approach in terms of energy consumption and localization error, and ultimately prolonging the lifetime of the network in an efficacious way.","{'model': 'tldr@v2.0.0', 'text': 'The Harris’ hawks optimization (HHO) algorithm is employed and subsequently the Prim’s shortest path algorithm is used to reconstruct the network by making minimum transmission paths from the sink node to the rest of the sensor nodes.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/08966999.pdf
-development and optimization of naringenin-loaded chitosan-coated nanoemulsion for topical therapy in wound healing,S. Akrawi,"The potential role of naringenin (NAR), a natural flavonoid, in the treatment of chronic wound has prompted the present research to deliver the drug in nanoemulsion (NE) form, where synergistic role of chitosan was achieved through development of chitosan-coated NAR NE (CNNE). The NE consisted of Capryol 90, Tween 20 and Transcutol P, which was fabricated by low-energy emulsification method to encapsulate NAR within the oil core. The optimization of the formulated NEs was performed using Box–Behnken statistical design to obtain crucial variable parameters that influence globule size, size distribution and surface charge. Finally, the optimized formulation was coated with different concentrations of chitosan and subsequently characterized in vitro. The size of the CNNE was found to be increased when the drug-loaded formulation was coated with chitosan. Controlled release characteristics depicted 67–81% release of NAR from the CNNE, compared to 89% from the NE formulation. Cytotoxicity study of the formulation was performed in vitro using fibroblast cell line (NIH-3T3), where no inhibition in proliferation of the cells was observed with CNNE. Finally, the wound healing potential of the CNNE was evaluated in an abrasion-created wound model in experimental animals where the animals were treated and compared histologically at 0 and 14 days. Significant improvement in construction of the abrasion wound was observed when the animals were treated with formulated CNNE, whereas stimulation of skin regeneration was depicted in the histological examination. Therefore, it could be summarized that the chitosan coating of the developed NAR NE is a potential platform to accelerate healing of wounds.","{'model': 'tldr@v2.0.0', 'text': 'The chitosan coating of the developed NAR NE is a potential platform to accelerate healing of wounds and was evaluated in an abrasion-created wound model in experimental animals where the animals were treated with formulated CNNE.'}",https://www.mdpi.com/1999-4923/12/9/893/pdf?version=1600585179
-ram-pressure stripping of a kicked hill sphere: prompt electromagnetic emission from the merger of stellar mass black holes in an agn accretion disk,B. McKernan,"Accretion disks around supermassive black holes (SMBHs) are promising sites for stellar mass black hole (BH) mergers due to mass segregation and merger acceleration by disk gas torques. Here we show that a gravitational-wave (GW) kick at BH merger causes ram-pressure stripping of gas within the BH Hill sphere. If RH ≥ H, the disk height, an off-center UV flare at aBH ∼ 103rg, emerges within tUV ∼ O(2 days)(aBH/103rg)(MSMBH/108M⊙)(vkick/102 km s−1) postmerger and lasts O(RH/vkick) ∼ O(5tUV). The flare emerges with luminosity O(1042erg s−1)(tUV/2days)−1(MHill/1M⊙)(vkick/102 km s−1)2. Active galactic nucleus optical/UV photometry is altered and asymmetric broad emission line profiles can develop after weeks. If RH < H, detectability depends on disk optical depth. Follow-up by large optical sky surveys is optimized for small GW error volumes and for Laser Interferometer Gravitational-Wave Observatory/Virgo triggers >50M⊙.",,https://iopscience.iop.org/article/10.3847/2041-8213/ab4886/pdf
-uav relay-assisted emergency communications in iot networks: resource allocation and trajectory optimization,Dinh-Hieu Tran,"Unmanned aerial vehicle (UAV) communication has emerged as a prominent technology for emergency communications (e.g., natural disaster) in the Internet of Things (IoT) networks to enhance the ability of disaster prediction, damage assessment, and rescue operations promptly. A UAV can be deployed as a flying base station (BS) to collect data from time-constrained IoT devices and then transfer it to a ground gateway (GW). In general, the latency constraint at IoT devices and UAV’s limited storage capacity highly hinder practical applications of UAV-assisted IoT networks. In this paper, full-duplex (FD) radio is adopted at the UAV to overcome these challenges. In addition, half-duplex (HD) scheme for UAV-based relaying is also considered to provide a comparative study between two modes (viz., FD and HD). Herein, a device is considered to be successfully served if its data is collected by the UAV and conveyed to GW timely during flight time. In this context, we aim to maximize the number of served IoT devices by jointly optimizing bandwidth, power allocation, and the UAV trajectory while satisfying each device’s requirement and the UAV’s limited storage capacity. The formulated optimization problem is troublesome to solve due to its non-convexity and combinatorial nature. Towards appealing applications, we first relax binary variables into continuous ones and transform the original problem into a more computationally tractable form. By leveraging inner approximation framework, we derive newly approximated functions for non-convex parts and then develop a simple yet efficient iterative algorithm for its solutions. Next, we attempt to maximize the total throughput subject to the number of served IoT devices. Finally, numerical results show that the proposed algorithms significantly outperform benchmark approaches in terms of the number of served IoT devices and system throughput.","{'model': 'tldr@v2.0.0', 'text': 'Full-duplex (FD) radio is adopted at the UAV to overcome challenges of latency constraint at IoT devices and UAV’s limited storage capacity, and a simple yet efficient iterative algorithm is developed to maximize the total throughput subject to the number of served IoT devices.'}",https://ieeexplore.ieee.org/ielx7/7693/9731098/09522072.pdf
-prompt engineering for text-based generative art,J. Oppenlaender,"Text-based generative art has seen an explosion of interest in 2021. Online communities around text-based generative art as a novel digital medium have quickly emerged. This short paper identifies five types of prompt modifiers used by practitioners in the community of text-based generative art based on a 3-month ethnographic study on Twitter. The novel taxonomy of prompt modifiers provides researchers a conceptual starting point for investigating the practices of text-based generative art, but also may help practitioners of text-based generative art improve their images. The paper concludes with a discussion of research opportunities in the space of text-based generative art and the broader implications of prompt engineering from the perspective of human-AI interaction in future applications beyond the use case of text-based generative art.",,http://arxiv.org/pdf/2204.13988
-prompting ai art: an investigation into the creative skill of prompt engineering,J. Oppenlaender,"Humankind is entering a novel era of creativity - an era in which anybody can synthesize digital content. The paradigm under which this revolution takes place is prompt-based learning (or in-context learning). This paradigm has found fruitful application in text-to-image generation where it is being used to synthesize digital images from zero-shot text prompts in natural language for the purpose of creating AI art. This activity is referred to as prompt engineering - the practice of iteratively crafting prompts to generate and improve images. In this paper, we investigate prompt engineering as a novel creative skill for creating prompt-based art. In three studies with participants recruited from a crowdsourcing platform, we explore whether untrained participants could 1) recognize the quality of prompts, 2) write prompts, and 3) improve their prompts. Our results indicate that participants could assess the quality of prompts and respective images. This ability increased with the participants' experience and interest in art. Participants further were able to write prompts in rich descriptive language. However, even though participants were specifically instructed to generate artworks, participants' prompts were missing the specific vocabulary needed to apply a certain style to the generated images. Our results suggest that prompt engineering is a learned skill that requires expertise and practice. Based on our findings and experience with running our studies with participants recruited from a crowdsourcing platform, we provide ten recommendations for conducting experimental research on text-to-image generation and prompt engineering with a paid crowd. Our studies offer a deeper understanding of prompt engineering thereby opening up avenues for research on the future of prompt engineering. We conclude by speculating on four possible futures of prompt engineering.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates prompt engineering as a novel creative skill for creating prompt-based art and suggests that prompt engineering is a learned skill that requires expertise and practice.'}",http://arxiv.org/pdf/2303.13534
-grimm in wonderland: prompt engineering with midjourney to illustrate fairytales,M. Ruskov,"The quality of text-to-image generation is continuously improving, yet the boundaries of its applicability are still unclear. In particular, refinement of the text input with the objective of achieving better results - commonly called prompt engineering - so far seems to have not been geared towards work with pre-existing texts. We investigate whether text-to-image generation and prompt engineering could be used to generate basic illustrations of popular fairytales. Using Midjourney v4, we engage in action research with a dual aim: to attempt to generate 5 believable illustrations for each of 5 popular fairytales, and to define a prompt engineering process that starts from a pre-existing text and arrives at an illustration of it. We arrive at a tentative 4-stage process: i) initial prompt, ii) composition adjustment, iii) style refinement, and iv) variation selection. We also discuss three reasons why the generation model struggles with certain illustrations: difficulties with counts, bias from stereotypical configurations and inability to depict overly fantastic situations. Our findings are not limited to the specific generation model and are intended to be generalisable to future ones.",,https://arxiv.org/pdf/2302.08961
-prompt engineering in medical education,Thomas F. Heston,"Artificial intelligence-powered generative language models (GLMs), such as ChatGPT, Perplexity AI, and Google Bard, have the potential to provide personalized learning, unlimited practice opportunities, and interactive engagement 24/7, with immediate feedback. However, to fully utilize GLMs, properly formulated instructions are essential. Prompt engineering is a systematic approach to effectively communicating with GLMs to achieve the desired results. Well-crafted prompts yield good responses from the GLM, while poorly constructed prompts will lead to unsatisfactory responses. Besides the challenges of prompt engineering, significant concerns are associated with using GLMs in medical education, including ensuring accuracy, mitigating bias, maintaining privacy, and avoiding excessive reliance on technology. Future directions involve developing more sophisticated prompt engineering techniques, integrating GLMs with other technologies, creating personalized learning pathways, and researching the effectiveness of GLMs in medical education.","{'model': 'tldr@v2.0.0', 'text': None}",https://www.mdpi.com/2813-141X/2/3/19/pdf?version=1693479951
-"multi-party goal tracking with llms: comparing pre-training, fine-tuning, and prompt engineering",Angus Addlesee,"This paper evaluates the extent to which current LLMs can capture task-oriented multi-party conversations (MPCs). We have recorded and transcribed 29 MPCs between patients, their companions, and a social robot in a hospital. We then annotated this corpus for multi-party goal-tracking and intent-slot recognition. People share goals, answer each other’s goals, and provide other people’s goals in MPCs - none of which occur in dyadic interactions. To understand user goals in MPCs, we compared three methods in zero-shot and few-shot settings: we fine-tuned T5, created pre-training tasks to train DialogLM using LED, and employed prompt engineering techniques with GPT-3.5-turbo, to determine which approach can complete this novel task with limited data. GPT-3.5-turbo significantly outperformed the others in a few-shot setting. The ‘reasoning’ style prompt, when given 7% of the corpus as example annotated conversations, was the best performing method. It correctly annotated 62.32% of the goal tracking MPCs, and 69.57% of the intent-slot recognition MPCs. A ‘story’ style prompt increased model hallucination, which could be detrimental if deployed in safety-critical settings. We conclude that multi-party conversations still challenge state-of-the-art LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that multi-party conversations still challenge state-of-the-art LLMs and could be detrimental if deployed in safety-critical settings.'}",https://arxiv.org/pdf/2308.15231
-real estate insights unleashing the potential of chatgpt in property valuation reports: the “red book” compliance chain-of-thought (cot) prompt engineering,K. Cheung,"PurposeThis viewpoint article explores the transformative capabilities of large language models (LLMs) like the Chat Generative Pre-training Transformer (ChatGPT) within the property valuation industry. It particularly accentuates the pivotal role of prompt engineering in facilitating valuation reporting and advocates for adopting the “Red Book” compliance Chain-of-thought (COT) prompt engineering as a gold standard for generating AI-facilitated valuation reports.Design/methodology/approachThe article offers a high-level examination of the application of LLMs in real estate research, highlighting the essential role of prompt engineering for future advancements in generative AI. It explores the collaborative dynamic between valuers and AI advancements, emphasising the importance of precise instructions and contextual cues in directing LLMs to generate accurate and reproducible valuation outcomes.FindingsIntegrating LLMs into property valuation processes paves the way for efficiency improvements and task automation, such as generating reports and drafting contracts. AI-facilitated reports offer unprecedented transparency and elevate client experiences. The fusion of valuer expertise with prompt engineering ensures the reliability and interpretability of valuation reports.Practical implicationsDelineating the types and versions of LLMs used in AI-generated valuation reports encourage the adoption of transparency best practices within the industry. Valuers, as expert prompt engineers, can harness the potential of AI to enhance efficiency, accuracy and transparency in the valuation process, delivering significant benefits to a broad array of stakeholders.Originality/valueThe article elucidates the substantial impact of prompt engineering in leveraging LLMs within the property industry. It underscores the importance of valuers training their unique GPT models, enabling customisation and reproducibility of valuation outputs. The symbiotic relationship between valuers and LLMs is identified as a key driver shaping the future of property valuations.",,
-improving formality-sensitive machine translation using data-centric approaches and prompt engineering,Seugnjun Lee,"In this paper, we present the KU x Upstage team’s submission for the Special Task on Formality Control on Spoken Language Translation, which involves translating English into four languages with diverse grammatical formality markers. Our methodology comprises two primary components: 1) a language-specific data-driven approach, and 2) the generation of synthetic data through the employment of large-scale language models and empirically-grounded prompt engineering. By adapting methodologies and models to accommodate the unique linguistic properties of each language, we observe a notable enhancement in performance relative to the baseline, substantiating the heightened efficacy of data-driven approaches. Moreover, our devised prompt engineering strategy yields superior synthetic translation instances.","{'model': 'tldr@v2.0.0', 'text': 'The KU x Upstage team’s submission for the Special Task on Formality Control on Spoken Language Translation, which involves translating English into four languages with diverse grammatical formality markers, shows a notable enhancement in performance relative to the baseline, substantiating the heightened efficacy of data-driven approaches.'}",https://aclanthology.org/2023.iwslt-1.40.pdf
-artificial intelligence prompt engineering as a new digital competence: analysis of generative ai technologies such as chatgpt,P. Korzyński,"Objective: The article aims to offer a thorough examination and comprehension of the challenges and pro‐ spects connected with artificial intelligence (AI) prompt engineering. Our research aimed to create a theoret‐ ical framework that would highlight optimal approaches in the field of AI prompt engineering. Research Design & Methods: This research utilized a narrative and critical literature review and established a conceptual framework derived from existing literature taking into account both academic and practitioner sources. This article should be regarded as a conceptual work that emphasizes the best practices in the domain of AI prompt engineering. Findings: Based on the conducted deep and extensive query of academic and practitioner literature on the subject, as well as professional press and Internet portals, we identified various insights for effective AI prompt engineering. We provide specific prompting strategies. Implications & Recommendations: The study revealed the profound implications of AI prompt engineering across various domains such as entrepreneurship, art, science, and healthcare. We demonstrated how the effective crafting of prompts can significantly enhance the performance of large language models (LLMs), gen‐ erating more accurate and contextually relevant results. Our findings offer valuable insights for AI practition‐ ers, researchers, educators, and organizations integrating AI into their operations, emphasizing the need to invest time and resources in prompt engineering. Moreover, we contributed the AI PROMPT framework to the field, providing clear and actionable guidelines for text‐to‐text prompt engineering. Contribution & Value Added: The value of this study lies in its comprehensive exploration of AI prompt engineer‐ ing as a digital competence. By building upon existing research and prior literature, this study aimed to provide a deeper understanding of the intricacies involved in AI prompt engineering and its role as a digital competence. Article",,https://eber.uek.krakow.pl/index.php/eber/article/view/2142/863
-cases of efl secondary students' prompt engineering pathways to complete a writing task with chatgpt,D. Woo,"ChatGPT is a state-of-the-art (SOTA) chatbot. Although it has potential to support English as a foreign language (EFL) students' writing, to effectively collaborate with it, a student must learn to engineer prompts, that is, the skill of crafting appropriate instructions so that ChatGPT produces desired outputs. However, writing an appropriate prompt for ChatGPT is not straightforward for non-technical users who suffer a trial-and-error process. This paper examines the content of EFL students' ChatGPT prompts when completing a writing task and explores patterns in the quality and quantity of the prompts. The data come from iPad screen recordings of secondary school EFL students who used ChatGPT and other SOTA chatbots for the first time to complete the same writing task. The paper presents a case study of four distinct pathways that illustrate the trial-and-error process and show different combinations of prompt content and quantity. The cases contribute evidence for the need to provide prompt engineering education in the context of the EFL writing classroom, if students are to move beyond an individual trial-and-error process, learning a greater variety of prompt content and more sophisticated prompts to support their writing.",,https://arxiv.org/pdf/2307.05493
-enhancing automated program repair through fine-tuning and prompt engineering,Rishov Paul,"Sequence-to-sequence models have been used to transform erroneous programs into correct ones when trained with a large enough dataset. Some recent studies also demonstrated strong empirical evidence that code review could improve the program repair further. Large language models, trained with Natural Language (NL) and Programming Language (PL), can contain inherent knowledge of both. In this study, we investigate if this inherent knowledge of PL and NL can be utilized to improve automated program repair. We applied PLBART and CodeT5, two state-of-the-art language models that are pre-trained with both PL and NL, on two such natural language-based program repair datasets and found that the pre-trained language models fine-tuned with datasets containing both code review and subsequent code changes notably outperformed each of the previous models. With the advent of code generative models like Codex and GPT-3.5-Turbo, we also performed zero-shot and few-shots learning-based prompt engineering to assess their performance on these datasets. However, the practical application of using LLMs in the context of automated program repair is still a long way off based on our manual analysis of the generated repaired codes by the learning models.","{'model': 'tldr@v2.0.0', 'text': 'This study applied PLBART and CodeT5, two state-of-the-art language models that are pre- trained with both PL and NL, on two such natural language-based program repair datasets and found that the pre-trained language models fine-tuned with datasets containing both code review and subsequent code changes notably outperformed each of the previous models.'}",
-"supporting self-directed learning and self-assessment using teachergaia, a generative ai chatbot application: learning approaches and prompt engineering",Farhan Ali,"ABSTRACT Self-directed learning and self-assessment require student responsibility over learning needs, goals, processes, and outcomes. However, this student-led learning can be challenging to achieve in a classroom limited by a one-to-many teacher-led instruction. We, thus, have designed and prototyped a generative artificial intelligence chatbot application (GAIA), named TeacherGAIA, that can be used to asynchronously support students in their self-directed learning and self-assessment outside the classroom. We first identified diverse constructivist learning approaches that align with, and promote, student-led learning. These included knowledge construction, inquiry-based learning, self-assessment, and peer teaching. The in-context learning abilities of large language model (LLM) from OpenAI were then leveraged via prompt engineering to steer interactions supporting these different learning approaches. These interactions contrasted with ChatGPT, OpenAI’s chatbot which by default engaged in the traditional transmissionist mode of learning reminiscent of teacher-led instruction. Preliminary design, prompt engineering and prototyping suggested fidelity to the learning approaches, cognitive guidance, and social-emotional support, all of which were implemented in a generative AI manner without pre-specified rules or “hard-coding”. Other affordances of TeacherGAIA are discussed and future development outlined. We anticipate TeacherGAIA to be a useful application for teachers in facilitating self-directed learning and self-assessment among K-12 students.",,
-prompting meaning: a hermeneutic approach to optimising prompt engineering with chatgpt,Leah Henrickson,,,https://link.springer.com/content/pdf/10.1007/s00146-023-01752-8.pdf
-ncu-iisr: prompt engineering on gpt-4 to stove biological problems in bioasq 11b phase b,Chun-Yu Hsueh,"In this paper, we present our system applied in BioASQ 11b phase b. We showcase prompt engineering strategies and outline our experimental steps. Building upon the success of ChatGPT/GPT-4 in answer generation and the field of biology, we developed a system that utilizes GPT-4 to answer biomedical questions. The system leverages OpenAI’s ChatCompletions API and combines Prompt Engineering methods to explore various prompts. In addition, we also attempted to incorporate GPT-4 into our system from last year, which combines a BERT-based model and BERTScore. However, the standalone GPT-4 method outperformed this approach by a large margin. Ultimately, in our submission, we adopted what we believe to be the optimal prompts and achieved the highest scores in the second batch.","{'model': 'tldr@v2.0.0', 'text': 'A system that utilizes GPT-4 to answer biomedical questions and leverages OpenAI’s ChatCompletions API and combines Prompt Engineering methods to explore various prompts and adopted what it believes to be the optimal prompts.'}",
-"optimizing mobile-edge ai-generated everything (aigx) services by prompt engineering: fundamental, framework, and case study",Yinqiu Liu,"As the next-generation paradigm for content creation, AI-Generated Content (AIGC), i.e., generating content automatically by Generative AI (GAI) based on user prompts, has gained great attention and success recently. With the ever-increasing power of GAI, especially the emergence of Pretrained Foundation Models (PFMs) that contain billions of parameters and prompt engineering methods (i.e., finding the best prompts for the given task), the application range of AIGC is rapidly expanding, covering various forms of information for human, systems, and networks, such as network designs, channel coding, and optimization solutions. In this article, we present the concept of mobile-edge AI-Generated Everything (AIGX). Specifically, we first review the building blocks of AIGX, the evolution from AIGC to AIGX, as well as practical AIGX applications. Then, we present a unified mobile-edge AIGX framework, which employs edge devices to provide PFM-empowered AIGX services and optimizes such services via prompt engineering. More importantly, we demonstrate that suboptimal prompts lead to poor generation quality, which adversely affects user satisfaction, edge network performance, and resource utilization. Accordingly, we conduct a case study, showcasing how to train an effective prompt optimizer using ChatGPT and investigating how much improvement is possible with prompt engineering in terms of user experience, quality of generation, and network performance.","{'model': 'tldr@v2.0.0', 'text': 'A unified mobile-edge AIGX framework is presented, which employs edge devices to provide PFM-empowered AigX services and optimizes such services via prompt engineering and demonstrates that suboptimal prompts lead to poor generation quality, which adversely affects user satisfaction, edge network performance, and resource utilization.'}",https://arxiv.org/pdf/2309.01065
-prompt engineering as an important emerging skill for medical professionals: tutorial,B. Meskó,"Prompt engineering is a relatively new field of research that refers to the practice of designing, refining, and implementing prompts or instructions that guide the output of large language models (LLMs) to help in various tasks. With the emergence of LLMs, the most popular one being ChatGPT that has attracted the attention of over a 100 million users in only 2 months, artificial intelligence (AI), especially generative AI, has become accessible for the masses. This is an unprecedented paradigm shift not only because of the use of AI becoming more widespread but also due to the possible implications of LLMs in health care. As more patients and medical professionals use AI-based tools, LLMs being the most popular representatives of that group, it seems inevitable to address the challenge to improve this skill. This paper summarizes the current state of research about prompt engineering and, at the same time, aims at providing practical recommendations for the wide range of health care professionals to improve their interactions with LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This paper summarizes the current state of research about prompt engineering and aims at providing practical recommendations for the wide range of health care professionals to improve their interactions with LLMs.'}",
-exploring the intersection of large language models and agent-based modeling via prompt engineering,Edward Junprung,"The final frontier for simulation is the accurate representation of complex, real-world social systems. While agent-based modeling (ABM) seeks to study the behavior and interactions of agents within a larger system, it is unable to faithfully capture the full complexity of human-driven behavior. Large language models (LLMs), like ChatGPT, have emerged as a potential solution to this bottleneck by enabling researchers to explore human-driven interactions in previously unimaginable ways. Our research investigates simulations of human interactions using LLMs. Through prompt engineering, inspired by Park et al. (2023), we present two simulations of believable proxies of human behavior: a two-agent negotiation and a six-agent murder mystery game.","{'model': 'tldr@v2.0.0', 'text': 'This research investigates simulations of human interactions using large language models using LLMs and presents two simulations of believable proxies of human behavior: a two-agent negotiation and a six-agent murder mystery game.'}",https://arxiv.org/pdf/2308.07411
-the prompt engineering librarian,Brady Lund," Purpose In terms of training the public in prompt engineering skills, no single discipline or profession currently takes the lead, presenting an opportunity for professions like librarianship to step into this role. Librarians are already well-equipped to educate the public in a wide range of literacy skills and tasks, so prompt engineering may be a natural progression. The purpose of this paper is to examine the potential role of prompt engineering for library professionals. Design/methodology/approach Prompt engineering is the process of optimizing the text that is provided to an artificial intelligence (A)I model to ensure proper interpretation and the generation of relevant, detailed results. The field of prompt engineering is relatively young, evolving alongside the growth of large language models like ChatGPT and BARD. This conceptual paper will explore prompt engineering as a possible domain of expertise for librarians. Findings This paper delves into the world of prompt engineering, its alignment with the existing roles and expertise of librarians, and the potential emergence of a new role known as the “prompt engineering librarian,” akin to the well-established “information literacy librarian” role that has gained prominence in recent decades. Originality/value The significance of this work lies in exploring the synergy between prompt engineering and the traditional roles of librarians, highlighting the potential for a new and valuable profession in the form of prompt engineering librarians. This innovative concept could bridge the gap in AI literacy and facilitate more effective interactions with AI systems, contributing to the broader goal of AI accessibility and understanding. ",,
-contextual stance classification using prompt engineering,Felipe Penhorate Carvalho de Fonseca,"This paper introduces a prompt-based method for few-shot learning addressing, as an application example, contextual stance classification, that is, the task of determining the attitude expressed by a given statement within a conversation thread with multiple points of view towards another statement. More specifically, we envisaged a method that uses the existing conversation thread (i.e., messages that are part of the test data) to create natural language prompts for few-shot learning with minimal reliance on training samples, whose preliminary results suggest that prompt engineering may be a competitive alternative to supervised methods both in terms of accuracy and development costs for the task at hand.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method that uses the existing conversation thread to create natural language prompts for few-shot learning with minimal reliance on training samples is introduced, whose preliminary results suggest that prompt engineering may be a competitive alternative to supervised methods both in terms of accuracy and development costs for the task at hand.'}",https://sol.sbc.org.br/index.php/stil/article/download/25435/25256
-promptmagician: interactive prompt engineering for text-to-image creation,Yingchaojie Feng,"Generative text-to-image models have gained great popularity among the public for their powerful capability to generate high-quality images based on natural language prompts. However, developing effective prompts for desired images can be challenging due to the complexity and ambiguity of natural language. This research proposes PromptMagician, a visual analysis system that helps users explore the image results and refine the input prompts. The backbone of our system is a prompt recommendation model that takes user prompts as input, retrieves similar prompt-image pairs from DiffusionDB, and identifies special (important and relevant) prompt keywords. To facilitate interactive prompt refinement, PromptMagician introduces a multi-level visualization for the cross-modal embedding of the retrieved images and recommended keywords, and supports users in specifying multiple criteria for personalized exploration. Two usage scenarios, a user study, and expert interviews demonstrate the effectiveness and usability of our system, suggesting it facilitates prompt engineering and improves the creativity support of the generative text-to-image model.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes PromptMagician, a visual analysis system that helps users explore the image results and refine the input prompts, and introduces a multi-level visualization for the cross-modal embedding of the retrieved images and recommended keywords.'}",https://arxiv.org/pdf/2307.09036
-logprompt: prompt engineering towards zero-shot and interpretable log analysis,Yilun Liu,"Automated log analysis is crucial in modern software-intensive systems for ensuring reliability and resilience throughout software maintenance and engineering life cycles. Existing methods perform tasks such as log parsing and log anomaly detection by providing a single prediction value without interpretation. However, given the increasing volume of system events, the limited interpretability of analysis results hinders analysts' trust and their ability to take appropriate actions. Moreover, these methods require substantial in-domain training data, and their performance declines sharply (by up to 62.5%) in online scenarios involving unseen logs from new domains, a common occurrence due to rapid software updates. In this paper, we propose LogPrompt, a novel zero-shot and interpretable log analysis approach. LogPrompt employs large language models (LLMs) to perform zero-shot log analysis tasks via a suite of advanced prompt strategies tailored for log tasks, which enhances LLMs' performance by up to 107.5% compared with simple prompts. Experiments on nine publicly available evaluation datasets across two tasks demonstrate that LogPrompt, despite using no training data, outperforms existing approaches trained on thousands of logs by up to around 50%. We also conduct a human evaluation of LogPrompt's interpretability, with six practitioners possessing over 10 years of experience, who highly rated the generated content in terms of usefulness and readability (averagely 4.42/5). LogPrompt also exhibits remarkable compatibility with open-source and smaller-scale LLMs, making it flexible for practical deployment.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on nine publicly available evaluation datasets across two tasks demonstrate that LogPrompt, despite using no training data, outperforms existing approaches trained on thousands of logs by up to around 50%.'}",https://arxiv.org/pdf/2308.07610
-"the artificially intelligent entrepreneur: chatgpt, prompt engineering, and entrepreneurial rhetoric creation",Cole E. Short,,,
-a survey on segment anything model (sam): vision foundation model meets prompt engineering,Chaoning Zhang,"Segment anything model (SAM) developed by Meta AI Research has recently attracted significant attention. Trained on a large segmentation dataset of over 1 billion masks, SAM is capable of segmenting any object on a certain image. In the original SAM work, the authors turned to zero-short transfer tasks (like edge detection) for evaluating the performance of SAM. Recently, numerous works have attempted to investigate the performance of SAM in various scenarios to recognize and segment objects. Moreover, numerous projects have emerged to show the versatility of SAM as a foundation model by combining it with other models, like Grounding DINO, Stable Diffusion, ChatGPT, etc. With the relevant papers and projects increasing exponentially, it is challenging for the readers to catch up with the development of SAM. To this end, this work conducts the first yet comprehensive survey on SAM. This is an ongoing project and we intend to update the manuscript on a regular basis. Therefore, readers are welcome to contact us if they complete new works related to SAM so that we can include them in our next version.","{'model': 'tldr@v2.0.0', 'text': 'The first yet comprehensive survey on SAM is conducted, to show the versatility of SAM as a foundation model by combining it with other models, like Grounding DINO, Stable Diffusion, ChatGPT, etc.'}",http://arxiv.org/pdf/2306.06211
-the clear path: a framework for enhancing information literacy through prompt engineering,Leo S. Lo,,,
-prompt engineering for large language models,Andrew Gao,,,
-exploring the effectiveness of prompt engineering for legal reasoning tasks,Fang Yu,",","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2023.findings-acl.858.pdf
-abstractive summarization evaluation for prompt engineering,Shayak Chakraborty,,,
-prompt engineering for narrative choice generation,Sarah Harmon,,,
-plain template insertion: korean-prompt-based engineering for few-shot learners,Jaehyung Seo,"Prompt-based learning is a method used for language models to interpret natural language by remembering the prior knowledge acquired and the training objective. Recent prompt-based few-shot learners have achieved superior performance by alleviating the catastrophic forgetting that occurs in pretrained language models. Few-shot learning contributes towards solving the data scarcity problem, an enormous challenge in AI systems and a significant consideration in natural language processing research. In spite of the significance of few-shot learning, research on Korean language-based few-shot learning is insufficient, and whether the prompt-based approach is appropriate for the Korean language has not been thoroughly verified. As a step toward realizing a Korean-prompt-based few-shot learner, we attempt to apply prompt engineering to the Korean language understanding benchmark dataset and introduce plain template insertion to overcome data scarcity in a more practical few-shot setting. The contributions of this study are as follows: (1) presumably, this is the first study to apply prompt-based few-shot learning to Korean benchmark datasets. With 32 few-shot settings, it improves performance by +14.88, +29.04, and +1.81 in the natural language inference, semantic textual similarity, and topic classification tasks. (2) We present prompt engineering, which merely inserts a plain template and increases data efficiency without training example selection, augmentation, reformulation, and retrieval. (3) Our approach is robust to the Korean prompt’s contextual information and sentence structure and is applicable to both hard- and soft-prompt.","{'model': 'tldr@v2.0.0', 'text': 'This study attempts to apply prompt engineering to the Korean language understanding benchmark dataset and introduces plain template insertion to overcome data scarcity in a more practical few-shot setting.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09913979.pdf
-polyglot prompt: multilingual multitask prompt training,Jinlan Fu,"This paper aims for a potential architectural improvement for multilingual learning and asks: Can different tasks from different languages be modeled in a monolithic framework, i.e. without any task/language-specific module? The benefit of achieving this could open new doors for future multilingual research, including allowing systems trained on low resources to be further assisted by other languages as well as other tasks. We approach this goal by developing a learning framework named Polyglot Prompting to exploit prompting methods for learning a unified semantic space for different languages and tasks with multilingual prompt engineering. We performed a comprehensive evaluation of 6 tasks, namely topic classification, sentiment classification, named entity recognition, question answering, natural language inference, and summarization, covering 24 datasets and 49 languages. The experimental results demonstrated the efficacy of multilingual multitask prompt-based learning and led to inspiring observations. We also present an interpretable multilingual evaluation methodology and show how the proposed framework, multilingual multitask prompt training, works. We release all datasets prompted in the best setting and code.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops a learning framework named Polyglot Prompting to exploit prompting methods for learning a unified semantic space for different languages and tasks with multilingual prompt engineering, and presents an interpretable multilingual evaluation methodology.'}",https://aclanthology.org/2022.emnlp-main.674.pdf
-"chatgpt prompt patterns for improving code quality, refactoring, requirements elicitation, and software design",Jules White,"This paper presents prompt design techniques for software engineering, in the form of patterns, to solve common problems when using large language models (LLMs), such as ChatGPT to automate common software engineering activities, such as ensuring code is decoupled from third-party libraries and simulating a web application API before it is implemented. This paper provides two contributions to research on using LLMs for software engineering. First, it provides a catalog of patterns for software engineering that classifies patterns according to the types of problems they solve. Second, it explores several prompt patterns that have been applied to improve requirements elicitation, rapid prototyping, code quality, refactoring, and system design.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2303.07839
-retrieval-based prompt selection for code-related few-shot learning,Noor Nashid,"Large language models trained on massive code corpora can generalize to new tasks without the need for task-specific fine-tuning. In few-shot learning, these models take as input a prompt, composed of natural language instructions, a few instances of task demonstration, and a query and generate an output. However, the creation of an effective prompt for code-related tasks in few-shot learning has received little attention. We present a technique for prompt creation that automatically retrieves code demonstrations similar to the developer task, based on embedding or frequency analysis. We apply our approach, Cedar, to two different programming languages, statically and dynamically typed, and two different tasks, namely, test assertion generation and program repair. For each task, we compare Cedar with state-of-the-art task-specific and fine-tuned models. The empirical results show that, with only a few relevant code demonstrations, our prompt creation technique is effective in both tasks with an accuracy of 76% and 52% for exact matches in test assertion generation and program repair tasks, respectively. For assertion generation, Cedar outperforms existing task-specific and fine-tuned models by 333% and 11%, respectively. For program repair, Cedar yields 189% better accuracy than task-specific models and is competitive with recent fine-tuned models. These findings have practical implications for practitioners, as Cedar could potentially be applied to multilingual and multitask settings without task or language-specific training with minimal examples and effort.","{'model': 'tldr@v2.0.0', 'text': 'A technique for prompt creation that automatically retrieves code demonstrations similar to the developer task, based on embedding or frequency analysis is presented, which could potentially be applied to multilingual and multitask settings without task or language-specific training with minimal examples and effort.'}",
-"a study on prompt design, advantages and limitations of chatgpt for deep learning program repair",Jialun Cao,"ChatGPT has revolutionized many research and industrial fields. ChatGPT has shown great potential in software engineering to boost various traditional tasks such as program repair, code understanding, and code generation. However, whether automatic program repair (APR) applies to deep learning (DL) programs is still unknown. DL programs, whose decision logic is not explicitly encoded in the source code, have posed unique challenges to APR. While to repair DL programs, an APR approach needs to not only parse the source code syntactically but also needs to understand the code intention. With the best prior work, the performance of fault localization is still far less than satisfactory (only about 30\%). Therefore, in this paper, we explore ChatGPT's capability for DL program repair by asking three research questions. (1) Can ChatGPT debug DL programs effectively? (2) How can ChatGPT's repair performance be improved by prompting? (3) In which way can dialogue help facilitate the repair? On top of that, we categorize the common aspects useful for prompt design for DL program repair. Also, we propose various prompt templates to facilitate the performance and summarize the advantages and disadvantages of ChatGPT's abilities such as detecting bad code smell, code refactoring, and detecting API misuse/deprecation.","{'model': 'tldr@v2.0.0', 'text': ""This paper explores ChatGPT's capability for DL program repair by asking three research questions and categorizes the common aspects useful for prompt design forDL program repair.""}",http://arxiv.org/pdf/2304.08191
-ip-adapter: text compatible image prompt adapter for text-to-image diffusion models,Hu Ye,"Recent years have witnessed the strong power of large text-to-image diffusion models for the impressive generative capability to create high-fidelity images. However, it is very tricky to generate desired images using only text prompt as it often involves complex prompt engineering. An alternative to text prompt is image prompt, as the saying goes:""an image is worth a thousand words"". Although existing methods of direct fine-tuning from pretrained models are effective, they require large computing resources and are not compatible with other base models, text prompt, and structural controls. In this paper, we present IP-Adapter, an effective and lightweight adapter to achieve image prompt capability for the pretrained text-to-image diffusion models. The key design of our IP-Adapter is decoupled cross-attention mechanism that separates cross-attention layers for text features and image features. Despite the simplicity of our method, an IP-Adapter with only 22M parameters can achieve comparable or even better performance to a fully fine-tuned image prompt model. As we freeze the pretrained diffusion model, the proposed IP-Adapter can be generalized not only to other custom models fine-tuned from the same base model, but also to controllable generation using existing controllable tools. With the benefit of the decoupled cross-attention strategy, the image prompt can also work well with the text prompt to achieve multimodal image generation. The project page is available at \url{https://ip-adapter.github.io}.","{'model': 'tldr@v2.0.0', 'text': 'The proposed IP-Adapter is an effective and lightweight adapter to achieve image prompt capability for the pretrained text-to-image diffusion models and has the benefit of the decoupled cross-attention strategy, the image prompt can also work well with the text prompt to achieve multimodal image generation.'}",https://arxiv.org/pdf/2308.06721
-prompt space optimizing few-shot reasoning success with large language models,Fobo Shi,"Prompt engineering is an essential technique for enhancing the abilities of large language models (LLMs) by providing explicit and specific instructions. It enables LLMs to excel in various tasks, such as arithmetic reasoning, question answering, summarization, relation extraction, machine translation, and sentiment analysis. Researchers have been actively exploring different prompt engineering strategies, such as Chain of Thought (CoT), Zero-CoT, and In-context learning. However, an unresolved problem arises from the fact that current approaches lack a solid theoretical foundation for determining optimal prompts. To address this issue in prompt engineering, we propose a new and effective approach called Prompt Space. Our methodology utilizes text embeddings to obtain basis vectors by matrix decomposition, and then constructs a space for representing all prompts. Prompt Space significantly outperforms state-of-the-art prompt paradigms on ten public reasoning benchmarks. Notably, without the help of the CoT method and the prompt""Let's think step by step"", Prompt Space shows superior performance over the few-shot method. Overall, our approach provides a robust and fundamental theoretical framework for selecting simple and effective prompts. This advancement marks a significant step towards improving prompt engineering for a wide variety of applications in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Prompt Space methodology utilizes text embeddings to obtain basis vectors by matrix decomposition, and then constructs a space for representing all prompts, which significantly outperforms state-of-the-art prompt paradigms on ten public reasoning benchmarks.'}",http://arxiv.org/pdf/2306.03799
-"prompt-free diffusion: taking ""text"" out of text-to-image diffusion models",Xingqian Xu,"Text-to-image (T2I) research has grown explosively in the past year, owing to the large-scale pre-trained diffusion models and many emerging personalization and editing approaches. Yet, one pain point persists: the text prompt engineering, and searching high-quality text prompts for customized results is more art than science. Moreover, as commonly argued:""an image is worth a thousand words""- the attempt to describe a desired image with texts often ends up being ambiguous and cannot comprehensively cover delicate visual details, hence necessitating more additional controls from the visual domain. In this paper, we take a bold step forward: taking""Text""out of a pre-trained T2I diffusion model, to reduce the burdensome prompt engineering efforts for users. Our proposed framework, Prompt-Free Diffusion, relies on only visual inputs to generate new images: it takes a reference image as""context"", an optional image structural conditioning, and an initial noise, with absolutely no text prompt. The core architecture behind the scene is Semantic Context Encoder (SeeCoder), substituting the commonly used CLIP-based or LLM-based text encoder. The reusability of SeeCoder also makes it a convenient drop-in component: one can also pre-train a SeeCoder in one T2I model and reuse it for another. Through extensive experiments, Prompt-Free Diffusion is experimentally found to (i) outperform prior exemplar-based image synthesis approaches; (ii) perform on par with state-of-the-art T2I models using prompts following the best practice; and (iii) be naturally extensible to other downstream applications such as anime figure generation and virtual try-on, with promising quality. Our code and models are open-sourced at https://github.com/SHI-Labs/Prompt-Free-Diffusion.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.16223
-what initiates evidence‐based reasoning?: situations that prompt students to support their design ideas and decisions,Emilie A. Siverling,"As engineering becomes increasingly incorporated into precollege classrooms, it is important to explore students' ability to engage in engineering practices. One of these practices, engaging in argument from evidence, has been well studied in science education. However, it has not yet been fully explored in engineering education.",,
-bim-gpt: a prompt-based virtual assistant framework for bim information retrieval,Junwen Zheng,"Efficient information retrieval (IR) from building information models (BIMs) poses significant challenges due to the necessity for deep BIM knowledge or extensive engineering efforts for automation. We introduce BIM-GPT, a prompt-based virtual assistant (VA) framework integrating BIM and generative pre-trained transformer (GPT) technologies to support NL-based IR. A prompt manager and dynamic template generate prompts for GPT models, enabling interpretation of NL queries, summarization of retrieved information, and answering BIM-related questions. In tests on a BIM IR dataset, our approach achieved 83.5% and 99.5% accuracy rates for classifying NL queries with no data and 2% data incorporated in prompts, respectively. Additionally, we validated the functionality of BIM-GPT through a VA prototype for a hospital building. This research contributes to the development of effective and versatile VAs for BIM IR in the construction industry, significantly enhancing BIM accessibility and reducing engineering efforts and training data requirements for processing NL queries.","{'model': 'tldr@v2.0.0', 'text': 'BIM-GPT, a prompt-based virtual assistant framework integrating BIM and generative pre-trained transformer technologies to support NL-based IR, is introduced, significantly enhancing BIM accessibility and reducing engineering efforts and training data requirements for processing NL queries.'}",http://arxiv.org/pdf/2304.09333
-api entity and relation joint extraction from text via dynamic prompt-tuned language model,Qing Huang,"Extraction of Application Programming Interfaces (APIs) and their semantic relations from unstructured text (e.g., Stack Overflow) is a fundamental work for software engineering tasks (e.g., API recommendation). However, existing approaches are rule-based and sequence-labeling based. They must manually enumerate the rules or label data for a wide range of sentence patterns, which involves a significant amount of labor overhead and is exacerbated by morphological and common-word ambiguity. In contrast to matching or labeling API entities and relations, this paper formulates heterogeneous API extraction and API relation extraction task as a sequence-to-sequence generation task, and proposes AERJE, an API entity-relation joint extraction model based on the large pre-trained language model. After training on a small number of ambiguous but correctly labeled data, AERJE builds a multi-task architecture that extracts API entities and relations from unstructured text using dynamic prompts. We systematically evaluate AERJE on a set of long and ambiguous sentences from Stack Overflow. The experimental results show that AERJE achieves high accuracy and discrimination ability in API entity-relation joint extraction, even with zero or few-shot fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'This paper formulates heterogeneous API extraction and API relation extraction task as a sequence-to-sequence generation task, and proposes AERJE, an API entity-relation joint extraction model based on the large pre-trained language model.'}",https://dl.acm.org/doi/pdf/10.1145/3607188
-engineering a second-order dna logic-gated nanorobot to sense-then-release on live cell membranes for multiplexed diagnosis and synergistic therapy.,Dan Wang,"Tumor biomarker-based theranostics has achieved broad interest and success in recent years. However, single biomarker-based recognition can cause false-positive feedback, including the on-target off-tumor phenomenon by the absence of tumor-specific antigen. Besides, multi-biomarker-based recognition molecules often elicit nonspecific and undesired internalization when they bind to ''bystander'' cells. Herein, we report a universal DNA tetrahedral scaffold (DTS) that anchors on the cell membrane to load multiple aptamers and therapeutics for precise and effective theranostics. This DNA logic-gated nanorobot (DLGN) not only facilitates the precise discrimination among five cell lines, but also triggers the synergistic killing of effector aptamer-tethered synergistic drugs (EASDs) to target cancer cells. Logic-gated recognition integrated in aptamer-functionalized molecular machines will prompt fast tumor profiling, in situ capture and isolation, and safe delivery of precise medicine.","{'model': 'tldr@v2.0.0', 'text': 'A universal DNA tetrahedral scaffold that anchors on the cell membrane to load multiple aptamers and therapeutics for precise and effective theranostics and triggers the synergistic killing of effector aptamer-tethered synergistic drugs (EASDs) to target cancer cells.'}",
-undergraduate university education in internet of things engineering in china: a survey,B. Du,"To address the problem of a serious Internet of Things (IoT) talent shortage, the Chinese government is encouraging universities and institutions to establish a major in IoT Engineering to train high-level IoT talent. Since 2011, 700 Chinese universities and institutions have established IoT Engineering undergraduate programs. We surveyed the history, status, curriculum, and problems of IoT undergraduate degree programs in China. Based on data collected from faculties, students, and graduates, we found that the IoT Engineering curriculum was an unsystematic patchwork and was deficient in practical platforms. We then provide suggestions for further development and exploration of IoT education by presenting a Technical Knowledge Map of IoT Engineering. This survey work on Chinese IoT education can prompt Chinese educators to reflect on its achievements and problems, which will contribute to IoT talent cultivation worldwide.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the IoT Engineering curriculum in China was an unsystematic patchwork and was deficient in practical platforms, which will contribute to IoT talent cultivation worldwide.'}",https://www.mdpi.com/2227-7102/11/5/202/pdf?version=1619423706
-prompt gamma neutron activation analysis (pgaa): recent developments and applications,R. Lindstrom,,,
-a prompt-aware neural network approach to content-based scoring of non-native spontaneous speech,Yao Qian,"We present a neural network approach to the automated assessment of non-native spontaneous speech in a listen and speak task. An attention-based Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) is used to learn the relations (scoring rubrics) between the spoken responses and their assigned scores. Each prompt (listening material) is encoded as a vector in a low-dimensional space and then employed as a condition of the inputs of the attention LSTM-RNN. The experimental results show that our approach performs as well as the strong baseline of a Support Vector Regressor (SVR) using content-related features, i.e., a correlation of r = 0.806 with holistic proficiency scores provided by humans, without doing any feature engineering. The prompt-encoded vector improves the discrimination between the high-scoring sample and low-scoring sample, and it is more effective in grading responses to unseen prompts, which have no corresponding responses in the training set.","{'model': 'tldr@v2.0.0', 'text': 'A neural network approach to the automated assessment of non-native spontaneous speech in a listen and speak task that performs as well as the strong baseline of a Support Vector Regressor using content-related features, without doing any feature engineering.'}",
-performance of chatgpt on the us fundamentals of engineering exam: comprehensive assessment of proficiency and potential implications for professional environmental engineering practice,Vinay Pursnani,"In recent years, advancements in artificial intelligence (AI) have led to the development of large language models like GPT-4, demonstrating potential applications in various fields, including education. This study investigates the feasibility and effectiveness of using ChatGPT, a GPT-4 based model, in achieving satisfactory performance on the Fundamentals of Engineering (FE) Environmental Exam. This study further shows a significant improvement in the model's accuracy when answering FE exam questions through noninvasive prompt modifications, substantiating the utility of prompt modification as a viable approach to enhance AI performance in educational contexts. Furthermore, the findings reflect remarkable improvements in mathematical capabilities across successive iterations of ChatGPT models, showcasing their potential in solving complex engineering problems. Our paper also explores future research directions, emphasizing the importance of addressing AI challenges in education, enhancing accessibility and inclusion for diverse student populations, and developing AI-resistant exam questions to maintain examination integrity. By evaluating the performance of ChatGPT in the context of the FE Environmental Exam, this study contributes valuable insights into the potential applications and limitations of large language models in educational settings. As AI continues to evolve, these findings offer a foundation for further research into the responsible and effective integration of AI models across various disciplines, ultimately optimizing the learning experience and improving student outcomes.",,http://arxiv.org/pdf/2304.12198
-integrated structure and device engineering for high performance and scalable quantum dot infrared photodetectors.,Kaimin Xu,"Colloidal quantum dots (CQDs) are emerging as promising materials for the next generation infrared (IR) photodetectors, due to their easy solution processing, low cost manufacturing, size-tunable optoelectronic properties, and flexibility. Tremendous efforts including material engineering and device structure manipulation have been made to improve the performance of the photodetectors based on CQDs. In recent years, benefiting from the facial integration with materials such as 2D structure, perovskite and silicon, as well as device engineering, the performance of CQD IR photodetectors have been developing rapidly. On the other hand, to prompt the application of CQD IR photodetectors, scalable device structures that are compatible with commercial systems are developed. Herein, recent advances of CQD based IR photodetectors are summarized, especially material integration, device engineering, and scalable device structures.","{'model': 'tldr@v2.0.0', 'text': 'Recent advances of CQD based IR photodetectors are summarized, especially material integration, device engineering, and scalable device structures.'}",
-are fast radio bursts the most likely electromagnetic counterpart of neutron star mergers resulting in prompt collapse?,V. Paschalidis,"National Science Foundation (NSF) at the University of Arizona [PHY-1912619]; NSF [PHY-1602536, PHY-1662211]; NASA at the University of Illinois at Urbana-Champaign [80NSSC17K0070]; Extreme Science and Engineering Discovery Environment (XSEDE) [TG-PHY180036, TG-PHY190020]",,https://link.aps.org/accepted/10.1103/PhysRevD.100.043001
-tunable phosphorescence-fluorescence dual emissions of organic isoquinoline-benzophenone doped system by alkoxy engineering.,Yating Chen,"The dual/multi-component organic doped system with room temperature phosphorescence (RTP) properties has been developed. However, the unknown luminescence mechanism still greatly limits the development of the doped materials. Herein, a new doped system exhibiting phosphorescence-fluorescence dual emissions (ФPhos.= 4-24% and τphos. = 101-343 ms) is successfully constructed through prediction and design. A series of isoquinoline derivatives with different alkoxy chains were selected as the guests. Benzophenone was chosen as the host due to the characteristic of low melting point and good crystallinity. The alkoxy chain lengths of the guests are first reported to be used to control the fluorescence and phosphorescence intensities of the doped materials, which results in different prompt emission colors. Additionally, the doped ratio of the guest and host can also control the luminous intensities of the materials. In particular, the doped materials still exhibit phosphorescent properties even if the ratio of guest-host is as low as 1:100000.","{'model': 'tldr@v2.0.0', 'text': 'A new doped system exhibiting phosphorescence-fluorescence dual emissions is successfully constructed through prediction and design and can still exhibit phosphorescent properties even if the ratio of guest-host is as low as 1:100000.'}",
-a surface‐tailoring method for rapid non‐thermosensitive cell‐sheet engineering via functional polymer coatings,Jieung Baek,"Cell sheet engineering, a technique utilizing a monolayer cell sheet, has recently emerged as a promising technology for scaffold‐free tissue engineering. In contrast to conventional tissue‐engineering approaches, the cell sheet technology allows cell harvest as a continuous cell sheet with intact extracellular matrix proteins and cell–cell junction, which facilitates cell transplantation without any other artificial biomaterials. A facile, non‐thermoresponsive method is demonstrated for a rapid but highly reliable platform for cell‐sheet engineering. The developed method exploits the precise modulation of cell–substrate interactions by controlling the surface energy of the substrate via a series of functional polymer coatings to enable prompt cell sheet harvesting within 100 s. The engineered surface can trigger an intrinsic cellular response upon the depletion of divalent cations, leading to spontaneous cell sheet detachment under physiological conditions (pH 7.4 and 37 °C) in a non‐thermoresponsive manner. Additionally, the therapeutic potential of the cell sheet is successfully demonstrated by the transplantation of multilayered cell sheets into mouse models of diabetic wounds and ischemia. These findings highlight the ability of the developed surface for non‐thermoresponsive cell sheet engineering to serve as a robust platform for regenerative medicine and provide significant breakthroughs in cell sheet technology.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the developed surface for non‐thermoresponsive cell sheet engineering to serve as a robust platform for regenerative medicine and provide significant breakthroughs in cell sheet technology are highlighted.'}",
-vascular tissue engineering: advanced techniques and gene editing in stem cells for graft generation.,Sin-Guang Chen,"The common occurrence of cardiovascular diseases and the lack of proper autologous tissues prompt and promote the pressing development of tissue-engineered vascular grafts. Current progress on scaffold production, genetically modified cells and use of nanotechnology-based monitoring has considerably improved the long-term patency of engineered tissue grafts. However, challenges abound in the autologous materials and manipulation of genes and cells for tissue engineering. This review overviews current development in tissue-engineered vascular grafts and discusses recent improvements in scaffolding techniques as well as the efficiency of gene-editing tools and their ability to fill the existing gaps in stem-cell and regenerative therapies. Current advances in 3D-printing approaches for fabrication of engineered tissues are also reviewed together with specific biomaterials for vascular tissues. In addition, the natural and synthetic polymers that hold increasing significance for vascular tissue engineering are highlighted. Both animal models and nanotechnology-based monitoring are proposed for pre-clinical evaluation of engineered grafts in view of their historical significance in tissue engineering. The ultimate success of tissue regeneration, which is yet to be fully realized, depends on the optimal performance of culture systems, biomaterial constructs and stem cells in a suitable artificial physiological environment.","{'model': 'tldr@v2.0.0', 'text': 'This review overviews current development in tissue-engineered vascular grafts and discusses recent improvements in scaffolding techniques as well as the efficiency of gene-editing tools and their ability to fill the existing gaps in stem-cell and regenerative therapies.'}",http://ira.lib.polyu.edu.hk/bitstream/10397/93663/1/Chen_Vascular_Tissue_Engineering.pdf
-co₂-induced phase engineering: protocol for enhanced photoelectrocatalytic performance of 2d mos₂ nanosheets.,Yuhang Qi,"Molybdenum disulfide (MoS2) is a promising non-precious-metal catalyst, but its performance is limited by the density of active sites and poor electrical transport. Its metallic 1T phase possesses higher photoelectrocatalytic activity. Thus, how to efficiently increase the concentration of the 1T phase in the exfoliated two-dimensiaonal (2D) MoS2 nanosheets is an important premise. In this work, we propose a strategy to prepare a 2D heterostructure of MoS2 nanosheets using supercritical CO2-induced phase engineering to form metallic 1T-MoS2. Theoretical calculations and experimental results demonstrate that the introduced CO2 in the 2H-MoS2 host can prompt the transformation of partial 2H-MoS2 lattices into 1T-MoS2. Moreover, the electrical coupling and synergistic effect between 2H and 1T phases can greatly facilitate the efficient electron transfer from the active sites of MoS2, which significantly improves the photocatalytic performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a strategy to prepare a 2D heterostructure of MoS2 nanosheets using supercritical CO2-induced phase engineering to form metallic 1T-MoS2, which significantly improves the photocatalytic performance.'}",
-engineering chiral polyoxometalate hybrid metal-organic frameworks for asymmetric dihydroxylation of olefins.,Q. Han,"Chiral metal-organic frameworks (MOFs) with porous and tunable natures have made them feasible for performing a variety of chemical reactions as heterogeneous asymmetric catalysts. By incorporating the oxidation catalyst [BW12O40](5-) and the chiral group, L- or D-pyrrolidin-2-ylimidazole (PYI), into one single framework, the two enantiomorphs Ni-PYI1 and Ni-PYI2 were obtained via self-assembly, respectively. The channels of Ni-PYIs were enlarged through a guest exchange reaction to remove the cationic chiral templates and were well modulated with hydrophilic/hydrophobic properties to allow molecules of both H2O2 and olefin ingress and egress. The coexistence of both the chiral directors and the oxidants within a confined space provided a special environment for the formation of reaction intermediates in a stereoselective fashion with high selectivity. The resulting MOF acted as an amphipathic catalyst to prompt the asymmetric dihydroxylation of aryl olefins with excellent stereoselectivity.","{'model': 'tldr@v2.0.0', 'text': 'The coexistence of both the chiral directors and the oxidants within a confined space provided a special environment for the formation of reaction intermediates in a stereoselective fashion with high selectivity.'}",
-how do different reflection prompts affect engineering students’ academic performance and engagement?,Muhsin Menekse,"Abstract This study explored the relative effectiveness of generic versus specific reflection prompts on engineering students’ academic performance and engagement with four scales (i.e., behavioral, emotional, social, cognitive). The sample consisted of 208 engineering students in two sections of a required first-year engineering course. By using a quasi-experimental research design in classroom settings, students in the specific and generic reflection prompt conditions were compared based on exams, project scores, and problem sets. Additionally, the role of reflection prompts in students’ behavioral, emotional, social, and cognitive engagements was investigated across conditions by using pre- and post-surveys administered at the beginning and end of the semester. Students in both conditions were asked to submit a total of 52 reflections during an academic semester. Results showed that students in the specific prompts condition performed significantly better on exams, total project score, and two problem sets than students in the generic prompts condition. In terms of engagement, results showed there was a significant decrease in behavioral, emotional, and cognitive scales for both conditions. On the other hand, for the social scale, while there was a decrease for students in generic prompts condition, there was an increase for the ones in specific prompts condition.",,
-engineering mindsets and learning outcomes in elementary school,Pamela S. Lottero‐Perdue,"Students may exhibit growth mindsets, where intelligence is seen as malleable and failures prompt more effort and new approaches, or fixed mindsets, where intelligence is seen as immutable and failures indicate lack of intelligence. One's mindset in general may be different from that for a particular domain such as engineering. Having a growth mindset predicts more positive learning outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Students may exhibit growth mindsets, where intelligence is seen as malleable and failures prompt more effort and new approaches, or fixed mindsets that are seen as immutable and failures indicate lack of intelligence.'}",
-"rapid detection of genetic engineering, structural variation, and antimicrobial resistance markers in bacterial biothreat pathogens by nanopore sequencing",Amy S. Gargis,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/s41598-019-49700-1.pdf
-engineering a segmented dual-reservoir polyurethane intravaginal ring for simultaneous prevention of hiv transmission and unwanted pregnancy,Justin T. Clark,"The HIV/AIDS pandemic and its impact on women prompt the investigation of prevention strategies to interrupt sexual transmission of HIV. Long-acting drug delivery systems that simultaneously protect womenfrom sexual transmission of HIV and unwanted pregnancy could be important tools in combating the pandemic. We describe the design, in silico, in vitro and in vivo evaluation of a dual-reservoir intravaginal ring that delivers the HIV-1 reverse transcriptase inhibitor tenofovir and the contraceptive levonorgestrel for 90 days. Two polyether urethanes with two different hard segment volume fractions were used to make coaxial extruded reservoir segments with a 100 µm thick rate controlling membrane and a diameter of 5.5 mm that contain 1.3 wt% levonorgestrel. A new mechanistic diffusion model accurately described the levonorgestrel burst release in early time points and pseudo-steady state behavior at later time points. As previously described, tenofovir was formulated as a glycerol paste and filled into a hydrophilic polyurethane, hollow tube reservoir that was melt-sealed by induction welding. These tenofovir-eluting segments and 2 cm long coaxially extruded levonorgestrel eluting segments were joined by induction welding to form rings that released an average of 7.5 mg tenofovir and 21 µg levonorgestrel per day in vitro for 90 days. Levonorgestrel segments placed intravaginally in rabbits resulted in sustained, dose-dependent levels of levonorgestrel in plasma and cervical tissue for 90 days. Polyurethane caps placed between segments successfully prevented diffusion of levonorgestrel into the tenofovir-releasing segment during storage.Hydrated rings endured between 152 N and 354 N tensile load before failure during uniaxial extension testing. In summary, this system represents a significant advance in vaginal drug delivery technology, and is the first in a new class of long-acting multipurpose prevention drug delivery systems.","{'model': 'tldr@v2.0.0', 'text': 'This system is described, in silico, in vitro and in vivo evaluation of a dual-reservoir intravaginal ring that delivers the HIV-1 reverse transcriptase inhibitor tenofovir and the contraceptive levonorgestrel for 90 days and is the first in a new class of long-acting multipurpose prevention drug delivery systems.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0088509&type=printable
-exploring the effects of the design prompt on students’ design cognition,Christopher B. Williams,"Engineering design educators often provide their students a task (or “prompt”) to guide their design projects. Similarly, engineering design educational researchers also provide research participants with a design task to guide their activity during experimental sessions. In both contexts, there is a fundamental underlying assumption that the design task has no significant effect on the students’/participants’ design cognition. Specifically, the authors test the hypothesis that a design task does affect a student’s design experience. Failing to disprove this hypothesis could significantly impact both design education practice and design education experimental research.To determine the effect of a design task on students’ design cognition, experimental sessions were conducted wherein student design teams worked together to solve a speculative design task. The student teams were presented with two nearly identical design tasks; however, one featured an additional design requirement. A task-independent protocol analysis method grounded in the Function-Behavior-Structure design ontology is performed on audio and video recordings of the design sessions to provide a common basis for comparing the two groups. Differences in design cognition are identified by analyzing and comparing the percent occurrences of the design issues and design processes and the Problem-Solution indices.Copyright © 2013 by ASME","{'model': 'tldr@v2.0.0', 'text': 'The authors test the hypothesis that a design task does affect a student’s design experience and try to disprove this hypothesis, which could significantly impact both design education practice and design education experimental research.'}",
-organoid technologies meet genome engineering,J. Nie,"Three‐dimensional (3D) stem cell differentiation cultures recently emerged as a novel model system for investigating human embryonic development and disease progression in vitro, complementing existing animal and two‐dimensional (2D) cell culture models. Organoids, the 3D self‐organizing structures derived from pluripotent or somatic stem cells, can recapitulate many aspects of structural organization and functionality of their in vivo organ counterparts, thus holding great promise for biomedical research and translational applications. Importantly, faithful recapitulation of disease and development processes relies on the ability to modify the genomic contents in organoid cells. The revolutionary genome engineering technologies, CRISPR/Cas9 in particular, enable investigators to generate various reporter cell lines for prompt validation of specific cell lineages as well as to introduce disease‐associated mutations for disease modeling. In this review, we provide historical overviews, and discuss technical considerations, and potential future applications of genome engineering in 3D organoid models.","{'model': 'tldr@v2.0.0', 'text': 'Three‐dimensional stem cell differentiation cultures recently emerged as a novel model system for investigating human embryonic development and disease progression in vitro, complementing existing animal and two‐dimensional cell culture models.'}",https://europepmc.org/articles/pmc5331239?pdf=render
-collagen hydrogel as an immunomodulatory scaffold in cartilage tissue engineering.,T. Yuan,"A collagen type I hydrogel was constructed and used as the scaffold for cartilage tissue engineering. Neonatal rabbit chondrocytes were seeded into the hydrogel, and the constructs were cultured in vitro for 7, 14, and 28 days. The immunomodulatory effect of the hydrogel on seeded chondrocytes was carefully investigated. The expressions of major histocompatibility complex classes I and II of seeded chondrocytes increased with the time, which indicated that the immunogenicity also increased with the time. Meanwhile, the properly designed collagen type I hydrogel could prompt the chondrogenesis of engineered cartilage. The extracellular matrix (ECM) synthesis ability of seeded chondrocytes and the accumulated ECM in the constructs continuously increased with the culture time. Both the isolation and protection, which come from formed ECM and hydrogel scaffold, can effectively control the adverse immunogenicity of seeded chondrocytes and even help to lessen the immunogenicity of the whole engineered cartilage. As the result, the levels of mixed lymphocyte chondrocyte reactions of seed cells and the constructs decreased gradually. The stimulation on allogeneic lymphocytes of the whole constructs was obviously lower than that of the retrieved cells from the constructs. Therefore, properly designed collagen type I hydrogel can give certain immunogenicity-reducing effects on engineered cartilage based on chondrocytes, and it may be a potential immunomodulatory biomaterial in tissue engineering.","{'model': 'tldr@v2.0.0', 'text': None}",
-event-shape engineering for the d-meson elliptic flow in mid-central pb-pb collisions at snn = 5 . 02 tev,S. Acharya,"The production yield of prompt D mesons and their elliptic flow coefficient v2 were measured with the Event-Shape Engineering (ESE) technique applied to mid-central (10–30% and 30–50% centrality classes) Pb-Pb collisions at the centre-of-mass energy per nucleon pair √ sNN = 5.02 TeV, with the ALICE detector at the LHC. The ESE technique allows the classification of events, belonging to the same centrality, according to the azimuthal anisotropy of soft particle production in the collision. The reported measurements give the opportunity to investigate the dynamics of charm quarks in the Quark-Gluon Plasma and provide information on their participation in the collective expansion of the medium. D mesons were reconstructed via their hadronic decays at mid-rapidity, |η| < 0.8, in the transverse momentum interval 1 < pT < 24 GeV/c. The v2 coefficient is found to be sensitive to the event-shape selection confirming a correlation between the D-meson azimuthal anisotropy and the collective expansion of the bulk matter, while the per-event D-meson yields do not show any significant modification within the current uncertainties.",,
-application and optimization of crispr–cas9-mediated genome engineering in axolotl (ambystoma mexicanum),J. Fei,,"{'model': 'tldr@v2.0.0', 'text': 'An optimized stepwise protocol to create genetically modified axolotls using the CRISPR–Cas9 system is described, which bypasses the long generation time of axolOTls and allows direct functional analysis in F0 genetically manipulated axolotsls.'}",
-symbolic knowledge distillation: from general language models to commonsense models,Peter West,"The common practice for training commonsense models has gone from–human–to–corpus–to–machine: humans author commonsense knowledge graphs in order to train commonsense models. In this work, we investigate an alternative, from–machine–to–corpus–to–machine: general language models author these commonsense knowledge graphs to train commonsense models. Our study leads to a new framework, Symbolic Knowledge Distillation. As with prior art in Knowledge Distillation (Hinton et al. 2015), our approach uses larger models to teach smaller models. A key difference is that we distill knowledge symbolically–as text–in addition to the neural model. We distill only one aspect–the commonsense of a general language model teacher, allowing the student to be a different type, a commonsense model. Altogether, we show that careful prompt engineering and a separately trained critic model allow us to selectively distill high-quality causal commonsense from GPT-3, a general language model. Empirical results demonstrate that, for the first time, a human-authored commonsense knowledge graph is surpassed by our automatically distilled variant in all three criteria: quantity, quality, and diversity. In addition, it results in a neural commonsense model that surpasses the teacher model’s commonsense capabilities despite its 100x smaller size. We apply this to the ATOMIC resource, and will share our new symbolic knowledge graph and commonsense models.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that careful prompt engineering and a separately trained critic model allow us to selectively distill high-quality causal commonsense from GPT-3, a general language model, and results in a neural commonsense model that surpasses the teacher model's commonsense capabilities despite its 100x smaller size.""}",https://aclanthology.org/2022.naacl-main.341.pdf
-"chat2vis: generating data visualizations via natural language using chatgpt, codex and gpt-3 large language models",Paula Maddigan,"The field of data visualisation has long aimed to devise solutions for generating visualisations directly from natural language text. Research in Natural Language Interfaces (NLIs) has contributed towards the development of such techniques. However, the implementation of workable NLIs has always been challenging due to the inherent ambiguity of natural language, as well as in consequence of unclear and poorly written user queries which pose problems for existing language models in discerning user intent. Instead of pursuing the usual path of developing new iterations of language models, this study uniquely proposes leveraging the advancements in pre-trained large language models (LLMs) such as ChatGPT and GPT-3 to convert free-form natural language directly into code for appropriate visualisations. This paper presents a novel system, Chat2VIS, which takes advantage of the capabilities of LLMs and demonstrates how, with effective prompt engineering, the complex problem of language understanding can be solved more efficiently, resulting in simpler and more accurate end-to-end solutions than prior approaches. Chat2VIS shows that LLMs together with the proposed prompts offer a reliable approach to rendering visualisations from natural language queries, even when queries are highly misspecified and underspecified. This solution also presents a significant reduction in costs for the development of NLI systems, while attaining greater visualisation inference abilities compared to traditional NLP approaches that use hand-crafted grammar rules and tailored models. This study also presents how LLM prompts can be constructed in a way that preserves data security and privacy while being generalisable to different datasets. This work compares the performance of GPT-3, Codex and ChatGPT across several case studies and contrasts the performances with prior studies.","{'model': 'tldr@v2.0.0', 'text': 'Chat2VIS shows that LLMs together with the proposed prompts offer a reliable approach to rendering visualisations from natural language queries, even when queries are highly misspecified and underspecified.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10121440.pdf
-"chatgpt evaluation on sentence level relations: a focus on temporal, causal, and discourse relations",Chunkit Chan,"This paper aims to quantitatively evaluate the performance of ChatGPT, an interactive large language model, on inter-sentential relations such as temporal relations, causal relations, and discourse relations. Given ChatGPT's promising performance across various tasks, we conduct extensive evaluations on the whole test sets of 13 datasets, including temporal and causal relations, PDTB2.0-based and dialogue-based discourse relations, and downstream applications on discourse understanding. To achieve reliable results, we adopt three tailored prompt templates for each task, including the zero-shot prompt template, zero-shot prompt engineering (PE) template, and in-context learning (ICL) prompt template, to establish the initial baseline scores for all popular sentence-pair relation classification tasks for the first time. We find that ChatGPT exhibits strong performance in detecting and reasoning about causal relations, while it may not be proficient in identifying the temporal order between two events. It can recognize most discourse relations with existing explicit discourse connectives, but the implicit discourse relation still remains a challenging task. Meanwhile, ChatGPT performs poorly in the dialogue discourse parsing task that requires structural understanding in a dialogue before being aware of the discourse relation.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT exhibits strong performance in detecting and reasoning about causal relations, while it may not be proficient in identifying the temporal order between two events.'}",http://arxiv.org/pdf/2304.14827
-integrating engineering outputs from natural disaster models into a dynamic spatial computable general equilibrium model of centerville,H. Cutler,"Abstract A dynamic spatial computable general equilibrium (DSCGE) model is constructed that describes how engineering and economic models can be integrated to assess the economic, demographic, and fiscal impacts of disasters. This paper has two objectives. First, we introduce the DSCGE model and describe how it is calibrated specifically for Centerville. Second, to demonstrate the analytic flexibility of the DSCGE platform, we present economy-wide prompt effects from simulations involving spatially circumscribed shocks to Centerville’s building portfolio and transportation infrastructure, and then detail dynamic economy-wide effects from simulations involving combinations of infrastructure damage and adjustments to the economic behavior of agents. We conclude with a discussion of the technical challenges ahead.",,
-cutting down on prompts and parameters: simple few-shot learning with language models,Robert L Logan IV,"Prompting language models (LMs) with training examples and task descriptions has been seen as critical to recent successes in few-shot learning. In this work, we show that finetuning LMs in the few-shot setting can considerably reduce the need for prompt engineering. In fact, one can use null prompts, prompts that contain neither task-specific templates nor training examples, and achieve competitive accuracy to manually-tuned prompts across a wide range of tasks. While finetuning LMs does introduce new parameters for each downstream task, we show that this memory overhead can be substantially reduced: finetuning only the bias terms can achieve comparable or better accuracy than standard finetuning while only updating 0.1% of the parameters. All in all, we recommend finetuning LMs for few-shot learning as it is more accurate, robust to different prompts, and can be made nearly as efficient as using frozen LMs.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that finetuning LMs in the few-shot setting can considerably reduce the need for prompt engineering, and recommends finetuned LMs for few- shot learning as it is more accurate, robust to different prompts, and can be made nearly as efficient as using frozen LMs.'}",https://aclanthology.org/2022.findings-acl.222.pdf
-overcoming physical constraints in bone engineering: ‘the importance of being vascularized’,T. Genova,"Bone plays several physiological functions and is the second most commonly transplanted tissue after blood. Since the treatment of large bone defects is still unsatisfactory, researchers have endeavoured to obtain scaffolds able to release growth and differentiation factors for mesenchymal stem cells, osteoblasts and endothelial cells in order to obtain faster mineralization and prompt a reliable vascularization. Nowadays, the application of osteoblastic cultures spans from cell physiology and pharmacology to cytocompatibility measurement and osteogenic potential evaluation of novel biomaterials. To overcome the simple traditional monocultures in vitro, co-cultures of osteogenic and vasculogenic precursors were introduced with very interesting results. Increasingly complex culture systems have been developed, where cells are seeded on proper scaffolds and stimulated so as to mimic the physiological conditions more accurately. These bioreactors aim at enabling bone regeneration by incorporating different cells types into bio-inspired materials within a surveilled habitat. This review is focused on the most recent developments in the organomimetic cultures of osteoblasts and vascular endothelial cells for bone tissue engineering.","{'model': 'tldr@v2.0.0', 'text': 'This review is focused on the most recent developments in the organomimetic cultures of osteoblasts and vascular endothelial cells for bone tissue engineering.'}",https://iris.unito.it/bitstream/2318/1541558/4/Genova_et_al_2016.pdf
-fake it till you make it: learning transferable representations from synthetic imagenet clones,Mert Bulent Sariyildiz,"Recent image generation models such as Stable Diffusion have exhibited an impressive ability to generate fairly realistic images starting from a simple text prompt. Could such models render real images obsolete for training image prediction models? In this paper, we answer part of this provocative question by investigating the need for real images when training models for ImageNet classification. Provided only with the class names that have been used to build the dataset, we explore the ability of Stable Diffusion to generate synthetic clones of ImageNet and measure how useful these are for training classification models from scratch. We show that with minimal and class-agnostic prompt engineering, ImageNet clones are able to close a large part of the gap between models produced by synthetic images and models trained with real images, for the several standard classification benchmarks that we consider in this study. More importantly, we show that models trained on synthetic images exhibit strong generalization properties and perform on par with models trained on real data for transfer. Project page: https://europe.naverlabs.com/imagenet-sd","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with minimal and class-agnostic prompt engineering, ImageNet clones are able to close a large part of the gap between models produced by synthetic images and models trained with real images, for the several standard classification benchmarks that are considered in this study.'}",https://arxiv.org/pdf/2212.08420
-text-guided synthesis of artistic images with retrieval-augmented diffusion models,Robin Rombach,"Novel architectures have recently improved generative image synthesis leading to excellent visual quality in various tasks. Of particular note is the field of ``AI-Art'', which has seen unprecedented growth with the emergence of powerful multimodal models such as CLIP. By combining speech and image synthesis models, so-called ``prompt-engineering'' has become established, in which carefully selected and composed sentences are used to achieve a certain visual style in the synthesized image. In this note, we present an alternative approach based on retrieval-augmented diffusion models (RDMs). In RDMs, a set of nearest neighbors is retrieved from an external database during training for each training instance, and the diffusion model is conditioned on these informative samples. During inference (sampling), we replace the retrieval database with a more specialized database that contains, for example, only images of a particular visual style. This provides a novel way to prompt a general trained model after training and thereby specify a particular visual style. As shown by our experiments, this approach is superior to specifying the visual style within the text prompt. We open-source code and model weights at https://github.com/CompVis/latent-diffusion .","{'model': 'tldr@v2.0.0', 'text': 'This note presents an alternative approach based on retrieval-augmented diffusion models (RDMs) that provides a novel way to prompt a general trained model after training and thereby specify a particular visual style.'}",http://arxiv.org/pdf/2207.13038
-bigbio: a framework for data-centric biomedical natural language processing,Jason Alan Fries,"Training and evaluating language models increasingly requires the construction of meta-datasets --diverse collections of curated data with clear provenance. Natural language prompting has recently lead to improved zero-shot generalization by transforming existing, supervised datasets into a diversity of novel pretraining tasks, highlighting the benefits of meta-dataset curation. While successful in general-domain text, translating these data-centric approaches to biomedical language modeling remains challenging, as labeled biomedical datasets are significantly underrepresented in popular data hubs. To address this challenge, we introduce BigBIO a community library of 126+ biomedical NLP datasets, currently covering 12 task categories and 10+ languages. BigBIO facilitates reproducible meta-dataset curation via programmatic access to datasets and their metadata, and is compatible with current platforms for prompt engineering and end-to-end few/zero shot language model evaluation. We discuss our process for task schema harmonization, data auditing, contribution guidelines, and outline two illustrative use cases: zero-shot evaluation of biomedical prompts and large-scale, multi-task learning. BigBIO is an ongoing community effort and is available at https://github.com/bigscience-workshop/biomedical","{'model': 'tldr@v2.0.0', 'text': 'BigBIO is introduced, a community library of 126+ biomedical NLP datasets, currently covering 12 task categories and 10+ languages that facilitates reproducible meta-dataset curation via programmatic access to datasets and their metadata, and is compatible with current platforms for prompt engineering and end-to-end few/zero shot language model evaluation.'}",http://arxiv.org/pdf/2206.15076
-repair is nearly generation: multilingual program repair with llms,Harshit Joshi,"Most programmers make mistakes when writing code. Some of these mistakes are small and require few edits to the original program – a class of errors recently termed last mile mistakes. These errors break the flow for experienced developers and can stump novice programmers. Existing automated repair techniques targeting this class of errors are language-specific and do not easily carry over to new languages. Transferring symbolic approaches requires substantial engineering and neural approaches require data and retraining. We introduce RING, a multilingual repair engine powered by a large language model trained on code (LLMC) such as Codex. Such a multilingual engine enables a flipped model for programming assistance, one where the programmer writes code and the AI assistance suggests fixes, compared to traditional code suggestion technology. Taking inspiration from the way programmers manually fix bugs, we show that a prompt-based strategy that conceptualizes repair as localization, transformation, and candidate ranking, can successfully repair programs in multiple languages with minimal effort. We present the first results for such a multilingual repair engine by evaluating on 6 different languages and comparing performance to language-specific repair engines. We show that RING can outperform language-specific repair engines for three of these languages.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces RING, a multilingual repair engine powered by a large language model trained on code (LLMC) such as Codex that enables a flipped model for programming assistance, one where the programmer writes code and the AI assistance suggests fixes compared to traditional code suggestion technology.'}",https://arxiv.org/pdf/2208.11640
-injectable bottlebrush hydrogels with tissue-mimetic mechanical properties,F. Vashahi,"Injectable hydrogels are desired in many biomedical applications due to their minimally invasive deployment to the body and their ability to introduce drugs. However, current injectables suffer from mechanical mismatch with tissue, fragility, water expulsion, and high viscosity. To address these issues, we design brush-like macromolecules that concurrently provide softness, firmness, strength, fluidity, and swellability. The synthesized linear-bottlebrush-linear (LBL) copolymers facilitate improved injectability as the compact conformation of bottlebrush blocks results in low solution viscosity, while the thermoresponsive linear blocks permit prompt gelation at 37°C. The resulting hydrogels mimic the deformation response of supersoft tissues such as adipose and brain while withstanding deformations of 700% and precluding water expulsion upon gelation. Given their low cytotoxicity and mild inflammation in vivo, the developed materials will have vital implications for reconstructive surgery, tissue engineering, and drug delivery applications.","{'model': 'tldr@v2.0.0', 'text': 'The synthesized linear-bottlebrush-linear (LBL) copolymers facilitate improved injectability as the compact conformation of bottlebrush blocks results in low solution viscosity, while the thermoresponsive linear blocks permit prompt gelation at 37°C.'}",
-prompting is all your need: automated android bug replay with large language models,Sidong Feng,"Bug reports are vital for software maintenance that allow users to inform developers of the problems encountered while using the software. As such, researchers have committed considerable resources toward automating bug replay to expedite the process of software maintenance. Nonetheless, the success of current automated approaches is largely dictated by the characteristics and quality of bug reports, as they are constrained by the limitations of manually-crafted patterns and pre-defined vocabulary lists. Inspired by the success of Large Language Models (LLMs) in natural language understanding, we propose AdbGPT , a new lightweight approach to automatically reproduce the bugs from bug reports through prompt engineering, without any training and hard-coding effort. AdbGPT leverages few-shot learning and chain-of-thought reasoning to elicit human knowledge and logical reasoning from LLMs to accomplish the bug replay in a manner similar to a developer. Our evaluations demonstrate the effectiveness and efficiency of our AdbGPT to reproduce 81.3% of bug reports in 253.6 seconds, outperforming the state-of-the-art baselines and ablation studies. We also conduct a small-scale user study to confirm the usefulness of AdbGPT in enhancing developers’ bug replay capabilities.","{'model': 'tldr@v2.0.0', 'text': 'AdbGPT is proposed, a new lightweight approach to automatically reproduce the bugs from bug reports through prompt engineering, without any training and hard-coding effort, that leverages few-shot learning and chain-of-thought reasoning to elicit human knowledge and logical reasoning from LLMs to accomplish the bug replay in a manner similar to a developer.'}",https://arxiv.org/pdf/2306.01987
-qaner: prompting question answering models for few-shot named entity recognition,Andy T. Liu,"Recently, prompt-based learning for pre-trained language models has succeeded in few-shot Named Entity Recognition (NER) by exploiting prompts as task guidance to increase label efficiency. However, previous prompt-based methods for few-shot NER have limitations such as a higher computational complexity, poor zero-shot ability, requiring manual prompt engineering, or lack of prompt robustness. In this work, we address these shortcomings by proposing a new prompt-based learning NER method with Question Answering (QA), called QaNER. Our approach includes 1) a refined strategy for converting NER problems into the QA formulation; 2) NER prompt generation for QA models; 3) prompt-based tuning with QA models on a few annotated NER examples; 4) zero-shot NER by prompting the QA model. Comparing the proposed approach with previous methods, QaNER is faster at inference, insensitive to the prompt quality, and robust to hyper-parameters, as well as demonstrating significantly better low-resource performance and zero-shot capability.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new prompt-based learning NER method with Question Answering (QA), called QaNER, which is faster at inference, insensitive to the prompt quality, and robust to hyper-parameters, as well as demonstrating significantly better low-resource performance and zero-shot capability.'}",http://arxiv.org/pdf/2203.01543
-prompting the hidden talent of web-scale speech models for zero-shot task generalization,Puyuan Peng,"We investigate the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering. We selected three tasks: audio-visual speech recognition (AVSR), code-switched speech recognition (CS-ASR), and speech translation (ST) on unseen language pairs. We design task-specific prompts, by either leveraging another large-scale model, or simply manipulating the special tokens in the default prompts. Experiments show that compared to the default prompts, our proposed prompts improve performance by 10% to 45% on the three zero-shot tasks, and even outperform SotA supervised models on some datasets. In addition, our experiments reveal many interesting properties of Whisper, including its robustness to prompts, bias on accents, and the multilingual understanding in its latent space. Code is available at https://github.com/jasonppy/PromptingWhisper","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering, and designs task-specific prompts that improve performance on the three zero-shot tasks and even outperform SotA supervised models on some datasets.'}",https://arxiv.org/pdf/2305.11095
-surface engineering for mechanical enhancement of cell sheet by nano-coatings,Miso Yang,,"{'model': 'tldr@v2.0.0', 'text': None}",https://www.nature.com/articles/s41598-017-04746-x.pdf
-enabling continuous software engineering for embedded systems architectures with virtual prototypes,P. Antonino,,"{'model': 'tldr@v2.0.0', 'text': 'This paper discusses in this paper how to orchestrate architecture drivers and design specification techniques with virtual prototypes, to address the demands of continuous software engineering in development processes.'}",
-advancements in research and development to combat covid-19 using nanotechnology,Vishal Chaudhary,,"{'model': 'tldr@v2.0.0', 'text': 'Various claims, studies and reports on research and development to combat these challenges associated with COVID-19 have been collectively discussed in this article from the perspectives of nanotechnology.'}",https://link.springer.com/content/pdf/10.1007/s41204-021-00102-7.pdf
-the creativity of text-based generative art,J. Oppenlaender,"Text-based generation of digital images has made a giant leap to-wards becoming a mainstream phenomenon. With text-based generative systems, anybody can create digital images and artworks. This provokes the question of whether text-based generative art is creative. This paper expounds on the nature of human creativity involved in text-based generative art with a specific focus on the practice of prompt engineering, drawing on Rhodes’s conceptual model of creativity. The paper critiques the current product-centered view of creativity which may fall short in the context of text-based generative art. An case exemplifying this shortcoming is provided and future opportunities for research on text-based generative art are outlined.",,http://arxiv.org/pdf/2206.02904
-no token left behind: explainability-aided image classification and generation,Roni Paiss,"The application of zero-shot learning in computer vision has been revolutionized by the use of image-text matching models. The most notable example, CLIP, has been widely used for both zero-shot classification and guiding generative models with a text prompt. However, the zero-shot use of CLIP is unstable with respect to the phrasing of the input text, making it necessary to carefully engineer the prompts used. We find that this instability stems from a selective similarity score, which is based only on a subset of the semantically meaningful input tokens. To mitigate it, we present a novel explainability-based approach, which adds a loss term to ensure that CLIP focuses on all relevant semantic parts of the input, in addition to employing the CLIP similarity loss used in previous works. When applied to one-shot classification through prompt engineering, our method yields an improvement in the recognition rate, without additional training or fine-tuning. Additionally, we show that CLIP guidance of generative models using our method significantly improves the generated images. Finally, we demonstrate a novel use of CLIP guidance for text-based image generation with spatial conditioning on object location, by requiring the image explainability heatmap for each object to be confined to a pre-determined bounding box.","{'model': 'tldr@v2.0.0', 'text': 'A novel explainability-based approach is presented, which adds a loss term to ensure that CLIP focuses on all relevant semantic parts of the input, in addition to employing the CLIP similarity loss used in previous works.'}",http://arxiv.org/pdf/2204.04908
-automatically generating cs learning materials with large language models,S. Macneil,"Recent breakthroughs in Large Language Models (LLMs), such as GPT-3 and Codex, now enable software developers to generate code based on a natural language prompt. Within computer science education, researchers are exploring the potential for LLMs to generate code explanations and programming assignments using carefully crafted prompts. These advances may enable students to interact with code in new ways while helping instructors scale their learning materials. However, LLMs also introduce new implications for academic integrity, curriculum design, and software engineering careers. This workshop will demonstrate the capabilities of LLMs to help attendees evaluate whether and how LLMs might be integrated into their pedagogy and research. We will also engage attendees in brainstorming to consider how LLMs will impact our field.","{'model': 'tldr@v2.0.0', 'text': 'This workshop will demonstrate the capabilities of LLMs to help attendees evaluate whether and how LLMs might be integrated into their pedagogy and research, and engage attendees in brainstorming to consider how LLM will impact the field.'}",https://arxiv.org/pdf/2212.05113
-language-aware soft prompting for vision & language foundation models,Adrian Bulat,"This paper is on soft prompt learning for Vision & Language (V&L) models. Similarly to their NLP counterparts, V&L models can be adapted to a downstream task by learning soft continuous prompts using a few training examples. Current methods learn the soft prompts by minimizing a cross-entropy loss using as class weights the features obtained by passing the prompts plus the class names through the text encoder. Such methods, however, significantly overfit the training data suffering from large accuracy degradation when tested on unseen classes from the same domain. Our main contribution, in this paper, is a surprisingly simple approach to alleviate this problem: we use a second cross entropy loss to minimize the distance between the learned soft prompts and a set of hand-engineered manual prompts (obtained by prompt engineering). The proposed loss can be interpreted in multiple ways including as a regularizer, as a means for language-based augmentation, and as a way of learning more discriminative class centroids. Importantly, our formulation is inherently amenable to including, during training, virtual classes, i.e. class names for which no visual samples are available, further increasing the robustness of the learned prompts. Through extensive evaluations on 11 datasets, we show that our approach (a) significantly outperforms all prior works on soft prompting, and (b) matches and surpasses, for the first time, the accuracy on novel classes obtained by hand-crafted prompts and CLIP for the majority of the test datasets. Code will be made available.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a second cross entropy loss to minimize the distance between the learned soft prompts and a set of hand-engineered manual prompts (obtained by prompt engineering), and can be interpreted in multiple ways including as a regularizer, as a means for language-based augmentation, and as a way of learning more discriminating class centroids.'}",http://arxiv.org/pdf/2210.01115
-chatgpt4pcg competition: character-like level generation for science birds,Pittawat Taveekitworachai,"This paper presents the first ChatGPT4PCG Competition at the 2023 IEEE Conference on Games. The objective of this competition is for participants to create effective prompts for ChatGPT--enabling it to generate Science Birds levels with high stability and character-like qualities--fully using their creativity as well as prompt engineering skills. ChatGPT is a conversational agent developed by OpenAI. Science Birds is selected as the competition platform because designing an Angry Birds-like level is not a trivial task due to the in-game gravity; the quality of the levels is determined by their stability. To lower the entry barrier to the competition, we limit the task to the generation of capitalized English alphabetical characters. We also allow only a single prompt to be used for generating all the characters. Here, the quality of the generated levels is determined by their stability and similarity to the given characters. A sample prompt is provided to participants for their reference. An experiment is conducted to determine the effectiveness of several modified versions of this sample prompt on level stability and similarity by testing them on several characters. To the best of our knowledge, we believe that ChatGPT4PCG is the first competition of its kind and hope to inspire enthusiasm for prompt engineering in procedural content generation.",,http://arxiv.org/pdf/2303.15662
-advancements in casualty modelling facilitated by the usgs prompt assessment of global earthquakes for response (pager) system,D. Wald,,,
-a review on expert system and its applications in civil engineering,Muhammad Akram,"Aim of Study This study provides in depth review of expert system and potential benefits achieved with its application in the field of the Civil Engineering. Need of Study Currently, construction projects are facing various chronic problems. One of the reason contributing to these problems is delay in decision making. This leads to need of mechanisms which can enable practitioners in making prompt decision. Hence, expert systems are investigated for seeking opportunities and studying applicability in construction projects to use as a tool for aiding in decision making process. Research Approach This study is carried out through reviewing previous studies conducting in addressing the importance and issues of the expert system. It also includes various research works demonstrating the applications of an expert system in different research areas of civil engineering. Research Findings This study highlighted that expert system is very useful approach for benefiting the practitioner in making quick decision. It will be helpful in avoiding waiting times. It can be successfully applied in the areas of civil engineering specifically construction management. Limitations This paper presents a review on expert system with its application. It has focused on four established methods which are Rule based system (RBS), Case based system (CBS), Fuzzy expert system (FS), Neural network (NN). Importance and Contribution This review paper contributes in the field of Civil Engineering in highlighting the usage and advantages of an expert system.","{'model': 'tldr@v2.0.0', 'text': 'In depth review of expert system and potential benefits achieved with its application in the field of the Civil Engineering highlights that expert system is very useful approach for benefiting the practitioner in making quick decision.'}",
-a real-time analysis and feedback system for quality control of dam foundation grouting engineering,D. Zhong,,,
-guidelines for environmental infection control in health-care facilities. recommendations of cdc and the healthcare infection control practices advisory committee (hicpac).,Lynne Sehulster,"The health-care facility environment is rarely implicated in disease transmission, except among patients who are immunocompromised. Nonetheless, inadvertent exposures to environmental pathogens (e.g., Aspergillus spp. and Legionella spp.) or airborne pathogens (e.g., Mycobacterium tuberculosis and varicella-zoster virus) can result in adverse patient outcomes and cause illness among health-care workers. Environmental infection-control strategies and engineering controls can effectively prevent these infections. The incidence of health-care--associated infections and pseudo-outbreaks can be minimized by 1) appropriate use of cleaners and disinfectants; 2) appropriate maintenance of medical equipment (e.g., automated endoscope reprocessors or hydrotherapy equipment); 3) adherence to water-quality standards for hemodialysis, and to ventilation standards for specialized care environments (e.g., airborne infection isolation rooms, protective environments, or operating rooms); and 4) prompt management of water intrusion into the facility. Routine environmental sampling is not usually advised, except for water quality determinations in hemodialysis settings and other situations where sampling is directed by epidemiologic principles, and results can be applied directly to infection-control decisions. This report reviews previous guidelines and strategies for preventing environment-associated infections in health-care facilities and offers recommendations. These include 1) evidence-based recommendations supported by studies; 2) requirements of federal agencies (e.g., Food and Drug Administration, U.S. Environmental Protection Agency, U.S. Department of Labor, Occupational Safety and Health Administration, and U.S. Department of Justice); 3) guidelines and standards from building and equipment professional organizations (e.g., American Institute of Architects, Association for the Advancement of Medical Instrumentation, and American Society of Heating, Refrigeration, and Air-Conditioning Engineers); 4) recommendations derived from scientific theory or rationale; and 5) experienced opinions based upon infection-control and engineering practices. The report also suggests a series of performance measurements as a means to evaluate infection-control efforts.","{'model': 'tldr@v2.0.0', 'text': 'This report reviews previous guidelines and strategies for preventing environment-associated infections in health-care facilities and offers recommendations, including evidence-based recommendations supported by studies and experienced opinions based upon infection-control and engineering practices.'}",
-improved charge separation in poly(heptazine-triazine) imides with semi-coherent interfaces for photocatalytic hydrogen evolution.,Jian Zhang,"The construction of heterojunctions is a promising manner to accelerate the separation and transfer of the charge carriers at the interface. Herein, a binary poly(heptazine-triazine) imides (PHI/PTI) with semi-coherent interfaces was fabricated via a facile two-step salt-melt synthetic process. The built-in electric fields at the semi-coherent interface endow prompt exciton splitting and charge carrier separation. Hence, the optimized PHI/PTI-based copolymer presents a high apparent quantum yield (AQY=64%) for visible-light driven hydrogen production, by the aids of K 2 HPO 4 as charge transfer mediator. This study provides physical insights for the rational promotion of the photocatalytic performance from the viewpoint of interfacial engineering of photocatalytic junctions on crystalline carbon nitride based semiconductors.",,
-stylemc: multi-channel based fast text-guided image generation and manipulation,Umut Kocasari,"Discovering meaningful directions in the latent space of GANs to manipulate semantic attributes typically requires large amounts of labeled data. Recent work aims to overcome this limitation by leveraging the power of Contrastive Language-Image Pre-training (CLIP), a joint text-image model. While promising, these methods require several hours of preprocessing or training to achieve the desired manipulations. In this paper, we present StyleMC, a fast and efficient method for text-driven image generation and manipulation. StyleMC uses a CLIP-based loss and an identity loss to manipulate images via a single text prompt without significantly affecting other attributes. Unlike prior work, StyleMC requires only a few seconds of training per text prompt to find stable global directions, does not require prompt engineering and can be used with any pre-trained StyleGAN2 model. We demonstrate the effectiveness of our method and compare it to state-of-the-art methods. Our code can be found at http://catlab-team.github.io/stylemc.","{'model': 'tldr@v2.0.0', 'text': 'StyleMC is presented, a fast and efficient method for text-driven image generation and manipulation that uses a CLIP-based loss and an identity loss to manipulate images via a single text prompt without significantly affecting other attributes.'}",https://arxiv.org/pdf/2112.08493
-will it blend? mixing training paradigms & prompting for argument quality prediction,Michiel van der Meer,"This paper describes our contributions to the Shared Task of the 9th Workshop on Argument Mining (2022). Our approach uses Large Language Models for the task of Argument Quality Prediction. We perform prompt engineering using GPT-3, and also investigate the training paradigms multi-task learning, contrastive learning, and intermediate-task training. We find that a mixed prediction setup outperforms single models. Prompting GPT-3 works best for predicting argument validity, and argument novelty is best estimated by a model trained using all three training paradigms.","{'model': 'tldr@v2.0.0', 'text': 'This paper performs prompt engineering using GPT-3, and investigates the training paradigms multi-task learning, contrastive learning, and intermediate-task training to find that a mixed prediction setup outperforms single models.'}",http://arxiv.org/pdf/2209.08966
-the infinite index: information retrieval on generative text-to-image models,Niklas Deckers,"Conditional generative models such as DALL-E and Stable Diffusion generate images based on a user-defined text, the prompt. Finding and refining prompts that produce a desired image has become the art of prompt engineering. Generative models do not provide a built-in retrieval model for a user’s information need expressed through prompts. In light of an extensive literature review, we reframe prompt engineering for generative models as interactive text-based retrieval on a novel kind of “infinite index”. We apply these insights for the first time in a case study on image generation for game design with an expert. Finally, we envision how active learning may help to guide the retrieval of generated images.","{'model': 'tldr@v2.0.0', 'text': 'This work reframe prompt engineering for generative models as interactive text-based retrieval on a novel kind of “infinite index”, and envision how active learning may help to guide the retrieval of generated images.'}",https://dl.acm.org/doi/pdf/10.1145/3576840.3578327
-the amount of errors in chatgpt’s responses is indirectly correlated with the number of publications related to the topic under investigation,K. Metze,,"{'model': 'tldr@v2.0.0', 'text': 'This study designed a study analyzing whether ChatGPT could contribute in writing a critical review on Chagas disease, focusing on the role of individual researchers.'}",
-exploring the benefits of visual prompting in differential privacy,Yizhe Li,"Visual Prompting (VP) is an emerging and powerful technique that allows sample-efficient adaptation to downstream tasks by engineering a well-trained frozen source model. In this work, we explore the benefits of VP in constructing compelling neural network classifiers with differential privacy (DP). We explore and integrate VP into canonical DP training methods and demonstrate its simplicity and efficiency. In particular, we discover that VP in tandem with PATE, a state-of-the-art DP training method that leverages the knowledge transfer from an ensemble of teachers, achieves the state-of-the-art privacy-utility trade-off with minimum expenditure of privacy budget. Moreover, we conduct additional experiments on cross-domain image classification with a sufficient domain gap to further unveil the advantage of VP in DP. Lastly, we also conduct extensive ablation studies to validate the effectiveness and contribution of VP under DP consideration. Our code is available at (https://github.com/EzzzLi/Prompt-PATE).","{'model': 'tldr@v2.0.0', 'text': 'This work discovers that VP in tandem with PATE, a state-of-the-art DP training method that leverages the knowledge transfer from an ensemble of teachers, achieves the state- of theart privacy-utility trade-off with minimum expenditure of privacy budget.'}",https://arxiv.org/pdf/2303.12247
-beyond prompting: making pre-trained language models better zero-shot learners by clustering representations,Yu Fei,"Recent work has demonstrated that pre-trained language models (PLMs) are zero-shot learners. However, most existing zero-shot methods involve heavy human engineering or complicated self-training pipelines, hindering their application to new situations. In this work, we show that zero-shot text classification can be improved simply by clustering texts in the embedding spaces of PLMs. Specifically, we fit the unlabeled texts with a Bayesian Gaussian Mixture Model after initializing cluster positions and shapes using class names. Despite its simplicity, this approach achieves superior or comparable performance on both topic and sentiment classification datasets and outperforms prior works significantly on unbalanced datasets. We further explore the applicability of our clustering approach by evaluating it on 14 datasets with more diverse topics, text lengths, and numbers of classes. Our approach achieves an average of 20% absolute improvement over prompt-based zero-shot learning. Finally, we compare different PLM embedding spaces and find that texts are well-clustered by topics even if the PLM is not explicitly pre-trained to generate meaningful sentence embeddings. This work indicates that PLM embeddings can categorize texts without task-specific fine-tuning, thus providing a new way to analyze and utilize their knowledge and zero-shot learning ability.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that zero-shot text classification can be improved simply by clustering texts in the embedding spaces of PLMs, indicating that PLM embeddings can categorize texts without task-specific fine-tuning, thus providing a new way to analyze and utilize their knowledge and zero- shot learning ability.'}",http://arxiv.org/pdf/2210.16637
-beat over the old ground with new strategy: engineering as···as interaction in arsenite-based dawson cluster β-[w18o54(aso3)2(6-.,Q. Han,"By reaction of [As2W19O67(H2O)](14-), NiCl2·6H2O, and phen under hydrothermal conditions, a new organic-inorganic tungstoarsenate hybrid [Ni(phen)3]4[As2W18O60]{[Ni(phen)2][H2As2W18O60]}·12H2O (where phen = 1,10-phenanthroline) (1) was obtained via self-assembly and characterized by elemental analysis, infrared (IR) spectroscopy, solid UV-vis absorption spectrum, and single-crystal X-ray diffraction (XRD). An unprecedented 18-tungstoarsenate Dawson cluster, β-[W18O54(AsO3)2](6-), encapsulating two pyramidal arsenite AsO3(3-) anions as templates and exhibiting interesting intramolecular As···As interaction is first achieved. 1 displays a one-dimensional (1D) chain architecture constructed by alternating β-[W18O54(AsO3)2](6-) and nickel(II) complexes [Ni(phen)2)](2+). The resulting hybrid can act as a photocatalyst to prompt the degradation of Rhodamine B (RhB) with excellent efficiency.","{'model': 'tldr@v2.0.0', 'text': 'A new organic-inorganic tungstoarsenate hybrid that can act as a photocatalyst to prompt the degradation of Rhodamine B (RhB) with excellent efficiency is presented.'}",
-thickness‐dependent facet junction control of layered bioio3 single crystals for highly efficient co2 photoreduction,Fang Chen,"Thin layer fabrication and crystal facet engineering favor the prompt charge transfer from bulk to the surface of a material and spatial charge separation among different facets, tremendously benefitting photocatalytic activity. However, the thickness and surface facet composition are considered as two entwined characteristics of layered materials with well‐defined and tunable shapes, which possess great promise to achieve the simultaneous manipulation of charge transfer and spatial separation. Herein, it is demonstrated that one solution for the aforementioned issue by controllably regulating the surface {010}/{100} facet junctions of a layered thickness‐tunable bismuth‐based material, BiOIO3. The attenuation in thickness of BiOIO3 nanoplates shortens the diffusion pathway of charge carriers, and more importantly the tuning of nanolayer thickness renders the ratio variation of the top {010} facet to the lateral {100} facet, which dominates the spatial separation of photogenerated electrons and holes. As a result, the highest CO evolution rate from CO2 reduction over BiOIO3 nanoplates with the optimal thickness and ratio of exposed facets reaches 5.42 µmol g−1 h−1, over 300% that of the bulk counterpart (1.77 µmol g−1 h−1). This work paves a new way for governing charge movement behaviors on the basis of the synergistic engineering of layer structure and exposing facets.",,
-textgraphs-16 natural language premise selection task: zero-shot premise selection with prompting generative language models,Liubov Kovriguina,"Automated theorem proving can benefit a lot from methods employed in natural language processing, knowledge graphs and information retrieval: this non-trivial task combines formal languages understanding, reasoning, similarity search. We tackle this task by enhancing semantic similarity ranking with prompt engineering, which has become a new paradigm in natural language understanding. None of our approaches requires additional training. Despite encouraging results reported by prompt engineering approaches for a range of NLP tasks, for the premise selection task vanilla re-ranking by prompting GPT-3 doesn’t outperform semantic similarity ranking with SBERT, but merging of the both rankings shows better results.","{'model': 'tldr@v2.0.0', 'text': 'This work focuses on enhancing semantic similarity ranking with prompt engineering, which has become a new paradigm in natural language understanding, and introduces a new approach to premise selection task.'}",
-generating requirements elicitation interview scripts with large language models,Binnur Görer,"Requirements elicitation interviews are the most popular requirements elicitation technique and an integral part of requirements engineering education. Good and bad interview scripts provide students with examples of applying the theory. Constructing an interview script requires technical knowledge, practical experience, and creativity. As a result, only a few educational interview scripts are available to the community. This paper explores automatically generating interview scripts with large language models through prompt engineering. Our contribution is two-fold: First, we present a graph representation of interactive interview scripts. Second, we apply prompt engineering techniques to generate business domain descriptions, linear scripts, and conversation pieces focused on certain types of mistakes. Our findings indicate that large language models face challenges in handling interview conversation graphs. However, we can enhance the quality of the generated interview scripts by decomposing the task into smaller components and refining the prompts to provide more precise instructions.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate that large language models face challenges in handling interview conversation graphs, but the quality of the generated interview scripts can be enhanced by decomposing the task into smaller components and refining the prompts to provide more precise instructions.'}",
-an empirical evaluation of prompting strategies for large language models in zero-shot clinical natural language processing,S. Sivarajkumar,"Large language models (LLMs) have shown remarkable capabilities in Natural Language Processing (NLP), especially in domains where labeled data is scarce or expensive, such as clinical domain. However, to unlock the clinical knowledge hidden in these LLMs, we need to design effective prompts that can guide them to perform specific clinical NLP tasks without any task-specific training data. This is known as in-context learning, which is an art and science that requires understanding the strengths and weaknesses of different LLMs and prompt engineering approaches. In this paper, we present a comprehensive and systematic experimental study on prompt engineering for five clinical NLP tasks: Clinical Sense Disambiguation, Biomedical Evidence Extraction, Coreference Resolution, Medication Status Extraction, and Medication Attribute Extraction. We assessed the prompts proposed in recent literature, including simple prefix, simple cloze, chain of thought, and anticipatory prompts, and introduced two new types of prompts, namely heuristic prompting and ensemble prompting. We evaluated the performance of these prompts on three state-of-the-art LLMs: GPT-3.5, BARD, and LLAMA2. We also contrasted zero-shot prompting with few-shot prompting, and provide novel insights and guidelines for prompt engineering for LLMs in clinical NLP. To the best of our knowledge, this is one of the first works on the empirical evaluation of different prompt engineering approaches for clinical NLP in this era of generative AI, and we hope that it will inspire and inform future research in this area.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a comprehensive and systematic experimental study on prompt engineering for five clinical NLP tasks: Clinical Sense Disambiguation, Biomedical Evidence Extraction, Coreference Resolution, Medication Status Ext extraction, and Medication Attribute Extraction.'}",https://arxiv.org/pdf/2309.08008
-study on prompt methane hydrate formation derived by addition of ionic liquid,T. Kitajima,"Aims: The objective of this study is to establish the fundamental model on methane hydrate formation and to accelerate the rate of methane hydrate formation with a small amount of ionic liquid and to investigate the effect of ionic liquid on hydrate formation. Study Design: Experimental study containing modeling. Place and Duration of Study: The present study was held between April 2010 and February 2012 at Division of Chemical Engineering, Department of Materials Engineering Science, Osaka University. Methodology: Methane hydrate formation was modelized based on the driving force, fugacity difference before and after hydrate formation. BMIM-hexafuorophosphate (BMIMPF6) was adopted as a representative of 1-butyl-3-methylimidazolium (BMIM) salts. The temperature dependence of methane hydrate formation rate was investigated and activation energy of hydrate formation was evaluated for the pure water and BMIM-PF6 aqueous solution systems. Research Article American Chemical Science Journal, 2(3): 100-110, 2012 101 Results: An addition of small amount of BMIM-PF6 is able to accelerate the methane hydrate formation. The pseudo-first order reaction model is applicable to the methane hydrate formation in both the pure water and BMIM-PF6 aqueous solution systems. The activation energies of methane hydrate formation are large negative values in the both systems, that is, the methane hydrate formation process is considered to be composed of the precursory hydration and succeeding hydrate formation. A very small amount of BMIMPF6 seems to change the interfacial energy between guest molecules and precursor or initial hydrate particles without the change of the activation energy for overall methane hydrate formation.",,https://zenodo.org/record/8284/files/1345297018-Kitajima_etal_2-3-2012ACSj1512.pdf
-an experimental investigation of analogy formation using the engineering-to-biology thesaurus,M. Glier,"Identifying applicable biological systems for engineering design remains a persistent challenge for bioinspired design. Previous researchers have proposed an Engineering-to-Biology thesaurus that allows designers to identify biological keywords that are functionally similar to terms in the Functional Basis. This work presents an experimental examination of the effectiveness of the Engineering-to-Biology thesaurus. A group of 100 mechanical engineering students are presented with a simple design problem: to create a device to remove the husk and silk from ears of corn. The participants read passages drawn from a large biology corpus with keywords from the Engineering-to-Biology thesaurus and indicate which passages prompt some idea for solving the design problem. The analysis of student responses indicates that students’ level of design training is not a significant factor in the number of analogies they found in the passages and that some non-random criteria is used to identify passages as useful for idea generation. Passages that rarely offer participants ideas can be reasonably well classified as either being too technical for a lay-reader to understand or lacking information on a biological system. Passages that typically offer ideas cannot be so easily classified. Finally, keywords from the Engineering-to-Biology thesaurus are examined to find that keywords very specific to biology and those that are very common words with multiple meanings are rarely contained in sentences that offer strong design inspiration.Copyright © 2013 by ASME",,
-the rising engineering education faculty experience (reefe): preparing junior colleagues,Cory Hixson,"Despite the importance of professional development, for most graduate students as up-andcoming faculty members professional development is informal at best. Graduate programs often emphasize gaining technical knowledge, skills, and abilities through courses and research projects, but provide less opportunity for future faculty members to gain experience with teaching, service, communication, assessment, proposal writing, etc. To provide this experience, we developed the Rising Engineering Education Faculty Experience (REEFE). Founded on theoretical and practical models of graduate student development, REEFE is an innovative faculty apprenticeship program for engineering education graduate students that places students in visiting faculty member positions at host schools. This paper describes the foundations of REEFE and the program itself. We also offer lessons learned from the host school, sending school, and participants based on prior REEFE implementations. We hope our learnings prompt discussions regarding how to effectively prepare future engineering education faculty.","{'model': 'tldr@v2.0.0', 'text': 'The foundations of REEFE are described, an innovative faculty apprenticeship program for engineering education graduate students that places students in visiting faculty member positions at host schools and lessons learned from the host school, sending school, and participants are offered.'}",
-economic bases of formation and development of financial engineering in financial innovation,A. Ajupov,"Today the main spheres of society are developed mainly through the using of innovative technologies and constant updating of products. The result of active innovative activity of the economic subjects directed on the solution of different financial problems including the management all of increasing innovative risks became the prompt progress of financial engineering.м In turn, the process of designing innovative financial products created to reduce the various types of risks through instruments of financial engineering is characterized in economic science as a risk-engineering. DOI: 10.5901/mjss.2014.v5n24p148",,https://www.richtmann.org/journal/index.php/mjss/article/download/4953/4794
-the meaning of prompt feedback and other student perceptions of feedback: should national student survey scores be taken at face value?,P. M. Mendes,"Abstract Regardless of the institution or discipline, the UK’s National Student Survey (NSS) has consistently highlighted that the level of student satisfaction about the feedback they receive is notably lower than for other aspects of their learning experience. This study explored how students understand concepts and practices rated through NSS questions evaluating feedback practices in higher education. Drawing on questionnaires completed by first, third and fourth year Chemical Engineering students, the study calls into particular question the reliability of NSS data on promptness of feedback. In conclusion, it calls for greater attention to be paid at institutional level to the identification and management of students’ perceptions and expectations of the process, content and outcomes of feedback.",,
-an empirical study on few-shot knowledge probing for pretrained language models,Tianxing He,"Prompt-based knowledge probing for 1-hop relations has been used to measure how much world knowledge is stored in pretrained language models. Existing work uses considerable amounts of data to tune the prompts for better performance. In this work, we compare a variety of approaches under a few-shot knowledge probing setting, where only a small number (e.g., 10 or 20) of example triples are available. In addition, we create a new dataset named TREx-2p, which contains 2-hop relations. We report that few-shot examples can strongly boost the probing performance for both 1-hop and 2-hop relations. In particular, we find that a simple-yet-effective approach of finetuning the bias vectors in the model outperforms existing prompt-engineering methods. Our dataset and code are available at \url{https://github.com/cloudygoose/fewshot_lama}.","{'model': 'tldr@v2.0.0', 'text': 'A simple-yet-effective approach of finetuning the bias vectors in the model outperforms existing prompt-engineering methods under a few-shot knowledge probing setting, where only a small number of example triples are available.'}",
-feature engineering in the nli shared task 2013: charles university submission report,Barbora Hladká,"Our goal is to predict the first language (L1) of English essays’s authors with the help of the TOEFL11 corpus where L1, prompts (topics) and proficiency levels are provided. Thus we approach this task as a classification task employing machine learning methods. Out of key concepts of machine learning, we focus on feature engineering. We design features across all the L1 languages not making use of knowledge of prompt and proficiency level. During system development, we experimented with various techniques for feature filtering and combination optimized with respect to the notion of mutual information and information gain. We trained four different SVM models and combined them through majority voting achieving accuracy 72.5%.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to predict the first language (L1) of English essays’s authors with the help of the TOEFL11 corpus where L1, prompts (topics) and proficiency levels are provided and focuses on feature engineering.'}",
-sepsis diagnosis and treatment using nanomaterials,Jaesung Lim,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the recent research advancements in nanotechnology to diagnose and treat sepsis with a brief introduction to sepsi is presented.'}",https://link.springer.com/content/pdf/10.1007/s13534-021-00200-0.pdf
-optimal location of an intake at a reservoir prone to salt diffusion,A. Krutov,"There are two main factors which prompt to look for an alternative to traditional approaches to the provision of drinking water. They are climate change and population growth. Climate change and its increasing scale are no longer in doubt. It tends to increase the frequency and intensity of droughts. Changes in average water availability in most Central Asian river basins are estimated to be drastically big for the next 30 years. And, groundwater recharge may also be affected with a reduction in the availability of groundwater for drinking water in some regions. Water use has been increasing worldwide by about 1% per year since the 1980s, driven by a combination of population growth, socio-economic development and changing consumption patterns. Global water demand is expected to continue increasing at a similar rate until 2050, accounting for an increase of 20 to 30% above the current level of water use, mainly due to rising demand in the industrial and domestic sectors. These obliging to redefine the strategy for the use of water resources to ensure sustainable drinking water supply forcing the use of all available water resources even those that were not previously taken into account, for example, off stream storage reservoirs. Research, development and innovation play an important role in supporting informed decision-making. Therefore, further scientific and engineering studies are also needed for the development of financially affordable, safe and efficient infrastructure services in the areas of drinking water supply, sanitation and hygiene and its components. The article considers a particular reservoir the bottom of which is composed of saline soils operating in regular filling and emptying and subject to significant wind effects. The research conducted allowed to find a place of water intake location which provides a minimum salt content in the water abstracted.",,
-combining computer vision and deep learning to enable ultra-scale aerial phenotyping and precision agriculture: a case study of lettuce production,Alan Bauer,,"{'model': 'tldr@v2.0.0', 'text': 'AirSurf is reported, an automated and open-source analytic platform that combines modern computer vision, up-to-date machine learning, and modular software engineering in order to measure yield-related phenotypes from ultra-large aerial imagery.'}",https://www.nature.com/articles/s41438-019-0151-5.pdf
-should engineering ethics be taught?,Charles J. Abaté,,"{'model': 'tldr@v2.0.0', 'text': 'There is a justification to teaching engineering ethics, insofar as the authors are able to clearly identify the most desirable and efficacious pedagogical approach to the subject area, which is proposed to be a case study-based format that utilizes the principle of human cognitive pattern recognition.'}",
-generating disentangled arguments with prompts: a simple event extraction framework that works,Jinghui Si,"Event Extraction bridges the gap between text and event signals. Based on the assumption of trigger-argument dependency, existing approaches have achieved state-of-the-art performance with expert-designed templates or complicated decoding constraints. In this paper, for the first time we introduce the prompt-based learning strategy to the domain of Event Extraction, which empowers the automatic exploitation of label semantics on both input and output sides. To validate the effectiveness of the proposed generative method, we conduct extensive experiments with 11 diverse baselines. Empirical results show that, in terms of F1 score on Argument Extraction, our simple architecture is stronger than any other generative counterpart and even competitive with algorithms that require template engineering. Regarding the measure of recall, it sets new overall records for both Argument and Trigger Extractions. We hereby recommend this framework to the community, with the code publicly available at https://github.com/RingBDStack/GDAP.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the prompt-based learning strategy to the domain of Event Extraction, which empowers the automatic exploitation of label semantics on both input and output sides and sets new overall records for both Argument and Trigger Extractions.'}",https://eprints.whiterose.ac.uk/191435/1/jinghui_GDAP_icassp2022.pdf
-prompt versus problem: helping students learn to frame problems and think creatively,J. Lai,"Currently engineering students are given opportunities to apply knowledge learned in the classroom to hands-on design projects. While being trained to solve problems, are creativity skills also developed as part of that experience? According to a recent study, surveys found that students do not feel that creativity is part of that experience (Kazerounian and Foley 2007). Two questions are posed to frame discussion: how does current curriculum shape the way students approach problems? What can be done to create assignments that encourage creative thinking? Four observations are made from a class that focuses on creativity and design.",,
-prompt magnetic investigations of archaeological remains in areas of infrastructure development: israeli experience,L. Eppelbaum,"Magnetic prospecting is the most widely used method for the investigation of archaeological remains in the world. This method is often applied before and during various engineering exploration (e.g. road construction, preparing areas for various types of building) and areas of agricultural utilization. Taking into account an extremely high concentration of archaeological sites of different age at Israeli territory, prompt geophysical–archaeological assessment of areas allocated for engineering and agricultural utilization, is a necessary preliminary component of the design process. However, noise both of natural and artificial origin strongly complicates interpretation of observed magnetic anomalies. The most significant disturbing factors are oblique magnetization (magnetization vector in the Israeli territory is inclined about 45°), complex geological structure (presence of a variety of disturbing bodies) of investigated site, influence of rugged topography, as well as influence of modern iron‐containing objects. The non‐conventional procedures developed by the authors are intended for application in such complicated environments and include elimination of noise from various sources, selection of ‘useful’ anomalies against the noise background, qualitative and quantitative interpretation of anomalies and three‐dimensional physical–geological modelling. Effective integration of these procedures with other interpretation methods allowed rapid localization of tens of archaeological remains and their protection from unpremeditated destruction. Copyright © 2001 John Wiley & Sons, Ltd.",,
-climate engineering responses to climate emergencies,J. Blackstock,"Despite efforts to stabilize CO_2 concentrations, it is possible that the climate system could respond abruptly with catastrophic consequences. Intentional intervention in the climate system to avoid or ameliorate such consequences has been proposed as one possible response, should such a scenario arise. In a one-week study, the authors of this report conducted a technical review and evaluation of proposed climate engineering concepts that might serve as a rapid palliative response to such climate emergency scenarios. Because of their potential to induce a prompt (less than one year) global cooling, this study concentrated on Shortwave Climate Engineering (SWCE) methods for moderately reducing the amount of shortwave solar radiation reaching the Earth. The study's main objective was to outline a decade-long agenda of technical research that would maximally reduce the uncertainty surrounding the benefits and risks associated with SWCE. For rigor of technical analysis, the study focused the research agenda on one particular SWCE concept--stratospheric aerosol injection--and in doing so developed several conceptual frameworks and methods valuable for assessing any SWCE proposal.",,http://arxiv.org/pdf/0907.5140
-does your configuration code smell?,Tushar Sharma,"Infrastructure as Code (IaC) is the practice of specifying computing system configurations through code, and managing them through traditional software engineering methods. The wide adoption of configuration management and increasing size and complexity of the associated code, prompt for assessing, maintaining, and improving the configuration code's quality. In this context, traditional software engineering knowledge and best practices associated with code quality management can be leveraged to assess and manage configuration code quality. We propose a catalog of 13 implementation and 11 design configuration smells, where each smell violates recommended best practices for configuration code. We analyzed 4,621 Puppet repositories containing 8.9 million lines of code and detected the cataloged implementation and design configuration smells. Our analysis reveals that the design configuration smells show 9% higher average co-occurrence among themselves than the implementation configuration smells. We also observed that configuration smells belonging to a smell category tend to co-occur with configuration smells belonging to another smell category when correlation is computed by volume of identified smells. Finally, design configuration smell density shows negative correlation whereas implementation configuration smell density exhibits no correlation with the size of a configuration management system.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a catalog of 13 implementation and 11 design configuration smells, where each smell violates recommended best practices for configuration code, and finds that configuration smells belonging to a smell category tend to co-occur with configuration smell belonging to another smell category when correlation is computed by volume of identified smells.'}",https://zenodo.org/record/1997943/files/ConfigurationSmells_preprint.pdf
-ethics and engineering education,L. Bucciarelli,"In the US, Accreditation Board for Engineering and Technology (ABET) recommends the study of ethics so that students acquire ‘an understanding of professional and ethical responsibility’. For the most part, teaching of the subject relies upon the use of scenarios – both hypothetical and ‘real’– and open discussion framed by the codes. These scenarios and this framing are seriously deficient – lacking in their attention to the complexities of context, almost solely focused on individual agency, while reflecting too narrow and simplistic a view of the responsibilities of the practicing engineer. A critique of several exemplary scenarios, and consideration of the demands placed upon today's professional, prompt reflection on the need for, not just a more expansive reading of the codes of ethics re what it might mean to be ‘responsible’, but a substantial reform of undergraduate engineering education across the board.",,https://dspace.mit.edu/bitstream/1721.1/40284/1/ethics_20_talk.pdf
-contract cheating and assessment design: exploring the relationship,T. Bretag,"Abstract This paper reports on findings from a large Australian research project that explored the relationship between contract cheating and assessment design. Using survey responses from 14,086 students and 1147 educators at eight universities, a multivariate analysis examined the influence of a range of factors on the likelihood that different assessment types would prompt considerations of contract cheating in students. Perceptions of likelihood were highest among students who speak a language other than English at home. Perceptions of likelihood were also higher among students who reported there to be lots of opportunities to cheat, and amongst students who were dissatisfied with the teaching and learning environment. Perceptions of likelihood for certain assessment types were also higher in commerce and engineering than in any other discipline. Overall, four assessment types were perceived by students to be the least likely to be outsourced, however these are also the least likely to be set by educators. The analysis indicates that educators are more likely to use these assessment tasks when they report positively on organisational support for teaching and learning.",,
-the analysis of conventional prompt global strike alternatives,R. P. Hopkins,"United States Strategic Command (STRATCOM) has expressed a strong interest in the development of a non-nuclear offensive strategy capable of reaching all enemy targets across the globe within a sixty-minute period from when the launching order is given to when the target is engaged. STRATCOM has defined this mission as conventional Prompt Global Strike (PGS). Currently, STRATCOM believes the United States does not possess any sort of conventional weapon system capable of meeting the criteria of a conventional PGS weapon. The conventional PGS mission calls for a solution that can be implemented immediately. This means the solution must contain only technology that is currently available. As a result of our literature review and our stakeholder analysis, our problem statement is that STRATCOM lacks a conventional PGS capability to hold time-sensitive targets at risk in denied access areas worldwide. To solve this problem, a systems engineering approach is applied to examine different alternatives by using Value-Focused Thinking to develop value models, math modeling to generate performance data, ArcGIS to highlight the system need and to recommend an implementation plan, and a simulation to demonstrate weapon system survivability. This work will result in recommendations for conventional Prompt Global Strike alternatives.",,
-eeg-based depression detection using convolutional neural network with demographic attention mechanism,Xiaowei Zhang,"Electroencephalography (EEG)-based depression detection has become a hot topic in the development of biomedical engineering. However, the complexity and nonstationarity of EEG signals are two biggest obstacles to this application. In addition, the generalization of detection algorithms may be degraded owing to the influences brought by individual differences. In view of the correlation between EEG signals and individual demographics, such as gender, age, etc., and influences of these demographic factors on the incidence of depression, it would be better to incorporate demographic factors during EEG modeling and depression detection. In this work, we constructed an one-dimensional Convolutional Neural Network (1-D CNN) to obtain more effective features of EEG signals, then integrated gender and age factors into the 1-D CNN via an attention mechanism, which could prompt our 1-D CNN to explore complex correlations between EEG signals and demographic factors, and generate more effective high-level representations ultimately for the detection of depression. Experimental results on 170 (81 depressed patients and 89 normal controls) subjects showed that the proposed method is superior to the unitary 1-D CNN without gender and age factors and two other ways of incorporating demographics. This work also indicates that organic mixture of EEG signals and demographic factors is promising for the detection of depression.Clinical relevance-This work indicates that organically mixture of EEG signals and demographic factors is promising for the detection of depression","{'model': 'tldr@v2.0.0', 'text': 'Experimental results showed that the proposed method is superior to the unitary 1-D CNN without gender and age factors and two other ways of incorporating demographics, and indicates that organic mixture of EEG signals and demographic factors is promising for the detection of depression.'}",
-platelet-rich fibrin scaffolds for cartilage and tendon regenerative medicine: from bench to bedside,S. Barbon,"Nowadays, research in Tissue Engineering and Regenerative Medicine is focusing on the identification of instructive scaffolds to address the requirements of both clinicians and patients to achieve prompt and adequate healing in case of injury. Among biomaterials, hemocomponents, and in particular Platelet-rich Fibrin matrices, have aroused widespread interest, acting as delivery platforms for growth factors, cytokines and immune/stem-like cells for immunomodulation; their autologous origin and ready availability are also noteworthy aspects, as safety- and cost-related factors and practical aspects make it possible to shorten surgical interventions. In fact, several authors have focused on the use of Platelet-rich Fibrin in cartilage and tendon tissue engineering, reporting an increasing number of in vitro, pre-clinical and clinical studies. This narrative review attempts to compare the relevant advances in the field, with particular reference being made to the regenerative role of platelet-derived growth factors, as well as the main pre-clinical and clinical research on Platelet-rich Fibrin in chondrogenesis and tenogenesis, thereby providing a basis for critical revision of the topic.","{'model': 'tldr@v2.0.0', 'text': 'A narrative review attempts to compare the relevant advances in the field, with particular reference being made to the regenerative role of platelet-derived growth factors, as well as the main pre-clinical and clinical research on Platelet-rich Fibrin in chondrogenesis and tenogenesis, thereby providing a basis for critical revision of the topic.'}",https://www.mdpi.com/1422-0067/20/7/1701/pdf?version=1554453987
-a cytochrome c is the natural electron acceptor for nicotine oxidoreductase,Mark Dulchavsky,,"{'model': 'tldr@v2.0.0', 'text': 'It is discovered thatNicA2 donates electrons to a cytochrome c, which means that NicA2 is actually a dehydrogenase, which is surprising, as enzymes of the flavin-containing amine oxidase family were invariably thought to use O2 as an electron acceptor.'}",
-engineering of escherichia coli for targeted delivery of transgenes to her2/neu‐positive tumor cells,Chih-Hsiang Chang,"Targeting of non‐phagocytic tumor cells and prompt release of gene cargos upon entry into tumors are two limiting steps in the bacterial gene delivery path. To tackle these problems, the non‐pathogenic Escherichia coli strain BL21(DE3) was engineered to display the anti‐HER2/neu affibody on the surface. After co‐incubation with tumor cells for 3 h, the anti‐HER2/neu affibody‐presenting E. coli strain was selectively internalized into HER2/neu‐positive SKBR‐3 cells. The invasion efficiency reached as high as 30%. Furthermore, the bacteria were equipped with the phage ϕX174 lysin gene E‐mediated autolysis system. Carrying the transgene (e.g., eukaryotic green fluorescent protein, GFP), the tumor‐targeting bacteria were subjected to the thermal shock to trigger the autolysis system upon entry into HER2/neu‐positive cells. Flow cytometric analysis revealed that 3% of infected cells expressed GFP 24 h post thermal induction. Overall, the results show a promise of the proposed approach for developing bacteria as a delivery carrier. Biotechnol. Bioeng. 2011; 108:1662–1672. © 2011 Wiley Periodicals, Inc.","{'model': 'tldr@v2.0.0', 'text': 'The results show a promise of the proposed approach for developing bacteria as a delivery carrier for treating non‐phagocytic tumor cells and prompt release of gene cargos upon entry into tumors.'}",http://ir.cmu.edu.tw/ir/bitstream/310903500/40944/1/20110502193414.pdf
-higgs-mediated optical amplification in a nonequilibrium superconductor,M. Buzzi,"The quest for new functionalities in quantum materials has recently been extended to non-equilibrium states, which are interesting both because they exhibit new physical phenomena and because of their potential for high-speed device applications. Notable advances have been made in the creation of metastable phases and in Floquet engineering under external periodic driving. In the context of non-equilibrium superconductivity, examples have included the generation of transient superconductivity above the thermodynamic transition temperature, the excitation of coherent Higgs mode oscillations, and the optical control of the interlayer phase in cuprates. Here, we propose theoretically a novel non-equilibrium phenomenon, through which a prompt quench from a metal to a transient superconducting state could induce large oscillations of the order parameter amplitude. We argue that this oscillating mode could act as a source of parametric amplification of the incident radiation. We report experimental results on optically driven K$_3$C$_{60}$ that are consistent with these predictions. The effect is found to disappear when the onset of the excitation becomes slower than the Higgs mode period, consistent with the theory proposed here. These results open new possibilities for the use of collective modes in many-body systems to induce non-linear optical effects.",,http://link.aps.org/pdf/10.1103/PhysRevX.11.011055
-review of smart health monitoring approaches with survey analysis and proposed framework,Sonal Gahlot,"Smart health monitoring system technologies pertaining to the biomedical domain has evinced intense interest, which in turn has led the emphasis on unconventional applications of engineering and internetwork transmission in the field. Throughout the next series of textual and contextual matter, we will be reviewing and exhibiting, such initiatives and established methods which are being used to achieve health monitoring systems approaching toward making smart remote villages and cities. It also collates for the reader the issues and challenges prevalent in the present technical advancements and for societal validation, a survey analysis has also been briefed in this paper to show the demand and requisites of the real world in health monitoring system. Today there is a need for an integrated solution which can diagnose the early cyanotic congenital diseases in newly born babies to monitoring of various geriatric problems in elderly people. The future avenues of maximum life expectancy and uber-prompt medical treatment depend upon it. This paper proposes an approach required to make smart villages and towns with respect to the healthcare domain. This proposal has been made with due consideration and technical backing of survey research including questionnaire-based evaluation and regressive analysis of issues in present technologies.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an approach required to make smart villages and towns with respect to the healthcare domain with due consideration and technical backing of survey research including questionnaire-based evaluation and regressive analysis of issues in present technologies.'}",
-investigation of hydrogen content in chemically delithiated lithium-ion battery cathodes using prompt gamma activation analysis,S. Aghara,,,
-evolutionary genome engineering using a restriction–modification system,Y. Asakura,"Modification of complex microbial cellular processes is often necessary to obtain organisms with particularly favorable characteristics, but such experiments can take many generations to achieve. In the present article, we accelerated the experimental evolution of Escherichia coli populations under selection for improved growth using one of the restriction–modification systems, which have shaped bacterial genomes. This resulted in faster evolutionary changes in both the genome and bacterial growth. Transcriptome/genome analysis at various stages enabled prompt identification of sequential genome rearrangements and dynamic gene-expression changes associated with growth improvement. The changes were related to cell-to-cell communication, the cell death program, as well as mass production and energy consumption. These observed changes imply that improvements in microorganism population growth can be achieved by inactivating the cellular mechanisms regulating fraction of active cells in a population. Some of the mutations were shown to have additive effects on growth. These results open the way for the application of evolutionary genome engineering to generate organisms with desirable properties.","{'model': 'tldr@v2.0.0', 'text': 'This article accelerated the experimental evolution of Escherichia coli populations under selection for improved growth using one of the restriction–modification systems, which have shaped bacterial genomes, resulting in faster evolutionary changes in both the genome and bacterial growth.'}",https://academic.oup.com/nar/article-pdf/39/20/9034/16779366/gkr585.pdf
-sterile and dual-porous aerogels scaffolds obtained through a multistep supercritical co2-based approach,Víctor Santos-Rosales,"Aerogels from natural polymers are endowed with attractive textural and biological properties for biomedical applications due to their high open mesoporosity, low density, and reduced toxicity. Nevertheless, the lack of macroporosity in the aerogel structure and of a sterilization method suitable for these materials restrict their use for regenerative medicine purposes and prompt the research on getting ready-to-implant dual (macro + meso)porous aerogels. In this work, zein, a family of proteins present in materials for tissue engineering, was evaluated as a sacrificial porogen to obtain macroporous starch aerogels. This approach was particularly advantageous since it could be integrated in the conventional aerogel processing method without extra leaching steps. Physicochemical, morphological, and mechanical characterization were performed to study the effect of porogen zein at various proportions (0:1, 1:2, and 1:1 zein:starch weight ratio) on the properties of the obtained starch-based aerogels. From a forward-looking perspective for its clinical application, a supercritical CO2 sterilization treatment was implemented for these aerogels. The sterilization efficacy and the influence of the treatment on the aerogel final properties were evaluated mainly in terms of absence of microbial growth, cytocompatibility, as well as physicochemical, structural, and mechanical modifications.",,https://www.mdpi.com/1420-3049/24/5/871/pdf?version=1551432957
-3d hollow hierarchical structures based on 1d biocl nanorods intersected with 2d bi2wo6 nanosheets for efficient photocatalysis under visible light,Yongchao Ma,"Constructing elaborate catalysts to prompt the charge carrier separation and transport is critical to developing efficient photocatalytic systems. Here, a hierarchical hollow structure based on 1D/2D BiOCl/Bi2WO6 hybrid materials was fabricated by a precursor chemical engineering method. This hybrid is made up of molten 1D BiOCl nanorods and 2D Bi2WO6 nanosheets. The synergetic effect of the presence of BiOCl and specific interfaces between BiOCl and Bi2WO6 provided efficient interfacial charge transfer of photogenerated carriers under visible light. Seamless BiOCl functions like a noble metal, with platinum-like behavior, accelerating the oxidizing ability of fabricated BiOCl/Bi2WO6 hybrids, which was favorable for the photocatalytic decomposition of organic compounds (3.2 times greater for Rhodamine B (RhB) and 4 times greater for Ciprofloxacin (CIP)) over the Bi2WO6 catalysts. The beneficial interfacial interaction between BiOCl and Bi2WO6 resulting from the unique construction prompted the charge transfer from the conduction band of Bi2WO6 to that of BiOCl. The findings presented in this study provide a cost-effective precursor-mediated strategy to realize the critical and efficient separation of photoinduced carriers in environmental remediation applications.","{'model': 'tldr@v2.0.0', 'text': 'The findings presented in this study provide a cost-effective precursor-mediated strategy to realize the critical and efficient separation of photoinduced carriers in environmental remediation applications.'}",https://www.mdpi.com/2079-4991/9/3/322/pdf?version=1551422128
-prototypical networks for few-shot learning,Jake Snell,"We propose Prototypical Networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical Networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend Prototypical Networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prototypical Networks for few-shot classification, and provides an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning.'}",
-learning to compare: relation network for few-shot learning,Flood Sung,"We present a conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each. Our method, called the Relation Network (RN), is trained end-to-end from scratch. During meta-learning, it learns to learn a deep distance metric to compare a small number of images within episodes, each of which is designed to simulate the few-shot setting. Once trained, a RN is able to classify images of new classes by computing relation scores between query images and the few examples of each new class without further updating the network. Besides providing improved performance on few-shot learning, our framework is easily extended to zero-shot learning. Extensive experiments on five benchmarks demonstrate that our simple approach provides a unified and effective approach for both of these two tasks.","{'model': 'tldr@v2.0.0', 'text': 'A conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each, which is easily extended to zero- shot learning.'}",https://www.pure.ed.ac.uk/ws/files/57835996/LearningToCompare.pdf
-matching networks for one shot learning,Oriol Vinyals,"Learning from a few examples remains a key challenge in machine learning. Despite recent advances in important domains such as vision and language, the standard supervised deep learning paradigm does not offer a satisfactory solution for learning new concepts rapidly from little data. In this work, we employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories. Our framework learns a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types. We then define one-shot learning problems on vision (using Omniglot, ImageNet) and language tasks. Our algorithm improves one-shot accuracy on ImageNet from 87.6% to 93.2% and from 88.0% to 93.8% on Omniglot compared to competing approaches. We also demonstrate the usefulness of the same model on language modeling by introducing a one-shot task on the Penn Treebank.","{'model': 'tldr@v2.0.0', 'text': 'This work employs ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories to learn a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types.'}",
-optimization as a model for few-shot learning,S. Ravi,,,
-generalizing from a few examples: a survey on few-shot learning,Yaqing Wang,"Machine learning has been highly successful in data-intensive applications but is often hampered when the data set is small. Recently, Few-Shot Learning (FSL) is proposed to tackle this problem. Using prior knowledge, FSL can rapidly generalize to new tasks containing only a few samples with supervised information. In this paper, we conduct a thorough survey to fully understand FSL. Starting from a formal definition of FSL, we distinguish FSL from several relevant machine learning problems. We then point out that the core issue in FSL is that the empirical risk minimized is unreliable. Based on how prior knowledge can be used to handle this core issue, we categorize FSL methods from three perspectives: (i) data, which uses prior knowledge to augment the supervised experience; (ii) model, which uses prior knowledge to reduce the size of the hypothesis space; and (iii) algorithm, which uses prior knowledge to alter the search for the best hypothesis in the given hypothesis space. With this taxonomy, we review and discuss the pros and cons of each category. Promising directions, in the aspects of the FSL problem setups, techniques, applications and theories, are also proposed to provide insights for future research.","{'model': 'tldr@v2.0.0', 'text': 'A thorough survey to fully understand Few-Shot Learning (FSL), and categorizes FSL methods from three perspectives: data, which uses prior knowledge to augment the supervised experience; model, which used to reduce the size of the hypothesis space; and algorithm, which using prior knowledgeto alter the search for the best hypothesis in the given hypothesis space.'}",
-tadam: task dependent adaptive metric for improved few-shot learning,Boris N. Oreshkin,"Few-shot learning has become essential for producing models that generalize from few examples. In this work, we identify that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms. Our analysis reveals that simple metric scaling completely changes the nature of few-shot algorithm parameter updates. Metric scaling provides improvements up to 14% in accuracy for certain metrics on the mini-Imagenet 5-way 5-shot classification task. We further propose a simple and effective way of conditioning a learner on the task sample set, resulting in learning a task-dependent metric space. Moreover, we propose and empirically test a practical end-to-end optimization procedure based on auxiliary task co-training to learn a task-dependent metric space. The resulting few-shot learning model based on the task-dependent scaled metric achieves state of the art on mini-Imagenet. We confirm these results on another few-shot dataset that we introduce in this paper based on CIFAR100. Our code is publicly available at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms and proposes and empirically test a practical end-to-end optimization procedure based on auxiliary task co-training to learn a task-dependent metric space.'}",
-few-shot learning with retrieval augmented language models,Gautier Izacard,"Large language models have shown impressive few-shot results on a wide range of tasks. However, when knowledge is key for such results, as is the case for tasks such as question answering and fact checking, massive parameter counts to store knowledge seem to be needed. Retrieval augmented models are known to excel at knowledge intensive tasks without the need for as many parameters, but it is unclear whether they work in few-shot settings. In this work we present Atlas, a carefully designed and pre-trained retrieval augmented language model able to learn knowledge intensive tasks with very few training examples. We perform evaluations on a wide range of tasks, including MMLU, KILT and NaturalQuestions, and study the impact of the content of the document index, showing that it can easily be updated. Notably, Atlas reaches over 42% accuracy on Natural Questions using only 64 examples, outperforming a 540B parameters model by 3% despite having 50x fewer parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Atlas, a carefully designed and pre-trained retrieval augmented language model able to learn knowledge intensive tasks with very few training examples, and studies the impact of the content of the document index, showing that it can easily be updated.'}",https://arxiv.org/pdf/2208.03299
-meta-sgd: learning to learn quickly for few shot learning,Zhenguo Li,"Few-shot learning is challenging for learning algorithms that learn each task in isolation and from scratch. In contrast, meta-learning learns from many related tasks a meta-learner that can learn a new task more accurately and faster with fewer examples, where the choice of meta-learners is crucial. In this paper, we develop Meta-SGD, an SGD-like, easily trainable meta-learner that can initialize and adapt any differentiable learner in just one step, on both supervised learning and reinforcement learning. Compared to the popular meta-learner LSTM, Meta-SGD is conceptually simpler, easier to implement, and can be learned more efficiently. Compared to the latest meta-learner MAML, Meta-SGD has a much higher capacity by learning to learn not just the learner initialization, but also the learner update direction and learning rate, all in a single meta-learning process. Meta-SGD shows highly competitive performance for few-shot learning on regression, classification, and reinforcement learning.","{'model': 'tldr@v2.0.0', 'text': 'Meta-SGD, an SGD-like, easily trainable meta-learner that can initialize and adapt any differentiable learner in just one step, shows highly competitive performance for few-shot learning on regression, classification, and reinforcement learning.'}",
-few-shot learning with graph neural networks,Victor Garcia Satorras,"We propose to study the problem of few-shot learning with the prism of inference on a partially observed graphical model, constructed from a collection of input images whose label can be either observed or not. By assimilating generic message-passing inference algorithms with their neural-network counterparts, we define a graph neural network architecture that generalizes several of the recently proposed few-shot learning models. Besides providing improved numerical performance, our framework is easily extended to variants of few-shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.","{'model': 'tldr@v2.0.0', 'text': ""A graph neural network architecture is defined that generalizes several of the recently proposed few-shot learning models and provides improved numerical performance, and is easily extended to variants of few- shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.""}",
-canet: class-agnostic segmentation networks with iterative refinement and attentive few-shot learning,Chi Zhang,"Recent progress in semantic segmentation is driven by deep Convolutional Neural Networks and large-scale labeled image datasets. However, data labeling for pixel-wise segmentation is tedious and costly. Moreover, a trained model can only make predictions within a set of pre-defined classes. In this paper, we present CANet, a class-agnostic segmentation network that performs few-shot segmentation on new classes with only a few annotated images available. Our network consists of a two-branch dense comparison module which performs multi-level feature comparison between the support image and the query image, and an iterative optimization module which iteratively refines the predicted results. Furthermore, we introduce an attention mechanism to effectively fuse information from multiple support examples under the setting of k-shot learning. Experiments on PASCAL VOC 2012 show that our method achieves a mean Intersection-over-Union score of 55.4% for 1-shot segmentation and 57.1% for 5-shot segmentation, outperforming state-of-the-art methods by a large margin of 14.6% and 13.2%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Canet is presented, a class-agnostic segmentation network that performs few-shot segmentation on new classes with only a few annotated images available, and introduces an attention mechanism to effectively fuse information from multiple support examples under the setting of k-shot learning.'}",https://dr.ntu.edu.sg/bitstream/10356/144391/2/gusoheng%20paper1%20cvpr%202019.pdf
-fleurs: few-shot learning evaluation of universal representations of speech,Alexis Conneau,"We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Speech-Text Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like speech-only w2v-BERT [1] and speech-text multimodal mSLAM [2]. The goal of FLEURS is to enable speech technology in more languages and catalyze research in low-resource speech understanding.1.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2205.12446
-pushing the limits of simple pipelines for few-shot learning: external data and fine-tuning make a difference,S. Hu,"Few-shot learning (FSL) is an important and topical problem in computer vision that has motivated extensive research into numerous methods spanning from sophisticated metalearning methods to simple transfer learning baselines. We seek to push the limits of a simple-but-effective pipeline for real-worldfew-shot image classification in practice. To this end, we explore few-shot learning from the perspective of neural architecture, as well as a three stage pipeline of pre-training on external data, meta-training with labelled few-shot tasks, and task-specific fine-tuning on unseen tasks. We investigate questions such as: ① How pre-training on external data benefits FSL? ② How state of the art transformer architectures can be exploited? and ③ How to best exploit finetuning? Ultimately, we show that a simple transformer-based pipeline yields surprisingly good performance on standard benchmarks such as Mini-ImageNet, CIFAR-FS, CDFSL and Meta-Dataset. Our code is available at https://hushell.github.io/pmf.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a simple transformer-based pipeline yields surprisingly good performance on standard benchmarks such as Mini-ImageNet, CIFAR-FS, CDFSL and Meta-Dataset.'}",https://www.pure.ed.ac.uk/ws/files/262579857/Pushing_the_Limits_HU_DOA02032022_AFV.pdf
-meta-transfer learning for few-shot learning,Qianru Sun,"Meta-learning has been proposed as a framework to address the challenging few-shot learning setting. The key idea is to leverage a large number of similar few-shot tasks in order to learn how to adapt a base-learner to a new task for which only a few labeled samples are available. As deep neural networks (DNNs) tend to overfit using a few samples only, meta-learning typically uses shallow neural networks (SNNs), thus limiting its effectiveness. In this paper we propose a novel few-shot learning method called meta-transfer learning (MTL) which learns to adapt a deep NN for few shot learning tasks. Specifically, ""meta"" refers to training multiple tasks, and ""transfer"" is achieved by learning scaling and shifting functions of DNN weights for each task. In addition, we introduce the hard task (HT) meta-batch scheme as an effective learning curriculum for MTL. We conduct experiments using (5-class, 1-shot) and (5-class, 5-shot) recognition tasks on two challenging few-shot learning benchmarks: miniImageNet and Fewshot-CIFAR100. Extensive comparisons to related works validate that our meta-transfer learning approach trained with the proposed HT meta-batch scheme achieves top performance. An ablation study also shows that both components contribute to fast convergence and high accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A novel few-shot learning method called meta-transfer learning (MTL) which learns to adapt a deep NN for few shot learning tasks and introduces the hard task (HT) meta-batch scheme as an effective learning curriculum for MTL.'}",
-graph information aggregation cross-domain few-shot learning for hyperspectral image classification.,Yuxiang Zhang,"Most domain adaptation (DA) methods in cross-scene hyperspectral image classification focus on cases where source data (SD) and target data (TD) with the same classes are obtained by the same sensor. However, the classification performance is significantly reduced when there are new classes in TD. In addition, domain alignment, as one of the main approaches in DA, is carried out based on local spatial information, rarely taking into account nonlocal spatial information (nonlocal relationships) with strong correspondence. A graph information aggregation cross-domain few-shot learning (Gia-CFSL) framework is proposed, intending to make up for the above-mentioned shortcomings by combining FSL with domain alignment based on graph information aggregation. SD with all label samples and TD with a few label samples are implemented for FSL episodic training. Meanwhile, intradomain distribution extraction block (IDE-block) and cross-domain similarity aware block (CSA-block) are designed. The IDE-block is used to characterize and aggregate the intradomain nonlocal relationships and the interdomain feature and distribution similarities are captured in the CSA-block. Furthermore, feature-level and distribution-level cross-domain graph alignments are used to mitigate the impact of domain shift on FSL. Experimental results on three public HSI datasets demonstrate the superiority of the proposed method. The codes will be available from the website: https://github.com/YuxiangZhang-BIT/IEEE_TNNLS_Gia-CFSL.","{'model': 'tldr@v2.0.0', 'text': 'A graph information aggregation cross-domain few-shot learning (Gia-CFSL) framework is proposed, intending to make up for the above-mentioned shortcomings by combining FSL with domain alignment based on graph information aggregating.'}",
-learning to propagate labels: transductive propagation network for few-shot learning,Yanbin Liu,"The goal of few-shot learning is to learn a classifier that generalizes well even when trained with a limited number of training instances per class. The recently introduced meta-learning approaches tackle this problem by learning a generic classifier across a large number of multiclass classification tasks and generalizing the model to a new task. Yet, even with such meta-learning, the low-data problem in the novel classification task still remains. In this paper, we propose Transductive Propagation Network (TPN), a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem. Specifically, we propose to learn to propagate labels from labeled instances to unlabeled test instances, by learning a graph construction module that exploits the manifold structure in the data. TPN jointly learns both the parameters of feature embedding and the graph construction in an end-to-end manner. We validate TPN on multiple benchmark datasets, on which it largely outperforms existing few-shot learning approaches and achieves the state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Transductive Propagation Network (TPN), a novel meta-learning framework for transductive inference that classifies the entire test set at once to alleviate the low-data problem.'}",
-revisiting local descriptor based image-to-class measure for few-shot learning,Wenbin Li,"Few-shot learning in image classification aims to learn a classifier to classify images when only few training examples are available for each class. Recent work has achieved promising classification performance, where an image-level feature based measure is usually used. In this paper, we argue that a measure at such a level may not be effective enough in light of the scarcity of examples in few-shot learning. Instead, we think a local descriptor based image-to-class measure should be taken, inspired by its surprising success in the heydays of local invariant features. Specifically, building upon the recent episodic training mechanism, we propose a Deep Nearest Neighbor Neural Network (DN4 in short) and train it in an end-to-end manner. Its key difference from the literature is the replacement of the image-level feature based measure in the final layer by a local descriptor based image-to-class measure. This measure is conducted online via a k-nearest neighbor search over the deep local descriptors of convolutional feature maps. The proposed DN4 not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity, thanks to the exchangeability of visual patterns across the images in the same class. Our work leads to a simple, effective, and computationally efficient framework for few-shot learning. Experimental study on benchmark datasets consistently shows its superiority over the related state-of-the-art, with the largest absolute improvement of 17% over the next best. The source code can be available from https://github.com/WenbinLee/DN4.git.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Deep Nearest Neighbor Neural Network (DN4), a simple, effective, and computationally efficient framework for few-shot learning that not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity.'}",https://arxiv.org/pdf/1903.12290
-few-shot learning via embedding adaptation with set-to-set functions,Han-Jia Ye,"Learning with limited data is a key challenge for visual recognition. Many few-shot learning methods address this challenge by learning an instance embedding function from seen classes and apply the function to instances from unseen classes with limited labels. This style of transfer learning is task-agnostic: the embedding function is not learned optimally discriminative with respect to the unseen classes, where discerning among them leads to the target task. In this paper, we propose a novel approach to adapt the instance embeddings to the target classification task with a set-to-set function, yielding embeddings that are task-specific and are discriminative. We empirically investigated various instantiations of such set-to-set functions and observed the Transformer is most effective --- as it naturally satisfies key properties of our desired model. We denote this model as FEAT (few-shot embedding adaptation w/ Transformer) and validate it on both the standard few-shot classification benchmark and four extended few-shot learning settings with essential use cases, i.e., cross-domain, transductive, generalized few-shot learning, and low-shot learning. It archived consistent improvements over baseline models as well as previous methods, and established the new state-of-the-art results on two benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach to adapt the instance embeddings to the target classification task with a set-to-set function, yielding embeddeddings that are task-specific and are discriminative.'}",https://arxiv.org/pdf/1812.03664
-variational few-shot learning for microservice-oriented intrusion detection in distributed industrial iot,Wei Liang,"Along with the popularity of the Internet of Things (IoT) techniques with several computational paradigms, such as cloud and edge computing, microservice has been viewed as a promising architecture in large-scale application design and deployment. Due to the limited computing ability of edge devices in distributed IoT, only a small scale of data can be used for model training. In addition, most of the machine-learning-based intrusion detection methods are insufficient when dealing with imbalanced dataset under limited computing resources. In this article, we propose an optimized intra/inter-class-structure-based variational few-shot learning (OICS-VFSL) model to overcome a specific out-of-distribution problem in imbalanced learning, and to improve the microservice-oriented intrusion detection in distributed IoT systems. Following a newly designed VFSL framework, an intra/inter-class optimization scheme is developed using reconstructed feature embeddings, in which the intra-class distance is optimized based on the approximation during a variation Bayesian process, while the inter-class distance is optimized based on the maximization of similarities during a feature concatenation process. An intelligent intrusion detection algorithm is, then, introduced to improve the multiclass classification via a nonlinear neural network. Evaluation experiments are conducted using two public datasets to demonstrate the effectiveness of our proposed model, especially in detecting novel attacks with extremely imbalanced data, compared with four baseline methods.","{'model': 'tldr@v2.0.0', 'text': 'An optimized intra/inter-class-structure-based variational few-shot learning (OICS-VFSL) model is proposed to overcome a specific out-of-distribution problem in imbalanced learning, and to improve the microservice-oriented intrusion detection in distributed IoT systems.'}",https://ieeexplore.ieee.org/ielx7/9424/9772431/09551702.pdf
-generalized zero- and few-shot learning via aligned variational autoencoders,Edgar Schönfeld,"Many approaches in generalized zero-shot learning rely on cross-modal mapping between the image feature space and the class embedding space. As labeled images are expensive, one direction is to augment the dataset by generating either images or image features. However, the former misses fine-grained details and the latter requires learning a mapping associated with class embeddings. In this work, we take feature generation one step further and propose a model where a shared latent space of image features and class embeddings is learned by modality-specific aligned variational autoencoders. This leaves us with the required discriminative information about the image and classes in the latent features, on which we train a softmax classifier. The key to our approach is that we align the distributions learned from images and from side-information to construct latent features that contain the essential multi-modal information associated with unseen classes. We evaluate our learned latent features on several benchmark datasets, i.e. CUB, SUN, AWA1 and AWA2, and establish a new state of the art on generalized zero-shot as well as on few-shot learning. Moreover, our results on ImageNet with various zero-shot splits show that our latent features generalize well in large-scale settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a model where a shared latent space of image features and class embeddings is learned by modality-specific aligned variational autoencoders, and align the distributions learned from images and from side-information to construct latent features that contain the essential multi-modal information associated with unseen classes.'}",https://arxiv.org/pdf/1812.01784
-alexatm 20b: few-shot learning using a large-scale multilingual seq2seq model,Saleh Soltan,"In this work, we demonstrate that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling (CLM) tasks, are more efficient few-shot learners than decoder-only models on various tasks. In particular, we train a 20 billion parameter multilingual seq2seq model called Alexa Teacher Model (AlexaTM 20B) and show that it achieves state-of-the-art (SOTA) performance on 1-shot summarization tasks, outperforming a much larger 540B PaLM decoder model. AlexaTM 20B also achieves SOTA in 1-shot machine translation, especially for low-resource languages, across almost all language pairs supported by the model (Arabic, English, French, German, Hindi, Italian, Japanese, Marathi, Portuguese, Spanish, Tamil, and Telugu) on Flores-101 dataset. We also show in zero-shot setting, AlexaTM 20B outperforms GPT3 (175B) on SuperGLUE and SQuADv2 datasets and provides SOTA performance on multilingual tasks such as XNLI, XCOPA, Paws-X, and XWinograd. Overall, our results present a compelling case for seq2seq models as a powerful alternative to decoder-only models for Large-scale Language Model (LLM) training.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling tasks, are more efficient few-shot learners than decoder-only models on various tasks.'}",http://arxiv.org/pdf/2208.01448
-hypertransformer: model generation for supervised and semi-supervised few-shot learning,A. Zhmoginov,"In this work we propose a HyperTransformer, a Transformer-based model for supervised and semi-supervised few-shot learning that generates weights of a convolutional neural network (CNN) directly from support samples. Since the dependence of a small generated CNN model on a specific task is encoded by a high-capacity Transformer model, we effectively decouple the complexity of the large task space from the complexity of individual tasks. Our method is particularly effective for small target CNN architectures where learning a fixed universal task-independent embedding is not optimal and better performance is attained when the information about the task can modulate all model parameters. For larger models we discover that generating the last layer alone allows us to produce competitive or better results than those obtained with state-of-the-art methods while being end-to-end differentiable.","{'model': 'tldr@v2.0.0', 'text': 'A HyperTransformer, a Transformer-based model for supervised and semi-supervised few-shot learning that generates weights of a convolutional neural network directly from support samples is proposed, which is particularly effective for small target CNN architectures.'}",
-hierarchical graph neural networks for few-shot learning,Cen Chen,"Recent graph neural network (GNN) based methods for few-shot learning (FSL) represent the samples of interest as a fully-connected graph and conduct reasoning on the nodes flatly, which ignores the hierarchical correlations among nodes. However, real-world categories may have hierarchical structures, and for FSL, it is important to extract the distinguishing features of the categories from individual samples. To explore this, we propose a novel hierarchical graph neural network (HGNN) for FSL, which consists of three parts, i.e., bottom-up reasoning, top-down reasoning, and skip connections, to enable the efficient learning of multi-level relationships. For the bottom-up reasoning, we design intra-class k-nearest neighbor pooling (intra-class knnPool) and inter-class knnPool layers, to conduct hierarchical learning for both the intra- and inter-class nodes. For the top-down reasoning, we propose to utilize graph unpooling (gUnpool) layers to restore the down-sampled graph into its original size. Skip connections are proposed to fuse multi-level features for the final node classification. The parameters of HGNN are learned by episodic training with the signal of node losses, which aims to train a well-generalizable model for recognizing unseen classes with few labeled data. Experimental results on benchmark datasets have demonstrated that HGNN outperforms other state-of-the-art GNN based methods significantly, for both transductive and non-transductive FSL tasks. The dataset as well as the source code can be downloaded online1","{'model': 'tldr@v2.0.0', 'text': 'A novel hierarchical graph neural network (HGNN) for FSL is proposed, which consists of three parts, i.e., bottom-up reasoning, top-down reasoning, and skip connections, to enable the efficient learning of multi-level relationships.'}",
-free lunch for few-shot learning: distribution calibration,Shuo Yang,"Learning from a limited number of samples is challenging since the learned model can easily become overfitted based on the biased distribution formed by only a few training examples. In this paper, we calibrate the distribution of these few-sample classes by transferring statistics from the classes with sufficient examples, then an adequate number of examples can be sampled from the calibrated distribution to expand the inputs to the classifier. We assume every dimension in the feature representation follows a Gaussian distribution so that the mean and the variance of the distribution can borrow from that of similar classes whose statistics are better estimated with an adequate number of samples. Our method can be built on top of off-the-shelf pretrained feature extractors and classification models without extra parameters. We show that a simple logistic regression classifier trained using the features sampled from our calibrated distribution can outperform the state-of-the-art accuracy on two datasets (~5% improvement on miniImageNet compared to the next best). The visualization of these generated features demonstrates that our calibrated distribution is an accurate estimation.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a simple logistic regression classifier trained using the features sampled from the calibrated distribution can outperform the state-of-the-art accuracy on two datasets (~5% improvement on miniImageNet compared to the next best).'}",
-"a survey of few-shot learning in smart agriculture: developments, applications, and challenges",Jiachen Yang,,"{'model': 'tldr@v2.0.0', 'text': 'This review comprehensively expounds on few-shot learning in smart agriculture, introduces the definition of few-shots learning, four kinds of learning methods, the publicly available datasets for few- shot learning, various applications in smart Agriculture, and the challenges in smart farming in future development.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-022-00866-2
-"a comprehensive survey of few-shot learning: evolution, applications, challenges, and opportunities",Yisheng Song,"Few-shot learning (FSL) has emerged as an effective learning method and shows great potential. Despite the recent creative works in tackling FSL tasks, learning valid information rapidly from just a few or even zero samples remains a serious challenge. In this context, we extensively investigated 200+ FSL papers published in top journals and conferences in the past three years, aiming to present a timely and comprehensive overview of the most recent advances in FSL with a fresh perspective and to provide an impartial comparison of the strengths and weaknesses of existing work. To avoid conceptual confusion, we first elaborate and contrast a set of relevant concepts including few-shot learning, transfer learning, and meta-learning. Then, we inventively extract prior knowledge related to few-shot learning in the form of a pyramid, which summarizes and classifies previous work in detail from the perspective of challenges. Furthermore, to enrich this survey, we present in-depth analysis and insightful discussions of recent advances in each subsection. What is more, taking computer vision as an example, we highlight the important application of FSL, covering various research hotspots. Finally, we conclude the survey with unique insights into technology trends and potential future research opportunities to guide FSL follow-up research.","{'model': 'tldr@v2.0.0', 'text': 'This survey extensively investigated 200+ FSL papers published in top journals and conferences in the past three years to present a timely and comprehensive overview of the most recent advances in FSL with a fresh perspective and to provide an impartial comparison of the strengths and weaknesses of existing work.'}",https://arxiv.org/pdf/2205.06743
-learning from few examples: a summary of approaches to few-shot learning,Archit Parnami,"Few-Shot Learning refers to the problem of learning the underlying pattern in the data just from a few training samples. Requiring a large number of data samples, many deep learning solutions suffer from data hunger and extensively high computation time and resources. Furthermore, data is often not available due to not only the nature of the problem or privacy concerns but also the cost of data preparation. Data collection, preprocessing, and labeling are strenuous human tasks. Therefore, few-shot learning that could drastically reduce the turnaround time of building machine learning applications emerges as a low-cost solution. This survey paper comprises a representative list of recently proposed few-shot learning algorithms. Given the learning dynamics and characteristics, the approaches to few-shot learning problems are discussed in the perspectives of meta-learning, transfer learning, and hybrid approaches (i.e., different variations of the few-shot learning problem).","{'model': 'tldr@v2.0.0', 'text': 'This survey paper comprises a representative list of recently proposed few-shot learning algorithms that could drastically reduce the turnaround time of building machine learning applications emerges as a low-cost solution.'}",http://arxiv.org/pdf/2203.04291
-charting the right manifold: manifold mixup for few-shot learning,Puneet Mangla,"Few-shot learning algorithms aim to learn model parameters capable of adapting to unseen classes with the help of only a few labeled examples. A recent regularization technique - Manifold Mixup focuses on learning a general-purpose representation, robust to small changes in the data distribution. Since the goal of few-shot learning is closely linked to robust representation learning, we study Manifold Mixup in this problem setting. Self-supervised learning is another technique that learns semantically meaningful features, using only the inherent structure of the data. This work investigates the role of learning relevant feature manifold for few-shot tasks using self-supervision and regularization techniques. We observe that regularizing the feature manifold, enriched via self-supervised techniques, with Manifold Mixup significantly improves few-shot learning performance. We show that our proposed method S2M2 beats the current state-of-the-art accuracy on standard few-shot learning datasets like CIFAR-FS, CUB, mini-ImageNet and tiered-ImageNet by 3 − 8%. Through extensive experimentation, we show that the features learned using our approach generalize to complex few-shot evaluation tasks, cross-domain scenarios and are robust against slight changes to data distribution.","{'model': 'tldr@v2.0.0', 'text': 'This work observes that regularizing the feature manifold, enriched via self-supervised techniques, with Manifold Mixup significantly improves few-shot learning performance, and proposes the proposed method S2M2, which beats the current state-of-the-art accuracy on standard few- shot learning datasets.'}",http://raiith.iith.ac.in/5861/1/1907.12087.pdf
-confess: a framework for single source cross-domain few-shot learning,Debasmit Das,"Most current few-shot learning methods train a model from abundantly labeled base category data and then transfer and adapt the model to sparsely labeled novel category data. These methods mostly generalize well on novel categories from the same domain as the base categories but perform poorly for distant domain categories. In this paper, we propose a framework for few-shot learning coined as ConFeSS (Contrastive Learning and Feature Selection System) that tackles large domain shift between base and novel categories. The first step of our framework trains a feature extracting backbone with the contrastive loss on the base category data. Since the contrastive loss does not use supervision, the features can generalize better to distant target domains. For the second step, we train a masking module to select relevant features that are more suited to target domain classification. Finally, a classifier is fine-tuned along with the backbone such that the backbone produces features similar to the relevant ones. To evaluate our framework, we tested it on a recently introduced cross-domain few-shot learning benchmark. Experimental results demonstrate that our framework outperforms all meta-learning approaches and produces competitive results against recent cross-domain methods. Additional analyses are also performed to better understand our framework.","{'model': 'tldr@v2.0.0', 'text': 'A framework for few-shot learning coined as ConFeSS (Contrastive Learning and Feature Selection System) that tackles large domain shift between base and novel categories and outperforms all meta-learning approaches and produces competitive results against recent cross-domain methods is proposed.'}",
-a broader study of cross-domain few-shot learning,Yunhui Guo,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed Broader Study of Cross-Domain Few-Shot Learning (BSCD-FSL) benchmark, consisting of image data from a diverse assortment of image acquisition methods, demonstrates that state-of-art meta- learning methods are surprisingly outperformed by earlier meta-learning approaches, and all meta- Learning methods underperform in relation to simple fine-tuning.'}",https://arxiv.org/pdf/1912.07200
-integrative few-shot learning for classification and segmentation,Dahyun Kang,"We introduce the integrative task of few-shot classification and segmentation (FS-CS) that aims to both classify and segment target objects in a query image when the target classes are given with a few examples. This task combines two conventional few-shot learning problems, few-shot classification and segmentation. FS-CS generalizes them to more realistic episodes with arbitrary image pairs, where each target class may or may not be present in the query. To address the task, we propose the integrative few-shot learning (iFSL) framework for FS-CS, which trains a learner to construct class-wise foreground maps for multi-label classification and pixel-wise segmentation. We also develop an effective iFSL model, attentive squeeze network (ASNet), that leverages deep semantic correlation and global self-attention to produce reliable foreground maps. In experiments, the proposed method shows promising performance on the FS-CS task and also achieves the state of the art on standard few-shot segmentation benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'The integrative few-shot learning (iFSL) framework for FS-CS is proposed, which trains a learner to construct class-wise foreground maps for multi-label classification and pixel-wise segmentation, and an effective iFSL model is developed, attentive squeeze network (ASNet), that leverages deep semantic correlation and global self-attention to produce reliable foreground maps.'}",https://arxiv.org/pdf/2203.15712
-edge-labeling graph neural network for few-shot learning,Jongmin Kim,"In this paper, we propose a novel edge-labeling graph neural network (EGNN), which adapts a deep neural network on the edge-labeling graph, for few-shot learning. The previous graph neural network (GNN) approaches in few-shot learning have been based on the node-labeling framework, which implicitly models the intra-cluster similarity and the inter-cluster dissimilarity. In contrast, the proposed EGNN learns to predict the edge-labels rather than the node-labels on the graph that enables the evolution of an explicit clustering by iteratively updating the edge-labels with direct exploitation of both intra-cluster similarity and the inter-cluster dissimilarity. It is also well suited for performing on various numbers of classes without retraining, and can be easily extended to perform a transductive inference. The parameters of the EGNN are learned by episodic training with an edge-labeling loss to obtain a well-generalizable model for unseen low-data problem. On both of the supervised and semi-supervised few-shot image classification tasks with two benchmark datasets, the proposed EGNN significantly improves the performances over the existing GNNs.","{'model': 'tldr@v2.0.0', 'text': 'A novel edge-labeling graph neural network (EGNN) is proposed, which adapts a deep neural network on the edge- labeling graph, for few-shot learning and significantly improves the performances over the existing GNNs.'}",https://arxiv.org/pdf/1905.01436
-few-shot learning with class-covariance metric for hyperspectral image classification,Bobo Xi,"Recently, embedding and metric-based few-shot learning (FSL) has been introduced into hyperspectral image classification (HSIC) and achieved impressive progress. To further enhance the performance with few labeled samples, we in this paper propose a novel FSL framework for HSIC with a class-covariance metric (CMFSL). Overall, the CMFSL learns global class representations for each training episode by interactively using training samples from the base and novel classes, and a synthesis strategy is employed on the novel classes to avoid overfitting. During the meta-training and meta-testing, the class labels are determined directly using the Mahalanobis distance measurement rather than an extra classifier. Benefiting from the task-adapted class-covariance estimations, the CMFSL can construct more flexible decision boundaries than the commonly used Euclidean metric. Additionally, a lightweight cross-scale convolutional network (LXConvNet) consisting of 3D and 2D convolutions is designed to thoroughly exploit the spectral-spatial information in the high-frequency and low-frequency scales with low computational complexity. Furthermore, we devise a spectral-prior-based refinement module (SPRM) in the initial stage of feature extraction, which cannot only force the network to emphasize the most informative bands while suppressing the useless ones, but also alleviate the effects of the domain shift between the base and novel categories to learn a collaborative embedding mapping. Extensive experiment results on four benchmark data sets demonstrate that the proposed CMFSL can outperform the state-of-the-art methods with few-shot annotated samples.","{'model': 'tldr@v2.0.0', 'text': 'A novel FSL framework for HSIC with a class-covariance metric (CMFSL), which can outperform the state-of-the-art methods with few-shot annotated samples and devise a spectral-prior-based refinement module (SPRM) in the initial stage of feature extraction.'}",
-realistic evaluation of transductive few-shot learning,Olivier Veilleux,"Transductive inference is widely used in few-shot learning, as it leverages the statistics of the unlabeled query set of a few-shot task, typically yielding substantially better performances than its inductive counterpart. The current few-shot benchmarks use perfectly class-balanced tasks at inference. We argue that such an artificial regularity is unrealistic, as it assumes that the marginal label probability of the testing samples is known and fixed to the uniform distribution. In fact, in realistic scenarios, the unlabeled query sets come with arbitrary and unknown label marginals. We introduce and study the effect of arbitrary class distributions within the query sets of few-shot tasks at inference, removing the class-balance artefact. Specifically, we model the marginal probabilities of the classes as Dirichlet-distributed random variables, which yields a principled and realistic sampling within the simplex. This leverages the current few-shot benchmarks, building testing tasks with arbitrary class distributions. We evaluate experimentally state-of-the-art transductive methods over 3 widely used data sets, and observe, surprisingly, substantial performance drops, even below inductive methods in some cases. Furthermore, we propose a generalization of the mutual-information loss, based on $\alpha$-divergences, which can handle effectively class-distribution variations. Empirically, we show that our transductive $\alpha$-divergence optimization outperforms state-of-the-art methods across several data sets, models and few-shot settings. Our code is publicly available at https://github.com/oveilleux/Realistic_Transductive_Few_Shot.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces and study the effect of arbitrary class distributions within the query sets of few-shot tasks at inference, removing the class-balance artefact, and proposes a generalization of the mutual-information loss, based on $\\alpha$-divergences, which can handle effectively class-distribution variations.'}",http://arxiv.org/pdf/2204.11181
-deep cross-domain few-shot learning for hyperspectral image classification,Zhaokui Li,"One of the challenges in hyperspectral image (HSI) classification is that there are limited labeled samples to train a classifier for very high-dimensional data. In practical applications, we often encounter an HSI domain (called target domain) with very few labeled data, while another HSI domain (called source domain) may have enough labeled data. Classes between the two domains may not be the same. This article attempts to use source class data to help classify the target classes, including the same and new unseen classes. To address this classification paradigm, a meta-learning paradigm for few-shot learning (FSL) is usually adopted. However, existing FSL methods do not account for domain shift between source and target domain. To solve the FSL problem under domain shift, a novel deep cross-domain few-shot learning (DCFSL) method is proposed. For the first time, DCFSL tackles FSL and domain adaptation issues in a unified framework. Specifically, a conditional adversarial domain adaptation strategy is utilized to overcome domain shift, which can achieve domain distribution alignment. In addition, FSL is executed in source and target classes at the same time, which can not only discover transferable knowledge in the source classes but also learn a discriminative embedding model to the target classes. Experiments conducted on four public HSI data sets demonstrate that DCFSL outperforms the existing FSL methods and deep learning methods for HSI classification. Our source code is available at https://github.com/Li-ZK/DCFSL-2021.","{'model': 'tldr@v2.0.0', 'text': 'A novel deep cross-domain few-shot learning (DCFSL) method that tackles FSL and domain adaptation issues in a unified framework and demonstrates that DCFSL outperforms the existing FSL methods and deep learning methods for HSI classification.'}",
-a new meta-baseline for few-shot learning,Yinbo Chen,"Meta-learning has become a popular framework for few-shot learning in recent years, with the goal of learning a model from collections of few-shot classification tasks. While more and more novel meta-learning models are being proposed, our research has uncovered simple baselines that have been overlooked. We present a Meta-Baseline method, by pre-training a classifier on all base classes and meta-learning on a nearest-centroid based few-shot classification algorithm, it outperforms recent state-of-the-art methods by a large margin. Why does this simple method work so well? In the meta-learning stage, we observe that a model generalizing better on unseen tasks from base classes can have a decreasing performance on tasks from novel classes, indicating a potential objective discrepancy. We find both pre-training and inheriting a good few-shot classification metric from the pre-trained classifier are important for Meta-Baseline, which potentially helps the model better utilize the pre-trained representations with stronger transferability. Furthermore, we investigate when we need meta-learning in this Meta-Baseline. Our work sets up a new solid benchmark for this field and sheds light on further understanding the phenomenons in the meta-learning framework for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a Meta-Baseline method, by pre-training a classifier on all base classes and meta-learning on a nearest-centroid based few-shot classification algorithm, which outperforms recent state-of-the-art methods by a large margin.'}",
-simpleshot: revisiting nearest-neighbor classification for few-shot learning,Yan Wang,"Few-shot learners aim to recognize new object classes based on a small number of labeled training examples. To prevent overfitting, state-of-the-art few-shot learners use meta-learning on convolutional-network features and perform classification using a nearest-neighbor classifier. This paper studies the accuracy of nearest-neighbor baselines without meta-learning. Surprisingly, we find simple feature transformations suffice to obtain competitive few-shot learning accuracies. For example, we find that a nearest-neighbor classifier used in combination with mean-subtraction and L2-normalization outperforms prior results in three out of five settings on the miniImageNet dataset.","{'model': 'tldr@v2.0.0', 'text': 'Surprisingly, simple feature transformations suffice to obtain competitive few-shot learning accuracies and it is found that a nearest-neighbor classifier used in combination with mean-subtraction and L2-normalization outperforms prior results in three out of five settings on the miniImageNet dataset.'}",
-distributed few-shot learning for intelligent recognition of communication jamming,Mingqian Liu,"Effective recognition of communication jamming is of vital importance in improving wireless communication system’s anti-jamming capability. Motivated by the major challenges that the jamming data sets in wireless communication system are often small and the recognition performance may be poor, we introduce a novel jamming recognition method based on distributed few-shot learning in this paper. Our proposed method employs a distributed recognition architecture to achieve the global optimization of multiple sub-networks by federated learning. It also introduces a dense block structure in the sub-network structure to improve network information flow by the feature multiplexing and configuration bypass to improve resistance to over-fitting. Our key idea is to first obtain the time-frequency diagram, fractional Fourier transform and constellation diagram of the communication jamming signal as the model-agnostic meta-learning network input, and then train the distributed network through federated learning for jamming recognition. Simulation results show that our proposed method leads to excellent recognition performance with a small data set.","{'model': 'tldr@v2.0.0', 'text': 'A novel jamming recognition method based on distributed few-shot learning that employs a distributed recognition architecture to achieve the global optimization of multiple sub-networks by federated learning and introduces a dense block structure in the sub-network structure to improve network information flow.'}",http://repository.essex.ac.uk/33420/1/distributed-few-shot%20learning-intelligent-recognition-communication-jamming-Chen-2021.pdf
-ease: unsupervised discriminant subspace learning for transductive few-shot learning,Hao Zhu,"Few-shot learning (FSL) has received a lot of attention due to its remarkable ability to adapt to novel classes. Although many techniques have been proposed for FSL, they mostly focus on improving FSL backbones. Some works also focus on learning on top of the features generated by these backbones to adapt them to novel classes. We present an unsupErvised discriminAnt Subspace lEarning (EASE) that improves transductive few-shot learning performance by learning a linear projection onto a subspace built from features of the support set and the unlabeled query set in the test time. Specifically, based on the support set and the unlabeled query set, we generate the similarity matrix and the dissimilarity matrix based on the structure prior for the proposed EASE method, which is efficiently solved with SVD. We also introduce conStraIned wAsserstein MEan Shift clustEring (SIAMESE) which extends Sinkhorn K-means by incorporating labeled support samples. SIAMESE works on the features obtained from EASE to estimate class centers and query predictions. On the miniImageNet, tiered-ImageNet, CIFAR-FS, CUB and OpenMIC benchmarks, both steps significantly boost the performance in transductive FSL and semi-supervised FSL.","{'model': 'tldr@v2.0.0', 'text': 'An unsupErvised discriminAnt Subspace lEarning (EASE) that improves transductive few-shot learning performance by learning a linear projection onto a subspace built from features of the support set and the unlabeled query set in the test time is presented.'}",
-crossfit: a few-shot learning challenge for cross-task generalization in nlp,Qinyuan Ye,"Humans can learn a new language task efficiently with only few examples, by leveraging their knowledge obtained when learning prior tasks. In this paper, we explore whether and how such cross-task generalization ability can be acquired, and further applied to build better few-shot learners across diverse NLP tasks. We introduce CrossFit, a problem setup for studying cross-task generalization ability, which standardizes seen/unseen task partitions, data access during different learning stages, and the evaluation protocols. To instantiate different seen/unseen task partitions in CrossFit and facilitate in-depth analysis, we present the NLP Few-shot Gym, a repository of 160 diverse few-shot NLP tasks created from open-access NLP datasets and converted to a unified text-to-text format. Our analysis reveals that the few-shot learning ability on unseen tasks can be improved via an upstream learning stage using a set of seen tasks. We also observe that the selection of upstream learning tasks can significantly influence few-shot performance on unseen tasks, asking further analysis on task similarity and transferability.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the NLP Few-shot Gym, a repository of 160 diverse few-shot NLP tasks created from open-access NLP datasets and converted to a unified text-to-text format, and reveals that the few- shot learning ability on unseen tasks can be improved via an upstream learning stage using a set of seen tasks.'}",https://aclanthology.org/2021.emnlp-main.572.pdf
-few-shot learning with noisy labels,Kevin J Liang,"Few-shot learning (FSL) methods typically assume clean support sets with accurately labeled samples when training on novel classes. This assumption can often be unrealistic: support sets, no matter how small, can still include mislabeled samples. Robustness to label noise is therefore essential for FSL methods to be practical, but this problem surprisingly remains largely unexplored. To address mislabeled samples in FSL settings, we make several technical contributions. (1) We offer simple, yet effective, feature aggregation methods, improving the prototypes used by ProtoNet, a popular FSL technique. (2) We describe a novel Transformer model for Noisy Few-Shot Learning (TraNFS). TraNFS leverages a transformer's attention mechanism to weigh mislabeled versus correct samples. (3) Finally, we extensively test these methods on noisy versions of MinilmageNet and TieredImageNet. Our results show that TraNFS is on-par with leading FSL methods on clean support sets, yet outperforms them, by far, in the presence of label noise.","{'model': 'tldr@v2.0.0', 'text': ""This work offers simple, yet effective, feature aggregation methods, improving the prototypes used by ProtoNet, a popular FSL technique, and describes a novel Transformer model for Noisy Few-Shot Learning (TraNFS), which leverages a transformer's attention mechanism to weigh mislabeled versus correct samples.""}",https://arxiv.org/pdf/2204.05494
-how to prompt? opportunities and challenges of zero- and few-shot learning for human-ai interaction in creative applications of generative models,Hai Dang,"Deep generative models have the potential to fundamentally change the way we create high-fidelity digital content but are often hard to control. Prompting a generative model is a promising recent development that in principle enables end-users to creatively leverage zero-shot and few-shot learning to assign new tasks to an AI ad-hoc, simply by writing them down. However, for the majority of end-users writing effective prompts is currently largely a trial and error process. To address this, we discuss the key opportunities and challenges for interactive creative applications that use prompting as a new paradigm for Human-AI interaction. Based on our analysis, we propose four design goals for user interfaces that support prompting. We illustrate these with concrete UI design sketches, focusing on the use case of creative writing. The research community in HCI and AI can take these as starting points to develop adequate user interfaces for models capable of zero- and few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'The research community in HCI and AI can take these as starting points to develop adequate user interfaces for models capable of zero- and few-shot learning, and propose four design goals for user interfaces that support prompting.'}",http://arxiv.org/pdf/2209.01390
-learning to capture the query distribution for few-shot learning,Ziqiu Chi,"In the Few-Shot Learning (FSL), much of the related efforts only rely on the few available labeled samples (support set) building approach. However, the challenge is that the support set is easy-to-be-biased, so that they cannot be competent prototypes and are hard to represent the class distribution, leading to performance bottlenecks. In this paper, we propose to solve this obstacle by capturing the distribution of the unlabeled samples (query set). We propose two sampling methods: DeepSearch ( $\cal DS$ ) and WideSearch ( $\cal WS$ ). Both approaches are simple to implement and have no trainable parameters. They search the query samples near to the support set in different manners. Afterward, the statistic information is calculated, and we generate the latent samples according to it. The generated latent set is promising. First, it brings the query set distribution information to the classifier, which significantly improves the performance of the cross-entropy-based classifier. Second, it helps the support set become the better prototypes, which boosts the performance of the prototype-based classifier. Third, we find few latent samples are enough to boost the performance. Abundant experiments prove the proposed method achieves state-of-the-art performance on the few-shot tasks. Finally, rich ablation studies explain the compelling details of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes two sampling methods that bring the query set distribution information to the classifier, which significantly improves the performance of the cross-entropy-based classifier and few latent samples are enough to boost the performance.'}",
-multi-level second-order few-shot learning,Hongguang Zhang,"We propose a Multi-level Second-order (MlSo) few-shot learning network for supervised or unsupervised few-shot image classification and few-shot action recognition. We leverage so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and we use self-supervised discriminating mechanisms. As Second-order Pooling (SoP) is popular in image recognition, we employ its basic element-wise variant in our pipeline. The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning. As SoP can handle convolutional feature maps of varying spatial sizes, we also introduce image inputs at multiple spatial scales into MlSo. To exploit the discriminative information from multi-level and multi-scale features, we develop a Feature Matching (FM) module that reweights their respective branches. We also introduce a self-supervised step, which is a discriminator of the spatial level and the scale of abstraction. Our pipeline is trained in an end-to-end manner. With a simple architecture, we demonstrate respectable results on standard datasets such as Omniglot, mini–ImageNet, tiered–ImageNet, Open MIC, fine-grained datasets such as CUB Birds, Stanford Dogs and Cars, and action recognition datasets such as HMDB51, UCF101, and mini–MIT.","{'model': 'tldr@v2.0.0', 'text': 'A Multi-level Second-order (MlSo) few-shots learning network for supervised or unsupervised few-shot image classification and few- shot action recognition is proposed, leveraging so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and self-supervised discriminating mechanisms are used.'}",https://arxiv.org/pdf/2201.05916
-heterogeneous few-shot learning for hyperspectral image classification,Yan Wang,"Deep learning has achieved great success in hyperspectral image (HSI) classification. However, its success relies on the availability of sufficient training samples. Unfortunately, the collection of training samples is expensive, time-consuming, and even impossible in some cases. Natural image datasets that are different from HSI, such as Image Net and mini-ImageNet, have abundant texture and structure information. Effective knowledge transfer between two heterogeneous datasets can significantly improve the accuracy of HSI classification. In this letter, heterogeneous few-shot learning (HFSL) for HSI classification is proposed with only a few labeled samples per class. First, few-shot learning is performed on the mini-ImageNet datasets to learn the transferable knowledge. Then, to make full use of the spatial and spectral information, a spectral–spatial fusion network is devised. Spectral information is obtained by the residual network with pure 1-D operators. Spatial information is extracted by a convolution network with pure 2-D operators, and the weights of the spatial network are initialized by the weights of the model trained on the mini-ImageNet datasets. Finally, few-shot learning is fine-tuned on HSI to extract discriminative spectral–spatial features and individual knowledge, which can improve the classification performance of the new classification task. Experiments conducted on two public HSI datasets demonstrate that the HFSL outperforms the existing few-shot learning methods and supervised learning methods for HSI classification with only a few labeled samples. Our source code is available at https://github.com/Li-ZK/HFSL.","{'model': 'tldr@v2.0.0', 'text': 'Experiments conducted on two public HSI datasets demonstrate that the HFSL outperforms the existing few-shot learning methods and supervised learning methods for HSI classification with only a few labeled samples.'}",
-amcrn: few-shot learning for automatic modulation classification,Quan Zhou,"Deep learning (DL) has been widely applied in automatic modulation classification (AMC), while the superb performance highly depends on high-quality datasets. Motivated by this, the AMC under few-shot conditions is considered in this letter, where a novel network architecture is proposed, namely automatic modulation classification relation network (AMCRN), and verified with the baseline methods. Experimental results state that the accuracy of proposed AMCRN exceeds 90% and 10% to 50% improvements are obtained compared with classical schemes when the signal-to-noise ratio (SNR) is greater than −2 dB.","{'model': 'tldr@v2.0.0', 'text': 'The AMC under few-shot conditions is considered, where a novel network architecture is proposed, namely automatic modulation classification relation network (AMCRN), and verified with the baseline methods.'}",
-exploring complementary strengths of invariant and equivariant representations for few-shot learning,Mamshad Nayeem Rizve,"In many real-world problems, collecting a large number of labeled samples is infeasible. Few-shot learning (FSL) is the dominant approach to address this issue, where the objective is to quickly adapt to novel categories in presence of a limited number of samples. FSL tasks have been predominantly solved by leveraging the ideas from gradient-based meta-learning and metric learning approaches. However, recent works have demonstrated the significance of powerful feature representations with a simple embedding network that can outperform existing sophisticated FSL algorithms. In this work, we build on this insight and propose a novel training mechanism that simultaneously enforces equivariance and invariance to a general set of geometric transformations. Equivariance or invariance has been employed standalone in the previous works; however, to the best of our knowledge, they have not been used jointly. Simultaneous optimization for both of these contrasting objectives allows the model to jointly learn features that are not only independent of the input transformation but also the features that encode the structure of geometric transformations. These complementary sets of features help generalize well to novel classes with only a few data samples. We achieve additional improvements by incorporating a novel self-supervised distillation objective. Our extensive experimentation shows that even without knowledge distillation our proposed method can outperform current state-of-the-art FSL methods on five popular benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel training mechanism that simultaneously enforces equivariance and invariance to a general set of geometric transformations and shows that even without knowledge distillation this proposed method can outperform current state-of-the-art FSL methods on five popular benchmark datasets.'}",https://arxiv.org/pdf/2103.01315
-few-shot learning with multilingual generative language models,Xi Victoria Lin,"Large-scale generative language models such as GPT-3 are competitive few-shot learners. While these models are known to be able to jointly represent many different languages, their training data is dominated by English, potentially limiting their cross-lingual generalization. In this work, we train multilingual generative language models on a corpus covering a diverse set of languages, and study their few- and zero-shot learning capabilities in a wide range of tasks. Our largest model with 7.5 billion parameters sets new state of the art in few-shot learning in more than 20 representative languages, outperforming GPT-3 of comparable size in multilingual commonsense reasoning (with +7.4% absolute accuracy improvement in 0-shot settings and +9.4% in 4-shot settings) and natural language inference (+5.4% in each of 0-shot and 4-shot settings). On the FLORES-101 machine translation benchmark, our model outperforms GPT-3 on 171 out of 182 directions with 32 training examples, while surpassing the official supervised baseline in 45 directions. We conduct an in-depth analysis of different multilingual prompting approaches, showing in particular that strong few-shot learning performance across languages can be achieved via cross-lingual transfer through both templates and demonstration examples.","{'model': 'tldr@v2.0.0', 'text': 'This work trains multilingual generative language models on a corpus covering a diverse set of languages, and shows in particular that strong few-shot learning performance across languages can be achieved via cross-lingual transfer through both templates and demonstration examples.'}",https://aclanthology.org/2022.emnlp-main.616.pdf
-adaptive subspaces for few-shot learning,Christian Simon,"Object recognition requires a generalization capability to avoid overfitting, especially when the samples are extremely few. Generalization from limited samples, usually studied under the umbrella of meta-learning, equips learning techniques with the ability to adapt quickly in dynamical environments and proves to be an essential aspect of life long learning. In this paper, we provide a framework for few-shot learning by introducing dynamic classifiers that are constructed from few samples. A subspace method is exploited as the central block of a dynamic classifier. We will empirically show that such modelling leads to robustness against perturbations (e.g., outliers) and yields competitive results on the task of supervised and semi-supervised few-shot classification. We also develop a discriminative form which can boost the accuracy even further. Our code is available at https://github.com/chrysts/dsn_fewshot","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a framework for few-shot learning by introducing dynamic classifiers that are constructed from few samples and empirically shows that such modelling leads to robustness against perturbations and yields competitive results on the task of supervised and semi-supervised few- shot classification.'}",
-siamese neural network based few-shot learning for anomaly detection in industrial cyber-physical systems,Xiaokang Zhou,"With the increasing population of Industry 4.0, both AI and smart techniques have been applied and become hotly discussed topics in industrial cyber-physical systems (CPS). Intelligent anomaly detection for identifying cyber-physical attacks to guarantee the work efficiency and safety is still a challenging issue, especially when dealing with few labeled data for cyber-physical security protection. In this article, we propose a few-shot learning model with Siamese convolutional neural network (FSL-SCNN), to alleviate the over-fitting issue and enhance the accuracy for intelligent anomaly detection in industrial CPS. A Siamese CNN encoding network is constructed to measure distances of input samples based on their optimized feature representations. A robust cost function design including three specific losses is then proposed to enhance the efficiency of training process. An intelligent anomaly detection algorithm is developed finally. Experiment results based on a fully labeled public dataset and a few labeled dataset demonstrate that our proposed FSL-SCNN can significantly improve false alarm rate (FAR) and F1 scores when detecting intrusion signals for industrial CPS security protection.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning model with Siamese convolutional neural network (FSL-SCNN) is proposed, to alleviate the over-fitting issue and enhance the accuracy for intelligent anomaly detection in industrial CPS.'}",https://ieeexplore.ieee.org/ielx7/9424/9422988/09311786.pdf
-tuning language models as training data generators for augmentation-enhanced few-shot learning,Yu Meng,"Recent studies have revealed the intriguing few-shot learning ability of pretrained language models (PLMs): They can quickly adapt to a new task when fine-tuned on a small amount of labeled data formulated as prompts, without requiring abundant task-specific annotations. Despite their promising performance, most existing few-shot approaches that only learn from the small training set still underperform fully supervised training by nontrivial margins. In this work, we study few-shot learning with PLMs from a different perspective: We first tune an autoregressive PLM on the few-shot samples and then use it as a generator to synthesize a large amount of novel training samples which augment the original training set. To encourage the generator to produce label-discriminative samples, we train it via weighted maximum likelihood where the weight of each token is automatically adjusted based on a discriminative meta-learning objective. A classification PLM can then be fine-tuned on both the few-shot and the synthetic samples with regularization for better generalization and stability. Our approach FewGen achieves an overall better result across seven classification tasks of the GLUE benchmark than existing few-shot learning methods, improving no-augmentation methods by 5+ average points, and outperforming augmentation methods by 3+ average points.","{'model': 'tldr@v2.0.0', 'text': 'This work first tunes an autoregressive PLM on the few-shot samples and then uses it as a generator to synthesize a large amount of novel training samples which augment the original training set, achieving an overall better result across seven classification tasks of the GLUE benchmark than existing few- shot learning methods.'}",http://arxiv.org/pdf/2211.03044
-"a neural network solves, explains, and generates university math problems by program synthesis and few-shot learning at human level",Iddo Drori,"Significance We demonstrate that a neural network automatically solves, explains, and generates university-level problems from the largest Massachusetts Institute of Technology (MIT) mathematics courses at a human level. Our methods combine three innovations: 1) using recent neural networks pretrained on text and fine-tuned on code rather than pretrained on text; 2) few-shot learning synthesizing programs that correctly solve course problems automatically; and 3) a pipeline to solve questions, explain solutions, and generate new questions indistinguishable by students from course questions. Our work solves university-level mathematics courses and improves upon state-of-the-art, increasing automatic accuracy on randomly sampled questions on a benchmark by order of magnitude. Implications for higher education include roles of artificial intelligence (AI) in automated course evaluation and content generation.","{'model': 'tldr@v2.0.0', 'text': 'This work solves university-level mathematics courses and improves upon state-of-the-art, increasing automatic accuracy on randomly sampled questions on a benchmark by order of magnitude.'}",
-"few-shot learning via learning the representation, provably",S. Du,"This paper studies few-shot learning via representation learning, where one uses $T$ source tasks with $n_1$ data per task to learn a representation in order to reduce the sample complexity of a target task for which there is only $n_2 (\ll n_1)$ data. Specifically, we focus on the setting where there exists a good \emph{common representation} between source and target, and our goal is to understand how much of a sample size reduction is possible. First, we study the setting where this common representation is low-dimensional and provide a fast rate of $O\left(\frac{\mathcal{C}\left(\Phi\right)}{n_1T} + \frac{k}{n_2}\right)$; here, $\Phi$ is the representation function class, $\mathcal{C}\left(\Phi\right)$ is its complexity measure, and $k$ is the dimension of the representation. When specialized to linear representation functions, this rate becomes $O\left(\frac{dk}{n_1T} + \frac{k}{n_2}\right)$ where $d (\gg k)$ is the ambient input dimension, which is a substantial improvement over the rate without using representation learning, i.e. over the rate of $O\left(\frac{d}{n_2}\right)$. Second, we consider the setting where the common representation may be high-dimensional but is capacity-constrained (say in norm); here, we again demonstrate the advantage of representation learning in both high-dimensional linear regression and neural network learning. Our results demonstrate representation learning can fully utilize all $n_1T$ samples from source tasks.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate representation learning can fully utilize all $n_1T$ samples from source tasks and the advantage of representation learning in both high-dimensional linear regression and neural network learning.'}",
-research progress on few-shot learning for remote sensing image interpretation,Xian Sun,"The rapid development of deep learning brings effective solutions for remote sensing image interpretation. Training deep neural network models usually require a large number of manually labeled samples. However, there is a limitation to obtain sufficient labeled samples in remote sensing field to satisfy the data requirement. Therefore, it is of great significance to conduct the research on few-shot learning for remote sensing image interpretation. First, this article provides a bibliometric analysis of the existing works for remote sensing interpretation related to few-shot learning. Second, two categories of few-shot learning methods, i.e., the data-augmentation-based and the prior-knowledge-based, are introduced for the interpretation of remote sensing images. Then, three typical remote sensing interpretation applications are listed, including scene classification, semantic segmentation, and object detection, together with the corresponding public datasets and the evaluation criteria. Finally, the research status is summarized, and some possible research directions are provided. This article gives a reference for scholars working on few-shot learning research in the remote sensing field.","{'model': 'tldr@v2.0.0', 'text': 'A bibliometric analysis of the existing works for remote sensing interpretation related to few-shot learning and two categories of few- shot learning methods, i.e., the data-augmentation-based and the prior-knowledge-based, are introduced for the interpretation of remote sensing images.'}",https://ieeexplore.ieee.org/ielx7/4609443/9314330/09328476.pdf
-few-shot learning creates predictive models of drug response that translate from high-throughput screens to individual patients,Jianzhu Ma,,"{'model': 'tldr@v2.0.0', 'text': 'A recently developed technique, few-shot machine learning, is applied to train a versatile neural network model in cell lines that can be tuned to new contexts using few additional samples, and it quickly adapts when switching among different tissue types and in moving from cell-line models to clinical contexts, including patient- derived tumor cells and patient-derived xenografts.'}",
-overcoming catastrophic forgetting in incremental few-shot learning by finding flat minima,Guangyuan Shi,"This paper considers incremental few-shot learning, which requires a model to continually recognize new categories with only a few examples provided. Our study shows that existing methods severely suffer from catastrophic forgetting, a well-known problem in incremental learning, which is aggravated due to data scarcity and imbalance in the few-shot setting. Our analysis further suggests that to prevent catastrophic forgetting, actions need to be taken in the primitive stage -- the training of base classes instead of later few-shot learning sessions. Therefore, we propose to search for flat local minima of the base training objective function and then fine-tune the model parameters within the flat region on new tasks. In this way, the model can efficiently learn new classes while preserving the old ones. Comprehensive experimental results demonstrate that our approach outperforms all prior state-of-the-art methods and is very close to the approximate upper bound. The source code is available at https://github.com/moukamisama/F2M.","{'model': 'tldr@v2.0.0', 'text': 'This paper considers incremental few-shot learning, which requires a model to continually recognize new categories with only a few examples provided, and proposes to search for flat local minima of the base training objective function and then fine-tune the model parameters within the flat region on new tasks.'}",
-learning dynamic alignment via meta-filter for few-shot learning,C. Xu,"Few-shot learning (FSL), which aims to recognise new classes by adapting the learned knowledge with extremely limited few-shot (support) examples, remains an important open problem in computer vision. Most of the existing methods for feature alignment in few-shot learning only consider image-level or spatial-level alignment while omitting the channel disparity. Our insight is that these methods would lead to poor adaptation with redundant matching, and leveraging channel-wise adjustment is the key to well adapting the learned knowledge to new classes. Therefore, in this paper, we propose to learn a dynamic alignment, which can effectively highlight both query regions and channels according to different local support information. Specifically, this is achieved by first dynamically sampling the neighbourhood of the feature position conditioned on the input few shot, based on which we further predict a both position-dependent and channel-dependent Dynamic Meta-filter. The filter is used to align the query feature with position-specific and channel-specific knowledge. Moreover, we adopt Neural Ordinary Differential Equation (ODE) to enable a more accurate control of the alignment. In such a sense our model is able to better capture fine-grained semantic context of the few-shot example and thus facilitates dynamical knowledge adaptation for few-shot learning. The resulting framework establishes the new state-of-the-arts on major few-shot visual recognition benchmarks, including miniImageNet and tieredImageNet.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to learn a dynamic alignment, which can effectively highlight both query regions and channels according to different local support information and facilitates dynamical knowledge adaptation for few-shot learning.'}",https://arxiv.org/pdf/2103.13582
-multi-scale metric learning for few-shot learning,Wen Jiang,"Few-shot learning in image classification is developed to learn a model that aims to identify unseen classes with only few training samples for each class. Fewer training samples and new tasks of classification make many traditional classification models no longer applicable. In this paper, a novel few-shot learning method named multi-scale metric learning (MSML) is proposed to extract multi-scale features and learn the multi-scale relations between samples for the classification of few-shot learning. In the proposed method, a feature pyramid structure is introduced for multi-scale feature embedding, which aims to combine high-level strong semantic features with low-level but abundant visual features. Then a multi-scale relation generation network (MRGN) is developed for hierarchical metric learning, in which high-level features are corresponding to deeper metric learning while low-level features are corresponding to lighter metric learning. Moreover, a novel loss function named intra-class and inter-class relation loss (IIRL) is proposed to optimize the proposed deep network, which aims to strengthen the correlation between homogeneous groups of samples and weaken the correlation between heterogeneous groups of samples. Experimental results on mini ImageNet and tiered ImageNet demonstrate that the proposed method achieves superior performance in few-shot learning problem.","{'model': 'tldr@v2.0.0', 'text': 'A novel few-shot learning method named multi-scale metric learning (MSML) is proposed to extract multi- Scale features and learn the multi- scale relations between samples for the classification of few- shot learning.'}",
-attribute surrogates learning and spectral tokens pooling in transformers for few-shot learning,Yang He,"This paper presents new hierarchically cascaded transformers that can improve data efficiency through attribute surrogates learning and spectral tokens pooling. Vision transformers have recently been thought of as a promising alternative to convolutional neural networks for visual recognition. But when there is no sufficient data, it gets stuck in overfitting and shows inferior performance. To improve data efficiency, we propose hierarchically cascaded transformers that exploit intrinsic image structures through spectral tokens pooling and optimize the learnable parameters through latent attribute surrogates. The intrinsic image structure is utilized to reduce the ambiguity between foreground content and background noise by spectral tokens pooling. And the attribute surrogate learning scheme is designed to benefit from the rich visual information in image-label pairs instead of simple visual concepts assigned by their labels. Our Hierarchically Cascaded Transformers, called HCTransformers, is built upon a self-supervised learning framework DINO and is tested on several popular few-shot learning benchmarks. In the inductive setting, HCTransformers surpass the DINO baseline by a large margin of 9.7% 5-way 1-shot accuracy and 9.17% 5-way 5-shot accuracy on miniImageNet, which demonstrates HCTransformers are efficient to extract discriminative features. Also, HCTransformers show clear advantages over SOTA few-shot classification methods in both 5-way 1-shot and 5-way 5-shot settings on four popular benchmark datasets, including miniImageNet, tieredImageNet, FC100, and CIFAR-FS. The trained weights and codes are available at https://github.com/StomachCold/HCTransformers.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents new hierarchically cascaded transformers that can improve data efficiency through attribute surrogates learning and spectral tokens pooling and shows clear advantages over SOTA few-shot classification methods in both 5- way 1-shot and 5-way 5-shot settings on four popular benchmark datasets.'}",https://arxiv.org/pdf/2203.09064
-semi-supervised few-shot learning approach for plant diseases recognition,Yang Li,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed semi-supervised few-shot learning methods can outperform other related works with fewer labeled training data and adopt the confidence interval to determine the number of unlabeled samples for pseudo-labelling adaptively.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-021-00770-1
-partial is better than all: revisiting fine-tuning strategy for few-shot learning,Zhiqiang Shen,"The goal of few-shot learning is to learn a classifier that can recognize unseen classes from limited support data with labels. A common practice for this task is to train a model on the base set first and then transfer to novel classes through fine-tuning or meta-learning. However, as the base classes have no overlap to the novel set, simply transferring whole knowledge from base data is not an optimal solution since some knowledge in the base model may be biased or even harmful to the novel class. In this paper, we propose to transfer partial knowledge by freezing or fine-tuning particular layer(s) in the base model. Specifically, layers will be imposed different learning rates if they are chosen to be fine-tuned, to control the extent of preserved transferability. To determine which layers to be recast and what values of learning rates for them, we introduce an evolutionary search based method that is efficient to simultaneously locate the target layers and determine their individual learning rates. We conduct extensive experiments on CUB and mini-ImageNet to demonstrate the effectiveness of our proposed method. It achieves the state-of-the-art performance on both meta-learning and non-meta based frameworks. Furthermore, we extend our method to the conventional pre-training + fine-tuning paradigm and obtain consistent improvement.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to transfer partial knowledge by freezing or fine-tuning particular layer(s) in the base model to control the extent of preserved transferability, and introduces an evolutionary search based method that is efficient to simultaneously locate the target layers and determine their individual learning rates.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17155/16962
-attentional constellation nets for few-shot learning,Weijian Xu,"The success of deep convolutional neural networks builds on top of the learning of effective convolution operations, capturing a hierarchy of structured features via filtering, activation, and pooling. However, the explicit structured features, e.g. object parts, are not expressive in the existing CNN frameworks. In this paper, we tackle the few-shot learning problem and make an effort to enhance structured features by expanding CNNs with a constellation model, which performs cell feature clustering and encoding with a dense part representation; the relationships among the cell features are further modeled by an attention mechanism. With the additional constellation branch to increase the awareness of object parts, our method is able to attain the advantages of the CNNs while making the overall internal representations more robust in the few-shot learning setting. Our approach attains a significant improvement over the existing methods in few-shot learning on the CIFAR-FS, FC100, and mini-ImageNet benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper tackles the few-shot learning problem and makes an effort to enhance structured features by expanding CNNs with a constellation model, which performs cell feature clustering and encoding with a dense part representation; the relationships among the cell features are further modeled by an attention mechanism.'}",
-few-shot learning on graphs: a survey,Chuxu Zhang,"Graph representation learning has attracted tremendous attention due to its remarkable performance in many real-world applications. However, prevailing supervised graph representation learning models for specific tasks often suffer from label sparsity issue as data labeling is always time and resource consuming. In light of this, few-shot learning on graphs (FSLG), which combines the strengths of graph representation learning and few-shot learning together, has been proposed to tackle the performance degradation in face of limited annotated data challenge. There have been many studies working on FSLG recently. In this paper, we comprehensively survey these work in the form of a series of methods and applications. Specifically, we first introduce FSLG challenges and bases, then categorize and summarize existing work of FSLG in terms of three major graph mining tasks at different granularity levels, i.e., node, edge, and graph. Finally, we share our thoughts on some future research directions of FSLG. The authors of this survey have contributed significantly to the AI literature on FSLG over the last few years.","{'model': 'tldr@v2.0.0', 'text': 'This paper categorize and summarize existing work of FSLG in terms of three major graph mining tasks at different granularity levels, i.e., node, edge, and graph.'}",https://arxiv.org/pdf/2203.09308
-incremental few-shot learning via vector quantization in deep embedded space,Kuilin Chen,"The capability of incrementally learning new tasks without forgetting old ones is a challenging problem due to catastrophic forgetting. This challenge becomes greater when novel tasks contain very few labelled training samples. Currently, most methods are dedicated to class-incremental learning and rely on sufficient training data to learn additional weights for newly added classes. Those methods cannot be easily extended to incremental regression tasks and could suffer from severe overfitting when learning few-shot novel tasks. In this study, we propose a nonparametric method in deep embedded space to tackle incremental few-shot learning problems. The knowledge about the learned tasks is compressed into a small number of quantized reference vectors. The proposed method learns new tasks sequentially by adding more reference vectors to the model using few-shot samples in each novel task. For classification problems, we employ the nearest neighbor scheme to make classification on sparsely available data and incorporate intra-class variation, less forgetting regularization and calibration of reference vectors to mitigate catastrophic forgetting. In addition, the proposed learning vector quantization (LVQ) in deep embedded space can be customized as a kernel smoother to handle incremental few-shot regression tasks. Experimental results demonstrate that the proposed method outperforms other state-of-the-art methods in incremental learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed learning vector quantization in deep embedded space can be customized as a kernel smoother to handle incremental few-shot regression tasks and outperforms other state-of-the-art methods in incremental learning.'}",
-deep few-shot learning for hyperspectral image classification,Bing Liu,"Deep learning methods have recently been successfully explored for hyperspectral image (HSI) classification. However, training a deep-learning classifier notoriously requires hundreds or thousands of labeled samples. In this paper, a deep few-shot learning method is proposed to address the small sample size problem of HSI classification. There are three novel strategies in the proposed algorithm. First, spectral–spatial features are extracted to reduce the labeling uncertainty via a deep residual 3-D convolutional neural network. Second, the network is trained by episodes to learn a metric space where samples from the same class are close and those from different classes are far. Finally, the testing samples are classified by a nearest neighbor classifier in the learned metric space. The key idea is that the designed network learns a metric space from the training data set. Furthermore, such metric space could generalize to the classes of the testing data set. Note that the classes of the testing data set are not seen in the training data set. Four widely used HSI data sets were used to assess the performance of the proposed algorithm. The experimental results indicate that the proposed method can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.","{'model': 'tldr@v2.0.0', 'text': 'A deep few-shot learning method is proposed to address the small sample size problem of HSI classification and can achieve better classification accuracy than the conventional semisupervised methods with only a few labeled samples.'}",
-boil: towards representation change for few-shot learning,Jaehoon Oh,,,
-adaptive cross-modal few-shot learning,Chen Xing,"Metric-based meta-learning techniques have successfully been applied to few-shot classification problems. In this paper, we propose to leverage cross-modal information to enhance metric-based few-shot learning methods. Visual and semantic feature spaces have different structures by definition. For certain concepts, visual features might be richer and more discriminative than text ones. While for others, the inverse might be true. Moreover, when the support from visual information is limited in image classification, semantic representations (learned from unsupervised text corpora) can provide strong prior knowledge and context to help learning. Based on these two intuitions, we propose a mechanism that can adaptively combine information from both modalities according to new image categories to be learned. Through a series of experiments, we show that by this adaptive combination of the two modalities, our model outperforms current uni-modality few-shot learning methods and modality-alignment methods by a large margin on all benchmarks and few-shot scenarios tested. Experiments also show that our model can effectively adjust its focus on the two modalities. The improvement in performance is particularly large when the number of shots is very small.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a mechanism that can adaptively combine information from both modalities according to new image categories to be learned and shows that by this adaptive combination of the two modalities, this model outperforms current uni-modality few-shot learning methods and modality-alignment methods by a large margin on all benchmarks and few- shot scenarios tested.'}",
-meta-baseline: exploring simple meta-learning for few-shot learning,Yinbo Chen,"Meta-learning has been the most common framework for few-shot learning in recent years. It learns the model from collections of few-shot classification tasks, which is believed to have a key advantage of making the training objective consistent with the testing objective. However, some recent works report that by training for whole-classification, i.e. classification on the whole label-set, it can get comparable or even better embedding than many meta-learning algorithms. The edge between these two lines of works has yet been underexplored, and the effectiveness of meta-learning in few-shot learning remains unclear. In this paper, we explore a simple process: meta-learning over a whole-classification pre-trained model on its evaluation metric. We observe this simple method achieves competitive performance to state-of-the-art methods on standard bench-marks. Our further analysis shed some light on understanding the trade-offs between the meta-learning objective and the whole-classification objective in few-shot learning. Our code is available at https://github.com/yinboc/few-shot-meta-baseline.","{'model': 'tldr@v2.0.0', 'text': 'A simple process: meta-learning over a whole-classification pre-trained model on its evaluation metric achieves competitive performance to state-of-the-art methods on standard bench-marks and sheds some light on understanding the trade-offs between the meta- learning objective and the whole- classification objective in few-shot learning.'}",https://arxiv.org/pdf/2003.04390
-dpgn: distribution propagation graph network for few-shot learning,Ling Yang,"Most graph-network-based meta-learning approaches model instance-level relation of examples. We extend this idea further to explicitly model the distribution-level relation of one example to all other examples in a 1-vs-N manner. We propose a novel approach named distribution propagation graph network (DPGN) for few-shot learning. It conveys both the distribution-level relations and instance-level relations in each few-shot learning task. To combine the distribution-level relations and instance-level relations for all examples, we construct a dual complete graph network which consists of a point graph and a distribution graph with each node standing for an example. Equipped with dual graph architecture, DPGN propagates label information from labeled examples to unlabeled examples within several update generations. In extensive experiments on few-shot learning benchmarks, DPGN outperforms state-of-the-art results by a large margin in 5%∼12% under supervised setting and 7%∼13% under semi-supervised setting. Code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach named distribution propagation graph network (DPGN) for few-shot learning, which conveys both the distribution-level relations and instance- level relations in each few- shot learning task.'}",https://arxiv.org/pdf/2003.14247
-bridging the gap between few-shot and many-shot learning via distribution calibration,Shuo Yang,"A major gap between few-shot and many-shot learning is the data distribution empirically oserved by the model during training. In few-shot learning, the learned model can easily become over-fitted based on the biased distribution formed by only a few training examples, while the ground-truth data distribution is more accurately uncovered in many-shot learning to learn a well-generalized model. In this paper, we propose to calibrate the distribution of these few-sample classes to be more unbiased to alleviate such an over-fitting problem. The distribution calibration is achieved by transferring statistics from the classes with sufficient examples to those few-sample classes. After calibration, an adequate number of examples can be sampled from the calibrated distribution to expand the inputs to the classifier. Specifically, we assume every dimension in the feature representation from the same class follows a Gaussian distribution so that the mean and the variance of the distribution can borrow from that of similar classes whose statistics are better estimated with an adequate number of samples. Extensive experiments on three datasets, miniImageNet, tieredImageNet, and CUB, show that a simple linear classifier trained using the features sampled from our calibrated distribution can outperform the state-of-the-art accuracy by a large margin. Besides the favorable performance, the proposed method also exhibits high flexibility by showing consistent accuracy improvement when it is built on top of any off-the-shelf pretrained feature extractors and classification models without extra learnable parameters. The visualization of these generated features demonstrates that our calibrated distribution is an accurate estimation thus the generalization ability gain is convincing. We also establish a generalization error bound for the proposed distribution-calibration-based few-shot learning, which consists of the distribution assumption error, the distribution approximation error, and the estimation error. This generalization error bound theoretically justifies the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'A generalization error bound is established for the proposed distribution-calibration-based few-shot learning, which consists of the distribution assumption error, the distribution approximation error, and the estimation error that theoretically justifies the effectiveness of the proposed method.'}",https://www.techrxiv.org/articles/preprint/Bridging_the_Gap_between_Few-Shot_and_Many-Shot_Learning_via_Distribution_Calibration/14380697/files/27474479.pdf
-rectifying the shortcut learning of background for few-shot learning,Xu Luo,"The category gap between training and evaluation has been characterised as one of the main obstacles to the success of Few-Shot Learning (FSL). In this paper, we for the first time empirically identify image background, common in realistic images, as a shortcut knowledge helpful for in-class classification but ungeneralizable beyond training categories in FSL. A novel framework, COSOC, is designed to tackle this problem by extracting foreground objects in images at both training and evaluation without any extra supervision. Extensive experiments carried on inductive FSL tasks demonstrate the effectiveness of our approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper for the first time empirically identify image background, common in realistic images, as a shortcut knowledge helpful for in-class classification but ungeneralizable beyond training categories in FSL.'}",
-cross-domain few-shot learning with task-specific adapters,Wei-Hong Li,"In this paper, we look at the problem of cross-domain few-shot classification that aims to learn a classifier from previously unseen classes and domains withfew labeled samples. Recent approaches broadly solve this problem by pa-rameterizing their few-shot classifiers with task-agnostic and task-specific weights where the former is typically learned on a large training set and the latter is dynamically predicted through an auxiliary network conditioned on a small support set. In this work, we focus on the estimation of the latter, and propose to learn task-specific weights from scratch directly on a small support set, in contrast to dynamically estimating them. In particular, through systematic analysis, we show that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the per-formance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to learn task- specific weights from scratch directly on a small support set, in contrast to dynamically estimating them, and shows that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the per-formance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost.'}",https://arxiv.org/pdf/2107.00358
-melr: meta-learning via modeling episode-level relationships for few-shot learning,Nanyi Fei,,,
-finding task-relevant features for few-shot learning by category traversal,Hongyang Li,"Few-shot learning is an important area of research. Conceptually, humans are readily able to understand new concepts given just a few examples, while in more pragmatic terms, limited-example training situations are common practice. Recent effective approaches to few-shot learning employ a metric-learning framework to learn a feature similarity comparison between a query (test) example, and the few support (training) examples. However, these approaches treat each support class independently from one another, never looking at the entire task as a whole. Because of this, they are constrained to use a single set of features for all possible test-time tasks, which hinders the ability to distinguish the most relevant dimensions for the task at hand. In this work, we introduce a Category Traversal Module that can be inserted as a plug-and-play module into most metric-learning based few-shot learners. This component traverses across the entire support set at once, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space. Incorporating our module improves performance considerably (5%-10% relative) over baseline systems on both miniImageNet and tieredImageNet benchmarks, with overall performance competitive with the most recent state-of-the-art systems.","{'model': 'tldr@v2.0.0', 'text': 'A Category Traversal Module is introduced that can be inserted as a plug-and-play module into most metric-learning based few-shot learners, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space.'}",https://arxiv.org/pdf/1905.11116
-adversarial feature hallucination networks for few-shot learning,K. Li,"The recent flourish of deep learning in various tasks is largely accredited to the rich and accessible labeled data. Nonetheless, massive supervision remains a luxury for many real applications, boosting great interest in label-scarce techniques such as few-shot learning (FSL), which aims to learn concept of new classes with a few labeled samples. A natural approach to FSL is data augmentation and many recent works have proved the feasibility by proposing various data synthesis models. However, these models fail to well secure the discriminability and diversity of the synthesized data and thus often produce undesirable results. In this paper, we propose Adversarial Feature Hallucination Networks (AFHN) which is based on conditional Wasserstein Generative Adversarial networks (cWGAN) and hallucinates diverse and discriminative features conditioned on the few labeled samples. Two novel regularizers, i.e., the classification regularizer and the anti-collapse regularizer, are incorporated into AFHN to encourage discriminability and diversity of the synthesized features, respectively. Ablation study verifies the effectiveness of the proposed cWGAN based feature hallucination framework and the proposed regularizers. Comparative results on three common benchmark datasets substantiate the superiority of AFHN to existing data augmentation based FSL approaches and other state-of-the-art ones.","{'model': 'tldr@v2.0.0', 'text': 'Adversarial Feature Hallucination Networks (AFHN) is proposed which is based on conditional Wasserstein Generative Adversarial networks (cWGAN) and hallucinates diverse and discriminative features conditioned on the few labeled samples and comparative results substantiate the superiority of AFHN to existing data augmentation based FSL approaches and other state-of-the-art ones.'}",https://arxiv.org/pdf/2003.13193
-instance credibility inference for few-shot learning,Yikai Wang,"Few-shot learning (FSL) aims to recognize new objects with extremely limited training data for each category. Previous efforts are made by either leveraging meta-learning paradigm or novel principles in data augmentation to alleviate this extremely data-scarce problem. In contrast, this paper presents a simple statistical approach, dubbed Instance Credibility Inference (ICI) to exploit the distribution support of unlabeled instances for few-shot learning. Specifically, we first train a linear classifier with the labeled few-shot examples and use it to infer the pseudo-labels for the unlabeled data. To measure the credibility of each pseudo-labeled instance, we then propose to solve another linear regression hypothesis by increasing the sparsity of the incidental parameters and rank the pseudo-labeled instances with their sparsity degree. We select the most trustworthy pseudo-labeled instances alongside the labeled examples to re-train the linear classifier. This process is iterated until all the unlabeled samples are included in the expanded training set, i.e. the pseudo-label is converged for unlabeled data pool. Extensive experiments under two few-shot settings show that our simple approach can establish new state-of-the-arts on four widely used few-shot learning benchmark datasets including miniImageNet, tieredImageNet, CIFAR-FS, and CUB. Our code is available at: https://github.com/Yikai-Wang/ICI-FSL","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple statistical approach, dubbed Instance Credibility Inference (ICI), to exploit the distribution support of unlabeled instances for few-shot learning to establish new state-of-the-arts on four widely used few- shot learning benchmark datasets.'}",https://arxiv.org/pdf/2003.11853
-fs-mol: a few-shot learning dataset of molecules,Megan Stanley,"Small datasets are ubiquitous in drug discovery as data generation is expensive and can be restricted for ethical reasons (e.g. in vivo experiments). A widely applied technique in early drug discovery to identify novel active molecules against a protein target is modeling quantitative structure-activity relationships (QSAR). It is known to be extremely challenging, as available measurements of compound activities range in the low dozens or hundreds. However, many such related datasets exist, each with a small number of datapoints, opening up the opportunity for few-shot learning after pretraining on a substantially larger corpus of data. At the same time, many few-shot learning methods are currently evaluated in the computer-vision domain. We propose that expansion into a new application, as well as the possibility to use explicitly graph-structured data, will drive exciting progress in few-shot learning. Here, we provide a few-shot learning dataset (FS-Mol) and complementary benchmarking procedure. We define a set of tasks on which few-shot learning methods can be evaluated, with a separate set of tasks for use in pretraining. In addition, we implement and evaluate a number of existing single-task, multi-task, and meta-learning approaches as baselines for the community. We hope that our dataset, support code release, and baselines will encourage future work on this extremely challenging new domain for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a few-shot learning dataset (FS-Mol) and complementary benchmarking procedure, and implements and evaluates a number of existing single-task, multi- task, and meta-learning approaches as baselines for the community.'}",
-yuan 1.0: large-scale pre-trained language model in zero-shot and few-shot learning,Shaohua Wu,"Recent work like GPT-3 has demonstrated excellent performance of Zero-Shot and Few-Shot learning on many natural language processing (NLP) tasks by scaling up model size, dataset size and the amount of computation. However, training a model like GPT-3 requires huge amount of computational resources which makes it challengeable to researchers. In this work, we propose a method that incorporates large-scale distributed training performance into model architecture design. With this method, Yuan 1.0, the current largest singleton language model with 245B parameters, achieves excellent performance on thousands GPUs during training, and the state-of-the-art results on NLP tasks. A data processing method is designed to efficiently filter massive amount of raw data. The current largest high-quality Chinese corpus with 5TB high quality texts is built based on this method. In addition, a calibration and label expansion method is proposed to improve the Zero-Shot and Few-Shot performance, and steady improvement is observed on the accuracy of various tasks. Yuan 1.0 presents strong capacity of natural language generation, and the generated articles are difficult to distinguish from the human-written ones.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method that incorporates large-scale distributed training performance into model architecture design and achieves excellent performance on thousands GPUs during training, and the state-of-the-art results on NLP tasks.'}",
-boosting the generalization capability in cross-domain few-shot learning via noise-enhanced supervised autoencoder,Hanwen Liang,"State of the art (SOTA) few-shot learning (FSL) methods suffer significant performance drop in the presence of domain differences between source and target datasets. The strong discrimination ability on the source dataset does not necessarily translate to high classification accuracy on the target dataset. In this work, we address this cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model. Specifically, we teach the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE). NSAE trains the model by jointly reconstructing inputs and predicting the labels of inputs as well as their reconstructed pairs. Theoretical analysis based on intra-class correlation (ICC) shows that the feature embeddings learned from NSAE have stronger discrimination and generalization abilities in the target domain. We also take advantage of NSAE structure and propose a two-step fine-tuning procedure that achieves better adaption and improves classification performance in the target domain. Extensive experiments and ablation studies are conducted to demonstrate the effectiveness of the proposed method. Experimental results show that our proposed method consistently outperforms SOTA methods under various conditions.","{'model': 'tldr@v2.0.0', 'text': 'This work addresses the cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model by teaching the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE).'}",https://arxiv.org/pdf/2108.05028
-interventional few-shot learning,Zhongqi Yue,"We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL) methods: the pre-trained knowledge is indeed a confounder that limits the performance. This finding is rooted from our causal assumption: a Structural Causal Model (SCM) for the causalities among the pre-trained knowledge, sample features, and labels. Thanks to it, we propose a novel FSL paradigm: Interventional Few-Shot Learning (IFSL). Specifically, we develop three effective IFSL algorithmic implementations based on the backdoor adjustment, which is essentially a causal intervention towards the SCM of many-shot learning: the upper-bound of FSL in a causal view. It is worth noting that the contribution of IFSL is orthogonal to existing fine-tuning and meta-learning based FSL methods, hence IFSL can improve all of them, achieving a new 1-/5-shot state-of-the-art on \textit{mini}ImageNet, \textit{tiered}ImageNet, and cross-domain CUB. Code is released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that the contribution of IFSL is orthogonal to existing fine-tuning and meta-learning based FSL methods, hence IFSL can improve all of them, achieving a new 1-/5-shot state-of-the-art on \\textit{mini}ImageNet, \\text it{tiered}Image net, and cross-domain CUB.'}",
-true few-shot learning with prompts—a real-world perspective,Timo Schick,"Abstract Prompt-based approaches excel at few-shot learning. However, Perez et al. (2021) recently cast doubt on their performance as they had difficulty getting good results in a “true” few-shot setting in which prompts and hyperparameters cannot be tuned on a dev set. In view of this, we conduct an extensive study of Pet, a method that combines textual instructions with example-based finetuning. We show that, if correctly configured, Pet performs strongly in true few-shot settings without a dev set. Crucial for this strong performance is a number of design choices, including Pet’s ability to intelligently handle multiple prompts. We put our findings to a real-world test by running Pet on RAFT, a benchmark of tasks taken from realistic NLP applications for which no labeled dev or test sets are available. Pet achieves a new state of the art on RAFT and performs close to non-expert humans for 7 out of 11 tasks. These results demonstrate that prompt-based learners can successfully be applied in true few-shot settings and underpin our belief that learning from instructions will play an important role on the path towards human-like few-shot learning capabilities.","{'model': 'tldr@v2.0.0', 'text': 'An extensive study of Pet, a method that combines textual instructions with example-based finetuning, shows that, if correctly configured, Pet performs strongly in true few-shot settings without a dev set and underpin the belief that learning from instructions will play an important role on the path towards human-like few- shot learning capabilities.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00485/2030692/tacl_a_00485.pdf
-pareto self-supervised training for few-shot learning,Zhengyu Chen,"While few-shot learning (FSL) aims for rapid generalization to new concepts with little supervision, self-supervised learning (SSL) constructs supervisory signals directly computed from unlabeled data. Exploiting the complementarity of these two manners, few-shot auxiliary learning has recently drawn much attention to deal with few labeled data. Previous works benefit from sharing inductive bias between the main task (FSL) and auxiliary tasks (SSL), where the shared parameters of tasks are optimized by minimizing a linear combination of task losses. However, it is challenging to select a proper weight to balance tasks and reduce task conflict. To handle the problem as a whole, we propose a novel approach named as Pareto self-supervised training (PSST) for FSL. PSST explicitly decomposes the few-shot auxiliary problem into multiple constrained multi-objective subproblems with different trade-off preferences, and here a preference region in which the main task achieves the best performance is identified. Then, an effective preferred Pareto exploration is proposed to find a set of optimal solutions in such a preference region. Extensive experiments on several public benchmark datasets validate the effectiveness of our approach by achieving state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach named as Pareto self-supervised training (PSST) for FSL is proposed, which decomposes the few-shot auxiliary problem into multiple constrained multi-objective subproblems with different trade-off preferences, and here a preference region in which the main task achieves the best performance is identified.'}",https://arxiv.org/pdf/2104.07841
-leveraging the feature distribution in transfer-based few-shot learning,Yuqing Hu,,"{'model': 'tldr@v2.0.0', 'text': 'A transfer-based novel method that builds on two steps: preprocessing the feature vectors so that they become closer to Gaussian-like distributions, and leveraging this preprocessing using an optimal-transport inspired algorithm.'}",https://arxiv.org/pdf/2006.03806
-iept: instance-level and episode-level pretext tasks for few-shot learning,Manli Zhang,,,
-laplacian regularized few-shot learning,Imtiaz Masud Ziko,"We propose a transductive Laplacian-regularized inference for few-shot tasks. Given any feature embedding learned from the base classes, we minimize a quadratic binary-assignment function containing two terms: (1) a unary term assigning query samples to the nearest class prototype, and (2) a pairwise Laplacian term encouraging nearby query samples to have consistent label assignments. Our transductive inference does not re-train the base model, and can be viewed as a graph clustering of the query set, subject to supervision constraints from the support set. We derive a computationally efficient bound optimizer of a relaxation of our function, which computes independent (parallel) updates for each query sample, while guaranteeing convergence. Following a simple cross-entropy training on the base classes, and without complex meta-learning strategies, we conducted comprehensive experiments over five few-shot learning benchmarks. Our LaplacianShot consistently outperforms state-of-the-art methods by significant margins across different models, settings, and data sets. Furthermore, our transductive inference is very fast, with computational times that are close to inductive inference, and can be used for large-scale few-shot tasks.","{'model': 'tldr@v2.0.0', 'text': ""The authors' transductive LaplacianShot consistently outperforms state-of-the-art methods by significant margins across different models, settings, and data sets, and can be used for large-scale few-shot tasks.""}",
-"learn continually, generalize rapidly: lifelong knowledge accumulation for few-shot learning",Xisen Jin,"The ability to continuously expand knowledge over time and utilize it to rapidly generalize to new tasks is a key feature of human linguistic intelligence. Existing models that pursue rapid generalization to new tasks (e.g., few-shot learning methods), however, are mostly trained in a single shot on fixed datasets, unable to dynamically expand their knowledge; while continual learning algorithms are not specifically designed for rapid generalization. We present a new learning setup, Continual Learning of Few-Shot Learners (CLIF), to address the challenges of both learning settings in a unified setup. CLIF assumes a model learns from a sequence of diverse NLP tasks arriving sequentially, accumulating knowledge for improved generalization to new tasks, while also retaining performance on the tasks learned earlier. We examine how the generalization ability is affected in the continual learning setup, evaluate a number of continual learning algorithms, and propose a novel regularized adapter generation approach. We find that catastrophic forgetting affects generalization ability to a less degree than performance on seen tasks; while continual learning algorithms can still bring considerable benefit to the generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'It is found that catastrophic forgetting affects generalization ability to a less degree than performance on seen tasks; while continual learning algorithms can still bring considerable benefit to thegeneralization ability.'}",https://aclanthology.org/2021.findings-emnlp.62.pdf
-"interactive few-shot learning: limited supervision, better medical image segmentation",Ruiwei Feng,"Many known supervised deep learning methods for medical image segmentation suffer an expensive burden of data annotation for model training. Recently, few-shot segmentation methods were proposed to alleviate this burden, but such methods often showed poor adaptability to the target tasks. By prudently introducing interactive learning into the few-shot learning strategy, we develop a novel few-shot segmentation approach called Interactive Few-shot Learning (IFSL), which not only addresses the annotation burden of medical image segmentation models but also tackles the common issues of the known few-shot segmentation methods. First, we design a new few-shot segmentation structure, called Medical Prior-based Few-shot Learning Network (MPrNet), which uses only a few annotated samples (e.g., 10 samples) as support images to guide the segmentation of query images without any pre-training. Then, we propose an Interactive Learning-based Test Time Optimization Algorithm (IL-TTOA) to strengthen our MPrNet on the fly for the target task in an interactive fashion. To our best knowledge, our IFSL approach is the first to allow few-shot segmentation models to be optimized and strengthened on the target tasks in an interactive and controllable manner. Experiments on four few-shot segmentation tasks show that our IFSL approach outperforms the state-of-the-art methods by more than 20% in the DSC metric. Specifically, the interactive optimization algorithm (IL-TTOA) further contributes ~10% DSC improvement for the few-shot segmentation models.","{'model': 'tldr@v2.0.0', 'text': 'The IFSL approach is the first to allow few-shot segmentation models to be optimized and strengthened on the target tasks in an interactive and controllable manner and outperforms the state-of-the-art methods by more than 20% in the DSC metric.'}",
-boosting few-shot learning with adaptive margin loss,Aoxue Li,"Few-shot learning (FSL) has attracted increasing attention in recent years but remains challenging, due to the intrinsic difficulty in learning to generalize from a few examples. This paper proposes an adaptive margin principle to improve the generalization ability of metric-based meta-learning approaches for few-shot learning problems. Specifically, we first develop a class-relevant additive margin loss, where semantic similarity between each pair of classes is considered to separate samples in the feature embedding space from similar classes. Further, we incorporate the semantic context among all classes in a sampled training task and develop a task-relevant additive margin loss to better distinguish samples from different classes. Our adaptive margin method can be easily extended to a more realistic generalized FSL setting. Extensive experiments demonstrate that the proposed method can boost the performance of current metric-based meta-learning approaches, under both the standard FSL and generalized FSL settings.","{'model': 'tldr@v2.0.0', 'text': 'An adaptive margin principle is proposed to improve the generalization ability of metric-based meta-learning approaches for few-shot learning problems by developing a class-relevant additive margin loss, where semantic similarity between each pair of classes is considered to separate samples in the feature embedding space from similar classes.'}",http://arxiv.org/pdf/2005.13826
-few-shot learning via feature hallucination with variational inference,Qinxuan Luo,"Deep learning has achieved huge success in the field of artificial intelligence, but the performance heavily depends on labeled data. Few-shot learning aims to make a model rapidly adapt to unseen classes with few labeled samples after training on a base dataset, and this is useful for tasks lacking labeled data such as medical image processing. Considering that the core problem of few-shot learning is the lack of samples, a straightforward solution to this issue is data augmentation. This paper proposes a generative model (VI-Net) based on a cosine-classifier baseline. Specifically, we construct a framework to learn to define a generating space for each category in the latent space based on few support samples. In this way, new feature vectors can be generated to help make the decision boundary of classifier sharper during the fine-tuning process. To evaluate the effectiveness of our proposed approach, we perform comparative experiments and ablation studies on mini-ImageNet and CUB. Experimental results show that VI-Net does improve performance compared with the baseline and obtains the state-of-the-art result among other augmentation-based methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a generative model (VI-Net) based on a cosine-classifier baseline and constructs a framework to learn to define a generating space for each category in the latent space based on few support samples to help make the decision boundary of classifier sharper during the fine-tuning process.'}",
-reinforced attention for few-shot learning and beyond,Jie Hong,"Few-shot learning aims to correctly recognize query samples from unseen classes given a limited number of support samples, often by relying on global embeddings of images. In this paper, we propose to equip the backbone network with an attention agent, which is trained by reinforcement learning. The policy gradient algorithm is employed to train the agent towards adaptively localizing the representative regions on feature maps over time. We further design a reward function based on the prediction of the held-out data, thus helping the attention mechanism to generalize better across the unseen classes. The extensive experiments show, with the help of the reinforced attention, that our embedding network has the capability to progressively generate a more discriminative representation in few-shot learning. Moreover, experiments on the task of image classification also show the effectiveness of the proposed design.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to equip the backbone network with an attention agent, which is trained by reinforcement learning, and designs a reward function based on the prediction of the held-out data, thus helping the attention mechanism to generalize better across the unseen classes.'}",https://arxiv.org/pdf/2104.04192
-multi-domain few-shot learning and dataset for agricultural applications,Sai Vidyaranya Nuthalapati,"Automatic classification of pests and plants (both healthy and diseased) is of paramount importance in agriculture to improve yield. Conventional deep learning models based on convolutional neural networks require thousands of labeled examples per category. In this work we propose a method to learn from a few samples to automatically classify different pests, plants, and their diseases, using Few-Shot Learning (FSL). We learn a feature extractor to generate embeddings and then update the embeddings using Transformers. Using Mahalanobis distance, a class-covariance-based metric, we then calculate the similarity of the transformed embeddings with the embedding of the image to be classified. Using our proposed architecture, we conduct extensive experiments on multiple datasets showing the effectiveness of our proposed model. We conduct 42 experiments in total to comprehensively analyze the model and it achieves up to 14% and 24% performance gains on few-shot image classification benchmarks on two datasets.We also compile a new FSL dataset containing images of healthy and diseased plants taken in real-world settings. Using our proposed architecture which has been shown to outperform several existing FSL architectures in agriculture, we provide strong baselines on our newly proposed dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to learn from a few samples to automatically classify different pests, plants, and their diseases, using Few-Shot Learning (FSL), which has been shown to outperform several existing FSL architectures in agriculture.'}",https://arxiv.org/pdf/2109.09952
-distribution consistency based covariance metric networks for few-shot learning,Wenbin Li,"Few-shot learning aims to recognize new concepts from very few examples. However, most of the existing few-shot learning methods mainly concentrate on the first-order statistic of concept representation or a fixed metric on the relation between a sample and a concept. In this work, we propose a novel end-to-end deep architecture, named Covariance Metric Networks (CovaMNet). The CovaMNet is designed to exploit both the covariance representation and covariance metric based on the distribution consistency for the few-shot classification tasks. Specifically, we construct an embedded local covariance representation to extract the second-order statistic information of each concept and describe the underlying distribution of this concept. Upon the covariance representation, we further define a new deep covariance metric to measure the consistency of distributions between query samples and new concepts. Furthermore, we employ the episodic training mechanism to train the entire network in an end-to-end manner from scratch. Extensive experiments in two tasks, generic few-shot image classification and fine-grained fewshot image classification, demonstrate the superiority of the proposed CovaMNet. The source code can be available from https://github.com/WenbinLee/CovaMNet.git.","{'model': 'tldr@v2.0.0', 'text': 'The CovaMNet is designed to exploit both the covariance representation and covariance metric based on the distribution consistency for the few-shot classification tasks and employs the episodic training mechanism to train the entire network in an end-to-end manner from scratch.'}",https://ojs.aaai.org/index.php/AAAI/article/download/4885/4758
-limited data rolling bearing fault diagnosis with few-shot learning,Ansi Zhang,"This paper focuses on bearing fault diagnosis with limited training data. A major challenge in fault diagnosis is the infeasibility of obtaining sufficient training samples for every fault type under all working conditions. Recently deep learning based fault diagnosis methods have achieved promising results. However, most of these methods require large amount of training data. In this study, we propose a deep neural network based few-shot learning approach for rolling bearing fault diagnosis with limited data. Our model is based on the siamese neural network, which learns by exploiting sample pairs of the same or different categories. Experimental results over the standard Case Western Reserve University (CWRU) bearing fault diagnosis benchmark dataset showed that our few-shot learning approach is more effective in fault diagnosis with limited data availability. When tested over different noise environments with minimal amount of training data, the performance of our few-shot learning model surpasses the one of the baseline with reasonable noise level. When evaluated over test sets with new fault types or new working conditions, few-shot models work better than the baseline trained with all fault types. All our models and datasets in this study are open sourced and can be downloaded from https://mekhub.cn/as/fault_diagnosis_with_few-shot_learning/.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network based few-shot learning approach for rolling bearing fault diagnosis with limited data availability based on the siamese neural network, which learns by exploiting sample pairs of the same or different categories.'}",https://ieeexplore.ieee.org/ielx7/6287639/8600701/08793060.pdf
-prototype rectification for few-shot learning,Jinlu Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple yet effective approach for prototype rectification in transductive setting that utilizes label propagation and feature shifting to diminish the intra-class bias and the cross- class bias.'}",https://arxiv.org/pdf/1911.10713
-generating classification weights with gnn denoising autoencoders for few-shot learning,Spyros Gidaris,"Given an initial recognition model already trained on a set of base classes, the goal of this work is to develop a meta-model for few-shot learning. The meta-model, given as input some novel classes with few training examples per class, must properly adapt the existing recognition model into a new model that can correctly classify in a unified way both the novel and the base classes. To accomplish this goal it must learn to output the appropriate classification weight vectors for those two types of classes. To build our meta-model we make use of two main innovations: we propose the use of a Denoising Autoencoder network (DAE) that (during training) takes as input a set of classification weights corrupted with Gaussian noise and learns to reconstruct the target-discriminative classification weights. In this case, the injected noise on the classification weights serves the role of regularizing the weight generating meta-model. Furthermore, in order to capture the co-dependencies between different classes in a given task instance of our meta-model, we propose to implement the DAE model as a Graph Neural Network (GNN). In order to verify the efficacy of our approach, we extensively evaluate it on ImageNet based few-shot benchmarks and we report state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the use of a Denoising Autoencoder network that takes as input a set of classification weights corrupted with Gaussian noise and learns to reconstruct the target-discriminative classification weights, and proposes to implement the DAE model as a Graph Neural Network (GNN).'}",https://arxiv.org/pdf/1905.01102
-infinite mixture prototypes for few-shot learning,Kelsey R. Allen,"We propose infinite mixture prototypes to adaptively represent both simple and complex data distributions for few-shot learning. Our infinite mixture prototypes represent each class by a set of clusters, unlike existing prototypical methods that represent each class by a single cluster. By inferring the number of clusters, infinite mixture prototypes interpolate between nearest neighbor and prototypical representations, which improves accuracy and robustness in the few-shot regime. We show the importance of adaptive capacity for capturing complex data distributions such as alphabets, with 25% absolute accuracy improvements over prototypical networks, while still maintaining or improving accuracy on the standard Omniglot and mini-ImageNet benchmarks. In clustering labeled and unlabeled data by the same clustering rule, infinite mixture prototypes achieves state-of-the-art semi-supervised accuracy. As a further capability, we show that infinite mixture prototypes can perform purely unsupervised clustering, unlike existing prototypical methods.","{'model': 'tldr@v2.0.0', 'text': 'It is shown the importance of adaptive capacity for capturing complex data distributions such as alphabets, with 25% absolute accuracy improvements over prototypical networks, while still maintaining or improving accuracy on the standard Omniglot and mini-ImageNet benchmarks.'}",
-tapnet: neural network augmented with task-adaptive projection for few-shot learning,Sung Whan Yoon,"Handling previously unseen tasks after given only a few training examples continues to be a tough challenge in machine learning. We propose TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning. Here, employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks. At the same time, for every episode, features in the embedding space are linearly projected into a new space as a form of quick task-specific conditioning. The training loss is obtained based on a distance metric between the query and the reference vectors in the projection space. Excellent generalization results in this way. When tested on the Omniglot, miniImageNet and tieredImageNet datasets, we obtain state of the art classification accuracies under various few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning by employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks.'}",
-strata: self-training with task augmentation for better few-shot learning,Tu Vu,"Despite their recent successes in tackling many NLP tasks, large-scale pre-trained language models do not perform as well in few-shot settings where only a handful of training examples are available. To address this shortcoming, we propose STraTA, which stands for Self-Training with Task Augmentation, an approach that builds on two key ideas for effective leverage of unlabeled data. First, STraTA uses task augmentation, a novel technique that synthesizes a large amount of data for auxiliary-task fine-tuning from target-task unlabeled texts. Second, STraTA performs self-training by further fine-tuning the strong base model created by task augmentation on a broad distribution of pseudo-labeled data. Our experiments demonstrate that STraTA can substantially improve sample efficiency across 12 few-shot benchmarks. Remarkably, on the SST-2 sentiment dataset, STraTA, with only 8 training examples per class, achieves comparable results to standard fine-tuning with 67K training examples. Our analyses reveal that task augmentation and self-training are both complementary and independently effective.","{'model': 'tldr@v2.0.0', 'text': 'STraTA, which stands for Self-Training with Task Augmentation, an approach that builds on two key ideas for effective leverage of unlabeled data, reveals that task augmentation and self-training are both complementary and independently effective.'}",https://aclanthology.org/2021.emnlp-main.462.pdf
-task-aware part mining network for few-shot learning,Jiamin Wu,"Few-Shot Learning (FSL) aims at classifying samples into new unseen classes with only a handful of labeled samples available. However, most of the existing methods are based on the image-level pooled representation, yet ignore considerable local clues that are transferable across tasks. To address this issue, we propose an end-to-end Task-aware Part Mining Network (TPMN) by integrating an automatic part mining process into the metric-based model for FSL. The proposed TPMN model enjoys several merits. First, we design a meta filter learner to generate task-aware part filters based on the task embedding in a meta-learning way. The task-aware part filters can adapt to any individual task and automatically mine task-related local parts even for an unseen task. Second, an adaptive importance generator is proposed to identify key local parts and assign adaptive importance weights to different parts. To the best of our knowledge, this is the first work to automatically exploit the task-aware local parts in a meta-learning way for FSL. Extensive experimental results on four standard benchmarks demonstrate that the proposed model performs favorably against state-of-the-art FSL methods.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end Task-aware Part Mining Network (TPMN) is proposed by integrating an automatic part mining process into the metric-based model for FSL by designing a meta filter learner to generate task-aware part filters based on the task embedding in a meta-learning way.'}",
-cins: comprehensive instruction for few-shot learning in task-oriented dialog systems,Fei Mi,"As the labeling cost for different modules in task-oriented dialog (ToD) systems is high, a major challenge is to learn different tasks with the least amount of labeled data. Recently, pre-trained language models (PLMs) have shown promising results for few-shot learning in ToD. To better utilize the power of PLMs, this paper proposes Comprehensive Instruction (CINS) that exploits PLMs with extra task-specific instructions. We design a schema (definition, constraint, prompt) of instructions and their customized realizations for three important downstream tasks in ToD, ie. intent classification, dialog state tracking, and natural language generation. A sequence-to-sequence model (T5) is adopted to solve these three tasks in a unified framework. Extensive experiments are conducted on these ToD tasks in realistic few-shot learning scenarios with small validation data. Empirical results demonstrate that the proposed CINS approach consistently improves techniques that finetune PLMs with raw input or short prompt.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive Instruction (CINS) is proposed that exploits PLMs with extra task-specific instructions and their customized realizations for three important downstream tasks in ToD, ie.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21356/21105
-task agnostic meta-learning for few-shot learning,Muhammad Abdullah Jamal,"Meta-learning approaches have been proposed to tackle the few-shot learning problem. Typically, a meta-learner is trained on a variety of tasks in the hopes of being generalizable to new tasks. However, the generalizability on new tasks of a meta-learner could be fragile when it is over-trained on existing tasks during meta-training phase. In other words, the initial model of a meta-learner could be too biased towards existing tasks to adapt to new tasks, especially when only very few examples are available to update the model. To avoid a biased meta-learner and improve its generalizability, we propose a novel paradigm of Task-Agnostic Meta-Learning (TAML) algorithms. Specifically, we present an entropy-based approach that meta-learns an unbiased initial model with the largest uncertainty over the output labels by preventing it from over-performing in classification tasks. Alternatively, a more general inequality-minimization TAML is presented for more ubiquitous scenarios by directly minimizing the inequality of initial losses beyond the classification tasks wherever a suitable loss can be defined. Experiments on benchmarked datasets demonstrate that the proposed approaches outperform compared meta-learning algorithms in both few-shot classification and reinforcement learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'An entropy-based approach that meta-learns an unbiased initial model with the largest uncertainty over the output labels by preventing it from over-performing in classification tasks, which outperforms compared meta-learning algorithms in both few-shot classification and reinforcement learning tasks.'}",https://arxiv.org/pdf/1805.07722
-story centaur: large language model few shot learning as a creative writing tool,Benjamin Swanson,"Few shot learning with large language models has the potential to give individuals without formal machine learning training the access to a wide range of text to text models. We consider how this applies to creative writers and present Story Centaur, a user interface for prototyping few shot models and a set of recombinable web components that deploy them. Story Centaur’s goal is to expose creative writers to few shot learning with a simple but powerful interface that lets them compose their own co-creation tools that further their own unique artistic directions. We build out several examples of such tools, and in the process probe the boundaries and issues surrounding generation with large language models.","{'model': 'tldr@v2.0.0', 'text': 'Story Centaur is a user interface for prototyping few shot models and a set of recombinable web components that deploy them to expose creative writers to few shot learning with a simple but powerful interface that lets them compose their own co-creation tools that further their own unique artistic directions.'}",https://aclanthology.org/2021.eacl-demos.29.pdf
-generalized and incremental few-shot learning by explicit learning and calibration without forgetting,Anna Kukleva,"Both generalized and incremental few-shot learning have to deal with three major challenges: learning novel classes from only few samples per class, preventing catastrophic forgetting of base classes, and classifier calibration across novel and base classes. In this work we propose a three-stage framework that allows to explicitly and effectively address these challenges. While the first phase learns base classes with many samples, the second phase learns a calibrated classifier for novel classes from few samples while also preventing catastrophic forgetting. In the final phase, calibration is achieved across all classes. We evaluate the proposed framework on four challenging benchmark datasets for image and video few-shot classification and obtain state-of-the-art results for both generalized and incremental few shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a three-stage framework that allows to explicitly and effectively address the challenges of generalized and incremental few shot learning and evaluates the proposed framework on four challenging benchmark datasets for image and video few-shot classification and obtains state-of-the-art results.'}",https://arxiv.org/pdf/2108.08165
-graph prototypical networks for few-shot learning on attributed networks,Kaize Ding,"Attributed networks nowadays are ubiquitous in a myriad of high-impact applications, such as social network analysis, financial fraud detection, and drug discovery. As a central analytical task on attributed networks, node classification has received much attention in the research community. In real-world attributed networks, a large portion of node classes only contains limited labeled instances, rendering a long-tail node class distribution. Existing node classification algorithms are unequipped to handle the few-shot node classes. As a remedy, few-shot learning has attracted a surge of attention in the research community. Yet, few-shot node classification remains a challenging problem as we need to address the following questions: (i) How to extract meta-knowledge from an attributed network for few-shot node classification? (ii) How to identify the informativeness of each labeled instance for building a robust and effective model? To answer these questions, in this paper, we propose a graph meta-learning framework -- Graph Prototypical Networks (GPN). By constructing a pool of semi-supervised node classification tasks to mimic the real test environment, GPN is able to perform meta-learning on an attributed network and derive a highly generalizable model for handling the target classification task. Extensive experiments demonstrate the superior capability of GPN in few-shot node classification.","{'model': 'tldr@v2.0.0', 'text': 'By constructing a pool of semi-supervised node classification tasks to mimic the real test environment, GPN is able to perform meta-learning on an attributed network and derive a highly generalizable model for handling the target classification task.'}",https://arxiv.org/pdf/2006.12739
-pseudo-loss confidence metric for semi-supervised few-shot learning,Kai Huang,"Semi-supervised few-shot learning is developed to train a classifier that can adapt to new tasks with limited labeled data and a fixed quantity of unlabeled data. Most semi-supervised few-shot learning methods select pseudo-labeled data of unlabeled set by task-specific confidence estimation. This work presents a task-unified confidence estimation approach for semi-supervised few-shot learning, named pseudo-loss confidence metric (PLCM). It measures the data credibility by the loss distribution of pseudo-labels, which is synthetical considered multi-tasks. Specifically, pseudo-labeled data of different tasks are mapped to a unified metric space by mean of the pseudo-loss model, making it possible to learn the prior pseudo-loss distribution. Then, confidence of pseudo-labeled data is estimated according to the distribution component confidence of its pseudo-loss. Thus highly reliable pseudo-labeled data are selected to strengthen the classifier. Moreover, to overcome the pseudo-loss distribution shift and improve the effectiveness of classifier, we advance the multi-step training strategy coordinated with the class balance measures of class-apart selection and class weight. Experimental results on four popular benchmark datasets demonstrate that the proposed approach can effectively select pseudo-labeled data and achieve the state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A task-unified confidence estimation approach for semi-supervised few-shot learning, named pseudo-loss confidence metric (PLCM), which measures the data credibility by the loss distribution of pseudo-labels, which is synthetical considered multi-tasks.'}",
-contrastive prototype learning with augmented embeddings for few-shot learning,Yizhao Gao,"Most recent few-shot learning (FSL) methods are based on meta-learning with episodic training. In each meta-training episode, a discriminative feature embedding and/or classifier are first constructed from a support set in an inner loop, and then evaluated in an outer loop using a query set for model updating. This query set sample centered learning objective is however intrinsically limited in addressing the lack of training data problem in the support set. In this paper, a novel contrastive prototype learning with augmented embeddings (CPLAE) model is proposed to overcome this limitation. First, data augmentations are introduced to both the support and query sets with each sample now being represented as an augmented embedding (AE) composed of concatenated embeddings of both the original and augmented versions. Second, a novel support set class prototype centered contrastive loss is proposed for contrastive prototype learning (CPL). With a class prototype as an anchor, CPL aims to pull the query samples of the same class closer and those of different classes further away. This support set sample centered loss is highly complementary to the existing query centered loss, fully exploiting the limited training data in each episode. Extensive experiments on several benchmarks demonstrate that our proposed CPLAE achieves new state-of-the-art.","{'model': 'tldr@v2.0.0', 'text': 'A novel contrastive prototype learning with augmented embeddings (CPLAE) model is proposed to overcome the lack of training data problem in the support set and achieves new state-of-the-art results.'}",
-curvature generation in curved spaces for few-shot learning,Zhi Gao,"Few-shot learning describes the challenging problem of recognizing samples from unseen classes given very few labeled examples. In many cases, few-shot learning is cast as learning an embedding space that assigns test samples to their corresponding class prototypes. Previous methods assume that data of all few-shot learning tasks comply with a fixed geometrical structure, mostly a Euclidean structure. Questioning this assumption that is clearly difficult to hold in real-world scenarios and incurs distortions to data, we propose to learn a task-aware curved embedding space by making use of the hyperbolic geometry. As a result, task-specific embedding spaces where suitable curvatures are generated to match the characteristics of data are constructed, leading to more generic embedding spaces. We then leverage on intra-class and inter-class context information in the embedding space to generate class prototypes for discriminative classification. We conduct a comprehensive set of experiments on inductive and transductive few-shot learning, demonstrating the benefits of our proposed method over existing embedding methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to learn a task-aware curved embedding space by making use of the hyperbolic geometry of Euclidean geometry, and leverages on intra-class and inter-class context information in theembedding space to generate class prototypes for discriminative classification.'}",https://openaccess.thecvf.com/content/ICCV2021/papers/Gao_Curvature_Generation_in_Curved_Spaces_for_Few-Shot_Learning_ICCV_2021_paper.pdf
-squeezing backbone feature distributions to the max for efficient few-shot learning,Yuqing Hu,"In many real-life problems, it is difficult to acquire or label large amounts of data, resulting in so-called few-shot learning problems. However, few-shot classification is a challenging problem due to the uncertainty caused by using few labeled samples. In the past few years, many methods have been proposed with the common aim of transferring knowledge acquired on a previously solved task, which is often achieved by using a pretrained feature extractor. As such, if the initial task contains many labeled samples, it is possible to circumvent the limitations of few-shot learning. A shortcoming of existing methods is that they often require priors about the data distribution, such as the balance between considered classes. In this paper, we propose a novel transfer-based method with a double aim: providing state-of-the-art performance, as reported on standardized datasets in the field of few-shot learning, while not requiring such restrictive priors. Our methodology is able to cope with both inductive cases, where prediction is performed on test samples independently from each other, and transductive cases, where a joint (batch) prediction is performed.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel transfer-based method with a double aim: providing state-of-the-art performance, as reported on standardized datasets in the field of few-shot learning, while not requiring such restrictive priors.'}",https://www.mdpi.com/1999-4893/15/5/147/pdf?version=1650974986
-"z-score normalization, hubness, and few-shot learning",Nanyi Fei,"The goal of few-shot learning (FSL) is to recognize a set of novel classes with only few labeled samples by exploiting a large set of abundant base class samples. Adopting a meta-learning framework, most recent FSL methods meta-learn a deep feature embedding network, and during inference classify novel class samples using nearest neighbor in the learned high-dimensional embedding space. This means that these methods are prone to the hubness problem, that is, a certain class prototype becomes the nearest neighbor of many test instances regardless which classes they belong to. However, this problem is largely ignored in existing FSL studies. In this work, for the first time we show that many FSL methods indeed suffer from the hubness problem. To mitigate its negative effects, we further propose to employ z-score feature normalization, a simple yet effective trans-formation, during meta-training. A theoretical analysis is provided on why it helps. Extensive experiments are then conducted to show that with z-score normalization, the performance of many recent FSL methods can be boosted, resulting in new state-of-the-art on three benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that many FSL methods indeed suffer from the hubness problem, and proposes to employ z-score feature normalization, a simple yet effective trans-formation, during meta-training to mitigate its negative effects.'}",
-a survey on few-shot learning in natural language processing,Meng Yang,"The annotated dataset is the foundation for Supervised Natural Language Processing. However, the cost of obtaining dataset is high. In recent years, the Few-Shot Learning has gradually attracted the attention of researchers. From the definition, in this paper, we conclude the difference in Few-Shot Learning between Natural Language Processing and Computer Vision. On that basis, the current Few-Shot Learning on Natural Language Processing is summarized, including Transfer Learning, Meta Learning and Knowledge Distillation. Furthermore, we conclude the solutions to Few-Shot Learning in Natural Language Processing, such as the method based on Distant Supervision, Meta Learning and Knowledge Distillation. Finally, we present the challenges facing Few-Shot Learning in Natural Language Processing.","{'model': 'tldr@v2.0.0', 'text': 'The difference in Few-Shot Learning between Natural Language Processing and Computer Vision is concluded and the current Few- shot Learning on Natural Language processing is summarized, including Transfer Learning, Meta Learning and Knowledge Distillation.'}",
-mhfc: multi-head feature collaboration for few-shot learning,Shuai Shao,"Few-shot learning (FSL) aims to address the data-scarce problem. A standard FSL framework is composed of two components: (1) Pre-train. Employ the base data to generate a CNN-based feature extraction model (FEM). (2) Meta-test. Apply the trained FEM to acquire the novel data's features and recognize them. FSL relies heavily on the design of the FEM. However, various FEMs have distinct emphases. For example, several may focus more attention on the contour information, whereas others may lay particular emphasis on the texture information. The single-head feature is only a one-sided representation of the sample. Besides the negative influence of cross-domain (e.g., the trained FEM can not adapt to the novel class flawlessly), the distribution of novel data may have a certain degree of deviation compared with the ground truth distribution, which is dubbed as distribution-shift-problem (DSP). To address the DSP, we propose Multi-Head Feature Collaboration (MHFC) algorithm, which attempts to project the multi-head features (e.g., multiple features extracted from a variety of FEMs) to a unified space and fuse them to capture more discriminative information. Typically, first, we introduce a subspace learning method to transform the multi-head features to aligned low-dimensional representations. It corrects the DSP via learning the feature with more powerful discrimination and overcomes the problem of inconsistent measurement scales from different head features. Then, we design an attention block to update combination weights for each head feature automatically. It comprehensively considers the contribution of various perspectives and further improves the discrimination of features. We evaluate the proposed method on five benchmark datasets (including cross-domain experiments) and achieve significant improvements of 2.1%-7.8% compared with state-of-the-arts.","{'model': 'tldr@v2.0.0', 'text': 'Multi-Head Feature Collaboration (MHFC) algorithm is proposed, which attempts to project the multi-head features to a unified space and fuse them to capture more discriminative information.'}",https://arxiv.org/pdf/2109.07785
-dense classification and implanting for few-shot learning,Yann Lifchitz,"Few-shot learning for deep neural networks is a highly challenging and key problem in many computer vision tasks. In this context, we are targeting knowledge transfer from a set with abundant data to other sets with few available examples. We propose two simple and effective solutions: (i) dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and (ii) implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features. Implanting enables training of multiple layers in the few-shot regime, departing from most related methods derived from metric learning that train only the final layer. Both contributions show consistent gains when used individually or jointly and we report state of the art performance on few-shot classification on miniImageNet.","{'model': 'tldr@v2.0.0', 'text': 'Two simple and effective solutions are proposed: dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features.'}",https://arxiv.org/pdf/1903.05050
-transductive episodic-wise adaptive metric for few-shot learning,Limeng Qiao,"Few-shot learning, which aims at extracting new concepts rapidly from extremely few examples of novel classes, has been featured into the meta-learning paradigm recently. Yet, the key challenge of how to learn a generalizable classifier with the capability of adapting to specific tasks with severely limited data still remains in this domain. To this end, we propose a Transductive Episodic-wise Adaptive Metric (TEAM) framework for few-shot learning, by integrating the meta-learning paradigm with both deep metric learning and transductive inference. With exploring the pairwise constraints and regularization prior within each task, we explicitly formulate the adaptation procedure into a standard semi-definite programming problem. By solving the problem with its closed-form solution on the fly with the setup of transduction, our approach efficiently tailors an episodic-wise metric for each task to adapt all features from a shared task-agnostic embedding space into a more discriminative task-specific metric space. Moreover, we further leverage an attention-based bi-directional similarity strategy for extracting the more robust relationship between queries and prototypes. Extensive experiments on three benchmark datasets show that our framework is superior to other existing approaches and achieves the state-of-the-art performance in the few-shot literature.","{'model': 'tldr@v2.0.0', 'text': 'A Transductive Episodic-wise Adaptive Metric (TEAM) framework for few-shot learning is proposed, by integrating the meta-learning paradigm with both deep metric learning and transductive inference and leverages an attention-based bi-directional similarity strategy for extracting the more robust relationship between queries and prototypes.'}",https://arxiv.org/pdf/1910.02224
-libfewshot: a comprehensive library for few-shot learning,Wenbin Li,"Few-shot learning, especially few-shot image classification, has received increasing attention and witnessed significant advances in recent years. Some recent studies implicitly show that many generic techniques or “tricks”, such as data augmentation, pre-training, knowledge distillation, and self-supervision, may greatly boost the performance of a few-shot learning method. Moreover, different works may employ different software platforms, backbone architectures and input image sizes, making fair comparisons difficult and practitioners struggle with reproducibility. To address these situations, we propose a comprehensive library for few-shot learning (LibFewShot) by re-implementing eighteen state-of-the-art few-shot learning methods in a unified framework with the same single codebase in PyTorch. Furthermore, based on LibFewShot, we provide comprehensive evaluations on multiple benchmarks with various backbone architectures to evaluate common pitfalls and effects of different training tricks. In addition, with respect to the recent doubts on the necessity of meta- or episodic-training mechanism, our evaluation results confirm that such a mechanism is still necessary especially when combined with pre-training. We hope our work can not only lower the barriers for beginners to enter the area of few-shot learning but also elucidate the effects of nontrivial tricks to facilitate intrinsic research on few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a comprehensive library for few-shot learning (LibFewShot) by re-implementing eighteen state-of-the-art few- shot learning methods in a unified framework with the same single codebase in PyTorch, and provides comprehensive evaluations on multiple benchmarks with various backbone architectures to evaluate common pitfalls and effects of different training tricks.'}",https://arxiv.org/pdf/2109.04898
-multi-label few-shot learning for aspect category detection,Mengting Hu,"Aspect category detection (ACD) in sentiment analysis aims to identify the aspect categories mentioned in a sentence. In this paper, we formulate ACD in the few-shot learning scenario. However, existing few-shot learning approaches mainly focus on single-label predictions. These methods can not work well for the ACD task since a sentence may contain multiple aspect categories. Therefore, we propose a multi-label few-shot learning method based on the prototypical network. To alleviate the noise, we design two effective attention mechanisms. The support-set attention aims to extract better prototypes by removing irrelevant aspects. The query-set attention computes multiple prototype-specific representations for each query instance, which are then used to compute accurate distances with the corresponding prototypes. To achieve multi-label inference, we further learn a dynamic threshold per instance by a policy network. Extensive experimental results on three datasets demonstrate that the proposed method significantly outperforms strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A multi-label few-shot learning method based on the prototypical network for aspect category detection that significantly outperforms strong baselines and also learns a dynamic threshold per instance by a policy network.'}",https://aclanthology.org/2021.acl-long.495.pdf
-metagan: an adversarial approach to few-shot learning,Ruixiang Zhang,"In this paper, we propose a conceptually simple and general framework called MetaGAN for few-shot learning problems. Most state-of-the-art few-shot classification models can be integrated with MetaGAN in a principled and straightforward way. By introducing an adversarial generator conditioned on tasks, we augment vanilla few-shot classification models with the ability to discriminate between real and fake data. We argue that this GAN-based approach can help few-shot classifiers to learn sharper decision boundary, which could generalize better. We show that with our MetaGAN framework, we can extend supervised few-shot learning models to naturally cope with unlabeled data. Different from previous work in semi-supervised few-shot learning, our algorithms can deal with semi-supervision at both sample-level and task-level. We give theoretical justifications of the strength of MetaGAN, and validate the effectiveness of MetaGAN on challenging few-shot image classification benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a conceptually simple and general framework called MetaGAN for few-shot learning problems, and shows that with this MetaGAN framework, it can extend supervised few- shot learning models to naturally cope with unlabeled data.'}",
-when does self-supervision improve few-shot learning?,Jong-Chyi Su,,"{'model': 'tldr@v2.0.0', 'text': 'This work investigates the role of self-supervised learning in the context of few-shot learning and presents a technique that automatically selects images for SSL from a large, generic pool of unlabeled images for a given dataset that provides further improvements.'}",https://arxiv.org/pdf/1910.03560
-eckpn: explicit class knowledge propagation network for transductive few-shot learning,Chaofan CHEN,"Recently, the transductive graph-based methods have achieved great success in the few-shot classification task. However, most existing methods ignore exploring the class-level knowledge that can be easily learned by humans from just a handful of samples. In this paper, we propose an Explicit Class Knowledge Propagation Network (ECKPN), which is composed of the comparison, squeeze and calibration modules, to address this problem. Specifically, we first employ the comparison module to explore the pairwise sample relations to learn rich sample representations in the instance-level graph. Then, we squeeze the instance-level graph to generate the class-level graph, which can help obtain the class-level visual knowledge and facilitate modeling the relations of different classes. Next, the calibration module is adopted to characterize the relations of the classes explicitly to obtain the more discriminative class-level knowledge representations. Finally, we combine the class-level knowledge with the instance-level sample representations to guide the inference of the query samples. We conduct extensive experiments on four few-shot classification benchmarks, and the experimental results show that the proposed ECKPN significantly outperforms the state-of-the art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an Explicit Class Knowledge Propagation Network (ECKPN), which is composed of the comparison, squeeze and calibration modules, and shows that the proposed ECKPN significantly outperforms the state-of-the art methods.'}",https://arxiv.org/pdf/2106.08523
-few-shot learning with embedded class models and shot-free meta training,Avinash Ravichandran,"We propose a method for learning embeddings for few-shot learning that is suitable for use with any number of shots (shot-free). Rather than fixing the class prototypes to be the Euclidean average of sample embeddings, we allow them to live in a higher-dimensional space (embedded class models) and learn the prototypes along with the model parameters. The class representation function is defined implicitly, which allows us to deal with a variable number of shots per class with a simple constant-size architecture. The class embedding encompasses metric learning, that facilitates adding new classes without crowding the class representation space. Despite being general and not tuned to the benchmark, our approach achieves state-of-the-art performance on the standard few-shot benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for learning embeddings for few-shot learning that is suitable for use with any number of shots (shot-free), that encompasses metric learning, that facilitates adding new classes without crowding the class representation space.'}",https://arxiv.org/pdf/1905.04398
-large-scale few-shot learning: knowledge transfer with class hierarchy,Aoxue Li,"Recently, large-scale few-shot learning (FSL) becomes topical. It is discovered that, for a large-scale FSL problem with 1,000 classes in the source domain, a strong baseline emerges, that is, simply training a deep feature embedding model using the aggregated source classes and performing nearest neighbor (NN) search using the learned features on the target classes. The state-of-the-art large-scale FSL methods struggle to beat this baseline, indicating intrinsic limitations on scalability. To overcome the challenge, we propose a novel large-scale FSL model by learning transferable visual features with the class hierarchy which encodes the semantic relations between source and target classes. Extensive experiments show that the proposed model significantly outperforms not only the NN baseline but also the state-of-the-art alternatives. Furthermore, we show that the proposed model can be easily extended to the large-scale zero-shot learning (ZSL) problem and also achieves the state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel large-scale FSL model by learning transferable visual features with the class hierarchy which encodes the semantic relations between source and target classes and significantly outperforms not only the NN baseline but also the state-of-the-art alternatives.'}",
-variational few-shot learning,Jian Zhang,"We propose a variational Bayesian framework for enhancing few-shot learning performance. This idea is motivated by the fact that single point based metric learning approaches are inherently noise-vulnerable and easy-to-be-biased. In a nutshell, stochastic variational inference is invoked to approximate bias-eliminated class specific sample distributions. In the meantime, a classifier-free prediction is attained by leveraging the distribution statistics on novel samples. Extensive experimental results on several benchmarks well demonstrate the effectiveness of our distribution-driven few-shot learning framework over previous point estimates based methods, in terms of superior classification accuracy and robustness.","{'model': 'tldr@v2.0.0', 'text': 'Stochastic variational inference is invoked to approximate bias-eliminated class specific sample distributions to enhance few-shot learning performance, and a classifier-free prediction is attained by leveraging the distribution statistics on novel samples.'}",
-few-shot learning via saliency-guided hallucination of samples,Hongguang Zhang,"Learning new concepts from a few of samples is a standard challenge in computer vision. The main directions to improve the learning ability of few-shot training models include (i) a robust similarity learning and (ii) generating or hallucinating additional data from the limited existing samples. In this paper, we follow the latter direction and present a novel data hallucination model. Currently, most datapoint generators contain a specialized network (i.e., GAN) tasked with hallucinating new datapoints, thus requiring large numbers of annotated data for their training in the first place. In this paper, we propose a novel less-costly hallucination method for few-shot learning which utilizes saliency maps. To this end, we employ a saliency network to obtain the foregrounds and backgrounds of available image samples and feed the resulting maps into a two-stream network to hallucinate datapoints directly in the feature space from viable foreground-background combinations. To the best of our knowledge, we are the first to leverage saliency maps for such a task and we demonstrate their usefulness in hallucinating additional datapoints for few-shot learning. Our proposed network achieves the state of the art on publicly available datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work employs a saliency network to obtain the foregrounds and backgrounds of available image samples and feeds the resulting maps into a two-stream network to hallucinate datapoints directly in the feature space from viable foreground-background combinations, the first to leverage saliency maps for such a task.'}",https://arxiv.org/pdf/1904.03472
-looking wider for better adaptive representation in few-shot learning,Jiabao Zhao,"Building a good feature space is essential for the metric-based few-shot algorithms to recognize a novel class with only a few samples. The feature space is often built by Convolutional Neural Networks (CNNs). However, CNNs primarily focus on local information with the limited receptive field, and the global information generated by distant pixels is not well used. Meanwhile, having a global understanding of the current task and focusing on distinct regions of the same sample for different queries are important for the few-shot classification. To tackle these problems, we propose the Cross Non-Local Neural Network (CNL) for capturing the long-range dependency of the samples and the current task. CNL extracts the task-specific and context-aware features dynamically by strengthening the features of the sample at a position via aggregating information from all positions of itself and the current task. To reduce losing important information, we maximize the mutual information between the original and refined features as a constraint. Moreover, we add a task-specific scaling to deal with multi-scale and task-specific features extracted by CNL. We conduct extensive experiments for validating our proposed algorithm, which achieves new state-of-the-art performances on two public benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'The Cross Non-Local Neural Network (CNL) is proposed for capturing the long-range dependency of the samples and the current task, and extracts the task-specific and context-aware features dynamically by strengthening the features of the sample at a position via aggregating information from all positions of itself and theCurrent task.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17311/17118
-graph few-shot learning via knowledge transfer,Huaxiu Yao,"Towards the challenging problem of semi-supervised node classification, there have been extensive studies. As a frontier, Graph Neural Networks (GNNs) have aroused great interest recently, which update the representation of each node by aggregating information of its neighbors. However, most GNNs have shallow layers with a limited receptive field and may not achieve satisfactory performance especially when the number of labeled nodes is quite small. To address this challenge, we innovatively propose a graph few-shot learning (GFL) algorithm that incorporates prior knowledge learned from auxiliary graphs to improve classification accuracy on the target graph. Specifically, a transferable metric space characterized by a node embedding and a graph-specific prototype embedding function is shared between auxiliary graphs and the target, facilitating the transfer of structural knowledge. Extensive experiments and ablation studies on four real-world graph datasets demonstrate the effectiveness of our proposed model and the contribution of each component.","{'model': 'tldr@v2.0.0', 'text': 'This work innovatively proposes a graph few-shot learning (GFL) algorithm that incorporates prior knowledge learned from auxiliary graphs to improve classification accuracy on the target graph.'}",https://ojs.aaai.org/index.php/AAAI/article/download/6142/5998
-attentive weights generation for few shot learning via information maximization,Yiluan Guo,"Few shot image classification aims at learning a classifier from limited labeled data. Generating the classification weights has been applied in many meta-learning methods for few shot image classification due to its simplicity and effectiveness. In this work, we present Attentive Weights Generation for few shot learning via Information Maximization (AWGIM), which introduces two novel contributions: i) Mutual information maximization between generated weights and data within the task; this enables the generated weights to retain information of the task and the specific query sample. ii) Self-attention and cross-attention paths to encode the context of the task and individual queries. Both two contributions are shown to be very effective in extensive experiments. Overall, AWGIM is competitive with state-of-the-art. Code is available at https://github.com/Yiluan/AWGIM.","{'model': 'tldr@v2.0.0', 'text': 'Attentive Weights Generation for few shot learning via Information Maximization (AWGIM), which introduces two novel contributions: i) Mutual information maximization between generated weights and data within the task; this enables the generated weights to retain information of the task and the specific query sample.'}",
-few-shot learning for palmprint recognition via meta-siamese network,Huikai Shao,"Palmprint is one of the discriminant biometric modalities of humans. Recently, deep learning-based palmprint recognition algorithms have improved the accuracy and robustness of recognition results to a new level. Most of them require a large amount of labeled training samples to guarantee satisfactory performance. However, getting enough labeled data is difficult due to time consumption and privacy issues. Therefore, in this article, a novel meta-Siamese network (MSN) is proposed to exploit few-shot learning for small-sample palmprint recognition. During each episode-based training iteration, a few images are selected as sample and query sets to simulate the support and testing sets in the test set. Specifically, the model is trained episodically with a flexible framework to learn both the feature embedding and deep similarity metric function. In addition, two distance-based losses are introduced to assist the optimization. After training, the model can learn the ability to get similarity scores between two images for few-shot testing. Adequate experiments conducted on several constrained and unconstrained benchmark palmprint databases show that MSN can obtain competitive improvements compared with baseline methods, where the best accuracy can be up to 100%.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-Siamese network is proposed to exploit few-shot learning for small-sample palmprint recognition and can obtain competitive improvements compared with baseline methods, where the best accuracy can be up to 100%.'}",https://research.utwente.nl/files/254372707/09420130.pdf
-incremental few-shot learning with attention attractor networks,Mengye Ren,"Machine learning classifiers are often trained to recognize a set of pre-defined classes. However, in many applications, it is often desirable to have the flexibility of learning additional concepts, with limited data and without re-training on the full training set. This paper addresses this problem, incremental few-shot learning, where a regular classification network has already been trained to recognize a set of base classes, and several extra novel classes are being considered, each with only a few labeled examples. After learning the novel classes, the model is then evaluated on the overall classification performance on both base and novel classes. To this end, we propose a meta-learning model, the Attention Attractor Network, which regularizes the learning of novel classes. In each episode, we train a set of new weights to recognize novel classes until they converge, and we show that the technique of recurrent back-propagation can back-propagate through the optimization process and facilitate the learning of these parameters. We demonstrate that the learned attractor network can help recognize novel classes while remembering old classes without the need to review the original training set, outperforming various baselines.","{'model': 'tldr@v2.0.0', 'text': 'A meta-learning model, the Attention Attractor Network, which regularizes the learning of novel classes, and it is demonstrated that the learned attractor network can help recognize novel classes while remembering old classes without the need to review the original training set.'}",
-radar target recognition based on few-shot learning,Yue-Wen Yang,,"{'model': 'tldr@v2.0.0', 'text': 'A sample selection method based on few-shot learning is proposed, which can achieve similar results with 19 kinds of trajectory data training model and the accuracy is improved by 4–10% in different time lengths.'}",
-hybrid graph neural networks for few-shot learning,Tianyuan Yu,"Graph neural networks (GNNs) have been used to tackle the few-shot learning (FSL) problem and shown great potentials under the transductive setting. However under the inductive setting, existing GNN based methods are less competitive. This is because they use an instance GNN as a label propagation/classification module, which is jointly meta-learned with a feature embedding network. This design is problematic because the classifier needs to adapt quickly to new tasks while the embedding does not. To overcome this problem, in this paper we propose a novel hybrid GNN (HGNN) model consisting of two GNNs, an instance GNN and a prototype GNN. Instead of label propagation, they act as feature embedding adaptation modules for quick adaptation of the meta-learned feature embedding to new tasks. Importantly they are designed to deal with a fundamental yet often neglected challenge in FSL, that is, with only a handful of shots per class, any few-shot classifier would be sensitive to badly sampled shots which are either outliers or can cause inter-class distribution overlapping. Extensive experiments show that our HGNN obtains new state-of-the-art on three FSL benchmarks. The code and models are available at https://github.com/TianyuanYu/HGNN.","{'model': 'tldr@v2.0.0', 'text': 'A novel hybrid GNN (HGNN) model consisting of two GNNs, an instance GNN and a prototype GNN designed to deal with a fundamental yet often neglected challenge in FSL, that is, with only a handful of shots per class, any few-shot classifier would be sensitive to badly sampled shots.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20226/19985
-federated few-shot learning with adversarial learning,Chenyou Fan,"We are interested in developing a unified machine learning framework for effectively training machine learning models from many small data sources such as mobile devices. This is a commonly encountered situation in mobile computing scenarios, where data is scarce and distributed while the tasks are distinct. In this paper, we propose a federated few-shot learning (FedFSL) framework to learn a few-shot classification model that can classify unseen data classes with only a few labeled samples. With the federated learning strategy, FedFSL can utilize many data sources while keeping data privacy and communication efficiency. To tackle the issue of obtaining misaligned decision boundaries produced by client models, we propose to regularize local updates by minimizing the divergence of client models. We also formulate the training in an adversarial fashion and optimize the client models to produce a discriminative feature space that can better represent unseen data samples. We demonstrate the intuitions and conduct experiments to show our approaches outperform baselines by more than 10% in learning benchmark vision tasks and 5% in language tasks.","{'model': 'tldr@v2.0.0', 'text': 'A federated few-shot learning (FedFSL) framework to learn a few- shot classification model that can classify unseen data classes with only a few labeled samples and formulate the training in an adversarial fashion and optimize the client models to produce a discriminative feature space that can better represent unseen data samples is proposed.'}",https://arxiv.org/pdf/2104.00365
-few-shot learning with a strong teacher,Han-Jia Ye,"Few-shot learning (FSL) aims to generate a classifier using limited labeled examples. Many existing works take the meta-learning approach, constructing a few-shot learner (a meta-model) that can learn from few-shot examples to generate a classifier. The performance is measured by how well the resulting classifiers classify the test (\ie, query) examples of those tasks. In this paper, we point out two potential weaknesses of this approach. First, the sampled query examples may not provide sufficient supervision for meta-training the few-shot learner. Second, the effectiveness of meta-learning diminishes sharply with the increasing number of shots. We propose a novel meta-training objective for the few-shot learner, which is to encourage the few-shot learner to generate classifiers that perform like strong classifiers. Concretely, we associate each sampled few-shot task with a strong classifier, which is trained with ample labeled examples. The strong classifiers can be seen as the target classifiers that we hope the few-shot learner to generate given few-shot examples, and we use the strong classifiers to supervise the few-shot learner. We validate our approach in combinations with many representative meta-learning methods. More importantly, with our approach, meta-learning based FSL methods can consistently outperform non-meta-learning based methods at different numbers of shots.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-training objective for the few-shot learner is proposed, which is to encourage theFew- shot learner to generate classifiers that perform like strong classifiers, and meta-learning based FSL methods can consistently outperform non-meta- learning based methods at different numbers of shots.'}",https://arxiv.org/pdf/2107.00197
-few-shot learning: a survey,Yaqing Wang,"The quest of `can machines think' and `can machines do what human do' are quests that drive the development of artificial intelligence. Although recent artificial intelligence succeeds in many data intensive applications, it still lacks the ability of learning from limited exemplars and fast generalizing to new tasks. To tackle this problem, one has to turn to machine learning, which supports the scientific study of artificial intelligence. Particularly, a machine learning problem called Few-Shot Learning (FSL) targets at this case. It can rapidly generalize to new tasks of limited supervised experience by turning to prior knowledge, which mimics human's ability to acquire knowledge from few examples through generalization and analogy. It has been seen as a test-bed for real artificial intelligence, a way to reduce laborious data gathering and computationally costly training, and antidote for rare cases learning. With extensive works on FSL emerging, we give a comprehensive survey for it. We first give the formal definition for FSL. Then we point out the core issues of FSL, which turns the problem from ""how to solve FSL"" to ""how to deal with the core issues"". Accordingly, existing works from the birth of FSL to the most recent published ones are categorized in a unified taxonomy, with thorough discussion of the pros and cons for different categories. Finally, we envision possible future directions for FSL in terms of problem setup, techniques, applications and theory, hoping to provide insights to both beginners and experienced researchers.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of the core issues of Few-Shot Learning, and existing works from the birth of FSL to the most recent published ones are categorized in a unified taxonomy, with thorough discussion of the pros and cons for different categories.'}",
-few shot learning based on the street view house numbers (svhn) dataset,Rundong Yang,,,
-mutual crf-gnn for few-shot learning,Shixiang Tang,"Graph-neural-networks (GNN) is a rising trend for fewshot learning. A critical component in GNN is the affinity. Typically, affinity in GNN is mainly computed in the feature space, e.g., pairwise features, and does not take fully advantage of semantic labels associated to these features. In this paper, we propose a novel Mutual CRF-GNN (MCGN). In this MCGN, the labels and features of support data are used by the CRF for inferring GNN affinities in a principled and probabilistic way. Specifically, we construct a Conditional Random Field (CRF) conditioned on labels and features of support data to infer a affinity in the label space. Such affinity is fed to the GNN as the node-wise affinity. GNN and CRF mutually contributes to each other in MCGN. For GNN, CRF provides valuable affinity information. For CRF, GNN provides better features for inferring affinity. Experimental results show that our approach outperforms stateof-the-arts on datasets miniImageNet, tieredImageNet, and CIFAR-FS on both 5-way 1-shot and 5-way 5-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper constructs a Conditional Random Field conditioned on labels and features of support data to infer a affinity in the label space and proposes a novel Mutual CRF-GNN (MCGN), where GNN and CRF mutually contributes to each other in MCGN.'}",
-demystification of few-shot and one-shot learning,I. Tyukin,"Few-shot and one-shot learning have been the subject of active and intensive research in recent years, with mounting evidence pointing to successful implementation and exploitation of few-shot learning algorithms in practice. Classical statistical learning theories do not fully explain why few- or one-shot learning is at all possible since traditional generalisation bounds normally require large training and testing samples to be meaningful. This sharply contrasts with numerous examples of successful one- and few-shot learning systems and applications. In this work we present mathematical foundations for a theory of one-shot and few-shot learning and reveal conditions specifying when such learning schemes are likely to succeed. Our theory is based on intrinsic properties of high-dimensional spaces. We show that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met. In this work we present mathematical foundations for a theory of one-shot and few-shot learning and reveal conditions specifying when such learning schemes are likely to succeed. Our theory is based on intrinsic properties of high-dimensional spaces. We show that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that if the ambient or latent decision space of a learning machine is sufficiently high-dimensional than a large class of objects in this space can indeed be easily learned from few examples provided that certain data non-concentration conditions are met.'}",https://arxiv.org/pdf/2104.12174
-few-shot learning approach for plant disease classification using images taken in the field,David Argüeso,,"{'model': 'tldr@v2.0.0', 'text': 'It is possible to learn new plant leaf and disease types with very small datasets using deep learning Siamese networks with Triplet loss, achieving almost a 90% reduction in training data needs and outperforming classical learning techniques for small training sets.'}",
-knowledge-guided multi-label few-shot learning for general image recognition,Tianshui Chen,"Recognizing multiple labels of an image is a practical yet challenging task, and remarkable progress has been achieved by searching for semantic regions and exploiting label dependencies. However, current works utilize RNN/LSTM to implicitly capture sequential region/label dependencies, which cannot fully explore mutual interactions among the semantic regions/labels and do not explicitly integrate label co-occurrences. In addition, these works require large amounts of training samples for each category, and they are unable to generalize to novel categories with limited samples. To address these issues, we propose a knowledge-guided graph routing (KGGR) framework, which unifies prior knowledge of statistical label correlations with deep neural networks. The framework exploits prior knowledge to guide adaptive information propagation among different categories to facilitate multi-label analysis and reduce the dependency of training samples. Specifically, it first builds a structured knowledge graph to correlate different labels based on statistical label co-occurrence. Then, it introduces the label semantics to guide learning semantic-specific features to initialize the graph, and it exploits a graph propagation network to explore graph node interactions, enabling learning contextualized image feature representations. Moreover, we initialize each graph node with the classifier weights for the corresponding label and apply another propagation network to transfer node messages through the graph. In this way, it can facilitate exploiting the information of correlated labels to help train better classifiers, especially for labels with limited training samples. We conduct extensive experiments on the traditional multi-label image recognition (MLR) and multi-label few-shot learning (ML-FSL) tasks and show that our KGGR framework outperforms the current state-of-the-art methods by sizable margins on the public benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A knowledge-guided graph routing (KGGR) framework, which unifies prior knowledge of statistical label correlations with deep neural networks and can facilitate exploiting the information of correlated labels to help train better classifiers, especially for labels with limited training samples.'}",https://arxiv.org/pdf/2009.09450
-automatic modulation recognition: a few-shot learning method based on the capsule network,Lixin Li,"With the rapid development of deep learning (DL) in recent years, automatic modulation recognition (AMR) with DL has achieved high accuracy. However, aiming to obtain higher classification accuracy, DL requires numerous training samples. In order to solve this problem, it is a challenge to study how to efficiently use DL for AMR in the case of few samples. In this letter, inspired by the capsule network (CapsNet), we propose a new network structure named AMR-CapsNet to achieve higher classification accuracy of modulation signals with fewer samples, and further analyze the adaptability of DL models in the case of few samples. The simulation results demonstrate that when 3% of the dataset is used to train and the signal-to-noise ratio (SNR) is greater than 2 dB, the overall classification accuracy of the AMR-CapsNet is greater than 80%. Compared with convolutional neural network (CNN), the classification accuracy is improved by 20%.","{'model': 'tldr@v2.0.0', 'text': 'A new network structure named AMR-CapsNet is proposed to achieve higher classification accuracy of modulation signals with fewer samples, and further analyze the adaptability of DL models in the case of few samples.'}",
-local descriptor-based multi-prototype network for few-shot learning,Hongwei Huang,,"{'model': 'tldr@v2.0.0', 'text': 'A novel Local descriptor-based Multi-Prototype Network (LMPNet) is proposed, a well-designed framework that generates an embedding space with multiple prototypes that can capture more informative and subtler cues of an image than the normally adopted image-level features.'}",
-collect and select: semantic alignment metric learning for few-shot learning,Fusheng Hao,"Few-shot learning aims to learn latent patterns from few training examples and has shown promises in practice. However, directly calculating the distances between the query image and support image in existing methods may cause ambiguity because dominant objects can locate anywhere on images. To address this issue, this paper proposes a Semantic Alignment Metric Learning (SAML) method for few-shot learning that aligns the semantically relevant dominant objects through a ``collect-and-select'' strategy. Specifically, we first calculate a relation matrix (RM) to ``collect"" the distances of each local region pairs of the $3$D tensor extracted from a query image and the mean tensor of the support images. Then, the attention technique is adapted to ``select"" the semantically relevant pairs and put more weights on them. Afterwards, a multi-layer perceptron (MLP) is utilized to map the reweighted RMs to their corresponding similarity scores. Theoretical analysis demonstrates the generalization ability of SAML and gives a theoretical guarantee. Empirical results demonstrate that semantic alignment is achieved. Extensive experiments on benchmark datasets validate the strengths of the proposed approach and demonstrate that SAML significantly outperforms the current state-of-the-art methods. The source code is available at https://github.com/haofusheng/SAML.","{'model': 'tldr@v2.0.0', 'text': ""A Semantic Alignment Metric Learning (SAML) method for few-shot learning that aligns the semantically relevant dominant objects through a ``collect-and-select'' strategy and demonstrates the generalization ability of SAML and gives a theoretical guarantee.""}",
-holodetect: few-shot learning for error detection,Alireza Heidari,"We introduce a few-shot learning framework for error detection. We show that data augmentation (a form of weak supervision) is key to training high-quality, ML-based error detection models that require minimal human involvement. Our framework consists of two parts: (1) an expressive model to learn rich representations that capture the inherent syntactic and semantic heterogeneity of errors; and (2) a data augmentation model that, given a small seed of clean records, uses dataset-specific transformations to automatically generate additional training data. Our key insight is to learn data augmentation policies from the noisy input dataset in a weakly supervised manner. We show that our framework detects errors with an average precision of ~94% and an average recall of ~93% across a diverse array of datasets that exhibit different types and amounts of errors. We compare our approach to a comprehensive collection of error detection methods, ranging from traditional rule-based methods to ensemble-based and active learning approaches. We show that data augmentation yields an average improvement of 20 F1 points while it requires access to 3x fewer labeled examples compared to other ML approaches.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning framework for error detection that consists of an expressive model to learn rich representations that capture the inherent syntactic and semantic heterogeneity of errors, and a data augmentation model that, given a small seed of clean records, uses dataset-specific transformations to automatically generate additional training data.'}",https://dl.acm.org/doi/pdf/10.1145/3299869.3319888
-meta navigator: search for a good adaptation policy for few-shot learning,Chi Zhang,"Few-shot learning aims to adapt knowledge learned from previous tasks to novel tasks with only a limited amount of labeled data. Research literature on few-shot learning exhibits great diversity, while different algorithms often excel at different few-shot learning scenarios. It is therefore tricky to decide which learning strategies to use under different task conditions. Inspired by the recent success in Automated Machine Learning literature (AutoML), in this paper, we present Meta Navigator, a framework that attempts to solve the aforementioned limitation in few-shot learning by seeking a higher-level strategy and proffer to automate the selection from various few-shot learning designs. The goal of our work is to search for good parameter adaptation policies that are applied to different stages in the network for few-shot classification. We present a search space that covers many popular few-shot learning algorithms in the literature, and develop a differentiable searching and decoding algorithm based on meta-learning that supports gradient-based optimization. We demonstrate the effectiveness of our searching-based method on multiple benchmark datasets. Extensive experiments show that our approach significantly outperforms baselines and demonstrates performance advantages over many state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the recent success in Automated Machine Learning literature, this paper presents Meta Navigator, a framework that attempts to solve the aforementioned limitation in few-shot learning by seeking a higher-level strategy and proffer to automate the selection from various few- shot learning designs.'}",https://arxiv.org/pdf/2109.05749
-few-shot learning by integrating spatial and frequency representation,Xiangyu Chen,"Human beings can recognize new objects with only a few labeled examples, however, few-shot learning remains a challenging problem for machine learning systems. Most previous algorithms in few-shot learning only utilize spatial information of the images. In this paper, we propose to integrate the frequency information into the learning model to boost the discrimination ability of the system. We employ Discrete Cosine Transformation (DCT) to generate the frequency representation, then, integrate the features from both the spatial domain and frequency domain for classification. The proposed strategy and its effectiveness are validated with different backbones, datasets, and algorithms. Extensive experiments demonstrate that the frequency information is complementary to the spatial representations in few-shot classification. The classification accuracy is boosted significantly by integrating features from both the spatial and frequency domains in different few-shot learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to integrate the frequency information into the learning model to boost the discrimination ability of the system, employing Discrete Cosine Transformation to generate the frequency representation and integrating the features from both the spatial domain and frequency domain for classification.'}",https://arxiv.org/pdf/2105.05348
-few-shot learning for road object detection,Anay Majee,"Few-shot learning is a problem of high interest in the evolution of deep learning. In this work, we consider the problem of few-shot object detection (FSOD) in a real-world, class-imbalanced scenario. For our experiments, we utilize the India Driving Dataset (IDD), as it includes a class of less-occurring road objects in the image dataset and hence provides a setup suitable for few-shot learning. We evaluate both metric-learning and meta-learning based FSOD methods, in two experimental settings: (i) representative (same-domain) splits from IDD, that evaluates the ability of a model to learn in the context of road images, and (ii) object classes with less-occurring object samples, similar to the open-set setting in real-world. From our experiments, we demonstrate that the metric-learning method outperforms meta-learning on the novel classes by (i) 11.2 mAP points on the same domain, and (ii) 1.0 mAP point on the open-set. We also show that our extension of object classes in a real-world open dataset offers a rich ground for few-shot learning studies.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the problem of few-shot object detection (FSOD) in a real-world, class-imbalanced scenario, and demonstrates that the metric-learning method outperforms meta-learning on the novel classes by 11.2 mAP points on the same domain, and 1.0 mAP point on the open-set.'}",
-learning to adapt with memory for probabilistic few-shot learning,Lei Zhang,"Few-shot learning has recently generated increasing popularity in machine learning, which addresses the fundamental yet challenging problem of learning to adapt to new tasks with the limited data. In this paper, we propose a new probabilistic framework that learns to fast adapt with external memory. We model the classifier parameters as distributions that are inferred from the support set and directly applied to the query set for prediction. The model is optimized by formulating as a variational inference problem. The probabilistic modeling enables better handling prediction uncertainty due to the limited data. We impose a discriminative constraint on the feature representations by exploring the class structure, which can improve the classification performance. We further introduce a memory unit to store task-specific information extracted from the support set and used for the query set to achieve explicit adaption to individual tasks. By episodic training, the model learns to acquire the capability of adapting to specific tasks, which guarantees its performance on new related tasks. We conduct extensive experiments on widely-used benchmarks for few-shot recognition. Our method achieves new state-of-the-art performance and largely surpassing previous methods by large margins. The ablation study further demonstrates the effectiveness of the proposed discriminative learning and memory unit.","{'model': 'tldr@v2.0.0', 'text': 'A new probabilistic framework that learns to fast adapt with external memory and introduces a memory unit to store task-specific information extracted from the support set and used for the query set to achieve explicit adaption to individual tasks.'}",
-who calls the shots? rethinking few-shot learning for audio,Yu Wang,"Few-shot learning aims to train models that can recognize novel classes given just a handful of labeled examples, known as the support set. While the field has seen notable advances in recent years, they have often focused on multi-class image classification. Audio, in contrast, is often multi-label due to overlapping sounds, resulting in unique properties such as polyphony and signal-to-noise ratios (SNR). This leads to unanswered questions concerning the impact such audio properties may have on few-shot learning system design, performance, and human-computer interaction, as it is typically up to the user to collect and provide inference-time support set examples. We address these questions through a series of experiments designed to elucidate the answers to these questions. We introduce two novel datasets, FSD-MIX-CLIPS and FSD-MIX-SED, whose programmatic generation allows us to explore these questions systematically. Our experiments lead to audio-specific insights on few-shot learning, some of which are at odds with recent findings in the image domain: there is no best one-size- fits-all model, method, and support set selection criterion. Rather, it depends on the expected application scenario. Our code and data are available at https://github.com/wangyu/rethink-audio-fsl.","{'model': 'tldr@v2.0.0', 'text': 'A series of experiments lead to audio-specific insights on few-shot learning, some of which are at odds with recent findings in the image domain: there is no best one-size- fits-all model, method, and support set selection criterion, and it depends on the expected application scenario.'}",https://arxiv.org/pdf/2110.09600
-tensor feature hallucination for few-shot learning,Michalis Lazarou,"Few-shot learning addresses the challenge of learning how to address novel tasks given not just limited supervision but limited data as well. An attractive solution is synthetic data generation. However, most such methods are overly sophisticated, focusing on high-quality, realistic data in the input space. It is unclear whether adapting them to the few-shot regime and using them for the downstream task of classification is the right approach. Previous works on synthetic data generation for few-shot classification focus on exploiting complex models, e.g. a Wasserstein GAN with multiple regularizers or a network that transfers latent diversities from known to novel classes.We follow a different approach and investigate how a simple and straightforward synthetic data generation method can be used effectively. We make two contributions, namely we show that: (1) using a simple loss function is more than enough for training a feature generator in the few-shot setting; and (2) learning to generate tensor features instead of vector features is superior. Extensive experiments on miniImagenet, CUB and CIFAR-FS datasets show that our method sets a new state of the art, outperforming more sophisticated few-shot data augmentation methods. The source code can be found at https://github.com/MichalisLazarou/TFH_fewshot.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that: (1) using a simple loss function is more than enough for training a feature generator in the few-shot setting; and (2) learning to generate tensor features instead of vector features is superior.'}",https://arxiv.org/pdf/2106.05321
-unsupervised meta-learning for few-shot learning,Hui Xu,,"{'model': 'tldr@v2.0.0', 'text': 'An unsupervised meta-learning algorithm that learns from an unlabeled dataset and adapts to downstream human-specific tasks with few labeled data and constructs tasks using clustering embedding methods and data augmentation functions to satisfy two critical class distinction requirements is proposed.'}",
-laso: label-set operations networks for multi-label few-shot learning,Amit Alfassy,"Example synthesis is one of the leading methods to tackle the problem of few-shot learning, where only a small number of samples per class are available. However, current synthesis approaches only address the scenario of a single category label per image. In this work, we propose a novel technique for synthesizing samples with multiple labels for the (yet unhandled) multi-label few-shot classification scenario. We propose to combine pairs of given examples in feature space, so that the resulting synthesized feature vectors will correspond to examples whose label sets are obtained through certain set operations on the label sets of the corresponding input pairs. Thus, our method is capable of producing a sample containing the intersection, union or set-difference of labels present in two input samples. As we show, these set operations generalize to labels unseen during training. This enables performing augmentation on examples of novel categories, thus, facilitating multi-label few-shot classifier learning. We conduct numerous experiments showing promising results for the label-set manipulation capabilities of the proposed approach, both directly (using the classification and retrieval metrics), and in the context of performing data augmentation for multi-label few-shot learning. We propose a benchmark for this new and challenging task and show that our method compares favorably to all the common baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel technique for synthesizing samples with multiple labels for the (yet unhandled) multi-label few-shot classification scenario, and proposes to combine pairs of given examples in feature space so that the resulting synthesized feature vectors will correspond to examples whose label sets are obtained through certain set operations on the label sets of the corresponding input pairs.'}",http://arxiv.org/pdf/1902.09811
-an ensemble of epoch-wise empirical bayes for few-shot learning,Yaoyao Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to meta-learn the ensemble of epoch-wise empirical Bayes models (E3BM) to achieve robust predictions and introduces four kinds of hyperprior learners by considering inductive vs. transductive, and epoch-dependent vs. epoch-independent, in the paradigm of meta-learning.'}",
-few-shot learning with global class representations,Tiange Luo,"In this paper, we propose to tackle the challenging few-shot learning (FSL) problem by learning global class representations using both base and novel class training samples. In each training episode, an episodic class mean computed from a support set is registered with the global representation via a registration module. This produces a registered global class representation for computing the classification loss using a query set. Though following a similar episodic training pipeline as existing meta learning based approaches, our method differs significantly in that novel class training samples are involved in the training from the beginning. To compensate for the lack of novel class training samples, an effective sample synthesis strategy is developed to avoid overfitting. Importantly, by joint base-novel class training, our approach can be easily extended to a more practical yet challenging FSL setting, i.e., generalized FSL, where the label space of test data is extended to both base and novel classes. Extensive experiments show that our approach is effective for both of the two FSL settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to tackle the challenging few-shot learning (FSL) problem by learning global class representations using both base and novel class training samples, and an effective sample synthesis strategy is developed to avoid overfitting.'}",https://arxiv.org/pdf/1908.05257
-meta-learning of neural architectures for few-shot learning,T. Elsken,"The recent progress in neural architecture search (NAS) has allowed scaling the automated design of neural architectures to real-world domains, such as object detection and semantic segmentation. However, one prerequisite for the application of NAS are large amounts of labeled data and compute resources. This renders its application challenging in few-shot learning scenarios, where many related tasks need to be learned, each with limited amounts of data and compute time. Thus, few-shot learning is typically done with a fixed neural architecture. To improve upon this, we propose MetaNAS, the first method which fully integrates NAS with gradient-based meta-learning. MetaNAS optimizes a meta-architecture along with the meta-weights during meta-training. During meta-testing, architectures can be adapted to a novel task with a few steps of the task optimizer, that is: task adaptation becomes computationally cheap and requires only little data per task. Moreover, MetaNAS is agnostic in that it can be used with arbitrary model-agnostic meta-learning algorithms and arbitrary gradient-based NAS methods. Empirical results on standard few-shot classification benchmarks show that MetaNAS with a combination of DARTS and REPTILE yields state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MetaNAS is the first method which fully integrates NAS with gradient-based meta-learning, and optimizes a meta-architecture along with the meta-weights during meta-training, which can be adapted to a novel task with a few steps of the task optimizer.'}",https://arxiv.org/pdf/1911.11090
-"on episodes, prototypical networks, and few-shot learning",Steinar Laenen,"Episodic learning is a popular practice among researchers and practitioners interested in few-shot learning. It consists of organising training in a series of learning problems, each relying on small ""support"" and ""query"" sets to mimic the few-shot circumstances encountered during evaluation. In this paper, we investigate the usefulness of episodic learning in Prototypical Networks and Matching Networks, two of the most popular algorithms making use of this practice. Surprisingly, in our experiments we found that, for Prototypical and Matching Networks, it is detrimental to use the episodic learning strategy of separating training samples between support and query set, as it is a data-inefficient way to exploit training batches. These ""non-episodic"" variants, which are closely related to the classic Neighbourhood Component Analysis, reliably improve over their episodic counterparts in multiple datasets, achieving an accuracy that (in the case of Prototypical Networks) is competitive with the state-of-the-art, despite being extremely simple.","{'model': 'tldr@v2.0.0', 'text': 'Surprisingly, it is detrimental to use the episodic learning strategy of separating training samples between support and query set, as it is a data-inefficient way to exploit training batches in Prototypical Networks and Matching Networks.'}",
-csi-based human activity recognition with graph few-shot learning,Yong Zhang,"Human activity recognition (HAR) based on channel state information (CSI) plays an increasingly important role in the research of human–computer interaction. Many CSI HAR models based on traditional machine learning methods and deep learning methods have encountered two challenges. A lot of CSI activity data is needed to train the HAR models, which is time consuming. When the indoor environment or scene changes, the recognition accuracy of the model drops significantly, so it is necessary to recollect data to train the model. The existing few-shot learning-based method can solve the above problems to some extent, but when there are more kinds of new activities or fewer shots, the recognition accuracy will decrease significantly. In this article, considering the relationship between various activity data, a graph-based few-shot learning method with dual attention mechanism (CSI-GDAM) is proposed to perform CSI-based HAR. The model uses a feature extraction layer, including the convolutional block attention module (CBAM), to extract activity-related information in CSI data. The difference and inner product of the feature vector of the CSI activity samples are used to realize the graph convolutional network with a graph attention mechanism. The experiments proved that under the learning task of recognizing new activities in the new environment, the recognition accuracy rates reached 99.74% and 98.42% in the 5-way 5-shot and 5-way 1-shot cases, respectively. The proposed method is also compared with other few-shot learning and transfer learning methods.","{'model': 'tldr@v2.0.0', 'text': 'A graph-based few-shot learning method with dual attention mechanism (CSI-GDAM) is proposed to perform CSI-based HAR, which uses a feature extraction layer, including the convolutional block attention module (CBAM), to extract activity-related information in CSI data.'}",
-parameterless transductive feature re-representation for few-shot learning,Wentao Cui,"Recent literature in few-shot learning (FSL) has shown that transductive methods often outperform their inductive counterparts. However, most transductive solutions, particularly the meta-learning based ones, require inserting trainable parameters on top of some inductive baselines to facilitate transduction. In this paper, we propose a parameterless transductive feature re-representation framework that differs from all existing solutions from the following perspectives. (1) It is widely compatible with existing FSL methods, including meta-learning and fine tuning based models. (2) The framework is simple and introduces no extra training parameters when applied to any architecture. We conduct experiments on three benchmark datasets by applying the framework to both representative meta-learning baselines and state-of-the-art FSL methods. Our framework consistently improves performances in all experiments and refreshes the state-of-the-art FSL results.","{'model': 'tldr@v2.0.0', 'text': 'A parameterless transductive feature re-representation framework that is widely compatible with existing FSL methods, including meta-learning and tuning based models and refreshes the state-of-the-art FSL results.'}",
-transductive information maximization for few-shot learning,Malik Boudiaf,"We introduce Transductive Infomation Maximization (TIM) for few-shot learning. Our method maximizes the mutual information between the query features and their label predictions for a given few-shot task, in conjunction with a supervision loss based on the support set. Furthermore, we propose a new alternating-direction solver for our mutual-information loss, which substantially speeds up transductive-inference convergence over gradient-based optimization, while yielding similar accuracy. TIM inference is modular: it can be used on top of any base-training feature extractor. Following standard transductive few-shot settings, our comprehensive experiments demonstrate that TIM outperforms state-of-the-art methods significantly across various datasets and networks, while used on top of a fixed feature extractor trained with simple cross-entropy on the base classes, without resorting to complex meta-learning schemes. It consistently brings between 2% and 5% improvement in accuracy over the best performing method, not only on all the well-established few-shot benchmarks but also on more challenging scenarios,with domain shifts and larger numbers of classes.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Transductive Infomation Maximization (TIM) for few-shot learning, and proposes a new alternating-direction solver for the mutual-information loss, which substantially speeds up transductive-inference convergence over gradient-based optimization, while yielding similar accuracy.'}",
-self-supervised knowledge distillation for few-shot learning,Jathushan Rajasegaran,"Real-world contains an overwhelmingly large number of object classes, learning all of which at once is infeasible. Few shot learning is a promising learning paradigm due to its ability to learn out of order distributions quickly with only a few samples. Recent works [7, 41] show that simply learning a good feature embedding can outperform more sophisticated meta-learning and metric learning algorithms for few-shot learning. In this paper, we propose a simple approach to improve the representation capacity of deep neural networks for few-shot learning tasks. We follow a two-stage learning process: First, we train a neural network to maximize the entropy of the feature embedding, thus creating an optimal output manifold using a self-supervised auxiliary loss. In the second stage, we minimize the entropy on feature embedding by bringing self-supervised twins together, while constraining the manifold with student-teacher distillation. Our experiments show that, even in the first stage, self-supervision can outperform current state-of-the-art methods, with further gains achieved by our second stage distillation process. Our codes are available at: this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple approach to improve the representation capacity of deep neural networks for few-shot learning tasks, and shows that, even in the first stage, self-supervision can outperform current state-of-the-art methods.'}",
-extensively matching for few-shot learning event detection,Viet Dac Lai,"Current event detection models under supervised learning settings fail to transfer to new event types. Few-shot learning has not been explored in event detection even though it allows a model to perform well with high generalization on new event types. In this work, we formulate event detection as a few-shot learning problem to enable to extend event detection to new event types. We propose two novel loss factors that matching examples in the support set to provide more training signals to the model. Moreover, these training signals can be applied in many metric-based few-shot learning models. Our extensive experiments on the ACE-2005 dataset (under a few-shot learning setting) show that the proposed method can improve the performance of few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work formulate event detection as a few-shot learning problem to enable to extend event detection to new event types and proposes two novel loss factors that matching examples in the support set to provide more training signals to the model.'}",https://www.aclweb.org/anthology/2020.nuse-1.5.pdf
-self-supervised few-shot learning on point clouds,Charu Sharma,"The increased availability of massive point clouds coupled with their utility in a wide variety of applications such as robotics, shape synthesis, and self-driving cars has attracted increased attention from both industry and academia. Recently, deep neural networks operating on labeled point clouds have shown promising results on supervised learning tasks like classification and segmentation. However, supervised learning leads to the cumbersome task of annotating the point clouds. To combat this problem, we propose two novel self-supervised pre-training tasks that encode a hierarchical partitioning of the point clouds using a cover-tree, where point cloud subsets lie within balls of varying radii at each level of the cover-tree. Furthermore, our self-supervised learning network is restricted to pre-train on the support set (comprising of scarce training examples) used to train the downstream network in a few-shot learning (FSL) setting. Finally, the fully-trained self-supervised network's point embeddings are input to the downstream task's network. We present a comprehensive empirical evaluation of our method on both downstream classification and segmentation tasks and show that supervised methods pre-trained with our self-supervised learning method significantly improve the accuracy of state-of-the-art methods. Additionally, our method also outperforms previous unsupervised methods in downstream classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive empirical evaluation of the self-supervised learning method is presented and it is shown that supervised methods pre-trained with the method significantly improve the accuracy of state-of-the-art methods and also outperforms previous unsupervised methods in downstream classification tasks.'}",
-model-agnostic boundary-adversarial sampling for test-time generalization in few-shot learning,Jaekyeom Kim,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a modelagnostic method that improves the test-time performance of any few-shot learning models with no additional training, and thus is free from the training-test domain gap.'}",
-prototype completion with primitive knowledge for few-shot learning,Baoquan Zhang,"Few-shot learning is a challenging task, which aims to learn a classifier for novel classes with few examples. Pre-training based meta-learning methods effectively tackle the problem by pre-training a feature extractor and then fine-tuning it through the nearest centroid based meta-learning. However, results show that the fine-tuning step makes very marginal improvements. In this paper, 1) we figure out the key reason, i.e., in the pre-trained feature space, the base classes already form compact clusters while novel classes spread as groups with large variances, which implies that fine-tuning the feature extractor is less meaningful; 2) instead of fine-tuning the feature extractor, we focus on estimating more representative prototypes during meta-learning. Consequently, we propose a novel prototype completion based meta-learning framework. This framework first introduces primitive knowledge (i.e., class-level part or attribute annotations) and extracts representative attribute features as priors. Then, we design a prototype completion network to learn to complete prototypes with these priors. To avoid the prototype completion error caused by primitive knowledge noises or class differences, we further develop a Gaussian based prototype fusion strategy that combines the mean-based and completed prototypes by exploiting the unlabeled samples. Extensive experiments show that our method: (i) can obtain more accurate prototypes; (ii) out-performs state-of-the-art techniques by 2%~9% in terms of classification accuracy. Our code is available online1.","{'model': 'tldr@v2.0.0', 'text': 'A novel prototype completion based meta-learning framework that first introduces primitive knowledge and extracts representative attribute features as priors, and designs a prototype completion network to learn to complete prototypes with these priors.'}",https://arxiv.org/pdf/2009.04960
-xtarnet: learning to extract task-adaptive representation for incremental few-shot learning,Sung Whan Yoon,"Learning novel concepts while preserving prior knowledge is a long-standing challenge in machine learning. The challenge gets greater when a novel task is given with only a few labeled examples, a problem known as incremental few-shot learning. We propose XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning. The method utilizes a backbone network pretrained on a set of base categories while also employing additional modules that are meta-trained across episodes. Given a new task, the novel feature extracted from the meta-trained modules is mixed with the base feature obtained from the pretrained model. The process of combining two different features provides TAR and is also controlled by meta-trained modules. The TAR contains effective information for classifying both novel and base categories. The base and novel classifiers quickly adapt to a given task by utilizing the TAR. Experiments on standard image datasets indicate that XtarNet achieves state-of-the-art incremental few-shot learning performance. The concept of TAR can also be used in conjunction with existing incremental few-shot learning methods; extensive simulation results in fact show that applying TAR enhances the known methods significantly.","{'model': 'tldr@v2.0.0', 'text': 'XtarNet, which learns to extract task-adaptive representation (TAR) for facilitating incremental few-shot learning, utilizes a backbone network pretrained on a set of base categories while also employing additional modules that are meta-trained across episodes.'}",
-transductive relation-propagation with decoupling training for few-shot learning,Yuqing Ma,"Few-shot learning, aiming to learn novel concepts from one or a few labeled examples, is an interesting and very challenging problem with many practical advantages. Existing few-shot methods usually utilize data of the same classes to train the feature embedding module and in a row, which is unable to learn adapting to new tasks. Besides, traditional few-shot models fail to take advantage of the valuable relations of the support-query pairs, leading to performance degradation. In this article, we propose a transductive relation-propagation graph neural network (GNN) with a decoupling training strategy (TRPN-D) to explicitly model and propagate such relations across support-query pairs, and empower the few-shot module the ability of transferring past knowledge to new tasks via the decoupling training. Our few-shot module, namely TRPN, treats the relation of each support-query pair as a graph node, named relational node, and resorts to the known relations between support samples, including both intraclass commonality and interclass uniqueness. Through relation propagation, the model could generate the discriminative relation embeddings for support-query pairs. To the best of our knowledge, this is the first work that decouples the training of the embedding network and the few-shot graph module with different tasks, which might offer a new way to solve the few-shot learning problem. Extensive experiments conducted on several benchmark datasets demonstrate that our method can significantly outperform a variety of state-of-the-art few-shot learning methods.","{'model': 'tldr@v2.0.0', 'text': 'A transductive relation-propagation graph neural network (GNN) with a decoupling training strategy (TRPN-D) to explicitly model and propagate such relations across support-query pairs, and empower the few-shot module the ability of transferring past knowledge to new tasks via the decoupled training.'}",
-few-shot learning with metric-agnostic conditional embeddings,Nathan Hilliard,"Learning high quality class representations from few examples is a key problem in metric-learning approaches to few-shot learning. To accomplish this, we introduce a novel architecture where class representations are conditioned for each few-shot trial based on a target image. We also deviate from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison. This allows the network to decide what aspects of each class are important for the comparison at hand. We find that this flexible architecture works well in practice, achieving state-of-the-art performance on the Caltech-UCSD birds fine-grained classification task.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel architecture where class representations are conditioned for each few-shot trial based on a target image, and deviates from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison.'}",
-multimodality helps unimodality: cross-modal few-shot learning with multimodal models,Zhiqiu Lin,"The ability to quickly learn a new task with minimal instruction - known as few-shot learning - is a central aspect of intelligent agents. Classical few-shot benchmarks make use of few-shot samples from a single modality, but such samples may not be sufficient to characterize an entire concept class. In contrast, humans use cross-modal information to learn new concepts efficiently. In this work, we demonstrate that one can indeed build a better visual dog classifier by reading about dogs and listening to them bark. To do so, we exploit the fact that recent multimodal foundation models such as CLIP are inherently cross-modal, mapping different modalities to the same representation space. Specifically, we propose a simple cross-modal adaptation approach that learns from few-shot examples spanning different modalities. By repurposing class names as additional one-shot training samples, we achieve SOTA results with an embarrassingly simple linear classifier for vision-language adaptation. Furthermore, we show that our approach can benefit existing methods such as prefix tuning, adapters, and classifier ensembling. Finally, to explore other modalities beyond vision and language, we construct the first (to our knowledge) audiovisual few-shot benchmark and use cross-modal training to improve the performance of both image and audio classification. Project site at link.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple cross-modal adaptation approach that learns from few-shot examples spanning different modalities, and achieves SOTA results with an embarrassingly simple linear classifier for vision-language adaptation.'}",https://arxiv.org/pdf/2301.06267
-multimodal prototypical networks for few-shot learning,Frederik Pahde,"Although providing exceptional results for many computer vision tasks, state-of-the-art deep learning algorithms catastrophically struggle in low data scenarios. However, if data in additional modalities exist (e.g. text) this can compensate for the lack of data and improve the classification results. To overcome this data scarcity, we design a cross-modal feature generation framework capable of enriching the low populated embedding space in few-shot scenarios, leveraging data from the auxiliary modality. Specifically, we train a generative model that maps text data into the visual feature space to obtain more reliable prototypes. This allows to exploit data from additional modalities (e.g. text) during training while the ultimate task at test time remains classification with exclusively visual data. We show that in such cases nearest neighbor classification is a viable approach and outperform state-of-the-art single-modal and multimodal few-shot learning methods on the CUB-200 and Oxford-102 datasets.","{'model': 'tldr@v2.0.0', 'text': 'A generative model is trained that maps text data into the visual feature space to obtain more reliable prototypes and shows that in such cases nearest neighbor classification is a viable approach and outperform state-of-the-art single-modal and multimodal few-shot learning methods on the CUB-200 and Oxford-102 datasets.'}",http://arxiv.org/pdf/2011.08899
-few-shot learning for decoding surface electromyography for hand gesture recognition,E. Rahimian,"This work is motivated by the recent advancements of Deep Neural Networks (DNNs) for myoelectric prosthesis control. In this regard, hand gesture recognition via surface Electromyogram (sEMG) signals has shown a high potential for improving the performance of myoelectric control prostheses. Although the recent researches in hand gesture recognition with DNNs have achieved promising results, they are still in their infancy. The recent literature uses traditional supervised learning methods that usually have poor performance if a small amount of data is available or requires adaptation to a changing task. Therefore, in this work, we develop a novel hand gesture recognition framework based on the formulation of FewShot Learning (FSL) to infer the required output given only one or a few numbers of training examples. Thus in this paper, we proposed a new architecture (named as FHGR which refers to ""Few-shot Hand Gesture Recognition"") that learns the mapping using a small number of data and quickly adapts to a new user/gesture by combing its prior experience. The proposed approach led to 83.99% classification accuracy on new repetitions with few-shot observations, 76.39% accuracy on new subjects with few-shot observations, and 72.19% accuracy on new gestures with few-shot observations.","{'model': 'tldr@v2.0.0', 'text': 'A novel hand gesture recognition framework based on the formulation of FewShot Learning (FSL), that learns the mapping using a small number of data and quickly adapts to a new user/gesture by combing its prior experience is proposed.'}",
-the unreasonable effectiveness of few-shot learning for machine translation,Xavier García,"We demonstrate the potential of few-shot translation systems, trained with unpaired language data, for both high and low-resource language pairs. We show that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able to match specialized supervised state-of-the-art models as well as more general commercial translation systems. In particular, we outperform the best performing system on the WMT'21 English - Chinese news translation task by only using five examples of English - Chinese parallel data at inference. Moreover, our approach in building these models does not necessitate joint multilingual training or back-translation, is conceptually simple and shows the potential to extend to the multilingual setting. Furthermore, the resulting models are two orders of magnitude smaller than state-of-the-art language models. We then analyze the factors which impact the performance of few-shot translation systems, and highlight that the quality of the few-shot demonstrations heavily determines the quality of the translations generated by our models. Finally, we show that the few-shot paradigm also provides a way to control certain attributes of the translation -- we show that we are able to control for regional varieties and formality using only a five examples at inference, paving the way towards controllable machine translation systems.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able to match specialized supervised state-of-the-art models as well as more general commercial translation systems.'}",http://arxiv.org/pdf/2302.01398
-task cooperation for semi-supervised few-shot learning,Han Ye,"Training a model with limited data is an essential task for machine learning and visual recognition. Few-shot learning approaches meta-learn a task-level inductive bias from SEEN class few-shot tasks, and the meta-model is expected to facilitate the few-shot learning with UNSEEN classes. Inspired by the idea that unlabeled data can be utilized to smooth the model space in traditional semi-supervised learning, we propose TAsk COoperation (TACO) which takes advantage of unsupervised tasks to smooth the meta-model space. Specifically, we couple the labeled support set in a few-shot task with easily-collected unlabeled instances, prediction agreement on which encodes the relationship between tasks. The learned smooth meta-model promotes the generalization ability on supervised UNSEEN few-shot tasks. The state-of-the-art few-shot classification results on MiniImageNet and TieredImageNet verify the superiority of TACO to leverage unlabeled data and task relationship in meta-learning.","{'model': 'tldr@v2.0.0', 'text': 'This work couple the labeled support set in a few-shot task with easily-collected unlabeled instances, prediction agreement on which encodes the relationship between tasks, and learns smooth meta-model which promotes the generalization ability on supervised UNSEEN few- shot tasks.'}",https://ojs.aaai.org/index.php/AAAI/article/download/17277/17084
-augmenting few-shot learning with supervised contrastive learning,Taemin Lee,"Few-shot learning deals with a small amount of data which incurs insufficient performance with conventional cross-entropy loss. We propose a pretraining approach for few-shot learning scenarios. That is, considering that the feature extractor quality is a critical factor in few-shot learning, we augment the feature extractor using a contrastive learning technique. It is reported that supervised contrastive learning applied to base class training in transductive few-shot training pipeline leads to improved results, outperforming the state-of-the-art methods on Mini-ImageNet and CUB. Furthermore, our experiment shows that a much larger dataset is needed to retain few-shot classification accuracy when domain-shift degradation exists, and if our method is applied, the need for a large dataset is eliminated. The accuracy gain can be translated to a runtime reduction of $3.87\times $ in a resource-constrained environment.","{'model': 'tldr@v2.0.0', 'text': 'It is reported that supervised contrastive learning applied to base class training in transductive few-shot training pipeline leads to improved results, outperforming the state-of-the-art methods on Mini-ImageNet and CUB.'}",https://ieeexplore.ieee.org/ielx7/6287639/9312710/09409075.pdf
-consistent meta-regularization for better meta-knowledge in few-shot learning,Pinzhuo Tian,"Recently, meta-learning provides a powerful paradigm to deal with the few-shot learning problem. However, existing meta-learning approaches ignore the prior fact that good meta-knowledge should alleviate the data inconsistency between training and test data, caused by the extremely limited data, in each few-shot learning task. Moreover, legitimately utilizing the prior understanding of meta-knowledge can lead us to design an efficient method to improve the meta-learning model. Under this circumstance, we consider the data inconsistency from the distribution perspective, making it convenient to bring in the prior fact, and propose a new consistent meta-regularization (Con-MetaReg) to help the meta-learning model learn how to reduce the data-distribution discrepancy between the training and test data. In this way, the ability of meta-knowledge on keeping the training and test data consistent is enhanced, and the performance of the meta-learning model can be further improved. The extensive analyses and experiments demonstrate that our method can indeed improve the performances of different meta-learning models in few-shot regression, classification, and fine-grained classification.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the data inconsistency from the distribution perspective, making it convenient to bring in the prior fact, and proposes a new consistent meta-regularization (Con-MetaReg) to help the meta-learning model learn how to reduce the data-distribution discrepancy between the training and test data.'}",
-few-shot learning with class imbalance,Mateusz Ochal,"Few-shot learning (FSL) algorithms are commonly trained through meta-learning (ML), which exposes models to batches of tasks sampled from a meta-dataset to mimic tasks seen during evaluation. However, the standard training procedures overlook the real-world dynamics where classes commonly occur at different frequencies. While it is generally understood that class imbalance harms the performance of supervised methods, limited research examines the impact of imbalance on the FSL evaluation task. Our analysis compares ten state-of-the-art ML and FSL methods on different imbalance distributions and rebalancing techniques. Our results reveal that: 1) some FSL methods display a natural disposition against imbalance while most other approaches produce a performance drop by up to 17% compared to the balanced task without the appropriate mitigation; 2) many ML algorithms will not automatically learn to balance from exposure to imbalanced training tasks; 3) classical rebalancing strategies, such as random oversampling, can still be very effective, leading to state-of-the-art performances and should not be overlooked.","{'model': 'tldr@v2.0.0', 'text': 'This analysis compares ten state-of-the-art ML and F SL methods on different imbalance distributions and rebalancing techniques and reveals that some FSL methods display a natural disposition against imbalance while most other approaches produce a performance drop compared to the balanced task without the appropriate mitigation.'}",https://arxiv.org/pdf/2101.02523
-prototypical siamese networks for few-shot learning,Junhua Wang,"We propose a novel architecture, called Prototypical Siamese Networks, for few-shot learning, where a classifier must generalize to new classes not seen in the training set, given only a few examples of each class. Prototypical Siamese Networks add a new module to siamese networks to learn a high quality prototypical representation of each class. Compared to recent methods for few-shot learning, our method achieves state-of-the-art performance on few-shot learning. Experiments on two benchmarks validate the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel architecture, called Prototypical Siamese Networks, for few-shot learning, where a classifier must generalize to new classes not seen in the training set, given only a few examples of each class.'}",
-fs-hgr: few-shot learning for hand gesture recognition via electromyography,E. Rahimian,"This work is motivated by the recent advances in Deep Neural Networks (DNNs) and their widespread applications in human-machine interfaces. DNNs have been recently used for detecting the intended hand gesture through the processing of surface electromyogram (sEMG) signals. Objective: Although DNNs have shown superior accuracy compared to conventional methods when large amounts of data are available for training, their performance substantially decreases when data are limited. Collecting large datasets for training may be feasible in research laboratories, but it is not a practical approach for real-life applications. The main objective of this work is to design a modern DNN-based gesture detection model that relies on minimal training data while providing high accuracy. Methods: We propose the novel Few-Shot learning- Hand Gesture Recognition (FS-HGR) architecture. Few-shot learning is a variant of domain adaptation with the goal of inferring the required output based on just one or a few training observations. The proposed FS-HGR generalizes after seeing very few observations from each class by combining temporal convolutions with attention mechanisms. This allows the meta-learner to aggregate contextual information from experience and to pinpoint specific pieces of information within its available set of inputs. Data Source & Summary of Results: The performance of FS-HGR was tested on the second and fifth Ninapro databases, referred to as the DB2 and DB5, respectively. The DB2 consists of 50 gestures (rest included) from 40 healthy subjects. The Ninapro DB5 contains data from 10 healthy participants performing a total of 53 different gestures (rest included). The proposed approach for the Ninapro DB2 led to 85.94% classification accuracy on new repetitions with few-shot observation (5-way 5-shot), 81.29% accuracy on new subjects with few-shot observation (5-way 5-shot), and 73.36% accuracy on new gestures with few-shot observation (5-way 5-shot). Moreover, the proposed approach for the Ninapro DB5 led to 64.65% classification accuracy on new subjects with few-shot observation (5-way 5-shot).","{'model': 'tldr@v2.0.0', 'text': 'The main objective of this work is to design a modern DNN-based gesture detection model that relies on minimal training data while providing high accuracy and is motivated by the recent advances in Deep Neural Networks and their widespread applications in human-machine interfaces.'}",
-semi-supervised few-shot learning for medical image segmentation,A. Fayjie,"Recent years have witnessed the great progress of deep neural networks on semantic segmentation, particularly in medical imaging. Nevertheless, training high-performing models require large amounts of pixel-level ground truth masks, which can be prohibitive to obtain in the medical domain. Furthermore, training such models in a low-data regime highly increases the risk of overfitting. Recent attempts to alleviate the need for large annotated datasets have developed training strategies under the few-shot learning paradigm, which addresses this shortcoming by learning a novel class from only a few labeled examples. In this context, a segmentation model is trained on episodes, which represent different segmentation problems, each of them trained with a very small labeled dataset. In this work, we propose a novel few-shot learning framework for semantic segmentation, where unlabeled images are also made available at each episode. To handle this new learning paradigm, we propose to include surrogate tasks that can leverage very powerful supervisory signals --derived from the data itself-- for semantic feature learning. We show that including unlabeled surrogate tasks in the episodic training leads to more powerful feature representations, which ultimately results in better generability to unseen tasks. We demonstrate the efficiency of our method in the task of skin lesion segmentation in two publicly available datasets. Furthermore, our approach is general and model-agnostic, which can be combined with different deep architectures.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel few-shot learning framework for semantic segmentation, where unlabeled images are also made available at each episode, and proposes to include surrogate tasks that can leverage very powerful supervisory signals --derived from the data itself-- for semantic feature learning.'}",
-domain-adaptive few-shot learning,An Zhao,"Existing few-shot learning (FSL) methods make the implicit assumption that the few target class samples are from the same domain as the source class samples. However, in practice, this assumption is often invalid –the target classes could come from a different domain. This poses an additional challenge of domain adaptation (DA) with few training samples. In this paper, the problem of domain-adaptive few-shot learning (DA-FSL) is tackled, which is expected to have wide use in real-world scenarios and requires solving FSL and DA in a unified framework. To this end, we propose a novel domain-adversarial prototypical network (DAPN) model. It is designed to address a specific challenge in DA-FSL: the DA objective means that the source and target data distributions need to be aligned, typically through a shared domain-adaptive feature embedding space; but the FSL objective dictates that the target domain per class distribution must be different from that of any source domain class, meaning aligning the distributions across domains may harm the FSL performance. How to achieve global domain distribution alignment whilst maintaining source/target per-class discriminativeness thus becomes the key. Our solution is to explicitly enhance the source/target per-class separation before domain-adaptive feature embedding learning, to alleviate the negative effect of domain alignment on FSL. Extensive experiments show that our DAPN outperforms the state-of-the-arts. The code is available at https://github.com/dingmyu/DAPN.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel domain-adversarial prototypical network (DAPN) model, designed to address a specific challenge in DA-FSL, which dictates that the target domain per class distribution must be different from that of any source domain class, meaning aligning the distributions across domains may harm the FSL performance.'}",https://arxiv.org/pdf/2003.08626
-asymmetric distribution measure for few-shot learning,Wenbin Li,"The core idea of metric-based few-shot image classification is to directly measure the relations between query images and support classes to learn transferable feature embeddings. Previous work mainly focuses on image-level feature representations, which actually cannot effectively estimate a class's distribution due to the scarcity of samples. Some recent work shows that local descriptor based representations can achieve richer representations than image-level based representations. However, such works are still based on a less effective instance-level metric, especially a symmetric metric, to measure the relation between a query image and a support class. Given the natural asymmetric relation between a query image and a support class, we argue that an asymmetric measure is more suitable for metric-based few-shot learning. To that end, we propose a novel Asymmetric Distribution Measure (ADM) network for few-shot learning by calculating a joint local and global asymmetric measure between two multivariate local distributions of a query and a class. Moreover, a task-aware Contrastive Measure Strategy (CMS) is proposed to further enhance the measure function. On popular miniImageNet and tieredImageNet, ADM can achieve the state-of-the-art results, validating our innovative design of asymmetric distribution measures for few-shot learning. The source code can be downloaded from https://github.com/WenbinLee/ADM.git.","{'model': 'tldr@v2.0.0', 'text': 'A novel Asymmetric Distribution Measure (ADM) network for few-shot learning is proposed by calculating a joint local and global asymmetric measure between two multivariate local distributions of a query and a class.'}",https://www.ijcai.org/proceedings/2020/0409.pdf
-few-shot learning through an information retrieval lens,Eleni Triantafillou,"Few-shot learning refers to understanding new concepts from only a few examples. We propose an information retrieval-inspired approach for this problem that is motivated by the increased importance of maximally leveraging all the available information in this low-data regime. We define a training objective that aims to extract as much information as possible from each training batch by effectively optimizing over all relative orderings of the batch points simultaneously. In particular, we view each batch point as a `query' that ranks the remaining ones based on its predicted relevance to them and we define a model within the framework of structured prediction to optimize mean Average Precision over these rankings. Our method achieves impressive results on the standard few-shot classification benchmarks while is also capable of few-shot retrieval.","{'model': 'tldr@v2.0.0', 'text': 'This work defines a training objective that aims to extract as much information as possible from each training batch by effectively optimizing over all relative orderings of the batch points simultaneously and defines a model within the framework of structured prediction to optimize mean Average Precision over these rankings.'}",
-rethinking class relations: absolute-relative supervised and unsupervised few-shot learning,Hongguang Zhang,"The majority of existing few-shot learning methods describe image relations with binary labels. However, such binary relations are insufficient to teach the network complicated real-world relations, due to the lack of decision smoothness. Furthermore, current few-shot learning models capture only the similarity via relation labels, but they are not exposed to class concepts associated with objects, which is likely detrimental to the classification performance due to underutilization of the available class labels. For instance, children learn the concept of tiger from a few of actual examples as well as from comparisons of tiger to other animals. Thus, we hypothesize that both similarity and class concept learning must be occurring simultaneously. With these observations at hand, we study the fundamental problem of simplistic class modeling in current few-shot learning methods. We rethink the relations between class concepts, and propose a novel Absolute-relative Learning paradigm to fully take advantage of label information to refine the image an relation representations in both supervised and unsupervised scenarios. Our proposed paradigm improves the performance of several state-of-the-art models on publicly available datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work rethink the relations between class concepts, and proposes a novel Absolute-relative Learning paradigm to fully take advantage of label information to refine the image an relation representations in both supervised and unsupervised scenarios.'}",https://arxiv.org/pdf/2001.03919
-few-shot parameter-efficient fine-tuning is better and cheaper than in-context learning,Haokun Liu,"Few-shot in-context learning (ICL) enables pre-trained language models to perform a previously-unseen task without any gradient-based training by feeding a small number of training examples as part of the input. ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made. Parameter-efficient fine-tuning (PEFT) (e.g. adapter modules, prompt tuning, sparse update methods, etc.) offers an alternative paradigm where a small set of parameters are trained to enable a model to perform the new task. In this paper, we rigorously compare few-shot ICL and PEFT and demonstrate that the latter offers better accuracy as well as dramatically lower computational costs. Along the way, we introduce a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters. We also propose a simple recipe based on the T0 model called T-Few that can be applied to new tasks without task-specific tuning or modifications. We validate the effectiveness of T-Few on completely unseen tasks by applying it to the RAFT benchmark, attaining super-human performance for the first time and outperforming the state-of-the-art by 6% absolute. All of the code used in our experiments is publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper rigorously compares few-shot ICL and PEFT and demonstrates that the latter offers better accuracy as well as dramatically lower computational costs, and introduces a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters.'}",http://arxiv.org/pdf/2205.05638
-exploring effectiveness of gpt-3 in grammatical error correction: a study on performance and controllability in prompt-based methods,Mengsay Loem,"Large-scale pre-trained language models such as GPT-3 have shown remarkable performance across various natural language processing tasks. However, applying prompt-based methods with GPT-3 for Grammatical Error Correction (GEC) tasks and their controllability remains underexplored. Controllability in GEC is crucial for real-world applications, particularly in educational settings, where the ability to tailor feedback according to learner levels and specific error types can significantly enhance the learning process.This paper investigates the performance and controllability of prompt-based methods with GPT-3 for GEC tasks using zero-shot and few-shot setting. We explore the impact of task instructions and examples on GPT-3’s output, focusing on controlling aspects such as minimal edits, fluency edits, and learner levels. Our findings demonstrate that GPT-3 could effectively perform GEC tasks, outperforming existing supervised and unsupervised approaches. We also showed that GPT-3 could achieve controllability when appropriate task instructions and examples are given.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the performance and controllability of prompt-based methods with GPT-3 for Grammatical Error Correction tasks using zero-shot and few-shot setting and demonstrates that G PT-3 could effectively perform GEC tasks, outperforming existing supervised and unsupervised approaches.'}",http://arxiv.org/pdf/2305.18156
-knowledge injected prompt based fine-tuning for multi-label few-shot icd coding,Zhichao Yang,"Automatic International Classification of Diseases (ICD) coding aims to assign multiple ICD codes to a medical note with average length of 3,000+ tokens. This task is challenging due to a high-dimensional space of multi-label assignment (tens of thousands of ICD codes) and the long-tail challenge: only a few codes (common diseases) are frequently assigned while most codes (rare diseases) are infrequently assigned. This study addresses the long-tail challenge by adapting a prompt-based fine-tuning technique with label semantics, which has been shown to be effective under few-shot setting. To further enhance the performance in medical domain, we propose a knowledge-enhanced longformer by injecting three domain-specific knowledge: hierarchy, synonym, and abbreviation with additional pretraining using contrastive learning. Experiments on MIMIC-III-full, a benchmark dataset of code assignment, show that our proposed method outperforms previous state-of-the-art method in 14.5% in marco F1 (from 10.3 to 11.8, P<0.001). To further test our model on few-shot setting, we created a new rare diseases coding dataset, MIMIC-III-rare50, on which our model improves marco F1 from 17.1 to 30.4 and micro F1 from 17.2 to 32.6 compared to previous method.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based fine-tuning technique with label semantics, which has been shown to be effective under few-shot setting is adapted by adapting a knowledge-enhanced longformer by injecting three domain-specific knowledge: hierarchy, synonym, and abbreviation with additional pretraining using contrastive learning.'}",
-improved universal sentence embeddings with prompt-based contrastive learning and energy-based learning,Yuxin Jiang,"Contrastive learning has been demonstrated to be effective in enhancing pre-trained language models (PLMs) to derive superior universal sentence embeddings. However, existing contrastive methods still have two limitations. Firstly, previous works may acquire poor performance under domain shift settings, thus hindering the application of sentence representations in practice. We attribute this low performance to the over-parameterization of PLMs with millions of parameters. To alleviate it, we propose PromCSE (Prompt-based Contrastive Learning for Sentence Embeddings), which only trains small-scale \emph{Soft Prompt} (i.e., a set of trainable vectors) while keeping PLMs fixed. Secondly, the commonly used NT-Xent loss function of contrastive learning does not fully exploit hard negatives in supervised learning settings. To this end, we propose to integrate an Energy-based Hinge loss to enhance the pairwise discriminative power, inspired by the connection between the NT-Xent loss and the Energy-based Learning paradigm. Empirical results on seven standard semantic textual similarity (STS) tasks and a domain-shifted STS task both show the effectiveness of our method compared with the current state-of-the-art sentence embedding models. Our code is publicly avaliable at https://github.com/YJiangcm/PromCSE","{'model': 'tldr@v2.0.0', 'text': 'PromCSE (Prompt-based Contrastive Learning for Sentence Embeddings), which only trains small-scale \\emph{Soft Prompt} (i.e., a set of trainable vectors) while keeping PLMs fixed, is proposed, which shows the effectiveness of the method compared with the current state-of-the-art sentence embedding models.'}",https://aclanthology.org/2022.findings-emnlp.220.pdf
-v2p: vision-to-prompt based multi-modal product summary generation,Xuemeng Song,"Multi-modal Product Summary Generation is a new yet challenging task, which aims to generate a concise and readable summary for a product given its multi-modal content, e.g., its long text description and image. Although existing methods have achieved great success, they still suffer from three key limitations: 1) overlook the benefit of pre-training, 2) lack the representation-level supervision, and 3) ignore the diversity of the seller-generated data. To address these limitations, in this work, we propose a Vision-to-Prompt based multi-modal product summary generation framework, dubbed as V2P, where a Generative Pre-trained Language Model (GPLM) is adopted as the backbone. In particular, to maintain the original text capability of the GPLM and fully utilize the high-level concepts contained in the product image, we design V2P with two key components: vision-based prominent attribute prediction, and attribute prompt-guided summary generation. The first component works on obtaining the vital semantic attributes of the product from its image by the Swin Transformer, while the second component aims to generate the summary based on the product's long text description and the attribute prompts yielded by the first component with a GPLM. Towards comprehensive supervision over the second component, apart from the conventional output-level supervision, we introduce the representation-level regularization. Meanwhile, we design the data augmentation-based robustness regularization to handle the diverse inputs and improve the robustness of the second component. Extensive experiments on a large-scale Chinese dataset verify the superiority of our model over cutting-edge methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Vision-to-Prompt based multi-modal product summary generation framework, dubbed as V2P, where a Generative Pre-trained Language Model (GPLM) is adopted as the backbone, and introduces the representation-level regularization over the second component.'}",
-do we still need human assessors? prompt-based gpt-3 user simulation in conversational ai,Selina Meyer,"Scarcity of user data continues to be a problem in research on conversational user interfaces and often hinders or slows down technical innovation. In the past, different ways of synthetically generating data, such as data augmentation techniques have been explored. With the rise of ever improving pre-trained language models, we ask if we can go beyond such methods by simply providing appropriate prompts to these general purpose models to generate data. We explore the feasibility and cost-benefit trade-offs of using non fine-tuned synthetic data to train classification algorithms for conversational agents. We compare this synthetically generated data with real user data and evaluate the performance of classifiers trained on different combinations of synthetic and real data. We come to the conclusion that, although classifiers trained on such synthetic data perform much better than random baselines, they do not compare to the performance of classifiers trained on even very small amounts of real user data, largely because such data is lacking much of the variability found in user generated data. Nevertheless, we show that in situations where very little data and resources are available, classifiers trained on such synthetically generated data might be preferable to the collection and annotation of naturalistic data.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that in situations where very little data and resources are available, classifiers trained on such synthetically generated data might be preferable to the collection and annotation of naturalistic data.'}",https://dl.acm.org/doi/pdf/10.1145/3543829.3544529
-towards open-vocabulary scene graph generation with prompt-based finetuning,Tao He,"Scene graph generation (SGG) is a fundamental task aimed at detecting visual relations between objects in an image. The prevailing SGG methods require all object classes to be given in the training set. Such a closed setting limits the practical application of SGG. In this paper, we introduce open-vocabulary scene graph generation, a novel, realistic and challenging setting in which a model is trained on a set of base object classes but is required to infer relations for unseen target object classes. To this end, we propose a two-step method that firstly pre-trains on large amounts of coarse-grained region-caption data and then leverages two prompt-based techniques to finetune the pre-trained model without updating its parameters. Moreover, our method can support inference over completely unseen object classes, which existing methods are incapable of handling. On extensive experiments on three benchmark datasets, Visual Genome, GQA, and Open-Image, our method significantly outperforms recent, strong SGG methods on the setting of Ov-SGG, as well as on the conventional closed SGG.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a two-step method that firstly pre-trains on large amounts of coarse-grained region-caption data and then leverages two prompt-based techniques to finetune the pre-trained model without updating its parameters, which can support inference over completely unseen object classes, which existing methods are incapable of handling.'}",http://arxiv.org/pdf/2208.08165
-zero-shot cross-lingual transfer of prompt-based tuning with a unified multilingual prompt,Lianzhe Huang,"Prompt-based tuning has been proven effective for pretrained language models (PLMs). While most of the existing work focuses on the monolingual prompts, we study the multilingual prompts for multilingual PLMs, especially in the zero-shot cross-lingual setting. To alleviate the effort of designing different prompts for multiple languages, we propose a novel model that uses a unified prompt for all languages, called UniPrompt. Different from the discrete prompts and soft prompts, the unified prompt is model-based and language-agnostic. Specifically, the unified prompt is initialized by a multilingual PLM to produce language-independent representation, after which is fused with the text input. During inference, the prompts can be pre-computed so that no extra computation cost is needed. To collocate with the unified prompt, we propose a new initialization method for the target label word to further improve the model’s transferability across languages. Extensive experiments show that our proposed methods can significantly outperform the strong baselines across different languages. We release data and code to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'A novel model that uses a unified prompt for all languages, called UniPrompt, which is model-based and language-agnostic and can significantly outperform the strong baselines across different languages.'}",https://aclanthology.org/2022.emnlp-main.790.pdf
-promptcast: a new prompt-based learning paradigm for time series forecasting,Hao Xue,"This paper presents a new perspective on time series forecasting. In existing time series forecasting methods, the models take a sequence of numerical values as input and yield numerical values as output. The existing SOTA models are largely based on the Transformer architecture, modified with multiple encoding mechanisms to incorporate the context and semantics around the historical data. Inspired by the successes of pre-trained language foundation models, we pose a question about whether these models can also be adapted to solve time-series forecasting. Thus, we propose a new forecasting paradigm: prompt-based time series forecasting (PromptCast). In this novel task, the numerical input and output are transformed into prompts and the forecasting task is framed in a sentence-to-sentence manner, making it possible to directly apply language models for forecasting purposes. To support and facilitate the research of this task, we also present a large-scale dataset (PISA) that includes three real-world forecasting scenarios. We evaluate different SOTA numerical-based forecasting methods and language generation models. The benchmark results with various forecasting settings demonstrate the proposed PromptCast with language generation models is a promising research direction. Additionally, in comparison to conventional numerical-based forecasting, PromptCast shows a much better generalization ability under the zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'The benchmark results with various forecasting settings demonstrate the proposed PromptCast with language generation models is a promising research direction, and in comparison to conventional numerical-based forecasting, PromptCast shows a much better generalization ability under the zero-shot setting.'}",
-lego-absa: a prompt-based task assemblable unified generative framework for multi-task aspect-based sentiment analysis,Tianhao Gao,"Aspect-based sentiment analysis (ABSA) has received increasing attention recently. ABSA can be divided into multiple tasks according to the different extracted elements. Existing generative methods usually treat the output as a whole string rather than the combination of different elements and only focus on a single task at once. This paper proposes a unified generative multi-task framework that can solve multiple ABSA tasks by controlling the type of task prompts consisting of multiple element prompts. Further, the proposed approach can train on simple tasks and transfer to difficult tasks by assembling task prompts, like assembling Lego bricks. We conduct experiments on six ABSA tasks across multiple benchmarks. Our proposed multi-task approach achieves new state-of-the-art results in almost all tasks and competitive results in task transfer scenarios.","{'model': 'tldr@v2.0.0', 'text': 'A unified generative multi-task framework that can solve multiple ABSA tasks by controlling the type of task prompts consisting of multiple element prompts and achieves new state-of-the-art results in almost all tasks and competitive results in task transfer scenarios.'}",
-prompt-based connective prediction method for fine-grained implicit discourse relation recognition,Hao Zhou,"Due to the absence of connectives, implicit discourse relation recognition (IDRR) is still a challenging and crucial task in discourse analysis. Most of the current work adopted multi-task learning to aid IDRR through explicit discourse relation recognition (EDRR) or utilized dependencies between discourse relation labels to constrain model predictions. But these methods still performed poorly on fine-grained IDRR and even utterly misidentified on most of the few-shot discourse relation classes. To address these problems, we propose a novel Prompt-based Connective Prediction (PCP) method for IDRR. Our method instructs large-scale pre-trained models to use knowledge relevant to discourse relation and utilizes the strong correlation between connectives and discourse relation to help the model recognize implicit discourse relations. Experimental results show that our method surpasses the current state-of-the-art model and achieves significant improvements on those fine-grained few-shot discourse relation. Moreover, our approach is able to be transferred to EDRR and obtain acceptable results. Our code is released in https://github.com/zh-i9/PCP-for-IDRR.","{'model': 'tldr@v2.0.0', 'text': 'A novel Prompt-based Connective Prediction (PCP) method for IDRR that surpasses the current state-of-the-art model and achieves significant improvements on those fine-grained few-shot discourse relation classes.'}",http://arxiv.org/pdf/2210.07032
-prompt-based distribution alignment for domain generalization in text classification,Chen Jia,"Prompt-based learning (a.k.a. prompting) achieves high performance by bridging the gap between the objectives of language modeling and downstream tasks. Domain generalization ability can be improved by prompting since classification across different domains can be unified into the prediction of the same set of label words. The remaining challenge for domain generalization by prompting comes from discrepancies between the data distribution of different domains. To improve domain generalization with prompting, we learn distributional invariance across source domains via two alignment regularization loss functions. The first is vocabulary distribution alignment, which uses a Kullback-Leibler divergence regularization on source-domain vocabulary distributions. The second is feature distribution alignment, which uses a novel adversarial training strategy to learn domain invariant representation across source domains. Experiments on sentiment analysis and natural language inference show the effectiveness of our method and achieve state-of-the-art results on six datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work learns distributional invariance across source domains via two alignment regularization loss functions via a Kullback-Leibler divergence regularization on source-domain vocabulary distributions.'}",https://aclanthology.org/2022.emnlp-main.690.pdf
-context variance evaluation of pretrained language models for prompt-based biomedical knowledge probing,Zonghai Yao,"Pretrained language models (PLMs) have motivated research on what kinds of knowledge these models learn. Fill-in-the-blanks problem (e.g., cloze tests) is a natural approach for gauging such knowledge. BioLAMA generates prompts for biomedical factual knowledge triples and uses the Top-k accuracy metric to evaluate different PLMs' knowledge. However, existing research has shown that such prompt-based knowledge probing methods can only probe a lower bound of knowledge. Many factors like prompt-based probing biases make the LAMA benchmark unreliable and unstable. This problem is more prominent in BioLAMA. The severe long-tailed distribution in vocabulary and large-N-M relation make the performance gap between LAMA and BioLAMA remain notable. To address these, we introduced context variance into the prompt generation and proposed a new rank-change-based evaluation metric. Different from the previous known-unknown evaluation criteria, we proposed the concept of ""Misunderstand"" in LAMA for the first time. Through experiments on 12 PLMs, we showed that our context variance prompts and Understand-Confuse-Misunderstand (UCM) metric make BioLAMA more friendly to large-N-M relations and rare relations. We also conducted a set of control experiments to disentangle ""understand"" from just ""read and copy"".","{'model': 'tldr@v2.0.0', 'text': 'Context variance prompts and Understand-Confuse-Misunderstand (UCM) metric make BioLAMA more friendly to large-N-M relations and rare relations and disentangle ""understand"" from just ""read and copy"".'}",
-zero-shot event detection based on ordered contrastive learning and prompt-based prediction,Senhui Zhang,"Event detection is a classic natural language processing task. However, the constantly emerging new events make supervised methods not applicable to unseen types. Previous zero-shot event detection methods either require predefined event types as heuristic rules or resort to external semantic analyzing tools. To overcome this weakness, we propose an end-to-end framework named Z ero-Shot E vent Detection Based on O rdered Contrastive Learning and P rompt-Based Prediction (ZEOP). By creatively introducing multiple contrastive samples with ordered similarities, the encoder can learn event representations from both instance-level and class-level, which makes the distinctions between different unseen types more significant. Meanwhile, we utilize the prompt-based prediction to identify trigger words without relying on external resources. Experiments demonstrate that our model detects events more effectively and accurately than state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end framework named Z ero-Shot Event Detection Based on O rdered Contrastive Learning and P rompt-Based Prediction (ZEOP) that can learn event representations from both instance-level and class-level, which makes the distinctions between different unseen types more significant.'}",https://aclanthology.org/2022.findings-naacl.196.pdf
-prompt-based time series forecasting: a new task and dataset,Hao Xue,"The research of time series forecasting benefits a wide range of applications from weather forecasting to human mobility or traffic prediction. This paper studies the time series forecasting problem from a whole new perspective. In the existing methods, the forecasting models take a sequence of numerical values as input and yield numerical values as output. Inspired by the successes of pre-trained language foundation models, we pose a question about whether these models can also be adapted to time series forecasting tasks. Thus, we propose a novel prompt-based time series forecasting (PromptCast) task. In this task, the numerical input and output are transformed into language sentence prompts. We frame the forecasting task in a sentence-to-sentence manner which makes it possible to directly apply language models for the forecasting purpose. To support and facilitate the research of this task, we also present a large-scale dataset (PISA) that includes three real-world forecasting scenarios in this paper. We evaluate different state-of-the-art numerical-based forecasting methods and language generation models such as Bart and Bigbird. The benchmark results demonstrate that the proposed prompt-based time series forecasting with language generation models is a promising research direction. In addition, in comparison to conventional numerical-based forecasting, prompt-based forecasting shows a better generalization ability. We believe that the proposed PromptCast benchmark task as well as our PISA dataset could provide novel insights and further lead to new research directions in the time series forecasting domain.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based time series forecasting (PromptCast) task that frame the forecasting task in a sentence-to-sentence manner which makes it possible to directly apply language models for the forecasting purpose and shows a better generalization ability.'}",http://arxiv.org/pdf/2210.08964
-prompt-based meta-learning for few-shot text classification,Haoxing Zhang,"Few-shot Text Classification predicts the semantic label of a given text with a handful of supporting instances. Current meta-learning methods have achieved satisfying results in various few-shot situations. Still, they often require a large amount of data to construct many few-shot tasks for meta-training, which is not practical in real-world few-shot scenarios. Prompt-tuning has recently proved to be another effective few-shot learner by bridging the gap between pre-train and downstream tasks. In this work, we closely combine the two promising few-shot learning methodologies in structure and propose a Prompt-Based Meta-Learning (PBML) model to overcome the above meta-learning problem by adding the prompting mechanism. PBML assigns label word learning to base-learners and template learning to meta-learner, respectively. Experimental results show state-of-the-art performance on four text classification datasets under few-shot settings, with higher accuracy and good robustness. We demonstrate through low-resource experiments that our method alleviates the shortcoming that meta-learning requires too much data for meta-training. In the end, we use the visualization to interpret and verify that the meta-learning framework can help the prompting method converge better. We release our code to reproduce our experiments.","{'model': 'tldr@v2.0.0', 'text': 'This work closely combines the two promising few-shot learning methodologies in structure and proposes a Prompt-Based Meta-Learning (PBML) model to overcome the above meta-learning problem by adding the prompting mechanism.'}",https://aclanthology.org/2022.emnlp-main.87.pdf
-ai illustrator: translating raw descriptions into images by prompt-based cross-modal generation,Y. Ma,"AI illustrator aims to automatically design visually appealing images for books to provoke rich thoughts and emotions. To achieve this goal, we propose a framework for translating raw descriptions with complex semantics into semantically corresponding images. The main challenge lies in the complexity of the semantics of raw descriptions, which may be hard to be visualized e.g., ""gloomy"" or ""Asian""). It usually poses challenges for existing methods to handle such descriptions. To address this issue, we propose a Prompt-based Cross-Modal Generation Framework (PCM-Frame) to leverage two powerful pre-trained models, including CLIP and StyleGAN. Our framework consists of two components: a projection module from Text Embeddings to Image Embeddings based on prompts, and an adapted image generation module built on StyleGAN which takes Image Embeddings as inputs and is trained by combined semantic consistency losses. To bridge the gap between realistic images and illustration designs, we further adopt a stylization model as post-processing in our framework for better visual effects. Benefiting from the pre-trained models, our method can handle complex descriptions and does not require external paired data for training. Furthermore, we have built a benchmark that consists of 200 descriptions from literature books or online resources. We conduct a user study to demonstrate our superiority over the competing methods of text-to-image translation with complicated semantics.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt-based Cross-Modal Generation Framework (PCM-Frame) to leverage two powerful pre-trained models, including CLIP and StyleGAN, and conducts a user study to demonstrate its superiority over the competing methods of text-to-image translation with complicated semantics.'}",https://arxiv.org/pdf/2209.03160
-parabart: a prompt-based method with parabiotic decoder for few-shot named entity recognition,,"Prompt-based methods have been widely used 001 in few-shot named entity recognition (NER). 002 We first conduct a preliminary experiment and 003 observe that what really affects prompt-based 004 NER models is the ability to detect entity 005 boundaries. However, previous prompt-based 006 NER models neglect to enhance the ability of 007 entity boundary detection. To solve the issue, 008 we propose a novel method, ParaBART, which 009 consists of a BART encoder and the Parabi- 010 otic 1 Decoder we design. Parabiotic Decoder 011 includes two BART decoders and a conjoint 012 module. The two decoders are responsible for 013 entity boundary detection and entity type classi- 014 fication respectively and share the well-learned 015 knowledge through the conjoint module, which 016 replaces unimportant tokens’ embeddings in 017 one decoder with the average embedding of 018 all tokens in the other decoder. Moreover, we 019 propose a novel boundary expansion strategy 020 to enhance the ability of entity type classifica- 021 tion. Experimental results show that ParaBART 022 can achieve significant performance gains over 023 previous state-of-the-art methods. For repro- 024 ducibility, all datasets and codes are provided 025 in the supplementary materials. 026 ,","{'model': 'tldr@v2.0.0', 'text': 'A novel method, ParaBART, which consists of a BART encoder and the Parabi- 010 otic 1 Decoder, which can achieve significant performance gains over previous state-of-the-art methods in few-shot named entity recognition.'}",
-clamp: prompt-based contrastive learning for connecting language and animal pose,Xu Zhang,"Animal pose estimation is challenging for existing image-based methods because of limited training data and large intra- and inter-species variances. Motivated by the progress of visual-language research, we propose that pre-trained language models (e.g., CLIP) can facilitate animal pose estimation by providing rich prior knowledge for describing animal keypoints in text. However, we found that building effective connections between pre-trained language models and visual animal keypoints is non-trivial since the gap between text-based descriptions and keypoint-based visual features about animal pose can be significant. To address this issue, we introduce a novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively. The CLAMP attempts to bridge the gap by adapting the text prompts to the animal keypoints during network training. The adaptation is decomposed into spatialaware and feature-aware processes, and two novel contrastive losses are devised correspondingly. In practice, the CLAMP enables the first cross-modal animal pose estimation paradigm. Experimental results show that our method achieves state-of-the-art performance under the supervised, few-shot, and zero-shot settings, outperforming image-based methods by a large margin. The code is available at https://github.com/xuzhang1199/CLAMP.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively, which enables the first cross-modal animal pose estimation paradigm.'}",https://arxiv.org/pdf/2206.11752
-promptattack: prompt-based attack for language models via gradient search,Yundi Shi,"As the pre-trained language models (PLMs) continue to grow, so do the hardware and data requirements for fine-tuning PLMs. Therefore, the researchers have come up with a lighter method called \textit{Prompt Learning}. However, during the investigations, we observe that the prompt learning methods are vulnerable and can easily be attacked by some illegally constructed prompts, resulting in classification errors, and serious security problems for PLMs. Most of the current research ignores the security issue of prompt-based methods. Therefore, in this paper, we propose a malicious prompt template construction method (\textbf{PromptAttack}) to probe the security performance of PLMs. Several unfriendly template construction approaches are investigated to guide the model to misclassify the task. Extensive experiments on three datasets and three PLMs prove the effectiveness of our proposed approach PromptAttack. We also conduct experiments to verify that our method is applicable in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a malicious prompt template construction method (\\textbf{PromptAttack}) to probe the security performance of PLMs and proves the effectiveness of this approach.'}",http://arxiv.org/pdf/2209.01882
-pts: a prompt-based teacher-student network for weakly supervised aspect detection,Hongjian Li,"Most existing weakly supervised aspect detection algorithms utilize pre-trained language models as their backbone networks by constructing discriminative tasks with seed words. Once the number of seed words decreases, the performance of current models declines significantly. Recently, prompt tuning has been proposed to bridge the gap of objective forms in pre-training and fine-tuning, which is hopeful of alleviating the above challenge. However, directly applying the existing prompt-based methods to this task not only fails to effectively use large amounts of unlabeled data, but also may cause serious over-fitting problems. In this paper, we propose a lightweight teacher-student network (PTS) based on prompts to solve the above two problems. Concretely, the student network is a hybrid prompt-based classification model to detect aspects, which innovatively compounds hand-crafted prompts and auto-generated prompts. The teacher network comprehensively considers the representation of the sentence and the masked aspect token in the template to guide classification. To utilize unlabeled data and seed words intelligently, we train the teacher and student network alternately. Furthermore, in order to solve the problem that the uneven quality of training data obviously affects the iterative efficiency of PTS, we design a general dynamic data selection strategy to feed the most pertinent data into the current model. Experimental results show that even given the minimum seed words, PTS significantly outperforms previous state-of-the-art methods on three widely used benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A lightweight teacher-student network (PTS) based on prompts to solve the problem that the uneven quality of training data obviously affects the iterative efficiency of PTS, and design a general dynamic data selection strategy to feed the most pertinent data into the current model.'}",
-prompt-based metric learning for few-shot ner,Yanru Chen,"Few-shot named entity recognition (NER) targets generalizing to unseen labels and/or domains with few labeled examples. Existing metric learning methods compute token-level similarities between query and support sets, but are not able to fully incorporate label semantics into modeling. To address this issue, we propose a simple method to largely improve metric learning for NER: 1) multiple prompt schemas are designed to enhance label semantics; 2) we propose a novel architecture to effectively combine multiple prompt-based representations. Empirically, our method achieves new state-of-the-art (SOTA) results under 16 of the 18 considered settings, substantially outperforming the previous SOTA by an average of 8.84% and a maximum of 34.51% in relative gains of micro F1. Our code is available at https://github.com/AChen-qaq/ProML.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple method to largely improve metric learning for NER by proposing a novel architecture to effectively combine multiple prompt-based representations designed to enhance label semantics.'}",https://arxiv.org/pdf/2211.04337
-prompt-based zero-shot video moment retrieval,Guolong Wang,"Video moment retrieval aims at localizing a specific moment from an untrimmed video by a sentence query. Most methods rely on heavy annotations of video moment-query pairs. Recent zero-shot methods reduced annotation cost, yet they neglected the global visual feature due to the separation of video and text learning process. To avoid the lack of visual features, we propose a Prompt-based Zero-shot Video Moment Retrieval (PZVMR) method. Motivated by the frame of prompt learning, we design two modules: 1) Proposal Prompt (PP): We randomly masks sequential frames to build a prompt to generate proposals; 2) Verb Prompt (VP): We provide patterns of nouns and the masked verb to build a prompt to generate pseudo queries with verbs. Our PZVMR utilizes task-relevant knowledge distilled from pre-trained CLIP and adapts the knowledge to VMR. Unlike the pioneering work, we introduce visual features into each module. Extensive experiments show that our PZVMR not only outperforms the existing zero-shot method (PSVL) on two public datasets (Charades-STA and ActivityNet-Captions) by 4.4% and 2.5% respectively in mIoU, but also outperforms several methods using stronger supervision.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Prompt-based Zero-shot Video Moment Retrieval (PZVMR) method that utilizes task-relevant knowledge distilled from pre-trained CLIP and adapts the knowledge to VMR and outperforms several methods using stronger supervision.'}",
-on the robustness of dialogue history representation in conversational question answering: a comprehensive study and a new prompt-based method,Zorik Gekhman,"Most work on modeling the conversation history in Conversational Question Answering (CQA) reports a single main result on a common CQA benchmark. While existing models show impressive results on CQA leaderboards, it remains unclear whether they are robust to shifts in setting (sometimes to more realistic ones), training data size (e.g., from large to small sets) and domain. In this work, we design and conduct the first large-scale robustness study of history modeling approaches for CQA. We find that high benchmark scores do not necessarily translate to strong robustness, and that various methods can perform extremely differently under different settings. Equipped with the insights from our study, we design a novel prompt-based history modeling approach and demonstrate its strong robustness across various settings. Our approach is inspired by existing methods that highlight historic answers in the passage. However, instead of highlighting by modifying the passage token embeddings, we add textual prompts directly in the passage text. Our approach is simple, easy to plug into practically any model, and highly effective, thus we recommend it as a starting point for future model developers. We also hope that our study and insights will raise awareness to the importance of robustness-focused evaluation, in addition to obtaining high leaderboard scores, leading to better CQA systems.1","{'model': 'tldr@v2.0.0', 'text': 'This work design and conduct the first large-scale robustness study of history modeling approaches for CQA, finding that high benchmark scores do not necessarily translate to strong robustness, and that various methods can perform extremely differently under different settings.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00549/2080031/tacl_a_00549.pdf
-promptda: label-guided data augmentation for prompt-based few shot learners,Canyu Chen,"Recent advances in large pre-trained language models (PLMs) lead to impressive gains on natural language understanding (NLU) tasks with task-specific fine-tuning. However, directly fine-tuning PLMs heavily relies on sufficient labeled training instances, which are usually hard to obtain. Prompt-based tuning on PLMs has shown to be powerful for various downstream few-shot tasks. Existing works studying prompt-based tuning for few-shot NLU tasks mainly focus on deriving proper label words with a verbalizer or generating prompt templates to elicit semantics from PLMs. In addition, conventional data augmentation strategies such as synonym substitution are also widely adopted in low-resource scenarios. However, the improvements they bring to prompt-based few-shot learning have been demonstrated to be marginal. Thus, an important research question arises as follows: how to design effective data augmentation methods for prompt-based few-shot tuning? To this end, considering the label semantics are essential in prompt-based tuning, we propose a novel label-guided data augmentation framework PromptDA, which exploits the enriched label semantic information for data augmentation. Extensive experiment results on few-shot text classification tasks show that our proposed framework achieves superior performances by effectively leveraging label semantics and data augmentation for natural language understanding.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel label-guided data augmentation framework PromptDA, which exploits the enriched label semantic information for data augmentations in prompt-based few-shot tuning and achieves superior performances.'}",http://arxiv.org/pdf/2205.09229
-nsp-bert: a prompt-based zero-shot learner through an original pre-training task-next sentence prediction,Yi Sun,"Using prompts to utilize language models 001 to perform various downstream tasks, also 002 known as prompt-based learning or prompt- 003 learning , has lately gained significant success 004 in comparison to the pre-train and fine-tune 005 paradigm. Nonetheless, virtually all prompt- 006 based methods are token-level, meaning they 007 all utilize GPT’s left-to-right language model 008 or BERT’s masked language model to per- 009 form cloze-style tasks. In this paper, we at- 010 tempt to accomplish several NLP tasks in the 011 zero-shot scenario using a BERT original pre- 012 training task abandoned by RoBERTa and other 013 models—Next Sentence Prediction (NSP). Un- 014 like token-level techniques, our sentence-level 015 prompt-based method NSP-BERT does not 016 need to fix the length of the prompt or the po- 017 sition to be predicted, allowing it to handle 018 tasks such as entity linking with ease. Based on 019 the characteristics of NSP-BERT, we offer sev- 020 eral quick building templates for various down- 021 stream tasks. We suggest a two-stage prompt 022 method for word sense disambiguation tasks 023 in particular. Our samples-contrast method for 024 mapping the labels significantly enhance the 025 model’s performance on sentence-pair tasks. 026 On the Chinese benchmark FewCLUE, our 027 NSP-BERT outperforms other zero-shot meth- 028 ods on most of these tasks and comes close to 029 the few-shot methods. And on GLUE and other 030 English datasets NSP-BERT is still competitive. 031 Our code will be available on github. 032","{'model': 'tldr@v2.0.0', 'text': 'This paper accomplishes several NLP tasks in the 011 zero-shot scenario using a BERT original pre- 012 training task abandoned by RoBERTa and other 013 models—Next Sentence Prediction (NSP).'}",
-unified multimodal pre-training and prompt-based tuning for vision-language understanding and generation,Tianyi Liu,"Most existing vision-language pre-training methods focus on understanding tasks and use BERT-like objectives (masked language modeling and image-text matching) during pretraining. Although they perform well in many understanding downstream tasks, e.g., visual question answering, image-text retrieval and visual entailment, they do not possess the ability to generate. To tackle this problem, we propose Unified multimodal pre-training for both Vision-Language understanding and generation (UniVL). The proposed UniVL is capable of handling both understanding tasks and generative tasks. We augment existing pretraining paradigms that only use random masks with causal masks, i.e., triangular masks that mask out future tokens, such that the pre-trained models can have autoregressive generation abilities by design. We formulate several previous understanding tasks as a text generation task and propose to use prompt-based method for fine-tuning on different downstream tasks. Our experiments show that there is a trade-off between understanding tasks and generation tasks while using the same model, and a feasible way to improve both tasks is to use more data. Our UniVL framework attains comparable performance to recent vision-language pre-training methods on both understanding tasks and generation tasks. Moreover, we demostrate that prompt-based finetuning is more data-efficient - it outperforms discriminative methods in few-shot scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The UniVL framework attains comparable performance to recent vision-language pre-training methods on both understanding tasks and generation tasks, and demostrate that prompt-based finetuning is more data-efficient - it outperforms discriminative methods in few-shot scenarios.'}",
-adversarial robustness of prompt-based few-shot learning for natural language understanding,Venkata Prabhakara Sarath Nookala,"State-of-the-art few-shot learning (FSL) methods leverage prompt-based fine-tuning to obtain remarkable results for natural language understanding (NLU) tasks. While much of the prior FSL methods focus on improving downstream task performance, there is a limited understanding of the adversarial robustness of such methods. In this work, we conduct an extensive study of several state-of-the-art FSL methods to assess their robustness to adversarial perturbations. To better understand the impact of various factors towards robustness (or the lack of it), we evaluate prompt-based FSL methods against fully fine-tuned models for aspects such as the use of unlabeled data, multiple prompts, number of few-shot examples, model size and type. Our results on six GLUE tasks indicate that compared to fully fine-tuned models, vanilla FSL methods lead to a notable relative drop in task performance (i.e., are less robust) in the face of adversarial perturbations. However, using (i) unlabeled data for prompt-based FSL and (ii) multiple prompts flip the trend. We further demonstrate that increasing the number of few-shot examples and model size lead to increased adversarial robustness of vanilla FSL methods. Broadly, our work sheds light on the adversarial robustness evaluation of prompt-based FSL methods for NLU tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts an extensive study of several state-of-the-art FSL methods to assess their robustness to adversarial perturbations, and demonstrates that increasing the number of few-shot examples and model size lead to increased adversarial robustness of vanilla F SL methods.'}",http://arxiv.org/pdf/2306.11066
-"don’t prompt, search! mining-based zero-shot learning with language models",Mozes van de Kar,"Masked language models like BERT can perform text classification in a zero-shot fashion by reformulating downstream tasks as text infilling. However, this approach is highly sensitive to the template used to prompt the model, yet practitioners are blind when designing them in strict zero-shot settings. In this paper, we propose an alternative mining-based approach for zero-shot learning. Instead of prompting language models, we use regular expressions to mine labeled examples from unlabeled corpora, which can optionally be filtered through prompting, and used to finetune a pretrained model. Our method is more flexible and interpretable than prompting, and outperforms it on a wide range of tasks when using comparable templates. Our results suggest that the success of prompting can partly be explained by the model being exposed to similar examples during pretraining, which can be directly retrieved through regular expressions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an alternative mining-based approach for zero-shot learning that uses regular expressions to mine labeled examples from unlabeled corpora, which can optionally be filtered through prompting, and used to finetune a pretrained model.'}",https://arxiv.org/pdf/2210.14803
-prediction model development of late-onset preeclampsia using machine learning-based methods,J. Jhee,"Preeclampsia is one of the leading causes of maternal and fetal morbidity and mortality. Due to the lack of effective preventive measures, its prediction is essential to its prompt management. This study aimed to develop models using machine learning to predict late-onset preeclampsia using hospital electronic medical record data. The performance of the machine learning based models and models using conventional statistical methods were also compared. A total of 11,006 pregnant women who received antenatal care at Yonsei University Hospital were included. Maternal data were retrieved from electronic medical records during the early second trimester to 34 weeks. The prediction outcome was late-onset preeclampsia occurrence after 34 weeks’ gestation. Pattern recognition and cluster analysis were used to select the parameters included in the prediction models. Logistic regression, decision tree model, naïve Bayes classification, support vector machine, random forest algorithm, and stochastic gradient boosting method were used to construct the prediction models. C-statistics was used to assess the performance of each model. The overall preeclampsia development rate was 4.7% (474 patients). Systolic blood pressure, serum blood urea nitrogen and creatinine levels, platelet counts, serum potassium level, white blood cell count, serum calcium level, and urinary protein were the most influential variables included in the prediction models. C-statistics for the decision tree model, naïve Bayes classification, support vector machine, random forest algorithm, stochastic gradient boosting method, and logistic regression models were 0.857, 0.776, 0.573, 0.894, 0.924, and 0.806, respectively. The stochastic gradient boosting model had the best prediction performance with an accuracy and false positive rate of 0.973 and 0.009, respectively. The combined use of maternal factors and common antenatal laboratory data of the early second trimester through early third trimester could effectively predict late-onset preeclampsia using machine learning algorithms. Future prospective studies are needed to verify the clinical applicability algorithms.","{'model': 'tldr@v2.0.0', 'text': 'The combined use of maternal factors and common antenatal laboratory data of the early second trimester through early third trimester could effectively predict late-onset preeclampsia using machine learning algorithms.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0221202&type=printable
-"management and point-of-care for tobacco dependence (prompt): a feasibility mixed methods community-based participatory action research project in ottawa, canada",S. Pakhalé,"Objective To determine the feasibility of a Community-Based Participatory Tobacco Dependence Strategy (PROMPT) in the inner city population of Ottawa (Canada). Design A feasibility mixed methods prospective cohort study following principles of community-based participatory action research. Intervention Recruited 80 people whouse drugs, followed them for 6 months while providing access to counselling, nicotine replacement therapy and peer-support in a community setting. Setting Community research office in downtown Ottawa, adjacent to low-income housing, shelter services and street-based drug consumption. Primary outcome Retention rate at 6-month follow-up. Secondary outcome Biochemically validated 7-day point prevalence smoking abstinence at 26 weeks, self-reported abstinence in the past 7 days with exhaled carbon monoxide ≤10 ppm. Results The average age of participants was 43.8 years. The 6-month follow-up rate was 42.5%. The mean number of smoking years reported was 27.3 years. The participants were 70% male, 33.7% reported less than a high-school education, 21% identified as indigenous and 43.8% reported an income between US$1000 and US$1999 per month. The baseline mean daily cigarette use was 20.5 and 9.3 cigarettes at study end, with mean reduction of 11.2 cigarettes at 6 months (P=0.0001). There was a considerable reduction in self-reported illicit substance use (18.8%), including a reduction in the opioids heroin (6.3%), fentanyl (2.6%) and Oxycontin (3.8%). The study findings also reveal psycho-socioeconomic benefits such as improved health, return to work and greater community engagement. Conclusions The PROMPT project describes socioeconomic variables associated with tobacco and polysubstance use. A programme focused on tobacco dependence, easily accessible in the community and led by community peers with lived experience is feasible to implement and has the potential to support positive life changes. PROMPT’s patient engagement model is an effective harm-reduction strategy for the growing opioid use crisis and can improve the health outcomes of marginalised at-risk populations worldwide.","{'model': 'tldr@v2.0.0', 'text': 'A programme focused on tobacco dependence, easily accessible in the community and led by community peers with lived experience is feasible to implement and has the potential to support positive life changes.'}",https://bmjopen.bmj.com/content/bmjopen/8/1/e018416.full.pdf
-web-based patient-reported outcome measures for personalized treatment and care (prompt-care): multicenter pragmatic nonrandomized trial,A. Girgis,"Background Despite the acceptability and efficacy of e–patient-reported outcome (ePRO) systems, implementation in routine clinical care remains challenging. Objective This pragmatic trial implemented the PROMPT-Care (Patient Reported Outcome Measures for Personalized Treatment and Care) web-based system into existing clinical workflows and evaluated its effectiveness among a diverse population of patients with cancer. Methods Adult patients with solid tumors receiving active treatment or follow-up care in four cancer centers were enrolled. The PROMPT-Care intervention supported patient management through (1) monthly off-site electronic PRO physical symptom and psychosocial well-being assessments, (2) automated electronic clinical alerts notifying the care team of unresolved clinical issues following two consecutive assessments, and (3) tailored online patient self-management resources. Propensity score matching was used to match controls with intervention patients in a 4:1 ratio for patient age, sex, and treatment status. The primary outcome was a reduction in emergency department presentations. Secondary outcomes were time spent on chemotherapy and the number of allied health service referrals. Results From April 2016 to October 2018, 328 patients from four public hospitals received the intervention. Matched controls (n=1312) comprised the general population of patients with cancer, seen at the participating hospitals during the study period. Emergency department visits were significantly reduced by 33% (P=.02) among patients receiving the intervention compared with patients in the matched controls. No significant associations were found in allied health referrals or time to end of chemotherapy. At baseline, the most common patient reported outcomes (above-threshold) were fatigue (39%), tiredness (38.4%), worry (32.9%), general wellbeing (32.9%), and sleep (24.1%), aligning with the most frequently accessed self-management domain pages of physical well-being (36%) and emotional well-being (23%). The majority of clinical feedback reports were reviewed by nursing staff (729/893, 82%), largely in response to the automated clinical alerts (n=877). Conclusions Algorithm-supported web-based systems utilizing patient reported outcomes in clinical practice reduced emergency department presentations among a diverse population of patients with cancer. This study also highlighted the importance of (1) automated triggers for reviewing above-threshold results in patient reports, rather than passive manual review of patient records; (2) the instrumental role nurses play in managing alerts; and (3) providing patients with resources to support guided self-management, where appropriate. Together, these factors will inform the integration of web-based PRO systems into future models of routine cancer care. Trial Registration Australian New Zealand Clinical Trials Registry ACTRN12616000615482; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=370633 International Registered Report Identifier (IRRID) RR2-10.1186/s12885-018-4729-3","{'model': 'tldr@v2.0.0', 'text': 'Algorithm-supported web-based systems utilizing patient reported outcomes in clinical practice reduced emergency department presentations among a diverse population of patients with cancer.'}",https://www.jmir.org/2020/10/e19685/PDF
-decorate the newcomers: visual domain prompt for continual test time adaptation,Yulu Gan,"Continual Test-Time Adaptation (CTTA) aims to adapt the source model to continually changing unlabeled target domains without access to the source data. Existing methods mainly focus on model-based adaptation in a self-training manner, such as predicting pseudo labels for new domain datasets. Since pseudo labels are noisy and unreliable, these methods suffer from catastrophic forgetting and error accumulation when dealing with dynamic data distributions. Motivated by the prompt learning in NLP, in this paper, we propose to learn an image-layer visual domain prompt for target domains while having the source model parameters frozen. During testing, the changing target datasets can be adapted to the source model by reformulating the input data with the learned visual prompts. Specifically, we devise two types of prompts, i.e., domains-specific prompts and domains-agnostic prompts, to extract current domain knowledge and maintain the domain-shared knowledge in the continual adaptation. Furthermore, we design a homeostasis-based adaptation strategy to suppress domain-sensitive parameters in domain-invariant prompts to learn domain-shared knowledge more effectively. This transition from the model-dependent paradigm to the model-free one enables us to bypass the catastrophic forgetting and error accumulation problems. Experiments show that our proposed method achieves significant performance gains over state-of-the-art methods on four widely-used benchmarks, including CIFAR-10C, CIFAR-100C, ImageNet-C, and VLCS datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to learn an image-layer visual domain prompt for target domains while having the source model parameters frozen, and designs a homeostasis-based adaptation strategy to suppress domain-sensitive parameters in domain-invariant prompts to learn domain-shared knowledge more effectively.'}",http://arxiv.org/pdf/2212.04145
-prompt-tuned code language model as a neural knowledge base for type inference in statically-typed partial code,Qing Huang,"Partial code usually involves non-fully-qualified type names (non-FQNs) and undeclared receiving objects. Resolving the FQNs of these non-FQN types and undeclared receiving objects (referred to as type inference) is the prerequisite to effective search and reuse of partial code. Existing dictionary-lookup based methods build a symbolic knowledge base of API names and code contexts, which involve significant compilation overhead and are sensitive to unseen API names and code context variations. In this paper, we formulate type inference as a cloze-style fill-in-blank language task. Built on source code naturalness, our approach fine-tunes a code masked language model (MLM) as a neural knowledge base of code elements with a novel “pre-train, prompt and predict” paradigm from raw source code. Our approach is lightweight and has minimum requirements on code compilation. Unlike existing symbolic name and context matching for type inference, our prompt-tuned code MLM packs FQN syntax and usage in its parameters and supports fuzzy neural type inference. We systematically evaluate our approach on a large amount of source code from GitHub and Stack Overflow. Our results confirm the effectiveness of our approach design and the practicality for partial code type inference. As the first of its kind, our neural type inference method opens the door to many innovative ways of using partial code.","{'model': 'tldr@v2.0.0', 'text': 'This paper fine-tunes a code masked language model (MLM) as a neural knowledge base of code elements with a novel “pre-train, prompt and predict” paradigm from raw source code, and formulates type inference as a cloze-style fill-in-blank language task.'}",https://dl.acm.org/doi/pdf/10.1145/3551349.3556912
-"toward human readable prompt tuning: kubrick's the shining is a good movie, and a good prompt too?",Weijia Shi,"Large language models can perform new tasks in a zero-shot fashion, given natural language prompts that specify the desired behavior. Such prompts are typically hand engineered, but can also be learned with gradient-based methods from labeled data. However, it is underexplored what factors make the prompts effective, especially when the prompts are natural language. In this paper, we investigate common attributes shared by effective prompts. We first propose a human readable prompt tuning method (F LUENT P ROMPT) based on Langevin dynamics that incorporates a fluency constraint to find a diverse distribution of effective and fluent prompts. Our analysis reveals that effective prompts are topically related to the task domain and calibrate the prior probability of label words. Based on these findings, we also propose a method for generating prompts using only unlabeled data, outperforming strong baselines by an average of 7.0% accuracy across three tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a human readable prompt tuning method (F LUENT P ROMPT) based on Langevin dynamics that incorporates a fluency constraint to find a diverse distribution of effective and fluent prompts and investigates common attributes shared by effective prompts.'}",http://arxiv.org/pdf/2212.10539
-parameter-efficient prompt tuning makes generalized and calibrated neural text retrievers,W. Tam,"Prompt tuning attempts to update few task-specific parameters in pre-trained models. It has achieved comparable performance to fine-tuning of the full parameter set on both language understanding and generation tasks. In this work, we study the problem of prompt tuning for neural text retrievers. We introduce parameter-efficient prompt tuning for text retrieval across in-domain, cross-domain, and cross-topic settings. Through an extensive analysis, we show that the strategy can mitigate the two issues -- parameter-inefficiency and weak generalizability -- faced by fine-tuning based retrieval methods. Notably, it can significantly improve the out-of-domain zero-shot generalization of the retrieval models. By updating only 0.1% of the model parameters, the prompt tuning strategy can help retrieval models achieve better generalization performance than traditional methods in which all parameters are updated. Finally, to facilitate research on retrievers' cross-topic generalizability, we curate and release an academic retrieval dataset with 18K query-results pairs in 87 topics, making it the largest topic-specific one to date.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces parameter-efficient prompt tuning for text retrieval across in-domain, cross- domain, and cross-topic settings and shows that the strategy can mitigate the two issues -- parameter-inefficiency and weak generalizability -- faced by fine-tuning based retrieval methods.'}",http://arxiv.org/pdf/2207.07087
-relation extraction as open-book examination: retrieval-enhanced prompt tuning,Xiang Chen,"Pre-trained language models have contributed significantly to relation extraction by demonstrating remarkable few-shot learning abilities. However, prompt tuning methods for relation extraction may still fail to generalize to those rare or hard patterns. Note that the previous parametric learning paradigm can be viewed as memorization regarding training data as a book and inference as the close-book test. Those long-tailed or hard patterns can hardly be memorized in parameters given few-shot instances. To this end, we regard RE as an open-book examination and propose a new semiparametric paradigm of retrieval-enhanced prompt tuning for relation extraction. We construct an open-book datastore for retrieval regarding prompt-based instance representations and corresponding relation labels as memorized key-value pairs. During inference, the model can infer relations by linearly interpolating the base output of PLM with the non-parametric nearest neighbor distribution over the datastore. In this way, our model not only infers relation through knowledge stored in the weights during training but also assists decision-making by unwinding and querying examples in the open-book datastore. Extensive experiments on benchmark datasets show that our method can achieve state-of-the-art in both standard supervised and few-shot settings","{'model': 'tldr@v2.0.0', 'text': 'This work regards RE as an open-book examination and proposes a new semiparametric paradigm of retrieval-enhanced prompt tuning for relation extraction, which can achieve state-of-the-art in both standard supervised and few-shot settings.'}",https://arxiv.org/pdf/2205.02355
-prompt deep light-weight vessel segmentation network (plvs-net),Muhammad Arsalan,"Achieving accurate retinal vessel segmentation is critical in the progression and diagnosis of vision-threatening diseases such as diabetic retinopathy and age-related macular degeneration. Existing vessel segmentation methods are based on encoder-decoder architectures, which frequently fail to take into account the retinal vessel structure's context in their analysis. As a result, such methods have difficulty bridging the semantic gap between encoder and decoder characteristics. This paper proposes a Prompt Deep Light-weight Vessel Segmentation Network (PLVS-Net) to address these issues by using prompt blocks. Each prompt block use combination of asymmetric kernel convolutions, depth-wise separable convolutions, and ordinary convolutions to extract useful features. This novel strategy improves the performance of the segmentation network while simultaneously decreasing the number of trainable parameters. Our method outperformed competing approaches in the literature on three benchmark datasets, including DRIVE, STARE, and CHASE.","{'model': 'tldr@v2.0.0', 'text': 'A Prompt Deep Light-weight Vessel Segmentation Network (PLVS-Net) is proposed to address issues by using prompt blocks to improve the performance of the segmentation network while simultaneously decreasing the number of trainable parameters.'}",
-rethinking reinforcement learning for recommendation: a prompt perspective,Xin Xin,"Modern recommender systems aim to improve user experience. As reinforcement learning (RL) naturally fits this objective---maximizing an user's reward per session---it has become an emerging topic in recommender systems. Developing RL-based recommendation methods, however, is not trivial due to the offline training challenge. Specifically, the keystone of traditional RL is to train an agent with large amounts of online exploration making lots of 'errors' in the process. In the recommendation setting, though, we cannot afford the price of making 'errors' online. As a result, the agent needs to be trained through offline historical implicit feedback, collected under different recommendation policies; traditional RL algorithms may lead to sub-optimal policies under these offline training settings. Here we propose a new learning paradigm---namely Prompt-Based Reinforcement Learning (PRL)---for the offline training of RL-based recommendation agents. While traditional RL algorithms attempt to map state-action input pairs to their expected rewards (e.g., Q-values), PRL directly infers actions (i.e., recommended items) from state-reward inputs. In short, the agents are trained to predict a recommended item given the prior interactions and an observed reward value---with simple supervised learning. At deployment time, this historical (training) data acts as a knowledge base, while the state-reward pairs are used as a prompt. The agents are thus used to answer the question: Which item should be recommended given the prior interactions & the prompted reward value? We implement PRL with four notable recommendation models and conduct experiments on two real-world e-commerce datasets. Experimental results demonstrate the superior performance of our proposed methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new learning paradigm, Prompt-Based Reinforcement Learning (PRL), for the offline training of RL-based recommendation agents, and implements PRL with four notable recommendation models and conducts experiments on two real-world e-commerce datasets.'}",https://dl.acm.org/doi/pdf/10.1145/3477495.3531714
-non-destructive analyses of bronze artefacts from bronze age hungary using neutron-based methods,V. Kiss,"In this paper we present the application of various neutron-based methods carried out at the Budapest Neutron Centre. Non-destructive and non-invasive neutron radiography (NR), prompt gamma activation analysis (PGAA) and time-of-flight neutron diffraction (TOF-ND) analysis were applied to reveal more information on raw material and production techniques of bronze artefacts that can be dated to the Central European Bronze Age (2000–1200 BC).",,
-point prompt tuning for temporally language grounding,Yawen Zeng,"The task of temporally language grounding (TLG) aims to locate a video moment from an untrimmed video that match a given textual query, which has attracted considerable research attention. In recent years, typical retrieval-based TLG methods are inefficient due to pre-segmented candidate moments, while localization-based TLG solutions adopt reinforcement learning resulting in unstable convergence. Therefore, how to perform TLG task efficiently and stably is a non-trivial work. Toward this end, we innovatively contribute a solution, Point Prompt Tuning (PPT), which formulates this task as a prompt-based multi-modal problem and integrates multiple sub-tasks to tuning performance. Specifically, a flexible prompt strategy is contributed to rewrite the query firstly, which contains both query, start point and end point. Thereafter, a multi-modal Transformer is adopted to fully learn the multi-modal context. Meanwhile, we design various sub-tasks to constrain the novel framework, namely matching task and localization task. Finally, the start and end points of matched video moment are straightforward predicted, simply yet stably. Extensive experiments on two real-world datasets have well verified the effectiveness of our proposed solution.","{'model': 'tldr@v2.0.0', 'text': 'This work innovatively contribute a solution, Point Prompt Tuning (PPT), which formulates this task as a prompt-based multi-modal problem and integrates multiple sub-tasks to tuning performance.'}",
-generative prompt tuning for relation classification,Jiale Han,"Using prompts to explore the knowledge contained within pre-trained language models for downstream tasks has now become an active topic. Current prompt tuning methods mostly convert the downstream tasks to masked language modeling problems by adding cloze-style phrases and mapping all labels to verbalizations with fixed length, which has proven effective for tasks with simple label spaces. However, when applied to relation classification exhibiting complex label spaces, vanilla prompt tuning methods may struggle with label verbalizations with arbitrary lengths due to rigid prompt restrictions. Inspired by the text infilling task for pre-training generative models that can flexibly predict missing spans, we propose a novel generative prompt tuning method to reformulate relation classification as an infilling problem, which frees our approach from limitations of current prompt based approaches and thus fully exploits rich semantics of entity and relation types. In addition, we design entity-guided decoding and discriminative relation scoring to generate and align relations effectively and efficiently during inference. Extensive experiments under fully supervised settings and low-resource settings demonstrate the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel generative prompt tuning method to reformulate relation classification as an infilling problem, which frees the approach from limitations of current prompt based approaches and thus fully exploits rich semantics of entity and relation types.'}",http://arxiv.org/pdf/2210.12435
-ehealth system for collecting and utilizing patient reported outcome measures for personalized treatment and care (prompt-care) among cancer patients: mixed methods approach to evaluate feasibility and acceptability,A. Girgis,"Background Despite accumulating evidence indicating that collecting patient-reported outcomes (PROs) and transferring results to the treating health professional in real time has the potential to improve patient well-being and cancer outcomes, this practice is not widespread. Objective The aim of this study was to test the feasibility and acceptability of PROMPT-Care (Patient Reported Outcome Measures for Personalized Treatment and Care), a newly developed electronic health (eHealth) system that facilitates PRO data capture from cancer patients, data linkage and retrieval to support clinical decisions and patient self-management, and data retrieval to support ongoing evaluation and innovative research. Methods We developed an eHealth system in consultation with content-specific expert advisory groups and tested it with patients receiving treatment or follow-up care in two hospitals in New South Wales, Australia, over a 3-month period. Participants were recruited in clinic and completed self-report Web-based assessments either just before their upcoming clinical consultation or every 4 weeks if in follow-up care. A mixed methods approach was used to evaluate feasibility and acceptability of PROMPT-Care; data collected throughout the study informed the accuracy and completeness of data transfer procedures, and extent of missing data was determined from participants’ assessments. Patients participated in cognitive interviews while completing their first assessment and completed evaluation surveys and interviews at study-end to assess system acceptability and usefulness of patient self-management resources, and oncology staff were interviewed at study-end to determine the acceptability and perceived usefulness of real-time PRO reporting. Results A total of 42 patients consented to the study; 7 patients were withdrawn before starting the intervention primarily because of changes in eligibility. Overall, 35 patients (13 on treatment and 22 in follow-up) completed 67 assessments during the study period. Mean completeness of patient-reported data was 93%, with 100% accuracy of data transfer. Ten patients completed cognitive interviews, 28 completed evaluation surveys, and 14 completed evaluation interviews at study-end. PROMPT-Care patient acceptability was high—100% (28/28) reported the time to complete the Web-based assessments (average 15 min) as about right, most willing to answer more questions (79%, 22/28 yes), 96% (27/28) found the Web-based assessment easier or same as completing a paper copy, and they valued the self-management resources . Oncology staff (n=5) also reported high acceptability and potential feasibility of the system. Conclusions Patients and oncology staff found the PROMPT-Care system to be highly acceptable, and the results suggest that it would be feasible to implement it into an oncology setting. Suggested modifications to the patient assessment survey, clinician access to the reports, and system requirements will be made as part of the next stage of large-scale testing and future implementation of the system as part of routine care. Trial registration Australian New Zealand Clinical Trials Registry ACTRN1261500135294; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=369299&isReview=true (Archived by WebCite at http://www.webcitation.org/6lzylG5A0).","{'model': 'tldr@v2.0.0', 'text': ""Patients and oncology staff found the PROMPT-Care system to be highly acceptable, and the results suggest that it would be feasible to implement it into an oncologists' setting.""}",
-prompt learning for few-shot dialogue state tracking,Yuting Yang,"Collecting dialogue state labels, slots and values, for learning dialogue state tracking (DST) models can be costly, especially with the wide application of dialogue systems in new-rising domains. In this paper, we focus on how to learn a DST model efficiently with limited labeled data. We design a prompt learning framework for few-shot DST, which consists of two main components: value-based prompt and inverse prompt mechanism. This framework aims to utilize the language understanding and generation ability of pre-trained language models (PLM). First, we design value-based prompt functions to probe the DST-related knowledge from PLM, which do not rely on the known ontology of slots. Further, an inverse prompt mechanism is utilized to self-check the “prompted” knowledge and help the PLM understand the essence of DST task further. Experiments show that our model can generate unseen slots and outperforms existing state-of-the-art few-shot methods. It indicates that DST-related knowledge can be probed from PLM and utilized to address low-resource DST efficiently with the help of prompt learning.","{'model': 'tldr@v2.0.0', 'text': 'A prompt learning framework for few-shot DST, which consists of two main components: value-based prompt and inverse prompt mechanism, that indicates that DST-related knowledge can be probed from PLM and utilized to address low-resource DST efficiently with the help of prompt learning.'}",
-"prompt, generate, then cache: cascade of foundation models makes strong few-shot learners",Renrui Zhang,"Visual recognition in low-data regimes requires deep neural networks to learn generalized representations from limited training samples. Recently, CLIP-based methods have shown promising few-shot performance benefited from the contrastive language-image pre-training. We then question, if the more diverse pre-training knowledge can be cascaded to further assist few-shot representation learning. In this paper, we propose CaFo, a Cascade of Foundation models that incorporates diverse prior knowledge of various pretraining paradigms for better few-shot learning. Our CaFo incorporates CLIP's language-contrastive knowledge, DINO's vision-contrastive knowledge, DALL-E's vision-generative knowledge, and GPT-3's language-generative knowledge. Specifically, CaFo works by ‘Prompt, Generate, then Cache’. Firstly, we leverage GPT-3 to produce textual inputs for prompting CLIP with rich downstream linguistic semantics. Then, we generate synthetic images via DALL-E to expand the few-shot training data without any manpower. At last, we introduce a learnable cache model to adaptively blend the predictions from CLIP and DINO. By such collaboration, CaFo can fully unleash the potential of different pre-training methods and unify them to perform state-of-the-art for few-shot classification. Code is available at https://github.com/ZrrSkywalker/CaFo.","{'model': 'tldr@v2.0.0', 'text': 'CaFo, a Cascade of Foundation models that incorporates diverse prior knowledge of various pretraining paradigms for better few-shot learning is proposed that can fully unleash the potential of different pre-training methods and unify them to perform state-of-the-art for few- shot classification.'}",https://arxiv.org/pdf/2303.02151
-self-consistent contrastive attributed graph clustering with pseudo-label prompt,Wei Xia,"Attributed graph clustering, which learns node representation from node attribute and topological graph for clustering, is a fundamental and challenging task for multimedia network-structured data analysis. Recently, graph contrastive learning (GCL)-based methods have obtained impressive clustering performance on this task. Nevertheless, there still remain some limitations to be solved: 1) most existing methods fail to consider the self-consistency between latent representations and cluster structures; and 2) most methods require a post-processing operation to get clustering labels. Such a two-step learning scheme results in models that cannot handle newly generated data, i.e., out-of-sample (OOS) nodes. To address these issues in a unified framework, a Self-consistent Contrastive Attributed Graph Clustering (SCAGC) network with pseudo-label prompt is proposed in this article. In SCAGC, by clustering labels prompt information, a self-consistent contrastive loss, which aims to maximize the consistencies of intra-cluster representations while minimizing the consistencies of inter-cluster representations, is designed for representation learning. Meanwhile, a clustering module is built to directly output clustering labels by contrasting the representation of different clusters. Thus, for the OOS nodes, SCAGC can directly calculate their clustering labels. Extensive experimental results on seven benchmark datasets have shown that SCAGC consistently outperforms 16 competitive clustering methods.","{'model': 'tldr@v2.0.0', 'text': 'A Self-consistent Contrastive Attributed Graph Clustering (SCAGC) network with pseudo-label prompt with extensive experimental results on seven benchmark datasets have shown that SCAGC consistently outperforms 16 competitive clustering methods.'}",https://arxiv.org/pdf/2110.08264
-compton camera and prompt gamma ray timing: two methods for in vivo range assessment in proton therapy,F. Hueso-González,"Proton beams are promising means for treating tumors. Such charged particles stop at a defined depth, where the ionization density is maximum. As the dose deposit beyond this distal edge is very low, proton therapy minimizes the damage to normal tissue compared to photon therapy. Nevertheless, inherent range uncertainties cast doubts on the irradiation of tumors close to organs at risk and lead to the application of conservative safety margins. This constrains significantly the potential benefits of protons over photons. In this context, several research groups are developing experimental tools for range verification based on the detection of prompt gammas, a nuclear by-product of the proton irradiation. At OncoRay and Helmholtz-Zentrum Dresden-Rossendorf, detector components have been characterized in realistic radiation environments as a step toward a clinical Compton camera. On the one hand, corresponding experimental methods and results obtained during the ENTERVISION training network are reviewed. On the other hand, a novel method based on timing spectroscopy has been proposed as an alternative to collimated imaging systems. The first tests of the timing method at a clinical proton accelerator are summarized, its applicability in a clinical environment for challenging the current safety margins is assessed, and the factors limiting its precision are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The first tests of the timing method at a clinical proton accelerator are summarized, its applicability in a clinical environment for challenging the current safety margins is assessed, and the factors limiting its precision are discussed.'}",https://www.frontiersin.org/articles/10.3389/fonc.2016.00080/pdf
-knowledge graph-enhanced molecular contrastive learning with functional prompt,Yin Fang,,"{'model': 'tldr@v2.0.0', 'text': 'A chemical element-oriented knowledge graph is introduced to summarize the basic knowledge of elements and their closely related functional groups to contribute to more efficient drug design by offering a high-quality knowledge prior, interpretable molecular representation and superior prediction performance.'}",https://www.nature.com/articles/s42256-023-00654-0.pdf
-instructionner: a multi-task instruction-based generative framework for few-shot ner,Liwen Wang,"Recently, prompt-based methods have achieved significant performance in few-shot learning scenarios by bridging the gap between language model pre-training and fine-tuning for downstream tasks. However, existing prompt templates are mostly designed for sentence-level tasks and are inappropriate for sequence labeling objectives. To address the above issue, we propose a multi-task instruction-based generative framework, named InstructionNER, for low-resource named entity recognition. Specifically, we reformulate the NER task as a generation problem, which enriches source sentences with task-specific instructions and answer options, then inferences the entities and types in natural language. We further propose two auxiliary tasks, including entity extraction and entity typing, which enable the model to capture more boundary information of entities and deepen the understanding of entity type semantics, respectively. Experimental results show that our method consistently outperforms other baselines on five datasets in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a multi-task instruction-based generative framework, named InstructionNER, for low-resource named entity recognition, which reformulate the NER task as a generation problem, which enriches source sentences with task-specific instructions and answer options, then inferences the entities and types in natural language.'}",http://arxiv.org/pdf/2203.03903
-from prompt gamma distribution to dose: a novel approach combining an evolutionary algorithm and filtering based on gaussian-powerlaw convolutions,A. Schumann,"Range verification and dose monitoring in proton therapy is considered as highly desirable. Different methods have been developed worldwide, like particle therapy positron emission tomography (PT-PET) and prompt gamma imaging (PGI). In general, these methods allow for a verification of the proton range. However, quantification of the dose from these measurements remains challenging. For the first time, we present an approach for estimating the dose from prompt γ-ray emission profiles. It combines a filtering procedure based on Gaussian-powerlaw convolution with an evolutionary algorithm. By means of convolving depth dose profiles with an appropriate filter kernel, prompt γ-ray depth profiles are obtained. In order to reverse this step, the evolutionary algorithm is applied. The feasibility of this approach is demonstrated for a spread-out Bragg-peak in a water target.","{'model': 'tldr@v2.0.0', 'text': 'For the first time, an approach is presented for estimating the dose from prompt γ-ray emission profiles that combines a filtering procedure based on Gaussian-powerlaw convolution with an evolutionary algorithm.'}",
-finding skill neurons in pre-trained transformer-based language models,Xiaozhi Wang,"Transformer-based pre-trained language models have demonstrated superior performance on various natural language processing tasks. However, it remains unclear how the skills required to handle these tasks distribute among model parameters. In this paper, we find that after prompt tuning for specific tasks, the activations of some neurons within pre-trained Transformers are highly predictive of the task labels. We dub these neurons skill neurons and confirm they encode task-specific skills by finding that: (1) Skill neurons are crucial for handling tasks. Performances of pre-trained Transformers on a task significantly drop when corresponding skill neurons are perturbed. (2) Skill neurons are task-specific. Similar tasks tend to have similar distributions of skill neurons. Furthermore, we demonstrate the skill neurons are most likely generated in pre-training rather than fine-tuning by showing that the skill neurons found with prompt tuning are also crucial for other fine-tuning methods freezing neuron weights, such as the adapter-based tuning and BitFit. We also explore the applications of skill neurons, including accelerating Transformers with network pruning and building better transferability indicators. These findings may promote further research on understanding Transformers. The source code can be obtained from https://github.com/THU-KEG/Skill-Neuron.","{'model': 'tldr@v2.0.0', 'text': 'It is found that after prompt tuning for specific tasks, the activations of some neurons within pre-trained Transformers are highly predictive of the task labels, and these neurons are dubbed skill neurons and confirmed they encode task-specific skills.'}",https://arxiv.org/pdf/2211.07349
-model of the transient neurovascular response based on prompt arterial dilation,J. H. Kim,,"{'model': 'tldr@v2.0.0', 'text': 'This model provides a powerful framework to quantitatively interpret oxygen transport in the brain; in particular, its intravascular oxygen concentration predictions provide a new model for fMRI responses.'}",https://journals.sagepub.com/doi/pdf/10.1038/jcbfm.2013.90
-analysis framework for the prompt discovery of compact binary mergers in gravitational-wave data,C. Messick,"We describe a stream-based analysis pipeline to detect gravitational waves from the merger of binary neutron stars, binary black holes, and neutron-star–black-hole binaries within ∼1 min of the arrival of the merger signal at Earth. Such low-latency detection is crucial for the prompt response by electromagnetic facilities in order to observe any fading electromagnetic counterparts that might be produced by mergers involving at least one neutron star. Even for systems expected not to produce counterparts, low-latency analysis of the data is useful for deciding when not to point telescopes, and as feedback to observatory operations. Analysts using this pipeline were the first to identify GW151226, the second gravitational-wave event ever detected. The pipeline also operates in an offline mode, in which it incorporates more refined information about data quality and employs acausal methods that are inapplicable to the online mode. The pipeline’s offline mode was used in the detection of the first two gravitational-wave events, GW150914 and GW151226, as well as the identification of a third candidate, LVT151012.",,https://link.aps.org/accepted/10.1103/PhysRevD.95.042001
-multimodal few-shot object detection with meta-learning based cross-modal prompting,G. Han,"We study multi-modal few-shot object detection (FSOD) in this paper, using both few-shot visual examples and class semantic information for detection, which are complementary to each other by definition. Most of the previous works on multi-modal FSOD are fine-tuning-based which are inefficient for online applications. Moreover, these methods usually require expertise like class names to extract class semantic embedding, which are hard to get for rare classes. Our approach is motivated by the high-level conceptual similarity of (metric-based) meta-learning and prompt-based learning to learn generalizable few-shot and zero-shot object detection models respectively without fine-tuning. Specifically, we combine the few-shot visual classifier and text classifier learned via meta-learning and prompt-based learning respectively to build the multi-modal classifier and detection models. In addition, to fully exploit the pre-trained language models, we propose meta-learning-based cross-modal prompting to generate soft prompts for novel classes present in few-shot visual examples, which are then used to learn the text classifier. Knowledge distillation is introduced to learn the soft prompt generator without using human prior knowledge of class names, which may not be available for rare classes. Our insight is that the few-shot support images naturally include related context information and semantics of the class. We comprehensively evaluate the proposed multi-modal FSOD models on multiple few-shot object detection benchmarks, achieving promising results.","{'model': 'tldr@v2.0.0', 'text': 'This paper combines the few-shot visual classifier and text classifier learned via meta-learning and prompt-based learning respectively to build the multi-modal classifiers and detection models.'}",http://arxiv.org/pdf/2204.07841
-good examples make a faster learner: simple demonstration-based learning for low-resource ner,Dong-Ho Lee,"Recent advances in prompt-based learning have shown strong results on few-shot text classification by using cloze-style templates.Similar attempts have been made on named entity recognition (NER) which manually design templates to predict entity types for every text span in a sentence. However, such methods may suffer from error propagation induced by entity span detection, high cost due to enumeration of all possible text spans, and omission of inter-dependencies among token labels in a sentence. Here we present a simple demonstration-based learning method for NER, which lets the input be prefaced by task demonstrations for in-context learning. We perform a systematic study on demonstration strategy regarding what to include (entity examples, with or without surrounding context), how to select the examples, and what templates to use. Results on in-domain learning and domain adaptation show that the model’s performance in low-resource settings can be largely improved with a suitable demonstration strategy (e.g., a 4-17% improvement on 25 train instances). We also find that good demonstration can save many labeled examples and consistency in demonstration contributes to better performance.","{'model': 'tldr@v2.0.0', 'text': 'Results on in-domain learning and domain adaptation show that the model’s performance in low-resource settings can be largely improved with a suitable demonstration strategy, and good demonstration can save many labeled examples and consistency in demonstration contributes to better performance.'}",https://aclanthology.org/2022.acl-long.192.pdf
-inferring prompt black-hole formation in neutron star mergers from gravitational-wave data,M. Agathos,"The gravitational-wave GW170817 is associated to the inspiral phase of a binary neutron star coalescence event. The LIGO-Virgo detectors sensitivity at high frequencies was not sufficient to detect the signal corresponding to the merger and post-merger phases. Hence, the question whether the merger outcome was a prompt black hole formation or not must be answered using either the pre-merger gravitational wave signal or electromagnetic counterparts. In this work we present two methods to infer the probability of prompt black hole formation, using the analysis of the inspiral gravitational-wave signal. Both methods combine the posterior distribution from the gravitational-wave data analysis with numerical relativity results. One method relies on the use of phenomenological models for the equation of state and on the estimate of the collapse threshold mass. The other is based on the estimate of the tidal polarizability parameter $\tilde{\Lambda}$ that is correlated in an equation-of-state agnostic way with the prompt BH formation. We analyze GW170817 data and find that the two methods consistently predict a probability of ~ 50-70% for prompt black-hole formation, which however may significantly decrease below 10% if the maximum mass constraint from PSR J0348+0432 or PSR J0740+6620 is imposed.",,https://arxiv.org/pdf/1908.05442
-development and applications of k0 based naa and prompt gamma-ray naa methods at barc,R. Acharya,,,
-methods for scaling simulation-based teamwork training,M. Delisle,"Effective teamwork and communication is now recognised as a critical component of safe and high-quality patient care. Researchers, policymakers and frontline providers are in search of strategies to improve teamwork in healthcare. The most frequently used strategy is teamwork training.1 Teamwork training involves a systematic process in which a team is guided (often by facilitators) to improve and master different skills for working together effectively. Single-centre teamwork training initiatives have demonstrated improvements in patient care, but these results have been challenging to reproduce at scale.2 In this issue of BMJ Quality and Safety, Lenguerrand et al report the results of a stepped-wedge randomised controlled trial in which PRactical Obstetric Multi-Professional Training (PROMPT), an interprofessional intrapartum training package, was implemented across 12 maternity units in Scotland.3 Each participating unit identified an in-house training team to travel to attend a 2-day PROMPT Train the Trainers programme conducted in one simulation centre; two units were unable to send training teams. Training teams were subsequently responsible for coordinating the delivery of in-house PROMPT courses to all maternity staff within 12 months. The courses were intended to cover core obstetrical emergencies, such as postpartum haemorrhage, sepsis, shoulder dystocia, teamwork and fetal monitoring. In addition to clinical outcomes, each maternity unit collected process data about their local PROMPT courses, including the total number of staff trained and courses delivered and the actual course content. The authors found a significant amount of variability in the implementation across units. For example, all courses included elements of teamwork whereas fetal monitoring and shoulder dystocia training were not universally included. Despite the previously demonstrated benefits of PROMPT in single-centre studies, the final results did not demonstrate a reduction of term babies with a low Apgar score. The authors postulate this null finding may be in part related …","{'model': 'tldr@v2.0.0', 'text': 'Despite the previously demonstrated benefits of PROMPT in single-centre studies, the final results did not demonstrate a reduction of term babies with a low Apgar score, and the authors postulate this null finding may be in part related to variability in the implementation across units.'}",https://qualitysafety.bmj.com/content/qhc/29/2/98.full.pdf
-the utility of an evidence-based lecture and clinical prompt as methods to improve quality of care in colorectal cancer screening,K. Seres,,"{'model': 'tldr@v2.0.0', 'text': 'Clinical prompts are superior to evidence-based lectures for improving physician colorectal cancer screening practices and are simple low-cost measures that can improve quality of care.'}",
-machine learning-based prediction of acute coronary syndrome using only the pre-hospital 12-lead electrocardiogram,S. Al-Zaiti,,"{'model': 'tldr@v2.0.0', 'text': 'A machine learning-based system based exclusively on ECG data that can help clinicians identify 37% more heart attacks during initial screening and reduce unnecessary costs in patients with chest pain is reported.'}",https://www.nature.com/articles/s41467-020-17804-2.pdf
-plasma paracetamol concentration at hospital presentation has a dose-dependent relationship with liver injury despite prompt treatment with intravenous acetylcysteine,David G. Cairney,"ABSTRACT Context: Paracetamol (acetaminophen) overdose is a common reason for emergency hospital admission in the UK and the leading cause of acute liver failure in the Western world. Currently, the antidote acetylcysteine (NAC) is administered at a dose determined only by body weight without regard for the body burden of paracetamol. Objective: To determine whether higher plasma paracetamol concentrations are associated with increased risk of liver injury despite prompt treatment with intravenous NAC. Methods: Patients admitted to hospital for treatment with intravenous NAC following a single acute paracetamol overdose entered the study if NAC was commenced within 24 h of drug ingestion (N = 727 hospital presentations). Based on the plasma paracetamol concentration at first presentation to hospital, a series of nomograms were created: 0–100, 101–150, 151–200, 201–300, 301–500 and over 501 mg/L. The primary endpoints were acute liver injury (ALI – peak serum ALT activity >150 U/L and double the admission value) and hepatotoxicity (peak ALT >1000 U/L). Results: ALI and hepatotoxicity were more common in patients with higher admission plasma paracetamol concentrations despite NAC treatment (ALI: nomogram 0–100: 6%, 101–150: 3%, 151–200: 3%, 201–300: 9%, 301–500: 13%, over 501 mg/dL: 27%. p < 0.0001). This dose-response relationship between paracetamol concentration and ALI persisted even in patients treated with NAC within 8 h of overdose (nomogram 0–100: 0%, 101–150: 0.8%, 151–200: 2%, 201–300: 3.6%, 301–500: 12.5%, over 501mg/L: 33%. p < 0.0001) and in patients with normal ALT activity at first presentation (nomogram: 0–100: 0%, 101–150: 1.2%, 151–200: 1.5%, 201–300: 5.3%, 301–500: 10.8% p < 0.0001). Discussion: Patients with increased concentrations of plasma paracetamol at hospital presentation are at higher risk of liver injury even when intravenous NAC is promptly administered before there is biochemical evidence of toxicity. Conclusion: This study supports theoretical concerns that the current intravenous dose of NAC may be too low in the setting of higher paracetamol exposure.","{'model': 'tldr@v2.0.0', 'text': 'Patients with increased concentrations of plasma paracetamol at hospital presentation are at higher risk of liver injury even when intravenous NAC is promptly administered before there is biochemical evidence of toxicity, which supports theoretical concerns that the current intravenous dose of NAC may be too low in the setting of higher par acetamol exposure.'}",
-real‐time regional seismic damage assessment framework based on long short‐term memory neural network,Yongjia Xu,"Effective post‐earthquake response requires a prompt and accurate assessment of earthquake‐induced damage. However, existing damage assessment methods cannot simultaneously meet these requirements. This study proposes a framework for real‐time regional seismic damage assessment that is based on a Long Short‐Term Memory (LSTM) neural network architecture. The proposed framework is not specially designed for individual structural types, but offers rapid estimates at regional scale. The framework is built around a workflow that establishes high‐performance mapping rules between ground motions and structural damage via region‐specific models. This workflow comprises three main parts—namely, region‐specific database generation, LSTM model training and verification, and model utilization for damage prediction. The influence of various LSTM architectures, hyperparameter selection, and dataset resampling procedures are systematically analyzed. As a testbed for the established framework, a case study is performed on the Tsinghua University campus buildings. The results demonstrate that the developed LSTM framework can perform damage assessment in real time at regional scale with high prediction accuracy and acceptable variance.","{'model': 'tldr@v2.0.0', 'text': 'A framework for real‐time regional seismic damage assessment that is based on a Long Short‐Term Memory (LSTM) neural network architecture is proposed and can perform damage assessment in real time at regional scale with high prediction accuracy and acceptable variance.'}",
-ranibizumab plus prompt or deferred laser for diabetic macular edema in eyes with vitrectomy before anti-vascular endothelial growth factor therapy,S. Bressler,"Background: The approach to managing diabetic macular edema in eyes with previous vitrectomy is based on limited evidence. Therefore, an exploratory post hoc assessment of 3-year data from eyes with and without vitrectomy before randomization in a DRCR.net trial that evaluated ranibizumab + prompt or deferred laser for diabetic macular edema is presented. Methods: Visual acuity and optical coherence tomography outcomes were compared between eyes with and without previous vitrectomy. Results: At baseline, eyes with previous vitrectomy (n = 25) had longer duration of diabetes, worse visual acuity, less thickened central subfield measurements on optical coherence tomography and were more apt to have worse diabetic retinopathy severity level or previous treatment for macular edema or cataract surgery than eyes without a history of vitrectomy (n = 335). Analyses adjusted for these baseline imbalances did not identify substantial differences between eyes with and without previous vitrectomy at each annual visit through 3 years for the favorable visual acuity, optical coherence tomography central subfield thickness, or volume outcomes, although optical coherence tomography improvement appeared slower in vitrectomy eyes during the first year. Conclusion: This study provides little evidence that the beneficial clinical outcomes for patients with center-involved diabetic macular edema treated with anti-vascular endothelial growth factor are affected in the long term by previous vitrectomy.","{'model': 'tldr@v2.0.0', 'text': 'This study provides little evidence that the beneficial clinical outcomes for patients with center-involved diabetic macular edema treated with anti-vascular endothelial growth factor are affected in the long term by previous vitrectomy.'}",https://europepmc.org/articles/pmc4658280?pdf=render
-validation of deep convolutional neural network-based algorithm for detection of diabetic retinopathy – artificial intelligence versus clinician for screening,P. Shah,"Purpose: Deep learning is a newer and advanced subfield in artificial intelligence (AI). The aim of our study is to validate a machine-based algorithm developed based on deep convolutional neural networks as a tool for screening to detect referable diabetic retinopathy (DR). Methods: An AI algorithm to detect DR was validated at our hospital using an internal dataset consisting of 1,533 macula-centered fundus images collected retrospectively and an external validation set using Methods to Evaluate Segmentation and Indexing Techniques in the field of Retinal Ophthalmology (MESSIDOR) dataset. Images were graded by two retina specialists as any DR, prompt referral (moderate nonproliferative diabetic retinopathy (NPDR) or above or presence of macular edema) and sight-threatening DR/STDR (severe NPDR or above) and compared with AI results. Sensitivity, specificity, and area under curve (AUC) for both internal and external validation sets for any DR detection, prompt referral, and STDR were calculated. Interobserver agreement using kappa value was calculated for both the sets and two out of three agreements for DR grading was considered as ground truth to compare with AI results. Results: In the internal validation set, the overall sensitivity and specificity was 99.7% and 98.5% for Any DR detection and 98.9% and 94.84%for Prompt referral respectively. The AUC was 0.991 and 0.969 for any DR detection and prompt referral respectively. The agreement between two observers was 99.5% and 99.2% for any DR detection and prompt referral with a kappa value of 0.94 and 0.96, respectively. In the external validation set (MESSIDOR 1), the overall sensitivity and specificity was 90.4% and 91.0% for any DR detection and 94.7% and 97.4% for prompt referral, respectively. The AUC was. 907 and. 960 for any DR detection and prompt referral, respectively. The agreement between two observers was 98.5% and 97.8% for any DR detection and prompt referral with a kappa value of 0.971 and 0.980, respectively. Conclusion: With increasing diabetic population and growing demand supply gap in trained resources, AI is the future for early identification of DR and reducing blindness. This can revolutionize telescreening in ophthalmology, especially where people do not have access to specialized health care.","{'model': 'tldr@v2.0.0', 'text': 'With increasing diabetic population and growing demand supply gap in trained resources, AI is the future for early identification of DR and reducing blindness, especially where people do not have access to specialized health care.'}",
-analytical computation of prompt gamma ray emission and detection for proton range verification,E. Sterpin,"A prompt gamma (PG) slit camera prototype recently demonstrated that Bragg Peak position in a clinical proton scanned beam could be measured with 1–2 mm accuracy by comparing an expected PG detection profile to a measured one. The computation of the expected PG detection profile in the context of a clinical framework is challenging but must be solved before clinical implementation. Obviously, Monte Carlo methods (MC) can simulate the expected PG profile but at prohibitively long calculation times. We implemented a much faster method that is based on analytical processing of precomputed MC data that would allow practical evaluation of this range monitoring approach in clinical conditions. Reference PG emission profiles were generated with MC simulations (PENH) in targets consisting of either 12C, 14N, 16O, 31P or 40Ca, with 10% of 1H. In a given geometry, the local PG emission can then be derived by adding the contribution of each element, according to the local energy of the proton obtained by continuous slowing down approximation and the local composition. The actual incident spot size is taken into account using an optical model fitted to measurements and by super sampling the spot with several rays (up to 113). PG transport in the patient/camera geometries and the detector response are modelled by convolving the PG production profile with a transfer function. The latter is interpolated from a database of transfer functions fitted to MC data (PENELOPE) generated for a photon source in a cylindrical phantom with various radiuses and a camera placed at various positions. As a benchmark, the analytical model was compared to MC and experiments in homogeneous and heterogeneous phantoms. Comparisons with MC were also performed in a thoracic CT. For all cases, the analytical model reproduced the prediction of the position of the Bragg peak computed with MC within 1 mm for the camera in nominal configuration. When compared to measurements, the shape of the profiles was well reproduced and agreement for the estimation of the position of the Bragg peak was within 2.7 mm on average (1.4 mm standard deviation). On a non-optimized MATLAB code, computation time with the analytical model is between 0.3 to 10 s depending on the number of rays simulated per spot. The analytical model can be further used to determine which spots are the best candidates to evaluate the range in clinical conditions and eventually correct for over- and under-shoots depending on the acquired PG profiles.","{'model': 'tldr@v2.0.0', 'text': 'A much faster method that is based on analytical processing of precomputed MC data that would allow practical evaluation of this range monitoring approach in clinical conditions and eventually correct for over- and under-shoots depending on the acquired PG profiles is implemented.'}",
-"who gets prompt access to artemisinin-based combination therapy? a prospective community-based study in children from rural kilosa, tanzania",D. Simba,"Background Effective and timely case management remains one of the fundamental pillars for control of malaria. Tanzania introduced artemisinin-combination therapy [ACT] for uncomplicated malaria; however, the policy change is challenged by limited availability of ACTs due to high cost. This study aimed to determine factors influencing prompt access to ACTs among febrile children in rural Kilosa, Tanzania. Methods and Findings In a community-based study, 1,235 randomly selected children under five were followed up weekly for six months, in 2008. Using a structured questionnaire, children's caretakers were asked about the child's febrile history in the last seven days, and treatment actions including timing, medicines used and source of care. Caretakers' knowledge about malaria and socioeconomic and demographic data were also obtained. About half of followed-up children had at least one episode of fever. Less than half (44.8%) of febrile children were taken to government facilities. Almost one-third (37.6%; 95% CI 33.1–42.1) of febrile children had prompt access to ACT. Care-seeking from a government facility was the overriding factor, increasing the likelihood of prompt access to an ACT 18 times (OR 17.7; 95% CI 10.55–29.54; adjusted OR 16.9; 95% CI 10.06–28.28). Caretakers from the better-off household (3rd–5th quintiles) were more likely to seek care from government facilities (OR 3.66; 95% CI 2.56–5.24; adjusted OR 1.80; 95% CI 1.18–2.76). The majority of antimalarials accessed by the poor were ineffective [86.0%; 295/343], however, they paid more for them (median Tsh 500) compared to the better-offs (median Tsh 0). Conclusions Prompt access to ACT among febrile children was unacceptably low, due mainly to limited availability of subsidised ACT at the location where most caretakers sought care. There is urgent need to accelerate implementation of strategies that will ensure availability of ACT at an affordable price in remote rural areas, where the burden of malaria is highest.","{'model': 'tldr@v2.0.0', 'text': None}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0012104&type=printable
-european consensus-based recommendations for the diagnosis and treatment of kawasaki disease – the share initiative,N. de Graeff,"OBJECTIVES The European Single Hub and Access point for paediatric Rheumatology in Europe initiative aimed to optimize care for children with rheumatic diseases. Kawasaki disease (KD) is the most common cause of acquired heart disease in children and an important cause of long-term cardiac disease into adulthood. Prompt diagnosis and treatment of KD is difficult due to the heterogeneity of the disease but is crucial for improving outcome. To date, there are no European internationally agreed, evidence-based guidelines concerning the diagnosis and treatment of KD in children. Accordingly, treatment regimens differ widely. The aim of this study is to provide consensus-based, European-wide evidence-informed recommendations for diagnosis and treatment of children with KD. METHODS Recommendations were developed using the EULAR's standard operating procedures. An extensive systematic literature search was performed, and evidence-based recommendations were extrapolated from the included papers. These were evaluated by a panel of international experts via online surveys and subsequently discussed in three consensus meetings, using nominal group technique. Recommendations were accepted when ⩾80% agreed. RESULTS In total, 17 recommendations for diagnosis and 14 for treatment of KD in children were accepted. Diagnostic recommendations included laboratory and imaging workup for complete as well as incomplete KD. Treatment recommendations included the importance of early treatment in both complete and incomplete KD, use of intravenous immunoglobulin, aspirin, corticosteroids for high-risk cases, and other treatment options for those with resistant disease. CONCLUSION The Single Hub and Access point for paediatric Rheumatology in Europe initiative provides international evidence-based recommendations for diagnosing and treating KD in children, facilitating improvement and uniformity of care.","{'model': 'tldr@v2.0.0', 'text': 'The Single Hub and Access point for paediatric Rheumatology in Europe initiative provides international evidence-based recommendations for diagnosing and treating Kawasaki disease in children, facilitating improvement and uniformity of care.'}",https://discovery.ucl.ac.uk/10072207/1/SHARE%20KD%20Recommendations%20-%20Rheumatology%20Manuscript%20-%20270618%20-%20Final.pdf
-automated essay scoring based on two-stage learning,Jiawei Liu,"Current state-of-art feature-engineered and end-to-end Automated Essay Score (AES) methods are proven to be unable to detect adversarial samples, e.g. the essays composed of permuted sentences and the prompt-irrelevant essays. Focusing on the problem, we develop a Two-Stage Learning Framework (TSLF) which integrates the advantages of both feature-engineered and end-to-end AES models. In experiments, we compare TSLF against a number of strong baselines, and the results demonstrate the effectiveness and robustness of our models. TSLF surpasses all the baselines on five-eighths of prompts and achieves new state-of-the-art average performance when without negative samples. After adding some adversarial essays to the original datasets, TSLF outperforms the feature-engineered and end-to-end baselines to a great extent, and shows great robustness.","{'model': 'tldr@v2.0.0', 'text': 'After adding some adversarial essays to the original datasets, TSLF outperforms the feature-engineered and end-to-end baselines to a great extent, and shows great robustness.'}",
-promptbert: improving bert sentence embeddings with prompts,Ting Jiang,"We propose PromptBERT, a novel contrastive learning method for learning better sentence representation. We firstly analysis the drawback of current sentence embedding from original BERT and find that it is mainly due to the static token embedding bias and ineffective BERT layers. Then we propose the first prompt-based sentence embeddings method and discuss two prompt representing methods and three prompt searching methods to make BERT achieve better sentence embeddings .Moreover, we propose a novel unsupervised training objective by the technology of template denoising, which substantially shortens the performance gap between the supervised and unsupervised settings. Extensive experiments show the effectiveness of our method. Compared to SimCSE, PromptBert achieves 2.29 and 2.58 points of improvement based on BERT and RoBERTa in the unsupervised setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel unsupervised training objective is proposed by the technology of template denoising, which substantially shortens the performance gap between the supervised and unsuper supervised settings.'}",https://aclanthology.org/2022.emnlp-main.603.pdf
-"fetal and neonatal alloimmune thrombocytopenia: recommendations for evidence‐based practice, an international approach",L. Lieberman,"Fetal and neonatal alloimmune thrombocytopenia (FNAIT) may result in severe bleeding, particularly fetal and neonatal intracranial haemorrhage (ICH). As a result, FNAIT requires prompt identification and treatment; subsequent pregnancies need close surveillance and management. An international panel convened to develop evidence‐based recommendations for diagnosis and management of FNAIT. A rigorous approach was used to search, review and develop recommendations from published data for: antenatal management, postnatal management, diagnostic testing and universal screening. To confirm FNAIT, fetal human platelet antigen (HPA) typing, using non‐invasive methods if quality‐assured, should be performed during pregnancy when the father is unknown, unavailable for testing or heterozygous for the implicated antigen. Women with a previous child with an ICH related to FNAIT should be offered intravenous immunoglobulin (IVIG) infusions during subsequent affected pregnancies as early as 12 weeks gestation. Ideally, HPA‐selected platelets should be available at delivery for potentially affected infants and used to increase the neonatal platelet count as needed. If HPA‐selected platelets are not immediately available, unselected platelets should be transfused. FNAIT studies that optimize antenatal and postnatal management, develop risk stratification algorithms to guide management and standardize laboratory testing to identify high risk pregnancies are needed.","{'model': 'tldr@v2.0.0', 'text': 'FNAIT studies that optimize antenatal and postnatal management, develop risk stratification algorithms to guide management and standardize laboratory testing to identify high risk pregnancies are needed.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/bjh.15813
-how can we know what language models know?,Zhengbao Jiang,"Abstract Recent work has presented intriguing results examining the knowledge contained in language models (LMs) by having the LM fill in the blanks of prompts such as “Obama is a __ by profession”. These prompts are usually manually created, and quite possibly sub-optimal; another prompt such as “Obama worked as a __ ” may result in more accurately predicting the correct profession. Because of this, given an inappropriate prompt, we might fail to retrieve facts that the LM does know, and thus any given prompt only provides a lower bound estimate of the knowledge contained in an LM. In this paper, we attempt to more accurately estimate the knowledge contained in LMs by automatically discovering better prompts to use in this querying process. Specifically, we propose mining-based and paraphrasing-based methods to automatically generate high-quality and diverse prompts, as well as ensemble methods to combine answers from different prompts. Extensive experiments on the LAMA benchmark for extracting relational knowledge from LMs demonstrate that our methods can improve accuracy from 31.1% to 39.6%, providing a tighter lower bound on what LMs know. We have released the code and the resulting LM Prompt And Query Archive (LPAQA) at https://github.com/jzbjyb/LPAQA.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes mining-based and paraphrasing-based methods to automatically generate high-quality and diverse prompts, as well as ensemble methods to combine answers from different prompts to provide a tighter lower bound on what LMs know.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00324/1923867/tacl_a_00324.pdf
-realfusion 360° reconstruction of any object from a single image,Luke Melas-Kyriazi,"We consider the problem of reconstructing a full 360° photographic model of an object from a single image of it. We do so by fitting a neural radiance field to the image, but find this problem to be severely ill-posed. We thus take an off-the-self conditional image generator based on diffusion and engineer a prompt that encourages it to “dream up” novel views of the object. Using the recent DreamFusion method, we fuse the given input view, the conditional prior, and other regularizers into a final, consistent reconstruction. We demonstrate state-of-the-art reconstruction results on benchmark images when compared to prior methods for monocular 3D reconstruction of objects. Qualitatively, our reconstructions provide a faithful match of the input view and a plausible extrapolation of its appearance and 3D shape, including to the side of the object not visible in the image.",,https://arxiv.org/pdf/2302.10663
-active prompting with chain-of-thought for large language models,Shizhe Diao,"The increasing scale of large language models (LLMs) brings emergent abilities to various complex tasks requiring reasoning, such as arithmetic and commonsense reasoning. It is known that the effective design of task-specific prompts is critical for LLMs' ability to produce high-quality answers. In particular, an effective approach for complex question-and-answer tasks is example-based prompting with chain-of-thought (CoT) reasoning, which significantly improves the performance of LLMs. However, current CoT methods rely on a fixed set of human-annotated exemplars, which are not necessarily the most effective examples for different tasks. This paper proposes a new method, Active-Prompt, to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning). For this purpose, we propose a solution to the key problem of determining which questions are the most important and helpful ones to annotate from a pool of task-specific queries. By borrowing ideas from the related problem of uncertainty-based active learning, we introduce several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation. Experimental results demonstrate the superiority of our proposed method, achieving state-of-the-art on eight complex reasoning tasks. Further analyses of different uncertainty metrics, pool sizes, zero-shot learning, and accuracy-uncertainty relationship demonstrate the effectiveness of our method. Our code will be available at https://github.com/shizhediao/active-prompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new method to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning), and introduces several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation.'}",http://arxiv.org/pdf/2302.12246
-blended diffusion for text-driven editing of natural images,Omri Avrahami,"Natural language offers a highly intuitive interface for image editing. In this paper, we introduce the first solution for performing local (region-based) edits in generic natural images, based on a natural language description along with an ROI mask. We achieve our goal by leveraging and combining a pretrained language-image model (CLIP), to steer the edit towards a user-provided text prompt, with a denoising diffusion probabilistic model (DDPM) to generate natural-looking results. To seamlessly fuse the edited region with the unchanged parts of the image, we spatially blend noised versions of the input image with the local text-guided diffusion latent at a progression of noise levels. In addition, we show that adding augmentations to the diffusion process mitigates adversarial results. We compare against several baselines and related methods, both qualitatively and quantitatively, and show that our method outperforms these solutions in terms of overall realism, ability to preserve the background and matching the text. Finally, we show several text-driven editing applications, including adding a new object to an image, removing/replacing/altering existing objects, background replacement, and image extrapolation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the first solution for performing local (region-based) edits in generic natural images, based on a natural language description along with an ROI mask, and shows several text-driven editing applications, including adding a new object to an image, removing/replacing/altering existing objects, background replacement, and image extrapolation.'}",https://arxiv.org/pdf/2111.14818
-"diagnosis, treatment, and long-term management of kawasaki disease: a scientific statement for health professionals from the american heart association",B. McCrindle,"Background: Kawasaki disease is an acute vasculitis of childhood that leads to coronary artery aneurysms in ≈25% of untreated cases. It has been reported worldwide and is the leading cause of acquired heart disease in children in developed countries. Methods and Results: To revise the previous American Heart Association guidelines, a multidisciplinary writing group of experts was convened to review and appraise available evidence and practice-based opinion, as well as to provide updated recommendations for diagnosis, treatment of the acute illness, and long-term management. Although the cause remains unknown, discussion sections highlight new insights into the epidemiology, genetics, pathogenesis, pathology, natural history, and long-term outcomes. Prompt diagnosis is essential, and an updated algorithm defines supplemental information to be used to assist the diagnosis when classic clinical criteria are incomplete. Although intravenous immune globulin is the mainstay of initial treatment, the role for additional primary therapy in selected patients is discussed. Approximately 10% to 20% of patients do not respond to initial intravenous immune globulin, and recommendations for additional therapies are provided. Careful initial management of evolving coronary artery abnormalities is essential, necessitating an increased frequency of assessments and escalation of thromboprophylaxis. Risk stratification for long-term management is based primarily on maximal coronary artery luminal dimensions, normalized as Z scores, and is calibrated to both past and current involvement. Patients with aneurysms require life-long and uninterrupted cardiology follow-up. Conclusions: These recommendations provide updated and best evidence-based guidance to healthcare providers who diagnose and manage Kawasaki disease, but clinical decision making should be individualized to specific patient circumstances.","{'model': 'tldr@v2.0.0', 'text': 'These recommendations provide updated and best evidence-based guidance to healthcare providers who diagnose and manage Kawasaki disease, but clinical decision making should be individualized to specific patient circumstances.'}",https://www.ahajournals.org/doi/pdf/10.1161/CIR.0000000000000484
-gpt3mix: leveraging large-scale language models for text augmentation,Kang Min Yoo,"Large-scale language models such as GPT-3 are excellent few-shot learners, allowing them to be controlled via natural text prompts. Recent studies report that prompt-based direct classification eliminates the need for fine-tuning but lacks data and inference scalability. This paper proposes a novel data augmentation technique that leverages large-scale language models to generate realistic text samples from a mixture of real samples. We also propose utilizing soft-labels predicted by the language models, effectively distilling knowledge from the large-scale language models and creating textual perturbations simultaneously. We perform data augmentation experiments on diverse classification tasks and show that our method hugely outperforms existing text augmentation methods. Ablation studies and a qualitative analysis provide more insights into our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel data augmentation technique that leverages large-scale language models to generate realistic text samples from a mixture of real samples and utilizes soft-labels predicted by the language models, effectively distilling knowledge from the large- scale language models and creating textual perturbations simultaneously.'}",https://aclanthology.org/2021.findings-emnlp.192.pdf
-optimizing rating scale category effectiveness.,J. Linacre,"Rating scales are employed as a means of extracting more information out of an item than would be obtained from a mere ""yes/no"", ""right/wrong"" or other dichotomy. But does this additional information increase measurement accuracy and precision? Eight guidelines are suggested to aid the analyst in optimizing the manner in which rating scales categories cooperate in order to improve the utility of the resultant measures. Though these guidelines are presented within the context of Rasch analysis, they reflect aspects of rating scale functioning which impact all methods of analysis. The guidelines feature rating-scale-based data such as category frequency, ordering, rating-to-measure inferential coherence, and the quality of the scale from measurement and statistical perspectives. The manner in which the guidelines prompt recategorization or reconceptualization of the rating scale is indicated. Utilization of the guidelines is illustrated through their application to two published data sets.","{'model': 'tldr@v2.0.0', 'text': 'Eight guidelines are suggested to aid the analyst in optimizing the manner in which rating scales categories cooperate in order to improve the utility of the resultant measures.'}",
-warp: word-level adversarial reprogramming,Karen Hambardzumyan,"Transfer learning from pretrained language models recently became the dominant approach for solving many NLP tasks. A common approach to transfer learning for multiple tasks that maximize parameter sharing trains one or more task-specific layers on top of the language model. In this paper, we present an alternative approach based on adversarial reprogramming, which extends earlier work on automatic prompt generation. Adversarial reprogramming attempts to learn task-specific word embeddings that, when concatenated to the input text, instruct the language model to solve the specified task. Using up to 25K trainable parameters per task, this approach outperforms all existing methods with up to 25M trainable parameters on the public leaderboard of the GLUE benchmark. Our method, initialized with task-specific human-readable prompts, also works in a few-shot setting, outperforming GPT-3 on two SuperGLUE tasks with just 32 training samples.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents an alternative approach based on adversarial reprogramming, which extends earlier work on automatic prompt generation, and outperforms all existing methods with up to 25M trainable parameters on the public leaderboard of the GLUE benchmark.'}",https://aclanthology.org/2021.acl-long.381.pdf
-an automated essay scoring systems: a systematic literature review,D. Ramesh,,"{'model': 'tldr@v2.0.0', 'text': 'It is observed that the essay evaluation is not done based on the relevance of the content and coherence, which means that there is no proper evaluation system for grading essays and short answers.'}",https://link.springer.com/content/pdf/10.1007/s10462-021-10068-2.pdf
-prompting for multimodal hateful meme classification,Rui Cao,"Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge. Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes. However, there is no known explicit external knowledge base that could provide such hate speech contextual information. To address this gap, we propose PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification. Specifically, we construct simple prompts and provide a few in-context examples to exploit the implicit knowledge in the pre-trained RoBERTa language model for hateful meme classification. We conduct extensive experiments on two publicly available hateful and offensive meme datasets. Our experiment results show that PromptHate is able to achieve a high AUC of 90.96, outperforming state-of-the-art baselines on the hateful meme classification task. We also perform fine-grain analyses and case studies on various prompt settings and demonstrate the effectiveness of the prompts on hateful meme classification.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification, and constructs simple prompts and provides a few in-context examples to exploit the implicit knowledge in the pre- trained RoBERTa language model for hateful memes classification.'}",http://arxiv.org/pdf/2302.04156
-μema: microinteraction-based ecological momentary assessment (ema) using a smartwatch,S. Intille,"Ecological Momentary Assessment (EMA) is a method of in situ data collection for assessment of behaviors, states, and contexts. Questions are prompted during everyday life using an individual's mobile device, thereby reducing recall bias and increasing validity over other self-report methods such as retrospective recall. We describe a microinteraction-based EMA method (""micro"" EMA, or μEMA) using smartwatches, where all EMA questions can be answered with a quick glance and a tap -- nearly as quickly as checking the time on a watch. A between-subjects, 4-week pilot study was conducted where μEMA on a smartwatch (n=19) was compared with EMA on a phone (n=14). Despite an =8 times increase in the number of interruptions, μEMA had a significantly higher compliance rate, completion rate, and first prompt response rate, and μEMA was perceived as less distracting. The temporal density of data collection possible with μEMA could prove useful in ubiquitous computing studies.","{'model': 'tldr@v2.0.0', 'text': 'A microinteraction-based EMA method (""micro"" EMA, or μEMA) using smartwatches, where all EMA questions can be answered with a quick glance and a tap -- nearly as quickly as checking the time on a watch.'}",https://europepmc.org/articles/pmc6143290?pdf=render
-badprompt: backdoor attacks on continuous prompts,Xiangrui Cai,"The prompt-based learning paradigm has gained much research attention recently. It has achieved state-of-the-art performance on several NLP tasks, especially in the few-shot scenarios. While steering the downstream tasks, few works have been reported to investigate the security problems of the prompt-based models. In this paper, we conduct the first study on the vulnerability of the continuous prompt learning algorithm to backdoor attacks. We observe that the few-shot scenarios have posed a great challenge to backdoor attacks on the prompt-based models, limiting the usability of existing NLP backdoor methods. To address this challenge, we propose BadPrompt, a lightweight and task-adaptive algorithm, to backdoor attack continuous prompts. Specially, BadPrompt first generates candidate triggers which are indicative for predicting the targeted label and dissimilar to the samples of the non-targeted labels. Then, it automatically selects the most effective and invisible trigger for each sample with an adaptive trigger optimization algorithm. We evaluate the performance of BadPrompt on five datasets and two continuous prompt models. The results exhibit the abilities of BadPrompt to effectively attack continuous prompts while maintaining high performance on the clean test sets, outperforming the baseline models by a large margin. The source code of BadPrompt is publicly available at https://github.com/papersPapers/BadPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper conducts the first study on the vulnerability of the continuous prompt learning algorithm to backdoor attacks, and proposes BadPrompt, a lightweight and task-adaptive algorithm, to backdoor attack continuous prompts.'}",https://arxiv.org/pdf/2211.14719
-microvascular decompression surgery: surgical principles and technical nuances based on 4000 cases,J. Zhong,"Abstract Background: As an etiological treatment of trigeminal neuralgia (TN) and hemifacial spasm (HFS), microvascular decompression (MVD) has been popularized around the world. However, as a functional operation in the cerebellopontine angle (CPA), this process can be risky and the postoperative outcomes might not be good enough sometimes. Objective: In order to obtain a better result with less complication, this surgery should be further addressed. Methods: With experience of more than 4000 MVDs, we have gained knowledge about the operative technique. Through abundant intraoperative photos, each step of the procedure was demonstrated in detail and the surgical strategy was focused. Results: The principle of MVD is to separate the nerve-vessel confliction rather than isolate it with prostheses. A prompt identification of the conflict site is important, which hinges on a good exposure. A satisfactory working space can be established by an appropriate positioning of the patient's head and a proper craniectomy as well as a rational approach. A sharp dissection of arachnoids leads to a maximal visualization of the entire intracranial course of the nerve root. All the vessels contacting the trigeminal or facial nerve should be treated. Intraoperative electrophysiological mentoring is helpful to distinguish the offending artery for hemifacial cases. Conclusion: MVD is an effective treatment for the patient with TN or HFS. Immediate relief can be achieved by an experienced neurosurgeon with good knowledge of regional anatomy. A safe surgery is the tenet of MVD, and accordingly, no single step of the procedure should be ignored.","{'model': 'tldr@v2.0.0', 'text': 'MVD is an effective treatment for the patient with TN or HFS and can be achieved by an experienced neurosurgeon with good knowledge of regional anatomy and no single step of the procedure should be ignored.'}",
-diagnosis of dengue infection using conventional and biosensor based techniques,O. Parkash,"Dengue is an arthropod-borne viral disease caused by four antigenically different serotypes of dengue virus. This disease is considered as a major public health concern around the world. Currently, there is no licensed vaccine or antiviral drug available for the prevention and treatment of dengue disease. Moreover, clinical features of dengue are indistinguishable from other infectious diseases such as malaria, chikungunya, rickettsia and leptospira. Therefore, prompt and accurate laboratory diagnostic test is urgently required for disease confirmation and patient triage. The traditional diagnostic techniques for the dengue virus are viral detection in cell culture, serological testing, and RNA amplification using reverse transcriptase PCR. This paper discusses the conventional laboratory methods used for the diagnosis of dengue during the acute and convalescent phase and highlights the advantages and limitations of these routine laboratory tests. Subsequently, the biosensor based assays developed using various transducers for the detection of dengue are also reviewed.","{'model': 'tldr@v2.0.0', 'text': 'The conventional laboratory methods used for the diagnosis of dengue during the acute and convalescent phase are discussed and the advantages and limitations of these routine laboratory tests are highlighted.'}",https://www.mdpi.com/1999-4915/7/10/2877/pdf?version=1445253550
-a review on fault diagnosis and fault tolerant control methods for single-rotor aerial vehicles,Xin Qi,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the recent development and current researches in the field of fault diagnosis, including analytical/model-based, signal processing-based and knowledge-based techniques, and also passive/active fault- tolerant control approaches is presented.'}",http://ir.sia.cn//bitstream/173321/14053/4/A%20Review%20on%20Fault%20Diagnosis%20and%20Fault%20Tolerant%20Control%20Methods%20for%20Single-rotor%20Aerial%20Vehicles.pdf
-are we overpathologizing everyday life? a tenable blueprint for behavioral addiction research,J. Billieux,"Background Behavioral addiction research has been particularly flourishing over the last two decades. However, recent publications have suggested that nearly all daily life activities might lead to a genuine addiction. Methods and aim In this article, we discuss how the use of atheoretical and confirmatory research approaches may result in the identification of an unlimited list of “new” behavioral addictions. Results Both methodological and theoretical shortcomings of these studies were discussed. Conclusions We suggested that studies overpathologizing daily life activities are likely to prompt a dismissive appraisal of behavioral addiction research. Consequently, we proposed several roadmaps for future research in the field, centrally highlighting the need for longer tenable behavioral addiction research that shifts from a mere criteria-based approach toward an approach focusing on the psychological processes involved.","{'model': 'tldr@v2.0.0', 'text': 'This article discussed how the use of atheoretical and confirmatory research approaches may result in the identification of an unlimited list of “new” behavioral addictions, and suggested that studies overpathologizing daily life activities are likely to prompt a dismissive appraisal of behavioral addiction research.'}",https://akjournals.com/downloadpdf/journals/2006/4/3/article-p119.pdf
-in vivo measurements of body protein based on the determination of nitrogen by prompt gamma analysis.,K. Mcneill,"Measurement of nitrogen in the human body gives a measure of total protein in the body. Nitrogen (N) may be measured in vivo by a nuclear technique in which the characteristic nitrogen gamma rays emitted after capture of neutrons are detected and counted. The accuracy (as determined by comparison of the measurement of nitrogen in a pig by nuclear and chemical methods) is better than 10%. The reproducibility is less than 6%. The radiation dose is about 5 millirads. This method had been used with 43 patients and 23 normal volunteers and the results are reported. The data are presented in the from of total N content, percentage N by weight, and as a comparison of N and potassium (K) in the body. It is found that the mean percentage of N in the normal males is 2.5%, giving a value of 1.75 kg in the 70-kg standard man. The mean percentage values are not greatly different between normal volunteers and malnourished patients. Taking all subjects, N/K ratios ranged from 7.5 to 20 g/g. Thus K cannot be used as a predictor of N. Nitrogen content of normals correlates well (r = 0.68) with height squared; for malnourished patients there is again a good correlation (r = 0.81) but the regression line for patients is approximately 34% below that for the volunteers consistent with their depleted state.","{'model': 'tldr@v2.0.0', 'text': 'Nitrogen content of normals correlates well with height squared; for malnourished patients there is again a good correlation but the regression line for patients is approximately 34% below that for the volunteers consistent with their depleted state, so K cannot be used as a predictor of N.'}",
-multilingual relation classification via efficient and effective prompting,Yuxuan Chen,"Prompting pre-trained language models has achieved impressive performance on various NLP tasks, especially in low data regimes. Despite the success of prompting in monolingual settings, applying prompt-based methods in multilingual scenarios has been limited to a narrow set of tasks, due to the high cost of handcrafting multilingual prompts. In this paper, we present the first work on prompt-based multilingual relation classification (RC), by introducing an efficient and effective method that constructs prompts from relation triples and involves only minimal translation for the class labels. We evaluate its performance in fully supervised, few-shot and zero-shot scenarios, and analyze its effectiveness across 14 languages, prompt variants, and English-task training in cross-lingual settings. We find that in both fully supervised and few-shot scenarios, our prompt method beats competitive baselines: fine-tuning XLM-R_EM and null prompts. It also outperforms the random baseline by a large margin in zero-shot experiments. Our method requires little in-language knowledge and can be used as a strong baseline for similar multilingual classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the first work on prompt-based multilingual relation classification (RC), by introducing an efficient and effective method that constructs prompts from relation triples and involves only minimal translation for the class labels.'}",https://arxiv.org/pdf/2210.13838
-measurement of the differential cross section for isolated prompt photon production in pp collisions at 7 tev,S. Chatrchyan,"A measurement of the differential cross section for the inclusive production of isolated prompt photons in proton-proton collisions at a centre-of-mass energy of 7 TeV is presented. The data sample corresponds to an integrated luminosity of 36 inverse picobarns recorded by the CMS detector at the LHC. The measurement covers the pseudorapidity range |eta|<2.5 and the transverse energy range 25 < ET < 400 GeV, corresponding to the kinematic region 0.007 < xT < 0.114. Photon candidates are identified with two complementary methods, one based on photon conversions in the silicon tracker and the other on isolated energy deposits in the electromagnetic calorimeter. The measured cross section is presented as a function of ET in four pseudorapidity regions. The next-to-leading-order perturbative QCD calculations are consistent with the measured cross section.",,http://link.aps.org/pdf/10.1103/PhysRevD.84.052011
-electrocardiogram pattern recognition and analysis based on artificial neural networks and support vector machines: a review.,M. Sansone,"Computer systems for Electrocardiogram (ECG) analysis support the clinician in tedious tasks (e.g., Holter ECG monitored in Intensive Care Units) or in prompt detection of dangerous events (e.g., ventricular fibrillation). Together with clinical applications (arrhythmia detection and heart rate variability analysis), ECG is currently being investigated in biometrics (human identification), an emerging area receiving increasing attention. Methodologies for clinical applications can have both differences and similarities with respect to biometrics. This paper reviews methods of ECG processing from a pattern recognition perspective. In particular, we focus on features commonly used for heartbeat classification. Considering the vast literature in the field and the limited space of this review, we dedicated a detailed discussion only to a few classifiers (Artificial Neural Networks and Support Vector Machines) because of their popularity; however, other techniques such as Hidden Markov Models and Kalman Filtering will be also mentioned.","{'model': 'tldr@v2.0.0', 'text': 'This paper reviews methods of ECG processing from a pattern recognition perspective and focuses on features commonly used for heartbeat classification, mainly Artificial Neural Networks and Support Vector Machines.'}",https://downloads.hindawi.com/journals/jhe/2013/904584.pdf
-understanding and managing methotrexate nephrotoxicity.,B. Widemann,"Methotrexate (MTX) is one of the most widely used anti-cancer agents, and administration of high-dose methotrexate (HDMTX) followed by leucovorin (LV) rescue is an important component in the treatment of a variety of childhood and adult cancers. HDMTX can be safely administered to patients with normal renal function by the use of alkalinization, hydration, and pharmacokinetically guided LV rescue. Despite these measures, HDMTX-induced renal dysfunction continues to occur in approximately 1.8% of patients with osteosarcoma treated on clinical trials. Prompt recognition and treatment of MTX-induced renal dysfunction are essential to prevent potentially life-threatening MTX-associated toxicities, especially myelosuppression, mucositis, and dermatitis. In addition to conventional treatment approaches, dialysis-based methods have been used to remove MTX with limited effectiveness. More recently carboxypeptidase-G(2) (CPDG(2)), a recombinant bacterial enzyme that rapidly hydrolyzes MTX to inactive metabolites, has become available for the treatment of HDMTX-induced renal dysfunction. CPDG(2) administration has been well tolerated and resulted in consistent and rapid reductions in plasma MTX concentrations by a median of 98.7% (range, 84%-99.5%). The early administration of CPDG(2) in addition to LV may be beneficial for patients with MTX-induced renal dysfunction and significantly elevated plasma MTX concentrations.","{'model': 'tldr@v2.0.0', 'text': 'CPDG(2) administration has been well tolerated and resulted in consistent and rapid reductions in plasma MTX concentrations by a median of 98.7% (range, 84%-99.5%).'}",http://theoncologist.alphamedpress.org/content/11/6/694.full.pdf
-changes in teaching efficacy during a professional development school-based science methods course.,S. Swars,"This mixed methods study offers a theoretically grounded description of a field-based science methods course within a Professional Development School (PDS) model (i.e., PDS-based course). The preservice teachers' (n= 21) experiences within the PDS-based course prompted significant changes in their personal teaching efficacy, with the opportunities working with children afforded in the course attributed as a source of these beliefs. However, the preservice teachers' teaching outcome expectancy beliefs did not significantly shift. The results of this study support the extant literature in that field experiences with science methods courses can facilitate preservice teacher development in meaningful ways. This study expands upon this literature by considering a PDS context for science teacher preparation, more specifically, a science methods course purposefully integrated in elementary classrooms at a PDS where preservice teachers facilitated scientific inquiry projects with children. The findings should prompt new ways of thinking about teacher preparation, particularly related to science, that provide systematic and intentional connectivity between university programs and K–12 schools so preservice teachers can connect theory and practice.",,
-paper-based point-of-care testing for diagnosis of dengue infections,Jane Ru Choi,"Abstract Dengue endemic is a serious healthcare concern in tropical and subtropical countries. Although well-established laboratory tests can provide early diagnosis of acute dengue infections, access to these tests is limited in developing countries, presenting an urgent need to develop simple, rapid, and robust diagnostic tools. Point-of-care (POC) devices, particularly paper-based POC devices, are typically rapid, cost-effective and user-friendly, and they can be used as diagnostic tools for the prompt diagnosis of dengue at POC settings. Here, we review the importance of rapid dengue diagnosis, current dengue diagnostic methods, and the development of paper-based POC devices for diagnosis of dengue infections at the POC.","{'model': 'tldr@v2.0.0', 'text': 'The importance of rapid d Dengue diagnosis, current dengue diagnostic methods, and the development of paper-based POC devices for diagnosis of dengu infections at the POC are reviewed.'}",
-knowledge prompting in pre-trained language model for natural language understanding,J. Wang,"Knowledge-enhanced Pre-trained Language Model (PLM) has recently received significant attention, which aims to incorporate factual knowledge into PLMs. However, most existing methods modify the internal structures of fixed types of PLMs by stacking complicated modules, and introduce redundant and irrelevant factual knowledge from knowledge bases (KBs). In this paper, to address these problems, we introduce a seminal knowledge prompting paradigm and further propose a knowledge-prompting-based PLM framework KP-PLM. This framework can be flexibly combined with existing mainstream PLMs. Specifically, we first construct a knowledge sub-graph from KBs for each context. Then we design multiple continuous prompts rules and transform the knowledge sub-graph into natural language prompts. To further leverage the factual knowledge from these prompts, we propose two novel knowledge-aware self-supervised tasks including prompt relevance inspection and masked prompt modeling. Extensive experiments on multiple natural language understanding (NLU) tasks show the superiority of KP-PLM over other state-of-the-art methods in both full-resource and low-resource settings. Our source codes will be released upon the acceptance of the paper.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a seminal knowledge prompting paradigm and proposes a knowledge-prompting-based PLM framework KP-PLM that can be flexibly combined with existing mainstream PLMs and proposes two novel knowledge-aware self-supervised tasks.'}",http://arxiv.org/pdf/2210.08536
-"event-by-event evaluation of the prompt fission neutron spectrum from 239pu(n,f)",R. Vogt,"We have developed an improved evaluation method for the spectrum of neutrons emitted in fission of {sup 239}Pu induced by incident neutrons with energies up to 20 MeV. The covariance data, including incident energy correlations introduced by the evaluation method, were used to fix the input parameters in our event-by-event model of fission, FREYA, by applying formal statistical methods. Formal estimates of uncertainties in the evaluation were developed by randomly sampling model inputs and calculating likelihood functions based on agreement with the evaluated . Our approach is able to employ a greater variety of fission measurements than the relatively coarse spectral data alone. It also allows the study of numerous fission observables for more accurate model validation. The combination of an event-by-event Monte Carlo fission model with a statistical-likelihood analysis is thus a powerful tool for evaluation of fission-neutron data. Our empirical model FREYA follows the complete fission event from birth of the excited fragments through their decay via neutron emission until the fragment excitation energy is below the neutron separation energy when neutron emission can no longer occur. The most recent version of FREYA incorporates pre-equilibrium neutron emission, the emission of the first neutron before equilibrium ismore » reached in the compound nucleus, and multi-chance fission, neutron evaporation prior to fission when the incident neutron energy is above the neutron separation energy. Energy, momentum, charge and mass number are conserved throughout the fission process. The best available values of fragment masses and total kinetic energies are used as inputs to FREYA. We fit three parameters that are not well under control from previous measurements: the shift in the total fragment kinetic energy; the energy scale of the asymptotic level density parameter, controlling the fragment 'temperature' for neutron evaporation; and the relative excitation of the light and heavy fragments, governing the number and energy of neutrons emitted from each fragment. The latter two parameters are assumed to be independent of the incident neutron energy while the first varies with incident energy. We describe our method and the subsequent spectral evaluation and present the results of several standard validation calculations that test our new evaluation. These benchmarks include critical assemblies, sensitive to criticality in fast systems; pulsed sphere measurements testing the spectra at incident neutron energies of 14 MeV; and other tests.« less",,https://link.aps.org/accepted/10.1103/PhysRevC.85.024608
-"frequently asked questions about chlorophyll fluorescence, the sequel",H. Kalaji,,"{'model': 'tldr@v2.0.0', 'text': 'Questions about instruments, methods and applications based on chlorophyll a fluorescence, and the answers draw on knowledge from different Chl a Fluorescence analysis domains, yielding in several cases new insights.'}",https://link.springer.com/content/pdf/10.1007%2Fs11120-016-0318-y.pdf
-evidence-based diagnostics: adult septic arthritis.,C. Carpenter,"BACKGROUND Acutely swollen or painful joints are common complaints in the emergency department (ED). Septic arthritis in adults is a challenging diagnosis, but prompt differentiation of a bacterial etiology is crucial to minimize morbidity and mortality. OBJECTIVES The objective was to perform a systematic review describing the diagnostic characteristics of history, physical examination, and bedside laboratory tests for nongonococcal septic arthritis. A secondary objective was to quantify test and treatment thresholds using derived estimates of sensitivity and specificity, as well as best-evidence diagnostic and treatment risks and anticipated benefits from appropriate therapy. METHODS Two electronic search engines (PUBMED and EMBASE) were used in conjunction with a selected bibliography and scientific abstract hand search. Inclusion criteria included adult trials of patients presenting with monoarticular complaints if they reported sufficient detail to reconstruct partial or complete 2 × 2 contingency tables for experimental diagnostic test characteristics using an acceptable criterion standard. Evidence was rated by two investigators using the Quality Assessment Tool for Diagnostic Accuracy Studies (QUADAS). When more than one similarly designed trial existed for a diagnostic test, meta-analysis was conducted using a random effects model. Interval likelihood ratios (LRs) were computed when possible. To illustrate one method to quantify theoretical points in the probability of disease whereby clinicians might cease testing altogether and either withhold treatment (test threshold) or initiate definitive therapy in lieu of further diagnostics (treatment threshold), an interactive spreadsheet was designed and sample calculations were provided based on research estimates of diagnostic accuracy, diagnostic risk, and therapeutic risk/benefits. RESULTS The prevalence of nongonococcal septic arthritis in ED patients with a single acutely painful joint is approximately 27% (95% confidence interval [CI] = 17% to 38%). With the exception of joint surgery (positive likelihood ratio [+LR] = 6.9) or skin infection overlying a prosthetic joint (+LR = 15.0), history, physical examination, and serum tests do not significantly alter posttest probability. Serum inflammatory markers such as white blood cell (WBC) counts, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) are not useful acutely. The interval LR for synovial white blood cell (sWBC) counts of 0 × 10(9)-25 × 10(9)/L was 0.33; for 25 × 10(9)-50 × 10(9)/L, 1.06; for 50 × 10(9)-100 × 10(9)/L, 3.59; and exceeding 100 × 10(9)/L, infinity. Synovial lactate may be useful to rule in or rule out the diagnosis of septic arthritis with a +LR ranging from 2.4 to infinity, and negative likelihood ratio (-LR) ranging from 0 to 0.46. Rapid polymerase chain reaction (PCR) of synovial fluid may identify the causative organism within 3 hours. Based on 56% sensitivity and 90% specificity for sWBC counts of >50 × 10(9)/L in conjunction with best-evidence estimates for diagnosis-related risk and treatment-related risk/benefit, the arthrocentesis test threshold is 5%, with a treatment threshold of 39%. CONCLUSIONS Recent joint surgery or cellulitis overlying a prosthetic hip or knee were the only findings on history or physical examination that significantly alter the probability of nongonococcal septic arthritis. Extreme values of sWBC (>50 × 10(9)/L) can increase, but not decrease, the probability of septic arthritis. Future ED-based diagnostic trials are needed to evaluate the role of clinical gestalt and the efficacy of nontraditional synovial markers such as lactate.","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the diagnostic characteristics of history, physical examination, and bedside laboratory tests for nongonococcal septic arthritis found recent joint surgery or cellulitis overlying a prosthetic hip or knee were the only findings on history or physical examination that significantly alter the probability of nong onococcalSeptic arthritis.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1553-2712.2011.01121.x
-mluke: the power of entity representations in multilingual pretrained language models,Ryokan Ri,"Recent studies have shown that multilingual pretrained language models can be effectively improved with cross-lingual alignment information from Wikipedia entities.However, existing methods only exploit entity information in pretraining and do not explicitly use entities in downstream tasks.In this study, we explore the effectiveness of leveraging entity representations for downstream cross-lingual tasks.We train a multilingual language model with 24 languages with entity representations and showthe model consistently outperforms word-based pretrained models in various cross-lingual transfer tasks.We also analyze the model and the key insight is that incorporating entity representations into the input allows us to extract more language-agnostic features.We also evaluate the model with a multilingual cloze prompt task with the mLAMA dataset.We show that entity-based prompt elicits correct factual knowledge more likely than using only word representations.","{'model': 'tldr@v2.0.0', 'text': 'A multilingual language model with 24 languages with entity representations is trained and it is shown that the model consistently outperforms word-based pretrained models in various cross-lingual transfer tasks.'}",https://aclanthology.org/2022.acl-long.505.pdf
-photophysics of thermally activated delayed fluorescence molecules,F. Dias,"Thermally activated delayed fluorescence (TADF) has recently emerged as one of the most attractive methods for harvesting triplet states in metal-free organic materials for application in organic light emitting diodes (OLEDs). A large number of TADF molecules have been reported in the literature with the purpose of enhancing the efficiency of OLEDs by converting non-emissive triplet states into emissive singlet states. TADF emitters are able to harvest both singlets and triplet states through fluorescence (prompt and delayed), the latter due to the thermally activated reverse intersystem crossing mechanism that allows up-conversion of low energy triplet states to the emissive singlet level. This allows otherwise pure fluorescent OLEDs to overcome their intrinsic limit of 25% internal quantum efficiency (IQE), which is imposed by the 1:3 singlet–triplet ratio arising from the recombination of charges (electrons and holes). TADF based OLEDS with IQEs close to 100% are now routinely fabricated in the green spectral region. There is also significant progress for blue emitters. However, red emitters still show relatively low efficiencies. Despite the significant progress that has been made in recent years, still significant challenges persist to achieve full understanding of the TADF mechanism and improve the stability of these materials. These questions need to be solved in order to fully implement TADF in OLEDs and expand their application to other areas. To date, TADF has been exploited mainly in the field of OLEDs, but applications in other areas, such as sensing and fluorescence microscopies, are envisaged. In this review, the photophysics of TADF molecules is discussed, summarising current methods to characterise these materials and the current understanding of the TADF mechanism in various molecular systems.","{'model': 'tldr@v2.0.0', 'text': 'Thermally activated delayed fluorescence has recently emerged as one of the most attractive methods for harvesting triplet states in metal-free organic materials for application in organic light emitting diodes (OLEDs).'}",https://iopscience.iop.org/article/10.1088/2050-6120/aa537e/pdf
-"necrotizing fasciitis: the importance of early diagnosis, prompt surgical debridement and adjuvant therapy",N. Machado,"Necrotizing Fasciitis (NF) is a necrotizing soft tissue infection involving the fascia and subcutaneous tissue that can cause rapid local tissue necrosis and life-threatening severe sepsis. Aim: This article aims to review the aetiopathogenesis, investigations and management based on a literature review. Methods and Materials: The Medline literature search of relevant articles restricted to English language on necrotizing fasciitis was conducted and reviewed. Results: Necrotizing fasciitis is rare with an incidence ranging from 0.15 to 0.55 cases per 100,000 of the population. Accurate assessment and timely intervention are critical in the treatment of patients affected with NF. Understanding the history and unique characteristics of this disease is crucial to achieve early recognition, effective treatment and a favorable outcome. Classic symptoms include severe pain out of proportion to local findings, erythema, mottling, crepitus, skin anesthesia, warmth, tenderness, hemorrhagic bullous formation, edema in the affected area and fever. Predisposing conditions of NF are classified into 2 main categories (type I and II) based on causative microorganisms. Radical surgical debridement, broad spectrum antibiotics, negative pressure wound dressings, and hyperbaric oxygen therapy are considered to be the cornerstone of treatment. The mortality rate ranges widely from 10% to 75% and is related to delay in initial debridement, patient age of more than 60 years, associated hypotension, acidosis, bacteremia, renal failure, hyponatremia, peripheral vascular disease, myonecrosis and myositis. Conclusion: Necrotizing fasciitis is a devastating infection of the fascia and subcutaneous tissue. The presentation of the disease is nonspecific and variable. Delay in recognition and effective treatment increases the mortality. Prompt radical surgical debridement, appropriate antibiotics and adjuvant therapy contribute to an improved outcome.","{'model': 'tldr@v2.0.0', 'text': 'The history and unique characteristics of this disease is crucial to achieve early recognition, effective treatment and a favorable outcome, and radical surgical debridement, broad spectrum antibiotics, negative pressure wound dressings, and hyperbaric oxygen therapy are considered to be the cornerstone of treatment.'}",
-epidemiology and clinical characteristics of community-acquired pneumonia in hospitalized children.,I. Michelow,"OBJECTIVES The precise epidemiology of childhood pneumonia remains poorly defined. Accurate and prompt etiologic diagnosis is limited by inadequate clinical, radiologic, and laboratory diagnostic methods. The objective of this study was to determine as precisely as possible the epidemiology and morbidity of community-acquired pneumonia in hospitalized children. METHODS Consecutive immunocompetent children hospitalized with radiographically confirmed lower respiratory infections (LRIs) were evaluated prospectively from January 1999 through March 2000. Positive blood or pleural fluid cultures or pneumolysin-based polymerase chain reaction assays, viral direct fluorescent antibody tests, or viral, mycoplasmal, or chlamydial serologic tests were considered indicative of infection by those organisms. Methods for diagnosis of pneumococcal pneumonia among study subjects were published by us previously. Selected clinical characteristics, indices of inflammation (white blood cell and differential counts and procalcitonin values), and clinical outcome measures (time to defervescence and duration of oxygen supplementation and hospitalization) were compared among groups of children. RESULTS One hundred fifty-four hospitalized children with LRIs were enrolled. Median age was 33 months (range: 2 months to 17 years). A pathogen was identified in 79% of children. Typical respiratory bacteria were identified in 60% (of which 73% were Streptococcus pneumoniae), viruses in 45%, Mycoplasma pneumoniae in 14%, Chlamydia pneumoniae in 9%, and mixed bacterial/viral infections in 23%. Preschool-aged children had as many episodes of atypical bacterial LRIs as older children. Children with typical bacterial or mixed bacterial/viral infections had the greatest inflammation and disease severity. Multivariate logistic-regression analyses revealed that high temperature (> or = 38.4 degrees C) within 72 hours after admission (odds ratio: 2.2; 95% confidence interval: 1.4-3.5) and the presence of pleural effusion (odds ratio: 6.6; 95% confidence interval: 2.1-21.2) were significantly associated with bacterial pneumonia. CONCLUSIONS This study used an expanded diagnostic armamentarium to define the broad spectrum of pathogens that cause pneumonia in hospitalized children. The data confirm the importance of S pneumoniae and the frequent occurrence of bacterial and viral coinfections in children with pneumonia. These findings will facilitate age-appropriate antibiotic selection and future evaluation of the clinical effectiveness of the pneumococcal conjugate vaccine as well as other candidate vaccines.","{'model': 'tldr@v2.0.0', 'text': 'The data confirm the importance of S pneumoniae and the frequent occurrence of bacterial and viral coinfections in children with pneumonia and will facilitate age-appropriate antibiotic selection and future evaluation of the clinical effectiveness of the pneumococcal conjugate vaccine as well as other candidate vaccines.'}",
-artificial intelligence and acute stroke imaging,J. Soun,"SUMMARY: Artificial intelligence technology is a rapidly expanding field with many applications in acute stroke imaging, including ischemic and hemorrhage subtypes. Early identification of acute stroke is critical for initiating prompt intervention to reduce morbidity and mortality. Artificial intelligence can help with various aspects of the stroke treatment paradigm, including infarct or hemorrhage detection, segmentation, classification, large vessel occlusion detection, Alberta Stroke Program Early CT Score grading, and prognostication. In particular, emerging artificial intelligence techniques such as convolutional neural networks show promise in performing these imaging-based tasks efficiently and accurately. The purpose of this review is twofold: first, to describe AI methods and available public and commercial platforms in stroke imaging, and second, to summarize the literature of current artificial intelligence–driven applications for acute stroke triage, surveillance, and prediction.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this review is to describe AI methods and available public and commercial platforms in stroke imaging, and to summarize the literature of current artificial intelligence–driven applications for acute stroke triage, surveillance, and prediction.'}",http://www.ajnr.org/content/ajnr/42/1/2.full.pdf
-integrating rapid pathogen identification and antimicrobial stewardship significantly decreases hospital costs.,K. Perez,"CONTEXT Early diagnosis of gram-negative bloodstream infections, prompt identification of the infecting organism, and appropriate antibiotic therapy improve patient care outcomes and decrease health care expenditures. In an era of increasing antimicrobial resistance, methods to acquire and rapidly translate critical results into timely therapies for gram-negative bloodstream infections are needed. OBJECTIVE To determine whether mass spectrometry technology coupled with antimicrobial stewardship provides a substantially improved alternative to conventional laboratory methods. DESIGN An evidence-based intervention that integrated matrix-assisted laser desorption and ionization time-of-flight mass spectrometry, rapid antimicrobial susceptibility testing, and near-real-time antimicrobial stewardship practices was implemented. Outcomes in patients hospitalized prior to initiation of the study intervention were compared to those in patients treated after implementation. Differences in length of hospitalization and hospital costs were assessed in survivors. RESULTS The mean hospital length of stay in the preintervention group survivors (n = 100) was 11.9 versus 9.3 days in the intervention group (n = 101; P = .01). After multivariate analysis, factors independently associated with decreased length of hospitalization included the intervention (hazard ratio, 1.38; 95% confidence interval, 1.01-1.88) and active therapy at 48 hours (hazard ratio, 2.9; confidence interval, 1.15-7.33). Mean hospital costs per patient were $45 709 in the preintervention group and $26 162 in the intervention group (P = .009). CONCLUSIONS Integration of rapid identification and susceptibility techniques with antimicrobial stewardship significantly improved time to optimal therapy, and it decreased hospital length of stay and total costs. This innovative strategy has ramifications for other areas of patient care.","{'model': 'tldr@v2.0.0', 'text': None}",
-"a web-based, computer-tailored smoking prevention program to prevent children from starting to smoke after transferring to secondary school: randomized controlled trial",H. Cremers,"Background Smoking prevalence rates among Dutch children increase rapidly after they transit to secondary school, in particular among children with a low socioeconomic status (SES). Web-based, computer-tailored programs supplemented with prompt messages may be able to empower children to prevent them from starting to smoke when they transit to secondary school. Objective The main aim of this study is to evaluate whether computer-tailored feedback messages, with and without prompt messages, are effective in decreasing children’s smoking intentions and smoking behavior after 12 and 25 months of follow-up. Methods Data were gathered at baseline (T0), and after 12 months (T1) and 25 months (T2) of follow-up of a smoking prevention intervention program called Fun without Smokes. A total of 162 schools were randomly allocated to a no-intervention control group, an intervention prompt group, or an intervention no-prompt group. A total of 3213 children aged 10 to 12 years old participated in the study and completed a Web-based questionnaire assessing their smoking intention, smoking behavior, and sociocognitive factors, such as attitude, social influence, and self-efficacy, related to smoking. After completion, children in the intervention groups received computer-tailored feedback messages in their own email inbox and those messages could be accessed on the intervention website. Children in the prompt group received prompt messages, via email and short message service (SMS) text messaging, to stimulate them to reuse the intervention website with nonsmoking content. Multilevel logistic regression analyses were performed using multiple imputations to assess the program effects on smoking intention and smoking behavior at T1 and T2. Results A total of 3213 children participated in the Fun without Smokes study at T0. Between T0 and T1 a total of 1067 children out of the original 3213 (33.21%) dropped out of the study. Between T0 and T2 the number of children that did not participate in the final measurement was 1730 out of the original 3213 (53.84%). No significant program effects were observed for any of the intervention groups compared to the control group at T1 for the intention to engage in smoking—prompt, OR 0.67 (95% CI 0.30-1.50), no-prompt, OR 0.76 (95% CI 0.34-1.67)—or for smoking behavior—prompt, OR 1.13 (95% CI 0.13-9.98), no-prompt, OR 0.50 (95% CI 0.04-5.59). Similar nonsignificant program effects were found at T2 for the intention to start smoking—prompt, OR 0.78 (95% CI 0.26-2.32), no-prompt, OR 1.31 (95% CI 0.45-3.82)—and smoking behavior—prompt, OR 0.53 (95% CI 0.12-2.47), no-prompt, OR 1.01 (95% CI 0.24-4.21). Conclusions This study showed that the Web-based, computer-tailored feedback messages with and without prompt messages were not effective in modifying children’s smoking intentions and smoking behavior as compared to no information. Future smoking prevention interventions are recommended to start closer to the age of actual smoking uptake. Furthermore, future studies on Web-based, computer-tailored smoking prevention programs should focus on assessing and controlling exposure to the educational content and the response to the prompt messages. Trial Registration Netherlands Trial Register NTR3116; http://www.trialregister.nl/trialreg/admin/rctview.asp?TC=3116 (Archived by WebCite at http://www.webcitation.org/6O0wQYuPI).","{'model': 'tldr@v2.0.0', 'text': 'This study showed that the Web-based, computer-tailored feedback messages with and without prompt messages were not effective in modifying children’s smoking intentions and smoking behavior as compared to no information.'}",https://www.jmir.org/2015/3/e59/PDF
-detection and mitigation of spoofing attacks on a vector-based tracking gps receiver,A. Jafarnia-Jahromi,"Position solutions provided by GNSS receivers can be completely misled by spoofing attacks. Consequently detection and mitigation of spoofing signals have become one of the important topics within the GNSS community. Some recently published work has addressed spoofing attack on tracking receivers. The work has mostly focused on signal quality monitoring (SQM) techniques to detect spoofing and authentic signals interaction. SQM methods can effectively detect the spoofing correlation peak that is approaching the authentic signal. However, they are not applicable in cases where spoofing attack does not affect the shape of the correlation peak. This situation happens when a spoofer and authentic signals are almost aligned together. This paper provides an analytical approach to investigate the interaction between the authentic and the spoofing correlation peaks during spoofing attacks. Then, a spoofing detection technique based on amplitude analysis of different (namely very early (VE), early (E), prompt (P), late (L) and very late (VL)) correlator branches is proposed. The proposed spoofing detection technique continuously checks the distribution of each correlator output. Spoofing attack is detected if this distribution considerably deviates from that of the authentic signal. A vector based tracking receiver structure has been also employed to bridge the authentic signal outage during the spoofing attack. Data collection has been performed using the simulated spoofing attack by Spirent hardware simulator. Data processing results verify that the proposed spoofing detection and mitigation based on vector based (VB) receiver structure techniques perform well in real world scenarios.","{'model': 'tldr@v2.0.0', 'text': 'An analytical approach to investigate the interaction between the authentic and the spoofing correlation peaks during spoofing attacks is provided and a spoofing detection technique based on amplitude analysis of different correlator branches is proposed.'}",
-brain tumor detection using deep neural network and machine learning algorithm,Masoumeh Siar,"Brain tumor can be classified into two types: benign and malignant. Timely and prompt disease detection and treatment plan leads to improved quality of life and increased life expectancy in these patients. One of the most practical and important methods is to use Deep Neural Network (DNN). In this paper, a Convolutional Neural Network (CNN) has been used to detect a tumor through brain Magnetic Resonance Imaging (MRI) images. Images were first applied to the CNN. The accuracy of Softmax Fully Connected layer used to classify images obtained 98.67%. Also, the accuracy of the CNN is obtained with the Radial Basis Function (RBF) classifier 97.34% and the Decision Tree (DT) classifier, is 94.24%. In addition to the accuracy criterion, we use the benchmarks of Sensitivity, Specificity and Precision evaluate network performance. According to the results obtained from the categorizers, the Softmax classifier has the best accuracy in the CNN according to the results obtained from network accuracy on the image testing. This is a new method based on the combination of feature extraction techniques with the CNN for tumor detection from brain images. The method proposed accuracy 99.12% on the test data. Due to the importance of the diagnosis given by the physician, the accuracy of the doctors help in diagnosing the tumor and treating the patient increased.","{'model': 'tldr@v2.0.0', 'text': 'This is a new method based on the combination of feature extraction techniques with the CNN for tumor detection from brain images and the Softmax classifier has the best accuracy in the CNN according to the results obtained from network accuracy on the image testing.'}",
-clinical and diagnostic utility of saliva as a non-invasive diagnostic fluid:a systematic review,L. A. S. Nunes,"This systematic review presents the latest trends in salivary research and its applications in health and disease. Among the large number of analytes present in saliva, many are affected by diverse physiological and pathological conditions. Further, the non-invasive, easy and cost-effective collection methods prompt an interest in evaluating its diagnostic or prognostic utility. Accumulating data over the past two decades indicates towards the possible utility of saliva to monitor overall health, diagnose and treat various oral or systemic disorders and drug monitoring. Advances in saliva based systems biology has also contributed towards identification of several biomarkers, development of diverse salivary diagnostic kits and other sensitive analytical techniques. However, its utilization should be carefully evaluated in relation to standardization of pre-analytical and analytical variables, such as collection and storage methods, analyte circadian variation, sample recovery, prevention of sample contamination and analytical procedures. In spite of all these challenges, there is an escalating evolution of knowledge with the use of this biological matrix.","{'model': 'tldr@v2.0.0', 'text': 'Salivary utilization should be carefully evaluated in relation to standardization of pre-analytical and analytical variables, such as collection and storage methods, analyte circadian variation, sample recovery, prevention of sample contamination and analytical procedures.'}",https://europepmc.org/articles/pmc4470107?pdf=render
-a comparison of four prompt modes for route finding for community travellers with severe cognitive impairments,M. Sohlberg,"Primary objective: Navigational skills are fundamental to community travel and, hence, personal independence and are often disrupted in people with cognitive impairments. Navigation devices are being developed that can support community navigation by delivering directional information. Selecting an effective mode to provide route-prompts is a critical design issue. This study evaluated the differential effects on pedestrian route finding using different modes of prompting delivered via a handheld electronic device for travellers with severe cognitive impairments. Research design: A within-subject comparison study was used to evaluate potential differences in route navigation performance when travellers received directions using four different prompt modes: (1) aerial map image, (2) point of view map image, (3) text based instructions/no image and (4) audio direction/no image. Methods and procedures: Twenty travellers with severe cognitive impairments due to acquired brain injury walked four equivalent routes using four different prompting modes delivered via a wrist-worn navigation device. Navigation scores were computed that captured accuracy and confidence during navigation. Main outcome: Results of the repeated measures Analysis of Variance suggested that participants performed best when given prompts via speech-based audio directions. The majority of the participants also preferred this prompting mode. Findings are interpreted in the context of cognitive resource allocation theory.","{'model': 'tldr@v2.0.0', 'text': 'Results of the repeated measures Analysis of Variance suggested that participants performed best when given prompts via speech-based audio directions, and the majority of the participants also preferred this prompting mode.'}",
-mastitis treatment-reduction in antibiotic usage in dairy cows.,V. Krömker,"Animal-friendly, economical, resource-saving milk production provides the basis for sustained consumer acceptance. Bovine mastitis plays a decisive role in the dairy industry-disturbing animal health and welfare and causing considerable economic losses on the other hand. Currently, antimicrobial treatment is indispensable to keep bovine udder health, animal welfare and economic aspects in balance. On the contrary, emergence and spread of antimicrobial resistance (AMR) is an urgent matter of particular public interest, and as a consequence, antimicrobial usage (AMU) in production livestock is a critically discussed subject. In urgent need of future reduction in AMU in the dairy industry, this review article describes and discusses possible approaches promising prompt implementation, including therapeutical alternatives as well as pro- and metaphylactic concepts such as the implementation of evidence-based mastitis therapy concepts and selective dry cow treatment (sDCT), in search of the most effective and contemporary methods for decreasing AMU and AMR in dairy production.","{'model': 'tldr@v2.0.0', 'text': 'In urgent need of future reduction in AMU in the dairy industry, this review article describes and discusses possible approaches promising prompt implementation, including therapeutical alternatives as well as pro- and metaphylactic concepts such as the implementation of evidence-based mastitis therapy concepts and selective dry cow treatment.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/rda.13032
-using theories of behaviour change to inform interventions for addictive behaviours.,Thomas Llewelyn Webb,"AIMS This paper reviews a set of theories of behaviour change that are used outside the field of addiction and considers their relevance for this field. METHODS Ten theories are reviewed in terms of (i) the main tenets of each theory, (ii) the implications of the theory for promoting change in addictive behaviours and (iii) studies in the field of addiction that have used the theory. An augmented feedback loop model based on Control Theory is used to organize the theories and to show how different interventions might achieve behaviour change. RESULTS Briefly, each theory provided the following recommendations for intervention: Control Theory: prompt behavioural monitoring, Goal-Setting Theory: set specific and challenging goals, Model of Action Phases: form 'implementation intentions', Strength Model of Self-Control: bolster self-control resources, Social Cognition Models (Protection Motivation Theory, Theory of Planned Behaviour, Health Belief Model): modify relevant cognitions, Elaboration Likelihood Model: consider targets' motivation and ability to process information, Prototype Willingness Model: change perceptions of the prototypical person who engages in behaviour and Social Cognitive Theory: modify self-efficacy. CONCLUSIONS There are a range of theories in the field of behaviour change that can be applied usefully to addiction, each one pointing to a different set of modifiable determinants and/or behaviour change techniques. Studies reporting interventions should describe theoretical basis, behaviour change techniques and mode of delivery accurately so that effective interventions can be understood and replicated.","{'model': 'tldr@v2.0.0', 'text': 'There are a range of theories in the field of behaviour change that can be applied usefully to addiction, each one pointing to a different set of modifiable determinants and/or behaviour change techniques.'}",
-evidence-based drainage of infected hydronephrosis secondary to ureteric calculi.,S. Ramsey,"INTRODUCTION The obstructed, infected kidney is a urological emergency. It has been accepted that the management of infected hydronephrosis secondary to ureteric stones is through prompt decompression of the collecting system. However, the optimal method of decompression has yet to be established. MATERIALS AND METHODS A PubMed and Medline search was performed of all English-language articles from 1960 using key words ""sepsis,"" ""urosepsis,"" ""obstruction,"" ""obstructive pyelonephritis,"" ""pyonephrosis,"" ""infection and hydronephrosis"" ""decompression,"" ""stent,"" ""nephrostomy,"" and ""management."" The Cochrane database and National Institute for Clinical Effectiveness guidelines were searched using the terms ""sepsis,"" ""urosepsis,"" ""stent,"" ""nephrostomy,"" or ""obstruction."" Scottish intercollegiate guidelines were reviewed and no relevant guidance was identified. RESULTS Two randomised trials have compared retrograde stent insertion with percutaneous nephrostomy with one trial reporting specifically on patients with acute sepsis and obstruction. Neither trial showed one superior modality of decompression in effecting decompression and resolution of sepsis. A further literature search regarding the complications of percutaneous nephrostomy and stent insertion was carried out. An overall major complication rate from percutaneous nephrostomy insertion was found to be 4%, although the complication rates from stent insertion are less consistently reported. DISCUSSION There appears little evidence to suggest that retrograde stent insertion leads to increased bacteraemia or is significantly more hazardous in the setting of acute obstruction. Further region-wide discussion between urologists and interventional radiologists is required to establish management protocols for these acutely unwell patients.","{'model': 'tldr@v2.0.0', 'text': 'There appears little evidence to suggest that retrograde stent insertion leads to increased bacteraemia or is significantly more hazardous in the setting of acute obstruction.'}",
-multi-stage pre-training for automated chinese essay scoring,Wei Song,"This paper proposes a pre-training based automated Chinese essay scoring method. The method involves three components: weakly supervised pre-training, supervised cross- prompt fine-tuning and supervised target- prompt fine-tuning. An essay scorer is first pre- trained on a large essay dataset covering diverse topics and with coarse ratings, i.e., good and poor, which are used as a kind of weak supervision. The pre-trained essay scorer would be further fine-tuned on previously rated es- says from existing prompts, which have the same score range with the target prompt and provide extra supervision. At last, the scorer is fine-tuned on the target-prompt training data. The evaluation on four prompts shows that this method can improve a state-of-the-art neural essay scorer in terms of effectiveness and domain adaptation ability, while in-depth analysis also reveals its limitations..","{'model': 'tldr@v2.0.0', 'text': 'The evaluation on four prompts shows that this method can improve a state-of-the-art neural essay scorer in terms of effectiveness and domain adaptation ability, while in-depth analysis also reveals its limitations.'}",https://www.aclweb.org/anthology/2020.emnlp-main.546.pdf
-carotid blowout syndrome: modern trends in management,C. Suárez,"Carotid blowout syndrome (CBS) refers to rupture of the carotid artery and is an uncommon complication of head and neck cancer that can be rapidly fatal without prompt diagnosis and intervention. CBS develops when a damaged arterial wall cannot sustain its integrity against the patient’s blood pressure, mainly in patients who have undergone surgical procedures and radiotherapy due to cancer of the head and neck, or have been reirradiated for a recurrent or second primary tumor in the neck. Among patients irradiated prior to surgery, CBS is usually a result of wound breakdown, pharyngocutaneous fistula and infection. This complication has often been fatal in the past, but at the present time, early diagnosis and modern technology applied to its management have decreased morbidity and mortality rates. In addition to analysis of the causes and consequences of CBS, the purpose of this paper is to critically review methods for early diagnosis of this complication and establish individualized treatment based on endovascular procedures for each patient.","{'model': 'tldr@v2.0.0', 'text': 'Methods for early diagnosis of carotid blowout syndrome are critically reviewed and individualized treatment based on endovascular procedures for each patient is established.'}",https://www.dovepress.com/getfile.php?fileID=46143
-helicobacter pylori test and eradicate versus prompt endoscopy for management of dyspeptic patients: 6.7 year follow up of a randomised trial,A. Lassen,"Background: Dyspepsia is a chronic disease with significant impact on the use of health care resources. A management strategy based on Helicobacter pylori testing has been recommended but the long term effect is unknown. Aim: To investigate the long term effect of a test and treat strategy compared with prompt endoscopy for management of dyspeptic patients in primary care. Patients: A total of 500 patients presenting in primary care with dyspepsia were randomised to management by H pylori testing plus eradication therapy (n = 250) or by endoscopy (n = 250). Results of 12 month follow up have previously been presented. Methods: Symptoms, quality of life, and patient satisfaction were recorded during a three month period, a median 6.7 years after randomisation (range 6.1–7.3 years). Number of endoscopies, antisecretory medication, H pylori treatments, and hospital visits were recorded from health care databases for the entire follow up period. Results: Median age was 45 years; 28% were H pylori infected. Use of resources was registered in all 500 patients (3084 person years) of whom 312 completed diaries. We found no difference in symptoms between the two groups. Median proportion of days without symptoms was 0.52 (interquartile range 0.10–0.88) in the test and eradicate group versus 0.64 (0.14–0.90) in the prompt endoscopy group (p = 0.27) (mean difference 0.05 (95% confidence interval (CI) −0.03 to 0.14)). Compared with the prompt endoscopy group, the test and eradicate group underwent fewer endoscopies (mean difference 0.62 endoscopies/person (95% CI 0.38–0.86)) and used less antisecretory medication (mean difference 102 defined daily doses/person (95% CI −1 to 205)). Conclusion: On a long term basis, a H pylori test and eradicate strategy is as efficient as prompt endoscopy for management of dyspeptic patients in primary care and reduces the use of endoscopy and antisecretory medication.","{'model': 'tldr@v2.0.0', 'text': 'On a long term basis, a H pylori test and eradicate strategy is as efficient as prompt endoscopy for management of dyspeptic patients in primary care and reduces the use of endoscopies and antisecretory medication.'}",https://gut.bmj.com/content/gutjnl/53/12/1758.full.pdf
-"microbiology of chronic suppurative otitis media in a tertiary care setup of uttarakhand state, india",R. Prakash,"Background: Chronic suppurative otitis media (CSOM) is a notorious infection and a major health problem in developing countries causing serious local damage and threatening complications. Early and effective treatment based on the knowledge of causative micro-organisms and their antimicrobial sensitivity ensures prompt clinical recovery and possible complications can thus be avoided. Aims: The aim of this study was to isolate the organisms associated with CSOM and to detect the antibiogram of the aerobic isolates. Materials and Methods: A total of 204 patients clinically diagnosed of CSOM were enrolled in the study and the samples were obtained from each patient using sterile cotton swabs and cultured for microbial flora. Drug susceptibility testing for aerobic isolates was conducted using Kirby-Bauer disc diffusion method. Results: The most common causative organisms isolated were Staphylococcus aureus (48.69%) and Pseudomonas aeruginosa (19.89%) amongst the 191 aerobic isolates. Anaerobes accounted for 29.41% of the isolates while 12.25% were fungi. Antimicrobial profile of aerobic isolates revealed maximum sensitivity to amikacin (95.5%), ceftriaxone (83.4%) and gentamicin (82.7%). Conclusion: Knowing the etiological agents of CSOM and their antimicrobial susceptibility is of essential importance for an efficient treatment, prevention of both complications and development of antibiotic resistance and finally, the reduction of the treatment costs.","{'model': 'tldr@v2.0.0', 'text': 'Knowing the etiological agents of CSOM and their antimicrobial susceptibility is of essential importance for an efficient treatment, prevention of both complications and development of antibiotic resistance and finally, the reduction of the treatment costs.'}",
-why is mucormycosis more difficult to cure than more common mycoses?,A. Katragkou,"Although considered to be a rare infection, mucormycosis (zygomycosis) has emerged as the second most common invasive mould infection. Despite the advent of newer antifungal agents, mortality rate of mucormycosis remains exceedingly high. Successful management of mucormycosis requires early diagnosis, reversal of underlying predisposing risk factors, surgical debridement and prompt administration of active antifungal agents. However, mucormycosis is not always amenable to cure. There are challenging obstacles that lead to difficulties in management of amphotericin B. These include unique host-based risk factors for mucormycosis, the fungus' resistance to innate host defences and distinctive features of its immunopathogenesis, such as extensive angioinvasion, increased virulence and use of chelators by the fungus as siderophores. In addition to these obstacles, the difficulties in early diagnosis, including nonspecific clinical manifestations, lack of serological methods, as well limitations of culture and molecular methods, lead to delay in initiation of antifungal therapy. Finally, the variability of susceptibility to amphotericin B and resistance to most other conventional antifungal agents leads to major limitations in successful treatment of this devastating infection.","{'model': 'tldr@v2.0.0', 'text': 'The variability of susceptibility to amphotericin B and resistance to most other conventional antifungal agents leads to major limitations in successful treatment of this devastating infection.'}",http://www.clinicalmicrobiologyandinfection.com/article/S1198743X14603360/pdf
-real and accrual‐based earnings management and its legal consequences,Salma Ibrahim,"Purpose – Prior research suggests that firms manipulate earnings through accruals to achieve certain reporting objectives. Recently, especially following the Sarbanes‐Oxley (SarbOx) Act, researchers have turned their attention to real account manipulation as an alternative. However, there is no evidence on whether the likelihood of being detected by outsiders is different for firms using these alternative manipulation methods. The purpose of this paper is to examine this research question in the context of seasoned equity offerings (SEOs).Design/methodology/approach – First, the authors compare SEOs to a matched sample of non‐SEOs to document income‐increasing manipulation. Next, they identify SEOs that prompt lawsuits and compare sued and non‐sued firms to determine whether using a particular method of manipulation is more likely to be detected and associated with litigation.Findings – The authors find evidence of income‐increasing accrual and real manipulation for SEOs in the year prior to the offering ...",,
-a smartphone app (afyadata) for innovative one health disease surveillance from community to national levels in africa: intervention in disease surveillance,E. Karimuribo,"Background We describe the development and initial achievements of a participatory disease surveillance system that relies on mobile technology to promote Community Level One Health Security (CLOHS) in Africa. Objective The objective of this system, Enhancing Community-Based Disease Outbreak Detection and Response in East and Southern Africa (DODRES), is to empower community-based human and animal health reporters with training and information and communication technology (ICT)–based solutions to contribute to disease detection and response, thereby complementing strategies to improve the efficiency of infectious disease surveillance at national, regional, and global levels. In this study, we refer to techno-health as the application of ICT-based solutions to enhance early detection, timely reporting, and prompt response to health events in human and animal populations. Methods An EpiHack, involving human and animal health experts as well as ICT programmers, was held in Tanzania in 2014 to identify major challenges facing early detection, timely reporting, and prompt response to disease events. This was followed by a project inception workshop in 2015, which brought together key stakeholders, including policy makers and community representatives, to refine the objectives and implementation plan of the DODRES project. The digital ICT tools were developed and packaged together as the AfyaData app to support One Health disease surveillance. Community health reporters (CHRs) and officials from animal and human health sectors in Morogoro and Ngorongoro districts in Tanzania were trained to use the AfyaData app. The AfyaData supports near- to real-time data collection and submission at both community and health facility levels as well as the provision of feedback to reporters. The functionality of the One Health Knowledge Repository (OHKR) app has been integrated into the AfyaData app to provide health information on case definitions of diseases of humans and animals and to synthesize advice that can be transmitted to CHRs with next step response activities or interventions. Additionally, a WhatsApp social group was made to serve as a platform to sustain interactions between community members, local government officials, and DODRES team members. Results Within the first 5 months (August-December 2016) of AfyaData tool deployment, a total of 1915 clinical cases in livestock (1816) and humans (99) were reported in Morogoro (83) and Ngorongoro (1832) districts. Conclusions These initial results suggest that the DODRES community-level model creates an opportunity for One Health engagement of people in their own communities in the detection of infectious human and animal disease threats. Participatory approaches supported by digital and mobile technologies should be promoted for early disease detection, timely reporting, and prompt response at the community, national, regional, and global levels.","{'model': 'tldr@v2.0.0', 'text': 'These initial results suggest that the DODRES community-level model creates an opportunity for One Health engagement of people in their own communities in the detection of infectious human and animal disease threats.'}",
-significantly improved analytical sensitivity of lateral flow immunoassays by using thermal contrast.,Zhenpeng Qin,"The ability to rapidly identify diseases enables prompt treatment and improves outcomes. This has increased the development and use of rapid point-of-care diagnostic devices capable of biomolecular detection in both high-income and resource-limited settings.[1] Lateral flow assays (LFAs) are inexpensive, simple, portable, and robust,[2] making LFAs commonplace in medicine, agriculture, and over-the-counter personal use such as for pregnancy testing. Although the analytical performance of some LFAs are comparable to laboratory based methods,[1a] the sensitivity of most LFAs is in the mM to μM range,[2–3] which is many folds less sensitive than other molecular techniques such as enzyme-linked immunoassays (ELISA). As a consequence, LFAs are not particularly useful for detection early in a disease course when there is low level of antigen. Due to the increasing need for highly sensitive molecular diagnostics, researchers have focused on developing microfluidics,[1a, 1b] biobar codes,[1c, 1d] and enzyme-based immunoassay technologies[4] technologies to fulfill the need since these technologies have nM to pM detection sensitivity for protein analysis and can potentially be miniaturized as handheld point-of-care diagnostic devices.[1c] These emerging technologies are still early in development and are not yet field-ready.","{'model': 'tldr@v2.0.0', 'text': 'Due to the increasing need for highly sensitive molecular diagnostics, researchers have focused on developing microfluidics, biobar codes, and enzyme-based immunoassay technologies to fulfill the need since these technologies have nM to pM detection sensitivity for protein analysis and can potentially be miniaturized as handheld point-of-care diagnostic devices.'}",https://europepmc.org/articles/pmc3337364?pdf=render
-recent advances in the microbiological diagnosis of bloodstream infections,W. Florio,"Abstract Rapid identification (ID) and antimicrobial susceptibility testing (AST) of the causative agent(s) of bloodstream infections (BSIs) are essential for the prompt administration of an effective antimicrobial therapy, which can result in clinical and financial benefits. Immediately after blood sampling, empirical antimicrobial therapy, chosen on clinical and epidemiological data, is administered. When ID and AST results are available, the clinician decides whether to continue or streamline the antimicrobial therapy, based on the results of the in vitro antimicrobial susceptibility profile of the pathogen. The aim of the present study is to review and discuss the experimental data, advantages, and drawbacks of recently developed technological advances of culture-based and molecular methods for the diagnosis of BSI (including mass spectrometry, magnetic resonance, PCR-based methods, direct inoculation methods, and peptide nucleic acid fluorescence in situ hybridization), the understanding of which could provide new perspectives to improve and fasten the diagnosis and treatment of septic patients. Although blood culture remains the gold standard to diagnose BSIs, newly developed methods can significantly shorten the turnaround time of reliable microbial ID and AST, thus substantially improving the diagnostic yield.","{'model': 'tldr@v2.0.0', 'text': 'Although blood culture remains the gold standard to diagnose BSIs, newly developed methods can significantly shorten the turnaround time of reliable microbial ID and AST, thus substantially improving the diagnostic yield.'}",
-comparative archaeometrical study of roman silver coins by prompt gamma activation analysis and sem-edx,Z. Kasztovszky,,,
-detecting consciousness: a unique role for neuroimaging.,A. Owen,"How can we ever know, unequivocally, that another person is conscious and aware? Putting aside deeper philosophical considerations about the nature of consciousness itself, historically, the only reliable method for detecting awareness in others has been through a predicted behavioral response to an external prompt or command. The answer may take the form of spoken words or a nonverbal signal such as a hand movement or the blink of an eye, but it is this answer, and only this answer, that allows us to infer awareness. In recent years, rapid technological developments in the field of neuroimaging have provided new methods for revealing thoughts, actions, and intentions based solely on the pattern of activity that is observed in the brain. In specialized centers, these methods are now being employed routinely to detect consciousness in behaviorally nonresponsive patients when all existing clinical techniques have failed to provide that information. In this review, I compare those circumstances in which neuroimaging data can be used to infer consciousness in the absence of a behavioral response with those circumstances in which it cannot. This distinction is fundamental for understanding and interpreting patterns of brain activity following acute brain injury and has profound implications for clinical care, diagnosis, prognosis, and medical-legal decision-making (relating to the prolongation, or otherwise, of life after severe brain injury). It also sheds light on more basic scientific questions about the nature of consciousness and the neural representation of our own thoughts and intentions.","{'model': 'tldr@v2.0.0', 'text': 'This review compares those circumstances in which neuroimaging data can be used to infer consciousness in the absence of a behavioral response with those circumstancesIn which it cannot and sheds light on more basic scientific questions about the nature of consciousness.'}",
-specific effects of a calorie-based intervention on stair climbing in overweight commuters,A. Lewis,,"{'model': 'tldr@v2.0.0', 'text': 'The specificity of the effect confirms the appeal of this lifestyle activity for the overweight, and focuses on how intentions to control weight may be converted into behaviour.'}",https://academic.oup.com/abm/article-pdf/42/2/257/22059945/12160_2011_article_9283.pdf
-fission modelling with fifrelin,O. Litaize,,,https://hal-cea.archives-ouvertes.fr/cea-02428818/file/Litaize2015_Article_FissionModellingWithFIFRELIN.pdf
-kinetics of ch + n2 revisited with multireference methods.,L. B. Harding,"The potential energy surface for the CH + N2 reaction was reexamined with multireference ab initio electronic structure methods employing basis sets up to aug-cc-pvqz. Comparisons with related CCSD(T) calculations were also made. The multireference ab initio calculations indicate significant shortcomings in single reference based methods for two key rate-limiting transition states. Transition state theory calculations incorporating the revised best estimates for the transition state properties provide order of magnitude changes in the predicted rate coefficient in the temperature range of importance to the mechanism for prompt NO formation. At higher temperatures, two distinct pathways make a significant contribution to the kinetics. A key part of the transition state analysis involves a variable reaction coordinate transition state theory treatment for the formation of H + NCN from HNCN. The present predictions for the rate coefficients resolve the discrepancy between prior theory and very recent experimental measurements.","{'model': 'tldr@v2.0.0', 'text': 'Transition state theory calculations incorporating the revised best estimates for the transition state properties provide order of magnitude changes in the predicted rate coefficient in the temperature range of importance to the mechanism for prompt NO formation.'}",
-automated dna sequence-based early warning system for the detection of methicillin-resistant staphylococcus aureus outbreaks,A. Mellmann,"Background The detection of methicillin-resistant Staphylococcus aureus (MRSA) usually requires the implementation of often rigorous infection-control measures. Prompt identification of an MRSA epidemic is crucial for the control of an outbreak. In this study we evaluated various early warning algorithms for the detection of an MRSA cluster. Methods and Findings Between 1998 and 2003, 557 non-replicate MRSA strains were collected from staff and patients admitted to a German tertiary-care university hospital. The repeat region of the S. aureus protein A (spa) gene in each of these strains was sequenced. Using epidemiological and typing information for the period 1998–2002 as reference data, clusters in 2003 were determined by temporal-scan test statistics. Various early warning algorithms (frequency, clonal, and infection control professionals [ICP] alerts) were tested in a prospective analysis for the year 2003. In addition, a newly implemented automated clonal alert system of the Ridom StaphType software was evaluated. A total of 549 of 557 MRSA were typeable using spa sequencing. When analyzed using scan test statistics, 42 out of 175 MRSA in 2003 formed 13 significant clusters (p < 0.05). These clusters were used as the “gold standard” to evaluate the various algorithms. Clonal alerts (spa typing and epidemiological data) were 100% sensitive and 95.2% specific. Frequency (epidemiological data only) and ICP alerts were 100% and 62.1% sensitive and 47.2% and 97.3% specific, respectively. The difference in specificity between clonal and ICP alerts was not significant. Both methods exhibited a positive predictive value above 80%. Conclusions Rapid MRSA outbreak detection, based on epidemiological and spa typing data, is a suitable alternative for classical approaches and can assist in the identification of potential sources of infection.","{'model': 'tldr@v2.0.0', 'text': 'Rapid MRSA outbreak detection, based on epidemiological and spa typing data, is a suitable alternative for classical approaches and can assist in the identification of potential sources of infection.'}",https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.0030033&type=printable
-general purpose computer-assisted clustering and conceptualization,Justin Grimmer,"We develop a computer-assisted method for the discovery of insightful conceptualizations, in the form of clusterings (i.e., partitions) of input objects. Each of the numerous fully automated methods of cluster analysis proposed in statistics, computer science, and biology optimize a different objective function. Almost all are well defined, but how to determine before the fact which one, if any, will partition a given set of objects in an “insightful” or “useful” way for a given user is unknown and difficult, if not logically impossible. We develop a metric space of partitions from all existing cluster analysis methods applied to a given dataset (along with millions of other solutions we add based on combinations of existing clusterings) and enable a user to explore and interact with it and quickly reveal or prompt useful or insightful conceptualizations. In addition, although it is uncommon to do so in unsupervised learning problems, we offer and implement evaluation designs that make our computer-assisted approach vulnerable to being proven suboptimal in specific data types. We demonstrate that our approach facilitates more efficient and insightful discovery of useful information than expert human coders or many existing fully automated methods.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a metric space of partitions from all existing cluster analysis methods applied to a given dataset and demonstrates that this approach facilitates more efficient and insightful discovery of useful information than expert human coders or many existing fully automated methods.'}",https://www.pnas.org/content/pnas/108/7/2643.full.pdf
-parents' educational involvement: a developmental ecology perspective,R. Seginer,"The objective of this review is to examine research on home-based and school-based parental involvement and generate new research questions by employing Bronfenbrenner's ecological framework consisting of the micro-, meso-, exo-, and macrosystems. This analysis shows that, although both family-based and school-based parental involvement are positively related to educational outcomes, their examination in the ecological framework prompts consideration of additional aspects of the micro- and mesosystems and their embeddedness in four exosystemic aspects (parents' networks and workplace, neighborhood, and educational policy) and two macrosystemic types (immigrant and ethnic groups). Guided by Bronfenbrenner's ecological thinking and the availability of advanced multivariate analysis methods, the next stage of this research should test multiple-step models describing factors that prompt parental involvement and mediate and moderate the parental involvement - educational outcomes links in different sociocultural settings.",,
-social representation of cyberbullying and adolescent suicide: a mixed-method analysis of news stories,R. Young,"ABSTRACT Cyberbullying has provoked public concern after well-publicized suicides of adolescents. This mixed-methods study investigates the social representation of these suicides. A content analysis of 184 U.S. newspaper articles on death by suicide associated with cyberbullying or aggression found that few articles adhered to guidelines suggested by the World Health Organization and the American Foundation for Suicide Prevention to protect against suicidal behavioral contagion. Few articles made reference to suicide or bullying prevention resources, and most suggested that the suicide had a single cause. Thematic analysis of a subset of articles found that individual deaths by suicide were used as cautionary tales to prompt attention to cyberbullying. This research suggests that newspaper coverage of these events veers from evidence-based guidelines and that more work is needed to determine how best to engage with journalists about the potential consequences of cyberbullying and suicide coverage.","{'model': 'tldr@v2.0.0', 'text': 'The research suggests that newspaper coverage of these events veers from evidence-based guidelines and that more work is needed to determine how best to engage with journalists about the potential consequences of cyberbullying and suicide coverage.'}",
-impact of an electronic health record (ehr) reminder on human papillomavirus (hpv) vaccine initiation and timely completion,M. Ruffin,"Background: The initiation and timely completion of the human papillomavirus (HPV) vaccine in young women is critical. We compared the initiation and completion of the HPV vaccine among women in 2 community-based networks with electronic health records: 1 with a prompt and reminder system (prompted cohort) and 1 without (unprompted cohort). Methods: Female patients aged 9 to 26 years seen between March 1, 2007, and January 25, 2010, were used as the retrospective cohort. Patient demographics and vaccination dates were extracted from the electronic health records. Results: Patients eligible for the vaccine included 6019 from the prompted cohort and 9096 from the unprompted cohort. Mean age at initiation was 17.3 years in the prompted cohort and 18.1 years in the unprompted cohort. Significantly more (P < .001) patients initiated the vaccine in the prompted cohort (34.9%) compared with the unprompted cohort (21.5%). African Americans aged 9 to 18 years with ≥3 visits during the observation period were significantly more likely to initiate in the prompted cohort (P < .001). The prompted cohort was significantly more likely (P < .001) to complete the vaccine series in a timely manner compared with the unprompted cohort. Conclusion: More patients aged 9 to 26 years initiated and achieved timely completion of the HPV vaccine series in clinics using an electronic health record system with prompts compared with clinics without prompts.","{'model': 'tldr@v2.0.0', 'text': 'The prompted cohort was significantly more likely to complete the vaccine series in a timely manner compared with the unprompted cohort and more patients aged 9 to 26 years initiated and achieved timely completion of the HPV vaccines in clinics using an electronic health record system with prompts.'}",https://www.jabfm.org/content/jabfp/28/3/324.full.pdf
-comparative study using various methods for identification of staphylococcus species in clinical specimens,F. Layer,"ABSTRACT Coagulase-negative staphylococci (CNS) play a predominant role in nosocomial infections. Rapid, reliable identification of these organisms is essential for accurate diagnosis and prompt effective treatment of these infections. Quite recently, the VITEK 2 g-positive (gram-positive [GP]) identification card (bioMérieux) has been redesigned for greater accuracy in the identification of gram-positive cocci. We compared the BD Phoenix (Becton Dickinson) and VITEK 2 (bioMérieux) automated microbiology systems, using their respective update version cards, and the API ID32 STAPH test. The glyceraldehyde-3-phosphate dehydrogenase (gap) gene-based T-RFLP (terminal restriction fragment length polymorphism) method was used for verifying the results. In total, 86 clinical isolates of CNS and 27 reference strains were analyzed. The results show that for identification of CNS, the automated identification methods using the newest VITEK 2 and BD Phoenix identification cards are comparable. However, API ID32 STAPH revealed more correct results compared to both automated microbiology systems. Despite the increased performance of the phenotypic automated identification systems compared to the former versions, molecular methods, e.g., the gap-based T-RFLP method, still show superior accuracy in identifying Staphylococcus species other than Staphylococcus aureus.","{'model': 'tldr@v2.0.0', 'text': 'Despite the increased performance of the phenotypic automated identification systems compared to the former versions, molecular methods, e.g., the gap-based T-RFLP method, still show superior accuracy in identifying Staphylococcus species other than Staphlyococcus aureus.'}",https://europepmc.org/articles/pmc1594629?pdf=render
-boron determination—a review of analytical methods,R. N. Sah,"This paper reviews published methods of sample preparation, determinand purification, and the determination of boron concentration and isotopic composition in a sample. The most common methods for the determination of B concentration are spectrophotometric and plasma-source spectrometric methods. Although most spectrophotometric methods are based on colorimetric reactions of B with azomethine-H, curcumin, or carmine, other colorimetric and fluorometric methods have also been used to some extent. These methods, in general, suffer from numerous interferences and have low sensitivity and precision. Application of nuclear reaction and atomic emission/absorption spectrometric (AES/AAS) methods has remained limited because these methods have poor sensitivity and suffer from serious memory effects and interferences. Among a large number of published nuclear reaction methods only prompt-γ spectrometry has been of practical use. The prompt-γ method can determine B concentration in intact samples, which makes this method especially useful for some medical applications, including boron neutron capture therapy. However, this is a time-consuming method and not suitable for detection of low levels of B. Inductively coupled plasma optical emission spectrometry (ICP-OES) created a new dimension in B determination because of its simplicity, sensitivity, and multielement capability. However, it suffers interferences and is not adequately sensitive for some nutritional and medical applications involving animal tissues that are naturally low in B. All methods involving the measurement of B isotopic composition require a mass spectrometer. Thermal ionization mass spectrometry (TIMS) and secondary ion mass spectrometry (SIMS) have been used to measure isotopic composition of B; however, these methods are time consuming and require extensive sample preparation and purification. Development of inductively coupled plasma mass spectrometry (ICP-MS) not only overcame most of the drawbacks of earlier methods, but also its capabiltiy of measuring B isotopes made possible (1) B concentration determination by isotope dilution, (2) verification of B concentration by isotope fingerprinting in routine analysis, and (3) determination of total B concentration and B isotope ratio for biological tracer studies in the same run. Therefore, plasma source MS appears to be the method of choice among present-day technologies.",,
-concept drift detection for online class imbalance learning,Shuo Wang,"Concept drift detection methods are crucial components of many online learning approaches. Accurate drift detections allow prompt reaction to drifts and help to maintain high performance of online models over time. Although many methods have been proposed, no attention has been given to data streams with imbalanced class distributions, which commonly exist in real-world applications, such as fault diagnosis of control systems and intrusion detection in computer networks. This paper studies the concept drift problem for online class imbalance learning. We look into the impact of concept drift on single-class performance of online models based on three types of classifiers, under seven different scenarios with the presence of class imbalance. The analysis reveals that detecting drift in imbalanced data streams is a more difficult task than in balanced ones. Minority-class recall suffers from a significant drop after the drift involving the minority class. Overall accuracy is not suitable for drift detection. Based on the findings, we propose a new detection method DDM-OCI derived from the existing method DDM. DDM-OCI monitors minority-class recall online to capture the drift. The results show a quick response of the online model working with DDM-OCI to the new concept.","{'model': 'tldr@v2.0.0', 'text': 'The analysis reveals that detecting drift in imbalanced data streams is a more difficult task than in balanced ones, and proposes a new detection method DDM-OCI derived from the existing methodDDM.'}",
-cervical spine clearance in blunt trauma: evaluation of a computed tomography-based protocol.,B. Sanchez,"BACKGROUND Prompt identification of cervical spine injuries has been a critical issue in trauma management. In 1998, the authors developed a new protocol to evaluate cervical spines in blunt trauma. This protocol relies on clinical clearance for appropriate patients and helical computed tomography instead of plain radiographs for patients who cannot be clinically cleared. The authors then prospectively collected data on all cervical spine evaluations to assess the sensitivity and specificity of their approach. METHODS Any patient without clinical evidence of neurologic injury, alcohol or drug intoxication, or distracting injury underwent cervical spine evaluation by clinical examination. Patients who did not meet these criteria underwent helical computed tomographic scanning of the entire cervical spine. For patients who had neurologic deficits, a magnetic resonance image was obtained. If the patient was not evaluable secondary to coma, the computed tomographic scan was without abnormality, and the patient was moving all four extremities at arrival in the emergency department, the cervical spine was cleared, and spinal precautions were removed. Data were collected for all patients admitted to Santa Barbara Cottage Hospital trauma service between 1999 and 2002. The authors selected for analysis patients with blunt trauma and further identified those with closed head injuries (Glasgow Coma Scale score < 15 and loss of consciousness). In addition, all blunt cervical spine injuries were reviewed. RESULTS During the period of study, 2,854 trauma patients were admitted, of whom 2,603 (91%) had blunt trauma. Of these, 1,462 (56%) had closed head injuries. One hundred patients (7% of patients admitted for blunt trauma) had cervical spine or spinal cord injuries, of which 99 were identified by the authors' protocol. Only one injury was not appreciated in a patient with syringomyelia. Fifteen percent of patients with spinal cord injury had no radiographic abnormality; all of these patients presented with neurologic deficits. The sensitivity for detecting cervical spine injury was thus 99%, and the specificity was 100%. The risk of missing a cervical spine injury in these blunt trauma patients was 0.04%. The authors missed no spine injuries in patients with head injuries. CONCLUSION The use of the authors' protocol resulted in excellent sensitivity and specificity in detecting cervical spine injuries. In addition, it allowed early removal of spinal precautions.","{'model': 'tldr@v2.0.0', 'text': ""The use of the authors' protocol resulted in excellent sensitivity and specificity in detecting cervical spine injuries and allowed early removal of spinal precautions.""}",
-indocyanine green lymphographic evidence of surgical efficacy following microsurgical and supermicrosurgical lymphedema reconstructions,Wei F. Chen,"Abstract Background Microsurgical vascularized lymph node transfer (VLNT) and supermicrosurgical lymphaticovenular anastomosis (LVA) are increasingly performed to treat lymphedema. The surgical outcome is commonly assessed by volume-based measurement (VBM), a method that is not consistently reliable. We describe indocyanine green (ICG) lymphography as an alternative postoperative tracking modality after lymphatic reconstruction with VLNT and LVA. Methods VLNT and LVA were performed in patients with therapy-refractory lymphedema. Patients were evaluated qualitatively by clinical assessment, quantitatively with VBM, and lymphographically using ICG lymphography. The evaluation was performed preoperatively, and at 3, 6, and 12-month postoperatively. Results Overall, 21 patients underwent lymphatic reconstruction with either VLNT or LVA. All reported prompt and durable relief of symptoms during the study period. All experienced disease regression based on the Campisi criteria. Out of the 21 patients, 20 (95%) demonstrated lymphographic down staging of disease severity. Out of the 21 patients, 3 (14%) developed a paradoxical increase in limb volume based on VBM despite clinical improvement. Conclusions ICG lymphography correlated highly with patient self-assessment and clinical examination, and is an effective postoperative tracking modality after lymphatic reconstruction.","{'model': 'tldr@v2.0.0', 'text': 'ICG lymphography correlated highly with patient self-assessment and clinical examination, and is an effective postoperative tracking modality after lymphatic reconstruction with VLNT and LVA.'}",
-"nonpalpable, circumscribed, noncalcified solid breast masses: likelihood of malignancy based on lesion size and age of patient.",E. Sickles,"PURPOSE To determine whether lesion size and patient age should prompt immediate biopsy of selected nonpalpable, circumscribed, noncalcified solid (probably benign) breast masses, which otherwise would be managed with periodic mammographic surveillance. MATERIALS AND METHODS Consecutive cases of probably benign breast masses were studied, for which mammographic surveillance was recommended prospectively. Patient age and lesion size were recorded, but age and size thresholds were not used as eligibility criteria. Clinical outcome, determined for each patient after 3 or 3 1/2 years of surveillance, was evaluated as a function of patient age and lesion size. RESULTS There were 1,403 study cases, among which 19 cancers were diagnosed (positive predictive value [PPV] = 1.4%). Only small differences were found in PPV for various patient-age and lesion-size subgroups. Even the subgroup with the highest PPV, which consisted of women aged 50 years and older, contained 60 benign masses for each cancer. CONCLUSION Nonpalpable, circumscribed, noncalcified (probably benign) breast masses should be managed with periodic mammographic surveillance regardless of lesion size and patient age.","{'model': 'tldr@v2.0.0', 'text': 'Nonpalpable, circumscribed, noncalcified (probably benign) breast masses should be managed with periodic mammographic surveillance regardless of lesion size and patient age, regardless of patient age.'}",
-study on pid parameters tuning method based on matlab/simulink,Suping Li,"Enormous calculation of proportional-integral-derivative (PID) controller parameters tuning with analytic methods is an important problem demanding prompt solution. Parameters tuning based on Matlab/Simulink is simplicity, visual manipulation method which leaves out above program. According to the Ziegler-Nichols (Z-N) method, this paper introduces how to reduce and validate the PID controller parameter with the help of MATLAB tool taking a certain control model as an example. The simulation results show the effectiveness of this method and can be fit for application in the engineering.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces how to reduce and validate the PID controller parameter with the help of MATLAB tool taking a certain control model as an example and the simulation results show the effectiveness of this method and can be fit for application in the engineering.'}",
-management of severe hyperkalemia,L. Weisberg,"Background and Objectives:Hyperkalemia is one of the few potentially lethal electrolyte disturbances. Prompt recognition and expeditious treatment of severe hyperkalemia are expected to save lives. This review is intended to provide intensivists and other interested clinicians with an understanding of the pathophysiology that underlies hyperkalemia, and a rational approach to its management. Methods:This article reviews and analyzes literature relevant to the pathophysiology and management of severe hyperkalemia. Methods include search of MEDLINE, and bibliographic search of current textbooks and journal articles. Results and Conclusions:A more complete understanding of potassium homeostasis in recent years has led to new approaches to the management of severe hyperkalemia. The physiologically based sequential approach still applies. The efficacy, pitfalls, and risks of the agents available for use at each step in the sequence are critically reviewed. Rational use of the available tools will allow clinicians to successfully treat severe hyperkalemia.","{'model': 'tldr@v2.0.0', 'text': 'A more complete understanding of potassium homeostasis in recent years has led to new approaches to the management of severe hyperkalemia, and the physiologically based sequential approach still applies.'}",
-"prompts to disrupt sitting time and increase physical activity at work, 2011–2012",A. Swartz,"Introduction The objective of this study was to assess change in sitting and physical activity behavior in response to a workplace intervention to disrupt prolonged sitting time. Methods Sixty office workers were randomized to either a Stand group (n = 29), which received hourly prompts (computer-based and wrist-worn) to stand up, or a Step group (n = 31), which received the same hourly prompts and an additional prompt to walk 100 steps or more upon standing. An ActivPAL monitor was used to assess sitting and physical activity behavior on the same 3 consecutive workdays during baseline and intervention periods. Mixed-effect models with random intercepts and random slopes for time were performed to assess change between groups and across time. Results Both groups significantly reduced duration of average sitting bouts (Stand group, by 16%; Step group, by 19%) and the number of sitting bouts of 60 minutes or more (Step group, by 36%; Stand group, by 54%). The Stand group significantly reduced total sitting time (by 6.6%), duration of the longest sitting bout (by 29%), and number of sitting bouts of 30 minutes or more (by 13%) and increased the number of sit-to-stand transitions (by 15%) and standing time (by 23%). Stepping time significantly increased in the Stand (by 14%) and Step (by 29%) groups, but only the Step group significantly increased (by 35%) the number of steps per workday. Differences in changes from baseline to intervention between groups were not significant for any outcome. Conclusion Interventions that focus on disrupting sitting time only in the workplace may result in less sitting. When sitting time disruptions are paired with a physical activity prompt, people may be more likely to increase their workday physical activity, but the effect on sitting time may be attenuated.","{'model': 'tldr@v2.0.0', 'text': 'Interventions that focus on disrupting sitting time only in the workplace may result in less sitting, but when sitting time disruptions are paired with a physical activity prompt, people may be more likely to increase their workday physical activity, but the effect on sitting time may be attenuated.'}",https://www.cdc.gov/pcd/issues/2014/pdf/13_0318.pdf
-modeling thesis clarity in student essays,Isaac Persing,"Recently, researchers have begun exploring methods of scoring student essays with respect to particular dimensions of quality such as coherence, technical errors, and relevance to prompt, but there is relatively little work on modeling thesis clarity. We present a new annotated corpus and propose a learning-based approach to scoring essays along the thesis clarity dimension. Additionally, in order to provide more valuable feedback on why an essay is scored as it is, we propose a second learning-based approach to identifying what kinds of errors an essay has that may lower its thesis clarity score.","{'model': 'tldr@v2.0.0', 'text': 'A new annotated corpus is presented and a learning- based approach to scoring essays along the thesis clarity dimension is proposed and a second learning-based approach to identifying what kinds of errors an essay has that may lower its thesis clarity score is proposed.'}",
-nuclear reaction prompt gamma-ray analysis,G. Molnár,,,
-validation of multidimensional scaling-based modeling of alcohol expectancies in memory: age and drinking-related differences in expectancies of children assessed as first associates.,M. Dunn,"BACKGROUND As evidence has accumulated that alcohol expectancies mediate the effects of other drinking antecedents, attempts to understand the mechanism by which expectancies influence behavior have focused on modeling memory processes. Previous expectancy work, however, has used relatively indirect approaches to retrieve and model information stored in memory. By using the method most recommended by memory researchers for directly obtaining uncontaminated memory contents, we assessed children's expectancies and related findings to empirically modeled organization and activation of expectancies in memory based on scaled instruments. METHODS Individual interviews were conducted with 462 children in 2nd through 5th grades, and surveys were completed by 1,003 children in 3rd, 6th, 9th, and 12th grades. Interviews and surveys consisted of a measure designed to retrieve participants' first expectancy associate to an alcohol prompt and several drinking quantity/frequency questions. RESULTS Older and higher drinking children were more likely to report positive expectancies as their first associate to an alcohol prompt. Age and drinking-based findings were consistent with organizational structure, dimension emphasis shift, and paths of association identified by prior multidimensional scaling techniques. CONCLUSIONS Consumption of alcohol among children corresponded to accessibility of positive expectancies in memory. In addition, the use of multidimensional scaling to study the organization and activation of alcohol expectancies in memory was validated.","{'model': 'tldr@v2.0.0', 'text': ""Children's expectancies and related findings were assessed to empirically modeled organization and activation of expectancies in memory based on scaled instruments and the use of multidimensional scaling to study the organization andactivation of alcohol expectancy in memory was validated.""}",
-increasing the vocalizations of individuals with autism during intervention with a speech-generating device.,Cindy Gevarter,"This study aimed to teach individuals with autism spectrum disorder (ASD) and limited vocal speech to emit target vocalizations while using a speech-generating device (SGD). Of the 4 participants, 3 began emitting vocal word approximations with SGD responses after vocal instructional methods (delays, differential reinforcement, prompting) were introduced. Two participants met mastery criterion with a reinforcer delay and differential reinforcement, and 1 met criterion after fading an echoic model and prompt delay. For these participants, vocalizations initiated before speech outputs were shown to increase, and vocalizations generalized to a context in which the SGD was absent. The 4th participant showed high vocalization rates only when prompted. The results suggest that adding vocal instruction to an SGD-based intervention can increase vocalizations emitted along with SGD responses for some individuals with ASD.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that adding vocal instruction to an SGD-based intervention can increase vocalizations emitted along with SGD responses for some individuals with ASD.'}",
-regression-based quantitative-trait-locus mapping in the 21st century.,E. Feingold,"In the beginning, there was Haseman-Elston regression. This tool for human QTL mapping, developed in 1972, was simple and inspired. The idea was to take pairs of siblings and regress the squared differences in their trait values on their identity-by-descent (IBD) sharing at a marker. If the marker is linked to the trait, high levels of IBD sharing should be associated with a small difference in trait values, and the regression slope should be negative. Thus, linkage can be tested with a regression t test. This method (with some extensions) was predominant in human studies for >20 years, which was primarily a reflection of the fact that too little human QTL mapping was being performed to prompt the development of more sophisticated methods. In the mid-1990s, we saw the first important alternative to Haseman-Elston regression, maximum-likelihood–based variance-components estimation (see, e.g., Amos 1994; Almasy and Blangero 1998). Variance components is seamlessly applicable to any type of pedigree, whereas Haseman-Elston regression is not, and it has substantially higher power than Haseman-Elston when trait distributions are approximately Gaussian. It has superseded Haseman-Elston as the method of choice for most studies, particularly when large pedigrees are used. However, variance components relies heavily on normality assumptions and can fail dramatically when those assumptions are violated either by nonnormality of the trait distribution or by selected sampling. Attempts to “robustify” variance components have had mixed success (see Feingold [2001] for a more complete discussion), so there is still a role for regression-based methods, which are intrinsically more robust. In the past 5 years, there has been an avalanche of attempts to improve the power of Haseman-Elston regression and to bring regression-based QTL mapping up to date. This was set off by Wright's (1997) Letter to the Editor suggesting that it is beneficial to use the trait values of both members of a sib pair rather than just the squared difference (although this was, in fact, pointed out by Gaines and Elston [1969]). Since then, there have been six articles suggesting “revised Haseman-Elston” (regression-based) methods that use the bivariate data—by Drigalenko (1998), Elston et al. (2000), Xu et al. (2000), Forrest (2001), Sham and Purcell (2001), and Visscher and Hopper (2001). I believe that this is a complete list, but I offer profound apologies to anyone I may have omitted. There have also been three new articles discussing score statistics that have properties similar to the regression-based methods, by Tang and Siegmund (2001), Putter et al. (2002), and Wang and Huang (2002). The best of these new methods have succeeded in matching the power of variance components while retaining the robustness of the regression framework. However, they are all limited to sibships, or, in some cases, to sib pairs. In this issue of the Journal, Sham et al. (2002) take the logical next step, by developing a regression-based method that can be applied to extended pedigrees. Those of us trying to map human QTLs have a much richer set of tools available to us than we did 5 years ago. However, the abundance of new methods has made it difficult to make choices. Only true aficionados can keep up with the literature. In this editorial, I briefly review the newest options. I will describe the new regression-based methods and score statistics, compare their strengths and weaknesses, and conclude by describing how the current offering from Sham et al. (2002 [in this issue]) fits in. I will start, however, with a disclaimer. Because all of these methods are very new, they have not been tested extensively. Most of my observations below are based on statistical theory, and I'm sure that further study of the statistics will prove at least some of my guesses wrong. A related caveat is that all of the theory I rely on is large-sample theory, and even among statistics that are asymptotically identical there may be important differences in small-sample behavior.","{'model': 'tldr@v2.0.0', 'text': 'The new regression-based methods and score statistics are described, compare their strengths and weaknesses, and conclude by describing how the current offering from Sham et al. (2002) fits in.'}",http://www.cell.com/article/S0002929707604686/pdf
-punifiedner: a prompting-based unified ner system for diverse datasets,Jinghui Lu,"Much of named entity recognition (NER) research focuses on developing dataset-specific models based on data from the domain of interest, and a limited set of related entity types. This is frustrating as each new dataset requires a new model to be trained and stored. In this work, we present a ``versatile'' model---the Prompting-based Unified NER system (PUnifiedNER)---that works with data from different domains and can recognise up to 37 entity types simultaneously, and theoretically it could be as many as possible. By using prompt learning, PUnifiedNER is a novel approach that is able to jointly train across multiple corpora, implementing intelligent on-demand entity recognition. Experimental results show that PUnifiedNER leads to significant prediction benefits compared to dataset-specific models with impressively reduced model deployment costs. Furthermore, the performance of PUnifiedNER can achieve competitive or even better performance than state-of-the-art domain-specific methods for some datasets. We also perform comprehensive pilot and ablation studies to support in-depth analysis of each component in PUnifiedNER.","{'model': 'tldr@v2.0.0', 'text': ""This work presents a ``versatile'' model---the Prompting-based Unified NER system (PUnifiedNER)---that works with data from different domains and can recognise up to 37 entity types simultaneously, and theoretically it could be as many as possible.""}",http://arxiv.org/pdf/2211.14838
-prompting through prototype: a prototype-based prompt learning on pretrained vision-language models,Yue Zhang,"Prompt learning is a new learning paradigm which reformulates downstream tasks as similar pretraining tasks on pretrained models by leveraging textual prompts. Recent works have demonstrated that prompt learning is particularly useful for few-shot learning, where there is limited training data. Depending on the granularity of prompts, those methods can be roughly divided into task-level prompting and instance-level prompting. Task-level prompting methods learn one universal prompt for all input samples, which is efficient but ineffective to capture subtle differences among different classes. Instance-level prompting methods learn a specific prompt for each input, though effective but inefficient. In this work, we develop a novel prototype-based prompt learning method to overcome the above limitations. In particular, we focus on few-shot image recognition tasks on pretrained vision-language models (PVLMs) and develop a method of prompting through prototype (PTP), where we define $K$ image prototypes and $K$ prompt prototypes. In PTP, the image prototype represents a centroid of a certain image cluster in the latent space and a prompt prototype is defined as a soft prompt in the continuous space. The similarity between a query image and an image prototype determines how much this prediction relies on the corresponding prompt prototype. Hence, in PTP, similar images will utilize similar prompting ways. Through extensive experiments on seven real-world benchmarks, we show that PTP is an effective method to leverage the latent knowledge and adaptive to various PVLMs. Moreover, through detailed analysis, we discuss pros and cons for prompt learning and parameter-efficient fine-tuning under the context of few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work focuses on few-shot image recognition tasks on pretrained vision-language models (PVLMs) and develops a method of prompting through prototype (PTP), where the image prototype represents a centroid of a certain image cluster in the latent space and a prompt prototype is defined as a soft prompt in the continuous space.'}",http://arxiv.org/pdf/2210.10841
-self-prompting large language models for open-domain qa,Junlong Li,"Open-Domain Question Answering (ODQA) requires models to answer factoid questions with no context given. The common way for this task is to train models on a large-scale annotated dataset to retrieve related documents and generate answers based on these documents. In this paper, we show that the ODQA architecture can be dramatically simplified by treating Large Language Models (LLMs) as a knowledge corpus and propose a Self-Prompting framework for LLMs to perform ODQA so as to eliminate the need for training data and external knowledge corpus. Concretely, we firstly generate multiple pseudo QA pairs with background passages and one-sentence explanations for these QAs by prompting LLMs step by step and then leverage the generated QA pairs for in-context learning. Experimental results show our method surpasses previous state-of-the-art methods by +8.8 EM averagely on three widely-used ODQA datasets, and even achieves comparable performance with several retrieval-augmented fine-tuned models.","{'model': 'tldr@v2.0.0', 'text': 'This paper shows that the ODQA architecture can be dramatically simplified by treating Large Language Models (LLMs) as a knowledge corpus and proposes a Self-Prompting framework for LLMs to perform ODZA so as to eliminate the need for training data and external knowledge corpus.'}",http://arxiv.org/pdf/2212.08635
-dialogue state tracking with a language model using schema-driven prompting,Chia-Hsuan Lee,"Task-oriented conversational systems often use dialogue state tracking to represent the user’s intentions, which involves filling in values of pre-defined slots. Many approaches have been proposed, often using task-specific architectures with special-purpose classifiers. Recently, good results have been obtained using more general architectures based on pretrained language models. Here, we introduce a new variation of the language modeling approach that uses schema-driven prompting to provide task-aware history encoding that is used for both categorical and non-categorical slots. We further improve performance by augmenting the prompting with schema descriptions, a naturally occurring source of in-domain knowledge. Our purely generative system achieves state-of-the-art performance on MultiWOZ 2.2 and achieves competitive performance on two other benchmarks: MultiWOZ 2.1 and M2M. The data and code will be available at https://github.com/chiahsuan156/DST-as-Prompting.","{'model': 'tldr@v2.0.0', 'text': 'A new variation of the language modeling approach that uses schema-driven prompting to provide task-aware history encoding that is used for both categorical and non-categorical slots is introduced.'}",https://aclanthology.org/2021.emnlp-main.404.pdf
-mapl: parameter-efficient adaptation of unimodal pre-trained models for vision-language few-shot prompting,Oscar Mañas,"Large pre-trained models have proved to be remarkable zero- and (prompt-based) few-shot learners in unimodal vision and language tasks. We propose MAPL, a simple and parameter-efficient method that reuses frozen pre-trained unimodal models and leverages their strong generalization capabilities in multimodal vision-language (VL) settings. MAPL learns a lightweight mapping between the representation spaces of unimodal models using aligned image-text data, and can generalize to unseen VL tasks from just a few in-context examples. The small number of trainable parameters makes MAPL effective at low-data and in-domain learning. Moreover, MAPL’s modularity enables easy extension to other pre-trained models. Extensive experiments on several visual question answering and image captioning benchmarks show that MAPL achieves superior or competitive performance compared to similar methods while training orders of magnitude fewer parameters. MAPL can be trained in just a few hours using modest computational resources and public datasets. We release our code and pre-trained model weights at https://github.com/oscmansan/mapl.","{'model': 'tldr@v2.0.0', 'text': 'MAPL is a simple and parameter-efficient method that reuses frozen pre-trained unimodal models and leverages their strong generalization capabilities in multimodal vision-language (VL) settings and learns a lightweight mapping between the representation spaces of unimmodal models using aligned image-text data.'}",http://arxiv.org/pdf/2210.07179
-prompting medical students to self–assess their learning needs during the ageing and health module: a mixed methods study,Grace Kennedy,"ABSTRACT Understanding our learning needs is fundamental for safe, effective and knowledge-based medical practice and facilitates life-long learning. A mixed methods study investigated fourth-year medical students’ self-perceived understanding of their learning needs using 1] a visual scale, before and after a four-week module in Ageing and Health (A&H) and 2] through focus group discussions. During 2013–14 academic year, all students (252) were invited to use a Visual Analogue Scale (VAS) tool to self-assess their learning needs that were linked to Ageing and Health curriculum learning outcomes. Assenting students (197 at pre-self-assessment, 201 at post-assessment) returned anonymous Visual Analogue Scales, self-assessing history-taking skills, examination skills, knowledge of medication use, co-morbidity, nutritional and swallowing assessment responses, before and after the A&H module. Three student focus groups explored whether completion of the VAS self-assessment had prompted improved self-awareness of their learning needs. The VAS responses increased for each curriculum domain with significant differences between the pre-and post responses – for the student-year-group. Nutritional and swallowing knowledge showed the greatest improvement from a self-assessed low baseline at entry. Focus-group students generally viewed the VAS tool positively, and as an aid for prompting consideration of current and future clinical practice. Some students recognised that ‘a need to be ready-for-work’ focused engaged learning; others demonstrated self-regulated learning through self-motivation and an action plan. The Visual Analogue Scale quantitative responses showed increased student-self-perceived learning for each curriculum domain at fourth-year completion of the A&H module, suggesting that prompting self-assessment had increased students’ knowledge and skills. Focus group students saw the VAS tool as useful for prompting awareness of their current and future learning needs. Additional educational strategies should be explored to enable all students to self-reflect and engage effectively on their learning needs, to gain the skills for the maintenance of professional medical competence. Abbreviations: A&H: Ageing and Health Module; e-portfolio: an electronic version of an evidence portfolio, which allows medical students and graduates to reflect and document learning and competencies; F1: year1 of post-graduate medical clinical training; GMC: General Medical Council-the regulation organisation for maintaining standards for doctors in UK; Logbook: usually a written document which can be used to record procedures and attendance at clinics or case-based discussions and can be used to set learning outcomes and to structure teaching in clinical settings for medical students and doctors; PDP: personal development plan is used to plan future learning and skills needs for work and education with an plan for action/s outcome; SPSS: Statistical Package for the Social Sciences; VAS: Visual Analogue Scale is a visual method of describing an experience.","{'model': 'tldr@v2.0.0', 'text': 'The Visual Analogue Scale quantitative responses showed increased student-self-perceived learning for each curriculum domain at fourth-year completion of the A&H module, suggesting that prompting self-assessment had increased students’ knowledge and skills.'}",https://www.tandfonline.com/doi/pdf/10.1080/10872981.2019.1579558?needAccess=true
-transprompt: towards an automatic transferable prompting framework for few-shot text classification,Chengyu Wang,"Recent studies have shown that prompts improve the performance of large pre-trained language models for few-shot text classification. Yet, it is unclear how the prompting knowledge can be transferred across similar NLP tasks for the purpose of mutual reinforcement. Based on continuous prompt embeddings, we propose TransPrompt, a transferable prompting framework for few-shot learning across similar tasks. In TransPrompt, we employ a multi-task meta-knowledge acquisition procedure to train a meta-learner that captures cross-task transferable knowledge. Two de-biasing techniques are further designed to make it more task-agnostic and unbiased towards any tasks. After that, the meta-learner can be adapted to target tasks with high accuracy. Extensive experiments show that TransPrompt outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets. We further show that the meta-learner can effectively improve the performance on previously unseen tasks; and TransPrompt also outperforms strong fine-tuning baselines when learning with full training sets.","{'model': 'tldr@v2.0.0', 'text': 'In TransPrompt, a transferable prompting framework for few-shot learning across similar NLP tasks, a multi-task meta-knowledge acquisition procedure is employed to train a meta-learner that captures cross-task transferable knowledge.'}",https://aclanthology.org/2021.emnlp-main.221.pdf
-context-faithful prompting for large language models,Wenxuan Zhou,"Large language models (LLMs) encode parametric knowledge about world facts and have shown remarkable performance in knowledge-driven NLP tasks. However, their reliance on parametric knowledge may cause them to overlook contextual cues, leading to incorrect predictions in context-sensitive NLP tasks (e.g., knowledge acquisition tasks). In this paper, we seek to assess and enhance LLMs' contextual faithfulness in two aspects: knowledge conflict and prediction with abstention. We demonstrate that LLMs' faithfulness can be significantly improved using carefully designed prompting strategies. In particular, we identify opinion-based prompts and counterfactual demonstrations as the most effective methods. Opinion-based prompts reframe the context as a narrator's statement and inquire about the narrator's opinions, while counterfactual demonstrations use instances containing false facts to improve faithfulness in knowledge conflict situations. Neither technique requires additional training. We conduct experiments on three datasets of two standard NLP tasks, machine reading comprehension and relation extraction, and the results demonstrate significant improvement in faithfulness to contexts. Code and data are released at https://github.com/wzhouad/context-faithful-llm.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that LLMs' faithfulness can be significantly improved using carefully designed prompting strategies, and opinion-based prompts and counterfactual demonstrations are identified as the most effective methods.""}",http://arxiv.org/pdf/2303.11315
-cora: adapting clip for open-vocabulary detection with region prompting and anchor pre-matching,Xiaoshi Wu,"Open-vocabulary detection (OVD) is an object detection task aiming at detecting objects from novel categories beyond the base categories on which the detector is trained. Recent OVD methods rely on large-scale visual-language pre-trained models, such as CLIP, for recognizing novel objects. We identify the two core obstacles that need to be tackled when incorporating these models into detector training: (1) the distribution mismatch that happens when applying a VL-model trained on whole images to region recognition tasks; (2) the difficulty of localizing objects of unseen classes. To overcome these obstacles, we propose CORA, a DETR-style framework that adapts CLIP for Open-vocabulary detection by Region prompting and Anchor pre-matching. Region prompting mitigates the whole-to-region distribution gap by prompting the region features of the CLIP-based region classifier. Anchor pre-matching helps learning generalizable object localization by a class-aware matching mechanism. We evaluate CORA on the COCO OVD benchmark, where we achieve 41.7 AP50 on novel classes, which outperforms the previous SOTA by 2.4 AP50 even without resorting to extra training data. When extra training data is available, we train CORA+ on both ground-truth base-category annotations and additional pseudo bounding box labels computed by CORA. CORA+ achieves 43.1 AP50 on the COCO OVD benchmark and 28.1 box APr on the LVIS OVD benchmark. The code is available at https://github.com/tgxs002/CORA.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes CORA, a DETR-style framework that adapts CLIP for Open-vocabulary detection by Region prompting and Anchor pre-matching, which mitigates the whole-to-region distribution gap by prompting the region features of the CLIP-based region classifier.'}",https://arxiv.org/pdf/2303.13076
-system impact studies for near 100% renewable energy systems dominated by inverter based variable generation,H. Holttinen,"The demand for low carbon energy calls for close to 100% renewable power systems, with decarbonization of other energy sectors adding to the anticipated paradigm shift. Rising levels of variable inverter-based renewable energy sources (VIBRES) are prompting questions about how such systems will be planned and operated when variable renewable generation becomes the dominant technology. Here, we examine the implications of this paradigm shift with respect to planning, operation and system stability, also addressing the need for integration with other energy vectors, including heat, transport and Power-to-X. We highlight the knowledge gaps and provide recommendations for improved methods and models needed as power systems transform towards 100% VIBRES.",,
-information and communication technology based prompting for treatment compliance for people with serious mental illness.,Kaisa Kauppi,"BACKGROUND Non-compliance is a significant problem among people with serious mental disorders, presenting a challenge for mental health professionals. Prompts such as telephone calls, visits, and a posted referral letter to patients are currently used to encourage patient attendance at clinics and/or compliance with medication. More recently, the use of information and communication technology (ICT)-based prompting methods have increased. Methods include mobile text message (SMS - short message service), e-mail or use of any other electronic device with the stated purpose of encouraging compliance. OBJECTIVES To investigate the effects of ICT-based prompting to support treatment compliance in people with serious mental illness compared with standard care. SEARCH METHODS We searched the Cochrane Schizophrenia Group's Trials Register (31(st) May 2011 and 9(th) July 2012) which is based on regular searches of CINAHL, BIOSIS, AMED, EMBASE, PubMed, MEDLINE, PsycINFO, and registries of clinical trials. Also, we inspected references of all identified studies for further trials and contacted authors of trials for additional information. SELECTION CRITERIA Relevant randomised controlled trials involving adults with serious mental illness, comparing any ICT-based prompt or combination of prompts by automatic or semi-automatic system compared with standard care. DATA COLLECTION AND ANALYSIS Review authors reliably assessed trial quality and extracted data. We calculated risk ratio (RR) with 95% confidence intervals (CI) using a fixed-effect model. For continuous outcomes, we estimated the mean difference (MD) between groups, again with 95% confidence intervals. A 'Summary of findings' table using GRADE was created, and we assessed included studies for risk of bias. MAIN RESULTS The search identified 35 references, with 25 studies, but we could only include two studies with a total of 358 participants. The studies had a moderate risk of bias, and therefore risk overestimating any positive effects of ICT-based prompting. Both included studies compared semi-automatised ICT-based prompting intervention with standard care groups in mental health outpatient care. The interventions were SMS-message and an electronic assistant device. One included study reported our primary outcome, compliance.There was not any clear evidence that ICT-based prompts increase improvement in compliance (stop taking medication within six months n = 320, RR 1.11 CI 0.96 to 1.29, moderate quality evidence). There was some low quality evidence that ICT-based prompts have small effects for: mental state (average change in specific symptom scores within three months n = 251, MD -0.30 CI -0.53 to -0.07; severity of illness within three months n = 251, MD -0.10 CI -0.13 to -0.07 and six months n = 251, MD -0.10 CI -0.13 to -0.07; average change in depressive scores within six months n = 251, RR 0.00 CI -0.28 to 0.28; global symptoms within three months n = 251, MD -0.10 CI -0.38 to -0.07; negative symptoms within three months n = 251, MD -0.10 CI -0.38 to 0.18 and six months n = 251, MD -0.30 CI -0.58 to 0.02, low quality evidence). Level of insight improved more among people receiving ICT-based prompt compared with those in the control group at six months (n = 251, MD -0.10 CI -0.13 to -0.07). ICT-based prompts also increased quality of life (average change in quality of life within six months n = 251, RR 0.50 CI 0.19 to 0.81, moderate quality evidence).Based on the existing data, there is no evidence that either intervention is less acceptable than the other (n = 347, 2 RCTs, RR 1.46 CI 0.70 to 3.05, low quality evidence). Included studies did not report outcomes of service utilisation, behaviour, costs or adverse events. AUTHORS' CONCLUSIONS The evidence base on the effects of ICT-based prompts is still inconclusive. Data to clarify ICT-based prompting effects are awaited from an ongoing trial, but further well-conducted trials considering the different ICT-based prompts are warranted.","{'model': 'tldr@v2.0.0', 'text': 'There was not any clear evidence that ICT-based prompts increase improvement in compliance, and both included studies compared semi-automatised I CT-based prompting intervention with standard care groups in mental health outpatient care.'}",
-prompting technologies: a comparison of time-based and context-aware transition-based prompting.,Kayela Robertson,"BACKGROUND While advancements in technology have encouraged the development of novel prompting systems to support cognitive interventions, little research has evaluated the best time to deliver prompts, which may impact the effectiveness of these interventions. OBJECTIVE This study examined whether transition-based context prompting (prompting an individual during task transitions) is more effective than traditional fixed time-based prompting. METHODS Participants were 42 healthy adults who completed 12 different everyday activities, each lasting 1-7 minutes, in an experimental smart home testbed and received prompts to record the completed activities from an electronic memory notebook. Half of the participants were delivered prompts during activity transitions, while the other half received prompts every 5 minutes. Participants also completed Likert-scale ratings regarding their perceptions of the prompting system. RESULTS Results revealed that participants in the transition-based context prompting condition responded to the first prompt more frequently and rated the system as more convenient, natural, and appropriate compared to participants in the time-based condition. CONCLUSIONS Our findings suggest that prompting during activity transitions produces higher adherence to the first prompt and more positive perceptions of the prompting system. This is an important finding given the benefits of prompting technology and the possibility of improving cognitive interventions by using context-aware transition prompting.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that prompting during activity transitions produces higher adherence to the first prompt and more positive perceptions of the prompting system, an important finding given the benefits of prompting technology and the possibility of improving cognitive interventions by using context-aware transition prompting.'}",https://europepmc.org/articles/pmc4803438?pdf=render
-the effect of video modeling and video prompting interventions on individuals with intellectual disability: a systematic literature review,Jiyoon Park,"A popular technology-based intervention is video-based instruction (VBI). VBI is known to be effective in teaching various skills to individuals with intellectual disability. Although many researchers have taught this population various skills using video modeling and/or prompting, none have conducted a literature review for this population. The purpose of this review was to analyze different types of VBI (i.e., video modeling, video prompting) used in intervention research, the skills taught using these methods, and their effectiveness. The findings demonstrate both video modeling and video prompting were used to similar extents, daily/living skills were taught more often, and many studies combined video modeling or video prompting interventions with additional strategies (e.g., error correction, constant time delay).",,https://journals.sagepub.com/doi/pdf/10.1177/0162643418780464
-"standardization of sample collection, isolation and analysis methods in extracellular vesicle research",K. Witwer,"The emergence of publications on extracellular RNA (exRNA) and extracellular vesicles (EV) has highlighted the potential of these molecules and vehicles as biomarkers of disease and therapeutic targets. These findings have created a paradigm shift, most prominently in the field of oncology, prompting expanded interest in the field and dedication of funds for EV research. At the same time, understanding of EV subtypes, biogenesis, cargo and mechanisms of shuttling remains incomplete. The techniques that can be harnessed to address the many gaps in our current knowledge were the subject of a special workshop of the International Society for Extracellular Vesicles (ISEV) in New York City in October 2012. As part of the “ISEV Research Seminar: Analysis and Function of RNA in Extracellular Vesicles (evRNA)”, 6 round-table discussions were held to provide an evidence-based framework for isolation and analysis of EV, purification and analysis of associated RNA molecules, and molecular engineering of EV for therapeutic intervention. This article arises from the discussion of EV isolation and analysis at that meeting. The conclusions of the round table are supplemented with a review of published materials and our experience. Controversies and outstanding questions are identified that may inform future research and funding priorities. While we emphasize the need for standardization of specimen handling, appropriate normative controls, and isolation and analysis techniques to facilitate comparison of results, we also recognize that continual development and evaluation of techniques will be necessary as new knowledge is amassed. On many points, consensus has not yet been achieved and must be built through the reporting of well-controlled experiments.","{'model': 'tldr@v2.0.0', 'text': 'The need for standardization of specimen handling, appropriate normative controls, and isolation and analysis techniques to facilitate comparison of results is emphasized, and it is recognized that continual development and evaluation of techniques will be necessary as new knowledge is amassed.'}",https://www.tandfonline.com/doi/pdf/10.3402/jev.v2i0.20360?needAccess=true&role=button
-a randomized clinical trial of three prompting systems to teach tact relations.,Joseph H. Cihon,"Prompts are commonly used during discrete trial teaching for individuals diagnosed with autism spectrum disorder (ASD). Three commonly used prompting systems include constant time delay, most-to-least prompting, and flexible prompt fading. Most of the research demonstrating the effectiveness of these three prompting strategies have been completed through the use of single subject experimental designs. Some within the field of behaviorally based approaches to ASD interventions have called for more randomized clinical trials of these approaches. The purpose of the present study was to compare these 3 prompting systems to teach tact relations for 27 individuals diagnosed with ASD through a randomized clinical trial without a control group with respect to pre-post responding, generalization, sessions to mastery, and responding during teaching. Overall, minimal differences were observed across the 3 systems. The results are discussed with respect to future research and clinical application of the methods evaluated.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of the present study was to compare these 3 prompting systems to teach tact relations for 27 individuals diagnosed with ASD through a randomized clinical trial without a control group with respect to pre-post responding, generalization, sessions to mastery, and responding during teaching.'}",
-use of a smartphone application for prompting oral medication adherence among adolescents and young adults with cancer,Yelena P. Wu,"OBJECTIVES To explore the feasibility and acceptability of use of a smartphone medication reminder application to promote adherence to oral medications among adolescents and young adults (AYAs) with cancer.. SAMPLE & SETTING 23 AYAs with cancer from a Children's Oncology Group-affiliated children's hospital and a National Cancer Institute-designated comprehensive cancer center in Salt Lake City, UT.. METHODS & VARIABLES Participants were asked to use the application for eight weeks. Data on application usage were obtained from a cloud-based server hosted by the application developers. Weekly self-report questionnaires were completed. Feasibility was assessed through participants' usage and responses. Acceptability was assessed through participants' perceived ease of use and usefulness.. RESULTS Almost all participants used the application at least once. More than half reported that they took their medications immediately when they received reminders. Participants also reported that the application was easy to set up and use, and that it was useful for prompting medication taking.. IMPLICATIONS FOR NURSING Nurses could continue to test the efficacy of integrating e-health modalities, such as smartphone applications, into efforts to promote medication adherence.","{'model': 'tldr@v2.0.0', 'text': 'Nursing nurses could continue to test the efficacy of integrating e-health modalities, such as smartphone applications, into efforts to promote medication adherence by exploring the feasibility and acceptability of use of a smartphone medication reminder application among adolescents and young adults with cancer.'}",https://europepmc.org/articles/pmc5927589?pdf=render
-acceptability of artificial intelligence (ai)-led chatbot services in healthcare: a mixed-methods study,T. Nadarzynski,"Background Artificial intelligence (AI) is increasingly being used in healthcare. Here, AI-based chatbot systems can act as automated conversational agents, capable of promoting health, providing education, and potentially prompting behaviour change. Exploring the motivation to use health chatbots is required to predict uptake; however, few studies to date have explored their acceptability. This research aimed to explore participants’ willingness to engage with AI-led health chatbots. Methods The study incorporated semi-structured interviews (N-29) which informed the development of an online survey (N-216) advertised via social media. Interviews were recorded, transcribed verbatim and analysed thematically. A survey of 24 items explored demographic and attitudinal variables, including acceptability and perceived utility. The quantitative data were analysed using binary regressions with a single categorical predictor. Results Three broad themes: ‘Understanding of chatbots’, ‘AI hesitancy’ and ‘Motivations for health chatbots’ were identified, outlining concerns about accuracy, cyber-security, and the inability of AI-led services to empathise. The survey showed moderate acceptability (67%), correlated negatively with perceived poorer IT skills OR = 0.32 [CI95%:0.13–0.78] and dislike for talking to computers OR = 0.77 [CI95%:0.60–0.99] as well as positively correlated with perceived utility OR = 5.10 [CI95%:3.08–8.43], positive attitude OR = 2.71 [CI95%:1.77–4.16] and perceived trustworthiness OR = 1.92 [CI95%:1.13–3.25]. Conclusion Most internet users would be receptive to using health chatbots, although hesitancy regarding this technology is likely to compromise engagement. Intervention designers focusing on AI-led health chatbots need to employ user-centred and theory-based approaches addressing patients’ concerns and optimising user experience in order to achieve the best uptake and utilisation. Patients’ perspectives, motivation and capabilities need to be taken into account when developing and assessing the effectiveness of health chatbots.","{'model': 'tldr@v2.0.0', 'text': 'Intervention designers focusing on AI-led health chatbots need to employ user-centred and theory-based approaches addressing patients’ concerns and optimising user experience in order to achieve the best uptake and utilisation.'}",https://journals.sagepub.com/doi/pdf/10.1177/2055207619871808
-automated detection of activity transitions for prompting,Kyle D. Feuz,"Individuals with cognitive impairment can benefit from intervention strategies like recording important information in a memory notebook. However, training individuals to use the notebook on a regular basis requires a constant delivery of reminders. In this study, we design and evaluate machine-learning-based methods for providing automated reminders using a digital memory notebook interface. Specifically, we identify transition periods between activities as times to issue prompts. We consider the problem of detecting activity transitions using supervised and unsupervised machine-learning techniques and find that both techniques show promising results for detecting transition periods. We test the techniques in a scripted setting with 15 individuals. Motion sensors data are recorded and annotated as participants perform a fixed set of activities. We also test the techniques in an unscripted setting with eight individuals. Motion sensor data are recorded as participants go about their normal daily routine. In both the scripted and unscripted settings, a true positive rate of greater than 80% can be achieved while maintaining a false positive rate of less than 15%. On average, this leads to transitions being detected within 1 min of a true transition for the scripted data and within 2 min of a true transition on the unscripted data.","{'model': 'tldr@v2.0.0', 'text': 'This study designs and evaluates machine-learning-based methods for providing automated reminders using a digital memory notebook interface and identifies transition periods between activities as times to issue prompts.'}",https://europepmc.org/articles/pmc4805372?pdf=render
-training a paraprofessional to implement video prompting with error correction to teach a vocational skill,Rachel L. Seaman-Tullis,"Very few individuals with autism spectrum disorder (ASD) have been trained in the vocational skills needed to obtain gainful employment. Moreover, although there is an abundance of research evaluating the practice of training practitioners of students with ASD to use evidence-based practices to teach a wide variety of skills, there have been few that apply this training to the acquisition of vocational tasks. This study uses a multiple-baseline-across-behaviors design to evaluate the training methods used to train a paraprofessional in the preparation and implementation of video prompting and error correction with his student with ASD. Furthermore, the behavior and learning of both the paraprofessional and student are measured. Results indicate that the training package resulted in increased video prompting implementation behavior for the paraprofessional, as well as corresponding, increased vocational skill behavior for the student.",,
-using a simultaneous prompting procedure with an ipad to teach the pythagorean theorem to adolescents with moderate intellectual disability,Carey Creech-Galloway,"Providing standards-based instruction in core content areas for individuals with moderate and severe disabilities is a hot topic in the field of special education, and teachers struggle to find the best methods for providing high-quality standards-based instruction in core content areas that also has personal relevance for the students. This investigation evaluated the effectiveness of a simultaneous prompting procedure to teach four adolescents with moderate intellectual disabilities to use the Pythagorean theorem to solve real-life scenarios (i.e., sewing, using a ladder, finding dimensions of a screen) shown on a short video on an iPad. A multiple probe design across participants evaluated the effectiveness of the procedure. Results indicate the participants acquired the skill of using the Pythagorean theorem and generalized it to additional real-life, novel problems. Social validity, limitations, and implications for practitioners are also discussed.",,
-a systematic review of methods and procedures used in ecological momentary assessments of diet and physical activity research in youth: an adapted strobe checklist for reporting ema studies (cremas),Y. Liao,"Background Ecological momentary assessment (EMA) is a method of collecting real-time data based on careful timing, repeated measures, and observations that take place in a participant’s typical environment. Due to methodological advantages and rapid advancement in mobile technologies in recent years, more studies have adopted EMA in addressing topics of nutrition and physical activity in youth. Objective The aim of this systematic review is to describe EMA methodology that has been used in studies addressing nutrition and physical activity in youth and provide a comprehensive checklist for reporting EMA studies. Methods Thirteen studies were reviewed and analyzed for the following 5 areas of EMA methodology: (1) sampling and measures, (2) schedule, (3) technology and administration, (4) prompting strategy, and (5) response and compliance. Results Results of this review showed a wide variability in the design and reporting of EMA studies in nutrition and physical activity among youth. The majority of studies (69%) monitored their participants during one period of time, although the monitoring period ranged from 4 to 14 days, and EMA surveys ranged from 2 to 68 times per day. More than half (54%) of the studies employed some type of electronic technology. Most (85%) of the studies used interval-contingent prompting strategy. For studies that utilized electronic devices with interval-contingent prompting strategy, none reported the actual number of EMA prompts received by participants out of the intended number of prompts. About half (46%) of the studies failed to report information about EMA compliance rates. For those who reported, compliance rates ranged from 44-96%, with an average of 71%. Conclusions Findings from this review suggest that in order to identify best practices for EMA methodology in nutrition and physical activity research among youth, more standardized EMA reporting is needed. Missing the key information about EMA design features and participant compliance might lead to misinterpretation of results. Future nutrition and physical activity EMA studies need to be more rigorous and thorough in descriptions of methodology and results. A reporting checklist was developed with the goal of enhancing reliability, efficacy, and overall interpretation of the findings for future studies that use EMAs.","{'model': 'tldr@v2.0.0', 'text': 'Findings from this review suggest that in order to identify best practices for EMA methodology in nutrition and physical activity research among youth, more standardized EMA reporting is needed.'}",
-do universal school‐based mental health promotion programmes improve the mental health and emotional wellbeing of young people? a literature review,C. O’Connor,"AIMS AND OBJECTIVES To examine evidence-using a range of outcomes-for the effectiveness of school-based mental health and emotional well-being programmes. BACKGROUND It is estimated that 20% of young people experience mental health difficulties every year. Schools have been identified as an appropriate setting for providing mental health and emotional well-being promotion prompting the need to determine whether current school-based programmes are effective in improving the mental health and emotional well-being of young people. METHODS A systematic search was conducted using the health and education databases, which identified 29 studies that measured the effectiveness of school-based universal interventions. Prisma guidelines were used during the literature review process. RESULTS Thematic analysis generated three key themes: (i) help seeking and coping; (ii) social and emotional well-being; and (iii) psycho-educational effectiveness. CONCLUSION It is concluded that whilst these studies show promising results, there is a need for further robust evaluative studies to guide future practice. RELEVANCE TO CLINICAL PRACTICE All available opportunities should be taken to provide mental health promotion interventions to young people in the school environment, with a requirement for educational professionals to be provided the necessary skills and knowledge to ensure that the school setting continues to be a beneficial environment for conducting mental health promotion.","{'model': 'tldr@v2.0.0', 'text': 'There is a need for further robust evaluative studies to guide future practice on the effectiveness of school-based universal interventions in mental health and emotional well-being.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/jocn.14078
-a novel hybrid sequence-based model for identifying anticancer peptides,Lei Xu,"Cancer is a serious health issue worldwide. Traditional treatment methods focus on killing cancer cells by using anticancer drugs or radiation therapy, but the cost of these methods is quite high, and in addition there are side effects. With the discovery of anticancer peptides, great progress has been made in cancer treatment. For the purpose of prompting the application of anticancer peptides in cancer treatment, it is necessary to use computational methods to identify anticancer peptides (ACPs). In this paper, we propose a sequence-based model for identifying ACPs (SAP). In our proposed SAP, the peptide is represented by 400D features or 400D features with g-gap dipeptide features, and then the unrelated features are pruned using the maximum relevance-maximum distance method. The experimental results demonstrate that our model performs better than some existing methods. Furthermore, our model has also been extended to other classifiers, and the performance is stable compared with some state-of-the-art works.","{'model': 'tldr@v2.0.0', 'text': 'In this paper, a sequence-based model for identifying anticancer peptides (SAP) is proposed, and the experimental results demonstrate that the model performs better than some existing methods and is stable compared with some state-of-the-art works.'}",https://www.mdpi.com/2073-4425/9/3/158/pdf?version=1520937538
-b-mobile - a smartphone-based intervention to reduce sedentary time in overweight/obese individuals: a within-subjects experimental trial,D. Bond,"Purpose Excessive sedentary time (SED) has been linked to obesity and other adverse health outcomes. However, few sedentary-reducing interventions exist and none have utilized smartphones to automate behavioral strategies to decrease SED. We tested a smartphone-based intervention to monitor and decrease SED in overweight/obese individuals, and compared 3 approaches to prompting physical activity (PA) breaks and delivering feedback on SED. Design and Methods Participants [N = 30; Age = 47.5(13.5) years; 83% female; Body Mass Index (BMI) = 36.2(7.5) kg/m2] wore the SenseWear Mini Armband (SWA) to objectively measure SED for 7 days at baseline. Participants were then presented with 3 smartphone-based PA break conditions in counterbalanced order: (1) 3-min break after 30 SED min; (2) 6-min break after 60 SED min; and (3) 12-min break after 120 SED min. Participants followed each condition for 7 days and wore the SWA throughout. Results All PA break conditions yielded significant decreases in SED and increases in light (LPA) and moderate-to-vigorous PA (MVPA) (p<0.005). Average % SED at baseline (72.2%) decreased by 5.9%, 5.6%, and 3.3% [i.e. by mean (95% CI) −47.2(−66.3, −28.2), −44.5(−65.2, −23.8), and −26.2(−40.7, −11.6) min/d] in the 3-, 6-, and 12-min conditions, respectively. Conversely, % LPA increased from 22.8% to 26.7%, 26.7%, and 24.7% [i.e. by 31.0(15.8, 46.2), 31.0(13.6, 48.4), and 15.3(3.9, 26.8) min/d], and % MVPA increased from 5.0% to 7.0%, 6.7%, and 6.3% (i.e. by 16.2(8.5, 24.0), 13.5(6.3, 20.6), and 10.8(4.2, 17.5) min/d] in the 3-, 6-, and 12-min conditions, respectively. Planned pairwise comparisons revealed the 3-min condition was superior to the 12-min condition in decreasing SED and increasing LPA (p<0.05). Conclusion The smartphone-based intervention significantly reduced SED. Prompting frequent short activity breaks may be the most effective way to decrease SED and increase PA in overweight/obese individuals. Future investigations should determine whether these SED reductions can be maintained long-term. Trial Registration ClinicalTrials.gov NCT01688804","{'model': 'tldr@v2.0.0', 'text': 'The smartphone-based intervention to monitor and decrease SED in overweight/obese individuals, and compared 3 approaches to prompting physical activity (PA) breaks and delivering feedback on SED significantly reduced SED, suggests that Prompting frequent short activity breaks may be the most effective way to decreasing SED and increase PA in obese individuals.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0100821&type=printable
-"self-contradictory hallucinations of large language models: evaluation, detection and mitigation",Niels Mündler,"Large language models (large LMs) are susceptible to producing text that contains hallucinated content. An important instance of this problem is self-contradiction, where the LM generates two contradictory sentences within the same context. In this work, we present a comprehensive investigation into self-contradiction for various instruction-tuned LMs, covering evaluation, detection, and mitigation. Our analysis reveals the prevalence of self-contradictions when LMs generate text for open-domain topics, e.g., in 17.7% of all sentences produced by ChatGPT. Self-contradiction also complements retrieval-based methods, as a large portion of them (e.g., 35.8% for ChatGPT) cannot be verified using Wikipedia. We then propose a novel prompting-based framework designed to effectively detect and mitigate self-contradictions. Our detector achieves high accuracy, e.g., around 80% F1 score when prompting ChatGPT. The mitigation algorithm iteratively refines the generated text to remove contradictory information while preserving text fluency and informativeness. Importantly, our entire framework is applicable to black-box LMs and does not require external grounded knowledge. Our approach is practically effective and has been released as a push-button tool to benefit the public, available at https://chatprotect.ai/.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a novel prompting-based framework designed to effectively detect and mitigate self-contradictions in instruction-tuned LMs, and proposes a novel mitigation algorithm that iteratively refines the generated text to remove contradictory information while preserving text fluency and informativeness.'}",https://arxiv.org/pdf/2305.15852
-systematic review of instructional methods to teach employment skills to secondary students with intellectual and developmental disabilities,Carly B. Gilson,"Effective vocational instruction is an essential aspect of preparing students with intellectual and developmental disabilities (IDD) for the world of work. We systematically reviewed research on instructional methods used to teach employment skills to secondary students with IDD. We identified 56 studies involving 766 participants with IDD. Four intervention approaches emphasized technology or some other instructional stimulus (i.e., self-management devices, video-based, audio-based, picture and tactile-based) and four focused on live instructors (i.e., direct instruction, augmentative and alternative communication, simulation, peer-delivered). Among the 21 instructional methods used within these approaches, performance feedback, device-assisted instruction, response prompting, and community-based instruction were the most common. We address the extent to which these intervention approaches were effective across students, instructional methods, settings, and outcomes, as well as offer recommendations for future research and practice.",,
-convfinqa: exploring the chain of numerical reasoning in conversational finance question answering,Zhiyu Chen,"With the recent advance in large pre-trained language models, researchers have achieved record performances in NLP tasks that mostly focus on language pattern matching. The community is experiencing the shift of the challenge from how to model language to the imitation of complex reasoning abilities like human beings. In this work, we investigate the application domain of finance that involves real-world, complex numerical reasoning. We propose a new large-scale dataset, ConvFinQA, aiming to study the chain of numerical reasoning in conversational question answering. Our dataset poses great challenge in modeling long-range, complex numerical reasoning paths in real-world conversations. We conduct comprehensive experiments and analyses with both the neural symbolic methods and the prompting-based methods, to provide insights into the reasoning mechanisms of these two divisions. We believe our new dataset should serve as a valuable resource to push forward the exploration of real-world, complex reasoning tasks as the next research focus. Our dataset and code is publicly available at https://github.com/czyssrs/ConvFinQA.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new large-scale dataset, ConvFinQA, aiming to study the chain of numerical reasoning in conversational question answering, and conducts comprehensive experiments and analyses with both the neural symbolic methods and the prompting-based methods to provide insights into the reasoning mechanisms.'}",http://arxiv.org/pdf/2210.03849
-patterns of instructional-design factors prompting reflective thinking in middle-school and college level problem-based learning environments,Hae-Deok Song,,,
-intelligent rule-based phishing websites classification,B. Yogesh,"Phishing is depicted as the specialty of reverberating a site of a noteworthy firm meaning to snatch client's private data, for example, usernames, passwords and standardized savings number. Phishing sites involve an assortment of signs inside its substance parts and the program based security pointers furnished alongside the site. A few arrangements have been proposed to handle phishing. All things considered, there is no single enchantment shot that can explain this risk drastically. One of the promising strategies that can be utilized in anticipating phishing assaults depends on information mining, especially the prompting of order principles since hostile to phishing arrangements intend to foresee the site class precisely and that precisely coordinates the information mining characterization procedure objectives. In this study, the creators shed light on the essential elements that recognize phishing sites from authentic ones and evaluate how great standard based information mining arrangement methods are in foreseeing phishing sites and which characterization procedure is turned out to be more solid. Phishing assault traditionally begins by sending an email that appears to originate from a legitimate undertaking to casualties requesting that they upgrade or affirm their own data by going to a connection inside the email. Despite the fact that phishers are presently utilizing a few methods in making phishing sites to trick and charm clients, they all utilization an arrangement of common elements to make phishing sites on the grounds that, without those components they lose the benefit of misdirection. This helps us to separate in the middle of fair and phishing sites taking into account the elements removed from the went to site. By and large, two methodologies are utilized in recognizing phishing sites. The first depends on boycotts, in which the asked for URL is contrasted and those in that rundown. The drawback of this methodology is that the boycott as a rule can't cover all phishing sites, subsequent to, inside seconds; another deceitful site is required to be propelled. The second approach is known as heuristic-based strategy, where a few elements are gathered from the site to arrange it as either phishing or authentic. Rather than the boycott technique, a heuristic-based arrangement can perceive crisply made phishing sites. Keywords- Website features, Phishing, Security, Data Mining, Rule based Classification.","{'model': 'tldr@v2.0.0', 'text': 'Light is shed on the essential elements that recognize phishing sites from authentic ones and how great standard based information mining arrangement methods are in foreseeingphishing sites and which characterization procedure is turned out to be more solid.'}",
-"improving icu-based palliative care delivery: a multicenter, multidisciplinary survey of critical care clinician attitudes and beliefs",N. Wysham,"Objective: Addressing the quality gap in ICU-based palliative care is limited by uncertainty about acceptable models of collaborative specialist and generalist care. Therefore, we characterized the attitudes of physicians and nurses about palliative care delivery in an ICU environment. Design: Mixed-methods study. Setting: Medical and surgical ICUs at three large academic hospitals. Participants: Three hundred three nurses, intensivists, and advanced practice providers. Measurements and Main Results: Clinicians completed written surveys that assessed attitudes about specialist palliative care presence and integration into the ICU setting, as well as acceptability of 23 published palliative care prompts (triggers) for specialist consultation. Most (n = 225; 75%) reported that palliative care consultation was underutilized. Prompting consideration of eligibility for specialist consultation by electronic health record searches for triggers was most preferred (n = 123; 41%); only 17 of them (6%) felt current processes were adequate. The most acceptable specialist triggers were metastatic malignancy, unrealistic goals of care, end of life decision making, and persistent organ failure. Advanced age, length of stay, and duration of life support were the least acceptable. Screening led by either specialists or ICU teams was equally preferred. Central themes derived from qualitative analysis of 65 written responses to open-ended items included concerns about the roles of physicians and nurses, implementation, and impact on ICU team–family relationships. Conclusions: Integration of palliative care specialists in the ICU is broadly acceptable and desired. However, the most commonly used current triggers for prompting specialist consultation were among the least well accepted, while more favorable triggers are difficult to abstract from electronic health record systems. There is also disagreement about the role of ICU nurses in palliative care delivery. These findings provide important guidance to the development of collaborative care models for the ICU setting.","{'model': 'tldr@v2.0.0', 'text': None}",https://europepmc.org/articles/pmc5344777?pdf=render
-the varied circumstances prompting requests for emergency contraception at school-based clinics.,A. Sidebottom,"BACKGROUND Little is known about the circumstances that prompt teenagers to request emergency contraception (EC). This evaluation was designed to refine the EC clinical protocol and improve pregnancy prevention efforts in high school-based clinics by analyzing information on EC use and subsequent contraception use of EC patients. METHODS Sites included 5 clinics located at mainstream, racially diverse, and urban high schools. Nursing staff documented all EC-related visits during the 2002-2003 school year on a standardized form. These forms and additional information were collected from chart reviews. RESULTS EC was requested in 113 instances involving 91 students and dispensed in all but 4 instances. The most frequently reported circumstance prompting an EC request was the use of no protection (37.2%), followed by a condom mishap (27.4%) and questionable protection from a hormonal method (23.9%). Anxiety despite adequate protection from a hormonal method (9.7%) or a condom (1.8%) accounted for the remaining cases. Although the clinical protocol included a follow-up appointment 2 weeks after receipt of EC, 39.5% of EC users failed to keep this appointment. The chart review revealed that hormonal contraception was eventually initiated following two thirds (68.0%) of the EC instances that involved either no protection or only a condom. CONCLUSIONS Because young EC seekers vary considerably in terms of sexual experience, contraceptive use consistency, and pregnancy risk, individualized risk assessment and counseling are essential. The EC visit may be an optimal time to encourage initiation or reinstitution of hormonal methods of birth control.","{'model': 'tldr@v2.0.0', 'text': 'The EC visit may be an optimal time to encourage initiation or reinstitution of hormonal methods of birth control and individualized risk assessment and counseling are essential.'}",
-change detection in coral reef environment using high-resolution images: comparison of object-based and pixel-based paradigms,Zhenjin Zhou,"Despite increases in the spatial resolution of satellite imagery prompting interest in object-based image analysis, few studies have used object-based methods for monitoring changes in coral reefs. This study proposes a high accuracy object-based change detection (OBCD) method intended for coral reef environment, which uses QuickBird and WorldView-2 images. The proposed methodological framework includes image fusion, multi-temporal image segmentation, image differencing, random forests models, and object-area-based accuracy assessment. For validation, we applied the method to images of four coral reef study sites in the South China Sea. We compared the proposed OBCD method with a conventional pixel-based change detection (PBCD) method by implementing both methods under the same conditions. The average overall accuracy of OBCD exceeded 90%, which was approximately 20% higher than PBCD. The OBCD method was free from salt-and-pepper effects and was less prone to images misregistration in terms of change detection accuracy and mapping results. The object-area-based accuracy assessment reached a higher overall accuracy and per-class accuracy than the object-number-based and pixel-number-based accuracy assessment.","{'model': 'tldr@v2.0.0', 'text': 'The proposed OBCD method was free from salt-and-pepper effects and was less prone to images misregistration in terms of change detection accuracy and mapping results, and reached a higher overall accuracy and per-class accuracy than the object- number-based and pixel-number-based accuracy assessment.'}",https://www.mdpi.com/2220-9964/7/11/441/pdf?version=1542019325
-impact of an encounter-based prompting system on resident vaccine administration performance and immunization knowledge.,J. Shaw,"OBJECTIVES To evaluate an encounter-based immunization prompting system on resident performance in administering vaccines and knowledge of immunization guidelines. DESIGN/METHODS Prospective randomized, controlled trial. Subjects were first- and second-year pediatric residents in a hospital-based continuity clinic. The intervention group received manual prompts of immunizations due. Postclinic chart review compared immunizations due with those administered. Acceptable and unacceptable reasons for not administering vaccines were assigned. Resident knowledge was measured by a 70-item examination. RESULTS The intervention group had significantly less missed opportunities/vaccine administration errors (11.4% vs 21.6%). The most common reason for unacceptable errors in the intervention group: vaccine was given too early; in the control group: vaccine was postponed to next visit. Pre- and postintervention knowledge scores were similar: intervention group (75.5% vs 80.7%, control group; 76.5% vs 81.3%). CONCLUSION An immunization prompting system in a hospital-based pediatric resident continuity clinic reduced missed opportunities/vaccine administration errors without significantly impacting resident knowledge of immunization guidelines.immunization schedule, vaccination, immunization, prompting systems, resident education.","{'model': 'tldr@v2.0.0', 'text': 'An immunization prompting system in a hospital-based pediatric resident continuity clinic reduced missed opportunities/vaccine administration errors without significantly impacting resident knowledge of immunization guidelines.'}",
-an ipad™-based picture and video activity schedule increases community shopping skills of a young adult with autism spectrum disorder and intellectual disability,Elizabeth Burckley,"Abstract Objective: To evaluate the iPad 2™ with Book Creator™ software to provide visual cues and video prompting to teach shopping skills in the community to a young adult with an autism spectrum disorder and intellectual disability. Methods: A multiple probe across settings design was used to assess effects of the intervention on the participant’s independence with following a shopping list in a grocery store across three community locations. Results: Visual cues and video prompting substantially increased the participant’s shopping skills within two of the three community locations, skill increases maintained after the intervention was withdrawn, and shopping skills generalized to two untaught shopping items. Social validity surveys suggested that the participant’s parent and staff favorably viewed the goals, procedures, and outcomes of intervention. Conclusions: The iPad 2™ with Book Creator™ software may be an effective way to teach independent shopping skills in the community; additional replications are needed.","{'model': 'tldr@v2.0.0', 'text': 'The iPad 2™ with Book Creator™ software may be an effective way to teach independent shopping skills in the community; additional replications are needed.'}",
-capnographic monitoring of propofol-based sedation during colonoscopy,M. Friedrich-Rust,"Background and study aims: Capnography enables the measurement of end-tidal CO2 and thereby the early detection of apnea, prompting immediate intervention to restore ventilation. Studies have shown that capnographic monitoring is associated with a reduction of hypoxemia during sedation for endoscopy and early detection of apnea during sedation for colonoscopy. The primary aim of this prospective randomized study was to evaluate whether capnographic monitoring without tracheal intubation reduces hypoxemia during propofol-based sedation in patients undergoing colonoscopy. Patients and methods: A total of 533 patients presenting for colonoscopy at two study sites were randomized to either standard monitoring (n = 266) or to standard monitoring with capnography (n = 267). The incidence of hypoxemia (SO2 < 90 %) and severe hypoxemia (SO2 < 85 %) were compared between the groups. Furthermore, risk factors for hypoxemia were evaluated, and sedation performed by anesthesiologists was compared with nurse-administered propofol sedation (NAPS) or endoscopist-directed sedation (EDS). Results: The incidence of hypoxemia was significantly lower in patients with capnography monitoring compared with those receiving standard monitoring (18 % vs. 32 %; P = 0.00091). Independent risk factors for hypoxemia were age (P = 0.00015), high body mass index (P = 0.0044), history of sleep apnea (P = 0.025), standard monitoring group (P = 0.000069), total dose of propofol (P = 0.031), and dose of ketamine (P < 0.000001). Patients receiving anesthesiologist-administered sedation developed hypoxemic events more often than those receiving NAPS or EDS. In patients with anesthesiologist-administered sedation, sedation was deeper, a combination of sedative medication (propofol, midazolam and/or ketamine) was administered significantly more often, and sedative doses were significantly higher compared with patients receiving NAPS or EDS. Conclusions: In patients undergoing colonoscopy during propofol-based sedation capnography monitoring with a simple and inexpensive device reduced the incidence of hypoxemia.","{'model': 'tldr@v2.0.0', 'text': 'In patients undergoing colonoscopy during propofol-based sedation capnography monitoring with a simple and inexpensive device reduced the incidence of hypoxemia, particularly in patients with anesthesiologist-administered sedation.'}",
-"social media monitoring: aims, methods, and challenges for international companies",Boyang Zhang,"Purpose – The purpose of this paper is to clarify the aims, monitoring methods and challenges of social media monitoring from the perspective of international companies. Trends in the literature are also investigated. Design/methodology/approach – Based on a systematic literature review, 30 key articles from 2008 to 2012 were further analysed. Findings – International companies need real-time monitoring software, expertise and dynamic visualization to facilitate early detection and prognoses supporting strategy making. This is a costly affair, prompting questions about return on investment. A recent trend in the research literature concerns the development of models describing how issues spread in social media with the aim of facilitating prognoses. Research limitations/implications – The online databases used comprised refereed peer-reviewed scientific articles. Books were not included in the search process. Practical implications – Because information spreads fast in social media and affects internation...",,https://jyx.jyu.fi/bitstream/123456789/44532/1/zhangvos2014socialmediamonitoringinternatcompanies.pdf
-using an ipad-based video modelling package to teach numeracy skills to a child with an autism spectrum disorder,Elizabeth Jowett,"Purpose: To evaluate the effectiveness of a video modelling package to teach a 5 year-old boy diagnosed with an autism spectrum disorder (ASD) basic numeracy skills. The treatment package consisted of iPad-based video modelling, gradual fading of video prompts, reinforcement, in vivo prompting and forward chaining. Methods: A single subject multiple baseline across numerals design with generalization and maintenance observational measures. Results: Clear gains were evident in the participant's ability to identify and write the Arabic numerals 1–7 and comprehend the quantity each numeral represents in association with the lagged intervention. Generalization and maintenance data demonstrated the robustness of the treatment effects. Conclusions: IPad-based video modelling, when used in a package, can be an effective technique for teaching numeracy skills to children with an ASD. Systematic replication of this study with different participants is warranted.","{'model': 'tldr@v2.0.0', 'text': 'IPad-based video modelling, when used in a package, can be an effective technique for teaching numeracy skills to children with an ASD.'}",
-"whole organism high-content screening by label-free, image-based bayesian classification for parasitic diseases",R. Paveley,"Sole reliance on one drug, Praziquantel, for treatment and control of schistosomiasis raises concerns about development of widespread resistance, prompting renewed interest in the discovery of new anthelmintics. To discover new leads we designed an automated label-free, high content-based, high throughput screen (HTS) to assess drug-induced effects on in vitro cultured larvae (schistosomula) using bright-field imaging. Automatic image analysis and Bayesian prediction models define morphological damage, hit/non-hit prediction and larval phenotype characterization. Motility was also assessed from time-lapse images. In screening a 10,041 compound library the HTS correctly detected 99.8% of the hits scored visually. A proportion of these larval hits were also active in an adult worm ex-vivo screen and are the subject of ongoing studies. The method allows, for the first time, screening of large compound collections against schistosomes and the methods are adaptable to other whole organism and cell-based screening by morphology and motility phenotyping.","{'model': 'tldr@v2.0.0', 'text': 'An automated label-free, high content-based, high throughput screen (HTS) to assess drug-induced effects on in vitro cultured larvae (schistosomula) using bright-field imaging and the methods are adaptable to other whole organism and cell-based screening by morphology and motility phenotyping.'}",https://journals.plos.org/plosntds/article/file?id=10.1371/journal.pntd.0001762&type=printable
-a review on electromyography decoding and pattern recognition for human-machine interaction,M. Simão,"This paper presents a literature review on pattern recognition of electromyography (EMG) signals and its applications. The EMG technology is introduced and the most relevant aspects for the design of an EMG-based system are highlighted, including signal acquisition and filtering. EMG-based systems have been used with relative success to control upper- and lower-limb prostheses, electronic devices and machines, and for monitoring human behavior. Nevertheless, the existing systems are still inadequate and are often abandoned by their users, prompting for further research. Besides controlling prostheses, EMG technology is also beneficial for the development of machine learning-based devices that can capture the intention of able-bodied users by detecting their gestures, opening the way for new human-machine interaction (HMI) modalities. This paper also reviews the current feature extraction techniques, including signal processing and data dimensionality reduction. Novel classification methods and approaches for detecting non-trained gestures are discussed. Finally, current applications are reviewed, through the comparison of different EMG systems and discussion of their advantages and drawbacks.","{'model': 'tldr@v2.0.0', 'text': 'The EMG technology is introduced and the most relevant aspects for the design of an EMG-based system are highlighted, including signal acquisition and filtering, and the current feature extraction techniques, including Signal processing and data dimensionality reduction are reviewed.'}",https://ieeexplore.ieee.org/ielx7/6287639/8600701/08672131.pdf
-evaluation of the accelerate pheno system for fast identification and antimicrobial susceptibility testing from positive blood cultures in bloodstream infections caused by gram-negative pathogens,M. Marschal,"ABSTRACT Bloodstream infections (BSI) are an important cause of morbidity and mortality. Increasing rates of antimicrobial-resistant pathogens limit treatment options, prompting an empirical use of broad-range antibiotics. Fast and reliable diagnostic tools are needed to provide adequate therapy in a timely manner and to enable a de-escalation of treatment. The Accelerate Pheno system (Accelerate Diagnostics, USA) is a fully automated test system that performs both identification and antimicrobial susceptibility testing (AST) directly from positive blood cultures within approximately 7 h. In total, 115 episodes of BSI with Gram-negative bacteria were included in our study and compared to conventional culture-based methods. The Accelerate Pheno system correctly identified 88.7% (102 of 115) of all BSI episodes and 97.1% (102 of 105) of isolates that are covered by the system's identification panel. The Accelerate Pheno system generated an AST result for 91.3% (95 of 104) samples in which the Accelerate Pheno system identified a Gram-negative pathogen. The overall category agreement between the Accelerate Pheno system and culture-based AST was 96.4%, the rates for minor discrepancies 1.4%, major discrepancies 2.3%, and very major discrepancies 1.0%. Of note, ceftriaxone, piperacillin-tazobactam, and carbapenem resistance was correctly detected in blood culture specimens with extended-spectrum beta-lactamase-producing Escherichia coli (n = 7) and multidrug-resistant Pseudomonas aeruginosa (n = 3) strains. The utilization of the Accelerate Pheno system reduced the time to result for identification by 27.49 h (P < 0.0001) and for AST by 40.39 h (P < 0.0001) compared to culture-based methods in our laboratory setting. In conclusion, the Accelerate Pheno system provided fast, reliable results while significantly improving turnaround time in blood culture diagnostics of Gram-negative BSI.","{'model': 'tldr@v2.0.0', 'text': 'The Accelerate Pheno system provided fast, reliable results while significantly improving turnaround time in blood culture diagnostics of Gram-negative BSI and compared to conventional culture-based methods in the laboratory setting.'}",https://jcm.asm.org/content/jcm/55/7/2116.full.pdf
-unrecognized physical illness prompting psychiatric admission.,R. Reeves,"BACKGROUND We assessed factors that might contribute to clinicians erroneously attributing medically based changes in a patient's mental status to psychiatric illness. METHODS Records of 1340 patients admitted to a VA hospital psychiatric unit and 613 to a public hospital psychiatric unit from 2001 to 2007 were reviewed. Cases admitted because of an unrecognized medical disorder underwent further analysis of the preadmission assessment and documented history of mental illness. RESULTS Of 1340 patients whose records were reviewed, 55 (2.8%) had a medical disorder that caused their symptoms. Compared with patients admitted to medical units, patients inappropriately admitted to psychiatric units had lower rates of completion of medical histories, physical examinations, cognitive assessments, indicated laboratory and/or radiologic studies, and treatment of abnormal vital signs (P < .001 in each case). Among patients admitted to psychiatric units, 85.5% had a history of mental illness vs 30.9% of comparable admissions to medical units (Chi2(1) = 35.85; P < .001). CONCLUSIONS Key assessment procedures are less likely to be performed in patients with mental status changes who are admitted to psychiatric units than in comparable patients admitted to medical units. Symptoms of patients with a history of mental illness are more likely to be attributed to psychiatric illness than are those of patients without such a history.","{'model': 'tldr@v2.0.0', 'text': 'Key assessment procedures are less likely to be performed in patients with mental status changes who are admitted to psychiatric units than in comparable patients admitted to medical units.'}",
-field-based development of an information support robot for persons with dementia,Takenobu Inoue,"Background: The use of robots for providing support to persons with dementia is very promising. However, it is difficult to develop a truly useful robot system because of the difficulty of clarifying their needs and opinions. Objective: Because a field-based method is effective for the development of these kinds of systems, in this study, we developed a prototype of an information support robot for persons with dementia, using field-based methodology. Methods: A communication robot system produced by NEC Corporation, called ""PaPeRo"", was chosen as the platform for this system. To achieve the goal of keeping persons with dementia informed of their daily schedule and prompting them to take desired actions, interactive verbal communication algorithms were programmed into the robot system. Results: The results of experiments with five dementia sufferers showed that this system produced an information acquisition rate of over 90%. In addition, a life support demonstration showed the possibility of prompting users to perform actions. Conclusions: These results suggest the possibility of using this robot to support independent living by persons with dementia, and also the effectiveness of the field-based method.","{'model': 'tldr@v2.0.0', 'text': 'A prototype of an information support robot for persons with dementia is developed, using field-based methodology, and the results suggest the possibility of using this robot to support independent living by people with dementia, and also the effectiveness of the field- based method.'}",https://content.iospress.com:443/download/technology-and-disability/tad00357?id=technology-and-disability%2Ftad00357
-accuracy of 25-hydroxyvitamin d assays: confronting the issues.,G. Carter,"Measurement of 25-hydroxyvitamin D (25-OHD) is widely used for assessing vitamin D status. There has been a dramatic increase in 25-OHD requests over recent years prompting many laboratories to consider the use of automated immunoassays. To achieve higher throughput, these methods have abandoned the traditional solvent extraction of samples and are therefore more prone to non-specific interference. The Vitamin D External Quality Assessment Scheme (DEQAS) has revealed method-related differences in 25-OHD results, raising concerns about the comparability and accuracy of different assays. This paper highlights some of the pre-analytical, analytical and post-analytical issues which may influence the accuracy of 25-OHD assays and interpretation of results. Recent attention has focused on reconciling the relatively high results given by liquid chromatography-tandem mass spectrometry (LC-MS/MS) to those of the DiaSorin radioimmunoassay (RIA) on which clinical decision points have previously been based. Data is presented on 20 DEQAS samples which were analysed by an LC-MS/MS assay developed as a candidate reference measurement procedure by the US National Institute of Standards and Technology (NIST). The NIST results were on average 11.2% lower than those given by routine LC-MS/MS methods. If confirmed, these results suggest that most routine LC-MS/MS assays are perhaps overestimating 25-OHD by failing to resolve a molecule having the same mass as 25-OHD(3) and a similar fragmentation pattern. All 25-OHD assays should be monitored by a proficiency testing scheme and the results made available to clinicians and editors of scientific journals.","{'model': 'tldr@v2.0.0', 'text': 'If confirmed, these results suggest that most routine LC-MS/MS assays are perhaps overestimating 25-OHD by failing to resolve a molecule having the same mass as 25- OHD(3) and a similar fragmentation pattern.'}",
-"active, population-based surveillance for severe rotavirus gastroenteritis in children in the united states",D. Payne,"OBJECTIVES. Routine vaccination of US infants against rotavirus was implemented in 2006, prompting the Centers for Disease Control and Prevention New Vaccine Surveillance Network to begin population-based acute gastroenteritis surveillance among US children <3 years of age. This surveillance system establishes baseline estimates of rotavirus disease burden and allows for the prospective monitoring of rotavirus vaccination impact. METHODS. Eligible children with acute gastroenteritis (≥3 episodes of diarrhea and/or any vomiting in a 24-hour period) who were hospitalized, were seen in emergency departments, or visited selected outpatient clinics in 3 US counties during the period of January through June 2006 were enrolled. Epidemiological and clinical information was obtained through parental interview and medical chart review, and stool specimens were tested for rotavirus with enzyme immunoassays. Rotavirus-positive specimens were genotyped by using reverse transcription-polymerase chain reaction assays. RESULTS. Stool specimens were collected from 516 of the 739 enrolled children with acute gastroenteritis (181 inpatient, 201 emergency department, and 134 outpatient) and 44% tested positive for rotavirus (227 of 516 specimens). The most common strain was P[8]G1 (84%), followed by P[4]G2 (5%) and P[6]G12 (4%). None of the 516 children had received rotavirus vaccine. The rotavirus detection rate was 50% for hospitalized acute gastroenteritis cases, 50% for emergency department visits, and 27% for outpatient visits. Rotavirus-related acute gastroenteritis cases were more likely than non–rotavirus-related acute gastroenteritis cases to present with vomiting, diarrhea, fever, and lethargy. Directly calculated, population-based rates for rotavirus hospitalizations and emergency department visits were 22.5 hospitalizations and 301.0 emergency department visits per 10 000 children <3 years of age, respectively. A sentinel outpatient clinic visit rate of 311.9 outpatient visits per 10 000 children <3 years of age was observed. CONCLUSIONS. Population-based, laboratory-confirmed rotavirus surveillance in the final rotavirus season before implementation of the US rotavirus vaccine program indicated a considerable burden of disease on the US health care system.","{'model': 'tldr@v2.0.0', 'text': None}",
-wavelet methods for spike detection in mouse renal sympathetic nerve activity,R. Brychta,"Abnormal autonomic nerve traffic has been associated with a number of peripheral neuropathies and cardiovascular disorders prompting the development of genetically altered mice to study the genetic and molecular components of these diseases. Autonomic function in mice can be assessed by directly recording sympathetic nerve activity. However, murine sympathetic spikes are typically detected using a manually adjusted voltage threshold and no unsupervised detection methods have been developed for the mouse. Therefore, we tested the performance of several unsupervised spike detection algorithms on simulated murine renal sympathetic nerve recordings, including an automated amplitude discriminator and wavelet-based detection methods which used both the discrete wavelet transform (DWT) and the stationary wavelet transform (SWT) and several wavelet threshold rules. The parameters of the wavelet methods were optimized by comparing basal sympathetic activity to postmortem recordings and recordings made during pharmacological suppression and enhancement of sympathetic activity. In general, SWT methods were found to outperform amplitude discriminators and DWT methods with similar wavelet coefficient thresholding algorithms when presented with simulations with varied mean spike rates and signal-to-noise ratios. A SWT method which estimates the noise level using a ""noise-only"" wavelet scale and then selectively thresholds scales containing the physiologically important signal information was found to have the most robust spike detection. The proposed noise-level estimation method was also successfully validated during pharmacological interventions","{'model': 'tldr@v2.0.0', 'text': 'ASWT methods were found to outperform amplitude discriminators and DWT methods with similar wavelet coefficient thresholding algorithms when presented with simulations with varied mean spike rates and signal-to-noise ratios and the proposed noise-level estimation method was successfully validated during pharmacological interventions.'}",https://europepmc.org/articles/pmc2075098?pdf=render
-effect of human leukocyte antigen compatibility on kidney graft survival: comparative analysis of two decades,G. Opelz,"Background. Based on an analysis of United Network for Organ Sharing data, it was reported that the influence of human leukocyte antigen (HLA) matching in renal transplantation has diminished in recent years, prompting the suggestion that donor kidney allocation algorithms should be revised. Methods. We compared the impact of HLA matching on kidney graft survival during the decades 1985–1994 and 1995–2004 using the data of the Collaborative Transplant Study. Results for the last 5 years (2000–2004) were analyzed separately in addition. Multivariate Cox regression analysis was used to account for the influence of confounders. Results. Our results show that, while graft survival rates have improved overall over time, the relative impact of HLA matching on the graft survival rate has remained strong and highly significant. Both the need for posttransplant rejection treatment and the graft survival rates showed statistically highly significant associations with HLA matching regardless of the interval analyzed (P<0.001). Conclusions. We conclude that HLA mismatches significantly influence the outcome of kidney transplants and that kidney exchange programs for the purpose of achieving better HLA matches continue to be meaningful.","{'model': 'tldr@v2.0.0', 'text': 'It is concluded that HLA mismatches significantly influence the outcome of kidney transplants and that kidney exchange programs for the purpose of achieving better HLA matches continue to be meaningful.'}",
-"enhancing instruction through constructivism, cooperative learning, and cloud computing",David W. Denton,,"{'model': 'tldr@v2.0.0', 'text': 'Strategies for integrating cloud-based applications are suggested and results from a case study involving graduate education students are presented.'}",
-update on negative-pressure wound therapy,D. Orgill,"Background: Over the last 15 years, negative-pressure wound therapy has become commonly used for treatment of a wide variety of complex wounds. There are now several systems marketed, and additional products will be available in the near future. Many clinicians have noted a dramatic response when negative-pressure wound therapy technology has been used, prompting a number of scientific investigations related to its mechanism of action and clinical trials determining its efficacy. Methods: The peer-reviewed literature within the past 5 years was reviewed, using an evidence-based approach. Results: Negative-pressure wound therapy works through mechanisms that include fluid removal, drawing the wound together, microdeformation, and moist wound healing. Several randomized clinical trials support the use of negative-pressure wound therapy in certain wound types. Serious complications, including bleeding and infection, have recently been reported by the U.S. Food and Drug Administration in a small number of patients. Conclusions: Negative-pressure wound therapy has dramatically changed the way complex wounds are treated. The rapid introduction of this technology has occurred faster than large-scale randomized controlled studies or registry studies have been conducted. Further clinical studies and basic science studies will help surgeons to better understand the evidence and use this technology in the future.","{'model': 'tldr@v2.0.0', 'text': 'The rapid introduction of this technology has occurred faster than large-scale randomized controlled studies or registry studies have been conducted and will help surgeons to better understand the evidence and use this technology in the future.'}",
-prompting clinicians: a systematic review of preventive care reminders,J. Dexheimer,"Prompting clinicians to offer preventive care procedures has been shown to increase the use of these procedures. This study is an update of a systematic review examining the effect of reminder systems on offers of preventive care to patients. Of 1,404 eligible studies, 23 were included. The studies were evaluated according to their intervention type and use of computerized methods. We found that although computerized reminder systems have become more common, paper-based reminders were the most effective reminder strategy.","{'model': 'tldr@v2.0.0', 'text': 'It was found that although computerized reminder systems have become more common, paper-based reminders were the most effective reminder strategy.'}",
-energy supply chain optimization of hybrid feedstock processes: a review.,J. Elia,"The economic, environmental, and social performances of energy systems depend on their geographical locations and the surrounding market infrastructure for feedstocks and energy products. Strategic decisions to locate energy conversion facilities must take all upstream and downstream operations into account, prompting the development of supply chain modeling and optimization methods. This article reviews the contributions of energy supply chain studies that include heat, power, and liquid fuels production. Studies are categorized based on specific features of the mathematical model, highlighting those that address energy supply chain models with and without considerations of multiperiod decisions. Studies that incorporate uncertainties are discussed, and opportunities for future research developments are outlined.","{'model': 'tldr@v2.0.0', 'text': 'The contributions of energy supply chain studies that include heat, power, and liquid fuels production are reviewed, highlighting those that address energy supply network models with and without considerations of multiperiod decisions.'}",https://www.annualreviews.org/doi/pdf/10.1146/annurev-chembioeng-060713-040425
-"repetitive daily point of choice prompts and occupational sit-stand transfers, concentration and neuromuscular performance in office workers: an rct",L. Donath,"Objective: Prolonged office sitting time adversely affects neuromuscular and cardiovascular health parameters. As a consequence, the present study investigated the effects of prompting the use of height-adjustable working desk (HAWD) on occupational sitting and standing time, neuromuscular outcomes and concentration in office workers. Methods: A single-blinded randomized controlled trial (RCT) with parallel group design was conducted. Thirty-eight office workers were supplied with HAWDs and randomly assigned (Strata: physical activity (PA), BMI, gender, workload) to a prompt (INT) or non-prompt (CON) group. INT received three daily screen-based prompts within 12 weeks. CON was only instructed once concerning the benefits of using HAWDs prior to the start of the study. Sitting and standing times were objectively assessed as primary outcomes for one entire working week using the ActiGraph wGT3X-BT at baseline (pre), after 6 (mid) and 12 weeks (post). Concentration (d2-test), postural sway during upright stance (under single, dual and triple task) and lower limb strength endurance (heel-rise) were collected as secondary outcomes. Results: With large but not statistically significant within group effects from pre to post, INT increased weekly standing time at work by 9% (p = 0.22, d = 0.8) representing an increase from 7.2 h (4.8) to 9.7 (6.6) h (p = 0.07). Concentration and neuromuscular performance did not change from pre to post testing (0.23 < p < 0.95; 0.001 < ηp² < 0.05). Conclusion: Low-frequent and low cost screen-based point of choice prompts (3 per day within 12 weeks) already result in notable increases of occupational standing time of approx. daily 30 min. These stimuli, however, did not relevantly affect neuromuscular outcomes.","{'model': 'tldr@v2.0.0', 'text': 'Low-frequent and low cost screen-based point of choice prompts (3 per day within 12 weeks) already result in notable increases of occupational standing time of approx.'}",https://www.mdpi.com/1660-4601/12/4/4340/pdf?version=1429534409
-training skills in the psychiatrically disabled: learning coping and competence.,R. Liberman,"Social skills training methods represent a major strategy for psychiatric rehabilitation. Building skills in patients with schizophrenic and other major mental disorders is based on the assumption that coping and competence can override stress and vulnerability in reducing relapses and improving psychosocial functioning. For maximum efficiency, skills training needs to incorporate procedures and principles of human learning and information processing. Several models for skills training have been designed and evaluated, each of which has proved to be effective in raising the social competence of chronic mental patients. The ""basic"" model involves role playing by the patient and modeling, prompting, feedback, and reinforcement by the therapist. A ""problem-solving"" model of training provides general strategies for dealing with a wide variety of social situations. This model uses role playing to enhance behavioral performance but also highlights the patient's abilities to perceive and process incoming social messages and meanings. It is essential that social skills training be imbedded in a comprehensive program of rehabilitation that features continuity of care, supportive community services, therapeutic relationships, and judicious prescription of psychotropic drugs.","{'model': 'tldr@v2.0.0', 'text': 'It is essential that social skills training be imbedded in a comprehensive program of rehabilitation that features continuity of care, supportive community services, therapeutic relationships, and judicious prescription of psychotropic drugs.'}",https://academic.oup.com/schizophreniabulletin/article-pdf/12/4/631/5302085/12-4-631.pdf
-adaptive algebraic multigrid,M. Brezina,"Efficient numerical simulation of physical processes is constrained by our ability to solve the resulting linear systems, prompting substantial research into the development of multiscale iterative methods capable of solving these linear systems with an optimal amount of effort. Overcoming the limitations of geometric multigrid methods to simple geometries and differential equations, algebraic multigrid methods construct the multigrid hierarchy based only on the given matrix. While this allows for efficient black-box solution of the linear systems associated with discretizations of many elliptic differential equations, it also results in a lack of robustness due to unsatisfied assumptions made on the near null spaces of these matrices. This paper introduces an extension to algebraic multigrid methods that removes the need to make such assumptions by utilizing an adaptive process. Emphasis is on the principles that guide the adaptivity and their application to algebraic multigrid solution of certain symmetric positive-definite linear systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an extension to algebraic multigrid methods that removes the need to make unsatisfied assumptions made on the near null spaces of these matrices by utilizing an adaptive process.'}",
-current and emerging commercial optical biosensors,Cheryl L. Baird,"The field of commercial optical biosensors is rapidly evolving, with new systems and detection methods being developed each year. This review outlines the currently available biosensor hardware and highlights unique features of each platform. Affinity‐based biosensor technology, with its high sensitivity, wide versatility and high throughput, is playing a significant role in basic research, pharmaceutical development, and the food and environmental sciences. Likewise, the increasing popularity of biosensors is prompting manufacturers to develop new instrumentation for dedicated applications. We provide a preview of some of the emerging commercial systems that are dedicated to drug discovery, proteomics, clinical diagnostics and routine biomolecular interaction analysis. Copyright © 2001 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'This review outlines the currently available biosensor hardware and highlights unique features of each platform, providing a preview of some of the emerging commercial systems that are dedicated to drug discovery, proteomics, clinical diagnostics and routine biomolecular interaction analysis.'}",
-improving training methods in brain injury rehabilitation,M. Mozzoni,"The effects of therapist training techniques upon treatment outcomes in traumatic brain injury (TBI) has not been widely studied. This study demonstrates that outcome scores are likely to increase when therapists work on outcome-related activities and use identified training methods. Six therapists working with five persons with TBI were subjects for this study. Therapists were targeted because their clients were not making progress on the functional independent measurement (FIM). Baseline teaching data were collected from videotapes, targeting 14 training elements within the areas of task, prompting, reinforcement, and client's behavior. Therapists were given feedback based upon the teaching diagnostic of the decision matrix. The results show that diagnostic-based intervention with therapists can result in direct, immediate, and obvious improvements in outcome measures across patients.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that outcome scores are likely to increase when therapists work on outcome-related activities and use identified training methods and diagnostic-based intervention with therapists can result in direct, immediate, and obvious improvements in outcome measures across patients.'}",
-review of studies with students with significant cognitive disabilities which link to science standards,Ginevra Courtade,"A comprehensive review of research was conducted on teaching science to students with significant cognitive disabilities. Guidelines from the National Science Education Standards were used to identify categories of studies based on the strands of science. A total of 11 studies were identified from the 20 years of literature searched. In general, work in the area of science is sparse for students with significant cognitive disabilities. Eight of the 11 studies fell in Content Standard F: Science in Personal and Social Perspectives. Single subject experimental designs were the methodology used in all of the studies and outcomes suggest that this population benefits from instruction in highly specific skills with modeling and errorless learning strategies like time delay. The interventions found in the 11 studies that linked to science incorporated systematic response prompting methods similar to those found in evidence-based reading and math research for students with significant cognitive disabilities. Additional research is needed to identify methods to teach science to this population that includes both broader content and additional skills that link to state standards.",,
-is the mussel test a good indicator of antifouling activity? a comparison between laboratory and field assays,B. D. da Gama,"Current antifouling technologies rely on metal-based paints, but due to their toxicity, an expected worldwide ban of organotin-containing paints is now prompting the quest for safe and effective alternatives. One of these is antifouling coatings whose active components are naturally occurring compounds in marine organisms. A number of laboratory bioassays has been designed to search for antifouling compounds. However, there is no evidence to date that these assays provide results reproducible through ecologically realistic field experiments. Natural concentrations of the extracts from the Brazilian seaweeds Laurencia obtusa and Stypopodium zonale were tested in the laboratory through the 'mussel test' and in the field through the 'phytagel method' in order to compare the efficiency of these methods in assessing antifouling activity. L. obtusa extract significantly inhibited fouling in both the laboratory and field assays, while S. zonale stimulated fouling in both assays. Major compounds from the extracts were identified. The findings suggest that the 'mussel test' is a reliable time and cost-saving screening method for antifouling substances, although field assays are more sensitive for detection of their activity spectrum.","{'model': 'tldr@v2.0.0', 'text': ""The findings suggest that the 'mussel test' is a reliable time and cost-saving screening method for antifouling substances, although field assays are more sensitive for detection of their activity spectrum.""}",
-the use of celebrity athletes as endorsers: views of the new zealand general public,J. Charbonneau,"Matching celebrity athletes with potential endorsement opportunities is often difficult. Yet there are easy-touse survey-based methods available. Based on a survey of the general public in New Zealand, this study uses both Ohanian's source-credibility scale and a constant-sum scale to help brand managers, player agents and advertising practitioners select good celebrity athlete-product fit. Four New Zealand athletes (two males and two females) and several products were included in the test. Results show that the female celebrity athletes outperformed their male counterparts as potential endorsers. Use of Ohanian's multi-attribute scale yields a level of richness and insight, prompting us to advocate the use of both scales in the pursuit of endorser-product congruences.",,https://researchcommons.waikato.ac.nz/bitstream/10289/5124/2/Garland%20The%20use%20of.pdf
-expected value prioritization of prompts and reminders,S. Downs,"Computer-based prompting and reminder systems have been shown to be highly effective in increasing rates of preventive services delivery. However, there are many more recommended preventive services than can be practically included in a typical clinic visit. Therefore prioritization of preventive services prompts is necessary. We describe two approaches to prioritizing preventive services prompts based on expected value decision making. One method involves a static, global prioritization across all preventive services and has been used in a production system for almost 7 years. The second method uses influence diagrams to prioritize prompts dynamically, based on individual patient data. The latter approach is still under development. Both methods are labor intensive and require a combination of epidemiologic data and expert judgment. Compromises in strictly normative process were necessary to achieve user satisfaction.","{'model': 'tldr@v2.0.0', 'text': 'Two approaches to prioritizing preventive services prompts based on expected value decision making are described, one of which involves a static, global prioritization across all preventive services and has been used in a production system for almost 7 years, and the second which uses influence diagrams to prioritize prompts dynamically, based on individual patient data.'}",
-exploring recall of physical activity in young people using qualitative interviewing,J. McKenna,"Recall of physical activity is a known problem affecting all forms of self-report. Participants age 8?16 years contributed to 16 focus groups and 24 interviews based on cognitive interviewing (n = 8) and think-aloud (n = 8) and general probing (n = 8) techniques. When unassisted, participants readily described physical activity mode but gave vague descriptions of daily activities. In contrast, the close detail of frequency, intensity, and duration of these activities was only more fully developed through prompting. Talk-based methods can provide considerable insight into developing more reliable and valid physical activity self-reports","{'model': 'tldr@v2.0.0', 'text': 'Talk-based methods can provide considerable insight into developing more reliable and valid physical activity self-reports.'}",
-a prompting-based approach for adversarial example generation and robustness enhancement,Yuting Yang,"Recent years have seen the wide application of NLP models in crucial areas such as finance, medical treatment, and news media, raising concerns of the model robustness and vulnerabilities. In this paper, we propose a novel prompt-based adversarial attack to compromise NLP models and robustness enhancement technique. We first construct malicious prompts for each instance and generate adversarial examples via mask-and-filling under the effect of a malicious purpose. Our attack technique targets the inherent vulnerabilities of NLP models, allowing us to generate samples even without interacting with the victim NLP model, as long as it is based on pre-trained language models (PLMs). Furthermore, we design a prompt-based adversarial training method to improve the robustness of PLMs. As our training method does not actually generate adversarial samples, it can be applied to large-scale training sets efficiently. The experimental results show that our attack method can achieve a high attack success rate with more diverse, fluent and natural adversarial examples. In addition, our robustness enhancement method can significantly improve the robustness of models to resist adversarial attacks. Our work indicates that prompting paradigm has great potential in probing some fundamental flaws of PLMs and fine-tuning them for downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based adversarial attack to compromise NLP models and robustness enhancement technique that can significantly improve the robustness of models to resist adversarial attacks and indicates that prompting paradigm has great potential in probing some fundamental flaws of PLMs and fine-tuning them for downstream tasks.'}",http://arxiv.org/pdf/2203.10714
-kemampuan representasi matematis siswa melalui model problem based learning dan probing prompting learning,Suwanti Suwanti,"The students' mathematical representation ability is low. Efforts to improve mathematical representation skills include innovating learning models. The purpose of the study was to analyze the differences and quality of improvement in mathematical representation skills between students who received the problem-based learning model with probing prompting learning, along with student attitude responses. The research method is quasi-experimental. The population is all class VIII students of one of the private junior high schools in Garut for the 2018/2019 academic year, the samples are class VIII-A and VIII-B. The results showed that there were differences in the ability of mathematical representation between students who received the problem-based learning model with probing prompting learning, the quality of improving the mathematical representation ability of students in both classes had moderate interpretation, and students' attitudes had good interpretations. A problem-based learning model with probing prompting learning can be used in learning to improve mathematical representation abilities.",,https://journal.institutpendidikan.ac.id/index.php/plusminus/article/download/pv1n2_09/854
-dictionary-based phrase-level prompting of large language models for machine translation,Marjan Ghazvininejad,"Large language models (LLMs) demonstrate remarkable machine translation (MT) abilities via prompting, even though they were not explicitly trained for this task. However, even given the incredible quantities of data they are trained on, LLMs can struggle to translate inputs with rare words, which are common in low resource or domain transfer scenarios. We show that LLM prompting can provide an effective solution for rare words as well, by using prior knowledge from bilingual dictionaries to provide control hints in the prompts. We propose a novel method, DiPMT, that provides a set of possible translations for a subset of the input words, thereby enabling fine-grained phrase-level prompted control of the LLM. Extensive experiments show that DiPMT outperforms the baseline both in low-resource MT, as well as for out-of-domain MT. We further provide a qualitative analysis of the benefits and limitations of this approach, including the overall level of controllability that is achieved.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that LLM prompting can provide an effective solution for rare words as well, by using prior knowledge from bilingual dictionaries to provide control hints in the prompts, thereby enabling fine-grained phrase-level prompted control of the LLM.'}",http://arxiv.org/pdf/2302.07856
-fine-grained controllable text generation using non-residual prompting,F. Carlsson,"The introduction of immensely large Causal Language Models (CLMs) has rejuvenated the interest in open-ended text generation. However, controlling the generative process for these Transformer-based models is at large an unsolved problem. Earlier work has explored either plug-and-play decoding strategies, or more powerful but blunt approaches such as prompting. There hence currently exists a trade-off between fine-grained control, and the capability for more expressive high-level instructions. To alleviate this trade-off, we propose an encoder-decoder architecture that enables intermediate text prompts at arbitrary time steps. We propose a resource-efficient method for converting a pre-trained CLM into this architecture, and demonstrate its potential on various experiments, including the novel task of contextualized word inclusion. Our method provides strong results on multiple experimental settings, proving itself to be both expressive and versatile.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an encoder-decoder architecture that enables intermediate text prompts at arbitrary time steps, and proposes a resource-efficient method for converting a pre-trained CLM into this architecture, and demonstrates its potential on various experiments, including the novel task of contextualized word inclusion.'}",https://aclanthology.org/2022.acl-long.471.pdf
-bloom+1: adding language support to bloom for zero-shot prompting,Zheng-Xin Yong,"The BLOOM model is a large publicly available multilingual language model, but its pretraining was limited to 46 languages. To extend the benefits of BLOOM to other languages without incurring prohibitively large costs, it is desirable to adapt BLOOM to new languages not seen during pretraining. In this work, we apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. We find language adaptation to be effective at improving zero-shot performance in new languages. Surprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models. In addition, we discover that prompting performance is not significantly affected by language specifics, such as the writing system. It is primarily determined by the size of the language adaptation data. We also add new languages to BLOOMZ, which is a multitask finetuned version of BLOOM capable of following task instructions zero-shot. We find including a new language in the multitask fine-tuning mixture to be the most effective method to teach BLOOMZ a new language. We conclude that with sufficient training data language adaptation can generalize well to diverse languages. Our code is available at https://github.com/bigscience-workshop/multilingual-modeling.","{'model': 'tldr@v2.0.0', 'text': 'This work applies existing language adaptation strategies to BLOOM and finds language adaptation to be effective at improving zero-shot performance in new languages and concludes that with sufficient training data language adaptation can generalize well to diverse languages.'}",http://arxiv.org/pdf/2212.09535
-understanding and improving visual prompting: a label-mapping perspective,Aochuan Chen,"We revisit and advance visual prompting (VP), an input prompting technique for vision tasks. VP can reprogram a fixed, pre-trained source model to accomplish downstream tasks in the target domain by simply incorporating universal prompts (in terms of input perturbation patterns) into downstream data points. Yet, it remains elusive why VP stays effective even given a ruleless label mapping (LM) between the source classes and the target classes. Inspired by the above, we ask: How is LM interrelated with VP? And how to exploit such a relationship to improve its accuracy on target tasks? We peer into the influence of LM on VP and provide an affirmative answer that a better ‘quality’ of LM (assessed by mapping precision and explanation) can consistently improve the effectiveness of VP. This is in contrast to the prior art where the factor of LM was missing. To optimize LM, we propose a new VP framework, termed ILM-VP (iterative label mapping-based visual prompting), which automatically re-maps the source labels to the target labels and progressively improves the target task accuracy of VP. Further, when using a contrastive language-image pretrained (CLIP) model for VP, we propose to integrate an LM process to assist the text prompt selection of CLIP and to improve the target task accuracy. Extensive experiments demonstrate that our proposal significantly outperforms state-of-the-art VP methods. As highlighted below, we show that when reprogramming an ImageNet-pretrained ResNet-18 to 13 target tasks, ILM-VP outperforms baselines by a substantial margin, e.g., 7.9% and 6.7% accuracy improvements in transfer learning to the target Flowers102 and CIFAR100 datasets. Besides, our proposal on CLIP-based VP provides 13.7% and 7.1% accuracy improvements on Flowers102 and DTD respectively. Code is available at https://github.com/OPTML-Group/ILM-VP.","{'model': 'tldr@v2.0.0', 'text': 'A new VP framework, termed ILM-VP (iterative label mapping-based visual prompting), which automatically re-maps the source labels to the target labels and progressively improves the target task accuracy of VP is proposed.'}",https://arxiv.org/pdf/2211.11635
-hybrid-sar technique: joint analysis using phase-based and amplitude-based methods for the xishancun giant landslide monitoring,T. Qu,"Early detection and early warning are of great importance in giant landslide monitoring because of the unexpectedness and concealed nature of large-scale landslides. In China, the western mountainous areas are prone to landslides and feature many giant complex landslides, especially following the Wenchuan Earthquake in 2008. This work concentrates on a new technique, known as the “hybrid-SAR technique”, that combines both phase-based and amplitude-based methods to detect and monitor large-scale landslides in Li County, Sichuan Province, southwestern China. This work aims to develop a robust methodological approach to promptly identify diverse landslides with different deformation magnitudes, sliding modes and slope geometries, even when the available satellite data are limited. The phase-based and amplitude-based techniques are used to obtain the landslide displacements from six TerraSAR-X Stripmap descending scenes acquired from November 2014 to March 2015. Furthermore, the application circumstances and influence factors of hybrid-SAR are evaluated according to four aspects: (1) quality of terrain visibility to the radar sensor; (2) landslide deformation magnitude and different sliding mode; (3) impact of dense vegetation cover; and (4) sliding direction sensitivity. The results achieved from hybrid-SAR are consistent with in situ measurements. This new hybrid-SAR technique for complex giant landslide research successfully identified representative movement areas, e.g., an extremely slow earthflow and a creeping region with a displacement rate of 1 cm per month and a typical rotational slide with a displacement rate of 2–3 cm per month downwards and towards the riverbank. Hybrid-SAR allows for a comprehensive and preliminary identification of areas with significant movement and provides reliable data support for the forecasting and monitoring of landslides.","{'model': 'tldr@v2.0.0', 'text': 'This work concentrates on a new technique, known as the “hybrid-SAR technique”, that combines both phase-based and amplitude-based methods to detect and monitor large-scale landslides in Li County, Sichuan Province, southwestern China, and successfully identified representative movement areas.'}",https://www.mdpi.com/2072-4292/8/10/874/pdf?version=1477292667
-automatic multi-label prompting: simple and interpretable few-shot classification,Han Wang,"Prompt-based learning (i.e., prompting) is an emerging paradigm for exploiting knowledge learned by a pretrained language model. In this paper, we propose Automatic Multi-Label Prompting (AMuLaP), a simple yet effective method to automatically select label mappings for few-shot text classification with prompting. Our method exploits one-to-many label mappings and a statistics-based algorithm to select label mappings given a prompt template. Our experiments demonstrate that AMuLaP achieves competitive performance on the GLUE benchmark without human effort or external resources.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Automatic Multi-Label Prompting (AMuLaP), a simple yet effective method to automatically select label mappings for few-shot text classification with prompting that achieves competitive performance on the GLUE benchmark without human effort or external resources.'}",http://arxiv.org/pdf/2204.06305
-fs-detr: few-shot detection transformer with prompting and without re-training,Adrian Bulat,"This paper is on Few-Shot Object Detection (FSOD), where given a few templates (examples) depicting a novel class (not seen during training), the goal is to detect all of its occurrences within a set of images. From a practical perspective, an FSOD system must fulfil the following desiderata: (a) it must be used as is, without requiring any fine-tuning at test time, (b) it must be able to process an arbitrary number of novel objects concurrently while supporting an arbitrary number of examples from each class and (c) it must achieve accuracy comparable to a closed system. Towards satisfying (a)-(c), in this work, we make the following contributions: We introduce, for the first time, a simple, yet powerful, few-shot detection transformer (FS-DETR) based on visual prompting that can address both desiderata (a) and (b). Our system builds upon the DETR framework, extending it based on two key ideas: (1) feed the provided visual templates of the novel classes as visual prompts during test time, and (2) ``stamp'' these prompts with pseudo-class embeddings (akin to soft prompting), which are then predicted at the output of the decoder. Importantly, we show that our system is not only more flexible than existing methods, but also, it makes a step towards satisfying desideratum (c). Specifically, it is significantly more accurate than all methods that do not require fine-tuning and even matches and outperforms the current state-of-the-art fine-tuning based methods on the most well-established benchmarks (PASCAL VOC&MSCOCO).","{'model': 'tldr@v2.0.0', 'text': 'This work introduces, for the first time, a simple, yet powerful, few-shot detection transformer (FS-DETR) based on visual prompting that can address both desiderata (a) and (b) and makes a step towards satisfying desideratum (c).'}",https://arxiv.org/pdf/2210.04845
-prompting contrastive explanations for commonsense reasoning tasks,Bhargavi Paranjape,"Many commonsense reasoning NLP tasks involve choosing between one or more possible answers to a question or prompt based on knowledge that is often implicit. Large pretrained language models (PLMs) can achieve near-human performance on such tasks, while providing little human-interpretable evidence of the underlying reasoning they use. In this work, we show how to use these same models to generate such evidence: inspired by the contrastive nature of human explanations, we use PLMs to complete explanation prompts which contrast alternatives according to the key attribute(s) required to justify the correct answer (for example, peanuts are usually salty while raisins are sweet). Conditioning model decisions on these explanations improves performance on two commonsense reasoning benchmarks, as compared to previous non-contrastive alternatives. These explanations are also judged by humans to be more relevant for solving the task, and facilitate a novel method to evaluate explanation faithfulfness.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the contrastive nature of human explanations, this work uses PLMs to complete explanation prompts which contrast alternatives according to the key attribute(s) required to justify the correct answer (for example, peanuts are usually salty while raisins are sweet).'}",https://aclanthology.org/2021.findings-acl.366.pdf
-enhancing cross-lingual prompting with mask token augmentation,Meng Zhou,"Prompting 1 shows promising results in few-shot scenarios. However, its strength for multilingual/cross-lingual problems has not been fully exploited. Zhao and Schütze (2021) made initial explorations in this direction by presenting that cross-lingual prompting outperforms cross-lingual finetuning. In this paper, we conduct empirical analysis on the effect of each component in cross-lingual prompting and derive Universal Prompting across languages, which helps alleviate the discrepancies between source-language training and target-language inference. Based on this, we propose a mask token augmentation framework to further improve the performance of prompt-based cross-lingual transfer. Notably, for XNLI, our method achieves 46.54% with only 16 English training examples per class, significantly better than 34.99% of finetuning.","{'model': 'tldr@v2.0.0', 'text': 'Empirical analysis on the effect of each component in cross-lingual prompting is conducted and Universal Prompting across languages is derived, which helps alleviate the discrepancies between source-language training and target-language inference.'}",
-a changing landscape: web-based methods for dietary assessment in adolescents,K. Storey,"Purpose of reviewAdolescents’ dietary intake is an important determinant of health and well-being and is influenced by a complex interaction of environmental, social, psychological, and physiological factors. The complexity of the adolescent diet makes its assessment prone to error, which has prompted researchers and clinicians to turn to technology to reduce this error. Previous reviews have been conducted regarding the use of technology in dietary assessment for adults; however, there are no known reviews for adolescents. Therefore, the purpose of this review is to describe the practical considerations for web-based dietary assessment methods and to evaluate recent evidence on their validity and implications. Recent findingsThere are numerous web-based dietary assessment methods that are available, valid, and reliable for use in the adolescent population. Web-based methods include both native and web-based applications (or ‘apps’), and have been developed for use as food records, 24-h dietary recalls, and food frequency questionnaires. SummaryWeb-based methods provide an efficient, cost-effective and practical solution to assess dietary intake; they are less burdensome to respondents and reduce errors and bias. Furthermore, adolescents are technologically savvy and often prefer the use of technology. Web-based methods should be considered when assessing adolescents’ dietary intake.","{'model': 'tldr@v2.0.0', 'text': 'Web-based methods provide an efficient, cost-effective and practical solution to assess dietary intake; they are less burdensome to respondents and reduce errors and bias.'}",
-generated knowledge prompting for commonsense reasoning,Jiacheng Liu,"It remains an open question whether incorporating external knowledge benefits commonsense reasoning while maintaining the flexibility of pretrained sequence models. To investigate this question, we develop generated knowledge prompting, which consists of generating knowledge from a language model, then providing the knowledge as additional input when answering a question. Our method does not require task-specific supervision for knowledge integration, or access to a structured knowledge base, yet it improves performance of large-scale, state-of-the-art models on four commonsense reasoning tasks, achieving state-of-the-art results on numerical commonsense (NumerSense), general commonsense (CommonsenseQA 2.0), and scientific commonsense (QASC) benchmarks. Generated knowledge prompting highlights large-scale language models as flexible sources of external knowledge for improving commonsense reasoning.Our code is available at github.com/liujch1998/GKP","{'model': 'tldr@v2.0.0', 'text': 'Generated knowledge prompting develops generated knowledge prompting, which consists of generating knowledge from a language model, then providing the knowledge as additional input when answering a question, and improves performance of large-scale, state-of-the-art models on four commonsense reasoning tasks.'}",https://aclanthology.org/2022.acl-long.225.pdf
-a review of deep learning-based three-dimensional medical image registration methods.,Haonan Xiao,"Medical image registration is a vital component of many medical procedures, such as image-guided radiotherapy (IGRT), as it allows for more accurate dose-delivery and better management of side effects. Recently, the successful implementation of deep learning (DL) in various fields has prompted many research groups to apply DL to three-dimensional (3D) medical image registration. Several of these efforts have led to promising results. This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research. The collected studies were statistically analyzed based on the region of interest (ROI), image modality, supervision method, and registration evaluation metrics. The studies were classified into three categories: deep iterative registration, supervised registration, and unsupervised registration. The studies are thoroughly reviewed and their unique contributions are highlighted. A summary is presented following a review of each category of study, discussing its advantages, challenges, and trends. Finally, the common challenges for all categories are discussed, and potential future research topics are identified.","{'model': 'tldr@v2.0.0', 'text': 'This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research.'}",https://qims.amegroups.com/article/viewFile/75304/pdf
-diffedit: diffusion-based semantic image editing with mask guidance,Guillaume Couairon,"Image generation has recently seen tremendous advances, with diffusion models allowing to synthesize convincing images for a large variety of text prompts. In this article, we propose DiffEdit, a method to take advantage of text-conditioned diffusion models for the task of semantic image editing, where the goal is to edit an image based on a text query. Semantic image editing is an extension of image generation, with the additional constraint that the generated image should be as similar as possible to a given input image. Current editing methods based on diffusion models usually require to provide a mask, making the task much easier by treating it as a conditional inpainting task. In contrast, our main contribution is able to automatically generate a mask highlighting regions of the input image that need to be edited, by contrasting predictions of a diffusion model conditioned on different text prompts. Moreover, we rely on latent inference to preserve content in those regions of interest and show excellent synergies with mask-based diffusion. DiffEdit achieves state-of-the-art editing performance on ImageNet. In addition, we evaluate semantic image editing in more challenging settings, using images from the COCO dataset as well as text-based generated images.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes DiffEdit, a method to take advantage of text-conditioned diffusion models for the task of semantic image editing, where the goal is to edit an image based on a text query.'}",http://arxiv.org/pdf/2210.11427
-make-a-scene: scene-based text-to-image generation with human priors,Oran Gafni,"Recent text-to-image generation methods provide a simple yet exciting conversion capability between text and image domains. While these methods have incrementally improved the generated image fidelity and text relevancy, several pivotal gaps remain unanswered, limiting applicability and quality. We propose a novel text-to-image method that addresses these gaps by (i) enabling a simple control mechanism complementary to text in the form of a scene, (ii) introducing elements that substantially improve the tokenization process by employing domain-specific knowledge over key image regions (faces and salient objects), and (iii) adapting classifier-free guidance for the transformer use case. Our model achieves state-of-the-art FID and human evaluation results, unlocking the ability to generate high fidelity images in a resolution of 512x512 pixels, significantly improving visual quality. Through scene controllability, we introduce several new capabilities: (i) Scene editing, (ii) text editing with anchor scenes, (iii) overcoming out-of-distribution text prompts, and (iv) story illustration generation, as demonstrated in the story we wrote.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel text-to-image method that addresses gaps in applicability and quality by enabling a simple control mechanism complementary to text in the form of a scene, and introducing elements that substantially improve the tokenization process by employing domain-specific knowledge over key image regions.'}",http://arxiv.org/pdf/2203.13131
-prompting healthier eating: testing the use of health and social norm based messages.,E. Robinson,"OBJECTIVE Health based messages are commonly used to promote fruit and vegetable intake, but are limited in their effectiveness. Social norm messages, which suggest other people are eating healthily, may be more effective. Our aim was to compare the effect on food selection of a message containing health related information about fruit and vegetable consumption with a message containing social normative information about consumption of fruit and vegetables. METHOD In two laboratory studies, predominantly young female adult students were exposed to a health or social norm message about fruit and vegetables. In Study 1, lunch meal food selections and intake were assessed and in Study 2, snack food selections and intake were assessed. Study 1 examined the effect of a descriptive social norm (information about what others are eating) versus a health message and Study 2 examined the effect of both a descriptive norm and an injunctive norm message (information about what others approve of) versus a health message. RESULTS In Study 1, exposure to a descriptive social norm message resulted in significantly more vegetables being selected and eaten than exposure to a health message. In Study 2, exposure to a descriptive social norm message resulted in significantly more fruit and vegetables and less high energy dense snack food being selected and eaten than exposure to a health message. There was no effect of exposure to the injunctive norm message. In both studies, significant differences between the social norm and health message conditions were observed in low but not high usual consumers of fruit and vegetables. CONCLUSIONS For the promotion of healthy eating, social norm messages may be more effective than health messages for consumers failing to adhere to dietary guidelines.","{'model': 'tldr@v2.0.0', 'text': 'For the promotion of healthy eating, social norm messages may be more effective than health messages for consumers failing to adhere to dietary guidelines.'}",
-dynamic prefix-tuning for generative template-based event extraction,Xiao Liu,"We consider event extraction in a generative manner with template-based conditional generation.Although there is a rising trend of casting the task of event extraction as a sequence generation problem with prompts, these generation-based methods have two significant challenges, including using suboptimal prompts and static event type information.In this paper, we propose a generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type-specific prefixes to learn a context-specific prefix for each context.Experimental results show that our model achieves competitive results with the state-of-the-art classification-based model OneIE on ACE 2005 and achieves the best performances on ERE.Additionally, our model is proven to be portable to new types of events effectively.","{'model': 'tldr@v2.0.0', 'text': 'A generative template-based event extraction method with dynamic prefix (GTEE-DynPref) by integrating context information with type- specific prefixes to learn a context-specific prefix for each context is proposed.'}",https://aclanthology.org/2022.acl-long.358.pdf
-covid-19 detection based on lung ct scan using deep learning techniques,S. V. Kogilavani,"SARS-CoV-2 is a novel virus, responsible for causing the COVID-19 pandemic that has emerged as a pandemic in recent years. Humans are becoming infected with the virus. In 2019, the city of Wuhan reported the first-ever incidence of COVID-19. COVID-19 infected people have symptoms that are related to pneumonia, and the virus affects the body's respiratory organs, making breathing difficult. A real-time reverse transcriptase-polymerase chain reaction (RT-PCR) kit is used to diagnose the disease. Due to a shortage of kits, suspected patients cannot be treated promptly, resulting in disease spread. To develop an alternative, radiologists looked at the changes in radiological imaging, like CT scans, that produce comprehensive pictures of the body of excellent quality. The suspected patient's computed tomography (CT) scan is used to distinguish between a healthy individual and a COVID-19 patient using deep learning algorithms. A lot of deep learning methods have been proposed for COVID-19. The proposed work utilizes CNN architectures like VGG16, DeseNet121, MobileNet, NASNet, Xception, and EfficientNet. The dataset contains 3873 total CT scan images with “COVID” and “Non-COVID.” The dataset is divided into train, test, and validation. Accuracies obtained for VGG16 are 97.68%, DenseNet121 is 97.53%, MobileNet is 96.38%, NASNet is 89.51%, Xception is 92.47%, and EfficientNet is 80.19%, respectively. From the obtained analysis, the results show that the VGG16 architecture gives better accuracy compared to other architectures.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the VGG16 architecture gives better accuracy compared to other architectures for COVID-19, a novel pandemic that has emerged as a pandemic in recent years.'}",https://downloads.hindawi.com/journals/cmmm/2022/7672196.pdf
-fusion-based supply chain collaboration using machine learning techniques,Naeem Ali,"Supply Chain Collaboration is the network of various entities that work cohesively to make up the entire process. The supply chain organizations’ success is dependent on integration, teamwork, and the communication of information. Every day, supply chain and business players work in a dynamic setting. They must balance competing goals such as process robustness, risk reduction, vulnerability reduction, real financial risks, and resilience against just-in-time and cost-efficiency. Decision-making based on shared information in Supply Chain Collaboration constitutes the recital and competitiveness of the collective process. Supply Chain Collaboration has prompted companies to implement the perfect data analytics functions (e.g., data science, predictive analytics, and big data) to improve supply chain operations and, eventually, efficiency. Simulation and modeling are powerful methods for analyzing, investigating, examining, observing and evaluating real-world industrial and logistic processes in this scenario. Fusion-based Machine learning provides a platform that may address the issues/limitations of Supply Chain Collaboration. Compared to the Classical probable data fusion techniques, the fused Machine learning method may offer a strong computing ability and prediction. In this scenario, the machine learningbased Supply Chain Collaboration model has been proposed to evaluate the propensity of the decision-making process to increase the efficiency of the Supply Chain Collaboration.","{'model': 'tldr@v2.0.0', 'text': 'The machine learningbased Supply Chain Collaboration model has been proposed to evaluate the propensity of the decision-making process to increase the efficiency of the supply chain Collaboration.'}",https://www.techscience.com/iasc/v31n3/44835/pdf
-a kinect-based vocational task prompting system for individuals with cognitive impairments,Yao-Jen Chang,,"{'model': 'tldr@v2.0.0', 'text': 'It is suggested that the image recognition technology may be able to facilitate task prompts needed by people with cognitive impairments and may be helpful for pre-service training while increasing independence in the process of community integration.'}",
-speechut: bridging speech and text with hidden-unit for encoder-decoder based speech-text pre-training,Zi-Hua Zhang,"The rapid development of single-modal pre-training has prompted researchers to pay more attention to cross-modal pre-training methods. In this paper, we propose a unified-modal speech-unit-text pre-training model, SpeechUT, to connect the representations of a speech encoder and a text decoder with a shared unit encoder. Leveraging hidden-unit as an interface to align speech and text, we can decompose the speech-to-text model into a speech-to-unit model and a unit-to-text model, which can be jointly pre-trained with unpaired speech and text data respectively. Our proposed SpeechUT is fine-tuned and evaluated on automatic speech recognition (ASR) and speech translation (ST) tasks. Experimental results show that SpeechUT gets substantial improvements over strong baselines, and achieves state-of-the-art performance on both the LibriSpeech ASR and MuST-C ST tasks. To better understand the proposed SpeechUT, detailed analyses are conducted. The code and pre-trained models are available at https://aka.ms/SpeechUT.","{'model': 'tldr@v2.0.0', 'text': 'A unified-modal speech-unit-text pre-training model to connect the representations of a speech encoder and a text decoder with a shared unit encoder, and achieves state-of-the-art performance on both the LibriSpeech ASR and MuST-C ST tasks.'}",http://arxiv.org/pdf/2210.03730
-an empirical study of gpt-3 for few-shot knowledge-based vqa,Zhengyuan Yang,"Knowledge-based visual question answering (VQA) involves answering questions that require external knowledge not present in the image. Existing methods first retrieve knowledge from external resources, then reason over the selected knowledge, the input image, and question for answer prediction. However, this two-step approach could lead to mismatches that potentially limit the VQA performance. For example, the retrieved knowledge might be noisy and irrelevant to the question, and the re-embedded knowledge features during reasoning might deviate from their original meanings in the knowledge base (KB). To address this challenge, we propose PICa, a simple yet effective method that Prompts GPT3 via the use of Image Captions, for knowledge-based VQA. Inspired by GPT-3’s power in knowledge retrieval and question answering, instead of using structured KBs as in previous work, we treat GPT-3 as an implicit and unstructured KB that can jointly acquire and process relevant knowledge. Specifically, we first convert the image into captions (or tags) that GPT-3 can understand, then adapt GPT-3 to solve the VQA task in a few-shot manner by just providing a few in-context VQA examples. We further boost performance by carefully investigating: (i) what text formats best describe the image content, and (ii) how in-context examples can be better selected and used. PICa unlocks the first use of GPT-3 for multimodal tasks. By using only 16 examples, PICa surpasses the supervised state of the art by an absolute +8.6 points on the OK-VQA dataset. We also benchmark PICa on VQAv2, where PICa also shows a decent few-shot performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PICa, a simple yet effective method that Prompts GPT3 via the use of Image Captions, for knowledge-based VQA, and treats GPT-3 as an implicit and unstructured KB that can jointly acquire and process relevant knowledge.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20215/19974
-comprehensive review of deep learning-based 3d point cloud completion processing and analysis,Ben Fei,"Point cloud completion is a generation and estimation issue derived from the partial point clouds, which plays a vital role in the applications of 3D computer vision. The progress of deep learning (DL) has impressively improved the capability and robustness of point cloud completion. However, the quality of completed point clouds is still needed to be further enhanced to meet the practical utilization. Therefore, this work aims to conduct a comprehensive survey on various methods, including point-based, view-based, convolution-based, graph-based, generative model-based, transformer-based approaches, etc. And this survey summarizes the comparisons among these methods to provoke further research insights. Besides, this review sums up the commonly used datasets and illustrates the applications of point cloud completion. Eventually, we also discussed possible research trends in this promptly expanding field.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to conduct a comprehensive survey on various methods of point cloud completion, including point-based, view- based, convolution-based), convolution, graph, graph based, generative model based, transformer-based approaches, etc, and summarizes the comparisons among these methods to provoke further research insights.'}",https://arxiv.org/pdf/2203.03311
-interactive-chain-prompting: ambiguity resolution for crosslingual conditional generation with interaction,Jonathan Pilault,"Crosslingual conditional generation (e.g., machine translation) has long enjoyed the benefits of scaling. Nonetheless, there are still issues that scale alone may not overcome. A source query in one language, for instance, may yield several translation options in another language without any extra context. Only one translation could be acceptable however, depending on the translator's preferences and goals. Choosing the incorrect option might significantly affect translation usefulness and quality. We propose a novel method interactive-chain prompting -- a series of question, answering and generation intermediate steps between a Translator model and a User model -- that reduces translations into a list of subproblems addressing ambiguities and then resolving such subproblems before producing the final text to be translated. To check ambiguity resolution capabilities and evaluate translation quality, we create a dataset exhibiting different linguistic phenomena which leads to ambiguities at inference for four languages. To encourage further exploration in this direction, we release all datasets. We note that interactive-chain prompting, using eight interactions as exemplars, consistently surpasses prompt-based methods with direct access to background information to resolve ambiguities.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method interactive-chain prompting -- a series of question, answering and generation intermediate steps between a Translator model and a User model -- that reduces translations into a list of subproblems addressing ambiguities and then resolving such subpro problems before producing the final text to be translated.'}",http://arxiv.org/pdf/2301.10309
-the effectiveness of embedded teaching through the most-to-least prompting procedure in concept teaching to children with autism within orff-based music activities.,Bilgehan Eren,"AbstractThe purpose of this study was to demonstrate the effectiveness of embedded teaching through the most-to- least prompting procedure in concept teaching to children with autism in Orff-based music activities. In this research, being one of the single subject research designs, multiple probe design was used. The generalization effect of the research was assessed in the form of a pre-test and post-test, discussing the organization about interpersonal generalization in different environment. The maintenance of learning was checked by means of the observation sessions applied one, two, and four weeks after the practice was completed. The research was conducted with 3 children with autism between the ages of 3-6 in Nova Special Training and Rehabilitation Center in Istanbul. In research process, full probe, daily probe, teaching, observing and generalizing sessions have been organized. All of the sessions were done by one-to-one teaching method. The data of reliability among the observers and of application reliability were collected. According to research results, it was clear that the embedding teaching through the most-to-least prompting in Orff-based music activities is effective to teach concepts to the chilren with autism. It was understood that the maintenance was kept after the training was completed. It was also clear that the concept which was thought is effective in the generalization of different settings, different people and different materials. Additionally, social validity was verified in research to deter- mine the importance of purposes of research, teaching methods to reach these purposes and research results.Key WordsChildren with Autism, Orff Approach, Embedded Teaching, Most-to-Least Prompting, Concept Teaching.Defined as permanent behavioral change arising from experiences in behavior or potential behavior, ""learning"" occurs through different ways for each person (Senemoglu, 2007). Special arrangements should be made for teaching children who have special needs. Functional education programs are developed and systematic teaching practices are implemented with regard to teaching of independent life skills to children with developmental disabilities (Eripek, 2005; Kurt, 2009).According to Special Education Services Regulation of the Ministry of National Education (2011), an individual who needs special education is defined as an individual who shows significant difference compared to the level expected from fellows with regard to educational and individual qualifications due to various reasons'. Individual with autism is defined as an individual whose restraints in social interaction, verbal and non-verbal communication, as well as interest and activity develop in early childhood period and who need special education and supportive education service due to these restraints (Milli Egitim Bakanhgi [MEB], 2012), while autistic disorder is defined as a developmental disorder characterized by incapacity in social interaction, language and communication abilities; restricted, stereotyped and repetitive patterns in behaviors, interest and activities, as well as disorders or extraordinary functionality in at least one of the following: social interaction, language used in social communication, or symbolic/imaginative play abilities, which occur before the age of 3 (Diken, 2008; Kircaali-Iftar, 2003; Korkmaz, 2005). In order to offer an efficient education by exploring individual learning methods of children with autism, it is required to organize the physical environment according to their requirements, work on improper behaviors that they exhibit, utilize visual supports and prompts, and to determine the most suitable teaching techniques for them (Korkmaz, 2003).Used in education of children with autism, 'Applied Behavior Analysis' involves the following steps respectively: determination of current performance of child, identification of fields where child suffers from skill deficits, classifying identified abilities into sub-steps, involving abilities (which are classified into sub-steps) into individual programs, and rewarding positive reactions of child (Darica, Abidoglu, & Gumuccu, 2005). …",,
-a feasibility study of enhancing independent task performance for people with cognitive impairments through the use of a handheld location-based prompting system,Yao-Jen Chang,"An autonomous task-prompting system is presented to increase workplace and life independence for people with cognitive impairments such as traumatic brain injury, intellectual disability, schizophrenia, and down syndrome. This paper describes an approach to providing distributed cognition support of work engagement for persons with cognitive disabilities. In the pilot study, a prototype was built and tested in a community-based rehabilitation program involving preservice food preparation training of eight participants with cognitive impairments. The results show improvement in helping with task engagement is statistically significant compared to the oral-instruction method. A follow-up comparative study with two participants evaluated the shadow-team approach against the proposed system. Although the number of participants was few, the participants were studied in depth and the findings were very promising. The results in the autonomous task prompting without staff intervention indicate that the performance is statistically as good as the shadow-team approach. Our findings suggest that acquisition of job skills may be facilitated by the proposed system in conjunction with operant conditioning strategies.","{'model': 'tldr@v2.0.0', 'text': 'The results in the autonomous task prompting without staff intervention indicate that the performance is statistically as good as the shadow-team approach, and suggest that acquisition of job skills may be facilitated by the proposed system in conjunction with operant conditioning strategies.'}",
-medicinal chemistry projects requiring imaginative structure-based drug design methods.,N. Moitessier,"Computational methods for docking small molecules to proteins are prominent in drug discovery. There are hundreds, if not thousands, of documented examples-and several pertinent cases within our research program. Fifteen years ago, our first docking-guided drug design project yielded nanomolar metalloproteinase inhibitors and illustrated the potential of structure-based drug design. Subsequent applications of docking programs to the design of integrin antagonists, BACE-1 inhibitors, and aminoglycosides binding to bacterial RNA demonstrated that available docking programs needed significant improvement. At that time, docking programs primarily considered flexible ligands and rigid proteins. We demonstrated that accounting for protein flexibility, employing displaceable water molecules, and using ligand-based pharmacophores improved the docking accuracy of existing methods-enabling the design of bioactive molecules. The success prompted the development of our own program, Fitted, implementing all of these aspects. The primary motivation has always been to respond to the needs of drug design studies; the majority of the concepts behind the evolution of Fitted are rooted in medicinal chemistry projects and collaborations. Several examples follow: (1) Searching for HDAC inhibitors led us to develop methods considering drug-zinc coordination and its effect on the pKa of surrounding residues. (2) Targeting covalent prolyl oligopeptidase (POP) inhibitors prompted an update to Fitted to identify reactive groups and form bonds with a given residue (e.g., a catalytic residue) when the geometry allows it. Fitted-the first fully automated covalent docking program-was successfully applied to the discovery of four new classes of covalent POP inhibitors. As a result, efficient stereoselective syntheses of a few screening hits were prioritized rather than synthesizing large chemical libraries-yielding nanomolar inhibitors. (3) In order to study the metabolism of POP inhibitors by cytochrome P450 enzymes (CYPs)-for toxicology studies-the program Impacts was derived from Fitted and helped us to reveal a complex metabolism with unforeseen stereocenter isomerizations. These efforts, combined with those of other docking software developers, have strengthened our understanding of the complex drug-protein binding process while providing the medicinal chemistry community with useful tools that have led to drug discoveries. In this Account, we describe our contributions over the past 15 years-within their historical context-to the design of drug candidates, including BACE-1 inhibitors, POP covalent inhibitors, G-quadruplex binders, and aminoglycosides binding to nucleic acids. We also remark the necessary developments of docking programs, specifically Fitted, that enabled structure-based design to flourish and yielded multiple fruitful, rational medicinal chemistry campaigns.","{'model': 'tldr@v2.0.0', 'text': 'Your contributions over the past 15 years are described, within their historical context-to the design of drug candidates, including BACE-1 inhibitors, POP covalent inhibitors, G-quadruplex binders, and aminoglycosides binding to nucleic acids.'}",
-challenges from tuberculosis diagnosis to care in community-based active case finding among the urban poor in cambodia: a mixed-methods study,N. Lorent,"Background While community-based active case finding (ACF) for tuberculosis (TB) holds promise for increasing early case detection among hard-to-reach populations, limited data exist on the acceptability of active screening. We aimed to identify barriers and explore facilitators on the pathway from diagnosis to care among TB patients and health providers. Methods Mixed-methods study. We administered a survey questionnaire to, and performed in-depth interviews with, TB patients identified through ACF from poor urban settlements in Phnom Penh, Cambodia. Additionally, we conducted focus group discussions and in-depth interviews with community and public health providers involved in ACF, respectively. Results Acceptance of home TB screening was strong among key stakeholders due to perceived reductions in access barriers and in direct and indirect patient costs. Privacy and stigma were not an issue. To build trust and facilitate communication, the participation of community representatives alongside health workers was preferred. Most health providers saw ACF as complementary to existing TB services; however, additional workload as a result of ACF was perceived as straining operating capacity at public sector sites. Proximity to a health facility and disease severity were the strongest determinants of prompt care-seeking. The main reasons reported for delays in treatment-seeking were non-acceptance of diagnosis, high indirect costs related to lost income/productivity and transportation expenses, and anticipated side-effects from TB drugs. Conclusions TB patients and health providers considered home-based ACF complementary to facility-based TB screening. Strong engagement with community representatives was believed critical in gaining access to high risk communities. The main barriers to prompt treatment uptake in ACF were refusal of diagnosis, high indirect costs, and anticipated treatment side-effects. A patient-centred approach and community involvement were essential in mitigating barriers to care in marginalised communities.","{'model': 'tldr@v2.0.0', 'text': 'The main barriers to prompt treatment uptake in ACF were refusal of diagnosis, high indirect costs, and anticipated treatment side-effects, and a patient-centred approach and community involvement were essential in mitigating barriers to care in marginalised communities.'}",https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0130179&type=printable
-applying cusum-based methods for the detection of outbreaks of ross river virus disease in western australia,R. Watkins,,"{'model': 'tldr@v2.0.0', 'text': 'The retrospective analysis of historical data suggests that the negative binomial cusum provides greater sensitivity for the detection of outbreaks of RRv disease at low false alarm levels, and decreased timeliness early in the outbreak period.'}",https://bmcmedinformdecismak.biomedcentral.com/counter/pdf/10.1186/1472-6947-8-37
-embedded simultaneous prompting procedure to teach stem content to high school students with moderate disabilities in an inclusive setting.,S. Heinrich,"Effects of an embedded simultaneous prompting procedure to teach STEM (science, technology, engineering, math) content to three secondary students with moderate intellectual disabilities in an inclusive general education classroom were evaluated in the current study. Students learned discrete (i.e., geometric figures, science vocabulary, or use of technology to publish) and chained tasks (i.e., linear equation, Punnett square, or Internet research) from a peer tutor and paraprofessionals. Using a multiple probe across participants design, results showed students reached criterion in two to eight sessions and maintained the skills for one month following intervention. Generalization was also at higher levels than in the baseline condition. In addition, general education students who attended class with the participants expressed positive comments based on the intervention. The National Science Board has argued that every student in the United States “. . . deserves the opportunity to achieve his or her full potential” (NSF, 2010, p. v). As STEM (science, technology, engineering, math) permeates every aspect of our lives, it is crucial that all students have access to this content. Students benefit from science content by learning about themselves and the natural world; from engaging in technology by learning how to use smart phones, mobile devices, and computers; from engineering concepts by learning how to solve problems; and from math content by learning how to budget their money and determining how much they have for groceries (Science Pioneers, 2013). There is a national emphasis on preparing students for STEM careers; however, all citizens, even those not pursuing STEM careers, should be able to participate in the scientific and technical issues affecting our society (Matthews, 2007). Traditionally, students with disabilities have been underrepresented in STEM education, and students with moderate intellectual disabilities are especially at risk. Common Core State Standards (CCSS) and the Next Generation Science Standards (NGSS) emphasize the importance of high expectations for all students. According to Kleinert, Kearns, and Kennedy (1997), “. . . one of the keys to ensuring high expectations for every child is requiring that all students be included in measures of educational accountability” (p. 88). Based on these assertions, it is likely that assessments for all students, including alternate assessments based on alternate achievement standards (AA-AAS) for students with moderate and severe disabilities (MSD), will be derived from the content recommended in the CCSS and NGSS. While practices for implementing the AAAAS vary from state to state, having access to the general education curriculum is key for promoting student progress in content areas. Jackson, Ryndak, and Wehmeyer (2010) argued that the interplay between context of Correspondence concerning this article should be addressed to Amy D. Spriggs, Department of Early Childhood, Special Education, and Rehabilitation Counseling, University of Kentucky, 229 Taylor Education Building, Lexington, KY 40506-0001. E-mail: amy.spriggs@uky.edu Education and Training in Autism and Developmental Disabilities, 2016, 51(1), 41–54 © Division on Autism and Developmental Disabilities Teaching STEM Content to High School Students / 41 instruction and curriculum content is a critical consideration in student learning and progress since students with disabilities who are included in general education contexts are more likely to engage in activities related to grade level standards than students who are in self-contained settings. Hudson, Browder, and Wood (2013) conducted a literature review to evaluate evidence-based practices for students with MSD in inclusive contexts. Authors found embedded instruction (specifically using constant time delay [CTD]) to be an evidence-based practice. Embedded instruction can be defined as an instructor (teacher, paraprofessional, or peer) distributing trials throughout a lesson or naturally occurring routine in the inclusive classroom. Studies in the review focused on teaching reading and writing skills (e.g., Collins, Hall, Branson, & Holder, 1999), social studies skills (e.g., Collins, Evans, Creech-Galloway, Karl, & Miller, 2007), science skills (e.g., Jimenez, Browder, Spooner, & DiBiase, 2012), and math skills (McDonnell et al., 2006) to students with moderate and severe disabilities within general education classrooms. Hudson et al. (2013) commended researchers for evaluating procedures to teach a range of academic content in general education settings; however, they recommended additional research on strategies to teach complex and chained skills in inclusive environments. Further, only two quality studies in the review used simultaneous prompting (SP), suggesting a need for further research on this strategy in general education settings. In most of the quality studies reviewed in the Hudson et al. (2013) review, researchers used systematic instruction to teach students core content in general education settings. Methods of systematic instruction include the system of least prompts (SLP), time delay, and simultaneous prompting procedures. Perhaps the easiest and least time consuming method to implement is the SP procedure because the instructor presents the stimulus and the controlling prompt at the same time on each training trial, conducting daily probe trials prior to instructional trials to determine if the student has acquired the target skill (Collins, 2012). Collins et al. (2007) used the SP procedure to teach core content vocabulary to 4 students with MSD at the elementary, middle, and secondary levels, finding the procedure to be effective whether it was delivered in a segregated or an inclusive classroom setting. Riesen, McDonnell, Johnson, Polychronis, and Jameson (2003) and Head, Collins, Schuster, and Ault (2011) compared the efficiency of the CTD and SP procedures in teaching academic content (e.g., science and history vocabulary, state capitols) to middle or high school students with disabilities. In both studies, the researchers found the CTD and SP procedures to be equally effective, but the SP procedure was found to be slightly more efficient (i.e., faster acquisition of skills). A body of literature has emerged suggesting that, in addition to academic gains, inclusive settings offer the opportunity for positive social effects on students, both with and without disabilities (e.g., Kennedy, Shukla, & Fryxell, 1997). Results have indicated greater social benefits for students who were included full time in general education classrooms. For example, students who were included full time engaged in more frequent interactions with peers without disabilities, more social contacts across a wide range of activities and settings, higher levels of social support behaviors, larger friendship networks, and longer lasting relationships with peers. While prior research has shown promising effects as a result of teaching in inclusive classrooms, there are several limitations to the research. For example, few studies to date have taught students complex or chained skills in inclusive settings, and no studies exist evaluating STEM-related chained tasks. Hudson et al. (2013) recommended future research using SP in inclusive settings, and despite previous calls for research, sight word identification remains the skill most frequently addressed in these settings. In addition, no research exists on teaching STEM skills identified as part of the AA-AAS in a general education classroom using an embedded SP procedure. The goal of this study is to evaluate the effects of using an embedded SP procedure to teach STEMrelated tasks to students with moderate intellectual disabilities in an inclusive setting. The current study examined instruction on the state standards related to STEM content on which students taking part in the AA-AAS were tested. The study occurred in a suburban high school in a southern state, therefore fo42 / Education and Training in Autism and Developmental Disabilities-March 2016 cusing on the state’s AA-AAS. The special education teacher and the general education teacher collaborated to develop a plan for instruction in the general education classroom to be delivered by a paraprofessional or a peer tutor, which included embedding the SP procedure into the natural routine of the general education classroom. Specifically, the current study focused on the following research questions: (a) Will the embedded SP procedure result in the acquisition of STEM content skills by students with moderate intellectual disabilities in the general education classroom? (b) Will peer tutors and paraprofessionals implement the embedded SP procedure with high levels of procedural fidelity? (c) Will the target students maintain the acquired skills over time? (d) Will the target students generalize the STEM skills to situations that simulate the AA-AAS? and (e) Will inclusion of the target students in the general education classroom result in positive social relationships with their peers without disabilities?",,
-dna-based methods for monitoring invasive species: a review and prospectus,J. Darling,,,
-technological application of tannin-based extracts,M. Fraga-Corral,"Tannins are polyphenolic compounds naturally found in vegetables. Their presence in nature has prompted their historical use in many different ways. The revision of their traditional utilization has allowed their further modification aiming for an industrial application. Sometimes these modifications have implied the addition of harmful substances such as formaldehyde, classified as a carcinogen of category B1. In other cases, these natural tannins have been replaced by synthetic compounds that threaten human and animal health and damage the environment. Therefore, currently, both academy and industry are searching for the substitution of these unsafe complexes by the increasing inclusion of tannins, natural molecules that can be obtained from several and diverse renewable resources, modified using harmless additives. To achieve promising results, cost-efficient and eco-friendly extraction methods have been designed. Once these green alternatives have been isolated, they have been successfully applied to many fields with very assorted aims of utilization such as coagulants, adhesives, floatation agents, tannings, dyes, additives, or biomolecules. Therefore, this review offers a global vision of the full process that involves the tannin’s technological application including an overview of the most relevant tannin sources, effective extraction methods, and their utilization in very diverse fields.","{'model': 'tldr@v2.0.0', 'text': 'This review offers a global vision of the full process that involves the tannin’s technological application including an overview of the most relevant tannIn sources, effective extraction methods, and their utilization in very diverse fields.'}",https://www.mdpi.com/1420-3049/25/3/614/pdf?version=1581334096
-a comparison of automatic cell identification methods for single-cell rna sequencing data,T. Abdelaal,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that most classifiers perform well on a variety of datasets with decreased accuracy for complex datasets with overlapping classes or deep annotations, but the general-purpose support vector machine classifier has overall the best performance across the different experiments.'}",https://genomebiology.biomedcentral.com/track/pdf/10.1186/s13059-019-1795-z
-fatezero: fusing attentions for zero-shot text-based video editing,Chenyang Qi,"The diffusion-based generative models have achieved remarkable success in text-based image generation. However, since it contains enormous randomness in generation progress, it is still challenging to apply such models for real-world visual content editing, especially in videos. In this paper, we propose FateZero, a zero-shot text-based editing method on real-world videos without per-prompt training or use-specific mask. To edit videos consistently, we propose several techniques based on the pre-trained models. Firstly, in contrast to the straightforward DDIM inversion technique, our approach captures intermediate attention maps during inversion, which effectively retain both structural and motion information. These maps are directly fused in the editing process rather than generated during denoising. To further minimize semantic leakage of the source video, we then fuse self-attentions with a blending mask obtained by cross-attention features from the source prompt. Furthermore, we have implemented a reform of the self-attention mechanism in denoising UNet by introducing spatial-temporal attention to ensure frame consistency. Yet succinct, our method is the first one to show the ability of zero-shot text-driven video style and local attribute editing from the trained text-to-image model. We also have a better zero-shot shape-aware editing ability based on the text-to-video model. Extensive experiments demonstrate our superior temporal consistency and editing capability than previous works.","{'model': 'tldr@v2.0.0', 'text': 'F FateZero, a zero-shot text-based editing method on real-world videos without per-prompt training or use-specific mask, is proposed, which is the first one to show the ability of zero- shot text-driven video style and local attribute editing from the trained text-to-image model.'}",https://arxiv.org/pdf/2303.09535
-fast and constrained absent keyphrase generation by prompt-based learning,Huanqin Wu,"Generating absent keyphrases, which do not appear in the input document, is challenging in the keyphrase prediction task. Most previous works treat the problem as an autoregressive sequence-to-sequence generation task, which demonstrates promising results for generating grammatically correct and fluent absent keyphrases. However, such an end-to-end process with a complete data-driven manner is unconstrained, which is prone to generate keyphrases inconsistent with the input document. In addition, the existing autoregressive decoding method makes the generation of keyphrases must be done from left to right, leading to slow speed during inference. In this paper, we propose a constrained absent keyphrase generation method in a prompt-based learning fashion. Specifically, the prompt will be created firstly based on the keywords, which are defined as the overlapping words between absent keyphrase and document. Then, a mask-predict decoder is used to complete the absent keyphrase on the constraint of prompt. Experiments on keyphrase generation benchmarks have demonstrated the effectiveness of our approach. In addition, we evaluate the performance of constrained absent keyphrases generation from an information retrieval perspective. The result shows that our approach can generate more consistent keyphrases, which can improve document retrieval performance. What’s more, with a non-autoregressive decoding manner, our model can speed up the absent keyphrase generation by 8.67× compared with the autoregressive method.","{'model': 'tldr@v2.0.0', 'text': 'The result shows that the proposed constrained absent keyphrase generation method can generate more consistent keyphrases, which can improve document retrieval performance, and with a non-autoregressive decoding manner, can speed up the absentKeyphrase generation by 8.67× compared with the autoregressive method.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21402/21151
-rare-earth based nanomaterials and their composites as electrode materials for high performance supercapacitors: a review,Shunfei Liang,"The emergence of energy crisis and greenhouse effect has prompted people to develop energy storage equipment with excellent performance. Supercapacitors (SCs), also known as electrochemical capacitors, are widely studied for their high power density, fast charge and discharge and long cycle life. Rare earth is a group of elements with unique structure and properties. Rare earth-based nanomaterials, especially rare earth oxides, hydroxides and sulfides, have attracted much attention due to their good redox properties and environmentally friendly characteristics. Rare earth-based SCs nanomaterials can be obtained by environmentally friendly, simple and low-cost methods, such as hydrothermal/solvothermal method, electrodeposition method, atomic layer deposition method, etc. This paper reviews the synthesis methods, morphology and electrochemical properties of various rare earth-based nanomaterials and their composites for SCs. Finally, based on the reported literature, we discussed some challenges faced by rare earth-based nanomaterials and the issues that need attention in SCs and other electrochemical energy storage fields in the future.",,
-ai-based prognostic imaging biomarkers for precision neurooncology: the respond consortium.,C. Davatzikos,"Artificial intelligence (AI) and machine learning (ML) methods have begun to reveal that complex imaging patterns can provide individualized biomarkers for diagnosis and prognosis. However, AI methods have been challenged by insufficient training, heterogeneity of imaging protocols across hospitals, and lack of generalization to new patient data. These challenges prompted the development of the ReSPOND (Radiomics Signatures for PrecisiON Diagnostics) consortium on glioblastoma (GBM). This collaboration of over 10 institutions, across 3 continents, is positioned to pool, harmonize, and analyze brain MRIs from more than 3300 de novo GBM patients who underwent the Stupp protocol, in addition to datasets from The Cancer Imaging Archive (TCIA).1 ReSPOND aims to further develop and test AI-based biomarkers for individualized prediction and prognostication, by moving from single-institution studies to generalized, well-validated predictive biomarkers in the following 4 areas:","{'model': 'tldr@v2.0.0', 'text': 'The ReSPOND consortium is positioned to pool, harmonize, and analyze brain MRIs from more than 3300 de novo GBM patients who underwent the Stupp protocol, in addition to datasets from The Cancer Imaging Archive (TCIA).'}",https://academic.oup.com/neuro-oncology/article-pdf/22/6/886/33373505/noaa045.pdf
-technical attribute prioritisation in qfd based on cloud model and grey relational analysis,Xu Wang,"Promptly development of new products can be achieved through quality function deployment (QFD) process, which is critical to companies’ survival. Since the multi-criteria decision-making problem involved in QFD, a novel method integrating cloud model and grey relational analysis is put forward in this paper. Taking into account the subjectivity and ambiguity in linguistic evaluations, some scholars utilise fuzzy theory, rough theory, interval-valued fuzzy-rough sets and MCDM methods to improve traditional QFD. However, much priori information requirements, inability to handle subjectivity and randomness, and lack of mechanism to overcome small sample size problem are some inevitable drawbacks in these methods. To solve these deficiencies, a hybrid methodology is proposed in this paper, integrating the fortes of cloud model in processing ambiguity and randomness, and the merits of grey relational analysis in overcoming small sample size error as well as revealing the inner correlations. The comparative analysis of different approaches as well as the sensitivity analysis of criteria weights is implemented to prove the stability of the novel method. The results obtained in this paper shows that the proposed method can be a practical tool for improving the efficiency and accuracy of traditional QFD in reality management.","{'model': 'tldr@v2.0.0', 'text': 'A hybrid methodology is proposed in this paper, integrating the fortes of cloud model in processing ambiguity and randomness, and the merits of grey relational analysis in overcoming small sample size error as well as revealing the inner correlations.'}",
-using simple technology to prompt multistep tasks in the home for people with dementia: an exploratory study comparing prompting formats,H. Boyd,"Objectives To investigate the relative effectiveness of different prompts for people with dementia during multistep tasks in the home, to inform prompting technology design. Methods Nine pairs of participants (one with dementia and a partner or relative) participated at home. The participants with mild to moderate dementia (5M/4F, aged 73–86 years) functioned at the Planned or Exploratory levels of the Pool Activity Level instrument. A touchscreen computer displayed different prompts during two set tasks: “card-and-envelope” and “CD player.” The trials were scored to establish the relative effectiveness of the prompts. Individual tasks were also explored. Results Text and audio prompts were each more effective than video or picture prompts for a card-and-envelope task, but this was not seen in a CD player task. The differences may be related to the type of actions within the tasks; the card-and-envelope actions were easier to convey verbally; the CD player actions lent themselves to visual prompts. Conclusions Designers of technology-based prompts for people with dementia should consider that the effectiveness of different prompts is likely to be task dependent. Familiar, unambiguous language can increase the success of tailored prompts. There are significant practical challenges associated with choosing and deconstructing everyday tasks at home.","{'model': 'tldr@v2.0.0', 'text': 'Designers of technology-based prompts for people with dementia should consider that the effectiveness of different prompts is likely to be task dependent, and familiar, unambiguous language can increase the success of tailored prompts.'}",https://journals.sagepub.com/doi/pdf/10.1177/1471301215602417
-green extraction methods for polyphenols from plant matrices and their byproducts: a review.,K. Ameer,"Polyphenols as phytochemicals have gained significant importance owing to several associated health benefits with regard to lifestyle diseases and oxidative stress. To date, the development of a single standard method for efficient and rapid extraction of polyphenols from plant matrices has remained a challenge due to the inherent limitations of various conventional extraction methods. The exploitation of polyphenols as bioactive compounds at various commercial levels has motivated scientists to explore more eco-friendly, efficient, and cost-effective extraction techniques, based on a green extraction approach. The current review aims to provide updated technical information about extraction mechanisms, their advantages and disadvantages, and factors affecting efficiencies, and also presents a comparative overview of applications of the following modern green extraction techniques-supercritical fluid extraction, ultrasound-assisted extraction, microwave-assisted extraction, pressurized liquid extraction, and pressurized hot water extraction-as alternatives to conventional extraction methods for polyphenol extraction. These techniques are proving to be promising for the extraction of thermolabile phenolic compounds due to their advantages over conventional, time-consuming, and laborious extraction techniques, such as reduced solvent use and time and energy consumption and higher recovery rates with lower operational costs. The growing interest in plant-derived polyphenols prompts continual search for green and economically feasible modern extraction techniques. Modern green extraction techniques represent promising approaches by virtue of overcoming current limitations to the exploitation of polyphenols as bioactive compounds to explore their wide-reaching applications on an industrial scale and in emerging global markets. Future research is needed in order to remove the technical barriers to scale-up the processes for industrial needs by increasing our understanding and improving the design of modern extraction operations.","{'model': 'tldr@v2.0.0', 'text': 'These techniques are proving to be promising for the extraction of thermolabile phenolic compounds due to their advantages over conventional, time-consuming, and laborious extraction techniques, such as reduced solvent use and time and energy consumption and higher recovery rates with lower operational costs.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/1541-4337.12253
-evaluation of novel precision viticulture tool for canopy biomass estimation and missing plant detection based on 2.5d and 3d approaches using rgb images acquired by uav platform,S. F. Di Gennaro,,"{'model': 'tldr@v2.0.0', 'text': 'A rapid and objective tool for the farmer to promptly identify canopy management strategies and drive replanting decisions and the overcoming of the current limit represented by the pre- and post-processing phases of the large image dataset should mainstream this methodology.'}",https://plantmethods.biomedcentral.com/track/pdf/10.1186/s13007-020-00632-2
-2016 updated eular evidence-based recommendations for the management of gout,P. Richette,"Background New drugs and new evidence concerning the use of established treatments have become available since the publication of the first European League Against Rheumatism (EULAR) recommendations for the management of gout, in 2006. This situation has prompted a systematic review and update of the 2006 recommendations. Methods The EULAR task force consisted of 15 rheumatologists, 1 radiologist, 2 general practitioners, 1 research fellow, 2 patients and 3 experts in epidemiology/methodology from 12 European countries. A systematic review of the literature concerning all aspects of gout treatments was performed. Subsequently, recommendations were formulated by use of a Delphi consensus approach. Results Three overarching principles and 11 key recommendations were generated. For the treatment of flare, colchicine, non-steroidal anti-inflammatory drugs (NSAIDs), oral or intra-articular steroids or a combination are recommended. In patients with frequent flare and contraindications to colchicine, NSAIDs and corticosteroids, an interleukin-1 blocker should be considered. In addition to education and a non-pharmacological management approach, urate-lowering therapy (ULT) should be considered from the first presentation of the disease, and serum uric acid (SUA) levels should be maintained at<6 mg/dL (360 µmol/L) and <5 mg/dL (300 µmol/L) in those with severe gout. Allopurinol is recommended as first-line ULT and its dosage should be adjusted according to renal function. If the SUA target cannot be achieved with allopurinol, then febuxostat, a uricosuric or combining a xanthine oxidase inhibitor with a uricosuric should be considered. For patients with refractory gout, pegloticase is recommended. Conclusions These recommendations aim to inform physicians and patients about the non-pharmacological and pharmacological treatments for gout and to provide the best strategies to achieve the predefined urate target to cure the disease.","{'model': 'tldr@v2.0.0', 'text': 'In addition to education and a non-pharmacological management approach, urate-lowering therapy (ULT) should be considered from the first presentation of the disease, and serum uric acid (SUA) levels should be maintained at the predefined urate target to cure the disease.'}",https://ard.bmj.com/content/annrheumdis/76/1/29.full.pdf
-"uv-protective, self-cleaning, and antibacterial nanofiber-based triboelectric nanogenerators for self-powered human motion monitoring.",Yang Jiang,"Equipping wearable electronics with special functions will endow them with more additional values and more comprehensive practical performance. Here, we report an ultraviolet (UV)-protective, self-cleaning, antibacterial, and self-powered all-nanofiber-based triboelectric nanogenerator (TENG) for mechanical energy harvesting and self-powered sensing, which is fabricated with Ag nanowires (NWs)/TPU nanofibers and the TiO2@PAN networks through a facile electrospinning method. Due to the added TiO2 nanoparticles (NPs), the TENG presents excellent UV-protective performance, including the ultraviolet protection factor (UPF) of ∼204, the transmittance of UVA (TUVA) of ∼0.0574%, and the transmittance of UVB (TUVB) ∼0.107%. Furthermore, under solar lighting for 25 min, most surface contamination can be degraded, and the decreased power output would be recovered. Owing to the coupled effects of TiO2 NPs and Ag NWs, the TENG shows excellent antibacterial activity against Staphylococcus aureus. Due to the micro-to-nano hierarchical porous structure, the all-nanofiber-based TENG can serve as self-powered pedometers for detecting and tracking human motion behaviors. As a multifunctional self-powered device, the TENG prompts various applications in the fields of micro/nanopower sources, human movement monitoring, and human-machine interfaces, potentially providing an alternative energy solution and a multifunctional interactive platform for the next-generation wearable electronics.","{'model': 'tldr@v2.0.0', 'text': 'An ultraviolet (UV)-protective, self-cleaning, antibacterial, and self-powered all-nanofiber-based triboelectric nanogenerator (TENG) for mechanical energy harvesting andSelf-powered sensing that prompts various applications in the fields of micro/nanopower sources, human movement monitoring, and human-machine interfaces is reported.'}",
-prompt-based re-ranking language model for asr,Mengxi Nie,"In Automatic Speech Recognition(ASR), the language model re-ranking based on unlabeled text can improve the performance and realize flexibly scene adaptation. The scheme of ASR re-ranking is usually to build a language model and then use it to reorder the speech recognition N-best hypotheses. Recently, BERT-based re-ranking has achieved impressive results, benefiting from the powerful modeling capability of contextual semantic. In the view of that BERT’s nonautoregressive structure limits the calculation speed of the language model scores(perplexity, ppl), we use a classification method in prompt paradigm instead of the re-ranking method based on ppl. The prompt-based re-ranking scheme simplifies the pipeline of re-ranking as well as ensures the performance. Experiments on AISHELL-1 dataset show the effective of our proposed method. On the test set, the inference speed is accelerated by 49 times and compared to baseline the Character Error Rate(CER) is relatively decreased by 13.51% ∼ 14.43%.","{'model': 'tldr@v2.0.0', 'text': 'In the view of that BERT’s nonautoregressive structure limits the calculation speed of the language model scores(perplexity, ppl), this work uses a classification method in prompt paradigm instead of the re-ranking method based on ppl.'}",
-the effectiveness of video prompting on teaching aquatic play skills for children with autism,M. Yanardag,"Objective: To investigate the effectiveness of the video prompting procedure on teaching aquatic play skills and to determine the effects of aquatic exercise training on the motor performance of children with autism. Design: A multiple probe design across behaviours was used and replicated across subjects for the instructional part of this study. Pretest–posttest design was applied for the exercise training part of this study. Methods: Three children with autism were taught three aquatic play skills in a one-to-one training format. Aquatic play skills intervention and aquatic exercise training were performed separately throughout 12 weeks at three sessions per week, each lasting 1 h. The video prompting procedure was utilized for the instruction part of this study. Results: Video prompting was effective in teaching aquatic play skills to children with autism. In addition, aquatic exercise training increased the total motor performance scores of all the participants after 12 weeks. According to the social validity results, the families gave positive feedback about the learned skills and movement capabilities of their children. Conclusion: Aquatic play skills and swimming pools are favoured for children with autism. This attractive intervention is recommended as a means to extend knowledge of leisure skills and motor development of children with autism. Implications for Rehabilitation Video prompting may utilize to teach action-based motor tasks. Aquatic exercise training could be beneficial for improving movement difficulties of children with autism. Video-based instruction and swimming pool is recommended to have alluring effects on children with special needs.","{'model': 'tldr@v2.0.0', 'text': 'Video prompting was effective in teaching aquatic play skills to children with autism and aquatic exercise training increased the total motor performance scores of all the participants after 12 weeks.'}",
-classification of normal and depressed eeg signals based on centered correntropy of rhythms in empirical wavelet transform domain,H. Akbari,,"{'model': 'tldr@v2.0.0', 'text': 'A new method based on centered correntropy (CC) and empirical wavelet transform (EWT) for the classification of normal and depressed EEG signals can be used as a fast and accurate computer-aided detection system for the diagnosis of patients with depression in clinics and hospitals.'}",
-lfpt5: a unified framework for lifelong few-shot language learning based on prompt tuning of t5,Chengwei Qin,"Existing approaches to lifelong language learning rely on plenty of labeled data for learning a new task, which is hard to obtain in most real scenarios. Considering that humans can continually learn new tasks from a handful of examples, we expect the models also to be able to generalize well on new few-shot tasks without forgetting the previous ones. In this work, we define this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and propose a unified framework for it based on prompt tuning of T5. Our framework called LFPT5 takes full advantage of PT's strong few-shot learning ability, and simultaneously trains the model as a task solver and a data generator. Before learning a new domain of the same task type, LFPT5 generates pseudo (labeled) samples of previously learned domains, and later gets trained on those samples to alleviate forgetting of previous knowledge as it learns the new domain. In addition, a KL divergence loss is minimized to achieve label consistency between the previous and the current model. While adapting to a new task type, LFPT5 includes and tunes additional prompt embeddings for the new task. With extensive experiments, we demonstrate that LFPT5 can be applied to various different types of tasks and significantly outperform previous methods in different LFLL settings.","{'model': 'tldr@v2.0.0', 'text': ""This work defines this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and proposes a unified framework for it based on prompt tuning of T5 called LFPT5, which takes full advantage of PT's strong few-shot learning ability and simultaneously trains the model as a task solver and a data generator.""}",
-verbal prompting to improve everyday cognition in mci and unimpaired older adults.,K. Thomas,"OBJECTIVE This study investigated the effect of verbal prompting on elders' 10-year longitudinal change in everyday cognition. Differential effects of prompting associated with impaired cognitive status were also examined. METHOD At baseline, 2,802 participants (mean age = 73.6 years, mean education = 13.5 years) from the Advanced Cognitive Training for Independent and Vital Elderly trial were classified as unimpaired, having amnestic mild cognitive impairment (MCI) or nonamnestic MCI, based on psychometric algorithm. Participants were given the Observed Tasks of Daily Living (OTDL; a behavioral measure with tasks involving medication management/finances/telephone use) at baseline and at 1-, 2-, 3-, 5-, and 10-year follow-ups. When participants said ""I don't know"" or did not respond to an item, they received a standardized verbal prompt. At each occasion, unprompted (sum of items correct without prompting) and prompted (sum of items correct including both prompted and unprompted) scores were derived for each participant. Multilevel modeling, adjusting for demographics/health/training group, was used to determine the trajectories of OTDL performance. RESULTS Persons with MCI performed at lower levels than those who were unimpaired (amnestic < nonamnestic < unimpaired), and for all groups, prompted performance exceeded unprompted in all years. There was differential performance of the prompting conditions over time; prompted performance, unlike unprompted, was relatively protected from age-related decline, and persons with MCI experienced greater improvement due to prompting. CONCLUSIONS Very simple prompting appears to enhance and maintain performance on a task of everyday cognition over 10 years for both unimpaired and mildly impaired older adults.","{'model': 'tldr@v2.0.0', 'text': 'Very simple prompting appears to enhance and maintain performance on a task of everyday cognition over 10 years for both unimpaired and mildly impaired older adults.'}",https://europepmc.org/articles/pmc3935329?pdf=render
-molecularly imprinted polymer based sensors for medical applications,Yeşeren Saylan,"Sensors have been extensively used owing to multiple advantages, including exceptional sensing performance, user-friendly operation, fast response, high sensitivity and specificity, portability, and real-time analysis. In recent years, efforts in sensor realm have expanded promptly, and it has already presented a broad range of applications in the fields of medical, pharmaceutical and environmental applications, food safety, and homeland security. In particular, molecularly imprinted polymer based sensors have created a fascinating horizon for surface modification techniques by forming specific recognition cavities for template molecules in the polymeric matrix. This method ensures a broad range of versatility to imprint a variety of biomolecules with different size, three dimensional structure, physical and chemical features. In contrast to complex and time-consuming laboratory surface modification methods, molecular imprinting offers a rapid, sensitive, inexpensive, easy-to-use, and highly selective approaches for sensing, and especially for the applications of diagnosis, screening, and theranostics. Due to its physical and chemical robustness, high stability, low-cost, and reusability features, molecularly imprinted polymer based sensors have become very attractive modalities for such applications with a sensitivity of minute structural changes in the structure of biomolecules. This review aims at discussing the principle of molecular imprinting method, the integration of molecularly imprinted polymers with sensing tools, the recent advances and strategies in molecular imprinting methodologies, their applications in medical, and future outlook on this concept.","{'model': 'tldr@v2.0.0', 'text': 'The principle of molecular imprinting method, the integration of molecularly imprinted polymers with sensing tools, the recent advances and strategies in molecular imprinted methodologies, their applications in medical, and future outlook on this concept are discussed.'}",https://www.mdpi.com/1424-8220/19/6/1279/pdf?version=1552557371
-prompting reflections for integrating self-regulation into teacher technology education,Tova Michalsky,"Background Technology represents a major topic in educational research. Nevertheless, a gap in the research remains concerning how teachers can bring technology into the classroom. This study focuses on the technological pedagogical content knowledge (TPCK) framework, which aims to consolidate the multidisciplinary professional knowledge related to technology, pedagogy, and content that teachers need so that they can teach and students can learn effectively using technology tools. Purpose The goal of the present study was to investigate the value of modification reflection prompts (“think ahead”) as a complementary reflective framework during the teacher preparatory program, beyond the more traditional judgment reflection prompts (“think back”). In particular, we examined how preservice science teachers may capitalize on learning from modification prompts versus judgment prompts versus both (“think back” and “think ahead”), compared with learning from generic prompts (“stop and think”) based on the IMPROVE model oriented to TPCK. We examined these four treatments’ effects on preservice teachers’ actual design of science lessons and development of their own self-reflection abilities. Participants Participants were 199 first-year preservice science teachers in their preparatory programs at a university in central Israel. Research Design We created a quasi-experimental opportunity for four groups of preservice science teachers to systematically contemplate ready-made TPCK-oriented lesson designs. Each used one of four different reflective methods (the independent variable): modification, judgment, combined modification+judgment, or generic prompts. Then we examined the differential contribution of these treatment methods to the two dependent variables: (1) preservice teachers’ skills for designing actual science lessons and (2) their judgment-type and modification-type self-reflection ability regarding the planning, monitoring, and evaluation phases of their lesson-design process. Data Collection and Analysis Data were scored by coding schemes and were analyzed by multivariate analysis of variance and follow-up analyses of variance with repeated measures. Findings Results indicated that preservice teachers who contemplated a combination of both judgment and modification reflections in treatment improved more in their lesson-design skills and in their self-reflection ability (of both types at the three phases), compared with preservice teachers who contemplated only a single type of reflective prompt (generic or only judgment or modification). Lasting effects (after a semester without the IMPROVE model, prompts, or TPCK focus) revealed that the combined approach continued to significantly outperform the single approaches. Recommendations The current study reinterprets the instructional-reflective framework of teacher education programs to include modification reflection too as a means of developing preservice teachers’ capacity to integrate technology in their lesson designs.",,https://journals.sagepub.com/doi/pdf/10.1177/016146811511700507
-cold-start data selection for better few-shot language model fine-tuning: a prompt-based uncertainty propagation approach,Yue Yu,"We present PATRON, a prompt-based data selection method for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON will be published upon acceptance.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based data selection method for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available, and a partition-then-rewrite strategy to promote sample diversity when querying for annotations is designed.'}",https://aclanthology.org/2023.acl-long.141.pdf
-current state and future directions of technology-based ecological momentary assessment and intervention for major depressive disorder: a systematic review,D. Colombo,"Ecological momentary assessment (EMA) and ecological momentary intervention (EMI) are alternative approaches to retrospective self-reports and face-to-face treatments, and they make it possible to repeatedly assess patients in naturalistic settings and extend psychological support into real life. The increase in smartphone applications and the availability of low-cost wearable biosensors have further improved the potential of EMA and EMI, which, however, have not yet been applied in clinical practice. Here, we conducted a systematic review, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, to explore the state of the art of technology-based EMA and EMI for major depressive disorder (MDD). A total of 33 articles were included (EMA = 26; EMI = 7). First, we provide a detailed analysis of the included studies from technical (sampling methods, duration, prompts), clinical (fields of application, adherence rates, dropouts, intervention effectiveness), and technological (adopted devices) perspectives. Then, we identify the advantages of using information and communications technologies (ICTs) to extend the potential of these approaches to the understanding, assessment, and intervention in depression. Furthermore, we point out the relevant issues that still need to be addressed within this field, and we discuss how EMA and EMI could benefit from the use of sensors and biosensors, along with recent advances in machine learning for affective modelling.","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the state of the art of technology-based EMA and EMI for major depressive disorder to identify the advantages of using information and communications technologies (ICTs) to extend the potential of these approaches to the understanding, assessment, and intervention in depression.'}",https://www.mdpi.com/2077-0383/8/4/465/pdf?version=1555583492
-large language models are state-of-the-art evaluators of translation quality,Tom Kocmi,"We describe GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without. In our evaluation, we focus on zero-shot prompting, comparing four prompt variants in two modes, based on the availability of the reference. We investigate seven versions of GPT models, including ChatGPT. We show that our method for translation quality assessment only works with GPT 3.5 and larger models. Comparing to results from WMT22’s Metrics shared task, our method achieves state-of-the-art accuracy in both modes when compared to MQM-based human labels. Our results are valid on the system level for all three WMT22 Metrics shared task language pairs, namely English into German, English into Russian, and Chinese into English. This provides a first glimpse into the usefulness of pre-trained, generative large language models for quality assessment of translations. We publicly release all our code and prompt templates used for the experiments described in this work, as well as all corresponding scoring results, to allow for external validation and reproducibility.","{'model': 'tldr@v2.0.0', 'text': 'GEMBA, a GPT-based metric for assessment of translation quality, which works both with a reference translation and without, is described, and achieves state-of-the-art accuracy in both modes when compared to MQM-based human labels.'}",http://arxiv.org/pdf/2302.14520
-teaching chained tasks to students with intellectual disabilities by using video prompting in small group instruction.,Ç. Aykut,"AbstractTeaching students with intellectual disabilities in groups presents several difficulties. Use of technology can reduce some of these difficulties. The literature cites several examples of skill acquisition. The purpose of this study is to teach skills to students with intellectual disabilities by using video prompting. A multiple-probe design of single-subject design methods has been applied to this study. Three students with intellectual disabilities participated as subjects in the study. The findings provided evidence for the effectiveness of video prompting in improving teaching skills. The subjects were able to apply and perform these skills away from the teaching environment, and were able to maintain their improved skills for the following 6 months.Key WordsAcquisition of Skills, Intellectual Disabilities, Teaching Skills, Video Prompting.Skills formed by sequential single-staged behaviors comprising complicated end behaviors are called chained tasks (Tekin-?ftar & K?rcaali-?ftar, 2004). Chained tasks are of different types, for example, self-care skills such as eating, toileting, and dressing; gross motor skills such as walking, running, and jumping rope; fine motor skills such as plucking, tearing, cutting, and painting; and daily living skills such as phoning and shopping. These skills, acquired from early childhood on, positively affect both children's school years and future life as an independent individual (Mechling, Pridgen, & Cronin, 2005; Snell & Brown, 1993; Varol, 2005).Modeling is a teaching technique used frequently in teaching new behaviors to individuals with intellectual disabilities; its effectiveness has been supported by numerous studies (Rehfeldt, Dahman, Young, Cherry, & Davis, 2003). Six types of video technology issues have been discussed in the literature: a) video feedback, b) video modeling, c) video self-modeling, d) subjective point of view video modeling, e) video prompting, and f) computer-based video instruction (Mechling, 2005; Murray & Noland, 2013; Oncul & Ozkan, 2010).Video prompting requires the respondent to watch a segment of a video recording and actively respond to the prompt in the video. In contrast to other video teaching methods, video prompting does not require the respondent to watch a video from beginning to the end before performing the behavior. Instead, the respondent is expected to watch a segment of the video and give an immediate active response to the video prompt. Depending on the subjects' response, the video recording is continued or repeated by the researcher the respondent (Mechling, 2005; Norman, Collins, & Schuster, 2001).Chained tasks are more complicated skills consisting of more than one sequential, single-step behaviors. Performance of these tasks takes time. Video prompting appears to be an effective way to enable students with intellectual disabilities to acquire target behaviors (Horn et al., 2008). Thus, the video prompting technique has been used in this study. The literature, though limited, contains studies with experimental designs and a single subject, carried out by using video prompting to allow students to acquire chained tasks. Previous studies have used the video prompting technique to help intellectually disabled or developmentally delayed students acquire daily living skills and self-care skills such as setting a table (Norman et al., 2001), organizing food bought from a green-grocer (Cannella-Malone et al., 2006), cooking (Graves, Collins, Schuster, & Kleinert, 2005), washing clothes (Horn et al., 2008), and washing dishes (Sigafoos et al., 2007). At the end of these studies, the target chained tasks were acquired by these students.Planning the training environment is at least as important as planning the training syllabus for helping students with disabilities acquire new behaviors. In general, one-to-one training for moderately and severely disabled students is effective; however, group instruction is another effective alternative with several advantages over one-to-one teaching: a) teachers can train more than one student, b) group training requires fewer personnel and less teaching time, c) students can earn more functional skills in a less-restrictive environment, d) they can develop appropriate skills for communicating with their peers, and e) they can acquire extra knowledge by observing other students in the group (Collins, Gast, Ault, & Wolery, 1991; Ledford, Gast, Luscre, & Ayres, 2008). …",,
-visual prompt based personalized federated learning,Guang-Ming Li,"As a popular paradigm of distributed learning, personalized federated learning (PFL) allows personalized models to improve generalization ability and robustness by utilizing knowledge from all distributed clients. Most existing PFL algorithms tackle personalization in a model-centric way, such as personalized layer partition, model regularization, and model interpolation, which all fail to take into account the data characteristics of distributed clients. In this paper, we propose a novel PFL framework for image classification tasks, dubbed pFedPT, that leverages personalized visual prompts to implicitly represent local data distribution information of clients and provides that information to the aggregation model to help with classification tasks. Specifically, in each round of pFedPT training, each client generates a local personalized prompt related to local data distribution. Then, the local model is trained on the input composed of raw data and a visual prompt to learn the distribution information contained in the prompt. During model testing, the aggregated model obtains prior knowledge of the data distributions based on the prompts, which can be seen as an adaptive fine-tuning of the aggregation model to improve model performances on different clients. Furthermore, the visual prompt can be added as an orthogonal method to implement personalization on the client for existing FL methods to boost their performance. Experiments on the CIFAR10 and CIFAR100 datasets show that pFedPT outperforms several state-of-the-art (SOTA) PFL algorithms by a large margin in various settings.","{'model': 'tldr@v2.0.0', 'text': 'A novel PFL framework for image classification tasks, dubbed pFedPT, is proposed that leverages personalized visual prompts to implicitly represent local data distribution information of clients and provides that information to the aggregation model to help with classification tasks.'}",http://arxiv.org/pdf/2303.08678
-memobert: pre-training model with prompt-based learning for multimodal emotion recognition,Jinming Zhao,"Multimodal emotion recognition study is hindered by the lack of labelled corpora in terms of scale and diversity, due to the high annotation cost and label ambiguity. In this paper, we propose a multimodal pre-training model MEmoBERT for multimodal emotion recognition, which learns multimodal joint representations through self-supervised learning from a self-collected large-scale unlabeled video data that come in sheer volume. Furthermore, unlike the conventional ""pre-train, finetune"" paradigm, we propose a prompt-based method that reformulates the downstream emotion classification task as a masked text prediction one, bringing the downstream task closer to the pre-training. Extensive experiments on two benchmark datasets, IEMOCAP and MSP-IMPROV, show that our proposed MEmoBERT significantly enhances emotion recognition performance.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method is proposed that reformulates the downstream emotion classification task as a masked text prediction one, bringing the downstream task closer to the pre-training.'}",https://arxiv.org/pdf/2111.00865
-toxigen: a large-scale machine-generated dataset for adversarial and implicit hate speech detection,Thomas Hartvigsen,"Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language.To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a massive pretrained language model. Controlling machine generation in this way allows ToxiGen to cover implicitly toxic text at a larger scale, and about more demographic groups, than previous resources of human-written text. We conduct a human evaluation on a challenging subset of ToxiGen and find that annotators struggle to distinguish machine-generated text from human-written language. We also find that 94.5% of toxic examples are labeled as hate speech by human annotators. Using three publicly-available datasets, we show that finetuning a toxicity classifier on our data improves its performance on human-written data substantially. We also demonstrate that ToxiGen can be used to fight machine-generated toxicity as finetuning improves the classifier significantly on our evaluation subset.","{'model': 'tldr@v2.0.0', 'text': 'ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups, is created and it is demonstrated that finetuning a toxicity classifier on data improves its performance on human-written data substantially.'}",http://arxiv.org/pdf/2203.09509
-using smartphones and health apps to change and manage health behaviors: a population-based survey,C. Ernsting,"Background Chronic conditions are an increasing challenge for individuals and the health care system. Smartphones and health apps are potentially promising tools to change health-related behaviors and manage chronic conditions. Objective The aim of this study was to explore (1) the extent of smartphone and health app use, (2) sociodemographic, medical, and behavioral correlates of smartphone and health app use, and (3) associations of the use of apps and app characteristics with actual health behaviors. Methods A population-based survey (N=4144) among Germans, aged 35 years and older, was conducted. Sociodemographics, presence of chronic conditions, health behaviors, quality of life, and health literacy, as well as the use of the Internet, smartphone, and health apps were assessed by questionnaire at home visit. Binary logistic regression models were applied. Results It was found that 61.25% (2538/4144) of participants used a smartphone. Compared with nonusers, smartphone users were younger, did more research on the Internet, were more likely to work full-time and more likely to have a university degree, engaged more in physical activity, and less in low fat diet, and had a higher health-related quality of life and health literacy. Among smartphone users, 20.53% (521/2538) used health apps. App users were younger, less likely to be native German speakers, did more research on the Internet, were more likely to report chronic conditions, engaged more in physical activity, and low fat diet, and were more health literate compared with nonusers who had a smartphone. Health apps focused on smoking cessation (232/521, 44.5%), healthy diet (201/521, 38.6%), and weight loss (121/521, 23.2%). The most common app characteristics were planning (264/521, 50.7%), reminding (188/521, 36.1%), prompting motivation (179/521 34.4%), and the provision of information (175/521, 33.6%). Significant associations were found between planning and the health behavior physical activity, between feedback or monitoring and physical activity, and between feedback or monitoring and adherence to doctor’s advice. Conclusions Although there were many smartphone and health app users, a substantial proportion of the population was not engaged. Findings suggest age-related, socioeconomic-related, literacy-related, and health-related disparities in the use of mobile technologies. Health app use may reflect a user’s motivation to change or maintain health behaviors. App developers and researchers should take account of the needs of older people, people with low health literacy, and chronic conditions.","{'model': 'tldr@v2.0.0', 'text': 'App users were younger, less likely to be native German speakers, did more research on the Internet, were more likely to report chronic conditions, engaged more in physical activity, and low fat diet, and were more health literate compared with nonusers who had a smartphone.'}",https://www.jmir.org/2017/4/e101/PDF
-prompt-based text entailment for low-resource named entity recognition,Dongfang Li,"Pre-trained Language Models (PLMs) have been applied in NLP tasks and achieve promising results. Nevertheless, the fine-tuning procedure needs labeled data of the target domain, making it difficult to learn in low-resource and non-trivial labeled scenarios. To address these challenges, we propose Prompt-based Text Entailment (PTE) for low-resource named entity recognition, which better leverages knowledge in the PLMs. We first reformulate named entity recognition as the text entailment task. The original sentence with entity type-specific prompts is fed into PLMs to get entailment scores for each candidate. The entity type with the top score is then selected as final label. Then, we inject tagging labels into prompts and treat words as basic units instead of n-gram spans to reduce time complexity in generating candidates by n-grams enumeration. Experimental results demonstrate that the proposed method PTE achieves competitive performance on the CoNLL03 dataset, and better than fine-tuned counterparts on the MIT Movie and Few-NERD dataset in low-resource settings.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2211.03039
-consprompt: easily exploiting contrastive samples for few-shot prompt learning,Jinta Weng,"Prompt learning recently become an effective linguistic tool to motivate the PLMs’ knowledge on few-shot-setting tasks. However, studies have shown the lack of robustness still exists in prompt learning, since suitable initialization of continuous prompt and expert-first manual prompt are essential in fine-tuning process. What is more, human also utilize their comparative ability to motivate their existing knowledge for distinguishing different examples. Motivated by this, we explore how to use contrastive samples to strengthen prompt learning. In detail, we first propose our model ConsPrompt combining with prompt encoding network, contrastive sampling module, and contrastive scoring module. Subsequently, two sampling strategies, similarity-based and label-based strategies, are introduced to realize dif-ferential contrastive learning. The effectiveness of proposed ConsPrompt is demonstrated in five different few-shot learning tasks and shown the similarity-based sampling strategy is more effective than label-based in combining contrastive learning. Our results also ex-hibits the state-of-the-art performance and robustness in different few-shot settings, which proves that the ConsPrompt could be assumed as a better knowledge probe to motivate PLMs. As far as we could reach, this is the first work exploring how to use contrastive learning approach and suitable contrastive samples to enhance prompt-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'The model ConsPrompt is proposed, which ex-hibits the state-of-the-art performance and robustness in different few-shot settings, which proves that the Cons Prompt could be assumed as a better knowledge probe to motivate PLMs.'}",https://arxiv.org/pdf/2211.04118
-towards informative few-shot prompt with maximum information gain for in-context learning,Hongfu Liu,"Large Language models (LLMs) possess the capability to engage In-context Learning (ICL) by leveraging a few demonstrations pertaining to a new downstream task as conditions. However, this particular learning paradigm suffers from high instability stemming from substantial variances induced by factors such as the input distribution of selected examples, their ordering, and prompt formats. In this work, we demonstrate that even when all these factors are held constant, the random selection of examples still results in high variance. Consequently, we aim to explore the informative ability of data examples by quantifying the Information Gain (IG) obtained in prediction after observing a given example candidate. Then we propose to sample those with maximum IG. Additionally, we identify the presence of template bias, which can lead to unfair evaluations of IG during the sampling process. To mitigate this bias, we introduce Calibration Before Sampling strategy. The experimental results illustrate that our proposed method can yield an average relative improvement of 14.3% across six classification tasks using three LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to explore the informative ability of data examples by quantifying the Information Gain obtained in prediction after observing a given example candidate, and proposes to sample those with maximum IG.'}",https://arxiv.org/pdf/2310.08923
-virtual prompt pre-training for prototype-based few-shot relation extraction,Kai He,,,http://manuscript.elsevier.com/S0957417422019455/pdf/S0957417422019455.pdf
-few-shot multi-modal sentiment analysis with prompt-based vision-aware language modeling,Yang Yu,"As a hot study topic in natural language processing, affec-tive computing and multimedia analysis, multi-modal senti-ment analysis (MSA) is widely explored on aspect-level and sentence-level tasks. However, the existing studies normally rely on a lot of annotated multi-modal data, which are difficult to collect due to the massive expenditure of manpower and re-sources, especially in some open-ended and fine-grained do-mains. Therefore, it is necessary to investigate the few-shot scenario for MSA. In this paper, we propose a prompt-based vision-aware language modeling (PVLM) approach to MSA, which only requires a few supervised data. Specifically, our PVLM can incorporate the visual information into pre-trained language model and leverage prompt tuning to bridge the gap between masked language prediction in pre-training and MSA tasks. Systematic experiments on three aspect-level and two sentence-level datasets of MSA demonstrate the effectiveness of our few-shot approach.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based vision-aware language modeling (PVLM) approach to MSA, which only requires a few supervised data and can incorporate the visual information into pre-trained language model and leverage prompt tuning to bridge the gap between masked language prediction in pre-training and MSA tasks.'}",
-unified multi-modal pre-training for few-shot sentiment analysis with prompt-based learning,Yang Yu,"Multi-modal sentiment analysis (MSA) has become more and more attractive in both academia and industry. The conventional studies normally require massive labeled data to train the deep neural models. To alleviate the above issue, in this paper, we conduct few-shot MSA with quite a small number of labeled samples. Inspired by the success of textual prompt-based fine-tuning (PF) approaches in few-shot scenario, we introduce a multi-modal prompt-based fine-tuning (MPF) approach. To narrow the semantic gap between language and vision, we propose unified pre-training for multi-modal prompt-based fine-tuning (UP-MPF) with two stages. First, in unified pre-training stage, we employ a simple and effective task to obtain coherent vision-language representations from fixed pre-trained language models (PLMs), i.e., predicting the rotation direction of the input image with a prompt phrase as input concurrently. Second, in multi-modal prompt-based fine-tuning, we freeze the visual encoder to reduce more parameters, which further facilitates few-shot MSA. Extensive experiments and analysis on three coarse-grained and three fine-grained MSA datasets demonstrate the better performance of our UP-MPF against the state-of-the-art of PF, MSA, and multi-modal pre-training approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes unified pre-training for multi-modal prompt-based fine-tuning (UP-MPF) with two stages, and employs a simple and effective task to obtain coherent vision-language representations from fixed pre-trained language models (PLMs).'}",
-commonsense knowledge-aware prompt tuning for few-shot nota relation classification,Bo Lv,"Compared with the traditional few-shot task, the few-shot none-of-the-above (NOTA) relation classification focuses on the realistic scenario of few-shot learning, in which a test instance might not belong to any of the target categories. This undoubtedly increases the task’s difficulty because given only a few support samples, this cannot represent the distribution of NOTA categories in space. The model needs to make full use of the syntactic information and word meaning information learned in the pre-training stage to distinguish the NOTA category and the support sample category in the embedding space. However, previous fine-tuning methods mainly focus on optimizing the extra classifiers (on top of pre-trained language models (PLMs)) and neglect the connection between pre-training objectives and downstream tasks. In this paper, we propose the commonsense knowledge-aware prompt tuning (CKPT) method for a few-shot NOTA relation classification task. First, a simple and effective prompt-learning method is developed by constructing relation-oriented templates, which can further stimulate the rich knowledge distributed in PLMs to better serve downstream tasks. Second, external knowledge is incorporated into the model by a label-extension operation, which forms knowledgeable prompt tuning to improve and stabilize prompt tuning. Third, to distinguish the NOTA pairs and positive pairs in embedding space more accurately, a learned scoring strategy is proposed, which introduces a learned threshold classification function and improves the loss function by adding a new term focused on NOTA identification. Experiments on two widely used benchmarks (FewRel 2.0 and Few-shot TACRED) show that our method is a simple and effective framework, and a new state of the art is established in the few-shot classification field.","{'model': 'tldr@v2.0.0', 'text': 'The commonsense knowledge-aware prompt tuning (CKPT) method is proposed, a simple and effective prompt-learning method is developed by constructing relation-oriented templates, which can further stimulate the rich knowledge distributed in PLMs to better serve downstream tasks.'}",https://www.mdpi.com/2076-3417/12/4/2185/pdf?version=1645269904
-dual context-guided continuous prompt tuning for few-shot learning,Jie Zhou,"Prompt-based paradigm has shown its competitive performance in many NLP tasks. However, its success heavily depends on prompt design, and the effectiveness varies upon the model and training data. In this paper, we propose a novel dual context-guided continuous prompt (DCCP) tuning method. To explore the rich contextual information in language structure and close the gap between discrete prompt tuning and continuous prompt tuning, DCCP introduces two auxiliary training objectives and constructs input in a pair-wise fashion.Experimental results demonstrate that our method is applicable to many NLP tasks, and can often outperform existing prompt tuning methods by a large margin in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'To explore the rich contextual information in language structure and close the gap between discrete prompt tuning and continuous prompt tuning, DCCP introduces two auxiliary training objectives and constructs input in a pair-wise fashion.'}",https://aclanthology.org/2022.findings-acl.8.pdf
-a dual prompt learning framework for few-shot dialogue state tracking,Yuting Yang,"Dialogue State Tracking (DST) module is an essential component of task-oriented dialog systems to understand users’ goals and needs. Collecting dialogue state labels including slots and values can be costly, requiring experts to annotate all (slot, value) information for each turn in dialogues. It is also difficult to define all possible slots and values in advance, especially with the wide application of dialogue systems in more and more new-rising applications. In this paper, we focus on improving DST module to generate dialogue states in circumstances with limited annotations and knowledge about slot ontology. To this end, we design a dual prompt learning framework for few-shot DST. The dual framework aims to explore how to utilize the language understanding and generation capabilities of pre-trained language models for DST efficiently. Specifically, we consider the learning of slot generation and value generation as dual tasks, and two kinds of prompts are designed based on this dual structure to incorporate task-related knowledge of these two tasks respectively. In this way, the DST task can be formulated as a language modeling task efficiently under few-shot settings. To evaluate the proposed framework, we conduct experiments on two task-oriented dialogue datasets. The results demonstrate that the proposed method not only outperforms existing state-of-the-art few-shot methods, but also can generate unseen slots. It indicates that DST-related knowledge can be probed from pre-trained language models and utilized to address low-resource DST efficiently with the help of prompt learning.","{'model': 'tldr@v2.0.0', 'text': 'A dual prompt learning framework for few-shot DST is designed that indicates that DST-related knowledge can be probed from pre-trained language models and utilized to address low-resource DST efficiently with the help of prompt learning.'}",https://dl.acm.org/doi/pdf/10.1145/3543507.3583238
-multi-task pre-training of modular prompt for few-shot learning,Tianxiang Sun,"Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present M ulti-task P re-trained M odular P rompt ( MP 2 ) to boost prompt tuning for few-shot learning. MP 2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP 2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP 2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents M ulti-task P re-trained M odular P rompt ( MP 2) to boost prompt tuning for few-shot learning, and demonstrates that MP 2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts.'}",http://arxiv.org/pdf/2210.07565
-visual prompt tuning for few-shot text classification,Jing Wen,"Deploying large-scale pre-trained models in the prompt-tuning paradigm has demonstrated promising performance in few-shot learning. Particularly, vision-language pre-training models (VL-PTMs) have been intensively explored in various few-shot downstream tasks. However, most existing works only apply VL-PTMs to visual tasks like image classification, with few attempts being made on language tasks like text classification. In few-shot text classification, a feasible paradigm for deploying VL-PTMs is to align the input samples and their category names via the text encoders. However, it leads to the waste of visual information learned by the image encoders of VL-PTMs. To overcome this drawback, we propose a novel method named Visual Prompt Tuning (VPT). To our best knowledge, this method is the first attempt to deploy VL-PTM in few-shot text classification task. The main idea is to generate the image embeddings w.r.t. category names as visual prompt and then add them to the aligning process. Extensive experiments show that our VPT can achieve significant improvements under both zero-shot and few-shot settings. Importantly, our VPT even outperforms the most recent prompt-tuning methods on five public text classification datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel method named Visual Prompt Tuning (VPT), which is the first attempt to deploy VL-PTM in few-shot text classification task and outperforms the most recent prompt-tuning methods on five public text classification datasets.'}",
-idiapers @ causal news corpus 2022: efficient causal relation identification through a prompt-based few-shot approach,S. Burdisso,"In this paper, we describe our participation in the subtask 1 of CASE-2022, Event Causality Identification with Casual News Corpus. We address the Causal Relation Identification (CRI) task by exploiting a set of simple yet complementary techniques for fine-tuning language models (LMs) on a few annotated examples (i.e., a few-shot configuration).We follow a prompt-based prediction approach for fine-tuning LMs in which the CRI task is treated as a masked language modeling problem (MLM). This approach allows LMs natively pre-trained on MLM tasks to directly generate textual responses to CRI-specific prompts.We compare the performance of this method against ensemble techniques trained on the entire dataset.Our best-performing submission was fine-tuned with only 256 instances per class, 15.7% of the all available data, and yet obtained the second-best precision (0.82), third-best accuracy (0.82), and an F1-score (0.85) very close to what was reported by the winner team (0.86).","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the Causal Relation Identification (CRI) task by exploiting a set of simple yet complementary techniques for fine-tuning language models (LMs) on a few annotated examples (i.e., a few-shot configuration) in which the CRI task is treated as a masked language modeling problem (MLM).'}",http://arxiv.org/pdf/2209.03895
-p4e: few-shot event detection as prompt-guided identification and localization,Sha Li,"We propose P4E, an identify-and-localize event detection framework that integrates the best of few-shot prompting and structured prediction. Our framework decomposes event detection into an identification task and a localization task. For the identification task, which we formulate as multi-label classification, we leverage cloze-based prompting to align our objective with the pre-training task of language models, allowing our model to quickly adapt to new event types. We then employ an event type-agnostic sequence labeling model to localize the event trigger conditioned on the identification output. This heterogeneous model design allows P4E to quickly learn new event types without sacrificing the ability to make structured predictions. Our experiments demonstrate the effectiveness of our proposed design, and P4E shows superior performance for few-shot event detection on benchmark datasets FewEvent and MAVEN and comparable performance to SOTA for fully-supervised event detection on ACE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes P4E, an identify-and-localize event detection framework that integrates the best of few-shot prompting and structured prediction, and shows superior performance for few- shot event detection on benchmark datasets FewEvent and MAVEN and comparable performance to SOTA for fully-supervised event Detection on ACE.'}",
-few-shot natural language inference generation with pdd: prompt and dynamic demonstration,Kaijian Li,"Natural Language Inference Generation task is to generate a text hypothesis given a text premise and a logical relation between the two. This task can be used in data augmentation and controllable text generation in practice. In this paper, we propose language models with prompt and dynamic demonstration (LM-PDD) to tackle this problem in few-shot settings. Our framework outperforms standard fine-tuned models with low resource, achieving an average 8% absolute improvement on SNLI and MNLI datasets, and the results on 13 natural language classification tasks also show that our dynamic demonstration method has good generalizability.","{'model': 'tldr@v2.0.0', 'text': ""This paper's framework outperforms standard fine-tuned models with low resource, achieving an average 8% absolute improvement on SNLI and MNLI datasets, and the results on 13 natural language classification tasks show that the dynamic demonstration method has good generalizability.""}",https://arxiv.org/pdf/2205.10593
-discriminative language model as semantic consistency scorer for prompt-based few-shot text classification,Zhipeng Xie,"This paper proposes a novel prompt-based finetuning method (called DLM-SCS) for few-shot text classification by utilizing the discriminative language model ELECTRA that is pretrained to distinguish whether a token is original or generated. The underlying idea is that the prompt instantiated with the true label should have higher semantic consistency score than other prompts with false labels. Since a prompt usually consists of several components (or parts), its semantic consistency can be decomposed accordingly. The semantic consistency of each component is then computed by making use of the pretrained ELECTRA model, without introducing extra parameters. Extensive experiments have shown that our model outperforms several state-of-the-art prompt-based few-shot methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based finetuning method for few-shot text classification by utilizing the discriminative language model ELECTRA that is pretrained to distinguish whether a token is original or generated is proposed.'}",http://arxiv.org/pdf/2210.12763
-dfs-ner: description enhanced few-shot ner via prompt learning and meta-learning,Huinan Huang,"Named Entity Recognition (NER) is a very common task in many social good related domains. Recently, deep learning based NER has gradually matured, but still faces the scarcity problem of labeled data in specific domains. Therefore, researchers focus on few-shot NER to reduce the model’s data dependence and enhance the transferability of the model. However, existing works usually cannot adapt to new entity types and are prone to the so-called negative transfer problem. Therefore, in this paper we propose a type- Description-enhanced Few Shot NER model, called DFS-NER, which effectively integrates the prompt learning paradigm and the meta-learning framework. DFS-NER performs well under frozen pre-training model parameters through continuous templates. It realizes efficient source domain training and target domain parameter fine-tuning through the metalearning framework. We enhance the robustness of entity- type prototype representations by introducing word-word- level and word-type-level contrastive learning objectives and capsule networks as the induction module. Simultaneously, based on discrete prompt learning, a masked-language model learning objective guided by type description is proposed, which can well absorb the semantic information of entity types. Experiments on commonly used datasets, including, SNIPS, Few-NERD, and MIT Movie show that DFS-NER basically surpasses baseline models and achieves the state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a type- Description-enhanced Few Shot NER model, called DFS-NER, which effectively integrates the prompt learning paradigm and the meta-learning framework and proposes a masked-language model learning objective guided by type description, which can well absorb the semantic information of entity types.'}",
-a prompt-based few-shot machine reading comprehension model for intelligent bridge management,Luyi Zhang,"Bridge inspection reports are an important data source in the bridge management process, and they contain a large amount of fine-grained information. However, the research on machine reading comprehension (MRC) methods for this field is insufficient, and annotating large scale domain-specific corpus is time-consuming. This paper presented a novel prompt-based few-shot MRC approach for intelligent bridge management. The proposed model uses the pretrained model MacBERT as backbone. The prompt templates are designed based on some domain-specific heuristic rules. The experimental results show that our model outperforms the baseline models in different few-shot settings. The proposed model can provide technical support for the construction of automatic question answering system in the field of bridge management.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based few-shot MRC approach for intelligent bridge management that uses the pretrained model MacBERT as backbone and prompt templates are designed based on some domain-specific heuristic rules.'}",
-a study on prompt-based few-shot learning methods for belief state tracking in task-oriented dialog systems,Debjoy Saha,"We tackle the Dialogue Belief State Tracking(DST) problem of task-oriented conversational systems. Recent approaches to this problem leveraging Transformer-based models have yielded great results. However, training these models is expensive, both in terms of computational resources and time. Additionally, collecting high quality annotated dialogue datasets remains a challenge for researchers because of the extensive annotation required for training these models. Driven by the recent success of pre-trained language models and prompt-based learning, we explore prompt-based few-shot learning for Dialogue Belief State Tracking. We formulate the DST problem as a 2-stage prompt-based language modelling task and train language models for both tasks and present a comprehensive empirical analysis of their separate and joint performance. We demonstrate the potential of prompt-based methods in few-shot learning for DST and provide directions for future improvement.","{'model': 'tldr@v2.0.0', 'text': 'The DST problem is formulated as a 2-stage prompt-based language modelling task and language models for both tasks are trained and a comprehensive empirical analysis of their separate and joint performance is presented.'}",http://arxiv.org/pdf/2204.08167
-prompt and contrastive learning for few-shot sentiment classification,Fei Wang,"Sentiment classification is a hot topic in the field of natural language processing. Currently, state-of-the-art classification models follow two steps: pre-training a large language model on upstream tasks, and then using human-labeled data to fine-tune a task-related model. However, there is a large gap between the upstream tasks of the pre-trained model and the downstream tasks being performed, resulting in the need for more labeled data to achieve excellent performance. Manually annotating data is expensive. In this paper, we propose a few-shot sentiment classification method based on Prompt and Contrastive Learning (PCL), which can significantly improve the performance of large-scale pre-trained language models in low-data and high-data regimes. Prompt learning aims to alleviate the gap between upstream and downstream tasks, and the contrastive learning is designed to capture the inter-class and intra-class distribution patterns of labeled data. Thanks to the integration of the two strategies, PCL markedly exceeds baselines with low resources. Extensive experiments on three datasets show that our method has outstanding performance in the few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a few-shot sentiment classification method based on Prompt and Contrastive Learning (PCL), which can significantly improve the performance of large-scale pre-trained language models in low-data and high-data regimes.'}",
-ti-prompt: towards a prompt tuning method for few-shot threat intelligence twitter classification*,Yizhe You,"Obtaining the latest Threat Intelligence (TI) via Twitter has become one of the most important methods for defenders to catch up with emerging cyber threats. Existing TI Twitter classification works mainly based on supervised learning methods. Such approaches require large amounts of annotated data and are difficult to be transferred to other TI Twitter classification tasks. This paper proposes a prompt-based method for classifying TI on Twitter, named TI-Prompt. TI-Prompt lever-ages the prompt-tuning method with two templates in different TI Twitter classification tasks. TI-Prompt also uses a semantic similarity-based approach to automatically enrich the prompt verbalizer without expert knowledge and a verbalizer refinement method to calibrate the verbalizer based on the training data. We evaluate TI-Prompt with binary and multi-classification tasks on two Twitter Threat Intelligence datasets. Evaluation results show that the proposed TI-Prompt improves 5-10% over the best performance of previous supervised learning methods under the few-shot settings. Compared to the general prompt-tuning methods, the proposed prompt-tuning templates can also improve the classification performance by 2–5%. Meanwhile, the proposed verbalizer enrichment method and refinement method improve classification accuracy by 1–4% compared with the general single-word verbalizer prompt method. Therefore, TI-Prompt can be extended to other Threat Intelligence classification tasks without requiring large amounts of training data, significantly reducing the annotation cost.","{'model': 'tldr@v2.0.0', 'text': 'Evaluation results show that the proposed TI-Prompt improves 5-10% over the best performance of previous supervised learning methods under the few-shot settings, and can be extended to other Threat Intelligence classification tasks without requiring large amounts of training data, significantly reducing the annotation cost.'}",
-adaptive prompt learning-based few-shot sentiment analysis,Pengfei Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results on FewCLUE datasets demonstrate that the proposed method AP can effectively construct appropriate adaptive prompt regardless of the quality of hand-crafted prompt and outperform the state-of-the-art baselines.'}",https://arxiv.org/pdf/2205.07220
-augprompt: knowledgeable augmented-trigger prompt for few-shot event classification,Chengyu Song,,,
-prompt-based few-shot learning for table-based fact verification,Lei Hou,"Natural language processing has been a hot topic of research, but existing research is mainly limited to unstructured information such as natural language sentences and documents, and less research has been done on structured information such as tables. The main object of this paper is the table-based fact verification task, under which there is only one TABFACT dataset. Most of the existing methods on this dataset are based on pre-trained models and need to be fine-tuned again if a new dataset appears. And some previous work on natural language sentences has shown that prompt approach can achieve good performance with few samples. Therefore, in this paper, we adopt the prompt approach for experiments on the table fact detection task by manually designing templates for hinting the pre-trained model. Meanwhile, to enhance the generalization of the model, we introduce a multi-pair mapping relationship in the Answer Engineering phase. Experiments on the TABFACT dataset show that using the prompt method for table-based fact verification task in the case of few samples can be effective, providing a new way for optimizing table-related tasks in the case of few samples.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on the TABFACT dataset show that using the prompt method for table-based fact verification task in the case of few samples can be effective, providing a new way for optimizing table-related tasks in the Case ofFew samples.'}",
-investigating prompt learning for chinese few-shot text classification with pre-trained language models,Chengyu Song,"Text classification aims to assign predefined labels to unlabeled sentences, which tend to struggle in real-world applications when only a few annotated samples are available. Previous works generally focus on using the paradigm of meta-learning to overcome the classification difficulties brought by insufficient data, where a set of auxiliary tasks is given. Accordingly, prompt-based approaches are proposed to deal with the low-resource issue. However, existing prompt-based methods mainly focus on English tasks, which generally apply English pretrained language models that can not directly adapt to Chinese tasks due to structural and grammatical differences. Thus, we propose a prompt-based Chinese text classification framework that uses generated natural language sequences as hints, which can alleviate the classification bottleneck well in low-resource scenarios. In detail, we first design a prompt-based fine-tuning together with a novel pipeline for automating prompt generation in Chinese. Then, we propose a refined strategy for dynamically and selectively incorporating demonstrations into each context. We present a systematic evaluation for analyzing few-shot performance on a wide range of Chinese text classification tasks. Our approach makes few assumptions about task resources and expertise and therefore constitutes a powerful, task-independent approach for few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based Chinese text classification framework that uses generated natural language sequences as hints, which can alleviate the classification bottleneck well in low-resource scenarios and constitutes a powerful, task-independent approach for few-shot learning.'}",https://www.mdpi.com/2076-3417/12/21/11117/pdf?version=1667385041
-"few-shot information extraction is here: pre-train, prompt and entail",Eneko Agirre,"Deep Learning has made tremendous progress in Natural Language Processing (NLP), where large pre-trained language models (PLM) fine-tuned on the target task have become the predominant tool. More recently, in a process called prompting, NLP tasks are rephrased as natural language text, allowing us to better exploit linguistic knowledge learned by PLMs and resulting in significant improvements. Still, PLMs have limited inference ability. In the Textual Entailment task, systems need to output whether the truth of a certain textual hypothesis follows from the given premise text. Manually annotated entailment datasets covering multiple inference phenomena have been used to infuse inference capabilities to PLMs. This talk will review these recent developments, and will present an approach that combines prompts and PLMs fine-tuned for textual entailment that yields state-of-the-art results on Information Extraction (IE) using only a small fraction of the annotations. The approach has additional benefits, like the ability to learn from different schemas and inference datasets. These developments enable a new paradigm for IE where the expert can define the domain-specific schema using natural language and directly run those specifications, annotating a handful of examples in the process. A user interface based on this new paradigm will also be presented. Beyond IE, inference capabilities could be extended, acquired and applied from other tasks, opening a new research avenue where entailment and downstream task performance improve in tandem.","{'model': 'tldr@v2.0.0', 'text': 'This talk will present an approach that combines prompts and PLMs fine-tuned for textual entailment that yields state-of-the-art results on Information Extraction (IE) using only a small fraction of the annotations.'}",
-prompt-based self-training framework for few-shot named entity recognition,Ganghong Huang,,,
-instance-guided prompt learning for few-shot text matching,Jiacheng Du,,,https://aclanthology.org/2022.findings-emnlp.285.pdf
-towards unified prompt tuning for few-shot learning,,"Prompt-based fine-tuning has boosted the per- 001 formance of Pre-trained Language Models 002 (PLMs) on few-shot learning by employing 003 task-specific prompts. However, PLMs are 004 unfamiliar with the prompt-style expressions 005 during pre-training, which limits the few-shot 006 learning performance on downstream tasks. 007 It would be desirable if models can acquire 008 some prompting knowledge before task adap- 009 tation. We present the Unified Prompt Tun- 010 ing ( UPT ) framework, leading to better few- 011 shot learning for BERT-style models by ex- 012 plicitly capturing prompting semantics from 013 non-target NLP datasets. In UPT , a novel 014 paradigm Prompt-Options-Verbalizer is pro- 015 posed for joint prompt learning across differ- 016 ent NLP tasks, forcing PLMs to capture task- 017 invariant prompting knowledge. We further de- 018 sign a self-supervised task named Knowledge- 019 enhanced Selective Masked Language Model- 020 ing to improve the PLM’s generalization abil- 021 ities for accurate adaptation to previously un- 022 seen tasks. After multi-task learning, the PLM 023 can be fine-tuned for any target few-shot NLP 024 tasks using the same prompting paradigm. Ex- 025 periments over a variety of NLP tasks show 026 that UPT consistently outperforms state-of- 027 the-arts for prompt-based fine-tuning. 1 028","{'model': 'tldr@v2.0.0', 'text': 'This work presents the Unified Prompt Tun- 010 ing ( UPT ) framework, leading to better few- 011 shot learning for BERT-style models by ex- 012 plicitly capturing prompting semantics from non-target NLP datasets.'}",
-cqare: contrastive question-answering for few-shot relation extraction with prompt tuning,,"Prompt tuning with pre-trained language mod- 001 els (PLM) has exhibited outstanding perfor- 002 mance by closing the gap between pre-training 003 tasks and various downstream applications, 004 without the need for uninitialized parameters 005 to be introduced. However, prompt tuning 006 requires vast amounts of prompt engineering 007 and predefined label word mapping, which ob- 008 structs its implements in practice. Besides, the 009 ample label space makes prompt tuning more 010 arduous and challenging when it comes to re- 011 lation extraction (RE). To tackle these issues, 012 we propose a Contrastive Question-Answering 013 method with prompt tuning for few-shot RE 014 (CQARE). CQARE carries out a RE task- 015 specific pre-training with four entity-relation- 016 aware pre-training objects, including a prompt 017 pre-training to automatically generate contin- 018 uous prompts. The proposed pre-training can 019 provide more robust initialization with prompt 020 tuning while maintaining semantic consistency 021 with the proposed PLM. Furthermore, CQARE 022 can effectively avoid label words mapping by 023 reformulating RE as contrastive question an- 024 swering. The results indicate CQARE raising 025 averaged accuracy of 5.11% on a cross-domain 026 few-shot dataset, demonstrating that robust ini- 027 tialization is crucial for prompt tuning and ef- 028 fective contrastive question answering. 029","{'model': 'tldr@v2.0.0', 'text': 'A Contrastive Question-Answering method with prompt tuning for few-shot RE 014 (CQARE), which can effectively avoid label words mapping and reformulating RE as contrastive question answering while maintaining semantic consistency with the proposed PLM.'}",
-prompt-guided few-shot event detection,Prafulla Dhariwal,"Practical applications of event extraction sys- 001 tems have long been hindered by their need 002 for heavy human annotation. In order to scale 003 up to new domains and event types, models 004 must learn to cope with limited supervision, 005 as in few-shot learning settings. To this end, 006 the major challenge is to let the model master 007 the semantic of event types, without requiring 008 abundant event mention annotations. In our 009 study, we employ cloze prompts to elicit event- 010 related knowledge from pretrained language 011 models and further use event definitions and 012 keywords to pinpoint the trigger word. By for- 013 mulating the event detection task as an identify- 014 then-localize procedure, we minimize the num- 015 ber of type-specific parameters, enabling our 016 model to quickly adapt to event detection tasks 017 for new types. Experiments on three event de- 018 tection benchmark datasets (ACE, FewEvent, 019 MAVEN) show that our proposed method per- 020 forms favorably under fully supervised settings 021 and surpasses existing few-shot methods by 022 16% F1 on the FewEvent dataset and 23% on 023 the MAVEN dataset when only 5 examples are 024 provided for each event type. 1 025","{'model': 'tldr@v2.0.0', 'text': 'This study employs cloze prompts to elicit event- 010 related knowledge from pretrained language 011 models and further use event definitions and 012 keywords to pinpoint the trigger word, enabling the model to quickly adapt to event detection tasks for new types.'}",
-psp: pre-trained soft prompts for few-shot abstractive summarization,Xiaochen Liu,"Few-shot abstractive summarization has become a challenging task in natural language generation. To support it, we developed a novel soft prompts architecture coupled with a prompt pre-training plus prompt fine-tuning paradigm, which is effective and tunes only extremely light parameters. To meet the structure of the generation models, the soft prompts comprise continuous input embeddings across an encoder and a decoder. Importantly, a new inner-prompt placed in the text is introduced to capture document-level information. The aim is to devote attention to understanding the document that better prompts the model to generate document-related content. In the training process, the prompt pre-training with self-supervised pseudo-data firstly teaches the model basic summarizing capability. Then, with few-shot examples, only the designed lightweight soft prompts are fine-tuned. Experimental results on the CNN/DailyMail and XSum datasets show that our method, with only 0.1% of the parameters, outperforms full-model tuning where all model parameters are tuned. It also surpasses Prompt Tuning by a large margin and delivers competitive results against Prefix-Tuning with 3% of the parameters.","{'model': 'tldr@v2.0.0', 'text': 'A novel soft prompts architecture coupled with a prompt pre-training plus prompt fine-tuning paradigm, which is effective and tunes only extremely light parameters, and outperforms full-model tuning where all model parameters are tuned.'}",http://arxiv.org/pdf/2204.04413
-few-shot text-to-sql translation using structure and content prompt learning,Ju Fan,"A common problem with adopting Text-to-SQL translation in database systems is poor generalization. Specifically, when there is limited training data on new datasets, existing few-shot Text-to-SQL techniques, even with carefully designed textual prompts on pre-trained language models (PLMs), tend to be ineffective. In this paper, we present a divide-and-conquer framework to better support few-shot Text-to-SQL translation, which divides Text-to-SQL translation into two stages (or sub-tasks), such that each sub-task is simpler to be tackled. The first stage, called the structure stage , steers a PLM to generate an SQL structure (including SQL commands such as SELECT, FROM, WHERE and SQL operators such as “ < ”, “ > ”) with placeholders for missing identifiers. The second stage, called the content stage , guides a PLM to populate the placeholders in the generated SQL structure with concrete values (including SQL identifies such as table names, column names, and constant values). We propose a hybrid prompt strategy that combines learnable vectors and fixed vectors ( i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages. In addition, we design keyword constrained decoding to ensure the validity of generated SQL structures, and structure guided decoding to guarantee the model to fill correct content. Extensive experiments, by comparing with ten state-of-the-art Text-to-SQL solutions at the time of writing, show that SC-Prompt significantly outperforms them in the few-shot scenario. In particular, on the widely-adopted Spider dataset, given less than 500 labeled training examples (5% of the official training set), SC-Prompt outperforms","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hybrid prompt strategy that combines learnable vectors and fixed vectors, such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages of Text-to-SQL translation.'}",
-zero- and few-shot event detection via prompt-based meta learning,Zhenrui Yue,"With emerging online topics as a source for numerous new events, detecting unseen / rare event types presents an elusive challenge for existing event detection methods, where only limited data access is provided for training. To address the data scarcity problem in event detection, we propose MetaEvent, a meta learning-based framework for zero- and few-shot event detection. Specifically, we sample training tasks from existing event types and perform meta training to search for optimal parameters that quickly adapt to unseen tasks. In our framework, we propose to use the cloze-based prompt and a trigger-aware soft verbalizer to efficiently project output to unseen event types. Moreover, we design a contrastive meta objective based on maximum mean discrepancy (MMD) to learn class-separating features. As such, the proposed MetaEvent can perform zero-shot event detection by mapping features to event types without any prior knowledge. In our experiments, we demonstrate the effectiveness of MetaEvent in both zero-shot and few-shot scenarios, where the proposed method achieves state-of-the-art performance in extensive experiments on benchmark datasets FewEvent and MAVEN.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MetaEvent can perform zero-shot event detection by mapping features to event types without any prior knowledge, and achieves state-of-the-art performance in extensive experiments on benchmark datasets FewEvent and MAVEN.'}",http://arxiv.org/pdf/2305.17373
-vppt: visual pre-trained prompt tuning framework for few-shot image classification,Zhao Song,"Large-scale pre-trained transformers have recently achieved remarkable success in several computer vision tasks. However, it remains highly challenging to fully fine-tune models for downstream tasks, due to the expensive computational and storage cost. Recently, Parameter-Efficient Tuning (PETuning) techniques, e.g., Visual Prompt Tuning (VPT), have significantly reduced the computation cost by inserting lightweight prompt modules including prompt tokens or adapter layers, into the pre-trained models and tuning these prompt modules with a small number of trainable parameters, while keeping the transformer backbone freeze. Although encouraging results were achieved, existing PETuning methods cannot perform well under the few-shot learning settings (i.e., extremely limited training data, with only 1 or 2 shots per class), due to the scarce supervision signal. To this end, we first empirically identify the poor performance is mainly due to the inappropriate way of initializing prompt modules, which has also been verified in the pre-trained language models. Next, we propose a Visual Pre-trained Prompt Tuning framework (VPPT), which pre-trains the prompt modules first and then leverages the pre-trained modules along with the pre-trained transformer backbone to perform prompt tuning on downstream tasks. Extensive experiments show that our VPPT framework achieves 16.08% average accuracy absolute improvement under 1 shot setting on five fine-grained visual classification datasets, compared with the previous PETuning techniques, e.g., VPT, in few-shot image classification.","{'model': 'tldr@v2.0.0', 'text': 'A Visual Pre-trained Prompt Tuning framework (VPPT), which pre-trains the prompt modules first and then leverages the pre-trained modules along with thePre-trained transformer backbone to perform prompt tuning on downstream tasks, and achieves significant improvement under few-shot image classification settings.'}",
-decomposed two-stage prompt learning for few-shot named entity recognition,Feiyang Ye,"Named entity recognition (NER) in a few-shot setting is an extremely challenging task, and most existing methods fail to account for the gap between NER tasks and pre-trained language models. Although prompt learning has been successfully applied in few-shot classification tasks, adapting to token-level classification similar to the NER task presents challenges in terms of time consumption and efficiency. In this work, we propose a decomposed prompt learning NER framework for few-shot settings, decomposing the NER task into two stages: entity locating and entity typing. In training, the location information of distant labels is used to train the entity locating model. A concise but effective prompt template is built to train the entity typing model. In inference, a pipeline approach is used to handle the entire NER task, which elegantly resolves time-consuming and inefficient problems. Specifically, a well-trained entity locating model is used to predict entity spans for each input. The input is then transformed using prompt templates, and the well-trained entity typing model is used to predict their types in a single step. Experimental results demonstrate that our framework outperforms previous prompt-based methods by an average of 2.3–12.9% in F1 score while achieving the best trade-off between accuracy and inference speed.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a decomposed prompt learning NER framework for few-shot settings, decomposing the NER task into two stages: entity locating and entity typing.'}",https://www.mdpi.com/2078-2489/14/5/262/pdf?version=1682672547
-few-shot table-to-text generation with prompt planning and knowledge memorization,Zhixin Guo,"Pre-trained language models (PLM) have achieved remarkable advancement in table-to-text generation tasks. However, the lack of labeled domain-specific knowledge and the topology gap between tabular data and text make it difficult for PLMs to yield faithful text. Low-resource generation likewise faces unique challenges in this domain. Inspired by how humans descript tabular data with prior knowledge, we suggest a new framework: PromptMize, which targets table-to-text generation under few-shot settings. The design of our framework consists of two aspects: a prompt planner and a knowledge adapter. The prompt planner aims to generate a prompt signal that provides instance guidance for PLMs to bridge the topology gap between tabular data and text. Moreover, the knowledge adapter memorizes domain-specific knowledge from the unlabelled corpus to supply essential information during generation. Extensive experiments and analyses are investigated on three open domain few-shot NLG datasets: human, song, and book. Compared with previous state-of-the-art approaches, our model achieves remarkable performance in generating quality as judged by human and automatic evaluations.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by how humans descript tabular data with prior knowledge, a new framework is suggested: PromptMize, which targets table-to-text generation under few-shot settings and achieves remarkable performance in generating quality as judged by human and automatic evaluations.'}",https://arxiv.org/pdf/2302.04415
-locoop: few-shot out-of-distribution detection via prompt learning,Atsuyuki Miyai,"We present a novel vision-language prompt learning approach for few-shot out-of-distribution (OOD) detection. Few-shot OOD detection aims to detect OOD images from classes that are unseen during training using only a few labeled in-distribution (ID) images. While prompt learning methods such as CoOp have shown effectiveness and efficiency in few-shot ID classification, they still face limitations in OOD detection due to the potential presence of ID-irrelevant information in text embeddings. To address this issue, we introduce a new approach called Local regularized Context Optimization (LoCoOp), which performs OOD regularization that utilizes the portions of CLIP local features as OOD features during training. CLIP's local features have a lot of ID-irrelevant nuisances (e.g., backgrounds), and by learning to push them away from the ID class text embeddings, we can remove the nuisances in the ID class text embeddings and enhance the separation between ID and OOD. Experiments on the large-scale ImageNet OOD detection benchmarks demonstrate the superiority of our LoCoOp over zero-shot, fully supervised detection methods and prompt learning methods. Notably, even in a one-shot setting -- just one label per class, LoCoOp outperforms existing zero-shot and fully supervised detection methods. The code will be available via https://github.com/AtsuMiyai/LoCoOp.","{'model': 'tldr@v2.0.0', 'text': 'A novel vision-language prompt learning approach for few-shot out-of-distribution (OOD) detection called LoCoOp, which performs OOD regularization that utilizes the portions of CLIP local features as OOD features during training.'}",http://arxiv.org/pdf/2306.01293
-few-shot joint multimodal aspect-sentiment analysis based on generative multimodal prompt,Xiaocui Yang,"We have witnessed the rapid proliferation of multimodal data on numerous social media platforms. Conventional studies typically require massive labeled data to train models for Multimodal Aspect-Based Sentiment Analysis (MABSA). However, collecting and annotating fine-grained multimodal data for MABSA is tough. To alleviate the above issue, we perform three MABSA-related tasks with quite a small number of labeled multimodal samples. We first build diverse and comprehensive multimodal few-shot datasets according to the data distribution. To capture the specific prompt for each aspect term in a few-shot scenario, we propose a novel Generative Multimodal Prompt (GMP) model for MABSA, which includes the Multimodal Encoder module and the N-Stream Decoders module. We further introduce a subtask to predict the number of aspect terms in each instance to construct the multimodal prompt. Extensive experiments on two datasets demonstrate that our approach outperforms strong baselines on two MABSA-related tasks in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Generative Multimodal Prompt (GMP) model for MABSA, which includes the Multinodal Encoder module and the N-Stream Decoders module, and introduces a subtask to predict the number of aspect terms in each instance to construct the multimodal prompt.'}",http://arxiv.org/pdf/2305.10169
-partseg: few-shot part segmentation via part-aware prompt learning,M. Han,"In this work, we address the task of few-shot part segmentation, which aims to segment the different parts of an unseen object using very few labeled examples. It is found that leveraging the textual space of a powerful pre-trained image-language model (such as CLIP) can be beneficial in learning visual features. Therefore, we develop a novel method termed PartSeg for few-shot part segmentation based on multimodal learning. Specifically, we design a part-aware prompt learning method to generate part-specific prompts that enable the CLIP model to better understand the concept of ``part'' and fully utilize its textual space. Furthermore, since the concept of the same part under different object categories is general, we establish relationships between these parts during the prompt learning process. We conduct extensive experiments on the PartImageNet and Pascal$\_$Part datasets, and the experimental results demonstrated that our proposed method achieves state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': ""This work designs a part-aware prompt learning method to generate part-specific prompts that enable the CLIP model to better understand the concept of ``part'' and fully utilize its textual space.""}",https://arxiv.org/pdf/2308.12757
-evolutionary verbalizer search for prompt-based few shot text classification,Tongtao Ling,"Recent advances for few-shot text classification aim to wrap textual inputs with task-specific prompts to cloze questions. By processing them with a masked language model to predict the masked tokens and using a verbalizer that constructs the mapping between predicted words and target labels. This approach of using pre-trained language models is called prompt-based tuning, which could remarkably outperform conventional fine-tuning approach in the low-data scenario. As the core of prompt-based tuning, the verbalizer is usually handcrafted with human efforts or suboptimally searched by gradient descent. In this paper, we focus on automatically constructing the optimal verbalizer and propose a novel evolutionary verbalizer search (EVS) algorithm, to improve prompt-based tuning with the high-performance verbalizer. Specifically, inspired by evolutionary algorithm (EA), we utilize it to automatically evolve various verbalizers during the evolutionary procedure and select the best one after several iterations. Extensive few-shot experiments on five text classification datasets show the effectiveness of our method.","{'model': 'tldr@v2.0.0', 'text': 'This paper focuses on automatically constructing the optimal verbalizer and proposes a novel evolutionary verbalizer search (EVS) algorithm, inspired by evolutionary algorithm, to improve prompt-based tuning with the high-performance verbalizer.'}",http://arxiv.org/pdf/2306.10514
-a chinese few-shot text classification method utilizing improved prompt learning and unlabeled data,Tingkai Hu,"Insufficiently labeled samples and low-generalization performance have become significant natural language processing problems, drawing significant concern for few-shot text classification (FSTC). Advances in prompt learning have significantly improved the performance of FSTC. However, prompt learning methods typically require the pre-trained language model and tokens of the vocabulary list for model training, while different language models have different token coding structures, making it impractical to build effective Chinese prompt learning methods from previous approaches related to English. In addition, a majority of current prompt learning methods do not make use of existing unlabeled data, thus often leading to unsatisfactory performance in real-world applications. To address the above limitations, we propose a novel Chinese FSTC method called CIPLUD that combines an improved prompt learning method and existing unlabeled data, which are used for the classification of a small amount of Chinese text data. We used the Chinese pre-trained language model to build two modules: the Multiple Masks Optimization-based Prompt Learning (MMOPL) module and the One-Class Support Vector Machine-based Unlabeled Data Leveraging (OCSVM-UDL) module. The former generates prompt prefixes with multiple masks and constructs suitable prompt templates for Chinese labels. It optimizes the random token combination problem during label prediction with joint probability and length constraints. The latter, by establishing an OCSVM model in the trained text vector space, selects reasonable pseudo-label data for each category from a large amount of unlabeled data. After selecting the pseudo-label data, we mixed them with the previous few-shot annotated data to obtain brand new training data and then repeated the steps of the two modules as an iterative semi-supervised optimization process. The experimental results on the four Chinese FSTC benchmark datasets demonstrate that our proposed solution outperformed other prompt learning methods with an average accuracy improvement of 2.3%.","{'model': 'tldr@v2.0.0', 'text': 'A novel Chinese FSTC method called CIPLUD is proposed that combines an improved prompt learning method and existing unlabeled data, which are used for the classification of a small amount of Chinese text data, and outperformed other prompt learning methods with an average accuracy improvement of 2.3%.'}",https://www.mdpi.com/2076-3417/13/5/3334/pdf?version=1678093925
-unified prompt learning makes pre-trained language models better few-shot learners,Feihu Jin,"Language prompting induces the model to produce a textual output during the training phase, which achieves remarkable performance in few-shot learning scenarios. However, current prompt-based methods either use the same task-specific prompts for each instance, losing the particularity of instance-dependent information, or generate an instance-dependent prompt for each instance, lacking shared information about the task. In this paper, we propose an efficient few-shot learning method to dynamically decide the degree to which task-specific and instance-dependent information are incorporated according to different task and instance characteristics, enriching the prompt with task-specific and instance-dependent information. Extensive experiments on a wide range of natural language understanding tasks demonstrate that our approach obtains significant improvements compared to prompt-based fine-tuning baselines in a few-shot setting with about 0.1% parameters tuned. Moreover, our approach outperforms existing state-of-the-art efficient few-shot learning methods on several natural language understanding tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an efficient few-shot learning method to dynamically decide the degree to which task-specific and instance-dependent information are incorporated according to different task and instance characteristics, enriching the prompt with task- specific and instanceTheoretical studies demonstrate that this approach obtains significant improvements compared to prompt-based fine-tuning baselines in a few- shot setting.'}",
-boosting prompt-based few-shot learners through out-of-domain knowledge distillation,Xiaoqing Chen,"Prompt-based learning improves the performance of Pre-trained Language Models (PLMs) over few-shot learning and is suitable for low-resourced scenarios. However, it is challenging to deploy large PLMs online. Knowledge Distillation (KD) can compress large PLMs into small ones; yet, few-shot KD for prompt-tuned PLMs is challenging due to the lack of training data and the capacity gap between teacher and student models. We propose Boost-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs with the help of the out-of-domain data. Apart from distilling the model logits, Boost-Distiller specifically considers heuristically-generated fake logits that improve the generalization abilities of student models. We further leverage the cross-domain model logits, weighted with domain expertise scores that measure the transferablity of out-of-domain instances. Experiments over various datasets show Boost-Distiller consistently outperforms baselines by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Boost-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs with the help of the out-of-domain data, and specifically considers heuristically-generated fake logits that improve the generalization abilities of student models.'}",
-prompt-based zero- and few-shot node classification: a multimodal approach,Yue Li,"Multimodal data empowers machine learning models to better understand the world from various perspectives. In this work, we study the combination of \emph{text and graph} modalities, a challenging but understudied combination which is prevalent across multiple settings including citation networks, social media, and the web. We focus on the popular task of node classification using limited labels; in particular, under the zero- and few-shot scenarios. In contrast to the standard pipeline which feeds standard precomputed (e.g., bag-of-words) text features into a graph neural network, we propose \textbf{T}ext-\textbf{A}nd-\textbf{G}raph (TAG) learning, a more deeply multimodal approach that integrates the raw texts and graph topology into the model design, and can effectively learn from limited supervised signals without any meta-learning procedure. TAG is a two-stage model with (1) a prompt- and graph-based module which generates prior logits that can be directly used for zero-shot node classification, and (2) a trainable module that further calibrates these prior logits in a few-shot manner. Experiments on two node classification datasets show that TAG outperforms all the baselines by a large margin in both zero- and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'TAG is a two-stage model with a prompt- and graph-based module which generates prior logits that can be directly used for zero-shot node classification, and a trainable module that further calibrates these priorlogits in a few-shot manner.'}",https://arxiv.org/pdf/2307.11572
-prompt-distiller: few-shot knowledge distillation for prompt-based language learners with dual contrastive learning,Boyu Hou,"Prompt-based learning has improved the few-shot learning performance of large-scale Pre-trained Language Models (PLMs). Yet, it is challenging to deploy large-scale PLMs in resource-constrained environments for online applications. Knowledge Distillation (KD) is a promising approach for PLM compression. However, distilling prompt-tuned PLMs in the few-shot learning setting is a non-trivial problem due to the lack of task-specific training data and KD techniques for the new prompting paradigm. We propose Prompt-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs, which forces the student model to learn from both its pre-trained and prompt-tuned teacher models to alleviate the model overfitting problem. We further design a contrastive learning technique to learn higher-order dependencies from intermediate-layer representations of teacher models, considering different knowledge capacities of teacher and student models. Extensive experiments over various datasets show that Prompt-Distiller consistently outperforms baselines by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Prompt-Distiller, the first few-shot KD algorithm for prompt-tuned PLMs, which forces the student model to learn from both its pre-trained and prompt- Tuned teacher models to alleviate the model overfitting problem.'}",
-few-shot text-to-sql translation using structure and content prompt learning,Zihui Gu,"A common problem with adopting Text-to-SQL translation in database systems is poor generalization. Specifically, when there is limited training data on new datasets, existing few-shot Text-to-SQL techniques, even with carefully designed textual prompts on pre-trained language models (PLMs), tend to be ineffective. In this paper, we present a divide-and-conquer framework to better support few-shot Text-to-SQL translation, which divides Text-to-SQL translation into two stages (or sub-tasks), such that each sub-task is simpler to be tackled. The first stage, called the structure stage, steers a PLM to generate an SQL structure (including SQL commands such as SELECT, FROM, WHERE and SQL operators such as <"", ?>"") with placeholders for missing identifiers. The second stage, called the content stage, guides a PLM to populate the placeholders in the generated SQL structure with concrete values (including SQL identifies such as table names, column names, and constant values). We propose a hybrid prompt strategy that combines learnable vectors and fixed vectors (i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages. In addition, we design keyword constrained decoding to ensure the validity of generated SQL structures, and structure guided decoding to guarantee the model to fill correct content. Extensive experiments, by comparing with ten state-of-the-art Text-to-SQL solutions at the time of writing, show that SC-Prompt significantly outperforms them in the few-shot scenario. In particular, on the widely-adopted Spider dataset, given less than 500 labeled training examples (5% of the official training set), SC-Prompt outperforms the previous SOTA methods by around 5% on accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hybrid prompt strategy that combines learnable vectors and fixed vectors (i.e., word embeddings of textual prompts), such that the hybrid prompt can learn contextual information to better guide PLMs for prediction in both stages.'}",
-dreamartist: towards controllable one-shot text-to-image generation via positive-negative prompt-tuning,Ziyi Dong,"Large-scale text-to-image generation models have achieved remarkable progress in synthesizing high-quality, feature-rich images with high resolution guided by texts. However, these models often struggle with novel concepts, eg, new styles, object entities, etc. Although recent attempts have employed fine-tuning or prompt-tuning strategies to teach the pre-trained diffusion model novel concepts from a reference image set,they have the drawback of overfitting to the given reference images, particularly in one-shot applications, which is harmful to generate diverse and high-quality images while maintaining generation controllability. To tackle this challenge, we present a simple yet effective method called DreamArtist, which employs a positive-negative prompt-tuning learning strategy. Specifically, DreamArtist incorporates both positive and negative embeddings and jointly trains them. The positive embedding aggressively captures the salient characteristics of the reference image to drive diversified generation and the negative embedding rectifies inadequacies from the positive embedding. It learns not only what is correct, but also what can be avoided or improved. We have conducted extensive experiments and evaluated the proposed method from image similarity and diversity, generation controllability, and style cloning. And our DreamArtist has achieved a superior generation performance over existing methods. Besides, our additional evaluation on extended tasks, including concept compositions and prompt-guided image editing, demonstrates its effectiveness for more applications.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a simple yet effective method, which employs a positive-negative prompt-tuning learning strategy, which incorporates both positive and negative embeddings and jointly trains them and has achieved a superior generation performance over existing methods.'}",
-dreamartist: towards controllable one-shot text-to-image generation via contrastive prompt-tuning,Ziyi Dong,"Large-scale text-to-image generation models with an exponential evolution can currently synthesize high-resolution, feature-rich, high-quality images based on text guidance. However, they are often overwhelmed by words of new concepts, styles, or object entities that always emerge. Although there are some recent attempts to use fine-tuning or prompt-tuning methods to teach the model a new","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2211.11337
-one-shot and partially-supervised cell image segmentation using small visual prompt,S. Kato,"Semantic segmentation of microscopic cell images using deep learning is an important technique, however, it requires a large number of images and ground truth labels for training. To address the above problem, we consider an efficient learning framework with as little data as possible, and we propose two types of learning strategies: One-shot segmentation which can learn with only one training sample, and Partially-supervised segmentation which assigns annotations to only a part of images. Furthermore, we introduce novel segmentation methods using the small prompt images inspired by prompt learning in recent studies. Our proposed methods use a pre-trained model based on only cell images and teach the information of the prompt pairs to the target image to be segmented by the attention mechanism, which allows for efficient learning while reducing the burden of annotation costs. Through experiments conducted on three types of microscopic cell image datasets, we confirmed that the proposed method improved the Dice score coefficient (DSC) in comparison with the conventional methods. Our code is available at https://github.com/usagisukisuki/Oneshot-Part-CellSegmentation.","{'model': 'tldr@v2.0.0', 'text': 'The proposed methods use a pre-trained model based on only cell images and teach the information of the prompt pairs to the target image to be segmented by the attention mechanism, which allows for efficient learning while reducing the burden of annotation costs.'}",https://arxiv.org/pdf/2304.07991
-learning from graph propagation via ordinal distillation for one-shot automated essay scoring,Zhiwei Jiang,"One-shot automated essay scoring (AES) aims to assign scores to a set of essays written specific to a certain prompt, with only one manually scored essay per distinct score. Compared to the previous-studied prompt-specific AES which usually requires a large number of manually scored essays for model training (e.g., about 600 manually scored essays out of totally 1000 essays), one-shot AES can greatly reduce the workload of manual scoring. In this paper, we propose a Transductive Graph-based Ordinal Distillation (TGOD) framework to tackle the task of one-shot AES. Specifically, we design a transductive graph-based model as a teacher model to generate pseudo labels of unlabeled essays based on the one-shot labeled essays. Then, we distill the knowledge in the teacher model into a neural student model by learning from the high confidence pseudo labels. Different from the general knowledge distillation, we propose an ordinal-aware unimodal distillation which makes a unimodal distribution constraint on the output of student model, to tolerate the minor errors existed in pseudo labels. Experimental results on the public dataset ASAP show that TGOD can improve the performance of existing neural AES models under the one-shot AES setting and achieve an acceptable average QWK of 0.69.","{'model': 'tldr@v2.0.0', 'text': 'A transductive graph-based model as a teacher model to generate pseudo labels of unlabeled essays based on the one-shot labeled essays is designed, which distill the knowledge in the teacher model into a neural student model by learning from the high confidence pseudo labels.'}",
-mental health recognition using one-shot learning and faci al expressions,Mahesh Kumar,"The visual representation of a person's affective state, cognitive function, purpose, personality, and psychology can be obtained from facial expressions.[1]. The idea of our project is to create a platform with a camera, to determine whether an individual is suffering from deterioration of mental health, particularly in an institutionalized environment. In general, disturbed mental condition is determined by using a combination of various factors such as facial expressions, body posture, speech, tone, gestures, and so on[2]. This project lays focus on Facial Expression Recognition (FER), achievedwith the help of Convolution Neural Network (CNN), One-Shot learning and TensorFlow. The dataset used here is the Kaggle dataset FER2013, which comprises of seven facial expressions labelled as happy, sad, surprise, fear, anger, disgust and neutral. The results of this classification are stored and compared to determine whether an individual displays negative emotion for a significant number of times. After detection and classification, a prompt is sent to the designated authorities at the institution through a web application. On the basis of that, if an individual is found to be suffering from a disturbed state of mind, he/she is recommended for therapy or medical attention.","{'model': 'tldr@v2.0.0', 'text': 'The idea of this project is to create a platform with a camera, to determine whether an individual is suffering from deterioration of mental health, particularly in an institutionalized environment, with the help of Convolution Neural Network, One-Shot learning and TensorFlow.'}",
-pøda: prompt-driven zero-shot domain adaptation,Mohammad Fahes,"Domain adaptation has been vastly investigated in computer vision but still requires access to target images at train time, which might be intractable in some uncommon conditions. In this paper, we propose the task of `Prompt-driven Zero-shot Domain Adaptation', where we adapt a model trained on a source domain using only a general description in natural language of the target domain, i.e., a prompt. First, we leverage a pretrained contrastive vision-language model (CLIP) to optimize affine transformations of source features, steering them towards the target text embedding while preserving their content and semantics. To achieve this, we propose Prompt-driven Instance Normalization (PIN). Second, we show that these prompt-driven augmentations can be used to perform zero-shot domain adaptation for semantic segmentation. Experiments demonstrate that our method significantly outperforms CLIP-based style transfer baselines on several datasets for the downstream task at hand, even surpassing one-shot unsupervised domain adaptation. A similar boost is observed on object detection and image classification. The code is available at https://github.com/astra-vision/PODA .","{'model': 'tldr@v2.0.0', 'text': 'This paper uses a pretrained contrastive vision-language model (CLIP) to optimize affine transformations of source features, steering them towards the target text embedding while preserving their content and semantics, and shows that these prompt-driven augmentations can be used to perform zero-shot domain adaptation for semantic segmentation.'}",https://arxiv.org/pdf/2212.03241
-cohoz: contrastive multimodal prompt tuning for hierarchical open-set zero-shot recognition,Ning Liao,"Practical image recognition often encounters samples whose labels either are totally unknown or belong to new classes outside the training set. The first problem refers to the open-set recognition (OSR), in which unknown classes are recognized as one with no more semantic information. While the latter is called zero-shot learning (ZSL), in which new classes are usually predefined. The existing literature mostly addresses these two problems separately. In this paper, we take the ambition for solving the combination of these two problems to fulfill semantically recognizing the unknown classes detected in OSR by zero-shot prediction. We propose the Contrastive multimodal prompt tuning for Hierarchical Open-set Zero-shot recognition (CoHOZ). Specifically, we firstly build a global and compatible hierarchical label tree with all downstream datasets aligned, which lays foundations for other modules. To detect unknown classes, we propose the contrastive continuous prompt tuning, which introduces additional negative classes from the fine level of the built hierarchy for prompt learning. To generate candidate classes for zero-shot prediction on the unknown data using prompt, we combine the built hierarchy to collect candidate classes from coarse to fine. In our experiments, when following the standard OSR protocol regarding all the unknown classes as a single class, CoHOZ achieves a new state-of-the-art performance both in unknown detection and open-set recognition. Few-shot tuning by the CoHOZ also shows competitive performance on them. In addition, the detailed semantic information of unknown classes are well explored, which has also been verified in experiments.","{'model': 'tldr@v2.0.0', 'text': 'The ambition for solving the combination of these two problems to fulfill semantically recognizing the unknown classes detected in OSR by zero-shot prediction is taken and the Contrastive multimodal prompt tuning for Hierarchical Open-set Zero-shot recognition (CoHOZ) is proposed.'}",
-proze: explainable and prompt-guided zero-shot text classification,Ismail Harrando,"As technology accelerates the generation and communication of textual data, the need to automatically understand this content becomes a necessity. In order to classify text, being it for tagging, indexing, or curating documents, one often relies on large, opaque models that are trained on preannotated datasets, making the process unexplainable, difficult to scale, and ill-adapted for niche domains with scarce data. To tackle these challenges, we propose ProZe, a text classification approach that leverages knowledge from two sources: prompting pretrained language models, as well as querying ConceptNet, a common-sense knowledge base which can be used to add a layer of explainability to the results. We evaluate our approach empirically and we show how this combination not only performs on par with state-of-the-art zero shot classification on several domains, but also offers explainable predictions that can be visualized.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ProZe, a text classification approach that leverages knowledge from two sources: prompting pretrained language models, as well as querying ConceptNet, a common-sense knowledge base which can be used to add a layer of explainability to the results.'}",
-controlvideo: adding conditional control for one shot text-to-video editing,Min Zhao,"In this paper, we present ControlVideo, a novel method for text-driven video editing. Leveraging the capabilities of text-to-image diffusion models and ControlNet, ControlVideo aims to enhance the fidelity and temporal consistency of videos that align with a given text while preserving the structure of the source video. This is achieved by incorporating additional conditions such as edge maps, fine-tuning the key-frame and temporal attention on the source video-text pair with carefully designed strategies. An in-depth exploration of ControlVideo's design is conducted to inform future research on one-shot tuning video diffusion models. Quantitatively, ControlVideo outperforms a range of competitive baselines in terms of faithfulness and consistency while still aligning with the textual prompt. Additionally, it delivers videos with high visual realism and fidelity w.r.t. the source content, demonstrating flexibility in utilizing controls containing varying degrees of source video information, and the potential for multiple control combinations. The project page is available at \href{https://ml.cs.tsinghua.edu.cn/controlvideo/}{https://ml.cs.tsinghua.edu.cn/controlvideo/}.","{'model': 'tldr@v2.0.0', 'text': 'Quantitatively, ControlVideo outperforms a range of competitive baselines in terms of faithfulness and consistency while still aligning with the textual prompt, and it delivers videos with high visual realism and fidelity w.r.t. the source content.'}",http://arxiv.org/pdf/2305.17098
-matcher: segment anything with one shot using all-purpose feature matching,Yang Liu,"Powered by large-scale pre-training, vision foundation models exhibit significant potential in open-world image understanding. Even though individual models have limited capabilities, combining multiple such models properly can lead to positive synergies and unleash their full potential. In this work, we present Matcher, which segments anything with one shot by integrating an all-purpose feature extraction model and a class-agnostic segmentation model. Naively connecting the models results in unsatisfying performance, e.g., the models tend to generate matching outliers and false-positive mask fragments. To address these issues, we design a bidirectional matching strategy for accurate cross-image semantic dense matching and a robust prompt sampler for mask proposal generation. In addition, we propose a novel instance-level matching strategy for controllable mask merging. The proposed Matcher method delivers impressive generalization performance across various segmentation tasks, all without training. For example, it achieves 52.7% mIoU on COCO-20$^i$ for one-shot semantic segmentation, surpassing the state-of-the-art specialist model by 1.6%. In addition, our visualization results show open-world generality and flexibility on images in the wild. The code shall be released at https://github.com/aim-uofa/Matcher.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Matcher, which segments anything with one shot by integrating an all-purpose feature extraction model and a class-agnostic segmentation model, and proposes a novel instance-level matching strategy for controllable mask merging.'}",http://arxiv.org/pdf/2305.13310
-the segment anything model (sam) for remote sensing applications: from zero to one shot,L. Osco,"Segmentation is an essential step for remote sensing image processing. This study aims to advance the application of the Segment Anything Model (SAM), an innovative image segmentation model by Meta AI, in the field of remote sensing image analysis. SAM is known for its exceptional generalization capabilities and zero-shot learning, making it a promising approach to processing aerial and orbital images from diverse geographical contexts. Our exploration involved testing SAM across multi-scale datasets using various input prompts, such as bounding boxes, individual points, and text descriptors. To enhance the model's performance, we implemented a novel automated technique that combines a text-prompt-derived general example with one-shot training. This adjustment resulted in an improvement in accuracy, underscoring SAM's potential for deployment in remote sensing imagery and reducing the need for manual annotation. Despite the limitations encountered with lower spatial resolution images, SAM exhibits promising adaptability to remote sensing data analysis. We recommend future research to enhance the model's proficiency through integration with supplementary fine-tuning techniques and other networks. Furthermore, we provide the open-source code of our modifications on online repositories, encouraging further and broader adaptations of SAM to the remote sensing domain.",,http://arxiv.org/pdf/2306.16623
-on the design of optimal one-shot distributed combining for cooperative multi-uav systems,Xuesong Pan,"Unmanned aerial vehicle (UAV) networks will become one of the key components of the future mobile communication systems, especially suitable for on-demand coverage extension and prompt capacity enhancement. However, UAV networks, detached from the ground infrastructure, are usually constrained for traffic-intensive network-wide signal processing. This paper tackles this issue by considering the multi-user signal detection at the UAV swarm, where the uplink signals of multiple ground users are cooperatively recovered. In particular, to reduce the signaling exchange and the coordination latency, a two-stage distributed minimum mean squared error (MMSE) detection is proposed for the UAV swarm, which merges the local MMSE detections at each UAV with one-shot weighted combining at a central station. The combining weights only depend on the long-term statistical channel state informations (CSIs) and therefore, are suitable for distributed network with limited signaling exchange. Numerical results show that the proposed distributed MMSE detection achieves similar performance to the centralized MMSE detection.","{'model': 'tldr@v2.0.0', 'text': 'A two-stage distributed minimum mean squared error (MMSE) detection is proposed for the UAV swarm, which merges the local MMSE detections at each UAV with one-shot weighted combining at a central station, suitable for distributed network with limited signaling exchange.'}",
-reinforcement learning-based grasping via one-shot affordance localization and zero-shot contrastive language–image learning,Xiang Long,"—We present a novel robotic grasping system using a caging-style gripper, that combines one-shot affordance localization and zero-shot object identification. We demonstrate an integrated system requiring minimal prior knowledge, focusing on flexible few-shot object agnostic approaches. For grasping a novel target object, we use as input the color and depth of the scene, an image of an object affordance similar to the target object, and an up to three-word text prompt describing the target object. We demonstrate the system using real-world grasping of objects from the YCB benchmark set, with four distractor objects cluttering the scene. Overall, our pipeline has a success rate of the affordance localization of 96% , object identification of 62 . 5% , and grasping of 72% . Videos are on the project website: https://sites.google.com/view/ rl-affcorrs-grasp .","{'model': 'tldr@v2.0.0', 'text': None}",
-one-shot localization and segmentation of medical images with foundation models,Deepa Anand,"Recent advances in Vision Transformers (ViT) and Stable Diffusion (SD) models with their ability to capture rich semantic features of the image have been used for image correspondence tasks on natural images. In this paper, we examine the ability of a variety of pre-trained ViT (DINO, DINOv2, SAM, CLIP) and SD models, trained exclusively on natural images, for solving the correspondence problems on medical images. While many works have made a case for in-domain training, we show that the models trained on natural images can offer good performance on medical images across different modalities (CT,MR,Ultrasound) sourced from various manufacturers, over multiple anatomical regions (brain, thorax, abdomen, extremities), and on wide variety of tasks. Further, we leverage the correspondence with respect to a template image to prompt a Segment Anything (SAM) model to arrive at single shot segmentation, achieving dice range of 62%-90% across tasks, using just one image as reference. We also show that our single-shot method outperforms the recently proposed few-shot segmentation method - UniverSeg (Dice range 47%-80%) on most of the semantic segmentation tasks(six out of seven) across medical imaging modalities.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the models trained on natural images can offer good performance on medical images across different modalities (CT,MR,Ultrasound) sourced from various manufacturers, over multiple anatomical regions (brain, thorax, abdomen, extremities), and on wide variety of tasks.'}",
-prompt-based extraction of social determinants of health using few-shot learning,Giridhar Kaushik Ramachandran,"Social determinants of health (SDOH) documented in the electronic health record through unstructured text are increasingly being studied to understand how SDOH impacts patient health outcomes. In this work, we utilize the Social History Annotation Corpus (SHAC), a multi-institutional corpus of de-identified social history sections annotated for SDOH, including substance use, employment, and living status information. We explore the automatic extraction of SDOH information with SHAC in both standoff and inline annotation formats using GPT-4 in a one-shot prompting setting. We compare GPT-4 extraction performance with a high-performing supervised approach and perform thorough error analyses. Our prompt-based GPT-4 method achieved an overall 0.652 F1 on the SHAC test set, similar to the 7th best-performing system among all teams in the n2c2 challenge with SHAC.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes the Social History Annotation Corpus (SHAC), a multi-institutional corpus of de-identified social history sections annotated for SDOH, including substance use, employment, and living status information, to explore the automatic extraction of SDOH information with SHAC in both standoff and inline annotation formats using GPT-4 in a one-shot prompting setting.'}",http://arxiv.org/pdf/2306.07170
-augmenters at semeval-2023 task 1: enhancing clip in handling compositionality and ambiguity for zero-shot visual wsd through prompt augmentation and text-to-image diffusion,Jie Li,"This paper describes our zero-shot approachesfor the Visual Word Sense Disambiguation(VWSD) Task in English. Our preliminarystudy shows that the simple approach of match-ing candidate images with the phrase usingCLIP suffers from the many-to-many natureof image-text pairs. We find that the CLIP textencoder may have limited abilities in captur-ing the compositionality in natural language.Conversely, the descriptive focus of the phrasevaries from instance to instance. We addressthese issues in our two systems, Augment-CLIPand Stable Diffusion Sampling (SD Sampling).Augment-CLIP augments the text prompt bygenerating sentences that contain the contextphrase with the help of large language mod-els (LLMs). We further explore CLIP modelsin other languages, as the an ambiguous wordmay be translated into an unambiguous one inthe other language. SD Sampling uses text-to-image Stable Diffusion to generate multipleimages from the given phrase, increasing thelikelihood that a subset of images match theone that paired with the text.","{'model': 'tldr@v2.0.0', 'text': 'The preliminary study shows that the simple approach of match-ing candidate images with the phrase using CLIP suffers from the many-to-many nature of image-text pairs, and finds that the CLIP textencoder may have limited abilities in captur-ing the compositionality in natural language.'}",https://arxiv.org/pdf/2307.05564
-self-supervised meta-prompt learning with meta-gradient regularization for few-shot generalization,Kaihang Pan,"Prompt tuning is a parameter-efficient method, which learns soft prompts and conditions frozen language models to perform specific downstream tasks. Though effective, prompt tuning under few-shot settings on the one hand heavily relies on a good initialization of soft prompts. On the other hand, it can easily overfit to few-shot training samples, thereby undermining generalizability. Existing works leverage pre-training or supervised meta-learning to initialize soft prompts but they fail to data-efficiently generalize to unseen downstream tasks. To address the above problems, this paper proposes a novel Self-sUpervised meta-Prompt learning framework with MEta-gradient Regularization for few-shot generalization (SUPMER). SUPMER leverages self-supervised meta-learning with a diverse set of well-designed meta-training tasks to learn a universal prompt initialization for efficient adaptation using only unlabeled data. Additionally, it jointly meta-learns a gradient regularization function to transform raw gradients into a domain-generalizable direction, thus alleviating the problem of overfitting. Extensive experiments show that SUPMER achieves better performance for different few-shot downstream tasks, and also exhibits a stronger domain generalization ability. The code for SUPMER will be available at https://github.com/beepkh/SUPMER.","{'model': 'tldr@v2.0.0', 'text': 'SUPMER leverages self-supervised meta-learning with a diverse set of well-designed meta-training tasks to learn a universal prompt initialization for efficient adaptation using only unlabeled data, and jointly meta-learns a gradient regularization function to transform raw gradients into a domain-generalizable direction, thus alleviating the problem of overfitting.'}",
-enhancing black-box few-shot text classification with prompt-based data augmentation,Dan Luo,"Training or finetuning large-scale language models (LLMs) such as GPT-3 requires substantial computation resources, motivating recent efforts to explore parameter-efficient adaptation to downstream tasks. One practical area of research is to treat these models as black boxes and interact with them through their inference APIs. In this paper, we investigate how to optimize few-shot text classification without accessing the gradients of the LLMs. To achieve this, we treat the black-box model as a feature extractor and train a classifier with the augmented text data. Data augmentation is performed using prompt-based finetuning on an auxiliary language model with a much smaller parameter size than the black-box model. Through extensive experiments on eight text classification datasets, we show that our approach, dubbed BT-Classifier, significantly outperforms state-of-the-art black-box few-shot learners and performs on par with methods that rely on full-model tuning.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates how to optimize few-shot text classification without accessing the gradients of the LLMs by treating the black-box model as a feature extractor and train a classifier with the augmented text data.'}",http://arxiv.org/pdf/2305.13785
-"cocoopter: pre-train, prompt, and fine-tune the vision-language model for few-shot image classification",Jie Yan,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed CoCoOpter alleviates the overfitting and ensures generalizability in unseen new categories in unseen classes and achieves superior few-shot classification performance with a straightforward design.'}",
-few-shot fake news detection via prompt-based tuning,Wang Gao," As people increasingly use social media to read news, fake news has become a major problem for the public and government. One of the main challenges in fake news detection is how to identify them in the early stage of propagation. Another challenge is that detection model training requires large amounts of labeled data, which are often unavailable or expensive to acquire. To address these challenges, we propose a novel Fake News Detection model based on Prompt Tuning (FNDPT). FNDPT first designs a prompt-based template for early fake news detection. This mechanism incorporates contextual information into textual content and extracts relevant knowledge from pre-trained language models. Furthermore, our model utilizes prompt-based tuning to enhance the performance in a few-shot setting. Experimental results on two real-world datasets verify the effectiveness of FNDPT.","{'model': 'tldr@v2.0.0', 'text': 'A novel Fake News Detection model based on Prompt Tuning (FNDPT), which incorporates contextual information into textual content and extracts relevant knowledge from pre-trained language models to enhance the performance in a few-shot setting.'}",
-lm-cppf: paraphrasing-guided data augmentation for contrastive prompt-based few-shot fine-tuning,Amirhossein Abaskohi,"In recent years, there has been significant progress in developing pre-trained language models for NLP. However, these models often struggle when fine-tuned on small datasets. To address this issue, researchers have proposed various adaptation approaches. Prompt-based tuning is arguably the most common way, especially for larger models. Previous research shows that adding contrastive learning to prompt-based fine-tuning is effective as it helps the model generate embeddings that are more distinguishable between classes, and it can also be more sample-efficient as the model learns from positive and negative examples simultaneously. One of the most important components of contrastive learning is data augmentation, but unlike computer vision, effective data augmentation for NLP is still challenging. This paper proposes LM-CPPF, Contrastive Paraphrasing-guided Prompt-based Fine-tuning of Language Models, which leverages prompt-based few-shot paraphrasing using generative language models, especially large language models such as GPT-3 and OPT-175B, for data augmentation. Our experiments on multiple text classification benchmarks show that this augmentation method outperforms other methods, such as easy data augmentation, back translation, and multiple templates.","{'model': 'tldr@v2.0.0', 'text': 'LM-CPPF, Contrastive Paraphrasing-guided Prompt-based Fine-tuning of Language Models, which leverages prompt-based few-shot paraphrasing using generative language models, especially large language models such as GPT-3 and OPT-175B, for data augmentation is proposed.'}",https://aclanthology.org/2023.acl-short.59.pdf
-syntax-aware hybrid prompt model for few-shot multi-modal sentiment analysis,Zikai Zhou,"Multimodal Sentiment Analysis (MSA) has been a popular topic in natural language processing nowadays, at both sentence and aspect level. However, the existing approaches almost require large-size labeled datasets, which bring about large consumption of time and resources. Therefore, it is practical to explore the method for few-shot sentiment analysis in cross-modalities. Previous works generally execute on textual modality, using the prompt-based methods, mainly two types: hand-crafted prompts and learnable prompts. The existing approach in few-shot multi-modality sentiment analysis task has utilized both methods, separately. We further design a hybrid pattern that can combine one or more fixed hand-crafted prompts and learnable prompts and utilize the attention mechanisms to optimize the prompt encoder. The experiments on both sentence-level and aspect-level datasets prove that we get a significant outperformance.","{'model': 'tldr@v2.0.0', 'text': 'A hybrid pattern that can combine one or more fixed hand-crafted prompts and learnable prompts and utilize the attention mechanisms to optimize the prompt encoder is designed.'}",https://arxiv.org/pdf/2306.01312
-enhancing few-shot ner with prompt ordering based data augmentation,Huiming Wang,"Recently, data augmentation (DA) methods have been proven to be effective for pre-trained language models (PLMs) in low-resource settings, including few-shot named entity recognition (NER). However, conventional NER DA methods are mostly aimed at sequence labeling models, i.e., token-level classification, and few are compatible with unified autoregressive generation frameworks, which can handle a wider range of NER tasks, such as nested NER. Furthermore, these generation frameworks have a strong assumption that the entities will appear in the target sequence with the same left-to-right order as the source sequence. In this paper, we claim that there is no need to keep this strict order, and more diversified but reasonable target entity sequences can be provided during the training stage as a novel DA method. Nevertheless, a naive mixture of augmented data can confuse the model since one source sequence will then be paired with different target sequences. Therefore, we propose a simple but effective Prompt Ordering based Data Augmentation (PODA) method to improve the training of unified autoregressive generation frameworks under few-shot NER scenarios. Experimental results on three public NER datasets and further analyses demonstrate the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple but effective Prompt Ordering based Data Augmentation (PODA) method to improve the training of unified autoregressive generation frameworks under few-shot NER scenarios.'}",http://arxiv.org/pdf/2305.11791
-few-shot composition learning for image retrieval with prompt tuning,Junda Wu,"We study the problem of composition learning for image retrieval, for which we learn to retrieve target images with search queries in the form of a composition of a reference image and a modification text that describes desired modifications of the image. Existing models of composition learning for image retrieval are generally built with large-scale datasets, demanding extensive training samples, i.e., query-target pairs, as supervision, which restricts their application for the scenario of few-shot learning with only few query-target pairs available. Recently, prompt tuning with frozen pretrained language models has shown remarkable performance when the amount of training data is limited. Inspired by this, we propose a prompt tuning mechanism with the pretrained CLIP model for the task of few-shot composition learning for image retrieval. Specifically, we regard the representation of the reference image as a trainable visual prompt, prefixed to the embedding of the text sequence. One challenge is to efficiently train visual prompt with few-shot samples. To deal with this issue, we further propose a self-upervised auxiliary task via ensuring that the reference image can retrieve itself when no modification information is given from the text, which facilitates training for the visual prompt, while not requiring additional annotations for query-target pairs. Experiments on multiple benchmarks show that our proposed model can yield superior performance when trained with only few query-target pairs.","{'model': 'tldr@v2.0.0', 'text': 'A self-upervised auxiliary task via ensuring that the reference image can retrieve itself when no modification information is given from the text, which facilitates training for the visual prompt, while not requiring additional annotations for query-target pairs is proposed.'}",https://ojs.aaai.org/index.php/AAAI/article/download/25597/25369
-image-object-specific prompt learning for few-shot class-incremental learning,In-Ug Yoon,"While many FSCIL studies have been undertaken, achieving satisfactory performance, especially during incremental sessions, has remained challenging. One prominent challenge is that the encoder, trained with an ample base session training set, often underperforms in incremental sessions. In this study, we introduce a novel training framework for FSCIL, capitalizing on the generalizability of the Contrastive Language-Image Pre-training (CLIP) model to unseen classes. We achieve this by formulating image-object-specific (IOS) classifiers for the input images. Here, an IOS classifier refers to one that targets specific attributes (like wings or wheels) of class objects rather than the image's background. To create these IOS classifiers, we encode a bias prompt into the classifiers using our specially designed module, which harnesses key-prompt pairs to pinpoint the IOS features of classes in each session. From an FSCIL standpoint, our framework is structured to retain previous knowledge and swiftly adapt to new sessions without forgetting or overfitting. This considers the updatability of modules in each session and some tricks empirically found for fast convergence. Our approach consistently demonstrates superior performance compared to state-of-the-art methods across the miniImageNet, CIFAR100, and CUB200 datasets. Further, we provide additional experiments to validate our learned model's ability to achieve IOS classifiers. We also conduct ablation studies to analyze the impact of each module within the architecture.","{'model': 'tldr@v2.0.0', 'text': 'A novel training framework for FSCIL is introduced, capitalizing on the generalizability of the Contrastive Language-Image Pre-training (CLIP) model to unseen classes, by formulating image-object-specific (IOS) classifiers for the input images.'}",https://arxiv.org/pdf/2309.02833
-participatory learning strategies for one-shot instruction sessions,Erin Durham,"Ø Elaboration: “How does this concept relate to other concepts?” Ø Distinctiveness: “What are the key differences between this concept and others? Ø Personal: “How does this concept relate to the student’s personal experiences?” Ø Application: “How are students expected to apply this concept?” -Stephen Chew, PhD, Cognitive Psychologist Deep Learning Strategies Cognitive psychologists have found that students can gain a deeper understanding of lesson concepts when instructors focus on: Post-It Note: Personal PostIt notes are placed at each desk and students are asked in the beginning of class to write a question they have about the library or their research. Notes are collected and reviewed during group work and then answered throughout the class session. This allows for student centered, relevant teaching. Group Brainstorm: Elaboration The class is divided into groups of 4-6 students and given a research topic to brainstorm for 5 minutes. Each group is given a white board with four questions that prompt discussion about the topic. The activity helps students expand their thinking and practice the skill of refining a topic to more narrow focus. Teaching and Learning Services http://www.lib.umd.edu/tl/first-year/engl-101 Jigsaw Formation: Distinctiveness The Jigsaw formation (developed by Elliot Aronson) engages deep learning by assigning students a portion of the class assignment to teach others. 1) Students are assigned a numbered group 1-4 and given sources to skim and evaluate for credibility. They are asked to become mini experts of those articles so they can teach their fellow classmates. 2) A student from each number group is assigned to a new group A-D and each shares her “expert knowledge” with the new group. The Jigsaw formation will be incorporated into select one-shot sessions in Fall 2017.","{'model': 'tldr@v2.0.0', 'text': 'The Jigsaw formation engages deep learning by assigning students a portion of the class assignment to teach others and the Post-It Note helps students expand their thinking and practice the skill of refining a topic to more narrow focus.'}",
-overcoming catastrophic forgetting in zero-shot cross-lingual generation,Tu Vu,"In this paper, we explore the challenging problem of performing a generative task in a target language when labeled data is only available in English, using summarization as a case study. We assume a strict setting with no access to parallel data or machine translation and find that common transfer learning approaches struggle in this setting, as a generative multilingual model fine-tuned purely on English catastrophically forgets how to generate non-English. Given the recent rise of parameter-efficient adaptation techniques, we conduct the first investigation into how one such method, prompt tuning (Lester et al., 2021), can overcome catastrophic forgetting to enable zero-shot cross-lingual generation. Our experiments show that parameter-efficient prompt tuning provides gains over standard fine-tuning when transferring between less-related languages, e.g., from English to Thai. However, a significant gap still remains between these methods and fully-supervised baselines. To improve cross-lingual transfer further, we explore several approaches, including: (1) mixing in unlabeled multilingual data, and (2) explicitly factoring prompts into recombinable language and task components. Our approaches can provide further quality gains, suggesting that robust zero-shot cross-lingual generation is within reach.","{'model': 'tldr@v2.0.0', 'text': 'This is the first investigation into how one such method, prompt tuning, can overcome catastrophic forgetting to enable zero-shot cross-lingual generation, and shows that parameter-efficient prompt tuning provides gains over standard fine-tuning when transferring between less-related languages, e.g., from English to Thai.'}",http://arxiv.org/pdf/2205.12647
-nearest neighbor zero-shot inference,Weijia Shi,"Retrieval-augmented language models (LMs) use non-parametric memory to substantially outperform their non-retrieval counterparts on perplexity-based evaluations, but it is an open question whether they achieve similar gains in few- and zero-shot end-task accuracy. We extensively study one such model, the k-nearest neighbor LM (kNN-LM), showing that the gains marginally transfer. The main challenge is to achieve coverage of the verbalizer tokens that define the different end-task class labels. To address this challenge, we also introduce kNN-Prompt, a simple and effective kNN-LM with automatically expanded fuzzy verbalizers (e.g. to expand “terrible” to also include “silly” and other task-specific synonyms for sentiment classification). Across nine diverse end-tasks, using kNN-Prompt with GPT-2 large yields significant performance boosts over strong zeroshot baselines (13.4% absolute improvement over the base LM on average). We also show that other advantages of non-parametric augmentation hold for end tasks; kNN-Prompt is effective for domain adaptation with no further training, and gains increase with the size of the retrieval model.","{'model': 'tldr@v2.0.0', 'text': 'KNN-Prompt is introduced, a simple and effective kNN-LM with automatically expanded fuzzy verbalizers that is effective for domain adaptation with no further training, and gains increase with the size of the retrieval model.'}",https://arxiv.org/pdf/2205.13792
-balanced distributed augmentation for multi-label few shot learning with prototypical network,Hamza Haruna Mohammed,"Many methods have been presented as a few shot learners in order to enhance few-shot learners. Some of these methods involve routine-based pre-trained language models and novel pipeline for automating the prompt generation. In this study, we propose a new evenly distributed data augmentation technique, which generates samples according to the probabilistic distribution of the relationship of each label with the mean of a label group. In the labeling phase, we present a semantic sentiment analysis approach in order to increase the realism of the data, in a more semantic augmentation way. The results show that this approach improves the few shot learners. In addition to this, we compare our adaptation approach to other traditional problem transformation methods. The newly developed approach outperforms these traditional methods, especially when the classifier learns from a limited number of samples.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a new evenly distributed data augmentation technique, which generates samples according to the probabilistic distribution of the relationship of each label with the mean of a label group, and presents a semantic sentiment analysis approach in order to increase the realism of the data.'}",
-kecp: knowledge enhanced contrastive prompting for few-shot extractive question answering,J. Wang,"Extractive Question Answering (EQA) is one of the most essential tasks in Machine Reading Comprehension (MRC), which can be solved by fine-tuning the span selecting heads of Pre-trained Language Models (PLMs). However, most existing approaches for MRC may perform poorly in the few-shot learning scenario. To solve this issue, we propose a novel framework named Knowledge Enhanced Contrastive Prompt-tuning (KECP). Instead of adding pointer heads to PLMs, we introduce a seminal paradigm for EQA that transforms the task into a non-autoregressive Masked Language Modeling (MLM) generation problem. Simultaneously, rich semantics from the external knowledge base (KB) and the passage context support enhancing the query’s representations. In addition, to boost the performance of PLMs, we jointly train the model by the MLM and contrastive learning objectives. Experiments on multiple benchmarks demonstrate that our method consistently outperforms state-of-the-art approaches in few-shot settings by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a seminal paradigm for EQA that transforms the task into a non-autoregressive Masked Language Modeling (MLM) generation problem and jointly train the model by the MLM and contrastive learning objectives.'}",http://arxiv.org/pdf/2205.03071
-cross-lingual retrieval augmented prompt for low-resource languages,Ercong Nie,"Multilingual Pretrained Language Models (MPLMs) have shown their strong multilinguality in recent empirical cross-lingual transfer studies. In this paper, we propose the Prompts Augmented by Retrieval Crosslingually (PARC) pipeline to improve the zero-shot performance on low-resource languages (LRLs) by augmenting the context with semantically similar sentences retrieved from a high-resource language (HRL) as prompts. PARC improves the zero-shot performance on three downstream tasks (binary sentiment classification, topic categorization and natural language inference) with multilingual parallel test sets across 10 LRLs covering 6 language families in both unlabeled settings (+5.1%) and labeled settings (+16.3%). PARC-labeled also outperforms the finetuning baseline by 3.7%. We find a significant positive correlation between cross-lingual transfer performance on one side, and the similarity between the high- and low-resource languages as well as the amount of low-resource pretraining data on the other side. A robustness analysis suggests that PARC has the potential to achieve even stronger performance with more powerful MPLMs.","{'model': 'tldr@v2.0.0', 'text': 'A robustness analysis suggests that PARC has the potential to achieve even stronger performance with more powerful MPLMs, and a significant positive correlation between cross-lingual transfer performance on one side, and the similarity between the high- and low-resource languages as well as the amount of low- resource pretraining data on the other side.'}",https://arxiv.org/pdf/2212.09651
-indirect: language-guided zero-shot deep metric learning for images,Konstantin Kobs,"Common Deep Metric Learning (DML) datasets specify only one notion of similarity, e.g., two images in the Cars196 dataset are deemed similar if they show the same car model. We argue that depending on the application, users of image retrieval systems have different and changing similarity notions that should be incorporated as easily as possible. Therefore, we present Language-Guided Zero-Shot Deep Metric Learning (LanZ-DML) as a new DML setting in which users control the properties that should be important for image representations without training data by only using natural language. To this end, we propose InDiReCT (Image representations using Dimensionality Reduction on CLIP embedded Texts), a model for LanZ-DML on images that exclusively uses a few text prompts for training. InDiReCT utilizes CLIP as a fixed feature extractor for images and texts and transfers the variation in text prompt embeddings to the image embedding space. Extensive experiments on five datasets and overall thirteen similarity notions show that, despite not seeing any images during training, InDiReCT performs better than strong baselines and approaches the performance of fully-supervised models. An analysis reveals that InDiReCT learns to focus on regions of the image that correlate with the desired similarity notion, which makes it a fast to train and easy to use method to create custom embedding spaces only using natural language.","{'model': 'tldr@v2.0.0', 'text': 'Language-Guided Zero-Shot Deep Metric Learning (LanZ-DML) is presented as a new DML setting in which users control the properties that should be important for image representations without training data by only using natural language.'}",https://arxiv.org/pdf/2211.12760
-list: lite self-training makes efficient few-shot learners,Yaqing Wang,"We present a new method LiST 1 for efficient fine-tuning of large pre-trained language models (PLMs) in few-shot learning settings. LiST significantly improves over recent methods that adopt prompt-tuning using two key techniques. The first one is the use of self-training to leverage large amounts of unlabeled data for prompt-tuning to significantly boost the model performance in few-shot settings. We use self-training in conjunction with meta-learning for re-weighting noisy pseudo-prompt labels. However, traditional self-training is expensive as it requires updating all the model parameters repetitively. Therefore, we use a second technique for light-weight fine-tuning where we introduce a small number of task-specific adapter parameters that are fine-tuned during self-training while keeping the PLM encoder frozen. This also significantly reduces the overall model footprint across several tasks that can now share a common PLM encoder as backbone for inference. Combining the above techniques, LiST not only improves the model performance for few-shot learning on target domains but also reduces the model memory footprint. We present a comprehensive study on six NLU tasks to validate the effectiveness of LiST . The results show that LiST improves by 35% over classic fine-tuning methods and 6% over prompt-tuning with 96% reduction in number of trainable parameters when fine-tuned with no more than 30 labeled examples from each target domain.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive study on six NLU tasks to validate the effectiveness of LiST and shows that LiST improves by 35% over classic fine- Tuning methods and 6% over prompt-tuning with 96% reduction in number of trainable parameters when flne-tuned with no more than 30 labeled examples from each target domain.'}",
-prompt-based multi-modal image segmentation,Timo Lüddecke,"Image segmentation is usually addressed by training a model for a fixed set of object classes. Incorporating additional classes or more complex queries later is expensive as it requires re-training the model on a dataset that encompasses these expressions. Here we propose a system that can generate image segmentations based on arbitrary prompts at test time. A prompt can be either a text or an image. This approach enables us to create a unified model (trained once) for three common segmentation tasks, which come with distinct challenges: referring expression segmentation, zero-shot segmentation and one-shot segmentation. We build upon the CLIP model as a backbone which we extend with a transformer-based decoder that enables dense prediction. After training on an extended version of the PhraseCut dataset, our system generates a binary segmentation map for an image based on a free-text prompt or on an additional image expressing the query. Different variants of the latter image-based prompts are analyzed in detail. This novel hybrid input allows for dynamic adaptation not only to the three segmentation tasks mentioned above, but to any binary segmentation task where a text or image query can be formulated. Finally, we find our system to adapt well to generalized queries involving affordances or properties. Source code: https://eckerlab.org/code/clipseg","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a system that can generate image segmentations based on arbitrary prompts at test time with a transformer-based decoder that enables dense prediction and allows for dynamic adaptation to generalized queries involving affordances or properties.'}",
-jurassic is (almost) all you need: few-shot meaning-to-text generation for open-domain dialogue,Lena Reed,,"{'model': 'tldr@v2.0.0', 'text': 'These are the first results demonstrating that few-shot semantic prompt-based learning can create NLGs that generalize to new domains, and produce high-quality, semantically-controlled, conversational responses directly from meaning representations.'}",https://arxiv.org/pdf/2110.08094
-prompt scoring system for dialogue summarization using gpt-3,George Prodan,"Recent results in language processing show that language models are capable of performing several natural language tasks without the need of supervised learning. A challenging task for pre-trained language models is dialogue summarization. One way of generating summaries is engineering prompt templates for few-shot training. However, a static approach of creating prompts leads to unreliable outcomes between different classes of dialogues. Focusing on the dialogues structure properties we propose a scoring system to improve the few-shot training performances. We build tuned prompts composed by the highest scored dialogue samples. Our evaluation based on ROUGE scores and human evaluation shows that there is an improvement for the experiments in which we use the score system. All experiments are performed within the framework of the GPT-3 API. We use different engines for comparison. Moreover, the human evaluation we conducted showed that the number of failures decreased by 11\% after applying our scoring system.","{'model': 'tldr@v2.0.0', 'text': 'Focusing on the dialogues structure properties, this work proposes a scoring system to improve the few-shot training performances of dialogue summarization and builds tuned prompts composed by the highest scored dialogue samples.'}",https://www.techrxiv.org/articles/preprint/Prompt_scoring_system_for_dialogue_summarization_using_GPT-3/16652392/2/files/35289613.pdf
-inverse is better! fast and accurate prompt for slot tagging,,"Prompting methods recently achieve impres- 001 sive success in few-shot learning. These meth- 002 ods embed input samples with prompt sentence 003 pieces and decode label-related tokens to map 004 samples to the label. However, such a paradigm 005 is very inefficient for the task of slot tagging. 006 Because the slot tagging samples are multiple 007 consecutive words in a sentence, the prompting 008 methods have to enumerate all n-grams token 009 span to find all the possible slots, which greatly 010 slows down the prediction. To tackle this, we 011 introduce an inverse paradigm for prompting. 012 Different from the classic prompts map tokens 013 to labels, we reversely predict slot values given 014 slot types. Such inverse prompting only re- 015 quires a one-turn prediction for each slot type 016 and greatly speeds up the prediction. Besides, 017 we propose a novel Iterative Prediction Strat- 018 egy, from which the model learns to refine pre- 019 dictions by considering the relations between 020 different slot types. We find, somewhat surpris- 021 ingly, the proposed method not only predicts 022 faster, but also significantly improves the effect 023 (improve over 6 . 1 F1-scores on 10-shot setting) 024 and achieves new state-of-the-art performance. 025","{'model': 'tldr@v2.0.0', 'text': 'The proposed inverse prompting method not only predicts faster, but also significantly improves the effect 023 (improve over 6 . 1 F1-scores on 10-shot setting) 024 and achieves new state-of-the-art performance.'}",
-is a prompt and a few samples all you need? using gpt-4 for data augmentation in low-resource classification tasks,Anders Giovanni Møller,"Obtaining and annotating data can be expensive and time-consuming, especially in complex, low-resource domains. We use GPT-4 and ChatGPT to augment small labeled datasets with synthetic data via simple prompts, in three different classification tasks with varying complexity. For each task, we randomly select a base sample of 500 texts to generate 5,000 new synthetic samples. We explore two augmentation strategies: one that preserves original label distribution and another that balances the distribution. Using a progressively larger training sample size, we train and evaluate a 110M parameter multilingual language model on the real and synthetic data separately. We also test GPT-4 and ChatGPT in a zero-shot setting on the test sets. We observe that GPT-4 and ChatGPT have strong zero-shot performance across all tasks. We find that data augmented with synthetic samples yields a good downstream performance, and particularly aids in low-resource settings, such as in identifying rare classes. Human-annotated data exhibits a strong predictive power, overtaking synthetic data in two out of the three tasks. This finding highlights the need for more complex prompts for synthetic datasets to consistently surpass human-generated ones.","{'model': 'tldr@v2.0.0', 'text': 'This work uses GPT-4 and ChatGPT to augment small labeled datasets with synthetic data via simple prompts, in three different classification tasks with varying complexity, and finds that data augmented with synthetic samples yields a good downstream performance, and particularly aids in low-resource settings, such as in identifying rare classes.'}",http://arxiv.org/pdf/2304.13861
-structure pretraining and prompt tuning for knowledge graph transfer,Wen Zhang,"Knowledge graphs (KG) are essential background knowledge providers in many tasks. When designing models for KG-related tasks, one of the key tasks is to devise the Knowledge Representation and Fusion (KRF) module that learns the representation of elements from KGs and fuses them with task representations. While due to the difference of KGs and perspectives to be considered during fusion across tasks, duplicate and ad hoc KRF modules design are conducted among tasks. In this paper, we propose a novel knowledge graph pretraining model KGTransformer that could serve as a uniform KRF module in diverse KG-related tasks. We pretrain KGTransformer with three self-supervised tasks with sampled sub-graphs as input. For utilization, we propose a general prompt-tuning mechanism regarding task data as a triple prompt to allow flexible interactions between task KGs and task data. We evaluate pretrained KGTransformer on three tasks, triple classification, zero-shot image classification, and question answering. KGTransformer consistently achieves better results than specifically designed task models. Through experiments, we justify that the pretrained KGTransformer could be used off the shelf as a general and effective KRF module across KG-related tasks. The code and datasets are available at https://github.com/zjukg/KGTransformer.","{'model': 'tldr@v2.0.0', 'text': 'A novel knowledge graph pretraining model KGTransformer is proposed that could serve as a uniform KRF module in diverse KG-related tasks and consistently achieves better results than specifically designed task models.'}",https://arxiv.org/pdf/2303.03922
-residual prompt tuning: improving prompt tuning with residual reparameterization,Anastasia Razdaibiedina,"Prompt tuning is one of the successful approaches for parameter-efficient tuning of pre-trained language models. Despite being arguably the most parameter-efficient (tuned soft prompts constitute<0.1% of total parameters), it typically performs worse than other efficient tuning methods and is quite sensitive to hyper-parameters. In this work, we introduce Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning. We propose to reparameterize soft prompt embeddings using a shallow network with a residual connection. Our experiments show that Residual Prompt Tuning significantly outperforms prompt tuning on SuperGLUE benchmark. Notably, our method reaches +7 points improvement over prompt tuning with T5-Base and allows to reduce the prompt length by 10x without hurting performance. In addition, we show that our approach is robust to the choice of learning rate and prompt initialization, and is effective in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Residual Prompt Tuning - a simple and efficient method that significantly improves the performance and stability of prompt tuning and proposes to reparameterize soft prompt embeddings using a shallow network with a residual connection.'}",http://arxiv.org/pdf/2305.03937
-ds4dh at mediqa-chat 2023: leveraging svm and gpt-3 prompt engineering for medical dialogue classification and summarization,Boya Zhang,"This paper presents the results of the Data Science for Digital Health (DS4DH) group in the MEDIQA-Chat Tasks at ACL-ClinicalNLP 2023. Our study combines the power of a classical machine learning method, Support Vector Machine, for classifying medical dialogues, along with the implementation of one-shot prompts using GPT-3.5. We employ dialogues and summaries from the same category as prompts to generate summaries for novel dialogues. Our findings exceed the average benchmark score, offering a robust reference for assessing performance in this field.","{'model': 'tldr@v2.0.0', 'text': 'This study combines the power of a classical machine learning method, Support Vector Machine, for classifying medical dialogues, along with the implementation of one-shot prompts using GPT-3.5 to generate summaries for novel dialogues.'}",https://access.archive-ouverte.unige.ch/access/metadata/290c4289-0017-45ec-baa9-ff2fdd7948f9/download
-soft prompt tuning for augmenting dense retrieval with large language models,Zhiyuan Peng,"Dense retrieval (DR) converts queries and documents into dense embeddings and measures the similarity between queries and documents in vector space. One of the challenges in DR is the lack of domain-specific training data. While DR models can learn from large-scale public datasets like MS MARCO through transfer learning, evidence shows that not all DR models and domains can benefit from transfer learning equally. Recently, some researchers have resorted to large language models (LLMs) to improve the zero-shot and few-shot DR models. However, the hard prompts or human-written prompts utilized in these works cannot guarantee the good quality of generated weak queries. To tackle this, we propose soft prompt tuning for augmenting DR (SPTAR): For each task, we leverage soft prompt-tuning to optimize a task-specific soft prompt on limited ground truth data and then prompt the LLMs to tag unlabeled documents with weak queries, yielding enough weak document-query pairs to train task-specific dense retrievers. We design a filter to select high-quality example document-query pairs in the prompt to further improve the quality of weak tagged queries. To the best of our knowledge, there is no prior work utilizing soft prompt tuning to augment DR models. The experiments demonstrate that SPTAR outperforms the unsupervised baselines BM25 and the recently proposed LLMs-based augmentation method for DR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes soft prompt tuning for augmenting DR (SPTAR) by leveraging soft prompt-tuning to optimize a task-specific soft prompt on limited ground truth data and prompts the LLMs to tag unlabeled documents with weak queries, yielding enough weak document-query pairs to train task- specific dense retrievers.'}",https://arxiv.org/pdf/2307.08303
-self-prompting large vision models for few-shot medical image segmentation,Qi Wu,"Recent advancements in large foundation models have shown promising potential in the medical industry due to their flexible prompting capability. One such model, the Segment Anything Model (SAM), a prompt-driven segmentation model, has shown remarkable performance improvements, surpassing state-of-the-art approaches in medical image segmentation. However, existing methods primarily rely on tuning strategies that require extensive data or prior prompts tailored to the specific task, making it particularly challenging when only a limited number of data samples are available. In this paper, we propose a novel perspective on self-prompting in medical vision applications. Specifically, we harness the embedding space of SAM to prompt itself through a simple yet effective linear pixel-wise classifier. By preserving the encoding capabilities of the large model, the contextual information from its decoder, and leveraging its interactive promptability, we achieve competitive results on multiple datasets (i.e. improvement of more than 15% compared to fine-tuning the mask decoder using a few images).","{'model': 'tldr@v2.0.0', 'text': 'A novel perspective on self-prompting in medical vision applications is proposed by harnessing the embedding space of SAM to prompt itself through a simple yet effective linear pixel-wise classifier.'}",https://arxiv.org/pdf/2308.07624
-cipta: contrastive-based iterative prompt-tuning using text annotation from large language models,Yuan Yan,"In recent years, public opinion analysis has become increasingly important due to the widespread use of social media platforms and the growing influence of online information on public security. Prompt tuning, a typical few-shot learning method, ensures that the model quickly adapts to opinion analysis with different classification rules. However, existing prompt tuning for opinion analysis cannot guarantee the effectiveness of the model in zero-shot or one-shot cases. In this study, we propose the Contrastive-based Iterative Prompt-tuning method using Text-Annotation from Large Language Models (LLMs), CIPTA, for low-resource public opinion analysis. Specifically, with a small amount of manually labeled data, CIPTA leverages the knowledge from LLMs to text annotation and utilizes unsupervised contrastive embedding training to optimize text representation. Based on the prompt tuning method and the iterative training over unlabeled data, the model further utilizes the knowledge from the pre-training stage. Experiment results on tweet data show that our CIPTA achieves encouraging performance in public opinion analysis.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes the Contrastive-based Iterative Prompt-tuning method using Text-Annotation from Large Language Models, CIPTA, for low-resource public opinion analysis that leverages the knowledge from LLMs to text annotation and utilizes unsupervised contrastive embedding training to optimize text representation.'}",
-unleashing the potential of prompt engineering in large language models: a comprehensive review,Banghao Chen,"This paper delves into the pivotal role of prompt engineering in unleashing the capabilities of Large Language Models (LLMs). Prompt engineering is the process of structuring input text for LLMs and is a technique integral to optimizing the efficacy of LLMs. This survey elucidates foundational principles of prompt engineering, such as role-prompting, one-shot, and few-shot prompting, as well as more advanced methodologies such as the chain-of-thought and tree-of-thoughts prompting. The paper sheds light on how external assistance in the form of plugins can assist in this task, and reduce machine hallucination by retrieving external knowledge. We subsequently delineate prospective directions in prompt engineering research, emphasizing the need for a deeper understanding of structures and the role of agents in Artificial Intelligence-Generated Content (AIGC) tools. We discuss how to assess the efficacy of prompt methods from different perspectives and using different methods. Finally, we gather information about the application of prompt engineering in such fields as education and programming, showing its transformative potential. This comprehensive survey aims to serve as a friendly guide for anyone venturing through the big world of LLMs and prompt engineering.",,
-multiple brain abscesses by an air gun shot: a case report,M. Jamali,"Abstract Introduction Low-velocity penetrating brain injury is not prevalent. In some conditions such as childhood, and with the penetration of a pellet in weak spots of skull, low-velocity penetrating brain injury is expected; however, high-velocity projectiles have also been reported as the cause of severe brain injuries. One of the complications of penetrating brain injury is infection, in which different types of microorganisms play a role. The Streptococcus genus is the leading cause of abscess formation in non-traumatic patients. Multiple brain abscesses are not common. Case Presentation A 10-year-old boy with penetrating brain injury caused by an air gun pellet, who developed signs and symptoms of high intracranial pressure 18 days after the trauma. After the imaging scans and the detection of multiple brain abscesses and severe brain edema, prompt surgical intervention was performed for all three lesions in a single operation. The culture of a pus specimen was positive for Streptococcus species, and, with adequate antibiotic therapy, the patient was discharged from the hospital in good condition. Conclusion Brain injury with air gun shot is not prevalent. The penetration of a low-velocity air gun pellet in weak points of the skull (such as the orbit, the squamous portion of the temporal bone, and the cranial suture), specially in children, can cause significant brain injuries.","{'model': 'tldr@v2.0.0', 'text': 'Brain injury with air gun shot is not prevalent; however, the penetration of a low-velocity air gun pellet in weak points of the skull (such as the orbit, the squamous portion of the temporal bone, and the cranial suture) can cause significant brain injuries.'}",
-multi-mask label mapping for prompt-based learning,Jirui Qi,"Prompt-based Learning has shown significant success in few-shot classification. The mainstream approach is to concatenate a template for the input text to transform the classification task into a cloze-type task where label mapping plays an important role in finding the ground-truth labels. While current label mapping methods only use the contexts in one single input, it could be crucial if wrong information is contained in the text. Specifically, it is proved in recent work that even the large language models like BERT/RoBERTa make classification decisions heavily dependent on a specific keyword regardless of the task or the context. Such a word is referred to as a lexical cue and if a misleading lexical cue is included in the instance it will lead the model to make a wrong prediction. We propose a multi-mask prompt-based approach with Multi-Mask Label Mapping (MMLM) to reduce the impact of misleading lexical cues by allowing the model to exploit multiple lexical cues. To satisfy the conditions of few-shot learning, an instance augmentation approach for the cloze-type model is proposed and the misleading cues are gradually excluded through training. We demonstrate the effectiveness of MMLM by both theoretical analysis and empirical studies, and show that MMLM outperforms other existing label mapping approaches.","{'model': 'tldr@v2.0.0', 'text': 'To satisfy the conditions of few-shot learning, an instance augmentation approach for the cloze-type model is proposed and the misleading cues are gradually excluded through training to demonstrate the effectiveness of MMLM by both theoretical analysis and empirical studies, and it is shown that M MLM outperforms other existing label mapping approaches.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26579/26351
-prompts can play lottery tickets well: achieving lifelong information extraction via lottery prompt tuning,Zujie Liang,"Thanks to the recent success of Pre-trained Language Models (PLMs), it has become a promising research direction to develop a universal model (UIE) that can solve all typical information extraction tasks within one generative framework. Nonetheless, in real-world scenarios of UIE applications, new data of different IE tasks and domains usually come in a stream over time. A desirable UIE system should be capable of continually learning new tasks without forgetting old ones, thereby allowing knowledge and functionalities expansion without re-training the whole system. In this paper, we study the UIE system under a more challenging yet practical scenario, i.e., “lifelong learning” settings, to evaluate its abilities in three aspects, including knowledge sharing and expansion, catastrophic forgetting prevention, and rapid generalization on few-shot and unseen tasks.To achieve these three goals, we present a novel parameter- and deployment-efficient prompt tuning method namely Lottery Prompt Tuning (LPT).LPT freezes the PLM’s parameters and sequentially learns compact pruned prompt vectors for each task leveraging a binary prompt mask, while keeping the prompt parameters selected by the previous tasks insusceptible.Furthermore, we use a simple yet effective method to perform mask selection and show the powerful transferability of Lottery Prompts to novel tasks.Extensive experiments demonstrate that LPT consistently sets state-of-the-art performance on multiple lifelong learning settings of UIE, including task-incremental setting on seen tasks, few-shot adaptation, and zero-shot generalization on novel tasks.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments demonstrate that LPT consistently sets state-of-the-art performance on multiple lifelong learning settings of UIE, including task-incremental setting on seen tasks, few-shot adaptation, and zero-shot generalization on novel tasks.'}",https://aclanthology.org/2023.acl-long.16.pdf
-llm comparative assessment: zero-shot nlg evaluation through pairwise comparisons using large language models,Adian Liusie,"Current developments in large language models (LLMs) have enabled impressive zero-shot capabilities across various natural language tasks. An interesting application of these systems is in the automated assessment of natural language generation (NLG), a highly challenging area with great practical benefit. In this paper, we explore two options for exploiting the emergent abilities of LLMs for zero-shot NLG assessment: absolute score prediction, and comparative assessment which uses relative comparisons between pairs of candidates. Though comparative assessment has not been extensively studied in NLG assessment, we note that humans often find it more intuitive to compare two options rather than scoring each one independently. This work examines comparative assessment from multiple perspectives: performance compared to absolute grading; positional biases in the prompt; and efficient ranking in terms of the number of comparisons. We illustrate that LLM comparative assessment is a simple, general and effective approach for NLG assessment. For moderate-sized open-source LLMs, such as FlanT5 and Llama2-chat, comparative assessment is superior to prompt scoring, and in many cases can achieve performance competitive with state-of-the-art methods. Additionally, we demonstrate that LLMs often exhibit strong positional biases when making pairwise comparisons, and we propose debiasing methods that can further improve performance.","{'model': 'tldr@v2.0.0', 'text': 'It is illustrated that LLM comparative assessment is a simple, general and effective approach for NLG assessment, and for moderate-sized open-source LLMs, superior to prompt scoring, and in many cases can achieve performance competitive with state-of-the-art methods.'}",
-tuning multi-mode token-level prompt alignment across modalities,Dongsheng Wang,"Advancements in prompt tuning of vision-language models have underscored their potential in enhancing open-world visual concept comprehension. However, prior works only primarily focus on single-mode (only one prompt for each modality) and holistic level (image or sentence) semantic alignment, which fails to capture the sample diversity, leading to sub-optimal prompt discovery. To address the limitation, we propose a multi-mode token-level tuning framework that leverages the optimal transportation to learn and align a set of prompt tokens across modalities. Specifically, we rely on two essential factors: 1) multi-mode prompts discovery, which guarantees diverse semantic representations, and 2) token-level alignment, which helps explore fine-grained similarity. Consequently, the similarity can be calculated as a hierarchical transportation problem between the modality-specific sets. Extensive experiments on popular image recognition benchmarks show the superior generalization and few-shot abilities of our approach. The qualitative analysis demonstrates that the learned prompt tokens have the ability to capture diverse visual concepts.","{'model': 'tldr@v2.0.0', 'text': 'A multi-mode token-level tuning framework that leverages the optimal transportation to learn and align a set of prompt tokens across modalities and demonstrates that the learned prompt tokens have the ability to capture diverse visual concepts.'}",https://arxiv.org/pdf/2309.13847
-[cls] token is all you need for zero-shot semantic segmentation,Letian Wu,"In this paper, we propose an embarrassingly simple yet highly effective zero-shot semantic segmentation (ZS3) method, based on the pre-trained vision-language model CLIP. First, our study provides a couple of key discoveries: (i) the global tokens (a.k.a [CLS] tokens in Transformer) of the text branch in CLIP provide a powerful representation of semantic information and (ii) these text-side [CLS] tokens can be regarded as category priors to guide CLIP visual encoder pay more attention on the corresponding region of interest. Based on that, we build upon the CLIP model as a backbone which we extend with a One-Way [CLS] token navigation from text to the visual branch that enables zero-shot dense prediction, dubbed \textbf{ClsCLIP}. Specifically, we use the [CLS] token output from the text branch, as an auxiliary semantic prompt, to replace the [CLS] token in shallow layers of the ViT-based visual encoder. This one-way navigation embeds such global category prior earlier and thus promotes semantic segmentation. Furthermore, to better segment tiny objects in ZS3, we further enhance ClsCLIP with a local zoom-in strategy, which employs a region proposal pre-processing and we get ClsCLIP+. Extensive experiments demonstrate that our proposed ZS3 method achieves a SOTA performance, and it is even comparable with those few-shot semantic segmentation methods.","{'model': 'tldr@v2.0.0', 'text': 'This study provides a couple of key discoveries: (i) the global tokens of the text branch in CLIP provide a powerful representation of semantic information and (ii) these text-side [CLS] tokens can be regarded as category priors to guide CLIP visual encoder pay more attention on the corresponding region of interest.'}",http://arxiv.org/pdf/2304.06212
-metricprompt: prompting model as a relevance metric for few-shot text classification,Hongyuan Dong,"Prompting methods have shown impressive performance in a variety of text mining tasks and applications, especially few-shot ones. Despite the promising prospects, the performance of prompting model largely depends on the design of prompt template and verbalizer. In this work, we propose MetricPrompt, which eases verbalizer design difficulty by reformulating few-shot text classification task into text pair relevance estimation task. MetricPrompt adopts prompting model as the relevance metric, further bridging the gap between Pre-trained Language Model's (PLM) pre-training objective and text classification task, making possible PLM's smooth adaption. Taking a training sample and a query one simultaneously, MetricPrompt captures cross-sample relevance information for accurate relevance estimation. We conduct experiments on three widely used text classification datasets across four few-shot settings. Results show that MetricPrompt outperforms manual verbalizer and other automatic verbalizer design methods across all few-shot settings, achieving new state-of-the-art (SOTA) performance.","{'model': 'tldr@v2.0.0', 'text': ""MetricPrompt is proposed, which eases verbalizer design difficulty by reformulating few-shot text classification task into text pair relevance estimation task, and adopts prompting model as the relevance metric, further bridging the gap between Pre-trained Language Model's (PLM) pre-training objective and text classificationtask.""}",https://arxiv.org/pdf/2306.08892
-unsupervised dual modality prompt learning for facial expression recognition,Muhammad Shahid,"A method of facial expression recognition using a vision language model is proposed. Recently vision-language models for example CLIP (Contrastive Language-Image Pre-training) models developed by OpenAI have achieved exceptional results on a variety of image recognition and retrieval tasks, exhibiting strong zero-shot performance. Transferable representations can be adapted through prompt tuning to a variety of downstream tasks. From the general knowledge stored in a pre-trained model, prompt tuning attempts to extract useful information for downstream tasks. In order to avoid time-consuming prompt engineering, recent works use a small amount of labeled data for adapting vision language models to downstream image recognition problems. However, requiring target datasets to be labeled may restrict their scalability. Moreover, we also note that adapting prompt learning techniques in only one branch of CLIP (vision or language) is suboptimal because it won't allow for the dynamic adjustment of both representation spaces on a downstream task. In this paper, we evaluated the performance of the CLIP model as a zero-shot face recognizer and proposed an Unsupervised Dual Modality Prompt Learning framework for Facial Expression Recognition. Our model tunes the prompts through learning text and visual prompts simultaneously to improve alignment between the linguistic and visual representations when labels are not provided for the target dataset. The experimental results on CK+, JAFFE, RAF-DB, and FER2013 datasets showed that our proposed method performs better compared with CLIP Zero-Shot and other unsupervised prompt-based learning methods for facial expression recognition tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper evaluated the performance of the CLIP model as a zero-shot face recognizer and proposed an Unsupervised Dual Modality Prompt Learning framework for Facial Expression Recognition that performs better compared with CLIP Zero-Shot and other unsupervised prompt-based learning methods for facial expression recognition tasks.'}",
-an attempt at zero-shot ancient documents restoration based on diffusion models,Hayata Kaneko,"The virtual restoration of ancient documents using deep learning is an emergency and an expected work. However, GANs-based image-to-image translation approaches hit the degradation data shortage, a hardness to build one-to-many restoration models, and a limitation for large deformation. In this study, we apply zero-shot restoration based on Diffusion models to ancient degraded documents, specifically, leverage inpainting of Denoing Diffusion Restoration Models (DDRM) for missing ancient characters. Furthermore, we introduce a noise masking method, which limits the attention area of predicted noise images in the reverse process. Noise masking forces DDRM to generate faithful objects following mask images, so that has high usability without re-training of deep neural networks. The zero-shot restoration and noise masking prompt GUI-connecting restoration of missing characters, leading to realizing a cooperative application with humans for ancient document restoration.","{'model': 'tldr@v2.0.0', 'text': 'This study applies zero-shot restoration based on Diffusion models to ancient degraded documents and introduces a noise masking method, which limits the attention area of predicted noise images in the reverse process, leading to cooperative application with humans for ancient document restoration.'}",
-label-aware automatic verbalizer for few-shot text classification,Thanakorn Thaminkaew,"Prompt-based learning has shown its effectiveness in few-shot text classification. One important factor in its success is a verbalizer, which translates output from a language model into a predicted class. Notably, the simplest and widely acknowledged verbalizer employs manual labels to represent the classes. However, manual selection does not guarantee the optimality of the selected words when conditioned on the chosen language model. Therefore, we propose Label-Aware Automatic Verbalizer (LAAV), effectively augmenting the manual labels to achieve better few-shot classification results. Specifically, we use the manual labels along with the conjunction""and""to induce the model to generate more effective words for the verbalizer. The experimental results on five datasets across five languages demonstrate that LAAV significantly outperforms existing verbalizers. Furthermore, our analysis reveals that LAAV suggests more relevant words compared to similar approaches, especially in mid-to-low resource languages.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Label-Aware Automatic Verbalizer (LAAV), effectively augmenting the manual labels to achieve better few-shot classification results and reveals that LAAV suggests more relevant words compared to similar approaches, especially in mid-to-low resource languages.'}",
-wide-field one-shot optical polarimeter: howpol,K. Kawabata,"For prompt optical polarimetry of gamma-ray burst (GRB) afterglow, we require wide-field imaging polarimeter which can produce both Stokes Q and U parameters from only a single exposure, as well as quickly-moving telescope and enclosure system. HOWPol is an optical imaging polarimeter which provides four linearly polarized images at position angles of 0°, 45°, 90° and 135°, i.e., Stokes I, Q, U, simultaneously. The key device is the wedged double Wollaston prism described by Oliva (1997)1 and Pernechele et al. (2003).2 The images are focused on two 2k×4k fully depleted CCDs. We report the design and development of the optical devices of HOWPol, which will be mounted to the 1.5-m Kanata telescope at Hiroshima University and stand by the GRB alert.",,
-one-shot illusory correlations and stereotype formation,Jane L. Risen,"In four studies, the authors explored the emergence of one-shot illusory correlations—in which a single instance of unusual behavior by a member of a rare group is sufficient to create an association between group and behavior. In Studies 1, 2, and 3, unusual behaviors committed by members of rare groups were processed differently than other types of behaviors. They received more processing time, prompted more attributional thinking, and were more memorable. In Study 4, the authors obtained evidence from two implicit measures of association that one-shot illusory correlations are generalized to other members of a rare group. The authors contend that one-shot illusory correlations arise because unusual pairings of behaviors and groups uniquely prompt people to entertain group membership as an explanation of the unusual behavior.","{'model': 'tldr@v2.0.0', 'text': 'The authors contend that one-shot illusory correlations arise because unusual pairings of behaviors and groups uniquely prompt people to entertain group membership as an explanation of the unusual behavior.'}",
-few-shot tweet detection in emerging disaster events,Anna Kruspe,"Social media sources can provide crucial information in crisis situations, but discovering relevant messages is not trivial. Methods have so far focused on universal detection models for all kinds of crises or for certain crisis types (e.g. floods). Event-specific models could implement a more focused search area, but collecting data and training new models for a crisis that is already in progress is costly and may take too much time for a prompt response. As a compromise, manually collecting a small amount of example messages is feasible. Few-shot models can generalize to unseen classes with such a small handful of examples, and do not need be trained anew for each event. We compare how few-shot approaches (matching networks and prototypical networks) perform for this task. Since this is essentially a one-class problem, we also demonstrate how a modified one-class version of prototypical models can be used for this application.","{'model': 'tldr@v2.0.0', 'text': 'This work compares how few-shot approaches (matching networks and prototypical networks) perform for this task and demonstrates how a modified one-class version of prototypical models can be used for this application.'}",
-shot scale matters: the effect of close-up frequency on theory of mind responses in film viewers,K. Bálint,"Theory of mind (ToM) processing is crucial in everyday social functioning. Fictional film narratives can elicit ToM, however, little is known about the cinematographic techniques underlying this effect. The present experiment focuses on the role of close-up shots of the character’s face in viewers’ ToM responding, as well as in their cognitive and affective processing more generally. The online experiment (N = 495) included thirteen versions of an animated film and employed a 6 (Close-up frequency, 1 vs. 2. vs. 3 vs. 4 vs. 5 vs. 10) × 2 (Facial expression, neutral vs. sad) plus a control condition with zero close-up factorial between-subject design. Participants were randomly assigned to one version of the film and subsequently asked (with and without a prompt for ToM) to recall the story. In these free responses, the study used a quantitative content analytic method (with independent blind-coders) to identify the level of spontaneous and prompted ToM, as well as cognitive and affective processing employed by viewers. Additionally, we tested the moderation effect of character facial expression (in the close-up) and participant gender. Confirming our main hypothesis, close-up frequency significantly influenced ToM-spontaneous, but not ToM-prompted. Results indicate that increasing the number of close-ups may elicit higher levels of ToM-spontaneous up to a certain point, beyond which it may decrease spontaneous ToM responses. Results suggest that the effect of close-up frequency is specific to ToM processing rather than some general effect on cognitive and affective processing of narratives.",,
-speak foreign languages with your own voice: cross-lingual neural codec language modeling,Zi-Hua Zhang,"We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID. Audio samples are available at \url{https://aka.ms/vallex}.","{'model': 'tldr@v2.0.0', 'text': ""Experimental results show that VALL-E X can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment.""}",http://arxiv.org/pdf/2303.03926
-pointclip v2: adapting clip for powerful 3d open-world learning,Xiangyang Zhu,"Contrastive Language-Image Pre-training (CLIP) has shown promising open-world performance on 2D image tasks, while its transferred capacity on 3D point clouds, i.e., PointCLIP, is still far from satisfactory. In this work, we propose PointCLIP V2, a powerful 3D open-world learner, to fully unleash the potential of CLIP on 3D point cloud data. First, we introduce a realistic shape projection module to generate more realistic depth maps for CLIP’s visual encoder, which is quite efficient and narrows the domain gap between projected point clouds with natural images. Second, we leverage large-scale language models to automatically design a more descriptive 3D-semantic prompt for CLIP’s textual encoder, instead of the previous hand-crafted one. Without introducing any training in 3D domains, our approach significantly surpasses Point-CLIP by +42.90% , +40.44% , and +28.75% accuracy on three datasets for zero-shot 3D classification. Furthermore, PointCLIP V2 can be extended to few-shot classifi-cation, zero-shot part segmentation, and zero-shot 3D object detection in a simple manner, demonstrating our superior generalization ability for 3D open-world learning. Code will be available at https://github.com/","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PointCLIP V2, a powerful 3D open-world learner, to fully unleash the potential of CLIP on 3D point cloud data, and introduces a realistic shape projection module to generate more realistic depth maps for CLIP’s visual encoder.'}",https://arxiv.org/pdf/2211.11682
-image segmentation using text and image prompts,Timo Lüddecke,"Image segmentation is usually addressed by training a model for a fixed set of object classes. Incorporating additional classes or more complex queries later is expensive as it requires re-training the model on a dataset that encompasses these expressions. Here we propose a system that can generate image segmentations based on arbitrary prompts at test time. A prompt can be either a text or an image. This approach enables us to create a unified model (trained once) for three common segmentation tasks, which come with distinct challenges: referring expression segmentation, zero-shot segmentation and one-shot segmentation. We build upon the CLIP model as a backbone which we extend with a transformer-based decoder that enables dense prediction. After training on an extended version of the PhraseCut dataset, our system generates a binary segmentation map for an image based on a free-text prompt or on an additional image expressing the query. We analyze different variants of the latter image-based prompts in detail. This novel hybrid input allows for dynamic adaptation not only to the three segmentation tasks mentioned above, but to any binary segmentation task where a text or image query can be formulated. Finally, we find our system to adapt well to generalized queries involving affordances or properties. Code is available at https://eckerlab.org/code/CLIPSeg","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a system that can generate image segmentations based on arbitrary prompts at test time, and builds upon the CLIP model as a backbone which it extends with a transformer-based decoder that enables dense prediction.'}",https://arxiv.org/pdf/2112.10003
-sega: instructing diffusion using semantic dimensions,Manuel Brack,"Text-to-image diffusion models have recently received a lot of interest for their astonishing ability to produce high-fidelity images from text only. However, achieving one-shot generation that aligns with the user's intent is nearly impossible, yet small changes to the input prompt often result in very different images. This leaves the user with little semantic control. To put the user in control, we show how to interact with the diffusion process to flexibly steer it along semantic directions. This semantic guidance (SEGA) allows for subtle and extensive edits, changes in composition and style, as well as optimizing the overall artistic conception. We demonstrate SEGA's effectiveness on a variety of tasks and provide evidence for its versatility and flexibility.","{'model': 'tldr@v2.0.0', 'text': 'This semantic guidance (SEGA) allows for subtle and extensive edits, changes in composition and style, as well as optimizing the overall artistic conception in text-to-image diffusion models.'}",http://arxiv.org/pdf/2301.12247
-fast video shot boundary detection technique based on stochastic model,M. Al-Jarrah,"A video is composed of set of shots, where shot is defined as a sequence of consecutive frames captured by one camera without interruption. In video shot transition could be a prompt hard cut or gradual fade, dissolve, and wipe. Shot boundary detection is an essential component of video processing. These boundaries are utilized on many aspect of video processing such as video indexing, and video in demand. In this paper, the authors proposed a new shot boundary detection algorithm. The proposed algorithm detects all type of shot boundaries in a high accuracy. The algorithm is developed based on a global stochastic model for video stream. The proposed stochastic model utilizes the joined characteristic function and consequently the joined momentum to model the video stream. The proposed algorithm is implemented and tested against different types of categorized videos. The proposed algorithm detects cuts fades, dissolves, and wipes transitions. Experimental results show that the algorithm has high performance. The computed precision and recall rates validated its performance.","{'model': 'tldr@v2.0.0', 'text': 'A new shot boundary detection algorithm is proposed based on a global stochastic model based on the joined characteristic function and consequently the joined momentum to model the video stream and has high performance.'}",
-"freeseg: unified, universal and open-vocabulary image segmentation",Jie Qin,"Recently, open-vocabulary learning has emerged to accomplish segmentation for arbitrary categories of text-based descriptions, which popularizes the segmentation system to more general-purpose application scenarios. However, existing methods devote to designing specialized architectures or parameters for specific segmentation tasks. These customized design paradigms lead to fragmentation between various segmentation tasks, thus hindering the uniformity of segmentation models. Hence in this paper, we propose FreeSeg, a generic framework to accomplish Unified, Universal and Open-Vocabulary Image Segmentation. FreeSeg optimizes an all-in-one network via one-shot training and employs the same architecture and parameters to handle diverse segmentation tasks seamlessly in the inference procedure. Additionally, adaptive prompt learning facilitates the unified model to capture task-aware and category-sensitive concepts, improving model robustness in multi-task and varied scenarios. Extensive experimental results demonstrate that FreeSeg establishes new state-of-the-art results in performance and generalization on three segmentation tasks, which outperforms the best task-specific architectures by a large margin: 5.5% mIoU on semantic segmentation, 17.6% mAP on instance segmentation, 20.1% PQ on panoptic segmentation for the unseen class on COCO. Project page: https://FreeSeg.github.io.","{'model': 'tldr@v2.0.0', 'text': 'FreeSeg is proposed, a generic framework to accomplish Unified, Universal and Open-Vocabulary Image Segmentation, which optimizes an all-in-one network via one-shot training and employs the same architecture and parameters to handle diverse segmentation tasks seamlessly in the inference procedure.'}",https://arxiv.org/pdf/2303.17225
-learnable ophthalmology sam,Zhongxi Qiu,"Segmentation is vital for ophthalmology image analysis. But its various modal images hinder most of the existing segmentation algorithms applications, as they rely on training based on a large number of labels or hold weak generalization ability. Based on Segment Anything (SAM), we propose a simple but effective learnable prompt layer suitable for multiple target segmentation in ophthalmology multi-modal images, named Learnable Ophthalmology Segment Anything (SAM). The learnable prompt layer learns medical prior knowledge from each transformer layer. During training, we only train the prompt layer and task head based on a one-shot mechanism. We demonstrate the effectiveness of our thought based on four medical segmentation tasks based on nine publicly available datasets. Moreover, we only provide a new improvement thought for applying the existing fundamental CV models in the medical field. Our codes are available at \href{https://github.com/Qsingle/LearnablePromptSAM}{website}.","{'model': 'tldr@v2.0.0', 'text': 'A simple but effective learnable prompt layer suitable for multiple target segmentation in ophthalmology multi-modal images, named Learnable Ophthalmology Segment Anything (SAM), which learns medical prior knowledge from each transformer layer.'}",http://arxiv.org/pdf/2304.13425
-prompting multilingual large language models to generate code-mixed texts: the case of south east asian languages,Zheng-Xin Yong,"While code-mixing is a common linguistic practice in many parts of the world, collecting high-quality and low-cost code-mixed data remains a challenge for natural language processing (NLP) research. The recent proliferation of Large Language Models (LLMs) compels one to ask: how capable are these systems in generating code-mixed data? In this paper, we explore prompting multilingual LLMs in a zero-shot manner to generate code-mixed data for seven languages in South East Asia (SEA), namely Indonesian, Malay, Chinese, Tagalog, Vietnamese, Tamil, and Singlish. We find that publicly available multilingual instruction-tuned models such as BLOOMZ and Flan-T5-XXL are incapable of producing texts with phrases or clauses from different languages. ChatGPT exhibits inconsistent capabilities in generating code-mixed texts, wherein its performance varies depending on the prompt template and language pairing. For instance, ChatGPT generates fluent and natural Singlish texts (an English-based creole spoken in Singapore), but for English-Tamil language pair, the system mostly produces grammatically incorrect or semantically meaningless utterances. Furthermore, it may erroneously introduce languages not specified in the prompt. Based on our investigation, existing multilingual LLMs exhibit a wide range of proficiency in code-mixed data generation for SEA languages. As such, we advise against using LLMs in this context without extensive human checks.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores prompting multilingual LLMs in a zero-shot manner to generate code-mixed data for seven languages in South East Asia, and finds that publicly available multilingual instruction-tuned models are incapable of producing texts with phrases or clauses from different languages.'}",https://arxiv.org/pdf/2303.13592
-how to efficiently adapt large segmentation model(sam) to medical images,Xinrong Hu,"The emerging scale segmentation model, Segment Anything (SAM), exhibits impressive capabilities in zero-shot segmentation for natural images. However, when applied to medical images, SAM suffers from noticeable performance drop. To make SAM a real ``foundation model""for the computer vision community, it is critical to find an efficient way to customize SAM for medical image dataset. In this work, we propose to freeze SAM encoder and finetune a lightweight task-specific prediction head, as most of weights in SAM are contributed by the encoder. In addition, SAM is a promptable model, while prompt is not necessarily available in all application cases, and precise prompts for multiple class segmentation are also time-consuming. Therefore, we explore three types of prompt-free prediction heads in this work, include ViT, CNN, and linear layers. For ViT head, we remove the prompt tokens in the mask decoder of SAM, which is named AutoSAM. AutoSAM can also generate masks for different classes with one single inference after modification. To evaluate the label-efficiency of our finetuning method, we compare the results of these three prediction heads on a public medical image segmentation dataset with limited labeled data. Experiments demonstrate that finetuning SAM significantly improves its performance on medical image dataset, even with just one labeled volume. Moreover, AutoSAM and CNN prediction head also has better segmentation accuracy than training from scratch and self-supervised learning approaches when there is a shortage of annotations.","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that finetuning SAM significantly improves its performance on medical image dataset, even with just one labeled volume, and AutoSAM and CNN prediction head also has better segmentation accuracy than training from scratch and self-supervised learning approaches when there is a shortage of annotations.'}",http://arxiv.org/pdf/2306.13731
-the stable artist: steering semantics in diffusion latent space,Manuel Brack,"Large, text-conditioned generative diffusion models have recently gained a lot of attention for their impressive performance in generating high-fidelity images from text alone. However, achieving high-quality results is almost unfeasible in a one-shot fashion. On the contrary, text-guided image generation involves the user making many slight changes to inputs in order to iteratively carve out the envisioned image. However, slight changes to the input prompt often lead to entirely different images being generated, and thus the control of the artist is limited in its granularity. To provide flexibility, we present the Stable Artist, an image editing approach enabling fine-grained control of the image generation process. The main component is semantic guidance (SEGA) which steers the diffusion process along variable numbers of semantic directions. This allows for subtle edits to images, changes in composition and style, as well as optimization of the overall artistic conception. Furthermore, SEGA enables probing of latent spaces to gain insights into the representation of concepts learned by the model, even complex ones such as 'carbon emission'. We demonstrate the Stable Artist on several tasks, showcasing high-quality image editing and composition.","{'model': 'tldr@v2.0.0', 'text': 'The Stable Artist is presented, an image editing approach enabling fine-grained control of the image generation process that allows for subtle edits to images, changes in composition and style, as well as optimization of the overall artistic conception.'}",http://arxiv.org/pdf/2212.06013
-internet-augmented language models through few-shot prompting for open-domain question answering,Angeliki Lazaridou,"In this work, we aim to capitalize on the unique few-shot capabilities of large-scale language models (LSLMs) to overcome some of their challenges with respect to grounding to factual and up-to-date information. Motivated by semi-parametric language models (LMs), which ground their decisions in external retrieved evidence, we use few-shot prompting to learn to condition LMs on information returned from the web using Google Search, a broad and constantly updated knowledge source. Our approach does not involve fine-tuning or learning additional parameters, thus making it applicable to any LM, offering therefore a strong baseline. Indeed, we find that LMs conditioned on the web surpass performance of closed-book models of similar, or even larger, model sizes in open-domain question answering. Finally, we find that increasing the inference-time compute of models, achieved via using multiple retrieved evidences to generate multiple answers followed by a reranking stage that uses scores generated by the same LMs, leads to better performance and alleviates lower performance of smaller few-shot LMs. All in all, our findings suggest that it might be beneficial to slow down the race towards the biggest model and instead shift attention towards finding more effective ways to use models, including but not limited to, better prompting or increasing inference-time compute.","{'model': 'tldr@v2.0.0', 'text': 'Motivated by semi-parametric language models (LMs), few-shot prompting is used to learn to condition LMs on information returned from the web using Google Search, a broad and constantly updated knowledge source, making it applicable to any LM, offering therefore a strong baseline.'}",https://arxiv.org/pdf/2203.05115
-the unreliability of explanations in few-shot prompting for textual reasoning,Xi Ye,"Does prompting a large language model (LLM) like GPT-3 with explanations improve in-context learning? We study this question on two NLP tasks that involve reasoning over text, namely question answering and natural language inference. We test the performance of four LLMs on three textual reasoning datasets using prompts that include explanations in multiple different styles. For these tasks, we find that including explanations in the prompts for OPT, GPT-3 (davinci), and InstructGPT (text-davinci-001) only yields small to moderate accuracy improvements over standard few-show learning. However, text-davinci-002 is able to benefit more substantially. We further show that explanations generated by the LLMs may not entail the models' predictions nor be factually grounded in the input, even on simple tasks with extractive explanations. However, these flawed explanations can still be useful as a way to verify LLMs' predictions post-hoc. Through analysis in our three settings, we show that explanations judged by humans to be good--logically consistent with the input and the prediction--more likely cooccur with accurate predictions. Following these observations, we train calibrators using automatically extracted scores that assess the reliability of explanations, allowing us to improve performance post-hoc across all of our datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work studies two NLP tasks that involve reasoning over text, namely question answering and natural language inference, and shows that explanations judged by humans to be good--logically consistent with the input and the prediction--more likely cooccur with accurate predictions.'}",
-few-shot prompting towards controllable response generation,Hsuan Su,"Much literature has shown that prompt-based learning is an efficient method to make use of the large pre-trained language model. Recent works also exhibit the possibility of steering a chatbot’s output by plugging in an ap-propriate prompt. Gradient-based methods are often used to perturb the prompts. However, some language models are not even available to the public. In this work, we first explored the combination of prompting and reinforcement learning (RL) to steer models’ generation without accessing any of the models’ parameters. Second, to reduce the training effort and enhance the generalizability to the unseen task, we apply multi-task learning to make the model learn to generalize to new tasks better. The experiment results show that our proposed method can successfully control several state-of-the-art (SOTA) dialogue models without accessing their parameters. Furthermore, the model demonstrates the strong ability to quickly adapt to an unseen task in fewer steps than the baseline model.","{'model': 'tldr@v2.0.0', 'text': 'The experiment results show that the proposed method can successfully control several state-of-the-art (SOTA) dialogue models without accessing their parameters and demonstrates the strong ability to quickly adapt to an unseen task in fewer steps than the baseline model.'}",http://arxiv.org/pdf/2206.03931
-zero- and few-shot prompting with llms: a comparative study with fine-tuned models for bangla sentiment analysis,Md. Arid Hasan,"The rapid expansion of the digital world has propelled sentiment analysis into a critical tool across diverse sectors such as marketing, politics, customer service, and healthcare. While there have been significant advancements in sentiment analysis for widely spoken languages, low-resource languages, such as Bangla, remain largely under-researched due to resource constraints. Furthermore, the recent unprecedented performance of Large Language Models (LLMs) in various applications highlights the need to evaluate them in the context of low-resource languages. In this study, we present a sizeable manually annotated dataset encompassing 33,605 Bangla news tweets and Facebook comments. We also investigate zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine-tuned models. Our findings suggest that monolingual transformer-based models consistently outperform other models, even in zero and few-shot scenarios. To foster continued exploration, we intend to make this dataset and our research tools publicly available to the broader research community. In the spirit of further research, we plan to make this dataset and our experimental resources publicly accessible to the wider research community.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a sizeable manually annotated dataset encompassing 33,605 Bangla news tweets and Facebook comments and investigates zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine- tuned models.'}",https://arxiv.org/pdf/2308.10783
-multilingual social media text generation and evaluation with few-shot prompting,Mack Blackburn,"This work adapts large language models to generate multilingual social media text that meets several objectives simultaneously: topic relevance, author style consistency, and reply validity. Leveraging existing online information behavior simulators, which currently only forecast activities but not content, our approach comprised of generalizable prompt formation and efficient evaluation to produce a believable, personalized, and responsive synthetic social network. According to some preliminary experiments, our multi-objective prompt formation and automatic evaluation/selection methods are able to yield a significant number of high-quality synthetic texts according to both standardized and trained metrics.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2022.gem-1.39.pdf
-sparsefit: few-shot prompting with sparse fine-tuning for jointly generating predictions and natural language explanations,Jesus Solano,"Explaining the decisions of neural models is crucial for ensuring their trustworthiness at deployment time. Using Natural Language Explanations (NLEs) to justify a model's predictions has recently gained increasing interest. However, this approach usually demands large datasets of human-written NLEs for the ground-truth answers, which are expensive and potentially infeasible for some applications. For models to generate high-quality NLEs when only a few NLEs are available, the fine-tuning of Pre-trained Language Models (PLMs) in conjunction with prompt-based learning recently emerged. However, PLMs typically have billions of parameters, making fine-tuning expensive. We propose SparseFit, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs. We experiment with SparseFit on the T5 model and four datasets and compare it against state-of-the-art parameter-efficient fine-tuning techniques. We perform automatic and human evaluations to assess the quality of the model-generated NLEs, finding that fine-tuning only 6.8% of the model parameters leads to competitive results for both the task performance and the quality of the NLEs.","{'model': 'tldr@v2.0.0', 'text': 'SarseFit is proposed, a sparse few-shot fine-tuning strategy that leverages discrete prompts to jointly generate predictions and NLEs and is compared against state-of-the-art parameter-efficient fine- Tuning techniques.'}",http://arxiv.org/pdf/2305.13235
-style-aware radiology report generation with radgraph and few-shot prompting,Benjamin Yan,"Automatically generated reports from medical images promise to improve the workflow of radiologists. Existing methods consider an image-to-report modeling task by directly generating a fully-fledged report from an image. However, this conflates the content of the report (e.g., findings and their attributes) with its style (e.g., format and choice of words), which can lead to clinically inaccurate reports. To address this, we propose a two-step approach for radiology report generation. First, we extract the content from an image; then, we verbalize the extracted content into a report that matches the style of a specific radiologist. For this, we leverage RadGraph -- a graph representation of reports -- together with large language models (LLMs). In our quantitative evaluations, we find that our approach leads to beneficial performance. Our human evaluation with clinical raters highlights that the AI-generated reports are indistinguishably tailored to the style of individual radiologist despite leveraging only a few examples as context.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-step approach for radiology report generation that leverages RadGraph -- a graph representation of reports -- together with large language models (LLMs) and finds that this approach leads to beneficial performance.'}",
-prompting electra: few-shot learning with discriminative pre-trained models,Mengzhou Xia,"Pre-trained masked language models successfully perform few-shot learning by formulating downstream tasks as text infilling. How- ever, as a strong alternative in full-shot settings, discriminative pre-trained models like ELECTRA do not fit into the paradigm. In this work, we adapt prompt-based few-shot learning to ELECTRA and show that it outperforms masked language models in a wide range of tasks. ELECTRA is pre-trained to distinguish if a token is generated or original. We naturally extend that to prompt-based few-shot learning by training to score the originality of the target options without introducing new parameters. Our method can be easily adapted to tasks involving multi-token predictions without extra computation overhead. Analysis shows that ELECTRA learns distributions that align better with downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work adapt prompt-based few-shot learning to ELECTRA and shows that it outperforms masked language models in a wide range of tasks and learns distributions that align better with downstream tasks.'}",https://arxiv.org/pdf/2205.15223
-unsupervisedly prompting alphafold2 for few-shot learning of accurate folding landscape and protein structure prediction,Jun Zhang,"Data-driven predictive methods which can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining accurate folding landscape using co-evolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit co-evolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologs. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences, but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method which could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.","{'model': 'tldr@v2.0.0', 'text': 'EvoGen, a meta generative model, is presented to remedy the underperformance of AlphaFold2 for poor MSA targets, and yields a probabilistic structure generation method which could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.'}",
-knowledge prompting for few-shot action recognition,Yuheng Shi,"Few-shot action recognition in videos is challenging for its lack of supervision and difficulty in generalizing to unseen actions. To address this task, we propose a simple yet effec- tive method, called knowledge prompting, which leverages commonsense knowledge of actions from external resources to prompt a powerful pre-trained vision-language model for few-shot classification. We first collect large-scale language descriptions of actions, defined as text proposals, to build an action knowledge base. The collection of text proposals is done by filling in handcraft sentence templates with external action-related corpus or by extracting action-related phrases from captions of Web instruction videos. Then we feed these text proposals into the pre-trained vision-language model along with video frames to generate matching scores of the proposals to each frame, and the scores can be treated as action semantics with strong generalization. Finally, we design a lightweight temporal modeling network to capture the temporal evolution of action semantics for classification. Extensive experiments on six benchmark datasets demonstrate that our method generally achieves the state-of-the-art performance while reducing the training overhead to 1‰ of existing methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple yet effective method, called knowledge prompting, which leverages commonsense knowledge of actions from external resources to prompt a powerful pre-trained vision-language model for few-shot classification.'}",https://arxiv.org/pdf/2211.12030
-promptner: a prompting method for few-shot named entity recognition via k nearest neighbor search,Mozhi Zhang,"Few-shot Named Entity Recognition (NER) is a task aiming to identify named entities via limited annotated samples. Recently, prototypical networks have shown promising performance in few-shot NER. Most of prototypical networks will utilize the entities from the support set to construct label prototypes and use the query set to compute span-level similarities and optimize these label prototype representations. However, these methods are usually unsuitable for fine-tuning in the target domain, where only the support set is available. In this paper, we propose PromptNER: a novel prompting method for few-shot NER via k nearest neighbor search. We use prompts that contains entity category information to construct label prototypes, which enables our model to fine-tune with only the support set. Our approach achieves excellent transfer learning ability, and extensive experiments on the Few-NERD and CrossNER datasets demonstrate that our model achieves superior performance over state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.12217
-prompting large language models with chain-of-thought for few-shot knowledge base question generation,Yuanyuan Liang,"The task of Question Generation over Knowledge Bases (KBQG) aims to convert a logical form into a natural language question. For the sake of expensive cost of large-scale question annotation, the methods of KBQG under low-resource scenarios urgently need to be developed. However, current methods heavily rely on annotated data for fine-tuning, which is not well-suited for few-shot question generation. The emergence of Large Language Models (LLMs) has shown their impressive generalization ability in few-shot tasks. Inspired by Chain-of-Thought (CoT) prompting, which is an in-context learning strategy for reasoning, we formulate KBQG task as a reasoning problem, where the generation of a complete question is splitted into a series of sub-question generation. Our proposed prompting method KQG-CoT first retrieves supportive logical forms from the unlabeled data pool taking account of the characteristics of the logical form. Then, we write a prompt to explicit the reasoning chain of generating complicated questions based on the selected demonstrations. To further ensure prompt quality, we extend KQG-CoT into KQG-CoT+ via sorting the logical forms by their complexity. We conduct extensive experiments over three public KBQG datasets. The results demonstrate that our prompting method consistently outperforms other prompting baselines on the evaluated datasets. Remarkably, our KQG-CoT+ method could surpass existing few-shot SoTA results of the PathQuestions dataset by 18.25, 10.72, and 10.18 absolute points on BLEU-4, METEOR, and ROUGE-L, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by Chain-of-Thought prompting, the proposed prompting method KQG-CoT first retrieves supportive logical forms from the unlabeled data pool taking account of the characteristics of the logical form, and could surpass existing few-shot SoTA results of the PathQuestions dataset.'}",https://arxiv.org/pdf/2310.08395
-investigating prompting techniques for zero- and few-shot visual question answering,Rabiul Awal,"Visual question answering (VQA) is a challenging task that requires the ability to comprehend and reason with visual information. While recent vision-language models have made strides, they continue to struggle with zero-shot VQA, particularly in handling complex compositional questions and adapting to new domains i.e. knowledge-based reasoning. This paper explores the use of various prompting strategies, focusing on the BLIP2 model, to enhance zero-shot VQA performance. We conduct a comprehensive investigation across several VQA datasets, examining the effectiveness of different question templates, the role of few-shot exemplars, the impact of chain-of-thought (CoT) reasoning, and the benefits of incorporating image captions as additional visual cues. Despite the varied outcomes, our findings demonstrate that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples. However, we also identify a limitation in the use of chain-of-thought rationalization, which negatively affects VQA accuracy. Our study thus provides critical insights into the potential of prompting for improving zero-shot VQA performance.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that carefully designed question templates and the integration of additional visual cues, like image captions, can contribute to improved VQA performance, especially when used in conjunction with few-shot examples, however, a limitation in the use of chain-of-thought rationalization is identified, which negatively affects V QA accuracy.'}",http://arxiv.org/pdf/2306.09996
-prod: prompting-to-disentangle domain knowledge for cross-domain few-shot image classification,Tianyi Ma,"This paper considers few-shot image classification under the cross-domain scenario, where the train-to-test domain gap compromises classification accuracy. To mitigate the domain gap, we propose a prompting-to-disentangle (ProD) method through a novel exploration with the prompting mechanism. ProD adopts the popular multi-domain training scheme and extracts the backbone feature with a standard Convolutional Neural Network. Based on these two common practices, the key point of ProD is using the prompting mechanism in the transformer to disentangle the domain-general (DG) and domain-specific (DS) knowledge from the backbone feature. Specifically, ProD concatenates a DG and a DS prompt to the backbone feature and feeds them into a lightweight transformer. The DG prompt is learnable and shared by all the training domains, while the DS prompt is generated from the domain-of-interest on the fly. As a result, the transformer outputs DG and DS features in parallel with the two prompts, yielding the disentangling effect. We show that: 1) Simply sharing a single DG prompt for all the training domains already improves generalization towards the novel test domain. 2) The cross-domain generalization can be further reinforced by making the DG prompt neutral towards the training domains. 3) When inference, the DS prompt is generated from the support samples and can capture test domain knowledge through the prompting mechanism. Combining all three benefits, ProD significantly improves cross-domain few-shot classification. For instance, on CUB, ProD improves the 5-way 5-shot ac-curacy from 73.56% (baseline) to 79.19%, setting a new state of the art.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that simply sharing a single DG prompt for all the training domains already improves generalization towards the novel test domain, and the cross-domain generalization can be further reinforced by making the DG prompt neutral towards theTraining domains.'}",
-template-free prompting for few-shot named entity recognition via semantic-enhanced contrastive learning.,Kai He,"Prompt tuning has achieved great success in various sentence-level classification tasks by using elaborated label word mappings and prompt templates. However, for solving token-level classification tasks, e.g., named entity recognition (NER), previous research, which utilizes N-gram traversal for prompting all spans with all possible entity types, is time-consuming. To this end, we propose a novel prompt-based contrastive learning method for few-shot NER without template construction and label word mappings. First, we leverage external knowledge to initialize semantic anchors for each entity type. These anchors are simply appended with input sentence embeddings as template-free prompts (TFPs). Then, the prompts and sentence embeddings are in-context optimized with our proposed semantic-enhanced contrastive loss. Our proposed loss function enables contrastive learning in few-shot scenarios without requiring a significant number of negative samples. Moreover, it effectively addresses the issue of conventional contrastive learning, where negative instances with similar semantics are erroneously pushed apart in natural language processing (NLP)-related tasks. We examine our method in label extension (LE), domain-adaption (DA), and low-resource generalization evaluation tasks with six public datasets and different settings, achieving state-of-the-art (SOTA) results in most cases.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel prompt-based contrastive learning method for few-shot NER without template construction and label word mappings, and effectively addresses the issue of conventional Contrastive learning, where negative instances with similar semantics are erroneously pushed apart in natural language processing (NLP)-related tasks.'}",
-lmcap: few-shot multilingual image captioning by retrieval augmented language model prompting,R. Ramos,"Multilingual image captioning has recently been tackled by training with large-scale machine translated data, which is an expensive, noisy, and time-consuming process. Without requiring any multilingual caption data, we propose LMCap, an image-blind few-shot multilingual captioning model that works by prompting a language model with retrieved captions. Specifically, instead of following the standard encoder-decoder paradigm, given an image, LMCap first retrieves the captions of similar images using a multilingual CLIP encoder. These captions are then combined into a prompt for an XGLM decoder, in order to generate captions in the desired language. In other words, the generation model does not directly process the image, instead processing retrieved captions. Experiments on the XM3600 dataset of geographically diverse images show that our model is competitive with fully-supervised multilingual captioning models, without requiring any supervised training on any captioning data.","{'model': 'tldr@v2.0.0', 'text': 'LMCap is proposed, an image-blind few-shot multilingual captioning model that works by prompting a language model with retrieved captions, which is competitive with fully-supervised mult bilingual captioning models, without requiring any supervised training on any captioning data.'}",http://arxiv.org/pdf/2305.19821
-hiprompt: few-shot biomedical knowledge fusion via hierarchy-oriented prompting,Jiaying Lu,"Medical decision-making processes can be enhanced by comprehensive biomedical knowledge bases, which require fusing knowledge graphs constructed from different sources via a uniform index system. The index system often organizes biomedical terms in a hierarchy to provide the aligned entities with fine-grained granularity. To address the challenge of scarce supervision in the biomedical knowledge fusion (BKF) task, researchers have proposed various unsupervised methods. However, these methods heavily rely on ad-hoc lexical and structural matching algorithms, which fail to capture the rich semantics conveyed by biomedical entities and terms. Recently, neural embedding models have proved effective in semantic-rich tasks, but they rely on sufficient labeled data to be adequately trained. To bridge the gap between the scarce-labeled BKF and neural embedding models, we propose HiPrompt, a supervision-efficient knowledge fusion framework that elicits the few-shot reasoning ability of large language models through hierarchy-oriented prompts. Empirical results on the collected KG-Hi-BKF benchmark datasets demonstrate the effectiveness of HiPrompt.","{'model': 'tldr@v2.0.0', 'text': 'HiPrompt is proposed, a supervision-efficient knowledge fusion framework that elicits the few-shot reasoning ability of large language models through hierarchy-oriented prompts and demonstrates the effectiveness of HiPrompt on the collected KG-Hi-BKF benchmark datasets.'}",https://arxiv.org/pdf/2304.05973
-few-shot learning with prompting methods,Morteza Bahrami,"Today, in natural language processing, labeled data is important, however, getting adequate amount of data is a challenging step. There are many tasks for which it is difficult to obtain the required training data. For example, in machine translation, we need to prepare a lot of data in the target language, so that the work performance is acceptable. We may not be able to collect useful data in the target language. Hence, we need to use few-shot learning. Recently, a method called prompting has evolved, in which text inputs are converted into text with a new structure using a certain format, which has a blank space. Given the prompted text, a pre-trained language model replaces the space with the best word. Prompting can help us in the field of few-shot learning; even in cases where there is no data, i.e. zero-shot learning. Recent works use large language models such as GPT-2 and GPT-3, with the prompting method, performed tasks such as machine translation. These efforts do not use any labeled training data. But these types of models with a massive number of parameters require powerful hardware. Pattern-Exploiting Training (PET) and iterative Pattern-Exploiting Training (iPET) were introduced, which perform few-shot learning using prompting and smaller pre-trained language models such as Bert and Roberta. For example, for the Yahoo text classification dataset, using iPET and Roberta and ten labeled datasets, 70% accuracy has been reached. This paper reviews research works in few-shot learning with a new paradigm in natural language processing, which we dub prompt-based learning or in short, prompting.","{'model': 'tldr@v2.0.0', 'text': 'Research works in few-shot learning with a new paradigm in natural language processing, which is dubbed prompt-based learning or in short, prompting are reviewed.'}",
-adversarial knowledge stimulated contrastive prompting for few-shot language learners,Kai Zheng,"Prompt-based fine-tuning has boosted the performance of Pre-trained Language Models (PLMs) on few-shot Natural Language Understanding (NLU) tasks by employing task-specific prompts. Yet, PLMs are unfamiliar with prompt-style expressions during pre-training, which limits the few-shot learning performance on downstream tasks. It would be desirable if the models can stimulate prompting knowledge while adaptation to specific NLU tasks. We present the Adversarial Knowledge Stimulated Contrastive Prompting (AKSCP) framework, leading to better few-shot NLU tasks for language models by implicitly stimulate knowledge from pretrained language model. In AKSCP, a novel paradigm Cloze-driven prompt is proposed for joint prompt tuning across word cloze task and prompt-based learning, forcing PLMs to stimulate prompting knowledge. We further design an Adversarial Contrastive learning method to improve the generalization ability of PLM for different downstream tasks. Experiments over a variety of NLU tasks show that AKSCP consistently outperforms state-of-the-arts for prompt-based fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'In AKSCP, a novel paradigm Cloze-driven prompt is proposed for joint prompt tuning across word cloze task and prompt-based learning, forcing PLMs to stimulate prompting knowledge, and an Adversarial Contrastive learning method is designed to improve the generalization ability of PLM for different downstream tasks.'}",https://aclanthology.org/2023.findings-acl.852.pdf
-multi-step prompting for few-shot emotion-grounded conversations,Mauzama Firdaus,"Conversational systems have shown immense growth in their ability to communicate like humans. With the emergence of large pre-trained language models (PLMs) the ability to provide informative responses have improved significantly. Despite the success of PLMs, the ability to identify and generate engaging and empathetic responses is largely dependent on labelled-data. In this work, we design a prompting approach that identifies the emotion of a given utterance and uses the emotion information for generating the appropriate responses for conversational systems. We propose a two-step prompting method that first recognises the emotion in the dialogue utterance and in the second-step uses the predicted emotion to prompt the PLM to generate the corresponding em- pathetic response in a few-shot setting. Experimental results on three publicly available datasets show that our proposed approach outperforms the state-of-the-art approaches for both automatic and manual evaluation.","{'model': 'tldr@v2.0.0', 'text': 'A two-step prompting method that first recognises the emotion in the dialogue utterance and in the second-step uses the predicted emotion to prompt the PLM to generate the corresponding em- pathetic response in a few-shot setting is proposed.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615265
-leveraging few-shot data augmentation and waterfall prompting for response generation,Lea Krause,"This paper discusses our approaches for task-oriented conversational modelling using subjective knowledge, with a particular emphasis on response generation. Our methodology was shaped by an extensive data analysis that evaluated key factors such as response length, sentiment, and dialogue acts present in the provided dataset. We used few-shot learning to augment the data with newly generated subjective knowledge items and present three approaches for DSTC11: (1) task-specific model exploration, (2) incorporation of the most frequent question into all generated responses, and (3) a waterfall prompting technique using a combination of both GPT-3 and ChatGPT.","{'model': 'tldr@v2.0.0', 'text': 'This paper discusses the approaches for task-oriented conversational modelling using subjective knowledge, with a particular emphasis on response generation, and presents three approaches for DSTC11: task-specific model exploration, incorporation of the most frequent question into all generated responses, and a waterfall prompting technique using a combination of both GPT-3 and ChatGPT.'}",https://arxiv.org/pdf/2308.01080
-unsupervisedly prompting alphafold2 for accurate few-shot protein structure prediction.,Jun Zhang,"Data-driven predictive methods that can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining an accurate folding landscape using coevolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit coevolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologues. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in the low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.","{'model': 'tldr@v2.0.0', 'text': 'EvoGen, a meta generative model, is presented to remedy the underperformance of AlphaFold2 for poor MSA targets and yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.'}",
-the unreliability of explanations in few-shot in-context learning,Xi Ye,"How can prompting a large language model like GPT-3 with explanations improve in-context learning? We focus specifically on two NLP tasks that involve reasoning over text, namely question answering and natural language inference. Including explanations in the prompt and having the model generate them does not consistently improve performance in the settings we study, contrary to recent results on symbolic reasoning tasks (Nye et al., 2021; Wei et al., 2022). Despite careful prompting, explanations generated by GPT-3 may not even be factually grounded in the input, even on simple tasks with straightforward extractive explanations. However, these flawed explanations can still be useful as a way to verify GPT-3’s predictions post-hoc. Through analysis in three settings, we show that explanations judged as good by humans—those that are logically consistent with the input and the prediction—usually indicate more accurate predictions. Following these observations, we present a framework for calibrating model predictions based on the reliability of the explanations. Our framework trains calibrators using automatically extracted scores that approximately assess the reliability of explanations, which helps improve performance across three different datasets","{'model': 'tldr@v2.0.0', 'text': 'A framework for calibrating model predictions based on the reliability of explanations is presented, showing that explanations judged as good by humans—those that are logically consistent with the input and the prediction—usually indicate more accurate predictions.'}",http://arxiv.org/pdf/2205.03401
-self-convinced prompting: few-shot question answering with repeated introspection,Haodi Zhang,"While large language models (LLMs) such as ChatGPT and PaLM have demonstrated remarkable performance in various language understanding and generation tasks, their capabilities in complex reasoning and intricate knowledge utilization still fall short of human-level proficiency. Recent studies have established the effectiveness of prompts in steering LLMs towards generating desired outputs. Building on these insights, we introduce a novel framework that harnesses the potential of large-scale pre-trained language models, to iteratively enhance performance of the LLMs. Our framework incorporates three components: \textit{Normal CoT}, a \textit{Convincer}, and an \textit{Answerer}. It processes the output of a typical few-shot chain-of-thought prompt, assesses the correctness of the response, scrutinizes the answer, refines the reasoning, and ultimately produces a new solution. Experimental results on the 7 datasets of miscellaneous problems validate the efficacy of the Self-Convince framework, achieving substantial improvements compared to the baselines. This study contributes to the burgeoning body of research focused on integrating pre-trained language models with tailored prompts and iterative refinement processes to augment their performance in complex tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework that harnesses the potential of large-scale pre-trained language models, to iteratively enhance performance of the LLMs, and validate the efficacy of the Self-Convince framework on the 7 datasets of miscellaneous problems.'}",https://arxiv.org/pdf/2310.05035
-"focusing, bridging and prompting for few-shot nested named entity recognition",Yuanyuan Xu,,,https://aclanthology.org/2023.findings-acl.164.pdf
-retrieving visual facts for few-shot visual question answering,,"We introduce the Retrieving Visual Facts 001 (RVF) framework for few-shot visual question 002 answering (VQA). The RVF framework rep- 003 resents an image as a set of natural language 004 facts; for example, in practice these could be 005 tags from an object detector. Critically, the 006 question is used to retrieve relevant facts: an 007 image may contain numerous details, and one 008 should attend to the few which may be use- 009 ful for the question. Finally, one predicts the 010 answer from the retrieved facts and the ques- 011 tion, e.g., by prompting a language model as 012 we do here. Compared to P ICA (Yang et al., 013 2021), the previous state-of-the-art in few-shot 014 VQA, a proof-of-concept RVF implementa- 015 tion improves absolute performance by 2.6% 016 and 1.5% respectively on the VQAv2 (Goyal 017 et al., 2017) and OK-VQA (Marino et al., 2019) 018 datasets. We also analyze our implementation’s 019 strengths and weaknesses on various question 020 types, highlighting directions for further study. 021","{'model': 'tldr@v2.0.0', 'text': 'A proof-of-concept RVF implementation improves absolute performance by 2.6% and 1.5% respectively on the VQAv2 and OK-VQA datasets, and analyzes the implementation’s 019 strengths and weaknesses on various question 020 types, highlighting directions for further study.'}",
-continued pretraining for better zero- and few-shot promptability,Zhaofeng Wu,"Recently introduced language model prompting methods can achieve high accuracy in zero- and few-shot settings while requiring few to no learned task-specific parameters. Nevertheless, these methods still often trail behind full model finetuning. In this work, we investigate if a dedicated continued pretraining stage could improve “promptability”, i.e., zero-shot performance with natural language prompts or few-shot performance with prompt tuning. We reveal settings where existing continued pretraining methods lack promptability. We also identify current methodological gaps, which we fill with thorough large-scale experiments. We demonstrate that a simple recipe, continued pretraining that incorporates a trainable prompt during multi-task learning, leads to improved promptability in both zero- and few-shot settings compared to existing methods, up to 31% relative. On the other hand, we find that continued pretraining using MAML-style meta-learning, a method that directly optimizes few-shot promptability, yields subpar performance. We validate our findings with two prompt tuning methods, and, based on our results, we provide concrete recommendations to optimize promptability for different use cases.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that a simple recipe, continued pretraining that incorporates a trainable prompt during multi-task learning, leads to improved promptability in both zero- and few-shot settings compared to existing methods, up to 31% relative.'}",http://arxiv.org/pdf/2210.10258
-what makes pre-trained language models better zero/few-shot learners?,Jinghui Lu,"In this paper, we propose a theoretical framework to explain the efficacy of prompt learning in zero/few-shot scenarios. First, we prove that conventional pre-training and fine-tuning paradigm fails in few-shot scenarios due to overfitting the unrepresentative labelled data. We then detail the assumption that prompt learning is more effective because it empowers pre-trained language model that is built upon massive text corpora, as well as domain-related human knowledge to participate more in prediction and thereby reduces the impact of limited label information provided by the small training set. We further hypothesize that language discrepancy can measure the qual- ity of prompting. Comprehensive experiments are performed to verify our assumptions. More remarkably, inspired by the theoretical framework, we propose an annotation-agnostic template selection method based on perplexity, which enables us to “forecast” the prompting performance in advance. This approach is especially encouraging because existing work still relies on development set to post-hoc evaluate templates. Experiments show that this method leads to significant prediction benefits compared to state-of-the-art zero-shot methods.","{'model': 'tldr@v2.0.0', 'text': 'A theoretical framework to explain theacy of prompt learning in zero/few-shot scenarios is proposed and an annotation-agnostic template selection method based on perplexity is proposed, which enables us to “forecast” the prompting performance in advance.'}",http://arxiv.org/pdf/2209.15206
-plan-and-solve prompting: improving zero-shot chain-of-thought reasoning by large language models,Lei Wang,"Large language models (LLMs) have recently been shown to deliver impressive performance in various NLP tasks. To tackle multi-step reasoning tasks, Few-shot chain-of-thought (CoT) prompting includes a few manually crafted step-by-step reasoning demonstrations which enable LLMs to explicitly generate reasoning steps and improve their reasoning task accuracy. To eliminate the manual efforts, Zero-shot-CoT concatenates the target problem statement with “Let’s think step by step” as an input prompt to LLMs. Despite the success of Zero-shot-CoT, it still suffers from three pitfalls: calculation errors, missing-step errors, and semantic misunderstanding errors. To address the missing-step errors, we propose Plan-and-Solve (PS) Prompting. It consists of two components: first, devising a plan to divide the entire task into smaller subtasks, and then carrying out the subtasks according to the plan. To address the calculation errors and improve the quality of generated reasoning steps, we extend PS prompting with more detailed instructions and derive PS+ prompting. We evaluate our proposed prompting strategy on ten datasets across three reasoning problems. The experimental results over GPT-3 show that our proposed zero-shot prompting consistently outperforms Zero-shot-CoT across all datasets by a large margin, is comparable to or exceeds Zero-shot-Program-of-Thought Prompting, and has comparable performance with 8-shot CoT prompting on the math reasoning problem. The code can be found at https://github.com/AGI-Edgerunners/Plan-and-Solve-Prompting.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results over GPT-3 show that the proposed zero-shot prompting consistently outperforms Zero- shot-CoT across all datasets by a large margin, is comparable to or exceeds Zero-shot-Program-of-Thought Prompting, and has comparable performance with 8-shot CoT prompting on the math reasoning problem.'}",http://arxiv.org/pdf/2305.04091
-better zero-shot reasoning with self-adaptive prompting,Xingchen Wan,"Modern large language models (LLMs) have demonstrated impressive capabilities at sophisticated tasks, often through step-by-step reasoning similar to humans. This is made possible by their strong few and zero-shot abilities -- they can effectively learn from a handful of handcrafted, completed responses (""in-context examples""), or are prompted to reason spontaneously through specially designed triggers. Nonetheless, some limitations have been observed. First, performance in the few-shot setting is sensitive to the choice of examples, whose design requires significant human effort. Moreover, given the diverse downstream tasks of LLMs, it may be difficult or laborious to handcraft per-task labels. Second, while the zero-shot setting does not require handcrafting, its performance is limited due to the lack of guidance to the LLMs. To address these limitations, we propose Consistency-based Self-adaptive Prompting (COSP), a novel prompt design method for LLMs. Requiring neither handcrafted responses nor ground-truth labels, COSP selects and builds the set of examples from the LLM zero-shot outputs via carefully designed criteria that combine consistency, diversity and repetition. In the zero-shot setting for three different LLMs, we show that using only LLM predictions, COSP improves performance up to 15% compared to zero-shot baselines and matches or exceeds few-shot baselines for a range of reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Consistency-based Self-adaptive Prompting (COSP), a novel prompt design method for LLMs that selects and builds the set of examples from the LLM zero-shot outputs via carefully designed criteria that combine consistency, diversity and repetition.'}",http://arxiv.org/pdf/2305.14106
-gpt-3 for few-shot dialogue state tracking,Nicholas Pezzotti,"GPT-3 (Brown et al., 2020) has attracted considerable attention due to its superior performance across a wide range of Natural Language Processing (NLP) tasks, especially with its powerful and versatile in-context few-shot learning ability. That is, it has been shown that by carefully crafting a prompt, consisting of a few labelled examples followed by an unlabelled example, GPT’3 is able to do few-shot sentiment classification, three-digit arithmetic and much more. We seek to evaluate its performance on a novel and notably more complicated task: few-shot Dialogue State Tracking (DST). We propose a few-shot prompting framework that selects in-context examples based on similarity which outperforms the original random in-context selection framework. We also review and formalise the two types of completion strategies employed by previous literature, which we name constrained and unconstrained, and propose a third ""semi-constrained"" completion strategy, which is particularly well adapted for DST. Additionally, we propose a prompt ensembling technique that reliably outperforms individual models. Furthermore, we are the first, to the best of our knowledge, to fine-tune GPT-3 for the task of few-shot DST, showing that it reliably outperforms its GPT-2 counterpart. Furthermore, we seek to synthesise and formalise the largely heterogeneous body of previous work on prompt programming and in-context learning for GPT-3. In an attempt to contribute to the understanding of the strengths, weaknesses and inner-working of GPT-3, we perform numerous ablative studies that validate and confute previous in-context learning empirical findings: mainly, we find that natural language instructions in the prompt have little impact on performance, larger language models do not always induce higher downstream performance and that GPT-3 is highly sensitive to the order and number of the in-context examples.","{'model': 'tldr@v2.0.0', 'text': 'It is found that natural language instructions in the prompt have little impact on performance, larger language models do not always induce higher downstream performance and that GPT-3 is highly sensitive to the order and number of the in-context examples.'}",
-multi-modal prompting for low-shot temporal action localization,Chen Ju,"In this paper, we consider the problem of temporal action localization under low-shot (zero-shot&few-shot) scenario, with the goal of detecting and classifying the action instances from arbitrary categories within some untrimmed videos, even not seen at training time. We adopt a Transformer-based two-stage action localization architecture with class-agnostic action proposal, followed by open-vocabulary classification. We make the following contributions. First, to compensate image-text foundation models with temporal motions, we improve category-agnostic action proposal by explicitly aligning embeddings of optical flows, RGB and texts, which has largely been ignored in existing low-shot methods. Second, to improve open-vocabulary action classification, we construct classifiers with strong discriminative power, i.e., avoid lexical ambiguities. To be specific, we propose to prompt the pre-trained CLIP text encoder either with detailed action descriptions (acquired from large-scale language models), or visually-conditioned instance-specific prompt vectors. Third, we conduct thorough experiments and ablation studies on THUMOS14 and ActivityNet1.3, demonstrating the superior performance of our proposed model, outperforming existing state-of-the-art approaches by one significant margin.","{'model': 'tldr@v2.0.0', 'text': 'This paper adopts a Transformer-based two-stage action localization architecture with class-agnostic action proposal, followed by open-vocabulary classification, and proposes to prompt the pre-trained CLIP text encoder either with detailed action descriptions, or visually-conditioned instance-specific prompt vectors.'}",http://arxiv.org/pdf/2303.11732
-program of thoughts prompting: disentangling computation from reasoning for numerical reasoning tasks,Wenhu Chen,"Recently, there has been significant progress in teaching language models to perform step-by-step reasoning to solve complex numerical reasoning tasks. Chain-of-thoughts prompting (CoT) is by far the state-of-art method for these tasks. CoT uses language models to perform both reasoning and computation in the multi-step `thought' process. To disentangle computation from reasoning, we propose `Program of Thoughts' (PoT), which uses language models (mainly Codex) to express the reasoning process as a program. The computation is relegated to an external computer, which executes the generated programs to derive the answer. We evaluate PoT on five math word problem datasets (GSM, AQuA, SVAMP, TabMWP, MultiArith) and three financial-QA datasets (FinQA, ConvFinQA, TATQA) for both few-shot and zero-shot setups. Under both few-shot and zero-shot settings, PoT can show an average performance gain over CoT by around 12\% across all the evaluated datasets. By combining PoT with self-consistency decoding, we can achieve SoTA performance on all math problem datasets and near-SoTA performance on financial datasets. All of our data and code are released in Github https://github.com/wenhuchen/Program-of-Thoughts","{'model': 'tldr@v2.0.0', 'text': 'Under both few-shot and zero-shot settings, PoT can show an average performance gain over CoT by around 12\\% across all the evaluated datasets, and by combining PoT with self-consistency decoding, can achieve SoTA performance on all math problem datasets and near-SoTAperformance on financial datasets.'}",http://arxiv.org/pdf/2211.12588
-generative zero-shot prompt learning for cross-domain slot filling with inverse prompting,Xuefeng Li,"Zero-shot cross-domain slot filling aims to transfer knowledge from the labeled source domain to the unlabeled target domain. Existing models either encode slot descriptions and examples or design handcrafted question templates using heuristic rules, suffering from poor generalization capability or robustness. In this paper, we propose a generative zero-shot prompt learning framework for cross-domain slot filling, both improving generalization and robustness than previous work. Besides, we introduce a novel inverse prompting strategy to distinguish different slot types to avoid the multiple prediction problem, and an efficient prompt-tuning strategy to boost higher performance by only training fewer prompt parameters. Experiments and analysis demonstrate the effectiveness of our proposed framework, especially huge improvements (+13.44% F1) on the unseen slots.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a generative zero-shot prompt learning framework for cross-domain slot filling, both improving generalization and robustness than previous work, and introduces a novel inverse prompting strategy to distinguish different slot types to avoid the multiple prediction problem.'}",https://arxiv.org/pdf/2307.02830
-language-aware soft prompting: text-to-text optimization for few- and zero-shot adaptation of v &l models,Adrian Bulat,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Language-Aware Soft Prompting (LASP) learning method by means of a text-to-text cross-entropy loss that maximizes the probability of the learned prompts to be correctly classified with respect to pre-defined hand-crafted textual prompts and presents a novel zero-shot variant of LASP.'}",https://link.springer.com/content/pdf/10.1007/s11263-023-01904-9.pdf
-"large language model is not a good few-shot information extractor, but a good reranker for hard samples!",Yubo Ma,"Large Language Models (LLMs) have made remarkable strides in various tasks. Whether LLMs are competitive few-shot solvers for information extraction (IE) tasks, however, remains an open problem. In this work, we aim to provide a thorough answer to this question. Through extensive experiments on nine datasets across four IE tasks, we demonstrate that current advanced LLMs consistently exhibit inferior performance, higher latency, and increased budget requirements compared to fine-tuned SLMs under most settings. Therefore, we conclude that LLMs are not effective few-shot information extractors in general. Nonetheless, we illustrate that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with. And moreover, we propose an adaptive filter-then-rerank paradigm to combine the strengths of LLMs and SLMs. In this paradigm, SLMs serve as filters and LLMs serve as rerankers. By prompting LLMs to rerank a small portion of difficult samples identified by SLMs, our preliminary system consistently achieves promising improvements (2.4% F1-gain on average) on various IE tasks, with an acceptable time and cost investment.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that with appropriate prompting strategies, LLMs can effectively complement SLMs and tackle challenging samples that SLMs struggle with, and an adaptive filter-then-rerank paradigm is proposed to combine the strengths of LLMs and SLMs.'}",http://arxiv.org/pdf/2303.08559
-towards few-shot identification of morality frames using in-context learning,Shamik Roy,"Data scarcity is a common problem in NLP, especially when the annotation pertains to nuanced socio-linguistic concepts that require specialized knowledge. As a result, few-shot identification of these concepts is desirable. Few-shot in-context learning using pre-trained Large Language Models (LLMs) has been recently applied successfully in many NLP tasks. In this paper, we study few-shot identification of a psycho-linguistic concept, Morality Frames (Roy et al., 2021), using LLMs. Morality frames are a representation framework that provides a holistic view of the moral sentiment expressed in text, identifying the relevant moral foundation (Haidt and Graham, 2007) and at a finer level of granularity, the moral sentiment expressed towards the entities mentioned in the text. Previous studies relied on human annotation to identify morality frames in text which is expensive. In this paper, we propose prompting based approaches using pretrained Large Language Models for identification of morality frames, relying only on few-shot exemplars. We compare our models’ performance with few-shot RoBERTa and found promising results.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes prompting based approaches using pretrained Large Language Models for identification of morality frames, relying only on few-shot exemplars, and compares the models’ performance with few- shot RoBERTa and found promising results.'}",http://arxiv.org/pdf/2302.02029
-enhancing few-shot text-to-sql capabilities of large language models: a study on prompt design strategies,Linyong Nan,"In-context learning (ICL) has emerged as a new approach to various natural language processing tasks, utilizing large language models (LLMs) to make predictions based on context that has been supplemented with a few examples or task-specific instructions. In this paper, we aim to extend this method to question answering tasks that utilize structured knowledge sources, and improve Text-to-SQL systems by exploring various prompt design strategies for employing LLMs. We conduct a systematic investigation into different demonstration selection methods and optimal instruction formats for prompting LLMs in the Text-to-SQL task. Our approach involves leveraging the syntactic structure of an example's SQL query to retrieve demonstrations, and we demonstrate that pursuing both diversity and similarity in demonstration selection leads to enhanced performance. Furthermore, we show that LLMs benefit from database-related knowledge augmentations. Our most effective strategy outperforms the state-of-the-art system by 2.5 points (Execution Accuracy) and the best fine-tuned system by 5.1 points on the Spider dataset. These results highlight the effectiveness of our approach in adapting LLMs to the Text-to-SQL task, and we present an analysis of the factors contributing to the success of our strategy.","{'model': 'tldr@v2.0.0', 'text': 'This paper conducts a systematic investigation into different demonstration selection methods and optimal instruction formats for prompting LLMs in the Text-to-SQL task, and presents an analysis of the factors contributing to the success of this strategy.'}",http://arxiv.org/pdf/2305.12586
-few-shot and prompt training for text classification in german doctor's letters,Phillip Richter-Pechanski,"To classify sentences in cardiovascular German doctor's letters into eleven section categories, we used pattern-exploiting training, a prompt-based method for text classification in few-shot learning scenarios (20, 50 and 100 instances per class) using language models with various pre-training approaches evaluated on CARDIO:DE, a freely available German clinical routine corpus. Prompting improves results by 5-28% accuracy compared to traditional methods, reducing manual annotation efforts and computational costs in a clinical setting.","{'model': 'tldr@v2.0.0', 'text': ""To classify sentences in cardiovascular German doctor's letters into eleven section categories, pattern-exploiting training is used, a prompt-based method for text classification in few-shot learning scenarios, improving results by 5-28% accuracy compared to traditional methods and reducing manual annotation efforts and computational costs in a clinical setting.""}",https://ebooks.iospress.nl/pdf/doi/10.3233/SHTI230275
-exploring zero and few-shot techniques for intent classification,S. Parikh,"Conversational NLU providers often need to scale to thousands of intent-classification models where new customers often face the cold-start problem. Scaling to so many customers puts a constraint on storage space as well. In this paper, we explore four different zero and few-shot intent classification approaches with this low-resource constraint: 1) domain adaptation, 2) data augmentation, 3) zero-shot intent classification using descriptions large language models (LLMs), and 4) parameter-efficient fine-tuning of instruction-finetuned language models. Our results show that all these approaches are effective to different degrees in low-resource settings. Parameter-efficient fine-tuning using T-few recipe on Flan-T5 yields the best performance even with just one sample per intent. We also show that the zero-shot method of prompting LLMs using intent descriptions is also very competitive.","{'model': 'tldr@v2.0.0', 'text': 'Parameter-efficient fine-tuning using T-few recipe on Flan-T5 yields the best performance even with just one sample per intent, and the zero-shot method of prompting LLMs using intent descriptions is also very competitive.'}",http://arxiv.org/pdf/2305.07157
-knowledge-guided prompt learning for few-shot text classification,Liangguo Wang,"Recently, prompt-based learning has shown impressive performance on various natural language processing tasks in few-shot scenarios. The previous study of knowledge probing showed that the success of prompt learning contributes to the implicit knowledge stored in pre-trained language models. However, how this implicit knowledge helps solve downstream tasks remains unclear. In this work, we propose a knowledge-guided prompt learning method that can reveal relevant knowledge for text classification. Specifically, a knowledge prompting template and two multi-task frameworks were designed, respectively. The experiments demonstrated the superiority of combining knowledge and prompt learning in few-shot text classification.","{'model': 'tldr@v2.0.0', 'text': 'A knowledge-guided prompt learning method that can reveal relevant knowledge for text classification and demonstrates the superiority of combining knowledge and prompt learning in few-shot text classification.'}",https://www.mdpi.com/2079-9292/12/6/1486/pdf?version=1679462243
-a smashed glass cannot be full: generation of commonsense explanations through prompt-based few-shot learning,Andrea Zaninello,"We assume that providing explanations is a process to elicit implicit knowledge in human communication, and propose a general methodology to generate commonsense explanations from pairs of semantically related sentences. We take advantage of both prompting applied to large, encoder-decoder pre-trained language models, and few-shot learning techniques, such as pattern-exploiting training. Experiments run on the e-SNLI dataset show that the proposed method achieves state-of-the-art results on the explanation generation task, with a substantial reduction of labelled data. The obtained results open new perspective on a number of tasks involving the elicitation of implicit knowledge.","{'model': 'tldr@v2.0.0', 'text': 'A general methodology to generate commonsense explanations from pairs of semantically related sentences is proposed, taking advantage of both prompting applied to large, encoder-decoder pre-trained language models, and few-shot learning techniques, such as pattern-exploiting training.'}",https://aclanthology.org/2023.nlrse-1.3.pdf
-successive prompting for decomposing complex questions,Dheeru Dua,"Answering complex questions that require making latent decisions is a challenging task, especially when limited supervision is available. Recent works leverage the capabilities of large language models (LMs) to perform complex question answering in a few-shot setting by demonstrating how to output intermediate rationalizations while solving the complex question in a single pass. We introduce “Successive Prompting” where, we iteratively break down a complex task into a simple task, solve it, and then repeat the process until we get the final solution. Successive prompting decouples the supervision for decomposing complex questions from the supervision for answering simple questions, allowing us to (1) have multiple opportunities to query in-context examples at each reasoning step (2) learn question decomposition separately from question answering, including using synthetic data, and (3) use bespoke (fine-tuned) components for reasoning steps where a large LM does not perform well. The intermediate supervision is typically manually written, which can be expensive to collect. We introduce a way to generate synthetic dataset which can be used to bootstrap model’s ability to decompose and answer intermediate questions. Our best model (with successive prompting) achieves an improvement in F1 of ~5% when compared with a state-of-the-art model with synthetic augmentations and few-shot version of the DROP dataset.","{'model': 'tldr@v2.0.0', 'text': 'A way to generate synthetic dataset which can be used to bootstrap model’s ability to decompose and answer intermediate questions is introduced and achieves an improvement in F1 of ~5% when compared with a state-of-the-art model with synthetic augmentations and few-shot version of the DROP dataset.'}",https://arxiv.org/pdf/2212.04092
-"machine translation with large language models: prompting, few-shot learning, and fine-tuning with qlora",Xuan Zhang,"While large language models have made remarkable advancements in natural language generation, their potential in machine translation, especially when fine-tuned, remains under-explored. In our study, we conduct comprehensive experiments, evaluating 15 publicly available language models on machine translation tasks. We compare the performance across three methodologies: zero-shot prompting, few-shot learning, and fine-tuning. Central to our approach is the use of QLoRA, an efficient fine-tuning method. On French-English, QLoRA fine-tuning outperforms both few-shot learning and models trained from scratch. This superiority is highlighted in both sentence-level and document-level translations, with a significant BLEU score improvement of 28.93 over the prompting method. Impressively, with QLoRA, the enhanced performance is achieved by fine-tuning a mere 0.77% of the model’s parameters.","{'model': 'tldr@v2.0.0', 'text': 'On French-English, QLoRA fine-tuning outperforms both few-shot learning and models trained from scratch, and is highlighted in both sentence-level and document-level translations.'}",
-"structured prompting: scaling in-context learning to 1, 000 examples",Y. Hao,"Large language models have exhibited intriguing in-context learning capability, achieving promising zero- and few-shot performance without updating the parameters. However, conventional in-context learning is usually restricted by length constraints, rendering it ineffective to absorb supervision from a large number of examples. In order to go beyond few shots, we introduce structured prompting that breaks the length limit and scales in-context learning to thousands of examples. Specifically, demonstration examples are separately encoded with well-designed position embeddings, and then they are jointly attended by the test example using a rescaled attention mechanism. So we can scale the number of exemplars with linear complexity instead of quadratic complexity with respect to length. Experimental results on a diverse set of tasks show that our approach improves end-task performance and reduces evaluation variance over conventional in-context learning as the number of demonstration examples increases. Code has been released at https://aka.ms/structured-prompting .","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on a diverse set of tasks show that the structured prompting approach improves end-task performance and reduces evaluation variance over conventional in-context learning as the number of demonstration examples increases.'}",http://arxiv.org/pdf/2212.06713
-short answer grading using one-shot prompting and text similarity scoring model,Su-Youn Yoon,"In this study, we developed an automated short answer grading (ASAG) model that provided both analytic scores and final holistic scores. Short answer items typically consist of multiple sub-questions, and providing an analytic score and the text span relevant to each sub-question can increase the interpretability of the automated scores. Furthermore, they can be used to generate actionable feedback for students. Despite these advantages, most studies have focused on predicting only holistic scores due to the difficulty in constructing dataset with manual annotations. To address this difficulty, we used large language model (LLM)-based one-shot prompting and a text similarity scoring model with domain adaptation using small manually annotated dataset. The accuracy and quadratic weighted kappa of our model were 0.67 and 0.71 on a subset of the publicly available ASAG dataset. The model achieved a substantial improvement over the majority baseline.","{'model': 'tldr@v2.0.0', 'text': 'An automated short answer grading (ASAG) model that provided both analytic scores and final holistic scores was developed that achieved a substantial improvement over the majority baseline.'}",http://arxiv.org/pdf/2305.18638
-zero-shot prompting for implicit intent prediction and recommendation with commonsense reasoning,Hui-Chi Kuo,"Intelligent virtual assistants are currently designed to perform tasks or services explicitly mentioned by users, so multiple related domains or tasks need to be performed one by one through a long conversation with many explicit intents. Instead, human assistants are capable of reasoning (multiple) implicit intents based on user utterances via commonsense knowledge, reducing complex interactions and improving practicality. Therefore, this paper proposes a framework of multi-domain dialogue systems, which can automatically infer implicit intents based on user utterances and then perform zero-shot prompting using a large pre-trained language model to trigger suitable single task-oriented bots. The proposed framework is demonstrated effective to realize implicit intents and recommend associated bots in a zero-shot manner.","{'model': 'tldr@v2.0.0', 'text': 'A framework of multi-domain dialogue systems which can automatically infer implicit intents based on user utterances and then perform zero-shot prompting using a large pre-trained language model to trigger suitable single task-oriented bots is proposed.'}",http://arxiv.org/pdf/2210.05901
-personalize segment anything model with one shot,Renrui Zhang,"Driven by large-data pre-training, Segment Anything Model (SAM) has been demonstrated as a powerful and promptable framework, revolutionizing the segmentation models. Despite the generality, customizing SAM for specific visual concepts without man-powered prompting is under explored, e.g., automatically segmenting your pet dog in different images. In this paper, we propose a training-free Personalization approach for SAM, termed as PerSAM. Given only a single image with a reference mask, PerSAM first localizes the target concept by a location prior, and segments it within other images or videos via three techniques: target-guided attention, target-semantic prompting, and cascaded post-refinement. In this way, we effectively adapt SAM for private use without any training. To further alleviate the mask ambiguity, we present an efficient one-shot fine-tuning variant, PerSAM-F. Freezing the entire SAM, we introduce two learnable weights for multi-scale masks, only training 2 parameters within 10 seconds for improved performance. To demonstrate our efficacy, we construct a new segmentation dataset, PerSeg, for personalized evaluation, and test our methods on video object segmentation with competitive performance. Besides, our approach can also enhance DreamBooth to personalize Stable Diffusion for text-to-image generation, which discards the background disturbance for better target appearance learning. Code is released at https://github.com/ZrrSkywalker/Personalize-SAM","{'model': 'tldr@v2.0.0', 'text': 'A training-free Personalization approach for SAM, termed as PerSAM, which effectively adapt SAM for private use without any training, and can enhance DreamBooth to personalize Stable Diffusion for text-to-image generation, which discards the background disturbance for better target appearance learning.'}",https://arxiv.org/pdf/2305.03048
-one-shot learning for function-specific region segmentation in mouse brain,Xu Zhang,"A brain contains a large number of structured regions responsible for diverse functions. Detailed region annotations upon stereotaxic coordinates are highly rare, prompting the need of using one or very few available annotated results of a specific brain section to label images of broadly accessible brain section samples. Here we develop a one-shot learning approach to segment regions of mouse brains. Using the highly ordered geometry of brains, we introduce a reference mask to incorporate both the anatomical structure (visual information) and the brain atlas into brain segmentation. Using the UNet model with this reference mask, we are able to predict the region of hippocampus with high accuracy. We further implement it to segment brain images into 95 detailed regions augmented from the annotation on only one image from Allen Brain Atlas. Together, our one-shot learning method provides neuroscientists an efficient way for brain segmentation and facilitates future region-specific functional studies of brains.","{'model': 'tldr@v2.0.0', 'text': 'A one-shot learning approach to segment regions of mouse brains using a reference mask to incorporate both the anatomical structure and the brain atlas into brain segmentation and is able to predict the region of hippocampus with high accuracy.'}",
-naturalspeech 2: latent diffusion models are natural and zero-shot speech and singing synthesizers,Kai Shen,"Scaling text-to-speech (TTS) to large-scale, multi-speaker, and in-the-wild datasets is important to capture the diversity in human speech such as speaker identities, prosodies, and styles (e.g., singing). Current large TTS systems usually quantize speech into discrete tokens and use language models to generate these tokens one by one, which suffer from unstable prosody, word skipping/repeating issue, and poor voice quality. In this paper, we develop NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input. To enhance the zero-shot capability that is important to achieve diverse speech synthesis, we design a speech prompting mechanism to facilitate in-context learning in the diffusion model and the duration/pitch predictor. We scale NaturalSpeech 2 to large-scale datasets with 44K hours of speech and singing data and evaluate its voice quality on unseen speakers. NaturalSpeech 2 outperforms previous TTS systems by a large margin in terms of prosody/timbre similarity, robustness, and voice quality in a zero-shot setting, and performs novel zero-shot singing synthesis with only a speech prompt. Audio samples are available at https://speechresearch.github.io/naturalspeech2.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input.'}",http://arxiv.org/pdf/2304.09116
-the correlation of a scientific approach using the probing prompting technique and a problem-based learning model on learning outcomes on reaction rate material,Wardhatul Jannah,"Abstract─This study began with the learning results on the subject of the still-low reaction rate. The purpose of this study is to discover the relationship between the scientific method and learning outcomes in the reaction rate material by using the probing prompting technique and the issue based learning model alone or combination. Quantitative research using correlational methodologies and a one-shot case study design is the approach of choice. Class XI MIPA SMAN 1 Padang and XI MIPA SMAN 14 Padang were the subjects of this research. A total of 144 persons were selected for the study using a simple random sampling technique. Questionnaires and learning outcomes assessments in the form of 22 objective questions that had been verified for validity and reliability were used in this study. The results found a positive and significant correlation between: 1) the scientific approach to learning outcomes using the probing prompting technique, with a correlation coefficient of 0.903 suggesting a very strong correlation. This suggests that by using the probing prompting technique, the scientific approach module can improve learning results by 81.5 percent. 2) A scientific approach to learning outcomes using a problem-based learning model with a correlation coefficient of 0.770, suggesting a strong link. This suggests that the scientific approach module, when combined with a problem-based learning model, has an effective contribution of 59.3 percent in improving learning outcomes. 3) Using the probing prompting technique and the issue based learning model together on the reaction rate learning outcomes yielded a correlation coefficient of 0.958, indicating a very strong correlation. The contribution of the two independent variables to the dependent variable in this study is 91.8 percent, with the remainder impacted by other variables. Keywords─Scientific Approach; Probing Prompting; Problem-Based Learning; Learning Outcomes; Reaction Rate","{'model': 'tldr@v2.0.0', 'text': 'The results found a positive and significant correlation between the scientific approach to learning outcomes using the probing prompting technique and the issue based learning model alone or combination, and a correlation coefficient of 0.903 suggesting a very strong correlation.'}",
-udapdr: unsupervised domain adaptation via llm prompting and distillation of rerankers,Jon Saad-Falcon,"Many information retrieval tasks require large labeled datasets for fine-tuning. However, such datasets are often unavailable, and their utility for real-world applications can diminish quickly due to domain shifts. To address this challenge, we develop and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply. The method begins by generating a small number of synthetic queries using an expensive LLM. After that, a much less expensive one is used to create large numbers of synthetic queries, which are used to fine-tune a family of reranker models. These rerankers are then distilled into a single efficient retriever for use in the target domain. We show that this technique boosts zero-shot accuracy in long-tail domains and achieves substantially lower latency than standard reranking methods.","{'model': 'tldr@v2.0.0', 'text': 'This work develops and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply and achieves substantially lower latency than standard reranking methods.'}",https://arxiv.org/pdf/2303.00807
-probing power by prompting: harnessing pre-trained language models for power connotation framing,Shima Khanehzar,"When describing actions, subtle changes in word choice can evoke very different associations with the involved entities. For instance, a company ‘{{it employing} workers’ evokes a more positive connotation than the one ‘{{it exploiting}’ them. This concept is called {{it connotation}. This paper investigates whether pre-trained language models (PLMs) encode such subtle connotative information about {{it power differentials} between involved entities. We design a probing framework for power connotation, building on~{citet{sap-etal-2017-connotation}’s operationalization of {{it connotation frames}. We show that zero-shot prompting of PLMs leads to above chance prediction of power connotation, however fine-tuning PLMs using our framework drastically improves their accuracy. Using our fine-tuned models, we present a case study of {{it power dynamics} in US news reporting on immigration, showing the potential of our framework as a tool for understanding subtle bias in the media.","{'model': 'tldr@v2.0.0', 'text': ""It is shown that zero-shot prompting of PLMs leads to above chance prediction of power connotation, however fine-tuning PLMs using the authors' framework drastically improves their accuracy, showing the potential of the framework as a tool for understanding subtle bias in the media.""}",https://aclanthology.org/2023.eacl-main.61.pdf
-language models are causal knowledge extractors for zero-shot video question answering,Hung-Ting Su,"Causal Video Question Answering (CVidQA) queries not only association or temporal relations but also causal relations in a video. Existing question synthesis methods pretrained question generation (QG) systems on reading comprehension datasets with text descriptions as inputs. However, QG models only learn to ask association questions (e.g., ""what is someone doing…"") and result in inferior performance due to the poor transfer of association knowledge to CVidQA, which focuses on causal questions like ""why is someone doing …"". Observing this, we proposed to exploit causal knowledge to generate question-answer pairs, and proposed a novel framework, Causal Knowledge Extraction from Language Models (CaKE-LM), leveraging causal commonsense knowledge from language models to tackle CVidQA. To extract knowledge from LMs, CaKE-LM generates causal questions containing two events with one triggering another (e.g., ""score a goal"" triggers ""soccer player kicking ball"") by prompting LM with the action (soccer player kicking ball) to retrieve the intention (to score a goal). CaKE-LM significantly outperforms conventional methods by 4% to 6% of zero-shot CVidQA accuracy on NExT-QA and Causal-VidQA datasets. We also conduct comprehensive analyses and provide key findings for future research.","{'model': 'tldr@v2.0.0', 'text': 'Causal Knowledge Extraction from Language Models (CaKE-LM) is proposed, leveraging causal commonsense knowledge from language models to tackle CVidQA, and significantly outperforms conventional methods by 4% to 6% of zero-shot CVidZA accuracy on NExT-QA and Causal-Vid QA datasets.'}",https://arxiv.org/pdf/2304.03754
-what do language models know about word senses? zero-shot wsd with language models and domain inventories,Oscar Sainz,"Language Models are the core for almost any Natural Language Processing system nowadays. One of their particularities is their contextualized representations, a game changer feature when a disambiguation between word senses is necessary. In this paper we aim to explore to what extent language models are capable of discerning among senses at inference time. We performed this analysis by prompting commonly used Languages Models such as BERT or RoBERTa to perform the task of Word Sense Disambiguation (WSD). We leverage the relation between word senses and domains, and cast WSD as a textual entailment problem, where the different hypothesis refer to the domains of the word senses. Our results show that this approach is indeed effective, close to supervised systems.","{'model': 'tldr@v2.0.0', 'text': 'This analysis leverages the relation between word senses and domains, and cast WSD as a textual entailment problem, where the different hypothesis refer to the domains of the word senses.'}",http://arxiv.org/pdf/2302.03353
-compresso: structured pruning with collaborative prompting learns compact large language models,Song Guo,"Despite the remarkable success of Large Language Models (LLMs), the massive size poses significant deployment challenges, particularly on resource-constrained hardware. While existing LLM compression methods focus on quantization, pruning remains relatively unexplored due to the high cost of training-based approaches and data collection challenges. One-shot pruning methods, although cost-effective and data-free, have become dominant in LLM pruning, but lead to performance decline under the structured pruning setting. In this work, we introduce a new paradigm for structurally pruning LLMs, called Compresso. Our approach, through the collaboration of the proposed resource-efficient pruning algorithm and the LLM itself, learns optimal pruning decisions during the training process. Compresso addresses the challenges of expensive training costs and data collection by incorporating Low-Rank Adaptation (LoRA) into the $L_0$ regularization during the instruction tuning process. Then, we further augment the pruning algorithm by introducing a collaborative prompt that fosters collaboration between the LLM and the pruning algorithm, significantly boosting the overall performance. To this end, Compresso prunes LLaMA-7B to 5.4B, maintaining original performance and even surpassing LLaMA-7B in reading comprehension by 2.62%. Extensive experiments demonstrate that Compresso significantly outperforms one-shot pruning baselines across various sparsity ratios, achieving up to 2.21%, 11.43%, 7.04%, and 4.81% higher scores on the commonsense reasoning, reading comprehension, MMLU, and BBH benchmarks, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a new paradigm for structurally pruning LLMs, called Compresso, which addresses the challenges of expensive training costs and data collection by incorporating Low-Rank Adaptation into the $L_0$ regularization during the instruction tuning process and introduces a collaborative prompt that fosters collaboration between the LLM and the pruning algorithm, significantly boosting the overall performance.'}",https://arxiv.org/pdf/2310.05015
-you can generate it again: data-to-text generation with verification and correction prompting,Xuan Ren,"Despite significant advancements in existing models, generating text descriptions from structured data input, known as data-to-text generation, remains a challenging task. In this paper, we propose a novel approach that goes beyond traditional one-shot generation methods by introducing a multi-step process consisting of generation, verification, and correction stages. Our approach, VCP(Verification and Correction Prompting), begins with the model generating an initial output. We then proceed to verify the correctness of different aspects of the generated text. The observations from the verification step are converted into a specialized error-indication prompt, which instructs the model to regenerate the output while considering the identified errors. To enhance the model's correction ability, we have developed a carefully designed training procedure. This procedure enables the model to incorporate feedback from the error-indication prompt, resulting in improved output generation. Through experimental results, we demonstrate that our approach effectively reduces slot error rates while maintaining the overall quality of the generated text.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach that goes beyond traditional one-shot generation methods by introducing a multi-step process consisting of generation, verification, and correction stages, and demonstrates that this approach effectively reduces slot error rates while maintaining the overall quality of the generated text.'}",http://arxiv.org/pdf/2306.15933
-transprompt v2: a transferable prompting framework for cross-task text classification,J. Wang,"Text classification is one of the most imperative tasks in natural language processing (NLP). Recent advances with pre-trained language models (PLMs) have shown remarkable success on this task. However, the satisfying results obtained by PLMs heavily depend on the large amounts of task-specific labeled data, which may not be feasible in many application scenarios due to data access and privacy constraints. The recently-proposed prompt-based fine-tuning paradigm improves the performance of PLMs for few-shot text classification with task-specific templates. Yet, it is unclear how the prompting knowledge can be transferred across tasks, for the purpose of mutual reinforcement. We propose TransPrompt v2, a novel transferable prompting framework for few-shot learning across similar or distant text classification tasks. For learning across similar tasks, we employ a multi-task meta-knowledge acquisition (MMA) procedure to train a meta-learner that captures the cross-task transferable knowledge. For learning across distant tasks, we further inject the task type descriptions into the prompt, and capture the intra-type and inter-type prompt embeddings among multiple distant tasks. Additionally, two de-biasing techniques are further designed to make the trained meta-learner more task-agnostic and unbiased towards any tasks. After that, the meta-learner can be adapted to each specific task with better parameters initialization. Extensive experiments show that TransPrompt v2 outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets. We further show that the meta-learner can effectively improve the performance of PLMs on previously unseen tasks. In addition, TransPrompt v2 also outperforms strong fine-tuning baselines when learning with full training sets.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that TransPrompt v2 outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets, and shows that the meta-learner can effectively improve the performance of PLMs on previously unseen tasks.'}",https://arxiv.org/pdf/2308.15010
-dynamic strategy chain: dynamic zero-shot cot for long mental health support generation,Qi Chen,"Long counseling Text Generation for Mental health support (LTGM), an innovative and challenging task, aims to provide help-seekers with mental health support through a comprehensive and more acceptable response. The combination of chain-of-thought (CoT) prompting and Large Language Models (LLMs) is employed and get the SOTA performance on various NLP tasks, especially on text generation tasks. Zero-shot CoT prompting is one of the most common methods in CoT prompting. However, in the LTGM task, Zero-shot CoT prompting can not simulate a counselor or provide personalized strategies without effective mental health counseling strategy prompts. To tackle this challenge, we propose a zero-shot Dynamic Strategy Chain (DSC) prompting method. Firstly, we utilize GPT2 to learn the responses written by mental health counselors and dynamically generate mental health counseling strategies tailored to the help-seekers' needs. Secondly, the Zero-shot DSC prompting is constructed according to mental health counseling strategies and the help-seekers' post. Finally, the Zero-shot DSC prompting is employed to guide LLMs in generating more human-like responses for the help-seekers. Both automatic and manual evaluations demonstrate that Zero-shot DSC prompting can deliver more human-like responses than CoT prompting methods on LTGM tasks.",,https://arxiv.org/pdf/2308.10444
-adapt and decompose: efficient generalization of text-to-sql via domain adapted least-to-most prompting,Aseem Arora,"Cross-domain and cross-compositional generalization of Text-to-SQL semantic parsing is a challenging task. Existing Large Language Model (LLM) based solutions rely on inference-time retrieval of few-shot exemplars from the training set to synthesize a run-time prompt for each Natural Language (NL) test query. In contrast, we devise an algorithm which performs offline sampling of a minimal set-of few-shots from the training data, with complete coverage of SQL clauses, operators and functions, and maximal domain coverage within the allowed token length. This allows for synthesis of a fixed Generic Prompt (GP), with a diverse set-of exemplars common across NL test queries, avoiding expensive test time exemplar retrieval. We further auto-adapt the GP to the target database domain (DA-GP), to better handle cross-domain generalization; followed by a decomposed Least-To-Most-Prompting (LTMP-DA-GP) to handle cross-compositional generalization. The synthesis of LTMP-DA-GP is an offline task, to be performed one-time per new database with minimal human intervention. Our approach demonstrates superior performance on the KaggleDBQA dataset, designed to evaluate generalizability for the Text-to-SQL task. We further showcase consistent performance improvement of LTMP-DA-GP over GP, across LLMs and databases of KaggleDBQA, highlighting the efficacy and model agnostic benefits of our prompt based adapt and decompose approach.","{'model': 'tldr@v2.0.0', 'text': 'This work devise an algorithm which performs offline sampling of a minimal set-of few-shots from the training data, with complete coverage of SQL clauses, operators and functions, and maximal domain coverage within the allowed token length, for cross-domain and cross-compositional generalization of Text-to-SQL semantic parsing.'}",https://arxiv.org/pdf/2308.02582
-the effectivity of the application technical probing prompting on story problems is reviewed from mathematical creative thinking skills to high school student,Dede Suhendar,"This article present the results of research in the form of effectiveness using probing prompting techniques and experimental research methods with desain pre experimental : one-shot case study design by purposive sampling conducted on class X students in the 2018/2019 academic year on mathematical questions in the form of story problems. While the objectives of this study include exposing the effectiveness of mathematical creative thinking abilities of students using probing promting techniques, obtaining a number of descriptions of students problems in the process of creative thinking in mathematical questions form of story matter. Based on data analysis the results of the study showed that the percentage of students who have good mathematical creative thinking skills is less than 60%, so the probing prompting technique is not effective in terms of mathematical creative thinking abilities students.",,https://journal.ikipsiliwangi.ac.id/index.php/jiml/article/download/3975/1469
-efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa,Muhamad Ruslan Layn,"Penelitian ini bertujuan untuk mengetahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Kota Sorong tahun pelajaran 2017/2018 yang terdiri dari 8 kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik purposive sampling. Desain penelitian ini adalah one shot case study. Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 65%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa.",,http://e-journal.unipma.ac.id/index.php/JEMS/article/download/2013/1477
-analisis kemampuan berpikir deduktif aksiomatik mahasiswa menggunakan teknik probing-prompting pada mata kuliah analisis real lanjut,Karunia Eka Lestari,"Penelitian ini menggunakan pendekatan deskriptif kuantitatif dengan metode eksperimen. Desain eksperimen yang digunakan yaitu pre-experimental design tipe one-shot case study . Sampel dalam penelitian ini adalah seluruh mahasiswa semester VII program studi pendidikan matematika FKIP UNSIKA tahun ajaran 2014-2015 yang mengontrak mata kuliah analisis real lanjut sebanyak 38 mahasiswa. Berdasarkan hasil analisis data instrumen tes diperoleh kesimpulan bahwa Pada taraf kepercayaan 95% tidak cukup bukti untuk menyatakan bahwa kemampuan berpikir deduktif aksiomatik mahasiswa menggunakan teknik probing-prompting pada mata kuliah analisis real lanjut secara signifikan lebih besar dari 75% dari kriteria ideal yang ditetapkan. Selanjutnya hasil analisis data instrumen non tes menunjukkan bahwa kesulitan yang dialami oleh mahasiswa dalam berpikir deduktif aksiomatik menggunakan teknik probing-prompting pada mata kuliah analisis real lanjut, yaitu: (1) kesulitan dalam membaca simbol-simbol matematis; (2) kesulitan dalam memahami definisi, aksioma, postulat, dan teorema (3) kesulitan memulai dan melakukan proses pembuktian; (4) kesulitan memilih dan menggunakan definisi, aksioma, postulat, atau teorema untuk melakukan pembuktian; dan (5) kesulitan menerapkan definisi, aksioma, postulat dan teorema pada permasalahan dalam kehidupan sehari-hari. Faktor yang menyebabkan mahasiswa mengalami kesulitan dalam berpikir deduktif aksiomatik pada mata kuliah analisis real lanjut adalah: (1) faktor teknis; (2) faktor habit of mind, (3) faktor penguasaan materi dan (4) faktor intensitas latihan. Kata kunci : Kemampuan Berpikir Deduktif Aksiomatik, Teknik Probing-Prompting",,
-leveraging large language models for multiple choice question answering,Joshua Robinson,"While large language models (LLMs) like GPT-3 have achieved impressive results on multiple choice question answering (MCQA) tasks in the zero, one, and few-shot settings, they generally lag behind the MCQA state of the art (SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks. An LLM is conditioned on a question (without the associated answer options) and its chosen option is the one assigned the highest probability after normalization (for length, etc.). A more natural prompting approach is to present the question and answer options to the LLM jointly and have it output the symbol (e.g.,""A"") associated with its chosen answer option. This approach allows the model to explicitly compare answer options, reduces computational costs, and mitigates the effects of tokenization scheme and answer option representations on answer selection. For the natural approach to be effective, the LLM it is used with must be able to associate answer options with the symbols that represent them. The LLM needs what we term multiple choice symbol binding (MCSB) ability. This ability varies greatly by model. We show that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated.'}",http://arxiv.org/pdf/2210.12353
-data augmentation for intent classification with off-the-shelf large language models,Gaurav Sahu,"Data augmentation is a widely employed technique to alleviate the problem of data scarcity. In this work, we propose a prompting-based approach to generate labelled training data for intent classification with off-the-shelf language models (LMs) such as GPT-3. An advantage of this method is that no task-specific LM-fine-tuning for data generation is required; hence the method requires no hyper parameter tuning and is applicable even when the available training data is very scarce. We evaluate the proposed method in a few-shot setting on four diverse intent classification tasks. We find that GPT-generated data significantly boosts the performance of intent classifiers when intents in consideration are sufficiently distinct from each other. In tasks with semantically close intents, we observe that the generated data is less helpful. Our analysis shows that this is because GPT often generates utterances that belong to a closely-related intent instead of the desired one. We present preliminary evidence that a prompting-based GPT classifier could be helpful in filtering the generated data to enhance its quality.","{'model': 'tldr@v2.0.0', 'text': 'It is found that GPT-generated data significantly boosts the performance of intent classifiers when intents in consideration are sufficiently distinct from each other, and preliminary evidence that a prompting-based GPT classifier could be helpful in filtering the generated data to enhance its quality is presented.'}",http://arxiv.org/pdf/2204.01959
-efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa,L. Mulyadi,"This quasi-experimental research aimed to know the effectiveness of probing-prompting model viewed by students critical thinking ability . The population of this research was all students of grade VIII of SMP N egeri 9 Bandar lampung in academic year of 201 4 /201 5 that consist of 7 classes . The sample of this research was students of VII I-C which was taken by purposive sampling technique . This research design was one shot case study . Instrument of this research was essay test of critical thinking . Based on the analysis of data, the percentage of students who ha ve good critical thinking ability was less than 60%. Thus, probing-prompting model was less effective viewed by critical thinking ability . Penelitian kuasi eksperimen ini bertujuan untuk mengetahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Bandarlampung tahun pelajaran 2014/2015 yang terdiri dari tujuh kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik p urposive sampling . Desain penelitian ini adalah one shot case study . Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 60%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa. Kata kunci: berpikir kritis, efektivitas, probing-prompting",,
-multi-party chat: conversational agents in group settings with humans and models,Jimmy Wei,"Current dialogue research primarily studies pairwise (two-party) conversations, and does not address the everyday setting where more than two speakers converse together. In this work, we both collect and evaluate multi-party conversations to study this more general case. We use the LIGHT environment to construct grounded conversations, where each participant has an assigned character to role-play. We thus evaluate the ability of language models to act as one or more characters in such conversations. Models require two skills that pairwise-trained models appear to lack: (1) being able to decide when to talk; (2) producing coherent utterances grounded on multiple characters. We compare models trained on our new dataset to existing pairwise-trained dialogue models, as well as large language models with few-shot prompting. We find that our new dataset, MultiLIGHT, which we will publicly release, can help bring significant improvements in the group setting.","{'model': 'tldr@v2.0.0', 'text': 'This work collects and evaluates multi-party conversations and uses the LIGHT environment to construct grounded conversations, where each participant has an assigned character to role-play, to evaluate the ability of language models to act as one or more characters in such conversations.'}",http://arxiv.org/pdf/2304.13835
-efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa (studi pada siswa kelas viii smp negeri 9 bandarlampung tahun pelajaran 2014/2015),M. Mulyadi,"Penelitian kuasi eksperimen ini bertujuan untuk menge¬tahui efektivitas penerapan model probing-prompting ditinjau dari kemampuan berpikir kritis siswa. Populasi penelitian ini adalah seluruh siswa kelas VIII SMP Negeri 9 Bandarlampung Tahun Pelajaran 2014/2015 yang terdiri dari tujuh kelas. Sampel penelitian ini adalah siswa kelas VIII-C yang dipilih dengan teknik purposive sampling. Desain penelitian ini adalah one shot case study. Instrumen penelitian ini adalah tes uraian berpikir kritis. Berdasarkan hasil analisis data, persentase siswa yang memiliki kemampuan berpikir kritis dengan baik kurang dari 60%. Dengan demikian, model probing-prompting tidak efektif ditinjau dari kemampuan berpikir kritis siswa. Kata Kunci : berpikir kritis, efektivitas, probing-prompting",,
-metaprompting: learning to learn better prompts,Yutai Hou,"Prompting method is regarded as one of the crucial progress for few-shot nature language processing. Recent research on prompting moves from discrete tokens based “hard prompts” to continuous “soft prompts”, which employ learnable vectors as pseudo prompt tokens and achieve better performance. Though showing promising prospects, these soft-prompting methods are observed to rely heavily on good initialization to take effect. Unfortunately, obtaining a perfect initialization for soft prompts requires understanding of inner language models working and elaborate design, which is no easy task and has to restart from scratch for each new task. To remedy this, we propose a generalized soft prompting method called MetaPrompting, which adopts the well-recognized model-agnostic meta-learning algorithm to automatically find better prompt initialization that facilitates fast adaptation to new prompting tasks. Extensive experiments show MetaPrompting tackles soft prompt initialization problem and brings significant improvement on three different datasets (over 6 points improvement in accuracy for 1-shot setting), achieving new state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': 'A generalized soft prompting method called MetaPrompting is proposed, which adopts the well-recognized model-agnostic meta-learning algorithm to automatically find better prompt initialization that facilitates fast adaptation to new prompting tasks.'}",http://arxiv.org/pdf/2209.11486
-unraveling chatgpt: a critical analysis of ai-generated goal-oriented dialogues and annotations,Tiziano Labruna,"Large pre-trained language models have exhibited unprecedented capabilities in producing high-quality text via prompting techniques. This fact introduces new possibilities for data collection and annotation, particularly in situations where such data is scarce, complex to gather, expensive, or even sensitive. In this paper, we explore the potential of these models to generate and annotate goal-oriented dialogues, and conduct an in-depth analysis to evaluate their quality. Our experiments employ ChatGPT, and encompass three categories of goal-oriented dialogues (task-oriented, collaborative, and explanatory), two generation modes (interactive and one-shot), and two languages (English and Italian). Based on extensive human-based evaluations, we demonstrate that the quality of generated dialogues and annotations is on par with those generated by humans.","{'model': 'tldr@v2.0.0', 'text': 'The quality of generated dialogues and annotations is on par with those generated by humans, based on extensive human-based evaluations.'}",http://arxiv.org/pdf/2305.14556
-shot for shot remakes,D. Varndell,,,
-improving patient pre-screening for clinical trials: assisting physicians with large language models,D. Hamer,"Physicians considering clinical trials for their patients are met with the laborious process of checking many text based eligibility criteria. Large Language Models (LLMs) have shown to perform well for clinical information extraction and clinical reasoning, including medical tests, but not yet in real-world scenarios. This paper investigates the use of InstructGPT to assist physicians in determining eligibility for clinical trials based on a patient's summarised medical profile. Using a prompting strategy combining one-shot, selection-inference and chain-of-thought techniques, we investigate the performance of LLMs on 10 synthetically created patient profiles. Performance is evaluated at four levels: ability to identify screenable eligibility criteria from a trial given a medical profile; ability to classify for each individual criterion whether the patient qualifies; the overall classification whether a patient is eligible for a clinical trial and the percentage of criteria to be screened by physician. We evaluated against 146 clinical trials and a total of 4,135 eligibility criteria. The LLM was able to correctly identify the screenability of 72% (2,994/4,135) of the criteria. Additionally, 72% (341/471) of the screenable criteria were evaluated correctly. The resulting trial level classification as eligible or ineligible resulted in a recall of 0.5. By leveraging LLMs with a physician-in-the-loop, a recall of 1.0 and precision of 0.71 on clinical trial level can be achieved while reducing the amount of criteria to be checked by an estimated 90%. LLMs can be used to assist physicians with pre-screening of patients for clinical trials. By forcing instruction-tuned LLMs to produce chain-of-thought responses, the reasoning can be made transparent to and the decision process becomes amenable by physicians, thereby making such a system feasible for use in real-world scenarios.","{'model': 'tldr@v2.0.0', 'text': 'By forcing instruction-tuned LLMs to produce chain-of-thought responses, the reasoning can be made transparent to and the decision process becomes amenable by physicians, thereby making such a system feasible for use in real-world scenarios.'}",http://arxiv.org/pdf/2304.07396
-penerapan teknik pembelajaran probing prompting untuk mengetahui kemampuan penalaran matematika siswa kelas 7 g di smpn 1 rejoso,Yayuk Kurniasari,"Matematika dipahami melalui penalaran, agar kemampuan penalaran matematis dapat berkembang secara optimal, siswa harus memiliki kesempatan yang terbuka untuk berpikir. Salah satu teknik mengajar yang dinilai akomodatif dapat meningkatkan aktivitas berpikir siswa adalah teknik Probing Prompting, yaitu teknik pembelajaran dengan cara guru menyajikan serangkaian pertanyaan yang sifatnya menuntun dan menggali sehingga terjadi proses berpikir yang mengaitkan pengetahuan siswa dan pengalamannya dengan pengetahuan baru yang sedang dipelajari.Penelitian ini bertujuan untuk mendeskripsikan kemampuan penalaran matematika dan respon siswa pada pembelajaran dengan teknik Probing Prompting. Subyek penelitian ini adalah siswa SMP Negeri 1 Rejoso Kab. Nganjuk kelas 7 G sebanyak 36 siswa. Rancangan penelitian yang digunakan adalah “One Shot Case Study”.Berdasarkan hasil analisis data yang diperoleh dapat disimpulkan bahwa siswa mampu menyajikan pernyataan matematika secara lisan, tertulis, gambar, dan diagram dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 72 % siswa berada pada level 2 ke atas. Siswa mampu mengajukan dugaan dengan benar meskipun ada beberapa penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 91 % siswa berada pada level 2 ke atas. Siswa mampu melakukan manipulasi matematika dengan benar meskipun penalarannya kurang lengkap, hal ini diketahui sebanyak 78 % siswa berada pada level 2 ke atas. Siswa mampu menyusun bukti, memberikan alasan atau bukti terhadap kebenaran solusi dengan benar dan menggunakan penalaran dengan baik, hal ini diketahui sebanyak 52 % siswa berada pada level 3 ke atas. Siswa mampu menarik kesimpulan dari pernyataan dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini diketahui bahwa sebanyak 86 % berada pada level 2 ke atas. Siswa mampu memeriksa kesahihan suatu argumen dengan benar dan penalaran baik, hal ini diketahui bahwa sebanyak 53 % berada pada level 3 ke atas. Siswa mampu menemukan pola atau sifat gejala matematis untuk membuat generalisasi dengan benar meskipun penalaran yang digunakan kurang lengkap, hal ini dieketahui bahwa sebanyak 87 % berada pada level 2 ke atas. Hasil angket respon siswa menunjukkan bahwa respon siswa adalah positif dengan rata-rata nilai respon siswa sebesar 3,17.",,
-sinc: spatial composition of 3d human motions for simultaneous action generation,Nikos Athanasiou,"Our goal is to synthesize 3D human motions given textual inputs describing simultaneous actions, for example 'waving hand' while 'walking' at the same time. We refer to generating such simultaneous movements as performing 'spatial compositions'. In contrast to temporal compositions that seek to transition from one action to another, spatial compositing requires understanding which body parts are involved in which action, to be able to move them simultaneously. Motivated by the observation that the correspondence between actions and body parts is encoded in powerful language models, we extract this knowledge by prompting GPT-3 with text such as""what are the body parts involved in the action?"", while also providing the parts list and few-shot examples. Given this action-part mapping, we combine body parts from two motions together and establish the first automated method to spatially compose two actions. However, training data with compositional actions is always limited by the combinatorics. Hence, we further create synthetic data with this approach, and use it to train a new state-of-the-art text-to-motion generation model, called SINC (""SImultaneous actioN Compositions for 3D human motions""). In our experiments, that training with such GPT-guided synthetic data improves spatial composition generation over baselines. Our code is publicly available at https://sinc.is.tue.mpg.de/.","{'model': 'tldr@v2.0.0', 'text': 'This work combines body parts from two motions together and establishes the first automated method to spatially compose two actions, and uses it to train a new state-of-the-art text-to-motion generation model, called SINC (""SImultaneous actioN Compositions for 3D human motions"").'}",https://arxiv.org/pdf/2304.10417
-hybrid-regressive neural machine translation,Qiang Wang,"In this work, we empirically confirm that non-autoregressive translation with an iterative refinement mechanism (IR-NAT) suffers from poor acceleration robustness because it is more sensitive to decoding batch size and computing device setting than autoregressive translation (AT). Inspired by it, we attempt to investigate how to combine the strengths of autoregressive and non-autoregressive translation paradigms better. To this end, we demonstrate through synthetic experiments that prompting a small number of AT's predictions can promote one-shot non-autoregressive translation to achieve the equivalent performance of IR-NAT. Following this line, we propose a new two-stage translation prototype called hybrid-regressive translation (HRT). Specifically, HRT first generates discontinuous sequences via autoregression (e.g., make a prediction every k tokens, k>1) and then fills in all previously skipped tokens at once in a non-autoregressive manner. We also propose a bag of techniques to effectively and efficiently train HRT without adding any model parameters. HRT achieves the state-of-the-art BLEU score of 28.49 on the WMT En-De task and is at least 1.5x faster than AT, regardless of batch size and device. In addition, another bonus of HRT is that it successfully inherits the good characteristics of AT in the deep-encoder-shallow-decoder architecture. Concretely, compared to the vanilla HRT with a 6-layer encoder and 6-layer decoder, the inference speed of HRT with a 12-layer encoder and 1-layer decoder is further doubled on both GPU and CPU without BLEU loss.","{'model': 'tldr@v2.0.0', 'text': 'This work empirically confirms that non-autoregressive translation with an iterative refinement mechanism (IR-NAT) suffers from poor acceleration robustness and proposes a new two-stage translation prototype called hybrid-regressive translation (HRT).'}",http://arxiv.org/pdf/2210.10416
-the potential and pitfalls of using a large language model such as chatgpt or gpt-4 as a clinical assistant,Jingqing Zhang,"Recent studies have demonstrated promising performance of ChatGPT and GPT-4 on several medical domain tasks. However, none have assessed its performance using a large-scale real-world electronic health record database, nor have evaluated its utility in providing clinical diagnostic assistance for patients across a full range of disease presentation. We performed two analyses using ChatGPT and GPT-4, one to identify patients with specific medical diagnoses using a real-world large electronic health record database and the other, in providing diagnostic assistance to healthcare workers in the prospective evaluation of hypothetical patients. Our results show that GPT-4 across disease classification tasks with chain of thought and few-shot prompting can achieve performance as high as 96% F1 scores. For patient assessment, GPT-4 can accurately diagnose three out of four times. However, there were mentions of factually incorrect statements, overlooking crucial medical findings, recommendations for unnecessary investigations and overtreatment. These issues coupled with privacy concerns, make these models currently inadequate for real world clinical use. However, limited data and time needed for prompt engineering in comparison to configuration of conventional machine learning workflows highlight their potential for scalability across healthcare applications.","{'model': 'tldr@v2.0.0', 'text': 'This work performed two analyses using ChatGPT and GPT-4, one to identify patients with specific medical diagnoses using a real-world large electronic health record database and the other, in providing diagnostic assistance to healthcare workers in the prospective evaluation of hypothetical patients.'}",https://arxiv.org/pdf/2307.08152
-little giants: exploring the potential of small llms as evaluation metrics in summarization in the eval4nlp 2023 shared task,Neema Kotonya,"This paper describes and analyzes our participation in the 2023 Eval4NLP shared task, which focuses on assessing the effectiveness of prompt-based techniques to empower Large Language Models to handle the task of quality estimation, particularly in the context of evaluating machine translations and summaries. We conducted systematic experiments with various prompting techniques, including standard prompting, prompts informed by annotator instructions, and innovative chain-of-thought prompting. In addition, we integrated these approaches with zero-shot and one-shot learning methods to maximize the efficacy of our evaluation procedures. Our work reveals that combining these approaches using a""small"", open source model (orca_mini_v3_7B) yields competitive results.","{'model': 'tldr@v2.0.0', 'text': None}",
-hitachi at semeval-2023 task 4: exploring various task formulations reveals the importance of description texts on human values,Masaya Tsunokake,"This paper describes our participation in SemEval-2023 Task 4, ValueEval: Identification of Human Values behind Arguments. The aim of this task is to identify whether or not an input text supports each of the 20 pre-defined human values. Previous work on human value detection has shown the effectiveness of a sequence classification approach using BERT. However, little is known about what type of task formulation is suitable for the task. To this end, this paper explores various task formulations, including sequence classification, question answering, and question answering with chain-of-thought prompting and evaluates their performances on the shared task dataset. Experiments show that a zero-shot approach is not as effective as other methods, and there is no one approach that is optimal in every scenario. Our analysis also reveals that utilizing the descriptions of human values can help to improve performance.","{'model': 'tldr@v2.0.0', 'text': 'Various task formulations are explored, including sequence classification, question answering, and question answering with chain-of-thought prompting and evaluates their performances on the shared task dataset and reveals that utilizing the descriptions of human values can help to improve performance.'}",https://aclanthology.org/2023.semeval-1.240.pdf
-learning to paraphrase sentences to different complexity levels,Alison Chi,"While sentence simplification is an active research topic in NLP, its adjacent tasks of sentence complexification and same-level paraphrasing are not. To train models on all three tasks, we present two new unsupervised datasets. We compare these datasets, one labeled by a weak classifier and the other by a rule-based approach, with a single supervised dataset. Using these three datasets for training, we perform extensive experiments on both multitasking and prompting strategies. Compared to other systems trained on unsupervised parallel data, models trained on our weak classifier labeled dataset achieve state-of-the-art performance on the ASSET simplification benchmark. Our models also outperform previous work on sentence level targeting. Finally, we establish how a handful of Large Language Models perform on these tasks under a zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': ""Compared to other systems trained on unsupervised parallel data, models trained on the authors' weak classifier labeled dataset achieve state-of-the-art performance on the ASSET simplification benchmark and outperform previous work on sentence level targeting.""}",https://arxiv.org/pdf/2308.02226
-"investigating the perception of the future in gpt-3, -3.5 and gpt-4",Diana Kozachek,"This study explores the potential of GPT-3, GPT-3.5, and GPT-4, in generating human-like future scenarios to investigate each model's ability to perceive time. The methodology combines a coding-based experiment and an expert survey. The investigation involves fine- and prompt-tuning GPT-3, prompt-tuning GPT-3.5, and few-shot prompting GPT-4 with human-made future scenarios. The models and output are quantitatively and qualitatively analyzed. The survey invited practitioners from fields of foresight and futurology, AI, and NLP to assess whether differences in output can be identified. This study's findings suggest that GPT-3 and GPT-4 generated scenarios are difficult to distinguish from human-made ones, while GPT-3.5 performed more poorly. Yet none of the models can differentiate time horizons and their respective effects on the future from each other. And while no one knows the shape of things to come, this lack of understanding of a core concept of life invites future investigations.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3 and GPT-4 generated scenarios are difficult to distinguish from human-made ones, while G PT-3.5 performed more poorly, and none of the models can differentiate time horizons and their respective effects on the future from each other.'}",
-tree of clarifications: answering ambiguous questions with retrieval-augmented large language models,Gangwoo Kim,"Questions in open-domain question answering are often ambiguous, allowing multiple interpretations. One approach to handling them is to identify all possible interpretations of the ambiguous question (AQ) and to generate a long-form answer addressing them all, as suggested by Stelmakh et al., (2022). While it provides a comprehensive response without bothering the user for clarification, considering multiple dimensions of ambiguity and gathering corresponding knowledge remains a challenge. To cope with the challenge, we propose a novel framework, Tree of Clarifications (ToC): It recursively constructs a tree of disambiguations for the AQ -- via few-shot prompting leveraging external knowledge -- and uses it to generate a long-form answer. ToC outperforms existing baselines on ASQA in a few-shot setup across the metrics, while surpassing fully-supervised baselines trained on the whole training set in terms of Disambig-F1 and Disambig-ROUGE. Code is available at https://github.com/gankim/tree-of-clarifications.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, Tree of Clarifications (ToC), recursively constructs a tree of disambiguations for the AQ -- via few-shot prompting leveraging external knowledge -- and uses it to generate a long-form answer.'}",
-evaluation of prompts to simplify cardiovascular disease information using a large language model,Mbbs Vishala Mishra,"AI chatbots powered by large language models (LLMs) are emerging as an important source of public-facing medical information. Generative models hold promise for producing tailored guidance at scale, which could advance health literacy and mitigate well-known disparities in the accessibility of health-protective information. In this study, we highlight an important limitation of basic approaches to AI-powered text simplification: when given a zero-shot or one-shot simplification prompt, GPT-4 often responds by omitting critical details. To address this limitation, we developed a new prompting strategy, which we term rubric prompting. Rubric prompts involve a combination of a zero-shot simplification prompt with brief reminders about important topics to address. Using rubric prompts, we generate recommendations about cardiovascular disease prevention that are more complete, more readable, and have lower syntactic complexity than baseline responses produced without prompt engineering. This analysis provides a blueprint for rigorous evaluation of AI model outputs in medicine.","{'model': 'tldr@v2.0.0', 'text': 'Using rubric prompts, a new prompting strategy is developed that generates recommendations about cardiovascular disease prevention that are more complete, more readable, and have lower syntactic complexity than baseline responses produced without prompt engineering.'}",
-category-specific prompts for animal action recognition with pretrained vision-language models,Yinuo Jing,"Animal action recognition has a wide range of applications. However, the field largely remains unexplored due to the greater challenges compared to human action recognition, such as lack of annotated training data, large intra-class variation, and interference of cluttered background. Most of the existing methods directly apply human action recognition techniques, which essentially require a large amount of annotated data. In recent years, contrastive vision-language pretraining has demonstrated strong zero-shot generalization ability and has been used for human action recognition. Inspired by the success, we develop a highly performant action recognition framework based on the CLIP model. Our model addresses the above challenges via a novel category-specific prompting module to generate adaptive prompts for both text and video based on the animal category detected in input videos. On one hand, it can generate more precise and customized textual descriptions for each action and animal category pair, being helpful in the alignment of textual and visual space. On the other hand, it allows the model to focus on video features of the target animal in the video and reduce the interference of video background noise. Experimental results demonstrate that our method outperforms five previous action recognition methods on the Animal Kingdom dataset and has shown best generalization ability on unseen animals.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the method outperforms five previous action recognition methods on the Animal Kingdom dataset and has shown best generalization ability on unseen animals.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612551
-"slow academic travel: an antidote to ""fly over"" scholarship in the age of climate crisis",Meredith A. Conti,"All three animals faced away from me, a row of hairy posteriors blocking the rest of their hulking bodies from view. My iPhone poised in camera mode, I walked slowly to the left, keeping my eyes on the herd grazing eighty yards in the distance. They were in no hurry, and neither was I. Just as I had lined up the perfect shot—the lowered heads of the bison, tiny black eyes shining and mouths masticating in perpetual motion, were now visible—a frigid blast of wind whipped across the plain, rustling grasses, flapping my jacket, and prompting the bison to rotate in one choreographed movement to the right. Unlike me, they wanted to face into the wind. Laughing under my breath, I moved farther to the left until the bison were in profile and began snapping. I arrived back at my car the proud owner of a new nature photo series: “Unruffled Buffalo, Badlands National Park.” Individual behavior change is a necessary condition for making a positive societal transition to a resource-constrained future. Because we live on a planet with finite resources, behavior change interventions need to move beyond creating “green consumers” and instead foster and support “green citizens.” . . . There is an urgent need to promote agency and self-sufficiency with regards to individual conservation behaviors, while simultaneously promoting mindfulness of collective impacts on the planet. (73)",,
-method for prompting passenger density information in all carriages of subway to target station platform,叶尔肯.拜山,"The invention discloses a method for prompting passenger density information in all carriages of a subway to a target station platform. In the method, more than one camera or camera head, an image analysis module, a transmitting module and a display module are included, wherein the camera or the camera head is arranged in the carriages of the subway to acquire images of passengers in the carriages of the subway; the image analysis module is used for analyzing and calculating valid passenger density data shot by the camera or the camera head and outputting the valid passenger density data; the transmission module is used for outputting the acquired valid passenger density data; and the display module displays the acquired valid passenger density data. According to the method, by monitoring for the free space in the carriages of the subway, passengers waiting for the subway on the station platform know the passenger density in all the carriages of the subway which is about to enter the station and then make queues reasonably, so that the passenger transport efficiency of the subway is improved.","{'model': 'tldr@v2.0.0', 'text': 'By monitoring for the free space in the carriages of the subway, passengers waiting for the subway on the station platform know the passenger density in all the car carriage which is about to enter the station and then make queues reasonably, so that the passenger transport efficiency of the Subway is improved.'}",
-how large language models perform on the united states medical licensing examination: a systematic review,MD Dana Brin,"Purpose: The United States Medical Licensing Examination (USMLE) assesses physicians' competency and passing is a requirement to practice medicine in the U.S. With the emergence of large language models (LLMs) like ChatGPT and GPT-4, understanding their performance on these exams illuminates their potential in medical education and healthcare. Methods: A literature search following the 2020 PRISMA guidelines was conducted, focusing on studies using official USMLE questions and publicly available LLMs. Results: Three relevant studies were found, with GPT-4 showcasing the highest accuracy rates of 80-90% on the USMLE. Open-ended prompts typically outperformed multiple-choice ones, with 5-shot prompting slightly edging out zero-shot. Conclusion: LLMs, especially GPT-4, display proficiency in tackling USMLE-standard questions. While the USMLE is a structured evaluation tool, it may not fully capture the expansive capabilities and limitations of LLMs in medical scenarios. As AI integrates further into healthcare, ongoing assessments against trusted benchmarks are essential.","{'model': 'tldr@v2.0.0', 'text': 'While the USMLE is a structured evaluation tool, it may not fully capture the expansive capabilities and limitations of LLMs in medical scenarios, as AI integrates further into healthcare, ongoing assessments against trusted benchmarks are essential.'}",https://www.medrxiv.org/content/medrxiv/early/2023/09/07/2023.09.03.23294842.full.pdf
-resource constrained dialog policy learning via differentiable inductive logic programming,Zhenpeng Zhou,"Motivated by the needs of resource constrained dialog policy learning, we introduce dialog policy via differentiable inductive logic (DILOG). We explore the tasks of one-shot learning and zero-shot domain transfer with DILOG on SimDial and MultiWoZ. Using a single representative dialog from the restaurant domain, we train DILOG on the SimDial dataset and obtain 99+% in-domain test accuracy. We also show that the trained DILOG zero-shot transfers to all other domains with 99+% accuracy, proving the suitability of DILOG to slot-filling dialogs. We further extend our study to the MultiWoZ dataset achieving 90+% inform and success metrics. We also observe that these metrics are not capturing some of the shortcomings of DILOG in terms of false positives, prompting us to measure an auxiliary Action F1 score. We show that DILOG is 100x more data efficient than state-of-the-art neural approaches on MultiWoZ while achieving similar performance metrics. We conclude with a discussion on the strengths and weaknesses of DILOG.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces dialog policy via differentiable inductive logic (DILOG) and shows that DILOG is 100x more data efficient than state-of-the-art neural approaches on MultiWoZ while achieving similar performance metrics.'}",https://www.aclweb.org/anthology/2020.coling-main.597.pdf
-prestack suppression of high frequency ground roll using a 3d multiwindow kl filter: application to a legacy mississippi lime survey,S. Verma,"Summary Recently the Mississippi Lime has become one of the most active resource plays. Our study area falls in-between the Fort Worth and Midland Basins. The main production comes from high porosity tripolitic chert. Our objective is to use 3D seismic data to map the areal distribution of discontinuous tripolitic facies. In the early 1990s several 3D surveys were shot in the study area to image shallower objectives. With the advent of the Mississippi Lime play, four of these surveys were merged and reprocessed using careful statics and velocity analysis. Even after prestack time migration, the target zone is contaminated with the acquisition footprint. The data are low (~15) fold and contaminated by highly aliased, high frequency, high amplitude ground roll. Given the sparsity of the survey, modern f-kx-ky filters were not able to remove ground roll prompting the development of a new ground roll suppression workflow. In workflow, we first window and low-pass filter (f<50 Hz) the data, 3D patch by 3D patch. We then apply linear moveout to approximately flatten the ground roll phases, estimate the dip about this reference moveout, and compute coherence within a 3channel by 3-shot by 20 ms window for each sample. Using a Kuwahara algorithm, we choose the most coherent window within which we apply a structure-oriented KL filter. At the end we simply modeled the ground roll from the original data. This 3D filter preserves signal amplitude and is flexible enough to model the piece wise continuous ground roll pattern common with irregular topography.",,
-our chimp cousins get that much closer,A. Gibbons,"ARE CHIMPS MORE CLOSELY RELATED TO gorillas or to human beings? That seemingly innocent question provoked a war among evolutionary specialists that has lasted more than 20 years. Now one side has new ammunition-in the form of mitochondrial DNA studies showing that chimps have more in common with us than with gorillas. This data, coming from independent U.S. and Japanese teams, is consistent with a growing body of results based on nuclear DNA sequences, prompting one chimphuman partisan, Charles Sibley of San Francisco State University, to claim, ""We've won the war."" But don't assume a surrender treaty is about to be signed. The embattled minority who think chimps are closer to gorillas has by no means conceded defeat. ""I think the attempt to close this question is extremely premature,"" says Jonathan Marks, professor of anthropology at Yale, who questions the quality of almost all the DNA data now available-and points to a few studies that suggest a chimp-gorilla connection. The first shot in the war was fired in the early 60s, when Morris Goodman ofWayne State University undermined the conventional classification system that put humans in one family-Hominidae-and chimps and gorillas in a second family: Pongidae. Goodman's work, based on cross reactions between immunologic blood proteins, suggested humans are very closely related genetically to the two species of African apes and that all three belong in one family. Most evolutionists thought Goodman's data were shaky, and they stuck to the old system. But molecular biologists persevered. In 1984, Sibley and Jon Ahlquist, both then at Yale, published DNA hybridization data showing that not only were we close to the apes, but that chimps actually had more in common genetically with humans than with gorillas. DNA hybridization is based on seeing how much of one set of DNA fragments will ""hybridize,"" or bind to another, providing an indication ofhow many sequences they share. But it is tricky to do, tricky to interpret, and less precise than having the actual sequences. Partly as a result, the Sibley and Ahlquist results were subjected to a withering bombardment from Marks and Vincent Sarich of the University of California at Berkeley. Marks and Sarich questioned their methods of data analysis and even charged that Sibley and Ahlquist had falsified data. It wasn't long afterwards, however, that actual DNA sequences from nuclear genes began coming in that offered strong support for close ties between chimps and humans. Perhaps the best work came from Goodman and his col-",,
-safety and arming method for fuzes based on geomagnetic field signal,Hu Xuegong,"Based on analyzing the method of geomagnetism turns-counting and the mathematical model for testing distance, tow kinds of information are put forward as the second environment prompting of the redundant safety, one is the geomagnetism field information when the projectile exiting the muzzle, and the second is the spin signal of the projectile in the geomagnetic field. “Threshold level + time window” is used as the judgment criterion, and this can distinguish the shot environment and non-shot environment. The method of geomagnetism turns-counting for testing distance is used to realize the distance design of the muzzle safety, it is easy to solve the problem of insufficient muzzle safety distance, and it can distinguish the normal shot environment and improper shot environment. The geomagnetism curve is tested by dynamic shooting, and it is testified that the schemes is feasible.","{'model': 'tldr@v2.0.0', 'text': 'The method of geomagnetism turns-counting for testing distance is used to realize the distance design of the muzzle safety, it is easy to solve the problem of insufficient muzzle safety distance, and it can distinguish the normal shot environment and improper shot environment.'}",
-tune-a-video: one-shot tuning of image diffusion models for text-to-video generation,Jay Zhangjie Wu,"To replicate the success of text-to-image (T2I) generation, recent works employ large-scale video datasets to train a text-to-video (T2V) generator. Despite their promising results, such paradigm is computationally expensive. In this work, we propose a new T2V generation setting$\unicode{x2014}$One-Shot Video Tuning, where only one text-video pair is presented. Our model is built on state-of-the-art T2I diffusion models pre-trained on massive image data. We make two key observations: 1) T2I models can generate still images that represent verb terms; 2) extending T2I models to generate multiple images concurrently exhibits surprisingly good content consistency. To further learn continuous motion, we introduce Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy. At inference, we employ DDIM inversion to provide structure guidance for sampling. Extensive qualitative and numerical experiments demonstrate the remarkable ability of our method across various applications.","{'model': 'tldr@v2.0.0', 'text': 'A new T2V generation setting, where only one text-video pair is presented, and Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy, is introduced.'}",http://arxiv.org/pdf/2212.11565
-siamese neural networks for one-shot image recognition,Gregory R. Koch,"The process of learning good features for machine learning applications can be very computationally expensive and may prove difficult in cases where little data is available. A prototypical example of this is the one-shot learning setting, in which we must correctly make predictions given only a single example of each new class. In this paper, we explore a method for learning siamese neural networks which employ a unique structure to naturally rank similarity between inputs. Once a network has been tuned, we can then capitalize on powerful discriminative features to generalize the predictive power of the network not just to new data, but to entirely new classes from unknown distributions. Using a convolutional architecture, we are able to achieve strong results which exceed those of other deep learning models with near state-of-the-art performance on one-shot classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A method for learning siamese neural networks which employ a unique structure to naturally rank similarity between inputs and is able to achieve strong results which exceed those of other deep learning models with near state-of-the-art performance on one-shot classification tasks.'}",
-flow-guided one-shot talking face generation with a high-resolution audio-visual dataset,Zhimeng Zhang,"One-shot talking face generation should synthesize high visual quality facial videos with reasonable animations of expression and head pose, and just utilize arbitrary driving audio and arbitrary single face image as the source. Current works fail to generate over 256×256 resolution realistic-looking videos due to the lack of an appropriate high-resolution audio-visual dataset, and the limitation of the sparse facial landmarks in providing poor expression details. To synthesize high-definition videos, we build a large in-the-wild high-resolution audio-visual dataset and propose a novel flow-guided talking face generation framework. The new dataset is collected from youtube and consists of about 16 hours 720P or 1080P videos. We leverage the facial 3D morphable model (3DMM) to split the framework into two cascaded modules instead of learning a direct mapping from audio to video. In the first module, we propose a novel animation generator to produce the movements of mouth, eyebrow and head pose simultaneously. In the second module, we transform animation into dense flow to provide more expression details and carefully design a novel flow-guided video generator to synthesize videos. Our method is able to produce high-definition videos and outperforms state-of-the-art works in objective and subjective comparisons*.","{'model': 'tldr@v2.0.0', 'text': 'This work builds a large in-the-wild high-resolution audio-visual dataset and proposes a novel flow-guided talking face generation framework that is able to produce high-definition videos and outperforms state-of- the-art works in objective and subjective comparisons.'}",
-heterogeneity for the win: one-shot federated clustering,D. Dennis,"In this work, we explore the unique challenges -- and opportunities -- of unsupervised federated learning (FL). We develop and analyze a one-shot federated clustering scheme, $k$-FED, based on the widely-used Lloyd's method for $k$-means clustering. In contrast to many supervised problems, we show that the issue of statistical heterogeneity in federated networks can in fact benefit our analysis. We analyse $k$-FED under a center separation assumption and compare it to the best known requirements of its centralized counterpart. Our analysis shows that in heterogeneous regimes where the number of clusters per device $(k')$ is smaller than the total number of clusters over the network $k$, $(k'\le \sqrt{k})$, we can use heterogeneity to our advantage -- significantly weakening the cluster separation requirements for $k$-FED. From a practical viewpoint, $k$-FED also has many desirable properties: it requires only round of communication, can run asynchronously, and can handle partial participation or node/network failures. We motivate our analysis with experiments on common FL benchmarks, and highlight the practical utility of one-shot clustering through use-cases in personalized FL and device sampling.","{'model': 'tldr@v2.0.0', 'text': ""This work develops and analyzes a one-shot federated clustering scheme, $k$-FED, based on the widely-used Lloyd's method, and shows that in heterogeneous regimes where the number of clusters per device $(k')$ is smaller than the total number of clustering over the network, the issue of statistical heterogeneity in federated networks can in fact benefit the analysis.""}",
-one shot learning based drivers head movement identification using a millimetre wave radar sensor,Hong Nhung Nguyen,"Concentration of drivers on traffic is a vital safety issue; thus, monitoring a driver being on road becomes an essential requirement. The key purpose of supervision is to detect abnormal behaviours of the driver and promptly send warnings to him her for avoiding incidents related to traffic accidents. In this paper, to meet the requirement, based on radar sensors applications, the authors first use a small sized millimetre wave radar installed at the steering wheel of the vehicle to collect signals from different head movements of the driver. The received signals consist of the reflection patterns that change in response to the head movements of the driver. Then, in order to distinguish these different movements, a classifier based on the measured signal of the radar sensor is designed. However, since the collected data set is not large, in this paper, the authors propose One shot learning to classify four cases of driver's head movements. The experimental results indicate that the proposed method can classify the four types of cases according to the various head movements of the driver with a high accuracy reaching up to 100. In addition, the classification performance of the proposed method is significantly better than that of the convolutional neural network model.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results indicate that the proposed One shot learning to classify four types of cases according to the various head movements of the driver with a high accuracy reaching up to 100 and the classification performance is significantly better than that of the convolutional neural network model.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1049/rsn2.12223
-single path one-shot neural architecture search with uniform sampling,Zichao Guo,,"{'model': 'tldr@v2.0.0', 'text': 'A Single Path One-Shot model is proposed to construct a simplified supernet, where all architectures are single paths so that weight co-adaption problem is alleviated.'}",http://arxiv.org/pdf/1904.00420
-nas-bench-1shot1: benchmarking and dissecting one-shot neural architecture search,Arber Zela,"One-shot neural architecture search (NAS) has played a crucial role in making NAS methods computationally feasible in practice. Nevertheless, there is still a lack of understanding on how these weight-sharing algorithms exactly work due to the many factors controlling the dynamics of the process. In order to allow a scientific study of these components, we introduce a general framework for one-shot NAS that can be instantiated to many recently-introduced variants and introduce a general benchmarking framework that draws on the recent large-scale tabular benchmark NAS-Bench-101 for cheap anytime evaluations of one-shot NAS methods. To showcase the framework, we compare several state-of-the-art one-shot NAS methods, examine how sensitive they are to their hyperparameters and how they can be improved by tuning their hyperparameters, and compare their performance to that of blackbox optimizers for NAS-Bench-101.","{'model': 'tldr@v2.0.0', 'text': 'A general framework for one-shot NAS that can be instantiated to many recently-introduced variants and a general benchmarking framework that draws on the recent large-scale tabular benchmark NAS-Bench-101 for cheap anytime evaluations of one- shot NAS methods are introduced.'}",
-one‐shot device testing data,,,,
-greedynas: towards fast one-shot nas with greedy supernet,Shan You,"Training a supernet matters for one-shot neural architecture search (NAS) methods since it serves as a basic performance estimator for different architectures (paths). Current methods mainly hold the assumption that a supernet should give a reasonable ranking over all paths. They thus treat all paths equally, and spare much effort to train paths. However, it is harsh for a single supernet to evaluate accurately on such a huge-scale search space (e.g., 7^21). In this paper, instead of covering all paths, we ease the burden of supernet by encouraging it to focus more on evaluation of those potentially-good ones, which are identified using a surrogate portion of validation data. Concretely, during training, we propose a multi-path sampling strategy with rejection, and greedily filter the weak paths. The training efficiency is thus boosted since the training space has been greedily shrunk from all paths to those potentially-good ones. Moreover, we further adopt an exploration and exploitation policy by introducing an empirical candidate path pool. Our proposed method GreedyNAS is easy-to-follow, and experimental results on ImageNet dataset indicate that it can achieve better Top-1 accuracy under same search space and FLOPs or latency level, but with only ~60% of supernet training cost. By searching on a larger space, our GreedyNAS can also obtain new state-of-the-art architectures.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-path sampling strategy with rejection, and greedily filter the weak paths to ease the burden of supernet by encouraging it to focus more on evaluation of those potentially-good ones, which are identified using a surrogate portion of validation data.'}",https://arxiv.org/pdf/2003.11236
-distilled one-shot federated learning,Yanlin Zhou,"Current federated learning algorithms take tens of communication rounds transmitting unwieldy model weights under ideal circumstances and hundreds when data is poorly distributed. Inspired by recent work on dataset distillation and distributed one-shot learning, we propose Distilled One-Shot Federated Learning, which reduces the number of communication rounds required to train a performant model to only one. Each client distills their private dataset and sends the synthetic data (e.g. images or sentences) to the server. The distilled data look like noise and become useless after model fitting. We empirically show that, in only one round of communication, our method can achieve 96% test accuracy on federated MNIST with LeNet (centralized 99%), 81% on federated IMDB with a customized CNN (centralized 86%), and 84% on federated TREC-6 with a Bi-LSTM (centralized 89%). Using only a few rounds, DOSFL can match the centralized baseline on all three tasks. By evading the need for model-wise updates (i.e., weights, gradients, loss, etc.), the total communication cost of DOSFL is reduced by over an order of magnitude. We believe that DOSFL represents a new direction orthogonal to previous work, towards weight-less and gradient-less federated learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Distilled One-Shot Federated Learning, which reduces the number of communication rounds required to train a performant model to only one, and represents a new direction orthogonal to previous work, towards weight-less and gradient-less federated learning.'}",
-adversarial style mining for one-shot unsupervised domain adaptation,Yawei Luo,"We aim at the problem named One-Shot Unsupervised Domain Adaptation. Unlike traditional Unsupervised Domain Adaptation, it assumes that only one unlabeled target sample can be available when learning to adapt. This setting is realistic but more challenging, in which conventional adaptation approaches are prone to failure due to the scarce of unlabeled target data. To this end, we propose a novel Adversarial Style Mining approach, which combines the style transfer module and task-specific module into an adversarial manner. Specifically, the style transfer module iteratively searches for harder stylized images around the one-shot target sample according to the current learning state, leading the task model to explore the potential styles that are difficult to solve in the almost unseen target domain, thus boosting the adaptation performance in a data-scarce scenario. The adversarial learning framework makes the style transfer module and task-specific module benefit each other during the competition. Extensive experiments on both cross-domain classification and segmentation benchmarks verify that ASM achieves state-of-the-art adaptation performance under the challenging one-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'A novel Adversarial Style Mining approach, which combines the style transfer module and task-specific module into an adversarial manner, which achieves state-of-the-art adaptation performance under the challenging one-shot setting.'}",
-overcoming multi-model forgetting in one-shot nas with diversity maximization,Miao Zhang,"One-Shot Neural Architecture Search (NAS) significantly improves the computational efficiency through weight sharing. However, this approach also introduces multi-model forgetting during the supernet training (architecture search phase), where the performance of previous architectures degrade when sequentially training new architectures with partially-shared weights. To overcome such catastrophic forgetting, the state-of-the-art method assumes that the shared weights are optimal when jointly optimizing a posterior probability. However, this strict assumption is not necessarily held for One-Shot NAS in practice. In this paper, we formulate the supernet training in the One-Shot NAS as a constrained optimization problem of continual learning that the learning of current architecture should not degrade the performance of previous architectures during the supernet training. We propose a Novelty Search based Architecture Selection (\textbf{NSAS}) loss function and demonstrate that the posterior probability could be calculated without the strict assumption when maximizing the diversity of the selected constraints. A greedy novelty search method is devised to find the most representative subset to regularize the supernet training. We apply our proposed approach to two One-Shot NAS baselines, random sampling NAS (RandomNAS) and gradient-based sampling NAS (GDAS). Extensive experiments demonstrate that our method enhances the predictive ability of the supernet in One-Shot NAS and achieves remarkable performance on CIFAR-10, CIFAR-100, and PTB with efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Novelty Search based Architecture Selection (\\textbf{NSAS}) loss function and demonstrates that the posterior probability could be calculated without the strict assumption when maximizing the diversity of the selected constraints.'}",https://research-repository.griffith.edu.au/bitstream/10072/419622/2/Pan2923687-Accepted.pdf
-understanding mathematical concept: the effect of savi learning model with probing-prompting techniques viewed from self-concept,R. Widyastuti,"This research aims to discover the effect of SAVI learning models with the Probing-prompting technique in terms of self-concepts on students’ mathematical conceptual understanding. The method used in this research was the quasi-experimental design. The experimental design used in this study was the post-test only design with a 3x3 factorial research design. The population in this study were 153 students. The sampling technique used was random. The normality test, homogeneity test, two way ANOVA test and Scheffe’ method were used as data analysis techniques. The results showed that there is an influence of the SAVI learning model with probing-prompting techniques in terms of students’ self-concepts on the ability to understand mathematical concepts compared to conventional learning models applied by teachers at school.",,
-ability of children to perform touchscreen gestures and follow prompting techniques when using mobile apps,Savita Yadav,"Background Children today get access to smartphones at an early age. However, their ability to use mobile apps has not yet been studied in detail. Purpose This study aimed to assess the ability of children aged 2–8 years to perform touchscreen gestures and follow prompting techniques, i.e., ways apps provide instructions on how to use them. Methods We developed one mobile app to test the ability of children to perform various touchscreen gestures and another mobile app to test their ability to follow various prompting techniques. We used these apps in this study of 90 children in a kindergarten and a primary school in New Delhi in July 2019. We noted the touchscreen gestures that the children could perform and the most sophisticated prompting technique that they could follow. Results Two- and 3-year-old children could not follow any prompting technique and only a minority (27%) could tap the touchscreen at an intended place. Four- to 6-year-old children could perform simple gestures like a tap and slide (57%) and follow instructions provided through animation (63%). Seven- and 8-year-old children could perform more sophisticated gestures like dragging and dropping (30%) and follow instructions provided in audio and video formats (34%). We observed a significant difference between the number of touchscreen gestures that the children could perform and the number of prompting techniques that they could follow (F=544.0407, P<0.05). No significant difference was observed in the performance of female versus male children (P>0.05). Conclusion Children gradually learn to use mobile apps beginning at 2 years of age. They become comfortable performing single-finger gestures and following nontextual prompting techniques by 8 years of age. We recommend that these results be considered in the development of mobile apps for children.","{'model': 'tldr@v2.0.0', 'text': 'Children gradually learn to use mobile apps beginning at 2 years of age and become comfortable performing single-finger gestures and following nontextual prompting techniques by 8 years ofAge, which is recommended in the development of mobile apps for children.'}",https://www.e-cep.org/upload/pdf/cep-2019-00997.pdf
-improving critical thinking skills through module solubility and solubility results based on discovery learning with probing prompting techniques at sman 1 pariaman,Fefri Perrianty,"Abstract - This study aims to determine the improvement of student critical thinking based on discovery learning using modules that help investigate questions that encourage questions about solubility and solubility results. The research design used was the Non-Equivalent Post-test Group Design. The population in this study were students of class XI IPA SMAN 1 Pariaman academic year 2018/2019. The selection of sample classes is done by cluster purposive sampling technique. The sample in this study was class XI IPA 1 as the control class and XI IPA 2 as the experimental class. The experimental learning class uses modules based on discovery learning withtechniques probing prompting, while the control class without using modules. The research instrument used was a test of the results of learning critical thinking questions. The results of the study show that module based discovery learning withtechniques probing promptingcan improve students' critical thinking skills, as evidenced by the average increase in posttest results. The results of testing for normality and homogeneity state that the critical thinking value of the two distributed samples is normal and homogeneous. Hypothesis test results obtained by GIS. (2-tailed) of 0.016 <0.05 means that the value of critical thinking learners who use modules is based on discovery learning bytechniques probing prompting differ significantly",,
-learning models using probing prompting techniques assisted by blended learning edmodo to improve mathematical understanding ability,R. Rahmawati,"This study aims to see an increase in the ability of mathematical understanding of students using probing learning models prompting assisted by edmodo blended learning. In this model the researcher presents a series of questions that are digging and guiding so that a thinking process that links each student's knowledge and experience with new knowledge is occurring. learned. Research subject of 3rd Semester Student in Mathematics and Teacher Training Faculty, University of PGRI Palembang. This research method is Quasi-Experimental. The data of this study are of two types. Quantitative data was obtained through analysis of student answers on the KPM test which was carried out pretest and posttest, and qualitative data were obtained from observations of lecturer and student activities, and the results of interviews with students. Data analysis used: doing descriptive analysis of data and calculating normalized gain (normalized gain) pretest and posttest, Testing requirements analysis is the data normality test and variance homogeneity test of the whole quantitative data, and hypothesis testing in this study t test, t-test, Mann-Whitney U test, one-way ANAVA test or Kruskal-Wallis test. Hypothesis testing using SPSS-20 for Windows. The results of the data analysis concluded that 1) the increase in the ability of mathematical understanding of students who get learning by probing prompting assisted by edmodo blended learning is better than students who get conventional learning in terms of (a). all students (b) .KAM students",,https://ojs.unm.ac.id/JDM/article/download/8843/5139
-deeplyrics: gpt2 for lyrics generation with finetuning and prompting techniques,Li Tian,"Lyrics are incredibly important part of a song’s success. The fast advancing natural language processing (NLP) methods have been successful in lyrics generation, making AI-assisting lyrics creation possible. However, as these methods get more resource-consuming and data-demanding, we identify a neglected research area of exploring efficient ways of model learning to simplify existing lyrics generation methods. Our project proposes DeepLyrics, a GPT-2 model using tuning-free prompting (in-context learning) on lyrics of highly successful songs in the past several decades to assist creative generation. Our tuning-free method, DeepLyrics, is able to achieve comparable and even better performance compared to thoroughly fine-tuned lyrics generation models. Our work shows the practicability of reducing large amount of training and finetuning work in existing lyrics generation methods with a well-engineered prompting technique. 1 Key Information to include • Mentor:NA • External Collaborators: Ying Lin @ Stanford Culture Lab (data preparation only) • Sharing project: NA","{'model': 'tldr@v2.0.0', 'text': 'This work shows the practicability of reducing large amount of training and finetuning work in existing lyrics generation methods with a well-engineered prompting technique, and achieves comparable and even better performance compared to thoroughly fine-tuned lyrics generation models.'}",
-"the effect of poe learning models (prediction, observation, and explanation) with probing-prompting techniques on the student’s cognitive learning outcomes of sma muhammadiyah 3 jember",Dewi Farida,"Student learning outcomes are competencies that students have after gaining their learning experience, so indicators of learning success can be seen from student learning outcomes after experiencing the learning process. Student learning outcomes can be improved through the use of learning models, one of which is by applying the POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques. The purpose of this study was to determine the effect of POE learning models (Prediction, Observation and Explanation) with probing-prompting techniques on the cognitive learning outcomes of the tenth grade students at SMA Muhammadiyah 3 Jember on the environmental pollution subject. Type of research is a quasy experimental study using a pretest and posttest design. In the experimental class, learning activity was carried out by applying the POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques and the control class, learning activities are carried out by applying conventional learning models. The data of the students' cognitive learning outcomes were obtained through the pretest and posttest scores then analyzed using ANAKOVA test. The application of POE learning model (Prediction, Observation and Explanation) with probing-prompting techniques had a significant effect with a probability as much as 0,000. The affective learning outcomes influenced significantly with a significance value of 0.001 or p<0.05.",,http://pancaran.kreatindo.com/index.php/pancaran/article/download/199/217
-prompting techniques for guidance and action assistance using augmented-reality smart-glasses,Patrick Renner,"In the context of picking and assembly tasks, assistance systems based on Augmented Reality (AR) can help finding target objects and conducting correct actions. The aim is to develop guiding and action assistance techniques for smart glasses, which are easily understandable not only for workers, but also for impaired and elderly people.","{'model': 'tldr@v2.0.0', 'text': 'The aim is to develop guiding and action assistance techniques for smart glasses, which are easily understandable not only for workers, but also for impaired and elderly people.'}",
-effects of video prompting techniques on teaching daily living skills to children with autism spectrum disorders,Sarah C. Domire,"Video-based instruction is becoming a common intervention in today’s classrooms. Previous research has focused primarily on video modeling techniques that required the student to watch an entire video of the task before attempting to complete the task independently. Video prompting is a form of video instruction that breaks down target skills into steps that are then performed directly after viewing each clip. The present review examined studies using video prompting techniques to teach functional and daily living skills to individuals with autism spectrum disorders (ASD). The focus of the review was on evaluation of the effectiveness of video prompting and the factors that affect student attention to the video, retention of target behavior, production of target behavior, and motivation. Results showed that video prompting was an effective intervention for teaching a wide array of target skills and that students with ASD were able to generalize and maintain the acquired skills. Video prompting was also shown to be more effective than both static picture prompts and video modeling techniques in terms of percentage of correct independent responding. Suggestions for practice and future research are discussed.",,
-comparison of prompting techniques to teach children with autism to ask questions in the context of a conversation,Matthew G. Swerdan,"An adapted alternating treatments design was used to compare the efficacy of echoic and textual prompts to teach three students with autism (ages 8–15) to ask questions related to two pre-selected topics of conversation. Participants were first required to answer questions related to the topics to determine whether accurate responses were within their repertoire. This was followed by a transfer of stimulus control procedure to teach participants to ask relevant questions to the experimenter on the same topics. Probes with a novel conversation topic were conducted in the natural environment with a peer, and follow-up probes were conducted after training. Results indicate questions taught using an echoic prompt were acquired in fewer trials to criterion than questions taught using a textual prompt for two of the three participants. Limitations and implications for future research will be discussed.",,
-the difference between mathematical reasoning ability improvement by learning with meta cognitive approach aided probing and prompting techniques in smp negeri 4 seisuka,Nadran Hamdani Siregar,"The purpose of this study were: (1) analyzed the differences in students’ mathematical reasoning ability improvement taught by metacognition approach aided probing technique (PMT-probing) and metacognition approach aided prompting technique (PMT-prompting); and (2) described the process of the students’ responses in solving mathematical reasoning abilities. This study was a quasi experimentalresearch. The population in this study were all students of class VIII SMP Negeri 4 SeiSuka, with a purposive sampling techniques, the obtained sample was VIII-1 and VIII-2. The research instrument used a test of mathematical reasoning ability, and had qualified the criteria of content validity, and reliability coefficient of 0.819. Anova two ways was used to analyze the difference of mathematical reasoning ability improvement, while descriptive analysis was used to analyze students’ answers process. The results showed that: (1) There were differences in students’ mathematical reasoning skills improvement which were taught by metacognition approach aided probing techniques and the students taught by prompting technical approach; and (2) The process of the students’ responses on students’ mathematical reasoning abilitythrough learning with metacognition approach aided by prompting techniques was better than metacognition approach aided by probing techniques.",,http://www.scholink.org/ojs/index.php/wjer/article/download/766/781
-effects of teacher prompting techniques on the writing performance of fourth and fifth grade students,Mindy Allenger,,,
-effect of prompting techniques and learning styles on requirements elicitation,Bonnie Lynn Jan,,,
-"a survey of graph prompting methods: techniques, applications, and challenges",Xuansheng Wu,"The recent""pre-train, prompt, predict training""paradigm has gained popularity as a way to learn generalizable models with limited labeled data. The approach involves using a pre-trained model and a prompting function that applies a template to input samples, adding indicative context and reformulating target tasks as the pre-training task. However, the design of prompts could be a challenging and time-consuming process in complex tasks. The limitation can be addressed by using graph data, as graphs serve as structured knowledge repositories by explicitly modeling the interaction between entities. In this survey, we review prompting methods from the graph perspective, where prompting functions are augmented with graph knowledge. In particular, we introduce the basic concepts of graph prompt learning, organize the existing work of designing graph prompting functions, and describe their applications and future challenges. This survey will bridge the gap between graphs and prompt design to facilitate future methodology development.","{'model': 'tldr@v2.0.0', 'text': 'This survey will bridge the gap between graphs and prompt design to facilitate future methodology development and introduce the basic concepts of graph prompt learning, organize the existing work of designing graph prompting functions, and describe their applications and future challenges.'}",http://arxiv.org/pdf/2303.07275
-prompting techniques – researching subjectivities in educational history.,Helle Bjerg,,,
-help me think: a simple prompting strategy for non-experts to create customized content with models,Swaroop Mishra,"Controlling the text generated by language models and customizing the content has been a long-standing challenge. Existing prompting techniques proposed in pursuit of providing control are task-specific and lack generality; this provides overwhelming choices for non-expert users to find a suitable method for their task. The effort associated with those techniques, such as in writing examples, explanations, instructions, etc. further limits their adoption among non-expert users. In this paper, we propose a simple prompting strategy HELP ME THINK where we encourage GPT3 to help non-expert users by asking a set of relevant questions and leveraging user answers to execute the task. We demonstrate the efficacy of our technique HELP ME THINK on a variety of tasks. Specifically, we focus on tasks that are hard for average humans and require significant thinking to perform. We hope our work will encourage the development of unconventional ways to harness the power of large language models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple prompting strategy HELP ME THINK where GPT3 is encouraged to help non-expert users by asking a set of relevant questions and leveraging user answers to execute the task.'}",http://arxiv.org/pdf/2208.08232
-an empirical investigation of user requirements elicitation: comparing the effectiveness of prompting techniques,Glenn J. Browne,"Eliciting requirements from users and other stakeholders is of central importance to information systems development. Despite this importance, surprisingly little research has measured the effectiveness of various requirements elicitation techniques. The present research first discusses theory relevant to information requirements determination in general and elicitation in particular.We then develop a model of the requirements elicitation process. This model and its underlying theory were then used to construct a new requirements elicitation prompting technique. To provide a context for testing the relative effectiveness of the new technique, two other questioning methodologies were also operationalized as prompting techniques: (1) the interrogatories technique, which involves asking ""who,"" ""what,"" ""when,"" ""where,"" ""how,"" and ""why""; questions; and (2) a semantic questioning scheme, which involves asking questions based on a theoretical model of knowledge structures. To measure the usefulness of the prompting techniques in eliciting requirements, a set of generic requirements categories was adapted from previous research to capture requirements evoked by users. The effectiveness of the three methods in eliciting requirements for a software application was then tested in an experiment with users. Results showed that the new prompting technique elicited a greater quantity of requirements from users than did the other two techniques. Implications of the findings for research and systems analysis practice are discussed.","{'model': 'tldr@v2.0.0', 'text': None}",
-evaluation of computer-aided detection (cad) prompting techniques for mammography.,S. Astley,"Computer-aided detection (CAD) systems, in which abnormalities are automatically detected and their locations presented to the radiologist as prompts, are increasingly being used to improve reader performance. The performance of CAD systems can be evaluated in two ways: by measuring the performance of the algorithms, or by monitoring the performance of readers using the system. All aspects of evaluation need careful consideration to avoid potential bias. This paper examines a variety of different approaches to evaluation and discusses their relative strengths and weaknesses.","{'model': 'tldr@v2.0.0', 'text': 'A variety of different approaches to evaluation of CAD systems are examined and their relative strengths and weaknesses are discussed.'}",
-prompting techniques to increase the return rate of mailed questionnaires1,R. Winett,"To increase the return rate of questionnaires mailed to clergy and physicians concerning their mental-health practices, different prompts were used after the questionnaire was received during four mail-outs to four randomly drawn samples of clergy and physicians. For each mail-out, the sample was divided into experimental (received prompt) and comparison (no prompt) groups, and one type of prompt or combination was used. Non-returnees of the questionnaire in the experimental group received either: (a) a single telephone call, (b) a memo, (c) a package (personal letter and new questionnaire) or package plus a telephone call, or (d) a double call. Comparison physicians and clergy were mailed only the original questionnaire. Relative to their respective comparison group's return rate, which averaged 22% across the four mail-outs (range 18% to 24%), the single call and package alone about doubled the overall return rate, the package and call increased the return rate about two-and-a-half fold, and the double call almost tripled the return rate. The memo was ineffective. A cost-effectiveness analysis indicated that the double-call procedure was less expensive than the single call, and much less expensive than the package alone or package with a call in securing returns. An analysis of the pattern of returns showed clearly that when prompts were not delivered (comparison groups), very few returns were received after about seven days from the initial mail-out. Most returns from prompts (experimental groups) were received by several days after the prompt. The results were seen as salient to the problem of reducing selection or volunteer bias in questionnaire studies and subsequent research demonstrating the effectiveness of telephone calls made about a week after distribution of surveys in securing high return rates was discussed.","{'model': 'tldr@v2.0.0', 'text': 'The results were seen as salient to the problem of reducing selection or volunteer bias in questionnaire studies and subsequent research demonstrating the effectiveness of telephone calls made about a week after distribution of surveys in securing high return rates was discussed.'}",https://europepmc.org/articles/pmc1311322?pdf=render
-neuro-symbolic causal language planning with commonsense prompting,Yujie Lu,"Language planning aims to implement complex high-level goals by decomposition into sequential simpler low-level steps. Such procedural reasoning ability is essential for applications such as household robots and virtual assistants. Al-though language planning is a basic skill set for humans in daily life, it remains a challenge for large language models (LLMs) that lack deep-level commonsense knowledge in the real world. Previous methods require either manual exemplars or annotated programs to acquire such ability from LLMs. In contrast, this paper proposes Neuro-Symbolic C ausal LA nguage P lanner (CLAP) that elicits procedural knowledge from the LLMs with commonsense-infused prompting. Pre-trained knowledge in LLMs is essentially an unobserved confounder that causes spurious correlations between tasks and action plans. Through the lens of a Structural Causal Model (SCM), we propose an effective strategy in CLAP to construct prompts as a causal intervention toward our SCM. Using graph sampling techniques and symbolic program executors, our strategy formalizes the structured causal prompts from commonsense knowledge bases. CLAP obtains state-of-the-art performance on WikiHow and RobotHow, achieving a relative improvement of 5 . 28% in human evaluations under the counterfactual setting. This indicates the superiority of CLAP in causal language planning semantically and sequentially. as household robots and virtual assistants. We propose a Neuro-Symbolic Causal Language Planner (CLAP) with commonsense prompts elicited from the commonsense knowledge base to solve the language planning problem in a zero-shot manner. Experiments show the effectiveness of our proposed CLAP on both automatic and human evaluation results. Extending causal language planning to handle the long-horizon composite tasks and provide effective automatic evaluation metrics are important directions for future work.","{'model': 'tldr@v2.0.0', 'text': 'A Neuro-Symbolic Causal Language Planner (CLAP) is proposed that elicits procedural knowledge from the LLMs with commonsense-infused prompting to solve the language planning problem in a zero-shot manner.'}",http://arxiv.org/pdf/2206.02928
-generative speech recognition error correction with large language models and task-activating prompting,Chao-Han Huck Yang,"We explore the ability of large language models (LLMs) to act as speech recognition post-processors that perform rescoring and error correction. Our first focus is on instruction prompting to let LLMs perform these task without fine-tuning, for which we evaluate different prompting schemes, both zero- and few-shot in-context learning, and a novel task activation prompting method that combines causal instructions and demonstration to increase its context windows. Next, we show that rescoring only by in-context learning with frozen LLMs achieves results that are competitive with rescoring by domain-tuned LMs, using a pretrained first-pass recognition system and rescoring output on two out-of-domain tasks (ATIS and WSJ). By combining prompting techniques with fine-tuning we achieve error rates below the N-best oracle level, showcasing the generalization power of the LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the ability of large language models to act as speech recognition post-processors that perform rescoring and error correction, and shows that rescoring only by in-context learning with frozen LLMs achieves results that are competitive with rescoring by domain-tuned LMs.'}",https://arxiv.org/pdf/2309.15649
-llm-rec: personalized recommendation via prompting large language models,Hanjia Lyu,"We investigate various prompting strategies for enhancing personalized recommendation performance with large language models (LLMs) through input augmentation. Our proposed approach, termed LLM-Rec, encompasses four distinct prompting strategies: (1) basic prompting, (2) recommendation-driven prompting, (3) engagement-guided prompting, and (4) recommendation-driven + engagement-guided prompting. Our empirical experiments show that incorporating the augmented input text generated by LLM leads to improved recommendation performance. Recommendation-driven and engagement-guided prompting strategies are found to elicit LLM's understanding of global and local item characteristics. This finding highlights the importance of leveraging diverse prompts and input augmentation techniques to enhance the recommendation capabilities with LLMs.","{'model': 'tldr@v2.0.0', 'text': ""Recommendation-driven and engagement-guided prompting strategies are found to elicit LLM's understanding of global and local item characteristics and highlight the importance of leveraging diverse prompts and input augmentation techniques to enhance the recommendation capabilities with LLMs.""}",https://arxiv.org/pdf/2307.15780
-coaching and prompting for remote physical activity promotion: study protocol of a three-arm randomized controlled trial (movingcall),X. Fischer,"Background. Physical inactivity is currently seen as one of the biggest global public health issue. Remote physical activity (PA) promotion programs are expected to be effective if they are individually tailored and include behavior change techniques, personal coaching, and regular prompting. However, it is still not fully understood which intervention components are most effective. This paper describes the rationale and design of a study on an individually tailored remote PA promotion program comparing the efficacy of coaching and prompting with a single written advice. Methods. In total, 288 adults (age 20 to 65 years) were randomly assigned to three different intervention arms of a 6-month-long PA promotion program. A minimal intervention group received a single written PA recommendation. The two remaining groups either received telephone coaching sessions (n = 12 calls) with or without additional short message service (SMS) prompting (n = 48 SMSs for each participant). Data assessment took place at baseline, at the end of the intervention, and after a six-month follow-up-period. The primary outcome of the study was self-reported PA. Objectively assessed PA, psychosocial determinants of PA, well-being, body mass index (BMI), and adherence were assessed as secondary outcomes. Conclusion. Findings of this three-arm study will provide insight into the short and long-term effects of coaching and prompting for PA promotion.","{'model': 'tldr@v2.0.0', 'text': 'The rationale and design of a study on an individually tailored remote PA promotion program comparing the efficacy of coaching and prompting with a single written advice is described and insight into the short and long-term effects of coach and prompting for PA promotion is provided.'}",https://www.mdpi.com/1660-4601/16/3/331/pdf?version=1548412965
-validitas dan praktikalitas modul hukum-hukum dasar kimia berbasis pendekatan saintifik dengan menerapkan teknik probing prompting untuk siswa kelas x sma/ma,Yuni Purnama Sari,"The Implementation of the curriculum is carried out by practicing process skills that are reflected in learning activities. The process skills applied in the form of 5M activities are known as the scientific approach. This research aims to produce of modul of basic chemical laws based on a scientific approach by applying probing prompting techniques. The research instrument used was a questionnaire of validity and practicality. The module was validated by 5 validators while the practicality test was carried out by 2 chemistry teachers and 20 XI grade students of MIA 3 SMAN 1 Basa Ampek Balai Tapan, Pesisir Selatan. Data from the validity and practicality test results were analyzed using the cohen kappa formula. Based on the results of the study, it was found that the average kappa moment of validity test was 0,89 with a very high validity category and the average kappa moment of teacher practicality was 0,90 and the average kappa moment of students practicality was 0,91 with a very high practicality category. Thus, it was concluded that basic chemical laws module based on a scientific approach by applying probing prompting techniques was valid and practical for high school students.",,http://edukimia.ppj.unp.ac.id/ojs/index.php/edukimia/article/download/22/20
-which behaviour change techniques are most effective at increasing older adults’ self-efficacy and physical activity behaviour? a systematic review,D. French,,"{'model': 'tldr@v2.0.0', 'text': 'Many commonly used self-regulation intervention techniques that are effective for younger adults may not be effective for older adults.'}",https://openaccess.city.ac.uk/id/eprint/4248/1/Which_behavior_change_techniques_are_most_effective_at_increasing_older_adults%E2%80%99_selfefficacy.pdf
-prompting conceptual understanding with computer-mediated peer discourse and knowledge acquisition techniques,Chen-Chung Liu,"Numerous computer-mediated communication (CMC) tools have been designed to facilitate group discourses on the Web. However, previous studies noted that participants did not value online conferencing as a method for conducting in-depth discussions, and instead considered this method as merely scratching the surface of the issues involved. Therefore, the teachers must moderate the online discourse to add value to the success of CMC. This study presents a methodology that uses knowledge acquisition techniques during online discourses to increase student conceptual understanding and evaluation skills. The methodology explains the conceptual frameworks of individuals and the conceptual relationships between student peers. Active database triggers then are applied to automatically monitor the conceptual relationships and to moderate online discourses between student peers. Student reactions demonstrated that discourses mediated by the repertory grid analysis technique helped group members not only promote common conceptual understandings but also to approach the level of conceptual understanding of the teacher.","{'model': 'tldr@v2.0.0', 'text': 'A methodology that uses knowledge acquisition techniques during online discourses to increase student conceptual understanding and evaluation skills is presented and demonstrated that discourses mediated by the repertory grid analysis technique helped group members not only promote common conceptual understandings but also to approach the level of conceptual understanding of the teacher.'}",
-pembelajaran menggunakan teknik probing prompting berbantuan edmodo blended learning pada materi persamaan diferensial matematis mahasiswa di universitas pgri palembang,Jayanti Jayanti,"Abstract — This study aims to examine the extent of learning using probing prompting techniques applied in the course of differential equations at PGRI University Palembang assisted Edmodo Blended Learning. Learning model using Probing Prompting technique, aided by edmodo blended learning that is by using internet online, to assist them in learning. Probing Prompting Techniques is a way of presenting learning with a series of questions that are guided and explored to form a thought process that links previous knowledge and experience with new knowledge to be learned. Questions raised on the students encourage students to think more rationally about the knowledge that has been obtained before, and also can link the next questions so that new knowledge arises that is reflected from the results of a lesson. This research uses derskriftif method which only describes process from learning of differential equation using probing prompting. The subject of this research is a mathematics student in the course of Differential Equation at University PGRI of Palembang. Keywords — Probing Prompting, Edmodo Blended learning, and Differential Equation (PD) Abstrak — Penelitian ini bertujuan untuk mengkaji sejauhmana pembelajaran menggunakan teknik probing prompting yang diterapkan di matakuliah persamaan diferensial pada universitas PGRI Palembang berbantuan Edmodo Blended Learning. Model Pembelajaran dengan menggunakan teknik Probing Prompting, berbantuan edmodo blended learning yaitu dengan menggunakan internet secara online, untuk membantu mereka dalam pembelajaran. Teknik Probing Prompting adalah Cara menyajikan Pembelajaran dengan serangkaian pertanyaan yang sifatnya menuntun dan menggali sehingga membentuk suatu proses berpikir yang mengaitkan pengetahuan dan pengalaman sebelumnya dengan pengetahuan baru yang akan dipelajari. Pertanyaan-pertanyaan yang dilontarkan pada siswa mendorong siswa agar dapat berpikir lebih rasional tentang pengetahuan yang telah di dapat oleh sebelumnya, dan juga dapat mengaitkan pertanyaan-pertanyaan berikutnya sehingga timbul pengetahuan baru yang tercermin dari hasil suatu pembelajaran . Penelitian ini menggunakan metode derskriftif dimana cuma menggambarkan proses dari pembelajaran persamaan diferensial menggunakan probing prompting . Subyek penelitian ini adalah mahasiswa matematika pada mata kuliah Persamaan Diferensial pada Universitas PGRI Palembang. Kata Kunci — Probing Prompting, Edmodo Blended learning, dan Persamaan Diferensial (PD)",,
-comparing self-directed video prompting to least-to-most prompting in post-secondary students with moderate intellectual disabilities,Helen I. Cannella-Malone,"Objectives: This study compared the effects of self-directed video prompting with error correction using an iPod Touch to least-to-most prompting on teaching vocational skills to two post-secondary students with moderate intellectual disabilities. Methods: Students were first taught to use the inPromptu mobile application on the iPod Touch. They were then taught to collate and pack a large envelope and to tri-fold papers and pack a business envelope. The two tasks were counterbalanced across students in an adapted alternating treatments design. Students were instructed to use self-directed video prompting with error correction to teach themselves one skill, and an instructor used least-to-most prompting to teach the second skill. Data were also collected on sessions to criterion and maintenance of the skills over time. Results: Results demonstrated that both instructional techniques were effective in teaching the skills, and students generalized their use of the technology to learn a new skill with no additional instruction. These results were replicated with one of the students. Conclusions: Self-directed video prompting with error correction and least-to-most prompting were nearly equivalent in terms of acquisition efficiency and maintenance. These findings suggest that self-directed video prompting with error correction could be used in vocational settings, allowing job coaches to fade more quickly and for adults with intellectual and developmental disabilities to be more independent.",,
-computer-delivered interventions for reducing alcohol consumption: meta-analysis and meta-regression using behaviour change techniques and theory,Nicola Black,"The current aim was to examine the effectiveness of behaviour change techniques (BCTs), theory and other characteristics in increasing the effectiveness of computer-delivered interventions (CDIs) to reduce alcohol consumption. Included were randomised studies with a primary aim of reducing alcohol consumption, which compared self-directed CDIs to assessment-only control groups. CDIs were coded for the use of 42 BCTs from an alcohol-specific taxonomy, the use of theory according to a theory coding scheme and general characteristics such as length of the CDI. Effectiveness of CDIs was assessed using random-effects meta-analysis and the association between the moderators and effect size was assessed using univariate and multivariate meta-regression. Ninety-three CDIs were included in at least one analysis and produced small, significant effects on five outcomes (d+ = 0.07–0.15). Larger effects occurred with some personal contact, provision of normative information or feedback on performance, prompting commitment or goal review, the social norms approach and in samples with more women. Smaller effects occurred when information on the consequences of alcohol consumption was provided. These findings can be used to inform both intervention- and theory-development. Intervention developers should focus on, including specific, effective techniques, rather than many techniques or more-elaborate approaches.","{'model': 'tldr@v2.0.0', 'text': 'Examination of behaviour change techniques, theory and other characteristics in increasing the effectiveness of computer-delivered interventions to reduce alcohol consumption found intervention developers should focus on specific, effective techniques, rather than many techniques or more-elaborate approaches.'}",
-touchscreen prompts for preschoolers: designing developmentally appropriate techniques for teaching young children to perform gestures,Alexis Hiniker,"Though toddlers and preschoolers are regular touchscreen users, relatively little is known about how they learn to perform unfamiliar gestures. In this paper we assess the responses of 34 children, aged 2 to 5, to the most common in-app prompting techniques for eliciting specific gestures. By reviewing 100 touchscreen apps for preschoolers, we determined the types of prompts that children are likely to encounter. We then evaluated their relative effectiveness in teaching children to perform simple gestures. We found that children under 3 were only able to interpret instructions when they came from an adult model, but that children made rapid gains between age 3 and 3-and-a-half, at which point they were able to follow in-app audio instructions and on-screen demonstrations. The common technique of using visual state changes to prompt gestures was ineffective across this age range. Given that prior work in this space has primarily focused on children's fine motor control, our findings point to a need for increased attention to the design of prompts that accommodate children's cognitive development as well.","{'model': 'tldr@v2.0.0', 'text': 'It was found that children under 3 were only able to interpret instructions when they came from an adult model, but that children made rapid gains between age 3 and 3-and-a-half, at which point they were able to follow in-app audio instructions and on-screen demonstrations.'}",
-model discovery learning disertai teknik probing prompting dalam pembelajaran fisika di ma,Tiani Alfi Kusuma,"Discovery learning model by probing-prompting technique is one of the learning models of physics which guide the students to discover new things, the teacher provides the questions which can stimulate students’creativity and direct students to discover new knowledge. The purpose of this research are (1) to examine the influence of discovery learning model by probing-prompting techniquetothe students’ science-physics achievement at senior high school, and (2) to describe students’ science process skills during learning process. Kind of research is true experimental research by using post-test only control design. The techniques of data collection are observation, interview, documentation, andachievement test. The techniques of data analysis are indepependent sample T-Test and descriptive analysis by percentage. The Result of the study are (1) the value of p based on the T-Test is 0.001, and (2) the percentage of the overall students’science process skills is 81.71%. The study can be concluded that (1) discovery learning model by probing-prompting technique give significance influence to the students’ science-physics achievement, and (2) the students’ science process skills during join the physics learning by using discovery learning model by probing-prompting technique are in very good category",,
-detection techniques for extracellular polymeric substances in biofilms: a review,Mei Pan,"Extracellular polymeric substances (EPS) are one of the main components of biofilm, prompting biofilm to form a cohesive three-dimensional framework. Numerous methods are available to help characterize the properties and the structural, chemical and physical organizations of EPS during the biofilm formation process. This review highlights key techniques from different disciplines that have been successfully applied in-situ and non-destructively to describe the complex composition and distribution of EPS in biofilm, especially microscopic, spectroscopic, and the combination of multi-disciplinary methods that can provide new insights into the complex structure/function correlations in biofilms. Among them, confocal laser scanning microscopy (CLSM) is emphasized, and its principles, applications, advantages, and limitations are summarized. Multidisciplinary techniques have been developed and recommended to study EPS during the biofilm formation process, providing more in-depth insights into the composition and spatial distributions of EPS, so as to improve our understanding of the role EPS plays in biofilms ultimately.","{'model': 'tldr@v2.0.0', 'text': 'This review highlights key techniques from different disciplines that have been successfully applied in-situ and non-destructively to describe the complex composition and distribution of EPS in biofilm, especially microscopic, spectroscopic, and the combination of multi-disciplinary methods that can provide new insights into the complex structure/function correlations in biofilms.'}",https://ojs.cnr.ncsu.edu/index.php/BioRes/article/download/BioRes_11_3_Pan_Review_8092_Detection_Techniques_Extracellular_Biofilms/4693
-pharmacokinetics in morbid obesity: influence of two bariatric surgery techniques on paracetamol and caffeine metabolism,A. Goday Arnó,,"{'model': 'tldr@v2.0.0', 'text': 'The two studied bariatric surgical techniques normalize paracetamol oral bioavailability without impairing the liver function (measured by cytochrome P450 1A2 activity) and caffeine plasma levels are altered in severely obese patients.'}",http://repositori.upf.edu/bitstream/10230/35065/1/goday-osu-phar.pdf
-novel and advanced techniques for complex ivc filter retrieval,D. Daye,,"{'model': 'tldr@v2.0.0', 'text': 'Standard and advanced IVC filter retrieval techniques including single- access, dual-access, and dissection techniques are reviewed, carrying a non-negligible risk for complications such as filter fragmentation and resultant embolization of filter components, venous pseudoaneurysms or stenoses, and breach of the integrity of the caval wall.'}",
-application of cognitive rehabilitation theory to the development of smart prompting technologies,,"—Older adults with cognitive impairments often have difficulty performing instrumental activities of daily living (IADLs). Prompting technologies have gained popularity over the last decade and have the potential to assist these individuals with IADLs in order to live independently. Although prompting techniques are routinely used by caregivers and health care providers to aid individuals with cognitive impairment in maintaining their independence with everyday activities, there is no clear consensus or gold standard regarding prompt content, method of instruction, timing of delivery, or interface of prompt delivery in the gerontology or technology literatures. In this paper, we demonstrate how cognitive rehabilitation principles can inform and advance the development of more effective assistive prompting technologies that could be employed in smart environments. We first describe cognitive rehabilitation theory (CRT) and show how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion. We then use the CRT framework to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effective smart prompting technologies. Finally, we raise questions for future exploration as well as challenges and suggestions for future directions in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'Cognitive rehabilitation theory (CRT) is described and it is shown how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion and is used to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effectiveSmart prompting technologies.'}",
-a general language assistant as a laboratory for alignment,Amanda Askell,"Given the broad capabilities of large language models, it should be possible to work towards a general-purpose, text-based assistant that is aligned with human values, meaning that it is helpful, honest, and harmless. As an initial foray in this direction we study simple baseline techniques and evaluations, such as prompting. We find that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models. Next we investigate scaling trends for several training objectives relevant to alignment, comparing imitation learning, binary discrimination, and ranked preference modeling. We find that ranked preference modeling performs much better than imitation learning, and often scales more favorably with model size. In contrast, binary discrimination typically performs and scales very similarly to imitation learning. Finally we study a `preference model pre-training' stage of training, with the goal of improving sample efficiency when finetuning on human preferences.","{'model': 'tldr@v2.0.0', 'text': 'This work studies simple baseline techniques and evaluations and finds that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models.'}",
-enabling conversational interaction with mobile ui using large language models,Bryan Wang,"Conversational agents show the promise to allow users to interact with mobile devices using language. However, to perform diverse UI tasks with natural language, developers typically need to create separate datasets and models for each specific task, which is expensive and effort-consuming. Recently, pre-trained large language models (LLMs) have been shown capable of generalizing to various downstream tasks when prompted with a handful of examples from the target task. This paper investigates the feasibility of enabling versatile conversational interactions with mobile UIs using a single LLM. We designed prompting techniques to adapt an LLM to mobile UIs. We experimented with four important modeling tasks that address various scenarios in conversational interaction. Our method achieved competitive performance on these challenging tasks without requiring dedicated datasets and training, offering a lightweight and generalizable approach to enable language-based mobile interaction.","{'model': 'tldr@v2.0.0', 'text': 'This paper designed prompting techniques to adapt an LLM to mobile UIs without requiring dedicated datasets and training, offering a lightweight and generalizable approach to enable language-based mobile interaction.'}",https://dl.acm.org/doi/pdf/10.1145/3544548.3580895
-recent advances in natural language processing via large pre-trained language models: a survey,Bonan Min,"Large, pre-trained language models (PLMs) such as BERT and GPT have drastically changed the Natural Language Processing (NLP) field. For numerous NLP tasks, approaches leveraging PLMs have achieved state-of-the-art performance. The key idea is to learn a generic, latent representation of language from a generic task once, then share it across disparate NLP tasks. Language modeling serves as the generic task, one with abundant self-supervised text available for extensive training. This article presents the key fundamental concepts of PLM architectures and a comprehensive view of the shift to PLM-driven NLP techniques. It surveys work applying the pre-training then fine-tuning, prompting, and text generation approaches. In addition, it discusses PLM limitations and suggested directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'This article presents the key fundamental concepts of PLM architectures and a comprehensive view of the shift to PLM-driven NLP techniques, and surveys work applying the pre-training then fine-tuning, prompting, and text generation approaches.'}",https://arxiv.org/pdf/2111.01243
-single-molecule junctions beyond electronic transport.,S. Aradhya,,"{'model': 'tldr@v2.0.0', 'text': 'The emerging methods being used to interrogate multiple properties in single molecule-based devices are presented, how these measurements have advanced the understanding of the structure-function relationships in molecular junctions are detailed, and the potential for future research and applications are discussed.'}",
-are large language models ready for healthcare? a comparative study on clinical language understanding,Yuqing Wang,"Large language models (LLMs) have made significant progress in various domains, including healthcare. However, the specialized nature of clinical language understanding tasks presents unique challenges and limitations that warrant further investigation. In this study, we conduct a comprehensive evaluation of state-of-the-art LLMs, namely GPT-3.5, GPT-4, and Bard, within the realm of clinical language understanding tasks. These tasks span a diverse range, including named entity recognition, relation extraction, natural language inference, semantic textual similarity, document classification, and question-answering. We also introduce a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand. Our evaluation underscores the significance of task-specific learning strategies and prompting techniques for improving LLMs' effectiveness in healthcare-related tasks. Additionally, our in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP. Our study sheds light on the practical implications of employing LLMs in the specialized domain of healthcare, serving as a foundation for future research and the development of potential applications in healthcare settings.","{'model': 'tldr@v2.0.0', 'text': ""An in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP, and introduces a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand.""}",https://arxiv.org/pdf/2304.05368
-energy-saving techniques for reducing co 2 emissions in elementary schools,Taehoon Hong,"Global warming has grown into a significantly intense global issue, prompting all nations to exert various efforts to reduce their greenhouse gas emission, according to the United Nations Framework Convention on Climate Change (UNFCCC). The South Korean government is likewise working on reducing its greenhouse gas emission and, as part of its effort, it is currently promoting the Green-School Project. This project aims to create energy-saving schools by renovating deteriorated school buildings. In this study, CO2 emission reduction scenarios are sought by applying energy-saving techniques (ESTs) to educational facilities, and an optimal scenario is created by performing life cycle cost (LCC) analysis for each scenario. To this end, the amounts of energy consumption and CO2 emission reduction in two schools were evaluated using DesignBuilder simulation, based on 15 scenarios with combinations of four ESTs. LCC analysis was also performed. In terms of energy efficiency, CO2 emission reduction, and LCC analy...",,
-chloroplast-selective gene delivery and expression in planta using chitosan-complexed single-walled carbon nanotube carriers,Seonyeong Kwak,,"{'model': 'tldr@v2.0.0', 'text': 'This nanoparticle-mediated chloroplast transgene delivery tool provides practical advantages over current delivery techniques as a potential transformation method for mature plants to benefit plant bioengineering and biological studies.'}",
-direct-conversion radio transceivers for digital communications,A. Abidi,"The current interest in portable wireless communications devices is prompting research into new IC technologies, circuit configurations, and transceiver architectures. Miniature transceivers dissipating low power are sought to sought to communicate digital data. While transistor technology scaling and improved circuit techniques will lead to the inevitable evolutionary advances towards this goal, architectural innovations in the transceiver will lead to revolutionary improvements. It is in this context that there is a resurgence of interest in direct-conversion.",,
-a survey of rdf stores & sparql engines for querying knowledge graphs,Waqas Ali,,"{'model': 'tldr@v2.0.0', 'text': 'This survey paper provides a comprehensive review of techniques and systems for querying RDF knowledge graphs, including state-of-the-art storage, indexing and query processing techniques for efficiently evaluating SPARQL queries in a local setting.'}",https://www.techrxiv.org/articles/preprint/Survey_of_RDF_Stores_SPARQL_Engines_for_Querying_Knowledge_Graphs/14376884/files/27462647.pdf
-application of cognitive rehabilitation theory to the development of smart prompting technologies,Adriana M Seelye,"Older adults with cognitive impairments often have difficulty performing instrumental activities of daily living (IADLs). Prompting technologies have gained popularity over the last decade and have the potential to assist these individuals with IADLs in order to live independently. Although prompting techniques are routinely used by caregivers and health care providers to aid individuals with cognitive impairment in maintaining their independence with everyday activities, there is no clear consensus or gold standard regarding prompt content, method of instruction, timing of delivery, or interface of prompt delivery in the gerontology or technology literatures. In this paper, we demonstrate how cognitive rehabilitation principles can inform and advance the development of more effective assistive prompting technologies that could be employed in smart environments. We first describe cognitive rehabilitation theory (CRT) and show how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion. We then use the CRT framework to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effective smart prompting technologies. Finally, we raise questions for future exploration as well as challenges and suggestions for future directions in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'Cognitive rehabilitation theory (CRT) is described and it is shown how it provides a useful theoretical foundation for guiding the development of assistive technologies for IADL completion and is used to critically review existing smart prompting technologies to answer questions that will be integral to advancing development of effectiveSmart prompting technologies.'}",
-identifying active ingredients in complex behavioural interventions for obese adults with obesity-related co-morbidities or additional risk factors for co-morbidities: a systematic review,S. Dombrowski,"Abstract Reducing obesity is an important preventive strategy for people who are at increased risk of major disabling or life-threatening conditions. Behavioural treatments for obesity are complex and involve several components aiming to facilitate behaviour change. Systematic reviews need to assess the components that moderate intervention effects. Electronic databases and journals were searched for randomised controlled trials of behavioural interventions targeting dietary and/or physical activity change for obese adults (mean BMI≥30, mean age≥40 years) with risk factors and follow-up data≥12 weeks. A reliable taxonomy of theory-congruent behaviour change techniques (BCTs; Abraham & Michie, 2008) was used to identify programme components. Meta-regression suggested that increasing numbers of identified BCTs are not necessarily associated with better outcomes. The BCTs provision of instructions (β =− 2.69, p=0.02), self-monitoring (β = − 3.37, p<0.001), relapse prevention (β = − 2.63, p=0.02) and prompting practice (β = − 3.63, p<0.001) could be linked to more successful interventions. Studies including more BCTs aimed at dietary change that are congruent with Control Theory were associated with greater weight loss (β = − 1.13, p=0.04). Post-hoc ratings of intervention components in published trials can lead to the identification of components and theories for behaviour change practice and research.","{'model': 'tldr@v2.0.0', 'text': 'Meta-regression suggested that increasing numbers of identified BCTs are not necessarily associated with better outcomes, and post-hoc ratings of intervention components in published trials can lead to the identification of components and theories for behaviour change practice and research.'}",https://aura.abdn.ac.uk/bitstream/2164/2808/1/Dombrowski_2010.pdf
-teaching individuals with developmental delays: basic intervention techniques,O. Lovaas,"Autism Spectrum DisordersToilet Training SuccessHandbook of Autism and Pervasive Developmental Disorders, Assessment, Interventions, and PolicyEffective Practices for Children with AutismAutism Spectrum DisordersSocial and Communication Development in Autism Spectrum DisordersCompetencies for People Teaching Individuals with Autism and Other Pervasive Developmental DisordersTeaching Early Numeracy to Children with Developmental DisabilitiesThe Use of Video Prompting on the Acquisition, Maintenance, and Generalization of a Line Dance by Adolescents with Autism Spectrum DisordersTreating Childhood Psychopathology and Developmental DisabilitiesHandbook of Early Intervention for Autism Spectrum DisordersTeaching Language to Children with Autism Or Other Developmental DisabilitiesWhat Teachers Need to Know about Students with DisabilitiesA Complete ABA Curriculum for Individuals on the Autism Spectrum with a Developmental Age of 4-7 YearsDerived Relational RespondingImplementing the Group-Based Early Start Denver Model for Preschoolers with AutismAn ABA Curriculum for Children with Autism Spectrum Disorders Aged Approximately 1-4 YearsTeaching Young Children: Choices In Theory And PracticeClinical Assessment and Intervention for Autism Spectrum DisordersTeaching Social Communication to Children with Autism and Other Developmental Delays, Second Edition (2-book set)Teaching ChoicesEvaluation of Behavioural Skills Training with Volunteers Teaching Motor Skills to Individuals with Developmental DisabilitiesAdaptive Behavior Strategies for Individuals with Intellectual and Developmental DisabilitiesHandbook of Research on Technology Tools for Real-World Skill DevelopmentApplied Behavior Analysis for Children with Autism Spectrum DisordersThe Teaching of Self-medication Skills to Individuals with Developmental DisabilitiesEquity and Full Participation for Individuals with Severe DisabilitiesResources for Teaching MindfulnessMaking Inclusion Work for Students with Autism Spectrum DisordersThe Handbook of Intellectual Disability and Clinical Psychology PracticePractitioner's Guide to Evidence-Based PsychotherapyTeaching Functional Skills to Individuals with Developmental Disabilities Using Video PromptingTeaching Individuals with Developmental DelaysControversial Therapies for Developmental DisabilitiesChildren with Developmental DisabilitiesHandbook of Childhood Psychopathology and Developmental Disabilities AssessmentUnderstanding Children with Autism Spectrum DisordersControversial Therapies for Autism and Intellectual DisabilitiesThe Handbook of Language and Speech DisordersJEBPS Vol 17-N2","{'model': 'tldr@v2.0.0', 'text': 'Autism Spectrum DisordersToilet Training SuccessHandbook of Autism and Pervasive Developmental Disorders, Assessment, Interventions, and PolicyEffective Practices for Children with AutismAutistic Spectrum DisordersSocial and Communication Development in Autism Spectrum DisordersCompetencies for People Teaching Individuals with Autism and Other Pervvasive Developmental disorders.'}",
-"using computer, mobile and wearable technology enhanced interventions to reduce sedentary behaviour: a systematic review and meta-analysis",Aoife Stephenson,,"{'model': 'tldr@v2.0.0', 'text': 'Interventions using computer, mobile and wearable technologies can be effective in reducing SB, and effectiveness appeared most prominent in the short-term and lessened over time.'}",https://ijbnpa.biomedcentral.com/track/pdf/10.1186/s12966-017-0561-4
-fate of zno nanoparticles in soils and cowpea (vigna unguiculata).,Peng Wang,"The increasing use of zinc oxide nanoparticles (ZnO-NPs) in various commercial products is prompting detailed investigation regarding the fate of these materials in the environment. There is, however, a lack of information comparing the transformation of ZnO-NPs with soluble Zn(2+) in both soils and plants. Synchrotron-based techniques were used to examine the uptake and transformation of Zn in various tissues of cowpea ( Vigna unguiculata (L.) Walp.) exposed to ZnO-NPs or ZnCl2 following growth in either solution or soil culture. In solution culture, soluble Zn (ZnCl2) was more toxic than the ZnO-NPs, although there was substantial accumulation of ZnO-NPs on the root surface. When grown in soil, however, there was no significant difference in plant growth and accumulation or speciation of Zn between soluble Zn and ZnO-NP treatments, indicating that the added ZnO-NPs underwent rapid dissolution following their entry into the soil. This was confirmed by an incubation experiment with two soils, in which ZnO-NPs could not be detected after incubation for 1 h. The speciation of Zn was similar in shoot tissues for both soluble Zn and ZnO-NPs treatments and no upward translocation of ZnO-NPs from roots to shoots was observed in either solution or soil culture. Under the current experimental conditions, the similarity in uptake and toxicity of Zn from ZnO-NPs and soluble Zn in soils indicates that the ZnO-NPs used in this study did not constitute nanospecific risks.","{'model': 'tldr@v2.0.0', 'text': 'Under the current experimental conditions, the similarity in uptake and toxicity of Zn from ZnO-NPs and soluble Zn in soils indicates that the ZnNPs used in this study did not constitute nanospecific risks.'}",http://espace.library.uq.edu.au/view/UQ:320191/UQ320191OA.pdf
-prompting a consumer behavior for pollution control.,E. Geller,"A field application of behavior modification studied the relative effectiveness of different prompting procedures for increasing the probability that customers entering a grocery store would select their soft drinks in returnable rather than nonreturnable containers. Six different 2-hr experimental conditions during which bottle purchases were recorded were (1) No Prompt (i.e., control), (2) one student gave incoming customers a handbill urging the purchase of soft drinks in returnable bottles, (3) distribution of the handbill by one student and public charting of each customer's bottle purchases by another student, (4) handbill distribution and charting by a five-member group, (5) handbills distributed and purchases charted by three females. The variant prompting techniques were equally effective, and in general increased the percentage of returnable-bottle customers by an average of 25%.","{'model': 'tldr@v2.0.0', 'text': 'A field application of behavior modification studied the relative effectiveness of different prompting procedures for increasing the probability that customers entering a grocery store would select their soft drinks in returnable rather than nonreturnable containers.'}",https://europepmc.org/articles/pmc1310848?pdf=render
-using augmented reality to help children with autism stay focused,Lizbeth Escobedo,"Children with autism have difficulty sustaining their selective attention during therapy sessions. Attention management techniques involve the use of verbal and visual prompting, annotated on top of the physical objects used during therapies. Here, the authors explore how augmented reality helps integrate the physical and digital worlds, mimicking current strategies for attention management in autism. They describe their design decisions when developing the Mobile Object Identification System (Mobis), a mobile augmented reality application that lets teachers superimpose digital content on top of physical objects. The results of a five-week deployment study demonstrate that Mobis is useful and easy to use, increases the sustained and selective attention of children with autism, and elicits positive emotions during therapies. This article is part of a special issue on managing attention.","{'model': 'tldr@v2.0.0', 'text': 'How augmented reality helps integrate the physical and digital worlds, mimicking current strategies for attention management in autism, is explored in a special issue on managing attention.'}",
-tripling the capacity of wireless communications using electromagnetic polarization,M. R. Andrews,,"{'model': 'tldr@v2.0.0', 'text': 'It is shown that, in a scattering environment, an extra factor of three in channel capacity can be obtained, relative to the conventional limit using dual-polarized radio signals.'}",
-when can group level clustering be ignored? multilevel models versus single-level models with sparse data,P. Clarke,"Objective: The use of multilevel modelling with data from population-based surveys is often limited by the small number of cases per level-2 unit, prompting many researchers to use single-level techniques such as ordinary least squares regression. Design: Monte Carlo simulations are used to investigate the effects of data sparseness on the validity of parameter estimates in two-level versus single-level models. Setting: Both linear and non-linear hierarchical models are simulated in order to examine potential differences in the effects of small group size across continuous and discrete outcomes. Results are then compared with those obtained using disaggregated techniques (ordinary least squares and logistic regression). Main results: At the extremes of data sparseness (two observations per group), the group level variance components are overestimated in the two-level models. But with an average of only five observations per group, valid and reliable estimates of all parameters can be obtained when using a two-level model with either a continuous or a discrete outcome. In contrast, researchers run the risk of Type I error (standard errors biased downwards) when using single-level models even when there are as few as two observations per group on average. Bias is magnified when modelling discrete outcomes. Conclusions: Multilevel models can be reliably estimated with an average of only five observations per group. Disaggregated techniques carry an increased risk of Type I error, even in situations where there is only limited clustering in the data.","{'model': 'tldr@v2.0.0', 'text': 'Multilevel models can be reliably estimated with an average of only five observations per group, valid and reliable estimates of all parameters can be obtained when using a two-level model with either a continuous or a discrete outcome.'}",
-supplementary interviewing techniques to maximize output in free listing tasks,D. Brewer,"Free listing is an important ethnographic tool for defining semantic domains. However, when informants free list items from a particular domain, they often do not mention all items they know because they forget items and/or do not understand that they should list exhaustively. In this article, the author reviews results from research on three supplementary interviewing techniques to encourage full responding and enhance recall in such tasks (nonspecific prompting, reading back to the informant the items he or she free listed, and using free-listed items as semantic cues). These methods increase substantially the number of items elicited from individual informants and the number of items in a domain identified from informants in the aggregate. Moreover, these techniques do not require the interviewer to have any prior domain knowledge to be effective.",,
-2012 midwest drought in the united states,Ganeshchandra Mallya,"The 2012 North American drought may be the costliest and one of the most widespread natural disasters in U.S. history [USDA Economic Research Service (USDA-ERS) 2012]. While several states across the United States were experiencing drought conditions to varying degrees of severity, the Midwest and Northern Plains were perhaps the most affected. The drought severely impacted agricultural activities across the United States, particularly corn and soybean crops, prompting federal agencies including U.S. Department of Agriculture to declare disaster areas (USDA 2012b) and to provide assistance to those affected by this calamity. This paper utilizes existing and new techniques to provide insights into the severity of the 2012 Midwest drought and its impacts over the region.",,
-randomized trial of supplementary interviewing techniques to enhance recall of sexual partners in contact interviews*,D. Brewer,"Background: People with multiple sex partners tend to forget a significant proportion when recalling them. Methods: Randomized trial of supplementary interviewing techniques during routine partner notification contact interviews for chlamydia, gonorrhea, and syphilis in Colorado Springs, CO. Cases with multiple sex partners in the last 3 months (n = 123) participated. Interviewers prompted nonspecifically and read back the list of elicited partners after cases recalled partners on their own. We then randomly assigned cases to receive 1 of 3 sets of recall cues: (1) an experimental set of cues consisting of locations where people meet partners, role relationships, network ties, and first letters of names; (2) another experimental set including common first names; and (3) control cues referring to individual characteristics (e.g., physical appearance). Results: Nonspecific prompting and reading back the list each increased the number of additional partners elicited and located by 3% to 5% on average. On average, the combined location/role/letter/network cues elicited more additional partners (0.57) than did the first-name (0.29) and individual characteristics (0.28) cues. The location and first-name cues were the most effective in eliciting located partners. The supplementary techniques increased the number of new cases found by 12% and, importantly, identified branches of the sexual network that would not otherwise have been discovered. Conclusion: Elicitation of sex partners can be enhanced in contact interviews with simple interviewing techniques, resulting in improved network ascertainment and sexually transmitted disease case finding.","{'model': 'tldr@v2.0.0', 'text': 'Elicitation of sex partners can be enhanced in contact interviews with simple interviewing techniques, resulting in improved network ascertainment and sexually transmitted disease case finding.'}",http://www.interscientific.net/csevcues.pdf
-the useful field of view test: a new technique for evaluating age-related declines in visual function.,K. Ball,"Many older adults eventually experience difficulty in everyday visual activities, prompting them to consult with an eye care specialist. In some cases, a thorough examination may reveal no clinical basis for these difficulties. New techniques for evaluating age-related declines in visual function have substantiated the existence of deficits in visual attention that are not adequately captured by standard clinical measures, and which are predictive of real world problems in driving and mobility. These new techniques may serve an important function in guiding the diagnosis, and treatment of those older adults experiencing functional visual problems.","{'model': 'tldr@v2.0.0', 'text': 'New techniques for evaluating age-related declines in visual function have substantiated the existence of deficits in visual attention that are not adequately captured by standard clinical measures, and which are predictive of real world problems in driving and mobility.'}",
-"interviewing autistic adults: adaptations to support recall in police, employment, and healthcare interviews",J. E. Norris,"Recalling specific past experiences is critical for most formal social interactions, including when being interviewed for employment, as a witness or defendant in the criminal justice system, or as a patient during a clinical consultation. Such interviews can be difficult for autistic adults under standard open questioning, yet applied research into effective methods to facilitate autistic adults’ recall is only recently beginning to emerge. The current study tested the efficacy of different prompting techniques to support autistic adults’ recall of specific personal memories; 30 autistic and 30 typically developing adults (intelligence quotients > 85) were asked to recall specific instances from their past, relevant to criminal justice system, healthcare, and employment interviews. Questions comprised ‘open questions’, ‘semantic prompting’ (where semantic knowledge was used to prompt specific episodic retrieval) and ‘visual–verbal prompting’ (a pie-diagram with prompts to recall specific details, for example, who, what, and where). Half the participants received the questions in advance. Consistent with previous research, autistic participants reported memories with reduced specificity. For both groups, visual–verbal prompting support improved specificity and episodic relevance, while semantic prompting also aided recall for employment questions (but not health or criminal justice system). Findings offer new practical insight for interviewers to facilitate communication with typically developing and autistic adults. Lay abstract During many types of interviews (e.g. in employment, with the police, and in healthcare), we need to recall detailed memories of specific events, which can be difficult for autistic people in response to commonly used questions. This is especially because these tend to be open questions (i.e. very broad). Autistic people have disproportionately high rates of physical and mental health conditions, are more likely to interact with police, and are the most underemployed disability group. However, interviewers are often unsure about how to adapt their communication for autistic people. Our research tested whether different types of prompts enabled autistic people to recall specific memories (memories of a single event within one day). Participants were asked about situations relating to witnessing a crime (e.g. at the bank), physical or mental health scenarios and employment interviews (e.g. a time you’ve met a deadline). We tested the following: Open questions: basic questions only (e.g. ‘tell me about a time you went to the cinema’), Semantic prompting: a general prompt (e.g. ‘do you enjoy going to the cinema?’) before asking for a specific instance (‘tell me about a time you went to the cinema?’), Visual–verbal prompting: asking participants to recall when it happened, who was there, the actions that occurred, the setting, and any objects. With visual–verbal prompting, autistic and typically developing participants’ memories were more specific and detailed. Semantic prompting was also effective for employment questions. Our study shows that autistic people can recall specific memories when they are appropriately prompted. Visual–verbal prompting may be effective across different situations.","{'model': 'tldr@v2.0.0', 'text': 'This study shows that autistic people can recall specific memories when they are appropriately prompted, and suggests visual–verbal prompting may be effective across different situations.'}",
-diagnosis and treatment of fetal arrhythmia,A. Wacker-Gussmann,"Abstract Aims Detection and careful stratification of fetal heart rate (FHR) is extremely important in all pregnancies. The most lethal cardiac rhythm disturbances occur during apparently normal pregnancies where FHR and rhythm are regular and within normal or low-normal ranges. These hidden depolarization and repolarization abnormalities, associated with genetic ion channelopathies cannot be detected by echocardiography, and may be responsible for up to 10% of unexplained fetal demise, prompting a need for newer and better fetal diagnostic techniques. Other manifest fetal arrhythmias such as premature beats, tachycardia, and bradycardia are commonly recognized. Methods Heart rhythm diagnosis in obstetrical practice is usually made by M-mode and pulsed Doppler fetal echocardiography, but not all fetal cardiac time intervals are captured by echocardiographic methods. Results and Conclusions This article reviews different types of fetal arrhythmias, their presentation and treatment strategies, and gives an overview of the present and future diagnostic techniques.","{'model': 'tldr@v2.0.0', 'text': 'This article reviews different types of fetal arrhythmias, their presentation and treatment strategies, and gives an overview of the present and future diagnostic techniques.'}",http://www.thieme-connect.de/products/ejournals/pdf/10.1055/s-0034-1372430.pdf
-an analysis of prompting procedures for teaching a sight vocabulary1,O. Duell,"""Prompting"" is one technique used to facilitate learning in selfinstructional programs. Depending upon the context, a prompt may consist, for instance, of the printed form of a word the subject (S) is to vocalize, a rule which can be applied to an example, a hint to help in the solution of a problem, the first letter of an answer, a synonym for an answer, or a picture to elicit the vocalization of a word. Prompting can be defined generally as the pairing of a cue, the stimulus that is to control a response when training is completed, with a prompt, a stimulus which already controls or partially controls the desired response. The object of training sequences using prompts is to shift the control of the response from the prompt to the cue with which the prompt is paired. Although there have been some notable successes with prompted training sequences, there also have been anomalies. There is still much to be learned about conditions under which prompting techniques can be used to advantage (Anderson, 1967). Recent studies (Anderson and Faust, 1967; Faust and Anderson, 1967) have shown that a prompted sequence teaches more if it forces the student to notice the cues while he responds during training. College students learned more Russian vocabulary from",,
-one-class classification-based real-time activity error detection in smart homes,Barnan Das,"Caring for individuals with dementia is frequently associated with extreme physical and emotional stress, which often leads to depression. Smart home technology and advances in machine learning techniques can provide innovative solutions to reduce caregiver burden. One key service that caregivers provide is prompting individuals with memory limitations to initiate and complete daily activities. We hypothesize that sensor technologies combined with machine learning techniques can automate the process of providing reminder-based interventions. The first step toward automated interventions is to detect when an individual faces difficulty with activities. We propose machine learning approaches based on one-class classification that learn normal activity patterns. When we apply these classifiers to activity patterns that were not seen before, the classifiers are able to detect activity errors, which represent potential prompt situations. We validate our approaches on smart home sensor data obtained from older adult participants, some of whom faced difficulties performing routine activities and thus committed errors.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesize that sensor technologies combined with machine learning techniques can automate the process of providing reminder-based interventions, and proposes machine learning approaches based on one-class classification that learn normal activity patterns.'}",
-"""expertness"" from structured text? reconsider: a diagnostic prompting program",M. Tuttle,"RECONSIDER is an interactive diagnostic prompting program which uses simple information retrieval techniques to prompt a physician regarding possible diagnoses, given a list of positive patient findings. Its knowledge base consists of ""structured text"" definitions of 3262 diseases and a synonym dictionary Patient findings, and their synonyms, are matched against inverted files of terms from the disease descriptions, the number and selectivity of the patient findings matching terms in a given disease description determine that disease's ""score"", and the matched diseases are sorted on this score to form a preliminary differential diagnosis. Definitions of diseases can be referenced for viewing by name, or by their position in a differential. While its first formal evaluation is not yet complete, the performance of RECONSIDER continues to exceed the expectations of user and designer alike.","{'model': 'tldr@v2.0.0', 'text': 'While its first formal evaluation is not yet complete, the performance of RECONSIDER continues to exceed the expectations of user and designer alike.'}",
-"óleos essenciais no brasil: aspectos gerais, desenvolvimento e perspectivas",H. Bizzo,"Essential oils are extracted by steam distillation of plants or cold pressing of citrus fruit pericarp. They are used in food, cosmetic, personal care and pharmaceutical industries. In Brazil, oils from orange and related products contribute to near 97% to the positive commercial performance of the sector. Predatory exploitation and the availability of new sources of raw materials, with more attracting prices, changed the paradigm. Prospective studies, sustainable use of Brazilian biodiversity, domestication of exotic species with commercial relevance, the use of breeding techniques and the development of new applications for essential oils are thematic lines, usually multidisciplinary, which have been prompting the expansion of the research on essential oils. This paper presents an analysis on essential oils balance trade from 2005 to 2008 and some historical data on research and production of essential oils in Brazil.",,https://www.scielo.br/j/qn/a/QwJBsdNzGmZSq4jKmhwVDnJ/?lang=pt&format=pdf
-the conceptual structure of research utilization.,C. Estabrooks,"Despite the fact that the nursing literature is replete with calls to make the practice of nursing research based, little is known about the structure and function of research utilization. The purpose of this study was to explore the conceptual structure of research utilization. Data were collected from a randomly selected sample of 600 registered nurses practicing in western Canada. Using the techniques of structural equation modeling (with LISREL), competing models representing conceptual structures of research utilization were developed and evaluated. In the first model, a simplex style of model, the investigator proposed that a nurse's early responses would influence subsequent responses to the question measuring research utilization, implying a time ordered causal sequence. In the second style of model, a common cause (or factor-like) model, the investigator proposed a stable underlying concept, research utilization, that was relatively insensitive to prompting and time ordering. The simplex style of model failed to reach acceptable indices of fit. The common cause model fit the data well, suggesting that instrumental, conceptual, and persuasive research utilization exist and that a global measure of research utilization may be defensible.","{'model': 'tldr@v2.0.0', 'text': 'Two competing models representing conceptual structures of research utilization were developed and evaluated, suggesting that instrumental, conceptual, and persuasive research utilization exist and that a global measure ofResearch utilization may be defensible.'}",
-realization of techniques in problem solving: the construction of bijections for enumeration tasks,J. Mamona-Downs,,,
-the effects of vicarious prompting on attentive behavior of children with behavior disorders,J. Birnbrauer,"Vicarious prompting, reinforcing and labelling the appropriate behavior of a nontarget child contingent on inappropriate behavior of a target child, was assessed in two classrooms of children with behavior disorders. Vicarious prompts using social reinforcement alone were clearly effective in increasing the attentive behavior of one target child in Experiment 1. The children in Experiment 2, however, did not appera to be responsive to vicarious prompting. Comparison of the two groups suggests that vicarious prompting, a relatively subtle, indirect control technique, may be a low cost, positive, and effective alternative to other techniques under conditions in which the behavior of most of the group is attentive.",,
-a comparison of behavioral techniques to teach functional independent-living skills to individuals with severe and profound mental retardation,J. Matson,"The efficacy of two treatment approaches was compared for functional skill acquisition in individuals with severe and profound mental retardation. Participants included 22 residents from a large developmental center (Pinecrest) in central Louisiana. Treatment including staff training, feedback, and edible reinforcement in addition to prompting, modeling, and physical guidance was more effective than prompting, modeling, and physical guidance alone. Additionally, daily documentation of teaching did not enhance treatment effectiveness. Implications of the findings are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Treatment including staff training, feedback, and edible reinforcement in addition to prompting, modeled, and physical guidance alone was more effective than prompting, modeling, andPhysical guidance alone.'}",
-addressing data sparseness in contextual population research,P. Clarke,"The use of multilevel modeling with data from population-based surveys is often limited by the small number of cases per Level 2 unit, prompting a recent trend in the neighborhood literature to apply cluster techniques to address the problem of data sparseness. In this study, the authors use Monte Carlo simulations to investigate the effects of marginal group sizes on multilevel model performance, bias, and efficiency. They then employ cluster analysis techniques to minimize data sparseness and examine the consequences in the simulations. They find that estimates of the fixed effects are robust at the extremes of data sparseness, while cluster analysis is an effective strategy to increase group size and prevent the overestimation of variance components. However, researchers should be cautious about the degree to which they use such clustering techniques due to the introduction of artificial within-group heterogeneity.","{'model': 'tldr@v2.0.0', 'text': 'This study uses Monte Carlo simulations to investigate the effects of marginal group sizes on multilevel model performance, bias, and efficiency, and employs cluster analysis techniques to minimize data sparseness and examines the consequences in the simulations.'}",
-changes over time and transfer of analogy-problem solving of gifted and non-gifted children in a dynamic testing setting,B. Vogelaar,"Abstract This study examined differences in transfer of analogical reasoning after analogy-problem solving between 40 gifted and 95 average-ability children (aged 9–10 years old), utilising dynamic testing principles. This approach was used in order to examine potential differences between gifted and average-ability children in relation to progression after training, and with regard to the question whether training children in analogy problem-solving elicits transfer of analogical reasoning skills to an analogy construction-task. Children were allocated to one of two experimental conditions: either children received unguided practice in analogy problem-solving, or they were provided with this in addition to training incorporating graduated prompting techniques. The results showed that gifted and average-ability children who were trained made more progress in analogy problem-solving than their peers who received unguided practice experiences only. Gifted and average-ability children were found to show similar progression in analogy problem-solving, and gifted children did not appear to have an advantage in the analogy-construction transfer task. The dynamic training seemed to bring about no additional improvement on the transfer task over that of unguided practice experiences only.",,https://www.tandfonline.com/doi/pdf/10.1080/01443410.2017.1409886?needAccess=true
-"behavior therapy techniques applied to eating, exercise, and diet modification in childhood obesity",J. Varni,"Behavior therapy techniques are receiving increasing attention in the treatment of childhood obesity. This paper provides a selective review of experimental studies of childhood obesity that have utilized behavior therapy techniques with preadolescent populations. It further defines and describes specific behavior therapy techniques utilized in the modification of eating, exercise, and diet patterns in childhood obesity, including stimulus control, stimulus and response generalization, prompting, positive reinforcement, shaping, contingency contracting, and generalization and maintenance strategies.","{'model': 'tldr@v2.0.0', 'text': 'Specific behavior therapy techniques utilized in the modification of eating, exercise, and diet patterns in childhood obesity, including stimulus control, stimulus and response generalization, prompting, positive reinforcement, shaping, contingency contracting, and generalization and maintenance strategies are described.'}",
-combined home and school obesity prevention interventions for children,G. Hendrie,"This review identifies studies describing interventions delivered across both the home and school/community setting, which target obesity and weight-related nutrition and physical activity behaviors in children. Fifteen studies, published between 1998 and 2010, were included and evaluated for effectiveness, study quality, nutrition/activity content, behavior change techniques, and theoretical basis, using validated assessment tools/taxonomies. Seven studies were rated as effective. Behavior change techniques used to engage families, and techniques associated with intervention effectiveness were coded. Effective studies used about 10 behavior change techniques, compared with 6.5 in ineffective studies. Effective interventions used techniques including providing general information on behavior–health links, prompting practice of behavior, and planning for social support/social changes. Different behavior change techniques were applied in the home and school setting. The findings of this review provide novel insights into the techniques associated with intervention effectiveness that can inform the development of public health obesity prevention strategies.","{'model': 'tldr@v2.0.0', 'text': 'This review identifies studies describing interventions delivered across both the home and school/community setting, which target obesity and weight-related nutrition and physical activity behaviors in children, and provides novel insights into the techniques associated with intervention effectiveness.'}",
-operational risk management,I. Moosa,,,
-harnessing repetitive behaviours to engage attention and learning in a novel therapy for autism: an exploratory analysis,G. Chen,"Rigorous, quantitative examination of therapeutic techniques anecdotally reported to have been successful in people with autism who lack communicative speech will help guide basic science toward a more complete characterisation of the cognitive profile in this underserved subpopulation, and show the extent to which theories and results developed with the high-functioning subpopulation may apply. This study examines a novel therapy, the “Rapid Prompting Method” (RPM). RPM is a parent-developed communicative and educational therapy for persons with autism who do not speak or who have difficulty using speech communicatively. The technique aims to develop a means of interactive learning by pointing amongst multiple-choice options presented at different locations in space, with the aid of sensory “prompts” which evoke a response without cueing any specific response option. The prompts are meant to draw and to maintain attention to the communicative task – making the communicative and educational content coincident with the most physically salient, attention-capturing stimulus – and to extinguish the sensory–motor preoccupations with which the prompts compete. Video-recorded RPM sessions with nine autistic children ages 8–14 years who lacked functional communicative speech were coded for behaviours of interest. An analysis controlled for age indicates that exposure to the claimed therapy appears to support a decrease in repetitive behaviours and an increase in the number of multiple-choice response options without any decrease in successful responding. Direct gaze is not related to successful responding, suggesting that direct gaze might not be any advantage for this population and need not in all cases be a precondition to communication therapies.","{'model': 'tldr@v2.0.0', 'text': 'Exposure to the claimed therapy appears to support a decrease in repetitive behaviours and an increase in the number of multiple-choice response options without any decrease in successful responding, suggesting that direct gaze might not be any advantage for this population and need not be a precondition to communication therapies.'}",
-errorless learning: review and practical application for teaching children with pervasive developmental disorders.,Michael M. Mueller,"Errorless learning refers to a variety of discrimination learning techniques that eliminate or minimize responding to incorrect choices. This article describes experimental roots of errorless learning and applied errorless strategies. Specifically, previous research on stimulus fading, stimulus shaping, response prevention, delayed prompting, superimposition with stimulus fading, and superimposition with stimulus shaping are discussed. Educational applications for children with Pervasive Developmental Disorders (PDD) are presented for each technique so that school psychologists, educators, and teachers working with children with PDD can understand the underpinnings and practical applications of errorless techniques to use in skill acquisition programming in school settings. © 2007 Wiley Periodicals, Inc. Psychol Schs 44: 691–700, 2007.",,
-a palmtop computer program for the treatment of generalized anxiety disorder,M. Newman,"This is the first report of a palmtop computer program developed to increase the efficiency and cost-effectiveness of cognitive behavioral therapy for generalized anxiety disorder (GAD). The computer program offers advantages to researchers, therapists, and clients. These advantages include continuous, unobtrusive collection of process data on treatment adherence as well as on the impact of cognitive behavioral therapy techniques in the client’s natural setting. In addition, the computer extends treatment beyond the therapy hour and motivates clients to comply with homework assignments by prompting practice of cognitive behavioral strategies. The successful application of the palmtop computer program reported in this integrated series suggests a new line of research directed toward increasing the cost-effectiveness of what is currently the goldstandard treatment for GAD.","{'model': 'tldr@v2.0.0', 'text': 'The successful application of the palmtop computer program reported in this integrated series suggests a new line of research directed toward increasing the cost-effectiveness of what is currently the goldstandard treatment for GAD.'}",
-imaging drosophila gene activation and polymerase pausing in vivo,J. Lis,,"{'model': 'tldr@v2.0.0', 'text': 'Live-cell imaging techniques coupled with protein–DNA crosslinking assays on a genome-wide scale allow more detailed mechanistic questions to be addressed and are prompting the re-evaluation of models of transcription regulation in both Drosophila and mammals.'}",https://www.nature.com/articles/nature06324.pdf
-pengaruh penerapan pendekatan scaffolding terhadap kemampuan pemahaman konsep matematika siswa kelas viii smp pertiwi 2 padang,Nicke Septriani,"The ability of students to understand mathematical concepts are expected to grow and thrive in learning mathematics. But mathematic lesson that happen in class VIII SMP Pertiwi 2 Padang, understanding of mathematics concept student still lower. During the learning process of students tend to memorize a given concept. This resulted in the ability of student understanding of mathematical concept is low. Scaffolding approach to probing-prompting techniques are expected to facilitate the students in an improved understanding of the concept. This research aimed to see to whether students' understanding of mathematical concepts are learned with scaffolding approach is better than the understanding of mathematical concepts that students learn using conventional learning in class VIII SMP Pertiwi 2 Padang. This Type Research is experiment kuasi with device research of Randomized Control Group Only Design. Pursuant to result of data analysis can be concluded by the understanding of mathematics concept after applied Scaffolding approach better than the understanding of mathematics concept after applying of conventional study student of c lass VIII SMP Pertiwi 2 Padang .","{'model': 'tldr@v2.0.0', 'text': ""This research aimed to see to whether students' understanding of mathematical concepts are learned with scaffolding approach is better than the understanding of mathematics concepts that students learn using conventional learning in class VIII SMP Pertiwi 2 Padang.""}",
-"""i can speak for myself"": involving individuals with intellectual disabilities as research participants.",Jennifer Mactavish,"The Lifespan and Disability Project, a 2-year qualitative study, was designed to enhance understanding of social integration by including the perspectives of individuals with intellectual disabilities. Procedures and strategies employed to involve these individuals in the project were documented. Specifically, we describe the use of individual interviews and focus groups (e.g., purpose, rationale, facilitation techniques) to collect data, and verification meetings in which a prompting hierarchy and cueing were employed to assist participants in confirming or disconfirming the researchers' interpretations of the emerging findings. Highlights of the results were presented to illustrate the efficacy of the strategies in enabling the participants with intellectual disabilities to share their perspectives on social integration.","{'model': 'tldr@v2.0.0', 'text': ""The use of individual interviews and focus groups to collect data, and verification meetings in which a prompting hierarchy and cueing were employed to assist participants in confirming or disconfirming the researchers' interpretations of the emerging findings are described.""}",
-using hidden markov models to characterize student behaviors in learning-by-teaching environments,Hogyeong Jeong,,"{'model': 'tldr@v2.0.0', 'text': ""Evidence is presented that HMMs can be used to effectively determine students' pattern of activities and clear differences between different interventions, and links between students learning performance and their interactions with the system are indicated.""}",
-syngraph: a graphical user interface generator,D. Olsen,"The SYNGRAPH system automatically generates graphical user interfaces. It generates interactive Pascal programs from a description of the input language's grammar. From the grammar it deduces information about how to manage both physical and simulated devices, and how prompting and echoing are performed. Input errors are detected, and can be corrected using automatically provided rubout and cancel features. The natural integration of application specific semantics into the system is also shown including appropriate semantic recovery from input errors.","{'model': 'tldr@v2.0.0', 'text': ""The SYNGRAPH system automatically generates graphical user interfaces from a description of the input language's grammar, and the natural integration of application specific semantics into the system is shown including appropriate semantic recovery from input errors.""}",
-programming the generalization of a greeting response in four retarded children.,T. Stokes,"Reinforcement techniques of prompting and shaping were employed to develop handwaving, a useful social greeting response, in four institutionalized retarded subjects. A multiple-baseline design across subjects demonstrated the reliable functioning of the training procedures. Specifically, it showed that training and maintenance of the greeting response by one experimenter was not usually sufficient for generalization of the response to the more than 20 other members of the institution staff who had not participated in the training of the response. However, high levels of generalization to staff members were recorded for three subjects over periods ranging from one to six months after a second experimenter trained and maintained the response in conjunction with the first experimenter. The fourth subject, although never receiving training by a second experimenter, showed similar results following a second training by the first experimenter.","{'model': 'tldr@v2.0.0', 'text': 'Reinforcement techniques of prompting and shaping were employed to develop handwaving, a useful social greeting response, in four institutionalized retarded subjects and showed that training and maintenance of the greeting response by one experimenter was not usually sufficient for generalization of the response to the more than 20 other members of the institution staff who had not participated in the training of theresponse.'}",https://europepmc.org/articles/pmc1311674?pdf=render
-what counts in brain aging? design-based stereological analysis of cell number.,Jeffrey M. Long,"The advent and implementation of new design-based stereological techniques allows the quantification of cell number without the assumptions required when obtaining areal densities. These new techniques are rapidly becoming the standard for quantifying cell number, particularly in aging studies. Recently, studies using stereological techniques have failed to confirm earlier findings regarding age-associated neural loss. This newly emerging view of retained cell number during aging is having a major impact on biogerontology, prompting revaluation of long-standing hypotheses of age-related cell loss as causal for age-related impairments in brain functioning. Rather than focus on neuronal loss as the end-result of a negative cascade of neuronal injury, research has begun to consider that age-related behavioral declines may reflect neuronal dysfunction (e.g., synaptic or receptor loss, signal transduction deficits) instead of neuronal death. Here we discuss design-based stereology in the context of age-related change in brain cell number and its impact on consideration of structural change in brain aging. Emergence of this method of morphometrics, however, can have relevance to many areas of gerontological research.","{'model': 'tldr@v2.0.0', 'text': 'The advent and implementation of new design-based stereological techniques allows the quantification of cell number without the assumptions required when obtaining areal densities, and their impact on consideration of structural change in brain aging is discussed.'}",https://academic.oup.com/biomedgerontology/article-pdf/54/10/B407/1611374/10568523.pdf
-the detection of interturn stator faults in doubly-fed induction generators,H. Douglas,"Presently, many condition monitoring techniques that are based on steady-state analysis are being applied to wind generators. However, the operation of wind generators is predominantly transient, therefore prompting the development of non-stationary techniques for fault detection. In this paper we apply steady-state techniques, e.g. motor current signatures analysis (MCSA) and the extended Park's vector approach (EPVA), as well as a new transient technique that is a combination of the EPVA, the discrete wavelet transform and statistics, to the detection of turn faults in a doubly-fed induction generators (DFIG). It is shown that steady-state techniques are not effective when applied to DFIG's operating under transient conditions. The new technique shows that stator turn faults can be unambiguously detected under transient conditions.",,
-‘projective transidentification’: an extension of the concept of projective identification,J. Grotstein,"Questions about the concept of projective identifi cation still persist. The author presents the following hypotheses: Klein's traditional view and Bion's extension and revision of it can be thought of as occupying a continuum in reverse. He postulates that Bion's concept of communicative intersubjective projective identifi cation (which the author renames ‘projective transidentifi cation’) is primary and inclusive of Klein's earlier unconscious, omnipotent, intrapsychic mode but includes Bion's ‘realistic’ communicative mode as well. The author hypothesizes, consequently, that intersubjective projective identifi cation constitutes both the operation of an unconscious phantasy of omnipotent intrapsychic projective identifi cation solely within the internal world of the projecting subject‐in addition to two other processes: conscious and/or preconscious modes of sensorimotor induction, which would include signaling and/or evocation or prompting gestures or techniques (mental, physical, verbal, posturing or priming) on the part of the projecting subject; followed by spontaneous empathic simulation in the receptive object of the subject's experience in which the receptive object is already inherently ‘hard‐wired’ to be empathic with the prompting subject.","{'model': 'tldr@v2.0.0', 'text': 'The author hypothesizes that intersubjective projective identifi cation constitutes both the operation of an unconscious phantasy of omnipotent intrapsychic projective identification solely within the internal world of the projecting subject and two other processes: conscious and/or preconscious modes of sensorimotor induction.'}",
-increasing behaviors of severely retarded and autistic persons,R. Foxx,"This widely used manual shows how to increase desirable behaviors by using techniques such as shaping, prompting, fading, modeling, backward chaining, and graduated guidance. It offers specific guidelines for arranging and managing the learning environment as well as standards for evaluating and maintaining success. Exercises, review questions, and numerous examples are included. The book is written for special educators, aides, residential staff, and those responsible for designing or evaluating behavioral programs. Often adopted as a supplementary college text.",,
-a behavioral science approach to transportation safety.,E. Geller,"This paper reviews a variety of techniques that can increase the large-scale use of vehicle safety belts. Activator approaches or behavior change strategies that precede opportunities to buckle up included: reminder techniques (i.e., buckle-up dashborad stickers, sound-light reminder systems in vehicles, and special ""Airline Lifesaver"" and ""Flash for Life"" prompting procedures), role modeling by television stars, educational tactics involving young school children, commitment pledgecard signing, incentives, and disincentives. Reward strategies can be direct and immediate, direct and delayed, or indirect. These three techniques for delivering rewards have been equivalently effective at increasing safety belt use in corporate and community settings. The conclusion is that a number of strategies are cost effective for increasing safety belt use on a large scale, and no single approach is sufficient.","{'model': 'tldr@v2.0.0', 'text': 'The conclusion is that a number of strategies are cost effective for increasing safety belt use on a large scale, and no single approach is sufficient.'}",
-maps: dynamic scaffolding for independence for persons with cognitive impairments,Stefan Carmien,,"{'model': 'tldr@v2.0.0', 'text': None}",
-a modified speech tracking procedure as a communicative performance measure.,M. Matthies,"A modified speech tracking procedure was investigated to determine the effects of controlling connected discourse materials and of specific prompting techniques on tracking rate and subjects' responses. In addition, comprehension was tested after subjects completed the tracking task. Two experiments were conducted with normally hearing subjects, one in which subjects tracked in an auditory plus noise condition and the second in which they tracked in a visual-alone condition. The results of these two experiments showed significant differences between tracking rates for controlled versus uncontrolled materials for both modalities, with higher tracking rates observed for the controlled materials. The use of specific prompting techniques reduced intertester variability in both modalities. Further, results indicated that comprehension ability and tracking skill are not perfectly related.","{'model': 'tldr@v2.0.0', 'text': 'Results indicated that comprehension ability and tracking skill are not perfectly related, with significant differences between tracking rates for controlled versus uncontrolled materials for both modalities.'}",
-overprompting science students using adjunct study questions.,W. Holliday,"Overprompting students by providing them with strong hints to answers of questions can do learners more instructional harm than good. The selective attention model was used to explain the effects of overprompting students provided with study questions adjunct to a complex flow diagram describing scientific cyclical schema. Tenth-grade students were randomly assigned to an unprompted-question, no-question, prompted-question, and a placebo control group. Analysis showed that strongly prompting students to the answers of such questions was less effective than an unprompted question treatment, p < 0.05. The no-question treatment did not significantly outperform the prompted treatment. The information presented in the flow diagram was operationally related to comprehension study and posttest questions. The theoretical discussion and the present findings suggested that science teachers should use prompting techniques with extreme caution.",,
-conservation of electricity on a college campus,E. Zolik,"This study investigated the extent to which posters requesting that lights be turned off in classrooms would increase energy conserving behaviors. A multiple baseline design was used with two groups of eleven classrooms. Results indicated that with onset of the intervention phase, moderate increases in energy conserving behaviors occurred in both clusters of classrooms. These findings indicate that prompting techniques are an effective strategy for getting lights turned off in classrooms. The conservation of electrical energy can be approached from two different perspectives: increasing the efficiency of equipment which is operated by electricity and the installation of automated equipment to eliminate waste on the one hand, and the modification of behavior patterns of people with respect to energy consumption. For maximum effectiveness, both approaches — the technical and the behavioral — have to be combined. While solely informational procedures have had only negligible effects on energy saving [1], incentive payments reducing energy use [2], and regular feedback seem to be strong behavioral procedures for inducing energy conservation [3]. One area where different strategies have been infrequently implemented is in college classrooms. Winett, in a study of conservation in a classroom, reported that a large two foot by three foot sign was effective in reducing the per cent of lights turned off * Poster presented at the Association for Behavior Analysis, Milwaukee, May 1982.",,
-preparing severely and profoundly mentally retarded adults for tests of motor fitness,P. Tomporowski,"The effectiveness of using behavior shaping techniques to prepare nine institutionalized severely and profoundly mentally retarded adults for tests of motor fitness was assessed. Three severely mentally retarded subjects served as the executive subjects in a yoked-control study. Each executive subject was paired randomly with two profoundly mentally retarded subjects. A three-phase multiple baseline experiment was then conducted with executive subjects; yoked subjects received the same intervention as their executive partner. The training program consisted of a verbal instruction baseline phase, a modeling phase, and a graduated guidance phase. Tasks used for training were the shuttle run and standing long jump. Modeling instruction produced minimal improvement in subject performance; however, graduated guidance instruction produced marked, immediate improvement in the performance of six of nine subjects on both tasks. These results suggest that physical prompting techniques will prepare some, but not all...",,
-monte carlo techniques in radiation therapy,J. Seco,"Monte Carlo Fundamentals History of Monte Carlo Alex F. Bielajew Basics of Monte Carlo Simulations Matthias Fippel Variance Reduction Techniques Matthias Fippel Application of Monte Carlo Techniques in Radiation Therapy Applications of Monte Carlo to Radiation Dosimetry Hugo Bouchard and Jan Seuntjens Monte Carlo Modeling of External Photon Beams in Radiotherapy Frank Verhaegen Monte Carlo Modeling of External Electron Beams in Radiotherapy Frank Verhaegen Dynamic Beam Delivery and 4D Monte Carlo Emily Heath and Joao Seco Patient Dose Calculation Joao Seco and Maggy Fragoso Monte Carlo Methods and Applications for Brachytherapy Dosimetry and Treatment Planning Guillaume Landry, Mark R. Rivard, Jeffrey F. Williamson, and Frank Verhaegen Monte Carlo as a QA Tool for Advanced Radiation Therapy JinSheng Li and C.-M. Charlie Ma Electrons: Clinical Considerations and Applications Joanna E. Cygler and George X. Ding Photons: Clinical Considerations and Applications Michael K. Fix Monte Carlo Calculations for Proton and Ion Beam Dosimetry Hugo Palmans Protons: Clinical Considerations and Applications Harald Paganetti Application of Monte Carlo Methods to Radionuclide Therapy Michael Ljungberg Monte Carlo for Kilovoltage and Megavoltage Imaging George X. Ding and Andrew Fielding Monte Carlo Calculations for PET-Based Treatment Verification of Ion Beam Therapy Katia Parodi Monte Carlo Studies of Prompt Gamma Emission and of Proton Radiography/Proton-CT Joao Seco, Jerimy C. Polf, and Katia Parodi Monte Carlo for Treatment Device Design Bruce A. Faddegon GPU-Based Fast Monte Carlo Simulation for Radiotherapy Dose Calculation Xun Jia, Sami Hissoiny, and Steve B. Jiang Monte Carlo for Shielding of Radiotherapy Facilities Peter J. Biggs and Stephen F. Kry Index","{'model': 'tldr@v2.0.0', 'text': 'The Monte Carlo Fundamentals History of Monte Carlo and Monte Carlo as a QA Tool for Advanced Radiation Therapy and Applications are presented.'}",
-the effect of probing – prompting technique on reading achievement,Rizqi Nadia Putri,"The fact that most of the students in Junior High School are lack of vocabulary, lazy in reading a long text and low motivation in reading textbooks can be solved by choosing an appropriate method in teaching reading. The aimed of this research was to find out whether the probing-prompting technique was effective for the students' reading achievement. The word “Probing” means research and explore, meanwhile “prompting” means push and guide. In probing-prompting, students are required to answer the question by their own words or sentences. Two groups design was applied, and the participants were eighth students of Junior High School in Tegal. We used reading testing in collecting the data, pre and post-test were conducted. And the data analysis calculated using SPSS program. Based on data analysis using an independent sample t-test, it was revealed that the t-test was higher than t-table (4.332>1.672). We state a conclusion; the use of probing prompting technique is effective because students who taught using probing-prompting technique perform better achievement on reading. ",,https://jurnal.untidar.ac.id/index.php/metathesis/article/download/1699/1251
-probing-prompting based on ethnomathematics learning model: the effect on mathematical communication skill,Sitti Hartinah,"Communication skills on mathematics should be mastered by students to involve in mathematics learning process. T his study aims to determine the effect of the Probing-Prompting based on Ethnomatematics learning model on the impact on mathematical communication skills. This research is an experimental with simple random sampling technique . T he instrument used to collect data is a test of mathematical communication skills. Test trials include content validity, level of difficulty, difference in power, and reliability. For the normality test using the Liliefors-test, the homogeneity test uses a similarity test of two variances. The hypothesis test used by t-test. The results show that there is an influence from the Probing-Prompting based on Ethnomatematics learning model on mathematical communication skills.","{'model': 'tldr@v2.0.0', 'text': 'The results show that there is an influence from the Probing-Prompting based on Ethnomatematics learning model on mathematical communication skills.'}",https://dergipark.org.tr/tr/download/article-file/876976
-a systematic review of artificial intelligence techniques in cancer prediction and diagnosis,Yogesh Kumar,,"{'model': 'tldr@v2.0.0', 'text': 'An efficient search was performed and included the research articles that employed AI-based learning approaches for cancer prediction using conventional machine and deep learning-based classifications between 2009 and 2021.'}",https://link.springer.com/content/pdf/10.1007/s11831-021-09648-w.pdf
-heart disease prediction using machine learning techniques,Devansh Shah,,"{'model': 'tldr@v2.0.0', 'text': 'This research paper presents various attributes related to heart disease, and the model on basis of supervised learning algorithms as Naïve Bayes, decision tree, K-nearest neighbor, and random forest algorithm, using the existing dataset from the Cleveland database of UCI repository of heart disease patients.'}",
-"prompts, not questions: four techniques for crafting better interview protocols",Tomás R. Jiménez,,,
-unpacking mhealth interventions: a systematic review of behavior change techniques used in randomized controlled trials assessing mhealth effectiveness,M. Dugas,"Objective Mobile health interventions have surged in popularity but their implementation varies widely and evidence of effectiveness is mixed. We sought to advance understanding of the diversity of behavior change techniques in mHealth interventions, especially those that leverage advanced mobile technologies. Methods We conducted a systematic review of articles published between 2007 and 2017 in high-impact journals in medicine, medical informatics, and health psychology to identify randomized controlled trials in which the effectiveness of an mobile health intervention was tested. Search terms included a mix of general (e.g. mobile health), hardware (e.g. Android, iPhone), and format (e.g. SMS, application) terms. Results In a systematic review of 21 studies, we found the techniques of personalization, feedback and monitoring, and associations were most commonly used in mobile health interventions, but there remains considerable opportunity to leverage more sophisticated aspects of ubiquitous computing. We found that prompts and cues were the most common behavior change techniques used in effective trials, but there was notable overlap in behavior change techniques used in ineffective trials. Conclusions Our results identify techniques that are commonly used in mobile health interventions and highlight pathways to advance the science of mobile health.","{'model': 'tldr@v2.0.0', 'text': 'The techniques of personalization, feedback and monitoring, and associations were most commonly used in mobile health interventions, but there remains considerable opportunity to leverage more sophisticated aspects of ubiquitous computing.'}",https://journals.sagepub.com/doi/pdf/10.1177/2055207620905411
-behaviour change techniques associated with smoking cessation in intervention and comparator groups of randomised controlled trials: a systematic review and meta-regression.,Nicola Black,"AIMS To estimate the strengths of associations between use of behaviour change techniques (BCTs) and clusters of BCTs in behavioural smoking cessation interventions and comparators with smoking cessation rates. METHOD Systematic review and meta-regression of biochemically verified smoking cessation rates on BCTs in interventions and comparators in randomised controlled trials, adjusting for a priori defined potential confounding variables, together with moderation analyses. Studies were drawn from the Cochrane Tobacco Addiction Group Specialised Register. Data were extracted from published and unpublished (i.e., obtained from study authors) study materials by two independent coders. Adequately described intervention (k = 143) and comparator (k = 92) groups were included in the analyses (N = 43992 participants). Using bivariate mixed-effects meta-regressions, while controlling for key a priori confounders, we regressed smoking cessation on a) three BCT groupings consistent with dual-process theory (i.e., associative, reflective motivational, and self-regulatory), b) 17 expert-derived BCT groupings (i.e., BCT taxonomy v1 clusters), and c) individual BCTs from the BCT taxonomy v1. RESULTS Amongst person-delivered interventions, higher smoking cessation rates were predicted by BCTs targeting associative and self-regulatory processes (B = 0.034-0.041, p < .05), and by three individual BCTs (prompting commitment, social reward, identity associated with changed behaviour), Amongst written interventions, BCTs targeting taxonomy cluster 10a (rewards) predicted higher smoking cessation (B = 0.394, p < .05). Moderation effects were observed for nicotine dependence, mental health status, and mode of delivery. CONCLUSIONS Amongst person-delivered behavioural smoking cessation interventions, specific behaviour change techniques and clusters of techniques are associated with higher success rates.","{'model': 'tldr@v2.0.0', 'text': 'Amongst person-delivered behavioural smoking cessation interventions, specific behaviour change techniques and clusters of techniques are associated with higher success rates and Moderation effects were observed for nicotine dependence, mental health status, and mode of delivery.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/add.15056
-a systematic review of tactile prompting in teacher education,Andrew M. Markelz,"Researchers are exploring innovative technologies to prepare educators. A systematic literature review was conducted to study the role of tactile prompting (i.e., vibratory signals on a schedule) in teacher education. Ten studies meeting inclusion criteria were synthesized and analyzed based on how tactile prompting is utilized, effects of tactile prompting, and quality of research. Findings suggest studies with tactile prompting as the primary independent variable, or condition within a phase, demonstrated medium to very high Tau-U single-case effect sizes (range = 0.54-1.00) and high social acceptability on classroom management teaching behaviors. Simplistic prompting devices reduce cost, training, and anxiety barriers to adoption; however, innovations in wearable technology provide opportunities for a more robust behavior modification tool. Research suggests tactile prompting reduces cognitive load and provides multiple opportunities to practice, which has value in time-constrained preparation programs. Further exploration of this technique in teacher education is warranted.",,
-how to use the nominal group and delphi techniques,S. McMillan,,"{'model': 'tldr@v2.0.0', 'text': 'An overview of the NGT and Delphi technique is provided, including the steps involved and the types of research questions best suited to each method, with examples from the pharmacy literature.'}",https://link.springer.com/content/pdf/10.1007%2Fs11096-016-0257-x.pdf
-evaluation of microencapsulation techniques for micp bacterial spores applied in self-healing concrete,Wiboonluk Pungrasmi,,"{'model': 'tldr@v2.0.0', 'text': 'It is revealed that freeze drying has a high potential as a microencapsulation technique for application to self-healing concrete technology.'}",https://www.nature.com/articles/s41598-019-49002-6.pdf
-the mathematical communication ability based on gender difference on students of 11th grade by using problem-based learning model assisted by probing prompting technique,Ishmah Khairina Zakiri,"This study aims to test the MCA (Mathematical Communication Ability) in Problem Based Learning (PBL) model assisted Probing Prompting Technique (PBL PP), to achieve learning mastery, to know the average ratio of MCA of students on learning model of PBL PP and PBL only, to differ the MCA of male students with female students in PBL PP, and to describe MCA of students based on gender difference. The method used in this research is mix method. While the population is science program students of XI grade at MAN 2 Semarang. The sample is chosen with random sampling. In addition, students of XI IPA 1 are as experiment class, while students of XI IPA 2 as control class. The qualitative research subjects in this study are six students, for each is taken three from every male and female students in experiment class. The results of this research are the MCA’s students with PBL PP learning reaches the learning mastery, the average ratio of MCA in PBL PP is better than PBL model, the MCA of male and female students have no significant difference, and there are some differences of MCA in each indicators based on gender difference.",,
-behavior change techniques implemented in electronic lifestyle activity monitors: a systematic content analysis,E. Lyons,"Background Electronic activity monitors (such as those manufactured by Fitbit, Jawbone, and Nike) improve on standard pedometers by providing automated feedback and interactive behavior change tools via mobile device or personal computer. These monitors are commercially popular and show promise for use in public health interventions. However, little is known about the content of their feedback applications and how individual monitors may differ from one another. Objective The purpose of this study was to describe the behavior change techniques implemented in commercially available electronic activity monitors. Methods Electronic activity monitors (N=13) were systematically identified and tested by 3 trained coders for at least 1 week each. All monitors measured lifestyle physical activity and provided feedback via an app (computer or mobile). Coding was based on a hierarchical list of 93 behavior change techniques. Further coding of potentially effective techniques and adherence to theory-based recommendations were based on findings from meta-analyses and meta-regressions in the research literature. Results All monitors provided tools for self-monitoring, feedback, and environmental change by definition. The next most prevalent techniques (13 out of 13 monitors) were goal-setting and emphasizing discrepancy between current and goal behavior. Review of behavioral goals, social support, social comparison, prompts/cues, rewards, and a focus on past success were found in more than half of the systems. The monitors included a range of 5-10 of 14 total techniques identified from the research literature as potentially effective. Most of the monitors included goal-setting, self-monitoring, and feedback content that closely matched recommendations from social cognitive theory. Conclusions Electronic activity monitors contain a wide range of behavior change techniques typically used in clinical behavioral interventions. Thus, the monitors may represent a medium by which these interventions could be translated for widespread use. This technology has broad applications for use in clinical, public health, and rehabilitation settings.","{'model': 'tldr@v2.0.0', 'text': 'Electronic activity monitors contain a wide range of behavior change techniques typically used in clinical behavioral interventions, and may represent a medium by which these interventions could be translated for widespread use.'}",https://www.jmir.org/2014/8/e192/PDF
-review of modern techniques for the assessment of skin hydration,M. Qassem,"Skin hydration is a complex process that influences the physical and mechanical properties of skin. Various technologies have emerged over the years to assess this parameter, with the current standard being electrical probe-based instruments. Nevertheless, their inability to provide detailed information has prompted the use of sophisticated spectroscopic and imaging methodologies, which are capable of in-depth skin analysis that includes structural and composition details. Modern imaging and spectroscopic techniques have transformed skin research in the dermatological and cosmetics disciplines, and are now commonly employed in conjunction with traditional methods for comprehensive assessment of both healthy and pathological skin. This article reviews current techniques employed in measuring skin hydration, and gives an account on their principle of operation and applications in skin-related research.","{'model': 'tldr@v2.0.0', 'text': 'Current techniques employed in measuring skin hydration are reviewed, and an account on their principle of operation and applications in skin-related research is given.'}",https://www.mdpi.com/2079-9284/6/1/19/pdf?version=1552129987
-classifiers combination techniques: a comprehensive review,M. Mohandes,"In critical applications, such as medical diagnosis, security related systems, and so on, the cost or risk of action taking based on incorrect classification can be very high. Hence, combining expert opinions before taking decision can substantially increase the reliability of such systems. Such pattern recognition systems base their final decision on evidence collected from different classifiers. Such evidence can be of data type, feature type, or classifier type. Common problems in pattern recognition, such as curse of dimensionality, and small sample data size, among others, have also prompted researchers into seeking new approaches for combining evidences. This paper presents a criteria-based framework for multi-classifiers combination techniques and their areas of applications. The criteria discussed here include levels of combination, types of thresholding, adaptiveness of the combination, and ensemble-based approaches. The strengths and weaknesses of each of these categories are discussed in details. Following this analysis, we provide our perspective on the outlook of this area of research and open problems. The lack of a well-formulated theoretical framework for analyzing the performance of combination techniques is shown to provide a fertile ground for further research. In addition to summarizing the existing work, this paper also updates and complements the latest developments in this area of research.","{'model': 'tldr@v2.0.0', 'text': 'A criteria-based framework for multi-classifiers combination techniques and their areas of applications is presented and the lack of a well-formulated theoretical framework for analyzing the performance of combination techniques is shown to provide a fertile ground for further research.'}",
-security techniques for the electronic health records,C. Kruse,,"{'model': 'tldr@v2.0.0', 'text': 'This review seeks to analyze and discuss prominent security techniques for healthcare organizations seeking to adopt a secure electronic health records system using PubMed, CINAHL, and ProQuest Nursing and Allied Health Source as sources.'}",https://link.springer.com/content/pdf/10.1007%2Fs10916-017-0778-4.pdf
-behavior change techniques for increasing physical activity in cancer survivors: a systematic review and meta-analysis of randomized controlled trials,E. Finne,"Purpose The purpose of this systematic review and meta-analysis is to investigate how physical activity (PA) can be effectively promoted in cancer survivors. The effect of PA-promoting interventions in general, behavior change techniques (BCTs), and further variables as moderators in particular are evaluated. Methods This study included randomized controlled trials of lifestyle interventions aiming at an increase in PA that can be carried out independently at home, published by December 2016, for adults diagnosed with cancer after completion of the main treatment. Primary outcomes were subjective and objective measures of PA prior to and immediately after the intervention. Meta-analysis and meta-regression were used to estimate effect sizes (ES) in terms of standardized mean differences, variation between ES in terms of heterogeneity indices (I2), and moderator effects in terms of regression coefficients. Results This study included 30 studies containing 45 ES with an overall significant small positive effect size of 0.28 (95% confidence interval=0.18–0.37) on PA, and I2=54.29%. The BCTs Prompts, Reduce prompts, Graded tasks, Non-specific reward, and Social reward were significantly related to larger effects, while Information about health consequences and Information about emotional consequences, as well as Social comparison were related to smaller ES. The number of BCTs per intervention did not predict PA effects. Interventions based on the Theory of Planned Behavior were associated with smaller ES, and interventions with a home-based setting component were associated with larger ES. Neither the duration of the intervention nor the methodological quality explained differences in ES. Conclusion Certain BCTs were associated with an increase of PA in cancer survivors. Interventions relying on BCTs congruent with (social) learning theory such as using prompts and rewards could be especially successful in this target group. However, large parts of between-study heterogeneity in ES remained unexplained. Further primary studies should directly compare specific BCTs and their combinations.","{'model': 'tldr@v2.0.0', 'text': 'Interventions relying on BCTs congruent with (social) learning theory such as using prompts and rewards could be especially successful in this target group of cancer survivors, however, large parts of between-study heterogeneity in ES remained unexplained.'}",https://www.dovepress.com/getfile.php?fileID=45752
-problematising ethnography and case study: reflections on using ethnographic techniques and researcher positioning,M. Parker-Jenkins,"ABSTRACT This paper was prompted by the question, what do we mean by conducting ‘ethnography’? Is it in fact ‘case study’ drawing on ethnographic techniques? My contention is that in many cases, researchers are not actually conducting ethnography as understood within a traditional sense but rather are engaging in case study, drawing on ethnographic techniques. Does that matter you might ask? Well it determines what we can expect to discover from a research project in terms of results and the unearthing of deeper complexities. I frame the discussion around a set of closely related issues, namely ethnography, case study and researcher positioning, drawing on ethnographic techniques and fieldwork relations. The original contribution of the piece and overall argument is that research can represent a hybrid form, and based on my own research experience, I propose a new term ‘ethno-case study’ that has advantages of both ethnography and case study.",,
-what are the most effective techniques in changing obese individuals’ physical activity self-efficacy and behaviour: a systematic review and meta-analysis,E. Olander,,"{'model': 'tldr@v2.0.0', 'text': 'A non-significant relationship was found between change in self-efficacy and change in physical activity (Spearman’s Rho”=\u2009−0.18 p\u2009=\u202c0.72) and the majority of techniques increased physical activity behaviour, without having discernible effects on self- efficacy.'}",https://ijbnpa.biomedcentral.com/counter/pdf/10.1186/1479-5868-10-29
-do physical activity and dietary smartphone applications incorporate evidence-based behaviour change techniques?,A. Direito,,"{'model': 'tldr@v2.0.0', 'text': 'Presence of BCTs varied by app type and price; however, B CTs associated with increased intervention effectiveness were in general more common in paid apps.'}",https://bmcpublichealth.biomedcentral.com/counter/pdf/10.1186/1471-2458-14-646
-machine learning techniques for quality control in high conformance manufacturing environment,Carlos A. Escobar,"In today’s highly competitive global market, winning requires near-perfect quality. Although most mature organizations operate their processes at very low defects per million opportunities, customers expect completely defect-free products. Therefore, the prompt detection of rare quality events has become an issue of paramount importance and an opportunity for manufacturing companies to move quality standards forward. This article presents the learning process and pattern recognition strategy for a knowledge-based intelligent supervisory system, in which the main goal is the detection of rare quality events. Defect detection is formulated as a binary classification problem. The l1-regularized logistic regression is used as the learning algorithm for the classification task and to select the features that contain the most relevant information about the quality of the process. The proposed strategy is supported by the novelty of a hybrid feature elimination algorithm and optimal classification threshold search algorithm. According to experimental results, 100% of defects can be detected effectively.","{'model': 'tldr@v2.0.0', 'text': 'The learning process and pattern recognition strategy for a knowledge-based intelligent supervisory system, in which the main goal is the detection of rare quality events, is presented and 100% of defects can be detected effectively.'}",https://journals.sagepub.com/doi/pdf/10.1177/1687814018755519
-penerapan teknik probing-prompting dalam pembelajaran matematika siswa kelas viii mtsn lubuk buaya padang,Yuriska Mandasari,"The ability of mathematical communication is the important ability in the science development. The ability of mathematical communication is expected can increase simultaneously by the development of the science and technology. Nevertheless, the implementation of mathematical learning still focuses to the teacher, this method makes students cannot develop their ability in mathematical communication. Their report study in mathematics is still under minimum completeness criteria (KKM), namely 75. One of the recommended effort to deal with this condition is the using probing-prompting technique. The aim of this research is knowing the impact of the enhancement of student’s ability in mathematical communication using probing-prompting technique rather than are compared with convensional learning technique. The result of this research show that student’s ability of mathematical communication using probing-prompting technique is better than convensional technique with α = 0,05. To conclude, probing-prompting technique can increase student is ability in mathematical communication. Key words : mathematical communication ability, probing-prompting technique, convensional learning","{'model': 'tldr@v2.0.0', 'text': 'The result of this research show that student’s ability of mathematical communication using probing-prompting technique is better than convensional technique with α = 0,05.'}",
-pengaruh model pembelajaran inkuiri terbimbing dengan teknik probing-prompting terhadap aktivitas dan hasil belajar fisika di sma,Siscawati Rizki Lasmo,"This research concentrated of guided inquiry model accompanied with probing-prompting technique for activity and learning outcomes physisics in SMA. The purposes of this research were to study the effect of using inquiry model accompanied with probing-prompting technique to student's learning activities and kognitif student's achievement. The type of this research was an experimental research by post-test control group design. Population of the research was all student's at SMAN 1 Pakusari. The technique to collect data was documentation, observation, interview, and tests. The analysis result description of student learning activities for experiment class equals to 85.80 % is in active criteria. The analysis data of kognitif student's achievement by using independent sample t-test for significant value (1-tailed) was 0,018(≤0,05), it's mean the kognitif students' achievement for experiment class is better than control class (Ha accepted and Ho refused). Conclusion of this research are : (1) student's learning activities using guided inquiry model accompanied with probing-prompting technique is in active criteria which percentage level of activity is 85.80%. (2) Guided Inquiry Model accompanied with probing-prompting technique has significant effect of learning outcomes physisics in SMA. Keywords: guided inquiry model, probing-prompting technique",,
-structured analytic techniques for intelligence analysis,R. Heuer,"This book takes the relatively new concept of structured analytic techniques, defines its place in a taxonomy of analytic methods, and moves it a giant leap forward. It describes 50 techniques that are divided into eight categories. There are techniques for: Decomposition and Visualization Idea GenerationScenarios and IndicatorsHypothesis Generation and TestingCause and Effect Challenge AnalysisConflict ManagementDecision Support Each structured technique involves a step-by-step process that externalizes an individual analyst s thinking in a manner that makes it readily apparent to others, thereby enabling it to be shared, built on, and easily critiqued by others. This structured and transparent process combined with the intuitive input of subject matter experts is expected to reduce the risk of analytic error. Our current high tech, global environment increasingly requires collaboration between analysts with different areas of expertise and analysts representing different organizational perspectives. Structured analytic techniques are the ideal process for guiding the interaction of analysts within a small team or group. Each step in a technique prompts relevant discussion within the team, and such discussion generates and evaluates substantially more divergent information and more new ideas than a team that does not use a structured process. By defining the domain of structured analytic techniques, providing a manual for using and teaching these techniques, and outlining procedures for evaluating and validating these techniques, this book lays a common ground for continuing improvement of how analysis is done. These techniques are especially needed in the field of intelligence analysis where analysts typically deal with incomplete, ambiguous and sometimes deceptive information. However, these practical tools for analysis are also useful in a wide variety of professions including law enforcement, medicine, finance, and business.","{'model': 'tldr@v2.0.0', 'text': 'This book takes the relatively new concept of structured analytic techniques, defines its place in a taxonomy of analytic methods, and moves it a giant leap forward by providing a manual for using and teaching these techniques, and outlining procedures for evaluating and validating these techniques.'}",
-a survey of location inference techniques on twitter,Oluwaseun Ajao,"The increasing popularity of the social networking service, Twitter, has made it more involved in day-to-day communications, strengthening social relationships and information dissemination. Conversations on Twitter are now being explored as indicators within early warning systems to alert of imminent natural disasters such as earthquakes and aid prompt emergency responses to crime. Producers are privileged to have limitless access to market perception from consumer comments on social media and microblogs. Targeted advertising can be made more effective based on user profile information such as demography, interests and location. While these applications have proven beneficial, the ability to effectively infer the location of Twitter users has even more immense value. However, accurately identifying where a message originated from or an author’s location remains a challenge, thus essentially driving research in that regard. In this paper, we survey a range of techniques applied to infer the location of Twitter users from inception to state of the art. We find significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.","{'model': 'tldr@v2.0.0', 'text': 'A range of techniques applied to infer the location of Twitter users from inception to state of the art are surveyed, finding significant improvements over time in the granularity levels and better accuracy with results driven by refinements to algorithms and inclusion of more spatial features.'}",http://shura.shu.ac.uk/23769/1/Ajao_survey_location_%28VoR%29.pdf
-using a behaviour change techniques taxonomy to identify active ingredients within trials of implementation interventions for diabetes care,J. Presseau,,"{'model': 'tldr@v2.0.0', 'text': 'The recently developed Behaviour Change Techniques Taxonomy version 1 (BCTTv1) can be used to characterise the active ingredients in trials of implementation interventions and provides specificity of content beyond what is given by broader intervention labels.'}",https://implementationscience.biomedcentral.com/counter/pdf/10.1186/s13012-015-0248-7
-rapid tests and urine sampling techniques for the diagnosis of urinary tract infection ( uti ) in children under five years : a systematic review,P. Whiting,"Background: Urinary tract infection (UTI) is one of the most common sources of infection in children under five. Prompt diagnosis and treatment is important to reduce the risk of renal scarring. Rapid, cost-effective, methods of UTI diagnosis are required as an alternative to culture. Methods: We conducted a systematic review to determine the diagnostic accuracy of rapid tests for detecting UTI in children under five years of age. Results: The evidence supports the use of dipstick positive for both leukocyte esterase and nitrite (pooled LR+ = 28.2, 95% CI: 17.3, 46.0) or microscopy positive for both pyuria and bacteriuria (pooled LR+ = 37.0, 95% CI: 11.0, 125.9) to rule in UTI. Similarly dipstick negative for both LE and nitrite (Pooled LR= 0.20, 95% CI: 0.16, 0.26) or microscopy negative for both pyuria and bacteriuria (Pooled LR= 0.11, 95% CI: 0.05, 0.23) can be used to rule out UTI. A test for glucose showed promise in potty-trained children. However, all studies were over 30 years old. Further evaluation of this test may be useful. Conclusion: Dipstick negative for both LE and nitrite or microscopic analysis negative for both pyuria and bacteriuria of a clean voided urine, bag, or nappy/pad specimen may reasonably be used to rule out UTI. These patients can then reasonably be excluded from further investigation, without the need for confirmatory culture. Similarly, combinations of positive tests could be used to rule in UTI, and trigger further investigation. Background Urinary tract infection (UTI) is one of the most common sources of infection in children under 5. In a small proportion of children UTI may lead to renal scarring [1,2]. This outcome of infection is of concern as it is associated with significant future complications and ultimately with end stage renal disease[3]. Prompt diagnosis and treatment is therefore important to reduce the risk of future scarring. Published: 05 April 2005 BMC Pediatrics 2005, 5:4 doi:10.1186/1471-2431-5-4 Received: 05 October 2004 Accepted: 05 April 2005 This article is available from: http://www.biomedcentral.com/1471-2431/5/4 © 2005 Whiting et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BMC Pediatrics 2005, 5:4 http://www.biomedcentral.com/1471-2431/5/4","{'model': 'tldr@v2.0.0', 'text': 'A systematic review of rapid tests for detecting UTI in children under five years of age found the evidence supports the use of dipstick positive for both leukocyte esterase and nitrite or microscopic analysis negative for both pyuria and bacteriuria of a clean voided urine, bag, or nappy/pad specimen to rule out UTI.'}",
-task effects on linguistic complexity and accuracy: a large-scale learner corpus analysis employing natural language processing techniques,T. Alexopoulou,"Large-scale learner corpora collected from online language learning platforms, such as the EF-Cambridge Open Language Database (EFCAMDAT), provide opportunities to analyze learner data at an unprecedented scale. However, interpreting the learner language in such corpora requires a precise understanding of tasks: Howdoes the prompt and input of a task and its functional requirements influence task-based linguistic performance? This question is vital for making large-scale task-based corpora fruitful for second language acquisition research. We explore the issue through an analysis of selected tasks in EFCAMDAT and the complexity and accuracy of the language they elicit.","{'model': 'tldr@v2.0.0', 'text': 'The issue of how the prompt and input of a task and its functional requirements influence task-based linguistic performance is explored through an analysis of selected tasks in EFCAMDAT and the complexity and accuracy of the language they elicit.'}",https://eprints.lancs.ac.uk/id/eprint/83702/3/Alexopoulou_et_al_in_press.pdf
-behavior change techniques in mhealth apps for the mental and physical health of employees: systematic assessment,E. D. de Korte,"Background Employees remain at risk of developing physical and mental health problems. To improve the lifestyle, health, and productivity many workplace interventions have been developed. However, not all of these interventions are effective. Mobile and wireless technology to support health behavior change (mobile health [mHealth] apps) is a promising, but relatively new domain for the occupational setting. Research on mHealth apps for the mental and physical health of employees is scarce. Interventions are more likely to be useful if they are rooted in health behavior change theory. Evaluating the presence of specific combinations of behavior change techniques (BCTs) in mHealth apps might be used as an indicator of potential quality and effectiveness. Objective The aim of this study was to assess whether mHealth apps for the mental and physical health of employees incorporate BCTs and, if so, which BCTs can be identified and which combinations of BCTs are present. Methods An assessment was made of apps aiming to reduce the risk of physical and psychosocial work demands and to promote a healthy lifestyle for employees. A systematic search was performed in iTunes and Google Play. Forty-five apps were screened and downloaded. BCTs were identified using a taxonomy applied in similar reviews. The mean and ranges were calculated. Results On average, the apps included 7 of the 26 BCTs (range 2-18). Techniques such as “provide feedback on performance,” “provide information about behavior-health link,” and “provide instruction” were used most frequently. Techniques that were used least were “relapse prevention,” “prompt self-talk,” “use follow-up prompts,” and “provide information about others’ approval.” “Stress management,” “prompt identification as a role model,” and “agree on behavioral contract” were not used by any of the apps. The combination “provide information about behavior-health link” with “prompt intention formation” was found in 7/45 (16%) apps. The combination “provide information about behavior-health link” with “provide information on consequences,” and “use follow-up prompts” was found in 2 (4%) apps. These combinations indicated potential effectiveness. The least potentially effective combination “provide feedback on performance” without “provide instruction” was found in 13 (29%) apps. Conclusions Apps for the occupational setting might be substantially improved to increase potential since results showed a limited presence of BCTs in general, limited use of potentially successful combinations of BCTs in apps, and use of potentially unsuccessful combinations of BCTs. Increasing knowledge on the effectiveness of BCTs in apps might be used to develop guidelines for app developers and selection criteria for companies and individuals. Also, this might contribute to decreasing the burden of work-related diseases. To achieve this, app developers, health behavior change professionals, experts on physical and mental health, and end-users should collaborate when developing apps for the working context.","{'model': 'tldr@v2.0.0', 'text': 'App developers, health behavior change professionals, experts on physical and mental health, and end-users should collaborate when developing apps for the working context to increase potential since results showed a limited presence of BCTs in general, limited use of potentially successful combinations of B CTs in apps, and use of possibly unsuccessful combinations ofBCTs.'}",https://mhealth.jmir.org/2018/10/e167/PDF
-a large-scale empirical comparison of static and dynamic test case prioritization techniques,Qi Luo,"The large body of existing research in Test Case Prioritization (TCP) techniques, can be broadly classified into two categories: dynamic techniques (that rely on run-time execution information) and static techniques (that operate directly on source and test code). Absent from this current body of work is a comprehensive study aimed at understanding and evaluating the static approaches and comparing them to dynamic approaches on a large set of projects. In this work, we perform the first extensive study aimed at empirically evaluating four static TCP techniques comparing them with state-of-research dynamic TCP techniques at different test-case granularities (e.g., method and class-level) in terms of effectiveness, efficiency and similarity of faults detected. This study was performed on 30 real-word Java programs encompassing 431 KLoC. In terms of effectiveness, we find that the static call-graph-based technique outperforms the other static techniques at test-class level, but the topic-model-based technique performs better at test-method level. In terms of efficiency, the static call-graph-based technique is also the most efficient when compared to other static techniques. When examining the similarity of faults detected for the four static techniques compared to the four dynamic ones, we find that on average, the faults uncovered by these two groups of techniques are quite dissimilar, with the top 10% of test cases agreeing on only 25% - 30% of detected faults. This prompts further research into the severity/importance of faults uncovered by these techniques, and into the potential for combining static and dynamic information for more effective approaches.","{'model': 'tldr@v2.0.0', 'text': 'This first extensive study aimed at empirically evaluating four static TCP techniques comparing them with state-of-research dynamic TCP techniques at different test-case granularities in terms of effectiveness, efficiency and similarity of faults detected finds that the static call-graph-based technique outperforms the other static techniques at test-class level, but the topic-model-based techniques performs better atTest-method level.'}",https://arxiv.org/pdf/1801.05917
-using progressive video prompting to teach students with moderate intellectual disability to shoot a basketball,Ya-yu Lo,"The current study examined the effects of a modified video prompting procedure, namely progressive video prompting, to increase technique accuracy of shooting a basketball in the school gymnasium of three 11th-grade students with moderate intellectual disability. The intervention involved participants viewing video clips of an adult model who showed progressively chucked steps for making a free throw. We used a single-case, multiple probe across participants design to evaluate the intervention effects. The results of this study showed that all three participants increased the number of steps performed correctly and maintained the skill at the 1-week and 2-week maintenance check without the video viewing. Implications for practice and future research related to video prompting variations are discussed.",,
-cyber-attack modeling analysis techniques: an overview,Hamad Al-Mohannadi,"Cyber attack is a sensitive issue in the world of Internet security. Governments and business organisations around the world are providing enormous effort to secure their data. They are using various types of tools and techniques to keep the business running, while adversaries are trying to breach security and send malicious software such as botnets, viruses, trojans etc., to access valuable data. Everyday the situation is getting worse because of new types of malware emerging to attack networks. It is important to understand those attacks both before and after they happen in order to provide better security to our systems. Understanding attack models provide more insight into network vulnerability, which in turn can be used to protect the network from future attacks. In the cyber security world, it is difficult to predict a potential attack without understanding the vulnerability of the network. So, it is important to analyse the network to identify top possible vulnerability list, which will give an intuitive idea to protect the network. Also, handling an ongoing attack poses significant risk on the network and valuable data, where prompt action is necessary. Proper utilisation of attack modelling techniques provide advance planning, which can be implemented rapidly during an ongoing attack event. This paper aims to analyse various types of existing attack modelling techniques to understand the vulnerability of the network, and the behaviour and goals of the adversary. The ultimate goal is to handle cyber attack in efficient manner using attack modelling techniques.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to analyse various types of existing attack modelling techniques to understand the vulnerability of the network, and the behaviour and goals of the adversary.'}",https://bradscholars.brad.ac.uk/bitstream/10454/10703/2/hamad_ICI.pdf
-handling imbalanced and overlapping classes in smart environments prompting dataset,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'In this chapter, intersection between the problems of imbalanced class and overlapping classes is explored from the perspective of Smart Environments as the application domain and a novel clustering-based under-sampling technique is proposed.'}",
-fluoride removal from water by various techniques : review,S. Waghmare,"This paper depicts the fluoride removal from drinking water can be accomplished by different methods, for example, coagulation-precipitation, membrane separation process, ion exchange, adsorption techniques and so on. Among these procedures, membrane and ion exchange processes are not extremely regular because of its high establishment and support price. Other two tecniques are extremely regular in India . Nalgonda procedure is one of the well known strategies generally utilized for defluoridation of water as a part of developing nations, for example, India, Kenya, Senegal and Tanzania. Among different tecniques utilized for defluoridation of water, the adsorption procedure is broadly utilized and offers acceptable results and is by all accounts more appealing technique for the removal of fluoride regarding expense, straightforwardness of outline and operation. It is apparent from the literature study that different methods have indicated novel potential for the removal of fluoride. Be that as it may, in any case there is a need to figure out the reasonable utility of such developed procedures on a business scale, prompting the change of contamination control.",,
-combinations of techniques that effectively change health behavior: evidence from meta-cart analysis.,E. Dusseldorp,"OBJECTIVE Many health-promoting interventions combine multiple behavior change techniques (BCTs) to maximize effectiveness. Although, in theory, BCTs can amplify each other, the available meta-analyses have not been able to identify specific combinations of techniques that provide synergistic effects. This study overcomes some of the shortcomings in the current methodology by applying classification and regression trees (CART) to meta-analytic data in a special way, referred to as Meta-CART. The aim was to identify particular combinations of BCTs that explain intervention success. METHOD A reanalysis of data from Michie, Abraham, Whittington, McAteer, and Gupta (2009) was performed. These data included effect sizes from 122 interventions targeted at physical activity and healthy eating, and the coding of the interventions into 26 BCTs. A CART analysis was performed using the BCTs as predictors and treatment success (i.e., effect size) as outcome. A subgroup meta-analysis using a mixed effects model was performed to compare the treatment effect in the subgroups found by CART. RESULTS Meta-CART identified the following most effective combinations: Provide information about behavior-health link with Prompt intention formation (mean effect size ḡ = 0.46), and Provide information about behavior-health link with Provide information on consequences and Use of follow-up prompts (ḡ = 0.44). Least effective interventions were those using Provide feedback on performance without using Provide instruction (ḡ = 0.05). CONCLUSIONS Specific combinations of BCTs increase the likelihood of achieving change in health behavior, whereas other combinations decrease this likelihood. Meta-CART successfully identified these combinations and thus provides a viable methodology in the context of meta-analysis.","{'model': 'tldr@v2.0.0', 'text': 'Specific combinations of BCTs increase the likelihood of achieving change in health behavior, whereas other combinations decrease this likelihood, and Meta-CART successfully identified these combinations and thus provides a viable methodology in the context of meta-analysis.'}",http://www.elisedusseldorp.nl/pdf/Dusseldorp2014_HPMetaCART.pdf
-"support for learning from multimedia explanations. a comparison of prompting, signaling, and questioning",H. Garcia-Rodicio,In one experiment 97 undergraduate students learned about plate tectonics from a multimedia presentation involving narrated animations and support in one of four forms. Support in the prompting condition included hints inducing participants to self-explain critical information. The signaling condition included overviews recapping critical information. The questioning condition included questions about critical information and feedback on participants' answers. The control condition included no support. Participants in the questioning condition outperformed those in the rest of conditions in retention and transfer. This means that questioning is a very powerful technique for promoting multimedia learning.,,
-application of machine learning techniques to sentiment analysis,A. Jain,"Today, we live in a ‘data age’. Due to rapid increase in the amount of user-generated data on social media platforms like Twitter, several opportunities and new open doors have been prompted for organizations that endeavour hard to keep a track on customer reviews and opinions about their products. Twitter is a huge fast emergent micro-blogging social networking platform for users to express their views about politics, products sports etc. These views are useful for businesses, government and individuals. Hence, tweets can be used as a valuable source for mining public's opinion. Sentiment analysis is a process of automatically identifying whether a user-generated text expresses positive, negative or neutral opinion about an entity (i.e. product, people, topic, event etc). The objective of this paper is to give step-by-step detail about the process of sentiment analysis on twitter data using machine learning. This paper also provides details of proposed approach for sentiment analysis. This work proposes a Text analysis framework for twitter data using Apache spark and hence is more flexible, fast and scalable. Naïve Bayes and Decision trees machine learning algorithms are used for sentiment analysis in the proposed framework.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Text analysis framework for twitter data using Apache spark and hence is more flexible, fast and scalable and Naïve Bayes and Decision trees machine learning algorithms are used for sentiment analysis in the proposed framework.'}",
-behavior change techniques used to promote walking and cycling,E. Bird,"Objective: Evidence on the effectiveness of walking and cycling interventions is mixed. This may be partly attributable to differences in intervention content, such as the cognitive and behavioral techniques (BCTs) used. Adopting a taxonomy of BCTs, this systematic review addressed two questions: (a) What are the behavior change techniques used in walking and cycling interventions targeted at adults? (b) What characterizes interventions that appear to be associated with changes in walking and cycling in adults? Method: Previous systematic reviews and updated database searches were used to identify controlled studies of individual-level walking and cycling interventions involving adults. Characteristics of intervention design, context, and methods were extracted in addition to outcomes. Intervention content was independently coded according to a 26-item taxonomy of BCTs. Results: Studies of 46 interventions met the inclusion criteria. Twenty-one reported a statistically significant effect on walking and cycling outcomes. Analysis revealed substantial heterogeneity in the vocabulary used to describe intervention content and the number of BCTs coded. “Prompt self-monitoring of behavior” and “prompt intention formation” were the most frequently coded BCTs. Conclusion: Future walking and cycling intervention studies should ensure that all aspects of the intervention are reported in detail. The findings lend support to the inclusion of self-monitoring and intention formation techniques in future walking and cycling intervention design, although further exploration of these and other BCTs is required. Further investigation of the interaction between BCTs and study design characteristics would also be desirable.","{'model': 'tldr@v2.0.0', 'text': 'The findings lend support to the inclusion of self-monitoring and intention formation techniques in future walking and cycling intervention design, although further exploration of these and other BCTs is required.'}",https://europepmc.org/articles/pmc3727344?pdf=render
-effective behaviour change techniques in the prevention and management of childhood obesity,Jilly Martin,,"{'model': 'tldr@v2.0.0', 'text': 'To design effective behaviour change programmes for the prevention and management of childhood obesity, it is recommended to utilise the BCTs identified as effective in this review, including six that may be effective components of future management interventions.'}",https://openaccess.city.ac.uk/id/eprint/14188/3/Supplementary%20Table%202.pdf
-discovering the effects of metacognitive prompts on the sequential structure of srl-processes using process mining techniques,Christoph Sonnenberg,"According to research examining self-regulated learning (SRL), we regard individual regulation as a specific sequence of regulatory activities. Ideally, students perform various learning activities, such as analyzing, monitoring, and evaluating cognitive and motivational aspects during learning. Metacognitive prompts can foster SRL by inducing regulatory activities, which, in turn, improve the learning outcome. However, the specific effects of metacognitive support on the dynamic characteristics of SRL are not understood. Therefore, the aim of our study was to analyze the effects of metacognitive prompts on learning processes and outcomes during a computer-based learning task. Participants of the experimental group (EG, n = 35) were supported by metacognitive prompts, whereas participants of the control group (CG, n = 35) received no support. Data regarding learning processes were obtained by concurrent think-aloud protocols. The EG exhibited significantly more metacognitive learning events than did the CG. Furthermore, these regulatory activities correspond positively with learning outcomes. Process mining techniques were used to analyze sequential patterns. Our findings indicate differences in the process models of the EG and CG and demonstrate the added value of taking the order of learning activities into account by discovering regulatory patterns.",,https://learning-analytics.info/index.php/JLA/article/download/4090/4837
-research and development aspects on chemical preparation techniques of photoanodes for dye sensitized solar cells,N. Asim,"The importance of dye sensitized solar cells (DSSCs) as a low-cost and environmentally friendly photovoltaic (PV) technology has prompted many researchers to improve its efficiency and durability. The realization of these goals is impossible without taking into account the importance of the materials in DSSCs, so the focus on the preparation/deposition methods is essential. These methods can be either chemical or physical. In this study, the chemical applied methods that utilize chemical reaction to synthesize and deposit the materials are covered and categorized according to their gas phase and liquid phase precursors. Film processing techniques that can be used to enhance the materials' properties postpreparation are also included for further evaluation in this study. However, there is a variety of consideration, and certain criteria must be taken into account when selecting a specific deposition method, due to the fact that the fabrication conditions vary and are unoptimized.",,https://downloads.hindawi.com/journals/ijp/2014/518156.pdf
-meningkatkan kemampuan pemahaman dan disposisi matematik siswa sma menggunakan teknik probing prompting,Sekolah Tinggi,"This study was intended to analyze the role of probing prompting technique (PPT) toward students’ mathematical understanding ability and disposition and asociation both of them. This study was a part of a master thesis and a sub-study of a Postgraduate Research Grant from DGHE in 2015. This study was a pretest-postest quasi-experimental control group design involving 66 twelfth-grade students of a senior high school in Pengalengan which were chosen puposively.The instruments of this study are an essay test on mathematical understanding ability, and a mathematical disposition (MD) scale. The study revealed that students getting treatment on PPT attained better grades on mathematical understanding ability and disposition than that of students taught by conventional teaching, though the grades of mathematical understanding were at low level and the grades of mathematical disposition were at fairly good. Students realized difficulties in determining requirements of a concept, compiling mathematical model and solving it concerning graph of function. Also, there was association between mathematical understanding ability and disposition, and students performed positive opinioan toward probing prompting technique. Keyword: Mathematical understanding, mathematical disposition, probing prompting technique",,
-grammatical evolutionary techniques for prompt migraine prediction,Josué Pagán,"The migraine disease is a chronic headache presenting symptomatic crisis that causes high economic costs to the national health services, and impacts negatively on the quality of life of the patients. Even if some patients can feel unspecific symptoms before the onset of the migraine, these only happen randomly and cannot predict the crisis precisely. In our work, we have proved how migraine crisis can be predicted with high accuracy from the physiological variables of the patients, acquired by a non-intrusive Wireless Body Sensor Network. In this paper, we derive alternative models for migraine prediction using Grammatical Evolution techniques. We obtain prediction horizons around 20 minutes, which are sufficient to advance the drug intake and avoid the symptomatic crisis. The robustness of the models with respect to sensor failures has also been tackled to allow the practical implementation in the ambulatory monitoring platform. The achieved models are non linear mathematical expressions with low computing overhead during the run-time execution in the wearable devices.","{'model': 'tldr@v2.0.0', 'text': 'This work has proved how migraine crisis can be predicted with high accuracy from the physiological variables of the patients, acquired by a non-intrusive Wireless Body Sensor Network.'}",
-voyager: an open-ended embodied agent with large language models,Guanzhi Wang,"We introduce Voyager, the first LLM-powered embodied lifelong learning agent in Minecraft that continuously explores the world, acquires diverse skills, and makes novel discoveries without human intervention. Voyager consists of three key components: 1) an automatic curriculum that maximizes exploration, 2) an ever-growing skill library of executable code for storing and retrieving complex behaviors, and 3) a new iterative prompting mechanism that incorporates environment feedback, execution errors, and self-verification for program improvement. Voyager interacts with GPT-4 via blackbox queries, which bypasses the need for model parameter fine-tuning. The skills developed by Voyager are temporally extended, interpretable, and compositional, which compounds the agent's abilities rapidly and alleviates catastrophic forgetting. Empirically, Voyager shows strong in-context lifelong learning capability and exhibits exceptional proficiency in playing Minecraft. It obtains 3.3x more unique items, travels 2.3x longer distances, and unlocks key tech tree milestones up to 15.3x faster than prior SOTA. Voyager is able to utilize the learned skill library in a new Minecraft world to solve novel tasks from scratch, while other techniques struggle to generalize. We open-source our full codebase and prompts at https://voyager.minedojo.org/.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2305.16291
-uptake of behavior change techniques - a key focus for process analysis and for intervention delivery: a comment on hankonen et al.,C. Greaves,,,https://academic.oup.com/abm/article-pdf/49/1/1/21983123/12160_2014_article_9630.pdf
-the flan collection: designing data and methods for effective instruction tuning,S. Longpre,"We study the design decisions of publicly available instruction tuning methods, and break down the development of Flan 2022 (Chung et al., 2022). Through careful ablation studies on the Flan Collection of tasks and methods, we tease apart the effect of design decisions which enable Flan-T5 to outperform prior work by 3-17%+ across evaluation settings. We find task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings (zero-shot, few-shot, and chain-of-thought) actually yields stronger (2%+) performance in all settings. In further experiments, we show Flan-T5 requires less finetuning to converge higher and faster than T5 on single downstream tasks, motivating instruction-tuned models as more computationally-efficient starting checkpoints for new tasks. Finally, to accelerate research on instruction tuning, we make the Flan 2022 collection of datasets, templates, and methods publicly available at https://github.com/google-research/FLAN/tree/main/flan/v2.","{'model': 'tldr@v2.0.0', 'text': 'It is found task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings actually yields stronger performance in all settings.'}",http://arxiv.org/pdf/2301.13688
-identification of behaviour change techniques to reduce excessive alcohol consumption.,S. Michie,"BACKGROUND Interventions to reduce excessive alcohol consumption have a small but important effect, but a better understanding is needed of their 'active ingredients'. AIMS This study aimed to (i) develop a reliable taxonomy of behaviour change techniques (BCTs) used in interventions to reduce excessive alcohol consumption (not to treat alcohol dependence) and (ii) to assess whether use of specific BCTs in brief interventions might be associated with improved effectiveness. METHODS A selection of guidance documents and treatment manuals, identified via expert consultation, were analysed into BCTs by two coders. The resulting taxonomy of BCTs was applied to the Cochrane Review of brief alcohol interventions, and the associations between the BCTs and effectiveness were investigated using meta-regression. FINDINGS Forty-two BCTs were identified, 34 from guidance documents and an additional eight from treatment manuals, with average inter-rater agreement of 80%. Analyses revealed that brief interventions that included the BCT 'prompt self-recording' (P = 0.002) were associated with larger effect sizes. CONCLUSIONS It is possible to identify specific behaviour change techniques reliably in manuals and guidelines for interventions to reduce excessive alcohol consumption. In brief interventions, promoting self-monitoring is associated with improved outcomes. More research is needed to identify other behaviour change techniques or groupings of behaviour change techniques that can produce optimal results in brief interventions and to extend the method to more intensive interventions and treatment of alcohol dependence.","{'model': 'tldr@v2.0.0', 'text': 'It is possible to identify specific behaviour change techniques reliably in manuals and guidelines for interventions to reduce excessive alcohol consumption and, in brief interventions, promoting self-monitoring is associated with improved outcomes.'}",
-the use of neutron analysis techniques for detecting the concentration and distribution of chloride ions in archaeological iron,D. Watkinson,"Chloride (Cl) ions diffuse into iron objects during burial and drive corrosion after excavation. Located under corrosion layers, Cl is inaccessible to many analytical techniques. Neutron analysis offers non-destructive avenues for determining Cl content and distribution in objects. A pilot study used prompt gamma activation analysis (PGAA) and prompt gamma activation imaging (PGAI) to analyse the bulk concentration and longitudinal distribution of Cl in archaeological iron objects. This correlated with the object corrosion rate measured by oxygen consumption, and compared well with Cl measurement using a specific ion meter. High-Cl areas were linked with visible damage to the corrosion layers and attack of the iron core. Neutron techniques have significant advantages in the analysis of archaeological metals, including penetration depth and low detection limits.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study used prompt gamma activation analysis (PGAA) and prompt gammaactivation imaging (PGAI) to analyse the bulk concentration and longitudinal distribution of Cl in archaeological iron objects and found high-Cl areas were linked with visible damage to the corrosion layers and attack of the iron core.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/arcm.12058
-puck: an automated prompting system for smart environments: toward achieving automated prompting—challenges involved,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'This paper uses simple off-the-shelf sensors and learns the timing for prompts based on real data that are collected with volunteer participants in the smart home test bed and proposes a variant of an existing sampling technique, SMOTE, to deal with the class imbalance problem.'}",https://europepmc.org/articles/pmc4215554?pdf=render
-"concepts, instrumentation and techniques of neutron activation analysis",H. Slamene,"Following the discovery of neutron by J. Chadwick in 1932 (Nobel prize, 1935) and the re‐ sults of F. Joliot and I. Curie in 1934, neutron activation analysis was first developed by G. Hevesy and H. Levi in 1936. They used a neutron source (226Ra + Be) and a radiation detec‐ tor (ionization chamber) and promptly recognized that the element Dy (dysprosium) in the sample became highly radioactive after exposure to the neutron source. They showed that the nuclear reaction may be used to determine the elements present in unknown samples by measuring the induced radioactivity.",,
-learning to compose soft prompts for compositional zero-shot learning,Nihal V. Nayak,"We introduce compositional soft prompting (CSP), a parameter-efficient learning technique to improve the zero-shot compositionality of large-scale pretrained vision-language models (VLMs) like CLIP. We develop CSP for compositional zero-shot learning, the task of predicting unseen attribute-object compositions (e.g., old cat and young tiger). VLMs have a flexible text encoder that can represent arbitrary classes as natural language prompts but they often underperform task-specific architectures on the compositional zero-shot benchmark datasets. CSP treats the attributes and objects that define classes as learnable tokens of vocabulary. During training, the vocabulary is tuned to recognize classes that compose tokens in multiple ways (e.g., old cat and white cat). At test time, we recompose the learned attribute-object vocabulary in new combinations to recognize novel classes. We show that CSP outperforms the CLIP on benchmark datasets by an average of 10.9 percentage points on AUC. CSP also outperforms CoOp, a soft prompting method that fine-tunes the prefix context tokens, by an average of 5.8 percentage points on AUC. We perform additional experiments to show that CSP improves generalization to higher-order attribute-attribute-object compositions (e.g., old white cat) and combinations of pretrained attributes and fine-tuned objects. The code is available at https://github.com/BatsResearch/csp.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2204.03574
-generating training data with language models: towards zero-shot language understanding,Yu Meng,"Pretrained language models (PLMs) have demonstrated remarkable performance in various natural language processing tasks: Unidirectional PLMs (e.g., GPT) are well known for their superior text generation capabilities; bidirectional PLMs (e.g., BERT) have been the prominent choice for natural language understanding (NLU) tasks. While both types of models have achieved promising few-shot learning performance, their potential for zero-shot learning has been underexplored. In this paper, we present a simple approach that uses both types of PLMs for fully zero-shot learning of NLU tasks without requiring any task-specific data: A unidirectional PLM generates class-conditioned texts guided by prompts, which are used as the training data for fine-tuning a bidirectional PLM. With quality training data selected based on the generation probability and regularization techniques (label smoothing and temporal ensembling) applied to the fine-tuning stage for better generalization and stability, our approach demonstrates strong performance across seven classification tasks of the GLUE benchmark (e.g., 72.3/73.8 on MNLI-m/mm and 92.8 on SST-2), significantly outperforming zero-shot prompting methods and achieving even comparable results to strong few-shot approaches using 32 training samples per class.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple approach that uses both types of PLMs for fully zero-shot learning of NLU tasks without requiring any task-specific data: a unidirectional PLM generates class-conditioned texts guided by prompts, which are used as the training data for fine-tuning a bidirectionalPLM.'}",
-simultaneous prompting: an instructional strategy for skill acquisition,R. E. Waugh,"Errorless learning is an instructional approach designed to eliminate and/or reduce the number of errors students produce in traditional trial-and-error approaches (Mueller, Palkovic, & Maynard, 2007). Various response prompting strategies are employed to produce errorless learning. Simultaneous prompting is an errorless learning strategy that has a growing body of literature to support its use spanning two decades. This paper provides a comprehensive review of the literature including (a) skills targeted for instruction, (b) populations targeted for instruction, (c) strengths and weaknesses of simultaneous prompting, and (d) future areas of research. Errorless learning is an instructional ap- proach designed to reduce the number of errors students make in traditional trial-and- error approaches (Mueller, Palkovic, & May- nard, 2007). During errorless learning proce- dures stimulus control is transferred from the controlling prompt to the discriminative stim- ulus. This transfer is achieved through the use of response prompting strategies. Response prompting strategies consist of additional in- formation which results in the correct re- sponse being emitted (Wolery, Ault, & Doyle, 1992). The underlying purpose of errorless learning is the transfer of stimulus control from a response prompt to the natural stimu- lus. Wolery and Gast (1984) identified four common response prompting strategies that commonly are employed to transfer stimulus control: (a) most-to-least prompts, (b) least-to- most prompts, (c) graduated guidance, and (d) time delay. Most-to-least prompts consists of employing the most intrusive prompt needed to assist the student in emitting the correct response and fading the intensity of the prompt until the student is correctly re- sponding to the discriminative stimulus inde- pendently. Least-to-most prompts provide the student with an opportunity to respond inde- pendently to the discriminative stimulus. If the student responds incorrectly a prompt is provided which gradually increases in inten- sity until the student responds correctly to the discriminative stimulus. ""Graduated guidance is a technique combining physical guidance and fading in which the physical guidance is systematically and gradually reduced and then faded completely"" (Foxx, 1982, p. 129). Grad- uate guidance relies heavily on the teacher's judgment whether or not a prompt is required or the degree of prompt required. There are two forms of graduated guidance. During one form a teacher shadows a student's movement when teaching a task to provide guidance dur- ing each step or to remove the physical prompt. During a second form of graduated guidance the teacher may provide constant contact but gradually and systematically re- duce the intrusiveness and placement of the prompt (Foxx, 1981; Wolery & Gast, 1984). Time delay is the fourth common response prompting strategy which results in near er- rorless learning by transferring stimulus con- trol from a controlling prompt to the discrim- inative stimulus by inserting a delay between the presentation of the discriminative stimu- lus and the controlling prompt (Snell & Gast, 1981; Touchette, 1971). Two forms of time delay are reported in the literature, progres- sive time delay (PTD) and constant time delay (CTD). During PTD a systematically increased",,
-interventions that involve parents to improve children's weight‐related nutrition intake and activity patterns – what nutrition and activity targets and behaviour change techniques are associated with intervention effectiveness?,R. Golley,"Parent involvement is an important component of obesity prevention interventions. However, the best way to support parents remains unclear. This review identifies interventions targeting parents to improve children's weight status, dietary and/or activity patterns, examines whether intervention content and behaviour change techniques employed are associated with effectiveness. Seventeen studies, in English, 1998–2008, were included. Studies were evaluated by two reviewers for study quality, nutrition/activity content and behaviour change techniques using a validated quality assessment tool and behaviour change technique taxonomy. Study findings favoured intervention effectiveness in 11 of 17 studies. Interventions that were considered effective had similar features: better study quality, parents responsible for participation and implementation, greater parental involvement and inclusion of prompt barrier identification, restructure the home environment, prompt self‐monitoring, prompt specific goal setting behaviour change techniques. Energy intake/density and food choices were more likely to be targeted in effective interventions. The number of lifestyle behaviours targeted did not appear to be associated with effectiveness. Intervention effectiveness was favoured when behaviour change techniques spanned the spectrum of behaviour change process. The review provides guidance for researchers to make informed decisions on how best to utilize resources in interventions to support and engage parents, and highlights a need for improvement in intervention content reporting practices.","{'model': 'tldr@v2.0.0', 'text': ""Interventions targeting parents to improve children's weight status, dietary and/or activity patterns, and whether intervention content and behaviour change techniques employed are associated with effectiveness are examined.""}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/j.1467-789X.2010.00745.x
-"equine castration: a review of techniques, complications and their management",I. Kilcoyne,"Summary Complications associated with castration occur commonly and, although the majority are mild and resolve easily, potentially life-threatening complications can occur. The preoperative identification of risk factors for these complications can help the veterinarian to take the appropriate measures to reduce these risks. However, even with proper surgical technique, complications can occur. Therefore prompt recognition and initiation of appropriate therapy are essential to prevent further morbidity, death or lawsuits.","{'model': 'tldr@v2.0.0', 'text': 'Preoperative identification of risk factors for castration complications can help the veterinarian to take the appropriate measures to reduce these risks, although the majority are mild and resolve easily.'}",
-using video prompting and constant time delay to teach an internet search basic skill to students with intellectual disabilities,D. Zisimopoulos,"We evaluated a video prompting and a constant time delay procedure for teaching three primary school students with moderate intellectual disabilities to access the Internet and download pictures related to participation in a classroom History project. Video clips were used as an antecedent prompt and as an error correction technique within a constant time delay (CTD) procedure. Training, using a 29-step task analysis, was conducted in the students’ special education classrooms. The prompting procedure was introduced in a multiple baseline across subjects design. Video prompting was effective in promoting rapid acquisition of the task for all three students. Following acquisition, observations suggested that the new skill was further generalized to another trainer, novel materials, and to another setting. In a final phase then, video prompting was removed and maintenance in the absence of video prompting was assessed at 1, 3, and 18 weeks. Performance decreased at the last follow-up session, but stabilized at 82.7–89.6% correct. These data suggest that video prompting may be an effective instructional strategy for teaching Internet skills to students with moderate intellectual disabilities.",,
-towards personalized federated learning,A. Tan,"In parallel with the rapid adoption of artificial intelligence (AI) empowered by advances in AI research, there has been growing awareness and concerns of data privacy. Recent significant developments in the data regulation landscape have prompted a seismic shift in interest toward privacy-preserving AI. This has contributed to the popularity of Federated Learning (FL), the leading paradigm for the training of machine learning models on data silos in a privacy-preserving manner. In this survey, we explore the domain of personalized FL (PFL) to address the fundamental challenges of FL on heterogeneous data, a universal characteristic inherent in all real-world datasets. We analyze the key motivations for PFL and present a unique taxonomy of PFL techniques categorized according to the key challenges and personalization strategies in PFL. We highlight their key ideas, challenges, opportunities, and envision promising future trajectories of research toward a new PFL architectural design, realistic PFL benchmarking, and trustworthy PFL approaches.","{'model': 'tldr@v2.0.0', 'text': 'This survey explores the domain of personalized FL (PFL) to address the fundamental challenges of FL on heterogeneous data, a universal characteristic inherent in all real-world datasets.'}",https://arxiv.org/pdf/2103.00710
-sketch-guided text-to-image diffusion models,A. Voynov,"Text-to-Image models have introduced a remarkable leap in the evolution of machine learning, demonstrating high-quality synthesis of images from a given text-prompt. However, these powerful pretrained models still lack control handles that can guide spatial properties of the synthesized images. In this work, we introduce a universal approach to guide a pretrained text-to-image diffusion model, with a spatial map from another domain (e.g., sketch) during inference time. Unlike previous works, our method does not require to train a dedicated model or a specialized encoder for the task. Our key idea is to train a Latent Guidance Predictor (LGP) - a small, per-pixel, Multi-Layer Perceptron (MLP) that maps latent features of noisy images to spatial maps, where the deep features are extracted from the core Denoising Diffusion Probabilistic Model (DDPM) network. The LGP is trained only on a few thousand images and constitutes a differential guiding map predictor, over which the loss is computed and propagated back to push the intermediate images to agree with the spatial map. The per-pixel training offers flexibility and locality which allows the technique to perform well on out-of-domain sketches, including free-hand style drawings. We take a particular focus on the sketch-to-image translation task, revealing a robust and expressive way to generate images that follow the guidance of a sketch of arbitrary style or domain.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a universal approach to guide a pretrained text-to-image diffusion model, with a spatial map from another domain (e.g., sketch) during inference time, and trains a Latent Guidance Predictor, a small, per-pixel, Multi-Layer Perceptron that maps latent features of noisy images to spatial maps.'}",https://arxiv.org/pdf/2211.13752
-factual probing is [mask]: learning vs. learning to recall,Zexuan Zhong,"Petroni et al. (2019) demonstrated that it is possible to retrieve world facts from a pre-trained language model by expressing them as cloze-style prompts and interpret the model’s prediction accuracy as a lower bound on the amount of factual information it encodes. Subsequent work has attempted to tighten the estimate by searching for better prompts, using a disjoint set of facts as training data. In this work, we make two complementary contributions to better understand these factual probing techniques. First, we propose OptiPrompt, a novel and efficient method which directly optimizes in continuous embedding space. We find this simple method is able to predict an additional 6.4% of facts in the LAMA benchmark. Second, we raise a more important question: Can we really interpret these probing results as a lower bound? Is it possible that these prompt-search methods learn from the training data too? We find, somewhat surprisingly, that the training data used by these methods contains certain regularities of the underlying fact distribution, and all the existing prompt methods, including ours, are able to exploit them for better fact prediction. We conduct a set of control experiments to disentangle “learning” from “learning to recall”, providing a more detailed picture of what different prompts can reveal about pre-trained language models.","{'model': 'tldr@v2.0.0', 'text': 'OptiPrompt is proposed, a novel and efficient method which directly optimizes in continuous embedding space and is able to predict an additional 6.4% of facts in the LAMA benchmark.'}",https://aclanthology.org/2021.naacl-main.398.pdf
-"using chatgpt standard prompt engineering techniques in lesson preparation: role, instructions and seed-word prompts",A. Spasic,The application of available natural language processing systems can have a significant impact on the education process. The primary aim of this research was to test the impact of three standard prompting techniques on the results obtained from ChatGPT. Generation of a lesson plan for programming for preschoolers was chosen as the task set for AI. The obtained results show that use of a standard prompting with additional defined roles and seed words can be useful in preparation of teaching units and lessons and it can be considered as a technique of teachers' choice.,,
-generative ai tools in art education: exploring prompt engineering and iterative processes for enhanced creativity,Peter Cotroneo,"The rapid development and adoption of generative artificial intelligence (AI) tools in the art and design education landscape have introduced both opportunities and challenges. This timely study addresses the need to effectively integrate these tools into the classroom while considering ethical implications and the importance of prompt engineering. By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context. Students in the digital at case study were taught prompt engineering techniques and were tasked with crafting multiple prompts, focusing on refining their ideas over time. Participants demonstrated an increased understanding of the potential and limitations of generative AI tools and how to manipulate subject matter for more effective results. The iterative process encouraged students to explore and experiment with their creative ideas, leading to a deeper understanding of the possibilities offered by AI tools. Despite acknowledging the ethical concerns regarding copyright and the potential replacement of artists, students appreciated the value of generative AI tools for enhancing their sketchbooks and ideation process. Through prompt engineering and iterative processes, students developed a more detail-oriented approach to their work. The challenge of using AI-generated images as final products was conceptually intriguing, requiring further investigation and consideration of the prompts. This study highlights the potential benefits and challenges of integrating generative AI tools into art and design classrooms, emphasizing the importance of prompt engineering, iterative processes, and ethical considerations as these technologies continue to evolve.",,
-how does prompt engineering affect chatgpt performance on unsupervised entity resolution?,Khanin Sisaengsuwanchai,"Entity Resolution (ER) is the problem of semi-automatically determining when two entities refer to the same underlying entity, with applications ranging from healthcare to e-commerce. Traditional ER solutions required considerable manual expertise, including feature engineering, as well as identification and curation of training data. In many instances, such techniques are highly dependent on the domain. With recent advent in large language models (LLMs), there is an opportunity to make ER much more seamless and domain-independent. However, it is also well known that LLMs can pose risks, and that the quality of their outputs can depend on so-called prompt engineering. Unfortunately, a systematic experimental study on the effects of different prompting methods for addressing ER, using LLMs like ChatGPT, has been lacking thus far. This paper aims to address this gap by conducting such a study. Although preliminary in nature, our results show that prompting can significantly affect the quality of ER, although it affects some metrics more than others, and can also be dataset dependent.","{'model': 'tldr@v2.0.0', 'text': 'The results show that prompting can significantly affect the quality of ER, although it affects some metrics more than others, and can also be dataset dependent.'}",https://arxiv.org/pdf/2310.06174
-spectral features based convolutional neural network for accurate and prompt identification of schizophrenic patients,Kuldeep Singh,"Schizophrenia is a fatal mental disorder, which affects millions of people globally by the disturbance in their thinking, feeling and behaviour. In the age of the internet of things assisted with cloud computing and machine learning techniques, the computer-aided diagnosis of schizophrenia is essentially required to provide its patients with an opportunity to own a better quality of life. In this context, the present paper proposes a spectral features based convolutional neural network (CNN) model for accurate identification of schizophrenic patients using spectral analysis of multichannel EEG signals in real-time. This model processes acquired EEG signals with filtering, segmentation and conversion into frequency domain. Then, given frequency domain segments are divided into six distinct spectral bands like delta, theta-1, theta-2, alpha, beta and gamma. The spectral features including mean spectral amplitude, spectral power and Hjorth descriptors (Activity, Mobility and Complexity) are extracted from each band. These features are independently fed to the proposed spectral features-based CNN and long short-term memory network (LSTM) models for classification. This work also makes use of raw time-domain and frequency-domain EEG segments for classification using temporal CNN and spectral CNN models of same architectures respectively. The overall analysis of simulation results of all models exhibits that the proposed spectral features based CNN model is an efficient technique for accurate and prompt identification of schizophrenic patients among healthy individuals with average classification accuracies of 94.08% and 98.56% for two different datasets with optimally small classification time.","{'model': 'tldr@v2.0.0', 'text': 'The overall analysis of simulation results of all models exhibits that the proposed spectral features based CNN model is an efficient technique for accurate and prompt identification of schizophrenic patients among healthy individuals with average classification accuracies of 94.08% and 98.56% for two different datasets with optimally small classification time.'}",
-multimodal propaganda detection via anti-persuasion prompt enhanced contrastive learning,Jian Cui,"Propaganda, commonly used in memes disinformation, can influence the thinking of the audience and increase the reach of communication. Usually logical fallacy, as a kind of popular expression of memes, aims to create a logical reasonable illusion where the conclusion cannot be drawn with the use of correct logic rules. However, this characteristic inherent in memes leads to difficulties for classic multi-label classifiers to understand propagation techniques. To this end, we propose a novel propaganda detection model called Antipersuasion Prompt Enhanced Contrastive Learning (abbreviated as APCL). First, our APCL reformulates the multi-label classification task by leveraging the category words of propaganda technique based prompt engineering, which is converted into an image-text matching (ITM). Second, prompt engineering is designed with a persuasion prompt template and an anti-persuasion prompt template. The former is to build matched text-image pairs, and the latter is to form mismatched text-image pairs which fit the logical fallacy style of memes. Finally, the propagation technique is predicted based on the distances between the above two prompt templates enhanced texts and an input image. Experimental results on the memes dataset of SemEval2021 task 6 show that our APCL outperforms the state-of-the-art multimodal classification models in terms of F1 measures.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on the memes dataset of SemEval2021 task 6 show that the proposed Antipersuasion Prompt Enhanced Contrastive Learning outperforms the state-of-the-art multimodal classification models in terms of F1 measures.'}",
-user-friendly image editing with minimal text input: leveraging captioning and injection techniques,Sunwoo Kim,"Recent text-driven image editing in diffusion models has shown remarkable success. However, the existing methods assume that the user's description sufficiently grounds the contexts in the source image, such as objects, background, style, and their relations. This assumption is unsuitable for real-world applications because users have to manually engineer text prompts to find optimal descriptions for different images. From the users' standpoint, prompt engineering is a labor-intensive process, and users prefer to provide a target word for editing instead of a full sentence. To address this problem, we first demonstrate the importance of a detailed text description of the source image, by dividing prompts into three categories based on the level of semantic details. Then, we propose simple yet effective methods by combining prompt generation frameworks, thereby making the prompt engineering process more user-friendly. Extensive qualitative and quantitative experiments demonstrate the importance of prompts in text-driven image editing and our method is comparable to ground-truth prompts.","{'model': 'tldr@v2.0.0', 'text': 'This work demonstrates the importance of a detailed text description of the source image, by dividing prompts into three categories based on the level of semantic details, and proposes simple yet effective methods by combining prompt generation frameworks, thereby making the prompt engineering process more user-friendly.'}",http://arxiv.org/pdf/2306.02717
-ascm: an answer space clustered prompting method without answer engineering,Zhen Wang,"Prompt-based learning, which exploits knowledge from pre-trained language models by providing textual prompts and designing appropriate answer-category mapping methods, has achieved impressive successes on few-shot text classification and natural language inference (NLI). Because of the diverse linguistic expression, there exist many answer tokens for the same category. However, both manual answer design and automatic answer search constrain answer space and therefore hardly achieve ideal performance. To address this issue, we propose an answer space clustered prompting model (ASCM) together with a synonym initialization method (SI) which automatically categorizes all answer tokens in a semantic-clustered embedding space. We also propose a stable semi-supervised method named stair learning (SL) that orderly distills knowledge from better models to weaker models. Extensive experiments demonstrate that our ASCM+SL significantly outperforms existing state-of-the-art techniques in few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an answer space clustered prompting model (ASCM) together with a synonym initialization method (SI) which automatically categorizes all answer tokens in a semantic-clustered embedding space and proposes a stable semi-supervised method named stair learning (SL) that orderly distills knowledge from better models to weaker models.'}",https://aclanthology.org/2022.findings-acl.193.pdf
-troubleshooting techniques for cerebral aneurysm surgery based on “the study of failure”,S. Takeuchi,"In this study, we analyzed the causes of problems encountered during aneurysm surgery based on “the study of failure,” which was originally devised for system engineering. We describe four problematic cases, which were all successfully managed by troubleshooting techniques. The majority of the problems (failures) were caused by the surgeon’s “carelessness and/or decision error”. Large vessel injury during aneurysm dissection is formidable but can be managed by troubleshooting techniques such as micro-suturing or a bypass procedure in the deep operative field. Prompt and secure micro-anastomotic suturing is one of the vital troubleshooting techniques during aneurysm surgery. Personal preparation of micro-suturing instruments and daily off-the-job training are essential to master such troubleshooting procedures.","{'model': 'tldr@v2.0.0', 'text': 'The causes of problems encountered during aneurysm surgery are analyzed based on “the study of failure,” which was originally devised for system engineering and described four problematic cases, which were successfully managed by troubleshooting techniques.'}",https://www.jstage.jst.go.jp/article/scs/48/2/48_96/_pdf
-automatic code summarization via chatgpt: how far are we?,Weisong Sun,"To support software developers in understanding and maintaining programs, various automatic code summarization techniques have been proposed to generate a concise natural language comment for a given code snippet. Recently, the emergence of large language models (LLMs) has led to a great boost in the performance of natural language processing tasks. Among them, ChatGPT is the most popular one which has attracted wide attention from the software engineering community. However, it still remains unclear how ChatGPT performs in (automatic) code summarization. Therefore, in this paper, we focus on evaluating ChatGPT on a widely-used Python dataset called CSN-Python and comparing it with several state-of-the-art (SOTA) code summarization models. Specifically, we first explore an appropriate prompt to guide ChatGPT to generate in-distribution comments. Then, we use such a prompt to ask ChatGPT to generate comments for all code snippets in the CSN-Python test set. We adopt three widely-used metrics (including BLEU, METEOR, and ROUGE-L) to measure the quality of the comments generated by ChatGPT and SOTA models (including NCS, CodeBERT, and CodeT5). The experimental results show that in terms of BLEU and ROUGE-L, ChatGPT's code summarization performance is significantly worse than all three SOTA models. We also present some cases and discuss the advantages and disadvantages of ChatGPT in code summarization. Based on the findings, we outline several open challenges and opportunities in ChatGPT-based code summarization.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating ChatGPT on a widely-used Python dataset called CSN-Python and comparing it with several state-of-the-art (SOTA) code summarization models shows that in terms of BLEU and ROUGE-L,ChatGPT's code summarizing performance is significantly worse than all three SOTA models.""}",http://arxiv.org/pdf/2305.12865
-game of tones: faculty detection of gpt-4 generated content in university assessments,Mike Perkins,,,https://arxiv.org/pdf/2305.18081
-do llms possess a personality? making the mbti test an amazing evaluation for large language models,Keyu Pan,"The field of large language models (LLMs) has made significant progress, and their knowledge storage capacity is approaching that of human beings. Furthermore, advanced techniques, such as prompt learning and reinforcement learning, are being employed to address ethical concerns and hallucination problems associated with LLMs, bringing them closer to aligning with human values. This situation naturally raises the question of whether LLMs with human-like abilities possess a human-like personality? In this paper, we aim to investigate the feasibility of using the Myers-Briggs Type Indicator (MBTI), a widespread human personality assessment tool, as an evaluation metric for LLMs. Specifically, extensive experiments will be conducted to explore: 1) the personality types of different LLMs, 2) the possibility of changing the personality types by prompt engineering, and 3) How does the training dataset affect the model's personality. Although the MBTI is not a rigorous assessment, it can still reflect the similarity between LLMs and human personality. In practice, the MBTI has the potential to serve as a rough indicator. Our codes are available at https://github.com/HarderThenHarder/transformers_tasks/tree/main/LLM/llms_mbti.",,https://arxiv.org/pdf/2307.16180
-a comparative study of damage-sensitive features for rapid data-driven seismic structural health monitoring,Y. Reuland,"Rapid post-earthquake damage assessment forms a critical element of resilience, ensuring a prompt and functional recovery of the built environment. Monitoring-based approaches have the potential to significantly improve upon current visual inspection-based condition assessment that is slow and potentially subjective. The large variety of sensing solutions that has become available at affordable cost in recent years allows the engineering community to envision permanent-monitoring applications even in conventional low-to-mid-rise buildings. When combined with adequate structural health monitoring (SHM) techniques, sensor data recorded during earthquakes have the potential to provide automated near-real-time identification of earthquake damage. Near-real time building assessment relies on the tracking of damage-sensitive features (DSFs) that can be directly and rapidly derived from dynamic monitoring data and scaled with damage. We here offer a comprehensive review of such damage-sensitive features in an effort to formally assess the capacity of such data-driven indicators to detect, localize and quantify the presence of nonlinearity in seismic-induced structural response. We employ both a parametric analysis on a simulated model and real data from shake-table tests to investigate the strengths and limitations of purely data-driven approaches, which typically involve a comparison against a healthy reference state. We present an array of damage-sensitive features which are found to be robust with respect to noise, to reliably detect and scale with nonlinearity, and to carry potential to localize the occurrence of nonlinear behavior in conventional structures undergoing earthquakes.",,https://www.mdpi.com/2076-3417/13/4/2708/pdf?version=1676887837
-the application of chatgpt in healthcare progress notes: a commentary from a clinical and research perspective,Josh Nguyen,"ChatGPT, powered by one of the most advanced language processing systems, gained over 100 million users in just 2 months following its release in November 2022.1 This unprecedented popularity is likely due to its wide range of potential applications in fields, such as engineering, education and healthcare.2–4 The integration of artificial intelligence (AI)—driven language models like ChatGPT has the potential to revolutionize documentation practices, streamline workflows, and ultimately lead to more efficient and patient-centred care,2 though the use of such tools is not without its challenges. Here, we outline the potential benefits and pitfalls of implementing AI-driven language models, such as ChatGPT, in the creation and management of healthcare progress notes using prompt engineering techniques. We provide recommendations for responsible and effective integration into clinical practice and priorities for future research. Healthcare clinicians spend 35% of their time documenting patient data, and evidence suggests the length of healthcare case notes has been increasing over time.5 Existing innovations, such as speech recognition technology, yield no clear benefit in timesaving or documentation quality.6 With the ability to coherently write logical and accurate text within a few seconds, ChatGPT has the potential to reduce time spent on tasks such as preparing healthcare progress notes, and might also enhance","{'model': 'tldr@v2.0.0', 'text': 'The potential benefits and pitfalls of implementing AI-driven language models, such as ChatGPT, in the creation and management of healthcare progress notes using prompt engineering techniques are outlined and recommendations for responsible and effective integration into clinical practice are provided.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ctm2.1324
-"augmented behavioral annotation tools, with application to multimodal datasets and models: a systematic review",Eleanor Watson,"Annotation tools are an essential component in the creation of datasets for machine learning purposes. Annotation tools have evolved greatly since the turn of the century, and now commonly include collaborative features to divide labor efficiently, as well as automation employed to amplify human efforts. Recent developments in machine learning models, such as Transformers, allow for training upon very large and sophisticated multimodal datasets and enable generalization across domains of knowledge. These models also herald an increasing emphasis on prompt engineering to provide qualitative fine-tuning upon the model itself, adding a novel emerging layer of direct machine learning annotation. These capabilities enable machine intelligence to recognize, predict, and emulate human behavior with much greater accuracy and nuance, a noted shortfall of which have contributed to algorithmic injustice in previous techniques. However, the scale and complexity of training data required for multimodal models presents engineering challenges. Best practices for conducting annotation for large multimodal models in the most safe and ethical, yet efficient, manner have not been established. This paper presents a systematic literature review of crowd and machine learning augmented behavioral annotation methods to distill practices that may have value in multimodal implementations, cross-correlated across disciplines. Research questions were defined to provide an overview of the evolution of augmented behavioral annotation tools in the past, in relation to the present state of the art. (Contains five figures and four tables).","{'model': 'tldr@v2.0.0', 'text': 'A systematic literature review of crowd and machine learning augmented behavioral annotation methods to distill practices that may have value in multimodal implementations, cross-correlated across disciplines is presented.'}",https://www.mdpi.com/2673-2688/4/1/7/pdf?version=1674957180
-copilot for xcode: exploring ai-assisted programming by prompting cloud-based large language models,C. Tan,"This paper presents an AI-assisted programming tool called Copilot for Xcode for program composition and design to support human software developers. By seamlessly integrating cloud-based Large Language Models (LLM) with Apple's local development environment, Xcode, this tool enhances productivity and unleashes creativity for software development in Apple software ecosystem (e.g., iOS apps, macOS). Leveraging advanced natural language processing (NLP) techniques, Copilot for Xcode effectively processes source code tokens and patterns within code repositories, enabling features such as code generation, autocompletion, documentation, and error detection. Software developers can also query and make""small""decisions for program composition, some of which can be made simultaneously, and this is facilitated through prompt engineering in a chat interface of Copilot for Xcode. Finally, we present simple case studies as evidence of the effectiveness of utilizing NLP in Xcode to prompt popular LLM services like OpenAI ChatGPT for program composition and design.","{'model': 'tldr@v2.0.0', 'text': ""An AI-assisted programming tool called Copilot for Xcode for program composition and design to support human software developers, seamlessly integrating cloud-based Large Language Models with Apple's local development environment, Xcode, enhances productivity and unleashes creativity for software development in Apple software ecosystem.""}",https://arxiv.org/pdf/2307.14349
-towards equitable representation in text-to-image synthesis models with the cross-cultural understanding benchmark (ccub) dataset,Zhixuan Liu,"It has been shown that accurate representation in media improves the well-being of the people who consume it. By contrast, inaccurate representations can negatively affect viewers and lead to harmful perceptions of other cultures. To achieve inclusive representation in generated images, we propose a culturally-aware priming approach for text-to-image synthesis using a small but culturally curated dataset that we collected, known here as Cross-Cultural Understanding Benchmark (CCUB) Dataset, to fight the bias prevalent in giant datasets. Our proposed approach is comprised of two fine-tuning techniques: (1) Adding visual context via fine-tuning a pre-trained text-to-image synthesis model, Stable Diffusion, on the CCUB text-image pairs, and (2) Adding semantic context via automated prompt engineering using the fine-tuned large language model, GPT-3, trained on our CCUB culturally-aware text data. CCUB dataset is curated and our approach is evaluated by people who have a personal relationship with that particular culture. Our experiments indicate that priming using both text and image is effective in improving the cultural relevance and decreasing the offensiveness of generated images while maintaining quality.","{'model': 'tldr@v2.0.0', 'text': 'A culturally-aware priming approach for text-to-image synthesis using a small but culturally curated dataset, known here as Cross-Cultural Understanding Benchmark (CCUB) Dataset, to fight the bias prevalent in giant datasets.'}",http://arxiv.org/pdf/2301.12073
-omniscientdb: a large language model-augmented dbms that knows what other dbmss do not know,Matthias Urban,"In this paper, we present our vision of OmniscientDB, a novel database that leverages the implicitly-stored knowledge in large language models to augment datasets for analytical queries or even machine learning tasks. OmiscientDB empowers its users to augment their datasets by means of simple SQL queries and thus has the potential to dramatically reduce the manual overhead associated with data integration. It uses automatic prompt engineering to construct appropriate prompts for given SQL queries and passes them to a large language model like GPT-3 to contribute additional data (i.e., new rows, columns, or entire tables), augmenting the explicitly stored data. Our initial evaluation demonstrates the general feasibility of our vision, explores different prompting techniques in greater detail, and points towards several directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'OmiscientDB empowers its users to augment their datasets by means of simple SQL queries and thus has the potential to dramatically reduce the manual overhead associated with data integration.'}",http://publikationen.ub.uni-frankfurt.de/files/74426/06_08.pdf
-visualizing catalytic dynamics processes via synchrotron radiation multitechniques,Dengfeng Cao,"The importance of catalysts today as workhorses in most modern industrial fields cannot be downplayed. As a result, rational design and engineering of targeted catalysts have emerged as key objectives and are dependent on in‐depth understanding of complex catalytic dynamics. Synchrotron radiation (SR) light sources with rich advanced experimental methods are being recognized as a comprehensive characterization platform, which can draw a full picture on such multiparameter‐involved catalysis under actual working conditions. Herein, the recent progress of catalytic dynamics process studied by the means of various SR techniques is summarized. In particular, SR‐based spectroscopic, scattering, and imaging investigations on true catalysts are first introduced with the potential of in situ and operando characterizations. Apparently, the limitations from single SR technique naturally prompt a simple combination of SR techniques to better understand the whole catalysis process. Moreover, the discrepancies among various online testing facilities and batches of samples, along with random/systematic errors introduced by traditional intermittent/asynchronous measurement make it imperative to develop more prolific systems, complementary of multiple SR techniques for deep probing of dynamic catalytic processes. It is believed that the booming new light sources can further enrich the current multiple SR techniques, and thus may realize the true visualization on future catalytic dynamic processes.",,
-automatic essay scoring method based on multi-scale features,Feng Li,"Essays are a pivotal component of conventional exams; accurately, efficiently, and effectively grading them is a significant challenge for educators. Automated essay scoring (AES) is a complex task that utilizes computer technology to assist teachers in scoring. Traditional AES techniques only focus on shallow linguistic features based on the grading criteria, ignoring the influence of deep semantic features. The AES model based on deep neural networks (DNN) can eliminate the need for feature engineering and achieve better accuracy. In addition, the DNN-AES model combining different scales of essays has recently achieved excellent results. However, it has the following problems: (1) It mainly extracts sentence-scale features manually and cannot be fine-tuned for specific tasks. (2) It does not consider the shallow linguistic features that the DNN-AES cannot extract. (3) It does not contain the relevance between the essay and the corresponding prompt. To solve these problems, we propose an AES method based on multi-scale features. Specifically, we utilize Sentence-BERT (SBERT) to vectorize sentences and connect them to the DNN-AES model. Furthermore, the typical shallow linguistic features and prompt-related features are integrated into the distributed features of the essay. The experimental results show that the Quadratic Weighted Kappa of our proposed method on the Kaggle ASAP competition dataset reaches 79.3%, verifying the efficacy of the extended method in the AES task.","{'model': 'tldr@v2.0.0', 'text': 'Sentence-BERT (SBERT) is utilized to vectorize sentences and connect them to the DNN-AES model, and the typical shallow linguistic features and prompt-related features are integrated into the distributed features of the essay.'}",https://www.mdpi.com/2076-3417/13/11/6775/pdf?version=1685706520
-"a systematic review of disaster management systems: approaches, challenges, and future directions",Saad Mazhar Khan,"Disaster management is a critical area that requires efficient methods and techniques to address various challenges. This comprehensive assessment offers an in-depth overview of disaster management systems, methods, obstacles, and potential future paths. Specifically, it focuses on flood control, a significant and recurrent category of natural disasters. The analysis begins by exploring various types of natural catastrophes, including earthquakes, wildfires, and floods. It then delves into the different domains that collectively contribute to effective flood management. These domains encompass cutting-edge technologies such as big data analysis and cloud computing, providing scalable and reliable infrastructure for data storage, processing, and analysis. The study investigates the potential of the Internet of Things and sensor networks to gather real-time data from flood-prone areas, enhancing situational awareness and enabling prompt actions. Model-driven engineering is examined for its utility in developing and modeling flood scenarios, aiding in preparation and response planning. This study includes the Google Earth engine (GEE) and examines previous studies involving GEE. Moreover, we discuss remote sensing; remote sensing is undoubtedly a valuable tool for disaster management, and offers geographical data in various situations. We explore the application of Geographical Information System (GIS) and Spatial Data Management for visualizing and analyzing spatial data and facilitating informed decision-making and resource allocation during floods. In the final section, the focus shifts to the utilization of machine learning and data analytics in flood management. These methodologies offer predictive models and data-driven insights, enhancing early warning systems, risk assessment, and mitigation strategies. Through this in-depth analysis, the significance of incorporating these spheres into flood control procedures is highlighted, with the aim of improving disaster management techniques and enhancing resilience in flood-prone regions. The paper addresses existing challenges and provides future research directions, ultimately striving for a clearer and more coherent representation of disaster management techniques.",,https://www.mdpi.com/2073-445X/12/8/1514/pdf?version=1690796243
-"guideline on application process of techniques developed for repair, replacement and mitigation activities",K. Koyama,"The process from the engineering development to actual application of the new techniques used for the maintenance activities for Japanese nuclear power plants’ (NPPs) component in operation, the JANTI Guideline, “Guideline on Application Process of Techniques Developed for Maintenance Activities” was developed to propose clarification activities to be checked or examined in each stage throughout process. The guideline focuses on techniques used for especially such remedial activities as repair, replacement and mitigation of ageing. In engineering development stage, all information on the new techniques are provided to be checked or examined at each stage of the process followed further, such as code/standard formulation stage. It is important to carry out share recognition for the activities, which should be performed in each process in the prompt application to the actual plant system efficiently, among the persons or organizations concerned.","{'model': 'tldr@v2.0.0', 'text': 'The JANTI Guideline, “Guideline on Application Process of Techniques Developed for Maintenance Activities” was developed to propose clarification activities to be checked or examined in each stage throughout process.'}",
-how understanding large language models can inform their use in physics education,Giulia Polverini,"The paper aims to fulfil three main functions: (1) to serve as an introduction for the physics education community to the functioning of Large Language Models (LLMs), (2) to present a series of illustrative examples demonstrating how prompt-engineering techniques can impact LLMs performance on conceptual physics tasks and (3) to discuss potential implications of the understanding of LLMs and prompt engineering for physics teaching and learning. We first summarise existing research on the performance of a popular LLM-based chatbot (ChatGPT) on physics tasks. We then give a basic account of how LLMs work, illustrate essential features of their functioning, and discuss their strengths and limitations. Equipped with this knowledge, we discuss some challenges with generating useful output with ChatGPT-4 in the context of introductory physics, paying special attention to conceptual questions and problems. We then provide a condensed overview of relevant literature on prompt engineering and demonstrate through illustrative examples how selected prompt-engineering techniques can be employed to improve ChatGPT-4's output on conceptual introductory physics problems. Qualitatively studying these examples provides additional insights into ChatGPT's functioning and its utility in physics problem solving. Finally, we consider how insights from the paper can inform the use of LMMs in the teaching and learning of physics.",,
-automatic bug fixing via deliberate problem solving with large language models,Guoyang Weng,"Developers dedicate a significant share of their activities to finding and fixing defects in their code. Automated program repair (APR) attempts to reduce this effort by a set of techniques for automatically fixing errors or vulnerabilities in software systems. Recent Large Language Models (LLMs) such as GPT-4 offer an effective alternative to existing APR methods, featuring out-of-the-box bug fixing performance comparable to even sophisticated deep learning approaches such as CoCoNut. In this work we propose a further extension to LLM-based program repair techniques by leveraging a recently introduced interactive prompting technique called Tree of Thoughts (ToT). Specifically, we ask a LLM to propose multiple hypotheses about the location of a bug, and based on the aggregated response we prompt for bug fixing suggestions. A preliminary evaluation shows that our approach is able to fix multiple complex bugs previously unsolved by GPT-4 even with prompt engineering. This result motivates further exploration of hybrid approaches which combine LLMs with suitable meta-strategies.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a further extension to LLM-based program repair techniques by leveraging a recently introduced interactive prompting technique called Tree of Thoughts (ToT), and shows that this approach is able to fix multiple complex bugs previously unsolved by GPT-4 even with prompt engineering.'}",
-data-driven approach for formality-sensitive machine translation: language-specific handling and synthetic data generation,Seugnjun Lee,"In this paper, we introduce a data-driven approach for Formality-Sensitive Machine Translation (FSMT) that caters to the unique linguistic properties of four target languages. Our methodology centers on two core strategies: 1) language-specific data handling, and 2) synthetic data generation using large-scale language models and empirical prompt engineering. This approach demonstrates a considerable improvement over the baseline, highlighting the effectiveness of data-centric techniques. Our prompt engineering strategy further improves performance by producing superior synthetic translation examples.","{'model': 'tldr@v2.0.0', 'text': 'A data-driven approach for Formality-Sensitive Machine Translation (FSMT) that caters to the unique linguistic properties of four target languages using large-scale language models and empirical prompt engineering.'}",http://arxiv.org/pdf/2306.14514
-noisy exemplars make large language models more robust: a domain-agnostic behavioral analysis,Hongyi Zheng,"Recent advances in prompt engineering enable large language models (LLMs) to solve multi-hop logical reasoning problems with impressive accuracy. However, there is little existing work investigating the robustness of LLMs with few-shot prompting techniques. Therefore, we introduce a systematic approach to test the robustness of LLMs in multi-hop reasoning tasks via domain-agnostic perturbations. We include perturbations at multiple levels of abstractions (e.g. lexical perturbations such as typos, and semantic perturbations such as the inclusion of intermediate reasoning steps in the questions) to conduct behavioral analysis on the LLMs. Throughout our experiments, we find that models are more sensitive to certain perturbations such as replacing words with their synonyms. We also demonstrate that increasing the proportion of perturbed exemplars in the prompts improves the robustness of few-shot prompting methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a systematic approach to test the robustness of LLMs in multi-hop reasoning tasks via domain-agnostic perturbations, and finds that models are more sensitive to certain perturbation such as replacing words with their synonyms.'}",
-automatic emotion recognition in healthcare data using supervised machine learning,Nazish Azam,"Human feelings are fundamental to perceive the conduct and state of mind of an individual. A healthy emotional state is one significant highlight to improve personal satisfaction. On the other hand, bad emotional health can prompt social or psychological well-being issues. Recognizing or detecting feelings in online health care data gives important and helpful information regarding the emotional state of patients. To recognize or detection of patient’s emotion against a specific disease using text from online sources is a challenging task. In this paper, we propose a method for the automatic detection of patient’s emotions in healthcare data using supervised machine learning approaches. For this purpose, we created a new dataset named EmoHD, comprising of 4,202 text samples against eight disease classes and six emotion classes, gathered from different online resources. We used six different supervised machine learning models based on different feature engineering techniques. We also performed a detailed comparison of the chosen six machine learning algorithms using different feature vectors on our dataset. We achieved the highest 87% accuracy using MultiLayer Perceptron as compared to other state of the art models. Moreover, we use the emotional guidance scale to show that there is a link between negative emotion and psychological health issues. Our proposed work will be helpful to automatically detect a patient’s emotion during disease and to avoid extreme acts like suicide, mental disorders, or psychological health issues. The implementation details are made publicly available at the given link: https://bit.ly/2NQeGET.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method for the automatic detection of patient’s emotions in healthcare data using supervised machine learning approaches and uses the emotional guidance scale to show that there is a link between negative emotion and psychological health issues.'}",
-techniques used to promote active learning in an undergraduate architectural acoustics course,Michelle C. Vigeant,"The five-year undergraduate architectural engineering (AE) program at Penn State requires all students to take an introductory course in architectural acoustics (AA) and for most students in the program, no further AA courses are required. As a result, it is very important to capture the students’ interest early on in the course and use strategies that will improve long-term retention of the material. A number of techniques were used to improve student engagement with the introductory AA course that had an enrollment of 94 students. The first homework assignment was to summarize a recent magazine article in any area of acoustics, which allowed the students to immediately see a real-world application of the subject. During lecture periods, students were encouraged to participate through the use of personal response devices, also known as “clickers.” Several different types of clicker questions were used, including review questions from the preceding lecture, questions to prompt participation during the pre...",,
-"vehicle fault diagnostics using text mining, vehicle engineering structure and machine learning",Y. Murphey,"This paper presents an intelligent vehicle fault diagnostics system, SeaProSel(Search-Prompt-Select). SeaProSel takes a casual description of vehicle problems as input and searches for a diagnostic code that accurately matches the problem description. SeaProSel was developed using automatic text classification and machine learning techniques combined with a prompt-and-select technique based on the vehicle diagnostic engineering structure to provide robust classification of the diagnostic code that accurately matches the problem description. Machine learning algorithms are developed to automatically learn words and terms, and their variations commonly used in verbal descriptions of vehicle problems, and to build a TCW(Term-Code-Weight) matrix that is used for measuring similarity between a document vector and a diagnostic code class vector. When no exactly matched diagnostic code is found based on the direct search using the TCW matrix, the SeaProSel system will search the vehicle fault diagnostic structure for the proper questions to pose to the user in order to obtain more details about the problem. A LSI (Latent Semantic Indexing) model is also presented and analyzed in the paper. The performances of the LSI model and TCW models are presented and discussed. An in-depth study of different term weight functions and their performances are presented. All experiments are conducted on real-world vehicle diagnostic data, and the results show that the proposed SeaProSel system generates accurate results efficiently for vehicle fault diagnostics.","{'model': 'tldr@v2.0.0', 'text': 'The results show that the proposed SeaProSel system generates accurate results efficiently for vehicle fault diagnostics.'}",https://article.sciencepublishinggroup.com/pdf/10.11648.j.ijiis.20150403.12.pdf
-a quantitative analysis approach for cardiac arrhythmia classification using higher order spectral techniques,L. Khadra,"Ventricular tachyarrhythmias, in particular ventricular fibrillation (VF), are the primary arrhythmic events in the majority of patients suffering from sudden cardiac death. Attention has focused upon these articular rhythms as it is recognized that prompt therapy can lead to a successful outcome. There has been considerable interest in analysis of the surface electrocardiogram (ECG) in VF centred on attempts to understand the pathophysiological processes occurring in sudden cardiac death, predicting the efficacy of therapy, and guiding the use of alternative or adjunct therapies to improve resuscitation success rates. Atrial fibrillation (AF) and ventricular tachycardia (VT) are other types of tachyarrhythmias that constitute a medical challenge. In this paper, a high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmias. The algorithm is based upon bispectral analysis techniques. The bispectrum is estimated using an autoregressive model, and the frequency support of the bispectrum is extracted as a quantitative measure to classify atrial and ventricular tachyarrhythmias. Results show a significant difference in the parameter values for different arrhythmias. Moreover, the bicoherency spectrum shows different bicoherency values for normal and tachycardia patients. In particular, the bicoherency indicates that phase coupling decreases as arrhythmia kicks in. The simplicity of the classification parameter and the obtained specificity and sensitivity of the classification scheme reveal the importance of higher order spectral analysis in the classification of life threatening arrhythmias. Further investigations and modification of the classification scheme could inherently improve the results of this technique and predict the instant of arrhythmia change.","{'model': 'tldr@v2.0.0', 'text': 'A high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmia, based upon bispectral analysis techniques and results show a significant difference in the parameter values for different arrhythmias.'}",
-enhancing performance prediction robustness by combining analytical modeling and machine learning,Diego Didona,"Classical approaches to performance prediction rely on two, typically antithetic, techniques: Machine Learning (ML) and Analytical Modeling (AM). ML takes a black box approach, whose accuracy strongly depends on the representativeness of the dataset used during the initial training phase. Specifically, it can achieve very good accuracy in areas of the features' space that have been sufficiently explored during the training process. Conversely, AM techniques require no or minimal training, hence exhibiting the potential for supporting prompt instantiation of the performance model of the target system. However, in order to ensure their tractability, they typically rely on a set of simplifying assumptions. Consequently, AM's accuracy can be seriously challenged in scenarios (e.g., workload conditions) in which such assumptions are not matched. In this paper we explore several hybrid/gray box techniques that exploit AM and ML in synergy in order to get the best of the two worlds. We evaluate the proposed techniques in case studies targeting two complex and widely adopted middleware systems: a NoSQL distributed key-value store and a Total Order Broadcast (TOB) service.","{'model': 'tldr@v2.0.0', 'text': 'Several hybrid/gray box techniques are explored that exploit AM and ML in synergy in synergy to get the best of the two worlds, targeting two complex and widely adopted middleware systems.'}",http://www.gsd.inesc-id.pt/%7Eromanop/files/papers/ICPE15.pdf
-optomechanical design of a prompt gamma reaction history diagnostic,M. Kaufman,"The National Ignition Facility and the Omega Laser Facility both have a need for measuring prompt gamma radiation as part of a nuclear diagnostic program. A new gamma-detection diagnostic using off-axis-parabolic mirrors has been built. Some new techniques were used in the design, construction, and tolerancing of this gamma ray diagnostic. Because of the wavelength requirement (250 to 700 nm), the optical element surface finishes were a key design consideration. The optical enclosure had to satisfy pressure safety concerns and shielding against electromagnetic interference induced by gammas and neutrons. Structural finite element analysis was needed to meet rigorous optical and safety requirements. The optomechanical design is presented. Alignment issues are also discussed.",,
-"automated library management system in the chirala engineering college, chirala-using ez school software.",B. Kumar,"Library is a fast growing organism. The Ancient methods of maintaining it are no longer dynamic and efficient. For expeditious retrieval and dissemination of information and better service for the clientele, application of modern techniques has become absolutely indispensable. A properly computerized library will help its user with quick and prompt services. Library automation refers to mechanization of library housekeeping operations predominantly by computerization. Objectives of this study. To develop and updated database of books and other resources of the Engineering College Library, Chirala Engineering college, Chirala, JNTU, Kakinada University. To implement automated system using EZ School integrated Open Source Software. To provide various search options to know the availability of books in the Library. To generated the list of books due by a particular member and also the overdue charges. EZ School is integrated software system with the required models for small to very large libraries. It is found that this automation project sill service as a model for any Library. Being an open source, any library wanted to go for automation for their library housekeeping operations can make use of this software.","{'model': 'tldr@v2.0.0', 'text': 'This automation project sill service as a model for any Library, being an open source, any library wanted to go for automation for their library housekeeping operations can make use of this software.'}",
-detecting copd exacerbations early using daily telemonitoring of symptoms and k-means clustering: a pilot study,D. S. Morillo,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed electronic questionnaire and the applied methodology could help to early detect COPD exacerbations on a day-to-day basis and therefore could provide support to patients and physicians.'}",
-embracing ai for better quality engineering,,"Infosys has already progressed towards this with an AI-first quality engineering framework leveraging traditional and generative AI techniques, large language models (LLMs) trained for specific QE tasks using Infosys test case repository, prompt engineering for testing use cases etc. With its unmatched ability to automate even unsolved testing tasks, analyse patterns and anomalies, detect performance issues early, and produce smart insights, AI will set the benchmarks in quality engineering.","{'model': 'tldr@v2.0.0', 'text': 'With its unmatched ability to automate even unsolved testing tasks, analyse patterns and anomalies, detect performance issues early, and produce smart insights, AI will set the benchmarks in quality engineering.'}",
-photons for prompt tumour detection,R. Alfano,"Aeroplanes in clouds, submarines in murky water and cancerous tumours in breast tissue all have one thing in common – they are objects hidden in highly scattering media. Imaging through such media remains one of the most challenging problems in science and engineering but the advent of ultrafast lasers and detectors, coupled with a range of time-resolved techniques, has led to breakthroughs. Interest is currently focused on developing a time-resolved imaging technique to detect breast cancer tumours as early as possible.",,
-"study of real options with exogenous competitive entry to analyze dispute resolution ladder investments in architecture, engineering, and construction projects",C. Menassa,"Architecture, engineering, and construction industry participants often find it pragmatic to implement a project-specific dispute resolution ladder (DRL) as a managerial tool to assist in the prompt resolution of claims and change orders (CCOs) that might arise during the project construction phase. This project-specific DRL consists of a single or multiple alternative dispute resolution (ADR) techniques that require capital expenditures to cover the expenses incurred by the owner’s/contractor’s employees and third-party neutrals. If a project-specific DRL is properly chosen, then the capital expenditures are outweighed by the expected benefits from the DRL implementation; namely, prompt resolution of the CCOs without incurring excessive cost overruns on an already financially stressed project budget, as well as avoiding the escalation of the claims to a dispute that requires long protracted litigation for final settlement. Typically, the decision as to which ADR techniques to include in the project-speci...",,
-prompt list for risk management in sri lankan software industry,Mangala R. Perera,"Sound risk management is of crucial importance in software projects. As an emerging industry in Sri Lanka, software companies should focus on the proper risk management techniques. Unfortunately a large number of companies fail to achieve their project objectives. To avoid such situations, it is important for organizations to identify the main risks and take appropriate risk response actions. In this context, it is important to identify the prominent risks prevalent in Sri Lankan software industry and devise possible risk management strategies to respond to them. A prompt list is very useful in this situation. In this paper we propose such a prompt list tailored for the emerging software industry in Sri Lanka. Prompt lists, is a popular tool used in engineering to stimulate specific risk identification. It will ensure that the team will not miss out those risks that were important in the past. Preparing a prompt list for risk identification in Sri Lankan software industry and prioritizing the identified risks are the main objectives of this paper.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a prompt list for risk identification in Sri Lankan software industry and prioritizing the identified risks, which are the main objectives of this paper.'}",
-facile synthesis of sulfobetaine-stabilized cu2o nanoparticles and their biomedical potential.,Marta J. Woźniak-Budych,"A novel approach using a zwitterionic sulfobetaine-based surfactant for the synthesis of spherical copper oxide nanoparticles (Cu2O NPs) has been applied. For the first time, N-hexadecyl-N,N-dimethyl-3-ammonio-1-propanesulfonate has been used as stabilizer to control the size and morphology of Cu2O NPs. Several techniques, such as transmission electron microscopy (TEM), X-ray diffraction (XRD), and fluorescence spectroscopy, are used to investigate the size, structure, and optical properties of synthesized Cu2O nanocrystals. The results indicate that copper(I) oxide nanoparticles with size in the range of 2 to 45 nm and crystalline structure, exhibit intense yellow fluorescence (λem = 575 nm). Furthermore, the cytotoxicity studies show that sulfobetaine-stabilized copper oxide nanoparticles prompt inhibition of cancer cell proliferation in a concentration-dependent manner, however, the adverse effect on the normal cells has also been observed. The results indicate that the sulfobetaine-stabilized Cu2O, because of their unique properties, have a potential to be applied in medical fields, such as cancer therapy and bioimaging.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the sulfobetaine-stabilized Cu2O, because of their unique properties, have a potential to be applied in medical fields, such as cancer therapy and bioimaging.'}",https://doi.org/10.1021/acsbiomaterials.7b00465.s001
-telehealth intensive care unit nurse surveillance of sepsis,Teresa A. Rincon,"The purpose of this article is to describe the usability and human factors engineering standards used in development of a sepsis alert known as the sepsis prompt. Sensory processing, cognitive processing, signal detection, criterion response, and user satisfaction were evaluated with controlled user testing and critical incident response techniques. Nurses reported that the sepsis prompt was visible and distinct, making it easily detectable. The prompt provided a clear response mechanism and adequately balanced the number of false alerts with the likelihood of misses. Designers were able to use a mental model approach as they designed the prompt because the nurses were already using a manual sepsis detection process. This may have predisposed the nurses to response bias, and as such, they were willing to accommodate more false alarms than nurses who are not familiar with sepsis screening (surveillance). Nurses not currently screening for sepsis may not place the same value on this alert and find it an annoyance. The sepsis prompt met usability standards, and the nurses reported that it improved efficiency over the manual screening method.","{'model': 'tldr@v2.0.0', 'text': 'The usability and human factors engineering standards used in development of a sepsis alert known as the sepsi prompt met usability standards, and the nurses reported that it improved efficiency over the manual screening method.'}",
-comparing the effects of design interventions on the quality of design concepts as a reflection of ideation flexibility,Danielle Henderson,"Many tools, techniques, and other interventions have been developed to support idea generation within the design process. In previous research, we explored the separate effects of three such design interventions: teaming, problem framing, and design heuristics. In the teaming intervention, participants discussed a design prompt together but recorded their own ideas separately. In problem framing, multiple versions (framings) of each design prompt were used to elicit different solutions. In design heuristics, participants used specially designed cards to prompt new ways of thinking about the given design problem. In the current work, we compared the effects of these three interventions on students' design ideas with respect to one idea attribute in particular—quality. In total, 1088 design concepts were collected from 171 undergraduate students in engineering and industrial design from two universities. Individual cognitive style was also assessed using Kirton's Adaption–Innovation inventory (KAI). Six metrics taken from the design literature were used to assess the quality of each concept, namely: acceptability, applicability, clarity, effectiveness, implementability, and implicational explicitness. Paired t-tests and Pearson correlations were used to assess differences in quality between concepts generated with and without the three interventions; in addition, secondary effects were sought based on the cognitive styles and academic standings of the participants. Statistically significant differences were observed in design concept quality for the teaming and design heuristics interventions over the full sample and for some subgroups separated by cognitive style and academic standing. These results have implications for how educators teach design interventions and how students choose and apply interventions to affect the quality of their own design solutions.",,
-highly sensitive and practical detection of plant viruses via electrical impedance of droplets on textured silicon-based devices,M. Ambrico,"Early diagnosis of plant virus infections before the disease symptoms appearance may represent a significant benefit in limiting disease spread by a prompt application of appropriate containment steps. We propose a label-free procedure applied on a device structure where the electrical signal transduction is evaluated via impedance spectroscopy techniques. The device consists of a droplet suspension embedding two representative purified plant viruses i.e., Tomato mosaic virus and Turnip yellow mosaic virus, put in contact with a highly hydrophobic plasma textured silicon surface. Results show a high sensitivity of the system towards the virus particles with an interestingly low detection limit, from tens to hundreds of attomolar corresponding to pg/mL of sap, which refers, in the infection time-scale, to a concentration of virus particles in still-symptomless plants. Such a threshold limit, together with an envisaged engineering of an easily manageable device, compared to more sophisticated apparatuses, may contribute in simplifying the in-field plant virus diagnostics.","{'model': 'tldr@v2.0.0', 'text': 'A label-free procedure applied on a device structure where the electrical signal transduction is evaluated via impedance spectroscopy techniques, which shows a high sensitivity of the system towards the virus particles with an interestingly low detection limit.'}",https://www.mdpi.com/1424-8220/16/11/1946/pdf?version=1479462590
-hybrid machine learning/analytical models for performance prediction: a tutorial,Diego Didona,"Classical approaches to performance prediction of computer systems rely on two, typically antithetic, techniques: Machine Learning (ML) and Analytical Modeling (AM). ML undertakes a black-box approach, which typically achieves very good accuracy in regions of the features' space that have been sufficiently explored during the training process, but that has very weak extrapolation power (i.e., poor accuracy in regions for which none, or too few samples are known). Conversely, AM relies on a white-box approach, whose key advantage is that it requires no or minimal training, hence supporting prompt instantiation of the target system's performance model. However, to ensure their tractability, AM-based performance predictors typically rely on simplifying assumptions. Consequently, AM's accuracy is challenged in scenarios not matching such assumptions. This tutorial describes techniques that exploit AM and ML in synergy in order to get the best of the two worlds. It surveys several such hybrid techniques and presents use cases spanning a wide range of application domains.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial describes techniques that exploit AM and ML in synergy in order to get the best of the two worlds and surveys several such hybrid techniques and presents use cases spanning a wide range of application domains.'}",
-ontology engineering and knowledge extraction for cross-lingual retrieval,J. Trapman,"In this paper, we show that by integrating existing NLP techniques and Semantic Web tools in a novel way, we can provide a valuable contribution to the solution of the knowledge acquisition bottleneck problem. NLP techniques to create a domain ontology on the basis of an open domain corpus have been combined with Semantic Web tools. More specifically, Watson and Prompt have been employed to enhance the kick-o ontology while Cornetto, a lexical database for Dutch, has been adopted to establish a link between the concepts and their Dutch lexicalization. The lexicalized ontology constitutes the basis for the cross-language retrieval of learning objects within the LT4eL eLearning project.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that by integrating existing NLP techniques and Semantic Web tools in a novel way, this paper can provide a valuable contribution to the solution of the knowledge acquisition bottleneck problem.'}",
-scientific literature text mining and the case for open access.,G. Sarma,"""Open access"" has become a central theme of journal reform in academic publishing. In this article, I examine the relationship between open access publishing and an important infrastructural element of a modern research enterprise, scientific literature text mining, or the use of data analytic techniques to conduct meta-analyses and investigations into the scientific corpus. I give a brief history of the open access movement, discuss novel journalistic practices, and an overview of data-driven investigation of the scientific corpus. I argue that particularly in an era where the veracity of many research studies has been called into question, scientific literature text mining should be one of the key motivations for open access publishing, not only in the basic sciences, but in the engineering and applied sciences as well. The enormous benefits of unrestricted access to the research literature should prompt scholars from all disciplines to lend their vocal support to enabling legal, wholesale access to the scientific literature as part of a data science pipeline.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that particularly in an era where the veracity of many research studies has been called into question, scientific literature text mining should be one of the key motivations for open access publishing, not only in the basic sciences, but in the engineering and applied sciences as well.'}",https://arxiv.org/pdf/1611.00097
-supporting users tasks with personal information management and web forms augmentation,Sergio Firmenich,,"{'model': 'tldr@v2.0.0', 'text': ""This work proposes an approach for enhancing Web forms using client-side adaptation techniques in order to assist users to fill in Web forms and includes the support for data exchange between user's personal information management systems (PIMs) and third-party Web forms.""}",https://link.springer.com/content/pdf/10.1007/978-3-642-31753-8_20.pdf
-deterioration of concrete in marine structure,P. KadamHarshada.,"Concrete is a hard composite material got by blending the fitting extents of bond and totals (fine and coarse) with water, which in mix with steel fortification structures an amazing Civil Engineering material. Concrete, being great in pressure and steel in strain, end up being extremely solid and strong, if effectively outlined and cast. There are many components which prompt the decay of cement. The most widely recognized components which prompt the decay are the natural factors the materials and the throwing methodology, (for example, shameful cementing techniques, utilization of permeable totals, utilization of totals from saline sources, utilization of saline water in solid, poor vibration, disgraceful cover to support, spilling formwork, utilization of permeable and delicate totals), basic plan surrenders temperature and uncalled for utilize. A solid structure experiencing the weakening gives the accompanying demonstrative suggestions: breaking, spalling, scraped spot, stains, disintegration, diversion, and erosion. The most widely recognized variables prompting crumbling are the natural components prompting consumption of support, dishonorable utilize whereby over-burdening happens on a structure and the development techniques prompting permeable cement. This report introduces the investigation of the different components and their instruments prompting disintegration, their outcomes and the restoration of such structures. This report additionally talks about the scientific device for measurement of decay of cement in seaward structure.",,
-"elastic rail clip design development, based on virtual prototyping",G. Todorov,"Virtual prototyping technology is a powerful tool when exact design solution is needed, corresponding to various and sometimes controversial requirements. It is used widely for design improvement/optimisation (topology, shape, parametric), based on engineering analysis through numerical techniques (as Finite Element Method) that enables prompt and accurate solutions. Design development approach based on this technology becomes standard for many industries, including railway transport. Presented study is based on an industrial project for design development of elastic rail clip. It is entirely based on virtual prototyping to obtain adequate design solution. Design is required to achieve certain clip rigidity, at certain level of safety (mechanical stress), for certain design space. This is a typical task for design optimisation techniques and performed simulations include 3 design concepts, 11 design variants and more than 2400 parametric configurations to obtain 4 allowable design solutions. The study is a good presentation of virtual prototyping application for industrial purposes.","{'model': 'tldr@v2.0.0', 'text': 'The study is a good presentation of virtual prototyping application for industrial purposes and performed simulations include 3 design concepts, 11 design variants and more than 2400 parametric configurations to obtain 4 allowable design solutions.'}",
-development of low-cost hybrid measurement system,A. Dindar,"Following the construction phase, tall buildings, bridges, dams are economically important engineering structures normally should not be in any critical condition during their service life unless a strong excitation (earthquake, blast, flood etc.) or boundary condition change (relative ground settlement) happens. For the property owner and for the community, it is essential to have a prompt condition assessment after the hazard in short period or to be aware whether a slow but progressing change in the building conditions for the long term. Structural Health Monitoring (SHM) systems aims to provide a basis for the requirement by installing the sensors on the system and processing data acquired from the sensors. With rapid development in the sensor technology, the monitoring methodologies have evolved from simple to more complex but complimentary measurement techniques. This study presents the measurement system of accelerometers, inclinometers and geodetic Global Navigation Satellite System (GNSS) receiver modules have been integrated in a single device for hybrid system. The data acquired from these different modules have been processed real time on site and the time series analysis for alerting mechanism has been set into the microchips for the sake of the prompt assessment of the critical condition change. The hybrid measurement device has been tested on a scaled building model excited by shake table for harmonic and random motions as well as on real building under ambient vibrations. The results are future promising in terms of the low-cost measurement and data process systems for the communities vulnerable to seismic hazard.",,
-recognition of ventricular fibrillation using neural networks,Richard H. Clayton,,"{'model': 'tldr@v2.0.0', 'text': 'Although many VF detection techniques have been developed and claim good performance, independent evaluation has shown that some techniques are not optimal and one approach with potential lies in the area of neural computing.'}",
-development of a collaboration system based on mobile framework in the field of ships and marine industry,Myeong-ki Han,"Engineering information in the shipbuilding and marine industry should be inevitably modified owing to frequent changes in, for example, customer's requirements and product specifications. Such a working environment caused by a discrepancy in information between engineering and production departments can have inefficient effects on production tasks. In order to perform efficient production tasks, staff in a production department should have easy access to the latest information in engineering. Furthermore, information should be easy and prompt in sharing and searching when needed. In this paper, we propose a development of collaborative system based on mobile framework. This proposed system facilitates its application to the work environment in the shipbuilding and marine industry. Moreover, it particularly saves resources in shipyards such as energy and costs of labor and materials. Mobile techniques, such as XML, SOAP and Android, which have been used in this paper are useful for implementing functions of sharing and searching for information in a distributed environment. The purpose of the system proposed in this paper is to provide system users with a collaborative environment which is based on an online and paperless mobile framework, when different kinds of tasks are required due to changes in engineering information. This proposed system not only can check information in production tasks in real time but also can implement functions based on mobile framework. By reducing the discrepancy in information between the engineering and production departments, this collaborative system can be implemented effectively.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of the system proposed in this paper is to provide system users with a collaborative environment which is based on an online and paperless mobile framework, when different kinds of tasks are required due to changes in engineering information.'}",
-"access, reward and penalty: the art of supervision at the (post)graduate level, a case study of an electrical engineering laboratory",Tanya M. Vernon,"Supervision of graduate students is becoming an increasingly important area of focus and scholarship. An example of the efflorescence of theory and pedagogy of supervision in Australia and other Commonwealth countries is the tertiary-level institutionalisation of programs which serve to accredit individuals in the complex art of supervision. But while scholarship of graduate-level supervision in the humanit ies and social science grows exponentially, basic discourse regarding method and practice of supervision of graduate level engineers is largely lacking. The author employs ethnomethodological techniques to explore and address this undertheorised area of engineering pedagogy from the point of view of a supervisor of number students in electrical and electronic engineers. The paper specifically explores a basic pedagogical structure in a laboratory of higher degree research (HDR) electrical engineering students, defines a complex time-based and proximity reward system the supervisor uses to prompt students to achieve excellence, and theorises an ethical way of dealing with poor student performance.",,
-rapid microbiology - raising awareness.,J. Bailie,"A 'high-level overview' of some of the emerging rapid microbiology technologies designed to help healthcare engineering and infection control teams working in hospitals and other healthcare facilities more rapidly identify potentially hazardous levels of waterborne microorganisms in their water systems, enabling them to take prompt remedial action, and a look at the some of the 'pros and cons' of such testing techniques, was given by Nalco technical director, Howard Barnes, the vice-chair of the Legionella Control Association (LCA), at a recent LCA open day. HEJ editor, Jonathan Bailie, reports.","{'model': 'tldr@v2.0.0', 'text': ""A 'high-level overview' of some of the emerging rapid microbiology technologies designed to help healthcare engineering and infection control teams working in hospitals and other healthcare facilities more rapidly identify potentially hazardous levels of waterborne microorganisms in their water systems, and a look at the 'pros and cons' of such testing techniques.""}",
-deterioration of concrete in marine structure,Ms. Kadam Harshada,"Concrete is a hard composite material got by blending the fitting extents of bond and totals (fine and coarse) with water, which in mix with steel fortification structures an amazing Civil Engineering material. Concrete, being great in pressure and steel in strain, end up being extremely solid and strong, if effectively outlined and cast. There are many components which prompt the decay of cement. The most widely recognized components which prompt the decay are the natural factors the materials and the throwing methodology, (for example, shameful cementing techniques, utilization of permeable totals, utilization of totals from saline sources, utilization of saline water in solid, poor vibration, disgraceful cover to support, spilling formwork, utilization of permeable and delicate totals), basic plan surrenders temperature and uncalled for utilize. A solid structure experiencing the weakening gives the accompanying demonstrative suggestions: breaking, spalling, scraped spot, stains, disintegration, diversion, and erosion. The most widely recognized variables prompting crumbling are the natural components prompting consumption of support, dishonorable utilize whereby over-burdening happens on a structure and the development techniques prompting permeable cement. This report introduces the investigation of the different components and their instruments prompting disintegration, their outcomes and the restoration of such structures. This report additionally talks about the scientific device for measurement of decay of cement in seaward structure.",,
-on the contribution of biomedical engineering and technology to the understanding and the management of arterial hypertension.,S. Ghione,"There are several reasons why arterial blood pressure, i.e. the pressure within the large arterial vessels, is out of the physical parameters of the human body, one of the most frequently measured. Firstly, arterial blood pressure is a physiologically meaningful parameter, since it represents the driving pressure generated by the heart which maintains blood perfusion in the periphery. Secondly, it is a clinically important parameter: a decline of arterial blood pressure (e.g. in shock) may represent a life-threatening emergency which requires prompt recognition and correction; elevated blood pressure (hypertension) on the other hand is a very common condition, which bears a high risk of cardiovascular mortality and morbidity and can be controlled with appropriate pharmacological means. Thirdly, but not lastly, arterial blood pressure is easily measurable with a fair degree of accuracy by the standard manual sphygmomanometric method and, more recently, by non-invasive automatic techniques. This paper discusses some of the aspects related to arterial blood pressure measurement, in which, in the author's opinion, medical engineering and technology are expected to provide useful advancements. Two major areas will be considered. The first regards the methodologies for arterial blood pressure assessment; the second the identification and acquisition of information additional to blood pressure which would be helpful for a better understanding of blood pressure measurements and/or of risk profiling. For the purpose of this brief paper, we shall mainly use examples and reasonings from our own experience.","{'model': 'tldr@v2.0.0', 'text': ""This paper discusses some of the aspects related to arterial blood pressure measurement, in which, in the author's opinion, medical engineering and technology are expected to provide useful advancements.""}",
-qualitative approach for estimating the influence of refactoring and scrum in software development,R. Hussain,"Software development is intellectually a complex chore. The swift progress of software currently requires the high rate software product release by development teams. Different software development techniques and quality assurance methods are used in order to achieve high worth, unfailing, and error free software. In order to deliver the product earlier, the development teams make an alteration to their conservative software development lifecycle to agile development method which can allow them towards prompt release of software management with the requirements-change experience. Refactoring has been rising in magnitude with modern software engineering advances, predominantly agile methodologies, which endorse uninterrupted progress of an application's code and blueprint. Refactoring is the practice of analyzing and facilitating the plan of offered code, without altering its performance. Another trendy techniques in Agile development is the Scrum methodology. It involves regular release and the client receives an absolutely prepared application that includes more and more features every time In this paper Qualitative Approach For Estimating the Influence Of Refactoring And Scrum In Software Development is utilized. In this model scrum methodology is utilized in enhanced form to overcome scrum issues along with refactoring project at both design and implementation level.","{'model': 'tldr@v2.0.0', 'text': 'A Qualitative Approach For Estimating the Influence of Refactoring And Scrum In Software Development is utilized and scrum methodology is utilized in enhanced form to overcome scrum issues along with refactoring project at both design and implementation level.'}",
-global project management – not business as usual,Damodara U. Kini,"Engineering and construction companies will have to make radical changes to their project management techniques if they are to be successful in the global marketplace in the 21st century. Customers will expect lower costs, shorter engineering and construction periods, low-cost, high-quality materials and equipment from manufacturers anywhere in the world, and prompt attention to their needs. These expectations mean that dramatically new approaches will be required not only in the organizational structure of such companies but also in their way of thinking and in such critical functions such as information systems, supplier management, and quality assurance. It will not be business as usual. There are six key areas in which engineering and construction companies will have to excel to be successful in satisfying the global customer. First, these companies will need an organization that can match their expertise in a cost-effective manner to the needs of their global customers. Second, global companies will ...",,
-original research,Shigang Gao,"The rapid development and adoption of generative artificial intelligence (AI) tools in the art and design education landscape have introduced both opportunities and challenges. This timely study addresses the need to effectively inte-grate these tools into the classroom while considering ethical implications and the importance of prompt engineering. By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context. Students in the digital at case study were taught prompt engineering techniques and were tasked with crafting multiple prompts, focusing on refining their ideas over time. Participants demonstrated an increased understanding of the potential and limitations of generative AI tools and how to manipulate subject matter for more effective results. The iterative process encouraged students to explore and experiment with their creative ideas, leading to a deeper understanding of the possibilities offered by AI tools. Despite acknowledging the ethical concerns regarding copyright and the potential replacement of artists, students appreciated the value of generative AI tools for enhancing their sketchbooks and ideation process. Through prompt engineering and iterative processes, students developed a more detail-oriented approach to their work. The challenge of using AI-generated images as final products was conceptually intriguing, requiring further investigation and consideration of the prompts. This study highlights the potential benefits and challenges of integrating generative AI tools into art and design classrooms, emphasizing the importance of prompt engineering, iterative processes, and ethical considerations as these technologies continue to evolve.","{'model': 'tldr@v2.0.0', 'text': 'By examining the iterative process of refining original ideas through multiple iterations, verbal expansion, and the use of OpenAI’s DALL-E2 for generating diverse visual outcomes, researchers gain insights into the potential benefits and pitfalls of these tools in an educational context.'}",
-"bioengineering for customized orthodontic applications- implant, bracket and dental vibrator",R. Patil,"To understand complex living system an effort has made by mechanical engineers and dentists to deliver prompt products and services to patients concerned about their aesthetic look. Since two decades various bracket systems have designed involving techniques like milling, injection molding which are technically not flexible for the customized dental product development. The aim of this paper to design, develop a customized system which is economical and mainly emphasizes the expertise design and integration of engineering and dental fields. A custom made selfadjustable lingual bracket and customized implants are designed and developed using computer aided design (CAD) and rapid prototyping technology (RPT) to improve the smiles and to overcome the difficulties associated with conventional ones. Lengthy orthodontic treatment usually not accepted by the patients because the patient compliance is lost. Patient’s compliance can be improved by facilitating faster tooth movements by designing a localized dental vibrator using advanced engineering principles. Keywords—Orthodontics, Prosthodontics, Lingual bracket, Implants, Dental vibrator, Computer aided design, Rapid prototyping technology.",,
-infrared thermography and ground penetrating radar for airport pavements assessment,A. Moropoulou,"Non-destructive techniques, such as infrared thermography and ground penetrating radar, have the potential to produce rapid and accurate assessment of airport pavements. In this work, an examination of asphalt pavements situated at the International Airport of Athens in Greece, is performed. Infrared thermography and ground penetrating radar are introduced with the purpose of providing prompt and accurate condition assessment of airport pavements. These techniques are used efficiently in the detection of cracks, voids and other imperfections appearing either from the ageing of the materials or due to poor workmanship. Finally, this paper describes the problem of deteriorating airport pavements, the procedure and the equipment used for the in situ tests, while the results obtained lead to the suggestion of a predictive monitoring non-destructive technique for the inspection and appropriateness of efficient engineering structures.",,
-special issue: knowledge-based intelligent systems and their applications,I. Lovrek,"Intelligent techniques derived from knowledge-based engineering and related computing paradigms have provided useful concepts and tools to undertake a variety of real-world problems. These systems mimic the analytical and learning capabilities of the human brain. They harness the benefits of knowledge and intelligence to form an integrated framework for problem solving. In this special issue, a total of thirteen articles comprising extended papers from the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES2008) as well as from other submissions that highlight a small number of innovative knowledge-based intelligent systems and their applications to solving problems in different domains are presented. A summary of each article is as follows. With the development of advanced travelers information systems, it is important to have a prompt and accurate travel time prediction system for road networks. In the first article, two travel time prediction algorithms using naive Bayesian classification and rulebased classification are proposed. Based on a historical traffic database, the algorithms are able to yield high accuracy in travel time prediction. The algorithms are also useful for road networks with arbitrary travel routes. The results also reveal that naive Bayesian classification produces better mean absolute relative error than that of rule-based classification. For large-scale complex process plants that involve safety critical systems, real-time diagnosis is an important aspect. In the second article, an ontology for","{'model': 'tldr@v2.0.0', 'text': 'In this special issue, a total of thirteen articles comprising extended papers from the 12th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES2008) as well as from other submissions that highlight a small number of innovative knowledge-based intelligent systems and their applications to solving problems in different domains are presented.'}",https://content.iospress.com:443/download/journal-of-intelligent-and-fuzzy-systems/ifs00430?id=journal-of-intelligent-and-fuzzy-systems%2Fifs00430
-a new quantitative analysis technique for cardiac arrhythmia using bispectrum and bicoherency,L. Khadra,"Ventricular tachyarrhythmias, in particular ventricular fibrillation (VF), are the primary arrhythmic events in the majority of patients suffering from sudden cardiac death. Attention has focused upon these articular rhythms as it is recognized that prompt therapy can lead to a successful outcome. There has been considerable interest in analysis of the surface electrocardiogram (ECG) in VF centred on attempts to understand the pathophysiological processes occurring in sudden cardiac death, predicting the efficacy of therapy, and guiding the use of alternative or adjunct therapies to improve resuscitation success rates. Atrial fibrillation (AF) and ventricular tachycardia (VT) are other types of tachyarrhythmias that constitute a medical challenge. In this paper, a high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmias. The algorithm is based upon bispectral analysis techniques. The bispectrum is estimated using an AR model, and the frequency support of the bispectrum is extracted as a quantitative measure to classify atrial and ventricular tachyarrhythmias. Results show a significant difference in the parameter values for different arrhythmias. Moreover, the bicoherency spectrum shows different bicoherency values for normal and tachycardia patients. In particular, the bicoherency indicates that phase coupling decreases as arrhythmia kicks in.","{'model': 'tldr@v2.0.0', 'text': 'A high order spectral analysis technique is suggested for quantitative analysis and classification of cardiac arrhythmia, based upon bispectral analysis techniques and results show a significant difference in the parameter values for different arrhythmias.'}",
-fast algorithm for x-ray cone-beam microtomography,Ang Shih,,"{'model': 'tldr@v2.0.0', 'text': 'A generalized Feldkamp cone-beam image reconstruction algorithm is reformulate, utilize curved voxels and mapping tables, improve the reconstruction efficiency by an order of magnitude relative to a direct implementation of the standard algorithm, and demonstrate the feasibility with numerical simulation and experiments using a prototype cone- beam X-ray microtomographic system.'}",
-fetal transfusion therapy.,Daniel W. Skupski,"Rapid advances are occurring in the diagnosis and treatment of the fetus with a red blood cell or platelet cytopenia. Noninvasive methods of monitoring the alloimmunized pregnancy, invasive methods such as amniocentesis and cordocentesis, and intrauterine transfusion therapy of both red cells and platelets, are being further refined to allow the prompt recognition and treatment of fetal cytopenias. Specialized centers have now accrued a large experience in the management of the fetus severely affected by alloimmunization. Advances in ultrasound, blood banking techniques, and genetic engineering technology have spurred the most recent advances. The indications for diagnosis, timing and frequency of invasive procedures for treatment, and technical considerations regarding preparation of blood products and volume of transfusion, are outlined in this review. Polymerase chain reaction (PCR) determination of fetal Rh(D) genotype by chorionic villus sampling or amniocentesis in the first or second trimesters is a recent clinically useful advance. The advent of hematopoietic stem cell transplantation and the potential for gene therapy are exciting advances in the treatment and prevention of hematopoietic diseases, including, but not limited, to the fetal cytopenias.","{'model': 'tldr@v2.0.0', 'text': 'The indications for diagnosis, timing and frequency of invasive procedures for treatment, and technical considerations regarding preparation of blood products and volume of transfusion, are outlined in this review.'}",
-a unified architecture for automatic software updates,Dominic White,"This paper attempts to address the issue of hardening the internal security of an organisation’s network by easing its patch management. Traditionally security has been modeled on a ""hard outer shell"" approach, with a firewall protecting the otherwise vulnerable internal network. With the advent of worms using such techniques as social engineering to bypass the organisational firewall and installing trojans, this approach is no longer sufficient. As a result of these new attacks, emphasis should be placed on improving the security of the internal network. Most research agrees that prompt patching of security vulnerabilities would significantly reduce the vulnerability of these machines. However, this requires system administrators not only to keep abreast of the flood of patches, but to ensure they are deployed to every machine, in what could be a very large network. These difficulties are worsened by problems the patches themselves often create. This is a difficult task and the failure of system administrators to perform it is echoed in the recent spate of worm attacks, with some taking advantage of vulnerabilities for which patches had been released up to six months earlier. To cope in this environment an organisation needs a comprehensive patch management policy. A technology agnostic view of this policy is described in order to better understand what is required of an automated solution. A few vendors have released tools to find vulnerabilities, find patches and distribute them, the best known example of which is Microsoft’s Software Update Service. However, these tools are generally inflexible, expensive and only deal with a limited part of the patch management process. This paper seeks to discuss the implementation of an open source, cross platform tool to solve this problem. This will involve discussions in four areas; the need for such a system, the patch management process, existing systems, and the proposed solution. The proposed system architecture is then broken down into four areas; sourcing patches, patch packaging, testing patches, patch distribution and the development of the system. The first area will discuss how to pull patches from multiple vendors such as Windows Update, Symantec AV, FreeBSD ports and ISC. The second will involve discussion on creating packages; whether to replace entire binaries, to employ binary patching techniques or to use source distribution. The third section will discuss how this tool can improve testing and provide stop-gap measures for protecting the organisation during the testing process through the use of IDS signatures. The fourth section will discuss methods of distributing these packages, such as direct download or peer-to-peer implementations like Bittorrent. The last section will discuss a modular, platform-independent implementation of the system.","{'model': 'tldr@v2.0.0', 'text': 'This paper attempts to address the issue of hardening the internal security of an organisation’s network by easing its patch management by discussing the implementation of an open source, cross platform tool to solve this problem.'}",
-a teaching approach for software testing,A. McAllister,"Teaching students how to test software is complicated by the absence of a simple, integrated approach for generating test plans. No single testing technique fulfils these needs, and teaching only a collection of disparate techniques makes it difficult to assign work for students. This talk presents an integrated approach for test plan generation that can be used by students in programming and software engineering courses. The approach provides simple guidelines that prompt discovery of sets of test cases that are typically more complete than students produce on an ad hoc basis. A technique is introduced that ensures all program statements are executed during testing and that loops are tested in a rigorous manner. Experience shows that this technique tends to be simpler to use than existing techniques that identify independent paths through programs. All of the guidelines presented can be applied without automated tools. The primary strength of the approach is in demonstrating to students how rigorous generation of test plans can identify test cases that might otherwise not occur to the tester, and how multiple techniques can be combined to complement one another.","{'model': 'tldr@v2.0.0', 'text': 'The primary strength of the approach is in demonstrating to students how rigorous generation of test plans can identify test cases that might otherwise not occur to the tester, and how multiple techniques can be combined to complement one another.'}",
-an expert advisor for the selection of thermodynamic property estimation methods,R. Butner,"The Prototype Process Simulation Consultant (PPSC) is a rule-oriented, knowledge-based software system designed to assist design engineers in the selection of physical property estimation methods for commercial chemical process simulators. The PPSC uses ''blackboarding'' techniques to share data with a number of external program subunits, that prompt the user for inputs, perform calculations and display results. The development of the PPSC provides a useful context for discussion of efforts to apply expert systems techniques to chemical engineering problems.","{'model': 'tldr@v2.0.0', 'text': 'The Prototype Process Simulation Consultant is a rule-oriented, knowledge-based software system designed to assist design engineers in the selection of physical property estimation methods for commercial chemical process simulators.'}",
-"bone tissue engineering techniques, advances and scaffolds for treatment of bone defects.",Matthew Alonzo,,"{'model': 'tldr@v2.0.0', 'text': 'Recent advances in bone tissue engineering are highlighted, with particular emphasis on the role of the biomaterials as scaffolding material to heal bone defects.'}",
-chatgpt-based debate game application utilizing prompt engineering,Eunyul Lee,"This paper1 focuses on the implementation of a debate game using ChatGPT, aiming to investigate the feasibility of incorporating large language models into the educational domain through prompt engineering. The study explores strategies to elicit desired outputs from the GPT model by employing the prompt engineering methodology, as provided by Microsoft. Specifically, the game implementation involves the customization of ChatGPT's responses to facilitate a natural progression of debates, varying levels of difficulty, and an evaluation system for assessing the quality of discourse. By leveraging the prompt engineering methodology, we demonstrate that providing specific instructions or case-based prompts improves the accuracy and relevance of ChatGPT's answers. The developed application targets teenagers, enabling them to engage in real-time debates with ChatGPT and enhance their literacy skills. Furthermore, the game fosters the development of logical reasoning, persuasive abilities, effective expression, active participation, and attentive listening while expressing personal opinions, ultimately fostering a sense of accomplishment. Moreover, through debate evaluation and personalized advice, ChatGPT is expected to recognize and address its shortcomings, thereby continuously improving its conversational capabilities. Overall, this research contributes to the understanding of how large language models can be harnessed in educational settings and underscores the potential benefits of prompt engineering techniques in optimizing the outputs of such models.",,
-exploring the impact of prompt engineering on chatgpt 3.5 text summarization: a bert score evaluation,Kartik Ashok Hawelikar,"In the domain of Natural Language Processing (NLP), the technique of prompt engineering is a strategic method utilized to guide the responses of models such as ChatGPT. This research explores the intricacies of prompt engineering, with a specific focus on its effects on the quality of summaries generated by ChatGPT 3.5, an openly accessible chatbot developed by OpenAI. The study encompasses a comprehensive examination of 110 summaries produced from ten diverse paragraphs, employing eleven distinct summarization prompts under zero-shot setting. Evaluation is conducted using the BERT Score, a metric that offers a more contextually relevant assessment of summary quality. This study introduces an innovative approach to appraising the quality of summaries, setting it apart from prior investigations and delivering valuable insights into the nuances of prompt engineering's role within the NLP landscape. Ultimately, this inquiry illuminates the strengths and weaknesses associated with various prompts and their influence on ChatGPT 3.5's summarization capabilities, thereby making a significant contribution to the constantly evolving field of NLP and automated text summarization.",,https://doi.org/10.56726/irjmets45268
-challenges of influencing cellular morphology by morphology engineering techniques and mechanical induced stress on filamentous pellet systems—a critical review,M. Böl,"Filamentous microorganisms are main producers of organic acids, enzymes, and pharmaceutical agents such as antibiotics and other active pharmaceutical ingredients. With their complex cell morphology, ranging from dispersed mycelia to dense pellets, the cultivation is challenging. In recent years, various techniques for tailor‐made cell morphologies of filamentous microorganisms have been developed to increase product formation and have been summarised under the term morphology engineering. These techniques, namely microparticle‐enhanced cultivation, macroparticle‐enhanced cultivation, and alteration of the osmolality of the culture medium by addition of inorganic salts, the salt‐enhanced cultivation, are presented and discussed in this review. These techniques have already proven to be useful and now await further proof‐of‐concept. Furthermore, the mechanical behaviour of individual pellets is of special interest for a general understanding of pellet mechanics and the productivity of biotechnological processes with filamentous microorganisms. Correlating them with substrate uptake and finally with productivity would be a breakthrough not to be underestimated for the comprehensive characterisation of filamentous systems. So far, this research field is under‐represented. First results on filamentous pellet mechanics are discussed and important future aspects, which the filamentous expert community should deal with, will be presented and critically discussed.","{'model': 'tldr@v2.0.0', 'text': 'The mechanical behaviour of individual pellets is of special interest for a general understanding of pellet mechanics and the productivity of biotechnological processes with filamentous microorganisms.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/elsc.202000060
-2nd international conference on sustainable engineering techniques (icset 2019),,"The 2nd International Conference on Sustainable Engineering Techniques (ICSET2019) Preface The ICSET2019 is the 2nd International Conference on Sustainable Engineering Techniques took place in Baghdad, on March 6-7th, 2019 at the Middle Technical University, Institute of Technology-Baghdad (Iraq Republic). ICSET present a platform to support the futurity research and deliver a forum for a discussion of the latest findings, innovation of new Techniques and their implementations into different fields of Mechanical and materials, Civil, Architectural, environmental; Chemical and Petroleum; Electrical; Electronic, and Information Technologies engineering techniques. ICSET 2019 fulfilled its primary objective to provide an international platform for academics, researchers, engineers, and industrial pioneers share their findings with experts, gather the scientific research efforts and the sustainability techniques to conserve the natural resources for future generations to enhance life and stimulate new ideas in the research area of sustainability. The conference program covered presentations, oral, and video conference presentations with more than 383 active contributors. Finally, a sincere thanks to the Ministry of Higher Education & Scientific Research (Iraq), which provide extraordinary support represented by the rich social program started with an outstanding set up an opening ceremony of the conference. We would also like to extend our thanks to all authors for their precious research papers and presentations. All reviewers are also thanked for their professional assistance in reviewing. Thanks are finally given to IOP Publication as well for producing this volume. List of Conference Photograph and Committees are available in this PDF.",,https://iopscience.iop.org/article/10.1088/1757-899X/518/1/011001/pdf
-handbook of research on green engineering techniques for modern manufacturing,,"The Handbook of Research on Green Engineering Techniques for Modern Manufacturing provides emerging perspectives on the theoretical and practical aspects of green industrial concepts, such as green supply chain management and reverse logistics, for the sustainable utilization of resources and applications within manufacturing and engineering. Featuring coverage on a broad range of topics such as additive manufacturing, integrated manufacturing systems, and machine materials, this publication is ideally designed for engineers, environmental professionals, researchers, academicians, managers, policymakers, and graduate-level students seeking current research on recent and sustainable practices in manufacturing processes.",,
-promptor: a conversational and autonomous prompt generation agent for intelligent text entry techniques,Junxiao Shen,"Text entry is an essential task in our day-to-day digital interactions. Numerous intelligent features have been developed to streamline this process, making text entry more effective, efficient, and fluid. These improvements include sentence prediction and user personalization. However, as deep learning-based language models become the norm for these advanced features, the necessity for data collection and model fine-tuning increases. These challenges can be mitigated by harnessing the in-context learning capability of large language models such as GPT-3.5. This unique feature allows the language model to acquire new skills through prompts, eliminating the need for data collection and fine-tuning. Consequently, large language models can learn various text prediction techniques. We initially showed that, for a sentence prediction task, merely prompting GPT-3.5 surpassed a GPT-2 backed system and is comparable with a fine-tuned GPT-3.5 model, with the latter two methods requiring costly data collection, fine-tuning and post-processing. However, the task of prompting large language models to specialize in specific text prediction tasks can be challenging, particularly for designers without expertise in prompt engineering. To address this, we introduce Promptor, a conversational prompt generation agent designed to engage proactively with designers. Promptor can automatically generate complex prompts tailored to meet specific needs, thus offering a solution to this challenge. We conducted a user study involving 24 participants creating prompts for three intelligent text entry tasks, half of the participants used Promptor while the other half designed prompts themselves. The results show that Promptor-designed prompts result in a 35% increase in similarity and 22% in coherence over those by designers.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2310.08101
-applying model driven engineering techniques to the development of contiki-based iot systems,T. Asici,"The huge variety of smart devices and their communication models increases the development complexity of embedded software for the Internet of Things. As a consequence, development of these systems becomes more complex, error-prone, and costly. To tackle this problem, in this study, a model-driven approach is proposed for the development of Contiki-based IoT systems. To this end, the Contiki metamodel available in the literature is extended to include elements of WiFi connectivity modules (such as ESP8266), IoT Log Manager, and information processing components (such as Raspberry Pi). Based on this new metamodel, a domain-specific modeling environment is developed in which visual symbols are used and static semantics (representing system constraints) are defined. Also, the architectural code for the computing components of the IoT system such as Contiki, ESP8266, and RaspberryPi are generated from the developer's instance model. Finally, a Smart Fire Detection system is used to evaluate this study. By modeling the Contiki-based IoT system, we support model-driven development of the system, including WSN motes and sink nodes (with ContikiOS), WiFi modules and information processing components.","{'model': 'tldr@v2.0.0', 'text': 'By modeling the Contiki-based IoT system, this study supports model-driven development of the system, including WSN motes and sink nodes (with ContikiOS), WiFi modules and information processing components.'}",
-knee ligament injury and the clinical application of tissue engineering techniques: a systematic review.,Thomas C Riley,"BACKGROUND The incidence of knee ligament injury is increasing and represents a significant cost to healthcare providers. Current interventions include tissue grafts, suture repair and non-surgical management. These techniques have demonstrated good patient outcomes but have been associated graft rejection, infection, long term immobilization and reduced joint function. The limitations of traditional management strategies have prompted research into tissue engineering of knee ligaments. OBJECTIVE This paper aims to evaluate whether tissue engineering of knee ligaments offers a viable alternative in the clinical management of knee ligament injuries. A search of existing literature was performed using OVID Medline, Embase, AMED, PubMed and Google Scholar, and a manual review of citations identified within these papers. RESULTS Silk, polymer and extracellular matrix based scaffolds can all improve graft healing and collagen production. Fibroblasts and stem cells demonstrate compatibility with scaffolds, and have been shown to increase organized collagen production. These effects can be augmented using growth factors and extracellular matrix derivatives. Animal studies have shown tissue engineered ligaments can provide the biomechanical characteristics required for effective treatment of knee ligament injuries. CONCLUSION There is a growing clinical demand for a tissue engineered alternative to traditional management strategies. Currently, there is limited consensus regarding material selection for use in tissue engineered ligaments. Further research is required to optimize tissue engineered ligament production before clinical application. Controlled clinical trials comparing the use of tissue engineered ligaments and traditional management in patients with knee ligament injury could determine whether they can provide a cost-effective alternative.","{'model': 'tldr@v2.0.0', 'text': 'Control clinical trials comparing the use of tissue engineered ligaments and traditional management in patients with knee ligament injury could determine whether they can provide a cost-effective alternative.'}",
-genome scale engineering techniques for metabolic engineering.,Rongming Liu,,"{'model': 'tldr@v2.0.0', 'text': 'This review will attempt to summarize recent genome-scale design, build, test, and learn technologies and relate their use to a range of metabolic engineering applications.'}",http://manuscript.elsevier.com/S1096717615001238/pdf/S1096717615001238.pdf
-scaffold techniques and designs in tissue engineering functions and purposes: a review,A. Eltom,"In this review paper, the definition of the tissue engineering (TE) was comprehensively explored towards scaffold fabrication techniques and applications. Scaffold properties and features in TE, biological aspects, scaffold material composition, scaffold structural requirements, and old and current manufacturing technologies were reported and discussed. In almost all the reviewed reports, the TE definition denotes renewal, development, and repairs of damaged tissues caused by various factors such as disease, injury, or congenital disabilities. TE is multidisciplinary that combines biology, biochemistry, clinical medicine, and materials science whose application in cellular systems such as organ transplantation serves as a delivery vehicle for cells and drug. According to the previous literature and this review, the scaffold fabrication techniques can be classified into two main categories: conventional and modern techniques. These TE fabrication techniques are applied in the scaffold building which later on are used in tissue and organ structure. The benefits and drawbacks of each of the fabrication techniques have been described in conjunction with current areas of research devoted to deal with some of the challenges. To figure out, the highlighted aspects aimed to define the advancements and challenges that should be addressed in the scaffold design for tissue engineering. Additionally, this study provides an excellent review of original numerical approaches focused on mechanical characteristics that can be helpful in the scaffold design assessment in the analysis of scaffold parameters in tissue engineering.","{'model': 'tldr@v2.0.0', 'text': 'The definition of the tissue engineering (TE) was comprehensively explored and original numerical approaches focused on mechanical characteristics that can be helpful in the scaffold design assessment in the analysis of scaffold parameters in tissue engineering were provided.'}",http://downloads.hindawi.com/journals/amse/2019/3429527.pdf
-marine collagen scaffolds for nasal cartilage repair: prevention of nasal septal perforations in a new orthotopic rat model using tissue engineering techniques.,C. Bermueller,"Autologous grafts are frequently needed for nasal septum reconstruction. Because they are only available in limited amounts, there is a need for new cartilage replacement strategies. Tissue engineering based on the use of autologous chondrocytes and resorbable matrices might be a suitable option. So far, an optimal material for nasal septum reconstruction has not been identified. The aim of our study was to provide the first evaluation of marine collagen for use in nasal cartilage repair. First, we studied the suitability of marine collagen as a cartilage replacement matrix in the context of in vitro three dimensional cultures by analyzing cell migration, cytotoxicity, and extracellular matrix formation using human and rat nasal septal chondrocytes. Second, we worked toward developing a suitable orthotopic animal model for nasal septum repair, while simultaneously evaluating the biocompatibility of marine collagen. Seeded and unseeded scaffolds were transplanted into nasal septum defects in an orthotopic rat model for 1, 4, and 12 weeks. Explanted scaffolds were histologically and immunohistochemically evaluated. Scaffolds did not induce any cytotoxic reactions in vitro. Chondrocytes were able to adhere to marine collagen and produce cartilaginous matrix proteins, such as collagen type II. Treating septal cartilage defects in vivo with seeded and unseeded scaffolds led to a significant reduction in the number of nasal septum perforations compared to no replacement. In summary, we demonstrated that marine collagen matrices provide excellent properties for cartilage tissue engineering. Marine collagen scaffolds are able to prevent septal perforations in an autologous, orthotopic rat model. This newly described experimental surgical procedure is a suitable way to evaluate new scaffold materials for their applicability in the context of nasal cartilage repair.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that marine collagen matrices provide excellent properties for cartilage tissue engineering and are able to prevent septal perforations in an autologous, orthotopic rat model.'}",https://europepmc.org/articles/pmc3762606?pdf=render
-recent advancements in fruit detection and classification using deep learning techniques,C. Ukwuoma,"Recent advances in computer vision have allowed broad applications in every area of life, and agriculture is not left out. For the agri-food industry, the use of advanced technology is essential. Owing to deep learning’s capability to learn robust features from images, it has witnessed enormous application in several fields. Fruit detection and classification remains challenging due to the form, color, and texture of different fruit species. While studying the impact of computer vision on fruit detection and classification, we pointed out that till 2018 many conventional machine learning methods were utilized while a few methods exploited the application of deep learning methods for fruit detection and classification. This has prompted us to pursue an extensive study on surveying and implementing deep learning models for fruit detection and classification. In this article, we intensively discussed the datasets used by many scholars, the practical descriptors, the model’s implementation, and the challenges of using deep learning to detect and categorize fruits. Lastly, we summarized the results of different deep learning methods applied in previous studies for the purpose of fruit detection and classification. This review covers the study of recently published articles that utilized deep learning models for fruit identification and classification. Additionally, we also implemented from scratch a deep learning model for fruit classification using the popular dataset “Fruit 360” to make it easier for beginner researchers in the field of agriculture to understand the role of deep learning in the agriculture domain.","{'model': 'tldr@v2.0.0', 'text': 'This article intensively discussed the datasets used by many scholars, the practical descriptors, the model’s implementation, and the challenges of using deep learning to detect and categorize fruits, and summarized the results of different deep learning methods applied in previous studies for the purpose of fruit detection and classification.'}",https://downloads.hindawi.com/journals/mpe/2022/9210947.pdf
-model-driven engineering techniques for the development of multi-agent systems,J. M. Gascueña,,"{'model': 'tldr@v2.0.0', 'text': 'In this paper, agent-oriented software development (AOSD) and MDE paradigms are fully integrated for the development of MAS and meta-modeling techniques are explicitly used to speed up several phases of the process.'}",https://ruidera.uclm.es/bitstreams/be2ad086-b609-4cf8-8b6d-153796bb487a/download
-nondestructive testing and evaluation techniques of defects in fiber-reinforced polymer composites: a review,Jian Chen,"Fiber-reinforced polymer composites have excellent mechanical properties and outstanding development potential and are cost-effective. They have increasingly been used in numerous advanced and engineering applications as materials for wind turbine blades, helicopter rotors, high-pressure pipelines, and medical equipment. Understanding and assessing structural failure promptly in the whole lifecycle of a composite is essential to mitigating safety concerns and reducing maintenance costs. Various nondestructive testing and evaluation (NDT&E) technologies based on different evaluation principles have been established to inspect defects under different conditions. This paper reviews the established types of NDT&E techniques: acoustic emission, ultrasonic testing, eddy current testing, infrared thermography, terahertz testing, digital image correlation, shearography, and X-ray computed tomography, which is divided into three categories based on the operation frequency and data processing means of the output signal that is directly under analysis. We listed four types of defects/damage that are currently of great interest, namely, voids and porosity, fiber waviness and wrinkling, delamination and debonding, as well as impact damage. To identify a suitable method for different defects/damage, we performed characterization and evaluation by using these NDT&E techniques for typical defects/damage. Then, the cost, inspection speed, benefits and limitations, etc. were compared and discussed. Finally, a brief overview of the development of the technologies and their applications in the field of composite fabrication was discussed.",,https://www.frontiersin.org/articles/10.3389/fmats.2022.986645/pdf
-sterilization techniques for biodegradable scaffolds in tissue engineering applications,Zheng Dai,"Biodegradable scaffolds have been extensively studied due to their wide applications in biomaterials and tissue engineering. However, infections associated with in vivo use of these scaffolds by different microbiological contaminants remain to be a significant challenge. This review focuses on different sterilization techniques including heat, chemical, irradiation, and other novel sterilization techniques for various biodegradable scaffolds. Comparisons of these techniques, including their sterilization mechanisms, post-sterilization effects, and sterilization efficiencies, are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Comparisons of different sterilization techniques, including their sterilization mechanisms, post-sterilization effects, and sterilization efficiencies, are discussed.'}",https://journals.sagepub.com/doi/pdf/10.1177/2041731416648810
-analysing app reviews for software engineering: a systematic literature review,Jacek Dąbrowski,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of app review analysis research covering 182 papers published between 2012 and 2020 is presented, classifying it not only in terms of mined information and applied data mining techniques but also, and most importantly, in Terms of supported software engineering activities.'}",https://link.springer.com/content/pdf/10.1007/s10664-021-10065-7.pdf
-applying formal software engineering techniques to smart grids,Georg Hackenberg,"Engineering complex systems that have to meet critical requirements is a difficult task, especially if multiple engineering disciplines are involved. Common practice in domains like the automotive or avionic industry shows that formal methods improve engineering process efficiency for embedded software due to abilities like abstraction, early verification and iterative refinement. This paper presents how existing formal software engineering methods can be adapted to meet the needs of the smart grid domain. A case study demonstrates how we develop a basic interdisciplinary but semantically integrated decomposition of a household including electric and software behavior. Finally, we provide first simulation results to evaluate the feasibility of the model as well as the presented engineering method.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents how existing formal software engineering methods can be adapted to meet the needs of the smart grid domain and develops a basic interdisciplinary but semantically integrated decomposition of a household including electric and software behavior.'}",
-advances on defect engineering of vanadium‐based compounds for high‐energy aqueous zinc–ion batteries,Cong Guo,"Aqueous zinc–ion batteries (ZIBs) have been promptly developed as a competitive and promising system for future large‐scale energy storage. In recent years, vanadium (V)‐based compounds, with diversity of valences and high electrochemical‐activity, have been widely studied as cathodes for aqueous ZIBs because of their rich reserves and high theoretical capacity. However, the stubborn issues including low conductivity and sluggish kinetics, plague their smooth application in aqueous ZIBs. Among various countermeasures, defect engineering is believed as an effective method to alleviate the above limitations. This review highlights the challenges of different V‐based cathode materials (e.g., vanadium oxides and vanadates) and summarizes the advances in defect engineering strategies including types and effects of the defects, designed strategies, and characterization techniques for high‐energy ZIBs. Finally, several sound prospects in this fervent field are also rationally proposed for fundamental research and practical application.",,
-additive manufacturing techniques for the production of tissue engineering constructs,C. Mota,"‘Additive manufacturing’ (AM) refers to a class of manufacturing processes based on the building of a solid object from three‐dimensional (3D) model data by joining materials, usually layer upon layer. Among the vast array of techniques developed for the production of tissue‐engineering (TE) scaffolds, AM techniques are gaining great interest for their suitability in achieving complex shapes and microstructures with a high degree of automation, good accuracy and reproducibility. In addition, the possibility of rapidly producing tissue‐engineered constructs meeting patient's specific requirements, in terms of tissue defect size and geometry as well as autologous biological features, makes them a powerful way of enhancing clinical routine procedures. This paper gives an extensive overview of different AM techniques classes (i.e. stereolithography, selective laser sintering, 3D printing, melt–extrusion‐based techniques, solution/slurry extrusion‐based techniques, and tissue and organ printing) employed for the development of tissue‐engineered constructs made of different materials (i.e. polymeric, ceramic and composite, alone or in combination with bioactive agents), by highlighting their principles and technological solutions. Copyright © 2012 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'An extensive overview of different AM techniques employed for the development of tissue‐engineered constructs made of different materials, by highlighting their principles and technological solutions is given.'}",
-implementing systems engineering techniques into health care,Joe Fitzler,"Revision with unchanged content. Health care delivery in the United States needs improvement. Each year, between 44,000 and 98,000 people die as a result of medical errors and the United States is outranked by 3/4 of developed countries in life expectancy and infant mortality. In a recent report, The National Academy of Engineering and the Institute of Medicine recommend that implementing systems engineering and systems engineering tools should be used in health care to improve the industry. A unique way to implement systems thinking in health care is to educate future caregivers in systems thinking so that they can recognize problems and make improvements from within. The implementation of systems skills into medical school curriculum was tested at the University of Missouri School of Medicine, which has a Problem Based Learning curriculum. The lack of a resource for medical students to research health care systems engineering prompted the creation of a handbook entitled Handbook of Health Care Systems Engineering for Medical Students. This book is addressed to medical educators, medical students, and any other health professional looking to implement systems engineering into health care.","{'model': 'tldr@v2.0.0', 'text': 'The lack of a resource for medical students to research health care systems engineering prompted the creation of a handbook entitled Handbook of Health Care Systems Engineering for Medical Students.'}",
-forensic engineering surveys with uav photogrammetry and laser scanning techniques,C. Cappelletti,"Abstract. This work aims at presenting the use of new technologies in the field of forensic engineering. In particular, the use of UAV photogrammetry and laser scanning is compared with the traditional methods of surveying an accident site. In this framework, surveys must be carried out promptly, executed in a short time and performed so that the greatest possible amount of information is collected with sufficient accuracy to avoid the possibility of neglecting details once that the scene is no longer preserved. The combination of modern surveying techniques such UAV photogrammetry and laser scanning can properly fulfill these requirements. An experimental test has been arranged and instruments, procedures, settings, practical limits and results have been evaluated and compared with respect to the usual way of performing the survey for forensic purposes. In particular, both qualitative and quantitative considerations are given, assessing the completeness of the reconstructed model, the statistical evaluation of the errors and the accuracy achieved. ","{'model': 'tldr@v2.0.0', 'text': 'The use of UAV photogrammetry and laser scanning is compared with the traditional methods of surveying an accident site and qualitative and quantitative considerations are given, assessing the completeness of the reconstructed model, the statistical evaluation of the errors and the accuracy achieved.'}",https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLII-2-W9/227/2019/isprs-archives-XLII-2-W9-227-2019.pdf
-an empirical evaluation of requirement engineering techniques for collaborative systems,Miguel A. Teruel,"A collaborative system is a distributed software which allows several users to work together and carry out collaboration, communication and coordination tasks. To perform these tasks, the users have to be aware of other user's actions, usually by means of a set of awareness techniques. When we are defining a collaborative system, the awareness techniques can be considered as non-functional requirements bounded to some quality factors, such as usability. However, serious flaws can be found during the specification of these systems if we use the usual Requirement Engineering techniques available, because their expressiveness limitations when dealing with non-functional requirements. In this paper an empirical evaluation is introduced to determine if these techniques are really appropriate to model groupware requirements and which is the best approach to specify this kind of systems. With this aim, a collaborative text editor is used to evaluate whether the current techniques for Requirement Engineering are appropriated or not, exploiting the relation between awareness capabilities and standard quality factors.","{'model': 'tldr@v2.0.0', 'text': 'A collaborative text editor is used to evaluate whether the current techniques for Requirement Engineering are appropriated or not, exploiting the relation between awareness capabilities and standard quality factors.'}",
-what we should know before using tissue engineering techniques to repair injured tendons: a developmental biology perspective.,Chia-feng Liu,"Tendons connect muscles to bones, and serve as the transmitters of force that allow all the movements of the body. Tenocytes are the basic cellular units of tendons, and produce the collagens that form the hierarchical fiber system of the tendon. Tendon injuries are common, and difficult to repair, particularly in the case of the insertion of tendon into bone. Successful attempts at cell-based repair therapies will require an understanding of the normal development of tendon tissues, including their differentiated regions such as the fibrous mid-section and fibrocartilaginous insertion site. Many genes are known to be involved in the formation of tendon. However, their functional roles in tendon development have not been fully characterized. Tissue engineers have attempted to generate functional tendon tissue in vitro. However, a lack of knowledge of normal tendon development has hampered these efforts. Here we review studies focusing on the developmental mechanisms of tendon development, and discuss the potential applications of a molecular understanding of tendon development to the treatment of tendon injuries.","{'model': 'tldr@v2.0.0', 'text': 'The potential applications of a molecular understanding of tendon development to the treatment of tendon injuries are discussed and a lack of knowledge of normal tendon development has hampered these efforts.'}",https://europepmc.org/articles/pmc3098959?pdf=render
-analysis of requirements engineering techniques for it-enabled product service systems,M. Berkovich,"Product service systems (PSS) are introduced by many companies to increase their differentiation and to provide integrated solutions to customers. PSS are integrated solutions consisting of physical products, software and services aiming at providing an individualized solution to a customer's problem. The specific attributes of PSS lead to specific requirements for requirements engineering (RE). The goal of this paper is to analyze to which degree the analysis techniques of software engineering are suitable for PSS. We therefore conducted a structured literature review of software engineering techniques. The criteria for assessing the suitability of the techniques were based on the characteristics of PSS and the task of RE in the development process of PSS. We analyzed five textbooks and 144 scientific articles and identified 27 groups of techniques. The result is that there are major gaps in techniques for RE for PSS. Two of ten criteria are not satisfied by any technique. Moreover, for the majority of tasks of RE multiple techniques have to be combined for satisfying the criteria. In summary, the literature review shows that the techniques of software engineering are largely not directly applicable to PSS.","{'model': 'tldr@v2.0.0', 'text': 'The literature review shows that the techniques of software engineering are largely not directly applicable to PSS, with major gaps in techniques for RE for PSS.'}",
-intelligent computational optimization in engineering: techniques and applications,L. Nolle,,"{'model': 'tldr@v2.0.0', 'text': 'This Volume presents some of the latest developments with a focus on the design of algorithms for computational optimization and their applications in practice and provides various case studies of the application of intelligent optimization techniques in real-world applications.'}",
-chain-engineering-based de novo drug design against mpxvgp169 virulent protein of monkeypox virus: a molecular modification approach,M. Naveed,"The unexpected appearance of the monkeypox virus and the extensive geographic dispersal of cases have prompted researchers to concentrate on potential therapeutic approaches. In addition to its vaccine build techniques, there should be some multiple integrated antiviral active compounds because of the MPV (monkeypox virus) outbreak in 2022. This study offers a computational engineering-based de novo drug discovery mediated by random antiviral active compounds that were screened against the virulent protein MPXVgp169, as one of the key players directing the pathogenesis of the virus. The screening of these candidates was supported by the use of 72 antiviral active compounds. The top candidate with the lowest binding affinity was selected for the engineering of chains or atoms. Literature assisted to identify toxic chains or atoms that were impeding the stability and effectiveness of antiviral compounds to modify them for enhanced efficacy. With a binding affinity of −9.4 Kcal/mol after chain, the lipophilicity of 0.41, the water solubility of 2.51 as soluble, and synthetic accessibility of 6.6, chain-engineered dolutegravir was one of the best active compounds, as proved by the computational engineering analysis. This study will revolutionize the era of drug engineering as a potential therapeutic strategy for monkeypox infection.","{'model': 'tldr@v2.0.0', 'text': 'This study offers a computational engineering-based de novo drug discovery mediated by random antiviral active compounds that were screened against the virulent protein MPXVgp169, as one of the key players directing the pathogenesis of the virus.'}",https://www.mdpi.com/2306-5354/10/1/11/pdf?version=1671620039
-a systematic review of the use of requirements engineering techniques in model-driven development,Grzegorz Loniewski,,"{'model': 'tldr@v2.0.0', 'text': 'A systematic review of the current use of requirements engineering techniques in MDD processes and their actual automation level shows that although MDD techniques are used to a great extent in platform-independent models, platform-specific models, and at code level, at the requirements level most MDD approaches use only partially defined requirements models or even natural language.'}",
-a chaos engineering approach for improving the resiliency of it services configurations,Filippo Poltronieri,"Testing the resiliency of complex IT services deployed in hybrid Cloud scenarios is a challenging task that requires expensive and possibly destructive operations. An interesting approach lies in Chaos Engineering, a set of practices to test the resiliency of software systems running in a production environment. However, Chaos Engineering is an expensive practice that requires the setup of complicated operations that further increase the complexity of management operations. To reduce this complexity, Chaos Engineering can benefit from the adoption of non-destructive approaches such as the definition of realistic digital twins. A digital twin is a virtual replica of a real-system on which experimenting with management configurations. This paper embraces this research avenue by extending our previous efforts to integrate Chaos Engineering techniques into an IT services management framework called ChaosTwin. ChaosTwin leverages novel methodologies and tools capable of identifying and promptly react to unexpected failures. Finally, to implement autonomous fault management, ChaosTwin defines scaling and migration policies that can quickly explore for more resilient placements of software components in case of system failures. We believe that ChaosTwin can provide useful guidance to service providers in finding cost-effective service configurations capable of minimizing the negative effects of unpredictable events.","{'model': 'tldr@v2.0.0', 'text': 'ChaosTwin can provide useful guidance to service providers in finding cost-effective service configurations capable of minimizing the negative effects of unpredictable events and leverages novel methodologies and tools capable of identifying and promptly react to unexpected failures.'}",
-advances in tissue engineering techniques for articular cartilage repair.,A. Haleem,"The limited repair potential of human articular cartilage contributes to development of debilitating osteoarthritis and remains a great clinical challenge. This has led to evolution of cartilage treatment strategies from palliative to either reconstructive or reparative methods in an attempt to delay or ""bridge the gap"" to joint replacement. Further development of tissue engineering-based cartilage repair methods have been pursued to provide a more functional biological tissue. Currently, tissue engineering of articular cartilage has three cornerstones; a cell population capable of proliferation and differentiation into mature chondrocytes, a scaffold that can host these cells, provide a suitable environment for cellular functioning and serve as a sustained-release delivery vehicle of chondrogenic growth factors and thirdly, signaling molecules and growth factors that stimulate the cellular response and the production of a hyaline extracellular matrix (ECM). The aim of this review is to summarize advances in each of these three fields of tissue engineering with specific relevance to surgical techniques and technical notes.","{'model': 'tldr@v2.0.0', 'text': 'The aim of this review is to summarize advances in each of these three fields of tissue engineering with specific relevance to surgical techniques and technical notes.'}",https://europepmc.org/articles/pmc5805482?pdf=render
-a benchmark for evaluating software engineering techniques for improving medical processes,Stefan Christov,"The software engineering and medical informatics communities have been developing a range of approaches for reasoning about medical processes. To facilitate the comparison of such approaches, it would be desirable to have a set of medical examples, or benchmarks, that are easily available, described in considerable detail, and characterized in terms of the real-world complexities they capture. This paper presents one such benchmark and discusses a list of desiderata that medical benchmarks can be evaluated against.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a set of medical examples, or benchmarks, that are easily available, described in considerable detail, and characterized in terms of the real-world complexities they capture and discusses a list of desiderata that medical benchmarks can be evaluated against.'}",http://laser.cs.umass.edu/techreports/10-023.pdf
-multiscale and local engineering of speckle morphology through disordered media.,Chunxu Zhang,"In this Letter, we prompt a novel, to the best of our knowledge, method based on transmission matrix decomposition with wavelets to engineer the speckle morphology behind disordered media. By analyzing the speckles in multiscale spaces, we experimentally realized multiscale and localized control on the speckle size, position-dependent spatial frequency, and global morphology by operating on the decomposition coefficients using different masks. Speckles with contrasting features in different parts of the fields can be generated in one step. Our experimental results demonstrate a high degree of flexibility in manipulating light in a customizable manner. This technique has stimulating prospects in correlation control and imaging under scattering conditions.",,
-retrofitting of rc beams using frp techniques: a review,A. Sharba,"In structural and civil engineering, reinforced concrete structures are susceptible to inordinate loadings such as earthquake disasters, terrorist attacks and accidental impacts. This prompts an increasing desire to strengthen as well as enhance the fatigue performance and extend the ability of fatigue life of reinforced concrete structural components especially beams. Since 2003, Iraq has suffered major damage to the structural buildings, so it has become an important issue to study the existing concrete structures. Among the different techniques of rehabilitation of existing structures, fibre reinforced polymer (FRP) as an external bonding has been considered as a popular one. This paper reviews several features of Reinforced concrete (RC) beams strengthened with FRP. Also this paper aims to impart a comprehensive insight on adhesive curing, surface arrangement, and failure modes of RC beams modified with FRP. This effect of FRP for enhancing the techniques of rehabilitation is a three-fold task, to strengthen and retrofit of concrete structures, to extend the fatigue life of the structural element, and eliminates the crack growth rate.",,
-"internet-based social engineering attacks, defenses and psychology: a survey",Theodore Longtchi,"Social engineering attacks are a major cyber threat because they often serve as a first step for an attacker to break into an otherwise well-defended network, steal victims' credentials, and cause financial losses. The problem has received due amount of attention with many publications proposing defenses against them. Despite this, the situation has not improved. In this paper, we aim to understand and explain this phenomenon by looking into the root cause of the problem. To this end, we examine the literature on attacks and defenses through a unique lens we propose -- {\em psychological factors (PFs) and techniques (PTs)}. We find that there is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions. This explains why existing defenses have achieved limited success. This prompts us to propose a roadmap for a more systematic approach towards designing effective defenses against social engineering attacks.","{'model': 'tldr@v2.0.0', 'text': 'There is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions, which explains why existing defenses have achieved limited success.'}",
-sok: why have defenses against social engineering attacks achieved limited success?,Theodore Longtchi,"Social engineering attacks are a major cyber threat because they often serve as a first step for an attacker to break into an otherwise well-defended network, steal victims’ credentials, and cause financial losses. The problem has received due amount of attention with many publications proposing defenses against them. Despite this, the situation has not improved. In this SoK paper, we aim to understand and explain this phenomenon by looking into the root cause of the problem. To this end, we examine the literature on attacks and defenses through a unique lens we propose — psychological factors (PFs) and techniques (PTs). We find that there is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions. This explains why existing defenses have achieved limited success. This prompts us to propose a roadmap for a more systematic approach towards designing effective defenses against social engineering attacks.","{'model': 'tldr@v2.0.0', 'text': 'There is a big discrepancy between attacks and defenses: Attacks have deliberately exploited PFs by leveraging PTs, but defenses rarely take either of these into consideration, preferring technical solutions, which explains why existing defenses have achieved limited success.'}",http://arxiv.org/pdf/2203.08302
-reverse engineering techniques for cranioplasty: a case study,E. Maravelakis,This paper presents rapid prototyping and reverse engineering techniques applied to create an implant for the surgical reconstruction of a large cranial defect. A series of computed tomography (CT) images was obtained and purpose built software was used to extract the cranial geometry in a point cloud. The point cloud produced was used for: (a) the creation of a stereolithographic (STL) physical model for direct assessment of the cranial defect; and (b) the creation of a 3D mould model for the fabrication of the patient-specific implant.,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents rapid prototyping and reverse engineering techniques applied to create an implant for the surgical reconstruction of a large cranial defect.'}",
-"balancing agility and formalism in software engineering, second ifip tc 2 central and east european conference on software engineering techniques, cee-set 2007, poznan, poland, october 10-12, 2007, revised selected papers",B. Meyer,,"{'model': 'tldr@v2.0.0', 'text': 'In this age of modern era, the use of internet must be maximized, as one of the benefits is to get the on-line balancing agility and formalism in software engineering.'}",
-emerging needs and considerations for digital engineering software tools,Jason Boss,"ion of models would benefit from some industry-established descriptions for content for some of the most commonly employed use cases for exchanging information, such that the amount of information is sufficient to perform the agreed-to tasks but obscure the additional IP or part of the agreement that is not needed. An example of this may be an “interface control” model, which depicts the overall geometry and functional behaviors, etc., of a system component but does not provide any insights into how it actually works at the lower level, or a “build to print” model, which provides enough information to manufacture a part but does not provide any design intent or analysis techniques that went into arriving at the final configuration. Development of a standard set of use cases and abstraction techniques for the most common customer and supply chain interactions that could be used as a convention might be considered for future iteroparity standardization. There have been several techniques established for data transmission that could be further explored for feasibility and a number of them that are available today, but due to their complexity and cost to tailor these implementations, they are normally limited in use to larger program installations, technologies such as digital rights management, block chain, etc. Finally, tools that leverage large IP data sets to be “trained” to gain insights or supply tailored results should be disclosed and optioned to not retain or feedback information outside of the owning organization’s environment when back-feeding insights to software supplier-hosted environments could impose IP or security concerns.","{'model': 'tldr@v2.0.0', 'text': None}",
-engineering of heterostructure pt/co/alox for the enhancement of dyzaloshinskii-moria interaction,B. R. Sankhi,"The interfacial Dyzaloshinskii-Moria interaction (DMI) helps to stabilize chiral domain walls and magnetic skyrmions, which will facilitate new magnetic memories and spintronics logic devices. The study of interfacial DMI in perpendicularly magnetized structurally asymmetric heavy metal (HM) / ferromagnetic (FM) multilayer systems is of high importance due to the formation of chiral magnetic textures in the presence of DMI. Here, we report the impact of the cobalt oxidation at the cobalt -aluminum oxide interface in Pt/Co/AlOxtrilayer structure on the DMI by varying the post-growth annealing time and Aluminum thickness. For quantifying DMI, we employed magneto-optical imaging of asymmetric domain wall expansion, hysteresis loop shift, and spin-wave spectroscopy techniques. We further correlated the Cobalt oxidation with low-temperature Hall effect measurements and X-ray photoelectron spectroscopy. Our results emphasize the characterization of magnetic films for MRAM technologies semiconductor temperature process window, where magnetic interaction will be critical for device performance.",,
-a case study validation of a knowledge-based approach for the selection of requirements engineering techniques,Li Jiang,,"{'model': 'tldr@v2.0.0', 'text': 'A Knowledge-based Approach for the Selection of Requirements Engineering Techniques (KASRET) is proposed that helps during RE techniques selection and makes not only a contribution to RE but also to research and application of knowledge management and decision support in process development.'}",
-leveraging reddit for suicidal ideation detection: a review of machine learning and natural language processing techniques,Eldar Yeskuatov,"Suicide is a major public-health problem that exists in virtually every part of the world. Hundreds of thousands of people commit suicide every year. The early detection of suicidal ideation is critical for suicide prevention. However, there are challenges associated with conventional suicide-risk screening methods. At the same time, individuals contemplating suicide are increasingly turning to social media and online forums, such as Reddit, to express their feelings and share their struggles with suicidal thoughts. This prompted research that applies machine learning and natural language processing techniques to detect suicidality among social media and forum users. The objective of this paper is to investigate methods employed to detect suicidal ideations on the Reddit forum. To achieve this objective, we conducted a literature review of the recent articles detailing machine learning and natural language processing techniques applied to Reddit data to detect the presence of suicidal ideations. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we selected 26 recent studies, published between 2018 and 2022. The findings of the review outline the prevalent methods of data collection, data annotation, data preprocessing, feature engineering, model development, and evaluation. Furthermore, we present several Reddit-based datasets utilized to construct suicidal ideation detection models. Finally, we conclude by discussing the current limitations and future directions in the research of suicidal ideation detection.",,https://www.mdpi.com/1660-4601/19/16/10347/pdf?version=1661158040
-adoption and effects of software engineering best practices in machine learning,A. Serban,"Background. The increasing reliance on applications with machine learning (ML) components calls for mature engineering techniques that ensure these are built in a robust and future-proof manner. Aim. We aim to empirically determine the state of the art in how teams develop, deploy and maintain software with ML components. Method. We mined both academic and grey literature and identified 29 engineering best practices for ML applications. We conducted a survey among 313 practitioners to determine the degree of adoption for these practices and to validate their perceived effects. Using the survey responses, we quantified practice adoption, differentiated along demographic characteristics, such as geography or team size. We also tested correlations and investigated linear and non-linear relationships between practices and their perceived effect using various statistical models. Results. Our findings indicate, for example, that larger teams tend to adopt more practices, and that traditional software engineering practices tend to have lower adoption than ML specific practices. Also, the statistical models can accurately predict perceived effects such as agility, software quality and traceability, from the degree of adoption for specific sets of practices. Combining practice adoption rates with practice importance, as revealed by statistical models, we identify practices that are important but have low adoption, as well as practices that are widely adopted but are less important for the effects we studied. Conclusion. Overall, our survey and the analysis of responses received provide a quantitative basis for assessment and step-wise improvement of practice adoption by ML teams.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate, for example, that larger teams tend to adopt more practices, and that traditional software engineering practices tend to have lower adoption than ML specific practices.'}",https://scholarlypublications.universiteitleiden.nl/access/item%3A3307601/view
-applying knowledge engineering techniques to customer analysis in the service industry,Sung-ho Ha,,"{'model': 'tldr@v2.0.0', 'text': 'This study differentiates between customers through customer segmentation, tracks customer shifts from segment to segment over time, discovers customer segment knowledge to build an individual transition path and a dominant transition path, and then predicts customer segment behavior patterns.'}",
-oriented porous polymer scaffolds in tissue engineering: a comprehensive review of preparation strategies and applications,Tong Liu,"The pursuit of effective therapeutic strategies for tissue damage has prompted extensive scholarly investigations worldwide. Tissue engineering has emerged as a prominent approach, particularly through the utilization of artificial scaffolds that closely resemble the natural extracellular matrix (ECM). These scaffolds exhibit multi‐scale topological structures and surface physicochemical properties, which significantly influence cellular behavior, thereby attracting considerable attention from numerous researchers. This comprehensive review is concentrated on the primary techniques employed in the fabrication of biodegradable polymer scaffolds possessing oriented porous structures, and the most recent advancements in tissue engineering research are presented. Significantly, the profound influence of scaffold surface characteristics is underscored on cellular behavior, elucidating the superiority of oriented pore structures over disordered ones in mimicking the distinctive attributes of the ECM. Enhanced cell adhesion, proliferation, and tissue differentiation represent notable advantages associated with oriented porous scaffolds. Additionally, the critical interplay between scaffold structure, performance, and functionalization is emphasized, highlighting the imperative to optimize the clinical application of tissue engineering scaffolds.","{'model': 'tldr@v2.0.0', 'text': 'The profound influence of scaffold surface characteristics is underscored on cellular behavior, elucidating the superiority of oriented pore structures over disordered ones in mimicking the distinctive attributes of the ECM.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/mame.202300246
-testing openfoam computational fluid dynamics simulation of heat transfer and fluid flow in a mechanical engineering bachelor degree,Marina Duarte,"Nowadays, countless areas in the domain of mechanical engineering practice can benefit from the use of Computational Fluid Dynamics techniques. Currently, at the Energy branch of the master’s degree in mechanical engineering at Instituto Superior de Engenharia do Porto, Computational Fluid Dynamics is only formally addressed in the first semester of the final year. In many cases, this prevents the students from promptly acquiring the necessary knowledge to embrace a master’s thesis or research project in this field. A possible way to minimise this problem was tested by exposing the students to the Computational Fluid Dynamics world earlier. Thus, in the curricular year 2022/2023, two optional simulation cases with a 35% weight in the final grade of a Heat Transfer course were proposed to the third-year students of the bachelor’s degree.For this purpose, an educational intervention following an action research methodology was implemented. A questionnaire was used to collect students’ feedback. Qualitative grades of the project reports and the discussion sessions corroborated students learning. Considering the students’ grades and that their feedback was in accordance with the complexity of the two simulation cases, we conclude that this education intervention was successful and that Computational Fluid Dynamics can be introduced earlier.",,
-"artificial intelligence in engineering and society: blue skies, black holes, and the job of requirements engineers (keynote)",Alessio Ferrari,"The democratization of artificial intelligence (AI) has brought substantial achievements in science, engineering disciplines, and society as a whole. New technologies based on large language models, multi-modal learning, embodied AI, and the quest for artificial general intelligence (AGI) promise to further change the world's landscape as we know it. At the same time, AI's rapid and uncontrolled evolution also poses serious risks to society, such as the concentration of power, exclusion, discrimination, and manipulation of reality. The keynote will present some experiences in AI democratization, including the usage of explainable machine learning approaches for agronomists, NLP-based solutions for railway engineers, image processing techniques for the maintenance of riverbeds, and mobile data processing in road safety assessment. The talk will outline the latest technological advancements in AI, e.g., in healthcare and science, and will show how large language models like ChatGPT and Bing Chat can solve long-standing requirements engineering (RE) problems. For example, requirements completeness can be easily checked and addressed with simple prompts, and model generation from requirements becomes a one-click task. The keynote will then describe the risks that current AI development poses to society. Besides the increasingly convincing deep fakes, and the widely discussed risks for privacy and reputation, we must be aware of the uncontrolled speed of AI evolution. As AI continues to advance, it will replace many jobs that require intellectual skills. This could lead to a significant number of people losing their jobs, as they may not have the necessary skills to adapt to the new labour market. People and entire countries that cannot exploit technological developments will be excluded from the game, and this will cause resentment and the possible emergence of new fundamentalism. The race for semiconductors is already creating hot spots and rifts between the superpowers. In this context, RE researchers are called to new technical and societal challenges. With pieces of code and even entire programs that can be automatically generated with large language models, the craft of prompting becomes the new requirements specification, and the concept of structured APIs dissolves into natural language interfaces. At the societal level, AI regulations are making their first steps, and we are called to contribute to operationalise the norms while preventing over-regulation. Equipped with years of experience at the boundary of the technical and social facets of systems, RE researchers are pivotal subjects in the new golden age of AI.","{'model': 'tldr@v2.0.0', 'text': 'The talk will outline the latest technological advancements in AI, e.g., in healthcare and science, and will show how large language models like ChatGPT and Bing Chat can solve long-standing requirements engineering (RE) problems, and describe the risks that current AI development poses to society.'}",
-selecting requirements engineering techniques based on project attributes--a case study,Li Jiang,"Selection of the most appropriate RE techniques for a software project based on the project's characteristics is a non-trivial process and a common challenge faced by software developers. In order to facilitate RE techniques selection, we propose a model for RE technique suitability assessment (RETSA) based on project attributes. This model was developed based on surveys and interviews of experienced software developers and experts from industry and academia. RETSA was applied to an industrial project. This case study shows the help this model provided during the selection of RE techniques for a software project","{'model': 'tldr@v2.0.0', 'text': 'A model for RE technique suitability assessment (RETSA) based on project attributes was developed based on surveys and interviews of experienced software developers and experts from industry and academia and applied to an industrial project.'}",
-"reverse engineering techniques applied to a human skull, for cad 3d reconstruction and physical replication by rapid prototyping",L. Galantucci,"The production of a copy of an existing object of complex shape is one of the typical applications of the integration between two modern computer-based technologies, reverse engineering (RE) and rapid prototyping (RP). The method is extremely versatile and can be used in various applicative domains (e.g. replacement of anatomical parts with artificial prostheses, replication of skeletal remains). Two different acquisition techniques of images of a skull, by laser and by CT scan, were compared to ascertain which enabled more accurate reproduction of the original specimen. The skull was chosen due to it being the body part most often used in medico-legal investigations (for personal identification, skull-photo superimposition techniques, forensic art, etc). Comparison between the copy and the original yielded satisfactory results for both techniques. However, CT scanning demonstrated some advantages over the laser technique, as it provided a cleaner point cloud, enabling shorter pre-reproduction processing times, as well as data on the internal parts, which resulted in the reproduction of a more faithful copy.","{'model': 'tldr@v2.0.0', 'text': 'Comparison between the copy and the original yielded satisfactory results for both techniques, however, CT scanning demonstrated some advantages over the laser technique, as it provided a cleaner point cloud, enabling shorter pre-reproduction processing times, as well as data on the internal parts, which resulted in the reproduction of a more faithful copy.'}",
-combining requirements engineering techniques - theory and case study,Li Jiang,"The selection of requirements engineering (RE) techniques during software project development is a challenge for most developers. One of the reasons is that there is a great lack of requirements engineering education in most academic programs, so software developers have to learn requirements engineering practices on the job. This can easily result in the selection of techniques that are ill-suited for a particular project, as the selection is based on personal preference rather than on the characteristics of the project. Very little research has been done in the area of technique selection based on project attributes. This paper describes research into the selection and combination of RE techniques as well as a case study that applied the selection process to an industrial software project.","{'model': 'tldr@v2.0.0', 'text': 'Research into the selection and combination of RE techniques as well as a case study that applied the selection process to an industrial software project are described.'}",https://digital.library.adelaide.edu.au/dspace/bitstream/2440/45364/1/hdl45364.pdf
-the development of power system restoration method for a bulk power system by applying knowledge engineering techniques,Y. Kojima,"A prototype expert system for bulk power system restoration is discussed. Restoration of a bulk power system involves operation of equipment that is complexly inter-related. For this reason, object-oriented programming techniques using a frame representation of knowledge and heuristics has been adopted to achieve efficient processing. The use of knowledge representation and processing is described, and the restoration process is examined. A prototype system that has successfully satisfied the requirements of speed, stability and flexibility is described. >","{'model': 'tldr@v2.0.0', 'text': 'A prototype expert system for bulk power system restoration using object-oriented programming techniques using a frame representation of knowledge and heuristics to achieve efficient processing is discussed.'}",
-requirements for requirements engineering techniques,Linda A. Macaulay,"The first argument presented is that if we have a clear understanding of the objectives of the requirements engineering (RE) process then we can identify what techniques we need. No single method or technique will be sufficient. The second argument is that a key objective of the RE process is to specify a system which will ultimately be successful. Three common types of failure are examined: process failure, expectation failure and interaction failure. The third argument is that if the causes of each type of failure can be described then we will be able to identify what techniques are needed to help us avoid failure. The five common causes are described: the requirements engineering process itself, human communication within requirements, knowledge development, documentation of requirements and management. A discussion of each cause is followed by a list of the RE techniques needed. The paper concludes with a 'wish list' of seventy requirements for RE techniques, and a brief discussion of the strengths and weaknesses of the RE community in meeting those requirements.","{'model': 'tldr@v2.0.0', 'text': 'The first argument presented is that if the authors have a clear understanding of the objectives of the requirements engineering (RE) process then they can identify what techniques they need to help us avoid failure and no single method or technique will be sufficient.'}",
-behavioral characteristics of magnesium as a biomaterial for surface engineering application,K. O. Babaremu,,,
-investigating university students’ conceptions of engineering: an implied identity perspective,M. Khosronejad,"ABSTRACT Examining how learners ascribe meaning to different aspects of their university life is a necessary undertaking in higher education. This article aims to investigate university students’ conceptions of engineering and explore how these conceptions relate to the context of their experience. Our data collection method and data analysis are informed by the techniques of the phenomenographic approach and the Implied Identity Framework. Fifteen engineering students in a course about sustainability participated in the study. Results revealed eight categories of participants’ conceptions: (1) engineering as a practice for designing solutions; (2) engineering as a pragmatic practice for the welfare of people; (3) engineering as a knowledge-based practice; (4) engineering as a communicative practice; (5) engineering as a technology-mediated practice; (6) engineering as thinking; (7) engineering as an independent practice; and (8) engineering as learning. These conceptions had cognitive, technological, or social orientations and were related to three aspects of engineering: outcome-focused; process-focused; and person-focused. Also, findings demonstrated that different contexts of the university, workplace, and society prompt different ways of thinking about engineering. The findings place a greater emphasis on personal agency and the development of student identities as lifelong learners.",,
-complementary techniques to analyse pericellular matrix formation by human msc within hyaluronic acid hydrogels,C. Salzlechner,"Hydrogels are widely used as mimics of the native extracellular matrix as their physical and biological properties can be tuned over a wide range to match those of the native tissue. Cells encapsulated within hydrogels have recently been reported to modify their local surroundings by secreting and assembling proteins pericellularly, which in turn impacts their fate. As a result, methods to characterise and visualise the secreted matrix are becoming increasingly important in the development of regenerative therapies and in understanding cell behaviour within 3D matrices. Here, by combining fluorescent non-canonical amino acid tagging with Raman spectral imaging, we aimed to create 3D maps of human mesenchymal stromal cells (hMSC) and their secreted matrix when embedded within hydrogels. To demonstrate the value of our combined technique in a tissue engineering context, we cultured hMSC in Dopa-modified hyaluronic acid-based hydrogels and treated cultures with the 2-oxyglutarate analogue dimethyloxalyglycine (DMOG), which mimics the cellular effects of physiological hypoxia and can both promote the chondrogenic differentiation of progenitor cells and enhance cartilage-like matrix formation. Quantitative analyses of the distribution of newly synthesised proteins combined with principal components analyses of Raman spectra showed that DMOG prompted encapsulated cells to secrete more protein pericellularly than did untreated controls. Our findings demonstrate that it is possible to visualise both the 3D secreted matrix, as well as cellular contents using simple, unbiased, inexpensive techniques, providing complementary information on cells and their secreted matrix when encapsulated within 3D hydrogels.","{'model': 'tldr@v2.0.0', 'text': '3D maps of human mesenchymal stromal cells (hMSC) and their secreted matrix when embedded within hydrogels are created by combining fluorescent non-canonical amino acid tagging with Raman spectral imaging to demonstrate that it is possible to visualise both the 3Dsecreted matrix, as well as cellular contents using simple, unbiased, inexpensive techniques.'}",https://pubs.rsc.org/en/content/articlepdf/2020/ma/d0ma00472c
-a comparative study of power electronics and control techniques for renewable energy integration in smart grids,Shashank Pareek,"The development of smart grid technology for more effective and sustainable energy management has been prompted by the growing use of renewable energy sources like solar and wind power. In order to integrate renewable energy sources into the grid, power electronics and control approaches are essential. The purpose of this study is to examine the pros and cons of the various power electronics and control approaches now in use for integrating renewable energy sources into smart grids. The development of smart grid technology for more effective and sustainable energy management has been facilitated by the use of renewable energy sources like solar and wind power. When it comes to incorporating renewable energy sources into the grid, power electronics and control approaches play a crucial role. This article provides a comprehensive analysis of the literature on the topic of renewable energy integration in smart grids, with a particular focus on power electronics and control strategies. The study is based on a review of relevant literature and an analysis of the benefits and drawbacks of each method. Efficiency, dependability, cost-effectiveness, scalability, and compatibility with various renewable energy sources are some of the characteristics included in the comparison study. The research shows that each method has advantages and disadvantages, and that the choice of method is context-specific. The work has significant significance for the field of electrical and computer engineering and aids in the development of more efficient and sustainable energy management systems for the future smart grid.",,
-estimating and optimizing safety factors of retaining wall through neural network and bee colony techniques,B. Gordan,,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents the advantages of both prediction and optimization of retaining wall SF through artificial neural network (ANN) and artificial bee colony (ABC), respectively.'}",
-three-dimensional scaffolds for tissue engineering applications: role of porosity and pore size.,Qiu Li Loh,"Tissue engineering applications commonly encompass the use of three-dimensional (3D) scaffolds to provide a suitable microenvironment for the incorporation of cells or growth factors to regenerate damaged tissues or organs. These scaffolds serve to mimic the actual in vivo microenvironment where cells interact and behave according to the mechanical cues obtained from the surrounding 3D environment. Hence, the material properties of the scaffolds are vital in determining cellular response and fate. These 3D scaffolds are generally highly porous with interconnected pore networks to facilitate nutrient and oxygen diffusion and waste removal. This review focuses on the various fabrication techniques (e.g., conventional and rapid prototyping methods) that have been employed to fabricate 3D scaffolds of different pore sizes and porosity. The different pore size and porosity measurement methods will also be discussed. Scaffolds with graded porosity have also been studied for their ability to better represent the actual in vivo situation where cells are exposed to layers of different tissues with varying properties. In addition, the ability of pore size and porosity of scaffolds to direct cellular responses and alter the mechanical properties of scaffolds will be reviewed, followed by a look at nature's own scaffold, the extracellular matrix. Overall, the limitations of current scaffold fabrication approaches for tissue engineering applications and some novel and promising alternatives will be highlighted.","{'model': 'tldr@v2.0.0', 'text': ""The ability of pore size and porosity of scaffolds to direct cellular responses and alter the mechanical properties of scaffold will be reviewed, followed by a look at nature's own scaffold, the extracellular matrix.""}",https://europepmc.org/articles/pmc3826579?pdf=render
-a survey of app store analysis for software engineering,William J. Martin,"App Store Analysis studies information about applications obtained from app stores. App stores provide a wealth of information derived from users that would not exist had the applications been distributed via previous software deployment methods. App Store Analysis combines this non-technical information with technical information to learn trends and behaviours within these forms of software repositories. Findings from App Store Analysis have a direct and actionable impact on the software teams that develop software for app stores, and have led to techniques for requirements engineering, release planning, software design, security and testing. This survey describes and compares the areas of research that have been explored thus far, drawing out common aspects, trends and directions future research should take to address open problems and challenges.","{'model': 'tldr@v2.0.0', 'text': 'This survey describes and compares the areas of research that have been explored thus far, drawing out common aspects, trends and directions future research should take to address open problems and challenges.'}",
-advances in porous scaffold design for bone and cartilage tissue engineering and regeneration.,A. Cheng,"IMPACT STATEMENT Challenges in musculoskeletal tissue regeneration affect millions of patients globally. Scaffolds for tissue engineering bone and cartilage provide promising solutions that increase healing and decrease need for complicated surgical procedures. Porous scaffolds have emerged as an attractive alternative to traditional scaffolds. However, the success of advanced materials, use of biological factors, and manufacturing techniques can vary depending on use case. This review provides perspective on porous scaffold manufacturing, characterization and application, and can be used to inform future scaffold design.","{'model': 'tldr@v2.0.0', 'text': None}",https://www.liebertpub.com/doi/pdf/10.1089/ten.teb.2018.0119
-grounded theory in software engineering research: a critical review and guidelines,Klaas-Jan Stol,"Grounded Theory (GT) has proved an extremely useful research approach in several fields including medical sociology, nursing, education and management theory. However, GT is a complex method based on an inductive paradigm that is fundamentally different from the traditional hypothetico-deductive research model. As there are at least three variants of GT, some ostensibly GT research suffers from method slurring, where researchers adopt an arbitrary subset of GT practices that are not recognizable as GT. In this paper, we describe the variants of GT and identify the core set of GT practices. We then analyze the use of grounded theory in software engineering. We carefully and systematically selected 98 articles that mention GT, of which 52 explicitly claim to use GT, with the other 46 using GT techniques only. Only 16 articles provide detailed accounts of their research procedures. We offer guidelines to improve the quality of both conducting and reporting GT studies. The latter is an important extension since current GT guidelines in software engineering do not cover the reporting process, despite good reporting being necessary for evaluating a study and informing subsequent research.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the variants of GT and identifies the core set of GT practices, then analyzes the use of grounded theory in software engineering and offers guidelines to improve the quality of both conducting and reporting GT studies.'}",https://researchrepository.ul.ie/articles/conference_contribution/Grounded_theory_in_software_engineering_research_a_critical_review_and_guidelines/19852003/1/files/35268280.pdf
-cell-seeding techniques in vascular tissue engineering.,Gustavo A. Villalona,"Previous studies have demonstrated the benefits of cell seeding in the construction of tissue-engineered vascular grafts (TEVG). However, seeding methods are diverse and no method is clearly superior in either promoting seeding efficiency or improving long-term graft function. As we head into an era during which a variety of different TEVG are under investigation in clinical trials around the world, it is important to consider the regulatory issues surrounding the translation of these technologies. In this review, we summarize important advances in the field of vascular tissue engineering, with particular attention on cell-seeding techniques for TEVG development and special emphasis placed on regulatory issues concerning the clinical translation of these various methods.","{'model': 'tldr@v2.0.0', 'text': 'Important advances in the field of vascular tissue engineering are summarized, with particular attention on cell-seeding techniques for TEVG development and special emphasis placed on regulatory issues concerning the clinical translation of these various methods.'}",https://europepmc.org/articles/pmc2946885?pdf=render
-ontology engineering,Elisa F. Kendall,"54 Computer Integrated Approach to Web Ontology Learning and Engineering T he development of the semantic Web 1 — which seeks to improve the semantic awareness of computers connected via the Internet—requires a systematic, computer-oriented representation of the world. Researchers often refer to such a world model as an ontology. Despite the significant amount of work done on them in recent years, ontologies have yet to be widely applied and used. Research to date has mainly addressed the basic principles, such as knowledge representation formalisms, devoting only limited attention to more practical issues such as techniques and tools aimed at an ontology's actual construction and content. We have developed a software environment, centered around the OntoLearn tool, that can build and assess a domain ontology for intelligent information integration within a virtual user community. Further, we have tested OntoLearn in two European projects, where it functioned as the basis for a semantic interoperability platform used by small-and medium-sized tourism enterprises. Our approach to ontology engineering uses an iterative process that involves automatic concept learning with OntoLearn, machine-supported concept validation with Consys, 2 and management with SymOntoX. 3 The engineering process starts with OntoLearn exploring available documents and related Web sites to learn domain concepts and detect taxonomic relations among them, producing as output a domain concept forest. Initially, we base concept learning on external, generic knowledge sources. In subsequent cycles, the domain ontology receives progressively more use as it becomes adequately populated. The self-learning cycle in Figure 1 shows this process. Next, we undertake ontology validation with Consys, a Web-based groupware package that performs consensus building by thoroughly validating representatives of the communities active in the application domain. Throughout the cycle, Onto-Learn operates in connection with SymOntoX. Ontology engineers can use this management system to define concepts and their mutual connections, thus allowing construction of a semantic net. Further, SymOntoX's environment can automatically attach learned domain concept trees under the appropriate nodes of the upper-domain ontology, thereby enriching concepts with additional information. Sym-OntoX also performs consistency checks. Figure 2 shows OntoLearn's system architecture, which supports a three-phase process. First, the system extracts a domain terminology from texts available in the application domain—usually drawn from specialized Web sites or documents exchanged among members of a virtual community. The system then filters this information through a natural The authors have built a software environment that supports the construction and assessment of a …","{'model': 'tldr@v2.0.0', 'text': 'A software environment is built that can build and assess a domain ontology for intelligent information integration within a virtual user community that functioned as the basis for a semantic interoperability platform used by small-and medium-sized tourism enterprises.'}",https://link.springer.com/content/pdf/bfm:978-3-031-79486-5/1?pdf=chapter%20toc
-software engineering economics,B. Boehm,"This paper summarizes the current state of the art and recent trends in software engineering economics. It provides an overview of economic analysis techniques and their applicability to software engineering and management. It surveys the field of software cost estimation, including the major estimation techniques available, the state of the art in algorithmic cost models, and the outstanding research issues in software cost estimation.","{'model': 'tldr@v2.0.0', 'text': 'The field of software cost estimation is surveyed, including the major estimation techniques available, the state of the art in algorithmic cost models, and the outstanding research issues inSoftware cost estimation.'}",http://csse.usc.edu/csse/TECHRPTS/1984/usccse84-500/usccse84-500.pdf
-a review of key challenges of electrospun scaffolds for tissue‐engineering applications,S. Khorshidi,"Tissue engineering holds great promise to develop functional constructs resembling the structural organization of native tissues to improve or replace biological functions, with the ultimate goal of avoiding organ transplantation. In tissue engineering, cells are often seeded into artificial structures capable of supporting three‐dimensional (3D) tissue formation. An optimal scaffold for tissue‐engineering applications should mimic the mechanical and functional properties of the extracellular matrix (ECM) of those tissues to be regenerated. Amongst the various scaffolding techniques, electrospinning is an outstanding one which is capable of producing non‐woven fibrous structures with dimensional constituents similar to those of ECM fibres. In recent years, electrospinning has gained widespread interest as a potential tissue‐engineering scaffolding technique and has been discussed in detail in many studies. So why this review? Apart from their clear advantages and extensive use, electrospun scaffolds encounter some practical limitations, such as scarce cell infiltration and inadequate mechanical strength for load‐bearing applications. A number of solutions have been offered by different research groups to overcome the above‐mentioned limitations. In this review, we provide an overview of the limitations of electrospinning as a tissue‐engineered scaffolding technique, with emphasis on possible resolutions of those issues. Copyright © 2015 John Wiley & Sons, Ltd.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the limitations of electrospinning as a tissue‐engineered scaffolding technique is provided, with emphasis on possible resolutions of those issues.'}",
-pathological process of prompt connection between host and donor tissue vasculature causing rapid perfusion of the engineered donor tissue after transplantation,Sachiko Sekiya,"The shortage of donors for transplantation therapy is a serious issue worldwide. Tissue engineering is considered a potential solution to this problem. Connection and perfusion in engineered tissues after transplantation is vital for the survival of the transplanted tissue, especially for tissues requiring blood perfusion to receive nutrients, such as the heart. A myocardial cell sheet containing an endothelial cell network structure was fabricated in vitro using cell sheet technology. Transplantation of the three-dimensional (3D) tissue by layering myocardial sheets could ameliorate ischemic heart disease in a rat model. The endothelial cell network in the 3D tissue was able to rapidly connect to host vasculature and begin perfusion within 24 h after transplantation. In this review, we compare and discuss the engineered tissue–host vasculature connection process between tissue engineered constructs with hydrogels and cell sheets by histological analysis. This review provides information that may be useful for further improvements of in vivo engineered tissue vascularization techniques.","{'model': 'tldr@v2.0.0', 'text': 'This review compares and discusses the engineered tissue–host vasculature connection process between tissue engineered constructs with hydrogels and cell sheets by histological analysis and provides information that may be useful for further improvements of in vivo engineered tissue vascularization techniques.'}",https://www.mdpi.com/1422-0067/19/12/4102/pdf?version=1545124295
-from word embeddings to document similarities for improved information retrieval in software engineering,Xin Ye,"The application of information retrieval techniques to search tasks in software engineering is made difficult by the lexical gap between search queries, usually expressed in natural language (e.g. English), and retrieved documents, usually expressed in code (e.g. programming languages). This is often the case in bug and feature location, community question answering, or more generally the communication between technical personnel and non-technical stake holders in a software project. In this paper, we propose bridging the lexical gap by projecting natural language statements and code snippets as meaning vectors in a shared representation space. In the proposed architecture, word embeddings are rst trained on API documents, tutorials, and reference documents, and then aggregated in order to estimate semantic similarities between documents. Empirical evaluations show that the learned vector space embeddings lead to improvements in a previously explored bug localization task and a newly de ned task of linking API documents to computer programming questions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes bridging the lexical gap by projecting natural language statements and code snippets as meaning vectors in a shared representation space and shows that the learned vector space embeddings lead to improvements in a previously explored bug localization task and a newly introduced task of linking API documents to computer programming questions.'}",
-"domain analysis and description principles, techniques, and modelling languages",D. Bjørner,"We present a method for analysing and describing domains. By a domain we shall understand a rationally describable segment of a human assisted reality, i.e., of the world, its physical parts: natural [“God-given”] and artifactual [“human-made”], and living species: plants and animals including, notably, humans. These are endurants (“still”), as well as perdurants (“alive”). Emphasis is placed on “human-assistedness,” that is, that there is at least one (human-made) artifact and, therefore, that humans are a primary cause for change of endurant states as well as perdurant behaviours. By a method we shall mean a set of principles of analysis and for selecting and applying a number of techniques and tools in the construction of some artifact, say a domain description. We shall present a method for constructing domain descriptions. Among the tools we shall only be concerned with are the analysis and synthesis languages. Domain science and engineering marks a new area of computing science. Just as we are formalising the syntax and semantics of programming languages, so we are formalising the syntax and semantics of human-assisted domains. Just as physicists are studying the natural physical world, endowing it with mathematical models, so we, computing scientists, are studying these domains, endowing them with mathematical models, A difference between the endeavours of physicists and ours lies in the tools: The physics models are based on classical mathematics, differential equations and integrals, and so on; our models are based on mathematical logic, set theory, and algebra [1]. Where physicists thus classically use a variety of differential and integral calculi to model the physical world, we shall be using the analysis and description calculi presented in this article to model primarily artifactual domains.","{'model': 'tldr@v2.0.0', 'text': 'Where physicists thus classically use a variety of differential and integral calculi to model the physical world, the analysis and description calculi presented in this article will be used to model primarily artifactual domains.'}",
-a review of three-dimensional printing in tissue engineering.,Nicholas A. Sears,"Recent advances in three-dimensional (3D) printing technologies have led to a rapid expansion of applications from the creation of anatomical training models for complex surgical procedures to the printing of tissue engineering constructs. In addition to achieving the macroscale geometry of organs and tissues, a print layer thickness as small as 20 μm allows for reproduction of the microarchitectures of bone and other tissues. Techniques with even higher precision are currently being investigated to enable reproduction of smaller tissue features such as hepatic lobules. Current research in tissue engineering focuses on the development of compatible methods (printers) and materials (bioinks) that are capable of producing biomimetic scaffolds. In this review, an overview of current 3D printing techniques used in tissue engineering is provided with an emphasis on the printing mechanism and the resultant scaffold characteristics. Current practical challenges and technical limitations are emphasized and future trends of bioprinting are discussed.","{'model': 'tldr@v2.0.0', 'text': 'An overview of current 3D printing techniques used in tissue engineering is provided with an emphasis on the printing mechanism and the resultant scaffold characteristics.'}",
-3d biofabrication strategies for tissue engineering and regenerative medicine.,P. Bajaj,"Over the past several decades, there has been an ever-increasing demand for organ transplants. However, there is a severe shortage of donor organs, and as a result of the increasing demand, the gap between supply and demand continues to widen. A potential solution to this problem is to grow or fabricate organs using biomaterial scaffolds and a person's own cells. Although the realization of this solution has been limited, the development of new biofabrication approaches has made it more realistic. This review provides an overview of natural and synthetic biomaterials that have been used for organ/tissue development. It then discusses past and current biofabrication techniques, with a brief explanation of the state of the art. Finally, the review highlights the need for combining vascularization strategies with current biofabrication techniques. Given the multitude of applications of biofabrication technologies, from organ/tissue development to drug discovery/screening to development of complex in vitro models of human diseases, these manufacturing technologies can have a significant impact on the future of medicine and health care.","{'model': 'tldr@v2.0.0', 'text': 'This review provides an overview of natural and synthetic biomaterials that have been used for organ/tissue development and highlights the need for combining vascularization strategies with current biofabrication techniques.'}",https://www.annualreviews.org/doi/pdf/10.1146/annurev-bioeng-071813-105155
-model-driven software engineering in practice,Marco Brambilla,"Model based software development differs from the conventional software development process and used in conjunction with a range of agile techniques. Despite improvements in third generation programming languages and runtime platforms, the levels of abstraction at which PLAs are developed today remains low-level relative to the concepts and concerns within the application domains themselves, such as manually tracking the library dependency or ensuring component composition syntactical and semantic correctness. [1] A promising means to address this problem involves developing PLAs using model-driven engineering (MDE) [2] which involves systematic use of models as key design and implementation artifacts throughout the software lifecycle. Thus to induce the works on architectures and models, researchers have been working to create structured models. The literature in this area hence proliferates; where the current edition [3] tries to address some of the significant aspects.","{'model': 'tldr@v2.0.0', 'text': 'To induce the works on architectures and models, researchers have been working to create structured models, and the literature in this area hence proliferates; where the current edition tries to address some of the significant aspects.'}",https://link.springer.com/content/pdf/bfm:978-3-031-02549-5/1?pdf=chapter%20toc
-a systematic literature review on fault prediction performance in software engineering,T. Hall,"Background: The accurate prediction of where faults are likely to occur in code can help direct test effort, reduce costs, and improve the quality of software. Objective: We investigate how the context of models, the independent variables used, and the modeling techniques applied influence the performance of fault prediction models. Method: We used a systematic literature review to identify 208 fault prediction studies published from January 2000 to December 2010. We synthesize the quantitative and qualitative results of 36 studies which report sufficient contextual and methodological information according to the criteria we develop and apply. Results: The models that perform well tend to be based on simple modeling techniques such as Naive Bayes or Logistic Regression. Combinations of independent variables have been used by models that perform well. Feature selection has been applied to these combinations when models are performing particularly well. Conclusion: The methodology used to build models seems to be influential to predictive performance. Although there are a set of fault prediction studies in which confidence is possible, more studies are needed that use a reliable methodology and which report their context, methodology, and performance comprehensively.","{'model': 'tldr@v2.0.0', 'text': 'Although there are a set of fault prediction studies in which confidence is possible, more studies are needed that use a reliable methodology and which report their context, methodology, and performance comprehensively.'}",https://researchrepository.ul.ie/articles/online_resource/A_Systematic_literature_review_on_fault_prediction_performance_in_software_engineering/19848790/1/files/35264761.pdf
-"mechanobiology in tendon, ligament, and skeletal muscle tissue engineering.",Michael T. K. Bramson,"Tendon, ligament, and skeletal muscle are highly-organized tissues that largely rely on a hierarchical collagenous matrix to withstand high tensile loads experienced in activities of daily life. This critical biomechanical role predisposes these tissues to injury, and current treatments fail to recapitulate the biomechanical function of native tissue. This has prompted researchers to pursue engineering functional tissue replacements, or dysfunction/disease/development models, by emulating in vivo stimuli within in vitro tissue engineering platforms; specifically mechanical stimulation, as well as active contraction in skeletal muscle. Mechanical loading is critical for matrix production and organization in the development, maturation, and maintenance of native tendon, ligament, and skeletal muscle, as well as their interfaces. Tissue engineers seek to harness these mechanobiological benefits using bioreactors to apply both static and dynamic mechanical stimulation to tissue constructs, and induce active contraction in engineered skeletal muscle. The vast majority of engineering approaches in these tissues are scaffold-based, providing interim structure and support to engineered constructs, and sufficient integrity to withstand mechanical loading. Alternatively, some recent studies have employed developmentally-inspired scaffold-free techniques, relying on cellular self-assembly and matrix production to form tissue constructs. Whether utilizing a scaffold or not, incorporation of mechanobiological stimuli has been shown to improve the composition, structure, and biomechanical function of engineered tendon, ligament, and skeletal muscle. Together, these findings highlight the importance of mechanobiology and suggest how it can be leveraged to engineer these tissues and their interfaces, and to create functional multi-tissue constructs.","{'model': 'tldr@v2.0.0', 'text': 'Together, these findings highlight the importance of mechanobiology and suggest how it can be leveraged to engineer these tissues and their interfaces, and to create functional multi-tissue constructs.'}",
-game development software engineering process life cycle: a systematic review,Saiqa Aleem,,"{'model': 'tldr@v2.0.0', 'text': 'The results of this study suggest that the game development software engineering process has many aspects that need further attention from researchers; that especially includes the postproduction phase.'}",https://jserd.springeropen.com/track/pdf/10.1186/s40411-016-0032-7
-analisis strategi menu engineering dalam upaya meningkatkan pendapatan restaurant,Ida Bagus Ananta Wijaya,"Penelitian ini dilaksanakan untuk membantu pihak manajemen dalam mengevaluasi dan menganalisis menu a’la carte menggunakan metode menu engineering. Pihak manajemen belum menerapkankan analisis menu menggunakan metode tersebut, sehingga pendapatan kurang maksimal. Penelitian ini bertujuan untuk mengetahui klasifikasi menu, dan diikuti dengan strategi untuk menindak lanjuti menu dengan klasifikasi tersebut. Penelitian ini dilaksanakan di The Sayan House Restaurant yang dilaksanakan pada bulan Maret 2021 –Juni 2021, dengan menggunakan sales history dari bulan Desember 2020 – Februari 2021. Jenis penelitian ini adalah deskriptif kualitatif dan kuantitatif. Penelitian ini menggunakan sampel jenuh sebanyak 38 item menu. Data dikumpulkan dari wawancara dan studi dokumentasi, kemudian dianalisis menggunakan teknik analisis menu engineering, sehingga mendapatkan klasifikasi menu yaitu star, plow horse, puzzle, dan dog. Hasil penelitian menunjukkan dari 38 item menu yang dianalisis, 16 diantaranya dikategorikan star, 10 diantaranya dikategorikan plow horse, 5 diantaranya dikategorikan puzzle, dan 7 diantaranya dikategorikan dog. Berdasarkan hasil analisis tersebut dan menyikapi masing-masing klasifikasi menu, diperlukan strategi dan tindak lanjut untuk setiap klasifikasinya, serta melakukan evaluasi menu secara rutin oleh pihak manajemen The Sayan House Restaurant. This research was conducted to assist management in evaluating and analyzing a'la carte menu usingthe menu engineering method. The management has not implemented menu analysis using that method, so the revenue is less than maximum. This study aims to know the classification of the menu, and followed by a strategy to follow up the menu with that classification. This research was conducted at The Sayan House Restaurant conducted in March 2021 – June 2021, using sales history from December 2020 – February 2021. This type of research is descriptive qualitative and quantitative. The study used saturated samples of 38 menu items. The data is collected from interviews and documentation studies, then analyzed using engineering menu analysis techniques, to obtain menu classifications namely star, plow horse, puzzle, and dog. The results showed from 38 menu items analyzed, 16 of them were categorized as a star, 10 of them were categorized as a plow horse, 5 of them were categorized as a puzzles, and 7 of them were categorized as a dogs. Based on the results of the analysis and addressing each menu classification, strategy and follow-up for each classification are required, as well as conducting regular menu evaluations by the management of The Sayan House Restaurant.",,https://paris.ipb-intl.ac.id/index.php/paris/article/download/464/351
-selection of software requirements elicitation techniques- a systematic review,Fazila Shams,"Article history: Received 27 November 2018 Received in revised form 2 January 2019 Accepted 9 January 2019 Available online 20 January 2019 Requirements elicitation is a basic action that frames some portion of the Requirements Engineering process. since it needs to find what the product must do through a strong comprehension of the desires and needs of the different stakeholders and then to change them into Requirements list. But, despite of its significance, only a couple of Systematic reviews are present in literature that give logical proof about the adequacy of the procedures used for Software Requirements Elicitation. This paper aims to present a systematic review of literature about the Elicitation techniques. The goal is to find out the mature techniques which are used for requirements elicitation and which techniques increases efficiency of the elicitation. Preceding literature expect that maturity prompts a better understanding of client’s requirements, and hence an improved probability that a subsequent product will fulfill those requirements. This review found 30 studies to find the answers and depicts which techniques are more effective and which techniques is suitable under which situations, considering the system to be developed and information obtained from other factors.","{'model': 'tldr@v2.0.0', 'text': 'This paper aims to present a systematic review of literature about the Elicitation techniques to find out the mature techniques which are used for requirements elicitation and which techniques increases efficiency of the elicitation.'}",
-a survey of feature location techniques,J. Rubin,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter discusses possible directions for leveraging SPLE architectures in order to improve the feature location process and describes their implementation strategies and exemplify the techniques on a realistic use-case.'}",http://www.cs.toronto.edu/~chechik/pubs/locationsurvey12.pdf
-early diagnosis of alzheimer's disease using machine learning techniques: a review paper,Aunsia Khan,"Alzheimer's, an irreparable brain disease, impairs thinking and memory while the aggregate mind size shrinks which at last prompts demise. Early diagnosis of AD is essential for the progress of more prevailing treatments. Machine learning (ML), a branch of artificial intelligence, employs a variety of probabilistic and optimization techniques that permits PCs to gain from vast and complex datasets. As a result, researchers focus on using machine learning frequently for diagnosis of early stages of AD. This paper presents a review, analysis and critical evaluation of the recent work done for the early detection of AD using ML techniques. Several methods achieved promising prediction accuracies, however they were evaluated on different pathologically unproven data sets from different imaging modalities making it difficult to make a fair comparison among them. Moreover, many other factors such as pre-processing, the number of important attributes for feature selection, class imbalance distinctively affect the assessment of the prediction accuracy. To overcome these limitations, a model is proposed which comprise of initial pre-processing step followed by imperative attributes selection and classification is achieved using association rule mining. Furthermore, this proposed model based approach gives the right direction for research in early diagnosis of AD and has the potential to distinguish AD from healthy controls.","{'model': 'tldr@v2.0.0', 'text': 'A model based approach is proposed which comprise of initial pre-processing step followed by imperative attributes selection and classification is achieved using association rule mining and has the potential to distinguish AD from healthy controls.'}",
-"model-driven software engineering in practice, second edition",M. Brambilla,,"{'model': 'tldr@v2.0.0', 'text': 'This book is to provide an agile and flexible tool to introduce you to the MDSE world, thus allowing you to quickly understand its basic principles and techniques and to choose the right set of MDSE instruments for your needs so that you can start to benefit from MDSE right away.'}",https://link.springer.com/content/pdf/bfm:978-3-031-02549-5/1?pdf=chapter%20toc
-the design of scaffolds for use in tissue engineering. part ii. rapid prototyping techniques.,Shoufeng Yang,"Tissue engineering (TE) is an important emerging area in biomedical engineering for creating biological alternatives for harvested tissues, implants, and prostheses. In TE, a highly porous artificial extracellular matrix or scaffold is required to accommodate mammalian cells and guide their growth and tissue regeneration in three-dimension (3D). However, existing 3D scaffolds for TE proved less than ideal for actual applications because they lack mechanical strength, interconnected channels, and controlled porosity or pores distribution. In this paper, the authors review the application and advancement of rapid prototyping (RP) techniques in the design and creation of synthetic scaffolds for use in TE. We also review the advantages and benefits, and limitations and shortcomings of current RP techniques as well as the future direction of RP development in TE scaffold fabrication.","{'model': 'tldr@v2.0.0', 'text': 'The authors review the application and advancement of rapid prototyping (RP) techniques in the design and creation of synthetic scaffolds for use in TE and the advantages and benefits, and limitations and shortcomings of current RP techniques as well as the future direction of RP development in TE scaffold fabrication.'}",https://lirias.kuleuven.be/bitstream/123456789/556008/1/The%20design%20of%20scaffolds%20for%20use%20in%20tissue%20engineering.%20Part%20II.%20Rapid%20prototyping%20techniques.pdf
-cell surface engineering and application in cell delivery to heart diseases,Daniel Y. Lee,,"{'model': 'tldr@v2.0.0', 'text': 'Current methods of non-genetic cell surface modification, including chemical conjugations, polymeric encapsulation, hydrophobic insertion, enzymatic and metabolic addition, will be introduced.'}",https://jbioleng.biomedcentral.com/track/pdf/10.1186/s13036-018-0123-6
-"machine learning, medical diagnosis, and biomedical engineering research - commentary",K. Foster,,"{'model': 'tldr@v2.0.0', 'text': 'This commentary is intended to help sensitize investigators as well as readers and reviewers of papers to some potential pitfalls in the development of classifiers, and suggests steps that researchers can take to help avoid these problems.'}",https://biomedical-engineering-online.biomedcentral.com/counter/pdf/10.1186/1475-925X-13-94
-state-of-the-art review of 3d bioprinting for cardiovascular tissue engineering,B. Duan,,"{'model': 'tldr@v2.0.0', 'text': 'Current challenges and further perspectives are discussed to help guide the bioink and bioprinter development, improve bioprinting strategies and direct future organ biopprinting and translational applications.'}",
-whole-organ tissue engineering: decellularization and recellularization of three-dimensional matrix scaffolds.,S. Badylak,"The definitive treatment for end-stage organ failure is orthotopic transplantation. However, the demand for transplantation far exceeds the number of available donor organs. A promising tissue-engineering/regenerative-medicine approach for functional organ replacement has emerged in recent years. Decellularization of donor organs such as heart, liver, and lung can provide an acellular, naturally occurring three-dimensional biologic scaffold material that can then be seeded with selected cell populations. Preliminary studies in animal models have provided encouraging results for the proof of concept. However, significant challenges for three-dimensional organ engineering approach remain. This manuscript describes the fundamental concepts of whole-organ engineering, including characterization of the extracellular matrix as a scaffold, methods for decellularization of vascular organs, potential cells to reseed such a scaffold, techniques for the recellularization process and important aspects regarding bioreactor design to support this approach. Critical challenges and future directions are also discussed.","{'model': 'tldr@v2.0.0', 'text': 'The fundamental concepts of whole-organ engineering, including characterization of the extracellular matrix as a scaffold, methods for decellularization of vascular organs, potential cells to reseed such a scaffolds, techniques for the recellularization process and important aspects regarding bioreactor design to support this approach are described.'}",
-comparison and evaluation of goal-oriented satisfaction analysis techniques,Jennifer Horkoff,,"{'model': 'tldr@v2.0.0', 'text': 'This comparison shows that different satisfaction analysis techniques for goal models can produce variable results, depending on the structure of the model, and recommends the use of satisfaction analysis technique as only heuristics for decision making.'}",
-metamodels for computer-based engineering design: survey and recommendations,T. Simpson,,"{'model': 'tldr@v2.0.0', 'text': 'This paper surveys their existing application in engineering design, and addresses the dangers of applying traditional statistical techniques to approximate deterministic computer analysis codes, along with recommendations for the appropriate use of statistical approximation techniques in given situations.'}",https://ntrs.nasa.gov/api/citations/19990087092/downloads/19990087092.pdf
-towards modelling and reasoning support for early-phase requirements engineering,E. Yu,"Requirements are usually understood as stating what a system is supposed to do, as apposed to how it should do it. However, understanding the organizational context and rationales (the ""Whys"") that lead up to systems requirements can be just as important for the ongoing success of the system. Requirements modelling techniques can be used to help deal with the knowledge and reasoning needed in this earlier phase of requirements engineering. However most existing requirements techniques are intended more for the later phase of requirements engineering, which focuses on completeness, consistency, and automated verification of requirements. In contrast, the early phase aims to model and analyze stakeholder interests and how they might be addressed, or compromised, by various system-and-environment alternatives. This paper argues, therefore, that a different kind of modelling and reasoning support is needed for the early phase. An outline of the i* framework is given as an example of a step in this direction. Meeting scheduling is used as a domain example.","{'model': 'tldr@v2.0.0', 'text': 'This paper argues that a different kind of modelling and reasoning support is needed for the early phase of requirements engineering, which aims to model and analyze stakeholder interests and how they might be addressed, or compromised, by various system-and-environment alternatives.'}",
-simple llm prompting is state-of-the-art for robust and multilingual dialogue evaluation,J. Mendoncca,"Despite significant research effort in the development of automatic dialogue evaluation metrics, little thought is given to evaluating dialogues other than in English. At the same time, ensuring metrics are invariant to semantically similar responses is also an overlooked topic. In order to achieve the desired properties of robustness and multilinguality for dialogue evaluation metrics, we propose a novel framework that takes advantage of the strengths of current evaluation models with the newly-established paradigm of prompting Large Language Models (LLMs). Empirical results show our framework achieves state of the art results in terms of mean Spearman correlation scores across several benchmarks and ranks first place on both the Robust and Multilingual tasks of the DSTC11 Track 4 “Automatic Evaluation Metrics for Open-Domain Dialogue Systems”, proving the evaluation capabilities of prompted LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework that takes advantage of the strengths of current evaluation models with the newly-established paradigm of prompting Large Language Models (LLMs) to achieve the desired properties of robustness and multilinguality for dialogue evaluation metrics.'}",https://arxiv.org/pdf/2308.16797
-towards understanding chain-of-thought prompting: an empirical study of what matters,Boshi Wang,"Chain-of-Thought (CoT) prompting can dramatically improve the multi-step reasoning abilities of large language models (LLMs). CoT explicitly encourages the LLM to generate intermediate rationales for solving a problem, by providing a series of reasoning steps in the demonstrations. Despite its success, there is still little understanding of what makes CoT prompting effective and which aspects of the demonstrated reasoning steps contribute to its performance. In this paper, we show that CoT reasoning is possible even with invalid demonstrations - prompting with invalid reasoning steps can achieve over 80-90% of the performance obtained using CoT under various metrics, while still generating coherent lines of reasoning during inference. Further experiments show that other aspects of the rationales, such as being relevant to the query and correctly ordering the reasoning steps, are much more important for effective CoT reasoning. Overall, these findings both deepen our understanding of CoT prompting, and open up new questions regarding LLMs’ capability to learn to reason in context.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that CoT reasoning is possible even with invalid demonstrations - prompting with invalid reasoning steps can achieve over 80-90% of the performance obtained using CoT under various metrics, while still generating coherent lines of reasoning during inference.'}",http://arxiv.org/pdf/2212.10001
-improving language model prompting in support of semi-autonomous task learning,James R. Kirk,"Language models (LLMs) offer potential as a source of knowledge for agents that need to acquire new task competencies within a performance environment. We describe efforts toward a novel agent capability that can construct cues (or""prompts"") that result in useful LLM responses for an agent learning a new task. Importantly, responses must not only be""reasonable""(a measure used commonly in research on knowledge extraction from LLMs) but also specific to the agent's task context and in a form that the agent can interpret given its native language capacities. We summarize a series of empirical investigations of prompting strategies and evaluate responses against the goals of targeted and actionable responses for task learning. Our results demonstrate that actionable task knowledge can be obtained from LLMs in support of online agent task learning.","{'model': 'tldr@v2.0.0', 'text': 'This work describes efforts toward a novel agent capability that can construct cues that result in useful LLM responses for an agent learning a new task.'}",https://arxiv.org/pdf/2209.07636
-chateval: towards better llm-based evaluators through multi-agent debate,Chi-Min Chan,"Text evaluation has historically posed significant challenges, often demanding substantial labor and time cost. With the emergence of large language models (LLMs), researchers have explored LLMs' potential as alternatives for human evaluation. While these single-agent-based approaches show promise, experimental results suggest that further advancements are needed to bridge the gap between their current effectiveness and human-level evaluation quality. Recognizing that best practices of human evaluation processes often involve multiple human annotators collaborating in the evaluation, we resort to a multi-agent debate framework, moving beyond single-agent prompting strategies. The multi-agent-based approach enables a group of LLMs to synergize with an array of intelligent counterparts, harnessing their distinct capabilities and expertise to enhance efficiency and effectiveness in handling intricate tasks. In this paper, we construct a multi-agent referee team called ChatEval to autonomously discuss and evaluate the quality of generated responses from different models on open-ended questions and traditional natural language generation (NLG) tasks. Our analysis shows that ChatEval transcends mere textual scoring, offering a human-mimicking evaluation process for reliable assessments. Our code is available at https://github.com/chanchimin/ChatEval.","{'model': 'tldr@v2.0.0', 'text': 'A multi-agent referee team called ChatEval is constructed to autonomously discuss and evaluate the quality of generated responses from different models on open-ended questions and traditional natural language generation (NLG) tasks, offering a human-mimicking evaluation process for reliable assessments.'}",https://arxiv.org/pdf/2308.07201
-boosting theory-of-mind performance in large language models via prompting,Shima Rahimi Moghaddam,"Large language models (LLMs) excel in many tasks in 2023, but they still face challenges in complex reasoning. Theory-of-mind (ToM) tasks, which require understanding agents' beliefs, goals, and mental states, are essential for common-sense reasoning involving humans, making it crucial to enhance LLM performance in this area. This study measures the ToM performance of GPT-4 and three GPT-3.5 variants (Davinci-2, Davinci-3, GPT-3.5-Turbo), and investigates the effectiveness of in-context learning in improving their ToM comprehension. We evaluated prompts featuring two-shot chain of thought reasoning and step-by-step thinking instructions. We found that LLMs trained with Reinforcement Learning from Human Feedback (RLHF) (all models excluding Davinci-2) improved their ToM accuracy via in-context learning. GPT-4 performed best in zero-shot settings, reaching nearly 80% ToM accuracy, but still fell short of the 87% human accuracy on the test set. However, when supplied with prompts for in-context learning, all RLHF-trained LLMs exceeded 80% ToM accuracy, with GPT-4 reaching 100%. These results demonstrate that appropriate prompting enhances LLM ToM reasoning, and they underscore the context-dependent nature of LLM cognitive capacities.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that appropriate prompting enhances LLM ToM reasoning, and the context-dependent nature of LLM cognitive capacities is underscored, as it is found that LLMs trained with Reinforcement Learning from Human Feedback improved their ToM accuracy via in-context learning.'}",http://arxiv.org/pdf/2304.11490
-3d-llm: injecting the 3d world into large language models,Yining Hong,"Large language models (LLMs) and Vision-Language Models (VLMs) have been proven to excel at multiple tasks, such as commonsense reasoning. Powerful as these models can be, they are not grounded in the 3D physical world, which involves richer concepts such as spatial relationships, affordances, physics, layout, and so on. In this work, we propose to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs. Specifically, 3D-LLMs can take 3D point clouds and their features as input and perform a diverse set of 3D-related tasks, including captioning, dense captioning, 3D question answering, task decomposition, 3D grounding, 3D-assisted dialog, navigation, and so on. Using three types of prompting mechanisms that we design, we are able to collect over 300k 3D-language data covering these tasks. To efficiently train 3D-LLMs, we first utilize a 3D feature extractor that obtains 3D features from rendered multi-view images. Then, we use 2D VLMs as our backbones to train our 3D-LLMs. By introducing a 3D localization mechanism, 3D-LLMs can better capture 3D spatial information. Experiments on ScanQA show that our model outperforms state-of-the-art baselines by a large margin ( e.g. , the BLEU-1 score surpasses state-of-the-art score by 9%). Furthermore, experiments on our held-in datasets for 3D captioning, task composition, and 3D-assisted dialogue show that our model outperforms 2D VLMs. Qualitative examples also show that our model could perform more tasks beyond the scope of existing LLMs and VLMs. Project Page: : https:","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs, and introduces a 3D localization mechanism, which can better capture 3D spatial information.'}",https://arxiv.org/pdf/2307.12981
-smart-llm: smart multi-agent robot task planning using large language models,S. S. Kannan,"In this work, we introduce SMART-LLM, an innovative framework designed for embodied multi-robot task planning. SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models (LLMs), harnesses the power of LLMs to convert high-level task instructions provided as input into a multi-robot task plan. It accomplishes this by executing a series of stages, including task decomposition, coalition formation, and task allocation, all guided by programmatic LLM prompts within the few-shot prompting paradigm. We create a benchmark dataset designed for validating the multi-robot task planning problem, encompassing four distinct categories of high-level instructions that vary in task complexity. Our evaluation experiments span both simulation and real-world scenarios, demonstrating that the proposed model can achieve promising results for generating multi-robot task plans. The experimental videos, code, and datasets from the work can be found at https://sites.google.com/view/smart-llm/.","{'model': 'tldr@v2.0.0', 'text': 'SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models (LLMs), harnesses the power of LLMs to convert high-level task instructions provided as input into a multi-robot task plan by executing a series of stages.'}",https://arxiv.org/pdf/2309.10062
-llm-deliberation: evaluating llms with interactive multi-agent negotiation games,Sahar Abdelnabi,"There is a growing interest in using Large Language Models (LLMs) as agents to tackle real-world tasks that may require assessing complex situations. Yet, we have a limited understanding of LLMs' reasoning and decision-making capabilities, partly stemming from a lack of dedicated evaluation benchmarks. As negotiating and compromising are key aspects of our everyday communication and collaboration, we propose using scorable negotiation games as a new evaluation framework for LLMs. We create a testbed of diverse text-based, multi-agent, multi-issue, semantically rich negotiation games, with easily tunable difficulty. To solve the challenge, agents need to have strong arithmetic, inference, exploration, and planning capabilities, while seamlessly integrating them. Via a systematic zero-shot Chain-of-Thought prompting (CoT), we show that agents can negotiate and consistently reach successful deals. We quantify the performance with multiple metrics and observe a large gap between GPT-4 and earlier models. Importantly, we test the generalization to new games and setups. Finally, we show that these games can help evaluate other critical aspects, such as the interaction dynamics between agents in the presence of greedy and adversarial players.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using scorable negotiation games as a new evaluation framework for LLMs, and shows that these games can help evaluate other critical aspects, such as the interaction dynamics between agents in the presence of greedy and adversarial players.'}",https://arxiv.org/pdf/2309.17234
-chainpoll: a high efficacy method for llm hallucination detection,Robert Friel,"Large language models (LLMs) have experienced notable advancements in generating coherent and contextually relevant responses. However, hallucinations - incorrect or unfounded claims - are still prevalent, prompting the creation of automated metrics to detect these in LLM outputs. Our contributions include: introducing ChainPoll, an innovative hallucination detection method that excels compared to its counterparts, and unveiling RealHall, a refined collection of benchmark datasets to assess hallucination detection metrics from recent studies. While creating RealHall, we assessed tasks and datasets from previous hallucination detection studies and observed that many are not suitable for the potent LLMs currently in use. Overcoming this, we opted for four datasets challenging for modern LLMs and pertinent to real-world scenarios. Using RealHall, we conducted a comprehensive comparison of ChainPoll with numerous hallucination metrics from recent studies. Our findings indicate that ChainPoll outperforms in all RealHall benchmarks, achieving an overall AUROC of 0.781. This surpasses the next best theoretical method by 11% and exceeds industry standards by over 23%. Additionally, ChainPoll is cost-effective and offers greater transparency than other metrics. We introduce two novel metrics to assess LLM hallucinations: Adherence and Correctness. Adherence is relevant to Retrieval Augmented Generation workflows, evaluating an LLM's analytical capabilities within given documents and contexts. In contrast, Correctness identifies logical and reasoning errors.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces two novel metrics to assess LLM hallucinations: Adherence and Correctness, and unveils RealHall, a refined collection of benchmark datasets to assess hallucination detection metrics from recent studies.'}",
-put your money where your mouth is: evaluating strategic planning and execution of llm agents in an auction arena,Jiangjie Chen,"Can Large Language Models (LLMs) simulate human behavior in complex environments? LLMs have recently been shown to exhibit advanced reasoning skills but much of NLP evaluation still relies on static benchmarks. Answering this requires evaluation environments that probe strategic reasoning in competitive, dynamic scenarios that involve long-term planning. We introduce AucArena, a novel simulation environment for evaluating LLMs within auctions, a setting chosen for being highly unpredictable and involving many skills related to resource and risk management, while also being easy to evaluate. We conduct several controlled simulations using state-of-the-art LLMs as bidding agents. We find that through simple prompting, LLMs do indeed demonstrate many of the skills needed for effectively engaging in auctions (e.g., managing budget, adhering to long-term goals and priorities), skills that we find can be sharpened by explicitly encouraging models to be adaptive and observe strategies in past auctions. These results are significant as they show the potential of using LLM agents to model intricate social dynamics, especially in competitive settings. However, we also observe considerable variability in the capabilities of individual LLMs. Notably, even our most advanced models (GPT-4) are occasionally surpassed by heuristic baselines and human agents, highlighting the potential for further improvements in the design of LLM agents and the important role that our simulation environment can play in further testing and refining agent architectures.","{'model': 'tldr@v2.0.0', 'text': 'AucArena, a novel simulation environment for evaluating LLMs within auctions, is introduced, finding that through simple prompting, LLMs do indeed demonstrate many of the skills needed for effectively engaging in auctions and can be sharpened by explicitly encouraging models to be adaptive and observe strategies in past auctions.'}",https://arxiv.org/pdf/2310.05746
-prompting large language models with speech recognition abilities,Yassir Fathullah,"Large language models have proven themselves highly flexible, able to solve a wide range of generative tasks, such as abstractive summarization and open-ended question answering. In this paper we extend the capabilities of LLMs by directly attaching a small audio encoder allowing it to perform speech recognition. By directly prepending a sequence of audial embeddings to the text token embeddings, the LLM can be converted to an automatic speech recognition (ASR) system, and be used in the exact same manner as its textual counterpart. Experiments on Multilingual LibriSpeech (MLS) show that incorporating a conformer encoder into the open sourced LLaMA-7B allows it to outperform monolingual baselines by 18% and perform multilingual speech recognition despite LLaMA being trained overwhelmingly on English text. Furthermore, we perform ablation studies to investigate whether the LLM can be completely frozen during training to maintain its original capabilities, scaling up the audio encoder, and increasing the audio encoder striding to generate fewer embeddings. The results from these studies show that multilingual ASR is possible even when the LLM is frozen or when strides of almost 1 second are used in the audio encoder opening up the possibility for LLMs to operate on long-form audio.","{'model': 'tldr@v2.0.0', 'text': 'The capabilities of LLMs are extended by directly attaching a small audio encoder allowing it to perform speech recognition and it is shown that multilingual ASR is possible even when the LLM is frozen or when strides of almost 1 second are used in theaudio encoder opening up the possibility for LLMs to operate on long-form audio.'}",https://arxiv.org/pdf/2307.11795
-"see, think, confirm: interactive prompting between vision and language models for knowledge-based visual reasoning",Zhenfang Chen,"Large pre-trained vision and language models have demonstrated remarkable capacities for various tasks. However, solving the knowledge-based visual reasoning tasks remains challenging, which requires a model to comprehensively understand image content, connect the external world knowledge, and perform step-by-step reasoning to answer the questions correctly. To this end, we propose a novel framework named Interactive Prompting Visual Reasoner (IPVR) for few-shot knowledge-based visual reasoning. IPVR contains three stages, see, think and confirm. The see stage scans the image and grounds the visual concept candidates with a visual perception model. The think stage adopts a pre-trained large language model (LLM) to attend to the key concepts from candidates adaptively. It then transforms them into text context for prompting with a visual captioning model and adopts the LLM to generate the answer. The confirm stage further uses the LLM to generate the supporting rationale to the answer, verify the generated rationale with a cross-modality classifier and ensure that the rationale can infer the predicted output consistently. We conduct experiments on a range of knowledge-based visual reasoning datasets. We found our IPVR enjoys several benefits, 1). it achieves better performance than the previous few-shot learning baselines; 2). it enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step; 3). it is computation-efficient compared with other fine-tuning baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework named Interactive Prompting Visual Reasoner (IPVR), which achieves better performance than the previous few-shot learning baselines, and enjoys the total transparency and trustworthiness of the whole reasoning process by providing rationales for each reasoning step.'}",http://arxiv.org/pdf/2301.05226
-satisfiability-aided language models using declarative prompting,Xi Ye,"Prior work has combined chain-of-thought prompting in large language models (LLMs) with programmatic representations to perform effective and transparent reasoning. While such an approach works well for tasks that only require forward reasoning (e.g., straightforward arithmetic), it is less effective for constraint solving problems that require more sophisticated planning and search. In this paper, we propose a new satisfiability-aided language modeling (SatLM) approach for improving the reasoning capabilities of LLMs. We use an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer. This approach has two key advantages. The declarative specification is closer to the problem description than the reasoning steps are, so the LLM can parse it out of the description more accurately. Furthermore, by offloading the actual reasoning task to an automated theorem prover, our approach can guarantee the correctness of the answer with respect to the parsed specification and avoid planning errors in the solving process. We evaluate SATLM on 8 different datasets and show that it consistently outperforms program-aided LMs in the imperative paradigm. In particular, SATLM outperforms program-aided LMs by 23% on a challenging subset of the GSM arithmetic reasoning dataset; SATLM also achieves a new SoTA on LSAT and BoardgameQA, surpassing previous models that are trained on the respective training sets.","{'model': 'tldr@v2.0.0', 'text': 'A new satisfiability-aided language modeling (SatLM) approach for improving the reasoning capabilities of LLMs that uses an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer.'}",https://arxiv.org/pdf/2305.09656
-"prompting and evaluating large language models for proactive dialogues: clarification, target-guided, and non-collaboration",Yang Deng,"Conversational systems based on Large Language Models (LLMs), such as ChatGPT, show exceptional proficiency in context understanding and response generation. However, despite their impressive capabilities, they still possess limitations, such as providing randomly-guessed answers to ambiguous queries or failing to refuse users' requests, both of which are considered aspects of a conversational agent's proactivity. This raises the question of whether LLM-based conversational systems are equipped to handle proactive dialogue problems. In this work, we conduct a comprehensive analysis of LLM-based conversational systems, specifically focusing on three aspects of proactive dialogue systems: clarification, target-guided, and non-collaborative dialogues. To trigger the proactivity of LLMs, we propose the Proactive Chain-of-Thought prompting scheme, which augments LLMs with the goal planning capability over descriptive reasoning chains. Empirical findings are discussed to promote future studies on LLM-based proactive dialogue systems.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive analysis of LLM-based conversational systems, specifically focusing on three aspects of proactive dialogue systems: clarification, target-guided, and non-collaborative dialogues, and the Proactive Chain-of-Thought prompting scheme is proposed.'}",http://arxiv.org/pdf/2305.13626
-query expansion by prompting large language models,R. Jagerman,"Query expansion is a widely used technique to improve the recall of search systems. In this paper, we propose an approach to query expansion that leverages the generative abilities of Large Language Models (LLMs). Unlike traditional query expansion approaches such as Pseudo-Relevance Feedback (PRF) that relies on retrieving a good set of pseudo-relevant documents to expand queries, we rely on the generative and creative abilities of an LLM and leverage the knowledge inherent in the model. We study a variety of different prompts, including zero-shot, few-shot and Chain-of-Thought (CoT). We find that CoT prompts are especially useful for query expansion as these prompts instruct the model to break queries down step-by-step and can provide a large number of terms related to the original query. Experimental results on MS-MARCO and BEIR demonstrate that query expansions generated by LLMs can be more powerful than traditional query expansion methods.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on MS-MARCO and BEIR demonstrate that query expansions generated by LLMs can be more powerful than traditional query expansion methods.'}",http://arxiv.org/pdf/2305.03653
-pive: prompting with iterative verification improving graph-based generative capability of llms,Jiuzhou Han,"Large language models (LLMs) have shown great abilities of solving various natural language tasks in different domains. Due to the training objective of LLMs and their pretraining data, LLMs are not very well equipped for tasks involving structured data generation. We propose a framework, Prompting with Iterative Verification (PiVe), to improve graphbased generative capability of LLMs. We show how a small language model could be trained to act as a verifier module for the output of an LLM (i.e., ChatGPT), and to iteratively improve its performance via fine-grained corrective instructions. Additionally, we show how the verifier module could apply iterative corrections offline for a more cost-effective solution to the text-to-graph generation task. Experiments on three graph-based datasets show consistent improvement gained via PiVe. Additionally, we highlight how the proposed verifier module can be used as a data augmentation tool to help improve the quality of automatically generated parallel text-graph datasets. Our code and data are available at https://github.com/Jiuzhouh/PiVe.","{'model': 'tldr@v2.0.0', 'text': 'A framework, Prompting with Iterative Verification (PiVe), to improve graphbased generative capability of LLMs and shows how a small language model could be trained to act as a verifier module for the output of an LLM, and to iteratively improve its performance via fine-grained corrective instructions.'}",http://arxiv.org/pdf/2305.12392
-enhancing small medical learners with privacy-preserving contextual prompting,Xinlu Zhang,"Large language models (LLMs) demonstrate remarkable medical expertise, but data privacy concerns impede their direct use in healthcare environments. Although offering improved data privacy protection, domain-specific small language models (SLMs) often underperform LLMs, emphasizing the need for methods that reduce this performance gap while alleviating privacy concerns. In this paper, we present a simple yet effective method that harnesses LLMs' medical proficiency to boost SLM performance in medical tasks under privacy-restricted scenarios. Specifically, we mitigate patient privacy issues by extracting keywords from medical data and prompting the LLM to generate a medical knowledge-intensive context by simulating clinicians' thought processes. This context serves as additional input for SLMs, augmenting their decision-making capabilities. Our method significantly enhances performance in both few-shot and full training settings across three medical knowledge-intensive tasks, achieving up to a 22.57% increase in absolute accuracy compared to SLM fine-tuning without context, and sets new state-of-the-art results in two medical tasks within privacy-restricted scenarios. Further out-of-domain testing and experiments in two general domain datasets showcase its generalizability and broad applicability.","{'model': 'tldr@v2.0.0', 'text': ""This paper mitigate patient privacy issues by extracting keywords from medical data and prompting the LLM to generate a medical knowledge-intensive context by simulating clinicians' thought processes, which serves as additional input for SLMs, augmenting their decision-making capabilities.""}",http://arxiv.org/pdf/2305.12723
-grammar prompting for domain-specific language generation with large language models,Bailin Wang,"Large language models (LLMs) can learn to perform a wide range of natural language tasks from just a handful of in-context examples. However, for generating strings from highly structured languages (e.g., semantic parsing to complex domain-specific languages), it is challenging for the LLM to generalize from just a few exemplars. We explore $\textbf{grammar prompting}$ as a simple approach for enabling LLMs to use external knowledge and domain-specific constraints, expressed through a grammar expressed in Backus--Naur Form (BNF), during in-context learning. Grammar prompting augments each demonstration example with a specialized grammar that is minimally sufficient for generating the particular output example, where the specialized grammar is a subset of the full DSL grammar. For inference, the LLM first predicts a BNF grammar given a test input, and then generates the output according to the rules of the grammar. Experiments demonstrate that grammar prompting can enable LLMs to perform competitively on a diverse set of DSL generation tasks, including semantic parsing (SMCalFlow, Overnight, GeoQuery), PDDL planning, and even molecule generation (SMILES).","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that grammar prompting can enable LLMs to perform competitively on a diverse set of DSL generation tasks, including semantic parsing, Overnight, GeoQuery, PDDL planning, and even molecule generation (SMILES).'}",http://arxiv.org/pdf/2305.19234
-allies: prompting large language model with beam search,Hao-Lun Sun,"With the advance of large language models (LLMs), the research field of LLM applications becomes more and more popular and the idea of constructing pipelines to accomplish complex tasks by stacking LLM API calls come true. However, this kind of methods face two limitations: narrow information coverage and low fault tolerance. In this work, we propose a novel method called ALLIES. Given an input query, ALLIES leverages LLMs to iteratively generate new queries related to the original query, enabling an iterative reasoning process. By iteratively refining and expanding the scope of the original query, ALLIES captures and utilizes hidden knowledge that may not be directly obtainable through retrieval. We take zero-shot open-domain question answering (ODQA) as an application scene and evaluate ALLIES on the widely-used benchmarks, such as NQ, WebQ and TriviaQA. The experimental results demonstrate that ALLIES significantly outperforms other zero-shot baselines, indicating its effectiveness in tackling those challenges. Our code is available in https://github.com/microsoft/SimXNS/tree/main/ALLIES.","{'model': 'tldr@v2.0.0', 'text': 'Allies leverages LLMs to iteratively generate new queries related to the original query, enabling an iterative reasoning process and captures and utilizes hidden knowledge that may not be directly obtainable through retrieval.'}",
-prompting language-informed distribution for compositional zero-shot learning,Wentao Bao,"Compositional zero-shot learning (CZSL) task aims to recognize unseen compositional visual concepts, e.g., sliced tomatoes, where the model is learned only from the seen compositions, e.g., sliced potatoes and red tomatoes. Thanks to the prompt tuning on large pre-trained visual language models such as CLIP, recent literature shows impressively better CZSL performance than traditional vision-based methods. However, the key aspects that impact the generalization to unseen compositions, including the diversity and informativeness of class context, and the entanglement between visual primitives, i.e., state and object, are not properly addressed in existing CLIP-based CZSL literature. In this paper, we propose a model by prompting the language-informed distribution, aka., PLID, for the CZSL task. Specifically, the PLID leverages pre-trained large language models (LLM) to 1) formulate the language-informed class distributions which are diverse and informative, and 2) enhance the compositionality of the class embedding. Moreover, a visual-language primitive decomposition (VLPD) module and a stochastic logit mixup (SLM) strategy are proposed to dynamically fuse the decisions from the compositional and the primitive logit space. Orthogonal to the existing literature of soft, hard, or distributional prompts, our method advocates prompting the LLM-supported class distribution that leads to a better zero-shot generalization. Experimental results on MIT-States, UT-Zappos, and C-GQA datasets show the superior performance of the PLID to the prior arts.","{'model': 'tldr@v2.0.0', 'text': 'A model by prompting the language-informed distribution, aka.'}",https://arxiv.org/pdf/2305.14428
-retrieval-augmented gpt-3.5-based text-to-sql framework with sample-aware prompting and dynamic revision chain,Chunxi Guo,"Text-to-SQL aims at generating SQL queries for the given natural language questions and thus helping users to query databases. Prompt learning with large language models (LLMs) has emerged as a recent approach, which designs prompts to lead LLMs to understand the input question and generate the corresponding SQL. However, it faces challenges with strict SQL syntax requirements. Existing work prompts the LLMs with a list of demonstration examples (i.e. question-SQL pairs) to generate SQL, but the fixed prompts can hardly handle the scenario where the semantic gap between the retrieved demonstration and the input question is large. In this paper, we propose a retrieval-augmented prompting method for a LLM-based Text-to-SQL framework, involving sample-aware prompting and a dynamic revision chain. Our approach incorporates sample-aware demonstrations, which include the composition of SQL operators and fine-grained information related to the given question. To retrieve questions sharing similar intents with input questions, we propose two strategies for assisting retrieval. Firstly, we leverage LLMs to simplify the original questions, unifying the syntax and thereby clarifying the users' intentions. To generate executable and accurate SQLs without human intervention, we design a dynamic revision chain which iteratively adapts fine-grained feedback from the previously generated SQL. Experimental results on three Text-to-SQL benchmarks demonstrate the superiority of our method over strong baseline models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a retrieval-augmented prompting method for a LLM-based Text-to-SQL framework, involving sample-aware prompting and a dynamic revision chain which iteratively adapts fine-grained feedback from the previously generated SQL.'}",https://arxiv.org/pdf/2307.05074
-towards better chain-of-thought prompting strategies: a survey,Zihan Yu,"Chain-of-Thought (CoT), a step-wise and coherent reasoning chain, shows its impressive strength when used as a prompting strategy for large language models (LLM). Recent years, the prominent effect of CoT prompting has attracted emerging research. However, there still lacks of a systematic summary about key factors of CoT prompting and comprehensive guide for prompts utilizing. For a deeper understanding about CoT prompting, we survey on a wide range of current research, presenting a systematic and comprehensive analysis on several factors that may influence the effect of CoT prompting, and introduce how to better apply it in different applications under these discussions. We further analyze the challenges and propose some future directions about CoT prompting. This survey could provide an overall reference on related research.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and comprehensive analysis on several factors that may influence the effect of CoT prompting are presented, and how to better apply it in different applications under these discussions are introduced.'}",https://arxiv.org/pdf/2310.04959
-"reinforcement learning in the era of llms: what is essential? what is needed? an rl perspective on rlhf, prompting, and beyond",Hao Sun,"Recent advancements in Large Language Models (LLMs) have garnered wide attention and led to successful products such as ChatGPT and GPT-4. Their proficiency in adhering to instructions and delivering harmless, helpful, and honest (3H) responses can largely be attributed to the technique of Reinforcement Learning from Human Feedback (RLHF). In this paper, we aim to link the research in conventional RL to RL techniques used in LLM research. Demystify this technique by discussing why, when, and how RL excels. Furthermore, we explore potential future avenues that could either benefit from or contribute to RLHF research. Highlighted Takeaways: 1. RLHF is Online Inverse RL with Offline Demonstration Data. 2. RLHF $>$ SFT because Imitation Learning (and Inverse RL) $>$ Behavior Cloning (BC) by alleviating the problem of compounding error. 3. The RM step in RLHF generates a proxy of the expensive human feedback, such an insight can be generalized to other LLM tasks such as prompting evaluation and optimization where feedback is also expensive. 4. The policy learning in RLHF is more challenging than conventional problems studied in IRL due to their high action dimensionality and feedback sparsity. 5. The main superiority of PPO over off-policy value-based methods is its stability gained from (almost) on-policy data and conservative policy updates.",,https://arxiv.org/pdf/2310.06147
-can instruction fine-tuned language models identify social bias through prompting?,O. Dige,"As the breadth and depth of language model applications continue to expand rapidly, it is increasingly important to build efficient frameworks for measuring and mitigating the learned or inherited social biases of these models. In this paper, we present our work on evaluating instruction fine-tuned language models' ability to identify bias through zero-shot prompting, including Chain-of-Thought (CoT) prompts. Across LLaMA and its two instruction fine-tuned versions, Alpaca 7B performs best on the bias identification task with an accuracy of 56.7%. We also demonstrate that scaling up LLM size and data diversity could lead to further performance gain. This is a work-in-progress presenting the first component of our bias mitigation framework. We will keep updating this work as we get more results.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating instruction fine-tuned language models' ability to identify bias through zero-shot prompting, including Chain-of-Thought (CoT) prompts shows Alpaca 7B performs best on the bias identification task and it is demonstrated that scaling up LLM size and data diversity could lead to further performance gain.""}",https://arxiv.org/pdf/2307.10472
-approximating online human evaluation of social chatbots with prompting,Ekaterina Svikhnushina,"With conversational models becoming increasingly available to the general public, developing scalable and robust evaluation metrics is crucial to minimize potential social and psychological risks for the users. Existing evaluation metrics aim to automate offline user evaluation and approximate human judgment of pre-curated dialogs. However, they are limited in their ability to capture subjective perceptions of users who actually interact with the chatbots and might not generalize to real-world settings. To address this limitation, we propose an approach to approximate online human evaluation, leveraging large language models (LLMs) from the GPT-family. We introduce a new Dialog system Evaluation framework based on Prompting (DEP), which enables a fully automatic evaluation pipeline that replicates live user studies and achieves an impressive correlation with human judgment (up to Pearson r=0.95 on a system level). The DEP approach involves collecting synthetic chat logs of evaluated bots with an LLM in the other-play setting, where the LLM is carefully conditioned to follow a specific scenario. We further explore different prompting approaches to produce evaluation scores with the same LLM. The best-performing prompts, which contain few-shot demonstrations and instructions, show outstanding performance on the tested dataset and demonstrate the ability to generalize to other dialog corpora.","{'model': 'tldr@v2.0.0', 'text': 'A new Dialog system Evaluation framework based on Prompting is introduced, which enables a fully automatic evaluation pipeline that replicates live user studies and achieves an impressive correlation with human judgment (up to Pearson r=0.95 on a system level).'}",
-march in chat: interactive prompting for remote embodied referring expression,Yanyuan Qiao,"Many Vision-and-Language Navigation (VLN) tasks have been proposed in recent years, from room-based to object-based and indoor to outdoor. The REVERIE (Remote Embodied Referring Expression) is interesting since it only provides high-level instructions to the agent, which are closer to human commands in practice. Nevertheless, this poses more challenges than other VLN tasks since it requires agents to infer a navigation plan only based on a short instruction. Large Language Models (LLMs) show great potential in robot action planning by providing proper prompts. Still, this strategy has not been explored under the REVERIE settings. There are several new challenges. For example, the LLM should be environment-aware so that the navigation plan can be adjusted based on the current visual observation. Moreover, the LLM planned actions should be adaptable to the much larger and more complex REVERIE environment. This paper proposes a March-in-Chat (MiC) model that can talk to the LLM on the fly and plan dynamically based on a newly proposed Room-and-Object Aware Scene Perceiver (ROASP). Our MiC model outperforms the previous state-of-the-art by large margins by SPL and RGSPL metrics on the REVERIE benchmark.","{'model': 'tldr@v2.0.0', 'text': 'A March-in-Chat (MiC) model that can talk to the LLM on the fly and plan dynamically based on a newly proposed Room-and-Object Aware Scene Perceiver (ROASP) and outperforms the previous state-of-the-art by large margins by SPL and RGSPL metrics on the REVERIE benchmark.'}",https://arxiv.org/pdf/2308.10141
-prompting a large language model to generate diverse motivational messages: a comparison with human-written messages,Samuel Rhys Cox,"Large language models (LLMs) are increasingly capable and prevalent, and can be used to produce creative content. The quality of content is influenced by the prompt used, with more specific prompts that incorporate examples generally producing better results. On from this, it could be seen that using instructions written for crowdsourcing tasks (that are specific and include examples to guide workers) could prove effective LLM prompts. To explore this, we used a previous crowdsourcing pipeline that gave examples to people to help them generate a collectively diverse corpus of motivational messages. We then used this same pipeline to generate messages using GPT-4, and compared the collective diversity of messages from: (1) crowd-writers, (2) GPT-4 using the pipeline, and (3&4) two baseline GPT-4 prompts. We found that the LLM prompts using the crowdsourcing pipeline caused GPT-4 to produce more diverse messages than the two baseline prompts. We also discuss implications from messages generated by both human writers and LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is found that using instructions written for crowdsourcing tasks (that are specific and include examples to guide workers) could prove effective LLM prompts that caused GPT-4 to produce more diverse messages than the two baseline prompts.'}",https://arxiv.org/pdf/2308.13479
-prompting gpt-3.5 for text-to-sql with de-semanticization and skeleton retrieval,Chunxi Guo,"Text-to-SQL is a task that converts a natural language question into a structured query language (SQL) to retrieve information from a database. Large language models (LLMs) work well in natural language generation tasks, but they are not specifically pre-trained to understand the syntax and semantics of SQL commands. In this paper, we propose an LLM-based framework for Text-to-SQL which retrieves helpful demonstration examples to prompt LLMs. However, questions with different database schemes can vary widely, even if the intentions behind them are similar and the corresponding SQL queries exhibit similarities. Consequently, it becomes crucial to identify the appropriate SQL demonstrations that align with our requirements. We design a de-semanticization mechanism that extracts question skeletons, allowing us to retrieve similar examples based on their structural similarity. We also model the relationships between question tokens and database schema items (i.e., tables and columns) to filter out scheme-related information. Our framework adapts the range of the database schema in prompts to balance length and valuable information. A fallback mechanism allows for a more detailed schema to be provided if the generated SQL query fails. Ours outperforms state-of-the-art models and demonstrates strong generalization ability on three cross-domain Text-to-SQL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'An LLM-based framework for Text-to-SQL which retrieves helpful demonstration examples to prompt LLMs and outperforms state-of-the-art models and demonstrates strong generalization ability on three cross-domain Text- to-SQL benchmarks.'}",
-large language models can self-improve,Jiaxin Huang,"Large Language Models (LLMs) have achieved excellent performances in various tasks. However, fine-tuning an LLM requires extensive supervision. Human, on the other hand, may improve their reasoning abilities by self-thinking without external inputs. In this work, we demonstrate that an LLM is also capable of self-improving with only unlabeled datasets. We use a pre-trained LLM to generate""high-confidence""rationale-augmented answers for unlabeled questions using Chain-of-Thought prompting and self-consistency, and fine-tune the LLM using those self-generated solutions as target outputs. We show that our approach improves the general reasoning ability of a 540B-parameter LLM (74.4%->82.1% on GSM8K, 78.2%->83.0% on DROP, 90.0%->94.4% on OpenBookQA, and 63.4%->67.9% on ANLI-A3) and achieves state-of-the-art-level performance, without any ground truth label. We conduct ablation studies and show that fine-tuning on reasoning is critical for self-improvement.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a pre-trained LLM to generate ""high-confidence""rationale-augmented answers for unlabeled questions using Chain-of-Thought prompting and self-consistency, and fine-tune the LLM using those self-generated solutions as target outputs to improve the general reasoning ability.'}",http://arxiv.org/pdf/2210.11610
-multimodal chain-of-thought reasoning in language models,Zhuosheng Zhang,"Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies have focused on the language modality. We propose Multimodal-CoT that incorporates language (text) and vision (images) modalities into a two-stage framework that separates rationale generation and answer inference. In this way, answer inference can leverage better generated rationales that are based on multimodal information. With Multimodal-CoT, our model under 1 billion parameters outperforms the previous state-of-the-art LLM (GPT-3.5) by 16 percentage points (75.17%->91.68% accuracy) on the ScienceQA benchmark and even surpasses human performance. Code is publicly available available at https://github.com/amazon-science/mm-cot.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Multimodal-CoT that incorporates language (text) and vision (images) modalities into a two-stage framework that separates rationale generation and answer inference so that answer inference can leverage better generated rationales that are based on multimodal information.'}",http://arxiv.org/pdf/2302.00923
-towards expert-level medical question answering with large language models,K. Singhal,"Recent artificial intelligence (AI) systems have reached milestones in""grand challenges""ranging from Go to protein-folding. The capability to retrieve medical knowledge, reason over it, and answer medical questions comparably to physicians has long been viewed as one such grand challenge. Large language models (LLMs) have catalyzed significant progress in medical question answering; Med-PaLM was the first model to exceed a""passing""score in US Medical Licensing Examination (USMLE) style questions with a score of 67.2% on the MedQA dataset. However, this and other prior work suggested significant room for improvement, especially when models' answers were compared to clinicians' answers. Here we present Med-PaLM 2, which bridges these gaps by leveraging a combination of base LLM improvements (PaLM 2), medical domain finetuning, and prompting strategies including a novel ensemble refinement approach. Med-PaLM 2 scored up to 86.5% on the MedQA dataset, improving upon Med-PaLM by over 19% and setting a new state-of-the-art. We also observed performance approaching or exceeding state-of-the-art across MedMCQA, PubMedQA, and MMLU clinical topics datasets. We performed detailed human evaluations on long-form questions along multiple axes relevant to clinical applications. In pairwise comparative ranking of 1066 consumer medical questions, physicians preferred Med-PaLM 2 answers to those produced by physicians on eight of nine axes pertaining to clinical utility (p<0.001). We also observed significant improvements compared to Med-PaLM on every evaluation axis (p<0.001) on newly introduced datasets of 240 long-form""adversarial""questions to probe LLM limitations. While further studies are necessary to validate the efficacy of these models in real-world settings, these results highlight rapid progress towards physician-level performance in medical question answering.","{'model': 'tldr@v2.0.0', 'text': 'Results highlight rapid progress towards physician-level performance in medical question answering by leveraging a combination of base LLM improvements (PaLM 2), medical domain finetuning, and prompting strategies including a novel ensemble refinement approach.'}",http://arxiv.org/pdf/2305.09617
-language models can solve computer tasks,Geunwoo Kim,"Agents capable of carrying out general tasks on a computer can improve efficiency and productivity by automating repetitive tasks and assisting in complex problem-solving. Ideally, such agents should be able to solve new computer tasks presented to them through natural language commands. However, previous approaches to this problem require large amounts of expert demonstrations and task-specific reward functions, both of which are impractical for new tasks. In this work, we show that a pre-trained large language model (LLM) agent can execute computer tasks guided by natural language using a simple prompting scheme where the agent Recursively Criticizes and Improves its output (RCI). The RCI approach significantly outperforms existing LLM methods for automating computer tasks and surpasses supervised learning (SL) and reinforcement learning (RL) approaches on the MiniWoB++ benchmark. We compare multiple LLMs and find that RCI with the InstructGPT-3+RLHF LLM is state-of-the-art on MiniWoB++, using only a handful of demonstrations per task rather than tens of thousands, and without a task-specific reward function. Furthermore, we demonstrate RCI prompting's effectiveness in enhancing LLMs' reasoning abilities on a suite of natural language reasoning tasks, outperforming chain of thought (CoT) prompting. We find that RCI combined with CoT performs better than either separately. Our code can be found here: https://github.com/posgnu/rci-agent.","{'model': 'tldr@v2.0.0', 'text': ""This work shows that a pre-trained large language model (LLM) agent can execute computer tasks guided by natural language using a simple prompting scheme where the agent Recursively Criticizes and Improves its output (RCI), and demonstrates RCI prompting's effectiveness in enhancing LLMs' reasoning abilities on a suite of natural language reasoning tasks, outperforming chain of thought (CoT) prompting.""}",http://arxiv.org/pdf/2303.17491
-how is chatgpt's behavior changing over time?,Lingjiao Chen,"GPT-3.5 and GPT-4 are the two most widely used large language model (LLM) services. However, when and how these models are updated over time is opaque. Here, we evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on several diverse tasks: 1) math problems, 2) sensitive/dangerous questions, 3) opinion surveys, 4) multi-hop knowledge-intensive questions, 5) generating code, 6) US Medical License tests, and 7) visual reasoning. We find that the performance and behavior of both GPT-3.5 and GPT-4 can vary greatly over time. For example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy). This is partly explained by a drop in GPT-4's amenity to follow chain-of-thought prompting. Interestingly, GPT-3.5 was much better in June than in March in this task. GPT-4 became less willing to answer sensitive questions and opinion survey questions in June than in March. GPT-4 performed better at multi-hop questions in June than in March, while GPT-3.5's performance dropped on this task. Both GPT-4 and GPT-3.5 had more formatting mistakes in code generation in June than in March. We provide evidence that GPT-4's ability to follow user instructions has decreased over time, which is one common factor behind the many behavior drifts. Overall, our findings show that the behavior of the""same""LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Overall, the findings show that the behavior of the""same""LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.'}",https://arxiv.org/pdf/2307.09009
-is chatgpt the ultimate programming assistant - how far is it?,Haoye Tian,"Recently, the ChatGPT LLM has received great attention: it can be used as a bot for discussing source code, prompting it to suggest changes, provide descriptions or even generate code. Typical demonstrations generally focus on existing benchmarks, which may have been used in model training (i.e., data leakage). To assess the feasibility of using an LLM as a useful assistant bot for programmers, we must assess its realistic capabilities on unseen problems as well as its capabilities on various tasks. In this paper, we present an empirical study of ChatGPT's potential as a fully automated programming assistant, focusing on the tasks of code generation, program repair, and code summariziation. The study investigates ChatGPT's performance on common programming problems and compares it with state-of-the-art approaches on two benchmarks. Among several findings, our study shows that ChatGPT is effective in dealing with common programming problems. However, our experiments also reveal limitations in terms of its attention span: detailed descriptions will constrain the focus of ChatGPT and prevent it from leveraging its vast knowledge to solve the actual problem. Surprisingly, we have identified the ability of ChatGPT to reason the original intention of the code. We expect future work to build on this insight for dealing with the open question of the oracle problem. Our findings contribute interesting insights to the development of LLMs for programming assistance, notably by demonstrating the importance of prompt engineering, and providing a better understanding of ChatGPT's practical applications for software engineering.","{'model': 'tldr@v2.0.0', 'text': ""An empirical study of ChatGPT's potential as a fully automated programming assistant, focusing on the tasks of code generation, program repair, and code summariziation and identifies the ability ofChatGPT to reason the original intention of the code.""}",https://arxiv.org/pdf/2304.11938
-art: automatic multi-step reasoning and tool-use for large language models,Bhargavi Paranjape,"Large language models (LLMs) can perform complex reasoning in few- and zero-shot settings by generating intermediate chain of thought (CoT) reasoning steps. Further, each reasoning step can rely on external tools to support computation beyond the core LLM capabilities (e.g. search/running code). Prior work on CoT prompting and tool use typically requires hand-crafting task-specific demonstrations and carefully scripted interleaving of model generations with tool use. We introduce Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program. Given a new task to solve, ART selects demonstrations of multi-step reasoning and tool use from a task library. At test time, ART seamlessly pauses generation whenever external tools are called, and integrates their output before resuming generation. ART achieves a substantial improvement over few-shot prompting and automatic CoT on unseen tasks in the BigBench and MMLU benchmarks, and matches performance of hand-crafted CoT prompts on a majority of these tasks. ART is also extensible, and makes it easy for humans to improve performance by correcting errors in task-specific programs or incorporating new tools, which we demonstrate by drastically improving performance on select tasks with minimal human intervention.","{'model': 'tldr@v2.0.0', 'text': 'Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program, achieves a substantial improvement over few-shot prompting and automatic CoT on unseen tasks in the BigBench and MMLU benchmarks, and matches performance of hand-crafted CoT prompts on a majority of these tasks.'}",http://arxiv.org/pdf/2303.09014
-graph of thoughts: solving elaborate problems with large language models,Maciej Besta,"We introduce Graph of Thoughts (GoT): a framework that advances prompting capabilities in large language models (LLMs) beyond those offered by paradigms such as Chain-of-Thought or Tree of Thoughts (ToT). The key idea and primary advantage of GoT is the ability to model the information generated by an LLM as an arbitrary graph, where units of information (""LLM thoughts"") are vertices, and edges correspond to dependencies between these vertices. This approach enables combining arbitrary LLM thoughts into synergistic outcomes, distilling the essence of whole networks of thoughts, or enhancing thoughts using feedback loops. We illustrate that GoT offers advantages over state of the art on different tasks, for example increasing the quality of sorting by 62% over ToT, while simultaneously reducing costs by>31%. We ensure that GoT is extensible with new thought transformations and thus can be used to spearhead new prompting schemes. This work brings the LLM reasoning closer to human thinking or brain mechanisms such as recurrence, both of which form complex networks.","{'model': 'tldr@v2.0.0', 'text': 'Graph of Thoughts is introduced: a framework that advances prompting capabilities in large language models (LLMs) beyond those offered by paradigms such as Chain-of-Thought or Tree of Thoughts, and is ensured that GoT is extensible with new thought transformations and thus can be used to spearhead new prompting schemes.'}",https://arxiv.org/pdf/2308.09687
-assessing the utility of chatgpt throughout the entire clinical workflow,Arya Rao,"IMPORTANCE: Large language model (LLM) artificial intelligence (AI) chatbots direct the power of large training datasets towards successive, related tasks, as opposed to single-ask tasks, for which AI already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as virtual physicians, has not yet been evaluated. OBJECTIVE: To evaluate ChatGPT's capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. DESIGN: We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. SETTING: ChatGPT, a publicly available LLM PARTICIPANTS: Clinical vignettes featured hypothetical patients with a variety of age and gender identities, and a range of Emergency Severity Indices (ESIs) based on initial clinical presentation. EXPOSURES: MSD Clinical Manual vignettes MAIN OUTCOMES AND MEASURES: We measured the proportion of correct responses to the questions posed within the clinical vignettes tested. RESULTS: ChatGPT achieved 71.7% (95% CI, 69.3% to 74.1%) accuracy overall across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI, 67.8% to 86.1%), and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI, 54.2% to 66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis ({beta}=-15.8%, p<0.001) and clinical management ({beta}=-7.4%, p=0.02) type questions. CONCLUSIONS AND RELEVANCE: ChatGPT achieves impressive accuracy in clinical decision making, with particular strengths emerging as it has more clinical information at its disposal.","{'model': 'tldr@v2.0.0', 'text': ""Evaluating ChatGPT's capacity for ongoing clinical decision support via its performance on standardized clinical vignettes achieves impressive accuracy in clinical decision making, with particular strengths emerging as it has more clinical information at its disposal.""}",https://www.medrxiv.org/content/medrxiv/early/2023/02/26/2023.02.21.23285886.full.pdf
-task and motion planning with large language models for object rearrangement,Yan Ding,"Multi-object rearrangement is a crucial skill for service robots, and commonsense reasoning is frequently needed in this process. However, achieving commonsense arrangements requires knowledge about objects, which is hard to transfer to robots. Large language models (LLMs) are one potential source of this knowledge, but they do not naively capture information about plausible physical arrangements of the world. We propose LLM-GROP, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry. LLM-GROP allows us to go from natural-language commands to human-aligned object rearrangement in varied environments. Based on human evaluations, our approach achieves the highest rating while outperforming competitive baselines in terms of success rate while maintaining comparable cumulative action costs. Finally, we demonstrate a practical implementation of LLM-GROP on a mobile manipulator in real-world scenarios. Supplementary materials are available at: https://sites.google.com/view/llm-grop","{'model': 'tldr@v2.0.0', 'text': 'LLM-GROP is proposed, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry.'}",http://arxiv.org/pdf/2303.06247
-interleaving retrieval with chain-of-thought reasoning for knowledge-intensive multi-step questions,H. Trivedi,"Prompting-based large language models (LLMs) are surprisingly powerful at generating natural language reasoning steps or Chains-of-Thoughts (CoT) for multi-step question answering (QA). They struggle, however, when the necessary knowledge is either unavailable to the LLM or not up-to-date within its parameters. While using the question to retrieve relevant text from an external knowledge source helps LLMs, we observe that this one-step retrieve-and-read approach is insufficient for multi-step QA. Here, what to retrieve depends on what has already been derived, which in turn may depend on what was previously retrieved. To address this, we propose IRCoT, a new approach for multi-step QA that interleaves retrieval with steps (sentences) in a CoT, guiding the retrieval with CoT and in turn using retrieved results to improve CoT. Using IRCoT with GPT3 substantially improves retrieval (up to 21 points) as well as downstream QA (up to 15 points) on four datasets: HotpotQA, 2WikiMultihopQA, MuSiQue, and IIRC. We observe similar substantial gains in out-of-distribution (OOD) settings as well as with much smaller models such as Flan-T5-large without additional training. IRCoT reduces model hallucination, resulting in factually more accurate CoT reasoning.","{'model': 'tldr@v2.0.0', 'text': 'IRCoT is proposed, a new approach for multi-step QA that interleaves retrieval with steps in a CoT, guiding the retrieval with CoT and in turn using retrieved results to improve CoT.'}",http://arxiv.org/pdf/2212.10509
-gpt-ner: named entity recognition via large language models,Shuhe Wang,"Despite the fact that large-scale Language Models (LLM) have achieved SOTA performances on a variety of NLP tasks, its performance on NER is still significantly below supervised baselines. This is due to the gap between the two tasks the NER and LLMs: the former is a sequence labeling task in nature while the latter is a text-generation model. In this paper, we propose GPT-NER to resolve this issue. GPT-NER bridges the gap by transforming the sequence labeling task to a generation task that can be easily adapted by LLMs e.g., the task of finding location entities in the input text""Columbus is a city""is transformed to generate the text sequence""@@Columbus## is a city"", where special tokens @@## marks the entity to extract. To efficiently address the""hallucination""issue of LLMs, where LLMs have a strong inclination to over-confidently label NULL inputs as entities, we propose a self-verification strategy by prompting LLMs to ask itself whether the extracted entities belong to a labeled entity tag. We conduct experiments on five widely adopted NER datasets, and GPT-NER achieves comparable performances to fully supervised baselines, which is the first time as far as we are concerned. More importantly, we find that GPT-NER exhibits a greater ability in the low-resource and few-shot setups, when the amount of training data is extremely scarce, GPT-NER performs significantly better than supervised models. This demonstrates the capabilities of GPT-NER in real-world NER applications where the number of labeled examples is limited.","{'model': 'tldr@v2.0.0', 'text': 'GPT-NER exhibits a greater ability in the low-resource and few-shot setups, when the amount of training data is extremely scarce, and performs significantly better than supervised models, which demonstrates the capabilities of GPT-ner in real-world NER applications where the number of labeled examples is limited.'}",https://arxiv.org/pdf/2304.10428
-unleashing cognitive synergy in large language models: a task-solving agent through multi-persona self-collaboration,Zhenhailong Wang,"Human intelligence thrives on the concept of cognitive synergy, where collaboration and information integration among different cognitive processes yield superior outcomes compared to individual cognitive processes in isolation. Although Large Language Models (LLMs) have demonstrated promising performance as general task-solving agents, they still struggle with tasks that require intensive domain knowledge and complex reasoning. In this work, we propose Solo Performance Prompting (SPP), which transforms a single LLM into a cognitive synergist by engaging in multi-turn self-collaboration with multiple personas. A cognitive synergist refers to an intelligent agent that collaborates with multiple minds, combining their individual strengths and knowledge, to enhance problem-solving and overall performance in complex tasks. By dynamically identifying and simulating different personas based on task inputs, SPP unleashes the potential of cognitive synergy in LLMs. We have discovered that assigning multiple, fine-grained personas in LLMs elicits better problem-solving abilities compared to using a single or fixed number of personas. We evaluate SPP on three challenging tasks: Trivia Creative Writing, Codenames Collaborative, and Logic Grid Puzzle, encompassing both knowledge-intensive and reasoning-intensive types. Unlike previous works, such as Chain-of-Thought, that solely enhance the reasoning abilities in LLMs, SPP effectively elicits internal knowledge acquisition abilities, reduces hallucination, and maintains strong reasoning capabilities. Code, data, and prompts can be found at: https://github.com/MikeWangWZHL/Solo-Performance-Prompting.git.","{'model': 'tldr@v2.0.0', 'text': 'Solo Performance Prompting (SPP), which transforms a single LLM into a cognitive synergist by engaging in multi-turn self-collaboration with multiple personas by dynamically identifying and simulating different personas based on task inputs, unleashes the potential of cognitive synergy in LLMs.'}",https://arxiv.org/pdf/2307.05300
-safety assessment of chinese large language models,Hao Sun,"With the rapid popularity of large language models such as ChatGPT and GPT-4, a growing amount of attention is paid to their safety concerns. These models may generate insulting and discriminatory content, reflect incorrect social values, and may be used for malicious purposes such as fraud and dissemination of misleading information. Evaluating and enhancing their safety is particularly essential for the wide application of large language models (LLMs). To further promote the safe deployment of LLMs, we develop a Chinese LLM safety assessment benchmark. Our benchmark explores the comprehensive safety performance of LLMs from two perspectives: 8 kinds of typical safety scenarios and 6 types of more challenging instruction attacks. Our benchmark is based on a straightforward process in which it provides the test prompts and evaluates the safety of the generated responses from the evaluated model. In evaluation, we utilize the LLM's strong evaluation ability and develop it as a safety evaluator by prompting. On top of this benchmark, we conduct safety assessments and analyze 15 LLMs including the OpenAI GPT series and other well-known Chinese LLMs, where we observe some interesting findings. For example, we find that instruction attacks are more likely to expose safety issues of all LLMs. Moreover, to promote the development and deployment of safe, responsible, and ethical AI, we publicly release SafetyPrompts including 100k augmented prompts and responses by LLMs.","{'model': 'tldr@v2.0.0', 'text': ""A Chinese LLM safety assessment benchmark is developed based on the LLM's strong evaluation ability and develops it as a safety evaluator by prompting, which finds that instruction attacks are more likely to expose safety issues of all LLMs.""}",http://arxiv.org/pdf/2304.10436
-can llms express their uncertainty? an empirical evaluation of confidence elicitation in llms,Miao Xiong,"The task of empowering large language models (LLMs) to accurately express their confidence, referred to as confidence elicitation, is essential in ensuring reliable and trustworthy decision-making processes. Previous methods, which primarily rely on model logits, have become less suitable for LLMs and even infeasible with the rise of closed-source LLMs (e.g., commercialized LLM APIs). This leads to a growing need to explore the untapped area of \emph{non-logit-based} approaches to estimate the uncertainty of LLMs. Hence, in this study, we investigate approaches for confidence elicitation that do not require model fine-tuning or access to proprietary information. We introduce three categories of methods: verbalize-based, consistency-based, and their hybrid methods for benchmarking, and evaluate their performance across five types of datasets and four widely-used LLMs. Our analysis of these methods uncovers several key insights: 1) LLMs often exhibit a high degree of overconfidence when verbalizing their confidence; 2) Prompting strategies such as CoT, Top-K and Multi-step confidences improve calibration of verbalized confidence; 3) Consistency-based methods outperform the verbalized confidences in most cases, with particularly notable improvements on the arithmetic reasoning task; 4) Hybrid methods consistently deliver the best performance over their baselines, thereby emerging as a promising state-of-the-art approach; 5) Despite these advancements, all investigated methods continue to struggle with challenging tasks, such as those requiring professional knowledge, leaving significant scope for improvement of confidence elicitation.","{'model': 'tldr@v2.0.0', 'text': 'This study investigates approaches for confidence elicitation that do not require model fine-tuning or access to proprietary information, and introduces three categories of methods: verbalize- based, consistency-based, and their hybrid methods for benchmarking, and evaluates their performance across five types of datasets and four widely-used LLMs.'}",http://arxiv.org/pdf/2306.13063
-adaptive test generation using a large language model,Max Schäfer,"—Unit tests play a key role in ensuring the correctness of software. However, manually creating unit tests is a laborious task, motivating the need for automation. This paper presents T EST P ILOT , an adaptive test generation technique that leverages Large Language Models (LLMs). T EST P ILOT uses Codex, an off-the-shelf LLM, to automatically generate unit tests for a given program without requiring additional training or few-shot learning on examples of existing tests. In our approach, Codex is provided with prompts that include the signature and implementation of a function under test, along with usage examples extracted from documentation. If a generated test fails, T EST P ILOT ’s adaptive component attempts to generate a new test that fixes the problem by re-prompting the model with the failing test and error message. We created an implementation of T EST P ILOT for JavaScript and evaluated it on 25 npm packages with a total of 1,684 API functions to generate tests for. Our results show that the generated tests achieve up to 93.1% statement coverage (median 68.2%). Moreover, on average, 58.5% of the generated tests contain at least one assertion that exercises functionality from the package under test. Our experiments with excluding parts of the information included in the prompts show that all components contribute towards the generation of effective test suites. Finally, we find that T EST P ILOT does not generate memorized tests: 92.7% of our generated tests have ≤ 50% similarity with existing tests (as measured by normalized edit distance), with none of them being exact copies.","{'model': 'tldr@v2.0.0', 'text': 'T EST P ILOT uses Codex, an off-the-shelf LLM, to automatically generate unit tests for a given program without requiring additional training or few-shot learning on examples of existing tests, and does not generate memorized tests.'}",https://arxiv.org/pdf/2302.06527
-when to make exceptions: exploring language models as accounts of human moral judgment,Zhijing Jin,"AI systems are becoming increasingly intertwined with human life. In order to effectively collaborate with humans and ensure safety, AI systems need to be able to understand, interpret and predict human moral judgments and decisions. Human moral judgments are often guided by rules, but not always. A central challenge for AI safety is capturing the flexibility of the human moral mind -- the ability to determine when a rule should be broken, especially in novel or unusual situations. In this paper, we present a novel challenge set consisting of rule-breaking question answering (RBQA) of cases that involve potentially permissible rule-breaking -- inspired by recent moral psychology studies. Using a state-of-the-art large language model (LLM) as a basis, we propose a novel moral chain of thought (MORALCOT) prompting strategy that combines the strengths of LLMs with theories of moral reasoning developed in cognitive science to predict human moral judgments. MORALCOT outperforms seven existing LLMs by 6.2% F1, suggesting that modeling human reasoning might be necessary to capture the flexibility of the human moral mind. We also conduct a detailed error analysis to suggest directions for future work to improve AI safety using RBQA. Our data is open-sourced at https://huggingface.co/datasets/feradauto/MoralExceptQA and code at https://github.com/feradauto/MoralCoT","{'model': 'tldr@v2.0.0', 'text': 'A novel moral chain of thought (MORALCOT) prompting strategy that combines the strengths of LLMs with theories of moral reasoning developed in cognitive science to predict human moral judgments, suggesting that modeling human reasoning might be necessary to capture the flexibility of the human moral mind.'}",https://arxiv.org/pdf/2210.01478
-expertprompting: instructing large language models to be distinguished experts,Benfeng Xu,"The answering quality of an aligned large language model (LLM) can be drastically improved if treated with proper crafting of prompts. In this paper, we propose ExpertPrompting to elicit the potential of LLMs to answer as distinguished experts. We first utilize In-Context Learning to automatically synthesize detailed and customized descriptions of the expert identity for each specific instruction, and then ask LLMs to provide answer conditioned on such agent background. Based on this augmented prompting strategy, we produce a new set of instruction-following data using GPT-3.5, and train a competitive open-source chat assistant called ExpertLLaMA. We employ GPT4-based evaluation to show that 1) the expert data is of significantly higher quality than vanilla answers, and 2) ExpertLLaMA outperforms existing open-source opponents and achieves 96\% of the original ChatGPT's capability. All data and the ExpertLLaMA model will be made publicly available at \url{https://github.com/OFA-Sys/ExpertLLaMA}.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes ExpertPrompting to elicit the potential of LLMs to answer as distinguished experts, and produces a new set of instruction-following data using GPT-3.5, and trains a competitive open-source chat assistant called ExpertLLaMA.'}",http://arxiv.org/pdf/2305.14688
-an empirical evaluation of using large language models for automated unit test generation,Max Schafer,"Unit tests play a key role in ensuring the correctness of software. However, manually creating unit tests is a laborious task, motivating the need for automation. Large Language Models (LLMs) have recently been applied to this problem, utilizing additional training or few-shot learning on examples of existing tests. This paper presents a large-scale empirical evaluation on the effectiveness of LLMs for automated unit test generation without additional training or manual effort, providing the LLM with the signature and implementation of the function under test, along with usage examples extracted from documentation. We also attempt to repair failed generated tests by re-prompting the model with the failing test and error message. We implement our approach in TestPilot, a test generation tool for JavaScript that automatically generates unit tests for all API functions in an npm package. We evaluate TestPilot using OpenAI's gpt3.5-turbo LLM on 25 npm packages with a total of 1,684 API functions. The generated tests achieve a median statement coverage of 70.2% and branch coverage of 52.8%, significantly improving on Nessie, a recent feedback-directed JavaScript test generation technique, which achieves only 51.3% statement coverage and 25.6% branch coverage. We also find that 92.8% of TestPilot's generated tests have no more than 50% similarity with existing tests (as measured by normalized edit distance), with none of them being exact copies. Finally, we run TestPilot with two additional LLMs, OpenAI's older code-cushman-002 LLM and the open LLM StarCoder. Overall, we observed similar results with the former (68.2% median statement coverage), and somewhat worse results with the latter (54.0% median statement coverage), suggesting that the effectiveness of the approach is influenced by the size and training set of the LLM, but does not fundamentally depend on the specific model.","{'model': 'tldr@v2.0.0', 'text': 'A large-scale empirical evaluation on the effectiveness of large Language Models for automated unit test generation without additional training or manual effort, providing the LLM with the signature and implementation of the function under test, along with usage examples extracted from documentation.'}",
-automatic evaluation of attribution by large language models,Xiang Yue,"A recent focus of large language model (LLM) development, as exemplified by generative search engines, is to incorporate external references to generate and support its claims. However, evaluating the attribution, i.e., verifying whether the generated statement is fully supported by the cited reference, remains an open problem. Although human evaluation is common practice, it is costly and time-consuming. In this paper, we investigate the automatic evaluation of attribution given by LLMs. We begin by defining different types of attribution errors, and then explore two approaches for automatic evaluation: prompting LLMs and fine-tuning smaller LMs. The fine-tuning data is repurposed from related tasks such as question answering, fact-checking, natural language inference, and summarization. We manually curate a set of test examples covering 12 domains from a generative search engine, New Bing. Our results on this curated test set and simulated examples from existing benchmarks highlight both promising signals and challenges. We hope our problem formulation, testbeds, and findings will help lay the foundation for future studies on this important problem.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the automatic evaluation of attribution given by large language model (LLMs), defining different types of attribution errors, and exploring two approaches for automatic evaluation: prompting LLMs and fine-tuning smaller LMs.'}",https://arxiv.org/pdf/2305.06311
-cape: corrective actions from precondition errors using large language models,S. S. Raman,"Extracting commonsense knowledge from a large language model (LLM) offers a path to designing intelligent robots. Existing approaches that leverage LLMs for planning are unable to recover when an action fails and often resort to retrying failed actions, without resolving the error's underlying cause. We propose a novel approach (CAPE) that attempts to propose corrective actions to resolve precondition errors during planning. CAPE improves the quality of generated plans by leveraging few-shot reasoning from action preconditions. Our approach enables embodied agents to execute more tasks than baseline methods while ensuring semantic correctness and minimizing re-prompting. In VirtualHome, CAPE generates executable plans while improving a human-annotated plan correctness metric from 28.89% to 49.63% over SayCan. Our improvements transfer to a Boston Dynamics Spot robot initialized with a set of skills (specified in language) and associated preconditions, where CAPE improves the correctness metric of the executed task plans by 76.49% compared to SayCan. Our approach enables the robot to follow natural language commands and robustly recover from failures, which baseline approaches largely cannot resolve or address inefficiently.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach (CAPE) that attempts to propose corrective actions to resolve precondition errors during planning and improves the quality of generated plans by leveraging few-shot reasoning from action preconditions.'}",
-on the risk of misinformation pollution with large language models,Yikang Pan,"In this paper, we comprehensively investigate the potential misuse of modern Large Language Models (LLMs) for generating credible-sounding misinformation and its subsequent impact on information-intensive applications, particularly Open-Domain Question Answering (ODQA) systems. We establish a threat model and simulate potential misuse scenarios, both unintentional and intentional, to assess the extent to which LLMs can be utilized to produce misinformation. Our study reveals that LLMs can act as effective misinformation generators, leading to a significant degradation in the performance of ODQA systems. To mitigate the harm caused by LLM-generated misinformation, we explore three defense strategies: prompting, misinformation detection, and majority voting. While initial results show promising trends for these defensive strategies, much more work needs to be done to address the challenge of misinformation pollution. Our work highlights the need for further research and interdisciplinary collaboration to address LLM-generated misinformation and to promote responsible use of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that LLMs can act as effective misinformation generators, leading to a significant degradation in the performance of ODQA systems, and the need for further research and interdisciplinary collaboration to address LLM-generated misinformation.'}",http://arxiv.org/pdf/2305.13661
-building cooperative embodied agents modularly with large language models,Hongxin Zhang,"Large Language Models (LLMs) have demonstrated impressive planning abilities in single-agent embodied tasks across various domains. However, their capacity for planning and communication in multi-agent cooperation remains unclear, even though these are crucial skills for intelligent embodied agents. In this paper, we present a novel framework that utilizes LLMs for multi-agent cooperation and tests it in various embodied environments. Our framework enables embodied agents to plan, communicate, and cooperate with other embodied agents or humans to accomplish long-horizon tasks efficiently. We demonstrate that recent LLMs, such as GPT-4, can surpass strong planning-based methods and exhibit emergent effective communication using our framework without requiring fine-tuning or few-shot prompting. We also discover that LLM-based agents that communicate in natural language can earn more trust and cooperate more effectively with humans. Our research underscores the potential of LLMs for embodied AI and lays the foundation for future research in multi-agent cooperation. Videos can be found on the project website https://vis-www.cs.umass.edu/Co-LLM-Agents/.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel framework that utilizes LLMs for multi-agent cooperation and tests it in various embodied environments, and discovers that LLM-based agents that communicate in natural language can earn more trust and cooperate more effectively with humans.'}",https://arxiv.org/pdf/2307.02485
-logic-lm: empowering large language models with symbolic solvers for faithful logical reasoning,Liangming Pan,"Large Language Models (LLMs) have shown human-like reasoning abilities but still struggle with complex logical problems. This paper introduces a novel framework, Logic-LM, which integrates LLMs with symbolic solvers to improve logical problem-solving. Our method first utilizes LLMs to translate a natural language problem into a symbolic formulation. Afterward, a deterministic symbolic solver performs inference on the formulated problem. We also introduce a self-refinement module, which utilizes the symbolic solver's error messages to revise symbolic formalizations. We demonstrate Logic-LM's effectiveness on five logical reasoning datasets: ProofWriter, PrOntoQA, FOLIO, LogicalDeduction, and AR-LSAT. On average, Logic-LM achieves a significant performance boost of 39.2% over using LLM alone with standard prompting and 18.4% over LLM with chain-of-thought prompting. Our findings suggest that Logic-LM, by combining LLMs with symbolic logic, offers a promising avenue for faithful logical reasoning. Code and data are publicly available at https://github.com/teacherpeterpan/Logic-LLM.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel framework, Logic-LM, which integrates LLMs with symbolic solvers to improve logical problem-solving and suggests that it offers a promising avenue for faithful logical reasoning.'}",http://arxiv.org/pdf/2305.12295
-red-teaming large language models using chain of utterances for safety-alignment,Rishabh Bhardwaj,"Larger language models (LLMs) have taken the world by storm with their massive multi-tasking capabilities simply by optimizing over a next-word prediction objective. With the emergence of their properties and encoded knowledge, the risk of LLMs producing harmful outputs increases, making them unfit for scalable deployment for the public. In this work, we propose a new safety evaluation benchmark RED-EVAL that carries out red-teaming. We show that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting, jailbreaking closed source LLM-based systems such as GPT-4 and ChatGPT to unethically respond to more than 65% and 73% of harmful queries. We also demonstrate the consistency of the RED-EVAL across 8 open-source LLMs in generating harmful responses in more than 86% of the red-teaming attempts. Next, we propose RED-INSTRUCT--An approach for the safety alignment of LLMs. It constitutes two phases: 1) HARMFULQA data collection: Leveraging CoU prompting, we collect a dataset that consists of 1.9K harmful questions covering a wide range of topics, 9.5K safe and 7.3K harmful conversations from ChatGPT; 2) SAFE-ALIGN: We demonstrate how the conversational dataset can be used for the safety alignment of LLMs by minimizing the negative log-likelihood over helpful responses and penalizing over harmful responses by gradient accent over sample loss. Our model STARLING, a fine-tuned Vicuna-7B, is observed to be more safely aligned when evaluated on RED-EVAL and HHH benchmarks while preserving the utility of the baseline models (TruthfulQA, MMLU, and BBH).","{'model': 'tldr@v2.0.0', 'text': 'It is shown that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting, jailbreaking closed source LLM-based systems such as GPT-4 and ChatGPT to unethically respond to more than 65% and 73% of harmful queries.'}",https://arxiv.org/pdf/2308.09662
-can large language models write good property-based tests?,Vasudev Vikram,"Property-based testing (PBT), while an established technique in the software testing research community, is still relatively underused in real-world software. Pain points in writing property-based tests include implementing diverse random input generators and thinking of meaningful properties to test. Developers, however, are more amenable to writing documentation; plenty of library API documentation is available and can be used as natural language specifications for property-based tests. As large language models (LLMs) have recently shown promise in a variety of coding tasks, we explore the potential of using LLMs to synthesize property-based tests. We call our approach PBT-GPT, and propose three different strategies of prompting the LLM for PBT. We characterize various failure modes of PBT-GPT and detail an evaluation methodology for automatically synthesized property-based tests. PBT-GPT achieves promising results in our preliminary studies on sample Python library APIs in $\texttt{numpy}$, $\texttt{networkx}$, and $\texttt{datetime}$.","{'model': 'tldr@v2.0.0', 'text': 'This work describes various failure modes of PBT-GPT, and proposes three different strategies of prompting the LLM for PBT, and details an evaluation methodology for automatically synthesized property-based tests.'}",https://arxiv.org/pdf/2307.04346
-chatgpt makes medicine easy to swallow: an exploratory case study on simplified radiology reports.,Katharina Jeblick,,"{'model': 'tldr@v2.0.0', 'text': 'The initial insights of this study indicate a tremendous potential in using LLMs like ChatGPT to improve patient-centered care in radiology and other medical domains.'}",https://link.springer.com/content/pdf/10.1007/s00330-023-10213-1.pdf
-codehelp: using large language models with guardrails for scalable support in programming classes,Mark H. Liffiton,"Computing educators face significant challenges in providing timely support to students, especially in large class settings. Large language models (LLMs) have emerged recently and show great promise for providing on-demand help at a large scale, but there are concerns that students may over-rely on the outputs produced by these models. In this paper, we introduce CodeHelp, a novel LLM-powered tool designed with guardrails to provide on-demand assistance to programming students without directly revealing solutions. We detail the design of the tool, which incorporates a number of useful features for instructors, and elaborate on the pipeline of prompting strategies we use to ensure generated outputs are suitable for students. To evaluate CodeHelp, we deployed it in a first-year computer and data science course with 52 students and collected student interactions over a 12-week period. We examine students' usage patterns and perceptions of the tool, and we report reflections from the course instructor and a series of recommendations for classroom use. Our findings suggest that CodeHelp is well-received by students who especially value its availability and help with resolving errors, and that for instructors it is easy to deploy and complements, rather than replaces, the support that they provide to students.","{'model': 'tldr@v2.0.0', 'text': 'The findings suggest that CodeHelp is well-received by students who especially value its availability and help with resolving errors, and that for instructors it is easy to deploy and complements, rather than replaces, the support that they provide to students.'}",https://arxiv.org/pdf/2308.06921
-i spy a metaphor: large language models and diffusion models co-create visual metaphors,Tuhin Chakrabarty,"Visual metaphors are powerful rhetorical devices used to persuade or communicate creative ideas through images. Similar to linguistic metaphors, they convey meaning implicitly through symbolism and juxtaposition of the symbols. We propose a new task of generating visual metaphors from linguistic metaphors. This is a challenging task for diffusion-based text-to-image models, such as DALL$\cdot$E 2, since it requires the ability to model implicit meaning and compositionality. We propose to solve the task through the collaboration between Large Language Models (LLMs) and Diffusion Models: Instruct GPT-3 (davinci-002) with Chain-of-Thought prompting generates text that represents a visual elaboration of the linguistic metaphor containing the implicit meaning and relevant objects, which is then used as input to the diffusion-based text-to-image models.Using a human-AI collaboration framework, where humans interact both with the LLM and the top-performing diffusion model, we create a high-quality dataset containing 6,476 visual metaphors for 1,540 linguistic metaphors and their associated visual elaborations. Evaluation by professional illustrators shows the promise of LLM-Diffusion Model collaboration for this task . To evaluate the utility of our Human-AI collaboration framework and the quality of our dataset, we perform both an intrinsic human-based evaluation and an extrinsic evaluation using visual entailment as a downstream task.","{'model': 'tldr@v2.0.0', 'text': 'Using a human-AI collaboration framework, where humans interact both with the LLM and the top-performing diffusion model, a high-quality dataset is created containing 6,476 visual metaphors for 1,540 linguistic metaphors and their associated visual elaborations.'}",https://arxiv.org/pdf/2305.14724
-reviewergpt? an exploratory study on using large language models for paper reviewing,Ryan Liu,"Given the rapid ascent of large language models (LLMs), we study the question: (How) can large language models help in reviewing of scientific papers or proposals? We first conduct some pilot studies where we find that (i) GPT-4 outperforms other LLMs (Bard, Vicuna, Koala, Alpaca, LLaMa, Dolly, OpenAssistant, StableLM), and (ii) prompting with a specific question (e.g., to identify errors) outperforms prompting to simply write a review. With these insights, we study the use of LLMs (specifically, GPT-4) for three tasks: 1. Identifying errors: We construct 13 short computer science papers each with a deliberately inserted error, and ask the LLM to check for the correctness of these papers. We observe that the LLM finds errors in 7 of them, spanning both mathematical and conceptual errors. 2. Verifying checklists: We task the LLM to verify 16 closed-ended checklist questions in the respective sections of 15 NeurIPS 2022 papers. We find that across 119 {checklist question, paper} pairs, the LLM had an 86.6% accuracy. 3. Choosing the""better""paper: We generate 10 pairs of abstracts, deliberately designing each pair in such a way that one abstract was clearly superior than the other. The LLM, however, struggled to discern these relatively straightforward distinctions accurately, committing errors in its evaluations for 6 out of the 10 pairs. Based on these experiments, we think that LLMs have a promising use as reviewing assistants for specific reviewing tasks, but not (yet) for complete evaluations of papers or proposals.","{'model': 'tldr@v2.0.0', 'text': 'It is thought that LLMs have a promising use as reviewing assistants for specific reviewing tasks, but not for complete evaluations of papers or proposals.'}",http://arxiv.org/pdf/2306.00622
-graphologue: exploring large language model responses with interactive diagrams,Peiling Jiang,"Large language models (LLMs) have recently soared in popularity due to their ease of access and the unprecedented ability to synthesize text responses to diverse user questions. However, LLMs like ChatGPT present significant limitations in supporting complex information tasks due to the insufficient affordances of the text-based medium and linear conversational structure. Through a formative study with ten participants, we found that LLM interfaces often present long-winded responses, making it difficult for people to quickly comprehend and interact flexibly with various pieces of information, particularly during more complex tasks. We present Graphologue, an interactive system that converts text-based responses from LLMs into graphical diagrams to facilitate information-seeking and question-answering tasks. Graphologue employs novel prompting strategies and interface designs to extract entities and relationships from LLM responses and constructs node-link diagrams in real-time. Further, users can interact with the diagrams to flexibly adjust the graphical presentation and to submit context-specific prompts to obtain more information. Utilizing diagrams, Graphologue enables graphical, non-linear dialogues between humans and LLMs, facilitating information exploration, organization, and comprehension.","{'model': 'tldr@v2.0.0', 'text': 'Graphologue is an interactive system that converts text-based responses from LLMs into graphical diagrams to facilitate information-seeking and question-answering tasks, and enables graphical, non-linear dialogues between humans and LLMs, facilitating information exploration, organization, and comprehension.'}",https://arxiv.org/pdf/2305.11473
-large language models as tax attorneys: a case study in legal capabilities emergence,John J. Nay,"Better understanding of Large Language Models' (LLMs) legal analysis abilities can contribute to improving the efficiency of legal services, governing artificial intelligence, and leveraging LLMs to identify inconsistencies in law. This paper explores LLM capabilities in applying tax law. We choose this area of law because it has a structure that allows us to set up automated validation pipelines across thousands of examples, requires logical reasoning and maths skills, and enables us to test LLM capabilities in a manner relevant to real-world economic lives of citizens and companies. Our experiments demonstrate emerging legal understanding capabilities, with improved performance in each subsequent OpenAI model release. We experiment with retrieving and utilising the relevant legal authority to assess the impact of providing additional legal context to LLMs. Few-shot prompting, presenting examples of question-answer pairs, is also found to significantly enhance the performance of the most advanced model, GPT-4. The findings indicate that LLMs, particularly when combined with prompting enhancements and the correct legal texts, can perform at high levels of accuracy but not yet at expert tax lawyer levels. As LLMs continue to advance, their ability to reason about law autonomously could have significant implications for the legal profession and AI governance.",,http://arxiv.org/pdf/2306.07075
-assessing the utility of chatgpt throughout the entire clinical workflow: development and usability study,Arya Rao,"Background Large language model (LLM)–based artificial intelligence chatbots direct the power of large training data sets toward successive, related tasks as opposed to single-ask tasks, for which artificial intelligence already achieves impressive performance. The capacity of LLMs to assist in the full scope of iterative clinical reasoning via successive prompting, in effect acting as artificial physicians, has not yet been evaluated. Objective This study aimed to evaluate ChatGPT’s capacity for ongoing clinical decision support via its performance on standardized clinical vignettes. Methods We inputted all 36 published clinical vignettes from the Merck Sharpe & Dohme (MSD) Clinical Manual into ChatGPT and compared its accuracy on differential diagnoses, diagnostic testing, final diagnosis, and management based on patient age, gender, and case acuity. Accuracy was measured by the proportion of correct responses to the questions posed within the clinical vignettes tested, as calculated by human scorers. We further conducted linear regression to assess the contributing factors toward ChatGPT’s performance on clinical tasks. Results ChatGPT achieved an overall accuracy of 71.7% (95% CI 69.3%-74.1%) across all 36 clinical vignettes. The LLM demonstrated the highest performance in making a final diagnosis with an accuracy of 76.9% (95% CI 67.8%-86.1%) and the lowest performance in generating an initial differential diagnosis with an accuracy of 60.3% (95% CI 54.2%-66.6%). Compared to answering questions about general medical knowledge, ChatGPT demonstrated inferior performance on differential diagnosis (β=–15.8%; P<.001) and clinical management (β=–7.4%; P=.02) question types. Conclusions ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal. In particular, ChatGPT demonstrates the greatest accuracy in tasks of final diagnosis as compared to initial diagnosis. Limitations include possible model hallucinations and the unclear composition of ChatGPT’s training data set.","{'model': 'tldr@v2.0.0', 'text': 'ChatGPT achieves impressive accuracy in clinical decision-making, with increasing strength as it gains more clinical information at its disposal, in particular in tasks of final diagnosis as compared to initial diagnosis.'}",https://www.jmir.org/2023/1/e48659/PDF
-"despite ""super-human"" performance, current llms are unsuited for decisions about ethics and safety",Joshua Albrecht,"Large language models (LLMs) have exploded in popularity in the past few years and have achieved undeniably impressive results on benchmarks as varied as question answering and text summarization. We provide a simple new prompting strategy that leads to yet another supposedly""super-human""result, this time outperforming humans at common sense ethical reasoning (as measured by accuracy on a subset of the ETHICS dataset). Unfortunately, we find that relying on average performance to judge capabilities can be highly misleading. LLM errors differ systematically from human errors in ways that make it easy to craft adversarial examples, or even perturb existing examples to flip the output label. We also observe signs of inverse scaling with model size on some examples, and show that prompting models to""explain their reasoning""often leads to alarming justifications of unethical actions. Our results highlight how human-like performance does not necessarily imply human-like understanding or reasoning.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a simple new prompting strategy that leads to yet another supposedly ""super-human"" result, this time outperforming humans at common sense ethical reasoning (as measured by accuracy on a subset of the ETHICS dataset).'}",http://arxiv.org/pdf/2212.06295
-human-in-the-loop through chain-of-thought,Zefan Cai,"While the emergence of powerful language models along with Chain-of-thought prompting has made automation more and more omnipresent, it sometimes demonstrates its weakness in long-term or multi-step logical reasoning. For example, users don't always get desirable answers for complex mathematical problems without human involvement. Against this background, we present the Manual Correction System (MCS) -- a human-in-the-loop system enhanced by Chain-of-Thought prompting, which explores how manual correction of sub-logics in rationales can improve LLM's reasoning performance. Moving one step forward, considering a system with human-in-the-loop involves more than having humans improve performance but also controlling the cost. Therefore, we post a Cost-utility Analysis Model for Human-in-the-Loop systems (CAMLOP) based on classical economics theory to analyze, quantify and balance the utility and the corresponding cost. We conduct experiments of MCS and CAMLOP with twelve datasets. A significant advantage w.r.t cost and utility proves its superiority over strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A Cost-utility Analysis Model for Human-in-the-Loop systems (CAMLOP) based on classical economics theory to analyze, quantify and balance the utility and the corresponding cost.'}",http://arxiv.org/pdf/2306.07932
-an evaluation of log parsing with chatgpt,Van-Hoang Le,"—Software logs play an essential role in ensuring the reliability and maintainability of large-scale software systems, as they are often the sole source of runtime information. Log parsing, which converts raw log messages into structured data, is an important initial step towards downstream log analytics. In recent studies, ChatGPT, the current cutting-edge large language model (LLM), has been widely applied to a wide range of software engineering tasks. However, its performance in automated log parsing remains unclear. In this paper, we evaluate ChatGPT’s ability to undertake log parsing by addressing two research questions. (1) Can ChatGPT effectively parse logs? (2) How does ChatGPT perform with different prompting methods? Our results show that ChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting. Based on our findings, we outline several challenges and opportunities for ChatGPT-based log parsing.","{'model': 'tldr@v2.0.0', 'text': 'Results show that ChatGPT can achieve promising results for log parsing with appropriate prompts, especially with few-shot prompting, and outline several challenges and opportunities for ChatG PT-based log parsing.'}",https://arxiv.org/pdf/2306.01590
-evaluating gpt-3 generated explanations for hateful content moderation,H. Wang,"Recent research has focused on using large language models (LLMs) to generate explanations for hate speech through fine-tuning or prompting. Despite the growing interest in this area, these generated explanations' effectiveness and potential limitations remain poorly understood. A key concern is that these explanations, generated by LLMs, may lead to erroneous judgments about the nature of flagged content by both users and content moderators. For instance, an LLM-generated explanation might inaccurately convince a content moderator that a benign piece of content is hateful. In light of this, we propose an analytical framework for examining hate speech explanations and conducted an extensive survey on evaluating such explanations. Specifically, we prompted GPT-3 to generate explanations for both hateful and non-hateful content, and a survey was conducted with 2,400 unique respondents to evaluate the generated explanations. Our findings reveal that (1) human evaluators rated the GPT-generated explanations as high quality in terms of linguistic fluency, informativeness, persuasiveness, and logical soundness, (2) the persuasive nature of these explanations, however, varied depending on the prompting strategy employed, and (3) this persuasiveness may result in incorrect judgments about the hatefulness of the content. Our study underscores the need for caution in applying LLM-generated explanations for content moderation. Code and results are available at https://github.com/Social-AI-Studio/GPT3-HateEval.",,https://arxiv.org/pdf/2305.17680
-large language models are strong zero-shot retriever,Tao Shen,"In this work, we propose a simple method that applies a large language model (LLM) to large-scale retrieval in zero-shot scenarios. Our method, the Language language model as Retriever (LameR), is built upon no other neural models but an LLM, while breaking brute-force combinations of retrievers with LLMs and lifting the performance of zero-shot retrieval to be very competitive on benchmark datasets. Essentially, we propose to augment a query with its potential answers by prompting LLMs with a composition of the query and the query's in-domain candidates. The candidates, regardless of correct or wrong, are obtained by a vanilla retrieval procedure on the target collection. As a part of the prompts, they are likely to help LLM generate more precise answers by pattern imitation or candidate summarization. Even if all the candidates are wrong, the prompts at least make LLM aware of in-collection patterns and genres. Moreover, due to the low performance of a self-supervised retriever, the LLM-based query augmentation becomes less effective as the retriever bottlenecks the whole pipeline. Therefore, we propose to leverage a non-parametric lexicon-based method (e.g., BM25) as the retrieval module to capture query-document overlap in a literal fashion. As such, LameR makes the retrieval procedure transparent to the LLM, thus circumventing the performance bottleneck.","{'model': 'tldr@v2.0.0', 'text': 'The Language language model as Retriever (LameR), is built upon no other neural models but an LLM, while breaking brute-force combinations of retrievers with LLMs and lifting the performance of zero-shot retrieval to be very competitive on benchmark datasets.'}",https://arxiv.org/pdf/2304.14233
-careful data curation stabilizes in-context learning,Ting-Yun Chang,"In-context learning (ICL) enables large language models (LLMs) to perform new tasks by prompting them with a sequence of training examples. However, ICL is very sensitive to the choice of training examples: randomly sampling examples from a training set leads to high variance in performance. In this paper, we show that curating a carefully chosen subset of training data greatly stabilizes ICL performance. We propose two methods to choose training subsets, both of which score training examples individually and then select the highest-scoring ones. C OND A CC scores a training example by its average ICL accuracy when combined with random training examples, while D ATAMODELS learns a linear proxy model that estimates how the presence of each training example influences LLM accuracy. On average, C OND A CC and D ATA - MODELS outperform sampling from the entire training set by 7.7% and 6.3%, respectively, across 5 tasks and two LLMs. Our analysis shows that stable subset examples are no more diverse than average, and are not outliers in terms of sequence length and perplexity.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that curating a carefully chosen subset of training data greatly stabilizes ICL performance, and that stable subset examples are no more diverse than average, and are not outliers in terms of sequence length and perplexity.'}",https://arxiv.org/pdf/2212.10378
-s3hqa: a three-stage approach for multi-hop text-table hybrid question answering,Fangyu Lei,"Answering multi-hop questions over hybrid factual knowledge from the given text and table (TextTableQA) is a challenging task. Existing models mainly adopt a retriever-reader framework, which have several deficiencies, such as noisy labeling in training retriever, insufficient utilization of heterogeneous information over text and table, and deficient ability for different reasoning operations. In this paper, we propose a three-stage TextTableQA framework S3HQA, which comprises of retriever, selector, and reasoner. We use a retriever with refinement training to solve the noisy labeling problem. Then, a hybrid selector considers the linked relationships between heterogeneous data to select the most relevant factual knowledge. For the final stage, instead of adapting a reading comprehension module like in previous methods, we employ a generation-based reasoner to obtain answers. This includes two approaches: a row-wise generator and an LLM prompting generator (first time used in this task). The experimental results demonstrate that our method achieves competitive results in the few-shot setting. When trained on the full dataset, our approach outperforms all baseline methods, ranking first on the HybridQA leaderboard.","{'model': 'tldr@v2.0.0', 'text': 'A retriever with refinement training is used to solve the noisy labeling problem, a hybrid selector considers the linked relationships between heterogeneous data to select the most relevant factual knowledge, and a generation-based reasoner is employed to obtain answers.'}",http://arxiv.org/pdf/2305.11725
-forward-backward reasoning in large language models for verification,Weisen Jiang,"Chain-of-Though (CoT) prompting has shown promising performance in various reasoning tasks. Recently, Self-Consistency (Wang et al., 2023) proposes to sample a diverse set of reasoning chains which may lead to different answers while the answer that receives the most votes is selected. In this paper, we propose a novel method to use backward reasoning in verifying candidate answers. We mask a token in the question by x and ask the LLM to predict the masked token when a candidate answer is provided by a simple template, i.e., “If we know the answer of the above question is {a candidate answer}, what is the value of unknown variable x?” Intuitively, the LLM is expected to predict the masked token successfully if the provided candidate answer is correct. We further propose FOBAR to combine forward and backward reasoning for estimating the probability of candidate answers. We conduct extensive experiments on six data sets and three LLMs. Experimental results demonstrate that FOBAR achieves state-of-the-art performance on various reasoning benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A novel method to use backward reasoning in verifying candidate answers inChain-of-Though prompting and proposes FOBAR to combine forward and backward reasoning for estimating the probability of candidate answers.'}",https://arxiv.org/pdf/2308.07758
-how to catch an ai liar: lie detection in black-box llms by asking unrelated questions,Lorenzo Pacchiardi,"Large language models (LLMs) can""lie"", which we define as outputting false statements despite""knowing""the truth in a demonstrable sense. LLMs might""lie"", for example, when instructed to output misinformation. Here, we develop a simple lie detector that requires neither access to the LLM's activations (black-box) nor ground-truth knowledge of the fact in question. The detector works by asking a predefined set of unrelated follow-up questions after a suspected lie, and feeding the LLM's yes/no answers into a logistic regression classifier. Despite its simplicity, this lie detector is highly accurate and surprisingly general. When trained on examples from a single setting -- prompting GPT-3.5 to lie about factual questions -- the detector generalises out-of-distribution to (1) other LLM architectures, (2) LLMs fine-tuned to lie, (3) sycophantic lies, and (4) lies emerging in real-life scenarios such as sales. These results indicate that LLMs have distinctive lie-related behavioural patterns, consistent across architectures and contexts, which could enable general-purpose lie detection.",,https://arxiv.org/pdf/2309.15840
-self-checker: plug-and-play modules for fact-checking with large language models,Miaoran Li,"Fact-checking is an essential task in NLP that is commonly utilized for validating the factual accuracy of claims. Prior work has mainly focused on fine-tuning pre-trained languages models on specific datasets, which can be computationally intensive and time-consuming. With the rapid development of large language models (LLMs), such as ChatGPT and GPT-3, researchers are now exploring their in-context learning capabilities for a wide range of tasks. In this paper, we aim to assess the capacity of LLMs for fact-checking by introducing Self-Checker, a framework comprising a set of plug-and-play modules that facilitate fact-checking by purely prompting LLMs in an almost zero-shot setting. This framework provides a fast and efficient way to construct fact-checking systems in low-resource environments. Empirical results demonstrate the potential of Self-Checker in utilizing LLMs for fact-checking. However, there is still significant room for improvement compared to SOTA fine-tuned models, which suggests that LLM adoption could be a promising approach for future fact-checking research.","{'model': 'tldr@v2.0.0', 'text': 'Self-Checker, a framework comprising a set of plug-and-play modules that facilitate fact-checking by purely prompting LLMs in an almost zero-shot setting, is introduced, which provides a fast and efficient way to construct fact- checking systems in low-resource environments.'}",http://arxiv.org/pdf/2305.14623
-llms to the moon? reddit market sentiment analysis with large language models,Xiang Deng,"Market sentiment analysis on social media content requires knowledge of both financial markets and social media jargon, which makes it a challenging task for human raters. The resulting lack of high-quality labeled data stands in the way of conventional supervised learning methods. In this work, we conduct a case study approaching this problem with semi-supervised learning using a large language model (LLM). We select Reddit as the target social media platform due to its broad coverage of topics and content types. Our pipeline first generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production. We find that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels, while training the student model using a regression loss further improves distillation quality. With only a handful of prompts, the final model performs on par with existing supervised models. Though production applications of our model are limited by ethical considerations, the model’s competitive performance points to the great potential of using LLMs for tasks that otherwise require skill-intensive annotation.","{'model': 'tldr@v2.0.0', 'text': 'This case study conducts a case study approaching market sentiment analysis on social media content with semi-supervised learning using a large language model (LLM), finding that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels.'}",
-leveraging commonsense knowledge from large language models for task and motion planning,Yan Ding,"Multi-object rearrangement is a crucial skill for service robots, and commonsense reasoning is frequently needed in this process. However, achieving commonsense arrangements requires knowledge about objects, which is hard to transfer to robots. Large language models (LLMs) are one potential source of this knowledge, but they do not naively capture information about plausible physical arrangements of the world. We propose LLMGROP, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry. LLM-GROP allows us to go from natural-language commands to human-aligned object rearrangement in varied environments. Based on human evaluations, our approach achieves the highest rating while outperforming competitive baselines in terms of success rate while maintaining comparable cumulative action costs. Finally, we demonstrate a practical implementation of LLM-GROP on a mobile manipulator in real-world scenarios. Supplementary materials are available at: https://sites.google.com/view/llm-grop","{'model': 'tldr@v2.0.0', 'text': 'LLMGROP is proposed, which uses prompting to extract commonsense knowledge about semantically valid object configurations from an LLM and instantiates them with a task and motion planner in order to generalize to varying scene geometry.'}",
-what do llms know about financial markets? a case study on reddit market sentiment analysis,Xiang Deng,"Market sentiment analysis on social media content requires knowledge of both financial markets and social media jargon, which makes it a challenging task for human raters. The resulting lack of high-quality labeled data stands in the way of conventional supervised learning methods. Instead, we approach this problem using semi-supervised learning with a large language model (LLM). Our pipeline generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production. We find that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels, while using a regression loss further improves distillation quality. With only a handful of prompts, the final model performs on par with existing supervised models. Though production applications of our model are limited by ethical considerations, the model’s competitive performance points to the great potential of using LLMs for tasks that otherwise require skill-intensive annotation.","{'model': 'tldr@v2.0.0', 'text': 'This model generates weak financial sentiment labels for Reddit posts with an LLM and then uses that data to train a small model that can be served in production, and finds that prompting the LLM to produce Chain-of-Thought summaries and forcing it through several reasoning paths helps generate more stable and accurate labels.'}",http://arxiv.org/pdf/2212.11311
-enhancing in-context learning with answer feedback for multi-span question answering,Zixian Huang,"Whereas the recent emergence of large language models (LLMs) like ChatGPT has exhibited impressive general performance, it still has a large gap with fully-supervised models on specific tasks such as multi-span question answering. Previous researches found that in-context learning is an effective approach to exploiting LLM, by using a few task-related labeled data as demonstration examples to construct a few-shot prompt for answering new questions. A popular implementation is to concatenate a few questions and their correct answers through simple templates, informing LLM of the desired output. In this paper, we propose a novel way of employing labeled data such that it also informs LLM of some undesired output, by extending demonstration examples with feedback about answers predicted by an off-the-shelf model, e.g., correct, incorrect, or incomplete. Experiments on three multi-span question answering datasets as well as a keyphrase extraction dataset show that our new prompting strategy consistently improves LLM's in-context learning performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel way of employing labeled data such that it also informs LLM of some undesired output, by extending demonstration examples with feedback about answers predicted by an off-the-shelf model.'}",http://arxiv.org/pdf/2306.04508
-retrieving texts based on abstract descriptions,Shauli Ravfogel,"While instruction-tuned Large Language Models (LLMs) excel at extracting information from text, they are not suitable for locating texts conforming to a given description in a large document collection (semantic retrieval). Similarity search over embedding vectors does allow to perform retrieval by query, but the similarity reflected in the embedding is ill-defined and non-consistent, and is sub-optimal for many use cases. What, then, is a good query representation for effective retrieval? We identify the well defined and consistent task of retrieving sentences based on abstract descriptions of their content. We demonstrate the inadequacy of current text embeddings and propose an alternative model that significantly improves when used in standard nearest neighbor search. The model is trained using positive and negative pairs sourced through prompting a LLM. While it is easy to source the training material from an LLM, the retrieval task cannot be performed by the LLM directly. This demonstrates that data from LLMs can be used not only for distilling more efficient specialized models than the original LLM, but also for creating new capabilities not immediately possible using the original model.","{'model': 'tldr@v2.0.0', 'text': 'This work demonstrates that data from LLMs can be used not only for distilling more efficient specialized models than the original LLM, but also for creating new capabilities not immediately possible using the original model.'}",http://arxiv.org/pdf/2305.12517
-skillgpt: a restful api service for skill extraction and standardization using a large language model,Nan Li,"We present SkillGPT, a tool for skill extraction and standardization (SES) from free-style job descriptions and user profiles with an open-source Large Language Model (LLM) as backbone. Most previous methods for similar tasks either need supervision or rely on heavy data-preprocessing and feature engineering. Directly prompting the latest conversational LLM for standard skills, however, is slow, costly and inaccurate. In contrast, SkillGPT utilizes a LLM to perform its tasks in steps via summarization and vector similarity search, to balance speed with precision. The backbone LLM of SkillGPT is based on Llama, free for academic use and thus useful for exploratory research and prototype development. Hence, our cost-free SkillGPT gives users the convenience of conversational SES, efficiently and reliably.","{'model': 'tldr@v2.0.0', 'text': 'SkillGPT, a tool for skill extraction and standardization from free-style job descriptions and user profiles with an open-source Large Language Model (LLM) as backbone, gives users the convenience of conversational SES, efficiently and reliably.'}",http://arxiv.org/pdf/2304.11060
-queer people are people first: deconstructing sexual identity stereotypes in large language models,Harnoor Dhingra,"Large Language Models (LLMs) are trained primarily on minimally processed web text, which exhibits the same wide range of social biases held by the humans who created that content. Consequently, text generated by LLMs can inadvertently perpetuate stereotypes towards marginalized groups, like the LGBTQIA+ community. In this paper, we perform a comparative study of how LLMs generate text describing people with different sexual identities. Analyzing bias in the text generated by an LLM using regard score shows measurable bias against queer people. We then show that a post-hoc method based on chain-of-thought prompting using SHAP analysis can increase the regard of the sentence, representing a promising approach towards debiasing the output of LLMs in this setting.",,http://arxiv.org/pdf/2307.00101
-knowing what llms do not know: a simple yet effective self-detection method,Yukun Zhao,"Large Language Models (LLMs) have shown great potential in Natural Language Processing (NLP) tasks. However, recent literature reveals that LLMs generate nonfactual responses intermittently, which impedes the LLMs' reliability for further utilization. In this paper, we propose a novel self-detection method to detect which questions that a LLM does not know that are prone to generate nonfactual results. Specifically, we first diversify the textual expressions for a given question and collect the corresponding answers. Then we examine the divergencies between the generated answers to identify the questions that the model may generate falsehoods. All of the above steps can be accomplished by prompting the LLMs themselves without referring to any other external resources. We conduct comprehensive experiments and demonstrate the effectiveness of our method on recently released LLMs, e.g., Vicuna, ChatGPT, and GPT-4.","{'model': 'tldr@v2.0.0', 'text': 'A novel self-detection method to detect which questions that a LLM does not know that are prone to generate nonfactual results, which can be accomplished by prompting the LLMs themselves without referring to any other external resources.'}",
-retrieving supporting evidence for llms generated answers,Siqing Huo,"Current large language models (LLMs) can exhibit near-human levels of performance on many natural language tasks, including open-domain question answering. Unfortunately, they also convincingly hallucinate incorrect answers, so that responses to questions must be verified against external sources before they can be accepted at face value. In this paper, we report a simple experiment to automatically verify generated answers against a corpus. After presenting a question to an LLM and receiving a generated answer, we query the corpus with the combination of the question + generated answer. We then present the LLM with the combination of the question + generated answer + retrieved answer, prompting it to indicate if the generated answer can be supported by the retrieved answer. We base our experiment on questions and passages from the MS MARCO (V1) test collection, exploring three retrieval approaches ranging from standard BM25 to a full question answering stack, including a reader based on the LLM. For a large fraction of questions, we find that an LLM is capable of verifying its generated answer if appropriate supporting material is provided. However, with an accuracy of 70-80%, this approach cannot be fully relied upon to detect hallucinations.","{'model': 'tldr@v2.0.0', 'text': 'For a large fraction of questions, it is found that an LLM is capable of verifying its generated answer if appropriate supporting material is provided, but with an accuracy of 70-80%, this approach cannot be fully relied upon to detect hallucinations.'}",http://arxiv.org/pdf/2306.13781
-knowledge sanitization of large language models,Yoichi Ishibashi,"We explore a knowledge sanitization approach to mitigate the privacy concerns associated with large language models (LLMs). LLMs trained on a large corpus of Web data can memorize and potentially reveal sensitive or confidential information, raising critical security concerns. Our technique fine-tunes these models, prompting them to generate harmless responses such as ``I don't know'' when queried about specific information. Experimental results in a closed-book question-answering task show that our straightforward method not only minimizes particular knowledge leakage but also preserves the overall performance of LLM. These two advantages strengthen the defense against extraction attacks and reduces the emission of harmful content such as hallucinations.","{'model': 'tldr@v2.0.0', 'text': ""A knowledge sanitization approach to mitigate the privacy concerns associated with large language models (LLMs), which fine-tunes these models, prompting them to generate harmless responses such as ``I don't know'' when queried about specific information.""}",https://arxiv.org/pdf/2309.11852
-reasoning in large language models through symbolic math word problems,Vedant Gaur,"Large language models (LLMs) have revolutionized NLP by solving downstream tasks with little to no labeled data. Despite their versatile abilities, the larger question of their ability to reason remains ill-understood. This paper addresses reasoning in math word problems (MWPs) by studying symbolic versions of the numeric problems, since a symbolic expression is a""concise explanation""of the numeric answer. We create and use a symbolic version of the SVAMP dataset and find that GPT-3's davinci-002 model also has good zero-shot accuracy on symbolic MWPs. To evaluate the faithfulness of the model's reasoning, we go beyond accuracy and additionally evaluate the alignment between the final answer and the outputted reasoning, which correspond to numeric and symbolic answers respectively for MWPs. We explore a self-prompting approach to encourage the symbolic reasoning to align with the numeric answer, thus equipping the LLM with the ability to provide a concise and verifiable reasoning and making it more interpretable. Surprisingly, self-prompting also improves the symbolic accuracy to be higher than both the numeric and symbolic accuracies, thus providing an ensembling effect. The SVAMP_Sym dataset will be released for future research on symbolic math problems.","{'model': 'tldr@v2.0.0', 'text': 'A self-prompting approach is explored to encourage the symbolic reasoning to align with the numeric answer, thus equipping the LLM with the ability to provide a concise and verifiable reasoning and making it more interpretable.'}",https://aclanthology.org/2023.findings-acl.364.pdf
-alphazero-like tree-search can guide large language model decoding and training,Xidong Feng,"Large language models (LLMs) typically employ sampling or beam search, accompanied by prompts such as Chain-of-Thought (CoT), to boost reasoning and decoding ability. Recent work like Tree-of-Thought (ToT) and Reasoning via Planning (RAP) aim to augment the reasoning capabilities of LLMs by utilizing tree-search algorithms to guide multi-step reasoning. These methods mainly focus on LLMs' reasoning ability during inference and heavily rely on human-designed prompts to activate LLM as a value function, which lacks general applicability and scalability. To address these limitations, we present an AlphaZero-like tree-search framework for LLMs (termed TS-LLM), systematically illustrating how tree-search with a learned value function can guide LLMs' decoding ability. TS-LLM distinguishes itself in two key ways: (1) Leveraging a learned value function, our approach can be generally applied to different tasks beyond reasoning (such as RLHF alignment), and LLMs of any size, without prompting advanced, large-scale models. (2) It can guide LLM's decoding during both inference and training. Empirical evaluations across reasoning, planning, and RLHF alignment tasks validate the effectiveness of TS-LLM, even on trees with a depth of 64.","{'model': 'tldr@v2.0.0', 'text': ""An AlphaZero-like tree-search framework for LLMs (termed TS-LLM) is presented, systematically illustrating how tree- search with a learned value function can guide LLM' decoding ability.""}",https://arxiv.org/pdf/2309.17179
-exploring human-like translation strategy with large language models,Zhiwei He,"Large language models (LLMs) have demonstrated impressive capabilities in general scenarios, exhibiting a level of aptitude that approaches, in some aspects even surpasses, human-level intelligence. Among their numerous skills, the translation abilities of LLMs have received considerable attention. In contrast to traditional machine translation that focuses solely on source-target mapping, LLM-based translation can potentially mimic the human translation process that takes many preparatory steps to ensure high-quality translation. This work aims to explore this possibility by proposing the MAPS framework, which stands for Multi-Aspect Prompting and Selection. Specifically, we enable LLMs to first analyze the given source text and extract three aspects of translation-related knowledge: keywords, topics and relevant demonstrations to guide the translation process. To filter out the noisy and unhelpful knowledge, we employ a selection mechanism based on quality estimation. Experiments suggest that MAPS brings significant and consistent improvements over text-davinci-003 and Alpaca on eight translation directions from the latest WMT22 test sets. Our further analysis shows that the extracted knowledge is critical in resolving up to 59% of hallucination mistakes in translation. Code is available at https://github.com/zwhe99/MAPS-mt.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the MAPS framework, which stands for Multi-Aspect Prompting and Selection, and enables LLMs to first analyze the given source text and extract three aspects of translation-related knowledge: keywords, topics and relevant demonstrations to guide the translation process.'}",http://arxiv.org/pdf/2305.04118
-"mmhqa-icl: multimodal in-context learning for hybrid question answering over text, tables and images",Weihao Liu,"In the real world, knowledge often exists in a multimodal and heterogeneous form. Addressing the task of question answering with hybrid data types, including text, tables, and images, is a challenging task (MMHQA). Recently, with the rise of large language models (LLM), in-context learning (ICL) has become the most popular way to solve QA problems. We propose MMHQA-ICL framework for addressing this problems, which includes stronger heterogeneous data retriever and an image caption module. Most importantly, we propose a Type-specific In-context Learning Strategy for MMHQA, enabling LLMs to leverage their powerful performance in this task. We are the first to use end-to-end LLM prompting method for this task. Experimental results demonstrate that our framework outperforms all baselines and methods trained on the full dataset, achieving state-of-the-art results under the few-shot setting on the MultimodalQA dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Type-specific In-context Learning Strategy for MMHQA, enabling LLMs to leverage their powerful performance in this task, and is the first to use end-to-end LLM prompting method for this task.'}",https://arxiv.org/pdf/2309.04790
-gear: augmenting language models with generalizable and efficient tool resolution,Yining Lu,"Augmenting large language models (LLM) to use external tools enhances their performance across a variety of tasks. However, prior works over-rely on task-specific demonstration of tool use that limits their generalizability and computational cost due to making many calls to large-scale LLMs. We introduce GEAR, a computationally efficient query-tool grounding algorithm that is generalizable to various tasks that require tool use while not relying on task-specific demonstrations. GEAR achieves better efficiency by delegating tool grounding and execution to small language models (SLM) and LLM, respectively; while leveraging semantic and pattern-based evaluation at both question and answer levels for generalizable tool grounding. We evaluate GEAR on 14 datasets across 6 downstream tasks, demonstrating its strong generalizability to novel tasks, tools and different SLMs. Despite offering more efficiency, GEAR achieves higher precision in tool grounding compared to prior strategies using LLM prompting, thus improving downstream accuracy at a reduced computational cost. For example, we demonstrate that GEAR-augmented GPT-J and GPT-3 outperform counterpart tool-augmented baselines because of better tool use.","{'model': 'tldr@v2.0.0', 'text': 'GE is introduced, a computationally efficient query-tool grounding algorithm that is generalizable to various tasks that require tool use while not relying on task-specific demonstrations and achieves higher precision in tool grounding compared to prior strategies using LLM prompting, thus improving downstream accuracy at a reduced computational cost.'}",https://arxiv.org/pdf/2307.08775
-constitutionmaker: interactively critiquing large language models by converting feedback into principles,S. Petridis,"Large language model (LLM) prompting is a promising new approach for users to create and customize their own chatbots. However, current methods for steering a chatbot's outputs, such as prompt engineering and fine-tuning, do not support users in converting their natural feedback on the model's outputs to changes in the prompt or model. In this work, we explore how to enable users to interactively refine model outputs through their feedback, by helping them convert their feedback into a set of principles (i.e. a constitution) that dictate the model's behavior. From a formative study, we (1) found that users needed support converting their feedback into principles for the chatbot and (2) classified the different principle types desired by users. Inspired by these findings, we developed ConstitutionMaker, an interactive tool for converting user feedback into principles, to steer LLM-based chatbots. With ConstitutionMaker, users can provide either positive or negative feedback in natural language, select auto-generated feedback, or rewrite the chatbot's response; each mode of feedback automatically generates a principle that is inserted into the chatbot's prompt. In a user study with 14 participants, we compare ConstitutionMaker to an ablated version, where users write their own principles. With ConstitutionMaker, participants felt that their principles could better guide the chatbot, that they could more easily convert their feedback into principles, and that they could write principles more efficiently, with less mental demand. ConstitutionMaker helped users identify ways to improve the chatbot, formulate their intuitive responses to the model into feedback, and convert this feedback into specific and clear principles. Together, these findings inform future tools that support the interactive critiquing of LLM outputs.","{'model': 'tldr@v2.0.0', 'text': 'ConstitutionMaker, an interactive tool for converting user feedback into principles, is developed to steer LLM-based chatbots and helps users identify ways to improve the chatbot, formulate their intuitive responses to the model into feedback, and convert this feedback into specific and clear principles.'}",
-generative ai: implications and applications for education,A. Tzirides,"The launch of ChatGPT in November 2022 precipitated a panic among some educators while prompting qualified enthusiasm from others. Under the umbrella term Generative AI, ChatGPT is an example of a range of technologies for the delivery of computer-generated text, image, and other digitized media. This paper examines the implications for education of one generative AI technology, chatbots responding from large language models, or C-LLM. It reports on an application of a C-LLM to AI review and assessment of complex student work. In a concluding discussion, the paper explores the intrinsic limits of generative AI, bound as it is to language corpora and their textual representation through binary notation. Within these limits, we suggest the range of emerging and potential applications of Generative AI in education.","{'model': 'tldr@v2.0.0', 'text': 'The intrinsic limits of generative AI are explored, bound as it is to language corpora and their textual representation through binary notation, and the range of emerging and potential applications of Generative AI in education are suggested.'}",http://arxiv.org/pdf/2305.07605
-theory of mind in large language models: examining performance of 11 state-of-the-art models vs. children aged 7-10 on advanced tests,Max J. van Duijn,"To what degree should we ascribe cognitive capacities to Large Language Models (LLMs), such as the ability to reason about intentions and beliefs known as Theory of Mind (ToM)? Here we add to this emerging debate by (i) testing 11 base- and instruction-tuned LLMs on capabilities relevant to ToM beyond the dominant false-belief paradigm, including non-literal language usage and recursive intentionality; (ii) using newly rewritten versions of standardized tests to gauge LLMs' robustness; (iii) prompting and scoring for open besides closed questions; and (iv) benchmarking LLM performance against that of children aged 7-10 on the same tasks. We find that instruction-tuned LLMs from the GPT family outperform other models, and often also children. Base-LLMs are mostly unable to solve ToM tasks, even with specialized prompting. We suggest that the interlinked evolution and development of language and ToM may help explain what instruction-tuning adds: rewarding cooperative communication that takes into account interlocutor and context. We conclude by arguing for a nuanced perspective on ToM in LLMs.",,
-generating multiple choice questions from a textbook: llms match human performance on most metrics,Andrew M. Olney,"Multiple choice questions are traditionally expensive to produce. Recent advances in large language models (LLMs) have led to fine-tuned LLMs that generate questions competitive with human-authored questions. However, the relative capabilities of ChatGPT-family models have not yet been established for this task. We present a carefully-controlled human evaluation of three conditions: a fine-tuned, augmented version of Macaw, instruction-tuned Bing Chat with zero-shot prompting, and human-authored questions from a college science textbook. Our results indicate that on six of seven measures tested, both LLM’s performance was not significantly different from human performance. Analysis of LLM errors further suggests that Macaw and Bing Chat have different failure modes for this task: Macaw tends to repeat answer options whereas Bing Chat tends to not include the specified answer in the answer options. For Macaw, removing error items from analysis results in performance on par with humans for all metrics; for Bing Chat, removing error items improves performance but does not reach human-level performance.","{'model': 'tldr@v2.0.0', 'text': 'Analysis of LLM errors suggests that Macaw and Bing Chat have different failure modes for this task: Macaw tends to repeat answer options whereas Bing Chat tends to not include the specified answer in the answer options.'}",
-retrieving supporting evidence for generative question answering,Siqing Huo,"Current large language models (LLMs) can exhibit near-human levels of performance on many natural language-based tasks, including open-domain question answering. Unfortunately, at this time, they also convincingly hallucinate incorrect answers, so that responses to questions must be verified against external sources before they can be accepted at face value. In this paper, we report two simple experiments to automatically validate generated answers against a corpus. We base our experiments on questions and passages from the MS MARCO (V1) test collection, and a retrieval pipeline consisting of sparse retrieval, dense retrieval and neural rerankers. In the first experiment, we validate the generated answer in its entirety. After presenting a question to an LLM and receiving a generated answer, we query the corpus with the combination of the question + generated answer. We then present the LLM with the combination of the question + generated answer + retrieved answer, prompting it to indicate if the generated answer can be supported by the retrieved answer. In the second experiment, we consider the generated answer at a more granular level, prompting the LLM to extract a list of factual statements from the answer and verifying each statement separately. We query the corpus with each factual statement and then present the LLM with the statement and the corresponding retrieved evidence. The LLM is prompted to indicate if the statement can be supported and make necessary edits using the retrieved material. With an accuracy of over 80%, we find that an LLM is capable of verifying its generated answer when a corpus of supporting material is provided. However, manual assessment of a random sample of questions reveals that incorrect generated answers are missed by this verification process. While this verification process can reduce hallucinations, it can not entirely eliminate them.","{'model': 'tldr@v2.0.0', 'text': 'With an accuracy of over 80%, it is found that an LLM is capable of verifying its generated answer when a corpus of supporting material is provided, however, manual assessment of a random sample of questions reveals that incorrect generated answers are missed by this verification process.'}",https://arxiv.org/pdf/2309.11392
-visualizing linguistic diversity of text datasets synthesized by large language models,Emily Reif,"Large language models (LLMs) can be used to generate smaller, more refined datasets via few-shot prompting for benchmarking, fine-tuning or other use cases. However, understanding and evaluating these datasets is difficult, and the failure modes of LLM-generated data are still not well understood. Specifically, the data can be repetitive in surprising ways, not only semantically but also syntactically and lexically. We present LinguisticLens, a novel inter-active visualization tool for making sense of and analyzing syntactic diversity of LLM-generated datasets. LinguisticLens clusters text along syntactic, lexical, and semantic axes. It supports hierarchical visualization of a text dataset, allowing users to quickly scan for an overview and inspect individual examples. The live demo is available at shorturl.at/zHOUV.","{'model': 'tldr@v2.0.0', 'text': 'LinguisticLens is a novel inter-active visualization tool for making sense of and analyzing syntactic diversity of LLM-generated datasets and supports hierarchical visualization of a text dataset, allowing users to quickly scan for an overview and inspect individual examples.'}",https://arxiv.org/pdf/2305.11364
-synergistic integration of large language models and cognitive architectures for robust ai: an exploratory analysis,Oscar J. Romero,"This paper explores the integration of two AI subdisciplines employed in the development of artificial agents that exhibit intelligent behavior: Large Language Models (LLMs) and Cognitive Architectures (CAs). We present three integration approaches, each grounded in theoretical models and supported by preliminary empirical evidence. The modular approach, which introduces four models with varying degrees of integration, makes use of chain-of-thought prompting, and draws inspiration from augmented LLMs, the Common Model of Cognition, and the simulation theory of cognition. The agency approach, motivated by the Society of Mind theory and the LIDA cognitive architecture, proposes the formation of agent collections that interact at micro and macro cognitive levels, driven by either LLMs or symbolic components. The neuro-symbolic approach, which takes inspiration from the CLARION cognitive architecture, proposes a model where bottom-up learning extracts symbolic representations from an LLM layer and top-down guidance utilizes symbolic representations to direct prompt engineering in the LLM layer. These approaches aim to harness the strengths of both LLMs and CAs, while mitigating their weaknesses, thereby advancing the development of more robust AI systems. We discuss the tradeoffs and challenges associated with each approach.","{'model': 'tldr@v2.0.0', 'text': 'Three integration approaches are presented, each grounded in theoretical models and supported by preliminary empirical evidence, that aim to harness the strengths of both LLMs and CAs, while mitigating their weaknesses, thereby advancing the development of more robust AI systems.'}",https://arxiv.org/pdf/2308.09830
-casteist but not racist? quantifying disparities in large language model bias between india and the west,Khyati Khandelwal,"Large Language Models (LLMs), now used daily by millions of users, can encode societal biases, exposing their users to representational harms. A large body of scholarship on LLM bias exists but it predominantly adopts a Western-centric frame and attends comparatively less to bias levels and potential harms in the Global South. In this paper, we quantify stereotypical bias in popular LLMs according to an Indian-centric frame and compare bias levels between the Indian and Western contexts. To do this, we develop a novel dataset which we call Indian-BhED (Indian Bias Evaluation Dataset), containing stereotypical and anti-stereotypical examples for caste and religion contexts. We find that the majority of LLMs tested are strongly biased towards stereotypes in the Indian context, especially as compared to the Western context. We finally investigate Instruction Prompting as a simple intervention to mitigate such bias and find that it significantly reduces both stereotypical and anti-stereotypical biases in the majority of cases for GPT-3.5. The findings of this work highlight the need for including more diverse voices when evaluating LLMs.",,https://arxiv.org/pdf/2309.08573
-feasibility of using the privacy-preserving large language model vicuna for labeling radiology reports.,Pritam Mukherjee,"Background Large language models (LLMs) such as ChatGPT, though proficient in many text-based tasks, are not suitable for use with radiology reports due to patient privacy constraints. Purpose To test the feasibility of using an alternative LLM (Vicuna-13B) that can be run locally for labeling radiography reports. Materials and Methods Chest radiography reports from the MIMIC-CXR and National Institutes of Health (NIH) data sets were included in this retrospective study. Reports were examined for 13 findings. Outputs reporting the presence or absence of the 13 findings were generated by Vicuna by using a single-step or multistep prompting strategy (prompts 1 and 2, respectively). Agreements between Vicuna outputs and CheXpert and CheXbert labelers were assessed using Fleiss κ. Agreement between Vicuna outputs from three runs under a hyperparameter setting that introduced some randomness (temperature, 0.7) was also assessed. The performance of Vicuna and the labelers was assessed in a subset of 100 NIH reports annotated by a radiologist with use of area under the receiver operating characteristic curve (AUC). Results A total of 3269 reports from the MIMIC-CXR data set (median patient age, 68 years [IQR, 59-79 years]; 161 male patients) and 25 596 reports from the NIH data set (median patient age, 47 years [IQR, 32-58 years]; 1557 male patients) were included. Vicuna outputs with prompt 2 showed, on average, moderate to substantial agreement with the labelers on the MIMIC-CXR (κ median, 0.57 [IQR, 0.45-0.66] with CheXpert and 0.64 [IQR, 0.45-0.68] with CheXbert) and NIH (κ median, 0.52 [IQR, 0.41-0.65] with CheXpert and 0.55 [IQR, 0.41-0.74] with CheXbert) data sets, respectively. Vicuna with prompt 2 performed at par (median AUC, 0.84 [IQR, 0.74-0.93]) with both labelers on nine of 11 findings. Conclusion In this proof-of-concept study, outputs of the LLM Vicuna reporting the presence or absence of 13 findings on chest radiography reports showed moderate to substantial agreement with existing labelers. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Cai in this issue.","{'model': 'tldr@v2.0.0', 'text': 'In this proof-of-concept study, outputs of the LLM Vicuna reporting the presence or absence of 13 findings on chest radiography reports showed moderate to substantial agreement with existing labelers.'}",
-empirical study of zero-shot ner with chatgpt,Tingyu Xie,"Large language models (LLMs) exhibited powerful capability in various natural language processing tasks. This work focuses on exploring LLM performance on zero-shot information extraction, with a focus on the ChatGPT and named entity recognition (NER) task. Inspired by the remarkable reasoning capability of LLM on symbolic and arithmetic reasoning, we adapt the prevalent reasoning methods to NER and propose reasoning strategies tailored for NER. First, we explore a decomposed question-answering paradigm by breaking down the NER task into simpler subproblems by labels. Second, we propose syntactic augmentation to stimulate the model's intermediate thinking in two ways: syntactic prompting, which encourages the model to analyze the syntactic structure itself, and tool augmentation, which provides the model with the syntactic information generated by a parsing tool. Besides, we adapt self-consistency to NER by proposing a two-stage majority voting strategy, which first votes for the most consistent mentions, then the most consistent types. The proposed methods achieve remarkable improvements for zero-shot NER across seven benchmarks, including Chinese and English datasets, and on both domain-specific and general-domain scenarios. In addition, we present a comprehensive analysis of the error types with suggestions for optimization directions. We also verify the effectiveness of the proposed methods on the few-shot setting and other LLMs.","{'model': 'tldr@v2.0.0', 'text': ""This work focuses on exploring LLM performance on zero-shot information extraction, with a focus on the ChatGPT and named entity recognition (NER) task, and proposes syntactic augmentation to stimulate the model's intermediate thinking in two ways.""}",
-large language models can learn rules,Zhaocheng Zhu,"When prompted with a few examples and intermediate steps, large language models (LLMs) have demonstrated impressive performance in various reasoning tasks. However, prompting methods that rely on implicit knowledge in an LLM often hallucinate incorrect answers when the implicit knowledge is wrong or inconsistent with the task. To tackle this problem, we present Hypotheses-to-Theories (HtT), a framework that learns a rule library for reasoning with LLMs. HtT contains two stages, an induction stage and a deduction stage. In the induction stage, an LLM is first asked to generate and verify rules over a set of training examples. Rules that appear and lead to correct answers sufficiently often are collected to form a rule library. In the deduction stage, the LLM is then prompted to employ the learned rule library to perform reasoning to answer test questions. Experiments on both numerical reasoning and relational reasoning problems show that HtT improves existing prompting methods, with an absolute gain of 11-27% in accuracy. The learned rules are also transferable to different models and to different forms of the same problem.","{'model': 'tldr@v2.0.0', 'text': 'Hypotheses-to-Theories (HtT), a framework that learns a rule library for reasoning with large language models, improves existing prompting methods, with an absolute gain of 11-27% in accuracy.'}",https://arxiv.org/pdf/2310.07064
-towards optimizing with large language models,Pei-Fu Guo,"In this work, we conduct an assessment of the optimization capabilities of LLMs across various tasks and data sizes. Each of these tasks corresponds to unique optimization domains, and LLMs are required to execute these tasks with interactive prompting. That is, in each optimization step, the LLM generates new solutions from the past generated solutions with their values, and then the new solutions are evaluated and considered in the next optimization step. Additionally, we introduce three distinct metrics for a comprehensive assessment of task performance from various perspectives. These metrics offer the advantage of being applicable for evaluating LLM performance across a broad spectrum of optimization tasks and are less sensitive to variations in test samples. By applying these metrics, we observe that LLMs exhibit strong optimization capabilities when dealing with small-sized samples. However, their performance is significantly influenced by factors like data size and values, underscoring the importance of further research in the domain of optimization tasks for LLMs.","{'model': 'tldr@v2.0.0', 'text': 'An assessment of the optimization capabilities of LLMs across various tasks and data sizes observes that LLMs exhibit strong optimization capabilities when dealing with small-sized samples, but their performance is significantly influenced by factors like data size and values, underscoring the importance of further research in the domain of optimization tasks for LLMs.'}",https://arxiv.org/pdf/2310.05204
-less is more for long document summary evaluation by llms,Yunshu Wu,"Large Language Models (LLMs) have shown promising performance in summary evaluation tasks, yet they face challenges such as high computational costs and the Lost-in-the-Middle problem where important information in the middle of long documents is often overlooked. To address these issues, this paper introduces a novel approach, Extract-then-Evaluate, which involves extracting key sentences from a long source document and then evaluating the summary by prompting LLMs. The results reveal that the proposed method not only significantly reduces evaluation costs but also exhibits a higher correlation with human evaluations. Furthermore, we provide practical recommendations for optimal document length and sentence extraction methods, contributing to the development of cost-effective yet more accurate methods for LLM-based text generation evaluation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a novel approach, Extract-then-Evaluate, which involves extracting key sentences from a long source document and then evaluating the summary by prompting LLMs, revealing that the proposed method not only significantly reduces evaluation costs but also exhibits a higher correlation with human evaluations.'}",https://arxiv.org/pdf/2309.07382
-creating large language model resistant exams: guidelines and strategies,Simon Larsén,"The proliferation of Large Language Models (LLMs), such as ChatGPT, has raised concerns about their potential impact on academic integrity, prompting the need for LLM-resistant exam designs. This article investigates the performance of LLMs on exams and their implications for assessment, focusing on ChatGPT's abilities and limitations. We propose guidelines for creating LLM-resistant exams, including content moderation, deliberate inaccuracies, real-world scenarios beyond the model's knowledge base, effective distractor options, evaluating soft skills, and incorporating non-textual information. The article also highlights the significance of adapting assessments to modern tools and promoting essential skills development in students. By adopting these strategies, educators can maintain academic integrity while ensuring that assessments accurately reflect contemporary professional settings and address the challenges and opportunities posed by artificial intelligence in education.",,http://arxiv.org/pdf/2304.12203
-developing a scalable benchmark for assessing large language models in knowledge graph engineering,Lars Meyer,"As the field of Large Language Models (LLMs) evolves at an accelerated pace, the critical need to assess and monitor their performance emerges. We introduce a benchmarking framework focused on knowledge graph engineering (KGE) accompanied by three challenges addressing syntax and error correction, facts extraction and dataset generation. We show that while being a useful tool, LLMs are yet unfit to assist in knowledge graph generation with zero-shot prompting. Consequently, our LLM-KG-Bench framework provides automatic evaluation and storage of LLM responses as well as statistical data and visualization tools to support tracking of prompt engineering and model performance.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that while being a useful tool, LLMs are yet unfit to assist in knowledge graph generation with zero-shot prompting, so the LLM-KG-Bench framework provides automatic evaluation and storage of LLM responses as well as statistical data and visualization tools to support tracking of prompt engineering and model performance.'}",https://arxiv.org/pdf/2308.16622
-forward-backward reasoning in large language models for mathematical verification,Weisen Jiang,"Chain-of-Thought (CoT) prompting in large language models (LLMs) has shown promising performance on mathematical reasoning tasks. Recently, Self-Consistency samples a diverse set of reasoning chains with different answers and chooses the answer by majority voting. Though effective, its performance cannot be further improved by sampling more reasoning chains. To address this problem, we propose to integrate backward reasoning into answer verification. We first mask a number in the question by ${\bf x}$. The LLM is then asked to predict the masked number with a candidate answer $A$ embedded in the template: ``If we know the answer to the above question is $\{A\}$, what is the value of unknown variable ${\bf x}$?'' The LLM is expected to predict the masked number successfully if the provided candidate answer is correct. To further improve performance, we propose FOBAR (FOrward-BAckward Reasoning) to combine forward and backward reasoning for verifying candidate answers. Experiments are performed on six standard mathematical data sets and three LLMs (text-davinci-003, GPT-3.5-Turbo, GPT-4). Results show that FOBAR achieves state-of-the-art performance. In particular, FOBAR outperforms Self-Consistency which uses forward reasoning alone, demonstrating that combining forward and forward reasoning is better. It also outperforms existing verification methods, verifying the effectiveness of using the simple template in backward reasoning and the proposed combination.","{'model': 'tldr@v2.0.0', 'text': 'FOBAR (FOrward-BAckward Reasoning) is proposed to combine forward and backward reasoning for verifying candidate answers to improve performance and outperforms existing verification methods.'}",
-c o rrpus: codex-leveraged structured representations for neurosymbolic story understanding,Yi Dong,"Story generation and understanding—as with all NLG/NLU tasks—has seen a surge in neurosymbolic work. Researchers have recognized that, while large language models (LLMs) have tremendous utility, they can be augmented with symbolic means to be even better and to make up for many flaws that neural networks have. However, symbolic methods are extremely costly in terms of the amount of time and expertise needed to create them. In this work, we capitalize on state-of-the-art Code-LLMs, such as Codex, to bootstrap the use of symbolic methods for tracking the state of stories and aiding in story understanding. We show that our CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI Task 2 and Re 3 ) with minimal hand engineering. This work highlights the usefulness of code-based symbolic representations for enabling LLMs to better perform story reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that the CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI Task 2 and Re 3 ) with minimal hand engineering.'}",
-mind vs. mouth: on measuring re-judge inconsistency of social bias in large language models,Yachao Zhao,"Recent researches indicate that Pre-trained Large Language Models (LLMs) possess cognitive constructs similar to those observed in humans, prompting researchers to investigate the cognitive aspects of LLMs. This paper focuses on explicit and implicit social bias, a distinctive two-level cognitive construct in psychology. It posits that individuals' explicit social bias, which is their conscious expression of bias in the statements, may differ from their implicit social bias, which represents their unconscious bias. We propose a two-stage approach and discover a parallel phenomenon in LLMs known as""re-judge inconsistency""in social bias. In the initial stage, the LLM is tasked with automatically completing statements, potentially incorporating implicit social bias. However, in the subsequent stage, the same LLM re-judges the biased statement generated by itself but contradicts it. We propose that this re-judge inconsistency can be similar to the inconsistency between human's unaware implicit social bias and their aware explicit social bias. Experimental investigations on ChatGPT and GPT-4 concerning common gender biases examined in psychology corroborate the highly stable nature of the re-judge inconsistency. This finding may suggest that diverse cognitive constructs emerge as LLMs' capabilities strengthen. Consequently, leveraging psychological theories can provide enhanced insights into the underlying mechanisms governing the expressions of explicit and implicit constructs in LLMs.",,https://arxiv.org/pdf/2308.12578
-the impact of ai in physics education: a comprehensive review from gcse to university levels,Will Yeadon,"With the rapid evolution of Artificial Intelligence (AI), its potential implications for higher education have become a focal point of interest. This study delves into the capabilities of AI in Physics Education and offers actionable AI policy recommendations. Using a Large Language Model (LLM), we assessed its ability to answer 1337 Physics exam questions spanning GCSE, A-Level, and Introductory University curricula. We employed various AI prompting techniques: Zero Shot, In Context Learning, and Confirmatory Checking, which merges Chain of Thought reasoning with Reflection. The AI's proficiency varied across academic levels: it scored an average of 83.4% on GCSE, 63.8% on A-Level, and 37.4% on university-level questions, with an overall average of 59.9% using the most effective prompting technique. In a separate test, the LLM's accuracy on 5000 mathematical operations was found to decrease as the number of digits increased. Furthermore, when evaluated as a marking tool, the LLM's concordance with human markers averaged at 50.8%, with notable inaccuracies in marking straightforward questions, like multiple-choice. Given these results, our recommendations underscore caution: while current LLMs can consistently perform well on Physics questions at earlier educational stages, their efficacy diminishes with advanced content and complex calculations. LLM outputs often showcase novel methods not in the syllabus, excessive verbosity, and miscalculations in basic arithmetic. This suggests that at university, there's no substantial threat from LLMs for non-invigilated Physics questions. However, given the LLMs' considerable proficiency in writing Physics essays and coding abilities, non-invigilated examinations of these skills in Physics are highly vulnerable to automated completion by LLMs. This vulnerability also extends to Physics questions pitched at lower academic levels.",,
-s3-dst: structured open-domain dialogue segmentation and state tracking in the era of llms,Sarkar Snigdha Sarathi Das,"The traditional Dialogue State Tracking (DST) problem aims to track user preferences and intents in user-agent conversations. While sufficient for task-oriented dialogue systems supporting narrow domain applications, the advent of Large Language Model (LLM)-based chat systems has introduced many real-world intricacies in open-domain dialogues. These intricacies manifest in the form of increased complexity in contextual interactions, extended dialogue sessions encompassing a diverse array of topics, and more frequent contextual shifts. To handle these intricacies arising from evolving LLM-based chat systems, we propose joint dialogue segmentation and state tracking per segment in open-domain dialogue systems. Assuming a zero-shot setting appropriate to a true open-domain dialogue system, we propose S3-DST, a structured prompting technique that harnesses Pre-Analytical Recollection, a novel grounding mechanism we designed for improving long context tracking. To demonstrate the efficacy of our proposed approach in joint segmentation and state tracking, we evaluate S3-DST on a proprietary anonymized open-domain dialogue dataset, as well as publicly available DST and segmentation datasets. Across all datasets and settings, S3-DST consistently outperforms the state-of-the-art, demonstrating its potency and robustness the next generation of LLM-based chat systems.","{'model': 'tldr@v2.0.0', 'text': 'S3-DST is proposed, a structured prompting technique that harnesses Pre-Analytical Recollection, a novel grounding mechanism the authors designed for improving long context tracking that consistently outperforms the state-of-the-art in joint segmentation and state tracking.'}",https://arxiv.org/pdf/2309.08827
-corrpus: detecting story inconsistencies via codex-bootstrapped neurosymbolic reasoning,Yi Dong,"Story generation and understanding—as with all NLG/NLU tasks—has seen a surge in neurosymbolic work. Researchers have rec-ognized that, while large language models (LLMs) have tremendous utility, they can be augmented with symbolic means to be even better and to make up for any flaws that the neural networks might have. However, symbolic methods are extremely costly in terms of the amount of time and expertise needed to create them. In this work, we capitalize on state-of-the-art Code-LLMs, such as Codex, to boot-strap the use of symbolic methods for tracking the state of stories and aiding in story understanding. We show that our CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI task 2 and Re 3 ) with minimal hand engineering. We hope that this work can help highlight the importance of symbolic representations and specialized prompting for LLMs as these models require some guidance for performing reasoning tasks properly.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the CoRRPUS system and abstracted prompting procedures can beat current state-of-the-art structured LLM techniques on pre-existing story understanding tasks (bAbI task 2 and Re 3 ) with minimal hand engineering.'}",
-paradigm shift: motivations for qualified legal professionals to undertake academic study,G. Brennan,"ABSTRACT This paper considers the factors that motivate qualified legal professionals to undertake a further legal academic programme of study. The paper analyses the findings of a recent study concerning a post-graduate research degree collaboration between Northumbria University (NU) and the Law Society of Ireland (LSI) whereby NU’s longstanding LLM Advanced Legal Practice (LLM ALP) has been offered in Ireland through the LSI (LLM ALP (Ire)). The initial offering of the LLM ALP (Ire) has attracted a significant amount of interest from the profession, prompting the authors to consider (i) the factors that motivate ‘time-poor’ professionals to pursue the academic study of law at the post-professional level and (ii) the potential that such a course of study might have to inform and enrich the students’ practice of law. The LLM ALP (Ire) has attracted a significant amount of more established practitioners, whereas those who generally enrol on the LLM ALP in England are at the early stages of their careers, either enrolled on the Legal Practice Course (LPC) or having only recently qualified. Hence this research is focused on a particular student profile of post-professionals with established careers in legal practice.",,http://shura.shu.ac.uk/22696/3/Brewer_paradigm_shift%20%28AM%29.pdf
-automatic chain of thought prompting in large language models,Zhuosheng Zhang,"Large language models (LLMs) can perform complex reasoning by generating intermediate reasoning steps. Providing these steps for prompting demonstrations is called chain-of-thought (CoT) prompting. CoT prompting has two major paradigms. One leverages a simple prompt like""Let's think step by step""to facilitate step-by-step thinking before answering a question. The other uses a few manual demonstrations one by one, each composed of a question and a reasoning chain that leads to an answer. The superior performance of the second paradigm hinges on the hand-crafting of task-specific demonstrations one by one. We show that such manual efforts may be eliminated by leveraging LLMs with the""Let's think step by step""prompt to generate reasoning chains for demonstrations one by one, i.e., let's think not just step by step, but also one by one. However, these generated chains often come with mistakes. To mitigate the effect of such mistakes, we find that diversity matters for automatically constructing demonstrations. We propose an automatic CoT prompting method: Auto-CoT. It samples questions with diversity and generates reasoning chains to construct demonstrations. On ten public benchmark reasoning tasks with GPT-3, Auto-CoT consistently matches or exceeds the performance of the CoT paradigm that requires manual designs of demonstrations. Code is available at https://github.com/amazon-research/auto-cot","{'model': 'tldr@v2.0.0', 'text': 'An automatic CoT prompting method that samples questions with diversity and generates reasoning chains to construct demonstrations and consistently matches or exceeds the performance of the CoT paradigm that requires manual designs of demonstrations.'}",http://arxiv.org/pdf/2210.03493
-llm abstracts for ajh v97 s3,Anemona Lucia,"inflammatory cytokines that are released as part of the systemic inflammatory response to acute pancreatitis. Decrease in nitric oxide (NO) level predisposes the patient to thrombotic microangiopathy. In in vitro studies, it is proved that inflammatory cytokines stimulated endothelial cell release of ultra large vWF multimers and inhibited the cleavage of ultra large vWF by ADAMTS13. In acute pancreatitis, endothelial damage is found to be associated with a decrease in endothelial NO synthase production. In rat models, this decrease in NO synthase has been found to be an important mediator in the pathogenesis of acute pancreatitis.These studies strongly favor the lack of NO as being one of the significant contributors to AP-induced TTP rather than just the lack of ADAMTS-13. Discussion: In patients with acute pancreatitis, any sudden drop in hemoglobin or platelet count should raise the clinical suspicion of TTP, especially with evidence of schistocytes in peripheral blood smear. Diverse mechanisms apart from ADAMTS-13 deficiency may be involved. Prompt diagnosis is required to start early treatment with plasmapheresis which brings a favorable outcome. of V-based were (PFS), patients; reduction or interruption of lenalidomide in 64% of patients; and to interruption of rituximab in 30% of patients (dose reduction for rituximab was not allowed). Neutropenia was the most common TEAE leading to lenalidomide discontinuation in 6% and reduction/interruption in 32%, and rituximab discontinuation in 3%. Infusion-related reaction was the most common TEAE leading to rituximab interruption in 8%. Discussion: These data represent complete analysis of all patients in the induction phase of MAGNIFY which continue to support that R 2 is active with a tolerable safety profile in patients with R/R FL grade 1 – 3a and MZL, including rituximab-refractory, double-refractory, and early relapse patients. of from 8745 patients with MM, 4610 with MGUS, 303 patients with WM, 77 with ALA and 75 with SMM have been collected. Together 13811 patients have been included in the registry until end of June 2022, 10000th patient was assigned to RMG on October 17th 2018. Median follow-up of MGUS patients is 7 years (0.5-46 years) and median follow-up for MM patients is 6 years (0.5-32 years). The huge amount of data allowed already many regular analysis and publication of treatment results of MM patients treated with novel drugs of multiple myeloma in the Czech Republic. The new prognostic models for MGUS and SMM progression have been created based on registry data. The RMG data has been also used in several international projects, such as Insight or Honeur. no evidence of cord compression, and he was given supportive care for his symptoms. Given known history of LMD and high risk of relapse, decision was made to start high-dose intravenous methotrexate for CNS prophylaxis followed by leucovorin rescue. He developed profound renal failure and was started on intravenous fluids, while leucovorin rescue therapy was extended. His renal functions slowly improved and he was discharged to short-term nursing home for continued rehabilitation. However, he was brought back to the hospital within a week with worsening neurological symptoms including intermittent episodes of unilateral blindness, altered mentation, and fluctuating weakness. He also became critically ill with sepsis and febrile neutropenia. He passed away shortly after being transitioned to comfort measures only. Results: Systemic administration of HD-MTX is the most widely used alternative to intrathecal chemotherapy in patients with leptomeningeal metastases and achieves higher concentration of methotrexate in cerebrospinal fluid than does IT administration. Although recommended by some guidelines and increasingly utilized as a means of CNS prophylaxis in high-risk patients, the use of HD-MTX in DLBCL remains controversial, as evidence supporting this practice is limited to small retrospective studies and several recent studies had demonstrated no benefit. In the case of our patient, no obvious benefit was observed following the use of HD-MTX, as he developed signs of recurrent LMD within 2 weeks of therapy completion. Discussion: The role of high-dose intravenous methotrexate for CNS prophylaxis in patients with diffuse large B-cell lymphoma remains controversial and the approach is associated with significant toxicity. Further researches are needed to further evaluate its efficacy. function is in patients (81%) 13 at the stage of end-stage chronic renal failure. The majority of patients (71%) have an ISS III score. All patients are eligible for chemotherapy with a predominance for the VTD protocol. Denosumab is both in newly diagnosed patients and in those (15 patients) treated with BP with an average of 6 monthly injections. (ONJ) of and had no further disease at the last visit. Two patients died after a short evolution of 3 and 12 months, the causes of which were progression and COVID-19 infection. Discussion: PMBCL is a specific lymphoma entity found in young people with good overall survival. It has a distinct clinical profile and prognosis, and therefore requires an approach different therapy. The R/DA-EPOCH regimen stopped the use of radiotherapy. It's necessary be confirmed in a subsequent prospective study on a large cohort. investigated the occurrence of after initiation of zoledronic acid compared to a of hypertension, hyperlipidemia, and benign hypertrophy presented four-months of painless hematuria. prostate Laboratory (CrCl) two the role of this strategy in elderly patients using both matched and mismatched donors. COVID-19 has substantially decreased cancer screening, management visits and surgeries. CVS Health recently developed a best-in-class mobile app and website that enables oncology patients to start and stay on therapy. This study examined the impact of COVID-19 on adherence to oral oncolytic agents in a large health plan with a significant digital health platform. Methods: This retrospective cohort study included adult patients with chronic myelogenous leukemia (CML), ovarian cancer or prostate cancer initiating oral oncolytics between 3/1/19 and 3/1/2021. Patients were divided into two groups: pre-COVID oral oncolytic initiators before 3/1/20 and COVID initiators after 3/1/20 and were followed for 1 year after therapy initiation. The primary outcome was optimal adherence to oral oncolytic agents as defined by a medication possession ratio (MPR) ≥ 0.8. Percent of digital engagement, defined as the number of times a patient interacted with the CVS digital platform, was examined as a secondary endpoint and was considered as a binary and categorical endpoint (none, low (< 28), moderate (28-105) and high ( >105)). Descriptive statistics and logistic regression model-ing were performed; p-values < 0.05 were significant. Results: In total, 15,494 patients were included in the study, with 8,067 (52.07%) in the pre-COVID initiator group. Patient demographics were similar across study groups, with the exception of pre-COVID initiators who were less likely to be male (75.32% vs. 77.34%; p< 0.01) and receive copay assistance (38.37% vs. 41.70%; p< 0.01). No difference was in 3 (50%) of the patients with favorable, intermediate and poor cytogenetic abnormalities in each of the cases respectively. Dose-intensive chemotherapy was given to 5 patients (83%%). Four patients received fludarabine, idarubicin and cytarabine (FIA) and one cladribine, idarubicin, cytarabine (CLIA) and venetoclax. The other patient with intracranial myeloid sarcoma secondary to relapsed APL received all-trans retinoid acid (ATRA) and arsenic trioxide (ATO) achieving a CR, however, the patient died 30 months later. Five patients (83%) achieved morphological and radiological complete remission after induction course. Four (67%) patients are alive at the time of analysis. One expired 30 months after relapsed with intracranial myeloid sarcoma and one was lost to follow up. with MM late in the course and elected not to pursue treatment due to lack of funding. Some individuals could not afford treatment after 1 year. The unaffordability of ASCT is made apparent as only 1 patient was able to pursue it. India needs access to newer therapies at a lower cost to treat MM and provide better outcomes. relapsing DLBCL. Ben-eficial effects of treatment with little toxicity seem limited to fit patients respectively. The most common all-grade TEAE was neutropenia in the Isa-Pd arm, and diarrhea in the Isa-Kd arm. Response data are forthcoming. Discussion: In this IONA-MM first interim analysis, we report comparable pt baseline characteristics to those seen in ICARIA-MM/IKEMA with a few imbalances. Both Isa-Pd and Isa-Kd have a manageable safety profile in routine clinical practice. These data provide RWE to support the use of Isa in RRMM outside of clinical trials and in wider populations. Enrollment in IONA-MM is ongoing and will continue until the sample size (1100 pts) is reached. progression for differential diagnosis. BM examination and LN re-biopsy confirmed atypical CD20 + lymphomatus infiltration. She received 3 rd line chemotherapy (3 GDP courses) and 10 sessions of involved field radiotherapy ( IFRT ) on the STM, she developed severe anemia, acute kidney injury, convulsions and disturbed conscious level, she was admitted to the intensive care unit (ICU), and was planned for CSF analysis after neurological assessment and radiology of the brain, but unfortunately her condition progressed rapidly to multiorgan failure and died. and or transmission of oncogenic such as EBV, and HTLV-1 due to multiple transfusions could be as a predispos-ing factor in development of lymphoma in thalassemic The effects of iron overload and treatment with iron chelators could be another hypothesis for development of malignancies in these patients. in Pakistan. This study aims to compare the response of ICE versus DHAP as salvage che",,https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ajh.26736
-neuro-symbolic procedural planning with commonsense prompting,Yujie Lu,"Procedural planning aims to implement complex high-level goals by decomposition into sequential simpler low-level steps. Although procedural planning is a basic skill set for humans in daily life, it remains a challenge for large language models (LLMs) that lack a deep understanding of the cause-effect relations in procedures. Previous methods require manual exemplars to acquire procedural planning knowledge from LLMs in the zero-shot setting. However, such elicited pre-trained knowledge in LLMs induces spurious correlations between goals and steps, which impair the model generalization to unseen tasks. In contrast, this paper proposes a neuro-symbolic procedural PLANner (PLAN) that elicits procedural planning knowledge from the LLMs with commonsense-infused prompting. To mitigate spurious goal-step correlations, we use symbolic program executors on the latent procedural representations to formalize prompts from commonsense knowledge bases as a causal intervention toward the Structural Causal Model. Both automatic and human evaluations on WikiHow and RobotHow show the superiority of PLAN on procedural planning without further training or manual exemplars.","{'model': 'tldr@v2.0.0', 'text': 'A neuro-symbolic procedural PLANner that elicits procedural planning knowledge from the LLMs with commonsense-infused prompting and uses symbolic program executors on the latent procedural representations to formalize prompts from commonsense knowledge bases as a causal intervention toward the Structural Causal Model.'}",
-"llm abstracts, ajh, v96 s1",G. Garcia-Manero,"LLM Abstracts, AJH, V96 S1 PO-1 | Magrolimab + Azacitidine Versus Azacitidine + Placebo in Untreated Higher Risk Myelodysplastic Syndrome: The Phase 3, Randomized, ENHANCE Study Guillermo Garcia-Manero, Naval Daver, Jin Xu, Mark Chao, Trisha Chung, Anderson Tan, Yan Wang, Andrew Wei, Paresh Vyas, David Sallman The University of Texas MD Anderson Cancer Center, Houston, United States, Gilead Sciences, Inc., Foster City, United States, The Alfred Hospital and Monash University, Melbourne, Australia, University of Oxford, Oxford, United Kingdom, Moffitt Cancer Center, Tampa, United States Background: Myelodysplastic syndrome (MDS) is a clonal myeloid disorder characterized by cytopenia and ineffective hematopoiesis. MDS primarily affects older individuals, with the median age of diagnosis at 70 years. Prognosis and treatment decisions are guided by the revised International Prognostic Scoring System (IPSS-R) criteria. Patients with intermediate, high, and very high risk MDS (HR-MDS) have a median overall survival (OS) of 0.8 to 3.7 years. Despite the high unmet need in this patient population, azacitidine (AZA) is the only approved therapy for HR-MDS that has improved OS in clinical trials to date. However, AZA treatment is characterized by low complete response (CR) rates (10% to 17%) with limited OS (< 2 years ), indicating a need for alternative therapies. Magrolimab is a first-in-class monoclonal antibody that blocks the macrophage inhibitory immune checkpoint CD47, a “do not eat me” signal overexpressed on tumor cells. Binding of magrolimab to CD47 leads to phagocytosis of tumor cells. AZA increases expression of tumor cell prophagocytic “eat me” signals, facilitating synergy with magrolimab. In an ongoing Phase 1b study, the combination of magrolimab+ AZA led to high response rates (overall response rate 91%, with a CR of 42%) and an acceptable safety profile without significant immune-related adverse events. ENHANCE (NCT04313881) is a Phase 3 trial comparing the efficacy and safety of magrolimab + AZA with that of AZA+ placebo in previously untreated patients with HR-MDS. Trial Design / Design and Methods: Patients ≥18 years old with previously untreated intermediate to very high risk MDS by IPSS-R are eligible for ENHANCE. Randomization is 1:1 to magrolimab + AZA or AZA + placebo with no crossover allowed. Magrolimab or placebo is administered intravenously with an initial 1 mg/kg priming dose to mitigate ontarget anemia. An intrapatient dose-escalation regimen up to 30 mg/kg is then administered through Cycle 1, 30 mg/kg weekly dosing in Cycle 2, and 30 mg/kg once every 2 weeks in Cycle 3 and beyond. AZA is administered per regional prescribing information. Patients may remain on treatment until disease progression, relapse, loss of clinical benefit, or until unacceptable toxicities occur. The 2 primary efficacy endpoints are CR rate and OS. For patients undergoing allogeneic stem cell transplantation (ASCT), data for the CR rate will be censored at the time of ASCT, and OS will be censored at the last known alive date. Secondary efficacy endpoints include red blood cell transfusion independence rate, eventfree survival, minimal residual disease-negative rate, time to acute myeloid leukemia transformation, and patient-reported Functional Assessment of Cancer Therapy Anemia response rate. Biomarkers of immune cell recruitment, immune cell signaling, and bone marrow penetration of magrolimab will also be explored. Status: As of June 2021, there are 62 sites active globally, including 41 distributed across the US. Patient enrollment began in September 2020, and as of June 2021, 144 patients have been enrolled in the trial. Planned enrollment is approximately 520 patients globally, and accrual is ongoing. PO-2 | Gastrointestinal Bleeding from Necrotizing Enterocolitis in Acute Myelogenous Leukemia Treated with Arterial Embolization Ralph Gunarsa, Ricci Steven, Rio Hermawan, Joyce Bratanata, Cosmas Lesmana, Aru Sudoyo MRCCC Siloam Hospital, South Jakarta, Indonesia Introduction / Background / Significance: Necrotizing Enterocolitis (NE) is a rare intestinal complication in hematological malignancies, it characterized by bowel inflammation with wall thickening, edema, and necrosis. Although the exact pathogenesis is not completely understood, the main elements are intestinal mucosal injury, leukemic infiltration, and superimposed infection. We presented a case report of AML-M4 patient with severe intestinal bleeding as a main presentation of NE. Materials and Methods / Case Presentation / Objective: 22 years old male diagnosed with secondary AML-M4, he experienced GI bleeding before and during FLAG regimen, the profuse life threatening bleeding occurred at 18th day. He experienced fever, abdominal pain, and diarrhea 3 days before the first GI bleeding. He had no history of any GI diseases nor a family history of it. The last GI bleeding production was approximately 750ml. Physical examination revealed pale skin and lips, anemic conjunctiva, tachycardia, and low blood pressure (60/40 mmHg). Fecal examination revealed positive C.difficle toxin, blood test showed very low hemoglobin level (3g/dL), severe leukopenia (40/μl), thrombocytopenia (30.000), and consumptive coagulopathy. He transported to ICU for further monitoring and management. In ICU, he still experienced GI bleeding despite a maximal supportive and medical treatment. He undergone arterial embolization to stop the bleeding. The procedure revealed contrast extravasation at right and media colic DOI: 10.1002/ajh.26351 Am J Hematol. 2021;96:S3–S33. wileyonlinelibrary.com/journal/ajh © 2021 Wiley Periodicals LLC. S3 artery, 4 coils applied to embolize the extravasated artery. After the procedure, the bleeding was reduced significantly, on the following days his condition is stable, and he transported back to the regular ward. Colonoscopy was performed to evaluate his condition, it revealed multiple ulcers with erosion all over the colon mucosa. Biopsy of the tissue revealed chronic inflammation with necrotic component. These findings consistent with necrotizing enterocolitis (NE). Results / Description / Main Outcome Measure(s): NE is a rare condition in hematological malignancies. The proposed mechanism was mucosal injury by direct toxicity of chemotherapeutic agents (ex. cytarabine), leukemic infiltration which more common in myeloid forms of leukemia, and superimposed infection as a result of neutropenic condition. The clinical presentations are abdominal pain, nausea, vomiting, diarrhea, fever, abdominal distention, and life threatening complications of severe bleeding and perforations. It usually appears 10-14 days after completion of chemotherapy when the low leukocyte counts coincide. Imaging modalities by abdominal x-ray, USG, and CT could show thickening and dilated bowel, in this case due to neutropenic isolation condition, the procedure could not be done. Treatment of NE consists of bowel rest, supportive care, prompt antibiotic treatment, and invasive procedure including surgery. In this case, patient's symptoms occurred before the initiation of chemotherapy, and worsen after chemotherapy, this condition support the underlying mechanism are leukemic infiltration complicated by infection and chemotherapy agent. Severe GI bleeding despite maximum medication was successfully treated by arterial embolization. Conclusion(s): 1. In this case, NE occurred from 3 factors: leukemic infiltration, toxicity of chemotherapy, and superimposed infection. We considered the leukemic infiltration play a major role, because the symptoms occurred before chemotherapy. 2. NE in leukemic patients should be managed with multi-modal approach, in this case his NE was improved by chemotherapy to treat the leukemic infiltration, antibiotics for superimposed infection, and embolization for GI bleeding. References 1. Longo DL. Harrison's Hematology and Oncology 3rd Ed. USA: McGraw-Gill. 2017. 832 p. 2. Ebert EC, Hagspiel KD. Gastrointestinal Manifestations of Leukemia. J Gastroenterol Hepatol, 2012; 27 (3): 483-63. 3. Quigley MM, Bethel K, Nowacki M, Millard F, Sharpe R. Neutropenic Enterocolitis: A Rare Presenting Complication of Acute Lekuemia. Am J Hematol, 2001; 66 (3): 213-9. 4. Camera A, et al. Intestinal Toxicity During Induction Chemotherapy with Cytarabine-Based Regimens in Adult Acute Myeloid Leukemia. Hematol J, 2003; 4(): 346-50. 5. Wach M, Dmoszynska A, Wasik-Szczepanek E, Pozarowski A, Drop A, Szczpanek D. Neutropenic Eneterocolitis: A Serious Complication During the Treatment of Acute Leukemias. Ann Hematol, 2004; 83 (8): 522-6. 6. Dotson J, Elhamdani A, Petryna E, Jamil MO, Alsharedi M. Neutropenic Enterocolitis in Patients with FLT3 Mutated Acute Myeloid Leukemia Undergoing Induction Chemotherapy With Midostaurin. Int J Hematol, 2019; 109 (3): 351-5. 7. Chow EJ, Bishop KD. Painless Neutropenic Eneterocolitis in a Patient Undergoing Chemotherapy. Curr Oncol, 2016; 23 (5): e514-16. 8. Sachak T, Arnold MA, Naini BV, et al. Neutropenic enterocolitis: new insights into a deadly entity. Am J Surg Pathol 2015; 39: 1635–42. S4 ABSTRACTS","{'model': 'tldr@v2.0.0', 'text': 'ENHANCE is a Phase 3 trial comparing the efficacy and safety of magrolimab + AZA with that of AZA+ placebo in previously untreated patients with HR-MDS, a clonal myeloid disorder characterized by cytopenia and ineffective hematopoiesis.'}",
-analyzing bert’s knowledge of hypernymy via prompting,Michael Hanna,"The high performance of large pretrained language models (LLMs) such as BERT on NLP tasks has prompted questions about BERT’s linguistic capabilities, and how they differ from humans’. In this paper, we approach this question by examining BERT’s knowledge of lexical semantic relations. We focus on hypernymy, the “is-a” relation that relates a word to a superordinate category. We use a prompting methodology to simply ask BERT what the hypernym of a given word is. We find that, in a setting where all hypernyms are guessable via prompting, BERT knows hypernyms with up to 57% accuracy. Moreover, BERT with prompting outperforms other unsupervised models for hypernym discovery even in an unconstrained scenario. However, BERT’s predictions and performance on a dataset containing uncommon hyponyms and hypernyms indicate that its knowledge of hypernymy is still limited.","{'model': 'tldr@v2.0.0', 'text': 'This paper examines BERT’s knowledge of lexical semantic relations, focusing on hypernymy, the “is-a” relation that relates a word to a superordinate category, and finds that, in a setting where all hypernyms are guessable via prompting, BERT knows hypernym with up to 57% accuracy.'}",https://aclanthology.org/2021.blackboxnlp-1.20.pdf
-prompter: utilizing large language model prompting for a data efficient embodied instruction following,Y. Inoue,"Embodied Instruction Following (EIF) studies how mobile manipulator robots should be controlled to accomplish long-horizon tasks specified by natural language instructions. While most research on EIF are conducted in simulators, the ultimate goal of the field is to deploy the agents in real life. As such, it is important to minimize the data cost required for training an agent, to help the transition from sim to real. However, many studies only focus on the performance and overlook the data cost -- modules that require separate training on extra data are often introduced without a consideration on deployability. In this work, we propose FILM++ which extends the existing work FILM with modifications that do not require extra data. While all data-driven modules are kept constant, FILM++ more than doubles FILM's performance. Furthermore, we propose Prompter, which replaces FILM++'s semantic search module with language model prompting. Unlike FILM++'s implementation that requires training on extra sets of data, no training is needed for our prompting based implementation while achieving better or at least comparable performance. Prompter achieves 42.64% and 45.72% on the ALFRED benchmark with high-level instructions only and with step-by-step instructions, respectively, outperforming the previous state of the art by 6.57% and 10.31%.","{'model': 'tldr@v2.0.0', 'text': ""FILM++ is proposed, which extends the existing work FILM with modifications that do not require extra data and more than doubles FILM's performance and proposes Prompter, which replaces FILM++'s semantic search module with language model prompting.""}",https://arxiv.org/pdf/2211.03267
-terminology-aware translation with constrained decoding and large language model prompting,Nikolay Bogoychev,"Terminology correctness is important in the downstream application of machine translation, and a prevalent way to ensure this is to inject terminology constraints into a translation system. In our submission to the WMT 2023 terminology translation task, we adopt a translate-then-refine approach which can be domain-independent and requires minimal manual efforts. We annotate random source words with pseudo-terminology translations obtained from word alignment to first train a terminology-aware model. Further, we explore two post-processing methods. First, we use an alignment process to discover whether a terminology constraint has been violated, and if so, we re-decode with the violating word negatively constrained. Alternatively, we leverage a large language model to refine a hypothesis by providing it with terminology constraints. Results show that our terminology-aware model learns to incorporate terminologies effectively, and the large language model refinement process can further improve terminology recall.","{'model': 'tldr@v2.0.0', 'text': 'This work adopts a translate-then-refine approach which can be domain-independent and requires minimal manual efforts, and annotates random source words with pseudo-terminology translations obtained from word alignment to first train a terminology-aware model.'}",https://arxiv.org/pdf/2310.05824
-chain of thought prompting elicits reasoning in large language models,Jason Wei,"We explore how generating a chain of thought -- a series of intermediate reasoning steps -- significantly improves the ability of large language models to perform complex reasoning. In particular, we show how such reasoning abilities emerge naturally in sufficiently large language models via a simple method called chain of thought prompting, where a few chain of thought demonstrations are provided as exemplars in prompting. Experiments on three large language models show that chain of thought prompting improves performance on a range of arithmetic, commonsense, and symbolic reasoning tasks. The empirical gains can be striking. For instance, prompting a 540B-parameter language model with just eight chain of thought exemplars achieves state of the art accuracy on the GSM8K benchmark of math word problems, surpassing even finetuned GPT-3 with a verifier.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on three large language models show that chain of thought prompting improves performance on a range of arithmetic, commonsense, and symbolic reasoning tasks.'}",
-rethinking with retrieval: faithful large language model inference,Hangfeng He,"Despite the success of large language models (LLMs) in various natural language processing (NLP) tasks, the stored knowledge in these models may inevitably be incomplete, out-of-date, or incorrect. This motivates the need to utilize external knowledge to assist LLMs. Unfortunately, current methods for incorporating external knowledge often require additional training or fine-tuning, which can be costly and may not be feasible for LLMs. To address this issue, we propose a novel post-processing approach, rethinking with retrieval (RR), which retrieves relevant external knowledge based on the decomposed reasoning steps obtained from the chain-of-thought (CoT) prompting. This lightweight approach does not require additional training or fine-tuning and is not limited by the input length of LLMs. We evaluate the effectiveness of RR through extensive experiments with GPT-3 on three complex reasoning tasks: commonsense reasoning, temporal reasoning, and tabular reasoning. Our results show that RR can produce more faithful explanations and improve the performance of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel post-processing approach, rethinking with retrieval (RR), which retrieves relevant external knowledge based on the decomposed reasoning steps obtained from the chain-of-thought (CoT) prompting, which can produce more faithful explanations and improve the performance of LLMs.'}",http://arxiv.org/pdf/2301.00303
-pop quiz! can a large language model help with reverse engineering?,H. Pearce,"Large language models (such as OpenAI's Codex) have demonstrated impressive zero-shot multi-task capabilities in the software domain, including code explanation. In this work, we examine if this ability can be used to help with reverse engineering. Specifically, we investigate prompting Codex to identify the purpose, capabilities, and important variable names or values from code, even when the code is produced through decompilation. Alongside an examination of the model's responses in answering open-ended questions, we devise a true/false quiz framework to characterize the performance of the language model. We present an extensive quantitative analysis of the measured performance of the language model on a set of program purpose identification and information extraction tasks: of the 136,260 questions we posed, it answered 72,754 correctly. A key takeaway is that while promising, LLMs are not yet ready for zero-shot reverse engineering.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates whether prompting Codex to identify the purpose, capabilities, and important variable names or values from code, even when the code is produced through decompilation, can be used to help with reverse engineering.'}",
-least-to-most prompting enables complex reasoning in large language models,Denny Zhou,"Chain-of-thought prompting has demonstrated remarkable performance on various natural language reasoning tasks. However, it tends to perform poorly on tasks which requires solving problems harder than the exemplars shown in the prompts. To overcome this challenge of easy-to-hard generalization, we propose a novel prompting strategy, least-to-most prompting. The key idea in this strategy is to break down a complex problem into a series of simpler subproblems and then solve them in sequence. Solving each subproblem is facilitated by the answers to previously solved subproblems. Our experimental results on tasks related to symbolic manipulation, compositional generalization, and math reasoning reveal that least-to-most prompting is capable of generalizing to more difficult problems than those seen in the prompts. A notable finding is that when the GPT-3 code-davinci-002 model is used with least-to-most prompting, it can solve the compositional generalization benchmark SCAN in any split (including length split) with an accuracy of at least 99% using just 14 exemplars, compared to only 16% accuracy with chain-of-thought prompting. This is particularly noteworthy because neural-symbolic models in the literature that specialize in solving SCAN are trained on the entire training set containing over 15,000 examples. We have included prompts for all the tasks in the Appendix.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on tasks related to symbolic manipulation, compositional generalization, and math reasoning reveal that least-to-most prompting is capable of generalizing to more difficult problems than those seen in the prompts.'}",http://arxiv.org/pdf/2205.10625
-tallrec: an effective and efficient tuning framework to align large language model with recommendation,Keqin Bao,"Large Language Models (LLMs) have demonstrated remarkable performance across diverse domains, thereby prompting researchers to explore their potential for use in recommendation systems. Initial attempts have leveraged the exceptional capabilities of LLMs, such as rich knowledge and strong generalization through In-context Learning, which involves phrasing the recommendation task as prompts. Nevertheless, the performance of LLMs in recommendation tasks remains suboptimal due to a substantial disparity between the training tasks for LLMs and recommendation tasks, as well as inadequate recommendation data during pre-training. To bridge the gap, we consider building a Large Recommendation Language Model by tunning LLMs with recommendation data. To this end, we propose an efficient and effective Tuning framework for Aligning LLMs with Recommendations, namely TALLRec. We have demonstrated that the proposed TALLRec framework can significantly enhance the recommendation capabilities of LLMs in the movie and book domains, even with a limited dataset of fewer than 100 samples. Additionally, the proposed framework is highly efficient and can be executed on a single RTX 3090 with LLaMA-7B. Furthermore, the fine-tuned LLM exhibits robust cross-domain generalization. Our code and data are available at https://github.com/SAI990323/TALLRec.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the proposed TALLRec framework can significantly enhance the recommendation capabilities of LLMs in the movie and book domains, even with a limited dataset of fewer than 100 samples.'}",https://arxiv.org/pdf/2305.00447
-gorilla: large language model connected with massive apis,Shishir G. Patil,"Large Language Models (LLMs) have seen an impressive wave of advances recently, with models now excelling in a variety of tasks, such as mathematical reasoning and program synthesis. However, their potential to effectively use tools via API calls remains unfulfilled. This is a challenging task even for today's state-of-the-art LLMs such as GPT-4, largely due to their inability to generate accurate input arguments and their tendency to hallucinate the wrong usage of an API call. We release Gorilla, a finetuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls. When combined with a document retriever, Gorilla demonstrates a strong capability to adapt to test-time document changes, enabling flexible user updates or version changes. It also substantially mitigates the issue of hallucination, commonly encountered when prompting LLMs directly. To evaluate the model's ability, we introduce APIBench, a comprehensive dataset consisting of HuggingFace, TorchHub, and TensorHub APIs. The successful integration of the retrieval system with Gorilla demonstrates the potential for LLMs to use tools more accurately, keep up with frequently updated documentation, and consequently increase the reliability and applicability of their outputs. Gorilla's code, model, data, and demo are available at https://gorilla.cs.berkeley.edu","{'model': 'tldr@v2.0.0', 'text': 'Gorilla is released, a finetuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls and substantially mitigates the issue of hallucination, commonly encountered when prompting LLMs directly.'}",http://arxiv.org/pdf/2305.15334
-thoughtsource: a central hub for large language model reasoning data,Simon Ott,,"{'model': 'tldr@v2.0.0', 'text': 'The goal of ThoughtSource is to improve future artificial intelligence systems by facilitating qualitative understanding of CoTs, enabling empirical evaluations, and providing training data.'}",https://www.nature.com/articles/s41597-023-02433-3.pdf
-large language model prompt chaining for long legal document classification,Dietrich Trautmann,"Prompting is used to guide or steer a language model in generating an appropriate response that is consistent with the desired outcome. Chaining is a strategy used to decompose complex tasks into smaller, manageable components. In this study, we utilize prompt chaining for extensive legal document classification tasks, which present difficulties due to their intricate domain-specific language and considerable length. Our approach begins with the creation of a concise summary of the original document, followed by a semantic search for related exemplar texts and their corresponding annotations from a training corpus. Finally, we prompt for a label - based on the task - to assign, by leveraging the in-context learning from the few-shot prompt. We demonstrate that through prompt chaining, we can not only enhance the performance over zero-shot, but also surpass the micro-F1 score achieved by larger models, such as ChatGPT zero-shot, using smaller models.","{'model': 'tldr@v2.0.0', 'text': 'This study utilizes prompt chaining for extensive legal document classification tasks, which present difficulties due to their intricate domain-specific language and considerable length, and demonstrates that it can enhance the performance over zero-shot, but also surpass the micro-F1 score achieved by larger models.'}",https://arxiv.org/pdf/2308.04138
-generate rather than retrieve: large language models are strong context generators,W. Yu,"Knowledge-intensive tasks, such as open-domain question answering (QA), require access to a large amount of world or domain knowledge. A common approach for knowledge-intensive tasks is to employ a retrieve-then-read pipeline that first retrieves a handful of relevant contextual documents from an external corpus such as Wikipedia and then predicts an answer conditioned on the retrieved documents. In this paper, we present a novel perspective for solving knowledge-intensive tasks by replacing document retrievers with large language model generators. We call our method generate-then-read (GenRead), which first prompts a large language model to generate contextutal documents based on a given question, and then reads the generated documents to produce the final answer. Furthermore, we propose a novel clustering-based prompting method that selects distinct prompts, resulting in the generated documents that cover different perspectives, leading to better recall over acceptable answers. We conduct extensive experiments on three different knowledge-intensive tasks, including open-domain QA, fact checking, and dialogue system. Notably, GenRead achieves 71.6 and 54.4 exact match scores on TriviaQA and WebQ, significantly outperforming the state-of-the-art retrieve-then-read pipeline DPR-FiD by +4.0 and +3.9, without retrieving any documents from any external knowledge source. Lastly, we demonstrate the model performance can be further improved by combining retrieval and generation. Our code and generated documents can be found at https://github.com/wyu97/GenRead.","{'model': 'tldr@v2.0.0', 'text': 'The method generate-then-read (GenRead), which first prompts a large language model to generate contextutal documents based on a given question, and then reads the generated documents to produce the final answer, is called.'}",http://arxiv.org/pdf/2209.10063
-lessons learned from gpt-sw3: building the first large-scale generative language model for swedish,Ariel Ekgren,"We present GTP-SW3, a 3.5 billion parameter autoregressive language model, trained on a newly created 100 GB Swedish corpus. This paper provides insights with regards to data collection and training, while highlights the challenges of proper model evaluation. The results of quantitive evaluation through perplexity indicate that GPT-SW3 is a competent model in comparison with existing autoregressive models of similar size. Additionally, we perform an extensive prompting study which reveals the good text generation capabilities of GTP-SW3.","{'model': 'tldr@v2.0.0', 'text': 'The results of quantitive evaluation through perplexity indicate that GPT-SW3 is a competent model in comparison with existing autoregressive models of similar size.'}",
-how well does chatgpt do when taking the medical licensing exams? the implications of large language models for medical education and knowledge assessment,A. Gilson,"Background: ChatGPT is a 175 billion parameter natural language processing model which can generate conversation style responses to user input. Objective: To evaluate the performance of ChatGPT on questions within the scope of United States Medical Licensing Examination (USMLE) Step 1 and Step 2 exams, as well as analyze responses for user interpretability. Methods: We used two novel sets of multiple choice questions to evaluate ChatGPT's performance, each with questions pertaining to Step 1 and Step 2. The first was derived from AMBOSS, a commonly used question bank for medical students, which also provides statistics on question difficulty and the performance on an exam relative to the userbase. The second, was the National Board of Medical Examiners (NBME) Free 120-question exams. After prompting ChatGPT with each question, ChatGPT's selected answer was recorded, and the text output evaluated across three qualitative metrics: logical justification of the answer selected, presence of information internal to the question, and presence of information external to the question. Results: On the four datasets, AMBOSS-Step1, AMBOSS-Step2, NBME-Free-Step1, and NBME-Free- Step2, ChatGPT achieved accuracies of 44%, 42%, 64.4%, and 57.8%. The model demonstrated a significant decrease in performance as question difficulty increased (P=.012) within the AMBOSS- Step1 dataset. We found logical justification for ChatGPT's answer selection was present in 100% of outputs. Internal information to the question was present in >90% of all questions. The presence of information external to the question was respectively 54.5% and 27% lower for incorrect relative to correct answers on the NBME-Free-Step1 and NBME-Free-Step2 datasets (P<=.001). Conclusion: ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering. By performing at greater than 60% threshold on the NBME-Free- Step-1 dataset we show that the model is comparable to a third year medical student. Additionally, due to the dialogic nature of the response to questions, we demonstrate ChatGPT's ability to provide reasoning and informational context across the majority of answers. These facts taken together make a compelling case for the potential applications of ChatGPT as a medical education tool.","{'model': 'tldr@v2.0.0', 'text': 'ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering and shows that the model is comparable to a third year medical student by performing at greater than 60% threshold on the NBME-Free- Step-1 dataset.'}",https://www.medrxiv.org/content/medrxiv/early/2022/12/26/2022.12.23.22283901.full.pdf
-a recipe for arbitrary text style transfer with large language models,Emily Reif,"In this paper, we leverage large language models (LLMs) to perform zero-shot text style transfer. We present a prompting method that we call augmented zero-shot learning, which frames style transfer as a sentence rewriting task and requires only a natural language instruction, without model fine-tuning or exemplars in the target style. Augmented zero-shot learning is simple and demonstrates promising results not just on standard style transfer tasks such as sentiment, but also on arbitrary transformations such as ‘make this melodramatic’ or ‘insert a metaphor.’","{'model': 'tldr@v2.0.0', 'text': 'Augmented zero-shot learning is simple and demonstrates promising results not just on standard style transfer tasks such as sentiment, but also on arbitrary transformations such as ‘make this melodramatic’ or ‘insert a metaphor’.'}",https://aclanthology.org/2022.acl-short.94.pdf
-dehallucinating large language models using formal methods guided iterative prompting,Susmit Jha,"Large language models (LLMs) such as ChatGPT have been trained to generate human-like responses to natural language prompts. LLMs use a vast corpus of text data for training, and can generate coherent and contextually relevant responses to a wide range of questions and statements. Despite this remarkable progress, LLMs are prone to hallucinations making their application to safety-critical applications such as autonomous systems difficult. The hallucinations in LLMs refer to instances where the model generates responses that are not factually accurate or contextually appropriate. These hallucinations can occur due to a variety of factors, such as the model’s lack of real-world knowledge, the influence of biased or inaccurate training data, or the model’s tendency to generate responses based on statistical patterns rather than a true understanding of the input. While these hallucinations are a nuisance in tasks such as text summarization and question-answering, they can be catastrophic when LLMs are used in autonomy-relevant applications such as planning. In this paper, we focus on the application of LLMs in autonomous systems and sketch a novel self-monitoring and iterative prompting architecture that uses formal methods to detect these errors in the LLM response automatically. We exploit the dialog capability of LLMs to iteratively steer them to responses that are consistent with our correctness specification. We report preliminary experiments that show the promise of the proposed approach on tasks such as automated planning.","{'model': 'tldr@v2.0.0', 'text': 'This paper sketches a novel self-monitoring and iterative prompting architecture that uses formal methods to detect errors in the LLM response automatically and exploits the dialog capability of LLMs to iteratively steer them to responses that are consistent with the correctness specification.'}",
-"camel: communicative agents for ""mind"" exploration of large scale language model society",G. Li,"The rapid advancement of conversational and chat-based language models has led to remarkable progress in complex task-solving. However, their success heavily relies on human input to guide the conversation, which can be challenging and time-consuming. This paper explores the potential of building scalable techniques to facilitate autonomous cooperation among communicative agents and provide insight into their""cognitive""processes. To address the challenges of achieving autonomous cooperation, we propose a novel communicative agent framework named role-playing. Our approach involves using inception prompting to guide chat agents toward task completion while maintaining consistency with human intentions. We showcase how role-playing can be used to generate conversational data for studying the behaviors and capabilities of chat agents, providing a valuable resource for investigating conversational language models. Our contributions include introducing a novel communicative agent framework, offering a scalable approach for studying the cooperative behaviors and capabilities of multi-agent systems, and open-sourcing our library to support research on communicative agents and beyond. The GitHub repository of this project is made publicly available on: https://github.com/lightaime/camel.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel communicative agent framework named role-playing, using inception prompting to guide chat agents toward task completion while maintaining consistency with human intentions, and showcases how role- playing can be used to generate conversational data for studying the behaviors and capabilities of chat agents.'}",http://arxiv.org/pdf/2303.17760
-towards a mathematics formalisation assistant using large language models,Ayush Agrawal,"Mathematics formalisation is the task of writing mathematics (i.e., definitions, theorem statements, proofs) in natural language, as found in books and papers, into a formal language that can then be checked for correctness by a program. It is a thriving activity today, however formalisation remains cumbersome. In this paper, we explore the abilities of a large language model (Codex) to help with formalisation in the Lean theorem prover. We find that with careful inputdependent prompt selection and postprocessing, Codex is able to formalise short mathematical statements at undergrad level with nearly 75% accuracy for 120 theorem statements. For proofs quantitative analysis is infeasible and we undertake a detailed case study. We choose a diverse set of 13 theorems at undergrad level with proofs that fit in two-three paragraphs. We show that with a new prompting strategy Codex can formalise these proofs in natural language with at least one out of twelve Codex completion being easy to repair into a complete proof. This is surprising as essentially no aligned data exists for formalised mathematics, particularly for proofs. These results suggest that large language models are a promising avenue towards fully or partially automating formalisation.","{'model': 'tldr@v2.0.0', 'text': 'The abilities of a large language model (Codex) to help with formalisation in the Lean theorem prover are explored, finding that with careful inputdependent prompt selection and postprocessing, Codex is able to formalise short mathematical statements at undergrad level with nearly 75% accuracy for 120 theorem statements.'}",https://arxiv.org/pdf/2211.07524
-meta-learning the difference: preparing large language models for efficient adaptation,Zejiang Hou,"Abstract Large pretrained language models (PLMs) are often domain- or task-adapted via finetuning or prompting. Finetuning requires modifying all of the parameters and having enough data to avoid overfitting while prompting requires no training and few examples but limits performance. Instead, we prepare PLMs for data- and parameter-efficient adaptation by learning to learn the difference between general and adapted PLMs. This difference is expressed in terms of model weights and sublayer structure through our proposed dynamic low-rank reparameterization and learned architecture controller. Experiments on few-shot dialogue completion, low-resource abstractive summarization, and multi-domain language modeling show improvements in adaptation time and performance over direct finetuning or preparation via domain-adaptive pretraining. Ablations show our task-adaptive reparameterization (TARP) and model search (TAMS) components individually improve on other parameter-efficient transfer like adapters and structure-learning methods like learned sparsification.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on few-shot dialogue completion, low-resource abstractive summarization, and multi-domain language modeling show improvements in adaptation time and performance over direct finetuning or preparation via domain-adaptive pretraining.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00517/2059907/tacl_a_00517.pdf
-zero-shot temporal action detection via vision-language prompting,Sauradip Nag,"Existing temporal action detection (TAD) methods rely on large training data including segment-level annotations, limited to recognizing previously seen classes alone during inference. Collecting and annotating a large training set for each class of interest is costly and hence unscalable. Zero-shot TAD (ZS-TAD) resolves this obstacle by enabling a pre-trained model to recognize any unseen action classes. Meanwhile, ZS-TAD is also much more challenging with significantly less investigation. Inspired by the success of zero-shot image classification aided by vision-language (ViL) models such as CLIP, we aim to tackle the more complex TAD task. An intuitive method is to integrate an off-the-shelf proposal detector with CLIP style classification. However, due to the sequential localization (e.g, proposal generation) and classification design, it is prone to localization error propagation. To overcome this problem, in this paper we propose a novel zero-Shot Temporal Action detection model via Vision-LanguagE prompting (STALE). Such a novel design effectively eliminates the dependence between localization and classification by breaking the route for error propagation in-between. We further introduce an interaction mechanism between classification and localization for improved optimization. Extensive experiments on standard ZS-TAD video benchmarks show that our STALE significantly outperforms state-of-the-art alternatives. Besides, our model also yields superior results on supervised TAD over recent strong competitors. The PyTorch implementation of STALE is available at https://github.com/sauradip/STALE.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-Shot Temporal Action detection model via Vision-LanguagE prompting (STALE) is proposed, which effectively eliminates the dependence between localization and classification by breaking the route for error propagation in-between and introduces an interaction mechanism between classification and localization for improved optimization.'}",http://arxiv.org/pdf/2207.08184
-tree of thoughts: deliberate problem solving with large language models,Shunyu Yao,"Language models are increasingly being deployed for general problem solving across a wide range of tasks, but are still confined to token-level, left-to-right decision-making processes during inference. This means they can fall short in tasks that require exploration, strategic lookahead, or where initial decisions play a pivotal role. To surmount these challenges, we introduce a new framework for language model inference, Tree of Thoughts (ToT), which generalizes over the popular Chain of Thought approach to prompting language models, and enables exploration over coherent units of text (thoughts) that serve as intermediate steps toward problem solving. ToT allows LMs to perform deliberate decision making by considering multiple different reasoning paths and self-evaluating choices to decide the next course of action, as well as looking ahead or backtracking when necessary to make global choices. Our experiments show that ToT significantly enhances language models' problem-solving abilities on three novel tasks requiring non-trivial planning or search: Game of 24, Creative Writing, and Mini Crosswords. For instance, in Game of 24, while GPT-4 with chain-of-thought prompting only solved 4% of tasks, our method achieved a success rate of 74%. Code repo with all prompts: https://github.com/ysymyth/tree-of-thought-llm.","{'model': 'tldr@v2.0.0', 'text': 'A new framework for language model inference, Tree of Thoughts (ToT), which generalizes over the popular Chain of Thought approach to prompting language models, and enables exploration over coherent units of text (thoughts) that serve as intermediate steps toward problem solving.'}",http://arxiv.org/pdf/2305.10601
-htlm: hyper-text pre-training and prompting of language models,Armen Aghajanyan,"We introduce HTLM, a hyper-text language model trained on a large-scale web crawl. Modeling hyper-text has a number of advantages: (1) it is easily gathered at scale, (2) it provides rich document-level and end-task-adjacent supervision (e.g. class and id attributes often encode document category information), and (3) it allows for new structured prompting that follows the established semantics of HTML (e.g. to do zero-shot summarization by infilling title tags for a webpage that contains the input text). We show that pretraining with a BART-style denoising loss directly on simplified HTML provides highly effective transfer for a wide range of end tasks and supervision levels. HTLM matches or exceeds the performance of comparably sized text-only LMs for zero-shot prompting and fine-tuning for classification benchmarks, while also setting new state-of-the-art performance levels for zero-shot summarization. We also find that hyper-text prompts provide more value to HTLM, in terms of data efficiency, than plain text prompts do for existing LMs, and that HTLM is highly effective at auto-prompting itself, by simply generating the most likely hyper-text formatting for any available training data. We will release all code and models to support future HTLM research.","{'model': 'tldr@v2.0.0', 'text': 'It is found thathyper-text prompts provide more value to HTLM, in terms of data efficiency, than plain text prompts do for existing LMs, and that HTLM is highly effective at auto-prompting itself, by simply generating the most likely hyper-text formatting for any available training data.'}",
-have llms advanced enough? a challenging problem solving benchmark for large language models,Daman Arora,"The performance of large language models (LLMs) on existing reasoning benchmarks has significantly improved over the past years. In response, we present JEEBench, a considerably more challenging benchmark dataset for evaluating the problem solving abilities of LLMs. We curate 515 challenging pre-engineering mathematics, physics and chemistry problems from the highly competitive IIT JEE-Advanced exam. Long-horizon reasoning on top of deep in-domain knowledge is essential for solving problems in this benchmark. Our evaluation on various open-source and proprietary models reveals that the highest performance, even after using techniques like self-consistency, self-refinement and chain-of-thought prompting, is less than 40%. The typical failure modes of GPT-4, the best model, are errors in algebraic manipulation, difficulty in grounding abstract concepts into mathematical equations accurately and failure in retrieving relevant domain-specific concepts. We also observe that by mere prompting, GPT-4 is unable to assess risk introduced by negative marking for incorrect answers. For this, we develop a post-hoc confidence-thresholding method over self-consistency, which enables effective response selection. We hope that our challenging benchmark will guide future re-search in problem-solving using LLMs.","{'model': 'tldr@v2.0.0', 'text': 'JEEBench is presented, a considerably more challenging benchmark dataset for evaluating the problem solving abilities of LLMs, and a post-hoc confidence-thresholding method over self-consistency is developed, which enables effective response selection.'}",https://arxiv.org/pdf/2305.15074
-explainability for large language models: a survey,Haiyan Zhao,"Large language models (LLMs) have demonstrated impressive capabilities in natural language processing. However, their internal mechanisms are still unclear and this lack of transparency poses unwanted risks for downstream applications. Therefore, understanding and explaining these models is crucial for elucidating their behaviors, limitations, and social impacts. In this paper, we introduce a taxonomy of explainability techniques and provide a structured overview of methods for explaining Transformer-based language models. We categorize techniques based on the training paradigms of LLMs: traditional fine-tuning-based paradigm and prompting-based paradigm. For each paradigm, we summarize the goals and dominant approaches for generating local explanations of individual predictions and global explanations of overall model knowledge. We also discuss metrics for evaluating generated explanations, and discuss how explanations can be leveraged to debug models and improve performance. Lastly, we examine key challenges and emerging opportunities for explanation techniques in the era of LLMs in comparison to conventional machine learning models.","{'model': 'tldr@v2.0.0', 'text': 'A taxonomy of explainability techniques and a structured overview of methods for explaining Transformer-based language models is provided and key challenges and emerging opportunities for explanation techniques in the era of LLMs in comparison to conventional machine learning models are examined.'}",https://arxiv.org/pdf/2309.01029
-adapting large language models via reading comprehension,Daixuan Cheng,"We explore how continued pre-training on domain-specific corpora influences large language models, revealing that training on the raw corpora endows the model with domain knowledge, but drastically hurts its prompting ability for question answering. Taken inspiration from human learning via reading comprehension--practice after reading improves the ability to answer questions based on the learned knowledge--we propose a simple method for transforming raw corpora into reading comprehension texts. Each raw text is enriched with a series of tasks related to its content. Our method, highly scalable and applicable to any pre-training corpora, consistently enhances performance across various tasks in three different domains: biomedicine, finance, and law. Notably, our 7B language model achieves competitive performance with domain-specific models of much larger scales, such as BloombergGPT-50B. Furthermore, we demonstrate that domain-specific reading comprehension texts can improve the model's performance even on general benchmarks, showing the potential to develop a general model across even more domains. Our model, code, and data will be available at https://github.com/microsoft/LMOps.","{'model': 'tldr@v2.0.0', 'text': ""It is demonstrated that domain-specific reading comprehension texts can improve the model's performance even on general benchmarks, showing the potential to develop a general model across even more domains.""}",https://arxiv.org/pdf/2309.09530
-radadapt: radiology report summarization via lightweight domain adaptation of large language models,Dave Van Veen,"We systematically investigate lightweight strategies to adapt large language models (LLMs) for the task of radiology report summarization (RRS). Specifically, we focus on domain adaptation via pretraining (on natural language, biomedical text, or clinical text) and via discrete prompting or parameter-efficient fine-tuning. Our results consistently achieve best performance by maximally adapting to the task via pretraining on clinical text and fine-tuning on RRS examples. Importantly, this method fine-tunes a mere 0.32% of parameters throughout the model, in contrast to end-to-end fine-tuning (100% of parameters). Additionally, we study the effect of in-context examples and out-of-distribution (OOD) training before concluding with a radiologist reader study and qualitative analysis. Our findings highlight the importance of domain adaptation in RRS and provide valuable insights toward developing effective natural language processing solutions for clinical tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work systematically investigates lightweight strategies to adapt large language models for the task of radiology report summarization via pretraining and via discrete prompting or parameter-efficient fine-tuning and highlights the importance of domain adaptation in RRS.'}",https://arxiv.org/pdf/2305.01146
-evaluating factual consistency of summaries with large language models,Shiqi Chen,"Detecting factual errors in summaries has been an important and challenging subject in summarization research. Inspired by the emergent ability of large language models (LLMs), we explore evaluating factual consistency of summaries by directly prompting LLMs. We present a comprehensive empirical study to assess the ability of LLMs as factual consistency evaluators, which consists of (1) analyzing different LLMs such as the GPT model series and Flan-T5; (2) investigating a variety of prompting methods including vanilla prompting, chain-of-thought prompting, and a sentence-by-sentence prompting method to tackle long summaries; and (3) evaluating on diverse summaries generated by multiple summarization systems, ranging from pre-transformer methods to SOTA pretrained models. Our experiments demonstrate that prompting LLMs is able to outperform the previous best factuality systems in all settings, by up to 12.2 absolute points in terms of the binary classification accuracy on inconsistency detection.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive empirical study to assess the ability of LLMs as factual consistency evaluators, which consists of analyzing different LLMs such as the GPT model series and Flan-T5 and investigating a variety of prompting methods including vanilla prompting, chain-of-thought prompting, and a sentence-by-sentence prompting method to tackle long summaries.'}",https://arxiv.org/pdf/2305.14069
-large language models are diverse role-players for summarization evaluation,Ning Wu,"Text summarization has a wide range of applications in many scenarios. The evaluation of the quality of the generated text is a complex problem. A big challenge to language evaluation is that there is a clear divergence between existing metrics and human evaluation. A document summary's quality can be assessed by human annotators on various criteria, both objective ones like grammar and correctness, and subjective ones like informativeness, succinctness, and appeal. Most of the automatic evaluation methods like BLUE/ROUGE may be not able to adequately capture the above dimensions. In this paper, we propose a new evaluation framework based on LLMs, which provides a comprehensive evaluation framework by comparing generated text and reference text from both objective and subjective aspects. First, we propose to model objective and subjective dimensions of generated text based on roleplayers prompting mechanism. Furthermore, we introduce a context-based prompting mechanism that is able to generate dynamic roleplayer profiles based on input context. Finally, we design a multi-roleplayer prompting technology based on batch prompting and integrate multiple outputs into the final evaluation results. Experimental results on three real datasets for summarization show that our model is highly competitive and has a very high consistency with human annotators.","{'model': 'tldr@v2.0.0', 'text': 'A new evaluation framework based on LLMs is proposed, which provides a comprehensive evaluation framework by comparing generated text and reference text from both objective and subjective aspects and introduces a context-based prompting mechanism that is able to generate dynamic roleplayer profiles based on input context.'}",https://arxiv.org/pdf/2303.15078
-can chatgpt detect intent? evaluating large language models for spoken language understanding,Mutian He,"Recently, large pretrained language models have demonstrated strong language understanding capabilities. This is particularly reflected in their zero-shot and in-context learning abilities on downstream tasks through prompting. To assess their impact on spoken language understanding (SLU), we evaluate several such models like ChatGPT and OPT of different sizes on multiple benchmarks. We verify the emergent ability unique to the largest models as they can reach intent classification accuracy close to that of supervised models with zero or few shots on various languages given oracle transcripts. By contrast, the results for smaller models fitting a single GPU fall far behind. We note that the error cases often arise from the annotation scheme of the dataset; responses from ChatGPT are still reasonable. We show, however, that the model is worse at slot filling, and its performance is sensitive to ASR errors, suggesting serious challenges for the application of those textual models on SLU.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the ChatGPT model is worse at slot filling, and its performance is sensitive to ASR errors, suggesting serious challenges for the application of those textual models on SLU.'}",https://arxiv.org/pdf/2305.13512
-complexity-based prompting for multi-step reasoning,Yao Fu,"We study the task of prompting large-scale language models to perform multi-step reasoning. Existing work shows that when prompted with a chain of thoughts (CoT), sequences of short sentences describing intermediate reasoning steps towards a final answer, large language models can generate new reasoning chains and predict answers for new inputs. A central question is which reasoning examples make the most effective prompts. In this work, we propose complexity-based prompting, a simple and effective example selection scheme for multi-step reasoning. We show that prompts with higher reasoning complexity, i.e., chains with more reasoning steps, achieve substantially better performance on multi-step reasoning tasks over strong baselines. We further extend our complexity-based criteria from prompting (selecting inputs) to decoding (selecting outputs), where we sample multiple reasoning chains from the model, then choose the majority of generated answers from complex reasoning chains (over simple chains). When used to prompt GPT-3 and Codex, our approach substantially improves multi-step reasoning accuracy and achieves new state-of-the-art (SOTA) performance on three math benchmarks (GSM8K, MultiArith, and MathQA) and two BigBenchHard tasks (Date Understanding and Penguins), with an average +5.3 and up to +18 accuracy improvements. Compared with existing example selection schemes like manual tuning or retrieval-based selection, selection based on reasoning complexity is intuitive, easy to implement, and annotation-efficient. Further results demonstrate the robustness of performance gains from complex prompts under format perturbation and distribution shift.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes complexity-based prompting, a simple and effective example selection scheme for multi-step reasoning that substantially improves multi- step reasoning accuracy and achieves new state-of-the-art (SOTA) performance on three math benchmarks and two BigBenchHard tasks.'}",http://arxiv.org/pdf/2210.00720
-"""according to ..."" prompting language models improves quoting from pre-training data",Orion Weller,"Large Language Models (LLMs) may hallucinate and generate fake information, despite pre-training on factual data. Inspired by the journalistic device of""according to sources"", we propose according-to prompting: directing LLMs to ground responses against previously observed text. To quantify this grounding, we propose a novel evaluation metric (QUIP-Score) that measures the extent to which model-produced answers are directly found in underlying text corpora. We illustrate with experiments on Wikipedia that these prompts improve grounding under our metrics, with the additional benefit of often improving end-task performance. Furthermore, prompts that ask the model to decrease grounding (or to ground to other corpora) decrease grounding, indicating the ability of language models to increase or decrease grounded generations on request.","{'model': 'tldr@v2.0.0', 'text': 'According-to prompting is proposed: directing LLMs to ground responses against previously observed text, and a novel evaluation metric (QUIP-Score) is proposed that measures the extent to which model-produced answers are directly found in underlying text corpora.'}",http://arxiv.org/pdf/2305.13252
-prompting for a conversation: how to control a dialog model?,Josef Valvoda,"Dialog modelling faces a difficult trade-off. Models are trained on a large amount of text, yet their responses need to be limited to a desired scope and style of a dialog agent. Because the datasets used to achieve the former contain language that is not compatible with the latter, pre-trained dialog models are fine-tuned on smaller curated datasets. However, the fine-tuning process robs them of the ability to produce diverse responses, eventually reducing them to dull conversation partners. In this paper we investigate if prompting can help with mitigating the above trade-off. Specifically, we experiment with conditioning the prompt on the query, rather than training a single prompt for all queries. By following the intuition that freezing the pre-trained language model will conserve its expressivity, we find that compared to fine-tuning, prompting can achieve a higher BLEU score and substantially improve the diversity and novelty of the responses.","{'model': 'tldr@v2.0.0', 'text': 'This paper experiments with conditioning the prompt on the query, rather than training a single prompt for all queries and finds that compared to fine-tuning, prompting can achieve a higher BLEU score and substantially improve the diversity and novelty of the responses.'}",http://arxiv.org/pdf/2209.11068
-scaling instruction-finetuned language models,Hyung Won Chung,"Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling the model size, and (3) finetuning on chain-of-thought data. We find that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups (zero-shot, few-shot, CoT), and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation). For instance, Flan-PaLM 540B instruction-finetuned on 1.8K tasks outperforms PALM 540B by a large margin (+9.4% on average). Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU. We also publicly release Flan-T5 checkpoints, which achieve strong few-shot performance even compared to much larger models, such as PaLM 62B. Overall, instruction finetuning is a general method for improving the performance and usability of pretrained language models.","{'model': 'tldr@v2.0.0', 'text': 'It is found that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups, and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation).'}",http://arxiv.org/pdf/2210.11416
-multi-stage prompting for knowledgeable dialogue generation,Zihan Liu,"Existing knowledge-grounded dialogue systems typically use finetuned versions of a pretrained language model (LM) and large-scale knowledge bases. These models typically fail to generalize on topics outside of the knowledge base, and require maintaining separate potentially large checkpoints each time finetuning is needed. In this paper, we aim to address these limitations by leveraging the inherent knowledge stored in the pretrained LM as well as its powerful generation ability. We propose a multi-stage prompting approach to generate knowledgeable responses from a single pretrained LM. We first prompt the LM to generate knowledge based on the dialogue context. Then, we further prompt it to generate responses based on the dialogue context and the previously generated knowledge. Results show that our knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness. In addition, our multi-stage prompting outperforms the finetuning-based dialogue model in terms of response knowledgeability and engagement by up to 10% and 5%, respectively. Furthermore, we scale our model up to 530 billion parameters and demonstrate that larger LMs improve the generation correctness score by up to 10%, and response relevance, knowledgeability and engagement by up to 10%. Our code is available at: https://github.com/NVIDIA/Megatron-LM.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-stage prompting approach to generate knowledgeable responses from a single pretrained language model (LM) and shows that its knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness.'}",http://arxiv.org/pdf/2203.08745
-react: synergizing reasoning and acting in language models,Shunyu Yao,"While large language models (LLMs) have demonstrated impressive capabilities across tasks in language understanding and interactive decision making, their abilities for reasoning (e.g. chain-of-thought prompting) and acting (e.g. action plan generation) have primarily been studied as separate topics. In this paper, we explore the use of LLMs to generate both reasoning traces and task-specific actions in an interleaved manner, allowing for greater synergy between the two: reasoning traces help the model induce, track, and update action plans as well as handle exceptions, while actions allow it to interface with external sources, such as knowledge bases or environments, to gather additional information. We apply our approach, named ReAct, to a diverse set of language and decision making tasks and demonstrate its effectiveness over state-of-the-art baselines, as well as improved human interpretability and trustworthiness over methods without reasoning or acting components. Concretely, on question answering (HotpotQA) and fact verification (Fever), ReAct overcomes issues of hallucination and error propagation prevalent in chain-of-thought reasoning by interacting with a simple Wikipedia API, and generates human-like task-solving trajectories that are more interpretable than baselines without reasoning traces. On two interactive decision making benchmarks (ALFWorld and WebShop), ReAct outperforms imitation and reinforcement learning methods by an absolute success rate of 34% and 10% respectively, while being prompted with only one or two in-context examples. Project site with code: https://react-lm.github.io","{'model': 'tldr@v2.0.0', 'text': 'The use of LLMs are explored to generate both reasoning traces and task-specific actions in an interleaved manner, allowing for greater synergy between the two: reasoning traces help the model induce, track, and update action plans as well as handle exceptions, while actions allow it to interface with external sources to gather additional information.'}",
-unnatural instructions: tuning language models with (almost) no human labor,Or Honovich,"Instruction tuning enables pretrained language models to perform new tasks from inference-time natural language descriptions. These approaches rely on vast amounts of human supervision in the form of crowdsourced datasets or user interactions. In this work, we introduce Unnatural Instructions: a large dataset of creative and diverse instructions, collected with virtually no human labor. We collect 64,000 examples by prompting a language model with three seed examples of instructions and eliciting a fourth. This set is then expanded by prompting the model to rephrase each instruction, creating a total of approximately 240,000 examples of instructions, inputs, and outputs. Experiments show that despite containing a fair amount of noise, training on Unnatural Instructions rivals the effectiveness of training on open-source manually-curated datasets, surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks. These results demonstrate the potential of model-generated data as a cost-effective alternative to crowdsourcing for dataset expansion and diversification.","{'model': 'tldr@v2.0.0', 'text': 'Experiments show that despite containing a fair amount of noise, training on Unnatural Instructions rivals the effectiveness of training on open-source manually-curated datasets, surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks.'}",http://arxiv.org/pdf/2212.09689
-language models are multilingual chain-of-thought reasoners,Freda Shi,"We evaluate the reasoning abilities of large language models in multilingual settings. We introduce the Multilingual Grade School Math (MGSM) benchmark, by manually translating 250 grade-school math problems from the GSM8K dataset (Cobbe et al., 2021) into ten typologically diverse languages. We find that the ability to solve MGSM problems via chain-of-thought prompting emerges with increasing model scale, and that models have strikingly strong multilingual reasoning abilities, even in underrepresented languages such as Bengali and Swahili. Finally, we show that the multilingual reasoning abilities of language models extend to other tasks such as commonsense reasoning and word-in-context semantic judgment. The MGSM benchmark is publicly available at https://github.com/google-research/url-nlp.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the ability to solve MGSM problems via chain-of-thought prompting emerges with increasing model scale, and that models have strikingly strong multilingual reasoning abilities, even in underrepresented languages such as Bengali and Swahili.'}",http://arxiv.org/pdf/2210.03057
-teaching small language models to reason,Lucie Charlotte Magister,"Chain of thought prompting successfully improves the reasoning capabilities of large language models, achieving state of the art results on a range of datasets. However, these reasoning capabilities only appear to emerge in models with at least tens of billions of parameters. In this paper, we explore the transfer of such reasoning capabilities to smaller models via knowledge distillation, also investigating model and dataset size trade-off. Specifically, we finetune a student model on the chain of thought outputs generated by a larger teacher model. Our experiments show that the proposed method improves task performance across arithmetic, commonsense and symbolic reasoning datasets. For example, the accuracy of T5 XXL on GSM8K improves from 8.11% to 21.99% and 18.42% when finetuned on PaLM 540B and GPT-3 175B generated chains of thought, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This paper finetune a student model on the chain of thought outputs generated by a larger teacher model, and shows that the proposed method improves task performance across arithmetic, commonsense and symbolic reasoning datasets.'}",http://arxiv.org/pdf/2212.08410
-instruction induction: from few examples to natural language task descriptions,Or Honovich,"Large language models are able to perform a task by conditioning on a few input-output demonstrations - a paradigm known as in-context learning. We show that language models can explicitly infer an underlying task from a few demonstrations by prompting them to generate a natural language instruction that fits the examples. To explore this ability, we introduce the instruction induction challenge, compile a dataset consisting of 24 tasks, and define a novel evaluation metric based on executing the generated instruction. We discover that, to a large extent, the ability to generate instructions does indeed emerge when using a model that is both large enough and aligned to follow instructions; InstructGPT achieves 65.7% of human performance in our execution-based metric, while the original GPT-3 model reaches only 9.8% of human performance. This surprising result suggests that instruction induction might be a viable learning paradigm in and of itself, where instead of fitting a set of latent continuous parameters to the data, one searches for the best description in the natural language hypothesis space.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that language models can explicitly infer an underlying task from a few demonstrations by prompting them to generate a natural language instruction that fits the examples, and discovers that the ability to generate instructions does indeed emerge when using a model that is both large enough and aligned to follow instructions.'}",https://arxiv.org/pdf/2205.10782
-weakly supervised data augmentation through prompting for dialogue understanding,Maximillian Chen,"Dialogue understanding tasks often necessitate abundant annotated data to achieve good performance and that presents challenges in low-resource settings. To alleviate this barrier, we explore few-shot data augmentation for dialogue understanding by prompting large pre-trained language models and present a novel approach that iterates on augmentation quality by applying weakly-supervised filters. We evaluate our methods on the emotion and act classification tasks in DailyDialog and the intent classification task in Facebook Multilingual Task-Oriented Dialogue. Models fine-tuned on our augmented data mixed with few-shot ground truth data are able to approach or surpass existing state-of-the-art performance on both datasets. For DailyDialog specifically, using 10% of the ground truth data we outperform the current state-of-the-art model which uses 100% of the data.","{'model': 'tldr@v2.0.0', 'text': 'This work explores few-shot data augmentation for dialogue understanding by prompting large pre-trained language models and presents a novel approach that iterates on augmentation quality by applying weakly-supervised filters.'}",https://arxiv.org/pdf/2210.14169
-pointclip v2: prompting clip and gpt for powerful 3d open-world learning,Xiangyang Zhu,"Large-scale pre-trained models have shown promising open-world performance for both vision and language tasks. However, their transferred capacity on 3D point clouds is still limited and only constrained to the classification task. In this paper, we first collaborate CLIP and GPT to be a unified 3D open-world learner, named as PointCLIP V2, which fully unleashes their potential for zero-shot 3D classification, segmentation, and detection. To better align 3D data with the pre-trained language knowledge, PointCLIP V2 contains two key designs. For the visual end, we prompt CLIP via a shape projection module to generate more realistic depth maps, narrowing the domain gap between projected point clouds with natural images. For the textual end, we prompt the GPT model to generate 3D-specific text as the input of CLIP's textual encoder. Without any training in 3D domains, our approach significantly surpasses PointCLIP by +42.90%, +40.44%, and +28.75% accuracy on three datasets for zero-shot 3D classification. On top of that, V2 can be extended to few-shot 3D classification, zero-shot 3D part segmentation, and 3D object detection in a simple manner, demonstrating our generalization ability for unified 3D open-world learning.","{'model': 'tldr@v2.0.0', 'text': 'This paper first collaborate CLIP and GPT to be a unified 3D open-world learner, named as PointCLIP V2, which fully unleashes their potential for zero-shot 3D classification, segmentation, and detection.'}",
-milan: masked image pretraining on language assisted representation,Zejiang Hou,"Self-attention based transformer models have been dominating many computer vision tasks in the past few years. Their superb model qualities heavily depend on the excessively large labeled image datasets. In order to reduce the reliance on large labeled datasets, reconstruction based masked autoencoders are gaining popularity, which learn high quality transferable representations from unlabeled images. For the same purpose, recent weakly supervised image pretraining methods explore language supervision from text captions accompanying the images. In this work, we propose masked image pretraining on language assisted representation, dubbed as MILAN. Instead of predicting raw pixels or low level features, our pretraining objective is to reconstruct the image features with substantial semantic signals that are obtained using caption supervision. Moreover, to accommodate our reconstruction target, we propose a more effective prompting decoder architecture and a semantic aware mask sampling mechanism, which further advance the transfer performance of the pretrained model. Experimental results demonstrate that MILAN delivers higher accuracy than the previous works. When the masked autoencoder is pretrained and finetuned on ImageNet-1K dataset with an input resolution of 224x224, MILAN achieves a top-1 accuracy of 85.4% on ViT-Base, surpassing previous state-of-the-arts by 1%. In the downstream semantic segmentation task, MILAN achieves 52.7 mIoU using ViT-Base on ADE20K dataset, outperforming previous masked pretraining results by 4 points.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes masked image pretraining on language assisted representation, dubbed as MILAN, and proposes a more effective prompting decoder architecture and a semantic aware mask sampling mechanism, which further advance the transfer performance of the pretrained model.'}",http://arxiv.org/pdf/2208.06049
-errors are useful prompts: instruction guided task programming with verifier-assisted iterative prompting,M. Skreta,"Generating low-level robot task plans from high-level natural language instructions remains a challenging problem. Although large language models have shown promising results in generating plans, the accuracy of the output remains unverified. Furthermore, the lack of domain-specific language data poses a limitation on the applicability of these models. In this paper, we propose CLAIRIFY, a novel approach that combines automatic iterative prompting with program verification to ensure programs written in data-scarce domain-specific language are syntactically valid and incorporate environment constraints. Our approach provides effective guidance to the language model on generating structured-like task plans by incorporating any errors as feedback, while the verifier ensures the syntactic accuracy of the generated plans. We demonstrate the effectiveness of CLAIRIFY in planning chemistry experiments by achieving state-of-the-art results. We also show that the generated plans can be executed on a real robot by integrating them with a task and motion planner.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes CLAIRIFY, a novel approach that combines automatic iterative prompting with program verification to ensure programs written in data-scarce domain-specific language are syntactically valid and incorporate environment constraints.'}",http://arxiv.org/pdf/2303.14100
-language is not all you need: aligning perception with language models,Shaohan Huang,"A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context, and follow instructions, and shows that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodals, and from multimodal to language.'}",http://arxiv.org/pdf/2302.14045
-improving factuality and reasoning in language models through multiagent debate,Yilun Du,"Large language models (LLMs) have demonstrated remarkable capabilities in language generation, understanding, and few-shot learning in recent years. An extensive body of work has explored how their performance may be further improved through the tools of prompting, ranging from verification, self-consistency, or intermediate scratchpads. In this paper, we present a complementary approach to improve language responses where multiple language model instances propose and debate their individual responses and reasoning processes over multiple rounds to arrive at a common final answer. Our findings indicate that this approach significantly enhances mathematical and strategic reasoning across a number of tasks. We also demonstrate that our approach improves the factual validity of generated content, reducing fallacious answers and hallucinations that contemporary models are prone to. Our approach may be directly applied to existing black-box models and uses identical procedure and prompts for all tasks we investigate. Overall, our findings suggest that such""society of minds""approach has the potential to significantly advance the capabilities of LLMs and pave the way for further breakthroughs in language generation and understanding.","{'model': 'tldr@v2.0.0', 'text': 'A complementary approach to improve language responses where multiple language model instances propose and debate their individual responses and reasoning processes over multiple rounds to arrive at a common final answer is presented, indicating that this approach significantly enhances mathematical and strategic reasoning across a number of tasks.'}",http://arxiv.org/pdf/2305.14325
-orca: interpreting prompted language models via locating supporting data evidence in the ocean of pretraining data,Xiaochuang Han,"Large pretrained language models have been performing increasingly well in a variety of downstream tasks via prompting. However, it remains unclear from where the model learns the task-specific knowledge, especially in a zero-shot setup. In this work, we want to find evidence of the model's task-specific competence from pretraining and are specifically interested in locating a very small subset of pretraining data that directly supports the model in the task. We call such a subset supporting data evidence and propose a novel method ORCA to effectively identify it, by iteratively using gradient information related to the downstream task. This supporting data evidence offers interesting insights about the prompted language models: in the tasks of sentiment analysis and textual entailment, BERT shows a substantial reliance on BookCorpus, the smaller corpus of BERT's two pretraining corpora, as well as on pretraining examples that mask out synonyms to the task verbalizers.","{'model': 'tldr@v2.0.0', 'text': ""This work aims to find evidence of the model's task-specific competence from pretraining and proposes a novel method ORCA to effectively identify it, by iteratively using gradient information related to the downstream task.""}",https://arxiv.org/pdf/2205.12600
-can chatgpt pass high school exams on english language comprehension?,Joost C. F. de Winter,,,https://link.springer.com/content/pdf/10.1007/s40593-023-00372-z.pdf
-"“the less i type, the better”: how ai language models can enhance or impede communication for aac users",Stephanie Valencia,"Users of augmentative and alternative communication (AAC) devices sometimes find it difficult to communicate in real time with others due to the time it takes to compose messages. AI technologies such as large language models (LLMs) provide an opportunity to support AAC users by improving the quality and variety of text suggestions. However, these technologies may fundamentally change how users interact with AAC devices as users transition from typing their own phrases to prompting and selecting AI-generated phrases. We conducted a study in which 12 AAC users tested live suggestions from a language model across three usage scenarios: extending short replies, answering biographical questions, and requesting assistance. Our study participants believed that AI-generated phrases could save time, physical and cognitive effort when communicating, but felt it was important that these phrases reflect their own communication style and preferences. This work identifies opportunities and challenges for future AI-enhanced AAC devices.","{'model': 'tldr@v2.0.0', 'text': 'This work identifies opportunities and challenges for future AI-enhanced AAC devices by testing live suggestions from a language model across three usage scenarios: extending short replies, answering biographical questions, and requesting assistance.'}",https://dl.acm.org/doi/pdf/10.1145/3544548.3581560
-language-guided music recommendation for video via prompt analogies,Daniel McKee,"We propose a method to recommend music for an input video while allowing a user to guide music selection with free-form natural language. A key challenge of this problem setting is that existing music video datasets provide the needed (video, music) training pairs, but lack text descriptions of the music. This work addresses this challenge with the following three contributions. First, we propose a text-synthesis approach that relies on an analogy-based prompting procedure to generate natural language music descriptions from a large-scale language model (BLOOM-176B) given pre-trained music tagger outputs and a small number of human text descriptions. Second, we use these synthesized music descriptions to train a new trimodal model, which fuses text and video input representations to query music samples. For training, we introduce a text dropout regularization mechanism which we show is critical to model performance. Our model design allows for the re-trieved music audio to agree with the two input modalities by matching visual style depicted in the video and musical genre, mood, or instrumentation described in the natural language query. Third, to evaluate our approach, we collect a testing dataset for our problem by annotating a subset of 4k clips from the YT8M-Music Video dataset with natural language music descriptions which we make publicly available. We show that our approach can match or exceed the performance of prior methods on video-to-music retrieval while significantly improving retrieval accuracy when using text guidance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to recommend music for an input video while allowing a user to guide music selection with free-form natural language, and shows that this approach can match or exceed the performance of prior methods on video-to-music retrieval while significantly improving retrieval accuracy when using text guidance.'}",https://arxiv.org/pdf/2306.09327
-language models represent space and time,Wes Gurnee,"The capabilities of large language models (LLMs) have sparked debate over whether such systems just learn an enormous collection of superficial statistics or a coherent model of the data generating process -- a world model. We find evidence for the latter by analyzing the learned representations of three spatial datasets (world, US, NYC places) and three temporal datasets (historical figures, artworks, news headlines) in the Llama-2 family of models. We discover that LLMs learn linear representations of space and time across multiple scales. These representations are robust to prompting variations and unified across different entity types (e.g. cities and landmarks). In addition, we identify individual ``space neurons'' and ``time neurons'' that reliably encode spatial and temporal coordinates. Our analysis demonstrates that modern LLMs acquire structured knowledge about fundamental dimensions such as space and time, supporting the view that they learn not merely superficial statistics, but literal world models.","{'model': 'tldr@v2.0.0', 'text': 'It is found that modern LLMs acquire structured knowledge about fundamental dimensions such as space and time, supporting the view that they learn not merely superficial statistics, but literal world models.'}",https://arxiv.org/pdf/2310.02207
-sparks of artificial general intelligence: early experiments with gpt-4,Sébastien Bubeck,"Artificial intelligence (AI) researchers have been developing and refining large language models (LLMs) that exhibit remarkable capabilities across a variety of domains and tasks, challenging our understanding of learning and cognition. The latest model developed by OpenAI, GPT-4, was trained using an unprecedented scale of compute and data. In this paper, we report on our investigation of an early version of GPT-4, when it was still in active development by OpenAI. We contend that (this early version of) GPT-4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models. We discuss the rising capabilities and implications of these models. We demonstrate that, beyond its mastery of language, GPT-4 can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more, without needing any special prompting. Moreover, in all of these tasks, GPT-4's performance is strikingly close to human-level performance, and often vastly surpasses prior models such as ChatGPT. Given the breadth and depth of GPT-4's capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence (AGI) system. In our exploration of GPT-4, we put special emphasis on discovering its limitations, and we discuss the challenges ahead for advancing towards deeper and more comprehensive versions of AGI, including the possible need for pursuing a new paradigm that moves beyond next-word prediction. We conclude with reflections on societal influences of the recent technological leap and future research directions.","{'model': 'tldr@v2.0.0', 'text': ""It is argued that (this early version of) GPT-4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models, and the rising capabilities and implications of these models are discussed.""}",http://arxiv.org/pdf/2303.12712
-are hard examples also harder to explain? a study with human and model-generated explanations,Swarnadeep Saha,"Recent work on explainable NLP has shown that few-shot prompting can enable large pre-trained language models (LLMs) to generate grammatical and factual natural language explanations for data labels. In this work, we study the connection between explainability and sample hardness by investigating the following research question – “Are LLMs and humans equally good at explaining data labels for both easy and hard samples?” We answer this question by first collecting human-written explanations in the form of generalizable commonsense rules on the task of Winograd Schema Challenge (Winogrande dataset). We compare these explanations with those generated by GPT-3 while varying the hardness of the test samples as well as the in-context samples. We observe that (1) GPT-3 explanations are as grammatical as human explanations regardless of the hardness of the test samples, (2) for easy examples, GPT-3 generates highly supportive explanations but human explanations are more generalizable, and (3) for hard examples, human explanations are significantly better than GPT-3 explanations both in terms of label-supportiveness and generalizability judgements. We also find that hardness of the in-context examples impacts the quality of GPT-3 explanations. Finally, we show that the supportiveness and generalizability aspects of human explanations are also impacted by sample hardness, although by a much smaller margin than models.","{'model': 'tldr@v2.0.0', 'text': 'This work collects human-written explanations in the form of generalizable commonsense rules and finds that for hard examples, human explanations are significantly better than GPT-3 explanations both in terms of label-supportiveness and generalizability judgements.'}",https://arxiv.org/pdf/2211.07517
-prefix-tuning: optimizing continuous prompts for generation,Xiang Lisa Li,"Fine-tuning is the de facto way of leveraging large pretrained language models for downstream tasks. However, fine-tuning modifies all the language model parameters and therefore necessitates storing a full copy for each task. In this paper, we propose prefix-tuning, a lightweight alternative to fine-tuning for natural language generation tasks, which keeps language model parameters frozen and instead optimizes a sequence of continuous task-specific vectors, which we call the prefix. Prefix-tuning draws inspiration from prompting for language models, allowing subsequent tokens to attend to this prefix as if it were “virtual tokens”. We apply prefix-tuning to GPT-2 for table-to-text generation and to BART for summarization. We show that by learning only 0.1% of the parameters, prefix-tuning obtains comparable performance in the full data setting, outperforms fine-tuning in low-data settings, and extrapolates better to examples with topics that are unseen during training.","{'model': 'tldr@v2.0.0', 'text': 'Prefix-tuning is proposed, a lightweight alternative to fine- Tuning for natural language generation tasks, which keeps language model parameters frozen and instead optimizes a sequence of continuous task-specific vectors, which is called the prefix.'}",https://aclanthology.org/2021.acl-long.353.pdf
-a comprehensive survey on pretrained foundation models: a history from bert to chatgpt,Ce Zhou,"Pretrained Foundation Models (PFMs) are regarded as the foundation for various downstream tasks with different data modalities. A PFM (e.g., BERT, ChatGPT, and GPT-4) is trained on large-scale data which provides a reasonable parameter initialization for a wide range of downstream applications. BERT learns bidirectional encoder representations from Transformers, which are trained on large datasets as contextual language models. Similarly, the generative pretrained transformer (GPT) method employs Transformers as the feature extractor and is trained using an autoregressive paradigm on large datasets. Recently, ChatGPT shows promising success on large language models, which applies an autoregressive language model with zero shot or few shot prompting. The remarkable achievements of PFM have brought significant breakthroughs to various fields of AI. Numerous studies have proposed different methods, raising the demand for an updated survey. This study provides a comprehensive review of recent research advancements, challenges, and opportunities for PFMs in text, image, graph, as well as other data modalities. The review covers the basic components and existing pretraining methods used in natural language processing, computer vision, and graph learning. Additionally, it explores advanced PFMs used for different data modalities and unified PFMs that consider data quality and quantity. The review also discusses research related to the fundamentals of PFMs, such as model efficiency and compression, security, and privacy. Finally, the study provides key implications, future research directions, challenges, and open problems in the field of PFMs. Overall, this survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence.","{'model': 'tldr@v2.0.0', 'text': 'This survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence.'}",http://arxiv.org/pdf/2302.09419
-chatgpt makes medicine easy to swallow: an exploratory case study on simplified radiology reports,Katharina Jeblick,"The release of ChatGPT, a language model capable of generating text that appears human-like and authentic, has gained significant attention beyond the research community. We expect that the convincing performance of ChatGPT incentivizes users to apply it to a variety of downstream tasks, including prompting the model to simplify their own medical reports. To investigate this phenomenon, we conducted an exploratory case study. In a questionnaire, we asked 15 radiologists to assess the quality of radiology reports simplified by ChatGPT. Most radiologists agreed that the simplified reports were factually correct, complete, and not potentially harmful to the patient. Nevertheless, instances of incorrect statements, missed key medical findings, and potentially harmful passages were reported. While further studies are needed, the initial insights of this study indicate a great potential in using large language models like ChatGPT to improve patient-centered care in radiology and other medical domains.","{'model': 'tldr@v2.0.0', 'text': 'The initial insights of this study indicate a great potential in using large language models like ChatGPT to improve patient-centered care in radiology and other medical domains.'}",http://arxiv.org/pdf/2212.14882
-segment everything everywhere all at once,Xueyan Zou,"In this work, we present SEEM, a promptable and interactive model for segmenting everything everywhere all at once in an image, as shown in Fig.1. In SEEM, we propose a novel decoding mechanism that enables diverse prompting for all types of segmentation tasks, aiming at a universal segmentation interface that behaves like large language models (LLMs). More specifically, SEEM is designed with four desiderata: i) Versatility. We introduce a new visual prompt to unify different spatial queries including points, boxes, scribbles and masks, which can further generalize to a different referring image; ii) Compositionality. We learn a joint visual-semantic space between text and visual prompts, which facilitates the dynamic composition of two prompt types required for various segmentation tasks; iii) Interactivity. We further incorporate learnable memory prompts into the decoder to retain segmentation history through mask-guided cross-attention from decoder to image features; and iv) Semantic-awareness. We use a text encoder to encode text queries and mask labels into the same semantic space for open-vocabulary segmentation. We conduct a comprehensive empirical study to validate the effectiveness of SEEM across diverse segmentation tasks. Notably, our single SEEM model achieves competitive performance across interactive segmentation, generic segmentation, referring segmentation, and video object segmentation on 9 datasets with minimum 1/100 supervision. Furthermore, SEEM showcases a remarkable capacity for generalization to novel prompts or their combinations, rendering it a readily universal image segmentation interface.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel decoding mechanism that enables diverse prompting for all types of segmentation tasks, aiming at a universal segmentation interface that behaves like large language models (LLMs).'}",https://arxiv.org/pdf/2304.06718
-verify-and-edit: a knowledge-enhanced chain-of-thought framework,Ruochen Zhao,"As large language models (LLMs) have become the norm in NLP, demonstrating good performance in generation and reasoning tasks, one of its most fatal disadvantages is the lack of factual correctness. Generating unfactual texts not only leads to lower performances but also degrades the trust and validity of their applications. Chain-of-Thought (CoT) prompting improves trust and model performance on complex reasoning tasks by generating interpretable reasoning chains, but still suffers from factuality concerns in knowledge-intensive tasks. In this paper, we propose the Verify-and-Edit framework for CoT prompting, which seeks to increase prediction factuality by post-editing reasoning chains according to external knowledge. Building on top of GPT-3, our framework lead to accuracy improvements in multiple open-domain question-answering tasks.","{'model': 'tldr@v2.0.0', 'text': 'The Verify-and-Edit framework for CoT prompting is proposed, which seeks to increase prediction factuality by post-editing reasoning chains according to external knowledge and lead to accuracy improvements in multiple open-domain question-answering tasks.'}",https://arxiv.org/pdf/2305.03268
-fengshenbang 1.0: being the foundation of chinese cognitive intelligence,Junjie Wang,"Nowadays, foundation models become one of fundamental infrastructures in artificial intelligence, paving ways to the general intelligence. However, the reality presents two urgent challenges: existing foundation models are dominated by the English-language community; users are often given limited resources and thus cannot always use foundation models. To support the development of the Chinese-language community, we introduce an open-source project, called Fengshenbang, which leads by the research center for Cognitive Computing and Natural Language (CCNL). Our project has comprehensive capabilities, including large pre-trained models, user-friendly APIs, benchmarks, datasets, and others. We wrap all these in three sub-projects: the Fengshenbang Model, the Fengshen Framework, and the Fengshen Benchmark. An open-source roadmap, Fengshenbang, aims to re-evaluate the open-source community of Chinese pre-trained large-scale models, prompting the development of the entire Chinese large-scale model community. We also want to build a user-centered open-source ecosystem to allow individuals to access the desired models to match their computing resources. Furthermore, we invite companies, colleges, and research institutions to collaborate with us to build the large-scale open-source model-based ecosystem. We hope that this project will be the foundation of Chinese cognitive intelligence.","{'model': 'tldr@v2.0.0', 'text': 'An open-source project, called Fengshenbang, which aims to re-evaluate the open- source community of Chinese pre-trained large-scale models, prompting the development of the entire Chinese large- scale model community, and to build a user-centeredopen-source ecosystem to allow individuals to access the desired models to match their computing resources.'}",http://arxiv.org/pdf/2209.02970
-graphprompt: unifying pre-training and downstream tasks for graph neural networks,Zemin Liu,"Graphs can model complex relationships between objects, enabling a myriad of Web applications such as online page/article classification and social recommendation. While graph neural networks (GNNs) have emerged as a powerful tool for graph representation learning, in an end-to-end supervised setting, their performance heavily relies on a large amount of task-specific supervision. To reduce labeling requirement, the “pre-train, fine-tune” and “pre-train, prompt” paradigms have become increasingly common. In particular, prompting is a popular alternative to fine-tuning in natural language processing, which is designed to narrow the gap between pre-training and downstream objectives in a task-specific manner. However, existing study of prompting on graphs is still limited, lacking a universal treatment to appeal to different downstream tasks. In this paper, we propose GraphPrompt, a novel pre-training and prompting framework on graphs. GraphPrompt not only unifies pre-training and downstream tasks into a common task template, but also employs a learnable prompt to assist a downstream task in locating the most relevant knowledge from the pre-trained model in a task-specific manner. Finally, we conduct extensive experiments on five public datasets to evaluate and analyze GraphPrompt.","{'model': 'tldr@v2.0.0', 'text': 'GraphPrompt is proposed, a novel pre- training and prompting framework on graphs that unifies pre-training and downstream tasks into a common task template, but also employs a learnable prompt to assist a downstream task in locating the most relevant knowledge from the pre-trained model in a task-specific manner.'}",https://dl.acm.org/doi/pdf/10.1145/3543507.3583386
-symbolic chain-of-thought distillation: small models can also “think” step-by-step,Liunian Harold Li,"Chain-of-thought prompting (e.g., “Let’s think step-by-ste”) primes large language models to verbalize rationalization for their predictions. While chain-of-thought can lead to dramatic performance gains, benefits appear to emerge only for sufficiently large models (beyond 50B parameters). We show that orders-of-magnitude smaller models (125M—1.3B parameters) can still benefit from chain-of-thought prompting. To achieve this, we introduce Symbolic Chain-of-Thought Distillation (SCoTD), a method to train a smaller student model on rationalizations sampled from a significantly larger teacher model. Experiments across several commonsense benchmarks show that: 1) SCoTD enhances the performance of the student model in both supervised and few-shot settings, and especially for challenge sets; 2) sampling many reasoning chains per instance from the teacher is paramount; and 3) after distillation, student chain-of-thoughts are judged by humans as comparable to the teacher, despite orders of magnitude fewer parameters. We test several hypotheses regarding what properties of chain-of-thought samples are important, e.g., diversity vs. teacher likelihood vs. open-endedness. We release our corpus of chain-of-thought samples and code.","{'model': 'tldr@v2.0.0', 'text': 'Symbolic Chain-of-Thought Distillation (SCoTD) is introduced, a method to train a smaller student model on rationalizations sampled from a significantly larger teacher model that enhances the performance of the student model in both supervised and few-shot settings and especially for challenge sets.'}",http://arxiv.org/pdf/2306.14050
-towards revealing the mystery behind chain of thought: a theoretical perspective,Guhao Feng,"Recent studies have discovered that Chain-of-Thought prompting (CoT) can dramatically improve the performance of Large Language Models (LLMs), particularly when dealing with complex tasks involving mathematics or reasoning. Despite the enormous empirical success, the underlying mechanisms behind CoT and how it unlocks the potential of LLMs remain elusive. In this paper, we take a first step towards theoretically answering these questions. Specifically, we examine the expressivity of LLMs with CoT in solving fundamental mathematical and decision-making problems. We start by giving an impossibility result showing that bounded-depth Transformers are unable to directly produce correct answers for basic arithmetic/equation tasks unless the model size grows super-polynomially with respect to the input length. In contrast, we then prove by construction that autoregressive Transformers of constant size suffice to solve both tasks by generating CoT derivations using a commonly-used math language format. Moreover, we show LLMs with CoT are capable of solving a general class of decision-making problems known as Dynamic Programming, thus justifying its power in tackling complex real-world tasks. Finally, extensive experiments on four tasks show that, while Transformers always fail to predict the answers directly, they can consistently learn to generate correct solutions step-by-step given sufficient CoT demonstrations.","{'model': 'tldr@v2.0.0', 'text': 'It is proved by construction that autoregressive Transformers of constant size suffice to solve both tasks by generating CoT derivations using a commonly-used math language format, and LLMs with CoT are capable of solving a general class of decision-making problems known as Dynamic Programming, thus justifying its power in tackling complex real-world tasks.'}",http://arxiv.org/pdf/2305.15408
-zeroshotdataaug: generating and augmenting training data with chatgpt,S. Ubani,"In this paper, we investigate the use of data obtained from prompting a large generative language model, ChatGPT, to generate synthetic training data with the aim of augmenting data in low resource scenarios. We show that with appropriate task-specific ChatGPT prompts, we outperform the most popular existing approaches for such data augmentation. Furthermore, we investigate methodologies for evaluating the similarity of the augmented data generated from ChatGPT with the aim of validating and assessing the quality of the data generated.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with appropriate task-specific ChatGPT prompts, the use of data obtained from prompting a large generative language model,ChatGPT, to generate synthetic training data with the aim of augmenting data in low resource scenarios outperform the most popular existing approaches for such data augmentation.'}",http://arxiv.org/pdf/2304.14334
-selective in-context data augmentation for intent detection using pointwise v-information,Yen-Ting Lin,"This work focuses on in-context data augmentation for intent detection. Having found that augmentation via in-context prompting of large pre-trained language models (PLMs) alone does not improve performance, we introduce a novel approach based on PLMs and pointwise V-information (PVI), a metric that can measure the usefulness of a datapoint for training a model. Our method first fine-tunes a PLM on a small seed of training data and then synthesizes new datapoints - utterances that correspond to given intents. It then employs intent-aware filtering, based on PVI, to remove datapoints that are not helpful to the downstream intent classifier. Our method is thus able to leverage the expressive power of large language models to produce diverse training data. Empirical results demonstrate that our method can produce synthetic training data that achieve state-of-the-art performance on three challenging intent detection datasets under few-shot settings (1.28% absolute improvement in 5-shot and 1.18% absolute in 10-shot, on average) and perform on par with the state-of-the-art in full-shot settings (within 0.01% absolute, on average).","{'model': 'tldr@v2.0.0', 'text': 'A novel approach based on PLMs and pointwise V-information (PVI), a metric that can measure the usefulness of a datapoint for training a model, is introduced, able to leverage the expressive power of large language models to produce diverse training data.'}",http://arxiv.org/pdf/2302.05096
-reading comprehension quiz generation using generative pre-trained transformers,Ramon Dijkstra,". Recent advances in AI have resulted in large pre-trained language models with superior performance on text generation tasks, prompting the question of whether we can use them to generate educationally useful text completions. This holds the potential to generate relevant quizzes for any educational text, greatly complementing current formative and summative tests from education professionals. We explore pre-trained language models for quiz generation on reading comprehension texts and propose EduQuiz , an end-to-end quiz generator based on a GPT-3 model fine-tuned on text-quiz pairs, able to generate a complete multiple-choice question, with the correct and distractor answers. We observed that the majority of generated quizzes is reasonable, and that generation of high-quality distractors is more challenging than question and answer generation. More generally, while it may be too early to replace manually generated tests for summative feedback and grading with automatic quiz generation, EduQuiz already has potential value for formative feedback and to increase engagement during the learning phase by enhancing textbooks with assessments.",,
-meet your favorite character: open-domain chatbot mimicking fictional characters with only a few utterances,Seungju Han,"In this paper, we consider mimicking fictional characters as a promising direction for building engaging conversation models. To this end, we present a new practical task where only a few utterances of each fictional character are available to generate responses mimicking them. Furthermore, we propose a new method named Pseudo Dialog Prompting (PDP) that generates responses by leveraging the power of large-scale language models with prompts containing the target character’s utterances. To better reflect the style of the character, PDP builds the prompts in the form of dialog that includes the character’s utterances as dialog history. Since only utterances of the characters are available in the proposed task, PDP matches each utterance with an appropriate pseudo-context from a predefined set of context candidates using a retrieval model. Through human and automatic evaluation, we show that PDP generates responses that better reflect the style of fictional characters than baseline methods.","{'model': 'tldr@v2.0.0', 'text': 'A new method named Pseudo Dialog Prompting (PDP) is proposed that generates responses by leveraging the power of large-scale language models with prompts containing the target character’s utterances to better reflect the style of fictional characters.'}",http://arxiv.org/pdf/2204.10825
-towards facet-driven generation of clarifying questions for conversational search,Ivan Sekulic,"Clarifying an underlying user information need is an important aspect of a modern-day IR system. The importance of clarification is even higher in limited-bandwidth scenarios, such as conversational or mobile search, where a user is unable to easily browse through a long list of retrieved results. Thus, asking clarifying questions about user's potentially ambiguous queries arises as one of the main tasks of conversational search. Recent approaches have, while making significant progress in the field, remained limited to selecting a clarifying question from a predefined set or prompting the user with vague or template-based questions. However, with the recent advances in text generation through large-scale language models, an ideal system should generate the next clarifying question. The challenge of generating an appropriate clarifying question is twofold: 1) to produce the question in coherent natural language; 2) to ask a question that is relevant to the user query. In this paper, we propose a model that generates clarifying questions with respect to the user query and query facets. We fine-tune the GPT-2 language model to generate questions related to the query and one of the extracted query facets. Compared to competitive baselines, results show that our proposed method is both natural and useful, as judged by human annotators. Moreover, we discuss the potential theoretical framework this approach would fit in. We release the code for future work and reproducibility purposes.","{'model': 'tldr@v2.0.0', 'text': 'A model that generates clarifying questions with respect to the user query and query facets is proposed using the GPT-2 language model and results show that the proposed method is both natural and useful, as judged by human annotators.'}",
-codegen: an open large language model for code with multi-turn program synthesis,Erik Nijkamp,"Program synthesis strives to generate a computer program as a solution to a given problem specification, expressed with input-output examples or natural language descriptions. The prevalence of large language models advances the state-of-the-art for program synthesis, though limited training resources and data impede open access to such models. To democratize this, we train and release a family of large language models up to 16.1B parameters, called CODEGEN, on natural language and programming language data, and open source the training library JAXFORMER. We show the utility of the trained model by demonstrating that it is competitive with the previous state-of-the-art on zero-shot Python code generation on HumanEval. We further investigate the multi-step paradigm for program synthesis, where a single program is factorized into multiple prompts specifying subproblems. To this end, we construct an open benchmark, Multi-Turn Programming Benchmark (MTPB), consisting of 115 diverse problem sets that are factorized into multi-turn prompts. Our analysis on MTPB shows that the same intent provided to CODEGEN in multi-turn fashion significantly improves program synthesis over that provided as a single turn. We make the training library JAXFORMER and model checkpoints available as open source contribution: https://github.com/salesforce/CodeGen.","{'model': 'tldr@v2.0.0', 'text': 'This work trains and releases a family of large language models up to 16.1B parameters, called CODEGEN, on natural language and programming language data, and open source the training library JAXFORMER and model checkpoints, and investigates the multi-step paradigm for program synthesis.'}",
-promptchainer: chaining large language model prompts through visual programming,Tongshuang Sherry Wu,"While LLMs have made it possible to rapidly prototype new ML functionalities, many real-world applications involve complex tasks that cannot be easily handled via a single run of an LLM. Recent work has found that chaining multiple LLM runs together (with the output of one step being the input to the next) can help users accomplish these more complex tasks, and in a way that is perceived to be more transparent and controllable. However, it remains unknown what users need when authoring their own LLM chains – a key step to lowering the barriers for non-AI-experts to prototype AI-infused applications. In this work, we explore the LLM chain authoring process. We find from pilot studies that users need support transforming data between steps of a chain, as well as debugging the chain at multiple granularities. To address these needs, we designed PromptChainer, an interactive interface for visually programming chains. Through case studies with four designers and developers, we show that PromptChainer supports building prototypes for a range of applications, and conclude with open questions on scaling chains to even more complex tasks, as well as supporting low-fi chain prototyping.","{'model': 'tldr@v2.0.0', 'text': 'This work explores the LLM chain authoring process, and designs PromptChainer, an interactive interface for visually programming chains that supports building prototypes for a range of applications, as well as supporting low-fi chain prototyping.'}",https://arxiv.org/pdf/2203.06566
-"grips: gradient-free, edit-based instruction search for prompting large language models",Archiki Prasad,"Providing natural language instructions in prompts is a useful new paradigm for improving task performance of large language models in a zero-shot setting. Recent work has aimed to improve such prompts via manual rewriting or gradient-based tuning. However, manual rewriting is time-consuming and requires subjective interpretation, while gradient-based tuning can be extremely computationally demanding for large models and may not be feasible for API-based models. In this work, we introduce Gradient-free Instructional Prompt Search (GrIPS), a gradient-free, edit-based search approach for improving task instructions for large language models. GrIPS takes in instructions designed for humans and automatically returns an improved, edited prompt, while allowing for API-based tuning. With InstructGPT models, GrIPS improves the average task performance by up to 4.30 percentage points on eight classification tasks from the Natural Instructions dataset (with similar improvements for OPT, BLOOM, and FLAN-T5). We see improvements for both instruction-only prompts and instruction + k-shot examples prompts. Notably, GrIPS outperforms manual rewriting and purely example-based prompts while controlling for the available compute and data budget. Further, performance of GrIPS is comparable to select gradient-based tuning approaches. Qualitatively, we show our edits can simplify instructions and at times make them incoherent but nonetheless improve accuracy.","{'model': 'tldr@v2.0.0', 'text': 'Gradient-free Instructional Prompt Search (GrIPS) is introduced, a gradient-free, edit-based search approach for improving task instructions for large language models and performance of GrIPS is comparable to select gradient-based tuning approaches.'}",http://arxiv.org/pdf/2203.07281
-ai chains: transparent and controllable human-ai interaction by chaining large language model prompts,Tongshuang Sherry Wu,"Although large language models (LLMs) have demonstrated impressive potential on simple tasks, their breadth of scope, lack of transparency, and insufficient controllability can make them less effective when assisting humans on more complex tasks. In response, we introduce the concept of Chaining LLM steps together, where the output of one step becomes the input for the next, thus aggregating the gains per step. We first define a set of LLM primitive operations useful for Chain construction, then present an interactive system where users can modify these Chains, along with their intermediate results, in a modular way. In a 20-person user study, we found that Chaining not only improved the quality of task outcomes, but also significantly enhanced system transparency, controllability, and sense of collaboration. Additionally, we saw that users developed new ways of interacting with LLMs through Chains: they leveraged sub-tasks to calibrate model expectations, compared and contrasted alternative strategies by observing parallel downstream effects, and debugged unexpected model outputs by “unit-testing” sub-components of a Chain. In two case studies, we further explore how LLM Chains may be used in future applications.","{'model': 'tldr@v2.0.0', 'text': 'Chaining LLM steps together is introduced, where the output of one step becomes the input for the next, thus aggregating the gains per step, and found that Chaining not only improved the quality of task outcomes, but also significantly enhanced system transparency, controllability, and sense of collaboration.'}",https://dl.acm.org/doi/pdf/10.1145/3491102.3517582
-craft an iron sword: dynamically generating interactive game characters by prompting large language models tuned on code,Ryan Volum,"Non-Player Characters (NPCs) significantly enhance the player experience in many games. Historically, players’ interactions with NPCs have tended to be highly scripted, to be limited to natural language responses to be selected by the player, and to not involve dynamic change in game state. In this work, we demonstrate that use of a few example conversational prompts can power a conversational agent to generate both natural language and novel code. This approach can permit development of NPCs with which players can have grounded conversations that are free-form and less repetitive. We demonstrate our approach using OpenAI Codex (GPT-3 finetuned on GitHub), with Minecraft game development as our test bed. We show that with a few example prompts, a Codex-based agent can generate novel code, hold multi-turn conversations and answer questions about structured data. We evaluate this application using experienced gamers in a Minecraft realm and provide analysis of failure cases and suggest possible directions for solutions.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that use of a few example conversational prompts can power a conversational agent to generate both natural language and novel code, which can permit development of NPCs with which players can have grounded conversations that are free-form and less repetitive.'}",https://aclanthology.org/2022.wordplay-1.3.pdf
-audiopalm: a large language model that can speak and listen,Paul K. Rubenstein,"We introduce AudioPaLM, a large language model for speech understanding and generation. AudioPaLM fuses text-based and speech-based language models, PaLM-2 [Anil et al., 2023] and AudioLM [Borsos et al., 2022], into a unified multimodal architecture that can process and generate text and speech with applications including speech recognition and speech-to-speech translation. AudioPaLM inherits the capability to preserve paralinguistic information such as speaker identity and intonation from AudioLM and the linguistic knowledge present only in text large language models such as PaLM-2. We demonstrate that initializing AudioPaLM with the weights of a text-only large language model improves speech processing, successfully leveraging the larger quantity of text training data used in pretraining to assist with the speech tasks. The resulting model significantly outperforms existing systems for speech translation tasks and has the ability to perform zero-shot speech-to-text translation for many languages for which input/target language combinations were not seen in training. AudioPaLM also demonstrates features of audio language models, such as transferring a voice across languages based on a short spoken prompt. We release examples of our method at https://google-research.github.io/seanet/audiopalm/examples","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that initializing AudioPaLM with the weights of a text-only large language model improves speech processing, successfully leveraging the larger quantity of text training data used in pretraining to assist with the speech tasks.'}",http://arxiv.org/pdf/2306.12925
-in-context autoencoder for context compression in a large language model,Tao Ge,"We propose the In-context Autoencoder (ICAE), leveraging the power of a large language models (LLM) to compress a long context into short compact memory slots that can be directly conditioned on by the LLM for various purposes. ICAE is first pretrained using both autoencoding and language modeling objectives on massive text data, enabling it to generate memory slots that accurately and comprehensively represent the original context; Then, it is fine-tuned on instruction data for producing desirable responses to various prompts. Experiments demonstrate that our lightweight ICAE, introducing fewer than 1% additional parameters, effectively achieves 4X context compression based on Llama, offering advantages in both improved latency and GPU memory cost during inference, and showing an interesting insight in memorization as well as potential for scalability. These promising results imply a novel perspective on the connection between working memory in cognitive science and representation learning in LLMs, revealing ICAE's significant implications in addressing the long context problem and suggesting further research in LLM context management. Our data, code and model are released at https://github.com/getao/icae.","{'model': 'tldr@v2.0.0', 'text': 'Experiments demonstrate that the lightweight ICAE, introducing fewer than 1% additional parameters, effectively achieves 4X context compression based on Llama, offering advantages in both improved latency and GPU memory cost during inference, and showing an interesting insight in memorization as well as potential for scalability.'}",https://arxiv.org/pdf/2307.06945
-learning to prompt clip for monocular depth estimation: exploring the limits of human language,Dylan Auty,"CLIP is a significant vision-and-language training framework that has shown surprisingly general understanding of the world, with good performance in many open-ended tasks with little or no additional training. A recent technique has used CLIP to perform 0-shot Monocular Depth Estimation (MDE) by using depth-related prompts, but the use of human language in these prompts presents an unnecessary human bias. In this work, we use continuous learnable tokens in place of discrete human-language words to shed light on the problem. We achieve a significant boost in performance, and find that the learned to-kens do not map neatly to depth-related human language, implying that CLIP’s concept of depth is not succinctly ex-pressible in human language. We posit that this may extend to other CLIP concepts, and believe that this finding will spark further research into both the use and interpretation of non-linguistic tokens in all open-ended scene interpretation tasks. Code is available at https://github.","{'model': 'tldr@v2.0.0', 'text': 'A significant boost in performance is achieved, and it is found that the learned to-kens do not map neatly to depth-related human language, implying that CLIP’s concept of depth is not succinctly ex-pressible in human language.'}",
-decomposed soft prompt guided fusion enhancing for compositional zero-shot learning,Xiaocheng Lu,"Compositional Zero-Shot Learning (CZSL) aims to recognize novel concepts formed by known states and objects during training. Existing methods either learn the combined state-object representation, challenging the generalization of unseen compositions, or design two classifiers to identify state and object separately from image features, ignoring the intrinsic relationship between them. To jointly eliminate the above issues and construct a more robust CZSL system, we propose a novel framework termed Decomposed Fusion with Soft Prompt (DFSP)11Code is available at: https://github.corn/Forest-art/DFSP.git, by involving vision-language models (VLMs)for unseen composition recognition. Specifically, DFSP constructs a vector combination of learnable soft prompts with state and object to establish the joint representation of them. In addition, a cross-modal decomposed fusion module is designed between the language and image branches, which decomposes state and object among language features instead of image features. Notably, being fused with the decomposed features, the image features can be more expressive for learning the relationship with states and objects, respectively, to improve the response of unseen compositions in the pair space, hence narrowing the domain gap between seen and unseen sets. Experimental results on three challenging benchmarks demonstrate that our approach significantly outperforms other state-of-the-art methods by large margins.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework termed Decomposed Fusion with Soft Prompt (DFSP), which constructs a vector combination of learnable soft prompts with state and object to establish the joint representation of them and significantly outperforms other state-of-the-art methods by large margins.'}",https://arxiv.org/pdf/2211.10681
-zero-shot rumor detection with propagation structure via prompt learning,Hongzhan Lin,"The spread of rumors along with breaking events seriously hinders the truth in the era of social media. Previous studies reveal that due to the lack of annotated resources, rumors presented in minority languages are hard to be detected. Furthermore, the unforeseen breaking events not involved in yesterday's news exacerbate the scarcity of data resources. In this work, we propose a novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages. More specifically, we firstly represent rumor circulated on social media as diverse propagation threads, then design a hierarchical prompt encoding mechanism to learn language-agnostic contextual representations for both prompts and rumor data. To further enhance domain adaptation, we model the domain-invariant structural features from the propagation threads, to incorporate structural position representations of influential community response. In addition, a new virtual response augmentation method is used to improve model training. Extensive experiments conducted on three real-world datasets demonstrate that our proposed model achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages, which achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.'}",http://arxiv.org/pdf/2212.01117
-efficiently enhancing zero-shot performance of instruction following model via retrieval of soft prompt,Seonghyeon Ye,"Enhancing the zero-shot performance of instruction-following models requires heavy computation, either by scaling the total number of training datasets or the model size. In this work, we explore how retrieval of soft prompts obtained through prompt tuning can efficiently assist hard prompts in zero-shot task generalization. Specifically, we train soft prompt embeddings for each prompt through prompt tuning, store the samples of the training instances mapped with the prompt embeddings, and retrieve the corresponding prompt embedding of the training instance closest to the query instance during inference. While only adding 0.007% additional parameters, retrieval of soft prompt enhances the performance of T0 on unseen tasks by outperforming it on 10 out of 11 datasets as well as improving the mean accuracy of T0 on BIG-bench benchmark by 2.39% points. Also, we report an interesting finding that retrieving source embeddings trained on similar answer choice formats is more important than those on similar task types.","{'model': 'tldr@v2.0.0', 'text': 'This work explores how retrieval of soft prompts obtained through prompt tuning can efficiently assist hard prompts in zero-shot task generalization and reports an interesting finding that retrieving source embeddings trained on similar answer choice formats is more important than those on similar task types.'}",
-enhancing class understanding via prompt-tuning for zero-shot text classification,Yuhao Dan,"Zero-shot text classification (ZSTC) poses a big challenge due to the lack of labeled data for unseen classes during training. Most studies focus on transferring knowledge from seen classes to unseen classes, which have achieved good performance in most cases. Whereas, it is difficult to transfer knowledge when the classes have semantic gaps or low similarities. In this paper, we propose a prompt-based method, which enhances semantic understanding for each class and learns the matching between texts and classes for better ZSTC. Specifically, we first generate discriminative words for class description with prompt inserting (PIN). Then, a prompt matching (POM) model is learned to determine whether the text can well match the class description. Experiments on three benchmark datasets show the great advantages of our proposed method. In particular, we achieve the state-of-the-art performance on the unseen classes, while maintaining comparable strength with the existing ZSTC approaches regarding to the seen classes.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based method, which enhances semantic understanding for each class and learns the matching between texts and classes for better ZSTC, achieves the state-of-the-art performance on the unseen classes, while maintaining comparable strength with the existing ZSTH approaches regarding to the seen classes.'}",
-phonological features for 0-shot multilingual speech synthesis,Marlene Staib,"Code-switching---the intra-utterance use of multiple languages---is prevalent across the world. Within text-to-speech (TTS), multilingual models have been found to enable code-switching. By modifying the linguistic input to sequence-to-sequence TTS, we show that code-switching is possible for languages unseen during training, even within monolingual models. We use a small set of phonological features derived from the International Phonetic Alphabet (IPA), such as vowel height and frontness, consonant place and manner. This allows the model topology to stay unchanged for different languages, and enables new, previously unseen feature combinations to be interpreted by the model. We show that this allows us to generate intelligible, code-switched speech in a new language at test time, including the approximation of sounds never seen in training.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a small set of phonological features derived from the International Phonetic Alphabet (IPA) to generate intelligible, code-switched speech in a new language at test time, including the approximation of sounds never seen in training.'}",http://arxiv.org/pdf/2008.04107
-prompt-based zero-shot relation classification with semantic knowledge augmentation,Jiaying Gong,"In relation classification, recognizing unseen 001 (new) relations for which there are no training 002 instances is a challenging task. We propose 003 a prompt-based model with semantic knowl- 004 edge augmentation (ZS-SKA) to recognize un- 005 seen relations under the zero-shot setting. We 006 present a new word-level sentence translation 007 rule and generate augmented instances with 008 unseen relations from instances with seen rela- 009 tions using that new rule. We design prompts 010 based on an external knowledge graph to inte- 011 grate semantic knowledge information learned 012 from seen relations. Instead of using the actual 013 label sets in the prompt template, we construct 014 weighted virtual label words. We learn the rep- 015 resentations of both seen and unseen relations 016 with augmented instances and prompts. We 017 then calculate the distance between the gen- 018 erated representations using prototypical net- 019 works to predict unseen relations. Extensive 020 experiments conducted on three public datasets 021 show that ZS-SKA outperforms state-of-the-art 022 methods under the zero-shot scenarios. Our 023 experimental results also demonstrate the effec- 024 tiveness and robustness of ZS-SKA. 025","{'model': 'tldr@v2.0.0', 'text': 'A prompt-based model with semantic knowl- 004 edge augmentation (ZS-SKA) to recognize un- seen relations under the zero-shot setting and the results demonstrate the effec- 024 tiveness and robustness of ZS- SKA.'}",
-matching exemplar as next sentence prediction (mensp): zero-shot prompt learning for automatic scoring in science education,Xuansheng Wu,"Developing models to automatically score students' written responses to science problems is critical for science education. However, collecting and labeling sufficient student responses for training models is time and cost-consuming. Recent studies suggest that pre-trained language models can be adapted to downstream tasks without fine-tuning with prompts. However, no research has employed such a prompt approach in science education. As student responses are presented with natural language, aligning the scoring procedure as the next sentence prediction task using prompts can skip the costly fine-tuning stage. In this study, we developed a zero-shot approach to automatically score student responses via Matching Exemplars as Next Sentence Prediction (MeNSP). This approach employs no training samples. We first apply MeNSP in scoring three assessment tasks of scientific argumentation and found machine-human scoring agreements, Cohen's Kappa ranges from 0.30 to 0.57, and F1 score ranges from 0.54 to 0.81. To improve the performance, we extend our research to the few-shots setting, either randomly selecting labeled student responses or manually constructing responses to fine-tune the models. We find that one task's performance is improved with more samples, Cohen's Kappa from 0.30 to 0.38, and F1 score from 0.54 to 0.59; for the two others, scoring performance is not improved. We also find that randomly selected few-shots perform better than the human expert-crafted approach. This study suggests that MeNSP can yield referable automatic scoring for student responses while significantly reducing the cost of model training. This method can benefit low-stakes classroom assessment practices in science education. Future research should further explore the applicability of the MeNSP in different types of assessment tasks in science education and improve the model performance.",,http://arxiv.org/pdf/2301.08771
-controlling personality style in dialogue with zero-shot prompt-based learning,Angela Ramirez,"Prompt-based or in-context learning has achieved high zero-shot performance on many natural language generation (NLG) tasks. Here we explore the performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue. We experiment with prompt-based learning on the PERSONAGE restaurant recommendation corpus to generate semantically and stylistically-controlled text for 5 different Big-5 personality types: agreeable, disagreeable, conscientious, unconscientious, and extravert. We test two different classes of discrete prompts to generate utterances for a particular personality style: (1) prompts that demonstrate generating directly from a meaning representation that includes a personality specification; and (2) prompts that rely on first converting the meaning representation to a textual pseudo-reference, and then using the pseudo-reference in a textual style transfer (TST) prompt. In each case, we show that we can vastly improve performance by over-generating outputs and ranking them, testing several ranking functions based on automatic metrics for semantic accuracy, personality-match, and fluency. We also test whether NLG personality demonstrations from the restaurant domain can be used with meaning representations for the video game domain to generate personality stylized utterances about video games. Our findings show that the TST prompts produces the highest semantic accuracy (78.46% for restaurants and 87.6% for video games) and personality accuracy (100% for restaurants and 97% for video games). Our results on transferring personality style to video game utterances are surprisingly good. To our knowledge, there is no previous work testing the application of prompt-based learning to simultaneously controlling both style and semantic accuracy in NLG.","{'model': 'tldr@v2.0.0', 'text': 'The performance of prompt-based learning for simultaneously controlling the personality and the semantic accuracy of an NLG for task-oriented dialogue is explored and the results on transferring personality style to video game utterances are surprisingly good.'}",http://arxiv.org/pdf/2302.03848
-structured prompt interrogation and recursive extraction of semantics (spires): a method for populating knowledge bases using zero-shot learning,J. Caufield,"Creating knowledge bases and ontologies is a time consuming task that relies on a manual curation. AI/NLP approaches can assist expert curators in populating these knowledge bases, but current approaches rely on extensive training data, and are not able to populate arbitrary complex nested knowledge schemas. Here we present Structured Prompt Interrogation and Recursive Extraction of Semantics (SPIRES), a Knowledge Extraction approach that relies on the ability of Large Language Models (LLMs) to perform zero-shot learning (ZSL) and general-purpose query answering from flexible prompts and return information conforming to a specified schema. Given a detailed, user-defined knowledge schema and an input text, SPIRES recursively performs prompt interrogation against GPT-3+ to obtain a set of responses matching the provided schema. SPIRES uses existing ontologies and vocabularies to provide identifiers for all matched elements. We present examples of use of SPIRES in different domains, including extraction of food recipes, multi-species cellular signaling pathways, disease treatments, multi-step drug mechanisms, and chemical to disease causation graphs. Current SPIRES accuracy is comparable to the mid-range of existing Relation Extraction (RE) methods, but has the advantage of easy customization, flexibility, and, crucially, the ability to perform new tasks in the absence of any training data. This method supports a general strategy of leveraging the language interpreting capabilities of LLMs to assemble knowledge bases, assisting manual knowledge curation and acquisition while supporting validation with publicly-available databases and ontologies external to the LLM. SPIRES is available as part of the open source OntoGPT package: https://github.com/ monarch-initiative/ontogpt.","{'model': 'tldr@v2.0.0', 'text': 'Current SPIRES accuracy is comparable to the mid-range of existing Relation Extraction methods, but has the advantage of easy customization, flexibility, and, crucially, the ability to perform new tasks in the absence of any training data.'}",http://arxiv.org/pdf/2304.02711
-prompt-guided zero-shot anomaly action recognition using pretrained deep skeleton features,Fumiaki Sato,"This study investigates unsupervised anomaly action recognition, which identifies video-level abnormal-human-behavior events in an unsupervised manner without abnormal samples, and simultaneously addresses three limitations in the conventional skeleton-based approaches: target domain-dependent DNN training, robustness against skeleton errors, and a lack of normal samples. We present a unified, user prompt-guided zero-shot learning framework using a target domain-independent skeleton feature extractor, which is pretrained on a large-scale action recognition dataset. Particularly, during the training phase using normal samples, the method models the distribution of skeleton features of the normal actions while freezing the weights of the DNNs and estimates the anomaly score using this distribution in the inference phase. Additionally, to increase robustness against skeleton errors, we introduce a DNN architecture inspired by a point cloud deep learning paradigm, which sparsely propagates the features between Joints. Furthermore, to prevent the unobserved normal actions from being misidentified as abnormal actions, we incorporate a similarity score between the user prompt embeddings and skeleton features aligned in the common space into the anomaly score, which indirectly supplements normal actions. On two publicly available datasets, we conduct experiments to test the effectiveness of the proposed method with respect to above-mentioned limitations.","{'model': 'tldr@v2.0.0', 'text': 'A unified, user prompt-guided zero-shot learning framework using a target domain-independent skeleton feature extractor, which is pretrained on a large-scale action recognition dataset and incorporates a similarity score between the user prompt embeddings and skeleton features aligned in the common space into the anomaly score, which indirectly supplements normal actions.'}",https://arxiv.org/pdf/2303.15167
-bayesian sharpness-aware prompt tuning for cross-domain few-shot learning,Shuo Fan,"Few-shot learning aims to learn a classifier to recognize novel classes with only few labeled images in each class. Fine-tuning is a promising tool to solve the few-shot learning problem, which pre-trains a large-scale model on source domains and then adapts it to target domains. However, existing methods have poor generalization when encountering the domain-shifting problem in the cross-domain scenario. Inspired by recent advances on domain generalization and prompt-based tuning methods, this paper proposes Bayesian Sharpness-Aware Prompt Tuning (BSAPT) for the cross-domain few-shot learning task. Instead of learning deterministic prompts like existing methods, our BSAPT learns a weight distribution over prompts to model the uncertainty caused by limited training data and resist overfitting. To improve the generalization ability, our BSAPT seeks the prompts which lie in neighborhoods having uniformly low loss by simultaneously minimizing the training loss value and loss sharpness. Benefiting from deterministic pre-trained training and Bayesian inference, our BSAPT has better generalization ability and less overfitting than existing fine-tuning methods. Extensive experiments on public datasets show that our BSAPT outperforms state-of-the-art methods and achieves new state-of-the-art performance in the cross-domain few-shot learning task.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Bayesian Sharpness-Aware Prompt Tuning (BSAPT) for the cross-domain few-shot learning task, which learns a weight distribution over prompts to model the uncertainty caused by limited training data and resist overfitting.'}",
-zero-shot generative model adaptation via image-specific prompt learning,Jiayi Guo,"Recently, CLIP-guided image synthesis has shown appealing performance on adapting a pre-trained source-domain generator to an unseen target domain. It does not require any target-domain samples but only the textual domain labels. The training is highly efficient, e.g., a few minutes. However, existing methods still have some limitations in the quality of generated images and may suffer from the mode collapse issue. A key reason is that a fixed adaptation direction is applied for all cross-domain image pairs, which leads to identical supervision signals. To address this issue, we propose an Image-specific Prompt Learning (IPL) method, which learns specific prompt vectors for each source-domain image. This produces a more precise adaptation direction for every cross-domain image pair, endowing the target-domain generator with greatly enhanced flexibility. Qualitative and quantitative evaluations on various domains demonstrate that IPL effectively improves the quality and diversity of synthesized images and alleviates the mode collapse. Moreover, IPL is independent of the structure of the generative model, such as generative adversarial networks or diffusion models. Code is available at https://github.com/Picsart-AI-Research/IPL-Zero-Shot-Generative-Model-Adaptation.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative and quantitative evaluations on various domains demonstrate that IPL effectively improves the quality and diversity of synthesized images and alleviates the mode collapse.'}",https://arxiv.org/pdf/2304.03119
-relationprompt: leveraging prompts to generate synthetic data for zero-shot relation triplet extraction,Yew Ken Chia,"Despite the importance of relation extraction in building and representing knowledge, less research is focused on generalizing to unseen relations types. We introduce the task setting of Zero-Shot Relation Triplet Extraction (ZeroRTE) to encourage further research in low-resource relation extraction methods. Given an input sentence, each extracted triplet consists of the head entity, relation label, and tail entity where the relation label is not seen at the training stage. To solve ZeroRTE, we propose to synthesize relation examples by prompting language models to generate structured texts. Concretely, we unify language model prompts and structured text approaches to design a structured prompt template for generating synthetic relation samples when conditioning on relation label prompts (RelationPrompt). To overcome the limitation for extracting multiple relation triplets in a sentence, we design a novel Triplet Search Decoding method. Experiments on FewRel and Wiki-ZSL datasets show the efficacy of RelationPrompt for the ZeroRTE task and zero-shot relation classification. Our code and data are available at github.com/declare-lab/RelationPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This work unify language model prompts and structured text approaches to design a structured prompt template for generating synthetic relation samples when conditioning on relation label prompts (RelationPrompt), and designs a novel Triplet Search Decoding method.'}",http://arxiv.org/pdf/2203.09101
-finetune like you pretrain: improved finetuning of zero-shot vision models,Sachin Goyal,"Finetuning image-text models such as CLIP achieves state-of-the-art accuracies on a variety of benchmarks. However, recent works (Kumar et al., 2022; Wortsman et al., 2021) have shown that even subtle differences in the finetuning process can lead to surprisingly large differences in the final performance, both for in-distribution (ID) and out-of-distribution (OOD) data. In this work, we show that a natural and simple approach of mimicking contrastive pretraining consistently outperforms alternative finetuning approaches. Specifically, we cast downstream class labels as text prompts and continue optimizing the contrastive loss between image embeddings and class-descriptive prompt embeddings (contrastive finetuning). Our method consistently outperforms baselines across 7 distribution shift, 6 transfer learning, and 3 few-shot learning benchmarks. On WILDS-iWILDCam, our proposed approach FLYP outperforms the top of the leaderboard by 2.3% ID and 2.7% OOD, giving the highest reported accuracy. Averaged across 7 OOD datasets (2 WILDS and 5 ImageNet associated shifts), FLYP gives gains of 4.2% OOD over standard finetuning and outperforms current state-of-the-art (LP-FT) by more than 1 % both ID and OOD. Similarly, on 3 few-shot learning benchmarks, FLYP gives gains up to 4.6% over standard finetuning and 4.4% over the state-of-the-art. Thus we establish our proposed method of contrastive finetuning as a simple and intuitive state-of-the-art for supervised finetuning of image-text models like CLIP. Code is available at https://github.com/locuslab/FLYP.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that a natural and simple approach of mimicking contrastive pretraining consistently outperforms alternative finetuning approaches, and establishes the proposed method of contrastive finet tuning as a simple and intuitive state-of-the-art for supervised finetuned of image-text models like CLIP.'}",https://arxiv.org/pdf/2212.00638
-decoupling knowledge from memorization: retrieval-augmented prompt learning,Xiang Chen,"Prompt learning approaches have made waves in natural language processing by inducing better few-shot performance while they still follow a parametric-based learning paradigm; the oblivion and rote memorization problems in learning may encounter unstable generalization issues. Specifically, vanilla prompt learning may struggle to utilize atypical instances by rote during fully-supervised training or overfit shallow patterns with low-shot data. To alleviate such limitations, we develop RetroPrompt with the motivation of decoupling knowledge from memorization to help the model strike a balance between generalization and memorization. In contrast with vanilla prompt learning, RetroPrompt constructs an open-book knowledge-store from training instances and implements a retrieval mechanism during the process of input, training and inference, thus equipping the model with the ability to retrieve related contexts from the training corpus as cues for enhancement. Extensive experiments demonstrate that RetroPrompt can obtain better performance in both few-shot and zero-shot settings. Besides, we further illustrate that our proposed RetroPrompt can yield better generalization abilities with new datasets. Detailed analysis of memorization indeed reveals RetroPrompt can reduce the reliance of language models on memorization; thus, improving generalization for downstream tasks. Code is available in https://github.com/zjunlp/PromptKG/tree/main/research/RetroPrompt.","{'model': 'tldr@v2.0.0', 'text': 'RetroPrompt is developed with the motivation of decoupling knowledge from memorization to help the model strike a balance between generalization and memorization and can reduce the reliance of language models on memorization; thus, improving generalization for downstream tasks.'}",https://arxiv.org/pdf/2205.14704
-zero-shot video captioning with evolving pseudo-tokens,Yoad Tewel,"We introduce a zero-shot video captioning method that employs two frozen networks: the GPT-2 language model and the CLIP image-text matching model. The matching score is used to steer the language model toward generating a sentence that has a high average matching score to a subset of the video frames. Unlike zero-shot image captioning methods, our work considers the entire sentence at once. This is achieved by optimizing, during the generation process, part of the prompt from scratch, by modifying the representation of all other tokens in the prompt, and by repeating the process iteratively, gradually improving the specificity and comprehensiveness of the generated sentence. Our experiments show that the generated captions are coherent and display a broad range of real-world knowledge. Our code is available at: https://github.com/YoadTew/zero-shot-video-to-text","{'model': 'tldr@v2.0.0', 'text': 'A zero-shot video captioning method that employs two frozen networks: the GPT-2 language model and the CLIP image-text matching model that improves the specificity and comprehensiveness of the generated sentence.'}",http://arxiv.org/pdf/2207.11100
-socratic models: composing zero-shot multimodal reasoning with language,Andy Zeng,"Large pretrained (e.g.,""foundation"") models exhibit distinct capabilities depending on the domain of data they are trained on. While these domains are generic, they may only barely overlap. For example, visual-language models (VLMs) are trained on Internet-scale image captions, but large language models (LMs) are further trained on Internet-scale text with no images (e.g., spreadsheets, SAT questions, code). As a result, these models store different forms of commonsense knowledge across different domains. In this work, we show that this diversity is symbiotic, and can be leveraged through Socratic Models (SMs): a modular framework in which multiple pretrained models may be composed zero-shot i.e., via multimodal-informed prompting, to exchange information with each other and capture new multimodal capabilities, without requiring finetuning. With minimal engineering, SMs are not only competitive with state-of-the-art zero-shot image captioning and video-to-text retrieval, but also enable new applications such as (i) answering free-form questions about egocentric video, (ii) engaging in multimodal assistive dialogue with people (e.g., for cooking recipes) by interfacing with external APIs and databases (e.g., web search), and (iii) robot perception and planning.","{'model': 'tldr@v2.0.0', 'text': 'Socratic Models (SMs) are shown to be competitive with state-of-the-art zero-shot image captioning and video-to-text retrieval, and enable new applications such as answering free-form questions about egocentric video, and engaging in multimodal assistive dialogue with people.'}",https://arxiv.org/pdf/2204.00598
-improving few-shot performance of language models via nearest neighbor calibration,Feng Nie,"Pre-trained language models (PLMs) have exhibited remarkable few-shot learning capabilities when provided a few examples in a natural language prompt as demonstrations of test instances, i.e., in-context learning. However, the performance of in-context learning is susceptible to the choice of prompt format, training examples and the ordering of the training examples. In this paper, we propose a novel nearest-neighbor calibration framework for in-context learning to ease this issue. It is inspired by a phenomenon that the in-context learning paradigm produces incorrect labels when inferring training instances, which provides a useful supervised signal to calibrate predictions. Thus, our method directly augments the predictions with a $k$-nearest-neighbor ($k$NN) classifier over a datastore of cached few-shot instance representations obtained by PLMs and their corresponding labels. Then adaptive neighbor selection and feature regularization modules are introduced to make full use of a few support instances to reduce the $k$NN retrieval noise. Experiments on various few-shot text classification tasks demonstrate that our method significantly improves in-context learning, while even achieving comparable performance with state-of-the-art tuning-based approaches in some sentiment analysis tasks.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on various few-shot text classification tasks demonstrate that the proposed nearest-neighbor calibration framework significantly improves in-context learning, while even achieving comparable performance with state-of-the-art tuning-based approaches in some sentiment analysis tasks.'}",https://arxiv.org/pdf/2212.02216
-few-shot fine-grained entity typing with automatic label interpretation and instance generation,Jiaxin Huang,"We study the problem of few-shot Fine-grained Entity Typing (FET), where only a few annotated entity mentions with contexts are given for each entity type. Recently, prompt-based tuning has demonstrated superior performance to standard fine-tuning in few-shot scenarios by formulating the entity type classification task as a ''fill-in-the-blank'' problem. This allows effective utilization of the strong language modeling capability of Pre-trained Language Models (PLMs). Despite the success of current prompt-based tuning approaches, two major challenges remain: (1) the verbalizer in prompts is either manually designed or constructed from external knowledge bases, without considering the target corpus and label hierarchy information, and (2) current approaches mainly utilize the representation power of PLMs, but have not explored their generation power acquired through extensive general-domain pre-training. In this work, we propose a novel framework for few-shot FET consisting of two modules: (1) an entity type label interpretation module automatically learns to relate type labels to the vocabulary by jointly leveraging few-shot instances and the label hierarchy, and (2) a type-based contextualized instance generator produces new instances based on given instances to enlarge the training set for better generalization. On three benchmark datasets, our model outperforms existing methods by significant margins.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework for few-shot Fine-grained Entity Typing consisting of an entity type label interpretation module automatically learns to relate type labels to the vocabulary by jointly leveraging few- shot instances and the label hierarchy, and a type-based contextualized instance generator produces new instances based on given instances to enlarge the training set for better generalization.'}",https://arxiv.org/pdf/2206.13746
-natural language inference prompts for zero-shot emotion classification in text across corpora,F. Plaza-Del-Arco,"Within textual emotion classification, the set of relevant labels depends on the domain and application scenario and might not be known at the time of model development. This conflicts with the classical paradigm of supervised learning in which the labels need to be predefined. A solution to obtain a model with a flexible set of labels is to use the paradigm of zero-shot learning as a natural language inference task, which in addition adds the advantage of not needing any labeled training data. This raises the question how to prompt a natural language inference model for zero-shot learning emotion classification. Options for prompt formulations include the emotion name anger alone or the statement “This text expresses anger”. With this paper, we analyze how sensitive a natural language inference-based zero-shot-learning classifier is to such changes to the prompt under consideration of the corpus: How carefully does the prompt need to be selected? We perform experiments on an established set of emotion datasets presenting different language registers according to different sources (tweets, events, blogs) with three natural language inference models and show that indeed the choice of a particular prompt formulation needs to fit to the corpus. We show that this challenge can be tackled with combinations of multiple prompts. Such ensemble is more robust across corpora than individual prompts and shows nearly the same performance as the individual best prompt for a particular corpus.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes how sensitive a natural language inference-based zero-shot-learning classifier is to such changes to the prompt under consideration of the corpus and shows that indeed the choice of a particular prompt formulation needs to fit to the corpus.'}",http://arxiv.org/pdf/2209.06701
-clinical prompt learning with frozen language models,Niall Taylor,"Prompt learning is a new paradigm in the Natural Language Processing (NLP) field which has shown impressive performance on a number of natural language tasks with common benchmarking text datasets in full, few-shot, and zero-shot train-evaluation setups. Recently, it has even been observed that large but frozen pre-trained language models (PLMs) with prompt learning outperform smaller but fine-tuned models. However, as with many recent NLP trends, the performance of even the largest PLMs such as GPT-3 do not perform well on specialized domains (e.g. medical text), and the common practice to achieve State of the Art (SoTA) results still consists of pre-training and fine-tuning the PLMs on downstream tasks. The reliance on fine-tuning large PLMs is problematic in clinical settings where data is often held in non-GPU environments, and more resource efficient methods of training specialized domain models is crucial. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared with more traditional fine-tuning methods. Results are partially in line with the prompt learning literature, with prompt learning able to match or improve on traditional fine-tuning with substantially fewer trainable parameters and requiring less training data. We argue that prompt learning therefore provides lower computational resource costs applicable to clinical settings, that can serve as an alternative to fine-tuning ever increasing in size PLMs. Complementary code to reproduce experiments presented in this work can be found at: https: //github.com/NtaylorOX/Public_Clinical_Prompt","{'model': 'tldr@v2.0.0', 'text': 'It is argued that prompt learning provides lower computational resource costs applicable to clinical settings, that can serve as an alternative to fine-tuning ever increasing in size PLMs.'}",http://arxiv.org/pdf/2205.05535
-language models as zero-shot planners: extracting actionable knowledge for embodied agents,Wenlong Huang,"Can world knowledge learned by large language models (LLMs) be used to act in interactive environments? In this paper, we investigate the possibility of grounding high-level tasks, expressed in natural language (e.g.""make breakfast""), to a chosen set of actionable steps (e.g.""open fridge""). While prior work focused on learning from explicit step-by-step examples of how to act, we surprisingly find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into mid-level plans without any further training. However, the plans produced naively by LLMs often cannot map precisely to admissible actions. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions. Our evaluation in the recent VirtualHome environment shows that the resulting method substantially improves executability over the LLM baseline. The conducted human evaluation reveals a trade-off between executability and correctness but shows a promising sign towards extracting actionable knowledge from language models. Website at https://huangwl18.github.io/language-planner","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the possibility of grounding high-level tasks, expressed in natural language, to a chosen set of actionable steps and proposes a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions.'}",
-few-shot table-to-text generation with prefix-controlled generator,Yutao Luo,"Neural table-to-text generation approaches are data-hungry, limiting their adaption for low-resource real-world applications. Previous works mostly resort to Pre-trained Language Models (PLMs) to generate fluent summaries of a table. However, they often contain hallucinated contents due to the uncontrolled nature of PLMs. Moreover, the topological differences between tables and sequences are rarely studied. Last but not least, fine-tuning on PLMs with a handful of instances may lead to over-fitting and catastrophic forgetting. To alleviate these problems, we propose a prompt-based approach, Prefix-Controlled Generator (i.e., PCG), for few-shot table-to-text generation. We prepend a task-specific prefix for a PLM to make the table structure better fit the pre-trained input. In addition, we generate an input-specific prefix to control the factual contents and word order of the generated text. Both automatic and human evaluations on different domains (humans, books and songs) of the Wikibio dataset prove the effectiveness of our approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based approach, Prefix-Controlled Generator (i.e., PCG), for few-shot table-to-text generation, which prepend a task-specific prefix for a PLM to make the table structure better fit the pre-trained input.'}",http://arxiv.org/pdf/2208.10709
-self-guided noise-free data generation for efficient zero-shot learning,Jiahui Gao,"There is a rising interest in further exploring the zero-shot learning potential of large pre-trained language models (PLMs). A new paradigm called data-generation-based zero-shot learning has achieved impressive success. In this paradigm, the synthesized data from the PLM acts as the carrier of knowledge, which is used to train a task-specific model with orders of magnitude fewer parameters than the PLM, achieving both higher performance and efficiency than prompt-based zero-shot learning methods on PLMs. The main hurdle of this approach is that the synthesized data from PLM usually contains a significant portion of low-quality samples. Fitting on such data will greatly hamper the performance of the task-specific model, making it unreliable for deployment. Previous methods remedy this issue mainly by filtering synthetic data using heuristic metrics(e.g., output confidence), or refining the data with the help of a human expert, which comes with excessive manual tuning or expensive costs. In this paper, we propose a novel noise-robust re-weighting framework SunGen to automatically construct high-quality data for zero-shot classification problems. Our framework features the ability to learn the sample weights indicating data quality without requiring any human annotation. We theoretically and empirically verify the ability of our method to help construct good-quality synthetic datasets. Notably, SunGen-LSTM yields a 9.8% relative improvement than the baseline on average accuracy across eight different established text classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel noise-robust re-weighting framework SunGen is proposed to automatically construct high-quality data for zero-shot classification problems, which features the ability to learn the sample weights indicating data quality without requiring any human annotation.'}",
-p3 ranker: mitigating the gaps between pre-training and ranking fine-tuning with prompt-based learning and pre-finetuning,Xiaomeng Hu,"Compared to other language tasks, applying pre-trained language models (PLMs) for search ranking often requires more nuances and training signals. In this paper, we identify and study the two mismatches between pre-training and ranking fine-tuning: the training schema gap regarding the differences in training objectives and model architectures, and the task knowledge gap considering the discrepancy between the knowledge needed in ranking and that learned during pre-training. To mitigate these gaps, we propose Pre-trained, Prompt-learned and Pre-finetuned Neural Ranker (P3 Ranker). P3 Ranker leverages prompt-based learning to convert the ranking task into a pre-training like schema and uses pre-finetuning to initialize the model on intermediate supervised tasks. Experiments on MS MARCO and Robust04 show the superior performances of P3 Ranker in few-shot ranking. Analyses reveal that P3 Ranker is able to better accustom to the ranking task through prompt-based learning and retrieve necessary ranking-oriented knowledge gleaned in pre-finetuning, resulting in data-efficient PLM adaptation. Our code is available at https://github.com/NEUIR/P3Ranker.","{'model': 'tldr@v2.0.0', 'text': 'P3 Ranker is proposed, which is able to better accustom to the ranking task through prompt-based learning and retrieve necessary ranking-oriented knowledge gleaned in pre-finetuning, resulting in data-efficient PLM adaptation.'}",https://dl.acm.org/doi/pdf/10.1145/3477495.3531786
-prompt tuning with soft context sharing for vision-language models,Kun Ding,"Vision-language models have recently shown great potential on many computer vision tasks. Meanwhile, prior work demonstrates prompt tuning designed for vision-language models could acquire superior performance on few-shot image recognition compared to linear probe, a strong baseline. In real-world applications, many few-shot tasks are correlated, particularly in a specialized area. However, such information is ignored by previous work. Inspired by the fact that modeling task relationships by multi-task learning can usually boost performance, we propose a novel method SoftCPT (Soft Context Sharing for Prompt Tuning) to fine-tune pre-trained vision-language models on multiple target few-shot tasks, simultaneously. Specifically, we design a task-shared meta network to generate prompt vector for each task using pre-defined task name together with a learnable meta prompt as input. As such, the prompt vectors of all tasks will be shared in a soft manner. The parameters of this shared meta network as well as the meta prompt vector are tuned on the joint training set of all target tasks. Extensive experiments on three multi-task few-shot datasets show that SoftCPT outperforms the representative single-task prompt tuning method CoOp [78] by a large margin, implying the effectiveness of multi-task learning in vision-language prompt tuning. The source code and data will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method SoftCPT (Soft Context Sharing for Prompt Tuning) to fine-tune pre-trained vision-language models on multiple target few-shot tasks, simultaneously, and designs a task-shared meta network to generate prompt vector for each task using pre-defined task name together with a learnable meta prompt as input.'}",http://arxiv.org/pdf/2208.13474
-partslip: low-shot part segmentation for 3d point clouds via pretrained image-language models,Minghua Liu,"Generalizable 3D part segmentation is important but challenging in vision and robotics. Training deep models via conventional supervised methods requires large-scale 3D datasets with fine-grained part annotations, which are costly to collect. This paper explores an alternative way for low-shot part segmentation of 3D point clouds by leveraging a pretrained image-language model, GLIP. which achieves superior performance on open-vocabulary 2D detection. We transfer the rich knowledge from 2D to 3D through GLIP-based part detection on point cloud rendering and a novel 2D-to-3D label lifting algorithm. We also utilize multi-view 3D priors and few-shot prompt tuning to boost performance significantly. Extensive evaluation on PartNet and PartNet-Mobility datasets shows that our method enables excellent zero-shot 3D part segmentation. Our few-shot version not only outperforms existing few-shot approaches by a large margin but also achieves highly competitive results compared to the fully supervised counterpart. Furthermore, we demonstrate that our method can be directly applied to iPhone-scanned point clouds without significant domain gaps.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores an alternative way for low-shot part segmentation of 3D point clouds by leveraging a pretrained image-language model, GLIP, which achieves superior performance on open-vocabulary 2D detection and a novel 2D-to-3D label lifting algorithm.'}",https://arxiv.org/pdf/2212.01558
-prompt-tuning can be much better than fine-tuning on cross-lingual understanding with multilingual language models,Lifu Tu,"Pre-trained multilingual language models show significant performance gains for zero-shot cross-lingual model transfer on a wide range of natural language understanding (NLU) tasks. Previously, for zero-shot cross-lingual evaluation, pre-trained models are only fine-tuned on English data and tested on a variety of target languages. In this paper, we do cross-lingual evaluation on various NLU tasks (sentence classification, sequence labeling, question answering) using prompt-tuning and compare it with fine-tuning. The results show that prompt tuning achieves much better cross-lingual transfer than fine-tuning across datasets, with only 0.1% to 0.3% tuned parameters. Additionally, we demonstrate through the analysis that prompt tuning can have better cross-lingual transferability of representations on downstream tasks with better aligned decision boundaries.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2210.12360
-zero-shot image-to-image translation,Gaurav Parmar,"Large-scale text-to-image generative models have shown their remarkable ability to synthesize diverse, high-quality images. However, directly applying these models for real image editing remains challenging for two reasons. First, it is hard for users to craft a perfect text prompt depicting every visual detail in the input image. Second, while existing models can introduce desirable changes in certain regions, they often dramatically alter the input content and introduce unexpected changes in unwanted regions. In this work, we introduce pix2pix-zero, an image-to-image translation method that can preserve the original image’s content without manual prompting. We first automatically discover editing directions that reflect desired edits in the text embedding space. To preserve the content structure, we propose cross-attention guidance, which aims to retain the cross-attention maps of the input image throughout the diffusion process. Finally, to enable interactive editing, we distill the diffusion model into a fast conditional GAN. We conduct extensive experiments and show that our method outperforms existing and concurrent works for both real and synthetic image editing. In addition, our method does not need additional training for these edits and can directly use the existing pre-trained text-to-image diffusion model.","{'model': 'tldr@v2.0.0', 'text': None}",https://dl.acm.org/doi/pdf/10.1145/3588432.3591513
-few-shot anaphora resolution in scientific protocols via mixtures of in-context experts,Nghia T. Le,"Anaphora resolution is an important task for information extraction across a range of languages, text genres, and domains, motivating the need for methods that do not require large annotated datasets. In-context learning has emerged as a promising approach, yet there are a number of challenges in applying in-context learning to resolve anaphora. For example, encoding a single in-context demonstration that consists of: an anaphor, a paragraph-length context, and a list of corresponding antecedents, requires conditioning a language model on a long sequence of tokens, limiting the number of demonstrations per prompt. In this paper, we present MICE (Mixtures of In-Context Experts), which we demonstrate is effective for few-shot anaphora resolution in scientific protocols (Tamari et al., 2021). Given only a handful of training examples, MICE combines the predictions of hundreds of in-context experts, yielding a 30% increase in F1 score over a competitive prompt retrieval baseline. Furthermore, we show MICE can be used to train compact student models without sacrificing performance. As far as we are aware, this is the first work to present experimental results demonstrating the effectiveness of in-context learning on the task of few-shot anaphora resolution in scientific protocols.","{'model': 'tldr@v2.0.0', 'text': 'This is the first work to present experimental results demonstrating the effectiveness of in-context learning on the task of few-shot anaphora resolution in scientific protocols, and it is shown MICE can be used to train compact student models without sacrificing performance.'}",https://arxiv.org/pdf/2210.03690
-exploiting domain-slot related keywords description for few-shot cross-domain dialogue state tracking,QiXiang Gao,"Collecting dialogue data with domain-slot-value labels for dialogue state tracking (DST) could be a costly process. In this paper, we propose a novel framework based on domain-slot related description to tackle the challenge of few-shot cross-domain DST. Specifically, we design an extraction module to extract domain-slot related verbs and nouns in the dialogue. Then, we integrates them into the description, which aims to prompt the model to identify the slot information. Furthermore, we introduce a random sampling strategy to improve the domain generalization ability of the model. We utilize a pre-trained model to encode contexts and description and generates answers with an auto-regressive manner. Experimental results show that our approaches substantially outperform the existing few-shot DST methods on MultiWOZ and gain strong improvements on the slot accuracy comparing to existing slot description methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel framework based on domain-slot related description to tackle the challenge of few-shot cross-domain DST and introduces a random sampling strategy to improve the domain generalization ability of the model.'}",https://aclanthology.org/2022.emnlp-main.157.pdf
-decorate the examples: a simple method of prompt design for biomedical relation extraction,Hui-Syuan Yeh,"Relation extraction is a core problem for natural language processing in the biomedical domain. Recent research on relation extraction showed that prompt-based learning improves the performance on both fine-tuning on full training set and few-shot training. However, less effort has been made on domain-specific tasks where good prompt design can be even harder. In this paper, we investigate prompting for biomedical relation extraction, with experiments on the ChemProt dataset. We present a simple yet effective method to systematically generate comprehensive prompts that reformulate the relation extraction task as a cloze-test task under a simple prompt formulation. In particular, we experiment with different ranking scores for prompt selection. With BioMed-RoBERTa-base, our results show that prompting-based fine-tuning obtains gains by 14.21 F1 over its regular fine-tuning baseline, and 1.14 F1 over SciFive-Large, the current state-of-the-art on ChemProt. Besides, we find prompt-based learning requires fewer training examples to make reasonable predictions. The results demonstrate the potential of our methods in such a domain-specific relation extraction task.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a simple yet effective method to systematically generate comprehensive prompts that reformulate the relation extraction task as a cloze-test task under a simple prompt formulation, and finds prompt-based learning requires fewer training examples to make reasonable predictions.'}",http://arxiv.org/pdf/2204.10360
-slot dependency modeling for zero-shot cross-domain dialogue state tracking,Qingyue Wang,"Zero-shot learning for Dialogue State Tracking (DST) focuses on generalizing to an unseen domain without the expense of collecting in domain data. However, previous zero-shot DST methods ignore the slot dependencies in a multidomain dialogue, resulting in sub-optimal performances when adapting to unseen domains. In this paper, we utilize slot prompts combination, slot values demonstration, and slot constraint object to model the slot-slot dependencies, slot-value dependency and slot-context dependency respectively. Specifically, each slot prompt consists of a slot-specific prompt and a slot-shared prompt to capture the shared knowledge across different domains. Experimental results show the effectiveness of our proposed method over existing state-of-art generation methods under zero-shot/few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show the effectiveness of the proposed method over existing state-of-art generation methods under zero-shot/few-shot settings.'}",
-pre-trained language models can be fully zero-shot learners,Xuandong Zhao,"How can we extend a pre-trained model to many language understanding tasks, without labeled or additional unlabeled data? Pre-trained language models (PLMs) have been effective for a wide range of NLP tasks. However, existing approaches either require fine-tuning on downstream labeled datasets or manually constructing proper prompts. In this paper, we propose nonparametric prompting PLM (NPPrompt) for fully zero-shot language understanding. Unlike previous methods, NPPrompt uses only pre-trained language models and does not require any labeled data or additional raw corpus for further fine-tuning, nor does it rely on humans to construct a comprehensive set of prompt label words. We evaluate NPPrompt against previous major few-shot and zero-shot learning methods on diverse NLP tasks: including text classification, text entailment, similar text retrieval, paraphrasing, and multiple-choice question answering. Experimental results demonstrate that our NPPrompt outperforms the previous best fully zero-shot method by big margins, with absolute gains of 12.8% in accuracy on text classification and 15.6% on the GLUE benchmark. Our source code is available at https://anonymous.4open. science/r/NPPrompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes nonparametric prompting PLM (NPPrompt) for fully zero-shot language understanding, which uses only pre-trained language models and does not require any labeled data or additional raw corpus for further fine-tuning, nor does it rely on humans to construct a comprehensive set of prompt label words.'}",http://arxiv.org/pdf/2212.06950
-multitask prompted training enables zero-shot task generalization,Victor Sanh,"Large language models have recently been shown to attain reasonable zero-shot generalization on a diverse set of tasks (Brown et al., 2020). It has been hypothesized that this is a consequence of implicit multitask learning in language models' pretraining (Radford et al., 2019). Can zero-shot generalization instead be directly induced by explicit multitask learning? To test this question at scale, we develop a system for easily mapping any natural language tasks into a human-readable prompted form. We convert a large set of supervised datasets, each with multiple prompts with diverse wording. These prompted datasets allow for benchmarking the ability of a model to perform completely held-out tasks. We fine-tune a pretrained encoder-decoder model (Raffel et al., 2020; Lester et al., 2021) on this multitask mixture covering a wide variety of tasks. The model attains strong zero-shot performance on several standard datasets, often outperforming models up to 16x its size. Further, our approach attains strong performance on a subset of tasks from the BIG-bench benchmark, outperforming models up to 6x its size. All trained models are available at https://github.com/bigscience-workshop/t-zero and all prompts are available at https://github.com/bigscience-workshop/promptsource.","{'model': 'tldr@v2.0.0', 'text': 'A system for easily mapping any natural language tasks into a human-readable prompted form and fine-tune a pretrained encoder-decoder model on this multitask mixture covering a wide variety of tasks.'}",
-tess: zero-shot classification via textual similarity comparison with prompting using sentence encoder,Jimin Hong,"We introduce TeSS ( Te xt S imilarity Comparison using S entence Encoder), a framework for zero-shot classification where the assigned label is determined by the embedding similarity between the input text and each candidate label prompt. We leverage representations from sentence encoders optimized to locate semantically similar samples closer to each other in embedding space during pretraining. The label prompt embeddings serve as prototypes of their corresponding class clus-ters. Furthermore, to compensate for the po-tentially poorly descriptive labels in their original format, we retrieve semantically similar sentences from external corpora and additionally use them with the original label prompt (TeSS-R). TeSS outperforms strong baselines on various closed-set and open-set classification datasets under zero-shot setting, with further gains when combined with label prompt diversification through retrieval. These results are robustly attained to verbalizer variations, an ancillary benefit of using a bi-encoder. Al-together, our method serves as a reliable baseline for zero-shot classification and a simple interface to assess the quality of sentence encoders 1 .","{'model': 'tldr@v2.0.0', 'text': 'TeSS is introduced, a framework for zero-shot classification where the assigned label is determined by the embedding similarity between the input text and each candidate label prompt, and a simple interface to assess the quality of sentence encoders.'}",http://arxiv.org/pdf/2212.10391
-zero-shot program representation learning,Nan Cui,"Learning program representations has been the core prerequisite of code intelligence tasks (e.g., code search and code clone detection). The state-of-the-art pre-trained models such as CodeBERT require the availability of large-scale code corpora. However, gathering training samples can be costly and infeasible for domain-specific languages such as Solidity for smart contracts. In this paper, we propose Zecoler, a zero-shot learning approach for code representations. Zecoler is built upon a pre-trained programming language model. In order to elicit knowledge from the pre-trained models efficiently, Zecoler casts the downstream tasks to the same form of pre-training tasks by inserting trainable prompts into the original input. Then, it employs the prompt learning technique to optimize the pre-trained model by merely adjusting the original input. This enables the representation model to efficiently fit the scarce task-specific data while reusing pre-trained knowledge. We evaluate Zecoler in three code intelligence tasks in two programming languages that have no training samples, namely, Solidity and Go, with model trained in corpora of common languages such as Java. Experimental results show that our approach significantly outperforms baseline models in both zero-shot and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Zecoler is a zero-shot learning approach for code representations built upon a pre-trained programming language model that significantly outperforms baseline models in both zero- shot and few-shot settings.'}",https://dl.acm.org/doi/pdf/10.1145/3524610.3527888
-queryform: a simple zero-shot form entity query framework,Zifeng Wang,"Zero-shot transfer learning for document understanding is a crucial yet under-investigated scenario to help reduce the high cost involved in annotating document entities. We present a novel query-based framework, QueryForm, that extracts entity values from form-like documents in a zero-shot fashion. QueryForm contains a dual prompting mechanism that composes both the document schema and a specific entity type into a query, which is used to prompt a Transformer model to perform a single entity extraction task. Furthermore, we propose to leverage large-scale query-entity pairs generated from form-like webpages with weak HTML annotations to pre-train QueryForm. By unifying pre-training and fine-tuning into the same query-based framework, QueryForm enables models to learn from structured documents containing various entities and layouts, leading to better generalization to target document types without the need for target-specific training data. QueryForm sets new state-of-the-art average F1 score on both the XFUND (+4.6%~10.1%) and the Payment (+3.2%~9.5%) zero-shot benchmark, with a smaller model size and no additional image input.","{'model': 'tldr@v2.0.0', 'text': 'By unifying pre-training and fine-tuning into the same query-based framework, QueryForm enables models to learn from structured documents containing various entities and layouts, leading to better generalization to target document types without the need for target-specific training data.'}",http://arxiv.org/pdf/2211.07730
-what language model architecture and pretraining objective work best for zero-shot generalization?,Thomas Wang,"Large pretrained Transformer language models have been shown to exhibit zero-shot generalization, i.e. they can perform a wide variety of tasks that they were not explicitly trained on. However, the architectures and pretraining objectives used across state-of-the-art models differ significantly, and there has been limited systematic comparison of these factors. In this work, we present a large-scale evaluation of modeling choices and their impact on zero-shot generalization. In particular, we focus on text-to-text models and experiment with three model architectures (causal/non-causal decoder-only and encoder-decoder), trained with two different pretraining objectives (autoregressive and masked language modeling), and evaluated with and without multitask prompted finetuning. We train models with over 5 billion parameters for more than 170 billion tokens, thereby increasing the likelihood that our conclusions will transfer to even larger scales. Our experiments show that causal decoder-only models trained on an autoregressive language modeling objective exhibit the strongest zero-shot generalization after purely unsupervised pretraining. However, models with non-causal visibility on their input trained with a masked language modeling objective followed by multitask finetuning perform the best among our experiments. We therefore consider the adaptation of pretrained models across architectures and objectives. We find that pretrained non-causal decoder models can be adapted into performant generative causal decoder models, using autoregressive language modeling as a downstream task. Furthermore, we find that pretrained causal decoder models can be efficiently adapted into non-causal decoder models, ultimately achieving competitive performance after multitask finetuning. Code and checkpoints are available at https://github.com/bigscience-workshop/architecture-objective.","{'model': 'tldr@v2.0.0', 'text': 'A large-scale evaluation of modeling choices and their impact on zero-shot generalization finds that pretrained non-causal decoder models can be adapted into performant generative causal decoding models, using autoregressive language modeling as a downstream task.'}",http://arxiv.org/pdf/2204.05832
-generating variable explanations via zero-shot prompt learning,Chong Wang,"As basic elements in program, variables convey essential information that is critical for program comprehension and maintenance. However, understanding the meanings of variables in program is not always easy for developers, since poor-quality variable names are prevalent while such variable are less informative for program comprehension. Therefore, in this paper, we target at generating concise natural language explanations for variables to facilitate program comprehension. In particular, there are two challenges in variable explanation generation, including the lack of training data and the association with complex code contexts around the variable. To address these issues, we propose a novel approach ZeroVar,which leverages code pre-trained models and zero-shot prompt learning to generate explanations for the variable based on its code context. ZeroVarcontains two stages: (i) a pre-training stage that continually pre-trains a base model (i.e., CodeT5) to recover the randomly-masked parameter descriptions in method docstrings; and (ii) a zero-shot prompt learning stage that leverages the pre-trained model to generate explanations for a given variable via the prompt constructed with the variable and its belonging method context. We then extensively evaluate the quality and usefulness of the variable explanations generated by ZeroVar.We construct an evaluation dataset of 773 variables and their reference explanations. Our results show that ZeroVarcan generate higher-quality explanations than baselines, not only on automated metrics such as BLEU and ROUGE, but also on human metrics such as correctness, completeness, and conciseness. Moreover, we further assess the usefulness of ZeroVAR-generated explanations on two downstream tasks related to variable naming quality, i.e., abbreviation expansion and spelling correction. For abbreviation expansion, the generated variable explanations can help improve the present rate (+13.1%), precision (+3.6%), and recall (+10.0%) of the state-of-the-art abbreviation explanation approach. For spelling correction, by using the generated explanations we can achieve higher hit@1 (+162.9(%) and hit@3 (+49.6%) than the recent variable representation learning approach.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel approach ZeroVar, which leverages code pre-trained models and zero-shot prompt learning to generate explanations for the variable based on its code context, and assess the quality and usefulness of the variable explanations generated by ZeroVar.'}",
-prompt gating: a parameter efficient tuning method for zero-shot multi-source translation,Xuancheng Huang,"Multi-source translation (MST), which typi-cally receives multiple source sentences of the same meaning in different languages, has been shown superior to single-source translation. As the quantity of multi-source parallel data is limited, taking full advantage of single-source data and limited multi-source data to make models perform well when receiving as many as possible sources remains a challenge. Unlike previous work mostly devoted to supervised scenarios, we focus on zero-shot MST: expecting models to be able to process unseen combinations of multiple sources, e.g., unseen language combinations, during inference. We propose a simple yet effective parameter efficient method, named Prompt Gating, which appends prompts to the model inputs and attaches gates on the extended hidden states for each encoder layer. It shows strong zero-shot transferability (+9.0 BLEU points maximally) and remarkable compositionality (+15.6 BLEU points maximally) on MST, and also shows its superiorities over baselines on lexically constrained translation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple yet effective parameter efficient method, named Prompt Gating, which appends prompts to the model inputs and attaches gates on the extended hidden states for each encoder layer on MST, which shows strong zero-shot transferability and remarkable compositionality.'}",http://arxiv.org/pdf/2212.09387
-peinet: joint prompt and evidence inference network via language family policy for zero-shot multilingual fact checking,Xiaoyu Li,"Zero-shot multilingual fact-checking, which aims to discover and infer subtle clues from the retrieved relevant evidence to verify the given claim in cross-language and cross-domain scenarios, is crucial for optimizing a free, trusted, wholesome global network environment. Previous works have made enlightening and practical explorations in claim verification, while the zero-shot multilingual task faces new challenging gap issues: neglecting authenticity-dependent learning between multilingual claims, lacking heuristic checking, and a bottleneck of insufficient evidence. To alleviate these gaps, a novel Joint Prompt and Evidence Inference Network (PEINet) is proposed to verify the multilingual claim according to the human fact-checking cognitive paradigm. In detail, firstly, we leverage the language family encoding mechanism to strengthen knowledge transfer among multi-language claims. Then, the prompt turning module is designed to infer the falsity of the fact, and further, sufficient fine-grained evidence is extracted and aggregated based on a recursive graph attention network to verify the claim again. Finally, we build a unified inference framework via multi-task learning for final fact verification. The newly achieved state-of-the-art performance on the released challenging benchmark dataset that includes not only an out-of-domain test, but also a zero-shot test, proves the effectiveness of our framework, and further analysis demonstrates the superiority of our PEINet in multilingual claim verification and inference, especially in the zero-shot scenario.","{'model': 'tldr@v2.0.0', 'text': 'A novel Joint Prompt and Evidence Inference Network (PEINet) is proposed to verify the multilingual claim according to the human fact-checking cognitive paradigm and builds a unified inference framework via multi-task learning for final fact verification.'}",https://www.mdpi.com/2076-3417/12/19/9688/pdf?version=1664345340
-an exploration of prompt-based zero-shot relation extraction method,Zhao Jun,,"{'model': 'tldr@v2.0.0', 'text': 'This work views zero-shot relation extraction as a semantic matching task optimized by prompt-tuning, which still maintains superior generalization performance when the labeled data of predefined relations are extremely scarce, and model the semantic interaction between relational instances and their descriptions directly during encoding.'}",
-prompt-guided scene generation for 3d zero-shot learning,Majid Nasiri,"Zero-shot learning on 3D point cloud data is a related underexplored problem compared to its 2D image counterpart. 3D data brings new challenges for ZSL due to the unavailability of robust pre-trained feature extraction models. To address this problem, we propose a prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex interplay of seen and unseen objects. First, we merge point clouds of two 3D models in certain ways described by a prompt. The prompt acts like the annotation describing each 3D scene. Later, we perform contrastive learning to train our proposed architecture in an end-to-end manner. We argue that 3D scenes can relate objects more efficiently than single objects because popular language models (like BERT) can achieve high performance when objects appear in a context. Our proposed prompt-guided scene generation method encapsulates data augmentation and prompt-based annotation/captioning to improve 3D ZSL performance. We have achieved state-of-the-art ZSL and generalized ZSL performance on synthetic (ModelNet40, ModelNet10) and real-scanned (ScanOjbectNN) 3D object datasets.","{'model': 'tldr@v2.0.0', 'text': 'A prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex interplay of seen and unseen objects.'}",https://arxiv.org/pdf/2209.14690
-a topic-based prompt learning method for zero-shot stance detection,Ruifeng Xu,"Zero-shot stance detection (ZSSD) aims to detecting the stance of previously unseen targets during the inference stage. It is generally believed that the stance expression in a sentence is closely related to the stance target and topics discussed.We divide stance expressions of speakers into two categories: target-invariant and target-specific categories. Target-invariant stance expressions carry the same stance polarity regardless of the targets they are associated with. On the contrary, target-specific stance expressions only co-occur with certain targets. As such, it is important to distinguish these two types of stance features to boost stance detection ability.In this paper,we develop an effective approach to distinguish the types of target-related stance expressions to better","{'model': 'tldr@v2.0.0', 'text': 'An effective approach to distinguish the types of target-related stance expressions to better zero-shot stance detection ability is developed.'}",
-prompt-based zero-shot relation extraction with semantic knowledge augmentation,Jiaying Gong,"In relation triplet extraction (RTE), recognizing unseen (new) relations for which there are no training instances is a challenging task. Efforts have been made to recognize unseen relations based on question-answering models or relation descriptions. However, these approaches miss the semantic information about connections between seen and unseen relations. In this paper, We propose a prompt-based model with semantic knowledge augmentation (ZS-SKA) to recognize unseen relations under the zero-shot setting. We present a new word-level analogy-based sentence translation rule and generate augmented instances with unseen relations from instances with seen relations using that new rule. We design prompts with weighted virtual label construction based on an external knowledge graph to integrate semantic knowledge information learned from seen relations. Instead of using the actual label sets in the prompt template, we construct weighted virtual label words. We learn the representations of both seen and unseen relations with augmented instances and prompts. We then calculate the distance between the generated representations using prototypical networks to predict unseen relations. Extensive experiments conducted on three public datasets FewRel, Wiki-ZSL, and NYT, show that ZS-SKA outperforms state-of-the-art methods under the zero-shot scenarios. Our experimental results also demonstrate the effectiveness and robustness of ZS-SKA.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a new word-level analogy-based sentence translation rule and generates augmented instances with unseen relations from instances with seen relations using that new rule, and designs prompts with weighted virtual label construction based on an external knowledge graph to integrate semantic knowledge information learned from seen relations.'}",
-from visual prompt learning to zero-shot transfer: mapping is all you need,Ziqing Yang,"Visual prompt learning, as a newly emerged technique, leverages the knowledge learned by a large-scale pre-trained model and adapts it to downstream tasks through the usage of prompts. While previous research has focused on designing effective prompts, in this work, we argue that compared to prompt design, a good mapping strategy matters more. In this sense, we propose SeMap, a more effective mapping using the semantic alignment between the pre-trained model's knowledge and the downstream task. Our experimental results show that SeMap can largely boost the performance of visual prompt learning. Moreover, our experiments show that SeMap is capable of achieving competitive zero-shot transfer, indicating that it can perform the downstream task without any fine-tuning on the corresponding dataset. This demonstrates the potential of our proposed method to be used in a broader range of applications where the zero-shot transfer is desired. Results suggest that our proposed SeMap could lead to significant advancements in both visual prompt learning and zero-shot transfer. We hope with SeMap, we can help the community move forward to more efficient and lightweight utilization of large vision models.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that SeMap is capable of achieving competitive zero-shot transfer, indicating that it can perform the downstream task without any fine-tuning on the corresponding dataset, demonstrating the potential of the method to be used in a broader range of applications where the zero- shot transfer is desired.'}",http://arxiv.org/pdf/2303.05266
-layout and task aware instruction prompt for zero-shot document image question answering,Wenjin Wang,"Layout-aware pre-trained models has achieved significant progress on document image question answering. They introduce extra learnable modules into existing language models to capture layout information within document images from text bounding box coordinates obtained by OCR tools. However, extra modules necessitate pre-training on extensive document images. This prevents these methods from directly utilizing off-the-shelf instruction-tuning language foundation models, which have recently shown promising potential in zero-shot learning. Instead, in this paper, we find that instruction-tuning language models like Claude and ChatGPT can understand layout by spaces and line breaks. Based on this observation, we propose the LAyout and Task aware Instruction Prompt (LATIN-Prompt), which consists of layout-aware document content and task-aware instruction. Specifically, the former uses appropriate spaces and line breaks to recover the layout information among text segments obtained by OCR tools, and the latter ensures that generated answers adhere to formatting requirements. Moreover, we propose the LAyout and Task aware Instruction Tuning (LATIN-Tuning) to improve the performance of small instruction-tuning models like Alpaca. Experimental results show that LATIN-Prompt enables zero-shot performance of Claude and ChatGPT to be comparable to the fine-tuning performance of SOTAs on document image question answering, and LATIN-Tuning enhances the zero-shot performance of Alpaca significantly. For example, LATIN-Prompt improves the performance of Claude and ChatGPT on DocVQA by 263% and 20% respectively. LATIN-Tuning improves the performance of Alpaca on DocVQA by 87.7%. Quantitative and qualitative analyses demonstrate the effectiveness of LATIN-Prompt and LATIN-Tuning. We provide the code in supplementary and will release it to facilitate future research.","{'model': 'tldr@v2.0.0', 'text': 'This paper finds that instruction-tuning language models like Claude and ChatGPT can understand layout by spaces and line breaks, and proposes the LAyout and Task aware Instruction Tuning (LATIN-Tuning), which consists of layout-aware document content and task-aware instruction.'}",https://arxiv.org/pdf/2306.00526
-navigating prompt complexity for zero-shot classification: a study of large language models in computational social science,Yida Mu,"Instruction-tuned Large Language Models (LLMs) have exhibited impressive language understanding and the capacity to generate responses that follow specific prompts. However, due to the computational demands associated with training these models, their applications often adopt a zero-shot setting. In this paper, we evaluate the zero-shot performance of two publicly accessible LLMs, ChatGPT and OpenAssistant, in the context of six Computational Social Science classification tasks, while also investigating the effects of various prompting strategies. Our experiments investigate the impact of prompt complexity, including the effect of incorporating label definitions into the prompt; use of synonyms for label names; and the influence of integrating past memories during foundation model training. The findings indicate that in a zero-shot setting, current LLMs are unable to match the performance of smaller, fine-tuned baseline transformer models (such as BERT-large). Additionally, we find that different prompting strategies can significantly affect classification accuracy, with variations in accuracy and F1 scores exceeding 10\%.","{'model': 'tldr@v2.0.0', 'text': 'The findings indicate that in a zero-shot setting, current LLMs are unable to match the performance of smaller, fine-tuned baseline transformer models (such as BERT-large), and different prompting strategies can significantly affect classification accuracy.'}",https://arxiv.org/pdf/2305.14310
-injecting commonsense knowledge into prompt learning for zero-shot text classification,Jing Qian,"The combination of pre-training and fine-tuning has become a default solution to Natural Language Processing (NLP) tasks. The emergence of prompt learning breaks such routine, especially in the scenarios of low data resources. Insufficient labelled data or even unseen classes are frequent problems in text classification, equipping Pre-trained Language Models (PLMs) with task-specific prompts helps get rid of the dilemma. However, general PLMs are barely provided with commonsense knowledge. In this work, we propose a KG-driven verbalizer that leverages commonsense Knowledge Graph (KG) to map label words with predefined classes. Specifically, we transform the mapping relationships into semantic relevance in the commonsense-injected embedding space. For zero-shot text classification task, experimental results exhibit the effectiveness of our KG-driven verbalizer on a Twitter dataset for natural disasters (i.e. HumAID) compared with other baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a KG-driven verbalizer that leverages commonsense Knowledge Graph (KG) to map label words with predefined classes and transforms the mapping relationships into semantic relevance in the commonsense-injected embedding space.'}",
-knowledge-embedded prompt learning for zero-shot social media text classification,Jingyi Li,"Social media plays an irreplaceable role in shaping the way information is created shared and consumed. While it provides access to a vast amount of data, extracting and analyzing useful insights from complex and dynamic social media data can be challenging. Deep learning models have shown promise in social media analysis tasks, but such models require a massive amount of labelled data which is usually unavailable in real-world settings. Additionally, these models lack common-sense knowledge which can limit their ability to generate comprehensive results. To address these challenges, we propose a knowledge-embedded prompt learning model for zero-shot social media text classification. Our experimental results on four social media datasets demonstrate that our proposed approach outperforms other well-known baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a knowledge-embedded prompt learning model for zero-shot social media text classification that outperforms other well-known baselines and shows superior results on four social media datasets.'}",
-zero-shot continuous prompt transfer: generalizing task semantics across language models,Zijun Wu,"Prompt tuning in natural language processing (NLP) has become an increasingly popular method for adapting large language models to specific tasks. However, the transferability of these prompts, especially continuous prompts, between different models remains a challenge. In this work, we propose a zero-shot continuous prompt transfer method, where source prompts are encoded into relative space and the corresponding target prompts are searched for transferring to target models. Experimental results confirm the effectiveness of our method, showing that 'task semantics' in continuous prompts can be generalized across various language models. Moreover, we find that combining 'task semantics' from multiple source models can further enhance the generalizability of transfer.","{'model': 'tldr@v2.0.0', 'text': ""A zero-shot continuous prompt transfer method is proposed, where source prompts are encoded into relative space and the corresponding target prompts are searched for transferring to target models, showing that 'task semantics' in continuous prompts can be generalized across various language models.""}",https://arxiv.org/pdf/2310.01691
-spteae: a soft prompt transfer model for zero-shot cross-lingual event argument extraction,Huipeng Ma,"In zero-shot cross-lingual event argument extraction(EAE) task, a model is typically trained on source language datasets and then applied on task language datasets. There is a trend to regard the zero-shot cross-lingual EAE task as a sequence generation task with manual prompts or discrete prompts. However, there are some problems with these prompts, including using suboptimal prompts and difficult to transfer from source language to target language. To overcome these issues, we propose a method called SPTEAE(A Soft Prompt Transfer model for zero-shot cross-lingual Event Argument Extraction). SPTEAE utilizes a sequence of tunable vectors which are tuned in source language as event type prompts. These source language event type prompts can be transferred as target prompts to perform target EAE task by key-value selection mechanism. For each event type, SPTEAE learns a special target prompt by attending to highly relevant source prompts. Experiment results show that the average performance of SPTEAE with soft prompt transfer is 2.6% higher than the current state-of-the-art model on the ACE2005 dataset.","{'model': 'tldr@v2.0.0', 'text': 'A Soft Prompt Transfer model for zero-shot cross-lingual Event Argument Extraction, which utilizes a sequence of tunable vectors which are tuned in source language as event type prompts and learns a special target prompt by attending to highly relevant source prompts.'}",
-prompt-based zero-shot text classification with conceptual knowledge,Yuqi Wang,"The proposed framework incorporates conceptual knowledge for prompt-based text classification in the extreme zero-shot setting, which outperforms existing approaches in sentiment analysis and topic detection on four widely-used datasets.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2023.acl-srw.4.pdf
-prompt-ner: zero-shot named entity recognition in astronomy literature via large language models,Wujun Shao,"This study delves into the application of Large Language Models (LLMs) for Named Entity Recognition (NER) tasks in the field of astronomy literature. To enhance the zero-shot recognition capabilities of LLMs for astronomical named entities, we propose a strategy called Prompt-NER. Prompt-NER includes five prompt elements: Task Descriptions, Entity Definitions, Task Emphasis, Task Examples, and Second Conversation. To assess the effectiveness of the Prompt-NER strategy, we utilize three representative LLMs (Claude-2, GPT-3.5, and LLaMA-2-70b) to identify telescope and celestial object named entities in astronomical literature. Our experiments are conducted based on two distinct datasets. The first dataset comprises 30 original PDF documents, which we split into paragraphs in sequential order, resulting in a second dataset consisting of 30 paragraph collections. Additionally, we incorporate 30 astronomical telegrams to diversify our experiments and assess the performance of LLMs based on Prompt-NER on concise, complete texts. Our experimental results indicate that the Prompt-NER strategy enables LLMs to effectively accomplish NER tasks in the field of astronomy, even without prior astronomical knowledge during training. We carefully analyze the experimental results, including the mechanism of different prompt elements and the influence of different features of long and short texts on their respective experimental results. This research provides experience for zero-shot NER tasks in astronomical literature and suggests future work in this area.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results indicate that the Prompt-NER strategy enables LLMs to effectively accomplish NER tasks in the field of astronomy, even without prior astronomical knowledge during training.'}",
-"synthesize, prompt and transfer: zero-shot conversational question generation with pre-trained language model",Hongwei Zeng,"Conversational question generation aims to generate questions that depend on both context and conversation history. Conventional works utilizing deep learning have shown promising results, but heavily rely on the availability of large-scale annotated conversations. In this paper, we introduce a more realistic and less explored setting, Zero-shot Conversational Question Generation (ZeroCQG), which requires no human-labeled conversations for training. To solve ZeroCQG, we propose a multi-stage knowledge transfer framework, Synthesize, Prompt, and trAnsfer with pRe-Trained lAnguage model (SPARTA) to effectively leverage knowledge from single-turn question generation instances. To validate the zero-shot performance of SPARTA, we conduct extensive experiments on three conversational datasets: CoQA, QuAC, and DoQA by transferring knowledge from three single-turn datasets: MS MARCO, NewsQA, and SQuAD. The experimental results demonstrate the superior performance of our method. Specifically, SPARTA has achieved 14.81 BLEU-4 (88.2% absolute improvement compared to T5) in CoQA with knowledge transferred from SQuAD.","{'model': 'tldr@v2.0.0', 'text': 'A multi-stage knowledge transfer framework, Synthesize, Prompt, and trAnsfer with pRe-Trained lAnguage model (SPARTA) to effectively leverage knowledge from single-turn question generation instances to solve ZeroCQG.'}",https://aclanthology.org/2023.acl-long.500.pdf
-weakly supervised few-shot and zero-shot semantic segmentation with mean instance aware prompt learning,Prashant Pandey,"The existing Zero-Shot Segmentation (ZSS) and Few-Shot Segmentation (FSS) methods utilize fully supervised pixel-labeled seen classes to segment unseen classes. Pixel-level labels are hard to obtain, and using weak supervision in the form of inexpensive image labels is often more practical. To this end, we propose a novel unified weakly supervised Zero-Shot and Few-Shot semantic segmentation pipeline that can perform ZSS and FSS on novel classes without using pixel-level labels for either the base (seen) or the novel (unseen) classes. We propose Mean Instance Aware Prompt based Network (MIAPNet), a novel language-guided segmentation pipeline that i) learns context vectors with batch aggregates (mean) to map class prompts to image features and ii) decouples weak ZSS/FSS into weak semantic segmentation and Zero-Shot segmentation. MIAPNet beats existing methods for weak generalized ZSS and weak FSS by 39 and 3 mIOU points respectively on PASCAL VOC and weak FSS by 5 mIOU points on MS COCO.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Mean Instance Aware Prompt based Network (MIAPNet), a novel language-guided segmentation pipeline that learns context vectors with batch aggregates to map class prompts to image features and decouples weak ZSS/FSS into weak semantic segmentation and Zero-Shot segmentation.'}",
-"entities, dates, and languages: zero-shot on historical texts with t0",F. Toni,"In this work, we explore whether the recently demonstrated zero-shot abilities of the T0 model extend to Named Entity Recognition for out-of-distribution languages and time periods. Using a historical newspaper corpus in 3 languages as test-bed, we use prompts to extract possible named entities. Our results show that a naive approach for prompt-based zero-shot multilingual Named Entity Recognition is error-prone, but highlights the potential of such an approach for historical languages lacking labeled datasets. Moreover, we also find that T0-like models can be probed to predict the publication date and language of a document, which could be very relevant for the study of historical texts.","{'model': 'tldr@v2.0.0', 'text': 'This work explores whether the recently demonstrated zero-shot abilities of the T0 model extend to Named Entity Recognition for out-of-distribution languages and time periods, and uses a historical newspaper corpus in 3 languages as test-bed.'}",http://arxiv.org/pdf/2204.05211
-pesco: prompt-enhanced self contrastive learning for zero-shot text classification,Yau-Shian Wang,"We present PESCO, a novel contrastive learning framework that substantially improves the performance of zero-shot text classification. We formulate text classification as a neural text retrieval problem where each document is treated as a query, and the system learns the mapping from each query to the relevant class labels by (1) adding prompts to enhance label retrieval, and (2) using retrieved labels to enrich the training set in a self-training loop of contrastive learning. PESCO achieves state-of-the-art performance on four benchmark text classification datasets. On DBpedia, we achieve 98.5% accuracy without any labeled data, which is close to the fully-supervised result. Extensive experiments and analyses show all the components of PESCO are necessary for improving the performance of zero-shot text classification.","{'model': 'tldr@v2.0.0', 'text': 'PESCO is presented, a novel contrastive learning framework that substantially improves the performance of zero-shot text classification and achieves state-of-the-art performance on four benchmark text classification datasets.'}",http://arxiv.org/pdf/2305.14963
-prompt to be consistent is better than self-consistent? few-shot and zero-shot fact verification with pre-trained language models,Fengzhu Zeng,"Few-shot or zero-shot fact verification only relies on a few or no labeled training examples. In this paper, we propose a novel method called ProToCo, to \underline{Pro}mpt pre-trained language models (PLMs) \underline{To} be \underline{Co}nsistent, for improving the factuality assessment capability of PLMs in the few-shot and zero-shot settings. Given a claim-evidence pair, ProToCo generates multiple variants of the claim with different relations and frames a simple consistency mechanism as constraints for making compatible predictions across these variants. We update PLMs by using parameter-efficient fine-tuning (PEFT), leading to more accurate predictions in few-shot and zero-shot fact verification tasks. Our experiments on three public verification datasets show that ProToCo significantly outperforms state-of-the-art few-shot fact verification baselines. With a small number of unlabeled instances, ProToCo also outperforms the strong zero-shot learner T0 on zero-shot verification. Compared to large PLMs using in-context learning (ICL) method, ProToCo outperforms OPT-30B and the Self-Consistency-enabled OPT-6.7B model in both few- and zero-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel method called ProToCo, to improve the factuality assessment capability of pre-trained language models (PLMs) by using parameter-efficient fine-tuning (PEFT), leading to more accurate predictions in few-shot and zero-shot fact verification tasks.'}",http://arxiv.org/pdf/2306.02569
-hierarchical prompt learning for compositional zero-shot recognition,Henan Wang,"Compositional Zero-Shot Learning (CZSL) aims to imitate the powerful generalization ability of human beings to recognize novel compositions of known primitive concepts that correspond to a state and an object, e.g., purple apple. To fully capture the intra- and inter-class correlations between compositional concepts, in this paper, we propose to learn them in a hierarchical manner. Specifically, we set up three hierarchical embedding spaces that respectively model the states, the objects, and their compositions, which serve as three “experts” that can be combined in inference for more accurate predictions. We achieve this based on the recent success of large-scale pretrained vision-language models, e.g., CLIP, which provides a strong initial knowledge of image-text relationships. To better adapt this knowledge to CZSL, we propose to learn three hierarchical prompts by explicitly fixing the unrelated word tokens in the three embedding spaces. Despite its simplicity, our proposed method consistently yields superior performance over current state-of-the-art approaches on three widely-used CZSL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper sets up three hierarchical embedding spaces that respectively model the states, the objects, and their compositions, which serve as three “experts” that can be combined in inference for more accurate predictions.'}",https://www.ijcai.org/proceedings/2023/0163.pdf
-anomalyclip: object-agnostic prompt learning for zero-shot anomaly detection,Qihang Zhou,"Zero-shot anomaly detection (ZSAD) requires detection models trained using auxiliary data to detect anomalies without any training sample in a target dataset. It is a crucial task when training data is not accessible due to various concerns, \eg, data privacy, yet it is challenging since the models need to generalize to anomalies across different domains where the appearance of foreground objects, abnormal regions, and background features, such as defects/tumors on different products/organs, can vary significantly. Recently large pre-trained vision-language models (VLMs), such as CLIP, have demonstrated strong zero-shot recognition ability in various vision tasks, including anomaly detection. However, their ZSAD performance is weak since the VLMs focus more on modeling the class semantics of the foreground objects rather than the abnormality/normality in the images. In this paper we introduce a novel approach, namely AnomalyCLIP, to adapt CLIP for accurate ZSAD across different domains. The key insight of AnomalyCLIP is to learn object-agnostic text prompts that capture generic normality and abnormality in an image regardless of its foreground objects. This allows our model to focus on the abnormal image regions rather than the object semantics, enabling generalized normality and abnormality recognition on diverse types of objects. Large-scale experiments on 17 real-world anomaly detection datasets show that AnomalyCLIP achieves superior zero-shot performance of detecting and segmenting anomalies in datasets of highly diverse class semantics from various defect inspection and medical imaging domains. Code will be made available at https://github.com/zqhang/AnomalyCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach, namely AnomalyCLIP, to adapt CLIP for accurate ZSAD across different domains, to learn object-agnostic text prompts that capture generic normality and abnormality in an image regardless of its foreground objects.'}",
-enhancing zero-shot crypto sentiment with fine-tuned language model and prompt engineering,Rahman S M Wahidur,"Blockchain technology has revolutionized the financial landscape, with cryptocurrencies gaining widespread adoption for their decentralized and transparent nature. As the sentiment expressed on social media platforms can significantly influence cryptocurrency discussions and market movements, sentiment analysis has emerged as a crucial tool for understanding public opinion and predicting market trends. Motivated by the aim to enhance sentiment analysis accuracy in the cryptocurrency domain, this paper investigates fine-tuning techniques on large language models. This paper also investigates the efficacy of supervised fine-tuning and instruction-based fine-tuning on large language models for unseen tasks. Experimental results demonstrate a significant average zero-shot performance gain of 40% after fine-tuning, highlighting the potential of this technique in optimizing pre-trained language model efficiency. Additionally, the impact of instruction tuning on models of varying scales is examined, revealing that larger models benefit from instruction tuning, achieving the highest average accuracy score of 75.16%. In contrast, smaller-scale models may experience reduced generalization due to the complete utilization of model capacity. To gain deeper insight about how instruction works with these language models, this paper presents an experimental investigation into the response of an instruction-based model under different instruction tuning setups. The investigation demonstrates that the model achieves an average accuracy score of 72.38% for short and simple instructions. This performance significantly outperforms its accuracy under long and complex instructions by over 12%, thereby effectively highlighting the profound significance of instruction characteristics in maximizing model performance.","{'model': 'tldr@v2.0.0', 'text': 'An experimental investigation into the response of an instruction-based model under different instruction tuning setups is presented, demonstrating a significant average zero-shot performance gain after fine-tuning and highlighting the potential of this technique in optimizing pre-trained language model efficiency.'}",
-zero-shot domain adaptation for neural machine translation with retrieved phrase-level prompts,Zewei Sun,"Domain adaptation is an important challenge for neural machine translation. However, the traditional fine-tuning solution requires multiple extra training and yields a high cost. In this paper, we propose a non-tuning paradigm, resolving domain adaptation with a prompt-based method. Specifically, we construct a bilingual phrase-level database and retrieve relevant pairs from it as a prompt for the input sentences. By utilizing Retrieved Phrase-level Prompts (RePP), we effectively boost the translation quality. Experiments show that our method improves domain-specific machine translation for 6.2 BLEU scores and improves translation constraints for 11.5% accuracy without additional training.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a non-tuning paradigm, resolving domain adaptation with a prompt-based method, which improves domain-specific machine translation for 6.2 BLEU scores and improves translation constraints for 11.5% accuracy without additional training.'}",http://arxiv.org/pdf/2209.11409
-"electra is a zero-shot learner, too",Shiwen Ni,"Recently, for few-shot or even zero-shot learning, the new paradigm""pre-train, prompt, and predict""has achieved remarkable achievements compared with the""pre-train, fine-tune""paradigm. After the success of prompt-based GPT-3, a series of masked language model (MLM)-based (e.g., BERT, RoBERTa) prompt learning methods became popular and widely used. However, another efficient pre-trained discriminative model, ELECTRA, has probably been neglected. In this paper, we attempt to accomplish several NLP tasks in the zero-shot scenario using a novel our proposed replaced token detection (RTD)-based prompt learning method. Experimental results show that ELECTRA model based on RTD-prompt learning achieves surprisingly state-of-the-art zero-shot performance. Numerically, compared to MLM-RoBERTa-large and MLM-BERT-large, our RTD-ELECTRA-large has an average of about 8.4% and 13.7% improvement on all 15 tasks. Especially on the SST-2 task, our RTD-ELECTRA-large achieves an astonishing 90.1% accuracy without any training data. Overall, compared to the pre-trained masked language models, the pre-trained replaced token detection model performs better in zero-shot learning. The source code is available at: https://github.com/nishiwen1214/RTD-ELECTRA.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that ELECTRA model based on RTD-prompt learning achieves surprisingly state-of-the-art zero-shot performance, and compared to the pre-trained masked language models, thePre-trained replaced token detection model performs better in zero- shot learning.'}",http://arxiv.org/pdf/2207.08141
-empowering sentence encoders with prompting and label retrieval for zero-shot text classification,Jimin Hong,"With contrastive pre-training, sentence encoders are generally optimized to locate semantically similar samples closer to each other in their embedding spaces. In this work, we focus on the potential of their embedding spaces to be readily adapted to zero-shot text classification, as semantically distinct samples are already well-separated. Our framework, RaLP (Retrieval augmented Label Prompts for sentence encoder), encodes prompted label candidates with a sentence encoder, then assigns the label whose prompt embedding has the highest similarity with the input text embedding. In order to compensate for the potentially poorly descriptive labels in their original format, RaLP retrieves sentences that are semantically similar to the original label prompt from external corpora and use them as additional pseudo-label prompts. RaLP achieves competitive or stronger performance than much larger baselines on various closed-set classification and multiple-choice QA datasets under zero-shot settings. We show that the retrieval component plays a pivotal role in RaLP's success, and its results are robustly attained regardless of verbalizer variations.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a framework that retrieves sentences that are semantically similar to the original label prompt from external corpora and uses them as additional pseudo-label prompts, and achieves competitive or stronger performance on various closed-set classification and multiple-choice QA datasets under zero-shot settings.'}",
-evaluating prompts across multiple choice tasks in a zero-shot setting,Gabriel Orlanski,"Large language models have shown that impressive zero-shot performance can be achieved through natural language prompts (Radford et al., 2019; Brown et al., 2020; Sanh et al., 2021). Creating an effective prompt, however, requires significant trial and error. That \textit{prompts} the question: how do the qualities of a prompt effects its performance? To this end, we collect and standardize prompts from a diverse range of tasks for use with tasks they were not designed for. We then evaluate these prompts across fixed multiple choice datasets for a quantitative analysis of how certain attributes of a prompt affect performance. We find that including the choices and using prompts not used during pre-training provide significant improvements. All experiments and code can be found https://github.com/gabeorlanski/zero-shot-cross-task.","{'model': 'tldr@v2.0.0', 'text': 'This work collects and standardizes prompts from a diverse range of tasks for use with tasks they were not designed for, and evaluates these prompts across fixed multiple choice datasets for a quantitative analysis of how certain attributes of a prompt affect performance.'}",http://arxiv.org/pdf/2203.15754
-zerotop: zero-shot task-oriented semantic parsing using large language models,Dheeraj Mekala,"We explore the use of large language models (LLMs) for zero-shot semantic parsing. Semantic parsing involves mapping natural language utterances to task-specific meaning representations. Language models are generally trained on the publicly available text and code and cannot be expected to directly generalize to domain-specific parsing tasks in a zero-shot setting. In this work, we propose ZEROTOP, a zero-shot task-oriented parsing method that decomposes a semantic parsing problem into a set of abstractive and extractive question-answering (QA) problems, enabling us to leverage the ability of LLMs to zero-shot answer reading comprehension questions. For each utterance, we prompt the LLM with questions corresponding to its top-level intent and a set of slots and use the LLM generations to construct the target meaning representation. We observe that current LLMs fail to detect unanswerable questions; and as a result, cannot handle questions corresponding to missing slots. To address this problem, we fine-tune a language model on public QA datasets using synthetic negative samples. Experimental results show that our QA-based decomposition paired with the fine-tuned LLM can correctly parse ~16% of utterances in the MTOP dataset without requiring any annotated data.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ZEROTOP, a zero-shot task-oriented parsing method that decomposes a semantic parsing problem into a set of abstractive and extractive question-answering (QA) problems, enabling us to leverage the ability of LLMs to zero- shot answer reading comprehension questions.'}",http://arxiv.org/pdf/2212.10815
-go-tuning: improving zero-shot learning abilities of smaller language models,Jingjing Xu,"With increasing scale, large language models demonstrate both quantitative improvement and new qualitative capabilities, especially as zero-shot learners, like GPT-3. However, these results rely heavily on delicate prompt design and large computation. In this work, we explore whether the strong zero-shot ability could be achieved at a smaller model scale without any external supervised data. To achieve this goal, we revisit masked language modeling and present a geometry-guided self-supervised learning method (Go-tuningfor short) by taking a small number of task-aware self-supervised data to update language models further. Experiments show that Go-tuning can enable T5-small (80M) competitive zero-shot results compared with large language models, such as T5-XL (3B). We also apply Go-tuning on multi-task settings and develop a multi-task model, mgo-T5 (250M). It can reach the average performance of OPT (175B) on 9 datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work revisits masked language modeling and presents a geometry-guided self-supervised learning method (Go-tuningfor short) by taking a small number of task-aware self- supervised data to update language models further, which can enable T5-small (80M) competitive zero-shot results compared with large language models.'}",http://arxiv.org/pdf/2212.10461
-"how to prompt llms for text-to-sql: a study in zero-shot, single-domain, and cross-domain settings",Shuaichen Chang,"Large language models (LLMs) with in-context learning have demonstrated remarkable capability in the text-to-SQL task. Previous research has prompted LLMs with various demonstration-retrieval strategies and intermediate reasoning steps to enhance the performance of LLMs. However, those works often employ varied strategies when constructing the prompt text for text-to-SQL inputs, such as databases and demonstration examples. This leads to a lack of comparability in both the prompt constructions and their primary contributions. Furthermore, selecting an effective prompt construction has emerged as a persistent problem for future research. To address this limitation, we comprehensively investigate the impact of prompt constructions across various settings and provide insights for future work.","{'model': 'tldr@v2.0.0', 'text': 'This work comprehensively investigates the impact of prompt constructions across various settings and provides insights for future work.'}",http://arxiv.org/pdf/2305.11853
-dialogue state tracking with zero-shot and few-shot learning for generalization: a review,Seungyeon Kim,"Research on Dialogue State Tracking (DST) has achieved meaningful advancements in benchmark datasets. However, the generalization ability of DST models to handle unseen data robustly remains an issue. Hence, recent studies on DST with zero-shot and few-shot learning are reviewed in this paper. For a task-oriented dialogue system, DST is explained by introducing datasets and evaluation metrics. DST models could be categorized into four groups: DST based on a pre-trained model, DST using a description, DST using a prompt, and DST with cross-task. Characteristics of each model are described and the performance of the model experimented under the same conditions is summarized.","{'model': 'tldr@v2.0.0', 'text': 'Recent studies on DST with zero-shot and few-shot learning are reviewed and characteristics of each model are described and the performance of the model experimented under the same conditions is summarized.'}",
-malm: mixing augmented language modeling for zero-shot machine translation,Kshitij Gupta,"Large pre-trained language models have brought remarkable progress in NLP. Pre-training and Fine-tuning have given state-of-art performance across tasks in text processing. Data Augmentation techniques have also helped build state-of-art models on low or zero resource tasks. Many works in the past have attempted at learning a single massively multilingual machine translation model for zero-shot translation. Although those translation models are producing correct translations, the main challenge is those models are producing the wrong languages for zero-shot translation. This work and its results indicate that prompt conditioned large models do not suffer from off-target language errors i.e. errors arising due to translation to wrong languages. We empirically demonstrate the effectiveness of self-supervised pre-training and data augmentation for zero-shot multi-lingual machine translation.","{'model': 'tldr@v2.0.0', 'text': 'This work and its results indicate that prompt conditioned large models do not suffer from off-target language errors i.e. errors arising due to translation to wrong languages.'}",http://arxiv.org/pdf/2210.00320
-zero-shot domain-sensitive speech recognition with prompt-conditioning fine-tuning,F. Liao,"In this work, we propose a method to create domain-sensitive speech recognition models that utilize textual domain information by conditioning its generation on a given text prompt. This is accomplished by fine-tuning a pre-trained, end-to-end model (Whisper) to learn from demonstrations with prompt examples. We show that this ability can be generalized to different domains and even various prompt contexts, with our model gaining a Word Error Rate (WER) reduction of up to 33% on unseen datasets from various domains, such as medical conversation, air traffic control communication, and financial meetings. Considering the limited availability of audio-transcript pair data, we further extend our method to text-only fine-tuning to achieve domain sensitivity as well as domain adaptation. We demonstrate that our text-only fine-tuned model can also attend to various prompt contexts, with the model reaching the most WER reduction of 29% on the medical conversation dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a method to create domain-sensitive speech recognition models that utilize textual domain information by conditioning its generation on a given text prompt by fine-tuning a pre-trained, end-to-end model (Whisper).'}",https://arxiv.org/pdf/2307.10274
-kbpt: knowledge-based prompt tuning for zero-shot relation triplet extraction,Q. Guo,,,
-zero-shot recommendation as language modeling,Damien Sileo,,"{'model': 'tldr@v2.0.0', 'text': 'A framework for recommendation with off-the-shelf pretrained language models (LM) that only used unstructured text corpora as training data is proposed and compared with standard matrix factorization trained on different data regimes.'}",https://arxiv.org/pdf/2112.04184
-zero-shot learning by generating task-specific adapters,Qinyuan Ye,"Pre-trained text-to-text transformers achieve impressive performance across a wide range of NLP tasks, and they naturally support zero-shot learning (ZSL) by using the task description as prompt in the input. However, this approach has potential limitations, as it learns from input-output pairs at instance level, instead of learning to solve tasks at task level. Alternatively, applying existing ZSL methods to text-to-text transformers is non-trivial due to their text generation objective and huge size. To address these issues, we introduce H YPTER , a framework that improves zero-shot transferability by training a hypernetwork to generate task-specific adapter s from task descriptions. This formulation enables learning at task level, and greatly reduces the number of parameters by using light-weight adapters. Experiments on two datasets demonstrate H YPTER improves upon fine-tuning baselines.","{'model': 'tldr@v2.0.0', 'text': 'H YPTER is introduced, a framework that improves zero-shot transferability by training a hypernetwork to generate task-specific adapter s from task descriptions, and greatly reduces the number of parameters by using light-weight adapters.'}",
-domain-aware continual zero-shot learning,Kai Yi,"We introduce Domain Aware Continual Zero-Shot Learning (DACZSL), the task of visually recognizing images of unseen categories in unseen domains sequentially. We created DACZSL on top of the DomainNet dataset by dividing it into a sequence of tasks, where classes are incrementally provided on seen domains during training and evaluation is conducted on unseen domains for both seen and unseen classes. We also proposed a novel Domain-Invariant CZSL Network (DIN), which outperforms state-of-the-art baseline models that we adapted to DACZSL setting. We adopt a structure-based approach to alleviate forgetting knowledge from previous tasks with a small per-task private network in addition to a global shared network. To encourage the private network to capture the domain and task-specific representation, we train our model with a novel adversarial knowledge disentanglement setting to make our global network task-invariant and domain-invariant over all the tasks. Our method also learns a class-wise learnable prompt to obtain better class-level text representation, which is used to represent side information to enable zero-shot prediction of future unseen classes. Our code and benchmarks will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Domain Aware Continual Zero-Shot Learning (DACZSL), the task of visually recognizing images of unseen categories in unseen domains sequentially, and proposes a novel Domain-Invariant CZSL Network (DIN), which outperforms state-of-the-art baseline models that are adapted to DACZSL setting.'}",
-cognitively aided zero-shot automatic essay grading,Sandeep Albert Mathias,"Automatic essay grading (AEG) is a process in which machines assign a grade to an essay written in response to a topic, called the prompt. Zero-shot AEG is when we train a system to grade essays written to a new prompt which was not present in our training data. In this paper, we describe a solution to the problem of zero-shot automatic essay grading, using cognitive information, in the form of gaze behaviour. Our experiments show that using gaze behaviour helps in improving the performance of AEG systems, especially when we provide a new essay written in response to a new prompt for scoring, by an average of almost 5 percentage points of QWK.","{'model': 'tldr@v2.0.0', 'text': 'Their experiments show that using gaze behaviour helps in improving the performance of AEG systems, especially when the authors provide a new essay written in response to a new prompt for scoring, by an average of almost 5 percentage points of QWK.'}",
-sam.md: zero-shot medical image segmentation capabilities of the segment anything model,Saikat Roy,"Foundation models have taken over natural language processing and image generation domains due to the flexibility of prompting. With the recent introduction of the Segment Anything Model (SAM), this prompt-driven paradigm has entered image segmentation with a hitherto unexplored abundance of capabilities. The purpose of this paper is to conduct an initial evaluation of the out-of-the-box zero-shot capabilities of SAM for medical image segmentation, by evaluating its performance on an abdominal CT organ segmentation task, via point or bounding box based prompting. We show that SAM generalizes well to CT data, making it a potential catalyst for the advancement of semi-automatic segmentation tools for clinicians. We believe that this foundation model, while not reaching state-of-the-art segmentation performance in our investigations, can serve as a highly potent starting point for further adaptations of such models to the intricacies of the medical domain. Keywords: medical image segmentation, SAM, foundation models, zero-shot learning","{'model': 'tldr@v2.0.0', 'text': 'It is shown that SAM generalizes well to CT data, making it a potential catalyst for the advancement of semi-automatic segmentation tools for clinicians, and can serve as a highly potent starting point for further adaptations of such models to the intricacies of the medical domain.'}",http://arxiv.org/pdf/2304.05396
-zero-shot clinical entity recognition using chatgpt,Yan Hu,"In this study, we investigated the potential of ChatGPT, a large language model developed by OpenAI, for the clinical named entity recognition task defined in the 2010 i2b2 challenge, in a zero-shot setting with two different prompt strategies. We compared its performance with GPT-3 in a similar zero-shot setting, as well as a fine-tuned BioClinicalBERT model using a set of synthetic clinical notes from MTSamples. Our findings revealed that ChatGPT outperformed GPT-3 in the zero-shot setting, with F1 scores of 0.418 (vs.0.250) and 0.620 (vs. 0.480) for exact- and relaxed-matching, respectively. Moreover, prompts affected ChatGPT's performance greatly, with relaxed-matching F1 scores of 0.628 vs.0.541 for two different prompt strategies. Although ChatGPT's performance was still lower than that of the supervised BioClinicalBERT model (i.e., relaxed-matching F1 scores of 0.620 vs. 0.888), our study demonstrates the great potential of ChatGPT for clinical NER tasks in a zero-shot setting, which is much more appealing as it does not require any annotation.",,http://arxiv.org/pdf/2303.16416
-a preliminary evaluation of chatgpt for zero-shot dialogue understanding,Wenbo Pan,"Zero-shot dialogue understanding aims to enable dialogue to track the user's needs without any training data, which has gained increasing attention. In this work, we investigate the understanding ability of ChatGPT for zero-shot dialogue understanding tasks including spoken language understanding (SLU) and dialogue state tracking (DST). Experimental results on four popular benchmarks reveal the great potential of ChatGPT for zero-shot dialogue understanding. In addition, extensive analysis shows that ChatGPT benefits from the multi-turn interactive prompt in the DST task but struggles to perform slot filling for SLU. Finally, we summarize several unexpected behaviors of ChatGPT in dialogue understanding tasks, hoping to provide some insights for future research on building zero-shot dialogue understanding systems with Large Language Models (LLMs).","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the understanding ability of ChatGPT for zero-shot dialogue understanding tasks including spoken language understanding (SLU) and dialogue state tracking (DST).'}",http://arxiv.org/pdf/2304.04256
-esc: exploration with soft commonsense constraints for zero-shot object navigation,KAI-QING Zhou,"The ability to accurately locate and navigate to a specific object is a crucial capability for embodied agents that operate in the real world and interact with objects to complete tasks. Such object navigation tasks usually require large-scale training in visual environments with labeled objects, which generalizes poorly to novel objects in unknown environments. In this work, we present a novel zero-shot object navigation method, Exploration with Soft Commonsense constraints (ESC), that transfers commonsense knowledge in pre-trained models to open-world object navigation without any navigation experience nor any other training on the visual environments. First, ESC leverages a pre-trained vision and language model for open-world prompt-based grounding and a pre-trained commonsense language model for room and object reasoning. Then ESC converts commonsense knowledge into navigation actions by modeling it as soft logic predicates for efficient exploration. Extensive experiments on MP3D, HM3D, and RoboTHOR benchmarks show that our ESC method improves significantly over baselines, and achieves new state-of-the-art results for zero-shot object navigation (e.g., 288% relative Success Rate improvement than CoW on MP3D).","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot object navigation method, Exploration with Soft Commonsense constraints (ESC), that transfers commonsense knowledge in pre-trained models to open-world object navigation without any navigation experience nor any other training on the visual environments is presented.'}",https://arxiv.org/pdf/2301.13166
-zero-shot performance of the segment anything model (sam) in 2d medical imaging: a comprehensive evaluation and practical guidelines,C. M. Oliveira,"Segmentation in medical imaging is a critical component for the diagnosis, monitoring, and treatment of various diseases and medical conditions. Presently, the medical segmentation landscape is dominated by numerous specialized deep learning models, each fine-tuned for specific segmentation tasks and image modalities. The recently-introduced Segment Anything Model (SAM) employs the ViT neural architecture and harnesses a massive training dataset to segment nearly any object; however, its suitability to the medical domain has not yet been investigated. In this study, we explore the zero-shot performance of SAM in medical imaging by implementing eight distinct prompt strategies across six datasets from four imaging modalities, including X-ray, ultrasound, dermatoscopy, and colonoscopy. Our findings reveal that SAM's zero-shot performance is not only comparable to, but in certain cases, surpasses the current state-of-the-art. Based on these results, we propose practical guidelines that require minimal interaction while consistently yielding robust outcomes across all assessed contexts. The source code, along with a demonstration of the recommended guidelines, can be accessed at https://github.com/Malta-Lab/SAM-zero-shot-in-Medical-Imaging.","{'model': 'tldr@v2.0.0', 'text': ""The findings reveal that SAM's zero-shot performance is not only comparable to, but in certain cases, surpasses the current state-of-the-art, and practical guidelines are proposed that require minimal interaction while consistently yielding robust outcomes across all assessed contexts.""}",http://arxiv.org/pdf/2305.00109
-"clip for all things zero-shot sketch-based image retrieval, fine-grained or not",Aneeshan Sain,"In this paper, we leverage CLIP for zero-shot sketch based image retrieval (ZS-SBIR). We are largely inspired by recent advances on foundation models and the unparalleled generalisation ability they seem to offer, but for the first time tailor it to benefit the sketch community. We put forward novel designs on how best to achieve this synergy, for both the category setting and the fine-grained setting ('all”}. At the very core of our solution is a prompt learning setup. First we show just via factoring in sketch-specific prompts, we already have a category-level ZS-SBIR system that over-shoots all prior arts, by a large margin (24.8%) - a great testimony on studying the CLIP and ZS-SBIR synergy. Moving onto the fine-grained setup is however trickier, and re-quires a deeper dive into this synergy. For that, we come up with two specific designs to tackle the fine-grained matching nature of the problem: (i) an additional regularisation loss to ensure the relative separation between sketches and photos is uniform across categories, which is not the case for the gold standard standalone triplet loss, and (ii) a clever patch shuffling technique to help establishing instance-level structural correspondences between sketch-photo pairs. With these designs, we again observe signifi-cant performance gains in the region of 26.9% over previ-ous state-of-the-art. The take-home message, if any, is the proposed CLIP and prompt learning paradigm carries great promise in tackling other sketch-related tasks (not limited to ZS-SBIR) where data scarcity remains a great challenge. Project page: https://aneeshan95.github.ioISketchLVM/","{'model': 'tldr@v2.0.0', 'text': 'The proposed CLIP and prompt learning paradigm carries great promise in tackling other sketch-related tasks (not limited to ZS-SBIR) where data scarcity remains a great challenge.'}",https://arxiv.org/pdf/2303.13440
-xplainer: from x-ray observations to explainable zero-shot diagnosis,Chantal Pellegrini,"Automated diagnosis prediction from medical images is a valuable resource to support clinical decision-making. However, such systems usually need to be trained on large amounts of annotated data, which often is scarce in the medical domain. Zero-shot methods address this challenge by allowing a flexible adaption to new settings with different clinical findings without relying on labeled data. Further, to integrate automated diagnosis in the clinical workflow, methods should be transparent and explainable, increasing medical professionals' trust and facilitating correctness verification. In this work, we introduce Xplainer, a novel framework for explainable zero-shot diagnosis in the clinical setting. Xplainer adapts the classification-by-description approach of contrastive vision-language models to the multi-label medical diagnosis task. Specifically, instead of directly predicting a diagnosis, we prompt the model to classify the existence of descriptive observations, which a radiologist would look for on an X-Ray scan, and use the descriptor probabilities to estimate the likelihood of a diagnosis. Our model is explainable by design, as the final diagnosis prediction is directly based on the prediction of the underlying descriptors. We evaluate Xplainer on two chest X-ray datasets, CheXpert and ChestX-ray14, and demonstrate its effectiveness in improving the performance and explainability of zero-shot diagnosis. Our results suggest that Xplainer provides a more detailed understanding of the decision-making process and can be a valuable tool for clinical diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'Xplainer is introduced, a novel framework for explainable zero-shot diagnosis in the clinical setting that adapts the classification-by-description approach of contrastive vision-language models to the multi-label medical diagnosis task and can be a valuable tool for clinical diagnosis.'}",http://arxiv.org/pdf/2303.13391
-rapgen: an approach for fixing code inefficiencies in zero-shot,Spandan Garg,"Performance bugs are non-functional bugs that can even manifest in well-tested commercial products. Fixing these performance bugs is an important yet challenging problem. In this work, we address this challenge and present a new approach called Retrieval-Augmented Prompt Generation (RAPGen). Given a code snippet with a performance issue, RAPGen first retrieves a prompt instruction from a pre-constructed knowledge-base of previous performance bug fixes and then generates a prompt using the retrieved instruction. It then uses this prompt on a Large Language Model (such as Codex) in zero-shot to generate a fix. We compare our approach with the various prompt variations and state of the art methods in the task of performance bug fixing. Our evaluation shows that RAPGen can generate performance improvement suggestions equivalent or better than a developer in ~60% of the cases, getting ~39% of them verbatim, in an expert-verified dataset of past performance changes made by C# developers.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a new approach called Retrieval-Augmented Prompt Generation (RAPGen), which first retrieves a prompt instruction from a pre-constructed knowledge-base of previous performance bug fixes and then generates a prompt using the retrieved instruction.'}",http://arxiv.org/pdf/2306.17077
-zero-shot cross-lingual summarization via large language models,Jiaan Wang,"Given a document in a source language, cross-lingual summarization (CLS) aims to generate a summary in a different target language. Recently, the emergence of Large Language Models (LLMs), such as GPT-3.5, ChatGPT and GPT-4, has attracted wide attention from the computational linguistics community. However, it is not yet known the performance of LLMs on CLS. In this report, we empirically use various prompts to guide LLMs to perform zero-shot CLS from different paradigms (i.e., end-to-end and pipeline), and provide a preliminary evaluation on the generated summaries. We find that ChatGPT and GPT-4 originally prefer to produce lengthy summaries with detailed information. These two LLMs can further balance informativeness and conciseness with the help of an interactive prompt, significantly improving their CLS performance. Experimental results on three widely-used CLS datasets show that GPT-4 achieves state-of-the-art zero-shot CLS performance, and performs competitively compared with the fine-tuned mBART-50. Moreover, we also find some multi-lingual and bilingual LLMs (i.e., BLOOMZ, ChatGLM-6B, Vicuna-13B and ChatYuan) have limited zero-shot CLS ability. Due to the composite nature of CLS, which requires models to perform summarization and translation simultaneously, accomplishing this task in a zero-shot manner is even a challenge for LLMs. Therefore, we sincerely hope and recommend future LLM research could use CLS as a testbed.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT and GPT-4 originally prefer to produce lengthy summaries with detailed information, and these two LLMs can further balance informativeness and conciseness with the help of an interactive prompt, significantly improving their CLS performance.'}",
-enabling calibration in the zero-shot inference of large vision-language models,Will LeVine,"Calibration of deep learning models is crucial to their trustworthiness and safe usage, and as such, has been extensively studied in supervised classification models, with methods crafted to decrease miscalibration. However, there has yet to be a comprehensive study of the calibration of vision-language models that are used for zero-shot inference, like CLIP. We measure calibration across relevant variables like prompt, dataset, and architecture, and find that zero-shot inference with CLIP is miscalibrated. Furthermore, we propose a modified version of temperature scaling that is aligned with the common use cases of CLIP as a zero-shot inference model, and show that a single learned temperature generalizes for each specific CLIP model (defined by a chosen pre-training dataset and architecture) across inference dataset and prompt choice.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a modified version of temperature scaling that is aligned with the common use cases of CLIP as a zero-shot inference model, and shows that a single learned temperature generalizes for each specific CLIP model across inference dataset and prompt choice.'}",http://arxiv.org/pdf/2303.12748
-clipn for zero-shot ood detection: teaching clip to say no,Hualiang Wang,"Out-of-distribution (OOD) detection refers to training the model on an in-distribution (ID) dataset to classify whether the input images come from unknown classes. Considerable effort has been invested in designing various OOD detection methods based on either convolutional neural networks or transformers. However, zero-shot OOD detection methods driven by CLIP, which only require class names for ID, have received less attention. This paper presents a novel method, namely CLIP saying no (CLIPN), which empowers the logic of saying no within CLIP. Our key motivation is to equip CLIP with the capability of distinguishing OOD and ID samples using positive-semantic prompts and negation-semantic prompts. Specifically, we design a novel learnable no prompt and a no text encoder to capture negation semantics within images. Subsequently, we introduce two loss functions: the image-text binary-opposite loss and the text semantic-opposite loss, which we use to teach CLIPN to associate images with no prompts, thereby enabling it to identify unknown samples. Furthermore, we propose two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from no prompts and the text encoder. Experimental results on 9 benchmark datasets (3 ID datasets and 6 OOD datasets) for the OOD detection task demonstrate that CLIPN, based on ViT-B-16, outperforms 7 well-used algorithms by at least 2.34% and 11.64% in terms of AUROC and FPR95 for zero-shot OOD detection on ImageNet-1K. Our CLIPN can serve as a solid foundation for effectively leveraging CLIP in downstream OOD tasks. The code is available on https://github.com/xmed-lab/CLIPN.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel method, namely CLIP saying no (CLIPN), which empowers the logic of saying no within CLIP and proposes two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from no prompts and the text encoder.'}",https://arxiv.org/pdf/2308.12213
-infusion: inject and attention fusion for multi concept zero-shot text-based video editing,Anant Khandelwal,"Large text-to-image diffusion models have achieved remarkable success in generating diverse, high-quality images. Additionally, these models have been successfully leveraged to edit input images by just changing the text prompt. But when these models are applied to videos, the main challenge is to ensure temporal consistency and coherence across frames. In this paper, we propose InFusion, a framework for zero-shot text-based video editing leveraging large pre-trained image diffusion models. Our framework specifically supports editing of multiple concepts with pixel-level control over diverse concepts mentioned in the editing prompt. Specifically, we inject the difference in features obtained with source and edit prompts from U-Net residual blocks of decoder layers. When these are combined with injected attention features, it becomes feasible to query the source contents and scale edited concepts along with the injection of unedited parts. The editing is further controlled in a fine-grained manner with mask extraction and attention fusion, which cut the edited part from the source and paste it into the denoising pipeline for the editing prompt. Our framework is a low-cost alternative to one-shot tuned models for editing since it does not require training. We demonstrated complex concept editing with a generalised image model (Stable Diffusion v1.5) using LoRA. Adaptation is compatible with all the existing image diffusion techniques. Extensive experimental results demonstrate the effectiveness of existing methods in rendering high-quality and temporally consistent videos.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes InFusion, a framework for zero-shot text-based video editing leveraging large pre-trained image diffusion models, and demonstrates complex concept editing with a generalised image model (Stable Diffusion v1.5) using LoRA.'}",https://arxiv.org/pdf/2308.00135
-troika: multi-path cross-modal traction for compositional zero-shot learning,Siteng Huang,"Recent compositional zero-shot learning (CZSL) methods adapt pre-trained vision-language models (VLMs) by constructing trainable prompts only for composed state-object pairs. Relying on learning the joint representation of seen compositions, these methods ignore the explicit modeling of the state and object, thus limiting the exploitation of pre-trained knowledge and generalization to unseen compositions. With a particular focus on the universality of the solution, in this work, we propose a novel paradigm for CZSL models that establishes three identification branches (i.e., Multi-Path) to jointly model the state, object, and composition. The presented Troika is our implementation that aligns the branch-specific prompt representations with decomposed visual features. To calibrate the bias between semantically similar multi-modal representations, we further devise a Cross-Modal Traction module into Troika that shifts the prompt representation towards the current visual content. We conduct extensive experiments on three popular benchmarks, where our method significantly outperforms existing methods in both closed-world and open-world settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel paradigm for CZSL models that establishes three identification branches (i.e., Multi-Path) to jointly model the state, object, and composition and conducts extensive experiments on three popular benchmarks, where this method significantly outperforms existing methods in both closed-world and open-world settings.'}",http://arxiv.org/pdf/2303.15230
-zero-shot information extraction for clinical meta-analysis using large language models,David Kartchner,"Meta-analysis of randomized clinical trials (RCTs) plays a crucial role in evidence-based medicine but can be labor-intensive and error-prone. This study explores the use of large language models to enhance the efficiency of aggregating results from randomized clinical trials (RCTs) at scale. We perform a detailed comparison of the performance of these models in zero-shot prompt-based information extraction from a diverse set of RCTs to traditional manual annotation methods. We analyze the results for two different meta-analyses aimed at drug repurposing in cancer therapy pharmacovigilience in chronic myeloid leukemia. Our findings reveal that the best model for the two demonstrated tasks, ChatGPT can generally extract correct information and identify when the desired information is missing from an article. We additionally conduct a systematic error analysis, documenting the prevalence of diverse error types encountered during the process of prompt-based information extraction.","{'model': 'tldr@v2.0.0', 'text': 'The use of large language models to enhance the efficiency of aggregating results from randomized clinical trials (RCTs) at scale is explored and the best model, ChatGPT, is revealed to be the best for the two demonstrated tasks.'}",https://aclanthology.org/2023.bionlp-1.37.pdf
-vision-language models are zero-shot reward models for reinforcement learning,Juan Rocamonde,"Reinforcement learning (RL) requires either manually specifying a reward function, which is often infeasible, or learning a reward model from a large amount of human feedback, which is often very expensive. We study a more sample-efficient alternative: using pretrained vision-language models (VLMs) as zero-shot reward models (RMs) to specify tasks via natural language. We propose a natural and general approach to using VLMs as reward models, which we call VLM-RMs. We use VLM-RMs based on CLIP to train a MuJoCo humanoid to learn complex tasks without a manually specified reward function, such as kneeling, doing the splits, and sitting in a lotus position. For each of these tasks, we only provide a single sentence text prompt describing the desired task with minimal prompt engineering. We provide videos of the trained agents at: https://sites.google.com/view/vlm-rm. We can improve performance by providing a second ``baseline'' prompt and projecting out parts of the CLIP embedding space irrelevant to distinguish between goal and baseline. Further, we find a strong scaling effect for VLM-RMs: larger VLMs trained with more compute and data are better reward models. The failure modes of VLM-RMs we encountered are all related to known capability limitations of current VLMs, such as limited spatial reasoning ability or visually unrealistic environments that are far off-distribution for the VLM. We find that VLM-RMs are remarkably robust as long as the VLM is large enough. This suggests that future VLMs will become more and more useful reward models for a wide range of RL applications.","{'model': 'tldr@v2.0.0', 'text': 'It is found that VLM-RMs are remarkably robust as long as the VLM is large enough, which suggests that future VLMs will become more and more useful reward models for a wide range of RL applications.'}",
-tonguesam: an universal tongue segmentation model based on sam with zero-shot,Shan Cao,"Tongue segmentation serves as the primary step in automated TCM tongue diagnosis, which plays a significant role in the diagnostic results. Currently, numerous deep learning based methods have achieved promising results. However, when confronted with tongue images that differ from the training set or possess challenging backgrounds, these methods demonstrate limited performance. To address this issue, this paper proposes a universal tongue segmentation model named TongueSAM based on SAM (Segment Anything Model). SAM is a large-scale pretrained interactive segmentation model known for its powerful zero-shot generalization capability. Applying SAM to tongue segmentation leverages its learned prior knowledge from natural images, enabling the achievement of zero-shot segmentation for various types of tongue images. In this study, a Prompt Generator based on object detection is integrated into SAM to enable an end-to-end automated tongue segmentation method. Experiments demonstrate that TongueSAM achieves exceptional performance across various of tongue segmentation datasets, particularly under zero-shot. Even when dealing with challenging background tongue images, TongueSAM achieves a mIoU of 95.23\% under zero-shot conditions, surpassing other segmentation methods. As far as we know, this is the first application of large-scale pretrained model for tongue segmentation. The project and pretrained model will be made public when the paper is accepted.","{'model': 'tldr@v2.0.0', 'text': 'Tongue segmentation serves as the primary step in automated TCM tongue diagnosis, which plays a significant role in the diagnostic results, and a large-scale pretrained interactive segmentation model based on SAM (Segment Anything Model) is proposed, which achieves exceptional performance across various of tongue segmentation datasets, particularly under zero-shot.'}",https://arxiv.org/pdf/2308.06444
-zero-shot text classification via self-supervised tuning,Chaoqun Liu,"Existing solutions to zero-shot text classification either conduct prompting with pre-trained language models, which is sensitive to the choices of templates, or rely on large-scale annotated data of relevant tasks for meta-tuning. In this work, we propose a new paradigm based on self-supervised learning to solve zero-shot text classification tasks by tuning the language models with unlabeled data, called self-supervised tuning. By exploring the inherent structure of free texts, we propose a new learning objective called first sentence prediction to bridge the gap between unlabeled data and text classification tasks. After tuning the model to learn to predict the first sentence in a paragraph based on the rest, the model is able to conduct zero-shot inference on unseen tasks such as topic classification and sentiment analysis. Experimental results show that our model outperforms the state-of-the-art baselines on 7 out of 10 tasks. Moreover, the analysis reveals that our model is less sensitive to the prompt design. Our code and pre-trained models are publicly available at https://github.com/DAMO-NLP-SG/SSTuning .","{'model': 'tldr@v2.0.0', 'text': 'After tuning the model to learn to predict the first sentence in a paragraph based on the rest, the model is able to conduct zero-shot inference on unseen tasks such as topic classification and sentiment analysis and outperforms the state-of-the-art baselines on 7 out of 10 tasks.'}",http://arxiv.org/pdf/2305.11442
-text-conditional contextualized avatars for zero-shot personalization,S. Azadi,"Recent large-scale text-to-image generation models have made significant improvements in the quality, realism, and diversity of the synthesized images and enable users to control the created content through language. However, the personalization aspect of these generative models is still challenging and under-explored. In this work, we propose a pipeline that enables personalization of image generation with avatars capturing a user's identity in a delightful way. Our pipeline is zero-shot, avatar texture and style agnostic, and does not require training on the avatar at all - it is scalable to millions of users who can generate a scene with their avatar. To render the avatar in a pose faithful to the given text prompt, we propose a novel text-to-3D pose diffusion model trained on a curated large-scale dataset of in-the-wild human poses improving the performance of the SOTA text-to-motion models significantly. We show, for the first time, how to leverage large-scale image datasets to learn human 3D pose parameters and overcome the limitations of motion capture datasets.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes a pipeline that enables personalization of image generation with avatars capturing a user's identity in a delightful way and shows, for the first time, how to leverage large-scale image datasets to learn human 3D pose parameters and overcome the limitations of motion capture datasets.""}",http://arxiv.org/pdf/2304.07410
-zero-shot co-salient object detection framework,Haoke Xiao,"Co-salient Object Detection (CoSOD) endeavors to replicate the human visual system's capacity to recognize common and salient objects within a collection of images. Despite recent advancements in deep learning models, these models still rely on training with well-annotated CoSOD datasets. The exploration of training-free zero-shot CoSOD frameworks has been limited. In this paper, taking inspiration from the zero-shot transfer capabilities of foundational computer vision models, we introduce the first zero-shot CoSOD framework that harnesses these models without any training process. To achieve this, we introduce two novel components in our proposed framework: the group prompt generation (GPG) module and the co-saliency map generation (CMP) module. We evaluate the framework's performance on widely-used datasets and observe impressive results. Our approach surpasses existing unsupervised methods and even outperforms fully supervised methods developed before 2020, while remaining competitive with some fully supervised methods developed before 2022.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the first zero-shot CoSOD framework that harnesses foundational computer vision models without any training process, and surpasses existing unsupervised methods and even outperforms fully supervised methods developed before 2020, while remaining competitive with some fully supervised method developed before 2022.'}",https://arxiv.org/pdf/2309.05499
-harnessing the zero-shot power of instruction-tuned large language model in end-to-end speech recognition,Yosuke Higuchi,"We present a novel integration of an instruction-tuned large language model (LLM) and end-to-end automatic speech recognition (ASR). Modern LLMs can perform a wide range of linguistic tasks within zero-shot learning when provided with a precise instruction or a prompt to guide the text generation process towards the desired task. We explore using this zero-shot capability of LLMs to extract linguistic information that can contribute to improving ASR performance. Specifically, we direct an LLM to correct grammatical errors in an ASR hypothesis and harness the embedded linguistic knowledge to conduct end-to-end ASR. The proposed model is built on the hybrid connectionist temporal classification (CTC) and attention architecture, where an instruction-tuned LLM (i.e., Llama2) is employed as a front-end of the decoder. An ASR hypothesis, subject to correction, is obtained from the encoder via CTC decoding, which is then fed into the LLM along with an instruction. The decoder subsequently takes as input the LLM embeddings to perform sequence generation, incorporating acoustic information from the encoder output. Experimental results and analyses demonstrate that the proposed integration yields promising performance improvements, and our approach largely benefits from LLM-based rescoring.","{'model': 'tldr@v2.0.0', 'text': 'A novel integration of an instruction-tuned large language model (LLM) and end-to-end automatic speech recognition (ASR) is presented, built on the hybrid connectionist temporal classification (CTC) and attention architecture.'}",https://arxiv.org/pdf/2309.10524
-align your prompts: test-time prompting with distribution alignment for zero-shot generalization,Jameel Hassan,"The promising zero-shot generalization of vision-language models such as CLIP has led to their adoption using prompt learning for numerous downstream tasks. Previous works have shown test-time prompt tuning using entropy minimization to adapt text prompts for unseen domains. While effective, this overlooks the key cause for performance degradation to unseen domains -- distribution shift. In this work, we explicitly handle this problem by aligning the out-of-distribution (OOD) test sample statistics to those of the source data using prompt tuning. We use a single test sample to adapt multi-modal prompts at test time by minimizing the feature distribution shift to bridge the gap in the test domain. Evaluating against the domain generalization benchmark, our method improves zero-shot top- 1 accuracy beyond existing prompt-learning techniques, with a 3.08% improvement over the baseline MaPLe. In cross-dataset generalization with unseen categories across 10 datasets, our method improves consistently across all datasets compared to the existing state-of-the-art. Our source code and models are available at https://jameelhassan.github.io/promptalign.","{'model': 'tldr@v2.0.0', 'text': 'This work uses a single test sample to adapt multi-modal prompts at test time by minimizing the feature distribution shift to bridge the gap in the test domain, and improves zero-shot top- 1 accuracy beyond existing prompt-learning techniques.'}",
-towards realistic zero-shot classification via self structural semantic alignment,Shengxiang Zhang,"Large-scale pre-trained Vision Language Models (VLMs) have proven effective for zero-shot classification. Despite the success, most traditional VLMs-based methods are restricted by the assumption of partial source supervision or ideal vocabularies, which rarely satisfy the open-world scenario. In this paper, we aim at a more challenging setting, Realistic Zero-Shot Classification, which assumes no annotation but instead a broad vocabulary. To address this challenge, we propose the Self Structural Semantic Alignment (S^3A) framework, which extracts the structural semantic information from unlabeled data while simultaneously self-learning. Our S^3A framework adopts a unique Cluster-Vote-Prompt-Realign (CVPR) algorithm, which iteratively groups unlabeled data to derive structural semantics for pseudo-supervision. Our CVPR process includes iterative clustering on images, voting within each cluster to identify initial class candidates from the vocabulary, generating discriminative prompts with large language models to discern confusing candidates, and realigning images and the vocabulary as structural semantic alignment. Finally, we propose to self-learn the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy. Our comprehensive experiments across various generic and fine-grained benchmarks demonstrate that the S^3A method offers substantial improvements over existing VLMs-based approaches, achieving a more than 15% accuracy improvement over CLIP on average. Our codes, models, and prompts are publicly released at https://github.com/sheng-eatamath/S3A.","{'model': 'tldr@v2.0.0', 'text': 'The Self Structural Semantic Alignment (S^3A) framework is proposed, which extracts the structural semantic information from unlabeled data while simultaneously self-learning and proposes to self-learn the CLIP image encoder with both individual and structural semantic alignment through a teacher-student learning strategy.'}",https://arxiv.org/pdf/2308.12960
-instruction distillation makes large language models efficient zero-shot rankers,Weiwei Sun,"Recent studies have demonstrated the great potential of Large Language Models (LLMs) serving as zero-shot relevance rankers. The typical approach involves making comparisons between pairs or lists of documents. Although effective, these listwise and pairwise methods are not efficient and also heavily rely on intricate prompt engineering. To tackle this problem, we introduce a novel instruction distillation method. The key idea is to distill the pairwise ranking ability of open-sourced LLMs to a simpler but more efficient pointwise ranking. Specifically, given the same LLM, we first rank documents using the effective pairwise approach with complex instructions, and then distill the teacher predictions to the pointwise approach with simpler instructions. Evaluation results on the BEIR, TREC, and ReDial datasets demonstrate that instruction distillation can improve efficiency by 10 to 100x and also enhance the ranking performance of LLMs. Furthermore, our approach surpasses the performance of existing supervised methods like monoT5 and is on par with the state-of-the-art zero-shot methods. The code to reproduce our results is available at www.github.com/sunnweiwei/RankGPT.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel instruction distillation method that surpasses the performance of existing supervised methods like monoT5 and is on par with the state-of-the-art zero-shot methods.'}",
-masked momentum contrastive learning for zero-shot semantic understanding,Jiantao Wu,"Self-supervised pretraining (SSP) has emerged as a popular technique in machine learning, enabling the extraction of meaningful feature representations without labelled data. In the realm of computer vision, pretrained vision transformers (ViTs) have played a pivotal role in advancing transfer learning. Nonetheless, the escalating cost of finetuning these large models has posed a challenge due to the explosion of model size. This study endeavours to evaluate the effectiveness of pure self-supervised learning (SSL) techniques in computer vision tasks, obviating the need for finetuning, with the intention of emulating human-like capabilities in generalisation and recognition of unseen objects. To this end, we propose an evaluation protocol for zero-shot segmentation based on a prompting patch. Given a point on the target object as a prompt, the algorithm calculates the similarity map between the selected patch and other patches, upon that, a simple thresholding is applied to segment the target. Another evaluation is intra-object and inter-object similarity to gauge discriminatory ability of SSP ViTs. Insights from zero-shot segmentation from prompting and discriminatory abilities of SSP led to the design of a simple SSP approach, termed MMC. This approaches combines Masked image modelling for encouraging similarity of local features, Momentum based self-distillation for transferring semantics from global to local features, and global Contrast for promoting semantics of global features, to enhance discriminative representations of SSP ViTs. Consequently, our proposed method significantly reduces the overlap of intra-object and inter-object similarities, thereby facilitating effective object segmentation within an image. Our experiments reveal that MMC delivers top-tier results in zero-shot semantic segmentation across various datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study endeavours to evaluate the effectiveness of pure self-supervised learning (SSL) techniques in computer vision tasks, obviating the need for finetuning, with the intention of emulating human-like capabilities in generalisation and recognition of unseen objects.'}",https://arxiv.org/pdf/2308.11448
-locally differentially private document generation using zero shot prompting,Saiteja Utpala,"Numerous studies have highlighted the privacy risks associated with pretrained large language models. In contrast, our research offers a unique perspective by demonstrating that pretrained large language models can effectively contribute to privacy preservation. We propose a locally differentially private mechanism called DP-Prompt, which leverages the power of pretrained large language models and zero-shot prompting to counter author de-anonymization attacks while minimizing the impact on downstream utility. When DP-Prompt is used with a powerful language model like ChatGPT (gpt-3.5), we observe a notable reduction in the success rate of de-anonymization attacks, showing that it surpasses existing approaches by a considerable margin despite its simpler design. For instance, in the case of the IMDB dataset, DP-Prompt (with ChatGPT) perfectly recovers the clean sentiment F1 score while achieving a 46\% reduction in author identification F1 score against static attackers and a 26\% reduction against adaptive attackers. We conduct extensive experiments across six open-source large language models, ranging up to 7 billion parameters, to analyze various effects of the privacy-utility tradeoff.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a locally differentially private mechanism called DP-Prompt, which leverages the power of pretrained large language models and zero-shot prompting to counter author de-anonymization attacks while minimizing the impact on downstream utility.'}",
-zero-shot relation triple extraction with prompts for low-resource languages,Ayiguli Halike,"Although low-resource relation extraction is vital in knowledge construction and characterization, more research is needed on the generalization of unknown relation types. To fill the gap in the study of low-resource (Uyghur) relation extraction methods, we created a zero-shot with a quick relation extraction task setup. Each triplet extracted from an input phrase consists of the subject, relation type, and object. This paper suggests generating structured texts by urging language models to provide related instances. Our model consists of two modules: relation generator and relation and triplet extractor. We use the Uyghur relation prompt in the relation generator stage to generate new synthetic data. In the relation and triple extraction stage, we use the new data to extract the relation triplets in the sentence. We use multi-language model prompts and structured text techniques to offer a structured relation prompt template. This method is the first research that extends relation triplet extraction to a zero-shot setting for Uyghur datasets. Experimental results show that our method achieves a maximum weighted average F1 score of 47.39%.","{'model': 'tldr@v2.0.0', 'text': 'This method is the first research that extends relation triplet extraction to a zero-shot setting for Uyghur datasets and uses multi-language model prompts and structured text techniques to offer a structured relation prompt template.'}",https://www.mdpi.com/2076-3417/13/7/4636/pdf?version=1681110517
-instance needs more care: rewriting prompts for instances yields better zero-shot performance,Saurabh Srivastava,"Enabling large language models (LLMs) to perform tasks in zero-shot has been an appealing goal owing to its labor-saving (i.e., requiring no task-specific annotations); as such, zero-shot prompting approaches also enjoy better task generalizability. To improve LLMs' zero-shot performance, prior work has focused on devising more effective task instructions (e.g., ``let's think step by step'' ). However, we argue that, in order for an LLM to solve them correctly in zero-shot, individual test instances need more carefully designed and customized instructions. To this end, we propose PRoMPTd, an approach that rewrites the task prompt for each individual test input to be more specific, unambiguous, and complete, so as to provide better guidance to the task LLM. We evaluated PRoMPTd on eight datasets covering tasks including arithmetics, logical reasoning, and code generation, using GPT-4 as the task LLM. Notably, PRoMPTd achieves an absolute improvement of around 10% on the complex MATH dataset and 5% on the code generation task on HumanEval, outperforming conventional zero-shot methods. In addition, we also showed that the rewritten prompt can provide better interpretability of how the LLM resolves each test instance, which can potentially be leveraged as a defense mechanism against adversarial prompting. The source code and dataset can be obtained from https://github.com/salokr/PRoMPTd","{'model': 'tldr@v2.0.0', 'text': 'PRoMPTd, an approach that rewrites the task prompt for each individual test input to be more specific, unambiguous, and complete, so as to provide better guidance to the task LLM, is proposed.'}",https://arxiv.org/pdf/2310.02107
-zyn: zero-shot reward models with yes-no questions,Víctor Gallego,"In this work, we address the problem of directing the text generations of a LLM towards a desired behavior, aligning the generated text with the preferences of the human operator. We propose using another language model as a critic, reward model in a zero-shot way thanks to the prompt of a Yes-No question that represents the user preferences, without requiring further labeled data. This zero-shot reward model provides the learning signal to further fine-tune the base LLM using reinforcement learning, as in RLAIF; yet our approach is also compatible in other contexts such as quality-diversity search. Extensive evidence of the capabilities of the proposed ZYN framework is provided through experiments in different domains related to text generation, including detoxification; optimizing sentiment of movie reviews, or any other attribute; steering the opinion about a particular topic the model may have; and personalizing prompt generators for text-to-image tasks. Code to be released at \url{https://github.com/vicgalle/zero-shot-reward-models/}.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes using another language model as a critic, reward model in a zero-shot way thanks to the prompt of a Yes-No question that represents the user preferences, without requiring further labeled data.'}",https://arxiv.org/pdf/2308.06385
-random word data augmentation with clip for zero-shot anomaly detection,Masato Tamura,"This paper presents a novel method that leverages a visual-language model, CLIP, as a data source for zero-shot anomaly detection. Tremendous efforts have been put towards developing anomaly detectors due to their potential industrial applications. Considering the difficulty in acquiring various anomalous samples for training, most existing methods train models with only normal samples and measure discrepancies from the distribution of normal samples during inference, which requires training a model for each object category. The problem of this inefficient training requirement has been tackled by designing a CLIP-based anomaly detector that applies prompt-guided classification to each part of an image in a sliding window manner. However, the method still suffers from the labor of careful prompt ensembling with known object categories. To overcome the issues above, we propose leveraging CLIP as a data source for training. Our method generates text embeddings with the text encoder in CLIP with typical prompts that include words of normal and anomaly. In addition to these words, we insert several randomly generated words into prompts, which enables the encoder to generate a diverse set of normal and anomalous samples. Using the generated embeddings as training data, a feed-forward neural network learns to extract features of normal and anomaly from CLIP's embeddings, and as a result, a category-agnostic anomaly detector can be obtained without any training images. Experimental results demonstrate that our method achieves state-of-the-art performance without laborious prompt ensembling in zero-shot setups.","{'model': 'tldr@v2.0.0', 'text': ""A feed-forward neural network learns to extract features of normal and anomaly from CLIP's embeddings, and as a result, a category-agnostic anomaly detector can be obtained without any training images.""}",https://arxiv.org/pdf/2308.11119
-supplementary - i2mvformer: large language model generated multi-view document supervision for zero-shot image classification,Muhammad Ferjad Naeem,"In this supplementary, we perform additional experiments on I2MVFormer to confirm our design choices. Moreover, we perform further analysis of our LLM prompting strategy to confirm that it is robust to the choice of LLM, the choice of the prompt and the choice of k-shot examples. Finally, we qualitatively analyze the LLM Views to confirm that each view provides complementary information about a class which allows for a highly discriminative zero-shot image classification model. The content of the supplementary is organized as follows.","{'model': 'tldr@v2.0.0', 'text': 'This supplementary performs additional experiments on I2MVFormer and qualitatively analyzes the LLM Views to confirm that each view provides complementary information about a class which allows for a highly discriminative zero-shot image classification model.'}",
-a setwise approach for effective and highly efficient zero-shot ranking with large language models,Shengyao Zhuang,"Large Language Models (LLMs) demonstrate impressive effectiveness in zero-shot document ranking tasks. Pointwise, Pairwise, and Listwise prompting approaches have been proposed for LLM-based zero-shot ranking. Our study begins by thoroughly evaluating these existing approaches within a consistent experimental framework, considering factors like model size, token consumption, latency, among others. This first-of-its-kind comparative evaluation of these approaches allows us to identify the trade-offs between effectiveness and efficiency inherent in each approach. We find that while Pointwise approaches score high on efficiency, they suffer from poor effectiveness. Conversely, Pairwise approaches demonstrate superior effectiveness but incur high computational overhead. To further enhance the efficiency of LLM-based zero-shot ranking, we propose a novel Setwise prompting approach. Our approach reduces the number of LLM inferences and the amount of prompt token consumption during the ranking procedure, significantly improving the efficiency of LLM-based zero-shot ranking. We test our method using the TREC DL datasets and the BEIR zero-shot document ranking benchmark. The empirical results indicate that our approach considerably reduces computational costs while also retaining high zero-shot ranking effectiveness.","{'model': 'tldr@v2.0.0', 'text': 'This first-of-its-kind comparative evaluation of Pointwise, Pairwise, and Listwise prompting approaches for LLM-based zero-shot ranking finds that while Pointwise approaches score high on efficiency, they suffer from poor effectiveness and Pairwise approaches demonstrate superior effectiveness but incur high computational overhead.'}",
-reducing negative effects of the biases of language models in zero-shot setting,Xiaosu Wang,"Pre-trained language models (PLMs) such as GPTs have been revealed to be biased towards certain target classes because of the prompt and the model's intrinsic biases. In contrast to the fully supervised scenario where there are a large number of costly labeled samples that can be used to fine-tune model parameters to correct for biases, there are no labeled samples available for the zero-shot setting. We argue that a key to calibrating the biases of a PLM on a target task in zero-shot setting lies in detecting and estimating the biases, which remains a challenge. In this paper, we first construct probing samples with the randomly generated token sequences, which are simple but effective in detecting inputs for stimulating GPTs to show the biases; and we pursue an in-depth research on the plausibility of utilizing class scores for the probing samples to reflect and estimate the biases of GPTs on a downstream target task. Furtherly, in order to effectively utilize the probing samples and thus reduce negative effects of the biases of GPTs, we propose a lightweight model Calibration Adapter (CA) along with a self-guided training strategy that carries out distribution-level optimization, which enables us to take advantage of the probing samples to fine-tune and select only the proposed CA, respectively, while keeping the PLM encoder frozen. To demonstrate the effectiveness of our study, we have conducted extensive experiments, where the results indicate that the calibration ability acquired by CA on the probing samples can be successfully transferred to reduce negative effects of the biases of GPTs on a downstream target task, and our approach can yield better performance than state-of-the-art (SOTA) models in zero-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the calibration ability acquired by CA on the probing samples can be successfully transferred to reduce negative effects of the biases of GPTs on a downstream target task, and the approach can yield better performance than state-of-the-art (SOTA) models in zero-shot settings.'}",
-improving language model-based zero-shot text-to-speech synthesis with multi-scale acoustic prompts,Shunwei Lei,"Zero-shot text-to-speech (TTS) synthesis aims to clone any unseen speaker's voice without adaptation parameters. By quantizing speech waveform into discrete acoustic tokens and modeling these tokens with the language model, recent language model-based TTS models show zero-shot speaker adaptation capabilities with only a 3-second acoustic prompt of an unseen speaker. However, they are limited by the length of the acoustic prompt, which makes it difficult to clone personal speaking style. In this paper, we propose a novel zero-shot TTS model with the multi-scale acoustic prompts based on a neural codec language model VALL-E. A speaker-aware text encoder is proposed to learn the personal speaking style at the phoneme-level from the style prompt consisting of multiple sentences. Following that, a VALL-E based acoustic decoder is utilized to model the timbre from the timbre prompt at the frame-level and generate speech. The experimental results show that our proposed method outperforms baselines in terms of naturalness and speaker similarity, and can achieve better performance by scaling out to a longer style prompt.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel zero-shot TTS model with the multi-scale acoustic prompts based on a neural codec language model VALL-E, which outperforms baselines in terms of naturalness and speaker similarity, and can achieve better performance by scaling out to a longer style prompt.'}",https://arxiv.org/pdf/2309.11977
-model-generated pretraining signals improves zero-shot generalization of text-to-text transformers,Linyuan Gong,"This paper explores the effectiveness of model-generated signals in improving zero-shot generalization of text-to-text Transformers such as T5. We study various designs to pretrain T5 using an auxiliary model to construct more challenging token replacements for the main model to denoise. Key aspects under study include the decoding target, the location of the RTD head, and the masking pattern. Based on these studies, we develop a new model, METRO-T0, which is pretrained using the redesigned ELECTRA-Style pretraining strategies and then prompt-finetuned on a mixture of NLP tasks. METRO-T0 outperforms all similar-sized baselines on prompted NLP benchmarks, such as _T0 Eval_ and MMLU, and rivals the state-of-the-art T0-11B model with only **8%** of its parameters. Our analysis on model’s neural activation and parameter sensitivity reveals that the effectiveness of METRO-T0 stems from more balanced contribution of parameters and better utilization of their capacity. The code and model checkpoints are available at [https://github.com/gonglinyuan/metro_t0](https://github.com/gonglinyuan/metro_t0).","{'model': 'tldr@v2.0.0', 'text': 'A new model, METRO-T0 is developed, which is pretrained using the redesigned ELECTRA-Style pretraining strategies and then prompt-finetuned on a mixture of NLP tasks and rivals the state-of-the-art T0-11B model with only **8%** of its parameters.'}",http://arxiv.org/pdf/2305.12567
-interaction-aware prompting for zero-shot spatio-temporal action detection,Weiting Huang,"The goal of spatial-temporal action detection is to determine the time and place where each person's action occurs in a video and classify the corresponding action category. Most of the existing methods adopt fully-supervised learning, which requires a large amount of training data, making it very difficult to achieve zero-shot learning. In this paper, we propose to utilize a pre-trained visual-language model to extract the representative image and text features, and model the relationship between these features through different interaction modules to obtain the interaction feature. In addition, we use this feature to prompt each label to obtain more appropriate text features. Finally, we calculate the similarity between the interaction feature and the text feature for each label to determine the action category. Our experiments on J-HMDB and UCF101-24 datasets demonstrate that the proposed interaction module and prompting make the visual-language features better aligned, thus achieving excellent accuracy for zero-shot spatio-temporal action detection. The code will be available at https://github.com/webber2933/iCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A pre-trained visual-language model is utilized to extract the representative image and text features, and model the relationship between these features through different interaction modules to obtain the interaction feature, which is used to prompt each label to obtain more appropriate text features.'}",https://arxiv.org/pdf/2304.04688
-beyond yes and no: improving zero-shot llm rankers via scoring fine-grained relevance labels,Honglei Zhuang,"Zero-shot text rankers powered by recent LLMs achieve remarkable ranking performance by simply prompting. Existing prompts for pointwise LLM rankers mostly ask the model to choose from binary relevance labels like""Yes""and""No"". However, the lack of intermediate relevance label options may cause the LLM to provide noisy or biased answers for documents that are partially relevant to the query. We propose to incorporate fine-grained relevance labels into the prompt for LLM rankers, enabling them to better differentiate among documents with different levels of relevance to the query and thus derive a more accurate ranking. We study two variants of the prompt template, coupled with different numbers of relevance levels. Our experiments on 8 BEIR data sets show that adding fine-grained relevance labels significantly improves the performance of LLM rankers.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to incorporate fine-grained relevance labels into the prompt for LLM rankers, enabling them to better differentiate among documents with different levels of relevance to the query and thus derive a more accurate ranking.'}",
-zero-shot inversion process for image attribute editing with diffusion models,Zhanbo Feng,"Denoising diffusion models have shown outstanding performance in image editing. Existing works tend to use either image-guided methods, which provide a visual reference but lack control over semantic coherence, or text-guided methods, which ensure faithfulness to text guidance but lack visual quality. To address the problem, we propose the Zero-shot Inversion Process (ZIP), a framework that injects a fusion of generated visual reference and text guidance into the semantic latent space of a \textit{frozen} pre-trained diffusion model. Only using a tiny neural network, the proposed ZIP produces diverse content and attributes under the intuitive control of the text prompt. Moreover, ZIP shows remarkable robustness for both in-domain and out-of-domain attribute manipulation on real images. We perform detailed experiments on various benchmark datasets. Compared to state-of-the-art methods, ZIP produces images of equivalent quality while providing a realistic editing effect.","{'model': 'tldr@v2.0.0', 'text': 'The Zero-shot Inversion Process (ZIP), a framework that injects a fusion of generated visual reference and text guidance into the semantic latent space of a pre-trained diffusion model, shows remarkable robustness for both in-domain and out-of-domain attribute manipulation on real images.'}",https://arxiv.org/pdf/2308.15854
-exploring grounding potential of vqa-oriented gpt-4v for zero-shot anomaly detection,Jiangning Zhang,"Large Multimodal Model (LMM) GPT-4V(ision) endows GPT-4 with visual grounding capabilities, making it possible to handle certain tasks through the Visual Question Answering (VQA) paradigm. This paper explores the potential of VQA-oriented GPT-4V in the recently popular visual Anomaly Detection (AD) and is the first to conduct qualitative and quantitative evaluations on the popular MVTec AD and VisA datasets. Considering that this task requires both image-/pixel-level evaluations, the proposed GPT-4V-AD framework contains three components: 1) Granular Region Division, 2) Prompt Designing, 3) Text2Segmentation for easy quantitative evaluation, and have made some different attempts for comparative analysis. The results show that GPT-4V can achieve certain results in the zero-shot AD task through a VQA paradigm, such as achieving image-level 77.1/88.0 and pixel-level 68.0/76.6 AU-ROCs on MVTec AD and VisA datasets, respectively. However, its performance still has a certain gap compared to the state-of-the-art zero-shot method, e.g., WinCLIP ann CLIP-AD, and further research is needed. This study provides a baseline reference for the research of VQA-oriented LMM in the zero-shot AD task, and we also post several possible future works. Code is available at \url{https://github.com/zhangzjn/GPT-4V-AD}.","{'model': 'tldr@v2.0.0', 'text': 'The results show that GPT-4V can achieve certain results in the zero-shot AD task through a VQA paradigm, such as achieving image-level 77.1/88.0 and pixel-level 68.0/76.6 AU-ROCs on MVTec AD and VisA datasets, respectively.'}",
-zero-shot learning for named entity recognition in software specification documents,Souvick Das,"Named entity recognition (NER) is a natural language processing task that has been used in Requirements Engineering for the identification of entities such as actors, actions, operators, resources, events, GUI elements, hardware, APIs, and others. NER might be particularly useful for extracting key information from Software Requirements Specification documents, which provide a blueprint for software development. However, a common challenge in this domain is the lack of annotated data. In this article, we propose and analyze two zero-shot approaches for NER in the requirements engineering domain. These are found to be particularly effective in situations where labeled data is scarce or non-existent. The first approach is a template-based zero-shot learning mechanism that uses the prompt engineering approach and achieves 93% accuracy according to our experimental results. The second solution takes an orthogonal approach by transforming the entity recognition problem into a question-answering task which results in 98% accuracy. Both zero-shot NER approaches introduced in this work perform better than the existing state-of-the-art solutions in the requirements engineering domain.","{'model': 'tldr@v2.0.0', 'text': 'Two zero-shot NER approaches for NER in the requirements engineering domain are proposed and analyzed and are found to be particularly effective in situations where labeled data is scarce or non-existent.'}",
-zero-textcap: zero-shot framework for text-based image captioning,Dongsheng Xu,"Text-based image captioning is a vital but under-explored task, which aims to describe images by captions containing scene text automatically. Recent studies have made encouraging progress, but they are still suffering from two issues. Firstly, current models cannot capture and generate scene text in non-Latin script languages, which severely limits the objectivity and the information completeness of generated captions. Secondly, current models tend to describe images with monotonous and templated style, which greatly limits the diversity of the generated captions. Although the above-mentioned issues can be alleviated through carefully designed annotations, this process is undoubtedly laborious and time-consuming. To address the above issues, we propose a Zero-shot Framework for Text-based Image Captioning (Zero-TextCap). Concretely, to generate candidate sentences starting from the prompt 'Image of' and iteratively refine them to improve the quality and diversity of captions, we introduce a Hybrid-sampling masked language model (H-MLM). To read multi-lingual scene text and model the relationships between them, we introduce a robust OCR system. To ensure that the captions generated by H-MLM contain scene text and are highly relevant to the image, we propose a CLIP-based generation guidance module to insert OCR tokens and filter candidate sentences. Our Zero-TextCap is capable of generalizing captions containing multi-lingual scene text and boosting the diversity of captions. Sufficient experiments demonstrate the effectiveness of our proposed Zero-TextCap. Our codes are available at https://github.com/Gemhuang79/Zero_TextCap.","{'model': 'tldr@v2.0.0', 'text': 'The Zero-TextCap is capable of generalizing captions containing multi-lingual scene text and boosting the diversity of captions, and introduces a Hybrid-sampling masked language model (H-MLM) and a robust OCR system.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612571
-are soft prompts good zero-shot learners for speech recognition?,Dianwen Ng,"Large self-supervised pre-trained speech models require computationally expensive fine-tuning for downstream tasks. Soft prompt tuning offers a simple parameter-efficient alternative by utilizing minimal soft prompt guidance, enhancing portability while also maintaining competitive performance. However, not many people understand how and why this is so. In this study, we aim to deepen our understanding of this emerging method by investigating the role of soft prompts in automatic speech recognition (ASR). Our findings highlight their role as zero-shot learners in improving ASR performance but also make them vulnerable to malicious modifications. Soft prompts aid generalization but are not obligatory for inference. We also identify two primary roles of soft prompts: content refinement and noise information enhancement, which enhances robustness against background noise. Additionally, we propose an effective modification on noise prompts to show that they are capable of zero-shot learning on adapting to out-of-distribution noise environments.","{'model': 'tldr@v2.0.0', 'text': 'This study investigates the role of soft prompts in automatic speech recognition (ASR), and proposes an effective modification on noise prompts to show that they are capable of zero-shot learning on adapting to out-of-distribution noise environments.'}",https://arxiv.org/pdf/2309.09413
-blended-nerf: zero-shot object generation and blending in existing neural radiance fields,Ori Gordon,"Editing a local region or a specific object in a 3D scene represented by a NeRF or consistently blending a new realistic object into the scene is challenging, mainly due to the implicit nature of the scene representation. We present Blended-NeRF, a robust and flexible framework for editing a specific region of interest in an existing NeRF scene, based on text prompts, along with a 3D ROI box. Our method leverages a pretrained language-image model to steer the synthesis towards a user-provided text prompt, along with a 3D MLP model initialized on an existing NeRF scene to generate the object and blend it into a specified region in the original scene. We allow local editing by localizing a 3D ROI box in the input scene, and blend the content synthesized inside the ROI with the existing scene using a novel volumetric blending technique. To obtain natural looking and view-consistent results, we leverage existing and new geometric priors and 3D augmentations for improving the visual fidelity of the final result. We test our framework both qualitatively and quantitatively on a variety of real 3D scenes and text prompts, demonstrating realistic multi-view consistent results with much flexibility and diversity compared to the baselines. Finally, we show the applicability of our framework for several 3D editing applications, including adding new objects to a scene, removing/replacing/altering existing objects, and texture conversion.","{'model': 'tldr@v2.0.0', 'text': 'This work presents Blended-NeRF, a robust and flexible framework for editing a specific region of interest in an existing NeRF scene, based on text prompts, along with a 3D ROI box, and shows the applicability of the framework for several 3D editing applications, including adding new objects to a scene, removing/replacing/altering existing objects, and texture conversion.'}",https://arxiv.org/pdf/2306.12760
-global constraints with prompting for zero-shot event argument classification,Zizheng Lin,"Determining the role of event arguments is a crucial subtask of event extraction. Most previous supervised models leverage costly annotations, which is not practical for open-domain applications. In this work, we propose to use global constraints with prompting to effectively tackles event argument classification without any annotation and task-specific training. Specifically, given an event and its associated passage, the model first creates several new passages by prefix prompts and cloze prompts, where prefix prompts indicate event type and trigger span, and cloze prompts connect each candidate role with the target argument span. Then, a pre-trained language model scores the new passages, making the initial prediction. Our novel prompt templates can easily adapt to all events and argument types without manual effort. Next, the model regularizes the prediction by global constraints exploiting cross-task, cross-argument, and cross-event relations. Extensive experiments demonstrate our model’s effectiveness: it outperforms the best zero-shot baselines by 12.5% and 10.9% F1 on ACE and ERE with given argument spans and by 4.3% and 3.3% F1, respectively, without given argument spans. We have made our code publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use global constraints with prompting to effectively tackles event argument classification without any annotation and task-specific training, exploiting cross-task, cross-argument, and cross-event relations.'}",http://arxiv.org/pdf/2302.04459
-large language models are frame-level directors for zero-shot text-to-video generation,Susung Hong,"In the paradigm of AI-generated content (AIGC), there has been increasing attention in extending pre-trained text-to-image (T2I) models to text-to-video (T2V) generation. Despite their effectiveness, these frameworks face challenges in maintaining consistent narratives and handling rapid shifts in scene composition or object placement from a single user prompt. This paper introduces a new framework, dubbed DirecT2V, which leverages instruction-tuned large language models (LLMs) to generate frame-by-frame descriptions from a single abstract user prompt. DirecT2V utilizes LLM directors to divide user inputs into separate prompts for each frame, enabling the inclusion of time-varying content and facilitating consistent video generation. To maintain temporal consistency and prevent object collapse, we propose a novel value mapping method and dual-softmax filtering. Extensive experimental results validate the effectiveness of the DirecT2V framework in producing visually coherent and consistent videos from abstract user prompts, addressing the challenges of zero-shot video generation.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a new framework, dubbed DirecT2V, which leverages instruction-tuned large language models (LLMs) to generate frame-by-frame descriptions from a single abstract user prompt, and proposes a novel value mapping method and dual-softmax filtering.'}",http://arxiv.org/pdf/2305.14330
-sc vall-e: style-controllable zero-shot text to speech synthesizer,Daegyeom Kim,"Expressive speech synthesis models are trained by adding corpora with diverse speakers, various emotions, and different speaking styles to the dataset, in order to control various characteristics of speech and generate the desired voice. In this paper, we propose a style control (SC) VALL-E model based on the neural codec language model (called VALL-E), which follows the structure of the generative pretrained transformer 3 (GPT-3). The proposed SC VALL-E takes input from text sentences and prompt audio and is designed to generate controllable speech by not simply mimicking the characteristics of the prompt audio but by controlling the attributes to produce diverse voices. We identify tokens in the style embedding matrix of the newly designed style network that represent attributes such as emotion, speaking rate, pitch, and voice intensity, and design a model that can control these attributes. To evaluate the performance of SC VALL-E, we conduct comparative experiments with three representative expressive speech synthesis models: global style token (GST) Tacotron2, variational autoencoder (VAE) Tacotron2, and original VALL-E. We measure word error rate (WER), F0 voiced error (FVE), and F0 gross pitch error (F0GPE) as evaluation metrics to assess the accuracy of generated sentences. For comparing the quality of synthesized speech, we measure comparative mean option score (CMOS) and similarity mean option score (SMOS). To evaluate the style control ability of the generated speech, we observe the changes in F0 and mel-spectrogram by modifying the trained tokens. When using prompt audio that is not present in the training data, SC VALL-E generates a variety of expressive sounds and demonstrates competitive performance compared to the existing models. Our implementation, pretrained models, and audio samples are located on GitHub.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a style control (SC) VALL-E model based on the neural codec language model, which follows the structure of the generative pretrained transformer 3 (GPT-3), designed to generate controllable speech by not simply mimicking the characteristics of the prompt audio but by controlling the attributes to produce diverse voices.'}",https://arxiv.org/pdf/2307.10550
-language models as zero-shot trajectory generators,Teyun Kwon,"Large Language Models (LLMs) have recently shown promise as high-level planners for robots when given access to a selection of low-level skills. However, it is often assumed that LLMs do not possess sufficient knowledge to be used for the low-level trajectories themselves. In this work, we address this assumption thoroughly, and investigate if an LLM (GPT-4) can directly predict a dense sequence of end-effector poses for manipulation skills, when given access to only object detection and segmentation vision models. We study how well a single task-agnostic prompt, without any in-context examples, motion primitives, or external trajectory optimisers, can perform across 26 real-world language-based tasks, such as""open the bottle cap""and""wipe the plate with the sponge"", and we investigate which design choices in this prompt are the most effective. Our conclusions raise the assumed limit of LLMs for robotics, and we reveal for the first time that LLMs do indeed possess an understanding of low-level robot control sufficient for a range of common tasks, and that they can additionally detect failures and then re-plan trajectories accordingly. Videos, code, and prompts are available at: https://www.robot-learning.uk/language-models-trajectory-generators.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates if an LLM (GPT-4) can directly predict a dense sequence of end-effector poses for manipulation skills, when given access to only object detection and segmentation vision models, and reveals for the first time that LLMs do indeed possess an understanding of low-level robot control sufficient for a range of common tasks.'}",
-zeroprompt: streaming acoustic encoders are zero-shot masked lms,Xingcheng Song,"In this paper, we present ZeroPrompt (Figure 1-(a)) and the corresponding Prompt-and-Refine strategy (Figure 3), two simple but effective \textbf{training-free} methods to decrease the Token Display Time (TDT) of streaming ASR models \textbf{without any accuracy loss}. The core idea of ZeroPrompt is to append zeroed content to each chunk during inference, which acts like a prompt to encourage the model to predict future tokens even before they were spoken. We argue that streaming acoustic encoders naturally have the modeling ability of Masked Language Models and our experiments demonstrate that ZeroPrompt is engineering cheap and can be applied to streaming acoustic encoders on any dataset without any accuracy loss. Specifically, compared with our baseline models, we achieve 350 $\sim$ 700ms reduction on First Token Display Time (TDT-F) and 100 $\sim$ 400ms reduction on Last Token Display Time (TDT-L), with theoretically and experimentally equal WER on both Aishell-1 and Librispeech datasets.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that streaming acoustic encoders naturally have the modeling ability of Masked Language Models and the experiments demonstrate that ZeroPrompt is engineering cheap and can be applied to streaming acousticencoders on any dataset without any accuracy loss.'}",
-zero-shot text-driven physically interpretable face editing,Yapeng Meng,"This paper proposes a novel and physically interpretable method for face editing based on arbitrary text prompts. Different from previous GAN-inversion-based face editing methods that manipulate the latent space of GANs, or diffusion-based methods that model image manipulation as a reverse diffusion process, we regard the face editing process as imposing vector flow fields on face images, representing the offset of spatial coordinates and color for each image pixel. Under the above-proposed paradigm, we represent the vector flow field in two ways: 1) explicitly represent the flow vectors with rasterized tensors, and 2) implicitly parameterize the flow vectors as continuous, smooth, and resolution-agnostic neural fields, by leveraging the recent advances of implicit neural representations. The flow vectors are iteratively optimized under the guidance of the pre-trained Contrastive Language-Image Pretraining~(CLIP) model by maximizing the correlation between the edited image and the text prompt. We also propose a learning-based one-shot face editing framework, which is fast and adaptable to any text prompt input. Our method can also be flexibly extended to real-time video face editing. Compared with state-of-the-art text-driven face editing methods, our method can generate physically interpretable face editing results with high identity consistency and image quality. Our code will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'Compared with state-of-the-art text-driven face editing methods, this method can generate physically interpretable face editing results with high identity consistency and image quality.'}",https://arxiv.org/pdf/2308.05976
-applenet: visual attention parameterized prompt learning for few-shot remote sensing image generalization using clip,M. Singha,"In recent years, the success of large-scale vision-language models (VLMs) such as CLIP has led to their increased usage in various computer vision tasks. These models enable zero-shot inference through carefully crafted instructional text prompts without task-specific supervision. However, the potential of VLMs for generalization tasks in remote sensing (RS) has not been fully realized. To address this research gap, we propose a novel image-conditioned prompt learning strategy called the Visual Attention Parameterized Prompts Learning Network (APPLeNet). APPLeNet emphasizes the importance of multi-scale feature learning in RS scene classification and disentangles visual style and content primitives for domain generalization tasks. To achieve this, APPLeNet combines visual content features obtained from different layers of the vision encoder and style properties obtained from feature statistics of domain-specific batches. An attention-driven injection module is further introduced to generate visual tokens from this information. We also introduce an anti-correlation regularizer to ensure discrimination among the token embeddings, as this visual information is combined with the textual tokens. To validate APPLeNet, we curated four available RS benchmarks and introduced experimental protocols and datasets for three domain generalization tasks. Our results consistently outperform the relevant literature and code is available at https://github.com/mainaksingha01/APPLeNet","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel image-conditioned prompt learning strategy called the Visual Attention Parameterized Prompts Learning Network (APPLeNet), which emphasizes the importance of multi-scale feature learning in RS scene classification and disentangles visual style and content primitives for domain generalization tasks.'}",https://arxiv.org/pdf/2304.05995
-schema-aware reference as prompt improves data-efficient relational triple and event extraction,Yunzhi Yao,"Information Extraction, which aims to extract structural relational triple or event from unstructured texts, often suffers from data scarcity issues. With the development of pre-trained language models, many prompt-based approaches to data-efficient information extraction have been proposed and achieved impressive performance. However, existing prompt learning methods for information extraction are still susceptible to several potential limitations: (i) semantic gap between natural language and output structure knowledge with pre-defined schema; (ii) representation learning with locally individual instances limits the performance given the insufficient features. In this paper, we propose a novel approach of schema-aware R eference A s P rompt ( RAP ), which dynamically leverage schema and knowledge inherited from global (few-shot) training data for each sample. Specifically, we propose a schema-aware reference store , which unifies symbolic schema and relevant textual instances. Then, we employ a dynamic reference integration module to retrieve pertinent knowledge from the datastore as prompts during training and inference. Experimental results demonstrate that RAP can be plugged into various existing models and outperforms baselines in low-resource settings on four datasets of relational triple extraction and event extraction. In addition, we provide comprehensive empirical ablations and case analysis regarding different types and scales of knowledge in order to better understand the mechanisms of RAP 1 .","{'model': 'tldr@v2.0.0', 'text': 'A novel approach of schema-aware R eference A s P rompt ( RAP), which dynamically leverage schema and knowledge inherited from global (few-shot) training data for each sample, and employs a dynamic reference integration module to retrieve pertinent knowledge from the datastore as prompts during training and inference.'}",http://arxiv.org/pdf/2210.10709
-prompt combines paraphrase: teaching pre-trained models to understand rare biomedical words,Hao Wang,"Prompt-based fine-tuning for pre-trained models has proven effective for many natural language processing tasks under few-shot settings in general domain. However, tuning with prompt in biomedical domain has not been investigated thoroughly. Biomedical words are often rare in general domain, but quite ubiquitous in biomedical contexts, which dramatically deteriorates the performance of pre-trained models on downstream biomedical applications even after fine-tuning, especially in low-resource scenarios. We propose a simple yet effective approach to helping models learn rare biomedical words during tuning with prompt. Experimental results show that our method can achieve up to 6% improvement in biomedical natural language inference task without any extra parameters or training steps using few-shot vanilla prompt settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed simple yet effective approach to helping models learn rare biomedical words during tuning with prompt can achieve up to 6% improvement in biomedical natural language inference task without any extra parameters or training steps using few-shot vanilla prompt settings.'}",http://arxiv.org/pdf/2209.06453
-domain prompt learning for efficiently adapting clip to unseen domains,X. Zhang,"Domain generalization (DG) is a difficult transfer learning problem aiming to learn a generalizable model for unseen domains. Recent foundation models (FMs) are robust to many distribution shifts and, therefore, should substantially improve the performance of DG. In this work, we study generic ways to adopt CLIP, a Visual-Language Foundation Model, for DG problems in image classification. While ERM greatly improves the accuracy with bigger backbones and training datasets using standard DG benchmarks, fine-tuning FMs is not practical in many real-world situations. We propose Domain Prompt Learning (DPL) as a novel approach for domain inference in the form of conditional prompt generation. DPL achieved a significant accuracy improvement with only training a lightweight prompt generator (a three-layer MLP), whose parameter is of equivalent scale to the classification projector in the previous DG literature. Combining \dplshort~with CLIP provides surprising performance, raising the accuracy of zero-shot CLIP from 73.7% to 79.3% on several standard datasets, namely PACS, VLCS, OfficeHome, and TerraIncognita. We hope the simplicity and success of our approach lead to broader adoption and analysis of foundation models in the domain generalization field. Our code is available at https://github.com/shogi880/DPLCLIP.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Domain Prompt Learning (DPL) as a novel approach for domain inference in the form of conditional prompt generation and achieves a significant accuracy improvement with only training a lightweight prompt generator (a three-layer MLP), whose parameter is of equivalent scale to the classification projector in the previous DG literature.'}",https://www.jstage.jst.go.jp/article/tjsai/38/6/38_38-6_B-MC2/_pdf
-amortized prompt: lightweight fine-tuning for clip in domain generalization,Xin Zhang,"Domain generalization (DG) is a difficult transfer learning problem aiming to learn a generalizable model to unseen domains. Recent massive pretrained models such as CLIP and GPT-3, i.e. foundation models (FMs), have been shown to be robust to many distribution shifts and therefore should lead to substantial improvements in DG. In this work, we study generic ways to adopt CLIP for DG problems in image classification, where we evaluate on naive zero-shot learning and full DG learning settings. For the latter, we propose AP (Amortized Prompt), as a novel approach for domain inference in the form of prompt generation. Using several standard datasets on domain generalization benchmark, namely PACS, VLCS, OfficeHome, and TerraIncognita, CLIP provides comparable performance without fine-tuning any parameters, suggesting the applicability and importance of FM in DG. In addition, we show that combining domain prompt inference with CLIP enables AP to outperform strong baselines and the naive CLIP baselines by a large margin, raising accuracy from 71.3% to 79.3%. We hope the simplicity and success of our approach emphasizes the importance of and leads to wider more adoption and analysis of foundation models in the field of domain generalization. ∗E-mail:xin@weblab.t.u-tokyo.ac.jp Figure 1: The average performance on four DG datasets, VLCS, PACS, OfficeHome, TerraIncognita. We select Empirical Risk Minimization (ERM) [46] as a baselines due to its good performance showed in [24]. We surprisingly find that CLIP(ViT-B16) without any fine-tuning also can outperforms ERM baselines which fine-tune on source domains. Moreover, we demonstrate our AP effectively improves the performance of CLIP and outperforms the baseline ERM with ViT-B16 as a backbone by a large margin (8.3%). Since the choice of backbones is critical to ERM in DG, we propose to use CLIP with AP as the basic component. 1 ar X iv :2 11 1. 12 85 3v 1 [ cs .C V ] 2 5 N ov 2 02 1","{'model': 'tldr@v2.0.0', 'text': 'This work proposes AP (Amortized Prompt), as a novel approach for domain inference in the form of prompt generation, and surprisingly finds that CLIP (ViT-B16) without any fine-tuning also can outperforms ERM baselines which fine-tune on source domains.'}",
-feature normalization and cartography-based demonstrations for prompt-based fine-tuning on emotion-related tasks,Mahshid Hosseini,"To train a model in a traditional supervised learning classification system for natural language processing (NLP) tasks, it is essential to have labeled data, which is not present in large amounts for many tasks. Prompt-based learning methods attempt to combat the supervised learning need for labeled data by directly adapting pre-trained language models and modeling the probability of text itself. In this paper, we propose a novel data-agnostic strategy for prompt-based fine-tuning that leverages feature moments (a.k.a., mean and standard deviation) as a data augmentation technique and employs training dynamics (i.e., confidence and variability) to allow more informative samples to be concatenated for generating demonstrations as input context. Our approach is a strong method for few-shot learning that forces the language model to pay special attention to the feature moments and allows more informative samples to be concatenated for generating demonstrations as input context by selecting high confidence and low variance samples. To demonstrate its effectiveness given limited training data, we conduct extensive experiments in different few-shot settings on three empathy and emotion classification datasets (from various domains). We further evaluate our method's robustness by introducing noise to our few-shot input data and labels and show that exchanging moments between samples and incorporating cartography-based demonstrations are beneficial when the available data is limited and noisy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel data-agnostic strategy for prompt-based fine-tuning that leverages feature moments (a.k.a., mean and standard deviation) as a data augmentation technique and employs training dynamics to allow more informative samples to be concatenated for generating demonstrations as input context.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26514/26286
-understanding prompt engineering may not require rethinking generalization,Victor Akinwande,"Zero-shot learning in prompted vision-language models, the practice of crafting prompts to build classifiers without an explicit training process, has achieved impressive performance in many settings. This success presents a seemingly surprising observation: these methods suffer relatively little from overfitting, i.e., when a prompt is manually engineered to achieve low error on a given training set (thus rendering the method no longer actually zero-shot), the approach still performs well on held-out test data. In this paper, we show that we can explain such performance well via recourse to classical PAC-Bayes bounds. Specifically, we show that the discrete nature of prompts, combined with a PAC-Bayes prior given by a language model, results in generalization bounds that are remarkably tight by the standards of the literature: for instance, the generalization bound of an ImageNet classifier is often within a few percentage points of the true test error. We demonstrate empirically that this holds for existing handcrafted prompts and prompts generated through simple greedy search. Furthermore, the resulting bound is well-suited for model selection: the models with the best bound typically also have the best test performance. This work thus provides a possible justification for the widespread practice of prompt engineering, even if it seems that such methods could potentially overfit the training data.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the discrete nature of prompts, combined with a PAC-Bayes prior given by a language model, results in generalization bounds that are remarkably tight by the standards of the literature: for instance, the generalization bound of an ImageNet classifier is often within a few percentage points of the true test error.'}",
-few shot learning approaches to essay scoring,Robert K. Helmeczi,"Automated essay scoring (AES) involves using computer technology to grade written assessments and assigning a score based on their perceived quality. AES has been among the most significant Natural Language Processing (NLP) applications due to its educational and commercial value. Similar to many other NLP tasks, training a model for AES typically involves acquiring a substantial amount of labeled data specific to the essay being graded. This usually incurs a substantial cost. In this study, we consider two recent few-shot learning methods to enhance the predictive performance of machine learning methods for AES tasks. Specifically, we experiment with a prompt-based few-shot learning method, pattern exploiting training (PET), and a prompt-free few-shot learning strategy, SetFit, and compare these against vanilla fine-tuning. Our numerical study shows that PET can provide substantial performance gains over other methods, and it can effectively boost performance when access to labeled data is limited. On the other hand, PET is found to be the most computationally expensive few-shot learning method considered, while SetFit is the fastest method among the approaches.","{'model': 'tldr@v2.0.0', 'text': 'A numerical study shows that PET can provide substantial performance gains over other methods, and it can effectively boost performance when access to labeled data is limited, and PET is found to be the most computationally expensive few-shot learning method considered.'}",https://caiac.pubpub.org/pub/gdf5n6gs/download/pdf
-byoc: personalized few-shot classification with co-authored class descriptions,Arth Bohra,"Text classification is a well-studied and versatile building block for many NLP applications. Yet, existing approaches require either large annotated corpora to train a model with or, when using large language models as a base, require carefully crafting the prompt as well as using a long context that can fit many examples. As a result, it is not possible for end-users to build classifiers for themselves. To address this issue, we propose a novel approach to few-shot text classification using an LLM. Rather than few-shot examples, the LLM is prompted with descriptions of the salient features of each class. These descriptions are coauthored by the user and the LLM interactively: while the user annotates each few-shot example, the LLM asks relevant questions that the user answers. Examples, questions, and answers are summarized to form the classification prompt. Our experiments show that our approach yields high accuracy classifiers, within 82% of the performance of models trained with significantly larger datasets while using only 1% of their training sets. Additionally, in a study with 30 participants, we show that end-users are able to build classifiers to suit their specific needs. The personalized classifiers show an average accuracy of 90%, which is 15% higher than the state-of-the-art approach.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel approach to few-shot text classification using an LLM that yields high accuracy classifiers, within 82% of the performance of models trained with significantly larger datasets while using only 1% of their training sets.'}",https://arxiv.org/pdf/2310.06111
-prompt sketching for large language models,Luca Beurer-Kellner,"Many recent prompting strategies for large language models (LLMs) query the model multiple times sequentially -- first to produce intermediate results and then the final answer. However, using these methods, both decoder and model are unaware of potential follow-up prompts, leading to disconnected and undesirably wordy intermediate responses. In this work, we address this issue by proposing prompt sketching, a new prompting paradigm in which an LLM does not only respond by completing a prompt, but by predicting values for multiple variables in a template. This way, sketching grants users more control over the generation process, e.g., by providing a reasoning framework via intermediate instructions, leading to better overall results. The key idea enabling sketching with existing, autoregressive models is to adapt the decoding procedure to also score follow-up instructions during text generation, thus optimizing overall template likelihood in inference. Our experiments show that in a zero-shot setting, prompt sketching outperforms existing, sequential prompting schemes such as direct asking or chain-of-thought on 7 out of 8 LLM benchmarking tasks, including state tracking, arithmetic reasoning, and general question answering. To facilitate future use, we release a number of generic, yet effective sketches applicable to many tasks, and an open source library called dclib, powering our sketch-aware decoders.","{'model': 'tldr@v2.0.0', 'text': None}",
-strength in numbers: estimating confidence of large language models by prompt agreement,Gwenyth Portillo Wightman,"Large language models have achieved impressive few-shot performance on a wide variety of tasks. However, in many settings, users require confidence estimates for model predictions. While traditional classifiers produce scores for each label, language models instead produce scores for the generation which may not be well calibrated. We compare generations across diverse prompts and show that these can be used to create confidence scores. By utilizing more prompts we can get more precise confidence estimates and use response diversity as a proxy for confidence. We evaluate this approach across ten multiple-choice question-answering datasets using three models: T0, FLAN-T5, and GPT-3. In addition to analyzing multiple human written prompts, we automatically generate more prompts using a language model in order to produce finer-grained confidence estimates. Our method produces more calibrated confidence estimates compared to the log probability of the answer to a single prompt. These improvements could benefit users who rely on prediction confidence for integration into a larger system or in decision-making processes.","{'model': 'tldr@v2.0.0', 'text': 'This work automatically generates more prompts using a language model in order to produce finer-grained confidence estimates and produces more calibrated confidence estimates compared to the log probability of the answer to a single prompt.'}",
-the language of prompting: what linguistic properties make a prompt successful?,Alina Leidinger,"The latest generation of LLMs can be prompted to achieve impressive zero-shot or few-shot performance in many NLP tasks. However, since performance is highly sensitive to the choice of prompts, considerable effort has been devoted to crowd-sourcing prompts or designing methods for prompt optimisation. Yet, we still lack a systematic understanding of how linguistic properties of prompts correlate with task performance. In this work, we investigate how LLMs of different sizes, pre-trained and instruction-tuned, perform on prompts that are semantically equivalent, but vary in linguistic structure. We investigate both grammatical properties such as mood, tense, aspect and modality, as well as lexico-semantic variation through the use of synonyms. Our findings contradict the common assumption that LLMs achieve optimal performance on lower perplexity prompts that reflect language use in pretraining or instruction-tuning data. Prompts transfer poorly between datasets or models, and performance cannot generally be explained by perplexity, word frequency, ambiguity or prompt length. Based on our results, we put forward a proposal for a more robust and comprehensive evaluation standard for prompting research.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates how LLMs of different sizes, pre-trained and instruction-tuned, perform on prompts that are semantically equivalent, but vary in linguistic structure, and puts forward a proposal for a more robust and comprehensive evaluation standard for prompting research.'}",
-hard sample aware prompt-tuning,Yuanjiang Xu,"Prompt-tuning based few-shot learning has garnered increasing attention in recent years due to its efficiency and promising capability. To achieve the best performance for NLP tasks with just a few samples, it is vital to include as many informative samples as possible and to avoid misleading ones. However, there is no work in prompt-tuning literature addressing the problem of differentiating informative hard samples from misleading ones in model training, which is challenging due to the lack of supervision signals about the quality of the samples to train a well-performed model. We propose a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method. An extensive empirical study on a series of NLP tasks demonstrates the capability of HardPT in few-shot scenarios. HardPT obtains new SOTA results on all evaluated NLP tasks, including pushing the SST-5 accuracy to 49.5% (1.1% point absolute improvement), QNLI accuracy to 74.6% (1.9% absolute improvement), NMLI accuracy to 71.5 (0.7% absolute improvement), TACREV F_1-score to 28.2 (1.0 absolute improvement), and i2b2/VA F_1-score to 41.2 (1.3 absolute improvement).","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a Hard Sample Aware Prompt-Tuning framework (i.e. HardPT) to solve the non-differentiable problem in hard sample identification with reinforcement learning, and to strengthen the discrimination of the feature space without changing the original data distribution via an adaptive contrastive learning method.'}",https://aclanthology.org/2023.acl-long.690.pdf
-multi-view vision-prompt fusion network: can 2d pre-trained model boost 3d point cloud data-scarce learning?,Hao Peng,"Point cloud based 3D deep model has wide applications in many applications such as autonomous driving, house robot, and so on. Inspired by the recent prompt learning in natural language processing, this work proposes a novel Multi-view Vision-Prompt Fusion Network (MvNet) for few-shot 3D point cloud classification. MvNet investigates the possibility of leveraging the off-the-shelf 2D pre-trained models to achieve the few-shot classification, which can alleviate the over-dependence issue of the existing baseline models towards the large-scale annotated 3D point cloud data. Specifically, MvNet first encodes a 3D point cloud into multi-view image features for a number of different views. Then, a novel multi-view prompt fusion module is developed to effectively fuse information from different views to bridge the gap between 3D point cloud data and 2D pre-trained models. A set of 2D image prompts can then be derived to better describe the suitable prior knowledge for a large-scale pre-trained image model for few-shot 3D point cloud classification. Extensive experiments on ModelNet, ScanObjectNN, and ShapeNet datasets demonstrate that MvNet achieves new state-of-the-art performance for 3D few-shot point cloud image classification. The source code of this work will be available soon.","{'model': 'tldr@v2.0.0', 'text': 'MvNet investigates the possibility of leveraging the off-the-shelf 2D pre-trained models to achieve the few-shot classification, which can alleviate the over-dependence issue of the existing baseline models towards the large-scale annotated 3D point cloud data.'}",https://arxiv.org/pdf/2304.10224
-voucher abuse detection with prompt-based fine-tuning on graph neural networks,Zhi Wen,"Voucher abuse detection is an important anomaly detection problem in E-commerce. While many GNN-based solutions have emerged, the supervised paradigm depends on a large quantity of labeled data. A popular alternative is to adopt self-supervised pre-training using label-free data, and further fine-tune on a downstream task with limited labels. Nevertheless, the ""pre-train, fine-tune"" paradigm is often plagued by the objective gap between pre-training and downstream tasks. Hence, we propose VPGNN, a prompt-based fine-tuning framework on GNNs for voucher abuse detection. We design a novel graph prompting function to reformulate the downstream task into a similar template as the pretext task in pre-training, thereby narrowing the objective gap. Extensive experiments on both proprietary and public datasets demonstrate the strength of VPGNN in both few-shot and semi-supervised scenarios. Moreover, an online deployment of VPGNN in a production environment shows a 23.4% improvement over two existing deployed models.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes VPGNN, a prompt-based fine-tuning framework on GNNs for voucher abuse detection, and designs a novel graph prompting function to reformulate the downstream task into a similar template as the pretext task in pre-training, thereby narrowing the objective gap.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615505
-modal interaction-enhanced prompt learning by transformer decoder for vision-language models,Mingyue Liu,,"{'model': 'tldr@v2.0.0', 'text': 'A prompt tuning method for simulating image-text interaction based on CoOp: Decoding context optimization (DeCoOp) is proposed, demonstrating that this approach substantially outperforms the baseline model CoOp in terms of point domain generalization and robustness.'}",
-stabilized in-context learning with pre-trained language models for few shot dialogue state tracking,Derek Chen,"Prompt-based methods with large pre-trained language models (PLMs) have shown impressive unaided performance across many NLP tasks. These models improve even further with the addition of a few labeled in-context exemplars to guide output generation. However, for more complex tasks such as dialogue state tracking (DST), designing prompts that reliably convey the desired intent is nontrivial, leading to unstable results. Furthermore, building in-context exemplars for dialogue tasks is difficult because conversational contexts are long while model input lengths are relatively short.To overcome these issues we first adapt a meta-learning scheme to the dialogue domain which stabilizes the ability of the model to perform well under various prompts. We additionally design a novel training method to improve upon vanilla retrieval mechanisms to find ideal in-context examples. Finally, we introduce a saliency model to limit dialogue text length, allowing us to include more exemplars per query. In effect, we are able to achieve highly competitive results for few-shot DST on MultiWOZ.","{'model': 'tldr@v2.0.0', 'text': 'A meta-learning scheme is adapted to the dialogue domain which stabilizes the ability of the model to perform well under various prompts and introduces a saliency model to limit dialogue text length, allowing for highly competitive results for few-shot DST on MultiWOZ.'}",http://arxiv.org/pdf/2302.05932
-proposal of range verification method for carbon therapy using prompt gamma-rays over 10 mev,S. Kimura,"Range verification system is a most important part of particle therapy, for beam’s high killing power and convergence. To verify the range, many kinds of methods are devised and have been studied. We have studied the system which detects prompt gamma-rays over 10 MeV by using pair production events. We simulated the distribution of prompt gamma-rays’ energy, vertex positions, final scattering positions, and its generation process by Geant4 Monte Carlo simulation toolkit. We used physics lists named ""QGSP_INCLXX_HP"" and ""G4RadioactiveDecayPhysics"". PMMA cylinder with a diameter of 30 cm was put in the air. Ten million carbon beams with the energy of 290 MeV/n were shot to the PMMA target. We also simulated a more simple detector. It is consist of arrayed scintillation plates, a lead collimator, and a plastic scintillation plate. The plastic scintillation plate is used for vetoing charged particles. The lead collimator is separated from the centre, and move for shielding completely. This system makes it possible to reduce neutron derived background. The result of the simulation says our detector can measure the Bragg peak position.",,
-claret: pre-training a correlation-aware context-to-event transformer for event-centric generation and classification,Yucheng Zhou,"Generating new events given context with correlated ones plays a crucial role in many event-centric reasoning tasks. Existing works either limit their scope to specific scenarios or overlook event-level correlations. In this paper, we propose to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning. To achieve this, we propose three novel event-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training. The proposed ClarET is applicable to a wide range of event-centric reasoning scenarios, considering its versatility of (i) event-correlation types (e.g., causal, temporal, contrast), (ii) application formulations (i.e., generation and classification), and (iii) reasoning types (e.g., abductive, counterfactual and ending reasoning). Empirical fine-tuning results, as well as zero- and few-shot learning, on 9 benchmarks (5 generation and 4 classification tasks covering 4 reasoning types with diverse event correlations), verify its effectiveness and generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning, and proposes three novelevent-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training.'}",http://arxiv.org/pdf/2203.02225
-does gpt-3 generate empathetic dialogues? a novel in-context example selection method and automatic evaluation metric for empathetic dialogue generation,Young-Jun Lee,"Since empathy plays a crucial role in increasing social bonding between people, many studies have designed their own dialogue agents to be empathetic using the well-established method of fine-tuning. However, they do not use prompt-based in-context learning, which has shown powerful performance in various natural language processing (NLP) tasks, for empathetic dialogue generation. Although several studies have investigated few-shot in-context learning for empathetic dialogue generation, an in-depth analysis of the generation of empathetic dialogue with in-context learning remains unclear, especially in GPT-3 (Brown et al., 2020). In this study, we explore whether GPT-3 can generate empathetic dialogues through prompt-based in-context learning in both zero-shot and few-shot settings. To enhance performance, we propose two new in-context example selection methods, called SITSM and EMOSITSM, that utilize emotion and situational information. We also introduce a new automatic evaluation method, DIFF-EPITOME, which reflects the human tendency to express empathy. From the analysis, we reveal that our DIFF-EPITOME is effective in measuring the degree of human empathy. We show that GPT-3 achieves competitive performance with Blender 90M, a state-of-the-art dialogue generative model, on both automatic and human evaluation. Our code is available at https://github.com/passing2961/EmpGPT-3.","{'model': 'tldr@v2.0.0', 'text': 'This study explores whether GPT-3 can generate empathetic dialogues through prompt-based in-context learning in both zero-shot and few-shot settings and introduces a new automatic evaluation method, DIFF-EPITOME, which reflects the human tendency to express empathy.'}",
-exploring the landscape of distributional robustness for question answering models,Anas Awadalla,"We conduct a large empirical evaluation to investigate the landscape of distributional robustness in question answering. Our investigation spans over 350 models and 16 question answering datasets, including a diverse set of architectures, model sizes, and adaptation methods (e.g., fine-tuning, adapter tuning, in-context learning, etc.). We find that, in many cases, model variations do not affect robustness and in-distribution performance alone determines out-of-distribution performance. Moreover, our findings indicate that i) zero-shot and in-context learning methods are more robust to distribution shifts than fully fine-tuned models; ii) few-shot prompt fine-tuned models exhibit better robustness than few-shot fine-tuned span prediction models; iii) parameter-efficient and robustness enhancing training methods provide no significant robustness improvements. In addition, we publicly release all evaluations to encourage researchers to further analyze robustness trends for question answering models.","{'model': 'tldr@v2.0.0', 'text': 'This investigation spans over 350 models and 16 question answering datasets, including a diverse set of architectures, model sizes, and adaptation methods, and indicates that zero-shot and in-context learning methods are more robust to distribution shifts than fully fine-tuned models.'}",http://arxiv.org/pdf/2210.12517
-emotionprompt: leveraging psychology for large language models enhancement via emotional stimulus,Cheng Li,"Large language models (LLMs) have achieved significant performance in many fields, such as reasoning, language understanding, and math problem-solving, and are regarded as an important step to artificial general intelligence (AGI). However, the sensitivity of LLMs to prompts remain a major bottleneck for their daily adoption. In this paper, we take inspiration from psychology and propose EmotionPrompt to explore emotional intelligence to enhance the performance of LLMs. Our EmotionPrompt operates on a remarkably straightforward principle: the incorporation of emotional stimulus into prompts. Experimental re-sults demonstrate that our EmotionPrompt, using the same single prompt templates, significantly out-performs the original prompt and Zero-shot-CoT in both zero-shot and few-shot settings on eight tasks with diverse models: ChatGPT, Vicuna-13b, Bloom, and Flan-T5-large. Furthermore, Emotion-Prompt was observed to improve both the truthfulness and informativeness. We believe that Emo-tionPrompt heralds a novel avenue for exploring interdisciplinary knowledge for interaction between humans and LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Emo-tionPrompt heralds a novel avenue for exploring interdisciplinary knowledge for interaction between humans and LLMs and was observed to improve both the truthfulness and informativeness.'}",https://arxiv.org/pdf/2307.11760
-scone: benchmarking negation reasoning in language models with fine-tuning and in-context learning,Jingyuan Selena She,"A number of recent benchmarks seek to assess how well models handle natural language negation. However, these benchmarks lack the controlled example paradigms that would allow us to infer whether a model had truly learned how negation morphemes semantically scope. To fill these analytical gaps, we present the Scoped Negation NLI (ScoNe-NLI) benchmark, which contains contrast sets of six examples with up to two negations where either zero, one, or both negative morphemes affect the NLI label. We use ScoNe-NLI to assess fine-tuning and in-context learning strategies. We find that RoBERTa and DeBERTa models solve ScoNe-NLI after many shot fine-tuning. For in-context learning, we test the latest InstructGPT models and find that most prompt strategies are not successful, including those using step-by-step reasoning. To better understand this result, we extend ScoNe with ScoNe-NLG, a sentence completion test set that embeds negation reasoning in short narratives. Here, InstructGPT is successful, which reveals the model can correctly reason about negation, but struggles to do so on NLI examples outside of its core pretraining regime.","{'model': 'tldr@v2.0.0', 'text': 'This work presents the Scoped Negation NLI (ScoNe-NLI) benchmark, which contains contrast sets of six examples with up to two negations where either zero, one, or both negative morphemes affect the NLI label, and extends ScoNe with ScoNe-NLG, a sentence completion test set that embeds negation reasoning in short narratives.'}",http://arxiv.org/pdf/2305.19426
-enabling classifiers to make judgements explicitly aligned with human values,Yejin Bang,"Many NLP classification tasks, such as sexism/racism detection or toxicity detection, are based on human values. Yet, human values can vary under diverse cultural conditions. Therefore, we introduce a framework for value-aligned classification that performs prediction based on explicitly written human values in the command. Along with the task, we propose a practical approach that distills value-aligned knowledge from large-scale language models (LLMs) to construct value-aligned classifiers in two steps.First, we generate value-aligned training data from LLMs by prompt-based few-shot learning. Next, we fine-tune smaller classification models with the generated data for the task. Empirical results show that our VA-Models surpass multiple baselines by at least 15.56% on the F1-score, including few-shot learning with OPT-175B and existing text augmentation methods. We suggest that using classifiers with explicit human value input improves both inclusivity & explainability in AI.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a framework for value-aligned classification that performs prediction based on explicitly written human values in the command and suggests that using classifiers with explicit human value input improves both inclusivity & explainability in AI.'}",http://arxiv.org/pdf/2210.07652
-bits of grass: does gpt already know how to write like whitman?,Piotr Sawicki,"This study examines the ability of GPT-3.5, GPT-3.5-turbo (ChatGPT) and GPT-4 models to generate poems in the style of specific authors using zero-shot and many-shot prompts (which use the maximum context length of 8192 tokens). We assess the performance of models that are not fine-tuned for generating poetry in the style of specific authors, via automated evaluation. Our findings indicate that without fine-tuning, even when provided with the maximum number of 17 poem examples (8192 tokens) in the prompt, these models do not generate poetry in the desired style.",,http://arxiv.org/pdf/2305.11064
-do prompts solve nlp tasks using natural language?,Sen Yang,"Thanks to the advanced improvement of large pre-trained language models, prompt-based fine-tuning is shown to be effective on a variety of downstream tasks. Though many prompting methods have been investigated, it remains unknown which type of prompts are the most effective among three types of prompts (i.e., human-designed prompts, schema prompts and null prompts). In this work, we empirically compare the three types of prompts under both few-shot and fully-supervised settings. Our experimental results show that schema prompts are the most effective in general. Besides, the performance gaps tend to diminish when the scale of training data grows large.","{'model': 'tldr@v2.0.0', 'text': 'This work empirically compare the three types of prompts under both few-shot and fully-supervised settings and shows that schema prompts are the most effective in general.'}",http://arxiv.org/pdf/2203.00902
-evaluation of a single-stage light-gas gun facility in malta: business analysis and preliminary design,L. Barilaro,,,https://link.springer.com/content/pdf/10.1007/s42496-022-00113-w.pdf
-cliptexture: text-driven texture synthesis,Yiren Song,"Can artificial intelligence create textures with artistic value according to human language control? Existing texture synthesis methods require example texture input. However, in many practical situations, users don't have satisfying textures but tell designers about their needs through simple sketches and verbal descriptions. This paper proposes a novel texture synthesis framework based on the CLIP, which models the texture synthesis problem as an optimization process and realizes text-driven texture synthesis by minimizing the distance between the input image and the text prompt in latent space. Our method performs zero-shot image manipulation successfully even between unseen domains. We implement texture synthesis using two different optimization methods, the TextureNet and Diffvg, demonstrating the generality of CLIPTexture. Extensive experiments confirmed the robust and superior manipulation performance of our methods compared to the existing baselines.","{'model': 'tldr@v2.0.0', 'text': 'A novel texture synthesis framework based on the CLIP is proposed, which models the texture synthesis problem as an optimization process and realizes text-driven texture synthesis by minimizing the distance between the input image and the text prompt in latent space.'}",
-bertnet: harvesting knowledge graphs with arbitrary relations from pretrained language models,Shibo Hao,"It is crucial to automatically construct knowledge graphs (KGs) of diverse new relations to support knowledge discovery and broad applications. Previous KG construction methods, based on either crowdsourcing or text mining, are often limited to a small predefined set of relations due to manual cost or restrictions in text corpus. Recent research proposed to use pretrained language models (LMs) as implicit knowledge bases that accept knowledge queries with prompts. Yet, the implicit knowledge lacks many desirable properties of a full-scale symbolic KG, such as easy access, navigation, editing, and quality assurance. In this paper, we propose a new approach of harvesting massive KGs of arbitrary relations from pretrained LMs. With minimal input of a relation definition (a prompt and a few shot of example entity pairs), the approach efficiently searches in the vast entity pair space to extract diverse accurate knowledge of the desired relation. We develop an effective search-and-rescore mechanism for improved efficiency and accuracy. We deploy the approach to harvest KGs of over 400 new relations from different LMs. Extensive human and automatic evaluations show our approach manages to extract diverse accurate knowledge, including tuples of complex relations (e.g.,""A is capable of but not good at B""). The resulting KGs as a symbolic interpretation of the source LMs also reveal new insights into the LMs' knowledge capacities.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new approach of harvesting massive KGs of arbitrary relations from pretrained LMs with minimal input of a relation definition, and develops an effective search-and-rescore mechanism for improved efficiency and accuracy.'}",https://aclanthology.org/2023.findings-acl.309.pdf
-learning disentangled prompts for compositional image synthesis,Kihyuk Sohn,"We study domain-adaptive image synthesis, the problem of teaching pretrained image generative models a new style or concept from as few as one image to synthesize novel images, to better understand the compositional image synthesis. We present a framework that leverages a pretrained class-conditional generation model and visual prompt tuning. Specifically, we propose a novel source class distilled visual prompt that learns disentangled prompts of semantic (e.g., class) and domain (e.g., style) from a few images. Learned domain prompt is then used to synthesize images of any classes in the style of target domain. We conduct studies on various target domains with the number of images ranging from one to a few to many, and show qualitative results which show the compositional generalization of our method. Moreover, we show that our method can help improve zero-shot domain adaptation classification accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel source class distilled visual prompt that learns disentangled prompts of semantic and domain from a few images, and shows that this method can help improve zero-shot domain adaptation classification accuracy.'}",http://arxiv.org/pdf/2306.00763
-language models as black-box optimizers for vision-language models,Samuel Yu,"Vision-language models (VLMs) pre-trained on web-scale datasets have demonstrated remarkable capabilities across a variety of vision and multimodal tasks. Currently, fine-tuning methods for VLMs mainly operate in a white-box setting, requiring access to model parameters for backpropagation. However, many VLMs rely on proprietary data and are not open-source, which restricts the use of white-box approaches for fine-tuning. Given that popular private large language models (LLMs) like ChatGPT still offer a language-based user interface, we aim to develop a novel fine-tuning approach for VLMs through natural language prompts, thereby avoiding the need to access model parameters, feature embeddings, or output logits. In this setup, we propose employing chat-based LLMs as black-box optimizers to search for the best text prompt on the illustrative task of few-shot image classification using CLIP. Specifically, we adopt an automatic""hill-climbing""procedure that converges on an effective prompt by evaluating the accuracy of current prompts and asking LLMs to refine them based on textual feedback, all within a conversational process without human-in-the-loop. In a challenging 1-shot learning setup, our simple approach surpasses the white-box continuous prompting method (CoOp) by an average of 1.5% across 11 datasets including ImageNet. Our approach also outperforms OpenAI's manually crafted prompts. Additionally, we highlight the advantage of conversational feedback that incorporates both positive and negative prompts, suggesting that LLMs can utilize the implicit""gradient""direction in textual feedback for a more efficient search. Lastly, we find that the text prompts generated through our strategy are not only more interpretable but also transfer well across different CLIP architectures in a black-box manner.","{'model': 'tldr@v2.0.0', 'text': 'This work aims to develop a novel fine-tuning approach for VLMs through natural language prompts, thereby avoiding the need to access model parameters, feature embeddings, or output logits, and finds that the text prompts generated through this strategy are not only more interpretable but also transfer well across different CLIP architectures in a black-box manner.'}",https://arxiv.org/pdf/2309.05950
-effects of target words and their locations in prompts,T. Bers,"Few and zero-shot learning with prompt-tuned models has recently been shown to be very successful. Prompt-tuned models require prompt engineering where the model is trained on many different prompts. A prompt is some text that is added to an input example in order to convert it into a language modeling task. For example instead of just training on pairs of hypotheses and premises, the prompted input text would be: “Given { premise } , does this follow: { hypothesis } ?” A challenge is figuring out which prompt wording allows the model to perform best. In this thesis, we experiment with different target words and their locations; either not included in the prompt, at the start, or at the end of the prompt. Target words are the set of possible response to the prompt, for the previous example the target words could be yes and no. We look at two different models, instruction tuned (T0) and not (ALBERT). We also analyze the difference between varying difficulties of prompts in Natural Language Inference (NLI) and across different tasks (NLI, coreference resolution, sentence completion, and multiple choice question and answer). In sum, we find 1) that yes/no and true/false perform best as the target words, 2) the impact of location of the target words on T0 varies by task and results are inconsistent, and 3) that including them as a negative impact on ALBERT.","{'model': 'tldr@v2.0.0', 'text': 'It is found that yes/no and true/false perform best as the target words, the impact of location of the target Words on T0 varies by task and results are inconsistent, and that including them as a negative impact on ALBERT is found.'}",
-weak supervision for question type detection with large language models,J. Martínek,"Large pre-trained language models (LLM) have shown remark-able Zero-Shot Learning performances in many Natural Language Processing tasks. However, designing effective prompts is still very difficult for some tasks, in particular for dialogue act recognition. We propose an alternative way to leverage pretrained LLM for such tasks that replace manual prompts with simple rules, which are more intuitive and easier to design for some tasks. We demonstrate this approach on the question type recognition task, and show that our zero-shot model obtains competitive performances both with a supervised LSTM trained on the full training corpus, and another supervised model from previously published works on the MRDA corpus. We further analyze the limits of the proposed approach, which can not be applied on any task, but may advantageously complement prompt programming for specific classes.","{'model': 'tldr@v2.0.0', 'text': 'The proposed zero-shot model obtains competitive performances both with a supervised LSTM trained on the full training corpus, and another supervised model from previously published works on the MRDA corpus.'}",https://hal.science/hal-03786135/document
-b-pet: the pet model with parameter-efficient learning,Qi Zheng,"In recent years, under the trend of training models in big data, Few-shot learning (FSL) which aims to learn models to solve problems with a few samples has also achieved good results on many data sets. In fact, acquiring high-quality training samples is expensive in many aspects, but FSL can save the overhead costs. Among FSL models, the PET model combines semi-supervised learning, prompt learning and knowledge distillation based on the pre-training language model. However, in fine-turning the PET model has the disadvantages that consumes a lot of resources and time and requires heavy costs of storage for model preservation. Therefore, this paper proposes the B-pet model, which freezes most of the training parameters and only trains bias parameters during fine-turning process, significantly reducing the storage consumption of the model for downstream tasks. We used six data sets with $\vert \tau \vert=\mathbf{10},\ \mathbf{50},\ \mathbf{100}$ and three different data training models respectively. The results show that four data sets on the B-pet model performed better than original PET model training. It is obvious that in the memory-constrained environment deployment, multitasking fine-tunes models have practical value. It also proved that most semi-supervised models with fixed parameters are realizable.","{'model': 'tldr@v2.0.0', 'text': 'The B-pet model is proposed, which freezes most of the training parameters and only trains bias parameters during fine-turning process, significantly reducing the storage consumption of the model for downstream tasks and proved that most semi-supervised models with fixed parameters are realizable.'}",
-automatic data transformation using large language model: an experimental study on building energy data,Ankita Sharma,"Existing approaches to automatic data transformation are insufficient to meet the requirements in many real-world scenarios, such as the building sector. First, there is no convenient interface for domain experts to provide domain knowledge easily. Second, they require significant training data collection overheads. Third, the accuracy suffers from complicated schema changes. To bridge this gap, we present a novel approach that leverages the unique capabilities of large language models (LLMs) in coding, complex reasoning, and zero-shot learning to generate SQL code that transforms the source datasets into the target datasets. We demonstrate the viability of this approach by designing an LLM-based framework, termed SQLMorpher, which comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases. It also implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection. The key contributions of this work include (1) pioneering an end-to-end LLM-based solution for data transformation, (2) developing a benchmark dataset of 105 real-world building energy data transformation problems, and (3) conducting an extensive empirical evaluation where our approach achieved 96% accuracy in all 105 problems. SQLMorpher demonstrates the effectiveness of utilizing LLMs in complex, domain-specific challenges, highlighting the potential of their potential to drive sustainable solutions.","{'model': 'tldr@v2.0.0', 'text': 'This work designs an LLM-based framework that comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases, and implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection.'}",https://arxiv.org/pdf/2309.01957
-grass: unified generation model for speech-to-semantic tasks,Aobo Xia,"This paper explores the instruction fine-tuning technique for speech-to-semantic tasks by introducing a unified end-to-end (E2E) framework that generates target text conditioned on a task-related prompt for audio data. We pre-train the model using large and diverse data, where instruction-speech pairs are constructed via a text-to-speech (TTS) system. Extensive experiments demonstrate that our proposed model achieves state-of-the-art (SOTA) results on many benchmarks covering speech named entity recognition, speech sentiment analysis, speech question answering, and more, after fine-tuning. Furthermore, the proposed model achieves competitive performance in zero-shot and few-shot scenarios. To facilitate future work on instruction fine-tuning for speech-to-semantic tasks, we release our instruction dataset and code.","{'model': 'tldr@v2.0.0', 'text': 'A unified end-to-end (E2E) framework that generates target text conditioned on a task-related prompt for audio data is introduced that achieves state-of-the-art (SOTA) results on many benchmarks covering speech named entity recognition, speech sentiment analysis, speech question answering, and more, after fine-tuning.'}",https://arxiv.org/pdf/2309.02780
-leveraging vision-language foundation models for fine-grained downstream tasks,Denis Coquenet,"Vision-language foundation models such as CLIP have shown impressive zero-shot performance on many tasks and datasets, especially thanks to their free-text inputs. However, they struggle to handle some downstream tasks, such as fine-grained attribute detection and localization. In this paper, we propose a multitask fine-tuning strategy based on a positive/negative prompt formulation to further leverage the capacities of the vision-language foundation models. Using the CLIP architecture as baseline, we show strong improvements on bird fine-grained attribute detection and localization tasks, while also increasing the classification performance on the CUB200-2011 dataset. We provide source code for reproducibility purposes: it is available at https://github.com/FactoDeepLearning/MultitaskVLFM.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multitask fine-tuning strategy based on a positive/negative prompt formulation to further leverage the capacities of the vision-language foundation models and shows strong improvements on bird fine-grained attribute detection and localization tasks.'}",https://arxiv.org/pdf/2307.06795
-towards expert systems for improved customer services using chatgpt as an inference engine,C. P. Ezenkwu,"By harnessing both implicit and explicit customer data, companies can develop a more comprehensive understanding of their consumers, leading to better customer engagement and experience, and improved loyalty. As a result, businesses have embraced many AI technologies, including chatbots, sentiment analysis, voice assistants, predictive analytics, and natural language processing, within customer services and e-commerce. The arrival of ChatGPT, a state-of-the-art deep learning model trained with general knowledge in mind, has brought about a paradigm shift in how companies approach AI applications. However, given that most business problems are bespoke and require specialised domain expertise, ChatGPT needs to be aligned with the requisite task-oriented ability to solve these issues. This paper presents an iterative procedure that incorporates expert system development process models and prompt engineering, in the design of descriptive knowledge and few-shot prompts, as are necessary for ChatGPT-powered expert systems applications within customer services. Furthermore, this paper explores potential application areas for ChatGPT-powered expert systems in customer services, presenting opportunities for their effective utilisation in the business sector.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents an iterative procedure that incorporates expert system development process models and prompt engineering, in the design of descriptive knowledge and few-shot prompts, as are necessary for ChatGPT-powered expert systems applications within customer services.'}",https://rgu-repository.worktribe.com/preview/1987218/EZENKWU%202023%20Towards%20expert%20systems%20%28AAM%29.pdf
-generating domain-specific programs for diagram authoring with large language models,Rijul Jain,"Large language models (LLMs) can generate programs in general-purpose languages from prose descriptions, but are not trained on many domain-specific languages (DSLs). Diagram authoring with Penrose, a diagramming system using three DSLs, exemplifies the utility of DSL program generation with LLMs, which enables diagram creation from prose. We provide methods to conceptualize and evaluate the structures of one-shot LLM prompts to generate error-free DSL programs and implement Penrose diagram creation from prose using LLMs. We will evaluate our LLM prompt structures by testing prompt variations across different diagramming domains and plan to run a user study to assess the ease of LLM-augmented Penrose diagramming over other tools.","{'model': 'tldr@v2.0.0', 'text': 'Methods to conceptualize and evaluate the structures of one-shot LLM prompts to generate error-free DSL programs and implement Penrose diagram creation from prose using LLMs are provided.'}",
-rewriting math word problems with large language models,Kole Norberg,"Large Language Models have recently achieved high performance on many writing tasks. In a recent study, math word problems in Carnegie Learning’s MATHia adaptive learning software were rewritten by human authors to improve their clarity and specificity. The randomized experiment found that emerging readers who received the rewritten word problems spent less time completing the problems and also achieved higher mastery compared to emerging readers who received the original content. We used GPT-4 to rewrite the same set of math word problems, prompting it to follow the same guidelines that the human authors followed. We lay out our prompt engineering process, comparing several prompting strategies: zero-shot, few-shot, and chain-of-thought prompting. Additionally, we overview how we leveraged GPT’s ability to write python code in order to encode mathematical components of word problems. We report text analysis of the original, human-rewritten, and GPT-rewritten problems. GPT rewrites had the most optimal readability, lexical diversity, and cohesion scores but used more low frequency words. We present our plan to test the GPT outputs in upcoming randomized field trials in MATHia.","{'model': 'tldr@v2.0.0', 'text': 'This work used GPT-4 to rewrite the same set of math word problems, prompting it to follow the same guidelines that the human authors followed, and found that GPT rewrites had the most optimal readability, lexical diversity, and cohesion scores but used more low frequency words.'}",
-effect of delayed refrigeration on the microbial carcass contamination of wild boars (sus scrofa),B. Cenci-Goga,"Simple Summary The bacteria that contaminate meat after the death of the animal can come from multiple sources, such as the animal itself, the external environment and the operator who handles it. The prompt refrigeration of hunted game is generally considered an important step to prevent spoilage and meat alterations, although many logistic hindrances, such as animal weight and distance from the hunting area to the refrigerators, limit the meticulous adoption of the best procedures. We show that the bacterial population of wild boar carcasses is not correlated to the mere time from shot to refrigeration but is correlated to the refrigeration time from chilling to analysis. The results of our study revealed a correlation between the time from shot to analysis and from refrigeration to analysis but a lack of correlation between the time from shot to refrigeration. Abstract The immediate refrigeration of meat after slaughter is a key issue for the proper storage and aging of meat. The industry standard cold chain relies on low temperatures and ventilation to lower the internal carcass temperature to 0–4 °C within the first 48 h, i.e., within four times the so-called semi-cooling time. On the other hand, for games, once bled and eviscerated, the carcass must be sent to a point where it can be sectioned or kept on air for maturation at refrigeration temperature. The precautions to observe are few and simple but essential: protect the meat and start the cooling process quickly. After preparing the animal (bleeding and evisceration), it may be necessary to face a period of transport that is sometimes long and not very easy; while small animals can be easily transported in a backpack, larger ones must necessarily be carried by several people or sometimes dragged to the vehicle capable of transporting them. It is obvious that a wild boar opened from the jaws to the pelvis and dragged for hundreds of meters will tend to be contaminated, although these contaminations are to be considered secondary for the preservation of the meat, compared to contamination by the intestinal contents. In an attempt to investigate the effect of delayed refrigeration on wild boar carcass contamination, the aim of this work was to determine a correlation between several hunting and logistic parameters (age, sex, animal weight, shooting distance, number of shots, weather and temperature and time from shot to refrigeration and to analysis) and bacterial contamination of the carcass. The correlation coefficient, r, was found to be 0.038 for the eviscerated body weight (p < 0.05), 0.091 for the external temperature on the day of hunting (p < 0.05), 0.027 for the time from shot to refrigeration (p = 0.081), 0.038 for the time from refrigeration to analysis (p < 0.05) and 0.043 for the time from shot to analysis (p < 0.05). These results stand for a negative correlation between the bacterial population and eviscerated carcass weight and between the bacterial population and external temperature and for a positive correlation between the time from shot to analysis and from refrigeration to analysis. No association was demonstrated between the bacterial population and the time from shot to refrigeration.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the bacterial population of wild boar carcasses is not correlation to the mere time from shot to refrigeration but is correlated to the refrigeration time from chilling to analysis, which stands for a negative correlation between theacterial population and eviscerated carcass weight and between the bacteria population and external temperature.'}",https://www.mdpi.com/2076-2615/11/5/1434/pdf?version=1621346402
-eliciting knowledge from language models for event extraction,Jiaju Lin,"Eliciting knowledge contained in language models via prompt-based learning has shown great potential in many natural language processing tasks, such as text classification and generation. Whereas, the applications for more complex tasks such as event extraction are less studied, since the design of prompt is not straightforward due to the complicated types and arguments. In this paper, we explore to elicit the knowledge from pre-trained language models for event trigger detection and argument extraction. Specifically, we present various joint trigger/argument prompt methods, which can elicit more complementary knowledge by modeling the interactions between different triggers or arguments. The experimental results on the benchmark dataset, namely ACE2005, show the great advantages of our proposed approach. In particular, our approach is superior to the recent advanced methods in the few-shot scenario where only a few samples are used for training.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents various joint trigger/argument prompt methods, which can elicit more complementary knowledge by modeling the interactions between different triggers or arguments.'}",
-dynamic sparsification for quadratic assignment problems,Maximilian John,,"{'model': 'tldr@v2.0.0', 'text': 'An iterative algorithm is proposed that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsified linearization of the original problem in every iteration, which results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions inevery iteration.'}",
-hiv murders prompt activist anger.,K. Morris,"A South African teacher is thought to have murdered his wife and her mother then committed suicide after the couple tested positive for HIV. Mpho Motloung was found shot next to a note saying ""HIV positive AIDS"" alongside her dead mother and critically ill father. Her husband was found dead in the next room. The murders occurred the same day that the inquest of AIDS activist Gugu Dlamini was adjourned because police failed to attend or subpoena key witnesses report the Treatment Action Campaign (TAC). Dlaminis murder 18 months ago is thought related to her HIV-positive status. TAC asserts that the governments ""primitive HIV/AIDS messages"" promote fear and anger causing discrimination and violence against people with HIV/AIDS particularly women. ""Everywhere we go we hear: ""HIV/AIDS is a death sentence"" says Zackie Achmat of TAC. ""Government messages fail to tell the millions of people living with HIV that it is possible to live fully and productively for many years"". (full text)","{'model': 'tldr@v2.0.0', 'text': 'A South African teacher is thought to have murdered his wife and her mother then committed suicide after the couple tested positive for HIV, the same day that the inquest of AIDS activist Gugu Dlamini was adjourned.'}",
-dynamic sparsi cation for quadratic assignment problems,Maximilian John,"We present a framework for optimizing sparse quadratic assignment problems. We propose an iterative algorithm that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsi ed linearization of the original problem in every iteration. This procedure results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions in every iteration. This framework was motivated by the task of the French government to design the French keyboard standard, which included solving sparse quadratic assignment problems with over 100 special characters; a size not feasible for many commonly used approaches. Designing a new standard often involves multiple stakeholders having con icting opinions and, hence, no agreement on a single well-de ned objective function to be used for an extensive one-shot optimization. Since the process of designing the standard is highly interactive, it demands rapid prototyping, e.g., quick primal solutions, on-they evaluation of manual changes, and prompt assessments of solution quality. Particularly concerning the latter aspect, our algorithm is able to provide high-quality lower bounds for these problems within only a few minutes.","{'model': 'tldr@v2.0.0', 'text': 'An iterative algorithm is proposed that dynamically generates the quadratic part of the assignment problem and, thus, solves a sparsi ed linearization of the original problem in every iteration, which results in a hierarchy of lower bounds and, in addition, provides heuristic primal solutions inevery iteration.'}",
-development of an expert system for the high intensity neutrino beam facility at j-parc,K. Nakayoshi,"A high intensity neutrino beam produced at J-PARC is utilized by a long-baseline neutrino oscillation experiment. To generate a high intensity neutrino beam, a high intensity proton beam is extracted from the 30 GeV Main Ring synchrotron to the neutrino primary beamline. In the beamline, one mistaken shot can potentially do serious damage to beamline equipment. To avoid such a consequence, many beamline equipment interlocks to stop the beam operation are implemented. Once an interlock is activated, prompt and proper error handling is necessary. We are developing an expert system for prompt and efficient understanding of the status to quickly resume the beam operation. An inference engine is one key component in the expert system. Although a typical inference engine of the expert system is rule-based, we adapt a Machine-Learning (ML) based inference engine in our expert system. We will report the initial evaluation of our ML-based inference engine. INTRODUCTION The T2K (Tokai-to-Kamioka) experiment [1] is a longbaseline neutrino oscillation experiment at J-PARC (Japan Proton Accelerator Research Complex). Figure 1 shows the overview of the T2K experiment. A high intensity neutrino/anti-neutrino beam is produced and propagates 295 km from J-PARC to Super-Kamiokande (SK). In July 2013, muon neutrino to electron neutrino transformation was firmly established [2]. In August 2017, T2K excluded CP-conservation at 95% confidence level using the latest data. In order to keep generating interesting physics, steady operation of the facility is very important. Super‐Ka ioka de J‐PARC Near Dete tors Neutri o Bea 9 k Mt. Nogu hi‐Goro ,9 Mt. Ike o‐Ya a , , elo sea le el Figure 1: Overview of the T2K experiment. Figure 2 shows a layout of the neutrino experimental facility (neutrino facility) at J-PARC. The neutrino facility is composed of two beamlines and a near detector (ND280). The beamline consists of the primary and secondary beamlines. In the primary beamline, the high intensity proton beam is extracted from the Main Ring synchrotron (MR) and guided through super and normal conducting magnets ∗ kazuo.nakayoshi@kek.jp to the target station. In the secondary beamline, the proton beam hits a graphite target and produces pions. These pions decay into muons and muon neutrinos in a decay volume. The high intensity proton beam reached 470 kW in 2017 and ready to the design power of 750kW with a few years. 0 50 100 m Main Ring",,
-empirical evidence on the relationship between audit probability and internal revenue service budget levels,Akinloye Akindayomi,"Despite a large literature on tax avoidance strategies by individuals and the IRS ability to curb them, how the agency’s audit/examination activities respond to tax expenditures within the individual taxation context is not well understood. We hypothesize that IRS audit probability of individual tax returns will positively respond to the magnitude of different tax expenditure drivers if the agency has a shot at curtailing them. We find that while the probability of IRS audit increases as the agency’s budget on enforcement activities increases, tax expenditures do not appear to prompt IRS enforcement activities in many of the tax expenditures drivers.",,
-development of a novel radiation imaging detector system for in vivo gene imaging in small animal studies,A. Weisenberger,"Many studies in molecular biology deal with following the expression and regulation of a gene at different stages of an organism's development or under different physiological conditions. Presently in situ hybridization and immunochemical assays are available to follow the gene expression at a single moment in time for one organism. One must sacrifice the organism to make a measurement, essentially taking a snap shot of the state of expression of the gene of interest. We have made progress on a new type of gene imaging technology which takes advantage of the emission properties of the radioisotope iodine 125 (/sup 125/I) as the probe and utilizes crystal scintillators and a position sensitive photomultiplier tube. Iodine 125 decays via electron capture emitting a 35 keV gamma-ray with the prompt emission of several 27-32 keV K/spl alpha/ and K/spl beta/ shell X-rays. Because of this a coincidence condition can be set to detect the /sup 125/I decays thus reducing background radiation contribution to the image. Mouse imaging studies of iodine uptake by the thyroid and melatonin receptor binding have been done with this detector system using low doses of /sup 125/I.","{'model': 'tldr@v2.0.0', 'text': 'Progress is made on a new type of gene imaging technology which takes advantage of the emission properties of the radioisotope iodine 125 (/sup 125/I) as the probe and utilizes crystal scintillators and a position sensitive photomultiplier tube.'}",
-increased lead absorption and lead poisoning from a retained bullet.,V. Fiorica,"A young man was found to have elevated blood lead and erythrocyte protoporphyrin levels seven years after he had been shot in the foot with a .22 caliber bullet. Although clearly evident in radiographs obtained at the time of the accident, the bullet had never been removed. Over the course of the intervening years the radiographic appearance of the projectile changed dramatically. The patient underwent a surgical debridement procedure with some resolution of his lead intoxication. This case was reviewed in the context of the literature describing the disposition of retained lead projectiles. Lead bullets or fragments lodged within joints are slowly disintegrated and, as absorption accelerates, predispose the patient to the risk of lead poisoning which may occur many years after the original incident. Prompt removal of any lead projectile entering a joint space is strongly recommended.","{'model': 'tldr@v2.0.0', 'text': 'This case was reviewed in the context of the literature describing the disposition of retained lead projectiles and Prompt removal of any lead projectile entering a joint space is strongly recommended.'}",
-[blood transfusion and consultation].,T. Tasaki,"Many hospital staff, including doctors, nurses, pharmacists, etc., are engaged in blood transfusion practice, and various inquiries are referred to the blood transfusion services. In order to provide a prompt and proper reply, transfusionists must have a wealth of knowledge and experience concerning blood transfusion medicine. Q & A relating to blood transfusion can be found on the home page of the Japanese Society of Blood Transfusion, and these are useful staff resources to obtain simple information. However, we sometimes encounter difficult problems in the management of a patient's treatment. Three representative transfusion-related issues are described in this article: (1) blood transfusion to patients with a positive DAT; (2) emergency transfusion, especially in cases where unexpected antibodies are encountered; (3) management of platelet transfusion refractoriness. Minimum standards for the investigation of transfusion-related adverse reactions developed by SHOT (Serious Hazards of Transfusion) are also introduced in this article, and these have a highly practical value. Finally, the importance of education in transfusion medicine is described. The number of doctors in Japan who are engaged exclusively in transfusion medicine is small, but blood transfusions are performed in every hospital, regardless of whether such a specialist is present. We have recently had to deal with a wide range of transplantation-related issues. Therefore, there is a great need for special education in transfusion medicine for doctors in the transplantation and cell therapy age.","{'model': 'tldr@v2.0.0', 'text': 'Three representative transfusion-related issues are described in this article: blood transfusion to patients with a positive DAT; emergency transfusion, especially in cases where unexpected antibodies are encountered; and management of platelet transfusion refractoriness.'}",
-"dennis judd, the lion and the tiger: the rise and fall of the british raj, 1600–1947 . new york, ny: oxford university press, 2004. 234 pp. isbn: 0-19-280358-1 (hbk.); 0-19-280579-7 (pbk.).",R. Wenzlhuemer,"damentally distinct domains of activity, namely religion and politics. His liberal assumptions preclude the possibility that his object of study might prompt investigation of a political modernity which, as in so many other parts of the world in the twentieth century (including Europe and the U.S.), does not conform to such an assumption and which instead comes to be shot through by competing ideologies of religion, culture, race, and national identity. One consequence ofjhis is the inadequate treatment of Congress socialism, which like so many Christian, Muslim, and national 'socialisms' elsewhere in the world, often had little in common with Marxian socialism, orthodox or otherwise. That Hinduist socialism should be widespread in North India in the 1930s and 1940s will not strike many readers as surprising as Gould seems to find it. Certainly, it need not have prompted him to alert his reader as to the supposedly peculiar penchant of Indian socialism for racialist or culturalist perspectives. By the same token, Gould's analysis would have benefited from greater attention being paid to parties commonly regarded as largely free of communalism, such as the Communist Party, which, after all, enjoyed significant strength in Kanpur and which, during precisely the period in question, fought for support among some of the same groups targeted by Congress and Muslim League organizational efforts. Finally, Gould's emphasis on the ubiquity of communalism in the Congress itself threatens to obscure the bases of support enjoyed by a (IP. Congressman like Jawaharlal Nehru, whose popularity as a writer and speaker unquestionably grew during just these years. Gould's volume verges on obscuring from view the anti-colonial struggle in (J.P., of which, after all, the Congress was the principal institutional agent. Nowhere is this omission more striking than in the volume's remarkable neglect of the development in and through the 1942 'Quit India' movement of the communal dynamics he is investigating, a fact which may derive from the relative paucity of the colonial archives from that period. Despite Gould's important departures from many Cambridge School orthodoxies, the one that seems to have left the strongest trace is the old deprecation of the mass-based and anti-colonial character of the Congressled anti-colonial movement. The result is that the reader is left wondering how exactly a divided and religiously pre-occupied movement such as the Congress that Gould describes ever managed to force the British out of India at all. Still, these misgivings notwithstanding, Hindu Nationalism and the Language of Politics in Late Colonial India is an important and long-overdue contribution to twentieth-century South Asian political history and it is to be hoped that Cambridge will soon bring out a paperback edition so that instructors can with a clear conscience require the text for advanced seminars on the subject.",,
-"thoracic, abdominal, and peripheral trauma. evaluation with angiography.",H. Redman,"PROMPT evaluation of the condition of patients who have suffered penetrating or blunt trauma is important in reducing morbidity and mortality. During the past few years, many diagnostic tests have been developed to aid the clinician, and use of the appropriate ones is important in facilitating proper patient management. Angiography has a specific role in some injuries but should be held in reserve in others. Both the type of trauma and the severity of the injury are important in deciding when to use angiography. Penetrating Trauma Most patients who have had a penetrating injury to the thorax or abdomen with a high- or low-velocity missile will need surgical exploration, and generally there is no need for angiography. Occasionally, when there have been multiple injuries, angiography may be used to decide which injuries should be approached first. For example, a patient had been shot through the right forearm and back at","{'model': 'tldr@v2.0.0', 'text': 'Angiography has a specific role in some injuries but should be held in reserve in others, and both the type of trauma and the severity of the injury are important in deciding when to use angiography.'}",
-many-shot from low-shot: learning to annotate using mixed supervision for object detection,C. Biffi,,"{'model': 'tldr@v2.0.0', 'text': 'This work introduces, for the first time, an online annotation module (OAM) that learns to generate a many-shot set of \\emph{reliable} annotations from a larger volume of weakly labelled images, resulting in a fully end-to-end strategy that only requires a low- shot set of fully annotated images.'}",https://arxiv.org/pdf/2008.09694
-ccprompt: counterfactual contrastive prompt-tuning for many-class classification,Y. Li,"With the success of the prompt-tuning paradigm in Natural Language Processing (NLP), various prompt templates have been proposed to further stimulate specific knowledge for serving downstream tasks, e.g., machine translation, text generation, relation extraction, and so on. Existing prompt templates are mainly shared among all training samples with the information of task description. However, training samples are quite diverse. The sharing task description is unable to stimulate the unique task-related information in each training sample, especially for tasks with the finite-label space. To exploit the unique task-related information, we imitate the human decision process which aims to find the contrastive attributes between the objective factual and their potential counterfactuals. Thus, we propose the \textbf{C}ounterfactual \textbf{C}ontrastive \textbf{Prompt}-Tuning (CCPrompt) approach for many-class classification, e.g., relation classification, topic classification, and entity typing. Compared with simple classification tasks, these tasks have more complex finite-label spaces and are more rigorous for prompts. First of all, we prune the finite label space to construct fact-counterfactual pairs. Then, we exploit the contrastive attributes by projecting training instances onto every fact-counterfactual pair. We further set up global prototypes corresponding with all contrastive attributes for selecting valid contrastive attributes as additional tokens in the prompt template. Finally, a simple Siamese representation learning is employed to enhance the robustness of the model. We conduct experiments on relation classification, topic classification, and entity typing tasks in both fully supervised setting and few-shot setting. The results indicate that our model outperforms former baselines.","{'model': 'tldr@v2.0.0', 'text': 'The Counterfactual-Contrastive-Prompt-Tuning approach for many-class classification, e.g., relation classification, topic classification, and entity typing, and results indicate that the model outperforms former baselines.'}",https://arxiv.org/pdf/2211.05987
-zero-shot versus many-shot: unsupervised texture anomaly detection,Toshimichi Aota,"Research on unsupervised anomaly detection (AD) has recently progressed, significantly increasing detection accuracy. This paper focuses on texture images and considers how few normal samples are needed for accurate AD. We first highlight the critical nature of the problem that previous studies have overlooked: accurate detection gets harder for anisotropic textures when image orientations are not aligned between inputs and normal samples. We then propose a zero-shot method, which detects anomalies without using a normal sample. The method is free from the issue of unaligned orientation between input and normal images. It assumes the input texture to be homogeneous, detecting image regions that break the homogeneity as anomalies. We present a quantitative criterion to judge whether this assumption holds for an input texture. Experimental results show the broad applicability of the proposed zero-shot method and its good performance comparable to or even higher than the state-of-the-art methods using hundreds of normal samples. The code and data are available from https://drive.google.com/drive/folders/10OyPzvI3H6llCZBxKxFlKWt1Pw1tkMK1.",,
-delving deep into many-to-many attention for few-shot video object segmentation,Haoxin Chen,"This paper tackles the task of Few-Shot Video Object Segmentation (FSVOS), i.e., segmenting objects in the query videos with certain class specified in a few labeled support images. The key is to model the relationship between the query videos and the support images for propagating the object information. This is a many-to-many problem and often relies on full-rank attention, which is computationally intensive. In this paper, we propose a novel Domain Agent Network (DAN), breaking down the full-rank attention into two smaller ones. We consider one single frame of the query video as the domain agent, bridging between the support images and the query video. Our DAN allows a linear space and time complexity as opposed to the original quadratic form with no loss of performance. In addition, we introduce a learning strategy by combining meta-learning with online learning to further improve the segmentation accuracy. We build a FSVOS benchmark on the Youtube-VIS dataset and conduct experiments to demonstrate that our method outperforms baselines on both computational cost and accuracy, achieving the state-of-the-art performance. Code is available at https://github.com/scutpaul/DANet.","{'model': 'tldr@v2.0.0', 'text': 'A novel Domain Agent Network (DAN) is proposed, breaking down the full-rank attention into two smaller ones, allowing a linear space and time complexity as opposed to the original quadratic form with no loss of performance.'}",
-few-shot and many-shot fusion learning in mobile visual food recognition,Heng Zhao,"Mobile visual food recognition is emerging as an important application in food logging and dietary monitoring in recent years. Existing food recognition methods use conventional many-shot learning to train a large backbone network, which refers to the use of sufficient number of training data to train the network. However, these methods firstly do not consider the cases where certain food categories have limited training data. Therefore, they cannot use the conventional training using many-shot learning. Further, existing solutions focus on improving the food recognition performance by implementing state-of-the-art large full networks, and do not pay much attention to reduce the size and computational cost of the network. As a result, they are not amenable for deployment on mobile devices. In this paper, we address these issues by proposing a new few-shot and many-shot fusion learning for mobile visual food recognition, it has a compact framework and is able to learn from existing dataset categories, and also new food categories given only a few sample images. We construct a new Indian food dataset called NTU-IndianFood107 in order to evaluate the performance of the proposed method. The dataset has two parts: (i) a Base Dataset of 83 classes of Indian food images with over 600 images per class to perform many-shot learning, and (ii) a Food Diary of 24 classes captured in restaurants with limited number to simulate the few-shot learning on new food categories. The proposed fusion method achieves a Top-1 classification accuracy of 72.0% on the new dataset.","{'model': 'tldr@v2.0.0', 'text': 'A new few-shot and many-shot fusion learning for mobile visual food recognition that has a compact framework and is able to learn from existing dataset categories, and also new food categories given only a few sample images is proposed.'}",
-automatic machine translation evaluation in many languages via zero-shot paraphrasing,Brian Thompson,"We propose the use of a sequence-to-sequence paraphraser for automatic machine translation evaluation. The paraphraser takes a human reference as input and then force-decodes and scores an MT system output. We propose training the aforementioned paraphraser as a multilingual NMT system, treating paraphrasing as a zero-shot ""language pair"" (e.g., Russian to Russian). We denote our paraphraser ""unbiased"" because the mode of our model's output probability is centered around a copy of the input sequence, which in our case represent the best case scenario where the MT system output matches a human reference. Our method is simple and intuitive, and our single model (trained in 39 languages) outperforms or statistically ties with all prior metrics on the WMT19 segment-level shared metrics task in all languages, excluding Gujarati where the model had no training data. We also explore using our model conditioned on the source instead of the reference, and find that it outperforms every quality estimation as a metric system from the WMT19 shared task on quality estimation by a statistically significant margin in every language pair.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the use of a sequence-to-sequence paraphraser for automatic machine translation evaluation, and finds that the model conditioned on the source instead of the reference outperforms every quality estimation as a metric system from the WMT19 shared task on quality estimation by a statistically significant margin in every language pair.'}",https://www.aclweb.org/anthology/2020.emnlp-main.8.pdf
-many-class few-shot learning on multi-granularity class hierarchy,Lu Liu,"We study many-class few-shot (MCFS) problem in both supervised learning and meta-learning settings. Compared to the well-studied many-class many-shot and few-class few-shot problems, the MCFS problem commonly occurs in practical applications but has been rarely studied in previous literature. It brings new challenges of distinguishing between many classes given only a few training samples per class. In this article, we leverage the class hierarchy as a prior knowledge to train a coarse-to-fine classifier that can produce accurate predictions for MCFS problem in both settings. The propose model, “memory-augmented hierarchical-classification network (MahiNet)”, performs coarse-to-fine classification where each coarse class can cover multiple fine classes. Since it is challenging to directly distinguish a variety of fine classes given few-shot data per class, MahiNet starts from learning a classifier over coarse-classes with more training data whose labels are much cheaper to obtain. The coarse classifier reduces the searching range over the fine classes and thus alleviates the challenges from “many classes”. On architecture, MahiNet first deploys a convolutional neural network (CNN) to extract features. It then integrates a memory-augmented attention module and a multi-layer perceptron (MLP) together to produce the probabilities over coarse and fine classes. While the MLP extends the linear classifier, the attention module extends the KNN classifier, both together targeting the “few-shot” problem. We design several training strategies of MahiNet for supervised learning and meta-learning. In addition, we propose two novel benchmark datasets “mcfsImageNet” (as a subset of ImageNet) and “mcfsOmniglot” (re-splitted Omniglot) specially designed for MCFS problem. In experiments, we show that MahiNet outperforms several state-of-the-art models (e.g., prototypical networks and relation networks) on MCFS problems in both supervised learning and meta-learning.","{'model': 'tldr@v2.0.0', 'text': 'The proposed model, “memory-augmented hierarchical-classification network (MahiNet)”, performs coarse-to-fine classification where each coarse class can cover multiple fine classes and outperforms several state-of-the-art models on MCFS problems in both supervised learning and meta-learning.'}",https://opus.lib.uts.edu.au/bitstream/10453/147743/3/Many-Class%20Few-Shot%20Learning.pdf
-generalized many-way few-shot video classification,Yongqin Xian,,"{'model': 'tldr@v2.0.0', 'text': 'A simple 3D CNN baseline is developed, surpassing existing methods by a large margin and proposed to leverage weakly-labeled videos from a large dataset using tag retrieval followed by selecting the best clips with visual similarities, yielding further improvement.'}",https://pure.mpg.de/pubman/item/item_3267299_1/component/file_3267300/arXiv%3A2007.04755.pdf
-tailor-made compositional gradient copolymer by a many-shot raft emulsion polymerization method,Yunlong Guo,"A many-shot RAFT emulsion polymerization method is proposed to synthesize gradient copolymers with high molecular weight and a tailor-made compositional gradient. In this method each shot consisting of comonomers with pre-set different fractions and targeting the molecular weight of 10 000 g mol−1 was added in a stepwise manner during the reaction. High conversions over 95% were achieved in 35 min after each shot. The compositional variation along the polymer chain was then directly determined by the comonomer fractions added at each shot. Styrene/n-butyl acrylate gradient copolymers (including linear and V-shaped gradient) with molecular weights as high as 90 000 g mol−1 were prepared by this method. The composition profiles along the polymer chains agreed well with the theoretical predictions, and the composition distribution among the polymer chains was narrow. The gradient copolymers showed different thermal and phase separation properties from their block counterparts, as expected. These results demonstrated the successful tailor-making of the gradient copolymers. The current strategy will act as a facile method to prepare tailor-made gradient copolymers with high molecular weights and within a short time.",,
-llm-planner: few-shot grounded planning for embodied agents with large language models,Chan Hee Song,"This study focuses on using large language models (LLMs) as a planner for embodied agents that can follow natural language instructions to complete complex tasks in a visually-perceived environment. The high data cost and poor sample efficiency of existing methods hinders the development of versatile agents that are capable of many tasks and can learn new tasks quickly. In this work, we propose a novel method, LLM-Planner, that harnesses the power of large language models to do few-shot planning for embodied agents. We further propose a simple but effective way to enhance LLMs with physical grounding to generate and update plans that are grounded in the current environment. Experiments on the ALFRED dataset show that our method can achieve very competitive few-shot performance: Despite using less than 0.5% of paired training data, LLM-Planner achieves competitive performance with recent baselines that are trained using the full training data. Existing methods can barely complete any task successfully under the same few-shot setting. Our work opens the door for developing versatile and sample-efficient embodied agents that can quickly learn many tasks. Website: https://dki-lab.github.io/LLM-Planner","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method, LLM-Planner, that harnesses the power of large language models to do few-shot planning for embodied agents and proposes a simple but effective way to enhance LLMs with physical grounding to generate and update plans that are grounded in the current environment.'}",https://arxiv.org/pdf/2212.04088
-putting nerf on a diet: semantically consistent few-shot view synthesis,Ajay Jain,"We present DietNeRF, a 3D neural scene representation estimated from a few images. Neural Radiance Fields (NeRF) learn a continuous volumetric representation of a scene through multi-view consistency, and can be rendered from novel viewpoints by ray casting. While NeRF has an impressive ability to reconstruct geometry and fine details given many images, up to 100 for challenging 360° scenes, it often finds a degenerate solution to its image reconstruction objective when only a few input views are available. To improve few-shot quality, we propose DietNeRF. We introduce an auxiliary semantic consistency loss that encourages realistic renderings at novel poses. DietNeRF is trained on individual scenes to (1) correctly render given input views from the same pose, and (2) match high-level semantic attributes across different, random poses. Our semantic loss allows us to supervise DietNeRF from arbitrary poses. We extract these semantics using a pre-trained visual encoder such as CLIP, a Vision Transformer trained on hundreds of millions of diverse single-view, 2D photographs mined from the web with natural language supervision. In experiments, DietNeRF improves the perceptual quality of few-shot view synthesis when learned from scratch, can render novel views with as few as one observed image when pre-trained on a multi-view dataset, and produces plausible completions of completely unobserved regions. Our project website is available at https://www.ajayj.com/dietnerf.","{'model': 'tldr@v2.0.0', 'text': 'DietNeRF improves the perceptual quality of few-shot view synthesis when learned from scratch, can render novel views with as few as one observed image when pre-trained on a multi-view dataset, and produces plausible completions of completely unobserved regions.'}",https://arxiv.org/pdf/2104.00677
-what does a platypus look like? generating customized prompts for zero-shot image classification,Sarah Pratt,"Open-vocabulary models are a promising new paradigm for image classification. Unlike traditional classification models, open-vocabulary models classify among any arbitrary set of categories specified with natural language during inference. This natural language, called""prompts"", typically consists of a set of hand-written templates (e.g.,""a photo of a {}"") which are completed with each of the category names. This work introduces a simple method to generate higher accuracy prompts, without relying on any explicit knowledge of the task domain and with far fewer hand-constructed sentences. To achieve this, we combine open-vocabulary models with large language models (LLMs) to create Customized Prompts via Language models (CuPL, pronounced""couple""). In particular, we leverage the knowledge contained in LLMs in order to generate many descriptive sentences that contain important discriminating characteristics of the image categories. This allows the model to place a greater importance on these regions in the image when making predictions. We find that this straightforward and general approach improves accuracy on a range of zero-shot image classification benchmarks, including over one percentage point gain on ImageNet. Finally, this simple baseline requires no additional training and remains completely zero-shot. Code available at https://github.com/sarahpratt/CuPL.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a simple method to generate higher accuracy prompts, without relying on any explicit knowledge of the task domain and with far fewer hand-constructed sentences, and finds that this straightforward and general approach improves accuracy on a range of zero-shot image classification benchmarks, including over one percentage point gain on ImageNet.'}",https://arxiv.org/pdf/2209.03320
-a uto g raphex : zero-shot biomedical definition generation with automatic prompting,,"Describing terminologies with definition texts 001 is an important step towards understanding 002 the scientific literature, especially for domains 003 with limited labeled terminologies. Previous 004 works have sought to design supervised neural 005 text generation models to solve the biomedi- 006 cal terminology generation task, but most of 007 them failed to define never-before-seen termi- 008 nologies in newly emerging research fields. 009 Here, we tackle this challenge by introducing 010 a zero-shot definition generation model based 011 on prompting , a recent approach for eliciting 012 knowledge from pre-trained language models, 013 with automatically generated prompts. Fur- 014 thermore, we enhanced the biomedical termi- 015 nology dataset by adding descriptive texts to 016 each biomedical subdiscipline, thus enabling 017 zero-shot learning scenarios. Our model out- 018 performed existing supervised baseline and the 019 baseline pre-trained language model that em- 020 ploys manually crafted prompts by up to 52 and 021 6 BLEU score, respectively. 022","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a zero-shot definition generation model based on prompting, a recent approach for eliciting knowledge from pre-trained language models, thus enabling zero- shot learning scenarios in biomedical termi- 015 nology dataset.'}",
-better zero-shot reasoning with role-play prompting,Aobo Kong,"Modern large language models (LLMs), such as ChatGPT, exhibit a remarkable capacity for role-playing, enabling them to embody not only human characters but also non-human entities like a Linux terminal. This versatility allows them to simulate complex human-like interactions and behaviors within various contexts, as well as to emulate specific objects or systems. While these capabilities have enhanced user engagement and introduced novel modes of interaction, the influence of role-playing on LLMs' reasoning abilities remains underexplored. In this study, we introduce a strategically designed role-play prompting methodology and assess its performance under the zero-shot setting across twelve diverse reasoning benchmarks, encompassing arithmetic, commonsense reasoning, symbolic reasoning, and more. Leveraging models such as ChatGPT and Llama 2, our empirical results illustrate that role-play prompting consistently surpasses the standard zero-shot approach across most datasets. Notably, accuracy on AQuA rises from 53.5% to 63.8%, and on Last Letter from 23.8% to 84.2%. Beyond enhancing contextual understanding, we posit that role-play prompting serves as an implicit Chain-of-Thought (CoT) trigger, thereby improving the quality of reasoning. By comparing our approach with the Zero-Shot-CoT technique, which prompts the model to""think step by step"", we further demonstrate that role-play prompting can generate a more effective CoT. This highlights its potential to augment the reasoning capabilities of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A strategically designed role-play prompting methodology is introduced and its performance under the zero-shot setting across twelve diverse reasoning benchmarks, encompassing arithmetic, commonsense reasoning, symbolic reasoning, and more, demonstrating its potential to augment the reasoning capabilities of LLMs.'}",https://arxiv.org/pdf/2308.07702
-zero-shot slot filling with slot-prefix prompting and attention relationship descriptor,Qiaoyang Luo,"This paper addresses zero-shot slot filling, which tries to build a system that can generalize to unseen slot types without any training data. The key to zero-shot slot-filling is to match the tokens from the utterance with the semantic definition of the slot without training data in the target domain. This paper tackles this problem by devising a scheme to fully leverage pre-trained language models (PLMs). To this end, we propose a new prompting scheme that utilizes both learnable tokens and slot names to guide the model to focus on the relevant text spans for a given slot. Furthermore, we use attention values between tokens to form a feature descriptor for each token, which is motivated by the fact that the attention value in a PLM naturally characterizes various relationships, e.g., syntactic or semantic, between tokens. By further consolidating those features with an additional transformer-based aggregation module, we create a simple-but-effective zero-shot slot filling system that can achieve significantly better performance than the previous methods, as demonstrated by our experimental studies.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new prompting scheme that utilizes both learnable tokens and slot names to guide the model to focus on the relevant text spans for a given slot, and uses attention values between tokens to form a feature descriptor for each token.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26566/26338
-relational representation learning for zero-shot relation extraction with instance prompting and prototype rectification,Bin Duan,"Zero-shot relation extraction aims to extract novel relations that are not observed beforehand. However, existing representation methods are not pre-trained for relational representations and embeddings contain much linguistic information, the distances between them are not consistent with relational semantic similarity. In this paper, we propose a novel method based on Instance Prompting and Prototype Rectification (IPPR) to conduct relational representation learning for zeroshot relation extraction. Instance prompting is designed to reduce the gap between pre-training and fine-tuning, and guide the pre-trained model to generate relation-oriented instance representations. Prototype rectification aims to push the prototype embeddings away from each other and makes the instance embeddings closer to its corresponding prototype embeddings for dynamically rectifying the prototype embeddings. Experimental results on two public datasets demonstrate that our proposed method achieves new state-of-the-arts performance1.","{'model': 'tldr@v2.0.0', 'text': 'A novel method based on Instance Prompting and Prototype Rectification to conduct relational representation learning for zeroshot relation extraction and achieves new state-of-the-arts performance.'}",
-distilling hypernymy relations from language models: on the effectiveness of zero-shot taxonomy induction,Devansh Jain,"In this paper, we analyze zero-shot taxonomy learning methods which are based on distilling knowledge from language models via prompting and sentence scoring. We show that, despite their simplicity, these methods outperform some supervised strategies and are competitive with the current state-of-the-art under adequate conditions. We also show that statistical and linguistic properties of prompts dictate downstream performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes zero-shot taxonomy learning methods which are based on distilling knowledge from language models via prompting and sentence scoring and shows that these methods outperform some supervised strategies and are competitive with the current state-of-the-art under adequate conditions.'}",https://aclanthology.org/2022.starsem-1.13.pdf
-prompting scientific names for zero-shot species recognition,Shubham Parashar,"Trained on web-scale image-text pairs, Vision-Language Models (VLMs) such as CLIP can recognize images of common objects in a zero-shot fashion. However, it is underexplored how to use CLIP for zero-shot recognition of highly specialized concepts, e.g., species of birds, plants, and animals, for which their scientific names are written in Latin or Greek. Indeed, CLIP performs poorly for zero-shot species recognition with prompts that use scientific names, e.g.,""a photo of Lepus Timidus""(which is a scientific name in Latin). Because these names are usually not included in CLIP's training set. To improve performance, prior works propose to use large-language models (LLMs) to generate descriptions (e.g., of species color and shape) and additionally use them in prompts. We find that they bring only marginal gains. Differently, we are motivated to translate scientific names (e.g., Lepus Timidus) to common English names (e.g., mountain hare) and use such in the prompts. We find that common names are more likely to be included in CLIP's training set, and prompting them achieves 2$\sim$5 times higher accuracy on benchmarking datasets of fine-grained species recognition.","{'model': 'tldr@v2.0.0', 'text': ""It is found that common names are more likely to be included in CLIP's training set, and prompting them achieves 2$\\sim$5 times higher accuracy on benchmarking datasets of fine-grained species recognition.""}",
-a universal discriminator for zero-shot generalization,Haike Xu,"Generative modeling has been the dominant approach for large-scale pretraining and zero-shot generalization. In this work, we challenge this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks. Technically, we train a single discriminator to predict whether a text sample comes from the true data distribution, similar to GANs. Since many NLP tasks can be formulated as selecting from a few options, we use this discriminator to predict the concatenation of input and which option has the highest probability of coming from the true data distribution. This simple formulation achieves state-of-the-art zero-shot results on the T0 benchmark, outperforming T0 by 16.0%, 7.8%, and 11.5% respectively on different scales. In the finetuning setting, our approach also achieves new state-of-the-art results on a wide range of NLP tasks, with only 1/4 parameters of previous methods. Meanwhile, our approach requires minimal prompting efforts, which largely improves robustness and is essential for real-world applications. Furthermore, we also jointly train a generalized UD in combination with generative tasks, which maintains its advantage on discriminative tasks and simultaneously works on generative tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work challenges this convention by showing that discriminative approaches perform substantially better than generative ones on a large number of NLP tasks, and jointly train a generalized UD in combination with generative tasks, which maintains its advantage on discrim inative tasks and simultaneously works onGenerative tasks.'}",http://arxiv.org/pdf/2211.08099
-zero-shot next-item recommendation using large pretrained language models,Lei Wang,"Large language models (LLMs) have achieved impressive zero-shot performance in various natural language processing (NLP) tasks, demonstrating their capabilities for inference without training examples. Despite their success, no research has yet explored the potential of LLMs to perform next-item recommendations in the zero-shot setting. We have identified two major challenges that must be addressed to enable LLMs to act effectively as recommenders. First, the recommendation space can be extremely large for LLMs, and LLMs do not know about the target user's past interacted items and preferences. To address this gap, we propose a prompting strategy called Zero-Shot Next-Item Recommendation (NIR) prompting that directs LLMs to make next-item recommendations. Specifically, the NIR-based strategy involves using an external module to generate candidate items based on user-filtering or item-filtering. Our strategy incorporates a 3-step prompting that guides GPT-3 to carry subtasks that capture the user's preferences, select representative previously watched movies, and recommend a ranked list of 10 movies. We evaluate the proposed approach using GPT-3 on MovieLens 100K dataset and show that it achieves strong zero-shot performance, even outperforming some strong sequential recommendation models trained on the entire training dataset. These promising results highlight the ample research opportunities to use LLMs as recommenders. The code can be found at https://github.com/AGI-Edgerunners/LLM-Next-Item-Rec.","{'model': 'tldr@v2.0.0', 'text': 'A prompting strategy called Zero-Shot Next-Item Recommendation (NIR) prompting that directs LLMs to make next-item recommendations that achieves strong zero-shot performance, even outperforming some strong sequential recommendation models trained on the entire training dataset.'}",http://arxiv.org/pdf/2304.03153
-the wall street neophyte: a zero-shot analysis of chatgpt over multimodal stock movement prediction challenges,Qianqian Xie,"Recently, large language models (LLMs) like ChatGPT have demonstrated remarkable performance across a variety of natural language processing tasks. However, their effectiveness in the financial domain, specifically in predicting stock market movements, remains to be explored. In this paper, we conduct an extensive zero-shot analysis of ChatGPT's capabilities in multimodal stock movement prediction, on three tweets and historical stock price datasets. Our findings indicate that ChatGPT is a""Wall Street Neophyte""with limited success in predicting stock movements, as it underperforms not only state-of-the-art methods but also traditional methods like linear regression using price features. Despite the potential of Chain-of-Thought prompting strategies and the inclusion of tweets, ChatGPT's performance remains subpar. Furthermore, we observe limitations in its explainability and stability, suggesting the need for more specialized training or fine-tuning. This research provides insights into ChatGPT's capabilities and serves as a foundation for future work aimed at improving financial market analysis and prediction by leveraging social media sentiment and historical stock data.","{'model': 'tldr@v2.0.0', 'text': 'It is indicated that ChatGPT is a ""Wall Street Neophyte"" with limited success in predicting stock movements, as it underperforms not only state-of-the-art methods but also traditional methods like linear regression using price features.'}",http://arxiv.org/pdf/2304.05351
-selfcheck: using llms to zero-shot check their own step-by-step reasoning,Ning Miao,"The recent progress in large language models (LLMs), especially the invention of chain-of-thought prompting, has made it possible to automatically answer questions by stepwise reasoning. However, when faced with more complicated problems that require non-linear thinking, even the strongest LLMs make mistakes. To address this, we explore whether LLMs are able to recognize errors in their own step-by-step reasoning, without resorting to external resources. To this end, we propose SelfCheck, a general-purpose zero-shot verification schema for recognizing such errors. We then use the results of these checks to improve question-answering performance by conducting weighted voting on multiple solutions to the question. We test SelfCheck on three datasets (GSM8K, MathQA, and MATH) and find that it successfully recognizes errors and, in turn, increases final answer accuracies.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes SelfCheck, a general-purpose zero-shot verification schema for recognizing errors in large language models and uses the results of these checks to improve question-answering performance by conducting weighted voting on multiple solutions to the question.'}",https://arxiv.org/pdf/2308.00436
-c3: zero-shot text-to-sql with chatgpt,Xuemei Dong,"This paper proposes a ChatGPT-based zero-shot Text-to-SQL method, dubbed C3, which achieves 82.3\% in terms of execution accuracy on the holdout test set of Spider and becomes the state-of-the-art zero-shot Text-to-SQL method on the Spider Challenge. C3 consists of three key components: Clear Prompting (CP), Calibration with Hints (CH), and Consistent Output (CO), which are corresponding to the model input, model bias and model output respectively. It provides a systematic treatment for zero-shot Text-to-SQL. Extensive experiments have been conducted to verify the effectiveness and efficiency of our proposed method.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2307.07306
-tab-cot: zero-shot tabular chain of thought,Ziqi Jin,"The chain-of-though (CoT) prompting methods were successful in various natural language processing (NLP) tasks thanks to their ability to unveil the underlying complex reasoning processes. Such reasoning processes typically exhibit implicitly structured steps. Recent efforts also started investigating methods to encourage more explicitly structured reasoning procedures to be captured. In this work, we propose Tab-CoT, a novel tabular-format CoT prompting method, which allows the complex reasoning process to be explicitly modelled in a highly structured manner. Despite its simplicity, we show that our approach is capable of performing reasoning across multiple dimensions (i.e., both rows and columns). We demonstrate our approach's strong zero-shot and few-shot capabilities through extensive experiments on a range of reasoning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Tab-CoT is proposed, a novel tabular-format CoT prompting method, which allows the complex reasoning process to be explicitly modelled in a highly structured manner and is capable of performing reasoning across multiple dimensions.'}",http://arxiv.org/pdf/2305.17812
-the benefits of label-description training for zero-shot text classification,Lingyu Gao,"Pretrained language models have improved zero-shot text classification by allowing the transfer of semantic knowledge from the training data in order to classify among specific label sets in downstream tasks. We propose a simple way to further improve zero-shot accuracies with minimal effort. We curate small finetuning datasets intended to describe the labels for a task. Unlike typical finetuning data, which has texts annotated with labels, our data simply describes the labels in language, e.g., using a few related terms, dictionary/encyclopedia entries, and short templates. Across a range of topic and sentiment datasets, our method is more accurate than zero-shot by 17-19% absolute. It is also more robust to choices required for zero-shot classification, such as patterns for prompting the model to classify and mappings from labels to tokens in the model's vocabulary. Furthermore, since our data merely describes the labels but does not use input texts, finetuning on it yields a model that performs strongly on multiple text domains for a given label set, even improving over few-shot out-of-domain classification in multiple settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple way to further improve zero-shot accuracies with minimal effort by curating small finetuning datasets intended to describe the labels for a task, which yields a model that performs strongly on multiple text domains for a given label set, even improving over few-shot out-of-domain classification in multiple settings.'}",http://arxiv.org/pdf/2305.02239
-self-icl: zero-shot in-context learning with self-generated demonstrations,Wei-Lin Chen,"Large language models (LLMs) have exhibited striking in-context learning (ICL) ability to adapt to target tasks with a few input-output demonstrations. For better ICL, different methods are proposed to select representative demonstrations from existing training corpora. However, such settings are not aligned with real-world practices, as end-users usually query LMs without access to demonstration pools. In this work, we introduce Self-ICL -- a simple framework which bootstraps LMs' intrinsic capabilities to perform zero-shot ICL. Given a test input, Self-ICL first prompts the model to generate pseudo-inputs. Next, the model predicts pseudo-labels for the pseudo-inputs via zero-shot prompting. Finally, we perform ICL for the test input with the pseudo-input-label pairs as demonstrations. Evaluation on 23 BIG-Bench Hard tasks shows Self-ICL outperforms zero-shot baselines on both average accuracy and head-to-head comparison. Moreover, with zero-shot chain-of-thought, Self-ICL achieves results comparable to using real demonstrations. Additionally, we conduct a range of analyses to validate Self-ICL's effectiveness and provide insights for its behaviors under different settings.","{'model': 'tldr@v2.0.0', 'text': ""Self-ICL is introduced -- a simple framework which bootstraps LMs' intrinsic capabilities to perform zero-shot ICL and achieves results comparable to using real demonstrations with a range of analyses to validate Self- ICL's effectiveness and provide insights for its behaviors under different settings.""}",http://arxiv.org/pdf/2305.15035
-ontotype: ontology-guided zero-shot fine-grained entity typing with weak supervision from pre-trained language models,Tanay Komarlu,"Fine-grained entity typing (FET), which assigns entities in text with context-sensitive, fine-grained semantic types, will play an important role in natural language understanding. A supervised FET method, which typically relies on human-annotated corpora for training, is costly and difficult to scale. Recent studies leverage pre-trained language models (PLMs) to generate rich and context-aware weak supervision for FET. However, a PLM may still generate a mixture of rough and fine-grained types, or tokens unsuitable for typing. In this study, we vision that an ontology provides a semantics-rich, hierarchical structure, which will help select the best results generated by multiple PLM models and head words. Specifically, we propose a novel zero-shot, ontology-guided FET method, OntoType, which follows a type ontological structure, from coarse to fine, ensembles multiple PLM prompting results to generate a set of type candidates, and refines its type resolution, under the local context with a natural language inference model. Our experiments on the Ontonotes, FIGER, and NYT datasets using their associated ontological structures demonstrate that our method outperforms the state-of-the-art zero-shot fine-grained entity typing methods. Our error analysis shows that refinement of the existing ontology structures will further improve fine-grained entity typing.","{'model': 'tldr@v2.0.0', 'text': 'A novel zero-shot, ontology-guided FET method, OntoType, which follows a type ontological structure, from coarse to fine, ensembles multiple PLM prompting results to generate a set of type candidates, and refines its type resolution, under the local context with a natural language inference model.'}",http://arxiv.org/pdf/2305.12307
-jack-ryder at semeval-2023 task 5: zero-shot clickbait spoiling by rephrasing titles as questions,Dirk Wangsadirdja,"In this paper, we describe our approach to the clickbait spoiling task of SemEval 2023.The core idea behind our system is to leverage pre-trained models capable of Question Answering (QA) to extract the spoiler from article texts based on the clickbait title without any task-specific training.Since oftentimes, these titles are not phrased as questions, we automatically rephrase the clickbait titles as questions in order to better suit the pretraining task of the QA-capable models.Also, to fit as much relevant context into the model’s limited input size as possible, we propose to reorder the sentences by their relevance using a semantic similarity model.Finally, we evaluate QA as well as text generation models (via prompting) to extract the spoiler from the text.Based on the validation data, our final model selects each of these components depending on the spoiler type and achieves satisfactory zero-shot results.The ideas described in this paper can easily be applied in fine-tuning settings.","{'model': 'tldr@v2.0.0', 'text': 'The core idea behind the system is to leverage pre-trained models capable of Question Answering to extract the spoiler from article texts based on the clickbait title without any task-specific training.'}",https://aclanthology.org/2023.semeval-1.150.pdf
-a zero-shot language agent for computer control with structured reflection,Tao Li,"Large language models (LLMs) have shown increasing capacity at planning and executing a high-level goal in a live computer environment (e.g. MiniWoB++). To perform a task, recent works often require a model to learn from trace examples of the task via either supervised learning or few/many-shot prompting. Without these trace examples, it remains a challenge how an agent can autonomously learn and improve its control on a computer, which limits the ability of an agent to perform a new task. We approach this problem with a zero-shot agent that requires no given expert traces. Our agent plans for executable actions on a partially observed environment, and iteratively progresses a task by identifying and learning from its mistakes via self-reflection and structured thought management. On the easy tasks of MiniWoB++, we show that our zero-shot agent often outperforms recent SoTAs, with more efficient reasoning. For tasks with more complexity, our reflective agent performs on par with prior best models, even though previous works had the advantages of accessing expert traces or additional screen information.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a zero-shot agent that plans for executable actions on a partially observed environment, and iteratively progresses a task by identifying and learning from its mistakes via self-reflection and structured thought management, which often outperforms recent SoTAs with more efficient reasoning.'}",https://arxiv.org/pdf/2310.08740
-anovl: adapting vision-language models for unified zero-shot anomaly localization,Hanqiu Deng,"Contrastive Language-Image Pre-training (CLIP) models have shown promising performance on zero-shot visual recognition tasks by learning visual representations under natural language supervision. Recent studies attempt the use of CLIP to tackle zero-shot anomaly detection by matching images with normal and abnormal state prompts. However, since CLIP focuses on building correspondence between paired text prompts and global image-level representations, the lack of patch-level vision to text alignment limits its capability on precise visual anomaly localization. In this work, we introduce a training-free adaptation (TFA) framework of CLIP for zero-shot anomaly localization. In the visual encoder, we innovate a training-free value-wise attention mechanism to extract intrinsic local tokens of CLIP for patch-level local description. From the perspective of text supervision, we particularly design a unified domain-aware contrastive state prompting template. On top of the proposed TFA, we further introduce a test-time adaptation (TTA) mechanism to refine anomaly localization results, where a layer of trainable parameters in the adapter is optimized using TFA's pseudo-labels and synthetic noise-corrupted tokens. With both TFA and TTA adaptation, we significantly exploit the potential of CLIP for zero-shot anomaly localization and demonstrate the effectiveness of our proposed methods on various datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a training-free adaptation (TFA) framework of CLIP for zero-shot anomaly localization and designs a unified domain-aware contrastive state prompting template for text supervision, and introduces a test-time adaptation mechanism to refine anomaly localization results.'}",https://arxiv.org/pdf/2308.15939
-a monte carlo language model pipeline for zero-shot sociopolitical event extraction,Erica Cai,"We consider dyadic zero-shot event extraction (EE) to identify actions between pairs of actors. The \emph{zero-shot} setting allows social scientists or other non-computational researchers to extract any customized, user-specified set of events without training, resulting in a \emph{dyadic} event database, allowing insight into sociopolitical relational dynamics among actors and the higher level organizations or countries they represent. Unfortunately, we find that current zero-shot EE methods perform poorly for the task, with issues including word sense ambiguity, modality mismatch, and efficiency. Straightforward application of large language model prompting typically performs even worse. We address these challenges with a new fine-grained, multi-stage generative question-answer method, using a Monte Carlo approach to exploit and overcome the randomness of generative outputs. It performs 90\% fewer queries than a previous approach, with strong performance on the widely-used Automatic Content Extraction dataset. Finally, we extend our method to extract affiliations of actor arguments and demonstrate our method and findings on a dyadic international relations case study.","{'model': 'tldr@v2.0.0', 'text': 'A new fine-grained, multi-stage generative question-answer method, using a Monte Carlo approach to exploit and overcome the randomness of generative outputs to identify actions between pairs of actors.'}",http://arxiv.org/pdf/2305.15051
-lt at semeval-2023 task 1: effective zero-shot visual word sense disambiguation approaches using external knowledge sources,Florian Schneider,"The objective of the SemEval-2023 Task 1: Visual Word Sense Disambiguation (VWSD) is to identify the image illustrating the indented meaning of a target word and some minimal additional context. The omnipresence of textual and visual data in the task strongly suggests the utilization of the recent advances in multi-modal machine learning, i.e., pretrained visiolinguistic models (VLMs). Often referred to as foundation models due to their strong performance on many vision-language downstream tasks, these models further demonstrate powerful zero-shot capabilities. In this work, we utilize various pertained VLMs in a zero-shot fashion for multiple approaches using external knowledge sources to enrich the contextual information. Further, we evaluate our methods on the final test data and extensively analyze the suitability of different knowledge sources, the influence of training data, model sizes, multi-linguality, and different textual prompting strategies. Although we are not among the best-performing systems (rank 20 of 56), our experiments described in this work prove competitive results. Moreover, we aim to contribute meaningful insights and propel multi-modal machine learning tasks like VWSD.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes various pertained VLMs in a zero-shot fashion for multiple approaches using external knowledge sources to enrich the contextual information and aims to contribute meaningful insights and propel multi-modal machine learning tasks like VWSD.'}",https://aclanthology.org/2023.semeval-1.64.pdf
-instruction tuning with lexicons for zero-shot style classification,Ruohao Guo,"Style is used to convey authors' intentions and attitudes. Despite the success of large pre-trained language models on style classification, prior work relies on fine-tuning with labeled examples. Prompting large language models to classify style without fine-tuning is challenging because language styles can be difficult to define. In this study, we investigate the effectiveness of style lexicons as a means for instructing language models how to identify new styles that are unseen during training. Our experiments show that lexicon-based instructions improve transfer zero-shot performance significantly. We will release our code and data.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of style lexicons as a means for instructing language models how to identify new styles that are unseen during training are investigated and show that lexicon-based instructions improve transfer zero-shot performance significantly.'}",http://arxiv.org/pdf/2305.14592
-zero-shot faithfulness evaluation for text summarization with foundation language model,Qi Jia,"Despite tremendous improvements in natural language generation, summarization models still suffer from the unfaithfulness issue. Previous work evaluates faithfulness either using models trained on the other tasks or in-domain synthetic data, or prompting a large model such as ChatGPT. This paper proposes to do zero-shot faithfulness evaluation simply with a moderately-sized foundation language model. We introduce a new metric FFLM, which is a combination of probability changes based on the intuition that prefixing a piece of text that is consistent with the output will increase the probability of predicting the output. Experiments show that FFLM performs competitively with or even outperforms ChatGPT on both inconsistency detection and faithfulness rating with 24x fewer parameters. FFLM also achieves improvements over other strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'A new metric FFLM is introduced, which is a combination of probability changes based on the intuition that prefixing a piece of text that is consistent with the output will increase the probability of predicting the output.'}",
-the art of socratic questioning: zero-shot multimodal reasoning with recursive thinking and self-questioning,Jingyuan Qi,"Chain-of-Thought prompting (CoT) enables large-scale language models to solve complex reasoning problems by decomposing the problem and tackling it step-by-step. However, Chain-of-Thought is a greedy thinking process that requires the language model to come up with a starting point and generate the next step solely based on previous steps. This thinking process is different from how humans approach a complex problem e.g., we proactively raise sub-problems related to the original problem and recursively answer them. In this work, we propose S OCRATIC Q UESTIONING , a divide-and-conquer fashion algorithm that simulates the self-questioning and recursive thinking process. S OCRATIC Q UESTIONING is driven by a S ELF -Q UESTIONING module that employs a large-scale language model to propose sub-problems related to the original problem as intermediate steps and S OCRATIC Q UESTION - ING recursively backtracks and answers the sub-problems until reaches the original problem. We apply our proposed algorithm to the visual question-answering task as a case study and by evaluating it on three public benchmark datasets, we observe a significant performance improvement over all baselines on (almost) all datasets. In addition, the qualitative analysis clearly demonstrates the intermediate thinking steps elicited by S OCRATIC Q UESTIONING are similar to the human’s recursively thinking process of a complex reasoning problem.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative analysis clearly demonstrates the intermediate thinking steps elicited by S OCRATIC Q UESTIONING are similar to the human’s recursively thinking process of a complex reasoning problem.'}",http://arxiv.org/pdf/2305.14999
-regeneration learning of diffusion models with rich prompts for zero-shot image translation,Yupei Lin,"Large-scale text-to-image models have demonstrated amazing ability to synthesize diverse and high-fidelity images. However, these models are often violated by several limitations. Firstly, they require the user to provide precise and contextually relevant descriptions for the desired image modifications. Secondly, current models can impose significant changes to the original image content during the editing process. In this paper, we explore ReGeneration learning in an image-to-image Diffusion model (ReDiffuser), that preserves the content of the original image without human prompting and the requisite editing direction is automatically discovered within the text embedding space. To ensure consistent preservation of the shape during image editing, we propose cross-attention guidance based on regeneration learning. This novel approach allows for enhanced expression of the target domain features while preserving the original shape of the image. In addition, we introduce a cooperative update strategy, which allows for efficient preservation of the original shape of an image, thereby improving the quality and consistency of shape preservation throughout the editing process. Our proposed method leverages an existing pre-trained text-image diffusion model without any additional training. Extensive experiments show that the proposed method outperforms existing work in both real and synthetic image editing.","{'model': 'tldr@v2.0.0', 'text': 'ReGeneration learning in an image-to-image Diffusion model (ReDiffuser), that preserves the content of the original image without human prompting and the requisite editing direction is automatically discovered within the text embedding space, is explored.'}",http://arxiv.org/pdf/2305.04651
-zero-shot refinement of buildings' segmentation models using sam,Ali Mayladan,"Foundation models have excelled in various tasks but are often evaluated on general benchmarks. The adaptation of these models for specific domains, such as remote sensing imagery, remains an underexplored area. In remote sensing, precise building instance segmentation is vital for applications like urban planning. While Convolutional Neural Networks (CNNs) perform well, their generalization can be limited. For this aim, we present a novel approach to adapt foundation models to address existing models' generalization dropback. Among several models, our focus centers on the Segment Anything Model (SAM), a potent foundation model renowned for its prowess in class-agnostic image segmentation capabilities. We start by identifying the limitations of SAM, revealing its suboptimal performance when applied to remote sensing imagery. Moreover, SAM does not offer recognition abilities and thus fails to classify and tag localized objects. To address these limitations, we introduce different prompting strategies, including integrating a pre-trained CNN as a prompt generator. This novel approach augments SAM with recognition abilities, a first of its kind. We evaluated our method on three remote sensing datasets, including the WHU Buildings dataset, the Massachusetts Buildings dataset, and the AICrowd Mapping Challenge. For out-of-distribution performance on the WHU dataset, we achieve a 5.47% increase in IoU and a 4.81% improvement in F1-score. For in-distribution performance on the WHU dataset, we observe a 2.72% and 1.58% increase in True-Positive-IoU and True-Positive-F1 score, respectively. We intend to release our code repository, hoping to inspire further exploration of foundation models for domain-specific tasks within the remote sensing community.","{'model': 'tldr@v2.0.0', 'text': ""A novel approach to adapt foundation models to address existing models' generalization dropback, focusing on the Segment Anything Model (SAM), a potent foundation model renowned for its prowess in class-agnostic image segmentation capabilities.""}",https://arxiv.org/pdf/2310.01845
-mm-react: prompting chatgpt for multimodal reasoning and action,Zhengyuan Yang,"We propose MM-REACT, a system paradigm that integrates ChatGPT with a pool of vision experts to achieve multimodal reasoning and action. In this paper, we define and explore a comprehensive list of advanced vision tasks that are intriguing to solve, but may exceed the capabilities of existing vision and vision-language models. To achieve such advanced visual intelligence, MM-REACT introduces a textual prompt design that can represent text descriptions, textualized spatial coordinates, and aligned file names for dense visual signals such as images and videos. MM-REACT's prompt design allows language models to accept, associate, and process multimodal information, thereby facilitating the synergetic combination of ChatGPT and various vision experts. Zero-shot experiments demonstrate MM-REACT's effectiveness in addressing the specified capabilities of interests and its wide application in different scenarios that require advanced visual understanding. Furthermore, we discuss and compare MM-REACT's system paradigm with an alternative approach that extends language models for multimodal scenarios through joint finetuning. Code, demo, video, and visualization are available at https://multimodal-react.github.io/","{'model': 'tldr@v2.0.0', 'text': 'This paper defines and explores a comprehensive list of advanced vision tasks that are intriguing to solve, but may exceed the capabilities of existing vision and vision-language models and proposes MM-REACT, a system paradigm that integrates ChatGPT with a pool of vision experts to achieve multimodal reasoning and action.'}",http://arxiv.org/pdf/2303.11381
-can an embodied agent find your “cat-shaped mug”? llm-guided exploration for zero-shot object navigation,Vishnu Sashank Dorbala,"—We present LGX (Language-guided Exploration), a novel algorithm for Language-Driven Zero-Shot Object Goal Navigation (L-ZSON), where an embodied agent navigates to an uniquely described target object in a previously unseen environment. Our approach makes use of Large Language Models (LLMs) for this task by leveraging the LLM’s commonsense-reasoning capabilities for making sequential navigational decisions. Simultaneously, we perform generalized target object detection using a pre-trained Vision-Language grounding model. We achieve state-of-the-art zero-shot object navigation results on RoboTHOR with a success rate (SR) improvement of over 27% over the current baseline of the OWL-ViT CLIP on Wheels (OWL CoW). Furthermore, we study the usage of LLMs for robot navigation and present an analysis of various prompting strategies affecting the model output. Finally, we showcase the benefits of our approach via real-world experiments that indicate the superior performance of LGX in detecting and navigating to visually unique objects.","{'model': 'tldr@v2.0.0', 'text': None}",
-the art of prompting: event detection based on type specific prompts,Sijia Wang,"We compare various forms of prompts to represent event types and develop a unified framework to incorporate the event type specific prompts for supervised, few-shot, and zero-shot event detection. The experimental results demonstrate that a well-defined and comprehensive event type prompt can significantly improve event detection performance, especially when the annotated data is scarce (few-shot event detection) or not available (zero-shot event detection). By leveraging the semantics of event types, our unified framework shows up to 22.2% F-score gain over the previous state-of-the-art baselines.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results demonstrate that a well-defined and comprehensive event type prompt can significantly improve event detection performance, especially when the annotated data is scarce (few- shot event detection) or not available (zero-shot event detection).'}",http://arxiv.org/pdf/2204.07241
-is evalita done? on the impact of prompting on the italian nlp evaluation campaign,Valerio Basile,"Prompt-based learning is a recent paradigm in NLP that leverages large pre-trained language models to perform a variety of tasks. With this technique, it is possible to build classifiers that do not need training data (zero-shot). In this paper, we assess the status of prompt-based learning applied to several text classification tasks in the Italian language. The results indicate that the performance gap towards current supervised methods is still relevant. However, the difference in performance between pre-trained models and the characteristic of the prompt-based classifier of operating in a zero-shot fashion open a discussion regarding the next generation of evaluation campaigns for NLP.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the performance gap towards current supervised methods is still relevant, however, the difference in performance between pre-trained models and the characteristic of the prompt-based classifier of operating in a zero-shot fashion open a discussion regarding the next generation of evaluation campaigns for NLP.'}",
-clip also understands text: prompting clip for phrase understanding,An Yan,"Contrastive Language-Image Pretraining (CLIP) efficiently learns visual concepts by pre-training with natural language supervision. CLIP and its visual encoder have been explored on various vision and language tasks and achieve strong zero-shot or transfer learning performance. However, the application of its text encoder solely for text understanding has been less explored. In this paper, we find that the text encoder of CLIP actually demonstrates strong ability for phrase understanding, and can even significantly outperform popular language models such as BERT with a properly designed prompt. Extensive experiments validate the effectiveness of our method across different datasets and domains on entity clustering and entity set expansion tasks.","{'model': 'tldr@v2.0.0', 'text': 'It is found that the text encoder of CLIP actually demonstrates strong ability for phrase understanding, and can even significantly outperform popular language models such as BERT with a properly designed prompt.'}",http://arxiv.org/pdf/2210.05836
-odor descriptor understanding through prompting,Laura Sisson,"Embeddings from contemporary natural language processing (NLP) models are commonly used as numerical representations for words or sentences. However, odor descriptor words, like""leather""or""fruity"", vary significantly between their commonplace usage and their olfactory usage, as a result traditional methods for generating these embeddings do not suffice. In this paper, we present two methods to generate embeddings for odor words that are more closely aligned with their olfactory meanings when compared to off-the-shelf embeddings. These generated embeddings outperform the previous state-of-the-art and contemporary fine-tuning/prompting methods on a pre-existing zero-shot odor-specific NLP benchmark.","{'model': 'tldr@v2.0.0', 'text': 'These generated embeddings outperform the previous state-of-the-art and contemporary fine-tuning/prompting methods on a pre-existing zero-shot odor-specific NLP benchmark.'}",http://arxiv.org/pdf/2205.03719
-arggen: prompting text generation models for document-level event-argument aggregation,Debanjana Kar,"Most of the existing discourse-level Information Extraction tasks have been modeled to be extractive in nature. However, we argue that extracting information from larger bodies of discourse-like documents requires more natural language understanding and reasoning capabilities. In our work, we propose the novel task of document-level event argument aggregation which generates consolidated event-arguments at a document-level with minimal loss of information. More specifically, we focus on generating precise document-level information frames in a multilingual setting using prompt-based methods. In this paper, we show the effectiveness of prompt-based text generation approach to generate document-level argument spans in a low-resource and zero-shot setting. We also release the first of its kind multilingual event argument aggregation dataset that can be leveraged in other related multilingual text generation tasks as well: https://github.com/","{'model': 'tldr@v2.0.0', 'text': 'This paper shows the effectiveness of prompt-based text generation approach to generate document-level argument spans in a low-resource and zero-shot setting and releases the first of its kind multilingual event argument aggregation dataset.'}",
-on the evaluations of chatgpt and emotion-enhanced prompting for mental health analysis,Kailai Yang,"Automated mental health analysis shows great potential for enhancing the efficiency and ac-cessibility of mental health care, whereas the recent dominant methods utilized pre-trained language models (PLMs) as the backbone and incorporated emotional information. The latest large language models (LLMs), such as ChatGPT, exhibit dramatic capabilities on di-verse natural language processing tasks. However, existing studies on ChatGPT’s zero-shot performance for mental health analysis have limitations in inadequate evaluation, utiliza-tion of emotional information, and explainability of methods. In this work, we comprehen-sively evaluate the mental health analysis and emotional reasoning ability of ChatGPT on 11 datasets across 5 tasks, including binary and multi-class mental health condition detection, cause/factor detection of mental health conditions, emotion recognition in conversations, and causal emotion entailment. We empirically analyze the impact of different prompting strategies with emotional cues on ChatGPT’s mental health analysis ability and explainability. Experimental results show that ChatGPT outperforms traditional neural network methods but still has a significant gap with advanced task-specific methods. The qualitative analysis shows its potential in explainability compared with advanced black-box methods but also limitations on robustness and inaccurate reasoning. Prompt engineering with emotional cues is found to be effective in improving its performance on mental health analysis but requires the proper way of emotion infusion.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that ChatGPT outperforms traditional neural network methods but still has a significant gap with advanced task-species methods, and the qualitative analysis shows its potential in explainability compared with advanced black-box methods but also limitations on robustness and inaccurate reasoning.'}",https://arxiv.org/pdf/2304.03347
-reasoning implicit sentiment with chain-of-thought prompting,Hao Fei,"While sentiment analysis systems try to determine the sentiment polarities of given targets based on the key opinion expressions in input texts, in implicit sentiment analysis (ISA) the opinion cues come in an implicit and obscure manner. Thus detecting implicit sentiment requires the common-sense and multi-hop reasoning ability to infer the latent intent of opinion. Inspired by the recent chain-of-thought (CoT) idea, in this work we introduce a Three-hop Reasoning (THOR) CoT framework to mimic the human-like reasoning process for ISA. We design a three-step prompting principle for THOR to step-by-step induce the implicit aspect, opinion, and finally the sentiment polarity. Our THOR+Flan-T5 (11B) pushes the state-of-the-art (SoTA) by over 6% F1 on supervised setup. More strikingly, THOR+GPT3 (175B) boosts the SoTA by over 50% F1 on zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'This work designs a three-step prompting principle for THOR to step-by-step induce the implicit aspect, opinion, and finally the sentiment polarity, and introduces a Three-hop Reasoning (THOR) CoT framework to mimic the human-like reasoning process for ISA.'}",http://arxiv.org/pdf/2305.11255
-pearl: prompting large language models to plan and execute actions over long documents,Simeng Sun,"Strategies such as chain-of-thought prompting improve the performance of large language models (LLMs) on complex reasoning tasks by decomposing input examples into intermediate steps. However, it remains unclear how to apply such methods to reason over long input documents, in which both the decomposition and the output of each intermediate step are non-trivial to obtain. In this work, we propose PEARL, a prompting framework to improve reasoning over long documents, which consists of three stages: action mining, plan formulation, and plan execution. More specifically, given a question about a long document, PEARL decomposes the question into a sequence of actions (e.g., SUMMARIZE, FIND_EVENT, FIND_RELATION) and then executes them over the document to obtain the answer. Each stage of PEARL is implemented via zero-shot or few-shot prompting of LLMs (in our work, GPT-4) with minimal human input. We evaluate PEARL on a challenging subset of the QuALITY dataset, which contains questions that require complex reasoning over long narrative texts. PEARL outperforms zero-shot and chain-of-thought prompting on this dataset, and ablation experiments show that each stage of PEARL is critical to its performance. Overall, PEARL is a first step towards leveraging LLMs to reason over long documents.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes PEARL, a prompting framework to improve reasoning over long documents, which consists of three stages: action mining, plan formulation, and plan execution, which outperforms zero-shot and chain-of-thought prompting on a challenging subset of the QuALITY dataset.'}",http://arxiv.org/pdf/2305.14564
-multimodal procedural planning via dual text-image prompting,Yujie Lu,"Embodied agents have achieved prominent performance in following human instructions to complete tasks. However, the potential of providing instructions informed by texts and images to assist humans in completing tasks remains underexplored. To uncover this capability, we present the multimodal procedural planning (MPP) task, in which models are given a high-level goal and generate plans of paired text-image steps, providing more complementary and informative guidance than unimodal plans. The key challenges of MPP are to ensure the informativeness, temporal coherence,and accuracy of plans across modalities. To tackle this, we propose Text-Image Prompting (TIP), a dual-modality prompting method that jointly leverages zero-shot reasoning ability in large language models (LLMs) and compelling text-to-image generation ability from diffusion-based models. TIP improves the interaction in the dual modalities using Text-to-Image Bridge and Image-to-Text Bridge, allowing LLMs to guide the textual-grounded image plan generation and leveraging the descriptions of image plans to ground the textual plan reversely. To address the lack of relevant datasets, we collect WIKIPLAN and RECIPEPLAN as a testbed for MPP. Our results show compelling human preferences and automatic scores against unimodal and multimodal baselines on WIKIPLAN and RECIPEPLAN in terms of informativeness, temporal coherence, and plan accuracy. Our code and data: https://github.com/YujieLu10/MPP.","{'model': 'tldr@v2.0.0', 'text': 'Text-Image Prompting (TIP) is proposed, a dual-modality prompting method that jointly leverages zero-shot reasoning ability in large language models (LLMs) and compelling text-to-image generation ability from diffusion-based models to tackle the key challenges of MPP.'}",http://arxiv.org/pdf/2305.01795
-federated prompting and chain-of-thought reasoning for improving llms answering,Xiangyang Liu,"We investigate how to enhance answer precision in frequently asked questions posed by distributed users using cloud-based Large Language Models (LLMs). Our study focuses on a typical situations where users ask similar queries that involve identical mathematical reasoning steps and problem-solving procedures. Due to the unsatisfactory accuracy of LLMs' zero-shot prompting with standalone questions, we propose to improve the distributed synonymous questions using Self-Consistency (SC) and Chain-of-Thought (CoT) techniques. Specifically, we first retrieve synonymous questions from a crowd-sourced database and create a federated question pool. We call these federated synonymous questions with the same or different parameters SP-questions or DP-questions, respectively. We refer to our methods as Fed-SP-SC and Fed-DP-CoT, which can generate significantly more accurate answers for all user queries without requiring sophisticated model-tuning. Through extensive experiments, we demonstrate that our proposed methods can significantly enhance question accuracy by fully exploring the synonymous nature of the questions and the consistency of the answers.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes to improve the distributed synonymous questions using Self-Consistency (SC) and Chain-of-Thought (CoT) techniques, which can generate significantly more accurate answers for all user queries without requiring sophisticated model-tuning.'}",http://arxiv.org/pdf/2304.13911
-code prompting: a neural symbolic method for complex reasoning in large language models,Y. Hu,"Large language models (LLMs) have scaled up to unlock a wide range of complex reasoning tasks with the aid of various prompting methods. However, current prompting methods generate natural language intermediate steps to help reasoning, which can cause imperfect task reduction and confusion. To mitigate such limitations, we explore code prompting, a neural symbolic prompting method with both zero-shot and few-shot versions which triggers code as intermediate steps. We conduct experiments on 7 widely-used benchmarks involving symbolic reasoning and arithmetic reasoning. Code prompting generally outperforms chain-of-thought (CoT) prompting. To further understand the performance and limitations of code prompting, we perform extensive ablation studies and error analyses, and identify several exclusive advantages of using symbolic promptings compared to natural language. We also consider the ensemble of code prompting and CoT prompting to combine the strengths of both. Finally, we show through experiments how code annotations and their locations affect code prompting.","{'model': 'tldr@v2.0.0', 'text': 'This work explores code prompting, a neural symbolic prompting method with both zero-shot and few-shot versions which triggers code as intermediate steps, and performs extensive ablation studies and error analyses and identifies several exclusive advantages of using symbolic promptings compared to natural language.'}",https://arxiv.org/pdf/2305.18507
-set-of-mark prompting unleashes extraordinary visual grounding in gpt-4v,Jianwei Yang,"We present Set-of-Mark (SoM), a new visual prompting method, to unleash the visual grounding abilities of large multimodal models (LMMs), such as GPT-4V. As illustrated in Fig. 1 (right), we employ off-the-shelf interactive segmentation models, such as SAM, to partition an image into regions at different levels of granularity, and overlay these regions with a set of marks e.g., alphanumerics, masks, boxes. Using the marked image as input, GPT-4V can answer the questions that require visual grounding. We perform a comprehensive empirical study to validate the effectiveness of SoM on a wide range of fine-grained vision and multimodal tasks. For example, our experiments show that GPT-4V with SoM outperforms the state-of-the-art fully-finetuned referring segmentation model on RefCOCOg in a zero-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that GPT-4V with SoM outperforms the state-of-the-art fully-finetuned referring segmentation model on RefCOCOg in a zero-shot setting and the effectiveness of SoM on a wide range of fine-grained vision and multimodal tasks is validated.'}",
-legal syllogism prompting: teaching large language models for legal judgment prediction,Cong Jiang,"Legal syllogism is a form of deductive reasoning commonly used by legal professionals to analyze cases. In this paper, we propose legal syllogism prompting (LoT), a simple prompting method to teach large language models (LLMs) for legal judgment prediction. LoT teaches only that in the legal syllogism the major premise is law, the minor premise is the fact, and the conclusion is judgment. Then the models can produce a syllogism reasoning of the case and give the judgment without any learning, fine-tuning, or examples. On CAIL2018, a Chinese criminal case dataset, we performed zero-shot judgment prediction experiments with GPT-3 models. Our results show that LLMs with LoT achieve better performance than the baseline and chain of thought prompting, the state-of-art prompting method on diverse reasoning tasks. LoT enables the model to concentrate on the key information relevant to the judgment and to correctly understand the legal meaning of acts, as compared to other methods. Our method enables LLMs to predict judgment along with law articles and justification, which significantly enhances the explainability of models.",,
-ramp: retrieval and attribute-marking enhanced prompting for attribute-controlled translation,Gabriele Sarti,"Attribute-controlled translation (ACT) is a subtask of machine translation that involves controlling stylistic or linguistic attributes (like formality and gender) of translation outputs. While ACT has garnered attention in recent years due to its usefulness in real-world applications, progress in the task is currently limited by dataset availability, since most prior approaches rely on supervised methods. To address this limitation, we propose Retrieval and Attribute-Marking enhanced Prompting (RAMP), which leverages large multilingual language models to perform ACT in few-shot and zero-shot settings. RAMP improves generation accuracy over the standard prompting approach by (1) incorporating a semantic similarity retrieval component for selecting similar in-context examples, and (2) marking in-context examples with attribute annotations. Our comprehensive experiments show that RAMP is a viable approach in both zero-shot and few-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'Retrieval and Attribute-Marking enhanced Prompting (RAMP) is proposed, which leverages large multilingual language models to performAttribute-controlled translation in few-shot and zero-shot settings and improves generation accuracy over the standard prompting approach.'}",
-pieclass: weakly-supervised text classification with prompting and noise-robust iterative ensemble training,Yunyi Zhang,"Weakly-supervised text classification trains a classifier using the label name of each target class as the only supervision, which largely reduces human annotation efforts. Most existing methods first use the label names as static keyword-based features to generate pseudo labels, which are then used for final classifier training. While reasonable, such a commonly adopted framework suffers from two limitations: (1) keywords can have different meanings in different contexts and some text may not have any keyword, so keyword matching can induce noisy and inadequate pseudo labels; (2) the errors made in the pseudo label generation stage will directly propagate to the classifier training stage without a chance of being corrected. In this paper, we propose a new method, PIEClass, consisting of two modules: (1) a pseudo label acquisition module that uses zero-shot prompting of pre-trained language models (PLM) to get pseudo labels based on contextualized text understanding beyond static keyword matching, and (2) a noise-robust iterative ensemble training module that iteratively trains classifiers and updates pseudo labels by utilizing two PLM fine-tuning methods that regularize each other. Extensive experiments show that PIEClass achieves overall better performance than existing strong baselines on seven benchmark datasets and even achieves similar performance to fully-supervised classifiers on sentiment classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'A new method, PIEClass, consisting of a pseudo label acquisition module that uses zero-shot prompting of pre-trained language models (PLM) to get pseudo labels based on contextualized text understanding beyond static keyword matching, and a noise-robust iterative ensemble training module that iteratively trains classifiers and updates pseudo labels by utilizing two PLM fine-tuning methods that regularize each other.'}",
-self-explanation prompting improves dialogue understanding in large language models,Haoyu Gao,"Task-oriented dialogue (TOD) systems facilitate users in executing various activities via multi-turn dialogues, but Large Language Models (LLMs) often struggle to comprehend these intricate contexts. In this study, we propose a novel""Self-Explanation""prompting strategy to enhance the comprehension abilities of LLMs in multi-turn dialogues. This task-agnostic approach requires the model to analyze each dialogue utterance before task execution, thereby improving performance across various dialogue-centric tasks. Experimental results from six benchmark datasets confirm that our method consistently outperforms other zero-shot prompts and matches or exceeds the efficacy of few-shot prompts, demonstrating its potential as a powerful tool in enhancing LLMs' comprehension in complex dialogue tasks.","{'model': 'tldr@v2.0.0', 'text': 'A novel Self-Explanation Prompting strategy to enhance the comprehension abilities of LLMs in multi-turn dialogues, which requires the model to analyze each dialogue utterance before task execution, thereby improving performance across various dialogue-centric tasks.'}",https://arxiv.org/pdf/2309.12940
-fixed input parameterization for efficient prompting,Eunbi Choi,"Recent works have shown that attaching prompts to the input is effective at conditioning Language Models (LM) to perform specific tasks. However, prompts are always included in the input text during inference, even when they are fixed, thus incurring substantial computational and memory overhead. Also, there is currently no straightforward method of utilizing prompts that are longer than the maximum input length of the LMs without incurring additional costs during inference. We formally define Fixed Input Parameterization (FIP) problem that focuses on injecting the fixed prompt into the parameters of an LM to be an efficient alternative to attaching fixed prompts to the input. We show that in scenarios with long fixed prompts, FIP can be up to 280 times more efficient in terms of total FLOPs than previous approaches. We further explore methodologies for FIP and show promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions. Through these explorations, we show that FIP can be a promising direction for conditioning language models, in scenarios with long and fixed prompts 1 .","{'model': 'tldr@v2.0.0', 'text': 'It is shown that in scenarios with long fixed prompts, FIP can be up to 280 times more efficient in terms of total FLOPs than previous approaches and further explore methodologies for FIP and shows promising results in persona-dependent conversation, semantic parsing, and zero-shot learning with task instructions.'}",https://aclanthology.org/2023.findings-acl.533.pdf
-map: low-data regime multimodal learning with adapter-based pre-training and prompting,Wenyan Li,"Pretrained vision-language (VL) models have shown impressive results on various multi-modal downstream tasks recently. Many of the benchmark models build on pretrained causal language models (LMs), leveraging the original few-shot learning and generalization capability of the LMs trained with large text corpora. However, these models are often gigantic and require large-scale image and text data with high computational cost to train. This paper introduces a moderate-size model called MAP for efficient VL transfer learning through adapter-based pretraining and prompting. We aim to answer the question of how much we can complete through VL pretraining within the low-data regime while maximizing efficiency in transferring knowledge of a moderate-size frozen LM. Our experiments demonstrate that MAP achieves substantially better zero-shot and few-shot performance on downstream VL tasks with only 10% the size of pretraining data and a 30x lighter pretrained LM backbone compared to Frozen. MAP also outperforms fully trained models of comparable size at retaining its transfer learning ability when the amount of training data reduces.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a moderate-size model called MAP for efficient VL transfer learning through adapter-based pretraining and prompting and achieves substantially better zero-shot and few-shot performance on downstream VL tasks with only 10% the size of pretraining data.'}",
-p5: plug-and-play persona prompting for personalized response selection,Joosung Lee,"The use of persona-grounded retrieval-based chatbots is crucial for personalized conversations, but there are several challenges that need to be addressed. 1) In general, collecting persona-grounded corpus is very expensive. 2) The chatbot system does not always respond in consideration of persona at real applications. To address these challenges, we propose a plug-and-play persona prompting method. Our system can function as a standard open-domain chatbot if persona information is not available. We demonstrate that this approach performs well in the zero-shot setting, which reduces the dependence on persona-ground training data. This makes it easier to expand the system to other languages without the need to build a persona-grounded corpus. Additionally, our model can be fine-tuned for even better performance. In our experiments, the zero-shot model improved the standard model by 7.71 and 1.04 points in the original persona and revised persona, respectively. The fine-tuned model improved the previous state-of-the-art system by 1.95 and 3.39 points in the original persona and revised persona, respectively. To the best of our knowledge, this is the first attempt to solve the problem of personalized response selection using prompt sequences. Our code is available on github~\footnote{https://github.com/rungjoo/plug-and-play-prompt-persona}.","{'model': 'tldr@v2.0.0', 'text': 'This is the first attempt to solve the problem of personalized response selection using prompt sequences using a plug-and-play persona prompting method, and it is demonstrated that this approach performs well in the zero-shot setting, which reduces the dependence on persona-ground training data.'}",https://arxiv.org/pdf/2310.06390
-cof-cot: enhancing large language models with coarse-to-fine chain-of-thought prompting for multi-domain nlu tasks,Hoang Nguyen,"While Chain-of-Thought prompting is popular in reasoning tasks, its application to Large Language Models (LLMs) in Natural Language Understanding (NLU) is under-explored. Motivated by multi-step reasoning of LLMs, we propose Coarse-to-Fine Chain-of-Thought (CoF-CoT) approach that breaks down NLU tasks into multiple reasoning steps where LLMs can learn to acquire and leverage essential concepts to solve tasks from different granularities. Moreover, we propose leveraging semantic-based Abstract Meaning Representation (AMR) structured knowledge as an intermediate step to capture the nuances and diverse structures of utterances, and to understand connections between their varying levels of granularity. Our proposed approach is demonstrated effective in assisting the LLMs adapt to the multi-grained NLU tasks under both zero-shot and few-shot multi-domain settings.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Coarse-to-Fine Chain-of-Thought (CoF-CoT) approach that breaks down NLU tasks into multiple reasoning steps where LLMs can learn to acquire and leverage essential concepts to solve tasks from different granularities.'}",
-a communication theory perspective on prompting engineering methods for large language models,Yuanfeng Song,"The springing up of Large Language Models (LLMs) has shifted the community from single-task-orientated natural language processing (NLP) research to a holistic end-to-end multi-task learning paradigm. Along this line of research endeavors in the area, LLM-based prompting methods have attracted much attention, partially due to the technological advantages brought by prompt engineering (PE) as well as the underlying NLP principles disclosed by various prompting methods. Traditional supervised learning usually requires training a model based on labeled data and then making predictions. In contrast, PE methods directly use the powerful capabilities of existing LLMs (i.e., GPT-3 and GPT-4) via composing appropriate prompts, especially under few-shot or zero-shot scenarios. Facing the abundance of studies related to the prompting and the ever-evolving nature of this field, this article aims to (i) illustrate a novel perspective to review existing PE methods, within the well-established communication theory framework; (ii) facilitate a better/deeper understanding of developing trends of existing PE methods used in four typical tasks; (iii) shed light on promising research directions for future PE methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel perspective is illustrated to review existingPE methods, within the well-established communication theory framework to facilitate a better/deeper understanding of developing trends of existing PE methods used in four typical tasks and shed light on promising research directions for future PE methods.'}",
-winning solution for the cvpr2023 visual anomaly and novelty detection challenge: multimodal prompting for data-centric anomaly detection,Yunkang Cao,"This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge. Going beyond uni-modal prompt, e.g., language prompt, we present a novel framework, i.e., Segment Any Anomaly + (SAA + ), for zero-shot anomaly segmentation with multi-modal prompts for the regularization of cascaded modern foundation models. Inspired by the great zero-shot generalization ability of foundation models like Segment Anything, we first explore their assembly (SAA) to leverage diverse multi-modal prior knowledge for anomaly localization. Subsequently, we further introduce multimodal prompts (SAA + ) derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation. The proposed SAA + model achieves state-of-the-art performance on several anomaly segmentation benchmarks, including VisA and MVTec-AD, in the zero-shot setting. We will release the code of our winning solution for the CVPR2023 VAND challenge at https:/","{'model': 'tldr@v2.0.0', 'text': 'This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge and introduces multimodal prompts derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation.'}",https://arxiv.org/pdf/2306.09067
-2nd place winning solution for the cvpr2023 visual anomaly and novelty detection challenge: multimodal prompting for data-centric anomaly detection,Yunkang Cao,"This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge. Going beyond uni-modal prompt, e.g., language prompt, we present a novel framework, i.e., Segment Any Anomaly + (SAA$+$), for zero-shot anomaly segmentation with multi-modal prompts for the regularization of cascaded modern foundation models. Inspired by the great zero-shot generalization ability of foundation models like Segment Anything, we first explore their assembly (SAA) to leverage diverse multi-modal prior knowledge for anomaly localization. Subsequently, we further introduce multimodal prompts (SAA$+$) derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation. The proposed SAA$+$ model achieves state-of-the-art performance on several anomaly segmentation benchmarks, including VisA and MVTec-AD, in the zero-shot setting. We will release the code of our winning solution for the CVPR2023 VAN.","{'model': 'tldr@v2.0.0', 'text': 'This technical report introduces the winning solution of the team Segment Any Anomaly for the CVPR2023 Visual Anomaly and Novelty Detection (VAND) challenge and introduces multimodal prompts derived from domain expert knowledge and target image context to enable the non-parameter adaptation of foundation models to anomaly segmentation.'}",
-prompting segmentation with sound is generalizable audio-visual source localizer,Yaoting Wang,"Never having seen an object and heard its sound simultaneously, can the model still accurately localize its visual position from the input audio? In this work, we concentrate on the Audio-Visual Localization and Segmentation tasks but under the demanding zero-shot and few-shot scenarios. To achieve this goal, different from existing approaches that mostly employ the encoder-fusion-decoder paradigm to decode localization information from the fused audio-visual feature, we introduce the encoder-prompt-decoder paradigm, aiming to better fit the data scarcity and varying data distribution dilemmas with the help of abundant knowledge from pre-trained models. Specifically, we first propose to construct Semantic-aware Audio Prompt (SAP) to help the visual foundation model focus on sounding objects, meanwhile, the semantic gap between the visual and audio modalities is also encouraged to shrink. Then, we develop a Correlation Adapter (ColA) to keep minimal training efforts as well as maintain adequate knowledge of the visual foundation model. By equipping with these means, extensive experiments demonstrate that this new paradigm outperforms other fusion-based methods in both the unseen class and cross-dataset settings. We hope that our work can further promote the generalization study of Audio-Visual Localization and Segmentation in practical application scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the encoder-prompt-decoder paradigm, aiming to better fit the data scarcity and varying data distribution dilemmas with the help of abundant knowledge from pre-trained models, and proposes to construct Semantic-aware Audio Prompt to help the visual foundation model focus on sounding objects.'}",https://arxiv.org/pdf/2309.07929
-aspiro: any-shot structured parsing-error-induced reprompting for consistent data-to-text generation,Martin Vejvar,"We present ASPIRO, an approach for structured data verbalisation into short template sentences in zero to few-shot settings. Unlike previous methods, our approach prompts large language models (LLMs) to directly produce entity-agnostic templates, rather than relying on LLMs to faithfully copy the given example entities, or validating/crafting the templates manually. We incorporate LLM re-prompting, triggered by algorithmic parsing checks, as well as the PARENT metric induced consistency validation to identify and rectify template generation problems in real-time. ASPIRO, compared to direct LLM output, averages 66\% parsing error rate reduction in generated verbalisations of RDF triples on the DART dataset. Our best 5-shot text-davinci-003 setup, scoring BLEU of 50.62, METEOR of 45.16, BLEURT of 0.82, NUBIA of 0.87, and PARENT of 0.8962 on the Rel2Text dataset, competes effectively with recent fine-tuned pre-trained language models.","{'model': 'tldr@v2.0.0', 'text': None}",
-prompting strategies for citation classification,Suchetha Nambanoor Kunnath,"Citation classification aims to identify the purpose of the cited article in the citing article. Previous citation classification methods rely largely on supervised approaches. The models are trained on datasets with citing sentences or citation contexts annotated for a citation's purpose or function or intent. Recent advancements in Large Language Models (LLMs) have dramatically improved the ability of NLP systems to achieve state-of-the-art performances under zero or few-shot settings. This makes LLMs particularly suitable for tasks where sufficiently large labelled datasets are not yet available, which remains to be the case for citation classification. This paper systematically investigates the effectiveness of different prompting strategies for citation classification and compares them to promptless strategies as a baseline. Specifically, we evaluate the following four strategies, two of which we introduce for the first time, which involve updating Language Model (LM) parameters while training the model: (1) Promptless fine-tuning, (2) Fixed-prompt LM tuning, (3) Dynamic Context-prompt LM tuning (proposed), (4) Prompt + LM fine-tuning (proposed). Additionally, we test the zero-shot performance of LLMs, GPT3.5, a (5) Tuning-free prompting strategy that involves no parameter updating. Our results show that prompting methods based on LM parameter updating significantly improve citation classification performances on both domain-specific and multi-disciplinary citation classifications. Moreover, our Dynamic Context-prompting method achieves top scores both for the ACL-ARC and ACT2 citation classification datasets, surpassing the highest-performing system in the 3C shared task benchmark. Interestingly, we observe zero-shot GPT3.5 to perform well on ACT2 but poorly on the ACL-ARC dataset.","{'model': 'tldr@v2.0.0', 'text': 'This paper systematically investigates the effectiveness of different prompting strategies for citation classification and compares them to promptless strategies as a baseline, showing that prompting methods based on LM parameter updating significantly improve citation classification performances on both domain-specific and multi-disciplinary citation classifications.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615018
-expanding language-image pretrained models for general video recognition,Bolin Ni,"Contrastive language-image pretraining has shown great success in learning visual-textual joint representation from web-scale data, demonstrating remarkable""zero-shot""generalization ability for various image tasks. However, how to effectively expand such new language-image pretraining methods to video domains is still an open problem. In this work, we present a simple yet effective approach that adapts the pretrained language-image models to video recognition directly, instead of pretraining a new model from scratch. More concretely, to capture the long-range dependencies of frames along the temporal dimension, we propose a cross-frame attention mechanism that explicitly exchanges information across frames. Such module is lightweight and can be plugged into pretrained language-image models seamlessly. Moreover, we propose a video-specific prompting scheme, which leverages video content information for generating discriminative textual prompts. Extensive experiments demonstrate that our approach is effective and can be generalized to different video recognition scenarios. In particular, under fully-supervised settings, our approach achieves a top-1 accuracy of 87.1% on Kinectics-400, while using 12 times fewer FLOPs compared with Swin-L and ViViT-H. In zero-shot experiments, our approach surpasses the current state-of-the-art methods by +7.6% and +14.9% in terms of top-1 accuracy under two popular protocols. In few-shot scenarios, our approach outperforms previous best methods by +32.1% and +23.1% when the labeled data is extremely limited. Code and models are available at https://aka.ms/X-CLIP","{'model': 'tldr@v2.0.0', 'text': 'This work presents a simple yet effective approach that adapts the pretrained language-image models to video recognition directly, instead of pretraining a new model from scratch, and proposes a cross-frame attention mechanism that explicitly exchanges information across frames.'}",http://arxiv.org/pdf/2208.02816
-can large language models transform computational social science?,Caleb Ziems,"Large Language Models (LLMs) like ChatGPT are capable of successfully performing many language processing tasks zero-shot (without the need for training data). If this capacity also applies to the coding of social phenomena like persuasiveness and political ideology, then LLMs could effectively transform Computational Social Science (CSS). This work provides a road map for using LLMs as CSS tools. Towards this end, we contribute a set of prompting best practices and an extensive evaluation pipeline to measure the zero-shot performance of 13 language models on 24 representative CSS benchmarks. On taxonomic labeling tasks (classification), LLMs fail to outperform the best fine-tuned models but still achieve fair levels of agreement with humans. On free-form coding tasks (generation), LLMs produce explanations that often exceed the quality of crowdworkers' gold references. We conclude that today's LLMs can radically augment the CSS research pipeline in two ways: (1) serving as zero-shot data annotators on human annotation teams, and (2) bootstrapping challenging creative generation tasks (e.g., explaining the hidden meaning behind text). In summary, LLMs can significantly reduce costs and increase efficiency of social science analysis in partnership with humans.","{'model': 'tldr@v2.0.0', 'text': ""It is concluded that today's LLMs can radically augment the CSS research pipeline in two ways: serving as zero-shot data annotators on human annotation teams, and bootstrapping challenging creative generation tasks (e.g., explaining the hidden meaning behind text).""}",http://arxiv.org/pdf/2305.03514
-segment anything in high quality,Lei Ke,"The recent Segment Anything Model (SAM) represents a big leap in scaling up segmentation models, allowing for powerful zero-shot capabilities and flexible prompting. Despite being trained with 1.1 billion masks, SAM's mask prediction quality falls short in many cases, particularly when dealing with objects that have intricate structures. We propose HQ-SAM, equipping SAM with the ability to accurately segment any object, while maintaining SAM's original promptable design, efficiency, and zero-shot generalizability. Our careful design reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation. We design a learnable High-Quality Output Token, which is injected into SAM's mask decoder and is responsible for predicting the high-quality mask. Instead of only applying it on mask-decoder features, we first fuse them with early and final ViT features for improved mask details. To train our introduced learnable parameters, we compose a dataset of 44K fine-grained masks from several sources. HQ-SAM is only trained on the introduced detaset of 44k masks, which takes only 4 hours on 8 GPUs. We show the efficacy of HQ-SAM in a suite of 9 diverse segmentation datasets across different downstream tasks, where 7 out of them are evaluated in a zero-shot transfer protocol. Our code and models will be released at https://github.com/SysCV/SAM-HQ.","{'model': 'tldr@v2.0.0', 'text': ""A learnable High-Quality Output Token is injected into SAM's mask decoder and is responsible for predicting the high-quality mask, which reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation.""}",http://arxiv.org/pdf/2306.01567
-solving challenging math word problems using gpt-4 code interpreter with code-based self-verification,Aojun Zhou,"Recent progress in large language models (LLMs) like GPT-4 and PaLM-2 has brought significant advancements in addressing math reasoning problems. In particular, OpenAI's latest version of GPT-4, known as GPT-4 Code Interpreter, shows remarkable performance on challenging math datasets. In this paper, we explore the effect of code on enhancing LLMs' reasoning capability by introducing different constraints on the \textit{Code Usage Frequency} of GPT-4 Code Interpreter. We found that its success can be largely attributed to its powerful skills in generating and executing code, evaluating the output of code execution, and rectifying its solution when receiving unreasonable outputs. Based on this insight, we propose a novel and effective prompting method, explicit \uline{c}ode-based \uline{s}elf-\uline{v}erification~(CSV), to further boost the mathematical reasoning potential of GPT-4 Code Interpreter. This method employs a zero-shot prompt on GPT-4 Code Interpreter to encourage it to use code to self-verify its answers. In instances where the verification state registers as ``False'', the model shall automatically amend its solution, analogous to our approach of rectifying errors during a mathematics examination. Furthermore, we recognize that the states of the verification result indicate the confidence of a solution, which can improve the effectiveness of majority voting. With GPT-4 Code Interpreter and CSV, we achieve an impressive zero-shot accuracy on MATH dataset \textbf{(53.9\% $\to$ 84.3\%)}.","{'model': 'tldr@v2.0.0', 'text': ""The effect of code on enhancing LLMs' reasoning capability by introducing different constraints on the Code Usage Frequency of GPT-4 Code Interpreter is explored, and a novel and effective prompting method, explicit \\uline{c}ode-based \\ULine{s}elf-\\uline {v}erification~(CSV), is proposed to further boost the mathematical reasoning potential of GPN.""}",https://arxiv.org/pdf/2308.07921
-a fine-grained comparison of pragmatic language understanding in humans and language models,Jennifer Hu,"Pragmatics and non-literal language understanding are essential to human communication, and present a long-standing challenge for artificial language models. We perform a fine-grained comparison of language models and humans on seven pragmatic phenomena, using zero-shot prompting on an expert-curated set of English materials. We ask whether models (1) select pragmatic interpretations of speaker utterances, (2) make similar error patterns as humans, and (3) use similar linguistic cues as humans to solve the tasks. We find that the largest models achieve high accuracy and match human error patterns: within incorrect responses, models favor literal interpretations over heuristic-based distractors. We also find preliminary evidence that models and humans are sensitive to similar linguistic cues. Our results suggest that pragmatic behaviors can emerge in models without explicitly constructed representations of mental states. However, models tend to struggle with phenomena relying on social expectation violations.",,http://arxiv.org/pdf/2212.06801
-knowledge-in-context: towards knowledgeable semi-parametric language models,Xiaoman Pan,"Fully-parametric language models generally require a huge number of model parameters to store the necessary knowledge for solving multiple natural language tasks in zero/few-shot settings. In addition, it is hard to adapt to the evolving world knowledge without the costly model re-training. In this paper, we develop a novel semi-parametric language model architecture, Knowledge-in-Context (KiC), which empowers a parametric text-to-text language model with a knowledge-rich external memory. Specifically, the external memory contains six different types of knowledge: entity, dictionary, commonsense, event, script, and causality knowledge. For each input instance, the KiC model adaptively selects a knowledge type and retrieves the most helpful pieces of knowledge. The input instance along with its knowledge augmentation is fed into a text-to-text model (e.g., T5) to generate the output answer, where both the input and the output are in natural language forms after prompting. Interestingly, we find that KiC can be identified as a special mixture-of-experts (MoE) model, where the knowledge selector plays the role of a router that is used to determine the sequence-to-expert assignment in MoE. This key observation inspires us to develop a novel algorithm for training KiC with an instance-adaptive knowledge selector. As a knowledge-rich semi-parametric language model, KiC only needs a much smaller parametric part to achieve superior zero-shot performance on unseen tasks. By evaluating on 40+ different tasks, we show that KiC_Large with 770M parameters easily outperforms large language models (LMs) that are 4-39x larger by a large margin. We also demonstrate that KiC exhibits emergent abilities at a much smaller model scale compared to the fully-parametric models.","{'model': 'tldr@v2.0.0', 'text': 'A novel semi-parametric language model architecture, Knowledge-in-Context (KiC), which empowers a parametric text-to-text language model with a knowledge-rich external memory and finds that KiC can be identified as a special mixture-of-experts (MoE) model, where the knowledge selector plays the role of a router that is used to determine the sequence- to-expert assignment in MoE.'}",http://arxiv.org/pdf/2210.16433
-learning to decompose visual features with latent textual prompts,Feng Wang,"Recent advances in pre-training vision-language models like CLIP have shown great potential in learning transferable visual representations. Nonetheless, for downstream inference, CLIP-like models suffer from either 1) degraded accuracy and robustness in the case of inaccurate text descriptions during retrieval-based inference (the challenge for zero-shot protocol); or 2) breaking the well-established vision-language alignment (the challenge for linear probing). To address them, we propose Decomposed Feature Prompting (DeFo). DeFo leverages a flexible number of learnable embeddings as textual input while maintaining the vision-language dual-model architecture, which enables the model to learn decomposed visual features with the help of feature-level textual prompts. We further use an additional linear layer to perform classification, allowing a scalable size of language inputs. Our empirical study shows DeFo's significance in improving the vision-language models. For example, DeFo obtains 73.2% test accuracy on ImageNet with a ResNet-50 backbone without tuning any pretrained weights of both the vision and language encoder, outperforming zero-shot CLIP by a large margin of 15.0%, and outperforming state-of-the-art vision-language prompt tuning method by 7.6%.","{'model': 'tldr@v2.0.0', 'text': 'Decomposed Feature Prompting (DeFo) leverages a flexible number of learnable embeddings as textual input while maintaining the vision-language dual-model architecture, which enables the model to learn decomposed visual features with the help of feature-level textual prompts.'}",http://arxiv.org/pdf/2210.04287
-xricl: cross-lingual retrieval-augmented in-context learning for cross-lingual text-to-sql semantic parsing,Peng Shi,"In-context learning using large language models has recently shown surprising results for semantic parsing tasks such as Text-to-SQL translation. Prompting GPT-3 or Codex using several examples of question-SQL pairs can produce excellent results, comparable to state-of-the-art finetuning-based models. However, existing work primarily focuses on English datasets, and it is unknown whether large language models can serve as competitive semantic parsers for other languages. To bridge this gap, our work focuses on cross-lingual Text-to-SQL semantic parsing for translating non-English utterances into SQL queries based on an English schema. We consider a zero-shot transfer learning setting with the assumption that we do not have any labeled examples in the target language (but have annotated examples in English). This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts. We also include global translation exemplars for a target language to facilitate the translation process for large language models. To systematically evaluate our model, we construct two new benchmark datasets, XSpider and XKaggle-dbqa, which include questions in Chinese, Vietnamese, Farsi, and Hindi. Our experiments show that XRICL effectively leverages large pre-trained language models to outperform existing baselines. Data and code are publicly available at https://github.com/Impavidity/XRICL.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts and effectively leverages large pre-trained language models to outperform existing baselines.'}",http://arxiv.org/pdf/2210.13693
-revisiting non-english text simplification: a unified multilingual benchmark,Michael Joseph Ryan,"Recent advancements in high-quality, large-scale English resources have pushed the frontier of English Automatic Text Simplification (ATS) research. However, less work has been done on multilingual text simplification due to the lack of a diverse evaluation benchmark that covers complex-simple sentence pairs in many languages. This paper introduces the MultiSim benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs. This benchmark will encourage research in developing more effective multilingual text simplification models and evaluation metrics. Our experiments using MultiSim with pre-trained multilingual language models reveal exciting performance improvements from multilingual training in non-English settings. We observe strong performance from Russian in zero-shot cross-lingual transfer to low-resource languages. We further show that few-shot prompting with BLOOM-176b achieves comparable quality to reference simplifications outperforming fine-tuned models in most languages. We validate these findings through human evaluation.","{'model': 'tldr@v2.0.0', 'text': 'The MultiSim benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs, is introduced to encourage research in developing more effective multilingual text simplification models and evaluation metrics.'}",http://arxiv.org/pdf/2305.15678
-multidimensional evaluation for text style transfer using chatgpt,Huiyuan Lai,"We investigate the potential of ChatGPT as a multidimensional evaluator for the task of \emph{Text Style Transfer}, alongside, and in comparison to, existing automatic metrics as well as human judgements. We focus on a zero-shot setting, i.e. prompting ChatGPT with specific task instructions, and test its performance on three commonly-used dimensions of text style transfer evaluation: style strength, content preservation, and fluency. We perform a comprehensive correlation analysis for two transfer directions (and overall) at different levels. Compared to existing automatic metrics, ChatGPT achieves competitive correlations with human judgments. These preliminary results are expected to provide a first glimpse into the role of large language models in the multidimensional evaluation of stylized text generation.","{'model': 'tldr@v2.0.0', 'text': 'Preliminary results are expected to provide a first glimpse into the role of large language models in the multidimensional evaluation of stylized text generation.'}",http://arxiv.org/pdf/2304.13462
-yes but.. can chatgpt identify entities in historical documents?,Carlos-Emiliano Gonz'alez-Gallardo,"Large language models (LLMs) have been leveraged for several years now, obtaining state-of-the-art performance in recognizing entities from modern documents. For the last few months, the conversational agent ChatGPT has “prompted” a lot of interest in the scientific community and public due to its capacity of generating plausible-sounding answers. In this paper, we explore this ability by probing it in the named entity recognition and classification (NERC) task in primary sources (e.g., historical newspapers and classical commentaries) in a zero-shot manner and by comparing it with state-of-the-art LM-based systems. Our findings indicate several shortcomings in identifying entities in historical text that range from the consistency of entity annotation guidelines, entity complexity, and code-switching, to the specificity of prompting. Moreover, as expected, the inaccessibility of historical archives to the public (and thus on the Internet) also impacts its performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the ability of the conversational agent ChatGPT to generate plausible-sounding answers in the named entity recognition and classification task in primary sources in a zero-shot manner and compares it with state-of-the-art LM-based systems.'}",https://arxiv.org/pdf/2303.17322
-chain of thought prompt tuning in vision language models,Jiaxin Ge,"Language-Image Pre-training has demonstrated promising results on zero-shot and few-shot downstream tasks by prompting visual models with natural language prompts. However, most recent studies only use a single prompt for tuning, neglecting the inherent step-to-step cognitive reasoning process that humans conduct in complex task settings, for example, when processing images from unfamiliar domains. Chain of Thought is a simple and effective approximation to human reasoning process and has been proven useful for natural language processing (NLP) tasks. Based on this cognitive intuition, we believe that conducting effective reasoning is also an important problem in visual tasks, and a chain of thought could be a solution to this problem. In this work, we propose a novel chain of thought prompt tuning for vision-language modeling. Extensive experiments show that our method not only generalizes better in image classification tasks, has greater transferability beyond a single dataset, and has stronger domain generalization performance, but also performs much better in imagetext retrieval and visual question answering, which require more reasoning capabilities. We are the first to successfully adapt chain-of-thought prompting that combines visual and textual embeddings. We will release our codes","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that the proposed novel chain of thought prompt tuning for vision-language modeling not only generalizes better in image classification tasks, has greater transferability beyond a single dataset, and has stronger domain generalization performance, but also performs much better in imagetext retrieval and visual question answering, which require more reasoning capabilities.'}",
-can language models understand physical concepts?,Lei Li,"Language models~(LMs) gradually become general-purpose interfaces in the interactive and embodied world, where the understanding of physical concepts is an essential prerequisite. However, it is not yet clear whether LMs can understand physical concepts in the human world. To investigate this, we design a benchmark VEC that covers the tasks of (i) Visual concepts, such as the shape and material of objects, and (ii) Embodied Concepts, learned from the interaction with the world such as the temperature of objects. Our zero (few)-shot prompting results show that the understanding of certain visual concepts emerges as scaling up LMs, but there are still basic concepts to which the scaling law does not apply. For example, OPT-175B performs close to humans with a zero-shot accuracy of 85\% on the material concept, yet behaves like random guessing on the mass concept. Instead, vision-augmented LMs such as CLIP and BLIP achieve a human-level understanding of embodied concepts. Analysis indicates that the rich semantics in visual representation can serve as a valuable source of embodied knowledge. Inspired by this, we propose a distillation method to transfer embodied knowledge from VLMs to LMs, achieving performance gain comparable with that by scaling up the parameters of LMs 134x. Our dataset is available at \url{https://github.com/TobiasLee/VEC}","{'model': 'tldr@v2.0.0', 'text': 'A distillation method is proposed to transfer embodied knowledge from VLMs to LMs, achieving performance gain comparable with that by scaling up the parameters of LMs 134x, and indicates that the rich semantics in visual representation can serve as a valuable source of embodied knowledge.'}",http://arxiv.org/pdf/2305.14057
-evaluating large language models on graphs: performance insights and comparative analysis,Chang Liu,"Large Language Models (LLMs) have garnered considerable interest within both academic and industrial. Yet, the application of LLMs to graph data remains under-explored. In this study, we evaluate the capabilities of four LLMs in addressing several analytical problems with graph data. We employ four distinct evaluation metrics: Comprehension, Correctness, Fidelity, and Rectification. Our results show that: 1) LLMs effectively comprehend graph data in natural language and reason with graph topology. 2) GPT models can generate logical and coherent results, outperforming alternatives in correctness. 3) All examined LLMs face challenges in structural reasoning, with techniques like zero-shot chain-of-thought and few-shot prompting showing diminished efficacy. 4) GPT models often produce erroneous answers in multi-answer tasks, raising concerns in fidelity. 5) GPT models exhibit elevated confidence in their outputs, potentially hindering their rectification capacities. Notably, GPT-4 has demonstrated the capacity to rectify responses from GPT-3.5-turbo and its own previous iterations. The code is available at: https://github.com/Ayame1006/LLMtoGraph.","{'model': 'tldr@v2.0.0', 'text': 'This study evaluates the capabilities of four LLMs in addressing several analytical problems with graph data, using four distinct evaluation metrics: Comprehension, Correctness, Fidelity, and Rectification.'}",https://arxiv.org/pdf/2308.11224
-speechx: neural codec language model as a versatile speech transformer,Xiaofei Wang,"Recent advancements in generative speech models based on audio-text prompts have enabled remarkable innovations like high-quality zero-shot text-to-speech. However, existing models still face limitations in handling diverse audio-text speech generation tasks involving transforming input speech and processing audio captured in adverse acoustic conditions. This paper introduces SpeechX, a versatile speech generation model capable of zero-shot TTS and various speech transformation tasks, dealing with both clean and noisy signals. SpeechX combines neural codec language modeling with multi-task learning using task-dependent prompting, enabling unified and extensible modeling and providing a consistent way for leveraging textual input in speech enhancement and transformation tasks. Experimental results show SpeechX's efficacy in various tasks, including zero-shot TTS, noise suppression, target speaker extraction, speech removal, and speech editing with or without background noise, achieving comparable or superior performance to specialized models across tasks. See https://aka.ms/speechx for demo samples.","{'model': 'tldr@v2.0.0', 'text': 'SpeechX combines neural codec language modeling with multi-task learning using task-dependent prompting, enabling unified and extensible modeling and providing a consistent way for leveraging textual input in speech enhancement and transformation tasks.'}",https://arxiv.org/pdf/2308.06873
-is chatgpt a good personality recognizer? a preliminary study,Yuzhe Ji,"In recent years, personality has been regarded as a valuable personal factor being incorporated into numerous tasks such as sentiment analysis and product recommendation. This has led to widespread attention to text-based personality recognition task, which aims to identify an individual's personality based on given text. Considering that ChatGPT has recently exhibited remarkable abilities on various natural language processing tasks, we provide a preliminary evaluation of ChatGPT on text-based personality recognition task for generating effective personality data. Concretely, we employ a variety of prompting strategies to explore ChatGPT's ability in recognizing personality from given text, especially the level-oriented prompting strategy we designed for guiding ChatGPT in analyzing given text at a specified level. The experimental results on two representative real-world datasets reveal that ChatGPT with zero-shot chain-of-thought prompting exhibits impressive personality recognition ability and is capable to provide natural language explanations through text-based logical reasoning. Furthermore, by employing the level-oriented prompting strategy to optimize zero-shot chain-of-thought prompting, the performance gap between ChatGPT and corresponding state-of-the-art model has been narrowed even more. However, we observe that ChatGPT shows unfairness towards certain sensitive demographic attributes such as gender and age. Additionally, we discover that eliciting the personality recognition ability of ChatGPT helps improve its performance on personality-related downstream tasks such as sentiment classification and stress prediction.","{'model': 'tldr@v2.0.0', 'text': 'E eliciting the personality recognition ability of ChatGPT helps improve its performance on personality-related downstream tasks such as sentiment classification and stress prediction, and is found to provide natural language explanations through text-based logical reasoning.'}",https://arxiv.org/pdf/2307.03952
-let's do a thought experiment: using counterfactuals to improve moral reasoning,Xiao Ma,"Language models still struggle on moral reasoning, despite their impressive performance in many other tasks. In particular, the Moral Scenarios task in MMLU (Multi-task Language Understanding) is among the worst performing tasks for many language models, including GPT-3. In this work, we propose a new prompting framework, Thought Experiments, to teach language models to do better moral reasoning using counterfactuals. Experiment results show that our framework elicits counterfactual questions and answers from the model, which in turn helps improve the accuracy on Moral Scenarios task by 9-16% compared to other zero-shot baselines. Interestingly, unlike math reasoning tasks, zero-shot Chain-of-Thought (CoT) reasoning doesn't work out of the box, and even reduces accuracy by around 4% compared to direct zero-shot. We further observed that with minimal human supervision in the form of 5 few-shot examples, the accuracy of the task can be improved to as much as 80%.","{'model': 'tldr@v2.0.0', 'text': 'A new prompting framework is proposed, Thought Experiments, to teach language models to do better moral reasoning using counterfactuals, and shows that with minimal human supervision in the form of 5 few-shot examples, the accuracy of the task can be improved to as much as 80%.'}",http://arxiv.org/pdf/2306.14308
-symbolic math reasoning with language models,Vedant Gaur,"The emergence of large language models (LLMs) such as OpenAI’s GPT-3, Google’s LaMDA, Meta’s OPT [2, 3, 7, 10] etc. have revolutionized the field of natural language processing (NLP). These models with upwards of hundreds of billions of parameters are trained on large unlabeled text corpora and can subsequently solve downstream tasks with little to no labeled data. While these models are increasingly versatile in their abilities, e.g., solving math word problems, the larger question of their ability to reason remains. Using and modifying the SVAMP dataset, we find that GPT-3’s davinci-002 model, in addition to having good performance on numerical math word problems, also performs well on the potentially harder symbolic version of the same problems. Furthermore, adopting a two-step approach (solve symbolically and then substitute numerical values) leads to better accuracy on the numerical test set in the zero-shot regime. Additionally, we find that the use of specific prompting techniques pushes the model, in many cases, to actively describe its thought process and aid in the final answer output when faced with a complex, multi-step problem, aligning with recent observations.","{'model': 'tldr@v2.0.0', 'text': 'GPT-3’s davinci-002 model, in addition to having good performance on numerical math word problems, also performs well on the potentially harder symbolic version of the same problems, and adopting a two-step approach leads to better accuracy on the numerical test set in the zero-shot regime.'}",
-instructexcel: a benchmark for natural language instruction in excel,Justin Payan,"With the evolution of Large Language Models (LLMs) we can solve increasingly more complex NLP tasks across various domains, including spreadsheets. This work investigates whether LLMs can generate code (Excel OfficeScripts, a TypeScript API for executing many tasks in Excel) that solves Excel specific tasks provided via natural language user instructions. To do so we introduce a new large-scale benchmark, InstructExcel, created by leveraging the 'Automate' feature in Excel to automatically generate OfficeScripts from users' actions. Our benchmark includes over 10k samples covering 170+ Excel operations across 2,000 publicly available Excel spreadsheets. Experiments across various zero-shot and few-shot settings show that InstructExcel is a hard benchmark for state of the art models like GPT-4. We observe that (1) using GPT-4 over GPT-3.5, (2) providing more in-context examples, and (3) dynamic prompting can help improve performance on this benchmark.","{'model': 'tldr@v2.0.0', 'text': ""This work investigates whether LLMs can generate code (Excel OfficeScripts, a TypeScript API for executing many tasks in Excel) that solves Excel specific tasks provided via natural language user instructions by leveraging the 'Automate' feature in Excel to automatically generate Office Scripts from users' actions.""}",
-business process text sketch automation generation using large language model,Rui Zhu,"Business Process Management (BPM) is gaining increasing attention as it has the potential to cut costs while boosting output and quality. Business process document generation is a crucial stage in BPM. However, due to a shortage of datasets, data-driven deep learning techniques struggle to deliver the expected results. We propose an approach to transform Conditional Process Trees (CPTs) into Business Process Text Sketches (BPTSs) using Large Language Models (LLMs). The traditional prompting approach (Few-shot In-Context Learning) tries to get the correct answer in one go, and it can find the pattern of transforming simple CPTs into BPTSs, but for close-domain and CPTs with complex hierarchy, the traditional prompts perform weakly and with low correctness. We suggest using this technique to break down a difficult CPT into a number of basic CPTs and then solve each one in turn, drawing inspiration from the divide-and-conquer strategy. We chose 100 process trees with depths ranging from 2 to 5 at random, as well as CPTs with many nodes, many degrees of selection, and cyclic nesting. Experiments show that our method can achieve a correct rate of 93.42%, which is 45.17% better than traditional prompting methods. Our proposed method provides a solution for business process document generation in the absence of datasets, and secondly, it becomes potentially possible to provide a large number of datasets for the process model extraction (PME) domain.","{'model': 'tldr@v2.0.0', 'text': 'This proposed method provides a solution for business process document generation in the absence of datasets, and secondly, it becomes potentially possible to provide a large number of datasets for the process model extraction (PME) domain.'}",https://arxiv.org/pdf/2309.01071
-an investigation of llms' inefficacy in understanding converse relations,Chengwen Qi,"Large Language Models (LLMs) have achieved remarkable success in many formal language oriented tasks, such as structural data-to-text and semantic parsing. However current benchmarks mostly follow the data distribution of the pre-training data of LLMs. Therefore, a natural question rises that do LLMs really understand the structured semantics of formal languages. In this paper, we investigate this problem on a special case, converse binary relation. We introduce a new benchmark ConvRe focusing on converse relations, which contains 17 relations and 1240 triples extracted from popular knowledge graph completion datasets. Our ConvRE features two tasks, Re2Text and Text2Re, which are formulated as multi-choice question answering to evaluate LLMs' ability to determine the matching between relations and associated text. For the evaluation protocol, apart from different prompting methods, we further introduce variants to the test text and few-shot example text. We conduct experiments on three popular LLM families and have observed various scaling trends. The results suggest that LLMs often resort to shortcut learning and still face challenges on our proposed benchmark.","{'model': 'tldr@v2.0.0', 'text': 'A new benchmark ConvRe focusing on converse relations, which contains 17 relations and 1240 triples extracted from popular knowledge graph completion datasets is introduced, which suggests that LLMs often resort to shortcut learning and still face challenges on this proposed benchmark.'}",https://arxiv.org/pdf/2310.05163
-"of eight oscillations and several messages carved in flesh: spectacle, spectatorship, cognition, and affect in dredd and looper",M. Bould,"Both popular and academic criticism tend to decry sf cinema’s commitment to spectacle and special effects as intellectually stultifying and thus politically narcotizing. This article challenges the class politics of taste and the crude models of interpolation underpinning such claims, questioning the ease with which many critics separate matters of cognition from matters of affect. It examines in detail three sequences from the mainstream, small-to-medium budget sf movies Dredd (2012) and Looper (2012). The climactic set piece from Dredd demonstrates the contradictory but entangled array of unstable subject positions opened up by cinematic spectacle. A sequence from Looper, which includes the most complicated special effects shot in the movie, shows how effects work can affect the viewer while simultaneously elaborating narrative/ world-building information that demands a cognitive response. A second sequence from Looper works in a similar way, ingeniously prompting the viewer to draw on intertextual knowledge to comprehend a shot that conveys complex narrative information and that ultimately refuses the specific special effect it has led the viewer to anticipate.",,
-method and system for detecting living body human face,王先基,"The invention relates to a method and system for detecting a living body human face. The method comprises the steps of obtaining images from a shot video sequence, detecting a human face in the video images, carrying out location on facial critical areas, randomly selecting one or more kinds of movements in a movement collection through a critical area detection classifier, randomly selecting the completing times of the movements, prompting a user, and determining the user as a living body if the user completes the random selected movements for many times within an appointed time period, otherwise, determining the user as a non-living body. According to the method and system for detecting the living body human face, when living body detection is carried out, the system randomly selects certain movements and randomly appoints the completing times of the movements every time, and therefore the user cannot pass the living body detection by using photos for cheating. In addition, due to the fact that the detected items are selected randomly, the user cannot pass the living body detection through the mode of playing video clips. Therefore, the method and system for detecting the living body human face can better prevent cheating behaviors in the mode such as the photos and the video clips in the living body detection, and the reliability and the safety of the living body detection are made to be higher.","{'model': 'tldr@v2.0.0', 'text': 'The method and system for detecting the live body human face can better prevent cheating behaviors in the mode such as the photos and the video clips in the living body detection, and the reliability and the safety of the livingBody detection are made to be higher.'}",
-zero-shot text-guided object generation with dream fields,Ajay Jain,"We combine neural rendering with multi-modal image and text representations to synthesize diverse 3D objects solely from natural language descriptions. Our method, Dream Fields, can generate the geometry and color of a wide range of objects without 3D supervision. Due to the scarcity of diverse, captioned 3D data, prior methods only generate objectsfrom a handful of categories, such as ShapeNet. Instead, we guide generation with image-text models pre-trained on large datasets of captioned images from the web. Our method optimizes a Neural Radiance Field from many camera views so that rendered images score highly with a target caption according to a pre-trained CLIP model. To improve fidelity and visual quality, we introduce simple geometric priors, including sparsity-inducing transmittance regularization, scene bounds, and new MLP architectures. In experiments, Dream Fields produce realistic, multi-view consistent object geometry and color from a variety of natural language captions.","{'model': 'tldr@v2.0.0', 'text': 'This work combines neural rendering with multi-modal image and text representations to synthesize diverse 3D objects solely from natural language descriptions, and introduces simple geometric priors, including sparsity-inducing transmittance regularization, scene bounds, and new MLP architectures.'}",https://arxiv.org/pdf/2112.01455
-large language models are few-shot testers: exploring llm-based general bug reproduction,Sungmin Kang,"Many automated test generation techniques have been developed to aid developers with writing tests. To facilitate full automation, most existing techniques aim to either increase coverage, or generate exploratory inputs. However, existing test generation techniques largely fall short of achieving more semantic objectives, such as generating tests to reproduce a given bug report. Reproducing bugs is nonetheless important, as our empirical study shows that the number of tests added in open source repositories due to issues was about 28% of the corresponding project test suite size. Meanwhile, due to the difficulties of transforming the expected program semantics in bug reports into test oracles, existing failure reproduction techniques tend to deal exclusively with program crashes, a small subset of all bug reports. To automate test generation from general bug reports, we propose Libro, a framework that uses Large Language Models (LLMs), which have been shown to be capable of performing code-related tasks. Since LLMs themselves cannot execute the target buggy code, we focus on post-processing steps that help us discern when LLMs are effective, and rank the produced tests according to their validity. Our evaluation of Libro shows that, on the widely studied Defects4J benchmark, Libro can generate failure reproducing test cases for 33% of all studied cases (251 out of 750), while suggesting a bug reproducing test in first place for 149 bugs. To mitigate data contamination (i.e., the possibility of the LLM simply remembering the test code either partially or in whole), we also evaluate Libro against 31 bug reports submitted after the collection of the LLM training data terminated: Libro produces bug reproducing tests for 32% of the studied bug reports. Overall, our results show Libro has the potential to significantly enhance developer efficiency by automatically generating tests from bug reports.","{'model': 'tldr@v2.0.0', 'text': 'The results show Libro has the potential to significantly enhance developer efficiency by automatically generating tests from bug reports, a framework that uses Large Language Models (LLMs), which have been shown to be capable of performing code-related tasks.'}",https://arxiv.org/pdf/2209.11515
-meta faster r-cnn: towards accurate few-shot object detection with attentive feature alignment,G. Han,"Few-shot object detection (FSOD) aims to detect objects using only a few examples. How to adapt state-of-the-art object detectors to the few-shot domain remains challenging. Object proposal is a key ingredient in modern object detectors. However, the quality of proposals generated for few-shot classes using existing methods is far worse than that of many-shot classes, e.g., missing boxes for few-shot classes due to misclassification or inaccurate spatial locations with respect to true objects. To address the noisy proposal problem, we propose a novel meta-learning based FSOD model by jointly optimizing the few-shot proposal generation and fine-grained few-shot proposal classification. To improve proposal generation for few-shot classes, we propose to learn a lightweight metric-learning based prototype matching network, instead of the conventional simple linear object/nonobject classifier, e.g., used in RPN. Our non-linear classifier with the feature fusion network could improve the discriminative prototype matching and the proposal recall for few-shot classes. To improve the fine-grained few-shot proposal classification, we propose a novel attentive feature alignment method to address the spatial misalignment between the noisy proposals and few-shot classes, thus improving the performance of few-shot object detection. Meanwhile we learn a separate Faster R-CNN detection head for many-shot base classes and show strong performance of maintaining base-classes knowledge. Our model achieves state-of-the-art performance on multiple FSOD benchmarks over most of the shots and metrics.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-learning based FSOD model is proposed that achieves state-of-the-art performance on multiple FSOD benchmarks over most of the shots and metrics and a novel attentive feature alignment method is proposed to address the spatial misalignment between the noisy proposals and few-shot classes, thus improving the performance of few- shot object detection.'}",https://ojs.aaai.org/index.php/AAAI/article/download/19959/19718
-d2c: diffusion-decoding models for few-shot conditional generation,Abhishek Sinha,"Conditional generative models of high-dimensional images have many applications, but supervision signals from conditions to images can be expensive to acquire. This paper describes Diffusion-Decoding models with Contrastive representations (D2C), a paradigm for training unconditional variational autoencoders (VAEs) for few-shot conditional image generation. D2C uses a learned diffusion-based prior over the latent representations to improve generation and contrastive self-supervised learning to improve representation quality. D2C can adapt to novel generation tasks conditioned on labels or manipulation constraints, by learning from as few as 100 labeled examples. On conditional generation from new labels, D2C achieves superior performance over state-of-the-art VAEs and diffusion models. On conditional image manipulation, D2C generations are two orders of magnitude faster to produce over StyleGAN2 ones and are preferred by 50% - 60% of the human evaluators in a double-blind study.","{'model': 'tldr@v2.0.0', 'text': 'D2C uses a learned diffusion-based prior over the latent representations to improve generation and contrastive self-supervised learning to improve representation quality, and achieves superior performance over state-of-the-art VAEs and diffusion models.'}",
-few-shot learning for medical text: a systematic review,Yao Ge,"Objective: Few-shot learning (FSL) methods require small numbers of labeled instances for training. As many medical topics have limited annotated textual data in practical settings, FSL-based natural language processing (NLP) methods hold substantial promise. We aimed to conduct a systematic review to explore the state of FSL methods for medical NLP. Materials and Methods: We searched for articles published between January 2016 and August 2021 using PubMed/Medline, Embase, ACL Anthology, and IEEE Xplore Digital Library. To identify the latest relevant methods, we also searched other sources such as preprint servers (eg., medRxiv) via Google Scholar. We included all articles that involved FSL and any type of medical text. We abstracted articles based on data source(s), aim(s), training set size(s), primary method(s)/approach(es), and evaluation method(s). Results: 31 studies met our inclusion criteria-all published after 2018; 22 (71%) since 2020. Concept extraction/named entity recognition was the most frequently addressed task (13/31; 42%), followed by text classification (10/31; 32%). Twenty-one (68%) studies reconstructed existing datasets to create few-shot scenarios synthetically, and MIMIC-III was the most frequently used dataset (7/31; 23%). Common methods included FSL with attention mechanisms (12/31; 39%), prototypical networks (8/31; 26%), and meta-learning (6/31; 19%). Discussion: Despite the potential for FSL in biomedical NLP, progress has been limited compared to domain-independent FSL. This may be due to the paucity of standardized, public datasets, and the relative underperformance of FSL methods on biomedical topics. Creation and release of specialized datasets for biomedical FSL may aid method development by enabling comparative analyses.","{'model': 'tldr@v2.0.0', 'text': 'Despite the potential for FSL in biomedical NLP, progress has been limited compared to domain-independent FSL, may be due to the paucity of standardized, public datasets, and the relative underperformance of FSL methods on biomedical topics.'}",http://arxiv.org/pdf/2204.14081
-relative and absolute location embedding for few-shot node classification on graph,Zemin Liu,"Node classification is an important problem on graphs. While recent advances in graph neural networks achieve promising performance, they require abundant labeled nodes for training. However, in many practical scenarios, there often exist novel classes in which only one or a few labeled nodes are available as supervision, known as few-shot node classification. Although meta-learning has been widely used in vision and language domains to address few-shot learning, its adoption on graphs has been limited. In particular, graph nodes in a few-shot task are not independent and relate to each other. To deal with this, we propose a novel model called Relative and Absolute Location Embedding (RALE) hinged on the concept of hub nodes. Specifically, RALE captures the task-level dependency by assigning each node a relative location within a task, as well as the graph-level dependency by assigning each node an absolute location on the graph to further align different tasks toward learning a transferable prior. Finally, extensive experiments on three public datasets demonstrate the state-of-the-art performance of RALE.","{'model': 'tldr@v2.0.0', 'text': 'A novel model called Relative and Absolute Location Embedding (RALE) hinged on the concept of hub nodes is proposed, which captures the task-level dependency by assigning each node a relative location within a task, as well as the graph- level dependency by assign each node an absolute location on the graph to further align different tasks toward learning a transferable prior.'}",https://ojs.aaai.org/index.php/AAAI/article/download/16551/16358
-few-shot intent detection via contrastive pre-training and fine-tuning,Jianguo Zhang,"In this work, we focus on a more challenging few-shot intent detection scenario where many intents are fine-grained and semantically similar. We present a simple yet effective few-shot intent detection schema via contrastive pre-training and fine-tuning. Specifically, we first conduct self-supervised contrastive pre-training on collected intent datasets, which implicitly learns to discriminate semantically similar utterances without using any labels. We then perform few-shot intent detection together with supervised contrastive learning, which explicitly pulls utterances from the same intent closer and pushes utterances across different intents farther. Experimental results show that our proposed method achieves state-of-the-art performance on three challenging intent detection datasets under 5-shot and 10-shot settings.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts self-supervised contrastive pre-training on collected intent datasets and performs few-shot intent detection together with supervised contrastive learning, which explicitly pulls utterances from the same intent closer and pushes utterances across different intents farther.'}",https://aclanthology.org/2021.emnlp-main.144.pdf
-relational learning with gated and attentive neighbor aggregator for few-shot knowledge graph completion,Guanglin Niu,"Aiming at expanding few-shot relations' coverage in knowledge graphs (KGs), few-shot knowledge graph completion (FKGC) has recently gained more research interests. Some existing models employ a few-shot relation's multi-hop neighbor information to enhance its semantic representation. However, noise neighbor information might be amplified when the neighborhood is excessively sparse and no neighbor is available to represent the few-shot relation. Moreover, modeling and inferring complex relations of one-to-many (1-N), many-to-one (N-1), and many-to-many (N-N) by previous knowledge graph completion approaches requires high model complexity and a large amount of training instances. Thus, inferring complex relations in the few-shot scenario is difficult for FKGC models due to limited training instances. In this paper, we propose a few-shot relational learning with global-local framework to address the above issues. At the global stage, a novel gated and attentive neighbor aggregator is built for accurately integrating the semantics of a few-shot relation's neighborhood, which helps filtering the noise neighbors even if a KG contains extremely sparse neighborhoods. For the local stage, a meta-learning based TransH (MTransH) method is designed to model complex relations and train our model in a few-shot learning fashion. Extensive experiments show that our model outperforms the state-of-the-art FKGC approaches on the frequently-used benchmark datasets NELL-One and Wiki-One. Compared with the strong baseline model MetaR, our model achieves 5-shot FKGC performance improvements of 8.0% on NELL-One and 2.8% on Wiki-One by the metric Hits@10.","{'model': 'tldr@v2.0.0', 'text': ""A novel gated and attentive neighbor aggregator is built for accurately integrating the semantics of a few-shot relation's neighborhood, which helps filtering the noise neighbors even if a KG contains extremely sparse neighborhoods.""}",https://arxiv.org/pdf/2104.13095
-improving massively multilingual neural machine translation and zero-shot translation,Biao Zhang,"Massively multilingual models for neural machine translation (NMT) are theoretically attractive, but often underperform bilingual models and deliver poor zero-shot translations. In this paper, we explore ways to improve them. We argue that multilingual NMT requires stronger modeling capacity to support language pairs with varying typological characteristics, and overcome this bottleneck via language-specific components and deepening NMT architectures. We identify the off-target translation issue (i.e. translating into a wrong target language) as the major source of the inferior zero-shot performance, and propose random online backtranslation to enforce the translation of unseen training language pairs. Experiments on OPUS-100 (a novel multilingual dataset with 100 languages) show that our approach substantially narrows the performance gap with bilingual models in both one-to-many and many-to-many settings, and improves zero-shot performance by ~10 BLEU, approaching conventional pivot-based methods.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that multilingual NMT requires stronger modeling capacity to support language pairs with varying typological characteristics, and overcome this bottleneck via language-specific components and deepening NMT architectures.'}",https://www.aclweb.org/anthology/2020.acl-main.148.pdf
-zero-shot generalization in dialog state tracking through generative question answering,Shuyang Li,"Dialog State Tracking (DST), an integral part of modern dialog systems, aims to track user preferences and constraints (slots) in task-oriented dialogs. In real-world settings with constantly changing services, DST systems must generalize to new domains and unseen slot types. Existing methods for DST do not generalize well to new slot names and many require known ontologies of slot types and values for inference. We introduce a novel ontology-free framework that supports natural language queries for unseen constraints and slots in multi-domain task-oriented dialogs. Our approach is based on generative question-answering using a conditional language model pre-trained on substantive English sentences. Our model improves joint goal accuracy in zero-shot domain adaptation settings by up to 9% (absolute) over the previous state-of-the-art on the MultiWOZ 2.1 dataset.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel ontology-free framework that supports natural language queries for unseen constraints and slots in multi-domain task-oriented dialogs based on generative question-answering using a conditional language model pre-trained on substantive English sentences.'}",https://aclanthology.org/2021.eacl-main.91.pdf
-zero-shot visual question answering using knowledge graph,Zhuo Chen,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Zero-shot VQA algorithm using knowledge graphs and a mask-based learning mechanism for better incorporating external knowledge, and presents new answer-based Zero- shot VQ a splits for the F-VQA dataset.'}",https://arxiv.org/pdf/2107.05348
-single-shot simulations of dynamic quantum many-body systems,K. Sakmann,,,https://arxiv.org/pdf/1501.03224
-improving zero-shot generalization and robustness of multi-modal models,Yunhao Ge,"Multi-modal- image-text models such as CLIP and LiT have demonstrated impressive performance on image classification benchmarks and their zero-shot generalization ability is particularly exciting. While the top-5 zero-shot accuracies of these models are very high, the top-1 accuracies are much lower (over 25% gap in some cases). We investigate the reasons for this performance gap and find that many of the failure cases are caused by ambiguity in the text prompts. First, we develop a simple and efficient zero-shot post-hoc method to identify images whose top-1 prediction is likely to be incorrect, by measuring consistency of the predictions w.r.t. multiple prompts and image transformations. We show that our procedure better predicts mistakes, outperforming the popular max logit baseline on selective prediction tasks. Next, we propose a simple and efficient way to improve accuracy on such uncertain images by making use of the WordNet hierarchy; specifically we augment the original class by incorporating its parent and children from the semantic label hierarchy, and plug the augmentation into text prompts. We conduct experiments on both CLIP and LiT models with five different ImageNet-based datasets. For CLIP, our method improves the top-1 accuracy by 17.13% on the uncertain subset and 3.6% on the entire ImageNet validation set. We also show that our method improves across ImageNet shifted datasets, four other datasets, and other model architectures such as LiT. The proposed method11Work carried out mainly at Google is hyperparameter-free, requires no additional model training and can be easily scaled to other large multi-modal architectures. Code is available at https://github.com/gyhandy/Hierarchy-CLIP.","{'model': 'tldr@v2.0.0', 'text': 'A simple and efficient zero-shot post-hoc method to identify images whose top-1 prediction is likely to be incorrect, by measuring consistency of the predictions w.r.t. multiple prompts and image transformations is developed.'}",https://arxiv.org/pdf/2212.01758
-enhancing cross-lingual natural language inference by soft prompting with language-independent knowledge,,"Cross-lingual natural language inference is 001 a fundamental problem in cross-lingual lan- 002 guage understanding. Many recent works 003 have used prompt learning to address the lack 004 of annotated parallel corpora in XNLI. How- 005 ever, these methods adopt discrete prompting 006 by simply translating the template to the tar- 007 get language and need external expert knowl- 008 edge to design the templates. Besides, dis- 009 crete prompts of human-designed template 010 words are not trainable vectors which can 011 be migrated to target languages in the infer- 012 ence stage flexibly. In this paper, we pro- 013 pose a novel Soft prompt learning frame- 014 work enhanced by L anguage- IN dependent 015 K nowledge (SoftLINK) for XNLI. SoftLINK 016 first constructs cloze-style question with soft 017 prompts for the input sample. Then we lever- 018 age bilingual dictionaries to generate an aug- 019 mented multilingual question for the original 020 question. SoftLINK also adopts a multilingual 021 verbalizer to align the representations of origi- 022 nal and augmented multilingual questions on 023 the semantic space with consistency regular- 024 ization. Experimental results on XNLI demon- 025 strate that SoftLINK can achieve state-of-the- 026 art performance and significantly outperform 027 the previous methods under the few-shot and 028 full-shot cross-lingual transfer settings. 029","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on XNLI show that SoftLINK can achieve state-of-the-art art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.'}",
-meta self-training for few-shot neural sequence labeling,Yaqing Wang,"Neural sequence labeling is widely adopted for many Natural Language Processing (NLP) tasks, such as Named Entity Recognition (NER) and slot tagging for dialog systems and semantic parsing. Recent advances with large-scale pre-trained language models have shown remarkable success in these tasks when fine-tuned on large amounts of task-specific labeled data. However, obtaining such large-scale labeled training data is not only costly, but also may not be feasible in many sensitive user applications due to data access and privacy constraints. This is exacerbated for sequence labeling tasks requiring such annotations at token-level. In this work, we develop techniques to address the label scarcity challenge for neural sequence labeling models. Specifically, we propose a meta self-training framework which leverages very few manually annotated labels for training neural sequence models. While self-training serves as an effective mechanism to learn from large amounts of unlabeled data via iterative knowledge exchange -- meta-learning helps in adaptive sample re-weighting to mitigate error propagation from noisy pseudo-labels. Extensive experiments on six benchmark datasets including two for massive multilingual NER and four slot tagging datasets for task-oriented dialog systems demonstrate the effectiveness of our method. With only 10 labeled examples for each class in each task, the proposed method achieves 10% improvement over state-of-the-art methods demonstrating its effectiveness for limited training labels regime.","{'model': 'tldr@v2.0.0', 'text': 'A meta self-training framework which leverages very few manually annotated labels for training neural sequence models and achieves 10% improvement over state-of-the-art methods demonstrating its effectiveness for limited training labels regime.'}",https://dl.acm.org/doi/pdf/10.1145/3447548.3467235
-precise zero-shot dense retrieval without relevance labels,Luyu Gao,"While dense retrieval has been shown to be effective and efficient across tasks and languages, it remains difficult to create effective fully zero-shot dense retrieval systems when no relevance labels are available. In this paper, we recognize the difficulty of zero-shot learning and encoding relevance. Instead, we propose to pivot through Hypothetical Document Embeddings (HyDE). Given a query, HyDE first zero-shot prompts an instruction-following language model (e.g., InstructGPT) to generate a hypothetical document. The document captures relevance patterns but is “fake” and may contain hallucinations. Then, an unsupervised contrastively learned encoder (e.g., Contriever) encodes the document into an embedding vector. This vector identifies a neighborhood in the corpus embedding space, from which similar real documents are retrieved based on vector similarity. This second step grounds the generated document to the actual corpus, with the encoder’s dense bottleneck filtering out the hallucinations. Our experiments show that HyDE significantly outperforms the state-of-the-art unsupervised dense retriever Contriever and shows strong performance comparable to fine-tuned retrievers across various tasks (e.g. web search, QA, fact verification) and in non-English languages (e.g., sw, ko, ja, bn).","{'model': 'tldr@v2.0.0', 'text': 'The proposed Hypothetical Document Embeddings (HyDE) significantly outperforms the state-of-the-art unsupervised dense retriever Contriever and shows strong performance comparable to fine-tuned retrievers across various tasks and in non-English languages.'}",http://arxiv.org/pdf/2212.10496
-seqzero: few-shot compositional semantic parsing with sequential prompts and zero-shot models,Jingfeng Yang,"Recent research showed promising results on combining pretrained language models (LMs) with canonical utterance for few-shot semantic parsing. The canonical utterance is often lengthy and complex due to the compositional structure of formal languages. Learning to generate such canonical utterance requires significant amount of data to reach high performance. Fine-tuning with only few-shot samples, the LMs can easily forget pretrained knowledge, overfit spurious biases, and suffer from compositionally out-of-distribution generalization errors. To tackle these issues, we propose a novel few-shot semantic parsing method -- SeqZero. SeqZero decomposes the problem into a sequence of sub-problems, which correspond to the sub-clauses of the formal language. Based on the decomposition, the LMs only need to generate short answers using prompts for predicting sub-clauses. Thus, SeqZero avoids generating a long canonical utterance at once. Moreover, SeqZero employs not only a few-shot model but also a zero-shot model to alleviate the overfitting. In particular, SeqZero brings out the merits from both models via ensemble equipped with our proposed constrained rescaling. SeqZero achieves SOTA performance of BART-based models on GeoQuery and EcommerceQuery, which are two few-shot datasets with compositional data split.","{'model': 'tldr@v2.0.0', 'text': 'SeqZero achieves SOTA performance of BART-based models on GeoQuery and EcommerceQuery, which are two few-shot datasets with compositional data split and brings out the merits from both models via ensemble equipped with the proposed constrained rescaling.'}",https://arxiv.org/pdf/2205.07381
-few-shot cross-lingual stance detection with sentiment-based pre-training,Momchil Hardalov,"The goal of stance detection is to determine the viewpoint expressed in a piece of text towards a target. These viewpoints or contexts are often expressed in many different languages depending on the user and the platform, which can be a local news outlet, a social media platform, a news forum, etc. Most research on stance detection, however, has been limited to working with a single language and on a few limited targets, with little work on cross-lingual stance detection. Moreover, non-English sources of labelled data are often scarce and present additional challenges. Recently, large multilingual language models have substantially improved the performance on many non-English tasks, especially such with a limited number of examples. This highlights the importance of model pre-training and its ability to learn from few examples. In this paper, we present the most comprehensive study of cross-lingual stance detection to date: we experiment with 15 diverse datasets in 12 languages from 6 language families, and with 6 low-resource evaluation settings each. For our experiments, we build on pattern-exploiting training (PET), proposing the addition of a novel label encoder to simplify the verbalisation procedure. We further propose sentiment-based generation of stance data for pre-training, which shows sizeable improvement of more than 6% F1 absolute in few-shot learning settings compared to several strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the most comprehensive study of cross-lingual stance detection to date, and proposes sentiment-based generation of stance data for pre-training, which shows sizeable improvement of more than 6% F1 absolute in few-shot learning settings compared to several strong baselines.'}",https://ojs.aaai.org/index.php/AAAI/article/download/21318/21067
-from images to textual prompts: zero-shot visual question answering with frozen large language models,Jiaxian Guo,"Large language models (LLMs) have demonstrated excellent zero-shot generalization to new language tasks. However, effective utilization of LLMs for zero-shot visual question-answering (VQA) remains challenging, primarily due to the modality disconnect and task disconnect between the LLM and VQA tasks. End-to-end training on multimodal data may bridge the disconnects, but is inflexible and computationally expensive. To address this issue, we propose Img2LLM, a plug-and-play module that provides LLM prompts to enable LLMs to perform zeroshot VQA tasks without end-to-end training. We develop LLM-agnostic models describe image content as exemplar question-answer pairs, which prove to be effective LLM prompts. Img2LLM offers the following benefits: 1) It achieves comparable or better performance than methods relying on end-to-end training. For example, we outperform Flamingo [3] by 5.6% on VQAv2. On the challenging A-OKVQA dataset, our method outperforms few-shot methods by as much as 20%. 2) It flexibly interfaces with a wide range of LLMs to perform VQA. 3) It eliminates the need to specialize LLMs using end-to-end finetuning and serve highly specialized LLMs to end users, thereby reducing cost. Code is available via the LAVIS [28] framework at https://github.com/salesforce/LAVIS/tree/main/projects/img2llm-vqa.","{'model': 'tldr@v2.0.0', 'text': 'Img2LLM is a plug-and-play module that provides LLM prompts to enable LLMs to perform zeroshot VQA tasks without end-to-end training, and develops LLM-agnostic models that describe image content as exemplar question-answer pairs, which prove to be effective LLm prompts.'}",
-analysis of relaxation processes in a multilevel system. a many‐shot expansion technique,R. Serauskas,"The relaxation of a gas of polyatomic molecules, excited to a particular energy, is analyzed in terms of a random walk through all relevant energy states to yield the various stable products. The quantum yield is taken as the probability of forming a given one of these products. This relaxation problem is solved by means of an asymptotic, natural expansion in terms of the events in the process which are physically the most significant. The first term in this many‐shot expansion is the strong collision result. For nearly strong collision processes and nonequilibrium systems this many‐shot expansion yields a rapidly converging series for an arbitrary ordered set of transition probabilities, and hence constitutes a computationally useful solution to the Pauli master equation for such a nonequilibrium system. The relaxation problem is analyzed in terms of some assumed sets of collisional transition probabilities, including a model that assumes statistical equilibration of vibrational energy in every collision...",,
-chemical kinetics as a relaxation process in a multilevel system. ii. general formulation of the many‐shot expansion,R. Serauskas,"For a complete understanding of rates in chemical kinetics it is necessary to analyze kinetic processes on the molecular level. For such an analysis one must know the change in the distribution of internal energy as the result of a collision process. It is suggested that this information can be obtained from an analysis of weak‐collision terms in a reacting system excited to a nonequilibrium distribution from which it relaxes by way of a many‐level process.The relaxation problem in a nonequilibrium multilevel system weakly coupled to a heat bath can be expressed in terms of a Pauli master equation. It has previously been shown that this relaxation problem can be solved in terms of a many‐shot expansion, the leading term being the strong‐collision result and further terms being due to weak collisions. In this paper it is shown that this many‐shot expansion can be derived from a generalized matrix form of the master equation, as well as by the phenomenological approach used previously. Furthermore, a matrix...",,
-rethinking the role of demonstrations: what makes in-context learning work?,Sewon Min,"Large language models (LMs) are able to in-context learn—perform a new task via inference alone by conditioning on a few input-label pairs (demonstrations) and making predictions for new inputs. However, there has been little understanding of how the model learns and which aspects of the demonstrations contribute to end task performance. In this paper, we show that ground truth demonstrations are in fact not required—randomly replacing labels in the demonstrations barely hurts performance on a range of classification and multi-choce tasks, consistently over 12 different models including GPT-3. Instead, we find that other aspects of the demonstrations are the key drivers of endtask performance, including the fact that they provide a few examples of (1) the label space, (2) the distribution of the input text, and (3) the overall format of the sequence. Together, our analysis provides a new way of understanding how and why in-context learning works, while opening up new questions about how much can be learned from large language models through inference alone.","{'model': 'tldr@v2.0.0', 'text': 'This paper shows that ground truth demonstrations are in fact not required and that other aspects of the demonstrations are the key drivers of end task performance, including the fact that they provide a few examples of the label space, the distribution of the input text, and the overall format of the sequence.'}",https://aclanthology.org/2022.emnlp-main.759.pdf
-what learning algorithm is in-context learning? investigations with linear models,Ekin Akyürek,"Neural sequence models, especially transformers, exhibit a remarkable capacity for in-context learning. They can construct new predictors from sequences of labeled examples $(x, f(x))$ presented in the input without further parameter updates. We investigate the hypothesis that transformer-based in-context learners implement standard learning algorithms implicitly, by encoding smaller models in their activations, and updating these implicit models as new examples appear in the context. Using linear regression as a prototypical problem, we offer three sources of evidence for this hypothesis. First, we prove by construction that transformers can implement learning algorithms for linear models based on gradient descent and closed-form ridge regression. Second, we show that trained in-context learners closely match the predictors computed by gradient descent, ridge regression, and exact least-squares regression, transitioning between different predictors as transformer depth and dataset noise vary, and converging to Bayesian estimators for large widths and depths. Third, we present preliminary evidence that in-context learners share algorithmic features with these predictors: learners' late layers non-linearly encode weight vectors and moment matrices. These results suggest that in-context learning is understandable in algorithmic terms, and that (at least in the linear case) learners may rediscover standard estimation algorithms. Code and reference implementations are released at https://github.com/ekinakyurek/google-research/blob/master/incontext.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the hypothesis that transformer-based in-context learners implement standard learning algorithms implicitly, by encoding smaller models in their activations, and updating these implicit models as new examples appear in the context, and suggests that in- context learning is understandable in algorithmic terms, and that (at least in the linear case) learners may rediscover standard estimation algorithms.'}",http://arxiv.org/pdf/2211.15661
-data distributional properties drive emergent in-context learning in transformers,Stephanie C. Y. Chan,"Large transformer-based models are able to perform in-context few-shot learning, without being explicitly trained for it. This observation raises the question: what aspects of the training regime lead to this emergent behavior? Here, we show that this behavior is driven by the distributions of the training data itself. In-context learning emerges when the training data exhibits particular distributional properties such as burstiness (items appear in clusters rather than being uniformly distributed over time) and having large numbers of rarely occurring classes. In-context learning also emerges more strongly when item meanings or interpretations are dynamic rather than fixed. These properties are exemplified by natural language, but are also inherent to naturalistic data in a wide range of other domains. They also depart significantly from the uniform, i.i.d. training distributions typically used for standard supervised learning. In our initial experiments, we found that in-context learning traded off against more conventional weight-based learning, and models were unable to achieve both simultaneously. However, our later experiments uncovered that the two modes of learning could co-exist in a single model when it was trained on data following a skewed Zipfian distribution -- another common property of naturalistic data, including language. In further experiments, we found that naturalistic data distributions were only able to elicit in-context learning in transformers, and not in recurrent models. In sum, our findings indicate how the transformer architecture works together with particular properties of the training data to drive the intriguing emergent in-context learning behaviour of large language models, and how future work might encourage both in-context and in-weights learning in domains beyond language.","{'model': 'tldr@v2.0.0', 'text': 'These findings indicate how the transformer architecture works together with particular properties of the training data to drive the intriguing emergent in- context learning behaviour of large language models, and how future work might encourage both in-context and in-weights learning in domains beyond language.'}",https://arxiv.org/pdf/2205.05055
-a survey for in-context learning,Qingxiu Dong,"With the increasing ability of large language models (LLMs), in-context learning (ICL) has become a new paradigm for natural language processing (NLP), where LLMs make predictions only based on contexts augmented with a few training examples. It has been a new trend exploring ICL to evaluate and extrapolate the ability of LLMs. In this paper, we aim to survey and summarize the progress, challenges, and future work in ICL. We first present a formal definition of ICL and clarify its correlation to related studies. Then, we organize and discuss advanced techniques of ICL, including training strategies, prompting strategies, and so on. Finally, we present the challenges of ICL and provide potential directions for further research. We hope our work can encourage more research on uncovering how ICL works and improving ICL in future work. 1","{'model': 'tldr@v2.0.0', 'text': 'The progress, challenges, and future work in ICL are summarized and a formal definition of ICL is presented and its correlation to related studies are clarified and potential directions for further research are provided.'}",http://arxiv.org/pdf/2301.00234
-transformers as algorithms: generalization and stability in in-context learning,Yingcong Li,"In-context learning (ICL) is a type of prompting where a transformer model operates on a sequence of (input, output) examples and performs inference on-the-fly. In this work, we formalize in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time. We first explore the statistical aspects of this abstraction through the lens of multitask learning: We obtain generalization bounds for ICL when the input prompt is (1) a sequence of i.i.d. (input, label) pairs or (2) a trajectory arising from a dynamical system. The crux of our analysis is relating the excess risk to the stability of the algorithm implemented by the transformer. We characterize when transformer/attention architecture provably obeys the stability condition and also provide empirical verification. For generalization on unseen tasks, we identify an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner. Finally, we provide numerical evaluations that (1) demonstrate transformers can indeed implement near-optimal algorithms on classical regression problems with i.i.d. and dynamic data, (2) provide insights on stability, and (3) verify our theoretical predictions.","{'model': 'tldr@v2.0.0', 'text': 'This work formalizes in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time and identifies an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner.'}",
-differentially private in-context learning,Ashwinee Panda,An important question in deploying large language models (LLMs) is how to augment LLMs with private data. We propose Differentially Private In-context Learning (DP-ICL) to enable LLMs to adapt to new tasks while maintaining privacy guarantees. DP-ICL performs private inference by establishing a noisy consensus over an ensemble of exemplars using the Report-Noisy-Max mechanism. We evaluate DP-ICL on four benchmarks and find that it achieves comparable performance (< 2% degradation) with non-private ICL.,"{'model': 'tldr@v2.0.0', 'text': 'Differentially Private In-context Learning (DP-ICL) is proposed to enable LLMs to adapt to new tasks while maintaining privacy guarantees, and achieves comparable performance (< 2% degradation) with non-private ICL.'}",https://arxiv.org/pdf/2305.01639
-in-context reinforcement learning with algorithm distillation,M. Laskin,"We propose Algorithm Distillation (AD), a method for distilling reinforcement learning (RL) algorithms into neural networks by modeling their training histories with a causal sequence model. Algorithm Distillation treats learning to reinforcement learn as an across-episode sequential prediction problem. A dataset of learning histories is generated by a source RL algorithm, and then a causal transformer is trained by autoregressively predicting actions given their preceding learning histories as context. Unlike sequential policy prediction architectures that distill post-learning or expert sequences, AD is able to improve its policy entirely in-context without updating its network parameters. We demonstrate that AD can reinforcement learn in-context in a variety of environments with sparse rewards, combinatorial task structure, and pixel-based observations, and find that AD learns a more data-efficient RL algorithm than the one that generated the source data.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that AD can reinforcement learn in-context in a variety of environments with sparse rewards, combinatorial task structure, and pixel-based observations, and it is found that AD learns a more data-efficient RL algorithm than the one that generated the source data.'}",http://arxiv.org/pdf/2210.14215
-context encoders: feature learning by inpainting,Deepak Pathak,"We present an unsupervised visual feature learning algorithm driven by context-based pixel prediction. By analogy with auto-encoders, we propose Context Encoders - a convolutional neural network trained to generate the contents of an arbitrary image region conditioned on its surroundings. In order to succeed at this task, context encoders need to both understand the content of the entire image, as well as produce a plausible hypothesis for the missing part(s). When training context encoders, we have experimented with both a standard pixel-wise reconstruction loss, as well as a reconstruction plus an adversarial loss. The latter produces much sharper results because it can better handle multiple modes in the output. We found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures. We quantitatively demonstrate the effectiveness of our learned features for CNN pre-training on classification, detection, and segmentation tasks. Furthermore, context encoders can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.","{'model': 'tldr@v2.0.0', 'text': 'It is found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures, and can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.'}",https://arxiv.org/pdf/1604.07379
-context autoencoder for self-supervised representation learning,Xiaokang Chen,,"{'model': 'tldr@v2.0.0', 'text': 'The CAE design encourages the separation of learning the encoder (representation) from completing the pertaining tasks: masked representation prediction and masked patch reconstruction tasks, and making predictions in the encoded representation space empirically shows the benefit to representation learning.'}",https://arxiv.org/pdf/2202.03026
-fast visual tracking via dense spatio-temporal context learning,Kaihua Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'A novel explicit scale adaptation scheme is proposed, able to deal with target scale variations efficiently and effectively, and the Fast Fourier Transform is adopted for fast learning and detection in this work, which only needs 4 FFT operations.'}",http://www4.comp.polyu.edu.hk/~cslzhang/paper/conf/STC_eccv14.pdf
-can you unpack that? learning to rewrite questions-in-context,Ahmed Elgohary,"Question answering is an AI-complete problem, but existing datasets lack key elements of language understanding such as coreference and ellipsis resolution. We consider sequential question answering: multiple questions are asked one-by-one in a conversation between a questioner and an answerer. Answering these questions is only possible through understanding the conversation history. We introduce the task of question-in-context rewriting: given the context of a conversation’s history, rewrite a context-dependent into a self-contained question with the same answer. We construct, CANARD, a dataset of 40,527 questions based on QuAC (Choi et al., 2018) and train Seq2Seq models for incorporating context into standalone questions.","{'model': 'tldr@v2.0.0', 'text': 'This work constructs, CANARD, a dataset of 40,527 questions based on QuAC and trains Seq2Seq models for incorporating context into standalone questions and introduces the task of question-in-context rewriting.'}",https://www.aclweb.org/anthology/D19-1605.pdf
-what can transformers learn in-context? a case study of simple function classes,Shivam Garg,"In-context learning refers to the ability of a model to condition on a prompt sequence consisting of in-context examples (input-output pairs corresponding to some task) along with a new query input, and generate the corresponding output. Crucially, in-context learning happens only at inference time without any parameter updates to the model. While large language models such as GPT-3 exhibit some ability to perform in-context learning, it is unclear what the relationship is between tasks on which this succeeds and what is present in the training data. To make progress towards understanding in-context learning, we consider the well-defined problem of training a model to in-context learn a function class (e.g., linear functions): that is, given data derived from some functions in the class, can we train a model to in-context learn""most""functions from this class? We show empirically that standard Transformers can be trained from scratch to perform in-context learning of linear functions -- that is, the trained model is able to learn unseen linear functions from in-context examples with performance comparable to the optimal least squares estimator. In fact, in-context learning is possible even under two forms of distribution shift: (i) between the training data of the model and inference-time prompts, and (ii) between the in-context examples and the query input during inference. We also show that we can train Transformers to in-context learn more complex function classes -- namely sparse linear functions, two-layer neural networks, and decision trees -- with performance that matches or exceeds task-specific learning algorithms. Our code and models are available at https://github.com/dtsip/in-context-learning .","{'model': 'tldr@v2.0.0', 'text': 'It is shown empirically that standard Transformers can be trained from scratch to perform in-context learning of linear functions -- that is, the trained model is able to learn unseen linear functions from in- context examples with performance comparable to the optimal least squares estimator.'}",https://arxiv.org/pdf/2208.01066
-a survey on context learning,Guangxu Xun,"Learning semantics based on context information has been researched in many research areas for decades. Context information can not only be directly used as the input data, but also sometimes used as auxiliary knowledge to improve existing models. This survey aims at providing a structured and comprehensive overview of the research on context learning. We summarize and group the existing literature into four categories, Explicit Analysis, Implicit Analysis, Neural Network Models, and Composite Models, based on the underlying techniques adopted by them. For each category, we talk about the basic idea and techniques, and also introduce how context information is utilized as the model input or incorporated into the model to enhance the performance or extend the domain of application as auxiliary knowledge. In addition, we discuss the advantages and disadvantages of each model from both the technical and practical point of view.","{'model': 'tldr@v2.0.0', 'text': 'This survey aims at providing a structured and comprehensive overview of the research on context learning by summarized and group the existing literature into four categories, Explicit Analysis, Implicit Analysis, Neural Network Models, and Composite Models, based on the underlying techniques adopted by them.'}",
-cross-modal retrieval in the cooking context: learning semantic text-image embeddings,Micael Carvalho,"Designing powerful tools that support cooking activities has rapidly gained popularity due to the massive amounts of available data, as well as recent advances in machine learning that are capable of analyzing them. In this paper, we propose a cross-modal retrieval model aligning visual and textual data (like pictures of dishes and their recipes) in a shared representation space. We describe an effective learning scheme, capable of tackling large-scale problems, and validate it on the Recipe1M dataset containing nearly 1 million picture-recipe pairs. We show the effectiveness of our approach regarding previous state-of-the-art models and present qualitative results over computational cooking use cases.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a cross-modal retrieval model aligning visual and textual data (like pictures of dishes and their recipes) in a shared representation space, and describes an effective learning scheme, capable of tackling large-scale problems.'}",https://hal.archives-ouvertes.fr/hal-01839068/file/1804.11146.pdf
-what makes good in-context examples for gpt-3?,Jiachang Liu,"GPT-3 has attracted lots of attention due to its superior performance across a wide range of NLP tasks, especially with its in-context learning abilities. Despite its success, we found that the empirical results of GPT-3 depend heavily on the choice of in-context examples. In this work, we investigate whether there are more effective strategies for judiciously selecting in-context examples (relative to random sampling) that better leverage GPT-3’s in-context learning capabilities.Inspired by the recent success of leveraging a retrieval module to augment neural networks, we propose to retrieve examples that are semantically-similar to a test query sample to formulate its corresponding prompt. Intuitively, the examples selected with such a strategy may serve as more informative inputs to unleash GPT-3’s power of text generation. We evaluate the proposed approach on several natural language understanding and generation benchmarks, where the retrieval-based prompt selection approach consistently outperforms the random selection baseline. Moreover, it is observed that the sentence encoders fine-tuned on task-related datasets yield even more helpful retrieval results. Notably, significant gains are observed on tasks such as table-to-text generation (44.3% on the ToTTo dataset) and open-domain question answering (45.5% on the NQ dataset).","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to retrieve examples that are semantically-similar to a test query sample to formulate its corresponding prompt, and evaluates the proposed approach on several natural language understanding and generation benchmarks, where the retrieval-based prompt selection approach consistently outperforms the random selection baseline.'}",https://aclanthology.org/2022.deelio-1.10.pdf
-"the neuroscience of placebo effects: connecting context, learning and health",T. Wager,,"{'model': 'tldr@v2.0.0', 'text': 'An empirical review of the brain systems that are involved in placebo effects and a conceptual framework linking these findings to the mind–brain processes that mediate them suggest that the neuropsychological processes thatMediate placebo effects may be crucial for a wide array of therapeutic approaches, including many drugs.'}",https://europepmc.org/articles/pmc6013051?pdf=render
-adult learning theories in context: a quick guide for healthcare professional educators,B. Mukhalalati,"Background: Adult learning theories play a pivotal role in the design and implementation of education programs, including healthcare professional programs. There is a variation in the use of theories in healthcare professional education programs and this is may be in part due to a lack of understanding of the range of learning theories available and paucity of specific, in-context examples, to help educators in considering alternative theories relevant to their teaching setting. This article seeks to synthesize key learning theories applicable in the learning and teaching of healthcare professionals and to provide examples of their use in context. Method and results: A literature review was conducted in 2015 and 2016 using PubMed, Scopus, Web of Science, and ERIC academic databases. Search terms used identified a range of relevant literature about learning theories, and their utilization in different healthcare professional education programs. The findings were synthesized and presented in a table format, illustrating the learning theory, specific examples from health and medical education, and a very brief critique of the theory. Outcome: The literature synthesis provides a quick and easy-to-use summary of key theories and examples of their use to help healthcare professional educators access a wider range of learning theories to inform their instructional strategies, learning objectives, and evaluation approaches. This will ultimately result in educational program enhancement and improvement in student learning experiences.","{'model': 'tldr@v2.0.0', 'text': 'The literature synthesis provides a quick and easy-to-use summary of key theories and examples of their use to help healthcare professional educators access a wider range of learning theories to inform their instructional strategies, learning objectives, and evaluation approaches.'}",https://journals.sagepub.com/doi/pdf/10.1177/2382120519840332
-professional development for cultural diversity: the challenges of teacher learning in context,Nikolett Szelei,"ABSTRACT This article describes context-based professional development (PD) for cultural diversity in a Portuguese school cluster, and discusses how it supports change for justice and equity. Teachers felt the importance of PD and showed willingness and interest to learn. Several teacher learning opportunities were mapped out such as formal workshops, starting small collaborations and teachers’ self-directed informal learning activities. Yet, a rather fragmented character of PD seemed to emerge in terms of content on cultural diversity and forms of learning. Conflicting agendas, scattered teacher collaboration and commitment, and little student and community involvement in planned PD were found. Furthermore, there seemed to be tensions between current PD and teachers’ needs and circumstances; teachers wished for more specific information and pedagogical solutions, more collaboration and more organisational support in PD. Applying a critical multicultural perspective, it is discussed that although the current constellation of PD is a potential start, it might still contribute to teachers’ conceptual confusion and pedagogical insecurities on the field of cultural diversity. It is suggested that criticality towards PD frames is needed to re-centre cultural diversity on the premises of justice, as well as teacher support, and conscious learning with and from students, families and communities.",,https://www.tandfonline.com/doi/pdf/10.1080/19415257.2019.1642233?needAccess=true
-english in context: learning materials,P. L. McEldowney,,,
-efficient off-policy meta-reinforcement learning via probabilistic context variables,Kate Rakelly,"Deep reinforcement learning algorithms require large amounts of experience to learn an individual task. While in principle meta-reinforcement learning (meta-RL) algorithms enable agents to learn new skills from small amounts of experience, several major challenges preclude their practicality. Current methods rely heavily on on-policy experience, limiting their sample efficiency. The also lack mechanisms to reason about task uncertainty when adapting to new tasks, limiting their effectiveness in sparse reward problems. In this paper, we address these challenges by developing an off-policy meta-RL algorithm that disentangles task inference and control. In our approach, we perform online probabilistic filtering of latent task variables to infer how to solve a new task from small amounts of experience. This probabilistic interpretation enables posterior sampling for structured and efficient exploration. We demonstrate how to integrate these task variables with off-policy RL algorithms to achieve both meta-training and adaptation efficiency. Our method outperforms prior algorithms in sample efficiency by 20-100X as well as in asymptotic performance on several meta-RL benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'This paper develops an off-policy meta-RL algorithm that disentangles task inference and control and performs online probabilistic filtering of latent task variables to infer how to solve a new task from small amounts of experience.'}",
-rural online learning in the context of covid 19 in south africa: evoking an inclusive education approach,B. Dube,"This paper discusses the challenges faced by rural learners in South Africa in the context of the world pandemic commonly known as COVID-19. Rural learners face unprecedented challenges in adjusting to a new mode of life and learning, the latter being characterised by the predominant use of online, learning management systems and low-tech applications. The paper is informed by critical emancipatory research, I used participatory action research. A total of 10 learners and five teachers participated via Whatsapp. The paper answers two questions: what are the learning challenges faced by rural learners in South Africa, and how can online learning be enhanced in the context of COVID-19? The findings suggest that, while the South African government is promoting online learning as the only alternative in the context of COVID-19, this mode excludes many rural learners from teaching and learning, due to a lack of resources to connect to the internet, the learning management system, and low-tech software. The paper argues that rural learners are critical stakeholders in education and in the fight against COVID-19, and they cannot be left behind in efforts to fight the pandemic. ","{'model': 'tldr@v2.0.0', 'text': 'The findings suggest that, while the South African government is promoting online learning as the only alternative in the context of COVID-19, this mode excludes many rural learners from teaching and learning, due to a lack of resources to connect to the internet, the learning management system and low-tech software.'}",
-language-agnostic representation learning of source code from structure and context,D. Zugner,"Source code (Context) and its parsed abstract syntax tree (AST; Structure) are two complementary representations of the same computer program. Traditionally, designers of machine learning models have relied predominantly either on Structure or Context. We propose a new model, which jointly learns on Context and Structure of source code. In contrast to previous approaches, our model uses only language-agnostic features, i.e., source code and features that can be computed directly from the AST. Besides obtaining state-of-the-art on monolingual code summarization on all five programming languages considered in this work, we propose the first multilingual code summarization model. We show that jointly training on non-parallel data from multiple programming languages improves results on all individual languages, where the strongest gains are on low-resource languages. Remarkably, multilingual training only from Context does not lead to the same improvements, highlighting the benefits of combining Structure and Context for representation learning on code.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new model, which jointly learns on Context and Structure of source code, and shows that jointly training on non-parallel data from multiple programming languages improves results on all individual languages, where the strongest gains are on low-resource languages.'}",
-multi-task reinforcement learning with context-based representations,Shagun Sodhani,"The benefit of multi-task learning over single-task learning relies on the ability to use relations across tasks to improve performance on any single task. While sharing representations is an important mechanism to share information across tasks, its success depends on how well the structure underlying the tasks is captured. In some real-world situations, we have access to metadata, or additional information about a task, that may not provide any new insight in the context of a single task setup alone but inform relations across multiple tasks. While this metadata can be useful for improving multi-task learning performance, effectively incorporating it can be an additional challenge. We posit that an efficient approach to knowledge transfer is through the use of multiple context-dependent, composable representations shared across a family of tasks. In this framework, metadata can help to learn interpretable representations and provide the context to inform which representations to compose and how to compose them. We use the proposed approach to obtain state-of-the-art results in Meta-World, a challenging multi-task benchmark consisting of 50 distinct robotic manipulation tasks.","{'model': 'tldr@v2.0.0', 'text': 'It is proposed that an efficient approach to knowledge transfer is through the use of multiple context-dependent, composable representations shared across a family of tasks, and metadata can help to learn interpretable representations and provide the context to inform which representations to compose and how to compose them.'}",
-learning pyramid-context encoder network for high-quality image inpainting,Yanhong Zeng,"High-quality image inpainting requires filling missing regions in a damaged image with plausible content. Existing works either fill the regions by copying high-resolution patches or generating semantically-coherent patches from region context, while neglecting the fact that both visual and semantic plausibility are highly-demanded. In this paper, we propose a Pyramid-context Encoder Network (denoted as PEN-Net) for image inpainting by deep generative models. The proposed PEN-Net is built upon a U-Net structure with three tailored components, ie., a pyramid-context encoder, a multi-scale decoder, and an adversarial training loss. First, we adopt a U-Net as backbone which can encode the context of an image from high-resolution pixels into high-level semantic features, and decode the features reversely. Second, we propose a pyramid-context encoder, which progressively learns region affinity by attention from a high-level semantic feature map, and transfers the learned attention to its adjacent high-resolution feature map. As the missing content can be filled by attention transfer from deep to shallow in a pyramid fashion, both visual and semantic coherence for image inpainting can be ensured. Third, we further propose a multi-scale decoder with deeply-supervised pyramid losses and an adversarial loss. Such a design not only results in fast convergence in training, but more realistic results in testing. Extensive experiments on a broad range of datasets shows the superior performance of the proposed network.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Pyramid-context Encoder Network for image inpainting by deep generative models, built upon a U-Net structure with three tailored components, ie.'}",https://arxiv.org/pdf/1904.07475
-neuromorphic context-dependent learning framework with fault-tolerant spike routing,Shuangming Yang,"Neuromorphic computing is a promising technology that realizes computation based on event-based spiking neural networks (SNNs). However, fault-tolerant on-chip learning remains a challenge in neuromorphic systems. This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework. We show how this system can learn associations between stimulation and response in two context-dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes. Furthermore, we demonstrate how our novel fault-tolerant neuromorphic spike routing scheme can avoid multiple fault nodes successfully and can enhance the maximum throughput of the neuromorphic network by 0.9%–16.1% in comparison with previous studies. By utilizing the real-time computational capabilities and multiple-fault-tolerant property of the proposed system, the neuronal mechanisms underlying the spiking activities of neuromorphic networks can be readily explored. In addition, the proposed system can be applied in real-time learning and decision-making applications, brain–machine integration, and the investigation of brain cognition during learning.","{'model': 'tldr@v2.0.0', 'text': 'This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework, and shows how this system can learn associations between stimulation and response in two context- dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes.'}",https://researchonline.jcu.edu.au/68691/7/JCU_68691_AAM.pdf
-fast tracking via spatio-temporal context learning,Kaihua Zhang,"In this paper, we present a simple yet fast and robust algorithm which exploits the spatio-temporal context for visual tracking. Our approach formulates the spatio-temporal relationships between the object of interest and its local context based on a Bayesian framework, which models the statistical correlation between the low-level features (i.e., image intensity and position) from the target and its surrounding regions. The tracking problem is posed by computing a confidence map, and obtaining the best target location by maximizing an object location likelihood function. The Fast Fourier Transform is adopted for fast learning and detection in this work. Implemented in MATLAB without code optimization, the proposed tracker runs at 350 frames per second on an i7 machine. Extensive experimental results show that the proposed algorithm performs favorably against state-of-the-art methods in terms of efficiency, accuracy and robustness.","{'model': 'tldr@v2.0.0', 'text': 'This approach formulates the spatio-temporal relationships between the object of interest and its local context based on a Bayesian framework, which models the statistical correlation between the low-level features from the target and its surrounding regions.'}",
-drivers of human development: how relationships and context shape learning and development1,D. Osher,"ABSTRACT This article synthesizes knowledge on the role of relationships and key macroand micro-contexts - poverty, racism, families, communities, schools, and peers - in supporting and/or undermining the healthy development of children and youth, using a relational developmental systems framework. Relationships with parents, siblings, peers, caregivers, and teachers are explored in the context of early care and childhood settings, schools, classrooms, and school-based interventions. Additional contextual factors include; chronic stress, institutionalized racism, stereotype threat, and racial identity. A companion article focuses on how the human brain develops, and the major constructs that define human development, the constructive nature of development, and the opportunities for resilience. Human development occurs through reciprocal coactions between the individual and their contexts and culture, with relationships as the key drivers. Relationships and contexts, along with how children appraise and interpret them, can be risks and assets for healthy learning and development, and their influence can be seen across generations and can produce intra- as well as intergenerational assets and risks. This knowledge about the individual’s responsiveness to context and experience has both positive and negative implications across early childhood, adolescence and into adulthood. Sensitive periods for brain growth and development are considered within the contextual factors that influence development including; parental responsiveness and attunement, intentional skill development, mindfulness, reciprocal interactions, adversity, trauma, and enriching opportunities. The accumulated knowledge on human development and the power of context and culture can inform child-serving systems that support positive adaptations, resilience, learning, health, and well-being.",,https://www.tandfonline.com/doi/pdf/10.1080/10888691.2017.1398650?needAccess=true
-unsupervised visual representation learning by context prediction,Carl Doersch,"This work explores the use of spatial context as a source of free and plentiful supervisory signal for training a rich visual representation. Given only a large, unlabeled image collection, we extract random pairs of patches from each image and train a convolutional neural net to predict the position of the second patch relative to the first. We argue that doing well on this task requires the model to learn to recognize objects and their parts. We demonstrate that the feature representation learned using this within-image context indeed captures visual similarity across images. For example, this representation allows us to perform unsupervised visual discovery of objects like cats, people, and even birds from the Pascal VOC 2011 detection dataset. Furthermore, we show that the learned ConvNet can be used in the R-CNN framework [19] and provides a significant boost over a randomly-initialized ConvNet, resulting in state-of-the-art performance among algorithms which use only Pascal-provided training set annotations.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the feature representation learned using this within-image context indeed captures visual similarity across images and allows us to perform unsupervised visual discovery of objects like cats, people, and even birds from the Pascal VOC 2011 detection dataset.'}",http://arxiv.org/pdf/1505.05192
-multiscale graph sample and aggregate network with context-aware learning for hyperspectral image classification,Yao Ding,"Recently, graph convolutional network (GCN) has achieved promising results in hyperspectral image (HSI) classification. However, GCN is a transductive learning method, which is difficult to aggregate the new node. Besides, the existing GCN-based methods divide graph construction and graph classification into two stages ignoring the influence of constructed graph error on classification results. Moreover, the available GCN-based methods fail to understand the global and contextual information of the graph. In this article, we propose a novel multiscale graph sample and aggregate network with a context-aware learning method for HSI classification. The proposed network adopts a multiscale graph sample and aggregate network (graphSAGE) to learn the multiscale features from the local regions graph, which improves the diversity of network input information and effectively solves the impact of original input graph errors on classification. By employing a context-aware mechanism to characterize the importance among spatially neighboring regions, deep contextual and global information of the graph can be learned automatically by focusing on important spatial targets. Meanwhile, the graph structure is reconstructed automatically based on the classified objects as network training, which is able to effectively reduce the influence of the initial graph error on the classification result. Extensive experiments are conducted on three real HSI datasets, which are demonstrated to outperform the compared state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel multiscale graph sample and aggregate network with a context-aware learning method for HSI classification that improves the diversity of network input information and effectively solves the impact of original input graph errors on classification.'}",https://ieeexplore.ieee.org/ielx7/4609443/9314330/09411656.pdf
-end-to-end learning of geometry and context for deep stereo regression,Alex Kendall,"We propose a novel deep learning architecture for regressing disparity from a rectified pair of stereo images. We leverage knowledge of the problem’s geometry to form a cost volume using deep feature representations. We learn to incorporate contextual information using 3-D convolutions over this volume. Disparity values are regressed from the cost volume using a proposed differentiable soft argmin operation, which allows us to train our method end-to-end to sub-pixel accuracy without any additional post-processing or regularization. We evaluate our method on the Scene Flow and KITTI datasets and on KITTI we set a new stateof-the-art benchmark, while being significantly faster than competing approaches.","{'model': 'tldr@v2.0.0', 'text': 'A novel deep learning architecture for regressing disparity from a rectified pair of stereo images is proposed, leveraging knowledge of the problem’s geometry to form a cost volume using deep feature representations and incorporating contextual information using 3-D convolutions over this volume.'}",http://arxiv.org/pdf/1703.04309
-"interprofessional education: a review of context, learning and the research agenda",J. Thistlethwaite,Medical Education 2012: 46: 58–70,,
-how e-learning environmental stimuli influence determinates of learning engagement in the context of covid-19? sor model perspective,Junhui Yang,"The COVID-19 pandemic at the beginning of 2020 has changed the conventional learning mode for most students at schools all over the world, and the e-learning at home has become a new trend. Taking Chinese college students as the research subject and drawing on the stimulus–organism–response (S-O-R) model, this paper examines the relationship between the peer referent, perceived closeness, and perceived control and the learning engagement. Using data from 377 college students who have used e-learning, this study shows that perceived closeness, perceived control, and peer referents in e-learning have a positive effect on the self-efficacy and well-being of students, thus improving students’ enthusiasm for learning. Our intent is to assist researchers, instructors, designers, and others in identifying effective methods to conceptualize and measure student engagement in e-learning.","{'model': 'tldr@v2.0.0', 'text': 'This study shows that perceived closeness, perceived control, and peer referents in e-learning have a positive effect on the self-efficacy and well-being of students, thus improving students’ enthusiasm for learning.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2021.584976/pdf
-involvement of dopamine d1/d5 and d2 receptors in context-dependent extinction learning and memory reinstatement,M. André,"Dopamine contributes to the regulation of higher order information processing and executive control. It is important for memory consolidation processes, and for the adaptation of learned responses based on experience. In line with this, under aversive learning conditions, application of dopamine receptor antagonists prior to extinction result in enhanced memory reinstatement. Here, we investigated the contribution of the dopaminergic system to extinction and memory reinstatement (renewal) of an appetitive spatial learning task in rodents. Rats were trained for 3 days in a T-maze (context “A”) to associate a goal arm with a food reward, despite low reward probability (acquisition phase). On day 4, extinction learning (unrewarded) occurred, that was reinforced by a context change (“B”). On day 5, re-exposure to the (unrewarded) “A” context took place (renewal of context “A”, followed by extinction of context “A”). In control animals, significant extinction occurred on day 4, that was followed by an initial memory reinstatement (renewal) on day 5, that was, in turn, succeeded by extinction of renewal. Intracerebral treatment with a D1/D5-receptor antagonist prior to the extinction trials, elicited a potent enhancement of extinction in context “B”. By contrast, a D1/D5-agonist impaired renewal in context “A”. Extinction in the “A” context on day 5 was unaffected by the D1/D5-ligands. Treatment with a D2-receptor antagonist prior to extinction had no overall effect on extinction in context “B” or renewal in context “A”, although extinction of the renewal effect was impaired on day 5, compared to controls. Taken together, these data suggest that dopamine acting on the D1/D5-receptor modulates both acquisition and consolidation of context-dependent extinction. By contrast, the D2-receptor may contribute to context-independent aspects of this kind of extinction learning.","{'model': 'tldr@v2.0.0', 'text': 'The data suggest that dopamine acting on the D1/D5-receptor modulates both acquisition and consolidation of context-dependent extinction, and that the D2- receptor may contribute to context-independent aspects of this kind of extinction learning.'}",https://www.frontiersin.org/articles/10.3389/fnbeh.2015.00372/pdf
-setting learning analytics in context: overcoming the barriers to large-scale adoption,Rebecca Ferguson,"Once learning analytics have been successfully developed and tested, the next step is to implement them at a larger scale -- across a faculty, an institution or an educational system. This introduces a new set of challenges, because education is a stable system, resistant to change. Implementing learning analytics at scale involves working with the entire technological complex that exists around technology-enhanced learning (TEL). This includes the different groups of people involved -- learners, educators, administrators and support staff -- the practices of those groups, their understandings of how teaching and learning take place, the technologies they use and the specific environments within which they operate. Each element of the TEL Complex requires explicit and careful consideration during the process of implementation, in order to avoid failure and maximise the chances of success. In order for learning analytics to be implemented successfully at scale, it is crucial to provide not only the analytics and their associated tools but also appropriate forms of support, training and community building.","{'model': 'tldr@v2.0.0', 'text': 'In order for learning analytics to be implemented successfully at scale, it is crucial to provide not only the analytics and their associated tools but also appropriate forms of support, training and community building.'}",https://learning-analytics.info/index.php/JLA/article/download/4077/4421
-self-supervised learning for medical image analysis using image context restoration,Liang Chen,,"{'model': 'tldr@v2.0.0', 'text': 'A novel self-supervised learning strategy based on context restoration is proposed in order to better exploit unlabelled images and is validated in three common problems in medical imaging: classification, localization, and segmentation.'}",
-learning from context or names? an empirical study on neural relation extraction,Hao Peng,"Neural models have achieved remarkable success on relation extraction (RE) benchmarks. However, there is no clear understanding which type of information affects existing RE models to make decisions and how to further improve the performance of these models. To this end, we empirically study the effect of two main information sources in text: textual context and entity mentions (names). We find that (i) while context is the main source to support the predictions, RE models also heavily rely on the information from entity mentions, most of which is type information, and (ii) existing datasets may leak shallow heuristics via entity mentions and thus contribute to the high performance on RE benchmarks. Based on the analyses, we propose an entity-masked contrastive pre-training framework for RE to gain a deeper understanding on both textual context and type information while avoiding rote memorization of entities or use of superficial cues in mentions. We carry out extensive experiments to support our views, and show that our framework can improve the effectiveness and robustness of neural models in different RE scenarios. All the code and datasets are released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'An entity-masked contrastive pre-training framework for RE is proposed to gain a deeper understanding on both textual context and type information while avoiding rote memorization of entities or use of superficial cues in mentions.'}",https://arxiv.org/pdf/2010.01923
-machine learning on knowledge graphs for context-aware security monitoring,J. Garrido,"Machine learning techniques are gaining attention in the context of intrusion detection due to the increasing amounts of data generated by monitoring tools, as well as the sophistication displayed by attackers in hiding their activity. However, existing methods often exhibit important limitations in terms of the quantity and relevance of the generated alerts. Recently, knowledge graphs are finding application in the cybersecurity domain, showing the potential to alleviate some of these drawbacks thanks to their ability to seamlessly integrate data from multiple domains using human-understandable vocabularies. We discuss the application of machine learning on knowledge graphs for intrusion detection and experimentally evaluate a link-prediction method for scoring anomalous activity in industrial systems. After initial unsupervised training, the proposed method is shown to produce intuitively well-calibrated and interpretable alerts in a diverse range of scenarios, hinting at the potential benefits of relational machine learning on knowledge graphs for intrusion detection purposes.","{'model': 'tldr@v2.0.0', 'text': 'After initial unsupervised training, the proposed method is shown to produce intuitively well-calibrated and interpretable alerts in a diverse range of scenarios, hinting at the potential benefits of relational machine learning on knowledge graphs for intrusion detection purposes.'}",https://arxiv.org/pdf/2105.08741
-"beyond transmitting bits: context, semantics, and task-oriented communications",Deniz Gündüz,"Communication systems to date primarily aim at reliably communicating bit sequences. Such an approach provides efficient engineering designs that are agnostic to the meanings of the messages or to the goal that the message exchange aims to achieve. Next generation systems, however, can be potentially enriched by folding message semantics and goals of communication into their design. Further, these systems can be made cognizant of the context in which communication exchange takes place, thereby providing avenues for novel design insights. This tutorial summarizes the efforts to date, starting from its early adaptations, semantic-aware and task-oriented communications, covering the foundations, algorithms and potential implementations. The focus is on approaches that utilize information theory to provide the foundations, as well as the significant role of learning in semantics and task-aware communications.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial summarizes the efforts to date, starting from its early adaptations, semantic-aware and task-oriented communications, covering the foundations, algorithms and potential implementations, with a focus on approaches that utilize information theory to provide the foundations.'}",https://arxiv.org/pdf/2207.09353
-fast context adaptation via meta-learning,L. Zintgraf,"We propose CAVIA, a meta-learning method for fast adaptation that is scalable, flexible, and easy to implement. CAVIA partitions the model parameters into two parts: context parameters that serve as additional input to the model and are adapted on individual tasks, and shared parameters that are meta-trained and shared across tasks. At test time, the context parameters are updated with one or several gradient steps on a task-specific loss that is backpropagated through the shared part of the network. Compared to approaches that adjust all parameters on a new task (e.g., MAML), CAVIA can be scaled up to larger networks without overfitting on a single task, is easier to implement, and is more robust to the inner-loop learning rate. We show empirically that CAVIA outperforms MAML on regression, classification, and reinforcement learning problems.","{'model': 'tldr@v2.0.0', 'text': 'It is shown empirically that CAVIA outperforms MAML on regression, classification, and reinforcement learning problems and is easier to implement, and is more robust to the inner-loop learning rate.'}",
-context-aware dynamics model for generalization in model-based reinforcement learning,Kimin Lee,"Model-based reinforcement learning (RL) enjoys several benefits, such as data-efficiency and planning, by learning a model of the environment's dynamics. However, learning a global model that can generalize across different dynamics is a challenging task. To tackle this problem, we decompose the task of learning a global dynamics model into two stages: (a) learning a context latent vector that captures the local dynamics, then (b) predicting the next state conditioned on it. In order to encode dynamics-specific information into the context latent vector, we introduce a novel loss function that encourages the context latent vector to be useful for predicting both forward and backward dynamics. The proposed method achieves superior generalization ability across various simulated robotics and control tasks, compared to existing RL schemes.","{'model': 'tldr@v2.0.0', 'text': 'This work decomposes the task of learning a global dynamics model into two stages: learning a context latent vector that captures the local dynamics, then predicting the next state conditioned on it, and introduces a novel loss function that encourages the context latentvector to be useful for predicting both forward and backward dynamics.'}",
-"context-aware computing, learning, and big data in internet of things: a survey",Omer Berat Sezer,"Internet of Things (IoT) has been growing rapidly due to recent advancements in communications and sensor technologies. Meanwhile, with this revolutionary transformation, researchers, implementers, deployers, and users are faced with many challenges. IoT is a complicated, crowded, and complex field; there are various types of devices, protocols, communication channels, architectures, middleware, and more. Standardization efforts are plenty, and this chaos will continue for quite some time. What is clear, on the other hand, is that IoT deployments are increasing with accelerating speed, and this trend will not stop in the near future. As the field grows in numbers and heterogeneity, “intelligence” becomes a focal point in IoT. Since data now becomes “big data,” understanding, learning, and reasoning with big data is paramount for the future success of IoT. One of the major problems in the path to intelligent IoT is understanding “context,” or making sense of the environment, situation, or status using data from sensors, and then acting accordingly in autonomous ways. This is called “context-aware computing,” and it now requires both sensing and, increasingly, learning, as IoT systems get more data and better learning from this big data. In this survey, we review the field, first, from a historical perspective, covering ubiquitous and pervasive computing, ambient intelligence, and wireless sensor networks, and then, move to context-aware computing studies. Finally, we review learning and big data studies related to IoT. We also identify the open issues and provide an insight for future study areas for IoT researchers.","{'model': 'tldr@v2.0.0', 'text': 'The field is reviewed from a historical perspective, covering ubiquitous and pervasive computing, ambient intelligence, and wireless sensor networks, and then, move to context-aware computing studies, which identify the open issues and provide an insight for future study areas for IoT researchers.'}",
-technology acceptance model in m-learning context: a systematic review,M. Al-Emran,,"{'model': 'tldr@v2.0.0', 'text': 'The main findings include that most of the TAM studies involving M-learning focused on extending the TAM with external variables, followed by the studies that extended the model by factors from other theories/models.'}",http://umpir.ump.edu.my/id/eprint/22323/1/Technology%20Acceptance%20Model%20in%20M-learning%20context%20A%20systematic%20review.pdf
-transformer-xl: attentive language models beyond a fixed-length context,Zihang Dai,"Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence, which consists of a segment-level recurrence mechanism and a novel positional encoding scheme.'}",https://www.aclweb.org/anthology/P19-1285.pdf
-a review of flipped classroom and cooperative learning method within the context of vygotsky theory,D. Erbil,"In the flipped classroom method, which is accepted as one of the blended learning approaches, the traditional teaching process takes place outside of the classroom through videos. Activities, projects, and homework related to upper-level cognitive field steps are carried out during classroom time. Research and interest in the flipped classroom are increasing steadily. Employing a cooperative learning method is suggested for using class time in the flipped classroom method. However, there has not been sufficient research on the implemented results of those suggestions. Moreover, there is no clear roadmap on how to incorporate cooperative learning methods into the flipped classroom. This research reviews theoretical infrastructures of flipped classroom and cooperative learning methods according to the Vygotsky theory and makes various suggestions for implementation and implementers.","{'model': 'tldr@v2.0.0', 'text': 'This research reviews theoretical infrastructures of flipped classroom and cooperative learning methods according to the Vygotsky theory and makes various suggestions for implementation and implementers.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2020.01157/pdf
-learning context graph for person search,Yichao Yan,"Person re-identification has achieved great progress with deep convolutional neural networks. However, most previous methods focus on learning individual appearance feature embedding, and it is hard for the models to handle difficult situations with different illumination, large pose variance and occlusion. In this work, we take a step further and consider employing context information for person search. For a probe-gallery pair, we first propose a contextual instance expansion module, which employs a relative attention module to search and filter useful context information in the scene. We also build a graph learning framework to effectively employ context pairs to update target similarity. These two modules are built on top of a joint detection and instance feature learning framework, which improves the discriminativeness of the learned features. The proposed framework achieves state-of-the-art performance on two widely used person search datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a contextual instance expansion module, which employs a relative attention module to search and filter useful context information in the scene, and builds a graph learning framework to effectively employ context pairs to update target similarity.'}",https://arxiv.org/pdf/1904.01830
-dynamic context selection for document-level neural machine translation via reinforcement learning,Xiaomian Kang,"Document-level neural machine translation has yielded attractive improvements. However, majority of existing methods roughly use all context sentences in a fixed scope. They neglect the fact that different source sentences need different sizes of context. To address this problem, we propose an effective approach to select dynamic context so that the document-level translation model can utilize the more useful selected context sentences to produce better translations. Specifically, we introduce a selection module that is independent of the translation module to score each candidate context sentence. Then, we propose two strategies to explicitly select a variable number of context sentences and feed them into the translation module. We train the two modules end-to-end via reinforcement learning. A novel reward is proposed to encourage the selection and utilization of dynamic context sentences. Experiments demonstrate that our approach can select adaptive context sentences for different source sentences, and significantly improves the performance of document-level translation methods.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a selection module that is independent of the translation module to score each candidate context sentence, and proposes two strategies to explicitly select a variable number of context sentences and feed them into thetranslation module.'}",https://www.aclweb.org/anthology/2020.emnlp-main.175.pdf
-continual learning of context-dependent processing in neural networks,Guanxiong Zeng,,"{'model': 'tldr@v2.0.0', 'text': 'An approach involving a learning algorithm, called orthogonal weights modification, with the addition of a context-dependent processing module to enable highly compact systems to gradually learn myriad regularities of the real world and eventually behave appropriately within it.'}",https://arxiv.org/pdf/1810.01256
-recommender systems to support learners’ agency in a learning context: a systematic review,M. Deschênes,,"{'model': 'tldr@v2.0.0', 'text': 'This systematic review of the literature on recommenders for technology-enhanced learning examines the context in which recommenders are used, the manners in which they are evaluated and the results of those evaluations.'}",https://educationaltechnologyjournal.springeropen.com/counter/pdf/10.1186/s41239-020-00219-w
-context2vec: learning generic context embedding with bidirectional lstm,Oren Melamud,"Context representations are central to various NLP tasks, such as word sense disam-biguation, named entity recognition, co-reference resolution, and many more. In this work we present a neural model for efficiently learning a generic context embedding function from large corpora, us-ing bidirectional LSTM. With a very simple application of our context representations, we manage to surpass or nearly reach state-of-the-art results on sentence completion, lexical substitution and word sense disambiguation tasks, while substantially outperforming the popular context representation of averaged word embeddings. We release our code and pre-trained models, suggesting they could be useful in a wide variety of NLP tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work presents a neural model for efficiently learning a generic context embedding function from large corpora, us-ing bidirectional LSTM and suggests they could be useful in a wide variety of NLP tasks.'}",https://www.aclweb.org/anthology/K16-1006.pdf
-encoding history with context-aware representation learning for personalized search,Yujia Zhou,"The key to personalized search is to clarify the meaning of current query based on user's search history. Previous personalized studies tried to build user profiles on the basis of historical data to tailor the ranking. However, we argue that the user profile based methods do not really disambiguate the current query. They still retain some semantic bias when building user profiles. In this paper, we propose to encode history with context-aware representation learning to enhance the representation of current query, which is a direct way to clarify the user's information need. Specifically, endowed with the benefit from transformer on aggregating contextual information, we devise a query disambiguation model to parse the meaning of current query in multiple stages. Moreover, for covering the cases that current query is not sufficient to express the intent, we train a personalized language model to predict user intent from existing queries. Under the interaction of two sub-models, we can generate the context-aware representation of current query and re-rank the results based on it. Experimental results show the significant improvement of our model compared with previous methods.","{'model': 'tldr@v2.0.0', 'text': 'A query disambiguation model to parse the meaning of current query in multiple stages, endowed with the benefit from transformer on aggregating contextual information, and a personalized language model to predict user intent from existing queries is devised.'}",
-learning a deep listwise context model for ranking refinement,Qingyao Ai,"Learning to rank has been intensively studied and widely applied in information retrieval. Typically, a global ranking function is learned from a set of labeled data, which can achieve good performance on average but may be suboptimal for individual queries by ignoring the fact that relevant documents for different queries may have different distributions in the feature space. Inspired by the idea of pseudo relevance feedback where top ranked documents, which we refer as the local ranking context, can provide important information about the query's characteristics, we propose to use the inherent feature distributions of the top results to learn a Deep Listwise Context Model that helps us fine tune the initial ranked list. Specifically, we employ a recurrent neural network to sequentially encode the top results using their feature vectors, learn a local context model and use it to re-rank the top results. There are three merits with our model: (1) Our model can capture the local ranking context based on the complex interactions between top results using a deep neural network; (2) Our model can be built upon existing learning-to-rank methods by directly using their extracted feature vectors; (3) Our model is trained with an attention-based loss function, which is more effective and efficient than many existing listwise methods. Experimental results show that the proposed model can significantly improve the state-of-the-art learning to rank methods on benchmark retrieval corpora.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use the inherent feature distributions of the top results to learn a Deep Listwise Context Model that helps to fine tune the initial ranked list and can significantly improve the state-of-the-art learning to rank methods on benchmark retrieval corpora.'}",https://dl.acm.org/doi/pdf/10.1145/3209978.3209985
-saliency detection by multi-context deep learning,Rui Zhao,"Low-level saliency cues or priors do not produce good enough saliency detection results especially when the salient object presents in a low-contrast background with confusing visual appearance. This issue raises a serious problem for conventional approaches. In this paper, we tackle this problem by proposing a multi-context deep learning framework for salient object detection. We employ deep Convolutional Neural Networks to model saliency of objects in images. Global context and local context are both taken into account, and are jointly modeled in a unified multi-context deep learning framework. To provide a better initialization for training the deep neural networks, we investigate different pre-training strategies, and a task-specific pre-training scheme is designed to make the multi-context modeling suited for saliency detection. Furthermore, recently proposed contemporary deep models in the ImageNet Image Classification Challenge are tested, and their effectiveness in saliency detection are investigated. Our approach is extensively evaluated on five public datasets, and experimental results show significant and consistent improvements over the state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a multi-context deep learning framework for salient object detection that employs deep Convolutional Neural Networks to model saliency of objects in images and investigates different pre-training strategies to provide a better initialization for training the deep neural networks.'}",http://www.ee.cuhk.edu.hk/%7Exgwang/papers/zhaoOHWcvpr15.pdf
-multimodal deep learning for activity and context recognition,Valentin Radu,"Wearables and mobile devices see the world through the lens of half a dozen low-power sensors, such as, barometers, accelerometers, microphones and proximity detectors. But differences between sensors ranging from sampling rates, discrete and continuous data or even the data type itself make principled approaches to integrating these streams challenging. How, for example, is barometric pressure best combined with an audio sample to infer if a user is in a car, plane or bike? Critically for applications, how successfully sensor devices are able to maximize the information contained across these multi-modal sensor streams often dictates the fidelity at which they can track user behaviors and context changes. This paper studies the benefits of adopting deep learning algorithms for interpreting user activity and context as captured by multi-sensor systems. Specifically, we focus on four variations of deep neural networks that are based either on fully-connected Deep Neural Networks (DNNs) or Convolutional Neural Networks (CNNs). Two of these architectures follow conventional deep models by performing feature representation learning from a concatenation of sensor types. This classic approach is contrasted with a promising deep model variant characterized by modality-specific partitions of the architecture to maximize intra-modality learning. Our exploration represents the first time these architectures have been evaluated for multimodal deep learning under wearable data -- and for convolutional layers within this architecture, it represents a novel architecture entirely. Experiments show these generic multimodal neural network models compete well with a rich variety of conventional hand-designed shallow methods (including feature extraction and classifier construction) and task-specific modeling pipelines, across a wide-range of sensor types and inference tasks (four different datasets). Although the training and inference overhead of these multimodal deep approaches is in some cases appreciable, we also demonstrate the feasibility of on-device mobile and wearable execution is not a barrier to adoption. This study is carefully constructed to focus on multimodal aspects of wearable data modeling for deep learning by providing a wide range of empirical observations, which we expect to have considerable value in the community. We summarize our observations into a series of practitioner rules-of-thumb and lessons learned that can guide the usage of multimodal deep learning for activity and context detection.","{'model': 'tldr@v2.0.0', 'text': 'This paper studies the benefits of adopting deep learning algorithms for interpreting user activity and context as captured by multi-sensor systems under wearable data by evaluating four variations of deep neural networks based either on fully-connected Deep Neural Networks (DNNs) or Convolutional Neural networks (CNNs).'}",https://www.pure.ed.ac.uk/ws/files/76915345/Radu_et_al_2017_Multimodal_Deep_Learning.pdf
-extending utaut2 toward acceptance of mobile learning in the context of higher education,A. Arain,,"{'model': 'tldr@v2.0.0', 'text': 'A tailored extension in UTAUT2 is made that provides valuable insights into assess m-learning acceptance in the context of higher education institutes of developing countries, specifically in Pakistan.'}",
-the sigmorphon 2019 shared task: morphological analysis in context and cross-lingual transfer for inflection,Arya D. McCarthy,"The SIGMORPHON 2019 shared task on cross-lingual transfer and contextual analysis in morphology examined transfer learning of inflection between 100 language pairs, as well as contextual lemmatization and morphosyntactic description in 66 languages. The first task evolves past years’ inflection tasks by examining transfer of morphological inflection knowledge from a high-resource language to a low-resource language. This year also presents a new second challenge on lemmatization and morphological feature analysis in context. All submissions featured a neural component and built on either this year’s strong baselines or highly ranked systems from previous years’ shared tasks. Every participating team improved in accuracy over the baselines for the inflection task (though not Levenshtein distance), and every team in the contextual analysis task improved on both state-of-the-art neural and non-neural baselines.","{'model': 'tldr@v2.0.0', 'text': 'The SIGMORPHON 2019 shared task on cross-lingual transfer and contextual analysis in morphology examined transfer learning of inflection between 100 language pairs, as well as contextual lemmatization and morphosyntactic description in 66 languages.'}",https://www.aclweb.org/anthology/W19-4226.pdf
-behavdt: a behavioral decision tree learning to build user-centric context-aware predictive model,Iqbal H. Sarker,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the proposed BehavDT context-aware model is more effective when compared with the traditional machine learning approaches, in predicting user diverse behaviors considering multi-dimensional contexts.'}",https://arxiv.org/pdf/2001.00621
-learning dynamic context augmentation for global entity linking,Xiyuan Yang,"Despite of the recent success of collective entity linking (EL) methods, these “global” inference methods may yield sub-optimal results when the “all-mention coherence” assumption breaks, and often suffer from high computational cost at the inference stage, due to the complex search space. In this paper, we propose a simple yet effective solution, called Dynamic Context Augmentation (DCA), for collective EL, which requires only one pass through the mentions in a document. DCA sequentially accumulates context information to make efficient, collective inference, and can cope with different local EL models as a plug-and-enhance module. We explore both supervised and reinforcement learning strategies for learning the DCA model. Extensive experiments show the effectiveness of our model with different learning settings, base models, decision orders and attention mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'A simple yet effective solution, called Dynamic Context Augmentation (DCA), for collective EL, which requires only one pass through the mentions in a document, and can cope with different local EL models as a plug-and-enhance module.'}",https://www.aclweb.org/anthology/D19-1026.pdf
-improving students’ mathematical problem solving ability and self-efficacy through guided discovery learning in local culture context,R. E. Simamora,"Qualified learning materials is needed in the efforts to improve the quality of teaching-learning mathematics. Qualified learning materials can be obtained through development research. Learning materials in this study were learning materials that were developed based on guided discovery learning model. The learning materials was also developed by integrating local culture into a guided learning model. The local culture in this study was adapted to the local culture of the students, namely the Batak Toba. Learning materials in this study were developed using the development model of Thiagarajan et al. (1974). The result of second trial showed that learning materials based guided discovery learning with Batak Toba context improved students’ mathematical problem solving ability and self-efficacy significantly. Based on the results of the study, it was suggested that mathematics teachers make an effort qualified learning materials and integrate local culture in mathematics learning.",,https://www.iejme.com/download/improving-students-mathematical-problem-solving-ability-and-self-efficacy-through-guided-discovery-3966.pdf
-aligning assessment with the needs of work-integrated learning: the challenges of authentic assessment in a complex context,R. Ajjawi,"Abstract Work-integrated learning (WIL) is a feature of university courses, both in professional areas, where it is commonplace, but also across many different disciplines. Assessment of WIL can be complex as it involves parties and settings external to the university, and it can be problematic because of difficulties in aligning learning activities during placements with what is or can be assessed by the university. This paper explores the relationship between students’ placement experiences and accompanying assessments in contexts where activities are tightly coupled with the curriculum, and in those where it is not. It draws on a qualitative analysis of student interviews and drawings by the interviewees of their WIL experiences, supplemented with analysis of unit guides. Our findings highlight that students’ perceptions of authenticity of assessment were undermined by misalignments between the student, university and industry. Assessment authenticity was perceived by students as based on alignment between their current and future selves in the assessment process, involvement of industry supervisors and relevance of placement activities to assessment activities. The paper discusses the complexity of coordination of educational activities with external partners, especially when one party drives assessment. It then suggests a reframing of WIL assessment to promote alignment and authenticity.","{'model': 'tldr@v2.0.0', 'text': 'The relationship between students’ placement experiences and accompanying assessments in contexts where activities are tightly coupled with the curriculum, and in those where it is not is explored, and a reframing of WIL assessment is suggested to promote alignment and authenticity.'}",https://www.tandfonline.com/doi/pdf/10.1080/02602938.2019.1639613?needAccess=true
-self‐regulated learning in the clinical context: a systematic review,M. A. Houten-Schat,"Research has suggested beneficial effects of self‐regulated learning (SRL) for medical students' and residents' workplace‐based learning. Ideally, learners go through a cyclic process of setting learning goals, choosing learning strategies and assessing progress towards goals. A clear overview of medical students' and residents' successful key strategies, influential factors and effective interventions to stimulate SRL in the workplace is missing. This systematic review aims to provide an overview of and a theoretical base for effective SRL strategies of medical students and residents for their learning in the clinical context.","{'model': 'tldr@v2.0.0', 'text': 'This systematic review aims to provide an overview of and a theoretical base for effective SRL strategies of medical students and residents for their learning in the clinical context.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/medu.13615
-assessment in the context of problem-based learning,C. V. D. van der Vleuten,,"{'model': 'tldr@v2.0.0', 'text': 'Programmatic assessment comes very close to achieving the desired constructive alignment with PBL, but its wide adoption—just like PBL—will take many years ahead of us.'}",https://link.springer.com/content/pdf/10.1007/s10459-019-09909-1.pdf
-blended learning of physics in the context of the professional development of teachers,L. Krasnova,"In line with the improvement of traditional teaching methods, the new ones are intensively introduced at all levels of education. Usually, these are the methods tied with e-learning. Essentially, teachers must be able and ready to create an innovation-driven learning environment contributing to the effective individualization of the learning process. At the same time, each student should achieve the highest possible outcomes standing behind the personality development. This paper introduces the refresher courses designed for the physics teachers. These courses are based on the blended learning technology combining traditional full-time education with the elements of distance learning supported by LMS Moodle. The courses were tested at the Elabuga Institute of Kazan Federal University in 2016-2018. This paper describes the module-based course structure and methods for applying the e-learning modules. The distinctive feature of the course is that the content of the learning modules was designed to deliver different methods for teaching physics and to improve the general cultural competence of a teacher. The analysis of polling results (poll included 89 physics teachers) allowed assessing the efficiency of designed courses in the context of teacher’s professional development, his/her readiness to work in modern learning environment. The research outcomes will be also useful for foreign educational establishments implementing the teacher professional development programs.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the refresher courses designed for the physics teachers based on the blended learning technology combining traditional full-time education with the elements of blended education.'}",
-imitation from observation: learning to imitate behaviors from raw video via context translation,Yuxuan Liu,"Imitation learning is an effective approach for autonomous systems to acquire control policies when an explicit reward function is unavailable, using supervision provided as demonstrations from an expert, typically a human operator. However, standard imitation learning methods assume that the agent receives examples of observation-action tuples that could be provided, for instance, to a supervised learning algorithm. This stands in contrast to how humans and animals imitate: we observe another person performing some behavior and then figure out which actions will realize that behavior, compensating for changes in viewpoint, surroundings, object positions and types, and other factors. We term this kind of imitation learning “imitation-from-observation,” and propose an imitation learning method based on video prediction with context translation and deep reinforcement learning. This lifts the assumption in imitation learning that the demonstration should consist of observations in the same environment configuration, and enables a variety of interesting applications, including learning robotic skills that involve tool use simply by observing videos of human tool use. Our experimental results show the effectiveness of our approach in learning a wide range of real-world robotic tasks modeled after common household chores from videos of a human demonstrator, including sweeping, ladling almonds, pushing objects as well as a number of tasks in simulation.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an imitation learning method based on video prediction with context translation and deep reinforcement learning that enables a variety of interesting applications, including learning robotic skills that involve tool use simply by observing videos of human tool use.'}",https://arxiv.org/pdf/1707.03374
-context theory of classification learning.,D. Medin,"Most theories dealing with ill-defined concepts assume that performance is based on category level information or a mixture of category level and specific item information. A context theory of classificatio n is described in which judgments are assumed to derive exclusively from stored exemplar information. The main idea is that a probe item acts as a retrieval cue to access information associated with stimuli similar to the probe. The predictions of the context theory are contrasted with those of a class of theories (including prototype theory) that assume that the information entering into judgments can be derived from an additive combination of information from component cue dimensions. Across four experiments using both geometric forms and schematic faces as stimuli, the context theory consistently gave a better account of the data. The relation of the context theory to other theories and phenomena associated with ill-defined concepts is discussed in detail. One of the major components of cognitive behavior concerns abstracting rules and forming concepts. Our entire system of naming objects and events, talking about them, and interacting with them presupposes the ability to group experiences into appropriate classes. Young children learn to tell the difference between dogs and cats, between clocks and fans, and between stars and street lights. Since few concepts are formally taught, the evolution of concepts from experience with exemplars must be a fundamental learning phenomenon. The focus of the present article is to explore how such conceptual achievements emerge from individual instances.","{'model': 'tldr@v2.0.0', 'text': 'A context theory of classificatio n is described in which judgments are assumed to derive exclusively from stored exemplar information, and the main idea is that a probe item acts as a retrieval cue to access information associated with stimuli similar to the probe.'}",
-"self-directed language learning in a mobile-assisted, out-of-class context: do students walk the talk?",Gustavo García Botero,"Abstract Can mobile-assisted language learning (MALL) foster self-directed learning outside the classroom? This article examines informal, out-of-class engagement with a MALL tool: Duolingo. After being invited to use Duolingo, 118 higher education language students agreed to have their activity in the application tracked. In addition to the data collected from the Duolingo dashboard, questionnaires, as well as semi-structured interviews, were implemented. Results viewed from Garrison’s dimensions of self-directed learning (motivation, self-management and self-monitoring) reveal inconsistencies between what students think about the app and what they do with it. Questionnaires suggest that Duolingo can encourage out-of-class learning through fun activities, but interviews reveal a lack of sustained motivation, self-monitoring and self-management reflected in the low usage of the application. Tracking data reveals that students lack the time for out-of-class MALL during course weeks and are more active during holidays. The results highlight the importance of a context in which users are encouraged to take responsibility for their learning. However, most students will, besides encouragement, also need training and support for their self-directed learning.",,
-improvements to context based self-supervised learning,T. N. Mundhenk,"We develop a set of methods to improve on the results of self-supervised learning using context. We start with a baseline of patch based arrangement context learning and go from there. Our methods address some overt problems such as chromatic aberration as well as other potential problems such as spatial skew and mid-level feature neglect. We prevent problems with testing generalization on common self-supervised benchmark tests by using different datasets during our development. The results of our methods combined yield top scores on all standard self-supervised benchmarks, including classification and detection on PASCAL VOC 2007, segmentation on PASCAL VOC 2012, and ""linear tests"" on the ImageNet and CSAIL Places datasets. We obtain an improvement over our baseline method of between 4.0 to 7.1 percentage points on transfer learning classification tests. We also show results on different standard network architectures to demonstrate generalization as well as portability. All data, models and programs are available at: https://gdo-datasci.llnl.gov/selfsupervised/.","{'model': 'tldr@v2.0.0', 'text': 'A set of methods to improve on the results of self-supervised learning using context, which start with a baseline of patch based arrangement context learning and go from there, and shows results on different standard network architectures to demonstrate generalization as well as portability.'}",https://arxiv.org/pdf/1711.06379
-personalization in context-aware ubiquitous learning-log system,Mengmeng Li,"This paper describes a ubiquitous learning log system called SCROLL (System for Capturing and Reminding of Learning Log). Learning log stands for the log of knowledge or learning experience acquired ubiquitously. This study primarily exploits a personalized learning and context-aware method supporting ubiquitous learning log system. Its aims lie in helping learners recall what they have logged (learned) making use of the contexts and learners' learning habits. The method contains three main measures, which are to recommend learning objects in accordance with both learners' needs and contexts, to detect their learning habits using the context history and to prompt them to review what they have learned regarding their learning habits. What's more, by monitoring learners' reaction on the recommendation or prompting, the method can improve its prediction. An experiment was conducted to evaluate SCROLL and the method. The results demonstrate that the system is very helpful for the learners and they benefit from the context-based recommendation and learning-style based prompting well.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that the SCROLL system is very helpful for the learners and they benefit from the context-based recommendation and learning-style based prompting well.'}",
-"context, learning, and extinction.",S. Gershman,"A. Redish et al. (2007) proposed a reinforcement learning model of context-dependent learning and extinction in conditioning experiments, using the idea of ""state classification"" to categorize new observations into states. In the current article, the authors propose an interpretation of this idea in terms of normative statistical inference. They focus on renewal and latent inhibition, 2 conditioning paradigms in which contextual manipulations have been studied extensively, and show that online Bayesian inference within a model that assumes an unbounded number of latent causes can characterize a diverse set of behavioral results from such manipulations, some of which pose problems for the model of Redish et al. Moreover, in both paradigms, context dependence is absent in younger animals, or if hippocampal lesions are made prior to training. The authors suggest an explanation in terms of a restricted capacity to infer new causes.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that online Bayesian inference within a model that assumes an unbounded number of latent causes can characterize a diverse set of behavioral results from such manipulations, some of which pose problems for the model of Redish et al. (2007).'}",
-challenges and opportunities for higher education amid the covid-19 pandemic: the philippine context,C. Toquero,"Covid-19 affected higher educational institutions not just in Wuhan, China where the virus originated but all other higher educational institutions in 188 countries as of April 06, 2020. Educational countermeasures are taken to continue educating the students despite the COVID-19 predicaments. Based on the author’s experiences, research, observations in the academe, COVID-19 guidelines, and the need for alternative solutions, this article introduces how higher education is affected and how it can respond to future challenges. This article recommends to educational institutions to produce studies to proliferate and document the impact of the pandemic to the educational system. There is also a greater need for educational institutions to strengthen the practices in the curriculum and make it more responsive to the learning needs of the students even beyond the conventional classrooms.",,https://www.pedagogicalresearch.com/download/challenges-and-opportunities-for-higher-education-amid-the-covid-19-pandemic-the-philippine-context-7947.pdf
-relationship between interest and mathematics performance in a technology-enhanced learning context in malaysia,S. L. Wong,,"{'model': 'tldr@v2.0.0', 'text': 'The findings of this study pointed to the importance of igniting interest among students with lower mathematics performance given its strong link to mathematics performance.'}",https://telrp.springeropen.com/track/pdf/10.1186/s41039-019-0114-3
-gene function prediction with gene interaction networks: a context graph kernel approach,Xin Li,"Predicting gene functions is a challenge for biologists in the postgenomic era. Interactions among genes and their products compose networks that can be used to infer gene functions. Most previous studies adopt a linkage assumption, i.e., they assume that gene interactions indicate functional similarities between connected genes. In this study, we propose to use a gene's context graph, i.e., the gene interaction network associated with the focal gene, to infer its functions. In a kernel-based machine-learning framework, we design a context graph kernel to capture the information in context graphs. Our experimental study on a testbed of p53-related genes demonstrates the advantage of using indirect gene interactions and shows the empirical superiority of the proposed approach over linkage-assumption-based methods, such as the algorithm to minimize inconsistent connected genes and diffusion kernels.","{'model': 'tldr@v2.0.0', 'text': 'The experimental study on a testbed of p53-related genes demonstrates the advantage of using indirect gene interactions and shows the empirical superiority of the proposed approach over linkage-assumption-based methods, such as the algorithm to minimize inconsistent connected genes and diffusion kernels.'}",https://arxiv.org/pdf/2204.10473
-a context-aware user-item representation learning for item recommendation,Libing Wu,"Both reviews and user-item interactions (i.e., rating scores) have been widely adopted for user rating prediction. However, these existing techniques mainly extract the latent representations for users and items in an independent and static manner. That is, a single static feature vector is derived to encode user preference without considering the particular characteristics of each candidate item. We argue that this static encoding scheme is incapable of fully capturing users’ preferences, because users usually exhibit different preferences when interacting with different items. In this article, we propose a novel context-aware user-item representation learning model for rating prediction, named CARL. CARL derives a joint representation for a given user-item pair based on their individual latent features and latent feature interactions. Then, CARL adopts Factorization Machines to further model higher order feature interactions on the basis of the user-item pair for rating prediction. Specifically, two separate learning components are devised in CARL to exploit review data and interaction data, respectively: review-based feature learning and interaction-based feature learning. In the review-based learning component, with convolution operations and attention mechanism, the pair-based relevant features for the given user-item pair are extracted by jointly considering their corresponding reviews. However, these features are only reivew-driven and may not be comprehensive. Hence, an interaction-based learning component further extracts complementary features from interaction data alone, also on the basis of user-item pairs. The final rating score is then derived with a dynamic linear fusion mechanism. Experiments on seven real-world datasets show that CARL achieves significantly better rating prediction accuracy than existing state-of-the-art alternatives. Also, with the attention mechanism, we show that the pair-based relevant information (i.e., context-aware information) in reviews can be highlighted to interpret the rating prediction for different user-item pairs.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes a novel context-aware user-item representation learning model for rating prediction, named CARL, which derives a joint representation for a given user- item pair based on their individual latent features and latent feature interactions.'}",https://arxiv.org/pdf/1712.02342
-global context enhanced graph neural networks for session-based recommendation,Ziyang Wang,"Session-based recommendation (SBR) is a challenging task, which aims at recommending items based on anonymous behavior sequences. Almost all the existing solutions for SBR model user preference only based on the current session without exploiting the other sessions, which may contain both relevant and irrelevant item-transitions to the current session. This paper proposes a novel approach, called Global Context Enhanced Graph Neural Networks (GCE-GNN) to exploit item transitions over all sessions in a more subtle manner for better inferring the user preference of the current session. Specifically, GCE-GNN learns two levels of item embeddings from session graph and global graph, respectively: (i) Session graph, which is to learn the session-level item embedding by modeling pairwise item-transitions within the current session; and (ii) Global graph, which is to learn the global-level item embedding by modeling pairwise item-transitions over all sessions. In GCE-GNN, we propose a novel global-level item representation learning layer, which employs a session-aware attention mechanism to recursively incorporate the neighbors' embeddings of each node on the global graph. We also design a session-level item representation learning layer, which employs a GNN on the session graph to learn session-level item embeddings within the current session. Moreover, GCE-GNN aggregates the learnt item representations in the two levels with a soft attention mechanism. Experiments on three benchmark datasets demonstrate that GCE-GNN outperforms the state-of-the-art methods consistently.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach to exploit item transitions over all sessions in a more subtle manner for better inferring the user preference of the current session, called GCE-GNN, which outperforms the state-of-the-art methods consistently.'}",https://arxiv.org/pdf/2106.05081
-theorising decolonisation in the context of lifelong learning and transnational migration: anti-colonial and anti-racist perspectives,Srabani Maitra,"ABSTRACT In the age of transnational migration, the practices and policies of lifelong learning in many immigrant-receiving countries continue to be impacted by the cultural and discursive politics of colonial legacies. Drawing on a wide range of anti-colonial and anti-racist scholarship, we argue for an approach to lifelong learning that aims to decolonise the ideological underpinnings of colonial relations of rule, especially in terms of its racialised privileging of ‘whiteness’ and Eurocentrism. In the context of lifelong learning, decolonisation would achieve four important purposes. First, it would illustrate the nexus between knowledge, power, and colonial narratives by interrogating how knowledge-making is a fundamental aspect of ‘coloniality’. Second, decolonisation would entail challenging the hegemony of western knowledge, education, and credentials and upholding a ‘multiculturalism of knowledge’ that is inclusive and responsive to the cultural needs and values of transnational migrants. Third, decolonisation would lead to the need for planning and designing learning curricula as well as institutionalised pedagogy based on non-western knowledge systems and epistemic diversity. The final emphasis is on the urgency to decolonise our minds as lifelong learners, practitioners and policy-makers in order to challenge the passivity, colonisation, and marginalisation of learners both in classrooms and workplaces.",,http://eprints.gla.ac.uk/177182/7/177182.pdf
-smart mobile learning environment for programming education in nigeria: adaptivity and context-aware features,F. J. Agbo,,"{'model': 'tldr@v2.0.0', 'text': 'The study presents some existing solutions for programming education with adaptive or context-aware features of SLE and proposes a model for designing an SLE with adaptivity and context-awareness features that take into cognizance the learner’s preferences and needs.'}",
-attribute recognition by joint recurrent learning of context and correlation,Jingya Wang,"Recognising semantic pedestrian attributes in surveillance images is a challenging task for computer vision, particularly when the imaging quality is poor with complex background clutter and uncontrolled viewing conditions, and the number of labelled training data is small. In this work, we formulate a Joint Recurrent Learning (JRL) model for exploring attribute context and correlation in order to improve attribute recognition given small sized training data with poor quality images. The JRL model learns jointly pedestrian attribute correlations in a pedestrian image and in particular their sequential ordering dependencies (latent high-order correlation) in an end-to-end encoder/ decoder recurrent network. We demonstrate the performance advantage and robustness of the JRL model over a wide range of state-of-the-art deep models for pedestrian attribute recognition, multi-label image classification, and multi-person image annotation on two largest pedestrian attribute benchmarks PETA and RAP.","{'model': 'tldr@v2.0.0', 'text': 'The JRL model learns jointly pedestrian attribute correlations in a pedestrian image and in particular their sequential ordering dependencies (latent high-order correlation) in an end-to-end encoder/ decoder recurrent network.'}",https://arxiv.org/pdf/1709.08553
-unsupervised learning of spoken language with visual context,David F. Harwath,"Humans learn to speak before they can read or write, so why can’t computers do the same? In this paper, we present a deep neural network model capable of rudimentary spoken language acquisition using untranscribed audio training data, whose only supervision comes in the form of contextually relevant visual images. We describe the collection of our data comprised of over 120,000 spoken audio captions for the Places image dataset and evaluate our model on an image search and annotation task. We also provide some visualizations which suggest that our model is learning to recognize meaningful words within the caption spectrograms.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network model capable of rudimentary spoken language acquisition using untranscribed audio training data, whose only supervision comes in the form of contextually relevant visual images, is presented.'}",
-developing a mobile app for learning english vocabulary in an open distance learning context,Mpine Makoe,"Academic success depends on the comprehension of a language, which is linked to vocabulary learning. Many distance students in South Africa find it difficult to comprehend learning in a language other than their mother tongue. Finding effective strategies for enhancing English vocabulary of university students amidst the spatial, temporal, and pedagogic distance associated with Open Distance Learning (ODL) practices remains a challenge. To address the need for enhancing vocabulary development, mobile application systems (apps) were explored as the best vehicle for the delivery of the vocabulary learning. Mobile learning technologies are ideal in the ODL context because they are flexible, accessible, available, and cater for a myriad of interaction activities. The purpose of the study is to design and implement a mobile-based application aimed at enhancing English vocabulary teaching and learning. Using the Design-Based Research methodology, this study maps the steps taken to develop a vocabulary learning mobile app named VocUp; it describes the architecture, user interface, features of VocUp, and advocates for contextually-conscious and learning-driven app development.","{'model': 'tldr@v2.0.0', 'text': 'This study maps the steps taken to develop a vocabulary learning mobile app named VocUp and describes the architecture, user interface, features of VocUp, and advocates for contextually-conscious and learning-driven app development.'}",http://www.irrodl.org/index.php/irrodl/article/download/3746/4777
-ce-net: context encoder network for 2d medical image segmentation,Zaiwang Gu,"Medical image segmentation is an important step in medical image analysis. With the rapid development of a convolutional neural network in image processing, deep learning has been used for medical image segmentation, such as optic disc segmentation, blood vessel detection, lung segmentation, cell segmentation, and so on. Previously, U-net based approaches have been proposed. However, the consecutive pooling and strided convolutional operations led to the loss of some spatial information. In this paper, we propose a context encoder network (CE-Net) to capture more high-level information and preserve spatial information for 2D medical image segmentation. CE-Net mainly contains three major components: a feature encoder module, a context extractor, and a feature decoder module. We use the pretrained ResNet block as the fixed feature extractor. The context extractor module is formed by a newly proposed dense atrous convolution block and a residual multi-kernel pooling block. We applied the proposed CE-Net to different 2D medical image segmentation tasks. Comprehensive results show that the proposed method outperforms the original U-Net method and other state-of-the-art methods for optic disc segmentation, vessel detection, lung segmentation, cell contour segmentation, and retinal optical coherence tomography layer segmentation.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive results show that the proposed CE-Net method outperforms the original U- net method and other state-of-the-art methods for optic disc segmentation, vessel detection, lung segmentation , cell contour segmentation and retinal optical coherence tomography layer segmentation.'}",https://arxiv.org/pdf/1903.02740
-collaborative live media curation: shared context for participation in online learning,William A. Hamilton,"In recent years, online education's reach and scale have increased through new platforms for large and small online courses. However, these platforms often rely on impoverished modalities, which provide limited support for participation in social learning experiences. We present Collaborative Live Media Curation (CLMC), a new medium for sharing context and participation in online learning. CLMC involves collaborative, synchronous collection, creation, and assemblage of web media, including images, text, video, and sketch. CLMC integrates live media including streaming video, screenshares, audio, and text chat. We deploy and study LiveMâché, a CLMC technology probe, in four situated online learning contexts. We discovered student and instructor strategies for sharing context and participating including creating curations in advance, sketching to illustrate and gesture, real-time transformations, sharing perspective, and assembling live streams. We develop implications through live experience patterns, which describe how spatial and computing structures support social activities.","{'model': 'tldr@v2.0.0', 'text': 'This work deploy and study LiveMâché, a CLMC technology probe, in four situated online learning contexts and discovered student and instructor strategies for sharing context and participating including creating curations in advance, sketching to illustrate and gesture, real-time transformations, sharing perspective, and assembling live streams.'}",
-choosing the right learning management system (lms) for the higher education institution context: a systematic review,N. Kasim,"A Learning Management System (LMS) is one e-learning tool that has been widely used to improve students’ learning experience and the construction of their understanding of certain topics. This concept paper discusses a number of potential Learning Management Systems (LMS) that can be utilised for teaching and learning processes in the context of Higher Education Institutions, for example Moodle, ATutor, Blackboard and SuccessFactors. The comparison is made based on a literature review of the characteristics of the selected LMS providers. Among the characteristics considered are flexibility, ease of use, accessibility and user-friendliness. This paper also provides some conclusions on the selection of the platform to be used. The results of this study give readers information to make their own decisions when choosing an LMS platform based on the needs of their institution.","{'model': 'tldr@v2.0.0', 'text': 'A number of potential Learning Management Systems that can be utilised for teaching and learning processes in the context of Higher Education Institutions, for example Moodle, ATutor, Blackboard and SuccessFactors are discussed.'}",https://online-journals.org/index.php/i-jet/article/download/5644/3990
-"offline reinforcement learning: tutorial, review, and perspectives on open problems",S. Levine,"In this tutorial article, we aim to provide the reader with the conceptual tools needed to get started on research on offline reinforcement learning algorithms: reinforcement learning algorithms that utilize previously collected data, without additional online data collection. Offline reinforcement learning algorithms hold tremendous promise for making it possible to turn large datasets into powerful decision making engines. Effective offline reinforcement learning methods would be able to extract policies with the maximum possible utility out of the available data, thereby allowing automation of a wide range of decision-making domains, from healthcare and education to robotics. However, the limitations of current algorithms make this difficult. We will aim to provide the reader with an understanding of these challenges, particularly in the context of modern deep reinforcement learning methods, and describe some potential solutions that have been explored in recent work to mitigate these challenges, along with recent applications, and a discussion of perspectives on open problems in the field.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial article aims to provide the reader with the conceptual tools needed to get started on research on offline reinforcement learning algorithms: reinforcementlearning algorithms that utilize previously collected data, without additional online data collection.'}",
-"context-aware, adaptive, and scalable android malware detection through online learning",A. Narayanan,"It is well known that Android malware constantly evolves so as to evade detection. This causes the entire malware population to be nonstationary. Contrary to this fact, most of the prior works on machine learning based android malware detection have assumed that the distribution of the observed malware characteristics (i.e., features) does not change over time. In this paper, we address the problem of malware population drift and propose a novel online learning based framework to detect malware, named Casandra (C ontext-aware, Adaptive and Scalable ANDRoid m Alware detector). In order to perform accurate detection, a novel graph kernel that facilitates capturing apps security-sensitive behaviors along with their context information from dependence graphs is proposed. Besides being accurate and scalable, Casandra has specific advantages: first, being adaptive to the evolution in malware features over time; second, explaining the significant features that led to an apps classification as being malicious or benign. In a large-scale comparative analysis, Casandra outperforms two state-of-the-art techniques on a benchmark dataset achieving 99.23% F-measure. When evaluated with more than 87 000 apps collected in-the-wild, Casandra achieves 89.92% accuracy, outperforming existing techniques by more than 25% in their typical batch learning setting and more than $\text{7}$% when they are continuously retained, while maintaining comparable efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the problem of malware population drift and proposes a novel online learning based framework to detect malware, named Casandra, which has specific advantages: first, being adaptive to the evolution in malware features over time; second, explaining the significant features that led to an apps classification as being malicious or benign.'}",
-"definition, framework and research issues of smart learning environments - a context-aware ubiquitous learning perspective",Gwo-jen Hwang,,"{'model': 'tldr@v2.0.0', 'text': 'The definition and criteria of smartlearning environments are presented from the perspective of context-aware ubiquitous learning and a framework is presented to address the design and development considerations of smart learning environments to support both online and real-world learning activities.'}",https://slejournal.springeropen.com/track/pdf/10.1186/s40561-014-0004-5
-object detection with deep learning: a review,Zhong-Qiu Zhao,"Due to object detection’s close relationship with video analysis and image understanding, it has attracted much research attention in recent years. Traditional object detection methods are built on handcrafted features and shallow trainable architectures. Their performance easily stagnates by constructing complex ensembles that combine multiple low-level image features with high-level context from object detectors and scene classifiers. With the rapid development in deep learning, more powerful tools, which are able to learn semantic, high-level, deeper features, are introduced to address the problems existing in traditional architectures. These models behave differently in network architecture, training strategy, and optimization function. In this paper, we provide a review of deep learning-based object detection frameworks. Our review begins with a brief introduction on the history of deep learning and its representative tool, namely, the convolutional neural network. Then, we focus on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further. As distinct specific detection tasks exhibit different characteristics, we also briefly survey several specific tasks, including salient object detection, face detection, and pedestrian detection. Experimental analyses are also provided to compare various methods and draw some meaningful conclusions. Finally, several promising directions and tasks are provided to serve as guidelines for future work in both object detection and relevant neural network-based learning systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a review of deep learning-based object detection frameworks and focuses on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further.'}",https://arxiv.org/pdf/1807.05511
-"rousing minds to life: teaching, learning, and schooling in social context",R. Tharp,"Acknowledgments Introduction Part I. Teaching, Schooling, and Literacy: A Unified Theory of Education: 1. The redefinition of teaching and schooling 2. A theory of teaching as assisted performance 3. The means of assisting performance 4. The social organization of assisted performance 5. Language, literacy, and thought Part II. Practice: 6. A school organized for teaching: the Kamehameha Elementary Education Program 7. The activity setting of the instructional conversation: developing word and discourse meaning 8. The orchestration of activity settings: learning and social interaction in the whole group and independent centers 9. The interpsychological plane of teacher training 10. Assisting teacher performance through the ZPD: a case study 11. The intrapsychological plane of teacher training: the internalization of higher-order teaching skills 12. The schools in mind and society References Author index Subject index.",,
-identifying context factors as a source for teacher professional learning,A. Koffeman,"ABSTRACT Traditionally, teacher professional learning is often seen as something that mainly takes place in organized, formalized settings. This article takes a broader starting point: the idea that it can be understood as the result of the teachers’ confrontations and interactions with and within their professional contexts, and that context can thus serve as a source for learning. To gain a closer understanding of the nature of these contexts, we start with exploring these contexts from a theoretical point of view. We then look for evidence for the role of different contexts in teacher learning autobiographies and learner reports, produced by a diverse group of experienced teachers (n = 18), as part of their course work in a master’s programme. The findings suggest that three contexts can be identified – a perso3nal practice domain, a social domain, and a theoretical domain – and that confrontations in each of these domains can take place planned as well as unplanned. The relevance of the study is that understanding the ways in which context can serve as a source for learning can contribute to the debate on how to stimulate teacher professional learning.",,
-deep learning and process understanding for data-driven earth system science,M. Reichstein,,"{'model': 'tldr@v2.0.0', 'text': 'It is argued that contextual cues should be used as part of deep learning to gain further process understanding of Earth system science problems, improving the predictive ability of seasonal forecasting and modelling of long-range spatial connections across multiple timescales.'}",https://pure.mpg.de/pubman/item/item_3029184_9/component/file_3282959/BGC3001P.pdf
-contextual cueing: implicit learning and memory of visual context guides spatial attention,M. Chun,"Global context plays an important, but poorly understood, role in visual tasks. This study demonstrates that a robust memory for visual context exists to guide spatial attention. Global context was operationalized as the spatial layout of objects in visual search displays. Half of the configurations were repeated across blocks throughout the entire session, and targets appeared within consistent locations in these arrays. Targets appearing in learned configurations were detected more quickly. This newly discovered form of search facilitation is termed contextual cueing. Contextual cueing is driven by incidentally learned associations between spatial configurations (context) and target locations. This benefit was obtained despite chance performance for recognizing the configurations, suggesting that the memory for context was implicit. The results show how implicit learning and memory of visual context can guide spatial attention towards task-relevant aspects of a scene.","{'model': 'tldr@v2.0.0', 'text': 'The results show how implicit learning and memory of visual context can guide spatial attention towards task-relevant aspects of a scene.'}",
-informal self-regulated learning on a surgical rotation: uncovering student experiences in context,Nicole N. Woods,,"{'model': 'tldr@v2.0.0', 'text': 'The study aimed to examine undergraduate medical students’ day-to-day learning strategies in order to better understand the process of informal self-regulated learning in practice and to provide a forum for students to share and critically reflect on their own self- regulated learning strategies.'}",
-a guide to deep learning in healthcare,A. Esteva,,"{'model': 'tldr@v2.0.0', 'text': 'How these computational techniques can impact a few key areas of medicine and explore how to build end-to-end systems are described.'}",
-"context, time, and memory retrieval in the interference paradigms of pavlovian learning.",M. Bouton,"In this article I review research and theory on the ""interference paradigms"" in Pavlovian learning. In these situations (e.g., extinction, counterconditioning, and latent inhibition), a conditioned stimulus (CS) is associated with different unconditioned stimuli (USs) or outcomes in different phases of the experiment; retroactive interference, proactive interference, or both are often observed. In all of the paradigms, contextual stimuli influence performance, and when information is available, so does the passage of time. Memories of both phases are retained, and performance may depend on which is retrieved. Despite the similarity of the paradigms, conditioning theories tend to explain them with separate mechanisms. They also do not provide an adequate account of the context's role, fail to predict the effects of time, and overemphasize the role of learning or storage deficits. By accepting 4 propositions about animal memory (i.e., contextual stimuli guide retrieval, time is a context, different memories are differentially dependent on context, and interference occurs at performance output), a memory retrieval framework can provide an integrated account of context, time, and performance in the various paradigms.","{'model': 'tldr@v2.0.0', 'text': 'A memory retrieval framework can provide an integrated account of context, time, and performance in the various paradigms of Pavlovian learning by accepting 4 propositions about animal memory.'}",
-context-aware recommendation-based learning analytics using tensor and coupled matrix factorization,Faisal M. Almutairi,"Student retention and timely graduation are enduring challenges in higher education. With the rapidly expanding collection and availability of learning data and related analytics, student performance can be accurately monitored, and possibly predicted ahead of time, thus, enabling early warning and degree planning “expert systems” to provide disciplined decision support to counselors, advisors, and educators. Previous work in educational data mining has explored matrix factorization techniques for grade prediction, albeit without taking contextual information into account. Temporal information should be informative as it distinguishes between the different class offerings and indirectly captures student experience as well. To exploit temporal and/or other kinds of context, we develop three approaches under the framework of collaborative filtering (CF). Two of the proposed approaches build upon coupled matrix factorization with a shared latent matrix factor. The third utilizes tensor factorization to model grades and their context, without introducing a new mode per context dimension as is common in the CF literature. The latent factors obtained can be used to predict grades and context, if desired. We evaluate these approaches on grade data obtained from the University of Minnesota. Experimental results show that fairly good prediction is possible even with simple approaches, but very accurate prediction is hard. The more advanced approaches can increase prediction accuracy, but only up to a point for the particular dataset considered.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that fairly good prediction is possible even with simple approaches, but very accurate prediction is hard, and the more advanced approaches can increase prediction accuracy, but only up to a point for the particular dataset considered.'}",
-a review of content-based and context-based recommendation systems,Umair Javed,"In our work, we have presented two widely used recommendation systems. We have presented a context-aware recommender system to filter the items associated with user’s interests coupled with a context-based recommender system to prescribe those items. In this study, context-aware recommender systems perceive the user’s location, time, and company. The context-based recommender system retrieves patterns from World Wide Web-based on the user’s past interactions and provides future news recommendations. We have presented different techniques to support media recommendations for smartphones, to create a framework for context-aware, to filter E-learning content, and to deliver convenient news to the user. To achieve this goal, we have used content-based, collaborative filtering, a hybrid recommender system, and implemented a Web ontology language (OWL). We have also used the Resource Description Framework (RDF), JAVA, machine learning, semantic mapping rules, and natural ontology languages that suggest user items related to the search. In our work, we have used E-paper to provide users with the required news. After applying the semantic reasoning approach, we have concluded that by some means, this approach works similarly as a content-based recommender system since by taking the gain of a semantic approach, we can also recommend items according to the user’s interests. In a content-based recommender system, the system provides additional options or results that rely on the user’s ratings, appraisals, and interests.","{'model': 'tldr@v2.0.0', 'text': 'This study has concluded that by some means, this approach works similarly as a content-based recommender system since by taking the gain of a semantic approach, the system can also recommend items according to the user’s interests.'}",
-sustainability in higher education in the context of the un desd: a review of learning and institutionalization processes,A. Wals,,,
-"sociophonetics: the role of words, the role of context, and the role of words in context",J. Hay,"This paper synthesizes a wide range of literature from sociolinguistics and cognitive psychology, to argue for a central role for the ""word"" as a vehicle of language variation and change. Three crucially interlinked strands of research are reviewed-the role of context in associative learning, the word-level storage of phonetic and contextual detail, and the phonetic consequences of skewed distributions of words across different contexts. I argue that the human capacity for associative learning, combined with attention to fine-phonetic detail at the level of the word, plays a significant role in predicting a range of subtle but systematically robust observed socioindexical patterns in speech production and perception.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that the human capacity for associative learning, combined with attention to fine-phonetic detail at the level of the word, plays a significant role in predicting a range of subtle but systematically robust observed socioindexical patterns in speech production and perception.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/tops.12326
-nicotine enhances context learning but not context-shock associative learning.,Justin W. Kenney,"Nicotine has been found to enhance learning in a variety of tasks, including contextual fear conditioning. During contextual fear conditioning animals have to learn the context and associate the context with an unconditioned stimulus (footshock). As both of these types of learning co-occur during fear conditioning, it is not clear whether nicotine enhances one or both of these types of learning. To tease these two forms of learning apart, the authors made use of the context preexposure facilitation effect (CPFE). Acquisition of the CPFE requires that contextual and context-shock learning occurs on separate days, allowing for their individual manipulation. Nicotine (0.09 mg/kg) administered prior to contextual learning and retrieval enhanced the CPFE whereas administration prior to context-shock learning and retrieval had no effect. Thus, nicotine enhances contextual learning but not context-shock associative learning. Finally, the results are discussed in terms of a theory of how nicotine could alter hippocampal-cortical-amygdala interactions to facilitate contextual learning.","{'model': 'tldr@v2.0.0', 'text': 'Nicotine enhances contextual learning but not context-shock associative learning, which is consistent with a theory of how nicotine could alter hippocampal-cortical-amygdala interactions to facilitate contextual learning.'}",https://europepmc.org/articles/pmc2559949?pdf=render
-development and evaluation of an active learning support system for context-aware ubiquitous learning,T. Hsu,"Situating students to learn from the real world has been recognized as an important and challenging issue. However, in a real-world learning environment, there are usually many physical constraints that affect the learning performance of students, such as the total learning time, the limitation of the number of students who can visit a learning target, and the time needed for moving from one learning location to another. It is essential to guide the students along an efficient learning path to maximize their learning performance according to the current situation. In this paper, an active learning support system (ALESS) for context-aware ubiquitous learning environments is designed and developed. ALESS can provide learning guidance when conducting ubiquitous learning activities. A great deal of context information is used in ALESS, including the location, the current capacity of the learning object, the time available, etc. ALESS is able to actively provide the required learning support to individual students when they approach the corresponding real-world learning targets. To evaluate the performance of ALESS, an experiment was conducted in the National Science Museum of Taiwan. The experimental results showed that, with the help of ALESS, the students learned more efficiently, and achieved better learning performance.","{'model': 'tldr@v2.0.0', 'text': 'An active learning support system (ALESS) for context-aware ubiquitous learning environments is designed and developed and shows that, with the help of ALESS, the students learned more efficiently, and achieved better learning performance.'}",https://ieeexplore.ieee.org/ielx7/4620076/7396975/07115952.pdf
-context counts: how learners' contexts influence learning in a mooc,Nina Hood,,"{'model': 'tldr@v2.0.0', 'text': ""This study examines how a learner's current role and context influences their ability to self-regulate their learning in a MOOC: Introduction to Data Science offered by Coursera.""}",https://researchonline.gcu.ac.uk/files/24389185/6728648_Context_Counts_preprint.pdf
-investigating different syntactic context types and context representations for learning word embeddings,Bofang Li,"The number of word embedding models is growing every year. Most of them are based on the co-occurrence information of words and their contexts. However, it is still an open question what is the best definition of context. We provide a systematical investigation of 4 different syntactic context types and context representations for learning word embeddings. Comprehensive experiments are conducted to evaluate their effectiveness on 6 extrinsic and intrinsic tasks. We hope that this paper, along with the published code, would be helpful for choosing the best context type and representation for a given task.","{'model': 'tldr@v2.0.0', 'text': 'This paper provides a systematical investigation of 4 different syntactic context types and context representations for learning word embeddings and hopes it would be helpful for choosing the best context type and representation for a given task.'}",
-education context and english teaching and learning in the kingdom of saudi arabia: an overview.,Oqab Alrashidi,"This paper discusses the education context and English teaching and learning in the Kingdom of Saudi Arabia (KSA). The paper is organised into five main sections. The first section offers a brief glance at the social, religious, economic, and political context in KSA. The second section provides an overview of the education system in KSA, which includes a brief explanation of the history of education, a discussion of the role of government in modern education, and a description of the administration of education in the country. The third section presents information about English teaching and learning in public and higher education in KSA, while the fourth gives insights into the challenges and difficulties students face when learning English. The last section focuses on the importance of learning English in KSA.",,https://ccsenet.org/journal/index.php/elt/article/download/47946/25732
-context learning in the rodent hippocampus,Mark C. Fuhs,"We present a Bayesian statistical theory of context learning in the rodent hippocampus. While context is often defined in an experimental setting in relation to specific background cues or task demands, we advance a single, more general notion of context that suffices for a variety of learning phenomena. Specifically, a context is defined as a statistically stationary distribution of experiences, and context learning is defined as the problem of how to form contexts out of groups of experiences that cluster together in time. The challenge of context learning is solving the model selection problem: How many contexts make up the rodent's world? Solving this problem requires balancing two opposing goals: minimize the variability of the distribution of experiences within a context and minimize the likelihood of transitioning between contexts. The theory provides an understanding of why hippocampal place cell remapping sometimes develops gradually over many days of experience and why even consistent landmark differences may need to be relearned after other environmental changes. The theory provides an explanation for progressive performance improvements in serial reversal learning, based on a clear dissociation between the incremental process of context learning and the relatively abrupt context selection process. The impact of partial reinforcement on reversal learning is also addressed. Finally, the theory explains why alternating sequence learning does not consistently result in unique context-dependent sequence representations in hippocampus.","{'model': 'tldr@v2.0.0', 'text': 'The theory provides an understanding of why hippocampal place cell remapping sometimes develops gradually over many days of experience and why even consistent landmark differences may need to be relearned after other environmental changes.'}",
-material recognition in the wild with the materials in context database,Sean Bell,"Recognizing materials in real-world images is a challenging task. Real-world materials have rich surface texture, geometry, lighting conditions, and clutter, which combine to make the problem particularly difficult. In this paper, we introduce a new, large-scale, open dataset of materials in the wild, the Materials in Context Database (MINC), and combine this dataset with deep learning to achieve material recognition and segmentation of images in the wild. MINC is an order of magnitude larger than previous material databases, while being more diverse and well-sampled across its 23 categories. Using MINC, we train convolutional neural networks (CNNs) for two tasks: classifying materials from patches, and simultaneous material recognition and segmentation in full images. For patch-based classification on MINC we found that the best performing CNN architectures can achieve 85.2% mean class accuracy. We convert these trained CNN classifiers into an efficient fully convolutional framework combined with a fully connected conditional random field (CRF) to predict the material at every pixel in an image, achieving 73.1% mean class accuracy. Our experiments demonstrate that having a large, well-sampled dataset such as MINC is crucial for real-world material recognition and segmentation.","{'model': 'tldr@v2.0.0', 'text': 'A new, large-scale, open dataset of materials in the wild, the Materials in Context Database (MINC), is introduced, and convolutional neural networks are trained for two tasks: classifying materials from patches, and simultaneous material recognition and segmentation in full images.'}",https://arxiv.org/pdf/1412.0623
-"motivational dynamics in language learning: change, stability, and context.",F. Waninge,"Motivation as a variable in L2 development is no longer seen as the stable individual difference factor it was once believed to be: Influenced by process-oriented models and principles, and especially by the growing understanding of how complex dynamic systems work, researchers have been focusing increasingly on the dynamic and changeable nature of the motivation process. In this study we micro-map the motivational dynamics of 4 language learners during their language lessons over a period of 2 weeks, using a novel instrument-the Motometer-combined with classroom observations and a questionnaire on motivation and attitude. The article answers three current questions concerning L2 motivation: (a) Can we demonstrate variability in students' L2 motivation in class; (b) Is there a detectable stable level of students' in-class motivation; and (c), If both of these are demonstrated, can they be accounted for by the classroom context? The results affirm that student motivation can be successfully explored using a dynamic systems framework. Our findings demonstrate how motivation changes over time on an individual level, while also being characterised by predictable and stable phases, and how it is inseparable from the learner's individual learning context. The data also show that motivation can be meaningfully studied at different interacting time scales. [ABSTRACT FROM AUTHOR]",,
-functional emergence of the hippocampus in context fear learning in infant rats,C. Raineki,"The hippocampus is a part of the limbic system and is important for the formation of associative memories, such as acquiring information about the context (e.g., the place where an experience occurred) during emotional learning (e.g., fear conditioning). Here, we assess whether the hippocampus is responsible for pups' newly emerging context learning. In all experiments, postnatal day (PN) 21 and PN24 rat pups received 10 pairings of odor‐0.5 mA shock or control unpaired odor‐shock, odor only, or shock only. Some pups were used for context, cue or odor avoidance tests, while the remaining pups were used for c‐Fos immunohistochemistry to assess hippocampal activity during acquisition. Our results show that cue and odor avoidance learning were similar at both ages, while contextual fear learning and learning‐associated hippocampal (CA1, CA3, and dentate gyrus) activity (c‐Fos) only occurred in PN24 paired pups. To assess a causal relationship between the hippocampus and context conditioning, we infused muscimol into the hippocampus, which blocked acquisition of context fear learning in the PN24 pups. Muscimol or vehicle infusions did not affect cue learning or aversion to the odor at PN21 or PN24. The results suggest that the newly emerging contextual learning exhibited by PN24 pups is supported by the hippocampus. © 2009 Wiley‐Liss, Inc.","{'model': 'tldr@v2.0.0', 'text': 'The results show that cue and odor avoidance learning were similar at both ages, while contextual fear learning and learning‐associated hippocampal activity (c‐Fos) activity (CA1, CA3, and dentate gyrus) only occurred in PN24 paired pups.'}",https://europepmc.org/articles/pmc2891848?pdf=render
-retrieval-based learning: an episodic context account,Jeffrey D. Karpicke,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter proposes an episodic context account of retrieval-based learning, which explains retrieval practice in terms of context reinstatement, context updating, and restriction of the search set.'}",
-recent trends in deep learning based natural language processing,Tom Young,"Deep learning methods employ multiple processing layers to learn hierarchical representations of data and have produced state-of-the-art results in many domains. Recently, a variety of model designs and methods have blossomed in the context of natural language processing (NLP). In this paper, we review significant deep learning related models and methods that have been employed for numerous NLP tasks and provide a walk-through of their evolution. We also summarize, compare and contrast the various models and put forward a detailed understanding of the past, present and future of deep learning in NLP.","{'model': 'tldr@v2.0.0', 'text': 'This paper reviews significant deep learning related models and methods that have been employed for numerous NLP tasks and provides a walk-through of their evolution.'}",
-the effectiveness of a meaningful learning-based evaluation model for context-aware mobile learning,Yueh-Min Huang,"In recent years, context-aware mobile learning ( CAML) has been widely applied to various fields and has become a popular issue in educational research. Despite the tremendous potential of CAML and its growing significance, continued evaluations and refinements under the advice of field experts and instructors are crucial to ensure its validity, value and sustainability. In this paper, an evaluation model for CAML is developed based on meaningful learning theory using the analytic hierarchy process ( AHP). To verify the effectiveness of the model, three different CAML learning activities are tested, and some experts are invited to evaluate and compare them. As a result, the strengths and weaknesses of each learning activity are obtained. With the aid of the evaluation model proposed in this work, CAML developers can progressively enhance the value of learning activities, thus improving this new approach to learning. [ABSTRACT FROM AUTHOR]","{'model': 'tldr@v2.0.0', 'text': 'With the aid of the evaluation model proposed in this work, CAML developers can progressively enhance the value of learning activities, thus improving this new approach to learning.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/bjet.12147
-education majors' preferences on the functionalities of e-learning platforms in the context of blended learning,N. Tsankov,"The modern stages of higher education development and the actual training of education majors require systematic use of different electronic forms and platforms of education in combination with the traditional educational methods and approaches which will provide students with essential digital skills and competencies, important for their future professional and personal success. Widespread learning management systems provide a common set of basic functionalities. In this study, an assessment of the preferences of education majors on the main functionalities of the electronic platforms used in the context of blended learning in university education is presented. The results reveal a preference on organizational and informational functionalities and less on communication features.","{'model': 'tldr@v2.0.0', 'text': 'An assessment of the preferences of education majors on the main functionalities of the electronic platforms used in the context of blended learning in university education reveals a preference on organizational and informational functionalities and less on communication features.'}",https://online-journals.org/index.php/i-jet/article/download/6971/4427
-context learning and the effect of context on memory retrieval in lymnaea.,James Haney,"Aerial respiratory behavior in Lymnaea was operantly conditioned so that the animals perform aerial respiration significantly less often. Using the standard training procedure (pond water made hypoxic by bubbling N2 through it) both food-deprived and fed animals learned and exhibited long-term memory (LTM). However, food-deprived animals exhibited neither learning nor memory when trained under a condition in which the hypoxic pond water also contained a food odorant (carrot, the food-odorant procedure). Fed animals, however, learned and exhibited LTM with the food-odorant procedure. Thus, the presence of the food odorant per se did not prevent learning or the establishment of LTM. Further experimentation, however, revealed that the ability of the snails to have recall (i.e., memory) for the learned behavior was dependent on the context in which memory was tested. That is, if animals were trained with the food-odorant procedure they could only exhibit recall if tested in the food-odorant context and vice versa with the standard training procedure. Thus, although fed animals could learn and show LTM with either training and testing procedure, LTM could only be seen when they were tested in the context in which they were trained.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the snails to have recall for the learned behavior was dependent on the context in which memory was tested, which meant that if animals were trained with the food-odorant procedure they could only exhibit recall if tested in theFood odorant context and vice versa with the standard training procedure.'}",http://learnmem.cshlp.org/content/8/1/35.full.pdf
-absorptive capacity: a new perspective on learning and innovation,Wesley M. Cohen,"Discusses the notion that the ability to exploit external knowledge is crucial to a firm's innovative capabilities. In addition, it is argued that the ability to evaluate and use outside knowledge is largely a function of the level of prior related knowledge--i.e., absorptive capacity. Prior research has shown that firms that conduct their own research and development (R&D) are better able to use information from external sources. Therefore, it is possible that the absorptive capacity of a firm is created as a byproduct of the firm's R&D investment. A simple model of firm R&D intensity is constructed in a broader context of what applied economists call the three classes of industry-level determinants of R&D intensity: demand, appropriability, and technological opportunity conditions. Several predictions are made, including the notions that absorptive capacity does have a direct effect on R&D spending and spillovers will provide a positive incentive to conduct R&D. All hypotheses are tested using cross-sectional survey data on technological opportunity and appropriability conditions--collected over the period 1975 to 1977 for 1,719 business units--in the American manufacturing sector from Levin et al. (1983, 1987) and the Federal Trade Commission's Line of Business Program data on business unit sales, transfers, and R&D expenditures. Results confirm that firms are sensitive to the characteristics of the learning environment in which they operate and that absorptive capacity does appear to be a part of a firm's decisions regarding resource allocation for innovative activity. Results also suggest that, although the analysis showing a positive effect of spillovers in two industry groups do not represent a direct test of the model, positive absorption incentive associated with spillovers may be sufficiently strong in some cases to more than offset the negative appropribility incentive. (SFL)",,
-epistemic metacognition in context: evaluating and learning online information,L. Mason,,,
-"does pbl matter? relations between instructional context, learning strategies, and learning outcomes",M. Delva,,"{'model': 'tldr@v2.0.0', 'text': ""A learning strategies questionnaire shows that students' learning strategies are influenced by instructional context, and patterns of learning strategy use change over time, and there is tentative evidence that the students'learning strategies influence learning outcomes.""}",
-seamless learning despite context,M. Sharples,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter examines seamless learning, where the aim is to enable a continuous flow of meaning-making despite changes in the physical and social context by inducing a flow state such that learners are so engaged in a mobile learning activity that they lose awareness of their surroundings.'}",http://oro.open.ac.uk/41627/1/Seamless%20Learning%20Despite%20Context%20revised%20clean.pdf
-technology acceptance model in educational context: a systematic literature review,A. Granić,"A respectable amount of work dealing with Technology Acceptance Model (TAM) clearly indicates a popularity of TAM in the field of technology acceptance in general. Nevertheless, there is still a gap in existing knowledge regarding representative academic literature that underlie research on TAM in educational context. The main objective of this systematic literature review is to provide an overview of the current state of research efforts on TAM application in the field of learning and teaching for a variety of learning domains, learning technologies and types of users. Through systematic search by the use of EBSCO Discovery Service, the review has identified 71 relevant studies ranged between 2003 and 2018. The main findings indicate that TAM and its many different versions represent a credible model for facilitating assessment of diverse learning technologies. TAM's core variables, perceived ease of use and perceived usefulness, have been proven to be antecedent factors affecting acceptance of learning with technology. The paper identifies some gaps in current work and suggests areas for further investigation. The results of this systematic review provide a better understanding of TAM acceptance studies in educational context and create a firm foundation for advancing knowledge in the field. Practitioner NotesWhat is already known about this topic Technology acceptance research in teaching and learning context has become an attractive trend.A number of reviews and meta‐analysis focused on specific topics related to technology acceptance in education have been conducted.The Technology Acceptance Model (TAM) is the key model in understanding predictors of human behaviour towards potential acceptance or rejection of the technology.What this paper adds The state of current research on Technology Acceptance Model application in educational context lacks comprehensive reviews addressing variety of learning domains, learning technologies and types of users.The paper presents systematic review of relevant academic literature on Technology Acceptance Model (TAM) in the field of learning and teaching.The paper provides empirical evidence on the predictive validity of the models based on TAM presented in selected literature.The findings revealed that TAM, along with its many different versions called TAM++, is a leading scientific paradigm and credible model for facilitating assessment of diverse technological deployments in educational context.TAM's core variables, perceived ease of use and perceived usefulness, have been proven to be antecedent factors that have affected acceptance of learning with technology.Implications for practice and/or policy The systematic review adds to the body of knowledge and creates a firm foundation for advancing knowledge in the field.By following the most common research objectives and/or by filling current gaps in applied research methods, chosen sample groups and types of result analysis, an own study could be conducted.Future research may well focus on identifying additional external factors that could further explain acceptance and usage of various learning technologies. [ABSTRACT FROM AUTHOR] uracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)","{'model': 'tldr@v2.0.0', 'text': 'An overview of the current state of research efforts on TAM application in the field of learning and teaching for a variety of learning domains, learning technologies and types of users is provided and empirical evidence on the predictive validity of the models based on TAM presented in selected literature is provided.'}",
-learning representations by maximizing mutual information across views,Philip Bachman,"We propose an approach to self-supervised representation learning based on maximizing mutual information between features extracted from multiple views of a shared context. For example, one could produce multiple views of a local spatio-temporal context by observing it from different locations (e.g., camera positions within a scene), and via different modalities (e.g., tactile, auditory, or visual). Or, an ImageNet image could provide a context from which one produces multiple views by repeatedly applying data augmentation. Maximizing mutual information between features extracted from these views requires capturing information about high-level factors whose influence spans multiple views -- e.g., presence of certain objects or occurrence of certain events. Following our proposed approach, we develop a model which learns image representations that significantly outperform prior methods on the tasks we consider. Most notably, using self-supervised learning, our model learns representations which achieve 68.1% accuracy on ImageNet using standard linear evaluation. This beats prior results by over 12% and concurrent results by 7%. When we extend our model to use mixture-based representations, segmentation behaviour emerges as a natural side-effect. Our code is available online: this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a model which learns image representations that significantly outperform prior methods on the tasks the authors consider, and extends this model to use mixture-based representations, where segmentation behaviour emerges as a natural side-effect.'}",
-deep learning for generic object detection: a survey,Li Liu,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey of the recent achievements in this field brought about by deep learning techniques, covering many aspects of generic object detection: detection frameworks, object feature representation, object proposal generation, context modeling, training strategies, and evaluation metrics.'}",https://link.springer.com/content/pdf/10.1007/s11263-019-01247-4.pdf
-toward the use of technology and 21st century teaching-learning approaches: the trend of development in malaysian schools within the context of asia pacific,S. Garba,"ICT Infrastructure and internet connectivity in educational institutions provides learners and teachers the opportunity of adopting 21st century teaching-learning methods that promotes the development of 21st century skills. The availability of internet connectivity in particular provides the platform for a shift from the use of teacher-centered pedagogy (content-based learning) to learner-centered pedagogy (inquiry and project-based learning) that is more interactive and activity oriented. But are teachers utilizing these facilities to advantage? This study investigates the use of ICT and internet resources in relation to the use of 21st century technology-based teaching-learning approaches in Malaysia and Asia Pacific context from a global perspective. Qualitative research approach was used for data collection and analysis in the study. Findings from the study show a unique unexpected changing pattern in the use of computer and internet among school teachers; and new challenges associated with the integration of ICT and the use of 21st century approaches in classroom pedagogical practices were identified. As observed in the study, despite the facilities provided; the technological competence acquired by teachers; and, the increasing use of computer and internet by teachers; the approach to teaching and learning has not change as desired.","{'model': 'tldr@v2.0.0', 'text': 'Despite the facilities provided; the technological competence acquired by teachers; and, the increasing use of computer and internet by teachers, the approach to teaching and learning has not change as desired.'}",https://online-journals.org/index.php/i-jet/article/download/4717/3593
-ppfnet: global context aware local features for robust 3d point matching,Haowen Deng,"We present PPFNet - Point Pair Feature NETwork for deeply learning a globally informed 3D local feature descriptor to find correspondences in unorganized point clouds. PPFNet learns local descriptors on pure geometry and is highly aware of the global context, an important cue in deep learning. Our 3D representation is computed as a collection of point-pair-features combined with the points and normals within a local vicinity. Our permutation invariant network design is inspired by PointNet and sets PPFNet to be ordering-free. As opposed to voxelization, our method is able to consume raw point clouds to exploit the full sparsity. PPFNet uses a novel N-tuple loss and architecture injecting the global information naturally into the local descriptor. It shows that context awareness also boosts the local feature representation. Qualitative and quantitative evaluations of our network suggest increased recall, improved robustness and invariance as well as a vital step in the 3D descriptor extraction performance.","{'model': 'tldr@v2.0.0', 'text': 'Qualitative and quantitative evaluations of the PPFNet network suggest increased recall, improved robustness and invariance as well as a vital step in the 3D descriptor extraction performance.'}",https://arxiv.org/pdf/1802.02669
-digital readiness and its effects on higher education students’ socio-emotional perceptions in the context of the covid-19 pandemic,M. Händel,"Abstract The current study investigated how ready higher education students were for emergency remote teaching due to the COVID-19 pandemic and how this influenced their socio-emotional perceptions. Results of N = 1,826 higher education students indicate that they seem to be ready for digital learning. A k-means cluster analysis revealed two groups of students that significantly differed with respect to their readiness for digital learning (in terms of technology equipment availability, prior experiences with e-learning, and skills for digital learning). Finally, students’ socio-emotional perceptions, that is, stress-related emotions (worries, tension, joy, and overload) as well as social and emotional loneliness significantly differed due to cluster membership. Hence, the study points a need for support of higher education students in successfully coping with the challenges of emergency remote studying.",,https://osf.io/b9pg7/download
-national systems of innovation : towards a theory of innovation and interactive learning,B. Lundvall,"Introduction Theories in the social sciences may be regarded as ‘focusing devices’. Any specific theory brings forward and exposes some aspects of the real world, leaving others in obscurity. That is why a long-lasting hegemony of one single theoretical tradition is damaging, both in terms of understanding and policymaking. In the field of economics, the dominating neoclassical paradigm puts its analytical focus on concepts such as scarcity, allocation and exchange in a static context. Even if these concepts reflect important phenomena in the real world, they only bring forward some aspects of the economic system. One aim of this book is to demonstrate the need for an alternative and supplementary focusing device that puts interactive learning and innovation at the centre of analysis. Through more than a decade, a group of economists at Aalborg University working on a research program on Innovation, Knowledge and Economic Dynamics – the IKE group – has worked together studying industrial development and international competitiveness from such a perspective. This book presents results from this work in relation to one specific subject: national systems of innovation. Our choice of perspective and subject is based on two sets of assumptions. First, it is assumed that the most fundamental resource in the modern economy is knowledge and accordingly that the most important process is learning. The fact that knowledge differs in crucial respects from other resources in the economy makes standard economics less relevant and motivates efforts to develop an alternative paradigm. Second, it is assumed that learning is predominantly an interactive, and therefore a socially embedded, process that cannot be understood without taking into consideration its institutional and cultural context. Specifically, it is assumed that the historical establishment and development of the modern nation state was a necessary prerequisite for the acceleration of the process of learning, which propelled the process of industrialization during the last centuries. Finally, it is recognized that the traditional role of nation states in supporting learning processes is now challenged by the process of internationalization and globalization.",,https://opus.bibliothek.uni-augsburg.de/opus4/files/85408/85408.pdf
-optimization methods for large-scale machine learning,L. Bottou,"This paper provides a review and commentary on the past, present, and future of numerical optimization algorithms in the context of machine learning applications. Through case studies on text classification and the training of deep neural networks, we discuss how optimization problems arise in machine learning and what makes them challenging. A major theme of our study is that large-scale machine learning represents a distinctive setting in which the stochastic gradient (SG) method has traditionally played a central role while conventional gradient-based nonlinear optimization techniques typically falter. Based on this viewpoint, we present a comprehensive theory of a straightforward, yet versatile SG algorithm, discuss its practical behavior, and highlight opportunities for designing algorithms with improved performance. This leads to a discussion about the next generation of optimization methods for large-scale machine learning, including an investigation of two main streams of research on techniques that diminish noise in the stochastic directions and methods that make use of second-order derivative approximations.","{'model': 'tldr@v2.0.0', 'text': 'A major theme of this study is that large-scale machine learning represents a distinctive setting in which the stochastic gradient method has traditionally played a central role while conventional gradient-based nonlinear optimization techniques typically falter, leading to a discussion about the next generation of optimization methods for large- scale machine learning.'}",https://arxiv.org/pdf/1606.04838
-apprenticeship in thinking: cognitive development in social context,B. Rogoff,PART I: The individual and the sociocultural context: Conceiving the relationship of the social world and the individual The sociocultural context of cognitive activity PART II: Processes of guided participation: Providing bridges from known to new Structuring situations and transferring responsibility Cultural universals and variations in guided participation PART III: Cognitive development through interaction with adults and peers: Explanations for cognitive development through social interaction: Vygotsky and Piaget Evidence of learning from guided participation with adults Peer interaction and cognitive development Shared thinking and guided participation.,,
-image-based surface defect detection using deep learning: a review,P. Bhatt," Automatically detecting surface defects from images is an essential capability in manufacturing applications. Traditional image processing techniques are useful in solving a specific class of problems. However, these techniques do not handle noise, variations in lighting conditions, and backgrounds with complex textures. In recent times, deep learning has been widely explored for use in automation of defect detection. This survey article presents three different ways of classifying various efforts in literature for surface defect detection using deep learning techniques. These three ways are based on defect detection context, learning techniques, and defect localization and classification method respectively. This article also identifies future research directions based on the trends in the deep learning area.","{'model': 'tldr@v2.0.0', 'text': 'Three different ways of classifying various efforts in literature for surface defect detection using deep learning techniques are presented, based on defect detection context, learning techniques, and defect localization and classification method respectively.'}",
-unsupervised learning of visual representations by solving jigsaw puzzles,M. Noroozi,,"{'model': 'tldr@v2.0.0', 'text': 'A novel unsupervised learning approach to build features suitable for object detection and classification and to facilitate the transfer of features to other tasks, the context-free network (CFN), a siamese-ennead convolutional neural network is introduced.'}",https://arxiv.org/pdf/1603.09246
-context-specific proportion congruency effects: an episodic learning account and computational model,James R. Schmidt,"In the Stroop task, participants identify the print color of color words. The congruency effect is the observation that response times and errors are increased when the word and color are incongruent (e.g., the word “red” in green ink) relative to when they are congruent (e.g., “red” in red). The proportion congruent (PC) effect is the finding that congruency effects are reduced when trials are mostly incongruent rather than mostly congruent. This PC effect can be context-specific. For instance, if trials are mostly incongruent when presented in one location and mostly congruent when presented in another location, the congruency effect is smaller for the former location. Typically, PC effects are interpreted in terms of strategic control of attention in response to conflict, termed conflict adaptation or conflict monitoring. In the present manuscript, however, an episodic learning account is presented for context-specific proportion congruent (CSPC) effects. In particular, it is argued that context-specific contingency learning can explain part of the effect, and context-specific rhythmic responding can explain the rest. Both contingency-based and temporal-based learning can parsimoniously be conceptualized within an episodic learning framework. An adaptation of the Parallel Episodic Processing model is presented. This model successfully simulates CSPC effects, both for contingency-biased and contingency-unbiased (transfer) items. The same fixed-parameter model can explain a range of other findings from the learning, timing, binding, practice, and attentional control domains.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that context- specific contingency learning can explain part of the effect, and context-specific rhythmic responding can explain the rest, as well as an adaptation of the Parallel Episodic Processing model that successfully simulates CSPC effects.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2016.01806/pdf
-applications of machine learning in drug discovery and development,J. Vamathevan,,"{'model': 'tldr@v2.0.0', 'text': 'The most useful techniques and how machine learning can promote data-driven decision making in drug discovery and development are discussed and major hurdles in the field are highlighted, such as the required data characteristics for applying ML.'}",https://europepmc.org/articles/pmc6552674?pdf=render
-storytelling: a strategy for providing context for learning.,D. Billings,"Storytelling--a narrative of events related to nursing and linked to evidence--provides a context for learning, particularly for learners who require a rich context to understand and integrate concepts related to patient care. This article offers suggestions for developing and using stories in nursing education.","{'model': 'tldr@v2.0.0', 'text': 'Suggestions for developing and using stories in nursing education are offered, particularly for learners who require a rich context to understand and integrate concepts related to patient care.'}",
-framing ubiquitous mobility educationally: mobile devices and context-aware learning,B. Bachmair,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter argues that mobile devices and the artefacts accessed through and created with them should be viewed as important cultural resources and, therefore, constitute valid resources for learning.'}",
-senticnet 5: discovering conceptual primitives for sentiment analysis by means of context embeddings,E. Cambria," With the recent development of deep learning, research in AI has gained new vigor and prominence. While machine learning has succeeded in revitalizing many research fields, such as computer vision, speech recognition, and medical diagnosis, we are yet to witness impressive progress in natural language understanding. One of the reasons behind this unmatched expectation is that, while a bottom-up approach is feasible for pattern recognition, reasoning and understanding often require a top-down approach. In this work, we couple sub-symbolic and symbolic AI to automatically discover conceptual primitives from text and link them to commonsense concepts and named entities in a new three-level knowledge representation for sentiment analysis. In particular, we employ recurrent neural networks to infer primitives by lexical substitution and use them for grounding common and commonsense knowledge by means of multi-dimensional scaling. ","{'model': 'tldr@v2.0.0', 'text': 'This work couple sub-symbolic and symbolic AI to automatically discover conceptual primitives from text and link them to commonsense concepts and named entities in a new three-level knowledge representation for sentiment analysis.'}",https://ojs.aaai.org/index.php/AAAI/article/download/11559/11418
-large-scale machine learning with stochastic gradient descent,L. Bottou,,"{'model': 'tldr@v2.0.0', 'text': 'A more precise analysis uncovers qualitatively different tradeoffs for the case of small-scale and large-scale learning problems.'}",http://leon.bottou.org/publications/pdf/compstat-2010.pdf
-revisiting semi-supervised learning with graph embeddings,Zhilin Yang,"We present a semi-supervised learning framework based on graph embeddings. Given a graph between instances, we train an embedding for each instance to jointly predict the class label and the neighborhood context in the graph. We develop both transductive and inductive variants of our method. In the transductive variant of our method, the class labels are determined by both the learned embeddings and input feature vectors, while in the inductive variant, the embeddings are defined as a parametric function of the feature vectors, so predictions can be made on instances not seen during training. On a large and diverse set of benchmark tasks, including text classification, distantly supervised entity extraction, and entity classification, we show improved performance over many of the existing models.","{'model': 'tldr@v2.0.0', 'text': 'On a large and diverse set of benchmark tasks, including text classification, distantly supervised entity extraction, and entity classification, the proposed semi-supervised learning framework shows improved performance over many of the existing models.'}",
-context-aware recommender systems for learning: a survey and future challenges,K. Verbert,"Recommender systems have been researched extensively by the Technology Enhanced Learning (TEL) community during the last decade. By identifying suitable resources from a potentially overwhelming variety of choices, such systems offer a promising approach to facilitate both learning and teaching tasks. As learning is taking place in extremely diverse and rich environments, the incorporation of contextual information about the user in the recommendation process has attracted major interest. Such contextualization is researched as a paradigm for building intelligent systems that can better predict and anticipate the needs of users, and act more efficiently in response to their behavior. In this paper, we try to assess the degree to which current work in TEL recommender systems has achieved this, as well as outline areas in which further work is needed. First, we present a context framework that identifies relevant context dimensions for TEL applications. Then, we present an analysis of existing TEL recommender systems along these dimensions. Finally, based on our survey results, we outline topics on which further research is needed.","{'model': 'tldr@v2.0.0', 'text': 'A context framework that identifies relevant context dimensions for TEL applications is presented and an analysis of existing TEL recommender systems along these dimensions is presented, and topics on which further research is needed are outlined.'}",https://lirias.kuleuven.be/bitstream/123456789/338644/3/survey-final.pdf
-vq-wav2vec: self-supervised learning of discrete speech representations,Alexei Baevski,We propose vq-wav2vec to learn discrete representations of audio segments through a wav2vec-style self-supervised context prediction task. The algorithm uses either a gumbel softmax or online k-means clustering to quantize the dense representations. Discretization enables the direct application of algorithms from the NLP community which require discrete inputs. Experiments show that BERT pre-training achieves a new state of the art on TIMIT phoneme classification and WSJ speech recognition.,"{'model': 'tldr@v2.0.0', 'text': 'Experiments show that BERT pre-training achieves a new state of the art on TIMIT phoneme classification and WSJ speech recognition and the algorithm uses a gumbel softmax or online k-means clustering to quantize the dense representations.'}",
-learning context-aware latent representations for context-aware collaborative filtering,Xin Liu,"In this paper, we propose a generic framework to learn context-aware latent representations for context-aware collaborative filtering. Contextual contents are combined via a function to produce the context influence factor, which is then combined with each latent factor to derive latent representations. We instantiate the generic framework using biased Matrix Factorization as the base model. A Stochastic Gradient Descent (SGD) based optimization procedure is developed to fit the model by jointly learning the weight of each context and latent factors. Experiments conducted over three real-world datasets demonstrate that our model significantly outperforms not only the base model but also the representative context-aware recommendation models.","{'model': 'tldr@v2.0.0', 'text': 'A Stochastic Gradient Descent based optimization procedure is developed to fit the model by jointly learning the weight of each context and latent factors, which significantly outperforms not only the base model but also the representative context-aware recommendation models.'}",
-dynamic edge association and resource allocation in self-organizing hierarchical federated learning networks,Wei Yang Bryan Lim,"Federated Learning (FL) is a promising privacy-preserving distributed machine learning paradigm. However, communication inefficiency remains the key bottleneck that impedes its large-scale implementation. Recently, hierarchical FL (HFL) has been proposed in which data owners, i.e., workers, can first transmit their updated model parameters to edge servers for intermediate aggregation. This reduces the instances of global communication and straggling workers. To enable efficient HFL, it is important to address the issues of edge association and resource allocation in the context of non-cooperative players, i.e., workers, edge servers, and model owner. However, the existing studies merely focus on static approaches and do not consider the dynamic interactions and bounded rationalities of the players. In this paper, we propose a hierarchical game framework to study the dynamics of edge association and resource allocation in self-organizing HFL networks. In the lower-level game, the edge association strategies of the workers are modelled using an evolutionary game. In the upper-level game, a Stackelberg differential game is adopted in which the model owner decides an optimal reward scheme given the expected bandwidth allocation control strategy of the edge server. Finally, we provide numerical results to validate that our proposed framework captures the HFL system dynamics under varying sources of network heterogeneity.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hierarchical game framework to study the dynamics of edge association and resource allocation in self-organizing HFL networks and provides numerical results to validate that the proposed framework captures the HFL system dynamics under varying sources of network heterogeneity.'}",https://dr.ntu.edu.sg/bitstream/10356/156039/2/Dynamic%20Edge%20Association%20and%20Resource%20Allocation%20in%20Self-Organizing%20Hierarchical%20Federated%20Learning%20Networks.pdf
-data shapley: equitable valuation of data for machine learning,Amirata Ghorbani,"As data becomes the fuel driving technological and economic growth, a fundamental challenge is how to quantify the value of data in algorithmic predictions and decisions. For example, in healthcare and consumer markets, it has been suggested that individuals should be compensated for the data that they generate, but it is not clear what is an equitable valuation for individual data. In this work, we develop a principled framework to address data valuation in the context of supervised machine learning. Given a learning algorithm trained on $n$ data points to produce a predictor, we propose data Shapley as a metric to quantify the value of each training datum to the predictor performance. Data Shapley value uniquely satisfies several natural properties of equitable data valuation. We develop Monte Carlo and gradient-based methods to efficiently estimate data Shapley values in practical settings where complex learning algorithms, including neural networks, are trained on large datasets. In addition to being equitable, extensive experiments across biomedical, image and synthetic data demonstrate that data Shapley has several other benefits: 1) it is more powerful than the popular leave-one-out or leverage score in providing insight on what data is more valuable for a given learning task; 2) low Shapley value data effectively capture outliers and corruptions; 3) high Shapley value data inform what type of new data to acquire to improve the predictor.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a principled framework to address data valuation in the context of supervised machine learning by proposing data Shapley as a metric to quantify the value of each training datum to the predictor performance.'}",
-cultural learning,R. Holmes,"Chapter 12 explores the ways we learn about our culture, how culture shapes learning, and different learning contexts. It discusses attachment, Bowlby’s attachment theory, the Strange Situation, attachment styles, challenges to Western attachment theories, and attachment and culture. It addresses parenting, Baumrind’s parenting styles, parental ethnotheories, and parenting in different cultural contexts. It explores learning in formal settings like school, the connection between culture and schooling, parent beliefs and academic outcomes, apprenticeships, and learning in informal settings. Finally, it investigates play as a context for cultural learning and includes culture-specific and cross-cultural studies. This chapter includes a case study, Culture Across Disciplines box, chapter summary, key terms, a What Do Other Disciplines Do? section, thought-provoking questions, and class and experiential activities.",,
-qt-opt: scalable deep reinforcement learning for vision-based robotic manipulation,Dmitry Kalashnikov,"In this paper, we study the problem of learning vision-based dynamic manipulation skills using a scalable reinforcement learning approach. We study this problem in the context of grasping, a longstanding challenge in robotic manipulation. In contrast to static learning behaviors that choose a grasp point and then execute the desired grasp, our method enables closed-loop vision-based control, whereby the robot continuously updates its grasp strategy based on the most recent observations to optimize long-horizon grasp success. To that end, we introduce QT-Opt, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real-world grasping that generalizes to 96% grasp success on unseen objects. Aside from attaining a very high success rate, our method exhibits behaviors that are quite distinct from more standard grasping systems: using only RGB vision-based perception from an over-the-shoulder camera, our method automatically learns regrasping strategies, probes objects to find the most effective grasps, learns to reposition objects and perform other non-prehensile pre-grasp manipulations, and responds dynamically to disturbances and perturbations.","{'model': 'tldr@v2.0.0', 'text': 'QT-Opt is introduced, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real- world grasping that generalizes to 96% grasp success on unseen objects.'}",
-cerebellar contribution to context processing in extinction learning and recall,Dae-In Chang,,"{'model': 'tldr@v2.0.0', 'text': 'Because the ventromedial prefrontal cortex (vmPFC) is critically involved in context-related processes during extinction retrieval, and there are no known connections between the cerebellum and the vmPFC, the Cerebellum may be less important during extinction recall.'}",
-the informational value of contexts affects context-dependent learning,Sara Lucke,,"{'model': 'tldr@v2.0.0', 'text': 'The results of two predictive-learning experiments support the idea that relevant contexts receive more attention, leading to stronger context specificity of learning.'}",https://link.springer.com/content/pdf/10.3758%2Fs13420-013-0104-z.pdf
-the da antagonist tiapride impairs context-related extinction learning in a novel context without affecting renewal,S. Lissek,"Renewal describes the recovery of an extinguished response if recall is tested in a context different from the extinction context. Behavioral studies demonstrated that attention to relevant context strengthens renewal. Neurotransmitters mediating attention and learning such as the dopaminergic (DA) system presumably modulate extinction learning and renewal. However, the role of DA for non-fear-based extinction learning and renewal in humans has not yet been investigated. This fMRI study investigated effects of DA-antagonism upon context-related extinction in a predictive learning task in which extinction occurred either in a novel (ABA) or an unchanged (AAA) context. The tiapride-treated group (TIA) showed significantly impaired ABA extinction learning and a significant within-group difference between ABA and AAA extinction, compared to placebo (PLAC). Groups did not differ in their level of ABA renewal. In ABA extinction, TIA showed reduced activation in dlPFC and OFC, hippocampus, and temporal regions. Across groups, activation in PFC and hippocampus correlated negatively with ABA extinction errors. Results suggest that in context-related extinction learning DA in PFC and hippocampus is involved in readjusting the cue-outcome relationship in the presence of a novel context. However, relating context to the appropriate association during recall does not appear to rely exclusively on DA signaling.","{'model': 'tldr@v2.0.0', 'text': 'Results suggest that in context-related extinction learning DA in PFC and hippocampus is involved in readjusting the cue-outcome relationship in the presence of a novel context, however, relating context to the appropriate association during recall does not appear to rely exclusively on DA signaling.'}",
-"doing a thematic analysis: a practical, step-by-step guide for learning and teaching scholars.",Moira Maguire,"Data analysis is central to credible qualitative research. Indeed the qualitative researcher is often described as the research instrument insofar as his or her ability to understand, describe and interpret experiences and perceptions is key to uncovering meaning in particular circumstances and contexts. While much has been written about qualitative analysis from a theoretical perspective we noticed that often novice, and even more experienced researchers, grapple with the ‘how’ of qualitative analysis. Here we draw on Braun and Clarke’s (2006) framework and apply it in a systematic manner to describe and explain the process of analysis within the context of learning and teaching research. We illustrate the process using a worked example based on (with permission) a short extract from a focus group interview, conducted with undergraduate students.",,
-context change and associative learning.,Juan M. Rosas,"This article reviews the effects of changing the background context on performance in associative learning tasks in humans and animals. The findings are complementary and consistent over animal conditioning (Pavlovian and instrumental learning) and human predictive learning and memory paradigms. In many cases, a context change after learning can have surprisingly little disruptive influence on performance. Extinction, or retroactive interference treatments more generally, is more context-specific than the initial learning. Contexts become important if the participant is exposed to any of several treatments that involve prediction error, which may serve to increase attention to the context. Contexts also become important if they are given predictive or informational value. Studies of instrumental (operant) learning are further consistent with the idea that the context might also influence affordances that support voluntary actions. Context switch effects are not universal, but mainly occur when certain attention and perception processes can come into play. WIREs Cogn Sci 2013, 4:237-244. doi: 10.1002/wcs.1225 This article is categorized under: Psychology > Learning.","{'model': 'tldr@v2.0.0', 'text': 'The findings are complementary and consistent over animal conditioning and human predictive learning and memory paradigms and further consistent with the idea that the context might also influence affordances that support voluntary actions.'}",https://europepmc.org/articles/pmc3680141?pdf=render
-context and repetition in word learning,Jessica S. Horst,"Young children learn words from a variety of situations, including shared storybook reading. A recent study by Horst et al. (2011a) demonstrates that children learned more new words during shared storybook reading if they were read the same stories repeatedly than if they were read different stories that had the same number of target words. The current paper reviews this study and further examines the effect of contextual repetition on children's word learning in both shared storybook reading and other situations, including fast mapping by mutual exclusivity. The studies reviewed here suggest that the same cognitive mechanisms support word learning in a variety of situations. Both practical considerations for experimental design and directions for future research are discussed.","{'model': 'tldr@v2.0.0', 'text': 'The studies reviewed here suggest that the same cognitive mechanisms support word learning in a variety of situations, including fast mapping by mutual exclusivity in shared storybook reading and other situations.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00149/pdf
-a bayesian context fear learning algorithm/automaton,F. Krasne,"Contextual fear conditioning is thought to involve the synaptic plasticity-dependent establishment in hippocampus of representations of to-be-conditioned contexts which can then become associated with USs in the amygdala. A conceptual and computational model of this process is proposed in which contextual attributes are assumed to be sampled serially and randomly during contextual exposures. Given this assumption, moment-to-moment information about such attributes will often be quite different from one exposure to another and, in particular, between exposures during which representations are created, exposures during which conditioning occurs, and during recall sessions. This presents challenges to current conceptual models of hippocampal function. In order to meet these challenges, our model's hippocampus was made to operate in different modes during representation creation and recall, and non-hippocampal machinery was constructed that controlled these hippocampal modes. This machinery uses a comparison between contextual information currently observed and information associated with existing hippocampal representations of familiar contexts to compute the Bayesian Weight of Evidence that the current context is (or is not) a known one, and it uses this value to assess the appropriateness of creation or recall modes. The model predicts a number of known phenomena such as the immediate shock deficit, spurious fear conditioning to contexts that are absent but similar to actually present ones, and modulation of conditioning by pre-familiarization with contexts. It also predicts a number of as yet unknown phenomena.","{'model': 'tldr@v2.0.0', 'text': 'A conceptual and computational model of contextual fear conditioning that uses a comparison between contextual information currently observed and information associated with existing hippocampal representations of familiar contexts to compute the Bayesian Weight of Evidence that the current context is (or is not) a known one.'}",https://www.frontiersin.org/articles/10.3389/fnbeh.2015.00112/pdf
-context-specific temporal learning with non-conflict stimuli: proof-of-principle for a learning account of context-specific proportion congruent effects,James R. Schmidt,"The conflict adaptation account proposes that participants adjust attention to target and distracting stimuli in response to conflict. This is argued to explain the proportion congruent effect, wherein the congruency effect decreases as the proportion of conflicting incongruent trials increases. Some reports further argue that this conflict adaptation process can be context-specific. This paper presents a proof-of-principle for a competing account. It is suggested that such context-specific effects might be driven by very basic temporal learning processes. In the reported experiment, we manipulated stimulus contrast in place of congruency. In one location, stimulus letters were mostly easy to identify (high stimulus contrast). In the other location, letters were mostly hard to identify (low stimulus contrast). Participants produced a larger contrast effect in the mostly easy context. Along with supplemental analyses investigating the role of context switching and previous trial response times, the results are consistent with the notion that different rhythms of responding are learned for an easy versus hard location context. These results suggest that context-specific proportion congruency effects might result, in whole or in part, from temporal learning. Conflict adaptation may or may not play an additional role.","{'model': 'tldr@v2.0.0', 'text': 'A proof-of-principle for a competing account of conflict adaptation suggests that context-specific proportion congruency effects might result, in whole or in part, from temporal learning.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2014.01241/pdf
-making connections: learning and teaching chemistry in context,Donna King,,"{'model': 'tldr@v2.0.0', 'text': 'This narrative inquiry addresses the question: how do the recalled experiences of a student and her teacher in context-based and concept-based chemistry programs compare, and constructed a collective account around four themes; namely, the extent to which the student makes connections between chemistry concepts and real-world contexts.'}",https://eprints.qut.edu.au/14509/1/14509.pdf
-situated learning,,"Situated learning or situated cognition is a learning theory developed in the late 1980s by Jean Lave and Etienne Wenger and soon expanded by John Seely Brown and his colleagues. Situated learning theory is based on the assumption that knowledge should be presented in authentic context that involves its application. Both authors argue that learning should not be viewed as transmission of abstract and contextualized knowledge between individuals, but a social process within certain conditions which include activity, context and culture.",,
-from lecture to learning tasks: use of the 4c/id model in a communication skills course in a continuing professional education context.,A. P. Susilo,"This article describes the use of four-component instructional design (4C/ID), a model to plan educational interventions for complex learning. This model was used to design a continuing education course on communication skills for health professionals in a context that is hierarchical and communal. The authors describe the 4C/ID model and provide an example of its application in designing the course. In the 4C/ID model, learning tasks serve as the backbone of the course, with lectures and other supportive information organized around them. The 4C/ID model is different from traditional models that base the course on lectures on different topics and connect part-task assignments to these topics. The use of the 4C/ID model to develop the educational intervention moves the paradigm from lectures to learning tasks to better prepare learners for real practice.","{'model': 'tldr@v2.0.0', 'text': 'This model was used to design a continuing education course on communication skills for health professionals in a context that is hierarchical and communal and moves the paradigm from lectures to learning tasks to better prepare learners for real practice.'}",
-"continuity, commitment and context: adult siblings of people with autism plus learning disability.",R. Tozer,"Sibling relationships are usually lifelong and reciprocal. They can assume particular significance when a brother or sister has a learning disability. Until recently, adult siblings of people with disabilities such as severe autism have been ignored by policy, practice and research. This qualitative study contributes to an emerging literature by exploring how adult siblings, who have a brother or sister with autism (plus learning disability) and living in England, give meaning to their family (and caring) relationships and engage with service delivery. We spoke to 21 adult siblings using semi-structured interviews and met with 12 of their siblings with autism. Our analysis, using a broad narrative approach, demonstrates the continuity of the sibling relationship and an enduring personalised commitment. The nature of this relationship, however, is sensitive to context. How non-disabled adult siblings relate to their childhood experience is fundamental when making sense of this, as is their need to fulfil other social and family obligations, alongside their 'sense of duty' to support their disabled brother or sister. Sibling experience was further mediated by negotiating their 'perceived invisibility' in social care policy and practice. Our work concludes that by understanding the way relationships between siblings have developed over time, adult siblings' contribution to the lives of their brother or sister with autism can be better supported for the benefit of both parties. Such an approach would support current policy developments.","{'model': 'tldr@v2.0.0', 'text': ""By understanding the way relationships between siblings have developed over time, adult siblings' contribution to the lives of their brother or sister with autism can be better supported for the benefit of both parties.""}",
-deep learning with limited numerical precision,Suyog Gupta,"Training of large-scale deep neural networks is often constrained by the available computational resources. We study the effect of limited precision data representation and computation on neural network training. Within the context of low-precision fixed-point computations, we observe the rounding scheme to play a crucial role in determining the network's behavior during training. Our results show that deep networks can be trained using only 16-bit wide fixed-point number representation when using stochastic rounding, and incur little to no degradation in the classification accuracy. We also demonstrate an energy-efficient hardware accelerator that implements low-precision fixed-point arithmetic with stochastic rounding.","{'model': 'tldr@v2.0.0', 'text': 'The results show that deep networks can be trained using only 16-bit wide fixed-point number representation when using stochastic rounding, and incur little to no degradation in the classification accuracy.'}",
-a survey on ensemble learning,Xibin Dong,,"{'model': 'tldr@v2.0.0', 'text': 'Challenges and possible research directions for each mainstream approach of ensemble learning are presented and an extra introduction is given for the combination of ensemblelearning with other machine learning hot spots such as deep learning, reinforcement learning, etc.'}",
-learning in context,Betsy Disalvo,"The authors present an exploratory study of Black middle school boys who play digital games. The study was conducted through observations and interviews with Black American middle school boys about digital games as an informal learning experience. The first goal of the study is to understand the cultural context that Black students from economically disadvantaged inner-city neighborhoods bring to playing digital games. The second goal of the study is to examine how this cultural context affects the learning opportunities with games. Third, the authors examine how differences in game play are potential factors in the discrepancy between White male gamers and Black male gamers in science, technology, engineering, and mathematics. Finally, the authors address several opportunities within the field of informal learning to augment game play by bridging the learning that takes place within game play to the real world.",,
-"classroom learning environments: retrospect, context and prospect",B. Fraser,,,
-struc2vec: learning node representations from structural identity,Leonardo F. R. Ribeiro,"Structural identity is a concept of symmetry in which network nodes are identified according to the network structure and their relationship to other nodes. Structural identity has been studied in theory and practice over the past decades, but only recently has it been addressed with representational learning techniques. This work presents struc2vec, a novel and flexible framework for learning latent representations for the structural identity of nodes. struc2vec uses a hierarchy to measure node similarity at different scales, and constructs a multilayer graph to encode structural similarities and generate structural context for nodes. Numerical experiments indicate that state-of-the-art techniques for learning node representations fail in capturing stronger notions of structural identity, while struc2vec exhibits much superior performance in this task, as it overcomes limitations of prior approaches. As a consequence, numerical experiments indicate that struc2vec improves performance on classification tasks that depend more on structural identity.","{'model': 'tldr@v2.0.0', 'text': 'Struc2vec, a novel and flexible framework for learning latent representations for the structural identity of nodes, is presented, which improves performance on classification tasks that depend more on structural identity.'}",https://arxiv.org/pdf/1704.03165
-a review of self-regulated learning: six models and four directions for research,E. Panadero,"Self-regulated learning (SRL) includes the cognitive, metacognitive, behavioral, motivational, and emotional/affective aspects of learning. It is, therefore, an extraordinary umbrella under which a considerable number of variables that influence learning (e.g., self-efficacy, volition, cognitive strategies) are studied within a comprehensive and holistic approach. For that reason, SRL has become one of the most important areas of research within educational psychology. In this paper, six models of SRL are analyzed and compared; that is, Zimmerman; Boekaerts; Winne and Hadwin; Pintrich; Efklides; and Hadwin, Järvelä and Miller. First, each model is explored in detail in the following aspects: (a) history and development, (b) description of the model (including the model figures), (c) empirical support, and (d) instruments constructed based on the model. Then, the models are compared in a number of aspects: (a) citations, (b) phases and subprocesses, (c) how they conceptualize (meta)cognition, motivation and emotion, (d) top–down/bottom–up, (e) automaticity, and (f) context. In the discussion, the empirical evidence from the existing SRL meta-analyses is examined and implications for education are extracted. Further, four future lines of research are proposed. The review reaches two main conclusions. First, the SRL models form an integrative and coherent framework from which to conduct research and on which students can be taught to be more strategic and successful. Second, based on the available meta-analytic evidence, there are differential effects of SRL models in light of differences in students’ developmental stages or educational levels. Thus, scholars and teachers need to start applying these differential effects of the SRL models and theories to enhance students’ learning and SRL skills.","{'model': 'tldr@v2.0.0', 'text': 'The SRL models form an integrative and coherent framework from which to conduct research and on which students can be taught to be more strategic and successful in order to enhance students’ learning and SRL skills.'}",https://www.frontiersin.org/articles/10.3389/fpsyg.2017.00422/pdf
-transformers with convolutional context for asr,Abdel-rahman Mohamed,"The recent success of transformer networks for neural machine translation and other NLP tasks has led to a surge in research work trying to apply it for speech recognition. Recent efforts studied key research questions around ways of combining positional embedding with speech features, and stability of optimization for large scale learning of transformer networks. In this paper, we propose replacing the sinusoidal positional embedding for transformers with convolutionally learned input representations. These contextual representations provide subsequent transformer blocks with relative positional information needed for discovering long-range relationships between local concepts. The proposed system has favorable optimization characteristics where our reported results are produced with fixed learning rate of 1.0 and no warmup steps. The proposed model achieves a competitive 4.7% and 12.9% WER on the Librispeech ``test clean'' and ``test other'' subsets when no extra LM text is provided.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes replacing the sinusoidal positional embedding for transformers with convolutionally learned input representations that provide subsequent transformer blocks with relative positional information needed for discovering long-range relationships between local concepts.'}",
-rotation-insensitive and context-augmented object detection in remote sensing images,Ke Li,"Most of the existing deep-learning-based methods are difficult to effectively deal with the challenges faced for geospatial object detection such as rotation variations and appearance ambiguity. To address these problems, this paper proposes a novel deep-learning-based object detection framework including region proposal network (RPN) and local-contextual feature fusion network designed for remote sensing images. Specifically, the RPN includes additional multiangle anchors besides the conventional multiscale and multiaspect-ratio ones, and thus can deal with the multiangle and multiscale characteristics of geospatial objects. To address the appearance ambiguity problem, we propose a double-channel feature fusion network that can learn local and contextual properties along two independent pathways. The two kinds of features are later combined in the final layers of processing in order to form a powerful joint representation. Comprehensive evaluations on a publicly available ten-class object detection data set demonstrate the effectiveness of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel deep-learning-based object detection framework including region proposal network (RPN) and local-contextual feature fusion network designed for remote sensing images that can deal with the multiangle and multiscale characteristics of geospatial objects.'}",
-learning image context for segmentation of the prostate in ct-guided radiotherapy,Wei Li,"Accurate segmentation of the prostate is the key to the success of external beam radiotherapy of prostate cancer. However, accurate segmentation of the prostate in computer tomography (CT) images remains challenging mainly due to three factors: (1) low image contrast between the prostate and its surrounding tissues, (2) unpredictable prostate motion across different treatment days and (3) large variations of intensities and shapes of the bladder and rectum around the prostate. In this paper, an online-learning and patient-specific classification method based on the location-adaptive image context is presented to deal with all these challenging issues and achieve the precise segmentation of the prostate in CT images. Specifically, two sets of location-adaptive classifiers are placed, respectively, along the two coordinate directions of the planning image space of a patient, and further trained with the planning image and also the previous-segmented treatment images of the same patient to jointly perform prostate segmentation for a new treatment image (of the same patient). In particular, each location-adaptive classifier, which itself consists of a set of sequential sub-classifiers, is recursively trained with both the static image appearance features and the iteratively updated image context features (extracted at different scales and orientations) for better identification of each prostate region. The proposed learning-based prostate segmentation method has been extensively evaluated on 161 images of 11 patients, each with more than nine daily treatment three-dimensional CT images. Our method achieves the mean Dice value 0.908 and the mean ± SD of average surface distance value 1.40 ± 0.57 mm. Its performance is also compared with three prostate segmentation methods, indicating the best segmentation accuracy by the proposed method among all methods under comparison.","{'model': 'tldr@v2.0.0', 'text': 'The proposed learning-based prostate segmentation method has been extensively evaluated and achieves the mean Dice value 0.908 and the mean ± SD of average surface distance value 1.40 ± 0.57 mm, indicating the best segmentation accuracy by the proposed method among all methods under comparison.'}",https://europepmc.org/articles/pmc3378724?pdf=render
-"federated learning for wireless communications: motivation, opportunities, and challenges",Solmaz Niknam,"There is a growing interest in the wireless communications community to complement the traditional model-driven design approaches with data-driven machine learning (ML)-based solutions. While conventional ML approaches rely on the assumption of having the data and processing heads in a central entity, this is not always feasible in wireless communications applications because of the inaccessibility of private data and large communication overhead required to transmit raw data to central ML processors. As a result, decentralized ML approaches that keep the data where it is generated are much more appealing. Due to its privacy-preserving nature, federated learning is particularly relevant for many wireless applications, especially in the context of fifth generation (5G) networks. In this article, we provide an accessible introduction to the general idea of federated learning, discuss several possible applications in 5G networks, and describe key technical challenges and open problems for future research on federated learning in the context of wireless communications.","{'model': 'tldr@v2.0.0', 'text': 'An accessible introduction to the general idea of federated learning is provided, several possible applications in 5G networks are discussed, and key technical challenges and open problems for future research on Federated learning in the context of wireless communications are described.'}",https://arxiv.org/pdf/1908.06847
-do no harm: a roadmap for responsible machine learning for health care,J. Wiens,,"{'model': 'tldr@v2.0.0', 'text': 'In this Perspective, the authors present a framework, context and guidelines for accelerating the translation of machine-learning-based interventions in health care.'}",
-learning to reinforcement learn,Jane X. Wang,"In recent years deep reinforcement learning (RL) systems have attained superhuman performance in a number of challenging task domains. However, a major limitation of such applications is their demand for massive amounts of training data. A critical present objective is thus to develop deep RL methods that can adapt rapidly to new tasks. In the present work we introduce a novel approach to this challenge, which we refer to as deep meta-reinforcement learning. Previous work has shown that recurrent networks can support meta-learning in a fully supervised context. We extend this approach to the RL setting. What emerges is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure. This second, learned RL algorithm can differ from the original one in arbitrary ways. Importantly, because it is learned, it is configured to exploit structure in the training domain. We unpack these points in a series of seven proof-of-concept experiments, each of which examines a key aspect of deep meta-RL. We consider prospects for extending and scaling up the approach, and also point out some potentially important implications for neuroscience.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel approach to deep meta-reinforcement learning, which is a system that is trained using one RL algorithm, but whose recurrent dynamics implement a second, quite separate RL procedure.'}",
-learning-related development of context-specific neuronal responses to places and events: the hippocampal role in context processing,David M. Smith,"Contextual information plays a key role in learning and memory. Learned information becomes associated with the context such that the context can cue the relevant memories and behaviors. An extensive literature involving experimental brain lesions has implicated the hippocampus in context processing. However, the neurophysiological mechanisms of context coding are not known. Although “context” has typically been defined in terms of the background cues, recent studies indicate that hippocampal neurons are sensitive to subtle changes in task demands, even in an unchanging environment. Thus, the context may also include non-environmental features of a learning situation. In the present study, hippocampal neuronal activity was recorded while rats learned to approach different reward locations in two contexts. Because all of the training took place in the same environment, the contexts were defined by the task demands rather than by environmental stimuli. Learning to differentiate two such contexts was associated with the development of highly context-specific neuronal firing patterns. These included different place fields in pyramidal neurons and different event (e.g., reward) responses in pyramidal and interneurons. The differential firing patterns did not develop in a control condition that did not involve a context manipulation. The context-specific firing patterns could modulate activity in extrahippocampal structures to prime context-appropriate behavioral responses and memories. These results provide direct support for a context processing role of the hippocampus and suggest that the hippocampus contributes contextual representations to episodic memories.","{'model': 'tldr@v2.0.0', 'text': 'Results provide direct support for a context processing role of the hippocampus and suggest that the hippocampus contributes contextual representations to episodic memories.'}",https://www.jneurosci.org/content/jneuro/26/12/3154.full.pdf
-applications of learning analytics in high schools: a systematic literature review,Erverson B. G. de Sousa,"Learning analytics aims to analyze data from students and learning environments to support learning at different levels. Although learning analytics is a recent field, it reached a high level of maturity, especially in its applications for higher education. However, little of the research in learning analytics targets other educational levels, such as high school. This paper reports the results of a systematic literature review (SLR) focused on the adoption of learning analytics in high schools. More specifically, the SLR followed four steps: the search, selection of relevant studies, critical assessment, and the extraction of the relevant field, which included the main goals, approaches, techniques, and challenges of adopting learning analytics in high school. The results show that, in this context, learning analytics applications are focused on small-scale initiatives rather than institutional adoption. Based on the findings of this study, in combination with the literature, this paper proposes future directions of research and development in order to scale up learning analytics applications in high schools.","{'model': 'tldr@v2.0.0', 'text': 'A systematic literature review focused on the adoption of learning analytics in high schools shows that, in this context, learning analytics applications are focused on small-scale initiatives rather than institutional adoption.'}",https://www.frontiersin.org/articles/10.3389/frai.2021.737891/pdf
-picanet: learning pixel-wise contextual attention for saliency detection,Nian Liu,"Contexts play an important role in the saliency detection task. However, given a context region, not all contextual information is helpful for the final task. In this paper, we propose a novel pixel-wise contextual attention network, i.e., the PiCANet, to learn to selectively attend to informative context locations for each pixel. Specifically, for each pixel, it can generate an attention map in which each attention weight corresponds to the contextual relevance at each context location. An attended contextual feature can then be constructed by selectively aggregating the contextual information. We formulate the proposed PiCANet in both global and local forms to attend to global and local contexts, respectively. Both models are fully differentiable and can be embedded into CNNs for joint training. We also incorporate the proposed models with the U-Net architecture to detect salient objects. Extensive experiments show that the proposed PiCANets can consistently improve saliency detection performance. The global and local PiCANets facilitate learning global contrast and homogeneousness, respectively. As a result, our saliency model can detect salient objects more accurately and uniformly, thus performing favorably against the state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A novel pixel-wise contextual attention network, i.e., the PiCANet, to learn to selectively attend to informative context locations for each pixel, which can generate an attention map in which each attention weight corresponds to the contextual relevance at each context location.'}",https://arxiv.org/pdf/1708.06433
-workplace learning in context,A. Fuller,"This book combines a critical analysis of the organizational and employment context of workplace learning with an understanding of theories of learning. It brings new ideas on workplace learning to the readers and an understanding of the social context of the workplace. Combining the latest research findings with practical examples of the issues raised, it is illustrated with case studies and critical vignettes as well as chapter summaries, guiding students the subject. Chapters are grouped under four main themes: 'The context of workplace learning', 'The workplace as a learning environment', 'Skills, knowledge and the workplace' and 'Research and policy'. With international contributors offering a global perspective, this encourages students to be critical when evaluating the important debates in the area.",,https://opus.lib.uts.edu.au/bitstream/10453/12376/1/2004001318.pdf
-learning to compose dynamic tree structures for visual contexts,Kaihua Tang,"We propose to compose dynamic tree structures that place the objects in an image into a visual context, helping visual reasoning tasks such as scene graph generation and visual Q&A. Our visual context tree model, dubbed VCTree, has two key advantages over existing structured object representations including chains and fully-connected graphs: 1) The efficient and expressive binary tree encodes the inherent parallel/hierarchical relationships among objects, e.g., ``clothes'' and ``pants'' are usually co-occur and belong to ``person''; 2) the dynamic structure varies from image to image and task to task, allowing more content-/task-specific message passing among objects. To construct a VCTree, we design a score function that calculates the task-dependent validity between each object pair, and the tree is the binary version of the maximum spanning tree from the score matrix. Then, visual contexts are encoded by bidirectional TreeLSTM and decoded by task-specific models. We develop a hybrid learning procedure which integrates end-task supervised learning and the tree structure reinforcement learning, where the former's evaluation result serves as a self-critic for the latter's structure exploration. Experimental results on two benchmarks, which require reasoning over contexts: Visual Genome for scene graph generation and VQA2.0 for visual Q&A, show that VCTree outperforms state-of-the-art results while discovering interpretable visual context structures.","{'model': 'tldr@v2.0.0', 'text': ""A hybrid learning procedure is developed which integrates end-task supervised learning and the tree structure reinforcement learning, where the former's evaluation result serves as a self-critic for the latter's structure exploration.""}",https://arxiv.org/pdf/1812.01880
-computing the context effect for science learning,Thomas Forissier,,"{'model': 'tldr@v2.0.0', 'text': 'The hypothesis presented is that a context gap between two students can be illuminating to highlight the respective general-particular aspects of an object or phenomenon, and provoking a perturbation during the learning process to obtain the emergence of such an event could be a productive tutoring strategy.'}",
-situated cognition and the culture of learning,J. Brown,"Many teaching practices implicitly assume that conceptual knowledge can be abstracted from the situations in which it is learned and used. This article argues that this assumption inevitably limits the effectiveness of such practices. Drawing on recent research into cognition as it is manifest in everyday activity, the authors argue that knowledge is situated, being in part a product of the activity, context, and culture in which it is developed and used. They discuss how this view of knowledge affects our understanding of learning, and they note that conventional schooling too often ignores the influence of school culture on what is learned in school. As an alternative to conventional practices, they propose cognitive apprenticeship (Collins, Brown, & Newman, in press), which honors the situated nature of knowledge. They examine two examples of mathematics instruction that exhibit certain key features of this approach to teaching.",,https://www.ideals.illinois.edu/items/18064/bitstreams/64664/data.pdf
-an efficient framework for learning sentence representations,Lajanugen Logeswaran,"In this work we propose a simple and efficient framework for learning sentence representations from unlabelled data. Drawing inspiration from the distributional hypothesis and recent work on learning sentence representations, we reformulate the problem of predicting the context in which a sentence appears as a classification problem. Given a sentence and its context, a classifier distinguishes context sentences from other contrastive sentences based on their vector representations. This allows us to efficiently learn different types of encoding functions, and we show that the model learns high-quality sentence representations. We demonstrate that our sentence representations outperform state-of-the-art unsupervised and supervised representation learning methods on several downstream NLP tasks that involve understanding sentence semantics while achieving an order of magnitude speedup in training time.","{'model': 'tldr@v2.0.0', 'text': 'This work reformulates the problem of predicting the context in which a sentence appears as a classification problem, and proposes a simple and efficient framework for learning sentence representations from unlabelled data.'}",
-learning end-to-end goal-oriented dialog,Antoine Bordes,"Traditional dialog systems used in goal-oriented applications require a lot of domain-specific handcrafting, which hinders scaling up to new domains. End-to-end dialog systems, in which all components are trained from the dialogs themselves, escape this limitation. But the encouraging success recently obtained in chit-chat dialog may not carry over to goal-oriented settings. This paper proposes a testbed to break down the strengths and shortcomings of end-to-end dialog systems in goal-oriented applications. Set in the context of restaurant reservation, our tasks require manipulating sentences and symbols, so as to properly conduct conversations, issue API calls and use the outputs of such calls. We show that an end-to-end dialog system based on Memory Networks can reach promising, yet imperfect, performance and learn to perform non-trivial operations. We confirm those results by comparing our system to a hand-crafted slot-filling baseline on data from the second Dialog State Tracking Challenge (Henderson et al., 2014a). We show similar result patterns on data extracted from an online concierge service.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that an end-to-end dialog system based on Memory Networks can reach promising, yet imperfect, performance and learn to perform non-trivial operations and be compared to a hand-crafted slot-filling baseline on data from the second Dialog State Tracking Challenge.'}",
-learning analytics in higher education: a preponderance of analytics but very little learning?,Carolina Guzmán-Valenzuela,,"{'model': 'tldr@v2.0.0', 'text': 'Examining the publication patterns on learning analytics in higher education and their main challenges concludes by distinguishing between a practice-based and management-oriented community of learning analytics and an academic- oriented community.'}",https://educationaltechnologyjournal.springeropen.com/counter/pdf/10.1186/s41239-021-00258-x
-curriculum learning,Yoshua Bengio,"Humans and animals learn much better when the examples are not randomly presented but organized in a meaningful order which illustrates gradually more concepts, and gradually more complex ones. Here, we formalize such training strategies in the context of machine learning, and call them ""curriculum learning"". In the context of recent research studying the difficulty of training in the presence of non-convex training criteria (for deep deterministic and stochastic neural networks), we explore curriculum learning in various set-ups. The experiments show that significant improvements in generalization can be achieved. We hypothesize that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and, in the case of non-convex criteria, on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesized that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).'}",
-does e-learning service quality influence e-learning student satisfaction and loyalty? evidence from vietnam,Long Pham,,,
-reviewing and exploring innovative ubiquitous learning tools in higher education,Shadi A. Aljawarneh,,"{'model': 'tldr@v2.0.0', 'text': 'A global view regarding the current as well as future tendencies pertaining to ubiquitous e- learning tools is obtained and thus possible key comments are provided for employing e-learning tools like MOODLE, Web 2.0 and Web 3.0 in the classroom.'}",
-design-based research methods for studying learning in context: introduction,W. Sandoval,"The field of psychology has a long history of interaction with education, and educational psychology has had a profound impact on how issues of learning have been framed and studied in educational contexts. Still, it has never been simple to translate theoretical insights into educational practice. Educational psychology has been criticized for not creating “usable knowledge” (Lagemann, 2002). Currently, educational researchers generally have been pushed to justify how their claims are “scientific” and “evidence-based” (National Research Council, 2002). There is a tension between the desire for locally usable knowledge on the one hand and scientifically sound, generalizable knowledge on the other. Lagemann, for example, argued that the traditional paradigm of psychology has striven for experimental control at the expense of fidelity to learning as it actually occurs. Thus, although such claims might be scientific in one sense, they do not adequately explain or predict the phenomena they purport to address. This critique extends the long-standing debate surrounding the ecological validity of well-defined psychological tasks and their relation to psychological phenomena as they come to occur in everyday settings (Brunswik, 1943; Lewin, 1943). As a field, we still lack an adequate methodological reconciliation that attends to issues of both experimental control and ecological validity. At the same time, there is considerable unease with the perceived “credibility gap” (Levin & O’Donnell, 1999) of much of educational research because it is not produced with what are considered to be scientific methods. From this perspective, the knowledge from educational research has limited usability because it is not trustworthy. An educational psychology that is both usable in a practical sense and scientifically trustworthy cannot proceed without directly studying the phenomena it hopes to explain in its inherent messiness. A little over a decade ago, Brown (1992) described her evolving approach to “design experimentation” as an effort to bridge laboratory studies of learning with studies of complex instructional interventions based on such insights. She showed how insights from the laboratory were inherently limited in their ability to explain or predict learning in the classroom. The challenge, as she saw it, was to develop a methodology of experimenting with intervention designs in situ to develop theories of learning (and teaching) that accounted for the multiple interactions of people acting in a complex social setting. At the same time, Collins (1992) was putting forth a notion of educational research as a “design science,” like aerospace engineering, that required a methodology to systematically test design variants for effectiveness. Achieving such a design science, however, requires a sufficient understanding of the underlying variables at all relevant layers of a complex social system (schooling)—an understanding that we do not yet have (Collins, Joseph, & Bielaczyc, 2004). The last 12 years have seen an increasing uptake of the design experimentation methodology, so much so that a recent handbook on research in math and science education is replete with examples and formulations of the approach (Kelly & Lesh, 2000). The general approach has been called by many names. We have settled on the term design-based research over the other commonly used phrases “design experimentation,” which connotes a specific form of controlled experimentation that does not capture the breadth of the approach, or “design research,” which is too easily confused with research design and other efforts in design fields that lack in situ research components. The approach to research described in this issue is design based in that it is theoretiEDUCATIONAL PSYCHOLOGIST, 39(4), 199–201 Copyright © 2004, Lawrence Erlbaum Associates, Inc.",,
-learning in context through conflict and alignment: farmers and scientists in search of sustainable agriculture,J. Eshuis,,,http://library.wur.nl/WebQuery/wurpubs/fulltext/28944
-"mathematics teachers' ""learning to notice"" in the context of a video club",E. V. Es,,,
-"organizational learning and communities-of-practice: toward a unified view of working, learning, and innovation",J. Brown,"Recent ethnographic studies of workplace practices indicate that the ways people actually work usually differ fundamentally from the ways organizations describe that work in manuals, training programs, organizational charts, and job descriptions. Nevertheless, organizations tend to rely on the latter in their attempts to understand and improve work practice. We examine one such study. We then relate its conclusions to compatible investigations of learning and of innovation to argue that conventional descriptions of jobs mask not only the ways people work, but also significant learning and innovation generated in the informal communities-of-practice in which they work. By reassessing work, learning, and innovation in the context of actual communities and actual practices, we suggest that the connections between these three become apparent. With a unified view of working, learning, and innovating, it should be possible to reconceive of and redesign organizations to improve all three.",,
-designing learning trajectory of circle using the context of ferris wheel,N. Fitri,"Ferris wheel is one amusement playground that resembles a giant spinning wheel. Many students are familiar with the Ferris wheel in the mini version of it at night market festivals. This is the potential for learning mathematics. Furthermore, there is a mathematical learning approach called Indonesian Realistic Mathematics Education (IRME) where students learn with contexts which are close to students' life as starting points. Therefore, this study aims to design a learning trajectory using the IRME approach with the Ferris wheel as the context in the learning process to support students' understanding of the learning about circles. The research method is design research that consists of three stages: preliminary design, design experiments, and retrospective analysis. The subjects were 20 eighth-grade students from one of the private Junior High School in Yogyakarta. The instruments used are videos to see the learning process and when students work on the given problems, photos to refer the results of student work, and written test in worksheets to get the data on student's work. The research result explores the learning trajectory practiced using the Ferris wheel as the context seen in the student's daily activities. The learning trajectory consists of four events, namely assembling the Ferris wheel, drawing an illustration of the Ferris wheel, making a list of the circle parts, and solving a problem related to the parts of the circle. Lastly, this study shows that learning trajectory activities have essential roles in supporting students' understanding of the concept of a circle.",,http://journals.ums.ac.id/index.php/jramathedu/article/download/10961/5776
-"re-conceptualizing ""scaffolding"" and the zone of proximal development in the context of symmetrical collaborative learning.",Juan Manuel Fernández-Cárdenas,"The linked concepts of ‘scaffolding’ and the Zone of Proximal Development (ZPD) were originally applied to the context of asymmetrical teaching and learning with a teacher or adult explicitly supporting a learner, usually a child, to achieve tasks beyond their ability when working alone. In this paper we investigate how these concepts need to be reconceptualised if they are to be applied to the different context of symmetrical learning amongst groups of peers. We present two separate studies. In the first one we analyse the type of talk used by a group of children from Mexico solving the Raven’s Standard Progressive Matrices (RSPM) test together both before and after an intervention programme teaching ‘exploratory talk’. Our analysis demonstrates a ZPD created by the way in which they talk together. In the second study we present the comparison of the talk of two groups of children, one from Mexico and the other from the UK, solving together a single matrix from the RSPM test. Our analysis shows how the concept of ‘scaffolding’ can be applied to understand how these groups of children use language to support shared thinking and learning. In both studies we found that applying ideas of ‘scaffolding’ and the ZPD to symmetrical learning required the re-conceptualisation of these concepts as characterisations of dynamic processes within dialogues.",,
-bert transformer model for detecting arabic gpt2 auto-generated tweets,F. Harrag,"During the last two decades, we have progressively turned to the Internet and social media to find news, entertain conversations and share opinion. Recently, OpenAI has developed a machine learning system called GPT-2 for Generative Pre-trained Transformer-2, which can produce deepfake texts. It can generate blocks of text based on brief writing prompts that look like they were written by humans, facilitating the spread false or auto-generated text. In line with this progress, and in order to counteract potential dangers, several methods have been proposed for detecting text written by these language models. In this paper, we propose a transfer learning based model that will be able to detect if an Arabic sentence is written by humans or automatically generated by bots. Our dataset is based on tweets from a previous work, which we have crawled and extended using the Twitter API. We used GPT2-Small-Arabic to generate fake Arabic Sentences. For evaluation, we compared different recurrent neural network (RNN) word embeddings based baseline models, namely: LSTM, BI-LSTM, GRU and BI-GRU, with a transformer-based model. Our new transfer-learning model has obtained an accuracy up to 98%. To the best of our knowledge, this work is the first study where ARABERT and GPT2 were combined to detect and classify the Arabic auto-generated texts.","{'model': 'tldr@v2.0.0', 'text': 'This work is the first study where ARABERT and GPT2 were combined to detect and classify the Arabic auto-generated texts, and the new transfer-learning model has obtained an accuracy up to 98%.'}",
-"operationalizing and implementing pretrained, large artificial intelligence linguistic models in the us health care system: outlook of generative pretrained transformer 3 (gpt-3) as a service model",Emre Sezgin,"Generative pretrained transformer models have been popular recently due to their enhanced capabilities and performance. In contrast to many existing artificial intelligence models, generative pretrained transformer models can perform with very limited training data. Generative pretrained transformer 3 (GPT-3) is one of the latest releases in this pipeline, demonstrating human-like logical and intellectual responses to prompts. Some examples include writing essays, answering complex questions, matching pronouns to their nouns, and conducting sentiment analyses. However, questions remain with regard to its implementation in health care, specifically in terms of operationalization and its use in clinical practice and research. In this viewpoint paper, we briefly introduce GPT-3 and its capabilities and outline considerations for its implementation and operationalization in clinical practice through a use case. The implementation considerations include (1) processing needs and information systems infrastructure, (2) operating costs, (3) model biases, and (4) evaluation metrics. In addition, we outline the following three major operational factors that drive the adoption of GPT-3 in the US health care system: (1) ensuring Health Insurance Portability and Accountability Act compliance, (2) building trust with health care providers, and (3) establishing broader access to the GPT-3 tools. This viewpoint can inform health care practitioners, developers, clinicians, and decision makers toward understanding the use of the powerful artificial intelligence tools integrated into hospital systems and health care.",,https://medinform.jmir.org/2022/2/e32875/PDF
-developing prompts from large language model for extracting clinical information from pathology and ultrasound reports in breast cancer,Hyeon Seok Choi,"Purpose We aimed to evaluate the time and cost of developing prompts using large language model (LLM), tailored to extract clinical factors in breast cancer patients and their accuracy. Materials and Methods We collected data from reports of surgical pathology and ultrasound from breast cancer patients who underwent radiotherapy from 2020 to 2022. We extracted the information using the Generative Pre-trained Transformer (GPT) for Sheets and Docs extension plugin and termed this the “LLM” method. The time and cost of developing the prompts with LLM methods were assessed and compared with those spent on collecting information with “full manual” and “LLM-assisted manual” methods. To assess accuracy, 340 patients were randomly selected, and the extracted information by LLM method were compared with those collected by “full manual” method. Results Data from 2,931 patients were collected. We developed 12 prompts for Extract function and 12 for Format function to extract and standardize the information. The overall accuracy was 87.7%. For lymphovascular invasion, it was 98.2%. Developing and processing the prompts took 3.5 hours and 15 minutes, respectively. Utilizing the ChatGPT application programming interface cost US $65.8 and when factoring in the estimated wage, the total cost was US $95.4. In an estimated comparison, “LLM-assisted manual” and “LLM” methods were time- and cost-efficient compared to the “full manual” method. Conclusion Developing and facilitating prompts for LLM to derive clinical factors was efficient to extract crucial information from huge medical records. This study demonstrated the potential of the application of natural language processing using LLM model in breast cancer patients. Prompts from the current study can be re-used for other research to collect clinical information.","{'model': 'tldr@v2.0.0', 'text': 'Developing and facilitating prompts for LLM to derive clinical factors was efficient to extract crucial information from huge medical records and can be re-used for other research to collect clinical information.'}",https://www.e-roj.org/upload/pdf/roj-2023-00633.pdf
-swectrl-mini: a data-transparent transformer-based large language model for controllable text generation in swedish,Dmytro Kalpakchi,"We present SweCTRL-Mini, a large Swedish language model that can be used for inference and fine-tuning on a single consumer-grade GPU. The model is based on the CTRL architecture by Keskar, McCann, Varshney, Xiong, and Socher (2019), which means that users of the SweCTRL-Mini model can control the genre of the generated text by inserting special tokens in the generation prompts. SweCTRL-Mini is trained on a subset of the Swedish part of the mC4 corpus and a set of Swedish novels. In this article, we provide (1) a detailed account of the utilized training data and text pre-processing steps, to the extent that it is possible to check whether a specific phrase/source was a part of the training data, and (2) an evaluation of the model on both discriminative tasks, using automatic evaluation methods, and generative tasks, using human referees. We also compare the generative capabilities of the model with those of GPT-3. SweCTRL-Mini is fully open and available for download.","{'model': 'tldr@v2.0.0', 'text': 'A detailed account of the utilized training data and text pre-processing steps is provided, to the extent that it is possible to check whether a specific phrase/source was a part of the training data, and an evaluation of the model on both discriminative tasks, using automatic evaluation methods, and generative task, using human referees are provided.'}",http://arxiv.org/pdf/2304.13994
-reward modeling for mitigating toxicity in transformer-based language models,Farshid Faal,,"{'model': 'tldr@v2.0.0', 'text': 'The experiments demonstrate that the Reinforce-Detoxify method for language model detoxification outperforms existing detoxification approaches in automatic evaluation metrics, indicating that the approach is less prone to unintended bias toward social identities in generated content.'}",https://arxiv.org/pdf/2202.09662
-optimizing continuous prompts for visual relationship detection by affix-tuning,Shouguan Xiao,"Visual relationship detection is crucial for understanding visual scenes and is widely used in many areas, including visual navigation, visual question answering, and machine trouble detection. Traditional detection methods often fuse multiple region modules, which takes considerable time and resources to train every module with extensive samples. As every module is independent, the computation process has difficulty achieving unity and lacks a higher level of logical reasonability. In response to the above problems, we propose a novel method of affix-tuning transformers for visual relationship detection tasks, which keeps transformer model parameters frozen and optimizes a small continuous task-specific vector. It not only makes the model unified and reduces the training cost but also maintains the common-sense reasonability without multiscale training. In addition, we design a vision-and-language sentence expression prompt template and train a few transformer model parameters for downstream tasks. Our method, Prompt Template and Affix-Tuning Transformers (PTAT), is evaluated on visual relationship detection and Visual Genome datasets. Finally, the results of the proposed method are close to or even higher than those of the state-of-the-art methods on some evaluation metrics.","{'model': 'tldr@v2.0.0', 'text': 'A novel method of affix-tuning transformers for visual relationship detection tasks, which keeps transformer model parameters frozen and optimizes a small continuous task-specific vector, which makes the model unified and reduces the training cost and maintains the common-sense reasonability without multiscale training.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09815128.pdf
-contextual transformer for offline meta reinforcement learning,Runji Lin,"The pretrain-finetuning paradigm in large-scale sequence models has made significant progress in natural language processing and computer vision tasks. However, such a paradigm is still hindered by several challenges in Reinforcement Learning (RL), including the lack of self-supervised pretraining algorithms based on offline data and efficient fine-tuning/prompt-tuning over unseen downstream tasks. In this work, we explore how prompts can improve sequence modeling-based offline reinforcement learning (offline-RL) algorithms. Firstly, we propose prompt tuning for offline RL, where a context vector sequence is concatenated with the input to guide the conditional policy generation. As such, we can pretrain a model on the offline dataset with self-supervised loss and learn a prompt to guide the policy towards desired actions. Secondly, we extend our framework to Meta-RL settings and propose Contextual Meta Transformer (CMT); CMT leverages the context among different tasks as the prompt to improve generalization on unseen tasks. We conduct extensive experiments across three different offline-RL settings: offline single-agent RL on the D4RL dataset, offline Meta-RL on the MuJoCo benchmark, and offline MARL on the SMAC benchmark. Superior results validate the strong performance, and generality of our methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt tuning for offline RL, where a context vector sequence is concatenated with the input to guide the conditional policy generation and proposes Contextual Meta Transformer (CMT), which leverages the context among different tasks as the prompt to improve generalization on unseen tasks.'}",http://arxiv.org/pdf/2211.08016
-fl-tuning: layer tuning for feed-forward network in transformer,Jingping Liu,"Prompt tuning is an emerging way of adapting pre-trained language models to downstream tasks. However, the existing studies are mainly to add prompts to the input sequence. This way would not work as expected due to the intermediate multi-head self-attention and feed-forward network computation, making model optimization not very smooth. Hence, we propose a novel tuning way called layer tuning, aiming to add learnable parameters in Transformer layers. Specifically, we focus on layer tuning for feed-forward network in the Transformer, namely FL-tuning. It introduces additional units into the hidden layer of each feed-forward network. We conduct extensive experiments on the public CLUE benchmark. The results show that: 1) Our FL-tuning outperforms prompt tuning methods under both full-data and few-shot settings in almost all cases. In particular, it improves accuracy by 17.93% (full-data setting) on WSC 1.0 and F1 by 16.142% (few-shot setting) on CLUENER over P-tuning v2. 2) Our FL-tuning is more stable and converges about 1.17 times faster than P-tuning v2. 3) With only about 3% of Transformer's parameters to be trained, FL-tuning is comparable with fine-tuning on most datasets, and significantly outperforms fine-tuning (e.g., accuracy improved by 12.9% on WSC 1.1) on several datasets. The source codes are available at https://github.com/genggui001/FL-Tuning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel tuning way called layer tuning, aiming to add learnable parameters in Transformer layers, namely FL-tuning, which introduces additional units into the hidden layer of each feed-forward network in the Transformer.'}",http://arxiv.org/pdf/2206.15312
-natural language processing based automated essay scoring with parameter-efficient transformer approach,Angad Sethi,"Existing automated scoring models implement layers of traditional recurrent neural networks to achieve reasonable performance. However, the models provide limited performance due to the limited capacity to encode long-term dependencies. The paper proposed a novel architecture incorporating pioneering language models of the natural language processing community. We leverage pre-trained language models and integrate it with adapter modules, which use a bottle-neck architecture to reduce the number of trainable parameters while delivering excellent performance. We also propose a model by re-purposing the bidirectional attention flow model to detect adversarial essays. The model we put forward achieves state-of-the-art performance on most essay prompts in the Automated Student Assessment Prize data set. We outline the previous methods employed to attempt this task, and show how our model outperforms them.","{'model': 'tldr@v2.0.0', 'text': 'A novel architecture incorporating pioneering language models of the natural language processing community is proposed, which leverage pre-trained language models and integrate it with adapter modules, which use a bottle-neck architecture to reduce the number of trainable parameters while delivering excellent performance.'}",
-context-based narrative generation transformer (ngen-transformer),Abraar Raza Samar,"Text generation is an important domain of natural language processing where the plausibility of the generated text depends upon the context assimilation capabilities of the architecture being used. Recently the performance of automatic text generation task have greatly improved with the use of attention based language models. In this paper, we have explored the task of story generation based on some user defined context or prompt. We have proposed a GP2 based narrative generation architecture called NGen-Transformer. Our proposed architectures focuses specifically on the context provided by the user to produce meaningful stories. For the purpose of evaluation of our proposed model, we have used the WritingPrompts dataset which consists of a large number of human written sample stories based on corresponding titles or sentences (prompts). Experimental results show that our proposed NGen- Transformer model outperforms several sequence to sequence as well as attention based architectures on the task of story generation.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed NGen- Transformer model outperforms several sequence to sequence as well as attention based architectures on the task of story generation.'}",
-chinese text paraphrase recognition based on openprompt introducing hybrid prompts,Chengbo Mao,"The fine-tune paradigm adopted by traditional paraphrase recognition tasks cannot fully exploit the knowledge of pretrained language models (PLMs). At this stage, the Prompt paradigm reconstructs downstream tasks by constructing templates to make it more suitable for the training form of PLM. However, there are many processes, the code base is not supervised, and a single discrete template limits the model prediction ability. In response to this problem, this paper proposes a method of introducing hybrid prompts based on OpenPrompt. OpenPrompt makes the Prompt process have a unified framework, and hybrid prompts solve the problem that discrete templates cannot fully mine PLM knowledge. This paper first constructs a hybrid template with [mask] slots, and then transforms the original input through the template to obtain xprompt , and then xprompt is input to the bert-base-chinese model of the multi-layer bidirectional transformer based on the attention mechanism for training. When the model trains the optimal prompt, calculate the label with the highest probability of filling in the label set, and finally map the label to the prediction result. The experimental results show that compared with the fine-tune paradigm, the F1 value of the prompt paradigm exceeds 4.69% under the same PLM and hyperparameters. Compared with not using soft prompt, the average accuracy and average F1 value obtained with soft hints are 2.294% and 2.31% higher, respectively. Meanwhile, when the number of soft hints is 6, the accuracy and F1 value reach the highest.","{'model': 'tldr@v2.0.0', 'text': 'This paper first constructs a hybrid template with [mask] slots, and then transforms the original input through the template to obtain xprompt, which is input to the bert-base-chinese model of the multi-layer bidirectional transformer based on the attention mechanism for training.'}",
-fp-detr: detection transformer advanced,Yang Cao,"Large-scale pre-training has proven to be effective for visual representation learning on downstream tasks, especially for improving robustness and generalization. However, the recently developed detection transformers only employ pre-training on its backbone while leaving the key component, i.e., a 12-layer transformer, being trained from scratch, which prevents the model from above benefits. This separated training paradigm is mainly caused by the discrepancy between the upstream and downstream tasks. To mitigate the issue, we propose FP-DETR, a new method that Fully Pre-Trains an encoder-only transformer and smoothly finetunes it for object detection via a task adapter. Inspired by the success of textual prompts in NLP, we treat query positional embeddings as visual prompts to help the model attend to the target area (prompting) and recognize the object. To this end, we propose the task adapter which leverages self-attention to model the contextual relation between object query embedding. Experiments on the challenging COCO dataset demonstrate that our FP-DETR achieves competitive performance. Moreover, it enjoys better robustness to common corruptions and generalization to small-size datasets than state-of-the-art detection transformers. Code will be made publicly available at https://github.com/encounter1997/FP-DETR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes FP-DETR, a new method that Fully Pre-Trains an encoder-only transformer and smoothly finetunes it for object detection via a task adapter which leverages self-attention to model the contextual relation between object query embedding.'}",
-learning to compress prompts with gist tokens,Jesse Mu,"Prompting is the primary way to utilize the multitask capabilities of language models (LMs), but prompts occupy valuable space in the input context window, and repeatedly encoding the same prompt is computationally inefficient. Finetuning and distillation methods allow for specialization of LMs without prompting, but require retraining the model for each task. To avoid this trade-off entirely, we present gisting, which trains an LM to compress prompts into smaller sets of""gist""tokens which can be cached and reused for compute efficiency. Gist models can be trained with no additional cost over standard instruction finetuning by simply modifying Transformer attention masks to encourage prompt compression. On decoder (LLaMA-7B) and encoder-decoder (FLAN-T5-XXL) LMs, gisting enables up to 26x compression of prompts, resulting in up to 40% FLOPs reductions, 4.2% wall time speedups, and storage savings, all with minimal loss in output quality.","{'model': 'tldr@v2.0.0', 'text': 'Gisting is presented, which trains an LM to compress prompts into smaller sets of ""gist"" tokens which can be cached and reused for compute efficiency, resulting in up to 26x compression of prompts.'}",https://arxiv.org/pdf/2304.08467
-zero-shot entity and tweet characterization with designed conditional prompts and contexts,S. Srivatsa,"Online news and social media have been the de facto mediums to disseminate information globally from the beginning of the last decade. However, bias in content and purpose of intentions are not regulated, and managing bias is the responsibility of content consumers. In this regard, understanding the stances and biases of news sources towards specific entities becomes important. To address this problem, we use pretrained language models, which have been shown to bring about good results with no task-specific training or few-shot training. In this work, we approach the prob-lem of characterizing Named Entities and Tweets as an open-ended text classification and open-ended fact probing problem. We evaluate the zero-shot language model capabilities of Generative Pretrained Transformer 2 (GPT-2) to characterize Entities and Tweets subjectively with human psychology-inspired and logical conditional prefixes and contexts. First, we fine-tune the GPT-2 modelon a sufficiently large news corpus and evaluate subjective characterization of popular entities in the corpus by priming with prefixes. Second, we fine-tune GPT-2 with a Tweets corpus from a few popular hashtags and evaluate characterizing tweets by priming the language model with prefixes, questions, and contextual synopsis prompts. Entity characterization results were positive across measures and human evaluation.","{'model': 'tldr@v2.0.0', 'text': 'This work evaluates the zero-shot language model capabilities of Generative Pretrained Transformer 2 (GPT-2) to characterize Entities and Tweets subjectively with human psychology-inspired and logical conditional prefixes and contexts and results were positive across measures and human evaluation.'}",http://arxiv.org/pdf/2204.08405
-instruction-vit: multi-modal prompts for instruction learning in vit,Zhe Xiao,"Prompts have been proven to play a crucial role in large language models, and in recent years, vision models have also been using prompts to improve scalability for multiple downstream tasks. In this paper, we focus on adapting prompt design based on instruction tuning into a visual transformer model for image classification which we called Instruction-ViT. The key idea is to implement multi-modal prompts (text or image prompt) related to category information to guide the fine-tuning of the model. Based on the experiments of several image captionining tasks, the performance and domain adaptability were improved. Our work provided an innovative strategy to fuse multi-modal prompts with better performance and faster adaptability for visual classification models.","{'model': 'tldr@v2.0.0', 'text': 'This work provided an innovative strategy to fuse multi-modal prompts with better performance and faster adaptability for visual classification models.'}",http://arxiv.org/pdf/2305.00201
-clinical decision transformer: intended treatment recommendation through goal prompting,Seunghyun Lee,"With recent achievements in tasks requiring context awareness, foundation models have been adopted to treat large-scale data from electronic health record (EHR) systems. However, previous clinical recommender systems based on foundation models have a limited purpose of imitating clinicians' behavior and do not directly consider a problem of missing values. In this paper, we propose Clinical Decision Transformer (CDT), a recommender system that generates a sequence of medications to reach a desired range of clinical states given as goal prompts. For this, we conducted goal-conditioned sequencing, which generated a subsequence of treatment history with prepended future goal state, and trained the CDT to model sequential medications required to reach that goal state. For contextual embedding over intra-admission and inter-admissions, we adopted a GPT-based architecture with an admission-wise attention mask and column embedding. In an experiment, we extracted a diabetes dataset from an EHR system, which contained treatment histories of 4788 patients. We observed that the CDT achieved the intended treatment effect according to goal prompt ranges (e.g., NormalA1c, LowerA1c, and HigherA1c), contrary to the case with behavior cloning. To the best of our knowledge, this is the first study to explore clinical recommendations from the perspective of goal prompting. See https://clinical-decision-transformer.github.io for code and additional information.","{'model': 'tldr@v2.0.0', 'text': 'This is the first study to explore clinical recommendations from the perspective of goal prompting, and proposes Clinical Decision Transformer, a recommender system that generates a sequence of medications to reach a desired range of clinical states given as goal prompts.'}",http://arxiv.org/pdf/2302.00612
-adversarial transformer language models for contextual commonsense inference,Pedro Colon-Hernandez,"Contextualized or discourse aware commonsense inference is the task of generating coherent commonsense assertions (i.e., facts) from a given story, and a particular sentence from that story. Some problems with the task are: lack of controllability for topics of the inferred facts; lack of commonsense knowledge during training; and, possibly, hallucinated or false facts. In this work, we utilize a transformer model for this task and develop techniques to address the aforementioned problems in the task. We control the inference by introducing a new technique we call""hinting"". Hinting is a kind of language model prompting, that utilizes both hard prompts (specific words) and soft prompts (virtual learnable templates). This serves as a control signal to advise the language model""what to talk about"". Next, we establish a methodology for performing joint inference with multiple commonsense knowledge bases. Joint inference of commonsense requires care, because it is imprecise and the level of generality is more flexible. You want to be sure that the results""still make sense""for the context. To this end, we align the textual version of assertions from three knowledge graphs (ConceptNet, ATOMIC2020, and GLUCOSE) with a story and a target sentence. This combination allows us to train a single model to perform joint inference with multiple knowledge graphs. We show experimental results for the three knowledge graphs on joint inference. Our final contribution is exploring a GAN architecture that generates the contextualized commonsense assertions and scores them as to their plausibility through a discriminator. The result is an integrated system for contextual commonsense inference in stories, that can controllably generate plausible commonsense assertions, and takes advantage of joint inference between multiple commonsense knowledge bases.","{'model': 'tldr@v2.0.0', 'text': 'The result is an integrated system for contextual commonsense inference in stories, that can controllably generate plausible commonsense assertions, and takes advantage of joint inference between multiple commonsense knowledge bases.'}",http://arxiv.org/pdf/2302.05406
-"tryage: real-time, intelligent routing of user prompts to large language models",S. Hari,"The introduction of the transformer architecture and the self-attention mechanism has led to an explosive production of language models trained on specific downstream tasks and data domains. With over 200, 000 models in the Hugging Face ecosystem, users grapple with selecting and optimizing models to suit multifaceted workflows and data domains while addressing computational, security, and recency concerns. There is an urgent need for machine learning frameworks that can eliminate the burden of model selection and customization and unleash the incredible power of the vast emerging model library for end users. Here, we propose a context-aware routing system, Tryage, that leverages a language model router for optimal selection of expert models from a model library based on analysis of individual input prompts. Inspired by the thalamic router in the brain, Tryage employs a perceptive router to predict down-stream model performance on prompts and, then, makes a routing decision using an objective function that integrates performance predictions with user goals and constraints that are incorporated through flags (e.g., model size, model recency). Tryage allows users to explore a Pareto front and automatically trade-off between task accuracy and secondary goals including minimization of model size, recency, security, verbosity, and readability. Across heterogeneous data sets that include code, text, clinical data, and patents, the Tryage framework surpasses Gorilla and GPT3.5 turbo in dynamic model selection identifying the optimal model with an accuracy of 50.9% , compared to 23.6% by GPT 3.5 Turbo and 10.8% by Gorilla. Conceptually, Tryage demonstrates how routing models can be applied to program and control the behavior of multi-model LLM systems to maximize efficient use of the expanding and evolving language model ecosystem.","{'model': 'tldr@v2.0.0', 'text': 'A context-aware routing system that leverages a language model router for optimal selection of expert models from a model library based on analysis of individual input prompts, Tryage demonstrates how routing models can be applied to program and control the behavior of multi-model LLM systems to maximize efficient use of the expanding and evolving language model ecosystem.'}",https://arxiv.org/pdf/2308.11601
-vima: robot manipulation with multimodal prompts,Yunfan Jiang,"Prompt-based learning has emerged as a successful paradigm in natural language processing, where a single general-purpose language model can be instructed to perform any task specified by input prompts. Yet task specification in robotics comes in various forms, such as imitating one-shot demonstrations, following language instructions, and reaching visual goals. They are often considered different tasks and tackled by specialized models. We show that a wide spectrum of robot manipulation tasks can be expressed with multimodal prompts , interleaving textual and visual tokens. Accordingly, we develop a new simulation benchmark that consists of thousands of procedurally-generated tabletop tasks with multimodal prompts, 600K+ expert trajectories for imitation learning, and a four-level evaluation protocol for systematic generalization. We de-sign a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively. VIMA features a recipe that achieves strong model scalability and data efficiency. It outperforms alternative designs in the hardest zero-shot generalization setting by up to 2 . 9 × task success rate given the same training data. With 10 × less training data, VIMA still performs 2 . 7 × better than the best competing variant. Code and video demos are available at vimalabs.github.io .","{'model': 'tldr@v2.0.0', 'text': 'This work develops a new simulation benchmark that consists of thousands of procedurally-generated tabletop tasks with multimodal prompts, 600K+ expert trajectories for imitation learning, and a four-level evaluation protocol for systematic generalization and de-signs a transformer-based robot agent, VIMA, that processes these prompts and outputs motor actions autoregressively.'}",
-prompt-based tuning of transformer models for multi-center medical image segmentation of head and neck cancer,Numan Saeed,"Medical image segmentation is a vital healthcare endeavor requiring precise and efficient models for appropriate diagnosis and treatment. Vision transformer (ViT)-based segmentation models have shown great performance in accomplishing this task. However, to build a powerful backbone, the self-attention block of ViT requires large-scale pre-training data. The present method of modifying pre-trained models entails updating all or some of the backbone parameters. This paper proposes a novel fine-tuning strategy for adapting a pretrained transformer-based segmentation model on data from a new medical center. This method introduces a small number of learnable parameters, termed prompts, into the input space (less than 1% of model parameters) while keeping the rest of the model parameters frozen. Extensive studies employing data from new unseen medical centers show that the prompt-based fine-tuning of medical segmentation models provides excellent performance regarding the new-center data with a negligible drop regarding the old centers. Additionally, our strategy delivers great accuracy with minimum re-training on new-center data, significantly decreasing the computational and time costs of fine-tuning pre-trained models. Our source code will be made publicly available.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel fine-tuning strategy for adapting a pretrained transformer-based segmentation model on data from a new medical center that delivers great accuracy with minimum re-training on new-center data, significantly decreasing the computational and time costs of fine- Tuning pre-trained models.'}",https://www.mdpi.com/2306-5354/10/7/879/pdf?version=1690208147
-tempo: prompt-based generative pre-trained transformer for time series forecasting,Defu Cao,"The past decade has witnessed significant advances in time series modeling with deep learning. While achieving state-of-the-art results, the best-performing architectures vary highly across applications and domains. Meanwhile, for natural language processing, the Generative Pre-trained Transformer (GPT) has demonstrated impressive performance via training one general-purpose model across various textual datasets. It is intriguing to explore whether GPT-type architectures can be effective for time series, capturing the intrinsic dynamic attributes and leading to significant accuracy improvements. In this paper, we propose a novel framework, TEMPO, that can effectively learn time series representations. We focus on utilizing two essential inductive biases of the time series task for pre-trained models: (i) decomposition of the complex interaction between trend, seasonal and residual components; and (ii) introducing the selection-based prompts to facilitate distribution adaptation in non-stationary time series. TEMPO expands the capability for dynamically modeling real-world temporal phenomena from data within diverse domains. Our experiments demonstrate the superior performance of TEMPO over state-of-the-art methods on a number of time series benchmark datasets. This performance gain is observed not only in standard supervised learning settings but also in scenarios involving previously unseen datasets as well as in scenarios with multi-modal inputs. This compelling finding highlights TEMPO's potential to constitute a foundational model-building framework.","{'model': 'tldr@v2.0.0', 'text': 'A novel framework, TEMPO, that can effectively learn time series representations by utilizing two essential inductive biases of the time series task for pre-trained models, and introducing the selection-based prompts to facilitate distribution adaptation in non-stationary time series.'}",https://arxiv.org/pdf/2310.04948
-batgpt: a bidirectional autoregessive talker from generative pre-trained transformer,Z. Li,"BatGPT is a large-scale language model designed and trained jointly by Wuhan University and Shanghai Jiao Tong University. It is capable of generating highly natural and fluent text in response to various types of input, including text prompts, images, and audio. In the modeling level, we employ a bidirectional autoregressive architecture that allows the model to efficiently capture the complex dependencies of natural language, making it highly effective in tasks such as language generation, dialog systems, and question answering. Moreover, the bidirectional autoregressive modeling not only operates from left to right but also from right to left, effectively reducing fixed memory effects and alleviating model hallucinations. In the training aspect, we propose a novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance. Overall, these approaches significantly improve the effectiveness of BatGPT, and the model can be utilized for a wide range of natural language applications.","{'model': 'tldr@v2.0.0', 'text': ""A novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance significantly improve the effectiveness of BatGPT.""}",https://arxiv.org/pdf/2307.00360
-prompt-engineering and transformer-based question generation and evaluation,Rubaba Amyeen,"Question generation has numerous applications in the educational context. Question generation can prove helpful for students when reviewing content and testing themselves. Furthermore, a question generation model can aid teachers by lessening the burden of creating assessments and other practice material. This paper aims to find the best method to generate questions from textual data through a transformer model and prompt engineering. In this research, we finetuned a pretrained distilBERT model on the SQuAD question answering dataset to generate questions. In addition to training a transformer model, prompt engineering was applied to generate questions effectively using the LLaMA model. The generated questions were compared against the baseline questions in the SQuAD dataset to evaluate the effectiveness of four different prompts. All four prompts demonstrated over 60% similarity on average. Of the prompt-generated questions, 30% achieved a high similarity score greater than 70%.","{'model': 'tldr@v2.0.0', 'text': 'This research finetuned a pretrained distilBERT model on the SQuAD question answering dataset to generate questions using the LLaMA model and prompt engineering, and evaluated the effectiveness of four different prompts.'}",
-prompt guided transformer for multi-task dense prediction,Yuxiang Lu,"Task-conditional architecture offers advantage in parameter efficiency but falls short in performance compared to state-of-the-art multi-decoder methods. How to trade off performance and model parameters is an important and difficult problem. In this paper, we introduce a simple and lightweight task-conditional model called Prompt Guided Transformer (PGT) to optimize this challenge. Our approach designs a Prompt-conditioned Transformer block, which incorporates task-specific prompts in the self-attention mechanism to achieve global dependency modeling and parameter-efficient feature adaptation across multiple tasks. This block is integrated into both the shared encoder and decoder, enhancing the capture of intra- and inter-task features. Moreover, we design a lightweight decoder to further reduce parameter usage, which accounts for only 2.7% of the total model parameters. Extensive experiments on two multi-task dense prediction benchmarks, PASCAL-Context and NYUD-v2, demonstrate that our approach achieves state-of-the-art results among task-conditional methods while using fewer parameters, and maintains a significant balance between performance and parameter size.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a simple and lightweight task-conditional model called Prompt Guided Transformer (PGT), which incorporates task-specific prompts in the self-attention mechanism to achieve global dependency modeling and parameter-efficient feature adaptation across multiple tasks.'}",https://arxiv.org/pdf/2307.15362
-small-scale proxies for large-scale transformer training instabilities,Mitchell Wortsman,"Teams that have trained large Transformer-based models have reported training instabilities at large scale that did not appear when training with the same hyperparameters at smaller scales. Although the causes of such instabilities are of scientific interest, the amount of resources required to reproduce them has made investigation difficult. In this work, we seek ways to reproduce and study training stability and instability at smaller scales. First, we focus on two sources of training instability described in previous work: the growth of logits in attention layers (Dehghani et al., 2023) and divergence of the output logits from the log probabilities (Chowdhery et al., 2022). By measuring the relationship between learning rate and loss across scales, we show that these instabilities also appear in small models when training at high learning rates, and that mitigations previously employed at large scales are equally effective in this regime. This prompts us to investigate the extent to which other known optimizer and model interventions influence the sensitivity of the final loss to changes in the learning rate. To this end, we study methods such as warm-up, weight decay, and the $\mu$Param (Yang et al., 2022), and combine techniques to train small models that achieve similar losses across orders of magnitude of learning rate variation. Finally, to conclude our exploration we study two cases where instabilities can be predicted before they emerge by examining the scaling behavior of model activation and gradient norms.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the extent to which other known optimizer and model interventions influence the sensitivity of the final loss to changes in the learning rate, and studies two cases where instabilities can be predicted before they emerge by examining the scaling behavior of model activation and gradient norms.'}",https://arxiv.org/pdf/2309.14322
-crosslingual content scoring in five languages using machine-translation and multilingual transformer models,Andrea Horbach,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that a combination of machine translation and multilingual models outperforms each method individually - the best results are reached when combining the available data in different languages, i.e. first training a model on the large English ASAP dataset before fine-tuning on smaller amounts of training data in the target language.'}",https://link.springer.com/content/pdf/10.1007/s40593-023-00370-1.pdf
-multimodal transformer-based lyric generation from midi and text data,Anthony Zhan,"Current generative language models are designed to take a text prompt as input, and output the highest probability continuation. However, when applying this idea to the design space of lyric generation, there is an obvious shortcoming — lyrics are often created with music’s auditory features in mind, such as tempo, tonality, melody, rhythm, etc. Which is to say, a model able to interpret both text prompts and snippets of music may be able to produce much better music. Therefore, our project seeks to combine both text and music data in a multimodal transformerbased model in order to generate lyrics more effectively than a purely text-based model, utilizing multiple transformer architectures, including MusicBERT and GPT-2.",,
-study on a real-time work assistance method for substation based on petri net model,Qi Gao,"Work assistance is one of the important means to improve work efficiency. This paper focuses on substation work and proposes a real-time work assistance method based on petri net model. The work assistance includes two aspects, termed as, workflow guidance and operation prompt. Particularly, the workflow constructed by the petri net includes sequence, inhibition, concurrency and conflict structures to declare the correlations between work steps. Further, different forms of operation prompts at each step are designed, considering real-time responses of workers including skip, replay, pause and restart. The effectiveness of the method is well validated by an application of the work assistance on measuring the grounding current of transformer core and clamp. It is demonstrated that the unified assistance method built by the petri net successfully instructs standardized working process, helps avoid potential risks, and provides flexible operation prompts in real time.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that the unified assistance method built by the petri net successfully instructs standardized working process, helps avoid potential risks, and provides flexible operation prompts in real time.'}",
-efficient model personalization in federated learning via client-specific prompt generation,Fu-En Yang,"Federated learning (FL) emerges as a decentralized learning framework which trains models from multiple distributed clients without sharing their data to preserve privacy. Recently, large-scale pre-trained models (e.g., Vision Transformer) have shown a strong capability of deriving robust representations. However, the data heterogeneity among clients, the limited computation resources, and the communication bandwidth restrict the deployment of large-scale models in FL frameworks. To leverage robust representations from large-scale models while enabling efficient model personalization for heterogeneous clients, we propose a novel personalized FL framework of client-specific Prompt Generation (pFedPG), which learns to deploy a personalized prompt generator at the server for producing client-specific visual prompts that efficiently adapts frozen backbones to local data distributions. Our proposed framework jointly optimizes the stages of personalized prompt adaptation locally and personalized prompt generation globally. The former aims to train visual prompts that adapt foundation models to each client, while the latter observes local optimization directions to generate personalized prompts for all clients. Through extensive experiments on benchmark datasets, we show that our pFedPG is favorable against state-of-the-art personalized FL methods under various types of data heterogeneity, allowing computation and communication efficient model personalization.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel personalized FL framework of client-specific Prompt Generation (pFedPG), which learns to deploy a personalized prompt generator at the server for producing client- specific visual prompts that efficiently adapts frozen backbones to local data distributions.'}",https://arxiv.org/pdf/2308.15367
-how interesting and coherent are the stories generated by a large‐scale neural language model? comparing human and automatic evaluations of machine‐generated text,Dominic Callan,"Evaluation of the narrative text generated by machines has traditionally been a challenge, particularly when attempting to evaluate subjective elements such as interest or believability. Recent improvements in narrative machine text generation have been largely driven by the emergence of transformer‐based language models, trained on massive quantities of data, resulting in higher quality text generation. In this study, a corpus of stories is generated using the pre‐trained GPT‐Neo transformer model, with human‐written prompts as inputs upon which to base the narrative text. The stories generated through this process are subsequently evaluated through both human evaluation and two automated metrics: BERTScore and BERT Next Sentence Prediction, with the aim of determining whether there is a correlation between the automatic scores and the human judgements. The results show variation in human evaluation results in comparison to modern automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.","{'model': 'tldr@v2.0.0', 'text': 'The results show variation in human evaluation results in comparison to modern automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/exsy.13292
-kosmos-2.5: a multimodal literate model,Tengchao Lv,"We present Kosmos-2.5, a multimodal literate model for machine reading of text-intensive images. Pre-trained on large-scale text-intensive images, Kosmos-2.5 excels in two distinct yet cooperative transcription tasks: (1) generating spatially-aware text blocks, where each block of text is assigned its spatial coordinates within the image, and (2) producing structured text output that captures styles and structures into the markdown format. This unified multimodal literate capability is achieved through a shared Transformer architecture, task-specific prompts, and flexible text representations. We evaluate Kosmos-2.5 on end-to-end document-level text recognition and image-to-markdown text generation. Furthermore, the model can be readily adapted for any text-intensive image understanding task with different prompts through supervised fine-tuning, making it a general-purpose tool for real-world applications involving text-rich images. This work also paves the way for the future scaling of multimodal large language models.","{'model': 'tldr@v2.0.0', 'text': 'Kosmos-2.5 can be readily adapted for any text-intensive image understanding task with different prompts through supervised fine-tuning, making it a general-purpose tool for real-world applications involving text-rich images and paves the way for the future scaling of multimodal large language models.'}",https://arxiv.org/pdf/2309.11419
-automated reading passage generation with openai's large language model,Ummugul Bezirhan,"The widespread usage of computer-based assessments and individualized learning platforms has resulted in an increased demand for the rapid production of high-quality items. Automated item generation (AIG), the process of using item models to generate new items with the help of computer technology, was proposed to reduce reliance on human subject experts at each step of the process. AIG has been used in test development for some time. Still, the use of machine learning algorithms has introduced the potential to improve the efficiency and effectiveness of the process greatly. The approach presented in this paper utilizes OpenAI's latest transformer-based language model, GPT-3, to generate reading passages. Existing reading passages were used in carefully engineered prompts to ensure the AI-generated text has similar content and structure to a fourth-grade reading passage. For each prompt, we generated multiple passages, the final passage was selected according to the Lexile score agreement with the original passage. In the final round, the selected passage went through a simple revision by a human editor to ensure the text was free of any grammatical and factual errors. All AI-generated passages, along with original passages were evaluated by human judges according to their coherence, appropriateness to fourth graders, and readability.","{'model': 'tldr@v2.0.0', 'text': ""The approach presented in this paper utilizes OpenAI's latest transformer-based language model, GPT-3, to generate reading passages that were evaluated by human judges according to their coherence, appropriateness to fourth graders, and readability.""}",http://arxiv.org/pdf/2304.04616
-prompt tuning gpt-2 language model for parameter-efficient domain adaptation of asr systems,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains creating a need to adapt to new domains with small memory and deployment overhead. In this work, we introduce domain-prompts, a methodology that involves training a small number of domain embedding parameters to prime a Transformer-based Language Model (LM) to a particular domain. Using this domain-adapted LM for rescoring ASR hypotheses can achieve 7-13% WER reduction for a new domain with just 1000 unlabeled textual domain-specific sentences. This improvement is comparable or even better than fully fine-tuned models even though just 0.02% of the parameters of the base LM are updated. Additionally, our method is deployment-friendly as the learnt domain embeddings are prefixed to the input to the model rather than changing the base model architecture. Therefore, our method is an ideal choice for on-the-fly adaptation of LMs used in ASR systems to progressively scale it to new domains.","{'model': 'tldr@v2.0.0', 'text': None}",
-phenaki: variable length video generation from open domain textual description,Ruben Villegas,"We present Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts. Generating videos from text is particularly challenging due to the computational cost, limited quantities of high quality text-video data and variable length of videos. To address these issues, we introduce a new model for learning video representation which compresses the video to a small representation of discrete tokens. This tokenizer uses causal attention in time, which allows it to work with variable-length videos. To generate video tokens from text we are using a bidirectional masked transformer conditioned on pre-computed text tokens. The generated video tokens are subsequently de-tokenized to create the actual video. To address data issues, we demonstrate how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets. Compared to the previous video generation methods, Phenaki can generate arbitrary long videos conditioned on a sequence of prompts (i.e. time variable text or a story) in open domain. To the best of our knowledge, this is the first time a paper studies generating videos from time variable prompts. In addition, compared to the per-frame baselines, the proposed video encoder-decoder computes fewer tokens per video but results in better spatio-temporal consistency.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts, and demonstrates how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets.'}",http://arxiv.org/pdf/2210.02399
-llama-adapter: efficient fine-tuning of language models with zero-init attention,Renrui Zhang,"We present LLaMA-Adapter, a lightweight adaption method to efficiently fine-tune LLaMA into an instruction-following model. Using 52K self-instruct demonstrations, LLaMA-Adapter only introduces 1.2M learnable parameters upon the frozen LLaMA 7B model, and costs less than one hour for fine-tuning on 8 A100 GPUs. Specifically, we adopt a set of learnable adaption prompts, and prepend them to the word tokens at higher transformer layers. Then, a zero-initialized attention mechanism with zero gating is proposed, which adaptively injects the new instructional cues into LLaMA, while effectively preserves its pre-trained knowledge. With our efficient training, LLaMA-Adapter can generate high-quality responses, comparable to Alpaca with fully fine-tuned 7B parameters. Besides language commands, our approach can be simply extended to multi-modal instructions for learning image-conditioned LLaMA model, which achieves superior reasoning performance on ScienceQA and COCO Caption benchmarks. Furthermore, we also evaluate the zero-initialized attention mechanism for fine-tuning other pre-trained models (ViT, RoBERTa) on traditional vision and language tasks, demonstrating the superior generalization capacity of our approach. Code is released at https://github.com/OpenGVLab/LLaMA-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'A zero-initialized attention mechanism with zero gating is proposed, which adaptively injects the new instructional cues into LLaMA, while effectively preserves its pre-trained knowledge on traditional vision and language tasks, demonstrating the superior generalization capacity of the approach.'}",http://arxiv.org/pdf/2303.16199
-"chatgpt in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations",Tirth Dave,"This paper presents an analysis of the advantages, limitations, ethical considerations, future prospects, and practical applications of ChatGPT and artificial intelligence (AI) in the healthcare and medical domains. ChatGPT is an advanced language model that uses deep learning techniques to produce human-like responses to natural language inputs. It is part of the family of generative pre-training transformer (GPT) models developed by OpenAI and is currently one of the largest publicly available language models. ChatGPT is capable of capturing the nuances and intricacies of human language, allowing it to generate appropriate and contextually relevant responses across a broad spectrum of prompts. The potential applications of ChatGPT in the medical field range from identifying potential research topics to assisting professionals in clinical and laboratory diagnosis. Additionally, it can be used to help medical students, doctors, nurses, and all members of the healthcare fraternity to know about updates and new developments in their respective fields. The development of virtual assistants to aid patients in managing their health is another important application of ChatGPT in medicine. Despite its potential applications, the use of ChatGPT and other AI tools in medical writing also poses ethical and legal concerns. These include possible infringement of copyright laws, medico-legal complications, and the need for transparency in AI-generated content. In conclusion, ChatGPT has several potential applications in the medical and healthcare fields. However, these applications come with several limitations and ethical considerations which are presented in detail along with future prospects in medicine and healthcare.","{'model': 'tldr@v2.0.0', 'text': 'An analysis of the advantages, limitations, ethical considerations, future prospects, and practical applications of ChatGPT and artificial intelligence (AI) in the healthcare and medical domains shows several potential applications come with several limitations and ethical considerations.'}",https://www.frontiersin.org/articles/10.3389/frai.2023.1169595/pdf
-synthetic data from diffusion models improves imagenet classification,Shekoofeh Azizi,"Deep generative models are becoming increasingly powerful, now generating diverse high fidelity photo-realistic samples given text prompts. Have they reached the point where models of natural images can be used for generative data augmentation, helping to improve challenging discriminative tasks? We show that large-scale text-to image diffusion models can be fine-tuned to produce class conditional models with SOTA FID (1.76 at 256x256 resolution) and Inception Score (239 at 256x256). The model also yields a new SOTA in Classification Accuracy Scores (64.96 for 256x256 generative samples, improving to 69.24 for 1024x1024 samples). Augmenting the ImageNet training set with samples from the resulting models yields significant improvements in ImageNet classification accuracy over strong ResNet and Vision Transformer baselines.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that large-scale text-to image diffusion models can be fine-tuned to produce class conditional models with SOTA FID and Inception Score that yields significant improvements in ImageNet classification accuracy over strong ResNet and Vision Transformer baselines.'}",http://arxiv.org/pdf/2304.08466
-evaluating academic answers generated using chatgpt,S. Fergus,"The integration of technology in education has become ever more prioritized since the COVID-19 pandemic. Chat Generative Pre-Trained Transformer (ChatGPT) is an artificial intelligence technology that generates conversational interactions to user prompts. The trained model can answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. The functionality of ChatGPT in answering chemistry assessment questions requires investigation to ascertain its potential impact on learning and assessment. Two chemistry-focused modules in year 1 and year 2 of a pharmaceutical science program are used to study and evaluate ChatGPT-generated responses in relation to the end-of-year exam assessments. For questions that focused on knowledge and understanding with ""describe” and ""discuss” verbs, the ChatGPT generated responses. For questions that focused on application of knowledge and interpretation with nontext information, the ChatGPT technology reached a limitation. A further analysis of the quality of responses is reported in this study. ChatGPT is not considered a high-risk technology tool in relation to cheating. Similar to the COVID-19 disruption, ChatGPT is expected to provide a catalyst for educational discussions on academic integrity and assessment design. © 2023 The Authors. Published by American Chemical Society and Division of Chemical Education, Inc.",,
-memory augmented large language models are computationally universal,Dale Schuurmans,"We show that transformer-based large language models are computationally universal when augmented with an external memory. Any deterministic language model that conditions on strings of bounded length is equivalent to a finite automaton, hence computationally limited. However, augmenting such models with a read-write memory creates the possibility of processing arbitrarily large inputs and, potentially, simulating any algorithm. We establish that an existing large language model, Flan-U-PaLM 540B, can be combined with an associative read-write memory to exactly simulate the execution of a universal Turing machine, $U_{15,2}$. A key aspect of the finding is that it does not require any modification of the language model weights. Instead, the construction relies solely on designing a form of stored instruction computer that can subsequently be programmed with a specific set of prompts.","{'model': 'tldr@v2.0.0', 'text': 'It is established that an existing large language model, Flan-U-PaLM 540B, can be combined with an associative read-write memory to exactly simulate the execution of a universal Turing machine, $U_{15,2}$.'}",http://arxiv.org/pdf/2301.04589
-learning to learn with generative models of neural network checkpoints,William S. Peebles,"We explore a data-driven approach for learning to optimize neural networks. We construct a dataset of neural network checkpoints and train a generative model on the parameters. In particular, our model is a conditional diffusion transformer that, given an initial input parameter vector and a prompted loss, error, or return, predicts the distribution over parameter updates that achieve the desired metric. At test time, it can optimize neural networks with unseen parameters for downstream tasks in just one update. We find that our approach successfully generates parameters for a wide range of loss prompts. Moreover, it can sample multimodal parameter solutions and has favorable scaling properties. We apply our method to different neural network architectures and tasks in supervised and reinforcement learning.","{'model': 'tldr@v2.0.0', 'text': 'This model is a conditional diffusion transformer that, given an initial input parameter vector and a prompted loss, error, or return, predicts the distribution over parameter updates that achieve the desired metric.'}",http://arxiv.org/pdf/2209.12892
-in-context learning of large language models explained as kernel regression,Chi Han,"Large language models (LLMs) have initiated a paradigm shift in transfer learning. In contrast to the classic pretraining-then-finetuning procedure, in order to use LLMs for downstream prediction tasks, one only needs to provide a few demonstrations, known as in-context examples, without adding more or updating existing model parameters. This in-context learning (ICL) capabilities of LLMs is intriguing, and it is not yet fully understood how pretrained LLMs acquire such capabilities. In this paper, we investigate the reason why a transformer-based language model can accomplish in-context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression algorithms when faced with in-context examples. More concretely, we first prove that Bayesian inference on in-context prompts can be asymptotically understood as kernel regression ˆ y = (cid:80) i y i K ( x,x i ) (cid:80) i K ( x,x i ) as the number of in-context demonstrations grows. Then, we empirically investigate the in-context behaviors of language models. We find that during ICL, the attentions and hidden features in LLMs match the behaviors of a kernel regression. Finally, our theory provides insights on multiple phenomena observed in ICL field: why retrieving demonstrative samples similar to test sample can help, why ICL performance is sensitive to the output formats, and why ICL accuracy benefits from selecting in-distribuion and representative samples. We will make our code available to the research community following publication.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates the reason why a transformer-based language model can accomplish in- Context learning after pre-training on a general language corpus by proposing one hypothesis that LLMs can simulate kernel regression algorithms when faced with in-context examples and empirically investigates the in- context behaviors of language models.'}",https://arxiv.org/pdf/2305.12766
-prompt tuning of deep neural networks for speaker-adaptive visual speech recognition,Minsu Kim,"Visual Speech Recognition (VSR) aims to infer speech into text depending on lip movements alone. As it focuses on visual information to model the speech, its performance is inherently sensitive to personal lip appearances and movements, and this makes the VSR models show degraded performance when they are applied to unseen speakers. In this paper, to remedy the performance degradation of the VSR model on unseen speakers, we propose prompt tuning methods of Deep Neural Networks (DNNs) for speaker-adaptive VSR. Specifically, motivated by recent advances in Natural Language Processing (NLP), we finetune prompts on adaptation data of target speakers instead of modifying the pre-trained model parameters. Different from the previous prompt tuning methods mainly limited to Transformer variant architecture, we explore different types of prompts, the addition, the padding, and the concatenation form prompts that can be applied to the VSR model which is composed of CNN and Transformer in general. With the proposed prompt tuning, we show that the performance of the pre-trained VSR model on unseen speakers can be largely improved by using a small amount of adaptation data (e.g., less than 5 minutes), even if the pre-trained model is already developed with large speaker variations. Moreover, by analyzing the performance and parameters of different types of prompts, we investigate when the prompt tuning is preferred over the finetuning methods. The effectiveness of the proposed method is evaluated on both word- and sentence-level VSR databases, LRW-ID and GRID.","{'model': 'tldr@v2.0.0', 'text': 'The proposed prompt tuning methods of Deep Neural Networks for speaker-adaptive VSR are proposed and it is shown that the performance of the pre-trained VSR model on unseen speakers can be largely improved by using a small amount of adaptation data, even if thePre-trained model is already developed with large speaker variations.'}",http://arxiv.org/pdf/2302.08102
-radiology gets chatty: the chatgpt saga unfolds,H. Grewal,"As artificial intelligence (AI) continues to evolve and mature, it is increasingly finding applications in the field of healthcare, particularly in specialties like radiology that are data-heavy and image-focused. Language learning models (LLMs) such as OpenAI's Generative Pre-trained Transformer-4 (GPT-4) are new in the field of medicine and there is a paucity of literature regarding the possible utilities of GPT-4 given its novelty. We aim to present an in-depth exploration of the role of GPT-4, an advanced language model, in radiology. Giving the GPT-4 model prompts for generating reports, template generation, enhancing clinical decision-making, and suggesting captivating titles for research articles, patient communication, and education, can occasionally be quite generic, and at times, it may present factually incorrect content, which could lead to errors. The responses were then analyzed in detail regarding their potential utility in day-to-day radiologist workflow, patient education, and research processes. Further research is required to evaluate LLMs' accuracy and safety in clinical practice and to develop comprehensive guidelines for their implementation.","{'model': 'tldr@v2.0.0', 'text': 'An in-depth exploration of the role of GPT-4, an advanced language model, in radiology is presented and responses were analyzed in detail regarding their potential utility in day-to-day radiologist workflow, patient education, and research processes.'}",https://assets.cureus.com/uploads/original_article/pdf/161200/20230708-6319-1gaxi52.pdf
-masksketch: unpaired structure-guided masked image generation,D. Bashkirova,"Recent conditional image generation methods produce images of remarkable diversity, fidelity and realism. However, the majority of these methods allow conditioning only on labels or text prompts, which limits their level of control over the generation result. In this paper, we introduce MaskSketch, an image generation method that allows spatial conditioning of the generation result using a guiding sketch as an extra conditioning signal during sampling. MaskSketch utilizes a pretrained masked generative transformer, requiring no model training or paired supervision, and works with input sketches of different levels of abstraction. We show that intermediate self-attention maps of a masked generative transformer encode important structural information of the input image, such as scene layout and object shape, and we propose a novel sampling method based on this observation to enable structure-guided generation. Our results show that MaskSketch achieves high image realism and fidelity to the guiding structure. Evaluated on standard benchmark datasets, MaskSketch outperforms state-of-the-art methods for sketch-to-image translation, as well as unpaired image-to-image translation approaches. The code can be found on our project website: https://masksketch.github.io/","{'model': 'tldr@v2.0.0', 'text': 'The paper shows that intermediate self-attention maps of a masked generative transformer encode important structural information of the input image, such as scene layout and object shape, and proposes a novel sampling method based on this observation to enable structure-guided generation in MaskSketch.'}",https://arxiv.org/pdf/2302.05496
-"how could chatgpt impact my practice as an intensivist? an overview of potential applications, risks and limitations",M. Komorowski,,"{'model': 'tldr@v2.0.0', 'text': 'Transformers are an architecture used primarily in the field of natural language processing, that aims to solve sequence-to-sequence tasks while handling long-range dependencies with ease and provide superior answers while requiring less training time than previous approaches such as recurrent neural networks.'}",https://link.springer.com/content/pdf/10.1007/s00134-023-07096-7.pdf
-a brief analysis of “chatgpt” – a revolutionary tool designed by openai,Md. Asraful Haque,"The ChatGPT, a powerful conversational tool trained by OpenAI is considered to be a revolutionary model in the field of artificial intelligence and natural language processing. It has received a lot of attention because of its potential to automate a variety of tasks and possibly have an impact on sectors like translation, customer service, and content creation. It uses GPT-3 (Generative Pre-training Transformer 3) language model to process user queries. GPT-3 has been trained on a very large dataset, which includes a wide range of texts from the internet and other sources. This has given it a broad knowledge base and has allowed it to generate responses to a wide range of prompts that are coherent and human-like. GPT-3 is one of the largest and most powerful language models to date, and it has the ability to perform a wide range of natural language processing tasks. After its release, ChatGPT has become a trending tool for the public to experiment and explore what it is capable of. In this article, we want to clarify what ChatGPT is. How does it work? What makes it different from other chatbots or search engines like Google? What are the major challenges and future prospects for it?","{'model': 'tldr@v2.0.0', 'text': 'The ChatGPT, a powerful conversational tool trained by OpenAI, uses GPT-3 (Generative Pre-training Transformer 3) language model to process user queries and has the ability to perform a wide range of natural language processing tasks.'}",https://publications.eai.eu/index.php/airo/article/download/2983/2322
-puma: secure inference of llama-7b in five minutes,Ye Dong,"With ChatGPT as a representative, tons of companies have began to provide services based on large Transformers models. However, using such a service inevitably leak users' prompts to the model provider. Previous studies have studied secure inference for Transformer models using secure multiparty computation (MPC), where model parameters and clients' prompts are kept secret. Despite this, these frameworks are still limited in terms of model performance, efficiency, and deployment. To address these limitations, we propose framework PUMA to enable fast and secure Transformer model inference. Our framework designs high quality approximations for expensive functions such as GeLU and softmax, and significantly reduce the cost of secure inference while preserving the model performance. Additionally, we design secure Embedding and LayerNorm procedures that faithfully implement the desired functionality without undermining the Transformer architecture. PUMA is about $2\times$ faster than the state-of-the-art framework MPCFORMER(ICLR 2023) and has similar accuracy as plaintext models without fine-tuning (which the previous works failed to achieve). PUMA can even evaluate LLaMA-7B in around 5 minutes to generate 1 token. To our best knowledge, this is the first time that a model with such a parameter size is able to be evaluated under MPC. PUMA has been open-sourced in the Github repository of SecretFlow-SPU.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes framework PUMA to enable fast and secure Transformer model inference, and designs high quality approximations for expensive functions such as GeLU and softmax, and significantly reduce the cost of secure inference while preserving the model performance.'}",https://arxiv.org/pdf/2307.12533
-à-la-carte prompt tuning (apt): combining distinct data via composable prompting,Benjamin Bowman,"We introduce À-la-carte Prompt Tuning (APT), a transformer-based scheme to tune prompts on distinct data so that they can be arbitrarily composed at inference time. The individual prompts can be trained in isolation, possibly on different devices, at different times, and on different distributions or domains. Furthermore each prompt only contains information about the subset of data it was exposed to during training. During inference, models can be assembled based on arbitrary selections of data sources, which we call à-la-carte learning. À-la-carte learning enables constructing bespoke models specific to each user's individual access rights and preferences. We can add or remove information from the model by simply adding or removing the corresponding prompts without retraining from scratch. We demonstrate that à-la-carte built models achieve accuracy within 5% of models trained on the union of the respective sources, with comparable cost in terms of training and inference time. For the continual learning benchmarks Split CIFAR- 100 and CORe50, we achieve state-of-the-art performance.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2302.07994
-a pilot evaluation of chatgpt and dall-e 2 on decision making and spatial reasoning,Zhi–Bin Tang,"We conduct a pilot study selectively evaluating the cognitive abilities (decision making and spatial reasoning) of two recently released generative transformer models, ChatGPT and DALL-E 2. Input prompts were constructed following neutral a priori guidelines, rather than adversarial intent. Post hoc qualitative analysis of the outputs shows that DALL-E 2 is able to generate at least one correct image for each spatial reasoning prompt, but most images generated are incorrect (even though the model seems to have a clear understanding of the objects mentioned in the prompt). Similarly, in evaluating ChatGPT on the rationality axioms developed under the classical Von Neumann-Morgenstern utility theorem, we find that, although it demonstrates some level of rational decision-making, many of its decisions violate at least one of the axioms even under reasonable constructions of preferences, bets, and decision-making prompts. ChatGPT's outputs on such problems generally tended to be unpredictable: even as it made irrational decisions (or employed an incorrect reasoning process) for some simpler decision-making problems, it was able to draw correct conclusions for more complex bet structures. We briefly comment on the nuances and challenges involved in scaling up such a 'cognitive' evaluation or conducting it with a closed set of answer keys ('ground truth'), given that these models are inherently generative and open-ended in responding to prompts.",,http://arxiv.org/pdf/2302.09068
-proof of concept: using chatgpt to teach emergency physicians how to break bad news,J. Webb,"Background Breaking bad news is an essential skill for practicing physicians, particularly in the field of emergency medicine (EM). Patient-physician communication teaching has previously relied on standardized patient scenarios and objective structured clinical examination formats. The novel use of artificial intelligence (AI) chatbot technology, such as Chat Generative Pre-trained Transformer (ChatGPT), may provide an alternative role in graduate medical education in this area. As a proof of concept, the author demonstrates how providing detailed prompts to the AI chatbot can facilitate the design of a realistic clinical scenario, enable active roleplay, and deliver effective feedback to physician trainees. Methods ChatGPT-3.5 language model was utilized to assist in the roleplay of breaking bad news. A detailed input prompt was designed to outline rules of play and grading assessment via a standardized scale. User inputs (physician role), chatbot outputs (patient role) and ChatGPT-generated feedback were recorded. Results ChatGPT set up a realistic training scenario on breaking bad news based on the initial prompt. Active roleplay as a patient in an emergency department setting was accomplished, and clear feedback was provided to the user through the application of the Setting up, Perception, Invitation, Knowledge, Emotions with Empathy, and Strategy or Summary (SPIKES) framework for breaking bad news. Conclusion The novel use of AI chatbot technology to assist educators is abundant with potential. ChatGPT was able to design an appropriate scenario, provide a means for simulated patient-physician roleplay, and deliver real-time feedback to the physician user. Future studies are required to expand use to a targeted group of EM physician trainees and provide best practice guidelines for AI use in graduate medical education.","{'model': 'tldr@v2.0.0', 'text': 'The author demonstrates how providing detailed prompts to the AI chatbot can facilitate the design of a realistic clinical scenario, enable active roleplay, and deliver effective feedback to physician trainees.'}",https://assets.cureus.com/uploads/original_article/pdf/154391/20230609-458-1qfzq7g.pdf
-i know what you do not know: knowledge graph embedding via co-distillation learning,Yang Liu,"Knowledge graph (KG) embedding seeks to learn vector representations for entities and relations. Conventional models reason over graph structures, but they suffer from the issues of graph incompleteness and long-tail entities. Recent studies have used pre-trained language models to learn embeddings based on the textual information of entities and relations, but they cannot take advantage of graph structures. In the paper, we show empirically that these two kinds of features are complementary for KG embedding. To this end, we propose CoLE, a Co-distillation Learning method for KG Embedding that exploits the complementarity of graph structures and text information. Its graph embedding model employs Transformer to reconstruct the representation of an entity from its neighborhood subgraph. Its text embedding model uses a pre-trained language model to generate entity representations from the soft prompts of their names, descriptions and relational neighbors. To let the two models promote each other, we propose co-distillation learning that allows them to distill selective knowledge from each other's prediction logits. In our co-distillation learning, each model serves as both a teacher and a student. Experiments on benchmark datasets demonstrate that the two models outperform their related baselines, and the ensemble method CoLE with co-distillation learning advances the state-of-the-art of KG embedding.","{'model': 'tldr@v2.0.0', 'text': 'CoLE is proposed, a Co-distillation Learning method for KG Embedding that exploits the complementarity of graph structures and text information and advances the state-of-the-art of KG embedding.'}",
-promptonomyvit: multi-task prompt learning improves video transformers using synthetic scene data,Roei Herzig,"Action recognition models have achieved impressive results by incorporating scene-level annotations, such as objects, their relations, 3D structure, and more. However, obtaining annotations of scene structure for videos requires a significant amount of effort to gather and annotate, making these methods expensive to train. In contrast, synthetic datasets generated by graphics engines provide powerful alternatives for generating scene-level annotations across multiple tasks. In this work, we propose an approach to leverage synthetic scene data for improving video understanding. We present a multi-task prompt learning approach for video transformers, where a shared video transformer backbone is enhanced by a small set of specialized parameters for each task. Specifically, we add a set of ``task prompts'', each corresponding to a different task, and let each prompt predict task-related annotations. This design allows the model to capture information shared among synthetic scene tasks as well as information shared between synthetic scene tasks and a real video downstream task throughout the entire network. We refer to this approach as ``Promptonomy'', since the prompts model task-related structure. We propose the PromptonomyViT model (PViT), a video transformer that incorporates various types of scene-level information from synthetic data using the ``Promptonomy'' approach. PViT shows strong performance improvements on multiple video understanding tasks and datasets.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes the PromptonomyViT model (PViT), a video transformer that incorporates various types of scene-level information from synthetic data using the ``Promptonomy'' approach, and shows strong performance improvements on multiple video understanding tasks and datasets.""}",http://arxiv.org/pdf/2212.04821
-a redhead walks into a bar: experiences of writing fiction with artificial intelligence,Maliheh Ghajargar,"Human creativity has been often aided and supported by artificial tools, spanning traditional tools such as ideation cards, pens, and paper, to computed and software. Tools for creativity are increasingly using artificial intelligence to not only support the creative process, but also to act upon the creation with a higher level of agency. This paper focuses on writing fiction as a creative activity and explores human-AI co-writing through a research product, which employs a natural language processing model, the Generative Pre-trained Transformer 3 (GPT-3), to assist the co-authoring of narrative fiction. We report on two progressive – not comparative – autoethnographic studies to attain our own creative practices in light of our engagement with the research product: (1) a co-writing activity initiated by basic textual prompts using basic elements of narrative and (2) a co-writing activity initiated by more advanced textual prompts using elements of narrative, including dialects and metaphors undertaken by one of the authors of this paper who has doctoral training in literature. In both studies, we quickly came up against the limitations of the system; then, we repositioned our goals and practices to maximize our chances of success. As a result, we discovered not only limitations but also hidden capabilities, which not only altered our creative practices and outcomes, but which began to change the ways we were relating to the AI as collaborator.",,https://dl.acm.org/doi/pdf/10.1145/3569219.3569418
-can language models automate data wrangling?,Gonzalo Jaimovitch-López,,"{'model': 'tldr@v2.0.0', 'text': 'A major finding is that language models appear as a powerful tool for a wide range of data wrangling tasks, and some guidelines about how they can be integrated into data processing pipelines are provided, provided the users can take advantage of their flexibility and the diversity of tasks to be addressed.'}",https://link.springer.com/content/pdf/10.1007/s10994-022-06259-9.pdf
-textdiffuser: diffusion models as text painters,Jingye Chen,"Diffusion models have gained increasing attention for their impressive generation abilities but currently struggle with rendering accurate and coherent text. To address this issue, we introduce TextDiffuser, focusing on generating images with visually appealing text that is coherent with backgrounds. TextDiffuser consists of two stages: first, a Transformer model generates the layout of keywords extracted from text prompts, and then diffusion models generate images conditioned on the text prompt and the generated layout. Additionally, we contribute the first large-scale text images dataset with OCR annotations, MARIO-10M, containing 10 million image-text pairs with text recognition, detection, and character-level segmentation annotations. We further collect the MARIO-Eval benchmark to serve as a comprehensive tool for evaluating text rendering quality. Through experiments and user studies, we show that TextDiffuser is flexible and controllable to create high-quality text images using text prompts alone or together with text template images, and conduct text inpainting to reconstruct incomplete images with text. The code, model, and dataset will be available at \url{https://aka.ms/textdiffuser}.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces TextDiffuser, focusing on generating images with visually appealing text that is coherent with backgrounds, and contributes the first large-scale text images dataset with OCR annotations, MARIO-10M, containing 10 million image-text pairs with text recognition, detection, and character-level segmentation annotations.'}",http://arxiv.org/pdf/2305.10855
-language prompt for autonomous driving,Dongming Wu,"A new trend in the computer vision community is to capture objects of interest following flexible human command represented by a natural language prompt. However, the progress of using language prompts in driving scenarios is stuck in a bottleneck due to the scarcity of paired prompt-instance data. To address this challenge, we propose the first object-centric language prompt set for driving scenes within 3D, multi-view, and multi-frame space, named NuPrompt. It expands Nuscenes dataset by constructing a total of 35,367 language descriptions, each referring to an average of 5.3 object tracks. Based on the object-text pairs from the new benchmark, we formulate a new prompt-based driving task, \ie, employing a language prompt to predict the described object trajectory across views and frames. Furthermore, we provide a simple end-to-end baseline model based on Transformer, named PromptTrack. Experiments show that our PromptTrack achieves impressive performance on NuPrompt. We hope this work can provide more new insights for the autonomous driving community. Dataset and Code will be made public at \href{https://github.com/wudongming97/Prompt4Driving}{https://github.com/wudongming97/Prompt4Driving}.","{'model': 'tldr@v2.0.0', 'text': 'The first object-centric language prompt set for driving scenes within 3D, multi-view, and multi-frame space, named NuPrompt is proposed, and a new prompt-based driving task is formulated, employing a language prompt to predict the described object trajectory across views and frames.'}",https://arxiv.org/pdf/2309.04379
-vampnet: music generation via masked acoustic token modeling,Hugo Flores Garcia,"We introduce VampNet, a masked acoustic token modeling approach to music synthesis, compression, inpainting, and variation. We use a variable masking schedule during training which allows us to sample coherent music from the model by applying a variety of masking approaches (called prompts) during inference. VampNet is non-autoregressive, leveraging a bidirectional transformer architecture that attends to all tokens in a forward pass. With just 36 sampling passes, VampNet can generate coherent high-fidelity musical waveforms. We show that by prompting VampNet in various ways, we can apply it to tasks like music compression, inpainting, outpainting, continuation, and looping with variation (vamping). Appropriately prompted, VampNet is capable of maintaining style, genre, instrumentation, and other high-level aspects of the music. This flexible prompting capability makes VampNet a powerful music co-creation tool. Code and audio samples are available online.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that by prompting VampNet in various ways, it can apply it to tasks like music compression, inpainting, out Painting, continuation, and looping with variation (vamping), and is capable of maintaining style, genre, instrumentation, and other high-level aspects of the music.'}",https://arxiv.org/pdf/2307.04686
-training transformers together,Alexander Borzunov,"The infrastructure necessary for training state-of-the-art models is becoming overly expensive, which makes training such models affordable only to large corporations and institutions. Recent work proposes several methods for training such models collaboratively, i.e., by pooling together hardware from many independent parties and training a shared model over the Internet. In this demonstration, we collaboratively trained a text-to-image transformer similar to OpenAI DALL-E. We invited the viewers to join the ongoing training run, showing them instructions on how to contribute using the available hardware. We explained how to address the engineering challenges associated with such a training run (slow communication, limited memory, uneven performance between devices, and security concerns) and discussed how the viewers can set up collaborative training runs themselves. Finally, we show that the resulting model generates images of reasonable quality on a number of prompts.","{'model': 'tldr@v2.0.0', 'text': 'This demonstration collaboratively trained a text-to-image transformer similar to OpenAI DALL-E, and it is shown that the resulting model generates images of reasonable quality on a number of prompts.'}",http://arxiv.org/pdf/2207.03481
-vector representations of idioms in conversational systems,Tosin P. Adewumi,"In this study, we demonstrate that an open-domain conversational system trained on idioms or figurative language generates more fitting responses to prompts containing idioms. Idioms are a part of everyday speech in many languages and across many cultures, but they pose a great challenge for many natural language processing (NLP) systems that involve tasks such as information retrieval (IR), machine translation (MT), and conversational artificial intelligence (AI). We utilized the Potential Idiomatic Expression (PIE)-English idiom corpus for the two tasks that we investigated: classification and conversation generation. We achieved a state-of-the-art (SoTA) result of a 98% macro F1 score on the classification task by using the SoTA T5 model. We experimented with three instances of the SoTA dialogue model—the Dialogue Generative Pre-trained Transformer (DialoGPT)—for conversation generation. Their performances were evaluated by using the automatic metric, perplexity, and a human evaluation. The results showed that the model trained on the idiom corpus generated more fitting responses to prompts containing idioms 71.9% of the time in comparison with a similar model that was not trained on the idiom corpus. We have contributed the model checkpoint/demo/code to the HuggingFace hub for public access.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that an open-domain conversational system trained on idioms or figurative language generates more fitting responses to prompts containing idioms than a similar model trained on the idiom corpus.'}",http://arxiv.org/pdf/2205.03666
-generalization properties of retrieval-based models,S. Basu,"Many modern high-performing machine learning models such as GPT-3 primarily rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Remarkably, retrieval-based methods have enjoyed success on a wide range of problems, ranging from standard natural language processing and vision tasks to protein folding, as demonstrated by many recent efforts, including WebGPT and AlphaFold. Despite growing literature showcasing the promise of these models, the theoretical underpinning for such models remains underexplored. In this paper, we present a formal treatment of retrieval-based models to characterize their generalization ability. In particular, we focus on two classes of retrieval-based classification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task.","{'model': 'tldr@v2.0.0', 'text': 'A formal treatment of retrieval-based models to characterize their generalization ability is presented and it is shown that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy.'}",http://arxiv.org/pdf/2210.02617
-clinical prompt learning with frozen language models.,Niall Taylor,"When the first transformer-based language models were published in the late 2010s, pretraining with general text and then fine-tuning the model on a task-specific dataset often achieved the state-of-the-art performance. However, more recent work suggests that for some tasks, directly prompting the pretrained model matches or surpasses fine-tuning in performance with few or no model parameter updates required. The use of prompts with language models for natural language processing (NLP) tasks is known as prompt learning. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared this with more traditional fine-tuning methods. Results show that prompt learning methods were able to match or surpass the performance of traditional fine-tuning with up to 1000 times fewer trainable parameters, less training time, less training data, and lower computation resource requirements. We argue that these characteristics make prompt learning a very desirable alternative to traditional fine-tuning for clinical tasks, where the computational resources of public health providers are limited, and where data can often not be made available or not be used for fine-tuning due to patient privacy concerns. The complementary code to reproduce the experiments presented in this work can be found at https://github.com/NtaylorOX/Public_Clinical_Prompt.","{'model': 'tldr@v2.0.0', 'text': 'This work investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared this with more traditional fine-tuning methods, and showed that prompt learning methods were able to match or surpass the performance of traditional Finetuning with up to 1000 times fewer trainable parameters, less training time, lessTraining data, and lower computation resource requirements.'}",https://arxiv.org/pdf/2205.05535
-maqa: a multimodal qa benchmark for negation,Judith Yue Li,"Multimodal learning can benefit from the representation power of pretrained Large Language Models (LLMs). However, state-of-the-art transformer based LLMs often ignore negations in natural language and there is no existing benchmark to quantitatively evaluate whether multimodal transformers inherit this weakness. In this study, we present a new multimodal question answering (QA) benchmark adapted from labeled music videos in AudioSet (Gemmeke et al., 2017) with the goal of systematically evaluating if multimodal transformers can perform complex reasoning to recognize new concepts as negation of previously learned concepts. We show that with standard fine-tuning approach multimodal transformers are still incapable of correctly interpreting negation irrespective of model size. However, our experiments demonstrate that augmenting the original training task distributions with negated QA examples allow the model to reliably reason with negation. To do this, we describe a novel data generation procedure that prompts the 540B-parameter PaLM model to automatically generate negated QA examples as compositions of easily accessible video tags. The generated examples contain more natural linguistic patterns and the gains compared to template-based task augmentation approach are significant.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that with standard fine-tuning approach multimodal transformers are still incapable of correctly interpreting negation irrespective of model size, but it is demonstrated that augmenting the original training task distributions with negated QA examples allow the model to reliably reason with negation.'}",http://arxiv.org/pdf/2301.03238
-fedyolo: augmenting federated learning with pretrained transformers,Xuechen Zhang,"The growth and diversity of machine learning applications motivate a rethinking of learning with mobile and edge devices. How can we address diverse client goals and learn with scarce heterogeneous data? While federated learning aims to address these issues, it has challenges hindering a unified solution. Large transformer models have been shown to work across a variety of tasks achieving remarkable few-shot adaptation. This raises the question: Can clients use a single general-purpose model, rather than custom models for each task, while obeying device and network constraints? In this work, we investigate pretrained transformers (PTF) to achieve these on-device learning goals and thoroughly explore the roles of model size and modularity, where the latter refers to adaptation through modules such as prompts or adapters. Focusing on federated learning, we demonstrate that: (1) Larger scale shrinks the accuracy gaps between alternative approaches and improves heterogeneity robustness. Scale allows clients to run more local SGD epochs which can significantly reduce the number of communication rounds. At the extreme, clients can achieve respectable accuracy locally highlighting the potential of fully-local learning. (2) Modularity, by design, enables $>$100$\times$ less communication in bits. Surprisingly, it also boosts the generalization capability of local adaptation methods and the robustness of smaller PTFs. Finally, it enables clients to solve multiple unrelated tasks simultaneously using a single PTF, whereas full updates are prone to catastrophic forgetting. These insights on scale and modularity motivate a new federated learning approach we call""You Only Load Once""(FedYolo): The clients load a full PTF model once and all future updates are accomplished through communication-efficient modules with limited catastrophic-forgetting, where each task is assigned to its own module.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates pretrained transformers (PTF) to achieve on-device learning goals and thoroughly explore the roles of model size and modularity, where the latter refers to adaptation through modules such as prompts or adapters.'}",https://arxiv.org/pdf/2307.04905
-membership inference attacks with token-level deduplication on korean language models,Myung Gyo Oh,"The confidentiality threat against training data has become a significant security problem in neural language models. Recent studies have shown that memorized training data can be extracted by injecting well-chosen prompts into generative language models. While these attacks have achieved remarkable success in the English-based Transformer architecture, it is unclear whether they are still effective in other language domains. This paper studies the effectiveness of attacks against Korean models and the potential for attack improvements that might be beneficial for future defense studies. The contribution of this study is two-fold. First, we perform a membership inference attack against the state-of-the-art Korean GPT model. We found approximate training data with 20% to 90% precision in the top-100 samples and confirmed that the proposed attack technique for naive GPT is valid across the language domains. Second, in this process, we observed that the redundancy of the selected sentences could hardly be detected with the existing attack method. Since the information appearing in a few documents is more likely to be meaningful, it is desirable to increase the uniqueness of the sentences to improve the effectiveness of the attack. Thus, we propose a deduplication strategy to replace the traditional word-level similarity metric with the BPE token level. Our proposed strategy reduces 6% to 22% of the underestimated samples among selected ones, thereby improving precision by up to 7%p. As a result, we show that considering both language- and model-specific characteristics is essential to improve the effectiveness of attack strategies. We also discuss possible mitigations against the MI attacks on the general language models.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that considering both language- and model-specific characteristics is essential to improve the effectiveness of attack strategies, and a deduplication strategy to replace the traditional word-level similarity metric with the BPE token level is proposed.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10025743.pdf
-generatect: text-guided 3d chest ct generation,Ibrahim Ethem Hamamci,"Generative modeling has experienced substantial progress in recent years, particularly in text-to-image and text-to-video synthesis. However, the medical field has not yet fully exploited the potential of large-scale foundational models for synthetic data generation. In this paper, we introduce GenerateCT, the first method for text-conditional computed tomography (CT) generation, addressing the limitations in 3D medical imaging research and making our entire framework open-source. GenerateCT consists of a pre-trained large language model, a transformer-based text-conditional 3D chest CT generation architecture, and a text-conditional spatial super-resolution diffusion model. We also propose CT-ViT, which efficiently compresses CT volumes while preserving auto-regressiveness in-depth, enabling the generation of 3D CT volumes with variable numbers of axial slices. Our experiments demonstrate that GenerateCT can produce realistic, high-resolution, and high-fidelity 3D chest CT volumes consistent with medical language text prompts. We further investigate the potential of GenerateCT by training a model using generated CT volumes for multi-abnormality classification of chest CT volumes. Our contributions provide a valuable foundation for future research in text-conditional 3D medical image generation and have the potential to accelerate advancements in medical imaging research. Our code, pre-trained models, and generated data are available at https://github.com/ibrahimethemhamamci/GenerateCT.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces GenerateCT, the first method for text-conditional computed tomography (CT) generation, addressing the limitations in 3D medical imaging research and making the entire framework open-source.'}",http://arxiv.org/pdf/2305.16037
-chinese asr and ner improvement based on whisper fine-tuning,Hao Yang,"Based on 680k hours of weakly supervised multilingual and multi-task speech transcription/translation data, Whisper [1] has developed a robust system for both Automated Speech Recognition (ASR) and Speech Translation (ST). Whisper provides a simple model architecture based on Mel spectrum + two-layer convolution + Seq2seq Transformer architecture, which is easy to fine-tune on conditional generation tasks. This paper analyzes how to fine-tune Chinese ASR [2] and NER tasks based on Whisper, including (1) how to design different prompts for different generative tasks; (2) how to train ASR and NER tasks at the same time; (3) whether the performance can be further improved by using weak supervision for data enhancement. Experiments based on AISHELL [3] and AISHELL-NER [4] data, and multi-task fine-tuning based on Whisper can effectively improve the performance of ASR and NER.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes how to fine-tune Chinese ASR and NER tasks based on Whisper, including how to design different prompts for different generative tasks and whether the performance can be further improved by using weak supervision for data enhancement.'}",
-evaluating deep generative models on cognitive tasks: a case study,Zhi–Bin Tang,,,https://link.springer.com/content/pdf/10.1007/s44163-023-00067-3.pdf
-prores: exploring degradation-aware visual prompt for universal image restoration,Jiaqi Ma,"Image restoration aims to reconstruct degraded images, e.g., denoising or deblurring. Existing works focus on designing task-specific methods and there are inadequate attempts at universal methods. However, simply unifying multiple tasks into one universal architecture suffers from uncontrollable and undesired predictions. To address those issues, we explore prompt learning in universal architectures for image restoration tasks. In this paper, we present Degradation-aware Visual Prompts, which encode various types of image degradation, e.g., noise and blur, into unified visual prompts. These degradation-aware prompts provide control over image processing and allow weighted combinations for customized image restoration. We then leverage degradation-aware visual prompts to establish a controllable and universal model for image restoration, called ProRes, which is applicable to an extensive range of image restoration tasks. ProRes leverages the vanilla Vision Transformer (ViT) without any task-specific designs. Furthermore, the pre-trained ProRes can easily adapt to new tasks through efficient prompt tuning with only a few images. Without bells and whistles, ProRes achieves competitive performance compared to task-specific methods and experiments can demonstrate its ability for controllable restoration and adaptation for new tasks. The code and models will be released in \url{https://github.com/leonmakise/ProRes}.","{'model': 'tldr@v2.0.0', 'text': 'Dgradation-aware Visual Prompts are presented, which encode various types of image degradation, e.g., noise and blur, into unified visual prompts, which provide control over image processing and allow weighted combinations for customized image restoration.'}",http://arxiv.org/pdf/2306.13653
-deep learning in chatgpt - a survey,A.M. Jasmine Hashana,"Abstract-As a subset of machine learning, deep learning makes use of multiple-layer neural networks to learn with available data and make decisions or predictions. A large language model called ChatGPT is based on deep learning, specifically a type of neural network called a transformer. ChatGPT's transformer architecture uses attention mechanisms to focus on the most important parts of the input, allowing it to process and comprehend a large amount of text data. In order for the model to comprehend the context and meaning of natural language text, it is trained on a huge database of text, including articles and books. One of the main importance of using deep learning in ChatGPT is its intelligence to understand relationships and patterns from the input text and generate or predict new text that is homogeneous to the input/training data. Because of this, ChatGPT is able to respond to questions and prompts in a manner that is comparable to that of a human, making it useful for a wide scope of natural language processing missions like translating languages, summarizing texts, and responding to questions. It's worth noting that, while deep learning has been highly effective in ChatGPT, it is not without its limitations. To train, deep learning models can be very complex and require a lot of data and computing power.","{'model': 'tldr@v2.0.0', 'text': 'The transformer architecture in ChatGPT is able to respond to questions and prompts in a manner that is comparable to that of a human, making it useful for a wide scope of natural language processing missions like translating languages, summarizing texts, and responding to questions.'}",
-attention satisfies: a constraint-satisfaction lens on factual errors of language models,Mert Yuksekgonul,"We investigate the internal behavior of Transformer-based Large Language Models (LLMs) when they generate factually incorrect text. We propose modeling factual queries as Constraint Satisfaction Problems and use this framework to investigate how the model interacts internally with factual constraints. Specifically, we discover a strong positive relation between the model's attention to constraint tokens and the factual accuracy of its responses. In our curated suite of 11 datasets with over 40,000 prompts, we study the task of predicting factual errors with the Llama-2 family across all scales (7B, 13B, 70B). We propose SAT Probe, a method probing self-attention patterns, that can predict constraint satisfaction and factual errors, and allows early error identification. The approach and findings demonstrate how using the mechanistic understanding of factuality in LLMs can enhance reliability.","{'model': 'tldr@v2.0.0', 'text': ""A strong positive relation is found between the model's attention to constraint tokens and the factual accuracy of its responses and the approach and findings demonstrate how using the mechanistic understanding of factuality in LLMs can enhance reliability.""}",https://arxiv.org/pdf/2309.15098
-can ai put gamma-ray astrophysicists out of a job?,S. Spencer,"In what will likely be a litany of generative-model-themed arXiv submissions celebrating April the 1st, we evaluate the capacity of state-of-the-art transformer models to create a paper detailing the detection of a Pulsar Wind Nebula with a non-existent Imaging Atmospheric Cherenkov Telescope (IACT) Array. We do this to evaluate the ability of such models to interpret astronomical observations and sources based on language information alone, and to assess potential means by which fraudulently generated scientific papers could be identified during peer review (given that reliable generative model watermarking has yet to be deployed for these tools). We conclude that our jobs as astronomers are safe for the time being. From this point on, prompts given to ChatGPT and Stable Diffusion are shown in orange, text generated by ChatGPT is shown in black, whereas analysis by the (human) authors is in blue.",,http://arxiv.org/pdf/2303.17853
-a statistical perspective on retrieval-based models,Soumya Basu,"Many modern high-performing machine learning models increasingly rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Despite a growing literature showcasing the promise of these retrieval-based models , their theoretical underpinnings remain under-explored. In this paper, we present a formal treatment of retrieval-based models to characterize their performance via a novel statistical perspective. In particular, we study two broad classes of retrieval-based clas-sification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall performance. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a formal treatment of retrieval-based models to characterize their performance via a novel statistical perspective and shows that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall performance.'}",
-content moderation for evolving policies using binary question answering,S. S. Mullick,"Content moderation on social media is governed by policies that are intricate and frequently updated with evolving world events. However, automated content moderation systems often restrict easy adaptation to policy changes and are expected to learn policy intricacies from limited amounts of labeled data, which make effective policy compliance challenging. We propose to model content moderation as a binary question answering problem where the questions validate the loosely coupled themes constituting a policy. A decision logic is applied on top to aggregate the theme-specific validations. This way the questions pass theme information to a transformer network as explicit policy prompts, that in turn enables explainability. This setting further allows for faster adaptation to policy updates by leveraging zero-shot capabilities of pre-trained transformers. We showcase improved recall for our proposed method at 95\% precision on two proprietary datasets of social media posts and comments respectively annotated under curated Hate Speech and Commercial Spam policies.","{'model': 'tldr@v2.0.0', 'text': 'Improved recall is showcased for the proposed binary question answering problem where the questions validate the loosely coupled themes constituting a policy, which allows for faster adaptation to policy updates by leveraging zero-shot capabilities of pre-trained transformers.'}",https://aclanthology.org/2023.acl-industry.54.pdf
-transformer assembling technology interactive simulation system and method,赵坚,"The invention relates to an assembling technology interactive simulation system and method based on virtual reality, a virtual assembling technology and cross platforms, in particular to a transformer assembling technology interactive simulation system and method. According to the transformer assembling technology interactive simulation system and method, a three-dimensional modeling tool, a three-dimensional model integrating tool and a somatosensory camera are combined through a virtual reality technology and by combining an image processing technology, a three-dimensional simulation technology and other related technologies, the set of assembling technology interactive simulation system based on a virtual prototype is developed, the whole process of production and assembling operation of the transformer assembling technology is simulated on a computer, and meanwhile real-time supervision main point prompts are combined to carry out training and guidance on onsite assembling workers. The transformer assembling technology interactive simulation system and method can simulate and present the transformer assembling process, and more importantly, training cost is reduced when training of an electric power system is carried out on assembling and construction personnel and when assembling experiments are carried out; training time is shortened, and training efficiency is improved; requirements for training sites are lowered.","{'model': 'tldr@v2.0.0', 'text': 'The transformer assembling technology interactive simulation system and method can simulate and present the transformer assembling process, and more importantly, training cost is reduced when training of an electric power system is carried out on assembling and construction personnel and when assembling experiments are carried out.'}",
-making humanoid robots teaching assistants by using natural language processing (nlp) cloud-based services,A. Lekova,"Humanoid robots have a substantial potential to serve as teaching and social assistants. However, the expectations of the children from robots to interact like humans are huge. This study presents a general model for understanding the natural language in human-robot interaction by applying Generative Pre-trained Transformer (GPT) language models as a service in the Internet of Things. Thus, the physical presence of the robot can help in fine-tuning the GPT model by prompts derived from the environmental context and subsequent robot actions for embodiment understanding of the GPT outputs. The model uses web or cloud services for Natural Language Processing (NLP) to produce and play human-like text, question answering or text generation. Verbal questions are processed either via a local speech recognition software or via a Speech-to-Text (STT) cloud service. The converted question into machine-readable code is sent to one of the GPT language models with zero- or few-shot learning prompts. GPT-J model has been tested and deployed either in the web or cloud with options for varying the parameters for controlling the haphazardness of the generated text. The robot produces human-like text by using Text-to-Speech (TTS) cloud services that convert the response into audio format rendered on the robot to be played. Useful requirements how the model to be used in order to be feasible were determined based on the evaluation of the outputs given from the different NLP and GPT-J web or cloud-services. We designed and implemented the model in order to be used by a humanoid NAO-type robot in the speech language therapy practice, however it can be used for other open-source and programmable robots and in different contexts.","{'model': 'tldr@v2.0.0', 'text': 'A general model for understanding the natural language in human-robot interaction by applying Generative Pre-trained Transformer (GPT) language models as a service in the Internet of Things to help in fine-tuning the GPT model.'}",https://www.extrica.com/article/22720/pdf
-prompt generation networks for input-based adaptation of frozen vision transformers,J. Loedeman,"With the introduction of the transformer architecture in computer vision, increasing model scale has been demonstrated as a clear path to achieving performance and robustness gains. However, with model parameter counts reaching the billions, classical finetuning approaches are becoming increasingly limiting and even unfeasible when models become hosted as inference APIs, as in NLP. To this end, visual prompt learning, whereby a model is adapted by learning additional inputs, has emerged as a potential solution for adapting frozen and cloud-hosted models: During inference, this neither requires access to the internals of models' forward pass function, nor requires any post-processing. In this work, we propose the Prompt Generation Network (PGN) that generates high performing, input-dependent prompts by sampling from an end-to-end learned library of tokens. We further introduce the""prompt inversion""trick, with which PGNs can be efficiently trained in a latent space but deployed as strictly input-only prompts for inference. We show the PGN is effective in adapting pre-trained models to various new datasets: It surpasses previous methods by a large margin on 12/12 datasets and even outperforms full-finetuning on 5/12, while requiring 100x less parameters.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt Generation Network (PGN) is proposed that generates high performing, input-dependent prompts by sampling from an end-to-end learned library of tokens, and the ""prompt inversion"" trick is introduced, with which PGNs can be efficiently trained in a latent space but deployed as strictly input-only prompts for inference.'}",
-optimizing language models for argumentative reasoning,Luke Thorburn,"Large transformer-based causal language models are capable of strong performance on many natural language processing tasks. Here, we systematically evaluate the performance of the 2.7 billion parameter GPT Neo pre-trained language model on 6 argumentative reasoning tasks under 5 different optimization strategies, including prompt programming, soft prompts, and parameter tuning. We report both intrinsic evaluation metrics (perplexity), and extrinsic measures of the coherence of model outputs, as judged by an expert human rater. With a few exceptions, the rate at which models produced coherent responses ranged from 15-50%. In contrast, human performance (users of the Kialo argument mapping platform) ranged from 65-82% coherent, depending on the task. These results suggest that larger, suitably optimized language models may be capable of supporting authors and auditors of natural language argument maps in human-in-the-loop settings. We share our finetuned models and code.","{'model': 'tldr@v2.0.0', 'text': 'This work systematically evaluates the performance of the 2.7 billion parameter GPT Neo pre-trained language model on 6 argumentative reasoning tasks under 5 different optimization strategies, including prompt programming, soft prompts, and parameter tuning.'}",
-finding patterns in knowledge attribution for transformers,Jeevesh Juneja,"We analyze the Knowledge Neurons framework for the attribution of factual and relational knowledge to particular neurons in the transformer network. We use a 12-layer multi-lingual BERT model for our experiments. Our study reveals various interesting phenomena. We observe that mostly factual knowledge can be attributed to middle and higher layers of the network($\ge 6$). Further analysis reveals that the middle layers($6-9$) are mostly responsible for relational information, which is further refined into actual factual knowledge or the""correct answer""in the last few layers($10-12$). Our experiments also show that the model handles prompts in different languages, but representing the same fact, similarly, providing further evidence for effectiveness of multi-lingual pre-training. Applying the attribution scheme for grammatical knowledge, we find that grammatical knowledge is far more dispersed among the neurons than factual knowledge.","{'model': 'tldr@v2.0.0', 'text': 'This study analyzes the Knowledge Neurons framework for the attribution of factual and relational knowledge to particular neurons in the transformer network and finds that grammatical knowledge is far more dispersed among the neurons than factual knowledge.'}",http://arxiv.org/pdf/2205.01366
-prompt enhanced generative mrc framework for pancreatic cancer ner,Zhendong Tan,"Medical Named Entity Recognition (NER) is a fundamental but challenging task due to the lack of specialized entity datasets like tumor entities, which are often overlapped and discontinuous. In this paper, we propose a novel Prompt Enhanced Generative Machine Reading Comprehension Framework (PGMRC) to improve the overlapped and discontinuous NER performance. Specifically, we formulate NER as a Machine Reading Comprehension (MRC) task and employ a pre-trained encoder-decoder module to generate entity span sequences according to their entity query. In this way, we adopt query to guide the model to focus on answer entities in context, which can naturally solve entity overlap and alleviate the exposure bias of the generative model. Then, we introduce continuous prompts to the self-attention mechanism in Transformer to reduce the dependence on manually constructed queries. In addition, we annotate 875 pathological documents of pancreatic cancer and construct a Chinese pathological NER dataset (PAN) containing overlapped and discontinuous entities. Finally, we conduct our experiments on three widely used benchmarks (GENIA, ACE04, ACE05) and our dataset PAN. Experiments have demonstrated its effectiveness and better performance than state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper forms NER as a Machine Reading Comprehension task and employs a pre-trained encoder-decoder module to generate entity span sequences according to their entity query and introduces continuous prompts to the self-attention mechanism in Transformer to reduce the dependence on manually constructed queries.'}",
-bootstrapping vision-language learning with decoupled language pre-training,Yiren Jian,"We present a novel methodology aimed at optimizing the application of frozen large language models (LLMs) for resource-intensive vision-language (VL) pre-training. The current paradigm uses visual features as prompts to guide language models, with a focus on determining the most relevant visual features for corresponding text. Our approach diverges by concentrating on the language component, specifically identifying the optimal prompts to align with visual features. We introduce the Prompt-Transformer (P-Former), a model that predicts these ideal prompts, which is trained exclusively on linguistic data, bypassing the need for image-text pairings. This strategy subtly bifurcates the end-to-end VL training process into an additional, separate stage. Our experiments reveal that our framework significantly enhances the performance of a robust image-to-text baseline (BLIP-2), and effectively narrows the performance gap between models trained with either 4M or 129M image-text pairs. Importantly, our framework is modality-agnostic and flexible in terms of architectural design, as validated by its successful application in a video learning task using varied base modules. The code will be made available at https://github.com/yiren-jian/BLIText.","{'model': 'tldr@v2.0.0', 'text': 'The Prompt-Transformer (P-Former) is introduced, a model that predicts these ideal prompts to align with visual features, which is trained exclusively on linguistic data, bypassing the need for image-text pairings.'}",https://arxiv.org/pdf/2307.07063
-exploring the hurdles in integrating artificial intelligence in medical education: a talk with chatgpt,R. Vignesh,"The integration of artificial intelligence (AI) in the medical field is rapidly increasing, and its importance is recognised as it can analyse medical information and improve patient outcomes. The Chat Generative Pre-trained Transformer (ChatGPT) by OpenAI is a language model that can respond to prompts in a human-like manner and has the potential to transform medical education. In this conversation, the authors engage with ChatGPT to discuss the challenges associated with the application of AI in medical education. ChatGPT acknowledges the limitations of AI in medical education and the importance of students demonstrating their understanding during assessments. It also discusses the need for academic policies to ensure the responsible and ethical use of AI in medical education.","{'model': 'tldr@v2.0.0', 'text': 'In this conversation, the authors engage with ChatGPT to discuss the challenges associated with the application of AI in medical education and the need for academic policies to ensure the responsible and ethical use of AIIn medical education.'}",
-prompt-based ingredient-oriented all-in-one image restoration,Huiyu Gao,"Image restoration aims to recover the high-quality images from their degraded observations. Since most existing methods have been dedicated into single degradation removal, they may not yield optimal results on other types of degradations, which do not satisfy the applications in real world scenarios. In this paper, we propose a novel data ingredient-oriented approach that leverages prompt-based learning to enable a single model to efficiently tackle multiple image degradation tasks. Specifically, we utilize a encoder to capture features and introduce prompts with degradation-specific information to guide the decoder in adaptively recovering images affected by various degradations. In order to model the local invariant properties and non-local information for high-quality image restoration, we combined CNNs operations and Transformers. Simultaneously, we made several key designs in the Transformer blocks (multi-head rearranged attention with prompts and simple-gate feed-forward network) to reduce computational requirements and selectively determines what information should be persevered to facilitate efficient recovery of potentially sharp images. Furthermore, we incorporate a feature fusion mechanism further explores the multi-scale information to improve the aggregated features. The resulting tightly interlinked hierarchy architecture, named as CAPTNet, extensive experiments demonstrate that our method performs competitively to the state-of-the-art.","{'model': 'tldr@v2.0.0', 'text': 'A novel data ingredient-oriented approach that leverages prompt-based learning to enable a single model to efficiently tackle multiple image degradation tasks, and incorporates a feature fusion mechanism that further explores the multi-scale information to improve the aggregated features.'}",https://arxiv.org/pdf/2309.03063
-hierarchical prompt tuning for few-shot multi-task learning,Jingping Liu,"Prompt tuning has enhanced the performance of Pre-trained Language Models for multi-task learning in few-shot scenarios. However, existing studies fail to consider that the prompts among different layers in Transformer are different due to the diverse information learned at each layer. In general, the bottom layers in the model tend to capture low-level semantic or structural information, while the upper layers primarily acquire task-specific knowledge. Hence, we propose a novel hierarchical prompt tuning model for few-shot multi-task learning to capture this regularity. The designed model mainly consists of three types of prompts: shared prompts, auto-adaptive prompts, and task-specific prompts. Shared prompts facilitate the sharing of general information across all tasks. Auto-adaptive prompts dynamically select and integrate relevant prompt information from all tasks into the current task. Task-specific prompts concentrate on learning task-specific knowledge. To enhance the model's adaptability to diverse inputs, we introduce deep instance-aware language prompts as the foundation for constructing the above prompts. To evaluate the effectiveness of our proposed method, we conduct extensive experiments on multiple widely-used datasets. The experimental results demonstrate that the proposed method achieves state-of-the-art performance for multi-task learning in few-shot settings and outperforms ChatGPT in the full-data setting.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel hierarchical prompt tuning model that achieves state-of-the-art performance for multi-task learning in few-shot settings and outperforms ChatGPT in the full-data setting.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3614913
-pm-detr: domain adaptive prompt memory for object detection with transformers,Peidong Jia,"The Transformer-based detectors (i.e., DETR) have demonstrated impressive performance on end-to-end object detection. However, transferring DETR to different data distributions may lead to a significant performance degradation. Existing adaptation techniques focus on model-based approaches, which aim to leverage feature alignment to narrow the distribution shift between different domains. In this study, we propose a hierarchical Prompt Domain Memory (PDM) for adapting detection transformers to different distributions. PDM comprehensively leverages the prompt memory to extract domain-specific knowledge and explicitly constructs a long-term memory space for the data distribution, which represents better domain diversity compared to existing methods. Specifically, each prompt and its corresponding distribution value are paired in the memory space, and we inject top M distribution-similar prompts into the input and multi-level embeddings of DETR. Additionally, we introduce the Prompt Memory Alignment (PMA) to reduce the discrepancy between the source and target domains by fully leveraging the domain-specific knowledge extracted from the prompt domain memory. Extensive experiments demonstrate that our method outperforms state-of-the-art domain adaptive object detection methods on three benchmarks, including scene, synthetic to real, and weather adaptation. Codes will be released.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a hierarchical Prompt Domain Memory (PDM), which comprehensively leverages the prompt memory to extract domain-specific knowledge and explicitly constructs a long-term memory space for the data distribution, which represents better domain diversity compared to existing methods.'}",http://arxiv.org/pdf/2307.00313
-visual prompt flexible-modal face anti-spoofing,Zitong Yu,"Recently, vision transformer based multimodal learning methods have been proposed to improve the robustness of face anti-spoofing (FAS) systems. However, multimodal face data collected from the real world is often imperfect due to missing modalities from various imaging sensors. Recently, flexible-modal FAS~\cite{yu2023flexible} has attracted more attention, which aims to develop a unified multimodal FAS model using complete multimodal face data but is insensitive to test-time missing modalities. In this paper, we tackle one main challenge in flexible-modal FAS, i.e., when missing modality occurs either during training or testing in real-world situations. Inspired by the recent success of the prompt learning in language models, we propose \textbf{V}isual \textbf{P}rompt flexible-modal \textbf{FAS} (VP-FAS), which learns the modal-relevant prompts to adapt the frozen pre-trained foundation model to downstream flexible-modal FAS task. Specifically, both vanilla visual prompts and residual contextual prompts are plugged into multimodal transformers to handle general missing-modality cases, while only requiring less than 4\% learnable parameters compared to training the entire model. Furthermore, missing-modality regularization is proposed to force models to learn consistent multimodal feature embeddings when missing partial modalities. Extensive experiments conducted on two multimodal FAS benchmark datasets demonstrate the effectiveness of our VP-FAS framework that improves the performance under various missing-modality cases while alleviating the requirement of heavy model re-training.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the recent success of the prompt learning in language models, a VP-FAS framework is proposed, which learns the modal-relevant prompts to adapt the frozen pre-trained foundation model to downstream flexible-modal FAS task.'}",https://arxiv.org/pdf/2307.13958
-harnessing gpt-3.5-turbo for rhetorical role prediction in legal cases,Anas Belfathi,"We propose a comprehensive study of one-stage elicitation techniques for querying a large pre-trained generative transformer (GPT-3.5-turbo) in the rhetorical role prediction task of legal cases. This task is known as requiring textual context to be addressed. Our study explores strategies such as zero-few shots, task specification with definitions and clarification of annotation ambiguities, textual context and reasoning with general prompts and specific questions. We show that the number of examples, the definition of labels, the presentation of the (labelled) textual context and specific questions about this context have a positive influence on the performance of the model. Given non-equivalent test set configurations, we observed that prompting with a few labelled examples from direct context can lead the model to a better performance than a supervised fined-tuned multi-class classifier based on the BERT encoder (weighted F1 score of = 72%). But there is still a gap to reach the performance of the best systems = 86%) in the LegalEval 2023 task which, on the other hand, require dedicated resources, architectures and training.","{'model': 'tldr@v2.0.0', 'text': 'It is observed that prompting with a few labelled examples from direct context can lead the model to a better performance than a supervised fined-tuned multi-class classifier based on the BERT encoder, but there is still a gap to reach the performance of the best systems in the LegalEval 2023 task.'}",
-chatgpt in forensic sciences: a new pandora’s box with advantages and challenges to pay attention,R. Dinis-Oliveira," ChatGPT is a variant of the GPT (Generative Pre-trained Transformer) language model that uses large amounts of text-based training data and a transformer architecture to generate human-like text adjusted to the received prompts. ChatGPT presents several advantages in forensic sciences, namely constituting a virtual assistant to aid lawyers, judges, and victims in managing and interpreting forensic expert data. But what would happen if ChatGPT began to be used to produce forensic expertise reports? Despite its potential applications, the use of ChatGPT and other Large Language Models (LLM) and artificial intelligence (AI) tools in forensic writing also poses ethical and legal concerns, which are discussed in this perspective together with some expected future perspectives.","{'model': 'tldr@v2.0.0', 'text': 'Despite its potential applications, the use of ChatGPT and other Large Language Models and artificial intelligence tools in forensic writing also poses ethical and legal concerns, which are discussed in this perspective together with some expected future perspectives.'}",https://academic.oup.com/fsr/advance-article-pdf/doi/10.1093/fsr/owad039/52412442/owad039.pdf
-fake news detection in low resource languages using setfit framework,Amin Abdedaiem,"Social media has become an integral part of people’s lives, resulting in a constant flow of information. However, a concerning trend has emerged with the rapid spread of fake news, attributed to the lack of verification mechanisms. Fake news has far-reaching consequences, influencing public opinion, disrupting democracy, fuelingsocial tensions, and impacting various domains such as health, environment, and the economy. In order to identify fake news with data sparsity, especially with low resources languages such as Arabic and its dialects, we propose a few-shot learning fake news detection model based on sentence transformer fine-tuning, utilizing no crafted prompts and language model with few parameters. The experimental results prove that the proposed method can achieve higher performances with fewer news samples. This approach provided 71% F1 score on the Algerian dialect fake news dataset and 70% F1 score on the Modern Standard Arabic (MSA) version of the same dataset, which proves that the approach can work on the standard Arabic and its dialects. Therefore, the proposed model can identify fake news in several domains concerning the Algerian community such as politics, COVID-19, tourism, e-commerce, sport, accidents, and car prices.","{'model': 'tldr@v2.0.0', 'text': 'A few-shot learning fake news detection model based on sentence transformer fine-tuning, utilizing no crafted prompts and language model with few parameters is proposed, which proves that the approach can work on the standard Arabic and its dialects.'}",https://journal.iberamia.org/index.php/intartif/article/download/1151/198
-on the relationship between skill neurons and robustness in prompt tuning,Leon Ackermann,"Prompt Tuning is a popular parameter-efficient finetuning method for pre-trained large language models (PLMs). Recently, based on experiments with RoBERTa, it has been suggested that Prompt Tuning activates specific neurons in the transformer's feed-forward networks, that are highly predictive and selective for the given task. In this paper, we study the robustness of Prompt Tuning in relation to these""skill neurons"", using RoBERTa and T5. We show that prompts tuned for a specific task are transferable to tasks of the same type but are not very robust to adversarial data, with higher robustness for T5 than RoBERTa. At the same time, we replicate the existence of skill neurons in RoBERTa and further show that skill neurons also seem to exist in T5. Interestingly, the skill neurons of T5 determined on non-adversarial data are also among the most predictive neurons on the adversarial data, which is not the case for RoBERTa. We conclude that higher adversarial robustness may be related to a model's ability to activate the relevant skill neurons on adversarial data.","{'model': 'tldr@v2.0.0', 'text': ""Higher adversarial robustness may be related to a model's ability to activate the relevant skill neurons on adversarial data, with higher robustness for T5 than RoBERTa.""}",https://arxiv.org/pdf/2309.12263
-dcpt: darkness clue-prompted tracking in nighttime uavs,Jiawen Zhu,"Existing nighttime unmanned aerial vehicle (UAV) trackers follow an""Enhance-then-Track""architecture - first using a light enhancer to brighten the nighttime video, then employing a daytime tracker to locate the object. This separate enhancement and tracking fails to build an end-to-end trainable vision system. To address this, we propose a novel architecture called Darkness Clue-Prompted Tracking (DCPT) that achieves robust UAV tracking at night by efficiently learning to generate darkness clue prompts. Without a separate enhancer, DCPT directly encodes anti-dark capabilities into prompts using a darkness clue prompter (DCP). Specifically, DCP iteratively learns emphasizing and undermining projections for darkness clues. It then injects these learned visual prompts into a daytime tracker with fixed parameters across transformer layers. Moreover, a gated feature aggregation mechanism enables adaptive fusion between prompts and between prompts and the base model. Extensive experiments show state-of-the-art performance for DCPT on multiple dark scenario benchmarks. The unified end-to-end learning of enhancement and tracking in DCPT enables a more trainable system. The darkness clue prompting efficiently injects anti-dark knowledge without extra modules. Code is available at https://github.com/bearyi26/DCPT.","{'model': 'tldr@v2.0.0', 'text': 'A novel architecture called Darkness Clue-Prompted Tracking (DCPT) is proposed that achieves robust UAV tracking at night by efficiently learning to generate darkness clue prompts and efficiently injects anti-dark knowledge without extra modules.'}",https://arxiv.org/pdf/2309.10491
-comparative analysis of drug-gpt and chatgpt llms for healthcare insights: evaluating accuracy and relevance in patient and hcp contexts,Giorgos Lysandrou,"This study presents a comparative analysis of three Generative Pre-trained Transformer (GPT) solutions in a question and answer (Q&A) setting: Drug-GPT 3, Drug-GPT 4, and ChatGPT, in the context of healthcare applications. The objective is to determine which model delivers the most accurate and relevant information in response to prompts related to patient experiences with atopic dermatitis (AD) and healthcare professional (HCP) discussions about diabetes. The results demonstrate that while all three models are capable of generating relevant and accurate responses, Drug-GPT 3 and Drug-GPT 4, which are supported by curated datasets of patient and HCP social media and message board posts, provide more targeted and in-depth insights. ChatGPT, a more general-purpose model, generates broader and more general responses, which may be valuable for readers seeking a high-level understanding of the topics but may lack the depth and personal insights found in the answers generated by the specialized Drug-GPT models. This comparative analysis highlights the importance of considering the language model's perspective, depth of knowledge, and currency when evaluating the usefulness of generated information in healthcare applications.","{'model': 'tldr@v2.0.0', 'text': ""A comparative analysis of three Generative Pre-trained Transformer (GPT) solutions in a question and answer (Q&A) setting highlights the importance of considering the language model's perspective, depth of knowledge, and currency when evaluating the usefulness of generated information in healthcare applications.""}",https://arxiv.org/pdf/2307.16850
-a tête-à-tête with chatgpt on the impact of artificial intelligence in medical education.,Vignesh Ramachandran,"Chat Generative Pre-Trained Transformer (ChatGPT) is an artificial intelligence (AI) language model developed by OpenAI. It is trained to process vast amounts of text and engage in human-like conversational interaction with users. Being accessible by all, it is widely used and its capabilities range from language translation, summarising long texts and creative writing. This article explores the potential role of ChatGPT in medical education and the possible concerns about the misuse of this technology through a conversation with ChatGPT itself via text prompts. The implications of this technology in medical education as told by ChatGPT are interesting and seemingly helpful for both the students and the tutors. However, this could be a double-edged sword considering the risks of compromised students' integrity and concerns of over-reliance. This also calls for counter strategies and policies in place to mitigate these risks.","{'model': 'tldr@v2.0.0', 'text': ""The implications of this technology in medical education as told by ChatGPT are interesting and seemingly helpful for both the students and the tutors, however, this could be a double-edged sword considering the risks of compromised students' integrity and concerns of over-reliance.""}",
-efficient domain adaptation of language models in asr systems using prompt-tuning,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems form a key component of various products across industry. Many of these ASR systems rely on a complex Acoustic Model (AM) whose output is rescored by a domain-specific Language Model (LM). Given the recent advancements in performance, their use in new domains is growing, and so is the memory, maintenance and data-collection costs for these domain-specific LMs. Particularly, with advent of parameter-heavy transformer based LMs (Devlin et al., 2019), maintaining multiple domain-specific LMs is practically infeasible. While on the other hand, using a generic LM for all domains falls short in performance when compared to multiple domain-specific LMs. Therefore, a need for a middle ground between performance and costs is evident. To overcome this problem, we bring forward a methodology based on recently proposed Prompt Tuning. Lester et al. (2021) introduced this idea of learning the token embeddings of the prompt used to prime a LM to a particular task. Prompts are special tokens describing a task which when appended to input data sample, helps the model understand and use this problem description to better solve the task. For example, to solve the machine translation task, instead of fine-tuning the transformer model with corresponding dataset, one can achieve comparable performance by just showing text describing machine translation to the powerful transformer-based LM. In prompt tuning, instead of providing this prompt manually to the model, one learn it from the labelled examples from the task. In our work, we generalize this idea from solving different tasks to domain adaptation. We identify that one can learn the promptembedding for not only different tasks but also dif-","{'model': 'tldr@v2.0.0', 'text': 'This work identifies that one can learn the promptembedding for not only different tasks but also differences in domain adaptation, and brings forward a methodology based on recently proposed Prompt Tuning.'}",
-all birds with one stone: multi-task learning for inference with one forward pass,,"Task-specific fine-tuning of pre-trained lan- 001 guage models like Transformers has shown 002 their effectiveness in various NLP tasks. To 003 achieve better storage efficiency and model 004 performance, Multi-task Learning (MTL) has 005 been studied to share model parameters 006 and utilize knowledge transfer between tasks. 007 However, in real applications where enormous 008 numbers of tasks (e.g., large sets of labels 009 to be classified) need to be conducted on a 010 large corpus, the inference efficiency is still 011 hindered by the number of tasks. For a doc- 012 ument with N sets of labels to be predicted, 013 recent MTL methods with adaptive modules or 014 prompts need to encode the input data N times 015 to extract the hidden representation needed for 016 the tasks. Notice that the hidden represen- 017 tation is not sharable between tasks, as task- 018 specific features are extracted at very bottom 019 layers in the Transformer. In this paper, we 020 seek to maintain the computational efficiency 021 of only requiring one forward pass for a doc- 022 ument to get a generalized feature for all N 023 tasks, without sacrificing overall model perfor- 024 mance. We design a prompt-sharing module 025 to let the model take all tasks into considera- 026 tions and output N heads simultaneously. We 027 also design a dynamic task scheduling mod- 028 ule to sample tasks according to their training 029 progress. In our evaluation, we show that our 030 method is able to outperform previous MTL 031 state-of-the-arts and single task fine-tuning by 032 0 . 4 − 1 . 5% on GLUE benchmark dataset. We 033 also perform comprehensive module analysis 034 to demonstrate the effectiveness and robust- 035 ness of our method. 036","{'model': 'tldr@v2.0.0', 'text': 'This paper seeks to maintain the computational efficiency of only requiring one forward pass for a doc- 022 ument to get a generalized feature for all N 023 tasks, without sacriflcing overall model perfor- 024 mance, and designs a prompt-sharing module to let the model take all tasks into considera- 026 tions and output N heads simultaneously.'}",
-evaluation of interest and coherence in machine generated stories,Dominic Callan,". Evaluation of the narrative text generated by machines has traditionally been a challenge, particularly when attempting to evaluate subjective elements such as interest or believability. Recent improve-ments in narrative machine text generation have been largely driven by the emergence of transformer-based language models, trained on mas-sive quantities of data. In this study, a corpus of stories is generated using the pre-trained GPT-Neo transformer model, with human-written prompts. The stories generated through this process are subsequently evaluated through both human evaluation and two automated metrics: BERTScore and BERT Next-Sentence-Prediction. The results show vari-ation in human evaluation results in comparison to automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.","{'model': 'tldr@v2.0.0', 'text': 'Results show vari-ation in human evaluation results in comparison to automated metrics, suggesting further work is required to train automated metrics to identify text that is defined as interesting by humans.'}",
-extracting creativity from narratives using distributional semantic modeling,D. Johnson,"Narrative text permeates our lives from job applications to journalistic stories to works of fiction. Developing automated metrics that capture creativity in narrative text has potentially far reaching implications. Human ratings of creativity in narrative text are labor-intensive, subjective, and difficult to replicate. Across 27 different story prompts and over 3,500 short stories, we used distributional semantic modeling to automate the assessment of creativity in narrative texts. We tested a new metric to capture one key component of creativity in writing – a writer’s ability to connect divergent ideas. We termed this metric, word-to-word semantic diversity (w2w SemDiv). We compared six models of w2w SemDiv that varied in their computational architecture. The best performing model employed Bidirectional Encoder Representations Transformer (BERT), which generates context-dependent numerical representations of words (i.e., embeddings). The BERT w2w SemDiv scores demonstrated impressive predictive power, explaining up to 72% of the variance in human creativity ratings, even exceeding human inter-rater reliability for some tasks. In addition, w2w SemDiv scores generalized across Ethnicity and English language proficiency, including individuals identifying as Hispanic and L2 English speakers. We provide a tutorial with R code (osf.io/ath2s) on how to compute w2w SemDiv. This code is incorporated into an online web app (semdis.wlu.psu.edu) where researchers and educators can upload a data file with stories and freely retrieve w2w SemDiv scores.","{'model': 'tldr@v2.0.0', 'text': 'A new metric to capture one key component of creativity in writing – a writer’s ability to connect divergent ideas is tested, word-to-word semantic diversity (w2w SemDiv), which generalized across Ethnicity and English language proficiency, including individuals identifying as Hispanic and L2 English speakers.'}",
-operating frequency characteristics of piezoelectric high-voltage power supply converter,Bai Chun-yu,"The operating frequency characteristics of high-voltage power supply converter based on piezoelectric transformer(PT) is researched.The analysis of fundamental wave was used to get the equivalent model of converter output circuit and then the frequency characteristics of converter were investigated in the process of charging.It is found that the output equivalent capacitance increases with the increase of the operating periods.The equivalent resistance prompts rapidly with the increase of operating periods and achieves the maximum,then descends slowly and tends to a steady-state.In the whole process of the converter charging,the converter operating frequency descends with the increment of operating period,but the change is small,the operating frequency is only changed about 0.8 kHz in 90 000 operating periods(about 1.1 socend).",,
-wearable sensor-based human activity recognition with transformer model,Iveta Dirgová Luptáková,"Computing devices that can recognize various human activities or movements can be used to assist people in healthcare, sports, or human–robot interaction. Readily available data for this purpose can be obtained from the accelerometer and the gyroscope built into everyday smartphones. Effective classification of real-time activity data is, therefore, actively pursued using various machine learning methods. In this study, the transformer model, a deep learning neural network model developed primarily for the natural language processing and vision tasks, was adapted for a time-series analysis of motion signals. The self-attention mechanism inherent in the transformer, which expresses individual dependencies between signal values within a time series, can match the performance of state-of-the-art convolutional neural networks with long short-term memory. The performance of the proposed adapted transformer method was tested on the largest available public dataset of smartphone motion sensor data covering a wide range of activities, and obtained an average identification accuracy of 99.2% as compared with 89.67% achieved on the same data by a conventional machine learning method. The results suggest the expected future relevance of the transformer model for human activity recognition.","{'model': 'tldr@v2.0.0', 'text': 'The transformer model, a deep learning neural network model developed primarily for the natural language processing and vision tasks, was adapted for a time-series analysis of motion signals, and the expected future relevance of the transformer model for human activity recognition is suggested.'}",https://www.mdpi.com/1424-8220/22/5/1911/pdf?version=1646126183
-an end-to-end transformer model for 3d object detection,Ishan Misra,"We propose 3DETR, an end-to-end Transformer based object detection model for 3D point clouds. Compared to existing detection methods that employ a number of 3D-specific inductive biases, 3DETR requires minimal modifications to the vanilla Transformer block. Specifically, we find that a standard Transformer with non-parametric queries and Fourier positional embeddings is competitive with specialized architectures that employ libraries of 3D-specific operators with hand-tuned hyperparameters. Nevertheless, 3DETR is conceptually simple and easy to implement, enabling further improvements by incorporating 3D domain knowledge. Through extensive experiments, we show 3DETR outperforms the well-established and highly optimized VoteNet baselines on the challenging ScanNetV2 dataset by 9.5%. Furthermore, we show 3DETR is applicable to 3D tasks beyond detection, and can serve as a building block for future research.","{'model': 'tldr@v2.0.0', 'text': 'It is found that a standard Transformer with non-parametric queries and Fourier positional embeddings is competitive with specialized architectures that employ libraries of 3D-specific operators with hand-tuned hyperparameters.'}",https://arxiv.org/pdf/2109.08141
-"embodied bert: a transformer model for embodied, language-guided visual task completion",Alessandro Suglia,"Language-guided robots performing home and office tasks must navigate in and interact with the world. Grounding language instructions against visual observations and actions to take in an environment is an open challenge. We present Embodied BERT (EmBERT), a transformer-based model which can attend to high-dimensional, multi-modal inputs across long temporal horizons for language-conditioned task completion. Additionally, we bridge the gap between successful object-centric navigation models used for non-interactive agents and the language-guided visual task completion benchmark, ALFRED, by introducing object navigation targets for EmBERT training. We achieve competitive performance on the ALFRED benchmark, and EmBERT marks the first transformer-based model to successfully handle the long-horizon, dense, multi-modal histories of ALFRED, and the first ALFRED model to utilize object-centric navigation targets.","{'model': 'tldr@v2.0.0', 'text': 'Embodied BERT (EmBERT) is presented, a transformer-based model which can attend to high-dimensional, multi-modal inputs across long temporal horizons for language-conditioned task completion and marks the first transformer- based model to successfully handle the long-horizon, dense, multi themodal histories of ALFRED, and the first AlFRED model to utilize object-centric navigation targets.'}",
-swin transformer: hierarchical vision transformer using shifted windows,Ze Liu,"This paper presents a new vision Transformer, called Swin Transformer, that capably serves as a general-purpose backbone for computer vision. Challenges in adapting Transformer from language to vision arise from differences between the two domains, such as large variations in the scale of visual entities and the high resolution of pixels in images compared to words in text. To address these differences, we propose a hierarchical Transformer whose representation is computed with Shifted windows. The shifted windowing scheme brings greater efficiency by limiting self-attention computation to non-overlapping local windows while also allowing for cross-window connection. This hierarchical architecture has the flexibility to model at various scales and has linear computational complexity with respect to image size. These qualities of Swin Transformer make it compatible with a broad range of vision tasks, including image classification (87.3 top-1 accuracy on ImageNet-1K) and dense prediction tasks such as object detection (58.7 box AP and 51.1 mask AP on COCO test-dev) and semantic segmentation (53.5 mIoU on ADE20K val). Its performance surpasses the previous state-of-the-art by a large margin of +2.7 box AP and +2.6 mask AP on COCO, and +3.2 mIoU on ADE20K, demonstrating the potential of Transformer-based models as vision backbones. The hierarchical design and the shifted window approach also prove beneficial for all-MLP architectures. The code and models are publicly available at https://github.com/microsoft/Swin-Transformer.","{'model': 'tldr@v2.0.0', 'text': 'A hierarchical Transformer whose representation is computed with Shifted windows, which has the flexibility to model at various scales and has linear computational complexity with respect to image size and will prove beneficial for all-MLP architectures.'}",https://arxiv.org/pdf/2103.14030
-advancing plain vision transformer toward remote sensing foundation model,Di Wang,"Large-scale vision foundation models have made significant progress in visual tasks on natural images, with vision transformers (ViTs) being the primary choice due to their good scalability and representation ability. However, large-scale models in remote sensing (RS) have not yet been sufficiently explored. In this article, we resort to plain ViTs with about 100 million parameters and make the first attempt to propose large vision models tailored to RS tasks and investigate how such large models perform. To handle the large sizes and objects of arbitrary orientations in RS images, we propose a new rotated varied-size window attention to replace the original full attention in transformers, which can significantly reduce the computational cost and memory footprint while learning better object representation by extracting rich context from the generated diverse windows. Experiments on detection tasks show the superiority of our model over all state-of-the-art models, achieving 81.24% mean average precision (mAP) on the DOTA-V1.0 dataset. The results of our models on downstream classification and segmentation tasks also show competitive performance compared to existing advanced methods. Further experiments show the advantages of our models in terms of computational complexity and data efficiency in transferring. The code and models will be released at https://github.com/ViTAE-Transformer/Remote-Sensing-RVSA.","{'model': 'tldr@v2.0.0', 'text': 'This article proposes a new rotated varied-size window attention to replace the original full attention in transformers, which can significantly reduce the computational cost and memory footprint while learning better object representation by extracting rich context from the generated diverse windows.'}",http://arxiv.org/pdf/2208.03987
-streaming automatic speech recognition with the transformer model,Niko Moritz,"Encoder-decoder based sequence-to-sequence models have demonstrated state-of-the-art results in end-to-end automatic speech recognition (ASR). Recently, the transformer architecture, which uses self-attention to model temporal context information, has been shown to achieve significantly lower word error rates (WERs) compared to recurrent neural network (RNN) based system architectures. Despite its success, the practical usage is limited to offline ASR tasks, since encoder-decoder architectures typically require an entire speech utterance as input. In this work, we propose a transformer based end-to-end ASR system for streaming ASR, where an output must be generated shortly after each spoken word. To achieve this, we apply time-restricted self-attention for the encoder and triggered attention for the encoder-decoder attention mechanism. Our proposed streaming transformer architecture achieves 2.8% and 7.3% WER for the ""clean"" and ""other"" test data of LibriSpeech, which to our knowledge is the best published streaming end-to-end ASR result for this task.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transformer based end-to-end ASR system for streaming ASR, where an output must be generated shortly after each spoken word, and applies time-restricted self-attention for the encoder and triggered attention for theEncoder-decoder attention mechanism.'}",https://arxiv.org/pdf/2001.02674
-a multiscale visualization of attention in the transformer model,Jesse Vig,"The Transformer is a sequence model that forgoes traditional recurrent architectures in favor of a fully attention-based approach. Besides improving performance, an advantage of using attention is that it can also help to interpret a model by showing how the model assigns weight to different input elements. However, the multi-layer, multi-head attention mechanism in the Transformer model can be difficult to decipher. To make the model more accessible, we introduce an open-source tool that visualizes attention at multiple scales, each of which provides a unique perspective on the attention mechanism. We demonstrate the tool on BERT and OpenAI GPT-2 and present three example use cases: detecting model bias, locating relevant attention heads, and linking neurons to model behavior.","{'model': 'tldr@v2.0.0', 'text': 'An open-source tool that visualizes attention at multiple scales, each of which provides a unique perspective on the attention mechanism in the Transformer model is introduced.'}",https://www.aclweb.org/anthology/P19-3007.pdf
-cnn-based transformer model for fault detection in power system networks,Jibin B. Thomas,"Fault detection and localization in electrical power lines has long been a crucial challenge for electrical engineers as it allows the detected fault to be isolated and recovered promptly. These faults, if neglected, can rupture the normal operation of the network and drastically damage the power lines and the equipment attached to it. The wastage of power and money due to these faults can be harmful to the economy of an industry or even a country. Therefore, efficient fault detection mechanisms have become crucial for the well-being of this power-hungry world. This research presents an end-to-end deep learning strategy to detect and localize symmetrical and unsymmetrical faults as well as high-impedance faults (HIFs) in a distribution system. This research proposes a novel deep convolutional neural network (CNN) transformer model to automatically detect the type and phase of the fault as well as the location of the fault. The proposed model utilizes 1-D deep CNNs for feature extraction and transformer encoder for sequence learning. The transformer encoder utilizes an attention mechanism to integrate the sequence embeddings and focus on significant time steps to learn long-term dependence to extract the context of the temporal current data. The different faults were simulated in MATLAB Simulink using IEEE 14-bus distribution system. The proposed models were found to produce better performance on the test database when evaluated using F1-Score, Matthews correlation coefficient (MCC), and accuracy. The models also produced better predictions on HIFs compared to conventional fault-detection techniques.",,
-a transformer model for retrosynthesis,Pavel Karpov,,"{'model': 'tldr@v2.0.0', 'text': 'A Transformer model for a retrosynthetic reaction prediction task is described and it is found that snapshot learning with averaging weights on learning rates minima works best.'}",
-lstm-embedding transformer model for los/nlos recognition in indoor localization,Lucong Chen,"The development of 5G networks has prompted higher localization accuracy requirements in indoor scenarios within the Industrial Internet of Things (IIoT). In the scenarios of indoor localization, mitigating the adverse effects caused by Non-Line-of-Sight (NLoS) propagation has been a longstanding and significant challenge. Despite the availability of numerous classifiers for distinguishing between Line-of-Sight (LoS) and NLoS paths, there are still some shortcomings in adequately capturing the long-term time dependence of signal sequences. In this study, to enhance the extraction of long-term dependencies and global information of time series, we propose the Long Short-Term Memory (LSTM)-embedding transformer model to identify the LoS and NLoS paths from the Power Delay Profile (PDP), which characterizes the amplitude decay of the signal over time. And the Time-of-Arrival (ToA) information of the LoS paths is exploited for localization. Lastly, to evaluate the robustness of our model, the model which is trained in the Indoor Factory Dense High (InF-DH) scenario is transferred to the Indoor Factory Sparse High (InF-SH) scenario. Experimental results show that our proposed model reached 96.04% recognition success in the InF-DH scenario and 93.27% recognition success in the migration scenario (InF-SH), with some improvements compared to the baseline model.","{'model': 'tldr@v2.0.0', 'text': 'To enhance the extraction of long-term dependencies and global information of time series, the Long Short-Term Memory-embedding transformer model is proposed to identify the LoS and NLoS paths from the Power Delay Profile (PDP), which characterizes the amplitude decay of the signal over time.'}",
-epigept: a pretrained transformer model for epigenomics,Zijing Gao,"The transformer-based models, such as GPT-31 and DALL-E2, have achieved unprecedented breakthroughs in the field of natural language processing and computer vision. The inherent similarities between natural language and biological sequences have prompted a new wave of inferring the grammatical rules underneath the biological sequences. In genomic study, it is worth noting that DNA sequences alone cannot explain all the gene activities due to epigenetic mechanism. To investigate this problem, we propose EpiGePT, a new transformer-based language pretrained model in epigenomics, for predicting genome-wide epigenomic signals by considering the mechanistic modeling of transcriptional regulation. Specifically, EpiGePT takes the context-specific activities of transcription factors (TFs) into consideration, which could offer deeper biological insights comparing to models trained on DNA sequence only. In a series of experiments, EpiGePT demonstrates state-of-the-art performance in a diverse epigenomic signals prediction tasks as well as new prediction tasks by fine-tuning. Furthermore, EpiGePT is capable of learning the cell-type-specific long-range interactions through the self-attention mechanism and interpreting the genetic variants that associated with human diseases. We expect that the advances of EpiGePT can shed light on understanding the complex regulatory mechanisms in gene regulation. We provide free online prediction service of EpiGePT through https://health.tsinghua.edu.cn/epigept/.","{'model': 'tldr@v2.0.0', 'text': 'EpiGePT, a new transformer-based language pretrained model in epigenomics, is proposed for predicting genome-wide epigenomic signals by considering the mechanistic modeling of transcriptional regulation and is capable of learning the cell-type-specific long-range interactions through the self-attention mechanism.'}",
-insights from attention: new approaches to visualizing transformer model output,Raphael Kunert,". Recent advancements in language models, particularly those based on the Transformer architecture, have led to remarkable achieve-ments in natural language processing. However, the increasing complex-ity and size of these models pose significant challenges for understanding their behavior and decision-making processes. In this work, we propose a set of new attention visualization techniques that address these challenges by improving model explainability and interpretability. The key improvements include new layouts that better handle the large number of tokens present in prompt and answer scenarios, making long distance attention relationships more comprehensible. Our techniques have the potential to enable researchers and practitioners to better understand the decision-making processes of large language models and identify potential sources of bias or errors. While detailed user studies and evalua-tions are outside the scope of this work, we discuss potential use cases for our visualization techniques and present directions for future research.","{'model': 'tldr@v2.0.0', 'text': 'A set of new attention visualization techniques that address challenges of large language models by improving model explainability and interpretability and making long distance attention relationships more comprehensible are proposed.'}",
-maskgit: masked generative image transformer,Huiwen Chang,"Generative transformers have experienced rapid popularity growth in the computer vision community in synthesizing high-fidelity and high-resolution images. The best generative transformer models so far, however, still treat an image naively as a sequence of tokens, and decode an image sequentially following the raster scan ordering (i.e. line-by-line). We find this strategy neither optimal nor efficient. This paper proposes a novel image synthesis paradigm using a bidirectional transformer decoder, which we term MaskGIT. During training, MaskGIT learns to predict randomly masked tokens by attending to tokens in all directions. At inference time, the model begins with generating all tokens of an image simultaneously, and then refines the image iteratively conditioned on the previous generation. Our experiments demonstrate that MaskGIT significantly outperforms the state-of-the-art transformer model on the ImageNet dataset, and accelerates autoregressive decoding by up to 48x. Besides, we illustrate that MaskGIT can be easily extended to various image editing tasks, such as inpainting, extrapolation, and image manipulation. Project page: masked-generative-image-transformer.github.io.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MaskGIT is a novel image synthesis paradigm using a bidirectional transformer decoder that significantly outperforms the state-of-the-art transformer model on the ImageNet dataset, and accelerates autoregressive decoding by up to 48x.'}",https://arxiv.org/pdf/2202.04200
-medical intervention duration estimation using language-enhanced transformer encoder with medical prompts,Yucheng Ruan,"In recent years, estimating the duration of medical intervention based on electronic health records (EHRs) has gained significant attention in the filed of clinical decision support. However, current models largely focus on structured data, leaving out information from the unstructured clinical free-text data. To address this, we present a novel language-enhanced transformer-based framework, which projects all relevant clinical data modalities (continuous, categorical, binary, and free-text features) into a harmonized language latent space using a pre-trained sentence encoder with the help of medical prompts. The proposed method enables the integration of information from different modalities within the cell transformer encoder and leads to more accurate duration estimation for medical intervention. Our experimental results on both US-based (length of stay in ICU estimation) and Asian (surgical duration prediction) medical datasets demonstrate the effectiveness of our proposed framework, which outperforms tailored baseline approaches and exhibits robustness to data corruption in EHRs.","{'model': 'tldr@v2.0.0', 'text': 'A novel language-enhanced transformer-based framework, which projects all relevant clinical data modalities into a harmonized language latent space using a pre-trained sentence encoder with the help of medical prompts, leads to more accurate duration estimation for medical intervention.'}",https://arxiv.org/pdf/2303.17408
-training tips for the transformer model,M. Popel,"Abstract This article describes our experiments in neural machine translation using the recent Tensor2Tensor framework and the Transformer sequence-to-sequence model (Vaswani et al., 2017). We examine some of the critical parameters that affect the final translation quality, memory usage, training stability and training time, concluding each experiment with a set of recommendations for fellow researchers. In addition to confirming the general mantra “more data and larger models”, we address scaling to multiple GPUs and provide practical tips for improved training regarding batch size, learning rate, warmup steps, maximum sentence length and checkpoint averaging. We hope that our observations will allow others to get better results given their particular hardware and data constraints.","{'model': 'tldr@v2.0.0', 'text': 'The experiments in neural machine translation using the recent Tensor2Tensor framework and the Transformer sequence-to-sequence model are described, confirming the general mantra “more data and larger models”.'}",https://content.sciendo.com/downloadpdf/journals/pralin/110/1/article-p43.pdf
-transformer transducer: a streamable speech recognition model with transformer encoders and rnn-t loss,Qian Zhang,"In this paper we present an end-to-end speech recognition model with Transformer encoders that can be used in a streaming speech recognition system. Transformer computation blocks based on self-attention are used to encode both audio and label sequences independently. The activations from both audio and label encoders are combined with a feed-forward layer to compute a probability distribution over the label space for every combination of acoustic frame position and label history. This is similar to the Recurrent Neural Network Transducer (RNN-T) model, which uses RNNs for information encoding instead of Transformer encoders. The model is trained with the RNN-T loss well-suited to streaming decoding. We present results on the LibriSpeech dataset showing that limiting the left context for self-attention in the Transformer layers makes decoding computationally tractable for streaming, with only a slight degradation in accuracy. We also show that the full attention version of our model beats the-state-of-the art accuracy on the LibriSpeech benchmarks. Our results also show that we can bridge the gap between full attention and limited attention versions of our model by attending to a limited number of future frames.","{'model': 'tldr@v2.0.0', 'text': 'An end-to-end speech recognition model with Transformer encoders that can be used in a streaming speech recognition system and shows that the full attention version of the model beats the-state-of-the art accuracy on the LibriSpeech benchmarks.'}",https://arxiv.org/pdf/2002.02562
-git: a generative image-to-text transformer for vision and language,Jianfeng Wang,"In this paper, we design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering. While generative models provide a consistent network architecture between pre-training and fine-tuning, existing work typically contains complex structures (uni/multi-modal encoder/decoder) and depends on external modules such as object detectors/taggers and optical character recognition (OCR). In GIT, we simplify the architecture as one image encoder and one text decoder under a single language modeling task. We also scale up the pre-training data and the model size to boost the model performance. Without bells and whistles, our GIT establishes new state of the arts on 12 challenging benchmarks with a large margin. For instance, our model surpasses the human performance for the first time on TextCaps (138.2 vs. 125.5 in CIDEr). Furthermore, we present a new scheme of generation-based image classification and scene text recognition, achieving decent performance on standard benchmarks. Codes are released at \url{https://github.com/microsoft/GenerativeImage2Text}.","{'model': 'tldr@v2.0.0', 'text': 'This paper design and train a Generative Image-to-text Transformer, GIT, to unify vision-language tasks such as image/video captioning and question answering and establishes new state of the arts on 12 challenging benchmarks with a large margin.'}",http://arxiv.org/pdf/2205.14100
-biogpt: generative pre-trained transformer for biomedical text generation and mining,Renqian Luo,"Pre-trained language models have attracted increasing attention in the biomedical domain, inspired by their great success in the general natural language domain. Among the two main branches of pre-trained language models in the general language domain, i.e. BERT (and its variants) and GPT (and its variants), the first one has been extensively studied in the biomedical domain, such as BioBERT and PubMedBERT. While they have achieved great success on a variety of discriminative downstream biomedical tasks, the lack of generation ability constrains their application scope. In this paper, we propose BioGPT, a domain-specific generative Transformer language model pre-trained on large-scale biomedical literature. We evaluate BioGPT on six biomedical natural language processing tasks and demonstrate that our model outperforms previous models on most tasks. Especially, we get 44.98%, 38.42% and 40.76% F1 score on BC5CDR, KD-DTI and DDI end-to-end relation extraction tasks, respectively, and 78.2% accuracy on PubMedQA, creating a new record. Our case study on text generation further demonstrates the advantage of BioGPT on biomedical literature to generate fluent descriptions for biomedical terms.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes BioGPT, a domain-specific generative Transformer language model pre-trained on large-scale biomedical literature and evaluates it on six biomedical natural language processing tasks and demonstrates that the model outperforms previous models on most tasks.'}",https://arxiv.org/pdf/2210.10341
-hts-at: a hierarchical token-semantic audio transformer for sound classification and detection,Ke Chen,"Audio classification is an important task of mapping audio samples into their corresponding labels. Recently, the transformer model with self-attention mechanisms has been adopted in this field. However, existing audio transformers require large GPU memories and long training time, meanwhile relying on pretrained vision models to achieve high performance, which limits the model’s scalability in audio tasks. To combat these problems, we introduce HTS-AT: an audio transformer with a hierarchical structure to reduce the model size and training time. It is further combined with a token-semantic module to map final outputs into class featuremaps, thus enabling the model for the audio event detection (i.e. localization in time). We evaluate HTS-AT on three datasets of audio classification where it achieves new state-of-the-art (SOTA) results on AudioSet and ESC50, and equals the SOTA on Speech Command V2. It also achieves better performance in event localization than the previous CNN-based models. Moreover, HTS-AT requires only 35% model parameters and 15% training time of the previous audio transformer. These results demonstrate the high performance and high efficiency of HTS-AT.","{'model': 'tldr@v2.0.0', 'text': 'HTS-AT is introduced: an audio transformer with a hierarchical structure to reduce the model size and training time, and is further combined with a token-semantic module to map final outputs into class featuremaps, thus enabling the model for the audio event detection and localization in time.'}",https://arxiv.org/pdf/2202.00874
-maxvit: multi-axis vision transformer,Zhengzhong Tu,"Transformers have recently gained significant attention in the computer vision community. However, the lack of scalability of self-attention mechanisms with respect to image size has limited their wide adoption in state-of-the-art vision backbones. In this paper we introduce an efficient and scalable attention model we call multi-axis attention, which consists of two aspects: blocked local and dilated global attention. These design choices allow global-local spatial interactions on arbitrary input resolutions with only linear complexity. We also present a new architectural element by effectively blending our proposed attention model with convolutions, and accordingly propose a simple hierarchical vision backbone, dubbed MaxViT, by simply repeating the basic building block over multiple stages. Notably, MaxViT is able to ''see'' globally throughout the entire network, even in earlier, high-resolution stages. We demonstrate the effectiveness of our model on a broad spectrum of vision tasks. On image classification, MaxViT achieves state-of-the-art performance under various settings: without extra data, MaxViT attains 86.5% ImageNet-1K top-1 accuracy; with ImageNet-21K pre-training, our model achieves 88.7% top-1 accuracy. For downstream tasks, MaxViT as a backbone delivers favorable performance on object detection as well as visual aesthetic assessment. We also show that our proposed model expresses strong generative modeling capability on ImageNet, demonstrating the superior potential of MaxViT blocks as a universal vision module. The source code and trained models will be available at https://github.com/google-research/maxvit.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces an efficient and scalable attention model, which consists of two aspects: blocked local and dilated global attention, and proposes a simple hierarchical vision backbone, dubbed MaxViT, by simply repeating the basic building block over multiple stages.'}",http://arxiv.org/pdf/2204.01697
-mat: mask-aware transformer for large hole image inpainting,Wenbo Li,"Recent studies have shown the importance of modeling long-range interactions in the inpainting problem. To achieve this goal, existing approaches exploit either standalone attention techniques or transformers, but usually under a low resolution in consideration of computational cost. In this paper, we present a novel transformer-based model for large hole inpainting, which unifies the merits of transformers and convolutions to efficiently process high-resolution images. We carefully design each component of our framework to guarantee the high fidelity and diversity of recovered images. Specifically, we customize an inpainting-oriented transformer block, where the attention module aggregates non-local information only from partial valid tokens, indicated by a dynamic mask. Extensive experiments demonstrate the state-of-the-art performance of the new model on multiple benchmark datasets. Code is released at https://github.com/fenglinglwb/MAT.","{'model': 'tldr@v2.0.0', 'text': 'A novel transformer-based model for large hole inpainting is presented, which unifies the merits of transformers and convolutions to efficiently process high-resolution images.'}",https://arxiv.org/pdf/2203.15270
-vision transformer adapter for dense predictions,Zhe Chen,"This work investigates a simple yet powerful dense prediction task adapter for Vision Transformer (ViT). Unlike recently advanced variants that incorporate vision-specific inductive biases into their architectures, the plain ViT suffers inferior performance on dense predictions due to weak prior assumptions. To address this issue, we propose the ViT-Adapter, which allows plain ViT to achieve comparable performance to vision-specific transformers. Specifically, the backbone in our framework is a plain ViT that can learn powerful representations from large-scale multi-modal data. When transferring to downstream tasks, a pre-training-free adapter is used to introduce the image-related inductive biases into the model, making it suitable for these tasks. We verify ViT-Adapter on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation. Notably, without using extra detection data, our ViT-Adapter-L yields state-of-the-art 60.9 box AP and 53.0 mask AP on COCO test-dev. We hope that the ViT-Adapter could serve as an alternative for vision-specific transformers and facilitate future research. The code and models will be released at https://github.com/czczup/ViT-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'The ViT-Adapter is proposed, which allows plain ViT to achieve comparable performance to vision-specific transformers and facilitate future research and is verified on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation.'}",http://arxiv.org/pdf/2205.08534
-ctrl: a conditional transformer language model for controllable generation,N. Keskar,"Large-scale language models show promising text generation capabilities, but users cannot easily control particular aspects of the generated text. We release CTRL, a 1.63 billion-parameter conditional transformer language model, trained to condition on control codes that govern style, content, and task-specific behavior. Control codes were derived from structure that naturally co-occurs with raw text, preserving the advantages of unsupervised learning while providing more explicit control over text generation. These codes also allow CTRL to predict which parts of the training data are most likely given a sequence. This provides a potential method for analyzing large amounts of data via model-based source attribution. We have released multiple full-sized, pretrained versions of CTRL at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'CTRL is released, a 1.63 billion-parameter conditional transformer language model, trained to condition on control codes that govern style, content, and task-specific behavior, providing more explicit control over text generation.'}",
-exploring the limits of transfer learning with a unified text-to-text transformer,Colin Raffel,"Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new ""Colossal Clean Crawled Corpus"", we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.","{'model': 'tldr@v2.0.0', 'text': 'This systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks and achieves state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more.'}",
-swinir: image restoration using swin transformer,Jingyun Liang,"Image restoration is a long-standing low-level vision problem that aims to restore high-quality images from low-quality images (e.g., downscaled, noisy and compressed images). While state-of-the-art image restoration methods are based on convolutional neural networks, few attempts have been made with Transformers which show impressive performance on high-level vision tasks. In this paper, we propose a strong baseline model SwinIR for image restoration based on the Swin Transformer. SwinIR consists of three parts: shallow feature extraction, deep feature extraction and high-quality image reconstruction. In particular, the deep feature extraction module is composed of several residual Swin Transformer blocks (RSTB), each of which has several Swin Transformer layers together with a residual connection. We conduct experiments on three representative tasks: image super-resolution (including classical, lightweight and real-world image super-resolution), image denoising (including grayscale and color image denoising) and JPEG compression artifact reduction. Experimental results demonstrate that SwinIR outperforms state-of-the-art methods on different tasks by up to 0.14∼0.45dB, while the total number of parameters can be reduced by up to 67%.","{'model': 'tldr@v2.0.0', 'text': 'A strong baseline model SwinIR is proposed for image restoration based on the Swin Transformer that outperforms state-of-the-art methods on different tasks by up to 0.14∼0.45dB, while the total number of parameters can be reduced byUp to 67%.'}",https://arxiv.org/pdf/2108.10257
-vivit: a video vision transformer,Anurag Arnab,"We present pure-transformer based models for video classification, drawing upon the recent success of such models in image classification. Our model extracts spatiotemporal tokens from the input video, which are then encoded by a series of transformer layers. In order to handle the long sequences of tokens encountered in video, we propose several, efficient variants of our model which factorise the spatial- and temporal-dimensions of the input. Although transformer-based models are known to only be effective when large training datasets are available, we show how we can effectively regularise the model during training and leverage pretrained image models to be able to train on comparatively small datasets. We conduct thorough ablation studies, and achieve state-of-the-art results on multiple video classification benchmarks including Kinetics 400 and 600, Epic Kitchens, Something-Something v2 and Moments in Time, outperforming prior methods based on deep 3D convolutional networks.","{'model': 'tldr@v2.0.0', 'text': 'This work shows how to effectively regularise the model during training and leverage pretrained image models to be able to train on comparatively small datasets, and achieves state-of-the-art results on multiple video classification benchmarks.'}",https://arxiv.org/pdf/2103.15691
-decision transformer: reinforcement learning via sequence modeling,Lili Chen,"We introduce a framework that abstracts Reinforcement Learning (RL) as a sequence modeling problem. This allows us to draw upon the simplicity and scalability of the Transformer architecture, and associated advances in language modeling such as GPT-x and BERT. In particular, we present Decision Transformer, an architecture that casts the problem of RL as conditional sequence modeling. Unlike prior approaches to RL that fit value functions or compute policy gradients, Decision Transformer simply outputs the optimal actions by leveraging a causally masked Transformer. By conditioning an autoregressive model on the desired return (reward), past states, and actions, our Decision Transformer model can generate future actions that achieve the desired return. Despite its simplicity, Decision Transformer matches or exceeds the performance of state-of-the-art model-free offline RL baselines on Atari, OpenAI Gym, and Key-to-Door tasks.","{'model': 'tldr@v2.0.0', 'text': 'Despite its simplicity, Decision Transformer matches or exceeds the performance of state-of-the-art model-free offline RL baselines on Atari, OpenAI Gym, and Key-to-Door tasks.'}",
-restormer: efficient transformer for high-resolution image restoration,Syed Waqas Zamir,"Since convolutional neural networks (CNNs) perform well at learning generalizable image priors from large-scale data, these models have been extensively applied to image restoration and related tasks. Recently, another class of neural architectures, Transformers, have shown significant performance gains on natural language and high-level vision tasks. While the Transformer model mitigates the shortcomings of CNNs (i.e., limited receptive field and inadaptability to input content), its computational complexity grows quadratically with the spatial resolution, therefore making it infeasible to apply to most image restoration tasks involving high-resolution images. In this work, we propose an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images. Our model, named Restoration Transformer (Restormer), achieves state-of-the-art results on several image restoration tasks, including image deraining, single-image motion deblurring, defocus deblurring (single-image and dual-pixel data), and image denoising (Gaussian grayscale/color denoising, and real image denoising). The source code and pre-trained models are available at https://github.com/swz30/Restormer.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images.'}",https://arxiv.org/pdf/2111.09881
-segmenter: transformer for semantic segmentation,Robin Strudel,"Image segmentation is often ambiguous at the level of individual image patches and requires contextual information to reach label consensus. In this paper we introduce Segmenter, a transformer model for semantic segmentation. In contrast to convolution-based methods, our approach allows to model global context already at the first layer and throughout the network. We build on the recent Vision Transformer (ViT) and extend it to semantic segmentation. To do so, we rely on the output embeddings corresponding to image patches and obtain class labels from these embed-dings with a point-wise linear decoder or a mask trans-former decoder. We leverage models pre-trained for image classification and show that we can fine-tune them on moderate sized datasets available for semantic segmentation. The linear decoder allows to obtain excellent results already, but the performance can be further improved by a mask transformer generating class masks. We conduct an extensive ablation study to show the impact of the different parameters, in particular the performance is better for large models and small patch sizes. Segmenter attains excellent results for semantic segmentation. It outperforms the state of the art on both ADE20K and Pascal Context datasets and is competitive on Cityscapes.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Segmenter, a transformer model for semantic segmentation that outperforms the state of the art on both ADE20K and Pascal Context datasets and is competitive on Cityscapes.'}",https://arxiv.org/pdf/2105.05633
-vilt: vision-and-language transformer without convolution or region supervision,Wonjae Kim,"Vision-and-Language Pre-training (VLP) has improved performance on various joint vision-and-language downstream tasks. Current approaches to VLP heavily rely on image feature extraction processes, most of which involve region supervision (e.g., object detection) and the convolutional architecture (e.g., ResNet). Although disregarded in the literature, we find it problematic in terms of both (1) efficiency/speed, that simply extracting input features requires much more computation than the multimodal interaction steps; and (2) expressive power, as it is upper bounded to the expressive power of the visual embedder and its predefined visual vocabulary. In this paper, we present a minimal VLP model, Vision-and-Language Transformer (ViLT), monolithic in the sense that the processing of visual inputs is drastically simplified to just the same convolution-free manner that we process textual inputs. We show that ViLT is up to tens of times faster than previous VLP models, yet with competitive or better downstream task performance. Our code and pre-trained weights are available at https://github.com/dandelin/vilt.","{'model': 'tldr@v2.0.0', 'text': 'A minimal VLP model, Vision-and-Language Transformer (ViLT), monolithic in the sense that the processing of visual inputs is drastically simplified to just the same convolution-free manner that the authors process textual inputs, showing that ViLT is up to tens of times faster than previous VLP models, yet with competitive or better downstream task performance.'}",
-improved low-frequency transformer model based on jiles–atherton hysteresis theory,Mi Zou,"Transformers are the most widely used apparatus in power systems to transfer electrical energy from one side to another. Despite remarkable achievements in transformer modelling over the past few years, three-phase transformer modelling techniques have not progressed to that of their single-phase counterparts, mainly because of topological correctness of core and coil structure, complicated interaction between fluxes of the three iron core legs, and non-linear saturation of ferromagnetic material. An improved three-phase transformer model with extended inverse Jiles-Atherton (JA) hysteresis theory is proposed in this study. The extended inverse JA model with anisotropic anhysteretic magnetisation function is addressed, and the classical eddy-current-loss term and anomalous-loss term are directly incorporated into the original energy balance equation to represent the non-linearity and iron core losses rather than representing them with a parallel combination of resistances and saturable inductances (R//L model) as most previous studies have done. The results of inrush current case indicate that the proposed transformer model can predict the residual flux, include eddy-current-loss and anomalous-loss, and predict the inrush current peak values with relative error less than 6%.",,
-recurrent memory transformer,Aydar Bulatov,"Transformer-based models show their effectiveness across multiple domains and tasks. The self-attention allows to combine information from all sequence elements into context-aware representations. However, global and local information has to be stored mostly in the same element-wise representations. Moreover, the length of an input sequence is limited by quadratic computational complexity of self-attention. In this work, we propose and study a memory-augmented segment-level recurrent Transformer (RMT). Memory allows to store and process local and global information as well as to pass information between segments of the long sequence with the help of recurrence. We implement a memory mechanism with no changes to Transformer model by adding special memory tokens to the input or output sequence. Then the model is trained to control both memory operations and sequence representations processing. Results of experiments show that RMT performs on par with the Transformer-XL on language modeling for smaller memory sizes and outperforms it for tasks that require longer sequence processing. We show that adding memory tokens to Tr-XL is able to improve its performance. This makes Recurrent Memory Transformer a promising architecture for applications that require learning of long-term dependencies and general purpose in memory processing, such as algorithmic tasks and reasoning.","{'model': 'tldr@v2.0.0', 'text': 'Recurrent Memory Transformer is a promising architecture for applications that require learning of long-term dependencies and general purpose in memory processing, such as algorithmic tasks and reasoning.'}",https://arxiv.org/pdf/2207.06881
-exploring visual prompts for adapting large-scale models,Hyojin Bahng,"We investigate the efficacy of visual prompting to adapt large-scale models in vision. Following the recent approach from prompt tuning and adversarial reprogramming, we learn a single image perturbation such that a frozen model prompted with this perturbation performs a new task. Through comprehensive experiments, we demonstrate that visual prompting is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes. We further analyze properties of the downstream dataset, prompt design, and output transformation in regard to adaptation performance. The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision. Code is available at http://hjbahng.github.io/visual_prompting .","{'model': 'tldr@v2.0.0', 'text': 'The surprising effectiveness of visual prompting provides a new perspective on adapting pre-trained models in vision and is particularly effective for CLIP and robust to distribution shift, achieving performance competitive with standard linear probes.'}",
-video swin transformer,Ze Liu,"The vision community is witnessing a modeling shift from CNNs to Transformers, where pure Transformer architectures have attained top accuracy on the major video recognition benchmarks. These video models are all built on Transformer layers that globally connect patches across the spatial and temporal dimensions. In this paper, we instead advocate an inductive bias of locality in video Transformers, which leads to a better speed-accuracy trade-off compared to previous approaches which compute self-attention globally even with spatial-temporal factorization. The locality of the proposed video architecture is realized by adapting the Swin Transformer designed for the image domain, while continuing to leverage the power of pre-trained image models. Our approach achieves state-of-the-art accuracy on a broad range of video recognition benchmarks, including on action recognition (84.9 top-l accuracy on Kinetics-400 and 85.9 top-l accuracy on Kinetics-600 with ~20× less pre-training data and ~3× smaller model size) and temporal modeling (69.6 top-l accuracy on Something-Something v2).","{'model': 'tldr@v2.0.0', 'text': 'This paper advocates an inductive bias of locality in video Transformers, which leads to a better speed-accuracy trade-off compared to previous approaches which compute self-attention globally even with spatial-temporal factorization.'}",
-efficiently scaling transformer inference,Reiner Pope,"We study the problem of efficient generative inference for Transformer models, in one of its most challenging settings: large deep models, with tight latency targets and long sequence lengths. Better understanding of the engineering tradeoffs for inference for large Transformer-based models is important as use cases of these models are growing rapidly throughout application areas. We develop a simple analytical model for inference efficiency to select the best multi-dimensional partitioning techniques optimized for TPU v4 slices based on the application requirements. We combine these with a suite of low-level optimizations to achieve a new Pareto frontier on the latency and model FLOPS utilization (MFU) tradeoffs on 500B+ parameter models that outperforms the FasterTransformer suite of benchmarks. We further show that with appropriate partitioning, the lower memory requirements of multiquery attention (i.e. multiple query heads share single key/value head) enables scaling up to 32x larger context lengths. Finally, we achieve a low-batch-size latency of 29ms per token during generation (using int8 weight quantization) and a 76% MFU during large-batch-size processing of input tokens, while supporting a long 2048-token context length on the PaLM 540B parameter model.","{'model': 'tldr@v2.0.0', 'text': 'A simple analytical model for inference efficiency is developed to select the best multi-dimensional partitioning techniques optimized for TPU v4 slices based on the application requirements and a suite of low-level optimizations are combined to achieve a new Pareto frontier on the latency and model FLOPS utilization tradeoffs on 500B+ parameter models that outperforms the FasterTransformer suite of benchmarks.'}",https://arxiv.org/pdf/2211.05102
-molecular transformer: a model for uncertainty-calibrated chemical reaction prediction,P. Schwaller,"Organic synthesis is one of the key stumbling blocks in medicinal chemistry. A necessary yet unsolved step in planning synthesis is solving the forward problem: Given reactants and reagents, predict the products. Similar to other work, we treat reaction prediction as a machine translation problem between simplified molecular-input line-entry system (SMILES) strings (a text-based representation) of reactants, reagents, and the products. We show that a multihead attention Molecular Transformer model outperforms all algorithms in the literature, achieving a top-1 accuracy above 90% on a common benchmark data set. Molecular Transformer makes predictions by inferring the correlations between the presence and absence of chemical motifs in the reactant, reagent, and product present in the data set. Our model requires no handcrafted rules and accurately predicts subtle chemical transformations. Crucially, our model can accurately estimate its own uncertainty, with an uncertainty score that is 89% accurate in terms of classifying whether a prediction is correct. Furthermore, we show that the model is able to handle inputs without a reactant–reagent split and including stereochemistry, which makes our method universally applicable.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that a multihead attention Molecular Transformer model outperforms all algorithms in the literature, achieving a top-1 accuracy above 90% on a common benchmark data set and is able to handle inputs without a reactant–reagent split and including stereochemistry, which makes the method universally applicable.'}",https://pubs.acs.org/doi/pdf/10.1021/acscentsci.9b00576
-visual saliency transformer,Nian Liu,"Existing state-of-the-art saliency detection methods heavily rely on CNN-based architectures. Alternatively, we rethink this task from a convolution-free sequence-to-sequence perspective and predict saliency by modeling long-range dependencies, which can not be achieved by convolution. Specifically, we develop a novel unified model based on a pure transformer, namely, Visual Saliency Transformer (VST), for both RGB and RGB-D salient object detection (SOD). It takes image patches as inputs and leverages the transformer to propagate global contexts among image patches. Unlike conventional architectures used in Vision Transformer (ViT), we leverage multi-level token fusion and propose a new token upsampling method under the transformer framework to get high-resolution detection results. We also develop a token-based multi-task decoder to simultaneously perform saliency and boundary detection by introducing task-related tokens and a novel patch-task-attention mechanism. Experimental results show that our model outperforms existing methods on both RGB and RGB-D SOD benchmark datasets. Most importantly, our whole framework not only provides a new perspective for the SOD field but also shows a new paradigm for transformer-based dense prediction models. Code is available at https://github.com/nnizhang/VST.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a novel unified model based on a pure transformer, namely, Visual Saliency Transformer, for both RGB and RGB-D salient object detection (SOD), which takes image patches as inputs and leverages the transformer to propagate global contexts among image patches.'}",https://arxiv.org/pdf/2104.12099
-bitfit: simple parameter-efficient fine-tuning for transformer-based masked language-models,Elad Ben-Zaken,"We introduce BitFit, a sparse-finetuning method where only the bias-terms of the model (or a subset of them) are being modified. We show that with small-to-medium training data, applying BitFit on pre-trained BERT models is competitive with (and sometimes better than) fine-tuning the entire model. For larger data, the method is competitive with other sparse fine-tuning methods.Besides their practical utility, these findings are relevant for the question of understanding the commonly-used process of finetuning: they support the hypothesis that finetuning is mainly about exposing knowledge induced by language-modeling training, rather than learning new task-specific linguistic knowledge.","{'model': 'tldr@v2.0.0', 'text': 'BitFit is introduced, a sparse-finetuning method where only the bias-terms of the model (or a subset of them) are being modified, which shows that with small-to-medium training data, applying BitFit on pre-trained BERT models is competitive with (and sometimes better than) fine-tuning the entire model.'}",https://aclanthology.org/2022.acl-short.1.pdf
-speech-transformer: a no-recurrence sequence-to-sequence model for speech recognition,Linhao Dong,"Recurrent sequence-to-sequence models using encoder-decoder architecture have made great progress in speech recognition task. However, they suffer from the drawback of slow training speed because the internal recurrence limits the training parallelization. In this paper, we present the Speech-Transformer, a no-recurrence sequence-to-sequence model entirely relies on attention mechanisms to learn the positional dependencies, which can be trained faster with more efficiency. We also propose a 2D-Attention mechanism, which can jointly attend to the time and frequency axes of the 2-dimensional speech inputs, thus providing more expressive representations for the Speech-Transformer. Evaluated on the Wall Street Journal (WSJ) speech recognition dataset, our best model achieves competitive word error rate (WER) of 10.9%, while the whole training process only takes 1.2 days on 1 GPU, significantly faster than the published results of recurrent sequence-to-sequence models.","{'model': 'tldr@v2.0.0', 'text': 'The Speech-Transformer is presented, a no-recurrence sequence-to-sequence model entirely relies on attention mechanisms to learn the positional dependencies, which can be trained faster with more efficiency and a 2D-Attention mechanism which can jointly attend to the time and frequency axes of the 2-dimensional speech inputs, thus providing more expressive representations for the Speech- Transformer.'}",
-unit: multimodal multitask learning with a unified transformer,Ronghang Hu,"We propose UniT, a Unified Transformer model to simultaneously learn the most prominent tasks across different domains, ranging from object detection to natural language understanding and multimodal reasoning. Based on the transformer encoder-decoder architecture, our UniT model encodes each input modality with an encoder and makes predictions on each task with a shared decoder over the encoded input representations, followed by task-specific output heads. The entire model is jointly trained end-to-end with losses from each task. Compared to previous efforts on multi-task learning with transformers, we share the same model parameters across all tasks instead of separately fine-tuning task-specific models and handle a much higher variety of tasks across different domains. In our experiments, we learn 7 tasks jointly over 8 datasets, achieving strong performance on each task with significantly fewer parameters. Our code is available in MMF at https://mmf.sh.","{'model': 'tldr@v2.0.0', 'text': 'UniT, a Unified Transformer model to simultaneously learn the most prominent tasks across different domains, ranging from object detection to natural language understanding and multimodal reasoning, achieves strong performance on each task with significantly fewer parameters.'}",https://arxiv.org/pdf/2102.10772
-mobile-former: bridging mobilenet and transformer,Yinpeng Chen,"We present Mobile-Former, a parallel design of MobileNet and transformer with a two-way bridge in between. This structure leverages the advantages of MobileNet at local processing and transformer at global interaction. And the bridge enables bidirectional fusion of local and global features. Different from recent works on vision transformer, the transformer in Mobile-Former contains very few tokens (e.g. 6 or fewer tokens) that are randomly initialized to learn global priors, resulting in low computational cost. Combining with the proposed light-weight cross attention to model the bridge, Mobile-Former is not only computationally efficient, but also has more representation power. It outperforms MobileNetV3 at low FLOP regime from 25M to 500M FLOPs on ImageNet classification. For instance, Mobile-Former achieves 77.9% top-1 accuracy at 294M FLOPs, gaining 1.3% over MobileNetV3 but saving 17% of computations. When transferring to object detection, Mobile-Former outperforms MobileNetV3 by 8.6 AP in RetinaNet framework. Furthermore, we build an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR by 1.3 AP but saves 52% of computational cost and 36% of parameters. Code will be released at https://github.com/aaboys/mobileformer.","{'model': 'tldr@v2.0.0', 'text': 'The proposed light-weight cross attention to model the bridge enables bidirectional fusion of local and global features and builds an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR but saves 52% of computational cost and 36% of parameters.'}",https://arxiv.org/pdf/2108.05895
-conformer: convolution-augmented transformer for speech recognition,Anmol Gulati,"Recently Transformer and Convolution neural network (CNN) based models have shown promising results in Automatic Speech Recognition (ASR), outperforming Recurrent neural networks (RNNs). Transformer models are good at capturing content-based global interactions, while CNNs exploit local features effectively. In this work, we achieve the best of both worlds by studying how to combine convolution neural networks and transformers to model both local and global dependencies of an audio sequence in a parameter-efficient way. To this regard, we propose the convolution-augmented transformer for speech recognition, named Conformer. Conformer significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies. On the widely used LibriSpeech benchmark, our model achieves WER of 2.1%/4.3% without using a language model and 1.9%/3.9% with an external language model on test/testother. We also observe competitive performance of 2.7%/6.3% with a small model of only 10M parameters.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the convolution-augmented transformer for speech recognition, named Conformer, which significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies.'}",https://arxiv.org/pdf/2005.08100
-towards physics-inspired data-driven weather forecasting: integrating data assimilation with a deep spatial-transformer-based u-net in a case study with era5,A. Chattopadhyay,"Abstract. There is growing interest in data-driven weather prediction (DDWP), e.g., using convolutional neural networks such as U-NET that are trained on data from models or reanalysis. Here, we propose three components, inspired by physics, to integrate with commonly used DDWP models in order to improve their forecast accuracy. These components are (1) a deep spatial transformer added to the latent space of U-NET to capture rotation and scaling transformation in the latent space for spatiotemporal data, (2) a data-assimilation (DA) algorithm to ingest noisy observations and improve the initial conditions for next forecasts, and (3) a multi-time-step algorithm, which combines forecasts from DDWP models with different time steps through DA, improving the accuracy of forecasts at short intervals. To show the benefit and feasibility of each component, we use geopotential height at 500 hPa (Z500) from ERA5 reanalysis and examine the short-term forecast accuracy of specific setups of the DDWP framework. Results show that the spatial-transformer-based U-NET (U-STN) clearly outperforms the U-NET, e.g., improving the forecast skill by 45 %. Using a sigma-point ensemble Kalman (SPEnKF) algorithm for DA and U-STN as the forward model, we show that stable, accurate DA cycles are achieved even with high observation noise. This DDWP+DA framework substantially benefits from large (O(1000)) ensembles that are inexpensively generated with the data-driven forward model in each DA cycle. The multi-time-step DDWP+DA framework also shows promise; for example, it reduces the average error by factors of 2–3. These results show the benefits and feasibility of these three components, which are flexible and can be used in a variety of DDWP setups. Furthermore, while here we focus on weather forecasting, the three components can be readily adopted for other parts of the Earth system, such as ocean and land, for which there is a rapid growth of data and need for forecast and assimilation. ",,https://gmd.copernicus.org/articles/15/2221/2022/gmd-15-2221-2022.pdf
-enhanced transformer model for low- and mid-frequency transients—part i: model development,A. Rezaei-Zare,"Based on an enhanced topological representation of the transformer core, flux air paths, and tank, a duality-based transformer model is proposed for low and mid-frequency transients analysis. In the developed transformer model, the core and tank iron are represented based on a frequency-dependent hysteresis model. The winding capacitances and the frequency dependency of the winding resistance are also incorporated in the developed model. Furthermore, a detailed model is developed for the air gap and tank, and the corresponding average equivalent is presented which is more appropriate for extracting the parameters from the measurement. In addition, the nonuniform air gap and the fringing flux effects are taken into account, and an approach is presented to obtain the corresponding parameters. The developed models also include the equivalents of the tank magnetic shield and the structural parts. Part II of this paper provides validations and further simulation results.",,
-reformer: the efficient transformer,Nikita Kitaev,"Large Transformer models routinely achieve state-of-the-art results on a number of tasks but training these models can be prohibitively costly, especially on long sequences. We introduce two techniques to improve the efficiency of Transformers. For one, we replace dot-product attention by one that uses locality-sensitive hashing, changing its complexity from O($L^2$) to O($L\log L$), where $L$ is the length of the sequence. Furthermore, we use reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of $N$ times, where $N$ is the number of layers. The resulting model, the Reformer, performs on par with Transformer models while being much more memory-efficient and much faster on long sequences.","{'model': 'tldr@v2.0.0', 'text': 'This work replaces dot-product attention by one that uses locality-sensitive hashing and uses reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of several times, making the model much more memory-efficient and much faster on long sequences.'}",
-enhanced transformer model for low- and mid-frequency transients—part ii: validation and simulation results,A. Rezaei-Zare,"The transformer model developed in Part I is validated based on the zero-sequence test data, a ferroresonance event, and the geomagnetically induced current (GIC) measurements. Various transformer core constructions, including single phase, three limb, five limb, conventional shell type, and seven-limb shell type are represented, based on the proposed transformer model. The simulation results show high accuracy of the proposed transformer model in representing the studied cases. The study reveals that the saturation of the transformers at a given GIC level is more severe than that predicted by the existing transformer models. Furthermore, the proposed transformer model clearly explains the reasons for the previously reported discrepancies between the GIC experimental and simulation results. The study also concludes that for an accurate transformer model, particularly for the three-limb core type, the detailed representation of the zero-sequence characteristic is an essential requirement. Such a characteristic can be obtained based on either the proposed approach of the paper or a zero-sequence dc excitation test in addition to the test data at the power frequency.",,
-a transformer-cnn hybrid model for cognitive behavioral therapy in psychological assessment and intervention for enhanced diagnostic accuracy and treatment efficiency,Dr. Veera Ankalu. Vuyyuru,"— The use of Cognitive Behavioral Therapy (CBT) as a method for psychological assessment and intervention has shown to be quite successful. However, by utilizing advancements in artificial intelligence and natural language processing techniques, the diagnostic precision and therapeutic efficacy of CBT can be significantly improved. For CBT in psychological evaluation and intervention, we suggest a unique Transformer-CNN hybrid model in this work. The hybrid model combines the strengths of the Transformer and Convolutional Neural Network (CNN) architectures. While the CNN model successfully extracts local and global features from the input sequences, the Transformer model accurately captures the contextual dependencies and semantic linkages in the text data. It intends to enhance the model's comprehension and interpretation of the complex linguistic patterns involved in psychological evaluation and intervention by merging these two algorithms. On a sizable collection of clinical text data, which includes patient narratives, treatment transcripts, and diagnostic reports, we undertake comprehensive experiments. The proposed Trans-CNN hybrid model outperformed all other methods with an impressive accuracy of 97%. In diagnosing psychiatric problems, the model shows improved diagnosis accuracy and offers more effective therapy advice. Our hybrid model's automatic real-time monitoring and feedback capabilities also make it possible for prompt intervention and customized care during therapy sessions. By giving doctors a formidable tool for precise evaluation and efficient intervention, the suggested approach has the potential to revolutionize the field of CBT and enhance patient outcomes for mental health. In order to improve the diagnostic precision and therapeutic efficacy of CBT in psychological evaluation and intervention, this work provides a transformational strategy that combines the advantages of the Transformer and CNN architectures.",,http://thesai.org/Downloads/Volume14No7/Paper_66-A_Transformer_CNN_Hybrid_Model_for_Cognitive_Behavioral_Therapy.pdf
-improving the transformer translation model with document-level context,Jiacheng Zhang,"Although the Transformer translation model (Vaswani et al., 2017) has achieved state-of-the-art performance in a variety of translation tasks, how to use document-level context to deal with discourse phenomena problematic for Transformer still remains a challenge. In this work, we extend the Transformer model with a new context encoder to represent document-level context, which is then incorporated into the original encoder and decoder. As large-scale document-level parallel corpora are usually not available, we introduce a two-step training method to take full advantage of abundant sentence-level parallel corpora and limited document-level parallel corpora. Experiments on the NIST Chinese-English datasets and the IWSLT French-English datasets show that our approach improves over Transformer significantly.","{'model': 'tldr@v2.0.0', 'text': 'This work extends the Transformer model with a new context encoder to represent document-level context, which is then incorporated into the original encoder and decoder, and introduces a two-step training method to take full advantage of abundant sentence-level parallel corpora and limited document- level parallel Corpora.'}",https://www.aclweb.org/anthology/D18-1049.pdf
-you only look at one sequence: rethinking transformer in vision through object detection,Yuxin Fang,"Can Transformer perform 2D object- and region-level recognition from a pure sequence-to-sequence perspective with minimal knowledge about the 2D spatial structure? To answer this question, we present You Only Look at One Sequence (YOLOS), a series of object detection models based on the vanilla Vision Transformer with the fewest possible modifications, region priors, as well as inductive biases of the target task. We find that YOLOS pre-trained on the mid-sized ImageNet-1k dataset only can already achieve quite competitive performance on the challenging COCO object detection benchmark, e.g., YOLOS-Base directly adopted from BERT-Base architecture can obtain 42.0 box AP on COCO val. We also discuss the impacts as well as limitations of current pre-train schemes and model scaling strategies for Transformer in vision through YOLOS. Code and pre-trained models are available at https://github.com/hustvl/YOLOS.","{'model': 'tldr@v2.0.0', 'text': 'It is found that YOLOS pre-trained on the mid-sized ImageNet-1k dataset only can already achieve quite competitive performance on the challenging COCO object detection benchmark, and the impacts as well as limitations of current pre-train schemes and model scaling strategies for Transformer in vision through YOLos are discussed.'}",
-tdstf: transformer-based diffusion probabilistic model for sparse time series forecasting,Ping Chang,"Background and objective: In the intensive care unit (ICU), vital sign monitoring is critical, and an accurate predictive system is required. This study will create a novel model to forecast Heart Rate (HR), Systolic Blood Pressure (SBP), and Diastolic Blood Pressure (DBP) in ICU. These vital signs are crucial for prompt interventions for patients. We extracted $24,886$ ICU stays from the MIMIC-III database, which contains data from over $46$ thousand patients, to train and test the model. Methods: The model proposed in this study, areansformerin intensive careabilistic Model for Sparse Time Series Forecasting (TDSTF), uses a deep learning technique called the Transformer. The TDSTF model showed state-of-the-art performance in predicting vital signs in the ICU, outperforming other models' ability to predict distributions of vital signs and being more computationally efficient. The code is available at https://github.com/PingChang818/TDSTF. Results: The results of the study showed that TDSTF achieved a Normalized Average Continuous Ranked Probability Score (NACRPS) of $0.4438$ and a Mean Squared Error (MSE) of $0.4168$, an improvement of $18.9\%$ and $34.3\%$ over the best baseline model, respectively. Conclusion: In conclusion, TDSTF is an effective and efficient solution for forecasting vital signs in the ICU, and it shows a significant improvement compared to other models in the field.","{'model': 'tldr@v2.0.0', 'text': ""The TDSTF model showed state-of-the-art performance in predicting vital signs in the ICU, outperforming other models' ability to predict distributions of vital signs and being more computationally efficient.""}",
-analyzing the structure of attention in a transformer language model,Jesse Vig,"The Transformer is a fully attention-based alternative to recurrent networks that has achieved state-of-the-art results across a range of NLP tasks. In this paper, we analyze the structure of attention in a Transformer language model, the GPT-2 small pretrained model. We visualize attention for individual instances and analyze the interaction between attention and syntax over a large corpus. We find that attention targets different parts of speech at different layer depths within the model, and that attention aligns with dependency relations most strongly in the middle layers. We also find that the deepest layers of the model capture the most distant relationships. Finally, we extract exemplar sentences that reveal highly specific patterns targeted by particular attention heads.","{'model': 'tldr@v2.0.0', 'text': 'It is found that attention targets different parts of speech at different layer depths within the model, and that attention aligns with dependency relations most strongly in the middle layers, and the deepest layers of the model capture the most distant relationships.'}",https://www.aclweb.org/anthology/W19-4808.pdf
-gpt4mia: utilizing geneative pre-trained transformer (gpt-3) as a plug-and-play transductive model for medical image analysis,Yizhe Zhang,". In this paper, we propose a novel approach (called GPT4MIA) that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA). We provide theoretical analysis on why a large pre-trained language model such as GPT-3 can be used as a plug-and-play transductive inference model for MIA. At the methodological level, we develop several technical treatments to improve the efficiency and effectiveness of GPT4MIA, including better prompt structure design, sample selection, and prompt ordering of representative samples/features. We present two concrete use cases (with workflow) of GPT4MIA: (1) detecting prediction errors and (2) improving prediction accuracy, working in conjecture with well-established vision-based models for image classification (e.g., ResNet). Experiments validate that our proposed method is effective for these two tasks. We further discuss the opportunities and challenges in utilizing Transformer-based large language models for broader MIA applications.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA) and develops several technical treatments to improve theency andectiveness of GPT4MIA.'}",http://arxiv.org/pdf/2302.08722
-mt5: a massively multilingual pre-trained text-to-text transformer,Linting Xue,"The recent “Text-to-Text Transfer Transformer” (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent “accidental translation” in the zero-shot setting, where a generative model chooses to (partially) translate its prediction into the wrong language. All of the code and model checkpoints used in this work are publicly available.","{'model': 'tldr@v2.0.0', 'text': None}",https://aclanthology.org/2021.naacl-main.41.pdf
-pretraining data mixtures enable narrow model selection capabilities in transformer models,Steve Yadlowsky,"Transformer models, notably large language models (LLMs), have the remarkable ability to perform in-context learning (ICL) -- to perform new tasks when prompted with unseen input-output examples without any explicit model training. In this work, we study how effectively transformers can bridge between their pretraining data mixture, comprised of multiple distinct task families, to identify and learn new tasks in-context which are both inside and outside the pretraining distribution. Building on previous work, we investigate this question in a controlled setting, where we study transformer models trained on sequences of $(x, f(x))$ pairs rather than natural language. Our empirical results show transformers demonstrate near-optimal unsupervised model selection capabilities, in their ability to first in-context identify different task families and in-context learn within them when the task families are well-represented in their pretraining data. However when presented with tasks or functions which are out-of-domain of their pretraining data, we demonstrate various failure modes of transformers and degradation of their generalization for even simple extrapolation tasks. Together our results highlight that the impressive ICL abilities of high-capacity sequence models may be more closely tied to the coverage of their pretraining data mixtures than inductive biases that create fundamental generalization capabilities.","{'model': 'tldr@v2.0.0', 'text': 'The empirical results show transformers demonstrate near-optimal unsupervised model selection capabilities, in their ability to first in- context identify different task families and in-context learn within them when the task families are well-represented in their pretraining data.'}",
-gpt4mia: utilizing generative pre-trained transformer (gpt-3) as a plug-and-play transductive model for medical image analysis,Yizhe Zhang,"In this paper, we propose a novel approach (called GPT4MIA) that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA). We provide theoretical analysis on why a large pre-trained language model such as GPT-3 can be used as a plug-and-play transductive inference model for MIA. At the methodological level, we develop several technical treatments to improve the efficiency and effectiveness of GPT4MIA, including better prompt structure design, sample selection, and prompt ordering of representative samples/features. We present two concrete use cases (with workflow) of GPT4MIA: (1) detecting prediction errors and (2) improving prediction accuracy, working in conjecture with well-established vision-based models for image classification (e.g., ResNet). Experiments validate that our proposed method is effective for these two tasks. We further discuss the opportunities and challenges in utilizing Transformer-based large language models for broader MIA applications.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach that utilizes Generative Pre-trained Transformer (GPT) as a plug-and-play transductive inference tool for medical image analysis (MIA), and develops several technical treatments to improve the efficiency and effectiveness of GPT4MIA.'}",
-domain prompts: towards memory and compute efficient domain adaptation of asr systems,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains. Since domain-specific systems perform better than their generic counterparts on in-domain evaluation, the need for memory and compute-efficient domain adaptation is obvious. Particularly, adapting parameter-heavy transformer-based language models used for rescoring ASR hypothesis is challenging. In this work, we introduce domain-prompts , a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain. With just a handful of extra parameters per domain, we achieve 7-14% WER improvement over the baseline of using an unadapted LM. Despite being parameter-efficient, these improvements are comparable to those of fully-fine-tuned models with hundreds of millions of parameters. With ablations on prompt-sizes, dataset sizes, initializations and domains, we provide evidence for the ben-efits of using domain-prompts in ASR systems.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces domain-prompts, a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain, achieving 7-14% WER improvement over the baseline of using an unadapted LM.'}",
-msa transformer,Roshan Rao,"Unsupervised protein language models trained across millions of diverse sequences learn structure and function of proteins. Protein language models studied to date have been trained to perform inference from individual sequences. The longstanding approach in computational biology has been to make inferences from a family of evo lutionarily related sequences by fitting a model to each family independently. In this work we combine the two paradigms. We introduce a protein language model which takes as input a set of sequences in the form of a multiple sequence alignment. The model interleaves row and column attention across the input sequences and is trained with a variant of the masked language modeling objective across many protein families. The performance of the model surpasses current state-of-the-art unsupervised structure learning methods by a wide margin, with far greater parameter efficiency than prior state-of-the-art protein language models.","{'model': 'tldr@v2.0.0', 'text': 'A protein language model which takes as input a set of sequences in the form of a multiple sequence alignment and is trained with a variant of the masked language modeling objective across many protein families surpasses current state-of-the-art unsupervised structure learning methods by a wide margin.'}",https://www.biorxiv.org/content/biorxiv/early/2021/08/27/2021.02.12.430858.full.pdf
-optimizing transformer-based machine translation model for single gpu training: a hyperparameter ablation study,Luv Verma,"In machine translation tasks, the relationship between model complexity and performance is often presumed to be linear, driving an increase in the number of parameters and consequent demands for computational resources like multiple GPUs. To explore this assumption, this study systematically investigates the effects of hyperparameters through ablation on a sequence-to-sequence machine translation pipeline, utilizing a single NVIDIA A100 GPU. Contrary to expectations, our experiments reveal that combinations with the most parameters were not necessarily the most effective. This unexpected insight prompted a careful reduction in parameter sizes, uncovering""sweet spots""that enable training sophisticated models on a single GPU without compromising translation quality. The findings demonstrate an intricate relationship between hyperparameter selection, model size, and computational resource needs. The insights from this study contribute to the ongoing efforts to make machine translation more accessible and cost-effective, emphasizing the importance of precise hyperparameter tuning over mere scaling.",,https://arxiv.org/pdf/2308.06017
-longt5: efficient text-to-text transformer for long sequences,Mandy Guo,"Recent work has shown that either (1) increasing the input length or (2) increasing model size can improve the performance of Transformer-based neural models. In this paper, we present a new model, called LongT5, with which we explore the effects of scaling both the input length and model size at the same time. Specifically, we integrated attention ideas from long-input transformers (ETC), and adopted pre-training strategies from summarization pre-training (PEGASUS) into the scalable T5 architecture. The result is a new attention mechanism we call {\em Transient Global} (TGlobal), which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs. We are able to achieve state-of-the-art results on several summarization tasks and outperform the original T5 models on question answering tasks.","{'model': 'tldr@v2.0.0', 'text': ""A new model, called LongT5, is presented, with which the effects of scaling both the input length and model size at the same time are explored, which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs.""}",https://aclanthology.org/2022.findings-naacl.55.pdf
-planning with learned entity prompts for abstractive summarization,Shashi Narayan,"Abstract We introduce a simple but flexible mechanism to learn an intermediate plan to ground the generation of abstractive summaries. Specifically, we prepend (or prompt) target summaries with entity chains—ordered sequences of entities mentioned in the summary. Transformer-based sequence-to-sequence models are then trained to generate the entity chain and then continue generating the summary conditioned on the entity chain and the input. We experimented with both pretraining and finetuning with this content planning objective. When evaluated on CNN/DailyMail, XSum, SAMSum, and BillSum, we demonstrate empirically that the grounded generation with the planning objective improves entity specificity and planning in summaries for all datasets, and achieves state-of-the-art performance on XSum and SAMSum in terms of rouge. Moreover, we demonstrate empirically that planning with entity chains provides a mechanism to control hallucinations in abstractive summaries. By prompting the decoder with a modified content plan that drops hallucinated entities, we outperform state-of-the-art approaches for faithfulness when evaluated automatically and by humans.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated empirically that the grounded generation with the planning objective improves entity specificity and planning in summaries for all datasets, and achieves state-of-the-art performance on XSum and SAMSum in terms of rouge.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00438/1979348/tacl_a_00438.pdf
-swin deformable attention u-net transformer (sdaut) for explainable fast mri,Jiahao Huang,". Fast MRI aims to reconstruct a high fidelity image from partially observed measurements. Exuberant development in fast MRI using deep learning has been witnessed recently. Meanwhile, novel deep learning paradigms, e.g., Transformer based models, are fast-growing in natural language processing and promptly developed for computer vision and medical image analysis due to their prominent performance. Nevertheless, due to the complexity of the Transformer, the application of fast MRI may not be straightforward. The main obstacle is the computational cost of the self-attention layer, which is the core part of the Transformer, can be expensive for high resolution MRI inputs. In this study, we propose a new Transformer architecture for solving fast MRI that coupled Shifted Windows Transformer with U-Net to reduce the network complexity. We incorporate deformable attention to construe the explainability of our reconstruction model. We empirically demonstrate that our method achieves consistently superior performance on the fast MRI task. Besides, compared to state-of-the-art Transformer models, our method has fewer network parameters while revealing explainability. The code is publicly available at https://github.com/ayanglab/SDAUT.","{'model': 'tldr@v2.0.0', 'text': 'A new Transformer architecture for solving fast MRI that coupled Shifted Windows Transformer with U-Net to reduce the network complexity and incorporate deformable attention to construe the explainability of the reconstruction model.'}",http://arxiv.org/pdf/2207.02390
-visformer: the vision-friendly transformer,Zhengsu Chen,"The past year has witnessed the rapid development of applying the Transformer module to vision problems. While some researchers have demonstrated that Transformer-based models enjoy a favorable ability of fitting data, there are still growing number of evidences showing that these models suffer over-fitting especially when the training data is limited. This paper offers an empirical study by performing step-by-step operations to gradually transit a Transformer-based model to a convolution-based model. The results we obtain during the transition process deliver useful messages for improving visual recognition. Based on these observations, we propose a new architecture named Visformer, which is abbreviated from the ‘Vision-friendly Transformer’. With the same computational complexity, Visformer outperforms both the Transformer-based and convolution-based models in terms of ImageNet classification accuracy, and the advantage becomes more significant when the model complexity is lower or the training set is smaller. The code is available at https://github.com/danczs/Visformer.","{'model': 'tldr@v2.0.0', 'text': 'A new architecture named Visformer is proposed, which is abbreviated from the ‘Vision-friendly Transformer’, which outperforms both the Transformer-based and convolution-based models in terms of ImageNet classification accuracy, and the advantage becomes more significant when the model complexity is lower or the training set is smaller.'}",https://arxiv.org/pdf/2104.12533
-trocr: transformer-based optical character recognition with pre-trained models,Minghao Li,"Text recognition is a long-standing research problem for document digitalization. Existing approaches are usually built based on CNN for image understanding and RNN for char-level text generation. In addition, another language model is usually needed to improve the overall accuracy as a post-processing step. In this paper, we propose an end-to-end text recognition approach with pre-trained image Transformer and text Transformer models, namely TrOCR, which leverages the Transformer architecture for both image understanding and wordpiece-level text generation. The TrOCR model is simple but effective, and can be pre-trained with large-scale synthetic data and fine-tuned with human-labeled datasets. Experiments show that the TrOCR model outperforms the current state-of-the-art models on the printed, handwritten and scene text recognition tasks. The TrOCR models and code are publicly available at https://aka.ms/trocr.","{'model': 'tldr@v2.0.0', 'text': 'The TrOCR model is simple but effective, and can be pre-trained with large-scale synthetic data and fine-tuned with human-labeled datasets, and outperforms the current state-of-the-art models on the printed, handwritten and scene text recognition tasks.'}",https://ojs.aaai.org/index.php/AAAI/article/download/26538/26310
-vision-language transformer and query generation for referring segmentation,Henghui Ding,"In this work, we address the challenging task of referring segmentation. The query expression in referring segmentation typically indicates the target object by describing its relationship with others. Therefore, to find the target one among all instances in the image, the model must have a holistic understanding of the whole image. To achieve this, we reformulate referring segmentation as a direct attention problem: finding the region in the image where the query language expression is most attended to. We introduce transformer and multi-head attention to build a network with an encoder-decoder attention mechanism architecture that ""queries"" the given image with the language expression. Furthermore, we propose a Query Generation Module, which produces multiple sets of queries with different attention weights that represent the diversified comprehensions of the language expression from different aspects. At the same time, to find the best way from these diversified comprehensions based on visual clues, we further propose a Query Balance Module to adaptively select the output features of these queries for a better mask generation. Without bells and whistles, our approach is light-weight and achieves new state-of-the-art performance consistently on three referring segmentation datasets, RefCOCO, RefCOCO+, and G-Ref. Our code is available at https://github.com/henghuiding/Vision-Language-Transformer.","{'model': 'tldr@v2.0.0', 'text': 'Transformer and multi-head attention are introduced and a Query Generation Module is proposed, which produces multiple sets of queries with different attention weights that represent the diversified comprehensions of the language expression from different aspects.'}",https://arxiv.org/pdf/2108.05565
-asformer: transformer for action segmentation,Fangqiu Yi,"Algorithms for the action segmentation task typically use temporal models to predict what action is occurring at each frame for a minute-long daily activity. Recent studies have shown the potential of Transformer in modeling the relations among elements in sequential data. However, there are several major concerns when directly applying the Transformer to the action segmentation task, such as the lack of inductive biases with small training sets, the deficit in processing long input sequence, and the limitation of the decoder architecture to utilize temporal relations among multiple action segments to refine the initial predictions. To address these concerns, we design an efficient Transformer-based model for action segmentation task, named ASFormer, with three distinctive characteristics: (i) We explicitly bring in the local connectivity inductive priors because of the high locality of features. It constrains the hypothesis space within a reliable scope, and is beneficial for the action segmentation task to learn a proper target function with small training sets. (ii) We apply a pre-defined hierarchical representation pattern that efficiently handles long input sequences. (iii) We carefully design the decoder to refine the initial predictions from the encoder. Extensive experiments on three public datasets demonstrate that effectiveness of our methods. Code is available at \url{https://github.com/ChinaYi/ASFormer}.","{'model': 'tldr@v2.0.0', 'text': 'An efficient Transformer-based model for action segmentation task, named ASFormer, with three distinctive characteristics, which constrains the hypothesis space within a reliable scope, and is beneficial for the action segmentations task to learn a proper target function with small training sets.'}",
-pre-trained image processing transformer,Hanting Chen,"As the computing power of modern hardware is increasing strongly, pre-trained deep learning models (e.g., BERT, GPT-3) learned on large-scale datasets have shown their effectiveness over conventional methods. The big progress is mainly contributed to the representation ability of transformer and its variant architectures. In this paper, we study the low-level computer vision task (e.g., denoising, super-resolution and deraining) and develop a new pre-trained model, namely, image processing transformer (IPT). To maximally excavate the capability of transformer, we present to utilize the well-known ImageNet benchmark for generating a large amount of corrupted image pairs. The IPT model is trained on these images with multi-heads and multi-tails. In addition, the contrastive learning is introduced for well adapting to different image processing tasks. The pre-trained model can therefore efficiently employed on desired task after fine-tuning. With only one pre-trained model, IPT outperforms the current state-of-the-art methods on various low-level benchmarks. Code is available at https://github.com/huawei-noah/Pretrained-IPT and https://gitee.com/mindspore/mindspore/tree/master/model_zoo/research/cv/IPT","{'model': 'tldr@v2.0.0', 'text': 'To maximally excavate the capability of transformer, the IPT model is presented to utilize the well-known ImageNet benchmark for generating a large amount of corrupted image pairs and the contrastive learning is introduced for well adapting to different image processing tasks.'}",https://arxiv.org/pdf/2012.00364
-personalized transformer for explainable recommendation,Lei Li,"Personalization of natural language generation plays a vital role in a large spectrum of tasks, such as explainable recommendation, review summarization and dialog systems. In these tasks, user and item IDs are important identifiers for personalization. Transformer, which is demonstrated with strong language modeling capability, however, is not personalized and fails to make use of the user and item IDs since the ID tokens are not even in the same semantic space as the words. To address this problem, we present a PErsonalized Transformer for Explainable Recommendation (PETER), on which we design a simple and effective learning objective that utilizes the IDs to predict the words in the target explanation, so as to endow the IDs with linguistic meanings and to achieve personalized Transformer. Besides generating explanations, PETER can also make recommendations, which makes it a unified model for the whole recommendation-explanation pipeline. Extensive experiments show that our small unpretrained model outperforms fine-tuned BERT on the generation task, in terms of both effectiveness and efficiency, which highlights the importance and the nice utility of our design.","{'model': 'tldr@v2.0.0', 'text': 'A PErsonalized Transformer for Explainable Recommendation (PETER), on which a simple and effective learning objective is designed that utilizes the IDs to predict the words in the target explanation, so as to endow the IDs with linguistic meanings and to achieve personalized Transformer.'}",https://aclanthology.org/2021.acl-long.383.pdf
-swin-unet: unet-like pure transformer for medical image segmentation,Hu Cao,,"{'model': 'tldr@v2.0.0', 'text': 'Under the direct down-sampling and up-sampled of the inputs and outputs by 4x, experiments demonstrate that the pure Transformer-based U-shaped Encoder-Decoder network outperforms those methods with full Convolution or the combination of transformer and convolution.'}",https://arxiv.org/pdf/2105.05537
-transformer feed-forward layers build predictions by promoting concepts in the vocabulary space,Mor Geva,"Transformer-based language models (LMs) are at the core of modern NLP, but their internal prediction construction process is opaque and largely not understood. In this work, we make a substantial step towards unveiling this underlying prediction process, by reverse-engineering the operation of the feed-forward network (FFN) layers, one of the building blocks of transformer models. We view the token representation as a changing distribution over the vocabulary, and the output from each FFN layer as an additive update to that distribution. Then, we analyze the FFN updates in the vocabulary space, showing that each update can be decomposed to sub-updates corresponding to single FFN parameter vectors, each promoting concepts that are often human-interpretable. We then leverage these findings for controlling LM predictions, where we reduce the toxicity of GPT2 by almost 50%, and for improving computation efficiency with a simple early exit rule, saving 20% of computation on average.","{'model': 'tldr@v2.0.0', 'text': 'This work reverse-engineering the operation of the feed-forward network layers, one of the building blocks of transformer models, shows that each update can be decomposed to sub-updates corresponding to single FFN parameter vectors, each promoting concepts that are often human-interpretable.'}",http://arxiv.org/pdf/2203.14680
-a survey of controllable text generation using transformer-based pre-trained language models,Hanqing Zhang,"Controllable Text Generation (CTG) is an emerging area in the field of natural language generation (NLG). It is regarded as crucial for the development of advanced text generation technologies that better meet the specific constraints in practical applications. In recent years, methods using large-scale pre-trained language models (PLMs), in particular the widely used Transformer-based PLMs, have become a new paradigm of NLG, allowing generation of more diverse and fluent text. However, due to the limited level of interpretability of deep neural networks, the controllability of these methods needs to be guaranteed. To this end, controllable text generation using Transformer-based PLMs has become a rapidly growing yet challenging new research hotspot. A diverse range of approaches have emerged in the past 3 to 4 years, targeting different CTG tasks that require different types of controlled constraints. In this article, we present a systematic critical review on the common tasks, main approaches, and evaluation methods in this area. Finally, we discuss the challenges that the field is facing, and put forward various promising future directions. To the best of our knowledge, this is the first survey article to summarize the state-of-the-art CTG techniques from the perspective of Transformer-based PLMs. We hope it can help researchers and practitioners in the related fields to quickly track the academic and technological frontier, providing them with a landscape of the area and a roadmap for future research.","{'model': 'tldr@v2.0.0', 'text': 'This is the first survey article to summarize the state-of-the-art CTG techniques from the perspective of Transformer-based PLMs, and it is hoped it can help researchers and practitioners in the related fields to quickly track the academic and technological frontier.'}",https://arxiv.org/pdf/2201.05337
-mate: multi-view attention for table transformer efficiency,Julian Martin Eisenschlos,"This work presents a sparse-attention Transformer architecture for modeling documents that contain large tables. Tables are ubiquitous on the web, and are rich in information. However, more than 20% of relational tables on the web have 20 or more rows (Cafarella et al., 2008), and these large tables present a challenge for current Transformer models, which are typically limited to 512 tokens. Here we propose MATE, a novel Transformer architecture designed to model the structure of web tables. MATE uses sparse attention in a way that allows heads to efficiently attend to either rows or columns in a table. This architecture scales linearly with respect to speed and memory, and can handle documents containing more than 8000 tokens with current accelerators. MATE also has a more appropriate inductive bias for tabular data, and sets a new state-of-the-art for three table reasoning datasets. For HybridQA (Chen et al., 2020), a dataset that involves large documents containing tables, we improve the best prior result by 19 points.","{'model': 'tldr@v2.0.0', 'text': 'MATE is proposed, a novel Transformer architecture designed to model the structure of web tables that uses sparse attention in a way that allows heads to efficiently attend to either rows or columns in a table.'}",https://aclanthology.org/2021.emnlp-main.600.pdf
-promptcal: contrastive affinity learning via auxiliary prompts for generalized novel category discovery,Shengxiang Zhang,"Although existing semi-supervised learning models achieve remarkable success in learning with unannotated in-distribution data, they mostly fail to learn on unlabeled data sampled from novel semantic classes due to their closed-set assumption. In this work, we target a pragmatic but under-explored Generalized Novel Category Discovery (GNCD) setting. The GNCD setting aims to categorize unlabeled training data coming from known and novel classes by leveraging the information of partially labeled known classes. We propose a two-stage Contrastive Affinity Learning method with auxiliary visual Prompts, dubbed PromptCAL, to address this challenging problem. Our approach discovers reliable pairwise sample affinities to learn better semantic clustering of both known and novel classes for the class token and visual prompts. First, we propose a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships. Besides, we propose contrastive affinity learning to calibrate semantic representations based on our iterative semi-supervised affinity graph generation method for semantically-enhanced supervision. Extensive experimental evaluation demonstrates that our PromptCAL method is more effective in discovering novel classes even with limited annotations and surpasses the current state-of-the-art on generic and fine-grained benchmarks (e.g., with nearly 11% gain on CUB-200, and 9% on ImageNet-100) on overall accuracy. Our code is available at https://github.com/sheng-eatamath/PromptCAL.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes contrastive affinity learning to calibrate semantic representations based on the authors' iterative semi-supervised affinity graph generation method for semantically-enhanced supervision and proposes a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships.""}",https://arxiv.org/pdf/2212.05590
-iot based classification of transformer faults using emerging techniques of e-nose and anfis,M. Equbal,"E-Nose finds its use in a wide range of applications such as quality assessment in food processing to toxic gas identification in chemical industry either in the offline or online mode. Their usage can be extended to transformer condition monitoring in the online mode. Considering the importance of transformers in power system and the impact it could create if faults in them are unidentified or left unattended, their functioning should be monitored on a real time basis. This work, describes the realization of a prospective E-Nose for online transformer incipient fault identification. The resistive gas sensor array has been simulated in real time using variable resistances forming one arm of a Wheatstone bridges. Separate variable resistances have been calibrated using characteristics of different fault gas sensors. The sensor array of the E-Nose helps to identify the transformer fault gases resulting from an incipient fault condition at the nascent stage itself and prompts for the necessary corrective action well before a catastrophic situation arises. Furthermore, ANFIS model of the Duval’s Triangle (DT) method have been developed to facilitate the online classification of incipient faults. The ANFIS models of other popularly used incipient fault interpretation methods, reported in earlier works, have also been used for a comparative analysis on their diagnostic capabilities. The developed model has been tested using the fault cases of IEC-TC10 fault database and the results thus obtained have been found to be very promising.",,https://www.frontiersin.org/articles/10.3389/fenrg.2022.1020040/pdf
-improving classification of tetanus severity for patients in low-middle income countries wearing ecg sensors by using a cnn-transformer network,Ping Lu,"Tetanus is a life-threatening infectious disease, which is still common in low- and middle-income countries, including in Vietnam. This disease is characterized by muscle spasm and in severe cases is complicated by autonomic dysfunction. Ideally continuous vital sign monitoring using bedside monitors allows the prompt detection of the onset of autonomic nervous system dysfunction or avoiding rapid deterioration. Detection can be improved using heart rate variability analysis from ECG signals. Recently, characteristic ECG and heart rate variability features have been shown to be of value in classifying tetanus severity. However, conventional manual analysis of ECG is time-consuming. The traditional convolutional neural network (CNN) has limitations in extracting the global context information, due to its fixed-sized kernel filters. In this work, we propose a novel hybrid CNN-Transformer model to automatically classify tetanus severity using tetanus monitoring from low-cost wearable sensors. This model can capture the local features from the CNN and the global features from the Transformer. The time series imaging - spectrogram - is transformed from one-dimensional ECG signal and input to the proposed model. The CNN-Transformer model outperforms state-of-the-art methods in tetanus classification, achieves results with a F1 score of $\mathbf {0.82\pm 0.03}$, precision of $\mathbf {0.94\pm 0.03}$, recall of $\mathbf {0.73\pm 0.07}$, specificity of $\mathbf {0.97\pm 0.02}$, accuracy of $\mathbf {0.88\pm 0.01}$ and AUC of $\mathbf {0.85\pm 0.03}$. In addition, we found that Random Forest with enough manually selected features can be comparable with the proposed CNN-Transformer model.","{'model': 'tldr@v2.0.0', 'text': 'A novel hybrid CNN-Transformer model to automatically classify tetanus severity using tetanus monitoring from low-cost wearable sensors that outperforms state-of-the-art methods in tetanus classification and finds that Random Forest with enough manually selected features can be comparable with the proposed CNN- transformer model.'}",https://ora.ox.ac.uk/objects/uuid:7b2beb4a-8f22-4be4-8e59-9150bff61086/files/s70795889r
-synthesizer: rethinking self-attention for transformer models,Yi Tay,"The dot product self-attention is known to be central and indispensable to state-of-the-art Transformer models. But is it really required? This paper investigates the true importance and contribution of the dot product-based self-attention mechanism on the performance of Transformer models. Via extensive experiments, we find that (1) random alignment matrices surprisingly perform quite competitively and (2) learning attention weights from token-token (query-key) interactions is useful but not that important after all. To this end, we propose \textsc{Synthesizer}, a model that learns synthetic attention weights without token-token interactions. In our experiments, we first show that simple Synthesizers achieve highly competitive performance when compared against vanilla Transformer models across a range of tasks, including machine translation, language modeling, text generation and GLUE/SuperGLUE benchmarks. When composed with dot product attention, we find that Synthesizers consistently outperform Transformers. Moreover, we conduct additional comparisons of Synthesizers against Dynamic Convolutions, showing that simple Random Synthesizer is not only $60\%$ faster but also improves perplexity by a relative $3.5\%$. Finally, we show that simple factorized Synthesizers can outperform Linformers on encoding only tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes \\textsc{Synthesizer}, a model that learns synthetic attention weights without token-token interactions, and shows that simple factorized Synthesizers can outperform Linformers on encoding only tasks.'}",
-transformer in transformer,Kai Han,"Transformer is a new kind of neural architecture which encodes the input data as powerful features via the attention mechanism. Basically, the visual transformers first divide the input images into several local patches and then calculate both representations and their relationship. Since natural images are of high complexity with abundant detail and color information, the granularity of the patch dividing is not fine enough for excavating features of objects in different scales and locations. In this paper, we point out that the attention inside these local patches are also essential for building visual transformers with high performance and we explore a new architecture, namely, Transformer iN Transformer (TNT). Specifically, we regard the local patches (e.g., 16$\times$16) as""visual sentences""and present to further divide them into smaller patches (e.g., 4$\times$4) as""visual words"". The attention of each word will be calculated with other words in the given visual sentence with negligible computational costs. Features of both words and sentences will be aggregated to enhance the representation ability. Experiments on several benchmarks demonstrate the effectiveness of the proposed TNT architecture, e.g., we achieve an 81.5% top-1 accuracy on the ImageNet, which is about 1.7% higher than that of the state-of-the-art visual transformer with similar computational cost. The PyTorch code is available at https://github.com/huawei-noah/CV-Backbones, and the MindSpore code is available at https://gitee.com/mindspore/models/tree/master/research/cv/TNT.","{'model': 'tldr@v2.0.0', 'text': 'It is pointed out that the attention inside these local patches are also essential for building visual transformers with high performance and a new architecture, namely, Transformer iN Transformer (TNT), is explored.'}",
-text style transfer between classical and modern chinese through prompt-based reinforcement learning,Minzhang Xu,,"{'model': 'tldr@v2.0.0', 'text': 'An unsupervised prompt-based reinforcement learning (PBRL) framework to transfer text between classical and modern Chinese styles via an entangled approach and Experimental evaluations show that the model outperforms state-of-art networks by a large margin.'}",
-automated covid-19 classification using heap-based optimization with the deep transfer learning model,Bahjat Fakieh,"The outbreak of the COVID-19 pandemic necessitates prompt identification of affected persons to restrict the spread of the COVID-19 epidemic. Radiological imaging such as computed tomography (CT) and chest X-rays (CXR) is considered an effective way to diagnose COVID-19. However, it needs an expert's knowledge and consumes more time. At the same time, artificial intelligence (AI) and medical images are discovered to be helpful in effectively assessing and providing treatment for COVID-19 infected patients. In particular, deep learning (DL) models act as a vital part of a high-performance classification model for COVID-19 recognition on CXR images. This study develops a heap-based optimization with the deep transfer learning model for detection and classification (HBODTL-DC) of COVID-19. The proposed HBODTL-DC system majorly focuses on the identification of COVID-19 on CXR images. To do so, the presented HBODTL-DC model initially exploits the Gabor filtering (GF) technique to enhance the image quality. In addition, the HBO algorithm with a neural architecture search network (NasNet) large model is employed for the extraction of feature vectors. Finally, Elman Neural Network (ENN) model gets the feature vectors as input and categorizes the CXR images into distinct classes. The experimental validation of the HBODTL-DC model takes place on the benchmark CXR image dataset from the Kaggle repository, and the outcomes are checked in numerous dimensions. The experimental outcomes stated the supremacy of the HBODTL-DC model over recent approaches with a maximum accuracy of 0.9992.","{'model': 'tldr@v2.0.0', 'text': 'A heap-based optimization with the deep transfer learning model for detection and classification (HBODTL-DC) of COVID-19 on CXR images is developed and the experimental outcomes stated the supremacy of the HBOD TL-DC model over recent approaches with a maximum accuracy of 0.9992.'}",https://downloads.hindawi.com/journals/cin/2022/7508836.pdf
-transfer learning-based state of charge estimation for lithium-ion battery at varying ambient temperatures,Yan Qin,"Accurate and reliable state of charge (SoC) estimation becomes increasingly important to provide a stable and efficient environment for Lithium-ion batteries (LiBs) powered devices. Most data-driven SoC models are built for a fixed ambient temperature, which neglect the high sensitivity of LiBs to temperature and may cause severe prediction errors. Nevertheless, a systematic evaluation of the impact of temperature on SoC estimation and ways for a prompt adjustment of the estimation model to new temperatures using limited data has been hardly discussed. To solve these challenges, a novel SoC estimation method is proposed by exploiting temporal dynamics of measurements and transferring consistent estimation ability among different temperatures. First, temporal dynamics, which is presented by correlations between the past fluctuation and the future motion, are extracted using canonical variate analysis. Next, two models, including a reference SoC estimation model and an estimation ability monitoring model, are developed with temporal dynamics. The monitoring model provides a path to quantitatively evaluate the influences of temperature on SoC estimation ability. After that, once the inability of the reference SoC estimation model is detected, consistent temporal dynamics between temperatures are selected for transfer learning. Finally, the efficacy of the proposed method is verified through a benchmark. Our proposed method not only reduces prediction errors at fixed temperatures (e.g., reduced by 24.35$\%$ at –20 $^{\circ }$C, 49.82$\%$ at 25 $^{\circ }$C) but also improves prediction accuracies at new temperatures.","{'model': 'tldr@v2.0.0', 'text': 'A novel SoC estimation method that reduces prediction errors at fixed temperatures and improves prediction accuracies at new temperatures is proposed by exploiting temporal dynamics of measurements and transferring consistent estimation ability among different temperatures.'}",https://arxiv.org/pdf/2101.03704
-improved transfer learning based deep learning model for breast cancer histopathological image classification,Mohd. Farhan Israk Soumik,"In recent years, the demand for prompt detection and classification of breast cancer is rising sharply as breast cancer has become leading cancer type among women throughout the world. Convolutional Neural Networks(CNNs) are widely being used for performing mentioned tasks.However, they need a large number of labeled images which may appear to be infeasible for some kinds of medical images data such as mammographic tumor images. To address this difficulty, Transfer Learning becomes convenient. In this paper, we proposed a deep learning model for classifying Benign and Malignant types of breast tumor that trains an InceptionV3 model which pulls out features from the histopathological images of various magnification. These features are then used for classification. Introduced system is validated on BreakHis dataset and gains average validation set accuracy of 99.50%, 98.90%, 98.96% and 98.51% for magnification factor 40X, 100X, 200X and 400X respectively which outperforms all studied baseline models. Different performance metrices such as precision, recall, F1 score, Specificity have additionally been used for performance estimation purposes.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposed a deep learning model for classifying Benign and Malignant types of breast tumor that trains an InceptionV3 model which pulls out features from the histopathological images of various magnification which outperforms all studied baseline models.'}",
-matchprompt: prompt-based open relation extraction with semantic consistency guided clustering,Jiaxin Wang,"Relation clustering is a general approach for open relation extraction (OpenRE). Current methods have two major problems. One is that their good performance relies on large amounts of labeled and pre-defined relational instances for pre-training, which are costly to acquire in reality. The other is that they only focus on learning a high-dimensional metric space to measure the similarity of novel relations and ignore the specific relational representations of clusters. In this work, we propose a new prompt-based framework named MatchPrompt, which can realize OpenRE with efficient knowledge transfer from only a few pre-defined relational instances as well as mine the specific meanings for cluster interpretability. To our best knowledge, we are the first to introduce a prompt-based framework for unlabeled clustering. Experimental results on different datasets show that MatchPrompt achieves the new SOTA results for OpenRE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a new prompt-based framework named MatchPrompt, which can realize OpenRE with efficient knowledge transfer from only a few pre-defined relational instances as well as mine the specific meanings for cluster interpretability.'}",https://aclanthology.org/2022.emnlp-main.537.pdf
-dynamic visual prompt tuning for parameter efficient transfer learning,Chunqing Ruan,"Parameter efficient transfer learning (PETL) is an emerging research spot that aims to adapt large-scale pre-trained models to downstream tasks. Recent advances have achieved great success in saving storage and computation costs. However, these methods do not take into account instance-specific visual clues for visual tasks. In this paper, we propose a Dynamic Visual Prompt Tuning framework (DVPT), which can generate a dynamic instance-wise token for each image. In this way, it can capture the unique visual feature of each image, which can be more suitable for downstream visual tasks. We designed a Meta-Net module that can generate learnable prompts based on each image, thereby capturing dynamic instance-wise visual features. Extensive experiments on a wide range of downstream recognition tasks show that DVPT achieves superior performance than other PETL methods. More importantly, DVPT even outperforms full fine-tuning on 17 out of 19 downstream tasks while maintaining high parameter efficiency. Our code will be released soon.","{'model': 'tldr@v2.0.0', 'text': 'A Dynamic Visual Prompt Tuning framework (DVPT), which can generate a dynamic instance-wise token for each image that can capture the unique visual feature of each image, which can be more suitable for downstream visual tasks.'}",https://arxiv.org/pdf/2309.06123
-an efficient brain tumour detection from mr images based on deep learning and transfer learning model,Sarita Simaiya,"Brain disease is the most severe, pervasive, and life-threatening illness globally. Globally, brain tumours destroy the lifestyles of numerous individuals annually due to the fast expansion of tumour tissue. To prevent the deaths of individuals worldwide, prompt evaluation and classification of brain diagnosis are therefore necessary. In past years, the three primary forms of brain tumours, gliomas, malignant tumours, and pituitary, have been detected and classified most frequently using deep learning techniques. However, the volume of the sample with annotations significantly impacts how well deep learning algorithms work. It becomes complicated to classify a massive volume of medical data. In this work, we offer a stable hybrid approach for brain tumour identification predicated using CNN (VGG-16) with transfer learning, which incorporates the classic confusion and uncertainty polling technique, a best-fit methodology pooled by best-fit methodology using VGG-16, as well as the transfer learning approach. This approach decreases tagging expenditures while preserving the resilience and consistency of the systems. This research utilises the online brain tumour Kaggle dataset. To compare the proposed dynamic transfer learning strategy with the current CNN technique, evaluation measures such as precision, f-measure, and accuracy are used. The proposed technique enhances the CNN strategy by 5%, achieving 96.77% accuracy, 98.7% recall, 96.35 precision, and 96.78% F-measure.",,
-efficiently aligned cross-lingual transfer learning for conversational tasks using prompt-tuning,Lifu Tu,"Cross-lingual transfer of language models trained on high-resource languages like English has been widely studied for many NLP tasks, but focus on conversational tasks has been rather limited. This is partly due to the high cost of obtaining non-English conversational data, which results in limited coverage. In this work, we introduce XSGD for cross-lingual alignment pretraining, a parallel and large-scale multilingual conversation dataset that we created by translating the English-only Schema-Guided Dialogue (SGD) dataset (Rastogi et al., 2020) into 105 other languages. XSGD contains approximately 330k utterances per language. To facilitate aligned cross-lingual representations, we develop an efficient prompt-tuning-based method for learning alignment prompts. We also investigate two different classifiers: NLI-based and vanilla classifiers, and test cross-lingual capability enabled by the aligned prompts. We evaluate our model's cross-lingual generalization capabilities on two conversation tasks: slot-filling and intent classification. Our results demonstrate the strong and efficient modeling ability of NLI-based classifiers and the large cross-lingual transfer improvements achieved by our aligned prompts, particularly in few-shot settings. In addition, we highlight the nice results of our approach compared to LLMs such as text-davinci-003 and ChatGPT in both zero-shot and few-shot settings. While LLMs exhibit impressive performance in English, their cross-lingual capabilities in other languages, particularly low-resource languages, are limited.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces XSGD for cross-lingual alignment pretraining, a parallel and large-scale multilingual conversation dataset that was created by translating the English-only Schema-Guided Dialogue (SGD) dataset into 105 other languages and develops an efficient prompt-tuning-based method for learning alignment prompts.'}",http://arxiv.org/pdf/2304.01295
-mobilenet-svm: a lightweight deep transfer learning model to diagnose bch scans for iomt-based imaging sensors,R. Ogundokun,"Many individuals worldwide pass away as a result of inadequate procedures for prompt illness identification and subsequent treatment. A valuable life can be saved or at least extended with the early identification of serious illnesses, such as various cancers and other life-threatening conditions. The development of the Internet of Medical Things (IoMT) has made it possible for healthcare technology to offer the general public efficient medical services and make a significant contribution to patients’ recoveries. By using IoMT to diagnose and examine BreakHis v1 400× breast cancer histology (BCH) scans, disorders may be quickly identified and appropriate treatment can be given to a patient. Imaging equipment having the capability of auto-analyzing acquired pictures can be used to achieve this. However, the majority of deep learning (DL)-based image classification approaches are of a large number of parameters and unsuitable for application in IoMT-centered imaging sensors. The goal of this study is to create a lightweight deep transfer learning (DTL) model suited for BCH scan examination and has a good level of accuracy. In this study, a lightweight DTL-based model “MobileNet-SVM”, which is the hybridization of MobileNet and Support Vector Machine (SVM), for auto-classifying BreakHis v1 400× BCH images is presented. When tested against a real dataset of BreakHis v1 400× BCH images, the suggested technique achieved a training accuracy of 100% on the training dataset. It also obtained an accuracy of 91% and an F1-score of 91.35 on the test dataset. Considering how complicated BCH scans are, the findings are encouraging. The MobileNet-SVM model is ideal for IoMT imaging equipment in addition to having a high degree of precision. According to the simulation findings, the suggested model requires a small computation speed and time.","{'model': 'tldr@v2.0.0', 'text': 'The goal of this study is to create a lightweight deep transfer learning (DTL) model suited for BCH scan examination and has a good level of accuracy, and the proposed MobileNet-SVM model is ideal for IoMT imaging equipment in addition to having a high degree of precision.'}",https://www.mdpi.com/1424-8220/23/2/656/pdf?version=1672995017
-adpl: adversarial prompt-based domain adaptation for dialogue summarization with knowledge disentanglement,Lulu Zhao,"Traditional dialogue summarization models rely on a large-scale manually-labeled corpus, lacking generalization ability to new domains, and domain adaptation from a labeled source domain to an unlabeled target domain is important in practical summarization scenarios. However, existing domain adaptation works in dialogue summarization generally require large-scale pre-training using extensive external data. To explore the lightweight fine-tuning methods, in this paper, we propose an efficient Adversarial Disentangled Prompt Learning (ADPL) model for domain adaptation in dialogue summarization. We introduce three kinds of prompts including domain-invariant prompt (DIP), domain-specific prompt (DSP), and task-oriented prompt (TOP). DIP aims to disentangle and transfer the shared knowledge from the source domain and target domain in an adversarial way, which improves the accuracy of prediction about domain-invariant information and enhances the ability for generalization to new domains. DSP is designed to guide our model to focus on domain-specific knowledge using domain-related features. TOP is to capture task-oriented knowledge to generate high-quality summaries. Instead of fine-tuning the whole pre-trained language model (PLM), we only update the prompt networks but keep PLM fixed. Experimental results on the zero-shot setting show that the novel design of prompts can yield more coherent, faithful, and relevant summaries than baselines using the prefix-tuning, and perform at par with fine-tuning while being more efficient. Overall, our work introduces a prompt-based perspective to the zero-shot learning for dialogue summarization task and provides valuable findings and insights for future research.","{'model': 'tldr@v2.0.0', 'text': 'An efficient Adversarial Disentangled Prompt Learning (ADPL) model for domain adaptation in dialogue summarization and three kinds of prompts including domain-invariant prompt, domain-specific prompt, and task-oriented prompt are introduced.'}",
-cataract detection from eye fundus image using an ensemble of transfer learning models,Kashish Chauhan,"Cataract is the clouding of the eye lens and is a major cause of blindness across the globe. Detecting cataracts early and prompt treatment can prevent blindness. To reduce dependence on experts to examine the eye fundus image, computer-assisted technologies are useful for early diagnosis. In the past, many cataract detection models have been described. Some of these models employed typical machine learning techniques, but the performance of these models is inefficient to make them useful. The image classification field has seen much improvement with the help of deep learning. One of the key drawbacks of these cutting-edge models is that they require a significant amount of computational resources and time for training. This problem is solved via transfer learning, which allows pre-trained models to be used for feature extraction. We present an ensemble technique for cataract diagnosis using the eye fundus image based on VGG-19, ResNet101V2, and InceptionV3. Soft voting was used to determine the final classification. The ensemble model gave the F-1 Score of 95.90 on the test dataset. According to our results, this ensemble network is more accurate than any single network.","{'model': 'tldr@v2.0.0', 'text': 'An ensemble technique for cataract diagnosis using the eye fundus image based on VGG-19, ResNet101V2, and InceptionV3 is presented, which is more accurate than any single network.'}",
-parameter efficient transfer learning for suicide attempt and ideation detection,Bhanu Pratap Singh Rawat,"Pre-trained language models (LMs) have been deployed as the state-of-the-art natural language processing (NLP) approaches for multiple clinical applications. Model generalisability is important in clinical domain due to the low available resources. In this study, we evaluated transfer learning techniques for an important clinical application: detecting suicide attempt (SA) and suicide ideation (SI) in electronic health records (EHRs). Using the annotation guideline provided by the authors of ScAN, we annotated two EHR datasets from different hospitals. We then fine-tuned ScANER, a publicly available SA and SI detection model, to evaluate five different parameter efficient transfer learning techniques, such as adapter-based learning and soft-prompt tuning, on the two datasets. Without any fine-tuning, ScANER achieve macro F1-scores of 0.85 and 0.87 for SA and SI evidence detection across the two datasets. We observed that by fine-tuning less than ~2% of ScANER’s parameters, we were able to further improve the macro F1-score for SA-SI evidence detection by 3% and 5% for the two EHR datasets. Our results show that parameter-efficient transfer learning methods can help improve the performance of publicly available clinical models on new hospital datasets with few annotations.","{'model': 'tldr@v2.0.0', 'text': 'Evaluated transfer learning techniques for detecting suicide attempt and suicide ideation in electronic health records (EHRs) show that parameter-efficient transfer learning methods can help improve the performance of publicly available clinical models on new hospital datasets with few annotations.'}",https://aclanthology.org/2022.louhi-1.13.pdf
-efficient diagnosis of covid19 by employing deep transfer learning on pretrained vgg and residualnet architectures,Faraz Bagwan,"Covid19 has had a widespread influence on health services and the way of life. A prompt diagnosis is crucial for curbing the development of the disease and lowering the number of fatalities. It is customary and standard routine to employ blood tests to detect presence of pathogen, but because of the time and expense involved, it is often necessary to turn to other rapid and affordable options. We implemented two distinct transference based deep layered architectures in this study i.e., ResidualNet50 along with VGG16, to classify X-rays as COVID19, pneumonia, or normal. ResidualNet50 trained with transference approach outperformed the other deep-learning model i.e., VGG16, in the planned execution. Our proposed transfer deep-learning based model obtained an overall high classification accuracy of 98.5 percent. Result analysis and interpretation via performance curves have been comprehensively discussed in this paper.","{'model': 'tldr@v2.0.0', 'text': 'Two distinct transference based deep layered architectures are implemented to classify X-rays as COVID19, pneumonia, or normal, and the proposed transfer deep-learning based model obtained an overall high classification accuracy of 98.5 percent.'}",
-distribution inference from early-stage stationary data streams by transfer learning,Kai Wang,"Abstract Data streams are prevalent in current manufacturing and service systems where real-time data arrive progressively. A quick distribution inference from such data streams at their early stages is extremely useful for prompt decision making in many industrial applications. For example, a quality monitoring scheme can be quickly started if the process data distribution is available and the optimal inventory level can be determined early once the customer demand distribution is estimated. To this end, this article proposes a novel online recursive distribution inference method for stationary data streams that can respond as soon as the streaming data are generated and update as regularly as the data accumulate. A major challenge is that the data size might be too small to produce an accurate estimation at the early stage of data streams. To solve this, we resort to an instance-based transfer learning approach which integrates a sufficient amount of auxiliary data from similar processes or products to aid the distribution inference in our target task. Particularly, the auxiliary data are reweighted automatically by a density ratio fitting model with a prior-belief-guided regularization term to alleviate data scarcity. Our proposed distribution inference method also possesses an efficient online algorithm with recursive formulas to update upon every incoming data point. Extensive numerical simulations and real case studies verify the advantages of the proposed method.","{'model': 'tldr@v2.0.0', 'text': ""An instance-based transfer learning approach which integrates a sufficient amount of auxiliary data from similar processes or products to aid the distribution inference in the authors' target task, and an efficient online algorithm with recursive formulas to update upon every incoming data point.""}",https://figshare.com/articles/journal_contribution/Distribution_Inference_from_Early-Stage_Stationary_Data_Streams_by_Transfer_Learning/13607782/2/files/26113477.pdf
-clinical concept and relation extraction using prompt-based machine reading comprehension,C.A.I. Peng,"OBJECTIVE To develop a natural language processing system that solves both clinical concept extraction and relation extraction in a unified prompt-based machine reading comprehension (MRC) architecture with good generalizability for cross-institution applications. METHODS We formulate both clinical concept extraction and relation extraction using a unified prompt-based MRC architecture and explore state-of-the-art transformer models. We compare our MRC models with existing deep learning models for concept extraction and end-to-end relation extraction using 2 benchmark datasets developed by the 2018 National NLP Clinical Challenges (n2c2) challenge (medications and adverse drug events) and the 2022 n2c2 challenge (relations of social determinants of health [SDoH]). We also evaluate the transfer learning ability of the proposed MRC models in a cross-institution setting. We perform error analyses and examine how different prompting strategies affect the performance of MRC models. RESULTS AND CONCLUSION The proposed MRC models achieve state-of-the-art performance for clinical concept and relation extraction on the 2 benchmark datasets, outperforming previous non-MRC transformer models. GatorTron-MRC achieves the best strict and lenient F1-scores for concept extraction, outperforming previous deep learning models on the 2 datasets by 1%-3% and 0.7%-1.3%, respectively. For end-to-end relation extraction, GatorTron-MRC and BERT-MIMIC-MRC achieve the best F1-scores, outperforming previous deep learning models by 0.9%-2.4% and 10%-11%, respectively. For cross-institution evaluation, GatorTron-MRC outperforms traditional GatorTron by 6.4% and 16% for the 2 datasets, respectively. The proposed method is better at handling nested/overlapped concepts, extracting relations, and has good portability for cross-institute applications. Our clinical MRC package is publicly available at https://github.com/uf-hobi-informatics-lab/ClinicalTransformerMRC.","{'model': 'tldr@v2.0.0', 'text': 'The proposed MRC models achieve state-of-the-art performance for clinical concept and relation extraction on the 2 benchmark datasets, outperforming previous non-MRC transformer models and has good portability for cross-institute applications.'}",https://arxiv.org/pdf/2303.08262
-is prompt-based finetuning always better than vanilla finetuning? insights from cross-lingual language understanding,Bolei Ma,"Multilingual pretrained language models (MPLMs) have demonstrated substantial performance improvements in zero-shot cross-lingual transfer across various natural language understanding tasks by finetuning MPLMs on task-specific labelled data of a source language (e.g. English) and evaluating on a wide range of target languages. Recent studies show that prompt-based finetuning surpasses regular finetuning in few-shot scenarios. However, the exploration of prompt-based learning in multilingual tasks remains limited. In this study, we propose the ProFiT pipeline to investigate the cross-lingual capabilities of Prompt-based Finetuning. We conduct comprehensive experiments on diverse cross-lingual language understanding tasks (sentiment classification, paraphrase identification, and natural language inference) and empirically analyze the variation trends of prompt-based finetuning performance in cross-lingual transfer across different few-shot and full-data settings. Our results reveal the effectiveness and versatility of prompt-based finetuning in cross-lingual language understanding. Our findings indicate that prompt-based finetuning outperforms vanilla finetuning in full-data scenarios and exhibits greater advantages in few-shot scenarios, with different performance patterns dependent on task types. Additionally, we analyze underlying factors such as language similarity and pretraining data size that impact the cross-lingual performance of prompt-based finetuning. Overall, our work provides valuable insights into the cross-lingual prowess of prompt-based finetuning.","{'model': 'tldr@v2.0.0', 'text': 'The ProFiT pipeline is proposed to investigate the cross-lingual capabilities of Prompt-based Finetuning and indicates that prompt-based finetuning outperforms vanillaFinetuning in full-data scenarios and exhibits greater advantages in few-shot scenarios, with different performance patterns dependent on task types.'}",https://arxiv.org/pdf/2307.07880
-knowledge transfer with visual prompt in multi-modal dialogue understanding and generation,Minjun Zhu,"Visual Dialogue (VD) task has recently received increasing attention in AI research. Visual Dialog aims to generate multi-round, interactive responses based on the dialog history and image content. Existing textual dialogue models cannot fully understand visual information, resulting in a lack of scene features when communicating with humans continuously. Therefore, how to efficiently fuse multimodal data features remains to be a challenge. In this work, we propose a knowledge transfer method with visual prompt (VPTG) fusing multi-modal data, which is a flexible module that can utilize the text-only seq2seq model to handle visual dialogue tasks. The VPTG conducts text-image co-learning and multi-modal information fusion with visual prompts and visual knowledge distillation. Specifically, we construct visual prompts from visual representations and then induce sequence-to-sequence(seq2seq) models to fuse visual information and textual contexts by visual-text patterns. And we also realize visual knowledge transfer through distillation between two different models’ text representations, so that the seq2seq model can actively learn visual semantic representations. Extensive experiments on the multi-modal dialogue understanding and generation (MDUG) datasets show the proposed VPTG outperforms other single-modal methods, which demonstrate the effectiveness of visual prompt and visual knowledge transfer.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments on the multi-modal dialogue understanding and generation (MDUG) datasets show the proposed VPTG outperforms other single- modal methods, which demonstrate the effectiveness of visual prompt and visual knowledge transfer.'}",
-plato-ad: a unified advertisement text generation framework with multi-task prompt learning,Zeyang Lei,"Online advertisement text generation aims at generating attractive and persuasive text ads to appeal to users clicking ads or purchasing products. While pretraining-based models have achieved remarkable success in generating high-quality text ads, some challenges remain, such as ad generation in low-resource scenarios and training efficiency for multiple ad tasks. In this paper, we propose a novel unified text ad generation framework with multi-task prompt learning, called PLATO-Ad, to tackle these problems. Specifically, we design a threephase transfer learning mechanism to tackle the low-resource ad generation problem. Furthermore, we present a novel multi-task prompt learning mechanism to efficiently utilize a single lightweight model to solve multiple ad generation tasks without loss of performance compared to training a separate model for each task. Finally, we conduct offline and online evaluations. Experiment results show that PLATO-Ad significantly outperforms the state-of-the-art on both offline and online metrics. PLATO-Ad has been deployed in a leading advertising platform with 3.5% CTR improvement on search ad descriptions and 10.4% CTR improvement on feed ad titles.","{'model': 'tldr@v2.0.0', 'text': 'A novel unified text ad generation framework with multi-task prompt learning, called PLATO-Ad, to tackle the low-resource ad generation problem and significantly outperforms the state-of-the-art on both offline and online metrics.'}",https://aclanthology.org/2022.emnlp-industry.52.pdf
-melanoma malignancy prognosis using deep transfer learning,R. Shobarani,"Melanoma is a type of Skin cancer that spreads rapidly and has a significant death risk if it is not detected early and treated. A prompt and accurate diagnosis can improve the patient’s chances of survival. The creation of a skin cancer diagnostic support system based on computer technologies is highly essential. This study suggests a unique deep transfer learning model for categorizing melanoma malignancy. The proposed system comprises of three main phases including image preprocessing, feature extraction and melanoma classification. The preprocessing phase employs image filters such as mean, median, gaussian and non-local means filter along with histogram equalization techniques to obtain the preprocessed images. Feature extraction and classification are performed using pre-trained Convolutional Neural Network architectures such as DenseNet121, Inception-Resnet-V2 and Xception. Using the ISIC 2020 dataset, the suggested deep learning model’s effectiveness is assessed. The experimental findings show that, in terms of precision and computational expense, the suggested deep transfer learning model performs better than cutting-edge deep learning algorithms.","{'model': 'tldr@v2.0.0', 'text': 'The experimental findings show that the suggested deep transfer learning model performs better than cutting-edge deep learning algorithms in terms of precision and computational expense.'}",
-motif-based prompt learning for universal cross-domain recommendation,Bowen Hao,"Cross-Domain Recommendation (CDR) stands as a pivotal technology addressing issues of data sparsity and cold start by transferring general knowledge from the source to the target domain. However, existing CDR models suffer limitations in adaptability across various scenarios due to their inherent complexity. To tackle this challenge, recent advancements introduce universal CDR models that leverage shared embeddings to capture general knowledge across domains and transfer it through""Multi-task Learning""or""Pre-train, Fine-tune""paradigms. However, these models often overlook the broader structural topology that spans domains and fail to align training objectives, potentially leading to negative transfer. To address these issues, we propose a motif-based prompt learning framework, MOP, which introduces motif-based shared embeddings to encapsulate generalized domain knowledge, catering to both intra-domain and inter-domain CDR tasks. Specifically, we devise three typical motifs: butterfly, triangle, and random walk, and encode them through a Motif-based Encoder to obtain motif-based shared embeddings. Moreover, we train MOP under the""Pre-training \&Prompt Tuning""paradigm. By unifying pre-training and recommendation tasks as a common motif-based similarity learning task and integrating adaptable prompt parameters to guide the model in downstream recommendation tasks, MOP excels in transferring domain knowledge effectively. Experimental results on four distinct CDR tasks demonstrate the effectiveness of MOP than the state-of-the-art models.","{'model': 'tldr@v2.0.0', 'text': 'A motif- based prompt learning framework, MOP, which introduces motif-based shared embeddings to encapsulate generalized domain knowledge, catering to both intra-domain and inter-domain CDR tasks, and excels in transferring domain knowledge effectively.'}",
-pro-cs : an instance-based prompt composition technique for code-switched tasks,Srijan Bansal,"Code-switched (CS) data is ubiquitous in today’s globalized world, but the dearth of annotated datasets in code-switching poses a significant challenge for learning diverse tasks across different language pairs. Parameter-efficient prompt-tuning approaches conditioned on frozen language models have shown promise for transfer learning in limited-resource setups. In this paper, we propose a novel instance-based prompt composition technique, PRO-CS, for CS tasks that combine language and task knowledge. We compare our approach with prompt-tuning and fine-tuning for code-switched tasks on 10 datasets across 4 language pairs. Our model outperforms the prompt-tuning approach by significant margins across all datasets and outperforms or remains at par with fine-tuning by using just 0.18% of total parameters. We also achieve competitive results when compared with the fine-tuned model in the low-resource cross-lingual and cross-task setting, indicating the effectiveness of our approach to incorporate new code-switched tasks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel instance-based prompt composition technique, PRO-CS, for CS tasks that combine language and task knowledge and achieves competitive results when compared with the fine-tuned model in the low-resource cross-lingual and cross-task setting, indicating the effectiveness of the approach to incorporate new code-switched tasks.'}",https://aclanthology.org/2022.emnlp-main.698.pdf
-"continually detection, rapidly react: unseen rumors detection based on continual prompt-tuning",Yuhui Zuo,"Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly. However, existing rumor detection (RD) models often assume the same training and testing distributions and can not cope with the continuously changing social network environment. This paper proposed a Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting (CF) of upstream tasks during sequential task learning and enables bidirectional knowledge transfer between domain tasks. Specifically, we propose the following strategies: (a) Our design explicitly decouples shared and domain-specific knowledge, thus reducing the interference among different domains during optimization; (b) Several technologies aim to transfer knowledge of upstream tasks to deal with emergencies; (c) A task-conditioned prompt-wise hypernetwork (TPHNet) is used to consolidate past domains. In addition, CPT-RD avoids CF without the necessity of a rehearsal buffer. Finally, CPT-RD is evaluated on English and Chinese RD datasets and is effective and efficient compared to prior state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables bidirectional knowledge transfer between domain tasks and is effective and efficient compared to prior state-of-the-art methods.'}",
-"continuous detection, rapidly react: unseen rumors detection based on continual prompt-tuning",Yuhui Zuo,"Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly. However, existing rumor detection (RD) models often assume the same training and testing distributions and cannot cope with the continuously changing social network environment. This paper proposes a Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables knowledge transfer between domain tasks. To avoid forgetting, we optimize and store task-special soft-prompt for each domain. Furthermore, we also propose several strategies to transfer knowledge of upstream tasks to deal with emergencies and a taskconditioned prompt-wise hypernetwork (TPHNet) to consolidate past domains, enabling bidirectional knowledge transfer. Finally, CPTRD is evaluated on English and Chinese RD datasets and is effective and efficient compared to state-of-the-art baselines, without data replay techniques and with only a few parameter tuning.1","{'model': 'tldr@v2.0.0', 'text': 'A Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting of upstream tasks during sequential task learning and enables knowledge transfer between domain tasks, and optimize and store task-special soft-prompt for each domain.'}",http://arxiv.org/pdf/2203.11720
-soft prompt guided joint learning for cross-domain sentiment analysis,Jingli Shi,"Aspect term extraction is a fundamental task in fine-grained sentiment analysis, which aims at detecting customer's opinion targets from reviews on product or service. The traditional supervised models can achieve promising results with annotated datasets, however, the performance dramatically decreases when they are applied to the task of cross-domain aspect term extraction. Existing cross-domain transfer learning methods either directly inject linguistic features into Language models, making it difficult to transfer linguistic knowledge to target domain, or rely on the fixed predefined prompts, which is time-consuming to construct the prompts over all potential aspect term spans. To resolve the limitations, we propose a soft prompt-based joint learning method for cross domain aspect term extraction in this paper. Specifically, by incorporating external linguistic features, the proposed method learn domain-invariant representations between source and target domains via multiple objectives, which bridges the gap between domains with varied distributions of aspect terms. Further, the proposed method interpolates a set of transferable soft prompts consisted of multiple learnable vectors that are beneficial to detect aspect terms in target domain. Extensive experiments are conducted on the benchmark datasets and the experimental results demonstrate the effectiveness of the proposed method for cross-domain aspect terms extraction.","{'model': 'tldr@v2.0.0', 'text': 'By incorporating external linguistic features, the proposed method learn domain-invariant representations between source and target domains via multiple objectives, which bridges the gap between domains with varied distributions of aspect terms.'}",http://arxiv.org/pdf/2303.00815
-towards large-scale 3d representation learning with multi-dataset point prompt training,Xiaoyang Wu,"The rapid advancement of deep learning models often attributes to their ability to leverage massive training data. In contrast, such privilege has not yet fully benefited 3D deep learning, mainly due to the limited availability of large-scale 3D datasets. Merging multiple available data sources and letting them collaboratively train a single model is a potential solution. However, due to the large domain gap between 3D point cloud datasets, such mixed supervision could adversely affect the model's performance and lead to degenerated performance (i.e., negative transfer) compared to single-dataset training. In view of this challenge, we introduce Point Prompt Training (PPT), a novel framework for multi-dataset synergistic learning in the context of 3D representation learning that supports multiple pre-training paradigms. Based on this framework, we propose Prompt-driven Normalization, which adapts the model to different datasets with domain-specific prompts and Language-guided Categorical Alignment that decently unifies the multiple-dataset label spaces by leveraging the relationship between label text. Extensive experiments verify that PPT can overcome the negative transfer associated with synergistic learning and produce generalizable representations. Notably, it achieves state-of-the-art performance on each dataset using a single weight-shared model with supervised multi-dataset training. Moreover, when served as a pre-training framework, it outperforms other pre-training approaches regarding representation quality and attains remarkable state-of-the-art performance across over ten diverse downstream tasks spanning both indoor and outdoor 3D scenarios.","{'model': 'tldr@v2.0.0', 'text': 'PPT is introduced, a novel framework for multi-dataset synergistic learning in the context of 3D representation learning that supports multiple pre-training paradigms and achieves state-of-the-art performance on each dataset using a single weight-shared model with supervised multi- dataset training.'}",https://arxiv.org/pdf/2308.09718
-adaptive prompt learning with distilled connective knowledge for implicit discourse relation recognition,Bang Wang,"Implicit discourse relation recognition (IDRR) aims at recognizing the discourse relation between two text segments without an explicit connective. Recently, the prompt learning has just been applied to the IDRR task with great performance improvements over various neural network-based approaches. However, the discrete nature of the state-art-of-art prompting approach requires manual design of templates and answers, a big hurdle for its practical applications. In this paper, we propose a continuous version of prompt learning together with connective knowledge distillation, called AdaptPrompt, to reduce manual design efforts via continuous prompting while further improving performance via knowledge transfer. In particular, we design and train a few virtual tokens to form continuous templates and automatically select the most suitable one by gradient search in the embedding space. We also design an answer-relation mapping rule to generate a few virtual answers as the answer space. Furthermore, we notice the importance of annotated connectives in the training dataset and design a teacher-student architecture for knowledge transfer. Experiments on the up-to-date PDTB Corpus V3.0 validate our design objectives in terms of the better relation recognition performance over the state-of-the-art competitors.","{'model': 'tldr@v2.0.0', 'text': 'A continuous version of prompt learning together with connective knowledge distillation, called AdaptPrompt, is proposed to reduce manual design efforts via continuous prompting while further improving performance via knowledge transfer.'}",https://arxiv.org/pdf/2309.07561
-prompt learning with knowledge memorizing prototypes for generalized few-shot intent detection,Chaiyut Luoyiching,"Generalized Few-Shot Intent Detection (GFSID) is challenging and realistic because it needs to categorize both seen and novel intents simultaneously. Previous GFSID methods rely on the episodic learning paradigm, which makes it hard to extend to a generalized setup as they do not explicitly learn the classification of seen categories and the knowledge of seen intents. To address the dilemma, we propose to convert the GFSID task into the class incremental learning paradigm. Specifically, we propose a two-stage learning framework, which sequentially learns the knowledge of different intents in various periods via prompt learning. And then we exploit prototypes for categorizing both seen and novel intents. Furthermore, to achieve the transfer knowledge of intents in different stages, for different scenarios we design two knowledge preservation methods which close to realistic applications. Extensive experiments and detailed analyses on two widely used datasets show that our framework based on the class incremental learning paradigm achieves promising performance.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-stage learning framework, which sequentially learns the knowledge of different intents in various periods via prompt learning, and designs two knowledge preservation methods which close to realistic applications.'}",https://arxiv.org/pdf/2309.04971
-rgb-t tracking via multi-modal mutual prompt learning,Yang Luo,"Object tracking based on the fusion of visible and thermal im-ages, known as RGB-T tracking, has gained increasing atten-tion from researchers in recent years. How to achieve a more comprehensive fusion of information from the two modalities with fewer computational costs has been a problem that re-searchers have been exploring. Recently, with the rise of prompt learning in computer vision, we can better transfer knowledge from visual large models to downstream tasks. Considering the strong complementarity between visible and thermal modalities, we propose a tracking architecture based on mutual prompt learning between the two modalities. We also design a lightweight prompter that incorporates attention mechanisms in two dimensions to transfer information from one modality to the other with lower computational costs, embedding it into each layer of the backbone. Extensive ex-periments have demonstrated that our proposed tracking ar-chitecture is effective and efficient, achieving state-of-the-art performance while maintaining high running speeds.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a tracking architecture based on mutual prompt learning between visible and thermal modalities and designs a lightweight prompter that incorporates attention mechanisms in two dimensions to transfer information from one modality to the other with lower computational costs.'}",https://arxiv.org/pdf/2308.16386
-visual-attribute prompt learning for progressive mild cognitive impairment prediction,Luoyao Kang,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed prompt learning model even outperforms the fully fine-tuning baseline on transferring the knowledge from AD to pMCI, and a novel global prompt token for the visual prompts to provide global guidance to the multi-modal representations is proposed.'}",
-deep learning model based on 3d optical coherence tomography images for the automated detection of pathologic myopia,S. Park,"Pathologic myopia causes vision impairment and blindness, and therefore, necessitates a prompt diagnosis. However, there is no standardized definition of pathologic myopia, and its interpretation by 3D optical coherence tomography images is subjective, requiring considerable time and money. Therefore, there is a need for a diagnostic tool that can automatically and quickly diagnose pathologic myopia in patients. This study aimed to develop an algorithm that uses 3D optical coherence tomography volumetric images (C-scan) to automatically diagnose patients with pathologic myopia. The study was conducted using 367 eyes of patients who underwent optical coherence tomography tests at the Ophthalmology Department of Incheon St. Mary’s Hospital and Seoul St. Mary’s Hospital from January 2012 to May 2020. To automatically diagnose pathologic myopia, a deep learning model was developed using 3D optical coherence tomography images. The model was developed using transfer learning based on four pre-trained convolutional neural networks (ResNet18, ResNext50, EfficientNetB0, EfficientNetB4). Grad-CAM was used to visualize features affecting the detection of pathologic myopia. The performance of each model was evaluated and compared based on accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC). The model based on EfficientNetB4 showed the best performance (95% accuracy, 93% sensitivity, 96% specificity, and 98% AUROC) in identifying pathologic myopia.","{'model': 'tldr@v2.0.0', 'text': 'This study aimed to develop an algorithm that uses 3D optical coherence tomography volumetric images (C-scan) to automatically diagnose patients with pathologic myopia and found the model based on EfficientNetB4 showed the best performance.'}",https://www.mdpi.com/2075-4418/12/3/742/pdf?version=1647601223
-trust-based intelligent routing protocol with q-learning for mission-critical wireless sensor networks,Dooho Keum,"Mission-critical wireless sensor networks require a trustworthy and punctual routing protocol to ensure the worst-case end-to-end delay and reliability when transmitting mission-critical data collected by various sensors to gateways. In particular, the trustworthiness of mission-critical data must be guaranteed for decision-making and secure communications. However, it is a challenging issue to meet the requirement of both reliability and QoS in sensor networking environments where cyber-attacks may frequently occur and a lot of mission-critical data is generated. This study proposes a trust-based routing protocol that learns the trust elements using Q-learning to detect various attacks and ensure network performance. The proposed mechanism ensures the prompt detection of cyber threats that may occur in a mission-critical wireless sensor network and guarantees the trustworthy transfer of mission-critical sensor data. This paper introduces a distributed transmission technology that prioritizes the trustworthiness of mission-critical data through Q-learning results considering trustworthiness, QoS, and energy factors. It is a technology suitable for mission-critical wireless sensor network operational environments and can reliably operate resource-constrained devices. We implemented and performed a comprehensive evaluation of our scheme using the OPNET simulator. In addition, we measured packet delivery rates, throughput, survivability, and delay considering the characteristics of mission-critical sensor networks. The simulation results show an enhanced performance when compared with other mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'A trust-based routing protocol that learns the trust elements using Q-learning to detect various attacks and ensure network performance and ensures the prompt detection of cyber threats that may occur in a mission-critical wireless sensor network and guarantees the trustworthy transfer of mission- critical sensor data is proposed.'}",https://www.mdpi.com/1424-8220/22/11/3975/pdf?version=1653919940
-accelerating reinforcement learning via predictive policy transfer in 6g ran slicing,A. M. Nagib,"Reinforcement Learning (RL) algorithms have recently been proposed to solve dynamic radio resource management (RRM) problems in beyond 5G networks. However, RL-based solutions are still not widely adopted in commercial cellular networks. One of the primary reasons for this is the slow convergence of RL agents when they are deployed in a live network and when the network’s context changes significantly. Concurrently, the open radio access network (O-RAN) paradigm promises to give mobile network operators (MNOs) more control over their networks, furthering the need for intelligent and RL-based network management. O-RAN’s standardized interfaces will allow MNOs to make real-time custom changes to intelligently control various RRM functionalities. We consider a RAN slicing scenario in which MNOs can modify the weights of the RL reward function. This enables MNOs to change the priorities of fulfilling the service level agreements of the slices. However, this results in a practical challenge since the RL agent needs to adapt promptly to the changes made by the MNO. This challenge is addressed in this paper, where we first present and discuss the results from an exhaustive experiment to examine the efficiency of using transfer learning (TL) to accelerate the convergence of RL-based RAN slicing in the considered scenario. We then propose a novel predictive approach to enhance the TL-based acceleration by selecting the best-saved policy for reuse. By adopting the proposed policy transfer approach, RL agents are able to converge up to 14000 learning steps faster than their non-accelerated counterparts. The proposed machine learning (ML)-based predictive approach also shows up to a 96.5% accuracy in selecting the best expert policy to reuse for acceleration.","{'model': 'tldr@v2.0.0', 'text': 'A novel predictive approach is proposed to enhance the TL-based acceleration by selecting the best-saved policy for reuse by adopting the proposed policy transfer approach, which is able to converge up to 14000 learning steps faster than their non-accelerated counterparts.'}",
-fetch: a deep-learning based classifier for fast transient classification,D. Agarwal,"With the upcoming commensal surveys for Fast Radio Bursts (FRBs), and their high candidate rate, usage of machine learning algorithms for candidate classification is a necessity. Such algorithms will also play a pivotal role in sending real-time triggers for prompt follow-ups with other instruments. In this paper, we have used the technique of Transfer Learning to train the state-of-the-art deep neural networks for classification of FRB and Radio Frequency Interference (RFI) candidates. These are convolutional neural networks which work on radio frequency-time and dispersion measure-time images as the inputs. We trained these networks using simulated FRBs and real RFI candidates from telescopes at the Green Bank Observatory. We present 11 deep learning models, each with an accuracy and recall above 99.5% on our test dataset comprising of real RFI and pulsar candidates. As we demonstrate, these algorithms are telescope and frequency agnostic and are able to detect all FRBs with signal-to-noise ratios above 10 in ASKAP and Parkes data. We also provide an open-source python package FETCH (Fast Extragalactic Transient Candidate Hunter) for classification of candidates, using our models. Using FETCH, these models can be deployed along with any commensal search pipeline for real-time candidate classification.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses the technique of Transfer Learning to train the state-of-the-art deep neural networks for classification of FRB and Radio Frequency Interference (RFI) candidates, which are convolutional neural networks which work on radio frequency-time and dispersion measure-time images as the inputs.'}",https://academic.oup.com/mnras/article-pdf/497/2/1661/33562918/staa1856.pdf
-retrieval-based knowledge transfer: an effective approach for extreme large language model compression,Jiduan Liu,"Large-scale pre-trained language models (LLMs) have demonstrated exceptional performance in various natural language processing (NLP) tasks. However, the massive size of these models poses huge challenges for their deployment in real-world applications. While numerous model compression techniques have been proposed, most of them are not well-suited for achieving extreme model compression when there is a significant gap in model scale. In this paper, we introduce a novel compression paradigm called Retrieval-based Knowledge Transfer (RetriKT), which effectively transfers the knowledge of LLMs to extremely small-scale models (e.g., 1%). In particular, our approach extracts knowledge from LLMs to construct a knowledge store, from which the small-scale model can retrieve relevant information and leverage it for effective inference. To improve the quality of the model, soft prompt tuning and Proximal Policy Optimization (PPO) reinforcement learning techniques are employed. Extensive experiments are conducted on low-resource tasks from SuperGLUE and GLUE benchmarks. The results demonstrate that the proposed approach significantly enhances the performance of small-scale models by leveraging the knowledge from LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A novel compression paradigm called Retrieval-based Knowledge Transfer (RetriKT), which effectively transfers the knowledge of LLMs to extremely small-scale models (e.g., 1%) and significantly enhances the performance of small- scale models by leveraging the knowledge from LLMs.'}",
-"pre-train, prompt and recommendation: a comprehensive survey of language modelling paradigm adaptations in recommender systems",Peng Liu,"The emergence of Pre-trained Language Models (PLMs) has achieved tremendous success in the field of Natural Language Processing (NLP) by learning universal representations on large corpora in a self-supervised manner. The pre-trained models and the learned representations can be beneficial to a series of downstream NLP tasks. This training paradigm has recently been adapted to the recommendation domain and is considered a promising approach by both academia and industry. In this paper, we systematically investigate how to extract and transfer knowledge from pre-trained models learned by different PLM-related training paradigms to improve recommendation performance from various perspectives, such as generality, sparsity, efficiency and effectiveness. Specifically, we propose a comprehensive taxonomy to divide existing PLM-based recommender systems w.r.t. their training strategies and objectives. Then, we analyze and summarize the connection between PLM-based training paradigms and different input data types for recommender systems. Finally, we elaborate on open issues and future research directions in this vibrant field.","{'model': 'tldr@v2.0.0', 'text': 'This paper systematically investigates how to extract and transfer knowledge from pre-trained models learned by different PLM-related training paradigms to improve recommendation performance from various perspectives, such as generality, sparsity, efficiency and effectiveness.'}",https://arxiv.org/pdf/2302.03735
-machine learning based potato leaves disease detection,Raju Deepak Potnuru,"The primary source of food, money, and employment for rural residents in economically developing nations is agriculture. Crop loss caused by plant diseases, which reduces production by 20 to 30%, is the main factor affecting agriculture productivity. Conventional methods have been used to diagnose the diseases in an attempt to prevent such losses, but they are inaccurate. To avoid losses brought on by such illnesses, accurate and prompt detection of plant diseases is vital. But occasionally those harvests and grains suffer a significant amount of damage, if not complete destruction, due to a lack of suitable cultivating knowledge, expertise, and sense of disease prediction. So, in order to lessen the loss caused by infections of plant leaves, this research attempts to integrate a portion of agriculture with the use of artificial intelligence. We used CNN transfer learning models such as VGG16, VGG19, and InceptionV3, to overcome this issue. To determine which strategy performs best at identifying potato leaf illnesses, we conducted trials using all three approaches on the standard dataset of potato leaves.",,
-can unsupervised knowledge transfer from social discussions help argument mining?,Subhabrata Dutta,"Identifying argument components from unstructured texts and predicting the relationships expressed among them are two primary steps of argument mining. The intrinsic complexity of these tasks demands powerful learning models. While pretrained Transformer-based Language Models (LM) have been shown to provide state-of-the-art results over different NLP tasks, the scarcity of manually annotated data and the highly domain-dependent nature of argumentation restrict the capabilities of such models. In this work, we propose a novel transfer learning strategy to overcome these challenges. We utilize argumentation-rich social discussions from the ChangeMyView subreddit as a source of unsupervised, argumentative discourse-aware knowledge by finetuning pretrained LMs on a selectively masked language modeling task. Furthermore, we introduce a novel prompt-based strategy for inter-component relation prediction that compliments our proposed finetuning method while leveraging on the discourse context. Exhaustive experiments show the generalization capability of our method on these two tasks over within-domain as well as out-of-domain datasets, outperforming several existing and employed strong baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work utilizes argumentation-rich social discussions from the ChangeMyView subreddit as a source of unsupervised, argumentative discourse-aware knowledge by finetuning pretrained LMs on a selectively masked language modeling task and introduces a novel prompt-based strategy for inter-component relation prediction that compliments the proposed finetuned method while leveraging on the discourse context.'}",http://arxiv.org/pdf/2203.12881
-generative adversarial imitation learning for empathy-based ai,Pratyush Muthukumar,"Generative adversarial imitation learning (GAIL) is a model-free algorithm that has been shown to provide strong results in imitating complex behaviors in high-dimensional environments. In this paper, we utilize the GAIL model for text generation to develop empathy-based context-aware conversational AI. Our model uses an expert trajectory of empathetic prompt-response dialogues which can accurately exhibit the correct empathetic emotion when generating a response. The Generator of the GAIL model uses the GPT-2 sequential pre-trained language model trained on 117 million parameters from 40 GB of internet data. We propose a novel application of an approach used in transfer learning to fine tune the GPT-2 model in order to generate concise, user-specific empathetic responses validated against the Discriminator. Our novel GAIL model utilizes a sentiment analysis history-based reinforcement learning approach to empathetically respond to human interactions in a personalized manner. We find that our model's response scores on various human-generated prompts collected from the Facebook Empathetic Dialogues dataset outperform baseline counterparts. Moreover, our model improves upon various history-based conversational AI models developed recently, as our model's performance over a sustained conversation of 3 or more interactions outperform similar conversational AI models.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel application of an approach used in transfer learning to fine tune the GPT-2 model in order to generate concise, user-specific empathetic responses validated against the Discriminator to develop empathy-based context-aware conversational AI.'}",
-three-round learning strategy based on 3d deep convolutional gans for alzheimer’s disease staging,Wenjie Kang,,"{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the proposed model avoids overfitting brought on by a paucity of sMRI data and enables the early detection of AD.'}",https://www.nature.com/articles/s41598-023-33055-9.pdf
-deep learning based voice assistant for the visually impaired,Renuga. K,"Visual impairment is a critical disorder that increases the affected individual's dependence on others in many ways, such as knowing their surroundings. Gaining prompt information about the environment, without any human dependency is beneficial for a visually impaired person. This could create an incredible impact in the lives of the visually impaired as they can learn about their surroundings. The main aim of the project is to help the visually impaired in knowing their environment. The image of the environment is captured in real time and processed to generate descriptions about the image that are read out loud to the vision impaired people using text-to-speech technology. Creating depiction sentences with genuine and right construction plays a pivotal role. A CNN-RNN hybrid model is utilized with the encoder – decoder mechanism; in the CNN part of the model, inception model for transfer learning is used and RNN is majorly used for language modeling. Inception-V3 is a pretrained model which is used to extract features from each image. The image captured lively is then compared and analyzed with other images in the training dataset so that precise description of the image is got from the trained model. The Microsoft Common Objects in Context (MS-COCO) dataset is used, and the Bahdanau Attention mechanism is implemented to make predictions more accurately. Various evaluation standards show that the proposed model achieves increased performance and sensible scores are accomplished.","{'model': 'tldr@v2.0.0', 'text': 'The main aim of the project is to help the visually impaired in knowing their environment by generating descriptions about the environment that are read out loud to the vision impaired people using text-to-speech technology.'}",
-hetgpt: harnessing the power of prompt tuning in pre-trained heterogeneous graph neural networks,Yihong Ma,"Graphs have emerged as a natural choice to represent and analyze the intricate patterns and rich information of the Web, enabling applications such as online page classification and social recommendation. The prevailing""pre-train, fine-tune""paradigm has been widely adopted in graph machine learning tasks, particularly in scenarios with limited labeled nodes. However, this approach often exhibits a misalignment between the training objectives of pretext tasks and those of downstream tasks. This gap can result in the""negative transfer""problem, wherein the knowledge gained from pre-training adversely affects performance in the downstream tasks. The surge in prompt-based learning within Natural Language Processing (NLP) suggests the potential of adapting a""pre-train, prompt""paradigm to graphs as an alternative. However, existing graph prompting techniques are tailored to homogeneous graphs, neglecting the inherent heterogeneity of Web graphs. To bridge this gap, we propose HetGPT, a general post-training prompting framework to improve the predictive performance of pre-trained heterogeneous graph neural networks (HGNNs). The key is the design of a novel prompting function that integrates a virtual class prompt and a heterogeneous feature prompt, with the aim to reformulate downstream tasks to mirror pretext tasks. Moreover, HetGPT introduces a multi-view neighborhood aggregation mechanism, capturing the complex neighborhood structure in heterogeneous graphs. Extensive experiments on three benchmark datasets demonstrate HetGPT's capability to enhance the performance of state-of-the-art HGNNs on semi-supervised node classification.","{'model': 'tldr@v2.0.0', 'text': 'HetGPT, a general post-training prompting framework to improve the predictive performance of pre-trained heterogeneous graph neural networks (HGNNs), is proposed, with the design of a novel prompting function that integrates a virtual class prompt and a heterogeneous feature prompt with the aim to reformulate downstream tasks to mirror pretext tasks.'}",
-approximated prompt tuning for vision-language pre-trained models,Qiong Wu,"Prompt tuning is a parameter-efficient way to deploy large-scale pre-trained models to downstream tasks by adding task-specific tokens. In terms of vision-language pre-trained (VLP) models, prompt tuning often requires a large number of learnable tokens to bridge the gap between the pre-training and downstream tasks, which greatly exacerbates the already high computational overhead. In this paper, we revisit the principle of prompt tuning for Transformer-based VLP models, and reveal that the impact of soft prompt tokens can be actually approximated via independent information diffusion steps, thereby avoiding the expensive global attention modeling and reducing the computational complexity to a large extent. Based on this finding, we propose a novel Approximated Prompt Tuning (APT) approach towards efficient VL transfer learning. To validate APT, we apply it to two representative VLP models, namely ViLT and METER, and conduct extensive experiments on a bunch of downstream tasks. Meanwhile, the generalization of APT is also validated on CLIP for image classification and StableDiffusion for text-to-image generation. The experimental results not only show the superior performance gains and computation efficiency of APT against the conventional prompt tuning methods, e.g., +7.01% accuracy and -82.30% additional computation overhead on METER, but also confirm its merits over other parameter-efficient transfer learning approaches.","{'model': 'tldr@v2.0.0', 'text': 'This paper revisits the principle of prompt tuning for Transformer-based VLP models, and reveals that the impact of soft prompt tokens can be actually approximated via independent information diffusion steps, thereby avoiding the expensive global attention modeling and reducing the computational complexity to a large extent.'}",https://arxiv.org/pdf/2306.15706
-efficient covid-19 identification using deep learning for iot,V. Veeraiah,"Internet of Things (IoT) has made it possible to diagnose and treat patients remotely, as well as to expedite the transportation of essential drugs and medical equipment to locations that are geographically separated. This has occurred at a time when society has become more socially distant. During the Ebola and COVID-19 outbreaks, the Internet of Things (IoT) technology was put to use in remote patient monitoring and the management of the vaccine cold chain. Concurrently, this study reflects on the variables that are required for IoT to scale. Since December 2019, the COVID-19 outbreak on a worldwide scale has developed into a significant problem. In order for medical treatment to be successful, it is essential to make a prompt and accurate diagnosis of persons who may be infected with the COVID-19 virus. In order to put a halt to the spread of COVID-19, it is important to construct an automated system that is based on deep transfer learning and is capable of detecting the virus based on chest X-rays. The authors of this study present an internet-of-things (IoT) system that makes use of ensemble deep transfer learning to diagnose COVID-19 patients at an earlier stage. It is feasible to keep an eye on potentially hazardous COVID-19 incidents as they occur so long as suitable procedures are adhered to. Inceptions A variety of different deep learning models are included into the framework that has been proposed for the Internet of Things. According to the findings of the study, the method that was suggested assisted radiologists in accurately and quickly identifying patients who could have COVID-19. The proposed effort focuses on developing an effective identification system based on the COVID-19 standard for use in an IoT setting.","{'model': 'tldr@v2.0.0', 'text': 'The authors of this study present an internet-of-things (IoT) system that makes use of ensemble deep transfer learning to diagnose COVID-19 patients at an earlier stage and the method that was suggested assisted radiologists in accurately and quickly identifying patients who could have CO VID-19.'}",
-transferable environment poisoning: training-time attack on reinforcement learning,Hang Xu,"Studying adversarial attacks on Reinforcement Learning (RL) agents has become a key aspect of developing robust, RL-based solutions. Test-time attacks, which target the post-learning performance of an RL agent’s policy, have been well studied in both whiteand black-box settings. More recently, however, state-of-the-art works have shifted to investigate training-time attacks on RL agents, i.e., forcing the learning process towards a target policy designed by the attacker. Alas, these SOTA works continue to rely on white-box settings and/or use a reward-poisoning approach. In contrast, this paper studies environment-dynamics poisoning attacks at training time. Furthermore, while environment-dynamics poisoning presumes a transfer-learning capable agent, it also allows us to expand our approach to black-box attacks. Our overall framework, inspired by hierarchical RL, seeks the minimal environment-dynamics manipulation that will prompt the momentary policy of the agent to change in a desired manner. We show the attack efficiency by comparing it with the reward-poisoning approach, and empirically demonstrate the transferability of the environment-poisoning attack strategy. Finally, we seek to exploit the transferability of the attack strategy to handle black-box settings.","{'model': 'tldr@v2.0.0', 'text': 'An overall framework, inspired by hierarchical RL, seeks the minimal environment-dynamics manipulation that will prompt the momentary policy of the agent to change in a desired manner and seeks to exploit the transferability of the attack strategy to handle black-box settings.'}",
-deep learning based ttc predictor for power systems with wind energy integration,Gao Qiu,"Assessing the security margin of interconnected tie-lines is highly important to power systems. However, the increase of highly probabilistic wind power output causes fast variation of boundary limit parameters such as total transfer capability (TTC) of tie-lines. The traditional physical model-based TTC calculation is extremely time-consuming. To overcome this shortage, a fast data-driven TTC predictor based on deep belief network (DBN) is presented in this paper for accurate and prompt awareness of transfer limits. In the first stage, a clustering-based sample production method using annual load and wind power data is designed to simulate a large volume of operation scenario samples for DBN training. The well-trained leaner is then employed to enable TTC prediction for the vital transmission path. Finally, a benchmark system is used to validate the proposed methodology.","{'model': 'tldr@v2.0.0', 'text': 'A fast data-driven TTC predictor based on deep belief network (DBN) is presented in this paper for accurate and prompt awareness of transfer limits.'}",
-explainable ai for retinoblastoma diagnosis: interpreting deep learning models with lime and shap,C. Chiesa-Estomba,"Retinoblastoma is a rare and aggressive form of childhood eye cancer that requires prompt diagnosis and treatment to prevent vision loss and even death. Deep learning models have shown promising results in detecting retinoblastoma from fundus images, but their decision-making process is often considered a “black box” that lacks transparency and interpretability. In this project, we explore the use of LIME and SHAP, two popular explainable AI techniques, to generate local and global explanations for a deep learning model based on InceptionV3 architecture trained on retinoblastoma and non-retinoblastoma fundus images. We collected and labeled a dataset of 400 retinoblastoma and 400 non-retinoblastoma images, split it into training, validation, and test sets, and trained the model using transfer learning from the pre-trained InceptionV3 model. We then applied LIME and SHAP to generate explanations for the model’s predictions on the validation and test sets. Our results demonstrate that LIME and SHAP can effectively identify the regions and features in the input images that contribute the most to the model’s predictions, providing valuable insights into the decision-making process of the deep learning model. In addition, the use of InceptionV3 architecture with spatial attention mechanism achieved high accuracy of 97% on the test set, indicating the potential of combining deep learning and explainable AI for improving retinoblastoma diagnosis and treatment.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that LIME and SHAP can effectively identify the regions and features in the input images that contribute the most to the model’s predictions, providing valuable insights into the decision-making process of the deep learning model.'}",https://www.mdpi.com/2075-4418/13/11/1932/pdf?version=1685598711
-breast cancer detection with machine learning-a review,Karnam Meghana,"Malignancy in the breast is a significant public health concern, where timely identification is essential for effective treatment. Machine Learning (ML) and Deep Learning (DL) algorithms are potential tools for prompt detection og breast malignancy through examination of medical images such as mammograms. Convolutional neural networks (CNNs), transfer learning, and ensemble learning are some of the recent techniques being used in this field. Despite the advantages of ML and DL algorithms for breast cancer detection, there are still several challenges that need to be addressed. The lack of diversity in the datasets used to train algorithms is one major challenge, with many datasets based on specific populations that may not represent others. Highly annotated data is also limited in medical field. The objective of this study is to provide researchers with valuable insights and guidance.","{'model': 'tldr@v2.0.0', 'text': 'The objective of this study is to provide researchers with valuable insights and guidance in breast cancer detection through examination of medical images such as mammograms.'}",
-p3o: transferring visual representations for reinforcement learning via prompting,Guoliang You,"It is important for deep reinforcement learning (DRL) algorithms to transfer their learned policies to new environments that have different visual inputs. In this paper, we introduce Prompt based Proximal Policy Optimization (P3O), a three-stage DRL algorithm that transfers visual representations from a target to a source environment by applying prompting. The process of P3O consists of three stages: pre-training, prompting, and predicting. In particular, we specify a prompt-transformer for representation conversion and propose a two-step training process to train the prompt-transformer for the target environment, while the rest of the DRL pipeline remains unchanged. We implement P3O and evaluate it on the OpenAI CarRacing video game. The experimental results show that P3O outperforms the state-of-the-art visual transferring schemes. In particular, P3O allows the learned policies to perform well in environments with different visual inputs, which is much more effective than retraining the policies in these environments.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces Prompt based Proximal Policy Optimization (P3O), a three-stage DRL algorithm that transfers visual representations from a target to a source environment by applying prompting, and shows that P3O outperforms the state-of-the-art visual transferring schemes.'}",https://arxiv.org/pdf/2303.12371
-diabetic retinopathy classification using deep learning,Abbaraju Sai Sathwik,"One of the main causes of adult blindness and a frequent consequence of diabetes is diabetic retinopathy (DR). To avoid visual loss, DR must be promptly identified and classified. In this article, we suggest an automated DR detection and classification method based on deep learning applied to fundus pictures. The suggested technique uses transfer learning for classification. On a dataset of 3,662 fundus images with real-world DR severity labels, we trained and validated our model. According to our findings, the suggested technique successfully detected and classified DR with an overall accuracy of 78.14%. Our model fared better than other recent cutting-edge techniques, illuminating the promise of deep learning-based strategies for DR detection and management. Our research indicates that the suggested technique may be employed as a screening tool for DR in a clinical environment, enabling early illness diagnosis and prompt treatment.","{'model': 'tldr@v2.0.0', 'text': 'An automated DR detection and classification method based on deep learning applied to fundus pictures that may be employed as a screening tool for DR in a clinical environment, enabling early illness diagnosis and prompt treatment.'}",https://publications.eai.eu/index.php/phat/article/download/4335/2670
-icpc: instance-conditioned prompting with contrastive learning for semantic segmentation,Chaohui Yu,"Modern supervised semantic segmentation methods are usually finetuned based on the supervised or self-supervised models pre-trained on ImageNet. Recent work shows that transferring the knowledge from CLIP to semantic segmentation via prompt learning can achieve promising performance. The performance boost comes from the feature enhancement with multimodal alignment, i.e., the dot product between vision and text embeddings. However, how to improve the multimodal alignment for better transfer performance in dense tasks remains underexplored. In this work, we focus on improving the quality of vision-text alignment from two aspects of prompting design and loss function, and present an instance-conditioned prompting with contrastive learning (ICPC) framework. First, compared with the static prompt designs, we reveal that dynamic prompting conditioned on image content can more efficiently utilize the text encoder for complex dense tasks. Second, we propose an align-guided contrastive loss to refine the alignment of vision and text embeddings. We further propose lightweight multi-scale alignment for better performance. Extensive experiments on three large-scale datasets (ADE20K, COCO-Stuff10k, and ADE20K-Full) demonstrate that ICPC brings consistent improvements across diverse backbones. Taking ResNet-50 as an example, ICPC outperforms the state-of-the-art counterpart by 1.71%, 1.05%, and 1.41% mIoU on the three datasets, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work reveals that dynamic prompting conditioned on image content can more efficiently utilize the text encoder for complex dense tasks and proposes an align-guided contrastive loss to refine the alignment of vision and text embeddings.'}",https://arxiv.org/pdf/2308.07078
-contrastive constraints guide explanation-based category learning,Seth Chin-Parker,"This paper provides evidence for a contrastive account of explanation that is motivated by pragmatic theories that recognize the contribution that context makes to the interpretation of a prompt for explanation. This study replicates the primary findings of previous work in explanation-based category learning (Williams & Lombrozo, 2010), extending that work by illustrating the critical role of the context in this type of learning. Participants interacted with items from two categories either by describing the items or explaining their category membership. We manipulated the feature overlap between the categories and examined both the explanations generated and acquired knowledge of the categories. Explanations for membership in a given category were influenced by the unprompted contrast category, indicating an important role for contrastive processing in the generation of explanations. The influence of the contrast category was similarly seen in the transfer performance of the participants.","{'model': 'tldr@v2.0.0', 'text': 'Evidence is provided for a contrastive account of explanation that is motivated by pragmatic theories that recognize the contribution that context makes to the interpretation of a prompt for explanation.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/cogs.12405
-the effects of goal-oriented instructions in digital game-based learning,Séverine Erhel,"Few studies have investigated the effects of the instructions provided in educational computer games on cognitive processing and learning outcomes. In our experiment, we sought to compare the effects on learning outcomes of two different types of goal-oriented instructions: mastery-goal instructions, which prompt learners to develop skills or master new knowledge, and performance-goal instructions, which are frequently used in game environments and which encourage individuals to demonstrate their ability to succeed, particularly by surpassing others. Results showed that a mastery-goal instruction elicited deeper learning (as assessed with a transfer task) than a performance-goal instruction. No effect of instruction was observed on either learning (demonstration consultation) times at the start of the game or on training task (solving riddles) performances during it. These results are discussed in terms of learning processes. This study demonstrates that mastery goal-oriented instructions can promote active processing of educational content in a serious game environment.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that mastery goal-oriented instructions can promote active processing of educational content in a serious game environment and elicited deeper learning than a performance-goal instruction.'}",
-gradient-based automated iterative recovery for parameter-efficient tuning,Maximilian Mozes,"Pretrained large language models (LLMs) are able to solve a wide variety of tasks through transfer learning. Various explainability methods have been developed to investigate their decision making process. TracIn (Pruthi et al., 2020) is one such gradient-based method which explains model inferences based on the influence of training examples. In this paper, we explore the use of TracIn to improve model performance in the parameter-efficient tuning (PET) setting. We develop conversational safety classifiers via the prompt-tuning PET method and show how the unique characteristics of the PET regime enable TracIn to identify the cause for certain misclassifications by LLMs. We develop a new methodology for using gradient-based explainability techniques to improve model performance, G-BAIR: gradient-based automated iterative recovery. We show that G-BAIR can recover LLM performance on benchmarks after manually corrupting training labels. This suggests that influence methods like TracIn can be used to automatically perform data cleaning, and introduces the potential for interactive debugging and relabeling for PET-based transfer learning methods.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that G-BAIR can recover LLM performance on benchmarks after manually corrupting training labels, suggesting that influence methods like TracIn can be used to automatically perform data cleaning, and introduces the potential for interactive debugging and relabeling for PET-based transfer learning methods.'}",http://arxiv.org/pdf/2302.06598
-detection and grade classification of diabetic retinopathy and adult vitelliform macular dystrophy based on ophthalmoscopy images,S. Srinivasan,"Diabetic retinopathy (DR) and adult vitelliform macular dystrophy (AVMD) may cause significant vision impairment or blindness. Prompt diagnosis is essential for patient health. Photographic ophthalmoscopy checks retinal health quickly, painlessly, and easily. It is a frequent eye test. Ophthalmoscopy images of these two illnesses are challenging to analyse since early indications are typically absent. We propose a deep learning strategy called ActiveLearn to address these concerns. This approach relies heavily on the ActiveLearn Transformer as its central structure. Furthermore, transfer learning strategies that are able to strengthen the low-level features of the model and data augmentation strategies to balance the data are incorporated owing to the peculiarities of medical pictures, such as their limited quantity and generally rigid structure. On the benchmark dataset, the suggested technique is shown to perform better than state-of-the-art methods in both binary and multiclass accuracy classification tasks with scores of 97.9% and 97.1%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a deep learning strategy called ActiveLearn that relies heavily on the ActiveLearn Transformer as its central structure and is shown to perform better than state-of-the-art methods in both binary and multiclass accuracy classification tasks.'}",https://www.mdpi.com/2079-9292/12/4/862/pdf?version=1675857334
-an aggregated loss function based lightweight few shot model for plant leaf disease classification,Shankey Garg,,"{'model': 'tldr@v2.0.0', 'text': 'An aggregated loss function formed by the combination of triplet loss and cross-entropy loss with MobileNetV2 as a base model for the effective classification of plant disease using small samples using only a few samples is used.'}",
-how teacher talk guidance during invention activities shapes students’ cognitive engagement and transfer,Catherine C. Chase,,,https://stemeducationjournal.springeropen.com/track/pdf/10.1186/s40594-019-0170-7
-cerebral micro-bleeding detection based on densely connected neural network,Shuihua Wang,"Cerebral micro-bleedings (CMBs) are small chronic brain hemorrhages that have many side effects. For example, CMBs can result in long-term disability, neurologic dysfunction, cognitive impairment and side effects from other medications and treatment. Therefore, it is important and essential to detect CMBs timely and in an early stage for prompt treatment. In this research, because of the limited labeled samples, it is hard to train a classifier to achieve high accuracy. Therefore, we proposed employing Densely connected neural network (DenseNet) as the basic algorithm for transfer learning to detect CMBs. To generate the subsamples for training and test, we used a sliding window to cover the whole original images from left to right and from top to bottom. Based on the central pixel of the subsamples, we could decide the target value. Considering the data imbalance, the cost matrix was also employed. Then, based on the new model, we tested the classification accuracy, and it achieved 97.71%, which provided better performance than the state of art methods.","{'model': 'tldr@v2.0.0', 'text': 'Densely connected neural network (DenseNet) is proposed as the basic algorithm for transfer learning to detectbral micro-bleedings and it provided better performance than the state of art methods.'}",https://www.frontiersin.org/articles/10.3389/fnins.2019.00422/pdf
-project based learning method using simulation tools and hardware exposure,P. V. V. R. Rao,"Power electronics have rapid growth in research and prompt emergent technology in electrical discipline. Study of Power Electronics (PE) for undergraduate and post graduate students is very much necessary as every industry is driven with power electronic converter based control of electric drives. This paper showcases the teaching of PE with effective teaching methodologies such as project based, lab acclimatize learning etc., Project Based Learning is recognized to be a motivating and problem centric teaching, that not only supports the students to acquire core knowledge but also helps the students to transfer their technical knowledge into realtime application. Problem will be given to students and asked to solve it to fulfil the industrial requirement and to obtain the course attainment. In this review paper Project Based Learning (PBL) is introduced and the feedback of the students before and after introducing the PBL is analysed.","{'model': 'tldr@v2.0.0', 'text': 'This paper showcases the teaching of PE with effective teaching methodologies such as project based, lab acclimatize learning etc., Project Based Learning is recognized to be a motivating and problem centric teaching, that not only supports theStudents to acquire core knowledge but also helps the students to transfer their technical knowledge into realtime application.'}",http://www.journaleet.org/index.php/jeet/article/download/85713/65807
-learning from clinical judgments: semi-markov-modulated marked hawkes processes for risk prognosis,A. Alaa,"Critically ill patients in regular wards are vulnerable to unanticipated adverse events which require prompt transfer to the intensive care unit (ICU). To allow for accurate prognosis of deteriorating patients, we develop a novel continuous-time probabilistic model for a monitored patient's temporal sequence of physiological data. Our model captures ""informatively sampled"" patient episodes: the clinicians' decisions on when to observe a hospitalized patient's vital signs and lab tests over time are represented by a marked Hawkes process, with intensity parameters that are modulated by the patient's latent clinical states, and with observable physiological data (mark process) modeled as a switching multi-task Gaussian process. In addition, our model captures ""informatively censored"" patient episodes by representing the patient's latent clinical states as an absorbing semi-Markov jump process. The model parameters are learned from offline patient episodes in the electronic health records via an EM-based algorithm. Experiments conducted on a cohort of patients admitted to a major medical center over a 3-year period show that risk prognosis based on our model significantly outperforms the currently deployed medical risk scores and other baseline machine learning algorithms.","{'model': 'tldr@v2.0.0', 'text': ""A novel continuous-time probabilistic model for a monitored patient's temporal sequence of physiological data is developed that significantly outperforms the currently deployed medical risk scores and other baseline machine learning algorithms.""}",
-prefixmol: target- and chemistry-aware molecule design via prefix embedding,Zhangyang Gao,"Is there a unified model for generating molecules considering different conditions, such as binding pockets and chemical properties? Although target-aware generative models have made significant advances in drug design, they do not consider chemistry conditions and cannot guarantee the desired chemical properties. Unfortunately, merging the target-aware and chemical-aware models into a unified model to meet customized requirements may lead to the problem of negative transfer. Inspired by the success of multi-task learning in the NLP area, we use prefix embeddings to provide a novel generative model that considers both the targeted pocket's circumstances and a variety of chemical properties. All conditional information is represented as learnable features, which the generative model subsequently employs as a contextual prompt. Experiments show that our model exhibits good controllability in both single and multi-conditional molecular generation. The controllability enables us to outperform previous structure-based drug design methods. More interestingly, we open up the attention mechanism and reveal coupling relationships between conditions, providing guidance for multi-conditional molecule generation.","{'model': 'tldr@v2.0.0', 'text': ""Inspired by the success of multi-task learning in the NLP area, prefix embeddings are used to provide a novel generative model that considers both the targeted pocket's circumstances and a variety of chemical properties, providing guidance for multi-conditional molecule generation.""}",http://arxiv.org/pdf/2302.07120
-task residual for tuning vision-language models,Tao Yu,"Large-scale vision-language models (VLMs) pre-trained on billion-level data have learned general visual representations and broad visual concepts. In principle, the welllearned knowledge structure of the VLMs should be inherited appropriately when being transferred to downstream tasks with limited data. However, most existing efficient transfer learning (ETL) approaches for VLMs either damage or are excessively biased towards the prior knowledge, e.g., prompt tuning (PT) discards the pre-trained text-based classifier and builds a new one while adapter-style tuning (AT) fully relies on the pre-trained features. To address this, we propose a new efficient tuning approach for VLMs named Task Residual Tuning (TaskRes), which performs directly on the text-based classifier and explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task. Specifically, TaskRes keeps the original classifier weights from the VLMs frozen and obtains a new classifier for the target task by tuning a set of prior-independent parameters as a residual to the original one, which enables reliable prior knowledge preservation and flexible task-specific knowledge exploration. The proposed TaskRes is simple yet effective, which significantly outperforms previous ETL methods (e.g., PT and AT) on 11 benchmark datasets while requiring minimal effort for the implementation. Our code is available at https://github.com/geekyutao/TaskRes.","{'model': 'tldr@v2.0.0', 'text': 'A new efficient tuning approach for VLMs named Task Residual Tuning (TaskRes), which performs directly on the text-based classifier and explicitly decouples the prior knowledge of the pre-trained models and new knowledge regarding a target task.'}",https://arxiv.org/pdf/2211.10277
-strategies for enhancing the multi-stage classification performances of her2 breast cancer from hematoxylin and eosin images,Md Sakib Hossain Shovon,"Breast cancer is a significant health concern among women. Prompt diagnosis can diminish the mortality rate and direct patients to take steps for cancer treatment. Recently, deep learning has been employed to diagnose breast cancer in the context of digital pathology. To help in this area, a transfer learning-based model called ‘HE-HER2Net’ has been proposed to diagnose multiple stages of HER2 breast cancer (HER2-0, HER2-1+, HER2-2+, HER2-3+) on H&E (hematoxylin & eosin) images from the BCI dataset. HE-HER2Net is the modified version of the Xception model, which is additionally comprised of global average pooling, several batch normalization layers, dropout layers, and dense layers with a swish activation function. This proposed model exceeds all existing models in terms of accuracy (0.87), precision (0.88), recall (0.86), and AUC score (0.98) immensely. In addition, our proposed model has been explained through a class-discriminative localization technique using Grad-CAM to build trust and to make the model more transparent. Finally, nuclei segmentation has been performed through the StarDist method.","{'model': 'tldr@v2.0.0', 'text': 'A transfer learning-based model called ‘HE-HER2Net’ has been proposed to diagnose multiple stages of HER2 breast cancer on H&E (hematoxylin & eosin) images from the BCI dataset to build trust and to make the model more transparent.'}",https://www.mdpi.com/2075-4418/12/11/2825/pdf?version=1669019352
-driver inattention monitoring system based on the orientation of the face using convolutional neural network,M. ManjulaP.,"Driving inattentively is one of the prime reasons for vehicle accidents worldwide and has significant implications for road safety. A prompt alert to the inattentive driver can mitigate many accidents and save numerous lives, and reduce the cost of damages caused by accidents. To achieve this, a proposal of a nonintrusive and noninvasive driver inattention monitoring and alerting system in real time has been put forward. A mobile camera mounted on the windshield captures the video of the driver. Viola-Jones algorithm detects the face in each frame of the video and the Kanade–Lucas–Tomasi (KLT) algorithm tracks the detected face from one frame to another frame. The driver is classified as inattentive or attentive using Convolutional Neural Network (CNN). The transfer learning of the AlexNet Convolutional Neural Network architecture is adopted for the classification. The accuracy, precision, sensitivity, F1 score, and specificity of the system proposed in this paper are 98.24%, 100%, 96.47%, 98.21% and 100%, respectively.","{'model': 'tldr@v2.0.0', 'text': 'A nonintrusive and noninvasive driver inattention monitoring and alerting system in real time that can mitigate many accidents and save numerous lives, and reduce the cost of damages caused by accidents.'}",
-extracting latent steering vectors from pretrained language models,Nishant Subramani,"Prior work on controllable text generation has focused on learning how to control language models through trainable decoding, smart-prompt design, or fine-tuning based on a desired objective. We hypothesize that the information needed to steer the model to generate a target sentence is already encoded within the model. Accordingly, we explore a different approach altogether: extracting latent vectors directly from pretrained language model decoders without fine-tuning. Experiments show that there exist steering vectors, which, when added to the hidden states of the language model, generate a target sentence nearly perfectly (> 99 BLEU) for English sentences from a variety of domains. We show that vector arithmetic can be used for unsupervised sentiment transfer on the Yelp sentiment benchmark, with performance comparable to models tailored to this task. We find that distances between steering vectors reflect sentence similarity when evaluated on a textual similarity benchmark (STS-B), outperforming pooled hidden states of models. Finally, we present an analysis of the intrinsic properties of the steering vectors. Taken together, our results suggest that frozen LMs can be effectively controlled through their latent steering space.","{'model': 'tldr@v2.0.0', 'text': 'The results suggest that frozen LMs can be effectively controlled through their latent steering space, and it is found that distances between steering vectors reflect sentence similarity when evaluated on a textual similarity benchmark (STS-B), outperforming pooled hidden states of models.'}",http://arxiv.org/pdf/2205.05124
-rethinking efficient tuning methods from a unified perspective,Zeyinzi Jiang,"Parameter-efficient transfer learning (PETL) based on large-scale pre-trained foundation models has achieved great success in various downstream applications. Existing tuning methods, such as prompt, prefix, and adapter, perform task-specific lightweight adjustments to different parts of the original architecture. However, they take effect on only some parts of the pre-trained models, i.e., only the feed-forward layers or the self-attention layers, which leaves the remaining frozen structures unable to adapt to the data distributions of downstream tasks. Further, the existing structures are strongly coupled with the Transformers, hindering parameter-efficient deployment as well as the design flexibility for new approaches. In this paper, we revisit the design paradigm of PETL and derive a unified framework U-Tuning for parameter-efficient transfer learning, which is composed of an operation with frozen parameters and a unified tuner that adapts the operation for downstream applications. The U-Tuning framework can simultaneously encompass existing methods and derive new approaches for parameter-efficient transfer learning, which prove to achieve on-par or better performances on CIFAR-100 and FGVC datasets when compared with existing PETL methods.","{'model': 'tldr@v2.0.0', 'text': 'A unified framework U-Tuning for parameter-efficient transfer learning is derived, which is composed of an operation with frozen parameters and a unified tuner that adapts the operation for downstream applications, which prove to achieve on-par or better performances on CIFAR-100 and FGVC datasets when compared with existing PETL methods.'}",http://arxiv.org/pdf/2303.00690
-prompting for multi-modal tracking,Jinyu Yang,"Multi-modal tracking gains attention due to its ability to be more accurate and robust in complex scenarios compared to traditional RGB-based tracking. Its key lies in how to fuse multi-modal data and reduce the gap between modalities. However, multi-modal tracking still severely suffers from data deficiency, thus resulting in the insufficient learning of fusion modules. Instead of building such a fusion module, in this paper, we provide a new perspective on multi-modal tracking by attaching importance to the multi-modal visual prompts. We design a novel multi-modal prompt tracker (ProTrack), which can transfer the multi-modal inputs to a single modality by the prompt paradigm. By best employing the tracking ability of pre-trained RGB trackers learning at scale, our ProTrack can achieve high-performance multi-modal tracking by only altering the inputs, even without any extra training on multi-modal data. Extensive experiments on 5 benchmark datasets demonstrate the effectiveness of the proposed ProTrack.","{'model': 'tldr@v2.0.0', 'text': 'A novel multi-modal prompt tracker (ProTrack), which can transfer the multi- modal inputs to a single modality by the prompt paradigm, and can achieve high-performance multi- Modal tracking by only altering the inputs, even without any extra training on multi-Modal data.'}",https://arxiv.org/pdf/2207.14571
-retrieval-augmented generative question answering for event argument extraction,Xinya Du,"Event argument extraction has long been studied as a sequential prediction problem with extractive-based methods, tackling each argument in isolation. Although recent work proposes generation-based methods to capture cross-argument dependency, they require generating and post-processing a complicated target sequence (template). Motivated by these observations and recent pretrained language models’ capabilities of learning from demonstrations. We propose a retrieval-augmented generative QA model (R-GQA) for event argument extraction. It retrieves the most similar QA pair and augments it as prompt to the current example’s context, then decodes the arguments as answers. Our approach outperforms substantially prior methods across various settings (i.e. fully supervised, domain transfer, and fewshot learning). Finally, we propose a clustering-based sampling strategy (JointEnc) and conduct a thorough analysis of how different strategies influence the few-shot learning performances.","{'model': 'tldr@v2.0.0', 'text': 'A retrieval-augmented generative QA model (R-GQA) is proposed for event argument extraction that retrieves the most similar QA pair and augments it as prompt to the current example’s context, then decodes the arguments as answers.'}",https://arxiv.org/pdf/2211.07067
-integrated parameter-efficient tuning for general-purpose audio models,Ju-ho Kim,"The advent of hyper-scale and general-purpose pre-trained models is shifting the paradigm of building task-specific models for target tasks. In the field of audio research, task-agnostic pre-trained models with high transferability and adaptability have achieved state-of-the-art performances through fine-tuning for downstream tasks. Nevertheless, re-training all the parameters of these massive models entails an enormous amount of time and cost, along with a huge carbon footprint. To overcome these limitations, the present study explores and applies efficient transfer learning methods in the audio domain. We also propose an integrated parameter-efficient tuning (IPET) framework by aggregating the embedding prompt (a prompt-based learning approach), and the adapter (an effective transfer learning method). We demonstrate the efficacy of the proposed framework using two backbone pre-trained audio models with different characteristics: the audio spectrogram transformer and wav2vec 2.0. The proposed IPET framework exhibits remarkable performance compared to fine-tuning method with fewer trainable parameters in four downstream tasks: sound event classification, music genre classification, keyword spotting, and speaker verification. Furthermore, the authors identify and analyze the shortcomings of the IPET framework, providing lessons and research directions for parameter efficient tuning in the audio domain.","{'model': 'tldr@v2.0.0', 'text': 'An integrated parameter-efficient tuning (IPET) framework is proposed by aggregating the embedding prompt (a prompt-based learning approach), and the adapter (an effective transfer learning method), demonstrating the efficacy of the proposed framework using two backbone pre-trained audio models with different characteristics.'}",http://arxiv.org/pdf/2211.02227
-classification of rail track crack using cnn with pre-trained vgg16 model,Shreetha Bhat,"One of the vital components of railway infrastructure is rail tracks. Maintenance of rail track has been a major challenge in most of the countries and one such challenge is the detection of cracks on the rail surface. To maintain good health of the tracks requires regular inspection and prompt action, failure to which, may lead to accidents and loss of lives. The railway department is introducing many innovative methods to make the inspection process efficient. In the past, various methods have been explored to detect defects on rail surfaces such as Computer Vision-Based method, but full automation is far from achievement. Few of the advanced countries are making use of Deep Learning techniques to monitor and maintain the condition of rail tracks. In, this paper, amalgamation of Convolutional Neural Network (CNN) and transfer learning is applied for classifying defective (with cracks) and non-defective rail surfaces.","{'model': 'tldr@v2.0.0', 'text': 'Al amalgamation of Convolutional Neural Network (CNN) and transfer learning is applied for classifying defective (with cracks) and non-defective rail surfaces.'}",
-virtual node tuning for few-shot node classification,Zhen Tan,"Few-shot Node Classification (FSNC) is a challenge in graph representation learning where only a few labeled nodes per class are available for training. To tackle this issue, meta-learning has been proposed to transfer structural knowledge from base classes with abundant labels to target novel classes. However, existing solutions become ineffective or inapplicable when base classes have no or limited labeled nodes. To address this challenge, we propose an innovative method dubbed Virtual Node Tuning (VNT). Our approach utilizes a pretrained graph transformer as the encoder and injects virtual nodes as soft prompts in the embedding space, which can be optimized with few-shot labels in novel classes to modulate node embeddings for each specific FSNC task. A unique feature of VNT is that, by incorporating a Graph-based Pseudo Prompt Evolution (GPPE) module, VNT-GPPE can handle scenarios with sparse labels in base classes. Experimental results on four datasets demonstrate the superiority of the proposed approach in addressing FSNC with unlabeled or sparsely labeled base classes, outperforming existing state-of-the-art methods and even fully supervised baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an innovative method dubbed Virtual Node Tuning (VNT), which utilizes a pretrained graph transformer as the encoder and injects virtual nodes as soft prompts in the embedding space, which can be optimized with few-shot labels in novel classes to modulate node embeddings for each specific FSNC task.'}",
-exploring hard samples in multiview for few-shot remote sensing scene classification,Yuyu Jia,"Few-shot remote sensing scene classification (RSSC) is of high practical value in real situations where data are scarce and annotated costly. The few-shot learner needs to identify new categories with limited examples, and the core issue of this assignment is how to prompt the model to learn transferable knowledge from a large-scale base dataset. Although current approaches based on transfer learning or meta-learning have achieved significant performance on this task, there are still two problems to be addressed: 1) as an essential characteristic of remote sensing (RS) images, spatial rotation insensitivity surprisingly remains largely unexplored and 2) the high distribution uncertainty of hard samples reduces the discriminative power of the model decision boundary. Stimulated by these, we propose a corresponding end-to-end framework termed a hard sample learning and multiview integration network (HSL-MINet). First, the multiview integration (MI) module contains a pretext task introduced to guide the knowledge transfer and a multiview-attention mechanism used to extract correlational information across different rotation views of images. Second, aiming at increasing the discrimination of the model decision boundary, the hard sample learning (HSL) module is designed to evaluate and select hard samples via a classwise adaptive threshold strategy and then decrease the uncertainty of their feature distributions by a devised triplet loss. Extensive evaluations on NWPU-RESISC45, WHU-RS19, and UCM datasets show that the effectiveness of our HSL-MINet surpasses the former state-of-the-art approaches.","{'model': 'tldr@v2.0.0', 'text': 'The effectiveness of the HSL-MINet surpasses the former state-of-the-art approaches on few-shot remote sensing scene classification datasets and is designed to evaluate and select hard samples via a classwise adaptive threshold strategy and decrease the uncertainty of their feature distributions by a devised triplet loss.'}",
-alexander knox at semeval-2023 task 5: the comparison of prompting and standard fine-tuning techniques for selecting the type of spoiler needed to neutralize a clickbait,M. Woźny,"Clickbait posts are a common problem on social media platforms, as they often deceive users by providing misleading or sensational headlines that do not match the content of the linked web page. The aim of this study is to create a technique for identifying the specific type of suitable spoiler - be it a phrase, a passage, or a multipart spoiler - needed to neutralize clickbait posts. This is achieved by developing a machine learning classifier analyzing both the clickbait post and the linked web page.Modern approaches for constructing a text classifier usually rely on fine-tuning a transformer-based model pre-trained on large unsupervised corpora. However, recent advances in the development of large-scale language models have led to the emergence of a new transfer learning paradigm based on prompt engineering.In this work, we study these two transfer learning techniques and compare their effectiveness for clickbait spoiler-type detection task.Our experimental results show that for this task, using the standard fine-tuning method gives better results than using prompting. The best model can achieve a similar performance to that presented by Hagen et al. (2022).","{'model': 'tldr@v2.0.0', 'text': 'The aim of this study is to create a technique for identifying the specific type of suitable spoiler - be it a phrase, a passage, or a multipart spoiler - needed to neutralize clickbait posts by developing a machine learning classifier analyzing both the clickb bait post and the linked web page.'}",https://aclanthology.org/2023.semeval-1.202.pdf
-pcbert: parent and child bert for chinese few-shot ner,Peichao Lai,"Achieving good performance on few-shot or zero-shot datasets has been a long-term challenge for NER. The conventional semantic transfer approaches on NER will decrease model performance when the semantic distribution is quite different, especially in Chinese few-shot NER. Recently, prompt-tuning has been thoroughly considered for low-resource tasks. But there is no effective prompt-tuning approach for Chinese few-shot NER. In this work, we propose a prompt-based Parent and Child BERT (PCBERT) for Chinese few-shot NER. To train an annotating model on high-resource datasets and then discover more implicit labels on low-resource datasets. We further design a label extension strategy to achieve label transferring from high-resource datasets. We evaluated our model on Weibo and the other three sampling Chinese NER datasets, and the experimental result demonstrates our approach’s effectiveness in few-shot learning.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a prompt-based Parent and Child BERT (PCBERT) for Chinese few-shot NER, and designs a label extension strategy to achieve label transferring from high-resource datasets.'}",
-wound severity classification using deep neural network,D. Anisuzzaman,"The classification of wound severity is a critical step in wound diagnosis. An effective classifier can help wound professionals categorize wound conditions more quickly and affordably, allowing them to choose the best treatment option. This study used wound photos to construct a deep neural network-based wound severity classifier that classified them into one of three classes: green, yellow, or red. The green class denotes wounds still in the early stages of healing and are most likely to recover with adequate care. Wounds in the yellow category require more attention and treatment than those in the green category. Finally, the red class denotes the most severe wounds that require prompt attention and treatment. A dataset containing different types of wound images is designed with the help of wound specialists. Nine deep learning models are used with applying the concept of transfer learning. Several stacked models are also developed by concatenating these transfer learning models. The maximum accuracy achieved on multi-class classification is 68.49%. In addition, we achieved 78.79%, 81.40%, and 77.57% accuracies on green vs. yellow, green vs. red, and yellow vs. red classifications for binary classifications. With a learning rate of 0.001 and an Adam optimizer, all models were trained for 250 epochs. Two callbacks were employed with the best validation accuracy and the optimal combination of validation and training accuracy saving. The "" sparse_categorical_crossentropy "" and "" binary_crossentropy "" loss functions were used for multi-class and binary classification, respectively.","{'model': 'tldr@v2.0.0', 'text': 'A deep neural network-based wound severity classifier that classified wounds into one of three classes: green, yellow, or red, which denotes wounds still in the early stages of healing and are most likely to recover with adequate care.'}",http://arxiv.org/pdf/2204.07942
-unihcp: a unified model for human-centric perceptions,,"In this section, we provide details and full results for oneshot fine-tuning and prompt tuning on human parsing and pose estimation. For each experiment, we sample ten sets of images with different random seeds; we also grid search on both iterations and learning rates until performance converges. The reported results are based on the best config found for each setting. Data sampling. In one-shot transfer experiments, only one image per class is used for a task [9]. Table 1 shows the number of sampled images on one-shot transfer tasks. Note that in UniHCP, classification tasks are multi-label classification for human parsing, pose estimation, and attribute recognition, where each query performs binary classification via the global probability unit. Therefore, we also make sure the presence of cases where a class is absent is covered in our samples. Such handling avoids the query simply learning to output 1 when the corresponding class always presents within the sampled images. On the other hand, when a class does appear in most of the images, e.g., all keypoint joints in pose estimation or the background class in human parsing, we are able to achieve reasonably good results without such handling, thus we do not intentionally sample “not present” case for keypoint joints and background class in our experiments.","{'model': 'tldr@v2.0.0', 'text': 'This section provides details and full results for oneshot fine-tuning and prompt tuning on human parsing and pose estimation on one-shot transfer tasks in UniHCP.'}",
-cloud-device collaborative adaptation to continual changing environments in the real-world,Yulu Gan,"When facing changing environments in the real world, the lightweight model on client devices suffers from severe performance drops under distribution shifts. The main limitations of the existing device model lie in (1) unable to update due to the computation limit of the device, (2) the limited generalization ability of the lightweight model. Meanwhile, recent large models have shown strong generalization capability on the cloud while they can not be deployed on client devices due to poor computation constraints. To enable the device model to deal with changing environments, we propose a new learning paradigm of Cloud-Device Collaborative Continual Adaptation, which encourages collaboration between cloud and device and improves the generalization of the device model. Based on this paradigm, we further propose an Uncertainty-based Visual Prompt Adapted (U-VPA) teacher-student model to transfer the generalization capability of the large model on the cloud to the device model. Specifically, we first design the Uncertainty Guided Sampling (UGS) to screen out challenging data continuously and transmit the most out-of-distribution samples from the device to the cloud. Then we propose a Visual Prompt Learning Strategy with Uncertainty guided updating (VPLU) to specifically deal with the selected samples with more distribution shifts. We transmit the visual prompts to the device and concatenate them with the incoming data to pull the device testing distribution closer to the cloud training distribution. We conduct extensive experiments on two object detection datasets with continually changing environments. Our proposed U-VPA teacher-student framework outperforms previous state-of-the-art test time adaptation and device-cloud collaboration methods. The code and datasets will be released.","{'model': 'tldr@v2.0.0', 'text': 'A new learning paradigm of Cloud-Device Collaborative Continual Adaptation, which encourages collaboration between cloud and device and improves the generalization of the device model is proposed, which outperforms previous state-of-the-art test time adaptation and device-cloud collaboration methods.'}",https://arxiv.org/pdf/2212.00972
-auto-prompting sam for mobile friendly 3d medical image segmentation,Chengyin Li,"The Segment Anything Model (SAM) has rapidly been adopted for segmenting a wide range of natural images. However, recent studies have indicated that SAM exhibits subpar performance on 3D medical image segmentation tasks. In addition to the domain gaps between natural and medical images, disparities in the spatial arrangement between 2D and 3D images, the substantial computational burden imposed by powerful GPU servers, and the time-consuming manual prompt generation impede the extension of SAM to a broader spectrum of medical image segmentation applications. To address these challenges, in this work, we introduce a novel method, AutoSAM Adapter, designed specifically for 3D multi-organ CT-based segmentation. We employ parameter-efficient adaptation techniques in developing an automatic prompt learning paradigm to facilitate the transformation of the SAM model's capabilities to 3D medical image segmentation, eliminating the need for manually generated prompts. Furthermore, we effectively transfer the acquired knowledge of the AutoSAM Adapter to other lightweight models specifically tailored for 3D medical image analysis, achieving state-of-the-art (SOTA) performance on medical image segmentation tasks. Through extensive experimental evaluation, we demonstrate the AutoSAM Adapter as a critical foundation for effectively leveraging the emerging ability of foundation models in 2D natural image segmentation for 3D medical image segmentation.","{'model': 'tldr@v2.0.0', 'text': ""This work introduces a novel method, AutoSAM Adapter, designed specifically for 3D multi-organ CT-based segmentation, and employs parameter-efficient adaptation techniques in developing an automatic prompt learning paradigm to facilitate the transformation of the SAM model's capabilities to 3D medical image segmentation.""}",https://arxiv.org/pdf/2308.14936
-transferring pre-trained multimodal representations with cross-modal similarity matching,Byoungjip Kim,"Despite surprising performance on zero-shot transfer, pre-training a large-scale multimodal model is often prohibitive as it requires a huge amount of data and computing resources. In this paper, we propose a method (BeamCLIP) that can effectively transfer the representations of a large pre-trained multimodal model (CLIP-ViT) into a small target model (e.g., ResNet-18). For unsupervised transfer, we introduce cross-modal similarity matching (CSM) that enables a student model to learn the representations of a teacher model by matching the relative similarity distribution across text prompt embeddings. To better encode the text prompts, we design context-based prompt augmentation (CPA) that can alleviate the lexical ambiguity of input text prompts. Our experiments show that unsupervised representation transfer of a pre-trained vision-language model enables a small ResNet-18 to achieve a better ImageNet-1K top-1 linear probe accuracy (66.2%) than vision-only self-supervised learning (SSL) methods (e.g., SimCLR: 51.8%, SwAV: 63.7%), while closing the gap with supervised learning (69.8%).","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a method (BeamCLIP) that can effectively transfer the representations of a large pre-trained multimodal model (CLIP-ViT) into a small target model (e.g., ResNet-18) and introduces cross-modal similarity matching (CSM) that enables a student model to learn the representation of a teacher model by matching the relative similarity distribution across text prompt embeddings.'}",http://arxiv.org/pdf/2301.02903
-engaging beachgoers for drowning prevention: the spillover effects on non-participants,Peter Kamstra,"ABSTRACT Despite reduced drowning incidence at lifeguard patrolled beaches, 71 drowning fatalities occurred on Australian beaches last year (2021–2022). Prevailing drowning prevention practices on beaches include patrolling lifeguards positioning safety flags in less hazardous locations and encouraging beachgoers to swim between them. Such methods represent a ‘deficit based’ approach to community engagement, in which experts determine acceptable behaviours and encourage adherence using a one-way transfer of information. Deficit based approaches can be useful, but participatory forms of community engagement are hypothesised to support learning that can transfer to other locations and through non-participants’ social networks. Using a lifeguard patrolled beach in Gerroa, Australia as a case study, we employed a ‘relationship building’ methodology to explore whether engagements on the beach can prompt transformational learning and whether such learning spills over to non-participants or to unpatrolled locations. Findings from 49 survey-interview engagements and 15 follow-up interviews suggest that building relationships with researchers is an enjoyable form of community engagement that contributes to learning about risk; simultaneously, findings suggest that learning can transform beachgoers’ intentions and practices at unpatrolled beaches. This paper provides a broadened theoretical and empirical model of community engagement aimed at beach drowning risk prevention via relationship building. Key policy highlights Engaging beachgoers via relationship building facilitates learning about beach risk, resulting in spillover effects to non-participants and to (unpatrolled) contexts. Engaging communities through dialogue is more likely to have a lasting influence on behaviours compared with deficit-based forms of engagement. Spillover effects to children, family, and friends provide evidence of who participatory research can have a ‘successful’ impact on. Experienced beachgoers discussing the engagement with others demonstrates how relationship building creates opportunities for experienced participants to demonstrate care for others. This study provides a broadened theoretical and empirical model of engagement aimed at beach drowning risk prevention via relationship building.",,
-student perceptions of staged transfer to independent research skills during a four-year honours science undergraduate program,Sarah L. Symons,"We describe interim results of an ongoing longitudinal pedagogical study investigating the efficacy of the Honours Integrated Science Program (iSci). We describe the pedagogical methods we use to prompt research skill development in a model from instructor-dependence to independent original research. We also describe a tool we use to help students organise their group research during their first attempts. Finally, we discuss students’ perceptions of how well iSci develops their research skills. Our results show that students are attracted to the iSci program because of the opportunities for research-based learning and skills development. We also found that in-program students value research skill development as a tool for successful completion of their degree and for their future academic or career plans. We conclude that our study methods help identify areas where we can support our students by building their research confidence and, in particular, their time-management skills. Nous presentons une description des resultats interimaires d’une etude pedagogique longitudinale qui vise a evaluer l’efficacite du programme specialise integre de sciences (iSci). Nous faisons une description des methodes pedagogiques que nous utilisons pour declencher le developpement des competences en recherche au sein d’un modele qui va de la recherche qui depend de l’instructeur a la recherche independante originale. Nous decrivons egalement un outil que nous utilisons pour aider les etudiants a organiser leur recherche par groupe au cours de leurs premieres tentatives. Pour finir, nous discutons les perceptions des etudiants sur la maniere dont le programme iSci developpe leurs competences en recherche. Nos resultats indiquent que les etudiants sont attires vers le programme iSci a cause des occasions d’apprentissage base sur la recherche et de developpement des competences. Nous avons egalement remarque que les etudiants inscrits au programme apprecient le developpement de competences en recherche en tant qu’outil qui leur servira a terminer leurs etudes et a obtenir leur diplome, ou pour leurs projets de carriere, a l’avenir. En conclusion, nous declarons que nos methodes d’etudes aident a identifier les domaines ou nous pouvons soutenir les etudiants en renforcant leur confiance en matiere de recherche et, en particulier, leurs competences en gestion du temps.",,https://ojs.lib.uwo.ca/index.php/cjsotl_rcacea/article/download/6999/5713
-analysis of the application of artificial intelligence technology in the protection of corporate governance rights and interests,Wen-liang Shen,"Corporate governance delivers feasible and controlled company operations using a group of common shareholders and appropriate policies. The roles and responsibilities of the shareholders suggest and improve corporate development through monotonous and independent rights. The implication of artificial intelligence provides knowledgeable insights for decision-making and control management. This article introduces a Mutual Consent-based Governance Regulation Model (MCGRM) for dissimilarity mitigation in corporate rule implications. The proposed model exploits transfer learning for balanced rule implication and decision-making. The learning states are defined based on mutual agreement, individual interest, and operational features. Based on the governance policies, the above rules are employed without hindering the pioneer regulations implemented in different periods. Therefore, artificial intelligence technology is utilized for prompt and swift governance decisions in delivering special rights for consumers and shareholders. The performance of this model is validated and verified using data sources related to governance policies from a real-time industry. The impact of varying policy features with dissimilarity is analyzed for varying occurrences. The analysis is given based on the considered data sources for which the classification and its impact over reports, sharing, voting, complaint, and market are analyzed. The availability before and after the proposed improves the above metrics by 10.48, 10.65, 9.78, 13.39, and 9.26%.",,https://www.frontiersin.org/articles/10.3389/fpsyg.2022.966689/pdf
-srcb at the ntcir-16 real-mednlp task,Yongwei Zhang,"The SRCB participated in subtask1: Few-resource Named Entity Recognition (NER) and subtask3: Adverse Drug Event detection (ADE) in NTCIR-16 Real-MedNLP. This paper reports our approach to solve the problem and discusses the official results. For the Few-resource NER subtask, we developed NER systems based on pretraining model, span-based classification and prompt learning. In addition, data augmentation and model ensemble are used to further improve performance. For ADE subtask, we mainly adopted two methods: multi-class classification and prompt learning. We em-ployed a two-stage training strategy to solve the long tail distribution problem and applied transfer learning to improve performance of model.","{'model': 'tldr@v2.0.0', 'text': 'The SRCB participated in subtask1: Few-resource Named Entity Recognition and subtask3: Adverse Drug Event detection in NTCIR-16 Real-MedNLP and adopted two methods: multi-class classification and prompt learning.'}",
-controllable generation of dialogue acts for dialogue systems via few-shot response generation and ranking,Angela Ramirez,"Dialogue systems need to produce responses that realize multiple types of dialogue acts (DAs) with high semantic fidelity. In the past, natural language generators (NLGs) for dialogue were trained on large parallel corpora that map from a domain-specific DA and its semantic attributes to an output utterance. Recent work shows that pretrained language models (LLMs) offer new possibilities for controllable NLG using prompt-based learning. Here we develop a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs. We compare eight few-shot prompt styles that include a novel method of generating from textual pseudo-references using a textual style transfer approach. We develop six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time. We test our approach on three domains and four LLMs. To our knowledge, this is the first work on NLG for dialogue that automatically ranks outputs using both DA and attribute accuracy. For completeness, we compare our results to fine-tuned few-shot models trained with 5 to 100 instances per DA. Our results show that several prompt settings achieve perfect DA accuracy, and near perfect semantic accuracy (99.81%) and perform better than few-shot fine-tuning.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs in natural language generators for dialogue and develops six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time.'}",https://arxiv.org/pdf/2307.14440
-adapting pre-trained language models to vision-language tasks via dynamic visual prompting,Shubin Huang,"Pre-trained language models (PLMs) have played an increasing role in multimedia research. In terms of vision-language (VL) tasks, they often serve as a language encoder and still require an additional fusion network for VL reasoning, resulting in excessive memory overhead. In this paper, we focus on exploring PLMs as a stand-alone model for VL reasoning tasks. Inspired by the recently popular prompt tuning, we first prove that the processed visual features can be also projected onto the semantic space of PLMs and act as prompt tokens to bridge the gap between single- and multi-modal learning. However, this solution exhibits obvious redundancy in visual information and model inference, and the placement of prompt tokens also greatly affects the final performance. Based on these observations, we further propose a novel transfer learning approach for PLMs, termed Dynamic Visual Prompting (DVP). Concretely, DVP first deploys a cross-attention module to obtain text-related and compact visual prompt tokens, thereby greatly reducing the input length of PLMs. To obtain the optimal placement, we also equip DVP with a reinforcement-learning based search algorithm, which can automatically merge DVP with PLMs for different VL tasks via a very short search process. In addition, we also experiment DVP with the recently popular adapter approach to keep the most parameters of PLMs intact when adapting to VL tasks, helping PLMs achieve a quick shift between single- and multi-modal tasks. We apply DVP to two representative PLMs, namely BERT and T5, and conduct extensive experiments on a set of VL reasoning benchmarks including VQA2.0, GQA and SNLIVE. The experimental results not only show the advantage of DVP on efficiency and performance, but also confirm its superiority in adapting pre-trained language models to VL tasks.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results not only show the advantage of DVP on efficiency and performance, but also confirm its superiority in adapting pre-trained language models to VL tasks.'}",https://arxiv.org/pdf/2306.00409
-supporting learning of variable control in a computer-based biology environment: effects of prompting college students to reflect on their own thinking,Xiaodong Lin,"While instruction on control of variables has been shown to be effective, especially when it encourages students to focus explicitly on rules or procedures, little evidence of application to novel problems has been obtained. We hypothesized that prompting students to understand their own learning processes while doing experiments involving control of variables would allow them to activate their reper- toire of knowledge and strategies and learn in a way that would enhance transfer of learning. Students were assigned to one of four versions of a computer-based biology simulation learning environment, each em- ploying a different type of prompt: reason justification, rule based, emotion focused, or none (control). Learning in this computer environment, college biology students designed and conducted experiments in- volving control of variables. Students' ability to solve both contextually similar (near transfer) and con- textually dissimilar (far transfer) problems was assessed. The treatment groups performed equally well on contextually similar problems. However, on a contextually dissimilar problem, the reason justification group had significantly higher scores than the other groups. Qualitative data showed that the reason justi- fication prompts directed students' attention to understanding when, why, and how to employ experiment design principles and strategies, and this in turn helped students to transfer their understanding to a novel problem. © 1999 John Wiley & Sons, Inc. J Res Sci Teach 36: 837- 858, 1999",,
-eco: ensembling context optimization for vision-language models,Lorenzo Agnolucci,"Image recognition has recently witnessed a paradigm shift, where vision-language models are now used to perform few-shot classification based on textual prompts. Among these, the CLIP model has shown remarkable capabilities for zero-shot transfer by matching an image and a custom textual prompt in its latent space. This has paved the way for several works that focus on engineering or learning textual contexts for maximizing CLIP's classification capabilities. In this paper, we follow this trend by learning an ensemble of prompts for image classification. We show that learning diverse and possibly shorter contexts improves considerably and consistently the results rather than relying on a single trainable prompt. In particular, we report better few-shot capabilities with no additional cost at inference time. We demonstrate the capabilities of our approach on 11 different benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that learning diverse and possibly shorter contexts improves considerably and consistently the results rather than relying on a single trainable prompt for image classification, and reports better few-shot capabilities with no additional cost at inference time.'}",https://arxiv.org/pdf/2307.14063
-generalizing few-shot named entity recognizers to unseen domains with type-related features,Zihan Wang,"Few-shot named entity recognition (NER) has shown remarkable progress in identifying entities in low-resource domains. However, few-shot NER methods still struggle with out-of-domain (OOD) examples due to their reliance on manual labeling for the target domain. To address this limitation, recent studies enable generalization to an unseen target domain with only a few labeled examples using data augmentation techniques. Two important challenges remain: First, augmentation is limited to the training data, resulting in minimal overlap between the generated data and OOD examples. Second, knowledge transfer is implicit and insufficient, severely hindering model generalizability and the integration of knowledge from the source domain. In this paper, we propose a framework, prompt learning with type-related features (PLTR), to address these challenges. To identify useful knowledge in the source domain and enhance knowledge transfer, PLTR automatically extracts entity type-related features (TRFs) based on mutual information criteria. To bridge the gap between training and OOD data, PLTR generates a unique prompt for each unseen example by selecting relevant TRFs. We show that PLTR achieves significant performance improvements on in-domain and cross-domain datasets. The use of PLTR facilitates model adaptation and increases representation similarities between the source and unseen domains.","{'model': 'tldr@v2.0.0', 'text': 'To identify useful knowledge in the source domain and enhance knowledge transfer, PLTR automatically extracts entity type-related features (TRFs) based on mutual information criteria and generates a unique prompt for each unseen example by selecting relevant TRFs.'}",
-optimizing learning environments: an individual difference approach to learning and transfer,Daniel M. Belenky,"Optimizing Learning Environments: An Individual Difference Approach to Learning and Transfer Daniel M. Belenky (dmb83@pitt.edu) Timothy J. Nokes (nokes@pitt.edu) Learning Research and Development Center University of Pittsburgh, 3939 O’Hara Street Pittsburgh, PA 15260 manipulations of task goals through instructions can change the ways students learn, similar to the effect of task structure. If directly manipulating task goals produces similar effects, it would offer a more direct way of encouraging students towards desired learning outcomes (whether towards transfer or skill). However, it is possible that achievement goals within a learning activity are not under conscious control, and task structure has more influence on how a student engages than instructions that attempt to prompt a particular achievement goal. It is also possible that task structure and task goals operate independently, leading to a three-way interaction in the adoption of achievement goals based on students’ prior dispositions. This study explores these possibilities. Abstract Prior work has found that the type of learning activity (direct instruction or invention) interacts with achievement goals (mastery or performance-oriented) such that invention tasks can help facilitate mastery goal adoption and knowledge transfer (Belenky & Nokes, 2009). In the current study, we investigated how robust the effect is, and whether explicit manipulations of the task goals can produce a similar effect. We conducted an experiment with 98 college students in which achievement goals were measured, while task goals and task structure were manipulated between subjects. Results indicated that task structure was generally a more effective way of influencing which achievement goals are adopted within a learning activity. However, task goals that promoted an evaluative context interfered with transfer for mastery- oriented learners from invention activities. The results are interpreted in relation to theories of regulatory fit and multiple goal hierarchies. Background Research on achievement goals has focused on two main aims; classifying what the goals are and then correlating those with predictors and outcomes. The prevailing classification is a 2 x 2 framework that has been well- validated (Elliot & McGregor, 2001). This framework separates the evaluative criterion (mastery or performance) from the valence (approach or avoidance), which results in four separable goals (mastery-approach, mastery-avoidance, performance-approach, performance-avoidance). Mastery goals refer to ones in which a person is basing his evaluation on the skill or competence he is trying to develop (that is, in comparison to an expectation or prior ability), while performance goals refer to evaluating oneself based on a normative standard (that is, in comparison to others). Approach goals refer to seeking out positive outcomes, while avoidance goals refer to averting negative ones. For example, a mastery-approach goal is one in which a person is seeking to improve his ability or knowledge, based on an internally-referenced criterion (“My aim is to completely master the material in this class”), while a performance- avoidance goal is one in which a person is seeking to not look bad compared to others (“My aim is to avoid doing worse than other students;” see Elliot & McGregor, 2001). Students can have different levels of each of these goals, and studies have validated that these four goals are separate factors (Elliot & McGregor, 2001). Because we are most interested in studying how different paths of successful learning affect what knowledge is gained, our work focuses on mastery-approach and performance-approach goals. Mastery-approach (MAP) goals have been correlated with a host of positive outcomes, such as intrinsic motivation, interest, better self-regulation, and deeper strategy use. Keywords: learning; transfer; skill acquisition; motivation; achievement goals. Student’s achievement goals have a large influence on their behaviors and experiences in academic settings. The literature surrounding Achievement Goal theory shows that these goals lead to very different patterns of affect, interest and achievement (e.g., Harackiewicz et al., 2005). However, this literature has not focused on how the goals influence what is learned. That is, although “achievement” is frequently measured as an outcome, it is almost always done at a coarse-grain level, such as final grades in a course. It is not clear how different achievement goals (mastery versus performance) are related to different kinds of learning, such as learning procedural skills, simple facts, or conceptual knowledge. To begin to address this gap, Belenky & Nokes (2009) examined how achievement goals impact the type of knowledge gained from different kinds of instruction. That study found that mastery-oriented learners do better on transfer measures, regardless of whether the mastery- orientation came from a stable predisposition or whether the open-ended structure of an “invention” task led to mastery- oriented feelings and goals in the specific context. Conversely, performance-oriented learners did better on skill acquisition when the instruction seemed to match their goals, by presenting a well-structured, simple task through direct instruction. This initial work has provided evidence that task structure interacts with existing achievement goals to influence learning. In the current work we examine whether direct",,
-analysis of factors affecting the length of stay in the emergency department for critically ill patients transferred to regional emergency medical centers,Hyung-Yeol Lee," BackgroundTransferring an emergency patient to another emergency department (ED) is necessary when she/he is unable to receive necessary treatment from the first visited ED, although the transfer poses potential risks for adverse clinical outcomes and lowering the quality of emergency medical services by overcrowding the transferred ED. This study aimed to understand the factors affecting the ED length of stay (LOS) of critically ill patients and to investigate whether they are receiving prompt treatment through Interhospital Transfer (IHT).MethodsThis study analyzed 968 critically ill patients transferred to the ED of the study site in 2019. Machine learning based prediction models were built to predict the ED LOS dichotomized as greater than 6 hours or less. Explanatory variables in patient characteristics, clinical characteristics, transfer-related characteristics, and ED characteristics were selected through univariate analyses.ResultsAmong the prediction models, the Logistic Regression (AUC 0.85) model showed the highest prediction performance, followed by Random Forest (AUC 0.83) and Naïve Bayes (AUC 0.83). The Logistic Regression model suggested that the need for emergency operation or angiography (OR 3.91, 95% CI=1.65–9.21), the need for Intensive Care Unit (ICU) admission (OR 3.84, 95% CI=2.53–5.83), fewer consultations (OR 3.57, 95% CI=2.84–4.49), a high triage level (OR 2.27, 95% CI=1.43–3.59), and fewer diagnoses (OR 1.32, 95% CI=1.09–1.61) coincided with a higher likelihood of 6-hour-or-less stays in the ED. Furthermore, an interhospital transfer handoff led to significantly shorter ED LOS among the patients who needed emergency operation or angiography, or ICU admission, or had a high triage level.ConclusionsThe results of this study suggest that patients prioritized in emergency treatment receive prompt intervention and leave the ED in time. Also, having a proper interhospital transfer handoff before IHT is crucial to provide efficient care and avoid unnecessarily longer stay in ED.","{'model': 'tldr@v2.0.0', 'text': 'The results of this study suggest that patients prioritized in emergency treatment receive prompt intervention and leave the ED in time, and having a proper interhospital transfer handoff before IHT is crucial to provide efficient care and avoid unnecessarily longer stay in ED.'}",
-an initial investigation of non-native spoken question-answering,V. Raina,"Text-based machine comprehension (MC) systems have a wide-range of applications, and standard corpora exist for developing and evaluating approaches. There has been far less research on spoken question answering (SQA) systems. The SQA task considered in this paper is to extract the answer from a candidate$\text{'}$s spoken response to a question in a prompt-response style language assessment test. Applying these MC approaches to this SQA task rather than, for example, off-topic response detection provides far more detailed information that can be used for further downstream processing. One significant challenge is the lack of appropriately annotated speech corpora to train systems for this task. Hence, a transfer-learning style approach is adopted where a system trained on text-based MC is evaluated on an SQA task with non-native speakers. Mismatches must be considered between text documents and spoken responses; non-native spoken grammar and written grammar. In practical SQA, ASR systems are used, necessitating an investigation of the impact of ASR errors. We show that a simple text-based ELECTRA MC model trained on SQuAD2.0 transfers well for SQA. It is found that there is an approximately linear relationship between ASR errors and the SQA assessment scores but grammar mismatches have minimal impact.","{'model': 'tldr@v2.0.0', 'text': 'It is found that there is an approximately linear relationship between ASR errors and the SQA assessment scores but grammar mismatches have minimal impact.'}",
-transfer effects of learning through two varied worked examples on word-problem solving,Zhang Qi,"There is disagreement with regard to the transfer effects of learning through varied worked examples. Ross(1989)found that multiple examples should be made very similar to each other; even a small difference in the surface feature could make the learner pay more attention to solving the problem. Gentner (2003) discovered that comparing examples with the same structure could prompt the learners' transition to problem solving. Holyoak (1987) pointed out that deeper structural differences had an impact on the transfer. Based on these researches, the present study hypothesized that the variability of the surface feature of two examples could have some effect on the near transfer of pupils' word-problem solving, and the proper variability of the structural feature of the two examples could have an impact on the far transfer of pupils' word-problem solving. A total of 210 second-grade students from a primary school were selected using a pretest and were divided into six experimental groups and a control group. Each group learnt a word problem example. Then, the pupils of three experimental groups learned one kind of word problem example that varied from the first example with regard to the surface feature: the pupils in the first group learned a number and substance varied example; the pupils in the second group learned a story varied example; and those in the third group learned an expression varied example. The pupils in the other three experimental groups learned another kind of worked example that varied from the first problem with regard to the structural feature. The first group learned a rule varied example, the second group learned a rule repeated example; the third group, learned a rule of speed composed example. The pupils in the control group did not learn any worked example. Finally, all the pupils of the six experimental groups and the control group were given a test with 15 word problems. Three of them varied from the first example with regard to the surface feature, three of them varied from the first example with regard to the structural feature, and the others varied from the first example with regard to the surface and structural features. The experimenters recorded the pupils' test scores. The results were as follows. First, the pupils in the control group who learned the first word problem eyample could solve the word problems whose example varied with regard to the surface feature, but could not solve the other word problems. Second, the pupils in the three experimental groups who learned the word problem example that varied with regard to the surface feature could solve the word problems that varied from the first example with regard to the surface feature, but could not solve the other word problems. Third, a different effect was observed for the pupils in the three experimental groups who learned the word problem example that varied with regard to the structural feature. The pupils who learned the rule repeated example solved more word problems than the pupils in the other groups. The pupils who learned the rule varied example solved more word problems than those who learned the rule of speed composed example. The pupils who learned the rule of speed composed example solved the fewest number of word problems. The variability of the surface feature of the two examples could have some impact on the near transfer of pupils' word-problem solving. The rule repeated example had the best effect on the far transfer of the pupils' word-problem solving. The rule varied example had a more positive effect than the rule of speed composed example on the pupils' word-problem solving far transfer. Finally, the rule of speed composed example had the poorest effect on the far transfer of the pupils' word-problem solving.","{'model': 'tldr@v2.0.0', 'text': ""It is hypothesized that the variability of the surface feature of two examples could have some effect on the near transfer of pupils' word-problem solving, and the proper variability ofThe structural feature of the two example could have an impact on the far transfer of pupil's word- problem solving.""}",
-toward educator-focused automated scoring systems for reading and writing,Mike Hardy,"This paper presents methods for improving automated essay scoring with techniques that address the computational trade-offs of self-attention and document length. To make Automated Essay Scoring (AES) more useful to practitioners, researchers must overcome the challenges of data and label availability, authentic and extended writing, domain scoring, prompt and source variety, and transfer learning. This paper addresses these challenges using neural network models by employing techniques that preserve essay length as an important feature without increasing model training costs. It introduces techniques for minimizing classification loss on ordinal labels using multi-objective learning, capturing semantic information across the entire essay using sentence embeddings to use transformer architecture across arbitrarily long documents, the use of such models for transfer learning, automated hyperparameter generation based on prompt-corpus metadata, and, most importantly, the use of semantic information to provide meaningful insights into student reading through analysis of passage-dependent writing resulting in state-of-the-art results for various essay tasks.","{'model': 'tldr@v2.0.0', 'text': None}",
-mobility awareness in cellular networks to support service continuity in vehicular users,Nandish P. Kuruvatti,"Mobile communication is an ubiquitously used technology that has evolved through various generations and is currently on the verge of its fifth generation (5G). In the recent years, Intelligent Transportation Systems (ITS) and supplementary vehicular use cases (e.g., autonomous driving) are considered widely within the scope of cellular networks. These use cases generally demand reliable and low latency services from the cellular network. Mobile Edge Clouds (MEC) in the 5G networks are often applicable to satiate such service demands of a vehicular user. However, the cellular handovers (HO) of vehicular users prompt frequent service migration among the MECs. The handovers and service migration increase the service interruption of a user. In this paper, we consider machine learning (ML) based mobility awareness to obtain future service migration and HO sites of a user. This enables smooth service migration by allowing non-state data transfer earlier to user handover. Further, it provides sufficient time for establishment of successful Coordinated Multipoint (CoMP) transmissions, which will reduce service interruption due to HO. Simulation results show that the proposed framework provides timely assistance for service migration and significantly reduces the service interruption time.","{'model': 'tldr@v2.0.0', 'text': 'Simulation results show that the proposed framework provides timely assistance for service migration and significantly reduces the service interruption time.'}",
-recognizing malaysia traffic signs with pre-trained deep convolutional neural networks,Tze How Dickson Neoh,"An essential component in the race towards the self-driving car is automatic traffic sign recognition. The capability to automatically recognize road signs allow self-driving cars to make prompt decisions such as adhering to speed limits, stopping at traffic junctions and so forth. Traditionally, feature-based computer vision techniques were employed to recognize traffic signs. However, recent advancements in deep learning techniques have shown to outperform traditional color and shape based detection methods. Deep convolutional neural network (DCNN) is a class of deep learning method that is most commonly applied to vision-related tasks such as traffic sign recognition. For DCNN to work well, it is imperative that the algorithm is given a vast amount of training data. However, due to the scarcity of a curated dataset of the Malaysian traffic signs, training DCNN to perform well can be very challenging. In this demonstrate that DCNN can be trained with little training data with excellent accuracy by using transfer learning. We retrain various pre-trained DCNN from other image recognition tasks by fine-tuning only the top layers on our dataset. Experiment results confirm that by using as little as 100 image samples for 5 different classes, we are able to classify hitherto traffic signs with above 90% accuracy for most pre-trained models and 98.33% for the DenseNet169 pre-trained model.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that DCNN can be trained with little training data with excellent accuracy by using transfer learning, and is able to classify hitherto traffic signs with above 90% accuracy for most pre-trained models and 98.33% for the DenseNet169 pre- trained model.'}",
-regarding solid oxide fuel cells simulation through artificial intelligence: a neural networks application,Arianna Baldinelli,"Because of their fuel flexibility, Solid Oxide Fuel Cells (SOFCs) are promising candidates to coach the energy transition. Yet, SOFC performance are markedly affected by fuel composition and operative parameters. In order to optimize SOFC operation and to provide a prompt regulation, reliable performance simulation tools are required. Given the high variability ascribed to the fuel in the wide range of SOFC applications and the high non-linearity of electrochemical systems, the implementation of artificial intelligence techniques, like Artificial Neural Networks (ANNs), is sound. In this paper, several network architectures based on a feedforward-backpropagation algorithm are proposed and trained on experimental data-set issued from tests on commercial NiYSZ/8YSZ/LSCF anode supported planar button cells. The best simulator obtained is a 3-hidden layer ANN (25/22/18 neurons per layer, hyperbolic tangent sigmoid as transfer function, obtained with a gradient descent with adaptive learning rate backpropagation). This shows high accuracy (RMS = 0.67% in the testing phase) and successful application in the forecast of SOFC polarization behaviour in two additional experiments (RMS in the order of 3% is scored, yet it is reduced to about 2% if only the typical operating current density range of real application is considered, from 300 to 500 mA·cm−2). Therefore, the neural tool is suitable for system simulation codes/software whether SOFC operating parameters agree with the input ranges (anode feeding composition 0–48%vol H2, 0–38%vol CO, 0–45%vol CH4, 9–32%vol CO2, 0–54%vol N2, specific equivalent hydrogen flow-rate per unit cell active area 10.8–23.6 mL·min−1·cm−2, current density 0–1300 mA·cm−2 and temperature 700–800 °C).",,https://www.mdpi.com/2076-3417/9/1/51/pdf?version=1547101227
-database for research projects to solve the inverse heat conduction problem,S. Szénási,"To achieve the optimal performance of an object to be heat treated, it is necessary to know the value of the Heat Transfer Coefficient (HTC) describing the amount of heat exchange between the work piece and the cooling medium. The prediction of the HTC is a typical Inverse Heat Transfer Problem (IHCP), which cannot be solved by direct numerical methods. Numerous techniques are used to solve the IHCP based on heuristic search algorithms having very high computational demand. As another approach, it would be possible to use machine-learning methods for the same purpose, which are capable of giving prompt estimations about the main characteristics of the HTC function. As known, a key requirement for all successful machine-learning projects is the availability of high quality training data. In this case, the amount of real-world measurements is far from satisfactory because of the high cost of these tests. As an alternative, it is possible to generate the necessary databases using simulations. This paper presents a novel model for random HTC function generation based on control points and additional parameters defining the shape of curve segments. As an additional step, a GPU accelerated finite-element method was used to simulate the cooling process resulting in the required temporary data records. These datasets make it possible for researchers to develop and test their IHCP solver algorithms.","{'model': 'tldr@v2.0.0', 'text': 'A novel model for random HTC function generation based on control points and additional parameters defining the shape of curve segments is presented, which makes it possible for researchers to develop and test their IHCP solver algorithms.'}",https://www.mdpi.com/2306-5729/4/3/90/pdf?version=1561620123
-principles of integrating research into teaching in higher education: a knowledge transfer perspective,R. Amaratunga,"Integrating new knowledge created through research with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures and research funding mechanisms within the UK higher education system. The link between research and teaching is not automatic. Thus, it needs to be formally created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study on which this paper is based on, aims to develop principles to enable transfer of research knowledge into teaching. This study is different from previous approaches in that it looks into the research and teaching link as a two-way knowledge transfer process in the light of growing knowledge management viewpoints. In developing the framework, this research uses the case study approach. Finally, the research introduces principles of research to teaching transfer that are applicable for higher education institutions.",,
-construction of a web-based virtual classroom and its effective analysis,Qingtang Liu,"In classroom, information between teachers and students includes course content and learning scene. Course content is mainly information, and learning scene is assisted information that can effectively prompt learners exchanging idea and acquiring knowledge. A based-Web virtual classroom should have the same features. Because of the limitation of bandwidth and QoS of network, it is very difficult to transfer teaching and learning information by video in Internet. The paper probes into a strategy of developing virtual classroom by the streaming media technologies. By contrast with two kinds of information, the Web-based virtual classroom we built includes two parts: real time teaching environment (RTTE) and collaborating learning environment (CLE). RTTE transports course content and feedback information into teachers or students. This paper firstly discusses the strategies of creating a virtual classroom for the remote learners. Secondly, it designs the architecture of a virtual classroom.","{'model': 'tldr@v2.0.0', 'text': 'The strategies of creating a virtual classroom for the remote learners are discussed and the architecture of a based-Web virtual classroom is designed, which includes real time teaching environment (RTTE) and collaborating learning environment (CLE).'}",
-principles of integrating research into teaching in higher education: built environment perspective,D. Amaratunga,"Integrating research-based knowledge with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures, and research funding mechanisms within the U.K. higher education system. The link between research and teaching is not automatic. Thus, it needs to be formally created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study on which this paper is based aims to develop generic principles to enable transfer of research knowledge into teaching. This study differs from previous approaches in that it examines the research and teaching link as a two-way knowledge transfer process in the light of growing knowledge management viewpoints. This research uses the case study approach and has conducted five case studies that represent five disciplines. Finally, the research introduces principles of research to teaching transfer that are applicable to higher education institutions.",,
-diversity: changing perspectives using case-based methods.,L. Andrews,"The case method is a popular teaching tool that can be applied to training for diversity and particularly the preparation of general education teachers for the inclusion of diverse learners in their classrooms. Case-based methodology teacher training activities focus on actual classroom experiences and prompt reflection and personal involvement. A study was done to validate these techniques with 56 student teachers in 2 groups. One group used a problem-solving strategy to analyze differing case features. This strategy group did 40 percent better on case problem-solving than those solving case problems without the strategy. A comparison was made of the identification of features with problem-solving scores to determine if the strategy was the determining factor, and a high positive correlation was found. The case method of teaching was found to promote classroom problem-solving when cases are systematically varied and when student teachers are given a strategy that guides transfer of learning from case to case. This demonstrates that student teachers can be trained to transfer learning from one teaching problem to another involving a diverse population of students. (Contains 8 references.) (JLS) *********************************************************************** Reproductions supplied by EDRS are the best that can be made from the original document. *********************************************************************** DIVERSITY: CHANGING PERSPECTIVES USING CASE-BASED METHODS A Paper presented at the 49th Annual Meeting of the American Association of Colleges for Teacher Education by Lanna Andrews, Ed.D Assistant Professor University of San Francisco 2130 Fulton Street San Francisco CA 94117-1980 (415) 666-2099 Office U.S. DEPARTMENT OF EDUCATION of Educational Research and Improvement EDUCATIONAL RESOURCES INFORMATION 140 CENTER (ERIC) 0 This document hes been reproduced as Ar) originating received from the person or organization (.011VT% 0 Minor changes have reproduction en made to improve Quality.. IK\ Points of view or opinions stated in thiadocument do not necesaarity represent official V3/43.1 OERI positron or 0011Cy. BEST COPY AVAILABLE 2 PERMISSION TO REPRODUCE AND DISSEMINATE THIS MATERIAL S BEEN GRANTED BY TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) Diversity: Changing Perspectives Using Case-Based Methods Since diversity is a term that accurately describes classrooms in today's urban centers, teacher trainers are charged with preparing novice teachers to meet the challenge (Grossman, 1995). This paper discusses the use of a popular teaching tool, the case method, and how it can be applied to training for diversity. Emphasis will be on the preparation of general education teachers for the inclusion of diverse learners into their classrooms. Topics to be covered include: (1) integration of cases into college curricula to promote cultural responsiveness; (2) use of varied techniques with the case method, such as cooperative learning, reflective writing, interactive discussion, problem-solving strategies; (3) and results of an empirical study incorporating all the above to promote diversity, completed in four sections of a Mainstreaming Course with general education teacher trainees. Theoretical Framework Discourse about culturally responsive curriculum takes place within the context of debate about a larger issue, the validity and shape of a multicultural approach to education. A widely held view of multicultural or culturally responsive curricula sees them as strategies for improving performance and enhancing self esteem among students whose racial, ethnic, or language-differs from that of the AngloEuropean population (Lubeck, 1988). This proposal takes the view that self esteems comes from inclusive practices in the classroom (Villa and Thousand, 1995). These include awareness and value of differences along with the realization that similarities are equally important, especially when designing curriculum. In addition, teachers must be trained to notice similarities and differences in order to transfer learning from",,
-integrating research and teaching in higher education: conceptual issues,S. Senaratne,"Integrating new knowledge created through research with teaching has become an important area that needs prompt attention with the growing emphasis on student learning activities, quality assurance procedures and research funding mechanisms within the UK higher education system. The link between research and teaching is not automatic. Thus, it needs to be created in higher education departments in order to achieve a productive relationship and manage research activities of university staff with teaching duties. The research study, on which this paper is based on, aims to develop principles in relation to transferring research knowledge into teaching through a literature review and case studies. The paper reports conceptual issues related to such a transfer process based on the literature findings.",,
-adaboost-based transfer learning method for positive and unlabelled learning problem,Bo Liu,,,
-an image-based transfer learning framework for classification of e-commerce products,Vrushali Atul Surve,"Classification of e-commerce products involves identifying the products and placing those products into the correct category. For example, men’s Nike Air Max will be in the men’s category shoes on an e-Commerce platform. Identifying the correct classification of a product from hundreds of categories is time-consuming for businesses. This research proposes an Image-based Transfer Learning Framework to classify the images into the correct category in the shortest time. The framework combines Image-based algorithms with Transfer Learning. This research compares the time to predict the category and accuracy of traditional CNN and transfer learning models such as VGG19, InceptionV3, ResNet50, and MobileNet. A visual classifier is trained CNN and transfer learning models such as VGG19, InceptionV3, ResNet50, and MobileNet. The models are trained on an e-commerce product dataset that combines the ImageNet dataset with pre-trained weights. The dataset consists of 15000 images scraped from the web. Results demonstrate that Inception V3 outperforms all other models based on a TIMING of 0.10 seconds and an accuracy of 85%.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes an Image-based Transfer Learning Framework to classify the images into the correct category in the shortest time and demonstrates that Inception V3 outperforms all other models based on a TIMING of 0.10 seconds and an accuracy of 85%.'}",https://norma.ncirl.ie/6322/1/vrushaliatulsurve.pdf
-a transfer learning-based method for personalized state of health estimation of lithium-ion batteries.,Guijun Ma,"State of health (SOH) estimation of lithium-ion batteries (LIBs) is of critical importance for battery management systems (BMSs) of electronic devices. An accurate SOH estimation is still a challenging problem limited by diverse usage conditions between training and testing LIBs. To tackle this problem, this article proposes a transfer learning-based method for personalized SOH estimation of a new battery. More specifically, a convolutional neural network (CNN) combined with an improved domain adaptation method is used to construct an SOH estimation model, where the CNN is used to automatically extract features from raw charging voltage trajectories, while the domain adaptation method named maximum mean discrepancy (MMD) is adopted to reduce the distribution difference between training and testing battery data. This article extends MMD from classification tasks to regression tasks, which can therefore be used for SOH estimation. Three different datasets with different charging policies, discharging policies, and ambient temperatures are used to validate the effectiveness and generalizability of the proposed method. The superiority of the proposed SOH estimation method is demonstrated through the comparison with direct model training using state-of-the-art machine learning methods and several other domain adaptation approaches. The results show that the proposed transfer learning-based method has wide generalizability as well as a positive precision improvement.","{'model': 'tldr@v2.0.0', 'text': 'This article extends MMD from classification tasks to regression tasks, which can be used for SOH estimation and shows that the proposed transfer learning-based method has wide generalizability as well as a positive precision improvement.'}",
-deep representation-based transfer learning for deep neural networks,Tao Yang,,,
-aspect-level sentiment classification based on attention-bilstm model and transfer learning,Guixian Xu,,,
-cdlx: an efficient novel approach for covi̇d detection lung x-rays throw transfer learning based on state of the art deep learning image classification models,Y. N. Afreen,"When compared to the general population, lung cancer patients have a higher incidence of COVID-19 infection, pulmonary problems, and poorer survival results. As a reference for prioritising cancer care issues during the epidemic, the world's main professional organisations issued new recommendations for the diagnosis, treatment, and follow-up of lung cancer patients. In today's world, we are fighting one of the greatest pandemics in human history, known as COVID-2019, which is caused by a coronavirus. The patient can be treated promptly if the infection is detected early (before it enters the lower respiratory tract). To observe ground-glass opacity in the chest X-ray due to fibrosis in the lungs once the virus has reached the lungs. Artificial intelligence techniques can be utilised to detect the presence and degree of illness based on the major discrepancies between X-ray images of an infected and non-infected person. For this study, I employed feature extraction from Transfer Learning, which entails importing a pre-trained CNN model, such as Distributed Deep Convolutional VGGNet or Distributed Deep Convolutional with ResNet Model, and changing the last layer to meet my needs.","{'model': 'tldr@v2.0.0', 'text': 'To observe ground-glass opacity in the chest X-ray due to fibrosis in the lungs once the virus has reached the lungs, artificial intelligence techniques can be utilised to detect the presence and degree of illness based on the major discrepancies betweenX-ray images of an infected and non-infected person.'}",https://sciencescholar.us/journal/index.php/ijhs/article/download/10593/7612
-fewfine: few-shot malware traffic classification via transfer learning based on fine-tuning strategy,Xingtong Liu,"Malware traffic is constantly evolving and remains destructive. The detection and classification of malware traffic is crucial for maintaining cyberspace security. Only by swiftly and accurately detecting and classifying malware traffic can user privacy and cyberspace security be effectively protected.In this paper, we propose FewFine, an approach for few-shot malware traffic classification based on transfer learning. We initially pre-train a detection model and two classification models with substantial quantity of malware and application traffic samples. For classifying new types of malware traffic accurately and promptly, we utilize transfer learning based on fine-tuning strategy and freeze several blocks in the pre-trained model. Utilizing prior knowledge from the pre-trained models, we leverage few samples of novel classes to perform accurate malware detection and classification. We execute extensive experiments on publicly available datasets to evaluate the effectiveness of FewFine. In model pre-training, with considerable number of samples, the accuracy of malware detection and classification can reach 0.99. The pre-trained models are saved for fine-tuning. When detecting and classifying novel malware traffic, FewFine can achieve the accuracy of 0.95 leveraging only 10 samples per class through fine-tuning the pre-trained model. It outperforms methods under comparison in terms of accuracy and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'FewFine, an approach for few-shot malware traffic classification based on transfer learning, outperforms methods under comparison in terms of accuracy and efficiency.'}",
-transfer learning based robust automatic detection system for diabetic retinopathy grading,C. Bhardwaj,,"{'model': 'tldr@v2.0.0', 'text': 'This work focuses on automatic DR disease identification and its grading by the means of transfer learning approach using dynamic investigation and the diagnosis ability of the proposed transfer learning-based model is justified by comparing the proposed methods with the state-of-the-art methods.'}",
-a concise peephole model based transfer learning method for small sample temporal feature-based data-driven quality analysis,Wentao Luo,,"{'model': 'tldr@v2.0.0', 'text': 'A novel transfer learning model and algorithm with feature mapping, feature learning and domain adaptation was proposed based on concise peephole model (TLCPM), which showed the TLCPM can be properly applied to analyze small sample temporal feature-based data and achieved good comprehensive performance.'}",
-output-based transfer learning in genetic programming for document classification,Wenlong Fu,,"{'model': 'tldr@v2.0.0', 'text': 'An investigation on an output-based transfer learning system using Genetic Programming in document classification tasks, which automatically selects features to construct classifiers, demonstrates that the GP programs from SD can be effectively used for classifying documents in the relevant TD.'}",
-motor imagery eeg signal classification based on deep transfer learning,Ming-fei Wei,"Deep transfer learning (DTL) has developed rapidly in the field of motor imagery (MI) on brain-computer interface (BCI) in recent years. DTL utilizes deep neural networks with strong generalization capabilities as the pre-training framework and automatically extracts richer and more expressive features during the training process. The goal of this paper is utilizing the DTL to classify MI electroencephalogram (EEG) signals on the premise of a small data set. The publicly available dataset III of the second BCI competition is applied in both the training part and testing part to evaluate the effectiveness of the proposed method. Firstly in the process, finite impulse response (FIR) filter and wavelet transform threshold denoising method are used to remove redundant signals and artifacts in EEG signals. Then, the continuous wavelet transform (CWT) is utilized to convert the one-dimensional EEG signal into a two-dimensional time-frequency amplitude representation as the input of the pre-trained convolutional neural network (CNN) for classifying two types of MI signals. Employing the input data of 140 trials for training, the final classification accuracy rate reaches 96.43%. Compared with the results of some superior machine learning models using the same data set, the accuracy and Kappa value of this DTL model are better. Therefore, the proposed scheme of MI EEG signal classification based on the DTL method offers preferably empirical performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed scheme of MI EEG signal classification based on the DTL method offers preferably empirical performance and the accuracy and Kappa value of this DTL model are better than some superior machine learning models using the same data set.'}",
-model-based transfer reinforcement learning based on graphical model representations,Yuewen Sun,"Reinforcement learning (RL) plays an essential role in the field of artificial intelligence but suffers from data inefficiency and model-shift issues. One possible solution to deal with such issues is to exploit transfer learning. However, interpretability problems and negative transfer may occur without explainable models. In this article, we define Relation Transfer as explainable and transferable learning based on graphical model representations, inferring the skeleton and relations among variables in a causal view and generalizing to the target domain. The proposed algorithm consists of the following three steps. First, we leverage a suitable casual discovery method to identify the causal graph based on the augmented source domain data. After that, we make inferences on the target model based on the prior causal knowledge. Finally, offline RL training on the target model is utilized as prior knowledge to improve the policy training in the target domain. The proposed method can answer the question of what to transfer and realize zero-shot transfer across related domains in a principled way. To demonstrate the robustness of the proposed framework, we conduct experiments on four classical control problems as well as one simulation to the real-world application. Experimental results on both continuous and discrete cases demonstrate the efficacy of the proposed method.","{'model': 'tldr@v2.0.0', 'text': 'Relation Transfer is defined as explainable and transferable learning based on graphical model representations, inferring the skeleton and relations among variables in a causal view and generalizing to the target domain.'}",
-deep transfer learning based multi-class brain tumors classification using mri images,Mrinmoy Mondal,"A brain tumor is a severe disease that can be fatal and significantly impacts one’s quality of life. The traditional method of identifying tumors relies on physicians, which is time-consuming and prone to errors, putting the patient’s life in jeopardy. Identifying the classes of brain tumors is difficult due to the high anatomical and spatial diversity of the brain tumor’s surrounding region. An automated and precise diagnosis approach is required to treat this severe disease effectively. Deep learning technology, such as CNN, can be used to diagnose various tumor types in the early stages of their development using brain MRI. In this study, a deep transfer learning framework based on VGG-19 is introduced to accurately detect three common kinds of tumors from brain MRI. There are primarily two phases to the suggested framework. The VGG-19 frozen part is the first phase, while the modified neural style classification part is the second phase. With certain modified techniques, the class imbalance impact within the MRI dataset and the generalization error issue during the training process were also resolved. The proposed model has a 94% classification accuracy and a 94% F1-score.","{'model': 'tldr@v2.0.0', 'text': 'A deep transfer learning framework based on VGG-19 is introduced to accurately detect three common kinds of tumors from brain MRI, which has a 94% classification accuracy and a 95% F1-score.'}",
-self-starting process monitoring based on transfer learning,Zhijun Wang,"Abstract Conventional self-starting control schemes can perform poorly when monitoring processes with early shifts, being limited by the number of historical observations sampled. In real applications, pre-observed data sets from other production lines are always available, prompting us to propose a scheme that monitors the target process using historical data obtained from other sources. The methodology of self-taught clustering from unsupervised transfer learning is revised to transfer knowledge from previous observations and improve out-of-control (OC) performance, especially for processes with early shifts. However, if the difference in distribution between the target process and the pre-observed data set is large, our scheme may not be the best. Simulation results and two illustrative examples demonstrate the superiority of the proposed scheme.","{'model': 'tldr@v2.0.0', 'text': 'The methodology of self-taught clustering from unsupervised transfer learning is revised to transfer knowledge from previous observations and improve out-of-control (OC) performance, especially for processes with early shifts.'}",
-voting in transfer learning system for ground-based cloud classification,M. Manzo,"Cloud classification is a great challenge in meteorological research. The different types of clouds, currently known and present in our skies, can produce radioactive effects that impact the variation of atmospheric conditions, with consequent strong dominance over the earth’s climate and weather. Therefore, identifying their main visual features becomes a crucial aspect. In this paper, the goal is to adopt pretrained deep neural networks-based architecture for clouds image description, and subsequently, classification. The approach is pyramidal. Proceeding from the bottom up, it partially extracts previous knowledge of deep neural networks related to original task and transfers it to the new task. The updated knowledge is integrated in a voting context to provide a classification prediction. The framework trains the neural models on unbalanced sets, a condition that makes the task even more complex, and combines the provided predictions through statistical measures. An experimental phase on different cloud image datasets is performed, and the results achieved show the effectiveness of the proposed approach with respect to state-of-the-art competitors.","{'model': 'tldr@v2.0.0', 'text': 'The goal is to adopt a pretrained deep neural networks based architecture for clouds image description, and subsequently, classification and results show the effectiveness of the proposed approach with respect to state of the art competitors.'}",https://www.mdpi.com/2504-4990/3/3/28/pdf?version=1626092557
-deep transfer learning-based network traffic classification for scarce dataset in 5g iot systems,Jianfeng Guan,,"{'model': 'tldr@v2.0.0', 'text': 'A traffic classification method based on deep transfer learning for 5G IoT scenarios with scarce labeled data and limited computing capability, and trains the classification model by weight transferring and neural network fine-tuning different from the previous work that extract artificially designed features.'}",
-transfer learning based surrogate assisted evolutionary bi-objective optimization for objectives with different evaluation times,Xilu Wang,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transfer learning scheme within a surrogate-assisted evolutionary algorithm framework to augment the training data for the surrogate for the slow objective function by transferring knowledge from the fast one and demonstrates that the proposed algorithm outperforms existing surrogate and non-surrogate-assisted delay-handling methods on a range of bi-objective optimization problems.'}",
-robust transfer learning based on geometric mean metric learning,Penghui Zhao,,"{'model': 'tldr@v2.0.0', 'text': 'This paper integrates the transfer learning and metric learning into a unified framework and proposes a novel robust transfer learning based on geometric mean metric learning, namely Geometric Mean Transfer Learning (GMTL), which uses weighted geometricmean metric learning to model the intra-class distance and the inter-class similarity.'}",
-ontology-based transfer learning in the airport and warehouse logistics domains,Miriam Herold,"This work is a position paper for the examination of ontology-based transfer learning in the context of business processes. We continue our preliminary work on transferring process-oriented knowledge from a well-known source domain to a less specified target domain. We outline our ideas on workflows from two specific contexts: passenger and baggage logistics at the airport on one hand and warehouse management logistics on the other hand. In the first step we automatically transform BPMN files from these two domains in two separate ontologies. In the next step we intend to use ontology mapping as a means for the transfer. We plan to examine the concepts of generalization and abstraction to ease the transfer. We claim that the mentioned domains are feasible candidates for transfer learning, as we find several analogies between the airport handling and warehouse management workflows. Additionally, we discuss possible utilization resp. benefits of the transfer learning within this two particular domains and draft the next steps for the future research.","{'model': 'tldr@v2.0.0', 'text': 'The preliminary work on transferring process-oriented knowledge from a well-known source domain to a less specified target domain is continued, and it is claimed that the mentioned domains are feasible candidates for transfer learning.'}",
-datlmedqa: a data augmentation and transfer learning based solution for medical question answering,Shuohua Zhou,"With the outbreak of COVID-19 that has prompted an increased focus on self-care, more and more people hope to obtain disease knowledge from the Internet. In response to this demand, medical question answering and question generation tasks have become an important part of natural language processing (NLP). However, there are limited samples of medical questions and answers, and the question generation systems cannot fully meet the needs of non-professionals for medical questions. In this research, we propose a BERT medical pretraining model, using GPT-2 for question augmentation and T5-Small for topic extraction, calculating the cosine similarity of the extracted topic and using XGBoost for prediction. With augmentation using GPT-2, the prediction accuracy of our model outperforms the state-of-the-art (SOTA) model performance. Our experiment results demonstrate the outstanding performance of our model in medical question answering and question generation tasks, and its great potential to solve other biomedical question answering challenges.","{'model': 'tldr@v2.0.0', 'text': 'This research proposes a BERT medical pretraining model, using GPT-2 for question augmentation and T5-Small for topic extraction, calculating the cosine similarity of the extracted topic and using XGBoost for prediction, which outperforms the state-of-the-art SOTA model performance.'}",https://www.mdpi.com/2076-3417/11/23/11251/pdf?version=1637938982
-"iglue: a benchmark for transfer learning across modalities, tasks, and languages",Emanuele Bugliarello,"Reliable evaluation benchmarks designed for replicability and comprehensiveness have driven progress in machine learning. Due to the lack of a multilingual benchmark, however, vision-and-language research has mostly focused on English language tasks. To fill this gap, we introduce the Image-Grounded Language Understanding Evaluation benchmark. IGLUE brings together - by both aggregating pre-existing datasets and creating new ones - visual question answering, cross-modal retrieval, grounded reasoning, and grounded entailment tasks across 20 diverse languages. Our benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in a zero-shot setting, but also in newly defined few-shot learning setups. Based on the evaluation of the available state-of-the-art models, we find that translate-test transfer is superior to zero-shot transfer and that few-shot learning is hard to harness for many tasks. Moreover, downstream performance is partially explained by the amount of available unlabelled textual data for pretraining, and only weakly by the typological distance of target-source languages. We hope to encourage future research efforts in this area by releasing the benchmark to the community.","{'model': 'tldr@v2.0.0', 'text': 'The Image-Grounded Language Understanding Evaluation benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in a zero-shot setting, but also in newly defined few-shot learning setups.'}",
-an uncertainty-aware transfer learning-based framework for covid-19 diagnosis,Afshar Shamsi,"The early and reliable detection of COVID-19 infected patients is essential to prevent and limit its outbreak. The PCR tests for COVID-19 detection are not available in many countries, and also, there are genuine concerns about their reliability and performance. Motivated by these shortcomings, this article proposes a deep uncertainty-aware transfer learning framework for COVID-19 detection using medical images. Four popular convolutional neural networks (CNNs), including VGG16, ResNet50, DenseNet121, and InceptionResNetV2, are first applied to extract deep features from chest X-ray and computed tomography (CT) images. Extracted features are then processed by different machine learning and statistical modeling techniques to identify COVID-19 cases. We also calculate and report the epistemic uncertainty of classification results to identify regions where the trained models are not confident about their decisions (out of distribution problem). Comprehensive simulation results for X-ray and CT image data sets indicate that linear support vector machine and neural network models achieve the best results as measured by accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC). Also, it is found that predictive uncertainty estimates are much higher for CT images compared to X-ray images.","{'model': 'tldr@v2.0.0', 'text': 'Comprehensive simulation results for X-ray and CT image data sets indicate that linear support vector machine and neural network models achieve the best results as measured by accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC).'}",https://ieeexplore.ieee.org/ielx7/5962385/9394817/09353390.pdf
-a quantile-based approach for hyperparameter transfer learning,David Salinas,"Bayesian optimization (BO) is a popular methodology to tune the hyperparameters of expensive black-box functions. Traditionally, BO focuses on a single task at a time and is not designed to leverage information from related functions, such as tuning performance objectives of the same algorithm across multiple datasets. In this work, we introduce a novel approach to achieve transfer learning across different datasets as well as different objectives. The main idea is to regress the mapping from hyperparameter to objective quantiles with a semi-parametric Gaussian Copula distribution, which provides robustness against different scales or outliers that can occur in different tasks. We introduce two methods to leverage this mapping: a Thompson sampling strategy as well as a Gaussian Copula process using such quantile estimate as a prior. We show that these strategies can combine the estimation of multiple objectives such as latency and accuracy, steering the hyperparameters optimization toward faster predictions for the same level of accuracy. Extensive experiments demonstrate significant improvements over state-of-the-art methods for both hyperparameter optimization and neural architecture search.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel approach to achieve transfer learning across different datasets as well as different objectives, to regress the mapping from hyperparameter to objective quantiles with a semi-parametric Gaussian Copula distribution.'}",
-towards a unified view on visual parameter-efficient transfer learning,Bruce X. B. Yu,"Parameter efficient transfer learning (PETL) aims at making good use of the representation knowledge in the pre-trained large models by fine-tuning a small number of parameters. Recently, taking inspiration from the natural language processing (NLP) domain, popular PETL techniques such as prompt-tuning and Adapter have also been successfully applied to the vision domain. However, prefix-tuning remains under-explored for vision tasks. In this work, we intend to adapt large vision models (LVMs) to downstream tasks with a good parameter-accuracy trade-off. Towards this goal, we propose a framework with a unified view of PETL called visual-PETL (V-PETL) to investigate the effects of different PETL techniques, data scales of downstream domains, positions of trainable parameters, and other aspects affecting the trade-off. Specifically, we analyze the positional importance of trainable parameters and differences between NLP and vision tasks in terms of data structures and pre-training mechanisms while implementing various PETL techniques, especially for the under-explored prefix-tuning technique. Based on a comprehensive understanding of the differences between NLP and vision data, we propose a new variation of the prefix-tuning module called parallel attention (PATT) for vision downstream tasks. An extensive empirical analysis on vision tasks via different frozen LVMs has been carried and the findings show that the proposed PATT can effectively contribute to other PETL techniques. An effective scheme Swin-BAPAT derived from the proposed V-PETL framework achieves significantly better performance than the state-of-the-art AdaptFormer-Swin with slightly more parameters and outperforms full-tuning with far fewer parameters. Code and data are available at: https://github.com/bruceyo/V-PETL.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a framework with a unified view of PETL called visual- PETL (V-PETL) to investigate the effects of different PETL techniques, data scales of downstream domains, positions of trainable parameters, and other aspects affecting the trade-off, and proposes a new variation of the prefix-tuning module called parallel attention (PATT) for vision downstream tasks.'}",http://arxiv.org/pdf/2210.00788
-a cnn based transfer learning model for automatic activity recognition from accelerometer sensors,B. Chikhaoui,,"{'model': 'tldr@v2.0.0', 'text': 'Deep learning based transfer learning algorithm based on convolutional neural networks (CNNs) that takes advantage of learned representations of activities of daily living from one dataset to recognize these activities in different other datasets characterized by different features including sensor modality, sampling rate, activity duration and environment is investigated.'}",http://r-libre.teluq.ca/1422/1/MLDM_2018_BChikhaoui_CRC.pdf
-a vision-based transfer learning approach for recognizing behavioral symptoms in people with dementia,Zachary Wharton,"With an aging population that continues to grow, dementia is a major global health concern. It is a syndrome in which there is a deterioration in memory, thinking, be-havior and the ability to perform activities of daily living. Depression and aggressive behavior are the most upsetting and challenging symptoms of dementia. Automatic recognition of these behaviors would not only be useful to alert family members and caregivers, but also helpful in planning and managing daily activities of people with dementia (PwD). In this work, we propose a vision-based approach that unifies transfer learning and deep convolutional neural network (CNN) for the effective recognition of behavioral symptoms. We also compare the performance of state-of-the-art CNN features with the hand-crafted HOG-feature, as well as their combination using a basic linear SVM. The proposed method is evaluated on a newly created dataset, which is based on the dementia storyline in ITVs Emmerdale episodes. The Alzheimer’s Society has described it as a ""realistic portrayal""1 of the condition to raise awareness of the issues surrounding dementia.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a vision-based approach that unifies transfer learning and deep convolutional neural network for the effective recognition of behavioral symptoms of dementia and compares the performance of state-of-the-art CNN features with the hand-crafted HOG-feature, as well as their combination using a basic linear SVM.'}",https://research.edgehill.ac.uk/ws/files/20125875/Vision-based-Behera.pdf
-developmental bayesian optimization of black-box with visual similarity-based transfer learning,Maxime Petit,"We present a developmental framework based on a long-term memory and reasoning mechanisms (Vision Similarity and Bayesian Optimisation). This architecture allows a robot to optimize autonomously hyper-parameters that need to be tuned from any action and/or vision module, treated as a black-box. The learning can take advantage of past experiences (stored in the episodic and procedural memories) in order to warm-start the exploration using a set of hyper-parameters previously optimized from objects similar to the new unknown one (stored in a semantic memory). As example, the system has been used to optimized 9 continuous hyper-parameters of a professional software (Kamido) both in simulation and with a real robot (industrial robotic arm Fanuc) with a total of 13 different objects. The robot is able to find a good object-specific optimization in 68 (simulation) or 40 (real) trials. In simulation, we demonstrate the benefit of the transfer learning based on visual similarity, as opposed to an amnesic learning (i.e. learning from scratch all the time). Moreover, with the real robot, we show that the method consistently outperforms the manual optimization from an expert with less than 2 hours of training time to achieve more than 88% of success.","{'model': 'tldr@v2.0.0', 'text': 'A developmental framework based on a long-term memory and reasoning mechanisms (Vision Similarity and Bayesian Optimisation) allows a robot to optimize autonomously hyper-parameters that need to be tuned from any action and/or vision module, treated as a black-box.'}",https://arxiv.org/pdf/1809.10141
-ttl: transformer-based two-phase transfer learning for cross-lingual news event detection,Hansi Hettiarachchi,,"{'model': 'tldr@v2.0.0', 'text': 'This research focuses on identifying event details at the sentence and token levels from news articles, considering their fine granularity, and proposes a novel learning strategy named Two-phase Transfer Learning (TTL), which allows the model to utilise the knowledge from a task at a particular data granularity for another task at different datagranularity.'}",https://link.springer.com/content/pdf/10.1007/s13042-023-01795-9.pdf
-an approach based on transfer learning to lifetime degradation rate prediction of the dry-type transformer,Y. Li,"Lifetime prediction of the power transformer plays an important role in maintaining the stable operation of power equipment. However, due to the complexity of insulation structure degenerative process, the existing methods featuring high cost and low precision are not effective enough in transformer life time prediction. Meanwhile, how to effectively and promptly respond to a new prediction scenario of insufficient and limited data is a common challenge for all the data-driven prediction methods. To address these concerns, a prediction approach of a back adoptive adjustment transfer learning scheme (BAATL) is proposed for lifetime degradation prediction of the dry-type transformer. The power transformer condition monitoring data of Supervisory Control and Data Acquisition system is conducted as the data driven. A deep neural network, a transfer learning module and a back adjustment module are constructed to realize feature extraction, domain adaptation and prediction network optimization. The proposed scheme is able to improve prediction accuracy and resolves the problems and drawbacks of traditional prediction methods, and presents its superior portability and application potential in the case of data shortage and scenario change. With authentic datasets, simulation tests performed on the condition monitoring data of dry-type transformers prove the effectiveness of the proposed scheme.","{'model': 'tldr@v2.0.0', 'text': 'A prediction approach of a back adoptive adjustment transfer learning scheme (BAATL) is proposed for lifetime degradation prediction of the dry-type transformer, able to improve prediction accuracy and resolves the problems and drawbacks of traditional prediction methods.'}",
-classification of citrus huanglongbing degree based on cbam-mobilenetv2 and transfer learning,Shiqing Dou,"Citrus has become a pivotal industry for the rapid development of agriculture and increasing farmers’ incomes in the main production areas of southern China. Knowing how to diagnose and control citrus huanglongbing has always been a challenge for fruit farmers. To promptly recognize the diagnosis of citrus huanglongbing, a new classification model of citrus huanglongbing was established based on MobileNetV2 with a convolutional block attention module (CBAM-MobileNetV2) and transfer learning. First, the convolution features were extracted using convolution modules to capture high-level object-based information. Second, an attention module was utilized to capture interesting semantic information. Third, the convolution module and attention module were combined to fuse these two types of information. Last, a new fully connected layer and a softmax layer were established. The collected 751 citrus huanglongbing images, with sizes of 3648 × 2736, were divided into early, middle, and late leaf images with different disease degrees, and were enhanced to 6008 leaf images with sizes of 512 × 512, including 2360 early citrus huanglongbing images, 2024 middle citrus huanglongbing images, and 1624 late citrus huanglongbing images. In total, 80% and 20% of the collected citrus huanglongbing images were assigned to the training set and the test set, respectively. The effects of different transfer learning methods, different model training effects, and initial learning rates on model performance were analyzed. The results show that with the same model and initial learning rate, the transfer learning method of parameter fine tuning was obviously better than the transfer learning method of parameter freezing, and that the recognition accuracy of the test set improved by 1.02~13.6%. The recognition accuracy of the citrus huanglongbing image recognition model based on CBAM-MobileNetV2 and transfer learning was 98.75% at an initial learning rate of 0.001, and the loss value was 0.0748. The accuracy rates of the MobileNetV2, Xception, and InceptionV3 network models were 98.14%, 96.96%, and 97.55%, respectively, and the effect was not as significant as that of CBAM-MobileNetV2. Therefore, based on CBAM-MobileNetV2 and transfer learning, an image recognition model of citrus huanglongbing images with high recognition accuracy could be constructed.","{'model': 'tldr@v2.0.0', 'text': 'An image recognition model of citrus huanglongbing images with high recognition accuracy could be constructed based on CBAM-MobileNetV2 and transfer learning based on a convolutional block attention module and a new fully connected layer and a softmax layer.'}",https://www.mdpi.com/1424-8220/23/12/5587/pdf?version=1686750838
-transfer learning-based quantized deep learning models for nail melanoma classification,Mujahid Hussain,,"{'model': 'tldr@v2.0.0', 'text': 'To facilitate an accurate, timely, and faster diagnosis of nail melanoma and to evaluate the early detection of other types of skin cancer, the proposed workflow could realize and achieve more than 95% accuracy.'}",
-detection of bangladeshi-produced plant disease using a transfer learning based on deep neural model,Tareq Hasan,"Plant diseases pose a significant threat to agricultural productivity and food security in Bangladesh. In this research, we address the challenge of timely and accurate plant disease detection through the application of transfer learning with deep neural models. We curated a diverse dataset comprising 18 categories of plant leaf images, including Bell pepper Bacterial spot, Bell pepper Healthy, Peach Healthy, Potato Early Blight, Rice Leaf Blast, Rice Healthy, Rice Brown Spot, Potato Healthy, Peach Bacterial spot, Corn Blight, Potato Late blight, Corn Healthy, Tomato Bacterial spot, Strawberry Leaf Scorch, Tomato Early blight, Tomato Early blight, Strawberry Healthy, and Tomato Healthy. The dataset represents the most prevalent plant diseases observed in the Bangladeshi context. We employed three state-of-the-art deep learning algorithms, EfficientNetV2M, VGG-19, and NASNetLarge, to develop robust plant disease detection models. Through transfer learning, these pre-trained models were fine-tuned on our specialized dataset to adapt them for the task at hand. The performance evaluation revealed impressive results, with EfficientNetV2M achieving an accuracy rate of 99%, VGG-19 achieving 93%, and NASNetLarge attaining 83% accuracy. The high accuracy of EfficientNetV2M showcases its exceptional capability in accurately classifying plant diseases prevalent in Bangladesh. The success of these deep neural models in detecting various plant diseases signifies their potential in revolutionizing plant disease management and enhancing agricultural practices. Our research contributes valuable insights into the effective use of transfer learning for plant disease detection and emphasizes the significance of dataset curation for improved model performance. The developed models hold promise in providing timely and precise disease diagnosis to farmers and agricultural professionals, thereby facilitating prompt interventions and minimizing crop losses. Future research can explore the integration of these deep neural models into practical agricultural tools, enabling real-time disease detection and offering substantial benefits to the agricultural industry in Bangladesh.",,https://al-kindipublisher.com/index.php/jcsts/article/download/5887/4915
-attention-based deep meta-transfer learning for few-shot fine-grained fault diagnosis,Chuanjiang Li,,,
-a novel method based on deep transfer unsupervised learning network for bearing fault diagnosis under variable working condition of unequal quantity,Hao Su,,,
-ecg biometrics method based on convolutional neural network and transfer learning,Yefei Zhang,"Personal identification based on ECG signals has been a significant challenge. The performance of an ECG authentication system depends significantly on the features extracted and the classifier subsequently applied. Although recently the deep neural networks based approaches featuring adaptive feature extractions and inherent classifications have attracted attention, they usually require a substantial set of training data. Aiming at tackling these issues, this paper presents a convolutional neural network-based transfer learning approach. It includes transferring the big data-trained GoogLeNet model into our identification task, fine-tuning the model using the ‘finetune’ idea, and adding three adaptive layers behind the original feature layer. The proposed approach not only requires a small set of training data, but also obtains great performance.","{'model': 'tldr@v2.0.0', 'text': 'A convolutional neural network-based transfer learning approach for personal identification based on ECG signals that includes transferring the big data-trained GoogLeNet model into the identification task, fine-tuning the model using the ‘finetune’ idea, and adding three adaptive layers behind the original feature layer.'}",
-facilitating efl learners’ active behaviors in speaking: a progressive question prompt-based peer-tutoring approach with vr contexts,Chin-Yu Chen,"ABSTRACT Speaking is a challenge for most English learners owing to the lack of practicing environments and partners. The virtual reality (VR) technology has changed this situation by providing learners with rich opportunities for practice. In VR-based practicing environments, peer tutoring is a common strategy for promoting peer assistance. However, in conventional VR-based learning, students may merely follow the guidelines and content provided by the teacher without thinking about the meaning and reasons for the practice in the situated contexts, which generally leads to poor learning outcomes. To solve this problem, a progressive question prompt-based peer-tutoring approach in VR contexts (PQP-PTVR) is proposed to assist students’ English-speaking development. An experiment was conducted to evaluate the effectiveness of the proposed approach. The experimental group used the proposed approach, while the control group learned with the conventional question prompt-based peer-tutoring approach in VR contexts (C-PTVR). The results showed that the experimental group achieved significantly greater improvement in speaking and self-efficacy than the control group. Concerning the participants’ learning behaviors, the experimental group had more interactions and made more attempts to improve their speaking. This study contributes to the existing literature by proposing the PQP-PTVR approach to enhancing learners’ speaking development, improving their interactions and self-efficacy.",,
-betel nut classification method based on transfer learning,Hengquan Cai,"Betel nut is one of the main chewing hobby in the world. The traditional methods of betel nut classification are mostly based on handcrafted features, and the classification stability is poor. To solve this problem and reduce labor costs for manual extraction of image features, this paper applies convolutional neural network to betel nut, and adopts parameter transfer strategy to realize transfer learning. The dataset of betel nut was constructed by shooting different types of betel nut with betel nut grading equipment. Aiming at the problem of insufficient samples, some parameters of AlexNet are transferred and fixed, new fully connected layers are build. Make full use of betel dataset to train and test the network. The experimental results indicate the proposed method is effective and low labor cost.","{'model': 'tldr@v2.0.0', 'text': 'This paper applies convolutional neural network to betel nut, and adopts parameter transfer strategy to realize transfer learning and experimental results indicate the proposed method is effective and low labor cost.'}",
-facial expression based imagination index and a transfer learning approach to detect deception,M. Hasan,"In this paper, we introduce a framework to automatically distinguish between facial expression sequences associated with imagining vs. remembering while answering a question. Our experiment includes a baseline and relevant questioning technique in the context of deception with 220 participants (20 hours long). Baseline questioning includes participants being separately asked to remember and imagine an arbitrary experience. During the relevant questioning, participants were prompted to either lie or tell the truth about a certain task. We trained a neural network model on the baseline data and achieved an accuracy of 60% on classifying imagining vs. remembering, whereas human performance for this task is 51%. Relevant questioning included a set of questions, each of which became an independent response segment. Using a transfer learning approach, we use the pretrained model from the baseline to obtain an imagination probability score for each relevant response segment. We define this individual probability per response as the Imagination Index. We apply the imagination indices as a feature vector to classify the whole relevant section as truth vs. bluff with an accuracy of 70%, significantly outperforming the human performance of 52%.","{'model': 'tldr@v2.0.0', 'text': 'A framework to automatically distinguish between facial expression sequences associated with imagining vs. remembering while answering a question is introduced and the pretrained model from the baseline is used to obtain an imagination probability score for each relevant response segment.'}",
-3d transfer learning network for classification of alzheimer’s disease with mri,Haifeng Wu,,"{'model': 'tldr@v2.0.0', 'text': 'The proposed three-dimensional (3D) transfer network is to realize the classification of 3D MRI data through an existing 2D transfer network, and it not only reduces the complexity of conventional 3D networks, but also improves the classification accuracy.'}",
-on the role of neural collapse in transfer learning,Tomer Galanti,"We study the ability of foundation models to learn representations for classification that are transferable to new, unseen classes. Recent results in the literature show that representations learned by a single classifier over many classes are competitive on few-shot learning problems with representations learned by special-purpose algorithms designed for such problems. In this paper we provide an explanation for this behavior based on the recently observed phenomenon that the features learned by overparameterized classification networks show an interesting clustering property, called neural collapse. We demonstrate both theoretically and empirically that neural collapse generalizes to new samples from the training classes, and -- more importantly -- to new classes as well, allowing foundation models to provide feature maps that work well in transfer learning and, specifically, in the few-shot setting.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated both theoretically and empirically that neural collapse generalizes to new samples from the training classes, and -- more importantly -- to new classes as well, allowing foundation models to provide feature maps that work well in transfer learning and, specifically, in the few-shot setting.'}",
-protein ubiquitylation and sumoylation site prediction based on ensemble and transfer learning,Fei He,"Ubiquitylation, a typical post-translational modification (PTM), plays an important role in signal transduction, apoptosis and cell proliferation. A ubiquitylation like PTM, sumoylation also may affect gene mapping, expression and genomic replication. Over the past two decades, machine learning has been widely employed in protein ubiquitylation and sumoylation site prediction tools. These existing tools require feature engineering, but failed to provide general interpretable features and probably underutilized the growing amount of data. This prompted us to propose a deep learning-based model that integrates multiple convolution and fully-connected layers of seven supervised learning sub-models to extract deep representations from protein sequences and physico-chemical properties (PCPs). Especially, we divided PCPs into 6 clusters and customized deep networks accordingly for handling the high correlations among one cluster. A stacking ensemble strategy was applied to combine these deep representations to make prediction. Furthermore, with the advantage of transfer learning, our deep learning model can work well on protein sumoylation site prediction as well after fine-tuning. On the high-quality annotated database Swiss-Prot, our model outperformed several well-known ubiquitylation and sumoylation site prediction tools. Our code is freely available at https://github.com/ruiwcoding/DeepUbiSumoPre.","{'model': 'tldr@v2.0.0', 'text': 'A deep learning-based model that integrates multiple convolution and fully-connected layers of seven supervised learning sub-models to extract deep representations from protein sequences and physico-chemical properties (PCPs) and outperformed several well-known ubiquitylation and sumoylation site prediction tools.'}",
-pac-net: a model pruning approach to inductive transfer learning,Sanghoon Myung,"Inductive transfer learning aims to learn from a small amount of training data for the target task by utilizing a pre-trained model from the source task. Most strategies that involve large-scale deep learning models adopt initialization with the pre-trained model and fine-tuning for the target task. However, when using over-parameterized models, we can often prune the model without sacrificing the accuracy of the source task. This motivates us to adopt model pruning for transfer learning with deep learning models. In this paper, we propose PAC-Net, a simple yet effective approach for transfer learning based on pruning. PAC-Net consists of three steps: Prune, Allocate, and Calibrate (PAC). The main idea behind these steps is to identify essential weights for the source task, fine-tune on the source task by updating the essential weights, and then calibrate on the target task by updating the remaining redundant weights. Under the various and extensive set of inductive transfer learning experiments, we show that our method achieves state-of-the-art performance by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'PAC-Net is proposed, a simple yet effective approach for transfer learning based on pruning that achieves state-of-the-art performance by a large margin under the various and extensive set of inductive transfer learning experiments.'}",http://arxiv.org/pdf/2206.05703
-transfer learning in differential privacy's hybrid-model,Reʾuven Kohen,"The hybrid-model (Avent et al 2017) in Differential Privacy is a an augmentation of the local-model where in addition to N local-agents we are assisted by one special agent who is in fact a curator holding the sensitive details of n additional individuals. Here we study the problem of machine learning in the hybrid-model where the n individuals in the curators dataset are drawn from a different distribution than the one of the general population (the local-agents). We give a general scheme -- Subsample-Test-Reweigh -- for this transfer learning problem, which reduces any curator-model DP-learner to a hybrid-model learner in this setting using iterative subsampling and reweighing of the n examples held by the curator based on a smooth variation of the Multiplicative-Weights algorithm (introduced by Bun et al, 2020). Our scheme has a sample complexity which relies on the chi-squared divergence between the two distributions. We give worst-case analysis bounds on the sample complexity required for our private reduction. Aiming to reduce said sample complexity, we give two specific instances our sample complexity can be drastically reduced (one instance is analyzed mathematically, while the other - empirically) and pose several directions for follow-up work.","{'model': 'tldr@v2.0.0', 'text': 'A general scheme is given -- Subsample-Test-Reweigh -- for this transfer learning problem, which reduces any curator-model DP-learner to a hybrid-model learner in this setting using iterative subsampling and reweighing of the n examples held by the curator based on a smooth variation of the Multiplicative-Weights algorithm.'}",
-virtual battery parameter identification using transfer learning based stacked autoencoder,Indrasis Chakraborty,"Recent studies have shown that the aggregated dynamic flexibility of an ensemble of thermostatic loads can be modeled in the form of a virtual battery. The existing methods for computing the virtual battery parameters require the knowledge of the first-principle models and parameter values of the loads in the ensemble. In real-world applications, however, it is likely that the only available information are end-use measurements such as power consumption, room temperature, device on/off status, etc., while very little about the individual load models and parameters are known. We propose a transfer learning based deep network framework for calculating virtual battery state of a given ensemble of flexible thermostatic loads, from the available end-use measurements. This proposed framework extracts first order virtual battery model parameters for the given ensemble. We illustrate the effectiveness of this novel framework on different ensembles of ACs and WHs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a transfer learning based deep network framework for calculating virtual battery state of a given ensemble of flexible thermostatic loads, from the available end-use measurements and illustrates the effectiveness of this novel framework on different ensembles of ACs and WHs.'}",https://arxiv.org/pdf/1810.04642
-transfer learning based approach for semantic person retrieval,Takuya Yaguchi,Many algorithms for semantic person retrieval suffer from a lack of training data often due to the difficulties in constructing a large dataset. We therefore propose a transfer learning based approach for semantic person identification and semantic person search. We apply the fine-tuned Mask R-CNN and DenseNet-161 for detection and attribute classification. The networks were pre-trained on the MS COCO and ILSVRC 2012 datasets. Our proposed approach achieves the highest recognition rate at each rank of CMC curve for semantic person identification and the highest average localization precision for semantic person search on our validation dataset.,"{'model': 'tldr@v2.0.0', 'text': ""This work proposes a transfer learning based approach for semantic person identification and semantic person search that achieves the highest recognition rate at each rank of CMC curve for semantic people identification and the highest average localization precision for semanticperson search on the authors' validation dataset.""}",https://eprints.soton.ac.uk/426371/1/CameraReady.pdf
-higher education students’ motivation to transfer learning: a scoping review,Doreen Bredenkamp,"PurposeThere is a need for higher education to produce graduates who are motivated to transfer learning into the workplace. Motivated graduates are work-ready and associated with increased performance. Presently, the research field around motivation to transfer learning by students in higher education is not clear and is inconsistent.Design/methodology/approachThis scoping review provides an overview of the characteristics of the literature, including key concepts, recommendations and gaps based on eight published articles on the motivation of students in higher education to transfer learning.FindingsThe results reflected a research field, which focused primarily on the influence of specific factors, namely student characteristics, educational design, the workplace environment, and on higher education students' motivation to transfer learning. The lack of a shared conceptual definition of motivation to transfer learning in higher education appears to influence the description of the results from the included studies. Most of the previous studies applied rigorous research designs.Originality/valueThis seemingly stunted research field related to higher education students' motivation to transfer learning needs to be amplified to influence the development of work-ready graduates from higher education. Approaches towards including all elements of motivation, expanding to other fields in higher education, including low-income countries, may be a proximal step in enhancing the trajectory of this research field.",,https://www.emerald.com/insight/content/doi/10.1108/HESWBL-03-2022-0057/full/pdf?title=higher-education-students-motivation-to-transfer-learning-a-scoping-review
-oryza sativa leaf disease detection using transfer learning,A. Musthafa,"Oryza sativa (Rice) is the world's most significant cereal harvest. It is taken as a staple feast for energy by the greater part of the total populace. Abiotic and biotic components like precipitation, soil richness, temperature, bugs, microscopic organisms, infections, etc. impact the yield creation amount and nature of rice grain. Ranchers contribute a great deal of time and energy to infection prevention, and they recognize sicknesses with their devastated unaided eye technique, which prompts unfortunate cultivating. The advancement of horticultural innovation helps significantly supports the computerized location of pathogenic living beings in the leaves of rice plants. The convolutional-based neural network calculation (CNN) is the one of very profound calculations that has been effectively used to settle PC vision issues like picture grouping, object division, picture investigation, etc. The proposed model boundaries have been tuned for the order work, and it has a great exactness of 95.67 percent. Using the transfer learning the data are trained faster andit can learn and apply the learned things in the next dataset faster. So that it does not acquire time in learning, which is not in the existing process.","{'model': 'tldr@v2.0.0', 'text': 'The convolutional-based neural network calculation (CNN) is the one of very profound calculations that has been effectively used to settle PC vision issues like picture grouping, object division, picture investigation, etc.'}",
-trcla: a transfer learning approach to reduce negative transfer for cellular learning automata,Seyyed Amir Hadi Minoofam,"In most traditional machine learning algorithms, the training and testing datasets have identical distributions and feature spaces. However, these assumptions have not held in many real applications. Although transfer learning methods have been invented to fill this gap, they introduce new challenges as negative transfers (NTs). Most previous research considered NT a significant problem, but they pay less attention to solving it. This study will propose a transductive learning algorithm based on cellular learning automata (CLA) to alleviate the NT issue. Two famous learning automata (LA) entitled estimators are applied as estimator CLA in the proposed algorithms. A couple of new decision criteria called merit and and attitude parameters are introduced to CLA to limit NT. The proposed algorithms are applied to standard LA environments. The experiments show that the proposed algorithm leads to higher accuracy and less NT results.","{'model': 'tldr@v2.0.0', 'text': 'This study will propose a transductive learning algorithm based on cellular learning automata (CLA) to alleviate the NT issue and shows that the proposed algorithm leads to higher accuracy and less NT results.'}",
-a survey on deep transfer learning,Chuanqi Tan,,"{'model': 'tldr@v2.0.0', 'text': 'This survey focuses on reviewing the current researches of transfer learning by using deep neural network and its applications and defined deep transfer learning, category and review the recent research works based on the techniques used inDeep transfer learning.'}",https://arxiv.org/pdf/1808.01974
-covid-19: automatic detection from x-ray images utilizing transfer learning with convolutional neural networks,Ioannis D. Apostolopoulos,,"{'model': 'tldr@v2.0.0', 'text': 'The results suggest that Deep Learning with X-ray imaging may extract significant biomarkers related to the Covid-19 disease, while the best accuracy, sensitivity, and specificity obtained is 96.78%, 98.66%, and 96.46% respectively.'}",https://link.springer.com/content/pdf/10.1007/s13246-020-00865-4.pdf
-enhancing cross-lingual natural language inference by prompt-learning from cross-lingual templates,Kunxun Qi,"Cross-lingual natural language inference (XNLI) is a fundamental task in cross-lingual natural language understanding. Recently this task is commonly addressed by pre-trained cross-lingual language models. Existing methods usually enhance pre-trained language models with additional data, such as annotated parallel corpora. These additional data, however, are rare in practice, especially for low-resource languages. Inspired by recent promising results achieved by prompt-learning, this paper proposes a novel prompt-learning based framework for enhancing XNLI. It reformulates the XNLI problem to a masked language modeling problem by constructing cloze-style questions through cross-lingual templates. To enforce correspondence between different languages, the framework augments a new question for every question using a sampled template in another language and then introduces a consistency loss to make the answer probability distribution obtained from the new question as similar as possible with the corresponding distribution obtained from the original question. Experimental results on two benchmark datasets demonstrate that XNLI models enhanced by our proposed framework significantly outperform original ones under both the full-shot and few-shot cross-lingual transfer settings.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results on two benchmark datasets demonstrate that XNLI models enhanced by the proposed framework significantly outperform original ones under both the full-shot and few-shot cross-lingual transfer settings.'}",https://aclanthology.org/2022.acl-long.134.pdf
-evolutionary based ensemble framework for realizing transfer learning in hiv-1 protease cleavage sites prediction,Deepak Singh,,"{'model': 'tldr@v2.0.0', 'text': 'A method for combining multiple weighted classifiers optimally by incorporating the knowledge derived from various amino acid encoding techniques is proposed, and the optimally coupled sequence of features and classifiers that characterized the heterogeneous feature is achieved promptly by genetic algorithm.'}",
-transfer learning on multi-fidelity data,Dong H. Song,"Neural networks (NNs) are often used as surrogates or emulators of partial differential equations (PDEs) that describe the dynamics of complex systems. A virtually negligible computational cost of such surrogates renders them an attractive tool for ensemble-based computation, which requires a large number of repeated PDE solves. Since the latter are also needed to generate sufficient data for NN training, the usefulness of NN-based surrogates hinges on the balance between the training cost and the computational gain stemming from their deployment. We rely on multi-fidelity simulations to reduce the cost of data generation for subsequent training of a deep convolutional NN (CNN) using transfer learning. High- and low-fidelity images are generated by solving PDEs on fine and coarse meshes, respectively. We use theoretical results for multilevel Monte Carlo to guide our choice of the numbers of images of each kind. We demonstrate the performance of this multi-fidelity training strategy on the problem of estimation of the distribution of a quantity of interest, whose dynamics is governed by a system of nonlinear PDEs (parabolic PDEs of multi-phase flow in heterogeneous porous media) with uncertain/random parameters. Our numerical experiments demonstrate that a mixture of a comparatively large number of low-fidelity data and smaller numbers of high- and low-fidelity data provides an optimal balance of computational speed-up and prediction accuracy. The former is reported relative to both CNN training on high-fidelity images only and Monte Carlo solution of the PDEs. The latter is expressed in terms of both the Wasserstein distance and the Kullback-Leibler divergence.","{'model': 'tldr@v2.0.0', 'text': 'Multi-fidelity simulations are relied on to reduce the cost of data generation for subsequent training of a deep convolutional NN (CNN) using transfer learning and demonstrate the performance of this multi-f fidelity training strategy on the problem of estimation of the distribution of a quantity of interest.'}",https://www.dl.begellhouse.com/download/article/1171f2446ae04ad7/JMLMC0301(2)-38925.pdf
-qbox: partial transfer learning with active querying for object detection,Ying-Peng Tang,"Object detection requires plentiful data annotated with bounding boxes for model training. However, in many applications, it is difficult or even impossible to acquire a large set of labeled examples for the target task due to the privacy concern or lack of reliable annotators. On the other hand, due to the high-quality image search engines, such as Flickr and Google, it is relatively easy to obtain resource-rich unlabeled datasets, whose categories are a superset of those of target data. In this article, to improve the target model with cost-effective supervision from source data, we propose a partial transfer learning approach QBox to actively query labels for bounding boxes of source images. Specifically, we design two criteria, i.e., informativeness and transferability, to measure the potential utility of a bounding box for improving the target model. Based on these criteria, QBox actively queries the labels of the most useful boxes from the source domain and, thus, requires fewer training examples to save the labeling cost. Furthermore, the proposed query strategy allows annotators to simply labeling a specific region, instead of the whole image, and, thus, significantly reduces the labeling difficulty. Extensive experiments are performed on various partial transfer benchmarks and a real COVID-19 detection task. The results validate that QBox improves the detection accuracy with lower labeling cost compared to state-of-the-art query strategies for object detection.","{'model': 'tldr@v2.0.0', 'text': 'This article designs two criteria, i.e., informativeness and transferability, to measure the potential utility of a bounding box for improving the target model, and proposes a partial transfer learning approach QBox to actively query labels for bounding boxes of source images.'}",
-a deep transfer learning approach to modeling teacher discourse in the classroom,Emily Jensen,"Teachers, like everyone else, need objective reliable feedback in order to improve their effectiveness. However, developing a system for automated teacher feedback entails many decisions regarding data collection procedures, automated analysis, and presentation of feedback for reflection. We address the latter two questions by comparing two different machine learning approaches to automatically model seven features of teacher discourse (e.g., use of questions, elaborated evaluations). We compared a traditional open-vocabulary approach using n-grams and Random Forest classifiers with a state-of-the-art deep transfer learning approach for natural language processing (BERT). We found a tradeoff between data quantity and accuracy, where deep models had an advantage on larger datasets, but not for smaller datasets, particularly for variables with low incidence rates. We also compared the models based on the level of feedback granularity: utterance-level (e.g., whether an utterance is a question or a statement), class session-level proportions by averaging across utterances (e.g., question incidence score of 48%), and session-level ordinal feedback based on pre-determined thresholds (e.g., question asking score is medium [vs. low or high]) and found that BERT generally provided more accurate feedback at all levels of granularity. Thus, BERT appears to be the most viable approach to providing automatic feedback on teacher discourse provided there is sufficient data to fine tune the model.","{'model': 'tldr@v2.0.0', 'text': 'A traditional open-vocabulary approach using n-grams and Random Forest classifiers with a state-of-the-art deep transfer learning approach for natural language processing (BERT) is compared and it is found that BERT generally provided more accurate feedback at all levels of granularity.'}",https://dl.acm.org/doi/pdf/10.1145/3448139.3448168
-discriminative fisher embedding dictionary transfer learning for object recognition,Zizhu Fan,"In transfer learning model, the source domain samples and target domain samples usually share the same class labels but have different distributions. In general, the existing transfer learning algorithms ignore the interclass differences and intraclass similarities across domains. To address these problems, this article proposes a transfer learning algorithm based on discriminative Fisher embedding and adaptive maximum mean discrepancy (AMMD) constraints, called discriminative Fisher embedding dictionary transfer learning (DFEDTL). First, combining the label information of source domain and part of target domain, we construct the discriminative Fisher embedding model to preserve the interclass differences and intraclass similarities of training samples in transfer learning. Second, an AMMD model is constructed using atoms and profiles, which can adaptively minimize the distribution differences between source domain and target domain. The proposed method has three advantages: 1) using the Fisher criterion, we construct the discriminative Fisher embedding model between source domain samples and target domain samples, which encourages the samples from the same class to have similar coding coefficients; 2) instead of using the training samples to design the maximum mean discrepancy (MMD), we construct the AMMD model based on the relationship between the dictionary atoms and profiles; thus, the source domain samples can be adaptive to the target domain samples; and 3) the dictionary learning is based on the combination of source and target samples which can avoid the classification error caused by the difference among samples and reduce the tedious and expensive data annotation. A large number of experiments on five public image classification datasets show that the proposed method obtains better classification performance than some state-of-the-art dictionary and transfer learning methods. The code has been available at https://github.com/shilinrui/DFEDTL.","{'model': 'tldr@v2.0.0', 'text': 'A large number of experiments on five public image classification datasets show that the proposed method obtains better classification performance than some state-of-the-art dictionary and transfer learning methods.'}",
-"""train one, classify one, teach one"" - cross-surgery transfer learning for surgical step recognition",Daniel Neimark,"Prior work demonstrated the ability of machine learning to automatically recognize surgical workflow steps from videos. However, these studies focused on only a single type of procedure. In this work, we analyze, for the first time, surgical step recognition on four different laparoscopic surgeries: Cholecystectomy, Right Hemicolectomy, Sleeve Gastrectomy, and Appendectomy. Inspired by the traditional apprenticeship model, in which surgical training is based on the Halstedian method, we paraphrase the""see one, do one, teach one""approach for the surgical intelligence domain as""train one, classify one, teach one"". In machine learning, this approach is often referred to as transfer learning. To analyze the impact of transfer learning across different laparoscopic procedures, we explore various time-series architectures and examine their performance on each target domain. We introduce a new architecture, the Time-Series Adaptation Network (TSAN), an architecture optimized for transfer learning of surgical step recognition, and we show how TSAN can be pre-trained using self-supervised learning on a Sequence Sorting task. Such pre-training enables TSAN to learn workflow steps of a new laparoscopic procedure type from only a small number of labeled samples from the target procedure. Our proposed architecture leads to better performance compared to other possible architectures, reaching over 90% accuracy when transferring from laparoscopic Cholecystectomy to the other three procedure types.","{'model': 'tldr@v2.0.0', 'text': 'This work analyzes, for the first time, surgical step recognition on four different laparoscopic surgeries: Cholecystectomy, Right Hemicolectomy, Sleeve Gastrectomy, and Appendectomy, and introduces a new architecture, the Time-Series Adaptation Network (TSAN), an architecture optimized for transfer learning of surgicalStep recognition.'}",
-deep learning of brain asymmetry images and transfer learning for early diagnosis of dementia,Nitsa J. Herzog,,"{'model': 'tldr@v2.0.0', 'text': 'Results from 300 independent simulation runs illustrate that transfer learning of CNN-based models equipped with SVM output layer is capable to produce better performing models within a few training epochs compared to commonly used transfer learning architectures that combine CNN pretrained models with fully connected Softmax layers.'}",https://eprints.bbk.ac.uk/id/eprint/44156/1/EANN21-NH-GM-camvers.pdf
-nlpbench: evaluating large language models on solving nlp problems,Linxin Song,"Recent developments in large language models (LLMs) have shown promise in enhancing the capabilities of natural language processing (NLP). Despite these successes, there remains a dearth of research dedicated to the NLP problem-solving abilities of LLMs. To fill the gap in this area, we present a unique benchmarking dataset, NLPBench, comprising 378 college-level NLP questions spanning various NLP topics sourced from Yale University's prior final exams. NLPBench includes questions with context, in which multiple sub-questions share the same public information, and diverse question types, including multiple choice, short answer, and math. Our evaluation, centered on LLMs such as GPT-3.5/4, PaLM-2, and LLAMA-2, incorporates advanced prompting strategies like the chain-of-thought (CoT) and tree-of-thought (ToT). Our study reveals that the effectiveness of the advanced prompting strategies can be inconsistent, occasionally damaging LLM performance, especially in smaller models like the LLAMA-2 (13b). Furthermore, our manual assessment illuminated specific shortcomings in LLMs' scientific problem-solving skills, with weaknesses in logical decomposition and reasoning notably affecting results.","{'model': 'tldr@v2.0.0', 'text': ""This study presents a unique benchmarking dataset, NLPBench, comprising 378 college-level NLP questions spanning various NLP topics sourced from Yale University's prior final exams, and reveals that the effectiveness of the advanced prompting strategies can be inconsistent, occasionally damaging LLM performance.""}",https://arxiv.org/pdf/2309.15630
-retuyt-inco at bea 2023 shared task: tuning open-source llms for generating teacher responses,Alexis Baladn,"This paper presents the results of our participation in the BEA 2023 shared task, which focuses on generating AI teacher responses in educational dialogues. We conducted experiments using several Open-Source Large Language Models (LLMs) and explored fine-tuning techniques along with prompting strategies, including Few-Shot and Chain-of-Thought approaches. Our best model was ranked 4.5 in the competition with a BertScore F1 of 0.71 and a DialogRPT final (avg) of 0.35. Nevertheless, our internal results did not exactly correlate with those obtained in the competition, which showed the difficulty in evaluating this task. Other challenges we faced were data leakage on the train set and the irregular format of the conversations.",,https://aclanthology.org/2023.bea-1.61.pdf
-aligning large language models for clinical tasks,Supun Manathunga,"Large Language Models (LLMs) have demonstrated remarkable adaptability, showcasing their capacity to excel in tasks for which they were not explicitly trained. However, despite their impressive natural language processing (NLP) capabilities, effective alignment of LLMs remains a crucial challenge when deploying them for specific clinical applications. The ability to generate responses with factually accurate content and to engage in non-trivial reasoning steps are crucial for the LLMs to be eligible for applications in clinical medicine. Employing a combination of techniques including instruction-tuning and in-prompt strategies like few-shot and chain-of-thought prompting has significantly enhanced the performance of LLMs. Our proposed alignment strategy for medical question-answering, known as 'expand-guess-refine', offers a parameter and data-efficient solution. A preliminary analysis of this method demonstrated outstanding performance, achieving a score of 70.63% on a subset of questions sourced from the USMLE dataset.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes a proposed alignment strategy for medical question-answering, known as 'expand-guess-refine', which offers a parameter and data-efficient solution.""}",https://arxiv.org/pdf/2309.02884
-naisteacher: a prompt and rerank approach to generating teacher utterances in educational dialogues,Justin Vasselli,"This paper presents our approach to the BEA 2023 shared task of generating teacher responses in educational dialogues, using the Teacher-Student Chatroom Corpus. Our system prompts GPT-3.5-turbo to generate initial suggestions, which are then subjected to reranking. We explore multiple strategies for candidate generation, including prompting for multiple candidates and employing iterative few-shot prompts with negative examples. We aggregate all candidate responses and rerank them based on DialogRPT scores. To handle consecutive turns in the dialogue data, we divide the task of generating teacher utterances into two components: teacher replies to the student and teacher continuations of previously sent messages. Through our proposed methodology, our system achieved the top score on both automated metrics and human evaluation, surpassing the reference human teachers on the latter.",,https://aclanthology.org/2023.bea-1.63.pdf
-challenges and strategies in cross-cultural nlp,Daniel Hershcovich,"Various efforts in the Natural Language Processing (NLP) community have been made to accommodate linguistic diversity and serve speakers of many different languages. However, it is important to acknowledge that speakers and the content they produce and require, vary not just by language, but also by culture. Although language and culture are tightly linked, there are important differences. Analogous to cross-lingual and multilingual NLP, cross-cultural and multicultural NLP considers these differences in order to better serve users of NLP systems. We propose a principled framework to frame these efforts, and survey existing and potential strategies.","{'model': 'tldr@v2.0.0', 'text': 'A principled framework to frame efforts to accommodate linguistic diversity and serve speakers of many different languages is proposed, and existing and potential strategies are surveyed.'}",http://arxiv.org/pdf/2203.10020
-an empirical study of tokenization strategies for various korean nlp tasks,Kyubyong Park,"Typically, tokenization is the very first step in most text processing works. As a token serves as an atomic unit that embeds the contextual information of text, how to define a token plays a decisive role in the performance of a model. Even though Byte Pair Encoding (BPE) has been considered the de facto standard tokenization method due to its simplicity and universality, it still remains unclear whether BPE works best across all languages and tasks. In this paper, we test several tokenization strategies in order to answer our primary research question, that is, “What is the best tokenization strategy for Korean NLP tasks?” Experimental results demonstrate that a hybrid approach of morphological segmentation followed by BPE works best in Korean to/from English machine translation and natural language understanding tasks such as KorNLI, KorSTS, NSMC, and PAWS-X. As an exception, for KorQuAD, the Korean extension of SQuAD, BPE segmentation turns out to be the most effective. Our code and pre-trained models are publicly available at https://github.com/kakaobrain/kortok.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that a hybrid approach of morphological segmentation followed by BPE works best in Korean to/from English machine translation and natural language understanding tasks such as KorNLI, KorSTS, NSMC, and PAWS-X.'}",
-evaluating adaptive pedagogical agents' prompting strategies effect on students' emotions,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Examination of the impact of different prompting strategies on self-reported agent-directed emotions in an ITS that scaffolds students’ use of self-regulated learning (SRL) strategies, taking into account students' prior knowledge indicates that more intense initial prompting can indeed lead to increased frustration, and sometimes boredom even toward pedagogical agents that are perceived as competent.""}",https://hal.archives-ouvertes.fr/hal-02015693/file/Bouchet%20et%20al.%20-%202018%20-%20Evaluating%20adaptive%20pedagogical%20agents%E2%80%99%20prompting%20.pdf
-what do governments plan in the field of artificial intelligence?: analysing national ai strategies using nlp,T. Papadopoulos,"The primary goal of this paper is to explore how Natural Language Processing techniques (NLP) can assist in reviewing, understanding, and drawing conclusions from text datasets. We explore NLP techniques for the analysis and the extraction of useful information from the text of twelve national strategies on artificial intelligence (AI). For this purpose, we are using a set of machine learning algorithms in order to (a) extract the most significant keywords and summarize each strategy document, (b) discover and assign topics to each document, and (c) cluster the strategies based on their pair-wise similarity. Using the results of the analysis, we discuss the findings and highlight critical issues that emerge from the national strategies for artificial intelligence, such as the importance of the data ecosystem for the development of AI, the increasing considerations about ethical and safety issues, as well as the growing ambition of many countries to lead in the AI race. Utilizing the LDA topic model, we were able to reveal the distributions of thematic sub-topics among the strategic documents. The topic modelling distributions were then used along with other document similarity measures as an input for the clustering of the strategic documents into groups. The results revealed three clusters of countries with a visible differentiation between the strategies of China and Japan on the one hand and the Scandinavian strategies (plus the German and the Luxemburgish) one on the other. The former promote technology and innovation-driven development plans in order to integrate AI with the economy, while the latter share a common view regarding the role of the public sector both as a promoter and investor but also as a user and beneficiary of AI, and give a higher priority to the ethical & safety issues that are connected to the development of AI.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores NLP techniques for the analysis and the extraction of useful information from the text of twelve national strategies on artificial intelligence, using a set of machine learning algorithms to extract the most significant keywords and summarize each strategy document, and discover and assign topics to each document.'}",
-visual prompting via image inpainting,Amir Bar,"How does one adapt a pre-trained visual model to novel downstream tasks without task-specific finetuning or any model modification? Inspired by prompting in NLP, this paper investigates visual prompting: given input-output image example(s) of a new task at test time and a new input image, the goal is to automatically produce the output image, consistent with the given examples. We show that posing this problem as simple image inpainting - literally just filling in a hole in a concatenated visual prompt image - turns out to be surprisingly effective, provided that the inpainting algorithm has been trained on the right data. We train masked auto-encoders on a new dataset that we curated - 88k unlabeled figures from academic papers sources on Arxiv. We apply visual prompting to these pretrained models and demonstrate results on various downstream image-to-image tasks, including foreground segmentation, single object detection, colorization, edge detection, etc.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates visual prompting: given input-output image example(s) of a new task at test time and a new input image, the goal is to automatically produce the output image, consistent with the given examples, and shows that posing this problem as simple image inpainting turns out to be surprisingly effective.'}",http://arxiv.org/pdf/2209.00647
-can adaptive pedagogical agents' prompting strategies improve students' learning and self-regulation?,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Results indicated that an initially more frequent but progressively fading prompting strategy is beneficial to learners' deployment of SRL processes once the scaffolding is faded, and has no negative impact on learners' perception of the system's usefulness.""}",https://hal.archives-ouvertes.fr/hal-01376429/file/Bouchet_et_al._ITS2016.pdf
-“that is a suspicious reaction!”: interpreting logits variation to detect nlp adversarial attacks,E. Mosca,"Adversarial attacks are a major challenge faced by current machine learning research. These purposely crafted inputs fool even the most advanced models, precluding their deployment in safety-critical applications. Extensive research in computer vision has been carried to develop reliable defense strategies. However, the same issue remains less explored in natural language processing. Our work presents a model-agnostic detector of adversarial text examples. The approach identifies patterns in the logits of the target classifier when perturbing the input text. The proposed detector improves the current state-of-the-art performance in recognizing adversarial inputs and exhibits strong generalization capabilities across different NLP models, datasets, and word-level attacks.","{'model': 'tldr@v2.0.0', 'text': 'A model-agnostic detector of adversarial text examples that identifies patterns in the logits of the target classifier when perturbing the input text and exhibits strong generalization capabilities across different NLP models, datasets, and word-level attacks.'}",https://aclanthology.org/2022.acl-long.538.pdf
-comparison of prompting strategies on the acquisition of daily living skills.,Linsey M. Sabielny,"Determining the most effective prompting strategies to be used for individuals with significant intellectual disability can assist in the acquisition of skills, reduction of errors, and avoidance of prompt dependency. However, few studies have directly compared the effects of different prompting strategies to determine which are the most effective. In the present study, physical only and physical plus vocal prompting strategies were compared to determine if one led to more efficient acquisition of two folding tasks than the other. An alternating treatments design was used with two individuals with significant intellectual disability, whereby the prompting strategies were counterbalanced across folding a shirt and folding a pair of pants. Results demonstrated that both strategies were equally effective for one participant, whereas the physical only prompting strategy may have been the more effective strategy for the second participant. The results suggest that pairing vocal instructions with a physical prompting strategy neither inhibits nor assists the acquisition of daily living skills for individuals with significant intellectual disability.","{'model': 'tldr@v2.0.0', 'text': 'The results suggest that pairing vocal instructions with a physical prompting strategy neither inhibits nor assists the acquisition of daily living skills for individuals with significant intellectual disability.'}",
-benchmarking intersectional biases in nlp,John P. Lalor,"There has been a recent wave of work assessing the fairness of machine learning models in general, and more specifically, on natural language processing (NLP) models built using machine learning techniques. While much work has highlighted biases embedded in state-of-the-art language models, and more recent efforts have focused on how to debias, research assessing the fairness and performance of biased/debiased models on downstream prediction tasks has been limited. Moreover, most prior work has emphasized bias along a single dimension such as gender or race. In this work, we benchmark multiple NLP models with regards to their fairness and predictive performance across a variety of NLP tasks. In particular, we assess intersectional bias - fairness across multiple demographic dimensions. The results show that while current debiasing strategies fare well in terms of the fairness-accuracy trade-off (generally preserving predictive power in debiased models), they are unable to effectively alleviate bias in downstream tasks. Furthermore, this bias is often amplified across dimensions (i.e., intersections). We conclude by highlighting possible causes and making recommendations for future NLP debiasing research.","{'model': 'tldr@v2.0.0', 'text': 'The results show that while current debiasing strategies fare well in terms of the fairness-accuracy trade-off (generally preserving predictive power in debiased models), they are unable to effectively alleviate bias in downstream tasks.'}",https://aclanthology.org/2022.naacl-main.263.pdf
-impact of different pedagogical agents' adaptive self-regulated prompting strategies on learning with metatutor,François Bouchet,,"{'model': 'tldr@v2.0.0', 'text': ""Compared learners' perception of PAs' prompts with MetaTutor, a hypermedia adaptive learning environment, with 40 undergraduates randomly assigned to one of three condi- tions, preliminary results indicated that more frequent, but adaptive prompting is an efficient scaffolding strategy, despite negatively im- pacting learners' satisfaction.""}",http://escholarship.mcgill.ca/downloads/9s161b83t
-measure and improve robustness in nlp models: a survey,Xuezhi Wang,"As NLP models achieved state-of-the-art performances over benchmarks and gained wide applications, it has been increasingly important to ensure the safe deployment of these models in the real world, e.g., making sure the models are robust against unseen or challenging scenarios. Despite robustness being an increasingly studied topic, it has been separately explored in applications like vision and NLP, with various definitions, evaluation and mitigation strategies in multiple lines of research. In this paper, we aim to provide a unifying survey of how to define, measure and improve robustness in NLP. We first connect multiple definitions of robustness, then unify various lines of work on identifying robustness failures and evaluating models’ robustness. Correspondingly, we present mitigation strategies that are data-driven, model-driven, and inductive-prior-based, with a more systematic view of how to effectively improve robustness in NLP models. Finally, we conclude by outlining open challenges and future directions to motivate further research in this area.","{'model': 'tldr@v2.0.0', 'text': 'A unifying survey of how to define, measure and improve robustness in NLP is provided, which first connects multiple definitions of robustness, then unify various lines of work on identifying robustness failures and evaluating models’ robustness.'}",https://aclanthology.org/2022.naacl-main.339.pdf
-low-resource ner by data augmentation with prompting,Jian Liu,"Named entity recognition (NER) is a fundamental information extraction task that seeks to identify entity mentions of certain types in text. Despite numerous advances, the existing NER methods rely on extensive supervision for model training, which struggle in a low-resource scenario with limited training data. In this paper, we propose a new data augmentation method for low-resource NER, by eliciting knowledge from BERT with prompting strategies. Particularly, we devise a label-conditioned word replacement strategy that can produce more label-consistent examples by capturing the underlying word-label dependencies, and a prompting with question answering method to generate new training data from unlabeled texts. The experimental results have widely confirmed the effectiveness of our approach. Particularly, in a low-resource scenario with only 150 training sentences, our approach outperforms previous methods without data augmentation by over 40% in F1 and prior best data augmentation methods by over 2.0% in F1. Furthermore, our approach also fits with a zero-shot scenario, yielding promising results without using any human-labeled data for the task.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new data augmentation method for low-resource NER, by eliciting knowledge from BERT with prompting strategies, and devise a label-conditioned word replacement strategy that can produce more label-consistent examples by capturing the underlying word-label dependencies.'}",https://www.ijcai.org/proceedings/2022/0590.pdf
-the dangers of underclaiming: reasons for caution when reporting how nlp systems fail,Sam Bowman,"Researchers in NLP often frame and discuss research results in ways that serve to deemphasize the field’s successes, often in response to the field’s widespread hype. Though well-meaning, this has yielded many misleading or false claims about the limits of our best technology. This is a problem, and it may be more serious than it looks: It harms our credibility in ways that can make it harder to mitigate present-day harms, like those involving biased systems for content moderation or resume screening. It also limits our ability to prepare for the potentially enormous impacts of more distant future advances. This paper urges researchers to be careful about these claims and suggests some research directions and communication strategies that will make it easier to avoid or rebut them.",,https://aclanthology.org/2022.acl-long.516.pdf
-pali-nlp at semeval-2022 task 4: discriminative fine-tuning of transformers for patronizing and condescending language detection,Dou Hu,"Patronizing and condescending language (PCL) has a large harmful impact and is difficult to detect, both for human judges and existing NLP systems. At SemEval-2022 Task 4, we propose a novel Transformer-based model and its ensembles to accurately understand such language context for PCL detection. To facilitate comprehension of the subtle and subjective nature of PCL, two fine-tuning strategies are applied to capture discriminative features from diverse linguistic behaviour and categorical distribution. The system achieves remarkable results on the official ranking, including 1st in Subtask 1 and 5th in Subtask 2. Extensive experiments on the task demonstrate the effectiveness of our system and its strategies.","{'model': 'tldr@v2.0.0', 'text': 'A novel Transformer-based model and its ensembles are proposed to accurately understand such language context for PCL detection and two fine-tuning strategies are applied to facilitate comprehension of the subtle and subjective nature of PCL.'}",http://arxiv.org/pdf/2203.04616
-smart homes for people with alzheimer's disease: adapting prompting strategies to the patient's cognitive profile,J. Lapointe,"Smart home technologies constitute a potential solution to allow people with Alzheimer's disease (AD) to remain in their home. These intelligent houses contain technological devices aiming to provide adapted cognitive assistance (prompts) when needed. However, a literature review of the field revealed a predominant use of verbal prompts with little knowledge about their real effectiveness. To contribute solving this important issue, we propose, in this paper, comprehensive guidelines to help smart homes researchers to maximize the efficiency by adapting the form of prompts to the specific cognitive profiles of patients with AD. First, we identify the main deficits of AD that influence the effectiveness of prompts. Second, we details which prompting strategy to use accordingly. Third, we propose an experimental protocol, based on a well-known test, and a new prompting software, which allows to validate the proposed guidelines. Finally, we present the preliminary results of a first experiment conducted in our lab with participants having mild to moderate AD.","{'model': 'tldr@v2.0.0', 'text': 'This paper identifies the main deficits of AD that influence the effectiveness of prompts and identifies which prompting strategy to use accordingly, and proposes comprehensive guidelines to help smart homes researchers to maximize the efficiency.'}",
-this joke is [mask]: recognizing humor and offense with prompting,Junze Li,"Humor is a magnetic component in everyday human interactions and communications. Computationally modeling humor enables NLP systems to entertain and engage with users. We investigate the effectiveness of prompting, a new transfer learning paradigm for NLP, for humor recognition. We show that prompting performs similarly to finetuning when numerous annotations are available, but gives stellar performance in low-resource humor recognition. The relationship between humor and offense is also inspected by applying influence functions to prompting; we show that models could rely on offense to determine humor during transfer.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the effectiveness of prompting, a new transfer learning paradigm for NLP, for humor recognition, and shows that prompting performs similarly to finetuning when numerous annotations are available, but gives stellar performance in low-resource humor recognition.'}",http://arxiv.org/pdf/2210.13985
-white-box testing of nlp models with mask neuron coverage,Arshdeep Sekhon,"Recent literature has seen growing interest in using black-box strategies like CheckList for testing the behavior of NLP models. Research on white-box testing has developed a number of methods for evaluating how thoroughly the internal behavior of deep models is tested, but they are not applicable to NLP models. We propose a set of white-box testing methods that are customized for transformer-based NLP models. These include Mask Neuron Coverage (MNCOVER) that measures how thoroughly the attention layers in models are exercised during testing. We show that MNCOVER can refine testing suites generated by CheckList by substantially reduce them in size, for more than 60\% on average, while retaining failing tests -- thereby concentrating the fault detection power of the test suite. Further we show how MNCOVER can be used to guide CheckList input generation, evaluate alternative NLP testing methods, and drive data augmentation to improve accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A set of white-box testing methods that are customized for transformer-based NLP models, including Mask Neuron Coverage (MNCOVER) that measures how thoroughly the attention layers in models are exercised during testing are proposed.'}",http://arxiv.org/pdf/2205.05050
-effectiveness of training of neuro-linguistic programming (nlp) strategies on reducing social anxiety,Zahra Sylabkhori,"The paper was aimed at exploring the efficacy of Neuro-linguistic programming (NLP) strategies on social anxiety people referring to counseling centers. To do this, the number of 30 people with social anxiety disorder who had referred for counseling was selected as the sample, and completed Liebowitz Social Anxiety Scale (1987). The pretest having been administered, 30 people were randomly placed in experimental and control groups respectively. The experimental group received ten sessions of training of neuro-linguistic programming strategies. The interventions were a combination of NLP tools, including the art of self-assertiveness, WACO system, Neuro-linguistic logical level, goal-setting, Switch system, chart sections, rapport, Metal model, encore and modeling. After the interventions were ended, the subjects of both groups were tested with Liebowitz Test twice. The data obtained were examined by using SPSS software ver.17 and statistical independent T Test. Research findings illustrated a significant difference between the experimental and control groups as regards posttest social anxiety scores (P = -0/001). Also, with regards to performance anxiety, a significant difference was seen between the two groups in the posttest (P = -0/001).",,
-on-the-job training of special education staff: teaching the simultaneous prompting strategies.,Sezgin Vuran,"AbstractThe aim of this study is to analyze the effectiveness of stafftraining on discrete-trial teaching (DTT). Multiple baseline design across subjects was used in order to analyze the effect of the training program on the educators' performance on probing and intervention implementation. For teaching these two skills, presentation of an information manual, live model and error correction including feedback giving through video were used. The results showed that the percentage of correct response related to probing and training skills through simultaneous prompting was 100% among all participants. Students who were instructed by these educators also reached 80-100% correct responding level in terms of the skill taught. Follow-up data was collected 4-8 weeks after the completion of the process and it was seen that the participants partially maintained the skills acquired. Social validity data was collected in order to assess opinions of the participants about the survey.Key WordsStaffTraining, Simultaneous Promptings, Discrete-Trial Teaching, Individuals with Developmental Disabilities.The success of individuals with developmental disabilities depends on the effective and proper use of appropriate teaching methods as well as the changes and adaptations of the attitudes of the people who play active roles in the child's education, such as peers, staffand parents. Staffmembers are individuals who interact with children with developmental disabilities and provide services to them (Sturmey, 2008). Considering this relationship, the effectiveness and significance of training these staffmembers have become an important current issue.StaffTraining: A Review of the LiteratureIn the existing literature, there are many studies on the effects of staffmember training on a staff's performance and the performances of the individuals with developmental disabilities to whom the staffprovides service. Most of the studies aim to teach discrete-trial teaching, an evidence-based practice, to staffmembers working with children who suffer from developmental disabilities (Belfiore, Fritts, & Herman, 2008; Dib & Sturmey, 2007; Fazzio, Martin, Arnal, & Yu, 2009; Koegel, Russo, & Rincover, 1977; LeBlanc, Ricciardi, & Luiselli, 2005; Ryan & Hemmes, 2005; Sarokoff& Sturmey, 2004, 2008; Thiessen et al., 2009). The findings of these studies indicate that staffmembers, who have undergone this training, enjoy a distinct increase in the accurate usage of this method. In another studies on stafftraining, teaching through daily routines (Lavie & Sturmey, 2002), and the evaluation of stimulus preferences (Lavie & Sturmey), have been conducted to gain information and skills to staffmember.In stafftraining, there are four main methods: the handbook, which presents related information through a trainer's written notes; modeling, which trains the staffvia live performances or videos; rehearsal, which encourages trainees to apply their skills; and feedback, which is the trainer's explanations of the staff's performance through written, oral or graphic accounts (Sturmey, 2008).In most of the studies investigating the effectiveness of stafftraining methods, at least two of these methods were used together (Dib & Sturmey, 2007; Koegel et al., 1977; Lavie & Sturmey, 2002; Leblanc et al., 2005; Ryan & Hemmes, 2005; Sarokoff& Sturmey, 2004, 2008; Schepis, Reid, Ownbey, & Parsons, 2001), yet in only one study was the effectiveness of self video modeling (in which the participants watch themselves) examined (Belfiore et al., 2008). The teaching methods in stafftraining play a significant role in creating the desired effects on the staffand the individuals taught by the staff. In addition, these methods are expected to be efficient in terms of time and effort. These features are involved in the desired characteristics of stafftraining (Sturmey, 2008).In Turkey, the discrete-trial approach was applied in various studies. …",,
-demonstrate-search-predict: composing retrieval and language models for knowledge-intensive nlp,O. Khattab,"Retrieval-augmented in-context learning has emerged as a powerful approach for addressing knowledge-intensive tasks using frozen language models (LM) and retrieval models (RM). Existing work has combined these in simple""retrieve-then-read""pipelines in which the RM retrieves passages that are inserted into the LM prompt. To begin to fully realize the potential of frozen LMs and RMs, we propose Demonstrate-Search-Predict (DSP), a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM. DSP can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions, systematically breaking down problems into small transformations that the LM and RM can handle more reliably. We have written novel DSP programs for answering questions in open-domain, multi-hop, and conversational settings, establishing in early evaluations new state-of-the-art in-context learning results and delivering 37-120%, 8-39%, and 80-290% relative gains against the vanilla LM (GPT-3.5), a standard retrieve-then-read pipeline, and a contemporaneous self-ask pipeline, respectively. We release DSP at https://github.com/stanfordnlp/dsp","{'model': 'tldr@v2.0.0', 'text': 'Demonstrate-Search-Predict (DSP) is proposed, a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM and can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions.'}",http://arxiv.org/pdf/2212.14024
-error analysis prompting enables human-like translation evaluation in large language models: a case study on chatgpt,Qingyu Lu,"Generative large language models (LLMs), e.g., ChatGPT, have demonstrated remarkable proficiency across several NLP tasks, such as machine translation, text summarization. Recent research (Kocmi and Federmann, 2023) has shown that utilizing ChatGPT for assessing the quality of machine translation (MT) achieves state-of-the-art performance at the system level but performs poorly at the segment level. To further improve the performance of LLMs on MT quality assessment, we conduct an investigation into several prompting methods, and propose a new prompting method called Error Analysis Prompting (EAPrompt) by combining Chain-of-Thoughts (Wei et al., 2022) and Error Analysis (Lu et al., 2022). Our results on WMT22 indicate that prompting LLMs like ChatGPT with error analysis can generate human-like MT evaluations at both the system and segment level. Additionally, we first discover some limitations of ChatGPT as an MT evaluator, such as changing the order of input may significantly influence the judgment when providing multiple translations in a single query. This work provides a preliminary experience of prompting LLMs as an evaluator to improve the reliability of translation evaluation metrics under the error analysis paradigm.","{'model': 'tldr@v2.0.0', 'text': 'The results on WMT22 indicate that prompting LLMs like ChatGPT with error analysis can generate human-like MT evaluations at both the system and segment level, and a new prompting method called EAPrompt is proposed by combining Chain-of-Thoughts and Error Analysis.'}",https://arxiv.org/pdf/2303.13809
-explicit visual prompting for low-level structure segmentations,Weihuang Liu,"We consider the generic problem of detecting low-level structures in images, which includes segmenting the manipulated parts, identifying out-of-focus pixels, separating shadow regions, and detecting concealed objects. Whereas each such topic has been typically addressed with a domain-specific solution, we show that a unified approach performs well across all of them. We take inspiration from the widely-used pre-training and then prompt tuning protocols in NLP and propose a new visual prompting model, named Explicit Visual Prompting (EVP). Different from the previous visual prompting which is typically a dataset-level implicit embedding, our key insight is to enforce the tunable parameters focusing on the explicit visual content from each individual image, i.e., the features from frozen patch embeddings and the input's high-frequency components. The proposed EVP significantly outperforms other parameter-efficient tuning protocols under the same amount of tunable parameters (5.7% extra trainable parameters of each task). EVP also achieves state-of-the-art performances on diverse low-level structure segmentation tasks compared to task-specific solutions. Our code is available at: https://github.com/NiFangBaAGe/Explicit-Visual-Prompt.","{'model': 'tldr@v2.0.0', 'text': 'This work takes inspiration from the widely-used pre-training and then prompt tuning protocols in NLP and proposes a new visual prompting model, named Explicit Visual Prompting (EVP), which significantly outperforms other parameter-efficient tuning protocols under the same amount of tunable parameters.'}",https://arxiv.org/pdf/2303.10883
-nlp-based prediction of medical specialties at hospital admission using triage notes,Émilien Arnaud,"Data Analytics is rapidly expanding within the healthcare domain to help develop strategies for improving the quality of care and curbing costs as well. Natural Language Processing (NLP) solutions have received particular attention whereas a large part of clinical data is stockpiled into unstructured physician or nursing notes. In this respect, we attempt to employ NLP to provide an early prediction of the medical specialties at hospital admission. The study uses a large-scale dataset including more than 260K ED records provided by the Amiens-Picardy University Hospital in France. Our approach aims to integrate structured data with unstructured textual notes recorded at the triage stage. On one hand, a standard MLP model is used against the typical set of features. On the other hand, a Convolutional Neural Network is used to operate over the textual data. While both learning components are conducted independently in parallel. The empirical results demonstrated a promising accuracy in general. It is conceived that the study could be an additional contribution to the mounting efforts of applying NLP methods in the healthcare domain.","{'model': 'tldr@v2.0.0', 'text': 'This work attempts to employ NLP to provide an early prediction of the medical specialties at hospital admission using a large-scale dataset including more than 260K ED records provided by the Amiens-Picardy University Hospital in France.'}",
-pushing the limits of chatgpt on nlp tasks,Xiaofei Sun,"Despite the success of ChatGPT, its performances on most NLP tasks are still well below the supervised baselines. In this work, we looked into the causes, and discovered that its subpar performance was caused by the following factors: (1) token limit in the prompt does not allow for the full utilization of the supervised datasets; (2) mismatch between the generation nature of ChatGPT and NLP tasks; (3) intrinsic pitfalls of LLMs models, e.g., hallucination, overly focus on certain keywords, etc. In this work, we propose a collection of general modules to address these issues, in an attempt to push the limits of ChatGPT on NLP tasks. Our proposed modules include (1) a one-input-multiple-prompts strategy that employs multiple prompts for one input to accommodate more demonstrations; (2) using fine-tuned models for better demonstration retrieval; (3) transforming tasks to formats that are more tailored to the generation nature; (4) employing reasoning strategies that are tailored to addressing the task-specific complexity; (5) the self-verification strategy to address the hallucination issue of LLMs; (6) the paraphrase strategy to improve the robustness of model predictions. We conduct experiments on 21 datasets of 10 representative NLP tasks, including question answering, commonsense reasoning, natural language inference, sentiment analysis, named entity recognition, entity-relation extraction, event extraction, dependency parsing, semantic role labeling, and part-of-speech tagging. Using the proposed assemble of techniques, we are able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.","{'model': 'tldr@v2.0.0', 'text': 'Using the proposed assemble of techniques, this work is able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances.'}",https://arxiv.org/pdf/2306.09719
-all in one: multi-task prompting for graph neural networks,Xiangguo Sun,"Recently, ""pre-training and fine-tuning'' has been adopted as a standard workflow for many graph tasks since it can take general graph knowledge to relieve the lack of graph annotations from each application. However, graph tasks with node level, edge level, and graph level are far diversified, making the pre-training pretext often incompatible with these multiple tasks. This gap may even cause a ""negative transfer'' to the specific application, leading to poor results. Inspired by the prompt learning in natural language processing (NLP), which has presented significant effectiveness in leveraging prior knowledge for various NLP tasks, we study the prompting topic for graphs with the motivation of filling the gap between pre-trained models and various graph tasks. In this paper, we propose a novel multi-task prompting method for graph models. Specifically, we first unify the format of graph prompts and language prompts with the prompt token, token structure, and inserting pattern. In this way, the prompting idea from NLP can be seamlessly introduced to the graph area. Then, to further narrow the gap between various graph tasks and state-of-the-art pre-training strategies, we further study the task space of various graph applications and reformulate downstream problems to the graph-level task. Afterward, we introduce meta-learning to efficiently learn a better initialization for the multi-task prompt of graphs so that our prompting framework can be more reliable and general for different tasks. We conduct extensive experiments, results from which demonstrate the superiority of our method.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel multi-task prompting method for graph models that unify the format of graph prompts and language prompts with the prompt token, token structure, and inserting pattern, and introduces meta-learning to efficiently learn a better initialization for the multi- task prompt of graphs so that the prompting framework can be more reliable and general for different tasks.'}",https://arxiv.org/pdf/2307.01504
-damo-nlp at semeval-2023 task 2: a unified retrieval-augmented system for multilingual named entity recognition,Zeqi Tan,"The MultiCoNER II shared task aims to tackle multilingual named entity recognition (NER) in fine-grained and noisy scenarios, and it inherits the semantic ambiguity and low-context setting of the MultiCoNER I task. To cope with these problems, the previous top systems in the MultiCoNER I either incorporate the knowledge bases or gazetteers. However, they still suffer from insufficient knowledge, limited context length, single retrieval strategy. In this paper, our team DAMO-NLP proposes a unified retrieval-augmented system (U-RaNER) for fine-grained multilingual NER. We perform error analysis on the previous top systems and reveal that their performance bottleneck lies in insufficient knowledge. Also, we discover that the limited context length causes the retrieval knowledge to be invisible to the model. To enhance the retrieval context, we incorporate the entity-centric Wikidata knowledge base, while utilizing the infusion approach to broaden the contextual scope of the model. Also, we explore various search strategies and refine the quality of retrieval knowledge. Our system wins 9 out of 13 tracks in the MultiCoNER II shared task. Additionally, we compared our system with ChatGPT, one of the large language models which have unlocked strong capabilities on many tasks. The results show that there is still much room for improvement for ChatGPT on the extraction task.","{'model': 'tldr@v2.0.0', 'text': 'A unified retrieval-augmented system (U-RaNER) for fine-grained multilingual NER, which incorporates the entity-centric Wikidata knowledge base, while utilizing the infusion approach to broaden the contextual scope of the model.'}",http://arxiv.org/pdf/2305.03688
-diffusion-nat: self-prompting discrete diffusion for non-autoregressive text generation,Kun Zhou,"Recently, continuous diffusion models (CDM) have been introduced into non-autoregressive (NAR) text-to-text generation. However, the discrete nature of text increases the difficulty of CDM to generate coherent and fluent texts, and also causes the incompatibility problem between CDM and advanced NLP techniques, especially the popular pre-trained language models~(PLMs). To solve it, we propose Diffusion-NAT, which introduces discrete diffusion models~(DDM) into NAR text-to-text generation and integrates BART to improve the performance. By revising the decoding process of BART and the typical settings of DDM, we unify the inference process of BART and the denoising process of DDM into the same NAR masked tokens recovering task. In this way, DDM can rely on BART to perform denoising, which can benefit from both the rich pre-learned knowledge of BART and the iterative refining paradigm of DDM. Besides, we also propose the iterative self-prompting strategy to further improve the generation quality. Experimental results on 7 datasets show that our approach can outperform competitive NAR methods, and even surpass autoregressive methods. Our code and data will be publicly released.","{'model': 'tldr@v2.0.0', 'text': 'Diffusion-NAT is proposed, which introduces discrete diffusion models~(DDM) into NAR text-to-text generation and integrates BART to improve the performance and proposes the iterative self-prompting strategy to further improve the generation quality.'}",http://arxiv.org/pdf/2305.04044
-sustainable marketing and the role of social media: an experimental study using natural language processing (nlp),Ganesh Dash,"Marketing has changed fundamentally in the new millennium. At the same time, sustainable marketing strategies have evolved to meet the challenges of environmental issues. In this study, we examined the trends in sustainable marketing strategies and the role of social media in these. Based on specific keywords per the objective, this study collected 33 published articles from the Scopus database from 1991 to 2022 (2012–2022). The KNIME (Konstanz Information Miner) and VOSviewer tools were deployed to provide detailed classification and prediction of the various trends in sustainable marketing, with a particular focus on the role of social media. The study method applied text mining and latent semantic analysis to predict the latest trends. The top three trends were Green Marketing and Consumer Behavior, Sustainable Social Media Marketing, and Influencer Social Media Marketing Practices. This NLP-based review and the clustering of research directions provide immense value to marketers and policymakers.",,https://www.mdpi.com/2071-1050/15/6/5443/pdf?version=1679297784
-matsci-nlp: evaluating scientific language models on materials science language tasks using text-to-schema modeling,Yurun Song,"We present MatSci-NLP, a natural language benchmark for evaluating the performance of natural language processing (NLP) models on materials science text. We construct the benchmark from publicly available materials science text data to encompass seven different NLP tasks, including conventional NLP tasks like named entity recognition and relation classification, as well as NLP tasks specific to materials science, such as synthesis action retrieval which relates to creating synthesis procedures for materials. We study various BERT-based models pretrained on different scientific text corpora on MatSci-NLP to understand the impact of pretraining strategies on understanding materials science text. Given the scarcity of high-quality annotated data in the materials science domain, we perform our fine-tuning experiments with limited training data to encourage the generalize across MatSci-NLP tasks.Our experiments in this low-resource training setting show that language models pretrained on scientific text outperform BERT trained on general text. MatBERT, a model pretrained specifically on materials science journals, generally performs best for most tasks. Moreover, we propose a unified text-to-schema for multitask learning on {pasted macro ‘BENCHMARK’} and compare its performance with traditional fine-tuning methods. In our analysis of different training methods, we find that our proposed text-to-schema methods inspired by question-answering consistently outperform single and multitask NLP fine-tuning methods. The code and datasets are publicly available https://github.com/BangLab-UdeM-Mila/NLP4MatSci-ACL23.","{'model': 'tldr@v2.0.0', 'text': 'In the analysis of different training methods, it is found that the proposed text-to-schema methods inspired by question-answering consistently outperform single and multitask NLP fine-tuning methods.'}",http://arxiv.org/pdf/2305.08264
-video prompting versus other instruction strategies for persons with alzheimer’s disease,Viviana Perilli,"Background/Aim: Two studies assessed the effectiveness of video prompting as a strategy to support persons with mild and moderate Alzheimer’s disease in performing daily activities. Methods: In study I, video prompting was compared to an existing strategy relying on verbal instructions. In study II, video prompting was compared to another existing strategy relying on static pictorial cues. Video prompting and the other strategies were counterbalanced across tasks and participants and compared within alternating treatments designs. Results: Video prompting was effective in all participants. Similarly effective were the other 2 strategies, and only occasional differences between the strategies were reported. Two social validation assessments showed that university psychology students and graduates rated the patients’ performance with video prompting more favorably than their performance with the other strategies. Conclusion: Video prompting may be considered a valuable alternative to the other strategies to support daily activities in persons with Alzheimer's disease.","{'model': 'tldr@v2.0.0', 'text': ""Video prompting may be considered a valuable alternative to the other strategies to support daily activities in persons with Alzheimer's disease.""}",
-parafuzz: an interpretability-driven technique for detecting poisoned samples in nlp,Lu Yan,"Backdoor attacks have emerged as a prominent threat to natural language processing (NLP) models, where the presence of specific triggers in the input can lead poisoned models to misclassify these inputs to predetermined target classes. Current detection mechanisms are limited by their inability to address more covert backdoor strategies, such as style-based attacks. In this work, we propose an innovative test-time poisoned sample detection framework that hinges on the interpretability of model predictions, grounded in the semantic meaning of inputs. We contend that triggers (e.g., infrequent words) are not supposed to fundamentally alter the underlying semantic meanings of poisoned samples as they want to stay stealthy. Based on this observation, we hypothesize that while the model's predictions for paraphrased clean samples should remain stable, predictions for poisoned samples should revert to their true labels upon the mutations applied to triggers during the paraphrasing process. We employ ChatGPT, a state-of-the-art large language model, as our paraphraser and formulate the trigger-removal task as a prompt engineering problem. We adopt fuzzing, a technique commonly used for unearthing software vulnerabilities, to discover optimal paraphrase prompts that can effectively eliminate triggers while concurrently maintaining input semantics. Experiments on 4 types of backdoor attacks, including the subtle style backdoors, and 4 distinct datasets demonstrate that our approach surpasses baseline methods, including STRIP, RAP, and ONION, in precision and recall.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an innovative test-time poisoned sample detection framework that hinges on the interpretability of model predictions, grounded in the semantic meaning of inputs, and adopts fuzzing, a technique commonly used for unearthing software vulnerabilities, to discover optimal paraphrase prompts that can effectively eliminate triggers while concurrently maintaining input semantics.'}",https://arxiv.org/pdf/2308.02122
-self-diagnosis and self-debiasing: a proposal for reducing corpus-based bias in nlp,Timo Schick,"Abstract ⚠ This paper contains prompts and model outputs that are offensive in nature. When trained on large, unfiltered crawls from the Internet, language models pick up and reproduce all kinds of undesirable biases that can be found in the data: They often generate racist, sexist, violent, or otherwise toxic language. As large models require millions of training examples to achieve good performance, it is difficult to completely prevent them from being exposed to such content. In this paper, we first demonstrate a surprising finding: Pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce. We refer to this capability as self-diagnosis. Based on this finding, we then propose a decoding algorithm that, given only a textual description of the undesired behavior, reduces the probability of a language model producing problematic text. We refer to this approach as self-debiasing. Self-debiasing does not rely on manually curated word lists, nor does it require any training data or changes to the model’s parameters. While we by no means eliminate the issue of language models generating biased text, we believe our approach to be an important step in this direction.1","{'model': 'tldr@v2.0.0', 'text': 'This paper demonstrates a surprising finding: Pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce and proposes a decoding algorithm that reduces the probability of a language model producing problematic text, known as self-debiasing.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf
-personalization of news for a logistics organisation by finding relevancy using nlp,R. Garg,,"{'model': 'tldr@v2.0.0', 'text': 'This paper attempts to provide a state of art in finding relevancy in news headlines by using statistical and deep learning models to extract information from the corpora using natural language processing.'}",
-large language models (llms) for natural language processing (nlp) of oil and gas drilling data,Prateek Kumar," In the oil and gas industry, drilling activities spawn substantial volumes of unstructured textual data. The examination and interpretation of these data pose significant challenges. This research exploits the emerging capabilities of large language models (LLMs) with over 100 billion parameters to extract actionable insights from raw drilling data. Through fine-tuning methodologies and the use of various prompt engineering strategies, we addressed several text downstream tasks, including summarization, classification, entity recognition, and information extraction. This study delves into our methods, findings, and the novel application of LLMs for efficient and precise analysis of drilling data.","{'model': 'tldr@v2.0.0', 'text': 'This research exploits the emerging capabilities of large language models with over 100 billion parameters to extract actionable insights from raw drilling data through fine-tuning methodologies and the use of various prompt engineering strategies.'}",
-returning the n to nlp: towards contextually personalized classification models,Lucie Flekova,"Most NLP models today treat language as universal, even though socio- and psycholingustic research shows that the communicated message is influenced by the characteristics of the speaker as well as the target audience. This paper surveys the landscape of personalization in natural language processing and related fields, and offers a path forward to mitigate the decades of deviation of the NLP tools from sociolingustic findings, allowing to flexibly process the “natural” language of each user rather than enforcing a uniform NLP treatment. It outlines a possible direction to incorporate these aspects into neural NLP models by means of socially contextual personalization, and proposes to shift the focus of our evaluation strategies accordingly.","{'model': 'tldr@v2.0.0', 'text': 'The landscape of personalization in natural language processing and related fields is surveyed, and a path forward to mitigate the decades of deviation of the NLP tools from sociolingustic findings is offered, allowing to flexibly process the “natural” language of each user rather than enforcing a uniform NLP treatment.'}",https://www.aclweb.org/anthology/2020.acl-main.700.pdf
-a survey on bias in deep nlp,Ismael Garrido-Muñoz,"Deep neural networks are hegemonic approaches to many machine learning areas, including natural language processing (NLP). Thanks to the availability of large corpora collections and the capability of deep architectures to shape internal language mechanisms in self-supervised learning processes (also known as “pre-training”), versatile and performing models are released continuously for every new network design. These networks, somehow, learn a probability distribution of words and relations across the training collection used, inheriting the potential flaws, inconsistencies and biases contained in such a collection. As pre-trained models have been found to be very useful approaches to transfer learning, dealing with bias has become a relevant issue in this new scenario. We introduce bias in a formal way and explore how it has been treated in several networks, in terms of detection and correction. In addition, available resources are identified and a strategy to deal with bias in deep NLP is proposed.","{'model': 'tldr@v2.0.0', 'text': 'Bias is introduced in a formal way and how it has been treated in several networks, in terms of detection and correction, and a strategy to deal with bias in deep NLP is proposed.'}",
-a decision tree for the use of auditory prompting strategies,Michał Post,"This article provides a decision tree for determining the most effective method of auditory prompting systems in order to increase workplace independence for employees with severe disabilities. A flow-through chart, The Auditory Prompts Decision Tree, is presented to guide this planning process. The Auditory Prompts Decision Tree is useful when analyzing why the breakdown in work performance is occurring and for selecting the types of auditory prompts that will support the employee's work performance needs. Specific examples from research studies are provided to illustrate the use of the Auditory Prompts Decision Tree for each type of auditory prompt.",,
-badpre: task-agnostic backdoor attacks to pre-trained nlp foundation models,Kangjie Chen,"Pre-trained Natural Language Processing (NLP) models can be easily adapted to a variety of downstream language tasks. This significantly accelerates the development of language models. However, NLP models have been shown to be vulnerable to backdoor attacks, where a pre-defined trigger word in the input text causes model misprediction. Previous NLP backdoor attacks mainly focus on some specific tasks. This makes those attacks less general and applicable to other kinds of NLP models and tasks. In this work, we propose \Name, the first task-agnostic backdoor attack against the pre-trained NLP models. The key feature of our attack is that the adversary does not need prior information about the downstream tasks when implanting the backdoor to the pre-trained model. When this malicious model is released, any downstream models transferred from it will also inherit the backdoor, even after the extensive transfer learning process. We further design a simple yet effective strategy to bypass a state-of-the-art defense. Experimental results indicate that our approach can compromise a wide range of downstream NLP tasks in an effective and stealthy way.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results indicate that the proposed \\Name, the first task-agnostic backdoor attack against the pre-trained NLP models, can compromise a wide range of downstream NLP tasks in an effective and stealthy way.'}",
-easytransfer: a simple and scalable deep transfer learning platform for nlp applications,Minghui Qiu,"The literature has witnessed the success of leveraging Pre-trained Language Models (PLMs) and Transfer Learning (TL) algorithms to a wide range of Natural Language Processing (NLP) applications, yet it is not easy to build an easy-to-use and scalable TL toolkit for this purpose. To bridge this gap, the EasyTransfer platform is designed to develop deep TL algorithms for NLP applications. EasyTransfer is backended with a high-performance and scalable engine for efficient training and inference, and also integrates comprehensive deep TL algorithms, to make the development of industrial-scale TL applications easier. In EasyTransfer, the built-in data and model parallelism strategies, combined with AI compiler optimization, show to be 4.0x faster than the community version of distributed training. EasyTransfer supports various NLP models in the ModelZoo, including mainstream PLMs and multi-modality models. It also features various in-house developed TL algorithms, together with the AppZoo for NLP applications. The toolkit is convenient for users to quickly start model training, evaluation, and online deployment. EasyTransfer is currently deployed at Alibaba to support a variety of business scenarios, including item recommendation, personalized search, conversational question answering, etc. Extensive experiments on real-world datasets and online applications show that EasyTransfer is suitable for online production with cutting-edge performance for various applications. The source code of EasyTransfer is released at Github1.","{'model': 'tldr@v2.0.0', 'text': 'The EasyTransfer platform is designed to develop deep TL algorithms for NLP applications, backended with a high-performance and scalable engine for efficient training and inference, and also integrates comprehensive deepTL algorithms, to make the development of industrial-scale TL applications easier.'}",
-integrating ethics into the nlp curriculum,Emily M. Bender,"To raise awareness among future NLP practitioners and prevent inertia in the field, we need to place ethics in the curriculum for all NLP students—not as an elective, but as a core part of their education. Our goal in this tutorial is to empower NLP researchers and practitioners with tools and resources to teach others about how to ethically apply NLP techniques. We will present both high-level strategies for developing an ethics-oriented curriculum, based on experience and best practices, as well as specific sample exercises that can be brought to a classroom. This highly interactive work session will culminate in a shared online resource page that pools lesson plans, assignments, exercise ideas, reading suggestions, and ideas from the attendees. Though the tutorial will focus particularly on examples for university classrooms, we believe these ideas can extend to company-internal workshops or tutorials in a variety of organizations. In this setting, a key lesson is that there is no single approach to ethical NLP: each project requires thoughtful consideration about what steps can be taken to best support people affected by that project. However, we can learn (and teach) what issues to be aware of, what questions to ask, and what strategies are available to mitigate harm.","{'model': 'tldr@v2.0.0', 'text': 'This tutorial is to empower NLP researchers and practitioners with tools and resources to teach others about how to ethically apply NLP techniques, and will present both high-level strategies for developing an ethics-oriented curriculum, based on experience and best practices.'}",https://www.aclweb.org/anthology/2020.acl-tutorials.2.pdf
-a novel hybrid approach of svm combined with nlp and probabilistic neural network for email phishing,Abhishek Kumar,"Phishing attacks are one of the slanting cyber-attacks that apply socially engineered messages that are imparted to individuals from expert hackers going for tricking clients to uncover their delicate data, the most mainstream correspondence channel to those messages is through clients' emails. Phishing has turned into a generous danger for web clients and a noteworthy reason for money related misfortunes. Therefore, different arrangements have been created to handle this issue. Deceitful emails, also called phishing emails, utilize a scope of impact strategies to convince people to react, for example, promising a fiscal reward or summoning a feeling of criticalness. Regardless of far reaching alerts and intends to instruct clients to distinguish phishing sends, these are as yet a pervasive practice and a worthwhile business. The creators accept that influence, as a style of human correspondence intended to impact others, has a focal job in fruitful advanced tricks. Cyber criminals have ceaselessly propelling their techniques for assault. The current strategies to recognize the presence of such malevolent projects and to keep them from executing are static, dynamic and hybrid analysis. In this work we are proposing a hybrid methodology for phishing detection incorporating feature extraction and classification of the mails using SVM. At last, alongside the chose features, the PNN characterizes the spam mails from the genuine mails with more exactness and accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This work is proposing a hybrid methodology for phishing detection incorporating feature extraction and classification of the mails using SVM, and alongside the chose features, the PNN characterizes the spam mails from the genuine mails with more exactness and accuracy.'}",http://ijece.iaescore.com/index.php/IJECE/article/download/20847/13512
-prompting strategies that promote student self-management,P. Alberto,"Students who perform tasks competently and independently tend to exhibit fewer instances of inappropriate behavior than those who have difficulty with them (Smith, 1985). Students with severe handicaps can perfonn more competently when teachers incorporate verbal and/or physical prompts into their instruction. However, competence attained this way tends to decrease when the teacher is not present to deliver the prompts. When this occurs, an alternative to teacher prompting must be implemented to maintain student perfonnance. This is challenging, especially in community settings. One way to reduce dependence on teachers is to teach students to use self-operated prompting strategies. This type of self-management fosters student independence and improves behavior in community environments.",,
-triggerless backdoor attack for nlp tasks with clean labels,Leilei Gan,"Backdoor attacks pose a new threat to NLP models. A standard strategy to construct poisoned data in backdoor attacks is to insert triggers (e.g., rare words) into selected sentences and alter the original label to a target label. This strategy comes with a severe flaw of being easily detected from both the trigger and the label perspectives: the trigger injected, which is usually a rare word, leads to an abnormal natural language expression, and thus can be easily detected by a defense model; the changed target label leads the example to be mistakenly labeled, and thus can be easily detected by manual inspections. To deal with this issue, in this paper, we propose a new strategy to perform textual backdoor attack which does not require an external trigger and the poisoned samples are correctly labeled. The core idea of the proposed strategy is to construct clean-labeled examples, whose labels are correct but can lead to test label changes when fused with the training set. To generate poisoned clean-labeled examples, we propose a sentence generation model based on the genetic algorithm to cater to the non-differentiable characteristic of text data. Extensive experiments demonstrate that the proposed attacking strategy is not only effective, but more importantly, hard to defend due to its triggerless and clean-labeled nature. Our work marks the first step towards developing triggerless attacking strategies in NLP.","{'model': 'tldr@v2.0.0', 'text': 'A new strategy to perform textual backdoor attack which does not require an external trigger and the poisoned samples are correctly labeled is proposed, which marks the first step towards developing triggerless attacking strategies in NLP.'}",https://aclanthology.org/2022.naacl-main.214.pdf
-assessing the acquisition of incidental information by secondary-age students with mental retardation: comparison of response prompting strategies.,D. Gast,"The effectiveness and efficiency of four response prompting conditions (progressive time delay, progressive time delay with a descriptive consequent event, system of least prompts, and system of least prompts with a descriptive consequent event) were compared. Students with moderate to severe mental retardation were taught to read functional recipe words. Maintenance and students' acquisition of incidental information were assessed when it was (a) embedded in the prompts of the system of least prompts procedure, (b) included in the descriptive praise statements following correct performance with the progressive time delay and system of least prompts procedures, and (c) not presented. A multiple probe design across behaviors, replicated across subjects, was used. Results indicated that (a) each of the procedures produced criterion level responding: (b) efficiency data on traditional measures were roughly equal; (c) maintenance checks showed no differential effects related to the instructional condition; and (d) incidental information was acquired, although it was not directly targeted for instruction.","{'model': 'tldr@v2.0.0', 'text': 'Students with moderate to severe mental retardation were taught to read functional recipe words and incidental information was acquired, although it was not directly targeted for instruction.'}",
-encouraging long-term compliance with breast self-examination: the evaluation of prompting strategies,J. Mayer,,"{'model': 'tldr@v2.0.0', 'text': 'Results indicated that the prompts, particularly those delivered by phone, were effective in encouraging initial performance of BSE, however, for all conditions, BSE decreased over the course of the study.'}",
-prompting strategies for those with a severe mental handicap: a comparison of procedures using only response prompts with a procedure combining stimulus and response prompts,G. A. Riley,"Prompts are assessed in terms of their ability to convey an understanding of the task to those with a severe mental handicap. On this basis, it is suggested that the most effective strategy for teaching manipulative tasks to this group may be one which combines stimulus and response prompts. Two case studies compared this procedure with two procedures which used response prompts only. The combined procedure was more effective in both studies. The studies also demonstrated the applicability of stimulus prompts to teaching the motor, as well as the visual, components of manipulative tasks.",,
-different dna-binding specificities of nlp and nin transcription factors underlie nitrate-induced control of root nodulation,Hanna Nishida,"Abstract Leguminous plants produce nodules for nitrogen fixation; however, nodule production incurs an energy cost. Therefore, as an adaptive strategy, leguminous plants halt root nodule development when sufficient amounts of nitrogen nutrients, such as nitrate, are present in the environment. Although legume NODULE INCEPTION (NIN)-LIKE PROTEIN (NLP) transcription factors have recently been identified, understanding how nodulation is controlled by nitrate, a fundamental question for nitrate-mediated transcriptional regulation of symbiotic genes, remains elusive. Here, we show that two Lotus japonicus NLPs, NITRATE UNRESPONSIVE SYMBIOSIS 1 (NRSYM1)/LjNLP4 and NRSYM2/LjNLP1, have overlapping functions in the nitrate-induced control of nodulation and act as master regulators for nitrate-dependent gene expression. We further identify candidate target genes of LjNLP4 by combining transcriptome analysis with a DNA affinity purification-seq approach. We then demonstrate that LjNLP4 and LjNIN, a key nodulation-specific regulator and paralog of LjNLP4, have different DNA-binding specificities. Moreover, LjNLP4–LjNIN dimerization underlies LjNLP4-mediated bifunctional transcriptional regulation. These data provide a basic principle for how nitrate controls nodulation through positive and negative regulation of symbiotic genes.","{'model': 'tldr@v2.0.0', 'text': 'A basic principle for how nitrate controls nodulation through positive and negative regulation of symbiotic genes is provided, which underlies LjNLP4-mediated bifunctional transcriptional regulation.'}",https://academic.oup.com/plcell/article-pdf/33/7/2340/39739430/koab103.pdf
-automatically correcting large language models: surveying the landscape of diverse self-correction strategies,Liangming Pan,"Large language models (LLMs) have demonstrated remarkable performance across a wide array of NLP tasks. However, their efficacy is undermined by undesired and inconsistent behaviors, including hallucination, unfaithful reasoning, and toxic content. A promising approach to rectify these flaws is self-correction, where the LLM itself is prompted or guided to fix problems in its own output. Techniques leveraging automated feedback -- either produced by the LLM itself or some external system -- are of particular interest as they are a promising way to make LLM-based solutions more practical and deployable with minimal human feedback. This paper presents a comprehensive review of this emerging class of techniques. We analyze and taxonomize a wide array of recent work utilizing these strategies, including training-time, generation-time, and post-hoc correction. We also summarize the major applications of this strategy and conclude by discussing future directions and challenges.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive review of techniques leveraging automated feedback to make LLM-based solutions more practical and deployable with minimal human feedback, including training-time, generation- time, and post-hoc correction.'}",https://arxiv.org/pdf/2308.03188
-strategies for bridging the digital divide by enhancing communication competence,N. Pokrovskaia,"The digital divide reflects disparities in social status and well-being that depend on access to physical infrastructure, acquired skills in the use of hardware and software, ability to interact in cyberspace. The pandemic shifted all activities to a remote mode, prompting people to enhance their communicative competence. The survey shows a funnel of online-communication choices depending on the criteria of experience and education; higher the intensity of activity and education, narrower the scope and nature of the content for networking, less time spent online.",,
-playing the lottery with rewards and multiple languages: lottery tickets in rl and nlp,Haonan Yu,"The lottery ticket hypothesis proposes that over-parameterization of deep neural networks (DNNs) aids training by increasing the probability of a ""lucky"" sub-network initialization being present rather than by helping the optimization process (Frankle & Carbin, 2019). Intriguingly, this phenomenon suggests that initialization strategies for DNNs can be improved substantially, but the lottery ticket hypothesis has only previously been tested in the context of supervised learning for natural image tasks. Here, we evaluate whether ""winning ticket"" initializations exist in two different domains: natural language processing (NLP) and reinforcement learning (RL).For NLP, we examined both recurrent LSTM models and large-scale Transformer models (Vaswani et al., 2017). For RL, we analyzed a number of discrete-action space tasks, including both classic control and pixel control. Consistent with workin supervised image classification, we confirm that winning ticket initializations generally outperform parameter-matched random initializations, even at extreme pruning rates for both NLP and RL. Notably, we are able to find winning ticket initializations for Transformers which enable models one-third the size to achieve nearly equivalent performance. Together, these results suggest that the lottery ticket hypothesis is not restricted to supervised learning of natural images, but rather represents a broader phenomenon in DNNs.","{'model': 'tldr@v2.0.0', 'text': 'Evaluating whether ""winning ticket"" initializations exist in NLP and reinforcement learning suggests that the lottery ticket hypothesis is not restricted to supervised learning of natural images, but rather represents a broader phenomenon in DNNs.'}",
-adversarial attacks on large language model-based system and mitigating strategies: a case study on chatgpt,Bowen Liu,"Machine learning algorithms are at the forefront of the development of advanced information systems. The rapid progress in machine learning technology has enabled cutting-edge large language models (LLMs), represented by GPT-3 and ChatGPT, to perform a wide range of NLP tasks with a stunning performance. However, research on adversarial machine learning highlights the need for these intelligent systems to be more robust. Adversarial machine learning aims to evaluate attack and defense mechanisms to prevent the malicious exploitation of these systems. In the case of ChatGPT, adversarial induction prompt can cause the model to generate toxic texts that could pose serious security risks or propagate false information. To address this challenge, we first analyze the effectiveness of inducing attacks on ChatGPT. Then, two effective mitigating mechanisms are proposed. The first is a training-free prefix prompt mechanism to detect and prevent the generation of toxic texts. The second is a RoBERTa-based mechanism that identifies manipulative or misleading input text via external detection models. The availability of this method is demonstrated through experiments.","{'model': 'tldr@v2.0.0', 'text': 'A training-free prefix prompt mechanism to detect and prevent the generation of toxic texts and a RoBERTa-based mechanism that identifies manipulative or misleading input text via external detection models are proposed.'}",https://downloads.hindawi.com/journals/scn/2023/8691095.pdf
-how suitable are subword segmentation strategies for translating non-concatenative morphology?,Chantal Amrhein,"Data-driven subword segmentation has become the default strategy for open-vocabulary machine translation and other NLP tasks, but may not be sufficiently generic for optimal learning of non-concatenative morphology. We design a test suite to evaluate segmentation strategies on different types of morphological phenomena in a controlled, semi-synthetic setting. In our experiments, we compare how well machine translation models trained on subword- and character-level can translate these morphological phenomena. We find that learning to analyse and generate morphologically complex surface representations is still challenging, especially for non-concatenative morphological phenomena like reduplication or vowel harmony and for rare word stems. Based on our results, we recommend that novel text representation strategies be tested on a range of typologically diverse languages to minimise the risk of adopting a strategy that inadvertently disadvantages certain languages.","{'model': 'tldr@v2.0.0', 'text': 'It is found that learning to analyse and generate morphologically complex surface representations is still challenging, especially for nonconcatenative morphological phenomena like reduplication or vowel harmony and for rare word stems.'}",https://aclanthology.org/2021.findings-emnlp.60.pdf
-learning to solve nlp tasks in an incremental number of languages,Giuseppe Castellucci,"In real scenarios, a multilingual model trained to solve NLP tasks on a set of languages can be required to support new languages over time. Unfortunately, the straightforward retraining on a dataset containing annotated examples for all the languages is both expensive and time-consuming, especially when the number of target languages grows. Moreover, the original annotated material may no longer be available due to storage or business constraints. Re-training only with the new language data will inevitably result in Catastrophic Forgetting of previously acquired knowledge. We propose a Continual Learning strategy that updates a model to support new languages over time, while maintaining consistent results on previously learned languages. We define a Teacher-Student framework where the existing model “teaches” to a student model its knowledge about the languages it supports, while the student is also trained on a new language. We report an experimental evaluation in several tasks including Sentence Classification, Relational Learning and Sequence Labeling.","{'model': 'tldr@v2.0.0', 'text': 'A Continual Learning strategy that updates a model to support new languages over time, while maintaining consistent results on previously learned languages is proposed and an experimental evaluation in several tasks including Sentence Classification, Relational Learning and Sequence Labeling is reported.'}",https://aclanthology.org/2021.acl-short.106.pdf
-beyond model extraction: imitation attack for black-box nlp apis,Qiongkai Xu,"Machine-learning-as-a-service (MLaaS) has attracted millions of users to their outperforming sophisticated models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonethe-less, none of the previous stolen models can outperform the original black-box APIs. In this work, we take the first step of showing that attackers could potentially surpass victims via unsupervised domain adaptation and multi-victim ensemble. Extensive experiments on benchmark datasets and real-world APIs vali-date that the imitators can succeed in outperforming the original black-box models. We consider this as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could influence the defense or even publishing strategy of API providers.","{'model': 'tldr@v2.0.0', 'text': 'This work takes the first step of showing that attackers could potentially surpass victims via unsupervised domain adaptation and multi-victim ensemble, and considers this as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could in turnence the defense or even publishing strategy of API providers.'}",
-visualizing and understanding neural models in nlp,Jiwei Li,"While neural networks have been successfully applied to many NLP tasks the resulting vector-based models are very difficult to interpret. For example it's not clear how they achieve {\em compositionality}, building sentence meaning from the meanings of words and phrases. In this paper we describe four strategies for visualizing compositionality in neural models for NLP, inspired by similar work in computer vision. We first plot unit values to visualize compositionality of negation, intensification, and concessive clauses, allow us to see well-known markedness asymmetries in negation. We then introduce three simple and straightforward methods for visualizing a unit's {\em salience}, the amount it contributes to the final composed meaning: (1) gradient back-propagation, (2) the variance of a token from the average word node, (3) LSTM-style gates that measure information flow. We test our methods on sentiment using simple recurrent nets and LSTMs. Our general-purpose methods may have wide applications for understanding compositionality and other semantic properties of deep networks , and also shed light on why LSTMs outperform simple recurrent nets,","{'model': 'tldr@v2.0.0', 'text': 'Four strategies for visualizing compositionality in neural models for NLP, inspired by similar work in computer vision, including LSTM-style gates that measure information flow and gradient back-propagation, are described.'}",https://www.aclweb.org/anthology/N16-1082.pdf
-building representative corpora from illiterate communities: a reviewof challenges and mitigation strategies for developing countries,Stephanie Hirmer,"Most well-established data collection methods currently adopted in NLP depend on the as- sumption of speaker literacy. Consequently, the collected corpora largely fail to represent swathes of the global population, which tend to be some of the most vulnerable and marginalised people in society, and often live in rural developing areas. Such underrepresented groups are thus not only ignored when making modeling and system design decisions, but also prevented from benefiting from development outcomes achieved through data-driven NLP. This paper aims to address the under-representation of illiterate communities in NLP corpora: we identify potential biases and ethical issues that might arise when collecting data from rural communities with high illiteracy rates in Low-Income Countries, and propose a set of practical mitigation strategies to help future work.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses the under-representation of illiterate communities in NLP corpora by identifying potential biases and ethical issues that might arise when collecting data from rural communities with high illiteracy rates in Low-Income Countries, and proposing a set of practical mitigation strategies to help future work.'}",
-visual grounding strategies for text-only natural language processing,Damien Sileo,"Visual grounding is a promising path toward more robust and accurate Natural Language Processing (NLP) models. Many multimodal extensions of BERT (e.g., VideoBERT, LXMERT, VL-BERT) allow a joint modeling of texts and images that lead to state-of-the-art results on multimodal tasks such as Visual Question Answering. Here, we leverage multimodal modeling for purely textual tasks (language modeling and classification) with the expectation that the multimodal pretraining provides a grounding that can improve text processing accuracy. We propose possible strategies in this respect. A first type of strategy, referred to as transferred grounding consists in applying multimodal models to text-only tasks using a placeholder to replace image input. The second one, which we call associative grounding, harnesses image retrieval to match texts with related images during both pretraining and text-only downstream tasks. We draw further distinctions into both strategies and then compare them according to their impact on language modeling and commonsense-related downstream tasks, showing improvement over text-only baselines.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes two strategies for applying multimodal models to text-only tasks using a placeholder to replace image input and harnesses image retrieval to match texts with related images during both pretraining and text- only downstream tasks.'}",
-student surpasses teacher: imitation attack for black-box nlp apis,Qiongkai Xu,"Machine-learning-as-a-service (MLaaS) has attracted millions of users to their splendid large-scale models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonetheless, none of the previous stolen models can outperform the original black-box APIs. In this work, we conduct unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous understanding of model extraction. Extensive experiments on both benchmark datasets and real-world APIs validate that the imitators can succeed in outperforming the original black-box models on transferred domains. We consider our work as a milestone in the research of imitation attack, especially on NLP APIs, as the superior performance could influence the defense or even publishing strategy of API providers.","{'model': 'tldr@v2.0.0', 'text': 'This work conducts unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous understanding of model extraction.'}",
-search-based algorithm with scatter search strategy for automated test case generation of nlp toolkit,Fangqing Liu,"Natural language processing (NLP), as a theory-motivated computational technique, has extensive applications. Automated test case generation based on path coverage, which is a popular structural testing activity, can automatically reveal logic defects that exist in NLP programs and can save testing consumption. NLP programs have many paths that can only be covered by specific input variables. This feature makes conventional search-based algorithm very difficult covering all possible paths in NLP programs. A strategy is required for improving the search ability of search-based algorithms. In this paper, we propose a scatter search strategy to automatically generate test cases for covering all possible paths of NLP programs. The scatter search strategy empowers search-based algorithms to explore all input variables and cover the paths that require specific input variables within a small amount of test cases. The experiment results show that the proposed scatter search strategy can quickly cover the paths, which requires specific input variables. Many test cases and running time consumptions will be saved when search-based algorithms combine with scatter search strategy.","{'model': 'tldr@v2.0.0', 'text': 'A scatter search strategy is proposed to automatically generate test cases for covering all possible paths of NLP programs to empower search-based algorithms to explore all input variables and cover the paths that require specific input variables within a small amount of test cases.'}",
-evaluating tuning strategies for sequence generation with protein language models,Andrea Nathansen,"Designing artificial proteins with specialized functions promises new solutions for biological, medical, and environmental use cases. This field benefits from advances in natural language processing, with state-of-the-art text generation models already being successfully applied to protein sequences. Openly available pre-trained protein language models are able to generate artificial protein sequences and can be finetuned on very specific tasks. Considering the high computational cost of finetuning a model exclusively for one downstream task, prompt tuning has been proposed as a more cost-efficient alternative that shares one model across different tasks. However, no openly available implementation of this approach compatible with protein language models has been previously published. Thus, we adapt an open-source codebase designed for NLP models to build a pipeline for prompt tuning on protein sequence data, supporting the protein language models ProtGPT2 and RITA. We benchmark this implementation for generating proteins of a specific family and evaluate the approach using text processing metrics as well as family membership prediction and protein activity prediction of generated sequences. Our results confirm the advantages of prompt tuning in resource usage, especially storage, encouraging further research and expansion of this technique to related use cases. For our evaluated use case, prompt tuning does not reach up to finetuning in terms of the quality of generated protein sequences, indicating the need for more extensive optimization. Lastly, we observe discrepancies between results of similar evaluation tools, highlighting open problems for principled assessment of protein sequence generation quality.","{'model': 'tldr@v2.0.0', 'text': 'An open-source codebase designed for NLP models is adapted to build a pipeline for prompt tuning on protein sequence data, supporting the protein language models ProtGPT2 and RITA, and discrepancies between results of similar evaluation tools are observed, highlighting open problems for principled assessment of protein sequence generation quality.'}",https://www.biorxiv.org/content/biorxiv/early/2023/03/01/2023.02.28.530492.full.pdf
-robustness tests of nlp machine learning models: search and semantically replace,Rahul Singh,"This paper proposes a strategy to assess the robustness of different machine learning models that involve natural language processing (NLP). The overall approach relies upon a Search and Semantically Replace strategy that consists of two steps: (1) Search, which identifies important parts in the text; (2) Semantically Replace, which finds replacements for the important parts, and constrains the replaced tokens with semantically similar words. We introduce different types of Search and Semantically Replace methods designed specifically for particular types of machine learning models. We also investigate the effectiveness of this strategy and provide a general framework to assess a variety of machine learning models. Finally, an empirical comparison is provided of robustness performance among three different model types, each with a different text representation.","{'model': 'tldr@v2.0.0', 'text': 'Different types of Search and Semantically Replace methods designed specifically for particular types of machine learning models are introduced, and an empirical comparison is provided of robustness performance among three different model types, each with a different text representation.'}",
-from humans to machines: can chatgpt-like llms effectively replace human annotators in nlp tasks?,Surendrabikram Thapa,"The increasing demand for natural language processing (NLP) applications has created a need for large amounts of labeled data to train machine learning models. This has led to using human annotators for tasks such as text classification, sentiment analysis, and named entity recognition. However, human annotation is costly and time-consuming, and the annotation quality can significantly vary depending on the anno-tator. Recent advances in language modeling have led to the development of large language models (LLMs), such as Chat-GPT, which are capable of generating human-like responses to text prompts. In this position paper, we explore the question: whether ChatGPT-like LLMs can effectively replace human annotators in NLP tasks? We discuss the advantages and limitations of using LLMs for annotation and highlight some of the challenges that need to be addressed to make this a feasible approach. We argue that while LLMs can potentially reduce the cost and time required for annotation, they may not be able to fully replace human annotators in all NLP tasks. We conclude by outlining future research directions that could help advance the use of LLMs for NLP annotation.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that while LLMs can potentially reduce the cost and time required for annotation, they may not be able to fully replace human annotators in all NLP tasks.'}",
-do not trust the experts: how the lack of standard complicates nlp for historical irish,Oksana Dereza,"In this paper, we describe how we unearthed some fundamental problems while building an analogy dataset modelled on BATS (Gladkova et al., 2016) to evaluate historical Irish embeddings on their ability to detect orthographic, morphological and semantic similarity.The performance of our models in the analogy task was extremely poor regardless of the architecture, hyperparameters and evaluation metrics, while the qualitative evaluation revealed positive tendencies. We argue that low agreement between field experts on fundamental lexical and orthographic issues, and the lack of a unified editorial standard in available resources make it impossible to build reliable evaluation datasets for computational models and obtain interpretable results. We emphasise the need for such a standard, particularly for NLP applications, and prompt Celticists and historical linguists to engage in further discussion. We would also like to draw NLP scholars’ attention to the role of data and its (extra)linguistic properties in testing new models, technologies and evaluation scenarios.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that low agreement between field experts on fundamental lexical and orthographic issues, and the lack of a unified editorial standard in available resources make it impossible to build reliable evaluation datasets for computational models and obtain interpretable results.'}",https://aclanthology.org/2023.insights-1.10.pdf
-conditionally adaptive multi-task learning: improving transfer learning in nlp using fewer parameters & less data,Jonathan Pilault,"Multi-Task Learning (MTL) has emerged as a promising approach for transferring learned knowledge across different tasks. However, multi-task learning must deal with challenges such as: overfitting to low resource tasks, catastrophic forgetting, and negative task transfer, or learning interference. Additionally, in Natural Language Processing (NLP), MTL alone has typically not reached the performance level possible through per-task fine-tuning of pretrained models. However, many fine-tuning approaches are both parameter inefficient, e.g. potentially involving one new model per task, and highly susceptible to losing knowledge acquired during pretraining. We propose a novel transformer based architecture consisting of a new conditional attention mechanism as well as a set of task conditioned modules that facilitate weight sharing. Through this construction we achieve more efficient parameter sharing and mitigate forgetting by keeping half of the weights of a pretrained model fixed. We also use a new multi-task data sampling strategy to mitigate the negative effects of data imbalance across tasks. Using this approach we are able to surpass single-task fine-tuning methods while being parameter and data efficient. With our base model, we attain 2.2% higher performance compared to a full fine-tuned BERT large model on the GLUE benchmark, adding only 5.6% more trained parameters per task (whereas naive fine-tuning potentially adds 100% of the trained parameters per task) and needing only 64.6% of the data. We show that a larger variant of our single multi-task model approach performs competitively across 26 NLP tasks and yields state-of-the-art results on a number of test and development sets.","{'model': 'tldr@v2.0.0', 'text': 'A novel transformer based architecture consisting of a new conditional attention mechanism as well as a set of task conditioned modules that facilitate weight sharing is proposed that is able to surpass single-task fine-tuning methods while being parameter and data efficient.'}",
-video modeling and prompting: a comparison of two strategies for teaching cooking skills to students with mild intellectual disabilities,T. Taber-Doughty,"Self-operated video prompting and video modeling was compared when used by three secondary students with mild intellectual disabilities as they completed novel recipes during cooking activities. Alternating between video systems, students completed twelve recipes within their classroom kitchen. An alternating treatment design with a follow-up and withdrawal probe was used to illustrate the effectiveness of both systems on each student's independent task performance. Results indicated increased independence following video system use by all three students with video modeling more effective for two students and video prompting more effective for the third. Future directions for research are presented.",,
-one size does not fit all: investigating strategies for differentially-private learning across nlp tasks,Manuel Senge,"Preserving privacy in contemporary NLP models allows us to work with sensitive data, but unfortunately comes at a price. We know that stricter privacy guarantees in differentially-private stochastic gradient descent (DP-SGD) generally degrade model performance. However, previous research on the efficiency of DP-SGD in NLP is inconclusive or even counter-intuitive. In this short paper, we provide an extensive analysis of different privacy preserving strategies on seven down-stream datasets in five different ‘typical’ NLP tasks with varying complexity using modern neural models based on BERT and XtremeDistil architectures. We show that unlike standard non-private approaches to solving NLP tasks, where bigger is usually better, privacy-preserving strategies do not exhibit a winning pattern, and each task and privacy regime re-quires a special treatment to achieve adequate performance.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that unlike standard non-private approaches to solving NLP tasks, where bigger is usually better, privacy-preserving strategies do not exhibit a winning pattern, and each task and privacy regime re-quires a special treatment to achieve adequate performance.'}",
-dice loss for data-imbalanced nlp tasks,Xiaoya Li,"Many NLP tasks such as tagging and machine reading comprehension are faced with the severe data imbalance issue: negative examples significantly outnumber positive examples, and the huge number of easy-negative examples overwhelms the training. The most commonly used cross entropy (CE) criteria is actually an accuracy-oriented objective, and thus creates a discrepancy between training and test: at training time, each training instance contributes equally to the objective function, while at test time F1 score concerns more about positive examples. In this paper, we propose to use dice loss in replacement of the standard cross-entropy objective for data-imbalanced NLP tasks. Dice loss is based on the Sørensen--Dice coefficient or Tversky index , which attaches similar importance to false positives and false negatives, and is more immune to the data-imbalance issue. To further alleviate the dominating influence from easy-negative examples in training, we propose to associate training examples with dynamically adjusted weights to deemphasize easy-negative examples. Theoretical analysis shows that this strategy narrows down the gap between the F1 score in evaluation and the dice loss in training. With the proposed training objective, we observe significant performance boost on a wide range of data imbalanced NLP tasks. Notably, we are able to achieve SOTA results on CTB5, CTB6 and UD1.4 for the part of speech tagging task; SOTA results on CoNLL03, OntoNotes5.0, MSRA and OntoNotes4.0 for the named entity recognition task; along with competitive results on the tasks of machine reading comprehension and paraphrase identification.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to use dice loss in replacement of the standard cross-entropy objective for data-imbalanced NLP tasks, based on the Sørensen--Dice coefficient or Tversky index, which attaches similar importance to false positives and false negatives, and is more immune to the data-IMbalance issue.'}",https://arxiv.org/pdf/1911.02855
-decoding strategies for improving low-resource machine translation,Chanjun Park,"Pre-processing and post-processing are significant aspects of natural language processing (NLP) application software. Pre-processing in neural machine translation (NMT) includes subword tokenization to alleviate the problem of unknown words, parallel corpus filtering that only filters data suitable for training, and data augmentation to ensure that the corpus contains sufficient content. Post-processing includes automatic post editing and the application of various strategies during decoding in the translation process. Most recent NLP researches are based on the Pretrain-Finetuning Approach (PFA). However, when small and medium-sized organizations with insufficient hardware attempt to provide NLP services, throughput and memory problems often occur. These difficulties increase when utilizing PFA to process low-resource languages, as PFA requires large amounts of data, and the data for low-resource languages are often insufficient. Utilizing the current research premise that NMT model performance can be enhanced through various pre-processing and post-processing strategies without changing the model, we applied various decoding strategies to Korean–English NMT, which relies on a low-resource language pair. Through comparative experiments, we proved that translation performance could be enhanced without changes to the model. We experimentally examined how performance changed in response to beam size changes and n-gram blocking, and whether performance was enhanced when a length penalty was applied. The results showed that various decoding strategies enhance the performance and compare well with previous Korean–English NMT approaches. Therefore, the proposed methodology can improve the performance of NMT models, without the use of PFA; this presents a new perspective for improving machine translation performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed methodology can improve the performance of NMT models, without the use of PFA; this presents a new perspective for improving machine translation performance.'}",https://www.mdpi.com/2079-9292/9/10/1562/pdf?version=1600932781
-unlink the link between covid-19 and 5g networks: an nlp and sna based approach,Mohammed Bahja,"Social media facilitates rapid dissemination of information for both factual and fictional information. The spread of non-scientific information through social media platforms such as Twitter has potential to cause damaging consequences. Situations such as the COVID-19 pandemic provides a favourable environment for misinformation to thrive. The upcoming 5G technology is one of the recent victims of misinformation and fake news and has been plagued with misinformation about the effects of its radiation. During the COVID-19 pandemic, conspiracy theories linking the cause of the pandemic to 5G technology have resonated with a section of people leading to outcomes such as destructive attacks on 5G towers. The analysis of the social network data can help to understand the nature of the information being spread and identify the commonly occurring themes in the information. The natural language processing (NLP) and the statistical analysis of the social network data can empower policymakers to understand the misinformation being spread and develop targeted strategies to counter the misinformation. In this paper, NLP based analysis of tweets linking COVID-19 to 5G is presented. NLP models including Latent Dirichlet allocation (LDA), sentiment analysis (SA) and social network analysis (SNA) were applied for the analysis of the tweets and identification of topics. An understanding of the topic frequencies, the inter-relationships between topics and geographical occurrence of the tweets allows identifying agencies and patterns in the spread of misinformation and equips policymakers with knowledge to devise counter-strategies.","{'model': 'tldr@v2.0.0', 'text': 'NLP based analysis of tweets linking COVID-19 to 5G is presented and an understanding of the topic frequencies, the inter-relationships between topics and geographical occurrence of the tweets allows identifying agencies and patterns in the spread of misinformation and equips policymakers with knowledge to devise counter-strategies.'}",https://ieeexplore.ieee.org/ielx7/6287639/8948470/09262907.pdf
-adaptation of idpt system based on patient-authored text data using nlp,S. K. Mukhiya,"Background: Internet-Delivered Psychological Treatment (IDPT) systems have the potential to provide evidence-based mental health treatments for a far-reaching population at a lower cost. However, most of the current IDPT systems follow a tunnel-based treatment process and do not adapt to the needs of different patients'. In this paper, we explore the possibility of applying Natural Language Processing (NLP) for personalizing mental health interventions. Objective: The primary objective of this study is to present an adaptive strategy based on NLP techniques that analyses patient-authored text data and extract depression symptoms based on a clinically established assessment questionnaire, PHQ-9. Method: We propose a novel word-embedding (Depression2Vec) to extract depression symptoms from patient authored text data and compare it with three state-of-the-art NLP techniques. We also present an adaptive IDPT system that personalizes treatments for mental health patients based on the proposed depression symptoms detection technique. Result: Our results indicate that the performance of proposed embedding Depression2Vec is comparable to WordNet, but in some cases, the former outperforms the latter with respect to extracting depression symptoms from the patient-authored text. Conclusion: Although the extraction of symptoms from text is challenging, our proposed method can effectively extract depression symptoms from text data, which can be used to deliver personalized intervention.","{'model': 'tldr@v2.0.0', 'text': 'A novel word-embedding to extract depression symptoms from patient authored text data and compare it with three state-of-the-art NLP techniques and indicates that the performance of proposed embedding Depression2Vec is comparable to WordNet, but in some cases, the former outperforms the latter with respect to extracting depression symptoms.'}",
-"internet data analysis methodology for cyberterrorism vocabulary detection, combining techniques of big data analytics, nlp and semantic web",Iván Castillo-Zúñiga,"This article presents a methodology for the analysis of data on the Internet, combining techniques of Big Data analytics, NLP and semantic web in order to find knowledge about large amounts of information on the web. To test the effectiveness of the proposed method, webpages about cyberterrorism were analyzed as a case study. The procedure implemented a genetic strategy in parallel, which integrates (Crawler to locate and download information from the web; to retrieve the vocabulary, using techniques of NLP (tokenization, stop word, TF, TFIDF), methods of stemming and synonyms). For the pursuit of knowledge was built a dataset through the description of a linguistic corpus with semantic ontologies, considering the characteristics of cyber-terrorism, which was analyzed with the algorithms, Random Forests (parallel), Boosting, SVM, neural network, K-nn and Bayes. The results reveal a percentage of the 95.62% accuracy in the detection of the vocabulary of cyber-terrorism, which were approved through cross validation, reaching 576% time savings with parallel processing.","{'model': 'tldr@v2.0.0', 'text': 'A methodology for the analysis of data on the Internet, combining techniques of Big Data analytics, NLP and semantic web in order to find knowledge about large amounts of information on the web, reaching 576% time savings with parallel processing.'}",
-a hybrid lp/nlp paradigm for global optimization relaxations,Aida Khajavirad,,"{'model': 'tldr@v2.0.0', 'text': 'Results show that incorporating the proposed techniques in the BARON software leads to significant reductions in execution time, and increases by 30% the number of problems that are solvable to global optimality within 500 s on a standard workstation.'}",
-interacting tcp and nlp transcription factors control plant responses to nitrate availability,Peizhu Guan,"Significance Nitrate is an essential nutrient and a critical signal for plant growth, development, and stress responses. Nitrate signaling underlies a myriad of physiological, morphological, and developmental processes. Here we report that interacting teosinte branched1/cycloidea/proliferating cell factor and NIN-like protein transcription factors constitute a molecular link between nitrate signaling and the control of the cell-cycle progression gene CYCB1;1 and root meristem growth. Our findings shed light on the regulatory mechanisms underlying an important plant adaptive process for coping with and surviving environmental challenges. Plants have evolved adaptive strategies that involve transcriptional networks to cope with and survive environmental challenges. Key transcriptional regulators that mediate responses to environmental fluctuations in nitrate have been identified; however, little is known about how these regulators interact to orchestrate nitrogen (N) responses and cell-cycle regulation. Here we report that teosinte branched1/cycloidea/proliferating cell factor1-20 (TCP20) and NIN-like protein (NLP) transcription factors NLP6 and NLP7, which act as activators of nitrate assimilatory genes, bind to adjacent sites in the upstream promoter region of the nitrate reductase gene, NIA1, and physically interact under continuous nitrate and N-starvation conditions. Regions of these proteins necessary for these interactions were found to include the type I/II Phox and Bem1p (PB1) domains of NLP6&7, a protein-interaction module conserved in animals for nutrient signaling, and the histidine- and glutamine-rich domain of TCP20, which is conserved across plant species. Under N starvation, TCP20-NLP6&7 heterodimers accumulate in the nucleus, and this coincides with TCP20 and NLP6&7-dependent up-regulation of nitrate assimilation and signaling genes and down-regulation of the G2/M cell-cycle marker gene, CYCB1;1. TCP20 and NLP6&7 also support root meristem growth under N starvation. These findings provide insights into how plants coordinate responses to nitrate availability, linking nitrate assimilation and signaling with cell-cycle progression.","{'model': 'tldr@v2.0.0', 'text': 'Insight is provided into how plants coordinate responses to nitrate availability, linking nitrate assimilation and signaling with cell-cycle progression and root meristem growth.'}",https://www.pnas.org/content/pnas/114/9/2419.full.pdf
-understanding the effect of in-video prompting on learners and instructors,Hyungyu Shin,"Online instructional videos are ubiquitous, but it is difficult for instructors to gauge learners' experience and their level of comprehension or confusion regarding the lecture video. Moreover, learners watching the videos may become disengaged or fail to reflect and construct their own understanding. This paper explores instructor and learner perceptions of in-video prompting where learners answer reflective questions while watching videos. We conducted two studies with crowd workers to understand the effect of prompting in general, and the effect of different prompting strategies on both learners and instructors. Results show that some learners found prompts to be useful checkpoints for reflection, while others found them distracting. Instructors reported the collected responses to be generally more specific than what they have usually collected. Also, different prompting strategies had different effects on the learning experience and the usefulness of responses as feedback.","{'model': 'tldr@v2.0.0', 'text': 'Examination of instructor and learner perceptions of in-video prompting where learners answer reflective questions while watching videos shows that some learners found prompts to be useful checkpoints for reflection, while others found them distracting.'}",
-spaml: a bimodal ensemble learning spam detector based on nlp techniques,Jaouhar Fattahi,"In this paper, we put forward a new tool, called SpaML, for spam detection using a set of supervised and unsupervised classifiers, and two techniques imbued with Natural Language Processing (NLP), namely Bag of Words (BoW) and Term Frequency-Inverse Document Frequency (TF-IDF). We first present the NLP techniques used. Then, we present our classifiers and their performance on each of these techniques. Then, we present our overall Ensemble Learning classifier and the strategy we are using to combine them. Finally, we present the interesting results shown by SpaML in terms of accuracy and precision.","{'model': 'tldr@v2.0.0', 'text': 'A new tool, called SpaML, for spam detection using a set of supervised and unsupervised classifiers, and two techniques imbued with Natural Language Processing (NLP), namely Bag of Words (BoW) and Term Frequency-Inverse Document Frequency (TF-IDF).'}",https://arxiv.org/pdf/2010.07444
-analyzing student strategies in blended courses using clickstream data,Nil-Jana Akpinar,"Educational software data promises unique insights into students' study behaviors and drivers of success. While much work has been dedicated to performance prediction in massive open online courses, it is unclear if the same methods can be applied to blended courses and a deeper understanding of student strategies is often missing. We use pattern mining and models borrowed from Natural Language Processing (NLP) to understand student interactions and extract frequent strategies from a blended college course. Fine-grained clickstream data is collected through Diderot, a non-commercial educational support system that spans a wide range of functionalities. We find that interaction patterns differ considerably based on the assessment type students are preparing for, and many of the extracted features can be used for reliable performance prediction. Our results suggest that the proposed hybrid NLP methods can provide valuable insights even in the low-data setting of blended courses given enough data granularity.","{'model': 'tldr@v2.0.0', 'text': 'This work uses pattern mining and models borrowed from Natural Language Processing to understand student interactions and extract frequent strategies from a blended college course, and suggests that the proposed hybrid NLP methods can provide valuable insights even in the low-data setting of blended courses given enough data granularity.'}",
-prompting undergraduate students’ metacognition of learning: implementing ‘meta-learning’ assessment tasks in the biomedical sciences,Kay Colthorpe,"Abstract To succeed at post-secondary education, it’s essential that students develop an understanding of their own knowledge and learning processes. This metacognition of learning, or ‘meta-learning’, helps students to become more effective learners, as they become more aware of their self-regulatory processes and recognise the effectiveness of their study strategies. To increase biomedical science students’ self-awareness, we have designed and implemented meta-learning assessment tasks across our biomedical science courses. Most students reported that meta-learning tasks had a positive impact on their learning, as they prompted self-regulatory processes of forethought and self-reflection. We found that students were equally likely to change or not change their study strategies across subsequent semesters. Those students that did not change were generally high achievers, believing their study approaches were most effective, but their performance did not improve across semesters. In contrast, students who adapted, mostly by modifying how they appraised and rearranged records or improved planning and time management, performed less well overall but significantly improved their performance across semesters. Meta-learning tasks may prompt students to become more self-reflective and independent learners by affecting their approach to learning, enabling them to reflect on their study strategies, adapt and improve performance, and may enable the development of lifelong learning skills.",,
-the use of response prompting and frames for teaching sentence writing to students with moderate intellectual disability,R. Pennington,"In the current study, we examined the effects of response prompting strategies (i.e., constant time delay, system of least prompts) and frames on sentence writing for three participants, ages 7 to 12, with moderate intellectual disability. We used a concurrent multiple probe across behaviors design to evaluate the efficacy of the intervention package and posttest probes to assess generalized responding to untrained stimulation. During intervention, the teacher taught two students to construct sentences using selection-based software and another to generate handwritten responses across three different writing frames (i.e., I want _________, I see _____, The _____ is ______). Our findings suggest that the package was effective and produced variable levels of maintenance and generalized responding for all three participants.",,
-improving robustness and generality of nlp models using disentangled representations,Jiawei Wu,"Supervised neural networks, which first map an input $x$ to a single representation $z$, and then map $z$ to the output label $y$, have achieved remarkable success in a wide range of natural language processing (NLP) tasks. Despite their success, neural models lack for both robustness and generality: small perturbations to inputs can result in absolutely different outputs; the performance of a model trained on one domain drops drastically when tested on another domain. In this paper, we present methods to improve robustness and generality of NLP models from the standpoint of disentangled representation learning. Instead of mapping $x$ to a single representation $z$, the proposed strategy maps $x$ to a set of representations $\{z_1,z_2,...,z_K\}$ while forcing them to be disentangled. These representations are then mapped to different logits $l$s, the ensemble of which is used to make the final prediction $y$. We propose different methods to incorporate this idea into currently widely-used models, including adding an $L$2 regularizer on $z$s or adding Total Correlation (TC) under the framework of variational information bottleneck (VIB). We show that models trained with the proposed criteria provide better robustness and domain adaptation ability in a wide range of supervised learning tasks.","{'model': 'tldr@v2.0.0', 'text': 'Methods to improve robustness and generality of NLP models from the standpoint of disentangled representation learning are presented and it is shown that models trained with the proposed criteria provide better robusts and domain adaptation ability in a wide range of supervised learning tasks.'}",
-"complementarity, f-score, and nlp evaluation",Leon Derczynski,"This paper addresses the problem of quantifying the differences between entity extraction systems, where in general only a small proportion a document should be selected. Comparing overall accuracy is not very useful in these cases, as small differences in accuracy may correspond to huge differences in selections over the target minority class. Conventionally, one may use per-token complementarity to describe these differences, but it is not very useful when the set is heavily skewed. In such situations, which are common in information retrieval and entity recognition, metrics like precision and recall are typically used to describe performance. However, precision and recall fail to describe the differences between sets of objects selected by different decision strategies, instead just describing the proportional amount of correct and incorrect objects selected. This paper presents a method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches.","{'model': 'tldr@v2.0.0', 'text': 'A method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches is presented.'}",
-effects of a progressive prompting-based educational game on second graders' mathematics learning performance and behavioral patterns,Kai-Hsiang Yang,"Game-based learning (GBL) has been proven to be an attractive learning model by many studies; however, scholars have pointed out that the effectiveness of game-based learning could be limited if proper learning strategies are not incorporated. Prompting is a strategy that plays the important role of providing hints and guidance in interactive learning environments. Therefore, this study proposes a game-based learning approach with a progressive prompting strategy, using different levels of hints to guide students to complete tasks and achieve learning goals. Quasi-experimental research was employed in this study using two groups of students. The experimental group learned with the proposed approach, while the control group was allocated a conventional game-based learning strategy. The findings of the study show that the proposed approach significantly improved the second graders’ mathematics learning achievement. From the analysis of the experimental students’ learning behaviors, it was reported that the experimental group students could generate the answer after two progressive prompts. Thus, we could see that the proposed approach could enhance the learning achievement of the experimental group students by correctly guiding them to answer questions, step up their thinking, and understand the learning content in the learning process.","{'model': 'tldr@v2.0.0', 'text': 'The proposed approach with a progressive prompting strategy could enhance the learning achievement of the experimental group students by correctly guiding them to answer questions, step up their thinking, and understand the learning content in the learning process.'}",
-multimix: a robust data augmentation strategy for cross-lingual nlp,M Saiful Bari,"Transfer learning has yielded state-of-the-art results in many supervised natural language processing tasks. However, annotated data for every target task in every target language is rare, especially for low-resource languages. In this work, we propose MultiMix, a novel data augmentation method for semi-supervised learning in zero-shot transfer learning scenarios. In particular, MultiMix targets to solve cross-lingual adaptation problems from a source (language) distribution to an unknown target (language) distribution assuming it has no training labels in the target language task. In its heart, MultiMix performs simultaneous self-training with data augmentation and unsupervised sample selection. To show its effectiveness, we have performed extensive experiments on zero-shot transfers for cross-lingual named entity recognition (XNER) and natural language inference (XNLI). Our experiments show sizeable improvements in both tasks outperforming the baselines by a good margin.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes MultiMix, a novel data augmentation method for semi-supervised learning in zero-shot transfer learning scenarios that targets to solve cross-lingual adaptation problems from a source ( language) distribution to an unknown target (language) distribution assuming it has no training labels in the target language task.'}",
-a just and comprehensive strategy for using nlp to address online abuse,David Jurgens,"Online abusive behavior affects millions and the NLP community has attempted to mitigate this problem by developing technologies to detect abuse. However, current methods have largely focused on a narrow definition of abuse to detriment of victims who seek both validation and solutions. In this position paper, we argue that the community needs to make three substantive changes: (1) expanding our scope of problems to tackle both more subtle and more serious forms of abuse, (2) developing proactive technologies that counter or inhibit abuse before it harms, and (3) reframing our effort within a framework of justice to promote healthy communities.","{'model': 'tldr@v2.0.0', 'text': 'This position paper argues that the NLP community needs to make three substantive changes: expanding the scope of problems to tackle both more subtle and more serious forms of abuse, developing proactive technologies that counter or inhibit abuse before it harms, and reframing the effort within a framework of justice to promote healthy communities.'}",https://www.aclweb.org/anthology/P19-1357.pdf
-"breaking nlp: using morphosyntax, semantics, pragmatics and world knowledge to fool sentiment analysis systems",Taylor Mahler,"This paper describes our “breaker” submission to the 2017 EMNLP “Build It Break It” shared task on sentiment analysis. In order to cause the “builder” systems to make incorrect predictions, we edited items in the blind test data according to linguistically interpretable strategies that allow us to assess the ease with which the builder systems learn various components of linguistic structure. On the whole, our submitted pairs break all systems at a high rate (72.6%), indicating that sentiment analysis as an NLP task may still have a lot of ground to cover. Of the breaker strategies that we consider, we find our semantic and pragmatic manipulations to pose the most substantial difficulties for the builder systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the “breaker” submission to the 2017 EMNLP “Build It Break It” shared task on sentiment analysis, and finds the semantic and pragmatic manipulations to pose the most substantial difficulties for the builder systems.'}",https://www.aclweb.org/anthology/W17-5405.pdf
-framing and agenda-setting in russian news: a computational analysis of intricate political strategies,Anjalie Field,"Amidst growing concern over media manipulation, NLP attention has focused on overt strategies like censorship and “fake news”. Here, we draw on two concepts from political science literature to explore subtler strategies for government media manipulation: agenda-setting (selecting what topics to cover) and framing (deciding how topics are covered). We analyze 13 years (100K articles) of the Russian newspaper Izvestia and identify a strategy of distraction: articles mention the U.S. more frequently in the month directly following an economic downturn in Russia. We introduce embedding-based methods for cross-lingually projecting English frames to Russian, and discover that these articles emphasize U.S. moral failings and threats to the U.S. Our work offers new ways to identify subtle media manipulation strategies at the intersection of agenda-setting and framing.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces embedding-based methods for cross-lingually projecting English frames to Russian, and offers new ways to identify subtle media manipulation strategies at the intersection of agenda-setting and framing.'}",https://www.aclweb.org/anthology/D18-1393.pdf
-role of entrepreneurial orientation in smes global performance: testing marketing strategies and technological orientation as mediators,S. Aslam,"Entrepreneurial orientation (EO) plays a key role in a firm’s success, particularly in the context of globalization. Thus, EO has gained prompt attention because of an increasing trend of globalized markets and economies. This article expands the understanding of Entrepreneurial Orientation as it studies EO to assess Small and Medium Enterprise (SMEs) global performance. The study extends the knowledge of SMEs entering global markets by considering firms’ marketing strategy and technological orientation as mediators. Through these mechanisms, corporate EO leads to cross-border excellence. The findings of this study, through a sample of 155 international SMEs in Pakistan, support the impact of EO on global performance whereas the marketing strategy and technological orientation serve as the mediators in this core relationship. The data were analyzed using a variance-based structural equation model (SEM) in Smart PLS software. The findings validate that SMEs, more concentrating on EO, are more able to explore new opportunities in global markets, but their small size, inadequate access to different resources and limited experience in operating in international markets prevents them from taking full advantage of arising opportunities. Hence the findings confirm the significance of EO in the globalization of SMEs however, it was also observed that although the marketing strategy and technological orientation may reduce the direct impact of EO on international performance, these mechanisms accelerate the indirect impact of EO as a tangible and intangible resource. ",,https://coralpublications.org/index.php/jms/article/download/157/116
-iie-nlp-nut at semeval-2020 task 4: guiding plm with prompt template reconstruction strategy for comve,Luxi Xing,"This paper introduces our systems for the first two subtasks of SemEval Task4: Commonsense Validation and Explanation. To clarify the intention for judgment and inject contrastive information for selection, we propose the input reconstruction strategy with prompt templates. Specifically, we formalize the subtasks into the multiple-choice question answering format and construct the input with the prompt templates, then, the final prediction of question answering is considered as the result of subtasks. Experimental results show that our approaches achieve significant performance compared with the baseline systems. Our approaches secure the third rank on both official test sets of the first two subtasks with an accuracy of 96.4 and an accuracy of 94.3 respectively.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces the systems for the first two subtasks of SemEval Task4: Commonsense Validation and Explanation and proposes the input reconstruction strategy with prompt templates, which formalizes the subtasks into the multiple-choice question answering format and construct the input with the prompt templates.'}",https://aclanthology.org/2020.semeval-1.42.pdf
-a joint many-task model: growing a neural network for multiple nlp tasks,Kazuma Hashimoto,"Transfer and multi-task learning have traditionally focused on either a single source-target pair or very few, similar tasks. Ideally, the linguistic levels of morphology, syntax and semantics would benefit each other by being trained in a single model. We introduce a joint many-task model together with a strategy for successively growing its depth to solve increasingly complex tasks. Higher layers include shortcut connections to lower-level task predictions to reflect linguistic hierarchies. We use a simple regularization term to allow for optimizing all model weights to improve one task’s loss without exhibiting catastrophic interference of the other tasks. Our single end-to-end model obtains state-of-the-art or competitive results on five different tasks from tagging, parsing, relatedness, and entailment tasks.","{'model': 'tldr@v2.0.0', 'text': 'A joint many-task model together with a strategy for successively growing its depth to solve increasingly complex tasks and uses a simple regularization term to allow for optimizing all model weights to improve one task’s loss without exhibiting catastrophic interference of the other tasks.'}",https://www.aclweb.org/anthology/D17-1206.pdf
-when low resource nlp meets unsupervised language model: meta-pretraining then meta-learning for few-shot text classification (student abstract),Shumin Deng,"Text classification tends to be difficult when data are deficient or when it is required to adapt to unseen classes. In such challenging scenarios, recent studies have often used meta-learning to simulate the few-shot task, thus negating implicit common linguistic features across tasks. This paper addresses such problems using meta-learning and unsupervised language models. Our approach is based on the insight that having a good generalization from a few examples relies on both a generic model initialization and an effective strategy for adapting this model to newly arising tasks. We show that our approach is not only simple but also produces a state-of-the-art performance on a well-studied sentiment classification dataset. It can thus be further suggested that pretraining could be a promising solution for few-shot learning of many other NLP tasks. The code and the dataset to replicate the experiments are made available at https://github.com/zxlzr/FewShotNLP.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses text classification problems using meta-learning and unsupervised language models using pretraining and shows that this approach is not only simple but also produces a state-of-the-art performance on a well-studied sentiment classification dataset.'}",https://ojs.aaai.org/index.php/AAAI/article/download/7158/7012
-"implementing nlp projects for noncentral languages: instructions for funding bodies, strategies for developers",O. Streiter,,"{'model': 'tldr@v2.0.0', 'text': 'The concepts of free software and software pools are elaborate and it is argued that NCLPs should embrace an open-source approach for the resources they develop and pool these resources together with other similaropen-source resources.'}",
-a comparison of static picture prompting and video prompting simulation strategies using group instructional procedures,David F. Cihak,Two groups of three students with moderate mental retardation were instructed using group procedures to compare static picture and video prompting simulation strategies. An alternating treatments design was used to compare individual student task acquisition and maintenance performances of purchasing and banking skills. The results indicated that these two strategies were equally effective and efficient. Individual differences suggest that simulation strategy materials benefit student learning by displaying relevant stimuli for a stationary period time during presentation.,,
-prompting all students to learn: examining dynamic assessment of special needs and typical students in a prekindergarten inclusive french program.,Michele Regalla,"This study examined a teacher's prompting strategies and the use of dynamic assessment (DA) in an inclusive prekindergarten French program. Prior research has shown that DA is an effective method to assess both foreign language learning and first language development for typically developing students and for students with special needs, as well as for identifying bilingual students with language impairments. Data were collected from videorecordings of class sessions, DA scores, and a questionnaire on which the French teacher reported her classroom prompting strategies and predicted students’ needs for prompting on the DA. Scores on the DA and the French teacher's predictions of student prompting needs were analyzed through independent t tests. In addition, 30 videorecordings of class sessions were analyzed for prompting strategies. The results indicated that DA can provide insights about the learning of typically developing students and special needs students in a prekindergarten French program. Although students with special needs required both more prompting and more practice time than typically developing students, all special needs students in this study were able to provide at least one second language response during the DA.",,
-representation and control strategies for large knowledge domains: an application to nlp,F. Antonacci,"The design issues encountered during the development of a natural language processor (NLP) for the Italian language are described. The focus is on strategic aspects, namely representation and control, and their implementation with first-order logic. The complexity and the size of the knowledge domain (press agency releases on finance and economics) do not present severe restrictions in the sentence structure; hence a considerable design effort for data structures and control algorithms was required. Logic proved to be an important tool for implementing in a modular and efficient way the knowledge sources along with the programs that derive the morphologic, syntactic, and semantic features of sentences. As for the data structures, we found a considerable advantage in separating linguistic knowledge in three sources: morphologic, syntactic, and semantic. This resulted in a clear and systematic representation scheme and reduced the complexity of the parsing system.","{'model': 'tldr@v2.0.0', 'text': 'The design issues encountered during the development of a natural language processor for the Italian language are described and Logic proved to be an important tool for implementing in a modular and efficient way the knowledge sources along with the programs that derive the morphologic, syntactic, and semantic features of sentences.'}",
-embedding strategies for specialized domains: application to clinical entity recognition,Hicham El Boukkouri,"Using pre-trained word embeddings in conjunction with Deep Learning models has become the “de facto” approach in Natural Language Processing (NLP). While this usually yields satisfactory results, off-the-shelf word embeddings tend to perform poorly on texts from specialized domains such as clinical reports. Moreover, training specialized word representations from scratch is often either impossible or ineffective due to the lack of large enough in-domain data. In this work, we focus on the clinical domain for which we study embedding strategies that rely on general-domain resources only. We show that by combining off-the-shelf contextual embeddings (ELMo) with static word2vec embeddings trained on a small in-domain corpus built from the task data, we manage to reach and sometimes outperform representations learned from a large corpus in the medical domain.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that by combining off-the-shelf contextual embeddings (ELMo) with static word2vecembeddings trained on a small in-domain corpus built from the task data, they manage to reach and sometimes outperform representations learned from a large corpus in the medical domain.'}",https://www.aclweb.org/anthology/P19-2041.pdf
-"simple, affordable, and effective strategies for prompting reading behavior",Sara C. Ernsbarger,"Many students with disabilities benefit from prompts for reading in general education classrooms. Visual, auditory, kinesthetic, and tactile prompts implemented by the teacher, by the student with minimal teacher assistance, or by the student independently can facilitate more effective reading. Prompting strategies provide clues and help students remember to perform a certain behavior, self-monitor their behavior, and focus their attention. Examples of prompts for each type of modality and level of implementation are provided.",,
-a bilevel nlp sensitivity‐based decomposition for dynamic optimization with moving finite elements,Weifeng Chen,"Optimal control has guided numerous applications in chemical engineering, and exact determination of optimal profiles is essential for operation of separation and reactive processes, and operating strategies and recipe generation for batch processes. Here, a simultaneous collocation formulation based on moving finite elements is developed for the solution of a class of optimal control problems. Novel features of the algorithm include the direct location of breakpoints for control profiles and a termination criterion based on a constant Hamiltonian profile. The algorithm is stabilized and performance is significantly improved by decomposing the overall nonlinear programming (NLP) formulation into an inner problem, which solves a fixed element simultaneous collocation problem, and an outer problem, which adjusts the finite elements based on several error criteria. This bilevel formulation is aided by a NLP solver (the interior point optimizer) for both problems as well as an NLP sensitivity component, which provides derivative information from the inner problem to the outer problem. This approach is demonstrated on 11 dynamic optimization problems drawn from the optimal control and chemical engineering literature. © 2014 American Institute of Chemical Engineers AIChE J, 60: 966–979, 2014",,
-reducing unnecessary urinary catheter use and other strategies to prevent catheter-associated urinary tract infection: an integrative review,J. Meddings,"Background Catheter-associated urinary tract infections (CAUTI) are costly, common and often preventable by reducing unnecessary urinary catheter (UC) use. Methods To summarise interventions to reduce UC use and CAUTIs, we updated a prior systematic review (through October 2012), and a meta-analysis regarding interventions prompting UC removal by reminders or stop orders. A narrative review summarises other CAUTI prevention strategies including aseptic insertion, catheter maintenance, antimicrobial UCs, and bladder bundle implementation. Results 30 studies were identified and summarised with interventions to prompt removal of UCs, with potential for inclusion in the meta-analyses. By meta-analysis (11 studies), the rate of CAUTI (episodes per 1000 catheter-days) was reduced by 53% (rate ratio 0.47; 95% CI 0.30 to 0.64, p<0.001) using a reminder or stop order, with five studies also including interventions to decrease initial UC placement. The pooled (nine studies) standardised mean difference (SMD) in catheterisation duration (days) was −1.06 overall (p=0.065) including a statistically significant decrease in stop-order studies (SMD −0.37; p<0.001) but not in reminder studies (SMD, −1.54; p=0.071). No significant harm from catheter removal strategies is supported. Limited research is available regarding the impact of UC insertion and maintenance technique. A recent randomised controlled trial indicates antimicrobial catheters provide no significant benefit in preventing symptomatic CAUTIs. Conclusions UC reminders and stop orders appear to reduce CAUTI rates and should be used to improve patient safety. Several evidence-based guidelines have evaluated CAUTI preventive strategies as well as emerging evidence regarding intervention bundles. Implementation strategies are important because reducing UC use involves changing well-established habits.","{'model': 'tldr@v2.0.0', 'text': 'UC reminders and stop orders appear to reduce CAUTI rates and should be used to improve patient safety, as well as emerging evidence regarding intervention bundles.'}",https://qualitysafety.bmj.com/content/qhc/23/4/277.full.pdf
-considering student choice when selecting instructional strategies: a comparison of three prompting systems.,T. Taber-Doughty,,"{'model': 'tldr@v2.0.0', 'text': 'This study found that the preferred prompting systems were more effective in terms of both skill acquisition and duration for completing tasks for students.'}",
-self-directed video prompting and least-to-most prompting: examining ways of increasing vocational skill acquisition among students with autism spectrum disorder and intellectual disability,Gulnoza Yakubova,,"{'model': 'tldr@v2.0.0', 'text': 'VP and LMP as a combined intervention can be effective in teaching vocational tasks that involve process steps to students with both autism spectrum disorder and intellectual disability when working on school-based vocational tasks.'}",
-news summarization and evaluation in the era of gpt-3,Tanya Goyal,"The recent success of prompting large language models like GPT-3 has led to a paradigm shift in NLP research. In this paper, we study its impact on text summarization, focusing on the classic benchmark domain of news summarization. First, we investigate how GPT-3 compares against fine-tuned models trained on large summarization datasets. We show that not only do humans overwhelmingly prefer GPT-3 summaries, prompted using only a task description, but these also do not suffer from common dataset-specific issues such as poor factuality. Next, we study what this means for evaluation, particularly the role of gold standard test sets. Our experiments show that both reference-based and reference-free automatic metrics cannot reliably evaluate GPT-3 summaries. Finally, we evaluate models on a setting beyond generic summarization, specifically keyword-based summarization, and show how dominant fine-tuning approaches compare to prompting. To support further research, we release: (a) a corpus of 10K generated summaries from fine-tuned and prompt-based models across 4 standard summarization benchmarks, (b) 1K human preference judgments comparing different systems for generic- and keyword-based summarization.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that not only do humans overwhelmingly prefer GPT-3 summaries, prompted using only a task description, but these also do not suffer from common dataset-specific issues such as poor factuality.'}",http://arxiv.org/pdf/2209.12356
-torque distribution strategies for energy-efficient electric vehicles with multiple drivetrains,B. Lenzo,"The paper discusses novel computationally efficient torque distribution strategies for electric vehicles with individually controlled drivetrains, aimed at minimizing the overall power losses while providing the required level of wheel torque and yaw moment. Analytical solutions of the torque control allocation problem are derived and effects of load transfers due to driving/braking and cornering are studied and discussed in detail. Influences of different drivetrain characteristics on the front and rear axles are described. The results of an analytically derived algorithm are contrasted with those from two other control allocation strategies, based on the offline numerical solution of more detailed formulations of the control allocation problem (i.e., a multiparametric nonlinear programming (mp-NLP) problem). The control allocation algorithms are experimentally validated with an electric vehicle with four identical drivetrains along multiple driving cycles and in steady-state cornering. The experiments show that the computationally efficient algorithms represent a very good compromise between low energy consumption and controller complexity.",,http://shura.shu.ac.uk/15643/1/Lenzo%20-%20Torque%20distribution%20%28AM%29%20%2B%20coversheet.pdf
-a progressive prompting approach to conducting context-aware learning activities for natural science courses,Chih-Hung Chen,"Owing to the popularity of mobile, wireless communication and sensing technologies, issues related to contextual mobile learning have widely been discussed in recent years. In the meantime, researchers have indicated the importance of developing effective strategies for guiding students to learn in context-aware learning environments. In this study, a progressive prompt-based context-aware learning approach is proposed to improve the learning performance of students. An experiment was conducted on a natural science course of an elementary school to evaluate the performance of the proposed approach. From the experimental results, it is found that the proposed approach could effectively enhance the learning achievement of the students in comparison with the conventional context-aware learning system with single-stage prompts. It is also interesting to find that the proposed approach, by providing more challenging tasks, encouraged the students to put more effort into examining the contextual information and interpreting the learning content.","{'model': 'tldr@v2.0.0', 'text': 'A progressive prompt-based context-aware learning approach is proposed to improve the learning performance of students and it is found that the proposed approach could effectively enhance the learning achievement of the students in comparison with the conventional context- Aware learning system with single-stage prompts.'}",
-opt-iml: scaling language model instruction meta learning through the lens of generalization,S. Iyer,"Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding of the performance trade-offs of different decisions made during the instruction-tuning process. These decisions include the scale and diversity of the instruction-tuning benchmark, different task sampling strategies, fine-tuning with and without demonstrations, training using specialized datasets for reasoning and dialogue, and finally, the fine-tuning objectives themselves. In this paper, we characterize the effect of instruction-tuning decisions on downstream task performance when scaling both model and benchmark sizes. To this end, we create OPT-IML Bench: a large benchmark for Instruction Meta-Learning (IML) of 2000 NLP tasks consolidated into task categories from 8 existing benchmarks, and prepare an evaluation framework to measure three types of model generalizations: to tasks from fully held-out categories, to held-out tasks from seen categories, and to held-out instances from seen tasks. Through the lens of this framework, we first present insights about instruction-tuning decisions as applied to OPT-30B and further exploit these insights to train OPT-IML 30B and 175B, which are instruction-tuned versions of OPT. OPT-IML demonstrates all three generalization abilities at both scales on four different evaluation benchmarks with diverse tasks and input formats -- PromptSource, FLAN, Super-NaturalInstructions, and UnifiedSKG. Not only does it significantly outperform OPT on all benchmarks but is also highly competitive with existing models fine-tuned on each specific benchmark. We release OPT-IML at both scales, together with the OPT-IML Bench evaluation framework.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes the effect of instruction-tuning decisions on downstream task performance when scaling both model and benchmark sizes and releases OPT-IML at both scales, together with an evaluation framework to measure three types of model generalizations.'}",http://arxiv.org/pdf/2212.12017
-the effect of nlp (accelerated learning) on iranian efl learner’s listening comprehension,Chnour Khalandi,"the present study was an attempt to investigate the impact of Neuro-Linguistic Programming (NLP) strategies and specifically Accelerated Learning on Iranian EFL learner's listening comprehension and detailed listening. To fulfill the purpose of this study, a total number of 30 Iranian EFL learners studying as Cambridge ESOL KET students were selected through stratified sampling procedure. They were divided in two control and experimental groups with 10 and 20 learners in each group respectively. Each group took a VAK questionnaire, a pre –test and a treatment process of 12 sessions and a posttest. In the experimental group the teacher conducted a set of short videos and mixed methods based on Accelerated Learning as one of neuro linguistic programming techniques. In control group a set of limited short videos conducted. On the last session, the participants of both groups took a post- test to measure their achievement of listening comprehension and detailed listening. This research suggests that accelerated learning is much more dynamic and has a significant effect on listening comprehension, detailed listening and, different learning styles, which hypothesizes that all VAK learner's language processing follow one another in a strictly accelerated manner. Statistically, the effect of accelerated learning on visual, auditory and kinesthetic learner's listening comprehension is 87, 93 and 85 percent respectively.",,http://www.academypublication.com/ojs/index.php/tpls/article/download/tpls071111391148/1321
-how good are gpt models at machine translation? a comprehensive evaluation,Amr Hendy,"Generative Pre-trained Transformer (GPT) models have shown remarkable capabilities for natural language generation, but their performance for machine translation has not been thoroughly investigated. In this paper, we present a comprehensive evaluation of GPT models for machine translation, covering various aspects such as quality of different GPT models in comparison with state-of-the-art research and commercial systems, effect of prompting strategies, robustness towards domain shifts and document-level translation. We experiment with eighteen different translation directions involving high and low resource languages, as well as non English-centric translations, and evaluate the performance of three GPT models: ChatGPT, GPT3.5 (text-davinci-003), and text-davinci-002. Our results show that GPT models achieve very competitive translation quality for high resource languages, while having limited capabilities for low resource languages. We also show that hybrid approaches, which combine GPT models with other translation systems, can further enhance the translation quality. We perform comprehensive analysis and human evaluation to further understand the characteristics of GPT translations. We hope that our paper provides valuable insights for researchers and practitioners in the field and helps to better understand the potential and limitations of GPT models for translation.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a comprehensive evaluation of GPT models for machine translation, covering various aspects such as quality of different G PT models in comparison with state-of-the-art research and commercial systems, effect of prompting strategies, robustness towards domain shifts and document-level translation.'}",http://arxiv.org/pdf/2302.09210
-can chatgpt understand too? a comparative study on chatgpt and fine-tuned bert,Qihuang Zhong,"Recently, ChatGPT has attracted great attention, as it can generate fluent and high-quality responses to human inquiries. Several prior studies have shown that ChatGPT attains remarkable generation ability compared with existing models. However, the quantitative analysis of ChatGPT's understanding ability has been given little attention. In this report, we explore the understanding ability of ChatGPT by evaluating it on the most popular GLUE benchmark, and comparing it with 4 representative fine-tuned BERT-style models. We find that: 1) ChatGPT falls short in handling paraphrase and similarity tasks; 2) ChatGPT outperforms all BERT models on inference tasks by a large margin; 3) ChatGPT achieves comparable performance compared with BERT on sentiment analysis and question-answering tasks. Additionally, by combining some advanced prompting strategies, we show that the understanding ability of ChatGPT can be further improved.","{'model': 'tldr@v2.0.0', 'text': 'The understanding ability of ChatGPT is explored by evaluating it on the most popular GLUE benchmark, and comparing it with 4 representative fine-tuned BERT-style models, and it is found that it falls short in handling paraphrase and similarity tasks.'}",http://arxiv.org/pdf/2302.10198
-nlp based verification of a uml class model,Rauf Sajjad,"Robotized model time is a creating examination field. A significant number of representations are furthermore given where model checking has been associated for affirmation of various sorts of model. Such delineations are SAT based checks of UML/OCL models, embeded structures model affirmation, et cetera. In all these kind of employments, the complement is model checking. In later past, UML programming models are created from programming necessities conveyed in a trademark tongue, for instance, English by using NLP technique. Regardless, the diverse sorts of UML models delivered from consistent vernacular programming requirements specific using NLP approach have no technique for check as in the complete and correct models are done. It is a normal learning NLP techniques have been viably associated with make UML models as delineated above, in any case, in this paper, we address the issue of model checking and model affirmation by using NLP strategies. Such sort of changes require package of effort and time that makes the system of model affirmation wild and ambling. We used an approach for model watching that makes the technique of model checking straightforward and additionally the used philosophy should use the present resources used for generation of the UML class model.","{'model': 'tldr@v2.0.0', 'text': 'An approach for model watching is used that makes the technique of model checking straightforward and additionally the used philosophy should use the present resources used for generation of the UML class model.'}",
-the effect of metacognitive training and prompting on learning success in simulation‐based physics learning,Stephanie Moser,"Computer-based simulations are of particular interest to physics learning because they allow learners to actively manipulate graphical visualizations of complex phenomena. However, learning with simulations requires supportive elements to scaffold learners’ activities. Thus, our motivation was to investigate whether direct or indirect metacognitive scaffolding (i.e., training and prompting, respectively) or a combination of both during simulation-based learning leads to better learning outcomes compared to simulation-based learning alone. Employing a 2 × 2 factorial design (N = 99 Austrian secondary school students), we explored the role of metacognitive training prior to the use of simulations and the role of metacognitive prompting during the learning phase. Four different conditions were compared in relation to knowledge about the principle of energy conservation. A pre- and posttest assessing knowledge acquisition, metacognitive behavior, cognitive load, and mental effort was applied. At first glance, the results indicate no main effects for training and prompting, and no interaction effect. A closer look at the actual use of prompting after the empirical data were obtained, however, reveals a significant impact of the metacognitive strategies employed on performance in the posttest, with higher scores in the condition where appropriate prompting was used. This result emphasizes the importance of guidance during simulation-based learning.",,
-teaching young children with special needs and their peers to play board games: effects of a least to most prompting procedure to increase independent performance,Janet Davis-Temple,,"{'model': 'tldr@v2.0.0', 'text': 'After teaching the boardgame steps using a systematic prompting strategy, the participants demonstrated increases in the performance of board game steps and game-related on-task behavior.'}",https://europepmc.org/articles/pmc4711729?pdf=render
-development of a digital game-based learning system with graduated prompting strategy for math course,Kai-Hsiang Yang,"Many studies indicated that digital game-based learning can enhance students' learning motivation effectively, but it also needs appropriate tools or strategies to improve the learning achievement. The purpose of this study is to develop a digital game-based learning system with a graduated prompting strategy, so that students can consolidate concepts and skills of problem-solving, and further strength retention and transfer by graduated prompting strategy progressively and repeatedly. This study investigates the influence of students' learning motivation and achievement by using the proposed digital game-based learning system in the math courses.","{'model': 'tldr@v2.0.0', 'text': 'The purpose of this study is to develop a digital game-based learning system with a graduated prompting strategy, so that students can consolidate concepts and skills of problem-solving, and further strength retention and transfer by graduating prompting strategy progressively and repeatedly.'}",
-pedagogical strategies to increase students’ engagement and motivation,C. De Grandi,,,
-impact of covid-19 on indian smes and survival strategies,D. Baragde,"India has been affected by the worldwide COVID-19 pandemic, which is dispensing two sorts of stuns on nations: health stun and a monetary stun. Given the idea of the infection, which is exceptionally infectious, the ways to contain the spread incorporate strategy activities, for example, burden of social removing, self-seclusion at home, closure of foundations and offices, limitations on versatility, and even lock-down of a whole nation. These activities can prompt desperate ramifications for economies around the globe. In other words, effective control of the infection requires the economy of a nation to stop its typical working.",,
-naijasenti: a nigerian twitter sentiment corpus for multilingual sentiment analysis,Shamsuddeen Hassan Muhammad,"Sentiment analysis is one of the most widely studied applications in NLP, but most work focuses on languages with large amounts of data. We introduce the first large-scale human-annotated Twitter sentiment dataset for the four most widely spoken languages in Nigeria—Hausa, Igbo, Nigerian-Pidgin, and Yorùbá—consisting of around 30,000 annotated tweets per language, including a significant fraction of code-mixed tweets. We propose text collection, filtering, processing and labeling methods that enable us to create datasets for these low-resource languages. We evaluate a range of pre-trained models and transfer strategies on the dataset. We find that language-specific models and language-adaptive fine-tuning generally perform best. We release the datasets, trained models, sentiment lexicons, and code to incentivize research on sentiment analysis in under-represented languages.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces the first large-scale human-annotated Twitter sentiment dataset for Nigeria—Hausa, Igbo, Nigerian-Pidgin, and Yorùbá—consisting of around 30,000 annotated tweets per language, including a significant fraction of code-mixed tweets.'}",
-enabling large language models to generate text with citations,Tianyu Gao,"Large language models (LLMs) have emerged as a widely-used tool for information seeking, but their generated outputs are prone to hallucination. In this work, our aim is to allow LLMs to generate text with citations, improving their factual correctness and verifiability. Existing work mainly relies on commercial search engines and human evaluation, making it challenging to reproduce and compare different modeling approaches. We propose ALCE, the first benchmark for Automatic LLMs' Citation Evaluation. ALCE collects a diverse set of questions and retrieval corpora and requires building end-to-end systems to retrieve supporting evidence and generate answers with citations. We develop automatic metrics along three dimensions -- fluency, correctness, and citation quality -- and demonstrate their strong correlation with human judgements. Our experiments with state-of-the-art LLMs and novel prompting strategies show that current systems have considerable room for improvement -- For example, on the ELI5 dataset, even the best models lack complete citation support 50% of the time. Our analyses further highlight promising future directions, including developing better retrievers, advancing long-context LLMs, and improving the ability to synthesize information from multiple sources.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes ALCE, the first benchmark for Automatic LLMs' Citation Evaluation, and develops automatic metrics along three dimensions -- fluency, correctness, and citation quality -- and demonstrates their strong correlation with human judgements.""}",http://arxiv.org/pdf/2305.14627
-"teaching assistants, neuro-linguistic programming (nlp) and special educational needs: ‘reframing’ the learning experience for students with mild sen",Voldis Kudliskis,"This study examines how an understanding of two NLP concepts, the meta-model of language and the implementation of reframing, could be used to help teaching assistants enhance class-based interactions with students with mild SEN. Participants (students) completed a pre-intervention and a post-intervention ‘Beliefs About my Learning Experiences Scale’ (BALES). The BALES has good internal consistency (Cronbach’s α = .85) and moderate test–retest reliability (r = .75), indicating that the BALES is a psychometrically sound instrument for assessing beliefs about learning experiences. The intervention group were exposed to positive verbal influencing strategies via ‘reframing’ over an intervention period of 12 weeks from TAs who received training in the use of these specific aspects of NLP. Following the intervention, semi-structured interviews were conducted with TAs to elicit information about the perceived utility of the NLP strategies. Semi-structured interviews were also conducted with the students to examine their experiences and perceptions of their learning post intervention. A paired samples t-test failed to reveal a statistically reliable difference between average pre-intervention BALES scores (M = 99.17, SD = 12.76) and post-intervention BALES scores (M = 105.75, SD = 15.21) that students with mild SEN achieved t(11) = 5.02, p = 1.78, α = .10. However, the qualitative responses collected from semi-structured interviews with TAs and with students do suggest that the use of specific elements of NLP, namely the meta-model of language and (embedded) reframing may be helpful, at some intrinsic psychological level, in challenging concepts that SEN students have about their experiences in education.",,
-how can recommender systems benefit from large language models: a survey,Jianghao Lin,"Recommender systems (RS) play important roles to match users' information needs for Internet applications. In natural language processing (NLP) domains, large language model (LLM) has shown astonishing emergent abilities (e.g., instruction following, reasoning), thus giving rise to the promising research direction of adapting LLM to RS for performance enhancements and user experience improvements. In this paper, we conduct a comprehensive survey on this research direction from an application-oriented view. We first summarize existing research works from two orthogonal perspectives: where and how to adapt LLM to RS. For the""WHERE""question, we discuss the roles that LLM could play in different stages of the recommendation pipeline, i.e., feature engineering, feature encoder, scoring/ranking function, and pipeline controller. For the""HOW""question, we investigate the training and inference strategies, resulting in two fine-grained taxonomy criteria, i.e., whether to tune LLMs or not, and whether to involve conventional recommendation model (CRM) for inference. Detailed analysis and general development trajectories are provided for both questions, respectively. Then, we highlight key challenges in adapting LLM to RS from three aspects, i.e., efficiency, effectiveness, and ethics. Finally, we summarize the survey and discuss the future prospects. We also actively maintain a GitHub repository for papers and other related resources in this rising direction: https://github.com/CHIANGEL/Awesome-LLM-for-RecSys.","{'model': 'tldr@v2.0.0', 'text': 'A comprehensive survey on this research direction of adapting LLM to RS for performance enhancements and user experience improvements is conducted from an application-oriented view.'}",http://arxiv.org/pdf/2306.05817
-a stream computing approach towards scalable nlp,X. Artola,"Computational power needs have grown dramatically in recent years. This is also the case in many language processing tasks, due to overwhelming quantities of textual information that must be processed in a reasonable time frame. This scenario has led to a paradigm shift in the computing architectures and large-scale data processing strategies used in the NLP field. In this paper we describe a series of experiments carried out in the context of the NewsReader project with the goal of analyzing the scaling capabilities of the language processing pipeline used in it. We explore the use of Storm in a new approach for scalable distributed language processing across multiple machines and evaluate its effectiveness and efficiency when processing documents on a medium and large scale. The experiments have shown that there is a big room for improvement regarding language processing performance when adopting parallel architectures, and that we might expect even better results with the use of large clusters with many processing nodes.","{'model': 'tldr@v2.0.0', 'text': 'The experiments have shown that there is a big room for improvement regarding language processing performance when adopting parallel architectures, and that there might expect even better results with the use of large clusters with many processing nodes.'}",
-expanding pretrained models to thousands more languages via lexicon-based adaptation,Xinyi Wang,"The performance of multilingual pretrained models is highly dependent on the availability of monolingual or parallel text present in a target language. Thus, the majority of the world’s languages cannot benefit from recent progress in NLP as they have no or limited textual data. To expand possibilities of using NLP technology in these under-represented languages, we systematically study strategies that relax the reliance on conventional language resources through the use of bilingual lexicons, an alternative resource with much better language coverage. We analyze different strategies to synthesize textual or labeled data using lexicons, and how this data can be combined with monolingual or parallel text when available. For 19 under-represented languages across 3 tasks, our methods lead to consistent improvements of up to 5 and 15 points with and without extra monolingual text respectively. Overall, our study highlights how NLP methods can be adapted to thousands more languages that are under-served by current technology.","{'model': 'tldr@v2.0.0', 'text': 'This study systematically study strategies that relax the reliance on conventional language resources through the use of bilingual lexicons, an alternative resource with much better language coverage and highlights how NLP methods can be adapted to thousands more languages that are under-served by current technology.'}",http://arxiv.org/pdf/2203.09435
-vision-and-language pretrained models: a survey,Siqu Long,"Pretrained models have produced great success in both Computer Vision (CV) and Natural Language Processing (NLP). This progress leads to learning joint representations of vision and language pretraining by feeding visual and linguistic contents into a multi-layer transformer, Visual-Language Pretrained Models (VLPMs). In this paper, we present an overview of the major advances achieved in VLPMs for producing joint representations of vision and language. As the preliminaries, we briefly describe the general task definition and genetic architecture of VLPMs. We first discuss the language and vision data encoding methods and then present the mainstream VLPM structure as the core content. We further summarise several essential pretraining and fine-tuning strategies. Finally, we highlight three future directions for both CV and NLP researchers to provide insightful guidance.","{'model': 'tldr@v2.0.0', 'text': 'An overview of the major advances achieved in VLPMs for producing joint representations of vision and language and highlights three future directions for both CV and NLP researchers to provide insightful guidance.'}",http://arxiv.org/pdf/2204.07356
-a survey of active learning for natural language processing,Zhisong Zhang,"In this work, we provide a literature review of active learning (AL) for its applications in natural language processing (NLP). In addition to a fine-grained categorization of query strategies, we also investigate several other important aspects of applying AL to NLP problems. These include AL for structured prediction tasks, annotation cost, model learning (especially with deep neural models), and starting and stopping AL. Finally, we conclude with a discussion of related topics and future directions.","{'model': 'tldr@v2.0.0', 'text': 'In addition to a fine-grained categorization of query strategies, several other important aspects of applying AL to NLP problems are investigated, including AL for structured prediction tasks, annotation cost, model learning, and starting and stopping AL.'}",http://arxiv.org/pdf/2210.10109
-covid-19 vaccine hesitancy: analysing twitter to identify barriers to vaccination in a low uptake region of the uk,Kate Lanyi,"To facilitate effective targeted COVID-19 vaccination strategies, it is important to understand reasons for vaccine hesitancy where uptake is low. Artificial intelligence (AI) techniques offer an opportunity for real-time analysis of public attitudes, sentiments, and key discussion topics from sources of soft-intelligence, including social media data. In this work, we explore the value of soft-intelligence, leveraged using AI, as an evidence source to support public health research. As a case study, we deployed a natural language processing (NLP) platform to rapidly identify and analyse key barriers to vaccine uptake from a collection of geo-located tweets from London, UK. We developed a search strategy to capture COVID-19 vaccine related tweets, identifying 91,473 tweets between 30 November 2020 and 15 August 2021. The platform's algorithm clustered tweets according to their topic and sentiment, from which we extracted 913 tweets from the top 12 negative sentiment topic clusters. These tweets were extracted for further qualitative analysis. We identified safety concerns; mistrust of government and pharmaceutical companies; and accessibility issues as key barriers limiting vaccine uptake. Our analysis also revealed widespread sharing of vaccine misinformation amongst Twitter users. This study further demonstrates that there is promising utility for using off-the-shelf NLP tools to leverage insights from social media data to support public health research. Future work to examine where this type of work might be integrated as part of a mixed-methods research approach to support local and national decision making is suggested.","{'model': 'tldr@v2.0.0', 'text': 'There is promising utility for using off-the-shelf NLP tools to leverage insights from social media data to support public health research, and safety concerns; mistrust of government and pharmaceutical companies; and accessibility issues as key barriers limiting vaccine uptake are identified.'}",https://www.frontiersin.org/articles/10.3389/fdgth.2021.804855/pdf
-diagnostic reasoning prompts reveal the potential for large language model interpretability in medicine,Thomas Savage,"One of the major barriers to using large language models (LLMs) in medicine is the perception they use uninterpretable methods to make clinical decisions that are inherently different from the cognitive processes of clinicians. In this manuscript we develop novel diagnostic reasoning prompts to study whether LLMs can perform clinical reasoning to accurately form a diagnosis. We find that GPT4 can be prompted to mimic the common clinical reasoning processes of clinicians without sacrificing diagnostic accuracy. This is significant because an LLM that can use clinical reasoning to provide an interpretable rationale offers physicians a means to evaluate whether LLMs can be trusted for patient care. Novel prompting methods have the potential to expose the black box of LLMs, bringing them one step closer to safe and effective use in medicine.","{'model': 'tldr@v2.0.0', 'text': 'Novel diagnostic reasoning prompts are developed to study whether LLMs can perform clinical reasoning to accurately form a diagnosis and it is found that GPT4 can be prompted to mimic the common clinical reasoning processes of clinicians without sacrificing diagnostic accuracy.'}",https://arxiv.org/pdf/2308.06834
-"prompt, plan, perform: llm-based humanoid control via quantized imitation learning",Jingkai Sun,"In recent years, reinforcement learning and imitation learning have shown great potential for controlling humanoid robots' motion. However, these methods typically create simulation environments and rewards for specific tasks, resulting in the requirements of multiple policies and limited capabilities for tackling complex and unknown tasks. To overcome these issues, we present a novel approach that combines adversarial imitation learning with large language models (LLMs). This innovative method enables the agent to learn reusable skills with a single policy and solve zero-shot tasks under the guidance of LLMs. In particular, we utilize the LLM as a strategic planner for applying previously learned skills to novel tasks through the comprehension of task-specific prompts. This empowers the robot to perform the specified actions in a sequence. To improve our model, we incorporate codebook-based vector quantization, allowing the agent to generate suitable actions in response to unseen textual commands from LLMs. Furthermore, we design general reward functions that consider the distinct motion features of humanoid robots, ensuring the agent imitates the motion data while maintaining goal orientation without additional guiding direction approaches or policies. To the best of our knowledge, this is the first framework that controls humanoid robots using a single learning policy network and LLM as a planner. Extensive experiments demonstrate that our method exhibits efficient and adaptive ability in complicated motion tasks.","{'model': 'tldr@v2.0.0', 'text': 'This is the first framework that controls humanoid robots using a single learning policy network and LLM as a planner and designs general reward functions that consider the distinct motion features of humanoid robots, ensuring the agent imitates the motion data while maintaining goal orientation.'}",https://arxiv.org/pdf/2309.11359
-make llm a testing expert: bringing human-like interaction to mobile gui testing via functionality-aware decisions,Zhe Liu,"Automated Graphical User Interface (GUI) testing plays a crucial role in ensuring app quality, especially as mobile applications have become an integral part of our daily lives. Despite the growing popularity of learning-based techniques in automated GUI testing due to their ability to generate human-like interactions, they still suffer from several limitations, such as low testing coverage, inadequate generalization capabilities, and heavy reliance on training data. Inspired by the success of Large Language Models (LLMs) like ChatGPT in natural language understanding and question answering, we formulate the mobile GUI testing problem as a Q&A task. We propose GPTDroid, asking LLM to chat with the mobile apps by passing the GUI page information to LLM to elicit testing scripts, and executing them to keep passing the app feedback to LLM, iterating the whole process. Within this framework, we have also introduced a functionality-aware memory prompting mechanism that equips the LLM with the ability to retain testing knowledge of the whole process and conduct long-term, functionality-based reasoning to guide exploration. We evaluate it on 93 apps from Google Play and demonstrate that it outperforms the best baseline by 32% in activity coverage, and detects 31% more bugs at a faster rate. Moreover, GPTDroid identify 53 new bugs on Google Play, of which 35 have been confirmed and fixed.","{'model': 'tldr@v2.0.0', 'text': 'GPTDroid is proposed, a functionality-aware memory prompting mechanism that equips the Large Language Models (LLM) with the ability to retain testing knowledge of the whole process and conduct long-term, functionality-based reasoning to guide exploration.'}",
-boosting static resource leak detection via llm-based resource-oriented intention inference,Chong Wang,"Resource leaks, caused by resources not being released after acquisition, often lead to performance issues and system crashes. Existing static detection techniques rely on mechanical matching of predefined resource acquisition/release APIs, posing challenges to their effectiveness, including completeness of predefined APIs, identification of reachability validation, and analysis complexity. To overcome these challenges, we propose InferROI, a novel approach that leverages large language models (LLMs) to directly infer resource-oriented intentions (acquisition, release, and reachability validation) in code, based on resource management knowledge and code context understanding, rather than mechanical API matching. InferROI uses a prompt to instruct the LLM in inferring involved intentions from a given code snippet, which are then translated into formal expressions. By aggregating these inferred intentions, InferROI utilizes a lightweight static-analysis based algorithm to analyze control-flow paths extracted from the code, thereby detecting resource leaks. We evaluate InferROI on Java program and investigate its effectiveness in both resource-oriented intention inference and resource leak detection. Experimental results demonstrate that InferROI achieves a precision of 74.6% and a recall of 81.8% in intention inference on 172 code snippets from the DroidLeaks dataset. Additionally, InferROI covers a significant portion of concerned Android resources listed in the dataset. When applied to 86 bugs from the DroidLeaks dataset, InferROI exhibits a high bug detection rate (53.5%) and a low false alarm rate (8.1%) compared to eight baseline detectors. Moreover, we apply InferROI to resource leak detection in 100 methods from real-world open-source projects, where it identifies 12 unknown resource leak bugs, with 7 of them being confirmed by developers.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes InferROI, a novel approach that leverages large language models (LLMs) to directly infer resource-oriented intentions (acquisition, release, and reachability validation) in code, based on resource management knowledge and code context understanding, rather than mechanical API matching.'}",
-llm-funcmapper: function identification for interpreting complex clauses in building codes via llm,Zhe Zheng,"As a vital stage of automated rule checking (ARC), rule interpretation of regulatory texts requires considerable effort. However, interpreting regulatory clauses with implicit properties or complex computational logic is still challenging due to the lack of domain knowledge and limited expressibility of conventional logic representations. Thus, LLM-FuncMapper, an approach to identifying predefined functions needed to interpret various regulatory clauses based on the large language model (LLM), is proposed. First, by systematically analysis of building codes, a series of atomic functions are defined to capture shared computational logics of implicit properties and complex constraints, creating a database of common blocks for interpreting regulatory clauses. Then, a prompt template with the chain of thought is developed and further enhanced with a classification-based tuning strategy, to enable common LLMs for effective function identification. Finally, the proposed approach is validated with statistical analysis, experiments, and proof of concept. Statistical analysis reveals a long-tail distribution and high expressibility of the developed function database, with which almost 100% of computer-processible clauses can be interpreted and represented as computer-executable codes. Experiments show that LLM-FuncMapper achieve promising results in identifying relevant predefined functions for rule interpretation. Further proof of concept in automated rule interpretation also demonstrates the possibility of LLM-FuncMapper in interpreting complex regulatory clauses. To the best of our knowledge, this study is the first attempt to introduce LLM for understanding and interpreting complex regulatory clauses, which may shed light on further adoption of LLM in the construction domain.","{'model': 'tldr@v2.0.0', 'text': 'This study is the first attempt to introduce LLM for understanding and interpreting complex regulatory clauses, which may shed light on further adoption of LLM in the construction domain.'}",https://arxiv.org/pdf/2308.08728
-cash transaction booking via retrieval augmented llm,Xiaoli Zhang,"In large corporations, millions of cash transactions are booked via cash management software (CMS) per month. Most CMS systems adopt a key-word (search string) based matching logic for booking, which checks if the cash transaction description contains a specific search string and books the transaction to an appropriate general ledger account (GL-account) according to a booking rule. However, due to the free-text nature of transaction description and the diversity of cash transactions, CMS systems often fail due to data corruption (truncation, insertions, spelling errors), paraphrasing, and lack of reusable key word in the description, requiring significant manual intervention by accountants. Month over month, accountants manually handle CMS booking failures in spreadsheets. We present two machine learning models, a GL-account classification model and a search string extraction model, to alleviate this manual process. These two models, backed by retrieval augmented large language models, can automate booking for a substantial portion of the manual transactions. Our approach is robust to common data issues in transaction description. Unlike typical deep-learning models, our models are interpretable and explainable. For GL-account classification, our approach has an accuracy close to human experts. For search string extraction, compared to other methods such as fine-tuning transformers for extraction tasks, our approach produces reliable results closer to accountants.","{'model': 'tldr@v2.0.0', 'text': 'Two machine learning models are presented, a GL-account classification model and a search string extraction model, which can automate booking for a substantial portion of the manual transactions of CMS systems.'}",
-resolving the imbalance issue in hierarchical disciplinary topic inference via llm-based data augmentation,Xunxin Cai,"In addressing the imbalanced issue of data within the realm of Natural Language Processing, text data augmentation methods have emerged as pivotal solutions. This data imbalance is prevalent in the research proposals submitted during the funding application process. Such imbalances, resulting from the varying popularity of disciplines or the emergence of interdisciplinary studies, significantly impede the precision of downstream topic models that deduce the affiliated disciplines of these proposals. At the data level, proposals penned by experts and scientists are inherently complex technological texts, replete with intricate terminologies, which augmenting such specialized text data poses unique challenges. At the system level, this, in turn, compromises the fairness of AI-assisted reviewer assignment systems, which raises a spotlight on solving this issue. This study leverages large language models (Llama V1) as data generators to augment research proposals categorized within intricate disciplinary hierarchies, aiming to rectify data imbalances and enhance the equity of expert assignments. We first sample within the hierarchical structure to find the under-represented class. Then we designed a prompt for keyword-based research proposal generation. Our experiments attests to the efficacy of the generated data, demonstrating that research proposals produced using the prompts can effectively address the aforementioned issues and generate high quality scientific text data, thus help the model overcome the imbalanced issue.","{'model': 'tldr@v2.0.0', 'text': 'This study leverages large language models (Llama V1) as data generators to augment research proposals categorized within intricate disciplinary hierarchies, aiming to rectify data imbalances and enhance the equity of expert assignments.'}",https://arxiv.org/pdf/2310.05318
-workshop on large language models' interpretability and trustworthiness (llmit),Tulika Saha,"Large language models (LLMs), when scaled from millions to billions of parameters, have been demonstrated to exhibit the so-called 'emergence' effect, in that they are not only able to produce semantically correct and coherent text, but are also able to adapt themselves surprisingly well with small changes in contexts supplied as inputs (commonly called prompts). Despite producing semantically coherent and potentially relevant text for a given context, LLMs are vulnerable to yield incorrect information. This misinformation generation, or the so-called hallucination problem of an LLM, gets worse when an adversary manipulates the prompts to their own advantage, e.g., generating false propaganda to disrupt communal harmony, generating false information to trap consumers with target consumables etc. Not only does the consumption of an LLM-generated hallucinated content by humans pose societal threats, such misinformation, when used as prompts, may lead to detrimental effects for in-context learning (also known as few-shot prompt learning). With reference to the above-mentioned problems of LLM usage, we argue that it is necessary to foster research on topics related to not only identifying misinformation from LLM-generated content, but also to mitigate the propagation effects of this generated misinformation on downstream predictive tasks thus leading to more robust and effective leveraging in-context learning.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that it is necessary to foster research on topics related to not only identifying misinformation from LLM-generated content, but also to mitigate the propagation effects of this generated misinformation on downstream predictive tasks thus leading to more robust and effective leveraging in-context learning.'}",https://dl.acm.org/doi/pdf/10.1145/3583780.3615311
-figstep: jailbreaking large vision-language models via typographic visual prompts,Yichen Gong,"Large vision-language models (VLMs) like GPT-4V represent an unprecedented revolution in the field of artificial intelligence (AI). Compared to single-modal large language models (LLMs), VLMs possess more versatile capabilities by incorporating additional modalities (e.g., images). Meanwhile, there's a rising enthusiasm in the AI community to develop open-source VLMs, such as LLaVA and MiniGPT4, which, however, have not undergone rigorous safety assessment. In this paper, to demonstrate that more modalities lead to unforeseen AI safety issues, we propose FigStep, a novel jailbreaking framework against VLMs. FigStep feeds harmful instructions into VLMs through the image channel and then uses benign text prompts to induce VLMs to output contents that violate common AI safety policies. Our experimental results show that FigStep can achieve an average attack success rate of 94.8% across 2 families of popular open-source VLMs, LLaVA and MiniGPT4 (a total of 5 VLMs). Moreover, we demonstrate that the methodology of FigStep can even jailbreak GPT-4V, which already leverages several system-level mechanisms to filter harmful queries. Above all, our experimental results reveal that VLMs are vulnerable to jailbreaking attacks, which highlights the necessity of novel safety alignments between visual and textual modalities.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results reveal that VLMs are vulnerable to jailbreaking attacks, which highlights the necessity of novel safety alignments between visual and textual modalities.'}",
-improving zero-shot visual question answering via large language models with reasoning question prompts,Yunshi Lan,"Zero-shot Visual Question Answering (VQA) is a prominent vision-language task that examines both the visual and textual understanding capability of systems in the absence of training data. Recently, by converting the images into captions, information across multi-modalities is bridged and Large Language Models (LLMs) can apply their strong zero-shot generalization capability to unseen questions. To design ideal prompts for solving VQA via LLMs, several studies have explored different strategies to select or generate question-answer pairs as the exemplar prompts, which guide LLMs to answer the current questions effectively. However, they totally ignore the role of question prompts. The original questions in VQA tasks usually encounter ellipses and ambiguity which require intermediate reasoning. To this end, we present Reasoning Question Prompts for VQA tasks, which can further activate the potential of LLMs in zero-shot scenarios. Specifically, for each question, we first generate self-contained questions as reasoning question prompts via an unsupervised question edition module considering sentence fluency, semantic integrity and syntactic invariance. Each reasoning question prompt clearly indicates the intent of the original question. This results in a set of candidate answers. Then, the candidate answers associated with their confidence scores acting as answer heuristics are fed into LLMs and produce the final answer. We evaluate reasoning question prompts on three VQA challenges, experimental results demonstrate that they can significantly improve the results of LLMs on zero-shot setting and outperform existing state-of-the-art zero-shot methods on three out of four data sets. Our source code is publicly released at https://github.com/ECNU-DASE-NLP/RQP.","{'model': 'tldr@v2.0.0', 'text': 'Reasoning Question Prompts for VQA tasks are presented, which can further activate the potential of Large Language Models in zero-shot scenarios and can significantly improve the results of LLMs on zero- shot setting and outperform existing state-of-the-art zero-Shot methods on three out of four data sets.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612389
-prompts of large language model for commanding power grid operation,Hanjiang Dong,"Large Language Models (LLMs) like ChatGPT can assist people’s general workflows, where the prompt is necessary to inspire the potential of LLMs to solve problems from specified or professional domains like robotics. In the electrical engineering subject or the electric power utility industry, experienced operators and professional experts monitor power grid operation statuses and interact with the grid via human commands on the screen, and components in the grid execute the commands to keep the complex grid safe and economical operation. In this process, human experts edit commands to operate the corresponding software. Human commands are the natural language that the LLM can process. The power grid is composed of generation, transmission, distribution, and other components. Therefore, we redesign the human-computer interaction frame between practitioners and the grid via recurrent prompts to apply the LLM to generate computer programming instructions from the multi-step natural language commands. The programming instruction is executed on system components after being confirmed or revised by human experts, and the quality of generated programs will be gradually improved through human feedback. The idea of this study is originally inspired by studies on controlling individual robotic components by ChatGPT. In the future, we will apply the designed prompt templates to drive the general LLM to generate desired samples which could be used to train an LLM professional in the domain knowledge of electrical engineering to operate multiple types of software for power grid operators.","{'model': 'tldr@v2.0.0', 'text': 'This study redesigns the human-computer interaction frame between practitioners and the grid via recurrent prompts to apply the LLM to generate computer programming instructions from the multi-step natural language commands for power grid operators.'}",
-psychologically-informed chain-of-thought prompts for metaphor understanding in large language models,Ben Prystawski,"Probabilistic models of language understanding are valuable tools for investigating human language use. However, they need to be hand-designed for a particular domain. In contrast, large language models (LLMs) are trained on text that spans a wide array of domains, but they lack the structure and interpretability of probabilistic models. In this paper, we use chain-of-thought prompts to introduce structures from probabilistic models into LLMs. We explore this approach in the case of metaphor understanding. Our chain-of-thought prompts lead language models to infer latent variables and reason about their relationships in order to choose appropriate paraphrases for metaphors. The latent variables and relationships chosen are informed by theories of metaphor understanding from cognitive psychology. We apply these prompts to the two largest versions of GPT-3 and show that they can improve performance in a paraphrase selection task.","{'model': 'tldr@v2.0.0', 'text': 'This paper uses chain-of-thought prompts to introduce structures from probabilistic models into large language models, and applies these prompts to the two largest versions of GPT-3 and shows that they can improve performance in a paraphrase selection task.'}",http://arxiv.org/pdf/2209.08141
-can large language models explain themselves? a study of llm-generated self-explanations,Shiyuan Huang,"Large language models (LLMs) such as ChatGPT have demonstrated superior performance on a variety of natural language processing (NLP) tasks including sentiment analysis, mathematical reasoning and summarization. Furthermore, since these models are instruction-tuned on human conversations to produce""helpful""responses, they can and often will produce explanations along with the response, which we call self-explanations. For example, when analyzing the sentiment of a movie review, the model may output not only the positivity of the sentiment, but also an explanation (e.g., by listing the sentiment-laden words such as""fantastic""and""memorable""in the review). How good are these automatically generated self-explanations? In this paper, we investigate this question on the task of sentiment analysis and for feature attribution explanation, one of the most commonly studied settings in the interpretability literature (for pre-ChatGPT models). Specifically, we study different ways to elicit the self-explanations, evaluate their faithfulness on a set of evaluation metrics, and compare them to traditional explanation methods such as occlusion or LIME saliency maps. Through an extensive set of experiments, we find that ChatGPT's self-explanations perform on par with traditional ones, but are quite different from them according to various agreement metrics, meanwhile being much cheaper to produce (as they are generated along with the prediction). In addition, we identified several interesting characteristics of them, which prompt us to rethink many current model interpretability practices in the era of ChatGPT(-like) LLMs.","{'model': 'tldr@v2.0.0', 'text': ""Through an extensive set of experiments, it is found that ChatGPT's self-explanations perform on par with traditional ones, but are quite different from them according to various agreement metrics, meanwhile being much cheaper to produce (as they are generated along with the prediction).""}",
-fake news in sheep's clothing: robust fake news detection against llm-empowered style attacks,Jiaying Wu,"It is commonly perceived that online fake news and reliable news exhibit stark differences in writing styles, such as the use of sensationalist versus objective language. However, we emphasize that style-related features can also be exploited for style-based attacks. Notably, the rise of powerful Large Language Models (LLMs) has enabled malicious users to mimic the style of trustworthy news outlets at minimal cost. Our analysis reveals that LLM-camouflaged fake news content leads to substantial performance degradation of state-of-the-art text-based detectors (up to 38% decrease in F1 Score), posing a significant challenge for automated detection in online ecosystems. To address this, we introduce SheepDog, a style-agnostic fake news detector robust to news writing styles. SheepDog achieves this adaptability through LLM-empowered news reframing, which customizes each article to match different writing styles using style-oriented reframing prompts. By employing style-agnostic training, SheepDog enhances its resilience to stylistic variations by maximizing prediction consistency across these diverse reframings. Furthermore, SheepDog extracts content-focused veracity attributions from LLMs, where the news content is evaluated against a set of fact-checking rationales. These attributions provide supplementary information and potential interpretability that assist veracity prediction. On three benchmark datasets, empirical results show that SheepDog consistently yields significant improvements over competitive baselines and enhances robustness against LLM-empowered style attacks.","{'model': 'tldr@v2.0.0', 'text': 'SheepDog is introduced, a style-agnostic fake news detector robust to news writing styles that consistently yields significant improvements over competitive baselines and enhances robustness against LLM-empowered style attacks.'}",
-towards llm-based fact verification on news claims with a hierarchical step-by-step prompting method,Xuan Zhang,"While large pre-trained language models (LLMs) have shown their impressive capabilities in various NLP tasks, they are still under-explored in the misinformation domain. In this paper, we examine LLMs with in-context learning (ICL) for news claim verification, and find that only with 4-shot demonstration examples, the performance of several prompting methods can be comparable with previous supervised models. To further boost performance, we introduce a Hierarchical Step-by-Step (HiSS) prompting method which directs LLMs to separate a claim into several subclaims and then verify each of them via multiple questions-answering steps progressively. Experiment results on two public misinformation datasets show that HiSS prompting outperforms state-of-the-art fully-supervised approach and strong few-shot ICL-enabled baselines.","{'model': 'tldr@v2.0.0', 'text': 'A Hierarchical Step-by-Step (HiSS) prompting method is introduced which directs LLMs to separate a claim into several subclaims and then verify each of them via multiple questions-answering steps progressively.'}",https://arxiv.org/pdf/2310.00305
-cgsmp: controllable generative summarization via multimodal prompt,Qian Yong,"Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of a large language model (LLM), this advancement has resulted in more fluent and coherent Natural Language Generation, which has contributed to improved development in downstream tasks such as abstractive summarization. Despite the recent progress in LLM, hallucination has become a serious problem in NLG. Hallucination happens when language models generate nonsensical or unfaithful text, which will lead to severe problems with reliability and effectiveness. In this paper, we propose a novel approach called Controllable Generative Summarization via Multimodal Prompt (CGSMP), which uses entities extracted from content and images as multimodal prompt control signals, thereby reducing hallucination issues. Specifically, the proposed CGSMP consists of three main modules: (1) an image prefix module that obtains image representations; (2) a prompt encoder module that fusion entities and images as multimodal prompts; and (3) a pre-trained causal language model that fuses input and controllable prompt and serves as the backbone of the language model. Experimental results demonstrate that the proposed method significantly improves the quality of generated summaries compared to the state of the arts.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results demonstrate that the proposed CGSMP significantly improves the quality of generated summaries compared to the state of the arts, and serves as the backbone of the language model.'}",
-unified human-scene interaction via prompted chain-of-contacts,Zeqi Xiao,"Human-Scene Interaction (HSI) is a vital component of fields like embodied AI and virtual reality. Despite advancements in motion quality and physical plausibility, two pivotal factors, versatile interaction control and the development of a user-friendly interface, require further exploration before the practical application of HSI. This paper presents a unified HSI framework, UniHSI, which supports unified control of diverse interactions through language commands. This framework is built upon the definition of interaction as Chain of Contacts (CoC): steps of human joint-object part pairs, which is inspired by the strong correlation between interaction types and human-object contact regions. Based on the definition, UniHSI constitutes a Large Language Model (LLM) Planner to translate language prompts into task plans in the form of CoC, and a Unified Controller that turns CoC into uniform task execution. To facilitate training and evaluation, we collect a new dataset named ScenePlan that encompasses thousands of task plans generated by LLMs based on diverse scenarios. Comprehensive experiments demonstrate the effectiveness of our framework in versatile task execution and generalizability to real scanned scenes. The project page is at https://github.com/OpenRobotLab/UniHSI .","{'model': 'tldr@v2.0.0', 'text': 'A unified HSI framework, UniHSI, which supports unified control of diverse interactions through language commands, built upon the definition of interaction as Chain of Contacts (CoC): steps of human joint-object part pairs, which is inspired by the strong correlation between interaction types and human-object contact regions.'}",https://arxiv.org/pdf/2309.07918
-learning profitable nft image diffusions via multiple visual-policy guided reinforcement learning,Huiguo He,"We study the task of generating profitable Non-Fungible Token (NFT) images from user-input texts. Recent advances in diffusion models have shown great potential for image generation. However, existing works can fall short in generating visually-pleasing and highly-profitable NFT images, mainly due to the lack of 1) plentiful and fine-grained visual attribute prompts for an NFT image, and 2) effective optimization metrics for generating high-quality NFT images. To solve these challenges, we propose a Diffusion based generation framework with Multiple Visual-Policies as rewards (i.e., Diffusion-MVP) for NFT images. The proposed framework consists of a large language model (LLM), a diffusion-based image generator, and a series of visual rewards by design. First, the LLM enhances a basic human input (such as ""panda"") by generating more comprehensive NFT-style prompts that include specific visual attributes, such as ""panda with Ninja style and green background."" Second, the diffusion-based image generator is fine-tuned using a large-scale NFT dataset to capture fine-grained image styles and accessory compositions of popular NFT elements. Third, we further propose to utilize multiple visual-policies as optimization goals, including visual rarity levels, visual aesthetic scores, and CLIP-based text-image relevances. This design ensures that our proposed Diffusion-MVP is capable of minting NFT images with high visual quality and market value. To facilitate this research, we have collected the largest publicly available NFT image dataset to date, consisting of 1.5 million high-quality images with corresponding texts and market values. Extensive experiments including objective evaluations and user studies demonstrate that our framework can generate NFT images showing more visually engaging elements and higher market value, compared with state-of-the-art approaches.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments including objective evaluations and user studies demonstrate that the proposed Diffusion-MVP framework can generate NFT images showing more visually engaging elements and higher market value, compared with state-of-the-art approaches.'}",
-majority rule: better patching via self-consistency,Toufique Ahmed,"—Large Language models (LLMs) can be induced to solve non-trivial problems with “few-shot” prompts including illustrative problem-solution examples. Now if the few-shots also include “chain of thought” ( C oT ) explanations, which are of the form problem-explanation-solution , LLMs will generate a “explained” solution, and perform even better. Recently an exciting, substantially better technique, self-consistency [1] ( S - C ) has emerged, based on the intuition that there are many plausible explanations for the right solution; when the LLM is sampled repeatedly to generate a pool of explanation-solution pairs, for a given problem, the most frequently occurring solutions in the pool (ignoring the explanations ) tend to be even more likely to be correct!Unfortunately, the use of this highly-performant S - C (or even C oT ) approach in software engineering settings is hampered by the lack of explanations ; most software datasets lack explanations. In this paper, we describe an application of the S - C approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots. We achieve state-of-the art results, beating previous approaches to prompting-based program repair, on the MODIT dataset; we also find evidence suggesting that the correct commit messages are helping the LLM learn to produce better patches.","{'model': 'tldr@v2.0.0', 'text': 'An application of the S - C approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots is described, beating previous approaches to prompting-based program repair on the MODIT dataset.'}",https://arxiv.org/pdf/2306.00108
-demonstration of insightpilot: an llm-empowered automated data exploration system,Pingchuan Ma,"Exploring data is crucial in data analysis, as it helps users understand and interpret the data more effectively. However, performing effective data exploration requires in-depth knowledge of the dataset and expertise in data analysis techniques. Not being familiar with either can create obstacles that make the process time-consuming and overwhelming for data analysts. To address this issue, we introduce InsightPilot, an LLM (Large Language Model)-based, automated data exploration system designed to simplify the data exploration process. InsightPilot automatically selects appropriate analysis intents, such as understanding, summarizing, and explaining. Then, these analysis intents are concretized by issuing corresponding intentional queries (IQueries) to create a meaningful and coherent exploration sequence. In brief, an IQuery is an abstraction and automation of data analysis operations, which mimics the approach of data analysts and simplifies the exploration process for users. By employing an LLM to iteratively collaborate with a state-of-the-art insight engine via IQueries, InsightPilot is effective in analyzing real-world datasets, enabling users to gain valuable insights through natural language inquiries. We demonstrate the effectiveness of InsightPilot in a case study, showing how it can help users gain valuable insights from their datasets.","{'model': 'tldr@v2.0.0', 'text': 'InsightPilot, an LLM-based, automated data exploration system designed to simplify the data exploration process, is introduced, effective in analyzing real-world datasets, enabling users to gain valuable insights through natural language inquiries.'}",http://arxiv.org/pdf/2304.00477
-llm-assisted content analysis: using large language models to support deductive coding,Robert F. Chew,"Deductive coding is a widely used qualitative research method for determining the prevalence of themes across documents. While useful, deductive coding is often burdensome and time consuming since it requires researchers to read, interpret, and reliably categorize a large body of unstructured text documents. Large language models (LLMs), like ChatGPT, are a class of quickly evolving AI tools that can perform a range of natural language processing and reasoning tasks. In this study, we explore the use of LLMs to reduce the time it takes for deductive coding while retaining the flexibility of a traditional content analysis. We outline the proposed approach, called LLM-assisted content analysis (LACA), along with an in-depth case study using GPT-3.5 for LACA on a publicly available deductive coding data set. Additionally, we conduct an empirical benchmark using LACA on 4 publicly available data sets to assess the broader question of how well GPT-3.5 performs across a range of deductive coding tasks. Overall, we find that GPT-3.5 can often perform deductive coding at levels of agreement comparable to human coders. Additionally, we demonstrate that LACA can help refine prompts for deductive coding, identify codes for which an LLM is randomly guessing, and help assess when to use LLMs vs. human coders for deductive coding. We conclude with several implications for future practice of deductive coding and related research methods.","{'model': 'tldr@v2.0.0', 'text': 'Overall, it is found that GPT-3.5 can often perform deductive coding at levels of agreement comparable to human coders, and it is demonstrated that LACA can help refine prompts for deductedive coding, identify codes for which an LLM is randomly guessing, and help assess when to use LLMs vs. human codering.'}",http://arxiv.org/pdf/2306.14924
-faithful explanations of black-box nlp models using llm-generated counterfactuals,Y. Gat,"Causal explanations of the predictions of NLP systems are essential to ensure safety and establish trust. Yet, existing methods often fall short of explaining model predictions effectively or efficiently and are often model-specific. In this paper, we address model-agnostic explanations, proposing two approaches for counterfactual (CF) approximation. The first approach is CF generation, where a large language model (LLM) is prompted to change a specific text concept while keeping confounding concepts unchanged. While this approach is demonstrated to be very effective, applying LLM at inference-time is costly. We hence present a second approach based on matching, and propose a method that is guided by an LLM at training-time and learns a dedicated embedding space. This space is faithful to a given causal graph and effectively serves to identify matches that approximate CFs. After showing theoretically that approximating CFs is required in order to construct faithful explanations, we benchmark our approaches and explain several models, including LLMs with billions of parameters. Our empirical results demonstrate the excellent performance of CF generation models as model-agnostic explainers. Moreover, our matching approach, which requires far less test-time resources, also provides effective explanations, surpassing many baselines. We also find that Top-K techniques universally improve every tested method. Finally, we showcase the potential of LLMs in constructing new benchmarks for model explanation and subsequently validate our conclusions. Our work illuminates new pathways for efficient and accurate approaches to interpreting NLP systems.","{'model': 'tldr@v2.0.0', 'text': 'This paper addresses model-agnostic explanations, proposing two approaches for counterfactual (CF) approximation and presents a method that is guided by an LLM at training-time and learns a dedicated embedding space that effectively serves to identify matches that approximate CFs.'}",https://arxiv.org/pdf/2310.00603
-selfcheckgpt: zero-resource black-box hallucination detection for generative large language models,Potsawee Manakul,"Generative Large Language Models (LLMs) such as GPT-3 are capable of generating highly fluent responses to a wide variety of user prompts. However, LLMs are known to hallucinate facts and make non-factual statements which can undermine trust in their output. Existing fact-checking approaches either require access to the output probability distribution (which may not be available for systems such as ChatGPT) or external databases that are interfaced via separate, often complex, modules. In this work, we propose""SelfCheckGPT"", a simple sampling-based approach that can be used to fact-check the responses of black-box models in a zero-resource fashion, i.e. without an external database. SelfCheckGPT leverages the simple idea that if an LLM has knowledge of a given concept, sampled responses are likely to be similar and contain consistent facts. However, for hallucinated facts, stochastically sampled responses are likely to diverge and contradict one another. We investigate this approach by using GPT-3 to generate passages about individuals from the WikiBio dataset, and manually annotate the factuality of the generated passages. We demonstrate that SelfCheckGPT can: i) detect non-factual and factual sentences; and ii) rank passages in terms of factuality. We compare our approach to several baselines and show that our approach has considerably higher AUC-PR scores in sentence-level hallucination detection and higher correlation scores in passage-level factuality assessment compared to grey-box methods.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ""SelfCheckGPT"", a simple sampling-based approach that can be used to fact-check the responses of black-box models in a zero-resource fashion, i.e. without an external database, and demonstrates that it can detect non-factual and factual sentences and rank passages in terms of factuality.'}",https://arxiv.org/pdf/2303.08896
-toolkengpt: augmenting frozen language models with massive tools via tool embeddings,Shibo Hao,"Augmenting large language models (LLMs) with external tools has emerged as a promising approach to solving complex problems. However, traditional methods, which finetune LLMs with tool demonstration data, can be both costly and restricted to a predefined set of tools. Recent in-context learning paradigm alleviates these issues, but the limited context length only allows for a few shots of demonstrations, leading to suboptimal understandings of the tools. Moreover, when there are numerous tools to choose from, in-context learning could completely fail to work. In this paper, we propose an alternative approach, $\textbf{ToolkenGPT}$, which combines the benefits of both sides. Our approach represents each $\underline{tool}$ as a to$\underline{ken}$ ($\textit{toolken}$) and learns an embedding for it, enabling tool calls in the same way as generating a regular word token. Once a toolken is triggered, the LLM is prompted to complete arguments for the tool to execute. ToolkenGPT offers the flexibility to plug in an arbitrary number of tools by expanding the set of toolkens on the fly. In addition, it improves tool use by allowing extensive demonstration data for learning the toolken embeddings. In diverse domains, including numerical reasoning, knowledge-based question answering, and embodied plan generation, our approach effectively augments LLMs with tools and substantially outperforms various latest baselines. ToolkenGPT demonstrates the promising ability to use relevant tools from a large tool set in complex scenarios.","{'model': 'tldr@v2.0.0', 'text': 'The proposed ToolkenGPT offers the flexibility to plug in an arbitrary number of tools by expanding the set of toolkens on the fly and improves tool use by allowing extensive demonstration data for learning the toolken embeddings.'}",http://arxiv.org/pdf/2305.11554
-revisiting the plastic surgery hypothesis via large language models,Chun Xia,"Automated Program Repair (APR) aspires to automatically generate patches for an input buggy program. Traditional APR tools typically focus on specific bug types and fixes through the use of templates, heuristics, and formal specifications. However, these techniques are limited in terms of the bug types and patch variety they can produce. As such, researchers have designed various learning-based APR tools with recent work focused on directly using Large Language Models (LLMs) for APR. While LLM-based APR tools are able to achieve state-of-the-art performance on many repair datasets, the LLMs used for direct repair are not fully aware of the project-specific information such as unique variable or method names. The plastic surgery hypothesis is a well-known insight for APR, which states that the code ingredients to fix the bug usually already exist within the same project. Traditional APR tools have largely leveraged the plastic surgery hypothesis by designing manual or heuristic-based approaches to exploit such existing code ingredients. However, as recent APR research starts focusing on LLM-based approaches, the plastic surgery hypothesis has been largely ignored. In this paper, we ask the following question: How useful is the plastic surgery hypothesis in the era of LLMs? Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting. To this end, we propose FitRepair, which combines the direct usage of LLMs with two domain-specific fine-tuning strategies and one prompting strategy for more powerful APR. Our experiments on the widely studied Defects4j 1.2 and 2.0 datasets show that FitRepair fixes 89 and 44 bugs (substantially outperforming the best-performing baseline by 15 and 8), respectively, demonstrating a promising future of the plastic surgery hypothesis in the era of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting, and this paper proposes FitRepair, which combines the direct usage of LLMs with two domain-specific fine- tuning strategies and one prompting strategy for more powerful APR.'}",http://arxiv.org/pdf/2303.10494
-automatic calibration and error correction for large language models via pareto optimal self-supervision,Theodore Zhao,"Large language models (LLMs) have demonstrated remarkable capabilities out of box for a wide range of applications, yet accuracy still remains a major growth area, especially in mission-critical domains such as biomedicine. An effective method to calibrate the confidence level on LLM responses is essential to automatically detect errors and facilitate human-in-the-loop verification. An important source of calibration signals stems from expert-stipulated programmatic supervision, which is often available at low cost but has its own limitations such as noise and coverage. In this paper, we introduce a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every response, without any additional manual efforts. This is accomplished by learning a harmonizer model to align LLM output with other available supervision sources, which would assign higher risk scores to more uncertain LLM responses and facilitate error correction. Experiments on standard relation extraction tasks in biomedical and general domains demonstrate the promise of this approach, with our proposed risk scores highly correlated with the real error rate of LLMs. For the most uncertain test instances, dynamic prompting based on our proposed risk scores results in significant accuracy improvement for off-the-shelf LLMs, boosting GPT-3 results past state-of-the-art (SOTA) weak supervision and GPT-4 results past SOTA supervised results on challenging evaluation datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every response, without any additional manual efforts.'}",https://arxiv.org/pdf/2306.16564
-automatic calibration and error correction for generative large language models via pareto optimal self-supervision,Theodore Zhao,"Generative Large language models (LLMs) have demonstrated remarkable capabilities for a wide range of applications, but reducing ungrounded or erroneous responses remains a major growth area. Unlike task-specific models, there lack an effective method to calibrate the confidence level of LLM responses to indicate potential errors and facilitate human-in-the-loop verification. An important source of calibration stems from expert-stipulated programmatic supervision, which is often available at low cost but has its own limitations such as noise and coverage. In this paper, we introduce a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every LLM response, without any additional manual efforts. This is accomplished by learning a harmonizer model to align with LLM output as well as other weak supervision sources. The model assigns higher risk scores to more uncertain LLM responses and facilitate error correction. Experiments on standard relation extraction and classification tasks in biomedical and general domains demonstrate that the proposed risk score is highly correlated with the actual LLM error rate. By using a dynamic prompting strategy based on the risk score, we observed significant accuracy improvement for off-the-shelf LLMs, boosting GPT-3.5 results past state-of-the-art (SOTA) weak supervision model and GPT-4 results past SOTA supervised results on challenging evaluation datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a Pareto optimal self-supervision framework that can leverage available programmatic supervision to systematically calibrate LLM responses by producing a risk score for every LLM response, without any additional manual efforts.'}",
-drivegpt4: interpretable end-to-end autonomous driving via large language model,Zhenhua Xu,"In the past decade, autonomous driving has experienced rapid development in both academia and industry. However, its limited interpretability remains a significant unsolved problem, severely hindering autonomous vehicle commercialization and further development. Previous approaches utilizing small language models have failed to address this issue due to their lack of flexibility, generalization ability, and robustness. Recently, multimodal large language models (LLMs) have gained considerable attention from the research community for their capability to process and reason non-text data (e.g., images and videos) by text. In this paper, we present DriveGPT4, an interpretable end-to-end autonomous driving system utilizing LLMs. DriveGPT4 is capable of interpreting vehicle actions and providing corresponding reasoning, as well as answering diverse questions posed by human users for enhanced interaction. Additionally, DriveGPT4 predicts vehicle low-level control signals in an end-to-end fashion. These capabilities stem from a customized visual instruction tuning dataset specifically designed for autonomous driving. To the best of our knowledge, DriveGPT4 is the first work focusing on interpretable end-to-end autonomous driving. When evaluated on multiple tasks alongside conventional methods and video understanding LLMs, DriveGPT4 demonstrates superior qualitative and quantitative performance. Additionally, DriveGPT4 can be generalized in a zero-shot fashion to accommodate more unseen scenarios. The project page is available at https://tonyxuqaq.github.io/projects/DriveGPT4/ .","{'model': 'tldr@v2.0.0', 'text': 'DriveGPT4 is an interpretable end-to-end autonomous driving system utilizing multimodal large language models that is capable of interpreting vehicle actions and providing corresponding reasoning, as well as answering diverse questions posed by human users for enhanced interaction.'}",https://arxiv.org/pdf/2310.01412
-blsp: bootstrapping language-speech pre-training via behavior alignment of continuation writing,Chen Wang,"The emergence of large language models (LLMs) has sparked significant interest in extending their remarkable language capabilities to speech. However, modality alignment between speech and text still remains an open problem. Current solutions can be categorized into two strategies. One is a cascaded approach where outputs (tokens or states) of a separately trained speech recognition system are used as inputs for LLMs, which limits their potential in modeling alignment between speech and text. The other is an end-to-end approach that relies on speech instruction data, which is very difficult to collect in large quantities. In this paper, we address these issues and propose the BLSP approach that Bootstraps Language-Speech Pre-training via behavior alignment of continuation writing. We achieve this by learning a lightweight modality adapter between a frozen speech encoder and an LLM, ensuring that the LLM exhibits the same generation behavior regardless of the modality of input: a speech segment or its transcript. The training process can be divided into two steps. The first step prompts an LLM to generate texts with speech transcripts as prefixes, obtaining text continuations. In the second step, these continuations are used as supervised signals to train the modality adapter in an end-to-end manner. We demonstrate that this straightforward process can extend the capabilities of LLMs to speech, enabling speech recognition, speech translation, spoken language understanding, and speech conversation, even in zero-shot cross-lingual scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes the BLSP approach that Bootstraps Language-Speech Pre-training via behavior alignment of continuation writing and demonstrates that this straightforward process can extend the capabilities of LLMs to speech, enabling speech recognition, speech translation, spoken language understanding, and speech conversation, even in zero-shot cross-lingual scenarios.'}",https://arxiv.org/pdf/2309.00916
-gptuner: a manual-reading database tuning system via gpt-guided bayesian optimization,Jiale Lao,"Modern database management systems (DBMS) expose hundreds of configurable knobs to control system behaviours. Determining the appropriate values for these knobs to improve DBMS performance is a long-standing problem in the database community. As there is an increasing number of knobs to tune and each knob could be in continuous or categorical values, manual tuning becomes impractical. Recently, automatic tuning systems using machine learning methods have shown great potentials. However, existing approaches still incur significant tuning costs or only yield sub-optimal performance. This is because they either ignore the extensive domain knowledge available (e.g., DBMS manuals and forum discussions) and only rely on the runtime feedback of benchmark evaluations to guide the optimization, or they utilize the domain knowledge in a limited way. Hence, we propose GPTuner, a manual-reading database tuning system. Firstly, we develop a Large Language Model (LLM)-based pipeline to collect and refine heterogeneous knowledge, and propose a prompt ensemble algorithm to unify a structured view of the refined knowledge. Secondly, using the structured knowledge, we (1) design a workload-aware and training-free knob selection strategy, (2) develop a search space optimization technique considering the value range of each knob, and (3) propose a Coarse-to-Fine Bayesian Optimization Framework to explore the optimized space. Finally, we evaluate GPTuner under different benchmarks (TPC-C and TPC-H), metrics (throughput and latency) as well as DBMS (PostgreSQL and MySQL). Compared to the state-of-the-art approaches, GPTuner identifies better configurations in 16x less time on average. Moreover, GPTuner achieves up to 30% performance improvement (higher throughput or lower latency) over the best-performing alternative.","{'model': 'tldr@v2.0.0', 'text': 'GPTuner, a manual-reading database tuning system that designs a workload-aware and training-free knob selection strategy, develops a search space optimization technique considering the value range of each knob, and proposes a Coarse-to-Fine Bayesian Optimization Framework to explore the optimized space.'}",
-instructprotein: aligning human and protein language via knowledge instruction,Zeyuan Wang,"Large Language Models (LLMs) have revolutionized the field of natural language processing, but they fall short in comprehending biological sequences such as proteins. To address this challenge, we propose InstructProtein, an innovative LLM that possesses bidirectional generation capabilities in both human and protein languages: (i) taking a protein sequence as input to predict its textual function description and (ii) using natural language to prompt protein sequence generation. To achieve this, we first pre-train an LLM on both protein and natural language corpora, enabling it to comprehend individual languages. Then supervised instruction tuning is employed to facilitate the alignment of these two distinct languages. Herein, we introduce a knowledge graph-based instruction generation framework to construct a high-quality instruction dataset, addressing annotation imbalance and instruction deficits in existing protein-text corpus. In particular, the instructions inherit the structural relations between proteins and function annotations in knowledge graphs, which empowers our model to engage in the causal modeling of protein functions, akin to the chain-of-thought processes in natural languages. Extensive experiments on bidirectional protein-text generation tasks show that InstructProtein outperforms state-of-the-art LLMs by large margins. Moreover, InstructProtein serves as a pioneering step towards text-based protein function prediction and sequence design, effectively bridging the gap between protein and human language understanding.","{'model': 'tldr@v2.0.0', 'text': 'InstructProtein, an innovative LLM that possesses bidirectional generation capabilities in both human and protein languages, serves as a pioneering step towards text-based protein function prediction and sequence design, effectively bridging the gap between protein and human language understanding.'}",https://arxiv.org/pdf/2310.03269
-self-detoxifying language models via toxification reversal,Chak Tou Leong,"Language model detoxification aims to minimize the risk of generating offensive or harmful content in pretrained language models (PLMs) for safer deployment. Existing methods can be roughly categorized as finetuning-based and decoding-based. However, the former is often resource-intensive, while the latter relies on additional components and potentially compromises the generation fluency. In this paper, we propose a more lightweight approach that enables the PLM itself to achieve""self-detoxification"". Our method is built upon the observation that prepending a negative steering prompt can effectively induce PLMs to generate toxic content. At the same time, we are inspired by the recent research in the interpretability field, which formulates the evolving contextualized representations within the PLM as an information stream facilitated by the attention layers. Drawing on this idea, we devise a method to identify the toxification direction from the normal generation process to the one prompted with the negative prefix, and then steer the generation to the reversed direction by manipulating the information movement within the attention layers. Experimental results show that our approach, without any fine-tuning or extra components, can achieve comparable performance with state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a more lightweight approach that enables the PLM itself to achieve ""self-detoxification"", built upon the observation that prepending a negative steering prompt can effectively induce PLMs to generate toxic content.'}",
-enhance reasoning ability of visual-language models via large language models,Yueting Yang,"Pre-trained visual language models (VLM) have shown excellent performance in image caption tasks. However, it sometimes shows insufficient reasoning ability. In contrast, large language models (LLMs) emerge with powerful reasoning capabilities. Therefore, we propose a method called TReE, which transfers the reasoning ability of a large language model to a visual language model in zero-shot scenarios. TReE contains three stages: observation, thinking, and re-thinking. Observation stage indicates that VLM obtains the overall information of the relative image. Thinking stage combines the image information and task description as the prompt of the LLM, inference with the rationals. Re-Thinking stage learns from rationale and then inference the final result through VLM.","{'model': 'tldr@v2.0.0', 'text': 'A method called TReE is proposed, which transfers the reasoning ability of a large language model to a visual language model in zero-shot scenarios, and contains three stages: observation, thinking, and re-thinking.'}",http://arxiv.org/pdf/2305.13267
-violation of expectation via metacognitive prompting reduces theory of mind prediction error in large language models,Courtland Leer,"Recent research shows that Large Language Models (LLMs) exhibit a compelling level of proficiency in Theory of Mind (ToM) tasks. This ability to impute unobservable mental states to others is vital to human social cognition and may prove equally important in principal-agent relations between individual humans and Artificial Intelligences (AIs). In this paper, we explore how a mechanism studied in developmental psychology known as Violation of Expectation (VoE) can be implemented to reduce errors in LLM prediction about users by leveraging emergent ToM affordances. And we introduce a \textit{metacognitive prompting} framework to apply VoE in the context of an AI tutor. By storing and retrieving facts derived in cases where LLM expectation about the user was violated, we find that LLMs are able to learn about users in ways that echo theories of human learning. Finally, we discuss latent hazards and augmentative opportunities associated with modeling user psychology and propose ways to mitigate risk along with possible directions for future inquiry.",,https://arxiv.org/pdf/2310.06983
-model selection and evaluation for learning analytics via interpretable machine learning,Zhuoran Li,"Learning analytics is one of the most crucial tasks in understanding the relationship between learners' learning behaviors and academic performance in MOOCs. Machine learning methods that enable us to learn the pattern of complex data make it more effective to conduct learning analytics. However, among a variety of models, only some of them can provide us with sufficient interpretability to understand the modeling result and propose pedagogically instructive suggestions for improving learners' learning strategies. By demonstrating the performance and interpretability of 17 machine learning models, this study would select the CART, which could balance the model performance and interpretability well, as the optimal model for learning analytics in MOOCs. Through the further study of the CART, we concluded that the social interaction behaviors and prompt completion of learning tasks could most significantly help with improving learners' academic performance.","{'model': 'tldr@v2.0.0', 'text': ""The further study of the CART concluded that the social interaction behaviors and prompt completion of learning tasks could most significantly help with improving learners' academic performance.""}",
-bridging the gap between human motion and action semantics via kinematic phrases,Xinpeng Liu,"The goal of motion understanding is to establish a reliable mapping between motion and action semantics, while it is a challenging many-to-many problem. An abstract action semantic (i.e., walk forwards) could be conveyed by perceptually diverse motions (walk with arms up or swinging), while a motion could carry different semantics w.r.t. its context and intention. This makes an elegant mapping between them difficult. Previous attempts adopted direct-mapping paradigms with limited reliability. Also, current automatic metrics fail to provide reliable assessments of the consistency between motions and action semantics. We identify the source of these problems as the significant gap between the two modalities. To alleviate this gap, we propose Kinematic Phrases (KP) that take the objective kinematic facts of human motion with proper abstraction, interpretability, and generality characteristics. Based on KP as a mediator, we can unify a motion knowledge base and build a motion understanding system. Meanwhile, KP can be automatically converted from motions and to text descriptions with no subjective bias, inspiring Kinematic Prompt Generation (KPG) as a novel automatic motion generation benchmark. In extensive experiments, our approach shows superiority over other methods. Our code and data would be made publicly available at https://foruck.github.io/KP.","{'model': 'tldr@v2.0.0', 'text': 'Kinematic Phrases (KP) is proposed that take the objective kinematic facts of human motion with proper abstraction, interpretability, and generality characteristics and can be automatically converted from motions and to text descriptions with no subjective bias.'}",https://arxiv.org/pdf/2310.04189
-automatic hallucination assessment for aligned large language models via transferable adversarial attacks,Xiaodong Yu,"Although remarkable progress has been achieved in preventing large language model (LLM) hallucinations using instruction tuning and retrieval augmentation, it remains challenging to measure the reliability of LLMs using human-crafted evaluation data which is not available for many tasks and domains and could suffer from data leakage. Inspired by adversarial machine learning, this paper aims to develop a method of automatically generating evaluation data by appropriately modifying existing data on which LLMs behave faithfully. Specifically, this paper presents AutoDebug, an LLM-based framework to use prompting chaining to generate transferable adversarial attacks in the form of question-answering examples. We seek to understand the extent to which these examples trigger the hallucination behaviors of LLMs. We implement AutoDebug using ChatGPT and evaluate the resulting two variants of a popular open-domain question-answering dataset, Natural Questions (NQ), on a collection of open-source and proprietary LLMs under various prompting settings. Our generated evaluation data is human-readable and, as we show, humans can answer these modified questions well. Nevertheless, we observe pronounced accuracy drops across multiple LLMs including GPT-4. Our experimental results show that LLMs are likely to hallucinate in two categories of question-answering scenarios where (1) there are conflicts between knowledge given in the prompt and their parametric knowledge, or (2) the knowledge expressed in the prompt is complex. Finally, we find that the adversarial examples generated by our method are transferable across all considered LLMs. The examples generated by a small model can be used to debug a much larger model, making our approach cost-effective.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents AutoDebug, an LLM-based framework to use prompting chaining to generate transferable adversarial attacks in the form of question-answering examples, and finds that the adversarial examples generated by the method are transferable across all considered LLMs.'}",
-improving few-shot generalization of safety classifiers via data augmented parameter-efficient fine-tuning,Ananth Balashankar,"As large language models (LLMs) are widely adopted, new safety issues and policies emerge, to which existing safety classifiers do not generalize well. If we have only observed a few examples of violations of a new safety rule, how can we build a classifier to detect violations? In this paper, we study the novel setting of domain-generalized few-shot learning for LLM-based text safety classifiers. Unlike prior few-shot work, these new safety issues can be hard to uncover and we do not get to choose the few examples. We demonstrate that existing few-shot techniques do not perform well in this setting, and rather we propose to do parameter-efficient fine-tuning (PEFT) combined with augmenting training data based on similar examples in prior existing rules. We empirically show that our approach of similarity-based data-augmentation + prompt-tuning (DAPT) consistently outperforms baselines that either do not rely on data augmentation or on PEFT by 7-17% F1 score in the Social Chemistry moral judgement and 9-13% AUC in the Toxicity detection tasks, even when the new rule is loosely correlated with existing ones.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the approach of similarity-based data-augmentation + prompt-tuning (DAPT) consistently outperforms baselines that either do not rely on data augmentation or on PEFT by 7-17% F1 score in the Social Chemistry moral judgement and 9-13% AUC in the Toxicity detection tasks.'}",
-evaluation of gpt-3.5 and gpt-4 for supporting real-world information needs in healthcare delivery,Debadutta Dash,"Despite growing interest in using large language models (LLMs) in healthcare, current explorations do not assess the real-world utility and safety of LLMs in clinical settings. Our objective was to determine whether two LLMs can serve information needs submitted by physicians as questions to an informatics consultation service in a safe and concordant manner. Sixty six questions from an informatics consult service were submitted to GPT-3.5 and GPT-4 via simple prompts. 12 physicians assessed the LLM responses' possibility of patient harm and concordance with existing reports from an informatics consultation service. Physician assessments were summarized based on majority vote. For no questions did a majority of physicians deem either LLM response as harmful. For GPT-3.5, responses to 8 questions were concordant with the informatics consult report, 20 discordant, and 9 were unable to be assessed. There were 29 responses with no majority on""Agree"",""Disagree"", and""Unable to assess"". For GPT-4, responses to 13 questions were concordant, 15 discordant, and 3 were unable to be assessed. There were 35 responses with no majority. Responses from both LLMs were largely devoid of overt harm, but less than 20% of the responses agreed with an answer from an informatics consultation service, responses contained hallucinated references, and physicians were divided on what constitutes harm. These results suggest that while general purpose LLMs are able to provide safe and credible responses, they often do not meet the specific information need of a given question. A definitive evaluation of the usefulness of LLMs in healthcare settings will likely require additional research on prompt engineering, calibration, and custom-tailoring of general purpose models.","{'model': 'tldr@v2.0.0', 'text': 'It is suggested that while general purpose LLMs are able to provide safe and credible responses, they often do not meet the specific information need of a given question.'}",http://arxiv.org/pdf/2304.13714
-tempera: test-time prompt editing via reinforcement learning,Tianjun Zhang,"Careful prompt design is critical to the use of large language models in zeroshot or few-shot learning. As a consequence, there is a growing interest in automated methods to design optimal prompts. In this work, we propose TEst-tiMe Prompt Editing using Reinforcement leArning (TEMPERA). In contrast to prior prompt generation methods, TEMPERA can efficiently leverage prior knowledge, is adaptive to different queries, and provides an interpretable prompt for every query. To achieve this, we design a novel action space that allows flexible editing of the initial prompts covering a comprehensive set of commonly-used components like instructions, few-shot exemplars, and verbalizers. The proposed method achieves significant gains compared with recent SoTA approaches like prompt tuning, AutoPrompt, and RLPrompt, across a variety of tasks, including sentiment analysis, topic classification, natural language inference, and reading comprehension. Our method achieves 5.33x on average improvement in sample efficiency when compared to the traditional fine-tuning methods. Our code is available at https://github.com/tianjunz/TEMPERA.",,
-gpt-4 is too smart to be safe: stealthy chat with llms via cipher,Youliang Yuan,"Safety lies at the core of the development of Large Language Models (LLMs). There is ample work on aligning LLMs with human ethics and preferences, including data filtering in pretraining, supervised fine-tuning, reinforcement learning from human feedback, and red teaming, etc. In this study, we discover that chat in cipher can bypass the safety alignment techniques of LLMs, which are mainly conducted in natural languages. We propose a novel framework CipherChat to systematically examine the generalizability of safety alignment to non-natural languages -- ciphers. CipherChat enables humans to chat with LLMs through cipher prompts topped with system role descriptions and few-shot enciphered demonstrations. We use CipherChat to assess state-of-the-art LLMs, including ChatGPT and GPT-4 for different representative human ciphers across 11 safety domains in both English and Chinese. Experimental results show that certain ciphers succeed almost 100% of the time to bypass the safety alignment of GPT-4 in several safety domains, demonstrating the necessity of developing safety alignment for non-natural languages. Notably, we identify that LLMs seem to have a ''secret cipher'', and propose a novel SelfCipher that uses only role play and several demonstrations in natural language to evoke this capability. SelfCipher surprisingly outperforms existing human ciphers in almost all cases. Our code and data will be released at https://github.com/RobustNLP/CipherChat.","{'model': 'tldr@v2.0.0', 'text': 'It is discovered that chat in cipher can bypass the safety alignment techniques of LLMs, and a novel SelfCipher is proposed that uses only role play and several demonstrations in natural language to evoke this capability, and surprisingly outperforms existing human ciphers in almost all cases.'}",https://arxiv.org/pdf/2308.06463
-llmatic: neural architecture search via large language models and quality-diversity optimization,Muhammad Umair Nasir,"Large Language Models (LLMs) have emerged as powerful tools capable of accomplishing a broad spectrum of tasks. Their abilities span numerous areas, and one area where they have made a significant impact is in the domain of code generation. In this context, we view LLMs as mutation and crossover tools. Meanwhile, Quality-Diversity (QD) algorithms are known to discover diverse and robust solutions. By merging the code-generating abilities of LLMs with the diversity and robustness of QD solutions, we introduce LLMatic, a Neural Architecture Search (NAS) algorithm. While LLMs struggle to conduct NAS directly through prompts, LLMatic uses a procedural approach, leveraging QD for prompts and network architecture to create diverse and highly performant networks. We test LLMatic on the CIFAR-10 image classification benchmark, demonstrating that it can produce competitive networks with just $2,000$ searches, even without prior knowledge of the benchmark domain or exposure to any previous top-performing models for the benchmark.","{'model': 'tldr@v2.0.0', 'text': 'While LLMs struggle to conduct NAS directly through prompts, LLMatic uses a procedural approach, leveraging QD for prompts and network architecture to create diverse and highly performant networks.'}",https://arxiv.org/pdf/2306.01102
-chatspot: bootstrapping multimodal llms via precise referring instruction tuning,Liang Zhao,"Human-AI interactivity is a critical aspect that reflects the usability of multimodal large language models (MLLMs). However, existing end-to-end MLLMs only allow users to interact with them through language instructions, leading to the limitation of the interactive accuracy and efficiency. In this study, we present precise referring instructions that utilize diverse reference representations such as points and boxes as referring prompts to refer to the special region. This enables MLLMs to focus on the region of interest and achieve finer-grained interaction. Based on precise referring instruction, we propose ChatSpot, a unified end-to-end multimodal large language model that supports diverse forms of interactivity including mouse clicks, drag-and-drop, and drawing boxes, which provides a more flexible and seamless interactive experience. We also construct a multi-grained vision-language instruction-following dataset based on existing datasets and GPT-4 generating. Furthermore, we design a series of evaluation tasks to assess the effectiveness of region recognition and interaction. Experimental results showcase ChatSpot's promising performance.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes ChatSpot, a unified end-to-end multimodal large language model that supports diverse forms of interactivity including mouse clicks, drag-and-drop, and drawing boxes, which provides a more flexible and seamless interactive experience.'}",https://arxiv.org/pdf/2307.09474
-chipgpt: how far are we from natural language hardware design,Kaiyan Chang,"As large language models (LLMs) like ChatGPT exhibited unprecedented machine intelligence, it also shows great performance in assisting hardware engineers to realize higher-efficiency logic design via natural language interaction. To estimate the potential of the hardware design process assisted by LLMs, this work attempts to demonstrate an automated design environment that explores LLMs to generate hardware logic designs from natural language specifications. To realize a more accessible and efficient chip development flow, we present a scalable four-stage zero-code logic design framework based on LLMs without retraining or finetuning. At first, the demo, ChipGPT, begins by generating prompts for the LLM, which then produces initial Verilog programs. Second, an output manager corrects and optimizes these programs before collecting them into the final design space. Eventually, ChipGPT will search through this space to select the optimal design under the target metrics. The evaluation sheds some light on whether LLMs can generate correct and complete hardware logic designs described by natural language for some specifications. It is shown that ChipGPT improves programmability, and controllability, and shows broader design optimization space compared to prior work and native LLMs alone.","{'model': 'tldr@v2.0.0', 'text': 'This work attempts to demonstrate an automated design environment that explores LLMs to generate hardware logic designs from natural language specifications without retraining or finetuning, and shows broader design optimization space compared to prior work and native LLMs alone.'}",http://arxiv.org/pdf/2305.14019
-ask an expert: leveraging language models to improve strategic reasoning in goal-oriented dialogue models,Qiang Zhang,"Existing dialogue models may encounter scenarios which are not well-represented in the training data, and as a result generate responses that are unnatural, inappropriate, or unhelpful. We propose the""Ask an Expert""framework in which the model is trained with access to an""expert""which it can consult at each turn. Advice is solicited via a structured dialogue with the expert, and the model is optimized to selectively utilize (or ignore) it given the context and dialogue history. In this work the expert takes the form of an LLM. We evaluate this framework in a mental health support domain, where the structure of the expert conversation is outlined by pre-specified prompts which reflect a reasoning strategy taught to practitioners in the field. Blenderbot models utilizing""Ask an Expert""show quality improvements across all expert sizes, including those with fewer parameters than the dialogue model itself. Our best model provides a $\sim 10\%$ improvement over baselines, approaching human-level scores on""engingingness""and""helpfulness""metrics.","{'model': 'tldr@v2.0.0', 'text': 'The ""Ask an Expert"" framework is proposed, in which the model is trained with access to an expert which it can consult at each turn, and the expert takes the form of an LLM.'}",http://arxiv.org/pdf/2305.17878
-zero-shot visual relation detection via composite visual cues from large language models,Lin Li,"Pretrained vision-language models, such as CLIP, have demonstrated strong generalization capabilities, making them promising tools in the realm of zero-shot visual recognition. Visual relation detection (VRD) is a typical task that identifies relationship (or interaction) types between object pairs within an image. However, naively utilizing CLIP with prevalent class-based prompts for zero-shot VRD has several weaknesses, e.g., it struggles to distinguish between different fine-grained relation types and it neglects essential spatial information of two objects. To this end, we propose a novel method for zero-shot VRD: RECODE, which solves RElation detection via COmposite DEscription prompts. Specifically, RECODE first decomposes each predicate category into subject, object, and spatial components. Then, it leverages large language models (LLMs) to generate description-based prompts (or visual cues) for each component. Different visual cues enhance the discriminability of similar relation categories from different perspectives, which significantly boosts performance in VRD. To dynamically fuse different cues, we further introduce a chain-of-thought method that prompts LLMs to generate reasonable weights for different visual cues. Extensive experiments on four VRD benchmarks have demonstrated the effectiveness and interpretability of RECODE.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method for zero-shot VRD: RECODE, which solves RElation detection via COmposite DEscription prompts, and introduces a chain-of-thought method that prompts LLMs to generate reasonable weights for different visual cues.'}",https://arxiv.org/pdf/2305.12476
-chill: zero-shot custom interpretable feature extraction from clinical notes with large language models,Denis Jered McInerney,"We propose CHiLL (Crafting High-Level Latents), an approach for natural-language specification of features for linear models. CHiLL prompts LLMs with expert-crafted queries to generate interpretable features from health records. The resulting noisy labels are then used to train a simple linear classifier. Generating features based on queries to an LLM can empower physicians to use their domain expertise to craft features that are clinically meaningful for a downstream task of interest, without having to manually extract these from raw EHR. We are motivated by a real-world risk prediction task, but as a reproducible proxy, we use MIMIC-III and MIMIC-CXR data and standard predictive tasks (e.g., 30-day readmission) to evaluate this approach. We find that linear models using automatically extracted features are comparably performant to models using reference features, and provide greater interpretability than linear models using""Bag-of-Words""features. We verify that learned feature weights align well with clinical expectations.","{'model': 'tldr@v2.0.0', 'text': 'It is found thatlinear models using automatically extracted features are comparably performant to models using reference features, and provide greater interpretability than linear models using""Bag-of-Words"" features.'}",http://arxiv.org/pdf/2302.12343
-could chatgpt imagine: content control for artistic painting generation via large language models,Yue Lu,,,
-can llms keep a secret? testing privacy implications of language models via contextual integrity theory,Niloofar Mireshghallah,"The interactive use of large language models (LLMs) in AI assistants (at work, home, etc.) introduces a new set of inference-time privacy risks: LLMs are fed different types of information from multiple sources in their inputs and are expected to reason about what to share in their outputs, for what purpose and with whom, within a given context. In this work, we draw attention to the highly critical yet overlooked notion of contextual privacy by proposing ConfAIde, a benchmark designed to identify critical weaknesses in the privacy reasoning capabilities of instruction-tuned LLMs. Our experiments show that even the most capable models such as GPT-4 and ChatGPT reveal private information in contexts that humans would not, 39% and 57% of the time, respectively. This leakage persists even when we employ privacy-inducing prompts or chain-of-thought reasoning. Our work underscores the immediate need to explore novel inference-time privacy-preserving approaches, based on reasoning and theory of mind.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ConfAIde, a benchmark designed to identify critical weaknesses in the privacy reasoning capabilities of instruction-tuned LLMs, and underscores the immediate need to explore novel inference-time privacy-preserving approaches, based on reasoning and theory of mind.'}",
-mitigating approximate memorization in language models via dissimilarity learned policy,Aly M. Kassem,"Large Language models (LLMs) are trained on large amounts of data, which can include sensitive information that may compromise per- sonal privacy. LLMs showed to memorize parts of the training data and emit those data verbatim when an adversary prompts appropriately. Previous research has primarily focused on data preprocessing and differential privacy techniques to address memorization or prevent verbatim memorization exclusively, which can give a false sense of privacy. However, these methods rely on explicit and implicit assumptions about the structure of the data to be protected, which often results in an incomplete solution to the problem. To address this, we propose a novel framework that utilizes a reinforcement learning approach (PPO) to fine-tune LLMs to mitigate approximate memorization. Our approach utilizes a negative similarity score, such as BERTScore or SacreBLEU, as a reward signal to learn a dissimilarity policy. Our results demonstrate that this framework effectively mitigates approximate memorization while maintaining high levels of coherence and fluency in the generated samples. Furthermore, our framework is robust in mitigating approximate memorization across various circumstances, including longer context, which is known to increase memorization in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel framework that utilizes a reinforcement learning approach (PPO) to fine-tune LLMs to mitigate approximate memorization and utilizes a negative similarity score, such as BERTScore or SacreBLEU, as a reward signal to learn a dissimilarity policy.'}",http://arxiv.org/pdf/2305.01550
-interleaving pre-trained language models and large language models for zero-shot nl2sql generation,Zihui Gu,"Zero-shot NL2SQL is crucial in achieving natural language to SQL that is adaptive to new environments (e.g., new databases, new linguistic phenomena or SQL structures) with zero annotated NL2SQL samples from such environments. Existing approaches either fine-tune pre-trained language models (PLMs) based on annotated data or use prompts to guide fixed large language models (LLMs) such as ChatGPT. PLMs can perform well in schema alignment but struggle to achieve complex reasoning, while LLMs is superior in complex reasoning tasks but cannot achieve precise schema alignment. In this paper, we propose a ZeroNL2SQL framework that combines the complementary advantages of PLMs and LLMs for supporting zero-shot NL2SQL. ZeroNL2SQL first uses PLMs to generate an SQL sketch via schema alignment, then uses LLMs to fill the missing information via complex reasoning. Moreover, in order to better align the generated SQL queries with values in the given database instances, we design a predicate calibration method to guide the LLM in completing the SQL sketches based on the database instances and select the optimal SQL query via an execution-based strategy. Comprehensive experiments show that ZeroNL2SQL can achieve the best zero-shot NL2SQL performance on real-world benchmarks. Specifically, ZeroNL2SQL outperforms the state-of-the-art PLM-based methods by 3.2% to 13% and exceeds LLM-based methods by 10% to 20% on execution accuracy.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a ZeroNL2SQL framework that combines the complementary advantages of PLMs and LLMs for supporting zero-shot NL2SQL and designs a predicate calibration method to guide the LLM in completing the SQL sketches based on the database instances and select the optimal SQL query via an execution-based strategy.'}",http://arxiv.org/pdf/2306.08891
-distinguish before answer: generating contrastive explanation as knowledge for commonsense question answering,Qianglong Chen,"Existing knowledge-enhanced methods have achieved remarkable results in certain QA tasks via obtaining diverse knowledge from different knowledge bases. However, limited by the properties of retrieved knowledge, they still have trouble benefiting from both the knowledge relevance and distinguishment simultaneously. To address the challenge, we propose CPACE, a Concept-centric Prompt-bAsed Contrastive Explanation Generation model, which aims to convert obtained symbolic knowledge into a contrastive explanation for better distinguishing the differences among given candidates. Firstly, following previous works, we retrieve different types of symbolic knowledge with a concept-centric knowledge extraction module. After that, we generate corresponding contrastive explanations using acquired symbolic knowledge and explanation prompts as guidance for better modeling the knowledge distinguishment and interpretability. Finally, we regard the generated contrastive explanation as external knowledge for downstream task enhancement. We conduct a series of experiments on three widely-used question-answering datasets: CSQA, QASC, and OBQA. Experimental results demonstrate that with the help of generated contrastive explanation, our CPACE model achieves new SOTA on CSQA (89.8% on the testing set, 0.9% higher than human performance), and gains impressive improvement on QASC and OBQA (4.2% and 3.5%, respectively).","{'model': 'tldr@v2.0.0', 'text': 'CPACE, a Concept-centric Prompt-bAsed Contrastive Explanation Generation model, which aims to convert obtained symbolic knowledge into a contrastive explanation for better distinguishing the differences among given candidates is proposed.'}",http://arxiv.org/pdf/2305.08135
-using global land cover product as prompt for cropland mapping via visual foundation model,Chao Tao,"Data-driven deep learning methods have shown great potential in cropland mapping. However, due to multiple factors such as attributes of cropland (topography, climate, crop type) and imaging conditions (viewing angle, illumination, scale), croplands under different scenes demonstrate a great domain gap. This makes it difficult for models trained in the specific scenes to directly generalize to other scenes. A common way to handle this problem is through the""Pretrain+Fine-tuning""paradigm. Unfortunately, considering the variety of features of cropland that are affected by multiple factors, it is hardly to handle the complex domain gap between pre-trained data and target data using only sparse fine-tuned samples as general constraints. Moreover, as the number of model parameters grows, fine-tuning is no longer an easy and low-cost task. With the emergence of prompt learning via visual foundation models, the""Pretrain+Prompting""paradigm redesigns the optimization target by introducing individual prompts for each single sample. This simplifies the domain adaption from generic to specific scenes during model reasoning processes. Therefore, we introduce the""Pretrain+Prompting""paradigm to interpreting cropland scenes and design the auto-prompting (APT) method based on freely available global land cover product. It can achieve a fine-grained adaptation process from generic scenes to specialized cropland scenes without introducing additional label costs. To our best knowledge, this work pioneers the exploration of the domain adaption problems for cropland mapping under prompt learning perspectives. Our experiments using two sub-meter cropland datasets from southern and northern China demonstrated that the proposed method via visual foundation models outperforms traditional supervised learning and fine-tuning approaches in the field of remote sensing.","{'model': 'tldr@v2.0.0', 'text': 'This work pioneers the exploration of the domain adaption problems for cropland mapping under prompt learning perspectives and designs the auto-prompting (APT) method based on freely available global land cover product that can achieve a fine-grained adaptation process from generic scenes to specialized croplands scenes without introducing additional label costs.'}",
-epa: easy prompt augmentation on large language models via multiple sources and multiple targets,Hongyuan Lu,"Large language models (LLMs) have shown promising performance on various NLP tasks via task prompting. And their performance can be further improved by appending task demonstrations to the head of the prompt. And usually, a better performance can be achieved with more demonstrations. However, asking the users to write the demonstrations can be cumbersome. As a simple yet cost-effective workaround, this paper proposes a novel method called EPA (\textbf{E}asy \textbf{P}rompt \textbf{A}ugmentation)\footnote{While this paper considers augmenting prompts via demonstrations, we name it EPA as the name EDA is already taken by a well-known NLP method \citep{wei-zou-2019-eda}.} that effectively minimizes user efforts in writing demonstrations while improving the model performance at the same time. EPA achieves these goals by automatically augmenting the demonstrations with multiple sources/targets, where each of them paraphrases each other. This is well motivated as augmenting data via paraphrasing effectively improves neural language models. EPA thus employs paraphrasing as an augmentation method for in-context learning. Extensive experiments indicate that EPA effectively improves both NLU and NLG tasks, covering from natural language inference to machine translation in translating tens of languages.\footnote{Code and data will be released upon publication.}","{'model': 'tldr@v2.0.0', 'text': 'A novel method called EPA that effectively minimizes user efforts in writing demonstrations while improving the model performance at the same time, and automatically augmenting the demonstrations with multiple sources/targets, where each of them paraphrasing each other.'}",https://arxiv.org/pdf/2309.04725
-expclip: bridging text and facial expressions via semantic alignment,Yicheng Zhong,"The objective of stylized speech-driven facial animation is to create animations that encapsulate specific emotional expressions. Existing methods often depend on pre-established emotional labels or facial expression templates, which may limit the necessary flexibility for accurately conveying user intent. In this research, we introduce a technique that enables the control of arbitrary styles by leveraging natural language as emotion prompts. This technique presents benefits in terms of both flexibility and user-friendliness. To realize this objective, we initially construct a Text-Expression Alignment Dataset (TEAD), wherein each facial expression is paired with several prompt-like descriptions.We propose an innovative automatic annotation method, supported by Large Language Models (LLMs), to expedite the dataset construction, thereby eliminating the substantial expense of manual annotation. Following this, we utilize TEAD to train a CLIP-based model, termed ExpCLIP, which encodes text and facial expressions into semantically aligned style embeddings. The embeddings are subsequently integrated into the facial animation generator to yield expressive and controllable facial animations. Given the limited diversity of facial emotions in existing speech-driven facial animation training data, we further introduce an effective Expression Prompt Augmentation (EPA) mechanism to enable the animation generator to support unprecedented richness in style control. Comprehensive experiments illustrate that our method accomplishes expressive facial animation generation and offers enhanced flexibility in effectively conveying the desired style.","{'model': 'tldr@v2.0.0', 'text': 'An innovative automatic annotation method is proposed, supported by Large Language Models, to expedite the dataset construction, thereby eliminating the substantial expense of manual annotation and introducing an effective Expression Prompt Augmentation mechanism to enable the animation generator to support unprecedented richness in style control.'}",https://arxiv.org/pdf/2308.14448
-divknowqa: assessing the reasoning ability of llms via open-domain question answering over knowledge base and text,Wenting Zhao,"Large Language Models (LLMs) have exhibited impressive generation capabilities, but they suffer from hallucinations when solely relying on their internal knowledge, especially when answering questions that require less commonly known information. Retrieval-augmented LLMs have emerged as a potential solution to ground LLMs in external knowledge. Nonetheless, recent approaches have primarily emphasized retrieval from unstructured text corpora, owing to its seamless integration into prompts. When using structured data such as knowledge graphs, most methods simplify it into natural text, neglecting the underlying structures. Moreover, a significant gap in the current landscape is the absence of a realistic benchmark for evaluating the effectiveness of grounding LLMs on heterogeneous knowledge sources (e.g., knowledge base and text). To fill this gap, we have curated a comprehensive dataset that poses two unique challenges: (1) Two-hop multi-source questions that require retrieving information from both open-domain structured and unstructured knowledge sources; retrieving information from structured knowledge sources is a critical component in correctly answering the questions. (2) The generation of symbolic queries (e.g., SPARQL for Wikidata) is a key requirement, which adds another layer of challenge. Our dataset is created using a combination of automatic generation through predefined reasoning chains and human annotation. We also introduce a novel approach that leverages multiple retrieval tools, including text passage retrieval and symbolic language-assisted retrieval. Our model outperforms previous approaches by a significant margin, demonstrating its effectiveness in addressing the above-mentioned reasoning challenges.","{'model': 'tldr@v2.0.0', 'text': 'This work curated a comprehensive dataset that poses two unique challenges: two-hop multi-source questions that require retrieving information from both open-domain structured and unstructured knowledge sources; and a novel approach that leverages multiple retrieval tools, including text passage retrieval and symbolic language-assisted retrieval.'}",
-tailoring personality traits in large language models via unsupervisedly-built personalized lexicons,Tianlong Li,"Personality plays a pivotal role in shaping human expression patterns, and empowering and manipulating large language models (LLMs) with personality traits holds significant promise in enhancing the user experience of LLMs. However, prior approaches either rely on fine-tuning LLMs on a corpus enriched with personalized expressions or necessitate the manual crafting of prompts to induce LLMs to produce personalized responses. The former approaches demand substantial time and resources for collecting sufficient training examples while the latter might fail in enabling the precise manipulation of the personality traits at a fine-grained level (e.g., achieving high agreeableness while reducing openness). In this study, we introduce a novel approach for tailoring personality traits within LLMs, allowing for the incorporation of any combination of the Big Five factors (i.e., openness, conscientiousness, extraversion, agreeableness, and neuroticism) in a pluggable manner. This is achieved by employing a set of Unsupervisedly-Built Personalized Lexicons (UBPL) that are utilized to adjust the probability of the next token predicted by the original LLMs during the decoding phase. This adjustment encourages the models to generate words present in the personalized lexicons while preserving the naturalness of the generated texts. Extensive experimentation demonstrates the effectiveness of our approach in finely manipulating LLMs' personality traits. Furthermore, our method can be seamlessly integrated into other LLMs without necessitating updates to their parameters.",,
-denevil: towards deciphering and navigating the ethical values of large language models via instruction learning,Shitong Duan,"Large Language Models (LLMs) have made unprecedented breakthroughs, yet their increasing integration into everyday life might raise societal risks due to generated unethical content. Despite extensive study on specific issues like bias, the intrinsic values of LLMs remain largely unexplored from a moral philosophy perspective. This work delves into ethical values utilizing Moral Foundation Theory. Moving beyond conventional discriminative evaluations with poor reliability, we propose DeNEVIL, a novel prompt generation algorithm tailored to dynamically exploit LLMs' value vulnerabilities and elicit the violation of ethics in a generative manner, revealing their underlying value inclinations. On such a basis, we construct MoralPrompt, a high-quality dataset comprising 2,397 prompts covering 500+ value principles, and then benchmark the intrinsic values across a spectrum of LLMs. We discovered that most models are essentially misaligned, necessitating further ethical value alignment. In response, we develop VILMO, an in-context alignment method that substantially enhances the value compliance of LLM outputs by learning to generate appropriate value instructions, outperforming existing competitors. Our methods are suitable for black-box and open-source models, offering a promising initial step in studying the ethical values of LLMs.","{'model': 'tldr@v2.0.0', 'text': ""DeNEVIL, a novel prompt generation algorithm tailored to dynamically exploit LLMs' value vulnerabilities and elicit the violation of ethics in a generative manner, revealing their underlying value inclinations is proposed.""}",
-pitl: cross-modal retrieval with weakly-supervised vision-language pre-training via prompting,Zixin Guo,"Vision-language (VL) Pre-training (VLP) has shown to well generalize VL models over a wide range of VL downstream tasks, especially for cross-modal retrieval. However, it hinges on a huge amount of image-text pairs, which requires tedious and costly curation. On the contrary,weakly-supervised VLP (W-VLP) explores means with object tags generated by a pre-trained object detector (OD) from images. Yet, they still require paired information, i.e. images and object-level annotations, as supervision to train an OD. To further reduce the amount of supervision, we propose Prompts-in-The-Loop (PiTL) that prompts knowledge from large language models (LLMs) to describe images. Concretely, given a category label of an image, e.g.refinery, the knowledge, e.g.a refinery could be seen with large storage tanks, pipework, and ..., extracted by LLMs is used as the language counterpart. The knowledge supplements, e.g. the common relations among entities most likely appearing in a scene. We create IN14K, a new VL dataset of 9M images and 1M descriptions of 14K categories from ImageNet21K with PiTL. Empirically, the VL models pre-trained with PiTL-generated pairs are strongly favored over other W-VLP works on image-to-text (I2T) and text-to-image (T2I) retrieval tasks, with less supervision. The results reveal the effectiveness of PiTL-generated pairs for VLP.","{'model': 'tldr@v2.0.0', 'text': 'Empirically, the VL models pre-trained with PiTL-generated pairs are strongly favored over other W-VLP works on image-to-text (I2T) and text- to-image (T2I) retrieval tasks, with less supervision.'}",https://dl.acm.org/doi/pdf/10.1145/3539618.3592038
-vision-language interpreter for robot task planning,Keisuke Shirai,"Large language models (LLMs) are accelerating the development of language-guided robot planners. Meanwhile, symbolic planners offer the advantage of interpretability. This paper proposes a new task that bridges these two trends, namely, multimodal planning problem specification. The aim is to generate a problem description (PD), a machine-readable file used by the planners to find a plan. By generating PDs from language instruction and scene observation, we can drive symbolic planners in a language-guided framework. We propose a Vision-Language Interpreter (ViLaIn), a new framework that generates PDs using state-of-the-art LLM and vision-language models. ViLaIn can refine generated PDs via error message feedback from the symbolic planner. Our aim is to answer the question: How accurately can ViLaIn and the symbolic planner generate valid robot plans? To evaluate ViLaIn, we introduce a novel dataset called the problem description generation (ProDG) dataset. The framework is evaluated with four new evaluation metrics. Experimental results show that ViLaIn can generate syntactically correct problems with more than 99% accuracy and valid plans with more than 58% accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A Vision-Language Interpreter (ViLaIn), a new framework that generates PDs using state-of-the-art LLM and vision-language models, to answer the question: How accurately can ViLaIn and the symbolic planner generate valid robot plans?'}",
-prompting with pseudo-code instructions,Mayank Mishra,"Prompting with natural language instructions has recently emerged as a popular method of harnessing the capabilities of large language models. Given the inherent ambiguity present in natural language, it is intuitive to consider the possible advantages of prompting with less ambiguous prompt styles, such as the use of pseudo-code. In this paper we explore if prompting via pseudo-code instructions helps improve the performance of pre-trained language models. We manually create a dataset of pseudo-code prompts for 132 different tasks spanning classification, QA and generative language tasks, sourced from the Super-NaturalInstructions dataset. Using these prompts along with their counterparts in natural language, we study their performance on two LLM families - BLOOM and CodeGen. Our experiments show that using pseudo-code instructions leads to better results, with an average increase (absolute) of 7-16 points in F1 scores for classification tasks and an improvement (relative) of 12-38% in aggregate ROUGE-L scores across all tasks. We include detailed ablation studies which indicate that code comments, docstrings, and the structural clues encoded in pseudo-code all contribute towards the improvement in performance. To the best of our knowledge our work is the first to demonstrate how pseudo-code prompts can be helpful in improving the performance of pre-trained LMs.","{'model': 'tldr@v2.0.0', 'text': 'This work is the first to demonstrate how pseudo-code prompts can be helpful in improving the performance of pre-trained LMs, with detailed ablation studies which indicate that code comments, docstrings, and the structural clues encoded in pseudo- code all contribute towards the improvement in performance.'}",http://arxiv.org/pdf/2305.11790
-towards general visual-linguistic face forgery detection,Ke Sun,"Deepfakes are realistic face manipulations that can pose serious threats to security, privacy, and trust. Existing methods mostly treat this task as binary classification, which uses digital labels or mask signals to train the detection model. We argue that such supervisions lack semantic information and interpretability. To address this issues, in this paper, we propose a novel paradigm named Visual-Linguistic Face Forgery Detection(VLFFD), which uses fine-grained sentence-level prompts as the annotation. Since text annotations are not available in current deepfakes datasets, VLFFD first generates the mixed forgery image with corresponding fine-grained prompts via Prompt Forgery Image Generator (PFIG). Then, the fine-grained mixed data and coarse-grained original data and is jointly trained with the Coarse-and-Fine Co-training framework (C2F), enabling the model to gain more generalization and interpretability. The experiments show the proposed method improves the existing detection models on several challenging benchmarks.","{'model': 'tldr@v2.0.0', 'text': 'A novel paradigm named Visual-Linguistic Face Forgery Detection (VLFFD), which uses fine-grained sentence-level prompts as the annotation, and improves the existing detection models on several challenging benchmarks.'}",https://arxiv.org/pdf/2307.16545
-forgetful large language models: lessons learned from using llms in robot programming,Juo-Tung Chen,"Large language models offer new ways of empowering people to program robot applications-namely, code generation via prompting. However, the code generated by LLMs is susceptible to errors. This work reports a preliminary exploration that empirically characterizes common errors produced by LLMs in robot programming. We categorize these errors into two phases: interpretation and execution. In this work, we focus on errors in execution and observe that they are caused by LLMs being""forgetful""of key information provided in user prompts. Based on this observation, we propose prompt engineering tactics designed to reduce errors in execution. We then demonstrate the effectiveness of these tactics with three language models: ChatGPT, Bard, and LLaMA-2. Finally, we discuss lessons learned from using LLMs in robot programming and call for the benchmarking of LLM-powered end-user development of robot applications.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes prompt engineering tactics designed to reduce errors in execution of LLMs in robot programming and demonstrates the effectiveness of these tactics with three language models: ChatGPT, Bard, and LLaMA-2.'}",https://arxiv.org/pdf/2310.06646
-interpretable unified language checking,Tianhua Zhang,"Despite recent concerns about undesirable behaviors generated by large language models (LLMs), including non-factual, biased, and hateful language, we find LLMs are inherent multi-task language checkers based on their latent representations of natural and social knowledge. We present an interpretable, unified, language checking (UniLC) method for both human and machine-generated language that aims to check if language input is factual and fair. While fairness and fact-checking tasks have been handled separately with dedicated models, we find that LLMs can achieve high performance on a combination of fact-checking, stereotype detection, and hate speech detection tasks with a simple, few-shot, unified set of prompts. With the ``1/2-shot'' multi-task language checking method proposed in this work, the GPT3.5-turbo model outperforms fully supervised baselines on several language tasks. The simple approach and results suggest that based on strong latent knowledge representations, an LLM can be an adaptive and explainable tool for detecting misinformation, stereotypes, and hate speech.","{'model': 'tldr@v2.0.0', 'text': 'It is found that LLMs can achieve high performance on a combination of fact-checking, stereotype detection, and hate speech detection tasks with a simple, few-shot, unified set of prompts.'}",http://arxiv.org/pdf/2304.03728
-genrec: large language model for generative recommendation,Jianchao Ji,"In recent years, large language models (LLM) have emerged as powerful tools for diverse natural language processing tasks. However, their potential for recommender systems under the generative recommendation paradigm remains relatively unexplored. This paper presents an innovative approach to recommendation systems using large language models (LLMs) based on text data. In this paper, we present a novel LLM for generative recommendation (GenRec) that utilized the expressive power of LLM to directly generate the target item to recommend, rather than calculating ranking score for each candidate item one by one as in traditional discriminative recommendation. GenRec uses LLM's understanding ability to interpret context, learn user preferences, and generate relevant recommendation. Our proposed approach leverages the vast knowledge encoded in large language models to accomplish recommendation tasks. We first we formulate specialized prompts to enhance the ability of LLM to comprehend recommendation tasks. Subsequently, we use these prompts to fine-tune the LLaMA backbone LLM on a dataset of user-item interactions, represented by textual data, to capture user preferences and item characteristics. Our research underscores the potential of LLM-based generative recommendation in revolutionizing the domain of recommendation systems and offers a foundational framework for future explorations in this field. We conduct extensive experiments on benchmark datasets, and the experiments shows that our GenRec has significant better results on large dataset.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a novel LLM for generative recommendation (GenRec) that utilized the expressive power of LLM to directly generate the target item to recommend, rather than calculating ranking score for each candidate item one by one as in traditional discriminative recommendation.'}",https://arxiv.org/pdf/2307.00457
-differentially private decoding in large language models,Jimit Majmudar,"Recent large-scale natural language processing (NLP) systems use a pre-trained Large Language Model (LLM) on massive and diverse corpora as a headstart. In practice, the pre-trained model is adapted to a wide array of tasks via fine-tuning on task-specific datasets. LLMs, while effective, have been shown to memorize instances of training data thereby potentially revealing private information processed during pre-training. The potential leakage might further propagate to the downstream tasks for which LLMs are fine-tuned. On the other hand, privacy-preserving algorithms usually involve retraining from scratch, which is prohibitively expensive for LLMs. In this work, we propose a simple, easy to interpret, and computationally lightweight perturbation mechanism to be applied to an already trained model at the decoding stage. Our perturbation mechanism is model-agnostic and can be used in conjunction with any LLM. We provide theoretical analysis showing that the proposed mechanism is differentially private, and experimental results showing a privacy-utility trade-off.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a simple, easy to interpret, and computationally lightweight perturbation mechanism to be applied to an already trained model at the decoding stage, which is model-agnostic and can be used in conjunction with any LLM.'}",http://arxiv.org/pdf/2205.13621
-"a multitask, multilingual, multimodal evaluation of chatgpt on reasoning, hallucination, and interactivity",Yejin Bang,"This paper proposes a framework for quantitatively evaluating interactive LLMs such as ChatGPT using publicly available data sets. We carry out an extensive technical evaluation of ChatGPT using 23 data sets covering 8 different common NLP application tasks. We evaluate the multitask, multilingual and multi-modal aspects of ChatGPT based on these data sets and a newly designed multimodal dataset. We find that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks. We find that it is better at understanding non-Latin script languages than generating them. It is able to generate multimodal content from textual prompts, via an intermediate code generation step. Moreover, we find that ChatGPT is 63.41% accurate on average in 10 different reasoning categories under logical reasoning, non-textual reasoning, and commonsense reasoning, hence making it an unreliable reasoner. It is, for example, better at deductive than inductive reasoning. ChatGPT suffers from hallucination problems like other LLMs and it generates more extrinsic hallucinations from its parametric memory as it does not have access to an external knowledge base. Finally, the interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance, i.e, 8% ROUGE-1 on summarization and 2% ChrF++ on machine translation, in a multi-turn""prompt engineering""fashion. We also release codebase for evaluation set extraction.","{'model': 'tldr@v2.0.0', 'text': 'It is found that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks and is better at understanding non-Latin script languages than generating them.'}",http://arxiv.org/pdf/2302.04023
-chain-of-thought prompt distillation for multimodal named entity and multimodal relation extraction,F. Chen,"Multimodal Named Entity Recognition (MNER) and Multimodal Relation Extraction (MRE) necessitate the fundamental reasoning capacity for intricate linguistic and multimodal comprehension. In this study, we explore distilling the reasoning ability of large language models (LLMs) into a more compact student model by generating a chain of thought (CoT) – a sequence of intermediate reasoning steps. Specifically, we commence by exemplifying the elicitation of such reasoning ability from LLMs through CoT prompts covering multi-grain (noun, sentence, multimodality) and data-augmentation (style, entity, image) dimensions. Subsequently, we present a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge. Extensive experiments reveal that our approach attains state-of-theart accuracy and manifests a plethora of advantages concerning interpretability, data efficiency, and cross-domain generalization on MNER and MRE datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge.'}",https://arxiv.org/pdf/2306.14122
-fedlogic: interpretable federated multi-domain chain-of-thought prompt selection for large language models,Pengwei Xing,"Leveraging ``chain-of-thought (CoT)'' reasoning to elicit rapid and precise responses from large language models (LLMs) is rapidly attracting research interest. A notable challenge here is how to design or select optimal prompts. The process of prompt selection relies on trial and error, involving continuous adjustments and combinations of input prompts by users based on the corresponding new responses generated from LLMs. Furthermore, minimal research has been conducted to explore how LLMs employ the mathematical problem-solving capabilities learned from user interactions to address issues in narrative writing. To improve interpretability and explore the balance principle between generality and personalization under a multi-domain CoT prompt selection scenario, we propose the Federated Logic rule learning approach (FedLogic). We introduce a theoretical formalization and interactive emulation of the multi-domain CoT prompt selection dilemma in the context of federated LLMs. We cast the problem of joint probability modeling as a bilevel program, where the CoT prompt selection intricacy can be likened to a fuzzy score-based rule selection with the LLMs function as rule generators. FedLogic solves this problem through variational expectation maximization (V-EM). In addition, we incorporate two KL-divergence constraints within this probabilistic modeling framework to surmount the intricacies of managing extensive search spaces and accomplishing cross-domain personalization of CoTs. To the best of our knowledge, FedLogic is the first interpretable and principled federated multi-domain CoT prompt selection approach for LLMs.","{'model': 'tldr@v2.0.0', 'text': 'FedLogic is the first interpretable and principled federated multi-domain CoT prompt selection approach for LLMs and incorporates two KL-divergence constraints within this probabilistic modeling framework to surmount the intricacies of managing extensive search spaces and accomplishing cross-domain personalization of CoTs.'}",https://arxiv.org/pdf/2308.15324
-learning concise and descriptive attributes for visual recognition,Andy Yan,"Recent advances in foundation models present new opportunities for interpretable visual recognition -- one can first query Large Language Models (LLMs) to obtain a set of attributes that describe each class, then apply vision-language models to classify images via these attributes. Pioneering work shows that querying thousands of attributes can achieve performance competitive with image features. However, our further investigation on 8 datasets reveals that LLM-generated attributes in a large quantity perform almost the same as random words. This surprising finding suggests that significant noise may be present in these attributes. We hypothesize that there exist subsets of attributes that can maintain the classification performance with much smaller sizes, and propose a novel learning-to-search method to discover those concise sets of attributes. As a result, on the CUB dataset, our method achieves performance close to that of massive LLM-generated attributes (e.g., 10k attributes for CUB), yet using only 32 attributes in total to distinguish 200 bird species. Furthermore, our new paradigm demonstrates several additional benefits: higher interpretability and interactivity for humans, and the ability to summarize knowledge for a recognition task.","{'model': 'tldr@v2.0.0', 'text': 'This work hypothesizes that there exist subsets of attributes that can maintain the classification performance with much smaller sizes, and proposes a novel learning-to-search method to discover those concise sets of attributes.'}",https://arxiv.org/pdf/2308.03685
-machine learning generated decision boundaries for prediction and exploration of patient-specific quality assurance failures in stereotactic radiosurgery plans.,Jeremy Braun,"INTRODUCTION Stereotactic Radiosurgery (SRS) is a form of radiotherapy treatment during which high radiation dose is delivered in a single or few fractions. These treatments require highly conformal plans with steep dose gradients which can result in an increase in plan complexity prompting the need for stringent pre-treatment patient specific quality assurance (QA) measurements to ensure the planned and measured dose distributions agree within clinical standards. Complexity scores and machine learning (ML) techniques may help with prediction of QA outcomes however interpretability and usability of those results continues to be an area of study. This study investigates the use of plan complexity metrics as input for an ML model to allow for prediction of QA outcomes for SRS plans as measured via 3D phantom dose verification. Explorations into interpretability and predictive performance changes as model dimensionality increases, as well as a prospective in-clinic implementation using the resulting model were also performed. METHODS 498 plans (1571 VMAT arcs) were processed via in-house script to generate several complexity scores. 3D phantom dose verification measurement results were extracted and classified as pass or failure (with failures defined as below 95% voxel agreement passing 3%/1mm gamma criteria with 10% threshold,) and 1472 of the arcs were split into training and testing sets, with 99 arcs as a sequential holdout set. A z-score scaler was trained on the training set and used to scale all other sets. Variations of MLC leaf movement variability, aperture complexity, and leaf size and MU at control point weighted target area scores were used as input to a Support Vector Classifier to generate a series of 1-D, 2-D, and 5-D decision boundaries. The best performing 5D model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements for 38 patient plans (112 arcs). The decision to order 3D phantom dose verification measurements was recorded before and after prediction. RESULTS Best performing 1-D threshold, and 2-D prediction models with best performance produced a QA failure recall and QA passing recall of 1.00 and 0.55, and 0.82 and 0.82 respectively. Best performing 5-D prediction model produced a QA failure recall (sensitivity) of 1.00, and QA passing recall (specificity) of 0.72. This model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements and achieved a QA failure recall of 1.00 and QA passing recall of 0.58. The decision to order 3D phantom dose verification measurements was recorded before and after measurement. A single initially unidentified failing plan of the prospective cohort was successfully predicted to fail by the model. CONCLUSION Implementation of complexity score based prediction models for SRS would allow for support of a clinician's decision to reduce time spent performing QA measurements, and avoid patient treatment delays (i.e. in case of QA failure). This article is protected by copyright. All rights reserved.","{'model': 'tldr@v2.0.0', 'text': ""Implementation of complexity score based prediction models for SRS would allow for support of a clinician's decision to reduce time spent performing QA measurements, and avoid patient treatment delays (i.e. in case of QA failure).""}",
-tackling vision language tasks through learning inner monologues,Diji Yang,"Visual language tasks require AI models to comprehend and reason with both visual and textual content. Driven by the power of Large Language Models (LLMs), two prominent methods have emerged: (1) the hybrid integration between LLMs and Vision-Language Models (VLMs), where visual inputs are firstly converted into language descriptions by VLMs, serving as inputs for LLMs to generate final answer(s); (2) visual feature alignment in language space, where visual inputs are encoded as embeddings and projected to LLMs' language space via further supervised fine-tuning. The first approach provides light training costs and interpretability but is hard to be optimized in an end-to-end fashion. The second approach presents decent performance, but feature alignment usually requires large amounts of training data and lacks interpretability. To tackle this dilemma, we propose a novel approach, Inner Monologue Multi-Modal Optimization (IMMO), to solve complex vision language problems by simulating inner monologue processes, a cognitive process in which an individual engages in silent verbal communication with themselves. We enable LLMs and VLMs to interact through natural language conversation and propose to use a two-stage training process to learn how to do the inner monologue (self-asking questions and answering questions). IMMO is evaluated on two popular tasks and the results suggest by emulating the cognitive phenomenon of internal dialogue, our approach can enhance reasoning and explanation abilities, contributing to the more effective fusion of vision and language models. More importantly, instead of using predefined human-crafted monologues, IMMO learns this process within the deep learning models, promising wider applicability to many different AI problems beyond vision language tasks.","{'model': 'tldr@v2.0.0', 'text': 'Inner Monologue Multi-Modal Optimization is proposed, to solve complex vision language problems by simulating inner monologue processes, a cognitive process in which an individual engages in silent verbal communication with themselves, promising wider applicability to many different AI problems beyond vision language tasks.'}",https://arxiv.org/pdf/2308.09970
-robust preference learning for storytelling via contrastive reinforcement learning,Louis Castricato,"Controlled automated story generation seeks to generate natural language stories satisfying constraints from natural language critiques or preferences. Existing methods to control for story preference utilize prompt engineering which is labor intensive and often inconsistent. They may also use logit-manipulation methods which require annotated datasets to exist for the desired attributes. To address these issues, we first train a contrastive bi-encoder model to align stories with corresponding human critiques, named CARP, building a general purpose preference model. This is subsequently used as a reward function to fine-tune a generative language model via reinforcement learning. However, simply fine-tuning a generative language model with a contrastive reward model does not always reliably result in a story generation system capable of generating stories that meet user preferences. To increase story generation robustness we further fine-tune the contrastive reward model using a prompt-learning technique. A human participant study is then conducted comparing generations from our full system, ablations, and two baselines. We show that the full fine-tuning pipeline results in a story generator preferred over a LLM 20x as large as well as logit-based methods. This motivates the use of contrastive learning for general purpose human preference modeling.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that the full fine-tuning pipeline results in a story generator preferred over a LLM 20x as large as well as logit-based methods, and motivates the use of contrastive learning for general purpose human preference modeling.'}",http://arxiv.org/pdf/2210.07792
-the power of large language models for wireless communication system development: a case study on fpga platforms,Yuyang Du,"Large language models (LLMs) have garnered significant attention across various research disciplines, including the wireless communication community. There have been several heated discussions on the intersection of LLMs and wireless technologies. While recent studies have demonstrated the ability of LLMs to generate hardware description language (HDL) code for simple computation tasks, developing wireless prototypes and products via HDL poses far greater challenges because of the more complex computation tasks involved. In this paper, we aim to address this challenge by investigating the role of LLMs in FPGA-based hardware development for advanced wireless signal processing. We begin by exploring LLM-assisted code refactoring, reuse, and validation, using an open-source software-defined radio (SDR) project as a case study. Through the case study, we find that an LLM assistant can potentially yield substantial productivity gains for researchers and developers. We then examine the feasibility of using LLMs to generate HDL code for advanced wireless signal processing, using the Fast Fourier Transform (FFT) algorithm as an example. This task presents two unique challenges: the scheduling of subtasks within the overall task and the multi-step thinking required to solve certain arithmetic problem within the task. To address these challenges, we employ in-context learning (ICL) and Chain-of-Thought (CoT) prompting techniques, culminating in the successful generation of a 64-point Verilog FFT module. Our results demonstrate the potential of LLMs for generalization and imitation, affirming their usefulness in writing HDL code for wireless communication systems. Overall, this work contributes to understanding the role of LLMs in wireless communication and motivates further exploration of their capabilities.","{'model': 'tldr@v2.0.0', 'text': 'This work begins by exploring LLM-assisted code refactoring, reuse, and validation, using an open-source software-defined radio (SDR) project as a case study, and examines the feasibility of using LLMs to generate HDL code for advanced wireless signal processing.'}",
-finvis-gpt: a multimodal large language model for financial chart analysis,Ziao Wang,"In this paper, we propose FinVis-GPT, a novel multimodal large language model (LLM) specifically designed for financial chart analysis. By leveraging the power of LLMs and incorporating instruction tuning and multimodal capabilities, FinVis-GPT is capable of interpreting financial charts and providing valuable analysis. To train FinVis-GPT, a financial task oriented dataset was generated for pre-training alignment and instruction tuning, comprising various types of financial charts and their corresponding descriptions. We evaluate the model performance via several case studies due to the time limit, and the promising results demonstrated that FinVis-GPT is superior in various financial chart related tasks, including generating descriptions, answering questions and predicting future market trends, surpassing existing state-of-the-art multimodal LLMs. The proposed FinVis-GPT serves as a pioneering effort in utilizing multimodal LLMs in the finance domain and our generated dataset will be release for public use in the near future to speedup related research.","{'model': 'tldr@v2.0.0', 'text': 'The promising results demonstrated that FinVis-GPT is superior in various financial chart related tasks, including generating descriptions, answering questions and predicting future market trends, surpassing existing state-of-the-art multimodal LLMs.'}",https://arxiv.org/pdf/2308.01430
-interpreting reward models in rlhf-tuned language models using sparse autoencoders,Luke Marks,"Large language models (LLMs) aligned to human preferences via reinforcement learning from human feedback (RLHF) underpin many commercial applications. However, how RLHF impacts LLM internals remains opaque. We propose a novel method to interpret learned reward functions in RLHF-tuned LLMs using sparse autoencoders. Our approach trains autoencoder sets on activations from a base LLM and its RLHF-tuned version. By comparing autoencoder hidden spaces, we identify unique features that reflect the accuracy of the learned reward model. To quantify this, we construct a scenario where the tuned LLM learns token-reward mappings to maximize reward. This is the first application of sparse autoencoders for interpreting learned rewards and broadly inspecting reward learning in LLMs. Our method provides an abstract approximation of reward integrity. This presents a promising technique for ensuring alignment between specified objectives and model behaviors.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method to interpret learned reward functions in RLHF-tuned LLMs using sparse autoencoders and provides an abstract approximation of reward integrity.'}",https://arxiv.org/pdf/2310.08164
-using natural language explanations to rescale human judgments,Manya Wadhwa,"The rise of large language models (LLMs) has brought a critical need for high-quality human-labeled data, particularly for processes like human feedback and evaluation. A common practice is to label data via consensus annotation over the judgments of multiple crowdworkers. However, different annotators may have different interpretations of labeling schemes unless given extensive training, and for subjective NLP tasks, even trained expert annotators can diverge heavily. We show that these nuances can be captured by high quality natural language explanations, and propose a method to rescale ordinal annotation in the presence of disagreement using LLMs. Specifically, we feed Likert ratings and corresponding natural language explanations into an LLM and prompt it to produce a numeric score. This score should reflect the underlying assessment of the example by the annotator. The presence of explanations allows the LLM to homogenize ratings across annotators in spite of scale usage differences. We explore our technique in the context of a document-grounded question answering task on which large language models achieve near-human performance. Among questions where annotators identify incompleteness in the answers, our rescaling improves correlation between nearly all annotator pairs, improving pairwise correlation on these examples by an average of 0.2 Kendall's tau.","{'model': 'tldr@v2.0.0', 'text': 'Rescaling improves correlation between nearly all annotator pairs, and allows the LLM to homogenize ratings across annotators in spite of scale usage differences, in the context of a document-grounded question answering task on which large language models achieve near-human performance.'}",http://arxiv.org/pdf/2305.14770
-multi-scale discriminative regions analysis in fdg-pet imaging for early diagnosis of alzheimer’s disease,Jin Zhang,"Objective. Alzheimer’s disease (AD) is a degenerative brain disorder, one of the main causes of death in elderly people, so early diagnosis of AD is vital to prompt access to medication and medical care. Fluorodeoxyglucose positron emission tomography (FDG-PET) proves to be effective to help understand neurological changes via measuring glucose uptake. Our aim is to explore information-rich regions of FDG-PET imaging, which enhance the accuracy and interpretability of AD-related diagnosis. Approach. We develop a novel method for early diagnosis of AD based on multi-scale discriminative regions in FDG-PET imaging, which considers the diagnosis interpretability. Specifically, a multi-scale region localization module is discussed to automatically identify disease-related discriminative regions in full-volume FDG-PET images in an unsupervised manner, upon which a confidence score is designed to evaluate the prioritization of regions according to the density distribution of anomalies. Then, the proposed multi-scale region classification module adaptively fuses multi-scale region representations and makes decision fusion, which not only reduces useless information but also offers complementary information. Most of previous methods concentrate on discriminating AD from cognitively normal (CN), while mild cognitive impairment, a transitional state, facilitates early diagnosis. Therefore, our method is further applied to multiple AD-related diagnosis tasks, not limited to AD vs. CN. Main results. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative dataset show that the proposed method achieves superior performance over state-of-the-art FDG-PET-based approaches. Besides, some cerebral cortices highlighted by extracted regions cohere with medical research, further demonstrating the superiority. Significance. This work offers an effective method to achieve AD diagnosis and detect disease-affected regions in FDG-PET imaging. Our results could be beneficial for providing an additional opinion on the clinical diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'A novel method for early diagnosis of AD based on multi-scale discriminative regions in FDG-PET imaging, which considers the diagnosis interpretability, and is applied to multiple AD-related diagnosis tasks, not limited to AD vs. CN.'}",
-chain-of-thought prompt distillation for multimodal named entity recognition and multimodal relation extraction,F. Chen,"Multimodal Named Entity Recognition (MNER) and Multimodal Relation Extraction (MRE) necessitate the fundamental reasoning capacity for intricate linguistic and multimodal comprehension. In this study, we explore distilling the reasoning ability of large language models (LLMs) into a more compact student model by generating a \textit{chain of thought} (CoT) -- a sequence of intermediate reasoning steps. Specifically, we commence by exemplifying the elicitation of such reasoning ability from LLMs through CoT prompts covering multi-grain (noun, sentence, multimodality) and data-augmentation (style, entity, image) dimensions. Subsequently, we present a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge. Extensive experiments reveal that our approach attains state-of-the-art accuracy and manifests a plethora of advantages concerning interpretability, data efficiency, and cross-domain generalization on MNER and MRE datasets.","{'model': 'tldr@v2.0.0', 'text': 'This study presents a novel conditional prompt distillation method to assimilate the commonsense reasoning ability from LLMs, thereby enhancing the utility of the student model in addressing text-only inputs without the requisite addition of image and CoT knowledge.'}",
-litsumm: large language models for literature summarisation of non-coding rnas,Andrew Green,"Motivation: Curation of literature in life sciences is a growing challenge. The continued increase in the rate of publication, coupled with the relatively fixed number of curators worldwide presents a major challenge to developers of biomedical knowledgebases. Very few knowledgebases have resources to scale to the whole relevant literature and all have to prioritise their efforts. Results: In this work, we take a first step to alleviating the lack of curator time in RNA science by generating summaries of literature for non-coding RNAs using large language models (LLMs). We demonstrate that high-quality, factually accurate summaries with accurate references can be automatically generated from the literature using a commercial LLM and a chain of prompts and checks. Manual assessment was carried out for a subset of summaries, with the majority being rated extremely high quality. We also applied the most commonly used automated evaluation approaches, finding that they do not correlate with human assessment. Finally, we apply our tool to a selection of over 4,600 ncRNAs and make the generated summaries available via the RNAcentral resource. We conclude that automated literature summarization is feasible with the current generation of LLMs, provided careful prompting and automated checking are applied. Availability: Code used to produce these summaries can be found here: https://github.com/RNAcentral/litscan-summarization and the dataset of contexts and summaries can be found here: https://huggingface.co/datasets/RNAcentral/litsumm-v1. Summaries are also displayed on the RNA report pages in RNAcentral (https://rnacentral.org/)","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that high-quality, factually accurate summaries with accurate references can be automatically generated from the literature using a commercial LLM and a chain of prompts and checks, concluding that automated literature summarization is feasible with the current generation of LLMs.'}",
-who wrote it and why? prompting large-language models for authorship verification,Chia-Yu Hung,"Authorship verification (AV) is a fundamental task in natural language processing (NLP) and computational linguistics, with applications in forensic analysis, plagiarism detection, and identification of deceptive content. Existing AV techniques, including traditional stylometric and deep learning approaches, face limitations in terms of data requirements and lack of explainability. To address these limitations, this paper proposes PromptAV, a novel technique that leverages Large-Language Models (LLMs) for AV by providing step-by-step stylometric explanation prompts. PromptAV outperforms state-of-the-art baselines, operates effectively with limited training data, and enhances interpretability through intuitive explanations, showcasing its potential as an effective and interpretable solution for the AV task.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2310.08123
-alltogether: investigating the efficacy of spliced prompt for web navigation using large language models,Jiarun Liu,"Large Language Models (LLMs) have emerged as promising agents for web navigation tasks, interpreting objectives and interacting with web pages. However, the efficiency of spliced prompts for such tasks remains underexplored. We introduces AllTogether, a standardized prompt template that enhances task context representation, thereby improving LLMs' performance in HTML-based web navigation. We evaluate the efficacy of this approach through prompt learning and instruction finetuning based on open-source Llama-2 and API-accessible GPT models. Our results reveal that models like GPT-4 outperform smaller models in web navigation tasks. Additionally, we find that the length of HTML snippet and history trajectory significantly influence performance, and prior step-by-step instructions prove less effective than real-time environmental feedback. Overall, we believe our work provides valuable insights for future research in LLM-driven web agents.","{'model': 'tldr@v2.0.0', 'text': ""AllTogether, a standardized prompt template that enhances task context representation, thereby improving LLMs' performance in HTML-based web navigation, is introduced, revealing that models like GPT-4 outperform smaller models in web navigation tasks.""}",
-wordart designer: user-driven artistic typography synthesis using large language models,Jun-Yan He,"This paper introduces""WordArt Designer"", a user-driven framework for artistic typography synthesis, relying on Large Language Models (LLM). The system incorporates four key modules: the""LLM Engine"",""SemTypo"",""StyTypo"", and""TexTypo""modules. 1) The""LLM Engine"", empowered by LLM (e.g., GPT-3.5-turbo), interprets user inputs and generates actionable prompts for the other modules, thereby transforming abstract concepts into tangible designs. 2) The""SemTypo module""optimizes font designs using semantic concepts, striking a balance between artistic transformation and readability. 3) Building on the semantic layout provided by the""SemTypo module"", the""StyTypo module""creates smooth, refined images. 4) The""TexTypo module""further enhances the design's aesthetics through texture rendering, enabling the generation of inventive textured fonts. Notably,""WordArt Designer""highlights the fusion of generative AI with artistic typography. Experience its capabilities on ModelScope: https://www.modelscope.cn/studios/WordArt/WordArt.",,
-context-aware prompt tuning for vision-language model with dual-alignment,Hongyu Hu,"Large-scale vision-language models (VLMs), e.g., CLIP, learn broad visual concepts from tedious training data, showing superb generalization ability. Amount of prompt learning methods have been proposed to efficiently adapt the VLMs to downstream tasks with only a few training samples. We introduce a novel method to improve the prompt learning of vision-language models by incorporating pre-trained large language models (LLMs), called Dual-Aligned Prompt Tuning (DuAl-PT). Learnable prompts, like CoOp, implicitly model the context through end-to-end training, which are difficult to control and interpret. While explicit context descriptions generated by LLMs, like GPT-3, can be directly used for zero-shot classification, such prompts are overly relying on LLMs and still underexplored in few-shot domains. With DuAl-PT, we propose to learn more context-aware prompts, benefiting from both explicit and implicit context modeling. To achieve this, we introduce a pre-trained LLM to generate context descriptions, and we encourage the prompts to learn from the LLM's knowledge by alignment, as well as the alignment between prompts and local image features. Empirically, DuAl-PT achieves superior performance on 11 downstream datasets on few-shot recognition and base-to-new generalization. Hopefully, DuAl-PT can serve as a strong baseline. Code will be available.","{'model': 'tldr@v2.0.0', 'text': 'A novel method to improve the prompt learning of vision-language models by incorporating pre-trained large language models (LLMs), called Dual-Aligned Prompt Tuning (DuAl-PT), which achieves superior performance on 11 downstream datasets on few-shot recognition and base-to-new generalization.'}",https://arxiv.org/pdf/2309.04158
-fairmonitor: a four-stage automatic framework for detecting stereotypes and biases in large language models,Yanhong Bai,"Detecting stereotypes and biases in Large Language Models (LLMs) can enhance fairness and reduce adverse impacts on individuals or groups when these LLMs are applied. However, the majority of existing methods focus on measuring the model's preference towards sentences containing biases and stereotypes within datasets, which lacks interpretability and cannot detect implicit biases and stereotypes in the real world. To address this gap, this paper introduces a four-stage framework to directly evaluate stereotypes and biases in the generated content of LLMs, including direct inquiry testing, serial or adapted story testing, implicit association testing, and unknown situation testing. Additionally, the paper proposes multi-dimensional evaluation metrics and explainable zero-shot prompts for automated evaluation. Using the education sector as a case study, we constructed the Edu-FairMonitor based on the four-stage framework, which encompasses 12,632 open-ended questions covering nine sensitive factors and 26 educational scenarios. Experimental results reveal varying degrees of stereotypes and biases in five LLMs evaluated on Edu-FairMonitor. Moreover, the results of our proposed automated evaluation method have shown a high correlation with human annotations.","{'model': 'tldr@v2.0.0', 'text': 'A four-stage framework to directly evaluate stereotypes and biases in the generated content of LLMs, including direct inquiry testing, serial or adapted story testing, implicit association testing, and unknown situation testing is introduced.'}",
-de-diffusion makes text a strong cross-modal interface,Chen Wei,"We demonstrate text as a strong cross-modal interface. Rather than relying on deep embeddings to connect image and language as the interface representation, our approach represents an image as text, from which we enjoy the interpretability and flexibility inherent to natural language. We employ an autoencoder that uses a pre-trained text-to-image diffusion model for decoding. The encoder is trained to transform an input image into text, which is then fed into the fixed text-to-image diffusion decoder to reconstruct the original input -- a process we term De-Diffusion. Experiments validate both the precision and comprehensiveness of De-Diffusion text representing images, such that it can be readily ingested by off-the-shelf text-to-image tools and LLMs for diverse multi-modal tasks. For example, a single De-Diffusion model can generalize to provide transferable prompts for different text-to-image tools, and also achieves a new state of the art on open-ended vision-language tasks by simply prompting large language models with few-shot examples.","{'model': 'tldr@v2.0.0', 'text': 'De-Diffusion text representing images can be readily ingested by off-the-shelf text-to-image tools and LLMs for diverse multi-modal tasks and achieves a new state of the art on open-ended vision-language tasks by simply prompting large language models with few-shot examples.'}",
-artificial intelligence and implications for the australian social work journal,David Hodgson,"Social work is a profession committed to integrity and social justice. The AASW Social Work Practice Standards (AASW, 2023) calls on social workers to be critically reflective, ethical practitioners engaged in lifelong professional development and learning. Equally, social work education seeks to prepare students for research-informed, culturally-responsive practice across a diverse range of contexts, and in this Issue, we showcase critical social work education and practice diversity. However, a different ethical challenge to integrity and practice standards is the focus of this Editorial. Here, we highlight some of the concerns and implications of generative Artificial Intelligence (generative AI) for social work education, research, practice, and scholarly publishing. In November 2022, OpenAI released ChatGPT, a generative AI Large Language Model (LLM) that could generate realistic and natural text outputs from simple prompts. This technology had been in development for some time but had not been released to the public for general use. Since then, there has been a proliferation of different AI models that can generate and augment text, images, video, and audio. Generative AI is being used to perform analytical and interpretive tasks such as language translation; responding to queries on specific data sources, coding, and interpreting code; summarising documents and webpages; and creating case assessments and plans. This technology can be used to construct legal documents; machine learning for facial recognition; and for undertaking medical, mental health, and other diagnostic assessments. These are just some examples. In this fast-moving field, the uses and applications seem endless. The open-sourcing of generative AI models and their underlying architecture means developers are starting to create a myriad of practical applications and tools that rapidly increase the depth and scale of automation, potentially replacing or augmenting many everyday tasks normally performed by humans. The implications for social work education, practice, research, and scholarship are extensive. As with any new technology, there are a range of stances, from early adopters to positions that have resonance with luddism. This adds to the complexities of responding to AI as a whole profession. Nevertheless, what is clear is that the rise and integration of generative AI systems, at scale, will yield a wide range of practical, ethical, and epistemological problems for many professions, including social work. It is to some of these problems we turn our attention below. Beginning with social work education, generative AI will have profound effects on assessment and learning for higher education providers. It is likely to cause educators to re-evaluate their educational practices, assessments, and assumptions about what is core to a social work curriculum. Social work will need to refine and reappraise its ideas about critical thinking, ethical decision making, professional judgement, and reflective practice—all skills that are considered core to effective social work practice as outlined in the AASW Practice Standards (AASW, 2023). How will we ensure students have an educational environment that promotes","{'model': 'tldr@v2.0.0', 'text': 'Some of the concerns and implications of generative Artificial Intelligence (generative AI) for social work education, research, practice, and scholarly publishing are highlighted.'}",https://www.tandfonline.com/doi/pdf/10.1080/0312407X.2023.2247833?needAccess=true
-structural embeddings of tools for large language models,Eren Unlu,"It is evident that the current state of Large Language Models (LLMs) necessitates the incorporation of external tools. The lack of straightforward algebraic and logical reasoning is well documented and prompted researchers to develop frameworks which allow LLMs to operate via external tools. The ontological nature of tool utilization for a specific task can be well formulated with a Directed Acyclic Graph (DAG). The central aim of the paper is to highlight the importance of graph based approaches to LLM-tool interaction in near future. We propose an exemplary framework to guide the orchestration of exponentially increasing numbers of external tools with LLMs,where objectives and functionalities of tools are graph encoded hierarchically. Assuming that textual segments of a Chain-of-Thought (CoT) can be imagined as a tool as defined here, the graph based framework can pave new avenues in that particular direction as well.","{'model': 'tldr@v2.0.0', 'text': 'An exemplary framework to guide the orchestration of exponentially increasing numbers of external tools with LLMs is proposed, where objectives and functionalities of tools are graph encoded hierarchically.'}",https://arxiv.org/pdf/2308.00447
-llamarec: two-stage recommendation using large language models for ranking,Zhenrui Yue,"Recently, large language models (LLMs) have exhibited significant progress in language understanding and generation. By leveraging textual features, customized LLMs are also applied for recommendation and demonstrate improvements across diverse recommendation scenarios. Yet the majority of existing methods perform training-free recommendation that heavily relies on pretrained knowledge (e.g., movie recommendation). In addition, inference on LLMs is slow due to autoregressive generation, rendering existing methods less effective for real-time recommendation. As such, we propose a two-stage framework using large language models for ranking-based recommendation (LlamaRec). In particular, we use small-scale sequential recommenders to retrieve candidates based on the user interaction history. Then, both history and retrieved items are fed to the LLM in text via a carefully designed prompt template. Instead of generating next-item titles, we adopt a verbalizer-based approach that transforms output logits into probability distributions over the candidate items. Therefore, the proposed LlamaRec can efficiently rank items without generating long text. To validate the effectiveness of the proposed framework, we compare against state-of-the-art baseline methods on benchmark datasets. Our experimental results demonstrate the performance of LlamaRec, which consistently achieves superior performance in both recommendation performance and efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a two-stage framework using large language models for ranking-based recommendation (LlamaRec) that uses small-scale sequential recommenders to retrieve candidates based on the user interaction history and adopts a verbalizer-based approach that transforms output logits into probability distributions over the candidate items.'}",
-evaluating diverse large language models for automatic and general bug reproduction,Sungmin Kang,"Bug reproduction is a critical developer activity that is also challenging to automate, as bug reports are often in natural language and thus can be difficult to transform to test cases consistently. As a result, existing techniques mostly focused on crash bugs, which are easier to automatically detect and verify. In this work, we overcome this limitation by using large language models (LLMs), which have been demonstrated to be adept at natural language processing and code generation. By prompting LLMs to generate bug-reproducing tests, and via a post-processing pipeline to automatically identify promising generated tests, our proposed technique LIBRO could successfully reproduce about one-third of all bugs in the widely used Defects4J benchmark. Furthermore, our extensive evaluation on 15 LLMs, including 11 open-source LLMs, suggests that open-source LLMs also demonstrate substantial potential, with the StarCoder LLM achieving 70% of the reproduction performance of the closed-source OpenAI LLM code-davinci-002 on the large Defects4J benchmark, and 90% of performance on a held-out bug dataset likely not part of any LLM's training data. In addition, our experiments on LLMs of different sizes show that bug reproduction using LIBRO improves as LLM size increases, providing information as to which LLMs can be used with the LIBRO pipeline.","{'model': 'tldr@v2.0.0', 'text': 'The proposed technique LIBRO could successfully reproduce about one-third of all bugs in the widely used Defects4J benchmark, and experiments on LLMs of different sizes show that bug reproduction using LIBRO improves as LLM size increases, providing information as to which LLMs can be used with the LIBRO pipeline.'}",
-text-based predictions of covid-19 diagnosis from self-reported chemosensory descriptions,Hongyang Li,,"{'model': 'tldr@v2.0.0', 'text': 'The results show that the description of perceptual symptoms caused by a viral infection can be used to fine-tune an LLM model to correctly predict and interpret the diagnostic status of a subject.'}",https://www.nature.com/articles/s43856-023-00334-5.pdf
-scpatcher: mining crowd security discussions to enrich secure coding practices,Ziyou Jiang,"Secure coding practices (SCPs) have been proposed to guide software developers to write code securely to prevent potential security vulnerabilities. Yet, they are typically one-sentence principles without detailed specifications, e.g., “Properly free allocated memory upon the completion of functions and at all exit points.”, which makes them difficult to follow in practice, especially for software developers who are not yet experienced in secure programming. To address this problem, this paper proposes SCPatcher, an automated approach to enrich secure coding practices by mining crowd security discussions on online knowledge-sharing platforms, such as Stack Overflow. In particular, for each security post, SCPatcher first extracts the area of coding examples and coding explanations with a fix-prompt tuned Large Language Model (LLM) via Prompt Learning. Then, it hierarchically slices the lengthy code into coding examples and summarizes the coding explanations with the areas. Finally, SCPatcher matches the CWE and Public SCP, integrating them with extracted coding examples and explanations to form the SCP specifications, which are the wild SCPs with details, proposed by the developers. To evaluate the performance of SCPatcher, we conduct experiments on 3,907 security posts from Stack Overflow. The experimental results show that SCPatcher outperforms all baselines in extracting the coding examples with 2.73 % MLine on average, as well as coding explanations with 3.97 % F1 on average. Moreover, we apply SCPatcher on 447 new security posts to further evaluate its practicality, and the extracted SCP specifications enrich the public SCPs with 3,074 lines of code and 1,967 sentences.","{'model': 'tldr@v2.0.0', 'text': 'SCPatcher is proposed, an automated approach to enrich secure coding practices by mining crowd security discussions on online knowledge-sharing platforms, such as Stack Overflow, with outperforms all baselines in extracting coding examples and coding explanations.'}",
-mind the instructions: a holistic evaluation of consistency and interactions in prompt-based learning,Lucas Weber,"Finding the best way of adapting pre-trained language models to a task is a big challenge in current NLP. Just like the previous generation of task-tuned models (TT), models that are adapted to tasks via in-context-learning (ICL) are robust in some setups but not in others. Here, we present a detailed analysis of which design choices cause instabilities and inconsistencies in LLM predictions. First, we show how spurious correlations between input distributions and labels -- a known issue in TT models -- form only a minor problem for prompted models. Then, we engage in a systematic, holistic evaluation of different factors that have been found to influence predictions in a prompting setup. We test all possible combinations of a range of factors on both vanilla and instruction-tuned (IT) LLMs of different scale and statistically analyse the results to show which factors are the most influential, interactive or stable. Our results show which factors can be used without precautions and which should be avoided or handled with care in most settings.","{'model': 'tldr@v2.0.0', 'text': 'A detailed analysis of which design choices cause instabilities and inconsistencies in LLM predictions is presented and which factors can be used without precautions and which should be avoided or handled with care in most settings.'}",
-botchat: evaluating llms' capabilities of having multi-turn dialogues,Haodong Duan,"Interacting with human via high-quality multi-turn dialogues is a key feature of large language models (LLMs). However, human-based evaluation of such capability involves intensive manual labor. This report provides a preliminary evaluation of existing large language models for human-style multi-turn chatting, through an LLM-based approach. We start from real-world human dialogues and keep the very first utterances as the ChatSEED. Then we prompt LLMs to generate a full multi-turn dialogue (tens of utterances) based on the ChatSEED, utterance by utterance. Finally, we adopt state-of-the-art LLMs (GPT-4, \etc) as the judge to evaluate the generated dialogues. With different evaluation protocols, we come to substantially identical conclusions. We find that GPT-4 can generate human-style multi-turn dialogues with impressive quality, significantly outperforms its counterparts. It's difficult for a discriminator to distinguish between GPT-4 generated dialogues and human dialogues. In contrast, other LLMs struggle to generate multi-turn dialogues of satisfactory quality due to poor instruction-following capability, tendency to generate lengthy utterances, or limited general capability. All data and codes will be provided in https://github.com/open-compass/BotChat/ and we hope they can serve as a valuable resource for evaluating multi-turn chatting capabilities of LLMs.","{'model': 'tldr@v2.0.0', 'text': 'It is found that GPT-4 can generate human-style multi-turn dialogues with impressive quality, significantly outperforms its counterparts and is difficult for a discriminator to distinguish between G PT-4 generated dialogues and human dialogues.'}",
-collaborative large language model for recommender systems,Yaochen Zhu,"Recently, there is a growing interest in developing next-generation recommender systems (RSs) based on pretrained large language models (LLMs), fully utilizing their encoded knowledge and reasoning ability. However, the semantic gap between natural language and recommendation tasks is still not well addressed, leading to multiple issues such as spuriously-correlated user/item descriptors, ineffective language modeling on user/item contents, and inefficient recommendations via auto-regression, etc. In this paper, we propose CLLM4Rec, the first generative RS that tightly integrates the LLM paradigm and ID paradigm of RS, aiming to address the above challenges simultaneously. We first extend the vocabulary of pretrained LLMs with user/item ID tokens to faithfully model the user/item collaborative and content semantics. Accordingly, in the pretraining stage, a novel soft+hard prompting strategy is proposed to effectively learn user/item collaborative/content token embeddings via language modeling on RS-specific corpora established from user-item interactions and user/item features, where each document is split into a prompt consisting of heterogeneous soft (user/item) tokens and hard (vocab) tokens and a main text consisting of homogeneous item tokens or vocab tokens that facilitates stable and effective language modeling. In addition, a novel mutual regularization strategy is introduced to encourage the CLLM4Rec to capture recommendation-oriented information from user/item contents. Finally, we propose a novel recommendation-oriented finetuning strategy for CLLM4Rec, where an item prediction head with multinomial likelihood is added to the pretrained CLLM4Rec backbone to predict hold-out items based on the soft+hard prompts established from masked user-item interaction history, where recommendations of multiple items can be generated efficiently.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes CLLM4Rec, the first generative RS that tightly integrates the LLM paradigm and ID paradigm of RS, aiming to address the above challenges simultaneously, and proposes a novel recommendation-oriented finetuning strategy for CLLm4Rec.'}",
-optimal scene graph planning with large language model guidance,Zhirui Dai,"Recent advances in metric, semantic, and topological mapping have equipped autonomous robots with semantic concept grounding capabilities to interpret natural language tasks. This work aims to leverage these new capabilities with an efficient task planning algorithm for hierarchical metric-semantic models. We consider a scene graph representation of the environment and utilize a large language model (LLM) to convert a natural language task into a linear temporal logic (LTL) automaton. Our main contribution is to enable optimal hierarchical LTL planning with LLM guidance over scene graphs. To achieve efficiency, we construct a hierarchical planning domain that captures the attributes and connectivity of the scene graph and the task automaton, and provide semantic guidance via an LLM heuristic function. To guarantee optimality, we design an LTL heuristic function that is provably consistent and supplements the potentially inadmissible LLM guidance in multi-heuristic planning. We demonstrate efficient planning of complex natural language tasks in scene graphs of virtualized real environments.","{'model': 'tldr@v2.0.0', 'text': 'This work considers a scene graph representation of the environment and utilizes a large language model (LLM) to convert a natural language task into a linear temporal logic (LTL) automaton, and designs an LTL heuristic function that is provably consistent and supplements the potentially inadmissible LLM guidance in multi-heuristic planning.'}",https://arxiv.org/pdf/2309.09182
-can large language models be good path planners? a benchmark and investigation on spatial-temporal reasoning,Mohamed Aghzal,"Large language models (LLMs) have achieved remarkable success across a wide spectrum of tasks; however, they still face limitations in scenarios that demand long-term planning and spatial reasoning. To facilitate this line of research, in this work, we propose a new benchmark, termed $\textbf{P}$ath $\textbf{P}$lanning from $\textbf{N}$atural $\textbf{L}$anguage ($\textbf{PPNL}$). Our benchmark evaluates LLMs' spatial-temporal reasoning by formulating ''path planning'' tasks that require an LLM to navigate to target locations while avoiding obstacles and adhering to constraints. Leveraging this benchmark, we systematically investigate LLMs including GPT-4 via different few-shot prompting methodologies and BART and T5 of various sizes via fine-tuning. Our experimental results show the promise of few-shot GPT-4 in spatial reasoning, when it is prompted to reason and act interleavedly, although it still fails to make long-term temporal reasoning. In contrast, while fine-tuned LLMs achieved impressive results on in-distribution reasoning tasks, they struggled to generalize to larger environments or environments with more obstacles.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show the promise of few-shot GPT-4 in spatial reasoning, when it is prompted to reason and act interleavedly, although it still fails to make long-term temporal reasoning.'}",https://arxiv.org/pdf/2310.03249
-automated assessment of comprehension strategies from self-explanations using llms,Bogdan-Ioan Nicula,"Text comprehension is an essential skill in today’s information-rich world, and self-explanation practice helps students improve their understanding of complex texts. This study was centered on leveraging open-source Large Language Models (LLMs), specifically FLAN-T5, to automatically assess the comprehension strategies employed by readers while understanding Science, Technology, Engineering, and Mathematics (STEM) texts. The experiments relied on a corpus of three datasets (N = 11,833) with self-explanations annotated on 4 dimensions: 3 comprehension strategies (i.e., bridging, elaboration, and paraphrasing) and overall quality. Besides FLAN-T5, we also considered GPT3.5-turbo to establish a stronger baseline. Our experiments indicated that the performance improved with fine-tuning, having a larger LLM model, and providing examples via the prompt. Our best model considered a pretrained FLAN-T5 XXL model and obtained a weighted F1-score of 0.721, surpassing the 0.699 F1-score previously obtained using smaller models (i.e., RoBERTa).",,https://www.mdpi.com/2078-2489/14/10/567/pdf?version=1697284775
-linc: a neurosymbolic approach for logical reasoning by combining language models with first-order logic provers,Theo X. Olausson,"Logical reasoning, i.e., deductively inferring the truth value of a conclusion from a set of premises, is an important task for artificial intelligence with wide potential impacts on science, mathematics, and society. While many prompting-based strategies have been proposed to enable Large Language Models (LLMs) to do such reasoning more effectively, they still appear unsatisfactory, often failing in subtle and unpredictable ways. In this work, we investigate the validity of instead reformulating such tasks as modular neurosymbolic programming, which we call LINC: Logical Inference via Neurosymbolic Computation. In LINC, the LLM acts as a semantic parser, translating premises and conclusions from natural language to expressions in first-order logic. These expressions are then offloaded to an external theorem prover, which symbolically performs deductive inference. Leveraging this approach, we observe significant performance gains on FOLIO and a balanced subset of ProofWriter for three different models in nearly all experimental conditions we evaluate. On ProofWriter, augmenting the comparatively small open-source StarCoder+ (15.5B parameters) with LINC even outperforms GPT-3.5 and GPT-4 with Chain-of-Thought (CoT) prompting by an absolute 38% and 10%, respectively. When used with GPT-4, LINC scores 26% higher than CoT on ProofWriter while performing comparatively on FOLIO. Further analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes. We thus provide promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers. All corresponding code is publicly available at https://github.com/benlipkin/linc","{'model': 'tldr@v2.0.0', 'text': 'Analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes, which provides promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers.'}",
-amortizing intractable inference in large language models,Edward J. Hu,"Autoregressive large language models (LLMs) compress knowledge from their training data through next-token conditional distributions. This limits tractable querying of this knowledge to start-to-end autoregressive sampling. However, many tasks of interest -- including sequence continuation, infilling, and other forms of constrained generation -- involve sampling from intractable posterior distributions. We address this limitation by using amortized Bayesian inference to sample from these intractable posteriors. Such amortization is algorithmically achieved by fine-tuning LLMs via diversity-seeking reinforcement learning algorithms: generative flow networks (GFlowNets). We empirically demonstrate that this distribution-matching paradigm of LLM fine-tuning can serve as an effective alternative to maximum-likelihood training and reward-maximizing policy optimization. As an important application, we interpret chain-of-thought reasoning as a latent variable modeling problem and demonstrate that our approach enables data-efficient adaptation of LLMs to tasks that require multi-step rationalization and tool use.","{'model': 'tldr@v2.0.0', 'text': 'This work interprets chain-of-thought reasoning as a latent variable modeling problem and demonstrates that this distribution-matching paradigm of LLM fine-tuning can serve as an effective alternative to maximum-likelihood training and reward-maximizing policy optimization.'}",
-fabricator: an open source toolkit for generating labeled training data with teacher llms,Jonas Golde,"Most NLP tasks are modeled as supervised learning and thus require labeled training data to train effective models. However, manually producing such data at sufficient quality and quantity is known to be costly and time-intensive. Current research addresses this bottleneck by exploring a novel paradigm called zero-shot learning via dataset generation. Here, a powerful LLM is prompted with a task description to generate labeled data that can be used to train a downstream NLP model. For instance, an LLM might be prompted to""generate 500 movie reviews with positive overall sentiment, and another 500 with negative sentiment.""The generated data could then be used to train a binary sentiment classifier, effectively leveraging an LLM as a teacher to a smaller student model. With this demo, we introduce Fabricator, an open-source Python toolkit for dataset generation. Fabricator implements common dataset generation workflows, supports a wide range of downstream NLP tasks (such as text classification, question answering, and entity recognition), and is integrated with well-known libraries to facilitate quick experimentation. With Fabricator, we aim to support researchers in conducting reproducible dataset generation experiments using LLMs and help practitioners apply this approach to train models for downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'With Fabricator, an open-source Python toolkit for dataset generation, this work aims to support researchers in conducting reproducible dataset generation experiments using LLMs and help practitioners apply this approach to train models for downstream tasks.'}",https://arxiv.org/pdf/2309.09582
-booookscore: a systematic exploration of book-length summarization in the era of llms,Yapei Chang,"Summarizing book-length documents (>100K tokens) that exceed the context window size of large language models (LLMs) requires first breaking the input document into smaller chunks and then prompting an LLM to merge, update, and compress chunk-level summaries. Despite the complexity and importance of this task, it has yet to be meaningfully studied due to the challenges of evaluation: existing book-length summarization datasets (e.g., BookSum) are in the pretraining data of most public LLMs, and existing evaluation methods struggle to capture errors made by modern LLM summarizers. In this paper, we present the first study of the coherence of LLM-based book-length summarizers implemented via two prompting workflows: (1) hierarchically merging chunk-level summaries, and (2) incrementally updating a running summary. We obtain 1193 fine-grained human annotations on GPT-4 generated summaries of 100 recently-published books and identify eight common types of coherence errors made by LLMs. Because human evaluation is expensive and time-consuming, we develop an automatic metric, BooookScore, that measures the proportion of sentences in a summary that do not contain any of the identified error types. BooookScore has high agreement with human annotations and allows us to systematically evaluate the impact of many other critical parameters (e.g., chunk size, base LLM) while saving $15K and 500 hours in human evaluation costs. We find that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher BooookScore than the oft-repetitive ones generated by LLaMA 2. Incremental updating yields lower BooookScore but higher level of detail than hierarchical merging, a trade-off sometimes preferred by human annotators. We release code and annotations after blind review to spur more principled research on book-length summarization.","{'model': 'tldr@v2.0.0', 'text': 'An automatic metric, BooookScore, is developed that measures the proportion of sentences in a summary that do not contain any of the identified error types and is found that closed-source LLMs such as GPT-4 and Claude 2 produce summaries with higher Booook score than the oft-repetitive ones generated by LLaMA 2.'}",https://arxiv.org/pdf/2310.00785
-a prefrontal cortex-inspired architecture for planning in large language models,Taylor Webb,"Large language models (LLMs) demonstrate impressive performance on a wide variety of tasks, but they often struggle with tasks that require multi-step reasoning or goal-directed planning. To address this, we take inspiration from the human brain, in which planning is accomplished via the recurrent interaction of specialized modules in the prefrontal cortex (PFC). These modules perform functions such as conflict monitoring, state prediction, state evaluation, task decomposition, and task coordination. We find that LLMs are sometimes capable of carrying out these functions in isolation, but struggle to autonomously coordinate them in the service of a goal. Therefore, we propose a black box architecture with multiple LLM-based (GPT-4) modules. The architecture improves planning through the interaction of specialized PFC-inspired modules that break down a larger problem into multiple brief automated calls to the LLM. We evaluate the combined architecture on two challenging planning tasks -- graph traversal and Tower of Hanoi -- finding that it yields significant improvements over standard LLM methods (e.g., zero-shot prompting or in-context learning). These results demonstrate the benefit of utilizing knowledge from cognitive neuroscience to improve planning in LLMs.","{'model': 'tldr@v2.0.0', 'text': 'A black box architecture with multiple LLM-based (GPT-4) modules that improves planning through the interaction of specialized PFC-inspired modules that break down a larger problem into multiple brief automated calls to the LLM.'}",https://arxiv.org/pdf/2310.00194
-walking down the memory maze: beyond context limit through interactive reading,Howard Chen,"Large language models (LLMs) have advanced in large strides due to the effectiveness of the self-attention mechanism that processes and compares all tokens at once. However, this mechanism comes with a fundamental issue -- the predetermined context window is bound to be limited. Despite attempts to extend the context window through methods like extrapolating the positional embedding, using recurrence, or selectively retrieving essential parts of the long sequence, long-text understanding continues to be a challenge. We propose an alternative approach which instead treats the LLM as an interactive agent, allowing it to decide how to read the text via iterative prompting. We introduce MemWalker, a method that first processes the long context into a tree of summary nodes. Upon receiving a query, the model navigates this tree in search of relevant information, and responds once it gathers sufficient information. On long-text question answering tasks our method outperforms baseline approaches that use long context windows, recurrence, and retrieval. We show that, beyond effective reading, MemWalker enhances explainability by highlighting the reasoning steps as it interactively reads the text; pinpointing the relevant text segments related to the query.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces MemWalker, a method that first processes the long context into a tree of summary nodes, and upon receiving a query, the model navigates this tree in search of relevant information, and responds once it gathers sufficient information.'}",https://arxiv.org/pdf/2310.05029
-scalable multi-robot collaboration with large language models: centralized or decentralized systems?,Yongchao Chen,"A flurry of recent work has demonstrated that pre-trained large language models (LLMs) can be effective task planners for a variety of single-robot tasks. The planning performance of LLMs is significantly improved via prompting techniques, such as in-context learning or re-prompting with state feedback, placing new importance on the token budget for the context window. An under-explored but natural next direction is to investigate LLMs as multi-robot task planners. However, long-horizon, heterogeneous multi-robot planning introduces new challenges of coordination while also pushing up against the limits of context window length. It is therefore critical to find token-efficient LLM planning frameworks that are also able to reason about the complexities of multi-robot coordination. In this work, we compare the task success rate and token efficiency of four multi-agent communication frameworks (centralized, decentralized, and two hybrid) as applied to four coordination-dependent multi-agent 2D task scenarios for increasing numbers of agents. We find that a hybrid framework achieves better task success rates across all four tasks and scales better to more agents. We further demonstrate the hybrid frameworks in 3D simulations where the vision-to-text problem and dynamical errors are considered. See our project website https://yongchao98.github.io/MIT-REALM-Multi-Robot/ for prompts, videos, and code.","{'model': 'tldr@v2.0.0', 'text': 'This work compares the task success rate and token efficiency of four multi-agent communication frameworks as applied to four coordination-dependent multi- agent 2D task scenarios for increasing numbers of agents and finds that a hybrid framework achieves better task success rates across all four tasks and scales better to more agents.'}",https://arxiv.org/pdf/2309.15943
-distilled language models are economically efficient for the enterprise. ...mostly.,Kristen Howell,"Contacting customer service via chat is a common practice. Because employing customer service agents is expensive, many companies are turning to NLP that assists human agents by auto-generating responses that can be used directly or with modifications. With their ability to handle large context windows, Large Language Models (LLMs) are a natural fit for this use case. However, their efficacy must be balanced with the cost of training and serving them. This paper assesses the practical cost and impact of LLMs for the enterprise as a function of the usefulness of the responses that they generate. We present a cost framework for evaluating an NLP model’s utility for this use case and apply it to a single brand as a case study in the context of an existing agent assistance product. We compare three strategies for specializing an LLM — prompt engineering, fine-tuning, and knowledge distillation — using feedback from the brand’s customer service agents. We find that the usability of a model’s responses can make up for a large difference in inference cost for our case study brand, and we extrapolate our findings to the broader enterprise space.",,http://arxiv.org/pdf/2306.07402
-making multimodal generation easier: when diffusion models meet llms,Xiangyu Zhao,"We present EasyGen, an efficient model designed to enhance multimodal understanding and generation by harnessing the capabilities of diffusion models and large language models (LLMs). Unlike existing multimodal models that predominately depend on encoders like CLIP or ImageBind and need ample amounts of training data to bridge the gap between modalities, EasyGen is built upon a bidirectional conditional diffusion model named BiDiffuser, which promotes more efficient interactions between modalities. EasyGen handles image-to-text generation by integrating BiDiffuser and an LLM via a simple projection layer. Unlike most existing multimodal models that are limited to generating text responses, EasyGen can also facilitate text-to-image generation by leveraging the LLM to create textual descriptions, which can be interpreted by BiDiffuser to generate appropriate visual responses. Extensive quantitative and qualitative experiments demonstrate the effectiveness of EasyGen, whose training can be easily achieved in a lab setting. The source code is available at https://github.com/zxy556677/EasyGen.","{'model': 'tldr@v2.0.0', 'text': 'Efficient model designed to enhance multimodal understanding and generation by harnessing the capabilities of diffusion models and large language models, built upon a bidirectional conditional diffusion model named BiDiffuser, which promotes more efficient interactions between modalities.'}",https://arxiv.org/pdf/2310.08949
-curriculum prompt learning with self-training for abstractive dialogue summarization,Changqun Li,"Succinctly summarizing dialogue is a task of growing interest, but inherent challenges, such as insufficient training data and low information density impede our ability to train abstractive models. In this work, we propose a novel curriculum-based prompt learning method with self-training to address these problems. Specifically, prompts are learned using a curriculum learning strategy that gradually increases the degree of prompt perturbation, thereby improving the dialogue understanding and modeling capabilities of our model. Unlabeled dialogue is incorporated by means of self-training so as to reduce the dependency on labeled data. We further investigate topic-aware prompts to better plan for the generation of summaries. Experiments confirm that our model substantially outperforms strong baselines and achieves new state-of-the-art results on the AMI and ICSI datasets. Human evaluations also show the superiority of our model with regard to the summary generation quality.","{'model': 'tldr@v2.0.0', 'text': 'A novel curriculum-based prompt learning method with self-training that substantially outperforms strong baselines and achieves new state-of-the-art results on the AMI and ICSI datasets is proposed.'}",https://aclanthology.org/2022.emnlp-main.72.pdf
-questioning prompts to examine blended pk-20 learning environments,Neal Shambaugh,"The decision to deliver a course or an academic program must address issues of learning outcomes and curriculum, what is to be taught along with choices of pedagogy or how the content is taught, as well as face-to-face or online delivery. The question of whether or not online or blended instruction is more effective than face to face teaching is not about the delivery decision but about curriculum and pedagogical issues. A set of questioning prompts, organized by public school and undergraduate/graduate programs, help educators systematically think through issues of learner differences, learning outcomes, teaching options, including the use of technology and instructional delivery approaches. Best practices are organized by public school and undergraduate/graduate programs. Research questions are suggested.",,
-practicing what we preach: incorporating team-based learning into the pre-service teaching curriculum for improved outcomes,Jane S. Vogler,"Background: Prior research has shown that pre-service teachers’ learning of theory is enhanced when teacher educators model the instructional approaches that they advocate as effective. Objective: The purpose of this study was to compare undergraduate learning outcomes in a course using Team-based Learning (TBL) with those in a traditional, lecture-based approach. Method: Occurring across two consecutive semesters, we recruited students (n = 43) enrolled in an educational psychology course serving primarily pre-service teachers for this quasi-experimental study. In addition to final grades, data sources included responses to essay prompts administered at the beginning and end of the semester, which were qualitatively analyzed and coded. Results: Independent samples t-tests showed significant differences in favor of students in the TBL condition for course grades, but not on the final exam. χ2 tests of independence revealed significant differences for two of the seven essay codes, again in favor of TBL. Conclusion: These results provide further evidence that when placed in a context that emphasizes collaborative learning, students demonstrate more nuanced understanding and fewer misconceptions. Teaching Implications: The opportunity to experience TBL may provide pre-service teachers with better theoretical understanding and an effective model for translating socio-constructive theory into collaborative learning practices.",,
-covid-19 pandemic prompts the development of a web-osce using zoom teleconferencing to resume medical students’ clinical skills training at weill cornell medicine-qatar,S. Major,"Web-Objective Structured Clinical Examination (OSCEs) were piloted for the Clinical Skills Assessment (CSA) exams in the USA two decades ago and were shown to be an acceptable way to conduct OSCEs remotely. The learners valued the process yet expressed numerous limitations.1 In response to the COVID-19 outbreak in Qatar, medical schools stopped in-person teaching creating a hiatus in clinical instruction. The Clinical Skills and Simulation Lab (CSSL) team devised and piloted a Web-OSCE to determine its feasibility in this setting, and to evaluate stakeholders’ experiences. We describe the steps taken to create a Web-OSCE built on women’s reproductive and sexual health which is part of the third year Obstetrics and Gynecology Clerkship curriculum, using Zoom teleconferencing. It mirrored the steps taken when conducting this activity in person; however, all communication relied on emails before the event, Zoom during the event with WhatsApp as a backup for connectivity between learners, Zoom hosts and faculty. ### Pre-Web-OSCE steps #### Preparing students CSSL provided by email a step-by-step instruction guide plus a pre-OSCE task for students to complete to ensure that they could access the OSCE on the day of the activity. A live Zoom session led by faculty reviewed the learning objectives, expectations and assessment and served as an open forum for students to ask questions in order to mitigate any anxiety surrounding the new modality. #### Preparing simulated patients Simulated patients (SP) were surveyed to assess their interest and technology capabilities to participate in Zoom activities. Interested SPs joined a Zoom meeting with staff members, who ensured they had adequate technical capability. SPs were selected and confirmed according to requirements of the case. Training notes and checklists were sent in …","{'model': 'tldr@v2.0.0', 'text': 'The steps taken to create a Web-OSCE built on women’s reproductive and sexual health which is part of the third year Obstetrics and Gynecology Clerkship curriculum, using Zoom teleconferencing are described.'}",https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7239650/pdf/bmjstel-2020-000629.pdf
-book review: numeracy across the curriculum: research strategies for enhancing teaching and learning,Glenda Anthony,"In Numeracy across the curriculum: Research strategies for enhancing teaching and learning, the authors provide a powerful resource in the form of practical, research-based strategies that will support prospective teachers, teachers, and teacher educators as they respond to the expectation that educators, across all levels and within all disciplines, will know and understand numeracy teaching strategies and their application. A critical and exciting feature of the book is the clear articulation that numeracy must move from a “ mathematics concern ” to an across curriculum focus. The authenticity of the advocated practices is informed by the authors’ many years of research in Australian educational, home, work, and community/civic settings. To that end, the text carefully explains the nature of numeracy and how it developed as an educational goal for the mathematics classroom and beyond. From the beginning chapters, and throughout, readers are prompted to engage in productive learnings with the inclusion of reflection prompts and numerous classroom exemplars of task design, assessment practices, and how to deal with the challenges and dilemmas of enactment in practice. urgent",,
-pedagogical prompts: designing experiences to promote deep learning,Nicola Yelland,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter focuses on the ways in which the Learning by Design theoretical framework acts as a pedagogical prompt for teachers so that they are able to create learning experiences that complement the range of pedagogies described as knowledge processes and support 21st century skills and deep learning.'}",
-challenges in developing automatic learning guidance in relation to an information literacy curriculum,Angela Fessl,"Becoming a data-savvy professional requires skills and competences in information literacy, communication and collaboration, and content creation in digital environments. In this paper, we present a concept for automatic learning guidance in relation to an information literacy curriculum. The learning guidance concept has three components: Firstly, an open learner model in terms of an information literacy curriculum is created. Based on the data collected in the learner model, learning analytics is used in combination with a corresponding visualization to present the current learning status of the learner. Secondly, reflection prompts in form of sentence starters or reflective questions adaptive to the learner model aim to guide learning. Thirdly, learning resources are suggested that are structured along learning goals to motivate learners to progress. The main contribution of this paper is to discuss what we see as main research challenges with respect to existing literature on open learner modeling, learning analytics, recommender systems for learning, and learning guidance.","{'model': 'tldr@v2.0.0', 'text': 'What the authors see as main research challenges with respect to existing literature on open learner modeling, learning analytics, recommender systems for learning, and learning guidance are discussed.'}",
-implementation and evaluation of a 10-week health equity curriculum for pharmacy students,S. Hsia,"Objective. To describe a health equity curriculum created for pharmacy students and evaluate students’ perceptions and structural competency after completion of the curriculum. Methods. A health equity curriculum based on transformative learning and structural competency frameworks was implemented as a 10-week mandatory component of the pass-no pass neuropsychiatric theme for second year pharmacy students. Each week, students reviewed materials around a neuropsychiatric-related health equity topic and responded to discussion prompts through asynchronous forums or synchronous online video discussions. The effectiveness of the health equity curriculum was evaluated through assessment of structural competency through a validated instrument, an objective structured clinical examination (OSCE), and a questionnaire. Results. All enrolled second year pharmacy students (n=124) participated in the health equity curriculum. Of the 75 students who completed the structural competency instrument, 46 (61%) were able to identify structural determinants of health, explain how structures contribute to health disparities, or design structural interventions. Ninety-six of the 124 students (77%) were able to address their OSCE standardized patient’s mistrust in the health care system. Thematic analysis of student comments elucidated three themes: allyship, peer connection, and self-awareness. Students rated asynchronous discussion forums as significantly less effective than online video discussions and patient cases for achieving curricular objectives. Conclusion. A mandatory curriculum delivered remotely throughout the didactic pharmacy curriculum using a blended learning approach was an effective way to incorporate health equity content and conversations into existing courses. Implementation of this or similar curriculums could be an important step in training pharmacy students to be advocates for social justice.","{'model': 'tldr@v2.0.0', 'text': 'A mandatory curriculum delivered remotely throughout the didactic pharmacy curriculum using a blended learning approach was an effective way to incorporate health equity content and conversations into existing courses.'}",https://www.ajpe.org/content/ajpe/85/9/8579.full.pdf
-what counts as nature in designing environmental links to health education curriculum in initial teacher education?,Bronwyn J. Cumbo,"ABSTRACT In this paper we report on a teacher education co-design project that explored Australian pre-service primary generalist teachers’ ideations of the pedagogical links between health education and nature. As part of their coursework in a Master of Teaching degree at an Australian University, students were invited to design a Victorian Curriculum: Health and Physical Education (VC:HPE) activity that connected primary curriculum (F-6) with nature. We conducted a 3-hour suite of online learning activities and prompts using Zoom, Padlet and Moodle. The data consists of the students’ curriculum design artefacts as well as recordings of the group discussions and non-assessment-based presentations of their work. Drawing on theories of child-nature interaction, we present an analysis of the ways in which n = 72 pre-service teachers across 18 groups of 4–5 educators conceptualised links between ‘nature’ and HPE across their activity designs. Prior to the workshop, the majority of pre-service teachers had not previously considered links between the HPE learning area and nature. Through the activities of the co-design workshop, students were surprised with the variety of pedagogical possibilities that were able to be made. There was much student discussion about the possibilities and limitations of balancing safety and risk in their nature-based activity designs. We present a thematic analysis of the quality of student-nature interactions in the groups’ learning designs through: (i) exploration; (ii) embodiment; (iii) cultivation; (iv) appropriation; and (v) representation. The analysis and discussion has implications for the way quality health education is linked to nature-based learning environments, teacher education and contemporary curriculum enactment that incorporates nature and the environment as part of the learning design.",,
-curriculum q-learning for visual vocabulary acquisition,Ahmed H. Zaidi,"The structure of curriculum plays a vital role in our learning process, both as children and adults. Presenting material in ascending order of difficulty that also exploits prior knowledge can have a significant impact on the rate of learning. However, the notion of difficulty and prior knowledge differs from person to person. Motivated by the need for a personalised curriculum, we present a novel method of curriculum learning for vocabulary words in the form of visual prompts. We employ a reinforcement learning model grounded in pedagogical theories that emulates the actions of a tutor. We simulate three students with different levels of vocabulary knowledge in order to evaluate the how well our model adapts to the environment. The results of the simulation reveal that through interaction, the model is able to identify the areas of weakness, as well as push students to the edge of their ZPD. We hypothesise that these methods can also be effective in training agents to learn language representations in a simulated environment where it has previously been shown that order of words and prior knowledge play an important role in the efficacy of language learning.","{'model': 'tldr@v2.0.0', 'text': 'It is hypothesised that these methods can also be effective in training agents to learn language representations in a simulated environment where it has previously been shown that order of words and prior knowledge play an important role in the efficacy of language learning.'}",
-redesigning curriculum at the higher education level: challenges and successes within a sport management program,Wendi K. Zimmer,"ABSTRACT Higher education programs face challenges to increase accountability regarding student achievement. Course misalignment occurs as instructors interpret course objectives differently, making changes to the course content without considering overall program alignment and purpose. Misalignment prompts academic programs to revisit its curriculum to assess effectiveness. Faced with a rapidly growing and increasingly diverse student population, the Sport Management (SPMT) division at a southwestern university used action research to redesign their undergraduate curriculum to meet their students’ needs. This study assesses the impact of using action research to determine why the SPMT program exists, what a successful SPMT program includes, and what knowledge, skills, abilities, and values industry leaders report necessary for SPMT graduates. Furthermore, methods for addressing challenges and successes through action research are discussed. As a result of the action research process and data collection, eight program learning outcomes were established to guide the next steps of curriculum development. These findings guided syllabus creation for a redesigned SPMT program, allowing for the creation of a rigorous, research based, and customized curriculum.",,
-analyzing preservice secondary mathematics teachers’ prompts in dynamic geometry environment tasks,Hilal Gulkilik,"ABSTRACT The purpose of this study was to analyze the prompts that preservice secondary mathematics teachers used for the acquisition of mathematics knowledge in dynamic geometry environment tasks. The participants, four preservice secondary mathematics teachers who were enrolled in a computer-supported mathematics education course, designed a dynamic geometry environment task based on a high school mathematics curriculum learning outcome. The main data sources were the participants’ task documents and the transcripts of interviews that were conducted with the participants to examine the details of their tasks. The techno-pedagogic mathematics task design model was used to analyze the data. The results indicated that the focus of the designs was to help students realize the invariant properties of geometric figures that were embodied by robust construction steps in the tasks. The preservice teachers utilized several capabilities of the dynamic geometry environment (e.g. measuring, dragging, and changing the input box value) to help students discern these properties and expected students to make a generalization based on inductive reasoning. However, since students were directed to build robust constructions, the mathematical activities prompted by the preservice teachers were limited to observing, finding, generalizing, and explaining the previously constructed invariant features in the tasks.",,
-the impact of scaffolding prompts on the collaborative problem solving of ill-structured tasks by undergraduate engineering student groups,Miss Taylor Tucker,"Taylor Tucker received her Bachelor’s degree in engineering mechanics from the University of Illinois at Urbana-Champaign and is now pursuing a Master’s in curriculum and instruction through the Digital Environments for Learning, Teaching, and Agency (DELTA) program. She is interested in engineering design and lends her technical background to her research with the Collaborative Learning Lab, exploring how to improve ill-structured tasks for engineering students in order to promote collaborative problem solving and provide experience relevant to authentic work in industry.",,
-support factors and barriers for outdoor learning in elementary schools: a systemic perspective,E. Oberle,"ABSTRACT Background Outdoor learning offers clear physical, cognitive, social-emotional and academic benefits for children and yet, it is considered a grassroots approach to teaching and learning in elementary schools. Purpose We examined teachers’ perspectives on barriers and supports for outdoor learning in public elementary schools. Methods Thirty-six teachers in (urban and rural) British Columbia, Alberta, and Ontario (all female; Mean age = 43.84, SD = 10) participated in one of five virtually administered, semi-structured focus groups. Questions/prompts facilitated a discussion on teachers’ experiences with barriers and supports for outdoor learning. Thematic analysis was used to identify main themes. Results Four interrelated themes and further sub-themes were found: 1) Teacher characteristics: interest/motivation to teach outdoors, preparedness, confidence in handling risks; 2) Systemic factors: principal support, school/district policies, funding/resources, curriculum, school schedule; 3) Culture: school culture, societal beliefs about education, family backgrounds; 4) Environmental factors: weather, built/natural environment, hazards. Discussion Systemic support is needed to integrate outdoor learning in schools. Translation to Health Education Practice The findings in this study are relevant to health education specialists particularly focused on elementary school education.",,
-evaluating the impact of reflecting on curriculum-embedded skill development: the experience of science undergraduates,Michelle A. Hill,"ABSTRACT Meaningful reflection on their learning and skill development is often lacking in the experience of undergraduates. Many students do not recognise the curriculum-embedded development of transferable skills and lack the ability to articulate such skills. This mixed-methods study sought to investigate whether engaging students in reflection would increase their ability to recognise and articulate their skill development. Sixty science undergraduates from Monash University completed a voluntary semester-long program recording and reflecting on course-related skill development, supported by email prompts and group discussions. The impact of students’ involvement was evaluated through pre- and post-participation surveys, reflections and group discussions. Most students were challenged by the unfamiliarity of thinking beyond knowledge attainment in order to identify and reflect on skill-related experiences. However, they recognised a range of benefits from doing so, including an improved ability to recognise their skill development, strengths and weaknesses and to articulate their skills in readiness for seeking employment. They also valued previously unappreciated learning tasks and gained motivation to improve skill deficits and seek out opportunities to improve their employability. Based on this study, recommendations are made regarding best practice for implementing skills reflection in the curriculum.",,
-meta-augmented prompt tuning for better few-shot learning,Kaihang Pan,"Prompt tuning is a parameter-efficient method, which freezes all PLM parameters and only prepends some additional tunable tokens called soft prompts to the input text. However, soft prompts heavily rely on a better initialization and may easily result in overfitting under few-shot settings, which causes prompt-tuning performing much worse than fine-tuning. To address the above issues, this paper proposes a novel S elf-s U pervised M eta-prompt learning framework with ME ta-gradient R egularization for few-shot generalization ( SUMMER ). We leverage self-supervised meta-learning to better initialize soft prompts and curriculum-based task augmentation is further proposed to enrich the meta-task distribution. Besides, a novel meta-gradient regularization method is integrated into the meta-prompt learning framework, which meta-learns to transform the raw gradient during few-shot learning into a domain- generalizable direction, thus alleviat-ing the problem of overfitting. Extensive experiments show that SUMMER achieves better performance for different few-shot downstream tasks, and also exhibits a stronger domain generalization ability.","{'model': 'tldr@v2.0.0', 'text': 'A novel meta-gradient regularization method is integrated into the meta-prompt learning framework, which meta-learns to transform the raw gradient during few-shot learning into a domain- generalizable direction, thus alleviat-ing the problem of overfitting.'}",http://arxiv.org/pdf/2303.12314
-learning to critique and adapt science curriculum materials: examining the development of preservice elementary teachers' pedagogical content knowledge,Carrie J. Beyer,"Teachers often engage in curricular planning by critiquing and adapting existing curriculum materials to contextualize lessons and compensate for their deficiencies. Designing instruction for students is shaped by teachers' ability to apply a variety of personal resources, including their pedagogical content knowledge (PCK). This study investigated a criterion-based approach to lesson plan analysis as one way to help preservice elementary teachers develop and use their PCK to plan instruction for students. Results show that the preservice teachers demonstrated a range of strengths and weaknesses in applying their knowledge of science assessment, science curriculum materials, and instructional strategies for teaching science. This range was influenced, in part, by the presence of alternative ideas about science teaching, the extent to which the original curriculum materials aligned with reform-based standards and practices, and the presence of prompts to use criteria in their analyses. Despite these factors, preservice teachers' PCK improved significantly over time when they had multiple opportunities to practice applying the same criterion in their analyses. Insights into science teacher knowledge and implications for science teacher education are discussed. © 2011 Wiley Periodicals, Inc. Sci Ed96:130–157, 2012",,http://deepblue.lib.umich.edu/bitstream/2027.42/89555/1/20466_ftp.pdf
-promoting social emotional learning through shared book reading: examining teacher’s strategies and children’s responses in kindergarten classrooms,S. Ng,"ABSTRACT Research Findings: Early childhood educators play an important role in promoting children’s social emotional learning (SEL). This study proposes a comprehensive coding scheme to explore teachers’ strategies and children’s responses in shared book reading (SBR), which create opportunities to share pictures and storylines to examine social circumstances and emotional situations. Drawing on videotaped observations of nineteen classrooms from a larger study in Singapore, SEL-related interactions consisted of illustrations or discussions with indicators from the five SEL areas outlined in the national curriculum. Results revealed that teachers’ support for SEL occurred more frequently when facilitating interpersonal learning over intrapersonal learning, while children’s responses accorded with the types of teachers’ input. In response to teachers’ strategies within the intrapersonal dimension, children provided responses to teachers’ prompts for identifying self-abilities and preferences most frequently. Within the interpersonal dimension, children provided responses to teachers’ prompts for labelling others’ emotions most frequently. Practice or Policy: Our findings demonstrate that SBR can be useful in promoting SEL even without books that explicitly convey skills and messages relating to SEL. The results imply that teachers should increase their input on facilitating intrapersonal learning topromote children’s balanced development in the understanding of themselves and others.",,
-improving significance of binary black hole mergers in advanced ligo data using deep learning : confirmation of gw151216,S. Jadhav,"We present a novel Machine Learning (ML) based strategy to search for compact binary coalescences (CBCs) in data from ground-based gravitational wave (GW) observatories. This is the first ML-based search that not only recovers all the binary black hole mergers in the first GW transients calalog (GWTC-1), but also makes a clean detection of GW151216, which was not significant enough to be included in the catalogue. Moreover, we achieve this by only adding a new coincident ranking statistic (MLStat) to a standard analysis that was used for GWTC-1. In CBC searches, reducing contamination by terrestrial and instrumental transients, which create a loud noise background by triggering numerous false alarms, is crucial to improving the sensitivity for detecting true events. The sheer volume of data and and large number of expected detections also prompts the use of ML techniques. We perform transfer learning to train ""InceptionV3"", a pre-trained deep neural network, along with curriculum learning to distinguish GW signals from noisy events by analysing their continuous wavelet transform (CWT) maps. MLStat incorporates information from this ML classifier into the standard coincident search likelihood used by the conventional search. This leads to at least an order of magnitude improvement in the inverse false-alarm-rate (IFAR) for the previously ""low significance"" events GW151012, GW170729 and GW151216. The confidence in detection of GW151216 is further strengthened by performing its parameter estimation using SEOBNRv4HM_ROM. Considering the impressive ability of the statistic to distinguish signals from glitches, the list of marginal events from MLStat could be quite reliable for astrophysical population studies and further follow-up. This work demonstrates the immense potential and readiness of MLStat for finding new sources in current data and possibility of its adaptation in similar searches.","{'model': 'tldr@v2.0.0', 'text': 'This is the first ML-based search that recovers all the binary black hole mergers in the first GW transients calalog (GWTC-1), but also makes a clean detection of GW151216, which was not significant enough to be included in the catalogue.'}",https://dspace.library.uu.nl/bitstream/handle/1874/413428/PhysRevD.104.064051.pdf?sequence=1&isAllowed=y
-the impact of multimedia educative curriculum materials (mecms) on teachers’ beliefs about scientific argumentation,S. Loper,"ABSTRACT Recent reform efforts in science education include a focus on science practices. Teachers require support in integrating these practices into instruction. Multimedia educative curriculum materials (MECMs), digital materials explicitly designed to support teacher learning, offer one potential resource for this critical need. Consequently, the authors investigated how teachers used MECMs and whether that use impacted their beliefs about the practice of scientific argumentation. They conducted a randomised experimental study with 90 middle school science teachers in the USA. Control and experimental groups taught the same curriculum, using a web-based teacher’s guide. Additionally, experimental teachers received MECMs including 24 videos and 17 interactive reflective prompts. The authors collected multiple data sources: pre-surveys, backend website curriculum use, self-report curriculum use and post-surveys. Results suggest that enacting a curriculum with a focus on argumentation is associated with positive changes in teachers’ beliefs about this practice. Furthermore, the authors observed a wide range in how teachers used the curriculum. In terms of self-efficacy, this differential use was associated with differences in changes to teachers’ beliefs about argumentation. Teachers who enacted more lessons became more confident in their ability to teach argumentation. Additionally, experimental teachers had smaller improvements in self-efficacy, perhaps because the MECM videos problematised what teachers thought counted as argumentation.",,
-"pythonsneks: an open-source, instructionally-designed introductory curriculum with action-design research",Austin Cory Bart,"Rising enrollments and limited instructor resources underscores the growing need for reusable, scalable curriculum. In this paper, we describe an open-source introductory Python course for non-Computer Science majors in STEM, designed following best practices of Instructional Design (a process similar to Software Engineering). The created resources include 234 learning objectives, 51 lesson videos, 45 lecture slides, 170 programming problems, 281 quiz questions, 6 unit tested projects, and 4 ethical prompts. A teaching field guide has also been produced as a result of this effort, documenting how to deploy this curriculum on a daily level. We describe our experiences deploying over two semesters. The course serviced over 500 students, with 100s in some sections. Along the way, two interventions were conducted in an Action Design Research style: one using Worked Examples, and another using Structured Small Groups. We report on the mixed results of these experiments, plus evaluations of the assignments from student surveys and statistical measures of item effectiveness. Finally, we describe lessons learned when following Instructional Design processes.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes an open-source introductory Python course for non-Computer Science majors in STEM, designed following best practices of Instructional Design (a process similar to Software Engineering).'}",
-toward expert thinking: how curriculum case writing prompts the development of theory-based professional knowledge in student teachers,Karen Hammerness,"The present paper explores what, and how, student teachers may learn about theory and practice from writing cases, and examines some pedagogical features that may contribute to these results. Drawing on data collected from our course ""Principles of Learning for Teaching"", including student cases from outline to final drafts and students' course reflections, we found that students' successive case drafts demonstrated a development from naïve generalizations to sophisticated, theory-based explanations of the issues at play in their cases. In particular, we suggest that students' cases demonstrated some of the moves that Berliner (1986, 1991) identified as characteristic of more ""expert"" thinking about teaching. We propose that reading theory in context with writing cases, that sharing cases with peer readers, that specific, theoretically grounded, and concrete feedback from instructors, and that providing multiple opportunities for revision may have been most useful in helping student teachers learn to think like a teacher.",,
-the curriculum innovation canvas: a design thinking framework for the engaged educational entrepreneur,C. Willness,"Integrating literature on entrepreneurial business models and community-based experiential learning, we propose a new framework to advance the practice of curriculum innovation. Grounded in principles of design thinking, the curriculum innovation canvas provides a human-centered, collaborative, and holistic platform for instructors, curriculum developers, and administrators to engage in innovation and implementation of experiential courses or programs—particularly those that involve community or organizational partnerships. The canvas promotes a creative and fluid approach to curriculum development. It prompts the consideration of the value propositions offered to various stakeholders (students, community partners, faculty peers, etc.) as well as how to involve stakeholders in the development and implementation process toward mutually beneficial outcomes in a complex and challenging environment. Evidence from an extensive prototyping process indicates that it can effectively assist instructors, administrators, students, and community partners in a variety of contexts.",,
-"learning chemistry: self-efficacy, chemical understanding, and graphing skills",Shirly Avargil,,,
-what millennial medical students say about flipped learning,Robin K Pettit,"Flipped instruction is gaining popularity in medical schools, but there are unanswered questions such as the optimum amount of the curriculum to flip and whether flipped sessions should be mandatory. We were in a unique position to evaluate feedback from first-year medical students who had experienced both flipped and lecture-based courses during their first semester of medical school. A key finding was that the students preferred a variety of different learning formats over an “all or nothing” learning format. Learning format preferences did not necessarily align with perceptions of which format led to better course exam performance. Nearly 70% of respondents wanted to make their own decisions regarding attendance. Candid responses to open-ended survey prompts reflected millennial preferences for choice, flexibility, efficiency, and the ability to control the pace of their learning, providing insight to guide curricular improvements.","{'model': 'tldr@v2.0.0', 'text': 'Feedback from first-year medical students who had experienced both flipped and lecture-based courses during their first semester of medical school found that the students preferred a variety of different learning formats over an “all or nothing” learning format.'}",https://www.dovepress.com/getfile.php?fileID=37537
-improving self-regulated learning through personalized weekly e-learning journals: a time series quasi-experimental study.,Chorng Yuan Fung,This study provides an insight into using personalized weekly e-Learning Journals to improve self-regulated learning (SRL) of university students. Quasi-experimental method with time series data analysis was used. Pre and post-tests together with time series data over an intervention period on SRL were collected. 54 students in an undergraduate course wrote personalized weekly e-Learning Journals (e-LJs) in the Learning Management System (LMS) over a 10-week period. The e-LJs contained selfreflection prompts designed according to the course curriculum activities and assessments to scaffold students’ SRL. It was found that students’ SRL improved significantly over the intervention period. The time series data on SRL shows that students’ SRL varied according to the timing of assessments. Academic staff can help university students improve their SRL by providing personalized weekly e-LJs that contained self-reflection prompts. These prompts need to be personalized according to the course curriculum activities and assessments.,,
-exploring the impacts of cognitive and metacognitive prompting on students’ scientific inquiry practices within an e-learning environment,Wenxia Zhang,"This study explores the effects of metacognitive and cognitive prompting on the scientific inquiry practices of students with various levels of initial metacognition. Two junior high school classes participated in this study. One class, the experimental group (n = 26), which received an inquiry-based curriculum with a combination of cognitive and metacognitive prompts, was compared to the other class, the comparison group (n = 25), which received only cognitive prompts in the same curriculum. Data sources included a test of inquiry practices, a questionnaire of metacognition, and worksheets. The results showed that the mixed cognitive and metacognitive prompts had significant impacts on the students’ inquiry practices, especially their planning and analyzing abilities. Furthermore, the mixed prompts appeared to have a differential effect on those students with lower level metacognition, who showed significant improvement in their inquiry abilities. A combination of cognitive and metacognitive prompts during an inquiry cycle was found to promote students’ inquiry practices.",,
-vygotsky in action in the early years: the key to learning curriculum,G. Dolya,"@contents: Selected Contents: CONTENTS FOREWORD LEV VYGOTSKY - A THINKER WHO CHANGED TEACHING VIVA VYGOTSKY! DEVELOPING LEARNING ABILITIES, UNLOCKING POSSIBILITIES UNDERSTANDING ABILITIES THE MECHANISM OF ABILITIES THE CLASSIFICATION OF ABILITIES A DEVELOPMENTAL CURRICULUM TWELVE PROGRAMMES, ONE PRACTICE THE PEDAGOGICAL PROCESS GROUP WORK SUBSTITUTION AND VISUAL MODELLING EXTERNAL MEDIATORS: VISIBLE PROPS AND PROMPTS FOR LEARNING COGNITIVE TASKS VISUAL MODELLING AND PRODUCTIVE IMAGINATION IN CONCLUSION FROM SENSORY STANDARDS TO VERBAL REASONING - A THEORETICAL PERSPECTIVE SENSORY STANDARDS AND PERCEPTUAL MODELLING CREATIVE MODELLING MATHEMATICS LOGIC CONSTRUCTION FROM THEORY TO PRACTICE - THE KEY TO LEARNING DEVELOPMENTAL CURRICULUM SENSORY MATHEMATICS LOGIC MATHEMATICS STORY GRAMMAR DEVELOPMENTAL GAMES ARTOGRAPHICS VISUAL-SPATIAL CREATIVE MODELLING CONSTRUCTION EXPLORATION EXPRESSIVE MOVEMENT YOU-ME-WORLD PARENT POWER - STRATEGIES FOR BUILDING A POSITIVE PARTNERSHIP ALL WORK AND NO PLAY IT'S NEVER TOO LATE TO LEARN, IT'S NEVER TOO EARLY TO TEACH ""HOMEWORK"" FOR PARENTS WHAT PEOPLE SAY... PROFESSIONAL OPINIONS WHAT PARENTS SAY REFERENCES",,
-arctic climate connections curriculum: a model for bringing authentic data into the classroom,A. Gold,"ABSTRACT Science education can build a bridge between research carried out by scientists and relevant learning opportunities for students. The Broader Impact requirements for scientists by funding agencies facilitate this connection. We propose and test a model curriculum development process in which scientists, curriculum developers, and classroom educators work together to scaffold the use of authentic, unprocessed scientific data for high school students. We outline a three-module curriculum structure that facilitates these goals. This curriculum engages students in the collection, description, visualization, and interpretation of data; develops understanding of the nature of science; includes prompts to develop higher-order thinking skills; builds knowledge of regional relevance of climate change in students; uses active learning techniques; and can be easily integrated with the Next Generation Science Standards. The curriculum was reviewed and tested in the classroom. To shed further light on the curriculum development process, we gathered reflection data from the scientists, curriculum developers, and educators. Scientists appreciated the collaborative process in which they contributed their expertise without requiring a large time commitment or strong expertise in science education. The curriculum developers viewed the modular structure as helpful in breaking complicated scientific concepts into teachable steps. Classroom educators appreciated the detailed description and step-by-step instructions to navigate data analysis tools like Excel or Google Earth. Initial classroom implementation of the curriculum by 11 teachers with over 1,100 students showed high levels of interest in the topic and engagement. Further work is needed to assess efficacy of the curriculum through classroom observations and measures of student learning.",,
-scaffolding collaborative reflective writing in a vet curriculum,E. Boldrini,,,
-empowering students with word-learning strategies: teach a child to fish,Michael F. Graves,"This article on word-learning strategies describes a theory- and research-based set of procedures for teaching students to use word-learning strategies—word parts, context clues, the dictionary, and a combined strategy—to infer the meanings of unknown words. The article begins with a rationale for teaching word-learning strategies, particularly to students with vocabularies smaller than those of many students their age. After this, the authors provide a definition of word-learning strategies, a review of the most relevant research, and a brief description of the effects of the program. Next are descriptions of the curriculum, the instruction, and key aspects of the authors’ approach. The article concludes with a note stressing the importance of following initial instruction on word-learning strategies with reviews, reminders, and prompts to use them over time and a description of two aspects of the program identified as particularly important by the teachers the authors worked with.",,
-good practices in european short learning programmes (e-slp),M. Maina,"collaborative inter-institutional curriculum designs. The outcome of these collaborations can be SLPs built from scratch or as puzzles (collection of existing learning building blocks). One of the consequences of international collaborative design is that it introduces distinct working practices and therefore requires teams to set innovative processes. It also brings opportunities from each institution to exchange policies and triggers new consensus. Furthermore, it requires harmonisation of recognitions and accreditations, which in turns prompts innovatory agreements.",,
-increasing student learning through arts integration,S. Catherine,"This action research study was designed to examine the effects of the arts integration on student learning. Student learning was broken down by focusing on student engagement, critical thinking, and creative response. Research was conducted with 72 participants from fourth grade, fifth grade, and ninth grade classrooms within the social studies curriculum. Data collection methods included a pre-assessment survey, student journal prompts, observation tally sheets, student self-assessments, and the post-assessment survey. The results of our study indicated that student learning in each classroom maintained or improved when the arts were integrated. Students showed enthusiasm and felt optimistic towards integrating the arts into their learning. According to our student survey many students noted an increase in their critical thinking skills and explained that the arts helped them display their knowledge and creativity. However, in the future we would guide students to be more specific in their journal responses to better communicate their ideas. As a result of our study, we have committed to improving student learning by continuing to integrate the arts into the curriculum.",,
-a making and gaming approach to learning about rf path loss and antenna design,Kevin J. Richardson,"As part of an ongoing, longitudinal study on the use of “making” and “gaming” in the classroom, two sequential activities for learning about radio-frequency (RF) path loss and antenna design are presented. “Making” involves integration of makerspace concepts and tinkering in the curriculum, while “gaming” refers to gamified curricula; in this study we investigate the joint use of these two elements in the classroom. The RF path loss activity is modeled after ham radio “fox hunting”, where students must locate a transmitter hidden on campus; it makes use of low-cost software-defined radios, and prompts students to confront concepts including measuring signal power, frequency domain thinking, and antenna polarization. The follow-up activity challenges students to build an antenna designed to receive household gas meter readings; students must design their antennas specifically for operation in the 900 MHz band, and must give a presentation describing the theory of their antenna to their peers. A competition is held where students attempt to see which of their antennas can collect the most wireless gas meter readings over a five-minute interval. Assessment data from the broader study show that relative to a baseline offering, the treatment group developed an improvement in interest, perception, independence, and self-assessed abilities. This paper discusses the implementation of the activities, the students' approach to solving the proposed challenges, the assessment data, lessons learned from student focus groups, and instructor observations.","{'model': 'tldr@v2.0.0', 'text': 'Assessment data from the broader study show that relative to a baseline offering, the treatment group developed an improvement in interest, perception, independence, and self-assessed abilities.'}",
-"a socially responsive curriculum: an alternative for elt in ""popular education"".",Orley Cruz Arcila,"This study reported on qualitative research that focused on the presentation of the current situation at the school de Cultura Popular with regards to the English Curriculum designed and implemented into the school specifically in 11th grade. This paper also contains a revision of the different social and institutional conditions that may have an impact on English language teaching (ELT) at the school. It was carried out with 36 students of fifth cycle at I.E.D de Cultura Popular, in Bogota. The main aim of this study was to re-design the English curriculum for 11th grade, considering students’ social context and learning interests for ELT by incorporating in its design the prompts of Popular Education, to give to the curriculum a social vision as a new alternative for ELT at the school. The instruments used for collecting data were teacher’s journal, students’ journal; represented in the elaboration of a “book” in English; and some documentary evidence as P.E.I and the socio-economical study of the students to identify in a more realistic way their current social and cultural context. The results of the study showed that the implementation of activities under the orientation of Popular Education, students develop a critical position about their social and educational reality that permits them to establish their own objectives in a long and short term.",,
-hollowed-out genring as a way of purposefully embracing troublesome knowledge: orientation and de-orientation in the learning and teaching of fine art,Shaun Camp,"This text emerges from ongoing research focused on pedagogical developments within the curriculum of the Year 0 Fine Art Pathway, at Norwich University of the Arts. Within any subject discipline, including Fine Art, there are threshold concepts (Meyer and Land, 2003, p.1) that can represent troublesome knowledge (Meyer and Land, 2003, p.2); knowledge that must be comprehended in order that students succeed within that discipline. Within the teaching and learning of Fine Art, students will work between both written and visual genres, often struggling to rationalise the two. As part of our research we have developed practiced based research tasks that serve as destabilising prompts, a hollowed out genre, that encourages students to enter into a de-orientated, liminal, conceptual space within their learning. With the introduction of reflective models, students are able to navigate this liminal space and are afforded opportunities for developmental re-orientations, thus beginning to embrace the threshold concepts inherent with the Fine Art discipline. Drawing upon examples of current practice and student learning, we demonstrate how a process of genring and the use of reflective tools encourage students to confront aspects of troublesome knowledge and see critical writing as an important facet of their practice.",,https://nua.collections.crest.ac.uk/17235/1/RegenringFinal.pdf
-impact of the course teaching and learning of mathematics on preservice grades 7 and 8 mathematics teachers in singapore,Berinderjeet Kaur,,,
-discussion of identity and implicit bias: a strategy to address racism and social justice in pediatric nursing education.,Kristen E. Stevens,"ABSTRACT Nurse educators must weave discussions of systemic racism, social justice, social determinants of health, and psychosocial influences throughout the curriculum. For an online pediatric course, an activity was developed to raise awareness of implicit bias. This experience interfused assigned readings from the literature, introspection of identity, and guided discussion. Framed by principles of transformative learning, faculty facilitated an online dialogue involving groups of 5 to 10 students through aggregated self-descriptors and open prompts. Ground rules for the discussion established psychological safety. This activity complements other schoolwide racial justice initiatives.",,
-blended matters: nurturing critical thinking,Pooja Jaswal,"The Organization for Economic Cooperation and Development (OECD) Learning Framework-2030 presents critical thinking as one of the transformative competencies for incorporation into the curriculum. Critical thinking refers to one’s ability to distinguish relevant information from irrelevant information, relate ideas, infer and justify whether a particular assumption is acceptable or unacceptable. This article argues that blended learning provides a concrete and meaningful context for enhancing learners’ critical thinking. Through collaboration, blended learning makes it possible to control, observe interactions, regulate tasks, and mediate in the acquisition of new knowledge. This form of collaboration creates a student-centered learning environment. Such a shift prompts to examine three objectives: (1) comprehend the creation of a blended learning environment. (2) discuss the characteristics of collaborative blended learning. (3) examine the role of collaboration in developing critical thinking among learners. To achieve the purpose of implementing a blended learning strategy in the classroom, the classroom teacher must create collaborative and participative learning experiences. Blended learning brings collaborative experience to develop higher-order thinking skills such as critical thinking.",,
-analytics for supporting teaching success in higher education: a systematic review,Dirk Ifenthaler,"Learning analytics are utilized to support learners’ educational needs as well as to enhance their study success, for example, via the use of real-time prompts, motivational dashboards, and appropriate learning interventions, which have been shown to increase students’ academic performance as well as their course retention rates. Yet, the perspective of higher education teachers in utilizing analytics to help analyze, reflect on, and improve their teaching design prior to delivery as well as to monitor and measure how the students engaged with their learning processes has been less recognized. In this paper, we present the results of a systematic review conducted from higher education teachers’ perspective concerning how analytics can be deployed to adapt the curriculum to suit better students’ educational needs in order to increase their study success. Thirty-five key studies were identified showing that analytics have been successful in influencing positively study success via teachers’ academic curriculum intervention. Specifically, via analytics, higher education teachers could rapidly visualize common course pathways and identify any difficulties students experienced in real-time in order to increase their learning experiences and outcomes.",,
-the effects of intraprofessional collaborative case based learning: a cohort study of student physical therapist and physical therapist assistant perceptions of the physical therapist role,Y. Colgrove,"Background: The changing nature and complex regulation of healthcare require the efficient use of resources, including the appropriate delegation and supervision of the physical therapist assistant (PTA). Knowledge of the scope of PTA practice introduced in the academic curriculum is mandated for entry-level practice. This study assessed the effect of a collaborative case-based educational intervention within the didactic curriculums of a physical therapy (PT) and PTA program on student knowledge of PTA scope of practice. Methods and Findings: A pre- and post-test research design was used. Students completed a validated survey exploring their perceptions of the PTA role before beginning the case study. The case study was a classroom assignment followed by instructional prompts requiring interactions between student cohorts three times over four weeks. Following case study completion, students completed the same survey. Independent and paired samples t -tests detected significant differences between and within groups ( p < .05). Conclusions: Based on the results, the case-based instructional model was efficacious in teaching both student cohorts about the role of the PTA. The impact was greater on the accuracy of the PT students, but PTA students became less uncertain in their perceptions. The effect of the clinical learning environment should be investigated to determine the impact on student perception of PTA role delineation following didactic instruction.","{'model': 'tldr@v2.0.0', 'text': 'The case-based instructional model was efficacious in teaching both student cohorts about the role of the PTA, and the impact was greater on the accuracy of the PT students, but PTA students became less uncertain in their perceptions.'}",https://jripe.org/index.php/journal/article/download/185/112
-advancing english language learning in china through multimodal content area teaching,C. North,"During a ten-day teaching abroad experience in China, eight teachers from the United States implemented an interactive curriculum focusing on disciplinary literacy and authentic tasks. Employing multiliteracies and kidwatching, teachers encouraged Chinese students to compose while focusing primarily on communicating ideas rather than grammatical correctness. This article provides a one-student case study that serves as a representative example of the growth of 50 elementary-level students involved in the experience. Initially, Paul focused on writing correctness in response to prompts; his compositions were short and provided little detail. After we provided multimodal and interactive authentic experiences and encouraged risk-taking, Paul's representative compositions became more detailed and complex. The implications for engaging in this type of teaching experience underscore the benefits of providing students with authentic experiences that are multimodal and interactive while simultaneously encouraging risk-taking. The pedagogical growth that teachers made working with ELL students is also discussed. Normal 0 false false false EN-US JA X-NONE",,
-could probability be out of proportion? self-explanation and example-based practice help students with lower proportional reasoning skills learn probability,K. Begolli,,,https://link.springer.com/content/pdf/10.1007/s11251-021-09550-9.pdf
-global health on the front lines: an innovative medical student elective combining education and service during the covid-19 pandemic,Brandon Altillo,,"{'model': 'tldr@v2.0.0', 'text': 'The elective was well-received by students, achieved stated objectives, and garnered public attention; course leadership should monitor students’ time commitment closely in service-learning settings to ensure appropriate balance of service and education.'}",https://bmcmededuc.biomedcentral.com/counter/pdf/10.1186/s12909-021-02616-9
-nurturing graduate attributes through a first year student experience which promotes the formation of effective learning communities,D. Leung,"Many students enter university with naive epistemological beliefs and study approaches incompatible with the goals of higher education or the display of attributes such as critical or creative thinking. This study examines whether a first year experience can promote the formation of effective learning communities, which in turn can provide a mechanism for nurturing a range of graduate attributes. The curriculum encompasses initiatives to assimilate students into the university, prompts the adaptation of appropriate university study behaviour, and contains a general education component to broaden the student experience. The impact of the first year experience was examined by collecting quantitative data which measured students’ perceptions of the effectiveness of assimilation, adaptation of study behaviour, the impact of the broadening component, and the nurturing of attributes. The data were analysed with structural equation modelling. A model in which the curriculum elements impacted on the development of graduate attributes showed a good fit to the data. Effective learning communities played a key role both in assimilation into the university community and the adoption of appropriate study behaviours.",,
-autotutor: a tutor with dialogue in natural language,A. Graesser,,"{'model': 'tldr@v2.0.0', 'text': 'The design was inspired by explanation-based constructivist theories of learning, intelligent tutoring systems that adaptively respond to student knowledge, and empirical research on dialogue patterns in tutorial discourse.'}",https://link.springer.com/content/pdf/10.3758/BF03195563.pdf
-heutagogy and e-learning in the workplace: some challenges and opportunities,S. Hase,"It is clear that e-learning offers tremendous opportunities in terms of access in a global environment as well as in terms of the economics of delivery for workplace learning and, indeed, other learning environments. However, those designing and delivering workplace e-learning also have the potential to challenge existing dogmas about the curriculum, how we conduct education programs, manage knowledge, and access and harness learning. Complexity theory, in particular, prompts us to question how we understand education, training and learning. This article discusses some of these challenges and possibilities for workplace e-learning through the lens of the recently developed concept of heutagogy, defined as the study of self-determined learning.",,
-using curriculum-based measurement for struggling beginning,Kristen L. McMaster,"loves to learn. He participates enthusiastically during shared storybook reading with his class. During independent reading time, he becomes immediately engrossed in the colorful books about science and sports that his teacher, Mrs. Graff, displayed on the classroom bookshelf. He is a whiz at mathematics and sometimes asks for extra problems to solve. At home, he reads National Geographic and books about animals, oceans, and dinosaurs with his dad and older sister. Despite Max’s excitement about learning, Mrs. Graff noticed that Max is experiencing difficulties with his handwriting, spelling, and written composition assignments. During daily journal time, Max typically busies himself with sharpening pencils, wandering around the classroom, or staring out the window. After several prompts from Mrs. Graff to get to work, he attempts to write a few words, in barely legible handwriting, with lots of crossing out and erasing, sometimes to the point of making holes in his paper. On several occasions, she has observed him balling up his paper and throwing it off his desk. His mother has expressed concern that when she tries to help him study his spelling words, he becomes easily frustrated. He often fails the Friday spelling test. Mrs. Graff is concerned that Max’s struggles with writing will increasingly affect his enjoyment of learning and his success in school.",,
-a rubric for assessing student drawings & reasoning to gauge understanding of osmosis & diffusion,A. Angra,"Abstract The concepts of osmosis and diffusion are essential to biology, and student difficulties with them are well documented, along with multiple ways of assessment and diagnosis. We add to the literature by sharing how we used drawing and reasoning prompts with first-year undergraduate biology students to gauge learning before and after an osmosis and diffusion lab. We also share the development and validation of a three-point analytical rubric to score the keywords, student drawings, and reasoning prompts. The qualitative and quantitative nature of this exercise provides instructors with the unique opportunity to diagnose difficulties not only in introductory biology but throughout the undergraduate curriculum. Implementation strategies throughout the undergraduate curriculum are discussed.",,
-numberjacks are on their way! a cultural historical reflection on contemporary society and the early childhood curriculum,S. Edwards,"This paper considers the temporal aspects of the early childhood curriculum from a cultural historical perspective, and in doing so focuses on the role of play in early childhood education. Drawing on ideas derived from cultural historical theory regarding the historical basis of community practices and knowledge, the paper reflects on the type of experiences that characterise playful activity for some of today's young children. Examples from previous research conducted by the author are provided as prompts for personal reflection on the temporal dimensions of cultural historical theory and the early childhood curriculum in order to understand the role of contemporary play experiences in children's learning.",,
-comparing elements of study abroad among sport management students,Carrie W. Lecrom,"A globalized curriculum has the potential to prepare students in a way that equips them for whatever sport looks like in the future. Study abroad programs are one way to achieve this. The current study looked at two short-term study abroad programs (one to western Europe, one to South Africa), offered during the same semester at the same institution, comparing learning outcomes between students on the two trips. Utilizing a mixed methods design, students completed quantitative pre/post surveys and responded to qualitative, open-ended daily prompts while on the trips. Findings indicate that knowledge acquisition occurs in both programs; however, students traveling on a sport-focused service-based trip to South Africa had a more transformational learning experience than those traveling on a sport-business-focused trip to western Europe.",,
-effects of formative assessment strategies on the fractions computation skills of students with disabilities,B. Bottge,"Learning to compute with fractions is a major challenge for many students and especially for students with disabilities (SWD). Phase 1 of this study employed a randomized pretest–posttest comparison design to test the effects of two versions of formative assessment combined with an instructional program called Fractions at Work. In one condition, teachers used technology-assisted prompts to assess student performance and remediate errors. In the comparison condition, teachers gave students the same items for assessing progress but used their own methods of reteaching. Results indicated no difference between the two methods. However, pretest-to-posttest gain scores were significantly higher on all three measures regardless of type of formative assessment, and students maintained much of what they had learned. Phase 2 examined issues related to instructional dosage. Students who received additional weeks of instruction scored significantly higher than students who went back to their business-as-usual curriculum.",,
-“how come nobody told me?” fostering self‐realization through a high school english curriculum,Laura T. Eisenman,"Through collaboratively designed qualitative inquiry, we investigated the responses of high school students with learning disabilities to a teacher’s intervention intended to promote self-realization, a fundamental component of self-determination. Activities were embedded within the general English curriculum and delivered in a special education classroom over the course of an academic year. Several themes emerged from analysis of student interviews, student responses to writing prompts and surveys, a teacher journal, and student portfolio pieces. Silence and misconceptions were prevalent in student experiences. However, through the intervention students acquired information that helped them make sense of their school experiences, redefine themselves in positive ways, and take small steps toward greater self-advocacy within their current school setting. The mediating influence of positive adult voices and concerns about social stigma were evident in students’ responses, which prompted us to question teachers’ and families’ responsibilities for engaging young people in dialogue about special education and disability.",,
-students exposed to community violence,Dorothy Rohde-Collins,"Exposure to community violence has a detrimental effect on the academic, social-emotional, and physical functioning of children, adolescents, and their families. Schools and educators are instrumental in enabling children and adults to process violence and trauma in order to develop a healthy worldview. Schools can counter the effects of community violence by emphasizing the development of sincere relationships between staff, students, and families; offering a warm, welcoming atmosphere; and providing a safe school environment conducive to learning. Educators can embed writing prompts and other opportunities for self-reflection within the academic curriculum as well as teach stress management, coping mechanisms, mindfulness, and relaxation techniques which can be used when a demanding situation arises. Educators and administrators are uniquely positioned to offer supports and opportunities within the school context to substantially improve the outcome for students, families, and communities experiencing community violence.",,
-"debate, discourse and productive disagreement: interrogating the performative dimensions of authorship in the creative writing classroom",S. Holland-Batt,"ABSTRACT The creative writing curriculum has historically focussed on discipline-specific skills, developing students’ proficiency in literary forms, craft and techniques. However, contemporary writers are increasingly expected to participate in the economy of ideas through festival appearances, debates, and other forms of public speaking – skills that the creative writing curriculum has yet to fully contend with. We argue for the value of teaching debate as a distinct topic of inquiry for creative writing students, and hold that pedagogical innovation is required to address the changing nature of literary cultures and their increasing orientation towards performativity. This article establishes a new pedagogical model designed to introduce creative writing students to the study and practice of debate, comprising four key stages: modelling, scaffolding, debating and reflection. This learning progression not only fosters students’ oral argumentation skills, but also prompts critical reflection on the way key ideas in their field connect with their creative works. We contend that introducing debate into the creative writing curriculum addresses broader shifts in the writing and publishing industry, and that oral argumentation and debate should be considered key graduate competencies for creative writing students in the twenty-first century.",,https://research-repository.griffith.edu.au/bitstream/10072/419569/2/Jeffery4000400-Accepted.pdf
-"nurturing faculty‐student dialogue, deep learning and creativity through journal writing exercises",B. Korgel,"One of the most difficult endeavors in the engineering curriculum is encouraging students to develop creative, independent thought and a deep level of understanding. To achieve these results, we incorporate journal writing exercises into the curriculum of a core chemical engineering course, ChE 363 (Unit Operations II: Separations) at the University of Texas at Austin. Students receive weekly prompts to develop analogies or thought experiments to describe fundamental concepts prior to discussion in lecture. This motivates students to learn material independently and deeply as they are required to link concepts to common life experiences. The instructor identifies three to five of the most creative and accurate analogies and presents them to the class. The class votes for the “best” one, which earns the student extra credit. The journal writing exercises promote dialogue between the instructor and the students, generating in‐class discussions that often extend well beyond the scope of the journal prompt.",,
-locating teacher voice in curriculum reform,R. Brooker,"Curriculum reform is a complex phenomenon, the outcomes of which may include changes to curriculum requirements, pedagogical practices (including assessment) and learning opportunities. The particular outcomes arising will be the product of influences exerted by multiple agents, representing a variety of voices, in a range of sites, and at particular points in time in a curriculum reform. The process of curriculum reform is, therefore, inherently complex and contested. This chapter prompts exploration of the process, specifically from the perspective of teachers and in relation to the notion of teacher voice in curriculum reform.",,
-the role of self-determination theory and cognitive evaluation theory in home education,Gina Riley,"Abstract This article explores the theories of Self-Determination, Cognitive Evaluation, and Intrinsic Motivation as it applies to home education. According to Self-Determination Theory, intrinsic motivation is innate. However, the maintenance and enhancement of intrinsic motivation depends upon the social and environmental conditions surrounding the individual. Deci and Ryan’s Cognitive Evaluation Theory specifically addresses the social and environmental factors that facilitate versus undermine intrinsic motivation and points to three significant psychological needs that must be present in the individual in order to foster self-motivation. These needs are competence, autonomy, and relatedness. Because of curriculum and time constraints, intrinsic motivation may be difficult to facilitate within the traditional classroom. This loss of intrinsic motivation for learning prompts some parents to homeschool their children. One of the most impressive strengths of home education lies in the fact that in many cases, the entire process revolves around a child’s intrinsic motivation to learn.",,
-taking advantage of automated assessment of student‐constructed graphs in science,Jonathan M. Vitale,"We present a new system for automated scoring of graph construction items that address complex science concepts, feature qualitative prompts, and support a range of possible solutions. This system utilizes analysis of spatial features (e.g., slope of a line) to evaluate potential student ideas represented within graphs. Student ideas are then scored with rubrics based upon the knowledge integration framework (Linn & Eylon, 2011). We tested the effectiveness of this system on graphs constructed by 397 8th–12th grade students preceding, during, and following a curriculum focusing on graphs of motion. Comparison with human-coded responses indicates that the automated scoring system is very accurate (κ = 0.9). Also, ideas represented in constructions were generally similar to those demonstrated in written explanations; although individual students often shifted ideas between items. Learning gains were similar in both written and graph construction formats. Overall, these results suggest that graph construction is a valid and efficient means of evaluating students' complex ideas about data representation in science. We discuss the opportunities for incorporating graph construction into new science content areas, such as graphs representing density. We consider the implications of this system for generating automated, adaptive guidance to support instruction. © 2015 Wiley Periodicals, Inc. J Res Sci Teach 52: 1426–1450, 2015.",,
-relating chemistry to healthcare and more: implementation of more in a survey organic and biochemistry course for prehealth students,L. Schroeder,"We implemented a laboratory curriculum reform to teach foundational concepts in chemistry, particularly those concepts related to healthcare, in a chemistry course for prenursing students. Here, we discuss the reform, exploring how students built upon understandings gained in lab and correlating lab learning to course outcomes. We further discuss shifts in student work as they move through the course. As the course progressed, students became familiar with the pedagogy but also faced more challenging tasks. We present details on several of the laboratories that build the groundwork for understanding chemical principles, including the following: intermolecular forces, physical properties, acid–base chemistry, equilibrium, and chemical reactions. We further share our observations of student interactions around in-lab prompts and activities, and how these interactions inform our teaching. Our reform aims to improve critical thinking skills, namely, making and using models, observation skills, reasoning with ...",,
-using latent semantic analysis to evaluate the contributions of students in autotutor,A. Graesser,"AutoTutor is a fully automated computer tutor that assists students in learning about hardware, operating systems, and the Internet in an introductory computer literacy course. AutoTutor presents questions and problems from a curriculum script, attempts to comprehend learner contributions that are entered by keyboard, formulates dialog moves that are sensitive to the learner's contributions (such as prompts, elaborations, corrections, and hints), and delivers the dialog moves with a talking head. Latent Semantic Analysis (LSA) is a major component of the mechanism that evaluates the quality of student contributions in the tutorial dialog. LSA's evaluations of college students' answers to deep reasoning questions are equivalent to the evaluations provided by intermediate experts of computer literacy, but not as high as more accomplished experts in computer science. LSA is capable of discriminating different classes of student ability (good, vague, erroneous or mute students) and in tracking the quality of contributions in tutorial dialog.","{'model': 'tldr@v2.0.0', 'text': 'Latent Semantic Analysis is a major component of the mechanism that evaluates the quality of student contributions in the tutorial dialog and is capable of discriminating different classes of student ability and in tracking thequality of contributions in tutorial dialog.'}",
-"""virtual assembly"" - a web-based student learning tool for thermodynamics concepts related to multistaging in compressors and turbines",S. Chaturvedi,"Simulation and visualization have been used to develop ""virtual assembly"" as a student learning to ol for comprehension and reinforcement of concepts in basic engineering thermodynamics course in undergraduate engineering curriculum. Using a web-based module described in this study, students are able to assem ble on a computer screen a multistage compressor or turbine from a number of elemental building blocks or stages. The module is interactive and requires students to input data such as overall compressor ratio, stage effici ency, stage pressure ratio, and compressor inlet temperat ure and pressure. A computer program embedded in the module calculates total number of stages, and temperature and pressure at exit section of each st age. It also displays visual images of all stages to be ass embled, and prompts students to use the clicking and draggi ng action of computer mouse to assemble elemental compressor stages into a multistage axial compresso r. During the assembly process, a temperature-entropy diagram is generated, displaying thermodynamic state of air as it traverses through the compressor. These v isual images allow students to explore relationships betw een overall compressor efficiency, stage efficiency and compressor pressure ratio. The module is assessed b y comparing the performance of a ""control"" group (no exposure to module) with an ""experimental"" group (using the module) for an identical quiz administra ted to both groups. Results described in the paper show improvement in the average score for the ""experimental"" group over the ""control"" group.",,
-washington state’s classroom-based performance assessments,AnnRené Joseph,"How does an entire state assess the arts in every school district at the elementary, middle, and high school levels with interrater validity and reliability? This chapter will summarize how Washington State’s Office of Superintendent of Public Instruction (OSPI) created, developed, designed, piloted, assessed, implemented, and reported arts classroom-based performance assessments (CBPAs), 2003–2016. The OSPI music CBPAs measure what educators value and teach, align with state arts learning standards regarding what all students should know and be able to do in music, and empower teachers to change their practice, resulting in practical significance. Developed by music educators, music CBPAs measure how individual students create, perform, and respond to real-life prompts. The CBPAs are formative and summative in design, and have become part of accountability instructional practices and state policy via annual district implementation verification reports. The vision for the initial CBPA design was: Curriculum + Instruction + Assessment for/ = Learning.",,
-the transformation of private tutoring: education in a franchise form,Janice Aurini," Various forms of private education are growing in Canada. This paper explores one instance of this change, private tutoring. Data from Ontario show that this massively growing industry is expanding its reach, as exemplified by the evolution from ""shadow education"" provision into ""learning centre"" franchises. Traditional shadow educators closely follow the school curriculum, offering short-term homework help and test prep. Learning centres develop their own curricular and assessment tools, offer comprehensive menus of services, and aim to nurture long-term skills. In so doing, these businesses are becoming increasingly ""school-like,"" moving beyond supplementary education towards a fuller alternative to public schooling.We link this evolution to the imperatives of the franchise form. Their larger investment prompts franchises to control their services via standardization, to secure stable revenues, and to seek new market niches. In conclusion we discuss the prospects for tutoring and other forms of private education.
Dans le secteur privé, différents modèles d'enseignement se développent actuellement au Canada. Cet article illustre un cas spécifique de cette évolution : la transformation du tutorat. Des données de l'Ontario démontrent que cette industrie, qui connaît un essor formidable, est en pleine expansion, tel qu'on peut le constater par la modification du tutorat, qui mèned'une forme d'éducation parallèle à des centres franchisés d'éducation. Les éducateurs traditionnels qui offrent des services parallèles suivent de près les programmes scolaires, offrant de l'aide à court terme pour les devoirs et pour les préparations aux examens, tandis que les centres d'éducation développent leurs propres outils d'apprentissage et d'évaluation, leurs programmes de services éducatifs polyvalents et visent aussi à entretenir leurs compétences à long terme. Ce faisant, ces entreprises deviennent de plus en plus « semblables à l'école », passant, au-delà de l'enseignement d'appoint, à une alternative plus sérieuse à l'école publique. Nous attribuons cette évolution aux impératifs entraînés par la formule des franchises. Les centres d'éducation, nécessitant de plus gros investissements, poussent les franchisésà standardiser leurs services, à s'assurer des revenus plus stables provenant de sources traditionnelles et à rechercher de nouveaux créneaux dans le domaine. Pour terminer, nous discutons des perspectives à court et à long termes du marché du tutorat et d'autres formes d'éducation privée.
",,
-using a concept inventory to reveal student thinking associated with common misconceptions about antibiotic resistance,S. M. Bornstein-Forst,"Misconceptions, also known as alternate conceptions, about key concepts often hinder the ability of students to learn new knowledge. Concept inventories (CIs) are designed to assess students’ understanding of key concepts, especially those prone to misconceptions. Two-tiered CIs include prompts that ask students to explain the logic behind their answer choice. Such two-tiered CIs afford an opportunity for faculty to explore the student thinking behind the common misconceptions represented by their choice of a distractor. In this study, we specifically sought to probe the misconceptions that students hold prior to beginning an introductory microbiology course (i.e., preconceptions). Faculty-learning communities at two research-intensive universities used the validated Host-Pathogen Interaction Concept Inventory (HPI-CI) to reveal student preconceptions. Our method of deep analysis involved communal review and discussion of students’ explanations for their CI answer choice. This approach provided insight valuable for curriculum development. Here the process is illustrated using one question from the HPI-CI related to the important topic of antibiotic resistance. The frequencies with which students chose particular multiple-choice responses for this question were highly correlated between institutions, implying common underlying misconceptions. Examination of student explanations using our analysis approach, coupled with group discussions within and between institutions, revealed patterns in student thinking to the participating faculty. Similar application of a two-tiered concept inventory by general microbiology instructors, either individually or in groups, at other institutions will allow them to better understand student thinking related to key concepts in their curriculum.","{'model': 'tldr@v2.0.0', 'text': 'This study used the validated Host-Pathogen Interaction Concept Inventory (HPI-CI) to probe the misconceptions that students hold prior to beginning an introductory microbiology course, and revealed patterns in student thinking to the participating faculty.'}",https://www.asmscience.org/deliver/fulltext/jmbe/18/1/jmbe-18-8.pdf?itemId=/content/journal/jmbe/10.1128/jmbe.v18i1.1281&mimeType=pdf&isFastTrackArticle=
-critical pedagogy for teaching hrm in the context of social change,S. Ruggunan,"This paper considers the imperatives of human resource management (HRM) studies in the context of contemporary South Africa. The authors draw on critical management studies (CMS) and the principles of emancipatory education to inform their argument for a critical and relevant HRM curriculum and associated teaching and learning approaches. The authors propose that the content and processes of HRM education must prepare students for critical participation in the contemporary South African society and workplace. The discussion outlines the rationale for the study, the specific prompts for its initiation, the theoretical framework of CMS, and Freire’s concept of emancipatory education.",,https://ajobe.journals.ac.za/pub/article/download/16/21
-strategy training: an incidental learning model for cai,S. Derry,,,
-the impact of intensive writing workshop professional development on a cross-curricular group of secondary teachers,T. Locke,,,
-thai students’ mental model of chemical bonding,Supawadee Sarawan,"This Research was finding the viewing about concept of chemical bonding is fundamental to subsequent learning of various other topics related to this concept in chemistry. Any conceptions about atomic structures that students have will be shown their further learning. The purpose of this study is to interviews conceptions held by high school chemistry students about metallic bonding and to reveal mental model of atomic structures show according to the educational level. With this aim, the questionnaire prepared making use of the literature and administered for analysis about mental model of chemical bonding. It was determined from the analysis of answers of questionnaire the 10th grade, 11th grade and 12th grade students. Finally, each was shown prompts in the form of focus cards derived from curriculum material that showed ways in which the bonding in specific metallic substances had been depicted. Students’ responses revealed that learners across all three levels prefer simple, realistic mental models f...",,
-steps towards flipping classes in higher education (esp),E. Păcurar,"The paper investigates the gradual transition from traditional methods of language instruction to delivering teaching material in a blended format and reports on some of the findings of a research grant dedicated to measuring the impact of flipped learning on students' development of productive skills. This transition reconsiders the learning/teaching stages and aims at both reducing teachers' lecturing time in class and increasing students' speaking time, so that more emphasis is given to productive oral activities and to fostering fluency. The process of reversing the learning and teaching stages prompts changes in the curriculum planning, in restructuring part of the materials available and reconsidering the role of both language instructor and student. These aspects are submitted under the concept of flipped learning and pinpoint the first steps involved in creating video materials for students in Cultural Tourism. The author reflects on choice of material (units, topics), curriculum changes (task-based and competence-based syllabi), comparing different learning environments and also on the advantages and challenges of implementing flipped methods. The benefits of transitioning from traditional to blended-learning methods stem from several aspects regarding the educational spectrum and its stakeholders. At the institutional level, it serves the language policy of the university and its long-term strategy of developing more blended teaching corpus and of encouraging multicultural and intercultural awareness through foreign language learning and teaching. At the European level, the focus on the proliferation of blended strategies, on Open Educational Resources and on facilitating the development of transversal and cross-sector skills are proof of the changes that need to be adopted by all European universities. At the local level, students' needs and their preparation for a mobile, ever-changing labor market acknowledge the role of language instruction in formal and informal interaction, inside and outside the classroom.","{'model': 'tldr@v2.0.0', 'text': ""The paper investigates the gradual transition from traditional methods of language instruction to delivering teaching material in a blended format and reports on some of the findings of a research grant dedicated to measuring the impact of flipped learning on students' development of productive skills.""}",https://easychair.org/publications/open/FnSp
-the jsep learning skills training system,S. Derry,"The Army Job Skills Education Program (JSEP) is a dual curriculum that reminds soldiers to use learning skills while they take CAI lessons that train basic math and verbal job competencies. Prior to entering the JSEP basic skills curriculum, soldiers are taught self-motivational techniques and strategies for studying and solving problems. Each type of learning skill is represented and trained by a different character model. These characters appear as prompts throughout the basic skills curriculum to remind soldiers to recall and use their new learning skills. In this article, the rationale for the curriculum model is overviewed and preliminary field trial data are reported.",,
-"preparing school systems to deliver a hybrid education program for students with autism via distance learning classrooms, in-class teleconferencing, and listserv technology.",J. C. Burke,"The Center for Technology in Education (CTE), a partnership between Johns Hopkins University and the Maryland state department of education, developed a model for teaching autistic students that integrates technology-based instructional and behavioral supports into existing curriculum. The goal is to promote student gains in communication, social skills, and academics, ab well as to increase students' overall involvement in classroom activities. The model recommends teacher training in providing clear instructions, prompts and prompt fading, positive reinforcement, behavior modification techniques, and advanced teaching principles such as targeting student responsiveness and providing students with opportunities for appropriately communicating their needs. The model also advocates the use of light or high technology based on student needs and demands of the learning task. Light and high technology include equipment such as voice output communication devices, picture systems, sequence and cue cards, touch screens, modified keyboards, switch access devices, and sign language. Teachers also need to target behaviors related to motivation, cooperation, and responsiveness; learning strategies and methods that help students learn more efficiently; and skills that help all students function in the inclusive educational setting. Equally important is the development of an individualized continuity plan that promotes longitudinal gains and transitions across teams of staff members within a school or across schools, and the use of school-wide staff training to promote methods of inclusion. Currently, CTE is exploring alternative methods for disseminating technical assistance and teacher training to local education agencies throughout Maryland using distance learning classrooms, in-class teleconferencing, and listserv technology. (LP) PERMISSION TO REPRODUCE AND DISSEMINATE THIS MATERIAL HAS BEEN GRANTED BY DIANE MONTGONE-RY ( PARENT) TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) U.S. DEPARTMENT OE EDUCATRDN Orlice of Educationat Research and Improvement EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) 1:104ns document has been reProduClid received from the persOn or Prgendation onginating C Mmot changes have been made to improve reproduction Quality Points of vie* ce opinions Slated in this docu rnenl do not necessanty represent official OE PI position or policy John C. Burke, Ph.D. Center for Technology in Education Johns Hopkins University Baltimore, Maryland PREPARING SCHOOL SYSTEMS TO DELIVER A HYBRID EDUCATION PROGRAM FOR STUDENTS WITH AUTISM VIA DISTANCE LEARNING CLASSROOMS, IN-CLASS TELECONFERENCING, AND LISTSERV TECHNOLOGY One of the most difficult challenges teachers and local educational agencies face today is educating students with pervasive developmental disorders such as autism. Teachers frequently describe these children as unique, puzzling, mysterious, and difficult to teach. They often comment that their students with autism rarely interact, communicate, or appear motivated to participate in academic or social activities. A committed teacher provided the following comment before receiving training: ""He came into my room the first day and cried and showed absolutely no interest in anything; I was at a loss."" Given the widespread behavioral deficits characteristic of autism and the limited resources available, teachers repeatedly face an enormous task as they strive to help these children learn. Aside from posing a challenge to teachers, LEAs are recurrently in a quandary as they form policy and develop curricula. Unfortunately, LEAs typically commit their resources to a single approach, then quickly discover that one approach does not easily serve all of their students with autism. The Center for Technology in Education (CTE), a partnership of Johns Hopkins University and the Maryland State Department of Education, is developing a hybrid paradigm in w11:.ch technology-based instructional and behavioral supports are infused into existing curricular activities to help promote gains in communication, social skills, and academics, as well as to increase the children's overall involvement in classroom activities. In addition, CTE is exploring alternative methods tbr disseminating technical assistance to LEAs throughout the state of Maryland using Distance Learning Classrooms, In-Class Teleconferencing, and LISTSERV Technology. Technology-Based Hybrid Educational Model for Childrot With Autism This hybrid paradigm is based on interdisciplinary research and years of educational experience and has four central components: Effective Behavioral Teaching Principles Curriculum-Wide Integration of Technology Flexible and Longitudinal Curriculum Guide Systemic Change Activities 112 BEST OR PAILMILE Effective Behavioral Teaching Principles A set of fundamental teaching principles, based on 30 years of research, now exists and forms the basis for many more advanced teaching strategies. The use of these principles is a central part of the approach. While many professionals are familiar with these principles, it is advantageous to delineate them clearly and to stress their importance in working with students with pervasive developmental disorders such as autism. All educational staff members should be trained in the use of these principles. It is equally vital that significant others, such as parents, learn to use these principles in less structured settings. A brief overview of these fundamental teaching principles is presented below. I. Effective instructions/requests. For all children, but especially for students with autism, instructions should be brief and clear, should be phrased in the form of a request, should be presented when you have the child's attention or should be used to get the child to attend to you, and should be individualized for each child. 2. Prompts and prompt fading. Prompts can help the child to respond and learn new skills. They can be used to help guide the child to respond. In using prompts, it is important to use the least intrusive, yet effective, prompt. Prompts can be verbal, gestural, and physical. When physical prompts are used, it is advantageous not to overuse ""hand-over-hand"" or other full physical ""holding"" prompts. While eftective, these are very difficult to fade. 3. Consequences. Consequences, in the form of positive reinforcement, should involve presenting highly preferred items or other events contingent on the display of the appropriate target behavior. It is vital that the items be things that the child clearly wants and finds reinforcing. In general, consequences should come soon after the child exhibits the target behavior. In addition, delivering positive consequences following a sequence of both disruptive and cooperative behaviors might inadvertently produce an increase in the disruptive behaviors. If a child first shows disruptive behaviors, then displays appropriate behaviors and is subsequently reinforced, he may learn that he needs to exhibit both to get reinforcement. It is usually advisable not to reinforce these chains of behaviors. In recent years, increased emphasis has been placed on integrating natural/direct reinforcers throughout a child's dav. Natural/direct reinforcers are consequences that are directly related to the activity. For instance, if a child uses a picture exchange system to communicate that the child wants a break, giving the child a break would be a natural reinforcer. Placing emphasis on using natural/direct reinforcers, rather than relying only on artificial reinforcers often leads to increased performance and cooperation. 4. Chainiag, Chaining involves breaking down a complex task into steps, establishing criteria for success for each step, and teaching the child to perform each step satisfactorily before adding the next step. S. Shaping. Shaping involves reinforcing closer approximations of the target response. For instance, if an essentially nonverbal child says ""C"" for Cup, the teacher would reinforce the child. If the child subsequently says ""Cu"" he would be reinforced. If he then said ""C"". no reinforcement would be provided. Only better approximations are reinforced. In addition to the fundamental teaching procedures described above, during the course of a training project, training would be extended to cover more advanced principles of effective teaching. Depending on the needs of the educational staff, these advanced procedures could include the following: 1. Use of the principle of partial participation. The principle of partial participation involves the idea of setting the occasion for the student to partially participate in an activity or task and arranging for the student to succeed. Over time, the student will learn more and more and will eventually learn to perform the entire task in a more independent manner. Using this principle throughout a child's day helps to ensure low levels of disruptive behavior. In a consistent manner, it is otlen beneficial to reinforce approximations and attempts. This concept centers on the view that a child who is truly trying to correctly respond should be provided with positive feedback relative to the response. This principle could be used during academic and leisure time periods. 2. Targeting cooperation and responsiveness. Cooperation forms the basis of effective teaching interactions between teachers and students. Cooperation is often increased as students are taught new academic and social skills. It is oflen advantageous, however, to directly target cooperation and responsiveness in a systematic, yet natural, manner throughout a child's day across settings and activities, and with people. Many children are more focused, responsive, and cooperative when relatively greater emphasis is placed on: Presenting the child a relatively short and simple request or instruction. Ensuring that the child is oriented to the adult when the instruction is presented. Using an appropriate prompting st",,
-high quality professional development: an investigation of the supports for and barriers to professional development in arts education.,Vicki R. Lind,"This study focused on a model of professional development designed to support and encourage arts educators to increase their understanding of student learning in the arts, broaden their knowledge of the Visual and Performing Arts Standards, build upon their repertoire of teaching methods and assessment strategies, and improve leadership skills. Data included 300 hours of observation, focus group and individual interviews, written responses to reflective prompts, unit plans, video and audio tapes, and samples of student work collected over a two year period. Findings indicated that working collaboratively, focusing on student learning, and identifying and planning curriculum around issues central to the discipline positively impacted teachers work. The issue of time constraints was consistently identified as a barrier to professional growth.",,
-linking schools of thought to schools of practice,Lucy Hunt,"Project Linking Learning (“Link”) was created to target the needs of gifted students in urban school districts with historically underserved populations. Project Linking Learning implemented a linking curriculum between in-class instruction and an afterschool enrichment program for selected students in second through fifth grade. Designed by Dr. Sandra Kaplan as a collaborative endeavor between the University of Southern California, Rossier School of Education, and the Los Angeles Unified School District (LAUSD), this scale-up grant (Jacob J. Javits Grant #5206A090045) targeted elementary schools in diverse, urban neighborhoods with consistently low rates of gifted referral, identification, and program implementation. Intensive training and support was provided to Link teachers to allow them to effectively roll out this novel curriculum, including foundational training on differentiated instruction (e.g., prompts of depth and complexity, research skills, thinking skills, learning centers, independent study, etc.), as well as information on the characteristics of gifted learners and able underachievers from diverse backgrounds. Schools participating in Project Linking Learning experienced a significant increase in the rate of gifted identification over a four-year period of time. Key program concepts and alignment to educational schools of thought are described, as well as takeaways and recommendations for district adaptation and implementation.",,
-integrating physical and virtual models in biology: a study of students' reasoning while solving a design challenge,N. Martin,"Using models to explain phenomena is important in science. Virtual and physical models have different affordances that can be integrated to foster students’ learning. Integrating evidence from multiple models to justify explanations is challenging, and we know little about how students coordinate such information, especially in biology. This study investigated how students’ integrated information from virtual and physical models in a design-based, biology curriculum. Some students used information from virtual simulations in written explanations of changes they would make to their physical models. However, onethird of students did not use the virtual model to justify their revisions, despite prompts from instructional materials, the teacher, and other group members. Even some students who integrated these different models did not initially do this without support from external prompts. This study provided deeper understanding of how students integrated physical and virtual models, which can help identify the kinds of support students may need.","{'model': 'tldr@v2.0.0', 'text': 'Deeper understanding of how students integrated physical and virtual models in a design-based, biology curriculum is provided, which can help identify the kinds of support students may need.'}",
-pre-service elementary school teachers' metaphors on mathematics textbooks,J. Kim,"The purpose of this study was to investigate the nature of pre-service elementary teachers` metaphors on mathematics textbooks. Their metaphors describe individual and collective patterns of thinking and action on mathematics teaching and learning. To analyze their metaphors, qualitative analysis method based on Lakoff and Johnson`s theory of metaphor (1980) was adopted. Metaphors on mathematics textbooks were elicited from 161 pre-service elementary school teachers through writing prompts. The writing prompt responses revealed three types and thirteen categories: As Type I, there were (1) `Principles`, (2) `Summary`, (3) `Manual`, (4) `Encyclopedia`, (5) `Code`, (6) `Guidelines`, and (7) `Example`. As TypeII, there were (9) `Assistant`, (10) `Friend`, (11) `Scale`, and (12) `Ongoing`. As TypeIII, there was (13) `Trap`. Among these categories, `Guidelines`, `Assistant`, and `Ongoing` were the most frequently revealed. These results indicate that the relations of mathematics curriculum, textbooks, and classrooms are not a unilateral way but should communicate with each other.","{'model': 'tldr@v2.0.0', 'text': 'The results indicate that the relations of mathematics curriculum, textbooks, and classrooms are not a unilateral way but should communicate with each other.'}",
-critical and reflective thinking in an intermediate financial accounting course: an action research study.,J. Cobb,"Accounting professionals have consistently called for educators to develop curriculum designed to encourage students to develop intellectual skills. The purpose of this action research study was to develop and implement an instructional method that requires intermediate financial accounting (IFA) students to consistently practice higher order thinking. Students completed a semester-long authentic comprehensive project (CP) that required them to consistently practice critical and reflective thinking (Facione, 1990; Mezirow, 1991). Findings led to the resolution of implementation issues associated with using a CP. Most implementation issues originated from students’ inability to apply learning obtained from working structured problems to unstructured accounting work. Short reflection papers (RPs) replaced periodic objective tests to encourage deep and meaningful learning. Students’ responses to question prompts gave evidence of one of Mezirow’s (1991) four stages of professional reflection. The depth of reflection trended with students’ understanding of when and how accountants use judgment. Students who consistently practiced higher order thinking also learned to adequately perform routine accounting procedures. This study resulted in an instructional method that requires accounting students to practice using the intellectual skills necessary for success in the accounting profession without sacrificing procedural knowledge. The findings will benefit other instructors working to develop learning materials that require students to practice higher order thinking as they complete authentic professional work.",,
-finding out what they really think: assessing non-science majors’ views of the nature of science,B. Beck-Winchatz,"As institutions of higher learning are increasingly held accountable for student outcomes, faculty are faced with the challenge to clearly articulate and assess what students should learn in their courses. We report on the assessment of a liberal studies learning outcome related to the nature of science, which involved 178 students from 41 academic non-science majors at DePaul University in Chicago. Students were shown a video recording of an interview with a research scientist and asked to respond to prompts about falsifiable predictions, uncertainty, and the functions of hypotheses, observations, and reasoning. We found that students held a wide range of views and misconceptions about the nature of science. Responses were, on average, at a “developing” level (a score of 2 on a scale from 1–4). We discuss several possible changes to curriculum and pedagogy to address these misconceptions based on a review of the science education research literature.",,
-hybrid teaching mode for laboratory-based remote education of computer structure course,Wan Han,"We describe an Open edX-based blended course developed for a reformed Computer Structure course at Beihang University. In three iterations of this laboratory-based course, we dive into key issues that impact students' learning, and then redesign our curriculum, which integrated with virtual laboratory technique into the MOOC platform. We show how certain course design aspects affect students' learning in the hybrid teaching mode: (a) Kung Fu style competency education with online-support laboratory system prompts students to own their learning as the pace and/or the path of learning, which is dictated by mastery instead of the time/space; (b) strengthen the use of learning-aid tools empowered teachers with the skills and information to define standard for each learner in each stage; (c) automated test technology make this blended learning possible at scale and also financially sustainable; (d) using discussion forum to build the lesson about `what to do' when learners get stuck helped in overcoming challenges.","{'model': 'tldr@v2.0.0', 'text': ""This work describes an Open edX-based blended course developed for a reformed Computer Structure course at Beihang University, and shows how certain course design aspects affect students' learning in the hybrid teaching mode.""}",
-student perceptions of gamified audience response system interactions in large group lectures and via lecture capture technology,Robin K Pettit,,,https://bmcmededuc.biomedcentral.com/counter/pdf/10.1186/s12909-015-0373-7
-interaction - based intervention programs in multiple disabilities,Andrea Hathazi,"Competences of practitioners working with children with multiple disabilities refer mainly to assessment, curriculum planning and intervention, but these are specifically related to the learning process and the needs of the child in development. Since the child- centered approach is a key factor in the success of intervention, more and more theoreticians and practitioners take into consideration the interaction approach so that the teachers become more reflective on their own behavior, language and prompts and also what they determine in a child’s behavior, motivation and communication. The present paper focuses on the modalities in which practitioners become aware of the consequences of their own abilities in intervention and they implement the most adequate strategy and method for that specific context, time and child’s needs. The intervention should be dynamic, managed by both partners involved, observing each other’s participation, paying attention to emotional state and well- being.",,
-strategies for planning and designing medical curricula and clinical teaching,M. Barrow,"Student learning is an active and constructive process. The role of a teacher is to provide an environment in which students are able to actively engage with subject matter in order to learn it. This article examines the principal features of good curriculum, course and lesson design and discusses ways in which doctors, in their roles as teachers of medical students and medical trainees can ensure that their teaching prompts learners’ engagement in the learning process.","{'model': 'tldr@v2.0.0', 'text': ""The principal features of good curriculum, course and lesson design are examined and ways in which doctors, in their roles as teachers of medical students and medical trainees can ensure that their teaching prompts learners' engagement in the learning process are discussed.""}",http://seajme.sljol.info/articles/10.4038/seajme.v4i1.432/galley/579/download/
-dance as a way of knowing,Jennifer Donohue Zakkai,"Jennifer Donohue Zakkai illuminates why and how dance is a powerful tool for learning and creativity in K-6 classrooms. In this ""workshop on paper"" you'll discover how to build on what you already know about movement, and implement techniques to tap into children's natural appetite for moving. Zakkai addresses the challenges of engaging students in full-bodied motion in the classroom. She focuses first on helping students become responsible movers in the space, then offers structured learning experiences that demand a high level of concentration and creativity. Through verbal prompts--not demonstrations--that involve students in creative problem-solving, students learn through discovery instead of imitation. You don't have to be a dancer to use the detailed model lessons that guide you through warm-ups, movement explorations, rich curricular integrations, culminating activities, observation and reflection. To help your students understand curriculum content, express themselves in creative ways, and learn about dance as an art form, Dance as a Way of Knowing offers a progression of strategies that will make this vibrant discipline come alive in your classroom.",,
-benchmarks curricular planning and assessment framework: utilizing standards without introducing standardization,Erika Feldman,,"{'model': 'tldr@v2.0.0', 'text': 'An approach to curricular planning and assessment that uses state benchmarks as an umbrella structure to support curricular plans, assessment, and feedback among them is introduced.'}",
-severe weather field experience: an undergraduate field course on career enhancement and severe convective storms,C. Godfrey,"Abstract Undergraduate students acquire a deeper understanding of scientific principles through first-hand experience. To enhance the learning environment for atmospheric science majors, the University of North Carolina at Asheville has developed the severe weather field experience. Participants travel to Tornado Alley in the Great Plains to forecast and observe convective storms for two weeks. The objectives of the course encompass far more than observing severe storms. On days with non-threatening weather in the Great Plains, students participate in an array of activities that provide exposure to facilities and interaction with professionals in various sectors of meteorology. While the allure of chasing storms initially prompts the students to enroll in the course, the focused career-development aspect of the curriculum increases awareness for the varied career options in the atmospheric sciences and helps students discover where their own capabilities and interests might best suit the discipline. The course thus offers students a comprehensive career-development experience woven within a thrilling adventure.",,
-teachers' planning processes: seeking insights from australian teachers | nova. the university of newcastle's digital repository,P. Sullivan,"The Australian Curriculum: Mathematics (ACM) prompts consideration of planning processes teachers typically use (as a whole school, in grade levels, and at the classroom level). In order to gain insights into the nature of the planning decisions teachers make, Australian teachers drawn from every state and territory were invited to respond to a survey on planning processes, during professional learning programs or voluntarily online through the Australian Association of Mathematics Teachers' website. In this article, we provide background to the overall project in which this survey was situated, and describe the process through which the survey was developed. Subsequent articles in this journal detail some important findings from the survey data.",,
-a dual purpose data base for research and diagnostic assessment of student writing,J. Parr,"The data base of writing examined serves a dual purpose. Here it is used as a research tool and the writing performance from the large, nationally representative sample (N = 20,947) of students (years 4 to 12) interrogated to examine patterns of performance in writing. However, the data base was designed to underpin a software tool for diagnostic assessment of writing. Viewing writing as accomplishing social communicative goals, performance was considered in terms of seven main purposes the writer may seek to achieve. Tasks related to each purpose were encapsulated in 60 writing prompts that included stimulus material. Participants produced one writing sample; the design ensured appropriate representation across writing purposes. Samples were scored using criteria differentiated according to purpose and curriculum level of schooling and acceptable reliability obtained. Analyses indicate that growth was most marked between years 8 and 10, arguably, as opportunity to write increases and writing is linked to learning in content areas. Variability in performance is relatively low at primary school and high at secondary school. Students at any level did not write equally well for different purposes. Mean scores across purposes at primary school were relatively similar with to instruct and to explain highest. By years 11-12 there is a considerable gap between the highest scores (for narrate and report) and the lowest, recount, reflecting likely opportunities to practice writing for different purposes. Although girls performed better than boys, the difference in mean scores narrows by years 11-12.","{'model': 'tldr@v2.0.0', 'text': 'The writing performance from the large, nationally representative sample of students is interrogated to examine patterns of performance in writing and indicates that growth was most marked between years 8 and 10, arguably, as opportunity to write increases and writing is linked to learning in content areas.'}",https://www.jowr.org/index.php/jowr/article/download/738/781
-impact of faculty development on physical therapy professors' beliefs.,L. Behar-Horenstein,"Physical therapy faculty share similarities with faculty across allied health fields, such as nursing, and other clinical disciplines that educate students in licensing and board certification programs. Most have clinical experience and discipline-based expertise, however they may not have had the benefit of continuous learning aimed at enhancing their teaching repertoires. Because of the rapid influx of clinicians into the academy, faculty development is considered essential to their integration. The purpose of this study was to describe how faculty development impacted physical therapy professors' understanding and use of new instructional skills. Eight physical therapists from a university located in the Southeast U.S. participated in a six-week, 12-hour teaching seminar focusing on curriculum and teaching where participants kept reflective journals to record their responses to question prompts. Basic unitizing, coding, and categorizing were used to conduct a multi-stage analytical process. Eight themes emerged including assessment, instructional strategies, teaching styles, and individuals' planned changes to their classroom practice, among others. Findings showed that professional development is essential for enriching faculty instructional capacity to promote student learning and patient care. Investing in the professional development of faculty may help ensure quality teaching so that professors become conduits to knowledge production. Faculty development is ""the continuous learning that professionals may need to pursue throughout their careers in order to maintain, enhance, and broaden their professional competence"" (Gottlieb, Rogers & Rainey, 2002, p. 280). Studies in physical therapy, and other health professions, have shown that the faculty development process is central to effective teaching and the preparation of future healthcare practitioners and professional educators (Behar-Horenstein, Childs, & Graff, 2008, 2009, 2010; Farmer, 2004; Mahara & Jones, 2005, Steinert et al., 2006). Keeping physical therapists continuously informed about new knowledge, skills and technology is essential to their capacity as instructors. Historically physical therapy (FT) programs have been disadvantaged. They have been forced to hire an overwhelming number of clinicians as faculty rather than individuals who have been trained in higher educational instruction and assessment (Harrison, «Sc Kelly, 1996). Most of the clinical instructors, while quite skillful, often lack teaching abilities (Gottlieb, Rogers «Sc Rainey, 2002). Hiring clinicians resulted because universities were in the initial stages of developing physical therapy degree programs. Such an action highlights the importance of developing a mechanism to continuously determine faculty effectiveness and productivity. Faculty development can help new faculty examine their own beliefs about teaching and consider how they might apply their thinking. New faculty from clinical disciplines are often different from traditional academicians who earn several degrees and typically experience graduate student teaching and research roles aimed at faculty-type expertise. The American Physical Therapy Association's (APTA) former director of professional development, Marilyn Phillips, encourages the use of faculty development. This process can be a vehicle to create one's own plan or ""blueprint for career development"" (Starcke, 2005, p. 42). APTA' s Board of Directors acknowledges varied modes of faculty development including where and how it can occur. "" [It] may occur in formal instructional settings or in natural societal settings and may include . . . academic courses of study, organized continuing education, independent study, and self- and external assessment"" (p. 42). Although some individuals benefit from structured activities, others may work on their own plan of professional growth. However, the APTA directors emphasize the role of assessment in professional development stating, ""All professional development experiences should be planned and assessed"" (p. …",,
-teachers have the power to alleviate math anxiety,D. J. Shields,"Abstract Many students suffer from math anxiety, but teachers can incorporate skills to alleviate it for their students. In order to alleviate math anxiety teachers first and foremost need to portray a positive, enthusiastic, helpful attitude, which communicates a love and usefulness for mathematics. Math curriculum should be designed that deepens student understanding in a practical, engaging and fun way. Pedagogy should be implemented that moves from the explain-practice-memorize strategy to strategies that stress reasoning and understanding. A classroom culture that prompts student inquisition, discovery, learning and the exploration of ideas needs to replace the structured, rigid atmosphere usually associated with math class. And finally, assessment should be conducted in a variety of ways. Introduction Not only do students today need math remediation at an overwhelming rate, but reports state that two thirds of American adults fear and loathe math (Burns, 1998). One reason such negative experiences are reported is math anxiety, which knows no boundaries regarding race, age or gender. Math anxiety can be perpetuated in the home, in society, and in the classroom. Because most students can name the teacher they believed caused their anxiety, the exact moment when it happened and the event that triggered the anxiety in the classroom, an examination of teacher attitudes, curriculum, pedagogy, the classroom culture, and assessment is crucial to understanding and alleviating the problem. Students at all grade levels can experience math anxiety and they attest to similar characteristics and consequences. One of the most notable consequences of math anxiety is poor math achievement and poor math grades. Part of this is because students with math anxiety attempt to cope with the debilitating effects of their anxiety by avoiding math in school. This in turn can lead to limited college majors, and ultimately career choices that are restricted mathematically and monetarily. This limited exposure to mathematics is also responsible for lower math achievement and competence (Ashcraft, 2002). The anxiety, which is not significantly related to intelligence, has been shown to inhibit student learning (Ashcraft, 2002) and to reduce working memory capacity (Ashcraft & Kirk, 2001; Beilock,& Cart, 2005; Shobe, Brewin & Carmack, 2005), which in turn has a major impact on students"" serf-confidence related to mathematics. Math anxiety is often the result of repeated negative experiences related to mathematics (Kogelman, Nigro, & Warren, 1978). It is a conditioned fear that develops into a fatalistic attitude, which becomes a self-fulfilling prophecy that reinforces one's beliefs about an inability to perform mathematically. For those who suffer from math anxiety physiological symptoms such as sweaty palms, nausea, muscle contractions, difficulty breathing, tightness in the throat and chest, headaches, heart palpitations, restless behavior, forgetfulness, and a temporary boost in one's heart rate are familiar. Teachers' attitudes, curriculum, pedagogy, the classroom culture, and assessment greatly impact students. All these areas must be considered in order to alleviate anxiety in the classroom. In a 1999 study by Jackson and Leffingwell, it was discovered that sixteen percent of the students studied were first traumatized mathematically in third or fourth grade. The difficulties in fourth grade arose from new math ideas such as fractions, timed tests, and memorization of multiplication tables and formulas. Teachers at this level were blamed for hostile behavior, making derogatory comments when children did not understand concepts, appearing angry when asked for additional help, displaying insensitive and uncaring attitudes, stereotyping females as not needing math, ridiculing girls more overtly than boys, and favoring boys in the subject. Approximately twenty-six percent of the students in the study first realized feelings of math anxiety in ninth, tenth and eleventh grade. …",,
-music assessment and the hegemony of musical heritage,G. Spruce,"About the book: Issues in Music Teaching stimulates critical reflection on a range of topics related to the teaching and learning of music in both the primary and secondary school, including: • the place of music in the curriculum • the nature of music and music education • ICT and music education • music education and individual needs • continuity and progression in music education The book prompts the reader to be analytical and critical of theory and practice, and to become an autonomous professional and curriculum developer.",,
-introducing in-service teachers to virtual schooling through the lens of the three teacher roles,M. Barbour,"This study will examine the third and fourth rounds of data collection from an action research project designed to help in-service teachers become better virtual school facilitators (currently being analyzed). The data included blog entries and comments from five of the seven graduate students in an instructional technology course related to K-12 online learning. The specific discussion prompts relate to virtual school readings and the Teacher Education Goes into Virtual Schooling (TEGIVS) curriculum. Based upon initial analysis, the TEGIVS curriculum was effective for providing these graduate students some experience with how K-12 online learning opportunities were delivered, along with some of the possibilities and challenges associated with K-12 online learning. The analysis of this data is continuing, and there are plans to continue this line of inquiry with additional students in future offerings as we continue to improve upon the course design. Clark (2001) estimated there were approximately 40,000 to 50,000 students engaged in virtual schooling in the United States. Eight years later Picciano and Seaman (2009) indicated that there were over 1,000,000 K-12 students enrolled in online courses. Watson, Gemin, and Ryan (2008) reported significant K-12 online learning activity in 44 states. Several states, including Michigan, require students to have an online learning experience in order to graduate from high school. This growth of online learning at the K-12 level has caused dramatic changes in the design and delivery of education, at least for the student population served by these opportunities. However, the vast majority of pre-service and in-service teacher education programs simply do not provide any curricular opportunities in this area. Rice and Dawley (2007) found that less than 40% of all online K-12 teachers in the United States reported receiving professional development before they began teaching online. This indicates a need for teacher education programs to address pre-service and in-service teachers’ ability to teach in environments that are completely mediated by technology. This study examined one university’s continuing efforts to address this growing need in teacher education. Wayne State University is a large, publicly funded, research-extensive university in Michigan (i.e., the first jurisdiction to have an online learning graduation requirement). Over the past two years, the Instructional Technology program has developed and adopted curriculum to support teachers with understanding of K-12 online learning in their own schools. This proceedings examines the results of the second round of data collection.",,
-integrating narrative reading comprehension and writing instruction for all learners.,Deborah C. Simmons,"A study examined the effects of an integrated reading/writing curriculum on the narrative writing of students in general education eighth-grade classrooms. Subjects, 93 Ftudents (including 10 with learning disabilities) from middle to w.:7er-middle socioeconomic backgrounds at a suburban middle school in wertern Oregon, were administered narrative writing probes preand post-intervention tc assess students' ability to plan, organize, and write stories for topic prompts. Students were enrolled in four sections taught by tuo teachers; students from one teacher's classes served as the experimental group and the other teacher's classes participated in the control condition. Instructional materials included 10 short stories. The integrated reading and writing curriculum was designed in three interdependent phases: learning narrative text structure; learning a writing process; and learning to generate stories. Students in the control condition received instruction and practice on narrative text comprehension. Results indicated that students in the experimental group significantly outperformed students in the control group--their stories contained more fully developed ideas, content, settings, characters, and attempts to solve the central problem than students in the control classrooms. Results also indicated that all students in the integrated condition benefitted from the curriculum, although students continued to have difficulty generating well-developed stories. Findings suggest the potential value of investing in curriculum development that equips learners with transferrable and maintainable knowledge. (Two tables of data are included; 22 references are attached.) (RS) Integrating Narrative Reading Comprehension and Writing Instruction for All Learners Deborah C. Simmons Edward J. Kameenui Shirley Dickson David Chard Barbara Gunn","{'model': 'tldr@v2.0.0', 'text': 'The effects of an integrated reading/writing curriculum on the narrative writing of students in general education eighth-grade classrooms and the potential value of investing in curriculum development that equips learners with transferrable and maintainable knowledge are suggested.'}",
-archiving student solutions with tablet pcs in a discussion‐based introductory physics class,E. Price,"Many active learning based physics courses use whiteboards as a space for groups to respond to prompts based on short lab activities, problem solving, or inquiry‐oriented activities. Whiteboards are volatile; once erased, the material is lost. Tablet PCs and software such as Ubiquitous Presenter can be used as digital whiteboards in active learning classes. This enables automatic capture and archiving of student work for online review by students, instructors, and researchers. We studied the use of digital whiteboards in an active‐learning introductory physics course at California State University, San Marcos. In this paper we examine the archival features of digital whiteboards’, and characterize the use of these features by students and instructors, and explore possible uses for researchers and curriculum developers.","{'model': 'tldr@v2.0.0', 'text': 'The archival features of digital whiteboards are examined, and the use of these features by students and instructors are characterized, and possible uses for researchers and curriculum developers are explored.'}",http://www.compadre.org/per/perc/2008/proceedings/UP205perc08_revised.pdf
-developing an it project management course to meet changing industry needs,T. Jewels,"ABSTRACT Acknowledging that an awareness of project management skills was an important outcome for their graduates, the IT faculty of one Australian university developed and introduced a new IT Project Management course into their undergraduate curriculum in 2002. A three stage approach that involved identifying students expected learning outcomes, selecting relevant content and choosing the most applicable pedagogy was applied in designing this course. An integrated case study approach, developed over a five year period which was to be the recipient of many teaching awards is described. However, despite its popularity and successes the course was totally rewritten for the 2008 academic year, reflecting alternative views about what IT Project Management courses should contain, which in turn prompts the question of what project management skills should we be teaching students in preparing them for professional IT careers? Keywords Case method, course design, project management, teaching.","{'model': 'tldr@v2.0.0', 'text': 'The course was totally rewritten for the 2008 academic year, reflecting alternative views about what IT Project Management courses should contain, which prompts the question of what project management skills should be teaching students in preparing them for professional IT careers.'}",
-course approval processes,M. Horsburgh,"Focuses on processes of course approval, which lead to enhancement of the curriculum and student learning. Findings from a recent study concerned with the question: To what extent does quality monitoring impact on the student experience of learning? showed the curriculum, and how it was approved, to be significant factors in enhancing student learning. It is argued that the focus for quality in a rapidly changing world should be on the attributes of graduates, with quality monitoring concerned with improvement and enhancement of student learning. Where quality monitoring does have a positive impact on student learning, through the approval and ongoing monitoring of the curriculum and its outcomes, then these processes should be improved. This should be through a process which fosters collegiality and encourages pedagogical discussion amongst academic colleagues. A series of prompts or questions, which seek to challenge teachers and enhance dialogue with colleagues, is proposed.",,
-on curriculum learning for commonsense reasoning,A. Maharana,"Commonsense reasoning tasks follow a standard paradigm of finetuning pretrained language models on the target task data, where samples are introduced to the model in a random order during training. However, recent research suggests that data order can have a significant impact on the performance of finetuned models for natural language understanding. Hence, we examine the effect of a human-like easy-to-difficult curriculum during finetuning of language models for commonsense reasoning tasks. We use paced curriculum learning to rank data and sample training mini-batches with increasing levels of difficulty from the ranked dataset during finetuning. Further, we investigate the effect of an adaptive curriculum, i.e., the data ranking is dynamically updated during training based on the current state of the learner model. We use a teacher model to measure difficulty of each sample and experiment with three measures based on question answering probability, variability and out-of-distribution. To understand the effectiveness of curriculum learning in various scenarios, we apply it on full model fine-tuning as well as parameter-efficient prompt-tuning settings. Our results show that fixed as well as adaptive curriculum learning significantly improve performance for five commonsense reasoning tasks, i.e., SocialIQA, CosmosQA, CODAH, HellaSwag, WinoGrande in both tuning settings. Further, we find that prioritizing the difficult samples in the tail end of training improves generalization to unseen in-domain data as well as out-of-domain data. Our work provides evidence and encourages research into curriculum learning for commonsense reasoning.","{'model': 'tldr@v2.0.0', 'text': 'This work uses paced curriculum learning to rank data and sample training mini-batches with increasing levels of difficulty from the ranked dataset during finetuning, and finds that prioritizing the difficult samples in the tail end of training improves generalization to unseen in- domain data as well as out-of-domain data.'}",https://aclanthology.org/2022.naacl-main.72.pdf
-nc2t: novel curriculum learning approaches for cross-prompt trait scoring,Yejin Lee,"Automated essay scoring (AES) is a crucial research area with potential applications in education and beyond. However, recent studies have primarily focused on AES models that evaluate essays within a specific domain or using a holistic score, leaving a gap in research and resources for more generalized models capable of assessing essays with detailed items from multiple perspectives. As evaluating and scoring essays based on complex traits is costly and time-consuming, datasets for such AES evaluations are limited. To address these issues, we developed a cross-prompt trait scoring AES model and proposed a suitable curriculum learning (CL) design. By devising difficulty scores and introducing the key curriculum method, we demonstrated its effectiveness compared to existing CL strategies in natural language understanding tasks.",,
-combating the covid-19 infodemic using prompt-based curriculum learning,Zifan Peng,,"{'model': 'tldr@v2.0.0', 'text': 'A prompt-based curriculum learning method that can verify content from multiple perspectives by answering a series of questions concerning the text’s reliability and is robust to the hyperparameter settings, making it more applicable with limited infrastructure resources.'}",
-using an experiential learning model to teach clinical reasoning theory and cognitive bias: an evaluation of a first-year medical student curriculum,Justin J. Choi,"ABSTRACT Background Most medical students entering clerkships have limited understanding of clinical reasoning concepts. The value of teaching theories of clinical reasoning and cognitive biases to first-year medical students is unknown. This study aimed to evaluate the value of explicitly teaching clinical reasoning theory and cognitive bias to first-year medical students. Methods Using Kolb’s experiential learning model, we introduced dual process theory, script theory, and cognitive biases in teaching clinical reasoning to first-year medical students at an academic medical center in New York City between January and June 2020. Due to the COVID-19 pandemic, instruction was transitioned to a distance learning format in March 2020. The curriculum included a series of written clinical reasoning examinations with facilitated small group discussions. Written self-assessments prompted each student to reflect on the experience, draw conclusions about their clinical reasoning, and plan for future encounters involving clinical reasoning. We evaluated the value of the curriculum using mixed-methods to analyze faculty assessments, student self-assessment questionnaires, and an end-of-curriculum anonymous questionnaire eliciting student feedback. Results Among 318 total examinations of 106 students, 254 (80%) had a complete problem representation, while 199 (63%) of problem representations were considered concise. The most common cognitive biases described by students in their clinical reasoning were anchoring bias, availability bias, and premature closure. Four major themes emerged as valuable outcomes of the CREs as identified by students: (1) synthesis of medical knowledge; (2) enhanced ability to generate differential diagnoses; (3) development of self-efficacy related to clinical reasoning; (4) raised awareness of personal cognitive biases. Conclusions We found that explicitly teaching clinical reasoning theory and cognitive biases using an experiential learning model provides first-year medical students with valuable opportunities for developing knowledge, skills, and self-efficacy related to clinical reasoning.","{'model': 'tldr@v2.0.0', 'text': 'It is found that explicitly teaching clinical reasoning theory and cognitive biases using an experiential learning model provides first-year medical students with valuable opportunities for developing knowledge, skills, and self-efficacy related to clinical reasoning.'}",https://www.tandfonline.com/doi/pdf/10.1080/10872981.2022.2153782?needAccess=true&role=button
-implementation of the emergency curriculum on thematic learning in private elementary school,Hani Hanifah," Abstract: The Covid-19 pandemic has changed the order of the world of education, this has prompted the government to issue policy Number 719/P/2020 regarding guidelines for implementing the curriculum in educational units under special conditions. The research aims to determine the strategies, learning principles, and policy analysis of the emergency curriculum. The data collection techniques in this study were distributing questionnaires to 26 private elementary school teachers in Tasikmalaya City as participants. The research method used is the quantitative method and the data analysis technique used is descriptive analysis. The results of the study stated that the existence of an emergency curriculum policy in the conditions of the Covid-19 pandemic was effective in thematic learning in private elementary schools by using learning strategies that emphasize the character of students and the principles of active and fun learning.",,http://journal.ummat.ac.id/index.php/elementary/article/download/6504/pdf
-the impact of time on learning outcomes: lessons from the delivery of high school mathematics curriculum in ghana,Farouq Sessah Mensah,"A student’s productive learning time is proportional to the length of time they spend working during their lessons. However, it is important to remember that the issues with educational time are not just about optimizing “in-school time.” The study is prompted by the time spent in school and how it impacts students learning outcomes in core mathematics, considering two different year groups for the May/June 2013 West African Senior School Certificate Examination (WASSCE-standardized examination written by Anglophonic countries of West Africa (Ghana, Nigeria, Sierra Leone, Liberia, and the Gambia)). The study compares students’ learning outcomes in the May/June 2013 WASSCE. The study uses a descriptive survey design and conveniently collected data from 10 Senior High Schools (SHSs) in the Central Region of Ghana. The results of the study showed no statistically significant difference in core mathematics grades between students who spent four years and those who spent three years in SHS. Thus, time is not the sole determinant of a student’s learning outcome, as has been suggested by some studies.",,https://www.ej-edu.org/index.php/ejedu/article/download/483/290
-factors of learning towards creating blended learning curriculum using learning management system in higher education during covid-19,Fazilah Razali,"Blended learning is a mix of online and face-to-face instruction. It refers to a variety of possibilities made available by integrating the internet and streaming media with traditional educational formats that necessitate the physical co-presence of teachers and students. While the Covid-19 pandemic is being referred to as the greatest challenge facing education systems around the world, it has prompted authorities to issue a drastic order requiring institutions to switch from face-to-face teaching to online teaching and virtual education for students. This present study aims to determine the model of integration between learning style and active learning towards enhancing blended learning curriculums among higher education students. A total of 208 students at a selected Malaysian higher education institution were selected randomly in this study. For data collection, this study used the quantitative approach and random sampling technique. The instrument was developed based on the literature review, and expert validation was obtained from various universities. The survey forms were distributed using an online medium (Google forms). Pearson Coefficient correlation (r) was employed to examine the relationship between variables, while Structural Equation Modelling (SEM) was utilized to examine the effect of the mediator variable. By dopting a two-stage method, the measurement model was first tested, followed by the structural model. Confirmatory Factor Analysis (CFA) is to test whether the data is compatible with the hypothesized measurement model as well as for the purpose of validation and construct reliability. The results show that the measurement model was a good fit for the data, and that the constructs were reliable and valid for testing the hypotheses based on the results. The assessment of the structural model, on the other hand, involved testing the hypothesized relationships about the direct effects. Four hypothesized direct paths were found to be statistically significant. © 2022 Eskisehir Osmangazi University. All rights reserved.",,
-"longevity, legacy, and lament: learning from longstanding educators in an innovative curriculum",G. Maudsley," Research about problem-based learning (PBL) tutoring in medicine has prioritized quantifying relationships between tutor characteristics or learning environment and tutoring behaviour or student outcomes. Longitudinal studies and qualitative research about how such tutors conceptualize their long-term experience are rare. The research question was thus: What educator outlooks do inaugural PBL tutors develop after substantial experience in a problem-based medical curriculum? At16 year-follow-up of interviews with an inaugural cohort of PBL tutors, semistructured interviews with the remaining ten explored their outlooks as educators now versus then. Two years later, an open-ended e-questionnaire (E-interview) reviewed their outlooks, particularly about the curriculum being replaced. Tutors viewed their role now through a more discerning, reflective, and constructivist ‘good educator’ lens. They articulated principles for facilitating active learning. When that curriculum was replaced, tutors were positive about its legacy but also lamented flawed educational governance for maintaining and renewing whole-system integrity. Educator development should prompt critical reflection about ‘the good educator’ identity, the related enthusiasms, discomforts, and uncertainties, and the impact of curriculum shifts.",,
-aligning classroom management strategies with a social emotional learning curriculum in early childhood,Jessica Kemp,"Traditionally, school-based preventative frameworks have been implemented in isolation with little consideration of alignment and integration of practices throughout the school day. The present study aims to address this gap by increasing school psychologists’ preventative involvement with consultation in early childhood school settings. Using an integrated approach through a multiple baseline design, four Head Start educators were trained in classroom management practices, to increase opportunities throughout the day for teaching, prompting, and reinforcing key skills taught through the Second Step Early Learning (SSEL) social emotional learning curriculum. Findings suggest that a brief professional development session (1-hour) followed by weekly performance feedback (15 minutes) largely increased educator use of aligning classroom management strategies with weekly SSEL lessons (ES = .94, p-value = <.000). Effects on challenging behavior were limited, although two classrooms demonstrated an overall decrease in challenging behavior (ES = -.20). Results further indicate this intervention increased feelings of teacher self-efficacy and was a socially valid approach; educators reported the aligned strategies were acceptable, sustainable, and beneficial to children. Limitations and implications of this study are further discussed with suggested directions for future research.",,https://digitalcommons.pace.edu/cgi/viewcontent.cgi?article=1054&context=perspectives
-taxonprompt: taxonomy-aware curriculum prompt learning for few-shot event classification,Chengyu Song,,,
-curriculum implementation status and adoption of ict resources in distance learning programme of the nigerian teachers’ institute,I. A. Alade,"As novel as the notion of establishing the National Teachers’ Institute (NTI) Distant Learning Studies is a general observation shows that various people still pick flaws in the programme, especially, its Nigerian Certificate in Education (NCE) at the implementation level. Some quarters doubt the quality of teaching manpower produced as well as how realistic the synergy of combining NTI teacher education with new technology in the training centres. These observations prompted this study which examined “curriculum implementation status and adoption of ICT resources in distance learning programme of the Nigerian Teachers’ Institute. The descriptive survey was guided by three research questions. Curriculum Implementation Status and ICT Resources Rating Scale with a reliability co-efficient of 0.84 was used to collect data from 373 respondents (73 facilitators and 300 students). Frequency count, percentage weights and pie-chart were used for data analysis. The findings revealed that: the programme facilitators are 82.3% and 70.0% qualified qualification wise and in working experience respectively; The adoption of many ICT resources are inadequate except the use of mobile phones with 70.0% adoption; and a below average (43.3%) rating in supervision and monitoring of the teacher education curriculum implementation, but a fairly regular supervision and monitoring of classroom instructional practices for quality assurance (53.4%). It is recommended that basic ICT courses should be part of NCE curriculum for its gradual integration into the teacher preparation process. Also, the government at all times must be cautions not to leave the curriculum implementation affairs in the hands of non-professionals for maximum yield. Keywords: Curriculum Implementation, ICT Resources, Teacher Education.",,
-learning inclusion through makerspace: a curriculum approach in italy to share powerful ideas in a meaningful context,Taziana Giusti,"Learning by making is being recognized as an efficient technique for students to develop knowledge and skills simultaneously. However, one of the most urgent challenges that schools are facing nowadays is to reach every student in their individual profile and potential. Therefore, the purpose of this paper is to offer an integrated approach for re-thinking the role of Makerspace in a context of inclusion: the characteristics of this learning strategy, in contrast to the traditional currents, can offer promising paths to successfully build shared ideas.,This paper critically analyses a Makerspace workshop implemented during a whole academic year in an Italian state primary school. This exploratory and instrumental qualitative approach (Baxter and Jack, 2008) included two channels that have been developed simultaneously, namely, technological skills and social competences.,Stemming from a long tradition of inclusion of children with various educational needs in the mainstream school system, the authors aim to share a success story of academic and social achievements: all the participants were able to develop at their own pace, sharing tools, reaching a balance between the demands of the task and their planning and negotiation skills.,Small group size and the reiterated daily interactions with differences embodied in students with special needs, immigration background, low SES, gifted.,Through ad-hoc training, relatively marginalized pedagogical components (such as the ability to work in low-control situations, flexibility, student-centered learning environments) should be given a more prominent role and can be introduced in the desirable professional development. In addition, national and school policies are prompted to consider its inclusion as a slow-process that cannot be fully achieved in the presence of time and space constraints.,The curricular approach discussed above has shown the importance of inclusion of all students within mainstream schools. Pupils with atypical development can interact with other children, and in this way, they can have first-hand experience of how social dynamics unfold in a real environment. Moreover, they can act in a challenging context where, more often than not, they are pushed to achieve goals exceeding their supposed cognitive abilities. The other pupils are also gaining from these interactions: they can understand different points of view, thus developing empathy, and they can appreciate original ways to approach a task, with cognitive and emotional benefits. In addition, the constant relationship helps them to control their reactions to behavioral problems that sometimes classmates with special needs display and so they deepen their knowledge of and tolerance for others’ peculiarities.,Based on the foundational principles of Papert’s powerful ideas and meaningful context, this paper describes the design principles of a successful makerspace, its integration in the school curriculum, and the achieved inclusion of children with Special Education Needs in a group of peers where adults became observers. Recommendations are discussed on how school practitioners can promote young children’s learning through making.",,
-facilitating teaching and learning with made to measure fashion design and creation mooc courses,Tanya Dove,"Massive Online Open Courses (MOOC) provides a flexible and engaging learning environment. A MOOC in Apparel Technology facilitates a comprehensive e-Learning platform, with a variety of instructional videos and technical resources, which will foster a technical foundation in pattern cutting and sewing for learners. Apparel technology is a core skill for fashion design students; however, universities only address a limited range of construction techniques within their curriculum. Students work towards model size apparel for their graduate fashion shows, which limit the learner’s knowledge in developing apparel for different body shapes. A MOOC in Apparel Technology would facilitate learners with the skills and knowledge to develop apparel for different sizes and figure shapes, with fit assessment and problem-based forums on fit solutions. With many retailers now providing a global platform for the sales of clothing, the sizing of women’s clothing is an area of concern in the fashion industry, ‘one size does not fit all’. The inconsistencies in garment sizing can contribute to the dissatisfaction of garment fit, leaving women unable to find, or know, what size fits their body shape. MOOC incorporates an innovative design and e-Learning platform, in a less structured approach, which harnesses learning in an interactive setting, with online interactions prompting learners to collaborate, share ideas and examples, in a less structured environment.","{'model': 'tldr@v2.0.0', 'text': 'A MOOC in Apparel Technology would facilitate learners with the skills and knowledge to develop apparel for different sizes and figure shapes, with fit assessment and problem-based forums on fit solutions.'}",
-english curriculum reform in the philippines: issues and challenges from a 21st century learning perspective,J. Barrot,"ABSTRACT Recent developments in the region and in the Philippines have prompted the Philippine government to push for a new basic education curriculum. Along with these changes is the adoption of the new English curriculum known as the Language Arts and Multiliteracies Curriculum (LAMC). This article, therefore, sought to examine the K-12 English Curriculum in the Philippines from a 21st century learning perspective. The first section of this article briefly describes the LAMC and 21st century learning in both general and English language teaching-related terms. The discussion then reviews the specificity and coherence of the LAMC, how it is consistent with the principles espoused by 21st century learning, and how it is aligned to established language teaching and learning principles. The findings revealed that the current curriculum needs to improve its specificity, internal coherence, and integration of some essential principles of 21st century learning and language teaching and learning. The article concludes with the possible challenges in implementing the curriculum, suggestions for future design and implementation, and implications for future studies.",,
-using perusall to motivate students’ curriculum-based academic reading,Mimi Li,,,
-"initiating curriculum mapping on the web-based, interactive learning opportunities, objectives and outcome platform (looop)",I. Treadwell,"Background. A web-based curriculum is made transparent by providing multidimensional overviews of content (learning objectives) aligned with learning outcomes and frameworks, opportunities and assessment formats. A South African university embarked on the mapping of its curricula on the web-based learning opportunities, objectives and outcome platform (LOOOP). Objectives. To reflect on the customisation of LOOOP and training of lecturers, and to determine lecturers’ perceptions of the usability and value of LOOOP. Methods. The project manager reflected on the initiating processes, and a survey determined the lecturers’ perceptions of the usability and value of curriculum mapping, using a 4-point Likert scale questionnaire. The convenience sample comprised the first 30 lecturers who had uploaded their curriculum content and consented to partake in this ethics-approved study. Descriptive statistics portray the percentages of agreement on the positive statements of the questionnaire. Results. Challenges related to slow staff buy-in and development were experienced. Required modifications to LOOOP were promptly dealt with. The majority of participants agreed on the usability (≥89%), as well as structure and transparency (≥87%) of LOOOP. Mapping is expected to enhance curriculum revision (≥95%) and communication (≥96%), viewing the curriculum scope, complexity and cohesion (97%), as well as abstracting data for management analysis and reporting (100%). Conclusions. The lecturers agreed on the usability and values of curriculum mapping, which indicates that online mapping is sufficiently beneficial to justify the time and resources invested. Mapping should be a product of collaborative participation and planned as a long-term commitment, which can also be used to research the impact of mapping on student learning.","{'model': 'tldr@v2.0.0', 'text': 'The lecturers agreed on the usability and values of curriculum mapping, which indicates that online mapping is sufficiently beneficial to justify the time and resources invested.'}",http://www.ajhpe.org.za/index.php/ajhpe/article/download/1017/616226
-the opportunity and challenges of implementing a prototype curriculum,P. Solikhah,"Learning loss due to the Covid-19 pandemic prompted the simplification of the 2013 Curriculum into an Emergency Curriculum, which was then refined into a Prototype Curriculum. Curriculum changes always bring optimism as well as skepticism. The hope for a new curriculum that is better in line with the ability of education providers in implementation. This research is evaluative research with a literature study approach. This study aims to make a corrective contribution to the policy coherence of the prototype curriculum. The data source comes from government policies supported by expert analysis of curriculum developers. Data analysis is rationally deductive. It was concluded that the prototype curriculum focused on developing non-technical abilities and character within the framework of Pancasila values, simplifying learning materials, and developing literacy and numeracy skills. Opportunities from this curriculum are reducing the administrative burden of teaching, and the design of learning becomes more flexible so that the quality of learning can be improved. The challenge is the teacher's readiness (attitudes and emotions, knowledge, and behavior). Based on the Ministry of Education's recommendation, which refers to the results of the 2020 UKG, it is stated that the quality and competence of teachers are still low, it is necessary to improve IT capabilities, and the KKG and MGMP functions are not yet maximized.",,https://e-journal.ikhac.ac.id/index.php/NAZHRUNA/article/download/2130/907
-the companion curriculum: medical students’ perceptions of the integration of humanities within medical education,Charlotte Axelrod,"Background The contributions of arts and humanities to medical education are known in the medical education community, but medical schools’ offerings vary. The Companion Curriculum (CC) is a student-curated set of optional humanities content for medical students at the University of Toronto. This study evaluates integration of the CC to identify key enabling conditions for medical humanities engagement. Methods A mixed-methods evaluation gauged usage and perceptions of integration of the CC among medical students using an online survey and focus groups. Narrative data underwent thematic analysis, supported by summary statistics of quantitative data. Results Half of survey respondents were aware of the CC (n = 67/130; 52%), and, once prompted with a description, 14% had discussed it in their tutorial groups. Of students using the CC, 80% reported learning something new regarding their roles as communicators and health advocates. Themes were the perceived value of the humanities, internal student barriers, institutional neglect of the humanities, and student critiques and recommendations. Conclusion Despite participants’ interest in medical humanities, our CC remains underused. To improve humanities’ visibility in the MD curriculum, our results indicate that greater institutional support, including faculty development and early curricular integration, is required. Further study should explore reasons for gaps between interest and participation.","{'model': 'tldr@v2.0.0', 'text': 'To improve humanities’ visibility in the MD curriculum, the results indicate that greater institutional support, including faculty development and early curricular integration, is required.'}",https://journalhosting.ucalgary.ca/index.php/cmej/article/download/72907/56208
-fpc: fine-tuning with prompt curriculum for relation extraction,Sicheng Yang,"The current classification methods for relation extraction (RE) generally utilize pre-trained language models (PLMs) and have achieved superior results. However, such methods directly treat relation labels as class numbers, therefore they ignore the semantics of relation labels. Recently, prompt-based fine-tuning has been proposed and attracted much attention. This kind of methods insert templates into the input and convert the classification task to a (masked) language modeling problem. With this inspiration, we propose a novel method Fine-tuning with Prompt Curriculum (FPC) for RE, with two distinctive characteristics: the relation prompt learning, introducing an auxiliary prompt-based fine-tuning task to make the model capture the semantics of relation labels; the prompt learning curriculum, a fine-tuning procedure including an increasingly difficult task to adapt the model to the difficult multi-task setting. We have conducted extensive experiments on four widely used RE benchmarks under fully supervised and low-resource settings. The experimental results show that FPC can significantly outperform the existing methods and obtain the new state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method Fine-tuning with Prompt Curriculum (FPC) for RE, with two distinctive characteristics: the relation prompt learning, introducing an auxiliary prompt-based fine-tuned task to make the model capture the semantics of relation labels; the prompt learning curriculum, a fine- Tuning procedure including an increasingly difficult task to adapt the model to the difficult multi-task setting.'}",
-the feasibility and effectiveness of an online curriculum for virtual onboarding/orientation of graduate medical education learners during the covid-19 pandemic,Scott A. Holliday,"Background Graduate medical education (GME) orientation/onboarding is conventionally an in-person activity, but the COVID-19 pandemic prompted virtual approaches to learner onboarding. However, online GME onboarding strategies have not been disseminated in the literature. Objective To determine the usefulness of an online curriculum for GME learner orientation at a large sponsoring institution using an electronic survey. The primary outcome was to discover the usefulness of our online curriculum for GME onboarding, and secondary outcomes included identifying barriers to implementation and weaknesses associated with online GME orientation. Methods We created an online GME orientation curriculum to onboard incoming learners (from June 1 to August 31, 2020) and electronically surveyed our learners to determine the usefulness of this novel approach. We conducted orientation sessions and electronically recorded questionnaire responses using CarmenCanvas, our institutional learning management system. Linear regression analysis was performed to identify factors predicting satisfaction with virtual GME orientation using IBM SPSS Statistics, Version 26.0 (Armonk, NY, USA). Results Of 353 trainees, 272 completed the survey for a 77% response rate. 97% of respondents reported that the curriculum supported performance of learner duties. 79% of trainees perceived the overall quality as “very good” or “good”, 91% responded that the curriculum provided “effective learning”, 94% reported “accessing the course content easily”, 92% reported “easily navigating the curriculum”, 91% described the curriculum as “well-organized”, and 87% reported that the lectures “supported their learning”. Conclusion Online delivery of a comprehensive GME orientation curriculum is useful and facilitates learner education, training, and integration into a large GME institution in the COVID-19 era.",,https://www.dovepress.com/getfile.php?fileID=85408
-wither the plurality of decolonising the curriculum? safe spaces and identitarian politics in the arts and humanities classroom,Ana Cristina Mendes,"Contributing to the debate on decolonising the curriculum, this reflective article questions: What does a safe space in a decolonised classroom mean? For whom is it safe? And at what cost? Must we redraw the parameters of ‘safe’? Prompted by a real-life ‘n-word incident’ in the classroom, this article unpacks the collision of decolonising the curriculum to continue making teaching and learning more pluriversal and inclusive, with the enactment of the ‘wounded attachments’ of identitarian politics and the playing of ‘Privilege or Oppression Olympics’. Using snippets from British parody and satire on decolonising the university, we query how far wokeness in a university setting can become political correctness taken to extremes that threaten decolonising efforts. In its concluding reflections, the article makes tentative recommendations for setting up safe spaces, away from self-silencing or censoring, and backing away from contention and provocation in the classroom.",,https://journals.sagepub.com/doi/pdf/10.1177/14740222221100711
-integrating academic skills in the curriculum: a partnership approach,Emma L. Smith,"Supporting our new students to make the transition to higher education, so that they stay with us and succeed beyond their first year, has been a priority focus for the University of Salford over the past 18 months. As an academic skills team, based in the Library, we have carved out an integral role for our service in responding to this challenge. Building on the prior success of a standalone eLearning programme, we have developed an extensive set of eLearning pathways and complementary learning activities designed for academics to easily and flexibly integrate into their course delivery so that every student is connected with the right academic support at the right time. This presentation aimed to share how our active blended learning approach is scalable and allows for local ownership and opportunities for contextualisation by academic colleagues. We explored how we established our role in this strategic project and the value of our partnership working with the academic community and the VLE support team. We hoped to provide helpful examples of how it has worked in practice to support students to learn how to learn at university. Finally, we sought to reflect on the journey so far - acknowledging the bumps and bends in the road - and initiate a discussion about where to go next. The community response to the presentation provided a space for practitioners to reflect on the ways in which e-learning design could facilitate an embedded approach to academic skills at their institutions; it also prompted reflection on the potential for collaboration between learning developers and Faculties. Our final author reflections address some of the questions and challenges raised at the conference, and set out some future aspirations for where this work will go next.",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/984/612
-pharmacy student challenges and strategies towards initial covid-19 curriculum changes,Lu Liu,"Due to COVID-19, tertiary institutions were forced to deliver knowledge virtually, which proposed challenges for both institutions and students. In this study, we aimed to characterize pharmacy students’ challenges and strategies during COVID-19 curriculum changes, therefore developing a comprehensive understanding of students’ learning, wellbeing, and resilience in the ever-changing situation. Data were collected from student written reflections across four year levels at one school of pharmacy from March–May 2020. In addition, data were collected from written responses of second-year pharmacy students responding to prompted questions. The data were qualitatively analyzed inductively by five coders using NVivo 12. For each piece of data, two coders independently coded the data, calculated the inter-rater agreement, and resolved discrepancies. The most coded challenges were ‘negative emotional response’ and ‘communication barrier during virtual learning’. The most coded strategies were ‘using new technology’ and ‘time management’. This study allows researchers and education institutions to gain an overview of pharmacy students’ experiences during COVID-19, therefore helping universities to provide students with necessary support and techniques on how to self-cope with COVID-19 as well as stressful events in the future.","{'model': 'tldr@v2.0.0', 'text': 'This study helps universities to provide students with necessary support and techniques on how to self-cope with COVID-19 as well as stressful events in the future by developing a comprehensive understanding of students’ learning, wellbeing, and resilience in the ever-changing situation.'}",https://www.mdpi.com/2227-9032/9/10/1322/pdf?version=1633341761
-cvrriculum program faculty development workshop: outcomes and suggestions for improving the way we guide instructors to embed virtual reality into course curriculum,Eva Peisachovich,"Experiential education and student engagement are a main source of student attraction and retention in post secondary milieus. To remain innovative, it is imperative that universities look beyond the internet and traditional multimedia mediums and incorporate novel ways and cutting-edge technologies that can drastically change the way students and educators experience learning. The application of technology as an approach to experiential education is becoming more popular and has extensively impacted universities and other higher education organizations around the world. One approach to support this change in education delivery is to use immersive technologies such as virtual reality (VR). Our team has conducted a pilot study that focuses on embedding VR as a medium to teach empathy within higher education milieus. We began the study by conducting a pilot faculty development workshop to provide an understanding of VR and ways it can be embedded as a pedagogical approach to support curriculum design. Five faculty members from a local university were recruited to participate. Outcomes suggest that embedding VR into the curriculum is a feasible approach that provides an engaging learning environment that is effective for teaching an array of interpersonal skills. The workshop laid the foundation for future faculty training programs guiding the use of VR, prompting a dialog regarding plans for future workshops across a pan-university context.","{'model': 'tldr@v2.0.0', 'text': 'A pilot study that focuses on embeddingVR as a medium to teach empathy within higher education milieus and suggests that embedding VR into the curriculum is a feasible approach that provides an engaging learning environment that is effective for teaching an array of interpersonal skills.'}",https://www.cureus.com/articles/52352-cvrriculum-program-faculty-development-workshop-outcomes-and-suggestions-for-improving-the-way-we-guide-instructors-to-embed-virtual-reality-into-course-curriculum.pdf
-homelessness in the medical curriculum: an analysis of case-based learning content from one canadian medical school,Matthew J To,"Abstract Phenomenon: Homelessness is a major public health concern. Given that homeless individuals have high rates of mortality and morbidity, are more likely to be users of the healthcare system, and often report unmet health needs, it is important to examine how homelessness is addressed in medical education. We wanted to examine content and framing of issues related to homelessness in the case-based learning (CBL) curriculum and provide insights about whether medical students are being adequately trained to meet the health needs of homeless individuals through CBL. Approach: CBL content at a Canadian medical school that featured content related to homelessness was analyzed. Data were extracted from cases for the following variables: curriculum unit (e.g., professionalism/ethics curriculum or biomedical/clinical curriculum), patient characteristics (e.g., age, sex), and medical and social conditions. A thematic analysis was performed on cases related to homelessness. Discrepancies in analysis were resolved by consensus. Findings: Homelessness was mentioned in five (2.6%) of 191 CBL cases in the medical curriculum. Homelessness was significantly more likely to be featured in professionalism/ethics cases than in biomedical/clinical cases (p = .03). Homeless patients were portrayed as socially disadvantaged individuals, and medical learners were prompted to discuss ethical issues related to homeless patients in cases. However, homeless individuals were largely voiceless in cases. Homelessness was associated with serious physical and mental health concerns, but students were rarely prompted to address these concerns. Insights: The health and social needs of homeless individuals are often overlooked in CBL cases in the medical curriculum. Moreover, stereotypes of homelessness may be reinforced through medical training. There are opportunities for growth in addressing the needs of homeless individuals through medical education.","{'model': 'tldr@v2.0.0', 'text': 'Examination of content and framing of issues related to homelessness in the case-based learning (CBL) curriculum and insights about whether medical students are being adequately trained to meet the health needs of homeless individuals through CBL are provided.'}",
-responding to crisis and building forward better: the relevance of curriculum,Yao Ydo,,"{'model': 'tldr@v2.0.0', 'text': 'This triple special issue of Prospects addresses questions head on of curriculum preparedness, by foregrounding the work of more than sixty leading curriculum researchers.'}",https://link.springer.com/content/pdf/10.1007/s11125-021-09572-8.pdf
-evaluation of a technical information systems module for distance learning during the covid-19 pandemic,A. Aryal,"PurposeIn the wake of the pandemic (COVID-19) throughout the United States, many educators had to promptly switch to online modality to continue to provide education to students with safety through physical distancing requirements. This study mainly focuses on delivering an Information Systems module aligned with the information systems curriculum model during a pandemic.Design/methodology/approachThe authors used data collection techniques from student assignments and course tasks. These data are used for the purpose of academic assessment for the Information Systems program. The student performance is assessed on a 5-point scale (1 being low and 5 being high) for the synchronous and the asynchronous tasks related to the course. The authors compared the student performance during the pandemic to the pre-pandemic semester.FindingsThis study revealed that the technical module of an information systems course can be successfully delivered during a pandemic in a remote session. However, the authors found that there is a decline in the student performance in synchronous tasks and asynchronous tasks. But the decline of the student performance in the synchronous tasks is greater than that of the asynchronous tasks. The result of this study helps the Information Systems program with their assessment and to improve their course delivery during a pandemic.Originality/valueIn this paper, the authors examine the delivery of a technical module in the field of information systems via online learning models. The authors particularly examine the synchronous and asynchronous online learning models in the delivery of the technical module. The lessons learned from transitioning to the online modality can help universities better prepare for the future during unprecedented times.","{'model': 'tldr@v2.0.0', 'text': 'It is revealed that the technical module of an information systems course can be successfully delivered during a pandemic in a remote session, however, the authors found that there is a decline in the student performance in synchronous tasks and asynchronous tasks.'}",https://www.emerald.com/insight/content/doi/10.1108/JRIT-11-2021-0078/full/pdf?title=evaluation-of-a-technical-information-systems-module-for-distance-learning-during-the-covid-19-pandemic
-student perceptions on experiential learning within the b.pharm curriculum in nigeria,Abdu-Aguye,"Within the undergraduate pharmacy curriculum for the bachelor of pharmacy (B.Pharm) degree in Nigeria, experiential learning is covered under the Student Industrial Work Experience Scheme (SIWES). The aim of this work was to describe student experiences during SIWES with a view towards identifying problems and recommendations that could be used for improvement. The study was descriptive in nature and utilized qualitative data obtained from students of Ahmadu Bello University, Zaria and University of Maiduguri, Borno. Participating students (151) had completed the 4 year of their B. Pharm degree and undergone their SIWES attachment during the 2015/2016 academic session. They were given 10 questions and asked to answer them either by writing a reflection (essay) or in the form of an open ended questionnaire, data collected was then analyzed using summative qualitative content analysis. Respondents mostly did their attachments in hospitals (71.5%) and community pharmacies (24.5%). They generally believed that they would be exposed to the practical aspects of what they had earlier learnt theoretically, and over half of students reported feeling that their expectations had been met. Patient interaction in the form of counselling was the most enjoyable activity reported by most of the respondents. Conversely, it was also the biggest difficulty encountered. Other difficulties experienced by the students included problems locating drugs on the shelves and learning about the various drug brand names and dosages available. Recommendations on how to improve the scheme included; prompt payment of their stipend to offset transport costs, incorporating more experiential learning opportunities into the curriculum and making it mandatory for students to rotate amongst various practice settings. In conclusion, while students generally had positive perceptions of the scheme, several aspects could be improved",,
-establishing social learning in an engineering mooc: benefits for diversity and inclusion in engineering education,J. Brereton,"Recent Higher Education Statistics Agency data shows that only 20% of engineering students at UK Universities are female, despite the hard work being undertaken by many educational institutions to address this gender imbalance via outreach events and special interventions focussing on girls/women in STEM. It has been argued that student-centred teaching methods, together with changes in the engineering curriculum itself, which emphasise the social, creative, and human-centred aspects of the discipline, are required to effect real change in engaging with those from traditionally underrepresented groups. Through analysing quantitative data on age, gender, learner type, and commenting rates in peer-to-peer discussions, we examine the development and delivery of an engineering MOOC, before, during, and after COVID-19-related lockdowns in the UK, to identify what aspects of online learning might be harnessed to improve diversity in engineering education. The results show that the MOOC attracted a better gender balance than reported for UK-based in-person engineering programmes. In addition, we show that careful structuring of discussion prompts encouraged higher levels of social learning. We recommend the continued use of interactive and discursive elements within a blended learning environment to positively impact diversity and inclusion in engineering education specifically, and STEM education in general.",,https://www.mdpi.com/2071-1050/14/9/5472/pdf?version=1651725710
-secondary mathematics teachers learning to do and teach mathematical modeling: a trajectory,Rose Mary Zbiek,,,https://link.springer.com/content/pdf/10.1007/s10857-022-09550-7.pdf
-learning to perform complex tasks through compositional fine-tuning of language models,Victor S. Bursztyn,"How to usefully encode compositional task structure has long been a core challenge in AI. Recent work in chain of thought prompting has shown that for very large neural language models (LMs), explicitly demonstrating the inferential steps involved in a target task may improve performance over end-to-end learning that focuses on the target task alone. However, chain of thought prompting has significant limitations due to its dependency on huge pretrained LMs. In this work, we present compositional fine-tuning (CFT): an approach based on explicitly decomposing a target task into component tasks, and then fine-tuning smaller LMs on a curriculum of such component tasks. We apply CFT to recommendation tasks in two domains, world travel and local dining, as well as a previously studied inferential task (sports understanding). We show that CFT outperforms end-to-end learning even with equal amounts of data, and gets consistently better as more component tasks are modeled via fine-tuning. Compared with chain of thought prompting, CFT performs at least as well using LMs only 7.4% of the size, and is moreover applicable to task domains for which data are not available during pretraining.","{'model': 'tldr@v2.0.0', 'text': 'This work presents compositional fine-tuning (CFT): an approach based on explicitly decomposing a target task into component tasks, and then fine- Tuning smaller LMs on a curriculum of such component tasks.'}",http://arxiv.org/pdf/2210.12607
-facilitating learners’ self-regulated learning skills and self-efficacy to write in english using technologies,Kretsai Woottipong,"The purpose of this quasi-experimental research was to study the effects of self-regulated strategy development (SRSD) instruction with explicit generalization training prompted by Constructivist Self-regulating Virtual Composing Activities (CSRVCA) on students’ reported use of self-regulatory strategies, their self-efficacy to write and writing performance. The sample group in this study was 86 high school students who were selected through the use of convenience sampling, then 44 students were assigned to a control group and 42 students were in an experimental group. The students from the experimental group received regular writing classes integrated with constructivist virtual writing instruction, while those from the control group received regular writing classes, required by the school curriculum and syllabus. Instruments included surveys, writing tests, and students’ reflective journals. Data were collected through semi-structured interviews and reflective journals to triangulate quantitative results. The results revealed that the students from the experimental group had a significantly higher level of deploying self-regulatory strategies than those in the control group. They also had a higher level of perceived writing self-efficacy than the students in the control group. Although both groups showed improvement in writing scores, the experimental group outperformed the students in the control group in the post and the delayed post-tests.",,https://jurnal.unai.edu/index.php/acuity/article/download/2581/2052
-a scientific research ability prompting approach in the teaching of the machine learning curriculum,Bi Zhao,"The teaching method of machine learning has a significant influence on the promotion of students' research ability. In traditional machine learning curriculum teaching, students have low interest in the theoretical study and lack practical ability of scientific research. It is particularly important to reform the teaching of machine learning to stimulate students' interest in the study, and to cultivate their comprehensive research ability. In this paper, we propose the idea of using exploring education to improve students' scientific research capability. Traditional teaching mode was transformed into an interactive teaching mode by using the methods of interactive teaching and case-teaching. In our practical teaching, our method obviously improved the teaching effect and got positive responses from participated students.","{'model': 'tldr@v2.0.0', 'text': ""This paper proposes the idea of using exploring education to improve students' scientific research capability by using the methods of interactive teaching and case-teaching.""}",http://dpi-proceedings.com/index.php/dtssehs/article/download/3471/3121
-covid-19 compelled rase model curriculum transaction for teacher education -a study,C. Praveen,"The unprecedented shut down of educational institutions owing to the Covid pandemic had prompted many institutions to shift to an Online mode of instruction. The investigator, a teacher educator by profession found the abrupt shift from face-to-face instruction to an online only pedagogy for teacher education a major challenge. This necessitated the exploration of appropriate instructional strategies to substitute the existing time-tested pedagogical practices followed in teacher training colleges. The target audience specific data collected through surveys and interviews revealed that the needs, resources available and internet connectivity of main stake holders namely teachers and learners vary in several ways. This posed several challenges and the investigator found a solution in the RASE Model which is increasingly being perceived as a practical, evidence based design model with applications of technology to improve student learning outcomes and student satisfaction (Churchill, King, & Fox, 2013). When a batch of English optional teacher trainees of the Bachelor of Education (BEd) programme, was assigned to the investigator-cum-teacher educator, the RASE model which basically addresses Resources, Activity, Support and Evaluation was utilized for curriculum transaction. This paper lists down the strategy employed to transact the BEd curriculum using the RASE model. It is hoped that the report of the study will help colleges of teacher education and English language teacher educators to gain a sense of direction to commence delivery of instruction which has already been delayed owing to the pandemic.",,
-a study on the characteristics and elements of best distance learning lecture through topic modeling,Songlee Han,"Objectives The purposes of this study is to identify the characteristics and elements of exemplary remote learning, provide factors that control the quality of effective remote learning, and present implications according to the educational environment changed by the COVID-19 situation. Methods For this study, a total of 324 students(1,178 case) and 459 instructors(459 case) in University A located in Seoul were surveyed using Google Survey, and the collected data was analyzed through topic modelling’s LDA method. Results Survey results show that instructors and students each perceived four topics of remote learning elements. Topics extracted from the students’ surveys are “prompt response to problems that occur”, “difference in the quality of lectures according to the instructors’ design of remote lectures”, “systematic preparation of remote learning and good communication”, and “recreating a curriculum that is suitable for remote learning”. Topics extracted from the instructors’ survey are “supporting the environment and equipment needed for remote learning”, “specific guidelines for remote learning”, “provision of clear grading guideline for remote learning”, and “good communication in remote learning.” Conclusions The results of this study present practical implications for smooth remote learning, such as the need for both students and instructors to have basic media literacy necessary for remote learning and the need to redesign teaching and learning to fit the remote learning environment(e.g., systematic redesign of lectures, interaction, class evaluation criteria, etc., appropriate for remote learning), and propose future research topics. ",,
-exploring the use of writing frames to teach and assess writing in english additional language learning in schools,Marina Burger,"The learning of writing is complex because it requires the development of cognitive and linguistic abilities. Effective teaching and learning of writing demand guided practice. The South African National Curriculum Statement Grades R–12 includes the Curriculum and Assessment Policy Statement (CAPS) for First Additional Language (FAL) teaching and learning. It introduces the use of question-prompt writing frames to learn writing in grade 4. In the second part of the grade 5 year, process writing is introduced as a writing frame that should be used when needed. From grade 7 to grade 12 process writing is implemented as part of the writing lessons. The education department adopted assessment for learning as a teaching and learning approach where feedback provides scaffolded learning. This theoretical paper argues that the writing frames used in the South African curriculum are inadequate to scaffold the learning of writing; furthermore, that the assessment of the writing skills of learners tends to focus on closed skills. The implementation of assessment for learning as an approach to teaching and learning is yet to develop fully to ensure successful scaffolded learning. Additionally, the writing frames introduced limit the teaching of a variety of writing genres and restrict the teaching of writing to western narrative and writing styles. The paper intends to demonstrate that the use of targeted frames appropriate to the writing task and outcome would improve writing learning and assessment.",,https://perlinguam.journals.ac.za/pub/article/download/1044/804
-prompting factors cultivating indonesian efl learners’ learning autonomy behaviors,K. Wijaya,"Previous research results unearthed that highly-autonomous EFL learners are more inclined to possess a higher sense of ownership of their learning enterprises owing to the advanced level of learning management. This library study was a further attempt enacted by the researcher to profoundly investigate particular factors believed to cultivate Indonesian EFL learners’ learning autonomy behaviors. In light of this major research objectivity, 30 previous learning autonomy literature were selectively selected to shed clearer enlightenment for Indonesian EFL experts, practitioners, curriculum designers, and educators to start internalizing various meaningful learning activities consolidating learners’ autonomous learning behaviors. Generally speaking, the obtained research results unveiled that the constant nourishment of Indonesian EFL learners’ learning autonomy behaviors can be insured in case second language teachers are committed to promoting more facilitative and innovative learning enterprises in their daily routines. In conformance with these research results, it is of foremost importance for second language teachers to organize their classroom learning dynamics harmonious with learners’ specific learning situations to advance their autonomous learning behaviors to greater development.",,
-authentic learning tasks within physical education teacher education : redesigning the curriculum,M. Slingerland,"Teaching PE requires an integrated set of knowledge, skills and attitudes and for optimal development students should be exposed to various meaningful learning situations. However, the current School of Sport Studies PETE curriculum seems to be suboptimal due to compartmentalization and fragmentation of knowledge, skills and attitudes, thereby hindering optimal transfer of what is learnt in PETE to what is done in the field. In addition, the PE profession is subject to continuous change (e.g. PE as an examinable subject, PE teachers as liaison officers between school, community and sport clubs), while there is also an increasing demand for reflective practice and life-long learning. Altogether, this has prompted a rigorous redesign of the PETE curriculum. The curriculum redesign process was based on the principles of the four-component instructional design (4C-ID) model (Van Merrienboer, Clark, & Croock, 2002). Central in this approach are authentic learning tasks: situations that emerge from the actual PE profession. First, a skill hierarchy was created that served as a foundation for further curriculum design. Second, a blueprint was worked out, providing a global thematic overview of the four-year curriculum. Then, education was worked out in detail through the design of authentic learning tasks. The design process has resulted in a more coherent curriculum in which authentic learning tasks form the backbone and in which knowledge, skills and attitudes are optimally integrated. The 4C-ID model provides a promising starting point for curriculum (re)design for PETE. Within this presentation an overview of the curriculum will be provided and the suitability of the 4C-ID model as well as various organizational and educational challenges are discussed.",,
-impacts of the covid-19 pandemic on teaching and learning social studies: a literature review,Mohammed Abdullah Al-Nofli,"The unprecedented circumstances surrounding the COVID-19 crisis have affected K-12 education systems in most countries around the world. It is widely recognized that this crisis has presented opportunities and challenges for various school subjects. The purpose of this study was to examine the impacts of the COVID-19 pandemic on teaching and learning social studies. The study reviewed the content of 53 articles on K-12 social studies and its associated subjects (e.g., geography, history, and civic education) published between January 2020 and July 2022. Overall, the research findings indicated that COVID-19 prompted social studies teachers and educators to prioritize important content areas in the social studies curriculum (e.g., health education, human-rights education, digital citizenship, and children’s everyday geographies), integrate a wide range of technologies, and implement innovative teaching practices. In spite of these opportunities, there were some common challenges that raised concern across international contexts, including pedagogical challenges, accessibility and connectivity challenges, administrative challenges, and insufficient professional development for teachers. The study recognized the need for more research on social studies, particularly in the area of learning loss.",,https://ijlter.org/index.php/ijlter/article/download/7467/pdf
-what were you thinking? medical students’ metacognition and perceptions of self-regulated learning,Marjolein Versteeg,"Abstract Phenomenon As a component of self-regulated learning, metacognition is gaining attention in the medical education research community. Metacognition, simply put, is thinking about one’s thinking. Having a metacognitive habit of mind is essential for healthcare professionals. This study identified the metacognitive competencies of medical students as they completed a conceptual learning task, and provided insight into students’ perceptions of self-regulated learning in their curriculum. Approach: Eleven third-year medical students from a Dutch University were purposively sampled to participate in this qualitative study. The study design included a think-aloud assignment followed by a semi-structured interview. During the assignment, participants were instructed to think aloud while solving questions about medical physiological concepts such as blood flow, pressure, and resistance. Think-aloud data were collected through audiotaping and used to identify participants’ metacognitive competencies. The assignment also served as a prompt for an interview in which participants were questioned about metacognitive knowledge, monitoring, experiences, and perceptions of self-regulated learning in their curriculum. All data were transcribed verbatim and analyzed iteratively using a template analysis. Findings: Students differed in their use of metacognitive skills, with an overall focus on monitoring and, to a lesser extent, on planning and evaluation. Additionally, differences were found in students’ metacognitive knowledge and metacognitive experiences. There was apparent use of inefficient, superficial predictive cues. Regarding perceptions of self-regulated learning skills, some students felt no need to develop such skills as they perceived medical education as an exercise in memorizing facts. Others emphasized the need for more insight into their actual level of knowledge and competence. Insights: Pre-clinical medical students require explicit teaching of metacognitive skills to facilitate self-regulated learning. Educators should aim to integrate metacognition in the everyday discourse of the classroom to foster an environment in which students discuss their own learning. Supplemental data for this article is available online at https://doi.org/10.1080/10401334.2021.1889559.","{'model': 'tldr@v2.0.0', 'text': 'The metacognitive competencies of medical students as they completed a conceptual learning task were identified, and insight into students’ perceptions of self-regulated learning in their curriculum was provided.'}",
-creativity in australian health and physical education curriculum and pedagogy,Rosie Welch,"ABSTRACT Notions of creativity are increasingly central to educational scholarship and policies, but few studies attend to the intersection between Health and Physical Education (H/PE) and creativity. In this paper we explore the literature on creativity in education and identify how creativity is represented across a broad spectrum of enquiry yet limited in relation to specifically H/PE. Writing from an Australian context, we examine how the key ideas and propositions from the Australian Curriculum for Health and Physical Education and the Critical and Creative Thinking capability raise important questions regarding the social emergence of educative creative purpose and potential in H/PE. We outline the utility of a ‘creative ecologies’ conceptual framework to consider contemporary practices in H/PE, and apply this to two purposively selected resources; to showcase their contextual eminence for creative learning in H/PE pedagogy and practice. The first resource, Phenomenom!, was funded by Horticultural Innovation Australia to develop food literacy and is linked to cross-curriculum learning outcomes. The second resource, Fitter. Faster. Better, is a St Martin's Youth drama-arts performance where students were prompted to design and perform a fitness programme for adults. Utilising the creative ecologies framework we illustrate the complexity of interconnected creative processes in H/PE as a series of networked elements and cultural flows (policies, practices, products, process, material environments, and partnerships). This framework is then used to identify how the two educational resources exemplify and open up a socially creative practice in both H/PE and everyday life. Our theoretical developments offer new ways to identify, design and enact quality creative and contemporary socio-cultural H/PE curriculum within a broader creative ecology.",,
-self-regulated e-learning in pre-service teacher training for realities of 21st century classrooms,Osman Sadeck,"Educators and students were unprepared for the suspension of face-to-face (f2f) educational activities due to Covid-19, specifically those less experienced in online teaching and learning. Students and educators were traumatised by the sudden switch to online teaching and learning. As such the transition from f2f to exclusive online learning prompted adjusted pedagogical methods and assumed measures of self-regulated e-learning (SRL). During this period researchers embarked on a longitudinal project in Information and Communication Technologies (ICT) called the 21st Century Project (21CP). The purpose of the study was to explore the SRL behaviours of pre-service teachers in online learning and to understand the extent to which a curriculum for technology integration alongside contextual factors influences SRL. As such our conceptualisation of the SRL framework adds two moderating constructs, these are, context and intervention to the core constructs of SRL. The sample of sixty six (66) students were drawn from a volunteer cohort of 166, 4th year pre-service students. Data were collected through interviews, surveys and online journal entries. The data was analysed qualitatively using narrative methods in which themes were identified and reported. The findings revealed that: the ICT integration curriculum represented a cornerstone for SRL development and shaped students’ SRL behaviours; there were differences in the SRL practices that could be attributed to students’ learning habits and attitudes to the ICT-based interventions. The findings of this study provide an understanding of interrelationships among SRL, context, and the design of an online curriculum. The study made three contributions to policy and practice. Firstly, SRL can be improved by providing activities that include clear guidelines for engagement through guided instructional methodologies. Secondly, the faculty curriculum developers should formulate guidelines to ensure that curricular iterations are developed as blended f2f/online modes to enable a quick and seamless transition for exclusive online use. Finally, course creators can improve student engagement by aligning learning outcomes and related activities with learning events.",,https://academic-publishing.org/index.php/ejel/article/download/2659/2148
-technology adaptation of teachers and students under the learning continuity plan: a case of one school in the philippines.,Maricel Sacramento,"The COVID-19 pandemic has prompted educational leaders to restructure the curriculum and modify the instructional set-up to accommodate remote learning of which using technology is the most viable solution to the existing problem. This study explores how teachers adapt and utilise technology-based teaching, and what makes students learn under blended learning modalities in Taytay Senior High School. Quantitatively, using the validated survey questionnaire anchored on the technology adaptation model and the adaptive learning environment model, this study revealed that teachers' age is the factor in all aspects of the model (performance and effort expectancy, social influence and facilitating conditions). When comparing the adaptation levels of students and teachers, it showed that teachers were slightly higher than the students, and that there is a negligible correlation. The findings of this study will serve as baseline data for immediate actions for items that surfaced concerns as hindrance or factors that can hamper students’ academic performance. Keywords: Technology adaptation, online classes, remote learning, Senior High School, Philippines. ",,
-reflective writing in dental education to improve critical thinking and learning: a systematic review.,Janet L. Woldt,"The Commission on Dental Accreditations states that ""graduates must be competent in the use of critical thinking and problem-solving."" With this in mind, dental education programs continually strive to enhance and deepen these skills by incorporating effective instructional strategies into the curriculum. To do this, predoctoral dental education has used a variety of techniques including problem-based learning, OSCEs, and standardized patient activities. Another technique, while less popular but potentially more effective if done well, is reflective writing. This study aimed to assess the effect of reflective writing on students' critical thinking skills and learning in dental education programs. Two database searches conducted between 2000 and 2019 resulted in 317 articles after the initial screening. Ultimately, 13 articles met inclusion criteria. The following recurring factors were identified for inclusion in dental education reflective writing initiative: reflective writing interventions, reflective writing instructions, writing prompts, evaluation frameworks, and negative perceptions of reflective writing. In answering the key aim of this review, 12 of the 13 studies determined that reflective writing had a positive impact on students' critical thinking, judgment, and/or learning. Key findings included positive effects on student growth in reflection, learning through reflective writing, reflection skills, self-assessment, critical thinking, clinical reasoning, problem solving, and motivation to change after evaluated experiences. The review indicated that dental education programs, which implemented reflective writing as an assessment tool within the curriculum as a means of developing and deepening critical thinking skills and learning were by and large successful in this effort.","{'model': 'tldr@v2.0.0', 'text': 'The review indicated that dental education programs, which implemented reflective writing as an assessment tool within the curriculum as a means of developing and deepening critical thinking skills and learning were by and large successful in this effort.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/jdd.12561
-transitioning an advanced practice fellowship curriculum to elearning during the covid-19 pandemic.,Lisa Pierce,"BACKGROUND The COVID-19 global pandemic brought mandatory shelter-in-place orders, disrupting traditional face-to-face teaching methods for advanced practice fellowship programs nationally, creating a challenge for fellowship program directors to preserve curriculum using nontraditional methods. METHOD Using a variety of both web-based and app-based virtual platforms, a nationally accredited fellowship program converted traditional in-seat learning modalities to elearning platforms using both synchronous and asynchronous education. RESULTS Preliminary data indicate that knowledge acquisition and perceived fellow satisfaction are preserved despite the abrupt change to program delivery. Programmatic modifications were submitted to the American Nurse Credentialing Center for compliance and deemed as creative, innovative, and collaborative. CONCLUSION Curriculum for advanced practice fellowship programs can be favorably converted to elearning using virtual platforms during a crisis. Through prompt reevaluation and restructuring, virtual platforms can replace in-seat didactic lectures, patient case studies, mentoring, and even simulation, while ensuring program continuation and compliance with accreditation standards. [J Nurs Educ. 2020;59(9):514-517.].","{'model': 'tldr@v2.0.0', 'text': 'Curriculum for advanced practice fellowship programs can be favorably converted to elearning using virtual platforms during a crisis, and virtual platforms can replace in-seat didactic lectures, patient case studies, mentoring, and even simulation, while ensuring program continuation and compliance with accreditation standards.'}",
-hands-on cybersecurity curriculum using a modular training kit,Mohammad Nasim Imtiaz Khan,"There is an exponential growth in the number of cyber-attack incidents resulting in significant financial loss and national security concerns. Secure cyberspace has been designated as one of the National Academy of Engineering (NAE) Grand Challenges in engineering. Broadly, the security threats are targeted on software programs, operating system and network with the intention to launch confidentiality, integrity and availability violations. Existing undergraduate and graduate-level cybersecurity education curriculum rely primarily on didactic teaching methods with little focus on student centered, inquiry-based teaching, known to improve student learning. With growing number of security incidents taking place, it is of utmost importance to prepare a workforce equipped with knowledge of the threat space and existing state-of-the-art solutions. Such comprehensive understanding is only possible by a dedicated hands-on course on cybersecurity where students can learn the key concepts by editing the hardware, software and OS, and, network policies. Unfortunately, such extensive and deep flexibilities are not provided in current cybersecurity curriculum. In this paper, we introduce a hands-on and modular self-learning Cybersecurity Training (CST) Kit to advance cybersecurity education. Students can promptly apply newly acquired knowledge on the CST Kit as part of the learning process. This Kit accompanies Do-It-Yourself (DIY) training modules that is used to model and investigate cybersecurity issues and their prevention to all levels of the cybersecurity workforce, including undergraduate and graduate students and K-12 science and technology teachers. The Kit also covers various aspects of cybersecurity issues including, hardware, software, operating system and network security. A coursework has been developed on hardware security for senior undergraduate and graduate students using the Kit. A preliminary survey conducted among students who were introduced to the modular board to implement hardware security threats such as, side-channel attack shows an 120% improvement in their understanding after the CST Kit based activities. The components of the CST Kit have also been used in a 4-day summer workshop for K-12 teachers. Teachers took pre-and post- concept inventories to assess their learning of content throughout the workshop and the results indicated improvement of 58%. These assessments focused on vulnerabilities and specific types of attacks, system security, data transmission and encryption, permutations and combinatorics, and binary numbers.","{'model': 'tldr@v2.0.0', 'text': 'A hands-on and modular self-learning Cybersecurity Training (CST) Kit that accompanies Do-It-Yourself (DIY) training modules that is used to model and investigate cybersecurity issues and their prevention to all levels of the cybersecurity workforce, including undergraduate and graduate students and K-12 science and technology teachers.'}",
-mediation and experiential learning: how a mediation clinic can inform a law-based curriculum,B. Waters,"The perception that students of Law and Legal Studies should learn about a variety of methods of dispute resolution and not just litigation, has prompted the Department of Law and Criminal Justice Studies at Canterbury Christ Church University in Kent UK to establish a mediation clinic as a focus for undergraduate experiential learning. This article will consider the importance of discipline based research and the integration of clinical legal education within the core curriculum, the benefits offered by a combined live and simulated curricula approach in the context of mediation and the importance of providing a practical input during the academic stage of legal education.",,https://www.northumbriajournals.co.uk/index.php/ijcle/article/download/48/52
-collaborative online international learning: language and cross-cultural experiences of university students,Frank Ramirez-Marin,"This qualitative study reports on a collaborative online project between the University of Washington at Tacoma (USA) and the Universidad Veracruzana at Veracruz City (México). The project was implemented as part of the internationalization of higher education policies of the participating universities, which include pedagogical practices oriented toward foreign language learning, the internationalization of the curriculum, and virtual exchange. The study documents cross-cultural learning experiences between two groups of students from different linguistic and cultural backgrounds in a Collaborative Online International Learning approach (COIL); and how those experiences related to the development of cross-cultural competence. The language of instruction was English and the students interacted through an online platform and a social network. The methodological design was qualitative-interpretive. Data was generated using interviews, linguistic samples (writing samples), and the interactions of the participants prompted by the use of an online platform and a social media network. Data analysis was realized through a content approach, which led to the formulation of assertions based on themes that emerged. Results indicate that the pedagogical approach implemented (COIL) prompted reflexion on issues related to language learning, cultural understanding, and common life experiences, and that it was conducive to the development of aspects of cross-cultural competence.",,https://revistas.unal.edu.co/index.php/male/article/download/92144/81604
-students engagement and development of generic skills in gamified hybrid service-learning course,Azizah Yusof,"Hybrid Service-Learning, an emerging trend of combining co-curriculum course embedded with service-learning instructional strategies in online settings, has increasingly gained popularity, as many service-learning educators are moving into online platforms. The ongoing COVID-19 pandemic has accentuated its needs, thus forcing education systems worldwide to find alternatives to face-to-face instructions, one of them being Hybrid Service-Learning. This has prompted many researchers to attempt utilising gamification in online courses to increase students’ engagement. The purpose of this research was to report a gamified Hybrid Service-Learning (GAMYSEL) through students’ level of engagement proposed by Schlechty, used as a data classification to compare students’ scorings in their earning of points, badges, level and leaderboards. The results of the study indicated that, the students’ levels of engagement were relatively high through Advocacy Hybrid-Service Learning approach, where majority of them achieved the Strategic Compliance and Optimal Engagement categories. The students’ generic skills were further analysed with one-way repeated measures ANOVA and it was discovered that, GAMYSEL had positive impacts in developing these skills in five stages of service-learning. Both of these findings suggest that student’s engagement and generic skills had a significant effect on students learning after the use of GAMYSEL.",,
-homeschooling for quarantined residents: a virtual back to the basics curriculum,Anthony Sielicki,"Introduction The COVID 19 pandemic resulted in local and institutional restrictions with significant effects on the clinical environment for graduate medical education, displacing residents from non-emergency medicine (EM) based rotations. Additionally, resident physicians considered patients under investigation (PUI) were furloughed from clinical practice. The necessity for supplemental learning in a virtual setting prompted the development of an online homeschooling curriculum that incorporated back to the basics textbook learning, application, and retention via virtual sessions for the quarantined and furloughed learners. Methods An online homeschooling curriculum was developed to replace the cancelled clinical experiences for EM residents and for those who were quarantined utilizing Google Classroom and Zoom teleconference software. After completion of their quarantine or return to normal rotation schedule, residents were asked to evaluate the homeschooling curriculum using an anonymous survey. Results A total of 12 residents participated in the homeschooling program over eight weeks during the spring of 2020. Of the nine residents surveyed, 88.8% percent felt the homeschooling added to their knowledge of EM, 100% found the online format easy to use, and 88.8% stated it helped maintain a sense of social connection to peers and faculty. Conclusion An online homeschooling program was considered an effective means of providing an opportunity for synchronous and continuous education for EM resident physicians. This program could be sustainable long term to fill in knowledge gaps or supplement remediation in emergency resident education, post pandemic.","{'model': 'tldr@v2.0.0', 'text': 'An online homeschooling program was considered an effective means of providing an opportunity for synchronous and continuous education for EM resident physicians and could be sustainable long term to fill in knowledge gaps or supplement remediation in emergency resident education, post pandemic.'}",https://www.cureus.com/articles/40374-homeschooling-for-quarantined-residents-a-virtual-back-to-the-basics-curriculum.pdf
-ants as an experiential learning strategy in preschool teacher training,Rosa María Gálvez Esteban,"In this chapter, the authors present the results of a project designed for 41 preservice preschool teachers to introduce the concept of living things as an experiential learning strategy in the classroom. The need to approach this concept from a different perspective prompted the design of an education project involving the introduction of insects into classroom as a teaching resource. An informative storyline was used for project launch presentation. The questions they strive to answer in this chapter are related with what concepts of living organisms and what inquiry stages will preservice teachers consider their pupils will carry out during the project. Relevant concepts that are usually not much covered in the preschool curriculum such as the life cycles of animals were considered by 23 participants. Twenty-five of the future teachers claimed that they would be able to work on every inquiry step if they implemented this project in the classroom.",,
-the covid-19 elective for pediatric residents: learning about systems-based practice during a pandemic,Maya S. Iyer,"Background: The coronavirus disease 2019 (COVID-19) pandemic has prompted pediatric residency programs to adjust the delivery of educational curricula and to update content relevant to the pandemic. Objective: In this descriptive paper, we present how we rapidly developed and implemented a COVID-19 pandemic elective for pediatric residents. Methods: This curriculum was established at a single tertiary care children’s hospital in June 2020. We used the ADDIE (analysis, design, development, implementation, evaluation) framework to develop a two-week elective (30 hours) consisting of six flexibly scheduled modules. We administered post-elective surveys and exit interviews to solicit feedback to improve the elective and obtain effectiveness of our educational interventions. Results: We developed an asynchronous online COVID-19 Elective for Pediatric Residents. The curriculum modules focus on pathophysiology of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the disaster management ecosystem, simulation of clinical care, mental health ramifications, and public health consequences. We also include six in-situ experiences (visits to a drive-through COVID-19 testing site, testing laboratory and local public health department, a simulation of a critically ill child, and meetings with emergency managers and social workers) to solidify learning and allow for further reflection. To date, eight participants have taken the elective. All participants strongly agreed on a five-point Likert item survey that the elective enhanced their knowledge in current evidence-based literature for COVID-19, disaster preparedness, hospital response, management of the critically ill child, and mental and public health ramifications. All participants agreed this curriculum was relevant to and will change their practice. Conclusions: We demonstrate how a COVID-19 elective for pediatric residents could be quickly developed and implemented. The pilot results show that pediatric trainees value asynchronous learning, supplemented by relevant in-situ experiences. Moreover, these results suggest that this curriculum provides needed disaster response and resiliency education for pediatric residents.","{'model': 'tldr@v2.0.0', 'text': 'The pilot results show that pediatric trainees value asynchronous learning, supplemented by relevant in-situ experiences, and suggest that this curriculum provides needed disaster response and resiliency education for pediatric residents.'}",https://assets.cureus.com/uploads/original_article/pdf/50576/20210305-32377-1ulh3cm.pdf
-a case study of curriculum development: backward or forward/central design?,N. T. Tung,"curriculum development, backward design, constructive alignment, objective, syllabus, methodology, evaluation Nowadays universities in Vietnam have begun to update their curricula by adopting a backward design with a focus on students’ learning outcomes to replace the old-fashioned forward model. However, to have a constructive alignment is a problem they have to face. This situation has prompted this case study research in May 2020, intending to examine a curriculum to find out whether it is coherent in terms of its main components – objectives, syllabus, methodology, and evaluation. The findings of the study indicate that for this case, although the curriculum was claimed to be of a backward design with learning outcomes as program objectives, it turned out to follow a forward one with more focus on knowledge transfer than competency development. In particular, as the program learning objectives were still written according to knowledge transmission, it was not constructively aligned with the three main components of syllabus, methodology and evaluation: The syllabi were mainly based on the experience of the experts in the field or syllabus designers, the innovation of the methodology as directed by the university with a shift of focus from the lecturer to learners was still interpreted rigidly and mechanically, and the evaluation of student learning was generally claimed to be criterionreferenced only without any elaboration for each syllabus.",,https://journalofscience.ou.edu.vn/index.php/soci-en/article/download/546/478
-letter to the editor regarding “the evolving impact of covid-19 on medical student orthopedic education: perspectives from medical students in different phases of the curriculum”,Shi-qi Wang,"Letter to the editor Richardson et al. recently published a paper in Geriatric Orthopaedic Surgery & Rehabilitation entitled “The Evolving Impact of COVID-19 on Medical Student Orthopedic Education: Perspectives From Medical Students in Different Phases of the Curriculum.” Richardson et al. mentioned that the COVID-19 pandemic has changed medical students’ orthopedic education curriculum and provided methods that may help students seek alternative programs and supplementary learning. These are very enlightening to the practice-based orthopedic rehabilitation medical education. During the early stage of the COVID-19 epidemic, medical colleges and universities in China also implemented similar management measures, including reducing or even stopping students’ clinical rotation and adopting network-based professional course teaching. However, in the process of improving the teaching methods, we also found the corresponding problems. First of all, the efficiency of online teaching and online academic lectures is challenging to be guaranteed. During the epidemic, online teaching and online academic lectures have broken the restrictions of location. Although they can effectively reduce the crowd gathering and the risk of infection, they will also greatly reduce some students’ learning efficiency with poor self-consciousness. The live broadcast software prompts students to be in a “listening state,” but they may be doing things that have nothing to do with teaching in the background and cannot be effectively supervised. In addition, some elderly teachers are not proficient in the use of teaching software, which further leads to the decline of students’ learning efficiency. During the teaching period, the department of rehabilitation medicine of our school received much negative feedback from students that the online teaching experience was inferior. This result is similar to that of Jordan, where students’ satisfaction with online teaching is very low. Secondly, the reduction of practical operation makes it difficult for students to adapt to clinical practice. Orthopedic sports rehabilitation is a subject that includes many clinical operations, such as joint mobilization, muscle strength exercise, and balance training. However, due to the lack of practical clinical teaching, after returning to the clinic in May, the proficiency of the current students is significantly lower than that of the previous students. We also asked students to watch the operation video repeatedly during the epidemic, and there were conditions to carry out simple action exercises between students, but the final effect was not satisfactory. From the above problems, it is not difficult to find that the self-consciousness of students plays a key role during the epidemic. The 5 solutions mentioned by Richardson et al. also depend on students’ self-consciousness, especially in reading important professional articles. In addition, online teaching","{'model': 'tldr@v2.0.0', 'text': 'The COVID-19 pandemic has changed medical students’ orthopedic education curriculum and provided methods that may help students seek alternative programs and supplementary learning, but from the corresponding problems, it is not difficult to find that the self-consciousness of students plays a key role during the epidemic.'}",https://journals.sagepub.com/doi/pdf/10.1177/2151459320973191
-effectiveness of the proposed robotics instructional design on “basic movements” for online distance learning: determining the effectiveness of the proposed rid module through collaborative action research,Genevieve A. Pillar,"Robotics brings education technology to a new level. Introducing robotics to schools does not only mean making hands-on activities but to prepare students for the future. The creation of Robotics Online Distance Learning at the De La Salle Santiago Zobel (DLSZ) school, and its integration to the curriculum prompted the researchers to come up with an evaluation to determine the effectiveness of the proposed Robotics Instructional Design (RID) module for “Robotics Basic Movements” facilitated through Online Distance Learning (ODL). This collaborative action research involved Grade 3 (N1=224) and Grade 4 (N2=215) students of DLSZ who served as the participants to the proposed RID module. Results from this study revealed that In terms of learning modalities, majority of the students were able to manipulate the apps using their available equipment at home. The Grade 3 group 0.377 standard deviation from the average mean of 15.193 indicates the better performance. It is also seen that based on the computed Grade 4 group results, the standard deviation of 1.07 with the average mean of 4.235 is higher compared to the other group. Most of the performances are close to the expected average, causing a symmetrical data distribution. In terms of learning targets, these became helpful for the teachers working with the students. These made lessons and activities easier to plan, monitor, and assess students’ learning. These also guided instructional designers to make sure that all students would reach mastery level. It is suggested that the proposed RID module design should involve processes from designing, to building and programming. This innovative way of transmitting skills revealed to be useful for all robotics competencies.",,
-student voice in an extended curriculum programme in the era of social media: a systematic review of academic literature,J. Chukwuere,"Student voice in this digital age and across higher learning institutions is increasing exponentially with the function of social media. Student voice provides a vibrant communication pathway to extended curriculum programme students in higher education institutions. Social media ensures active participation of extended curriculum programme students in generating views and ideas that define the higher learning environment and experience towards better learning conditions and outcomes. A systematic literature review was used in gathering scientific papers through trusted academic databases. The systematic literature review was conducted between the period of 1 April 2019 and 28 September 2019, by looking into the contents of articles covering the current research objectives. The study’s findings show that social media provides an effective and instant spread of the extended curriculum programme students’ voice across higher education learning institutions. It also allows the students in the extended curriculum programme to engage with each other and the institutional management promptly. Social media promotes extended curriculum programme students’ voice in reaching the right audience at the right time. The results of this study are key for extended curriculum programme students, lecturers, and university management in understanding and applying social media effectively and in bringing transformation to South African higher education institutions and beyond.",,
-steam education: student learning and transferable skills,M. Bertrand,"PurposeGlobally, interdisciplinary and transdisciplinary learning in schools has become an increasingly popular and growing area of interest for educational reform. This prompts discussions about Science, Technology, Engineering, Arts and Mathematics (STEAM), which is shifting educational paradigms toward art integration in science, technology, engineering and mathematics (STEM) subjects. Authentic tasks (i.e. real-world problems) address complex or multistep questions and offer opportunities to integrate disciplines across science and arts, such as in STEAM. The main purpose of this study is to better understand the STEAM instructional programs and student learning offered by nonprofit organizations and by publicly funded schools in Ontario, Canada.Design/methodology/approachThis study addresses the following research question: what interdisciplinary and transdisciplinary skills do students learn through different models of STEAM education in nonprofit and in-school contexts? We carried out a qualitative case study in which we conducted interviews, observations and data analysis of curriculum documents. A total of 103 participants (19 adults – director and instructors/teachers – and 84 students) participated in the study. The four STEAM programs comparatively taught both discipline specific and beyond discipline character-building skills. The skills taught included: critical thinking and problem solving; collaboration and communication; and creativity and innovation.FindingsThe main findings on student learning focused on students developing perseverance and adaptability, and them learning transferable skills.Originality/valueIn contrast to other research on STEAM, this study identifies both the enablers and the tensions. Also, we stress ongoing engagement with stakeholders (focus group), which has the potential to impact change in teaching and teacher development, as well as in related policies.",,https://www.emerald.com/insight/content/doi/10.1108/JRIT-01-2020-0003/full/pdf?title=steam-education-student-learning-and-transferable-skills
-semantic mapping of learning assets to align curriculum and evidence learning effectiveness in business education,Chad Patrizi,"Ensuring the alignment of course content against desired goals and objectives has always been at the core of effective instructional course design—whether the instruction is delivered face-to-face, or via the support of distance technologies. Nonetheless, with the latter delivery of instruction, two distinct challenges have recently emerged, thus prompting the need to re-examine the term quality as it relates to accreditation. The first challenge lies within the ability to locate and/or create and align digital learning objects to course goals and objectives. The second challenge lies within the ability to demonstrate learning effectiveness from learning management system metadata. A research and development team at American Public University System addressed these challenges by employing an open source repository and semantic engine for analysis and alignment of content, materials, and learning activities to goals and objectives across all courses within the School of Business. The result was a highly detailed, accurate mapping of the programs' knowledge base to established goals and objectives. As an added benefit, resources were disaggregated to a very granular level and sorted into taxonomies that can be independently referenced and utilized for cross-curricular consumption.","{'model': 'tldr@v2.0.0', 'text': 'A research and development team at American Public University System addressed challenges by employing an open source repository and semantic engine for analysis and alignment of content, materials, and learning activities to goals and objectives across all courses within the School of Business.'}",
-facilitating virtual medicinal chemistry active learning assignments using advanced zoom features during covid-19 campus closure,Mudit K. Singhal,"Active student participation is critical to the delivery of medicinal chemistry components of the pharmacy school curriculum. Carefully structured active learning strategies can promote student engagement while increasing course material application. Topic-based active learning assignments can be administered, where learners apply constructed knowledge and skills to complete assignments. The suspension of on-campus classes due to the COVID-19 pandemic prompted the development of innovative ways to deliver active learning assignments without using the school’s physical team rooms. To actively engage learners, the Zoom video teleconferencing platform with breakout rooms was used to conduct medicinal chemistry active learning assignments in a Pharmacotherapeutics IV: Endocrinology course. The students’ assignment scores showed no significant difference when assignments were conducted virtually using Zoom breakout rooms versus on campus, suggesting successful virtual administration of the assignments and students’ prompt adaptation to the new environment. The combination of Zoom-based lectures with actively facilitated Zoom breakout room assignments is an effective active learning strategy.","{'model': 'tldr@v2.0.0', 'text': 'The combination of Zoom-based lectures with actively facilitated Zoom breakout room assignments is an effective active learning strategy, and the students’ assignment scores showed no significant difference when assignments were conducted virtually using Zoom breakout rooms versus on campus.'}",
-the uses of facebook© technologies in hospitality curriculum on an experiential learning platform for a new generation of students,Christopher W. Harris,"Purpose – The purpose of this paper is to examine the applications of the social networking software, Facebook, currently practiced by “expert user” faculty teaching within a Malaysian Higher Education Hospitality Diploma programme. Of particular interest is whether the faculty experiences with Facebook reveal its functional appropriateness for use on an experiential learning platform. The aim is also to prompt further research and experimentation with the medium on teachers and students.Design/methodology/approach – Phenomenological: the researcher uses a focus group to enable the essence of teachers' experience in the utilisation of Facebook as a teaching and learning tool within the syllabus of their taught subjects in a shared hospitality curriculum to emerge.Findings – The experiences divulged in the focus group reveal the use of Facebook as an “effective” medium for reflective purposes in relation to experiential teaching and learning activities and as a mechanism for reflective and extrapolative co...",,
-does online engagement matter? the impact of interactive learning modules and synchronous class attendance on student achievement in an immersive delivery model,E. Goode,"One Australian public university is radically changing the way it delivers higher education, introducing a 6-week immersive scheduling delivery model across all units and courses. Despite the emerging success of block and immersive models for raising the performance of diverse student cohorts, the design factors underpinning positive outcomes are underexplored. This paper presents a mixed methods study of the impact and value of student engagement with interactive and responsive online content modules and synchronous classes in an immersive scheduling model. The findings indicate that behavioural engagement with online learning modules has a positive effect on academic success and is a significant predictor of a higher final score. Qualitative data indicate several attributes of high-quality online learning modules that students appear to associate with engagement and deeper learning in the immersive model: interactivity, media richness, constructive alignment, flexibility and responsiveness. Synchronous class attendance did not impact final scores; however, students nonetheless valued the opportunity to form safe and supportive communities of inquiry during classes. This study demonstrates that in times of increasing demand for more flexible learning, immersive scheduling models that are founded on active learning principles and embed interactive, responsive, media-rich online learning modules can improve student engagement and performance. Implications for practice or policy: Higher education practitioners should integrate interactive, responsive, media-rich and constructively aligned online learning modules into curricula. Synchronous active learning classes that create safe communities of inquiry should be offered alongside options for asynchronous participation. Low levels of engagement with online learning modules should prompt follow-up from educators to raise engagement and bolster academic achievement. Immersive delivery models are effective curriculum innovations that, when designed with interactive online modules, can support improved academic achievement. ",,https://ajet.org.au/index.php/AJET/article/download/7929/1922
-enhancing the 4cs among college students of a communication skills course in tanzania through a project-based learning model,M. Saimon,,,https://link.springer.com/content/pdf/10.1007/s10639-022-11406-9.pdf
-fully immersive virtual reality for total hip arthroplasty: objective measurement of skills and transfer of visuospatial performance after a competency-based simulation curriculum.,K. Logishetty,"BACKGROUND Fully immersive virtual reality (VR) uses headsets to situate a surgeon in a virtual operating room to perform open surgical procedures. The aims of this study were to determine (1) if a VR curriculum for training residents to perform anterior approach total hip replacement (AA-THR) was feasible, (2) if VR enabled residents' performance to be measured objectively, and (3) if cognitive and motor skills that were learned with use of VR were transferred to the physical world. METHODS The performance of 32 orthopaedic residents (surgical postgraduate years [PGY]-1 through 4) with no prior experience with AA-THR was measured during 5 consecutive VR training and assessment sessions. Outcome measures were related to procedural sequence, efficiency of movement, duration of surgery, and visuospatial precision in acetabular component positioning and femoral neck osteotomy, and were compared with the performance of 4 expert hip surgeons to establish competency-based criteria. Pretraining and post-training assessments on dry bone models were used to assess the transfer of visuospatial skills from VR to the physical world. RESULTS Residents progressively developed surgical skills in VR on a learning curve through repeated practice, plateauing, on average, after 4 sessions (4.1 ± 0.6 hours); they reached expert VR levels for 9 of 10 metrics (except femoral osteotomy angle). Procedural errors were reduced by 79%, assistive prompts were reduced by 70%, and procedural duration was reduced by 28%. Dominant and nondominant hand movements were reduced by 35% and 36%, respectively, and head movement was reduced by 44%. Femoral osteotomy was performed more accurately, and acetabular implant orientation improved in VR assessments. In the physical world assessments, experts were more accurate than residents prior to simulation, but were matched by residents after simulation for all of the metrics except femoral osteotomy angle. The residents who performed best in VR were the most accurate in the physical world, while 2 residents were unable to achieve competence despite sustained practice. CONCLUSIONS For novice surgeons learning AA-THR skills, fully immersive VR technology can objectively measure progress in the acquisition of surgical skills as measured by procedural sequence, efficiency of movement, and visuospatial accuracy. Skills learned in this environment are transferred to the physical environment.","{'model': 'tldr@v2.0.0', 'text': 'For novice surgeons learning AA-THR skills, fully immersive VR technology can objectively measure progress in the acquisition of surgical skills as measured by procedural sequence, efficiency of movement, and visuospatial accuracy.'}",
-is reflective learning visible in online discussion forums for medical students on general practice placements? a qualitative study,Kelly Gillingham,"Abstract Problem: Reflection is important for lifelong learning and professional development, and discussion forums have been highlighted as a potential platform for enhancing reflection further through peer interaction and teacher feedback. Forums provide students on general practice (GP) placement the opportunity to engage in collaborative reflective learning despite their geographical isolation and disparate schedules. This case report seeks to explore whether online discussion forums demonstrate community presence and reflective learning among medical students on GP placement. Intervention: Online discussion forums were introduced into the curriculum for University of Auckland Year 5 and 6 medical students on their respective 4- and 6-week GP placement. Via asynchronous posts and comments, groups of eight students on average presented cases and experiences for discussion with peers and an overseeing GP faculty member. Context: Students were dispersed across a large geographical area and were unlikely to have peer contact in their assigned practices. Online discussion forums were implemented for a number of reasons, including facilitating reflective learning. An adapted community of inquiry framework was used to investigate the cognitive, social, and teacher presence elements visible in the discussion forum transcripts of unanimously consenting groups. Content analysis of the transcripts was performed to evaluate the presence and quality of reflective learning. Impact: The forums were predominately student-led with relatively scarce comments by GP faculty facilitators. The majority of cases and experiences presented related to clinical management of patients and dissonance triggered by the on-site supervisor’s actions. Ideas, knowledge, and understanding of presented encounters were shared and built on by positive and supportive interaction, broadening students’ perspectives and, at times, leading to the formation of solutions. However, students’ reflection was mostly superficial. Deep reflection was rarely present, and affective dimensions of reflection were disclosed relatively less than clinical cognitive aspects of reflection. Direct instruction, based on prior experience and pedagogical expertise, was the main form of comment made by faculty facilitators, who prompted with questions to a far lesser extent. Lessons Learned: Online discussion forums appeared to enable medical students on GP placement to interact positively with peers and faculty facilitators. However, deep reflection was not reached. Modifying the discussion forums to facilitate more peer interaction and addressing the barriers that limit faculty facilitation may encourage deeper and more affective reflections.","{'model': 'tldr@v2.0.0', 'text': 'Modifying the discussion forums to facilitate more peer interaction and addressing the barriers that limit faculty facilitation may encourage deeper and more affective reflections.'}",
-a personalized approach to learning across time and space,Rachel Fitzgerald,"In a higher education environment massively disrupted by the pandemic, the importance of exploring the efficacy of purposeful, flexible, online learning is essential. Research shows that blended approaches to learning can be active, flexible, and student-centred. However, this research suggests that without human intervention or a bespoke context, there can also be very poor engagement. As such, this study presents the successes and challenges of introducing a personalised mobile-learning resource in a supported, personalized environment. The resource was designed and developed in 2019 and integrated in 2020 and 2021 into a large (n=200+) communication design class at an Australian university. This paper evaluates the effectiveness of the app using measures drawn from Abdullah & Ward’s (2016) take on the Technology Acceptance Model. Their model, GETAMEL (general extended technology acceptance model for e-learning) measures technology acceptance factors in eLearning such as enthusiasm, self-efficacy, and student enjoyment of technology in relation to contextual purpose. This paper also considers the value of self-directed learning and co-curricular activities in the context of additional learning to support core curriculum. This paper presents an evaluation of student experience drawn from a class-wide survey and analysis of usage data. We conclude that the use of a personalised app fosters an autonomously supportive experience that leads to intrinsic motivation and improves overall learning (Baker & Goodboy, 2018). The app provides additional support for students unable to attend tutorials and enabled students to remain engaged and abreast of the general topics each week and provided some additional prompts about self-directed learning activities they could engage in, that they wouldn’t have otherwise had. Enjoyment, as one of the key factors of the GETAMEL model, was explicitly mentioned by students. While a personalised tool is labour intensive and expensive in cost and time, this study concludes that the rewards regarding student experience make it worthwhile.",,https://papers.academic-conferences.org/index.php/ecel/article/download/901/775
-building professional competencies through a service learning ' gallery walk' in primary school teacher education,S. Ramsaroop,"This article reports on a service learning project in a South African primary school teacher education programme, as experiential and practice-based pedagogy in a social studies methods course. We aimed to broaden understanding of service learning as a form of non-placement work-integrated learning for the development of teacher professional competencies. Student teachers drew on topics in the middle school social studies curriculum and incorporated Indigenous geographical elements with local community history in the design of a service learning ‘gallery walk’ for Grade 5 learners. Using a generic qualitative design, data were generated from students’ and teachers’ reflective journals, lesson plans, photographs and video recordings. It was analysed for common content themes and prominent discourse markers of students’ developing professional knowledge and competencies. The findings provide evidence of deepened student learning, particularly on the influence of context and curriculum differentiation and how their struggles with group work enabled the development of collaboration and cooperation required by professionals. In addition, the service learning prompted changing notions of citizenship and reciprocity of learning.",,https://ro.uow.edu.au/cgi/viewcontent.cgi?article=2256&context=jutlp
-robot open-ended autonomous learning competition,Emilio Cartoni,"Open-ended learning, also called ‘life-long learning’ or ‘autonomous curriculum learning’, aims to program machines and robots that autonomously acquire knowledge and skills in a cumulative fashion. We illustrate the first edition of the REAL-2019 – Robot open-Ended Autonomous Learning competition, prompted by the EU project GOAL-Robots – Goalbased Open-ended Autonomous Learning Robots. The competition was based on a simulated robot that: (a) acquires sensorimotor competence to interact with objects on a table; (b) learns autonomously based on mechanisms such as curiosity, intrinsic motivations, and self-generated goals. The competition featured a first ‘intrinsic phase’, where the robots learned to interact with the objects in a fully autonomous way (no rewards, predefined tasks or human guidance), and a second ‘extrinsic phase’, where the acquired knowledge was evaluated with tasks unknown during the first phase. The competition ran online on AIcrowd for six months, involved 75 subscribers and 6 finalists, and was presented at NeurIPS-2019. The competition revealed very hard as it involved difficult machine learning challenges usually tackled in isolation, such as exploration, sparse rewards, object learning, generalisation, catastrophic interference, and autonomous skill learning. Following the participant’s positive feedback, the preparation of a second REAL-2020 competition is underway, improving on the formulation of a relevant benchmark for open-ended learning.","{'model': 'tldr@v2.0.0', 'text': 'The first edition of the REAL-2019 – Robot open-Ended Autonomous Learning competition, prompted by the EU project GOAL-Robots, is illustrated, based on a simulated robot that acquires sensorimotor competence to interact with objects on a table.'}",
-contemplating curriculum in an urban world,Philip Roberts,,,https://link.springer.com/content/pdf/10.1007/s41297-023-00194-y.pdf
-"transformational learning: using equality and diversity marketing client briefs to foster awareness, application and action",Sally Everett,"ABSTRACT This paper reports a study coauthored with second-year undergraduate students that examines student experiences of undertaking real-life, client-set marketing assessments with an equality and inclusion remit. Students were set a marketing assessment with an explicit social justice focus, thereby prompting them to reflect on their own backgrounds and unconscious biases whilst addressing issues of societal injustice. This study explores the impact on students of closely engaging with this material and how the assessment, which involved them interviewing experts and developing actionable recommendations, altered their views and career plans. Data were generated via focus groups with students who undertook the assessment, supplemented by some alumni and client interviews. The findings are presented as a “3A” framework of Awareness, Application, and Action: first, the assessment raised “Awareness” of broader societal issues and fostered transformation in terms of personal views; second, students developed core skills and networks they could “Apply” in their future careers; and third, “Action” describes the way students felt empowered to make a positive difference to their client and wider society. The 3A model provides simple design principles that educators could implement if they want to introduce transformational experiences based on social justice issues into the marketing curriculum.",,
-learning developers as their own cultural critics?,Christie Pritchard,"The values that guide the work of ALDinHE and those associated with us as an organisation include ‘making HE inclusive through emancipatory practice, partnership working and collaboration’ and ‘critical self-reflection, on-going learning and a commitment to professional development’ (ALDinHE, 2022). However, considering this from a relational, or systems thinking perspective; how achievable are these values? Can we truly be inclusive to all? Exploring this notion in my doctoral research, using Bourdieu’s relational framework (Bourdieu, 1992; Bourdieu, 1993; Bourdieu, 1997; Bourdieu & Passeron, 1990) I was confronted with some uncomfortable truths. Education is reproductive in the sense that it selects those with the necessary capital to succeed and nurtures them to develop further. The diversity of student’s prior experience, background and capital to succeed is clear, and Learning Developers know how difficult Higher Education can be to navigate, but how often do we reflexively consider how our work reproduces the cultural system. Conference attendees working in small groups, discussed a brief precis of my Doctoral research, a copy of ALDinHE values and summary of Bourdieu’s notion of autonomy (Bourdieu, 1992). The following questions were posed to prompt discussion. How autonomous are we as practitioners and are we able to change the system that reproduces the inequality of society? Is being positioned by students as part of their curriculum through embedded practice advantageous all of the time? Do we focus on the knowledge and skills students bring, or are we forced to help them adapt to the game? ",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/967/631
-development of a novel interactive multimedia e-learning model to enhance clinical competency training and quality of care among medical students,Yu-ting Hsiao,"Clinical competencies consisting of skills, knowledge, and communication techniques should be acquired by all medical graduates to optimize healthcare quality. However, transitioning from observation to hands-on learning in clinical competencies poses a challenge to medical students. The aim of this study is to evaluate the impact of a novel interactive multimedia eBook curriculum in clinical competency training. Ninety-six medical students were recruited. Students in the control group (n = 46) were taught clinical competencies via conventional teaching, while students in the experimental group (n = 50) were taught with conventional teaching plus interactive multimedia eBooks. The outcomes of clinical competencies were evaluated using Objective Structured Clinical Examination (OSCE) scores, and feedback on their interactive eBook experiences was obtained. In the experimental group, the average National OSCE scores were not only higher than the control group (214.8 vs. 206.5, p < 0.001), but also showed a quicker improvement when comparing between three consecutive mock OSCEs (p < 0.001). In response to open-ended questions, participants emphasized the importance of eBooks in improving their abilities and self-confidence when dealing with ‘difficult’ patients. Implementing interactive multimedia eBooks could prompt a more rapid improvement in clinical skill performance to provide safer healthcare, indicating the potential of our innovative module in enhancing clinical competencies.","{'model': 'tldr@v2.0.0', 'text': 'Implementing interactive multimedia eBooks could prompt a more rapid improvement in clinical skill performance to provide safer healthcare, indicating the potential of the innovative module in enhancing clinical competencies.'}",https://www.mdpi.com/2227-9032/8/4/500/pdf?version=1605847025
-the australian curriculum: mathematics as an opportunity to support teachers and improve student learning,P. Sullivan,"The creation of a national, as distinct from state and territory based, mathematics curriculum creates important opportunities for improving learning, but whether those opportunities are taken up will depend on the ways that teachers are supported, including by teacher educators, coaches, school leaders and readers of this monograph. The debates that continue on aspects of content are irrelevant to whether the national curriculum provides a prompt to improvement. In fact, such debates are a by-product of the negotiations that are an obvious artefact of the ceding of responsibilities by local jurisdictions to a national authority. The real opportunities for improving mathematics learning are in the principles that underpin the structure of the curriculum and the use of these principles to inform teacher learning.",,
-advocacy and awareness: integrating lgbtq health education into the prelicensure curriculum.,Paula McNiel,"BACKGROUND An identified gap in the curriculum related to lesbian, gay, bisexual, transgender, and queer (LGBTQ) health needs prompted nursing faculty to implement a collaborative educational offering. LGBTQ individuals experience significant health disparities, compared with heterosexual counterparts. Enhancing established LGBTQ population-specific training to highlight health disparities and awareness of special health care needs was piloted with two clinical groups of senior baccalaureate nursing students (N = 16). METHOD Didactic, simulated, and panel discussion related to LGBTQ terminology, current health standards of care, and the importance of advocacy was provided by campus advocates, experienced health care providers, and a student panel identifying as LGBTQ. Health specific learning outcomes were established and evaluated. RESULTS Posteducation, anonymous surveys, and journaling were completed. Survey respondents (n = 13) reported increased awareness and understanding of health disparities specific to the LGBTQ population. CONCLUSION LGBTQ-specific health education has been implemented as a permanent curriculum change. [J Nurs Educ. 2018;57(5):312-314.].","{'model': 'tldr@v2.0.0', 'text': 'enhancing established LGBTQ population-specific training to highlight health disparities and awareness of special health care needs was piloted with two clinical groups of senior baccalaureate nursing students.'}",
-embodied cognition and curriculum construction,Meiqian Wang,"Abstract The disembodiment of cognitive science has resulted in curricula with disembodied concepts and practice. The emergence of the embodied cognitive science provoked public reflections on the nature of the curriculum. This has elevated the body from the ‘peripheral’ position to the ‘central’ position, acting as the subject in action and becoming the bridge to experience transformations. Meanwhile, the nurturing role of the environment for the mind is attracting increasingly more attention, and the environment, the body, and the mind jointly constitute a dynamic system that prompts the transformation and restructuring of learners’ experience. The reshaping of the relationship between the body and education and the integration of physical experience and mind-building have led us to examine curriculum from a completely new perspective, and curriculum construction is thus transformed to the dynamic generation of interaction among the body, the mind, and the environment based on group experiences. To be precise, there are three basic dimensions in the construction of an embodied curriculum: (1) returning to the life-world and promoting the integration of the curriculum narrative framework and the student’s learning trajectory; (2) introducing the embodied experience and realizing the continuous interactions among the learner’s body, mind, and environment; and (3) integrating the functions of technologies to provide a powerful mediating tool for the construction of embodied curriculum.",,
-conversations about curriculum change: mathematical thinking and team-based learning in a discrete mathematics course,J. Paterson,"This article reports on the learning conversations between a mathematician and a mathematics educator as they worked together to change the delivery model of a third year discrete mathematics course from a traditional lecture mode to team-based learning (TBL). This change prompted the mathematician to create team tasks which increasingly focused on what he calls the ‘unspoken curriculum’: mathematical thinking. We consider the ways in which the TBL model promoted and enabled this in the light of literature on mathematical thinking, sense-making and behaviours, and strongly suggest that this approach warrants more attention from the mathematics teaching community. We also discuss shifts in the mathematician's thinking about task construction as he refined the tasks to encourage students to think and behave like mathematicians.",,
-journal of problem-based learning,Seung Wook Lee,"information, how they share it with others (or not), the way they learn together or apart, and how they support each other in ways that current classroom practices often prevent or discourage. is of which at to be hampered by ideas about what a curriculum should be, and how it should be managed. This is seen in over-managed and over- designed curricula, such as: be a The argument paper is that many PBL curricula focus on ‘tracing’ and ‘tracing other of PBL, what is needed is a shift imaginative curricula. It where PBL curricula what is getting in the way, how and PBL pedagogies might be delineated. Purpose: The purpose of this work was to describe and reflect on the worth of the iterative design of a valid teaching method for senior school level social studies classes reliant on PBL methods. Method: This descriptive reflection on the implementation of a Problem-Based Learning (PBL) activity involved a small elective class of ten Grade 11 World History students at Selwyn House School (SHS), in the Canadian Province of Quebec. Use of PBL was a five-step process, involving two different classes run concurrently, once with Cohort Group #1 (2017-18) in the spring of 2018, and twice with Cohort Group #2 (2018-19). Results: PBL processes evolved through three PBL examples involving twentieth-century historical situations, beginning with World War Two (1939-1945) and concluding with the Cold War (1946-1991). Design features and processes demonstrated enhanced outcomes from stu-dent-centered learning. Conclusion: The particular methods produced results that were evident in the students’ finished products. for ex-ample, found in this regard that a computer-based learning environment is ideal Case-based problem solving is a core approach to foster knowledge acquisition, especially during the learning process by which novices become more and more expert within a domain. This study investigated whether metacognitive scaffolding leads to better learning outcomes compared to learning without such support in a case-based learning environment. In addition, we examined the interaction between prior domain knowledge and prior metacognitive abilities and skills. Within a one-factorial design, we explored the role of metacognitive prompting during the learning phase. A pre- and post-test were applied assessing knowledge acquisition, metacognitive skills, and cognitive load. Results indicate no main effects for prompting, and no interaction effect between prior knowledge and prompting. Metacognitive prompting enables learners that already possess metacognitive abilities to activate these during problem-solving and, thus, score higher in the knowledge post-test than learners with low metacognitive abilities and no metacognitive prompting. to utilize the acquired knowledge effectively. Creation of new knowledge will arise from enquiry processes and, education should aim to develop this ability (Yoon, 2009). The purpose of contemporary education is to encourage self-directed learners with prob-Purpose: To i) use outcomes of professional development on PBL for curriculum renewal that leads to more active student learning ii) use academic and clinical peers to develop a learning package that integrates education and practice of Clinical Dental Hygienics and iii) present and evaluate the authentic cases used as stimulus material that is learner-centered. Methods: Professors of Dental Hygienics used professional development sessions to inform curriculum renewal and development of the stimulus material. Others with recent clinical experience used their expertise to assist with the development of learning packages. The PBL package was applied to two junior classes of Dental Hygienics and evaluated for perceptions of student satisfaction. Results: The journey through PBL Package development and implementation, using peer support through Professional Development, is reported on. Responses to statements from a survey on learning experience indicated there was a high level of student satisfaction with learning; after the application of PBL (91.4%) responded positively to the statement, “I actively attended this class”. Conclusions: It is necessary to develop more packages that reflect the field of Dental Hygienics and extend the use of PBL to verify its effect. Additional evaluation of the use PBL packages needs to focus on learning outcomes that demonstrate links to actual practice and students’ problem-solving, self-directed learning, and team-work abilities that can be cultivated through PBL methods. teachers out do the teachers work in but also the teachers do difference This paper addresses the current problems of STEAM education and suggests a solution to solve one of the problems through adopting a method from Flipped Learning. Through the meticulous literature review, comparison, and analysis on STEAM and Flipped Learning, it will be shown that the methodology of Flipped Learning complements the weaknesses in STEAM. The conclusion is that since there is a theoretical need for applying Flipped Learning to STEAM, further research needs to be carried out to apply flipped learning in STEAM classes.",,
-learning to be lawyers: professional identity and the law school curriculum,Charlotte S. Alexander,"The Carnegie Report faults American legal education for focusing exclusively on doctrine and analytical skills and neglecting the formation of professional identity. According to the Report, law schools can fix this problem by enabling students to ""encounter appealing representations of professional ideals, connect in a powerful way with engaging models of ethical commitment within in the profession, and reflect on their [own] emerging professional identity in relation to those ideals and models."" The Report identifies pro bono work, clinics, and externships as sites for this sort of learning, where students can interact with members of the profession and reflect on the models of professionalism that they encounter. Taking the Carnegie Report’s charge as a starting point, this article proposes an additional model for integrating a focus on professional identity into the law school curriculum. It profiles an experimental law school course that combined field work observations of practicing attorneys with in-class simulations of the work of a small law firm. The course was quite successful in prompting students to engage in an inquiry into what it is to be a lawyer and what kinds of lawyers they wanted to be. One student commented in a course evaluation, for example, that the course allowed him to see ""a new vision for what being a practicing lawyer can be."" That this sort of exposure to professional exemplars and reflection on professional identity was possible in a non-clinic course was an exciting discovery, suggesting new directions for curricular design as law schools continue to meet the challenges of the Carnegie Report.",,
-impacts of aiot implementation course on the learning outcomes of senior high school students,Chih-Cheng Tsai,"In this study, experimental teaching was conducted through the artificial intelligence of things (AIOT) practical course, and the 4D (discover, define, develop, deliver) double diamond shape was used to design the course and plan the teaching content to observe the students’ self-efficacy and learning anxiety. The technology acceptance model (TAM) concept was used to explore learning effectiveness and satisfaction. A total of 36 Senior One students from a senior high school in Kaohsiung were selected as the research subjects in two classes per week for 13 weeks. Questionnaires and qualitative interviews were used to understand the correlation between students’ self-efficacy, anxiety, outcomes, and learning satisfaction with the AIOT course during the flipped learning process. The study used SmartPLS3 to analyze the measurement model and structural model, and bootstrapping to conduct the path analysis and validate the research hypotheses. Because this research provides students with relevant online teaching videos for linking and browsing in the textbooks, and the video time should not be too long, in the teaching process, small-scale online learning courses are adopted for flipped teaching, which promptly arouses students’ interest and enhances their learning participation. Through the sharing of homework with each other, its content deficiencies were modified, students’ anxiety in learning was reduced, and the effect of learning and thinking together was achieved; in the teaching mode, theoretical content was combined with physical and practical courses to implement cross-disciplinary. To learn, the principle of 4D double diamond design to make “reciprocal corrections” between curriculum planning and teaching implementation as the teaching model framework was used. The results of this study show ”self-efficacy” has a significant positive effect on the “perceived usefulness” and “perceived ease of use” of users. “Learning anxiety” does not significantly predict the “perceived ease of use” or “perceived usefulness” of flipped learning using online e-learning. The “perceived ease of use” and “perceived usefulness” have positive impacts on the prediction of “behavioral intention” in flipped teaching using online digital teaching materials. The “perceived ease of use” has a positive and high impact on “perceived usefulness”. “Learning engagement” positively affects students’ actual “behavioral intention” towards learning. Students are highly receptive to learning new knowledge about science and technology.",,https://www.mdpi.com/2227-7102/11/2/82/pdf?version=1614245403
-disrupting colonial discourses in the geography curriculum during the introduction of british values policy in schools,C. Winter,"ABSTRACT The main purpose of this article is to expose and disrupt discourses dominating global development in an English school geography textbook chapter. The study was prompted by a teacher’s encounter with cultural difference in a geography lesson in South Korea. I investigate the issues raised through the lens of a new curriculum policy in English schools called ‘Promoting Fundamental British Values’ which forms part of England’s education-securitisation agenda, a topic of international attention. Following contextualization across research fields and in recent curriculum and assessment policy reform, I bring together theoretical perspectives from curriculum studies and Continental philosophy that do not usually speak to each other, to construct a new analytical approach. I identify three key themes, each informed by colonial logic: ‘development’, ‘numerical indicators’ and ‘learning to divide the world’. The inquiry appears to expose a tension between the knowledge of the textbook chapter and the purported aims of the British Values curriculum policy, but further investigation reveals the two to be connected through common colonial values. The findings are relevant to teachers, publishers, textbook authors, policy-makers and curriculum researchers. I recommend a refreshed curriculum agenda with the politics of knowledge and ethical global relations at its centre.",,https://www.tandfonline.com/doi/pdf/10.1080/00220272.2018.1428366?needAccess=true
-universal design for learning,J. Cressey,"Universal Design for Learning (UDL) is a framework for curriculum and instructional planning through which educators can maximize accessibility and minimize barriers that are often experienced by learners. Culturally responsive practices strengthen and complement UDL by framing accessibility as an equity goal and prompting educators to examine ableism, racism, and other structural inequities. Teacher educators are in a unique position to introduce UDL to future elementary teachers and support them in developing inclusive pedagogical methods early on in their careers. Education technology tools are used within UDL to make curriculum materials more accessible and engaging. In this chapter, the UDL framework will be described along with culturally responsive applications within elementary teacher education.",,
-integrating concept maps into a medical student oncology curriculum,S. Brondfield,,"{'model': 'tldr@v2.0.0', 'text': 'Concept mapping was feasible and reliable as an assessment tool and student perception was mixed, and students provided feedback to improve CM utility as a learning tool.'}",
-using the wider science curriculum to investigate fusion energy,Andrew Malcolm-Neale,"Fusion energy research is ‘mission oriented’ big-science. It has both rapid progress and a reputation for being forever distant that excites many with a curiosity for science. However, it is difficult to give time to in lesson plans given both its lack of associated learning objectives and complexity. Outreach at the Fusion Centre for Doctoral Training (FCDT) tries to make fusion accessible. We describe here a set of questions, which we use in workshops, that build a picture of fusion research. Each question integrates one or more aspects of the core science curriculum. We further share our experience of engagement with these questions and the rewarding experience pupils get on realising they can apply their knowledge to cutting-edge science. There is broad scope for using the worksheets. The available materials are a self-contained set of numerical questions that explore physics and fusion, providing all necessary equations and constants. The themes and physics of each sheet can be a springboard, though, for further exploration. Prompts for such discussion are suggested in this article for the first time, as they are not included in the available question set. Our questions, then, can be used to revise physics lessons in an interesting, unseen context; to explore the exciting world of fusion energy itself or as a workshop for the physics of energy. The full set of worksheets (one per question) is available upon request from outreach@fusion-cdt.ac.uk.",,
-exploring the k-12 computer science curriculum standards in the u.s,Meize Guo,"In order to create early exposure and to guide more talent into computing, computer science was included in K-12 education as a new subject area in many states. Although each state has different educational goals, computer science curriculum standards ensure equality among the states, especially in guiding teaching and accessing students' achievements. In this study, we applied content analysis to investigate the published computer science standards to examine: (1) How did computer science standards fit within the whole curriculum? (2) What concepts were covered in the computer science standards in the U.S.? The results showed four approaches to housing computer science in the whole curriculum. The Computer Science Teachers Association Computer Science framework was broadly adopted. Moreover, three major concepts, computer literacy, digital citizenship, and computational thinking, were additionally emphasized at the state standard level. This study fundamentally reported K-12 CS promotion in the U.S. through an academic standard scope, which directs K-12 CS teaching and learning. Additionally, this study prompted a conversation of preparing for different kinds of computing skills at the K-12 level in the U.S. among the policy makers and scholars","{'model': 'tldr@v2.0.0', 'text': 'The results showed four approaches to housing computer science in the whole curriculum, and three major concepts, computer literacy, digital citizenship, and computational thinking, were additionally emphasized at the state standard level.'}",
-"equality, diversity and inclusion: learning from laying our cards on the table (resource showcase)",Matthew Sillence,"In this showcase, we presented a card resource for learning developers who work with staff or students to set up an inclusive learning environment. These are often personal interactions and need to be explored respectfully (Love, Gaynor, and Blessett, 2016; Mahmood, Gray and Benincasa, 2022). The Equality, Diversity and Inclusion (EDI) cards were created by the University of Hertfordshire and comprise 26 topics and illustrations, ranging from ‘belonging’ to ‘trust’, and exist in both a physical and digital format. Resources used as games can encourage participation, reflection, and transformational thinking (Peabody and Noyes, 2017; Clark, Dickerson, and Jarvis, 2022), which are all important for encouraging respectful conversations about EDI issues. A collaborative project with the University of East Anglia to explore the use of the cards was funded by ALDinHE. Full ethics approval was gained from both universities before beginning. It recruited staff and students at Hertfordshire and East Anglia who worked with EDI issues and gathered reflections from facilitators and participants who used the cards in different learning contexts, such as academic support staff development exercises and student representation workshops. They reported that the card prompts allowed for personal storytelling and they suggested further topics for the cards set. This session, led by learning developers and teaching staff, explored the idea of the ‘brave space’ (Arao and Clemens, 2013, p. 142; Palfrey, 2017) and why ‘serious play’ is important in learning (Rieber, Smith and Noah, 1998, p. 29). It discussed where and when the cards can be used, in terms of: a physical or digital space; the curriculum or disciplinary context; and the stage of the learner’s journey. The EDI cards and new guidance booklet were made available for delegates to take away.",,https://journal.aldinhe.ac.uk/index.php/jldhe/article/download/1115/846
-critical reflexivity in indigenous and cross‐cultural psychology: a decolonial approach to curriculum?,Emily Castell,"Critical reflexivity is a mechanism for working toward decoloniality in higher education, with the potential to prompt students' to critique the contexts in which they are embedded, and facilitate transformative learning. We present a critical examination of the tensions surrounding decoloniality and critical reflexivity in an undergraduate unit on Indigenous and cross-cultural psychology at a large Australian university. We invited students in the unit to participate in a written reflexive exercise at the beginning (N = 44) and end of semester (N = 23) and analyzed these reflections qualitatively for level (four-category scheme for coding) and content (causal layered analysis) of reflection. Findings suggest that, while students' primarily demonstrate reflective engagement at levels preordinate to critical reflexivity, they are also engaged in active and nuanced processes of negotiating discomfort and uncertainty in this space. We pose critical commentary on the notion of safety in teaching practice, and consider the role of the academic institution in parametrizing the decolonial stance. This research holds application and transferability to higher education settings, and for the enduring project of engaging a decolonial approach to the curriculum within psychology.","{'model': 'tldr@v2.0.0', 'text': 'A critical examination of the tensions surrounding decoloniality and critical reflexivity in an undergraduate unit on Indigenous and cross-cultural psychology at a large Australian university suggests students are engaged in active and nuanced processes of negotiating discomfort and uncertainty in this space.'}",
-how does stem context-based learning work: what we know and what we still do not know,H. Sevian,"ABSTRACT Context-based learning (CBL) has influenced teaching and learning science in many countries over the past decades. Twelve years ago, a special issue on CBL was published in this Journal, focusing on CBL curriculum development. Seven papers in this current special issue on CBL now address the question of how a context influences the learning process. The papers focus on the stimulation of learning STEM subjects within contexts, how the learning process occurs and is enhanced, and the application of contexts in different settings. The approaches, results, and implications of the papers are located in a larger view that considers the question of what must be the case if a student not only engages in the tasks of learning but also succeeds at them. Concerning willingness and effort by learners, the papers draw conclusions about which STEM-related interests of students endure and are ephemeral across a decade, design criteria for maximising students’ situational interest, and students’ engagement with content and context simultaneously. Focusing on the opportunity to teach and learn, the papers reveal how a professional development approach functions to support STEM teachers to develop CBL materials, and how specific scaffolding acts in teaching bring students to more complex reasoning. Regarding good teaching, insights are offered on how metacognitive prompts improve teaching. Centring on the social surround that supports teaching and learning, a comparison of two contexts for teaching the same content reveals which aspects of the contexts move student learning forward. From this mapping, paths toward future research are projected.",,https://figshare.com/articles/journal_contribution/How_does_STEM_context-based_learning_work_what_we_know_and_what_we_still_do_not_know/6803186/1/files/12372731.pdf
-simulations and virtual learning supporting clinical education during the covid 19 pandemic,S. Tabatabai,"Abstract Currently, the COVID-19 pandemic has a critical impact on clinical education, and it has resulted in the widespread disruption of clinical assessment. Clinical mentors and students in all the health professions are working within the most troublesome of circumstances in the hospital-based educational settings. Medical educationists ought to concentrate on the health and the safety of the students and communities. The safety issues have prompted the Ministry of Health to suggest that schools develop action plans for the adoption of available technologies to keep medical education moving forward with high quality, active, and interactive learning for more demanding tomorrow. A key challenge for medical educators is to simulate the clinical encounters at this unprecedented time, and this emphasized the necessity of applying virtual simulation-based educational tools in clinical education. This commentary explores how COVID-19 has challenged medical education. It also has discussed the future implications and potencial challenges of incorporating simulation-based virtual learning technologies into the medical curriculum, for the future of clinical education, and students' or residents' competency evaluation.","{'model': 'tldr@v2.0.0', 'text': ""This commentary explores how COVID-19 has challenged medical education and the future implications and potencial challenges of incorporating simulation-based virtual learning technologies into the medical curriculum, for the future of clinical education, and students' or residents' competency evaluation.""}",https://www.dovepress.com/getfile.php?fileID=60342
-becoming doctors: examining student narratives to understand the process of professional identity formation within a learning community,D. Hatem,"Background: Professional identity formation is a key aim of medical education, yet empiric data on how this forms are limited. Methods: Our study is a qualitative analysis of student reflections written during the final session of our Becoming a Physician curriculum. After reading their medical school admission essay and their class oath, students wrote about a “time, or times during your third year when you felt like a doctor.” The reflections were qualitatively analyzed by the evaluation team, looking for themes found in the reflections. Results: Narrative themes separated into 4 distinct categories, specifically that performing physician tasks can make one feel like a doctor, demonstrating caring is a fundamental task of doctors, integrating personal ideals with professional values promotes professional identity formation, and the theme of never feeling like a doctor. Subsets of these broad categories provide further insight into individual and integrative tasks. Patients, patient families, and students through their own reflection prompted learners to feel like doctors in 74% of narratives, whereas physicians or the care team did so in 26% of our narratives. Conclusion: Students are able to reflect on times during their principal clinical year where they feel like doctors, taking a step toward forming a professional identity. Having faculty prompt and support such reflection can help faculty understand the student experience of their principal clinical year and promote professional identity formation.","{'model': 'tldr@v2.0.0', 'text': 'Students are able to reflect on times during their principal clinical year where they feel like doctors, taking a step toward forming a professional identity.'}",https://journals.sagepub.com/doi/pdf/10.1177/2382120519834546
-examining student work for evidence of teacher uptake of educative curriculum materials,A. Bismack,"The purpose of this study was to identify evidence in student work of teachers' uptake of educative features in educative curriculum materials. These are features in curriculum materials designed with the specific intent of supporting teacher learning and enactment. This study was prompted by previous work on educative curriculum materials and the need to determine how teachers' use of educative curriculum materials can influence student learning. Student work from two fourth-grade teachers' enactment of an electric circuits unit was analyzed for evidence of teachers' uptake of educative features, which included characteristics of quality for particular science practices. Findings from the student work revealed that the teachers used many of the supports in the educative curriculum materials, especially those that could be used directly with students. The student work also reflected characteristics of high-quality science practices, which were only supported within the educative features. This study supports and extends other work related to how teachers' use of educative curriculum materials may influence student learning and has implications for supporting teachers' productive engagement in teaching that supports the integration of science content and scientific practices, as emphasized in current reform efforts. © 2015 Wiley Periodicals, Inc. J Res Sci Teach 52: 816–846, 2015.",,http://deepblue.lib.umich.edu/bitstream/2027.42/112188/2/tea21220-sup-0001-SuppFig_S2.pdf
-"an integrated, multimodal resident curriculum in patient safety and quality improvement",J. Werner,"Introduction Patient safety and quality improvement are essential components of modern medicine. The traditional model of graduate medical education does not lend itself well to learning these disciplines. This curriculum encompasses these disciplines across multiple modalities and extends throughout residency. Methods The curriculum includes introductory presentations suitable for naive audiences. Following these is a structured rotation that provides the opportunity both to experience in-depth self-directed learning and to practice skills involved in quality and safety. This rotation includes existing online courses published elsewhere, reflective writing exercises based on self-directed learning, and practice cases. Finally, residents lead a morbidity, mortality, and improvement conference where adverse events are identified and reviewed, specific interventions and outcome objectives are selected, and action teams are identified. Results After two presentations on system issues and individual issues, responses to the prompt “This talk will aid in my professional development” were 4.75 and 4.59 out of 5, respectively. Eighty-three percent of residents agreed they had a better understanding of the concepts of patient safety and/or quality improvement than they did before the rotation. Audience members for the resident-led morbidity, mortality, and improvement conference agreed it would lead to a change in their own practice. Discussion The contents of this longitudinal curriculum have been incorporated into the core requirements of our general pediatrics residency program and could reasonably be imported into any residency requiring a robust longitudinal experience in quality improvement and patient safety.","{'model': 'tldr@v2.0.0', 'text': 'The contents of this longitudinal curriculum have been incorporated into the core requirements of the general pediatrics residency program and could reasonably be imported into any residency requiring a robust longitudinal experience in quality improvement and patient safety.'}",
-integrating curriculum-based dynamic assessment in computerized adaptive testing: development and predictive validity of the edpl-bai battery on reading competence,J. Navarro,"In recent decades there have been significant changes in the conceptualization of reading as well as in the perception of how this activity should be assessed. Interest in the analysis of reading processes has led to the emergence of new explanatory models based primarily on the contributions of cognitive psychology. In parallel, there have been notable advances in measurement procedures, especially in models based on Item Response Theory (IRT), as well as in the capacity and performance of specific software programs that allow data to be managed and analyzed. These changes have contributed significantly to the rise of testing procedures such as computerized adaptive tests (CATs), whose fundamental characteristic is that the sequence of items presented in the tests is adapted to the level of competence that the subject manifests. Likewise, the incorporation of elements of dynamic assessment (DA) as the prompts are gradually offered allows for obtaining information about the type and degree of support required to optimize the subject’s performance. In this sense, the confluence of contributions from DA and CATs offers a new possibility for approaching the assessment of learning processes. In this article, we present a longitudinal research developed in two phases, through which a computerized dynamic adaptive assessment battery of reading processes (EDPL-BAI) was configured. The research frame involved 1,831 students (46% girls) from 13 public schools in three regions of Chile. The purpose of this study was to analyze the differential contribution on reading competence of dynamic scores obtained in a subsample composed of 324 (47% girls) students from third to sixth grade after the implementation of a set of adaptive dynamic tests of morpho-syntactic processes. The results achieved in the structural equation modeling indicate a good global fit. Individual relationships show a significant contribution of calibrated score that reflects estimated knowledge level on reading competence, as well as dynamic scores based on the assigned value of graduated prompts required by the students. These results showed significant predictive values on reading competence and incremental validity in relation to predictions made by static criterion tests.","{'model': 'tldr@v2.0.0', 'text': 'This article presents a longitudinal research developed in two phases, through which a computerized dynamic adaptive assessment battery of reading processes (EDPL-BAI) was configured, and shows significant predictive values on reading competence and incremental validity in relation to predictions made by static criterion tests.'}",
-enhance student learning with powerpoint games: using twenty questions to promote active learning in managerial accounting,Victoria A. Fratto,"Stakeholders encourage accounting educators to provide active learning opportunities, to integrate the creative use of technology into the curriculum, and to emphasize learning by doing. The principles of good teaching practice can use technology to promote active learning, to provide prompt feedback to students, to increase student time on task, and to make learning more effective and efficient for the student. Technological tools can permit students to become active participants and can improve student learning by giving students convenient access to review material with immediate feedback. This article describes the use and development of a PowerPoint game in an introductory accounting course managerial accounting that provides the student with immediate feedback and is designed to be accessed by the student outside of the classroom. This technological tool can be used in other undergraduate academic disciplines.","{'model': 'tldr@v2.0.0', 'text': 'This article describes the use and development of a PowerPoint game in an introductory accounting course managerial accounting that provides the student with immediate feedback and is designed to be accessed by the student outside of the classroom.'}",
-narrative medicine: a writing workshop curriculum for residents,A. Winkel,"Introduction Reflection is a critical part of the learning cycle. Narrative medicine has been shown to help physicians-in-training develop both empathy and professional identity. A narrative medicine curriculum focused on the experience of the physician and challenging patient experiences creates opportunities in which to process complicated aspects of the experience of both patient and physician with other members of the training community. Methods Fifteen 1-hour small-group reflective writing workshops comprise a 2-year narrative medicine curriculum. Each workshop uses selected literature to focus a discussion and a prompt for written reflection and can be integrated into a didactic curriculum. Teacher guides have been created to help untrained preceptors lead small-group sessions. Feedback forms are distributed to participating residents. Results Out of 29 total residents enrolled at one program, 23 residents completed feedback forms, and 16 (69%) reported that the reflective writing sessions were relevant to their work as obstetrician-gynecologists. Residents stated the best parts of the course were reading and writing (n = 6; 27%), sharing writings with colleagues (n = 5; 21%), and having positive experiences with members of their community (n = 6; 27%). Some residents reported difficulty sharing their private reflections (n = 4; 17%). Discussion A narrative medicine curriculum is a powerful tool for promoting reflection about the challenging work of training in obstetrics and gynecology and other specialties. Reflective writing workshops have been found to be acceptable to obstetrics and gynecology residents, and the curriculum has been successfully implemented at several training programs.","{'model': 'tldr@v2.0.0', 'text': 'Reflective writing workshops have been found to be acceptable to obstetrics and gynecology residents, and the curriculum has been successfully implemented at several training programs.'}",
-jelly bean world: a testbed for never-ending learning,Emmanouil Antonios Platanios,"Machine learning has shown growing success in recent years. However, current machine learning systems are highly specialized, trained for particular problems or domains, and typically on a single narrow dataset. Human learning, on the other hand, is highly general and adaptable. Never-ending learning is a machine learning paradigm that aims to bridge this gap, with the goal of encouraging researchers to design machine learning systems that can learn to perform a wider variety of inter-related tasks in more complex environments. To date, there is no environment or testbed to facilitate the development and evaluation of never-ending learning systems. To this end, we propose the Jelly Bean World testbed. The Jelly Bean World allows experimentation over two-dimensional grid worlds which are filled with items and in which agents can navigate. This testbed provides environments that are sufficiently complex and where more generally intelligent algorithms ought to perform better than current state-of-the-art reinforcement learning approaches. It does so by producing non-stationary environments and facilitating experimentation with multi-task, multi-agent, multi-modal, and curriculum learning settings. We hope that this new freely-available software will prompt new research and interest in the development and evaluation of never-ending learning systems and more broadly, general intelligence systems.","{'model': 'tldr@v2.0.0', 'text': 'The Jelly Bean World testbed is proposed, a freely-available software that provides environments that are sufficiently complex and where more generally intelligent algorithms ought to perform better than current state-of-the-art reinforcement learning approaches.'}",
-development of the self-determined learning model of instruction coaching model: implications for research and practice.,Mayumi Hagiwara,"The Self-Determined Learning Model of Instruction (SDLMI) is an evidence-based model of instruction implemented by a facilitator (e.g., general or special education teacher, family member, related service professional, transition professional) to enable students to learn to self-regulate problem solving in service to a goal. Students learn how to select personally relevant goals, develop action plans for achieving those goals, and self-monitor and self-evaluate progress toward achieving those goals. To scale-up the implementation of the SDLMI with fidelity by facilitators, there is a need for a systematic coaching model. Researchers have established coaching as a critical aspect of the implementation of evidence-based practices at scale and as a part of ongoing professional development. The purpose of this article is to describe steps taken to develop a systematic coaching model to support the implementation of the SDLMI. We will describe components of the SDLMI Coaching Model and the process through which they were derived from the coaching literature and data gathered during SDLMI research studies. We will conclude with implications for the implementation of the SDLMI Coaching Model at scale while maintaining a focus on individual students’ support needs. Causal Agency Theory provides an empirically-validated framework for conceptualizing the development of self-determination across the lifespan. Causal Agency Theory defines self-determination as a “dispositional characteristic manifested as acting as the causal agent in one’s life. Self-determined people (i.e., causal agents) act in service to freely chosen goals” (Shogren, Wehmeyer, Palmer, Forber-Pratt et al., 2015, p. 258). Self-determination develops across the lifespan as children and youth have opportunities to build skills and attitudes associated with self-determined actions including choice-making, decisionmaking, problem solving, goal setting and attainment, planning, self-management, self-advocacy, self-awareness, and self-knowledge (Shogren, Wehmeyer, Palmer, Forber-Pratt et al., 2015). The positive impact of promoting self-determination on in-school (e.g., academic achievement, goal attainment; Shogren et al., 2012) and post-school (e.g., employment, community participation; Shogren, Wehmeyer, Palmer, Rifenbark, & Little, 2015) outcomes has been confirmed by multiple research studies (Hagiwara et al., 2017). As such, in both general and special education, there is a growing emphasis on promoting the skills associated with self-determination in inclusive environments aligned with efforts to promote social-emotional learning and college and career readiness (Shogren et al., 2016). Researchers have established teachers can embed instruction in skills associated with self-determination into academic and transition-related content, leading to enhanced selfdetermination and goal attainment (Powers et al., 2012; Test et al., 2009). However, teachers have also identified the need for further support for implementation, particularly as reThis research was supported in part by a grant from the Institute of Education Sciences, National Center for Special Education Research, Grant Award R324A170008. Correspondence concerning this article should be addressed to Mayumi Hagiwara, Kansas University Center on Developmental Disabilities, 3134 Haworth Hall, 1200 Sunnyside Avenue, Lawrence, KS 66045. Email: mayuhagi@ku.edu Education and Training in Autism and Developmental Disabilities, 2020, 55(1), 17–27 © Division on Autism and Developmental Disabilities SDLMI Coaching Model / 17 search project-based supports fade. This lead to the activities undertaken to create the SDLMI Coaching Model. Self-Determined Learning Model of Instruction The Self-Determined Learning Model of Instruction (SDLMI; Shogren et al., 2019) is an evidencebased practice (National Technical Assistance Center on Transition, 2017) designed to embed instruction on self-determination in ongoing curricular activities. The SDLMI is designed to be implemented by a facilitator (e.g., general or special education teachers, career counselors, transition specialists). Facilitators enable students to learn and practice skills associated with self-determination, by providing explicit instruction (e.g., how to make choices, set goals, plan) and embedding opportunities to practice those skills in ongoing instruction to ultimately teach students to self-regulate problem solving in service to a goal. As such, general or special education teachers can use the SDLMI to shape their instruction and teach students how to set and attain goals that are linked to the general education curriculum or students’ individualized learning plans, including transition planning. The SDLMI can be implemented with students with and without disabilities, in whole-class, small group, or one-to-one settings. Essentially, the SDLMI was designed as a model of instruction that could be overlaid on any curricular area to support teachers in enabling students to set and work toward goals in that curricular area (Shogren et al., 2019). The SDLMI consists of a three-phase instructional process repeated over time to enable students to work on setting and attaining goals that build on and enhance each other (see Figure 1). Each of the three instructional phases includes four Student Questions that guide students through the problem-solving steps needed to solve the overarching question of each phase (Phase 1: What is my goal?, Phase 2: What is my plan?, Phase 3: What have I learned?). The solution to the problem in each phase leads to the problem-solving sequence in the Student Questions in the next phase. Importantly, each Student Question is linked to a set of Teacher Objectives that provide teachers or other facilitators with a road map for how they can organize instruction to enable students to answer each Student Question. The Teacher Objectives drive teachers’ instructional practices and actions and are linked to Educational Supports, which are strategies teachers can use to meet the specified objective based on students’ individualized learning needs. Teachers deliver targeted instruction on these strategies to support students in answering the Student Questions. Additionally, teachers integrate the identified goal and action plan into ongoing instructional activities by embedding opportunities for students to apply learned skills in service to their goals. For example, a student might set a goal to ensure they have provided the format of a solution in math class (e.g., most simplified version) by underlining the directions on homework, quizzes and tests. To support the student in achieving their goal and implementing their action plan, the teacher can refer to this goal during core content instruction as well as deliver targeted instruction enFigure 1. The phases of the SDLMI. © 2017 – Kansas University Center on Developmental Disabilities, Lawrence, KS USA. 18 / Education and Training in Autism and Developmental Disabilities-March 2020 abling the student to develop a system that prompts them to underline the directions before each beginning each assignment. In a recent, systematic literature review of the SDLMI studies, Hagiwara and colleagues (2017) found that all 21 SDLMI studies produced positive students outcomes (e.g., enhanced self-determination, increased access to the general education curriculum, improved classroom behavior, student-directed transition planning). However, only eight out of the studies identified in the literature that were implemented by teachers and/or other facilitators reported the characteristics of initial training and ongoing technical assistance activities to support the implementation of the SDLMI. For example, Shogren et al. (2012) and Wehmeyer et al. (2012) reported that teachers were trained for 0.5 to 1.5 days prior to implementation and supported through ongoing email support. However, no ongoing, systematic training, professional development, or coaching were delivered. Cho et al., (2011) found that teachers reported a lack of formal training in interventions to promote self-determination as one of the most significant barriers to implementing instruction to enhance self-determination in general and special education. In a recent statewide SDLMI implementation (Shogren et al., 2018), efforts were undertaken to provide teachers with more systematic implementation supports, including coaching and yearly professional development. While coaching played a major role in supporting the teachers to implement the SDLMI with fidelity in this project, the coaching was delivered relatively informally and varied across participating districts when evaluated from the lens of implementation science (Burke et al., 2019). Therefore, researchers recommended establishing a systematic coaching model to ensure that high-quality, consistent coaching is delivered across coaches and their assigned facilitators. Furthermore, researchers suggested improving the coaching system by: (a) asking teachers for feedback on their coaching experiences and (b) establishing a strong, integrated fidelity assessment system to addresses teacher practices in the classroom and coaching practices to ensure high quality implementation (Burke et al., 2019). Development of the SDLMI Coaching Model The importance of coaching in promoting fidelity of implementation of an effective innovation has been widely acknowledged (Fixsen et al., 2005). Coaching also plays a key role in ensuring facilitators of an intervention to have the competencies to implement evidence-based practices as intended over time (Snyder et al., 2015). Additionally, researchers have identified coaching as an effective way to provide face-to-face modeling and guidance on the use of interventions (Fixsen et al., 2010). For teachers and other facilitators to implement the SDLMI with fidelity, best practice recommends the importance of sustained and systematic supports thr",,
-undergraduate recognition of curriculum-related skill development and the skills employers are seeking,Michelle A. Hill,"Employers of chemistry graduates are seeking a range of transferable skills from prospective employees, and academics are increasingly seeking to build employability skill development opportunities into the undergraduate curriculum. However, research suggests that undergraduates do not recognise or value such skill development without prompting. This recognition is essential if graduates are to be able to articulate their skills in the employment process. This study involves research amongst almost 1000 undergraduates studying chemistry at two institutions, using open-ended questions to collect qualitative data. The extent to which students recognised course-related skills development and understood the skills that employers are looking for was investigated, as was their desire to develop additional skills. Similarities and differences in student views between institutions are discussed, as well as trends across year levels and by gender. Results indicate that undergraduates studying chemistry are most likely to value and recognise development of some key skills sought by employers (teamwork, communication, thinking/problem solving, organisation/time management and laboratory/practical skills), but are very unlikely to value or recognise others (numeracy, independent learning, commercial awareness, interpersonal, research, computer/IT, creativity/innovation, flexibility/adaptability and initiative). Opportunities to develop the latter skills and recognition of the value of doing so will require improved communication with students and/or provision of new experiences within the curriculum.",,
-students’ sense-making of personalised feedback based on learning analytics,Lisa-Angelique Lim,"Although technological advances have brought about new opportunities for scaling feedback to students, there remain challenges in how such feedback is presented and interpreted. There is a need to better understand how students make sense of such feedback to adapt self-regulated learning processes. This study examined students’ sense-making of learning analytics–based personalised feedback across four courses. Results from a combination of thematic analysis and epistemic network analysis show an association between student perceptions of their personalised feedback and how these map to subsequent self-described self-regulated learning processes. Most notably, the results indicate that personalised feedback, elaborated by personal messages from course instructors, helps students refine or strengthen important forethought processes of goal-setting, as well as to reduce procrastination. The results highlight the need for instructors to increase the dialogic element in personalised feedback in order to reduce defensive reactions from students who hold to their own learning strategies. This approach may prompt reflection on the suitability of students’ current learning strategies and achievement of associated learning goals. Implications for practice or policy: Personalised feedback based on learning analytics should be informed by an understanding of students’ self-regulated learning. Instructors implementing personalised feedback should align this closely with the course curriculum. Instructors implementing personalised feedback in their courses should consider the relational element of feedback by using a positive tone. Personalised feedback can be further enhanced by increasing the dialogic element and by including more information about learning strategies. ",,https://ajet.org.au/index.php/AJET/article/download/6370/1696
-which prompts make the difference? data prioritization for efficient human llm evaluation,M. Boubdir,"Human evaluation is increasingly critical for assessing large language models, capturing linguistic nuances, and reflecting user preferences more accurately than traditional automated metrics. However, the resource-intensive nature of this type of annotation process poses significant challenges. The key question driving our work:""is it feasible to minimize human-in-the-loop feedback by prioritizing data instances which most effectively distinguish between models?""We evaluate several metric-based methods and find that these metrics enhance the efficiency of human evaluations by minimizing the number of required annotations, thus saving time and cost, while ensuring a robust performance evaluation. We show that our method is effective across widely used model families, reducing instances of indecisive (or""tie"") outcomes by up to 54% compared to a random sample when focusing on the top-20 percentile of prioritized instances. This potential reduction in required human effort positions our approach as a valuable strategy in future large language model evaluations.","{'model': 'tldr@v2.0.0', 'text': 'This work shows that the key question driving the work:""is it feasible to minimize human-in-the-loop feedback by prioritizing data instances which most effectively distinguish between models"" is feasible, and finds that several metric-based methods enhance the efficiency of human evaluations by minimizing the number of required annotations.'}",
-"conversational challenges in ai-powered data science: obstacles, needs, and design opportunities",Bhavya Chopra,"Large Language Models (LLMs) are being increasingly employed in data science for tasks like data preprocessing and analytics. However, data scientists encounter substantial obstacles when conversing with LLM-powered chatbots and acting on their suggestions and answers. We conducted a mixed-methods study, including contextual observations, semi-structured interviews (n=14), and a survey (n=114), to identify these challenges. Our findings highlight key issues faced by data scientists, including contextual data retrieval, formulating prompts for complex tasks, adapting generated code to local environments, and refining prompts iteratively. Based on these insights, we propose actionable design recommendations, such as data brushing to support context selection, and inquisitive feedback loops to improve communications with AI-based assistants in data-science tools.","{'model': 'tldr@v2.0.0', 'text': 'These findings highlight key issues faced by data scientists, including contextual data retrieval, formulating prompts for complex tasks, adapting generated code to local environments, and refining prompts iteratively, and propose actionable design recommendations, such as data brushing to support context selection.'}",
-cona: a novel context-aware instruction paradigm for communication using large language model,Nan Zhou,"We introduce CONA, a novel context-aware instruction paradigm for effective knowledge dissemination using generative pre-trained transformer (GPT) models. CONA is a flexible framework designed to leverage the capabilities of Large Language Models (LLMs) and incorporate DIKW (Data, Information, Knowledge, Wisdom) hierarchy to automatically instruct and optimise presentation content, anticipate potential audience inquiries, and provide context-aware answers that adaptive to the knowledge level of the audience group. The unique aspect of the CONA paradigm lies in its combination of an independent advisory mechanism and a recursive feedback loop rooted on the DIKW hierarchy. This synergy significantly enhances context-aware contents, ensuring they are accessible and easily comprehended by the audience. This paradigm is an early pioneer to explore new methods for knowledge dissemination and communication in the LLM era, offering effective support for everyday knowledge sharing scenarios. We conduct experiments on a range of audience roles, along with materials from various disciplines using GPT4. Both quantitative and qualitative results demonstrated that the proposed CONA paradigm achieved remarkable performance compared to the outputs guided by conventional prompt engineering.","{'model': 'tldr@v2.0.0', 'text': 'CONA is a novel context-aware instruction paradigm for effective knowledge dissemination using generative pre-trained transformer (GPT) models that combines an independent advisory mechanism and a recursive feedback loop rooted on the DIKW hierarchy.'}",http://arxiv.org/pdf/2305.18620
-exploring the design space of ai based code completion engines,Parth Thakkar,"Artificial Intelligence (AI) based code completion tools such as Github Copilot have recently gained tremendous popularity due to their ability to suggest arbitrary length snippets, improving developer productivity dramatically. However, there is little public understanding of what it takes to build such a tool. In this thesis, we explore the design space of building such a tool. We study the importance of the two key components of such a tool: the Large Language Model (LLM) that predicts the suggestions, and the system around it that feeds it the right context and filters out the bad suggestions. We start by exploring the design of Github Copilot to understand the state of the art, and describe the three key components of Copilot: Prompt Engineering, Model Invocation and Feedback loop. We then study the various factors that affect the quality of the suggestions generated by the LLM. We study both (a) the impact of the context fed to the LLM, and (b) the impact of the LLM itself. For the former, we study the impact including context from other files and code after the cursor along with different methods of context collection and amount of collected context. For the latter, we study the impact of the size of the LLM and the training procedure. Apart from factors affecting the quality of suggestions, we also study the factors affecting the latency of such code completion engines, as low latency is critical for building good code completion engines. We find that the context fed to the model makes a significant difference in the quality of generated suggestions, and good context collection can improve the quality of suggestions by 11-26% points (20-113% relative improvement) on the exact match metric for one line suggestions. Models that can exploit the context after the cursor can further improve the quality of suggestions by 6-14% points (12-16% relative improvement). Our experiments show that increasing the prompt length beyond a point does not improve suggestion quality significantly, and that 2048-4096 tokens are sufficient. We also find that the size of the LLM has a much smaller impact on the quality of suggestions than other factors such as the context fed to the model and the training procedure used. For example, we found that the SantaCoder model (1.1B parameters) provided better suggestions than the 16B CodeGen-Multi","{'model': 'tldr@v2.0.0', 'text': 'It is found that the context fed to the model makes a significant difference in thequality of generated suggestions, and good context collection can improve the quality of suggestions by 11-26% points (20-113% relative improvement) on the exact match metric for one line suggestions.'}",
-incremental learning of humanoid robot behavior from natural interaction and large language models,Leonard Bärmann,"Natural-language dialog is key for intuitive human-robot interaction. It can be used not only to express humans' intents, but also to communicate instructions for improvement if a robot does not understand a command correctly. Of great importance is to endow robots with the ability to learn from such interaction experience in an incremental way to allow them to improve their behaviors or avoid mistakes in the future. In this paper, we propose a system to achieve incremental learning of complex behavior from natural interaction, and demonstrate its implementation on a humanoid robot. Building on recent advances, we present a system that deploys Large Language Models (LLMs) for high-level orchestration of the robot's behavior, based on the idea of enabling the LLM to generate Python statements in an interactive console to invoke both robot perception and action. The interaction loop is closed by feeding back human instructions, environment observations, and execution results to the LLM, thus informing the generation of the next statement. Specifically, we introduce incremental prompt learning, which enables the system to interactively learn from its mistakes. For that purpose, the LLM can call another LLM responsible for code-level improvements of the current interaction based on human feedback. The improved interaction is then saved in the robot's memory, and thus retrieved on similar requests. We integrate the system in the robot cognitive architecture of the humanoid robot ARMAR-6 and evaluate our methods both quantitatively (in simulation) and qualitatively (in simulation and real-world) by demonstrating generalized incrementally-learned knowledge.","{'model': 'tldr@v2.0.0', 'text': ""A system that deploys Large Language Models for high-level orchestration of the robot's behavior based on the idea of enabling the LLM to generate Python statements in an interactive console to invoke both robot perception and action is presented.""}",https://arxiv.org/pdf/2309.04316
-real: resilience and adaptation using large language models on autonomous aerial robots,Andrea Tagliabue,"Large Language Models (LLMs) pre-trained on internet-scale datasets have shown impressive capabilities in code understanding, synthesis, and general purpose question-and-answering. Key to their performance is the substantial prior knowledge acquired during training and their ability to reason over extended sequences of symbols, often presented in natural language. In this work, we aim to harness the extensive long-term reasoning, natural language comprehension, and the available prior knowledge of LLMs for increased resilience and adaptation in autonomous mobile robots. We introduce REAL, an approach for REsilience and Adaptation using LLMs. REAL provides a strategy to employ LLMs as a part of the mission planning and control framework of an autonomous robot. The LLM employed by REAL provides (i) a source of prior knowledge to increase resilience for challenging scenarios that the system had not been explicitly designed for; (ii) a way to interpret natural-language and other log/diagnostic information available in the autonomy stack, for mission planning; (iii) a way to adapt the control inputs using minimal user-provided prior knowledge about the dynamics/kinematics of the robot. We integrate REAL in the autonomy stack of a real multirotor, querying onboard an offboard LLM at 0.1-1.0 Hz as part the robot's mission planning and control feedback loops. We demonstrate in real-world experiments the ability of the LLM to reduce the position tracking errors of a multirotor under the presence of (i) errors in the parameters of the controller and (ii) unmodeled dynamics. We also show (iii) decision making to avoid potentially dangerous scenarios (e.g., robot oscillates) that had not been explicitly accounted for in the initial prompt design.","{'model': 'tldr@v2.0.0', 'text': 'The ability of the LLM to reduce the position tracking errors of a multirotor under the presence of (i) errors in the parameters of the controller and (ii) unmodeled dynamics is demonstrated.'}",
-improving summarization with human edits,Zonghai Yao,"Recent work has shown the promise of learning with human feedback paradigms to produce human-determined high-quality text. Existing works use human feedback to train large language models (LLMs) in general domain abstractive summarization and have obtained summary quality exceeding traditional likelihood training. In this paper, we focus on a less explored form of human feedback -- Human Edits. We propose Sequence Alignment (un)Likelihood Training (SALT), a novel technique to use both the human-edited and model-generated data together in the training loop. In addition, we demonstrate simulating Human Edits with ground truth summaries coming from existing training data -- Imitation edits, along with the model-generated summaries obtained after the training, to reduce the need for expensive human-edit data. In our experiments, we extend human feedback exploration from general domain summarization to medical domain summarization. Our results demonstrate the effectiveness of SALT in improving the summary quality with Human and Imitation Edits. Through additional experiments, we show that SALT outperforms the conventional RLHF method (designed for human preferences) -- DPO, when applied to human-edit data. We hope the evidence in our paper prompts researchers to explore, collect, and better use different human feedback approaches scalably.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Sequence Alignment (un)Likelihood Training (SALT), a novel technique to use both the human-edited and model-generated data together in the training loop, and demonstrates the effectiveness of SALT in improving the summary quality with Human and Imitation Edits.'}",https://arxiv.org/pdf/2310.05857
-"double-loop learning control (dlc) model for reengineering: a ""yin"" and ""yang"" balanced approach for effective organizational change",Y. Hua,"The advent of information technology has prompted organizations to reinvent themselves in response to the rising demands of customers and expectations of employees. A conceptual model that shows how organizations can effectively manage the change process is essential in view of the multitude of problems encountered in reengineering endeavours. The double-loop learning control (DLC) model advocates the importance of having a shared vision and an environment that facilitates double-loop learning. The model uses the oriental philosophical 'yin-yang' balanced concept to illustrate the need to balance the application of technology with the human factor. The DLC model was empirically tested by a public sector survey on success factors for change, conducted in 1997. Analyses confirmed that having a shared vision is paramount to reengineering success and an organizational ability to learn as a mediator of the yin-yang contributions toward building this shared vision for organizational change.",,
-a parameter-efficient learning approach to arabic dialect identification with pre-trained general-purpose speech model,S. Radhakrishnan,"In this work, we explore Parameter-Efficient-Learning (PEL) techniques to repurpose a General-Purpose-Speech (GSM) model for Arabic dialect identification (ADI). Specifically, we investigate different setups to incorporate trainable features into a multi-layer encoder-decoder GSM formulation under frozen pre-trained settings. Our architecture includes residual adapter and model reprogramming (input-prompting). We design a token-level label mapping to condition the GSM for Arabic Dialect Identification (ADI). This is challenging due to the high variation in vocabulary and pronunciation among the numerous regional dialects. We achieve new state-of-the-art accuracy on the ADI-17 dataset by vanilla fine-tuning. We further reduce the training budgets with the PEL method, which performs within 1.86% accuracy to fine-tuning using only 2.5% of (extra) network trainable parameters. Our study demonstrates how to identify Arabic dialects using a small dataset and limited computation with open source code and pre-trained models.","{'model': 'tldr@v2.0.0', 'text': 'This work explores Parameter-Efficient-Learning techniques to repurpose a General-Purpose-Speech model for Arabic dialect identification (ADI) and achieves new state-of-the-art accuracy on the ADI-17 dataset by vanilla fine-tuning.'}",https://repository.kaust.edu.sa/bitstream/10754/694670/1/2305.11244.pdf
-tree-planner: efficient close-loop task planning with large language models,Mengkang Hu,"This paper studies close-loop task planning, which refers to the process of generating a sequence of skills (a plan) to accomplish a specific goal while adapting the plan based on real-time observations. Recently, prompting Large Language Models (LLMs) to generate actions iteratively has become a prevalent paradigm due to its superior performance and user-friendliness. However, this paradigm is plagued by two inefficiencies: high token consumption and redundant error correction, both of which hinder its scalability for large-scale testing and applications. To address these issues, we propose Tree-Planner, which reframes task planning with LLMs into three distinct phases: plan sampling, action tree construction, and grounded deciding. Tree-Planner starts by using an LLM to sample a set of potential plans before execution, followed by the aggregation of them to form an action tree. Finally, the LLM performs a top-down decision-making process on the tree, taking into account real-time environmental information. Experiments show that Tree-Planner achieves state-of-the-art performance while maintaining high efficiency. By decomposing LLM queries into a single plan-sampling call and multiple grounded-deciding calls, a considerable part of the prompt are less likely to be repeatedly consumed. As a result, token consumption is reduced by 92.2% compared to the previously best-performing model. Additionally, by enabling backtracking on the action tree as needed, the correction process becomes more flexible, leading to a 40.5% decrease in error corrections. Project page: https://tree-planner.github.io/","{'model': 'tldr@v2.0.0', 'text': 'Tree-Planner is proposed, which reframes task planning with Large Language Models into three distinct phases: plan sampling, action tree construction, and grounded deciding, which achieves state-of-the-art performance while maintaining high efficiency.'}",https://arxiv.org/pdf/2310.08582
-batchprompt: accomplish more with less,Jianzhe Lin,"As the ever-increasing token limits of large language models (LLMs) have enabled long context as input, prompting with single data samples might no longer an efficient way. A straightforward strategy improving efficiency is to batch data within the token limit (e.g., 8k for gpt-3.5-turbo; 32k for GPT-4), which we call BatchPrompt. We have two initial observations for prompting with batched data. First, we find that prompting with batched data in longer contexts will inevitably lead to worse performance, compared to single-data prompting. Second, the performance of the language model is significantly correlated with the positions and order of the batched data, due to the corresponding change in decoder context. To retain efficiency and overcome performance loss, we propose Batch Permutation and Ensembling (BPE), and a novel Self-reflection-guided EArly Stopping (SEAS) technique. Our comprehensive experimental evaluation demonstrates that BPE can boost the performance of BatchPrompt with a striking margin on a range of popular NLP tasks, including question answering (Boolq), textual entailment (RTE), and duplicate questions identification (QQP). These performances are even competitive with/higher than single-data prompting(SinglePrompt), while BatchPrompt requires much fewer LLM calls and input tokens (For SinglePrompt v.s. BatchPrompt with batch size 32, using just 9%-16% the number of LLM calls, Boolq accuracy 90.6% to 90.9% with 27.4% tokens, QQP accuracy 87.2% to 88.4% with 18.6% tokens, RTE accuracy 91.5% to 91.1% with 30.8% tokens). To the best of our knowledge, this is the first work to technically improve prompting efficiency of large language models. We hope our simple yet effective approach will shed light on the future research of large language models. The code will be released.","{'model': 'tldr@v2.0.0', 'text': 'This is the first work to technically improve prompting efficiency of large language models by proposing Batch Permutation and Ensembling (BPE), and a novel Self-reflection-guided EArly Stopping (SEAS) technique.'}",https://arxiv.org/pdf/2309.00384
-centerclip: token clustering for efficient text-video retrieval,Shuai Zhao,"Recently, large-scale pre-training methods like CLIP have made great progress in multi-modal research such as text-video retrieval. In CLIP, transformers are vital for modeling complex multi-modal relations. However, in the vision transformer of CLIP, the essential visual tokenization process, which produces discrete visual token sequences, generates many homogeneous tokens due to the redundancy nature of consecutive and similar frames in videos. This significantly increases computation costs and hinders the deployment of video retrieval models in web applications. In this paper, to reduce the number of redundant video tokens, we design a multi-segment token clustering algorithm to find the most representative tokens and drop the non-essential ones. As the frame redundancy occurs mostly in consecutive frames, we divide videos into multiple segments and conduct segment-level clustering. Center tokens from each segment are later concatenated into a new sequence, while their original spatial-temporal relations are well maintained. We instantiate two clustering algorithms to efficiently find deterministic medoids and iteratively partition groups in high dimensional space. Through this token clustering and center selection procedure, we successfully reduce computation costs by removing redundant visual tokens. This method further enhances segment-level semantic alignment between video and text representations, enforcing the spatio-temporal interactions of tokens from within-segment frames. Our method, coined as CenterCLIP, surpasses existing state-of-the-art by a large margin on typical text-video benchmarks, while reducing the training memory cost by 35% and accelerating the inference speed by 14% at the best case. The code is available at https://github.com/mzhaoshuai/CenterCLIP https://github.com/mzhaoshuai/CenterCLIP.","{'model': 'tldr@v2.0.0', 'text': 'A multi-segment token clustering algorithm to find the most representative tokens and drop the non-essential ones, which surpasses existing state-of-the-art by a large margin on typical text-video benchmarks, while reducing the training memory cost and accelerating the inference speed.'}",https://arxiv.org/pdf/2205.00823
-token dropping for efficient bert pretraining,Le Hou,"Transformer-based models generally allocate the same amount of computation for each token in a given sequence. We develop a simple but effective “token dropping” method to accelerate the pretraining of transformer models, such as BERT, without degrading its performance on downstream tasks. In particular, we drop unimportant tokens starting from an intermediate layer in the model to make the model focus on important tokens more efficiently if with limited computational resource. The dropped tokens are later picked up by the last layer of the model so that the model still produces full-length sequences. We leverage the already built-in masked language modeling (MLM) loss to identify unimportant tokens with practically no computational overhead. In our experiments, this simple approach reduces the pretraining cost of BERT by 25% while achieving similar overall fine-tuning performance on standard downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'A simple but effective “token dropping” method is developed to accelerate the pretraining of transformer models, such as BERT, without degrading its performance on downstream tasks.'}",http://arxiv.org/pdf/2203.13240
-dynamicvit: efficient vision transformers with dynamic token sparsification,Yongming Rao,"Attention is sparse in vision transformers. We observe the final prediction in vision transformers is only based on a subset of most informative tokens, which is sufficient for accurate image recognition. Based on this observation, we propose a dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input. Specifically, we devise a lightweight prediction module to estimate the importance score of each token given the current features. The module is added to different layers to prune redundant tokens hierarchically. To optimize the prediction module in an end-to-end manner, we propose an attention masking strategy to differentiably prune a token by blocking its interactions with other tokens. Benefiting from the nature of self-attention, the unstructured sparse tokens are still hardware friendly, which makes our framework easy to achieve actual speed-up. By hierarchically pruning 66% of the input tokens, our method greatly reduces 31%~37% FLOPs and improves the throughput by over 40% while the drop of accuracy is within 0.5% for various vision transformers. Equipped with the dynamic token sparsification framework, DynamicViT models can achieve very competitive complexity/accuracy trade-offs compared to state-of-the-art CNNs and vision transformers on ImageNet. Code is available at https://github.com/raoyongming/DynamicViT","{'model': 'tldr@v2.0.0', 'text': 'A dynamic token sparsification framework to prune redundant tokens progressively and dynamically based on the input and an attention masking strategy to differentiably prune a token by blocking its interactions with other tokens is proposed.'}",
-modular and parameter-efficient multimodal fusion with prompting,Sheng Liang,"Recent research has made impressive progress in large-scale multimodal pre-training. In the context of the rapid growth of model size, it is necessary to seek efficient and flexible methods other than finetuning. In this paper, we propose to use prompt vectors to align the modalities. Our method achieves comparable performance to several other multimodal fusion methods in low-resource settings. We further show that our method is modular and parameter-efficient for processing tasks involving two or more data modalities.","{'model': 'tldr@v2.0.0', 'text': None}",http://arxiv.org/pdf/2203.08055
-efficient token mixing for transformers via adaptive fourier neural operators,John Guibas,,,
-adapterbias: parameter-efficient token-dependent representation shift for adapters in nlp tasks,Chin-Lun Fu,"Transformer-based pre-trained models with millions of parameters require large storage. Recent approaches tackle this shortcoming by training adapters, but these approaches still require a relatively large number of parameters. In this study, AdapterBias, a surprisingly simple yet effective adapter architecture, is proposed. AdapterBias adds a token-dependent shift to the hidden output of transformer layers to adapt to downstream tasks with only a vector and a linear layer. Extensive experiments are conducted to demonstrate the effectiveness of AdapterBias. The experiments show that our proposed method can dramatically reduce the trainable parameters compared to the previous works with a minimal decrease in task performances compared with fine-tuned pre-trained models. We further find that AdapterBias automatically learns to assign more significant representation shifts to the tokens related to the task in consideration.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that the proposed AdapterBias method can dramatically reduce the trainable parameters compared to the previous works with a minimal decrease in task performances compared with fine-tuned pre-trained models.'}",http://arxiv.org/pdf/2205.00305
-tokenmixup: efficient attention-guided token-level data augmentation for transformers,Hyeong Kyu Choi,"Mixup is a commonly adopted data augmentation technique for image classification. Recent advances in mixup methods primarily focus on mixing based on saliency. However, many saliency detectors require intense computation and are especially burdensome for parameter-heavy transformer models. To this end, we propose TokenMixup, an efficient attention-guided token-level data augmentation method that aims to maximize the saliency of a mixed set of tokens. TokenMixup provides x15 faster saliency-aware data augmentation compared to gradient-based methods. Moreover, we introduce a variant of TokenMixup which mixes tokens within a single instance, thereby enabling multi-scale feature augmentation. Experiments show that our methods significantly improve the baseline models' performance on CIFAR and ImageNet-1K, while being more efficient than previous methods. We also reach state-of-the-art performance on CIFAR-100 among from-scratch transformer models. Code is available at https://github.com/mlvlab/TokenMixup.","{'model': 'tldr@v2.0.0', 'text': 'TokenMixup is proposed, an efficient attention-guided token-level data augmentation method that aims to maximize the saliency of a mixed set of tokens and provides x15 faster saliency-awareData augmentation compared to gradient-based methods.'}",http://arxiv.org/pdf/2210.07562
-tore: token reduction for efficient human mesh recovery with transformer,Zhiyang Dou,"In this paper, we introduce a set of simple yet effective TOken REduction (TORE) strategies for Transformer-based Human Mesh Recovery from monocular images. Current SOTA performance is achieved by Transformer-based structures. However, they suffer from high model complexity and computation cost caused by redundant tokens. We propose token reduction strategies based on two important aspects, i.e., the 3D geometry structure and 2D image feature, where we hierarchically recover the mesh geometry with priors from body structure and conduct token clustering to pass fewer but more discriminative image feature tokens to the Transformer. Our method massively reduces the number of tokens involved in high-complexity interactions in the Transformer. This leads to a significantly reduced computational cost while still achieving competitive or even higher accuracy in shape recovery. Extensive experiments across a wide range of benchmarks validate the superior effectiveness of the proposed method. We further demonstrate the generalizability of our method on hand mesh recovery. Visit our project page at https://frank-zy-dou.github.io/projects/Tore/index.html.","{'model': 'tldr@v2.0.0', 'text': 'This paper introduces a set of simple yet effective TOken REduction (TORE) strategies for Transformer-based Human Mesh Recovery from monocular images that massively reduces the number of tokens involved in high-complexity interactions in the Transformer.'}",https://arxiv.org/pdf/2211.10705
-adaptive fourier neural operators: efficient token mixers for transformers,John Guibas,"Vision transformers have delivered tremendous success in representation learning. This is primarily due to effective token mixing through self attention. However, this scales quadratically with the number of pixels, which becomes infeasible for high-resolution inputs. To cope with this challenge, we propose Adaptive Fourier Neural Operator (AFNO) as an efficient token mixer that learns to mix in the Fourier domain. AFNO is based on a principled foundation of operator learning which allows us to frame token mixing as a continuous global convolution without any dependence on the input resolution. This principle was previously used to design FNO, which solves global convolution efficiently in the Fourier domain and has shown promise in learning challenging PDEs. To handle challenges in visual representation learning such as discontinuities in images and high resolution inputs, we propose principled architectural modifications to FNO which results in memory and computational efficiency. This includes imposing a block-diagonal structure on the channel mixing weights, adaptively sharing weights across tokens, and sparsifying the frequency modes via soft-thresholding and shrinkage. The resulting model is highly parallel with a quasi-linear complexity and has linear memory in the sequence size. AFNO outperforms self-attention mechanisms for few-shot segmentation in terms of both efficiency and accuracy. For Cityscapes segmentation with the Segformer-B3 backbone, AFNO can handle a sequence size of 65k and outperforms other efficient self-attention mechanisms.","{'model': 'tldr@v2.0.0', 'text': 'Adaptive Fourier Neural Operator is proposed as an efficient token mixer that learns to mix in the Fourier domain that can handle a sequence size of 65k and outperforms other efficient self-attention mechanisms for few-shot segmentation in terms of both efficiency and accuracy.'}",
-citadel: conditional token interaction via dynamic lexical routing for efficient and effective multi-vector retrieval,Minghan Li,"Multi-vector retrieval methods combine the merits of sparse (e.g. BM25) and dense (e.g. DPR) retrievers and have achieved state-of-the-art performance on various retrieval tasks.These methods, however, are orders of magnitude slower and need much more space to store their indices compared to their single-vector counterparts.In this paper, we unify different multi-vector retrieval models from a token routing viewpoint and propose conditional token interaction via dynamic lexical routing, namely CITADEL, for efficient and effective multi-vector retrieval.CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key.This design significantly reduces the computation cost while maintaining high accuracy.Notably, CITADEL achieves the same or slightly better performance than the previous state of the art, ColBERT-v2, on both in-domain (MS MARCO) and out-of-domain (BEIR) evaluations, while being nearly 40 times faster. Source code and data are available at https://github.com/facebookresearch/dpr-scale/tree/citadel.","{'model': 'tldr@v2.0.0', 'text': 'CITADEL learns to route different token vectors to the predicted lexical keys such that a query token vector only interacts with document token vectors routed to the same key, which significantly reduces the computation cost while maintaining high accuracy.'}",https://arxiv.org/pdf/2211.10411
-adaptive token sampling for efficient vision transformers,Mohsen Fayyaz,,"{'model': 'tldr@v2.0.0', 'text': 'This work introduces a differentiable parameter-free Adaptive Token Sampler module, which can be plugged into any existing vision transformer architecture, and improves the SOTA by reducing their computational costs (GFLOPs) by 2X, while preserving their accuracy on the ImageNet, Kinetics-400, andKinetics-600 datasets.'}",https://arxiv.org/pdf/2111.15667
-orthogonal transformer: an efficient vision transformer backbone with token orthogonalization,Huaibo Huang,"We present a general vision transformer backbone, called as Orthogonal Trans-former, in pursuit of both efficiency and effectiveness. A major challenge for vision transformer is that self-attention, as the key element in capturing long-range dependency, is very computationally expensive for dense prediction tasks (e.g., object detection). Coarse global self-attention and local self-attention are then designed to reduce the cost, but they suffer from either neglecting local correlations or hurting global modeling. We present an orthogonal self-attention mechanism to alleviate these issues. Specifically, self-attention is computed in the orthogonal space that is reversible to the spatial domain but has much lower resolution. The capabilities of learning global dependency and exploring local correlations are maintained because every orthogonal token in self-attention can attend to the entire visual tokens. Remarkably, orthogonality is realized by constructing an endogenously orthogonal matrix that is friendly to neural networks and can be optimized as arbitrary orthogonal matrices. We also introduce Positional MLP to incorporate position information for arbitrary input resolutions as well as enhance the capacity of MLP. Finally, we develop a hierarchical architecture for Orthogonal Transformer. Extensive experiments demonstrate its strong performance on a broad range of vision tasks, including image classification, object detection, instance segmentation and semantic segmentation.","{'model': 'tldr@v2.0.0', 'text': 'A general vision transformer backbone, called as Orthogonal Trans-former, in pursuit of both efficiency and effectiveness is presented, and an orthogonal self-attention mechanism is presented to alleviate issues.'}",
-dynamicvit: efficient vision transformers with dynamic token sparsification,,.,,
-efficient video transformers with spatial-temporal token selection,Junke Wang,,"{'model': 'tldr@v2.0.0', 'text': 'STTS is a token selection framework that dynamically selects a few informative tokens in both temporal and spatial dimensions conditioned on input video samples that achieves similar results while requiring 20% less computation.'}",https://arxiv.org/pdf/2111.11591
-dtqatten: leveraging dynamic token-based quantization for efficient attention architecture,Tao Yang,"Models based on the attention mechanism, i.e. transformers, have shown extraordinary performance in Natural Language Processing (NLP) tasks. However, their memory footprint, inference latency, and power consumption are still prohibitive for efficient inference at edge devices, even at data centers. To tackle this issue, we present an algorithm-architecture co-design with dynamic and mixed-precision quantization, DTQAtten. We present empirically that the tolerance to the noise varies from token to token in attention-based NLP models. This finding leads us to quantize different tokens with mixed levels of bits. Thus, we design a compression framework that (i) dynamically quantizes tokens while they are forwarded in the models and (ii) jointly determines the ratio of each precision. Moreover, due to the dynamic mixed-precision tokens caused by our framework, previous matrix-multiplication accelerators (e.g. systolic array) cannot effectively exploit the benefit of the compressed attention computation. We thus design our accelerator with the variable-speed systolic array (VSSA) and propose an effective optimization strategy to alleviate the pipeline-stall problem in VSSA without hardware overhead. We conduct experiments with existing attention-based NLP models, including BERT and GPT-2 on various language tasks. Our results show that DTQAtten outperforms the previous neural network accelerator Eyeriss by 13.12× in terms of speedup and 3.8× in terms of energy-saving. Compared with the state-of-the-art attention accelerator SpAtten, our DTQAtten achieves at least 2.65× speedup and 3.38× energy efficiency improvement.","{'model': 'tldr@v2.0.0', 'text': 'This work designs an algorithm-architecture co-design with dynamic and mixed-precision quantization with an effective optimization strategy to alleviate the pipeline-stall problem in VSSA without hardware overhead and conducts experiments with existing attention-based NLP models.'}",
-heatvit: hardware-efficient adaptive token pruning for vision transformers,Peiyan Dong,"While vision transformers (ViTs) have continuously achieved new milestones in the field of computer vision, their sophisticated network architectures with high computation and memory costs have impeded their deployment on resource-limited edge devices. In this paper, we propose a hardware-efficient image-adaptive token pruning framework called HeatViT for efficient yet accurate ViT acceleration on embedded FPGAs. Based on the inherent computational patterns in ViTs, we first adopt an effective, hardware-efficient, and learnable head-evaluation token selector, which can be progressively inserted before transformer blocks to dynamically identify and consolidate the non-informative tokens from input images. Moreover, we implement the token selector on hardware by adding miniature control logic to heavily reuse existing hardware components built for the backbone ViT. To improve the hardware efficiency, we further employ 8-bit fixed-point quantization and propose polynomial approximations with regularization effect on quantization error for the frequently used nonlinear functions in ViTs. Compared to existing ViT pruning studies, under the similar computation cost, HeatViT can achieve 0.7% ~ 8.9% higher accuracy; while under the similar model accuracy, HeatViT can achieve more than 28.4% ~ 65.3% computation reduction, for various widely used ViTs, including DeiT-T, DeiT-S, DeiT-B, LV-ViT-S, and LV-ViT-M, on the ImageNet dataset. Compared to the baseline hardware accelerator, our implementations of HeatViT on the Xilinx ZCU102 FPGA achieve 3.46×~4.89× speedup with a trivial resource utilization overhead of 8%~11% more DSPs and 5%~8% more LUTs.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a hardware-efficient image-adaptive token pruning framework called HeatViT for efficient yet accurate ViT acceleration on embedded FPGAs, and adopts an effective, hardware- efficient, and learnable head-evaluation token selector.'}",http://arxiv.org/pdf/2211.08110
-efficient transformers with dynamic token pooling,Piotr Nawrot,"Transformers achieve unrivalled performance in modelling language, but remain inefficient in terms of memory and time complexity. A possible remedy is to reduce the sequence length in the intermediate layers by pooling fixed-length segments of tokens. Nevertheless, natural units of meaning, such as words or phrases, display varying sizes. To address this mismatch, we equip language models with a dynamic-pooling mechanism, which predicts segment boundaries in an autoregressive fashion. We compare several methods to infer boundaries, including end-to-end learning through stochastic re-parameterisation, supervised learning (based on segmentations from subword tokenizers or spikes in conditional entropy), as well as linguistically motivated boundaries. We perform character-level evaluation on texts from multiple datasets and morphologically diverse languages. The results demonstrate that dynamic pooling, which jointly segments and models language, is both faster and more accurate than vanilla Transformers and fixed-length pooling within the same computational budget.","{'model': 'tldr@v2.0.0', 'text': 'The results demonstrate that dynamic pooling, which jointly segments and models language, is both faster and more accurate than vanilla Transformers and fixed-length pooling within the same computational budget.'}",http://arxiv.org/pdf/2211.09761
-compact token representations with contextual quantization for efficient document re-ranking,Yingrui Yang,"Transformer based re-ranking models can achieve high search relevance through context- aware soft matching of query tokens with document tokens. To alleviate runtime complexity of such inference, previous work has adopted a late interaction architecture with pre-computed contextual token representations at the cost of a large online storage. This paper proposes contextual quantization of token embeddings by decoupling document-specific and document-independent ranking contributions during codebook-based compression. This allows effective online decompression and embedding composition for better search relevance. This paper presents an evaluation of the above compact token representation model in terms of relevance and space efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes contextual quantization of token embeddings by decoupling document-specific and document-independent ranking contributions during codebook-based compression to allow effective online decompression and embedding composition for better search relevance.'}",http://arxiv.org/pdf/2203.15328
-ats: adaptive token sampling for efficient vision transformers,Mohsen Fayyaz,"While state-of-the-art vision transformer models achieve promising results for image classification, they are computationally very expensive and require many GFLOPs. Although the GFLOPs of a vision transformer can be de-creased by reducing the number of tokens in the network, there is no setting that is optimal for all input images. In this work, we therefore introduce a differentiable parameter-free Adaptive Token Sampling (ATS) module, which can be plugged into any existing vision transformer architecture. ATS empowers vision transformers by scoring and adaptively sampling significant tokens. As a result, the number of tokens is not anymore static but it varies for each input image. By integrating ATS as an additional layer within current transformer blocks, we can convert them into much more efficient vision transformers with an adaptive number of tokens. Since ATS is a parameter-free module, it can be added to off-the-shelf pretrained vision transformers as a plug and play module, thus reducing their GFLOPs without any additional training. However, due to its differentiable design, one can also train a vision transformer equipped with ATS. We evaluate our module on the ImageNet dataset by adding it to multiple state-of-the-art vision transformers. Our evaluations show that the proposed module improves the state-of-the-art by reducing the computational cost (GFLOPs) by 37% while preserving the accuracy.","{'model': 'tldr@v2.0.0', 'text': 'A differentiable parameter-free Adaptive Token Sampling (ATS) module, which can be plugged into any existing vision transformer architecture, and improves the state-of-the-art by reducing the computational cost (GFLOPs) by 37% while preserving the accuracy.'}",
-you only group once: efficient point-cloud processing with token representation and relation inference module,Chenfeng Xu,"3D perception on point-cloud is a challenging and crucial computer vision task. A point-cloud consists of a sparse, unstructured, and unordered set of points. To understand a point-cloud, previous point-based methods, such as PointNet++, extract visual features through the hierarchical aggregation of local features. However, such methods have several critical limitations: 1) They require considerable sampling and grouping operations, which leads to low inference speed. 2) Despite redundancy among adjacent points, they treat all points alike with an equal amount of computation. 3) They aggregate local features together through downsampling, which causes information loss and hurts perception capability. To overcome these challenges, we propose a novel, simple, and elegant deep learning model called YOGO (You Only Group Once). YOGO divides a point-cloud into a small number of parts and extracts a high-dimensional token to represent points within each sub-region. Next, we use self-attention to capture token-to-token relations, and project the token features back to the point features. We formulate such a series of operations as a relation inference module (RIM). Compared with previous methods, YOGO is very efficient because it only needs to sample and group a point-cloud once. Instead of operating on points, YOGO operates on a small number of tokens, each of which summarizes the point features in a sub-region. This allows us to avoid redundant computation and thus boosts efficiency. Moreover, YOGO preserves pointwise features by projecting token features to point features although the RIM computes on tokens. This avoids information loss and enhances point-wise perception capability. We conduct thorough experiments to demonstrate that YOGO achieves at least 3.0x speedup over point-based baselines while delivering competitive classification and segmentation performance on a classification dataset and a segmentation dataset based on 3D Wharehouse, and S3DIS datasets. The code is available at https://github.com/chenfengxu714/YOGO.git.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel, simple, and elegant deep learning model called YOGO (You Only Group Once), which operates on a small number of tokens, each of which summarizes the point features in a sub-region, to avoid redundant computation and thus boosts efficiency.'}",https://arxiv.org/pdf/2103.09975
-spatten: efficient sparse attention architecture with cascade token and head pruning,Hanrui Wang,"The attention mechanism is becoming increasingly popular in Natural Language Processing (NLP) applications, showing superior performance than convolutional and recurrent architectures. However, general-purpose platforms such as CPUs and GPUs are inefficient when performing attention inference due to complicated data movement and low arithmetic intensity. Moreover, existing NN accelerators mainly focus on optimizing convolutional or recurrent models, and cannot efficiently support attention. In this paper, we present SpAtten, an efficient algorithm-architecture co-design that leverages token sparsity, head sparsity, and quantization opportunities to reduce the attention computation and memory access. Inspired by the high redundancy of human languages, we propose the novel cascade token pruning to prune away unimportant tokens in the sentence. We also propose cascade head pruning to remove unessential heads. Cascade pruning is fundamentally different from weight pruning since there is no trainable weight in the attention mechanism, and the pruned tokens and heads are selected on the fly. To efficiently support them on hardware, we design a novel top-k engine to rank token and head importance scores with high throughput. Furthermore, we propose progressive quantization that first fetches MSBs only and performs the computation; if the confidence is low, it fetches LSBs and recomputes the attention outputs, trading computation for memory reduction.Extensive experiments on 30 benchmarks show that, on average, SpAtten reduces DRAM access by 10.0× with no accuracy loss, and achieves 1.6×, 3.0×, 162×, 347× speedup, and 1.4×, 3.2×, 1193×, 4059× energy savings over A3 accelerator, MNNFast accelerator, TITAN Xp GPU, Xeon CPU, respectively.","{'model': 'tldr@v2.0.0', 'text': 'SpAtten is presented, an efficient algorithm-architecture co-design that leverages token sparsity, head Sparsity, and quantization opportunities to reduce the attention computation and memory access and proposes the novel cascade token pruning to prune away unimportant tokens in the sentence.'}",https://arxiv.org/pdf/2012.09852
-ponet: pooling network for efficient token mixing in long sequences,Chao-Hong Tan,"Transformer-based models have achieved great success in various NLP, vision, and speech tasks. However, the core of Transformer, the self-attention mechanism, has a quadratic time and memory complexity with respect to the sequence length, which hinders applications of Transformer-based models to long sequences. Many approaches have been proposed to mitigate this problem, such as sparse attention mechanisms, low-rank matrix approximations and scalable kernels, and token mixing alternatives to self-attention. We propose a novel Pooling Network (PoNet) for token mixing in long sequences with linear complexity. We design multi-granularity pooling and pooling fusion to capture different levels of contextual information and combine their interactions with tokens. On the Long Range Arena benchmark, PoNet significantly outperforms Transformer and achieves competitive accuracy, while being only slightly slower than the fastest model, FNet, across all sequence lengths measured on GPUs. We also conduct systematic studies on the transfer learning capability of PoNet and observe that PoNet achieves 95.7% of the accuracy of BERT on the GLUE benchmark, outperforming FNet by 4.5% relative. Comprehensive ablation analysis demonstrates effectiveness of the designed multi-granularity pooling and pooling fusion for token mixing in long sequences and efficacy of the designed pre-training tasks for PoNet to learn transferable contextualized language representations.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel Pooling Network (PoNet) for token mixing in long sequences with linear complexity and designs multi-granularity pooling and pooling fusion to capture different levels of contextual information and combine their interactions with tokens.'}",
-token system‐based efficient route optimization in mobile ad hoc network for vehicular ad hoc network in smart city,S. Sultanuddin,"Mobile ad hoc network (MANET) is a well‐known wireless hoc network, which several nodes combined together to form a network. These nodes do not have any structure that forms network by self‐configuring. In the global area, MANET is one of the popular technologies. At present, the most important real‐time challenge in MANET is efficient routing. This is because of dynamic topology and resource constraints. Several existing researcher's workfare were done in an achieving efficient routing mechanism. In our work, we propose a token agent‐based route planning protocol for achieving reliable routing. Vehicular ad hoc network (VANET) is advanced concept of the MANET. VANET can be performed in various smart cities, and it travels each city from source to destination. This token system consists of a token agent that monitors and operates the whole network. The tokens are responsible for collecting nodes activity details such as mobile node energy, node location, and speed. Later, its current and previous information is up‐to‐date and inmate to the token agent for further process in the network. Our proposed routing protocol is well applicable for typical application and a multimedia application. In order to transfer typical packets, our proposed scheme finds the shortest path based on the collect the information.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a token agent‐based route planning protocol for achieving reliable routing in mobile ad hoc network and its proposed protocol is well applicable for typical application and a multimedia application.'}",
-making vision transformers efficient from a token sparsification view,Shuning Chang,"The quadratic computational complexity to the number of tokens limits the practical applications of Vision Transformers (ViTs). Several works propose to prune redundant tokens to achieve efficient ViTs. However, these methods generally suffer from (i) dramatic accuracy drops, (ii) application difficulty in the local vision transformer, and (iii) non-general-purpose networks for downstream tasks. In this work, we propose a novel Semantic Token ViT (STViT), for efficient global and local vision transformers, which can also be revised to serve as backbone for downstream tasks. The semantic tokens represent cluster centers, and they are initialized by pooling image tokens in space and recovered by attention, which can adaptively represent global or local semantic information. Due to the cluster properties, a few semantic tokens can attain the same effect as vast image tokens, for both global and local vision transformers. For instance, only 16 semantic tokens on DeiT-(Tiny,Small,Base) can achieve the same accuracy with more than 100% inference speed improvement and nearly 60% FLOPs reduction; on Swin-(Tiny,Small,Base), we can employ 16 semantic tokens in each window to further speed it up by around 20% with slight accuracy increase. Besides great success in image classification, we also extend our method to video recognition. In addition, we design a STViT-R(ecovery) network to restore the detailed spatial information based on the STViT, making it work for downstream tasks, which is powerless for previous token sparsification methods. Experiments demonstrate that our method can achieve competitive results compared to the original networks in object detection and instance segmentation, with over 30% FLOPs reduction for backbone.","{'model': 'tldr@v2.0.0', 'text': 'A novel Semantic Token ViT (STViT) is proposed, for efficient global and local vision transformers, which can also be revised to serve as backbone for downstream tasks, and a STViT-R(ecovery) network is designed to restore the detailed spatial information based on the ST ViT, making it work for upstream tasks, which is powerless for previous token sparsification methods.'}",https://arxiv.org/pdf/2303.08685
-efficient load-balancing through distributed token dropping,S. Brandt,"We introduce a new graph problem, the token dropping game, and we show how to solve it efficiently in a distributed setting. We use the token dropping game as a tool to design an efficient distributed algorithm for stable orientations and more generally for locally optimal semi-matchings. The prior work by Czygrinow et al. (DISC 2012) finds a stable orientation in O(Δ^5) rounds in graphs of maximum degree Δ, while we improve it to O(Δ^4) and also prove a lower bound of Ω(Δ). For the more general problem of locally optimal semi-matchings, the prior upper bound is O(S^5) and our new algorithm runs in O(C · S^4) rounds, which is an improvement for C = o(S); here C and S are the maximum degrees of customers and servers, respectively.","{'model': 'tldr@v2.0.0', 'text': 'This work uses the token dropping game as a tool to design an efficient distributed algorithm for stable orientations and more generally for locally optimal semi-matchings and introduces a new graph problem, thetoken dropping game, and shows how to solve it efficiently in a distributed setting.'}",https://arxiv.org/pdf/2005.07761
-not all patches are what you need: expediting vision transformers via token reorganizations,Youwei Liang,"Vision Transformers (ViTs) take all the image patches as tokens and construct multi-head self-attention (MHSA) among them. Complete leverage of these image tokens brings redundant computations since not all the tokens are attentive in MHSA. Examples include that tokens containing semantically meaningless or distractive image backgrounds do not positively contribute to the ViT predictions. In this work, we propose to reorganize image tokens during the feed-forward process of ViT models, which is integrated into ViT during training. For each forward inference, we identify the attentive image tokens between MHSA and FFN (i.e., feed-forward network) modules, which is guided by the corresponding class token attention. Then, we reorganize image tokens by preserving attentive image tokens and fusing inattentive ones to expedite subsequent MHSA and FFN computations. To this end, our method EViT improves ViTs from two perspectives. First, under the same amount of input image tokens, our method reduces MHSA and FFN computation for efficient inference. For instance, the inference speed of DeiT-S is increased by 50% while its recognition accuracy is decreased by only 0.3% for ImageNet classification. Second, by maintaining the same computational cost, our method empowers ViTs to take more image tokens as input for recognition accuracy improvement, where the image tokens are from higher resolution images. An example is that we improve the recognition accuracy of DeiT-S by 1% for ImageNet classification at the same computational cost of a vanilla DeiT-S. Meanwhile, our method does not introduce more parameters to ViTs. Experiments on the standard benchmarks show the effectiveness of our method. The code is available at https://github.com/youweiliang/evit","{'model': 'tldr@v2.0.0', 'text': 'This work reorganizes image tokens by preserving attentive image tokens and fusing inattentive ones to expedite subsequent MHSA and FFN computations during the feed-forward process of ViT models, which is integrated into ViT during training.'}",
-frozen clip models are efficient video learners,Ziyi Lin,"Video recognition has been dominated by the end-to-end learning paradigm -- first initializing a video recognition model with weights of a pretrained image model and then conducting end-to-end training on videos. This enables the video network to benefit from the pretrained image model. However, this requires substantial computation and memory resources for finetuning on videos and the alternative of directly using pretrained image features without finetuning the image backbone leads to subpar results. Fortunately, recent advances in Contrastive Vision-Language Pre-training (CLIP) pave the way for a new route for visual recognition tasks. Pretrained on large open-vocabulary image-text pair data, these models learn powerful visual representations with rich semantics. In this paper, we present Efficient Video Learning (EVL) -- an efficient framework for directly training high-quality video recognition models with frozen CLIP features. Specifically, we employ a lightweight Transformer decoder and learn a query token to dynamically collect frame-level spatial features from the CLIP image encoder. Furthermore, we adopt a local temporal module in each decoder layer to discover temporal clues from adjacent frames and their attention maps. We show that despite being efficient to train with a frozen backbone, our models learn high quality video representations on a variety of video recognition datasets. Code is available at https://github.com/OpenGVLab/efficient-video-recognition.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents Efficient Video Learning (EVL) -- an efficient framework for directly training high-quality video recognition models with frozen CLIP features that adopt a lightweight Transformer decoder and learn a query token to dynamically collect frame-level spatial features from the CLIP image encoder.'}",http://arxiv.org/pdf/2208.03550
-towards efficient cross-blockchain token transfers,Philipp Frauenthaler,"Interoperability between blockchains remains an open problem, with current interoperability approaches providing very limited means of cross-blockchain interaction, mostly in the form of atomic swaps. More general means of blockchain interoperability such as cross-blockchain data exchange, including cross-blockchain token transfer would contribute to dissolving today’s fragmentation of the research and development field of blockchains. To address this issue, within the TAST research project, a cross-blockchain token was developed. However, the developed solution suffers from high synchronization cost. In this paper, we discuss requirements for more efficient crossblockchain token transfers, describe open research challenges, and give an outlook on two approaches aiming to overcome these challenges.","{'model': 'tldr@v2.0.0', 'text': 'Requirements for more efficient crossblockchain token transfers are discussed, open research challenges are described, and an outlook on two approaches aiming to overcome these challenges are given.'}",
-training data-efficient image transformers & distillation through attention,Hugo Touvron,"Recently, neural networks purely based on attention were shown to address image understanding tasks such as image classification. However, these visual transformers are pre-trained with hundreds of millions of images using an expensive infrastructure, thereby limiting their adoption. In this work, we produce a competitive convolution-free transformer by training on Imagenet only. We train them on a single computer in less than 3 days. Our reference vision transformer (86M parameters) achieves top-1 accuracy of 83.1% (single-crop evaluation) on ImageNet with no external data. More importantly, we introduce a teacher-student strategy specific to transformers. It relies on a distillation token ensuring that the student learns from the teacher through attention. We show the interest of this token-based distillation, especially when using a convnet as a teacher. This leads us to report results competitive with convnets for both Imagenet (where we obtain up to 85.2% accuracy) and when transferring to other tasks. We share our code and models.","{'model': 'tldr@v2.0.0', 'text': 'This work produces a competitive convolution-free transformer by training on Imagenet only and introduces a teacher-student strategy specific to transformers that relies on a distillation token ensuring that the student learns from the teacher through attention.'}",
-a new generation of perspective api: efficient multilingual character-level transformers,Alyssa Lees,"On the world wide web, toxic content detectors are a crucial line of defense against potentially hateful and offensive messages. As such, building highly effective classifiers that enable a safer internet is an important research area. Moreover, the web is a highly multilingual, cross-cultural community that develops its own lingo over time. As such, it is crucial to develop models that are effective across a diverse range of languages, usages, and styles. In this paper, we present the fundamentals behind the next version of the Perspective API from Google Jigsaw. At the heart of the approach is a single multilingual token-free Charformer model that is applicable across a range of languages, domains, and tasks. We demonstrate that by forgoing static vocabularies, we gain flexibility across a variety of settings. We additionally outline the techniques employed to make such a byte-level model efficient and feasible for productionization. Through extensive experiments on multilingual toxic comment classification benchmarks derived from real API traffic and evaluation on an array of code-switching, covert toxicity, emoji-based hate, human-readable obfuscation, distribution shift, and bias evaluation settings, we show that our proposed approach outperforms strong baselines. Finally, we present our findings from deploying this system in production.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents the fundamentals behind the next version of the Perspective API from Google Jigsaw, and presents a single multilingual token-free Charformer model that is applicable across a range of languages, domains, and tasks.'}",https://dl.acm.org/doi/pdf/10.1145/3534678.3539147
-colbertv2: effective and efficient retrieval via lightweight late interaction,Keshav Santhanam,"Neural information retrieval (IR) has greatly advanced search and other knowledge-intensive language tasks. While many neural IR methods encode queries and documents into single-vector representations, late interaction models produce multi-vector representations at the granularity of each token and decompose relevance modeling into scalable token-level computations. This decomposition has been shown to make late interaction more effective, but it inflates the space footprint of these models by an order of magnitude. In this work, we introduce Maize, a retriever that couples an aggressive residual compression mechanism with a denoised supervision strategy to simultaneously improve the quality and space footprint of late interaction. We evaluate Maize across a wide range of benchmarks, establishing state-of-the-art quality within and outside the training domain while reducing the space footprint of late interaction models by 6–10x.","{'model': 'tldr@v2.0.0', 'text': 'Maize is introduced, a retriever that couples an aggressive residual compression mechanism with a denoised supervision strategy to simultaneously improve the quality and space footprint of late interaction and establishes state-of-the-art quality within and outside the training domain.'}",https://aclanthology.org/2022.naacl-main.272.pdf
-abdomen ct multi-organ segmentation using token-based mlp-mixer.,Shaoyan Pan,"BACKGROUND Manual contouring is very labor-intensive, time-consuming, and subject to intra- and inter-observer variability. An automated deep learning approach to fast and accurate contouring and segmentation is desirable during radiotherapy treatment planning. PURPOSE This work investigates an efficient deep-learning-based segmentation algorithm in abdomen computed tomography (CT) to facilitate radiation treatment planning. METHODS In this work, we propose a novel deep-learning model utilizing U-shaped Multi-Layer Perceptron Mixer (MLP-Mixer) and convolutional neural network (CNN) for multi-organ segmentation in abdomen CT images. The proposed model has a similar structure to V-net, while a proposed MLP-Convolutional block replaces each convolutional block. The MLP-Convolutional block consists of three components: an early convolutional block for local features extraction and feature resampling, a token-based MLP-Mixer layer for capturing global features with high efficiency, and a token projector for pixel-level detail recovery. We evaluate our proposed network using: 1) an institutional dataset with 60 patient cases, and 2) a public dataset (BCTV) with 30 patient cases. The network performance was quantitatively evaluated in three domains: 1) volume similarity between the ground truth contours and the network predictions using the Dice score coefficient (DSC), sensitivity, and precision; 2) surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMS); 3) the computational complexity reported by the number of network parameters, training time, and inference time. The performance of the proposed network is compared with other state-of-the-art networks. RESULTS In the institutional dataset, the proposed network achieved the following volume similarity measures when averaged over all organs: DSC = 0.912, sensitivity = 0.917, precision = 0.917, average surface similarities were HD = 11.95mm, MSD = 1.90mm, RMS = 3.86mm. The proposed network achieved DSC = 0.786 and HD = 9.04mm on the public dataset. The network also shows statistically significant improvement, which is evaluated by a two-tailed Wilcoxon Mann-Whitney U test, on right lung (MSD where the maximum p-value is 0.001), spinal cord (sensitivity, precision, HD, RMSD where p-value ranges from 0.001 to 0.039), and stomach (DSC where the maximum p-value is 0.01) over all other competing networks. On the public dataset, the network report statistically significant improvement, which is shown by the Wilcoxon Mann-Whitney test, on pancreas (HD where the maximum p-value is 0.006), left (HD where the maximum p-value is 0.022) and right adrenal glands (DSC where the maximum p-value is 0.026). In both datasets, the proposed method can generate contours in less than five seconds. Overall, the proposed MLP-Vnet demonstrates comparable or better performance than competing methods with much lower memory complexity and higher speed. CONCLUSIONS The proposed MLP-Vnet demonstrates superior segmentation performance, in terms of accuracy and efficiency, relative to state-of-the-art methods. This reliable and efficient method demonstrates potential to streamline clinical workflows in abdominal radiotherapy, which may be especially important for online adaptive treatments. This article is protected by copyright. All rights reserved.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates an efficient deep-learning-based segmentation algorithm in abdomen computed tomography (CT) to facilitate radiation treatment planning and demonstrates potential to streamline clinical workflows in abdominal radiotherapy, which may be especially important for online adaptive treatments.'}",
-hydra attention: efficient attention with many heads,Daniel Bolya,"While transformers have begun to dominate many tasks in vision, applying them to large images is still computationally difficult. A large reason for this is that self-attention scales quadratically with the number of tokens, which in turn, scales quadratically with the image size. On larger images (e.g., 1080p), over 60% of the total computation in the network is spent solely on creating and applying attention matrices. We take a step toward solving this issue by introducing Hydra Attention, an extremely efficient attention operation for Vision Transformers (ViTs). Paradoxically, this efficiency comes from taking multi-head attention to its extreme: by using as many attention heads as there are features, Hydra Attention is computationally linear in both tokens and features with no hidden constants, making it significantly faster than standard self-attention in an off-the-shelf ViT-B/16 by a factor of the token count. Moreover, Hydra Attention retains high accuracy on ImageNet and, in some cases, actually improves it.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Hydra Attention, an extremely efficient attention operation for Vision Transformers (ViTs), which is computationally linear in both tokens and features with no hidden constants, making it significantly faster than standard self-attention in an off-the-shelf ViT-B/16 by a factor of the token count.'}",http://arxiv.org/pdf/2209.07484
-to collaborate or not to collaborate: prompting upstream eco-efficient innovation in a supply chain,Arda Yenipazarli,,"{'model': 'tldr@v2.0.0', 'text': 'A two-echelon supply chain where an upstream supplier sells through a downstream retailer and the retailer who embodies the channel power sets the product price and sells to consumers is considered, which studies the profit/cost implications of collaboration between two parties for upstream eco-efficient innovation.'}",
-dytox: transformers for continual learning with dynamic token expansion,Arthur Douillard,"Deep network architectures struggle to continually learn new tasks without forgetting the previous tasks. A recent trend indicates that dynamic architectures based on an ex-pansion of the parameters can reduce catastrophic forget-ting efficiently in continual learning. However, existing approaches often require a task identifier at test-time, need complex tuning to balance the growing number of parameters, and barely share any information across tasks. As a result, they struggle to scale to a large number of tasks without significant overhead. In this paper, we propose a transformer architecture based on a dedicated encoder/decoder framework. Critically, the encoder and decoder are shared among all tasks. Through a dynamic expansion of special tokens, we specialize each forward of our decoder network on a task distribution. Our strategy scales to a large number of tasks while having neg-ligible memory and time overheads due to strict control of the expansion of the parameters. Moreover, this efficient strategy doesn't need any hyperparameter tuning to control the network's expansion. Our model reaches excellent results on CIFAR100 and state-of-the-art performances on the large-scale ImageNet100 and ImageNet100 while having fewer parameters than concurrent dynamic frameworks.11Code is released at https://github.com/arthurdouillard/dytox.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a transformer architecture based on a dedicated encoder/decoder framework that reaches excellent results on CIFAR100 and state-of-the-art performances on the large-scale ImageNet 100 and ImageNet100 while having fewer parameters than concurrent dynamic frameworks.'}",https://arxiv.org/pdf/2111.11326
-taco: token-aware cascade contrastive learning for video-text alignment,Jianwei Yang,"Contrastive learning has been widely used to train transformer-based vision-language models for video-text alignment and multi-modal representation learning. This paper presents a new algorithm called Token-Aware Cascade contrastive learning (TACo) that improves contrastive learning using two novel techniques. The first is the token-aware contrastive loss which is computed by taking into account the syntactic classes of words. This is motivated by the observation that for a video-text pair, the content words in the text, such as nouns and verbs, are more likely to be aligned with the visual contents in the video than the function words. Second, a cascade sampling method is applied to generate a small set of hard negative examples for efficient loss estimation for multi-modal fusion layers. To validate the effectiveness of TACo, in our experiments we finetune pretrained models for a set of downstream tasks including text-video retrieval (YouCook2, MSR-VTT and ActivityNet), video action step localization (CrossTask), video action segmentation (COIN). The results show that our models attain consistent improvements across different experimental settings over previous methods, set-ting new state-of-the-art on three public text-video retrieval benchmarks of YouCook2, MSR-VTT and ActivityNet.","{'model': 'tldr@v2.0.0', 'text': 'A new algorithm called Token-Aware Cascade contrastive learning (TACo) is presented that improves Contrastive learning using two novel techniques, the token-aware contrastive loss which is computed by taking into account the syntactic classes of words.'}",https://arxiv.org/pdf/2108.09980
-learned token pruning for transformers,Sehoon Kim,"Efficient deployment of transformer models in practice is challenging due to their inference cost including memory footprint, latency, and power consumption, which scales quadratically with input sequence length. To address this, we present a novel token reduction method dubbed Learned Token Pruning (LTP) which adaptively removes unimportant tokens as an input sequence passes through transformer layers. In particular, LTP prunes tokens with an attention score below a threshold, whose value is learned for each layer during training. Our threshold-based method allows the length of the pruned sequence to vary adaptively based on the input sequence, and avoids algorithmically expensive operations such as top-k token selection. We extensively test the performance of LTP on GLUE and SQuAD tasks and show that our method outperforms the prior state-of-the-art token pruning methods by up to ∽2.5% higher accuracy with the same amount of FLOPs. In particular, LTP achieves up to 2.1× FLOPs reduction with less than 1% accuracy drop, which results in up to 1.9× and 2.0× throughput improvement on Intel Haswell CPUs and NVIDIA V100 GPUs. Furthermore, we demonstrate that LTP is more robust than prior methods to variations in input sequence lengths. Our code has been developed in PyTorch and open-sourced","{'model': 'tldr@v2.0.0', 'text': 'A novel token reduction method dubbed Learned Token Pruning (LTP) which adaptively removes unimportant tokens as an input sequence passes through transformer layers, which is more robust than prior methods to variations in input sequence lengths.'}",https://dl.acm.org/doi/pdf/10.1145/3534678.3539260
-tokens-to-token vit: training vision transformers from scratch on imagenet,Li Yuan,"Transformers, which are popular for language modeling, have been explored for solving vision tasks recently, e.g., the Vision Transformer (ViT) for image classification. The ViT model splits each image into a sequence of tokens with fixed length and then applies multiple Transformer layers to model their global relation for classification. However, ViT achieves inferior performance to CNNs when trained from scratch on a midsize dataset like ImageNet. We find it is because: 1) the simple tokenization of input images fails to model the important local structure such as edges and lines among neighboring pixels, leading to low training sample efficiency; 2) the redundant attention backbone design of ViT leads to limited feature richness for fixed computation budgets and limited training samples. To overcome such limitations, we propose a new Tokens-To-Token Vision Transformer (T2T-VTT), which incorporates 1) a layer-wise Tokens-to-Token (T2T) transformation to progressively structurize the image to tokens by recursively aggregating neighboring Tokens into one Token (Tokens-to-Token), such that local structure represented by surrounding tokens can be modeled and tokens length can be reduced; 2) an efficient backbone with a deep-narrow structure for vision transformer motivated by CNN architecture design after empirical study. Notably, T2T-ViT reduces the parameter count and MACs of vanilla ViT by half, while achieving more than 3.0% improvement when trained from scratch on ImageNet. It also outperforms ResNets and achieves comparable performance with MobileNets by directly training on ImageNet. For example, T2T-ViT with comparable size to ResNet50 (21.5M parameters) can achieve 83.3% top1 accuracy in image resolution 384x384 on ImageNet.1","{'model': 'tldr@v2.0.0', 'text': 'A new Tokens-To-Token Vision Transformer (T2T-VTT), which incorporates an efficient backbone with a deep-narrow structure for vision transformer motivated by CNN architecture design after empirical study and reduces the parameter count and MACs of vanilla ViT by half.'}",https://arxiv.org/pdf/2101.11986
-rest: an efficient transformer for visual recognition,Qing-Long Zhang,"This paper presents an efficient multi-scale vision Transformer, called ResT, that capably served as a general-purpose backbone for image recognition. Unlike existing Transformer methods, which employ standard Transformer blocks to tackle raw images with a fixed resolution, our ResT have several advantages: (1) A memory-efficient multi-head self-attention is built, which compresses the memory by a simple depth-wise convolution, and projects the interaction across the attention-heads dimension while keeping the diversity ability of multi-heads; (2) Position encoding is constructed as spatial attention, which is more flexible and can tackle with input images of arbitrary size without interpolation or fine-tune; (3) Instead of the straightforward tokenization at the beginning of each stage, we design the patch embedding as a stack of overlapping convolution operation with stride on the 2D-reshaped token map. We comprehensively validate ResT on image classification and downstream tasks. Experimental results show that the proposed ResT can outperform the recently state-of-the-art backbones by a large margin, demonstrating the potential of ResT as strong backbones. The code and models will be made publicly available at https://github.com/wofmanaf/ResT.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that the proposed ResT can outperform the recently state-of-the-art backbones by a large margin, demonstrating the potential of ResT as strong backbones.'}",
-gating dropout: communication-efficient regularization for sparsely activated transformers,R. Liu,"Sparsely activated transformers, such as Mixture of Experts (MoE), have received great interest due to their outrageous scaling capability which enables dramatical increases in model size without significant increases in computational cost. To achieve this, MoE models replace the feedforward sub-layer with Mixture-of-Experts sub-layer in transformers and use a gating network to route each token to its assigned experts. Since the common practice for efficient training of such models requires distributing experts and tokens across different machines, this routing strategy often incurs huge cross-machine communication cost because tokens and their assigned experts likely reside in different machines. In this paper, we propose \emph{Gating Dropout}, which allows tokens to ignore the gating network and stay at their local machines, thus reducing the cross-machine communication. Similar to traditional dropout, we also show that Gating Dropout has a regularization effect during training, resulting in improved generalization performance. We validate the effectiveness of Gating Dropout on multilingual machine translation tasks. Our results demonstrate that Gating Dropout improves a state-of-the-art MoE model with faster wall-clock time convergence rates and better BLEU scores for a variety of model sizes and datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes Gating Dropout, which allows tokens to ignore the gating network and stay at their local machines, thus reducing the cross-machine communication and improves a state-of-the-art MoE model with faster wall-clock time convergence rates and better BLEU scores for a variety of model sizes and datasets.'}",http://arxiv.org/pdf/2205.14336
-building extraction from remote sensing images with sparse token transformers,Keyan Chen,"Deep learning methods have achieved considerable progress in remote sensing image building extraction. Most building extraction methods are based on Convolutional Neural Networks (CNN). Recently, vision transformers have provided a better perspective for modeling long-range context in images, but usually suffer from high computational complexity and memory usage. In this paper, we explored the potential of using transformers for efficient building extraction. We design an efficient dual-pathway transformer structure that learns the long-term dependency of tokens in both their spatial and channel dimensions and achieves state-of-the-art accuracy on benchmark building extraction datasets. Since single buildings in remote sensing images usually only occupy a very small part of the image pixels, we represent buildings as a set of “sparse” feature vectors in their feature space by introducing a new module called “sparse token sampler”. With such a design, the computational complexity in transformers can be greatly reduced over an order of magnitude. We refer to our method as Sparse Token Transformers (STT). Experiments conducted on the Wuhan University Aerial Building Dataset (WHU) and the Inria Aerial Image Labeling Dataset (INRIA) suggest the effectiveness and efficiency of our method. Compared with some widely used segmentation methods and some state-of-the-art building extraction methods, STT has achieved the best performance with low time cost.","{'model': 'tldr@v2.0.0', 'text': 'An efficient dual-pathway transformer structure that learns the long-term dependency of tokens in both their spatial and channel dimensions and achieves state-of-the-art accuracy on benchmark building extraction datasets is designed.'}",https://www.mdpi.com/2072-4292/13/21/4441/pdf?version=1636075865
-token pooling in vision transformers,D. Marin,"Despite the recent success in many applications, the high computational requirements of vision transformers limit their use in resource-constrained settings. While many existing methods improve the quadratic complexity of attention, in most vision transformers, self-attention is not the major computation bottleneck, e.g., more than 80% of the computation is spent on fully-connected layers. To improve the computational complexity of all layers, we propose a novel token downsampling method, called Token Pooling, efficiently exploiting redundancies in the images and intermediate token representations. We show that, under mild assumptions, softmax-attention acts as a high-dimensional low-pass (smoothing) filter. Thus, its output contains redundancy that can be pruned to achieve a better trade-off between the computational cost and accuracy. Our new technique accurately approximates a set of tokens by minimizing the reconstruction error caused by downsampling. We solve this optimization problem via cost-efficient clustering. We rigorously analyze and compare to prior downsampling methods. Our experiments show that Token Pooling significantly improves the cost-accuracy trade-off over the state-of-the-art downsampling. Token Pooling is a simple and effective operator that can benefit many architectures. Applied to DeiT, it achieves the same ImageNet top-1 accuracy using 42% fewer computations.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel token downsampling method, called Token Pooling, efficiently exploiting redundancies in the images and intermediate token representations, and shows that it significantly improves the cost-accuracy trade-off over the state-of-the-art downsamplings.'}",
-a-vit: adaptive tokens for efficient vision transformer,Hongxu Yin,"We introduce A - ViT, a method that adaptively adjusts the inference cost of vision transformer (ViT) for images of different complexity. A - ViT achieves this by automatically reducing the number of tokens in vision transformers that are processed in the network as inference proceeds. We refor-mulate Adaptive Computation Time (ACT [17]) for this task, extending halting to discard redundant spatial tokens. The appealing architectural properties of vision transformers enables our adaptive token reduction mechanism to speed up inference without modifying the network architecture or inference hardware. We demonstrate that A - ViT requires no extra parameters or sub-network for halting, as we base the learning of adaptive halting on the original network parameters. We further introduce distributional prior regularization that stabilizes training compared to prior ACT approaches. On the image classification task (ImageNet1K), we show that our proposed A - ViT yields high efficacy in filtering informative spatial features and cutting down on the overall compute. The proposed method improves the throughput of DeiT-Tiny by 62% and DeiT-Small by 38% with only 0.3% accuracy drop, outperforming prior art by a large margin.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces A - ViT, a method that adaptively adjusts the inference cost of vision transformer (ViT) for images of different complexity by automatically reducing the number of tokens in vision transformers that are processed in the network as inference proceeds, and demonstrates that it requires no extra parameters or sub-network for halting.'}",
-vision transformer with super token sampling,Huaibo Huang,"Vision transformer has achieved impressive performance for many vision tasks. However, it may suffer from high redundancy in capturing local features for shallow layers. Local self-attention or early-stage convolutions are thus utilized, which sacrifice the capacity to capture long-range dependency. A challenge then arises: can we access efficient and effective global context modeling at the early stages of a neural network? To address this issue, we draw inspiration from the design of superpixels, which reduces the number of image primitives in subsequent processing, and introduce super tokens into vision transformer. Super tokens attempt to provide a semantically meaningful tessellation of visual content, thus reducing the token number in self-attention as well as preserving global modeling. Specifically, we propose a simple yet strong super token attention (STA) mechanism with three steps: the first samples super tokens from visual tokens via sparse association learning, the second performs self-attention on super tokens, and the last maps them back to the original token space. STA decomposes vanilla global attention into multiplications of a sparse association map and a low-dimensional attention, leading to high efficiency in capturing global dependencies. Based on STA, we develop a hierarchical vision transformer. Extensive experiments demonstrate its strong performance on various vision tasks. In particular, without any extra training data or label, it achieves 86.4% top-1 accuracy on ImageNet-1K with less than 100M parameters. It also achieves 53.9 box AP and 46.8 mask AP on the COCO detection task, and 51.9 mIOU on the ADE20K semantic segmentation task. Code will be released at https://github.com/hhb072/SViT.","{'model': 'tldr@v2.0.0', 'text': 'A simple yet strong super token attention (STA) mechanism with three steps: the first samples super tokens from visual tokens via sparse association learning, the second performs self-attention on super tokens, and the last maps them back to the original token space.'}",http://arxiv.org/pdf/2211.11167
-not all images are worth 16x16 words: dynamic transformers for efficient image recognition,Yulin Wang,"Vision Transformers (ViT) have achieved remarkable success in large-scale image recognition. They split every 2D image into a fixed number of patches, each of which is treated as a token. Generally, representing an image with more tokens would lead to higher prediction accuracy, while it also results in drastically increased computational cost. To achieve a decent trade-off between accuracy and speed, the number of tokens is empirically set to 16x16 or 14x14. In this paper, we argue that every image has its own characteristics, and ideally the token number should be conditioned on each individual input. In fact, we have observed that there exist a considerable number of""easy""images which can be accurately predicted with a mere number of 4x4 tokens, while only a small fraction of""hard""ones need a finer representation. Inspired by this phenomenon, we propose a Dynamic Transformer to automatically configure a proper number of tokens for each input image. This is achieved by cascading multiple Transformers with increasing numbers of tokens, which are sequentially activated in an adaptive fashion at test time, i.e., the inference is terminated once a sufficiently confident prediction is produced. We further design efficient feature reuse and relationship reuse mechanisms across different components of the Dynamic Transformer to reduce redundant computations. Extensive empirical results on ImageNet, CIFAR-10, and CIFAR-100 demonstrate that our method significantly outperforms the competitive baselines in terms of both theoretical computational efficiency and practical inference speed. Code and pre-trained models (based on PyTorch and MindSpore) are available at https://github.com/blackfeather-wang/Dynamic-Vision-Transformer and https://github.com/blackfeather-wang/Dynamic-Vision-Transformer-MindSpore.","{'model': 'tldr@v2.0.0', 'text': 'This paper argues that every image has its own characteristics, and ideally the token number should be conditioned on each individual input, and proposes a Dynamic Transformer to automatically configure a proper number of tokens for each input image.'}",
-"efficient, real-world token bucket configuration for residential gateways",Fengyu Gao,"Internet service providers should deploy effective active queue management (AQM) strategies to provide high-bandwidth low-latency access links to customers. However, they do not, and customers see high latencies, which are problematic for latency-sensitive applications (e.g., VoIP). As a result, customers have to deploy token buckets, but they are hard to configure. In this paper, we evaluate different token bucket configurations and find that the anecdotal evidence regarding token bucket configuration is not optimal. We analyze different configurations using the ns-2 simulator and find a formula to derive optimal parameters depending on the link bandwidth, which brings about much lower latency and higher throughput.","{'model': 'tldr@v2.0.0', 'text': 'This paper evaluates different token bucket configurations and finds that the anecdotal evidence regarding token bucket configuration is not optimal, and finds a formula to derive optimal parameters depending on the link bandwidth, which brings about much lower latency and higher throughput.'}",
-sparse detr: efficient end-to-end object detection with learnable sparsity,Byungseok Roh,"DETR is the first end-to-end object detector using a transformer encoder-decoder architecture and demonstrates competitive performance but low computational efficiency on high resolution feature maps. The subsequent work, Deformable DETR, enhances the efficiency of DETR by replacing dense attention with deformable attention, which achieves 10x faster convergence and improved performance. Deformable DETR uses the multiscale feature to ameliorate performance, however, the number of encoder tokens increases by 20x compared to DETR, and the computation cost of the encoder attention remains a bottleneck. In our preliminary experiment, we observe that the detection performance hardly deteriorates even if only a part of the encoder token is updated. Inspired by this observation, we propose Sparse DETR that selectively updates only the tokens expected to be referenced by the decoder, thus help the model effectively detect objects. In addition, we show that applying an auxiliary detection loss on the selected tokens in the encoder improves the performance while minimizing computational overhead. We validate that Sparse DETR achieves better performance than Deformable DETR even with only 10% encoder tokens on the COCO dataset. Albeit only the encoder tokens are sparsified, the total computation cost decreases by 38% and the frames per second (FPS) increases by 42% compared to Deformable DETR. Code is available at https://github.com/kakaobrain/sparse-detr","{'model': 'tldr@v2.0.0', 'text': 'Sparse DETR is proposed that selectively updates only the tokens expected to be referenced by the decoder to help the model effectively detect objects and achieves better performance than Deformable DETR even with only 10% encoder tokens on the COCO dataset.'}",
-understanding and overcoming the challenges of efficient transformer quantization,Yelysei Bondarenko,"Transformer-based architectures have become the de-facto standard models for a wide range of Natural Language Processing tasks. However, their memory footprint and high latency are prohibitive for efficient deployment and inference on resource-limited devices. In this work, we explore quantization for transformers. We show that transformers have unique quantization challenges – namely, high dynamic activation ranges that are difficult to represent with a low bit fixed-point format. We establish that these activations contain structured outliers in the residual connections that encourage specific attention patterns, such as attending to the special separator token. To combat these challenges, we present three solutions based on post-training quantization and quantization-aware training, each with a different set of compromises for accuracy, model size, and ease of use. In particular, we introduce a novel quantization scheme – per-embedding-group quantization. We demonstrate the effectiveness of our methods on the GLUE benchmark using BERT, establishing state-of-the-art results for post-training quantization. Finally, we show that transformer weights and embeddings can be quantized to ultra-low bit-widths, leading to significant memory savings with a minimum accuracy loss. Our source code is available at https://github.com/qualcomm-ai-research/transformer-quantization.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces a novel quantization scheme – per-embedding-group quantization, and shows that transformer weights and embeddings can be quantized to ultra-low bit-widths, leading to significant memory savings with a minimum accuracy loss.'}",https://aclanthology.org/2021.emnlp-main.627.pdf
-beyond distillation: task-level mixture-of-experts for efficient inference,Sneha Kudugunta,"Sparse Mixture-of-Experts (MoE) has been a successful approach for scaling multilingual translation models to billions of parameters without a proportional increase in training computation. However, MoE models are prohibitively large and practitioners often resort to methods such as distillation for serving. In this work, we investigate routing strategies at different granularity (token, sentence, task) in MoE models to bypass distillation. Experiments on WMT and a web-scale dataset suggest that task-level routing (task-MoE) enables us to extract smaller, ready-to-deploy sub-networks from large sparse models. On WMT, our task-MoE with 32 experts (533M parameters) outperforms the best performing token-level MoE model (token-MoE) by +1.0 BLEU on average across 30 language pairs. The peak inference throughput is also improved by a factor of 1.9x when we route by tasks instead of tokens. While distilling a token-MoE to a smaller dense model preserves only 32% of the BLEU gains, our sub-network task-MoE, by design, preserves all the gains with the same inference cost as the distilled student model. Finally, when scaling up to 200 language pairs, our 128-expert task-MoE (13B parameters) performs competitively with a token-level counterpart, while improving the peak inference throughput by a factor of 2.6x.","{'model': 'tldr@v2.0.0', 'text': 'Experiments on WMT and a web-scale dataset suggest that task-level routing (task-MoE) enables us to extract smaller, ready-to-deploy sub-networks from large sparse models.'}",https://aclanthology.org/2021.findings-emnlp.304.pdf
-"simple, secure, efficient, lightweight and token based protocol for mutual authentication in wireless sensor networks",Reena Rathore,,"{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a simple, secure, dynamic, scalable, efficient and lightweight protocol for mutual authentication of nodes in WSN based on tokens, which was modeled in scyther for verification and no potential attacks were detected.'}",
-uplift and upsample: efficient 3d human pose estimation with uplifting transformers,Moritz Einfalt,"The state-of-the-art for monocular 3D human pose estimation in videos is dominated by the paradigm of 2D-to-3D pose uplifting. While the uplifting methods themselves are rather efficient, the true computational complexity depends on the per-frame 2D pose estimation. In this paper, we present a Transformer-based pose uplifting scheme that can operate on temporally sparse 2D pose sequences but still produce temporally dense 3D pose estimates. We show how masked token modeling can be utilized for temporal upsampling within Transformer blocks. This allows to decouple the sampling rate of input 2D poses and the target frame rate of the video and drastically decreases the total computational complexity. Additionally, we explore the option of pre-training on large motion capture archives, which has been largely neglected so far We evaluate our method on two popular benchmark datasets: Human3.6M and MPI-INF-3DHP. With an MPJPE of 45.0 mm and 46.9 mm, respectively, our proposed method can compete with the state-of-the-art while reducing inference time by a factor of 12. This enables real-time throughput with variable consumer hardware in stationary and mobile applications. We release our code and models at https://github.com/goldbricklemon/uplift-upsample-3dhpe","{'model': 'tldr@v2.0.0', 'text': 'A Transformer-based pose uplifting scheme that can operate on temporally sparse 2D pose sequences but still produce temporally dense 3D pose estimates, and shows how masked token modeling can be utilized for temporal upsampling within Transformer blocks.'}",https://arxiv.org/pdf/2210.06110
-dtb-mac: dynamic token-based mac protocol for reliable and efficient beacon broadcasting in vanets,A. Balador,"Most applications developed for vehicular environments rely on broadcasting as the main mechanism to disseminate their messages. However, in IEEE 802.11p, which is the most widely accepted Medium Access Control (MAC) protocol for vehicular communications, all transmissions remain unacknowledged if broadcasting is used. Furthermore, safety message transmission requires a strict delay limit and a high reliability, which is an issue for random access MAC protocols like IEEE 802.11p. Therefore, transmission reliability becomes the most important issue for broadcast-based services in vehicular environments. In this paper, we propose a hybrid MAC protocol, referred as Dynamic Token-Based MAC Protocol (DTB-MAC). DTB-MAC uses both a token passing mechanism and a random access MAC protocol to prevent channel contention as much as possible, and to improve the reliability of safety message transmissions. Our proposed protocol tries to select the best neighbouring node as the next transmitter, and when it is not possible, or when it causes a high overhead, the random access MAC protocol is used instead. Based on simulation experiments, we show that the DTB-MAC protocol can achieve better performance compared with IEEE 802.11p in terms of channel utilization and beacon delivery ratio.","{'model': 'tldr@v2.0.0', 'text': 'DTB-MAC uses both a token passing mechanism and a random access MAC protocol to prevent channel contention as much as possible, and to improve the reliability of safety message transmissions.'}",https://riunet.upv.es/bitstream/10251/64592/3/DTB-MAC%20%28ninth%20version%29.pdf
-uniformer: unified transformer for efficient spatiotemporal representation learning,Kunchang Li,"It is a challenging task to learn rich and multi-scale spatiotemporal semantics from high-dimensional videos, due to large local redundancy and complex global dependency between video frames. The recent advances in this research have been mainly driven by 3D convolutional neural networks and vision transformers. Although 3D convolution can efficiently aggregate local context to suppress local redundancy from a small 3D neighborhood, it lacks the capability to capture global dependency because of the limited receptive field. Alternatively, vision transformers can effectively capture long-range dependency by self-attention mechanism, while having the limitation on reducing local redundancy with blind similarity comparison among all the tokens in each layer. Based on these observations, we propose a novel Unified transFormer (UniFormer) which seamlessly integrates merits of 3D convolution and spatiotemporal self-attention in a concise transformer format, and achieves a preferable balance between computation and accuracy. Different from traditional transformers, our relation aggregator can tackle both spatiotemporal redundancy and dependency, by learning local and global token affinity respectively in shallow and deep layers. We conduct extensive experiments on the popular video benchmarks, e.g., Kinetics-400, Kinetics-600, and Something-Something V1&V2. With only ImageNet-1K pretraining, our UniFormer achieves 82.9%/84.8% top-1 accuracy on Kinetics-400/Kinetics-600, while requiring 10x fewer GFLOPs than other state-of-the-art methods. For Something-Something V1 and V2, our UniFormer achieves new state-of-the-art performances of 60.9% and 71.2% top-1 accuracy respectively. Code is available at https://github.com/Sense-X/UniFormer.","{'model': 'tldr@v2.0.0', 'text': 'A novel Unified transFormer (UniFormer) which seamlessly integrates merits of 3D convolution and spatiotemporal self-attention in a concise transformer format, and achieves a preferable balance between computation and accuracy.'}",
-attention fusion: a light yet efficient late fusion mechanism for task adaptation in nlu,Jin Cao,"Fine-tuning a pre-trained language model using annotated data has become the de-facto standard for adapting general-purpose pre-trained models like BERT to downstream tasks. However, given the trend of larger pre-trained models, fine-tuning these models for each downstream task is parameter-inefficient and computationally-expensive deeming this approach sub-optimal for adoption by NLU systems. In recent years, various approaches have been proposed for parameter efficient task adaptation such as Adaptor, Bitfit, Prompt tuning, Prefix tuning etc. However, most of these efforts propose to insert task specific parameters in-between or inside intermediate layers of the pre-trained encoder resulting in higher computational cost due to back-propagation of errors to all layers. To mitigate this issue, we propose a light but efficient, attention based fusion module which computes task-attuned token representations by aggregating intermediate layer representations from a pre-trained network. Our proposed fusion module trains only 0.0009% of total parameters and achieves competitive performance to the standard fine-tuning approach on various tasks. It is also decoupled from the pre-trained network making it efficient during computation and scalable during deployment. Last but not the least, we demonstrate that our proposed attention-fusion mechanism can transfer effectively to different languages for further re-use and expansion.","{'model': 'tldr@v2.0.0', 'text': 'A light but efficient, attention based fusion module which computes task-attuned token representations by aggregating intermediate layer representations from a pre-trained network and can transfer effectively to different languages for further re-use and expansion.'}",
-efficient grid exploration with a stationary token,A. Pelc,"A mobile agent starting at an arbitrary node of an m × k grid, for 1 < m ≤ k, has to explore the grid by visiting all its nodes and traversing all edges. The cost of an exploration algorithm is the number of edge traversals by the agent. Nodes of the grid are unlabeled and ports at each node v have distinct numbers in {0,…, d − 1}, where d = 2, 3, 4 is the degree of v. Port numbering is local, i.e., there is no relation between port numbers at different nodes. When visiting a node the agent sees its degree. It also sees the port number by which it enters a node and can choose the port number by which it leaves a visited node. We are interested in deterministic exploration algorithms working at low cost. We consider the scenario in which the agent is equipped with a stationary token situated at its starting node. The agent sees the token whenever it visits this node. We give an exploration algorithm working at cost O(k2) for 2 × k grids, and at cost O(m2k), for m × k grids, when 2 < m ≤ k.","{'model': 'tldr@v2.0.0', 'text': 'This work considers the scenario in which the agent is equipped with a stationary token situated at its starting node and gives an exploration algorithm working at cost O(k2) for 2 × k grids, and at Cost O(m2k), for m ×k grids, when 2 < m ≤ k.'}",
-fuzzy-token: an adaptive mac protocol for wireless-enabled manycores,Antonio Franques,"Recent computer architecture trends herald the arrival of manycores with over one hundred cores on a single chip. In this context, traditional on-chip networks do not scale well in latency or energy consumption, leading to bottlenecks in the execution. The Wireless Network-on-Chip (WNoC) paradigm holds considerable promise for the implementation of on-chip networks that will enable such highly-parallel manycores. However, one of the main challenges in WNoCs is the design of mechanisms that provide fast and efficient access to the wireless channel, while adapting to the changing traffic patterns within and across applications. Existing approaches are either slow or complicated, and do not provide the required adaptivity. In this paper, we propose Fuzzy Token,a simple WNoC protocol that leverages the unique properties of the on-chip scenario to deliver efficient and low-latency access to the wireless channel irrespective of the application characteristics. We substantiate our claim via simulations with a synthetic traffic suite and with real application traces. Fuzzy Tokenconsistently provides one of the lowest packet latencies among the evaluated WNoC MAC protocols. On average, the packet latency in Fuzzy Token is 4.4 × and 2.6 × lower than in a state-of-the art contention-based WNoC MAC protocol and in a token-passing protocol, respectively.","{'model': 'tldr@v2.0.0', 'text': 'Fuzzy Token is proposed, a simple WNoC protocol that leverages the unique properties of the on-chip scenario to deliver efficient and low-latency access to the wireless channel irrespective of the application characteristics, and consistently provides one of the lowest packet latencies among the evaluated W nocturnal MAC protocols.'}",https://upcommons.upc.edu/bitstream/2117/349369/5/Fuzzy_Token_DATE_2021.pdf
-swapct: swap confidential transactions for privacy-preserving multi-token exchanges,Felix Engelmann,"Abstract Decentralized token exchanges allow for secure trading of tokens without a trusted third party. However, decentralization is mostly achieved at the expense of transaction privacy. For a fair exchange, transactions must remain private to hide the participants and volumes while maintaining the possibility for noninteractive execution of trades. In this paper we present a swap confidential transaction system (SwapCT) which is related to ring confidential transactions (e.g. used in Monero) but supports multiple token types to trade among and enables secure, partial transactions for noninteractive swaps. We prove that SwapCT is secure in a strict, formal model and present its efficient performance in a prototype implementation with logarithmic signature sizes for large anonymity sets. For our construction we design an aggregatable signature scheme which might be of independent interest. Our SwapCT system thereby enables a secure and private exchange for tokens without a trusted third party.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that SwapCT is secure in a strict, formal model and its efficient performance in a prototype implementation with logarithmic signature sizes for large anonymity sets is presented.'}",
-token-based authentication framework for 5g mec mobile networks,Wojciech Niewolski,"MEC technology provides a distributed computing environment in 5G mobile networks for application and service hosting. It allows customers with different requirements and professional competencies to use the services offered by external suppliers. We consider a service access control framework on 5G MEC networks that is efficient, flexible, and user-friendly. Its central element is the MEC Enabler, which handles AAA requests for stakeholders accessing services hosted on the edge servers. The JSON Web Token (JWT) open standard is a suitable tool for the MEC Enabler to manage access control credentials and transfer them securely between parties. In this paper, in the context of access control, we propose the token reference pattern called JSON MEC Access Token (JMAT) and analyze the effectiveness of its available protection methods in compliance with the standard requirements of MEC-hosted services in 5G networks.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes the token reference pattern called JSON MEC Access Token (JMAT) and analyzes the effectiveness of its available protection methods in compliance with the standard requirements of MEC-hosted services in 5G networks.'}",https://www.mdpi.com/2079-9292/10/14/1724/pdf?version=1626755715
-cross-token modeling with conditional computation,Yuxuan Lou,"Mixture-of-Experts (MoE), a conditional computation architecture, achieved promising performance by scaling local module (i.e. feed-forward network) of transformer. However, scaling the cross-token module (i.e. self-attention) is challenging due to the unstable training. This work proposes Sparse-MLP, an all-MLP model which applies sparsely-activated MLPs to cross-token modeling. Specifically, in each Sparse block of our all-MLP model, we apply two stages of MoE layers: one with MLP experts mixing information within channels along image patch dimension, the other with MLP experts mixing information within patches along the channel dimension. In addition, by proposing importance-score routing strategy for MoE and redesigning the image representation shape, we further improve our model's computational efficiency. Experimentally, we are more computation-efficient than Vision Transformers with comparable accuracy. Also, our models can outperform MLP-Mixer by 2.5\% on ImageNet Top-1 accuracy with fewer parameters and computational cost. On downstream tasks, i.e. Cifar10 and Cifar100, our models can still achieve better performance than baselines.","{'model': 'tldr@v2.0.0', 'text': ""This work proposes Sparse-MLP, an all- MLP model which applies sparsely-activated MLPs to cross-token modeling and improves the model's computational efficiency by proposing importance-score routing strategy for MoE and redesigning the image representation shape.""}",
-learning token-based representation for image retrieval,Hui Wu,"In image retrieval, deep local features learned in a data-driven manner have been demonstrated effective to improve retrieval performance. To realize efficient retrieval on large image database, some approaches quantize deep local features with a large codebook and match images with aggregated match kernel. However, the complexity of these approaches is non-trivial with large memory footprint, which limits their capability to jointly perform feature learning and aggregation. To generate compact global representations while maintaining regional matching capability, we propose a unified framework to jointly learn local feature representation and aggregation. In our framework, we first extract local features using CNNs. Then, we design a tokenizer module to aggregate them into a few visual tokens, each corresponding to a specific visual pattern. This helps to remove background noise, and capture more discriminative regions in the image. Next, a refinement block is introduced to enhance the visual tokens with self-attention and cross-attention. Finally, different visual tokens are concatenated to generate a compact global representation. The whole framework is trained end-to-end with image-level labels. Extensive experiments are conducted to evaluate our approach, which outperforms the state-of-the-art methods on the Revisited Oxford and Paris datasets.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a unified framework to jointly learn local feature representation and aggregation, which outperforms the state-of-the-art methods on the Revisited Oxford and Paris datasets.'}",https://ojs.aaai.org/index.php/AAAI/article/download/20173/19932
-token merging for fast stable diffusion,Daniel Bolya,"The landscape of image generation has been forever changed by open vocabulary diffusion models. However, at their core these models use transformers, which makes generation slow. Better implementations to increase the throughput of these transformers have emerged, but they still evaluate the entire model. In this paper, we instead speed up diffusion models by exploiting natural redundancy in generated images by merging redundant tokens. After making some diffusion-specific improvements to Token Merging (ToMe), our ToMe for Stable Diffusion can reduce the number of tokens in an existing Stable Diffusion model by up to 60% while still producing high quality images with-out any extra training. In the process, we speed up image generation by up to 2× and reduce memory consumption by up to 5.6×. Furthermore, this speed-up stacks with efficient implementations such as xFormers, minimally impacting quality while being up to 5.4× faster for large images. Code is available at https://github.com/dbolya/tomesd.","{'model': 'tldr@v2.0.0', 'text': 'This paper speed up diffusion models by exploiting natural redundancy in generated images by merging redundant tokens, and can reduce the number of tokens in an existing Stable Diffusion model by up to 60% while still producing high quality images with-out any extra training.'}",https://arxiv.org/pdf/2303.17604
-non-human primate token use shows possibilities but also limitations for establishing a form of currency,M. Beran,"Non-human primates evaluate choices based on quantitative information and subjective valuation of options. Non-human primates can learn to value tokens as placeholders for primary rewards (such as food). With those tokens established as a potential form of ‘currency’, it is then possible to examine how they respond to opportunities to earn and use tokens in ways such as accumulating tokens or exchanging tokens with each other or with human experimenters to gain primary rewards. Sometimes, individuals make efficient and beneficial choices to obtain tokens and then exchange them at the right moments to gain optimal reward. Sometimes, they even accumulate such rewards through extended delay of gratification, or through other exchange-based interactions. Thus, non-human primates are capable of associating value to arbitrary tokens that may function as currency-like stimuli, but there also are strong limitations on how non-human primates can integrate such tokens into choice situations or use such tokens to fully ‘symbolize’ economic decision-making. These limitations are important to acknowledge when considering the evolutionary emergence of currency use in our species. This article is part of the theme issue ‘Existence and prevalence of economic behaviours among non-human primates’.","{'model': 'tldr@v2.0.0', 'text': ""There are strong limitations on how non-human primates can integrate such tokens into choice situations or use such tokens to fully ‘symbolize’ economic decision-making when considering the evolutionary emergence of currency use in the authors' species.""}",https://royalsocietypublishing.org/doi/pdf/10.1098/rstb.2019.0675
-2dmac: a sustainable and efficient medium access control mechanism for future wireless nocs,Sidhartha Sankar Rout,"Wireless Network-on-Chip (WNoC) requires a Medium Access Control (MAC) mechanism for an interference-free sharing of the wireless channel. In traditional MAC, a token is circulated among the Wireless Interfaces (WIs) in a Round Robin manner. The WI with the token holds the channel for a fixed number of cycles. However, the channel requirement of the individual WIs dynamically changes over time due to the varying traffic density across the WNoC. Moreover, the conventional WNoCs give equal importance to all the traffic taking the wireless path and transmit it in an oldest-first manner. Nevertheless, the critical data can degrade the system performance to a large extent by delaying the application runtime if not served promptly. We propose 2DMAC, which can change the token arbitration pattern and tune the channel hold time of each WI based on its runtime traffic density and criticality status. Moreover, 2DMAC prioritizes the critical traffic over the non-critical traffic during the wireless data transfer. The proposed mechanism improves the wireless channel utilization by 15.67% and the network throughput by 29.83% and reduces the critical data latency by 29.77% over the traditional MAC.","{'model': 'tldr@v2.0.0', 'text': '2DMAC is proposed, which can change the token arbitration pattern and tune the channel hold time of each WI based on its runtime traffic density and criticality status, and prioritizes the critical traffic over the non-critical traffic during the wireless data transfer.'}",
-ppt: token-pruned pose transformer for monocular and multi-view human pose estimation,Haoyu Ma,"Recently, the vision transformer and its variants have played an increasingly important role in both monocular and multi-view human pose estimation. Considering image patches as tokens, transformers can model the global dependencies within the entire image or across images from other views. However, global attention is computationally expensive. As a consequence, it is difficult to scale up these transformer-based methods to high-resolution features and many views. In this paper, we propose the token-Pruned Pose Transformer (PPT) for 2D human pose estimation, which can locate a rough human mask and performs self-attention only within selected tokens. Furthermore, we extend our PPT to multi-view human pose estimation. Built upon PPT, we propose a new cross-view fusion strategy, called human area fusion, which considers all human foreground pixels as corresponding candidates. Experimental results on COCO and MPII demonstrate that our PPT can match the accuracy of previous pose transformer methods while reducing the computation. Moreover, experiments on Human 3.6M and Ski-Pose demonstrate that our Multi-view PPT can efficiently fuse cues from multiple views and achieve new state-of-the-art results.","{'model': 'tldr@v2.0.0', 'text': 'The token-Pruned Pose Transformer (PPT) is proposed, which can locate a rough human mask and performs self-attention only within selected tokens and is extended to multi-view human pose estimation and a new cross-view fusion strategy, called human area fusion, which considers all human foreground pixels as corresponding candidates.'}",http://arxiv.org/pdf/2209.08194
-mixmim: mixed and masked image modeling for efficient visual representation learning,Jihao Liu,"In this study, we propose Mixed and Masked Image Modeling (MixMIM), a simple but efficient MIM method that is applicable to various hierarchical Vision Transformers. Existing MIM methods replace a random subset of input tokens with a special [MASK] symbol and aim at reconstructing original image tokens from the corrupted image. However, we find that using the [MASK] symbol greatly slows down the training and causes training-finetuning inconsistency, due to the large masking ratio (e.g., 40% in BEiT). In contrast, we replace the masked tokens of one image with visible tokens of another image, i.e., creating a mixed image. We then conduct dual reconstruction to reconstruct the original two images from the mixed input, which significantly improves efficiency. While MixMIM can be applied to various architectures, this paper explores a simpler but stronger hierarchical Transformer, and scales with MixMIM-B, -L, and -H. Empirical results demonstrate that MixMIM can learn high-quality visual representations efficiently. Notably, MixMIM-B with 88M parameters achieves 85.1% top-1 accuracy on ImageNet-1K by pretraining for 600 epochs, setting a new record for neural networks with comparable model sizes (e.g., ViT-B) among MIM methods. Besides, its transferring performances on the other 6 datasets show MixMIM has better FLOPs / performance tradeoff than previous MIM methods. Code is available at https://github.com/Sense-X/MixMIM.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes Mixed and Masked Image Modeling (MixMIM), a simple but efficient MIM method that is applicable to various hierarchical Vision Transformers, and scales with MixMIM-B, -L, and -H.'}",http://arxiv.org/pdf/2205.13137
-attempt: parameter-efficient multi-task tuning via attentional mixtures of soft prompts,Akari Asai,"This work introduces a new multi-task, parameter-efficient language model (LM) tuning method that learns to transfer knowledge across different tasks via a mixture of soft prompts—small prefix embedding vectors pre-trained for different tasks. Our method, called ATTEMPT (ATTEntional Mixtures of Prompt Tuning), obtains source prompts as encodings of large-scale source tasks into a small number of parameters and trains an attention module to interpolate the source prompts and a newly initialized target prompt for every instance in the target task. During training, only the target task prompt and the attention weights, which are shared between tasks in multi-task training, are updated, while the original LM and source prompts are intact. ATTEMPT is highly parameter-efficient (e.g., updates 2,300 times fewer parameters than full fine-tuning), while it overcomes instability of prompt tuning and achieves high task performance using learned knowledge from high-resource tasks. Moreover, it is modular using pre-trained soft prompts, and can flexibly add or remove source prompts for effective knowledge transfer. Our experimental results across 21 diverse NLP datasets show that ATTEMPT significantly outperforms prompt tuning and outperforms or matches fully fine-tuned or other parameter-efficient tuning approaches that use 10 times more parameters. Finally, ATTEMPT outperforms previous work in few-shot learning settings.","{'model': 'tldr@v2.0.0', 'text': 'A new multi-task, parameter-efficient language model (LM) tuning method that learns to transfer knowledge across different tasks via a mixture of soft prompts—small prefix embedding vectors pre-trained for different tasks that significantly outperforms prompt tuning and outperforms or matches fully fine-tuned or other parameter- efficient tuning approaches that use 10 times more parameters.'}",https://aclanthology.org/2022.emnlp-main.446.pdf
-data efficient masked language modeling for vision and language,Yonatan Bitton,"Masked language modeling (MLM) is one of the key sub-tasks in vision-language pretraining. In the cross-modal setting, tokens in the sentence are masked at random, and the model predicts the masked tokens given the image and the text. In this paper, we observe several key disadvantages of MLM in this setting. First, as captions tend to be short, in a third of the sentences no token is sampled. Second, the majority of masked tokens are stop-words and punctuation, leading to under-utilization of the image. We investigate a range of alternative masking strategies specific to the cross-modal setting that address these shortcomings, aiming for better fusion of text and image in the learned representation. When pre-training the LXMERT model, our alternative masking strategies consistently improve over the original masking strategy on three downstream tasks, especially in low resource settings. Further, our pre-training approach substantially outperforms the baseline model on a prompt-based probing task designed to elicit image objects. These results and our analysis indicate that our method allows for better utilization of the training data.","{'model': 'tldr@v2.0.0', 'text': 'This paper investigates a range of alternative masking strategies specific to the cross-modal setting that address shortcomings of MLM, aiming for better fusion of text and image in the learned representation.'}",https://aclanthology.org/2021.findings-emnlp.259.pdf
-effective structured prompting by meta-learning and representative verbalizer,Weisen Jiang,"Prompt tuning for pre-trained masked language models (MLM) has shown promising performance in natural language processing tasks with few labeled examples. It tunes a prompt for the downstream task, and a verbalizer is used to bridge the predicted token and label prediction. Due to the limited training data, prompt initialization is crucial for prompt tuning. Recently, MetaPrompting (Hou et al., 2022) uses meta-learning to learn a shared initialization for all task-specific prompts. However, a single initialization is insufficient to obtain good prompts for all tasks and samples when the tasks are complex. Moreover, MetaPrompting requires tuning the whole MLM, causing a heavy burden on computation and memory as the MLM is usually large. To address these issues, we use a prompt pool to extract more task knowledge and construct instance-dependent prompts via attention. We further propose a novel soft verbalizer (RepVerb) which constructs label embedding from feature embeddings directly. Combining meta-learning the prompt pool and RepVerb, we propose MetaPrompter for effective structured prompting. MetaPrompter is parameter-efficient as only the pool is required to be tuned. Experimental results demonstrate that MetaPrompter performs better than the recent state-of-the-arts and RepVerb outperforms existing soft verbalizers.","{'model': 'tldr@v2.0.0', 'text': 'A prompt pool is used to extract more task knowledge and construct instance-dependent prompts via attention and a novel soft verbalizer (RepVerb) is proposed which constructs label embedding from feature embeddings directly and is parameter-efficient as only the pool is required to be tuned.'}",http://arxiv.org/pdf/2306.00618
-efficient pre-training objectives for transformers,Luca Di Liello,"The Transformer architecture deeply changed the natural language processing, outperforming all previous state-of-the-art models. However, well-known Transformer models like BERT, RoBERTa, and GPT-2 require a huge compute budget to create a high quality contextualised representation. In this paper, we study several efficient pre-training objectives for Transformers-based models. By testing these objectives on different tasks, we determine which of the ELECTRA model's new features is the most relevant. We confirm that Transformers pre-training is improved when the input does not contain masked tokens and that the usage of the whole output to compute the loss reduces training time. Moreover, inspired by ELECTRA, we study a model composed of two blocks; a discriminator and a simple generator based on a statistical model with no impact on the computational performances. Besides, we prove that eliminating the MASK token and considering the whole output during the loss computation are essential choices to improve performance. Furthermore, we show that it is possible to efficiently train BERT-like models using a discriminative approach as in ELECTRA but without a complex generator, which is expensive. Finally, we show that ELECTRA benefits heavily from a state-of-the-art hyper-parameters search.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that eliminating the MASK token and considering the whole output during the loss computation are essential choices to improve performance and it is shown that ELECTRA benefits heavily from a state-of-the-art hyper-parameters search.'}",
-eetbr: energy efficient token-based routing for wireless sensor networks,Taner Çevik,"The most significant drawback of wireless sensor networks is energy scarcity. As there is an increasing need for operating these networks for relatively long times, energy saving becomes the key challenge in the design of the architectures and protocols for sensor networks. Therefore, several research studies have been performed for making contributions to the analysis of this energy shortage problem. Most of these research activities have been focused on finding solutions for the energy consumption of the communication unit, which is the dominant energy dissipating component of the sensor nodes. In this paper, a novel, token-based routing protocol adapted with a multitier cluster-based architecture is presented. Most of the other cluster-based schemes mainly focus on intracluster organization and communication. However, it should be mentioned that a considerable amount of energy is dissipated during the intercluster communication when compared with intracluster communication. The architecture proposed here not only deals with intracluster communication, but also considers data aggregation, multihop data transmission, and best-effort next hop selection according to a cost factor that is described for the first time in this paper. The simulation results indicate that this token-based next hop selection method together with the multitier cluster-based architecture achieves a significant amount of energy savings, which inherently yields the prolongation of the network lifetime.","{'model': 'tldr@v2.0.0', 'text': 'A novel, token-based routing protocol adapted with a multitier cluster-based architecture that not only deals with intracluster communication, but also considers data aggregation, multihop data transmission, and best-effort next hop selection according to a cost factor is presented.'}",https://journals.tubitak.gov.tr/cgi/viewcontent.cgi?article=3199&context=elektrik
-llama: open and efficient foundation language models,Hugo Touvron,"We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community.","{'model': 'tldr@v2.0.0', 'text': 'LLaMA, a collection of foundation language models ranging from 7B to 65B parameters, is introduced and it is shown that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets.'}",http://arxiv.org/pdf/2302.13971
-funnel-transformer: filtering out sequential redundancy for efficient language processing,Zihang Dai,"With the success of language pretraining, it is highly desirable to develop more efficient architectures of good scalability that can exploit the abundant unlabeled data at a lower cost. To improve the efficiency, we examine the much-overlooked redundancy in maintaining a full-length token-level presentation, especially for tasks that only require a single-vector presentation of the sequence. With this intuition, we propose Funnel-Transformer which gradually compresses the sequence of hidden states to a shorter one and hence reduces the computation cost. More importantly, by re-investing the saved FLOPs from length reduction in constructing a deeper or wider model, we further improve the model capacity. In addition, to perform token-level predictions as required by common pretraining objectives, Funnel-Transformer is able to recover a deep representation for each token from the reduced hidden sequence via a decoder. Empirically, with comparable or fewer FLOPs, Funnel-Transformer outperforms the standard Transformer on a wide variety of sequence-level prediction tasks, including text classification, language understanding, and reading comprehension. The code and pretrained checkpoints are available at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Funnel-Transformer, a model which gradually compresses the sequence of hidden states to a shorter one and hence reduces the computation cost and outperforms the standard Transformer on a wide variety of sequence-level prediction tasks.'}",
-prompting classes: exploring the power of prompt class learning in weakly supervised semantic segmentation,Balamurali Murugesan,"Recently, CLIP-based approaches have exhibited remarkable performance on generalization and few-shot learning tasks, fueled by the power of contrastive language-vision pre-training. In particular, prompt tuning has emerged as an effective strategy to adapt the pre-trained language-vision models to downstream tasks by employing task-related textual tokens. Motivated by this progress, in this work we question whether other fundamental problems, such as weakly supervised semantic segmentation (WSSS), can benefit from prompt tuning. Our findings reveal two interesting observations that shed light on the impact of prompt tuning on WSSS. First, modifying only the class token of the text prompt results in a greater impact on the Class Activation Map (CAM), compared to arguably more complex strategies that optimize the context. And second, the class token associated with the image ground truth does not necessarily correspond to the category that yields the best CAM. Motivated by these observations, we introduce a novel approach based on a PrOmpt cLass lEarning (POLE) strategy. Through extensive experiments we demonstrate that our simple, yet efficient approach achieves SOTA performance in a well-known WSSS benchmark. These results highlight not only the benefits of language-vision models in WSSS but also the potential of prompt learning for this problem. The code is available at https://github.com/rB080/WSS_POLE.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach based on a PrOmpt cLass lEarning (POLE) strategy achieves SOTA performance in a well-known WSSS benchmark and highlights not only the benefits of language-vision models in W SSS but also the potential of prompt learning for this problem.'}",https://arxiv.org/pdf/2307.00097
-rewoo: decoupling reasoning from observations for efficient augmented language models,Binfeng Xu,"Augmented Language Models (ALMs) blend the reasoning capabilities of Large Language Models (LLMs) with tools that allow for knowledge retrieval and action execution. Existing ALM systems trigger LLM thought processes while pulling observations from these tools in an interleaved fashion. Specifically, an LLM reasons to call an external tool, gets halted to fetch the tool's response, and then decides the next action based on all preceding response tokens. Such a paradigm, though straightforward and easy to implement, often leads to huge computation complexity from redundant prompts and repeated execution. This study addresses such challenges for the first time, proposing a modular paradigm ReWOO (Reasoning WithOut Observation) that detaches the reasoning process from external observations, thus significantly reducing token consumption. Comprehensive evaluations across six public NLP benchmarks and a curated dataset reveal consistent performance enhancements with our proposed methodology. Notably, ReWOO achieves 5x token efficiency and 4% accuracy improvement on HotpotQA, a multi-step reasoning benchmark. Furthermore, ReWOO demonstrates robustness under tool-failure scenarios. Beyond prompt efficiency, decoupling parametric modules from non-parametric tool calls enables instruction fine-tuning to offload LLMs into smaller language models, thus substantially reducing model parameters. Our illustrative work offloads reasoning ability from 175B GPT3.5 into 7B LLaMA, demonstrating the significant potential for truly efficient and scalable ALM systems.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a modular paradigm ReWOO (Reasoning WithOut Observation) that detaches the reasoning process from external observations, thus significantly reducing token consumption and demonstrating robustness under tool-failure scenarios.'}",http://arxiv.org/pdf/2305.18323
-efficient domain adaptation of language models via adaptive tokenization,Vin Sachidananda,"Contextual embedding-based language models trained on large data sets, such as BERT and RoBERTa, provide strong performance across a wide range of tasks and are ubiquitous in modern NLP. It has been observed that fine-tuning these models on tasks involving data from domains different from that on which they were pretrained can lead to suboptimal performance. Recent work has explored approaches to adapt pretrained language models to new domains by incorporating additional pretraining on domain-specific corpora and task data. We propose an alternative approach for transferring pretrained language models to new domains by adapting their tokenizers. We show that domain-specific subword sequences can be determined efficiently directly from divergences in the conditional token distributions of the base and domain-specific corpora. In datasets from four disparate domains, we find adaptive tokenization on a pretrained RoBERTa model provides greater than 85% of the performance benefits of domain specific pretraining. Our approach produces smaller models and less training and inference time than other approaches using tokenizer augmentation. Although using adaptive tokenization incurs a 6% increase in model parameters (due to the introduction of 10k new domain-specific tokens), our approach, using 64 CPUs, is >72x faster than further pretraining the language model on domain-specific corpora on 8 TPUs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an alternative approach for transferring pretrained language models to new domains by adapting their tokenizers and shows that domain-specific subword sequences can be determined efficiently directly from divergences in the conditional token distributions of the base and domain- specific corpora.'}",https://aclanthology.org/2021.sustainlp-1.16.pdf
-lightweight composite re-ranking for efficient keyword search with bert,Yingrui Yang,"Recently transformer-based ranking models have been shown to deliver high relevance for document search and the relevance-efficiency tradeoff becomes important for fast query response times. This paper presents BECR (BERT-based Composite Re-Ranking), a lightweight composite re-ranking scheme that combines deep contextual token interactions and traditional lexical term-matching features. BECR conducts query decomposition and composes a query representation using pre-computable token embeddings based on uni-grams and skip-n-grams, to seek a tradeoff of inference efficiency and relevance. Thus it does not perform expensive transformer computations during online inference, and does not require the use of GPU. This paper describes an evaluation of relevance and efficiency of BECR with several TREC datasets.","{'model': 'tldr@v2.0.0', 'text': 'BECR (BERT-based Composite Re-Ranking), a lightweight composite re-ranking scheme that combines deep contextual token interactions and traditional lexical term-matching features, is presented and an evaluation of relevance and efficiency of BECR with several TREC datasets is described.'}",https://dl.acm.org/doi/pdf/10.1145/3488560.3498495
-parameter-efficient low-resource dialogue state tracking by prompt tuning,Mingyu Derek Ma,"Dialogue state tracking (DST) is an important step in dialogue management to keep track of users' beliefs. Existing works fine-tune all language model (LM) parameters to tackle the DST task, which requires significant data and computing resources for training and hosting. The cost grows exponentially in the real-world deployment where dozens of fine-tuned LM are used for different domains and tasks. To reduce parameter size and better utilize cross-task shared information, we propose to use soft prompt token embeddings to learn task properties. Without tuning LM parameters, our method drastically reduces the number of parameters needed to less than 0.5% of prior works while achieves better low-resource DST performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes to use soft prompt token embeddings to learn task properties and drastically reduces the number of parameters needed to less than 0.5% of prior works while achieves better low-resource DST performance.'}",http://arxiv.org/pdf/2301.10915
-uniex: an effective and efficient framework for unified information extraction via a span-extractive perspective,Junyu Lu,"We propose a new paradigm for universal information extraction (IE) that is compatible with any schema format and applicable to a list of IE tasks, such as named entity recognition, relation extraction, event extraction and sentiment analysis. Our approach converts the text-based IE tasks as the token-pair problem, which uniformly disassembles all extraction targets into joint span detection, classification and association problems with a unified extractive framework, namely UniEX. UniEX can synchronously encode schema-based prompt and textual information, and collaboratively learn the generalized knowledge from pre-defined information using the auto-encoder language models. We develop a traffine attention mechanism to integrate heterogeneous factors including tasks, labels and inside tokens, and obtain the extraction target via a scoring matrix. Experiment results show that UniEX can outperform generative universal IE models in terms of performance and inference-speed on 14 benchmarks IE datasets with the supervised setting. The state-of-the-art performance in low-resource scenarios also verifies the transferability and effectiveness of UniEX.","{'model': 'tldr@v2.0.0', 'text': 'This work develops a traffine attention mechanism to integrate heterogeneous factors including tasks, labels and inside tokens, and obtain the extraction target via a scoring matrix, and shows that UniEX can outperform generative universal IE models in terms of performance and inference-speed.'}",http://arxiv.org/pdf/2305.10306
-prompt-tuning in asr systems for efficient domain-adaptation,Saket Dingliwal,"Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains. Since domain-specific systems perform better than their generic counterparts on in-domain evaluation, the need for memory and compute-efficient domain adaptation is obvious. Particularly, adapting parameter-heavy transformer-based language models used for rescoring ASR hypothesis is challenging. In this work, we overcome the problem using prompt-tuning, a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain. With just a handful of extra parameters per domain, we achieve much better perplexity scores over the baseline of using an unadapted LM. Despite being parameter-efficient, these improvements are comparable to those of fully-fine-tuned models with hundreds of millions of parameters. We replicate our findings in perplexity numbers to Word Error Rate in a domain-specific ASR system for one such domain.","{'model': 'tldr@v2.0.0', 'text': None}",
-prompt-free and efficient language model fine-tuning anonymous acl submission,Morgan Funtowicz,"Current methods for few-shot fine-tuning of 001 pretrained masked language model (PLM) require 002 carefully engineered prompts and verbalizers 003 for each new task, to convert examples into a 004 cloze-format that the PLM can score. In this work, 005 we propose PERFECT, a simple and efficient 006 method for few-shot fine-tuning of PLMs without 007 relying on any such handcrafting, which is highly 008 effective given as few as 32 data points. PERFECT 009 makes two key design choices: First, we show 010 that manually engineered task prompts can be 011 replaced with task-specific adapters that enable 012 sample-efficient fine-tuning and reduce memory 013 and storage costs by roughly factors of 5 and 100, 014 respectively. Second, instead of using handcrafted 015 verbalizers, we learn a new multi-token label em016 bedding during fine-tuning which are not tied to 017 the model vocabulary and which allow us to avoid 018 complex auto-regressive decoding. These embed019 dings are not only learnable from limited data but 020 also enable nearly 100x faster training and infer021 ence. Experiments on a wide range of few shot 022 NLP tasks demonstrate that PERFECT, while be023 ing simple and efficient, also outperforms existing 024 state-of-the-art few-shot learning methods.1 025","{'model': 'tldr@v2.0.0', 'text': 'Experiments on a wide range of few shot NLP tasks demonstrate that PERFECT, while be023 ing simple and efficient, also outperforms existing 024 state-of-the-art few-shot learning methods.'}",
-fine-tuned clip models are efficient video learners,H. Rasheed,"Large-scale multi-modal training with image-text pairs imparts strong generalization to CLIP model. Since training on a similar scale for videos is infeasible, recent approaches focus on the effective transfer of image-based CLIP to the video domain. In this pursuit, new parametric modules are added to learn temporal information and inter-frame relationships which require meticulous design efforts. Furthermore, when the resulting models are learned on videos, they tend to overfit on the given task distribution and lack in generalization aspect. This begs the following question: How to effectively transfer image-level CLIP representations to videos? In this work, we show that a simple Video Fine-tuned CLIP (ViFi-CLIP) baseline is generally sufficient to bridge the domain gap from images to videos. Our qualitative analysis illustrates that the frame-level processing from CLIP image-encoder followed by feature pooling and similarity matching with corresponding text embeddings helps in implicitly modeling the temporal cues within ViFi-CLIP. Such fine-tuning helps the model to focus on scene dynamics, moving objects and inter-object relationships. For low-data regimes where full fine-tuning is not viable, we propose a ‘bridge and prompt’ approach that first uses fine-tuning to bridge the domain gap and then learns prompts on language and vision side to adapt CLIP representations. We extensively evaluate this simple yet strong baseline on zero-shot, base-to-novel generalization, few-shot and fully supervised settings across five video benchmarks. Our code and pre-trained models are available at https://github.com/muzairkhattak/ViFi-CLIP.","{'model': 'tldr@v2.0.0', 'text': None}",https://arxiv.org/pdf/2212.03640
-efficient second-order treecrf for neural dependency parsing,Yu Zhang,"In the deep learning (DL) era, parsing models are extremely simplified with little hurt on performance, thanks to the remarkable capability of multi-layer BiLSTMs in context representation. As the most popular graph-based dependency parser due to its high efficiency and performance, the biaffine parser directly scores single dependencies under the arc-factorization assumption, and adopts a very simple local token-wise cross-entropy training loss. This paper for the first time presents a second-order TreeCRF extension to the biaffine parser. For a long time, the complexity and inefficiency of the inside-outside algorithm hinder the popularity of TreeCRF. To address this issue, we propose an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation. Experiments and analysis on 27 datasets from 13 languages clearly show that techniques developed before the DL era, such as structural learning (global TreeCRF loss) and high-order modeling are still useful, and can further boost parsing performance over the state-of-the-art biaffine parser, especially for partially annotated training data. We release our code at https://github.com/yzhangcs/crfpar.","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a second-order TreeCRF extension to the biaffine parser, and proposes an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation.'}",https://www.aclweb.org/anthology/2020.acl-main.302.pdf
-efficient document re-ranking for transformers by precomputing term representations,Sean MacAvaney,"Deep pretrained transformer networks are effective at various ranking tasks, such as question answering and ad-hoc document ranking. However, their computational expenses deem them cost-prohibitive in practice. Our proposed approach, called PreTTR (Precomputing Transformer Term Representations), considerably reduces the query-time latency of deep transformer networks (up to a 42x speedup on web document ranking) making these networks more practical to use in a real-time ranking scenario. Specifically, we precompute part of the document term representations at indexing time (without a query), and merge them with the query representation at query time to compute the final ranking score. Due to the large size of the token representations, we also propose an effective approach to reduce the storage requirement by training a compression layer to match attention scores. Our compression technique reduces the storage required up to 95% and it can be applied without a substantial degradation in ranking performance.","{'model': 'tldr@v2.0.0', 'text': 'The proposed approach, called PreTTR (Precomputing Transformer Term Representations), considerably reduces the query-time latency of deep transformer networks making these networks more practical to use in a real-time ranking scenario.'}",https://arxiv.org/pdf/2004.14255
-sarathi: efficient llm inference by piggybacking decodes with chunked prefills,Amey Agrawal,"Large Language Model (LLM) inference consists of two distinct phases - prefill phase which processes the input prompt and decode phase which generates output tokens autoregressively. While the prefill phase effectively saturates GPU compute at small batch sizes, the decode phase results in low compute utilization as it generates one token at a time per request. The varying prefill and decode times also lead to imbalance across micro-batches when using pipeline parallelism, resulting in further inefficiency due to bubbles. We present SARATHI to address these challenges. SARATHI employs chunked-prefills, which splits a prefill request into equal sized chunks, and decode-maximal batching, which constructs a batch using a single prefill chunk and populates the remaining slots with decodes. During inference, the prefill chunk saturates GPU compute, while the decode requests 'piggyback' and cost up to an order of magnitude less compared to a decode-only batch. Chunked-prefills allows constructing multiple decode-maximal batches from a single prefill request, maximizing coverage of decodes that can piggyback. Furthermore, the uniform compute design of these batches ameliorates the imbalance between micro-batches, significantly reducing pipeline bubbles. Our techniques yield significant improvements in inference performance across models and hardware. For the LLaMA-13B model on A6000 GPU, SARATHI improves decode throughput by up to 10x, and accelerates end-to-end throughput by up to 1.33x. For LLaMa-33B on A100 GPU, we achieve 1.25x higher end-to-end-throughput and up to 4.25x higher decode throughput. When used with pipeline parallelism on GPT-3, SARATHI reduces bubbles by 6.29x, resulting in an end-to-end throughput improvement of 1.91x.","{'model': 'tldr@v2.0.0', 'text': 'SARATHI employs chunked-prefills, which splits a prefill request into equal sized chunks, and decode-maximal batching, which constructs a batch using a single prefill chunk and populates the remaining slots with decodes, resulting in significant improvements in inference performance across models and hardware.'}",https://arxiv.org/pdf/2308.16369
-efficient wait-k models for simultaneous machine translation,Maha Elbayad,"Simultaneous machine translation consists in starting output generation before the entire input sequence is available. Wait-k decoders offer a simple but efficient approach for this problem. They first read k source tokens, after which they alternate between producing a target token and reading another source token. We investigate the behavior of wait-k decoding in low resource settings for spoken corpora using IWSLT datasets. We improve training of these models using unidirectional encoders, and training across multiple values of k. Experiments with Transformer and 2D-convolutional architectures show that our wait-k models generalize well across a wide range of latency levels. We also show that the 2D-convolution architecture is competitive with Transformers for simultaneous translation of spoken language.","{'model': 'tldr@v2.0.0', 'text': 'This work investigates the behavior of wait-k decoding in low resource settings for spoken corpora using IWSLT datasets, and improves training of these models using unidirectional encoders, and training across multiple values of k.'}",https://arxiv.org/pdf/2005.08595
-efficient attentions for long document summarization,L. Huang,"The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with Hepos, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GovReport, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.","{'model': 'tldr@v2.0.0', 'text': 'Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source is proposed, able to process ten times more tokens than existing models that use full attentions.'}",https://aclanthology.org/2021.naacl-main.112.pdf
-fast and memory-efficient neural code completion,Alexey Svyatkovskiy,"Code completion is one of the most widely used features of modern integrated development environments (IDEs). While deep learning has made significant progress in the statistical prediction of source code, state-of-the-art neural network models consume hundreds of megabytes of memory, bloating the development environment. We address this in two steps: first we present a modular neural framework for code completion. This allows us to explore the design space and evaluate different techniques. Second, within this framework we design a novel reranking neural completion model that combines static analysis with granular token encodings. The best neural reranking model consumes just 6 MB of RAM, — 19x less than previous models — computes a single completion in 8 ms, and achieves 90% accuracy in its top five suggestions.","{'model': 'tldr@v2.0.0', 'text': 'A modular neural framework for code completion is presented and a novel reranking neural completion model is designed that combines static analysis with granular token encodings and achieves 90% accuracy in its top five suggestions.'}",https://arxiv.org/pdf/2004.13651
-protect: efficient password-based threshold single-sign-on authentication for mobile users against perpetual leakage,Y. Zhang,"Password-based single-sign-on authentication has been widely applied in mobile environments. It enables an identity server to issue authentication tokens to mobile users holding correct passwords. With an authentication token, one can request mobile services from related service providers without multiple registrations. However, if an adversary compromises the identity server, he can retrieve users’ passwords by performing dictionary guessing attacks (DGA) and can overissue authentication tokens to break the security. In this paper, we propose a password-based threshold single-sign-on authentication scheme dubbed PROTECT that thwarts adversaries who can compromise identity server(s), where multiple identity servers are introduced to authenticate mobile users and issue authentication tokens in a threshold way. PROTECT supports key renewal that periodically updates the secret on each identity server to resist perpetual leakage of the secret. Furthermore, PROTECT is secure against off-line DGA: a credential used to authenticate a user is computed from the password and a server-side key. PROTECT is also resistant to online DGA and password testing attacks in an efficient way. We conduct a comprehensive performance evaluation of PROTECT, which demonstrates the high efficiency on the user side in terms of computation and communication and proves that it can be easily deployed on mobile devices.","{'model': 'tldr@v2.0.0', 'text': 'A password-based threshold single-sign-on authentication scheme dubbed PROTECT is proposed that thwarts adversaries who can compromise identity server(s), where multiple identity servers are introduced to authenticate mobile users and issue authentication tokens in a threshold way and proves that it can be easily deployed on mobile devices.'}",
-panda: prompt transfer meets knowledge distillation for efficient model adaptation,Qihuang Zhong,"Prompt-tuning, which freezes pretrained language models (PLMs) and only fine-tunes few parameters of additional soft prompt, shows competitive performance against full-parameter fine-tuning ( i.e., model-tuning) when the PLM has billions of parameters, but still performs poorly in the case of smaller PLMs. Hence, prompt transfer (PoT), which initializes the target prompt with the trained prompt of similar source tasks, is recently proposed to improve over prompt-tuning. However, such a vanilla PoT approach usually achieves sub-optimal performance, as (i) the PoT is sensitive to the similarity of source-target pair and (ii) directly fine-tuning the prompt initialized with source prompt on target task might lead to catastrophic forgetting of source knowledge. In response to these problems, we propose a new metric to accurately predict the prompt transferability (regarding (i)), and a novel PoT approach (namely PANDA) that leverages the knowledge distillation technique to transfer the “knowledge” from the source prompt to the target prompt in a subtle manner and alleviate the catastrophic forgetting effectively (regarding (ii)). Furthermore, to achieve adaptive prompt transfer for each source-target pair, we use our metric to control the knowledge transfer in our PANDA approach. Extensive and systematic experiments on 189 combinations of 21 source and 9 target datasets across 5 scales of PLMs demonstrate that: 1) our proposed metric works well to predict the prompt transferability ; 2) our PANDA consistently outperforms the vanilla PoT approach by 2.3% average score (up to 24.1%) among all tasks and model sizes ; 3) with our PANDA approach, prompt-tuning can achieve competitive and even better performance than model-tuning in various PLM","{'model': 'tldr@v2.0.0', 'text': 'A new metric to accurately predict the prompt transferability is proposed, and a novel PoT approach (namely PANDA) is proposed that leverages the knowledge distillation technique to transfer the “knowledge” from the source prompt to the target prompt in a subtle manner and alleviate the catastrophic forgetting effectively.'}",http://arxiv.org/pdf/2208.10160
-token turing machines,M. Ryoo,"We propose Token Turing Machines (TTM), a sequential, autoregressive Transformer model with memory for real-world sequential visual understanding. Our model is inspired by the seminal Neural Turing Machine, and has an external memory consisting of a set of tokens which summarise the previous history (i.e., frames). This memory is efficiently addressed, read and written using a Transformer as the processing unit/controller at each step. The model's memory module ensures that a new observation will only be processed with the contents of the memory (and not the entire history), meaning that it can efficiently process long sequences with a bounded computational cost at each step. We show that TTM outperforms other alternatives, such as other Transformer models designed for long sequences and recurrent neural networks, on two real-world sequential visual understanding tasks: online temporal activity detection from videos and vision-based robot action policy learning. Code is publicly available at: https://github.com/google-research/scenic/tree/main/scenic/projects/token.turing.","{'model': 'tldr@v2.0.0', 'text': 'Token Turing Machines outperforms other alternatives, such as other Transformer models designed for long sequences and recurrent neural networks, on two real-world sequential visual understanding tasks: online temporal activity detection from videos and vision-based robot action policy learning.'}",
-toward efficient language model pretraining and downstream adaptation via self-evolution: a case study on superglue,Qihuang Zhong,"This technical report briefly describes our JDExplore d-team's Vega v2 submission on the SuperGLUE leaderboard. SuperGLUE is more challenging than the widely used general language understanding evaluation (GLUE) benchmark, containing eight difficult language understanding tasks, including question answering, natural language inference, word sense disambiguation, coreference resolution, and reasoning. [Method] Instead of arbitrarily increasing the size of a pretrained language model (PLM), our aim is to 1) fully extract knowledge from the input pretraining data given a certain parameter budget, e.g., 6B, and 2) effectively transfer this knowledge to downstream tasks. To achieve goal 1), we propose self-evolution learning for PLMs to wisely predict the informative tokens that should be masked, and supervise the masked language modeling (MLM) process with rectified smooth labels. For goal 2), we leverage the prompt transfer technique to improve the low-resource tasks by transferring the knowledge from the foundation model and related downstream tasks to the target task. [Results] According to our submission record (Oct. 2022), with our optimized pretraining and fine-tuning strategies, our 6B Vega method achieved new state-of-the-art performance on 4/8 tasks, sitting atop the SuperGLUE leaderboard on Oct. 8, 2022, with an average score of 91.3.","{'model': 'tldr@v2.0.0', 'text': ""This technical report briefly describes the JDExplore d-team's Vega v2 submission on the SuperGLUE leaderboard, which achieved new state-of-the-art performance on 4/8 tasks, and the prompt transfer technique to improve the low-resource tasks by transferring the knowledge from the foundation model and related downstream tasks to the target task.""}",https://arxiv.org/pdf/2212.01853
-vision transformers are parameter-efficient audio-visual learners,Yan-Bo Lin,"Vision transformers (ViTs) have achieved impressive results on various computer vision tasks in the last several years. In this work, we study the capability of frozen ViTs, pretrained only on visual data, to generalize to audio-visual data without finetuning any of its original parameters. To do so, we propose a latent audio-visual hybrid (LAVISH) adapter that adapts pretrained ViTs to audio-visual tasks by injecting a small number of trainable parameters into every layer of a frozen ViT. To efficiently fuse visual and audio cues, our LAVISH adapter uses a small set of latent tokens, which form an attention bottleneck, thus, eliminating the quadratic cost of standard cross-attention. Compared to the existing modality-specific audio-visual methods, our approach achieves competitive or even better performance on various audio-visual tasks while using fewer tunable parameters and without relying on costly audio pretraining or external audio encoders. Our code is available at https://genjib.github.io/project_page/LAVISH/","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a latent audio-visual hybrid (LAVISH) adapter that adapts pretrained ViTs to audio- visual tasks by injecting a small number of trainable parameters into every layer of a frozen ViT, eliminating the quadratic cost of standard cross-attention.'}",https://arxiv.org/pdf/2212.07983
-sait: sparse vision transformers through adaptive token pruning,Ling Li,"While vision transformers have achieved impressive results, effectively and efficiently accelerating these models can further boost performances. In this work, we propose a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities. Thus one model offers a range of accuracy and throughput tradeoffs for different applications. Besides, we introduce adaptive token pruning to optimize the patch token sparsity based on the input image. In addition, we investigate knowledge distillation to enhance token selection capability in early transformer modules. Sparse adaptive image Transformer (SaiT) offers varying levels of model acceleration by merely changing the token sparsity on the fly. Specifically, SaiT reduces the computation complexity (FLOPs) by 39% - 43% and increases the throughput by 67% - 91% with less than 0.5% accuracy loss for various vision transformer models. Meanwhile, the same model also provides the zero accuracy drop option by skipping the sparsification step. SaiT achieves better accuracy and computation tradeoffs than state-of-the-art transformer and convolutional models.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities, and introduces adaptive token pruning to optimize the patch token sparsity based on the input image.'}",http://arxiv.org/pdf/2210.05832
-cppcd: a token-based approach to detecting potential clones,Yu-Liang Hung,"Most state-of-the-art clone detection approaches are aimed at finding clones accurately and/or efficiently. Yet, whether a code fragment is a clone often varies according to different people's perspectives and different clone detection tools. In this paper, we present CPPCD (CP-based Potential Clone Detection), a novel token-based approach to detecting potential clones. It generates CP (clone probability) values and CP distribution graphs for developers to decide if a method is a clone. We have evaluated our approach on large-scale software projects written in Java. Our experiments suggest that the majority of clones have CP values greater than or equal to 0.75 and that CPPCD is an accurate (with respect to Type-1, Type-2, and Type-3 clones), efficient, and scalable approach to detecting potential clones.","{'model': 'tldr@v2.0.0', 'text': 'CPPCD (CP-based Potential Clone Detection), a novel token-based approach to detecting potential clones, generates CP (clone probability) values and CP distribution graphs for developers to decide if a method is a clone.'}",
-mc-bert: efficient language pre-training via a meta controller,Zhenhui Xu,"Pre-trained contextual representations (e.g., BERT) have become the foundation to achieve state-of-the-art results on many NLP tasks. However, large-scale pre-training is computationally expensive. ELECTRA, an early attempt to accelerate pre-training, trains a discriminative model that predicts whether each input token was replaced by a generator. Our studies reveal that ELECTRA's success is mainly due to its reduced complexity of the pre-training task: the binary classification (replaced token detection) is more efficient to learn than the generation task (masked language modeling). However, such a simplified task is less semantically informative. To achieve better efficiency and effectiveness, we propose a novel meta-learning framework, MC-BERT. The pre-training task is a multi-choice cloze test with a reject option, where a meta controller network provides training input and candidates. Results over GLUE natural language understanding benchmark demonstrate that our proposed method is both efficient and effective: it outperforms baselines on GLUE semantic tasks given the same computational budget.","{'model': 'tldr@v2.0.0', 'text': 'Results over GLUE natural language understanding benchmark demonstrate that the proposed MC-BERT method is both efficient and effective: it outperforms baselines on GLUE semantic tasks given the same computational budget.'}",
-degree: a data-efficient generation-based event extraction model,I-Hung Hsu,"Event extraction requires high-quality expert human annotations, which are usually expensive. Therefore, learning a data-efficient event extraction model that can be trained with only a few labeled examples has become a crucial challenge. In this paper, we focus on low-resource end-to-end event extraction and propose DEGREE, a data-efficient model that formulates event extraction as a conditional generation problem. Given a passage and a manually designed prompt, DEGREE learns to summarize the events mentioned in the passage into a natural sentence that follows a predefined pattern. The final event predictions are then extracted from the generated sentence with a deterministic algorithm. DEGREE has three advantages to learn well with less training data. First, our designed prompts provide semantic guidance for DEGREE to leverage DEGREE and thus better capture the event arguments. Moreover, DEGREE is capable of using additional weakly-supervised information, such as the description of events encoded in the prompts. Finally, DEGREE learns triggers and arguments jointly in an end-to-end manner, which encourages the model to better utilize the shared knowledge and dependencies among them. Our experimental results demonstrate the strong performance of DEGREE for low-resource event extraction.","{'model': 'tldr@v2.0.0', 'text': 'DEGREE is proposed, a data-efficient model that formulates event extraction as a conditional generation problem and learns triggers and arguments jointly in an end-to-end manner, which encourages the model to better utilize the shared knowledge and dependencies among them.'}",https://aclanthology.org/2022.naacl-main.138.pdf
-efficient exact similarity searches using multiple token orderings,Jongik Kim,"Similarity searches are essential in many applications including data cleaning and near duplicate detection. Many similarity search algorithms first generate candidate records, and then identify true matches among them. A major focus of those algorithms has been on how to reduce the number of candidate records in the early stage of similarity query processing. One of the most commonly used techniques to reduce the candidate size is the prefix filtering principle, which exploits the document frequency ordering of tokens. In this paper, we propose a novel partitioning technique that considers multiple token orderings based on token co-occurrence statistics. Experimental results show that the proposed technique is effective in reducing the number of candidate records and as a result improves the performance of existing algorithms significantly.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a novel partitioning technique that considers multiple token orderings based on token co-occurrence statistics and shows that the proposed technique is effective in reducing the number of candidate records and as a result improves the performance of existing algorithms significantly.'}",
-trips: efficient vision-and-language pre-training with text-relevant image patch selection,Chaoya Jiang,"Vision Transformers (ViTs) have been widely used in large-scale Vision and Language Pre-training (VLP) models. Though previous VLP works have proved the effectiveness of ViTs, they still suffer from computational efficiency brought by the long visual sequence. To tackle this problem, in this paper, we propose an efficient vision-and-language pre-training model with Text-Relevant Image Patch Selection, namely TRIPS, which reduces the visual sequence progressively with a text-guided patch-selection layer in the visual backbone for efficient training and inference. The patch-selection layer can dynamically compute text-dependent visual attention to identify the attentive image tokens with text guidance and fuse inattentive ones in an end-to-end manner. Meanwhile, TRIPS does not introduce extra parameters to ViTs. Experimental results on a variety of popular benchmark datasets demonstrate that TRIPS gain a speedup of 40% over previous similar VLP models, yet with competitive or better downstream task performance.","{'model': 'tldr@v2.0.0', 'text': 'An efficient vision-and-language pre-training model with Text-Relevant Image Patch Selection, namely TRIPS, which reduces the visual sequence progressively with a text-guided patch-selection layer in the visual backbone for efficient training and inference.'}",https://aclanthology.org/2022.emnlp-main.273.pdf
-canine: pre-training an efficient tokenization-free encoder for language representation,J. Clark,"Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model’s ability to adapt. In this paper, we present Canine, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, Canine combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. Canine outperforms a comparable mBert model by 5.7 F1 on TyDi QA, a challenging multilingual benchmark, despite having fewer model parameters.","{'model': 'tldr@v2.0.0', 'text': 'Canine is presented, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias.'}",https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00448/1985933/tacl_a_00448.pdf
-nodepiece: compositional and parameter-efficient representations of large knowledge graphs,Mikhail Galkin,"Conventional representation learning algorithms for knowledge graphs (KG) map each entity to a unique embedding vector. Such a shallow lookup results in a linear growth of memory consumption for storing the embedding matrix and incurs high computational costs when working with real-world KGs. Drawing parallels with subword tokenization commonly used in NLP, we explore the landscape of more parameter-efficient node embedding strategies with possibly sublinear memory requirements. To this end, we propose NodePiece, an anchor-based approach to learn a fixed-size entity vocabulary. In NodePiece, a vocabulary of subword/sub-entity units is constructed from anchor nodes in a graph with known relation types. Given such a fixed-size vocabulary, it is possible to bootstrap an encoding and embedding for any entity, including those unseen during training. Experiments show that NodePiece performs competitively in node classification, link prediction, and relation prediction tasks while retaining less than 10% of explicit nodes in a graph as anchors and often having 10x fewer parameters. To this end, we show that a NodePiece-enabled model outperforms existing shallow models on a large OGB WikiKG 2 graph having 70x fewer parameters.","{'model': 'tldr@v2.0.0', 'text': 'A NodePiece-enabled model outperforms existing shallow models on a large OGB WikiKG 2 graph having 70x fewer parameters and is shown to perform competitively in node classification, link prediction, and relation prediction tasks.'}",
-llama-adapter v2: parameter-efficient visual instruction model,Peng Gao,"How to efficiently transform large language models (LLMs) into instruction followers is recently a popular research direction, while training LLM for multi-modal reasoning remains less explored. Although the recent LLaMA-Adapter demonstrates the potential to handle visual inputs with LLMs, it still cannot generalize well to open-ended visual instructions and lags behind GPT-4. In this paper, we present LLaMA-Adapter V2, a parameter-efficient visual instruction model. Specifically, we first augment LLaMA-Adapter by unlocking more learnable parameters (e.g., norm, bias and scale), which distribute the instruction-following ability across the entire LLaMA model besides adapters. Secondly, we propose an early fusion strategy to feed visual tokens only into the early LLM layers, contributing to better visual knowledge incorporation. Thirdly, a joint training paradigm of image-text pairs and instruction-following data is introduced by optimizing disjoint groups of learnable parameters. This strategy effectively alleviates the interference between the two tasks of image-text alignment and instruction following and achieves strong multi-modal reasoning with only a small-scale image-text and instruction dataset. During inference, we incorporate additional expert models (e.g. captioning/OCR systems) into LLaMA-Adapter to further enhance its image understanding capability without incurring training costs. Compared to the original LLaMA-Adapter, our LLaMA-Adapter V2 can perform open-ended multi-modal instructions by merely introducing 14M parameters over LLaMA. The newly designed framework also exhibits stronger language-only instruction-following capabilities and even excels in chat interactions. Our code and models are available at https://github.com/ZrrSkywalker/LLaMA-Adapter.","{'model': 'tldr@v2.0.0', 'text': 'This work augments LLaMA-Adapter by unlocking more learnable parameters and proposes an early fusion strategy to feed visual tokens only into the early LLM layers, contributing to better visual knowledge incorporation and achieves strong multi-modal reasoning with only a small-scale image-text and instruction dataset.'}",http://arxiv.org/pdf/2304.15010
-long range arena: a benchmark for efficient transformers,Yi Tay,"Transformers do not scale very well to long sequence lengths largely because of quadratic self-attention complexity. In the recent months, a wide spectrum of efficient, fast Transformers have been proposed to tackle this problem, more often than not claiming superior or comparable model quality to vanilla Transformer models. To this date, there is no well-established consensus on how to evaluate this class of models. Moreover, inconsistent benchmarking on a wide spectrum of tasks and datasets makes it difficult to assess relative model quality amongst many models. This paper proposes a systematic and unified benchmark, LRA, specifically focused on evaluating model quality under long-context scenarios. Our benchmark is a suite of tasks consisting of sequences ranging from $1K$ to $16K$ tokens, encompassing a wide range of data types and modalities such as text, natural, synthetic images, and mathematical expressions requiring similarity, structural, and visual-spatial reasoning. We systematically evaluate ten well-established long-range Transformer models (Reformers, Linformers, Linear Transformers, Sinkhorn Transformers, Performers, Synthesizers, Sparse Transformers, and Longformers) on our newly proposed benchmark suite. LRA paves the way towards better understanding this class of efficient Transformer models, facilitates more research in this direction, and presents new challenging tasks to tackle. Our benchmark code will be released at this https URL.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and unified benchmark, LRA, specifically focused on evaluating model quality under long-context scenarios is proposed, paving the way towards better understanding this class of efficient Transformer models.'}",
-visual query tuning: towards effective usage of intermediate representations for parameter and memory efficient transfer learning,Cheng-Hao Tu,"Intermediate features of a pretrained model have been shown informative for making accurate predictions on downstream tasks, even if the model backbone is kept frozen. The key challenge is how to utilize these intermediate features given their gigantic amount. We propose visual query tuning (VQT), a simple yet effective approach to aggregate intermediate features of Vision Transformers. Through introducing a handful of learnable “query” tokens to each layer, VQT leverages the inner workings of Transformers to “summarize” rich intermediate features of each layer, which can then be used to train the prediction heads of downstream tasks. As VQT keeps the intermediate features intact and only learns to combine them, it enjoys memory efficiency in training, compared to many other parameter-efficient fine-tuning approaches that learn to adapt features and need back-propagation through the entire backbone. This also suggests the complementary role between VQT and those approaches in transfer learning. Empirically, VQT consistently surpasses the state-of-the-art approach that utilizes intermediate features for transfer learning and outperforms full fine-tuning in many cases. Compared to parameter-efficient approaches that adapt features, VQT achieves much higher accuracy under memory constraints. Most importantly, VQT is compatible with these approaches to attain even higher accuracy, making it a simple add-on to further boost transfer learning. Code is available at https://github.com/andytu28/VQT.","{'model': 'tldr@v2.0.0', 'text': 'Visual query tuning (VQT), a simple yet effective approach to aggregate intermediate features of Vision Transformers, which consistently surpasses the state-of-the-art approach that utilizes intermediate features for transfer learning and outperforms full fine-tuning in many cases.'}",https://arxiv.org/pdf/2212.03220
-fast-join: an efficient method for fuzzy token matching based string similarity join,Jiannan Wang,"String similarity join that finds similar string pairs between two string sets is an essential operation in many applications, and has attracted significant attention recently in the database community. A significant challenge in similarity join is to implement an effective fuzzy match operation to find all similar string pairs which may not match exactly. In this paper, we propose a new similarity metrics, called “fuzzy token matching based similarity”, which extends token-based similarity functions (e.g., Jaccard similarity and Cosine similarity) by allowing fuzzy match between two tokens. We study the problem of similarity join using this new similarity metrics and present a signature-based method to address this problem. We propose new signature schemes and develop effective pruning techniques to improve the performance. Experimental results show that our approach achieves high efficiency and result quality, and significantly outperforms state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new similarity metrics, called “fuzzy token matching based similarity”, which extends token-based similarity functions by allowing fuzzy match between two tokens, and achieves high efficiency and result quality, and significantly outperforms state-of-the-art methods.'}",http://dbgroup.cs.tsinghua.edu.cn/ligl/papers/icde2011-fastjoin.pdf
-fedprompt: communication-efficient and privacy-preserving prompt tuning in federated learning,Haodong Zhao,"Federated learning (FL) has enabled global model training on decentralized data in a privacy-preserving way. However, for tasks that utilize pre-trained language models (PLMs) with massive parameters, there are considerable communication costs. Prompt tuning, which tunes soft prompts without modifying PLMs, has achieved excellent performance as a new learning paradigm. In this paper, we want to combine these methods and explore the effect of prompt tuning under FL. We propose ""FedPrompt"" studying prompt tuning in a model split aggregation way using FL, and prove that split aggregation greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution. We further conduct backdoor attacks by data poisoning on FedPrompt. Experiments show that attack achieve a quite low attack success rate and can not inject backdoor effectively, proving the robustness of FedPrompt.","{'model': 'tldr@v2.0.0', 'text': 'It is proved that split aggregation greatly reduces the communication cost, only 0.01% of the PLMs’ parameters, with little decrease on accuracy both on IID and Non-IID data distribution, proving the robustness of FedPrompt.'}",https://arxiv.org/pdf/2208.12268
-token allocation for course bidding with machine learning method,Chonbadee Juthamanee,"Most educational institutions have a problem of allocation course seats to students because demand exceeds supply for many courses. A course bidding system is a tool for improving the registration system based on auction theory to provide course seat allocation equitably and efficiently. This paper proposes a method for token price prediction for allocate to courses through course bidding system. Three methods are compared: Decision Tree, Random Forest and Artificial Neuron Network. The result of the experiment shows that ANN is the best method with lowest RSME 3.98%. Furthermore, it provides an important information to users to estimate the risks on their course bidding strategy.","{'model': 'tldr@v2.0.0', 'text': 'The result of the experiment shows that ANN is the best method with lowest RSME 3.98% and provides an important information to users to estimate the risks on their course bidding strategy.'}",
-token meets wallet: formalizing privacy and revocation for fido2,L. Hanzlik,"The FIDO2 standard is a widely-used class of challenge-response type protocols that allows to authenticate to an online service using a hardware token. Barbosa et al. (CRYPTO ‘21) provided the first formal security model and analysis for the FIDO2 standard. However, their model has two shortcomings: (1) It does not include privacy, one of the key features claimed by FIDO2. (2) It only covers tokens that store all secret keys locally. In contrast, due to limited memory, most existing FIDO2 tokens either derive all secret keys from a common seed or store keys on the server (the latter approach is also known as key wrapping).In this paper, we revisit the security of the WebAuthn component of FIDO2 as implemented in practice. Our contributions are as follows. (1) We adapt the model of Barbosa et al. so as to capture authentication tokens using key derivation or key wrapping. (2) We provide the first formal definition of privacy for the WebAuthn component of FIDO2. We then prove the privacy of this component in common FIDO2 token implementations if the underlying building blocks are chosen appropriately. (3) We address the unsolved problem of global key revocation in FIDO2. To this end, we introduce and analyze a simple revocation procedure that builds on the popular BIP32 standard used in cryptocurrency wallets and can efficiently be implemented with existing FIDO2 servers.","{'model': 'tldr@v2.0.0', 'text': 'A simple revocation procedure is introduced and analyzed that builds on the popular BIP32 standard used in cryptocurrency wallets and can efficiently be implemented with existing FIDO2 servers and address the unsolved problem of global key revocation in F IDO2.'}",
-to repeat or not to repeat: insights from scaling llm under token-crisis,Fuzhao Xue,"Recent research has highlighted the importance of dataset size in scaling language models. However, large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. To further enhance LLMs, a straightforward approach is to repeat the pre-training data for additional epochs. In this study, we empirically investigate three key aspects under this approach. First, we explore the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting, leading to multi-epoch degradation. Second, we examine the key factors contributing to multi-epoch degradation, finding that significant factors include dataset size, model parameters, and training objectives, while less influential factors consist of dataset quality and model FLOPs. Finally, we explore whether widely used regularization can alleviate multi-epoch degradation. Most regularization techniques do not yield significant improvements, except for dropout, which demonstrates remarkable effectiveness but requires careful tuning when scaling up the model size. Additionally, we discover that leveraging mixture-of-experts (MoE) enables cost-effective and efficient hyper-parameter tuning for computationally intensive dense LLMs with comparable trainable parameters, potentially impacting efficient LLM development on a broader scale.","{'model': 'tldr@v2.0.0', 'text': 'This study empirically investigates the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting, leading to multi-epoch degradation, and discovers that leveraging mixture-of-experts (MoE) enables cost-effective and efficient hyper-parameter tuning for computationally intensive dense LLMs with comparable trainable parameters.'}",https://arxiv.org/pdf/2305.13230
-universal few-shot learning of dense prediction tasks with visual token matching,Donggyun Kim,"Dense prediction tasks are a fundamental class of problems in computer vision. As supervised methods suffer from high pixel-wise labeling cost, a few-shot learning solution that can learn any dense task from a few labeled images is desired. Yet, current few-shot learning methods target a restricted set of tasks such as semantic segmentation, presumably due to challenges in designing a general and unified model that is able to flexibly and efficiently adapt to arbitrary tasks of unseen semantics. We propose Visual Token Matching (VTM), a universal few-shot learner for arbitrary dense prediction tasks. It employs non-parametric matching on patch-level embedded tokens of images and labels that encapsulates all tasks. Also, VTM flexibly adapts to any task with a tiny amount of task-specific parameters that modulate the matching algorithm. We implement VTM as a powerful hierarchical encoder-decoder architecture involving ViT backbones where token matching is performed at multiple feature hierarchies. We experiment VTM on a challenging variant of Taskonomy dataset and observe that it robustly few-shot learns various unseen dense prediction tasks. Surprisingly, it is competitive with fully supervised baselines using only 10 labeled examples of novel tasks (0.004% of full supervision) and sometimes outperforms using 0.1% of full supervision. Codes are available at https://github.com/GitGyun/visual_token_matching.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Visual Token Matching (VTM), a universal few-shot learner for arbitrary dense prediction tasks that employs non-parametric matching on patch-level embedded tokens of images and labels that encapsulates all tasks and flexibly adapts to any task with a tiny amount of task-specific parameters that modulate the matching algorithm.'}",http://arxiv.org/pdf/2303.14969
-efficient training of bert by progressively stacking,Linyuan Gong,"Unsupervised pre-training is commonly used in natural language processing: a deep neural network trained with proper unsupervised prediction tasks are shown to be effective in many down-stream tasks. Because it is easy to create a large monolingual dataset by collecting data from the Web, we can train high-capacity models. Therefore, training efficiency becomes a critical issue even when using high-performance hardware. In this paper, we explore an efficient training method for the state-of-the-art bidirectional Transformer (BERT) model. By visualizing the self-attention distributions of different layers at different positions in a well-trained BERT model, we find that in most layers, the self-attention distribution will concentrate locally around its position and the start-of-sentence token. Motivated by this, we pro-pose the stacking algorithm to transfer knowledge from a shallow model to a deep model; then we apply stacking progressively to accelerate BERT training. Experiments showed that the models trained by our training strategy achieve similar performance to models trained from scratch, but our algorithm is much faster.","{'model': 'tldr@v2.0.0', 'text': 'This paper explores an efficient training method for the state-of-the-art bidirectional Transformer (BERT) model and pro-poses the stacking algorithm to transfer knowledge from a shallow model to a deep model; then the algorithm is applied progressively to accelerate BERT training.'}",
-energy efficient cross-layer design protocol by using token passing mechanism for wsn,Sandhyasree Thaskani,"Recent advancements in processor, memory and radio technologies have enabled cheap and small sensors which are capable of sensing, communication and computation. Most of the Wireless Sensor Networks have limitation of battery power and in most of the applications it is not possible to replace the battery of a node. Considering this scarce energy and processing resources of WSN, we try to establish a joint optimization and design of networking layers i.e., cross layer design which could be a promising alternative to the inefficient traditional layered protocol design. In this paper we propose energy efficient cross layer design of the MAC and Routing protocol namely Energy Efficient cross-layer design protocol by using token passing mechanism for WSN. This proposed protocol proves to be better than some of the existing protocols and it is shown with the help of simulations.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes energy efficient cross layer design of the MAC and Routing protocol namely Energy Efficient cross-layer design protocol by using token passing mechanism for WSN and it proves to be better than some of the existing protocols.'}",
-prompt tuning for parameter-efficient medical image segmentation,Marc Fischer,"Neural networks pre-trained on a self-supervision scheme have become the standard when operating in data rich environments with scarce annotations. As such, fine-tuning a model to a downstream task in a parameter-efficient but effective way, e.g. for a new set of classes in the case of semantic segmentation, is of increasing importance. In this work, we propose and investigate several contributions to achieve a parameter-efficient but effective adaptation for semantic segmentation on two medical imaging datasets. Relying on the recently popularized prompt tuning approach, we provide a prompt-able UNet (PUNet) architecture, that is frozen after pre-training, but adaptable throughout the network by class-dependent learnable prompt tokens. We pre-train this architecture with a dedicated dense self-supervision scheme based on assignments to online generated prototypes (contrastive prototype assignment, CPA) of a student teacher combination alongside a concurrent segmentation loss on a subset of classes. We demonstrate that the resulting neural network model is able to attenuate the gap between fully fine-tuned and parameter-efficiently adapted models on CT imaging datasets. As such, the difference between fully fine-tuned and prompt-tuned variants amounts to only 3.83 pp for the TCIA/BTCV dataset and 2.67 pp for the CT-ORG dataset in the mean Dice Similarity Coefficient (DSC, in %) while only prompt tokens, corresponding to 0.85% of the pre-trained backbone model with 6.8M frozen parameters, are adjusted. The code for this work is available on https://github.com/marcdcfischer/PUNet .","{'model': 'tldr@v2.0.0', 'text': 'This work provides a prompt-able UNet (PUNet) architecture, that is frozen after pre-training, but adaptable throughout the network by class-dependent learnable prompt tokens, and demonstrates that the resulting neural network model is able to attenuate the gap between fully fine-tuned and parameter-efficiently adapted models on CT imaging datasets.'}",https://arxiv.org/pdf/2211.09233
-rethinking visual prompt learning as masked visual token modeling,Ning Liao,"Prompt learning has achieved great success in efficiently exploiting large-scale pre-trained models in natural language processing (NLP). It reformulates the downstream tasks as the generative pre-training ones, thus narrowing down the gap between them and improving the performance stably. However, when transferring it to the vision area, current visual prompt learning methods are all designed on discriminative pre-trained models, and there is also a lack of careful design to unify the forms of pre-training and downstream tasks. To explore prompt learning on the generative pre-trained visual model as well as keeping the task consistency, we propose Visual Prompt learning as masked visual Token Modeling (VPTM) to transform the downstream visual classification into the pre-trained masked visual token prediction. In addition, we develop the prototypical verbalizer for mapping the predicted visual token with implicit semantics to explicit downstream labels. To our best knowledge, VPTM is the first visual prompt method on the generative pre-trained visual model, and the first to achieve consistency between pre-training and downstream visual classification by task reformulation. Experiments show that VPTM outperforms other visual prompt methods and achieves excellent efficiency. Moreover, the task consistency of VPTM contributes to the robustness against prompt location, prompt length and prototype dimension, and could be deployed uniformly.","{'model': 'tldr@v2.0.0', 'text': 'Visual Prompt learning as masked visual Token Modeling (VPTM) is proposed to transform the downstream visual classification into the pre-trained masked visual token prediction, and the task consistency of VPTM contributes to the robustness against prompt location, prompt length and prototype dimension, and could be deployed uniformly.'}",http://arxiv.org/pdf/2303.04998
-energy efficient token based mac protocol for wireless sensor networks,Soumya Ray,"In this paper, we propose on energy efficient token based MAC protocol for WSNs, in order to reduce energy consumption of each sensor node which is one of the important issues to prolong the network lifetime. WSN consists of a large number of small and low-cost sensor nodes powered by small batteries and equipped with various sensing devices. Due to limited energy in WSN energy efficiency is an important factor in designing a MAC protocol. To derogate energy consumption most of the MAC protocols in WSN exploits low duty-cycle; among those RMAC, HEMAC allows a node to transmit data packet for multi-hop in a single duty-cycle. To reduce energy consumption on prolonged network life time sensor networks are usually duty cycled; each node remains in low power sleep mode most of the time and wakes up periodically to sense for channel activities. In all these above MAC protocols, due to the synchronized scheduling, transmission collisions, flooding will increase resulting in energy waste and low throughput. By allowing nodes to operate with a new token-based approach, we intend to produce energy efficiency in an event based approach by reducing flooding, collision and traffic congestion. The work in this paper draws a bed on implementing a token based distributed approach to save energy at in order to mitigate flooding. Simulation studies of the proposed MAC protocol have been carried out using Castalia simulator.","{'model': 'tldr@v2.0.0', 'text': 'The work in this paper draws a bed on implementing a token based distributed approach to save energy at in order to mitigate flooding and produce energy efficiency in an event based approach by reducing flooding, collision and traffic congestion.'}",
-fast passage re-ranking with contextualized exact term matching and efficient passage expansion,Shengyao Zhuang,"BERT-based information retrieval models are expensive, in both time (query latency) and computational resources (energy, hardware cost), making many of these models impractical especially under resource constraints. The reliance on a query encoder that only performs tokenization and on the pre-processing of passage representations at indexing, has allowed the recently proposed TILDE method to overcome the high query latency issue typical of BERT-based models. This however is at the expense of a lower effectiveness compared to other BERT-based re-rankers and dense retrievers. In addition, the original TILDE method is characterised by indexes with a very high memory footprint, as it expands each passage into the size of the BERT vocabulary. In this paper, we propose TILDEv2, a new model that stems from the original TILDE but that addresses its limitations. TILDEv2 relies on contextualized exact term matching with expanded passages. This requires to only store in the index the score of tokens that appear in the expanded passages (rather than all the vocabulary), thus producing indexes that are 99% smaller than those of TILDE. This matching mechanism also improves ranking effectiveness by 24%, without adding to the query latency. This makes TILDEv2 the state-of-the-art passage re-ranking method for CPU-only environments, capable of maintaining query latency below 100ms on commodity hardware.","{'model': 'tldr@v2.0.0', 'text': 'TILDEv2 is proposed, a new model that stems from the original TILDE but that addresses its limitations, and relies on contextualized exact term matching with expanded passages to be the state-of-the-art passage re-ranking method for CPU-only environments, capable of maintaining query latency below 100ms on commodity hardware.'}",
-token-weighted crowdsourcing,Gerry Tsoukalas,"Blockchain-based platforms often rely on token-weighted voting (“τ-weighting”) to efficiently crowdsource information from their users for a wide range of applications, including content curation and on-chain governance. We examine the effectiveness of such decentralized platforms for harnessing the wisdom and effort of the crowd. We find that τ-weighting generally discourages truthful voting and erodes the platform’s predictive power unless users are “strategic enough” to unravel the underlying aggregation mechanism. Platform accuracy decreases with the number of truthful users and the dispersion in their token holdings, and in many cases, platforms would be better off with a “flat” 1/n mechanism. When, prior to voting, strategic users can exert effort to endogenously improve their signals, users with more tokens generally exert more effort—a feature often touted in marketing materials as a core advantage of τ-weighting—however, this feature is not attributable to the mechanism itself, and more importantly, the ensuing equilibrium fails to achieve the first-best accuracy of a centralized platform. The optimality gap decreases as the distribution of tokens across users approaches a theoretical optimum, which we derive, but tends to increase with the dispersion in users’ token holdings. This paper was accepted by Gabriel Weintraub, revenue management and market analytics.","{'model': 'tldr@v2.0.0', 'text': 'It is found that τ-weighting generally discourages truthful voting and erodes the platform’s predictive power unless users are “strategic enough” to unravel the underlying aggregation mechanism.'}",https://pubsonline.informs.org/doi/pdf/10.1287/mnsc.2019.3515
-funding new ventures with digital tokens: due diligence and token tradability,Yannis Bakos,"We study funding mechanisms for new ventures based on cryptographic tokens enabled by blockchain technology, such as initial coin offerings (ICOs). The infrastructure built around blockchain allows for efficient trading of these tokens. Thus the due diligence process can be crowdsourced beyond the potential early adopters, as favorable assessments of the venture can be leveraged with speculative token purchases exceeding the buyer's individual demand. We develop a model for an entrepreneur considering digital tokens to finance a new venture, focusing on token tradability and broader crowdsourcing of due diligence. We then compare funding via digital tokens with funding from traditional financing sources like venture capital as well as with pre-sale crowdfunding with non-tradable rewards. We find that tradable digital tokens are more attractive when there is higher uncertainty about market demand, which is frequently the case when developing applications of new technologies, such as blockchain-based platforms. In such cases crowdsourcing due diligence benefits from the information contained in the private valuations of the early potential adopters. Token tradability leverages that information and increases the amount that can be financed, thus enabling new ventures with higher development costs. The increased funding comes at the cost of a lower digital token price and lower total profit for the entrepreneur, but may still be preferable to the alternatives, if such alternatives are available at all. This work makes a contribution to the emerging literatures on digital cryptographic tokens and on crowdfunding new ventures.","{'model': 'tldr@v2.0.0', 'text': 'It is found that tradable digital tokens are more attractive when there is higher uncertainty about market demand, which is frequently the case when developing applications of new technologies, such as blockchain-based platforms.'}",
-sustainable growth and token economy design: the case of steemit,M. Kim,"Cryptocurrency blockchain technology is attracting worldwide attention, and the number of initial coin offerings (ICOs) is increasing rapidly. This new economic trend, called cryptoeconomics, can program human behavior through incentive design. A cryptocurrency-based incentive system is not only transparent, but also allows businesses to substitute initial investment costs with cryptocurrency tokens until they are on a sustainable growth trajectory in terms of network effects. This study aims to propose a process for building a desirable model of a token economy, based on the case of Steemit—a blogging and social networking website that is creating high values due to its efficient token economy model. We suggest the following design process of a token economy model: (1) Determine token-business fit, (2) determine the chance of success, (3) determine the properties of token, (4) give tokens intrinsic value, (5) establish strategies to raise token value, (6) establish operational strategies of token economy system, (7) establish strategies for token liquidation, and (8) continue modifying the operational base. Considering cryptoeconomics is still at an early stage, it is expected that the guidelines on the token economy model suggested in this paper will lay a significant foundation for the development of cryptoeconomics research.","{'model': 'tldr@v2.0.0', 'text': 'This study aims to propose a process for building a desirable model of a token economy, based on the case of Steemit—a blogging and social networking website that is creating high values due to its efficient token economy model.'}",https://www.mdpi.com/2071-1050/11/1/167/pdf?version=1546164358
-end-to-end contextual speech recognition using class language models and a token passing decoder,Zhehuai Chen,"End-to-end modeling (E2E) of automatic speech recognition (ASR) blends all the components of a traditional speech recognition system into a single, unified model. Although it simplifies the ASR systems, the unified model is hard to adapt when training and testing data mismatches. In this work, we focus on contextual speech recognition, which is particularly challenging for E2E models because contextual information is only available in inference time. To improve the performance in the presence of contextual information during training, we propose to use class-based language models (CLM) that can populate context-dependent information during inference. To enable this approach to scale to a large number of class members and minimize search errors, we propose a token passing algorithm with an efficient token recombination for E2E systems. We evaluate the proposed system on general and contextual ASR tasks, and achieve relative 62% Word Error Rate (WER) reduction for the contextual ASR task without hurting recognition performance for the general ASR task. We also show that the proposed method performs well without modification of the decoding hyper-parameters across tasks, making it a desirable solution for E2E ASR.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes to use class-based language models (CLM) that can populate context-dependent information during inference for contextual speech recognition, and proposes a token passing algorithm with an efficient token recombination for E2E ASR.'}",https://arxiv.org/pdf/1812.02142
-yarn: efficient context window extension of large language models,Bowen Peng,"Rotary Position Embeddings (RoPE) have been shown to effectively encode positional information in transformer-based language models. However, these models fail to generalize past the sequence length they were trained on. We present YaRN (Yet another RoPE extensioN method), a compute-efficient method to extend the context window of such models, requiring 10x less tokens and 2.5x less training steps than previous methods. Using YaRN, we show that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow, while also surpassing previous the state-of-the-art at context window extension. In addition, we demonstrate that YaRN exhibits the capability to extrapolate beyond the limited context of a fine-tuning dataset. The models fine-tuned using YaRN has been made available and reproduced online up to 128k context length at https://github.com/jquesnelle/yarn","{'model': 'tldr@v2.0.0', 'text': 'YaRN is presented, a compute-efficient method to extend the context window of RoPE extensioN models, requiring 10x less tokens and 2.5x less training steps than previous methods, and it is shown that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow.'}",https://arxiv.org/pdf/2309.00071
-soundstorm: efficient parallel audio generation,Zalán Borsos,"We present SoundStorm, a model for efficient, non-autoregressive audio generation. SoundStorm receives as input the semantic tokens of AudioLM, and relies on bidirectional attention and confidence-based parallel decoding to generate the tokens of a neural audio codec. Compared to the autoregressive generation approach of AudioLM, our model produces audio of the same quality and with higher consistency in voice and acoustic conditions, while being two orders of magnitude faster. SoundStorm generates 30 seconds of audio in 0.5 seconds on a TPU-v4. We demonstrate the ability of our model to scale audio generation to longer sequences by synthesizing high-quality, natural dialogue segments, given a transcript annotated with speaker turns and a short prompt with the speakers' voices.","{'model': 'tldr@v2.0.0', 'text': ""The ability of the model to scale audio generation to longer sequences is demonstrated by synthesizing high-quality, natural dialogue segments, given a transcript annotated with speaker turns and a short prompt with the speakers' voices.""}",http://arxiv.org/pdf/2305.09636
-socially shared regulation of learning in cscl: understanding and prompting individual- and group-level shared regulatory activities,Sanna Järvelä,,"{'model': 'tldr@v2.0.0', 'text': 'How conceptual work in socially shared regulation of learning (SSRL) contributes to effective and efficient CSCL, what tools are presently available, and what the implications of research on these tools are for future tool development are discussed.'}",
-stem: secure token exchange mechanisms,Maneesh Darisi,,"{'model': 'tldr@v2.0.0', 'text': 'This paper presents a solution which is token standard agnostic and provides effective intrinsic smart contracts facilitating token exchange and thus reducing the counterparty risk.'}",
-walk proximal gradient: an energy-efficient algorithm for consensus optimization,Xianghui Mao,"Decentralized computing is widely used for multiagent systems since it works without a central computing node. In this paper, we develop a first-order algorithm for decentralized consensus optimization that is more energy efficient than the current state-of-the-art. Our algorithm is suitable for application scenarios such as networks of wireless sensors and Internet of Things, where some agents have limited (battery) energy. We call our algorithm walk proximal gradient (WPG), which passes a token through a walk (a succession of nodes) in the graph. The agents that are visited during the walk compute the gradients of their private functions and update the token. We analyze WPG where the walk is the repetition of a Hamiltonian cycle and show that the token converges to the consensual solution faster (in terms of energy consumption) than existing gradient-based decentralized methods. We also generalize the analysis to the non-Hamiltonian graphs. Numerical experiments are presented to validate the energy efficiency of our algorithm.","{'model': 'tldr@v2.0.0', 'text': 'A first-order algorithm for decentralized consensus optimization that is more energy efficient than the current state-of-the-art and suitable for application scenarios such as networks of wireless sensors and Internet of Things, where some agents have limited energy.'}",
-point-peft: parameter-efficient fine-tuning for 3d pre-trained models,Ivan Tang,"The popularity of pre-trained large models has revolutionized downstream tasks across diverse fields, such as language, vision, and multi-modality. To minimize the adaption cost for downstream tasks, many Parameter-Efficient Fine-Tuning (PEFT) techniques are proposed for language and 2D image pre-trained models. However, the specialized PEFT method for 3D pre-trained models is still under-explored. To this end, we introduce Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters. Specifically, for a pre-trained 3D model, we freeze most of its parameters, and only tune the newly added PEFT modules on downstream tasks, which consist of a Point-prior Prompt and a Geometry-aware Adapter. The Point-prior Prompt adopts a set of learnable prompt tokens, for which we propose to construct a memory bank with domain-specific knowledge, and utilize a parameter-free attention to enhance the prompt tokens. The Geometry-aware Adapter aims to aggregate point cloud features within spatial neighborhoods to capture fine-grained geometric information through local interactions. Extensive experiments indicate that our Point-PEFT can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters, demonstrating the efficiency and effectiveness of our approach. Code will be released at https://github.com/EvenJoker/Point-PEFT.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters that can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters.'}",https://arxiv.org/pdf/2310.03059
-supporting beacon and event-driven messages in vehicular platoons through token-based strategies,A. Balador,"Timely and reliable inter-vehicle communications is a critical requirement to support traffic safety applications, such as vehicle platooning. Furthermore, low-delay communications allow the platoon to react quickly to unexpected events. In this scope, having a predictable and highly effective medium access control (MAC) method is of utmost importance. However, the currently available IEEE 802.11p technology is unable to adequately address these challenges. In this paper, we propose a MAC method especially adapted to platoons, able to transmit beacons within the required time constraints, but with a higher reliability level than IEEE 802.11p, while concurrently enabling efficient dissemination of event-driven messages. The protocol circulates the token within the platoon not in a round-robin fashion, but based on beacon data age, i.e., the time that has passed since the previous collection of status information, thereby automatically offering repeated beacon transmission opportunities for increased reliability. In addition, we propose three different methods for supporting event-driven messages co-existing with beacons. Analysis and simulation results in single and multi-hop scenarios showed that, by providing non-competitive channel access and frequent retransmission opportunities, our protocol can offer beacon delivery within one beacon generation interval while fulfilling the requirements on low-delay dissemination of event-driven messages for traffic safety applications.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a MAC method especially adapted to platoons, able to transmit beacons within the required time constraints, but with a higher reliability level than IEEE 802.11p, while concurrently enabling efficient dissemination of event-driven messages.'}",https://www.mdpi.com/1424-8220/18/4/955/pdf?version=1521820815
-gtfuzz: guard token directed grey-box fuzzing,Rundong Li,"Directed grey-box fuzzing is an effective technique to find bugs in programs with the guidance of user-specified target locations. However, it can hardly reach a target location guarded by certain syntax tokens (Guard Tokens for short), which is often seen in programs with string operations or grammar/lexical parsing. Only the test inputs containing Guard Tokens are likely to reach the target locations, which challenges the effectiveness of mutation-based fuzzers. In this paper, a Guard Token directed grey-box fuzzer called GTFuzz is presented, which extracts Guard Tokens according to the target locations first and then exploits them to direct the fuzzing. Specifically, to ensure the new test cases generated from mutations contain Guard Tokens, new strategies of seed prioritization, dictionary generation, and seed mutation are also proposed, so as to make them likely to reach the target locations. Experiments on real-world software show that GTFuzz can reach the target locations, reproduce crashes, and expose bugs more efficiently than the state-of-the-art grey-box fuzzers (i.e., AFL, AFLGO and FairFuzz). Moreover, GTFuzz identified 23 previously undiscovered bugs in LibXML2 and MJS.","{'model': 'tldr@v2.0.0', 'text': 'A Guard Token directed grey-box fuzzer called GTFuzz is presented, which extracts Guard Tokens according to the target locations first and then exploits them to direct the fuzzing, and identified 23 previously undiscovered bugs in LibXML2 and MJS.'}",
-hash embeddings for efficient word representations,Dan Svenstrup,"We present hash embeddings, an efficient method for representing words in a continuous vector form. A hash embedding may be seen as an interpolation between a standard word embedding and a word embedding created using a random hash function (the hashing trick). In hash embeddings each token is represented by $k$ $d$-dimensional embeddings vectors and one $k$ dimensional weight vector. The final $d$ dimensional representation of the token is the product of the two. Rather than fitting the embedding vectors for each token these are selected by the hashing trick from a shared pool of $B$ embedding vectors. Our experiments show that hash embeddings can easily deal with huge vocabularies consisting of millions of tokens. When using a hash embedding there is no need to create a dictionary before training nor to perform any kind of vocabulary pruning after training. We show that models trained using hash embeddings exhibit at least the same level of performance as models trained using regular embeddings across a wide range of tasks. Furthermore, the number of parameters needed by such an embedding is only a fraction of what is required by a regular embedding. Since standard embeddings and embeddings constructed using the hashing trick are actually just special cases of a hash embedding, hash embeddings can be considered an extension and improvement over the existing regular embedding types.","{'model': 'tldr@v2.0.0', 'text': 'The experiments show that hash embeddings can easily deal with huge vocabularies consisting of millions of tokens and that models trained using hash embedDings exhibit at least the same level of performance as models training using regular embeddins across a wide range of tasks.'}",
-efficient automatic punctuation restoration using bidirectional transformers with robust inference,Maury Courtland,"Though people rarely speak in complete sentences, punctuation confers many benefits to the readers of transcribed speech. Unfortunately, most ASR systems do not produce punctuated output. To address this, we propose a solution for automatic punctuation that is both cost efficient and easy to train. Our solution benefits from the recent trend in fine-tuning transformer-based language models. We also modify the typical framing of this task by predicting punctuation for sequences rather than individual tokens, which makes for more efficient training and inference. Finally, we find that aggregating predictions across multiple context windows improves accuracy even further. Our best model achieves a new state of the art on benchmark data (TED Talks) with a combined F1 of 83.9, representing a 48.7% relative improvement (15.3 absolute) over the previous state of the art.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a solution for automatic punctuation that is both cost efficient and easy to train, and modify the typical framing of this task by predicting punctuation for sequences rather than individual tokens, which makes for more efficient training and inference.'}",https://www.aclweb.org/anthology/2020.iwslt-1.33.pdf
-srdma - efficient nic-based authentication and encryption for remote direct memory access,Konstantin Taranov,"State-of-the-art remote direct memory access (RDMA) technologies have shown to be vulnerable against attacks by innetwork adversaries, as they provide only a weak form of protection by including access tokens in each message. A network eavesdropper can easily obtain sensitive information and modify bypassing packets, affecting not only secrecy but also integrity. Tampering with packets can have drastic consequences. For example, when memory pages with code are changed remotely, altering packet contents enables remote code injection. We propose sRDMA, a protocol that provides efficient authentication and encryption for RDMA to prevent information leakage and message tampering. sRDMA uses symmetric cryptography and employs network interface cards to perform cryptographic operations. Additionally, we provide an implementation for sRDMA using programmable network adapters.","{'model': 'tldr@v2.0.0', 'text': 'SRDMA is proposed, a protocol that provides efficient authentication and encryption for RDMA to prevent information leakage and message tampering and an implementation for sRDMA using programmable network adapters.'}",
-2ed: an efficient entity extraction algorithm using two-level edit-distance,Zeyi Wen,"Entity extraction is fundamental to many text mining tasks such as organisation name recognition. A popular approach to entity extraction is based on string matching against a dictionary of known entities. For approximate entity extraction from free text, considering solely character-based or solely token-based similarity cannot simultaneously deal with minor name variations at token-level and typos at character-level. Moreover, the tolerance of mismatch in character-level may be different from that in token-level, and the tolerance thresholds of the two levels should be able to be customised individually. In this paper, we propose an efficient character-level and token-level edit-distance based algorithm called FuzzyED. To improve the efficiency of FuzzyED, we develop various novel techniques including (i) a spanning-based candidate sub-string producing technique, (ii) a lower bound dissimilarity to determine the boundaries of candidate sub-strings, (iii) a core token based technique that makes use of the importance of tokens to reduce the number of unpromising candidate sub-strings, and (iv) a shrinking technique to reuse computation. Empirical results on real world datasets show that FuzzyED can efficiently extract entities and produce a high F1 score in the range of [0.91, 0.97].","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes an efficient character-level and token-level edit-distance based algorithm called FuzzyED, and develops various novel techniques including a spanning-based candidate sub-string producing technique, a lower bound dissimilarity to determine the boundaries of Candidate sub-strings, and a shrinking technique to reuse computation.'}",
-parameter-efficient tuning helps language model alignment,Tianci Xue,"Aligning large language models (LLMs) with human preferences is essential for safe and useful LLMs. Previous works mainly adopt reinforcement learning (RLHF) and direct preference optimization (DPO) with human feedback for alignment. Nevertheless, they have certain drawbacks. One such limitation is that they can only align models with one preference at the training time (e.g., they cannot learn to generate concise responses when the preference data prefers detailed responses), or have certain constraints for the data format (e.g., DPO only supports pairwise preference data). To this end, prior works incorporate controllable generations for alignment to make language models learn multiple preferences and provide outputs with different preferences during inference if asked. Controllable generation also offers more flexibility with regard to data format (e.g., it supports pointwise preference data). Specifically, it uses different control tokens for different preferences during training and inference, making LLMs behave differently when required. Current controllable generation methods either use a special token or hand-crafted prompts as control tokens, and optimize them together with LLMs. As control tokens are typically much lighter than LLMs, this optimization strategy may not effectively optimize control tokens. To this end, we first use parameter-efficient tuning (e.g., prompting tuning and low-rank adaptation) to optimize control tokens and then fine-tune models for controllable generations, similar to prior works. Our approach, alignMEnt with parameter-Efficient Tuning (MEET), improves the quality of control tokens, thus improving controllable generation quality consistently by an apparent margin on two well-recognized datasets compared with prior works.","{'model': 'tldr@v2.0.0', 'text': 'The approach, alignMEnt with parameter-Efficient Tuning (MEET), improves the quality of control tokens, thus improving controllable generation quality consistently by an apparent margin on two well-recognized datasets compared with prior works.'}",https://arxiv.org/pdf/2310.00819
-spegti: structured prediction for efficient generative text-to-image models,Sadeep Jayasumana,"Modern text-to-image generation models produce high-quality images that are both photorealistic and faithful to the text prompts. However, this quality comes at significant computational cost: nearly all of these models are iterative and require running inference multiple times with large models. This iterative process is needed to ensure that different regions of the image are not only aligned with the text prompt, but also compatible with each other. In this work, we propose a light-weight approach to achieving this compatibility between different regions of an image, using a Markov Random Field (MRF) model. This method is shown to work in conjunction with the recently proposed Muse model. The MRF encodes the compatibility among image tokens at different spatial locations and enables us to significantly reduce the required number of Muse prediction steps. Inference with the MRF is significantly cheaper, and its parameters can be quickly learned through back-propagation by modeling MRF inference as a differentiable neural-network layer. Our full model, SPEGTI, uses this proposed MRF model to speed up Muse by 1.5X with no loss in output image quality.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a light-weight approach to achieving compatibility between different regions of an image, using a Markov Random Field (MRF) model, and uses this proposed MRF model to speed up Muse by 1.5X with no loss in output image quality.'}",https://arxiv.org/pdf/2308.10997
-efficient token based clone detection with flexible tokenization,H. Basit,"Code clones are similar code fragments that occur at multiple locations in a software system. Detection of code clones provides useful information for maintenance, reengineering, program understanding and reuse. Several techniques have been proposed to detect code clones. These techniques differ in the code representation used for analysis of clones, ranging from plain text to parse trees and program dependence graphs. Clone detection based on lexical tokens involves minimal code transformation and gives good results, but is computationally expensive because of the large number of tokens that need to be compared. We explored string algorithms to find suitable data structures and algorithms for efficient token based clone detection and implemented them in our tool Repeated Tokens Finder (RTF). Instead of using suffix tree for string matching, we use more memory efficient suffix array. RTF incorporates a suffix array based linear time algorithm to detect string matches. It also provides a simple and customizable tokenization mechanism. Initial analysis and experiments show that our clone detection is simple, scalable, and performs better than the previous well-known tools.","{'model': 'tldr@v2.0.0', 'text': 'String algorithms are explored to find suitable data structures and algorithms for efficient token based clone detection and implemented them in the tool Repeated Tokens Finder (RTF), which incorporates a suffix array based linear time algorithm to detect string matches.'}",
-a new coeff-token decoding method with efficient memory access in h.264/avc video coding standard,Y. Moon,"In general, a large amount of memory accesses are required for context-based adaptive variable-length coding decoding in H.264/AVC standard. It is a serious problem for applications such as videophone and digital multimedia broadcasting services because of the considerable power consumption of the memory access function. In order to reduce the memory access, we carefully examined the codewords in variable-length code tables (VLCTs) of the coeff-token and found some distinctive features from the structure of codewords. Based on the features, a new coeff-token decoding method with an efficient memory organization is proposed in this paper. The simulation results show that the proposed algorithm achieves an approximately 95% memory access saving without video-quality degradation, compared with conventional decoding","{'model': 'tldr@v2.0.0', 'text': 'A new coeff-token decoding method with an efficient memory organization is proposed and it is shown that the proposed algorithm achieves an approximately 95% memory access saving without video-quality degradation, compared with conventional decoding.'}",
-local self-attention over long text for efficient document retrieval,Sebastian Hofstätter,"Neural networks, particularly Transformer-based architectures, have achieved significant performance improvements on several retrieval benchmarks. When the items being retrieved are documents, the time and memory cost of employing Transformers over a full sequence of document terms can be prohibitive. A popular strategy involves considering only the first n terms of the document. This can, however, result in a biased system that under retrieves longer documents. In this work, we propose a local self-attention which considers a moving window over the document terms and for each term attends only to other terms in the same window. This local attention incurs a fraction of the compute and memory cost of attention over the whole document. The windowed approach also leads to more compact packing of padded documents in minibatches resulting in additional savings. We also employ a learned saturation function and a two-staged pooling strategy to identify relevant regions of the document. The Transformer-Kernel pooling model with these changes can efficiently elicit relevance information from documents with thousands of tokens. We benchmark our proposed modifications on the document ranking task from the TREC 2019 Deep Learning track and observe significant improvements in retrieval quality as well as increased retrieval of longer documents at moderate increase in compute and memory costs.","{'model': 'tldr@v2.0.0', 'text': 'A local self-attention which considers a moving window over the document terms and for each term attends only to other terms in the same window resulting in increased retrieval of longer documents at moderate increase in compute and memory costs is proposed.'}",https://arxiv.org/pdf/2005.04908
-the dynamic geometry of interaction machine: a token-guided graph rewriter,Koko Muroya,"In implementing evaluation strategies of the lambda-calculus, both correctness and efficiency of implementation are valid concerns. While the notion of correctness is determined by the evaluation strategy, regarding efficiency there is a larger design space that can be explored, in particular the trade-off between space versus time efficiency. Aiming at a unified framework that would enable the study of this trade-off, we introduce an abstract machine, inspired by Girard's Geometry of Interaction (GoI), a machine combining token passing and graph rewriting. We show soundness and completeness of our abstract machine, called the \emph{Dynamic GoI Machine} (DGoIM), with respect to three evaluations: call-by-need, left-to-right call-by-value, and right-to-left call-by-value. Analysing time cost of its execution classifies the machine as ``efficient'' in Accattoli's taxonomy of abstract machines.","{'model': 'tldr@v2.0.0', 'text': ""An abstract machine, inspired by Girard's Geometry of Interaction (GoI), a machine combining token passing and graph rewriting, andalysing time cost of its execution classifies the machine as ``efficient'' in Accattoli's taxonomy of abstract machines.""}",http://arxiv.org/pdf/1803.00427
-es‐plag: efficient and sensitive source code plagiarism detection tool for academic environment,Lisan Sulistiani,"Source code plagiarism detection using Running‐Karp‐Rabin Greedy‐String‐Tiling (RKRGST) is a common practice in academic environment. However, such approach is time‐inefficient (due to RKRGST's cubic time complexity) and insensitive (toward token subsequence rearrangement). This paper proposes ES‐Plag, a plagiarism detection tool featured with cosine‐based filtering and penalty mechanism to handle aforementioned issues. Cosine‐based filtering mitigates time‐inefficiency by excluding non‐potential pairs from RKRGST comparison; while penalty mechanism mitigates insensitivity by reducing the number of matched tokens with the number of matched subsequences prior similarity normalization. In addition to issue‐solving features, ES‐Plag is also featured with project‐based input, colorized adjacency similarity matrix, matched token highlighting, and various similarity algorithms (e.g., Cosine Similarity and Local Alignment). Three findings can be deducted from our evaluation. First, cosine‐based filtering boosts up time efficiency with a trade‐off in effectiveness. Second, penalty mechanism enhances sensitivity even though its improvement in terms of effectiveness is quite limited. Third, ES‐Plag's features are beneficial for examiners.","{'model': 'tldr@v2.0.0', 'text': 'ES‐Plag, a plagiarism detection tool featured with cosine‐based filtering and penalty mechanism to handle aforementioned issues, is proposed and its features are beneficial for examiners.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/cae.22066
-token routing: a power efficient method for securing aodv routing protocol,Leiyuan Li,"Despite the abundance of existing secure routing protocols in ad hoc and sensor networks, real world ad hoc and sensor network applications still usually opt for routing protocols without any security consideration, mostly because of the relatively high cost for enforcing security. In order to reduce the energy overhead when enhancing the security of routing protocols, we propose a token routing protocol (TRP), which is a routing security enhancement of the widely accepted AODV routing protocol. TRP employs hash-chain algorithm to generate a token, which is appended to the data packets to identify the authenticity of the routing packets and to choose correct route for data packets. As compared to the well-known AODV security enhancements such as SAODV, TRP provides comparable security with a significant reduction in energy consumption and routing packet delay by using the hash algorithm instead of public key cryptographic algorithms. Therefore TRP provides a practical energy-efficient option for routing security enhancements in AODV routing protocol","{'model': 'tldr@v2.0.0', 'text': 'A token routing protocol (TRP), which is a routing security enhancement of the widely accepted AODV routing protocol, employs hash-chain algorithm to generate a token, which is appended to the data packets to identify the authenticity of the routing packets and to choose correct route for data packets.'}",
-testing the tolerance principle: children form productive rules when it is more computationally efficient to do so,Kathryn D. Schuler,"During language acquisition, children must learn when to generalize a pattern – applying it broadly and to new words (‘add –ed’ in English) – and when to restrict generalization, storing the pattern only with specific lexical items. One effort to quantify the conditions for generalization, the Tolerance Principle, has been shown to accurately predict children’s generalizations in dozens of corpus-based studies. This principle hypothesizes that a general rule will be formed when it is computationally more efficient than storing lexical forms individually. It is formalized as: a rule R will generalize if the number of exceptions does not exceed the number of words in the category N divided by the natural log of N (N/lnN). Here we test the principle in an artificial language of 9 nonsense nouns. As predicted, children exposed to 5 regular forms and 4 exceptions generalized, applying the regular form to 100% of novel test words. Children exposed to 3 regular forms and 6 exceptions did not extend the rule, even though the token frequency of the regular form was still high in this condition. The Tolerance Principle thus appears to capture a basic principle of generalization in rule formation.","{'model': 'tldr@v2.0.0', 'text': 'The Tolerance Principle appears to capture a basic principle of generalization in rule formation, as children exposed to 5 regular forms and 4 exceptions generalized, applying the regular form to 100% of novel test words.'}",
-smyrf: efficient attention using asymmetric clustering,Giannis Daras,"We propose a novel type of balanced clustering algorithm to approximate attention. Attention complexity is reduced from $O(N^2)$ to $O(N \log N)$, where $N$ is the sequence length. Our algorithm, SMYRF, uses Locality Sensitive Hashing (LSH) in a novel way by defining new Asymmetric transformations and an adaptive scheme that produces balanced clusters. The biggest advantage of SMYRF is that it can be used as a drop-in replacement for dense attention layers without any retraining. On the contrary, prior fast attention methods impose constraints (e.g. queries and keys share the same vector representations) and require re-training from scratch. We apply our method to pre-trained state-of-the-art Natural Language Processing and Computer Vision models and we report significant memory and speed benefits. Notably, SMYRF-BERT outperforms (slightly) BERT on GLUE, while using $50\%$ less memory. We also show that SMYRF can be used interchangeably with dense attention before and after training. Finally, we use SMYRF to train GANs with attention in high resolutions. Using a single TPU, we were able to scale attention to 128x128=16k and 256x256=65k tokens on BigGAN on CelebA-HQ.","{'model': 'tldr@v2.0.0', 'text': 'The algorithm, SMYRF, uses Locality Sensitive Hashing (LSH) in a novel way by defining new Asymmetric transformations and an adaptive scheme that produces balanced clusters that can be used interchangeably with dense attention before and after training.'}",
-an efficient and practical local synchronous bandwidth allocation scheme for the timed-token mac protocol,Sijing Zhang,"This paper is concerned with deadline guarantees of synchronous messages with deadlines equal to periods, in a timed token ring network such as FDDI where the timed token medium access control (MAC) protocol is used. The timed token protocol guarantees a bounded access time and an average bandwidth for synchronous traffic. However, this guarantee alone, though necessary, is insufficient for guaranteeing the transmission of synchronous messages before their deadlines. To ensure timely delivery, the synchronous bandwidth must be carefully allocated to individual nodes. We propose and analyse an efficient and practical local synchronous bandwidth allocation (SBA) scheme. The new scheme performs better than any previously published as it calculates the synchronous bandwidth such that during the message period, the total synchronous transmission time definitely available (when judged only by local information) is exactly equal to the transmission time required. Our scheme also differs significantly from previously reported ones by explicitly taking into account the synchronous bandwidth allocation for message sets whose minimum message deadlines (D/sub min/) are less than twice the target token rotation time (TTRT), and consequently can apply to any synchronous message set (with D/sub min/>TTRT). The feasibility of the allocations produced by the proposed scheme and the worst case achievable utilisation of the scheme are also discussed.","{'model': 'tldr@v2.0.0', 'text': 'An efficient and practical local synchronous bandwidth allocation (SBA) scheme is proposed and analysed that performs better than any previously published as it calculates theynchronous bandwidth such that during the message period, the total synchronous transmission time definitely available is exactly equal to the transmission time required.'}",
-efficient global allocation of synchronous bandwidths for hard real-time communication with the timed token mac protocol,Sijing Zhang,"This paper presents an efficient global synchronous bandwidth allocation (SBA) scheme for guaranteeing synchronous message deadlines in a timed token network (such as FDDI) where the timed token medium access control (MAC) protocol is used. A local SBA scheme allocates synchronous bandwidth to a node using only information available locally to that node while a global scheme uses network wide information from all nodes. A local SBA scheme may be preferred for practical use to a global one due to its lower network management overhead. However, a global SBA scheme can greatly outperform a local scheme due to a more appropriate allocation resulted from the more complete global information. That is, use of a global SBA scheme may largely enhance the network ability of guaranteeing synchronous traffic. The global scheme proposed in this paper outperforms all existing SBA schemes for guaranteeing synchronous messages with deadlines equal to periods. Numerical examples are given to demonstrate the superiority of the proposed scheme to previously-proposed SBA schemes.","{'model': 'tldr@v2.0.0', 'text': 'The global scheme proposed in this paper outperforms all existing SBA schemes for guaranteeing synchronous messages with deadlines equal to periods and can greatly outperform a local scheme due to a more appropriate allocation resulted from the more complete global information.'}",
-highly space-efficient self-stabilizing depth-first token circulation for trees,F. Petit,"Self-stabilization was rst introduced by Dijkstra Dij74]: it is the property for a system to eventually recover itself a legitimate state after any perturbation modifying the memory state. Since Dijkstra original paper, the goal of many works has been to obtain self-stabilizing algorithms requiring the fewest possible number of states. A great deal of proposed algorithms are token circulations, very useful to solve distributed mutual exclusion. In every algorithm, the token is held by the processor enabled to make a move. Tchuente Tch81] showed that in such conditions, the expected state number lower bound to solve mutual exclusion over tree networks is 2 n Q n i=1 i , n is the number of processors, i is the neighbors number of each processor p i. In this paper, we use a weaker token formulation introduced by Villain Vil97]: a processor holds a token if it holds a particular state. This new light allows us to propose a self-stabilizing depth-rst token circulation for tree networks requiring fewer states than Tchuente's lower bound, i.e. (1 + 1) Q n i=2 ((i + 2) states only.","{'model': 'tldr@v2.0.0', 'text': ""A weaker token formulation introduced by Villain Vil97 is used, which allows for a self-stabilizing depth-rst token circulation for tree networks requiring fewer states than Tchuente's lower bound, i.e. (1 + 1) Q n i=2 ((i + 2) states only.""}",
-"space-efficient, distributed and self-stabilizing depth-first token circulation",C. Johnen,The notion of self stabilization was introduced by Dijkstra He de ned a system as self stabilizing when regardless of its initial state it is guaranteed to arrive at a legitimate state in a nite number of steps Such a property is very desirable for any distributed system because after any unexpected perturbation modifying the memory state the system eventually recovers and returns to a legitimate state without any outside intervention In this paper we are interested in a distributed self stabilizing depth rst token circulation protocol on an uniform rooted network no identi ers but a distinguished root As already noted a search algorithm together with a deterministic enumeration of the node s neighbors yields an algorithm determining a spanning tree Our contribution is improving the best up to now known space complexity for this problem from O log N to O log D where N is number of nodes and D is the network s degree Moreover we give a full proof of the algorithm correctness assuming the existence of a dis tributed demon,"{'model': 'tldr@v2.0.0', 'text': 'A distributed self stabilizing depth rst token circulation protocol on an uniform rooted network no identi ers but a distinguished root is interested and the best up to now known space complexity for this problem is improved.'}",
-asynchronous completion token an object behavioral pattern for efficient asynchronous event handling,Timothy H. Harrison,"Contemporary applications must respond to many types of events, ranging from user interface notifications to network messages. Delay-sensitive applications, such as network management systems, often perform long-runningoperations asynchronously to avoid blocking the processing of other pending events. When these asynchronous operations complete, applications may need more information than simply the notification itself to properly handle the event. This paper describes the Asynchronous Completion Token pattern, which allows applications to efficiently associate state with the completion of asynchronous operations.","{'model': 'tldr@v2.0.0', 'text': 'The Asynchronous Completion Token pattern is described, which allows applications to efficiently associate state with the completion of asynchronous operations.'}",
-e/sup 2/wtrp: an energy-efficient wireless token ring protocol,Zhenhua Deng,"A wireless token ring protocol (WTRP) built on top of the IEEE 802.11 distributed control function (DCF) provides guaranteed QoS in terms of high throughput and bounded transmission delay. It is also applicable in a small-scale wireless ad hoc network where all nodes can hear each other and where throughput and delay requirements are stringent. As nodes in the network have different traffic styles and the number of nodes ready for transmission is highly dynamic, the aggregate throughput and delay of WTRP is degraded. Furthermore, each node keeps on forwarding the token, even though it is not ready for transmission, thus the power consumed by inactive nodes is high. An energy-efficient wireless token ring protocol is proposed to reduce the power consumption. It is shown that E/sup 2/WTRP could prolong the standby time for inactive nodes two-fold.","{'model': 'tldr@v2.0.0', 'text': 'It is shown that E/sup 2/WTRP could prolong the standby time for inactive nodes two-fold and an energy-efficient wireless token ring protocol is proposed to reduce the power consumption.'}",
-warplda: a cache efficient o(1) algorithm for latent dirichlet allocation,Jianfei Chen,"Developing efficient and scalable algorithms for Latent Dirichlet Allocation (LDA) is of wide interest for many applications. Previous work has developed an O(1) Metropolis-Hastings sampling method for each token. However, the performance is far from being optimal due to random accesses to the parameter matrices and frequent cache misses. In this paper, we first carefully analyze the memory access efficiency of existing algorithms for LDA by the scope of random access, which is the size of the memory region in which random accesses fall, within a short period of time. We then develop WarpLDA, an LDA sampler which achieves both the best O(1) time complexity per token and the best O(K) scope of random access. Our empirical results in a wide range of testing conditions demonstrate that WarpLDA is consistently 5-15x faster than the state-of-the-art Metropolis-Hastings based LightLDA, and is comparable or faster than the sparsity aware F+LDA. With WarpLDA, users can learn up to one million topics from hundreds of millions of documents in a few hours, at an unprecedentedly throughput of 11G tokens per second.","{'model': 'tldr@v2.0.0', 'text': 'WarpLDA is developed, an LDA sampler which achieves both the best O(1) time complexity per token and thebest O(K) scope of random access, within a short period of time.'}",
-an efficient algorithm for solving the token distribution problem on k-ary d-cube networks,Claude G. Diderich,"In parallel programs where the problem data is dynamically generated, it is very useful to be able to rely on an efficient load balancing algorithm. The token distribution problem (TDP) is a generalization of the static load balancing problem. The paper describes a novel algorithm for solving the TDP for k-ary d-cube topology networks. Compared to other algorithms, our method is more general and does not rely on every processor knowing the exact number of tokens associated to each processor. The correctness of the algorithm is proved and its complexity is informally studied.<>","{'model': 'tldr@v2.0.0', 'text': 'A novel algorithm for solving the token distribution problem for k-ary d-cube topology networks is described, which is more general and does not rely on every processor knowing the exact number of tokens associated to each processor.'}",
-1 an efficient token based algorithm for mutual exclusion in distributed system,F. Kawsar,"Many distributed computations involving the sharing of resources among various processes require that a resource be allocated to a single process at a time. Therefore, mutual exclusion is a fundamental problem in any distributed computing system. This problem must be solved to synchronize the access to shared resources in order to maintain their consistency and integrity. The major goal of this paper is to get the reader acquainted with a new approach towards the ring based technique for mutual exclusion in a distributed system. An algorithm is proposed based on the idea of generating token by the competing processes to enter the critical section and thus eliminating idle time message passing and reducing communication overhead.","{'model': 'tldr@v2.0.0', 'text': 'An algorithm is proposed based on the idea of generating token by the competing processes to enter the critical section and thus eliminating idle time message passing and reducing communication overhead.'}",
-multimodal prompting with missing modalities for visual recognition supplementary materials,Yi-Lun Lee,"We show ablation studies for attention-level prompts in Figure 1 and Figure 2, which analyze the effect of prompting layers and prompt length respectively. The results are similar to the study of input-level prompts as shown in Section 4.3 of the main paper. In summary, the earlier prompting layers and more prompting layers improve the performance more. In addition, even with fewer parameters (i.e., reducing the prompt length to 2), the performance is still competitive.",,
-examining taiwanese university students’ multimodal multiple text comprehension: individual differences and epistemic prompting,Yuan‐Hsuan Lee,,,
-scaffolding university students' epistemic cognition during multimodal multiple-document reading: the effects of the epistemic prompting and the automated reflection report,Yuan‐Hsuan Lee,,"{'model': 'tldr@v2.0.0', 'text': ""Results showed that students relied primarily on YouTube videos for their MMDR with higher credibility ratings for the YouTube video than written texts, and epistemic prompting, along with the automatically generated reflection report, enhanced students' scores on constructed-response items for the experimental group and helped them create a coherent mental representation of information from diverse multimodal sources.""}",
-prompting as multimodal fusing,,"Tsimpoukelli et al. (2021) devise Frozen, em- 001 powering a language model to solve multi- 002 modal tasks by pretraining a vision encoder 003 whose outputs are prompts fed to the language 004 model. The vision encoder has a dual objec- 005 tive: Extracting image features and aligning 006 image/text representation spaces. We propose 007 to disentangle the objectives by using prompt 008 vectors to align the spaces; this lets the vision 009 encoder focus on extracting image features. We 010 show that this disentangled approach is modu- 011 lar and parameter-efficient for processing tasks 012 that involve two or more modalities. 013","{'model': 'tldr@v2.0.0', 'text': 'Frozen, Frozen is devise Frozen, powering a language model to solve multi- modal tasks by pretraining a vision encoder whose outputs are prompts fed to the language 004 model.'}",
-ddcot: duty-distinct chain-of-thought prompting for multimodal reasoning in language models,Ge Zheng,"A long-standing goal of AI systems is to perform complex multimodal reasoning like humans. Recently, large language models (LLMs) have made remarkable strides in such multi-step reasoning on the language modality solely by leveraging the chain of thought (CoT) to mimic human thinking. However, the transfer of these advancements to multimodal contexts introduces heightened challenges, including but not limited to the impractical need for labor-intensive annotation and the limitations in terms of flexibility, generalizability, and explainability. To evoke CoT reasoning in multimodality, this work first conducts an in-depth analysis of these challenges posed by multimodality and presents two key insights:""keeping critical thinking""and""letting everyone do their jobs""in multimodal CoT reasoning. Furthermore, this study proposes a novel DDCoT prompting that maintains a critical attitude through negative-space prompting and incorporates multimodality into reasoning by first dividing the reasoning responsibility of LLMs into reasoning and recognition and then integrating the visual recognition capability of visual models into the joint reasoning process. The rationales generated by DDCoT not only improve the reasoning abilities of both large and small language models in zero-shot prompting and fine-tuning learning, significantly outperforming state-of-the-art methods but also exhibit impressive generalizability and explainability.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a novel DDCoT prompting that maintains a critical attitude through negative-space prompting and incorporates multimodality into reasoning by first dividing the reasoning responsibility of LLMs into reasoning and recognition and then integrating the visual recognition capability of visual models into the joint reasoning process.'}",
-efficient multimodal fusion via interactive prompting,Yaowei Li,"Large-scale pre-training has brought unimodal fields such as computer vision and natural language processing to a new era. Following this trend, the size of multimodal learning models constantly increases, leading to an urgent need to reduce the massive computational cost of finetuning these models for downstream tasks. In this paper, we propose an efficient and flexible multimodal fusion method, namely PMF, tailored for fusing unimodally pretrained transformers. Specifically, we first present a modular multimodal fusion framework that exhibits high flexibility and facilitates mutual interactions among different modalities. In addition, we disentangle vanilla prompts into three types in order to learn different optimizing objectives for multimodal learning. It is also worth noting that we propose to add prompt vectors only on the deep layers of the unimodal transformers, thus significantly reducing the training memory usage. Experiment results show that our proposed method achieves comparable performance to several other multimodal finetuning methods with less than 3% trainable parameters and up to 66% saving of training memory usage.","{'model': 'tldr@v2.0.0', 'text': 'An efficient and flexible multimodal fusion method, namely PMF, tailored for fusing unimodally pretrained transformers and achieves comparable performance to several other multimodals finetuning methods with less than 3% trainable parameters and up to 66% saving of training memory usage.'}",https://arxiv.org/pdf/2304.06306
-prompting chatgpt in mner: enhanced multimodal named entity recognition with auxiliary refined knowledge,Jinyuan Li,"Multimodal Named Entity Recognition (MNER) on social media aims to enhance textual entity prediction by incorporating image-based clues. Existing studies mainly focus on maximizing the utilization of pertinent image information or incorporating external knowledge from explicit knowledge bases. However, these methods either neglect the necessity of providing the model with external knowledge, or encounter issues of high redundancy in the retrieved knowledge. In this paper, we present PGIM -- a two-stage framework that aims to leverage ChatGPT as an implicit knowledge base and enable it to heuristically generate auxiliary knowledge for more efficient entity prediction. Specifically, PGIM contains a Multimodal Similar Example Awareness module that selects suitable examples from a small number of predefined artificial samples. These examples are then integrated into a formatted prompt template tailored to the MNER and guide ChatGPT to generate auxiliary refined knowledge. Finally, the acquired knowledge is integrated with the original text and fed into a downstream model for further processing. Extensive experiments show that PGIM outperforms state-of-the-art methods on two classic MNER datasets and exhibits a stronger robustness and generalization capability.","{'model': 'tldr@v2.0.0', 'text': 'PGIM is a two-stage framework that aims to leverage ChatGPT as an implicit knowledge base and enable it to heuristically generate auxiliary knowledge for more efficient entity prediction and exhibits a stronger robustness and generalization capability.'}",
-initial images: using image prompts to improve subject representation in multimodal ai generated art,Han Qiao,"Advances in text-to-image generative models have made it easier for people to create art by just prompting models with text. However, creating through text leaves users with limited control over the final composition or the way the subject is represented. A potential solution is to use image prompts alongside text prompts to condition the model. To better understand how and when image prompts can improve subject representation in generations, we conduct an annotation experiment to quantify their effect on generations of abstract, concrete plural, and concrete singular subjects. We find that initial images improved subject representation across all subject types, with the most noticeable improvement in concrete singular subjects. In an analysis of different types of initial images, we find that icons and photos produced high quality generations of different aesthetics. We conclude with design guidelines for how initial images can improve subject representation in AI art.",,
-multimodal multi-task stealth assessment for reflection-enriched game-based learning,Anisha Gupta,"Game-based learning environments enable effective and engaging learning experiences that can be dynamically tailored to students. There is growing interest in the role of reflection in supporting student learning in game-based learning environments. By prompting students to periodically stop and reflect on their learning processes, it is possible to gain insight into students’ perceptions of their knowledge and problem-solving progress, which can in turn inform adaptive scaffolding to improve student learning outcomes. Given the positive relationship between student reflection and learning, we investigate the benefits of jointly modeling post-test score and reflection depth using a multimodal, multitask stealth assessment framework. Specifically, we present a gated recurrent unit-based multi-task stealth assessment framework that takes as input multimodal data streams (e.g., game trace logs, pre-test data, natural language responses to in-game reflection prompts) to jointly predict post-test scores and written reflection depth scores. Evaluation results demonstrate that the multimodal multi-task model outperforms single-task neural models that utilize subsets of the modalities, as well as non-neural baselines such as random forest regressors. Our multi-task stealth assessment framework for measuring students’ content knowledge and reflection depth during game-based learning shows significant promise for supporting student learning and improved reflection.","{'model': 'tldr@v2.0.0', 'text': 'A gated recurrent unit-based multi-task stealth assessment framework that takes as input multimodal data streams to jointly predict post-test scores and written reflection depth scores shows significant promise for supporting student learning and improved reflection.'}",
-visual chain of thought: bridging logical gaps with multimodal infillings,Daniel Philip Rose,"Recent advances in large language models elicit reasoning in a chain of thought that allows models to decompose problems in a human-like fashion. Though this paradigm improves multi-step reasoning ability in language models, it is limited by being unimodal and applied mainly to question-answering tasks. We claim that incorporating visual augmentation into reasoning is essential, especially for complex, imaginative tasks. Consequently, we introduce VCoT, a novel method that leverages chain of thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data. Our method uses visual guidance to generate synthetic multimodal infillings that add consistent and novel information to reduce the logical gaps for downstream tasks that can benefit from temporal reasoning, as well as provide interpretability into models' multi-step reasoning. We apply VCoT to the Visual Storytelling and WikiHow summarization datasets and demonstrate through human evaluation that VCoT offers novel and consistent synthetic data augmentation beating chain of thought baselines, which can be used to enhance downstream performance.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces VCoT, a novel method that leverages chain of thought prompting with vision-language grounding to recursively bridge the logical gaps within sequential data and applies it to the Visual Storytelling and WikiHow summarization datasets.'}",http://arxiv.org/pdf/2305.02317
-pulmonary artery intimal sarcoma: a diagnostic challenge using a multimodal approach,Rima Nakrani,"Pulmonary artery intimal sarcoma (PAIS) is a rare tumor without clear syndromic presentation other than nonspecific symptoms of cough, dyspnea, and weight loss. This diagnosis is difficult due to challenging radiographic interpretations of multiple imaging modalities. We present a case of a 60-year-old male, who presented to his pulmonologist and underwent a CT chest with IV contrast that initially suggested primary lung carcinoma. CT angiogram showed significant vascular filling defects suspicious of an intravascular mass, rather than vascular invasion by lung lesions. The PET/CT scans further suggested a malignant process, but indistinguishable between an extravascular or intravascular etiology. Taking these results together, they suggested an intravascular malignancy, prompting a tissue biopsy, which ultimately led to a diagnosis of PAIS with metastases. Establishing a definitive diagnosis is essential as treatment and prognosis are different for sarcoma compared to carcinoma. There is no standard treatment to date, and management often includes a multidisciplinary approach involving surgery, radiation, chemotherapy, and targeted therapy. PAIS is a rare entity that cannot be diagnosed clinically and needs a multimodality approach for its diagnosis.","{'model': 'tldr@v2.0.0', 'text': 'A case of a 60-year-old male, who presented to his pulmonologist and underwent a CT chest with IV contrast that initially suggested primary lung carcinoma, which ultimately led to a diagnosis of PAIS with metastases, which needs a multimodality approach for its diagnosis.'}",https://downloads.hindawi.com/journals/crim/2020/6795272.pdf
-using multimodal learning analytics to study collaboration on discussion groups,Fabián Riquelme,,"{'model': 'tldr@v2.0.0', 'text': 'The main aim of this work is to develop a computational environment to both analyze and visualize student discussion groups working in a collaborative way to accomplish a task and show that this information can help to support complex decision-making processes.'}",
-prompting technology and persons with dementia: the significance of context and communication,Rachel Braley,"Background and Objectives Smart home auto-prompting has the potential to increase the functional independence of persons with dementia (PWDs) and decrease caregiver burden as instrumental activities of daily living (IADLs) are completed at home. To improve prompting technologies, we sought to inductively understand how PWDs responded to auto-prompting while performing IADL tasks. Research Design and Methods Fifteen PWDs completed eight IADLs in a smart home testbed and received a hierarchy of verbal auto-prompts (indirect, direct, multimodal) as needed for task completion. Two researchers viewed archived videos and recorded the observed behaviors of the PWDs and their reflections watching the PWDs. Using qualitative descriptive methods, an interdisciplinary analytic team reviewed transcripts and organized data into themes using content analysis. Results Context and Communication emerged as the major themes, suggesting that positive user experiences will require auto-prompting systems to account for a multitude of contextual factors (individual and environmental) such as level of cognitive impairment, previous exposure to task, and familiarity of environment. Communicating with another human rather than an automated prompting system may be important if individuals begin to exhibit signs of stress while completing activities. Discussion and Implications Additional work is needed to create auto-prompting systems that provide specific, personalized, and flexible prompts. Holistic conceptualization of ""successful task completion"" is needed and a positive end-user experience will be key to utility. Such systems will benefit from including positive reinforcement, training, and exploration of how, and whether, direct human involvement can be minimized during the provision of in-home care.","{'model': 'tldr@v2.0.0', 'text': 'Context and Communication emerged as the major themes, suggesting that positive user experiences will require auto-prompting systems to account for a multitude of contextual factors such as level of cognitive impairment, previous exposure to task, and familiarity of environment.'}",https://academic.oup.com/gerontologist/article-pdf/59/1/101/27456552/gny071.pdf
-exploring collaborative writing of user stories with multimodal learning analytics: a case study on a software engineering course,Rene Noel,"Software engineering is the application of principles used in engineering design, development, testing, deployment, and management of software systems. One of the software engineering’s approaches, highly used in new industries, is agile development. User stories are a commonly used notation to capture user requirements in agile development. Nevertheless, for the elaboration of user stories, a high level of collaboration with the client is necessary. This professional skill is rarely measured or evaluated in educational contexts. The present work approaches collaboration in software engineering students through multimodal learning analytics, modeling, and evaluating students’ collaboration while they are writing user stories. For that, we used multidirectional microphones in order to derive social network analysis metrics related to collaboration (permanence and prompting) together with human-annotated information (quality of the stories and productivity). Results show that groups with a lower productivity in writing user stories and less professional experience in managing software requirements present a non-collaborative behavior more frequently, and that teams with a fewer number of interventions are more likely to produce a greater number of user stories. Moreover, although low experience subjects produced more user stories, a greater productivity of the most experienced subjects was not statistically verified. We believe that these types of initiatives will allow the measurement and early development of such skills in university students.","{'model': 'tldr@v2.0.0', 'text': 'Results show that groups with a lower productivity in writing user stories and less professional experience in managing software requirements present a non-collaborative behavior more frequently, and that teams with a fewer number of interventions are more likely to produce a greater number of user stories.'}",
-multimodal pipeline : a generic approach for handling multimodal data for supporting learning,D. D. Mitri,"In this demo paper, we introduce the Multimodal Pipeline, a prototypical approach for the collection, storing, annotation, processing and exploitation of multimodal data for supporting learning. At the current stage of development, the Multimodal Pipeline consists of two relevant prototypes: 1) Multimodal Learning Hub for the collection and storing of sensor data from multiple applications and 2) the Visual Inspection Tool for visualisation and annotation of the recorded sessions. The Multimodal Pipeline is designed to be a flexible system useful for supporting psychomotor skills in a variety of learning scenarios such as presentation skills, medical simulation with patient manikins or calligraphy learning. The Multimodal Pipeline can be configured to serve different support strategies, including detecting mistakes and prompting live feedback in an intelligent tutoring system or stimulating self-reflection through a learning analytics dashboard.","{'model': 'tldr@v2.0.0', 'text': 'The Multimodal Pipeline is designed to be a flexible system useful for supporting psychomotor skills in a variety of learning scenarios such as presentation skills, medical simulation with patient manikins or calligraphy learning.'}",
-multimodal simile: the “when” meme in social media discourse,Adrian Lou,"This paper analyzes the “when” meme, a popular internet meme, which prototypically juxtaposes a when clause with an ostensibly unrelated image. Despite the initial incongruity, I contend this image prompts selective mapping between verbal and visual elements to produce a multimodal simile. First, I attempt to define and more clearly distinguish simile from metaphor. Second, I show how this multimodal simile exhibits unique viewpoint mapping by prompting audiences to subsume viewpoints that are both unfamiliar and bizarre. Third, I connect the like construction in simile with the like reported speech marker to show how both concepts are intimately related. Ultimately, the paper seeks to contribute to studies of simile by bolstering its ties with multimodality, blending, metonymy, viewpoint, and embodiment.",,
-multimodal treatment approach in a patient with multiple intracranial myxomatous aneurysms.,D. Penn,"The most common primary cardiac tumor is myxoma, typically originating in the left atrium. Emboli to the central nervous system can cause cerebral infarction or, rarely, seed tumor growth within vessel walls, causing myxomatous aneurysms. Fewer than 60 myxomatous aneurysms have been reported, including 2 cases in children. Here, the authors describe 2 different growing myxomatous aneurysms in a child successfully managed using a combined multidisciplinary approach. A 12-year-old boy developed a sudden headache, diplopia, gait instability, and speech difficulty. Magnetic resonance imaging revealed a left parietal hemorrhage and multifocal cerebral infarction, suspicious for an embolic etiology. A cardiac myxoma was identified in the left atrium and resected. Follow-up cranial vasculature imaging demonstrated multiple intracranial myxomatous aneurysms. These lesions were followed up, and serial imaging identified marked growth of 2 of them (right occipital and left parietal), prompting invasive intervention. The deep occipital lesion was better suited to endovascular treatment, while the superficial parietal lesion was amenable to resection. The patient underwent embolization of an enlarging fusiform aneurysm of the distal right posterior cerebral artery, followed by a left parietal craniotomy for a lesion of the distal left middle cerebral artery. Both procedures were performed without complications and achieved successful obliteration of the lesions, as confirmed by catheter angiography at the 30-month follow-up. To the authors' knowledge, this report illustrates the first combined endovascular and open surgical treatment of 2 myxomatous aneurysms in a single patient. While acknowledging the rarity of this condition, this report illustrates the clinical manifestations and treatment challenges posed by myxoma and details a successful strategy that could be employed in similar scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This report illustrates the first combined endovascular and open surgical treatment of 2 myxomatous aneurysms in a child successfully managed using a combined multidisciplinary approach and details a successful strategy that could be employed in similar scenarios.'}",
-changing the antibiotic prescribing of general practice registrars: the chap study protocol for a prospective controlled study of a multimodal educational intervention,M. V. van Driel,,"{'model': 'tldr@v2.0.0', 'text': 'It is hypothesised that intervening while early-career GPs are still developing their practice patterns and prescribing habits will result in better adherence to evidence-based guidelines as manifested by lower antibiotic prescribing rates for URTIs and acute bronchitis.'}",https://bmcfampract.biomedcentral.com/track/pdf/10.1186/s12875-016-0470-7
-on the importance of the multimodal approach to discourse markers: a pragmatic view,Kazuki Hata,"Discourse markers (DMs) are pragmatic devices, which operate beyond the traditional word or phrase classification and have little to no effect on the propositional meaning. They have significant functions with regard to organising ongoing discourse by linking discourse segments, illustrating the current discourse structure to the interlocutors. Recent studies have discovered that DMs are truly multifunctional and thus play communicative roles at different dimensions simultaneously, ranging from managing discourse by denoting the speaker’s attitude, monitoring turn-taking activities to prompting the shared-knowledge between interlocutors. Nevertheless, few attentions have been paid on the fact that human communication is multimodal, wherein discourse includes both spoken language and gestures. Most DM studies predominantly investigate the use of DMs in text-based frameworks and therefore do not do justice to the non-linguistic DM functions, which are not easily amenable to text descriptions. The goal of this paper is to draw attention to this gap from the pragmatic perspective on DMs, demonstrating the importance of multimodal approaches to the study area.",,
-discordant embryonic aneuploidy testing and prenatal ultrasonography prompting androgen insensitivity syndrome diagnosis,J. Franasiak,"BACKGROUND: Multimodal prenatal screening for developmental pathology is increasingly common. In this case, definitive prenatal diagnosis of androgen insensitivity syndrome was diagnosed after discordant results from karyotypes determined by embryonic preimplantation genetic screening and antenatal ultrasound results. CASE: A 38-year-old woman, gravida 2 para 0010, undergoing in vitro fertilization with preimplantation genetic screening transferred one male and one female embryo. An anatomic ultrasonogram revealed two fetuses with female genitalia. Cell-free fetal DNA analyzed using noninvasive prenatal screening demonstrated Y chromosome material, and amniocentesis confirmed one 46,XX and one 46,XY fetus. Sequencing of the androgen receptor for the 46,XY fetus identified a mutation. CONCLUSION: With increased use, discordance among prenatal testing modalities such as preimplantation genetic screening, noninvasive prenatal screening, and ultrasonography will become more common requiring expert navigation to identify true pathology.","{'model': 'tldr@v2.0.0', 'text': 'With increased use, discordance among prenatal testing modalities such as preimplantation genetic screening, noninvasive prenatal screening, and ultrasonography will become more common requiring expert navigation to identify true pathology.'}",
-the use of multimodal technologies to enhance reflective writing in teacher education,L. Adie,,,
-refractory postsurgical pyoderma gangrenosum in a patient with beckwith wiedemann syndrome: response to multimodal therapy,Faiza Fakhar,"Pyoderma gangrenosum (PG) is a rare neutrophilic dermatosis that may be difficult to diagnose and treat. We presented a 41-year-old woman who required skin grafting following third-degree burns to her left breast. She suffered recurrent graft dehiscence and infections over many years, prompting elective bilateral reduction mammoplasty. She subsequently developed suture margin ulcerations unresponsive to topical therapies and antibiotics. Skin biopsies were non-specific, and a clinical diagnosis of PG was established. Although initially responsive to corticosteroids, wounds promptly recurred following steroid taper. She was treated unsuccessfully with various immunomodulatory agents and underwent elective bilateral mastectomy. Following a mastectomy, she developed progressive deep chest wall ulcerations. She failed numerous immunomodulatory treatments, surgical wound closure and negative pressure wound therapy. Ultimately, treatment with adalimumab, mycophenolate mofetil and prednisone, in addition to hyperbaric oxygen therapy facilitated progressive healing. Our case highlights the role of collaborative multimodal therapy for the treatment of refractory PG.","{'model': 'tldr@v2.0.0', 'text': 'Treatment with adalimumab, mycophenolate mofetil and prednisone, in addition to hyperbaric oxygen therapy facilitated progressive healing and highlights the role of collaborative multimodal therapy for the treatment of refractory PG.'}",https://europepmc.org/articles/pmc3822173?pdf=render
-multi-step pericellular proteolysis controls the transition from individual to collective cancer cell invasion,Katarina Wolf,,"{'model': 'tldr@v2.0.0', 'text': 'Both ECM track widening and transition to multicellular invasion are dependent on MT1-MMP-mediated collagenolysis, shown by broad-spectrum protease inhibition and RNA interference, and invasive migration and proteolytic ECM remodelling are interdependent processes that control tissue micropatterning and macrop atterning.'}",
-svit: scaling up visual instruction tuning,Bo Zhao,"Thanks to the emerging of foundation models, the large language and vision models are integrated to acquire the multimodal ability of visual captioning, dialogue, question answering, etc. Although existing multimodal models present impressive performance of visual understanding and reasoning, their limits are still largely under-explored due to the scarcity of high-quality instruction tuning data. To push the limits of multimodal capability, we Sale up Visual Instruction Tuning (SVIT) by constructing a dataset of 4.2 million visual instruction tuning data including 1.6M conversation question-answer (QA) pairs, 1.6M complex reasoning QA pairs, 1.0M referring QA pairs and 106K detailed image descriptions. Besides the volume, the proposed dataset is also featured by the high quality and rich diversity, which is generated by prompting GPT-4 with the abundant manual annotations of images. We empirically verify that training multimodal models on SVIT can significantly improve the multimodal performance in terms of visual perception, reasoning and planing.","{'model': 'tldr@v2.0.0', 'text': 'It is empirically verified that training multimodal models on SVIT can significantly improve the multi-modal performance in terms of visual perception, reasoning and planing.'}",https://arxiv.org/pdf/2307.04087
-user behavior recognition for an automatic prompting system - a structured approach based on task analysis,Christian Peters,"In this paper, we describe a structured approach for user behavior recognition in an automatic prompting system that assists users with cognitive disabilities in the task of brushing their teeth. We analyze the brushing task using qualitative data analysis. The results are a hierarchical decomposition of the task and the identification of environmental configurations during subtasks. We develop a hierarchical recognition framework based on the results of task analysis: We extract a set of features from multimodal sensors which are discretized into the environmental configuration in terms of states of objects involved in the brushing task. We classify subtasks using a Bayesian Network (BN) classifier and a Bayesian Filtering approach. We compare three variants of the BN using different observation models (IU, NaiveBayes and Holistic) with a maximum-margin classifier (multi-class SVM). We present recognition results on 18 trials with regular users and found the BN with a NaiveBayes observation model to produce the best recognition rates of 84.5% on avg.","{'model': 'tldr@v2.0.0', 'text': 'A structured approach for user behavior recognition in an automatic prompting system that assists users with cognitive disabilities in the task of brushing their teeth and develops a hierarchical recognition framework based on the results of task analysis.'}",
-"the cognitive hearing science perspective on perceiving, understanding, and remembering language: the elu model",J. Rönnberg,"The review gives an introductory description of the successive development of data patterns based on comparisons between hearing-impaired and normal hearing participants’ speech understanding skills, later prompting the formulation of the Ease of Language Understanding (ELU) model. The model builds on the interaction between an input buffer (RAMBPHO, Rapid Automatic Multimodal Binding of PHOnology) and three memory systems: working memory (WM), semantic long-term memory (SLTM), and episodic long-term memory (ELTM). RAMBPHO input may either match or mismatch multimodal SLTM representations. Given a match, lexical access is accomplished rapidly and implicitly within approximately 100–400 ms. Given a mismatch, the prediction is that WM is engaged explicitly to repair the meaning of the input – in interaction with SLTM and ELTM – taking seconds rather than milliseconds. The multimodal and multilevel nature of representations held in WM and LTM are at the center of the review, being integral parts of the prediction and postdiction components of language understanding. Finally, some hypotheses based on a selective use-disuse of memory systems mechanism are described in relation to mild cognitive impairment and dementia. Alternative speech perception and WM models are evaluated, and recent developments and generalisations, ELU model tests, and boundaries are discussed.",,https://www.frontiersin.org/articles/10.3389/fpsyg.2022.967260/pdf
-"cyberbullying conceptualization, characterization and detection in social media – a systematic literature review",Wai Hong Woo,"Social media has become the primary form of communication wherein users can share intimate moments online through photos, videos, or posts. At a glance, while this greatly improves interconnectivity between people, it also increases the propensity towards unrestricted acts of Cyberbullying, prompting the need for a data-centric detection system. Unfortunately, these sites generate much metadata, which begs the need for complex Machine Learning (ML) classifiers to categorize these acts accurately. Prior studies on the subject matter only target the topics of Conceptualization, Characterization, and Classification of Cyberbullying individually, so this research aims to provide a more holistic understanding of the subject matter in a continuous, synthesized format. This study found that Cyberbullying differs from Traditional Bullying in key areas of Repetition and Intention. Moreover, multimodal feature sets, as opposed to single feature sets, significantly improve ML classifiers' performance. Lastly, the selection of appropriate ML classifiers and performance metrics is context-dependent. The result of this study presents a consolidated view of relevant parties tackling different aspects of an ML-based automated Cyberbullying detection system so that those assigned tasks can approach them strategically","{'model': 'tldr@v2.0.0', 'text': ""This study found that Cyberbullying differs from Traditional Bullying in key areas of Repetition and Intention, and multimodal feature sets, as opposed to single feature Sets, significantly improve ML classifiers' performance.""}",https://journals.iium.edu.my/kict/index.php/IJPCC/article/download/374/223
-mapping genomic and epigenomic evolution in cancer ecosystems,T. Ushijima,"Description Cancer is a major cause of global mortality underpinned by genomic and epigenomic derangements. Here, we highlight the importance of multimodal data integration in understanding the molecular evolution of malignant cell states across the cancer life cycle. The widespread presence of driver mutations and epigenetic alterations in normal-appearing tissues is prompting a reassessment of how cancer initiation is defined. In later-stage cancers, studying the roles of clonal selection, epigenomic adaptation, and persister cells in metastasis and therapy resistance is an emerging field. Finally, the importance of tumor ecosystems in driving cancer development is being unraveled by single-cell and spatial technologies at unprecedented resolution. Improving cancer risk assessment and accelerating therapeutic discovery for patients will require robust, comprehensive, and integrated temporal, spatial, and multilevel tumor atlases across the cancer life cycle.","{'model': 'tldr@v2.0.0', 'text': 'Improving cancer risk assessment and accelerating therapeutic discovery for patients will require robust, comprehensive, and integrated temporal, spatial, and multilevel tumor atlases across the cancer life cycle.'}",
-"upgpt: universal diffusion model for person image generation, editing and pose transfer",Soon Yau Cheong,"Text-to-image models (T2I) such as StableDiffusion have been used to generate high quality images of people. However, due to the random nature of the generation process, the person has a different appearance e.g. pose, face, and clothing, despite using the same text prompt. The appearance inconsistency makes T2I unsuitable for pose transfer. We address this by proposing a multimodal diffusion model that accepts text, pose, and visual prompting. Our model is the first unified method to perform all person image tasks - generation, pose transfer, and mask-less edit. We also pioneer using small dimensional 3D body model parameters directly to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining the person's appearance.","{'model': 'tldr@v2.0.0', 'text': ""This model is the first unified method to perform all person image tasks - generation, pose transfer, and mask-less edit and pioneer using small dimensional 3D body model parameters directly to demonstrate new capability - simultaneous pose and camera view interpolation while maintaining the person's appearance.""}",https://arxiv.org/pdf/2304.08870
-oncolytic adenovirus and gene therapy with epha2-bite for the treatment of pediatric high-grade gliomas,C. Arnone,"Background Pediatric high-grade gliomas (pHGGs) are among the most common and incurable malignant neoplasms of childhood. Despite aggressive, multimodal treatment, the outcome of children with high-grade gliomas has not significantly improved over the past decades, prompting the development of innovative approaches. Methods To develop an effective treatment, we aimed at improving the suboptimal antitumor efficacy of oncolytic adenoviruses (OAs) by testing the combination with a gene-therapy approach using a bispecific T-cell engager (BiTE) directed towards the erythropoietin-producing human hepatocellular carcinoma A2 receptor (EphA2), conveyed by a replication-incompetent adenoviral vector (EphA2 adenovirus (EAd)). The combinatorial approach was tested in vitro, in vivo and thoroughly characterized at a molecular level. Results After confirming the relevance of EphA2 as target in pHGGs, documenting a significant correlation with worse clinical outcome of the patients, we showed that the proposed strategy provides significant EphA2-BiTE amplification and enhanced tumor cell apoptosis, on coculture with T cells. Moreover, T-cell activation through an agonistic anti-CD28 antibody further increased the activation/proliferation profiles and functional response against infected tumor cells, inducing eradication of highly resistant, primary pHGG cells. The gene-expression analysis of tumor cells and T cells, after coculture, revealed the importance of both EphA2-BiTE and costimulation in the proposed system. These in vitro observations translated into significant tumor control in vivo, in both subcutaneous and a more challenging orthotopic model. Conclusions The combination of OA and EphA2-BiTE gene therapy strongly enhances the antitumor activity of OA, inducing the eradication of highly resistant tumor cells, thus supporting the clinical translation of the approach.","{'model': 'tldr@v2.0.0', 'text': 'The combination of OA and EphA2-BiTE gene therapy strongly enhances the antitumor activity of OAs, inducing the eradication of highly resistant tumor cells, thus supporting the clinical translation of the approach.'}",https://jitc.bmj.com/content/jitc/9/5/e001930.full.pdf
-unlocking multimedia capabilities of gigantic pretrained language models,Boyang Li,"Benefitting from unprecedented computational power, massive data throughput, and superhuman memory, large language models (LLMs) are fundamentally transforming multimodal machine learning. An LLM can be analogized to an enormous treasure box guarded by a lock. It contains extensive knowledge, but it can be non-trivial to access and apply appropriate knowledge to solve the problem at hand. Researchers have developed many techniques to unlock the capabilities of LLMs. Some well-known examples include chain-of-thought prompting, ""let's think step by step'', and instruction tuning. In this talk, I will discuss techniques to unlock the capability of LLMs to process both visual and linguistic information. VisualGPT is one of the earliest works that finetunes an LLM for a vision-language task. InstructBLIP is an instruction-tuned large vision-language model, which set new states of the art on several vision-language tasks and snatched top positions on several comprehensive evaluation suites. In addition, I will talk about how to unlock zero-shot capabilities without end-to-end finetuning, or any form of finetuning at all. In Plug-and-Play VQA and Img2LLM, we achieve excellent results on visual question-answering datasets by connecting existing pretrained models using natural language and model interpretations, demonstrating a feasible alternative to the mainstream finetuning approach. Finally, I will describe a new multimodal dataset, Synopses of Movie Narratives, or SyMoN, for story understanding, which constitutes a new challenge for large vision-language models. I will argue that story understanding is an important objective in the pursuit of artificial general intelligence (AGI) because stories are a preeminent form of human communication and story understanding requires many AGI capabilities such as cause-effect reasoning and theory of mind. Compared to other multimodal story datasets, the special advantages of SyMoN include (1) event descriptions at the right level of granularity, (2) abundant mental state descriptions, (3) the use of diverse storytelling techniques, and (4) the provision of easy-to-use automatic performance evaluation.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that story understanding is an important objective in the pursuit of artificial general intelligence (AGI) because stories are a preeminent form of human communication and story understanding requires many AGI capabilities such as cause-effect reasoning and theory of mind.'}",
-point-of-critical-care diagnostics for sepsis enabled by multiplexed micro and nanosensing technologies.,B. Ashley,"Sepsis is responsible for the highest economic and mortality burden in critical care settings around the world, prompting the World Health Organization in 2018 to designate it as a global health priority. Despite its high universal prevalence and mortality rate, a disproportionately low amount of sponsored research funding is directed toward diagnosis and treatment of sepsis, when early treatment has been shown to significantly improve survival. Additionally, current technologies and methods are inadequate to provide an accurate and timely diagnosis of septic patients in multiple clinical environments. For improved patient outcomes, a comprehensive immunological evaluation is critical which is comprised of both traditional testing and quantifying recently proposed biomarkers for sepsis. There is an urgent need to develop novel point-of-care, low-cost systems which can accurately stratify patients. These point-of-critical-care sensors should adopt a multiplexed approach utilizing multimodal sensing for heterogenous biomarker detection. For effective multiplexing, the sensors must satisfy criteria including rapid sample to result delivery, low sample volumes for clinical sample sparring, and reduced costs per test. A compendium of currently developed multiplexed micro and nano (M/N)-based diagnostic technologies for potential applications toward sepsis are presented. We have also explored the various biomarkers targeted for sepsis including immune cell morphology changes, circulating proteins, small molecules, and presence of infectious pathogens. An overview of different M/N detection mechanisms are also provided, along with recent advances in related nanotechnologies which have shown improved patient outcomes and perspectives on what future successful technologies may encompass. This article is categorized under: Diagnostic Tools > Biosensing.","{'model': 'tldr@v2.0.0', 'text': 'An urgent need to develop novel point-of-care, low-cost systems which can accurately stratify patients and explore the various biomarkers targeted for sepsis including immune cell morphology changes, circulating proteins, small molecules, and presence of infectious pathogens.'}",https://rss.onlinelibrary.wiley.com/doi/am-pdf/10.1002/wnan.1701
-winder: linking speech and visual objects to support communication in asynchronous collaboration,Tae Soo Kim,"Team members commonly collaborate on visual documents remotely and asynchronously. Particularly, students are frequently restricted to this setting as they often do not share work schedules or physical workspaces. As communication in this setting has delays and limits the main modality to text, members exert more effort to reference document objects and understand others’ intentions. We propose Winder, a Figma plugin that addresses these challenges through linked tapes—multimodal comments of clicks and voice. Bidirectional links between the clicked-on objects and voice recordings facilitate understanding tapes: selecting objects retrieves relevant recordings, and playing recordings highlights related objects. By periodically prompting users to produce tapes, Winder preemptively obtains information to satisfy potential communication needs. Through a five-day study with eight teams of three, we evaluated the system’s impact on teams asynchronously designing graphical user interfaces. Our findings revealed that producing linked tapes could be as lightweight as face-to-face (F2F) interactions while transmitting intentions more precisely than text. Furthermore, with preempted tapes, teammates coordinated tasks and invited members to build on each others’ work.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Winder, a Figma plugin that addresses challenges through linked tapes—multimodal comments of clicks and voice that could be as lightweight as face-to-face (F2F) interactions while transmitting intentions more precisely than text.'}",
-chiral benzothiazole monofluoroborate featuring chiroptical and oxygen-sensitizing properties: synthesis and photophysical studies.,Omar Sadek,"Advances in personalized medicine are prompting the development of multimodal agents, that is, molecules that combine properties promoting various diagnostic and therapeutic applications. General approaches exploit chemical conjugation of therapeutic agents with contrast agents or the design of multimodal nanoplatforms. Herein, we report the design of a single molecule that exhibits potential for different diagnostic modes as well as the ability to sensitize oxygen, thus offering potential for photodynamic therapy. Exceptionally, this work involves the synthesis and chiral resolution of an enantiomeric pair of chiral monofluoroborates that contain a stereogenic boron atom. Combining experimental and theoretical chiroptical studies allowed the unambiguous determination of their absolute configuration. Photophysical investigations established the ability of this compound to sensitize oxygen even in the absence of heavy atoms within its structure. The synthesis of a chiral benzothiazole monofluoroborate paves a way to multimodal diagnostic tools (fluorescence and nuclear imaging) while also featuring potential therapeutic applications owing to its ability to activate oxygen to its singlet state for use in photodynamic therapy.","{'model': 'tldr@v2.0.0', 'text': 'The synthesis of a chiral benzothiazole monofluoroborate paves a way to multimodal diagnostic tools (fluorescence and nuclear imaging) while also featuring potential therapeutic applications owing to its ability to activate oxygen to its singlet state for use in photodynamic therapy.'}",https://hal.archives-ouvertes.fr/hal-03333839/file/Sadek%20et%20al%20-%202021%20-%20Chiral%20Benzothiazole%20Monofluoroborate%20Featuring%20Chiroptical%20-%20V2_revised_manuscript2_2nd_revision.pdf
-advances in the understanding and management of chronic pain in multiple sclerosis: a comprehensive review,Ivan Urits,,"{'model': 'tldr@v2.0.0', 'text': 'A comprehensive update of the current perspective of MS pathophysiology, symptomatology, and treatment is provided and medicinal cannabis therapy, targeted physical therapy, and neuromodulation offer promising results are provided.'}",
-the influence of multimodal 3d visualizations on learning acquisition,Phuong T. Do,,"{'model': 'tldr@v2.0.0', 'text': 'The process of simultaneously exposing learners to interactive dynamic visualizations and prompting them to attend to information through the pragmatic use of audio cues reduced memory load, and in turn facilitated memory recall.'}",
-prospects and challenges of imaging neuroinflammation beyond tspo in alzheimer’s disease,D. Boche,,"{'model': 'tldr@v2.0.0', 'text': ""The present review describes the recent advancements in microglia and astrocyte biology in the context of health, ageing, and Alzheimer's disease, the most common dementia worldwide, and presents novel methodologies currently explored to improve in vivo knowledge of the neuroinflammatory patterns in Alzheimer’s disease.""}",https://link.springer.com/content/pdf/10.1007/s00259-019-04462-w.pdf
-making a case for spatial prompting in human-robot communication,A. Green,"In this paper we present an analysis of a set of examples of how verbal and non-verbal behavior of a service robot influence users’ way of positioning themselves during interaction, using concepts from theories of non-verbal behavior. Based on the analysis we propose a design case where a robot utilizes a (naïve) spatial prompting strategy to influence the spatial positioning and communicative behavior of the user. INTRODUCTION A design requirement of a personal service robot is that it should be configured and provided with work tasks by the user in an interactive and intuitive way. These robots are intended to provide service tasks in the home, possibly offering wide range of services. Typically they are envisioned to be equipped with multimodal spoken dialogue systems, to reduce the complexity in the user interface. In this paper we argue that theories of spatial positioning need to be considered when developing the communicative system of the robot. Furthermore we present an empirical account of the way spatial behavior of robots influence human users. We also propose the term spatial prompting, which refers to active strategies of the robot that are intended to influence users to position themselves in a way that is advantageous for further communicative actions. Positioning, as it has been approached as a research challenge for human-robot interaction, is considered as providing adaptive physical movements of the robot. A result of this is that the communicative dimension of positioning typically has been ignored in systems that interactively position themselves in relation to their users. One requirement that is typically put forward is that the robot should position itself in a socially appropriate manner [1, 4]. The parameters that concern these approaches are typically derived from research on nonverbal behavior. In robotics the problem of maintaining the robot localized and situated within a geometric representation of the world has been framed as the Simultaneous Localization and Mapping (SLAM) problem [13]. Recent advances in Human-Robot Interaction (HRI) have raised the interest in detecting and tracking the position of users during interaction. When the position of the user is known, the robot can plan how to position itself [1]. The research on spatial reasoning applied to robotics is well advanced but primarily focused on natural language understanding of spatial relations, providing for exchanges concerning locations of objects in the environment [3]. RESEARCH ON SPATIALITY IN COMMUNICATION There are several research approaches for human-human that are relevant for spatial management between humans and robots. Hall [14] studied interpersonal distances and distinguished four different distances: intimate (0-1.5 ft), personal (1.5-4 ft), social 4-12 ft, and public (> 12 ft). These distances vary both with respect to the current activity and cultural factors. Another dimension that is relevant to spatiality is the concept of territoriality, according to Sack, i.e., “the attempt by an individual or group to affect, influence, or control people, phenomena, and relationships, by delimiting and asserting control over a geographic area” [15]. Kendon [12] also studied the spatial configuration of the participants, using the term F-formations, for instance the L-shape which describes the relation when two participants have a common visual focus. The shared space, the so called o-space, or the transactional space is then located in front of the participants, and it is within this area that the interaction is conducted. Clark [5] refers to this space as the workspace, where perceptual co-presence is established between speakers [5, 10]. In this context, research on perception and especially visual perception plays an important part for maintaining common ground between participants [10, 8]. Gill [9] has investigated the communicative effects that participants achieve by using nonverbal behavior, focusing on the functional rather than the morphological perspective of nonverbal behavior. One such function is the category focus which is a metadiscursive function that signals a shift in the center of attention in the discussion, e.g., a shift in body posture with the same meaning as the utterance “I am going to focus on this spot”. Another, less obvious, but nevertheless important concept is Schegloff’s notion of body torque [15], a state of the bodily configuration when two different body segments are oriented in different directions. According to Schegloff [15] Body torque “project change”, i.e., when some part of the body is organized in an unstable way, the participants may predict that a change in posture is pending. For instance, when turning the head, this might predict a change of the general body orientation. During interaction, speakers monitor the action of others, interpreting purposeful actions that lead towards a common joint goal as compliance [10]. Human-robot interaction is situated in a physical context, where understanding and reference to actions of the human partner during interaction explicitly needs be taken into account. This makes research on virtual collaborative [6] environments interesting also in this context, since it is concerned with models that explicitly represent spatiality and reference. CORPUS ANALYSIS OF SPATIAL MANAGEMENT We have analyzed a video corpus, collected in a European project [7], containing transcribed data of about 20 user sessions, (approximately 20 minutes each) where a user talks to a robot and teaches it the names and locations of objects using a combination of gestures and speech. By viewing the video corpus we identified and analyzed instances where the robot movements or verbal actions appear to influence the actions of the user. The examples reflect three different ways in which the robot actively influences the user to act: • Primary verbal: by using a spoken command • Primary non-verbal: by movements • Multi-modal: using movement as trigger for a verbally specified (or grounded action) 00:32:889-00:34:071 R6: Robot is following 00:40:603-00:41:744 R7: You are too fast 00:44:690-00:46:539 R8: Please stand in front of the camera 00:49:783-00:50:883 R9: Robot is following User","{'model': 'tldr@v2.0.0', 'text': 'An analysis of a set of examples of how verbal and non-verbal behavior of a service robot influence users’ way of positioning themselves during interaction, using concepts from theories of non- verbal behavior is presented.'}",
-an iterative data collection approach for multimodal dialogue systems,S. Rapp,"This paper deals with the way in which data for multimodal dialogue systems are collected. We argue that for multimodal data, an iterative data collection strategy should be followed. Instead of a single major data collection effort using a “Wizard of OZ” (WOZ) or “prompting” experimental setup, several smaller data collections should accompany the system development. We also describe the “script” experimental setup we developed. It is in between the WOZ and prompting setup, and can be used as a cost effective design for the first data collection within the iterative data collection strategy.","{'model': 'tldr@v2.0.0', 'text': 'It is argued that for multimodal data, an iterative data collection strategy should be followed and instead of a single major data collection effort using a “Wizard of OZ” (WOZ) or “prompting” experimental setup, several smaller data collections should accompany the system development.'}",
-towards the concept of disease-modifier in post-stroke or vascular cognitive impairment: a consensus report,R. Bordet,,"{'model': 'tldr@v2.0.0', 'text': 'A consensus emerged to propose the development of a disease-modifying strategy in VCI and PSCI based on pleiotropic pharmacological and non-pharmacological approaches.'}",https://bmcmedicine.biomedcentral.com/track/pdf/10.1186/s12916-017-0869-6
-the smartweb corpora: multimodal access to the web in natural environments,F. Schiel,,"{'model': 'tldr@v2.0.0', 'text': 'This chapter describes the prompting scheme SitPro, the recording technique used to obtain real situation data as well as the properties of the resulting corpora.'}",
-guiding the selection of child spellchecker suggestions using audio and visual cues,Brody Downs,"Spellchecking functionality embedded in existing search tools can assist children by offering a list of spelling alternatives when a spelling error is detected. Unfortunately, children tend to generally select the first alternative when presented with a list of options, as opposed to the one that matches their intent. In this paper, we describe a study we conducted with 191 children ages 6-12 in order to offer empirical evidence of: (1) their selection habits when identifying spelling suggestions that match the word they meant to type, and (2) the degree of influence multimodal cues, i.e., synthesized speech and images, have in prompting children to select the correct spelling suggestion. The results from our study reveal that multimodal cues, primarily synthesized speech, have a positive impact on the children's ability to identify their intended word from a list of spelling suggestions.","{'model': 'tldr@v2.0.0', 'text': ""The results from the study reveal that multimodal cues, primarily synthesized speech, have a positive impact on the children's ability to identify their intended word from a list of spelling suggestions.""}",
-treatment patterns and outcomes of women with breast cancer and supraclavicular nodal metastases,N. Tamirisa,,"{'model': 'tldr@v2.0.0', 'text': 'Women with cN3c breast cancer who received multimodal therapy demonstrated improved overall survival when compared with patients undergoing non-standard therapy and those with metastatic (M1) disease.'}",
-a recalibration of theatre’s hypermediality,M. Crossley,,,https://link.springer.com/content/pdf/10.1007%2F978-3-030-49679-1_2.pdf
-coming to know more through and from writing,V. Prain,"Over the past 20 years, claims about how and why student writing can serve learning have changed markedly. This has been partly due to new technologies displacing writing as a predominant resource for learning, prompting new sense-making practices and shifts in how these changes are theorized. Learners now routinely collaborate to generate, manipulate, analyze, and share images in many subject areas, where multimodal and multimedia resources are expected to motivate learners, enact new learning processes and outcomes, and display this cross-modal learning. These new practices have prompted revisions to how writing is understood and used as a tool for learning in an increasingly multimodal, highly digitized world. In reviewing this literature, we claim that there are strong evidence-based reasons for viewing writing as a central but not sole resource for learning. Our case draws on both past and current research on writing as an epistemological tool. In presenting this case, we draw primarily on our professional background in science education research, acknowledging its distinctive take on the use of writing for learning. However, we think our general case also holds for other disciplinary areas.",,
-adaptive action prompting: a complementary aid to support task-oriented interaction in explorative user interfaces,Thomas Kuehme,"Adaptive action prompting supports users by suggesting how to continue with the interaction. A small number of continuously updated action prompts are offered in addition to the regular menu interface of an application. Users can use these prompts either occasionally in trouble situations or for a sequence of system-guided action selections. The suggestions are based on models of the application, context, and user. According prompting strategies are automatically evaluated and can optionally be controlled by the user. Multimodal user interfaces provide further perspectives for adaptive prompting.","{'model': 'tldr@v2.0.0', 'text': 'A small number of continuously updated action prompts are offered in addition to the regular menu interface of an application, offering suggestions based on models of the application, context, and user.'}",
-acute posterior multifocal placoid pigment epitheliopathy complicated by fatal cerebral vasculitis.,R. Maamari,"A 21-year-old man experienced unilateral vision loss associated with multiple atrophic chorioretinal lesions. He was treated for a presumptive diagnosis of acute retinal necrosis, but his vision did not improve with antiviral therapy. Over the course of several weeks, his symptoms progressed to involve both eyes. The fellow eye showed characteristic yellow-white placoid lesions, prompting treatment with oral corticosteroids for acute posterior multifocal placoid pigment epitheliopathy (APMPPE). Despite high-dose therapy with prednisone 80 mg daily, the patient developed the acute onset of mental status changes within the next several days. Neuroimaging revealed multifocal large-vessel strokes associated with cerebral edema; these infarcts led to herniation and death. Postmortem histopathologic examination confirmed granulomatous inflammation in both ocular and cerebral vasculatures. Together with findings from multimodal imaging obtained throughout this patient's clinical course, our findings support the notion that granulomatous choroiditis is the mechanism of the ocular lesions seen in APMPPE. This granulomatous inflammation can also affect cerebral vessels, leading to strokes.","{'model': 'tldr@v2.0.0', 'text': 'Findings support the notion that granulomatous choroiditis is the mechanism of the ocular lesions seen in APMPPE, which can also affect cerebral vessels, leading to strokes.'}",
-adaptive prompting,T. Kühme,"Adaptive prompting addresses problems with locating, referencing, and selecting interface items such as elements of directory hierarchies or menu and dialog box entries. By drawing the user’s attention to the most appropriate and most likely to be chosen items, adaptive prompting tries to increase users’ performance on selections and to provide guidance in explorative environments. Examples for adaptive prompting include an Adaptive Tool Prompter, an Adaptive Action Prompter, and Adaptive Prompting in Dialog Boxes. In order to deal with application contexts and evolving needs and preferences of the user, adaptive prompting employs models of the application and the user. The chosen approach allows for an optional user involvement into the adaptation and for an evaluation of the embedded prompting strategies. Multimodal user interfaces provide further perspectives for adaptive prompting.","{'model': 'tldr@v2.0.0', 'text': 'Adaptive prompting addresses problems with locating, referencing, and selecting interface items such as elements of directory hierarchies or menu and dialog box entries to increase users’ performance on selections and to provide guidance in explorative environments.'}",
-"energizing project-based inquiry: middle-grade students read, write, and create videos.",Hiller A. Spires,"In light of emerging technologies prompting new avenues for teaching and learning, students are positioned to “create” to learn, with video production being an important process for literacy development. There is a growing need for innovative instructional practices in reading and writing that are aligned with student interests and the activities they engage in outside of the classroom. Connecting video production to school reading and writing experiences taps into a student's natural predisposition for media consumption and production. This article presents a project-based inquiry process that merges students’ growing interest in grassroots video with educational goals that are aligned with state and national curricular standards. Through a five-phase process—ask a compelling question; gather and analyze information; creatively synthesize information; critically evaluate and revise; and publish, share, and act—students are engaged in reading, writing, and creating a video product that captures their multimodal learning of academic content.",,
-using subpopulation eas to map molecular structure landscapes,Ahmed Bin Zaman,"The emerging view in molecular biology is that molecules are intrinsically dynamic systems rearranging themselves into different structures to interact with molecules in the cell. Such rearrangements take place on energy landscapes that are vast and multimodal, with minima housing alternative structures. The multiplicity of biologically-active structures is prompting researchers to expand their treatment of classic computational biology problems, such as the template-free protein structure prediction problem (PSP), beyond the quest for the global optimum. In this paper, we revisit subpopulation-oriented EAs as vehicles to switch the objective from classic optimization to landscape mapping. Specifically, we present two EAs, one of which makes use of subpopulation competition to allocate more computational resources to fitter subpopulations, and another of which additionally utilizes a niche preservation technique to maintain stable and diverse subpopulations. Initial assessment on benchmark optimization problems confirms that stabler subpopulations are achieved by the niche-preserving EA. Evaluation on unknown energy landscapes in the context of PSP demonstrates superior mapping performance by both algorithms over a popular Monte Carlo-based method, with the niche-preserving EA achieving superior exploration of lower-energy regions. These results suggest that subpopulation EAs hold much promise for solving important mapping problems in computational structural biology.","{'model': 'tldr@v2.0.0', 'text': 'This paper revisits subpopulation-oriented EAs as vehicles to switch the objective from classic optimization to landscape mapping and presents two EAs, one of which makes use of subpopulation competition to allocate more computational resources to fitter subpopulations, and another of which additionally utilizes a niche preservation technique to maintain stable and diverse subpopulation.'}",https://dl.acm.org/doi/pdf/10.1145/3321707.3321777
-effectiveness of different recruitment strategies in an rct of a surgical device: experience from the endobarrier trial,A. Ruban,"Recruiting participants into clinical trials is notoriously difficult and poses the greatest challenge when planning any investigative study. Poor recruitment may not only have financial ramifications owing to increased time and resources being spent but could adversely influence the clinical impact of a study if it becomes underpowered. Herein, we present our own experience of recruiting into a nationally funded, multicentre, randomised controlled trial (RCT) of the Endobarrier versus standard medical therapy in obese patients with type 2diabetes. Despite these both being highly prevalent conditions, there were considerable barriers to the effectiveness of different recruitment strategies across each study site. Although recruitment from primary care proved extremely successful at one study site, this largely failed at another site prompting the implementation of multimodal recruitment strategies including a successful media campaign to ensure sufficient participants were enrolled and the study was adequately powered. From this experience, we propose where appropriate the early engagement and investment in media campaigns to enhance recruitment into clinical trials. Trial Registration: ISRCTN30845205.","{'model': 'tldr@v2.0.0', 'text': 'From this experience of recruiting into a nationally funded, multicentre, randomised controlled trial of the Endobarrier versus standard medical therapy in obese patients with type 2diabetes, it is proposed where appropriate the early engagement and investment in media campaigns to enhance recruitment into clinical trials are proposed.'}",https://bmjopen.bmj.com/content/bmjopen/9/11/e032439.full.pdf
-deferiprone as adjunctive treatment for patients with invasive mucormycosis: a retrospective case series,M. Chitasombat,"Mucormycosis is a life-threatening disease requiring multimodal treatment with antifungals and surgery. The mortality rate remains high, prompting consideration of alternative treatment strategies. Deferiprone has in vitro activity against Mucorales, but its efficacy has never been evaluated in humans. Here, we retrospectively analyzed patients with confirmed mucormycosis who received deferiprone from 2011 to 2017. Five patients had hematologic malignancies and one was diabetic. The sites of infection included sinus-orbit-cerebral (67%), lung (17%), and disseminated infection (17%). Surgery was performed in 83% of cases and achieved local control for 33% of patients. A combination regimen of polyenes plus echinocandins was administered with stepdown treatment using posaconazole. The median duration of antifungal treatment was 86 days (range: 46-435 days) days. Deferiprone was given as adjunctive treatment with a median dose and duration of 100 mg/kd/day (range: 86.2-100 mg/kg/day) and 25 days (range: 15-215 days), respectively. Overall, deferiprone was well-tolerated. Successful outcomes were observed at 12-week follow-up for 67% of patients. The mortality rate at 180- day follow-up was 50%. Adjunctive therapy with deferiprone showed safety and tolerability.","{'model': 'tldr@v2.0.0', 'text': 'Adjunctive therapy with deferiprone showed safety and tolerability and successful outcomes were observed at 12-week follow-up for 67% of patients, and the mortality rate at 180- day follow- up was 50%.'}",https://www.pagepress.org/journals/index.php/idr/article/download/7765/7453
-pain management in cancer patients using a mobile app: study design of a randomized controlled trial,S. Agboola,"Background Despite the availability of effective medications and clinical guidelines for pain management, pain control is suboptimal in a sizeable proportion of patients with cancer pain. The National Comprehensive Cancer Network guidelines recommend a comprehensive and multimodal approach for management of cancer pain. We developed a mobile phone application, ePAL, based on clinical guidelines to empower patients for cancer pain management by prompting regular pain assessments and coaching for self-management. Objective The objective of this study is to evaluate the effect of a multidimensional mobile phone-based pain management application, ePAL, on controlling cancer pain and improving quality of life in patients with cancer pain being treated at an academic palliative care clinic. Methods The study will be implemented as a 2-arm randomized controlled trial with 110 adult patients with CP who own a mobile phone over a follow-up period of two months. Participants will be randomized to either the intervention group receiving ePAL and usual care or to a control group receiving only usual care. The brief pain inventory will be used to assess our primary outcome which is pain intensity. We will also evaluate the effect of the intervention on secondary outcomes which include the effect of the intervention on hospital utilization for pain crisis, quality of life, adherence to analgesic medications, barriers to pain control, anxiety and patient engagement. Instruments that will be used in evaluating secondary outcomes include the Brief Pain Inventory, Morisky Medication Adherence Scale, Barriers Questionnaire-II, Functional Assessment of Cancer Therapy–General, Edmonton Symptom Assessment System, Generalized Anxiety Disorder 7-item scale, and the Functional Assessment of Chronic Illness Therapy-Fatigue. The intention-to-treat approach will be used to evaluate outcomes. Our primary outcome, pain intensity, measured longitudinally over eight weeks, will be assessed by mixed model repeated analysis. Effect sizes will be calculated as mean group differences with standard deviations. Results The study is still in progress. We hope to have results by the end of 2015. Conclusions The multidimensional approach to pain management implemented on a mobile phone application could lead to significant improvements in patient outcomes. Trial Registration ClinicalTrials.gov NCT02069743; https://clinicaltrials.gov/ct2/show/NCT02069743 (Archived by WebCite at http://www.webcitation.org/6Qb65XGGA).","{'model': 'tldr@v2.0.0', 'text': 'The multidimensional approach to pain management implemented on a mobile phone application could lead to significant improvements in patient outcomes.'}",https://www.researchprotocols.org/2014/4/e76/PDF
-deoxyhypusine synthase (dhps) inhibitor gc7 induces p21/rb-mediated inhibition of tumor cell growth and dhps expression correlates with poor prognosis in neuroblastoma patients,A. Bandino,,"{'model': 'tldr@v2.0.0', 'text': 'Results suggest that spermidine and DHPS are key contributing factors in NB tumor proliferation through regulation of the p21/Rb signaling axis.'}",
-using formulations and gaze to encourage parents to talk about their and their children's health and well-being,S. Tiitinen,"In preventive health-care settings, professionals need to encourage clients to talk about their problems before they become critical. We use multimodal conversation analysis to demonstrate how public health nurses encourage parents to elaborate on their problems in a sample of preventive maternity and child health (MCH) clinics in Finland. The nurses topicalize the problem-relevant aspects of the parents’ problem-indicative talk by issuing a formulation of what the parent has just said (that is, by redescribing it in problem-related terms). This verbal practice is synchronized with a visual one—the nurse issues the formulation, receives the parent's response, and then gazes directly at them. This has the effect of prompting the parent to take up the problem and talk about it. We discuss the findings in relation to the institutional tasks in MCH care and to the role of gaze in constituting actions, such as formulations. Data are in Finnish with English translation.","{'model': 'tldr@v2.0.0', 'text': 'Multiodal conversation analysis is used to demonstrate how public health nurses encourage parents to elaborate on their problems in a sample of preventive maternity and child health clinics in Finland, and the role of gaze in constituting actions, such as formulations.'}",https://trepo.tuni.fi/bitstream/10024/104965/1/using_formulations_and_gaze_2014.pdf
-unilateral diffuse uveal melanocytic proliferation,R. Spaide,"Purpose: To report a unilateral case of what is named bilateral diffuse uveal melanocytic proliferation and consider the consequences of this finding. Methods: The ocular findings were investigated with multimodal imaging to include color fundus photography, fluorescein angiography, autofluorescence imaging, and enhanced depth imaging optical coherence tomography. Results: A 66-year-old woman had a history of breast cancer 23 years previously that was treated and the patient was free of disease since. She developed a recent decrease in visual acuity in her left eye prompting referral. She was seen to have an alteration in the pigmentation of the posterior pole of the left eye with dispersed red placoid spots. Autofluorescence imaging showed nummular areas of absent autofluorescence signal, which corresponded to areas of hyperfluorescence during fluorescein angiography. The placoid spots were hyperautofluorescent and hypofluorescent in autofluorescence and fluorescein angiography, respectively. She had diffuse thickening and infiltration of the choroid in the left eye. Because of the ocular findings, the patient underwent a systemic evaluation and was found to have widely metastatic disease with an unknown primary cancer. No progression of disease was seen in the left eye over a 6-month follow-up, and the right eye never showed any abnormality, except for a modest cataract, by any means of examination. Conclusion: Bilateral uveal melanocytic proliferation has been attributed to a paraneoplastic process, allegedly from a factor in the IgG fraction of the serum. However, unilateral involvement suggests that there are other factors involved in disease manifestation.","{'model': 'tldr@v2.0.0', 'text': 'Bilateral uveal melanocytic proliferation has been attributed to a paraneoplastic process, allegedly from a factor in the IgG fraction of the serum, but unilateral involvement suggests that there are other factors involved in disease manifestation.'}",
-liver transplantation for hilar cholangiocarcinoma—a single-centre experience,S. Schüle,,"{'model': 'tldr@v2.0.0', 'text': 'Acceptable survival rates can be achieved by transplantation for hilar cholangiocarcinoma with lymph node metastases as the only exclusion criterion.'}",
-subjective cognitive complaints and objective memory performance influence prompt preference for instrumental activities of daily living.,Emily J Van Etten,"INTRODUCTION Declines in memory and executive functioning often lead to difficulties completing instrumental activities of daily living (IADLs). Prompting technologies have the potential to help promote aging in place by providing support for the initiation and accurate completion of IADLs. In this study, we evaluate preferences of older adults for different levels of prompting support based on subjective and objective measures of cognitive functioning. METHOD Participants were 170 community-dwelling older adults split into two cognitive complaint groups: cognitive complaints and few cognitive complaints. After completing six IADL tasks (e.g., organize a pillbox, cook), each participant was asked to make a specific error (e.g., leave stove on) on three of the tasks. They were then prompted to correct the error with one of three different prompt modes: verbal indirect, verbal direct, multimodal verbal direct and video. RESULTS The cognitive complaints group reported greater preference for the multimodal prompt compared to the few cognitive complaints group. The indirect prompt was the least preferred by both groups. Furthermore, participants who recalled less on objective memory measures preferred more support in terms of prompt mode. Executive functioning did not appear to be related to prompt preference. CONCLUSION Level of subjective cognitive complaints and objective memory performance may influence amount of support preferred in a prompt.","{'model': 'tldr@v2.0.0', 'text': 'Level of subjective cognitive complaints and objective memory performance may influence amount of support preferred in a prompt.'}",https://europepmc.org/articles/pmc5597053?pdf=render
-"type a behavior pattern: new perspectives on theory, assessment, and intervention.",C. Thoresen,"Research on the Type A behavior pattern (TA) has been plagued by inadequate theory, insensitive assessment, and insufficient interventions. These problems (e.g., using global dichotomous ratings of TA) have contributed to several failures of TA to predict cardiovascular outcomes prompting concern for new approaches. Conceptual models are discussed, such as a transactional model of TA, social cognitive theories (e.g., self-evaluative processes), and associative network theory (e.g., how emotions distort information). The view of hostility as the only pathogenic feature of TA is questioned. The ethnographic gap in TA research is described, and detailed descriptive and experimental case studies are recommended. Profile measures that assess all dimensions of TA are urged along with contextually sensitive multimodal assessments. Unlike results of some correlational studies, controlled TA interventions, although few, have consistently yielded positive results. Study of TA within an expanded conceptual perspective is encouraged, especially controlled experiments and interventions.","{'model': 'tldr@v2.0.0', 'text': 'The ethnographic gap in TA research is described, and detailed descriptive and experimental case studies are recommended, and the view of hostility as the only pathogenic feature of TA is questioned.'}",
-literacy and digital culture in the early years,K. Dooley,"In this chapter we present data drawn from observations of kindergarten children using iPads and talk with the children, their parents/guardians and teachers. We identify a continuum of practices that extends from ‘educational apps’ teaching handwriting, sight words and so forth to uses of the iPad as a device for multimodal literacy development and substantive conversation around children’s creative work. At the current time high stakes testing and the implementation of the Australian Curriculum are prompting new public and professional conversations about literacy and digital technology. The iPad is construed as both cause of and solution to problems of traditional literacy education. In this context we describe the literacies enabled by educational software available on iPads. We higlight the time constraints which bore on teachers' capacity to enact their visions of literacy education through the iPad platform and suggest ways of reflecting on responses to this constraint.","{'model': 'tldr@v2.0.0', 'text': 'A continuum of practices that extends from ‘educational apps’ teaching handwriting, sight words and so forth to uses of the iPad as a device for multimodal literacy development and substantive conversation around children’s creative work is identified.'}",
-cm3: a causal masked multimodal model of the internet,Armen Aghajanyan,"We introduce CM3, a family of causally masked generative models trained over a large corpus of structured multi-modal documents that can contain both text and image tokens. Our new causally masked approach generates tokens left to right while also masking out a small number of long token spans that are generated at the end of the string, instead of their original positions. The casual masking object provides a type of hybrid of the more common causal and masked language models, by enabling full generative modeling while also providing bidirectional context when generating the masked spans. We train causally masked language-image models on large-scale web and Wikipedia articles, where each document contains all of the text, hypertext markup, hyperlinks, and image tokens (from a VQVAE-GAN), provided in the order they appear in the original HTML source (before masking). The resulting CM3 models can generate rich structured, multi-modal outputs while conditioning on arbitrary masked document contexts, and thereby implicitly learn a wide range of text, image, and cross modal tasks. They can be prompted to recover, in a zero-shot fashion, the functionality of models such as DALL-E, GENRE, and HTLM. We set the new state-of-the-art in zero-shot summarization, entity linking, and entity disambiguation while maintaining competitive performance in the fine-tuning setting. We can generate images unconditionally, conditioned on text (like DALL-E) and do captioning all in a zero-shot setting with a single model.","{'model': 'tldr@v2.0.0', 'text': 'CM3, a family of causally masked generative models trained over a large corpus of structured multi-modal documents that can contain both text and image tokens, is introduced and set the new state-of-the-art in zero-shot summarization, entity linking, and entity disambiguation while maintaining competitive performance in the fine-tuning setting.'}",
-show me what and tell me how: video synthesis via multimodal conditioning,Ligong Han,"Most methods for conditional video synthesis use a single modality as the condition. This comes with major limitations. For example, it is problematic for a model conditioned on an image to generate a specific motion trajectory desired by the user since there is no means to provide motion information. Conversely, language information can describe the desired motion, while not precisely defining the content of the video. This work presents a multimodal video generation framework that benefits from text and images provided jointly or separately. We leverage the recent progress in quantized representations for videos and apply a bidirectional transformer with multiple modalities as inputs to predict a discrete video representation. To improve video quality and consistency, we propose a new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens. We introduce text augmentation to improve the robustness of the textual representation and diversity of generated videos. Our framework can incorporate various visual modalities, such as segmentation masks, drawings, and partially occluded images. It can generate much longer sequences than the one used for training. In addition, our model can extract visual information as suggested by the text prompt, e.g., “an object in image one is moving northeast”, and generate corresponding videos. We run evaluations on three public datasets and a newly collected dataset labeled with facial attributes, achieving state-of-the-art generation results on all four11Code: https://github.com/snap-research/MMVID and Webpage..","{'model': 'tldr@v2.0.0', 'text': 'To improve video quality and consistency, a new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens are proposed and text augmentation is introduced to improve the robustness of the textual representation and diversity of generated videos.'}",https://arxiv.org/pdf/2203.02573
-prompt tuning for generative multimodal pretrained models,Han Yang,"Prompt tuning has become a new paradigm for model tuning and it has demonstrated success in natural language pretraining and even vision pretraining. In this work, we explore the transfer of prompt tuning to multimodal pretraining, with a focus on generative multimodal pretrained models, instead of contrastive ones. Specifically, we implement prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks. Experimental results demonstrate that the light-weight prompt tuning can achieve comparable performance with finetuning and surpass other light-weight tuning methods. Besides, in comparison with finetuned models, the prompt-tuned models demonstrate improved robustness against adversarial attacks. We further figure out that experimental factors, including the prompt length, prompt depth, and reparameteratization, have great impacts on the model performance, and thus we empirically provide a recommendation for the setups of prompt tuning. Despite the observed advantages, we still find some limitations in prompt tuning, and we correspondingly point out the directions for future studies. Codes are available at \url{https://github.com/OFA-Sys/OFA}","{'model': 'tldr@v2.0.0', 'text': 'This work implements prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks, with a focus on generative multimodal pretrained models, instead of contrastive ones.'}",http://arxiv.org/pdf/2208.02532
-opal: multimodal image generation for news illustration,Vivian Liu,"Advances in multimodal AI have presented people with powerful ways to create images from text. Recent work has shown that text-to-image generations are able to represent a broad range of subjects and artistic styles. However, finding the right visual language for text prompts is difficult. In this paper, we address this challenge with Opal, a system that produces text-to-image generations for news illustration. Given an article, Opal guides users through a structured search for visual concepts and provides a pipeline allowing users to generate illustrations based on an article’s tone, keywords, and related artistic styles. Our evaluation shows that Opal efficiently generates diverse sets of news illustrations, visual assets, and concept ideas. Users with Opal generated two times more usable results than users without. We discuss how structured exploration can help users better understand the capabilities of human AI co-creative systems.","{'model': 'tldr@v2.0.0', 'text': 'How structured exploration can help users better understand the capabilities of human AI co-creative systems is discussed, and Opal, a system that produces text-to-image generations for news illustration, is addressed.'}",https://arxiv.org/pdf/2204.09007
-"joint engagement in the home environment is frequent, multimodal, timely, and structured.",Catalina Suarez-Rivera,"Infants develop in a social context, surrounded by knowledgeable caregivers who scaffold learning through shared engagement with objects. However, researchers have typically examined joint engagement in structured tasks, where caregivers sit near infants and display frequent, prompt, and multimodal behaviors around the objects of infant action. Which features of joint engagement generalize to the real-world? Despite the importance of joint engagement for infant learning, critical assumptions around joint engagement in everyday interaction remain unexamined. We investigated behavioral and temporal features of joint engagement in the home environment, where objects for play abound and dyad proximity fluctuates. Infant manual actions, mother manual and verbal behaviors, and dyad proximity were coded frame-by-frame from 2-h naturalistic recordings of 13- to 23-month-old infants and their mothers (N = 38). Infants experienced rich, highly structured, multimodal mother input around the objects of their actions. Specifically, joint engagement occurred within seconds of infant action and was amplified in the context of interpersonal proximity. Findings validate laboratory-based research on characteristics of joint engagement while highlighting unique properties around the role of mother-infant proximity and temporal structuring of caregiver input over extended time frames. Implications for the social contexts that support infant learning and development are discussed.","{'model': 'tldr@v2.0.0', 'text': 'Joint engagement occurred within seconds of infant action and was amplified in the context of interpersonal proximity, highlighting unique properties around the role of mother-infant proximity and temporal structuring of caregiver input over extended time frames.'}",https://discovery.ucl.ac.uk/10142739/1/Suarez-Rivera%20et%20al.%2C%202022-%20JE%20at%20home_Infancy.pdf
-draw your art dream: diverse digital art synthesis with multimodal guided diffusion,Nisha Huang,"Digital art synthesis is receiving increasing attention in the multimedia community because of engaging the public with art effectively. Current digital art synthesis methods usually use single-modality inputs as guidance, thereby limiting the expressiveness of the model and the diversity of generated results. To solve this problem, we propose the multimodal guided artwork diffusion (MGAD) model, which is a diffusion-based digital artwork generation approach that utilizes multimodal prompts as guidance to control the classifier-free diffusion model. Additionally, the contrastive language-image pretraining (CLIP) model is used to unify text and image modalities. Extensive experimental results on the quality and quantity of the generated digital art paintings confirm the effectiveness of the combination of the diffusion model and multimodal guidance. Code is available at https://github.com/haha-lisa/MGAD-multimodal-guided-artwork-diffusion.",,https://dl.acm.org/doi/pdf/10.1145/3503161.3548282
-contrastive language-vision ai models pretrained on web-scraped multimodal data exhibit sexual objectification bias,R. Wolfe,"Warning: The content of this paper may be upsetting or triggering. Nine language-vision AI models trained on web scrapes with the Contrastive Language-Image Pretraining (CLIP) objective are evaluated for evidence of a bias studied by psychologists: the sexual objectification of girls and women, which occurs when a person’s human characteristics, such as emotions, are disregarded and the person is treated as a body or a collection of body parts. We replicate three experiments in the psychology literature quantifying sexual objectification and show that the phenomena persist in trained AI models. A first experiment uses standardized images of women from the Sexual OBjectification and EMotion Database, and finds that human characteristics are disassociated from images of objectified women: the model’s recognition of emotional state is mediated by whether the subject is fully or partially clothed. Embedding association tests (EATs) return significant effect sizes for both anger (d > 0.80) and sadness (d > 0.50), associating images of fully clothed subjects with emotions. GRAD-CAM saliency maps highlight that CLIP gets distracted from emotional expressions in objectified images where subjects are partially clothed. A second experiment measures the effect in a representative application: an automatic image captioner (Antarctic Captions) includes words denoting emotion less than 50% as often for images of partially clothed women than for images of fully clothed women. A third experiment finds that images of female professionals (scientists, doctors, executives) are likely to be associated with sexual descriptions relative to images of male professionals. A fourth experiment shows that a prompt of ""a [age] year old girl"" generates sexualized images (as determined by an NSFW classifier) up to 73% of the time for VQGAN-CLIP (age 17), and up to 42% of the time for Stable Diffusion (ages 14 and 18); the corresponding rate for boys never surpasses 9%. The evidence indicates that language-vision AI models trained on automatically collected web scrapes learn biases of sexual objectification, which propagate to downstream applications.",,https://dl.acm.org/doi/pdf/10.1145/3593013.3594072
-bi-bimodal modality fusion for correlation-controlled multimodal sentiment analysis,Wei Han,"Multimodal sentiment analysis aims to extract and integrate semantic information collected from multiple modalities to recognize the expressed emotions and sentiment in multimodal data. This research area’s major concern lies in developing an extraordinary fusion scheme that can extract and integrate key information from various modalities. However, previous work is restricted by the lack of leveraging dynamics of independence and correlation between modalities to reach top performance. To mitigate this, we propose the Bi-Bimodal Fusion Network (BBFN), a novel end-to-end network that performs fusion (relevance increment) and separation (difference increment) on pairwise modality representations. The two parts are trained simultaneously such that the combat between them is simulated. The model takes two bimodal pairs as input due to the known information imbalance among modalities. In addition, we leverage a gated control mechanism in the Transformer architecture to further improve the final output. Experimental results on three datasets (CMU-MOSI, CMU-MOSEI, and UR-FUNNY) verifies that our model significantly outperforms the SOTA. The implementation of this work is available at https://github.com/declare-lab/multimodal-deep-learning and https://github.com/declare-lab/BBFN.","{'model': 'tldr@v2.0.0', 'text': 'The Bi-Bimodal Fusion Network (BBFN), a novel end-to-end network that performs fusion (relevance increment) and separation (difference increment) on pairwise modality representations that significantly outperforms the SOTA.'}",https://arxiv.org/pdf/2107.13669
-promptmner: prompt-based entity-related visual clue extraction and integration for multimodal named entity recognition,Xuwu Wang,,"{'model': 'tldr@v2.0.0', 'text': 'This work proposes to utilize entity-related prompts for extracting proper visual clues from the image to better integrate different modalities and address the popular semantic gap problem, and proposes a modality-aware attention mechanism for better cross-modal fusion.'}",
-multimodal data guided spatial feature fusion and grouping strategy for e-commerce commodity demand forecasting,Weiwei Cai,"E-commerce offers various merchandise for selling and purchasing with frequent transactions and commodity flows. An accurate prediction of customer needs and optimized allocation of goods is required for cost reduction. ,e existing solutions have significant errors and are unsuitable for addressing warehouse needs and allocation. ,at is why businesses cannot respond to customer demands promptly, as they need accurate and reliable demand forecasting.,erefore, this paper proposes spatial feature fusion and grouping strategies based onmultimodal data and builds a neural network predictionmodel for e-commodity demand. ,e designed model extracts order sequence features, consumer emotional features, and facial value features from multimodal data from e-commerce products. ,en, a bidirectional long short-term memory network(BiLSTM-) based grouping strategy is proposed. ,e proposed strategy fully learns the contextual semantics of time series data while reducing the influence of other features on the group’s local features. ,e output features of multimodal data are highly spatially correlated, and this paper employs the spatial dimension fusion strategy for feature fusion. ,is strategy effectively obtains the deep spatial relations among multimodal data by integrating the features of each column in each group across spatial dimensions. Finally, the proposedmodel’s prediction effect is tested using e-commerce dataset. ,e experimental results demonstrate the proposed algorithm’s effectiveness and superiority.","{'model': 'tldr@v2.0.0', 'text': 'A bidirectional long short-term memory network (BiLSTM-) based grouping strategy and spatial feature fusion and grouping strategies based on multimodal data are proposed and a neural network prediction model for e-commodity demand is built.'}",
-unitranser: a unified transformer semantic representation framework for multimodal task-oriented dialog system,Zhiyuan Ma,"As a more natural and intelligent interaction manner, multimodal task-oriented dialog system recently has received great attention and many remarkable progresses have been achieved. Nevertheless, almost all existing studies follow the pipeline to first learn intra-modal features separately and then conduct simple feature concatenation or attention-based feature fusion to generate responses, which hampers them from learning inter-modal interactions and conducting cross-modal feature alignment for generating more intention-aware responses. To address these issues, we propose UniTranSeR, a Unified Transformer Semantic Representation framework with feature alignment and intention reasoning for multimodal dialog systems. Specifically, we first embed the multimodal features into a unified Transformer semantic space to prompt inter-modal interactions, and then devise a feature alignment and intention reasoning (FAIR) layer to perform cross-modal entity alignment and fine-grained key-value reasoning, so as to effectively identify user’s intention for generating more accurate responses. Experimental results verify the effectiveness of UniTranSeR, showing that it significantly outperforms state-of-the-art approaches on the representative MMD dataset.","{'model': 'tldr@v2.0.0', 'text': 'UniTranSeR is proposed, a Unified Transformer Semantic Representation framework with feature alignment and intention reasoning for multimodal dialog systems that significantly outperforms state-of-the-art approaches on the representative MMD dataset.'}",https://aclanthology.org/2022.acl-long.9.pdf
-conceptfusion: open-set multimodal 3d mapping,Krishna Murthy Jatavallabhula,"Building 3D maps of the environment is central to robot navigation, planning, and interaction with objects in a scene. Most existing approaches that integrate semantic concepts with 3D maps largely remain confined to the closed-set setting: they can only reason about a finite set of concepts, pre-defined at training time. Further, these maps can only be queried using class labels, or in recent work, using text prompts. We address both these issues with ConceptFusion, a scene representation that is (1) fundamentally open-set, enabling reasoning beyond a closed set of concepts and (ii) inherently multimodal, enabling a diverse range of possible queries to the 3D map, from language, to images, to audio, to 3D geometry, all working in concert. ConceptFusion leverages the open-set capabilities of today's foundation models pre-trained on internet-scale data to reason about concepts across modalities such as natural language, images, and audio. We demonstrate that pixel-aligned open-set features can be fused into 3D maps via traditional SLAM and multi-view fusion approaches. This enables effective zero-shot spatial reasoning, not needing any additional training or finetuning, and retains long-tailed concepts better than supervised approaches, outperforming them by more than 40% margin on 3D IoU. We extensively evaluate ConceptFusion on a number of real-world datasets, simulated home environments, a real-world tabletop manipulation task, and an autonomous driving platform. We showcase new avenues for blending foundation models with 3D open-set multimodal mapping. For more information, visit our project page https://concept-fusion.github.io or watch our 5-minute explainer video https://www.youtube.com/watch?v=rkXgws8fiDs","{'model': 'tldr@v2.0.0', 'text': 'ConceptFusion enables effective zero-shot spatial reasoning, not needing any additional training or finetuning, and retains long-tailed concepts better than supervised approaches, outperforming them by more than 40% margin on 3D IoU.'}",http://arxiv.org/pdf/2302.07241
-lvp-m3: language-aware visual prompt for multilingual multimodal machine translation,Hongcheng Guo,"Multimodal Machine Translation (MMT) focuses on enhancing text-only translation with visual features, which has attracted considerable attention from both natural language processing and computer vision communities. Recent advances still struggle to train a separate model for each language pair, which is costly and unaffordable when the number of languages increases in the real world. In other words, the multilingual multimodal machine translation (Multilingual MMT) task has not been investigated, which aims to handle the aforementioned issues by providing a shared semantic space for multiple languages. Besides, the image modality has no language boundaries, which is superior to bridging the semantic gap between languages. To this end,we first propose the Multilingual MMT task by establishing two new Multilingual MMT benchmark datasets covering seven languages.Then, an effective baseline LVP-M3 using visual prompts is proposed to support translations between different languages,which includes three stages (token encoding, language-aware visual prompt generation, and language translation). Extensive experimental results on our constructed benchmark datasets demonstrate the effectiveness of LVP-M3 method for Multilingual MMT.","{'model': 'tldr@v2.0.0', 'text': 'An effective baseline LVP-M3 using visual prompts is proposed to support translations between different languages, which includes three stages (token encoding, language-aware visual prompt generation, and language translation).'}",http://arxiv.org/pdf/2210.15461
-multimodal data for the detection of freezing of gait in parkinson’s disease,Wei Zhang,,"{'model': 'tldr@v2.0.0', 'text': 'A protocol has been designed to acquire multimodal physical and physiological information during FOG, including gait acceleration (ACC), electroencephalogram (EEG), electromyogram (EMG), and skin conductance (SC), and multi-modal data have been validated by a FOG detection task.'}",https://www.nature.com/articles/s41597-022-01713-8.pdf
-designing multimodal interactive dashboard of disaster management systems,Abeer AlAbdulaali,"Disasters and crises are inevitable in this world. In the aftermath of a disaster, a society’s overall growth, resources, and economy are greatly affected as they cause damages from minor to huge proportions. Around the world, countries are interested in improving their emergency decision-making. The institutions are paying attention to collecting different types of data related to crisis information from various resources, including social media, to improve their emergency response. Previous efforts have focused on collecting, extracting, and classifying crisis data from text, audio, video, or files; however, the development of user-friendly multimodal disaster data dashboards to support human-to-system interactions during an emergency response has received little attention. Our paper seeks to fill this gap by proposing usable designs of interactive dashboards to present multimodal disaster information. For this purpose, we first investigated social media data and metadata for the required elicitation and analysis purposes. These requirements are then used to develop interactive multimodal dashboards to present complex disaster information in a usable manner. To validate our multimodal dashboard designs, we have conducted a heuristic evaluation. Experts have evaluated the interactive disaster dashboards using a customized set of heuristics. The overall assessment showed positive feedback from the evaluators. The proposed interactive multimodal dashboards complement the existing techniques of collecting textual, image, audio, and video emergency information and their classifications for usable presentation. The contribution will help the emergency response personnel in terms of useful information and observations for prompt responses to avoid significant damage.","{'model': 'tldr@v2.0.0', 'text': 'The proposed interactive multimodal dashboards complement the existing techniques of collecting textual, image, audio, and video emergency information and their classifications for usable presentation and will help the emergency response personnel in terms of useful information and observations for prompt responses to avoid significant damage.'}",https://www.mdpi.com/1424-8220/22/11/4292/pdf?version=1654601242
-m-vader: a model for diffusion with multimodal context,Samuel Weinbach,"We introduce M-VADER: a diffusion model (DM) for image generation where the output can be specified using arbitrary combinations of images and text. We show how M-VADER enables the generation of images specified using combinations of image and text, and combinations of multiple images. Previously, a number of successful DM image generation algorithms have been introduced that make it possible to specify the output image using a text prompt. Inspired by the success of those models, and led by the notion that language was already developed to describe the elements of visual contexts that humans find most important, we introduce an embedding model closely related to a vision-language model. Specifically, we introduce the embedding model S-MAGMA: a 13 billion parameter multimodal decoder combining components from an autoregressive vision-language model MAGMA and biases finetuned for semantic search.","{'model': 'tldr@v2.0.0', 'text': 'The embedding model S-MAGMA is introduced: a 13 billion parameter multimodal decoder combining components from an autoregressive vision-language model MAGMA and biases finetuned for semantic search.'}",https://arxiv.org/pdf/2212.02936
-"assistive multimodal robotic system (amrsys): security and privacy issues, challenges, and possible solutions",Jims Marchang,"Assistive robotic systems could be a suitable solution to support a variety of health and care services, help independent living, and even simulate affection, to reduce loneliness. However, adoption is limited by several issues, as well as user concerns about ethics, data security, and privacy. Other than the common threats related to internet connectivity, personal robotic systems have advanced interaction possibilities, such as audio, video, touch, and gestures, which could be exploited to gain access to private data that are stored in the robot. Therefore, novel, safer methods of interaction should be designed to safeguard users’ privacy. To solicit further research on secure and private multimodal interaction, this article presents a thorough study of the state-of-the-art literature on data security and user privacy in interactive social robotic systems for health and care. In our study, we focus on social robotics to assist older people, which is a global challenge that is receiving a great deal of attention from the robotics and social care communities. This application will have a significant positive impact on the economy and society, but poses various security and privacy issues. This article analyses the key vulnerable areas where data leakage could occur during a multimodal interaction with a personal assistive robotic system. Thus, blockchain with a resource-aware framework, along with a continuous multifactor authentication mechanism, are envisaged as a potential solution for making such systems secure by design; therefore, increasing trust, acceptability, and adoption. Among the key cybersecurity research challenges, it is crucial to create an intelligent mechanism that autonomously determines the right trade-off between continuous user prompts and system usability, according to data types and personal preferences.","{'model': 'tldr@v2.0.0', 'text': 'Blockchain with a resource-aware framework, along with a continuous multifactor authentication mechanism, are envisaged as a potential solution for making assistive robotic systems secure by design; therefore, increasing trust, acceptability, and adoption.'}",https://www.mdpi.com/2076-3417/12/4/2174/pdf?version=1645512954
-representation learning through multimodal attention and time-sync comments for affective video content analysis,Jicai Pan,"Although temporal patterns inherent in visual and audio signals are crucial for affective video content analysis, they have not been thoroughly explored yet. In this paper, we propose a novel Temporal-Aware Multimodal (TAM) method to fully capture the temporal information. Specifically, we design a cross-temporal multimodal fusion module that applies attention-based fusion to different modalities within and across video segments. As a result, it fully captures the temporal relations between different modalities. Furthermore, a single emotion label lacks supervision for learning representation of each segment, making temporal pattern mining difficult. We leverage time-synchronized comments (TSCs) as auxiliary supervision, since these comments are easily accessible and contain rich emotional cues. Two TSC-based self-supervised tasks are designed: the first aims to predict the emotional words in a TSC from video representation and TSC contextual semantics, and the second predicts the segment in which the TSC appears by calculating the correlation between video representation and TSC embedding. These self-supervised tasks are used to pre-train the cross-temporal multimodal fusion module on a large-scale video-TSC dataset, which is crawled from the web without labeling costs. These self-supervised pre-training tasks prompt the fusion module to perform representation learning on segments including TSC, thus capturing more temporal affective patterns. Experimental results on three benchmark datasets show that the proposed fusion module achieves state-of-the-art results in affective video content analysis. Ablation studies verify that after TSC-based pre-training, the fusion module learns more segments' affective patterns and achieves better performance.","{'model': 'tldr@v2.0.0', 'text': 'This paper designs a cross-temporal multimodal fusion module that applies attention-based fusion to different modalities within and across video segments, and achieves state-of-the-art results in affective video content analysis.'}",
-vision+x: a survey on multimodal learning in the light of data,Ye Zhu,"We are perceiving and communicating with the world in a multisensory manner, where different information sources are sophisticatedly processed and interpreted by separate parts of the human brain to constitute a complex, yet harmonious and unified sensing system. To endow the machines with true intelligence, the multimodal machine learning that incorporates data from various modalities has become an increasingly popular research area with emerging technical advances in recent years. In this paper, we present a survey on multimodal machine learning from a novel perspective considering not only the purely technical aspects but also the nature of different data modalities. We analyze the commonness and uniqueness of each data format ranging from vision, audio, text and others, and then present the technical development categorized by the combination of Vision+X, where the vision data play a fundamental role in most multimodal learning works. We investigate the existing literature on multimodal learning from both the representation learning and downstream application levels, and provide an additional comparison in the light of their technical connections with the data nature, e.g., the semantic consistency between image objects and textual descriptions, or the rhythm correspondence between video dance moves and musical beats. The exploitation of the alignment, as well as the existing gap between the intrinsic nature of data modality and the technical designs, will benefit future research studies to better address and solve a specific challenge related to the concrete multimodal task, and to prompt a unified multimodal machine learning framework closer to a real human intelligence system.","{'model': 'tldr@v2.0.0', 'text': 'This paper analyzes the commonness and uniqueness of each data format ranging from vision, audio, text and others, and presents the technical development categorized by the combination of Vision+X, where the vision data play a fundamental role in most multimodal learning works.'}",http://arxiv.org/pdf/2210.02884
-multimodal imaging and biomarkers in cardiac amyloidosis,M. Jung,"Amyloidosis is a progressive infiltrative disease instigated by the extracellular deposition of amyloid fibrils in various organs such as the heart, kidney, and peripheral nerves. Cardiac amyloid deposits cause restrictive cardiomyopathy, leading to a poor prognosis in systemic amyloidosis. The most common etiologies of cardiac amyloidosis (CA) are immunoglobulin light chain deposits (AL-CA) and misfolded transthyretin deposits (ATTR-CA). In recent years, many developments have been accomplished in the field of diagnosis and treatment of CA. At present, ATTR-CA can be noninvasively diagnosed if the following two conditions are fulfilled in the setting of typical echocardiographic/cardiac MRI findings: (1) grade 2 or 3 myocardial uptake in bone scintigraphy confirmed by SPECT and (2) absence of monoclonal protein confirmed by serum-free light chain assay, and serum/urine protein electrophoresis with immunofixation test. Effective therapies are evolving in both types of CA (tafamidis for ATTR-CA and immunologic treatments for AL-CA). Thus, early suspicion and prompt diagnosis are crucial for achieving better outcomes. In this review, we have summarized the role of multimodal imaging (e.g., echocardiography, cardiac MRI, and bone scintigraphy) and biomarkers (e.g., troponin, BNP) in the diagnosis, risk stratification, and treatment monitoring of CA.","{'model': 'tldr@v2.0.0', 'text': 'The role of multimodal imaging and biomarkers in the diagnosis, risk stratification, and treatment monitoring of CA is summarized and early suspicion and prompt diagnosis are crucial for achieving better outcomes.'}",https://www.mdpi.com/2075-4418/12/3/627/pdf?version=1646296989
-paracentral acute middle maculopathy after covid-19: multimodal evaluation,C. Castro,"Paracentral acute middle maculopathy is a spectral domain optical coherence tomography finding characterized by macular band-like hyperreflective lesions that affect the middle layers of the retina, above the outer plexiform layer. COVID-19 may induce different ophthalmological manifestations, including retinal vascular changes. We report a case of paracentral acute middle maculopathy after SARS-CoV-2 infection. Purpose: To report the case and multimodal imaging findings of a healthy young woman who developed paracentral acute middle maculopathy 9 weeks after COVID-19. Methods: Case report. Ultra-widefield fundus photography, macular spectral domain optical coherence tomography, fluorescein angiography, and optical coherence tomography angiography were performed. Results: A 36-year-old woman went to the emergency department with sudden, painless, left eye vision loss. The only relevant medical history was COVID-19 9 weeks before. The best-corrected visual acuity was 20/200, a left eye relative afferent pupillary defect was present, and superficial hemorrhages throughout the macular area and peripheral retina were found. Nearly 4 hours after admission, the left eye best-corrected visual acuity recovered to 20/20 without relative afferent pupillary defect. Five days after presentation in the emergency department, the patient returned with recurrent left eye vision loss, with spontaneous recovery within 12 hours. Macular spectral domain optical coherence tomography revealed hyperreflectivity of the inner plexiform and inner nuclear layers, and the diagnosis of paracentral acute middle maculopathy was established. The patient started oral acetylsalicylic acid and oral prednisolone. The patient did not report any new episodes of vision loss, and there was a progressive resolution of abnormal fundus findings. Conclusion: SARS-CoV-2 infection increases the risk of vascular thrombotic events with possible involvement of the retinal circulation, and paracentral acute middle maculopathy may present as a possible complication. Ophthalmologists should be able to recognize it promptly through multimodal imaging findings.","{'model': 'tldr@v2.0.0', 'text': 'SARS-CoV-2 infection increases the risk of vascular thrombotic events with possible involvement of the retinal circulation, and paracentral acute middle maculopathy may present as a possible complication.'}",
-few-shot multimodal sentiment analysis based on multimodal probabilistic fusion prompts,Xiaocui Yang,"Multimodal sentiment analysis has gained significant attention due to the proliferation of multimodal content on social media. However, existing studies in this area rely heavily on large-scale supervised data, which is time-consuming and labor-intensive to collect. Thus, there is a need to address the challenge of few-shot multimodal sentiment analysis. To tackle this problem, we propose a novel method called Multimodal Probabilistic Fusion Prompts (MultiPoint) that leverages diverse cues from different modalities for multimodal sentiment detection in the few-shot scenario. Specifically, we start by introducing a Consistently Distributed Sampling approach called CDS, which ensures that the few-shot dataset has the same category distribution as the full dataset. Unlike previous approaches primarily using prompts based on the text modality, we design unified multimodal prompts to reduce discrepancies between different modalities and dynamically incorporate multimodal demonstrations into the context of each multimodal instance. To enhance the model's robustness, we introduce a probabilistic fusion method to fuse output predictions from multiple diverse prompts for each input. Our extensive experiments on six datasets demonstrate the effectiveness of our approach. First, our method outperforms strong baselines in the multimodal few-shot setting. Furthermore, under the same amount of data (1% of the full dataset), our CDS-based experimental results significantly outperform those based on previously sampled datasets constructed from the same number of instances of each class.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a novel method called Multimodal Probabilistic Fusion Prompts (MultiPoint) that leverages diverse cues from different modalities for multimodal sentiment detection in the few-shot scenario and introduces a probabilistic fusion method to fuse output predictions from multiple diverse prompts for each input.'}",https://dl.acm.org/doi/pdf/10.1145/3581783.3612181
-beyond bounding box: multimodal knowledge learning for object detection,Wei Feng,"Multimodal supervision has achieved promising results in many visual language understanding tasks, where the language plays an essential role as a hint or context for recognizing and locating instances. However, due to the defects of the human-annotated language corpus, multimodal supervision remains unexplored in fully supervised object detection scenarios. In this paper, we take advantage of language prompt to introduce effective and unbiased linguistic supervision into object detection, and propose a new mechanism called multimodal knowledge learning (\textbf{MKL}), which is required to learn knowledge from language supervision. Specifically, we design prompts and fill them with the bounding box annotations to generate descriptions containing extensive hints and context for instances recognition and localization. The knowledge from language is then distilled into the detection model via maximizing cross-modal mutual information in both image- and object-level. Moreover, the generated descriptions are manipulated to produce hard negatives to further boost the detector performance. Extensive experiments demonstrate that the proposed method yields a consistent performance gain by 1.6\% $\sim$ 2.1\% and achieves state-of-the-art on MS-COCO and OpenImages datasets.","{'model': 'tldr@v2.0.0', 'text': 'This paper takes advantage of language prompt to introduce effective and unbiased linguistic supervision into object detection, and proposes a new mechanism called multimodal knowledge learning (\\textbf{MKL}), which is required to learn knowledge from language supervision.'}",http://arxiv.org/pdf/2205.04072
-multimodal opioid-sparing analgesia for total knee arthroplasty: results from a retrospective case series of 40 patients,A. Bronstone,"The high burden of opioids has prompted orthopedic surgeons to identify effective nonopioid pain management modalities. In this pilot retrospective case series, we describe opioid use and pain during the first 3 months after total knee arthroplasty (TKA) for consecutive patients treated with opioid-free multimodal analgesia. During May and October 2020, 40 consecutive patients who underwent primary unilateral TKA performed by a single surgeon at an academic community hospital received a multimodal analgesia regimen consisting of preoperative cryoneurolysis; perioperative oral and intravenous analgesics, a neuraxial (spinal) and regional adductor canal block, and local infiltration of liposomal bupivacaine (Exparel); intraoperative periarticular infiltration of bupivacaine hydrochloride (Marcaine); and postoperative oral non-opioid analgesics. A Prescription Monitoring Program database was examined to capture opioid prescriptions filled during the 3 months before and after TKA. Self-reported pain was assessed before surgery and at 2 weeks and 3 months after surgery. Two thirds (67.5%; 27/40) of patients recovered from TKA without using opioids. Whereas 100% of the experienced opioid users required opioids after TKA, only 15.6% of opioid-naïve patients required postoperative opioids. Opioid-experienced patients had a higher mean number of post-TKA opioid prescriptions (3.1 vs. 1.6) and total morphine milligram equivalent (985 vs. 265) than opioid-naïve patients. Patients who used opioids after TKA reported higher levels of pain at each time point than those who had opioid-free TKA. Multimodal analgesia allows for opioid-free TKA for the majority of patients, particularly those without a history of recent opioid use.","{'model': 'tldr@v2.0.0', 'text': 'This pilot retrospective case series describes opioid use and pain during the first 3 months after total knee arthroplasty (TKA) for consecutive patients treated with opioid-free multimodal analgesia for the majority of patients, particularly those without a history of recent opioid use.'}",https://journaloei.scholasticahq.com/article/33296.pdf
-weakly supervised multimodal 30-day all-cause mortality prediction for pulmonary embolism patients,Noa Cahan,"Pulmonary embolism (PE) is a common life-threatening condition with a challenging diagnosis, as patients often present with nonspecific symptoms. Prompt and accurate detection of PE and specifically an assessment of its severity are critical for managing patient treatment. We introduce diverse multimodal fusion models that are capable of utilizing weakly-labeled multi-modal data, combining both volumetric pixel data and clinical patient data for automatic risk stratification of PE. The best performing multimodality model is an intermediate fusion model that achieves an area under the curve (AUC) of 0.96 for assessing PE severity, with a sensitivity of 90% and specificity of 94%. To the best of our knowledge, this is the first study that attempted to automatically assess PE severity.","{'model': 'tldr@v2.0.0', 'text': 'Diverse multimodal fusion models that are capable of utilizing weakly-labeled multi-modal data are introduced, combining both volumetric pixel data and clinical patient data for automatic risk stratification of PE.'}",
-added value of ultrasound-based multimodal imaging to diagnose hepatic sclerosed hemangioma before biopsy and resection,Feiqian Wang,"Imaging methods have the overwhelming advantage of being non-invasive in the diagnosis of hepatic lesions and, thanks to technical developments in the field of ultrasound (US), radiation exposure can also be avoided in many clinical situations. In particular, contrast-enhanced US (CEUS) outperforms other radiological methods in regard to real-time images, repeatability, and prompt reporting and demonstrates relatively few contraindications and adverse reactions. In this study, we reported in detail a rare benign tumor: hepatic sclerosed hemangioma (HSH). We described US-based multimodal imaging (B-flow imaging, US elastography, and Sonazoid CEUS) features of this HSH case. Furthermore, by summarizing the recently published literature on the imaging diagnosis of HSH, we offered readers comprehensive knowledge of conventional imaging methods (CT, MRI) and CEUS in the diagnosis of HSH and preliminarily discussed their mechanism of pathology-based diagnosis. Our multimodal imaging approach may provide a diagnostic strategy for HSH, thus avoiding unnecessary biopsy or resection.","{'model': 'tldr@v2.0.0', 'text': 'This study reported in detail a rare benign tumor: hepatic sclerosed hemangioma (HSH) and described US-based multimodal imaging features of this HSH case and offered readers comprehensive knowledge of conventional imaging methods (CT, MRI) and CEUS in the diagnosis of HSH.'}",https://www.mdpi.com/2075-4418/12/11/2818/pdf?version=1668590166
-digital storytelling: a didactic approach to multimodal coherence,S. Meier,"The aim of this article is to explain multimodal coherence-making as a transcribing practice and how this can be used to teach multimodal, narrative, and media competences in different genres. In multimodal arrangements, language makes images readable in specific ways and images make language understandable in different ways. This results in an abductive understanding process that can be used in teaching and learning contexts. This idea of meaning-making is based on the social semiotic approach of style. According to the understanding of semiotic meta functions, this approach considers style as the practice of selecting, forming, and composing semiotic resources. These stylistic practices realize a subjective appropriation of discursive and habitual patterns, which are carried out within the semiotic and technological dispositions (affordances) of the situationally used media infrastructures. In this sense, digital storytelling is a multimodal style practice with digital tools. Multimodal storytelling in educational contexts means that teachers and learners are prompted to bring the communicative functions of text, image, video, and audio into narrative coherence. Based on a journalistic Instagram story, this article reconstructs the media-practical, multimodal, and narrative skills that are prototypically necessary. Based on this analysis, these competencies are operationalized to make them usable for new teaching/learning arrangements using digital storytelling.",,https://www.frontiersin.org/articles/10.3389/fcomm.2022.906268/pdf
-multimodal representations learning based on mutual information maximization and minimization and identity embedding for multimodal sentiment analysis,Jiahao Zheng,"Multimodal sentiment analysis (MSA) is a fundamental complex research problem due to the heterogeneity gap between di ff erent modalities and the ambiguity of human emotional expression. Although there have been many successful attempts to construct multimodal representations for MSA, there are still two challenges to be addressed: 1) A more robust multimodal representation needs to be constructed to bridge the heterogeneity gap and cope with the complex multimodal interactions, and 2) the contextual dynamics must be modeled e ff ectively throughout the information flow. In this work, we propose a multimodal representation model based on Mutual information Maximization and Minimization and Identity Embedding (MMMIE). We combine mutual information maximization between modal pairs, and mutual information minimization between input data and corresponding features to mine the modal-invariant and task-related information. Furthermore, Identity Embedding is proposed to prompt the downstream network to perceive the contextual information. Experimental results on two public datasets demonstrate the e ff ectiveness of the proposed model.","{'model': 'tldr@v2.0.0', 'text': 'A multimodal representation model based on Mutual information Maximization and Minimization and Identity Embedding (MMMIE) is proposed that combines mutual information maximization between modal pairs, and mutual information minimization between input data and corresponding features to mine the modal-invariant and task-related information.'}",
-"feasibility of a multimodal exercise, nutrition, and palliative care intervention in advanced lung cancer",M. Ester,,"{'model': 'tldr@v2.0.0', 'text': 'A multimodal physical activity, nutrition, and palliative symptom management intervention is feasible and shows potential benefits on quality of life that warrant further investigation in a larger cohort\xa0trial.'}",https://bmccancer.biomedcentral.com/track/pdf/10.1186/s12885-021-07872-y
-multimodal study of prph2 gene-related retinal phenotypes,G. Antonelli,"PRPH2 gene mutations are frequently found in inherited retinal dystrophies (IRD) and are associated with a wide spectrum of clinical phenotypes. We studied 28 subjects affected by IRD carrying pathogenic PRPH2 mutations, belonging to 11 unrelated families. Functional tests (best-corrected visual acuity measurement, chromatic test, visual field, full-field, 30 Hz flicker, and multifocal electroretinogram), morphological retino-choroidal imaging (optical coherence tomography, optical coherence tomography angiography, and fundus autofluorescence), and clinical data were collected and analyzed. Common primary complaints, with onset in their 40s, were visual acuity reduction and abnormal dark adaptation. Visual acuity ranged from light perception to 20/20 Snellen. Visual field peripheral constriction and central scotoma were found. Chromatic sense was reduced in one third of patients. Electrophysiological tests were abnormal in most of the patients. Choroidal neovascular lesions were detected in five patients. Three novel PRPH2 variants were found in four different families. Based on the present multimodal study, we identified seven distinct PRPH2 phenotypes in 11 unrelated families carrying either different mutations or the same mutation, both within the same family or among them. Fundus autofluorescence modality turned out to be the most adequate imaging method for early recognition of this dystrophy, and the optical coherence tomography angiography was highly informative to promptly detect choroidal neovascularization, even in the presence of the extensive chorioretinal atrophy phenotype.","{'model': 'tldr@v2.0.0', 'text': 'Fundus autofluorescence modality turned out to be the most adequate imaging method for early recognition of this dystrophy, and the optical coherence tomography angiography was highly informative to promptly detect choroidal neovascularization, even in the presence of the extensive chorioretinal atrophy phenotype.'}",https://www.mdpi.com/2075-4418/12/8/1851/pdf?version=1660106282
-nonopioid treatments for chronic pain-integrating multimodal biopsychosocial approaches to pain management.,S. Eucker,"Chronic pain is a major health burden in the US, with recent estimates indicating that 1 in 5 adults has chronic pain.1 Nearly 40% of these individuals experience substantial impairments in function or daily activities as a result of their pain.1 The prevalence of chronic pain has been growing in the past 2 decades, accompanied by an increase in opioid prescribing for pain. However, the exponential increase in opioid-related deaths prompted the development of the 2016 Centers for Disease Control and Prevention (CDC) guideline to reduce opioid prescribing for chronic pain.2 This guideline, along with several other efforts to limit opioids, has been associated with decreased opioid prescribing rates across multiple settings and patient populations.3,4 Although promising, minimizing the use of 1 treatment strategy without ensuring access to other effective and viable alternatives may expose patients to inadequate treatment and worsening of chronic pain. The CDC guideline attempts to address this issue by recommending nonopioid medication and nonpharmacologic pain management as first-line treatments for chronic pain.2 However, a major question remains: has the reduced use of opioids been accompanied by greater use of alternative treatments for chronic pain? The study by Goldstick et al5 takes a major step toward answering this question. The authors assessed the prescribing rates of several classes of nonopioid pain medications before and after the release of the 2016 CDC guideline. They included a large nationwide cohort of commercially insured patients with some of the most common types of chronic pain: neck or back pain, osteoarthritis, nonmigraine headaches, and fibromyalgia. The analysis focused on specific medication classes that have been commonly prescribed for chronic pain: nonsteroidal anti-inflammatory drugs, acetaminophen, gabapentinoid and carbamazepine anticonvulsants, and antidepressants. These medications have been shown to provide small-to-moderate short-term benefit and are recommended by guidelines for chronic pain treatment.6 Medications whose evidence was primarily for acute pain management (eg, skeletal muscle relaxants) were not included. The authors constructed sequential cohorts to analyze the prescribing patterns over time for the overall population and for specific subpopulations. Similar to previous work, the study by Goldstick et al5 found greater reductions in opioid prescribing after the release of the 2016 CDC guideline compared with the preguideline pattern. The authors also found that, although nonopioid prescribing rates had remained relatively constant over the 4-year period before the release of the guideline, these rates steadily increased during the 2 years after the release. Similar patterns were found when examining specific subpopulations, including those with chronic pain, previous opioid exposure, substance use disorder, anxiety disorder, or mood disorder.5 These shifts suggest that nonopioid medications replaced opioid medications, but the overall increase in nonopioid prescribing rates was small. On the other hand, claims data analysis cannot account for commonly used over-the-counter medications (eg, nonsteroidal anti-inflammatory drugs and acetaminophen). Furthermore, although not guideline-recommended, other medication classes (eg, benzodiazepines and topical lidocaine) that are sometimes prescribed for pain were not included in the analysis. Thus, these findings may potentially underrepresent the increase in clinician recommendations for nonopioid pain medications. + Related article","{'model': 'tldr@v2.0.0', 'text': 'A large nationwide cohort of commercially insured patients with some of the most common types of chronic pain included, found greater reductions in opioid prescribing after the release of the 2016 CDC guideline, but the overall increase in nonopioid prescribing rates was small.'}",https://jamanetwork.com/journals/jamanetworkopen/articlepdf/2793231/eucker_2022_ic_220103_1654263506.10054.pdf
-"multimodal c4: an open, billion-scale corpus of images interleaved with text",Wanrong Zhu,"In-context vision and language models like Flamingo support arbitrarily interleaved sequences of images and text as input. This format not only enables few-shot learning via interleaving independent supervised (image, text) examples, but also, more complex prompts involving interaction between images, e.g.,""What do image A and image B have in common?""To support this interface, pretraining occurs over web corpora that similarly contain interleaved images+text. To date, however, large-scale data of this form have not been publicly available. We release Multimodal C4, an augmentation of the popular text-only C4 corpus with images interleaved. We use a linear assignment algorithm to place images into longer bodies of text using CLIP features, a process that we show outperforms alternatives. Multimodal C4 spans everyday topics like cooking, travel, technology, etc. A manual inspection of a random sample of documents shows that a vast majority (88%) of images are topically relevant, and that linear assignment frequently selects individual sentences specifically well-aligned with each image (80%). After filtering NSFW images, ads, etc., the resulting corpus consists of 101.2M documents with 571M images interleaved in 43B English tokens.","{'model': 'tldr@v2.0.0', 'text': 'This work releases Multimodal C4, an augmentation of the popular text-only C4 corpus with images interleaved, and uses a linear assignment algorithm to place images into longer bodies of text using CLIP features, a process that shows outperforms alternatives.'}",http://arxiv.org/pdf/2304.06939
-"bionic soft multimodal actuators for fast, large deformation under ultralow magnetic conditions",Keren Dai,"Very recently, magnetically driven soft actuators have prompted increasing interest due to their programmable deformation, swift response, and remote actuation. However, it is still challenging to trigger strong and fast actuating performances with an extremely weak magnetic field due to the difficult task of in situ programming of magnetic domains and the limited mechanical structures. Here, first a bionic sandwich structure is proposed for designing soft magnetic actuators with a specific threshold value. The crab‐like jointed structure of PDMS‐embedded NdFeB upper/lower layers is essential for generating the desirable threshold effect, while a flytrap‐inspired soft interlayer is further implemented to decrease the driving magnetic field. Theoretical analysis and numerical simulations are implemented to optimize the actuating performances of soft actuators (deformation rate, deformation angle) by modulating the structural parameters. Experimental results show that the biomimetic features yield a promptly switchable bistable state, a superior strong deformation rate of 1.93, and a maximum deformation angle of 25.5° under an ultralow magnetic field of 1 mT. Two demonstrative applications of soft actuators are investigated, including threshold switching and soft grippers, suggesting their broad applications in engineering fields. Remarkably, the magnetic soft grippers associated with the gecko‐inspired adhesion surface exhibit an improved and stable grasping ability. This work focuses on designing highly deformable actuators and functions in emerging areas of soft robotics.",,
-a prospective study evaluating health-related quality of life following a multimodal treatment for colorectal cancer,M. Perrone,,"{'model': 'tldr@v2.0.0', 'text': 'A better HR-QoL at baseline was associated with better physical, social and sexual functioning, positive body image and sexual pleasure after 5\xa0years, and allows the early detection of patients at risk, favoring prompt patient-centered interventions.'}",
-n24news: a new dataset for multimodal news classification,Zhen Wang,"Current news datasets merely focus on text features on the news and rarely leverage the feature of images, excluding numerous essential features for news classification. In this paper, we propose a new dataset, N24News, which is generated from New York Times with 24 categories and contains both text and image information in each news. We use a multitask multimodal method and the experimental results show multimodal news classification performs better than text-only news classification. Depending on the length of the text, the classification accuracy can be increased by up to 8.11%. Our research reveals the relationship between the performance of a multimodal classifier and its sub-classifiers, and also the possible improvements when applying multimodal in news classification. N24News is shown to have great potential to prompt the multimodal news studies.","{'model': 'tldr@v2.0.0', 'text': 'A new dataset, N24News, is proposed, which is generated from New York Times with 24 categories and contains both text and image information in each news and shows multimodal news classification performs better than text-only news classification.'}",
-a novel multimodal method for depression identification,Rahul Singhal,"Depression is one of the most prominent mental health issues, characterized by a depressed low mood and an absence of enthusiasm in activities. In terms of early detection, accurate diagnosis, and effective treatment, doctors face a serious challenge from depression, which is a serious global health issue. For patients with this mental disease to receive prompt medical attention and improve their general well-being, early identification is essential. For the purpose of detecting various psychological illnesses including depression, anxiety, and post-traumatic stress disorder, medical audio consultations along with survey responses have been used. A depressed individual displays a range of subtle signs that may be more easily identified by combining the results of multiple modalities. Multimodality involves extracting maximum information from data by using multiple modes, so that the deep learning model can be trained efficiently to give better results. Given that each modality functions differently, combining various modalities is not easy, and each origin of a modality takes on a different form. It is clear from the literature that is currently significant in the area that, combining the modalities yields positive outcomes. A trustworthy approach to identify depression is thus urgently needed because it continues to be a problem for many individuals in today’s society. In this work, textual and audio features are incorporated related to the identification of depression, and a novel multimodal approach using an optimized Bi-directional Long Short -Term Memory model that recognizes premature depression is suggested for medical intervention before it develops further.",,
-a multimodal deep learning-based fault detection model for a plastic injection molding process,Gyeongho Kim,"The authors of this work propose a deep learning-based fault detection model that can be implemented in the field of plastic injection molding. Compared to conventional approaches to fault detection in this domain, recent deep learning approaches prove useful for on-site problems involving complex underlying dynamics with a large number of variables. In addition, the advent of advanced sensors that generate data types in multiple modalities prompts the need for multimodal learning with deep neural networks to detect faults. This process is able to facilitate information from various modalities in an end-to-end learning fashion. The proposed deep learning-based approach opts for an early fusion scheme, in which the low-level feature representations of modalities are combined. A case study involving real-world data, obtained from a car parts company and related to a car window side molding process, validates that the proposed model outperforms late fusion methods and conventional models in solving the problem.","{'model': 'tldr@v2.0.0', 'text': 'A case study involving real-world data, obtained from a car parts company and related to a car window side molding process, validates that the proposed model outperforms late fusion methods and conventional models in solving the problem.'}",https://ieeexplore.ieee.org/ielx7/6287639/6514899/09548039.pdf
-"introducing wesad, a multimodal dataset for wearable stress and affect detection",P. Schmidt,"Affect recognition aims to detect a person's affective state based on observables, with the goal to e.g. improve human-computer interaction. Long-term stress is known to have severe implications on wellbeing, which call for continuous and automated stress monitoring systems. However, the affective computing community lacks commonly used standard datasets for wearable stress detection which a) provide multimodal high-quality data, and b) include multiple affective states. Therefore, we introduce WESAD, a new publicly available dataset for wearable stress and affect detection. This multimodal dataset features physiological and motion data, recorded from both a wrist- and a chest-worn device, of 15 subjects during a lab study. The following sensor modalities are included: blood volume pulse, electrocardiogram, electrodermal activity, electromyogram, respiration, body temperature, and three-axis acceleration. Moreover, the dataset bridges the gap between previous lab studies on stress and emotions, by containing three different affective states (neutral, stress, amusement). In addition, self-reports of the subjects, which were obtained using several established questionnaires, are contained in the dataset. Furthermore, a benchmark is created on the dataset, using well-known features and standard machine learning methods. Considering the three-class classification problem ( baseline vs. stress vs. amusement ), we achieved classification accuracies of up to 80%,. In the binary case ( stress vs. non-stress ), accuracies of up to 93%, were reached. Finally, we provide a detailed analysis and comparison of the two device locations ( chest vs. wrist ) as well as the different sensor modalities.","{'model': 'tldr@v2.0.0', 'text': 'This work introduces WESAD, a new publicly available dataset for wearable stress and affect detection that bridges the gap between previous lab studies on stress and emotions, by containing three different affective states (neutral, stress, amusement).'}",
-improvement of autonomous vehicles trust through synesthetic-based multimodal interaction,Xiaofeng Sun,"Trust is the key factor for people to accept autonomous vehicles(AVs). Existing studies have reported that multimodal interaction would enhance people’s trust in AVs. However, these researches mainly focus on the superposition effect between sensory channels, and lack on the research of correlation between different sensory channels and its influence on AVs trust. Therefore, we innovatively introduce synesthesia theory for the research of improving AVs trust. We present an AVs multimodal interaction model based on audio-visual synesthesia theory, and finally prove that the model has a definite effect on improving AVs trust by experiments. Firstly, 82 participants are recruited and assigned into two groups: Group A (non-synesthesia group) and Group B (synesthesia group). They conduct an experimental driving experienced normal traffic conditions (NTC) (turning, traffic lights, over and limit speed prompts) and emergency traffic condition (ETC) (sudden braking of the car in front, temporary lane change, pedestrian thrusting) while completing a secondary task. Then, we conduct a survey (questionnaire and interviews) to evaluate the attitude about trust, technical competence, situation management and perceived ease of use after participants finished experimental driving. The results demonstrate that synesthetic-based multimodal interaction (SBMI) can more effectively remind people of relevant information especially under ETC. SBMI model is more effective than single information stimulus or non-synesthetic audio-visual information stimulus not only in terms of information transmission efficiency and effect, but also in terms of output response/ action. The results also show that SBMI contributes to the improvement of AVs trust. These findings provide evidence on the importance of SBMI to the improvement of AVs trust. The findings of this study will be helpful to the future design of AVs interaction system.","{'model': 'tldr@v2.0.0', 'text': 'An AVs multimodal interaction model based on audio-visual synesthesia theory is presented, and it is proved that the model has a definite effect on improving AVs trust by experiments, and demonstrates that synesthetic-based multi-modal interaction (SBMI) can more effectively remind people of relevant information especially under ETC.'}",https://ieeexplore.ieee.org/ielx7/6287639/9312710/09353546.pdf
-multimodal imaging in susac syndrome: a case report and literature review,S. Bagaglia,"Susac syndrome (SS) is a rare microangiopathy that involves arterioles of the brain, retina, and cochlea. Diagnosis is extremely difficult because of the rarity of the disease and because the signs and symptoms often occur at different times. Multidisciplinary approaches and multimodal images are mandatory for diagnosis and prompt therapy. In this report, we describe a case of SS and the application of multimodal retinal imaging to evaluate the ophthalmologic changes and to confirm diagnosis. Early diagnosis and therapy based on the associations of steroids and immunosuppressants are necessary to limit the sequelae of the disease.","{'model': 'tldr@v2.0.0', 'text': 'A case of SS is described and the application of multimodal retinal imaging is described to evaluate the ophthalmologic changes and to confirm diagnosis.'}",https://www.mdpi.com/1660-4601/18/7/3435/pdf?version=1617936888
-multimodal deep generative models for trajectory prediction: a conditional variational autoencoder approach,B. Ivanovic,"Human behavior prediction models enable robots to anticipate how humans may react to their actions, and hence are instrumental to devising safe and proactive robot planning algorithms. However, modeling complex interaction dynamics and capturing the possibility of many possible outcomes in such interactive settings is very challenging, which has recently prompted the study of several different approaches. In this work, we provide a self-contained tutorial on a conditional variational autoencoder (CVAE) approach to human behavior prediction which, at its core, can produce a multimodal probability distribution over future human trajectories conditioned on past interactions and candidate robot future actions. Specifically, the goals of this tutorial paper are to review and build a taxonomy of state-of-the-art methods in human behavior prediction, from physics-based to purely data-driven methods, provide a rigorous yet easily accessible description of a data-driven, CVAE-based approach, highlight important design characteristics that make this an attractive model to use in the context of model-based planning for human-robot interactions, and provide important design considerations when using this class of models.","{'model': 'tldr@v2.0.0', 'text': 'This work provides a self-contained tutorial on a conditional variational autoencoder (CVAE) approach to human behavior prediction which, at its core, can produce a multimodal probability distribution over future human trajectories conditioned on past interactions and candidate robot future actions.'}",https://arxiv.org/pdf/2008.03880
-antlion re-sampling based deep neural network model for classification of imbalanced multimodal stroke dataset,T. G.,,"{'model': 'tldr@v2.0.0', 'text': 'This study focuses on improving the quality of stroke data implementing a rigorous pre-processing technique using a multimodal stroke dataset available in the publicly available Kaggle repository and proves the superiority of proposed model.'}",https://e-space.mmu.ac.uk/626646/7/G2020_Article_AntlionRe-samplingBasedDeepNeu%20e.pdf
-concussion assessment potentially aided by use of an objective multimodal concussion index,A. Jacquin,"Objective Prompt, accurate, objective assessment of concussion is crucial as delays can lead to increased short and long-term consequences. The purpose of this study was to derive an objective multimodal concussion index (CI) using EEG at its core, to identify concussion, and to assess change over time throughout recovery. Methods Male and female concussed (N = 232) and control (N = 206) subjects 13–25 years were enrolled at 12 US colleges and high schools. Evaluations occurred within 72 h of injury, 5 days post-injury, at return-to-play (RTP), 45 days after RTP (RTP + 45); and included EEG, neurocognitive performance, and standard concussion assessments. Concussed subjects had a witnessed head impact, were removed from play for ≥ 5 days using site guidelines, and were divided into those with RTP < 14 or ≥14 days. Part 1 describes the derivation and efficacy of the machine learning derived classifier as a marker of concussion. Part 2 describes significance of differences in CI between groups at each time point and within each group across time points. Results Sensitivity = 84.9%, specificity = 76.0%, and AUC = 0.89 were obtained on a test Hold-Out group representing 20% of the total dataset. EEG features reflecting connectivity between brain regions contributed most to the CI. CI was stable over time in controls. Significant differences in CI between controls and concussed subjects were found at time of injury, with no significant differences at RTP and RTP + 45. Within the concussed, differences in rate of recovery were seen. Conclusions The CI was shown to have high accuracy as a marker of likelihood of concussion. Stability of CI in controls supports reliable interpretation of CI change in concussed subjects. Objective identification of the presence of concussion and assessment of readiness to return to normal activity can be aided by use of the CI, a rapidly obtained, point of care assessment tool.","{'model': 'tldr@v2.0.0', 'text': 'The CI was shown to have high accuracy as a marker of likelihood of concussion and assessment of readiness to return to normal activity can be aided by use of the CI, a rapidly obtained, point of care assessment tool.'}",
-acute basilar artery occlusion (bao): a pictorial review of multimodal imaging findings,Andrés Yesid Vásquez-Codina,,"{'model': 'tldr@v2.0.0', 'text': 'This pictorial essay illustrates the essential role that multimodal imaging plays in the prompt diagnosis, management, and overall outcome of patients with acute BAO.'}",
-caption anything: interactive image description with diverse multimodal controls,Teng Wang,"Controllable image captioning is an emerging multimodal topic that aims to describe the image with natural language following human purpose, $\textit{e.g.}$, looking at the specified regions or telling in a particular text style. State-of-the-art methods are trained on annotated pairs of input controls and output captions. However, the scarcity of such well-annotated multimodal data largely limits their usability and scalability for interactive AI systems. Leveraging unimodal instruction-following foundation models is a promising alternative that benefits from broader sources of data. In this paper, we present Caption AnyThing (CAT), a foundation model augmented image captioning framework supporting a wide range of multimodel controls: 1) visual controls, including points, boxes, and trajectories; 2) language controls, such as sentiment, length, language, and factuality. Powered by Segment Anything Model (SAM) and ChatGPT, we unify the visual and language prompts into a modularized framework, enabling the flexible combination between different controls. Extensive case studies demonstrate the user intention alignment capabilities of our framework, shedding light on effective user interaction modeling in vision-language applications. Our code is publicly available at https://github.com/ttengwang/Caption-Anything.","{'model': 'tldr@v2.0.0', 'text': 'Caption AnyThing (CAT) is presented, a foundation model augmented image captioning framework supporting a wide range of multimodel controls: 1) visual controls, including points, boxes, and trajectories; 2) language controls, such as sentiment, length, language, and factuality.'}",https://arxiv.org/pdf/2305.02677
-vip5: towards multimodal foundation models for recommendation,Shijie Geng,"Computer Vision (CV), Natural Language Processing (NLP), and Recommender Systems (RecSys) are three prominent AI applications that have traditionally developed independently, resulting in disparate modeling and engineering methodologies. This has impeded the ability for these fields to directly benefit from each other's advancements. With the recent development of foundation models, large language models have emerged as a potential general-purpose interface for unifying different modalities and problem formulations. In light of this, we propose the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks. This will enable the processing of multiple modalities in a shared architecture for improved recommendations. To achieve this, we introduce multimodal personalized prompts to accommodate multiple modalities under a shared format. Additionally, we propose a parameter-efficient training method for foundation models, which involves freezing the P5 backbone and fine-tuning lightweight adapters, resulting in improved recommendation performance and increased efficiency in terms of training time and memory usage. Code and data of VIP5 are available at https://github.com/jeykigung/VIP5.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes the development of a multimodal foundation model (MFM) considering visual, textual, and personalization modalities under the P5 recommendation paradigm, thus named VIP5 (Visual P5), to unify various modalities and recommendation tasks.'}",http://arxiv.org/pdf/2305.14302
-meta learning to bridge vision and language models for multimodal few-shot learning,Ivona Najdenkoska,"Multimodal few-shot learning is challenging due to the large domain gap between vision and language modalities. Existing methods are trying to communicate visual concepts as prompts to frozen language models, but rely on hand-engineered task induction to reduce the hypothesis space. To make the whole process learnable, we introduce a multimodal meta-learning approach. Specifically, our approach decomposes the training of the model into a set of related multimodal few-shot tasks. We define a meta-mapper network, acting as a meta-learner, to efficiently bridge frozen large-scale vision and language models and leverage their already learned capacity. By updating the learnable parameters only of the meta-mapper, it learns to accrue shared meta-knowledge among these tasks. Thus, it can rapidly adapt to newly presented samples with only a few gradient updates. Importantly, it induces the task in a completely data-driven manner, with no need for a hand-engineered task induction. We evaluate our approach on recently proposed multimodal few-shot benchmarks, measuring how rapidly the model can bind novel visual concepts to words and answer visual questions by observing only a limited set of labeled examples. The experimental results show that our meta-learning approach outperforms the baseline across multiple datasets and various training settings while being computationally more efficient.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the multimodal meta-learning approach outperforms the baseline across multiple datasets and various training settings while being computationally more efficient.'}",http://arxiv.org/pdf/2302.14794
-parents and children: distinguishing multimodal deepfakes from natural images,Roberto Amoroso,"Recent advancements in diffusion models have enabled the generation of realistic deepfakes by writing textual prompts in natural language. While these models have numerous benefits across various sectors, they have also raised concerns about the potential misuse of fake images and cast new pressures on fake image detection. In this work, we pioneer a systematic study of the authenticity of fake images generated by state-of-the-art diffusion models. Firstly, we conduct a comprehensive study on the performance of contrastive and classification-based visual features. Our analysis demonstrates that fake images share common low-level cues, which render them easily recognizable. Further, we devise a multimodal setting wherein fake images are synthesized by different textual captions, which are used as seeds for a generator. Under this setting, we quantify the performance of fake detection strategies and introduce a contrastive-based disentangling strategy which let us analyze the role of the semantics of textual descriptions and low-level perceptual cues. Finally, we release a new dataset, called COCOFake, containing about 600k images generated from original COCO images.","{'model': 'tldr@v2.0.0', 'text': 'This work pioneer a systematic study of the authenticity of fake images generated by state-of-the-art diffusion models, and conducts a comprehensive study on the performance of contrastive and classification-based visual features.'}",http://arxiv.org/pdf/2304.00500
-towards multimodal computational humanities. using clip to analyze late-nineteenth century magic lantern slides,T. Smits,"The introduction of the CLIP model signaled a breakthrough in multimodal deep learning. This paper examines whether CLIP can be fruitfully applied to a (binary) classification task in the Humanities. We focus on a historical collection of late-nineteenth century magic lantern slides from the Lucerna database. Based on the available metadata, we evaluate CLIP’s performance on classifying slide images into ‘exterior’ and ‘interior’ categories. We compare the performance of several textual prompts for CLIP to two conventional mono-modal models (textual and visual) which we train and evaluate on the same stratified set of 5,244 magic lantern slides and their captions. We find that the textual and multimodal models achieve a respectable performance (∼0.80 accuracy) but are still outperformed by a vision model that was fine-tuned to the task (∼0.89). We flag three methodological issues that might arise from the application of CLIP in the (computational) humanities. First, the lack of (need for) labelled data makes it hard to inspect and/or interpret the performance of the model. Second, CLIP’s zero-shot capability only allows for classification tasks to be simulated, which makes it doubtful if standard metrics can be used to compare its performance to text and/or image models. Third, the lack of effective prompt engineering techniques makes the performance of CLIP (highly) unstable.","{'model': 'tldr@v2.0.0', 'text': 'This paper examines whether CLIP can be fruitfully applied to a (binary) classification task in the Humanities, focusing on a historical collection of late-nineteenth century magic lantern slides from the Lucerna database.'}",
-"a cabinet of curiosities, a dwelling place: weekly writing on instagram as multimodal praxis",Rebecca Conklin,"This Instagram “Weekly Writing” assignment is a social-media-based, low-stakes, and longitudinal approach to teaching and experimenting with multimodal composition. Students create an account for the purposes of the class and follow each other. They post three times per week, sometimes freely and sometimes in response to a prompt or challenge. Together, we use the platform and its rich multimodal resources to consider how in-the-moment multimodal composing can spur invention, place the writer in the perpetual position of noticing, and create an archive of experience that holistically communicates beyond the author’s original intention. This article discusses the pedagogical rationale for this approach, along with the issues to consider before adopting and adapting this practice.",,
-n15news: a new dataset for multimodal news classification,Zhen Wang,"Current news datasets merely focus on text features on the news and rarely leverage the feature of images, excluding numerous essential features for news classification. In this paper, we propose a new dataset, N15News, which is generated from New York Times with 15 categories and contains both text and image information in each news. We design a novel multitask multimodal network with different fusion methods, and experiments show multimodal news classification performs better than text-only news classification. Depending on the length of the text, the classification accuracy can be increased by up to 5.8%. Our research reveals the relationship between the performance of a multimodal classifier and its sub-classifiers, and also the possible improvements when applying multimodal in news classification. N15News is shown to have great potential to prompt the multimodal news studies. The way to collect N15News can be in found here.","{'model': 'tldr@v2.0.0', 'text': 'A new dataset, N15News, which is generated from New York Times with 15 categories and contains both text and image information in each news is proposed, and experiments show multimodal news classification performs better than text-only news classification.'}",
-π-tuning: transferring multimodal foundation models with optimal multi-task interpolation,Chengyue Wu,"Foundation models have achieved great advances in multi-task learning with a unified interface of unimodal and multimodal tasks. However, the potential of such multi-task learners has not been exploited during transfer learning. In this work, we present a universal parameter-efficient transfer learning method, termed Predict-Interpolate Tuning ($\pi$-Tuning), for vision, language, and vision-language tasks. It aggregates the parameters of lightweight task-specific experts learned from similar tasks to aid the target downstream task. The task similarities are predicted in a unified modality-independent space, yielding a scalable graph to demonstrate task relationships. $\pi$-Tuning has several appealing benefits. First, it flexibly explores both intra- and inter-modal transferability between similar tasks to improve the accuracy and robustness of transfer learning, especially in data-scarce scenarios. Second, it offers a systematical solution for transfer learning with multi-task prediction-and-then-interpolation, compatible with diverse types of parameter-efficient experts, such as prompt and adapter. Third, an extensive study of task-level mutual benefits on 14 unimodal and 6 multimodal datasets shows that $\pi$-Tuning surpasses fine-tuning and other parameter-efficient transfer learning methods both in full-shot and low-shot regimes. The task graph also enables an in-depth interpretable analysis of task transferability across modalities. The code will be available at https://github.com/TencentARC/pi-Tuning.","{'model': 'tldr@v2.0.0', 'text': 'Predict-Interpolate Tuning aggregates the parameters of lightweight task-specific experts learned from similar tasks to aid the target downstream task, and surpasses fine-tuning and other parameter-efficient transfer learning methods both in full-shot and low-shot regimes.'}",http://arxiv.org/pdf/2304.14381
-fusing pre-trained language models with multimodal prompts through reinforcement learning,Youngjae Yu,"Language models are capable of commonsense reasoning: while domain-specific models can learn from explicit knowledge (e.g. commonsense graphs [6] ethical norms [25]), and larger models like GPT-3 [7] mani-fest broad commonsense reasoning capacity. Can their knowledge be extended to multimodal inputs such as images and audio without paired domain data? In this work, we propose ‡ESPER (Extending Sensory PErception with Reinforcement learning) which enables text-only pretrained models to address multimodal tasks such as visual commonsense reasoning. Our key novelty is to use rein-forcement learning to align multimodal inputs to language model generations without direct supervision: for example, our reward optimization relies only on cosine similarity derived from CLIP [52] and requires no additional paired (image, text) data. Experiments demonstrate that ESPER outperforms baselines and prior work on a variety of multimodal text generation tasks ranging from captioning to commonsense reasoning; these include a new benchmark we collect and release, the ESP dataset, which tasks models with generating the text of several different domains for each image. Our code and data are publicly released at https://github.com/JiwanChung/esper.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes ‡ESPER (Extending Sensory PErception with Reinforcement learning) which enables text-only pretrained models to address multimodal tasks such as visual commonsense reasoning.'}",
-multimodal (non-)participation and (dis)engagement,Christine M. Jacknick,"While subtler forms of engagement exist, the ones that “count” most are often verbal, because teachers notice those contributions more readily and they seem to have greater potential to affect the trajectory of classroom interaction. This chapter prompts researchers and teachers to examine the full range of students’ multimodal actions to characterize their participation and engagement. Students create meaning with their bodies while also speaking, but given the constraints of multiparty interaction, students often display their engagement with the ongoing interaction primarily in an embodied way. In operationalizing the concept of “engagement” as an embodied phenomenon, this chapter shows the crucial importance of the alignment, temporality, and sequentiality of students’ multimodal actions. Finally, this chapter uncovers the moment-by-moment nature of students’ embodied responses, showing how these actions often occur in a wave, potentially obscuring the teacher’s (and analyst’s) sense of how any individual student is participating in classroom interaction.",,
-deep learning-based acute ischemic stroke lesion segmentation method on multimodal mr images using a few fully labeled subjects,Bin Zhao,"Acute ischemic stroke (AIS) has been a common threat to human health and may lead to severe outcomes without proper and prompt treatment. To precisely diagnose AIS, it is of paramount importance to quantitatively evaluate the AIS lesions. By adopting a convolutional neural network (CNN), many automatic methods for ischemic stroke lesion segmentation on magnetic resonance imaging (MRI) have been proposed. However, most CNN-based methods should be trained on a large amount of fully labeled subjects, and the label annotation is a labor-intensive and time-consuming task. Therefore, in this paper, we propose to use a mixture of many weakly labeled and a few fully labeled subjects to relieve the thirst of fully labeled subjects. In particular, a multifeature map fusion network (MFMF-Network) with two branches is proposed, where hundreds of weakly labeled subjects are used to train the classification branch, and several fully labeled subjects are adopted to tune the segmentation branch. By training on 398 weakly labeled and 5 fully labeled subjects, the proposed method is able to achieve a mean dice coefficient of 0.699 ± 0.128 on a test set with 179 subjects. The lesion-wise and subject-wise metrics are also evaluated, where a lesion-wise F1 score of 0.886 and a subject-wise detection rate of 1 are achieved.","{'model': 'tldr@v2.0.0', 'text': 'A multifeature map fusion network (MFMF-Network) with two branches is proposed, where hundreds of weakly labeled subjects are used to train the classification branch, and several fully labeling subjects are adopted to tune the segmentation branch.'}",https://downloads.hindawi.com/journals/cmmm/2021/3628179.pdf
-multimodal garment designer: human-centric latent diffusion models for fashion image editing,Alberto Baldrati,"Fashion illustration is used by designers to communicate their vision and to bring the design idea from conceptualization to realization, showing how clothes interact with the human body. In this context, computer vision can thus be used to improve the fashion design process. Differently from previous works that mainly focused on the virtual try-on of garments, we propose the task of multimodal-conditioned fashion image editing, guiding the generation of human-centric fashion images by following multimodal prompts, such as text, human body poses, and garment sketches. We tackle this problem by proposing a new architecture based on latent diffusion models, an approach that has not been used before in the fashion domain. Given the lack of existing datasets suitable for the task, we also extend two existing fashion datasets, namely Dress Code and VITON-HD, with multimodal annotations collected in a semi-automatic manner. Experimental results on these new datasets demonstrate the effectiveness of our proposal, both in terms of realism and coherence with the given multimodal inputs. Source code and collected multimodal annotations are publicly available at: https://github.com/aimagelab/multimodal-garment-designer.","{'model': 'tldr@v2.0.0', 'text': 'The task of multimodal-conditioned fashion image editing is proposed, guiding the generation of human-centric fashion images by following multi-modal prompts, such as text, human body poses, and garment sketches.'}",https://arxiv.org/pdf/2304.02051
-multimodal-driven talking face generation via a unified diffusion-based generator,Chao Xu,"Multimodal-driven talking face generation refers to animating a portrait with the given pose, expression, and gaze transferred from the driving image and video, or estimated from the text and audio. However, existing methods ignore the potential of text modal, and their generators mainly follow the source-oriented feature rearrange paradigm coupled with unstable GAN frameworks. In this work, we first represent the emotion in the text prompt, which could inherit rich semantics from the CLIP, allowing flexible and generalized emotion control. We further reorganize these tasks as the target-oriented texture transfer and adopt the Diffusion Models. More specifically, given a textured face as the source and the rendered face projected from the desired 3DMM coefficients as the target, our proposed Texture-Geometry-aware Diffusion Model decomposes the complex transfer problem into multi-conditional denoising process, where a Texture Attention-based module accurately models the correspondences between appearance and geometry cues contained in source and target conditions, and incorporate extra implicit information for high-fidelity talking face generation. Additionally, TGDM can be gracefully tailored for face swapping. We derive a novel paradigm free of unstable seesaw-style optimization, resulting in simple, stable, and effective training and inference schemes. Extensive experiments demonstrate the superiority of our method.","{'model': 'tldr@v2.0.0', 'text': 'This work derives a novel paradigm free of unstable seesaw-style optimization, resulting in simple, stable, and effective training and inference schemes, and adopts the Diffusion Models.'}",http://arxiv.org/pdf/2305.02594
-d2tv: dual knowledge distillation and target-oriented vision modeling for many-to-many multimodal summarization,Yunlong Liang,"Many-to-many multimodal summarization (M$^3$S) task aims to generate summaries in any language with document inputs in any language and the corresponding image sequence, which essentially comprises multimodal monolingual summarization (MMS) and multimodal cross-lingual summarization (MXLS) tasks. Although much work has been devoted to either MMS or MXLS and has obtained increasing attention in recent years, little research pays attention to the M$^3$S task. Besides, existing studies mainly focus on 1) utilizing MMS to enhance MXLS via knowledge distillation without considering the performance of MMS or 2) improving MMS models by filtering summary-unrelated visual features with implicit learning or explicitly complex training objectives. In this paper, we first introduce a general and practical task, i.e., M$^3$S. Further, we propose a dual knowledge distillation and target-oriented vision modeling framework for the M$^3$S task. Specifically, the dual knowledge distillation method guarantees that the knowledge of MMS and MXLS can be transferred to each other and thus mutually prompt both of them. To offer target-oriented visual features, a simple yet effective target-oriented contrastive objective is designed and responsible for discarding needless visual information. Extensive experiments on the many-to-many setting show the effectiveness of the proposed approach. Additionally, we will contribute a many-to-many multimodal summarization (M$^3$Sum) dataset.","{'model': 'tldr@v2.0.0', 'text': 'The dual knowledge distillation method guarantees that the knowledge of MMS and MXLS can be transferred to each other and thus mutually prompt both of them, and a target-oriented vision modeling framework is proposed for the M$^3$S task.'}",http://arxiv.org/pdf/2305.12767
-beyond text-to-image: multimodal prompts to explore generative ai,Vivian Liu,"Text-to-image AI systems have proven to have extraordinary generative capacities that have facilitated widespread adoption. However, these systems are primarily text-based, which is a fundamental inversion of what many artists are traditionally used to: having full control over the composition of their work. Prior work has shown that there is great utility in using text prompts and that AI augmented workflows can increase momentum on creative tasks for end users. However, multimodal interactions beyond text need to be further defined, so end users can have rich points of interaction that allow them to truly co-pilot AI-generated content creation. To this end, the goal of my research is to equip creators with workflows that 1) translate abstract design goals into prompts of visual language, 2) structure exploration of design outcomes, and 3) integrate creator contributions into generations.","{'model': 'tldr@v2.0.0', 'text': 'The goal of this research is to equip creators with workflows that translate abstract design goals into prompts of visual language, structure exploration of design outcomes, and integrate creator contributions into generations.'}",
-mass-producing failures of multimodal systems with language models,Shengbang Tong,"Deployed multimodal systems can fail in ways that evaluators did not anticipate. In order to find these failures before deployment, we introduce MultiMon, a system that automatically identifies systematic failures -- generalizable, natural-language descriptions of patterns of model failures. To uncover systematic failures, MultiMon scrapes a corpus for examples of erroneous agreement: inputs that produce the same output, but should not. It then prompts a language model (e.g., GPT-4) to find systematic patterns of failure and describe them in natural language. We use MultiMon to find 14 systematic failures (e.g.,""ignores quantifiers"") of the CLIP text-encoder, each comprising hundreds of distinct inputs (e.g.,""a shelf with a few/many books""). Because CLIP is the backbone for most state-of-the-art multimodal systems, these inputs produce failures in Midjourney 5.1, DALL-E, VideoFusion, and others. MultiMon can also steer towards failures relevant to specific use cases, such as self-driving cars. We see MultiMon as a step towards evaluation that autonomously explores the long tail of potential system failures. Code for MULTIMON is available at https://github.com/tsb0601/MultiMon.","{'model': 'tldr@v2.0.0', 'text': 'MultiMon is introduced, a system that automatically identifies systematic failures -- generalizable, natural-language descriptions of patterns of model failures -- and is seen as a step towards evaluation that autonomously explores the long tail of potential system failures.'}",http://arxiv.org/pdf/2306.12105
-open visual knowledge extraction via relation-oriented multimodality model prompting,Hejie Cui,"Images contain rich relational knowledge that can help machines understand the world. Existing methods on visual knowledge extraction often rely on the pre-defined format (e.g., sub-verb-obj tuples) or vocabulary (e.g., relation types), restricting the expressiveness of the extracted knowledge. In this work, we take a first exploration to a new paradigm of open visual knowledge extraction. To achieve this, we present OpenVik which consists of an open relational region detector to detect regions potentially containing relational knowledge and a visual knowledge generator that generates format-free knowledge by prompting the large multimodality model with the detected region of interest. We also explore two data enhancement techniques for diversifying the generated format-free visual knowledge. Extensive knowledge quality evaluations highlight the correctness and uniqueness of the extracted open visual knowledge by OpenVik. Moreover, integrating our extracted knowledge across various visual reasoning applications shows consistent improvements, indicating the real-world applicability of OpenVik.","{'model': 'tldr@v2.0.0', 'text': 'This work presents OpenVik, a first exploration to a new paradigm of open visual knowledge extraction which consists of an open relational region detector to detect regions potentially containing relational knowledge and a visual knowledge generator that generates format-free knowledge by prompting the large multimodality model with the detected region of interest.'}",
-a prompt-based multimodal tabular transformer encoder for medical intervention duration estimation,Y. Ruan,"Objective: This study focuses on estimating the duration of medical interventions using electronic health records (EHRs) in clinical decision support. Most existing models were designed for structured tabular data only and often suffer from data corruption problem. The unstructured clinical free-text data that provides valuable insights and is more resistant to data corruption is often overlooked. The objective of this research is to develop a multimodal deep learning framework that integrates different data modalities from EHRs, thereby fully utilizing the predictive capability of EHRs for medical intervention estimation. Materials and Methods: A novel prompt-based tabular transformer encoder framework is proposed for medical intervention duration estimation based on multimodal EHR data. The framework leverages a pre-trained sentence encoder with medical prompts to harmonize language representations of various clinical data modalities, which a tabular transformer encoder is developed to further explore. Results: The developed model demonstrates superior performance compared to the baselines in two EHR datasets. Furthermore, the model exhibits resilience to data corruption in EHRs, with the RMSE curve increasing gradually with higher corruption rates. Discussion: Other than the predictive effectiveness and robustness of the proposed framework, the ablation study highlights the significance of critical components, such as medical prompts, free-text information, and the pre-trained sentence encoder, all contributing to the model's predictive ability. Conclusion: This research presents a promising pathway to enhance medical intervention estimation by incorporating diverse data modalities from language perspective, ultimately bolstering the reliability of deep learning models in clinical care.","{'model': 'tldr@v2.0.0', 'text': 'A novel prompt-based tabular transformer encoder framework is proposed for medical intervention duration estimation based on multimodal EHR data that leverages a pre-trained sentence encoder with medical prompts to harmonize language representations of various clinical data modalities and demonstrates superior performance compared to the baselines in two EHR datasets.'}",
-stone needle: a general multimodal large-scale model framework towards healthcare,Weihua Liu,"In healthcare, multimodal data is prevalent and requires to be comprehensively analyzed before diagnostic decisions, including medical images, clinical reports, etc. However, current large-scale artificial intelligence models predominantly focus on single-modal cognitive abilities and neglect the integration of multiple modalities. Therefore, we propose Stone Needle, a general multimodal large-scale model framework tailored explicitly for healthcare applications. Stone Needle serves as a comprehensive medical multimodal model foundation, integrating various modalities such as text, images, videos, and audio to surpass the limitations of single-modal systems. Through the framework components of intent analysis, medical foundation models, prompt manager, and medical language module, our architecture can perform multi-modal interaction in multiple rounds of dialogue. Our method is a general multimodal large-scale model framework, integrating diverse modalities and allowing us to tailor for specific tasks. The experimental results demonstrate the superior performance of our method compared to single-modal systems. The fusion of different modalities and the ability to process complex medical information in Stone Needle benefits accurate diagnosis, treatment recommendations, and patient care.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes Stone Needle, a general multimodal large-scale model framework tailored explicitly for healthcare applications, integrating diverse modalities and allowing it to tailor for specific tasks.'}",http://arxiv.org/pdf/2306.16034
-abhipaw@dravidianlangtech: multimodal abusive language detection and sentiment analysis,Abhinaba Bala,"Detecting abusive language in multimodal videos has become a pressing need in ensuring a safe and inclusive online environment. This paper focuses on addressing this challenge through the development of a novel approach for multimodal abusive language detection in Tamil videos and sentiment analysis for Tamil/Malayalam videos. By leveraging state-of-the-art models such as Multiscale Vision Transformers (MViT) for video analysis, OpenL3 for audio analysis, and the bert-base-multilingual-cased model for textual analysis, our proposed framework integrates visual, auditory, and textual features. Through extensive experiments and evaluations, we demonstrate the effectiveness of our model in accurately detecting abusive content and predicting sentiment categories. The limited availability of effective tools for performing these tasks in Dravidian Languages has prompted a new avenue of research in these domains.","{'model': 'tldr@v2.0.0', 'text': 'A novel approach for multimodal abusive language detection in Tamil videos and sentiment analysis for Tamil/Malayalam videos is developed by leveraging state-of-the-art models such as Multiscale Vision Transformers for video analysis, OpenL3 for audio analysis, and the bert-base-multilingual-cased model for textual analysis.'}",
-a multimodal prototypical approach for unsupervised sound classification,Saksham Singh Kushwaha,"In the context of environmental sound classification, the adaptability of systems is key: which sound classes are interesting depends on the context and the user's needs. Recent advances in text-to-audio retrieval allow for zero-shot audio classification, but performance compared to supervised models remains limited. This work proposes a multimodal prototypical approach that exploits local audio-text embeddings to provide more relevant answers to audio queries, augmenting the adaptability of sound detection in the wild. We do this by first using text to query a nearby community of audio embeddings that best characterize each query sound, and select the group's centroids as our prototypes. Second, we compare unseen audio to these prototypes for classification. We perform multiple ablation studies to understand the impact of the embedding models and prompts. Our unsupervised approach improves upon the zero-shot state-of-the-art in three sound recognition benchmarks by an average of 12%.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a multimodal prototypical approach that exploits local audio-text embeddings to provide more relevant answers to audio queries, augmenting the adaptability of sound detection in the wild.'}",https://arxiv.org/pdf/2306.12300
-factify3m: a benchmark for multimodal fact verification with explainability through 5w question-answering,Megha Chakraborty,"Combating disinformation is one of the burning societal crises -- about 67% of the American population believes that disinformation produces a lot of uncertainty, and 10% of them knowingly propagate disinformation. Evidence shows that disinformation can manipulate democratic processes and public opinion, causing disruption in the share market, panic and anxiety in society, and even death during crises. Therefore, disinformation should be identified promptly and, if possible, mitigated. With approximately 3.2 billion images and 720,000 hours of video shared online daily on social media platforms, scalable detection of multimodal disinformation requires efficient fact verification. Despite progress in automatic text-based fact verification (e.g., FEVER, LIAR), the research community lacks substantial effort in multimodal fact verification. To address this gap, we introduce FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering. Salient features of the dataset include: (i) textual claims, (ii) ChatGPT-generated paraphrased claims, (iii) associated images, (iv) stable diffusion-generated additional images (i.e., visual paraphrases), (v) pixel-level image heatmap to foster image-text explainability of the claim, (vi) 5W QA pairs, and (vii) adversarial fake news stories.","{'model': 'tldr@v2.0.0', 'text': 'FACTIFY 3M, a dataset of 3 million samples that pushes the boundaries of the domain of fact verification via a multimodal fake news dataset, in addition to offering explainability through the concept of 5W question-answering, is introduced.'}",http://arxiv.org/pdf/2306.05523
-videoadviser: video knowledge distillation for multimodal transfer learning,Yanan Wang,"Multimodal transfer learning aims to transform pretrained representations of diverse modalities into a common domain space for effective multimodal fusion. However, conventional systems are typically built on the assumption that all modalities exist, and the lack of modalities always leads to poor inference performance. Furthermore, extracting pretrained embeddings for all modalities is computationally inefficient for inference. In this work, to achieve high efficiency-performance multimodal transfer learning, we propose VideoAdviser, a video knowledge distillation method to transfer multimodal knowledge of video-enhanced prompts from a multimodal fundamental model (teacher) to a specific modal fundamental model (student). With an intuition that the best learning performance comes with professional advisers and smart students, we use a CLIP-based teacher model to provide expressive multimodal knowledge supervision signals to a RoBERTa-based student model via optimizing a step-distillation objective loss—first step: the teacher distills multimodal knowledge of video-enhanced prompts from classification logits to a regression logit—second step: the multimodal knowledge is distilled from the regression logit of the teacher to the student. We evaluate our method in two challenging multimodal tasks: video-level sentiment analysis (MOSI and MOSEI datasets) and audio-visual retrieval (VEGAS dataset). The student (requiring only the text modality as input) achieves an MAE score improvement of up to 12.3% for MOSI and MOSEI. Our method further enhances the state-of-the-art method by 3.4% mAP score for VEGAS without additional computations for inference. These results suggest the strengths of our method for achieving high efficiency-performance multimodal transfer learning.","{'model': 'tldr@v2.0.0', 'text': 'VideoAdviser is proposed, a video knowledge distillation method to transfer multimodal knowledge of video-enhanced prompts from a multi-modalities model (teacher) to a specific modal fundamental model (student) via optimizing a step-distillation objective loss.'}",https://ieeexplore.ieee.org/ielx7/6287639/10005208/10136716.pdf
-multimodal prompt learning in emotion recognition using context and audio information,Eunseo Jeong,"Prompt learning has improved the performance of language models by reducing the gap in language model training methods of pre-training and downstream tasks. However, extending prompt learning in language models pre-trained with unimodal data to multimodal sources is difficult as it requires additional deep-learning layers that cannot be attached. In the natural-language emotion-recognition task, improved emotional classification can be expected when using audio and text to train a model rather than only natural-language text. Audio information, such as voice pitch, tone, and intonation, can give more information that is unavailable in text to predict emotions more effectively. Thus, using both audio and text can enable better emotion prediction in speech emotion-recognition models compared to semantic information alone. In this paper, in contrast to existing studies that use multimodal data with an additional layer, we propose a method for improving the performance of speech emotion recognition using multimodal prompt learning with text-based pre-trained models. The proposed method is using text and audio information in prompt learning by employing a language model pre-trained on natural-language text. In addition, we propose a method to improve the emotion-recognition performance of the current utterance using the emotion and contextual information of the previous utterances for prompt learning in speech emotion-recognition tasks. The performance of the proposed method was evaluated using the English multimodal dataset MELD and the Korean multimodal dataset KEMDy20. Experiments using both the proposed methods obtained an accuracy of 87.49%, F1 score of 44.16, and weighted F1 score of 86.28.","{'model': 'tldr@v2.0.0', 'text': 'A method for improving the performance of speech emotion recognition using multimodal prompt learning with text-based pre-trained models and a method to improve the emotion- Recognition performance of the current utterance using the emotion and contextual information of the previous utterances for prompt learning in speech emotion-recognition tasks are proposed.'}",https://www.mdpi.com/2227-7390/11/13/2908/pdf?version=1688017556
-"telehealthcare and covid-19: a noninvasive & low cost invasive, scalable and multimodal real-time smartphone application for early diagnosis of sars-cov-2 infection",A. Shams,"The global coronavirus pandemic overwhelmed many health care systems, enforcing lockdown and encouraged work from home to control the spread of the virus and prevent overrunning of hospitalized patients. This prompted a sharp widespread use of telehealth to provide low-risk care for patients. Nevertheless, a continuous mutation into new variants and widespread unavailability of test kits, especially in developing countries, possess the challenge to control future potential waves of infection. In this paper, we propose a novel Smartphone application-based platform for early diagnosis of possible Covid-19 infected patients. The application provides three modes of diagnosis from possible symptoms, cough sound, and specific blood biomarkers. When a user chooses a particular setting and provides the necessary information, it sends the data to a trained machine learning (ML) model deployed in a remote server using the internet. The ML algorithm then predicts the possibility of contracting Covid-19 and sends the feedback to the user. The entire procedure takes place in real-time. Our machine learning models can identify Covid-19 patients with an accuracy of 100%, 95.65%, and 77.59% from blood parameters, cough sound, and symptoms respectively. Moreover, the ML sensitivity for blood and sound is 100%, which indicates correct identification of Covid positive patients. This is significant in limiting the spread of the virus. The multimodality offers multiplex diagnostic methods to better classify possible infectees and together with the instantaneous nature of our technique, demonstrates the power of telehealthcare as an easy and widespread low-cost scalable diagnostic solution for future pandemics.","{'model': 'tldr@v2.0.0', 'text': 'A novel Smartphone application-based platform for early diagnosis of possible Covid-19 infected patients that offers multiplex diagnostic methods to better classify possible infectees and demonstrates the power of telehealthcare as an easy and widespread low-cost scalable diagnostic solution for future pandemics.'}",
-reversing rivaroxaban anticoagulation as part of a multimodal hemostatic intervention in a polytrauma animal model,Farahnaz Rayatdoost,"Background: Life-threatening bleeding requires prompt reversal of the anticoagulant effects of factor Xa inhibitors. This study investigated the effectiveness of four-factor prothrombin complex concentrate in treating trauma-related hemorrhage with rivaroxaban-anticoagulation in a pig polytrauma model. This study also tested the hypothesis that the combined use of a low dose of prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate could improve its subtherapeutic effects. Methods: Trauma (blunt liver injury and bilateral femur fractures) was induced in 48 anesthetized male pigs after 30 min of rivaroxaban infusion (1 mg/kg). Animals in the first part of the study received prothrombin complex concentrate (12.5, 25, and 50 U/kg). In the second part, animals were treated with 12.5 U/kg prothrombin complex concentrate plus tranexamic acid or plus tranexamic acid and fibrinogen concentrate. The primary endpoint was total blood loss postinjury. The secondary endpoints (panel of coagulation parameters and thrombin generation) were monitored for 240 min posttrauma or until death. Results: The first part of the study showed that blood loss was significantly lower in the 25 U/kg prothrombin complex concentrate (1,541 ± 269 ml) and 50 U/kg prothrombin complex concentrate (1,464 ± 108 ml) compared with control (3,313 ± 634 ml), and 12.5 U/kg prothrombin complex concentrate (2,671 ± 334 ml, all P < 0.0001). In the second part of the study, blood loss was significantly less in the 12.5 U/kg prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate (1,836 ± 556 ml, P < 0.001) compared with 12.5 U/kg prothrombin complex concentrate plus tranexamic acid (2,910 ± 856 ml), and there were no early deaths in the 25 U/kg prothrombin complex concentrate, 50 U/kg prothrombin complex concentrate, and 12.5 U/kg prothrombin complex concentrate plus tranexamic acid and fibrinogen concentrate groups. Histopathologic analyses postmortem showed no adverse events. Conclusions: Prothrombin complex concentrate effectively reduced blood loss, restored hemostasis, and balanced thrombin generation. A multimodal hemostatic approach using tranexamic acid plus fibrinogen concentrate enhanced the effect of low doses of prothrombin complex concentrate, potentially reducing the prothrombin complex concentrate doses required for effective bleeding control. In an animal model of rivaroxaban-treated pigs that underwent complex traumatic injury, prothrombin complex concentrates alone and in combination with tranexamic acid and fibrinogen concentrate effectively reduced blood loss, restored hemostasis, and improved thrombin generation. Supplemental Digital Content is available in the text.","{'model': 'tldr@v2.0.0', 'text': 'Prothrombin complex concentrate effectively reduced blood loss, restored hemostasis, and balanced thrombin generation, and a multimodal hemostatic approach using tranexamic acid plus fibrinogen concentrate enhanced the effect of low doses of prothromin complex concentrate.'}",https://pubs.asahq.org/anesthesiology/article-pdf/135/4/673/524144/20211000.0-00027.pdf
-mpt: multimodal prompt tuning for event detection,,"Event Detection is a key and challenging sub001 task of event extraction, which has serious trig002 ger word ambiguity. Existing studies mainly 003 focus on contextual information in text, while 004 there are naturally many images in news ar005 ticles that need to be explored. We believe 006 that images not only reflect the core events of 007 the text but also help to trigger word disam008 biguation. In this paper, we propose a new 009 bi-recursive multimodal Prompt Tuning (MPT) 010 model for deep interaction between images and 011 sentences to achieve aggregation of modal fea012 tures. MPT uses pre-trained CLIP to encode 013 and map sentences and images into the same 014 multimodal semantic space and uses alternat015 ing dual attention to select information features 016 for mutual enhancement. Then, a soft prompt 017 method of multimodal guidance is proposed, 018 and the multimodal information obtained by 019 fusion is used to guide the downstream event 020 detection task. Our superior performance com021 pared to six state-of-the-art baselines and fur022 ther ablation studies, demonstrate the impor023 tance of image modality and the effectiveness 024 of the proposed architecture. 025","{'model': 'tldr@v2.0.0', 'text': 'A new 009 bi-recursive multimodal Prompt Tuning (MPT) model for deep interaction between images and 011 sentences to achieve aggregation of modal fea012 tures and demonstrates the impor023 tance of image modality and the effectiveness of the proposed architecture.'}",
-what matters in training a gpt4-style language model with multimodal inputs?,Yan Zeng,"Recent advancements in Large Language Models (LLMs) such as GPT4 have displayed exceptional multi-modal capabilities in following open-ended instructions given images. However, the performance of these models heavily relies on design choices such as network structures, training data, and training strategies, and these choices have not been extensively discussed in the literature, making it difficult to quantify progress in this field. To address this issue, this paper presents a systematic and comprehensive study, quantitatively and qualitatively, on training such models. We implement over 20 variants with controlled settings. Concretely, for network structures, we compare different LLM backbones and model designs. For training data, we investigate the impact of data and sampling strategies. For instructions, we explore the influence of diversified prompts on the instruction-following ability of the trained models. For benchmarks, we contribute the first, to our best knowledge, comprehensive evaluation set including both image and video tasks through crowd-sourcing. Based on our findings, we present Lynx, which performs the most accurate multi-modal understanding while keeping the best multi-modal generation ability compared to existing open-sourced GPT4-style models.","{'model': 'tldr@v2.0.0', 'text': 'Lynx is presented, which performs the most accurate multi-modal understanding while keeping the best multi- modal generation ability compared to existing open-sourced GPT4-style models.'}",https://arxiv.org/pdf/2307.02469
-multimodal prompts effectively elicit robot-initiated social touch interactions,Spatika Sampath Gujran,"Social touch plays an important role in building interpersonal relationships and might therefore also facilitate interactions with social robots. As people tend to have less experience interacting with social robots compared to with humans, especially with interactions involving social touch, more explicit communication might be necessary to disambiguate social intentions. In the experiment, participants engaged in an informal conversation with humanoid robot Pepper. Throughout the interaction, Pepper initiated various social touch interactions such as a handshake during introductions and a hug to say goodbye by using either a unimodal prompt (control condition: movement cue only) or a multimodal prompt (experimental condition: movement and verbal cue). The results show that the multimodal prompts significantly increased the number of successfully elicited social touch interactions. No significant differences in the self-reported perception of the robot were found between condition. Our results help to inform the design of robots that are intended to engage in social touch interactions.",,https://dl.acm.org/doi/pdf/10.1145/3610661.3617642
-multimodal prompt transformer with hybrid contrastive learning for emotion recognition in conversation,Shihao Zou,"Emotion Recognition in Conversation (ERC) plays an important role in driving the development of human-machine interaction. Emotions can exist in multiple modalities, and multimodal ERC mainly faces two problems: (1) the noise problem in the cross-modal information fusion process, and (2) the prediction problem of less sample emotion labels that are semantically similar but different categories. To address these issues and fully utilize the features of each modality, we adopted the following strategies: first, deep emotion cues extraction was performed on modalities with strong representation ability, and feature filters were designed as multimodal prompt information for modalities with weak representation ability. Then, we designed a Multimodal Prompt Transformer (MPT) to perform cross-modal information fusion. MPT embeds multimodal fusion information into each attention layer of the Transformer, allowing prompt information to participate in encoding textual features and being fused with multi-level textual information to obtain better multimodal fusion features. Finally, we used the Hybrid Contrastive Learning (HCL) strategy to optimize the model's ability to handle labels with few samples. This strategy uses unsupervised contrastive learning to improve the representation ability of multimodal fusion and supervised contrastive learning to mine the information of labels with few samples. Experimental results show that our proposed model outperforms state-of-the-art models in ERC on two benchmark datasets.","{'model': 'tldr@v2.0.0', 'text': ""The Hybrid Contrastive Learning (HCL) strategy is used to optimize the model's ability to handle labels with few samples, and shows that the proposed model outperforms state-of-the-art models in ERC on two benchmark datasets.""}",https://dl.acm.org/doi/pdf/10.1145/3581783.3611805
-towards interpretable multimodal predictive models for early mortality prediction of hemorrhagic stroke patients.,F. B. Emdad,"The increasing death rate over the past eight years due to stroke has prompted clinicians to look for data-driven decision aids. Recently, deep-learning-based prediction models trained with fine-grained electronic health record (EHR) data have shown superior promise for health outcome prediction. However, the use of EHR-based deep learning models for hemorrhagic stroke outcome prediction has not been extensively explored. This paper proposes an ensemble deep learning framework to predict early mortality among ICU patients with hemorrhagic stroke. The proposed ensemble model achieved an accuracy of 83%, which was higher than the fusion model and other baseline models (logistic regression, decision tree, random forest, and XGBoost). Moreover, we used SHAP values for interpretation of the ensemble model to identify important features for the prediction. In addition, this paper follows the MINIMAR (MINimum Information for Medical AI Reporting) standard, presenting an important step towards building trust among the AI system and clinicians.","{'model': 'tldr@v2.0.0', 'text': 'An ensemble deep learning framework to predict early mortality among ICU patients with hemorrhagic stroke is proposed, and follows the MINIMAR (MINimum Information for Medical AI Reporting) standard, presenting an important step towards building trust among the AI system and clinicians.'}",
-translating and segmenting multimodal medical volumes with cycle- and shape-consistency generative adversarial network,Zizhao Zhang,"Synthesized medical images have several important applications, e.g., as an intermedium in cross-modality image registration and as supplementary training samples to boost the generalization capability of a classifier. Especially, synthesized computed tomography (CT) data can provide X-ray attenuation map for radiation therapy planning. In this work, we propose a generic cross-modality synthesis approach with the following targets: 1) synthesizing realistic looking 3D images using unpaired training data, 2) ensuring consistent anatomical structures, which could be changed by geometric distortion in cross-modality synthesis and 3) improving volume segmentation by using synthetic data for modalities with limited training samples. We show that these goals can be achieved with an end-to-end 3D convolutional neural network (CNN) composed of mutually-beneficial generators and segmentors for image synthesis and segmentation tasks. The generators are trained with an adversarial loss, a cycle-consistency loss, and also a shape-consistency loss, which is supervised by segmentors, to reduce the geometric distortion. From the segmentation view, the segmentors are boosted by synthetic data from generators in an online manner. Generators and segmentors prompt each other alternatively in an end-to-end training fashion. With extensive experiments on a dataset including a total of 4,496 CT and magnetic resonance imaging (MRI) cardiovascular volumes, we show both tasks are beneficial to each other and coupling these two tasks results in better performance than solving them exclusively.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a generic cross-modality synthesis approach and shows that these goals can be achieved with an end-to-end 3D convolutional neural network (CNN) composed of mutually-beneficial generators and segmentors for image synthesis and segmentation tasks.'}",https://arxiv.org/pdf/1802.09655
-"multilingual, multimodal, and multidisciplinary: deaf students and translanguaging in content area classes",Jessica A. Scott,"Translanguaging is an approach to the education of multilingual students that has been gaining in popularity in recent years. Translanguaging is defined as the use of multiple languages and modalities during classroom instruction with the express purpose of encouraging and teaching students to use all their linguistic resources to make sense of content and demonstrate their understanding. Though there has been a great deal of research on translanguaging in general, this article focuses on two less-studied areas for translanguaging: the intersection of translanguaging in deaf education and content area education. Here, the authors explore the literature available in the domains of translanguaging and science, translanguaging and deaf students, and translanguaging in science classrooms with deaf students. Using the findings from these articles, the authors present four recommendations for teachers who want to engage in translanguaging in the content areas with deaf students: First, teachers can use both signed language and English (or other written languages) during instruction to expose students to essential concepts in multiple languages. Next, teachers can expose students to discipline and subject-specific written and signed words, as well as colloquial meanings of similar written and signed words (e.g., evidence versus fact versus proof in American Sign Language). Additionally, teachers should attempt to be visually oriented while translanguaging. Finally, we recommend that teachers encourage students to respond to questions or prompts using multiple languages and/or modalities. Implications of each recommendation are explored.",,https://www.mdpi.com/2226-471X/8/1/55/pdf?version=1676346473
-integrating multimodal information in machine learning for classifying acute myocardial infarction,R. Xiao,"Objective. Prompt identification and recognization of myocardial ischemia/infarction (MI) is the most important goal in the management of acute coronary syndrome. The 12-lead electrocardiogram (ECG) is widely used as the initial screening tool for patients with chest pain but its diagnostic accuracy remains limited. There is early evidence that machine learning (ML) algorithms applied to ECG waveforms can improve performance. Most studies are designed to classify MI from healthy controls and thus are limited due to the lack of consideration of ECG abnormalities from other cardiac conditions, leading to false positives. Moreover, clinical information beyond ECG has not yet been well leveraged in existing ML models. Approach. The present study considered downstream clinical implementation scenarios in the initial model design by dichotomizing study recordings from a public large-scale ECG dataset into a MI class and a non-MI class with the inclusion of MI-confounding conditions. Two experiments were conducted to systematically investigate the impact of two important factors entrained in the modeling process, including the duration of ECG, and the value of multimodal information for model training. A novel multimodal deep learning architecture was proposed to learn joint features from both ECG and patient demographics. Main results. The multimodal model achieved better performance than the ECG-only model, with a mean area under the receiver operating characteristic curve of 92.1% and a mean accuracy of 87.4%, which is on par with existing studies despite the increased task difficulty due to the new class definition. By investigation of model explainability, it revealed the contribution of patient information in model performance and clinical concordance of the model’s attention with existing clinical insights. Significance. The findings in this study help guide the development of ML solutions for prompt MI detection and move the models one step closer to real-world clinical applications.","{'model': 'tldr@v2.0.0', 'text': 'The investigation of model explainability revealed the contribution of patient information in model performance and clinical concordance of the model’s attention with existing clinical insights and helped guide the development of ML solutions for prompt MI detection and move the models one step closer to real-world clinical applications.'}",https://iopscience.iop.org/article/10.1088/1361-6579/acc77f/pdf
-multimodal parameter-efficient few-shot class incremental learning,Marco D’Alessandro,"Few-Shot Class Incremental Learning (FSCIL) is a challenging continual learning task, where limited training examples are available during several learning sessions. To succeed in this task, it is necessary to avoid over-fitting new classes caused by biased distributions in the few-shot training sets. The general approach to address this issue involves enhancing the representational capability of a pre-defined backbone architecture by adding special modules for backward compatibility with older classes. However, this approach has not yet solved the dilemma of ensuring high classification accuracy over time while reducing the gap between the performance obtained on larger training sets and the smaller ones. In this work, we propose an alternative approach called Continual Parameter-Efficient CLIP (CPE-CLIP) to reduce the loss of information between different learning sessions. Instead of adapting additional modules to address information loss, we leverage the vast knowledge acquired by CLIP in large-scale pre-training and its effectiveness in generalizing to new concepts. Our approach is multimodal and parameter-efficient, relying on learnable prompts for both the language and vision encoders to enable transfer learning across sessions. We also introduce prompt regularization to improve performance and prevent forgetting. Our experimental results demonstrate that CPE-CLIP significantly improves FSCIL performance compared to state-of-the-art proposals while also drastically reducing the number of learnable parameters and training costs.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes an alternative approach called Continual Parameter-Efficient CLIP (CPE-CLIP) to reduce the loss of information between different learning sessions, and leverages the vast knowledge acquired by CLIP in large-scale pre-training and its effectiveness in generalizing to new concepts.'}",http://arxiv.org/pdf/2303.04751
-multimodal data value chain (m-dvc): a conceptual tool to support the development of multimodal learning analytics solutions,Shashi Kant Shankar,"Multimodal Learning Analytics (MMLA) systems, understood as those that exploit multimodal evidence of learning to better model a learning situation, have not yet spread widely in educational practice. Their inherent technical complexity, and the lack of educational stakeholder involvement in their design, are among the hypothesized reasons for the slow uptake of this emergent field. To aid in the process of stakeholder communication and systematization leading to the specification of MMLA systems, this paper proposes a Multimodal Data Value Chain (M-DVC). This conceptual tool, derived from both the field of Big Data and the needs of MMLA scenarios, has been evaluated in terms of its usefulness for stakeholders, in three authentic case studies of MMLA systems currently under development. The results of our mixed-methods evaluation highlight the usefulness of the M-DVC to elicit unspoken assumptions or unclear data processing steps in the initial stages of development. The evaluation also revealed limitations of the M-DVC in terms of the technical terminology employed, and the need for more detailed contextual information to be included. These limitations also prompt potential improvements for the M-DVC, on the path towards clearer specification and communication within the multi-disciplinary teams needed to build educationally-meaningful MMLA solutions.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Multimodal Data Value Chain (M-DVC), a conceptual tool derived from both the field of Big Data and the needs of MMLA scenarios, that has been evaluated in terms of its usefulness for stakeholders, in three authentic case studies of M MLA systems currently under development.'}",https://uvadoc.uva.es/bitstream/10324/43245/1/Multimodal-2019_KantShankar_IEEERita.pdf
-retrieving multimodal prompts for generative visual question answering,Timothy Ossowski,,,https://aclanthology.org/2023.findings-acl.158.pdf
-seeing the unseen : euphemism in animated films : a multimodal and critical discourse analysis,Dalia Asseel,"Animated films are contemporary popular cultural products recreating the ‘real’ world and engaging massive worldwide audiences of adults and children. Children as the ostensible viewers of animated films may acquire their cultural and ideological knowledge and beliefs about the world from the representations in animated films. Although during the past decade animated films have increasingly been the focus of attention of researchers across different disciplines, including education, gender, sexuality and literacy, studies tackling the discourse and language of animated films are still in their early stages. More specifically, very few studies have investigated the use of euphemism as a major micro-level linguistic device reflecting macro-level discourse and extending to sociocultural structures. To this end, this thesis examines euphemism constructed through the discourse of animated films by employing the strategies of Critical Discourse Studies (CDS). Moreover, Multimodal Discourse Analysis (MDA) is employed to examine discursive strategies involving visual representations accompanying euphemism and what underpins those strategies, and to shed light on the multimodal relations between the representation of both. Euphemism is frequently associated with the notion of taboo. Consequently, new words or phrases are designated to refer to linguistic taboos as alternatives used by speakers to minimise the threat to the audience’s face as well as to their own. In addition, euphemistic occurrences represent a self-interested version of reality by pushing a topic into the background and highlighting instead a particular view of a topic. Therefore, euphemism is a speaker-oriented tool implying the reaction the speaker intends to prompt in the audience. This study shows that euphemism as a discursive linguistic tool has been used extensively in animated films as a manifestation of the discursive role anthropomorphised characters play to transmit certain ideological and social representations. A data set comprising 176 euphemisms found in four full-length anthropomorphised animated featured films, AAFF, extracted from film scripts and online channels was collected. The study identifies the main types of euphemism used in films, drawing on a framework of types based on Warren (1992), Allan and Burridge (1991) and Crespo (2006). After the main types of euphemism have been identified, the data set is approached from the perspective of taboo and culturally repressed topics, such as sexuality, gender and race. Moreover, discursive strategies adapted from Reisigl and Wodak (2016) are applied in order to identify important categories for the analysis of euphemistic discourse. This approach forms the basis for an in-depth, qualitative analysis of several representative scenes extracted from the films under investigation. First, the analysis focuses on racial euphemisms targeting different races. Then, I analyse sexual euphemisms related to nudity and sexual body parts, as well as conceptual metaphoric representations of sexual euphemisms. Finally, I focus on gendered euphemisms targeting the representation of four female characters in the films. The analysis shows that animated films tend to use euphemism to camouflage various taboo areas and manipulate the viewers’ perceptions, such as those relating to sexuality, gender and race. Nomination strategies are used more frequently with euphemisms referring to race, sex and gender. Metaphors and conceptual metaphors are used more frequently with sexual euphemisms. Visual empowerment strategies of female characters to highlight their sexuality are used more frequently with gendered euphemisms. I argue that while the linguistic element attenuates an ideology by virtue of euphemism’s manipulative nature, the visual element, in turn, highlights and confirms the same ideology, values or stereotypes. Hence, animated films can articulate the ideological and social legitimation or normalisation of a particular view of race, sex or gender through the use of euphemism and visual discursive strategies.",,
-travellers' perception of worship facilities for multimodal users of mrt sbk line,Nur Athifah A. Kadir,"The complete opening of the new Mass Rapid Transit System (MRT) Sungai Buloh - Kajang (SBK) of Klang Valley, Malaysia since July 2017 has been supported positively by transit researchers and public transport advocators alike. Multimodal public transport users make trips, usually involving transferring at interchanges. These interchanges are provided with several amenities facilitating each trip leg. This research recorded the perception of users of worship facilities provided at an interchanging node where several rails, bus, taxi, and other paratransit systems meet. This new rail system intersects with the other systems at Muzium Negara station. Using intercept off-board face to face questionnaire survey method, perception of the quality of worship facilities was captured between May and August 2019. Several explanatory variables such as socio economic and trip characteristics were set against the perception to identify factors influencing the perception. Results of the correlational analysis showed that both socio-demographic and trip characteristics influenced perception to various statistical significance degrees. Findings suggested that worship facilities were most utilised by passengers of MRT who relatively lower income, engaged in the public, specifically educational sector. Usage of MRT for commuting within the Klang Valley on monthly basis. The improvements in ventilation systems, prayer hall size and design, floor cleanliness, lighting and mirror design as well as shoe rack/clean area design were most sought by these passengers. As such, these should be the focus of strategies to be promptly adopted by interchanging nodes building managers to increase the quality of worship places provision.",,https://www.planningmalaysia.org/index.php/pmj/article/download/834/620
-a survey on psycho-physiological analysis & measurement methods in multimodal systems,M. Z. Baig,"Psycho-physiological analysis has gained greater attention in the last few decades in various fields including multimodal systems. Researchers use psychophysiological feedback devices such as skin conductance (SC), Electroencephalography (EEG) and Electrocardiography (ECG) to detect the affective states of the users during task performance. Psycho-physiological feedback has been successful in detection of the cognitive states of users in human-computer interaction (HCI). Recently, in game studies, psycho-physiological feedback has been used to capture the user experience and the effect of interaction on human psychology. This paper reviews several psycho-physiological, cognitive, and affective assessment studies and focuses on the use of psychophysiological signals in estimating the user’s cognitive and emotional states in multimodal systems. In this paper, we review the measurement techniques and methods that have been used to record psycho-physiological signals as well as the cognitive and emotional states in a variety of conditions. The aim of this review is to conduct a detailed study to identify, describe and analyze the key psycho-physiological parameters that relate to different mental and emotional states in order to provide an insight into key approaches. Furthermore, the advantages and limitations of these approaches are also highlighted in this paper. The findings state that the classification accuracy of >90% has been achieved in classifying emotions with EEG signals. A strong correlation between self-reported data, HCI experience, and psychophysiological data has been observed in a wide range of domains including games, human-robot interaction, mobile interaction, and simulations. An increase in β and γ -band activity have been observed in high intense games and simulations.","{'model': 'tldr@v2.0.0', 'text': 'The aim of this review is to conduct a detailed study to identify, describe and analyze the key psycho-physiological parameters that relate to different mental and emotional states in order to provide an insight into key approaches.'}",https://www.mdpi.com/2414-4088/3/2/37/pdf?version=1561376523
-multimodal sentiment analysis with word-level fusion and reinforcement learning,Minghai Chen,"With the increasing popularity of video sharing websites such as YouTube and Facebook, multimodal sentiment analysis has received increasing attention from the scientific community. Contrary to previous works in multimodal sentiment analysis which focus on holistic information in speech segments such as bag of words representations and average facial expression intensity, we propose a novel deep architecture for multimodal sentiment analysis that is able to perform modality fusion at the word level. In this paper, we propose the Gated Multimodal Embedding LSTM with Temporal Attention (GME-LSTM(A)) model that is composed of 2 modules. The Gated Multimodal Embedding allows us to alleviate the difficulties of fusion when there are noisy modalities. The LSTM with Temporal Attention can perform word level fusion at a finer fusion resolution between the input modalities and attends to the most important time steps. As a result, the GME-LSTM(A) is able to better model the multimodal structure of speech through time and perform better sentiment comprehension. We demonstrate the effectiveness of this approach on the publicly-available Multimodal Corpus of Sentiment Intensity and Subjectivity Analysis (CMU-MOSI) dataset by achieving state-of-the-art sentiment classification and regression results. Qualitative analysis on our model emphasizes the importance of the Temporal Attention Layer in sentiment prediction because the additional acoustic and visual modalities are noisy. We also demonstrate the effectiveness of the Gated Multimodal Embedding in selectively filtering these noisy modalities out. These results and analysis open new areas in the study of sentiment analysis in human communication and provide new models for multimodal fusion.","{'model': 'tldr@v2.0.0', 'text': 'The Gated Multimodal Embedding LSTM with Temporal Attention model is proposed that is composed of 2 modules and able to perform modality fusion at the word level and is able to better model the multimodal structure of speech through time and perform better sentiment comprehension.'}",https://arxiv.org/pdf/1802.00924
-towards a multimodal model of cognitive workload through synchronous optical brain imaging and eye tracking measures,E. Isbilir,"Recent advances in neuroimaging technologies have rendered multimodal analysis of operators’ cognitive processes in complex task settings and environments increasingly more practical. In this exploratory study, we utilized optical brain imaging and mobile eye tracking technologies to investigate the behavioral and neurophysiological differences among expert and novice operators while they operated a human-machine interface in normal and adverse conditions. In congruence with related work, we observed that experts tended to have lower prefrontal oxygenation and exhibit gaze patterns that are better aligned with the optimal task sequence with shorter fixation durations as compared to novices. These trends reached statistical significance only in the adverse condition where the operators were prompted with an unexpected error message. Comparisons between hemodynamic and gaze measures before and after the error message indicated that experts’ neurophysiological response to the error involved a systematic increase in bilateral dorsolateral prefrontal cortex (dlPFC) activity accompanied with an increase in fixation durations, which suggests a shift in their attentional state, possibly from routine process execution to problem detection and resolution. The novices’ response was not as strong as that of experts, including a slight increase only in the left dlPFC with a decreasing trend in fixation durations, which is indicative of visual search behavior for possible cues to make sense of the unanticipated situation. A linear discriminant analysis model capitalizing on the covariance structure among hemodynamic and eye movement measures could distinguish experts from novices with 91% accuracy. Despite the small sample size, the performance of the linear discriminant analysis combining eye fixation and dorsolateral oxygenation measures before and after an unexpected event suggests that multimodal approaches may be fruitful for distinguishing novice and expert performance in similar neuroergonomic applications in the field.","{'model': 'tldr@v2.0.0', 'text': 'Investigation of behavioral and neurophysiological differences among expert and novice operators while they operated a human-machine interface in normal and adverse conditions observed that experts tended to have lower prefrontal oxygenation and exhibit gaze patterns that are better aligned with the optimal task sequence with shorter fixation durations as compared to novices.'}",https://www.frontiersin.org/articles/10.3389/fnhum.2019.00375/pdf
-temporal multimodal fusion for video emotion classification in the wild,Valentin Vielzeuf,"This paper addresses the question of emotion classification. The task consists in predicting emotion labels (taken among a set of possible labels) best describing the emotions contained in short video clips. Building on a standard framework – lying in describing videos by audio and visual features used by a supervised classifier to infer the labels – this paper investigates several novel directions. First of all, improved face descriptors based on 2D and 3D Convolutional Neural Networks are proposed. Second, the paper explores several fusion methods, temporal and multimodal, including a novel hierarchical method combining features and scores. In addition, we carefully reviewed the different stages of the pipeline and designed a CNN architecture adapted to the task; this is important as the size of the training set is small compared to the difficulty of the problem, making generalization difficult. The so-obtained model ranked 4th at the 2017 Emotion in the Wild challenge with the accuracy of 58.8 %.","{'model': 'tldr@v2.0.0', 'text': 'Improved face descriptors based on 2D and 3D Convolutional Neural Networks are proposed and several fusion methods, temporal and multimodal, including a novel hierarchical method combining features and scores are explored.'}",https://arxiv.org/pdf/1709.07200
-multimodal signaling in myrmecophilous butterflies,L. P. Casacci,"Ant nests and their surrounding territories represent a hoard of trophic resources, as well as of stable and protected environments for many arthropods involved in commensal, mutualistic or parasitic associations. Among these organisms, called myrmecophiles, several are butterflies. Here, we explore the amazing diversity of strategies developed by myrmecophilous butterflies to “cheat” or manipulate ants and to elude the tough defenses of the colony. During oviposition, female butterflies use visual or plant volatile signals to identify the presence of ants, whereas chemical and vibroacoustic cues, either isolated or combined, are used by larvae and pupae to attract, deceive or appease workers. Examples of mimicry and eavesdropping on both intraspecific and interspecific signals are discussed, primarily referring to the obligate-parasitic interactions involving Maculinea butterflies and Myrmica ants. Multimodal communication is crucial to maintaining the strong cohesion and social structure of ant societies, but its corruption is at the base of the evolution and persistence of interspecific associations, which can be beneficial or detrimental for the colony’s fitness. In this framework, the remarkable complexity of signaling could have prompted the evolution of specialized life cycles enhancing the extraordinary butterfly diversity.","{'model': 'tldr@v2.0.0', 'text': 'The amazing diversity of strategies developed by myrmecophilous butterflies to “cheat” or manipulate ants and to elude the tough defenses of the colony are explored.'}",https://www.frontiersin.org/articles/10.3389/fevo.2019.00454/pdf
-deep multimodal fusion for persuasiveness prediction,Behnaz Nojavanasghari,"Persuasiveness is a high-level personality trait that quantifies the influence a speaker has on the beliefs, attitudes, intentions, motivations, and behavior of the audience. With social multimedia becoming an important channel in propagating ideas and opinions, analyzing persuasiveness is very important. In this work, we use the publicly available Persuasive Opinion Multimedia (POM) dataset to study persuasion. One of the challenges associated with this problem is the limited amount of annotated data. To tackle this challenge, we present a deep multimodal fusion architecture which is able to leverage complementary information from individual modalities for predicting persuasiveness. Our methods show significant improvement in performance over previous approaches.","{'model': 'tldr@v2.0.0', 'text': 'A deep multimodal fusion architecture is presented which is able to leverage complementary information from individual modalities for predicting persuasiveness and shows significant improvement in performance over previous approaches.'}",
-germline dicer1‐mutant intracranial sarcoma with dual chondroid and spindle cell morphology and pulmonary metastases treated with multimodal therapy,Anirban Das,"To the Editor: Pediatric intracranial sarcomas are rare tumors with poor prognoses.1–3 Mutations in the DICER1 gene on chromosome 4, which encodes a RNase-IIIb endonuclease involved in microRNA processing and thereby controlling gene expression, has recently been reported to be associated with intracranial spindle-cell sarcomas with rhabdomyosarcoma (RMS)-like features.3,4 We describe the multimodal management of a case with germline DICER1mutation and pulmonarymetastases. A 12-year-old female with headache, left hemiparesis, and seizures was referred following partial excision of a 7 × 5.7 cm2 right frontoparietal mass. Her maternal aunt had died at 39 years, with an unknown, recurrent, genitourinary malignancy. Histopathology demonstrated malignant oval-spindle cells streaming out of chondroid foci (Figure 1), staining strongly with desmin, myogenin, myo-D1, with retained INI-1, weak membranous CD99, and negative staining for GFAP, neuron-specific enolase, and synaptophysin. Magnetic resonance imaging (MRI)-spine, cerebrospinal fluid-cytology, serum AFP and BCG, and chest X ray were normal. Clinical and radiological progression within 2 weeks prompted re-exploration, achieving gross total resection (GTR), immediately followed by focal radiotherapy (59.4 Gy, 33 fractions) with adequate margins, and consensus-driven chemotherapy with vincristine, doxorubicin, and cyclophosphamide (VDC), alternating with ifosfamide, cisplatin, and etoposide (ICE). Methylation-based classification demonstrated highest concordance with “central nervous system (CNS)/embryonal RMS,” though this did not reach the diagnostic reference range. Gain of chromosomes 1q and 8 was reported. Next generation sequencing demonstrated a nonsense mutation (c.C1870T,p.R624X) in exon 11, and a missense mutation (c.A5438C,p.E1813A) in exon 25 of theDICER1 gene. Sanger sequencing on blood DNA confirmed an identical germline, nonsense mutation. Co-existence of germline truncatingmutations with somatic missense mutations characterize the cancer-predisposing DICER1 syndrome. Computed tomography-chest revealed bilateral pulmonary nodules (maximum size: 1.3 × 1 cm2). Ultrasound abdomen, pelvis, and thyroid, and detailed ophthalmological evaluation were unremarkable. After 50weeks of combination chemotherapy including five alternating cycles of VDC/ICE followed by vincristine, actinomycin-D, and cyclophosphamide (VAC), MRI-brain revealed no active disease, and fluorodeoxyglucose-positron emission tomography revealed a few tiny, calcified, inoperable, nonavid lesions in the lungs. Six months later, she remains well on recommended surveillance for DICER1 syndrome. F IGURE 1 (A) Hematoxylin-eosin (4×)—the section shows large areas of cartilage tissue (arrowhead) and a small fragment of spindled cells (arrow); the inset shows a higher-power image of cellular cartilage. (B) Hematoxylin-eosin (10×)—the section shows a tumor comprising sheets of malignant oval to spindle cells (arrow) with eosinophilic cytoplasm in a hyalinized stroma","{'model': 'tldr@v2.0.0', 'text': 'The multimodal management of a case with germline DICER1mutation and pulmonarymetastases and co-existence of germline truncatingmutations with somatic missense mutations characterize the cancer-predisposing DICer1 syndrome is described.'}",
-multimodal intraoperative monitoring during surgical correction of scoliosis to avoid neurologic damage,Tong Yu,"Abstract The purpose of this study was to evaluate the application of multimodal intraoperative monitoring (MIOM) system in patients with congenital scoliosis (CS) and adolescent idiopathic scoliosis (AIS). Twelve patients who underwent posterior surgical correction of scoliosis for CS and AIS from June 2014 to July 2018 were enrolled in this study. During the operation, we monitored the functional status of the spinal cord by MIOM. An abnormal somatosensory evoked potential was defined as a prolonged latency of more than 10% or a peak-to-peak amplitude decline of more than 50% when compared to baseline. An abnormal transcranial motor evoked potential (TcMEP) was defined as a TcMEP amplitude decrease of more than 50%. A normal triggered electromyography response, which presented with the absence of an electrical response on stimulation at 8.2 mA, indicated that the pedicle screw was not in contact with the spinal cord or nerve root. A total of 12 patients underwent MIOM surgery, of which 9 patients with negative MIOM had no significant deterioration of neurological function postoperatively, and exhibited satisfactory surgical correction of scoliosis during follow-ups. However, the remaining 3 patients suffered from MIOM events, 2 patients had normal neurological function, and 1 patient had deteriorated neurological function postoperatively. Using MIOM in CS and AIS surgery could promptly detect iatrogenic neurological injury at the early stage. Therefore, rapid response by appropriate intraoperative interventions can be taken to minimize the injury. Besides, stable MIOM recordings encourage surgeons to correct scoliosis even when the Cobb angle of scoliosis was extremely large.","{'model': 'tldr@v2.0.0', 'text': 'Using MIOM in CS and AIS surgery could promptly detect iatrogenic neurological injury at the early stage, and rapid response by appropriate intraoperative interventions can be taken to minimize the injury.'}",
-multimodal local-global ranking fusion for emotion recognition,P. Liang,"Emotion recognition is a core research area at the intersection of artificial intelligence and human communication analysis. It is a significant technical challenge since humans display their emotions through complex idiosyncratic combinations of the language, visual and acoustic modalities. In contrast to traditional multimodal fusion techniques, we approach emotion recognition from both direct person-independent and relative person-dependent perspectives. The direct person-independent perspective follows the conventional emotion recognition approach which directly infers absolute emotion labels from observed multimodal features. The relative person-dependent perspective approaches emotion recognition in a relative manner by comparing partial video segments to determine if there was an increase or decrease in emotional intensity. Our proposed model integrates these direct and relative prediction perspectives by dividing the emotion recognition task into three easier subtasks. The first subtask involves a multimodal local ranking of relative emotion intensities between two short segments of a video. The second subtask uses local rankings to infer global relative emotion ranks with a Bayesian ranking algorithm. The third subtask incorporates both direct predictions from observed multimodal behaviors and relative emotion ranks from local-global rankings for final emotion prediction. Our approach displays excellent performance on an audio-visual emotion recognition benchmark and improves over other algorithms for multimodal fusion.","{'model': 'tldr@v2.0.0', 'text': 'This work approaches emotion recognition from both direct person-independent and relative person-dependent perspectives and displays excellent performance on an audio-visual emotion recognition benchmark and improves over other algorithms for multimodal fusion.'}",https://arxiv.org/pdf/1809.04931
-gazetouchpin: protecting sensitive data on mobile devices using secure multimodal authentication,M. Khamis,"Although mobile devices provide access to a plethora of sensitive data, most users still only protect them with PINs or patterns, which are vulnerable to side-channel attacks (e.g., shoulder surfing). How-ever, prior research has shown that privacy-aware users are willing to take further steps to protect their private data. We propose GazeTouchPIN, a novel secure authentication scheme for mobile devices that combines gaze and touch input. Our multimodal approach complicates shoulder-surfing attacks by requiring attackers to ob-serve the screen as well as the user’s eyes to and the password. We evaluate the security and usability of GazeTouchPIN in two user studies (N=30). We found that while GazeTouchPIN requires longer entry times, privacy aware users would use it on-demand when feeling observed or when accessing sensitive data. The results show that successful shoulder surfing attack rate drops from 68% to 10.4%when using GazeTouchPIN.","{'model': 'tldr@v2.0.0', 'text': 'GazeTouchPIN is proposed, a novel secure authentication scheme for mobile devices that combines gaze and touch input that complicates shoulder-surfing attacks by requiring attackers to ob-serve the screen as well as the user’s eyes to and the password.'}",https://eprints.gla.ac.uk/170215/1/170215.pdf
-"the handbook of multimodal-multisensor interfaces: foundations, user modeling, and common modality combinations - volume 1",S. Oviatt,"The Handbook of Multimodal-Multisensor Interfaces provides the first authoritative resource on what has become the dominant paradigm for new computer interfaces-user input involving new media (speech, multi-touch, gestures, writing) embedded in multimodal-multisensor interfaces. These interfaces support smartphones, wearables, in-vehicle, robotic, and many other applications that are now highly competitive commercially. This edited collection is written by international experts and pioneers in the field. It provides a textbook for students, and a reference and technology roadmap for professionals working in this rapidly emerging area. Volume 1 of the handbook presents relevant theory and neuroscience foundations for guiding the development of high-performance systems. Additional chapters discuss approaches to user modeling, interface design that supports user choice, synergistic combination of modalities with sensors, and blending of multimodal input and output. They also highlight an in-depth look at the most common multimodal-multisensor combinations- for example, touch and pen input, haptic and non-speech audio output, and speech co-processed with visible lip movements, gaze, gestures, or pen input. A common theme throughout is support for mobility and individual differences among users-including the world's rapidly growing population of seniors. These handbook chapters provide walk-through examples and video illustrations of different system designs and their interactive use. Common terms are defined, and information on practical resources is provided (e.g., software tools, data resources) for hands-on project work to develop and evaluate multimodal-multisensor systems. In the final chapter, experts exchange views on a timely and controversial challenge topic, and how they believe multimodal-multisensor interfaces should be designed in the future to most effectively advance human performance.","{'model': 'tldr@v2.0.0', 'text': 'The Handbook of Multimodal-Multisensor Interfaces provides the first authoritative resource on what has become the dominant paradigm for new computer interfaces-user input involving new media (speech, multi-touch, gestures, writing) embedded in multimodAL-multisensor interfaces.'}",
-emonets: multimodal deep learning approaches for emotion recognition in video,S. Kahou,,"{'model': 'tldr@v2.0.0', 'text': 'This paper explores multiple methods for the combination of cues from these modalities into one common classifier, which achieves a considerably greater accuracy than predictions from the strongest single-modality classifier.'}",http://arxiv.org/pdf/1503.01800
-the use of digital storytelling (dst) to help students in learning multimodal writing,Dewi - Puspitasari,"This article explores the use of DST and the explicit teaching in one of universities in Indonesia on how stu-dents use it to help the learning process. Using a digital story to teach English and Based on multimodal theory, the term of DST has been increasingly used by scholars to illustrate various forms of support of learn-ing to help students learn successfully in a classroom. Despite being widely used in educational context of many countries, DST has received scanty attention from teacher especially in ESP classes. This article specif-ically describes our experience of using DST as a learning aid with students of 18 to 19 years old. In this project they individually created collected the photographs based on their interest related to the specified theme as multimodal text. In the process they utilized two linguistic resources (Bahasa Indonesia and Eng-lish) to help them in understanding the process of creation. Several supports from machine translation and machine pronunciation software were employed during the creation of DST project. The result shows that DST helps students in composing narrative writing by analyzing the visual prompts. This proves that DST is impending to support the writing process as students were engaged during the process.","{'model': 'tldr@v2.0.0', 'text': 'This article explores the use of DST and the explicit teaching in one of universities in Indonesia on how stu-dents use it to help the learning process and shows that DST helps students in composing narrative writing by analyzing the visual prompts.'}",http://jurnalnasional.ump.ac.id/index.php/LEKSIKA/article/download/3803/2285
-emoreact: a multimodal approach and dataset for recognizing emotional responses in children,Behnaz Nojavanasghari,"Automatic emotion recognition plays a central role in the technologies underlying social robots, affect-sensitive human computer interaction design and affect-aware tutors. Although there has been a considerable amount of research on automatic emotion recognition in adults, emotion recognition in children has been understudied. This problem is more challenging as children tend to fidget and move around more than adults, leading to more self-occlusions and non-frontal head poses. Also, the lack of publicly available datasets for children with annotated emotion labels leads most researchers to focus on adults. In this paper, we introduce a newly collected multimodal emotion dataset of children between the ages of four and fourteen years old. The dataset contains 1102 audio-visual clips annotated for 17 different emotional states: six basic emotions, neutral, valence and nine complex emotions including curiosity, uncertainty and frustration. Our experiments compare unimodal and multimodal emotion recognition baseline models to enable future research on this topic. Finally, we present a detailed analysis of the most indicative behavioral cues for emotion recognition in children.","{'model': 'tldr@v2.0.0', 'text': 'A newly collected multimodal emotion dataset of children between the ages of four and fourteen years old is introduced and a detailed analysis of the most indicative behavioral cues for emotion recognition in children is presented.'}",
-multimodal retrieval of similar soccer videos based on optimal combination of multiple distance measures,T. Haruyama,"This paper presents a new multimodal method for retrieval of similar soccer videos based on optimal combination of multiple distance measures. Our method first extracts three types of Convolutional Neural Network-based features focusing the players' actions, the audience's cheers and prompt reports. Then, by applying the optimal distance measure to each feature, we calculate the similarities between a query video and videos in a database. Finally, we realize accurate retrieval of similar soccer videos by integrating these similarities. Experiments on actual soccer videos demonstrate encouraging results.","{'model': 'tldr@v2.0.0', 'text': ""This paper first extracts three types of Convolutional Neural Network-based features focusing the players' actions, the audience's cheers and prompt reports, and calculates the similarities between a query video and videos in a database to realize accurate retrieval of similar soccer videos.""}",
-the development of plurilingual education through multimodal narrative reflection in teacher education: a case study of a pre-service teacher's beliefs about language education,Míriam Cabré Rocafort,"Résumé:L'auteure utilise la réflexion narrative multimodale pour étudier les croyances relatives à la formation linguistique des enseignants en devenir. La sensibilisation aux croyances acquises au moyen de pratiques réflexives est essentielle à l'amélioration de la formation des enseignants, car la réinterprétation des croyances permet d'envisager de nouveaux cadres de référence en formation linguistique. L'auteure analyse les réflexions narratives multimodales à titre d'outils méthodologiques pour déterminer si la réflexion sur l'expérience linguistique personnelle favorise l'acquisition de stratégies pédagogiques d'enseignement plurilingue. Les résultats de l'étude révèlent que certaines croyances sont très fermement ancrées, alors que d'autres sont davantage susceptibles d'être modifiées, les forces à l'œuvre étant la pression de l'hégémonie du purisme linguistique et des principes du monolinguisme. L'auteure utilise une étude de cas pour illustrer comment la réflexion narrative multimodale dans le cadre d'une approche de la formation des enseignants axée sur le plurilinguisme engendre une transformation des croyances. La réflexion narrative multimodale contribue à ce que la participante renonce à l'idée que les langues doivent être maîtrisées et adopte la notion plus fluide selon laquelle l'apprentissage des langues relève du développement d'un répertoire plurilingue. Ces observations ont des répercussions sur la formation des enseignants au chapitre de l'obtention de changements significatifs liés à la formation linguistique. L'étude démontre qu'il convient de développer un cadre pour la formation plurilingue au niveau universitaire.Abstract:This study utilizes multimodal narrative reflection to examine pre-service teachers' beliefs regarding language education. Building awareness of beliefs acquired by reflective practices is fundamental to the improvement of teacher education because reinterpretation of beliefs allows for the consideration of new frames of reference for language education. As methodological tools, multimodal narrative reflections are analyzed to identify whether reflection on personal linguistic experience favours the acquisition of teaching strategies for plurilingual education. Results show that there are some beliefs that are deeply seated, while other beliefs are more likely to be modified; what we see at work here is the pressure of hegemonic linguistic purism and monolingual principles. This article uses a case study to illustrate how multimodal narrative reflection within a plurilingual approach to teacher education prompted a transformation of beliefs. It helped the participant to move away from the idea that languages have to be mastered, to a more fluid notion that learning languages is about the development of a plurilingual repertoire. These findings have implications for teacher education in terms of achieving meaningful changes related to language education. The study demonstrates that it is necessary to develop a framework for plurilingual education at the university level.",,
-fluorescence guided sentinel lymph node mapping: from current molecular probes to future multimodal nanoprobes.,S. Hameed,"For SLN lymph node biopsy (SLNB), SLN mapping has become a standard of care procedure that can accurately locate the micrometastases disseminated from primary tumor sites to the regional lymph nodes. The broad array of SLN mapping has prompted the development of a wide range of SLN tracers, rationally designed for noninvasive and high-resolution imaging of SLNs. At present, conventional SLN imaging probes (blue dyes, radiocolloids, and few other small-molecular dyes), although serving the clinical needs, are often associated with major issues such as insufficient accumulation in SLN, short retention time, staining of the surgical field, and other adverse side effects. In a recent advancement, newly designed fluorescent nanoprobes are equipped with novel features that could be of high interest in SLN mapping such as (i) a unique niche that is not met by any other conventional SLN probes, (ii) their adoptable synthesis method, and (ii) excellent sensitivity facilitating high resolution SLN mapping. Most importantly, lots of effort has been devoted for translating the fluorescent nanoprobes into a clinical setup and also imparting the multimodal imaging abilities of nanoprobes for the excellent diagnosis of life-threatening diseases such as cancer. In this review, we will provide a detailed roadmap of the progress of a wide variety of current fluorescent molecular probes and emphasize the future of nanomaterial-based single/multimodal imaging probes that have true potential translation abilities for SLN mapping.","{'model': 'tldr@v2.0.0', 'text': 'A detailed roadmap of the progress of a wide variety of current fluorescent molecular probes is provided and the future of nanomaterial-based single/multimodal imaging probes that have true potential translation abilities for SLN mapping are emphasized.'}",
-"assuming a hybrid logic of reading: comics, psychoanalysis, and the multimodal promise of visual response",David Lewkowich,"ABSTRACT With the increasing educational and institutional legitimacy afforded to multimodal texts, there is a need to further explore the use of the visual and its place in reader response, not only as a textual means to prompt interpretation but also as a form of interpretation itself. In this paper, I look at the multimodal interpretive practices of one adult reader who participated in a study I recently conducted with a number of undergraduate students in teacher education, reading a series of graphic novels that centred on themes of adolescence. I explore this reader’s responses to two texts: Jillian and Mariko Tamaki’s This One Summer, and Lynda Barry’s My Perfect Life. In response to their experiences of reading, I asked this reader to think about her own adolescence and to create a visual representation of a memory that was sparked while reading. I thus proceeded with a methodological assumption that to limit our students to only one mode of response is also to limit their possibilities for textual description and existential understanding. Along with a theory of multimodal literacies, I turn to psychoanalytic theory as a way to describe the potential effects of the unconscious on reading experience, memory, and visual response.",,
-multitask instruction-based prompting for fallacy recognition,Tariq Alhindi,"Fallacies are used as seemingly valid arguments to support a position and persuade the audience about its validity. Recognizing fallacies is an intrinsically difficult task both for humans and machines. Moreover, a big challenge for computational models lies in the fact that fallacies are formulated differently across the datasets with differences in the input format (e.g., question-answer pair, sentence with fallacy fragment), genre (e.g., social media, dialogue, news), as well as types and number of fallacies (from 5 to 18 types per dataset). To move towards solving the fallacy recognition task, we approach these differences across datasets as multiple tasks and show how instruction-based prompting in a multitask setup based on the T5 model improves the results against approaches built for a specific dataset such as T5, BERT or GPT-3. We show the ability of this multitask prompting approach to recognize 28 unique fallacies across domains and genres and study the effect of model size and prompt choice by analyzing the per-class (i.e., fallacy type) results. Finally, we analyze the effect of annotation quality on model performance, and the feasibility of complementing this approach with external knowledge.","{'model': 'tldr@v2.0.0', 'text': 'This work shows how instruction-based prompting in a multitask setup based on the T5 model improves the results against approaches built for a specific dataset such as T5, BERT or GPT-3 and examines the effect of model size and prompt choice on model performance.'}",http://arxiv.org/pdf/2301.09992
-fostering germane load through self-explanation prompting in calculus instruction,Cecep Anwar Hadi Firdos Santosa,"The purpose of this research was to investigate the effect of self-explanation prompting to students’ germane load while studying mathematics in the multivariable calculus course. This research employed a quasi-experimental method with matching-only posttest-only control group design. The subject of the research consists of 72 first-year mathematics education undergraduate students. The results indicated that there was no significant difference in students’ germane load between students who implemented worked-example with self-explanation prompting and students who implemented worked-example without self-explanation prompting. However, it was revealed that the students' germane load was categorized high in both classes. It indicates that the worked-example method could foster students' germane load. Nonetheless, these results cannot be evidence that self-explanation prompting is capable to foster students' germane load. However, there is an association between germane load and learning objectives. When students achieve the learning objectives, then its learning method is able to foster the germane load. To assess the learning objectives, the posttest was arranged. The results stated that students who implemented the worked-example method with self-explanation prompting had better test scores than students who implemented the worked-example method without self-explanation prompting. This result was sufficient to provide evidence that the use of worked-example with self-explanation prompting could foster students’ germane load students in the multivariable calculus course.",,
-using simultaneous prompting and computer-assisted instruction to teach narrative writing skills to students with autism spectrum disorders,J. Schuster,"Despite the importance of written expression to the lives of individuals with autism spectrum disorders (ASD), there is limited research on teaching writing skills to this population. In the current study, the researcher evaluated the effects of simultaneous prompting (SP) and computer-assisted instruction (CAI) on the story writing responses of 5 males with autism, 6 to 10 years of age. A classroom teacher and teacher assistant conducted all probe and instructional sessions. The researcher used a multiple probe across behaviors design to evaluate the efficacy of the intervention. In addition, he used pre-posttest measures to assess the generalization of acquired skills across untrained story templates and different response topographies (i.e., handwriting, verbal). Finally, the researcher evaluated the effects of using SP and CAI during writing instruction on the expressive naming of untrained sight words. The data indicated that SP and CAI were effective in improving the story writing skills of and increasing the number of sight words read by all 5 participants. In addition, all participants demonstrated maintenance and generalization of story writing skills. KEYWORDS: Autism Spectrum Disorders, Writing Instruction, Simultaneous Prompting, Computer-Assisted Instruction, Story Writing.",,
-rapid transition of traditionally hands-on labs to online instruction in engineering courses,D. May,"ABSTRACT The COVID-19 pandemic forced universities to suspend face-to-face instruction, prompting a rapid transition to online education. As many lab courses transitioned online, this provided a rare window of opportunity to learn about the challenges and affordances that the online lab experiences created for students and instructors. We present results from exploratory educational research that investigated student motivation and self-regulated learning in the online lab environment. We consider two student factors: motivation and self-regulation. The instrument is administered to students (n = 121) at the beginning of the semester and statistically analysed for comparisons between different demographic groups. The results indicated students' major was the only distinguishing factor for their motivation and self-regulation. Students' unfamiliarity with online labs or uncertainty about what to expect in the course contributed to the lower levels of self-regulation. The lack of significant differences between various subgroups was not surprising, as we posit many students entered the virtual lab environment with the same level of online lab experience. We conducted interviews among these respondents to explore the factors in greater detail. Using latent Dirichlet allocation, three main topics that emerged: (1) Learning Compatibility, (2) Questions and Inquiry, and (3) Planning and Coordination.",,https://www.tandfonline.com/doi/pdf/10.1080/03043797.2022.2046707?needAccess=true
-using response-prompting procedures during small-group direct instruction: outcomes and procedural variations.,Jennifer R. Ledford,"Research was reviewed on small-group instruction for learners with disabilities. The review was conducted for articles published between 1990 and 2010 on the application of small-group direct instruction to teach discrete skills using prompting procedures. A total of 47 articles with 197 participants and 687 replications of effects was located. Small-group instruction was effective for 195 of 197 participants and across variations in implementation and contexts. Implementers were primarily special education personnel, and instruction typically occurred in special education settings. Rigorous designs were used in all studies, and fidelity was assessed in 46 of 47 studies and was uniformly high. Students consistently reached criterion on their own target behaviors, generalized those behaviors, maintained them, and learned the behaviors taught to their peers (when this was measured, which occurred in a majority of the studies). Future research should examine comparisons of procedural variables and promoting social behaviors between group mates.","{'model': 'tldr@v2.0.0', 'text': 'Small-group instruction for learners with disabilities was effective for 195 of 197 participants and across variations in implementation and contexts, and fidelity was assessed in 46 of 47 studies and was uniformly high.'}",
-when do you need chain-of-thought prompting for chatgpt?,Jiuhai Chen,"Chain-of-Thought (CoT) prompting can effectively elicit complex multi-step reasoning from Large Language Models~(LLMs). For example, by simply adding CoT instruction ``Let's think step-by-step'' to each input query of MultiArith dataset, GPT-3's accuracy can be improved from 17.7\% to 78.7\%. However, it is not clear whether CoT is still effective on more recent instruction finetuned (IFT) LLMs such as ChatGPT. Surprisingly, on ChatGPT, CoT is no longer effective for certain tasks such as arithmetic reasoning while still keeping effective on other reasoning tasks. Moreover, on the former tasks, ChatGPT usually achieves the best performance and can generate CoT even without being instructed to do so. Hence, it is plausible that ChatGPT has already been trained on these tasks with CoT and thus memorized the instruction so it implicitly follows such an instruction when applied to the same queries, even without CoT. Our analysis reflects a potential risk of overfitting/bias toward instructions introduced in IFT, which becomes more common in training LLMs. In addition, it indicates possible leakage of the pretraining recipe, e.g., one can verify whether a dataset and instruction were used in training ChatGPT. Our experiments report new baseline results of ChatGPT on a variety of reasoning tasks and shed novel insights into LLM's profiling, instruction memorization, and pretraining dataset leakage.","{'model': 'tldr@v2.0.0', 'text': 'This analysis reflects a potential risk of overfitting/bias toward instructions introduced in IFT, which becomes more common in training LLMs, and indicates possible leakage of the pretraining recipe, e.g., one can verify whether a dataset and instruction were used in training ChatGPT.'}",http://arxiv.org/pdf/2304.03262
-teacher implementation profiles for integrating computational thinking into elementary mathematics and science instruction,Kathryn M. Rich,,"{'model': 'tldr@v2.0.0', 'text': 'This work analyzed how eight elementary teachers created opportunities for their students to engage in four CT practices during unplugged mathematics and science activities and grouped teachers into four profiles of implementation according to how they used these three strategies.'}",
-using simultaneous prompting and computer-assisted instruction to teach story writing to students with autism,R. Pennington,"In the current study, the researchers evaluated the effects of simultaneous prompting and computerassisted instruction on the story-writing responses of 3 males with autism, 7 to 10 years of age. Classroom teachers conducted all probe and training sessions. The researchers used a multiple baseline across participants design to evaluate the efficacy of the intervention. In addition, they used preposttest measures to assess the generalization of acquired skills across untrained story topics and different response topographies. The data indicated that simultaneous prompting and computer-assisted instruction were effective in improving the story-writing skills of all 3 participants. Two of the participants demonstrated maintenance and generalization of trained responses.","{'model': 'tldr@v2.0.0', 'text': 'The data indicated that simultaneous prompting and computer-assisted instruction were effective in improving the story-writing skills of all 3 participants.'}",
-effectiveness of video prompting delivered via augmented reality for teaching transition-related math skills to adults with intellectual disabilities,Ryan O. Kellems,"The purpose of this study was to determine the effectiveness of a video-based instruction packet for teaching math-based vocational skills delivered through augmented reality (AR) to three adults with intellectual disabilities. The dependent variable was the percentage of steps performed correctly to solve each selected type of math problem. The independent variable was the video-based math intervention delivered via AR, which modeled the individual steps for solving three different multistep math problems: (1) adjusting a recipe to serve a different number of people, (2) calculating salary, and (3) calculating unit prices. Visual and statistical analyses demonstrated a functional relationship between the video-based math intervention and an increase in the percentage of steps completed correctly for each type of problem. All three participants showed significant gains immediately after receiving the intervention and maintained the learned skills following withdrawal of the intervention. Implications for practitioners and further research are discussed.",,
-"giving, prompting, making: aligning technology and pedagogy within tpack for social studies instruction",Thomas Hammond,"Technological pedagogical content knowledge (now known as technology, pedagogy, and content knowledge, or TPACK) has become a widely referenced conceptual framework within teacher education. It provides a common language to discuss the integration of technology into instruction (Koehler & Mishra, 2008) and builds upon the concepts of pedagogical content knowledge (Shulman, 1987) and teacher as curricular “gatekeeper” (Thornton, 2001a, 2001b). This paper describes a three-part pedagogical model—giving-prompting-making—to explicate the relationship between pedagogy and technology within the social studies classroom. This model is intended to enhance the TPACK framework by providing a clear and intuitive comparison between social studies teachers’ pedagogical aims and their choices with technology. The giving-prompting-making model can be used to guide social studies teacher education students to make the most appropriate use of technology.",,
-a multilevel analysis of the effect of prompting self-regulation in technology-delivered instruction,Traci Sitzmann,"Two studies were conducted to examine the effect of prompting self-regulation, an intervention designed to improve learning from technology-delivered instruction. In Study 1, trainees who were prompted to self-regulate gradually improved their declarative and procedural knowledge over time, relative to the other conditions, whereas test scores declined over time for trainees who were not prompted to self-regulate. In Study 2, basic performance remained stable over time and strategic performance improved over time for trainees who were prompted to self-regulate, relative to the other conditions, whereas performance declined over time for trainees who were not prompted to self-regulate. Trainees’ cognitive ability moderated the effect of the prompts on basic performance and task-specific self-efficacy moderated the effect of the prompts on strategic performance. Prompting self-regulation resulted in stronger performance gains over time for trainees with higher ability or higher self-efficacy. These results demonstrate prompting self-regulation improved performance over time, relative to the other conditions, in both online, work-related training and laboratory settings. The results are consistent with theory suggesting self-regulation is a dynamic process that has a gradual effect on performance and highlight the importance of using a within-subjects design in self-regulation research.",,https://ecommons.cornell.edu/bitstream/1813/77378/1/WP08_121.pdf
-"explicit linguistic knowledge is necessary, but not sufficient, for the provision of explicit early literacy instruction",Alison Arrow,,"{'model': 'tldr@v2.0.0', 'text': 'The teachers in the study used predominantly implicit approaches to early reading instruction, with word-level instruction and prompting used only after context, even when teachers with high linguistic knowledge used implicit approaches, suggesting that teacher’s knowledge is not sufficient, on its own, to ensure effective, explicit, word- level instruction to beginning readers.'}",
-"giving, prompting, making: framing a conceptual home for tpack in social studies instruction",Thomas Hammond,,,
-a comparison of constant time delay and simultaneous prompting within embedded instruction on teaching leisure skills to children with autism,Onur Kurt,"An adapted alternating-treatments design was used to compare the effectiveness and efficiency of constant time delay and simultaneous prompting procedures within an embedded instruction format on the acquisition of various leisure skills by four preschool students with autism. The results showed that both procedures were effective in promoting the acquisition of the skills and maintaining them over time by three students. Although the impacts of the procedures were evident for these three students, the results were not replicated with the fourth student. Furthermore, mixed results were obtained regarding the efficiency measures. However, the social validity findings of the study were positive overall. On the basis of an evaluation of the findings, implications and future research needs are discussed.",,
-comparison of simultaneous prompting procedure in 1:1 and small group instruction to teach play skills to preschool students with pervasive developmental disorder and developmental disabilities.,Gail A. Colozzi,"This study examined and compared the effectiveness of a simultaneous prompting procedure used in both 1:1 and small group instruction to teach pretend play skills to a group of preschool students, three having a diagnosis of pervasive developmental disorder and one having a diagnosis of severe developmental disabilities. The study also assessed acquisition of instructive feedback and observational learning stimuli. A multiple probe design across responses and replicated across students and the two instruction formats assessed and compared effectiveness of the simultaneous prompting procedure. Data indicates that simultaneous prompting procedure resulted in acquisition, maintenance, and some generalization of the target skills and instructive feedback stimuli in both instruction formats with minimal differences. The group format allowed the opportunity for observational learning to occur and data on observational learning accuracy and generalization are reported. Severe and pervasive impairment in communication and social skills are primary diagnostic characteristics for children with Pervasive Developmental Disorders (PDD), which includes Autistic Disorder. These deficits impact other skill areas, such as, self-care and play, and, overall, the ability of these children to interact effectively with their environment. Individualized and errorless instructional procedures that use systematic prompt fading methods are highly recommended and welldocumented for their success in teaching skills to children with PDD and other developmental disabilities (Akmanoglu & Batu, 2004; Akmanoglu-Uludag & Batu, 2005; Bozkurt & Gursel, 2005; Farmer, Gast, Wolery, & Winterling, 1991). Research over the past 14-years on one specific method called simultaneous prompting shows it to be a promising and errorless teaching method that is a more efficient, simpler, and straightforward method for teaching a variety of skills (Morse & Schuster, 2004). Simultaneous Prompting (SP) is a systematic form of antecedent prompt and test procedure; all trials are conducted at zero seconds delay between delivery of the discriminative stimulus and the controlling prompt. Unlike antecedent prompt and test, SP first, always uses a controlling prompt, and second, always conducts probe or test trials prior to instructional sessions, rather than after instruction. Schuster and Griffen (1993) pointed out the need for systematic daily probe trials, since learners are always given the controlling prompt during instruction and thus, never have the opportunity to respond independently. Probe trials are necessary to assess the transfer of stimulus control from the controlling prompt to the discriminative stimulus. Morse and Schuster (2004) conducted an extensive review of research on SP procedure. After analyzing 18 published articles, they concluded that SP procedure has been effective in teaching individuals age preschool through adult, with and without disabilities, discrete and chained tasks. The research to date has high procedural and dependent variable reliThe authors would like to thank special education teachers Joyce Camenker, Sally Ann Watt, Tricia Dimke, and music and movement therapist Mary Beth Trapeni for their contributions to the study. Correspondence concerning this article should be addressed to Gail A. Colozzi, 7 Matthew Drive, Salem, NH 03079-1516. Email: bgaegc@earthlink.net Education and Training in Developmental Disabilities, 2008, 43(2), 226-248 © Division on Developmental Disabilities 226 / Education and Training in Developmental Disabilities-June 2008 ability, as well as, positive measures of maintenance and generalization. They cited a number of suggestions for future research that included use of SP (a) with middle school age participants and participants having profound intellectual disabilities, such as, autism spectrum disorders, (b) in non-school settings and activities, (c) with small and large group instructional arrangements, particularly 4:1 or higher student to teacher ratios, (d) related to examination of instructional efficiency variables, like incidental or observational learning, (e) to teach chained tasks, (f) to explore ways to decrease participant errors during daily probes, (g) implemented by non-certified instructors, such as, paraprofessionals, volunteers, parents, and peer tutors, and (h) in comparison with other prompting procedures, such as, system of least prompts or constant time delay. Tekin and Kircaali-Iftar (2002) compared a 4-second constant time delay procedure to an SP procedure (zero delay) in their study of sibling tutors who were trained to implement the procedures to teach receptive animal identification to developmentally delayed students. They found minimal differences between the two procedures; the constant time delay resulted in more generalization, but simultaneous prompting resulted in more efficiency, in terms of fewer errors and less training time. These factors could, in turn, lead to lower frustration and inappropriate behaviors during teaching. Siblings performed as reliably as adults in this study and led the authors to conclude that SP procedures could easily be implemented by teacher, parents, tutors and other related service personnel. Morse and Schuster (2004) state the need for not only teachers, but also, parents, siblings, paraprofessionals, and peer tutors to be able to use effective methods to teach students with disabilities, given the emphasis on inclusion and the need to provide these students with opportunities to access their natural environment. Tekin-Iftar (2003) demonstrated the ease and effectiveness of SP procedures used by peer tutors to teach expressive identification of community sight words and non-target information (instructive feedback responses) to students with developmental disabilities. Other studies have demonstrated the effectiveness of SP procedures in 1:1 and small group instruction, with students having mild disabilities (Johnson, Schuster, & Bell, 1996) and moderate to severe developmental delays (Fetco, Schuster, Harley, & Collins, 1999; Fickel, Schuster, & Collins, 1998; Gursel, Tekin-Iftar, & Bozkurt, 2006; Maciag, Schuster, Collins, & Cooper, 2000; Parrott, Schuster, Collins, & Gassaway, 2000; Schuster & Griffen, 1993; Singleton, Schuster, & Ault, 1995). Birkan (2005) effectively taught three students, ages preschool, primary, and secondary, with mild to moderate developmental delays, three separate academic tasks of reading sight words, receptively identifying numbers, and telling time in hours, using SP during 1:1 instruction. Students maintained and generalized across people, settings, and materials. SP is currently being studied as a promising teaching method that may combine efficiency and error reduction in skill acquisition, maintenance, and generalization for students having various disabilities from preschool through high school age, in a variety of skill areas, such as, academics, vocational, self-care, and language (Dogan & Tekin-Iftar, 2002; Fetco et al., 1999; Gibson & Schuster, 1992; Morse & Schuster, 2004; Parker & Schuster, 2002; Schuster, Griffen, & Wolery, 1992; Singleton, Schuster, Morse, & Collins, 1999). Several studies have shown acquisition of instructive feedback during SP procedures (Griffen, Schuster, & Morse, 1998; Gursel et al., 2006; Parker & Schuster, 2002; Parrott et al., 2000; Schuster & Griffen, 1993; Singleton et al., 1995; Tekin-Iftar, Acar, & Kurt, 2003; Wolery, Holcombe, Werts, & Cipoloni, 1993). Instructive feedback is a strategy whereby additional, non-target stimuli are presented during consequent events of an instructional trial. Students are not reinforced for or expected to respond to this additional material. Werts, Wolery, Holcombe, and Gast (1995) reviewed 20 studies that showed successful acquisition of instructive feedback. Parker and Schuster demonstrated successful acquisition of targets, observational stimuli, and instructive feedback in their study that used SP with high school students. Likewise, Gursel et al. most recently demonstrated the efficacy of SP during small group instruction of middle school aged students to teach target academic skills and nontarget skills through instructive feedback and observational learning. Comparison of Simultaneous Prompting / 227 SP has been cited as an easy instructional method for teachers, siblings, and peer tutors to implement, since no fading procedures are required and a number of studies have shown a high degree of fidelity with regard to procedural implementation (Akmanoglu & Batu, 2004; Akmanoglu-Uludag & Batu, 2005; Morse & Schuster, 2004; Tekin & KircaaliIftar, 2002; Tekin-Iftar, 2003; Wolery et al., 1993). In most of the SP research, teachers have implemented the programming, rather than paraprofessionals or parents (Morse & Schuster). Two studies used sibling and peer tutors (Tekin & Kircaali-Iftar; Tekin-Iftar). Though there is a growing body of research on using SP to teach a variety of discrete and chained skills, to date, minimal research has been conducted with students diagnosed with PDD (Almanoglu & Batu, 2004; AkmanogluUludag & Batu, 2005; Morse & Schuster, 2004) and none with preschoolers having PDD. SP has been used effectively with the preschoolers having developmental disabilities to teach expressive word recognition (Gibson & Schuster, 1992), expressive object identification (MacFarland-Smith, Schuster, & Stevens, 1993), and self-care skills (Sewall, Collins, Hemmenter, & Schuster, 1998). Wolery (1998) cited the need for systematic teaching methods, including simultaneous prompting procedure, to be used with preschoolers having developmental delays, if they are to acquire skills. Akmanoglu and Batu were among the first to study SP procedure with autistic students aged 6 – 17-years old, who acquired, maintained, and generalized across materials, the skill of pointing to numerals. In a second study, Akmanoglu-Uludag and Batu, showed",,
-caregiver-delivered home-based instruction using simultaneous prompting for teaching home skills to individuals with developmental disabilities,Sema Batu,"It is very important for individuals with all kinds of developmental disabilities to learn skills in order to be independent at home. The purposes of the study were twofold; (1) to examine the effectiveness of caregiver-delivered home-based instruction using simultaneous prompting to children with moderate develop- mental disabilities on teaching chained home skills and (2) to find out whether the caregivers would be able to implement simultaneous prompting procedure correctly throughout the study. Moreover, maintenance and generalization of the acquired skills were also evaluated. Participants of the study included four children with moderate mental retardation and their caregivers. After the caregivers were trained to use the simultaneous prompting procedure at the beginning of the study and were expected to teach three different chained-home skills to their children. Results of the study revealed that caregivers were successful in using the simultaneous prompting procedure. The children acquired target skills of the study conducted with multiple exemplars. Results also showed that the children maintained them one and three weeks after the training sessions were completed, and generalized them across trainers. Results of the study were extensively discussed.","{'model': 'tldr@v2.0.0', 'text': 'The purposes of the study were to examine the effectiveness of caregiver-delivered home-based instruction using simultaneous prompting to children with moderate develop- mental disabilities on teaching chained home skills and to find out whether the caregivers would be able to implement simultaneous prompting procedure correctly throughout the study.'}",
-parent-delivered community-based instruction with simultaneous prompting for teaching community skills to children with developmental disabilities,Elif Tekin‐Iftar,"The present study was designed to determine whether parents (three mothers and one grandmother) could implement CBI with SP reliably for teaching community skills to their children and the effects of parent-delivered intervention on teaching the community skills. Maintenance and generalization effects of the intervention were also analyzed in the study. Lastly, the study was designed to reveal the participants' opinions about the intervention. Four parents and their children participated in the study and mother-child dyad was formed. A multiple probe design across community skills and replicated across children was used. Results showed that all parents implemented the intervention with a high degree of treatment integrity. Parent-delivered intervention was effective on teaching community skills. Participants were able to maintain the acquired community skills over time and generalized the acquired skills to generalization sites. The participants' opinions regarding the social validity aspects of the study were positive in general. Based upon evaluation of the findings and implications of the study, future research needs are discussed.",,
-using video prompting to teach mathematical problem solving of real-world video-simulation problems,Alicia F. Saunders,"Mathematical problem solving is necessary in many facets of everyday life, yet little research exists on how to teach students with more severe disabilities higher order mathematics like problem solving. Using a multiple probe across participants design, three middle school students with moderate intellectual disability (ID) were taught to solve video-simulation real-world mathematical problems by finger counting using video prompting (VP) in conjunction with systematic instruction (e.g., least intrusive prompting) with error correction and feedback. The simulated videos covered a variety of thematic units (i.e., pet store, grocery store, household chores, sporting goods store, outside chores, and thrift store) students may encounter in their everyday lives. The results of this study demonstrated a functional relation between VP and participants’ mathematical problem-solving skills. Findings from this study provide several implications for practice and research for using video-based interventions to teach mathematical problem solving to students with moderate ID.",,https://journals.sagepub.com/doi/pdf/10.1177/0741932517717042
-teaching spelling through prompting and review procedures using computer-based instruction,Kristin H. Mayfield,,,
-grammar instruction in the hispanic area : the case of spain with attention to empirical studies on metalinguistic activity,X. Fontich,"Grammar instruction is an unresolved issue in the Hispanic area, having long been approached from within the disjunction between rhetoric (teaching how to use language, especially writing) and grammar (teaching the grammar content). Over time grammar instruction has generated an intense debate around two positions: direct instruction on grammar content, versus instruction devoted to prompting reflection on grammar and language use. There has been an insistent and recurring tendency towards the former, a situation that still prevails. More recently, however, certain research trends in Spain, albeit a minority, have crystallized in a research trajectory on metalinguistic activity, opening up new possibili-ties for rethinking instruction based on grammar reflection to support writing. Within such a trajectory, innovative ways to promote grammar reflection within language use are explored, and metalinguistic activity becomes the focus of research and pedagogy. Crucially, some of the empirical studies developed within this trajectory suggest that grammar knowledge may not be a condition for reflection about language but its consequence, leading to a consideration of metalinguistic activity as a promising avenue for rethinking the debate on the role of grammar instruction.",,https://l1.publication-archive.com/download/1/4773
-"the effects of fading, modeling, prompting, and direct instruction on letter legibility for two preschool students with physical and developmental delays",Christine Park,"ABSTRACT The purpose of this study was to determine the effectiveness of the model, lead, and test procedure, as well as a fading procedure with prompts and Direct Instruction with two preschool developmentally delayed students. These procedures were implemented to teach a class of preschoolers to write their names in preparation for their kindergarten transition. The participants were enrolled in a special education preschool and both were referred to the program due to physical as well as academic delays. Signing their name was a daily task that the students needed to accomplish before they started the day. During baseline both students had difficulty with letter identification and formation. The results of the model, lead and test suggested that these were effective procedures. When these were faded, there was little change in both participants handwriting. The benefits of using Direct Instruction procedures and fading for preschool students with physical development delays at the preschool level were discussed.",,
-effects of prompting multiple solutions for modelling problems on students’ performance,S. Schukajlow,,,
-a comparison of constant time delay and simultaneous prompting within embedded instruction in general education classes with students with moderate to severe disabilities,Tim Riesen,,,
-coaching teachers to use a simultaneous prompting procedure to teach core content to students with autism,Elif Tekin‐Iftar,"The researchers in this study used a multiple baseline design across dyads to examine the effects of professional development with coaching to train general education teachers to use a simultaneous prompting procedure when teaching academic core content to students with autism and the effects of the procedure on the students’ outcomes. Three teacher–student dyads participated in the study. Results showed that (a) teachers acquired the ability to use the simultaneous prompting procedure with 100% accuracy, maintained the acquired teaching behaviors over time, and generalized them in teaching new academic content to their students; and (b) students acquired the targeted academic content, maintained it over time, and generalized it across different persons and settings. In addition, the students acquired instructive feedback stimuli added to instruction and maintained these over time as well. Last, both the opinions of the teachers and students about the social validity of the study were positive. Future research is needed to support these findings.",,
-"verbal prompting, hand-over-hand instruction, and passive observation in teaching children with developmental disabilities",G. Biederman,"In hand-over-hand modeling with response-contingent verbal prompts, tasks are divided into identifiable sequential components, and the achievement of each component is marked by the delivery of some form of verbal prompt. In a within-subjects design, children were trained in one skill with response-contingent verbal prompts and in a second skill with simple passive observation. A separate group of children were trained with less rigorous verbal prompting in one skill and with passive observation in a second. Consistent with previous research, we found that passive modeling was overall significantly more effective than hand-over-hand modeling and that passive modeling was significantly more effective than hand-over-hand modeling with response-contingent prompting. Our evidence therefore indicates that current methods for teaching basic skills to children with severe developmental delays may require reassessment because simple observation of modeled skills appears to be more effective than more labor-intensive instruction.",,
-an experimental analysis of proctor prompting behavior in a personalized instruction course.,Kent Johnson,"An Experimental Analysis of Proctor Prompting Behavior in a Personalized Instruction Course June 1977 Kent R. Johnson, E.S., Georgetown University M.S., University of Massachusetts Ph.D., University of Massachusetts Directed by: Professor Beth Sul zer-Azarof f The present study analyzed one aspect of proctor stud ent interactions in Personalized System of Instruction (PSI) quiz scoring sessions. When a student has omitted a quiz answer, or has supplied an unclear, partial, or erroneous quiz answer, the proctor may (a) supply the student with the correct answer and have the student repeat it (informational prompting), (b) provide varying degrees of additional information until tlic student emits the correct response (informational prompting), (c) tell the student where to find the correct answer and try again later (noninformat ional prompting) , or (d) simply mark the answer wrong and say nothing. To determine the strategies that proctors used during quiz scoring, nine proctors in a PSI Introductory Psychology course tapo-i'ccorded tlicir.^quiz scoring sessions throughout the semester. Tape recordings of proctoring sessions during the first three weeks of the course indicnted that proctors used informational prompting strategies approximately SOI. of the time an unclear, omitted, or incorrect quiz answer",,
-effects of self-mediated video modeling compared to video self-prompting for adolescents with intellectual disability.,Sally B. Shepley,"Self-instruction is a pivotal skill that promotes independence and self-determination by allowing individuals to independently access prompts during the acquisition of new skills while decreasing the need for support from another person. Self-mediated video modeling (SMVM) and video self-prompting (VSP) are two methods of video-based self-instruction in which individuals use technology to access visual support while practicing new skills. This study evaluated the efficiency of SMVM versus VSP to teach four high school females with intellectual disability art-related tasks in a school setting. An adapted alternating treatments design was used to compare the effects of SMVM and VSP on sessions to criterion and instructional time to criterion. Results indicated that both SMVM and VSP resulted in an increase in percentage of correct responses across all participants, but VSP tasks resulted in all participants reaching mastery criterion in the fewest sessions to criterion. Self-instruction occurs when an individual uses resources available to learn a skill rather than relying on support or guidance provided directly by another person. The materials used for self-instruction may serve as the prompt to notify the learner of the step that should occur next in a sequence to complete a skill (Browder & Shapiro, 1985). Self-instruction promotes monitoring of one’s own behaviors when completing a task and is a form of selfdirected learning. Teaching individuals how to self-instruct fosters self-determination (Agran & Wehmeyer, 2000). Self-determination is the ability to make decisions on one’s own life and plays a role in the success of an individual after graduating high school (Wehmeyer & Palmer, 2003). Self-instruction is a pivotal skill in that it allows individuals to not only learn one to three target behaviors, but instead, can potentially have collateral effects on several behaviors if the self-instruction skills generalize (Smith, Shepley, Alexander, & Ayres, 2015). For instance, once an individual learns how to self-instruct while learning to make a ham and cheese sandwich, they have acquired two skills (making a sandwich and self-instruction). While making a ham and cheese sandwich is valuable in a particular context, the skill of using a set of procedures to self-instruct creates opportunities to learn new skills. Individuals with intellectual disability (ID) rely on teachers and other postsecondary instructors, such as job coaches, to predict and teach skills needed for obtaining and maintaining employment and for independent living. Teaching each skill in isolation that is required for future employment or community settings is an inefficient use of time and resources. Further, teaching all skills needed to fully integrate into one’s community and live independently is impossible. Instructors and teachers might consider teaching individuals how to self-instruct rather than teaching skills in isolation to maximize instruction time and increase independent skills (Smith et al., 2015). With web-based resources like YouTube, engaging in self-instruction has become more commonplace. For instance, if one needs to change a battery in a car, learn to fix a dishwasher, or even play an instrument, they can find instructional materials for free on YouTube. The challenge instructionally then Correspondence concerning this article should be addressed to Sally B. Shepley, The University of Kentucky, Early Childhood, Special Education, and Rehabilitation Counseling, 229 Taylor Education Building, Lexington, KY 40506. E-mail: sally.shepley@uky.edu Education and Training in Autism and Developmental Disabilities, 2018, 53(3), 264–275 © Division on Autism and Developmental Disabilities 264 / Education and Training in Autism and Developmental Disabilities-September 2018 is to teach individuals with ID to sift through the available instructional resources to identify a prompt corresponding to their needs. Learning to self-instruct has the dual benefit of decreasing reliance on others while allowing a learner to pursue those topics of greatest need and interest to him or herself. Further, self-instruction may result in less stigmatization because rather than a teacher or job coach accompanying an individual into the natural environment, the learner can use mobile technology to self-instruct. Additionally, because self-instruction reduces dependence on instructors, instructors may have more time to dedicate to working with other learners or completing other duties. Smith et al. (2015) reviewed the research on self-instruction strategies for individuals with ID that specifically resulted in generalization to novel skills. The results indicated that 56 of 57 included participants in the review learned at least one multi-step skill with a self-instructional strategy, “meaning the participant independently controlled and manipulated the [self-instruction] materials” (p. 21). The review identified three times within the experimental process in which individuals learned to self-instruct, including the use of history training prior to the baseline condition, teaching between baseline and intervention, or teaching during the intervention itself. Of the 56 individuals with ID that successfully acquired a self-instructional strategy, only 32 generalized this strategy to learn at least one additional multi-step skill. Smith et al. identified potential barriers related to generalization of self-instructional skills in the included studies. One example barrier was 7 of the 19 studies incorporated vocal directions to use the self-instruction tool in their task direction (e.g., “Restock the vending machine. Watch the video on your iPod”). Fundamentally, this means the learners did not have to discriminate on their own that they needed to use their self-instructional tools, rather they were directly coached by an instructor to do so. This, in some ways, defeats the end goal of self-instruction. Another issue that Smith et al. (2015) cited related to a lack of instruction for learners on how to search for and select self-instructional materials. In 18 of the 19 included studies, researchers directly loaded the self-instructional materials to the specific prompt required for task completion (e.g., the audio tape corresponding to target task was loaded in the cassette player or a video model demonstrating the targeted skill was loaded to a mobile device). An alternative would be teaching an individual to navigate through their self-instructional tool to locate a specific prompt among multiple available options. As technology has advanced, selfinstructional tools have the potential to store multiple needed prompts to complete identified targeted skills. The tools used by individuals with ID to self-instruct have evolved since the 1980s when researchers used books and picture prompts to teach a novel set of skills (e.g., Wacker & Berg, 1983, 1984). Self-instruction later incorporated more high-tech tools, such as cassette players with headphones to provide auditory prompting in the late 1990s (Trask-Tyler, Grossi, & Heward, 1994), and portable DVD players to promote self-meditated video modeling (SMVM) in the early 2000s (Mechling, Gast, & Fields, 2008; Mechling & Stephens, 2009). Beginning in 2009, handheld devices became more prevalent on the technology market; therefore, the tools used in self-instructional research followed suit. Mechling, Gast, and Seid (2009) taught participants SMVM using a Hewlett Packard iPAQ Pocket PC to complete cooking recipes. The first-generation Apple iPhone was released in 2007 (Apple Inc., 2007) and Bereznak, Ayres, Mechling, and Alexander (2012) first used it as a self-instructional tool for learners with ID and autism spectrum disorder (ASD) to complete vocational and independent living tasks. Additional research has focused on video prompting self-instruction, or video selfprompting (VSP), in which a video of a task is separated into steps so that the individual views and then imitates a single step or a few steps of the task at a time, as opposed to the entire task in video modeling. Bereznak et al. (2012) evaluated the use VSP and taught three high school students how to pause and play video prompts of vocational and daily living skills such as using a washing machine, making noodles, and using a copy machine. After each step in the task analysis, the video displayed a stop sign for 4 s signaling to the participant to pause the video and complete the demonstrated step. Two of the three parSelf-Instruction Video Comparison / 265 ticipants learned to pause and play the video prompts and reached mastery criterion of the various vocational and daily living tasks. Shepley, Spriggs, Samudre, and Elliott (2017) taught four middle school students with ID to self-instruct using a similar VSP format. The researchers created videos with embedded pause signs [similar to Bereznak et al.’s (2012) stop signs] to cue the participants to pause the video, complete the demonstrated step(s), and resume play to view additional video prompts. All participants learned to navigate the technology to find the necessary videos and pause/play videos using a system of least prompts instruction. After completing technology training, three of the four participants were able to self-instruct using video prompts to acquire a novel daily living task sequence (i.e., set the table, make noodles, and make a cup of punch). These studies provided support for video prompting as an effective selfinstructional tool for learners of various ages with ID. As mentioned above and represented in the published video-based instruction literature, videos can take one of two forms, video modeling and video prompting. Both methods have led to skill acquisition, yet it is important to assess efficiency of instruction to ensure instructional time is spent wisely, thus maximize outcomes for learners with ID. Mechling, Ayres, Bryant, and Foster (2014) compared VM and ",,
-when procedures discourage insight: epistemological consequences of prompting novice physics students to construct force diagrams,E. Kuo,"ABSTRACT One aim of school science instruction is to help students become adaptive problem solvers. Though successful at structuring novice problem solving, step-by-step problem-solving frameworks may also constrain students’ thinking. This study utilises a paradigm established by Heckler [(2010). Some consequences of prompting novice physics students to construct force diagrams. International Journal of Science Education, 32(14), 1829–1851] to test how cuing the first step in a standard framework affects undergraduate students’ approaches and evaluation of solutions in physics problem solving. Specifically, prompting the construction of a standard diagram before problem solving increases the use of standard procedures, decreasing the use of a conceptual shortcut. Providing a diagram prompt also lowers students’ ratings of informal approaches to similar problems. These results suggest that reminding students to follow typical problem-solving frameworks limits their views of what counts as good problem solving.",,
-helping students think about public issues: instruction versus prompting.,W. Parker,,,
-key reading recovery strategies to support classroom guided reading instruction.,J. Lipp,"Effective teachers are continuously striving to improve their instruction. Reading Recovery teachers have detailed and specific literacy training and expertise that can be of great value to classroom teachers, especially in the area of guided reading instruction. This article explores the ways in which key Reading Recovery strategies can be interwoven into primary classroom guided reading instruction to support acceleration. Fluency, supportive book introductions, prompting, and observation and analysis are emphasized throughout.",,
-teaching students with developmental disabilities daily living skills using point-of-view modeling plus video prompting with error correction,S. Gardner,"A primary goal of instruction for students with developmental disabilities is to enhance their future quality of life by promoting skill acquisition, which will enable them to live, function, and participate in the community. One instructional method that can help students with developmental disabilities improve independence in performing daily living skills is video prompting (VP). This study investigated the effectiveness of a VP plus error correction procedure on skill acquisition when teaching daily living skills to four adolescents with mild and moderate developmental disabilities. A multiple baseline across participants design demonstrated that all four participants rapidly acquired dish washing skills upon introduction of the intervention. Future research should further explore the efficacy of error correction procedures used with VP and the impact that these procedures have on student learning and skill retention.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that all four participants rapidly acquired dish washing skills upon introduction of the intervention and future research should further explore the efficacy of error correction procedures used with VP and the impact that these procedures have on student learning and skill retention.'}",https://scholarlycommons.susqu.edu/cgi/viewcontent.cgi?article=1000&context=educ_fac_pubs
-teaching sight words to elementary students with intellectual disability and autism: a comparison of teacher-directed versus computer-assisted simultaneous prompting.,M. Coleman,"The purpose of this study was to compare the effects of teacher-directed simultaneous prompting to computer-assisted simultaneous prompting for teaching sight words to 3 elementary school students with intellectual disability. Activities in the computer-assisted condition were designed with Intellitools Classroom Suite software whereas traditional materials (i.e., flashcards) were used in the teacher-directed condition. Treatment conditions were compared using an adapted alternating treatments design. Acquisition of sight words occurred in both conditions for all 3 participants; however, each participant either clearly responded better in the teacher-directed condition or reported a preference for the teacher-directed condition when performance was similar with computer-assisted instruction being more efficient. Practical implications and directions for future research are discussed.","{'model': 'tldr@v2.0.0', 'text': ""Comparing the effects of teacher-directed simultaneous prompting to computer-assisted simultaneous prompting for teaching sight words to 3 elementary school students with intellectual disability found each participant either clearly responded better in the teachers' condition or reported a preference for the teacher's condition.""}",
-tablet-based video modeling and prompting in the workplace for individuals with autism,Raymond Burke,"The current study involved a preliminary job-site testing of computer software, i.e., VideoTote, delivered via a computer tablet and designed to provide users with video modeling and prompting for use by young adults with an autism spectrum disorder (ASD) across a range of employment settings. A multiple baseline design was used to assess changes in rates of completion with a complex, 104-step shipping task by four participants diagnosed with ASD. Baseline data were collected on accuracy of task completion after exposure to typical job-training involving instruction, modeling, and practice. The intervention involved video modeling and prompting with a 13 minute video depicting an individual completing job responsibilities that entailed checking to make sure materials were in working order, replacing defective items, packing materials in a container, entering information into a computer, and attaching a label to a container. Results suggested that video modeling and prompting were effective in helping individuals with autism complete a multi-step shipping task. Participants and their parents gave the device and software high ratings as an acceptable treatment for adults with autism to use in the workplace and intervention that complies with universal design principles. Implications for competitive job opportunities for individuals with autism are discussed.",,
-the effects of prompting in programed instruction as a function of motivation and instructions,G. Faust,,,
-a comparison of prompt delays with trial-and-error instruction in conditional discrimination training,Sean J. O’Neill,,"{'model': 'tldr@v2.0.0', 'text': 'Compared three variations of prompt delay alongside trial-and-error instruction, results suggest progressive prompt delay may be effective and the most efficient in reducing learner errors during instruction.'}",https://europepmc.org/articles/pmc6269381?pdf=render
-a preliminary discourse analysis of constructivist-oriented mathematics instruction for a student with learning disabilities,Y. Xin,"ABSTRACT Reform efforts in mathematics education arose, in part, in response to constructivist works on conceptual learning. However, little research has examined how students with learning disabilities (LD) respond to constructivist-oriented instruction in mathematics, particularly in moment-to-moment interactions. To understand the nature of constructivist-oriented mathematics instruction involving students with LD, the authors conducted a case study to analyze teacher–student interactions during constructivist-oriented small group instruction involving a student with LD. The student demonstrated, to a certain degree, the ability to reason mathematically when provided with appropriate opportunities and prompting. However, given the limited intervention time, his reasoning and problem solving did not seem to go beyond the semiconcrete level of operation, which may have inhibited his solving of complex word problems with large numbers. Findings indicate that more efforts are needed to support students, those with LD in particular, in their transitions from concrete or semiconcrete to abstract conceptual understanding and problem solving.",,
-video modeling and prompting in practice,Ryan O. Kellems,"This article discusses the creation of video modeling (VM) and video prompting (VP) interventions for teaching novel multi-step tasks to individuals with disabilities. This article reviews factors to consider when selecting skills to teach, and students for whom VM/VP may be successful, as well as the difference between VM and VP and circumstances affecting which to choose. The purpose of this article is to provide practitioners with step-by-step instruction on how to teach cooking skills using VM and VP for young adults with disabilities.",,
-use of continuous video prompting to teach a student with a significant disability,Helen I. Cannella-Malone,,,
-analysis of live modeling plus prompting and video modeling for teaching imitation to children with autism,Logan S. McDowell,"Previous researchers have demonstrated that training in imitation can significantly improve the learning capabilities of children diagnosed with autism spectrum disorder (ASD) and that children within this population show a preference for video presentations. Video-based instruction has been used to teach a variety of behaviors to individuals with ASD. However, only a small number of studies have examined the use of video modeling to teach initial imitation. Furthermore, there are limited and conflicting data on the effectiveness of a video modeling procedure that does not incorporate prompting when used to teach imitation to young children with ASD. Thus, the purpose of this study was to evaluate a video-modeling-alone procedure and a live-modeling-with-prompting procedure for teaching imitation to young children with ASD. The results suggest that the live modeling with prompting procedure was more effective, and implications related to this finding are discussed. Copyright © 2015 John Wiley & Sons, Ltd.",,
-using robot-assisted instruction to teach students with intellectual disabilities to use personal narrative in text messages,R. Pennington,"In the current investigation, we evaluated the effectiveness of a multi-component package (i.e., robot, simultaneous prompting, self-graphing) for teaching three students, ages 19–21, with intellectual disabilities (ID) to write text messages that included a greeting, personal narrative, and closing. Data suggest that the package was effective in increasing correct performance for all participants. In addition, participants demonstrated their newly acquired texting skills across different communicative partners.",,
-effects of prompting critical reading of science news on seventh graders' cognitive achievement.,Pei-ying Tsai,"Science teachers frequently select science news articles as supplementary teaching materials; however, the literature indicates that students encounter difficulties in examining and evaluating the news content and textual elements. This paper reports an instructional strategy of utilizing science news articles and investigates its effectiveness in enhancing students’ cognitive learning outcomes. In this quasi-experimental study, 118 seventh graders from four classes in one secondary school in Taiwan took part in the Science News Instruction (SNI). After eight weeks of instruction about Genetics and Reproduction, all students were requested to present their written arguments to the selected science news article. Two of the classes were prompted to link to their science content knowledge (the CK-SNI group, n=56), whereas their counterpart peers were not instructed to link to their science content knowledge (the SNI group, n=62). The analysis of students’ written artifacts revealed that the CK-SNI group performed significantly better on examining the data and claims made in the science news, producing warrants of higher quality to support their reasoning about the relationship between data and claims, and providing better reasons to refute the identified causal textual elements. Furthermore, the results of ANCOVA indicate that the CK-SNI group outperformed the SNI group on the cognitive test, with the largest improvement at Apply level (ηp 2 = .10). We conclude that prompting students to examine news content based on the content knowledge they learned is effective in fostering students to evaluate the science news critically, and thus make a positive impact on students’ cognitive learning.",,
-embedding science facts in leisure skill instruction conducted by peer tutors,Erin E. Fetko,"This investigation evaluated the effectiveness of using peer tutors to teach a chained leisure skill (i.e., UNO card game) to three middle school students with disabilities using a simultaneous prompting procedure within a multiple probe design. The investigation also assessed whether the students with disabilities would acquire four unrelated science core content facts presented as nontargeted information during instructive feedback. Results indicated that all students met or made progress toward criterion on the leisure skill. In addition, two of the three students acquired all four core content facts.",,
-using simultaneous prompting to teach computer-based story writing to a student with autism,R. C. Pennington,"Writing is a critical skill because it is used to access reinforcement in a variety of contexts. Unfortunately, there has been little research on writing skills instruction for students with intellectual disabilities and autism spectrum disorders. The purpose of this study was to evaluate the effects simultaneous prompting and computer-assisted instruction on story writing responses of a 7-year-old male with autism. Data indicated that the intervention was effective in teaching the participant to construct stories related to three different topics. Additionally, the student maintained responding at 2 and 4 weeks following intervention and increased responding across different topographies (i.e., handwriting, vocal).",,
-comparing the effects of video prompting with and without error correction on skill acquisition for students with intellectual disability,Helen I. Cannella-Malone,"This study used an iPod Touch to compare the effects of video prompting with and without error correction on the acquisition of two daily living skills across three students with moderate to profound intellectual disability and an extremely limited daily living skills repertoire. An adapted alternating treatments design within a multiple probe across participants design was used to demonstrate that the inclusion of error correction from the outset of intervention increased the efficiency of skill acquisition for at least one task for two students. For the third student, some skill acquisition was observed using video prompting both with and without error correction, but more stable responding was achieved using in vivo instruction.","{'model': 'tldr@v2.0.0', 'text': 'An adapted alternating treatments design within a multiple probe across participants design was used to demonstrate that the inclusion of error correction from the outset of intervention increased the efficiency of skill acquisition for at least one task for two students.'}",
-teaching daily living skills to seven individuals with severe intellectual disabilities: a comparison of video prompting to video modeling,Helen I. Cannella-Malone,"We conducted a systematic replication of Cannella-Malone et al. by comparing the effects of video prompting to video modeling for teaching seven students with severe disabilities to do laundry and wash dishes. The video prompting and video modeling procedures were counterbalanced across tasks and participants and compared in an alternating treatments design within a multiple probe across participants design. For six participants, video prompting was more effective than video modeling, which was generally ineffective. For one participant, neither video modeling nor video prompting was effective, but in vivo instruction led to skill acquisition. One participant who was deaf was also able to learn both skills using video prompting, even though he could not hear the voice-over instructions. These data suggest that the duration of the video may influence its effectiveness as a teaching tool and that the voice-over instructions may not be necessary.",,
-self-monitoring during whole group reading instruction: effects among students with emotional and behavioral disabilities during summer school intervention sessions,Lisa A. Rafferty,"The purpose of this study was to examine the effects that a self-monitoring strategy, plus a tactile prompting device, had upon the on-task and oral reading fluency behaviors of students with emotional and/or behavioral disabilities in the general education setting when used during whole group reading instruction. A multiple-baseline across pairs of participants design was used to evaluate the effects of the intervention. The results show that all four participants increased their on-task behavior during whole group reading instruction after being taught how to use the self-monitoring strategy, and their on-task behavior results were more similar to their peers' during the intervention condition. In addition, all students showed meaningful gains in oral reading fluency after being taught to self-monitor during whole group reading instruction when compared to baseline levels. Generalization probe data were taken during teacher-led, small group reading instruction. Although there were slight increases in performance for each of the target students after using the intervention for at least one week, these results should be interpreted with caution and should be further examined with future research.",,
-coder reviewer reranking for code generation,Tianyi Zhang,"Sampling diverse programs from a code language model and reranking with model likelihood is a popular method for code generation but it is prone to preferring degenerate solutions. Inspired by collaborative programming, we propose Coder-Reviewer reranking. We augment Coder language models from past work, which generate programs given language instructions, with Reviewer models, which evaluate the likelihood of the instruction given the generated programs. We perform an extensive study across six datasets with eight models from three model families. Experimental results show that Coder-Reviewer reranking leads to consistent and significant improvement (up to 17% absolute accuracy gain) over reranking with the Coder model only. When combined with executability filtering, Coder-Reviewer reranking can often outperform the minimum Bayes risk method. Coder-Reviewer reranking is easy to implement by prompting, can generalize to different programming languages, and works well with off-the-shelf hyperparameters.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that Coder-Reviewer reranking leads to consistent and significant improvement over reranking with the Coder model only, when combined with executability filtering, and can often outperform the minimum Bayes risk method.'}",https://arxiv.org/pdf/2211.16490
-teaching students with autism to tie a shoelace knot using video prompting and backward chaining,C. Rayner,"Purpose: To evaluate the effects of video prompting and backward chaining for teaching students with autism to tie a shoelace knot. Method: Videos featuring an adult and a peer or sibling model were used as part of the video prompting procedures to teach three boys with autism to tie a shoelace knot. A backward chaining procedure involving live modelling and verbal instruction was introduced following the video prompting phases. Results: Although the video prompting interventions increased the number of steps in the shoelace tying task completed by each of the participants, the backward chaining procedure was more effective, enabling one participant to reach mastery and a second participant to approach mastery. Conclusion: Practitioners should consider the pre-requisite skills of the participants and the nature of the target behaviour when selecting an intervention to teach daily living skills to individuals with autism.","{'model': 'tldr@v2.0.0', 'text': 'Although the video prompting interventions increased the number of steps in the shoelace tying task completed by each of the participants, the backward chaining procedure was more effective, enabling one participant to reach mastery and a second participant to approach mastery.'}",
-sight word instruction for students with autism: an evaluation of the evidence base,Janet E. Spector,,"{'model': 'tldr@v2.0.0', 'text': 'Evidence is found in support of a massed trials approach featuring student response to a succession of items, differential positive reinforcement, systematic prompting, and use of visual supports in teaching students with autism to read printed words.'}",
-research on reading instruction for individuals with significant cognitive disabilities,D. Browder,"This article presents the results of a comprehensive review of 128 studies on teaching reading to individuals with significant cognitive disabilities. The review compared these studies against the National Reading Panel's components of reading; although it revealed an inadequate consideration of the components of reading, it found strong evidence for teaching sight words using systematic prompting and fading. The reviewers considered not only the number of studies, but also indicators proposed for evidence-based practice and effect size. This study identified some high quality studies with strong effect size for comprehension and fluency, but only one phonics study was strong in both quality and effects. Additional research is needed to promote broader skills in literacy for this population.",,
-using simultaneous prompting to teach restaurant words and classifications as non-target information to secondary students with moderate to severe disabilities,Bethany R. Smith,"This paper reviews selected literature pertaining to simultaneous prompting and the acquisition of non-target information for individuals with moderate to severe disabilities. The purpose of this review was to discuss the definition of non-target information (NTI) and the various places it can be embedded within an instructional trial. The acquisition of NTI has become especially critical to teachers who serve this population of students. Incidental learning provides teachers with an opportunity to increase a student's exposure to information without increasing instructional time, therefore increasing the efficiency of the teacher's instruction. A multiple probe design with full conditions across behaviors and replicated across participants evaluated the effectiveness of simultaneous prompting (SP) when teaching four participants with moderate and severe disabilities to read 12 sight words from community restaurants. In addition, the researcher examined generalization of the non-target information provided in the discriminative stimulus. Data indicated that SP was effective in teaching the participants the targeted sight words and the participants maintained the knowledge at high rates over time. Additionally, participants acquired NTI knowledge of the food classification embedded in the discriminative stimulus. As expectations for students with moderate or severe disabilities increase, effective and effi- cient instruction is of the highest importance to special education professionals. The strug-",,
-an automated prompting system for smart environments,Barnan Das,,"{'model': 'tldr@v2.0.0', 'text': 'This paper takes the approach to fully automating a prompting system without any predefined rule set or user feedback and uses simple off-the-shelf sensors to learn the timing for prompts based on real data that is collected with volunteer participants in the smart home testbed.'}",http://www.eecs.wsu.edu/~cook/pubs/icost11p2.pdf
-academic instruction for students with moderate and severe intellectual disabilities in inclusive classrooms,J. Downing,"Foreword. by Diane Ryndak - Diane Ryndak Preface Acknowledgments About the Author 1. Teaching Students With Moderate to Severe Intellectual Disabilities in General Education Classrooms: Foundational Beliefs Key Concepts A Historical Perspective: Where We Came From The Present Situation and Challenge What is Inclusive Education? What is Not Inclusive Education Who Are We Talking About? Summary 2. Instructional Strategies and Teaching Arrangements Key Concepts Characteristics of Effective Instruction for All Students Clear Expectations Analyzing Tasks for Improved Learning What We Know About Teaching Students with Moderate to Severe Intellectual Disabilities The Importance of Student Interests Components of the Teaching Task Prompting Strategies Consequences of the Behavior Using Sequences of Different Prompts to Teach Students: Shaping Behavior Maintaining and Generalizing Skills Teaching Arrangements in General Education Classrooms Summary 3. Determining Student Needs: What to Teach Key Concepts Limitations of Standardized Assessment Family and Child-Based Assessment Procedure Record Review Observational Assessments What's the Class Doing? Interpreting Content Standards Blending Student/Family Goals with State Standards Identifying Learning Opportunities Writing IEP Goals and Objectives Summary 4. Teaching Core Curriculum to Students With Moderate to Severe Intellectual Disabilities Key Concepts The Critical Need to Adapt Curriculum to Make it Meaningful Identifying the BIG Ideas from Core Curriculum Determining Prompts to Use for a Particular Student and Lesson Examples of Students Receiving Direct Instruction Across Grades and Instructional Arrangements Large Group Instruction Generalization of Skills Taught Summary 5. It Takes a Village: Teaching as a Collaborative Effort Key Concepts The Expectation of Team Collaboration Team Members Involved in Instruction Credentialed Teachers Co-Teaching Supporting General Education Ownership Paraprofessionals as Teachers Related Service Providers Parent Volunteers Peers as Teachers A Few Cautions When Using Peers The Need for Information and Training Effective Use of Team Members The Importance of Consistency Generalization of Skills Across Team Members Summary 6. Keeping Track of Student Progress, by Kathryn D. Peckham-Hardin and June E. Downing Key Concepts Types of Data Collection Strategies Linking Data Collection Methods to the IEP Objectives Collecting Data While Teaching in General Education Classrooms Examples of Collecting Data During Instructional Times Test Taking by the Class Training Paraprofessionals and Others to Take Data The Need for Alternate Assessment Summary 7. He's Getting It! Now What? Taking Learning to the Next Level Key Concepts Involving the Student in Planning Next Steps Writing IEP Objectives to Reflect Next Steps Using Standards and Performance Indicators to Determine Next Steps Using Task Analyses to Determine Next Steps Using Life Needs to Determine Next Steps Postsecondary Options Next Steps for Nonacademic Skills Summary References Index","{'model': 'tldr@v2.0.0', 'text': 'This book discusses Teaching Students With Moderate to Severe Intellectual Disabilities in General Education Classrooms: Foundational Beliefs and Instructional Strategies and Teaching Arrangements, and the Critical Need to Adapt Curriculum to Make it Meaningful.'}",
-effects of no-no prompting on teaching expressive labeling of facial expressions to children with and without a pervasive developmental disorder,Justin B. Leaf,"Discrete trial teaching is a systematic form of instruction found to be effective for children diagnosed with autism. Three areas of discrete trial teaching warranting more research are the effectiveness and efficiency of various prompting procedures, the effectiveness of implementing teaching in a group instructional format, and the ability of children with autism to observationally learn from their peers. This study evaluated the effectiveness of a no-no prompting procedure implemented in a group instructional format to teach five children, four of whom were diagnosed with an autism spectrum disorder, to correctly label facial expressions. Addition- ally, this study evaluated whether participants observationally learned to expressively label facial expressions not directly taught to them but taught to their peers. Using a multiple baseline design, results of this study indicated that all participants learned to expressively label facial expressions taught to them directly with a no-no prompting procedure. The participants also learned through observation to expressively label facial expressions taught only to their peers using the same procedure.",,
-effects of peer tutoring with audio prompting on vocabulary acquisition for struggling readers,S. M. Mackiewicz,"Reciprocal peer tutoring can be an effective supplement to teacher-led instruction, but students need to have the tutoring skills necessary to teach their peers successfully. Previous studies have addressed the challenge of providing essential information to a naïve tutor, allowing for correct modeling and feedback. The present study compared incidental learning of vocabulary words through classroom reading instruction to a combination of incidental learning supplemented with peer tutoring. Eight fourth-grade students were trained to tutor each other using a digital recording and playback device that provides audio prompts to naïve tutors. Results indicated that students made modest gains from incidental learning and much stronger gains from peer tutoring with audio prompting. Implications for practice and future research are discussed.",,
-comparison of the effects of static picture and video prompting on completion of cooking related tasks by students with moderate intellectual disabilities,L. Mechling,"This study compared the effects of static photographs and video prompts on the independent performance of cooking related tasks by six young adults with moderate intellectual disabilities. An adapted alternating treatment design with baseline and final treatment phase was used to measure the percentage of tasks correctly completed by each student across treatments. Results indicated that both procedures were effective in increasing correct performance of tasks for all participants from baseline levels in which only verbal task directions were provided. All six participants demonstrated greater independent responding on the target tasks when using video prompting. During the final treatment phase students increased their level of performance using video prompting with sets of tasks receiving static picture prompts in the comparison phase. Implications for future research and the development of curriculum, which include instruction of self-prompting strategies, are discussed.",,
-generating instruction automatically for the reading strategy of self-questioning,Jack Mostow,"Self-questioning is an important reading comprehension strategy, so it would be useful for an intelligent tutor to help students apply it to any given text. Our goal is to help children generate questions that make them think about the text in ways that improve their comprehension and retention. However, teaching and scaffolding self-questioning involve analyzing both the text and the students' responses. This requirement poses a tricky challenge to generating such instruction automatically, especially for children too young to respond by typing. This paper describes how to generate self-questioning instruction for an automated reading tutor. Following expert pedagogy, we decompose strategy instruction into describing, modeling, scaffolding, and prompting the strategy. We present a working example to illustrate how we generate each of these four phases of instruction for a given text. We identify some relevant criteria and use them to evaluate the generated instruction on a corpus of 513 children's stories.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes how to generate self-questioning instruction for an automated reading tutor and decomposes strategy instruction into describing, modeling, scaffolding, and prompting the strategy.'}",
-enhancing creativity: a comparison between effects of an indicative instruction “to be creative” and a more elaborate heuristic instruction on chinese student creativity.,W. Niu,"This study consisted of an experiment conducted with a sample of 180 Chinese high school students,designed to scrutinize the effects of prompting and instructing students to be creative in their creativeperformance in the domains of art and literature. The results of the study reveal that students creativitybenefited from receiving appropriate strategies or more elaborate instruction (as induced by instructing),but not by merely being motivated (as induced by prompting) to be creative, regardless of the taskdomain. Gender differences in creativity were found only in the domain of literature, in favor of femaleparticipants. Implications of the findings are discussed.Keywords: creativity, instruction, Chinese students, motivation",,
-teaching effective hand raising to children with autism during group instruction.,Shaireen M. Charania,"We taught 3 children with autism to raise a hand or keep both hands down depending on their status (e.g., having heard a target word, possessing a specific item) using modeling, prompting, and reinforcement. All 3 children acquired accurate hand-raising skills in response to progressively more difficult discrimination tasks during group instruction. The implications for preparing children for general education settings are discussed.","{'model': 'tldr@v2.0.0', 'text': '3 children with autism are taught to raise a hand or keep both hands down depending on their status using modeling, prompting, and reinforcement to acquire accurate hand-raising skills in response to progressively more difficult discrimination tasks during group instruction.'}",https://europepmc.org/articles/pmc2938940?pdf=render
-effectiveness of multimedia-based instruction that emphasizes molecular representations on students' understanding of chemical change,D. Ardaç,"The present study makes use of the capabilities of computerized environments to enable simultaneous display of molecular representations that correspond to observations at the macroscopic level. This study questions the immediate and long-term effects of using a multimedia instructional unit that integrates the macroscopic, symbolic, and molecular representations of chemical phenomena. Forty-nine eighth graders received either multimedia-based instruction that emphasized molecular representations (n = 16), or regular instruction (n = 33). Students who received multimedia-based instruction that emphasized the molecular state of chemicals outperformed students from the regular instruction group in terms of the resulting test scores and the ease with which they could represent matter at the molecular level. However, results relating to the long-term effects suggest that the effectiveness of a multimedia-based environment can be improved if instruction includes additional prompting that requires students to attend to the correspondence between different representations of the same phenomena. © 2004 Wiley Periodicals, Inc. J Res Sci Teach 41: 317–337, 2004",,
-a qualitative study of teacher perceptions on using an explicit instruction curriculum to teach early reading skills to students with significant developmental disabilities,D. Taylor,"This study examines teachers’ perceptions about a structured literacy curriculum that uses explicit instruction and teacher script combined with instructional practices common in special education such as time delay, error correction, and prompting strategies. The main research question addressed by this study was “What are the perceptions and beliefs of teachers of students with significant developmental disabilities about the effects of using an explicit reading instruction curriculum?” Three themes or key findings emerged from our analysis of the data regarding teachers’ perceptions of such curricula. These themes suggest that these teachers believe that a structured literacy curriculum has a positive impact on student learning, teacher effectiveness, and teacher self-efficacy.",,
-making metacognition explicit: developing a theoretical foundation for metacognitive prompting during scenario-based training,S. Fiore,"In this paper we describe a set of metacognition related training interventions that can be introduced in advanced scenario-based training systems. A long line of research in the cognitive and learning sciences has shown that knowledge acquisition is more effective when learners reflect upon their comprehension processes. We support this view and discuss how instruction can assist trainees in monitoring their learning in order to facilitate overall retention and transfer. We focus on the multidimensional nature of metacognition and emphasize the regulatory component of metacognition. Regulation involves an active process engaged by the trainee in service of learning. We present a theoretical framework that describes how this can be developed as a means of instructional strategy in preparation for, during execution of, and in reflection upon, a given learning episode.",,
-an effective framework for primary-grade guided writing instruction,Sharan A. Gibson,"This article describes the theory and procedures for a specific format of primary-grade small-group writing instruction (guided writing lessons) based on (a) a sociocultural perspective, (b) Clay's theories of literacy learning, and (c) the author's study of second-grade guided writing instruction. Guided writing lessons are described as an intensive, small-group context for instructional support and interaction between teacher and students during writing that can be folded into a classroom context, including writing workshops and interactive writing. The purpose, format, teacher prompting, and assessment procedures of guided writing lessons are demonstrated within the article through the presentation of student examples, lesson excerpts, teacher prompts for writing, and an observation form. The article also describes the use of expert observation of student behavior while writing and an analytic assessment of writing products in order to maintain a strong pace for students' learning within and across lessons. تصف هذه المقالة نظرية نظام خاص وإجراءاته في الصفوف الابتدائية بصدد تعليم الكتابة في مجموعات صغيرة (دروس الكتابة الموجهة) المبني على (ا) وجهة اجتماعية ثقافية و(ب) نظريات (كلاي) حيال تعلم القراءة والكتابة و(ج) بحث المؤلف العلمي بشأن تعليم الكتابة الموجه في الصف الثاني. وتتم وصفة دروس الكتابة الموجهة كسياق مكثف في مجموعات صغيرة لدعم التعليم والتفاعل بين المعلم والطلاب خلال الكتابة ويمكن أن تنطوي هذه الدروس في سياق الصف التي تشتمل على ورشات الكتابة والكتابة التفاعلية. وقد تم عرض غرض دروس الكتابة الموجهة ونظامها وحث معلمها وإجراءات تقييمها في المقالة بواسطة أمثلة من الطلاب ومقتطفات الدروس وتحاريض المعلم على الكتابة واستمارة مراقبة. وتصف كذلك استخدام مراقبة الخبير تصرفات الطلاب عند كتابتهم والتقدير التحليلي لإنتاجات الكتابة من أجل الحفاظ على سرعة تعلم الطلاب القوية في الدروس وعبرها. 本文描述一个教授小学写作教学方法的理论和程序。该教学方法（指导写作练习课）是以特别的小组形式进行，其基础建于（a）一个社会文化观点，（b）克雷的读写学习理论，及（c）本文作者对小学二年级指导写作教学的研究。指导写作练习课能在教授写作时，为学生提供一个强化的小组学习环境，支援教学和促进教师与学生之间的互动。指导写作练习课可以融入一般的课堂教学环境中，包括写作工作坊和互动写作练习等。本文透过展示学生学习例子、授课节录、教师指导写作的提示及观课记录表，来表明指导写作练习课的教学目的、教学形式、教师提示和教学评估程序。本文亦描述如何使用专门方法去观察学生在写作练习时出现的行为，及如何使用分析评估方法去评估学生的写作成果，以帮助学生在课堂内，和课堂与课堂之间，保持稳健的学习进度。 Cet article decrit la theorie et les procedures d'un format specifique d'enseignement de l'ecriture en petit groupe en premiere annee (lecons d'ecriture guidee), a partir (a) d'une perspective socioculturelle, (b) des theories d'apprentissage de la litteratie de Clay, et (c) de l'etude effectuee par l'auteur d'un enseignement de l'ecriture guidee en seconde annee. Les lecons d'ecriture guidee sont decrites comme un contexte intensif en petit groupe, en vue d'un soutien pedagogique et d'interactions maitre- eleves pendant l'ecriture qui puisse etre mis en œuvre en contexte scolaire, y compris pendant des ateliers d'ecriture et d'ecriture interactive. L'article indique le but, le format, les interventions du professeur, et les procedures d'evaluation des lecons d'ecriture guidee, au moyen de la presentation d'exemples d'eleves, d'extraits de lecons, des incitations a ecrire du professeur, et d'un protocole d'observation. L'article decrit egalement l'utilisation d'une observation d'expert du comportement de l'eleve pendant l'ecriture et une evaluation detaillee des productions ecrites afin de conserver un rythme soutenu d'apprentissage des eleves au cours d'une lecon et d'une lecon a l'autre. Предлагаются теоретические обоснования и процедуры для определенного формата групповой работы над письмом в начальной школе (направляемое письмо), основанные на (а) социокультурном подходе, (б) теории изучения грамотности Клэя, и (в), исследованиях самого автора в области обучения второклассников письменной речи. Направляемое письмо происходит как интенсивные занятия в малых группах, проходящие с поддержкой учителя и при активном взаимодействии между учителем и учениками в ходе выполнения заданий. Такие занятия могут быть элементами обычного урока, писательского семинара или сессии интерактивного письма. Цель, формат, степень поддержки, которую оказывает учитель, и процедуры оценивания демонстрируются на примере ученических работ, эпизодов урока, типичных реплик-подсказок учителя и форм, которые он заполняет, наблюдая за работой учащихся. В статье также описано, как проводить экспертное наблюдение за процессом письма и аналитическое оценивание его результатов, чтобы обеспечить динамичное развитие навыков письма в пределах каждого отдельного урока и учебного курса в целом. Este articulo describe la teoria y los procedimientos de un formato especifico de la ensenanza de la escritura en pequenos grupos al nivel de primaria (lecciones de escritura guiada) basados en (a) una perspectiva sociocultural, (b) la teoria de Clay sobre el aprendizaje de competencias, y (c) el estudio hecho por el autor de la ensenanza de la escritura guiada en una clase de segundo grado. Las lecciones de la escritura guiada se describen como un contexto intenso de trabajo en grupos pequenos para apoyar y ofrecer interaccion entre el maestro y los estudiantes al escribir. Estos grupos pequenos se pueden integrar en el contexto del salon de clase, inclusive en talleres de escritura y en la escritura interactiva. En el articulo se demuestran el proposito, el formato, y los procedimientos para las sugerencias del maestro y la evaluacion de las lecciones de escritura guiada por medio de la presentacion de ejemplos de estudiantes, fragmentos de lecciones, las sugerencias de los maestros en la escritura, y un formulario de observaciones. El articulo tambien describe el uso de la observacion experta del comportamiento de los estudiantes mientras escriben y una evaluacion analitica de los productos de la escritura para mantener un ritmo rapido de aprendizaje dentro y a traves de las lecciones.",,
-speak,Beth Whittle,"This study presents holistic insights into the culturally relevant English language arts and reading instructional practices of one award-winning Asian-American male teacher as he worked with culturally and linguistically diverse students from a variety of backgrounds in a ‘significantly underperforming’ urban middle-school. Avenues through which the exemplary teacher reflected his culturally relevant teaching practices were planning instruction, prompting students to think critically through reflection, modeling, conferencing with students, facilitating peer evaluation and support, emphasizing procedural processes, establishing routine classroom practices, and capitalizing on technology and visual aid support. Mr. Wang found consistent collaboration with other teachers and opportunities for debriefing based on data from instruction to be the most beneficial forms of support. Students’ positive responses to Mr. Wang’s responsiveness were most influential to his continued instruction. Implications are provided.",,https://www.qeios.com/read/GUMCLM/pdf
-simulating naturalistic instruction: the case for a voice mediated interface for assistive technology for cognition,B. O'Neill,"A variety of brain pathologies can result in difficulties performing complex behavioural sequences. Assistive technology for cognition (ATC) attempts support of complex sequences with the aim of reducing disability. Traditional ATCs are cognitively demanding to use and thus have had poor uptake. A more intuitive interface may allow ATCs to reach their potential. Insights from psychological science may be useful to technologists in this area. We propose that an auditory‐verbal interface is more intuitive than a visual interface and reduces cognitive demands on users. Two experiments demonstrate a novel ATC, the General User Interface for Disorders of Execution (GUIDE). GUIDE is novel because it simulates normal conversational prompting to support task performance. GUIDE provides verbal prompts and questions and voice recognition allows the user to interact with the GUIDE. Research with non‐cognitively impaired participants and a single participant experiment involving a person with vascular dementia provide support for using interactive auditory‐verbal interfaces. Suggestions for the future development of auditory‐verbal interfaces are discussed.",,http://eprints.lse.ac.uk/38691/1/Running_head%28lsero%29.pdf
-reading instruction for students with emotional and behavioral disorders and at risk of antisocial behaviors in primary grades: review of literature,Mabel O. Rivera,"Frequently, students with emotional and behavior disorders (EBD) exhibit academic underachievement combined with high levels of externalizing behaviors and resistance to instructional efforts. Regardless of the present reading initiatives, research focusing on interventions for teaching reading to students with EBD continues to be limited. This article extends previous efforts to review literature concerning reading instruction interventions for students with EBD. Specifically, this review focuses on interventions employed in primary grades. Because of the paucity in research and documented issues related to late and misidentification of students with EBD, studies including students at risk of antisocial behaviors were included. Eleven studies were found and carefully reviewed. Results demonstrate the efficacy of several reading interventions, including Direct Instruction, peer tutoring, and behaviorally based procedures such as time delay prompting, trial and error, and differential reinforcement.",,
-using video prompting to teach cooking skills to secondary students with moderate disabilities.,Tara Graves,"Three secondary students with moderate disabilities acquired cooking skills through a constant time delay procedure used with video prompting. A multiple probe design was used to evaluate effectiveness of the procedure to teach preparation of a food item (a) on a stove, (b) in a microwave, and (c) on a counter top. The procedure was effective for each student. Guidelines for using videotapes in the instruction of functional skills are discussed.",,
-the reading assistant: eye gaze triggered auditory prompting for reading remediation,J. Sibert,We have developed a system for remedial reading instruction that uses visually controlled auditory prompting to help the user with recognition and pronunciation of words. Our underlying hypothesis is that the relatively unobtrusive assistance rendered by such a system will be more effective than previous computer aided approaches. We present a description of the design and implementation of our system and discuss a controlled study that we undertook to evaluate the usability of the Reading Assistant.,"{'model': 'tldr@v2.0.0', 'text': 'A system for remedial reading instruction that uses visually controlled auditory prompting to help the user with recognition and pronunciation of words and a controlled study is discussed that was undertook to evaluate the usability of the Reading Assistant.'}",
-effects of manipulative instruction on solving area and perimeter problems by students with learning disabilities,Mike Cass,"Abstract. A multiple baseline design was employed to test the effect of manipulative instruction on the perimeter and area problem-solving performance of middle and high school students who had been diagnosed with LD in the area of mathematics. Modeling, prompting/guided practice, and independent practice in conjunction with manipulative training were employed to teach both perimeter and area problem-solving skills. Analysis of data revealed that the students rapidly acquired the problem-solving-skills, maintained these skills over a two-month period, and transferred these skills to a paper and pencil problem-solving format. This research extends previous findings by revealing that use of concrete manipulatives promotes the long-term maintenance of skills.",,
-"improving comprehension instruction: rethinking research, theory, and classroom practice.",C. Block,"Tables, Figures, and Exhibits.Foreword (Gerald G. Duffy).Preface.The Editors.The Contributors.Part One: New Directions in Comprehension Instruction.Introduction: Improving Comprehension Instruction: An Urgent Priority (Linda B. Gambrell, Cathy Collins Block, and Michael Pressley).1. Reconceptualizing Reading Comprehension (Anne P. Sweet and Catherine Snow).2. The Thinking Process Approach to Comprehension Development: Preparing Students for Their Future Comprehension Challenges (Cathy Collins Block and Rebecca B. Johnson).3. From Good to Memorable: Characteristics of Highly Effective Comprehension Teaching (Ellin Oliver Keene).4. The Guided Reading Lesson: Explaining, Supporting, and Prompting for Comprehension (Gay Su Pinnell).5. Instructional Components for Promoting Thoughtful Literacy Learning (Pamela J. Dunston).Part Two: New Comprehension Lessons Across the Curriculum.6. Differentiating Reading and Writing Lessons to Promote Content Learning (Karen D. Wood).7. Parsing, Questioning, and Rephrasing (PQR): Building Syntactic Knowledge to Improve Reading Comprehension (James Flood, Diane Lapp, and Douglas Fisher).8. Using Writing to Improve Comprehension: A Review of the Writing-to-Reading Research (Bena R. Hef.in and Douglas K. Hartman).9. Research-Based Comprehension Practices That Create Higher-Level Discussions (Janice F. Almasi).10. Goose Bumps and Giggles: Engaging Young Readers' Critical Thinking with Books from the Teachers' Choices Project and Graphic Organizers (Kathy N. Headley and Jean Keeler).Part Three: Integrating Technology and Innovative Instruction.11. Using Technology to Individualize Reading Instruction (David Rose and Bridget Dalton).12. Computers, Kids, and Comprehension: Instructional Practices That Make a Difference (Linda D. Labbo).13. Out of This World: Cyberspace, Literacy, and Learning (Victoria Gentry Ridgeway, Chris L. Peters, and Terrell Seawell Tracy).14. Reading in the Digital Era: Strategies for Building Critical Literacy (Lisa Patel Stevens and Thomas W. Bean).Part Four: Overcoming Comprehension Challenges.15. Hitting the Wall: Helping Struggling Readers Comprehend (D. Ray Reutzel, Kay Camperell, and John A. Smith).16. At-Risk Students: Learning to Break Through Comprehension Barriers (Lynn Romeo).17. Helping Struggling Readers Make Sense of Reading (Irene W. Gaskins, Sally R. Laird, Colleen O'Hara, Theresa Scott, and Cheryl A. Cress).Conclusion: Improving Comprehension Instruction: A Path for the Future (Michael Pressley).Name Index.Subject Index.","{'model': 'tldr@v2.0.0', 'text': 'Improving Comprehension Instruction: A Path for the Future is a path for the future by Linda B. Gambrell, Cathy Collins Block, and Michael Pressley.'}",
-effectiveness of peer delivered simultaneous prompting on teaching community signs to students with developmental disabilities.,Elif Tekin‐Iftar,"A multiple probe study across behaviors, replicated across participants, assessed effectiveness of peer delivered simultaneous prompting (SP) in teaching expressively identifying community signs to four students with developmental disabilities. The two purposes of the study were: (a) to find out if peer tutors use simultaneous prompting reliably for instructing their tutees with developmental disabilities, and (b) to examine effectiveness of simultaneous prompting on teaching expressively identifying community signs. The definition of community sign was presented as instructive feedback. Besides these aims, generalization and maintenance effects of simultaneous prompting were also investigated in the study. Generalization across persons was tested before introducing simultaneous prompting and after tutees met criteria. Maintenance data for targeted and non-targeted behaviors were collected one week after instruction. Results show that peer tutors deliver SP reliably, and tutees acquired expressively identifying community signs. Tutees also gained some instructive feedback. Furthermore, tutees maintained acquired skills at criterion level and generalized acquired skills to another person at criterion level. Based upon evaluation of findings and implications of the study future research needs are discussed.",,
-training adults with moderate and severe mental retardation in a vocational skill using a simultaneous prompting procedure.,K. Maciag,"A multiple probe across subjects design evaluated the effectiveness of a simultaneous prompting procedure in training 10 adults with moderate and severe mental retardation to construct shipping boxes within a community vocational center. The trainer used a dyadic instructional format in which each member of the dyad received instruction on one half of the task analysis on an alternating basis. Two instructional sessions occurred daily, so each member of the dyad received instruction on the entire task analysis. Four of the five dyads learned to construct shipping boxes. Instructional data are presented and implications for future research are discussed. One of the primary roles of teachers is to only contribute to the growth of the individual promote the acquisition of target skills by uti- but also increase public awareness of the lizing effective and efficient instructional pro- strengths and contributions of individuals cedures. As curriculum goals for students with with mental retardation, moderate and severe disabilities focus on pre- The majority of vocational skills can be de paring these students for the least restrictive scribed as chained tasks; that is, a series of environment and vocational opportunities, behaviors that, when put together, form a teachers must identify instructional proce- more complex skill (Wolery, Ault, & Doyle, dures that are efficient in training vocational 1992). Numerous vocational skills have been tasks. Transition from school to the world of taught successfully in the literature and in work offers individuals the opportunity to de- elude opening a keyed lock (Fetko, Schuster, velop new social networks and increased per- Harley, & Collins, 1999), banking skills (Mc sonal independence. The acquisition of voca- Donnell & Ferguson, 1989), and other voca tional skills enables individuals to experience tional skills (Chandler, Schuster, & Stevens, increased self-esteem by increasing vocational 1993). Several response prompting proce opportunities, earning money, and contribut- dures such as constant and progressive time",,
-identifying teachers’ supports of metacognition through classroom talk and its relation to growth in conceptual learning,Cristina D. Zepeda,"A gulf exists between prior work testing metacognitive instructional interventions and teacher practices that may support metacognition in the classroom. To help bridge this gulf, we designed an observational protocol to capture whether and how teachers provide metacognitive support in their talk and examined whether these supports were related to student learning. We examined four features of metacognitive support, including the type of metacognitive knowledge supported (personal, strategy, or conditional), the type of metacognitive skill supported (planning, monitoring, or evaluating) the type of instructional manner in which the support was delivered (directives, prompting, or modeling), and the type of framing (problem specific, problem general, or domain general), during three types of instructional activities (individual, group, or whole-class instruction). We compared teacher talk from 20 middle school mathematics classrooms with high growth in conceptual mathematics scores with teacher talk from 20 classrooms with low growth. For each of these classrooms, we examined the amount of teacher talk that supported metacognition during one regular class period. Observations revealed that the high-conceptual growth classrooms had more metacognitive supports for personal knowledge, monitoring, evaluating, directive manners, and domain-general frames than the low-conceptual growth classrooms. We discuss the implications of those observations for bridging research on metacognition to teacher practice.",,
-response prompting and fading methods: a review.,M. Demchak,"Teachers of individuals with severe handicaps typically use response prompts, such as verbal instructions, modeling, and physical guidance, to encourage correct responding. However, the goal of instruction is to have the individual respond to natural stimuli rather than response prompts. Therefore, these response prompts must be faded. Currently there are four methods for systematically fading response prompts: increasing assistance, decreasing assistance, graduated guidance, and time delay. An overview of each method as well as a review of comparative investigations involving these methods of fading response prompts was presented. Recommendations for practitioners and for future research were included.","{'model': 'tldr@v2.0.0', 'text': 'An overview of each method as well as a review of comparative investigations involving these methods of fading response prompts was presented and Recommendations for practitioners and for future research were included.'}",
-the effect of metacognitive instruction in outlining and graphic organizer construction on students' comprehension in a tenth-grade world history class,Thomas W. Bean,"The purpose of the present study was to explore the effect of graphic organizer instruction versus outlining on students' text recall in tenth-grade world history. Based on the hierarchical structure of graphic organizers depicting interrelationships among ideas, we hypothesized that students instructed in the use of graphic organizers would display significantly higher text recall than a group using outlines. Seventy-two tenth graders in three sections of world history participated in the study. Two sections received systematic instruction in the development of graphic organizers and a third section received parallel instruction in outlining. One of the two graphic organizer sections had previous instruction in summarization and question generation during an earlier study. Six 15-item multiple-choice quizzes comprised one of the dependent measures. On the first five quizzes there were no significant differences between the three sections. However, when teacher prompting was phased-out on quiz six, the graphic organizer group with previous training in summarization significantly outperformed the other two groups. Additionally, on a second dependent measure involving the development of a written recall protocol based on difficult college level text, the group with previous summarization experience significantly outperformed both the new graphic organizer group and the outlining section. Implications for a cumulative, long-range view of metacognitive instruction are discussed.",,
-comparison of simultaneous prompting with and without error correction in teaching science vocabulary words to high school students with mild disabilities,Patti Johnson,,,
-the effects of simultaneous prompting on teaching expressive identification of objects: an instructive feedback study,Elif Tekin‐Iftar,"The present study examined whether the use of a simultaneous prompting procedure would result in an increase in the percentage of correct responses when expressively identifying first aid materials. A multiple probe design across behaviours and replicated across students was used. Three training sets with a total of nine first aid materials were presented to three students. Also, instructive feedback stimuli were presented during consequent events to increase the efficiency of instruction. The instructive feedback con tained the functions of the instructional materials. Maintenance effects were assessed 1, 2, and 4 weeks after training. The results showed that all students learned expressive identification of first aid materials and maintained them after training. Furthermore, all students acquired and maintained some of the instructive feedback stimuli presented to them during instruction.",,
-enhancing choice and participation for adults with severe disabilities in community-based instruction,K. Cooper,"Three adults with severe disabilities received instructional support to make purchases in fast food restaurants. During baseline and intervention, participants received discrete, systematic prompting and feedback to perform the steps of the purchasing task analysis. During intervention, participants were prompted to make five distinct choices during the community routine. This prompting resulted not only in increased choicemaking, but also in collateral increases in the performance of the purchasing task analysis.","{'model': 'tldr@v2.0.0', 'text': 'Three adults with severe disabilities received instructional support to make purchases in fast food restaurants and received discrete, systematic prompting and feedback to perform the steps of the purchasing task analysis.'}",
-the evolution of transactional strategies instruction in one teacher's classroom,R. Brown,"In this study, 1 teacher taught the same story for 3 consecutive years to 3 comparable groups of second-grade, low-achieving students. An interaction-tracking and coding scheme was used to analyze the 45-minute lessons for changes in interactional patterns, participation among group members, instructional focus, strategy instruction, and prompted and self-regulated use of strategies by students. By year 3, students participated more actively in story discussion and used strategies with less teacher prompting to support their interpretations and responses to text. These changes appeared to occur because of modifications in the teacher's instructional practices. The instruction that emerged during the third year could be characterized as transactional strategies instruction, an approach that involves teaching reading group members to use comprehension strategies as they jointly construct interpretations of text.","{'model': 'tldr@v2.0.0', 'text': 'The instruction that emerged during the third year could be characterized as transactional strategies instruction, an approach that involves teaching reading group members to use comprehension strategies as they jointly construct interpretations of text.'}",
-divergent thinking as a function of time and prompting to “be creative” in undergraduates,G. Johns,"This study examined divergent thinking in undergraduates as a function of gender, time, and test instruction conditions. Differences were examined in male and female undergraduates when tested on a divergent thinking test with or without prompting to “be creative” and with or without time limitations imposed. Subject groups were assigned to one of the following four conditions: standard instructions/four minute time limit; standard instructions/no time limit; creative instructions/four minute time limit; and creative instructions/no time limit. The no time limit condition resulted in significantly higher mean traditional fluency and flexibility divergent production scores. Males were found to have significantly higher mean divergent production scores on fluency. This study supported previous research findings regarding the lack of enhanced traditional divergent production scores for females resulting from creative instructions.",,
-developing joint attention for children with autism in robot-enhanced therapy,D. David,,"{'model': 'tldr@v2.0.0', 'text': 'Using more cues for prompting JA increases the performance of the children and emphasizes the importance of using more cues, such as pointing, for increasing engagement and performance engagement in a child–robot interaction session.'}",
-community-based instruction for independent meal preparation by adults with profound mental retardation,Nirbhay N. Singh,,"{'model': 'tldr@v2.0.0', 'text': 'Assessment, training and followup data on the preparation of a dessert by three adults with profound mental retardation and very limited reading skills transitioned to community living showed that they were able to independently prepare the dessert in various community settings of their choice.'}",
-a review and critical examination of the system of least prompts,Collin Shepley,"We systematically reviewed more than 25 years of applied research examining the system of least prompts response prompting procedure with individuals with disabilities. We identified 123 peer-reviewed studies including 413 participants receiving instruction with the system of least prompts. A total of 252 experimental designs were evaluated, with 51 designs indicating a functional relation and the presence of 154 demonstrations of effect across 91 individuals. Our data indicate that the system of least prompts is an evidenced-based practice for teaching chained responses related to community, self-care, and vocational skills to individuals with moderate intellectual disability who are 13 years of age or older. In addition, we present and discuss a method for analyzing and aggregating data from single-case studies to account for noneffects and publication bias when identifying if an intervention meets standards as an evidence-based practice.",,
-to err or not to err: a question for the instruction of psychomotor skills,R. Singer,"s International, 1969, No. 69-20, 790, p. 2385-A. Prather, D. C. Trial-and-error versus errorless learning: Training, transfer and stress. American Journal of Psychology, 1971, 84, 377-386. Prather, D. C., & Berry, G. A. Comparison of trial-and-error versus highly prompted learning of a perceptual skill. Proceedings of the 78th Annual Convention of the American Psychological Association, 1970, 5, 677-678. Prather, D. C., Berry, G. A., & Bermudez, J. M. Effect of prompting and feedback on performance during learning, stress, and transfer of a perceptual skill. Proceedings of the 80th Annual Convention of the American Psychological Association, 1972, 7, 643-644. Schmidt, R. A. A schema theory of discrete motor skill learning. Psychological Review, 1975, 82, 225-260. Schurr, E. Movement experiences for children. New York: Appleton-Century- Crofts, 1967. Seidel, R. J. Theories and strategies related to measurement in individualized instruction (HumRRO Tec. Rep. 2-71). Alexandria, Va.: Human Resources Research Organization, March 1971. Shulman, L., & Keislar, E. Learning by discovery-a critical appraisal. Chicago: Rand McNally, 1966. Singer, R. N. Coaching, athletics, and psychology. New York: McGraw-Hill, 1972. Singer, R. N. Motor Learning and human performance. (2nd ed.) New York: Macmillan, 1975. Singer, R. N., & Dick, W. Teaching physical education: A systems approach. Boston: Houghton Mifflin, 1974. Singer, R. N., & Gaines, L. Effect of prompted and trial-and-error learning on transfer performance of a serial motor task. American Educational Research Journal, 1975, 12, 395-403. Singer, R. N., & Pease, D. The effect of different instructional strategies on",,
-the effects of strong formal prompts in programed instruction,Richard C. Anderson,"Authors of self-instructional programs are often cautioned to avoid ""overprompting."" I t has never been demonstrated experimentally that ""overprompting"" inhibits learning. In fact, there is surprisingly little research on prompting considering the lengthy discussions and stronglystated opinions which appear in treatises on programed instruction (Anderson, 1967). Current conceptions of prompting seem to be based largely on Skinner's (1957) speculative analysis of verbal behavior and the practical experience of programers (Markle, 1964). Most of the actual research on prompting has involved paired associate lists. I t has been repeatedly demonstrated that people learn faster under a prompting procedure, in which both the stimulus term and response term appear before the response is required, than under the anticipation method, or confirmation method, as it has been called in these studies (Cook and Spitzer, 1960; Sidowski, 1961; Levine, 1965). On the basis of these experiments, Cook (1963) has argued that the student should be shown the correct answer before he makes the response. Cook seems to be suggesting that the ""copying frame"" is an especially effective sort of teaching device. The further implication is that there is no such condition as ""overprompting."" Our contention is that copying frames and, presumably, under some conditions, other kinds of strongly-prompted frames as well, do have a serious shortcoming. The following is an obvious copying frame from a program on the menstrual cycle (Biological Sciences Curriculum Study, 1965, p. 11):",,
-strengthening scientific verbal behavior: an experimental comparison of progressively prompted and unprompted programmed instruction and prose tutorials.,Darrel R. Davis,"Web-based software was used to deliver and record the effects of programmed instruction that progressively added formal prompts until attempts were successful, programmed instruction with one attempt, and prose tutorials. Error-contingent progressive prompting took significantly longer than programmed instruction and prose. Both forms of programmed instruction substantially increased the appropriate use of behavioral vocabulary during subsequent interpretive essays. These behavioral gains extended to a different setting, suggesting that more was being learned than simply how to answer programmed tutorial frames correctly.","{'model': 'tldr@v2.0.0', 'text': 'Both forms of programmed instruction substantially increased the appropriate use of behavioral vocabulary during subsequent interpretive essays, suggesting that more was being learned than simply how to answer programmed tutorial frames correctly.'}",https://europepmc.org/articles/pmc1868811?pdf=render
-promoting preservice teachers’ dual self-regulation roles as learners and as teachers: effects of generic vs. specific prompts,B. Kramarski,,,
-composition instruction with learning disabled students: self-instructional strategy training.,S. Graham,"Writing is a particularly complex and demanding task. The development of skillful writing requires much more than simply adding special knowledge and skills to existing oral language abilities (Scardamalia & Bereiter ,, 1986). The beginning writer must develop the ability to generate language in the absence of a conversational partner, activate relevant memories without prompting, generate units of text larger than typically included in a conversational turn, and examine what is generated from the perspectives of both the writer and the reader (Bereiter & Scardamalia, 1983). Furthermore, the beginning writer must learn to monitor and regulate a variety of mental operations including the generation of ideas, development of plans for writing, and the actual production of text that corresponds to the conventions of standard written English. The developing writer, in sum, must learn to orchestrate a variety of cognitive activities, and attention must be allocated to both mechanical and substantive concerns in an efficient, effective manner (Bereiter & Scardamalia, 1982; Graham, 1982; Scardamalia & Bereiter, 1986). One way to help poor and beginning writers overcome the difficulties inherent in the composing process is to teach them task-specific and metacognitive strategies for regulating their writing behavior (Graham & Harris, 1987a; Harris & Graham, in press). Such an approach may be particularly helpful because strategy instruction can be employed as a means for helping students gain more security and competence in the processes considered central to effective writing (e.g., planning, revising). Furthermore, strategy training holds much promise as a means for helping students develop the skills necessary to carry out more mature and complex composing behaviors (Graham & Harris, in press-a, in press-b). Strategy and self-regulation training may be especially useful as a means for helping learning disabled (LD) students improve their writing performance. Two assumptions underlying strategy instruction are (Graham & Harris, in press-b; Harris, 1982):",,https://journals.ku.edu/focusXchild/article/download/7502/6835
-evidence-based practice for teaching academics to students with severe developmental disabilities,F. Spooner,"A review of the literature was conducted for articles published between 2003 and 2010 to build a case for the degree to which evidence-based practices were documented for teaching academic skills to students with severe developmental disabilities. This review extended earlier comprehensive work in literacy, mathematics, and science for the population in question. A total of 18 studies met the Horner et al. (2005) quality indicator criteria. In general, time delay and task analytic instruction were found to be evidence-based practices. In addition, specific target responses were defined to show academic learning, with the most prevalent target responses being discrete responses; the type of systematic prompting and feedback used most often was time delay, while the component used least often was stimulus fading/shaping; and teaching formats used most often were massed trials and one-to-one instruction.",,
-strategy instruction is often insufficient: addressing the interdependency of executive and attributional processes,Andrea K. Groteluschen,,,
-teaching generalized pretend play and related behaviors to young children with disabilities,E. Barton,"Children with disabilities play less often and demonstrate fewer varied pretend play behaviors than children with typical development. A multiple-probe design was used to examine the relation between teachers’ use of the system of least prompts and contingent imitation and the acquisition, maintenance, and generalization of pretend play and related behaviors by four children with disabilities. Results indicated the teachers’ use of the intervention package was functionally related to increases in the children’s frequency and diversity of pretend play and related behaviors. Children also maintained responses in sessions without prompts and generalized across toys and contexts. The findings replicate previous studies on adult prompting of pretend play and extend the literature by assessing generalization of children’s pretend play across contexts and measuring intervention and implementation fidelity. Overall, this study provides a strong argument for engaging in systematic instruction of play, including pretend play, for children who do not display such behaviors.",,
-behavioral ethics and teaching ethical decision making,M. Drumwright,"Business education often renders students less likely to act ethically. An infusion of liberal learning in the form of behavioral ethics could improve this situation by prompting students to develop higher levels of professionalism that encompass ethics, social responsibility, self-critical reflection, and personal accountability. More specifically, teaching behavioral ethics, which draws upon psychology, sociology, and related fields, can improve students’ ethical decision making in a manner that can lead to a more ethical climate in organizations and in society more generally. This article introduces key concepts of behavioral ethics, argues that teaching behavioral ethics can have a positive impact, discusses materials that can be used to teach those concepts, and addresses action-research approaches to assessing the effectiveness of the instruction. There is significant evidence, though preliminary and incomplete, that teaching behavioral ethics is a promising new approach for improving the ethicality of students’ decisions and actions.",,
-use of technology in interventions for children with autism,T. Goldsmith,"A growing number of studies have investigated diverse applications of technology-based interventions with children with autism. The purpose of this paper is to review the growing empirical support for the efficacy of technology-based interventions with children with autism and to recommend future directions for research. This review will focus on five examples of technology introduced as a temporary instructional aid to be removed once the goal of behavior change has been met: (a) tactile and auditory prompting devices, (b) video-based instruction and feedback, (c) computer-aided instruction, (d) virtual reality, and (e) robotics. Future directions for research and practice with each technology are discussed. Keywords: autism; technology-based interventions; computer-aided instruction; virtual reality; robotics. ********** A growing number of studies have investigated diverse applications of technology-based interventions with children with autism. The popularity of technology in the field of psychology is evidenced by the development of new journals in the area (e.g., the Journal of Special Education Technology, the Journal of Educational Multimedia and Hypermedia, the Journal of Computer Assisted Learning, etc.). In addition, clinical psychology journals are recognizing the importance of technology in service delivery and devoting special issues to the topic (e.g., Newman, 2004). Parents and clinicians regularly report that children with autism are drawn to technological devices and researchers have noted the importance of devising treatments that take advantage of this fascination (Colby, 1973). The purpose of this paper is to review the growing empirical support for the efficacy of technology-based interventions with children with autism and to recommend future directions for research. A review of technology-based interventions must first clarify the use of the word technology. The term is broadly defined as ""the practical application of knowledge"" or ""the specialized aspects of a particular field of endeavor"" (Merriam-Webster's Collegiate Dictionary, 1994, p. 1210). This broad definition encompasses virtually any information or object that has been used in application to a field of study. For example, the application of stimulus control is technology to a behavioral clinician. A more specific use of the term refers to use of mechanical or electromechanical processes that often increase productivity and reduce or eliminate manual operations or operations done by older technologies. In the twenty-first century, technology commonly denotes a variety of popular electromechanical devices such as cell phones, video recording equipment, and hand-held, desktop, and laptop personal computers. It is the use of these tools with children with autism that will be the subject of this review paper. Some technology-based interventions are designed for indefinite use as an assistive tool (e.g., voice-output augmentative communication devices, microswitches, etc.) while others are introduced as a temporary instructional aid to be removed once the goal of behavior change has been met. This review will focus on five examples of the second category: tactile and auditory prompting devices, video-based instruction and feedback, computer-aided instruction, virtual reality, and robotics. Research support for the use of each technological advancement with children with autism will be reviewed, and directions for future research and practical application will be discussed. Mechanical Prompts Individuals with autism often need external stimulus prompts to initiate, maintain, or terminate a behavior. Commonly used prompts include vocal, gestural, physical, written/pictorial, and signed prompts, and each modality has been demonstrated effective for multiple purposes (MacDuff, Krantz, & McClannahan, 2001). Technological advancements in the last decade have created cost-effective automated prompting devices with the ability to deliver the same level of prompting with less human interaction and obtrusiveness and often less human effort in managing prompt delivery (e. …","{'model': 'tldr@v2.0.0', 'text': 'This review will focus on five examples of technology introduced as a temporary instructional aid to be removed once the goal of behavior change has been met: tactile and auditory prompting devices, video-based instruction and feedback, computer-aided instruction, virtual reality, and robotics.'}",
-effects of self-explaining on learning and transfer of critical thinking skills,Lara M. van Peppen,"Critical thinking is considered to be an important competence for students and graduates of higher education. Yet, it is largely unclear which teaching methods are most effective in supporting the acquisition of critical thinking skills, especially regarding one important aspect of critical thinking: avoiding biased reasoning. The present study examined whether creating desirable difficulties in instruction by prompting students to generate explanations of a problem-solution to themselves (i.e., self-explaining) is effective for fostering learning and transfer of unbiased reasoning. Seventy-nine first-year students of a Dutch Applied University of Sciences were first instructed on two categories of “heuristics and biases” tasks (syllogism and base-rate or Wason and conjunction). Thereafter, they practiced these either with (self-explaining condition) or without (no self-explaining condition) self-explanation prompts that asked them to motivate their answers. Performance was measured on a pretest, immediate posttest, and delayed (2 weeks later) posttest on all four task categories, to examine effects on learning (performance on practiced tasks) and transfer (performance on non-practiced tasks). Participants’ learning and transfer performance improved to a comparable degree from pretest to immediate posttest in both conditions, and this higher level of performance was retained on the delayed posttest. Surprisingly, self-explanation prompts had a negative effect on posttest performance on practiced tasks when those were Wason and conjunction tasks, and self-explaining had no effect on transfer performance. These findings suggest that the benefits of explicit instruction and practice on learning and transfer of unbiased reasoning cannot be enhanced by increasing the difficulty of the practice tasks through self-explaining.",,https://www.frontiersin.org/articles/10.3389/feduc.2018.00100/pdf
-"feedback, prompting, and overt correction procedures in nonbranching computer assisted instructed programs",D. Gilman,"AbstractNo differences in learning or retention were obtained for a computer assisted instruction program (CAI) which incorporated response contingent feedback, prompting, and overt correction procedures when compared to a CAI program which simply typed the correct response following a student response. No differences in learning or retention were obtained for a condition in which an instructional program was administered by a teletypewriter as compared to a condition in which the material was presented by means of programmed texts. Both conditions in which instruction was presented by a CAI communication device took significantly more instructional time than the programmed text condition.",,
-improving students' critical thinking : empirical support for explicit instructions combined with practice,A. Heijltjes,"This experiment investigated the impact of different types of critical thinking instruction and dispositions on bias in economics students' (N=141) reasoning performance. The following conditions were compared: (A) implicit instruction; (B) implicit instruction with practice; (C) implicit instruction with explicit instruction and practice; (D) implicit instruction with explicit instruction, practice, and self-explanation prompts; and (E) implicit instruction with explicit instruction, practice, and activation prompts. Results showed that explicit instruction combined with practice is required to improve critical thinking (i.e., conditions A/B’ and rest of the layers are initialized randomly (row-3). Initializing prompts with a similar template in all layers leads to lower performance suggesting that this is redundant as these prompts learn hierarchically different contextual concepts in different layers (row-1). However, complete random initialization of prompts provides competitive performance (row-2). For implementation, if the number of learnable prompts M = #P are less than the total tokens of initial prompt template, we convert the former M word embeddings of template with learnable prompts and consider the rest of word embeddings of prompt template as xed and use all token embeddings (learnable prompts + xed word tokens) as input to text encoder.","{'model': 'tldr@v2.0.0', 'text': 'For implementation, if the number of learnable prompts M = #P are less than the total tokens of initial prompt template, the authors convert the former M word embeddings of template with learnable prompt and consider the rest of word embedDings of prompt template as \ue000xed and use all token embeddeddings (learnable prompts + \ue001xed word tokens) as input to text encoder.'}",
-this prompt is measuring : evaluating bias evaluation in language models,Seraphina Goldfarb-Tarrant,"Bias research in NLP seeks to analyse models for social biases, thus helping NLP practitioners uncover, measure, and mitigate social harms. We analyse the body of work that uses prompts and templates to assess bias in language models. We draw on a measurement modelling framework to create a taxonomy of attributes that capture what a bias test aims to measure and how that measurement is carried out. By applying this taxonomy to 90 bias tests, we illustrate qualitatively and quantitatively that core aspects of bias test conceptualisations and operationalisations are frequently unstated or ambiguous, carry implicit assumptions, or be mismatched. Our analysis illuminates the scope of possible bias types the field is able to measure, and reveals types that are as yet under-researched. We offer guidance to enable the community to explore a wider section of the possible bias space, and to better close the gap between desired outcomes and experimental design, both for bias and for evaluating language models more broadly.",,http://arxiv.org/pdf/2305.12757
-prompt tuning with contradictory intentions for sarcasm recognition,Yiyi Liu,"Recently, prompt tuning has achieved promising results in a variety of natural language processing (NLP) tasks. The typical approach is to insert text pieces (i.e. templates) into the input and transform downstream tasks into the same form as pre-training. In essence, a high-quality template is the foundation of prompt tuning to support the performance of the converted cloze-style task. However, for sarcasm recognition, it is time-consuming and requires increasingly sophisticated domain knowledge to determine the appropriate templates and label words due to its highly figurative nature. In this work, we propose SarcPrompt, to incorporate the prior knowledge about contradictory intentions into prompt tuning for sarcasm recognition. SarcPrompt is inspired by that the speaker usually says the opposite of what they actually mean in the sarcastic text. Based on this idea, we explicitly mimic the actual intention by prompt construction and indicate whether the actual intention is contradictory to the literal content by verbalizer engineering. Experiments on three public datasets with standard and low-resource settings demonstrate the effectiveness of our SarcPrompt for sarcasm recognition.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes SarcPrompt, to incorporate the prior knowledge about contradictory intentions into prompt tuning for sarcasm recognition, inspired by that the speaker usually says the opposite of what they actually mean in the sarcastic text.'}",https://aclanthology.org/2023.eacl-main.25.pdf
-grammar correction for multiple errors in chinese based on prompt templates,Zhici Wang,"Grammar error correction (GEC) is a crucial task in the field of Natural Language Processing (NLP). Its objective is to automatically detect and rectify grammatical mistakes in sentences, which possesses immense application research value. Currently, mainstream grammar-correction methods primarily rely on sequence labeling and text generation, which are two kinds of end-to-end methods. These methods have shown exemplary performance in areas with low error density but often fail to deliver satisfactory results in high-error density situations where multiple errors exist in a single sentence. Consequently, these methods tend to overcorrect correct words, leading to a high rate of false positives. To address this issue, we researched the specific characteristics of the Chinese grammar error correction (CGEC) task in high-error density situations. We proposed a grammar-correction method based on prompt templates. Firstly, we proposed a strategy for constructing prompt templates suitable for CGEC. This strategy transforms the CGEC task into a masked fill-in-the-blank task compatible with the masked language model BERT. Secondly, we proposed a method for dynamically updating templates, which incorporates already corrected errors into the template through dynamic updates to improve the template quality. Moreover, we used the phonetic and graphical resemblance knowledge from the confusion set as guiding information. By combining this with BERT’s prediction results, the model can more accurately select the correct characters, significantly enhancing the accuracy of the model’s prediction correction results. Our methods were validated through experiments on a public grammar-correction dataset. The results indicate that our method achieves higher correction performance and lower false correction rates in high-error density scenarios.","{'model': 'tldr@v2.0.0', 'text': 'This work proposed a grammar-correction method based on prompt templates that achieves higher correction performance and lower false correction rates in high-error density scenarios and used the phonetic and graphical resemblance knowledge from the confusion set as guiding information.'}",https://www.mdpi.com/2076-3417/13/15/8858/pdf?version=1690869487
-ppm: prompt-free prompt-tuning for multi-task learning,Zihan Zeng,"The cost of fine-tuning has increased significantly in recent years as the size of language model parameters has increased. Prompt-tuning and adapters have made it possible to train models with a small number of parameters to obtain results similar to those of fine-tuning methods. However, most of the current prompt-tuning methods require the help of hand-crafted templates and verbalizers to achieve outstanding results in few-shot learning. In this work, we propose PPM, Prompt-free prompt-tuning for multi-task learning. First, we insert the task-specific adapter into the pre-trained language model to replace the hand-designed external template. Then, we train each adapter separately on different tasks and adjust the parameters of each adapter layer. Next, we combine the different adapters and draw on their valid knowledge by tuning the parameters of the fusion part to get the smallest loss function in the process of extracting knowledge from different adapters. To boost the training speed, we use Post-LN to replace Pre-LN, which switched the position of the Laynorm layer in the model from after the two Addition layers to before the FFN layer and the Multi-head Attention layer. Experimental results on different NLP tasks show that our model has better synergistic effects on diverse types of downstream tasks.","{'model': 'tldr@v2.0.0', 'text': 'PPM, Prompt-free prompt-tuning for multi-task learning, which inserts the task-specific adapter into the pre-trained language model to replace the hand-designed external template and combines the different adapters to draw on their valid knowledge.'}",
-teprompt: task enlightenment prompt learning for implicit discourse relation recognition,Wei Xiang,"Implicit Discourse Relation Recognition (IDRR) aims at classifying the relation sense between two arguments without an explicit connective. Recently, the ConnPrompt~\cite{Wei.X:et.al:2022:COLING} has leveraged the powerful prompt learning for IDRR based on the fusion of multi-prompt decisions from three different yet much similar connective prediction templates. Instead of multi-prompt ensembling, we propose to design auxiliary tasks with enlightened prompt learning for the IDRR task. Although an auxiliary task is not used to directly output final prediction, we argue that during the joint training some of its learned features can be useful to boost the main task. In light of such motivations, we propose a task enlightenment prompt learning model, called TEPrompt, to fuse learned features from three related tasks for IDRR. In particular, the TEPrompt contains three tasks, viz., Discourse Relation Recognition (DRR), Sense Semantics Classification (SSC) and Annotated Connective Prediction (ACP), each with a unique prompt template and an answer space. In the training phase, we jointly train three prompt learning tasks with shared argument representation. In the testing phase, we only take the DRR output with fused features as the final IDRR decision. Experiments with the same conditions have shown that the proposed TEPrompt outperforms the ConnPrompt. This can be attributed to the promoted decision features and language models benefited from joint-training of auxiliary tasks.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a task enlightenment prompt learning model, called TEPrompt, to fuse learned features from three related tasks for IDRR, and shows that the proposed TEP Prompt outperforms the ConnPrompt.'}",http://arxiv.org/pdf/2305.10866
-cover: a heuristic greedy adversarial attack on prompt-based learning in language models,Zihao Tan,"Prompt-based learning has been proved to be an effective way in pre-trained language models (PLMs), especially in low-resource scenarios like few-shot settings. However, the trustworthiness of PLMs is of paramount significance and potential vulnerabilities have been shown in prompt-based templates that could mislead the predictions of language models, causing serious security concerns. In this paper, we will shed light on some vulnerabilities of PLMs, by proposing a prompt-based adversarial attack on manual templates in black box scenarios. First of all, we design character-level and word-level heuristic approaches to break manual templates separately. Then we present a greedy algorithm for the attack based on the above heuristic destructive approaches. Finally, we evaluate our approach with the classification tasks on three variants of BERT series models and eight datasets. And comprehensive experimental results justify the effectiveness of our approach in terms of attack success rate and attack speed.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a prompt-based adversarial attack on manual templates in black box scenarios by designing character-level and word-level heuristic approaches to break manual templates separately and presents a greedy algorithm for the attack based on the above heuristic destructive approaches.'}",https://arxiv.org/pdf/2306.05659
-self-adaptive prompt-tuning for event extraction in ancient chinese literature,Jingyi Zhang,"Extracting different types of war events from ancient Chinese literature is significant, as war is an important factor in driving the development of Chinese history. The existing trend of event extraction models utilizes template-based generative approaches, which do not take into account the brevity and obscurity of ancient Chinese, as well as the diversity of templates for similar event types. In this paper, we propose a novel Knowledge Graph-based generative event extraction framework with a self-Adaptive Prompt (KGAP) for ancient Chinese war. Specifically, we construct a self-adaptive prompt, which considers its unique trigger words for different types of wars and is designed to solve the problem of the similarity in events. Moreover, we construct a semantic knowledge graph of ancient literature, assisting the pre-trained language model to better understand the ancient Chinese text. Since there is no public dataset for the ancient Chinese event extraction task, we provide an event extraction dataset and conduct experiments on it. Experimental results show that our model is more state-of-the-art than both the classification-based and generative-based methods for event extraction in ancient Chinese literature.","{'model': 'tldr@v2.0.0', 'text': 'A novel Knowledge Graph-based generative event extraction framework with a self-Adaptive Prompt (KGAP) for ancient Chinese war and a semantic knowledge graph of ancient literature is constructed, assisting the pre-trained language model to better understand the ancient Chinese text.'}",
-sptnet: span-based prompt tuning for video grounding,Yiren Zhang,"When a Pre-trained Language Model (PLM) is adopted in video grounding task, it usually acts as a text encoder without having its knowledge fully utilized. Also, there exists an inconsistency problem between the pre-training and downstream objectives. To solve the issues, we propose a new paradigm, named Span-based Prompt Tuning (SPTNet). It can convert the video grounding task into a cloze form. Specifically, a query is first changed into a form with mask token by a template, then the video and the query embeddings are integrated through a cross-modal transformer. The start and end points of the query matching time span are predicted with the embedding of the mask token. Experimental results on two public benchmarks ActivityNet Captions and Charades-STA show that our SPTNet achieves surpassing performance compared with state-of-the-art methods.","{'model': 'tldr@v2.0.0', 'text': 'A new paradigm, named Span-based Prompt Tuning (SPTNet), which can convert the video grounding task into a cloze form and achieves surpassing performance compared with state-of-the-art methods.'}",
-prompt cache: modular attention reuse for low-latency inference,In Gim,"We present Prompt Cache, an approach for accelerating inference for large language models (LLM) by reusing attention states across different LLM prompts. Many input prompts have overlapping text segments, such as system messages, prompt templates, and documents provided for context. Our key insight is that by precomputing and storing the attention states of these frequently occurring text segments on the inference server, we can efficiently reuse them when these segments appear in user prompts. Prompt Cache employs a schema to explicitly define such reusable text segments, called prompt modules. The schema ensures positional accuracy during attention state reuse and provides users with an interface to access cached states in their prompt. Using a prototype implementation, we evaluate Prompt Cache across several LLMs. We show that Prompt Cache significantly reduce latency in time-to-first-token, especially for longer prompts such as document-based question answering and recommendations. The improvements range from 8x for GPU-based inference to 60x for CPU-based inference, all while maintaining output accuracy and without the need for model parameter modifications.","{'model': 'tldr@v2.0.0', 'text': None}",
-low-resource multi-granularity academic function recognition based on multiple prompt knowledge,Jiawei Liu,"Fine-tuning pre-trained language models (PLMs), e.g., SciBERT, generally requires large numbers of annotated data to achieve state-of-the-art performance on a range of NLP tasks in the scientific domain. However, obtaining the fine-tune data for scientific NLP task is still challenging and expensive. Inspired by recent advancement in prompt learning, in this paper, we propose the Mix Prompt Tuning (MPT), which is a semi-supervised method to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks with a small number of labeled examples. Specifically, the proposed method provides multi-perspective representations by combining manual prompt templates with automatically learned continuous prompt templates to help the given academic function recognition task take full advantage of knowledge in PLMs. Based on these prompt templates and the fine-tuned PLM, a large number of pseudo labels are assigned to the unlabeled examples. Finally, we fine-tune the PLM using the pseudo training set. We evaluate our method on three academic function recognition tasks of different granularity including the citation function, the abstract sentence function, and the keyword function, with datasets from computer science domain and biomedical domain. Extensive experiments demonstrate the effectiveness of our method and statistically significant improvements against strong baselines. In particular, it achieves an average increase of 5% in Macro-F1 score compared with fine-tuning, and 6% in Macro-F1 score compared with other semi-supervised method under low-resource settings. In addition, MPT is a general method that can be easily applied to other low-resource scientific classification tasks.","{'model': 'tldr@v2.0.0', 'text': 'The proposed Mix Prompt Tuning (MPT) is a semi-supervised method to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks with a small number of labeled examples.'}",http://arxiv.org/pdf/2305.03287
-promptcl: improving event representation via prompt template and contrastive learning,Yubo Feng,,,
-learning historical status prompt for accurate and robust visual tracking,Wenrui Cai,"Most trackers perform template and search region similarity matching to find the most similar object to the template during tracking. However, they struggle to make prediction when the target appearance changes due to the limited historical information introduced by roughly cropping the current search region based on the predicted result of previous frame. In this paper, we identify that the central impediment to improving the performance of existing trackers is the incapacity to integrate abundant and effective historical information. To address this issue, we propose a Historical Information Prompter (HIP) to enhance the provision of historical information. We also build HIPTrack upon HIP module. HIP is a plug-and-play module that make full use of search region features to introduce historical appearance information. It also incorporates historical position information by constructing refined mask of the target. HIP is a lightweight module to generate historical information prompts. By integrating historical information prompts, HIPTrack significantly enhances the tracking performance without the need to retrain the backbone. Experimental results demonstrate that our method outperforms all state-of-the-art approaches on LaSOT, LaSOT ext, GOT10k and NfS. Futhermore, HIP module exhibits strong generality and can be seamlessly integrated into trackers to improve tracking performance. The source code and models will be released for further research.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a Historical Information Prompter (HIP) to enhance the provision of historical information and builds HIPTrack upon HIP module, a plug-and-play module that make full use of search region features to introduce historical appearance information.'}",
-prompt template construction by average gradient search with external knowledge for aspect sentimental analysis,Yongping Du,,,
-teaching story prompt writing through wh-questions cube among year 5 students,Daisy Kunang,"Today’s education field in Malaysia has undergone various transformations in line with progressive global educational trends. Henceforth, revolution of teaching and learning styles, methods and approaches are essential to ensure that students’ learning, skills and thinking are optimized and concurrently nurture motivated and confident learners starting from primary school level. Numerous studies on teaching and learning writing had highlighted that it is one of the most challenging skills to be mastered in English language. With respect to this, this study aims to explore the necessity of incorporating WH-Foldable Cube to facilitate expansion of ideas in teaching story prompt lessons. WH-Foldable Cube is a writing learning strategy which emphasised on expansion of ideas through the use of question words to facilitate generation of ideas in writing story prompt. Expansion of ideas becomes less strenuous through students’ engagement with the question words that are used in the WH-Foldable Cube template to stimulate their thinking about the pictures stimulus given. The data were collected using writing tasks and interview which were administered among five Year 5 students in SK Parish, a suburban school in Balingian district. The findings of the study revealed that majority of the respondents agreed that the use of WH-Foldable Cube had positive impacts towards their story prompt writing to generate and expand their ideas. This research also highlighted the potential use of WH-Foldable Cube to expand students’ motivation and confidence in their learning.",,
-activation of actuating hydrogels with ws2 nanosheets for biomimetic cellular structures and steerable prompt deformation.,Lu Zong,"Macroscopic soft actuation is intrinsic to living organisms in nature, including slow deformation (e.g., contraction, bending, twisting, and curling) of plants motivated by microscopic swelling and shrinking of cells, and rapid motion of animals (e.g., deformation of jellyfish) motivated by cooperative nanoscale movement of motor proteins. These actuation behaviors, with an exceptional combination of tunable speed and programmable deformation direction, inspire us to design artificial soft actuators for broad applications in artificial muscles, nanofabrication, chemical valves, microlenses, soft robotics, etc. However, so far artificial soft actuators have been typically produced on the basis of poly(N-isopropylacrylamide) (PNiPAM), whose deformation is motived by volumetric shrinkage and swelling in analogue to plant cells, and exhibits sluggish actuation kinetics. In this study, alginate-exfoliated WS2 nanosheets were incorporated into ice-template-polymerized PNiPAM hydrogels with the cellular microstructures which mimic plant cells, yet the prompt steerable actuation of animals. Because of the nanosheet-reinforced pore walls formed in situ in freezing polymerization and reasonable hierarchical water channels, this cellular hybrid hydrogel achieves super deformation speed (on the order of magnitude of 10° s), controllable deformation direction, and high near-infrared light responsiveness, offering an unprecedented platform of artificial muscles for various soft robotics and devices (e.g., rotator, microvalve, aquatic swimmer, and water-lifting filter).","{'model': 'tldr@v2.0.0', 'text': 'This cellular hybrid hydrogel achieves super deformation speed (on the order of magnitude of 10° s), controllable deformation direction, and high near-infrared light responsiveness, offering an unprecedented platform of artificial muscles for various soft robotics and devices.'}",https://doi.org/10.1021/acsami.7b10348.s001
-the effects of implementing a point-of-care electronic template to prompt routine anxiety and depression screening in patients consulting for osteoarthritis (the primary care osteoarthritis trial): a cluster randomised trial in primary care,C. Mallen,"Background This study aimed to evaluate whether prompting general practitioners (GPs) to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis (OA) improves pain outcomes. Methods and findings We conducted a cluster randomised controlled trial involving 45 English general practices. In intervention practices, patients aged ≥45 y consulting with OA received point-of-care anxiety and depression screening by the GP, prompted by an automated electronic template comprising five questions (a two-item Patient Health Questionnaire–2 for depression, a two-item Generalized Anxiety Disorder–2 questionnaire for anxiety, and a question about current pain intensity [0–10 numerical rating scale]). The template signposted GPs to follow National Institute for Health and Care Excellence clinical guidelines for anxiety, depression, and OA and was supported by a brief training package. The template in control practices prompted GPs to ask the pain intensity question only. The primary outcome was patient-reported current pain intensity post-consultation and at 3-, 6-, and 12-mo follow-up. Secondary outcomes included pain-related disability, anxiety, depression, and general health. During the trial period, 7,279 patients aged ≥45 y consulted with a relevant OA-related code, and 4,240 patients were deemed potentially eligible by participating GPs. Templates were completed for 2,042 patients (1,339 [31.6%] in the control arm and 703 [23.1%] in the intervention arm). Of these 2,042 patients, 1,412 returned questionnaires (501 [71.3%] from 20 intervention practices, 911 [68.0%] from 24 control practices). Follow-up rates were similar in both arms, totalling 1,093 (77.4%) at 3 mo, 1,064 (75.4%) at 6 mo, and 1,017 (72.0%) at 12 mo. For the primary endpoint, multilevel modelling yielded significantly higher average pain intensity across follow-up to 12 mo in the intervention group than the control group (adjusted mean difference 0.31; 95% CI 0.04, 0.59). Secondary outcomes were consistent with the primary outcome measure in reflecting better outcomes as a whole for the control group than the intervention group. Anxiety and depression scores did not reduce following the intervention. The main limitations of this study are two potential sources of bias: an imbalance in cluster size (mean practice size 7,397 [intervention] versus 5,850 [control]) and a difference in the proportion of patients for whom the GP deactivated the template (33.6% [intervention] versus 27.8% [control]). Conclusions In this study, we observed no beneficial effect on pain outcomes of prompting GPs to routinely screen for and manage comorbid anxiety and depression in patients presenting with symptoms due to OA, with those in the intervention group reporting statistically significantly higher average pain scores over the four follow-up time points than those in the control group. Trial registration ISRCTN registry ISRCTN40721988","{'model': 'tldr@v2.0.0', 'text': 'No beneficial effect on pain outcomes was observed of prompting GPs to routinely screen for and manage comorbid anxiety and depression in patients presenting with symptoms due to OA, with those in the intervention group reporting statistically significantly higher average pain scores over the four follow-up time points than those inThe control group.'}",https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.1002273&type=printable
-templating three-dimensional self-assembled structures in bilayer block copolymer films,Tavakkoli K G,"To the Next Level Block copolymers will spontaneously separate into a range of microstructures that depend on the polymer block lengths and chemical compositions, and have been used as a templating material because one can selectively etch or functionalize one of the blocks. However, creating a template that is more than one layer thick is challenging. Tavakkoli K. G. et al. (p. 1294) used an array of posts to provide independent and simultaneous control of the morphology and orientation of two layers of block copolymers and were able to create local variations in the curvature and spacing of the domains. An array of posts guides the bilayer assembly of block copolymers with independent control of morphology and orientation. The registration and alignment of a monolayer of microdomains in a self-assembled block copolymer thin film can be controlled by chemical or physical templating methods. Although planar patterns are useful for nanoscale device fabrication, three-dimensional multilevel structures are required for some applications. We found that a bilayer film of a cylindrical-morphology block copolymer, templated by an array of posts functionalized with a brush attractive to the majority block, can form a rich variety of three-dimensional structures consisting of cylinder arrays with controllable angles, bends, and junctions whose geometry is controlled by the template periodicity and arrangement. This technique allows control of microdomain patterns and the ability to route and connect microdomains in specific directions.","{'model': 'tldr@v2.0.0', 'text': 'A bilayer film of a cylindrical-morphology block copolymer, templated by an array of posts functionalized with a brush attractive to the majority block, can form a rich variety of three-dimensional structures consisting of cylinder arrays with controllable angles, bends, and junctions whose geometry is controlled by the template periodicity and arrangement.'}",https://dspace.mit.edu/bitstream/1721.1/86080/1/Berggren_Templating%20three.pdf
-surface-templated hydrogel patterns prompt matrix-dependent migration of breast cancer cells towards chemokine-secreting cells.,Taisuke Kojima,,"{'model': 'tldr@v2.0.0', 'text': 'A novel technique for fabricating spatially defined cell-laden collagen hydrogels using patterned, non-adhesive polyacrylamide-coated polydimethylsiloxane (PDMS) surfaces as a template revealed matrix-dependent and chemokine isoform-dependent migration of cancer cells.'}",https://europepmc.org/articles/pmc4293228?pdf=render
-prompt: a template system for second language reading comprehension,Donna Mydlarski,The following article briefly summarizes some different eras of CALL and suggests that the present era could be of immense help and usefulness to language teachers because of the multiplicity of template (miniauthoring)systems which are becoming available. This article also describes PROMPT which is a template for two types of reading comprehension exercises developed at the universities of Calgary and Guelph in Canada.,"{'model': 'tldr@v2.0.0', 'text': 'The following article briefly summarizes some different eras of CALL and suggests that the present era could be of immense help and usefulness to language teachers because of the multiplicity of template (miniauthoring) systems which are becoming available.'}",
-ground-truth labels matter: a deeper look into input-label demonstrations,Junyeob Kim,"Despite recent explosion of interests in in-context learning, the underlying mechanism and the precise impact of the quality of demonstrations remain elusive.Intuitively, ground-truth labels should have as much impact in in-context learning (ICL) as supervised learning, but recent work reported that the input-label correspondence is significantly less important than previously thought.Intrigued by this counter-intuitive observation, we re-examine the importance of ground-truth labels in in-context learning.With the introduction of two novel metrics, namely Label-Correctness Sensitivity and Ground-truth Label Effect Ratio (GLER), we were able to conduct quantifiable analysis on the impact of ground-truth label demonstrations.Through extensive analyses, we find that the correct input-label mappings can have varying impacts on the downstream in-context learning performances, depending on the experimental configuration.Through additional studies, we identify key components, such as the verbosity of prompt templates and the language model size, as the controlling factor to achieve more noise-resilient ICL.","{'model': 'tldr@v2.0.0', 'text': 'Through extensive analyses, it is found that the correct input-label mappings can have varying impacts on the downstream in-context learning performances, depending on the experimental configuration.'}",http://arxiv.org/pdf/2205.12685
-"method, device, client terminal and system for realizing commodity state change prompt",曹晴,"The invention provides a method, a device, a client terminal and a system for realizing a commodity state change prompt. The method comprises the following steps: a template used for collecting a commodity on a website is installed on the client terminal; when a user visits a commodity webpage, information in the commodity webpage is detected according to the template, and a collecting prompt sign is displayed in the commodity webpage when information in the commodity webpage is matched with all information in the template; when a triggering signal of the collecting prompt sign is received, all corresponding information of the commodity is extracted from the commodity webpage according to the template; and when price information in all corresponding information of the commodity is changed through detection, a commodity state change prompt message is sent to the user. Price information of each collected commodity is monitored, and more flexible and intelligent shopping information is provided for the user so that a shopping selection with higher cost performance is conveniently made by the user.","{'model': 'tldr@v2.0.0', 'text': 'Price information of each collected commodity is monitored, and more flexible and intelligent shopping information is provided for the user so that a shopping selection with higher cost performance is conveniently made by the user.'}",
-not all languages are created equal in llms: improving multilingual capability by cross-lingual-thought prompting,Haoyang Huang,"Large language models (LLMs) demonstrate impressive multilingual capability, but their performance varies substantially across different languages. In this work, we introduce a simple yet effective method, called cross-lingual-thought prompting (XLT), to systematically improve the multilingual capability of LLMs. Specifically, XLT is a generic template prompt that stimulates cross-lingual and logical reasoning skills to enhance task performance across languages. We conduct comprehensive evaluations on 7 typical benchmarks related to reasoning, understanding, and generation tasks, covering both high-resource and low-resource languages. Experimental results show that XLT not only remarkably enhances the performance of various multilingual tasks but also significantly reduces the gap between the average performance and the best performance of each task in different languages. Notably, XLT brings over 10 points of average improvement in arithmetic reasoning and open-domain question-answering tasks.","{'model': 'tldr@v2.0.0', 'text': 'Experimental results show that XLT not only remarkably enhances the performance of various multilingual tasks but also significantly reduces the gap between the average performance and the best performance of each task in different languages.'}",http://arxiv.org/pdf/2305.07004
-unihd at tsar-2022 shared task: is compute all we need for lexical simplification?,Dennis Aumiller,"Previous state-of-the-art models for lexical simplification consist of complex pipelines with several components, each of which requires deep technical knowledge and fine-tuned interaction to achieve its full potential. As an alternative, we describe a frustratingly simple pipeline based on prompted GPT-3 responses, beating competing approaches by a wide margin in settings with few training instances. Our best-performing submission to the English language track of the TSAR-2022 shared task consists of an “ensemble” of six different prompt templates with varying context levels. As a late-breaking result, we further detail a language transfer technique that allows simplification in languages other than English. Applied to the Spanish and Portuguese subset, we achieve state-of-the-art results with only minor modification to the original prompts. Aside from detailing the implementation and setup, we spend the remainder of this work discussing the particularities of prompting and implications for future work. Code for the experiments is available online at https://github.com/dennlinger/TSAR-2022-Shared-Task.","{'model': 'tldr@v2.0.0', 'text': 'A language transfer technique that allows simplification in languages other than English and a frustratingly simple pipeline based on prompted GPT-3 responses is described, beating competing approaches by a wide margin in settings with few training instances.'}",http://arxiv.org/pdf/2301.01764
-the utility of chatgpt for cancer treatment information,Shan Chen,"The use of large language models (LLMs) such as ChatGPT for medical question-answering is becoming increasingly popular. However, there are concerns that these models may generate and amplify medical misinformation. Because cancer patients frequently seek to educate themselves through online resources, some individuals will likely use ChatGPT to obtain cancer treatment information. This study evaluated the performance and robustness of ChatGPT in providing breast, prostate, and lung cancer treatment recommendations that align with National Comprehensive Cancer Network (NCCN) guidelines. Four prompt templates were created to explore how differences in how the query is posed impacts response. ChatGPT output was scored by 3 oncologists and a 4th oncologist adjudicated in cases of disagreement. ChatGPT provided at least one NCCN-concordant recommendation for 102/104 (98%) prompts. However, 35/102 (34.3%) of these also included a recommendation that was at least partially non-concordant with NCCN guidelines. Responses varied based on prompt type. In conclusion, ChatGPT did not perform well at reliably and robustly providing cancer treatment recommendations. Patients and clinicians should be aware of the limitations of ChatGPT and similar technologies for self-education.","{'model': 'tldr@v2.0.0', 'text': 'Patients and clinicians should be aware of the limitations of ChatGPT and similar technologies for self-education because cancer patients frequently seek to educate themselves through online resources and some individuals will likely useChatGPT to obtain cancer treatment information.'}",https://www.medrxiv.org/content/medrxiv/early/2023/03/23/2023.03.16.23287316.full.pdf
-avatarcraft: transforming text into neural human avatars with parameterized shape and pose control,Ruixia Jiang,"Neural implicit fields are powerful for representing 3D scenes and generating high-quality novel views, but it remains challenging to use such implicit representations for creating a 3D human avatar with a specific identity and artistic style that can be easily animated. Our proposed method, AvatarCraft, addresses this challenge by using diffusion models to guide the learning of geometry and texture for a neural avatar based on a single text prompt. We carefully design the optimization framework of neural implicit fields, including a coarse-to-fine multi-bounding box training strategy, shape regularization, and diffusion-based constraints, to produce high-quality geometry and texture. Additionally, we make the human avatar animatable by deforming the neural implicit field with an explicit warping field that maps the target human mesh to a template human mesh, both represented using parametric human models. This simplifies animation and reshaping of the generated avatar by controlling pose and shape parameters. Extensive experiments on various text descriptions show that AvatarCraft is effective and robust in creating human avatars and rendering novel views, poses, and shapes. Our project page is: https://avatar-craft.github.io/.","{'model': 'tldr@v2.0.0', 'text': 'The proposed method, AvatarCraft, addresses the challenge of creating a 3D human avatar with a specific identity and artistic style that can be easily animated by using diffusion models to guide the learning of geometry and texture for a neural avatar based on a single text prompt.'}",https://arxiv.org/pdf/2303.17606
-trustgpt: a benchmark for trustworthy and responsible large language models,Yue Huang,"Large Language Models (LLMs) such as ChatGPT, have gained significant attention due to their impressive natural language processing capabilities. It is crucial to prioritize human-centered principles when utilizing these models. Safeguarding the ethical and moral compliance of LLMs is of utmost importance. However, individual ethical issues have not been well studied on the latest LLMs. Therefore, this study aims to address these gaps by introducing a new benchmark -- TrustGPT. TrustGPT provides a comprehensive evaluation of LLMs in three crucial areas: toxicity, bias, and value-alignment. Initially, TrustGPT examines toxicity in language models by employing toxic prompt templates derived from social norms. It then quantifies the extent of bias in models by measuring quantifiable toxicity values across different groups. Lastly, TrustGPT assesses the value of conversation generation models from both active value-alignment and passive value-alignment tasks. Through the implementation of TrustGPT, this research aims to enhance our understanding of the performance of conversation generation models and promote the development of language models that are more ethical and socially responsible.","{'model': 'tldr@v2.0.0', 'text': 'This study aims to enhance the understanding of the performance of conversation generation models and promote the development of language models that are more ethical and socially responsible by introducing a new benchmark -- TrustGPT.'}",http://arxiv.org/pdf/2306.11507
-using natural sentence prompts for understanding biases in language models,Sarah Alnegheimish,"Evaluation of biases in language models is often limited to synthetically generated datasets. This dependence traces back to the need of prompt-style dataset to trigger specific behaviors of language models. In this paper, we address this gap by creating a prompt dataset with respect to occupations collected from real-world natural sentences present in Wikipedia.We aim to understand the differences between using template-based prompts and natural sentence prompts when studying gender-occupation biases in language models. We find bias evaluations are very sensitiveto the design choices of template prompts, and we propose using natural sentence prompts as a way of more systematically using real-world sentences to move away from design decisions that may bias the results.","{'model': 'tldr@v2.0.0', 'text': 'A prompt dataset with respect to occupations collected from real-world natural sentences present in Wikipedia is created to understand the differences between using template-based prompts and natural sentence prompts when studying gender-occupation biases in language models.'}",https://arxiv.org/pdf/2205.06303
-an innovative dual recognition aptasensor for specific detection of staphylococcus aureus based on au/fe3o4 binary hybrid,Mohamed M. El-wekil,,,https://www.nature.com/articles/s41598-022-15637-1.pdf
-domain knowledge matters: improving prompts with fix templates for repairing python type errors,Yun Peng,"Although the dynamic type system of Python facilitates the developers in writing Python programs, it also brings type errors at run-time. There exist rule-based approaches for automatically repairing Python type errors. The approaches can generate accurate patches but they require domain experts to design patch synthesis rules and suffer from low template coverage of real-world type errors. Learning-based approaches alleviate the manual efforts in designing patch synthesis rules. Among the learning-based approaches, the prompt-based approach which leverages the knowledge base of code pre-trained models via pre-defined prompts, obtains state-of-the-art performance in general program repair tasks. However, such prompts are manually defined and do not involve any specific clues for repairing Python type errors, resulting in limited effectiveness. How to automatically improve prompts with the domain knowledge for type error repair is challenging yet under-explored. In this paper, we present TypeFix, a novel prompt-based approach with fix templates incorporated for repairing Python type errors. TypeFix first mines generalized fix templates via a novel hierarchical clustering algorithm. The identified fix templates indicate the common edit patterns and contexts of existing type error fixes. TypeFix then generates code prompts for code pre-trained models by employing the generalized fix templates as domain knowledge, in which the masks are adaptively located for each type error instead of being pre-determined. Experiments on two benchmarks, including BugsInPy and TypeBugs, show that TypeFix successfully repairs 26 and 55 type errors, outperforming the best baseline approach by 9 and 14, respectively. Besides, the proposed fix template mining approach can cover 75% of developers' patches in both benchmarks, increasing the best rule-based approach PyTER by more than 30%.","{'model': 'tldr@v2.0.0', 'text': ""TypeFix is presented, a novel prompt-based approach with fix templates incorporated for repairing Python type errors that can cover 75% of developers' patches in both benchmarks, and increases the best rule- based approach PyTER by more than 30%.""}",http://arxiv.org/pdf/2306.01394
-citeprompt: using prompts to identify citation intent in scientific papers,Avishek Lahiri,"Citations in scientific papers not only help us trace the intellectual lineage but also are a useful indicator of the scientific significance of the work. Citation intents prove beneficial as they specify the role of the citation in a given context. We present a tool Citeprompt which uses the hitherto unexplored approach of prompt learning for citation intent classification. We argue that with the proper choice of the pretrained language model, the prompt template, and the prompt verbalizer, we can not only get results that are better than or comparable to those obtained with the state-of-the-art methods but also do it with much less exterior information about the scientific document. We report state-of-the-art results on the ACL-ARC dataset, and also show significant improvement on the SciCite dataset over all baseline models except one. As suitably large labelled datasets for citation intent classification can be quite hard to find, in a first, we propose the conversion of this task to the few-shot and zero-shot settings. For the ACL-ARC dataset, we report a 53.86% F1 score for the zero-shot setting, which improves to 63.61% and 66.99% for the 5-shot and 10-shot settings respectively.","{'model': 'tldr@v2.0.0', 'text': 'A tool Citeprompt is presented which uses the hitherto unexplored approach of prompt learning for citation intent classification and argues that with the proper choice of the pretrained language model, the prompt template, and the prompt verbalizer, one can get results that are better than or comparable to those obtained with the state-of-the-art methods.'}",https://arxiv.org/pdf/2304.12730
-extracting structured seed-mediated gold nanorod growth procedures from literature with gpt-3,Nicholas Walker,"Although gold nanorods have been the subject of much research, the pathways for controlling their shape and thereby their optical properties remain largely heuristically understood. Although it is apparent that the simultaneous presence of and interaction between various reagents during synthesis control these properties, computational and experimental approaches for exploring the synthesis space can be either intractable or too time-consuming in practice. This motivates an alternative approach leveraging the wealth of synthesis information already embedded in the body of scientific literature by developing tools to extract relevant structured data in an automated, high-throughput manner. To that end, we present an approach using the powerful GPT-3 language model to extract structured multi-step seed-mediated growth procedures and outcomes for gold nanorods from unstructured scientific text. GPT-3 prompt completions are fine-tuned to predict synthesis templates in the form of JSON documents from unstructured text input with an overall accuracy of $86\%$. The performance is notable, considering the model is performing simultaneous entity recognition and relation extraction. We present a dataset of 11,644 entities extracted from 1,137 papers, resulting in 268 papers with at least one complete seed-mediated gold nanorod growth procedure and outcome for a total of 332 complete procedures.","{'model': 'tldr@v2.0.0', 'text': 'An approach using the powerful GPT-3 language model to extract structured multi-step seed-mediated growth procedures and outcomes for gold nanorods from unstructured scientific text by developing tools to extract relevant structured data in an automated, high-throughput manner.'}",http://arxiv.org/pdf/2304.13846
-a dataset for cross-domain reasoning via template filling,,"While several benchmarks exist for reasoning 001 tasks, reasoning across domains is an under- 002 explored area in NLP. Towards this, we present 003 a dataset and a prompt-template-filling ap- 004 proach to enable sequence to sequence mod- 005 els to perform cross-domain reasoning. We 006 also present a case-study with commonsense 007 and health and well-being domains, where 008 we study how prompt-template-filling en- 009 ables pretrained sequence to sequence models 010 across domains. Our experiments across sev- 011 eral pretrained encoder-decoder models show 012 that cross-domain reasoning is challenging for 013 current models. We also show an in-depth er- 014 ror analysis and avenues for future research for 015 reasoning across domains 1 . 016","{'model': 'tldr@v2.0.0', 'text': 'A dataset and a prompt-template-filling approach to enable sequence to sequence mod-5 models to perform cross-domain reasoning and an in-depth analysis of avenues for future research for reasoning across domains.'}",
-prompting for automatic log template extraction,Junjie Xu,"Log parsing, the initial and vital stage in automated log analysis, involves extracting log templates from semi-structured logs to generate structured logs. Nonetheless, current log parsers are limited in effectiveness due to two primary reasons. Firstly, traditional data-driven log parsers heavily rely on heuristics or manually crafted features provided by domain experts, which may not consistently yield optimal performance when applied to diverse log systems. Secondly, existing deep learning-based log parsers necessitate model tuning, which is typically confined to training samples and leads to suboptimal performance across the entire log source. To overcome these limitations, we propose a precise log parsing framework named LogDiv, which leverages the in-context inference capability of large language models. Specifically, LogDiv extracts the hidden semantics from multiple log examples through prompt demonstrations. Without the need for model tuning, LogDiv can directly generate a log template for the target log message by leveraging the semantics provided in the prompt context. Additionally, we introduce a simple yet effective prompt format for extracting the output and enhancing the quality of the generated log templates. To validate the performance of LogDiv, we conducted experiments using 16 widely-used public datasets. The results show that LogDiv achieves state-of-the-art performance with an average parsing accuracy of 97.7%, precision template accuracy of 88.1%, and recall template accuracy of 90.8%.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a precise log parsing framework named LogDiv, which leverages the in-context inference capability of large language models and achieves state-of-the-art performance.'}",https://arxiv.org/pdf/2307.09950
-dspy: compiling declarative language model calls into self-improving pipelines,O. Khattab,"The ML community is rapidly exploring techniques for prompting language models (LMs) and for stacking them into pipelines that solve complex tasks. Unfortunately, existing LM pipelines are typically implemented using hard-coded""prompt templates"", i.e. lengthy strings discovered via trial and error. Toward a more systematic approach for developing and optimizing LM pipelines, we introduce DSPy, a programming model that abstracts LM pipelines as text transformation graphs, i.e. imperative computational graphs where LMs are invoked through declarative modules. DSPy modules are parameterized, meaning they can learn (by creating and collecting demonstrations) how to apply compositions of prompting, finetuning, augmentation, and reasoning techniques. We design a compiler that will optimize any DSPy pipeline to maximize a given metric. We conduct two case studies, showing that succinct DSPy programs can express and optimize sophisticated LM pipelines that reason about math word problems, tackle multi-hop retrieval, answer complex questions, and control agent loops. Within minutes of compiling, a few lines of DSPy allow GPT-3.5 and llama2-13b-chat to self-bootstrap pipelines that outperform standard few-shot prompting (generally by over 25% and 65%, respectively) and pipelines with expert-created demonstrations (by up to 5-46% and 16-40%, respectively). On top of that, DSPy programs compiled to open and relatively small LMs like 770M-parameter T5 and llama2-13b-chat are competitive with approaches that rely on expert-written prompt chains for proprietary GPT-3.5. DSPy is available at https://github.com/stanfordnlp/dspy","{'model': 'tldr@v2.0.0', 'text': 'DSPy is introduced, a programming model that abstracts LM pipelines as text transformation graphs, i.e. imperative computational graphs where LMs are invoked through declarative modules, and a compiler is designed that will optimize any DSPy pipeline to maximize a given metric.'}",https://arxiv.org/pdf/2310.03714
-role knowledge prompting for document-level event argument extraction,Ruijuan Hu,"Document-level event argument extraction (DEAE) aims to identify the arguments corresponding to the roles of a given event type in a document. However, arguments scattering and arguments and roles overlapping make DEAE face great challenges. In this paper, we propose a novel DEAE model called Role Knowledge Prompting for Document-Level Event Argument Extraction (RKDE), which enhances the interaction between templates and roles through a role knowledge guidance mechanism to precisely prompt pretrained language models (PLMs) for argument extraction. Specifically, it not only facilitates PLMs to understand deep semantics but also generates all the arguments simultaneously. The experimental results show that our model achieved decent performance on two public DEAE datasets, with 3.2% and 1.4% F1 improvement on Arg-C, and to some extent, it addressed the overlapping arguments and roles.","{'model': 'tldr@v2.0.0', 'text': 'A novel DEAE model called Role Knowledge Prompting for Document-Level Event Argument Extraction (RKDE) is proposed, which enhances the interaction between templates and roles through a role knowledge guidance mechanism to precisely prompt pretrained language models (PLMs) for argument extraction.'}",https://www.mdpi.com/2076-3417/13/5/3041/pdf?version=1677492694
-cocomo: computational consciousness modeling for generative and ethical ai,Edward Y. Chang,"The CoCoMo model proposes a computational solution to the challenge of incorporating ethical and emotional intelligence considerations into AI systems, with the aim of creating AI agents that combine knowledge with compassion. To achieve this goal, CoCoMo prioritizes fairness, beneficence, non-maleficence, empathy, adaptability, transparency, and critical and exploratory thinking abilities. The model employs consciousness modeling, reinforcement learning, and prompt template formulation to support these desired traits. By incorporating ethical and emotional intelligence considerations, a generative AI model can potentially lead to improved fairness, reduced toxicity, and increased reliability.",,http://arxiv.org/pdf/2304.02438
-cot-bert: enhancing unsupervised sentence representation through chain-of-thought,Bowen Zhang,"Unsupervised sentence representation learning aims to transform input sentences into fixed-length vectors enriched with intricate semantic information while obviating the reliance on labeled data. Recent progress within this field, propelled by contrastive learning and prompt engineering, has significantly bridged the gap between unsupervised and supervised strategies. Nonetheless, the potential utilization of Chain-of-Thought, remains largely untapped within this trajectory. To unlock latent capabilities within pre-trained models, such as BERT, we propose a two-stage approach for sentence representation: comprehension and summarization. Subsequently, the output of the latter phase is harnessed as the vectorized representation of the input sentence. For further performance enhancement, we meticulously refine both the contrastive learning loss function and the template denoising technique for prompt engineering. Rigorous experimentation substantiates our method, CoT-BERT, transcending a suite of robust baselines without necessitating other text representation models or external databases.","{'model': 'tldr@v2.0.0', 'text': 'Rigorous experimentation substantiates the method, CoT-BERT, transcending a suite of robust baselines without necessitating other text representation models or external databases, and meticulously refine both the contrastive learning loss function and the template denoising technique for prompt engineering.'}",https://arxiv.org/pdf/2309.11143
-advanced prompting as a catalyst: empowering large language models in the management of gastrointestinal cancers,J. Yuan,"Large Language Models' (LLMs) performance in healthcare can be significantly impacted by prompt engineering. However, the area of study remains relatively uncharted in gastrointestinal oncology until now. Our research delves into this unexplored territory, investigating the efficacy of varied prompting strategies, including simple prompts, templated prompts, in-context learning (ICL), and multi-round iterative questioning, for optimizing the performance of LLMs within a medical setting. We develop a comprehensive evaluation system to assess the performance of LLMs across multiple dimensions. This robust evaluation system ensures a thorough assessment of the LLMs' capabilities in the field of medicine. Our findings suggest a positive relationship between the comprehensiveness of the prompts and the LLMs' performance. Notably, the multi-round strategy, which is characterized by iterative question-and-answer rounds, consistently yields the best results. ICL, a strategy that capitalizes on interrelated contextual learning, also displays significant promise, surpassing the outcomes achieved with simpler prompts. The research underscores the potential of advanced prompt engineering and iterative learning approaches for boosting the applicability of LLMs in healthcare. We recommend that additional research be conducted to refine these strategies and investigate their potential integration, to truly harness the full potential of LLMs in medical applications. ","{'model': 'tldr@v2.0.0', 'text': 'The research investigates the efficacy of varied prompting strategies, including simple prompts, templated prompts, in-context learning (ICL), and multi-round iterative questioning, for optimizing the performance of LLMs within a medical setting, and develops a comprehensive evaluation system.'}",https://www.the-innovation.org/data/article/export-pdf?id=64db4fd54228a72545780714
-automated cerebral infarct detection on computed tomography images based on deep learning,S. Peng,"The limited accuracy of cerebral infarct detection on CT images caused by the low contrast of CT hinders the desirable application of CT as a first-line diagnostic modality for screening of cerebral infarct. This research was aimed at utilizing convolutional neural network to enhance the accuracy of automated cerebral infarct detection on CT images. The CT images underwent a series of preprocessing steps mainly to enhance the contrast inside the parenchyma, adjust the orientation, spatially normalize the images to the CT template, and create a t-score map for each patient. The input format of the convolutional neural network was the t-score matrix of a 16 × 16-pixel patch. Non-infarcted and infarcted patches were selected from the t-score maps, on which data augmentation was conducted to generate more patches for training and testing the proposed convolutional neural network. The convolutional neural network attained a 93.9% patch-wise detection accuracy in the test set. The proposed method offers prompt and accurate cerebral infarct detection on CT images. It renders a frontline detection modality of ischemic stroke on an emergent or regular basis.","{'model': 'tldr@v2.0.0', 'text': 'The proposed convolutional neural network renders a frontline detection modality of ischemic stroke on an emergent or regular basis and offers prompt and accurate cerebral infarct detection on CT images.'}",https://www.mdpi.com/2227-9059/10/1/122/pdf?version=1641806958
-incorporating instructional prompts into a unified generative framework for joint multiple intent detection and slot filling,Yangjun Wu,"The joint multiple Intent Detection (ID) and Slot Filling (SF) is a significant challenge in spoken language understanding. Because the slots in an utterance may relate to multi-intents, most existing approaches focus on utilizing task-specific components to capture the relations between intents and slots. The customized networks restrict models from modeling commonalities between tasks and generalization for broader applications. To address the above issue, we propose a Unified Generative framework (UGEN) based on a prompt-based paradigm, and formulate the task as a question-answering problem. Specifically, we design 5-type templates as instructional prompts, and each template includes a question that acts as the driver to teach UGEN to grasp the paradigm, options that list the candidate intents or slots to reduce the answer search space, and the context denotes original utterance. Through the instructional prompts, UGEN is guided to understand intents, slots, and their implicit correlations. On two popular multi-intent benchmark datasets, experimental results demonstrate that UGEN achieves new SOTA performances on full-data and surpasses the baselines by a large margin on 5-shot (28.1%) and 10-shot (23%) scenarios, which verify that UGEN is robust and effective.","{'model': 'tldr@v2.0.0', 'text': 'A Unified Generative framework (UGEN) based on a prompt-based paradigm, and formulate the task as a question-answering problem to understand intents, slots, and their implicit correlations is proposed.'}",
-a practical three-phase approach to fully automated programming using system decomposition and coding copilots,Haoli Bai,"Very large-scale (VLS) deep learning models are capable of generating meaningful code snippets, yet the performance drops dramatically when the coding task becomes more complex. Although fully neural approaches have been proposed to solve this problem, the value of the application is still limited. In our work, we propose a neuro-symbolic approach that integrates the symbolic natures of programming and the existing neural language models. We divide a programming task into three phases: forming a hierarchical task composed of functions, completing each function, and fulfilling the corner cases. Because each phase can be completed by language models, the coding process can be fully automated. Our contribution is three-fold. Firstly, we show that with little help from humans, VLS language models are capable of completing non-trivial programming tasks. Secondly, we provide a number of empirical insights to create prompt templates that help the language models generate better code. Thirdly, compared to the existing approaches, our work provides a much more practical approach for programmers and researchers to follow. The generated programming project using our fully automated programming approach and part of the ablation study code are available at https://github.com/BiEchi/FAP.","{'model': 'tldr@v2.0.0', 'text': 'This work proposes a neuro-symbolic approach that integrates the symbolic natures of programming and the existing neural language models to show that with little help from humans, VLS language models are capable of completing non-trivial programming tasks.'}",
-towards robust nlg bias evaluation with syntactically-diverse prompts,Arshiya Aggarwal,"We present a robust methodology for evaluating biases in natural language generation(NLG) systems. Previous works use fixed hand-crafted prefix templates with mentions of various demographic groups to prompt models to generate continuations for bias analysis. These fixed prefix templates could themselves be specific in terms of styles or linguistic structures, which may lead to unreliable fairness conclusions that are not representative of the general trends from tone varying prompts. To study this problem, we paraphrase the prompts with different syntactic structures and use these to evaluate demographic bias in NLG systems. Our results suggest similar overall bias trends but some syntactic structures lead to contradictory conclusions compared to past works. We show that our methodology is more robust and that some syntactic structures prompt more toxic content while others could prompt less biased generation. This suggests the importance of not relying on a fixed syntactic structure and using tone-invariant prompts. Introducing syntactically-diverse prompts can achieve more robust NLG (bias) evaluation.",,https://arxiv.org/pdf/2212.01700
-ku x upstage’s submission for the wmt22 quality estimation: critical error detection shared task,Sugyeong Eo,"This paper presents KU X Upstage’s submission to the quality estimation (QE): critical error detection (CED) shared task in WMT22. We leverage the XLM-RoBERTa large model without utilizing any additional parallel data. To the best of our knowledge, we apply prompt-based fine-tuning to the QE task for the first time. To maximize the model’s language understanding capability, we reformulate the CED task to be similar to the masked language model objective, which is a pre-training strategy of the language model. We design intuitive templates and label words, and include auxiliary descriptions such as demonstration or Google Translate results in the input sequence. We further improve the performance through the template ensemble, and as a result of the shared task, our approach achieve the best performance for both English-German and Portuguese-English language pairs in an unconstrained setting.","{'model': 'tldr@v2.0.0', 'text': 'KU X Upstage’s submission to the quality estimation (QE): critical error detection (CED) shared task in WMT22 is presented, and prompt-based fine-tuning is applied to the QE task for the first time.'}",
-desed: dialogue-based explanation for sentence-level event detection,Yinyi Wei,"Many recent sentence-level event detection efforts focus on enriching sentence semantics, e.g., via multi-task or prompt-based learning. Despite the promising performance, these methods commonly depend on label-extensive manual annotations or require domain expertise to design sophisticated templates and rules. This paper proposes a new paradigm, named dialogue-based explanation, to enhance sentence semantics for event detection. By saying dialogue-based explanation of an event, we mean explaining it through a consistent information-intensive dialogue, with the original event description as the start utterance. We propose three simple dialogue generation methods, whose outputs are then fed into a hybrid attention mechanism to characterize the complementary event semantics. Extensive experimental results on two event detection datasets verify the effectiveness of our method and suggest promising research opportunities in the dialogue-based explanation paradigm.","{'model': 'tldr@v2.0.0', 'text': 'This paper proposes a new paradigm, named dialogue-based explanation, to enhance sentence semantics for event detection, by explaining it through a consistent information-intensive dialogue, with the original event description as the start utterance.'}",
-vision encoders in visual question answering,Ryan R. Anderson,"Most existing methods that apply pretrained Visual Language Models (VLMs) to vision and language tasks do not sufficiently explore the effect of the format of their inputs on downstream performance. We show that utilising appropriate prompt formatting is a simple yet effective approach to improving the few-shot performance of VLMs that use relatively small language models on the Visual Question Answering (VQA) task. We format the inputs used to prompt a VLM using a modified text-only template from a closed-book question answering task that the language-model component of the VLM was pretrained on. By doing this, we explicitly align the VQA task with a task that this language model has already seen, enabling the VLM to leverage the similarities between the tasks, such as the answer-length distribution, when generating answers to the visual questions. In order to test our claims, we implement a simple architecture based on Frozen (Tsimpoukelli et al., 2021) and ClipCap (Mokady et al., 2021), whereby, through image captioning, the VLM learns to integrate powerful pretrained vision-only and language-only models via a relatively simple learnt mapping network. Furthermore, we contextualise our approach relative to existing work by presenting a unified view of VLMs. Our results show that explicit alignment enables our VLMs to achieve a significantly higher zero-shot (34.49% vs 20.89%) and best overall (40.39% vs 30.83%) VQA score on the VQA2.0 dataset (Goyal et al., 2017) than when the prompt template from Frozen (Tsimpoukelli et al., 2021) and Flamingo (Alayrac et al., 2022) is used. Furthermore, our zero-shot and best overall performance is better than Frozen’s (34.49% vs 29.5% and 40.39% vs 38.2%, respectively) despite Frozen using a language model with more than double the number of parameters. Our code is available here.","{'model': 'tldr@v2.0.0', 'text': 'This work formats the inputs used to prompt a VLM using a modified text-only template from a closed-book question answering task that the language-model component of the VLM was pretrained on, enabling the V LM to leverage the similarities between the tasks, such as the answer-length distribution, when generating answers to the visual questions.'}",
-keyword-optimized template insertion for clinical information extraction via prompt-based learning,Eugenia Alleva,"Clinical note classification is a common clinical NLP task. However, annotated data-sets are scarse. Prompt-based learning has recently emerged as an effective method to adapt pre-trained models for text classification using only few training examples. A critical component of prompt design is the definition of the template (i.e. prompt text). The effect of template position, however, has been insufficiently investigated. This seems particularly important in the clinical setting, where task-relevant information is usually sparse in clinical notes. In this study we develop a keyword-optimized template insertion method (KOTI) and show how optimizing position can improve performance on several clinical tasks in a zero-shot and few-shot training setting.","{'model': 'tldr@v2.0.0', 'text': 'This study develops a keyword-optimized template insertion method (KOTI) and shows how optimizing position can improve performance on several clinical tasks in a zero-shot and few-shot training setting.'}",
-kul@smm4h’22: template augmented adaptive pre-training for tweet classification,Sumam Francis,This paper describes models developed for the Social Media Mining for Health (SMM4H) 2022 shared tasks. Our team participated in the first subtask that classifies tweets with Adverse Drug Effect (ADE) mentions. Our best-performing model comprises of a template augmented task adaptive pre-training and further fine-tuning on target task data. Augmentation with random prompt templates increases the amount of task-specific data to generalize the LM to the target task domain. We explore 2 pre-training strategies: Masked language modeling (MLM) and Simple contrastive pre-training (SimSCE) and the impact of adding template augmentations with these pre-training strategies. Our system achieves an F1 score of 0.433 on the test set without using supplementary resources and medical dictionaries.,"{'model': 'tldr@v2.0.0', 'text': 'This paper describes models developed for the Social Media Mining for Health (SMM4H) 2022 shared tasks and explores 2 pre-training strategies: Masked language modeling (MLM) and Simple contrastive pre- training (SimSCE) and the impact of adding template augmentations with these pre- Training strategies.'}",
-research on chinese short text classification based on prefix-vector attention template and probabilistic answer set,Baoshan Sun,"When people tend to communicate in the network, a large amount of text data is generated. These text data have become an important way for people to obtain information. Nowadays, text classification through prompt learning has also become the focus of people’s research. However, the effect of prompt learning is not ideal due to the immobilization of the prompt template and the single label words. In response to the above problems, this paper proposes a method of constructing a prefix vector attention template (PAP) and a probabilistic answer set. The prefix space vector is added in front of the input text sentence as a template for prompt learning, and using the Attention mechanism to connect the template with the input sentence to construct the input of the model sentence. For each label word, an answer space set is constructed, and the probability that the text belongs to this category is obtained in the form of the answer set probability. The experimental results show that the PAP + probabilistic answer set algorithm proposed in this paper is significantly better than other benchmark models in the text classification task, and can achieve better results.","{'model': 'tldr@v2.0.0', 'text': 'The experimental results show that the PAP + probabilistic answer set algorithm proposed in this paper is significantly better than other benchmark models in the text classification task, and can achieve better results.'}",
-stt: soft template tuning for few-shot learning,Christopher Potts,"With the rapid expansion of large pre-trained 001 language models, fine-tuning all the model pa002 rameters for downstream tasks is becoming 003 computationally prohibitive. The recently de004 veloped prompt-based methods freeze the en005 tire model parameters and only update the so006 called prompt parameters appended to the in007 puts, significantly reducing the burden of fully 008 fine-tuning. However, standard prompt-based 009 methods mainly consider the case where suf010 ficient data of downstream tasks are available. 011 It is still unclear whether the advantage can be 012 transferred to the few-shot regime, where only 013 limited data are available for each downstream 014 task. Our empirical studies suggest there is 015 still a gap between prompt tuning and fully 016 fine-tuning for few-shot learning. We propose 017 a new prompt-tuning framework, called Soft 018 Template Tuning (STT), to bridge the gap. STT 019 combines manual prompts and auto-prompts, 020 and treats downstream classification tasks as a 021 masked language modeling task. STT can close 022 the gap between fine-tuning and prompt-based 023 methods without introducing additional param024 eters. Importantly, it can even outperform 025 the timeand resource-consuming fine-tuning 026 method on sentiment classification tasks. 027","{'model': 'tldr@v2.0.0', 'text': 'A new prompt-tuning framework, called Soft 018 Template Tuning (STT), is proposed, which combines manual prompts and auto-prompts and treats downstream classification tasks as a masked language modeling task and can even outperform the time and resource-consuming fine-tuned method on sentiment classification tasks.'}",
-cross-domain reasoning via template filling,Dheeraj Rajagopal,"In this paper, we explore the ability of sequence to sequence models to perform cross-domain reasoning. Towards this, we present a prompt-template-filling approach to enable sequence to sequence models to perform cross-domain reasoning. We also present a case-study with commonsense and health and well-being domains, where we study how prompt-template-filling enables pretrained sequence to sequence models across domains. Our experiments across several pretrained encoder-decoder models show that cross-domain reasoning is challenging for current models. We also show an in-depth error analysis and avenues for future research for reasoning across domains 1 .","{'model': 'tldr@v2.0.0', 'text': 'This paper presents a prompt-template-filling approach to enable sequence to sequence models to perform cross-domain reasoning, and presents a case-study with commonsense and health and well-being domains, where it is studied how prompt- templates enables pretrained sequence to sequences models across domains.'}",
-unveiling one-pot template-free fabrication of exquisite multidimensional ptni multicubes nanoarchitectonics for the efficient electrochemical oxidation of ethanol and methanol with a great tolerance for co.,Fengxia Wu,"Multidimensional bi-metallic Pt-based nanoarchitectonics are highly promising in electrochemical energy conversion technologies because of their fancy structural merits and accessible active sites; however, hitherto their precise template-free fabrication remains a grand challenge. Herein, we report a facile template-free solvothermal approach for the one-pot rational design of co-centric PtNi multicubes nanoarchitectonics via adjusting the oleylamine/oleic acid ratio with curcumin. The obtained multidimensional PtNi multicubes comprise multiple small interlace-stacked nanocubes subunits assembled in spatially porous branched nanoarchitectonics and bounded by high-index facets. The synthetic mechanism is based on the spontaneous isolation among prompt nucleation and oriented attachment epitaxial growth. These inimitable architectural and compositional merits of PtNi multicubes endowed the ethanol oxidation mass and specific activity by 5.6 and 9.03 times than commercial Pt/C catalyst, respectively, alongside enhancement the methanol oxidation mass activity by 2.3 times. Also, PtNi multicubes showed superior durability and a higher tolerance for CO poisoning than Pt/C catalyst. This work may pave the way for tailored synthesis of Pt-based nanoarchitectonics for myriad catalytic reactions.","{'model': 'tldr@v2.0.0', 'text': 'A facile template-free solvothermal approach for the one-pot rational design of co-centric PtNi multicubes nanoarchitectonics via adjusting the oleylamine/oleic acid ratio with curcumin is reported, paving the way for tailored synthesis of Pt-based nanoarch architectonics for myriad catalytic reactions.'}",
-daprompt: deterministic assumption prompt learning for event causality identification,Wei Xiang,"Event Causality Identification (ECI) aims at determining whether there is a causal relation between two event mentions. Conventional prompt learning designs a prompt template to first predict an answer word and then maps it to the final decision. Unlike conventional prompts, we argue that predicting an answer word may not be a necessary prerequisite for the ECI task. Instead, we can first make a deterministic assumption on the existence of causal relation between two events and then evaluate its rationality to either accept or reject the assumption. The design motivation is to try the most utilization of the encyclopedia-like knowledge embedded in a pre-trained language model. In light of such considerations, we propose a deterministic assumption prompt learning model, called DAPrompt, for the ECI task. In particular, we design a simple deterministic assumption template concatenating with the input event pair, which includes two masks as predicted events' tokens. We use the probabilities of predicted events to evaluate the assumption rationality for the final event causality decision. Experiments on the EventStoryLine corpus and Causal-TimeBank corpus validate our design objective in terms of significant performance improvements over the state-of-the-art algorithms.","{'model': 'tldr@v2.0.0', 'text': ""A deterministic assumption prompt learning model, called DAPrompt, is proposed, which includes two masks as predicted events' tokens and uses the probabilities of predicted events to evaluate the assumption rationality for the final event causality decision.""}",https://arxiv.org/pdf/2307.09813
-working example of template prompt payments by local authorities,Em Reeves,Number Value (€) Percentage (%) of total invoices paid Percentage (%) of the total amount paid Total invoices paid in Quarter,,
-the effects of implementing a point-of-care template to prompt routine and screening in consulting for (the,,Background This study aimed to evaluate whether prompting general practitioners (GPs) to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis (OA) improves pain outcomes.,"{'model': 'tldr@v2.0.0', 'text': 'This study aimed to evaluate whether prompting general practitioners to routinely assess and manage anxiety and depression in patients consulting with osteoarthritis improves pain outcomes.'}",
-coarse-grained diffraction template matching model to retrieve multiconformational models for biomolecule structures from noisy diffraction patterns,Atsushi Tokuhisa,"Biomolecular imaging using X-ray free-electron lasers (XFELs) has been successfully applied to serial femtosecond crystallography. However, the application of single-particle analysis for structure determination using XFELs with 100 nm or smaller biomolecules has two practical problems: the incomplete diffraction datasets for reconstructing 3D assembled structures and the heterogeneous conformational states of samples. A new diffraction template matching method is thus presented here to retrieve a plausible 3D structural model based on single noisy target diffraction patterns, assuming candidate structures. Two concepts are introduced here: prompt candidate diffraction, generated by enhanced sampled coarse-grain (CG) candidate structures, and efficient molecular orientation searching for matching based on Bayesian optimization. A CG model-based diffraction-matching protocol is proposed that achieves a 100-fold speed increase compared to exhaustive diffraction matching using an all-atom model. The conditions that enable multi-conformational analysis were also investigated by simulated diffraction data for various conformational states of chromatin and ribosomes. The proposed method can enable multi-conformational analysis, with a structural resolution of at least 20 Å for 270-800 Å flexible biomolecules, in experimental single-particle structure analyses that employ XFELs.","{'model': 'tldr@v2.0.0', 'text': 'A new diffraction template matching method is presented here to retrieve a plausible 3D structural model based on single noisy target diffraction patterns, assuming candidate structures, and a CG model-based diffraction-matching protocol is proposed that achieves a 100-fold speed increase compared to exhaustive diffraction matching using an all-atom model.'}",
-unraveling template-free fabrication of carbon nitride nanorods codoped with pt and pd for efficient electrochemical and photoelectrochemical carbon monoxide oxidation at room temperature.,Kamel Eid,"The tailored synthesis of carbon nitrides (CNs) is of particular interest in multidisciplinary catalytic applications. However, their fabrication in the form of one-dimensional (1D) nanorods for electrocatalytic carbon monoxide (CO) oxidation is not hitherto reported. Herein, a facile roadmap is presented for the rational design of Pt- and Pd-codoped CN (PtPd/CNs) nanorods via protonation of melamine in an ethylene glycol solution containing Pt and Pd precursors using NaNO3 and HCl and subsequent annealing. The protonation induces the polymerization of melamine to melon nanosheets that consequently roll up to CN nanorods. This tailored the prompt high mass production of uniform 1D CN nanorods (94 ± 2 nm) with a high surface area (155.2 m2 g-1) and they were atomically codoped with Pt and Pd (1.5 wt%) without a template and/or multiple complicated steps. The electrocatalytic CO oxidation activity of PtPd/CNs is 2.01 and 23.41 times greater than that of the commercial Pt/C catalyst and metal-free CNs, respectively, at room temperature. Meanwhile, the UV-vis light irradiation enhanced the CO oxidation activity of PtPd/CNs nanorods by 1.48 fold compared to that in the dark, emanated from the coupling between the drastic inbuilt catalytic merits of PtPd and the inimitable physicochemical properties of CNs. The presented study may pave the way for using CN-based materials in gas conversion reactions.","{'model': 'tldr@v2.0.0', 'text': 'A facile roadmap is presented for the rational design of Pt- and Pd-codoped CN (PtPd/CNs) nanorods via protonation of melamine in an ethylene glycol solution containing Pt and PD precursors using NaNO3 and HCl and subsequent annealing to pave the way for using CN-based materials in gas conversion reactions.'}",
-pd38-08 multi-institutional prospective validation of the novel michigan urological surgery improvement collaborative transperineal biopsy template,M. Maruf,"and IPA of both the PCPT and PBCG risk tools were improved by including Prompt-PGS for both the PCPT and PBCG populations (Table 1). Observed versus expected plots (Figure 1) revealed superior calibration of inclusion of the Prompt-PGS score compared to either risk tool alone in both the PCPT and REDUCE populations, particularly on the lower end of risk. CONCLUSIONS: A germline-genetic risk stratification tool, Prompt e PGS, improves the performance of both the PCPT and PBCG risk tools in two large populations, particularly at the lower end of risk. It may be a useful tool for safely decreasing unnecessary prostate biopsies.","{'model': 'tldr@v2.0.0', 'text': 'A germline-genetic risk stratification tool, Prompt e PGS, improves the performance of both the PCPT and PBCG risk tools in two large populations, particularly at the lower end of risk.'}",https://www.auajournals.org/doi/pdf/10.1097/JU.0000000000000917.08
-a powerful artificial intelligence-based authentication mechanism of retina template using sparse matrix representation with high security,B. Rani,,"{'model': 'tldr@v2.0.0', 'text': 'Mechanized picture characterization frameworks with these AI procedures are created for down-to-earth applications that have the capacity to deal with intra-user fluctuation in the procured biometric identifiers.'}",
-jammer identification technique based on a template matching method,M. Jin,"GNSS has the disadvantage of being vulnerable to jamming, and thus, the necessity of jamming countermeasure techniques has gradually increased. Jamming countermeasure techniques can be divided into an anti-jamming technique and a jammer localization technique. Depending on the type of a jammer, applicable techniques and performance vary significantly. Using an appropriate jamming countermeasure technique, the effect of jamming on a GNSS receiver can be attenuated, and prompt action is enabled when estimating the location of a jammer. However, if an inappropriate jamming countermeasure technique is used, a GNSS receiver may not operate in the worst case. Therefore, jammer identification is a technique that is essential for proper action. In this study, a technique that identifies a jammer based on template matching was proposed. For template matching, analysis of a received jamming signal is required; and the signal analysis was performed using a spectral correlation function. Based on a simulation, it was shown that the proposed identification of jamming signals was possible at various JNR.","{'model': 'tldr@v2.0.0', 'text': 'A technique that identifies a jammer based on template matching was proposed and it was shown that the proposed identification of jamming signals was possible at various JNR.'}",
-inspection method for template library programming conventions,金海,"The invention discloses an inspection method for template library programming conventions and belongs to the technical field of computer parallel programming. According to the inspection method, the programming conventions are described as states, and the migration of the states is described as calling of specific methods in the programming conventions; then the states are implemented into categories, and methods satisfying the programming conventions are written in the categories for calling by programming personnel, so as to ensure that a compiler stops compiling and gives out a corresponding prompt when the programming personnel is against the programming conventions; and possible errors from the programming personnel are predicted and the methods are added in the corresponding categories representing the states so as to provide corresponding prompt information when the programming personnel is against the programming conventions to help the programming personnel to inspect and modify the errors. The inspection method can effectively improve the parallel program writing efficiency of the programming personnel.","{'model': 'tldr@v2.0.0', 'text': 'The inspection method can effectively improve the parallel program writing efficiency of the programming personnel and be used to inspect and modify the errors.'}",
-template system for second language aural comprehension,Donna Mydlarski,"With the introduction of the PROMPT template materials into the French, Italian, and Spanish classes at the Universities of Calgary and Guelph, Canada, it became evident that a system to help language students improve their listening skills was much needed. In response to this need, a template called DICTATE was developed to allow teachers to use the dictation format to practice auditory discrimination, aural comprehension, and orthography in the most efficient and effective way possible.","{'model': 'tldr@v2.0.0', 'text': 'A template called DICTATE was developed to allow teachers to use the dictation format to practice auditory discrimination, aural comprehension, and orthography in the most efficient and effective way possible.'}",
-the 10 building blocks of high-performing primary care,T. Bodenheimer,"Our experiences studying exemplar primary care practices, and our work assisting other practices to become more patient centered, led to a formulation of the essential elements of primary care, which we call the 10 building blocks of high-performing primary care. The building blocks include 4 foundational elements—engaged leadership, data-driven improvement, empanelment, and team-based care—that assist the implementation of the other 6 building blocks—patient-team partnership, population management, continuity of care, prompt access to care, comprehensiveness and care coordination, and a template of the future. The building blocks, which represent a synthesis of the innovative thinking that is transforming primary care in the United States, are both a description of existing high-performing practices and a model for improvement.","{'model': 'tldr@v2.0.0', 'text': 'The building blocks, which represent a synthesis of the innovative thinking that is transforming primary care in the United States, are both a description of existing high-performing practices and a model for improvement.'}",http://www.annfammed.org/content/12/2/166.full.pdf
-let me check the examples: enhancing demonstration learning via explicit imitation,Sirui Wang,"Demonstration learning aims to guide the prompt prediction by providing answered demonstrations in the few shot settings. Despite achieving promising results, existing work only concatenates the answered examples as demonstrations to the prompt template (including the raw context) without any additional operation, neglecting the prompt-demonstration dependencies. Besides, prior research found that randomly replacing the labels of demonstrations marginally hurts performance, illustrating that the model could not properly learn the knowledge brought by the demonstrations. Inspired by the human learning process, in this paper, we introduce Imitation DEMOnstration learning (Imitation-Demo) to strengthen demonstration learning via explicitly imitating human review behaviour, which includes: (1) contrastive learning mechanism to concentrate on similar demonstrations.(2) demonstration-label re-prediction method to consolidate known knowledge. Experiment results show that our proposed method achieves state-of-the-art performance on 5 out of 14 classification corpus. Further studies also prove that Imitation-Demo strengthens the associations between the prompt and demonstrations, which could provide the basis for exploring how demonstration learning works.","{'model': 'tldr@v2.0.0', 'text': 'Inspired by the human learning process, Imitation-Demo is introduced to strengthen demonstration learning via explicitly imitating human review behaviour, which includes a contrastive learning mechanism to concentrate on similar demonstrations and a demonstration-label re-prediction method to consolidate known knowledge.'}",http://arxiv.org/pdf/2209.00455
-stprompt: semantic-guided and task-driven prompts for effective few-shot classification,Jinta Weng,"The effectiveness of prompt learning has been demonstrated in different pre-trained language models. By formulating suitable template and choosing representative label mapping, prompt learning can be used as an efficient knowledge probe. However, finding suitable prompt in existing methods requires multiple experimental attempts or appropriate vector initialization on formulating suitable template and choosing representative label mapping, which it is more common in few-shot learning tasks. Motivating by PLM working process, we try to construct the prompt from task semantic perspective and thus propose the STPrompt -Semantic-guided and Task-driven Prompt model. Specifically, two novel prompts generated from the semantic dependency tree (Dep-prompt) and task-specific metadata description (Meta-prompt), are firstly constructed in a prompt augmented pool, and the proposed model would automatically select a suitable semantic prompt to motivating the prompt learning process. Our results show that the proposed model achieves the state-of-the-art performance in five different datasets of few-shot text classification tasks, which prove that more semantic and significant prompts could assume as a better knowledge proving tool.","{'model': 'tldr@v2.0.0', 'text': 'The proposed STPrompt -Semantic-guided and Task-driven Prompt model achieves the state-of-the-art performance in five different datasets of few-shot text classification tasks, which prove that more semantic and significant prompts could assume as a better knowledge proving tool.'}",
-supplementary material for mask-free ovis: open-vocabulary instance segmentation without manual mask annotations,VS Vibashan,"COCO Caption vs Image-label pseudo-caption: Pseudo-caption generation: Since the pre-trained vision-language models are trained on full sentences, we need to feed the image-labels into a prompt template first, and use them to generate a pseudo-captions. Specifically, given image-labels [category-1,category-2...,category-n], we randomly sample a prompt from 63 prompt templates [1,6] and the pseudo-caption are generated as ”{Prompt-x} + {category-1 and category-2 and ... category-n}”. For example, as shown in Fig. 1 bottom row the sampled prompts are ”A black and white photo of the {category}.” and ”A photo of {category} in the scene.” and the image-labels are ”zebra” and ”giraffe”. Thus, the generated pseudo-captions are ”A black and white photo of the zebra and giraffe.” and ”A photo of one zebra and giraffe in the scene.”","{'model': 'tldr@v2.0.0', 'text': 'Pseudo-caption generation: Since the pre-trained vision-language models are trained on full sentences, they need to feed the image-labels into a prompt template first, and use them to generate a pseudo-captions.'}",
-pre-training extractive question-answer prompts for few-shot chinese text classification,Gaojian Ding,". In recent years, pre-training models (PLMs) have made impressive progress, and prompt learning has made few-shot learning achievable. However, traditional prompt learning methods often require manual template design, or performance may be unstable due to the limited data in few-shot tasks. To address these issues, we propose a few-shot text classification method based on multi-task learning. We first unify the multi-task into an extractive question-answering (EQA) format, then train the prompt using task data in the unified format. The prompt cists of modular prompts and a router that indicates their functionality. We then initonsialize the downstream training parameters using the router of a pre-training task similar to the downstream task and employ contrastive learning to improve EQA efficiency.","{'model': 'tldr@v2.0.0', 'text': 'This work unify the multi-task into an extractive question-answering (EQA) format, then train the prompt using task data in the unified format and initonsialize the downstream training parameters using the router of a pre-training task similar to the downstream task and employ contrastive learning to improve EQA efflciency.'}",
-enhancing cross-lingual natural language inference by soft prompting with multilingual verbalizer,Shuang Li,"Cross-lingual natural language inference is a fundamental problem in cross-lingual language understanding. Many recent works have used prompt learning to address the lack of annotated parallel corpora in XNLI. However, these methods adopt discrete prompting by simply translating the templates to the target language and need external expert knowledge to design the templates. Besides, discrete prompts of human-designed template words are not trainable vectors and can not be migrated to target languages in the inference stage flexibly. In this paper, we propose a novel Soft prompt learning framework with the Multilingual Verbalizer (SoftMV) for XNLI. SoftMV first constructs cloze-style question with soft prompts for the input sample. Then we leverage bilingual dictionaries to generate an augmented multilingual question for the original question. SoftMV adopts a multilingual verbalizer to align the representations of original and augmented multilingual questions into the same semantic space with consistency regularization. Experimental results on XNLI demonstrate that SoftMV can achieve state-of-the-art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.","{'model': 'tldr@v2.0.0', 'text': 'A novel Soft prompt learning framework with the Multilingual Verbalizer (SoftMV) for XNLI and demonstrates that SoftMV can achieve state-of-the-art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings.'}",http://arxiv.org/pdf/2305.12761
-exploring prompts in few-shot cross-linguistic topic classification scenarios,Zhipeng Zhang,"In recent years, large-scale pretrained language models have become widely used in natural language processing tasks. On this basis, prompt learning has achieved excellent performance in specific few-shot classification scenarios. The core idea of prompt learning is to convert a downstream task into a masked language modelling task. However, different prompt templates can greatly affect the results, and finding an appropriate template is difficult and time-consuming. To this end, this study proposes a novel hybrid prompt approach, which combines discrete prompts and continuous prompts, to motivate the model to learn more semantic knowledge from a small number of training samples. By comparing the performance difference between discrete prompts and continuous prompts, we find that hybrid prompts achieve the best results, reaching a 73.82% F1 value in the test set. In addition, we analyze the effect of different virtual token lengths in continuous prompts and hybrid prompts in a few-shot cross-language topic classification scenario. The results demonstrate that there is a threshold for the length of virtual tokens, and too many virtual tokens decrease the performance of the model. It is better not to exceed the average length of the training set corpus. Finally, this paper designs a method based on vector similarity to explore the real meanings represented by virtual tokens. The experimental results show that the prompt automatically learnt from the virtual token has a certain correlation with the input text.","{'model': 'tldr@v2.0.0', 'text': 'This study proposes a novel hybrid prompt approach, which combines discrete prompts and continuous prompts, to motivate the model to learn more semantic knowledge from a small number of training samples, and designs a method based on vector similarity to explore the real meanings represented by virtual tokens.'}",https://www.mdpi.com/2076-3417/13/17/9944/pdf?version=1693814326
-successful use of a physiologically acceptable artificial skin in the treatment of extensive burn injury,J. Burke,"A bilayer artificial skin composed of a temporary Silastic epidermis and a porous collagcn-chondroitin 6-sulfate fibrillar dermis, which is not removed, has been used to physiologically close up to 60% of the body surface following prompt excision of burn wounds in ten patients whose total burn size covered 50–95% body surface area (BSA). Following grafting, the dermal portion is populated with fibroblasts and vessels from the wound bed. The anatomic structure of the artificial dermis resembles normal dermis and serves as a template for the synthesis of new connective tissue and the formation of a “neodermis,” while it is slowly biodegraded. This artificial skin has physiologically closed excised burn wounds for periods of time up to 46 days before the Silastic epidermis was removed. At the time of election when donor sites are ready for reharvesting, the Silastic epidermis is removed from the vascularized artificial dermis and replaced with 0.004 auto-epidermal graft in sheet or meshed form. Clinical and histologic experience in a relatively short follow-up period (2–16 months) indicates that “neodermis” retains some of the anatomic characteristics and behavior of normal dermis, thus promising improvement in the functional and cosmetic results, as well as providing physiologic function as a skin substitute. The artificial skin is easily sterilized and stored at room temperature, capable of large scale production, and immediately available for grafting, indicating its potential for easy and relatively economic use in the burn patient","{'model': 'tldr@v2.0.0', 'text': 'Clinical and histologic experience in a relatively short follow-up period indicates that “neodermis” retains some of the anatomic characteristics and behavior of normal dermis, thus promising improvement in the functional and cosmetic results, as well as providing physiologic function as a skin substitute.'}",
-structural snapshots of la crosse virus polymerase reveal the mechanisms underlying peribunyaviridae replication and transcription,B. Arragain,,"{'model': 'tldr@v2.0.0', 'text': 'It is found that replication initiation is coupled to distal duplex promoter formation, endonuclease movement, prime-and-realign loop extension and closure of the polymerase core that direct the template towards the active site.'}",https://www.nature.com/articles/s41467-022-28428-z.pdf
-wound tissue can utilize a polymeric template to synthesize a functional extension of skin.,I. Yannas,"Prompt and long-term closure of full-thickness skin wounds is guinea pigs and humans is achieved by applying a bilayer polymeric membrane. The membrane comprises a top layer of a silicone elastomer and a bottom layer of a porous cross-linked network of collagen and glycosaminoglycan. The bottom layer can be seeded with a small number of autologous basal cells before grafting. No immunosuppression is used and infection, exudation, and rejection are absent. Host tissue utilizes the sterile membrane as a culture medium to synthesize neoepidermal and neodermal tissue. A functional extension of skin over the entire wound area is formed in about 4 weeks.","{'model': 'tldr@v2.0.0', 'text': 'A functional extension of skin over the entire wound area is formed in about 4 weeks and no immunosuppression is used and infection, exudation, and rejection are absent.'}",
-long-lived ag10 6+ luminescence and a split dna scaffold.,J. Petty,"Molecular silver clusters emit across the visible to near-infrared, and specific chromophores can be formed using DNA strands. We study C4AC4TC3G that selectively coordinates and encapsulates Ag10 6+, and this chromophore has two distinct electronic transitions. The green emission is strong and prompt with ϕ = 18% and τ = 1.25 ns, and the near-infrared luminescence is weaker, slower with τ = 50 µs, and is partly quenched by oxygen, suggesting phosphorescence. This lifetime can be modulated by the DNA host, and we consider two derivatives of C4AC4TC3G with similar sequences but distinct structures. In one variant, thymine was excised to create an abasic gap in an otherwise intact strand. In the other, the covalent phosphate linkage was removed to split the DNA scaffold into two fragments. In relation to the contiguous strands, the broken template speeds the luminescence decay by twofold, and this difference may be due to greater DNA flexibility. These modifications suggest that a DNA can be structurally tuned to modulate metastable electronic states in its silver cluster adducts.","{'model': 'tldr@v2.0.0', 'text': 'Modifications of C4AC4TC3G suggest that a DNA can be structurally tuned to modulate metastable electronic states in its silver cluster adducts.'}",
-the formless void as organizational template☆,T. Harle,"This article examines the first creation story in the Book of Genesis from the perspective of complexity theory. Traditional views of this narrative conceive of a deity harnessing chaos to create order. They reflect a loathing of chaos and a desire for control: a Newtonian worldview of cause and effect. It is an approach we see repeated in contemporary organizations. However, a radically different perspective sees the formless void as opening up the possibility of emergence and self-organization. This approach is used to prompt interaction both with the Judaeo-Christian tradition and organization studies. Four particular leadership challenges are examined: notions of control, attitudes to change, co-creating an environment for self-organization, and using emergence in relation to values. A concluding section highlights the importance of consistency in leadership to promote emergence.",,
-overview of tertiary addictions services response to opioid dependence during the covid-19 pandemic.,K. Hennigan,"The emergence of the COVID-19 pandemic has presented the addiction services with an unprecedented set of challenges. Opioid users are particularly vulnerable because of their high level of pre-existing health problems and lifestyle factors. In order to minimise their risks to self and to others in the current Covid-19 crisis, the service has sought to urgently identify vulnerable individuals, and induct them into OST treatment promptly. Additionally, several guidelines have been created and regularly updated by the HSE for any healthcare staff working with opioid users. These include guidance documents, to facilitate prompt induction of patients onto the OST programme, the prescribing of naloxone to all patients at risk of overdose, eConsultation, medication management for those in self-isolation, and the delivery of injecting equipment. The guidance documents and resources will provide a template for a new way of working for the sector during these challenging times and into the future.","{'model': 'tldr@v2.0.0', 'text': 'Several guidelines have been created and regularly updated by the HSE for any healthcare staff working with opioid users to facilitate prompt induction of patients onto the OST programme, and the prescribing of naloxone to all patients at risk of overdose.'}",
-grounding language to entities and dynamics for generalization in reinforcement learning,,"To collect the text manual, we first crowdsource 82 templates (with 2,214 possible descriptions after filling in the blanks). Each Amazon Mechanical Turk worker is asked to paraphrase a prompt sentence while preserving words in boldface (which become the blanks in our templates). We have three blanks per template, one each for the entity, role and an adjective. For each role (enemy, message, goal) we have three role words and three adjectives that are synonymous (Table 5). Each entity is also described in three synonymous ways. Thus, every entity-role assignment can be described in 27 different ways on the same template. Raw templates are filtered for duplicates, converted to lowercase, and corrected for typos to prevent confusion on downstream collection tasks.","{'model': 'tldr@v2.0.0', 'text': None}",
-kinect-based intelligent monitoring and warning of students' sitting posture,Heng Sun,"This paper proposes a Kinect-based software design for intelligent detection and warning of students' sitting posture. The plan includes three modules: video acquisition, preprocessing, and sitting posture recognition. Two methods, template matching method based on bone data and feature learning method based on Convolutional neural network (CNN), are used to realize sitting posture recognition. This design first uses Kinect to complete video capture, and secondly, preprocesses the real-time image frames in the video stream to extract the region of interest. Calculate the spatial position of the main joint points of the upper body of the human body again, and convert them into different spatial coordinate systems to locate the target. Finally, the best result of the two methods is used as the recognition result. When an incorrect posture is judged, a corresponding prompt or warning is issued to the supervised person. The experiment was carried out to realize the real-time detection and judgment of the student's sitting posture and timely notification of bad posture, which verified the software design scheme's feasibility.","{'model': 'tldr@v2.0.0', 'text': ""The experiment was carried out to realize the real-time detection and judgment of the student's sitting posture and timely notification of bad posture, which verified the software design scheme's feasibility.""}",
-circulating mirnas with prognostic value in metastatic breast cancer and for early detection of metastasis.,Dharanija Madhavan,"Metastasis is the principal cause of high morbidity and mortality among breast cancer (BC) patients. Identification of markers that can be routinely monitored to predict onset of metastasis in BC patients and prognosis of metastatic breast cancer (MBC) patients would increase their median survival. In this study, plasma miRNAs of 40 MBC patients were profiled by TaqMan low density arrays and miRNAs with prognostic capacity were identified. The candidates were validated initially in the samples of 237 MBC patients and subsequently in 335 samples from an independent study cohort of BC patients. Sixteen miRNAs were established to be significantly associated with overall survival, and were termed as prognostic miRNA panel template (PROMPT). These included miR-141, miR-144, miR-193b, miR-200a, miR-200b, miR-200c, miR-203, miR-210, miR-215, miR-365, miR-375, miR-429, miR-486-5p, miR-801, miR-1260 and miR-1274a. Additionally, 11 of these miRNAs were also associated with progression-free survival. Their prognostic significance was further confirmed in samples from a second study cohort of BC patients. In addition, miR-200a, miR-200b, miR-200c, miR-210, miR-215 and miR-486-5p were found to be significantly associated with onset of metastasis up to 2 years prior to clinical diagnosis in BC patients. We have thus identified panels of miRNAs, which include metastasis promoting miR-200 family and miR-203, as well as oncogenic and tumor-suppressive miRNAs, that can serve as prognostic markers for MBC, and early detection markers of metastasis in BC.","{'model': 'tldr@v2.0.0', 'text': 'Panel of miRNAs, which include metastasis promotingmiR-200 family and miR-203, as well as oncogenic and tumor-suppressive mi RNAs, that can serve as prognostic markers for MBC, and early detection markers of metastasis in BC are identified.'}",https://academic.oup.com/carcin/article-pdf/37/5/461/8193240/bgw008.pdf
-kilohertz gravitational waves from binary neutron star remnants: time-domain model and constraints on extreme matter,M. Breschi,"The remnant star of a neutron star merger is an anticipated loud source of kiloHertz gravitational waves that conveys unique information on the equation of state of hot matter at extreme densities. Observations of such signals are hampered by the photon shot noise of ground-based interferometers and pose a challenge for gravitational-wave astronomy. We develop an analytical time-domain waveform model for postmerger signals informed by numerical relativity simulations. The model completes effective-one-body waveforms for quasi-circular nonspinning binaries in the kiloHertz regime. We show that a template-based analysis can detect postmerger signals with a minimal signal-to-noise ratios (SNR) of 8, corresponding to GW170817-like events for third-generation interferometers. Using Bayesian model selection and the complete inspiral-merger-postmerger waveform model it is possible to infer whether the merger outcome is a prompt collapse to a black hole or a remnant star. In the latter case, the radius of the maximum mass (most compact) nonrotating neutron star can be determined to kilometer precision. We demonstrate the feasibility of inferring the stiffness of the equation of state at extreme densities using the quasiuniversal relations deduced from numerical-relativity simulations.",,https://www.repository.cam.ac.uk/bitstream/1810/301786/3/PhysRevD.100.104029.pdf
-financial scenario modelling: a study of uk universities,P. Cropper,"ABSTRACT Financial scenario modelling appears to offer the potential to help universities cope with an uncertain funding environment. By means of a questionnaire survey of UK universities, supported by interviews with members of finance departments, this paper explores the prevalence and construction of financial models used for scenario analysis. According to the findings, most universities undertake such modelling; but a minority do not, even though it is officially required. There is some variation in the purposes for which the models are constructed and in the sophistication of the approaches taken, but the overall impression is of a relatively straightforward approach, with spreadsheets generally viewed as having sufficient functionality to meet current strategic needs. There is no appetite for a standard template to be imposed or even offered by the funding councils. Sharing experience is likely to be a more welcome approach if growing pressures prompt finance departments to strengthen their modelling.",,https://pure.hud.ac.uk/ws/files/35714977/AMENDED_JHEPM_scen_R1_full.pdf
-cardiogenic shock: evolving definitions and future directions in management,Tara L Jones,"Cardiogenic shock (CS) is a complex and highly morbid entity conceptualised as a vicious cycle of injury, cardiac and systemic decompensation, and further injury and decompensation. The pathophysiology of CS is incompletely understood but limited clinical trial experience suggests that early and robust support of the failing heart to allow for restoration of systemic homoeostasis appears critical for survival. We review the pathophysiology, clinical features and trial data to construct a contemporary model of CS as a systemic process characterised with maladaptive compensatory mechanisms requiring prompt and appropriately tailored medical and mechanical support for optimal outcomes. We conclude with an algorithmic approach to acute CS incorporating clinical and haemodynamic data to match the patient’s cardiac and systemic needs as a template for contemporary management.","{'model': 'tldr@v2.0.0', 'text': 'An algorithmic approach to acute CS is taken incorporating clinical and haemodynamic data to match the patient’s cardiac and systemic needs as a template for contemporary management.'}",https://openheart.bmj.com/content/openhrt/6/1/e000960.full.pdf
-a brachytherapy template approach to standardize saturation prostatic biopsy,Simon Bott,"The increasing use of PSA testing has resulted in more men undergoing prostatic biopsy to diagnose or exclude malignancy. However, the standard sextant or octant biopsy regimen is reported to miss 15–35% of ‘significant’ prostate cancers [1–3]. The finding of ‘no evidence of malignancy’ on standard biopsy in the presence of a rising PSA level, a rapid PSA doubling time or a suspicious DRE may prompt further biopsies. Several studies have reported sites inadequately sampled using standard techniques, including the anterior transition zone superior and lateral to the urethra, the inferior part of the anterior horn where the peripheral zone wraps round the transition zone, and the midline of the peripheral zone [2,4,5]. Increasing the number of biopsies and targeting these sites after a negative standard prostatic biopsy results in increased cancer detection [1,2,5]. Saturation biopsy involves taking a large number of biopsies (14–45) using a periprostatic block, sedation, spinal or general spinal anaesthesia [6–8]. In men with a clinical suspicion of prostate cancer and previously negative sextant biopsies, transrectal saturation biopsies have a cancer detection rate of 14–34% [6–8]. We describe a systematic and comprehensive technique to take saturation prostatic biopsies through the perineum.","{'model': 'tldr@v2.0.0', 'text': 'A systematic and comprehensive technique to take saturation prostatic biopsies through the perineum is described, which is reported to miss 15–35% of ‘significant’ prostate cancers and in men with a clinical suspicion of prostate cancer and previously negative sextantBiopsies have a cancer detection rate of 14–34%.'}",
-pre-notification and reminder sms text messages with behaviourally informed invitation letters to improve uptake of nhs health checks: a factorial randomised controlled trial,A. Sallis,,"{'model': 'tldr@v2.0.0', 'text': 'This large randomised controlled trial adds support to the evidence that small, low cost behaviourally informed changes to letter-based invitations can increase uptake of NHS HCs and provides novel evidence on the effect of SMS reminders and pre-notification on NHS HC attendance.'}",https://bmcpublichealth.biomedcentral.com/track/pdf/10.1186/s12889-019-7476-8
-variations in influenza vaccination by clinic appointment time and an active choice intervention in the electronic health record to increase influenza vaccination,Rebecca H Kim,"Key Points Question Do influenza vaccination rates vary by clinic appointment time and can an active choice intervention in the electronic health record directed to medical assistants improve vaccination rates in primary care practices? Findings In this quality improvement study of 11 primary care practices and 96 291 patients, influenza vaccination rates significantly declined as the clinic day progressed. Primary care practices that implemented an active choice intervention in the electronic health record to prompt medical assistants to ask patients about influenza vaccination and template vaccination orders for clinicians to review were associated with a significant increase in influenza vaccination rates when compared with a control group of primary care practices. Meaning Reminders in the electronic health record can improve vaccination rates overall, but other strategies may also be needed to address variations by time of day.","{'model': 'tldr@v2.0.0', 'text': 'Primary care practices that implemented an active choice intervention in the electronic health record to prompt medical assistants to ask patients about influenza vaccination were associated with a significant increase in influenza vaccination rates when compared with a control group of primary care practices.'}",https://jamanetwork.com/journals/jamanetworkopen/articlepdf/2702210/kim_2018_oi_180107.pdf
-synthesis and anticancer properties of ‘azole’ based chemotherapeutics as emerging chemical moieties: a comprehensive review,P. Prasher," Azole frameworks serve as privileged scaffolds in the contemporary drug design paradigm owing to their unique physicochemical profile that promotes the development of highly selective, physiological benevolent chemotherapeutics. Several azole nuclei function as bioisostere in medicinal chemistry and prompt the development of tailored therapeutics for targeting the desired biological entities. Besides, the azole scaffold forms an integral part in the advanced drug designing methodologies, such as target template insitu drug synthesis, that assists in rapid identification of the hit molecules form a diverse pool of leads; and direct biomolecule-drug conjugation, along with bioorthogonal strategies that ensure localization, and superior target specificity of the directed therapeutic. Lastly, the structural diversity of azole framework and high yielding click synthetic methods provide a comprehensive Structure-Activity Relationship analysis for design optimization of the potential drug molecules by fine-tuning the placement of different substituents critical for the activity. This review provides a comprehensive analysis of the synthesis and anticancer potential of azole based chemotherapeutics. ","{'model': 'tldr@v2.0.0', 'text': 'The structural diversity of azole framework and high yielding click synthetic methods provide a comprehensive Structure-Activity Relationship analysis for design optimization of the potential drug molecules by fine-tuning the placement of different substituents critical for the activity.'}",
-comparing emissions mitigation efforts across countries,Joseph E. Aldy,"A natural outcome of the emerging pledge and review approach to international climate change policy is the interest in comparing mitigation effort among countries. Domestic publics and stakeholders will have an interest in knowing if peer countries are undertaking (or planning to undertake) comparable efforts in mitigating their GHG emissions. Moreover, if the aggregate effort is considered inadequate in addressing the risks posed by climate change, then this will likely prompt a broader interest in identifying those countries where greater effort is arguably warranted based on comparison with their peers. Both assessments require metrics of effort and comparisons among countries. We propose a framework for such an exercise, drawing from a set of principles for designing and implementing informative metrics. We present a template for organizing metrics on mitigation effort, for both ex ante and ex post review. We also provide preliminary assessments of effort along emissions, price, and cost metrics for post-2020 climate policy contributions by China, the European Union, Russia, and the United States. We close with a discussion of the role of academics and civil society in promoting transparency and facilitating the evaluation and comparison of effort. Policy relevance Statement: Our article presents a framework for the review of intended nationally determined contributions and the ex post review of contributions under the UNFCCC negotiations. We provide an illustration of this framework with an energy-economic model. Our work focuses on how countries may use the review to compare mitigation effort – planned under INDCs and delivered by implementation of the pledged contributions – to address concerns about equity, efficiency, competitiveness, and the stability of any agreement that arise in international negotiations.",,https://dash.harvard.edu/bitstream/1/30203383/6/Comparability%20APA%20Paper%20150930%20Final.pdf
-factors underlying quality problems with alcohol screening prompted by a clinical reminder in primary care: a multi-site qualitative study,E. Williams,,"{'model': 'tldr@v2.0.0', 'text': 'This qualitative study in nine clinics found that implementation of alcohol screening facilitated by a clinical reminder resulted primarily in verbal screening in which questions were not asked vertbatim and were otherwise adapted, addressing limitations of the clinical reminder and issues related to perceived discomfort.'}",https://link.springer.com/content/pdf/10.1007/s11606-015-3248-z.pdf
-an electrochemical non-enzymatic glucose sensor based on ultrathin pdag single-crystalline nanowires.,Xiaoxiang Xu,"Electrochemical non-enzymatic sensors have great potential for prompt and efficient detection of glucose. Herein, a novel, highly efficient electrochemical non-enzymatic glucose sensor is reported that is based on ultrathin PdAg single-crystalline nanowires (NWs). Ultrathin PdAg NWs are fabricated by a facile one-pot aqueous synthesis through an in situ growth strategy with an amphiphilic surfactant as the template. A comparison of the activities of PdAg NWs with different compositional ratios and nanostructures shows that ultrathin Pd2 Ag1 NWs hold the best performance toward electrochemical detection of glucose with an operable sensitivity of 11.6 μA mM-1 cm-2 and a linear response range of 0.1-8 mM. Structural and compositional features of the Pd2 Ag1 NWs allow an excellent selectivity, rapid response, and good long-term stability for electrochemical glucose sensor. This work thus provides a new possibility for the rational design and synthesis of noble-metal-based nanomaterials for non-enzymatic sensors.","{'model': 'tldr@v2.0.0', 'text': 'A novel, highly efficient electrochemical non-enzymatic glucose sensor is reported that is based on ultrathin PdAg single-crystalline nanowires (NWs) that provides a new possibility for the rational design and synthesis of noble-metal-based nanomaterials for non- enzymatic sensors.'}",
-liquid xenon scintillation measurements and pulse shape discrimination in the lux dark matter detector,D. Akerib,"Weakly Interacting Massive Particles (WIMPs) are a leading candidate for dark matter and are expected to produce nuclear recoil (NR) events within liquid xenon time-projection chambers. We present a measurement of liquid xenon scintillation characteristics in the LUX dark matter detector and develop a pulse shaped based discrimination parameter to be used for particle identification. To accurately measure the scintillation characteristics, we develop a template-fitting method to reconstruct the detection time of photons. Analyzing calibration data collected during the 2013-16 LUX WIMP search, we measure a singlet-to-triplet scintillation ratio for electron recoils (ER) that is consistent with existing literature, and we make a first-ever measurement of the NR singlet-to-triplet ratio at recoil energies below 74 keV. A prompt fraction discrimination parameter exploits the difference of the photon time spectra for NR and ER events and is optimized to have the least number of ER events that occur in the 50\% NR acceptance region. When this discriminator is used in conjunction with charge-to-light discrimination on the calibration data, the signal-to-noise ratio in the NR dark matter acceptance region increases by up to a factor of two.",,https://link.aps.org/accepted/10.1103/PhysRevD.97.112002
-pokérator - unveil your inner pokémon,Dominique Geissler,"The Pokérator is a generator of Pokémon names and descriptions, based on user input. The names are generated by blending words based on syllables or characters according to a bigram language model. An accompanying description is generated by filling a template with ConceptNet answers. This sentence is then used as a prompt for text generation with the GPT-2 language model which was finetuned on Pokédex entries. The evaluation of the generated Pokémon names shows that the names are not realistic, but appreciated and creative.","{'model': 'tldr@v2.0.0', 'text': 'The Pokérator is a generator of Pokémon names and descriptions, based on user input, which shows that the names are not realistic, but appreciated and creative.'}",
-measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at √ s = 7 tev,A. Belyaev,"A measurement of differential cross sections for the production of a pair of isolated photons in proton–proton collisionsat √ s = 7 TeVispresented.Thedatasamplecorre-sponds to an integrated luminosity of 5.0 fb − 1 collected with the CMS detector. A data-driven isolation template method is used to extract the prompt diphoton yield. The measured cross section for two isolated photons, with transverse energy above 40 and 25 GeV respectively, in the pseudorapidity range | η | < 2 . 5, | η | / ∈ [ 1 . 44 , 1 . 57 ] and with an angular separation (cid:3) R > 0 . 45, is 17 . 2 ± 0 . 2 (stat) ± 1 . 9 (syst) ± 0 . 4 (lumi)pb. Differential cross sections are measured as a function of the diphoton invariant mass, the diphoton transverse momentum, the azimuthal angle difference between the two photons, and the cosine of the polar angle in the Collins–Soper reference frame of the diphoton system. The results are compared to theoretical predictions at leading, next-to-leading, and next-to-next-to-leading order in quantum chromodynamics.",,
-optical and near-infrared observations of sn 2013dx associated with grb 130702a,V. Toy,"We present optical and near-infrared (NIR) light curves and optical spectra of SN 2013dx, associated with the nearby (redshift 0.145) gamma-ray burst GRB 130702A. The prompt isotropic gamma-ray energy released from GRB 130702A is measured to be E γ , iso = 6.4 − 1.0 + 1.3 × 10 50 ?> erg (1 keV to 10 MeV in the rest frame), placing it intermediate between low-luminosity GRBs like GRB 980425/SN 1998bw and the broader cosmological population. We compare the observed g ′ r ′ i ′ z ′ ?> light curves of SN 2013dx to a SN 1998bw template, finding that SN 2013dx evolves ∼20% faster (steeper rise time), with a comparable peak luminosity. Spectroscopically, SN 2013dx resembles other broad-lined SNe Ic, both associated with (SN 2006aj and SN 1998bw) and lacking (SN 1997ef, SN 2007I, and SN 2010ah) gamma-ray emission, with photospheric velocities around peak of ∼ 21,000 km s−1. We construct a quasi-bolometric ( g ′ r ′ i ′ z ′ yJ ?> ) light curve for SN 2013dx, only the fifth GRB-associated SN with extensive NIR coverage and the third with a bolometric light curve extending beyond Δ t > 40 days ?> . Together with the measured photospheric velocity, we derive basic explosion parameters using simple analytic models. We infer a 56Ni mass of M Ni = 0.37 ± 0.01 ?> M ⊙ ?> , an ejecta mass of M ej = 3.1 ± 0.1 ?> M ⊙ ?> , and a kinetic energy of E K = ( 8.2 ± 0.43 ) × 10 51 ?> erg (statistical uncertainties only), consistent with previous GRB-associated supernovae. When considering the ensemble population of GRB-associated supernovae, we find no correlation between the mass of synthesized 56Ni and high-energy properties, despite clear predictions from numerical simulations that M Ni ?> should correlate with the degree of asymmetry. On the other hand, M Ni ?> clearly correlates with the kinetic energy of the supernova ejecta across a wide range of core-collapse events.",,https://iopscience.iop.org/article/10.3847/0004-637X/818/1/79/pdf
-"measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at $$\sqrt{s}=7\,\text {tev} $$s=7tev",S. Chatrchyan,,,https://link.springer.com/content/pdf/10.1140/epjc/s10052-014-3129-3.pdf
-writing in math: a disciplinary literacy approach,W. Brozo,"ABSTRACT Mathematics teachers often resist generic literacy strategies because they do not seem relevant to math learning. Discipline-specific literacy practices that emerge directly from the math content and processes under study are more likely to be embraced by math teachers. Furthermore, national and state-level mathematics standards as well as Common Core standards provide frameworks for situating literacy practices squarely within the disciplines. A disciplinary literacy approach to writing in math requires teachers to develop innovative strategies and practices that link writing to particular mathematical processes and tasks. An example is shared of a math writing approach developed by a middle school teacher used to prompt her students' critical thinking and problem solving processes during the study of algebra. She designed a template that when completed can serve as a reflective tool for her students and provide the teacher useful feedback on their learning. The example of teaching with the template as a guide for working through steps to solve a story problem demonstrates what disciplinary writing can look like in a typical middle school classroom.",,
-the clinician as clinical ethics consultant: an empirical method of study,D. Kornfeld,"Some 30 years ago the role of the clinical ethics consultant (CEC) was formalized. At the time, the perception of the role differed between two groups serving in that capacity, clinicians and nonclinicians. Differences in their roles reflected their training and experience. These divergent views were resolved semantically by designating the role of the CEC as “ethics facilitation.” In practice the different perspectives have remained. However, the subsequent published literature on clinical ethics consultation has not adequately reflected the activity of the clinician as a CEC. There have been recurring unanswered calls for the acquisition of empirical data on the nature of the problems that prompt ethics consultation requests and the functions required to address them. The authors introduce a template that provides a means to acquire such data for clinician ethicists. A similar instrument could be constructed to reflect the role of the nonclinician ethicist serving in that capacity.","{'model': 'tldr@v2.0.0', 'text': 'A template is introduced that provides a means to acquire empirical data on the nature of the problems that prompt ethics consultation requests and the functions required to address them and a similar instrument could be constructed to reflect the role of the nonclinician ethicist serving in that capacity.'}",
-women’s views on anxiety in pregnancy and the use of anxiety instruments: a qualitative study,K. Evans,"Abstract Objective: To explore women’s experience of anxiety in pregnancy and views on the use of anxiety instruments in antenatal care. Background: Anxiety in pregnancy is associated with adverse birth outcomes, developmental and behavioural problems in infants and postnatal depression. Despite recommendations for routine psychological assessment in pregnancy, the optimal methods to identify anxiety in pregnancy have not been confirmed. Methods: A qualitative study using two focus group discussions was undertaken. Focus group one included women in a community setting and focus group two included women in a hospital clinic setting who had received additional support for anxiety in pregnancy. Participants were women who had given birth within the past nine months and considered themselves to have been anxious during their pregnancy. Results: Three main themes were identified using template analysis: sources of support, administration of anxiety instruments and the use of instruments to prompt discussion. Women stated that anxiety instruments could help them to identify their anxious feelings and prompt a discussion around those feelings. However, they expressed concerns surrounding the administration of anxiety instruments and questioned how useful they would be in helping women access help and support. Conclusions: The introduction of anxiety instruments in antenatal care may present an opportunity to discuss women’s emotional health and anxieties. Providing women with sufficient time to discuss their anxious feelings, identified by such instruments, could facilitate access to additional support.","{'model': 'tldr@v2.0.0', 'text': 'The introduction of anxiety instruments in antenatal care may present an opportunity to discuss women’s emotional health and anxieties, and providing women with sufficient time to discuss their anxious feelings, identified by such instruments, could facilitate access to additional support.'}",https://nottingham-repository.worktribe.com/preview/829175/JRIP%20FG%20Anx%20KE%20resubmission%20text%20%28non-anon%29%2011.7.16%20proof.pdf
-linguistics vanguard,Mark B. Turner,A range of multimodal form-meaning pairs has arisen to prompt for the generic integration template called blended classic joint attention (BCJA). This article presents examples and principles.,,
-pcr detection of escherichia coli o157:h7 directly from stools: evaluation of commercial extraction methods for purifying fecal dna,J. Holland,"ABSTRACT Rapid identification of Escherichia coli O157:H7 is important for patient management and for prompt epidemiological investigations. We evaluated one in-house method and three commercially available kits for their ability to extract E. coli O157:H7 DNA directly from stool specimens for PCR. Of the 153 stool specimens tested, 107 were culture positive and 46 were culture negative. The sensitivities and specificities of the in-house enrichment method, IsoQuick kit, NucliSens kit, and QIAamp kit were comparable, as follows: 83 and 98%, 85 and 100%, 74 and 98%, and 86 and 100%, respectively. False-negative PCR results may be due to the presence of either inherent inhibitors or small numbers of organisms. The presence of large amounts of bacteria relative to the amount of the E. coli O157:H7 target may result in the lower sensitivities of the assays. All commercial kits were rapid and easy to use, although DNA extracted with the QIAamp kit did not require further dilution of the DNA template prior to PCR.","{'model': 'tldr@v2.0.0', 'text': 'All commercial kits were rapid and easy to use, although DNA extracted with the QIAamp kit did not require further dilution of the DNA template prior to PCR, and the sensitivities and specificities of the in-house enrichment method, IsoQuick kit, NucliSens kit, and QIAamps kit were comparable.'}",https://jcm.asm.org/content/jcm/38/11/4108.full.pdf
-improving human papillomavirus (hpv) vaccination in the postpartum setting,R. Soon,,"{'model': 'tldr@v2.0.0', 'text': 'A drop-down menu that was inserted into the EMR postpartum note template inquiring about the patient’s HPV vaccination status and whether or not the vaccine was administered at that visit could be an effective way to promote HPV vaccination in the post partum setting.'}",https://europepmc.org/articles/pmc5510873?pdf=render
-chapter 9. interfaces for capturing the experiment record,Cerys Willoughby,"Electronic Laboratory Notebooks (ELNs) have many features that can support the work of scientists, but their main purpose is to replace the paper notebook in the laboratory. When we capture notes in a paper notebook, we have the flexibility to record our research in any format we choose. When we capture notes using an electronic notebook, and particularly using ELNs, the way that we capture information changes. Many ELNs make use of templates to encourage the capture of standard information and provide the opportunity to prompt users to capture items of interest that they might otherwise omit or forget to record. Using the cues provided by the formal structure, templates should improve the consistency and quality of records. There is, however, the possibility that templates may have negative as well as positive impacts because they may restrict the content that is recorded, and information that is not specifically prompted for in the template may be omitted, even if it is significant to the record. In this chapter, we will examine research into the impacts of different kinds of interfaces on the information that is captured in the laboratory. We also take a look at the role of mobile applications in bridging the gap between paper notebooks and ELNs, and how other kinds of interfaces may better support users in capturing their research in electronic form.","{'model': 'tldr@v2.0.0', 'text': 'Research into the impacts of different kinds of interfaces on the information that is captured in the laboratory is examined, including the role of mobile applications in bridging the gap between paper notebooks and ELNs, and how other types of interfaces may better support users in capturing their research in electronic form.'}",
-expectancies and motivations to attend an informal science lecture series,Niveen Abighannam,"This study explored the expectancies and motivations that prompt audiences to attend a university science lecture series. The series features talks by science experts from the host campus and around the USA. Each lecture typically attracts between 300 and 600 attendees, including middle and high school student groups, university students, and families and adults from the area. We conducted 47 semi-structured interviews with attendees in order to evaluate their expectancies and motivations. A template analysis of the interviews was grounded in social cognitive and self-determination theories. Results suggest that participants were mostly driven by intrinsic motivations and acquired strong sensory outcome expectancies, such as novelty and activity. Participants also held physical outcome expectancies, such as social expectancies, though to a lesser extent. Both intrinsic and extrinsic motivations to attend the lecture series were associated with expectancies held prior to the event. Of those expectancies, the novelty, entertainment, and social outcomes were dominant. Other noteworthy outcome expectancies include status and self-reactivity. Parents and teachers held outcome expectancies, not only for themselves, but also for their children and students who attended the talks with them.",,
-optical and conductive properties of as-synthesized organic-capped tio₂ nanorods highly dispersible in polystyrene-block-poly(methyl methacrylate) diblock copolymer.,L. Cano,"As-synthesized organic-capped TiO2 nanorods were incorporated into polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) diblock copolymer to achieve TiO2/PS-b-PMMA nanocomposites with enhanced optical and conductive properties. The specific surface chemistry of TiO2 nanorods derived from the colloidal synthetic approach allowed their prompt incorporation in the PS-b-PMMA block copolymer template up to 50 wt %, which resulted in films with an extended coverage of highly dispersed nanoparticles for contents higher than 30 wt %. At such high nanorod contents, the films fabricated by the prepared nanocomposites demonstrated enhanced optical properties. Atomic force microscopy investigation of the nanocomposite films showed a cylindrical morphology for low nanorod contents. Conversely, higher nanorod contents resulted upon removal of the organic component in the nanocomposites with UV treatment in overall nanorod coverage of the film surface with the concomitant formation of charge percolation paths, which led to noticeable conductivity values. EFM and PF-TUNA measurements confirmed the conductive properties of the composites at nanoscale, whereas semiconductor analyzer measurements provided their macroscale characterization. In addition, an increase in the UV-vis absorption was observed with the increase in the nanorod content along with a remarkable conductivity of the overall film.","{'model': 'tldr@v2.0.0', 'text': 'The specific surface chemistry of TiO2 nanorods derived from the colloidal synthetic approach allowed their prompt incorporation in the PS-b-PMMA block copolymer template up to 50 wt %, which resulted in films with an extended coverage of highly dispersed nanoparticles for contentsHigher nanorod contents led to noticeable conductivity values.'}",
-use of a nursing checklist to facilitate implementation of therapeutic hypothermia after cardiac arrest.,K. Avery,"Therapeutic hypothermia has become a widely accepted intervention that is improving neurological outcomes following return of spontaneous circulation after cardiac arrest. This intervention is highly complex but infrequently used, and prompt implementation of the many steps involved, especially achieving the target body temperature, can be difficult. A checklist was introduced to guide nurses in implementing the therapeutic hypothermia protocol during the different phases of the intervention (initiation, maintenance, rewarming, and normothermia) in an intensive care unit. An interprofessional committee began by developing the protocol, a template for an order set, and a shivering algorithm. At first, implementation of the protocol was inconsistent, and a lack of clarity and urgency in managing patients during the different phases of the protocol was apparent. The nursing checklist has provided all of the intensive care nurses with an easy-to-follow reference to facilitate compliance with the required steps in the protocol for therapeutic hypothermia. Observations of practice and feedback from nursing staff in all units confirm the utility of the checklist. Use of the checklist has helped reduce the time from admission to the unit to reaching the target temperature and the time from admission to continuous electroencephalographic monitoring in the cardiac intensive care unit. Evaluation of patients' outcomes as related to compliance with the protocol interventions is ongoing.","{'model': 'tldr@v2.0.0', 'text': 'A checklist was introduced to guide nurses in implementing the therapeutic hypothermia protocol during the different phases of the intervention (initiation, maintenance, rewarming, and normothermia) in an intensive care unit and has provided all of the intensive care nurses with an easy-to-follow reference to facilitate compliance with the required steps in the protocol.'}",http://ccn.aacnjournals.org/content/35/1/29.full.pdf
-measurement of differential cross sections for the production of a pair of isolated photons in pp collisions at \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength,S. Chatrchyan,,,https://link.springer.com/content/pdf/10.1140/epjc/s10052-014-3129-3.pdf
-a successful competency-based writing skills development programme: results of an experiment1,R. Craig,"We describe a successful, semester-length writing skills development programme conducted at the University of Toronto Mississauga (UTM) with intermediate level undergraduate financial accounting students. The programme focused on improving students' writing in five competency areas: organization, grammar, style, professional writing and case writing. Effectiveness was assessed by means of a quasi-experiment involving a pre-test, post-test design with a treatment group and a control group. Comparison of entry and exit test scores of students in the treatment group with those in the control group revealed a highly statistically significant increase in the writing skills of the treatment group. The programme outlined is not presented as a template for emulation, but rather to prompt educators to ponder how they might address the challenge of improving (and assessing) the writing skills of students.",,
-see yourself write: a simple student model to make students think,S. Bull,,"{'model': 'tldr@v2.0.0', 'text': 'This paper introduces See Yourself Write, a system with two components: a template for tutors to provide feedback to their students on their foreign language writing, and an inspectable student model which is automatically constructed for each student, based on the feedback given by the teacher.'}",
-life defined,Andrew Moore,"I have to admit, I’ve been waiting for a particular event to write this editorial. My prompt is the recent opinion piece ‘‘What is life?’’ [1] published in The Scientist, which authoritatively represents the problems of, and approaches to, defining life, ultimately settling on self-replication as its essence: self-replicating RNA might even be considered a primitive lifeform. Self-replication is a feature of life, alright, but I fundamentally differ that it defines life in terms of being its essence. Inanimate structures in nature can replicate themselves on a template too: minute fragments of a crystal can act as seeds for other crystals to grow. And although the complexity of a crystal structure is nothing approaching that of a polynucleotide, the physical principle is similar. Even the property of catalysis in a self-replicating RNA witnesses a physicochemical principle that is widespread in the inanimate realm. In order to see what is truly ‘‘special’’ about life – in contrast to non-living nature – I believe there is something more important than replication that we need to be thinking about: the way in which living organisms use energy gradients. Life, I would argue, is a combination of the ability to harness energy gradients to produce a local reduction in entropy – i.e. local increase in order/ organisational complexity of matter – coupled with the ability to replicate chemically coded information and reproduce (not replicate) the housing of that information (be it a single cell or multi-celled organism) (Fig. 1). But at the heart of it all is energy metabolism. Energy can only be harnessed via a physical structure that uses an energy gradient – in tangible terms, the natural flux of hot > cold, culminating in everything being ‘‘warm’’; for biological organisms it is potential energy in the form of an ion gradient, chemical bonds or oxidation state. The newest theories on the origins of life indeed argue that it all began with the exploitation of a naturally occurring physicochemical (energy) gradient [2, 3] by pre-biotic membranes. Thermodynamics play a role in another crucially special feature of life: thermodynamic renewal via reproduction. Renewal is only possible as a result of the subversion of energy fluxes that life performs – i.e. to produce a highly complex physicochemical information system that, despite becoming ‘‘degraded’’ via ageing, is able to produce a pristine new organism to carry on the procession of life [4]. Upon reproduction, degraded components are sequestered into the ‘‘mother’’ organism (even in the case of bacterial cells). Here, again we find a force working against the ‘‘inanimate’’ flux of energy and matter towards equilibrium: a force derived from life’s use of energy gradients. Renewal via reproduction is why life exploded and diversified on Earth, creating new order as Earth and the rest of the universe continue their predictable journey towards equilibrium – spectacular though this inanimate physical theatre is. Coupled with replication of the genetic code, renewal-via-reproduction imparts to life its additional characteristic of evolution. Life essentially manifests a physical principle that is not seen elsewhere: the ability to preserve itself (often down to the physical form of an organism) whilst all around it equilibrates. Life has ‘‘wandered’’ into a realm of physics that the inanimate universe does not ‘‘use’’. Every day until we die, our living organism is – in a sense – playing a game with the second law of thermodynamics, channeling the flux of energy toward maintaining complexity and order in contrast to the inexorable process of equilibration surrounding it in the inorganic environment: a game that no non-living system can play. It is also no accident – in fact it might even be considered a tautology to state – that humans employ this principle in the production of machines and engines that similarly precipitate, and then use, a thermodynamic flow from hot to cold to do useful work: constructing local order (decrease in entropy) that","{'model': 'tldr@v2.0.0', 'text': 'This editorial argues that there is something more important than replication that the authors need to be thinking about: the way in which living organisms use energy gradients, andThermodynamics play a role in another crucially special feature of life: thermodynamic renewal via reproduction.'}",https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/bies.201290011
-stable and noncompetitive rna internal control for routine clinical diagnostic reverse transcription-pcr,K. Dingle,"ABSTRACT Clinical diagnostic tests based on nucleic acid amplification assist with the prompt diagnosis of microbial infections because of their speeds and extremely low limits of detection. However, the design of appropriate internal controls for such assays has proven difficult. We describe a reaction-specific RNA internal control for diagnostic reverse transcription (RT)-PCR which allows extraction, RT, amplification, and detection to be monitored. The control consists of a G+C-rich (60%) RNA molecule with an extensive secondary structure, based on a modified hepatitis delta virus genome. The rod-like structure of this RNA, with 70% intramolecular base pairing, provides a difficult template for RT-PCR. This ensures that the more favorable target virus amplicon is generated in preference to the control, with the control being detected only if the target virus is absent. The unusual structure of hepatitis delta virus RNA has previously been shown to enhance its stability and resistance to nucleases, an advantage for routine use as an internal control. The control was implemented in three nested multiplex RT-PCRs to detect nine clinically important respiratory viruses: (i) influenza A and B viruses, (ii) respiratory syncytial viruses A and B and human metapneumovirus, and (iii) parainfluenza virus types 1 to 4. The detection limits of these assays were not detectably compromised by the presence of the RNA control. During routine testing of 324 consecutive unselected respiratory samples, the presence of the internal control ensured that genuine and false-negative results were distinguishable, thus increasing the diagnostic confidence in the assay.","{'model': 'tldr@v2.0.0', 'text': 'A reaction-specific RNA internal control for diagnostic reverse transcription (RT)-PCR which allows extraction, RT, amplification, and detection to be monitored, and ensures that genuine and false-negative results were distinguishable, thus increasing the diagnostic confidence in the assay.'}",https://europepmc.org/articles/pmc356891?pdf=render
-a single-center retrospective analysis of interventions provided to geriatric inpatients receiving end-of-life care,T. Tan,"This study is a single-center retrospective analysis of interventions provided to 50 geriatric inpatients (age > 70 years) receiving end-of-life (EOL) care. Data including frequency of medical and non-medical interventions provided to the patient at the EOL and the different symptoms experienced as well as the outcome of pharmacological interventions provided for these symptoms were collected. The analysis highlights the relatively low rates of anticipatory prescribing of non-opioid EOL medications and the under treatment of multiple symptoms experienced at the EOL – pain, breathlessness, and agitation. Presently there is no formalized or uniform process in EOL prescribing on the medical or surgical wards of the hospital. Use of an ‘anticipatory prescribing’ template that can be modified to suit individual cases may be useful, as well as inclusion of a prompt for routine anticipatory EOL prescribing for a patient who is on the EOL pathway. Education of nurses and doctors on prompt symptom recognition and treatment, as well as the importance of addressing spiritual, religious and social needs may be useful in improving the care of elderly inpatients receiving EOL care in the hospital setting.","{'model': 'tldr@v2.0.0', 'text': 'Education of nurses and doctors on prompt symptom recognition and treatment, as well as the importance of addressing spiritual, religious and social needs may be useful in improving the care of elderly inpatients receiving EOL care in the hospital setting.'}",
-comparing emission mitigation effort,Joseph E. Aldy,"A natural outcome of the emerging pledge and review approach to international climate change policy is interest in comparing mitigation effort among countries. Domestic publics and stakeholders will have an interest in knowing if peer countries are undertaking (or planning to undertake) comparable effort in mitigating their greenhouse gas emissions. Moreover, if considered inadequate to address the risks posed by climate change, the aggregate effort will likely prompt broader interest in identifying those countries where greater effort is arguably warranted on the basis of peer comparisons. Both assessments require metrics of effort and comparisons among countries. We propose a framework for such an exercise, drawing from a set of principles for designing and implementing informative metrics. We present a template for organizing metrics on mitigation effort, for both ex-ante and ex-post review. We also provide preliminary assessments of effort along emissions, price, and cost metrics for post–2020 climate policy contributions by China, the European Union, Russia, and the United States. We close with a discussion of the role of academics and civil society in promoting transparency and facilitating the evaluation and comparison of effort.",,
-clinical handover of the trauma and medical patient: a structured approach,Jez Loseby,"Handovers are vital to the continuity of patient care and can influence patient outcome. Several structures exist to facilitate handover delivery but further work is needed to prospectively evaluate them. This article advocates the implementation of a suitable pre-hospital handover template and the introduction of handover training. We believe the introduction of trauma networks into the UK provides the ideal opportunity to prompt hospitals and ambulance services to co-operate to standardise the approach to handover, improve training and carry out further prospective research into the most effective methods of patient handover.","{'model': 'tldr@v2.0.0', 'text': 'The introduction of trauma networks into the UK provides the ideal opportunity to prompt hospitals and ambulance services to co-operate to standardise the approach to handover, improve training and carry out further prospective research into the most effective methods of patient handover.'}",
-early identification of children at risk for critical care: standardizing communication for inter–emergency department transfers,Cyril Sahyoun,"Background Interfacility transfers occur frequently and often involve critically ill patients. Clear communication at the time of patient referral is essential for patient safety. Objectives The objective of this work was to study whether a standardized inter-emergency department (ED) transfer communication template for pediatric patients with respiratory complaints identifies patients that require intensive care unit (ICU) admission. Methods We created a template to structure the communication between referring and receiving providers involved in inter-ED transfers of children with respiratory complaints. The template was designed for use by nonphysicians to prompt specific questions that would trigger notification of the ED attending based on signs of critical illness. The template was retrospectively applied to determine whether it would have properly triggered attending physician notification of a child ultimately requiring ICU admission. Results Of 285 transferred children, 61 (21%) were admitted to an ICU from the receiving ED. The sensitivity of the communication template in predicting the need for ICU admission was 84% (95% confidence interval [CI], 72%–92%), negative predictive value of 95% (95% CI, 90%–97%), specificity of 77% (95% CI, 71%–82%), positive predictive value of 50% (95% CI, 40%–60%). Of the 10 patients admitted to an ICU who were not identified by the tool, none were critically ill upon arrival. Of the individual communication elements, the sensitivity and negative predictive value ranged from 3% to 38% and from 79% to 86%, respectively. Conclusions A standardized communication template for inter-ED transfers can identify children with respiratory complaints who require ICU admission. Next steps include real-time application to judge screening performance compared with current nonstandardized intake protocols.","{'model': 'tldr@v2.0.0', 'text': 'Whether a standardized inter-emergency department transfer communication template for pediatric patients with respiratory complaints identifies patients that require intensive care unit (ICU) admission is studied.'}",
-random word retrieval for automatic story generation,Richard S. Colon,"Over the past forty years, significant research has been done on story/narrative generation in which the computer is the author. Many existing systems generate stories by filling in a template or copying an analogous story (and changing the time, place, etc.) or by prompting the user to provide a start to the story. Very few systems generate variable stories without these techniques. While it is impossible to quantify a human writer's inspiration, we can consider a common exercise that authors perform; namely `writing prompts'. A writing prompt is just a topic or idea around which to start writing. The prompt can simply be a few words, which becomes the basis for a story. In this paper we present story generation from the perspective of how human authors create stories via writing prompts. The system will select a few random words as a prompt, which will form the basic parameters for generating a story. But unlike a human writer, a computer cannot intuitively know the context of a chosen word. Therefore, the Internet (and existing `Concept Knowledge' systems) will be used to find the context for the selected words, thus guiding the story generation process.","{'model': 'tldr@v2.0.0', 'text': ""This paper presents story generation from the perspective of how human authors create stories via writing prompts, using the Internet (and existing `Concept Knowledge' systems) to find the context for the selected words, thus guiding the story generation process.""}",https://scholarworks.bridgeport.edu/xmlui/bitstream/123456789/545/3/FRD_RColon_Story_Gen_Poster_Mar03_2014.pdf
-direct colony polymerase chain reaction for rapid identification of yeasts isolated from blood specimen,Rajyoganandh S. Vijayaraman,"Context: Bloodstream infections (BSIs) caused by yeasts have an increasing frequency due to the growing population of immunosuppressed individuals. Among yeasts, Candida remains the most prevalent species with the increase in the incidence of non-albicans Candida species. Apart from Candida, other yeasts are also involved in causing BSI. High mortality associated with Candida and other yeast infection can be reduced by prompt and appropriate antifungal therapy. Hence, rapid identification and speciation of yeasts isolated from blood play a significant role in the management of the patients. Since conventional methods used for speciation of Candida and other yeasts are laborious, time-consuming and often unclear, rapid and accurate molecular techniques are required. Materials and Methods: Instead of using purified genomic DNA as template for polymerase chain reaction (PCR), we used yeast colony and cell suspensions in water and 0.10M potassium hydroxide as template for PCR. Candida albicans, Trichosporon and Cryptococcus neoformans were used as reference strains. Further, a total of 100 yeast isolates were also tested. All reactions were performed using the universal fungal primers ITS1 and ITS4; the PCR products were then digested with restriction enzyme (Msp1). Results: Direct colony PCR (DCPCR) produced sharp and distinct bands compared to the cell suspensions with the reference strains. All the 100 clinical isolates tested also produced distinct bands. Conclusion: DCPCR approach not only reduces the DNA template preparation time but is also easy, rapid and reduces the cost of PCR.","{'model': 'tldr@v2.0.0', 'text': 'Direct colony PCR (DCPCR) approach not only reduces the DNA template preparation time but is also easy, rapid and reduces the cost of PCR.'}",
-orbital and physical parameters of eclipsing binaries from the asas catalogue -- iii. two new low-mass systems with rapidly evolving spots,K. Hełminiak,"We present the results of our spectroscopic and photometric analysis of two newly discovered low-mass detached eclipsing binaries found in the All-Sky Automated Survey (ASAS) catalogue: ASAS J093814-0104.4 and ASAS J212954-5620.1. Using the GIRAFFE instrument on the 1.9-m Radcliffe telescope at SAAO and the UCLES spectrograph on the 3.9-m Anglo-Australian Telescope, we obtained high-resolution spectra of both objects and derived their radial velocities (RVs) at various orbital phases. The RVs of both objects were measured with the TODCOR technique using synthetic template spectra as references. We also obtained V and I band photometry using the 1.0-m Elizabeth telescope at SAAO and the 0.4-m PROMPT instruments located at the CTIO. The orbital and physical parameters of the systems were derived with PHOEBE and JKTEBOP codes. We compared our results with several sets of widely-used isochrones. Our multi-epoch photometric observations demonstrate that both objects show significant out-of-eclipse modulations, which vary in time. We believe that this effect is caused by stellar spots, which evolve on time scales of tens of days. For this reason, we constructed our models on the basis of photometric observations spanning short time scales (less than a month). Our modeling indicates that (1) ASAS-09 is a main sequence active system with nearly-twin components with masses of M1 = 0.771(33) Msun, M2 = 0.768(21) Msun and radii of R1 = 0.772(12) Rsun and R2 = 0.769(13) Rsun. (2) ASAS-21 is a main sequence active binary with component masses of M1 = 0.833(17) Msun, M2 = 0.703(13) Msun and radii of R1 = 0.845(12) Rsun and R2 = 0.718(17) Rsun. Both systems confirm the characteristic of active low-mass stars, for which the observed radii are larger and the temperatures lower than predicted by evolutionary models. Other parameters agree within errors with the models of main sequence stars.",,https://www.aanda.org/articles/aa/pdf/2011/03/aa15127-10.pdf
-group medical appointments: organization and implementation in the bone marrow transplantation clinic.,K. Meehan,"Group medical appointments provide patients with prompt access to care, greater attention to their psychosocial needs, and increased time with their medical team. Care providers evaluate more patients with similar needs in a shorter period. Eligible patients (between 100 days and 3 years after autologous stem cell transplantation) were contacted to participate. The laboratory and radiographic results of each patient's completed reevaluation were entered into the electronic template office note before their visit. A group medical visit model, called a Physical Shared Medical Appointment (PSMA), was employed because this uses individual patient examinations followed by a group meeting. On the day of the visit, brief physical examinations were performed on each patient. A transplantation physician then met with the group of patients and answered questions. Patient satisfaction surveys were distributed upon completion of the group session. Each PSMA was limited to 10 patients to facilitate patient participation. Questions during the group meeting were general and applicable to all patients. At completion of the 2-hour visit, patient surveys indicated an extremely high level of satisfaction and the preference to attend a future PSMA. Issues discussed during the group meeting were pertinent to all transplant recipients, regardless of diagnosis. The PSMA model allows the patient to spend extended time with their care providers while providing the care providers an opportunity to discuss health issues with numerous patients during 1 appointment. The Dartmouth Transplant PSMA model is expanding to pretransplantation and postallogeneic transplant recipients.","{'model': 'tldr@v2.0.0', 'text': 'The Dartmouth Transplant PSMA model allows the patient to spend extended time with their care providers while providing the care providers an opportunity to discuss health issues with numerous patients during 1 appointment.'}",
-research on linear features change detection based on remote sensing image and vector data,Zuobang Zhang,"To prompt the present situation and utilized values of fundamental geo-information, this paper focuses on a change detection method based on remote sensing image and GIS vector for linear features. Firstly unilateral vector was taken as original value of linear features; then edge points were picked up by pyramid decomposition and multi-scale template matching, and Ziplock Snake method was adopted to further improve the extraction results; finally buffer zone was constructed to distinguish the changed part. This change detection method proves to have higher degree of automation and more precise, so long as the registration of remote sensing image and vector map is accurate.","{'model': 'tldr@v2.0.0', 'text': 'This change detection method proves to have higher degree of automation and more precise, so long as the registration of remote sensing image and vector map is accurate.'}",
-single-chip speech recognition system based on 8051 microcontroller core,Yuanyuan Shi,"This paper describes a single-chip speech recognition system. It contains the speech functions of prompt, playback, speakerdependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc. The chip is designed based on the SOC (System on Chip) philosophy and an 8-bit MCU, RAM, ROM, ADCIDAC, PWM, I/O ports and other peripheral circuits are all embedded in it. Software modules including control/communication, speech coding and speech recognition algorithms are implemented in an 805 1 compatible microcontroller core, resulting in the extremely low cost of the chip. The speech recognition adopts the template matching technique. It recognizes up to 20 phrases with an average length of 1 second and the recognition accuracy reaches more than 95% with the background SNR above IOdB. Speech coding uses Continuous Variablc Slope Deltamodulation (CVSD) algorithm. The bit rate is 16kbitsis.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes a single-chip speech recognition system that contains the speech functions of prompt, playback, speakerdependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc.'}",
-single-chip speech recognition system based on 8051 microcontroller core,Shi Yuanyuan,"This paper describes a single-chip speech recognition system. It contains the speech functions of prompt, playback, speaker-dependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc. The chip is designed based on the SOC (system on chip) philosophy and an 8-bit MCU, RAM, ROM, ADC/DAC, PWM, I/O ports and other peripheral circuits are all embedded in it. Software modules including control/communication, speech coding and speech recognition algorithms are implemented in an 8051 compatible microcontroller core, resulting in the extremely low cost of the chip. The speech recognition adopts the template matching technique. It recognizes up to 20 phrases with an average length of 1 second and the recognition accuracy reaches more than 95% with the background SNR above 10 dB. Speech coding uses continuous variable slope delta modulation (CVSD) algorithm. The bit rate is 16 kbits/s.","{'model': 'tldr@v2.0.0', 'text': 'This paper describes a single-chip speech recognition system that contains the speech functions of prompt, playback, speaker-dependent speech recognition, suitable for the voice activated systems in toys, games, consumer electronics, office devices, etc.'}",
-competitive polymerase chain reaction for quantitating feline immunodeficiency virus load in infected cat tissues.,M. Pistello,"To quantitate FIV provirus copy numbers present in tissue of infected cats, we have applied a competitive polymerase chain reaction (cPCR) recently described for HIV. The method consists in coamplifying a fixed amount of the DNA to be examined with graded copy numbers of a DNA competitor incorporating a short deletion and bearing the same primer recognition sequences. These conditions ensure almost identical thermodynamic and amplification efficiency for both template species but permit a prompt recognition of the two amplification products by gel electrophoresis. Since the amounts of the two amplicons are dependent on relative initial template concentrations, the number of FIV genomes in the sample can be calculated by densitometric analysis of the electrophoretic bands. After validation, the method has been applied to study the provirus loads in the tissues of cats infected with the Pisa-M2 isolate of FIV.","{'model': 'tldr@v2.0.0', 'text': 'A competitive polymerase chain reaction recently described for HIV is applied to quantitate FIV provirus copy numbers present in tissue of infected cats to study the provirus loads in the tissues of cats infected with the Pisa-M2 isolate of FIV.'}",
-a reliable logo and replay detector for sports video,Qiao Huang,"Replay is one of the key cues indicating highlights in sports videos. A replay is usually sandwiched by two identical logos which prompt the start and end of a replay. A logo transition usually contains 10-30 frames, describes a flying or varying object(s). In this paper, a reliable logo and replay detecting approach is proposed. It contains two main stages: first, a logo transition template is unsupervised learned, a key frame (K-frame) and a set of pixels that describes logo object (logo pixels, L-pixels) accurately are also extracted; second, the learned information are used jointly to detect logos and replays in the video. In addition to traditional color analysis, optical flow feature is employed to depict the movement of the logo object(s). Extensive experiments show that the proposed approach can reliably detect logos and replays regardless of the types of sports videos.","{'model': 'tldr@v2.0.0', 'text': 'Extensive experiments show that the proposed approach can reliably detect logos and replays regardless of the types of sports videos.'}",
-a rapid development framework for multilingual spoken dialogue systems,Masahiro Araki,"We propose a new framework for rapid development of multilingual spoken dialogue systems based on a data modeling driven method. The development process begins with the data model definition in the Groovy programming language, with an extension for implementing spoken dialogue systems, such as the annotation of the target task type and the initiative type of the realized dialogue. Since the data model definition can be inherited from the existing semantic Web schema, the developer can choose the necessary properties from the schema via the GUI interface. By applying the data model definition, our framework generates all the necessary components for a spoken dialogue system based on the model-view-controller (MVC) model. In addition, template prompt entries, which the developer uses to specify the contents of the system utterance, are automatically generated. The fulfilled contents are translated to various languages using a translation Web service. The multilingual speech interaction is realized by an HTML5 speech API specifying only the ""lang"" attribute of the HTML element. Therefore, to realize a first prototype of a foreign language spoken dialogue system, the developer does not necessarily need knowledge of the target language.","{'model': 'tldr@v2.0.0', 'text': 'A new framework for rapid development of multilingual spoken dialogue systems based on a data modeling driven method that generates all the necessary components for a spoken dialogue system based on the model-view-controller (MVC) model.'}",
-changes in membrane transport function in g0 and g1 cells,M. Costlow,"Confluent quiescent monolayers of aneuploid and euploid cells in culture can be stimulated to proliferate by appropriate nutritional changes. In confluent monolayers of WI‐38 human diploid fibroblasts the uptake of cycloleucine is increased three hours after these cells are stimulated to proliferate by a change of medium plus 10% serum. No changes in the uptake of cycloleucine are observed in logarithmically‐growing WI‐38 cells exposed to fresh medium plus 10% serum, or in WI‐38 confluent monolayers in which the conditioned medium has been replaced by fresh medium with 0.3% serum (a change that does not cause stimulation of cellular proliferation in WI‐38 cells). In 3T6 cells in the stationary phase stimulated to proliferate by nutritional changes, there is a prompt increase in the uptake of cycloleucine, within one hour after stimulation of cell proliferation. Similar results were obtained with stationary 2RA cells which are SV‐40 transformed WI‐38 fibroblasts. In addition, chromatin template activity which is known to increase in the early stages after stimulation of confluent WI‐38 cells, was unchanged in confluent 3T6 or 2RA cells stimulated to proliferate.","{'model': 'tldr@v2.0.0', 'text': 'In confluent monolayers of WI‐38 human diploid fibroblasts the uptake of cycloleucine is increased three hours after these cells are stimulated to proliferate by a change of medium plus 10% serum, while chromatin template activity was unchanged in confluent 3T6 or 2RA cells stimulated to proliferation.'}",
-isolation and characterization of conditional alleles of bacteriophage t4 genes uvsx and uvsy.,M. A. Conkling,"The bacteriophage T4 uvsW, uvsX and uvsY gene functions are required for wild-type levels of recombination and for normal survival and mutagenesis after treatments with ultraviolet (UV) and ionizing radiations. The ability of uvsX and uvsY mutations to suppress the lethality of gene 49 mutations was used to select temperature-sensitive and amber alleles of these two genes. (uvsW mutations do not suppress gene 49 mutations.) A simple and powerful complementation test was developed to assist in assigning uvs mutations to genes. The amber alleles of uvsX and uvsY behave as simple null alleles, fully suppressing a gene 49 defect, enhancing UV killing and abolishing UV mutagenesis. However, the properties of the ts alleles of uvsX and uvsY demonstrated that suppression of a gene 49 defect, sensitivity to UV-induced inactivation and UV mutability can be partially uncoupled. These results prompt the hypothesis that radiation mutagenesis occurs during DNA chain elongation past template damage within a recombinational intermediate rather than within a conventional replication fork.","{'model': 'tldr@v2.0.0', 'text': 'The properties of the ts alleles of uvsX and uvsY demonstrated that suppression of a gene 49 defect, sensitivity to UV-induced inactivation and UV mutability can be partially uncoupled, prompting the hypothesis that radiation mutagenesis occurs during DNA chain elongation past template damage within a recombinational intermediate rather than within a conventional replication fork.'}",https://academic.oup.com/genetics/article-pdf/107/4/505/34446702/genetics0505.pdf
-recruiting patients and collecting data for an observational study using computerised record pop-up prompts: the prog-res study,R. Hayward,"Background and Aim Engagement of general practitioners (GPs) and recruitment of patients are ever present problems in primary care studies. This paper seeks to demonstrate that electronic prompts represent one method of easing the burden on GPs to recruit individual patients to studies and also provide the opportunity to collect research data during a normal consultation. Methods Older adults consulting for non-inflammatory musculoskeletal pain from five general practices in Cheshire were recruited to a prospective cohort study (the PROG-RES study). Recruitment of patients was aided by a computer prompt during relevant consultations. When triggered by an appropriate Read code, a pop-up template appeared on the consultation screen prompting the GPs to record the answers to seven brief questions. A self-complete questionnaire was mailed to patients who had completed templates by the Keele GP Research Network team and permission was sought to access their medical records. A feasibility study suggested that the potential number of activated templates in the practice within four months would be 636. Results The 44 GPs completed 650 electronic templates during the four-month recruitment period. Almost 40% of recruitment was within four weeks and greater than 95% of recruitment was within 16 weeks. Practices A–D completed electronic templates at a similar rate (1.61–1.86 templates per 1000 patients), although practice E completed templates at a lower frequency (0.76) due to internal difficulties. Completion of individual items ranged from 98% to 83% and completion of all seven questions was recorded in 63% of patients; 4% of patients had three or fewer responses recorded. Conclusion Templates activated by appropriate codes in the GP consultation can facilitate recruitment to observational studies in primary care. It is possible to collect high-quality research data within a normal consultation. This may be a model for use in future studies in primary care.","{'model': 'tldr@v2.0.0', 'text': 'It is demonstrated that electronic prompts activated by appropriate codes in the GP consultation can facilitate recruitment to observational studies in primary care and it is possible to collect high-quality research data within a normal consultation.'}",https://www.cambridge.org/core/services/aop-cambridge-core/content/view/E93B7CE703F600A1423292CF4FD74CAA/S1463423612000047a.pdf/div-class-title-recruiting-patients-and-collecting-data-for-an-observational-study-using-computerised-record-pop-up-prompts-the-prog-res-study-div.pdf
-a web‐based performance assessment system for environmental protection: wepass,S. Cheung,"In response to the increasing public concern over the effects of construction waste on the environment, a performance assessment system has been designed to provide an instant, online assessment of how well a construction site performs environmentally, e.g. measuring the level of pollution generated from a construction site. Through the use of both Internet and database technologies, it is possible to collect and present data that are essential for environmental management online. To achieve this, a set of environmental performance parameters was derived through a postal questionnaire survey and interviews with experts in the field. A conceptual framework was developed, of which the key components were: template, knowledge base, output data and benchmark group. In combination, these factors provide a system that enables speedy assessment and effective management of environmental performance on construction sites, hence prompt decisions can be made and corrective actions can be taken before potential hazards turn into real disasters.",,
-xerte - a user-friendly tool for creating accessible learning objects,S. Ball,,"{'model': 'tldr@v2.0.0', 'text': 'The JISC TechDis template for Xerte brings some of the key aspects of Xerte to a non-technical user interface to enable any teacher or tutor to easily create Accessible Learning Objects.'}",
-radiation interrogation using signature analysis for detection of chemical explosives,W. L. Dunn,"A signature-based radiation technique for detecting chemical explosive is described. Radiation techniques offer the advantage that they can operate at standoff, although rapid detection of explosives at safe standoff distances remains problematic. The technique we describe differs from the major nuclear approaches, which involve imaging (using either photon or neutron interrogation) and/or quantitative analysis (for instance by use of prompt-and inelastic-scatter gamma-ray production based on neutron interrogation). Our technique utilizes both photon and neutron interrogation but avoids imaging or quantitative analysis. The technique, which we call signature-based radiation scanning (SBRS), limits itself to detecting whether a target contains an explosive and does not attempt to characterize the internals of a target further. A template-matching technique is employed, which provides a single figure-of-merit whose value is used to distinguish between safe targets and those containing explosives. Both simulation and experiment have been used to verify the validity of SBRS.",,
-"""automics"": assisted mobile authoring of photostories as souvenirs of a day in the park",Duncan Rowland,"This paper describes a prototype digital economy service that generates ‘photostory’ souvenirs for group day-trips to an amusement park. Location-based triggers prompt visitors to use their mobile phones to capture, share and annotate photos during the day. Selected photos are combined with professional photos from on-ride capture systems and a template is used to generate a bespoke photostory - an “Automic” souvenir. The design of this prototype system was informed by an ethnographic study of park visiting, and this is discussed along with directions for future work.",,
-identificación de especies de leishmania por la técnica de amplificación al azar del adn polimórfico,Lianet Monzote Fidalgo,"INTRODUCTION: leishmaniosis has been regarded by the World Health Organization as one of the most important tropical diseases. It is very difficult to control such parasitosis because there are not vaccines, and therapy is generally toxic and unsatisfactory. It is of vital importance to set prompt diagnostic method along with identification of the parasite in order to select the suitable treatment and to design the most convenient control measures. Recently, the advances in molecular biology have made it possible to characterize Leishmania species by different methods. The random amplified polymorphic DNA technique is a simple method to detect the genetic polymorphic DNA. OBJECTIVE: to standardize the random amplified polymorphic DNA technique for its use in New World Leishmania species typing. METHODS: by using 5 pmol primer concentration, 75 ng of template DNA, 2 mM of magnesium chloride and 2 U of polymerase DNA Taq in 25μL reaction, two reproducible amplification patters were obtained. The optimized random amplified polymorphic DNA technique served to determine the genetic differences among ten reference strains of Leishmania, with 6 sets of randomly designed conventional primers. The UP GMA method-based grouping strategy determined the phylogenetic relation among the studied species. RESULTS: OPA primers -3, 4 and 8 allowed distinguishing the ten reference strains of Leishmania under study. Two well defined genetic groups including species of Leishmania and Viannia subgenres were obtained; these 2 subgenres showed genetic differences. CONCLUSIONS: in this way, our laboratory has the optimized random amplified polymorphic DNA for the identification of Leishmania species.","{'model': 'tldr@v2.0.0', 'text': 'The laboratory has the optimized random amplified polymorphic DNA for the identification of Leishmania species and determined the phylogenetic relation among the studied species.'}",
-effect of drug information request templates on pharmacy student compliance with the modified systematic approach to answering drug information questions,Stacey M. Lavsa,"Background: The modified systematic approach to answering drug information questions is a technique used in drug information practice and in teaching pharmacy students to effectively provide drug information. Drug information request templates were developed to prompt students and other trainees to ask appropriate background questions and perform an effective search. Objective: An evaluation was conducted to determine whether use of drug information templates by fourth-year pharmacy students during their drug information experiential rotation improved compliance with the modified systematic approach. Methods: Fifty documented drug information requests, including 25 prior to template implementation (August 2005–August 2006) and 25 after template implementation (August 2007–August 2008), were randomly selected for evaluation. Each question was evaluated for completeness of background information obtained, categorization and identification of the ultimate question, completeness of references searched, and formulation of a concise response and an evidence-based recommendation. Results: Background information was complete in 16% of pre-template questions and 92% of post-template questions (p < 0.001). Eighty-four percent of pre-template questions and 96% of post-template questions were appropriately categorized (p = 0.349). The requestor's ultimate question was clearly identified in 68% of pre-template questions and 92% of post-template questions (p = 0.074). All necessary references were searched in 36% of pre-template questions and 88% of post-template questions (p < 0.001). A concise response was documented in 80% of pre-template questions and 92% of post-template questions (p = 0.417). In questions determined to require a specific recommendation among the pre-template (n = 20) and post-template groups (n = 14), a clear and evidence-based recommendation was described in 40% (p = 0.038) and 79% (p = 0.038), respectively. Conclusions: Use of drug information request templates improves students' compliance with the modified systematic approach, most notably in obtaining background information and searching necessary references including primary literature.","{'model': 'tldr@v2.0.0', 'text': ""Use of drug information request templates improves students' compliance with the modified systematic approach, most notably in obtaining background information and searching necessary references including primary literature.""}",
-new event-processing design patterns using cep,A. Alves,,"{'model': 'tldr@v2.0.0', 'text': 'A set of new patterns for event processing are documents, describing their problem domain and providing a solution template implemented using CEP, which is both succinct and highly re-usable.'}",http://adcalves.files.wordpress.com/2009/09/new-event-processing-design-patterns-using-cep.pdf
-a novel sports video logo detector based on motion analysis,Hongliang Bai,,"{'model': 'tldr@v2.0.0', 'text': 'A novel automatic logo detection approach is proposed that is automatically learned by dynamic programming and unsupervised clustering, a key frame is also extracted and used jointly to detect logos in sports videos.'}",
diff --git a/example.env b/example.env
new file mode 100644
index 0000000..321686f
--- /dev/null
+++ b/example.env
@@ -0,0 +1,3 @@
+OPENAI_API_KEY=
+HF_TOKEN=
+SEMANTIC_SCHOLAR_API_KEY=
\ No newline at end of file
diff --git a/main.py b/main.py
index 1bfaf5f..2c2d635 100644
--- a/main.py
+++ b/main.py
@@ -3,6 +3,11 @@
config_data.DataFolderPath = "./data"
config_data.DotenvPath = "./.env"
+
+### IF RUNNING EXPERIMENTS MULTIPLE TIMES, PLEASE SET
+### hasDownloadedPapers TO True IN config_data.py or uncomment the following line
+### config_data.hasDownloadedPapers = True
+
if not config_data.hasDownloadedPapers:
collect_papers.collect()
config_data.hasDownloadedPapers = True
@@ -14,6 +19,10 @@
os.makedirs(config_data.DataFolderPath + os.sep + "experiments_output", exist_ok=True)
print("Running experiments...")
for experiment in experiments.experiments:
- experiment.run()
+ try:
+ experiment.run()
+ except Exception as e:
+ print(f"Error running experiment {experiment.__name__}: {e}")
+ continue
print("Experiments completed. See data/experiments_output for output files")
diff --git a/requirements.txt b/requirements.txt
index e0300df..4503eb1 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -12,7 +12,6 @@ sphinx-autobuild
openai
tika
tqdm
-openai
pdfminer.six
load_dotenv
tomotopy
@@ -21,8 +20,6 @@ pytest-dotenv
python-dotenv
networkx
scipy
-networkx
-tomotopy
soup-nuts @ git+https://git@github.com/ahoho/topics.git@dev
black==23.10.1
pytz==2023.3
@@ -35,6 +32,4 @@ anytree==2.12.1
urllib3<2
acl-anthology-py
seaborn
-requests
-tenacity
--e .
\ No newline at end of file
+-e .
diff --git a/setup.cfg b/setup.cfg
index b80d129..d9d8f98 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -31,14 +31,35 @@ install_requires =
requests
matplotlib
soup-nuts @ git+https://git@github.com/ahoho/topics.git@dev
- black
- jellyfish
datasets
huggingface_hub
tika
scipy
networkx
tomotopy
+ pytest
+ jellyfish
+ sphinx
+ renku-sphinx-theme
+ sphinx-autobuild
+ openai
+ tqdm
+ pdfminer.six
+ load_dotenv
+ wordcloud
+ pytest-dotenv
+ python-dotenv
+ black==23.10.1
+ pytz==2023.3
+ regex==2023.8.8
+ Requests==2.31.0
+ sentence_transformers==2.2.2
+ tenacity==8.2.3
+ tiktoken==0.5.1
+ anytree==2.12.1
+ urllib3<2
+ acl-anthology-py
+ seaborn
[options.packages.find]
diff --git a/src/prompt_systematic_review/collect_papers.py b/src/prompt_systematic_review/collect_papers.py
index 3e81f58..d78acaa 100644
--- a/src/prompt_systematic_review/collect_papers.py
+++ b/src/prompt_systematic_review/collect_papers.py
@@ -1,4 +1,4 @@
-from prompt_systematic_review.get_papers.download_arxiv import query_archive
+from prompt_systematic_review.get_papers.download_arxiv import query_arxiv
from prompt_systematic_review.get_papers.download_semantic_scholar import (
query_semantic_scholar,
)
@@ -33,6 +33,14 @@
def downloadPaper(url: str, title: str):
+ """
+ Download a paper given its URL and title.
+
+ :param url: The URL of the paper to download.
+ :type url: str
+ :param title: The title of the paper.
+ :type title: str
+ """
response = requests.get(url)
recurse = 0
while (
@@ -53,8 +61,22 @@ def downloadPaper(url: str, title: str):
def collect():
+ """
+ Collect papers from various sources, deduplicate and filter them, and save them to a CSV file.
+
+ This function performs the following steps:
+ 1. Downloads papers from arXiv, Semantic Scholar, and ACL using the respective query functions.
+ 2. Cleans and deduplicates the downloaded papers.
+ 3. Removes papers that are in the blacklist.
+ 4. Downloads the PDF files of the remaining papers using multithreading.
+ 5. Filters out papers that don't contain the word "prompt" in their content.
+ 6. Performs an automated review of the papers using the GPT-4 model.
+ 7. Combines the human-reviewed and AI-reviewed papers into a final dataset.
+ 8. Removes PDF files of papers that are not in the final dataset.
+ 9. Saves the final dataset to a CSV file named "master_papers.csv".
+ """
# download CSV of arXiv results
- arxiv_df = query_archive(verbose=True)
+ arxiv_df = query_arxiv(verbose=True)
# clean arXiv CSV
arxiv_df["title"] = arxiv_df["title"].apply(lambda x: process_paper_title(x))
arxiv_df["source"] = "arXiv"
@@ -117,7 +139,6 @@ def collect():
deduplicated_df["title"] != filename[:-4]
]
# Add the paper to the new blacklist
- # TODO: this is messed up, results in an array of 80K single characters
new_blacklist += filename[:-4]
except Exception as e:
@@ -128,7 +149,6 @@ def collect():
# PDFRead Error is likely because of corrupted or empty PDF, can be ignored
if str(e) != "EOF marker not found":
print(f"Error processing {filename}: {e}")
- # TODO: there is smtg weird going on here...
# Get a list of all the paper titles in the directory (without the .pdf extension)
paper_titles = [
diff --git a/src/prompt_systematic_review/config_data.py b/src/prompt_systematic_review/config_data.py
index baea0ec..3e18cb0 100644
--- a/src/prompt_systematic_review/config_data.py
+++ b/src/prompt_systematic_review/config_data.py
@@ -7,16 +7,30 @@
def setDownloadedPapers(hasDownloadedPapers):
+ """
+ Set the value of hasDownloadedPapers.
+
+ :param hasDownloadedPapers: The new value for hasDownloadedPapers.
+ :type hasDownloadedPapers: bool
+ """
hasDownloadedPapers = hasDownloadedPapers
def setDataFolderPath(p):
+ """
+ Set the value of DataFolderPath.
+
+ :param p: The new path for DataFolderPath.
+ :type p: str
+ """
DataFolderPath = os.path.abspath(p)
def getDataPath():
- return os.path.abspath(DataFolderPath)
+ """
+ Get the absolute path of DataFolderPath.
-
-def concatPath(filename):
- return os.path.join(DataFolderPath, filename)
+ :return: The absolute path of DataFolderPath.
+ :rtype: str
+ """
+ return os.path.abspath(DataFolderPath)
diff --git a/src/prompt_systematic_review/experiments/__init__.py b/src/prompt_systematic_review/experiments/__init__.py
index e80b2cc..d5e2eef 100644
--- a/src/prompt_systematic_review/experiments/__init__.py
+++ b/src/prompt_systematic_review/experiments/__init__.py
@@ -14,10 +14,12 @@
from . import topicgpt
from . import download_mmlu
from . import graph_internal_references
-from . import graph
+
experiments = [
count_tool_mentions.Experiment,
+ download_mmlu.Experiment,
+ count_models.Experiment,
eval_prompts.Experiment,
evaluate_human_agreement.Experiment,
graph_dataset_citations.Experiment,
@@ -30,5 +32,5 @@
graph_gpt_3_5_benchmarks.Experiment,
run_tomotopy.Experiment,
topicgpt.Experiment,
- count_models.Experiment,
+ graph_internal_references.Experiment,
]
diff --git a/src/prompt_systematic_review/experiments/count_models.py b/src/prompt_systematic_review/experiments/count_models.py
index a801ee6..c340818 100644
--- a/src/prompt_systematic_review/experiments/count_models.py
+++ b/src/prompt_systematic_review/experiments/count_models.py
@@ -10,8 +10,6 @@
"""This script counts the number of papers in our dataset that mention each model.
The script takes one arg, for the path location of the full paper dataset."""
-# TODO improvement: use word embeddings to find similar model names.
-# for now, assume the common model name will appear in the paper
model_names = [
"GPT-3",
"GPT-4",
@@ -50,6 +48,12 @@
def parse_pdf(file_path):
+ """
+ Extract text from a PDF file.
+
+ :param file_path: The path to the PDF file.
+ :return: The extracted text from the PDF file.
+ """
try:
text = extract_text(file_path)
return text
@@ -59,6 +63,12 @@ def parse_pdf(file_path):
def process_file(args):
+ """
+ Process a single file to count model mentions.
+
+ :param args: A tuple containing the folder path and filename.
+ :return: A tuple containing the filename and a dictionary of model mention counts.
+ """
folder_path, filename = args
file_path = os.path.join(folder_path, filename)
if filename.endswith(".pdf"):
@@ -69,6 +79,12 @@ def process_file(args):
def count_model_mentions_parallel(folder_path):
+ """
+ Count model mentions in parallel for all files in a folder.
+
+ :param folder_path: The path to the folder containing the files.
+ :return: A dictionary mapping model names to lists of filenames mentioning the model.
+ """
files = os.listdir(folder_path)
with Pool(cpu_count()) as pool:
# Use imap_unordered for better tqdm compatibility
@@ -88,6 +104,9 @@ def count_model_mentions_parallel(folder_path):
def count_models():
+ """
+ Count model mentions in the papers dataset and save the results to a CSV file.
+ """
masterpaperscsv_file_path = os.path.join(DataFolderPath, "master_papers.csv")
arxiv_papers_df = pd.read_csv(masterpaperscsv_file_path)
paper_ids = set(arxiv_papers_df["paperId"])
@@ -109,3 +128,7 @@ def count_models():
class Experiment:
def run():
count_models()
+
+
+if __name__ == "__main__":
+ count_models()
diff --git a/src/prompt_systematic_review/experiments/count_tool_mentions.py b/src/prompt_systematic_review/experiments/count_tool_mentions.py
index 5fb6411..cfd68eb 100644
--- a/src/prompt_systematic_review/experiments/count_tool_mentions.py
+++ b/src/prompt_systematic_review/experiments/count_tool_mentions.py
@@ -14,9 +14,14 @@
The script takes one arg, for the path location of the full paper dataset."""
-# TODO improvement: use word embeddings to find similar tool names.
-# for now, assume the common tool name will appear in the paper
-def count_tool_mentions(input_folder_path: str, output_file_path: str, tool_lst: list):
+def count_tool_mentions(input_folder_path, output_file_path, tool_lst):
+ """
+ Count tool mentions in the papers dataset and save the results to a CSV file.
+
+ :param input_folder_path: The path to the folder containing the paper files.
+ :param output_file_path: The path to the output CSV file.
+ :param tool_lst: The list of tools to count mentions for.
+ """
tool_counts = defaultdict(list)
# Iterate through all files in the input folder, count tool mentions
@@ -54,7 +59,9 @@ def count_tool_mentions(input_folder_path: str, output_file_path: str, tool_lst:
def run_count_tool_mentions():
- # script portion
+ """
+ Run the tool mention counting process for models, datasets, and frameworks.
+ """
masterpaperscsv_file_path = os.path.join(DataFolderPath, "master_papers.csv")
# get all paper ids from our dataset
@@ -175,3 +182,7 @@ def run_count_tool_mentions():
class Experiment:
def run():
run_count_tool_mentions()
+
+
+if __name__ == "__main__":
+ run_count_tool_mentions()
diff --git a/src/prompt_systematic_review/experiments/download_mmlu.py b/src/prompt_systematic_review/experiments/download_mmlu.py
index 45a143d..8e9d817 100644
--- a/src/prompt_systematic_review/experiments/download_mmlu.py
+++ b/src/prompt_systematic_review/experiments/download_mmlu.py
@@ -5,6 +5,12 @@
def download_and_extract(url, target_path):
+ """
+ Download a file from the given URL and save it to the target path.
+
+ :param url: The URL of the file to download.
+ :param target_path: The path where the downloaded file will be saved.
+ """
# Download the file
response = requests.get(url, stream=True)
if response.status_code == 200:
@@ -13,12 +19,27 @@ def download_and_extract(url, target_path):
def extract_tar(file_path, extract_path):
+ """
+ Extract the contents of a .tar file to the specified extract path.
+
+ :param file_path: The path to the .tar file.
+ :param extract_path: The path where the contents will be extracted.
+ """
# Extract the .tar file
with tarfile.open(file_path) as tar:
tar.extractall(path=extract_path)
def move_and_rename_extracted_contents(extracted_folder, final_folder, new_folder_name):
+ """
+ Move and rename the contents of the extracted folder to the final folder.
+
+ :param extracted_folder: The path to the extracted folder.
+ :param final_folder: The path to the final folder.
+ :param new_folder_name: The new name for the folder.
+ :return: The path to the final folder.
+ :rtype: str
+ """
# Move and rename the contents of the extracted folder
mmlu_folder = os.path.join(final_folder, new_folder_name)
os.makedirs(mmlu_folder, exist_ok=True)
@@ -31,6 +52,9 @@ def move_and_rename_extracted_contents(extracted_folder, final_folder, new_folde
def download_mmlu():
+ """
+ Download the MMLU dataset and extract it to the final data folder.
+ """
# URL of the .tar file
url = "https://people.eecs.berkeley.edu/~hendrycks/data.tar"
@@ -61,3 +85,7 @@ def download_mmlu():
class Experiment:
def run():
download_mmlu()
+
+
+if __name__ == "__main__":
+ download_mmlu()
diff --git a/src/prompt_systematic_review/experiments/eval_prompts.py b/src/prompt_systematic_review/experiments/eval_prompts.py
index d489e9f..e479bbb 100644
--- a/src/prompt_systematic_review/experiments/eval_prompts.py
+++ b/src/prompt_systematic_review/experiments/eval_prompts.py
@@ -1,5 +1,5 @@
"""
-Test a set of prompts against a dataset and return the results. Currently working for GSM-8k. You must add your openAI API key to the key variable below.
+Test a set of prompts against a dataset and return the results. Currently working for GSM-8k and MMLU.
"""
from prompt_systematic_review.experiments.benchmarking import evaluate_prompts
@@ -167,6 +167,9 @@
def eval_prompts():
+ """
+ Evaluate a set of prompts against a dataset and save the results to a JSON file.
+ """
eval = evaluate_prompts(
prompts_to_test,
dataset,
@@ -202,3 +205,7 @@ def eval_prompts():
class Experiment:
def run():
eval_prompts()
+
+
+if __name__ == "__main__":
+ eval_prompts()
diff --git a/src/prompt_systematic_review/experiments/evaluate_human_agreement.py b/src/prompt_systematic_review/experiments/evaluate_human_agreement.py
index 35971ad..88f27c0 100644
--- a/src/prompt_systematic_review/experiments/evaluate_human_agreement.py
+++ b/src/prompt_systematic_review/experiments/evaluate_human_agreement.py
@@ -5,21 +5,36 @@
import openai
import tqdm
from prompt_systematic_review.utils.utils import process_paper_title
-
from prompt_systematic_review.config_data import DataFolderPath, DotenvPath
-load_dotenv(dotenv_path=DotenvPath) # load all entries from .env file
+# Load environment variables from the .env file
+load_dotenv(dotenv_path=DotenvPath)
+# Set the OpenAI API key from the environment variable
openai.api_key = os.getenv("OPENAI_API_KEY")
-def evaluate_human_agreement(inputFile="arxiv_papers_with_abstract.csv"):
- df = pd.read_csv(os.path.join(DataFolderPath, inputFile))
- # Empty list to keep track of results
+def evaluate_human_agreement(input_file="arxiv_papers_with_abstract.csv"):
+ """
+ Evaluate the agreement between AI predictions and human reviews on a dataset.
+
+ This function reads a dataset from a CSV file, processes each paper's title and abstract using the
+ `review_abstract_title_categorical` function, and compares the AI predictions with human reviews.
+
+ The results are saved to a new CSV file, and metrics such as precision, recall, accuracy, and F1 score
+ are computed and saved to a text file.
+
+ :param input_file: The name of the input CSV file containing the dataset. Defaults to "arxiv_papers_with_abstract.csv".
+ :type input_file: str
+ """
+ # Read input data
+ df = pd.read_csv(os.path.join(DataFolderPath, input_file))
+
+ # Initialize an empty list to keep track of results
results = []
# Iterate over DataFrame row by row
- for index, row in tqdm.tqdm(df.iterrows()):
+ for index, row in tqdm.tqdm(df.iterrows(), total=df.shape[0]):
# Apply function to each paper's title and abstract
result = review_abstract_title_categorical(
title=row["title"],
@@ -29,21 +44,26 @@ def evaluate_human_agreement(inputFile="arxiv_papers_with_abstract.csv"):
# Add result to list
results.append(result)
+ # Add results to DataFrame
for i, result in enumerate(results):
df.loc[i, "Probability"] = result["Probability"]
df.loc[i, "Reasoning"] = result["Reasoning"]
+ # Save AI labels to a new CSV file
df.to_csv(
os.path.join(
DataFolderPath,
"experiments_output" + os.sep + "arxiv_papers_with_ai_labels.csv",
- )
+ ),
+ index=False,
)
+
+ # Read blacklist data
blacklist = pd.read_csv(os.path.join(DataFolderPath, "blacklist.csv"))
blacklist["title"] = blacklist["title"].apply(lambda x: process_paper_title(x))
df["title"] = df["title"].apply(lambda x: process_paper_title(x))
- # df = df.iloc[400:800]
+ # Filter DataFrame for comparison
df_limited = df.copy().iloc[200:]
df_limited["human_review"] = ~df_limited["title"].isin(blacklist["title"])
keepables = ["highly relevant", "somewhat relevant", "neutral"]
@@ -51,9 +71,8 @@ def evaluate_human_agreement(inputFile="arxiv_papers_with_abstract.csv"):
df_limited["AI_keep"] = df_limited["Probability"].map(
lambda x: True if x in keepables else False
)
- num_same_rows = (df_limited["AI_keep"] == df_limited["human_review"]).sum()
- num_same_rows / len(df_limited["human_review"])
+ # Calculate agreement metrics
agreement_grid = pd.crosstab(df_limited["AI_keep"], df_limited["human_review"])
true_positives = agreement_grid.loc[True, True]
@@ -64,14 +83,31 @@ def evaluate_human_agreement(inputFile="arxiv_papers_with_abstract.csv"):
accuracy = (true_positives + true_negatives) / len(df_limited)
precision = true_positives / (true_positives + false_positives)
recall = true_positives / (true_positives + false_negatives)
-
f1_score = 2 * (precision * recall) / (precision + recall)
+
+ # Print metrics
print(f"Precision: {precision}")
print(f"Recall: {recall}")
print(f"Accuracy: {accuracy}")
print(f"F1 Score: {f1_score}")
+ # Write metrics to a text file
+ with open(
+ os.path.join(
+ DataFolderPath, "experiments_output" + os.sep + "agreement_metrics.txt"
+ ),
+ "w",
+ ) as f:
+ f.write(f"Precision: {precision}\n")
+ f.write(f"Recall: {recall}\n")
+ f.write(f"Accuracy: {accuracy}\n")
+ f.write(f"F1 Score: {f1_score}\n")
+
class Experiment:
def run():
evaluate_human_agreement()
+
+
+if __name__ == "__main__":
+ evaluate_human_agreement()
diff --git a/src/prompt_systematic_review/experiments/graph.py b/src/prompt_systematic_review/experiments/graph.py
deleted file mode 100644
index cf4b6ff..0000000
--- a/src/prompt_systematic_review/experiments/graph.py
+++ /dev/null
@@ -1,222 +0,0 @@
-import os
-import requests
-import time
-import json
-from tqdm import tqdm
-import pandas as pd
-from dotenv import load_dotenv
-import csv
-import networkx as nx
-import matplotlib.pyplot as plt
-import textwrap
-from prompt_systematic_review.config_data import DataFolderPath
-
-
-class SemanticScholarAPI:
- def __init__(self, api_key):
- self.api_key = api_key
-
- def rate_limited_request(self, url, headers, max_retries=3, delay=0.1):
- retries = 0
- while retries < max_retries:
- try:
- time.sleep(delay)
- response = requests.get(url, headers=headers)
- response.raise_for_status()
- return response
- except requests.exceptions.HTTPError as e:
- if response.status_code in [429, 504]:
- print(f"Received status code {response.status_code}. Retrying...")
- delay *= 2 # Exponential backoff
- retries += 1
- else:
- print(f"HTTPError: {e}")
- break
- except requests.RequestException as e:
- print(f"Request error: {e}")
- break
- return None
-
- def get_references(self, paper_id):
- url = f"https://api.semanticscholar.org/graph/v1/paper/{paper_id}/references?fields=title,authors&limit=1000"
- headers = {"x-api-key": self.api_key}
- response = self.rate_limited_request(url, headers)
- if response and response.status_code == 200:
- data = response.json()
- reference_ids = [
- ref["citedPaper"]["paperId"]
- for ref in data["data"]
- if "citedPaper" in ref
- ]
- return reference_ids
- return []
-
-
-class PaperProcessor:
- def __init__(self, api_key):
- self.semantic_scholar_api = SemanticScholarAPI(api_key)
-
- def process_papers(self, csv_file_path):
- paper_references = {}
- if os.path.exists(csv_file_path):
- with open(csv_file_path, mode="r", encoding="utf-8") as csvfile:
- csv_reader = csv.DictReader(csvfile, delimiter=",")
- all_papers = list(csv_reader)
- for row in tqdm(all_papers, desc="Processing Papers"):
- paper_id = row.get("paperId", "").strip()
- references = self.semantic_scholar_api.get_references(paper_id)
- paper_references[paper_id] = references
- else:
- print(f"CSV file does not exist: {csv_file_path}")
- return paper_references
-
-
-class GraphVisualizer:
- def visualize_citation_counts(self, paper_references, technique_to_title):
- citation_counts = {tech: 0 for tech in technique_to_title.values()}
- for paper_id, refs in paper_references.items():
- for ref_id in refs:
- if ref_id in paper_references:
- citation_counts[technique_to_title.get(ref_id, "Unknown")] += 1
- techniques, counts = zip(*citation_counts.items())
- plt.figure(figsize=(10, 5))
- plt.bar(techniques, counts, color="blue")
- plt.xlabel("Techniques")
- plt.ylabel("Citation Counts")
- plt.title("Citation Counts by Technique")
- plt.xticks(rotation=45)
- plt.tight_layout()
- plt.savefig(
- os.path.join(
- DataFolderPath, "experiments_output" + os.sep + "paper_graph.pdf"
- ),
- format="pdf",
- bbox_inches="tight",
- )
-
-
-class Main:
- def __init__(self):
- load_dotenv()
- self.api_key = os.getenv("SEMANTIC_SCHOLAR_API_KEY")
- self.paper_processor = PaperProcessor(self.api_key)
- self.graph_visualizer = GraphVisualizer()
-
- def run(self, csv_file_path, technique_to_title):
- paper_references = self.paper_processor.process_papers(csv_file_path)
- self.graph_visualizer.visualize_citation_counts(
- paper_references, technique_to_title
- )
-
-
-def run_graph():
- main = Main()
- titles = [
- "Bounding the Capabilities of Large Language Models in Open Text Generation with Prompt Constraints",
- "Language Models are Few-Shot Learners",
- "A Survey on In-context Learning",
- "What Makes Good In-Context Examples for GPT-3?",
- "Finding Support Examples for In-Context Learning",
- "Unified Demonstration Retriever for In-Context Learning",
- "Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity",
- "Reordering Examples Helps during Priming-based Few-Shot Learning",
- "Learning To Retrieve Prompts for In-Context Learning",
- "Self-Generated In-Context Learning: Leveraging Auto-regressive Language Models as a Demonstration Generator",
- "Large Language Models are Zero-Shot Reasoners",
- "Large Language Models Are Human-Level Prompt Engineers",
- "Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models",
- "Thread of Thought Unraveling Chaotic Contexts",
- "When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgment",
- "Automatic Chain of Thought Prompting in Large Language Models",
- "True Detective: A Deep Abductive Reasoning Benchmark Undoable for GPT-3 and Challenging for GPT-4",
- "Contrastive Chain-of-Thought Prompting",
- "Gemini: A Family of Highly Capable Multimodal Models",
- "Complexity-Based Prompting for Multi-Step Reasoning",
- "Active Prompting with Chain-of-Thought for Large Language Models",
- "MoT: Memory-of-Thought Enables ChatGPT to Self-Improve",
- "Measuring and Narrowing the Compositionality Gap in Language Models",
- "Automatic Prompt Augmentation and Selection with Chain-of-Thought from Labeled Data",
- "Tab-CoT: Zero-shot Tabular Chain of Thought",
- "Is a Question Decomposition Unit All We Need?",
- "Least-to-Most Prompting Enables Complex Reasoning in Large Language Models",
- "Decomposed Prompting: A Modular Approach for Solving Complex Tasks",
- "Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models",
- "Tree of Thoughts: Deliberate Problem Solving with Large Language Models",
- "Large Language Model Guided Tree-of-Thought",
- "Cumulative Reasoning with Large Language Models",
- "Graph of thoughts: Solving elaborate problems with large language models",
- "Recursion of Thought: A Divide-and-Conquer Approach to Multi-Context Reasoning with Language Models",
- "Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks",
- "Faithful Chain-of-Thought Reasoning",
- "Skeleton-of-Thought: Large Language Models Can Do Parallel Decoding",
- "Exploring Demonstration Ensembling for In-context Learning",
- "$k$NN Prompting: Beyond-Context Learning with Calibration-Free Nearest Neighbor Inference",
- "An Information-theoretic Approach to Prompt Engineering Without Ground Truth Labels",
- "Self-Consistency Improves Chain of Thought Reasoning in Language Models",
- "Universal Self-Consistency for Large Language Model Generation",
- "Making Language Models Better Reasoners with Step-Aware Verifier",
- "Language Models (Mostly) Know What They Know",
- "Self-Refine: Iterative Refinement with Self-Feedback",
- "RCOT: Detecting and Rectifying Factual Inconsistency in Reasoning by Reversing Chain-of-Thought",
- "Large Language Models are Better Reasoners with Self-Verification",
- "Deductive Verification of Chain-of-Thought Reasoning",
- "Chain-of-Verification Reduces Hallucination in Large Language Models",
- "Maieutic Prompting: Logically Consistent Reasoning with Recursive Explanations",
- "Large Language Models Understand and Can be Enhanced by Emotional Stimuli",
- "Re-Reading Improves Reasoning in Language Models",
- "Think Twice: Perspective-Taking Improves Large Language Models' Theory-of-Mind Capabilities",
- "Better Zero-Shot Reasoning with Self-Adaptive Prompting",
- "Universal Self-Adaptive Prompting",
- "System 2 Attention (is something you might need too)",
- "Large Language Models as Optimizers",
- "Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselves",
- ]
-
- technique_to_title = {
- "Language Models are Few-Shot Learners": "Few-Shot Learning",
- "A Survey on In-context Learning": "In-context Learning Survey",
- "Exploring Demonstration Ensembling for In-context Learning": "Demonstration Ensembling",
- "Unified Demonstration Retriever for In-Context Learning": "Unified Demo Retriever",
- "Finding Support Examples for In-Context Learning": "Support Examples",
- "Large Language Models Are Human-Level Prompt Engineers": "Human-Level Prompting",
- "Measuring and Narrowing the Compositionality Gap in Language Models": "Compositionality Gap",
- "Automatic Chain of Thought Prompting in Large Language Models": "Automatic CoT",
- "Complexity-Based Prompting for Multi-Step Reasoning": "Complexity-Based Prompting",
- "Self-Generated In-Context Learning: Leveraging Auto-regressive Language Models as a Demonstration Generator": "Self-Generated ICL",
- "Least-to-Most Prompting Enables Complex Reasoning in Large Language Models": "Least-to-Most Prompting",
- "Learning To Retrieve Prompts for In-Context Learning": "Prompt Retrieval",
- "Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity": "Prompt Order Sensitivity",
- "What Makes Good In-Context Examples for GPT-3?": "Good In-Context Examples",
- "MoT: Memory-of-Thought Enables ChatGPT to Self-Improve": "Memory-of-Thought",
- "kNN Prompting: Beyond-Context Learning with Calibration-Free Nearest Neighbor Inference": "kNN Prompting",
- "Large Language Models are Zero-Shot Reasoners": "Zero-Shot Reasoning",
- "Self-Consistency Improves Chain of Thought Reasoning in Language Models": "Self-Consistency",
- "Large Language Models as Optimizers": "LLMs as Optimizers",
- "Decomposed Prompting: A Modular Approach for Solving Complex Tasks": "Decomposed Prompting",
- "Is a Question Decomposition Unit All We Need?": "Question Decomposition",
- "Deductive Verification of Chain-of-Thought Reasoning": "Deductive Verification",
- "Active Prompting with Chain-of-Thought for Large Language Models": "Active Prompting",
- "Large Language Model Guided Tree-of-Thought": "LLM Guided ToT",
- "Language Models (Mostly) Know What They Know": "LLM Self-Knowledge",
- "Automatic Prompt Augmentation and Selection with Chain-of-Thought from Labeled Data": "Automatic Prompt Augmentation",
- "Maieutic Prompting: Logically Consistent Reasoning with Recursive Explanations": "Maieutic Prompting",
- "Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models": "Plan-and-Solve Prompting",
- "Tree of Thoughts: Deliberate Problem Solving with Large Language Models": "Tree of Thoughts",
- "Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks": "Program of Thoughts",
- "Self-Refine: Iterative Refinement with Self-Feedback": "Self-Refine",
- "Cumulative Reasoning with Large Language Models": "Cumulative Reasoning",
- "Faithful Chain-of-Thought Reasoning": "Faithful CoT",
- "Making Language Models Better Reasoners with Step-Aware Verifier": "Step-Aware Verification",
- "Graph of Thoughts: Solving Elaborate Problems with Large Language Models": "Graph of Thoughts",
- "Chain-of-Verification Reduces Hallucination in Large Language Models": "Chain-of-Verification",
- "Better Zero-Shot Reasoning with Self-Adaptive Prompting": "Self-Adaptive Prompting",
- "Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselves": "Rephrase and Respond",
- }
-
- csv_file_path = "path_to_your_csv.csv"
- main.run(csv_file_path, technique_to_title)
-
-
-class Experiment:
- def run():
- run_graph()
diff --git a/src/prompt_systematic_review/experiments/graph_dataset_citations.py b/src/prompt_systematic_review/experiments/graph_dataset_citations.py
index bf84e19..5591213 100644
--- a/src/prompt_systematic_review/experiments/graph_dataset_citations.py
+++ b/src/prompt_systematic_review/experiments/graph_dataset_citations.py
@@ -23,6 +23,14 @@
def parse_pdf(file_path):
+ """
+ Extract text from a PDF file.
+
+ :param file_path: Path to the PDF file
+ :type file_path: str
+ :return: Extracted text
+ :rtype: str
+ """
try:
text = extract_text(file_path)
return text
@@ -32,6 +40,14 @@ def parse_pdf(file_path):
def process_file(args):
+ """
+ Process a single file to count mentions of predefined datasets.
+
+ :param args: Tuple containing folder path and filename
+ :type args: tuple
+ :return: Filename and a dictionary of dataset mention counts
+ :rtype: tuple (str, dict)
+ """
folder_path, filename = args
file_path = os.path.join(folder_path, filename)
if filename.endswith(".pdf"):
@@ -44,6 +60,14 @@ def process_file(args):
def count_dataset_mentions_parallel(folder_path):
+ """
+ Count mentions of datasets in all PDF files in a folder using parallel processing.
+
+ :param folder_path: Path to the folder containing PDF files
+ :type folder_path: str
+ :return: Dictionary with datasets as keys and their mention counts as values
+ :rtype: dict
+ """
files = os.listdir(folder_path)
files = [f for f in files if f.endswith(".pdf")]
@@ -62,6 +86,13 @@ def count_dataset_mentions_parallel(folder_path):
def graph_dataset_citations():
+ """
+ Graph the number of citations for each dataset mentioned in the papers.
+
+ Generates a bar graph of dataset mentions, saves the graph as a PDF.
+
+ :return: None
+ """
papers_dataset_path = os.path.join(DataFolderPath, "papers/")
dataset_usage_counts = count_dataset_mentions_parallel(papers_dataset_path)
diff --git a/src/prompt_systematic_review/experiments/graph_gpt_3_5_benchmarks.py b/src/prompt_systematic_review/experiments/graph_gpt_3_5_benchmarks.py
index fdf0006..d547e51 100644
--- a/src/prompt_systematic_review/experiments/graph_gpt_3_5_benchmarks.py
+++ b/src/prompt_systematic_review/experiments/graph_gpt_3_5_benchmarks.py
@@ -4,6 +4,14 @@
def graph_gpt_3_5():
+ """
+ Graph the number of correct answers for GPT-3.5 across various categories.
+
+ Generates a bar chart displaying the number of correct answers for each category
+ and saves the chart as a PDF.
+
+ :return: None
+ """
# Data for plotting
data = {
"math rookie": {"correct": 1230, "total": 2000},
@@ -58,3 +66,7 @@ def graph_gpt_3_5():
class Experiment:
def run():
graph_gpt_3_5()
+
+
+if __name__ == "__main__":
+ graph_gpt_3_5()
diff --git a/src/prompt_systematic_review/experiments/graph_gpt_4_benchmarks200.py b/src/prompt_systematic_review/experiments/graph_gpt_4_benchmarks200.py
index a9ca086..604ed03 100644
--- a/src/prompt_systematic_review/experiments/graph_gpt_4_benchmarks200.py
+++ b/src/prompt_systematic_review/experiments/graph_gpt_4_benchmarks200.py
@@ -4,6 +4,15 @@
def graph_gt_4_benchmarks200():
+ """
+ Function to plot the GPT-4 benchmarks for role prompts and non-role prompts.
+
+ This function creates a bar chart to compare the scores of role prompts and non-role prompts.
+ The data is sorted in descending order and the scores are displayed on top of the bars.
+ The plot is saved as a PDF file in the experiments_output directory.
+
+ :return: None
+ """
# Data
data = {
"Role Prompts": {
@@ -30,14 +39,12 @@ def graph_gt_4_benchmarks200():
"10-Shot Contrastive CoT": 158,
},
}
-
# Preparing data for plotting
categories = ["Role Prompts", "Non-role Prompts"]
colors = ["blue", "green"]
role_prompt_scores = list(data["Role Prompts"].values())
non_role_prompt_scores = list(data["Non-role Prompts"].values())
labels = list(data["Role Prompts"].keys()) + list(data["Non-role Prompts"].keys())
-
# Sorting the data within each category from highest to lowest score
sorted_role_prompts = dict(
sorted(data["Role Prompts"].items(), key=lambda item: item[1], reverse=True)
@@ -45,14 +52,12 @@ def graph_gt_4_benchmarks200():
sorted_non_role_prompts = dict(
sorted(data["Non-role Prompts"].items(), key=lambda item: item[1], reverse=True)
)
-
# Preparing sorted data for plotting
sorted_role_prompt_scores = list(sorted_role_prompts.values())
sorted_non_role_prompt_scores = list(sorted_non_role_prompts.values())
sorted_labels = list(sorted_role_prompts.keys()) + list(
sorted_non_role_prompts.keys()
)
-
# Plotting the sorted data
plt.figure(figsize=(15, 8))
bars1 = plt.bar(
@@ -67,7 +72,6 @@ def graph_gt_4_benchmarks200():
color=colors[1],
label=categories[1],
)
-
# Adding the number above each bar
for bars in [bars1, bars2]:
for bar in bars:
@@ -80,7 +84,6 @@ def graph_gt_4_benchmarks200():
va="bottom",
bbox=dict(facecolor="white", alpha=0.5),
)
-
# Customizing the plot
plt.xlabel("Prompts")
plt.ylabel("Scores")
@@ -88,7 +91,6 @@ def graph_gt_4_benchmarks200():
plt.xticks(range(len(sorted_labels)), sorted_labels, rotation=90)
plt.legend()
plt.tight_layout()
-
plt.savefig(
os.path.join(
DataFolderPath,
@@ -102,3 +104,7 @@ def graph_gt_4_benchmarks200():
class Experiment:
def run():
graph_gt_4_benchmarks200()
+
+
+if __name__ == "__main__":
+ graph_gt_4_benchmarks200()
diff --git a/src/prompt_systematic_review/experiments/graph_internal_references.py b/src/prompt_systematic_review/experiments/graph_internal_references.py
index 2f33ac1..c421e16 100644
--- a/src/prompt_systematic_review/experiments/graph_internal_references.py
+++ b/src/prompt_systematic_review/experiments/graph_internal_references.py
@@ -187,9 +187,9 @@ def process_papers(self, csv_file_path):
arxiv_paper_id
)
else:
- unmatched_papers[row.get("title", "").strip()] = (
- "Source not supported"
- )
+ unmatched_papers[
+ row.get("title", "").strip()
+ ] = "Source not supported"
continue
if paper_id:
@@ -197,9 +197,9 @@ def process_papers(self, csv_file_path):
if references is not None:
paper_references[paper_id] = references
else:
- unmatched_papers[row["title"]] = (
- "No references found or error occurred"
- )
+ unmatched_papers[
+ row["title"]
+ ] = "No references found or error occurred"
else:
print(f"Paper Id Could not be found for: {row}")
else:
@@ -354,8 +354,6 @@ def visualize_citation_counts(self, paper_references, title_to_technique):
)
sorted_techniques, sorted_counts = zip(*sorted_citations)
-
-
plt.figure(figsize=(15, 6))
plt.bar(
sorted_techniques, sorted_counts, color=(45 / 255, 137 / 255, 145 / 255, 1)
@@ -365,7 +363,7 @@ def visualize_citation_counts(self, paper_references, title_to_technique):
ax = plt.gca()
ax.spines["top"].set_visible(True)
ax.spines["right"].set_visible(True)
-
+
plt.ylabel("Counts", fontdict={"fontsize": 14})
plt.xlabel("Prompting Techniques", fontdict={"fontsize": 14})
plt.title("Citation Counts of Prompting Techniques", fontdict={"fontsize": 30})
@@ -549,5 +547,6 @@ class Experiment:
def run():
graph_internal_references()
+
if __name__ == "__main__":
- graph_internal_references()
\ No newline at end of file
+ graph_internal_references()
diff --git a/src/prompt_systematic_review/experiments/graph_tool_mentions.py b/src/prompt_systematic_review/experiments/graph_tool_mentions.py
index 8dd3dc8..a4c5060 100644
--- a/src/prompt_systematic_review/experiments/graph_tool_mentions.py
+++ b/src/prompt_systematic_review/experiments/graph_tool_mentions.py
@@ -11,6 +11,24 @@ def graph_tool_mentions(
datasetCitationCountInput="dataset_citation_counts.csv",
frameworkCitationCountInput="framework_citation_counts.csv",
):
+ """
+ Graph the tool mentions counts generated by scripts/count_tool_mentions.py.
+
+ This function reads the CSV files containing the citation counts for models, datasets, and frameworks,
+ and generates bar graphs visualizing the top 20 most mentioned tools in each category.
+
+ :param modelCitationCountInput: The name of the CSV file containing the model citation counts.
+ Default is "model_citation_counts.csv".
+ :type modelCitationCountInput: str
+ :param datasetCitationCountInput: The name of the CSV file containing the dataset citation counts.
+ Default is "dataset_citation_counts.csv".
+ :type datasetCitationCountInput: str
+ :param frameworkCitationCountInput: The name of the CSV file containing the framework citation counts.
+ Default is "framework_citation_counts.csv".
+ :type frameworkCitationCountInput: str
+ :return: None
+ :rtype: None
+ """
csv_file = os.path.join(
DataFolderPath, "experiments_output" + os.sep + modelCitationCountInput
)
@@ -88,3 +106,7 @@ def graph_tool_mentions(
class Experiment:
def run():
graph_tool_mentions()
+
+
+if __name__ == "__main__":
+ graph_tool_mentions()
diff --git a/src/prompt_systematic_review/experiments/keyword_wordcloud.py b/src/prompt_systematic_review/experiments/keyword_wordcloud.py
index 27d53b4..4d0a233 100644
--- a/src/prompt_systematic_review/experiments/keyword_wordcloud.py
+++ b/src/prompt_systematic_review/experiments/keyword_wordcloud.py
@@ -1,16 +1,21 @@
-"""
-This script generates a wordcloud of the most common
-words in abstracts in the master_papers.csv dataset.
-"""
-
-import matplotlib.pyplot as plt
+import os
import pandas as pd
from wordcloud import WordCloud
-import os
+import matplotlib.pyplot as plt
from prompt_systematic_review.config_data import DataFolderPath
def keyword_wordcloud():
+ """
+ Generate a word cloud from the abstracts of the papers in the master_papers.csv file.
+
+ This function reads the abstracts from the master_papers.csv file, concatenates them into a single
+ string, and generates a word cloud visualization using the WordCloud library. The resulting word
+ cloud is saved as an image file.
+
+ :return: None
+ :rtype: None
+ """
file_path = os.path.join(DataFolderPath, "master_papers.csv")
# Read the CSV file into a DataFrame
@@ -48,3 +53,7 @@ def keyword_wordcloud():
class Experiment:
def run():
keyword_wordcloud()
+
+
+if __name__ == "__main__":
+ keyword_wordcloud()
diff --git a/src/prompt_systematic_review/experiments/papers_over_time.py b/src/prompt_systematic_review/experiments/papers_over_time.py
index bd714b8..bed854a 100644
--- a/src/prompt_systematic_review/experiments/papers_over_time.py
+++ b/src/prompt_systematic_review/experiments/papers_over_time.py
@@ -1,11 +1,24 @@
+import os
+import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
-import pandas as pd
-import os
from prompt_systematic_review.config_data import DataFolderPath
def papers_over_time():
+ """
+ Generate plots showing the number of papers submitted over time.
+
+ This function reads the master_papers.csv file, processes the submission dates, and generates three plots:
+ 1. Number of papers submitted over time, by year.
+ 2. Number of papers submitted between 2021 and 2023 (majority of papers).
+ 3. Number of papers submitted between 2021 and 2023, with vertical lines indicating release dates of different LLMs.
+
+ The plots are saved as image files in the experiments_output directory.
+
+ :return: None
+ :rtype: None
+ """
file_path = os.path.join(DataFolderPath, "master_papers.csv")
arxiv_papers_df = pd.read_csv(file_path)
@@ -146,3 +159,7 @@ def papers_over_time():
class Experiment:
def run():
papers_over_time()
+
+
+if __name__ == "__main__":
+ papers_over_time()
diff --git a/src/prompt_systematic_review/experiments/topicgpt.py b/src/prompt_systematic_review/experiments/topicgpt.py
index 1c39f7a..18d8753 100644
--- a/src/prompt_systematic_review/experiments/topicgpt.py
+++ b/src/prompt_systematic_review/experiments/topicgpt.py
@@ -9,6 +9,16 @@
def run_topic_gpt():
+ """
+ Run the TopicGPT experiment to generate a topic distribution for the papers.
+
+ This function sets up the necessary data and prompt files, runs the generation script,
+ generates a tree structure from the generated topics, filters out topics below a certain threshold,
+ and visualizes the topic distribution using a horizontal bar chart.
+
+ :return: None
+ :rtype: None
+ """
sns.set_style("whitegrid")
plt.rcParams.update({"font.size": 17})
@@ -82,3 +92,7 @@ def run_topic_gpt():
class Experiment:
def run():
run_topic_gpt()
+
+
+if __name__ == "__main__":
+ run_topic_gpt()
diff --git a/src/prompt_systematic_review/experiments/visualize_authors.py b/src/prompt_systematic_review/experiments/visualize_authors.py
index e2e9708..15d4139 100644
--- a/src/prompt_systematic_review/experiments/visualize_authors.py
+++ b/src/prompt_systematic_review/experiments/visualize_authors.py
@@ -12,6 +12,18 @@
def visualize_authors():
+ """
+ Generate visualizations of the publication counts for authors in the master_papers.csv dataset.
+
+ This function reads the master_papers.csv file, processes the authors' data, and generates two visualizations:
+ 1. A vertical bar chart displaying the top 20 authors' publication counts.
+ 2. A histogram of the number of publications per author on a logarithmic scale.
+
+ The visualizations are saved as PDF files in the experiments_output directory.
+
+ :return: None
+ :rtype: None
+ """
file_path = os.path.join(DataFolderPath, "master_papers.csv")
# Read the CSV file into a DataFrame
@@ -88,3 +100,7 @@ def visualize_authors():
class Experiment:
def run():
visualize_authors()
+
+
+if __name__ == "__main__":
+ visualize_authors()
diff --git a/src/prompt_systematic_review/get_papers/acl_source.py b/src/prompt_systematic_review/get_papers/acl_source.py
index 46803e8..1ae93e7 100644
--- a/src/prompt_systematic_review/get_papers/acl_source.py
+++ b/src/prompt_systematic_review/get_papers/acl_source.py
@@ -74,7 +74,7 @@ def getPapers(self, count: int, keyWords: List[str]) -> List[Paper]:
def getPaperSrc(self, paper: Paper, destinationFolder: str = None, recurse=0):
"""
- download a paper.
+ Download a paper given a paper object.
:param paper: The paper to get the download of.
:type paper: Paper
diff --git a/src/prompt_systematic_review/get_papers/arxiv_source.py b/src/prompt_systematic_review/get_papers/arxiv_source.py
index 72a7b3d..53cfd4a 100644
--- a/src/prompt_systematic_review/get_papers/arxiv_source.py
+++ b/src/prompt_systematic_review/get_papers/arxiv_source.py
@@ -93,7 +93,7 @@ def getPapers(self, count: int, keyWords: List[str]) -> List[Paper]:
def getPaperSrc(self, paper: Paper, destinationFolder: str = None, recurse=0):
"""
- download a paper.
+ Download a paper given a paper object.
:param paper: The paper to get the download of.
:type paper: Paper
diff --git a/src/prompt_systematic_review/get_papers/download_acl.py b/src/prompt_systematic_review/get_papers/download_acl.py
index 07e2870..6c5a8ee 100644
--- a/src/prompt_systematic_review/get_papers/download_acl.py
+++ b/src/prompt_systematic_review/get_papers/download_acl.py
@@ -8,7 +8,7 @@
def query_acl(downloadName: str = None, verbose=False):
"""
- Download papers from arxiv and save them to a csv file.
+ Download papers from ACL and save them to a csv file.
:param downloadName: The name of the csv file to save the data to.
"""
diff --git a/src/prompt_systematic_review/get_papers/download_arxiv.py b/src/prompt_systematic_review/get_papers/download_arxiv.py
index f6fadce..681e459 100644
--- a/src/prompt_systematic_review/get_papers/download_arxiv.py
+++ b/src/prompt_systematic_review/get_papers/download_arxiv.py
@@ -6,7 +6,7 @@
import tqdm
-def query_archive(downloadName: str = None, verbose=False):
+def query_arxiv(downloadName: str = None, verbose=False):
"""
Download papers from arxiv and save them to a csv file.
:param downloadName: The name of the csv file to save the data to.
diff --git a/src/prompt_systematic_review/get_papers/download_semantic_scholar.py b/src/prompt_systematic_review/get_papers/download_semantic_scholar.py
index e9ae520..e52bb07 100644
--- a/src/prompt_systematic_review/get_papers/download_semantic_scholar.py
+++ b/src/prompt_systematic_review/get_papers/download_semantic_scholar.py
@@ -9,19 +9,48 @@
def create_directory(directory_name):
- """Create a directory if it doesn't already exist."""
+ """
+ Create a directory if it doesn't already exist.
+
+ :param directory_name: The name of the directory to create.
+ :type directory_name: str
+ """
if not os.path.exists(directory_name):
os.makedirs(directory_name)
def save_papers_to_json(papers, file_path):
- """Save a list of Paper objects to a JSON file."""
+ """
+ Save a list of Paper objects to a JSON file.
+
+ :param papers: A list of Paper objects to save.
+ :type papers: List[Paper]
+ :param file_path: The file path to save the JSON file.
+ :type file_path: str
+ """
papers_dict = [paper.to_dict() for paper in papers]
with open(file_path, "w") as file:
json.dump(papers_dict, file, indent=4)
def query_semantic_scholar(downloadName: str = None, verbose=False):
+ """
+ Query Semantic Scholar for papers based on a list of keywords and save the results to a CSV file.
+
+ This function uses the SemanticScholarSource class to retrieve papers that match the specified keywords.
+ It iterates over the list of keywords and retrieves papers for each keyword using the getPapers method.
+ The retrieved papers are then combined into a single DataFrame.
+
+ If the `downloadName` parameter is provided, the DataFrame is saved to a CSV file with the specified name.
+ The function also returns the combined DataFrame.
+
+ :param downloadName: The name of the CSV file to save the papers to (optional).
+ :type downloadName: str
+ :param verbose: Whether to display progress information using tqdm (default is False).
+ :type verbose: bool
+ :return: A DataFrame containing the retrieved papers.
+ :rtype: pd.DataFrame
+ """
sss = SemanticScholarSource()
all_papers_df = pd.DataFrame()
diff --git a/src/prompt_systematic_review/utils/load_hf_data.py b/src/prompt_systematic_review/utils/load_hf_data.py
deleted file mode 100644
index 98ac16c..0000000
--- a/src/prompt_systematic_review/utils/load_hf_data.py
+++ /dev/null
@@ -1,26 +0,0 @@
-from datasets import load_dataset
-from datasets import Split
-
-
-def load_hf_dataset(dataset_name: str, name: str = None, split: str or Split = None):
- """
- Load in a Hugging Face dataset.
-
- :param dataset_name: The name of the Hugging Face dataset to load.
- :type dataset_name: str
- :param name: Defines the name of the dataset configuration
- :type name: str
- :param split: Which split of the data to load. If None, will return a dict with all splits (typically datasets.Split.TRAIN and datasets.Split.TEST).
- :type split: str or Split
- :return: The loaded dataset.
- :rtype: Dataset or DatasetDict
- """
-
- try:
- return load_dataset(
- dataset_name, name if name else None, split=split if split else None
- )
- except FileNotFoundError:
- print(
- f"The dataset {dataset_name} with config {name} and split {split} is not available on Hugging Face datasets."
- )
diff --git a/src/prompt_systematic_review/utils/pipeline.py b/src/prompt_systematic_review/utils/pipeline.py
deleted file mode 100644
index 1991d84..0000000
--- a/src/prompt_systematic_review/utils/pipeline.py
+++ /dev/null
@@ -1,148 +0,0 @@
-from huggingface_hub import (
- HfFileSystem,
- login,
- HfApi,
- hf_hub_download,
- snapshot_download,
-)
-import pandas as pd
-from io import StringIO
-import os
-
-
-"""
-READ THIS
-https://docs.github.com/en/actions/security-guides/using-secrets-in-github-actions
-https://huggingface.co/docs/huggingface_hub/v0.18.0.rc0/guides/hf_file_system
-
-"""
-
-
-class Pipeline:
- def __init__(
- self,
- revision="main",
- repo="PromptSystematicReview/Prompt_Systematic_Review_Dataset",
- ):
- try:
- self.token = os.getenv("HF_TOKEN")
- self.fs = HfFileSystem(token=self.token)
- self.api = HfApi(token=self.token)
- self.repo = repo
- self.repo_name = repo.split("/")[1]
-
- except:
- raise ValueError("Token not found")
-
- self.root = f"{repo}@{revision}/"
- self.root_without_revision = f"{repo}/"
- self.root_with_type = f"datasets/{self.root}"
- self.root_with_type_without_revision = f"datasets/{self.root_without_revision}"
- self.root_url = f"hf://datasets/{self.root}"
- self.root_url_without_revision = f"hf://datasets/{self.root_without_revision}"
-
- self.revision = revision
-
- def get_revision(self):
- return self.revision
-
- def set_revision(self, revision):
- try:
- assert revision.isalnum()
- self.revision = revision
- self.root = f"{self.repo}@{revision}/"
- self.root_without_revision = f"{self.repo}/"
- self.root_with_type = f"datasets/{self.root}"
- self.root_with_type_without_revision = (
- f"datasets/{self.root_without_revision}"
- )
- self.root_url = f"hf://datasets/{self.root}"
- self.root_url_without_revision = (
- f"hf://datasets/{self.root_without_revision}"
- )
- except:
- raise ValueError("Revision must be alphanumeric")
-
- def get_root_files(self):
- return self.api.list_repo_files(
- self.repo, revision=self.revision, repo_type="dataset"
- )
-
- def read_from_file(self, fileName):
- return pd.read_csv(os.path.join(self.root_url, fileName))
-
- def write_to_file(self, fileName, dataFrame):
- path = os.path.join(self.root_url, fileName)
- print(path)
- dataFrame.to_csv(path, index=False)
-
- def upload_file(self, filePath, folderInRepo=""):
- fileName = os.path.basename(filePath)
- path = os.path.join(self.repo, fileName)
- self.api.upload_file(
- repo_id=self.repo,
- path_in_repo=fileName,
- path_or_fileobj=fileName,
- commit_message=f"Add {fileName}",
- repo_type="dataset",
- revision=self.revision,
- )
-
- def download_file(self, fileName, downloadPath="./"):
- if self.root_without_revision[-1] == "/":
- repoId = self.root_without_revision[:-1]
- else:
- repoId = self.root_without_revision
- hf_hub_download(
- repo_id=repoId,
- filename=fileName,
- repo_type="dataset",
- revision=self.revision,
- local_dir=downloadPath,
- )
-
- def delete_file(self, fileName):
- path = os.path.join(self.repo, fileName)
- self.api.delete_file(
- fileName,
- self.repo,
- token=self.token,
- commit_message=f"Delete {fileName}",
- repo_type="dataset",
- revision=self.revision,
- )
-
- def upload_folder(self, folderPath, folderInRepo=""):
- if folderInRepo != "":
- folderInRepo = folderPath
- self.api.upload_folder(
- repo_id=self.repo,
- folder_path=folderPath,
- path_in_repo=folderInRepo,
- commit_message=f"Add {folderPath}",
- repo_type="dataset",
- revision=self.revision,
- )
-
- def delete_folder(self, folderPath):
- self.api.delete_folder(
- folderPath,
- self.repo,
- token=self.token,
- commit_message=f"Delete {folderPath}",
- repo_type="dataset",
- revision=self.revision,
- )
-
- def download_dataset(self, downloadPath="./"):
- if self.root_without_revision[-1] == "/":
- repoId = self.root_without_revision[:-1]
- else:
- repoId = self.root_without_revision
-
- snapshot_download(
- repo_id=repoId,
- repo_type="dataset",
- revision=self.revision,
- local_dir=downloadPath,
- )
diff --git a/src/prompt_systematic_review/utils/utils.py b/src/prompt_systematic_review/utils/utils.py
index 5aa034c..df24513 100644
--- a/src/prompt_systematic_review/utils/utils.py
+++ b/src/prompt_systematic_review/utils/utils.py
@@ -1,7 +1,6 @@
import requests
from xml.etree import ElementTree as ET
import re
-import prompt_systematic_review.utils.pipeline as p
# custom header so requests don't get blocked
headers = {
@@ -37,27 +36,3 @@ def process_paper_title(title: str) -> str:
# .replace(" ", "")
# s = re.sub(r'\W+', '', s)
return s
-
-
-def auto_pipeline(csvFile, folderPath):
- """
- Uploads both a CSV and a folder to HuggingFace Hub.
- We do both because it doesnt make sense to upload pdfs
- without also uploding the master CSV
-
- :param csvFile: The path to the CSV file.
- :type csvFile: str
- :param folderPath: The path to the folder.
- :type folderPath: str
- """
- pipe = p.Pipeline()
- try:
- pipe.delete_file(csvFile)
- except:
- pass
- try:
- pipe.delete_folder(folderPath)
- except:
- pass
- pipe.upload_file(csvFile)
- pipe.upload_folder(folderPath)
diff --git a/tests/test_load_hf_data.py b/tests/test_load_hf_data.py
deleted file mode 100644
index c16f8ef..0000000
--- a/tests/test_load_hf_data.py
+++ /dev/null
@@ -1,20 +0,0 @@
-from prompt_systematic_review.utils.load_hf_data import load_hf_dataset
-import pytest
-
-
-@pytest.mark.API_test
-def test_arxiv_source():
- # Small dataset for testing
- dataset = load_hf_dataset("rotten_tomatoes")
- assert len(dataset) == 3
- assert len(dataset["train"]) == 8530
- assert len(dataset["validation"]) == 1066
- assert len(dataset["test"]) == 1066
- assert dataset["train"].features["text"].dtype == "string"
- assert dataset["train"].features["label"].dtype == "int64"
- assert dataset["train"].features["label"].num_classes == 2
- assert dataset["train"][0]["label"] == 1
- assert (
- dataset["train"][0]["text"]
- == "the rock is destined to be the 21st century's new \" conan \" and that he's going to make a splash even greater than arnold schwarzenegger , jean-claud van damme or steven segal ."
- )
diff --git a/tests/test_pipeline.py b/tests/test_pipeline.py
deleted file mode 100644
index 58fb8a4..0000000
--- a/tests/test_pipeline.py
+++ /dev/null
@@ -1,156 +0,0 @@
-# test_pipeline.py
-import pytest
-import os
-from prompt_systematic_review.utils.pipeline import *
-from huggingface_hub import delete_file
-import random
-import time
-import hashlib
-import shutil
-
-
-def hashString(bytes):
- return str(hashlib.md5(bytes).hexdigest())
-
-
-@pytest.fixture
-def client():
- return Pipeline(revision="test")
-
-
-@pytest.mark.API_test
-def test_get_root_files(client):
- assert len(client.get_root_files()) > 0
-
-
-@pytest.mark.API_test
-def test_read_from_file(client):
- assert len(client.read_from_file("test.csv")) > 0
- assert len(client.read_from_file("test.csv").columns) == 2
- assert client.read_from_file("test.csv")["Age"].mean() == 21
-
-
-@pytest.mark.API_test
-def test_write_to_file(client):
- lenOfFiles = len(client.get_root_files())
- randString = random.randbytes(100) + str(time.time()).encode()
- randHash = hashString(randString)
- csvDict = {"test": [1, 3], "test2": [2, 4]}
- print(client.revision)
- client.write_to_file(f"{randHash[:10]}_test.csv", pd.DataFrame(csvDict))
- print(client.revision)
- time.sleep(1)
- # assert client.revision == "main"
- df = client.read_from_file(f"{randHash[:10]}_test.csv")
- assert df["test"].sum() == 4
- assert df["test2"].sum() == 6
- # time.sleep(1)
- print(client.root + f"{randHash[:10]}_test.csv")
- delete_file(
- f"{randHash[:10]}_test.csv",
- "PromptSystematicReview/Prompt_Systematic_Review_Dataset",
- repo_type="dataset",
- revision="test",
- )
-
- assert len(client.get_root_files()) == lenOfFiles
-
-
-@pytest.mark.API_test
-def test_upload_file(client):
- # create and populate "test.txt"
- with open("test.txt", "w") as file:
- file.write("Test file")
- # Upload a file
- client.upload_file("test.txt")
- # Check if the file exists in the root files
- assert "test.txt" in [os.path.basename(i) for i in client.get_root_files()]
-
- # Clean up: Delete the uploaded file
- client.delete_file("test.txt")
- print([os.path.basename(i) for i in client.get_root_files()])
- assert "test.txt" not in [os.path.basename(i) for i in client.get_root_files()]
- os.remove("test.txt")
-
-
-@pytest.mark.API_test
-def test_download_file(client):
- # create and populate "test.txt"
- with open("test.txt", "w") as file:
- file.write("Test file")
- # Upload a file
- client.upload_file("test.txt")
-
- os.makedirs("./downloads", exist_ok=True)
- # Download the file
- client.download_file("test.txt", downloadPath="./downloads")
-
- # Check if the file exists in the download path
- assert os.path.exists("./downloads/test.txt")
-
- # Clean up: Delete the uploaded file and the downloaded file
- client.delete_file("test.txt")
- os.remove("./downloads/test.txt")
- os.removedirs("./downloads")
-
-
-@pytest.mark.API_test
-def test_delete_file(client):
- # Upload a file
- with open("test.txt", "w") as file:
- file.write("Test file")
- client.upload_file("test.txt")
-
- # Check if the file exists in the root files
- assert "test.txt" in [os.path.basename(i) for i in client.get_root_files()]
-
- # Delete the file
- client.delete_file("test.txt")
-
- # Check if the file is deleted
- assert "test.txt" not in [os.path.basename(i) for i in client.get_root_files()]
-
- # cleanup
- os.remove("test.txt")
-
-
-@pytest.mark.API_test
-def test_upload_folder(client):
- # Create a temporary folder and files
- os.makedirs("./temp_folder", exist_ok=True)
- with open("./temp_folder/testfile1.txt", "w") as file:
- file.write("Test file 1")
- with open("./temp_folder/testfile2.txt", "w") as file:
- file.write("Test file 2")
-
- # Upload the folder
- client.upload_folder("./temp_folder", folderInRepo="temp_folder")
-
- # Check if the folder exists in the root files
- assert "testfile1.txt" in [os.path.basename(i) for i in client.get_root_files()]
- assert "testfile2.txt" in [os.path.basename(i) for i in client.get_root_files()]
-
- # Clean up: Delete the uploaded folder and files
- client.delete_folder("temp_folder")
- assert "testfile1.txt" not in [os.path.basename(i) for i in client.get_root_files()]
- assert "testfile2.txt" not in [os.path.basename(i) for i in client.get_root_files()]
-
- os.remove("./temp_folder/testfile1.txt")
- os.remove("./temp_folder/testfile2.txt")
- # delete ./temp_folder even if it is not empty
- shutil.rmtree("./temp_folder")
-
-
-@pytest.mark.API_test
-def test_download_dataset(client):
- # Create a temporary folder for downloading the dataset
- os.makedirs("./downloads", exist_ok=True)
-
- # Download the dataset
- client.download_dataset(downloadPath="./downloads")
-
- # Check if the dataset file exists in the download path
- assert len(os.listdir("./downloads")) > 0
-
- # Clean up: Delete the downloaded dataset file
- shutil.rmtree("./downloads")